Initial Contribution

52 files changed, 48654 insertions, 0 deletions

diff --git a/distrib/Makefile b/distrib/Makefile
new file mode 100644
index 0000000..677b0bb
--- /dev/null
+++ b/distrib/Makefile
@@ -0,0 +1,8 @@
+ZLIB_VERSION   := zlib-1.2.3
+LIBPNG_VERSION := libpng-1.2.19
+
+ZLIB_DIR    = $(SRC_PATH)/distrib/$(ZLIB_VERSION)
+LIBPNG_DIR := $(SRC_PATH)/distrib/$(LIBPNG_VERSION)
+
+include $(ZLIB_DIR)/Makefile
+include $(LIBPNG_DIR)/Makefile
diff --git a/distrib/README b/distrib/README
new file mode 100644
index 0000000..8a2cf52
--- /dev/null
+++ b/distrib/README
@@ -0,0 +1,50 @@
+This is source release of the Android emulator. simply run the "build-emulator.sh" script to
+generate a statically linked "emulator" binary in the current directory.
+
+you can also use the "--target=<path>" option to install the executable into a different location,
+
+At the moment, only Linux and Mac OS X are supported.
+
+This emulator is probably not usable without other support files provided by the Android project,
+like a specific kernel image, ramdisk, system and user disk images. Please go to the Android web
+site for more details.
+
+This emulator is licensed under the GNU General Public License (GPL) version 2, which can be
+found in the file "qemu/COPYING".
+
+it is based on QEMU 0.8.2 with many changes used to support the following features:
+
+  - additionnal hardware support for some Android reference boards.
+
+  - various OS-X related patches to make everything compile cleanly with GCC 4.1 and
+    beyond. this includes better support for the Mach-O binary format
+
+  - support for instruction-level profiling and data cache simulation. this allows the
+    emulator to generate "profile" files that can later be analyzed with external tools
+    to provide accurate information about what's happening in the system
+
+  - changes in the dynamic code generators, mainly to support concurrent generators in
+    a single binary (this allows us to use different generators for profiling and
+    non-profiling modes, and switch between them dynamically at runtime when needed)
+
+  - support for network throttling and latency simulation, used to better emulate the
+    network conditions of radio networks.
+
+  - a new graphical user interface capable of displaying and rotating "device skins"
+
+  - an optional (and disabled by default) "polling" runtime mode that doesn't use
+    SIGALRM signals to implement timers. this makes for much better timing accuracy
+    when using "old" emukated Linux kernels, at the cost of using 100% CPU, even when
+    the guest system is idle. This is now disabled since Linux 2.6.21 and beyond use
+    "dynamic ticks" that make this mode un-necessary for Android.
+
+
+it also uses a patched version of LibSDL-1.2.12 which implements the following:
+
+  - prevent a fatal bug in Quartz Extreme's QuickDraw emulation to crash the program
+    whenever SDL_WINDOW_POS is set in the environment before starting the program.
+    the patch implements a simple workaround to this system-level problem.
+
+  - new APIs: SDL_WM_GetPos() and SDL_WM_SetPos() are used to retrieve and set the emulator
+    window position. this allows us to implement a simple-yet-useful feature: the emulator remembers
+    its position among restarts.
diff --git a/distrib/build-emulator.sh b/distrib/build-emulator.sh
new file mode 100755
index 0000000..a904fad
--- /dev/null
+++ b/distrib/build-emulator.sh
@@ -0,0 +1,37 @@
+#!/bin/bash
+#
+#  this script is used to build a static version of the Android emulator
+#  from our distribution package.
+#
+cd $(dirname $0)
+CURDIR=$(pwd)
+
+show_help=
+TARGET=emulator
+for opt; do
+  optarg=`expr "x$opt" : 'x[^=]*=\(.*\)'`
+  case "$opt" in
+  --help|-h|-\?) show_help=yes
+  ;;
+  --target=*) TARGET=$optarg
+  ;;
+  esac
+done
+
+if [ -n "$show_help" ] ; then
+    echo "usage: build-emulator [--target=FILEPATH]"
+    exit 1
+fi
+
+# directory where we'll place the temporary SDL binaries
+LOCAL=$CURDIR/local
+
+cd $CURDIR/sdl
+if ! (./android-configure --prefix=$LOCAL && make && make install); then
+    echo "ERROR: could not build SDL library, please check their sources"
+fi
+
+cd $CURDIR/qemu
+if ! (./android-rebuild.sh --sdl-config=$LOCAL/bin/sdl-config --install=$CURDIR/emulator); then
+    echo "ERROR: could not build SDL library"
+fi
diff --git a/distrib/build_gcc_qemu_darwin.sh b/distrib/build_gcc_qemu_darwin.sh
new file mode 100755
index 0000000..4c82af4
--- /dev/null
+++ b/distrib/build_gcc_qemu_darwin.sh
@@ -0,0 +1,482 @@
+#!/bin/sh
+# APPLE LOCAL file B&I
+
+set -x
+
+# set BUILD_CPLUSPLUS to 'true' if you want to compile support for C++
+BUILD_CPLUSPLUS=false
+
+# set BUILD_DOCS to 'true' to build and install the documentation
+BUILD_DOCS=false
+
+# set BUILD_SYM to 'true' to build and install symbolic binaries
+BUILD_SYM=false
+
+# -arch arguments are different than configure arguments. We need to
+# translate them.
+
+TRANSLATE_ARCH="sed -e s/ppc/powerpc/ -e s/i386/i686/"
+
+# Build GCC the "Apple way".
+# Parameters:
+
+# The first parameter is a space-separated list of the architectures
+# the compilers will run on.  For instance, "ppc i386".  If the
+# current machine isn't in the list, it will (effectively) be added.
+#HOSTS=`echo $1 | $TRANSLATE_ARCH `
+HOSTS=i686
+
+# The second parameter is a space-separated list of the architectures the
+# compilers will generate code for.  If the current machine isn't in
+# the list, a compiler for it will get built anyway, but won't be
+# installed.
+#TARGETS=`echo $2 | $TRANSLATE_ARCH`
+TARGETS=i686
+
+# The GNU makefile target ('bootstrap' by default).
+BOOTSTRAP=${BOOTSTRAP-bootstrap}
+
+# The B&I build srcript (~rc/bin/buildit) accepts an '-othercflags'
+# command-line flag, and captures the argument to that flag in
+# $RC_NONARCH_CFLAGS (and mysteriously prepends '-pipe' thereto).
+# We will allow this to override the default $CFLAGS and $CXXFLAGS.
+
+CFLAGS="-g -O2 ${RC_NONARCH_CFLAGS/-pipe/}"
+
+# This isn't a parameter; it is the architecture of the current machine.
+BUILD=`arch | $TRANSLATE_ARCH`
+
+# The third parameter is the path to the compiler sources.  There should
+# be a shell script named 'configure' in this directory.  This script
+# makes a copy...
+#ORIG_SRC_DIR="$3"
+ORIG_SRC_DIR=`dirname $0`
+ORIG_SRC_DIR=`pwd`/$ORIG_SRC_DIR
+
+# The fourth parameter is the location where the compiler will be installed,
+# normally "/usr".  You can move it once it's built, so this mostly controls
+# the layout of $DEST_DIR.
+#DEST_ROOT="$4"
+DEST_ROOT=/
+
+# The fifth parameter is the place where the compiler will be copied once
+# it's built.
+#DEST_DIR="$5"
+DEST_DIR=/Volumes/Android/device/prebuilt/darwin-x86/gcc-qemu
+
+# The sixth parameter is a directory in which to place information (like
+# unstripped executables and generated source files) helpful in debugging
+# the resulting compiler.
+#SYM_DIR="$6"
+SYM_DIR=`pwd`/sym
+
+# The current working directory is where the build will happen.
+# It may already contain a partial result of an interrupted build,
+# in which case this script will continue where it left off.
+DIR=`pwd`
+
+# This isn't a parameter; it's the version of the compiler that we're
+# about to build.  It's included in the names of various files and
+# directories in the installed image.
+VERS=`sed -n -e '/version_string/s/.*\"\([^ \"]*\)[ \"].*/\1/p' \
+  < $ORIG_SRC_DIR/gcc/version.c || exit 1`
+
+# This isn't a parameter either, it's the major version of the compiler
+# to be built.  It's VERS but only up to the second '.' (if there is one).
+MAJ_VERS=`echo $VERS | sed 's/\([0-9]*\.[0-9]*\)[.-].*/\1/'`
+
+# This is the default architecture for i386 configurations.
+I386_CPU="--with-arch=pentium-m --with-tune=prescott"
+
+# This is the libstdc++ version to use.
+LIBSTDCXX_VERSION=4.0.0
+
+# Sniff to see if we can do ppc64 building.
+DARWIN_VERS=8
+if [ x"`file /usr/lib/crt1.o | grep 'architecture ppc64'`" == x ]; then
+    DARWIN_VERS=7
+fi
+
+echo DARWIN_VERS = $DARWIN_VERS
+
+########################################
+# Run the build.
+
+# Create the source tree we'll actually use to build, deleting
+# tcl since it doesn't actually build properly in a cross environment
+# and we don't really need it.
+SRC_DIR=$DIR/src
+rm -rf $SRC_DIR || exit 1
+mkdir $SRC_DIR || exit 1
+ln -s $ORIG_SRC_DIR/* $SRC_DIR/ || exit 1
+rm -rf $SRC_DIR/tcl $SRC_DIR/expect $SRC_DIR/dejagnu || exit 1
+# Also remove libstdc++ since it is built from a separate project.
+rm -rf $SRC_DIR/libstdc++-v3 || exit 1
+# Clean out old specs files
+rm -f /usr/lib/gcc/*/4.0.0/specs
+
+ENABLE_LANGUAGES="--enable-languages=c,objc"
+if [ $BUILD_CPLUSPLUS = true ] ; then
+    ENABLE_LANGUAGES="$ENABLE_LANGUAGES,c++,obj-c++"
+fi
+
+# These are the configure and build flags that are used.
+CONFIGFLAGS="--disable-checking -enable-werror \
+  --prefix=$DEST_ROOT \
+  --mandir=\${prefix}/share/man \
+  $ENABLE_LANGUAGES \
+  --program-transform-name=/^[cg][^.-]*$/s/$/-$MAJ_VERS/ \
+  --with-gxx-include-dir=\${prefix}/include/c++/$LIBSTDCXX_VERSION \
+  --with-slibdir=/usr/lib \
+  --build=$BUILD-apple-darwin$DARWIN_VERS
+  --disable-nls"
+
+# Figure out how many make processes to run.
+SYSCTL=`sysctl -n hw.activecpu`
+
+# hw.activecpu only available in 10.2.6 and later
+if [ -z "$SYSCTL" ]; then
+  SYSCTL=`sysctl -n hw.ncpu`
+fi
+
+# sysctl -n hw.* does not work when invoked via B&I chroot /BuildRoot. 
+# Builders can default to 2, since even if they are single processor, 
+# nothing else is running on the machine.
+if [ -z "$SYSCTL" ]; then
+  SYSCTL=2
+fi
+
+# The $LOCAL_MAKEFLAGS variable can be used to override $MAKEFLAGS.
+MAKEFLAGS=${LOCAL_MAKEFLAGS-"-j $SYSCTL"}
+
+# Build the native GCC.  Do this even if the user didn't ask for it
+# because it'll be needed for the bootstrap.
+mkdir -p $DIR/obj-$BUILD-$BUILD $DIR/dst-$BUILD-$BUILD || exit 1
+cd $DIR/obj-$BUILD-$BUILD || exit 1
+if [ \! -f Makefile ]; then
+ $SRC_DIR/configure $CONFIGFLAGS \
+   `if [ $BUILD = i686 ] ; then echo $I386_CPU ; fi` \
+   --host=$BUILD-apple-darwin$DARWIN_VERS --target=$BUILD-apple-darwin$DARWIN_VERS || exit 1
+fi
+make $MAKEFLAGS $BOOTSTRAP CFLAGS="$CFLAGS" CXXFLAGS="$CFLAGS" || exit 1
+if [ $BUILD_DOCS = "true" ] ; then
+make $MAKEFLAGS html CFLAGS="$CFLAGS" CXXFLAGS="$CFLAGS" || exit 1
+fi
+make $MAKEFLAGS DESTDIR=$DIR/dst-$BUILD-$BUILD install-gcc install-target \
+  CFLAGS="$CFLAGS" CXXFLAGS="$CFLAGS" || exit 1
+
+# Add the compiler we just built to the path, giving it appropriate names.
+D=$DIR/dst-$BUILD-$BUILD/$DEST_ROOT/bin
+ln -f $D/gcc-$MAJ_VERS $D/gcc || exit 1
+ln -f $D/gcc $D/$BUILD-apple-darwin$DARWIN_VERS-gcc || exit 1
+PATH=$DIR/dst-$BUILD-$BUILD/$DEST_ROOT/bin:$PATH
+
+# The cross-tools' build process expects to find certain programs
+# under names like 'i386-apple-darwin$DARWIN_VERS-ar'; so make them.
+# Annoyingly, ranlib changes behaviour depending on what you call it,
+# so we have to use a shell script for indirection, grrr.
+rm -rf $DIR/bin || exit 1
+mkdir $DIR/bin || exit 1
+for prog in ar nm ranlib strip lipo ; do
+  for t in `echo $TARGETS $HOSTS | sort -u`; do
+    P=$DIR/bin/${t}-apple-darwin$DARWIN_VERS-${prog}
+    echo '#!/bin/sh' > $P || exit 1
+    echo 'exec /usr/bin/'${prog}' $*' >> $P || exit 1
+    chmod a+x $P || exit 1
+  done
+done
+for t in `echo $1 $2 | sort -u`; do
+  gt=`echo $t | $TRANSLATE_ARCH`
+  P=$DIR/bin/${gt}-apple-darwin$DARWIN_VERS-as
+  echo '#!/bin/sh' > $P || exit 1
+  echo 'exec /usr/bin/as -arch '${t}' $*' >> $P || exit 1
+  chmod a+x $P || exit 1
+done
+PATH=$DIR/bin:$PATH
+
+# Build the cross-compilers, using the compiler we just built.
+for t in $TARGETS ; do
+ if [ $t != $BUILD ] ; then
+  mkdir -p $DIR/obj-$BUILD-$t $DIR/dst-$BUILD-$t || exit 1
+   cd $DIR/obj-$BUILD-$t || exit 1
+   if [ \! -f Makefile ]; then
+    $SRC_DIR/configure $CONFIGFLAGS --enable-werror-always \
+     `if [ $t = i686 ] ; then echo $I386_CPU ; fi` \
+      --program-prefix=$t-apple-darwin$DARWIN_VERS- \
+      --host=$BUILD-apple-darwin$DARWIN_VERS --target=$t-apple-darwin$DARWIN_VERS || exit 1
+   fi
+   make $MAKEFLAGS all CFLAGS="$CFLAGS" CXXFLAGS="$CFLAGS" || exit 1
+   make $MAKEFLAGS DESTDIR=$DIR/dst-$BUILD-$t install-gcc install-target \
+     CFLAGS="$CFLAGS" CXXFLAGS="$CFLAGS" || exit 1
+
+    # Add the compiler we just built to the path.
+   PATH=$DIR/dst-$BUILD-$t/$DEST_ROOT/bin:$PATH
+ fi
+done
+
+# Rearrange various libraries, for no really good reason.
+for t in $TARGETS ; do
+  DT=$DIR/dst-$BUILD-$t
+  D=`echo $DT/$DEST_ROOT/lib/gcc/$t-apple-darwin$DARWIN_VERS/$VERS`
+  mv $D/static/libgcc.a $D/libgcc_static.a || exit 1
+  mv $D/kext/libgcc.a $D/libcc_kext.a || exit 1
+  rm -r $D/static $D/kext || exit 1
+done
+
+# Build the cross-hosted compilers.
+for h in $HOSTS ; do
+  if [ $h != $BUILD ] ; then
+    for t in $TARGETS ; do
+      mkdir -p $DIR/obj-$h-$t $DIR/dst-$h-$t || exit 1
+      cd $DIR/obj-$h-$t || exit 1
+      if [ $h = $t ] ; then
+	pp=
+      else
+	pp=$t-apple-darwin$DARWIN_VERS-
+      fi
+
+      if [ \! -f Makefile ]; then
+        $SRC_DIR/configure $CONFIGFLAGS \
+	  `if [ $t = i686 ] ; then echo $I386_CPU ; fi` \
+          --program-prefix=$pp \
+          --host=$h-apple-darwin$DARWIN_VERS --target=$t-apple-darwin$DARWIN_VERS || exit 1
+      fi
+      make $MAKEFLAGS all-gcc CFLAGS="$CFLAGS" CXXFLAGS="$CFLAGS" || exit 1
+      make $MAKEFLAGS DESTDIR=$DIR/dst-$h-$t install-gcc \
+        CFLAGS="$CFLAGS" CXXFLAGS="$CFLAGS" || exit 1
+    done
+  fi
+done
+
+########################################
+# Construct the actual destination root, by copying stuff from
+# $DIR/dst-* to $DEST_DIR, with occasional 'lipo' commands.
+
+cd $DEST_DIR || exit 1
+
+# Clean out DEST_DIR in case -noclean was passed to buildit.
+rm -rf * || exit 1
+
+if [ $BUILD_DOCS = "true" ] ; then
+# HTML documentation
+HTMLDIR="/Developer/ADC Reference Library/documentation/DeveloperTools"
+mkdir -p ".$HTMLDIR" || exit 1
+cp -Rp $DIR/obj-$BUILD-$BUILD/gcc/HTML/* ".$HTMLDIR/" || exit 1
+
+# Manual pages
+mkdir -p .$DEST_ROOT/share || exit 1
+cp -Rp $DIR/dst-$BUILD-$BUILD$DEST_ROOT/share/man .$DEST_ROOT/share/ \
+  || exit 1
+fi
+
+# libexec
+cd $DIR/dst-$BUILD-$BUILD$DEST_ROOT/libexec/gcc/$BUILD-apple-darwin$DARWIN_VERS/$VERS \
+  || exit 1
+LIBEXEC_FILES=`find . -type f -print || exit 1` 
+LIBEXEC_DIRS=`find . -type d -print || exit 1`
+cd $DEST_DIR || exit 1
+for t in $TARGETS ; do
+  DL=$DEST_ROOT/libexec/gcc/$t-apple-darwin$DARWIN_VERS/$VERS
+  for d in $LIBEXEC_DIRS ; do
+    mkdir -p .$DL/$d || exit 1
+  done
+  for f in $LIBEXEC_FILES ; do
+    if file $DIR/dst-*-$t$DL/$f | grep -q 'Mach-O executable' ; then
+      lipo -output .$DL/$f -create $DIR/dst-*-$t$DL/$f || exit 1
+    else
+      cp -p $DIR/dst-$BUILD-$t$DL/$f .$DL/$f || exit 1
+    fi
+  done
+done
+
+# bin
+# The native drivers ('native' is different in different architectures).
+BIN_FILES=`ls $DIR/dst-$BUILD-$BUILD$DEST_ROOT/bin | grep '^[^-]*-[0-9.]*$' \
+  | grep -v gccbug | grep -v gcov || exit 1`
+mkdir .$DEST_ROOT/bin
+for f in $BIN_FILES ; do 
+  lipo -output .$DEST_ROOT/bin/$f -create $DIR/dst-*$DEST_ROOT/bin/$f || exit 1
+done
+# gcov, which is special only because it gets built multiple times and lipo
+# will complain if we try to add two architectures into the same output.
+TARG0=`echo $TARGETS | cut -d ' ' -f 1`
+lipo -output .$DEST_ROOT/bin/gcov-$MAJ_VERS -create \
+  $DIR/dst-*-$TARG0$DEST_ROOT/bin/*gcov* || exit 1
+# The fully-named drivers, which have the same target on every host.
+for t in $TARGETS ; do
+  lipo -output .$DEST_ROOT/bin/$t-apple-darwin$DARWIN_VERS-gcc-$VERS -create \
+    $DIR/dst-*-$t$DEST_ROOT/bin/$t-apple-darwin$DARWIN_VERS-gcc-$VERS || exit 1
+  if [ $BUILD_CPLUSPLUS = "true" ] ; then
+      lipo -output .$DEST_ROOT/bin/$t-apple-darwin$DARWIN_VERS-g++-$VERS -create \
+        $DIR/dst-*-$t$DEST_ROOT/bin/$t-apple-darwin$DARWIN_VERS-g++* || exit 1
+  fi
+done
+
+# lib
+mkdir -p .$DEST_ROOT/lib/gcc || exit 1
+for t in $TARGETS ; do
+  cp -Rp $DIR/dst-$BUILD-$t$DEST_ROOT/lib/gcc/$t-apple-darwin$DARWIN_VERS \
+    .$DEST_ROOT/lib/gcc || exit 1
+done
+
+SHARED_LIBS="libgcc_s.1.dylib libgcc_s.10.4.dylib libgcc_s.10.5.dylib"
+if echo $HOSTS | grep -q powerpc ; then
+  SHARED_LIBS="${SHARED_LIBS} libgcc_s_ppc64.1.dylib"
+fi
+for l in $SHARED_LIBS ; do
+  CANDIDATES=()
+  for t in $TARGETS ; do
+    if [ -e $DIR/dst-$t-$t$DEST_ROOT/lib/$l ] ; then
+      CANDIDATES[${#CANDIDATES[*]}]=$DIR/dst-$t-$t$DEST_ROOT/lib/$l
+    fi
+  done
+  if [ -L ${CANDIDATES[0]} ] ; then
+    ln -s `readlink ${CANDIDATES[0]}` .$DEST_ROOT/lib/$l || exit 1
+  else
+    lipo -output .$DEST_ROOT/lib/$l -create "${CANDIDATES[@]}" || exit 1
+  fi
+done
+
+if [ $BUILD_CPLUSPLUS = "true" ] ; then
+for t in $TARGETS ; do
+  ln -s ../../../libstdc++.6.dylib \
+    .$DEST_ROOT/lib/gcc/$t-apple-darwin$DARWIN_VERS/$VERS/libstdc++.dylib \
+    || exit 1
+done
+fi
+
+# include
+HEADERPATH=$DEST_ROOT/include/gcc/darwin/$MAJ_VERS
+mkdir -p .$HEADERPATH || exit 1
+
+# Some headers are installed from more-hdrs/.  They all share
+# one common feature: they shouldn't be installed here.  Sometimes,
+# they should be part of FSF GCC and installed from there; sometimes,
+# they should be installed by some completely different package; sometimes,
+# they only exist for codewarrior compatibility and codewarrior should provide
+# its own.  We take care not to install the headers if Libc is already
+# providing them.
+cd $SRC_DIR/more-hdrs
+for h in `echo *.h` ; do
+  if [ ! -f /usr/include/$h -o -L /usr/include/$h ] ; then
+    cp -R $h $DEST_DIR$HEADERPATH/$h || exit 1
+    for t in $TARGETS ; do
+      THEADERPATH=$DEST_DIR$DEST_ROOT/lib/gcc/${t}-apple-darwin$DARWIN_VERS/$VERS/include
+      [ -f $THEADERPATH/$h ] || \
+        ln -s ../../../../../include/gcc/darwin/$MAJ_VERS/$h $THEADERPATH/$h || \
+        exit 1
+    done
+  fi
+done
+mkdir -p $DEST_DIR$HEADERPATH/machine
+for h in `echo */*.h` ; do
+  if [ ! -f /usr/include/$h -o -L /usr/include/$h ] ; then
+    cp -R $h $DEST_DIR$HEADERPATH/$h || exit 1
+    for t in $TARGETS ; do
+      THEADERPATH=$DEST_DIR$DEST_ROOT/lib/gcc/${t}-apple-darwin$DARWIN_VERS/$VERS/include
+      mkdir -p $THEADERPATH/machine
+      # In fixincludes/fixinc.in we created this file...  always link for now
+      [ -f /disable/$THEADERPATH/$h ] || \
+        ln -f -s ../../../../../../include/gcc/darwin/$MAJ_VERS/$h $THEADERPATH/$h || \
+        exit 1
+    done
+  fi
+done
+
+if [ $BUILD_DOCS = "true" ] ; then
+if [ $BUILD_CPLUSPLUS = "true" ] ; then
+# Add extra man page symlinks for 'c++' and for arch-specific names.
+MDIR=$DEST_DIR$DEST_ROOT/share/man/man1
+ln -f $MDIR/g++-$MAJ_VERS.1 $MDIR/c++-$MAJ_VERS.1 || exit 1
+for t in $TARGETS ; do
+  ln -f $MDIR/gcc-$MAJ_VERS.1 $MDIR/$t-apple-darwin$DARWIN_VERS-gcc-$VERS.1 \
+      || exit 1
+  ln -f $MDIR/g++-$MAJ_VERS.1 $MDIR/$t-apple-darwin$DARWIN_VERS-g++-$VERS.1 \
+      || exit 1
+done
+fi
+fi
+
+# Build driver-driver using fully-named drivers
+for h in $HOSTS ; do
+    $DEST_DIR$DEST_ROOT/bin/$h-apple-darwin$DARWIN_VERS-gcc-$VERS                          \
+	$ORIG_SRC_DIR/gcc/config/darwin-driver.c                               \
+	-DPDN="\"-apple-darwin$DARWIN_VERS-gcc-$VERS\""                                    \
+	-DIL="\"$DEST_ROOT/bin/\"" -I  $ORIG_SRC_DIR/include                   \
+	-I  $ORIG_SRC_DIR/gcc -I  $ORIG_SRC_DIR/gcc/config                     \
+	-liberty -L$DIR/dst-$BUILD-$h$DEST_ROOT/lib/                           \
+	-L$DIR/dst-$BUILD-$h$DEST_ROOT/$h-apple-darwin$DARWIN_VERS/lib/                    \
+        -L$DIR/obj-$h-$BUILD/libiberty/                                        \
+	-o $DEST_DIR/$DEST_ROOT/bin/tmp-$h-gcc-$MAJ_VERS || exit 1
+
+    if [ $BUILD_CPLUSPLUS = "true" ] ; then
+    $DEST_DIR$DEST_ROOT/bin/$h-apple-darwin$DARWIN_VERS-gcc-$VERS                          \
+	$ORIG_SRC_DIR/gcc/config/darwin-driver.c                               \
+	-DPDN="\"-apple-darwin$DARWIN_VERS-g++-$VERS\""                                    \
+	-DIL="\"$DEST_ROOT/bin/\"" -I  $ORIG_SRC_DIR/include                   \
+	-I  $ORIG_SRC_DIR/gcc -I  $ORIG_SRC_DIR/gcc/config                     \
+	-liberty -L$DIR/dst-$BUILD-$h$DEST_ROOT/lib/                           \
+	-L$DIR/dst-$BUILD-$h$DEST_ROOT/$h-apple-darwin$DARWIN_VERS/lib/                    \
+        -L$DIR/obj-$h-$BUILD/libiberty/                                        \
+	-o $DEST_DIR/$DEST_ROOT/bin/tmp-$h-g++-$MAJ_VERS || exit 1
+    fi
+done
+
+lipo -output $DEST_DIR/$DEST_ROOT/bin/gcc-$MAJ_VERS -create \
+  $DEST_DIR/$DEST_ROOT/bin/tmp-*-gcc-$MAJ_VERS || exit 1
+
+if [ $BUILD_CPLUSPLUS = "true" ] ; then
+lipo -output $DEST_DIR/$DEST_ROOT/bin/g++-$MAJ_VERS -create \
+  $DEST_DIR/$DEST_ROOT/bin/tmp-*-g++-$MAJ_VERS || exit 1
+
+ln -f $DEST_DIR/$DEST_ROOT/bin/g++-$MAJ_VERS $DEST_DIR/$DEST_ROOT/bin/c++-$MAJ_VERS || exit 1
+fi
+
+rm $DEST_DIR/$DEST_ROOT/bin/tmp-*-gcc-$MAJ_VERS || exit 1 
+if [ $BUILD_CPLUPLUS = "true" ] ; then
+rm $DEST_DIR/$DEST_ROOT/bin/tmp-*-g++-$MAJ_VERS || exit 1
+fi
+
+########################################
+# Save the source files and objects needed for debugging
+if [ $BUILD_SYM = "true" ] ; then
+cd $SYM_DIR || exit 1
+
+# Clean out SYM_DIR in case -noclean was passed to buildit.
+rm -rf * || exit 1
+
+# Save executables and libraries.
+cd $DEST_DIR || exit 1
+find . \( -perm -0111 -or -name \*.a -or -name \*.dylib \) -type f -print \
+  | cpio -pdml $SYM_DIR || exit 1
+# Save source files.
+mkdir $SYM_DIR/src || exit 1
+cd $DIR || exit 1
+find obj-* -name \*.\[chy\] -print | cpio -pdml $SYM_DIR/src || exit 1
+fi
+
+########################################
+# Strip the executables and libraries
+find $DEST_DIR -perm -0111 \! -name \*.dylib \! -name fixinc.sh \
+    \! -name mkheaders -type f -print \
+  | xargs strip || exit 1
+find $DEST_DIR \( -name \*.a -or -name \*.dylib \) \
+    \! -name libgcc_s.10.*.dylib -type f -print \
+  | xargs strip -SX || exit 1
+find $DEST_DIR -name \*.a -type f -print \
+  | xargs ranlib || exit 1
+chgrp -h -R wheel $DEST_DIR
+chgrp -R wheel $DEST_DIR
+
+#########################################3
+# Rename the executables
+FILES="gcc cpp"
+if [ $BUILD_CPLUSPLUS = "true" ] ; then
+    FILES="$FILES g++"
+fi
+for ff in $FILES; do
+    ln -f $DEST_DIR/bin/$ff-$MAJ_VERS $DEST_DIR/bin/$ff || exit 1
+done
+
+# Done!
+exit 0
diff --git a/distrib/libpng-1.2.19/Makefile b/distrib/libpng-1.2.19/Makefile
new file mode 100644
index 0000000..ba59f45
--- /dev/null
+++ b/distrib/libpng-1.2.19/Makefile
@@ -0,0 +1,25 @@
+# Makefile used to compile libpng statically
+# you need to define ZLIB_INCLUDE to the Zlib include path
+# and PREFIX to the installation path
+#
+LIBPNG_LIB    := $(SRC_PATH)/libpng.a
+LIBPNG_CFLAGS := -I$(LIBPNG_DIR)
+
+HOST_ARCH := $(shell uname -p)
+HOST_OS   := $(shell uname -s)
+ifeq ($(HOST_OS),Darwin)
+    HOST_OS := darwin
+endif
+
+include $(LIBPNG_DIR)/sources.make
+
+LIBPNG_OBJS := $(LIBPNG_SOURCES:%.c=%.o)
+
+$(LIBPNG_LIB): $(LIBPNG_OBJS)
+	ar ru $@ $(LIBPNG_OBJS)
+
+$(LIBPNG_OBJS): CFLAGS += $(ZLIB_CFLAGS) $(LIBPNG_CFLAGS)
+
+clean-libpng:
+	rm -f $(LIBPNG_OBJS) $(LIBPNG_LIB)
+
diff --git a/distrib/libpng-1.2.19/png.c b/distrib/libpng-1.2.19/png.c
new file mode 100644
index 0000000..8aa4131
--- /dev/null
+++ b/distrib/libpng-1.2.19/png.c
@@ -0,0 +1,895 @@
+
+/* png.c - location for general purpose libpng functions
+ *
+ * Last changed in libpng 1.2.19 August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ */
+
+#define PNG_INTERNAL
+#define PNG_NO_EXTERN
+#include "png.h"
+
+/* Generate a compiler error if there is an old png.h in the search path. */
+typedef version_1_2_19 Your_png_h_is_not_version_1_2_19;
+
+/* Version information for C files.  This had better match the version
+ * string defined in png.h.  */
+
+#ifdef PNG_USE_GLOBAL_ARRAYS
+/* png_libpng_ver was changed to a function in version 1.0.5c */
+PNG_CONST char png_libpng_ver[18] = PNG_LIBPNG_VER_STRING;
+
+#ifdef PNG_READ_SUPPORTED
+
+/* png_sig was changed to a function in version 1.0.5c */
+/* Place to hold the signature string for a PNG file. */
+PNG_CONST png_byte FARDATA png_sig[8] = {137, 80, 78, 71, 13, 10, 26, 10};
+#endif /* PNG_READ_SUPPORTED */
+
+/* Invoke global declarations for constant strings for known chunk types */
+PNG_IHDR;
+PNG_IDAT;
+PNG_IEND;
+PNG_PLTE;
+PNG_bKGD;
+PNG_cHRM;
+PNG_gAMA;
+PNG_hIST;
+PNG_iCCP;
+PNG_iTXt;
+PNG_oFFs;
+PNG_pCAL;
+PNG_sCAL;
+PNG_pHYs;
+PNG_sBIT;
+PNG_sPLT;
+PNG_sRGB;
+PNG_tEXt;
+PNG_tIME;
+PNG_tRNS;
+PNG_zTXt;
+
+#ifdef PNG_READ_SUPPORTED
+/* arrays to facilitate easy interlacing - use pass (0 - 6) as index */
+
+/* start of interlace block */
+PNG_CONST int FARDATA png_pass_start[] = {0, 4, 0, 2, 0, 1, 0};
+
+/* offset to next interlace block */
+PNG_CONST int FARDATA png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1};
+
+/* start of interlace block in the y direction */
+PNG_CONST int FARDATA png_pass_ystart[] = {0, 0, 4, 0, 2, 0, 1};
+
+/* offset to next interlace block in the y direction */
+PNG_CONST int FARDATA png_pass_yinc[] = {8, 8, 8, 4, 4, 2, 2};
+
+/* width of interlace block (used in assembler routines only) */
+#if defined(PNG_HAVE_MMX_COMBINE_ROW) || defined(PNG_OPTIMIZED_CODE_SUPPORTED)
+PNG_CONST int FARDATA png_pass_width[] = {8, 4, 4, 2, 2, 1, 1};
+#endif
+
+/* Height of interlace block.  This is not currently used - if you need
+ * it, uncomment it here and in png.h
+PNG_CONST int FARDATA png_pass_height[] = {8, 8, 4, 4, 2, 2, 1};
+*/
+
+/* Mask to determine which pixels are valid in a pass */
+PNG_CONST int FARDATA png_pass_mask[] = {0x80, 0x08, 0x88, 0x22, 0xaa, 0x55, 0xff};
+
+/* Mask to determine which pixels to overwrite while displaying */
+PNG_CONST int FARDATA png_pass_dsp_mask[]
+   = {0xff, 0x0f, 0xff, 0x33, 0xff, 0x55, 0xff};
+
+#endif /* PNG_READ_SUPPORTED */
+#endif /* PNG_USE_GLOBAL_ARRAYS */
+
+/* Tells libpng that we have already handled the first "num_bytes" bytes
+ * of the PNG file signature.  If the PNG data is embedded into another
+ * stream we can set num_bytes = 8 so that libpng will not attempt to read
+ * or write any of the magic bytes before it starts on the IHDR.
+ */
+
+#ifdef PNG_READ_SUPPORTED
+void PNGAPI
+png_set_sig_bytes(png_structp png_ptr, int num_bytes)
+{
+   if(png_ptr == NULL) return;
+   png_debug(1, "in png_set_sig_bytes\n");
+   if (num_bytes > 8)
+      png_error(png_ptr, "Too many bytes for PNG signature.");
+
+   png_ptr->sig_bytes = (png_byte)(num_bytes < 0 ? 0 : num_bytes);
+}
+
+/* Checks whether the supplied bytes match the PNG signature.  We allow
+ * checking less than the full 8-byte signature so that those apps that
+ * already read the first few bytes of a file to determine the file type
+ * can simply check the remaining bytes for extra assurance.  Returns
+ * an integer less than, equal to, or greater than zero if sig is found,
+ * respectively, to be less than, to match, or be greater than the correct
+ * PNG signature (this is the same behaviour as strcmp, memcmp, etc).
+ */
+int PNGAPI
+png_sig_cmp(png_bytep sig, png_size_t start, png_size_t num_to_check)
+{
+   png_byte png_signature[8] = {137, 80, 78, 71, 13, 10, 26, 10};
+   if (num_to_check > 8)
+      num_to_check = 8;
+   else if (num_to_check < 1)
+      return (-1);
+
+   if (start > 7)
+      return (-1);
+
+   if (start + num_to_check > 8)
+      num_to_check = 8 - start;
+
+   return ((int)(png_memcmp(&sig[start], &png_signature[start], num_to_check)));
+}
+
+#if defined(PNG_1_0_X) || defined(PNG_1_2_X)
+/* (Obsolete) function to check signature bytes.  It does not allow one
+ * to check a partial signature.  This function might be removed in the
+ * future - use png_sig_cmp().  Returns true (nonzero) if the file is PNG.
+ */
+int PNGAPI
+png_check_sig(png_bytep sig, int num)
+{
+  return ((int)!png_sig_cmp(sig, (png_size_t)0, (png_size_t)num));
+}
+#endif
+#endif /* PNG_READ_SUPPORTED */
+
+#if defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED)
+/* Function to allocate memory for zlib and clear it to 0. */
+#ifdef PNG_1_0_X
+voidpf PNGAPI
+#else
+voidpf /* private */
+#endif
+png_zalloc(voidpf png_ptr, uInt items, uInt size)
+{
+   png_voidp ptr;
+   png_structp p=(png_structp)png_ptr;
+   png_uint_32 save_flags=p->flags;
+   png_uint_32 num_bytes;
+
+   if(png_ptr == NULL) return (NULL);
+   if (items > PNG_UINT_32_MAX/size)
+   {
+     png_warning (p, "Potential overflow in png_zalloc()");
+     return (NULL);
+   }
+   num_bytes = (png_uint_32)items * size;
+
+   p->flags|=PNG_FLAG_MALLOC_NULL_MEM_OK;
+   ptr = (png_voidp)png_malloc((png_structp)png_ptr, num_bytes);
+   p->flags=save_flags;
+
+#if defined(PNG_1_0_X) && !defined(PNG_NO_ZALLOC_ZERO)
+   if (ptr == NULL)
+       return ((voidpf)ptr);
+
+   if (num_bytes > (png_uint_32)0x8000L)
+   {
+      png_memset(ptr, 0, (png_size_t)0x8000L);
+      png_memset((png_bytep)ptr + (png_size_t)0x8000L, 0,
+         (png_size_t)(num_bytes - (png_uint_32)0x8000L));
+   }
+   else
+   {
+      png_memset(ptr, 0, (png_size_t)num_bytes);
+   }
+#endif
+   return ((voidpf)ptr);
+}
+
+/* function to free memory for zlib */
+#ifdef PNG_1_0_X
+void PNGAPI
+#else
+void /* private */
+#endif
+png_zfree(voidpf png_ptr, voidpf ptr)
+{
+   png_free((png_structp)png_ptr, (png_voidp)ptr);
+}
+
+/* Reset the CRC variable to 32 bits of 1's.  Care must be taken
+ * in case CRC is > 32 bits to leave the top bits 0.
+ */
+void /* PRIVATE */
+png_reset_crc(png_structp png_ptr)
+{
+   png_ptr->crc = crc32(0, Z_NULL, 0);
+}
+
+/* Calculate the CRC over a section of data.  We can only pass as
+ * much data to this routine as the largest single buffer size.  We
+ * also check that this data will actually be used before going to the
+ * trouble of calculating it.
+ */
+void /* PRIVATE */
+png_calculate_crc(png_structp png_ptr, png_bytep ptr, png_size_t length)
+{
+   int need_crc = 1;
+
+   if (png_ptr->chunk_name[0] & 0x20)                     /* ancillary */
+   {
+      if ((png_ptr->flags & PNG_FLAG_CRC_ANCILLARY_MASK) ==
+          (PNG_FLAG_CRC_ANCILLARY_USE | PNG_FLAG_CRC_ANCILLARY_NOWARN))
+         need_crc = 0;
+   }
+   else                                                    /* critical */
+   {
+      if (png_ptr->flags & PNG_FLAG_CRC_CRITICAL_IGNORE)
+         need_crc = 0;
+   }
+
+   if (need_crc)
+      png_ptr->crc = crc32(png_ptr->crc, ptr, (uInt)length);
+}
+
+/* Allocate the memory for an info_struct for the application.  We don't
+ * really need the png_ptr, but it could potentially be useful in the
+ * future.  This should be used in favour of malloc(png_sizeof(png_info))
+ * and png_info_init() so that applications that want to use a shared
+ * libpng don't have to be recompiled if png_info changes size.
+ */
+png_infop PNGAPI
+png_create_info_struct(png_structp png_ptr)
+{
+   png_infop info_ptr;
+
+   png_debug(1, "in png_create_info_struct\n");
+   if(png_ptr == NULL) return (NULL);
+#ifdef PNG_USER_MEM_SUPPORTED
+   info_ptr = (png_infop)png_create_struct_2(PNG_STRUCT_INFO,
+      png_ptr->malloc_fn, png_ptr->mem_ptr);
+#else
+   info_ptr = (png_infop)png_create_struct(PNG_STRUCT_INFO);
+#endif
+   if (info_ptr != NULL)
+      png_info_init_3(&info_ptr, png_sizeof(png_info));
+
+   return (info_ptr);
+}
+
+/* This function frees the memory associated with a single info struct.
+ * Normally, one would use either png_destroy_read_struct() or
+ * png_destroy_write_struct() to free an info struct, but this may be
+ * useful for some applications.
+ */
+void PNGAPI
+png_destroy_info_struct(png_structp png_ptr, png_infopp info_ptr_ptr)
+{
+   png_infop info_ptr = NULL;
+   if(png_ptr == NULL) return;
+
+   png_debug(1, "in png_destroy_info_struct\n");
+   if (info_ptr_ptr != NULL)
+      info_ptr = *info_ptr_ptr;
+
+   if (info_ptr != NULL)
+   {
+      png_info_destroy(png_ptr, info_ptr);
+
+#ifdef PNG_USER_MEM_SUPPORTED
+      png_destroy_struct_2((png_voidp)info_ptr, png_ptr->free_fn,
+          png_ptr->mem_ptr);
+#else
+      png_destroy_struct((png_voidp)info_ptr);
+#endif
+      *info_ptr_ptr = NULL;
+   }
+}
+
+/* Initialize the info structure.  This is now an internal function (0.89)
+ * and applications using it are urged to use png_create_info_struct()
+ * instead.
+ */
+#if defined(PNG_1_0_X) || defined(PNG_1_2_X)
+#undef png_info_init
+void PNGAPI
+png_info_init(png_infop info_ptr)
+{
+   /* We only come here via pre-1.0.12-compiled applications */
+   png_info_init_3(&info_ptr, 0);
+}
+#endif
+
+void PNGAPI
+png_info_init_3(png_infopp ptr_ptr, png_size_t png_info_struct_size)
+{
+   png_infop info_ptr = *ptr_ptr;
+
+   if(info_ptr == NULL) return;
+
+   png_debug(1, "in png_info_init_3\n");
+
+   if(png_sizeof(png_info) > png_info_struct_size)
+     {
+       png_destroy_struct(info_ptr);
+       info_ptr = (png_infop)png_create_struct(PNG_STRUCT_INFO);
+       *ptr_ptr = info_ptr;
+     }
+
+   /* set everything to 0 */
+   png_memset(info_ptr, 0, png_sizeof (png_info));
+}
+
+#ifdef PNG_FREE_ME_SUPPORTED
+void PNGAPI
+png_data_freer(png_structp png_ptr, png_infop info_ptr,
+   int freer, png_uint_32 mask)
+{
+   png_debug(1, "in png_data_freer\n");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+   if(freer == PNG_DESTROY_WILL_FREE_DATA)
+      info_ptr->free_me |= mask;
+   else if(freer == PNG_USER_WILL_FREE_DATA)
+      info_ptr->free_me &= ~mask;
+   else
+      png_warning(png_ptr,
+         "Unknown freer parameter in png_data_freer.");
+}
+#endif
+
+void PNGAPI
+png_free_data(png_structp png_ptr, png_infop info_ptr, png_uint_32 mask,
+   int num)
+{
+   png_debug(1, "in png_free_data\n");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+#if defined(PNG_TEXT_SUPPORTED)
+/* free text item num or (if num == -1) all text items */
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_TEXT) & info_ptr->free_me)
+#else
+if (mask & PNG_FREE_TEXT)
+#endif
+{
+   if (num != -1)
+   {
+     if (info_ptr->text && info_ptr->text[num].key)
+     {
+         png_free(png_ptr, info_ptr->text[num].key);
+         info_ptr->text[num].key = NULL;
+     }
+   }
+   else
+   {
+       int i;
+       for (i = 0; i < info_ptr->num_text; i++)
+           png_free_data(png_ptr, info_ptr, PNG_FREE_TEXT, i);
+       png_free(png_ptr, info_ptr->text);
+       info_ptr->text = NULL;
+       info_ptr->num_text=0;
+   }
+}
+#endif
+
+#if defined(PNG_tRNS_SUPPORTED)
+/* free any tRNS entry */
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_TRNS) & info_ptr->free_me)
+#else
+if ((mask & PNG_FREE_TRNS) && (png_ptr->flags & PNG_FLAG_FREE_TRNS))
+#endif
+{
+    png_free(png_ptr, info_ptr->trans);
+    info_ptr->valid &= ~PNG_INFO_tRNS;
+#ifndef PNG_FREE_ME_SUPPORTED
+    png_ptr->flags &= ~PNG_FLAG_FREE_TRNS;
+#endif
+    info_ptr->trans = NULL;
+}
+#endif
+
+#if defined(PNG_sCAL_SUPPORTED)
+/* free any sCAL entry */
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_SCAL) & info_ptr->free_me)
+#else
+if (mask & PNG_FREE_SCAL)
+#endif
+{
+#if defined(PNG_FIXED_POINT_SUPPORTED) && !defined(PNG_FLOATING_POINT_SUPPORTED)
+    png_free(png_ptr, info_ptr->scal_s_width);
+    png_free(png_ptr, info_ptr->scal_s_height);
+    info_ptr->scal_s_width = NULL;
+    info_ptr->scal_s_height = NULL;
+#endif
+    info_ptr->valid &= ~PNG_INFO_sCAL;
+}
+#endif
+
+#if defined(PNG_pCAL_SUPPORTED)
+/* free any pCAL entry */
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_PCAL) & info_ptr->free_me)
+#else
+if (mask & PNG_FREE_PCAL)
+#endif
+{
+    png_free(png_ptr, info_ptr->pcal_purpose);
+    png_free(png_ptr, info_ptr->pcal_units);
+    info_ptr->pcal_purpose = NULL;
+    info_ptr->pcal_units = NULL;
+    if (info_ptr->pcal_params != NULL)
+    {
+        int i;
+        for (i = 0; i < (int)info_ptr->pcal_nparams; i++)
+        {
+          png_free(png_ptr, info_ptr->pcal_params[i]);
+          info_ptr->pcal_params[i]=NULL;
+        }
+        png_free(png_ptr, info_ptr->pcal_params);
+        info_ptr->pcal_params = NULL;
+    }
+    info_ptr->valid &= ~PNG_INFO_pCAL;
+}
+#endif
+
+#if defined(PNG_iCCP_SUPPORTED)
+/* free any iCCP entry */
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_ICCP) & info_ptr->free_me)
+#else
+if (mask & PNG_FREE_ICCP)
+#endif
+{
+    png_free(png_ptr, info_ptr->iccp_name);
+    png_free(png_ptr, info_ptr->iccp_profile);
+    info_ptr->iccp_name = NULL;
+    info_ptr->iccp_profile = NULL;
+    info_ptr->valid &= ~PNG_INFO_iCCP;
+}
+#endif
+
+#if defined(PNG_sPLT_SUPPORTED)
+/* free a given sPLT entry, or (if num == -1) all sPLT entries */
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_SPLT) & info_ptr->free_me)
+#else
+if (mask & PNG_FREE_SPLT)
+#endif
+{
+   if (num != -1)
+   {
+      if(info_ptr->splt_palettes)
+      {
+          png_free(png_ptr, info_ptr->splt_palettes[num].name);
+          png_free(png_ptr, info_ptr->splt_palettes[num].entries);
+          info_ptr->splt_palettes[num].name = NULL;
+          info_ptr->splt_palettes[num].entries = NULL;
+      }
+   }
+   else
+   {
+       if(info_ptr->splt_palettes_num)
+       {
+         int i;
+         for (i = 0; i < (int)info_ptr->splt_palettes_num; i++)
+            png_free_data(png_ptr, info_ptr, PNG_FREE_SPLT, i);
+
+         png_free(png_ptr, info_ptr->splt_palettes);
+         info_ptr->splt_palettes = NULL;
+         info_ptr->splt_palettes_num = 0;
+       }
+       info_ptr->valid &= ~PNG_INFO_sPLT;
+   }
+}
+#endif
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+  if(png_ptr->unknown_chunk.data)
+  {
+    png_free(png_ptr, png_ptr->unknown_chunk.data);
+    png_ptr->unknown_chunk.data = NULL;
+  }
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_UNKN) & info_ptr->free_me)
+#else
+if (mask & PNG_FREE_UNKN)
+#endif
+{
+   if (num != -1)
+   {
+       if(info_ptr->unknown_chunks)
+       {
+          png_free(png_ptr, info_ptr->unknown_chunks[num].data);
+          info_ptr->unknown_chunks[num].data = NULL;
+       }
+   }
+   else
+   {
+       int i;
+
+       if(info_ptr->unknown_chunks_num)
+       {
+         for (i = 0; i < (int)info_ptr->unknown_chunks_num; i++)
+            png_free_data(png_ptr, info_ptr, PNG_FREE_UNKN, i);
+
+         png_free(png_ptr, info_ptr->unknown_chunks);
+         info_ptr->unknown_chunks = NULL;
+         info_ptr->unknown_chunks_num = 0;
+       }
+   }
+}
+#endif
+
+#if defined(PNG_hIST_SUPPORTED)
+/* free any hIST entry */
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_HIST)  & info_ptr->free_me)
+#else
+if ((mask & PNG_FREE_HIST) && (png_ptr->flags & PNG_FLAG_FREE_HIST))
+#endif
+{
+    png_free(png_ptr, info_ptr->hist);
+    info_ptr->hist = NULL;
+    info_ptr->valid &= ~PNG_INFO_hIST;
+#ifndef PNG_FREE_ME_SUPPORTED
+    png_ptr->flags &= ~PNG_FLAG_FREE_HIST;
+#endif
+}
+#endif
+
+/* free any PLTE entry that was internally allocated */
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_PLTE) & info_ptr->free_me)
+#else
+if ((mask & PNG_FREE_PLTE) && (png_ptr->flags & PNG_FLAG_FREE_PLTE))
+#endif
+{
+    png_zfree(png_ptr, info_ptr->palette);
+    info_ptr->palette = NULL;
+    info_ptr->valid &= ~PNG_INFO_PLTE;
+#ifndef PNG_FREE_ME_SUPPORTED
+    png_ptr->flags &= ~PNG_FLAG_FREE_PLTE;
+#endif
+    info_ptr->num_palette = 0;
+}
+
+#if defined(PNG_INFO_IMAGE_SUPPORTED)
+/* free any image bits attached to the info structure */
+#ifdef PNG_FREE_ME_SUPPORTED
+if ((mask & PNG_FREE_ROWS) & info_ptr->free_me)
+#else
+if (mask & PNG_FREE_ROWS)
+#endif
+{
+    if(info_ptr->row_pointers)
+    {
+       int row;
+       for (row = 0; row < (int)info_ptr->height; row++)
+       {
+          png_free(png_ptr, info_ptr->row_pointers[row]);
+          info_ptr->row_pointers[row]=NULL;
+       }
+       png_free(png_ptr, info_ptr->row_pointers);
+       info_ptr->row_pointers=NULL;
+    }
+    info_ptr->valid &= ~PNG_INFO_IDAT;
+}
+#endif
+
+#ifdef PNG_FREE_ME_SUPPORTED
+   if(num == -1)
+     info_ptr->free_me &= ~mask;
+   else
+     info_ptr->free_me &= ~(mask & ~PNG_FREE_MUL);
+#endif
+}
+
+/* This is an internal routine to free any memory that the info struct is
+ * pointing to before re-using it or freeing the struct itself.  Recall
+ * that png_free() checks for NULL pointers for us.
+ */
+void /* PRIVATE */
+png_info_destroy(png_structp png_ptr, png_infop info_ptr)
+{
+   png_debug(1, "in png_info_destroy\n");
+
+   png_free_data(png_ptr, info_ptr, PNG_FREE_ALL, -1);
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+   if (png_ptr->num_chunk_list)
+   {
+       png_free(png_ptr, png_ptr->chunk_list);
+       png_ptr->chunk_list=NULL;
+       png_ptr->num_chunk_list=0;
+   }
+#endif
+
+   png_info_init_3(&info_ptr, png_sizeof(png_info));
+}
+#endif /* defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED) */
+
+/* This function returns a pointer to the io_ptr associated with the user
+ * functions.  The application should free any memory associated with this
+ * pointer before png_write_destroy() or png_read_destroy() are called.
+ */
+png_voidp PNGAPI
+png_get_io_ptr(png_structp png_ptr)
+{
+   if(png_ptr == NULL) return (NULL);
+   return (png_ptr->io_ptr);
+}
+
+#if defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED)
+#if !defined(PNG_NO_STDIO)
+/* Initialize the default input/output functions for the PNG file.  If you
+ * use your own read or write routines, you can call either png_set_read_fn()
+ * or png_set_write_fn() instead of png_init_io().  If you have defined
+ * PNG_NO_STDIO, you must use a function of your own because "FILE *" isn't
+ * necessarily available.
+ */
+void PNGAPI
+png_init_io(png_structp png_ptr, png_FILE_p fp)
+{
+   png_debug(1, "in png_init_io\n");
+   if(png_ptr == NULL) return;
+   png_ptr->io_ptr = (png_voidp)fp;
+}
+#endif
+
+#if defined(PNG_TIME_RFC1123_SUPPORTED)
+/* Convert the supplied time into an RFC 1123 string suitable for use in
+ * a "Creation Time" or other text-based time string.
+ */
+png_charp PNGAPI
+png_convert_to_rfc1123(png_structp png_ptr, png_timep ptime)
+{
+   static PNG_CONST char short_months[12][4] =
+        {"Jan", "Feb", "Mar", "Apr", "May", "Jun",
+         "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"};
+
+   if(png_ptr == NULL) return (NULL);
+   if (png_ptr->time_buffer == NULL)
+   {
+      png_ptr->time_buffer = (png_charp)png_malloc(png_ptr, (png_uint_32)(29*
+         png_sizeof(char)));
+   }
+
+#if defined(_WIN32_WCE)
+   {
+      wchar_t time_buf[29];
+      wsprintf(time_buf, TEXT("%d %S %d %02d:%02d:%02d +0000"),
+          ptime->day % 32, short_months[(ptime->month - 1) % 12],
+        ptime->year, ptime->hour % 24, ptime->minute % 60,
+          ptime->second % 61);
+      WideCharToMultiByte(CP_ACP, 0, time_buf, -1, png_ptr->time_buffer, 29,
+          NULL, NULL);
+   }
+#else
+#ifdef USE_FAR_KEYWORD
+   {
+      char near_time_buf[29];
+      png_snprintf6(near_time_buf,29,"%d %s %d %02d:%02d:%02d +0000",
+          ptime->day % 32, short_months[(ptime->month - 1) % 12],
+          ptime->year, ptime->hour % 24, ptime->minute % 60,
+          ptime->second % 61);
+      png_memcpy(png_ptr->time_buffer, near_time_buf,
+          29*png_sizeof(char));
+   }
+#else
+   png_snprintf6(png_ptr->time_buffer,29,"%d %s %d %02d:%02d:%02d +0000",
+       ptime->day % 32, short_months[(ptime->month - 1) % 12],
+       ptime->year, ptime->hour % 24, ptime->minute % 60,
+       ptime->second % 61);
+#endif
+#endif /* _WIN32_WCE */
+   return ((png_charp)png_ptr->time_buffer);
+}
+#endif /* PNG_TIME_RFC1123_SUPPORTED */
+
+#endif /* defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED) */
+
+png_charp PNGAPI
+png_get_copyright(png_structp png_ptr)
+{
+   png_ptr = png_ptr;  /* silence compiler warning about unused png_ptr */
+   return ((png_charp) "\n libpng version 1.2.19 - August 18, 2007\n\
+   Copyright (c) 1998-2007 Glenn Randers-Pehrson\n\
+   Copyright (c) 1996-1997 Andreas Dilger\n\
+   Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc.\n");
+}
+
+/* The following return the library version as a short string in the
+ * format 1.0.0 through 99.99.99zz.  To get the version of *.h files
+ * used with your application, print out PNG_LIBPNG_VER_STRING, which
+ * is defined in png.h.
+ * Note: now there is no difference between png_get_libpng_ver() and
+ * png_get_header_ver().  Due to the version_nn_nn_nn typedef guard,
+ * it is guaranteed that png.c uses the correct version of png.h.
+ */
+png_charp PNGAPI
+png_get_libpng_ver(png_structp png_ptr)
+{
+   /* Version of *.c files used when building libpng */
+   png_ptr = png_ptr;  /* silence compiler warning about unused png_ptr */
+   return ((png_charp) PNG_LIBPNG_VER_STRING);
+}
+
+png_charp PNGAPI
+png_get_header_ver(png_structp png_ptr)
+{
+   /* Version of *.h files used when building libpng */
+   png_ptr = png_ptr;  /* silence compiler warning about unused png_ptr */
+   return ((png_charp) PNG_LIBPNG_VER_STRING);
+}
+
+png_charp PNGAPI
+png_get_header_version(png_structp png_ptr)
+{
+   /* Returns longer string containing both version and date */
+   png_ptr = png_ptr;  /* silence compiler warning about unused png_ptr */
+   return ((png_charp) PNG_HEADER_VERSION_STRING
+#ifdef PNG_READ_SUPPORTED
+#  ifdef PNG_USE_PNGGCCRD
+#    ifdef __x86_64__
+#      ifdef __PIC__
+   "     (PNGGCRD x86_64, PIC)\n"
+#      else
+#        ifdef PNG_THREAD_UNSAFE_OK
+   "     (PNGGCRD x86_64, Thread unsafe)\n"
+#        else
+   "     (PNGGCRD x86_64, Thread safe)\n"
+#        endif
+#      endif
+#    else
+#    ifdef PNG_THREAD_UNSAFE_OK
+   "     (PNGGCRD, Thread unsafe)\n"
+#      else
+   "     (PNGGCRD, Thread safe)\n"
+#      endif
+#    endif
+#  else
+#    ifdef PNG_USE_PNGVCRD
+#      ifdef __x86_64__
+   "     (x86_64 PNGVCRD)\n"
+#      else
+   "     (PNGVCRD)\n"
+#      endif
+#    else
+#      ifdef __x86_64__
+#        ifdef PNG_OPTIMIZED_CODE_SUPPORTED
+   "     (x86_64 OPTIMIZED)\n"
+#        else
+   "     (x86_64 NOT OPTIMIZED)\n"
+#        endif
+#      else
+#        ifdef PNG_OPTIMIZED_CODE_SUPPORTED
+   "     (OPTIMIZED)\n"
+#        else
+   "     (NOT OPTIMIZED)\n"
+#        endif
+#      endif
+#    endif
+#  endif
+#else
+   "     (NO READ SUPPORT)\n"
+#endif
+   );
+}
+
+#if defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED)
+#ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED
+int PNGAPI
+png_handle_as_unknown(png_structp png_ptr, png_bytep chunk_name)
+{
+   /* check chunk_name and return "keep" value if it's on the list, else 0 */
+   int i;
+   png_bytep p;
+   if(png_ptr == NULL || chunk_name == NULL || png_ptr->num_chunk_list<=0)
+      return 0;
+   p=png_ptr->chunk_list+png_ptr->num_chunk_list*5-5;
+   for (i = png_ptr->num_chunk_list; i; i--, p-=5)
+      if (!png_memcmp(chunk_name, p, 4))
+        return ((int)*(p+4));
+   return 0;
+}
+#endif
+
+/* This function, added to libpng-1.0.6g, is untested. */
+int PNGAPI
+png_reset_zstream(png_structp png_ptr)
+{
+   if (png_ptr == NULL) return Z_STREAM_ERROR;
+   return (inflateReset(&png_ptr->zstream));
+}
+#endif /* defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED) */
+
+/* This function was added to libpng-1.0.7 */
+png_uint_32 PNGAPI
+png_access_version_number(void)
+{
+   /* Version of *.c files used when building libpng */
+   return((png_uint_32) PNG_LIBPNG_VER);
+}
+
+
+#if defined(PNG_READ_SUPPORTED) && defined(PNG_ASSEMBLER_CODE_SUPPORTED)
+#if !defined(PNG_1_0_X)
+#if defined(PNG_MMX_CODE_SUPPORTED)
+/* this INTERNAL function was added to libpng 1.2.0 */
+void /* PRIVATE */
+png_init_mmx_flags (png_structp png_ptr)
+{
+    if(png_ptr == NULL) return;
+    png_ptr->mmx_rowbytes_threshold = 0;
+    png_ptr->mmx_bitdepth_threshold = 0;
+
+#  if (defined(PNG_USE_PNGVCRD) || defined(PNG_USE_PNGGCCRD))
+
+    png_ptr->asm_flags |= PNG_ASM_FLAG_MMX_SUPPORT_COMPILED;
+
+    if (png_mmx_support() > 0) {
+        png_ptr->asm_flags |= PNG_ASM_FLAG_MMX_SUPPORT_IN_CPU
+#    ifdef PNG_HAVE_MMX_COMBINE_ROW
+                              | PNG_ASM_FLAG_MMX_READ_COMBINE_ROW
+#    endif
+#    ifdef PNG_HAVE_MMX_READ_INTERLACE
+                              | PNG_ASM_FLAG_MMX_READ_INTERLACE
+#    endif
+#    ifndef PNG_HAVE_MMX_READ_FILTER_ROW
+                              ;
+#    else
+                              | PNG_ASM_FLAG_MMX_READ_FILTER_SUB
+                              | PNG_ASM_FLAG_MMX_READ_FILTER_UP
+                              | PNG_ASM_FLAG_MMX_READ_FILTER_AVG
+                              | PNG_ASM_FLAG_MMX_READ_FILTER_PAETH ;
+
+        png_ptr->mmx_rowbytes_threshold = PNG_MMX_ROWBYTES_THRESHOLD_DEFAULT;
+        png_ptr->mmx_bitdepth_threshold = PNG_MMX_BITDEPTH_THRESHOLD_DEFAULT;
+#    endif
+    } else {
+        png_ptr->asm_flags &= ~( PNG_ASM_FLAG_MMX_SUPPORT_IN_CPU
+                               | PNG_MMX_READ_FLAGS
+                               | PNG_MMX_WRITE_FLAGS );
+    }
+
+#  else /* !(PNGVCRD || PNGGCCRD) */
+
+    /* clear all MMX flags; no support is compiled in */
+    png_ptr->asm_flags &= ~( PNG_MMX_FLAGS );
+
+#  endif /* ?(PNGVCRD || PNGGCCRD) */
+}
+
+#endif /* !(PNG_MMX_CODE_SUPPORTED) */
+
+/* this function was added to libpng 1.2.0 */
+#if !defined(PNG_USE_PNGGCCRD) && \
+    !(defined(PNG_MMX_CODE_SUPPORTED) && defined(PNG_USE_PNGVCRD))
+int PNGAPI
+png_mmx_support(void)
+{
+    return -1;
+}
+#endif
+#endif /* PNG_1_0_X */
+#endif /* PNG_READ_SUPPORTED && PNG_ASSEMBLER_CODE_SUPPORTED */
+
+#if defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED)
+#ifdef PNG_SIZE_T
+/* Added at libpng version 1.2.6 */
+   PNG_EXTERN png_size_t PNGAPI png_convert_size PNGARG((size_t size));
+png_size_t PNGAPI
+png_convert_size(size_t size)
+{
+  if (size > (png_size_t)-1)
+     PNG_ABORT();  /* We haven't got access to png_ptr, so no png_error() */
+  return ((png_size_t)size);
+}
+#endif /* PNG_SIZE_T */
+#endif /* defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED) */
diff --git a/distrib/libpng-1.2.19/png.h b/distrib/libpng-1.2.19/png.h
new file mode 100644
index 0000000..04f786d
--- /dev/null
+++ b/distrib/libpng-1.2.19/png.h
@@ -0,0 +1,3525 @@
+
+/* png.h - header file for PNG reference library
+ *
+ * libpng version 1.2.19 - August 18, 2007
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ *
+ * Authors and maintainers:
+ *  libpng versions 0.71, May 1995, through 0.88, January 1996: Guy Schalnat
+ *  libpng versions 0.89c, June 1996, through 0.96, May 1997: Andreas Dilger
+ *  libpng versions 0.97, January 1998, through 1.2.19 - August 18, 2007: Glenn
+ *  See also "Contributing Authors", below.
+ *
+ * Note about libpng version numbers:
+ *
+ *    Due to various miscommunications, unforeseen code incompatibilities
+ *    and occasional factors outside the authors' control, version numbering
+ *    on the library has not always been consistent and straightforward.
+ *    The following table summarizes matters since version 0.89c, which was
+ *    the first widely used release:
+ *
+ *    source                 png.h  png.h  shared-lib
+ *    version                string   int  version
+ *    -------                ------ -----  ----------
+ *    0.89c "1.0 beta 3"     0.89      89  1.0.89
+ *    0.90  "1.0 beta 4"     0.90      90  0.90  [should have been 2.0.90]
+ *    0.95  "1.0 beta 5"     0.95      95  0.95  [should have been 2.0.95]
+ *    0.96  "1.0 beta 6"     0.96      96  0.96  [should have been 2.0.96]
+ *    0.97b "1.00.97 beta 7" 1.00.97   97  1.0.1 [should have been 2.0.97]
+ *    0.97c                  0.97      97  2.0.97
+ *    0.98                   0.98      98  2.0.98
+ *    0.99                   0.99      98  2.0.99
+ *    0.99a-m                0.99      99  2.0.99
+ *    1.00                   1.00     100  2.1.0 [100 should be 10000]
+ *    1.0.0      (from here on, the   100  2.1.0 [100 should be 10000]
+ *    1.0.1       png.h string is   10001  2.1.0
+ *    1.0.1a-e    identical to the  10002  from here on, the shared library
+ *    1.0.2       source version)   10002  is 2.V where V is the source code
+ *    1.0.2a-b                      10003  version, except as noted.
+ *    1.0.3                         10003
+ *    1.0.3a-d                      10004
+ *    1.0.4                         10004
+ *    1.0.4a-f                      10005
+ *    1.0.5 (+ 2 patches)           10005
+ *    1.0.5a-d                      10006
+ *    1.0.5e-r                      10100 (not source compatible)
+ *    1.0.5s-v                      10006 (not binary compatible)
+ *    1.0.6 (+ 3 patches)           10006 (still binary incompatible)
+ *    1.0.6d-f                      10007 (still binary incompatible)
+ *    1.0.6g                        10007
+ *    1.0.6h                        10007  10.6h (testing xy.z so-numbering)
+ *    1.0.6i                        10007  10.6i
+ *    1.0.6j                        10007  2.1.0.6j (incompatible with 1.0.0)
+ *    1.0.7beta11-14        DLLNUM  10007  2.1.0.7beta11-14 (binary compatible)
+ *    1.0.7beta15-18           1    10007  2.1.0.7beta15-18 (binary compatible)
+ *    1.0.7rc1-2               1    10007  2.1.0.7rc1-2 (binary compatible)
+ *    1.0.7                    1    10007  (still compatible)
+ *    1.0.8beta1-4             1    10008  2.1.0.8beta1-4
+ *    1.0.8rc1                 1    10008  2.1.0.8rc1
+ *    1.0.8                    1    10008  2.1.0.8
+ *    1.0.9beta1-6             1    10009  2.1.0.9beta1-6
+ *    1.0.9rc1                 1    10009  2.1.0.9rc1
+ *    1.0.9beta7-10            1    10009  2.1.0.9beta7-10
+ *    1.0.9rc2                 1    10009  2.1.0.9rc2
+ *    1.0.9                    1    10009  2.1.0.9
+ *    1.0.10beta1              1    10010  2.1.0.10beta1
+ *    1.0.10rc1                1    10010  2.1.0.10rc1
+ *    1.0.10                   1    10010  2.1.0.10
+ *    1.0.11beta1-3            1    10011  2.1.0.11beta1-3
+ *    1.0.11rc1                1    10011  2.1.0.11rc1
+ *    1.0.11                   1    10011  2.1.0.11
+ *    1.0.12beta1-2            2    10012  2.1.0.12beta1-2
+ *    1.0.12rc1                2    10012  2.1.0.12rc1
+ *    1.0.12                   2    10012  2.1.0.12
+ *    1.1.0a-f                 -    10100  2.1.1.0a-f (branch abandoned)
+ *    1.2.0beta1-2             2    10200  2.1.2.0beta1-2
+ *    1.2.0beta3-5             3    10200  3.1.2.0beta3-5
+ *    1.2.0rc1                 3    10200  3.1.2.0rc1
+ *    1.2.0                    3    10200  3.1.2.0
+ *    1.2.1beta1-4             3    10201  3.1.2.1beta1-4
+ *    1.2.1rc1-2               3    10201  3.1.2.1rc1-2
+ *    1.2.1                    3    10201  3.1.2.1
+ *    1.2.2beta1-6            12    10202  12.so.0.1.2.2beta1-6
+ *    1.0.13beta1             10    10013  10.so.0.1.0.13beta1
+ *    1.0.13rc1               10    10013  10.so.0.1.0.13rc1
+ *    1.2.2rc1                12    10202  12.so.0.1.2.2rc1
+ *    1.0.13                  10    10013  10.so.0.1.0.13
+ *    1.2.2                   12    10202  12.so.0.1.2.2
+ *    1.2.3rc1-6              12    10203  12.so.0.1.2.3rc1-6
+ *    1.2.3                   12    10203  12.so.0.1.2.3
+ *    1.2.4beta1-3            13    10204  12.so.0.1.2.4beta1-3
+ *    1.0.14rc1               13    10014  10.so.0.1.0.14rc1
+ *    1.2.4rc1                13    10204  12.so.0.1.2.4rc1
+ *    1.0.14                  10    10014  10.so.0.1.0.14
+ *    1.2.4                   13    10204  12.so.0.1.2.4
+ *    1.2.5beta1-2            13    10205  12.so.0.1.2.5beta1-2
+ *    1.0.15rc1-3             10    10015  10.so.0.1.0.15rc1-3
+ *    1.2.5rc1-3              13    10205  12.so.0.1.2.5rc1-3
+ *    1.0.15                  10    10015  10.so.0.1.0.15
+ *    1.2.5                   13    10205  12.so.0.1.2.5
+ *    1.2.6beta1-4            13    10206  12.so.0.1.2.6beta1-4
+ *    1.0.16                  10    10016  10.so.0.1.0.16
+ *    1.2.6                   13    10206  12.so.0.1.2.6
+ *    1.2.7beta1-2            13    10207  12.so.0.1.2.7beta1-2
+ *    1.0.17rc1               10    10017  10.so.0.1.0.17rc1
+ *    1.2.7rc1                13    10207  12.so.0.1.2.7rc1
+ *    1.0.17                  10    10017  10.so.0.1.0.17
+ *    1.2.7                   13    10207  12.so.0.1.2.7
+ *    1.2.8beta1-5            13    10208  12.so.0.1.2.8beta1-5
+ *    1.0.18rc1-5             10    10018  10.so.0.1.0.18rc1-5
+ *    1.2.8rc1-5              13    10208  12.so.0.1.2.8rc1-5
+ *    1.0.18                  10    10018  10.so.0.1.0.18
+ *    1.2.8                   13    10208  12.so.0.1.2.8
+ *    1.2.9beta1-3            13    10209  12.so.0.1.2.9beta1-3
+ *    1.2.9beta4-11           13    10209  12.so.0.9[.0]
+ *    1.2.9rc1                13    10209  12.so.0.9[.0]
+ *    1.2.9                   13    10209  12.so.0.9[.0]
+ *    1.2.10beta1-8           13    10210  12.so.0.10[.0]
+ *    1.2.10rc1-3             13    10210  12.so.0.10[.0]
+ *    1.2.10                  13    10210  12.so.0.10[.0]
+ *    1.2.11beta1-4           13    10211  12.so.0.11[.0]
+ *    1.0.19rc1-5             10    10019  10.so.0.19[.0]
+ *    1.2.11rc1-5             13    10211  12.so.0.11[.0]
+ *    1.0.19                  10    10019  10.so.0.19[.0]
+ *    1.2.11                  13    10211  12.so.0.11[.0]
+ *    1.0.20                  10    10020  10.so.0.20[.0]
+ *    1.2.12                  13    10212  12.so.0.12[.0]
+ *    1.2.13beta1             13    10213  12.so.0.13[.0]
+ *    1.0.21                  10    10021  10.so.0.21[.0]
+ *    1.2.13                  13    10213  12.so.0.13[.0]
+ *    1.2.14beta1-2           13    10214  12.so.0.14[.0]
+ *    1.0.22rc1               10    10022  10.so.0.22[.0]
+ *    1.2.14rc1               13    10214  12.so.0.14[.0]
+ *    1.0.22                  10    10022  10.so.0.22[.0]
+ *    1.2.14                  13    10214  12.so.0.14[.0]
+ *    1.2.15beta1-6           13    10215  12.so.0.15[.0]
+ *    1.0.23rc1-5             10    10023  10.so.0.23[.0]
+ *    1.2.15rc1-5             13    10215  12.so.0.15[.0]
+ *    1.0.23                  10    10023  10.so.0.23[.0]
+ *    1.2.15                  13    10215  12.so.0.15[.0]
+ *    1.2.16beta1-2           13    10216  12.so.0.16[.0]
+ *    1.2.16rc1               13    10216  12.so.0.16[.0]
+ *    1.0.24                  10    10024  10.so.0.24[.0]
+ *    1.2.16                  13    10216  12.so.0.16[.0]
+ *    1.2.17beta1-2           13    10217  12.so.0.17[.0]
+ *    1.0.25rc1               10    10025  10.so.0.25[.0]
+ *    1.2.17rc1-3             13    10217  12.so.0.17[.0]
+ *    1.0.25                  10    10025  10.so.0.25[.0]
+ *    1.2.17                  13    10217  12.so.0.17[.0]
+ *    1.0.26                  10    10026  10.so.0.26[.0]
+ *    1.2.18                  13    10218  12.so.0.18[.0]
+ *    1.2.19beta1-31          13    10219  12.so.0.19[.0]
+ *    1.0.27rc1-6             10    10027  10.so.0.27[.0]
+ *    1.2.19rc1-6             13    10219  12.so.0.19[.0]
+ *    1.0.27                  10    10027  10.so.0.27[.0]
+ *    1.2.19                  13    10219  12.so.0.19[.0]
+ *
+ *    Henceforth the source version will match the shared-library major
+ *    and minor numbers; the shared-library major version number will be
+ *    used for changes in backward compatibility, as it is intended.  The
+ *    PNG_LIBPNG_VER macro, which is not used within libpng but is available
+ *    for applications, is an unsigned integer of the form xyyzz corresponding
+ *    to the source version x.y.z (leading zeros in y and z).  Beta versions
+ *    were given the previous public release number plus a letter, until
+ *    version 1.0.6j; from then on they were given the upcoming public
+ *    release number plus "betaNN" or "rcN".
+ *
+ *    Binary incompatibility exists only when applications make direct access
+ *    to the info_ptr or png_ptr members through png.h, and the compiled
+ *    application is loaded with a different version of the library.
+ *
+ *    DLLNUM will change each time there are forward or backward changes
+ *    in binary compatibility (e.g., when a new feature is added).
+ *
+ * See libpng.txt or libpng.3 for more information.  The PNG specification
+ * is available as a W3C Recommendation and as an ISO Specification,
+ * <http://www.w3.org/TR/2003/REC-PNG-20031110/
+ */
+
+/*
+ * COPYRIGHT NOTICE, DISCLAIMER, and LICENSE:
+ *
+ * If you modify libpng you may insert additional notices immediately following
+ * this sentence.
+ *
+ * libpng versions 1.2.6, August 15, 2004, through 1.2.19, August 18, 2007, are
+ * Copyright (c) 2004, 2006-2007 Glenn Randers-Pehrson, and are
+ * distributed according to the same disclaimer and license as libpng-1.2.5
+ * with the following individual added to the list of Contributing Authors:
+ *
+ *    Cosmin Truta
+ *
+ * libpng versions 1.0.7, July 1, 2000, through 1.2.5, October 3, 2002, are
+ * Copyright (c) 2000-2002 Glenn Randers-Pehrson, and are
+ * distributed according to the same disclaimer and license as libpng-1.0.6
+ * with the following individuals added to the list of Contributing Authors:
+ *
+ *    Simon-Pierre Cadieux
+ *    Eric S. Raymond
+ *    Gilles Vollant
+ *
+ * and with the following additions to the disclaimer:
+ *
+ *    There is no warranty against interference with your enjoyment of the
+ *    library or against infringement.  There is no warranty that our
+ *    efforts or the library will fulfill any of your particular purposes
+ *    or needs.  This library is provided with all faults, and the entire
+ *    risk of satisfactory quality, performance, accuracy, and effort is with
+ *    the user.
+ *
+ * libpng versions 0.97, January 1998, through 1.0.6, March 20, 2000, are
+ * Copyright (c) 1998, 1999, 2000 Glenn Randers-Pehrson, and are
+ * distributed according to the same disclaimer and license as libpng-0.96,
+ * with the following individuals added to the list of Contributing Authors:
+ *
+ *    Tom Lane
+ *    Glenn Randers-Pehrson
+ *    Willem van Schaik
+ *
+ * libpng versions 0.89, June 1996, through 0.96, May 1997, are
+ * Copyright (c) 1996, 1997 Andreas Dilger
+ * Distributed according to the same disclaimer and license as libpng-0.88,
+ * with the following individuals added to the list of Contributing Authors:
+ *
+ *    John Bowler
+ *    Kevin Bracey
+ *    Sam Bushell
+ *    Magnus Holmgren
+ *    Greg Roelofs
+ *    Tom Tanner
+ *
+ * libpng versions 0.5, May 1995, through 0.88, January 1996, are
+ * Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.
+ *
+ * For the purposes of this copyright and license, "Contributing Authors"
+ * is defined as the following set of individuals:
+ *
+ *    Andreas Dilger
+ *    Dave Martindale
+ *    Guy Eric Schalnat
+ *    Paul Schmidt
+ *    Tim Wegner
+ *
+ * The PNG Reference Library is supplied "AS IS".  The Contributing Authors
+ * and Group 42, Inc. disclaim all warranties, expressed or implied,
+ * including, without limitation, the warranties of merchantability and of
+ * fitness for any purpose.  The Contributing Authors and Group 42, Inc.
+ * assume no liability for direct, indirect, incidental, special, exemplary,
+ * or consequential damages, which may result from the use of the PNG
+ * Reference Library, even if advised of the possibility of such damage.
+ *
+ * Permission is hereby granted to use, copy, modify, and distribute this
+ * source code, or portions hereof, for any purpose, without fee, subject
+ * to the following restrictions:
+ *
+ * 1. The origin of this source code must not be misrepresented.
+ *
+ * 2. Altered versions must be plainly marked as such and
+ * must not be misrepresented as being the original source.
+ *
+ * 3. This Copyright notice may not be removed or altered from
+ *    any source or altered source distribution.
+ *
+ * The Contributing Authors and Group 42, Inc. specifically permit, without
+ * fee, and encourage the use of this source code as a component to
+ * supporting the PNG file format in commercial products.  If you use this
+ * source code in a product, acknowledgment is not required but would be
+ * appreciated.
+ */
+
+/*
+ * A "png_get_copyright" function is available, for convenient use in "about"
+ * boxes and the like:
+ *
+ * printf("%s",png_get_copyright(NULL));
+ *
+ * Also, the PNG logo (in PNG format, of course) is supplied in the
+ * files "pngbar.png" and "pngbar.jpg (88x31) and "pngnow.png" (98x31).
+ */
+
+/*
+ * Libpng is OSI Certified Open Source Software.  OSI Certified is a
+ * certification mark of the Open Source Initiative.
+ */
+
+/*
+ * The contributing authors would like to thank all those who helped
+ * with testing, bug fixes, and patience.  This wouldn't have been
+ * possible without all of you.
+ *
+ * Thanks to Frank J. T. Wojcik for helping with the documentation.
+ */
+
+/*
+ * Y2K compliance in libpng:
+ * =========================
+ *
+ *    August 18, 2007
+ *
+ *    Since the PNG Development group is an ad-hoc body, we can't make
+ *    an official declaration.
+ *
+ *    This is your unofficial assurance that libpng from version 0.71 and
+ *    upward through 1.2.19 are Y2K compliant.  It is my belief that earlier
+ *    versions were also Y2K compliant.
+ *
+ *    Libpng only has three year fields.  One is a 2-byte unsigned integer
+ *    that will hold years up to 65535.  The other two hold the date in text
+ *    format, and will hold years up to 9999.
+ *
+ *    The integer is
+ *        "png_uint_16 year" in png_time_struct.
+ *
+ *    The strings are
+ *        "png_charp time_buffer" in png_struct and
+ *        "near_time_buffer", which is a local character string in png.c.
+ *
+ *    There are seven time-related functions:
+ *        png.c: png_convert_to_rfc_1123() in png.c
+ *          (formerly png_convert_to_rfc_1152() in error)
+ *        png_convert_from_struct_tm() in pngwrite.c, called in pngwrite.c
+ *        png_convert_from_time_t() in pngwrite.c
+ *        png_get_tIME() in pngget.c
+ *        png_handle_tIME() in pngrutil.c, called in pngread.c
+ *        png_set_tIME() in pngset.c
+ *        png_write_tIME() in pngwutil.c, called in pngwrite.c
+ *
+ *    All handle dates properly in a Y2K environment.  The
+ *    png_convert_from_time_t() function calls gmtime() to convert from system
+ *    clock time, which returns (year - 1900), which we properly convert to
+ *    the full 4-digit year.  There is a possibility that applications using
+ *    libpng are not passing 4-digit years into the png_convert_to_rfc_1123()
+ *    function, or that they are incorrectly passing only a 2-digit year
+ *    instead of "year - 1900" into the png_convert_from_struct_tm() function,
+ *    but this is not under our control.  The libpng documentation has always
+ *    stated that it works with 4-digit years, and the APIs have been
+ *    documented as such.
+ *
+ *    The tIME chunk itself is also Y2K compliant.  It uses a 2-byte unsigned
+ *    integer to hold the year, and can hold years as large as 65535.
+ *
+ *    zlib, upon which libpng depends, is also Y2K compliant.  It contains
+ *    no date-related code.
+ *
+ *       Glenn Randers-Pehrson
+ *       libpng maintainer
+ *       PNG Development Group
+ */
+
+#ifndef PNG_H
+#define PNG_H
+
+/* This is not the place to learn how to use libpng.  The file libpng.txt
+ * describes how to use libpng, and the file example.c summarizes it
+ * with some code on which to build.  This file is useful for looking
+ * at the actual function definitions and structure components.
+ */
+
+/* Version information for png.h - this should match the version in png.c */
+#define PNG_LIBPNG_VER_STRING "1.2.19"
+#define PNG_HEADER_VERSION_STRING \
+   " libpng version 1.2.19 - August 18, 2007\n"
+
+#define PNG_LIBPNG_VER_SONUM   0
+#define PNG_LIBPNG_VER_DLLNUM  13
+
+/* These should match the first 3 components of PNG_LIBPNG_VER_STRING: */
+#define PNG_LIBPNG_VER_MAJOR   1
+#define PNG_LIBPNG_VER_MINOR   2
+#define PNG_LIBPNG_VER_RELEASE 19
+/* This should match the numeric part of the final component of
+ * PNG_LIBPNG_VER_STRING, omitting any leading zero: */
+
+#define PNG_LIBPNG_VER_BUILD  0
+
+/* Release Status */
+#define PNG_LIBPNG_BUILD_ALPHA    1
+#define PNG_LIBPNG_BUILD_BETA     2
+#define PNG_LIBPNG_BUILD_RC       3
+#define PNG_LIBPNG_BUILD_STABLE   4
+#define PNG_LIBPNG_BUILD_RELEASE_STATUS_MASK 7
+  
+/* Release-Specific Flags */
+#define PNG_LIBPNG_BUILD_PATCH    8 /* Can be OR'ed with
+                                       PNG_LIBPNG_BUILD_STABLE only */
+#define PNG_LIBPNG_BUILD_PRIVATE 16 /* Cannot be OR'ed with
+                                       PNG_LIBPNG_BUILD_SPECIAL */
+#define PNG_LIBPNG_BUILD_SPECIAL 32 /* Cannot be OR'ed with
+                                       PNG_LIBPNG_BUILD_PRIVATE */
+
+#define PNG_LIBPNG_BUILD_BASE_TYPE PNG_LIBPNG_BUILD_STABLE
+
+/* Careful here.  At one time, Guy wanted to use 082, but that would be octal.
+ * We must not include leading zeros.
+ * Versions 0.7 through 1.0.0 were in the range 0 to 100 here (only
+ * version 1.0.0 was mis-numbered 100 instead of 10000).  From
+ * version 1.0.1 it's    xxyyzz, where x=major, y=minor, z=release */
+#define PNG_LIBPNG_VER 10219 /* 1.2.19 */
+
+#ifndef PNG_VERSION_INFO_ONLY
+/* include the compression library's header */
+#include "zlib.h"
+#endif
+
+/* include all user configurable info, including optional assembler routines */
+#include "pngconf.h"
+
+/*
+ * Added at libpng-1.2.8 */
+/* Ref MSDN: Private as priority over Special
+ * VS_FF_PRIVATEBUILD File *was not* built using standard release
+ * procedures. If this value is given, the StringFileInfo block must
+ * contain a PrivateBuild string. 
+ *
+ * VS_FF_SPECIALBUILD File *was* built by the original company using
+ * standard release procedures but is a variation of the standard
+ * file of the same version number. If this value is given, the
+ * StringFileInfo block must contain a SpecialBuild string. 
+ */
+
+#if defined(PNG_USER_PRIVATEBUILD)
+#  define PNG_LIBPNG_BUILD_TYPE \
+          (PNG_LIBPNG_BUILD_BASE_TYPE | PNG_LIBPNG_BUILD_PRIVATE)
+#else
+#  if defined(PNG_LIBPNG_SPECIALBUILD)
+#    define PNG_LIBPNG_BUILD_TYPE \
+            (PNG_LIBPNG_BUILD_BASE_TYPE | PNG_LIBPNG_BUILD_SPECIAL)
+#  else
+#    define PNG_LIBPNG_BUILD_TYPE (PNG_LIBPNG_BUILD_BASE_TYPE)
+#  endif
+#endif
+
+#ifndef PNG_VERSION_INFO_ONLY
+
+/* Inhibit C++ name-mangling for libpng functions but not for system calls. */
+#ifdef __cplusplus
+extern "C" {
+#endif /* __cplusplus */
+
+/* This file is arranged in several sections.  The first section contains
+ * structure and type definitions.  The second section contains the external
+ * library functions, while the third has the internal library functions,
+ * which applications aren't expected to use directly.
+ */
+
+#ifndef PNG_NO_TYPECAST_NULL
+#define int_p_NULL                (int *)NULL
+#define png_bytep_NULL            (png_bytep)NULL
+#define png_bytepp_NULL           (png_bytepp)NULL
+#define png_doublep_NULL          (png_doublep)NULL
+#define png_error_ptr_NULL        (png_error_ptr)NULL
+#define png_flush_ptr_NULL        (png_flush_ptr)NULL
+#define png_free_ptr_NULL         (png_free_ptr)NULL
+#define png_infopp_NULL           (png_infopp)NULL
+#define png_malloc_ptr_NULL       (png_malloc_ptr)NULL
+#define png_read_status_ptr_NULL  (png_read_status_ptr)NULL
+#define png_rw_ptr_NULL           (png_rw_ptr)NULL
+#define png_structp_NULL          (png_structp)NULL
+#define png_uint_16p_NULL         (png_uint_16p)NULL
+#define png_voidp_NULL            (png_voidp)NULL
+#define png_write_status_ptr_NULL (png_write_status_ptr)NULL
+#else
+#define int_p_NULL                NULL
+#define png_bytep_NULL            NULL
+#define png_bytepp_NULL           NULL
+#define png_doublep_NULL          NULL
+#define png_error_ptr_NULL        NULL
+#define png_flush_ptr_NULL        NULL
+#define png_free_ptr_NULL         NULL
+#define png_infopp_NULL           NULL
+#define png_malloc_ptr_NULL       NULL
+#define png_read_status_ptr_NULL  NULL
+#define png_rw_ptr_NULL           NULL
+#define png_structp_NULL          NULL
+#define png_uint_16p_NULL         NULL
+#define png_voidp_NULL            NULL
+#define png_write_status_ptr_NULL NULL
+#endif
+
+/* variables declared in png.c - only it needs to define PNG_NO_EXTERN */
+#if !defined(PNG_NO_EXTERN) || defined(PNG_ALWAYS_EXTERN)
+/* Version information for C files, stored in png.c.  This had better match
+ * the version above.
+ */
+#ifdef PNG_USE_GLOBAL_ARRAYS
+PNG_EXPORT_VAR (PNG_CONST char) png_libpng_ver[18];
+  /* need room for 99.99.99beta99z */
+#else
+#define png_libpng_ver png_get_header_ver(NULL)
+#endif
+
+#ifdef PNG_USE_GLOBAL_ARRAYS
+/* This was removed in version 1.0.5c */
+/* Structures to facilitate easy interlacing.  See png.c for more details */
+PNG_EXPORT_VAR (PNG_CONST int FARDATA) png_pass_start[7];
+PNG_EXPORT_VAR (PNG_CONST int FARDATA) png_pass_inc[7];
+PNG_EXPORT_VAR (PNG_CONST int FARDATA) png_pass_ystart[7];
+PNG_EXPORT_VAR (PNG_CONST int FARDATA) png_pass_yinc[7];
+PNG_EXPORT_VAR (PNG_CONST int FARDATA) png_pass_mask[7];
+PNG_EXPORT_VAR (PNG_CONST int FARDATA) png_pass_dsp_mask[7];
+#if defined(PNG_HAVE_MMX_COMBINE_ROW) || defined(PNG_OPTIMIZED_CODE_SUPPORTED)
+PNG_EXPORT_VAR (PNG_CONST int FARDATA) png_pass_width[7];
+#endif
+/* This isn't currently used.  If you need it, see png.c for more details.
+PNG_EXPORT_VAR (PNG_CONST int FARDATA) png_pass_height[7];
+*/
+#endif
+
+#endif /* PNG_NO_EXTERN */
+
+/* Three color definitions.  The order of the red, green, and blue, (and the
+ * exact size) is not important, although the size of the fields need to
+ * be png_byte or png_uint_16 (as defined below).
+ */
+typedef struct png_color_struct
+{
+   png_byte red;
+   png_byte green;
+   png_byte blue;
+} png_color;
+typedef png_color FAR * png_colorp;
+typedef png_color FAR * FAR * png_colorpp;
+
+typedef struct png_color_16_struct
+{
+   png_byte index;    /* used for palette files */
+   png_uint_16 red;   /* for use in red green blue files */
+   png_uint_16 green;
+   png_uint_16 blue;
+   png_uint_16 gray;  /* for use in grayscale files */
+} png_color_16;
+typedef png_color_16 FAR * png_color_16p;
+typedef png_color_16 FAR * FAR * png_color_16pp;
+
+typedef struct png_color_8_struct
+{
+   png_byte red;   /* for use in red green blue files */
+   png_byte green;
+   png_byte blue;
+   png_byte gray;  /* for use in grayscale files */
+   png_byte alpha; /* for alpha channel files */
+} png_color_8;
+typedef png_color_8 FAR * png_color_8p;
+typedef png_color_8 FAR * FAR * png_color_8pp;
+
+/*
+ * The following two structures are used for the in-core representation
+ * of sPLT chunks.
+ */
+typedef struct png_sPLT_entry_struct
+{
+   png_uint_16 red;
+   png_uint_16 green;
+   png_uint_16 blue;
+   png_uint_16 alpha;
+   png_uint_16 frequency;
+} png_sPLT_entry;
+typedef png_sPLT_entry FAR * png_sPLT_entryp;
+typedef png_sPLT_entry FAR * FAR * png_sPLT_entrypp;
+
+/*  When the depth of the sPLT palette is 8 bits, the color and alpha samples
+ *  occupy the LSB of their respective members, and the MSB of each member
+ *  is zero-filled.  The frequency member always occupies the full 16 bits.
+ */
+
+typedef struct png_sPLT_struct
+{
+   png_charp name;           /* palette name */
+   png_byte depth;           /* depth of palette samples */
+   png_sPLT_entryp entries;  /* palette entries */
+   png_int_32 nentries;      /* number of palette entries */
+} png_sPLT_t;
+typedef png_sPLT_t FAR * png_sPLT_tp;
+typedef png_sPLT_t FAR * FAR * png_sPLT_tpp;
+
+#ifdef PNG_TEXT_SUPPORTED
+/* png_text holds the contents of a text/ztxt/itxt chunk in a PNG file,
+ * and whether that contents is compressed or not.  The "key" field
+ * points to a regular zero-terminated C string.  The "text", "lang", and
+ * "lang_key" fields can be regular C strings, empty strings, or NULL pointers.
+ * However, the * structure returned by png_get_text() will always contain
+ * regular zero-terminated C strings (possibly empty), never NULL pointers,
+ * so they can be safely used in printf() and other string-handling functions.
+ */
+typedef struct png_text_struct
+{
+   int  compression;       /* compression value:
+                             -1: tEXt, none
+                              0: zTXt, deflate
+                              1: iTXt, none
+                              2: iTXt, deflate  */
+   png_charp key;          /* keyword, 1-79 character description of "text" */
+   png_charp text;         /* comment, may be an empty string (ie "")
+                              or a NULL pointer */
+   png_size_t text_length; /* length of the text string */
+#ifdef PNG_iTXt_SUPPORTED
+   png_size_t itxt_length; /* length of the itxt string */
+   png_charp lang;         /* language code, 0-79 characters
+                              or a NULL pointer */
+   png_charp lang_key;     /* keyword translated UTF-8 string, 0 or more
+                              chars or a NULL pointer */
+#endif
+} png_text;
+typedef png_text FAR * png_textp;
+typedef png_text FAR * FAR * png_textpp;
+#endif
+
+/* Supported compression types for text in PNG files (tEXt, and zTXt).
+ * The values of the PNG_TEXT_COMPRESSION_ defines should NOT be changed. */
+#define PNG_TEXT_COMPRESSION_NONE_WR -3
+#define PNG_TEXT_COMPRESSION_zTXt_WR -2
+#define PNG_TEXT_COMPRESSION_NONE    -1
+#define PNG_TEXT_COMPRESSION_zTXt     0
+#define PNG_ITXT_COMPRESSION_NONE     1
+#define PNG_ITXT_COMPRESSION_zTXt     2
+#define PNG_TEXT_COMPRESSION_LAST     3  /* Not a valid value */
+
+/* png_time is a way to hold the time in an machine independent way.
+ * Two conversions are provided, both from time_t and struct tm.  There
+ * is no portable way to convert to either of these structures, as far
+ * as I know.  If you know of a portable way, send it to me.  As a side
+ * note - PNG has always been Year 2000 compliant!
+ */
+typedef struct png_time_struct
+{
+   png_uint_16 year; /* full year, as in, 1995 */
+   png_byte month;   /* month of year, 1 - 12 */
+   png_byte day;     /* day of month, 1 - 31 */
+   png_byte hour;    /* hour of day, 0 - 23 */
+   png_byte minute;  /* minute of hour, 0 - 59 */
+   png_byte second;  /* second of minute, 0 - 60 (for leap seconds) */
+} png_time;
+typedef png_time FAR * png_timep;
+typedef png_time FAR * FAR * png_timepp;
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+/* png_unknown_chunk is a structure to hold queued chunks for which there is
+ * no specific support.  The idea is that we can use this to queue
+ * up private chunks for output even though the library doesn't actually
+ * know about their semantics.
+ */
+typedef struct png_unknown_chunk_t
+{
+    png_byte name[5];
+    png_byte *data;
+    png_size_t size;
+
+    /* libpng-using applications should NOT directly modify this byte. */
+    png_byte location; /* mode of operation at read time */
+}
+png_unknown_chunk;
+typedef png_unknown_chunk FAR * png_unknown_chunkp;
+typedef png_unknown_chunk FAR * FAR * png_unknown_chunkpp;
+#endif
+
+/* png_info is a structure that holds the information in a PNG file so
+ * that the application can find out the characteristics of the image.
+ * If you are reading the file, this structure will tell you what is
+ * in the PNG file.  If you are writing the file, fill in the information
+ * you want to put into the PNG file, then call png_write_info().
+ * The names chosen should be very close to the PNG specification, so
+ * consult that document for information about the meaning of each field.
+ *
+ * With libpng < 0.95, it was only possible to directly set and read the
+ * the values in the png_info_struct, which meant that the contents and
+ * order of the values had to remain fixed.  With libpng 0.95 and later,
+ * however, there are now functions that abstract the contents of
+ * png_info_struct from the application, so this makes it easier to use
+ * libpng with dynamic libraries, and even makes it possible to use
+ * libraries that don't have all of the libpng ancillary chunk-handing
+ * functionality.
+ *
+ * In any case, the order of the parameters in png_info_struct should NOT
+ * be changed for as long as possible to keep compatibility with applications
+ * that use the old direct-access method with png_info_struct.
+ *
+ * The following members may have allocated storage attached that should be
+ * cleaned up before the structure is discarded: palette, trans, text,
+ * pcal_purpose, pcal_units, pcal_params, hist, iccp_name, iccp_profile,
+ * splt_palettes, scal_unit, row_pointers, and unknowns.   By default, these
+ * are automatically freed when the info structure is deallocated, if they were
+ * allocated internally by libpng.  This behavior can be changed by means
+ * of the png_data_freer() function.
+ *
+ * More allocation details: all the chunk-reading functions that
+ * change these members go through the corresponding png_set_*
+ * functions.  A function to clear these members is available: see
+ * png_free_data().  The png_set_* functions do not depend on being
+ * able to point info structure members to any of the storage they are
+ * passed (they make their own copies), EXCEPT that the png_set_text
+ * functions use the same storage passed to them in the text_ptr or
+ * itxt_ptr structure argument, and the png_set_rows and png_set_unknowns
+ * functions do not make their own copies.
+ */
+typedef struct png_info_struct
+{
+   /* the following are necessary for every PNG file */
+   png_uint_32 width;       /* width of image in pixels (from IHDR) */
+   png_uint_32 height;      /* height of image in pixels (from IHDR) */
+   png_uint_32 valid;       /* valid chunk data (see PNG_INFO_ below) */
+   png_uint_32 rowbytes;    /* bytes needed to hold an untransformed row */
+   png_colorp palette;      /* array of color values (valid & PNG_INFO_PLTE) */
+   png_uint_16 num_palette; /* number of color entries in "palette" (PLTE) */
+   png_uint_16 num_trans;   /* number of transparent palette color (tRNS) */
+   png_byte bit_depth;      /* 1, 2, 4, 8, or 16 bits/channel (from IHDR) */
+   png_byte color_type;     /* see PNG_COLOR_TYPE_ below (from IHDR) */
+   /* The following three should have been named *_method not *_type */
+   png_byte compression_type; /* must be PNG_COMPRESSION_TYPE_BASE (IHDR) */
+   png_byte filter_type;    /* must be PNG_FILTER_TYPE_BASE (from IHDR) */
+   png_byte interlace_type; /* One of PNG_INTERLACE_NONE, PNG_INTERLACE_ADAM7 */
+
+   /* The following is informational only on read, and not used on writes. */
+   png_byte channels;       /* number of data channels per pixel (1, 2, 3, 4) */
+   png_byte pixel_depth;    /* number of bits per pixel */
+   png_byte spare_byte;     /* to align the data, and for future use */
+   png_byte signature[8];   /* magic bytes read by libpng from start of file */
+
+   /* The rest of the data is optional.  If you are reading, check the
+    * valid field to see if the information in these are valid.  If you
+    * are writing, set the valid field to those chunks you want written,
+    * and initialize the appropriate fields below.
+    */
+
+#if defined(PNG_gAMA_SUPPORTED) && defined(PNG_FLOATING_POINT_SUPPORTED)
+   /* The gAMA chunk describes the gamma characteristics of the system
+    * on which the image was created, normally in the range [1.0, 2.5].
+    * Data is valid if (valid & PNG_INFO_gAMA) is non-zero.
+    */
+   float gamma; /* gamma value of image, if (valid & PNG_INFO_gAMA) */
+#endif
+
+#if defined(PNG_sRGB_SUPPORTED)
+    /* GR-P, 0.96a */
+    /* Data valid if (valid & PNG_INFO_sRGB) non-zero. */
+   png_byte srgb_intent; /* sRGB rendering intent [0, 1, 2, or 3] */
+#endif
+
+#if defined(PNG_TEXT_SUPPORTED)
+   /* The tEXt, and zTXt chunks contain human-readable textual data in
+    * uncompressed, compressed, and optionally compressed forms, respectively.
+    * The data in "text" is an array of pointers to uncompressed,
+    * null-terminated C strings. Each chunk has a keyword that describes the
+    * textual data contained in that chunk.  Keywords are not required to be
+    * unique, and the text string may be empty.  Any number of text chunks may
+    * be in an image.
+    */
+   int num_text; /* number of comments read/to write */
+   int max_text; /* current size of text array */
+   png_textp text; /* array of comments read/to write */
+#endif /* PNG_TEXT_SUPPORTED */
+
+#if defined(PNG_tIME_SUPPORTED)
+   /* The tIME chunk holds the last time the displayed image data was
+    * modified.  See the png_time struct for the contents of this struct.
+    */
+   png_time mod_time;
+#endif
+
+#if defined(PNG_sBIT_SUPPORTED)
+   /* The sBIT chunk specifies the number of significant high-order bits
+    * in the pixel data.  Values are in the range [1, bit_depth], and are
+    * only specified for the channels in the pixel data.  The contents of
+    * the low-order bits is not specified.  Data is valid if
+    * (valid & PNG_INFO_sBIT) is non-zero.
+    */
+   png_color_8 sig_bit; /* significant bits in color channels */
+#endif
+
+#if defined(PNG_tRNS_SUPPORTED) || defined(PNG_READ_EXPAND_SUPPORTED) || \
+defined(PNG_READ_BACKGROUND_SUPPORTED)
+   /* The tRNS chunk supplies transparency data for paletted images and
+    * other image types that don't need a full alpha channel.  There are
+    * "num_trans" transparency values for a paletted image, stored in the
+    * same order as the palette colors, starting from index 0.  Values
+    * for the data are in the range [0, 255], ranging from fully transparent
+    * to fully opaque, respectively.  For non-paletted images, there is a
+    * single color specified that should be treated as fully transparent.
+    * Data is valid if (valid & PNG_INFO_tRNS) is non-zero.
+    */
+   png_bytep trans; /* transparent values for paletted image */
+   png_color_16 trans_values; /* transparent color for non-palette image */
+#endif
+
+#if defined(PNG_bKGD_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+   /* The bKGD chunk gives the suggested image background color if the
+    * display program does not have its own background color and the image
+    * is needs to composited onto a background before display.  The colors
+    * in "background" are normally in the same color space/depth as the
+    * pixel data.  Data is valid if (valid & PNG_INFO_bKGD) is non-zero.
+    */
+   png_color_16 background;
+#endif
+
+#if defined(PNG_oFFs_SUPPORTED)
+   /* The oFFs chunk gives the offset in "offset_unit_type" units rightwards
+    * and downwards from the top-left corner of the display, page, or other
+    * application-specific co-ordinate space.  See the PNG_OFFSET_ defines
+    * below for the unit types.  Valid if (valid & PNG_INFO_oFFs) non-zero.
+    */
+   png_int_32 x_offset; /* x offset on page */
+   png_int_32 y_offset; /* y offset on page */
+   png_byte offset_unit_type; /* offset units type */
+#endif
+
+#if defined(PNG_pHYs_SUPPORTED)
+   /* The pHYs chunk gives the physical pixel density of the image for
+    * display or printing in "phys_unit_type" units (see PNG_RESOLUTION_
+    * defines below).  Data is valid if (valid & PNG_INFO_pHYs) is non-zero.
+    */
+   png_uint_32 x_pixels_per_unit; /* horizontal pixel density */
+   png_uint_32 y_pixels_per_unit; /* vertical pixel density */
+   png_byte phys_unit_type; /* resolution type (see PNG_RESOLUTION_ below) */
+#endif
+
+#if defined(PNG_hIST_SUPPORTED)
+   /* The hIST chunk contains the relative frequency or importance of the
+    * various palette entries, so that a viewer can intelligently select a
+    * reduced-color palette, if required.  Data is an array of "num_palette"
+    * values in the range [0,65535]. Data valid if (valid & PNG_INFO_hIST)
+    * is non-zero.
+    */
+   png_uint_16p hist;
+#endif
+
+#ifdef PNG_cHRM_SUPPORTED
+   /* The cHRM chunk describes the CIE color characteristics of the monitor
+    * on which the PNG was created.  This data allows the viewer to do gamut
+    * mapping of the input image to ensure that the viewer sees the same
+    * colors in the image as the creator.  Values are in the range
+    * [0.0, 0.8].  Data valid if (valid & PNG_INFO_cHRM) non-zero.
+    */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   float x_white;
+   float y_white;
+   float x_red;
+   float y_red;
+   float x_green;
+   float y_green;
+   float x_blue;
+   float y_blue;
+#endif
+#endif
+
+#if defined(PNG_pCAL_SUPPORTED)
+   /* The pCAL chunk describes a transformation between the stored pixel
+    * values and original physical data values used to create the image.
+    * The integer range [0, 2^bit_depth - 1] maps to the floating-point
+    * range given by [pcal_X0, pcal_X1], and are further transformed by a
+    * (possibly non-linear) transformation function given by "pcal_type"
+    * and "pcal_params" into "pcal_units".  Please see the PNG_EQUATION_
+    * defines below, and the PNG-Group's PNG extensions document for a
+    * complete description of the transformations and how they should be
+    * implemented, and for a description of the ASCII parameter strings.
+    * Data values are valid if (valid & PNG_INFO_pCAL) non-zero.
+    */
+   png_charp pcal_purpose;  /* pCAL chunk description string */
+   png_int_32 pcal_X0;      /* minimum value */
+   png_int_32 pcal_X1;      /* maximum value */
+   png_charp pcal_units;    /* Latin-1 string giving physical units */
+   png_charpp pcal_params;  /* ASCII strings containing parameter values */
+   png_byte pcal_type;      /* equation type (see PNG_EQUATION_ below) */
+   png_byte pcal_nparams;   /* number of parameters given in pcal_params */
+#endif
+
+/* New members added in libpng-1.0.6 */
+#ifdef PNG_FREE_ME_SUPPORTED
+   png_uint_32 free_me;     /* flags items libpng is responsible for freeing */
+#endif
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+   /* storage for unknown chunks that the library doesn't recognize. */
+   png_unknown_chunkp unknown_chunks;
+   png_size_t unknown_chunks_num;
+#endif
+
+#if defined(PNG_iCCP_SUPPORTED)
+   /* iCCP chunk data. */
+   png_charp iccp_name;     /* profile name */
+   png_charp iccp_profile;  /* International Color Consortium profile data */
+                            /* Note to maintainer: should be png_bytep */
+   png_uint_32 iccp_proflen;  /* ICC profile data length */
+   png_byte iccp_compression; /* Always zero */
+#endif
+
+#if defined(PNG_sPLT_SUPPORTED)
+   /* data on sPLT chunks (there may be more than one). */
+   png_sPLT_tp splt_palettes;
+   png_uint_32 splt_palettes_num;
+#endif
+
+#if defined(PNG_sCAL_SUPPORTED)
+   /* The sCAL chunk describes the actual physical dimensions of the
+    * subject matter of the graphic.  The chunk contains a unit specification
+    * a byte value, and two ASCII strings representing floating-point
+    * values.  The values are width and height corresponsing to one pixel
+    * in the image.  This external representation is converted to double
+    * here.  Data values are valid if (valid & PNG_INFO_sCAL) is non-zero.
+    */
+   png_byte scal_unit;         /* unit of physical scale */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   double scal_pixel_width;    /* width of one pixel */
+   double scal_pixel_height;   /* height of one pixel */
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   png_charp scal_s_width;     /* string containing height */
+   png_charp scal_s_height;    /* string containing width */
+#endif
+#endif
+
+#if defined(PNG_INFO_IMAGE_SUPPORTED)
+   /* Memory has been allocated if (valid & PNG_ALLOCATED_INFO_ROWS) non-zero */
+   /* Data valid if (valid & PNG_INFO_IDAT) non-zero */
+   png_bytepp row_pointers;        /* the image bits */
+#endif
+
+#if defined(PNG_FIXED_POINT_SUPPORTED) && defined(PNG_gAMA_SUPPORTED)
+   png_fixed_point int_gamma; /* gamma of image, if (valid & PNG_INFO_gAMA) */
+#endif
+
+#if defined(PNG_cHRM_SUPPORTED) && defined(PNG_FIXED_POINT_SUPPORTED)
+   png_fixed_point int_x_white;
+   png_fixed_point int_y_white;
+   png_fixed_point int_x_red;
+   png_fixed_point int_y_red;
+   png_fixed_point int_x_green;
+   png_fixed_point int_y_green;
+   png_fixed_point int_x_blue;
+   png_fixed_point int_y_blue;
+#endif
+
+} png_info;
+
+typedef png_info FAR * png_infop;
+typedef png_info FAR * FAR * png_infopp;
+
+/* Maximum positive integer used in PNG is (2^31)-1 */
+#define PNG_UINT_31_MAX ((png_uint_32)0x7fffffffL)
+#define PNG_UINT_32_MAX ((png_uint_32)(-1))
+#define PNG_SIZE_MAX ((png_size_t)(-1))
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+/* PNG_MAX_UINT is deprecated; use PNG_UINT_31_MAX instead. */
+#define PNG_MAX_UINT PNG_UINT_31_MAX
+#endif
+
+/* These describe the color_type field in png_info. */
+/* color type masks */
+#define PNG_COLOR_MASK_PALETTE    1
+#define PNG_COLOR_MASK_COLOR      2
+#define PNG_COLOR_MASK_ALPHA      4
+
+/* color types.  Note that not all combinations are legal */
+#define PNG_COLOR_TYPE_GRAY 0
+#define PNG_COLOR_TYPE_PALETTE  (PNG_COLOR_MASK_COLOR | PNG_COLOR_MASK_PALETTE)
+#define PNG_COLOR_TYPE_RGB        (PNG_COLOR_MASK_COLOR)
+#define PNG_COLOR_TYPE_RGB_ALPHA  (PNG_COLOR_MASK_COLOR | PNG_COLOR_MASK_ALPHA)
+#define PNG_COLOR_TYPE_GRAY_ALPHA (PNG_COLOR_MASK_ALPHA)
+/* aliases */
+#define PNG_COLOR_TYPE_RGBA  PNG_COLOR_TYPE_RGB_ALPHA
+#define PNG_COLOR_TYPE_GA  PNG_COLOR_TYPE_GRAY_ALPHA
+
+/* This is for compression type. PNG 1.0-1.2 only define the single type. */
+#define PNG_COMPRESSION_TYPE_BASE 0 /* Deflate method 8, 32K window */
+#define PNG_COMPRESSION_TYPE_DEFAULT PNG_COMPRESSION_TYPE_BASE
+
+/* This is for filter type. PNG 1.0-1.2 only define the single type. */
+#define PNG_FILTER_TYPE_BASE      0 /* Single row per-byte filtering */
+#define PNG_INTRAPIXEL_DIFFERENCING 64 /* Used only in MNG datastreams */
+#define PNG_FILTER_TYPE_DEFAULT   PNG_FILTER_TYPE_BASE
+
+/* These are for the interlacing type.  These values should NOT be changed. */
+#define PNG_INTERLACE_NONE        0 /* Non-interlaced image */
+#define PNG_INTERLACE_ADAM7       1 /* Adam7 interlacing */
+#define PNG_INTERLACE_LAST        2 /* Not a valid value */
+
+/* These are for the oFFs chunk.  These values should NOT be changed. */
+#define PNG_OFFSET_PIXEL          0 /* Offset in pixels */
+#define PNG_OFFSET_MICROMETER     1 /* Offset in micrometers (1/10^6 meter) */
+#define PNG_OFFSET_LAST           2 /* Not a valid value */
+
+/* These are for the pCAL chunk.  These values should NOT be changed. */
+#define PNG_EQUATION_LINEAR       0 /* Linear transformation */
+#define PNG_EQUATION_BASE_E       1 /* Exponential base e transform */
+#define PNG_EQUATION_ARBITRARY    2 /* Arbitrary base exponential transform */
+#define PNG_EQUATION_HYPERBOLIC   3 /* Hyperbolic sine transformation */
+#define PNG_EQUATION_LAST         4 /* Not a valid value */
+
+/* These are for the sCAL chunk.  These values should NOT be changed. */
+#define PNG_SCALE_UNKNOWN         0 /* unknown unit (image scale) */
+#define PNG_SCALE_METER           1 /* meters per pixel */
+#define PNG_SCALE_RADIAN          2 /* radians per pixel */
+#define PNG_SCALE_LAST            3 /* Not a valid value */
+
+/* These are for the pHYs chunk.  These values should NOT be changed. */
+#define PNG_RESOLUTION_UNKNOWN    0 /* pixels/unknown unit (aspect ratio) */
+#define PNG_RESOLUTION_METER      1 /* pixels/meter */
+#define PNG_RESOLUTION_LAST       2 /* Not a valid value */
+
+/* These are for the sRGB chunk.  These values should NOT be changed. */
+#define PNG_sRGB_INTENT_PERCEPTUAL 0
+#define PNG_sRGB_INTENT_RELATIVE   1
+#define PNG_sRGB_INTENT_SATURATION 2
+#define PNG_sRGB_INTENT_ABSOLUTE   3
+#define PNG_sRGB_INTENT_LAST       4 /* Not a valid value */
+
+/* This is for text chunks */
+#define PNG_KEYWORD_MAX_LENGTH     79
+
+/* Maximum number of entries in PLTE/sPLT/tRNS arrays */
+#define PNG_MAX_PALETTE_LENGTH    256
+
+/* These determine if an ancillary chunk's data has been successfully read
+ * from the PNG header, or if the application has filled in the corresponding
+ * data in the info_struct to be written into the output file.  The values
+ * of the PNG_INFO_<chunk> defines should NOT be changed.
+ */
+#define PNG_INFO_gAMA 0x0001
+#define PNG_INFO_sBIT 0x0002
+#define PNG_INFO_cHRM 0x0004
+#define PNG_INFO_PLTE 0x0008
+#define PNG_INFO_tRNS 0x0010
+#define PNG_INFO_bKGD 0x0020
+#define PNG_INFO_hIST 0x0040
+#define PNG_INFO_pHYs 0x0080
+#define PNG_INFO_oFFs 0x0100
+#define PNG_INFO_tIME 0x0200
+#define PNG_INFO_pCAL 0x0400
+#define PNG_INFO_sRGB 0x0800   /* GR-P, 0.96a */
+#define PNG_INFO_iCCP 0x1000   /* ESR, 1.0.6 */
+#define PNG_INFO_sPLT 0x2000   /* ESR, 1.0.6 */
+#define PNG_INFO_sCAL 0x4000   /* ESR, 1.0.6 */
+#define PNG_INFO_IDAT 0x8000L  /* ESR, 1.0.6 */
+
+/* This is used for the transformation routines, as some of them
+ * change these values for the row.  It also should enable using
+ * the routines for other purposes.
+ */
+typedef struct png_row_info_struct
+{
+   png_uint_32 width; /* width of row */
+   png_uint_32 rowbytes; /* number of bytes in row */
+   png_byte color_type; /* color type of row */
+   png_byte bit_depth; /* bit depth of row */
+   png_byte channels; /* number of channels (1, 2, 3, or 4) */
+   png_byte pixel_depth; /* bits per pixel (depth * channels) */
+} png_row_info;
+
+typedef png_row_info FAR * png_row_infop;
+typedef png_row_info FAR * FAR * png_row_infopp;
+
+/* These are the function types for the I/O functions and for the functions
+ * that allow the user to override the default I/O functions with his or her
+ * own.  The png_error_ptr type should match that of user-supplied warning
+ * and error functions, while the png_rw_ptr type should match that of the
+ * user read/write data functions.
+ */
+typedef struct png_struct_def png_struct;
+typedef png_struct FAR * png_structp;
+
+typedef void (PNGAPI *png_error_ptr) PNGARG((png_structp, png_const_charp));
+typedef void (PNGAPI *png_rw_ptr) PNGARG((png_structp, png_bytep, png_size_t));
+typedef void (PNGAPI *png_flush_ptr) PNGARG((png_structp));
+typedef void (PNGAPI *png_read_status_ptr) PNGARG((png_structp, png_uint_32,
+   int));
+typedef void (PNGAPI *png_write_status_ptr) PNGARG((png_structp, png_uint_32,
+   int));
+
+#ifdef PNG_PROGRESSIVE_READ_SUPPORTED
+typedef void (PNGAPI *png_progressive_info_ptr) PNGARG((png_structp, png_infop));
+typedef void (PNGAPI *png_progressive_end_ptr) PNGARG((png_structp, png_infop));
+typedef void (PNGAPI *png_progressive_row_ptr) PNGARG((png_structp, png_bytep,
+   png_uint_32, int));
+#endif
+
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_LEGACY_SUPPORTED)
+typedef void (PNGAPI *png_user_transform_ptr) PNGARG((png_structp,
+    png_row_infop, png_bytep));
+#endif
+
+#if defined(PNG_USER_CHUNKS_SUPPORTED)
+typedef int (PNGAPI *png_user_chunk_ptr) PNGARG((png_structp, png_unknown_chunkp));
+#endif
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+typedef void (PNGAPI *png_unknown_chunk_ptr) PNGARG((png_structp));
+#endif
+
+/* Transform masks for the high-level interface */
+#define PNG_TRANSFORM_IDENTITY       0x0000    /* read and write */
+#define PNG_TRANSFORM_STRIP_16       0x0001    /* read only */
+#define PNG_TRANSFORM_STRIP_ALPHA    0x0002    /* read only */
+#define PNG_TRANSFORM_PACKING        0x0004    /* read and write */
+#define PNG_TRANSFORM_PACKSWAP       0x0008    /* read and write */
+#define PNG_TRANSFORM_EXPAND         0x0010    /* read only */
+#define PNG_TRANSFORM_INVERT_MONO    0x0020    /* read and write */
+#define PNG_TRANSFORM_SHIFT          0x0040    /* read and write */
+#define PNG_TRANSFORM_BGR            0x0080    /* read and write */
+#define PNG_TRANSFORM_SWAP_ALPHA     0x0100    /* read and write */
+#define PNG_TRANSFORM_SWAP_ENDIAN    0x0200    /* read and write */
+#define PNG_TRANSFORM_INVERT_ALPHA   0x0400    /* read and write */
+#define PNG_TRANSFORM_STRIP_FILLER   0x0800    /* WRITE only */
+
+/* Flags for MNG supported features */
+#define PNG_FLAG_MNG_EMPTY_PLTE     0x01
+#define PNG_FLAG_MNG_FILTER_64      0x04
+#define PNG_ALL_MNG_FEATURES        0x05
+
+typedef png_voidp (*png_malloc_ptr) PNGARG((png_structp, png_size_t));
+typedef void (*png_free_ptr) PNGARG((png_structp, png_voidp));
+
+/* The structure that holds the information to read and write PNG files.
+ * The only people who need to care about what is inside of this are the
+ * people who will be modifying the library for their own special needs.
+ * It should NOT be accessed directly by an application, except to store
+ * the jmp_buf.
+ */
+
+struct png_struct_def
+{
+#ifdef PNG_SETJMP_SUPPORTED
+   jmp_buf jmpbuf;            /* used in png_error */
+#endif
+   png_error_ptr error_fn;    /* function for printing errors and aborting */
+   png_error_ptr warning_fn;  /* function for printing warnings */
+   png_voidp error_ptr;       /* user supplied struct for error functions */
+   png_rw_ptr write_data_fn;  /* function for writing output data */
+   png_rw_ptr read_data_fn;   /* function for reading input data */
+   png_voidp io_ptr;          /* ptr to application struct for I/O functions */
+
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED)
+   png_user_transform_ptr read_user_transform_fn; /* user read transform */
+#endif
+
+#if defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED)
+   png_user_transform_ptr write_user_transform_fn; /* user write transform */
+#endif
+
+/* These were added in libpng-1.0.2 */
+#if defined(PNG_USER_TRANSFORM_PTR_SUPPORTED)
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED)
+   png_voidp user_transform_ptr; /* user supplied struct for user transform */
+   png_byte user_transform_depth;    /* bit depth of user transformed pixels */
+   png_byte user_transform_channels; /* channels in user transformed pixels */
+#endif
+#endif
+
+   png_uint_32 mode;          /* tells us where we are in the PNG file */
+   png_uint_32 flags;         /* flags indicating various things to libpng */
+   png_uint_32 transformations; /* which transformations to perform */
+
+   z_stream zstream;          /* pointer to decompression structure (below) */
+   png_bytep zbuf;            /* buffer for zlib */
+   png_size_t zbuf_size;      /* size of zbuf */
+   int zlib_level;            /* holds zlib compression level */
+   int zlib_method;           /* holds zlib compression method */
+   int zlib_window_bits;      /* holds zlib compression window bits */
+   int zlib_mem_level;        /* holds zlib compression memory level */
+   int zlib_strategy;         /* holds zlib compression strategy */
+
+   png_uint_32 width;         /* width of image in pixels */
+   png_uint_32 height;        /* height of image in pixels */
+   png_uint_32 num_rows;      /* number of rows in current pass */
+   png_uint_32 usr_width;     /* width of row at start of write */
+   png_uint_32 rowbytes;      /* size of row in bytes */
+   png_uint_32 irowbytes;     /* size of current interlaced row in bytes */
+   png_uint_32 iwidth;        /* width of current interlaced row in pixels */
+   png_uint_32 row_number;    /* current row in interlace pass */
+   png_bytep prev_row;        /* buffer to save previous (unfiltered) row */
+   png_bytep row_buf;         /* buffer to save current (unfiltered) row */
+   png_bytep sub_row;         /* buffer to save "sub" row when filtering */
+   png_bytep up_row;          /* buffer to save "up" row when filtering */
+   png_bytep avg_row;         /* buffer to save "avg" row when filtering */
+   png_bytep paeth_row;       /* buffer to save "Paeth" row when filtering */
+   png_row_info row_info;     /* used for transformation routines */
+
+   png_uint_32 idat_size;     /* current IDAT size for read */
+   png_uint_32 crc;           /* current chunk CRC value */
+   png_colorp palette;        /* palette from the input file */
+   png_uint_16 num_palette;   /* number of color entries in palette */
+   png_uint_16 num_trans;     /* number of transparency values */
+   png_byte chunk_name[5];    /* null-terminated name of current chunk */
+   png_byte compression;      /* file compression type (always 0) */
+   png_byte filter;           /* file filter type (always 0) */
+   png_byte interlaced;       /* PNG_INTERLACE_NONE, PNG_INTERLACE_ADAM7 */
+   png_byte pass;             /* current interlace pass (0 - 6) */
+   png_byte do_filter;        /* row filter flags (see PNG_FILTER_ below ) */
+   png_byte color_type;       /* color type of file */
+   png_byte bit_depth;        /* bit depth of file */
+   png_byte usr_bit_depth;    /* bit depth of users row */
+   png_byte pixel_depth;      /* number of bits per pixel */
+   png_byte channels;         /* number of channels in file */
+   png_byte usr_channels;     /* channels at start of write */
+   png_byte sig_bytes;        /* magic bytes read/written from start of file */
+
+#if defined(PNG_READ_FILLER_SUPPORTED) || defined(PNG_WRITE_FILLER_SUPPORTED)
+#ifdef PNG_LEGACY_SUPPORTED
+   png_byte filler;           /* filler byte for pixel expansion */
+#else
+   png_uint_16 filler;           /* filler bytes for pixel expansion */
+#endif
+#endif
+
+#if defined(PNG_bKGD_SUPPORTED)
+   png_byte background_gamma_type;
+#  ifdef PNG_FLOATING_POINT_SUPPORTED
+   float background_gamma;
+#  endif
+   png_color_16 background;   /* background color in screen gamma space */
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+   png_color_16 background_1; /* background normalized to gamma 1.0 */
+#endif
+#endif /* PNG_bKGD_SUPPORTED */
+
+#if defined(PNG_WRITE_FLUSH_SUPPORTED)
+   png_flush_ptr output_flush_fn;/* Function for flushing output */
+   png_uint_32 flush_dist;    /* how many rows apart to flush, 0 - no flush */
+   png_uint_32 flush_rows;    /* number of rows written since last flush */
+#endif
+
+#if defined(PNG_READ_GAMMA_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+   int gamma_shift;      /* number of "insignificant" bits 16-bit gamma */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   float gamma;          /* file gamma value */
+   float screen_gamma;   /* screen gamma value (display_exponent) */
+#endif
+#endif
+
+#if defined(PNG_READ_GAMMA_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+   png_bytep gamma_table;     /* gamma table for 8-bit depth files */
+   png_bytep gamma_from_1;    /* converts from 1.0 to screen */
+   png_bytep gamma_to_1;      /* converts from file to 1.0 */
+   png_uint_16pp gamma_16_table; /* gamma table for 16-bit depth files */
+   png_uint_16pp gamma_16_from_1; /* converts from 1.0 to screen */
+   png_uint_16pp gamma_16_to_1; /* converts from file to 1.0 */
+#endif
+
+#if defined(PNG_READ_GAMMA_SUPPORTED) || defined(PNG_sBIT_SUPPORTED)
+   png_color_8 sig_bit;       /* significant bits in each available channel */
+#endif
+
+#if defined(PNG_READ_SHIFT_SUPPORTED) || defined(PNG_WRITE_SHIFT_SUPPORTED)
+   png_color_8 shift;         /* shift for significant bit tranformation */
+#endif
+
+#if defined(PNG_tRNS_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED) \
+ || defined(PNG_READ_EXPAND_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+   png_bytep trans;           /* transparency values for paletted files */
+   png_color_16 trans_values; /* transparency values for non-paletted files */
+#endif
+
+   png_read_status_ptr read_row_fn;   /* called after each row is decoded */
+   png_write_status_ptr write_row_fn; /* called after each row is encoded */
+#ifdef PNG_PROGRESSIVE_READ_SUPPORTED
+   png_progressive_info_ptr info_fn; /* called after header data fully read */
+   png_progressive_row_ptr row_fn;   /* called after each prog. row is decoded */
+   png_progressive_end_ptr end_fn;   /* called after image is complete */
+   png_bytep save_buffer_ptr;        /* current location in save_buffer */
+   png_bytep save_buffer;            /* buffer for previously read data */
+   png_bytep current_buffer_ptr;     /* current location in current_buffer */
+   png_bytep current_buffer;         /* buffer for recently used data */
+   png_uint_32 push_length;          /* size of current input chunk */
+   png_uint_32 skip_length;          /* bytes to skip in input data */
+   png_size_t save_buffer_size;      /* amount of data now in save_buffer */
+   png_size_t save_buffer_max;       /* total size of save_buffer */
+   png_size_t buffer_size;           /* total amount of available input data */
+   png_size_t current_buffer_size;   /* amount of data now in current_buffer */
+   int process_mode;                 /* what push library is currently doing */
+   int cur_palette;                  /* current push library palette index */
+
+#  if defined(PNG_TEXT_SUPPORTED)
+     png_size_t current_text_size;   /* current size of text input data */
+     png_size_t current_text_left;   /* how much text left to read in input */
+     png_charp current_text;         /* current text chunk buffer */
+     png_charp current_text_ptr;     /* current location in current_text */
+#  endif /* PNG_PROGRESSIVE_READ_SUPPORTED && PNG_TEXT_SUPPORTED */
+
+#endif /* PNG_PROGRESSIVE_READ_SUPPORTED */
+
+#if defined(__TURBOC__) && !defined(_Windows) && !defined(__FLAT__)
+/* for the Borland special 64K segment handler */
+   png_bytepp offset_table_ptr;
+   png_bytep offset_table;
+   png_uint_16 offset_table_number;
+   png_uint_16 offset_table_count;
+   png_uint_16 offset_table_count_free;
+#endif
+
+#if defined(PNG_READ_DITHER_SUPPORTED)
+   png_bytep palette_lookup;         /* lookup table for dithering */
+   png_bytep dither_index;           /* index translation for palette files */
+#endif
+
+#if defined(PNG_READ_DITHER_SUPPORTED) || defined(PNG_hIST_SUPPORTED)
+   png_uint_16p hist;                /* histogram */
+#endif
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+   png_byte heuristic_method;        /* heuristic for row filter selection */
+   png_byte num_prev_filters;        /* number of weights for previous rows */
+   png_bytep prev_filters;           /* filter type(s) of previous row(s) */
+   png_uint_16p filter_weights;      /* weight(s) for previous line(s) */
+   png_uint_16p inv_filter_weights;  /* 1/weight(s) for previous line(s) */
+   png_uint_16p filter_costs;        /* relative filter calculation cost */
+   png_uint_16p inv_filter_costs;    /* 1/relative filter calculation cost */
+#endif
+
+#if defined(PNG_TIME_RFC1123_SUPPORTED)
+   png_charp time_buffer;            /* String to hold RFC 1123 time text */
+#endif
+
+/* New members added in libpng-1.0.6 */
+
+#ifdef PNG_FREE_ME_SUPPORTED
+   png_uint_32 free_me;       /* flags items libpng is responsible for freeing */
+#endif
+
+#if defined(PNG_USER_CHUNKS_SUPPORTED)
+   png_voidp user_chunk_ptr;
+   png_user_chunk_ptr read_user_chunk_fn; /* user read chunk handler */
+#endif
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+   int num_chunk_list;
+   png_bytep chunk_list;
+#endif
+
+/* New members added in libpng-1.0.3 */
+#if defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+   png_byte rgb_to_gray_status;
+   /* These were changed from png_byte in libpng-1.0.6 */
+   png_uint_16 rgb_to_gray_red_coeff;
+   png_uint_16 rgb_to_gray_green_coeff;
+   png_uint_16 rgb_to_gray_blue_coeff;
+#endif
+
+/* New member added in libpng-1.0.4 (renamed in 1.0.9) */
+#if defined(PNG_MNG_FEATURES_SUPPORTED) || \
+    defined(PNG_READ_EMPTY_PLTE_SUPPORTED) || \
+    defined(PNG_WRITE_EMPTY_PLTE_SUPPORTED)
+/* changed from png_byte to png_uint_32 at version 1.2.0 */
+#ifdef PNG_1_0_X
+   png_byte mng_features_permitted;
+#else
+   png_uint_32 mng_features_permitted;
+#endif /* PNG_1_0_X */
+#endif
+
+/* New member added in libpng-1.0.7 */
+#if defined(PNG_READ_GAMMA_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+   png_fixed_point int_gamma;
+#endif
+
+/* New member added in libpng-1.0.9, ifdef'ed out in 1.0.12, enabled in 1.2.0 */
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+   png_byte filter_type;
+#endif
+
+#if defined(PNG_1_0_X)
+/* New member added in libpng-1.0.10, ifdef'ed out in 1.2.0 */
+   png_uint_32 row_buf_size;
+#endif
+
+/* New members added in libpng-1.2.0 */
+#if defined(PNG_ASSEMBLER_CODE_SUPPORTED)
+#  if !defined(PNG_1_0_X)
+#    if defined(PNG_MMX_CODE_SUPPORTED)
+   png_byte     mmx_bitdepth_threshold;
+   png_uint_32  mmx_rowbytes_threshold;
+#    endif
+   png_uint_32  asm_flags;
+#  endif
+#endif
+
+/* New members added in libpng-1.0.2 but first enabled by default in 1.2.0 */
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_voidp mem_ptr;                /* user supplied struct for mem functions */
+   png_malloc_ptr malloc_fn;         /* function for allocating memory */
+   png_free_ptr free_fn;             /* function for freeing memory */
+#endif
+
+/* New member added in libpng-1.0.13 and 1.2.0 */
+   png_bytep big_row_buf;         /* buffer to save current (unfiltered) row */
+
+#if defined(PNG_READ_DITHER_SUPPORTED)
+/* The following three members were added at version 1.0.14 and 1.2.4 */
+   png_bytep dither_sort;            /* working sort array */
+   png_bytep index_to_palette;       /* where the original index currently is */
+                                     /* in the palette */
+   png_bytep palette_to_index;       /* which original index points to this */
+                                     /* palette color */
+#endif
+
+/* New members added in libpng-1.0.16 and 1.2.6 */
+   png_byte compression_type;
+
+#ifdef PNG_SET_USER_LIMITS_SUPPORTED
+   png_uint_32 user_width_max;
+   png_uint_32 user_height_max;
+#endif
+
+/* New member added in libpng-1.0.25 and 1.2.17 */
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+   /* storage for unknown chunk that the library doesn't recognize. */
+   png_unknown_chunk unknown_chunk;
+#endif
+};
+
+
+/* This triggers a compiler error in png.c, if png.c and png.h
+ * do not agree upon the version number.
+ */
+typedef png_structp version_1_2_19;
+
+typedef png_struct FAR * FAR * png_structpp;
+
+/* Here are the function definitions most commonly used.  This is not
+ * the place to find out how to use libpng.  See libpng.txt for the
+ * full explanation, see example.c for the summary.  This just provides
+ * a simple one line description of the use of each function.
+ */
+
+/* Returns the version number of the library */
+extern PNG_EXPORT(png_uint_32,png_access_version_number) PNGARG((void));
+
+/* Tell lib we have already handled the first <num_bytes> magic bytes.
+ * Handling more than 8 bytes from the beginning of the file is an error.
+ */
+extern PNG_EXPORT(void,png_set_sig_bytes) PNGARG((png_structp png_ptr,
+   int num_bytes));
+
+/* Check sig[start] through sig[start + num_to_check - 1] to see if it's a
+ * PNG file.  Returns zero if the supplied bytes match the 8-byte PNG
+ * signature, and non-zero otherwise.  Having num_to_check == 0 or
+ * start > 7 will always fail (ie return non-zero).
+ */
+extern PNG_EXPORT(int,png_sig_cmp) PNGARG((png_bytep sig, png_size_t start,
+   png_size_t num_to_check));
+
+/* Simple signature checking function.  This is the same as calling
+ * png_check_sig(sig, n) := !png_sig_cmp(sig, 0, n).
+ */
+extern PNG_EXPORT(int,png_check_sig) PNGARG((png_bytep sig, int num));
+
+/* Allocate and initialize png_ptr struct for reading, and any other memory. */
+extern PNG_EXPORT(png_structp,png_create_read_struct)
+   PNGARG((png_const_charp user_png_ver, png_voidp error_ptr,
+   png_error_ptr error_fn, png_error_ptr warn_fn));
+
+/* Allocate and initialize png_ptr struct for writing, and any other memory */
+extern PNG_EXPORT(png_structp,png_create_write_struct)
+   PNGARG((png_const_charp user_png_ver, png_voidp error_ptr,
+   png_error_ptr error_fn, png_error_ptr warn_fn));
+
+#ifdef PNG_WRITE_SUPPORTED
+extern PNG_EXPORT(png_uint_32,png_get_compression_buffer_size)
+   PNGARG((png_structp png_ptr));
+#endif
+
+#ifdef PNG_WRITE_SUPPORTED
+extern PNG_EXPORT(void,png_set_compression_buffer_size)
+   PNGARG((png_structp png_ptr, png_uint_32 size));
+#endif
+
+/* Reset the compression stream */
+extern PNG_EXPORT(int,png_reset_zstream) PNGARG((png_structp png_ptr));
+
+/* New functions added in libpng-1.0.2 (not enabled by default until 1.2.0) */
+#ifdef PNG_USER_MEM_SUPPORTED
+extern PNG_EXPORT(png_structp,png_create_read_struct_2)
+   PNGARG((png_const_charp user_png_ver, png_voidp error_ptr,
+   png_error_ptr error_fn, png_error_ptr warn_fn, png_voidp mem_ptr,
+   png_malloc_ptr malloc_fn, png_free_ptr free_fn));
+extern PNG_EXPORT(png_structp,png_create_write_struct_2)
+   PNGARG((png_const_charp user_png_ver, png_voidp error_ptr,
+   png_error_ptr error_fn, png_error_ptr warn_fn, png_voidp mem_ptr,
+   png_malloc_ptr malloc_fn, png_free_ptr free_fn));
+#endif
+
+/* Write a PNG chunk - size, type, (optional) data, CRC. */
+extern PNG_EXPORT(void,png_write_chunk) PNGARG((png_structp png_ptr,
+   png_bytep chunk_name, png_bytep data, png_size_t length));
+
+/* Write the start of a PNG chunk - length and chunk name. */
+extern PNG_EXPORT(void,png_write_chunk_start) PNGARG((png_structp png_ptr,
+   png_bytep chunk_name, png_uint_32 length));
+
+/* Write the data of a PNG chunk started with png_write_chunk_start(). */
+extern PNG_EXPORT(void,png_write_chunk_data) PNGARG((png_structp png_ptr,
+   png_bytep data, png_size_t length));
+
+/* Finish a chunk started with png_write_chunk_start() (includes CRC). */
+extern PNG_EXPORT(void,png_write_chunk_end) PNGARG((png_structp png_ptr));
+
+/* Allocate and initialize the info structure */
+extern PNG_EXPORT(png_infop,png_create_info_struct)
+   PNGARG((png_structp png_ptr));
+
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+/* Initialize the info structure (old interface - DEPRECATED) */
+extern PNG_EXPORT(void,png_info_init) PNGARG((png_infop info_ptr));
+#undef png_info_init
+#define png_info_init(info_ptr) png_info_init_3(&info_ptr,\
+    png_sizeof(png_info));
+#endif
+
+extern PNG_EXPORT(void,png_info_init_3) PNGARG((png_infopp info_ptr,
+    png_size_t png_info_struct_size));
+
+/* Writes all the PNG information before the image. */
+extern PNG_EXPORT(void,png_write_info_before_PLTE) PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+extern PNG_EXPORT(void,png_write_info) PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* read the information before the actual image data. */
+extern PNG_EXPORT(void,png_read_info) PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+#endif
+
+#if defined(PNG_TIME_RFC1123_SUPPORTED)
+extern PNG_EXPORT(png_charp,png_convert_to_rfc1123)
+   PNGARG((png_structp png_ptr, png_timep ptime));
+#endif
+
+#if !defined(_WIN32_WCE)
+/* "time.h" functions are not supported on WindowsCE */
+#if defined(PNG_WRITE_tIME_SUPPORTED)
+/* convert from a struct tm to png_time */
+extern PNG_EXPORT(void,png_convert_from_struct_tm) PNGARG((png_timep ptime,
+   struct tm FAR * ttime));
+
+/* convert from time_t to png_time.  Uses gmtime() */
+extern PNG_EXPORT(void,png_convert_from_time_t) PNGARG((png_timep ptime,
+   time_t ttime));
+#endif /* PNG_WRITE_tIME_SUPPORTED */
+#endif /* _WIN32_WCE */
+
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+/* Expand data to 24-bit RGB, or 8-bit grayscale, with alpha if available. */
+extern PNG_EXPORT(void,png_set_expand) PNGARG((png_structp png_ptr));
+#if !defined(PNG_1_0_X)
+extern PNG_EXPORT(void,png_set_expand_gray_1_2_4_to_8) PNGARG((png_structp
+  png_ptr));
+#endif
+extern PNG_EXPORT(void,png_set_palette_to_rgb) PNGARG((png_structp png_ptr));
+extern PNG_EXPORT(void,png_set_tRNS_to_alpha) PNGARG((png_structp png_ptr));
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+/* Deprecated */
+extern PNG_EXPORT(void,png_set_gray_1_2_4_to_8) PNGARG((png_structp png_ptr));
+#endif
+#endif
+
+#if defined(PNG_READ_BGR_SUPPORTED) || defined(PNG_WRITE_BGR_SUPPORTED)
+/* Use blue, green, red order for pixels. */
+extern PNG_EXPORT(void,png_set_bgr) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_GRAY_TO_RGB_SUPPORTED)
+/* Expand the grayscale to 24-bit RGB if necessary. */
+extern PNG_EXPORT(void,png_set_gray_to_rgb) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+/* Reduce RGB to grayscale. */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(void,png_set_rgb_to_gray) PNGARG((png_structp png_ptr,
+   int error_action, double red, double green ));
+#endif
+extern PNG_EXPORT(void,png_set_rgb_to_gray_fixed) PNGARG((png_structp png_ptr,
+   int error_action, png_fixed_point red, png_fixed_point green ));
+extern PNG_EXPORT(png_byte,png_get_rgb_to_gray_status) PNGARG((png_structp
+   png_ptr));
+#endif
+
+extern PNG_EXPORT(void,png_build_grayscale_palette) PNGARG((int bit_depth,
+   png_colorp palette));
+
+#if defined(PNG_READ_STRIP_ALPHA_SUPPORTED)
+extern PNG_EXPORT(void,png_set_strip_alpha) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_SWAP_ALPHA_SUPPORTED) || \
+    defined(PNG_WRITE_SWAP_ALPHA_SUPPORTED)
+extern PNG_EXPORT(void,png_set_swap_alpha) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_INVERT_ALPHA_SUPPORTED) || \
+    defined(PNG_WRITE_INVERT_ALPHA_SUPPORTED)
+extern PNG_EXPORT(void,png_set_invert_alpha) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_FILLER_SUPPORTED) || defined(PNG_WRITE_FILLER_SUPPORTED)
+/* Add a filler byte to 8-bit Gray or 24-bit RGB images. */
+extern PNG_EXPORT(void,png_set_filler) PNGARG((png_structp png_ptr,
+   png_uint_32 filler, int flags));
+/* The values of the PNG_FILLER_ defines should NOT be changed */
+#define PNG_FILLER_BEFORE 0
+#define PNG_FILLER_AFTER 1
+/* Add an alpha byte to 8-bit Gray or 24-bit RGB images. */
+#if !defined(PNG_1_0_X)
+extern PNG_EXPORT(void,png_set_add_alpha) PNGARG((png_structp png_ptr,
+   png_uint_32 filler, int flags));
+#endif
+#endif /* PNG_READ_FILLER_SUPPORTED || PNG_WRITE_FILLER_SUPPORTED */
+
+#if defined(PNG_READ_SWAP_SUPPORTED) || defined(PNG_WRITE_SWAP_SUPPORTED)
+/* Swap bytes in 16-bit depth files. */
+extern PNG_EXPORT(void,png_set_swap) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_PACK_SUPPORTED) || defined(PNG_WRITE_PACK_SUPPORTED)
+/* Use 1 byte per pixel in 1, 2, or 4-bit depth files. */
+extern PNG_EXPORT(void,png_set_packing) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED) || defined(PNG_WRITE_PACKSWAP_SUPPORTED)
+/* Swap packing order of pixels in bytes. */
+extern PNG_EXPORT(void,png_set_packswap) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_SHIFT_SUPPORTED) || defined(PNG_WRITE_SHIFT_SUPPORTED)
+/* Converts files to legal bit depths. */
+extern PNG_EXPORT(void,png_set_shift) PNGARG((png_structp png_ptr,
+   png_color_8p true_bits));
+#endif
+
+#if defined(PNG_READ_INTERLACING_SUPPORTED) || \
+    defined(PNG_WRITE_INTERLACING_SUPPORTED)
+/* Have the code handle the interlacing.  Returns the number of passes. */
+extern PNG_EXPORT(int,png_set_interlace_handling) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_INVERT_SUPPORTED) || defined(PNG_WRITE_INVERT_SUPPORTED)
+/* Invert monochrome files */
+extern PNG_EXPORT(void,png_set_invert_mono) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+/* Handle alpha and tRNS by replacing with a background color. */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(void,png_set_background) PNGARG((png_structp png_ptr,
+   png_color_16p background_color, int background_gamma_code,
+   int need_expand, double background_gamma));
+#endif
+#define PNG_BACKGROUND_GAMMA_UNKNOWN 0
+#define PNG_BACKGROUND_GAMMA_SCREEN  1
+#define PNG_BACKGROUND_GAMMA_FILE    2
+#define PNG_BACKGROUND_GAMMA_UNIQUE  3
+#endif
+
+#if defined(PNG_READ_16_TO_8_SUPPORTED)
+/* strip the second byte of information from a 16-bit depth file. */
+extern PNG_EXPORT(void,png_set_strip_16) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_DITHER_SUPPORTED)
+/* Turn on dithering, and reduce the palette to the number of colors available. */
+extern PNG_EXPORT(void,png_set_dither) PNGARG((png_structp png_ptr,
+   png_colorp palette, int num_palette, int maximum_colors,
+   png_uint_16p histogram, int full_dither));
+#endif
+
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+/* Handle gamma correction. Screen_gamma=(display_exponent) */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(void,png_set_gamma) PNGARG((png_structp png_ptr,
+   double screen_gamma, double default_file_gamma));
+#endif
+#endif
+
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+#if defined(PNG_READ_EMPTY_PLTE_SUPPORTED) || \
+    defined(PNG_WRITE_EMPTY_PLTE_SUPPORTED)
+/* Permit or disallow empty PLTE (0: not permitted, 1: permitted) */
+/* Deprecated and will be removed.  Use png_permit_mng_features() instead. */
+extern PNG_EXPORT(void,png_permit_empty_plte) PNGARG((png_structp png_ptr,
+   int empty_plte_permitted));
+#endif
+#endif
+
+#if defined(PNG_WRITE_FLUSH_SUPPORTED)
+/* Set how many lines between output flushes - 0 for no flushing */
+extern PNG_EXPORT(void,png_set_flush) PNGARG((png_structp png_ptr, int nrows));
+/* Flush the current PNG output buffer */
+extern PNG_EXPORT(void,png_write_flush) PNGARG((png_structp png_ptr));
+#endif
+
+/* optional update palette with requested transformations */
+extern PNG_EXPORT(void,png_start_read_image) PNGARG((png_structp png_ptr));
+
+/* optional call to update the users info structure */
+extern PNG_EXPORT(void,png_read_update_info) PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* read one or more rows of image data. */
+extern PNG_EXPORT(void,png_read_rows) PNGARG((png_structp png_ptr,
+   png_bytepp row, png_bytepp display_row, png_uint_32 num_rows));
+#endif
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* read a row of data. */
+extern PNG_EXPORT(void,png_read_row) PNGARG((png_structp png_ptr,
+   png_bytep row,
+   png_bytep display_row));
+#endif
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* read the whole image into memory at once. */
+extern PNG_EXPORT(void,png_read_image) PNGARG((png_structp png_ptr,
+   png_bytepp image));
+#endif
+
+/* write a row of image data */
+extern PNG_EXPORT(void,png_write_row) PNGARG((png_structp png_ptr,
+   png_bytep row));
+
+/* write a few rows of image data */
+extern PNG_EXPORT(void,png_write_rows) PNGARG((png_structp png_ptr,
+   png_bytepp row, png_uint_32 num_rows));
+
+/* write the image data */
+extern PNG_EXPORT(void,png_write_image) PNGARG((png_structp png_ptr,
+   png_bytepp image));
+
+/* writes the end of the PNG file. */
+extern PNG_EXPORT(void,png_write_end) PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* read the end of the PNG file. */
+extern PNG_EXPORT(void,png_read_end) PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+#endif
+
+/* free any memory associated with the png_info_struct */
+extern PNG_EXPORT(void,png_destroy_info_struct) PNGARG((png_structp png_ptr,
+   png_infopp info_ptr_ptr));
+
+/* free any memory associated with the png_struct and the png_info_structs */
+extern PNG_EXPORT(void,png_destroy_read_struct) PNGARG((png_structpp
+   png_ptr_ptr, png_infopp info_ptr_ptr, png_infopp end_info_ptr_ptr));
+
+/* free all memory used by the read (old method - NOT DLL EXPORTED) */
+extern void png_read_destroy PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_infop end_info_ptr));
+
+/* free any memory associated with the png_struct and the png_info_structs */
+extern PNG_EXPORT(void,png_destroy_write_struct)
+   PNGARG((png_structpp png_ptr_ptr, png_infopp info_ptr_ptr));
+
+/* free any memory used in png_ptr struct (old method - NOT DLL EXPORTED) */
+extern void png_write_destroy PNGARG((png_structp png_ptr));
+
+/* set the libpng method of handling chunk CRC errors */
+extern PNG_EXPORT(void,png_set_crc_action) PNGARG((png_structp png_ptr,
+   int crit_action, int ancil_action));
+
+/* Values for png_set_crc_action() to say how to handle CRC errors in
+ * ancillary and critical chunks, and whether to use the data contained
+ * therein.  Note that it is impossible to "discard" data in a critical
+ * chunk.  For versions prior to 0.90, the action was always error/quit,
+ * whereas in version 0.90 and later, the action for CRC errors in ancillary
+ * chunks is warn/discard.  These values should NOT be changed.
+ *
+ *      value                       action:critical     action:ancillary
+ */
+#define PNG_CRC_DEFAULT       0  /* error/quit          warn/discard data */
+#define PNG_CRC_ERROR_QUIT    1  /* error/quit          error/quit        */
+#define PNG_CRC_WARN_DISCARD  2  /* (INVALID)           warn/discard data */
+#define PNG_CRC_WARN_USE      3  /* warn/use data       warn/use data     */
+#define PNG_CRC_QUIET_USE     4  /* quiet/use data      quiet/use data    */
+#define PNG_CRC_NO_CHANGE     5  /* use current value   use current value */
+
+/* These functions give the user control over the scan-line filtering in
+ * libpng and the compression methods used by zlib.  These functions are
+ * mainly useful for testing, as the defaults should work with most users.
+ * Those users who are tight on memory or want faster performance at the
+ * expense of compression can modify them.  See the compression library
+ * header file (zlib.h) for an explination of the compression functions.
+ */
+
+/* set the filtering method(s) used by libpng.  Currently, the only valid
+ * value for "method" is 0.
+ */
+extern PNG_EXPORT(void,png_set_filter) PNGARG((png_structp png_ptr, int method,
+   int filters));
+
+/* Flags for png_set_filter() to say which filters to use.  The flags
+ * are chosen so that they don't conflict with real filter types
+ * below, in case they are supplied instead of the #defined constants.
+ * These values should NOT be changed.
+ */
+#define PNG_NO_FILTERS     0x00
+#define PNG_FILTER_NONE    0x08
+#define PNG_FILTER_SUB     0x10
+#define PNG_FILTER_UP      0x20
+#define PNG_FILTER_AVG     0x40
+#define PNG_FILTER_PAETH   0x80
+#define PNG_ALL_FILTERS (PNG_FILTER_NONE | PNG_FILTER_SUB | PNG_FILTER_UP | \
+                         PNG_FILTER_AVG | PNG_FILTER_PAETH)
+
+/* Filter values (not flags) - used in pngwrite.c, pngwutil.c for now.
+ * These defines should NOT be changed.
+ */
+#define PNG_FILTER_VALUE_NONE  0
+#define PNG_FILTER_VALUE_SUB   1
+#define PNG_FILTER_VALUE_UP    2
+#define PNG_FILTER_VALUE_AVG   3
+#define PNG_FILTER_VALUE_PAETH 4
+#define PNG_FILTER_VALUE_LAST  5
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED) /* EXPERIMENTAL */
+/* The "heuristic_method" is given by one of the PNG_FILTER_HEURISTIC_
+ * defines, either the default (minimum-sum-of-absolute-differences), or
+ * the experimental method (weighted-minimum-sum-of-absolute-differences).
+ *
+ * Weights are factors >= 1.0, indicating how important it is to keep the
+ * filter type consistent between rows.  Larger numbers mean the current
+ * filter is that many times as likely to be the same as the "num_weights"
+ * previous filters.  This is cumulative for each previous row with a weight.
+ * There needs to be "num_weights" values in "filter_weights", or it can be
+ * NULL if the weights aren't being specified.  Weights have no influence on
+ * the selection of the first row filter.  Well chosen weights can (in theory)
+ * improve the compression for a given image.
+ *
+ * Costs are factors >= 1.0 indicating the relative decoding costs of a
+ * filter type.  Higher costs indicate more decoding expense, and are
+ * therefore less likely to be selected over a filter with lower computational
+ * costs.  There needs to be a value in "filter_costs" for each valid filter
+ * type (given by PNG_FILTER_VALUE_LAST), or it can be NULL if you aren't
+ * setting the costs.  Costs try to improve the speed of decompression without
+ * unduly increasing the compressed image size.
+ *
+ * A negative weight or cost indicates the default value is to be used, and
+ * values in the range [0.0, 1.0) indicate the value is to remain unchanged.
+ * The default values for both weights and costs are currently 1.0, but may
+ * change if good general weighting/cost heuristics can be found.  If both
+ * the weights and costs are set to 1.0, this degenerates the WEIGHTED method
+ * to the UNWEIGHTED method, but with added encoding time/computation.
+ */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(void,png_set_filter_heuristics) PNGARG((png_structp png_ptr,
+   int heuristic_method, int num_weights, png_doublep filter_weights,
+   png_doublep filter_costs));
+#endif
+#endif /*  PNG_WRITE_WEIGHTED_FILTER_SUPPORTED */
+
+/* Heuristic used for row filter selection.  These defines should NOT be
+ * changed.
+ */
+#define PNG_FILTER_HEURISTIC_DEFAULT    0  /* Currently "UNWEIGHTED" */
+#define PNG_FILTER_HEURISTIC_UNWEIGHTED 1  /* Used by libpng < 0.95 */
+#define PNG_FILTER_HEURISTIC_WEIGHTED   2  /* Experimental feature */
+#define PNG_FILTER_HEURISTIC_LAST       3  /* Not a valid value */
+
+/* Set the library compression level.  Currently, valid values range from
+ * 0 - 9, corresponding directly to the zlib compression levels 0 - 9
+ * (0 - no compression, 9 - "maximal" compression).  Note that tests have
+ * shown that zlib compression levels 3-6 usually perform as well as level 9
+ * for PNG images, and do considerably fewer caclulations.  In the future,
+ * these values may not correspond directly to the zlib compression levels.
+ */
+extern PNG_EXPORT(void,png_set_compression_level) PNGARG((png_structp png_ptr,
+   int level));
+
+extern PNG_EXPORT(void,png_set_compression_mem_level)
+   PNGARG((png_structp png_ptr, int mem_level));
+
+extern PNG_EXPORT(void,png_set_compression_strategy)
+   PNGARG((png_structp png_ptr, int strategy));
+
+extern PNG_EXPORT(void,png_set_compression_window_bits)
+   PNGARG((png_structp png_ptr, int window_bits));
+
+extern PNG_EXPORT(void,png_set_compression_method) PNGARG((png_structp png_ptr,
+   int method));
+
+/* These next functions are called for input/output, memory, and error
+ * handling.  They are in the file pngrio.c, pngwio.c, and pngerror.c,
+ * and call standard C I/O routines such as fread(), fwrite(), and
+ * fprintf().  These functions can be made to use other I/O routines
+ * at run time for those applications that need to handle I/O in a
+ * different manner by calling png_set_???_fn().  See libpng.txt for
+ * more information.
+ */
+
+#if !defined(PNG_NO_STDIO)
+/* Initialize the input/output for the PNG file to the default functions. */
+extern PNG_EXPORT(void,png_init_io) PNGARG((png_structp png_ptr, png_FILE_p fp));
+#endif
+
+/* Replace the (error and abort), and warning functions with user
+ * supplied functions.  If no messages are to be printed you must still
+ * write and use replacement functions. The replacement error_fn should
+ * still do a longjmp to the last setjmp location if you are using this
+ * method of error handling.  If error_fn or warning_fn is NULL, the
+ * default function will be used.
+ */
+
+extern PNG_EXPORT(void,png_set_error_fn) PNGARG((png_structp png_ptr,
+   png_voidp error_ptr, png_error_ptr error_fn, png_error_ptr warning_fn));
+
+/* Return the user pointer associated with the error functions */
+extern PNG_EXPORT(png_voidp,png_get_error_ptr) PNGARG((png_structp png_ptr));
+
+/* Replace the default data output functions with a user supplied one(s).
+ * If buffered output is not used, then output_flush_fn can be set to NULL.
+ * If PNG_WRITE_FLUSH_SUPPORTED is not defined at libpng compile time
+ * output_flush_fn will be ignored (and thus can be NULL).
+ */
+extern PNG_EXPORT(void,png_set_write_fn) PNGARG((png_structp png_ptr,
+   png_voidp io_ptr, png_rw_ptr write_data_fn, png_flush_ptr output_flush_fn));
+
+/* Replace the default data input function with a user supplied one. */
+extern PNG_EXPORT(void,png_set_read_fn) PNGARG((png_structp png_ptr,
+   png_voidp io_ptr, png_rw_ptr read_data_fn));
+
+/* Return the user pointer associated with the I/O functions */
+extern PNG_EXPORT(png_voidp,png_get_io_ptr) PNGARG((png_structp png_ptr));
+
+extern PNG_EXPORT(void,png_set_read_status_fn) PNGARG((png_structp png_ptr,
+   png_read_status_ptr read_row_fn));
+
+extern PNG_EXPORT(void,png_set_write_status_fn) PNGARG((png_structp png_ptr,
+   png_write_status_ptr write_row_fn));
+
+#ifdef PNG_USER_MEM_SUPPORTED
+/* Replace the default memory allocation functions with user supplied one(s). */
+extern PNG_EXPORT(void,png_set_mem_fn) PNGARG((png_structp png_ptr,
+   png_voidp mem_ptr, png_malloc_ptr malloc_fn, png_free_ptr free_fn));
+/* Return the user pointer associated with the memory functions */
+extern PNG_EXPORT(png_voidp,png_get_mem_ptr) PNGARG((png_structp png_ptr));
+#endif
+
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_LEGACY_SUPPORTED)
+extern PNG_EXPORT(void,png_set_read_user_transform_fn) PNGARG((png_structp
+   png_ptr, png_user_transform_ptr read_user_transform_fn));
+#endif
+
+#if defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_LEGACY_SUPPORTED)
+extern PNG_EXPORT(void,png_set_write_user_transform_fn) PNGARG((png_structp
+   png_ptr, png_user_transform_ptr write_user_transform_fn));
+#endif
+
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_LEGACY_SUPPORTED)
+extern PNG_EXPORT(void,png_set_user_transform_info) PNGARG((png_structp
+   png_ptr, png_voidp user_transform_ptr, int user_transform_depth,
+   int user_transform_channels));
+/* Return the user pointer associated with the user transform functions */
+extern PNG_EXPORT(png_voidp,png_get_user_transform_ptr)
+   PNGARG((png_structp png_ptr));
+#endif
+
+#ifdef PNG_USER_CHUNKS_SUPPORTED
+extern PNG_EXPORT(void,png_set_read_user_chunk_fn) PNGARG((png_structp png_ptr,
+   png_voidp user_chunk_ptr, png_user_chunk_ptr read_user_chunk_fn));
+extern PNG_EXPORT(png_voidp,png_get_user_chunk_ptr) PNGARG((png_structp
+   png_ptr));
+#endif
+
+#ifdef PNG_PROGRESSIVE_READ_SUPPORTED
+/* Sets the function callbacks for the push reader, and a pointer to a
+ * user-defined structure available to the callback functions.
+ */
+extern PNG_EXPORT(void,png_set_progressive_read_fn) PNGARG((png_structp png_ptr,
+   png_voidp progressive_ptr,
+   png_progressive_info_ptr info_fn, png_progressive_row_ptr row_fn,
+   png_progressive_end_ptr end_fn));
+
+/* returns the user pointer associated with the push read functions */
+extern PNG_EXPORT(png_voidp,png_get_progressive_ptr)
+   PNGARG((png_structp png_ptr));
+
+/* function to be called when data becomes available */
+extern PNG_EXPORT(void,png_process_data) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_bytep buffer, png_size_t buffer_size));
+
+/* function that combines rows.  Not very much different than the
+ * png_combine_row() call.  Is this even used?????
+ */
+extern PNG_EXPORT(void,png_progressive_combine_row) PNGARG((png_structp png_ptr,
+   png_bytep old_row, png_bytep new_row));
+#endif /* PNG_PROGRESSIVE_READ_SUPPORTED */
+
+extern PNG_EXPORT(png_voidp,png_malloc) PNGARG((png_structp png_ptr,
+   png_uint_32 size));
+
+#if defined(PNG_1_0_X)
+#  define png_malloc_warn png_malloc
+#else
+/* Added at libpng version 1.2.4 */
+extern PNG_EXPORT(png_voidp,png_malloc_warn) PNGARG((png_structp png_ptr,
+   png_uint_32 size));
+#endif
+
+/* frees a pointer allocated by png_malloc() */
+extern PNG_EXPORT(void,png_free) PNGARG((png_structp png_ptr, png_voidp ptr));
+
+#if defined(PNG_1_0_X)
+/* Function to allocate memory for zlib. */
+extern PNG_EXPORT(voidpf,png_zalloc) PNGARG((voidpf png_ptr, uInt items,
+   uInt size));
+
+/* Function to free memory for zlib */
+extern PNG_EXPORT(void,png_zfree) PNGARG((voidpf png_ptr, voidpf ptr));
+#endif
+
+/* Free data that was allocated internally */
+extern PNG_EXPORT(void,png_free_data) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 free_me, int num));
+#ifdef PNG_FREE_ME_SUPPORTED
+/* Reassign responsibility for freeing existing data, whether allocated
+ * by libpng or by the application */
+extern PNG_EXPORT(void,png_data_freer) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, int freer, png_uint_32 mask));
+#endif
+/* assignments for png_data_freer */
+#define PNG_DESTROY_WILL_FREE_DATA 1
+#define PNG_SET_WILL_FREE_DATA 1
+#define PNG_USER_WILL_FREE_DATA 2
+/* Flags for png_ptr->free_me and info_ptr->free_me */
+#define PNG_FREE_HIST 0x0008
+#define PNG_FREE_ICCP 0x0010
+#define PNG_FREE_SPLT 0x0020
+#define PNG_FREE_ROWS 0x0040
+#define PNG_FREE_PCAL 0x0080
+#define PNG_FREE_SCAL 0x0100
+#define PNG_FREE_UNKN 0x0200
+#define PNG_FREE_LIST 0x0400
+#define PNG_FREE_PLTE 0x1000
+#define PNG_FREE_TRNS 0x2000
+#define PNG_FREE_TEXT 0x4000
+#define PNG_FREE_ALL  0x7fff
+#define PNG_FREE_MUL  0x4220 /* PNG_FREE_SPLT|PNG_FREE_TEXT|PNG_FREE_UNKN */
+
+#ifdef PNG_USER_MEM_SUPPORTED
+extern PNG_EXPORT(png_voidp,png_malloc_default) PNGARG((png_structp png_ptr,
+   png_uint_32 size));
+extern PNG_EXPORT(void,png_free_default) PNGARG((png_structp png_ptr,
+   png_voidp ptr));
+#endif
+
+extern PNG_EXPORT(png_voidp,png_memcpy_check) PNGARG((png_structp png_ptr,
+   png_voidp s1, png_voidp s2, png_uint_32 size));
+
+extern PNG_EXPORT(png_voidp,png_memset_check) PNGARG((png_structp png_ptr,
+   png_voidp s1, int value, png_uint_32 size));
+
+#if defined(USE_FAR_KEYWORD)  /* memory model conversion function */
+extern void *png_far_to_near PNGARG((png_structp png_ptr,png_voidp ptr,
+   int check));
+#endif /* USE_FAR_KEYWORD */
+
+/* Fatal error in PNG image of libpng - can't continue */
+extern PNG_EXPORT(void,png_error) PNGARG((png_structp png_ptr,
+   png_const_charp error_message));
+
+/* The same, but the chunk name is prepended to the error string. */
+extern PNG_EXPORT(void,png_chunk_error) PNGARG((png_structp png_ptr,
+   png_const_charp error_message));
+
+#ifndef PNG_NO_WARNINGS
+/* Non-fatal error in libpng.  Can continue, but may have a problem. */
+extern PNG_EXPORT(void,png_warning) PNGARG((png_structp png_ptr,
+   png_const_charp warning_message));
+
+#ifdef PNG_READ_SUPPORTED
+/* Non-fatal error in libpng, chunk name is prepended to message. */
+extern PNG_EXPORT(void,png_chunk_warning) PNGARG((png_structp png_ptr,
+   png_const_charp warning_message));
+#endif /* PNG_READ_SUPPORTED */
+#endif /* PNG_NO_WARNINGS */
+
+/* The png_set_<chunk> functions are for storing values in the png_info_struct.
+ * Similarly, the png_get_<chunk> calls are used to read values from the
+ * png_info_struct, either storing the parameters in the passed variables, or
+ * setting pointers into the png_info_struct where the data is stored.  The
+ * png_get_<chunk> functions return a non-zero value if the data was available
+ * in info_ptr, or return zero and do not change any of the parameters if the
+ * data was not available.
+ *
+ * These functions should be used instead of directly accessing png_info
+ * to avoid problems with future changes in the size and internal layout of
+ * png_info_struct.
+ */
+/* Returns "flag" if chunk data is valid in info_ptr. */
+extern PNG_EXPORT(png_uint_32,png_get_valid) PNGARG((png_structp png_ptr,
+png_infop info_ptr, png_uint_32 flag));
+
+/* Returns number of bytes needed to hold a transformed row. */
+extern PNG_EXPORT(png_uint_32,png_get_rowbytes) PNGARG((png_structp png_ptr,
+png_infop info_ptr));
+
+#if defined(PNG_INFO_IMAGE_SUPPORTED)
+/* Returns row_pointers, which is an array of pointers to scanlines that was
+returned from png_read_png(). */
+extern PNG_EXPORT(png_bytepp,png_get_rows) PNGARG((png_structp png_ptr,
+png_infop info_ptr));
+/* Set row_pointers, which is an array of pointers to scanlines for use
+by png_write_png(). */
+extern PNG_EXPORT(void,png_set_rows) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_bytepp row_pointers));
+#endif
+
+/* Returns number of color channels in image. */
+extern PNG_EXPORT(png_byte,png_get_channels) PNGARG((png_structp png_ptr,
+png_infop info_ptr));
+
+#ifdef PNG_EASY_ACCESS_SUPPORTED
+/* Returns image width in pixels. */
+extern PNG_EXPORT(png_uint_32, png_get_image_width) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+
+/* Returns image height in pixels. */
+extern PNG_EXPORT(png_uint_32, png_get_image_height) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+
+/* Returns image bit_depth. */
+extern PNG_EXPORT(png_byte, png_get_bit_depth) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+
+/* Returns image color_type. */
+extern PNG_EXPORT(png_byte, png_get_color_type) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+
+/* Returns image filter_type. */
+extern PNG_EXPORT(png_byte, png_get_filter_type) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+
+/* Returns image interlace_type. */
+extern PNG_EXPORT(png_byte, png_get_interlace_type) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+
+/* Returns image compression_type. */
+extern PNG_EXPORT(png_byte, png_get_compression_type) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+
+/* Returns image resolution in pixels per meter, from pHYs chunk data. */
+extern PNG_EXPORT(png_uint_32, png_get_pixels_per_meter) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+extern PNG_EXPORT(png_uint_32, png_get_x_pixels_per_meter) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+extern PNG_EXPORT(png_uint_32, png_get_y_pixels_per_meter) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+
+/* Returns pixel aspect ratio, computed from pHYs chunk data.  */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(float, png_get_pixel_aspect_ratio) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+#endif
+
+/* Returns image x, y offset in pixels or microns, from oFFs chunk data. */
+extern PNG_EXPORT(png_int_32, png_get_x_offset_pixels) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+extern PNG_EXPORT(png_int_32, png_get_y_offset_pixels) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+extern PNG_EXPORT(png_int_32, png_get_x_offset_microns) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+extern PNG_EXPORT(png_int_32, png_get_y_offset_microns) PNGARG((png_structp
+png_ptr, png_infop info_ptr));
+
+#endif /* PNG_EASY_ACCESS_SUPPORTED */
+
+/* Returns pointer to signature string read from PNG header */
+extern PNG_EXPORT(png_bytep,png_get_signature) PNGARG((png_structp png_ptr,
+png_infop info_ptr));
+
+#if defined(PNG_bKGD_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_bKGD) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_color_16p *background));
+#endif
+
+#if defined(PNG_bKGD_SUPPORTED)
+extern PNG_EXPORT(void,png_set_bKGD) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_color_16p background));
+#endif
+
+#if defined(PNG_cHRM_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(png_uint_32,png_get_cHRM) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, double *white_x, double *white_y, double *red_x,
+   double *red_y, double *green_x, double *green_y, double *blue_x,
+   double *blue_y));
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+extern PNG_EXPORT(png_uint_32,png_get_cHRM_fixed) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_fixed_point *int_white_x, png_fixed_point
+   *int_white_y, png_fixed_point *int_red_x, png_fixed_point *int_red_y,
+   png_fixed_point *int_green_x, png_fixed_point *int_green_y, png_fixed_point
+   *int_blue_x, png_fixed_point *int_blue_y));
+#endif
+#endif
+
+#if defined(PNG_cHRM_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(void,png_set_cHRM) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, double white_x, double white_y, double red_x,
+   double red_y, double green_x, double green_y, double blue_x, double blue_y));
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+extern PNG_EXPORT(void,png_set_cHRM_fixed) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_fixed_point int_white_x, png_fixed_point int_white_y,
+   png_fixed_point int_red_x, png_fixed_point int_red_y, png_fixed_point
+   int_green_x, png_fixed_point int_green_y, png_fixed_point int_blue_x,
+   png_fixed_point int_blue_y));
+#endif
+#endif
+
+#if defined(PNG_gAMA_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(png_uint_32,png_get_gAMA) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, double *file_gamma));
+#endif
+extern PNG_EXPORT(png_uint_32,png_get_gAMA_fixed) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_fixed_point *int_file_gamma));
+#endif
+
+#if defined(PNG_gAMA_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(void,png_set_gAMA) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, double file_gamma));
+#endif
+extern PNG_EXPORT(void,png_set_gAMA_fixed) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_fixed_point int_file_gamma));
+#endif
+
+#if defined(PNG_hIST_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_hIST) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_16p *hist));
+#endif
+
+#if defined(PNG_hIST_SUPPORTED)
+extern PNG_EXPORT(void,png_set_hIST) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_16p hist));
+#endif
+
+extern PNG_EXPORT(png_uint_32,png_get_IHDR) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 *width, png_uint_32 *height,
+   int *bit_depth, int *color_type, int *interlace_method,
+   int *compression_method, int *filter_method));
+
+extern PNG_EXPORT(void,png_set_IHDR) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 width, png_uint_32 height, int bit_depth,
+   int color_type, int interlace_method, int compression_method,
+   int filter_method));
+
+#if defined(PNG_oFFs_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_oFFs) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_int_32 *offset_x, png_int_32 *offset_y,
+   int *unit_type));
+#endif
+
+#if defined(PNG_oFFs_SUPPORTED)
+extern PNG_EXPORT(void,png_set_oFFs) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_int_32 offset_x, png_int_32 offset_y,
+   int unit_type));
+#endif
+
+#if defined(PNG_pCAL_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_pCAL) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_charp *purpose, png_int_32 *X0, png_int_32 *X1,
+   int *type, int *nparams, png_charp *units, png_charpp *params));
+#endif
+
+#if defined(PNG_pCAL_SUPPORTED)
+extern PNG_EXPORT(void,png_set_pCAL) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_charp purpose, png_int_32 X0, png_int_32 X1,
+   int type, int nparams, png_charp units, png_charpp params));
+#endif
+
+#if defined(PNG_pHYs_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_pHYs) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 *res_x, png_uint_32 *res_y, int *unit_type));
+#endif
+
+#if defined(PNG_pHYs_SUPPORTED)
+extern PNG_EXPORT(void,png_set_pHYs) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 res_x, png_uint_32 res_y, int unit_type));
+#endif
+
+extern PNG_EXPORT(png_uint_32,png_get_PLTE) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_colorp *palette, int *num_palette));
+
+extern PNG_EXPORT(void,png_set_PLTE) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_colorp palette, int num_palette));
+
+#if defined(PNG_sBIT_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_sBIT) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_color_8p *sig_bit));
+#endif
+
+#if defined(PNG_sBIT_SUPPORTED)
+extern PNG_EXPORT(void,png_set_sBIT) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_color_8p sig_bit));
+#endif
+
+#if defined(PNG_sRGB_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_sRGB) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, int *intent));
+#endif
+
+#if defined(PNG_sRGB_SUPPORTED)
+extern PNG_EXPORT(void,png_set_sRGB) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, int intent));
+extern PNG_EXPORT(void,png_set_sRGB_gAMA_and_cHRM) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, int intent));
+#endif
+
+#if defined(PNG_iCCP_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_iCCP) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_charpp name, int *compression_type,
+   png_charpp profile, png_uint_32 *proflen));
+   /* Note to maintainer: profile should be png_bytepp */
+#endif
+
+#if defined(PNG_iCCP_SUPPORTED)
+extern PNG_EXPORT(void,png_set_iCCP) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_charp name, int compression_type,
+   png_charp profile, png_uint_32 proflen));
+   /* Note to maintainer: profile should be png_bytep */
+#endif
+
+#if defined(PNG_sPLT_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_sPLT) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_sPLT_tpp entries));
+#endif
+
+#if defined(PNG_sPLT_SUPPORTED)
+extern PNG_EXPORT(void,png_set_sPLT) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_sPLT_tp entries, int nentries));
+#endif
+
+#if defined(PNG_TEXT_SUPPORTED)
+/* png_get_text also returns the number of text chunks in *num_text */
+extern PNG_EXPORT(png_uint_32,png_get_text) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_textp *text_ptr, int *num_text));
+#endif
+
+/*
+ *  Note while png_set_text() will accept a structure whose text,
+ *  language, and  translated keywords are NULL pointers, the structure
+ *  returned by png_get_text will always contain regular
+ *  zero-terminated C strings.  They might be empty strings but
+ *  they will never be NULL pointers.
+ */
+
+#if defined(PNG_TEXT_SUPPORTED)
+extern PNG_EXPORT(void,png_set_text) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_textp text_ptr, int num_text));
+#endif
+
+#if defined(PNG_tIME_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_tIME) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_timep *mod_time));
+#endif
+
+#if defined(PNG_tIME_SUPPORTED)
+extern PNG_EXPORT(void,png_set_tIME) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_timep mod_time));
+#endif
+
+#if defined(PNG_tRNS_SUPPORTED)
+extern PNG_EXPORT(png_uint_32,png_get_tRNS) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_bytep *trans, int *num_trans,
+   png_color_16p *trans_values));
+#endif
+
+#if defined(PNG_tRNS_SUPPORTED)
+extern PNG_EXPORT(void,png_set_tRNS) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_bytep trans, int num_trans,
+   png_color_16p trans_values));
+#endif
+
+#if defined(PNG_tRNS_SUPPORTED)
+#endif
+
+#if defined(PNG_sCAL_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(png_uint_32,png_get_sCAL) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, int *unit, double *width, double *height));
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+extern PNG_EXPORT(png_uint_32,png_get_sCAL_s) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, int *unit, png_charpp swidth, png_charpp sheight));
+#endif
+#endif
+#endif /* PNG_sCAL_SUPPORTED */
+
+#if defined(PNG_sCAL_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+extern PNG_EXPORT(void,png_set_sCAL) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, int unit, double width, double height));
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+extern PNG_EXPORT(void,png_set_sCAL_s) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, int unit, png_charp swidth, png_charp sheight));
+#endif
+#endif
+#endif /* PNG_sCAL_SUPPORTED || PNG_WRITE_sCAL_SUPPORTED */
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+/* provide a list of chunks and how they are to be handled, if the built-in
+   handling or default unknown chunk handling is not desired.  Any chunks not
+   listed will be handled in the default manner.  The IHDR and IEND chunks
+   must not be listed.
+      keep = 0: follow default behaviour
+           = 1: do not keep
+           = 2: keep only if safe-to-copy
+           = 3: keep even if unsafe-to-copy
+*/
+extern PNG_EXPORT(void, png_set_keep_unknown_chunks) PNGARG((png_structp
+   png_ptr, int keep, png_bytep chunk_list, int num_chunks));
+extern PNG_EXPORT(void, png_set_unknown_chunks) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_unknown_chunkp unknowns, int num_unknowns));
+extern PNG_EXPORT(void, png_set_unknown_chunk_location)
+   PNGARG((png_structp png_ptr, png_infop info_ptr, int chunk, int location));
+extern PNG_EXPORT(png_uint_32,png_get_unknown_chunks) PNGARG((png_structp
+   png_ptr, png_infop info_ptr, png_unknown_chunkpp entries));
+#endif
+#ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED
+PNG_EXPORT(int,png_handle_as_unknown) PNGARG((png_structp png_ptr, png_bytep
+   chunk_name));
+#endif
+
+/* Png_free_data() will turn off the "valid" flag for anything it frees.
+   If you need to turn it off for a chunk that your application has freed,
+   you can use png_set_invalid(png_ptr, info_ptr, PNG_INFO_CHNK); */
+extern PNG_EXPORT(void, png_set_invalid) PNGARG((png_structp png_ptr,
+   png_infop info_ptr, int mask));
+
+#if defined(PNG_INFO_IMAGE_SUPPORTED)
+/* The "params" pointer is currently not used and is for future expansion. */
+extern PNG_EXPORT(void, png_read_png) PNGARG((png_structp png_ptr,
+                        png_infop info_ptr,
+                        int transforms,
+                        png_voidp params));
+extern PNG_EXPORT(void, png_write_png) PNGARG((png_structp png_ptr,
+                        png_infop info_ptr,
+                        int transforms,
+                        png_voidp params));
+#endif
+
+/* Define PNG_DEBUG at compile time for debugging information.  Higher
+ * numbers for PNG_DEBUG mean more debugging information.  This has
+ * only been added since version 0.95 so it is not implemented throughout
+ * libpng yet, but more support will be added as needed.
+ */
+#ifdef PNG_DEBUG
+#if (PNG_DEBUG > 0)
+#if !defined(PNG_DEBUG_FILE) && defined(_MSC_VER)
+#include <crtdbg.h>
+#if (PNG_DEBUG > 1)
+#define png_debug(l,m)  _RPT0(_CRT_WARN,m)
+#define png_debug1(l,m,p1)  _RPT1(_CRT_WARN,m,p1)
+#define png_debug2(l,m,p1,p2) _RPT2(_CRT_WARN,m,p1,p2)
+#endif
+#else /* PNG_DEBUG_FILE || !_MSC_VER */
+#ifndef PNG_DEBUG_FILE
+#define PNG_DEBUG_FILE stderr
+#endif /* PNG_DEBUG_FILE */
+#if (PNG_DEBUG > 1)
+#define png_debug(l,m) \
+{ \
+     int num_tabs=l; \
+     fprintf(PNG_DEBUG_FILE,"%s"m,(num_tabs==1 ? "\t" : \
+       (num_tabs==2 ? "\t\t":(num_tabs>2 ? "\t\t\t":"")))); \
+}
+#define png_debug1(l,m,p1) \
+{ \
+     int num_tabs=l; \
+     fprintf(PNG_DEBUG_FILE,"%s"m,(num_tabs==1 ? "\t" : \
+       (num_tabs==2 ? "\t\t":(num_tabs>2 ? "\t\t\t":""))),p1); \
+}
+#define png_debug2(l,m,p1,p2) \
+{ \
+     int num_tabs=l; \
+     fprintf(PNG_DEBUG_FILE,"%s"m,(num_tabs==1 ? "\t" : \
+       (num_tabs==2 ? "\t\t":(num_tabs>2 ? "\t\t\t":""))),p1,p2); \
+}
+#endif /* (PNG_DEBUG > 1) */
+#endif /* _MSC_VER */
+#endif /* (PNG_DEBUG > 0) */
+#endif /* PNG_DEBUG */
+#ifndef png_debug
+#define png_debug(l, m)
+#endif
+#ifndef png_debug1
+#define png_debug1(l, m, p1)
+#endif
+#ifndef png_debug2
+#define png_debug2(l, m, p1, p2)
+#endif
+
+extern PNG_EXPORT(png_charp,png_get_copyright) PNGARG((png_structp png_ptr));
+extern PNG_EXPORT(png_charp,png_get_header_ver) PNGARG((png_structp png_ptr));
+extern PNG_EXPORT(png_charp,png_get_header_version) PNGARG((png_structp png_ptr));
+extern PNG_EXPORT(png_charp,png_get_libpng_ver) PNGARG((png_structp png_ptr));
+
+#ifdef PNG_MNG_FEATURES_SUPPORTED
+extern PNG_EXPORT(png_uint_32,png_permit_mng_features) PNGARG((png_structp
+   png_ptr, png_uint_32 mng_features_permitted));
+#endif
+
+/* For use in png_set_keep_unknown, added to version 1.2.6 */
+#define PNG_HANDLE_CHUNK_AS_DEFAULT   0
+#define PNG_HANDLE_CHUNK_NEVER        1
+#define PNG_HANDLE_CHUNK_IF_SAFE      2
+#define PNG_HANDLE_CHUNK_ALWAYS       3
+
+/* Added to version 1.2.0 */
+#if defined(PNG_ASSEMBLER_CODE_SUPPORTED)
+#if defined(PNG_MMX_CODE_SUPPORTED)
+#define PNG_ASM_FLAG_MMX_SUPPORT_COMPILED  0x01  /* not user-settable */
+#define PNG_ASM_FLAG_MMX_SUPPORT_IN_CPU    0x02  /* not user-settable */
+#define PNG_ASM_FLAG_MMX_READ_COMBINE_ROW  0x04
+#define PNG_ASM_FLAG_MMX_READ_INTERLACE    0x08
+#define PNG_ASM_FLAG_MMX_READ_FILTER_SUB   0x10
+#define PNG_ASM_FLAG_MMX_READ_FILTER_UP    0x20
+#define PNG_ASM_FLAG_MMX_READ_FILTER_AVG   0x40
+#define PNG_ASM_FLAG_MMX_READ_FILTER_PAETH 0x80
+#define PNG_ASM_FLAGS_INITIALIZED          0x80000000  /* not user-settable */
+
+#define PNG_MMX_READ_FLAGS ( PNG_ASM_FLAG_MMX_READ_COMBINE_ROW  \
+                           | PNG_ASM_FLAG_MMX_READ_INTERLACE    \
+                           | PNG_ASM_FLAG_MMX_READ_FILTER_SUB   \
+                           | PNG_ASM_FLAG_MMX_READ_FILTER_UP    \
+                           | PNG_ASM_FLAG_MMX_READ_FILTER_AVG   \
+                           | PNG_ASM_FLAG_MMX_READ_FILTER_PAETH )
+#define PNG_MMX_WRITE_FLAGS ( 0 )
+
+#define PNG_MMX_FLAGS ( PNG_ASM_FLAG_MMX_SUPPORT_COMPILED \
+                      | PNG_ASM_FLAG_MMX_SUPPORT_IN_CPU   \
+                      | PNG_MMX_READ_FLAGS                \
+                      | PNG_MMX_WRITE_FLAGS )
+
+#define PNG_SELECT_READ   1
+#define PNG_SELECT_WRITE  2
+#endif /* PNG_MMX_CODE_SUPPORTED */
+
+#if !defined(PNG_1_0_X)
+/* pngget.c */
+extern PNG_EXPORT(png_uint_32,png_get_mmx_flagmask)
+   PNGARG((int flag_select, int *compilerID));
+
+/* pngget.c */
+extern PNG_EXPORT(png_uint_32,png_get_asm_flagmask)
+   PNGARG((int flag_select));
+
+/* pngget.c */
+extern PNG_EXPORT(png_uint_32,png_get_asm_flags)
+   PNGARG((png_structp png_ptr));
+
+/* pngget.c */
+extern PNG_EXPORT(png_byte,png_get_mmx_bitdepth_threshold)
+   PNGARG((png_structp png_ptr));
+
+/* pngget.c */
+extern PNG_EXPORT(png_uint_32,png_get_mmx_rowbytes_threshold)
+   PNGARG((png_structp png_ptr));
+
+/* pngset.c */
+extern PNG_EXPORT(void,png_set_asm_flags)
+   PNGARG((png_structp png_ptr, png_uint_32 asm_flags));
+
+/* pngset.c */
+extern PNG_EXPORT(void,png_set_mmx_thresholds)
+   PNGARG((png_structp png_ptr, png_byte mmx_bitdepth_threshold,
+   png_uint_32 mmx_rowbytes_threshold));
+
+#endif /* PNG_1_0_X */
+
+#if !defined(PNG_1_0_X)
+/* png.c, pnggccrd.c, or pngvcrd.c */
+extern PNG_EXPORT(int,png_mmx_support) PNGARG((void));
+#endif /* PNG_ASSEMBLER_CODE_SUPPORTED */
+
+/* Strip the prepended error numbers ("#nnn ") from error and warning
+ * messages before passing them to the error or warning handler. */
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+extern PNG_EXPORT(void,png_set_strip_error_numbers) PNGARG((png_structp
+   png_ptr, png_uint_32 strip_mode));
+#endif
+
+#endif /* PNG_1_0_X */
+
+/* Added at libpng-1.2.6 */
+#ifdef PNG_SET_USER_LIMITS_SUPPORTED
+extern PNG_EXPORT(void,png_set_user_limits) PNGARG((png_structp
+   png_ptr, png_uint_32 user_width_max, png_uint_32 user_height_max));
+extern PNG_EXPORT(png_uint_32,png_get_user_width_max) PNGARG((png_structp
+   png_ptr));
+extern PNG_EXPORT(png_uint_32,png_get_user_height_max) PNGARG((png_structp
+   png_ptr));
+#endif
+
+/* Maintainer: Put new public prototypes here ^, in libpng.3, and project defs */
+
+#ifdef PNG_READ_COMPOSITE_NODIV_SUPPORTED
+/* With these routines we avoid an integer divide, which will be slower on
+ * most machines.  However, it does take more operations than the corresponding
+ * divide method, so it may be slower on a few RISC systems.  There are two
+ * shifts (by 8 or 16 bits) and an addition, versus a single integer divide.
+ *
+ * Note that the rounding factors are NOT supposed to be the same!  128 and
+ * 32768 are correct for the NODIV code; 127 and 32767 are correct for the
+ * standard method.
+ *
+ * [Optimized code by Greg Roelofs and Mark Adler...blame us for bugs. :-) ]
+ */
+
+ /* fg and bg should be in `gamma 1.0' space; alpha is the opacity          */
+
+#  define png_composite(composite, fg, alpha, bg)                            \
+     { png_uint_16 temp = (png_uint_16)((png_uint_16)(fg) * (png_uint_16)(alpha) \
+                        +        (png_uint_16)(bg)*(png_uint_16)(255 -       \
+                        (png_uint_16)(alpha)) + (png_uint_16)128);           \
+       (composite) = (png_byte)((temp + (temp >> 8)) >> 8); }
+
+#  define png_composite_16(composite, fg, alpha, bg)                         \
+     { png_uint_32 temp = (png_uint_32)((png_uint_32)(fg) * (png_uint_32)(alpha) \
+                        + (png_uint_32)(bg)*(png_uint_32)(65535L -           \
+                        (png_uint_32)(alpha)) + (png_uint_32)32768L);        \
+       (composite) = (png_uint_16)((temp + (temp >> 16)) >> 16); }
+
+#else  /* standard method using integer division */
+
+#  define png_composite(composite, fg, alpha, bg)                            \
+     (composite) = (png_byte)(((png_uint_16)(fg) * (png_uint_16)(alpha) +    \
+       (png_uint_16)(bg) * (png_uint_16)(255 - (png_uint_16)(alpha)) +       \
+       (png_uint_16)127) / 255)
+
+#  define png_composite_16(composite, fg, alpha, bg)                         \
+     (composite) = (png_uint_16)(((png_uint_32)(fg) * (png_uint_32)(alpha) + \
+       (png_uint_32)(bg)*(png_uint_32)(65535L - (png_uint_32)(alpha)) +      \
+       (png_uint_32)32767) / (png_uint_32)65535L)
+
+#endif /* PNG_READ_COMPOSITE_NODIV_SUPPORTED */
+
+/* Inline macros to do direct reads of bytes from the input buffer.  These
+ * require that you are using an architecture that uses PNG byte ordering
+ * (MSB first) and supports unaligned data storage.  I think that PowerPC
+ * in big-endian mode and 680x0 are the only ones that will support this.
+ * The x86 line of processors definitely do not.  The png_get_int_32()
+ * routine also assumes we are using two's complement format for negative
+ * values, which is almost certainly true.
+ */
+#if defined(PNG_READ_BIG_ENDIAN_SUPPORTED)
+#  define png_get_uint_32(buf) ( *((png_uint_32p) (buf)))
+#  define png_get_uint_16(buf) ( *((png_uint_16p) (buf)))
+#  define png_get_int_32(buf)  ( *((png_int_32p)  (buf)))
+#else
+extern PNG_EXPORT(png_uint_32,png_get_uint_32) PNGARG((png_bytep buf));
+extern PNG_EXPORT(png_uint_16,png_get_uint_16) PNGARG((png_bytep buf));
+extern PNG_EXPORT(png_int_32,png_get_int_32) PNGARG((png_bytep buf));
+#endif /* !PNG_READ_BIG_ENDIAN_SUPPORTED */
+extern PNG_EXPORT(png_uint_32,png_get_uint_31)
+  PNGARG((png_structp png_ptr, png_bytep buf));
+/* No png_get_int_16 -- may be added if there's a real need for it. */
+
+/* Place a 32-bit number into a buffer in PNG byte order (big-endian).
+ */
+extern PNG_EXPORT(void,png_save_uint_32)
+   PNGARG((png_bytep buf, png_uint_32 i));
+extern PNG_EXPORT(void,png_save_int_32)
+   PNGARG((png_bytep buf, png_int_32 i));
+
+/* Place a 16-bit number into a buffer in PNG byte order.
+ * The parameter is declared unsigned int, not png_uint_16,
+ * just to avoid potential problems on pre-ANSI C compilers.
+ */
+extern PNG_EXPORT(void,png_save_uint_16)
+   PNGARG((png_bytep buf, unsigned int i));
+/* No png_save_int_16 -- may be added if there's a real need for it. */
+
+/* ************************************************************************* */
+
+/* These next functions are used internally in the code.  They generally
+ * shouldn't be used unless you are writing code to add or replace some
+ * functionality in libpng.  More information about most functions can
+ * be found in the files where the functions are located.
+ */
+
+
+/* Various modes of operation, that are visible to applications because
+ * they are used for unknown chunk location.
+ */
+#define PNG_HAVE_IHDR               0x01
+#define PNG_HAVE_PLTE               0x02
+#define PNG_HAVE_IDAT               0x04
+#define PNG_AFTER_IDAT              0x08 /* Have complete zlib datastream */
+#define PNG_HAVE_IEND               0x10
+
+#if defined(PNG_INTERNAL)
+
+/* More modes of operation.  Note that after an init, mode is set to
+ * zero automatically when the structure is created.
+ */
+#define PNG_HAVE_gAMA               0x20
+#define PNG_HAVE_cHRM               0x40
+#define PNG_HAVE_sRGB               0x80
+#define PNG_HAVE_CHUNK_HEADER      0x100
+#define PNG_WROTE_tIME             0x200
+#define PNG_WROTE_INFO_BEFORE_PLTE 0x400
+#define PNG_BACKGROUND_IS_GRAY     0x800
+#define PNG_HAVE_PNG_SIGNATURE    0x1000
+#define PNG_HAVE_CHUNK_AFTER_IDAT 0x2000 /* Have another chunk after IDAT */
+
+/* flags for the transformations the PNG library does on the image data */
+#define PNG_BGR                0x0001
+#define PNG_INTERLACE          0x0002
+#define PNG_PACK               0x0004
+#define PNG_SHIFT              0x0008
+#define PNG_SWAP_BYTES         0x0010
+#define PNG_INVERT_MONO        0x0020
+#define PNG_DITHER             0x0040
+#define PNG_BACKGROUND         0x0080
+#define PNG_BACKGROUND_EXPAND  0x0100
+                          /*   0x0200 unused */
+#define PNG_16_TO_8            0x0400
+#define PNG_RGBA               0x0800
+#define PNG_EXPAND             0x1000
+#define PNG_GAMMA              0x2000
+#define PNG_GRAY_TO_RGB        0x4000
+#define PNG_FILLER             0x8000L
+#define PNG_PACKSWAP          0x10000L
+#define PNG_SWAP_ALPHA        0x20000L
+#define PNG_STRIP_ALPHA       0x40000L
+#define PNG_INVERT_ALPHA      0x80000L
+#define PNG_USER_TRANSFORM   0x100000L
+#define PNG_RGB_TO_GRAY_ERR  0x200000L
+#define PNG_RGB_TO_GRAY_WARN 0x400000L
+#define PNG_RGB_TO_GRAY      0x600000L  /* two bits, RGB_TO_GRAY_ERR|WARN */
+                       /*    0x800000L     Unused */
+#define PNG_ADD_ALPHA       0x1000000L  /* Added to libpng-1.2.7 */
+#define PNG_EXPAND_tRNS     0x2000000L  /* Added to libpng-1.2.9 */
+                       /*   0x4000000L  unused */
+                       /*   0x8000000L  unused */
+                       /*  0x10000000L  unused */
+                       /*  0x20000000L  unused */
+                       /*  0x40000000L  unused */
+
+/* flags for png_create_struct */
+#define PNG_STRUCT_PNG   0x0001
+#define PNG_STRUCT_INFO  0x0002
+
+/* Scaling factor for filter heuristic weighting calculations */
+#define PNG_WEIGHT_SHIFT 8
+#define PNG_WEIGHT_FACTOR (1<<(PNG_WEIGHT_SHIFT))
+#define PNG_COST_SHIFT 3
+#define PNG_COST_FACTOR (1<<(PNG_COST_SHIFT))
+
+/* flags for the png_ptr->flags rather than declaring a byte for each one */
+#define PNG_FLAG_ZLIB_CUSTOM_STRATEGY     0x0001
+#define PNG_FLAG_ZLIB_CUSTOM_LEVEL        0x0002
+#define PNG_FLAG_ZLIB_CUSTOM_MEM_LEVEL    0x0004
+#define PNG_FLAG_ZLIB_CUSTOM_WINDOW_BITS  0x0008
+#define PNG_FLAG_ZLIB_CUSTOM_METHOD       0x0010
+#define PNG_FLAG_ZLIB_FINISHED            0x0020
+#define PNG_FLAG_ROW_INIT                 0x0040
+#define PNG_FLAG_FILLER_AFTER             0x0080
+#define PNG_FLAG_CRC_ANCILLARY_USE        0x0100
+#define PNG_FLAG_CRC_ANCILLARY_NOWARN     0x0200
+#define PNG_FLAG_CRC_CRITICAL_USE         0x0400
+#define PNG_FLAG_CRC_CRITICAL_IGNORE      0x0800
+#define PNG_FLAG_FREE_PLTE                0x1000
+#define PNG_FLAG_FREE_TRNS                0x2000
+#define PNG_FLAG_FREE_HIST                0x4000
+#define PNG_FLAG_KEEP_UNKNOWN_CHUNKS      0x8000L
+#define PNG_FLAG_KEEP_UNSAFE_CHUNKS       0x10000L
+#define PNG_FLAG_LIBRARY_MISMATCH         0x20000L
+#define PNG_FLAG_STRIP_ERROR_NUMBERS      0x40000L
+#define PNG_FLAG_STRIP_ERROR_TEXT         0x80000L
+#define PNG_FLAG_MALLOC_NULL_MEM_OK       0x100000L
+#define PNG_FLAG_ADD_ALPHA                0x200000L  /* Added to libpng-1.2.8 */
+#define PNG_FLAG_STRIP_ALPHA              0x400000L  /* Added to libpng-1.2.8 */
+                                  /*      0x800000L  unused */
+                                  /*     0x1000000L  unused */
+                                  /*     0x2000000L  unused */
+                                  /*     0x4000000L  unused */
+                                  /*     0x8000000L  unused */
+                                  /*    0x10000000L  unused */
+                                  /*    0x20000000L  unused */
+                                  /*    0x40000000L  unused */
+
+#define PNG_FLAG_CRC_ANCILLARY_MASK (PNG_FLAG_CRC_ANCILLARY_USE | \
+                                     PNG_FLAG_CRC_ANCILLARY_NOWARN)
+
+#define PNG_FLAG_CRC_CRITICAL_MASK  (PNG_FLAG_CRC_CRITICAL_USE | \
+                                     PNG_FLAG_CRC_CRITICAL_IGNORE)
+
+#define PNG_FLAG_CRC_MASK           (PNG_FLAG_CRC_ANCILLARY_MASK | \
+                                     PNG_FLAG_CRC_CRITICAL_MASK)
+
+/* save typing and make code easier to understand */
+
+#define PNG_COLOR_DIST(c1, c2) (abs((int)((c1).red) - (int)((c2).red)) + \
+   abs((int)((c1).green) - (int)((c2).green)) + \
+   abs((int)((c1).blue) - (int)((c2).blue)))
+
+/* Added to libpng-1.2.6 JB */
+#define PNG_ROWBYTES(pixel_bits, width) \
+    ((pixel_bits) >= 8 ? \
+    ((width) * (((png_uint_32)(pixel_bits)) >> 3)) : \
+    (( ((width) * ((png_uint_32)(pixel_bits))) + 7) >> 3) )
+
+/* PNG_OUT_OF_RANGE returns true if value is outside the range
+   ideal-delta..ideal+delta.  Each argument is evaluated twice.
+   "ideal" and "delta" should be constants, normally simple
+   integers, "value" a variable. Added to libpng-1.2.6 JB */
+#define PNG_OUT_OF_RANGE(value, ideal, delta) \
+        ( (value) < (ideal)-(delta) || (value) > (ideal)+(delta) )
+
+/* variables declared in png.c - only it needs to define PNG_NO_EXTERN */
+#if !defined(PNG_NO_EXTERN) || defined(PNG_ALWAYS_EXTERN)
+/* place to hold the signature string for a PNG file. */
+#ifdef PNG_USE_GLOBAL_ARRAYS
+   PNG_EXPORT_VAR (PNG_CONST png_byte FARDATA) png_sig[8];
+#else
+#endif
+#endif /* PNG_NO_EXTERN */
+
+/* Constant strings for known chunk types.  If you need to add a chunk,
+ * define the name here, and add an invocation of the macro in png.c and
+ * wherever it's needed.
+ */
+#define PNG_IHDR png_byte png_IHDR[5] = { 73,  72,  68,  82, '\0'}
+#define PNG_IDAT png_byte png_IDAT[5] = { 73,  68,  65,  84, '\0'}
+#define PNG_IEND png_byte png_IEND[5] = { 73,  69,  78,  68, '\0'}
+#define PNG_PLTE png_byte png_PLTE[5] = { 80,  76,  84,  69, '\0'}
+#define PNG_bKGD png_byte png_bKGD[5] = { 98,  75,  71,  68, '\0'}
+#define PNG_cHRM png_byte png_cHRM[5] = { 99,  72,  82,  77, '\0'}
+#define PNG_gAMA png_byte png_gAMA[5] = {103,  65,  77,  65, '\0'}
+#define PNG_hIST png_byte png_hIST[5] = {104,  73,  83,  84, '\0'}
+#define PNG_iCCP png_byte png_iCCP[5] = {105,  67,  67,  80, '\0'}
+#define PNG_iTXt png_byte png_iTXt[5] = {105,  84,  88, 116, '\0'}
+#define PNG_oFFs png_byte png_oFFs[5] = {111,  70,  70, 115, '\0'}
+#define PNG_pCAL png_byte png_pCAL[5] = {112,  67,  65,  76, '\0'}
+#define PNG_sCAL png_byte png_sCAL[5] = {115,  67,  65,  76, '\0'}
+#define PNG_pHYs png_byte png_pHYs[5] = {112,  72,  89, 115, '\0'}
+#define PNG_sBIT png_byte png_sBIT[5] = {115,  66,  73,  84, '\0'}
+#define PNG_sPLT png_byte png_sPLT[5] = {115,  80,  76,  84, '\0'}
+#define PNG_sRGB png_byte png_sRGB[5] = {115,  82,  71,  66, '\0'}
+#define PNG_tEXt png_byte png_tEXt[5] = {116,  69,  88, 116, '\0'}
+#define PNG_tIME png_byte png_tIME[5] = {116,  73,  77,  69, '\0'}
+#define PNG_tRNS png_byte png_tRNS[5] = {116,  82,  78,  83, '\0'}
+#define PNG_zTXt png_byte png_zTXt[5] = {122,  84,  88, 116, '\0'}
+
+#ifdef PNG_USE_GLOBAL_ARRAYS
+PNG_EXPORT_VAR (png_byte FARDATA) png_IHDR[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_IDAT[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_IEND[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_PLTE[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_bKGD[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_cHRM[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_gAMA[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_hIST[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_iCCP[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_iTXt[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_oFFs[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_pCAL[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_sCAL[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_pHYs[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_sBIT[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_sPLT[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_sRGB[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_tEXt[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_tIME[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_tRNS[5];
+PNG_EXPORT_VAR (png_byte FARDATA) png_zTXt[5];
+#endif /* PNG_USE_GLOBAL_ARRAYS */
+
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+/* Initialize png_ptr struct for reading, and allocate any other memory.
+ * (old interface - DEPRECATED - use png_create_read_struct instead).
+ */
+extern PNG_EXPORT(void,png_read_init) PNGARG((png_structp png_ptr));
+#undef png_read_init
+#define png_read_init(png_ptr) png_read_init_3(&png_ptr, \
+    PNG_LIBPNG_VER_STRING,  png_sizeof(png_struct));
+#endif
+
+extern PNG_EXPORT(void,png_read_init_3) PNGARG((png_structpp ptr_ptr,
+    png_const_charp user_png_ver, png_size_t png_struct_size));
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+extern PNG_EXPORT(void,png_read_init_2) PNGARG((png_structp png_ptr,
+    png_const_charp user_png_ver, png_size_t png_struct_size, png_size_t
+    png_info_size));
+#endif
+
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+/* Initialize png_ptr struct for writing, and allocate any other memory.
+ * (old interface - DEPRECATED - use png_create_write_struct instead).
+ */
+extern PNG_EXPORT(void,png_write_init) PNGARG((png_structp png_ptr));
+#undef png_write_init
+#define png_write_init(png_ptr) png_write_init_3(&png_ptr, \
+    PNG_LIBPNG_VER_STRING, png_sizeof(png_struct));
+#endif
+
+extern PNG_EXPORT(void,png_write_init_3) PNGARG((png_structpp ptr_ptr,
+    png_const_charp user_png_ver, png_size_t png_struct_size));
+extern PNG_EXPORT(void,png_write_init_2) PNGARG((png_structp png_ptr,
+    png_const_charp user_png_ver, png_size_t png_struct_size, png_size_t
+    png_info_size));
+
+/* Allocate memory for an internal libpng struct */
+PNG_EXTERN png_voidp png_create_struct PNGARG((int type));
+
+/* Free memory from internal libpng struct */
+PNG_EXTERN void png_destroy_struct PNGARG((png_voidp struct_ptr));
+
+PNG_EXTERN png_voidp png_create_struct_2 PNGARG((int type, png_malloc_ptr
+  malloc_fn, png_voidp mem_ptr));
+PNG_EXTERN void png_destroy_struct_2 PNGARG((png_voidp struct_ptr,
+   png_free_ptr free_fn, png_voidp mem_ptr));
+
+/* Free any memory that info_ptr points to and reset struct. */
+PNG_EXTERN void png_info_destroy PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+
+#ifndef PNG_1_0_X
+/* Function to allocate memory for zlib. */
+PNG_EXTERN voidpf png_zalloc PNGARG((voidpf png_ptr, uInt items, uInt size));
+
+/* Function to free memory for zlib */
+PNG_EXTERN void png_zfree PNGARG((voidpf png_ptr, voidpf ptr));
+
+#ifdef PNG_SIZE_T
+/* Function to convert a sizeof an item to png_sizeof item */
+   PNG_EXTERN png_size_t PNGAPI png_convert_size PNGARG((size_t size));
+#endif
+
+/* Next four functions are used internally as callbacks.  PNGAPI is required
+ * but not PNG_EXPORT.  PNGAPI added at libpng version 1.2.3. */
+
+PNG_EXTERN void PNGAPI png_default_read_data PNGARG((png_structp png_ptr,
+   png_bytep data, png_size_t length));
+
+#ifdef PNG_PROGRESSIVE_READ_SUPPORTED
+PNG_EXTERN void PNGAPI png_push_fill_buffer PNGARG((png_structp png_ptr,
+   png_bytep buffer, png_size_t length));
+#endif
+
+PNG_EXTERN void PNGAPI png_default_write_data PNGARG((png_structp png_ptr,
+   png_bytep data, png_size_t length));
+
+#if defined(PNG_WRITE_FLUSH_SUPPORTED)
+#if !defined(PNG_NO_STDIO)
+PNG_EXTERN void PNGAPI png_default_flush PNGARG((png_structp png_ptr));
+#endif
+#endif
+#else /* PNG_1_0_X */
+#ifdef PNG_PROGRESSIVE_READ_SUPPORTED
+PNG_EXTERN void png_push_fill_buffer PNGARG((png_structp png_ptr,
+   png_bytep buffer, png_size_t length));
+#endif
+#endif /* PNG_1_0_X */
+
+/* Reset the CRC variable */
+PNG_EXTERN void png_reset_crc PNGARG((png_structp png_ptr));
+
+/* Write the "data" buffer to whatever output you are using. */
+PNG_EXTERN void png_write_data PNGARG((png_structp png_ptr, png_bytep data,
+   png_size_t length));
+
+/* Read data from whatever input you are using into the "data" buffer */
+PNG_EXTERN void png_read_data PNGARG((png_structp png_ptr, png_bytep data,
+   png_size_t length));
+
+/* Read bytes into buf, and update png_ptr->crc */
+PNG_EXTERN void png_crc_read PNGARG((png_structp png_ptr, png_bytep buf,
+   png_size_t length));
+
+/* Decompress data in a chunk that uses compression */
+#if defined(PNG_zTXt_SUPPORTED) || defined(PNG_iTXt_SUPPORTED) || \
+    defined(PNG_iCCP_SUPPORTED) || defined(PNG_sPLT_SUPPORTED)
+PNG_EXTERN png_charp png_decompress_chunk PNGARG((png_structp png_ptr,
+   int comp_type, png_charp chunkdata, png_size_t chunklength,
+   png_size_t prefix_length, png_size_t *data_length));
+#endif
+
+/* Read "skip" bytes, read the file crc, and (optionally) verify png_ptr->crc */
+PNG_EXTERN int png_crc_finish PNGARG((png_structp png_ptr, png_uint_32 skip));
+
+/* Read the CRC from the file and compare it to the libpng calculated CRC */
+PNG_EXTERN int png_crc_error PNGARG((png_structp png_ptr));
+
+/* Calculate the CRC over a section of data.  Note that we are only
+ * passing a maximum of 64K on systems that have this as a memory limit,
+ * since this is the maximum buffer size we can specify.
+ */
+PNG_EXTERN void png_calculate_crc PNGARG((png_structp png_ptr, png_bytep ptr,
+   png_size_t length));
+
+#if defined(PNG_WRITE_FLUSH_SUPPORTED)
+PNG_EXTERN void png_flush PNGARG((png_structp png_ptr));
+#endif
+
+/* simple function to write the signature */
+PNG_EXTERN void png_write_sig PNGARG((png_structp png_ptr));
+
+/* write various chunks */
+
+/* Write the IHDR chunk, and update the png_struct with the necessary
+ * information.
+ */
+PNG_EXTERN void png_write_IHDR PNGARG((png_structp png_ptr, png_uint_32 width,
+   png_uint_32 height,
+   int bit_depth, int color_type, int compression_method, int filter_method,
+   int interlace_method));
+
+PNG_EXTERN void png_write_PLTE PNGARG((png_structp png_ptr, png_colorp palette,
+   png_uint_32 num_pal));
+
+PNG_EXTERN void png_write_IDAT PNGARG((png_structp png_ptr, png_bytep data,
+   png_size_t length));
+
+PNG_EXTERN void png_write_IEND PNGARG((png_structp png_ptr));
+
+#if defined(PNG_WRITE_gAMA_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+PNG_EXTERN void png_write_gAMA PNGARG((png_structp png_ptr, double file_gamma));
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+PNG_EXTERN void png_write_gAMA_fixed PNGARG((png_structp png_ptr, png_fixed_point
+    file_gamma));
+#endif
+#endif
+
+#if defined(PNG_WRITE_sBIT_SUPPORTED)
+PNG_EXTERN void png_write_sBIT PNGARG((png_structp png_ptr, png_color_8p sbit,
+   int color_type));
+#endif
+
+#if defined(PNG_WRITE_cHRM_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+PNG_EXTERN void png_write_cHRM PNGARG((png_structp png_ptr,
+   double white_x, double white_y,
+   double red_x, double red_y, double green_x, double green_y,
+   double blue_x, double blue_y));
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+PNG_EXTERN void png_write_cHRM_fixed PNGARG((png_structp png_ptr,
+   png_fixed_point int_white_x, png_fixed_point int_white_y,
+   png_fixed_point int_red_x, png_fixed_point int_red_y, png_fixed_point
+   int_green_x, png_fixed_point int_green_y, png_fixed_point int_blue_x,
+   png_fixed_point int_blue_y));
+#endif
+#endif
+
+#if defined(PNG_WRITE_sRGB_SUPPORTED)
+PNG_EXTERN void png_write_sRGB PNGARG((png_structp png_ptr,
+   int intent));
+#endif
+
+#if defined(PNG_WRITE_iCCP_SUPPORTED)
+PNG_EXTERN void png_write_iCCP PNGARG((png_structp png_ptr,
+   png_charp name, int compression_type,
+   png_charp profile, int proflen));
+   /* Note to maintainer: profile should be png_bytep */
+#endif
+
+#if defined(PNG_WRITE_sPLT_SUPPORTED)
+PNG_EXTERN void png_write_sPLT PNGARG((png_structp png_ptr,
+   png_sPLT_tp palette));
+#endif
+
+#if defined(PNG_WRITE_tRNS_SUPPORTED)
+PNG_EXTERN void png_write_tRNS PNGARG((png_structp png_ptr, png_bytep trans,
+   png_color_16p values, int number, int color_type));
+#endif
+
+#if defined(PNG_WRITE_bKGD_SUPPORTED)
+PNG_EXTERN void png_write_bKGD PNGARG((png_structp png_ptr,
+   png_color_16p values, int color_type));
+#endif
+
+#if defined(PNG_WRITE_hIST_SUPPORTED)
+PNG_EXTERN void png_write_hIST PNGARG((png_structp png_ptr, png_uint_16p hist,
+   int num_hist));
+#endif
+
+#if defined(PNG_WRITE_TEXT_SUPPORTED) || defined(PNG_WRITE_pCAL_SUPPORTED) || \
+    defined(PNG_WRITE_iCCP_SUPPORTED) || defined(PNG_WRITE_sPLT_SUPPORTED)
+PNG_EXTERN png_size_t png_check_keyword PNGARG((png_structp png_ptr,
+   png_charp key, png_charpp new_key));
+#endif
+
+#if defined(PNG_WRITE_tEXt_SUPPORTED)
+PNG_EXTERN void png_write_tEXt PNGARG((png_structp png_ptr, png_charp key,
+   png_charp text, png_size_t text_len));
+#endif
+
+#if defined(PNG_WRITE_zTXt_SUPPORTED)
+PNG_EXTERN void png_write_zTXt PNGARG((png_structp png_ptr, png_charp key,
+   png_charp text, png_size_t text_len, int compression));
+#endif
+
+#if defined(PNG_WRITE_iTXt_SUPPORTED)
+PNG_EXTERN void png_write_iTXt PNGARG((png_structp png_ptr,
+   int compression, png_charp key, png_charp lang, png_charp lang_key,
+   png_charp text));
+#endif
+
+#if defined(PNG_TEXT_SUPPORTED)  /* Added at version 1.0.14 and 1.2.4 */
+PNG_EXTERN int png_set_text_2 PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_textp text_ptr, int num_text));
+#endif
+
+#if defined(PNG_WRITE_oFFs_SUPPORTED)
+PNG_EXTERN void png_write_oFFs PNGARG((png_structp png_ptr,
+   png_int_32 x_offset, png_int_32 y_offset, int unit_type));
+#endif
+
+#if defined(PNG_WRITE_pCAL_SUPPORTED)
+PNG_EXTERN void png_write_pCAL PNGARG((png_structp png_ptr, png_charp purpose,
+   png_int_32 X0, png_int_32 X1, int type, int nparams,
+   png_charp units, png_charpp params));
+#endif
+
+#if defined(PNG_WRITE_pHYs_SUPPORTED)
+PNG_EXTERN void png_write_pHYs PNGARG((png_structp png_ptr,
+   png_uint_32 x_pixels_per_unit, png_uint_32 y_pixels_per_unit,
+   int unit_type));
+#endif
+
+#if defined(PNG_WRITE_tIME_SUPPORTED)
+PNG_EXTERN void png_write_tIME PNGARG((png_structp png_ptr,
+   png_timep mod_time));
+#endif
+
+#if defined(PNG_WRITE_sCAL_SUPPORTED)
+#if defined(PNG_FLOATING_POINT_SUPPORTED) && !defined(PNG_NO_STDIO)
+PNG_EXTERN void png_write_sCAL PNGARG((png_structp png_ptr,
+   int unit, double width, double height));
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+PNG_EXTERN void png_write_sCAL_s PNGARG((png_structp png_ptr,
+   int unit, png_charp width, png_charp height));
+#endif
+#endif
+#endif
+
+/* Called when finished processing a row of data */
+PNG_EXTERN void png_write_finish_row PNGARG((png_structp png_ptr));
+
+/* Internal use only.   Called before first row of data */
+PNG_EXTERN void png_write_start_row PNGARG((png_structp png_ptr));
+
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+PNG_EXTERN void png_build_gamma_table PNGARG((png_structp png_ptr));
+#endif
+
+/* combine a row of data, dealing with alpha, etc. if requested */
+PNG_EXTERN void png_combine_row PNGARG((png_structp png_ptr, png_bytep row,
+   int mask));
+
+#if defined(PNG_READ_INTERLACING_SUPPORTED)
+/* expand an interlaced row */
+/* OLD pre-1.0.9 interface:
+PNG_EXTERN void png_do_read_interlace PNGARG((png_row_infop row_info,
+   png_bytep row, int pass, png_uint_32 transformations));
+ */
+PNG_EXTERN void png_do_read_interlace PNGARG((png_structp png_ptr));
+#endif
+
+/* GRR TO DO (2.0 or whenever):  simplify other internal calling interfaces */
+
+#if defined(PNG_WRITE_INTERLACING_SUPPORTED)
+/* grab pixels out of a row for an interlaced pass */
+PNG_EXTERN void png_do_write_interlace PNGARG((png_row_infop row_info,
+   png_bytep row, int pass));
+#endif
+
+/* unfilter a row */
+PNG_EXTERN void png_read_filter_row PNGARG((png_structp png_ptr,
+   png_row_infop row_info, png_bytep row, png_bytep prev_row, int filter));
+
+/* Choose the best filter to use and filter the row data */
+PNG_EXTERN void png_write_find_filter PNGARG((png_structp png_ptr,
+   png_row_infop row_info));
+
+/* Write out the filtered row. */
+PNG_EXTERN void png_write_filtered_row PNGARG((png_structp png_ptr,
+   png_bytep filtered_row));
+/* finish a row while reading, dealing with interlacing passes, etc. */
+PNG_EXTERN void png_read_finish_row PNGARG((png_structp png_ptr));
+
+/* initialize the row buffers, etc. */
+PNG_EXTERN void png_read_start_row PNGARG((png_structp png_ptr));
+/* optional call to update the users info structure */
+PNG_EXTERN void png_read_transform_info PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+
+/* these are the functions that do the transformations */
+#if defined(PNG_READ_FILLER_SUPPORTED)
+PNG_EXTERN void png_do_read_filler PNGARG((png_row_infop row_info,
+   png_bytep row, png_uint_32 filler, png_uint_32 flags));
+#endif
+
+#if defined(PNG_READ_SWAP_ALPHA_SUPPORTED)
+PNG_EXTERN void png_do_read_swap_alpha PNGARG((png_row_infop row_info,
+   png_bytep row));
+#endif
+
+#if defined(PNG_WRITE_SWAP_ALPHA_SUPPORTED)
+PNG_EXTERN void png_do_write_swap_alpha PNGARG((png_row_infop row_info,
+   png_bytep row));
+#endif
+
+#if defined(PNG_READ_INVERT_ALPHA_SUPPORTED)
+PNG_EXTERN void png_do_read_invert_alpha PNGARG((png_row_infop row_info,
+   png_bytep row));
+#endif
+
+#if defined(PNG_WRITE_INVERT_ALPHA_SUPPORTED)
+PNG_EXTERN void png_do_write_invert_alpha PNGARG((png_row_infop row_info,
+   png_bytep row));
+#endif
+
+#if defined(PNG_WRITE_FILLER_SUPPORTED) || \
+    defined(PNG_READ_STRIP_ALPHA_SUPPORTED)
+PNG_EXTERN void png_do_strip_filler PNGARG((png_row_infop row_info,
+   png_bytep row, png_uint_32 flags));
+#endif
+
+#if defined(PNG_READ_SWAP_SUPPORTED) || defined(PNG_WRITE_SWAP_SUPPORTED)
+PNG_EXTERN void png_do_swap PNGARG((png_row_infop row_info, png_bytep row));
+#endif
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED) || defined(PNG_WRITE_PACKSWAP_SUPPORTED)
+PNG_EXTERN void png_do_packswap PNGARG((png_row_infop row_info, png_bytep row));
+#endif
+
+#if defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+PNG_EXTERN int png_do_rgb_to_gray PNGARG((png_structp png_ptr, png_row_infop
+   row_info, png_bytep row));
+#endif
+
+#if defined(PNG_READ_GRAY_TO_RGB_SUPPORTED)
+PNG_EXTERN void png_do_gray_to_rgb PNGARG((png_row_infop row_info,
+   png_bytep row));
+#endif
+
+#if defined(PNG_READ_PACK_SUPPORTED)
+PNG_EXTERN void png_do_unpack PNGARG((png_row_infop row_info, png_bytep row));
+#endif
+
+#if defined(PNG_READ_SHIFT_SUPPORTED)
+PNG_EXTERN void png_do_unshift PNGARG((png_row_infop row_info, png_bytep row,
+   png_color_8p sig_bits));
+#endif
+
+#if defined(PNG_READ_INVERT_SUPPORTED) || defined(PNG_WRITE_INVERT_SUPPORTED)
+PNG_EXTERN void png_do_invert PNGARG((png_row_infop row_info, png_bytep row));
+#endif
+
+#if defined(PNG_READ_16_TO_8_SUPPORTED)
+PNG_EXTERN void png_do_chop PNGARG((png_row_infop row_info, png_bytep row));
+#endif
+
+#if defined(PNG_READ_DITHER_SUPPORTED)
+PNG_EXTERN void png_do_dither PNGARG((png_row_infop row_info,
+   png_bytep row, png_bytep palette_lookup, png_bytep dither_lookup));
+
+#  if defined(PNG_CORRECT_PALETTE_SUPPORTED)
+PNG_EXTERN void png_correct_palette PNGARG((png_structp png_ptr,
+   png_colorp palette, int num_palette));
+#  endif
+#endif
+
+#if defined(PNG_READ_BGR_SUPPORTED) || defined(PNG_WRITE_BGR_SUPPORTED)
+PNG_EXTERN void png_do_bgr PNGARG((png_row_infop row_info, png_bytep row));
+#endif
+
+#if defined(PNG_WRITE_PACK_SUPPORTED)
+PNG_EXTERN void png_do_pack PNGARG((png_row_infop row_info,
+   png_bytep row, png_uint_32 bit_depth));
+#endif
+
+#if defined(PNG_WRITE_SHIFT_SUPPORTED)
+PNG_EXTERN void png_do_shift PNGARG((png_row_infop row_info, png_bytep row,
+   png_color_8p bit_depth));
+#endif
+
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+PNG_EXTERN void png_do_background PNGARG((png_row_infop row_info, png_bytep row,
+   png_color_16p trans_values, png_color_16p background,
+   png_color_16p background_1,
+   png_bytep gamma_table, png_bytep gamma_from_1, png_bytep gamma_to_1,
+   png_uint_16pp gamma_16, png_uint_16pp gamma_16_from_1,
+   png_uint_16pp gamma_16_to_1, int gamma_shift));
+#else
+PNG_EXTERN void png_do_background PNGARG((png_row_infop row_info, png_bytep row,
+   png_color_16p trans_values, png_color_16p background));
+#endif
+#endif
+
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+PNG_EXTERN void png_do_gamma PNGARG((png_row_infop row_info, png_bytep row,
+   png_bytep gamma_table, png_uint_16pp gamma_16_table,
+   int gamma_shift));
+#endif
+
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+PNG_EXTERN void png_do_expand_palette PNGARG((png_row_infop row_info,
+   png_bytep row, png_colorp palette, png_bytep trans, int num_trans));
+PNG_EXTERN void png_do_expand PNGARG((png_row_infop row_info,
+   png_bytep row, png_color_16p trans_value));
+#endif
+
+/* The following decodes the appropriate chunks, and does error correction,
+ * then calls the appropriate callback for the chunk if it is valid.
+ */
+
+/* decode the IHDR chunk */
+PNG_EXTERN void png_handle_IHDR PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+PNG_EXTERN void png_handle_PLTE PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+PNG_EXTERN void png_handle_IEND PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+
+#if defined(PNG_READ_bKGD_SUPPORTED)
+PNG_EXTERN void png_handle_bKGD PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_cHRM_SUPPORTED)
+PNG_EXTERN void png_handle_cHRM PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_gAMA_SUPPORTED)
+PNG_EXTERN void png_handle_gAMA PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_hIST_SUPPORTED)
+PNG_EXTERN void png_handle_hIST PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_iCCP_SUPPORTED)
+extern void png_handle_iCCP PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif /* PNG_READ_iCCP_SUPPORTED */
+
+#if defined(PNG_READ_iTXt_SUPPORTED)
+PNG_EXTERN void png_handle_iTXt PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_oFFs_SUPPORTED)
+PNG_EXTERN void png_handle_oFFs PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_pCAL_SUPPORTED)
+PNG_EXTERN void png_handle_pCAL PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_pHYs_SUPPORTED)
+PNG_EXTERN void png_handle_pHYs PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_sBIT_SUPPORTED)
+PNG_EXTERN void png_handle_sBIT PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_sCAL_SUPPORTED)
+PNG_EXTERN void png_handle_sCAL PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_sPLT_SUPPORTED)
+extern void png_handle_sPLT PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif /* PNG_READ_sPLT_SUPPORTED */
+
+#if defined(PNG_READ_sRGB_SUPPORTED)
+PNG_EXTERN void png_handle_sRGB PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_tEXt_SUPPORTED)
+PNG_EXTERN void png_handle_tEXt PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_tIME_SUPPORTED)
+PNG_EXTERN void png_handle_tIME PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_tRNS_SUPPORTED)
+PNG_EXTERN void png_handle_tRNS PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+#if defined(PNG_READ_zTXt_SUPPORTED)
+PNG_EXTERN void png_handle_zTXt PNGARG((png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 length));
+#endif
+
+PNG_EXTERN void png_handle_unknown PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 length));
+
+PNG_EXTERN void png_check_chunk_name PNGARG((png_structp png_ptr,
+   png_bytep chunk_name));
+
+/* handle the transformations for reading and writing */
+PNG_EXTERN void png_do_read_transformations PNGARG((png_structp png_ptr));
+PNG_EXTERN void png_do_write_transformations PNGARG((png_structp png_ptr));
+
+PNG_EXTERN void png_init_read_transformations PNGARG((png_structp png_ptr));
+
+#ifdef PNG_PROGRESSIVE_READ_SUPPORTED
+PNG_EXTERN void png_push_read_chunk PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+PNG_EXTERN void png_push_read_sig PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+PNG_EXTERN void png_push_check_crc PNGARG((png_structp png_ptr));
+PNG_EXTERN void png_push_crc_skip PNGARG((png_structp png_ptr,
+   png_uint_32 length));
+PNG_EXTERN void png_push_crc_finish PNGARG((png_structp png_ptr));
+PNG_EXTERN void png_push_save_buffer PNGARG((png_structp png_ptr));
+PNG_EXTERN void png_push_restore_buffer PNGARG((png_structp png_ptr,
+   png_bytep buffer, png_size_t buffer_length));
+PNG_EXTERN void png_push_read_IDAT PNGARG((png_structp png_ptr));
+PNG_EXTERN void png_process_IDAT_data PNGARG((png_structp png_ptr,
+   png_bytep buffer, png_size_t buffer_length));
+PNG_EXTERN void png_push_process_row PNGARG((png_structp png_ptr));
+PNG_EXTERN void png_push_handle_unknown PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 length));
+PNG_EXTERN void png_push_have_info PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+PNG_EXTERN void png_push_have_end PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+PNG_EXTERN void png_push_have_row PNGARG((png_structp png_ptr, png_bytep row));
+PNG_EXTERN void png_push_read_end PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+PNG_EXTERN void png_process_some_data PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+PNG_EXTERN void png_read_push_finish_row PNGARG((png_structp png_ptr));
+#if defined(PNG_READ_tEXt_SUPPORTED)
+PNG_EXTERN void png_push_handle_tEXt PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 length));
+PNG_EXTERN void png_push_read_tEXt PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+#endif
+#if defined(PNG_READ_zTXt_SUPPORTED)
+PNG_EXTERN void png_push_handle_zTXt PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 length));
+PNG_EXTERN void png_push_read_zTXt PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+#endif
+#if defined(PNG_READ_iTXt_SUPPORTED)
+PNG_EXTERN void png_push_handle_iTXt PNGARG((png_structp png_ptr,
+   png_infop info_ptr, png_uint_32 length));
+PNG_EXTERN void png_push_read_iTXt PNGARG((png_structp png_ptr,
+   png_infop info_ptr));
+#endif
+
+#endif /* PNG_PROGRESSIVE_READ_SUPPORTED */
+
+#ifdef PNG_MNG_FEATURES_SUPPORTED
+PNG_EXTERN void png_do_read_intrapixel PNGARG((png_row_infop row_info,
+   png_bytep row));
+PNG_EXTERN void png_do_write_intrapixel PNGARG((png_row_infop row_info,
+   png_bytep row));
+#endif
+
+#if defined(PNG_ASSEMBLER_CODE_SUPPORTED)
+#if defined(PNG_MMX_CODE_SUPPORTED)
+/* png.c */ /* PRIVATE */
+PNG_EXTERN void png_init_mmx_flags PNGARG((png_structp png_ptr));
+#endif
+#endif
+
+#if defined(PNG_INCH_CONVERSIONS) && defined(PNG_FLOATING_POINT_SUPPORTED)
+PNG_EXTERN png_uint_32 png_get_pixels_per_inch PNGARG((png_structp png_ptr,
+png_infop info_ptr));
+
+PNG_EXTERN png_uint_32 png_get_x_pixels_per_inch PNGARG((png_structp png_ptr,
+png_infop info_ptr));
+
+PNG_EXTERN png_uint_32 png_get_y_pixels_per_inch PNGARG((png_structp png_ptr,
+png_infop info_ptr));
+
+PNG_EXTERN float png_get_x_offset_inches PNGARG((png_structp png_ptr,
+png_infop info_ptr));
+
+PNG_EXTERN float png_get_y_offset_inches PNGARG((png_structp png_ptr,
+png_infop info_ptr));
+
+#if defined(PNG_pHYs_SUPPORTED)
+PNG_EXTERN png_uint_32 png_get_pHYs_dpi PNGARG((png_structp png_ptr,
+png_infop info_ptr, png_uint_32 *res_x, png_uint_32 *res_y, int *unit_type));
+#endif /* PNG_pHYs_SUPPORTED */
+#endif  /* PNG_INCH_CONVERSIONS && PNG_FLOATING_POINT_SUPPORTED */
+
+/* Maintainer: Put new private prototypes here ^ and in libpngpf.3 */
+
+#endif /* PNG_INTERNAL */
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* PNG_VERSION_INFO_ONLY */
+/* do not put anything past this line */
+#endif /* PNG_H */
diff --git a/distrib/libpng-1.2.19/pngconf.h b/distrib/libpng-1.2.19/pngconf.h
new file mode 100644
index 0000000..43dfe04
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngconf.h
@@ -0,0 +1,1535 @@
+
+/* pngconf.h - machine configurable file for libpng
+ *
+ * libpng version 1.2.19 - August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ */
+
+/* Any machine specific code is near the front of this file, so if you
+ * are configuring libpng for a machine, you may want to read the section
+ * starting here down to where it starts to typedef png_color, png_text,
+ * and png_info.
+ */
+
+#ifndef PNGCONF_H
+#define PNGCONF_H
+
+#define PNG_1_2_X
+
+/* 
+ * PNG_USER_CONFIG has to be defined on the compiler command line. This
+ * includes the resource compiler for Windows DLL configurations.
+ */
+#ifdef PNG_USER_CONFIG
+#  ifndef PNG_USER_PRIVATEBUILD
+#    define PNG_USER_PRIVATEBUILD
+#  endif
+#include "pngusr.h"
+#endif
+
+/* PNG_CONFIGURE_LIBPNG is set by the "configure" script. */
+#ifdef PNG_CONFIGURE_LIBPNG
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+#endif
+
+/*
+ * Added at libpng-1.2.8
+ *  
+ * If you create a private DLL you need to define in "pngusr.h" the followings:
+ * #define PNG_USER_PRIVATEBUILD <Describes by whom and why this version of
+ *        the DLL was built>
+ *  e.g. #define PNG_USER_PRIVATEBUILD "Build by MyCompany for xyz reasons."
+ * #define PNG_USER_DLLFNAME_POSTFIX <two-letter postfix that serve to
+ *        distinguish your DLL from those of the official release. These
+ *        correspond to the trailing letters that come after the version
+ *        number and must match your private DLL name>
+ *  e.g. // private DLL "libpng13gx.dll"
+ *       #define PNG_USER_DLLFNAME_POSTFIX "gx"
+ * 
+ * The following macros are also at your disposal if you want to complete the 
+ * DLL VERSIONINFO structure.
+ * - PNG_USER_VERSIONINFO_COMMENTS
+ * - PNG_USER_VERSIONINFO_COMPANYNAME
+ * - PNG_USER_VERSIONINFO_LEGALTRADEMARKS
+ */
+
+#ifdef __STDC__
+#ifdef SPECIALBUILD
+#  pragma message("PNG_LIBPNG_SPECIALBUILD (and deprecated SPECIALBUILD)\
+ are now LIBPNG reserved macros. Use PNG_USER_PRIVATEBUILD instead.")
+#endif
+
+#ifdef PRIVATEBUILD
+# pragma message("PRIVATEBUILD is deprecated.\
+ Use PNG_USER_PRIVATEBUILD instead.")
+# define PNG_USER_PRIVATEBUILD PRIVATEBUILD
+#endif
+#endif /* __STDC__ */
+
+#ifndef PNG_VERSION_INFO_ONLY
+
+/* End of material added to libpng-1.2.8 */
+
+/* Added at libpng-1.2.19 */
+#ifndef PNG_NO_WARN_UNINITIALIZED_ROW
+#  ifndef PNG_WARN_UNINITIALIZED_ROW
+#    define PNG_WARN_UNINITIALIZED_ROW 1 /* 0: warning; 1: error */
+#  endif
+#endif
+/* End of material added at libpng-1.2.19 */
+
+/* This is the size of the compression buffer, and thus the size of
+ * an IDAT chunk.  Make this whatever size you feel is best for your
+ * machine.  One of these will be allocated per png_struct.  When this
+ * is full, it writes the data to the disk, and does some other
+ * calculations.  Making this an extremely small size will slow
+ * the library down, but you may want to experiment to determine
+ * where it becomes significant, if you are concerned with memory
+ * usage.  Note that zlib allocates at least 32Kb also.  For readers,
+ * this describes the size of the buffer available to read the data in.
+ * Unless this gets smaller than the size of a row (compressed),
+ * it should not make much difference how big this is.
+ */
+
+#ifndef PNG_ZBUF_SIZE
+#  define PNG_ZBUF_SIZE 8192
+#endif
+
+/* Enable if you want a write-only libpng */
+
+#ifndef PNG_NO_READ_SUPPORTED
+#  define PNG_READ_SUPPORTED
+#endif
+
+/* Enable if you want a read-only libpng */
+
+#ifndef PNG_NO_WRITE_SUPPORTED
+#  define PNG_WRITE_SUPPORTED
+#endif
+
+/* Enabled by default in 1.2.0.  You can disable this if you don't need to
+   support PNGs that are embedded in MNG datastreams */
+#if !defined(PNG_1_0_X) && !defined(PNG_NO_MNG_FEATURES)
+#  ifndef PNG_MNG_FEATURES_SUPPORTED
+#    define PNG_MNG_FEATURES_SUPPORTED
+#  endif
+#endif
+
+#ifndef PNG_NO_FLOATING_POINT_SUPPORTED
+#  ifndef PNG_FLOATING_POINT_SUPPORTED
+#    define PNG_FLOATING_POINT_SUPPORTED
+#  endif
+#endif
+
+/* If you are running on a machine where you cannot allocate more
+ * than 64K of memory at once, uncomment this.  While libpng will not
+ * normally need that much memory in a chunk (unless you load up a very
+ * large file), zlib needs to know how big of a chunk it can use, and
+ * libpng thus makes sure to check any memory allocation to verify it
+ * will fit into memory.
+#define PNG_MAX_MALLOC_64K
+ */
+#if defined(MAXSEG_64K) && !defined(PNG_MAX_MALLOC_64K)
+#  define PNG_MAX_MALLOC_64K
+#endif
+
+/* Special munging to support doing things the 'cygwin' way:
+ * 'Normal' png-on-win32 defines/defaults:
+ *   PNG_BUILD_DLL -- building dll
+ *   PNG_USE_DLL   -- building an application, linking to dll
+ *   (no define)   -- building static library, or building an
+ *                    application and linking to the static lib
+ * 'Cygwin' defines/defaults:
+ *   PNG_BUILD_DLL -- (ignored) building the dll
+ *   (no define)   -- (ignored) building an application, linking to the dll
+ *   PNG_STATIC    -- (ignored) building the static lib, or building an 
+ *                    application that links to the static lib.
+ *   ALL_STATIC    -- (ignored) building various static libs, or building an 
+ *                    application that links to the static libs.
+ * Thus,
+ * a cygwin user should define either PNG_BUILD_DLL or PNG_STATIC, and
+ * this bit of #ifdefs will define the 'correct' config variables based on
+ * that. If a cygwin user *wants* to define 'PNG_USE_DLL' that's okay, but
+ * unnecessary.
+ *
+ * Also, the precedence order is:
+ *   ALL_STATIC (since we can't #undef something outside our namespace)
+ *   PNG_BUILD_DLL
+ *   PNG_STATIC
+ *   (nothing) == PNG_USE_DLL
+ * 
+ * CYGWIN (2002-01-20): The preceding is now obsolete. With the advent
+ *   of auto-import in binutils, we no longer need to worry about 
+ *   __declspec(dllexport) / __declspec(dllimport) and friends.  Therefore,
+ *   we don't need to worry about PNG_STATIC or ALL_STATIC when it comes
+ *   to __declspec() stuff.  However, we DO need to worry about 
+ *   PNG_BUILD_DLL and PNG_STATIC because those change some defaults
+ *   such as CONSOLE_IO and whether GLOBAL_ARRAYS are allowed.
+ */
+#if defined(__CYGWIN__)
+#  if defined(ALL_STATIC)
+#    if defined(PNG_BUILD_DLL)
+#      undef PNG_BUILD_DLL
+#    endif
+#    if defined(PNG_USE_DLL)
+#      undef PNG_USE_DLL
+#    endif
+#    if defined(PNG_DLL)
+#      undef PNG_DLL
+#    endif
+#    if !defined(PNG_STATIC)
+#      define PNG_STATIC
+#    endif
+#  else
+#    if defined (PNG_BUILD_DLL)
+#      if defined(PNG_STATIC)
+#        undef PNG_STATIC
+#      endif
+#      if defined(PNG_USE_DLL)
+#        undef PNG_USE_DLL
+#      endif
+#      if !defined(PNG_DLL)
+#        define PNG_DLL
+#      endif
+#    else
+#      if defined(PNG_STATIC)
+#        if defined(PNG_USE_DLL)
+#          undef PNG_USE_DLL
+#        endif
+#        if defined(PNG_DLL)
+#          undef PNG_DLL
+#        endif
+#      else
+#        if !defined(PNG_USE_DLL)
+#          define PNG_USE_DLL
+#        endif
+#        if !defined(PNG_DLL)
+#          define PNG_DLL
+#        endif
+#      endif  
+#    endif  
+#  endif
+#endif
+
+/* This protects us against compilers that run on a windowing system
+ * and thus don't have or would rather us not use the stdio types:
+ * stdin, stdout, and stderr.  The only one currently used is stderr
+ * in png_error() and png_warning().  #defining PNG_NO_CONSOLE_IO will
+ * prevent these from being compiled and used. #defining PNG_NO_STDIO
+ * will also prevent these, plus will prevent the entire set of stdio
+ * macros and functions (FILE *, printf, etc.) from being compiled and used,
+ * unless (PNG_DEBUG > 0) has been #defined.
+ *
+ * #define PNG_NO_CONSOLE_IO
+ * #define PNG_NO_STDIO
+ */
+
+#if defined(_WIN32_WCE)
+#  include <windows.h>
+   /* Console I/O functions are not supported on WindowsCE */
+#  define PNG_NO_CONSOLE_IO
+#  ifdef PNG_DEBUG
+#    undef PNG_DEBUG
+#  endif
+#endif
+
+#ifdef PNG_BUILD_DLL
+#  ifndef PNG_CONSOLE_IO_SUPPORTED
+#    ifndef PNG_NO_CONSOLE_IO
+#      define PNG_NO_CONSOLE_IO
+#    endif
+#  endif
+#endif
+
+#  ifdef PNG_NO_STDIO
+#    ifndef PNG_NO_CONSOLE_IO
+#      define PNG_NO_CONSOLE_IO
+#    endif
+#    ifdef PNG_DEBUG
+#      if (PNG_DEBUG > 0)
+#        include <stdio.h>
+#      endif
+#    endif
+#  else
+#    if !defined(_WIN32_WCE)
+/* "stdio.h" functions are not supported on WindowsCE */
+#      include <stdio.h>
+#    endif
+#  endif
+
+/* This macro protects us against machines that don't have function
+ * prototypes (ie K&R style headers).  If your compiler does not handle
+ * function prototypes, define this macro and use the included ansi2knr.
+ * I've always been able to use _NO_PROTO as the indicator, but you may
+ * need to drag the empty declaration out in front of here, or change the
+ * ifdef to suit your own needs.
+ */
+#ifndef PNGARG
+
+#ifdef OF /* zlib prototype munger */
+#  define PNGARG(arglist) OF(arglist)
+#else
+
+#ifdef _NO_PROTO
+#  define PNGARG(arglist) ()
+#  ifndef PNG_TYPECAST_NULL
+#     define PNG_TYPECAST_NULL
+#  endif
+#else
+#  define PNGARG(arglist) arglist
+#endif /* _NO_PROTO */
+
+
+#endif /* OF */
+
+#endif /* PNGARG */
+
+/* Try to determine if we are compiling on a Mac.  Note that testing for
+ * just __MWERKS__ is not good enough, because the Codewarrior is now used
+ * on non-Mac platforms.
+ */
+#ifndef MACOS
+#  if (defined(__MWERKS__) && defined(macintosh)) || defined(applec) || \
+      defined(THINK_C) || defined(__SC__) || defined(TARGET_OS_MAC)
+#    define MACOS
+#  endif
+#endif
+
+/* enough people need this for various reasons to include it here */
+#if !defined(MACOS) && !defined(RISCOS) && !defined(_WIN32_WCE)
+#  include <sys/types.h>
+#endif
+
+#if !defined(PNG_SETJMP_NOT_SUPPORTED) && !defined(PNG_NO_SETJMP_SUPPORTED)
+#  define PNG_SETJMP_SUPPORTED
+#endif
+
+#ifdef PNG_SETJMP_SUPPORTED
+/* This is an attempt to force a single setjmp behaviour on Linux.  If
+ * the X config stuff didn't define _BSD_SOURCE we wouldn't need this.
+ */
+
+#  ifdef __linux__
+#    ifdef _BSD_SOURCE
+#      define PNG_SAVE_BSD_SOURCE
+#      undef _BSD_SOURCE
+#    endif
+#    ifdef _SETJMP_H
+     /* If you encounter a compiler error here, see the explanation
+      * near the end of INSTALL.
+      */
+         __png.h__ already includes setjmp.h;
+         __dont__ include it again.;
+#    endif
+#  endif /* __linux__ */
+
+   /* include setjmp.h for error handling */
+#  include <setjmp.h>
+
+#  ifdef __linux__
+#    ifdef PNG_SAVE_BSD_SOURCE
+#      define _BSD_SOURCE
+#      undef PNG_SAVE_BSD_SOURCE
+#    endif
+#  endif /* __linux__ */
+#endif /* PNG_SETJMP_SUPPORTED */
+
+#ifdef BSD
+#  include <strings.h>
+#else
+#  include <string.h>
+#endif
+
+/* Other defines for things like memory and the like can go here.  */
+#ifdef PNG_INTERNAL
+
+#include <stdlib.h>
+
+/* The functions exported by PNG_EXTERN are PNG_INTERNAL functions, which
+ * aren't usually used outside the library (as far as I know), so it is
+ * debatable if they should be exported at all.  In the future, when it is
+ * possible to have run-time registry of chunk-handling functions, some of
+ * these will be made available again.
+#define PNG_EXTERN extern
+ */
+#define PNG_EXTERN
+
+/* Other defines specific to compilers can go here.  Try to keep
+ * them inside an appropriate ifdef/endif pair for portability.
+ */
+
+#if defined(PNG_FLOATING_POINT_SUPPORTED)
+#  if defined(MACOS)
+     /* We need to check that <math.h> hasn't already been included earlier
+      * as it seems it doesn't agree with <fp.h>, yet we should really use
+      * <fp.h> if possible.
+      */
+#    if !defined(__MATH_H__) && !defined(__MATH_H) && !defined(__cmath__)
+#      include <fp.h>
+#    endif
+#  else
+#    include <math.h>
+#  endif
+#  if defined(_AMIGA) && defined(__SASC) && defined(_M68881)
+     /* Amiga SAS/C: We must include builtin FPU functions when compiling using
+      * MATH=68881
+      */
+#    include <m68881.h>
+#  endif
+#endif
+
+/* Codewarrior on NT has linking problems without this. */
+#if (defined(__MWERKS__) && defined(WIN32)) || defined(__STDC__)
+#  define PNG_ALWAYS_EXTERN
+#endif
+
+/* This provides the non-ANSI (far) memory allocation routines. */
+#if defined(__TURBOC__) && defined(__MSDOS__)
+#  include <mem.h>
+#  include <alloc.h>
+#endif
+
+/* I have no idea why is this necessary... */
+#if defined(_MSC_VER) && (defined(WIN32) || defined(_Windows) || \
+    defined(_WINDOWS) || defined(_WIN32) || defined(__WIN32__))
+#  include <malloc.h>
+#endif
+
+/* This controls how fine the dithering gets.  As this allocates
+ * a largish chunk of memory (32K), those who are not as concerned
+ * with dithering quality can decrease some or all of these.
+ */
+#ifndef PNG_DITHER_RED_BITS
+#  define PNG_DITHER_RED_BITS 5
+#endif
+#ifndef PNG_DITHER_GREEN_BITS
+#  define PNG_DITHER_GREEN_BITS 5
+#endif
+#ifndef PNG_DITHER_BLUE_BITS
+#  define PNG_DITHER_BLUE_BITS 5
+#endif
+
+/* This controls how fine the gamma correction becomes when you
+ * are only interested in 8 bits anyway.  Increasing this value
+ * results in more memory being used, and more pow() functions
+ * being called to fill in the gamma tables.  Don't set this value
+ * less then 8, and even that may not work (I haven't tested it).
+ */
+
+#ifndef PNG_MAX_GAMMA_8
+#  define PNG_MAX_GAMMA_8 11
+#endif
+
+/* This controls how much a difference in gamma we can tolerate before
+ * we actually start doing gamma conversion.
+ */
+#ifndef PNG_GAMMA_THRESHOLD
+#  define PNG_GAMMA_THRESHOLD 0.05
+#endif
+
+#endif /* PNG_INTERNAL */
+
+/* The following uses const char * instead of char * for error
+ * and warning message functions, so some compilers won't complain.
+ * If you do not want to use const, define PNG_NO_CONST here.
+ */
+
+#ifndef PNG_NO_CONST
+#  define PNG_CONST const
+#else
+#  define PNG_CONST
+#endif
+
+/* The following defines give you the ability to remove code from the
+ * library that you will not be using.  I wish I could figure out how to
+ * automate this, but I can't do that without making it seriously hard
+ * on the users.  So if you are not using an ability, change the #define
+ * to and #undef, and that part of the library will not be compiled.  If
+ * your linker can't find a function, you may want to make sure the
+ * ability is defined here.  Some of these depend upon some others being
+ * defined.  I haven't figured out all the interactions here, so you may
+ * have to experiment awhile to get everything to compile.  If you are
+ * creating or using a shared library, you probably shouldn't touch this,
+ * as it will affect the size of the structures, and this will cause bad
+ * things to happen if the library and/or application ever change.
+ */
+
+/* Any features you will not be using can be undef'ed here */
+
+/* GR-P, 0.96a: Set "*TRANSFORMS_SUPPORTED as default but allow user
+ * to turn it off with "*TRANSFORMS_NOT_SUPPORTED" or *PNG_NO_*_TRANSFORMS
+ * on the compile line, then pick and choose which ones to define without
+ * having to edit this file. It is safe to use the *TRANSFORMS_NOT_SUPPORTED
+ * if you only want to have a png-compliant reader/writer but don't need
+ * any of the extra transformations.  This saves about 80 kbytes in a
+ * typical installation of the library. (PNG_NO_* form added in version
+ * 1.0.1c, for consistency)
+ */
+
+/* The size of the png_text structure changed in libpng-1.0.6 when
+ * iTXt support was added.  iTXt support was turned off by default through
+ * libpng-1.2.x, to support old apps that malloc the png_text structure
+ * instead of calling png_set_text() and letting libpng malloc it.  It
+ * was turned on by default in libpng-1.3.0.
+ */
+
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+#  ifndef PNG_NO_iTXt_SUPPORTED
+#    define PNG_NO_iTXt_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_iTXt
+#    define PNG_NO_READ_iTXt
+#  endif
+#  ifndef PNG_NO_WRITE_iTXt
+#    define PNG_NO_WRITE_iTXt
+#  endif
+#endif
+
+#if !defined(PNG_NO_iTXt_SUPPORTED)
+#  if !defined(PNG_READ_iTXt_SUPPORTED) && !defined(PNG_NO_READ_iTXt)
+#    define PNG_READ_iTXt
+#  endif
+#  if !defined(PNG_WRITE_iTXt_SUPPORTED) && !defined(PNG_NO_WRITE_iTXt)
+#    define PNG_WRITE_iTXt
+#  endif
+#endif
+
+/* The following support, added after version 1.0.0, can be turned off here en
+ * masse by defining PNG_LEGACY_SUPPORTED in case you need binary compatibility
+ * with old applications that require the length of png_struct and png_info
+ * to remain unchanged.
+ */
+
+#ifdef PNG_LEGACY_SUPPORTED
+#  define PNG_NO_FREE_ME
+#  define PNG_NO_READ_UNKNOWN_CHUNKS
+#  define PNG_NO_WRITE_UNKNOWN_CHUNKS
+#  define PNG_NO_READ_USER_CHUNKS
+#  define PNG_NO_READ_iCCP
+#  define PNG_NO_WRITE_iCCP
+#  define PNG_NO_READ_iTXt
+#  define PNG_NO_WRITE_iTXt
+#  define PNG_NO_READ_sCAL
+#  define PNG_NO_WRITE_sCAL
+#  define PNG_NO_READ_sPLT
+#  define PNG_NO_WRITE_sPLT
+#  define PNG_NO_INFO_IMAGE
+#  define PNG_NO_READ_RGB_TO_GRAY
+#  define PNG_NO_READ_USER_TRANSFORM
+#  define PNG_NO_WRITE_USER_TRANSFORM
+#  define PNG_NO_USER_MEM
+#  define PNG_NO_READ_EMPTY_PLTE
+#  define PNG_NO_MNG_FEATURES
+#  define PNG_NO_FIXED_POINT_SUPPORTED
+#endif
+
+/* Ignore attempt to turn off both floating and fixed point support */
+#if !defined(PNG_FLOATING_POINT_SUPPORTED) || \
+    !defined(PNG_NO_FIXED_POINT_SUPPORTED)
+#  define PNG_FIXED_POINT_SUPPORTED
+#endif
+
+#ifndef PNG_NO_FREE_ME
+#  define PNG_FREE_ME_SUPPORTED
+#endif
+
+#if defined(PNG_READ_SUPPORTED)
+
+#if !defined(PNG_READ_TRANSFORMS_NOT_SUPPORTED) && \
+      !defined(PNG_NO_READ_TRANSFORMS)
+#  define PNG_READ_TRANSFORMS_SUPPORTED
+#endif
+
+#ifdef PNG_READ_TRANSFORMS_SUPPORTED
+#  ifndef PNG_NO_READ_EXPAND
+#    define PNG_READ_EXPAND_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_SHIFT
+#    define PNG_READ_SHIFT_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_PACK
+#    define PNG_READ_PACK_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_BGR
+#    define PNG_READ_BGR_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_SWAP
+#    define PNG_READ_SWAP_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_PACKSWAP
+#    define PNG_READ_PACKSWAP_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_INVERT
+#    define PNG_READ_INVERT_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_DITHER
+#    define PNG_READ_DITHER_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_BACKGROUND
+#    define PNG_READ_BACKGROUND_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_16_TO_8
+#    define PNG_READ_16_TO_8_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_FILLER
+#    define PNG_READ_FILLER_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_GAMMA
+#    define PNG_READ_GAMMA_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_GRAY_TO_RGB
+#    define PNG_READ_GRAY_TO_RGB_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_SWAP_ALPHA
+#    define PNG_READ_SWAP_ALPHA_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_INVERT_ALPHA
+#    define PNG_READ_INVERT_ALPHA_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_STRIP_ALPHA
+#    define PNG_READ_STRIP_ALPHA_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_USER_TRANSFORM
+#    define PNG_READ_USER_TRANSFORM_SUPPORTED
+#  endif
+#  ifndef PNG_NO_READ_RGB_TO_GRAY
+#    define PNG_READ_RGB_TO_GRAY_SUPPORTED
+#  endif
+#endif /* PNG_READ_TRANSFORMS_SUPPORTED */
+
+#if !defined(PNG_NO_PROGRESSIVE_READ) && \
+ !defined(PNG_PROGRESSIVE_READ_NOT_SUPPORTED) /* if you don't do progressive */
+#  define PNG_PROGRESSIVE_READ_SUPPORTED    /* reading.  This is not talking */
+#endif                              /* about interlacing capability!  You'll */
+             /* still have interlacing unless you change the following line: */
+
+#define PNG_READ_INTERLACING_SUPPORTED /* required in PNG-compliant decoders */
+
+#ifndef PNG_NO_READ_COMPOSITE_NODIV
+#  ifndef PNG_NO_READ_COMPOSITED_NODIV  /* libpng-1.0.x misspelling */
+#    define PNG_READ_COMPOSITE_NODIV_SUPPORTED  /* well tested on Intel, SGI */
+#  endif
+#endif
+
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+/* Deprecated, will be removed from version 2.0.0.
+   Use PNG_MNG_FEATURES_SUPPORTED instead. */
+#ifndef PNG_NO_READ_EMPTY_PLTE
+#  define PNG_READ_EMPTY_PLTE_SUPPORTED
+#endif
+#endif
+
+#endif /* PNG_READ_SUPPORTED */
+
+#if defined(PNG_WRITE_SUPPORTED)
+
+# if !defined(PNG_WRITE_TRANSFORMS_NOT_SUPPORTED) && \
+    !defined(PNG_NO_WRITE_TRANSFORMS)
+#  define PNG_WRITE_TRANSFORMS_SUPPORTED
+#endif
+
+#ifdef PNG_WRITE_TRANSFORMS_SUPPORTED
+#  ifndef PNG_NO_WRITE_SHIFT
+#    define PNG_WRITE_SHIFT_SUPPORTED
+#  endif
+#  ifndef PNG_NO_WRITE_PACK
+#    define PNG_WRITE_PACK_SUPPORTED
+#  endif
+#  ifndef PNG_NO_WRITE_BGR
+#    define PNG_WRITE_BGR_SUPPORTED
+#  endif
+#  ifndef PNG_NO_WRITE_SWAP
+#    define PNG_WRITE_SWAP_SUPPORTED
+#  endif
+#  ifndef PNG_NO_WRITE_PACKSWAP
+#    define PNG_WRITE_PACKSWAP_SUPPORTED
+#  endif
+#  ifndef PNG_NO_WRITE_INVERT
+#    define PNG_WRITE_INVERT_SUPPORTED
+#  endif
+#  ifndef PNG_NO_WRITE_FILLER
+#    define PNG_WRITE_FILLER_SUPPORTED   /* same as WRITE_STRIP_ALPHA */
+#  endif
+#  ifndef PNG_NO_WRITE_SWAP_ALPHA
+#    define PNG_WRITE_SWAP_ALPHA_SUPPORTED
+#  endif
+#  ifndef PNG_NO_WRITE_INVERT_ALPHA
+#    define PNG_WRITE_INVERT_ALPHA_SUPPORTED
+#  endif
+#  ifndef PNG_NO_WRITE_USER_TRANSFORM
+#    define PNG_WRITE_USER_TRANSFORM_SUPPORTED
+#  endif
+#endif /* PNG_WRITE_TRANSFORMS_SUPPORTED */
+
+#if !defined(PNG_NO_WRITE_INTERLACING_SUPPORTED) && \
+    !defined(PNG_WRITE_INTERLACING_SUPPORTED)
+#define PNG_WRITE_INTERLACING_SUPPORTED  /* not required for PNG-compliant
+                                            encoders, but can cause trouble
+                                            if left undefined */
+#endif
+
+#if !defined(PNG_NO_WRITE_WEIGHTED_FILTER) && \
+    !defined(PNG_WRITE_WEIGHTED_FILTER) && \
+     defined(PNG_FLOATING_POINT_SUPPORTED)
+#  define PNG_WRITE_WEIGHTED_FILTER_SUPPORTED
+#endif
+
+#ifndef PNG_NO_WRITE_FLUSH
+#  define PNG_WRITE_FLUSH_SUPPORTED
+#endif
+
+#if defined(PNG_1_0_X) || defined (PNG_1_2_X)
+/* Deprecated, see PNG_MNG_FEATURES_SUPPORTED, above */
+#ifndef PNG_NO_WRITE_EMPTY_PLTE
+#  define PNG_WRITE_EMPTY_PLTE_SUPPORTED
+#endif
+#endif
+
+#endif /* PNG_WRITE_SUPPORTED */
+
+#ifndef PNG_1_0_X
+#  ifndef PNG_NO_ERROR_NUMBERS
+#    define PNG_ERROR_NUMBERS_SUPPORTED
+#  endif
+#endif /* PNG_1_0_X */
+
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED)
+#  ifndef PNG_NO_USER_TRANSFORM_PTR
+#    define PNG_USER_TRANSFORM_PTR_SUPPORTED
+#  endif
+#endif
+
+#ifndef PNG_NO_STDIO
+#  define PNG_TIME_RFC1123_SUPPORTED
+#endif
+
+/* This adds extra functions in pngget.c for accessing data from the
+ * info pointer (added in version 0.99)
+ * png_get_image_width()
+ * png_get_image_height()
+ * png_get_bit_depth()
+ * png_get_color_type()
+ * png_get_compression_type()
+ * png_get_filter_type()
+ * png_get_interlace_type()
+ * png_get_pixel_aspect_ratio()
+ * png_get_pixels_per_meter()
+ * png_get_x_offset_pixels()
+ * png_get_y_offset_pixels()
+ * png_get_x_offset_microns()
+ * png_get_y_offset_microns()
+ */
+#if !defined(PNG_NO_EASY_ACCESS) && !defined(PNG_EASY_ACCESS_SUPPORTED)
+#  define PNG_EASY_ACCESS_SUPPORTED
+#endif
+
+/* PNG_ASSEMBLER_CODE was enabled by default in version 1.2.0 
+ * even when PNG_USE_PNGVCRD or PNG_USE_PNGGCCRD is not defined.
+ *
+ * PNG_NO_ASSEMBLER_CODE disables use of all assembler code,
+ * and removes several functions from the API.
+ *
+ * PNG_NO_MMX_CODE disables the use of MMX code without changing the API.
+ * When MMX code is off, then optimized C replacement functions are used,
+ * if PNG_NO_OPTIMIZED_CODE is not enabled.  This was added in version
+ * 1.2.19.
+*/
+
+#if defined(PNG_READ_SUPPORTED) && !defined(PNG_NO_OPTIMIZED_CODE)
+#  ifndef PNG_OPTIMIZED_CODE_SUPPORTED
+#    define PNG_OPTIMIZED_CODE_SUPPORTED
+#  endif
+#endif
+
+#if defined(PNG_READ_SUPPORTED) && !defined(PNG_NO_ASSEMBLER_CODE)
+#  ifndef PNG_ASSEMBLER_CODE_SUPPORTED
+#    define PNG_ASSEMBLER_CODE_SUPPORTED
+#  endif
+
+#  if !defined(PNG_MMX_CODE_SUPPORTED) && !defined(PNG_NO_MMX_CODE)
+#    define PNG_MMX_CODE_SUPPORTED
+#  endif
+
+#  if !defined(PNG_USE_PNGVCRD) && defined(PNG_MMX_CODE_SUPPORTED) && \
+     defined(_MSC_VER)
+#    define PNG_USE_PNGVCRD
+#  endif
+
+#  if !defined(PNG_USE_PNGGCCRD) && defined(PNG_MMX_CODE_SUPPORTED) && \
+     !defined(PNG_USE_PNGVCRD)
+#    define PNG_USE_PNGGCCRD
+     /* If you are sure that you don't need thread safety and you are compiling
+        with PNG_USE_PNGCCRD for an MMX application, you can define this for
+        faster execution.  See pnggccrd.c.
+#    define PNG_THREAD_UNSAFE_OK
+     */
+#  endif
+
+#endif
+
+#if !defined(PNG_1_0_X)
+#if !defined(PNG_NO_USER_MEM) && !defined(PNG_USER_MEM_SUPPORTED)
+#  define PNG_USER_MEM_SUPPORTED
+#endif
+#endif /* PNG_1_0_X */
+
+/* Added at libpng-1.2.6 */
+#if !defined(PNG_1_0_X)
+#ifndef PNG_SET_USER_LIMITS_SUPPORTED
+#if !defined(PNG_NO_SET_USER_LIMITS) && !defined(PNG_SET_USER_LIMITS_SUPPORTED)
+#  define PNG_SET_USER_LIMITS_SUPPORTED
+#endif
+#endif
+#endif /* PNG_1_0_X */
+
+/* Added at libpng-1.0.16 and 1.2.6.  To accept all valid PNGS no matter
+ * how large, set these limits to 0x7fffffffL
+ */
+#ifndef PNG_USER_WIDTH_MAX
+#  define PNG_USER_WIDTH_MAX 1000000L
+#endif
+#ifndef PNG_USER_HEIGHT_MAX
+#  define PNG_USER_HEIGHT_MAX 1000000L
+#endif
+
+/* These are currently experimental features, define them if you want */
+
+/* very little testing */
+/*
+#ifdef PNG_READ_SUPPORTED
+#  ifndef PNG_READ_16_TO_8_ACCURATE_SCALE_SUPPORTED
+#    define PNG_READ_16_TO_8_ACCURATE_SCALE_SUPPORTED
+#  endif
+#endif
+*/
+
+/* This is only for PowerPC big-endian and 680x0 systems */
+/* some testing */
+/*
+#ifndef PNG_READ_BIG_ENDIAN_SUPPORTED
+#  define PNG_READ_BIG_ENDIAN_SUPPORTED
+#endif
+*/
+
+/* Buggy compilers (e.g., gcc 2.7.2.2) need this */
+/*
+#define PNG_NO_POINTER_INDEXING
+*/
+
+/* These functions are turned off by default, as they will be phased out. */
+/*
+#define  PNG_USELESS_TESTS_SUPPORTED
+#define  PNG_CORRECT_PALETTE_SUPPORTED
+*/
+
+/* Any chunks you are not interested in, you can undef here.  The
+ * ones that allocate memory may be expecially important (hIST,
+ * tEXt, zTXt, tRNS, pCAL).  Others will just save time and make png_info
+ * a bit smaller.
+ */
+
+#if defined(PNG_READ_SUPPORTED) && \
+    !defined(PNG_READ_ANCILLARY_CHUNKS_NOT_SUPPORTED) && \
+    !defined(PNG_NO_READ_ANCILLARY_CHUNKS)
+#  define PNG_READ_ANCILLARY_CHUNKS_SUPPORTED
+#endif
+
+#if defined(PNG_WRITE_SUPPORTED) && \
+    !defined(PNG_WRITE_ANCILLARY_CHUNKS_NOT_SUPPORTED) && \
+    !defined(PNG_NO_WRITE_ANCILLARY_CHUNKS)
+#  define PNG_WRITE_ANCILLARY_CHUNKS_SUPPORTED
+#endif
+
+#ifdef PNG_READ_ANCILLARY_CHUNKS_SUPPORTED
+
+#ifdef PNG_NO_READ_TEXT
+#  define PNG_NO_READ_iTXt
+#  define PNG_NO_READ_tEXt
+#  define PNG_NO_READ_zTXt
+#endif
+#ifndef PNG_NO_READ_bKGD
+#  define PNG_READ_bKGD_SUPPORTED
+#  define PNG_bKGD_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_cHRM
+#  define PNG_READ_cHRM_SUPPORTED
+#  define PNG_cHRM_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_gAMA
+#  define PNG_READ_gAMA_SUPPORTED
+#  define PNG_gAMA_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_hIST
+#  define PNG_READ_hIST_SUPPORTED
+#  define PNG_hIST_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_iCCP
+#  define PNG_READ_iCCP_SUPPORTED
+#  define PNG_iCCP_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_iTXt
+#  ifndef PNG_READ_iTXt_SUPPORTED
+#    define PNG_READ_iTXt_SUPPORTED
+#  endif
+#  ifndef PNG_iTXt_SUPPORTED
+#    define PNG_iTXt_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_READ_oFFs
+#  define PNG_READ_oFFs_SUPPORTED
+#  define PNG_oFFs_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_pCAL
+#  define PNG_READ_pCAL_SUPPORTED
+#  define PNG_pCAL_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_sCAL
+#  define PNG_READ_sCAL_SUPPORTED
+#  define PNG_sCAL_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_pHYs
+#  define PNG_READ_pHYs_SUPPORTED
+#  define PNG_pHYs_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_sBIT
+#  define PNG_READ_sBIT_SUPPORTED
+#  define PNG_sBIT_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_sPLT
+#  define PNG_READ_sPLT_SUPPORTED
+#  define PNG_sPLT_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_sRGB
+#  define PNG_READ_sRGB_SUPPORTED
+#  define PNG_sRGB_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_tEXt
+#  define PNG_READ_tEXt_SUPPORTED
+#  define PNG_tEXt_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_tIME
+#  define PNG_READ_tIME_SUPPORTED
+#  define PNG_tIME_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_tRNS
+#  define PNG_READ_tRNS_SUPPORTED
+#  define PNG_tRNS_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_zTXt
+#  define PNG_READ_zTXt_SUPPORTED
+#  define PNG_zTXt_SUPPORTED
+#endif
+#ifndef PNG_NO_READ_UNKNOWN_CHUNKS
+#  define PNG_READ_UNKNOWN_CHUNKS_SUPPORTED
+#  ifndef PNG_UNKNOWN_CHUNKS_SUPPORTED
+#    define PNG_UNKNOWN_CHUNKS_SUPPORTED
+#  endif
+#  ifndef PNG_NO_HANDLE_AS_UNKNOWN
+#    define PNG_HANDLE_AS_UNKNOWN_SUPPORTED
+#  endif
+#endif
+#if !defined(PNG_NO_READ_USER_CHUNKS) && \
+     defined(PNG_READ_UNKNOWN_CHUNKS_SUPPORTED)
+#  define PNG_READ_USER_CHUNKS_SUPPORTED
+#  define PNG_USER_CHUNKS_SUPPORTED
+#  ifdef PNG_NO_READ_UNKNOWN_CHUNKS
+#    undef PNG_NO_READ_UNKNOWN_CHUNKS
+#  endif
+#  ifdef PNG_NO_HANDLE_AS_UNKNOWN
+#    undef PNG_NO_HANDLE_AS_UNKNOWN
+#  endif
+#endif
+#ifndef PNG_NO_READ_OPT_PLTE
+#  define PNG_READ_OPT_PLTE_SUPPORTED /* only affects support of the */
+#endif                      /* optional PLTE chunk in RGB and RGBA images */
+#if defined(PNG_READ_iTXt_SUPPORTED) || defined(PNG_READ_tEXt_SUPPORTED) || \
+    defined(PNG_READ_zTXt_SUPPORTED)
+#  define PNG_READ_TEXT_SUPPORTED
+#  define PNG_TEXT_SUPPORTED
+#endif
+
+#endif /* PNG_READ_ANCILLARY_CHUNKS_SUPPORTED */
+
+#ifdef PNG_WRITE_ANCILLARY_CHUNKS_SUPPORTED
+
+#ifdef PNG_NO_WRITE_TEXT
+#  define PNG_NO_WRITE_iTXt
+#  define PNG_NO_WRITE_tEXt
+#  define PNG_NO_WRITE_zTXt
+#endif
+#ifndef PNG_NO_WRITE_bKGD
+#  define PNG_WRITE_bKGD_SUPPORTED
+#  ifndef PNG_bKGD_SUPPORTED
+#    define PNG_bKGD_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_cHRM
+#  define PNG_WRITE_cHRM_SUPPORTED
+#  ifndef PNG_cHRM_SUPPORTED
+#    define PNG_cHRM_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_gAMA
+#  define PNG_WRITE_gAMA_SUPPORTED
+#  ifndef PNG_gAMA_SUPPORTED
+#    define PNG_gAMA_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_hIST
+#  define PNG_WRITE_hIST_SUPPORTED
+#  ifndef PNG_hIST_SUPPORTED
+#    define PNG_hIST_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_iCCP
+#  define PNG_WRITE_iCCP_SUPPORTED
+#  ifndef PNG_iCCP_SUPPORTED
+#    define PNG_iCCP_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_iTXt
+#  ifndef PNG_WRITE_iTXt_SUPPORTED
+#    define PNG_WRITE_iTXt_SUPPORTED
+#  endif
+#  ifndef PNG_iTXt_SUPPORTED
+#    define PNG_iTXt_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_oFFs
+#  define PNG_WRITE_oFFs_SUPPORTED
+#  ifndef PNG_oFFs_SUPPORTED
+#    define PNG_oFFs_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_pCAL
+#  define PNG_WRITE_pCAL_SUPPORTED
+#  ifndef PNG_pCAL_SUPPORTED
+#    define PNG_pCAL_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_sCAL
+#  define PNG_WRITE_sCAL_SUPPORTED
+#  ifndef PNG_sCAL_SUPPORTED
+#    define PNG_sCAL_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_pHYs
+#  define PNG_WRITE_pHYs_SUPPORTED
+#  ifndef PNG_pHYs_SUPPORTED
+#    define PNG_pHYs_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_sBIT
+#  define PNG_WRITE_sBIT_SUPPORTED
+#  ifndef PNG_sBIT_SUPPORTED
+#    define PNG_sBIT_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_sPLT
+#  define PNG_WRITE_sPLT_SUPPORTED
+#  ifndef PNG_sPLT_SUPPORTED
+#    define PNG_sPLT_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_sRGB
+#  define PNG_WRITE_sRGB_SUPPORTED
+#  ifndef PNG_sRGB_SUPPORTED
+#    define PNG_sRGB_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_tEXt
+#  define PNG_WRITE_tEXt_SUPPORTED
+#  ifndef PNG_tEXt_SUPPORTED
+#    define PNG_tEXt_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_tIME
+#  define PNG_WRITE_tIME_SUPPORTED
+#  ifndef PNG_tIME_SUPPORTED
+#    define PNG_tIME_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_tRNS
+#  define PNG_WRITE_tRNS_SUPPORTED
+#  ifndef PNG_tRNS_SUPPORTED
+#    define PNG_tRNS_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_zTXt
+#  define PNG_WRITE_zTXt_SUPPORTED
+#  ifndef PNG_zTXt_SUPPORTED
+#    define PNG_zTXt_SUPPORTED
+#  endif
+#endif
+#ifndef PNG_NO_WRITE_UNKNOWN_CHUNKS
+#  define PNG_WRITE_UNKNOWN_CHUNKS_SUPPORTED
+#  ifndef PNG_UNKNOWN_CHUNKS_SUPPORTED
+#    define PNG_UNKNOWN_CHUNKS_SUPPORTED
+#  endif
+#  ifndef PNG_NO_HANDLE_AS_UNKNOWN
+#     ifndef PNG_HANDLE_AS_UNKNOWN_SUPPORTED
+#       define PNG_HANDLE_AS_UNKNOWN_SUPPORTED
+#     endif
+#  endif
+#endif
+#if defined(PNG_WRITE_iTXt_SUPPORTED) || defined(PNG_WRITE_tEXt_SUPPORTED) || \
+    defined(PNG_WRITE_zTXt_SUPPORTED)
+#  define PNG_WRITE_TEXT_SUPPORTED
+#  ifndef PNG_TEXT_SUPPORTED
+#    define PNG_TEXT_SUPPORTED
+#  endif
+#endif
+
+#endif /* PNG_WRITE_ANCILLARY_CHUNKS_SUPPORTED */
+
+/* Turn this off to disable png_read_png() and
+ * png_write_png() and leave the row_pointers member
+ * out of the info structure.
+ */
+#ifndef PNG_NO_INFO_IMAGE
+#  define PNG_INFO_IMAGE_SUPPORTED
+#endif
+
+/* need the time information for reading tIME chunks */
+#if defined(PNG_tIME_SUPPORTED)
+#  if !defined(_WIN32_WCE)
+     /* "time.h" functions are not supported on WindowsCE */
+#    include <time.h>
+#  endif
+#endif
+
+/* Some typedefs to get us started.  These should be safe on most of the
+ * common platforms.  The typedefs should be at least as large as the
+ * numbers suggest (a png_uint_32 must be at least 32 bits long), but they
+ * don't have to be exactly that size.  Some compilers dislike passing
+ * unsigned shorts as function parameters, so you may be better off using
+ * unsigned int for png_uint_16.  Likewise, for 64-bit systems, you may
+ * want to have unsigned int for png_uint_32 instead of unsigned long.
+ */
+
+typedef unsigned long png_uint_32;
+typedef long png_int_32;
+typedef unsigned short png_uint_16;
+typedef short png_int_16;
+typedef unsigned char png_byte;
+
+/* This is usually size_t.  It is typedef'ed just in case you need it to
+   change (I'm not sure if you will or not, so I thought I'd be safe) */
+#ifdef PNG_SIZE_T
+   typedef PNG_SIZE_T png_size_t;
+#  define png_sizeof(x) png_convert_size(sizeof (x))
+#else
+   typedef size_t png_size_t;
+#  define png_sizeof(x) sizeof (x)
+#endif
+
+/* The following is needed for medium model support.  It cannot be in the
+ * PNG_INTERNAL section.  Needs modification for other compilers besides
+ * MSC.  Model independent support declares all arrays and pointers to be
+ * large using the far keyword.  The zlib version used must also support
+ * model independent data.  As of version zlib 1.0.4, the necessary changes
+ * have been made in zlib.  The USE_FAR_KEYWORD define triggers other
+ * changes that are needed. (Tim Wegner)
+ */
+
+/* Separate compiler dependencies (problem here is that zlib.h always
+   defines FAR. (SJT) */
+#ifdef __BORLANDC__
+#  if defined(__LARGE__) || defined(__HUGE__) || defined(__COMPACT__)
+#    define LDATA 1
+#  else
+#    define LDATA 0
+#  endif
+   /* GRR:  why is Cygwin in here?  Cygwin is not Borland C... */
+#  if !defined(__WIN32__) && !defined(__FLAT__) && !defined(__CYGWIN__)
+#    define PNG_MAX_MALLOC_64K
+#    if (LDATA != 1)
+#      ifndef FAR
+#        define FAR __far
+#      endif
+#      define USE_FAR_KEYWORD
+#    endif   /* LDATA != 1 */
+     /* Possibly useful for moving data out of default segment.
+      * Uncomment it if you want. Could also define FARDATA as
+      * const if your compiler supports it. (SJT)
+#    define FARDATA FAR
+      */
+#  endif  /* __WIN32__, __FLAT__, __CYGWIN__ */
+#endif   /* __BORLANDC__ */
+
+
+/* Suggest testing for specific compiler first before testing for
+ * FAR.  The Watcom compiler defines both __MEDIUM__ and M_I86MM,
+ * making reliance oncertain keywords suspect. (SJT)
+ */
+
+/* MSC Medium model */
+#if defined(FAR)
+#  if defined(M_I86MM)
+#    define USE_FAR_KEYWORD
+#    define FARDATA FAR
+#    include <dos.h>
+#  endif
+#endif
+
+/* SJT: default case */
+#ifndef FAR
+#  define FAR
+#endif
+
+/* At this point FAR is always defined */
+#ifndef FARDATA
+#  define FARDATA
+#endif
+
+/* Typedef for floating-point numbers that are converted
+   to fixed-point with a multiple of 100,000, e.g., int_gamma */
+typedef png_int_32 png_fixed_point;
+
+/* Add typedefs for pointers */
+typedef void            FAR * png_voidp;
+typedef png_byte        FAR * png_bytep;
+typedef png_uint_32     FAR * png_uint_32p;
+typedef png_int_32      FAR * png_int_32p;
+typedef png_uint_16     FAR * png_uint_16p;
+typedef png_int_16      FAR * png_int_16p;
+typedef PNG_CONST char  FAR * png_const_charp;
+typedef char            FAR * png_charp;
+typedef png_fixed_point FAR * png_fixed_point_p;
+
+#ifndef PNG_NO_STDIO
+#if defined(_WIN32_WCE)
+typedef HANDLE                png_FILE_p;
+#else
+typedef FILE                * png_FILE_p;
+#endif
+#endif
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+typedef double          FAR * png_doublep;
+#endif
+
+/* Pointers to pointers; i.e. arrays */
+typedef png_byte        FAR * FAR * png_bytepp;
+typedef png_uint_32     FAR * FAR * png_uint_32pp;
+typedef png_int_32      FAR * FAR * png_int_32pp;
+typedef png_uint_16     FAR * FAR * png_uint_16pp;
+typedef png_int_16      FAR * FAR * png_int_16pp;
+typedef PNG_CONST char  FAR * FAR * png_const_charpp;
+typedef char            FAR * FAR * png_charpp;
+typedef png_fixed_point FAR * FAR * png_fixed_point_pp;
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+typedef double          FAR * FAR * png_doublepp;
+#endif
+
+/* Pointers to pointers to pointers; i.e., pointer to array */
+typedef char            FAR * FAR * FAR * png_charppp;
+
+#if defined(PNG_1_0_X) || defined(PNG_1_2_X)
+/* SPC -  Is this stuff deprecated? */
+/* It'll be removed as of libpng-1.3.0 - GR-P */
+/* libpng typedefs for types in zlib. If zlib changes
+ * or another compression library is used, then change these.
+ * Eliminates need to change all the source files.
+ */
+typedef charf *         png_zcharp;
+typedef charf * FAR *   png_zcharpp;
+typedef z_stream FAR *  png_zstreamp;
+#endif /* (PNG_1_0_X) || defined(PNG_1_2_X) */
+
+/*
+ * Define PNG_BUILD_DLL if the module being built is a Windows
+ * LIBPNG DLL.
+ *
+ * Define PNG_USE_DLL if you want to *link* to the Windows LIBPNG DLL.
+ * It is equivalent to Microsoft predefined macro _DLL that is
+ * automatically defined when you compile using the share
+ * version of the CRT (C Run-Time library)
+ *
+ * The cygwin mods make this behavior a little different:
+ * Define PNG_BUILD_DLL if you are building a dll for use with cygwin
+ * Define PNG_STATIC if you are building a static library for use with cygwin,
+ *   -or- if you are building an application that you want to link to the
+ *   static library.
+ * PNG_USE_DLL is defined by default (no user action needed) unless one of
+ *   the other flags is defined.
+ */
+
+#if !defined(PNG_DLL) && (defined(PNG_BUILD_DLL) || defined(PNG_USE_DLL))
+#  define PNG_DLL
+#endif
+/* If CYGWIN, then disallow GLOBAL ARRAYS unless building a static lib.
+ * When building a static lib, default to no GLOBAL ARRAYS, but allow
+ * command-line override
+ */
+#if defined(__CYGWIN__)
+#  if !defined(PNG_STATIC)
+#    if defined(PNG_USE_GLOBAL_ARRAYS)
+#      undef PNG_USE_GLOBAL_ARRAYS
+#    endif
+#    if !defined(PNG_USE_LOCAL_ARRAYS)
+#      define PNG_USE_LOCAL_ARRAYS
+#    endif
+#  else
+#    if defined(PNG_USE_LOCAL_ARRAYS) || defined(PNG_NO_GLOBAL_ARRAYS)
+#      if defined(PNG_USE_GLOBAL_ARRAYS)
+#        undef PNG_USE_GLOBAL_ARRAYS
+#      endif
+#    endif
+#  endif
+#  if !defined(PNG_USE_LOCAL_ARRAYS) && !defined(PNG_USE_GLOBAL_ARRAYS)
+#    define PNG_USE_LOCAL_ARRAYS
+#  endif
+#endif
+
+/* Do not use global arrays (helps with building DLL's)
+ * They are no longer used in libpng itself, since version 1.0.5c,
+ * but might be required for some pre-1.0.5c applications.
+ */
+#if !defined(PNG_USE_LOCAL_ARRAYS) && !defined(PNG_USE_GLOBAL_ARRAYS)
+#  if defined(PNG_NO_GLOBAL_ARRAYS) || \
+      (defined(__GNUC__) && defined(PNG_DLL)) || defined(_MSC_VER)
+#    define PNG_USE_LOCAL_ARRAYS
+#  else
+#    define PNG_USE_GLOBAL_ARRAYS
+#  endif
+#endif
+
+#if defined(__CYGWIN__)
+#  undef PNGAPI
+#  define PNGAPI __cdecl
+#  undef PNG_IMPEXP
+#  define PNG_IMPEXP
+#endif  
+
+/* If you define PNGAPI, e.g., with compiler option "-DPNGAPI=__stdcall",
+ * you may get warnings regarding the linkage of png_zalloc and png_zfree.
+ * Don't ignore those warnings; you must also reset the default calling
+ * convention in your compiler to match your PNGAPI, and you must build
+ * zlib and your applications the same way you build libpng.
+ */
+
+#if defined(__MINGW32__) && !defined(PNG_MODULEDEF)
+#  ifndef PNG_NO_MODULEDEF
+#    define PNG_NO_MODULEDEF
+#  endif
+#endif
+
+#if !defined(PNG_IMPEXP) && defined(PNG_BUILD_DLL) && !defined(PNG_NO_MODULEDEF)
+#  define PNG_IMPEXP
+#endif
+
+#if defined(PNG_DLL) || defined(_DLL) || defined(__DLL__ ) || \
+    (( defined(_Windows) || defined(_WINDOWS) || \
+       defined(WIN32) || defined(_WIN32) || defined(__WIN32__) ))
+
+#  ifndef PNGAPI
+#     if defined(__GNUC__) || (defined (_MSC_VER) && (_MSC_VER >= 800))
+#        define PNGAPI __cdecl
+#     else
+#        define PNGAPI _cdecl
+#     endif
+#  endif
+
+#  if !defined(PNG_IMPEXP) && (!defined(PNG_DLL) || \
+       0 /* WINCOMPILER_WITH_NO_SUPPORT_FOR_DECLIMPEXP */)
+#     define PNG_IMPEXP
+#  endif
+
+#  if !defined(PNG_IMPEXP)
+
+#     define PNG_EXPORT_TYPE1(type,symbol)  PNG_IMPEXP type PNGAPI symbol
+#     define PNG_EXPORT_TYPE2(type,symbol)  type PNG_IMPEXP PNGAPI symbol
+
+      /* Borland/Microsoft */
+#     if defined(_MSC_VER) || defined(__BORLANDC__)
+#        if (_MSC_VER >= 800) || (__BORLANDC__ >= 0x500)
+#           define PNG_EXPORT PNG_EXPORT_TYPE1
+#        else
+#           define PNG_EXPORT PNG_EXPORT_TYPE2
+#           if defined(PNG_BUILD_DLL)
+#              define PNG_IMPEXP __export
+#           else
+#              define PNG_IMPEXP /*__import */ /* doesn't exist AFAIK in
+                                                 VC++ */
+#           endif                             /* Exists in Borland C++ for
+                                                 C++ classes (== huge) */
+#        endif
+#     endif
+
+#     if !defined(PNG_IMPEXP)
+#        if defined(PNG_BUILD_DLL)
+#           define PNG_IMPEXP __declspec(dllexport)
+#        else
+#           define PNG_IMPEXP __declspec(dllimport)
+#        endif
+#     endif
+#  endif  /* PNG_IMPEXP */
+#else /* !(DLL || non-cygwin WINDOWS) */
+#   if (defined(__IBMC__) || defined(__IBMCPP__)) && defined(__OS2__)
+#      ifndef PNGAPI
+#         define PNGAPI _System
+#      endif
+#   else
+#      if 0 /* ... other platforms, with other meanings */
+#      endif
+#   endif
+#endif
+
+#ifndef PNGAPI
+#  define PNGAPI
+#endif
+#ifndef PNG_IMPEXP
+#  define PNG_IMPEXP
+#endif
+
+#ifdef PNG_BUILDSYMS
+#  ifndef PNG_EXPORT
+#    define PNG_EXPORT(type,symbol) PNG_FUNCTION_EXPORT symbol END
+#  endif
+#  ifdef PNG_USE_GLOBAL_ARRAYS
+#    ifndef PNG_EXPORT_VAR
+#      define PNG_EXPORT_VAR(type) PNG_DATA_EXPORT
+#    endif
+#  endif
+#endif
+
+#ifndef PNG_EXPORT
+#  define PNG_EXPORT(type,symbol) PNG_IMPEXP type PNGAPI symbol
+#endif
+
+#ifdef PNG_USE_GLOBAL_ARRAYS
+#  ifndef PNG_EXPORT_VAR
+#    define PNG_EXPORT_VAR(type) extern PNG_IMPEXP type
+#  endif
+#endif
+
+/* User may want to use these so they are not in PNG_INTERNAL. Any library
+ * functions that are passed far data must be model independent.
+ */
+
+#ifndef PNG_ABORT
+#  define PNG_ABORT() abort()
+#endif
+
+#ifdef PNG_SETJMP_SUPPORTED
+#  define png_jmpbuf(png_ptr) ((png_ptr)->jmpbuf)
+#else
+#  define png_jmpbuf(png_ptr) \
+   (LIBPNG_WAS_COMPILED_WITH__PNG_SETJMP_NOT_SUPPORTED)
+#endif
+
+#if defined(USE_FAR_KEYWORD)  /* memory model independent fns */
+/* use this to make far-to-near assignments */
+#  define CHECK   1
+#  define NOCHECK 0
+#  define CVT_PTR(ptr) (png_far_to_near(png_ptr,ptr,CHECK))
+#  define CVT_PTR_NOCHECK(ptr) (png_far_to_near(png_ptr,ptr,NOCHECK))
+#  define png_snprintf _fsnprintf   /* Added to v 1.2.19 */
+#  define png_strcpy  _fstrcpy
+#  define png_strncpy _fstrncpy   /* Added to v 1.2.6 */
+#  define png_strlen  _fstrlen
+#  define png_memcmp  _fmemcmp    /* SJT: added */
+#  define png_memcpy  _fmemcpy
+#  define png_memset  _fmemset
+#else /* use the usual functions */
+#  define CVT_PTR(ptr)         (ptr)
+#  define CVT_PTR_NOCHECK(ptr) (ptr)
+#  ifndef PNG_NO_SNPRINTF
+#    ifdef _MSC_VER
+#      define png_snprintf _snprintf   /* Added to v 1.2.19 */
+#      define png_snprintf2 _snprintf
+#      define png_snprintf6 _snprintf
+#    else
+#      define png_snprintf snprintf   /* Added to v 1.2.19 */
+#      define png_snprintf2 snprintf
+#      define png_snprintf6 snprintf
+#    endif
+#  else
+     /* You don't have or don't want to use snprintf().  Caution: Using
+      * sprintf instead of snprintf exposes your application to accidental
+      * or malevolent buffer overflows.  If you don't have snprintf()
+      * as a general rule you should provide one (you can get one from
+      * Portable OpenSSH). */
+#    define png_snprintf(s1,n,fmt,x1) sprintf(s1,fmt,x1)
+#    define png_snprintf2(s1,n,fmt,x1,x2) sprintf(s1,fmt,x1,x2)
+#    define png_snprintf6(s1,n,fmt,x1,x2,x3,x4,x5,x6) \
+        sprintf(s1,fmt,x1,x2,x3,x4,x5,x6)
+#  endif
+#  define png_strcpy  strcpy
+#  define png_strncpy strncpy     /* Added to v 1.2.6 */
+#  define png_strlen  strlen
+#  define png_memcmp  memcmp      /* SJT: added */
+#  define png_memcpy  memcpy
+#  define png_memset  memset
+#endif
+/* End of memory model independent support */
+
+/* Just a little check that someone hasn't tried to define something
+ * contradictory.
+ */
+#if (PNG_ZBUF_SIZE > 65536L) && defined(PNG_MAX_MALLOC_64K)
+#  undef PNG_ZBUF_SIZE
+#  define PNG_ZBUF_SIZE 65536L
+#endif
+
+#ifdef PNG_READ_SUPPORTED
+/* Prior to libpng-1.0.9, this block was in pngasmrd.h */
+#if defined(PNG_INTERNAL)
+
+#if defined(PNG_USE_PNGGCCRD) || defined(PNG_USE_PNGVCRD)
+  /* Platform must be Pentium.  Makefile must assemble and load
+   * pnggccrd.c or  pngvcrd.c. MMX will be detected at run time and
+   * used if present.
+   */
+#  ifndef PNG_NO_MMX_COMBINE_ROW
+#    define PNG_HAVE_MMX_COMBINE_ROW
+#  endif
+#  ifndef PNG_NO_MMX_READ_INTERLACE
+#    define PNG_HAVE_MMX_READ_INTERLACE
+#  endif
+#  ifndef PNG_NO_MMX_READ_FILTER_ROW
+#    define PNG_HAVE_MMX_READ_FILTER_ROW
+#    ifndef PNG_NO_MMX_FILTER_SUB
+#      define PNG_MMX_READ_FILTER_SUB_SUPPORTED
+#    endif
+#    if !(defined(__GNUC__) && defined(__x86_64__) && (__GNUC__ < 4))
+       /* work around 64-bit gcc compiler bugs in gcc-3.x */
+#      ifndef PNG_NO_MMX_FILTER_UP
+#        define PNG_MMX_READ_FILTER_UP_SUPPORTED
+#      endif
+#      ifndef PNG_NO_MMX_FILTER_AVG
+#        define PNG_MMX_READ_FILTER_AVG_SUPPORTED
+#      endif
+#      ifndef PNG_NO_MMX_FILTER_PAETH
+#        define PNG_MMX_READ_FILTER_PAETH_SUPPORTED
+#      endif
+#    endif /* !((__x86_64__) && (GNUC < 4)) */
+#  endif
+  /* These are the default thresholds before the MMX code kicks in; if either
+   * rowbytes or bitdepth is below the threshold, plain C code is used.  These
+   * can be overridden at runtime via the png_set_mmx_thresholds() call in
+   * libpng 1.2.0 and later.  The values below were chosen by Intel.
+   */
+#  ifndef PNG_MMX_ROWBYTES_THRESHOLD_DEFAULT
+#    define PNG_MMX_ROWBYTES_THRESHOLD_DEFAULT  128  /*  >=  */
+#  endif
+#  ifndef PNG_MMX_BITDEPTH_THRESHOLD_DEFAULT
+#    define PNG_MMX_BITDEPTH_THRESHOLD_DEFAULT  9    /*  >=  */   
+#  endif
+#endif /* PNG_USE_PNGGCCRD || PNG_USE_PNGVCRD */
+/* - see pngvcrd.c or pnggccrd.c for info about what is currently enabled */
+
+#endif /* PNG_INTERNAL */
+#endif /* PNG_READ_SUPPORTED */
+
+/* Added at libpng-1.2.8 */
+#endif /* PNG_VERSION_INFO_ONLY */
+
+#endif /* PNGCONF_H */
diff --git a/distrib/libpng-1.2.19/pngerror.c b/distrib/libpng-1.2.19/pngerror.c
new file mode 100644
index 0000000..b1d5fbe
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngerror.c
@@ -0,0 +1,326 @@
+
+/* pngerror.c - stub functions for i/o and memory allocation
+ *
+ * Last changed in libpng 1.2.19 August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ *
+ * This file provides a location for all error handling.  Users who
+ * need special error handling are expected to write replacement functions
+ * and use png_set_error_fn() to use those functions.  See the instructions
+ * at each function.
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED)
+static void /* PRIVATE */
+png_default_error PNGARG((png_structp png_ptr,
+  png_const_charp error_message));
+#ifndef PNG_NO_WARNINGS
+static void /* PRIVATE */
+png_default_warning PNGARG((png_structp png_ptr,
+  png_const_charp warning_message));
+#endif /* PNG_NO_WARNINGS */
+
+/* This function is called whenever there is a fatal error.  This function
+ * should not be changed.  If there is a need to handle errors differently,
+ * you should supply a replacement error function and use png_set_error_fn()
+ * to replace the error function at run-time.
+ */
+void PNGAPI
+png_error(png_structp png_ptr, png_const_charp error_message)
+{
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+   char msg[16];
+   if (png_ptr != NULL)
+   {
+     if (png_ptr->flags&
+       (PNG_FLAG_STRIP_ERROR_NUMBERS|PNG_FLAG_STRIP_ERROR_TEXT))
+     {
+       if (*error_message == '#')
+       {
+           int offset;
+           for (offset=1; offset<15; offset++)
+              if (*(error_message+offset) == ' ')
+                  break;
+           if (png_ptr->flags&PNG_FLAG_STRIP_ERROR_TEXT)
+           {
+              int i;
+              for (i=0; i<offset-1; i++)
+                 msg[i]=error_message[i+1];
+              msg[i]='\0';
+              error_message=msg;
+           }
+           else
+              error_message+=offset;
+       }
+       else
+       {
+           if (png_ptr->flags&PNG_FLAG_STRIP_ERROR_TEXT)
+           {
+              msg[0]='0';
+              msg[1]='\0';
+              error_message=msg;
+           }
+       }
+     }
+   }
+#endif
+   if (png_ptr != NULL && png_ptr->error_fn != NULL)
+      (*(png_ptr->error_fn))(png_ptr, error_message);
+
+   /* If the custom handler doesn't exist, or if it returns,
+      use the default handler, which will not return. */
+   png_default_error(png_ptr, error_message);
+}
+
+#ifndef PNG_NO_WARNINGS
+/* This function is called whenever there is a non-fatal error.  This function
+ * should not be changed.  If there is a need to handle warnings differently,
+ * you should supply a replacement warning function and use
+ * png_set_error_fn() to replace the warning function at run-time.
+ */
+void PNGAPI
+png_warning(png_structp png_ptr, png_const_charp warning_message)
+{
+   int offset = 0;
+   if (png_ptr != NULL)
+   {
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+   if (png_ptr->flags&
+     (PNG_FLAG_STRIP_ERROR_NUMBERS|PNG_FLAG_STRIP_ERROR_TEXT))
+#endif
+     {
+       if (*warning_message == '#')
+       {
+           for (offset=1; offset<15; offset++)
+              if (*(warning_message+offset) == ' ')
+                  break;
+       }
+     }
+     if (png_ptr != NULL && png_ptr->warning_fn != NULL)
+        (*(png_ptr->warning_fn))(png_ptr, warning_message+offset);
+   }
+   else
+      png_default_warning(png_ptr, warning_message+offset);
+}
+#endif /* PNG_NO_WARNINGS */
+
+
+/* These utilities are used internally to build an error message that relates
+ * to the current chunk.  The chunk name comes from png_ptr->chunk_name,
+ * this is used to prefix the message.  The message is limited in length
+ * to 63 bytes, the name characters are output as hex digits wrapped in []
+ * if the character is invalid.
+ */
+#define isnonalpha(c) ((c) < 65 || (c) > 122 || ((c) > 90 && (c) < 97))
+static PNG_CONST char png_digit[16] = {
+   '0', '1', '2', '3', '4', '5', '6', '7', '8', '9',
+   'A', 'B', 'C', 'D', 'E', 'F'
+};
+
+static void /* PRIVATE */
+png_format_buffer(png_structp png_ptr, png_charp buffer, png_const_charp
+   error_message)
+{
+   int iout = 0, iin = 0;
+
+   while (iin < 4)
+   {
+      int c = png_ptr->chunk_name[iin++];
+      if (isnonalpha(c))
+      {
+         buffer[iout++] = '[';
+         buffer[iout++] = png_digit[(c & 0xf0) >> 4];
+         buffer[iout++] = png_digit[c & 0x0f];
+         buffer[iout++] = ']';
+      }
+      else
+      {
+         buffer[iout++] = (png_byte)c;
+      }
+   }
+
+   if (error_message == NULL)
+      buffer[iout] = 0;
+   else
+   {
+      buffer[iout++] = ':';
+      buffer[iout++] = ' ';
+      png_strncpy(buffer+iout, error_message, 63);
+      buffer[iout+63] = 0;
+   }
+}
+
+#ifdef PNG_READ_SUPPORTED
+void PNGAPI
+png_chunk_error(png_structp png_ptr, png_const_charp error_message)
+{
+   char msg[18+64];
+   if (png_ptr == NULL)
+     png_error(png_ptr, error_message);
+   else
+   {
+     png_format_buffer(png_ptr, msg, error_message);
+     png_error(png_ptr, msg);
+   }
+}
+
+#ifndef PNG_NO_WARNINGS
+void PNGAPI
+png_chunk_warning(png_structp png_ptr, png_const_charp warning_message)
+{
+   char msg[18+64];
+   if (png_ptr == NULL)
+     png_warning(png_ptr, warning_message);
+   else
+   {
+     png_format_buffer(png_ptr, msg, warning_message);
+     png_warning(png_ptr, msg);
+   }
+}
+#endif /* PNG_NO_WARNINGS */
+
+#endif /* PNG_READ_SUPPORTED */
+
+/* This is the default error handling function.  Note that replacements for
+ * this function MUST NOT RETURN, or the program will likely crash.  This
+ * function is used by default, or if the program supplies NULL for the
+ * error function pointer in png_set_error_fn().
+ */
+static void /* PRIVATE */
+png_default_error(png_structp png_ptr, png_const_charp error_message)
+{
+#ifndef PNG_NO_CONSOLE_IO
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+   if (*error_message == '#')
+   {
+     int offset;
+     char error_number[16];
+     for (offset=0; offset<15; offset++)
+     {
+         error_number[offset] = *(error_message+offset+1);
+         if (*(error_message+offset) == ' ')
+             break;
+     }
+     if((offset > 1) && (offset < 15))
+     {
+       error_number[offset-1]='\0';
+       fprintf(stderr, "libpng error no. %s: %s\n", error_number,
+          error_message+offset);
+     }
+     else
+       fprintf(stderr, "libpng error: %s, offset=%d\n", error_message,offset);
+   }
+   else
+#endif
+   fprintf(stderr, "libpng error: %s\n", error_message);
+#endif
+
+#ifdef PNG_SETJMP_SUPPORTED
+   if (png_ptr)
+   {
+#  ifdef USE_FAR_KEYWORD
+   {
+      jmp_buf jmpbuf;
+      png_memcpy(jmpbuf, png_ptr->jmpbuf, png_sizeof(jmp_buf));
+      longjmp(jmpbuf, 1);
+   }
+#  else
+   longjmp(png_ptr->jmpbuf, 1);
+#  endif
+   }
+#else
+   PNG_ABORT();
+#endif
+#ifdef PNG_NO_CONSOLE_IO
+   error_message = error_message; /* make compiler happy */
+#endif
+}
+
+#ifndef PNG_NO_WARNINGS
+/* This function is called when there is a warning, but the library thinks
+ * it can continue anyway.  Replacement functions don't have to do anything
+ * here if you don't want them to.  In the default configuration, png_ptr is
+ * not used, but it is passed in case it may be useful.
+ */
+static void /* PRIVATE */
+png_default_warning(png_structp png_ptr, png_const_charp warning_message)
+{
+#ifndef PNG_NO_CONSOLE_IO
+#  ifdef PNG_ERROR_NUMBERS_SUPPORTED
+   if (*warning_message == '#')
+   {
+     int offset;
+     char warning_number[16];
+     for (offset=0; offset<15; offset++)
+     {
+        warning_number[offset]=*(warning_message+offset+1);
+        if (*(warning_message+offset) == ' ')
+            break;
+     }
+     if((offset > 1) && (offset < 15))
+     {
+       warning_number[offset-1]='\0';
+       fprintf(stderr, "libpng warning no. %s: %s\n", warning_number,
+          warning_message+offset);
+     }
+     else
+       fprintf(stderr, "libpng warning: %s\n", warning_message);
+   }
+   else
+#  endif
+     fprintf(stderr, "libpng warning: %s\n", warning_message);
+#else
+   warning_message = warning_message; /* make compiler happy */
+#endif
+   png_ptr = png_ptr; /* make compiler happy */
+}
+#endif /* PNG_NO_WARNINGS */
+
+/* This function is called when the application wants to use another method
+ * of handling errors and warnings.  Note that the error function MUST NOT
+ * return to the calling routine or serious problems will occur.  The return
+ * method used in the default routine calls longjmp(png_ptr->jmpbuf, 1)
+ */
+void PNGAPI
+png_set_error_fn(png_structp png_ptr, png_voidp error_ptr,
+   png_error_ptr error_fn, png_error_ptr warning_fn)
+{
+   if (png_ptr == NULL)
+      return;
+   png_ptr->error_ptr = error_ptr;
+   png_ptr->error_fn = error_fn;
+   png_ptr->warning_fn = warning_fn;
+}
+
+
+/* This function returns a pointer to the error_ptr associated with the user
+ * functions.  The application should free any memory associated with this
+ * pointer before png_write_destroy and png_read_destroy are called.
+ */
+png_voidp PNGAPI
+png_get_error_ptr(png_structp png_ptr)
+{
+   if (png_ptr == NULL)
+      return NULL;
+   return ((png_voidp)png_ptr->error_ptr);
+}
+
+
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+void PNGAPI
+png_set_strip_error_numbers(png_structp png_ptr, png_uint_32 strip_mode)
+{
+   if(png_ptr != NULL)
+   {
+     png_ptr->flags &=
+       ((~(PNG_FLAG_STRIP_ERROR_NUMBERS|PNG_FLAG_STRIP_ERROR_TEXT))&strip_mode);
+   }
+}
+#endif
+#endif /* PNG_READ_SUPPORTED || PNG_WRITE_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pnggccrd.c b/distrib/libpng-1.2.19/pnggccrd.c
new file mode 100644
index 0000000..f63867c
--- /dev/null
+++ b/distrib/libpng-1.2.19/pnggccrd.c
@@ -0,0 +1,6041 @@
+
+/* pnggccrd.c - mixed C/assembler version of utilities to read a PNG file
+ *
+ * For Intel/AMD x86 or x86-64 CPU (Pentium-MMX or later) and GNU C compiler.
+ *
+ * Last changed in libpng 1.2.19 August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998 Intel Corporation
+ * Copyright (c) 1999-2002,2007 Greg Roelofs
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ *
+ * Based on MSVC code contributed by Nirav Chhatrapati, Intel Corp., 1998.
+ * Interface to libpng contributed by Gilles Vollant, 1999.
+ * GNU C port by Greg Roelofs, 1999-2001.
+ *
+ * References:
+ *
+ *     http://www.intel.com/drg/pentiumII/appnotes/916/916.htm
+ *     http://www.intel.com/drg/pentiumII/appnotes/923/923.htm
+ *       [Intel's performance analysis of the MMX vs. non-MMX code;
+ *        moved/deleted as of 2006, but text and some graphs still
+ *        available via WayBack Machine at archive.org]
+ *
+ *     http://www.ibiblio.org/gferg/ldp/GCC-Inline-Assembly-HOWTO.html
+ *     http://sam.zoy.org/blog/2007-04-13-shlib-with-non-pic-code-have-inline-assembly-and-pic-mix-well
+ *     http://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html
+ *     http://gcc.gnu.org/onlinedocs/gcc/Variable-Attributes.html
+ *     http://gcc.gnu.org/onlinedocs/gcc/Function-Attributes.html
+ *     AMD64 Architecture Programmer's Manual, volumes 1 and 5
+ *       [http://www.amd.com/us-en/Processors/TechnicalResources/0,,30_182_739_7044,00.html]
+ *     Intel 64 and IA-32 Software Developer's Manuals
+ *       [http://developer.intel.com/products/processor/manuals/]
+ *
+ * png_read_filter_row_mmx_*() were converted in place with intel2gas 1.3.1:
+ *
+ *     intel2gas -mdI pnggccrd.c.partially-msvc -o pnggccrd.c
+ *
+ * and then cleaned up by hand.  See http://hermes.terminal.at/intel2gas/ .
+ *
+ * NOTE:  A sufficiently recent version of GNU as (or as.exe under DOS/Windows)
+ * is required to assemble the newer asm instructions such as movq.  (Version
+ * 2.5.2l.15 is definitely too old.)  See ftp://ftp.gnu.org/pub/gnu/binutils/ .
+ */
+
+/*
+ * PORTING NOTES AND CHANGELOG (mostly by Greg Roelofs)
+ * ===========================
+ *
+ * 19991006:
+ *  - fixed sign error in post-MMX cleanup code (16- & 32-bit cases)
+ *
+ * 19991007:
+ *  - additional optimizations (possible or definite):
+ *     x [DONE] write MMX code for 64-bit case (pixel_bytes == 8) [not tested]
+ *     - write MMX code for 48-bit case (pixel_bytes == 6)
+ *     - figure out what's up with 24-bit case (pixel_bytes == 3):
+ *        why subtract 8 from width_mmx in the pass 4/5 case?
+ *        (only width_mmx case) (near line 2335)
+ *     x [DONE] replace pixel_bytes within each block with the true
+ *        constant value (or are compilers smart enough to do that?)
+ *     - rewrite all MMX interlacing code so it's aligned with
+ *        the *beginning* of the row buffer, not the end.  This
+ *        would not only allow one to eliminate half of the memory
+ *        writes for odd passes (that is, pass == odd), it may also
+ *        eliminate some unaligned-data-access exceptions (assuming
+ *        there's a penalty for not aligning 64-bit accesses on
+ *        64-bit boundaries).  The only catch is that the "leftover"
+ *        pixel(s) at the end of the row would have to be saved,
+ *        but there are enough unused MMX registers in every case,
+ *        so this is not a problem.  A further benefit is that the
+ *        post-MMX cleanup code (C code) in at least some of the
+ *        cases could be done within the assembler block.
+ *  x [DONE] the "v3 v2 v1 v0 v7 v6 v5 v4" comments are confusing,
+ *     inconsistent, and don't match the MMX Programmer's Reference
+ *     Manual conventions anyway.  They should be changed to
+ *     "b7 b6 b5 b4 b3 b2 b1 b0," where b0 indicates the byte that
+ *     was lowest in memory (i.e., corresponding to a left pixel)
+ *     and b7 is the byte that was highest (i.e., a right pixel).
+ *
+ * 19991016:
+ *  - Brennan's Guide notwithstanding, gcc under Linux does *not*
+ *     want globals prefixed by underscores when referencing them--
+ *     i.e., if the variable is const4, then refer to it as const4,
+ *     not _const4.  This seems to be a djgpp-specific requirement.
+ *     Also, such variables apparently *must* be declared outside
+ *     of functions; neither static nor automatic variables work if
+ *     defined within the scope of a single function, but both
+ *     static and truly global (multi-module) variables work fine.
+ *
+ * 19991017:
+ *  - replaced pixel_bytes in each png_memcpy() call with constant value for
+ *     inlining (png_do_read_interlace() "non-MMX/modified C code" block)
+ *
+ * 19991023:
+ *  - fixed png_combine_row() non-MMX replication bug (odd passes only?)
+ *  - switched from string-concatenation-with-macros to cleaner method of
+ *     renaming global variables for djgpp--i.e., always use prefixes in
+ *     inlined assembler code (== strings) and conditionally rename the
+ *     variables, not the other way around.  Hence _const4, _mask8_0, etc.
+ *
+ * 19991024:
+ *  - fixed mmxsupport()/png_do_read_interlace() first-row bug
+ *     This one was severely weird:  even though mmxsupport() doesn't touch
+ *     ebx (where "row" pointer was stored), it nevertheless managed to zero
+ *     the register (even in static/non-fPIC code--see below), which in turn
+ *     caused png_do_read_interlace() to return prematurely on the first row of
+ *     interlaced images (i.e., without expanding the interlaced pixels).
+ *     Inspection of the generated assembly code didn't turn up any clues,
+ *     although it did point at a minor optimization (i.e., get rid of
+ *     mmx_supported_local variable and just use eax).  Possibly the CPUID
+ *     instruction is more destructive than it looks?  (Not yet checked.)
+ *  - "info gcc" was next to useless, so compared fPIC and non-fPIC assembly
+ *     listings...  Apparently register spillage has to do with ebx, since
+ *     it's used to index the global offset table.  Commenting it out of the
+ *     input-reg lists in png_combine_row() eliminated compiler barfage, so
+ *     ifdef'd with __PIC__ macro:  if defined, use a global for unmask
+ *
+ * 19991107:
+ *  - verified CPUID clobberage:  12-char string constant ("GenuineIntel",
+ *     "AuthenticAMD", etc.) placed in ebx:ecx:edx.  Still need to polish.
+ *
+ * 19991120:
+ *  - made "diff" variable (now "_dif") global to simplify conversion of
+ *     filtering routines (running out of regs, sigh).  "diff" is still used
+ *     in interlacing routines, however.
+ *  - fixed up both versions of mmxsupport() (ORIG_THAT_USED_TO_CLOBBER_EBX
+ *     macro determines which is used); original not yet tested.
+ *
+ * 20000213:
+ *  - when compiling with gcc, be sure to use  -fomit-frame-pointer
+ *
+ * 20000319:
+ *  - fixed a register-name typo in png_do_read_interlace(), default (MMX) case,
+ *     pass == 4 or 5, that caused visible corruption of interlaced images
+ *
+ * 20000623:
+ *  - Various problems were reported with gcc 2.95.2 in the Cygwin environment,
+ *     many of the form "forbidden register 0 (ax) was spilled for class AREG."
+ *     This is explained at http://gcc.gnu.org/fom_serv/cache/23.html, and
+ *     Chuck Wilson supplied a patch involving dummy output registers.  See
+ *     http://sourceforge.net/bugs/?func=detailbug&bug_id=108741&group_id=5624
+ *     for the original (anonymous) SourceForge bug report.
+ *
+ * 20000706:
+ *  - Chuck Wilson passed along these remaining gcc 2.95.2 errors:
+ *       pnggccrd.c: In function `png_combine_row':
+ *       pnggccrd.c:525: more than 10 operands in `asm'
+ *       pnggccrd.c:669: more than 10 operands in `asm'
+ *       pnggccrd.c:828: more than 10 operands in `asm'
+ *       pnggccrd.c:994: more than 10 operands in `asm'
+ *       pnggccrd.c:1177: more than 10 operands in `asm'
+ *     They are all the same problem and can be worked around by using the
+ *     global _unmask variable unconditionally, not just in the -fPIC case.
+ *     Reportedly earlier versions of gcc also have the problem with more than
+ *     10 operands; they just don't report it.  Much strangeness ensues, etc.
+ *
+ * 20000729:
+ *  - enabled png_read_filter_row_mmx_up() (shortest remaining unconverted
+ *     MMX routine); began converting png_read_filter_row_mmx_sub()
+ *  - to finish remaining sections:
+ *     - clean up indentation and comments
+ *     - preload local variables
+ *     - add output and input regs (order of former determines numerical
+ *        mapping of latter)
+ *     - avoid all usage of ebx (including bx, bh, bl) register [20000823]
+ *     - remove "$" from addressing of Shift and Mask variables [20000823]
+ *
+ * 20000731:
+ *  - global union vars causing segfaults in png_read_filter_row_mmx_sub()?
+ *
+ * 20000822:
+ *  - ARGH, stupid png_read_filter_row_mmx_sub() segfault only happens with
+ *     shared-library (-fPIC) version!  Code works just fine as part of static
+ *     library.  Should have tested that sooner.
+ *     ebx is getting clobbered again (explicitly this time); need to save it
+ *     on stack or rewrite asm code to avoid using it altogether.  Blargh!
+ *
+ * 20000823:
+ *  - first section was trickiest; all remaining sections have ebx -> edx now.
+ *     (-fPIC works again.)  Also added missing underscores to various Shift*
+ *     and *Mask* globals and got rid of leading "$" signs.
+ *
+ * 20000826:
+ *  - added visual separators to help navigate microscopic printed copies
+ *     (http://pobox.com/~newt/code/gpr-latest.zip, mode 10); started working
+ *     on png_read_filter_row_mmx_avg()
+ *
+ * 20000828:
+ *  - finished png_read_filter_row_mmx_avg():  only Paeth left! (930 lines...)
+ *     What the hell, did png_read_filter_row_mmx_paeth(), too.  Comments not
+ *     cleaned up/shortened in either routine, but functionality is complete
+ *     and seems to be working fine.
+ *
+ * 20000829:
+ *  - ahhh, figured out last(?) bit of gcc/gas asm-fu:  if register is listed
+ *     as an input reg (with dummy output variables, etc.), then it *cannot*
+ *     also appear in the clobber list or gcc 2.95.2 will barf.  The solution
+ *     is simple enough...
+ *
+ * 20000914:
+ *  - bug in png_read_filter_row_mmx_avg():  16-bit grayscale not handled
+ *     correctly (but 48-bit RGB just fine)
+ *
+ * 20000916:
+ *  - fixed bug in png_read_filter_row_mmx_avg(), bpp == 2 case; three errors:
+ *     - "_ShiftBpp.use = 24;"      should have been   "_ShiftBpp.use = 16;"
+ *     - "_ShiftRem.use = 40;"      should have been   "_ShiftRem.use = 48;"
+ *     - "psllq _ShiftRem, %%mm2"   should have been   "psrlq _ShiftRem, %%mm2"
+ *
+ * 20010101:
+ *  - added new png_init_mmx_flags() function (here only because it needs to
+ *     call mmxsupport(), which should probably become global png_mmxsupport());
+ *     modified other MMX routines to run conditionally (png_ptr->asm_flags)
+ *
+ * 20010103:
+ *  - renamed mmxsupport() to png_mmx_support(), with auto-set of mmx_supported,
+ *     and made it public; moved png_init_mmx_flags() to png.c as internal func
+ *
+ * 20010104:
+ *  - removed dependency on png_read_filter_row_c() (C code already duplicated
+ *     within MMX version of png_read_filter_row()) so no longer necessary to
+ *     compile it into pngrutil.o
+ *
+ * 20010310:
+ *  - fixed buffer-overrun bug in png_combine_row() C code (non-MMX)
+ *
+ * 20010808:
+ *  - added PNG_THREAD_UNSAFE_OK around code using global variables [GR-P]
+ *
+ * 20011124:
+ *  - fixed missing save of Eflag in png_mmx_support() [Maxim Sobolev]
+ *
+ * 20020304:
+ *  - eliminated incorrect use of width_mmx in pixel_bytes == 8 case
+ *
+ * 20020407:
+ *  - fixed insufficient preservation of ebx register [Sami Farin]
+ *
+ * 20040724:
+ *  - more tinkering with clobber list at lines 4529 and 5033 to get it to
+ *     compile with gcc 3.4 [GR-P]
+ *
+ * 20040809:
+ *  - added "rim" definitions for CONST4 and CONST6 [GR-P]
+ *
+ * 20060303:
+ *  - added "OS2" to list of systems that don't need leading underscores [GR-P]
+ *
+ * 20060320:
+ *  - made PIC-compliant [Christian Aichinger]
+ *
+ * 20070313:
+ *  - finally applied Giuseppe Ghibò's 64-bit patch of 20060803 (completely
+ *     overlooked Dylan Alex Simon's similar patch of 20060414, oops...)
+ *
+ * 20070524:
+ *  - fixed link failure caused by asm-only variables being optimized out
+ *     (identified by Dimitri of Trolltech) with __attribute__((used)), which
+ *     also gets rid of warnings => nuked ugly png_squelch_warnings() hack
+ *  - dropped redundant ifdef
+ *  - moved png_mmx_support() back up where originally intended (as in
+ *     pngvcrd.c), using __attribute__((noinline)) in extra prototype
+ *
+ * 20070527:
+ *  - revised png_combine_row() to reuse mask in lieu of external _unmask
+ *  - moved 32-bit (RGBA) case to top of png_combine_row():  most common
+ *  - just about ready to give up on x86-64 -fPIC mode; can't even access 16
+ *     _mask*_* constants without triggering link error on shared library:
+ *       /usr/bin/ld: pnggccrd.pic.o: relocation R_X86_64_32S against `a local
+ *         symbol' can not be used when making a shared object; recompile with
+ *         -fPIC
+ *       pnggccrd.pic.o: could not read symbols: Bad value
+ *       ("objdump -x pnggccrd.pic.o | grep rodata" to verify)
+ *     [might be able to work around by doing within assembly code whatever
+ *     -fPIC does, but given problems to date, seems like long shot...]
+ *     [relevant ifdefs:  __x86_64__ && __PIC__ => C code only]
+ *  - changed #if 0 to #ifdef PNG_CLOBBER_MMX_REGS_SUPPORTED in case gcc ever
+ *     supports MMX regs (%mm0, etc.) in clobber list (not supported by gcc
+ *     2.7.2.3, 2.91.66 (egcs 1.1.2), 3.x, or 4.1.2)
+ *
+ * 20070603:
+ *  - revised png_combine_row() to use @GOTPCREL(%%rip) addressing on _c64
+ *     struct of _mask*_* constants for x86-64 -fPIC; see sam.zoy.org link
+ *     above for details
+ *  - moved _const4 and _const6 into _c64 struct, renamed to _amask5_3_0 and
+ *     _amask7_1_0, respectively
+ *  - can't figure out how to use _c64._mask*_* vars within asm code, so still
+ *     need single variables for non-x86-64/-fPIC half :-(
+ *  - replaced various __PIC__ ifdefs with *_GOT_ebx macros
+ *  - moved _LBCarryMask and _HBClearMask into _c64 struct
+ *  - conditionally replaced _p*temp variables with %r11d-%r13d (via p*_TEMP
+ *     and CLOBBER_r1*d macros)
+ *
+ * 20070604:
+ *  - replaced all _ActiveMask and _ActiveMaskEnd with new _amask*_*_* consts
+ *     (_amask naming convention:  numbers of 00-bytes, ff-bytes, 00-bytes)
+ *    - _ActiveMask     // (10) // avg/paeth/sub; read-only; consts; movq/pand
+ *       0x0000000000ffffffLL (bpp 3, avg)      _amask5_3_0
+ *       0xffffffffffffffffLL (bpp 4, 6, avg)   _amask0_8_0
+ *       0x000000000000ffffLL (bpp 2, avg)      _amask6_2_0
+ *       0x0000000000ffffffLL (bpp 3, paeth)    _amask5_3_0
+ *       0x00000000ffffffffLL (bpp 6, paeth)    _amask4_4_0
+ *       0x00000000ffffffffLL (bpp 4, paeth)    _amask4_4_0
+ *       0x00000000ffffffffLL (bpp 8, paeth)    _amask4_4_0
+ *       0x0000ffffff000000LL (bpp 3, sub)      _amask2_3_3
+ *       0x00000000ffff0000LL (bpp 2, sub)      _amask4_2_2
+ *    - _ActiveMaskEnd  // (1)  // paeth only; read-only; const; pand
+ *       0xffff000000000000LL (bpp 3, paeth)    _amask0_2_6
+ *  - changed all "#if defined(__x86_64__) // later // && defined(__PIC__)"
+ *     lines to "#ifdef PNG_x86_64_USE_GOTPCREL" for easier/safer testing
+ *
+ * 20070605:
+ *  - merged PNG_x86_64_USE_GOTPCREL, non-PNG_x86_64_USE_GOTPCREL code via
+ *     *MASK* and LOAD/RESTORE macros
+ *
+ * 20070607:
+ *  - replaced all constant instances of _ShiftBpp, _ShiftRem with immediates
+ *     (still have two shared cases in avg, sub routines)
+ *
+ * 20070609:
+ *  - replaced remaining instances of _ShiftBpp, _ShiftRem with immediates
+ *     (split sub and avg 4/6-bpp cases into separate blocks)
+ *  - fixed paeth bug due to clobbered r11/r12/r13 regs
+ *
+ * 20070610:
+ *  - made global "_dif" variable (avg/paeth/sub routines) local again (now
+ *     "diff"--see 19991120 entry above), using register constraints
+ *  - note that %ebp in clobber list doesn't actually work, at least for 32-bit
+ *     version and gcc 4.1.2; must save and restore manually.  (Seems to work
+ *     OK for 64-bit version and gcc 3.4.3, but gcc may not be using ebp/rbp
+ *     in that case.)
+ *  - started replacing direct _MMXLength accesses with register constraints
+ *
+ * 20070612:
+ *  - continued replacing direct _MMXLength accesses with register constraints
+ *
+ * 20070613:
+ *  - finished replacing direct _MMXLength accesses with register constraints;
+ *     switched to local variable (and renamed back to MMXLength)
+ *
+ * 20070614:
+ *  - fixed sub bpp = 1 bug
+ *  - started replacing direct _FullLength accesses with register constraints
+ *
+ * 20070615:
+ *  - fixed 64-bit paeth bpp 3 crash bug (misplaced LOAD_GOT_rbp)
+ *  - fixed 64-bit paeth bpp 1/2 and cleanup-block crash bugs (misplaced
+ *     RESTORE_r11_r12_r13)
+ *  - slightly optimized avg/paeth cleanup blocks and paeth bpp 1/2 block
+ *     (save/restore ebx only if needed)
+ *  - continued replacing direct _FullLength accesses with register constraints
+ *
+ * 20070616:
+ *  - finished replacing direct _FullLength accesses with register constraints
+ *     (*ugly* conditional clobber-separator macros for avg and paeth, sigh)
+ *
+ * 20070618:
+ *  - fixed misplaced PNG_THREAD_UNSAFE_OK endif (was missing LOAD_GOT_rbp/
+ *     RESTORE_rbp in 32-bit thread-safe case)
+ *  - changed all "ifdef *" to "if defined(*)" [GR-P]
+ *
+ * 20070619:
+ *  - rearranged most bitdepth-related case statements to put most frequent
+ *     cases at top (24-bit, 32-bit, 8-bit, rest)
+ *
+ * 20070623:
+ *  - cleaned up png_debug() warnings/formatting
+ *  - removed PNG_MMX_CODE_SUPPORTED ifdefs and added outer __GNUC__ ifdef
+ *     (module no longer used by non-x86/non-GCC builds as of libpng 1.2.19)
+ *  - removed single libpng-1.2.x PNG_DEBUG dependency on 1.0.x png_struct
+ *     member (row_buf_size)
+ *  - rearranged pass-related if-blocks in png_do_read_interlace() to put most
+ *     frequent cases (4, 5) at top [GR-P suggestion]
+ *
+ * 20070624-29:
+ *  - fixed 64-bit crash bug:  pointers -> rsi/rdi, not esi/edi (switched to
+ *     %0/%1/%2/%3/%4 notation; eliminated size suffixes from relevant add/
+ *     inc/sub/mov instructions; changed dummy vars to pointers)
+ *     - png_combine_row()
+ *     - png_do_read_interlace()
+ *     - png_read_filter_row_mmx_avg()
+ *     - png_read_filter_row_mmx_paeth()
+ *     - png_read_filter_row_mmx_sub()
+ *     - png_read_filter_row_mmx_up()
+ *  - NOTE:  this fix makes use of the fact that modifying a 32-bit reg (e.g.,
+ *     %%ebx) clears the top half of its corresponding 64-bit reg (%%rbx), so
+ *     it's safe to mix 32-bit operations with 64-bit base/index addressing
+ *     (see new PSI/PAX/PBX/PDX/PBP/etc. "pointer-register" macros); applies
+ *     also to clobber lists
+ *
+ * 20070630:
+ *  - cleaned up formatting, macros, minor png_read_filter_row_mmx_sub() 8-bpp
+ *     register-usage inefficiency
+ *  - fixed 32-bit png_do_read_interlace() bug (was using pointer size for
+ *     64-bit dummy values)
+ *
+ * 20070703:
+ *  - added check for (manual) PIC macro to fix OpenBSD crash bug
+ *
+ * 20070717:
+ *  - fixed 48-bit png_combine_row() bug (was acting like 32-bit):  copy 6
+ *     bytes per pixel, not 4, and use stride of 6, not 4, in the second loop
+ *     of interlace processing of 48-bit pixels [GR-P]
+ *
+ * 20070722:
+ *  - fixed 64-bit png_uint_32 bug with MMXLength/FullLength temp vars
+ *
+ * [still broken:  tops of all row-filter blocks (input/output constraints);
+ *  shows up on 64-bit dynamic (-fPIC) version with -O2, especially if debug-
+ *  printfs enabled, but at right edge of odd-width images even if disabled]
+ *
+ *
+ * STILL TO DO:
+ *  - fix final thread-unsafe code using stack vars and pointer? (paeth top,
+ *     default, bottom only:  default, bottom already 5 reg constraints; could
+ *     replace bpp with pointer and group bpp/patemp/pbtemp/pctemp in array)
+ *  - fix ebp/no-reg-constraint inefficiency (avg/paeth/sub top)
+ *  - test png_do_read_interlace() 64-bit case (pixel_bytes == 8)
+ *  - write MMX code for 48-bit case (pixel_bytes == 6)
+ *  - figure out what's up with 24-bit case (pixel_bytes == 3):
+ *     why subtract 8 from width_mmx in the pass 4/5 case?  due to
+ *     odd number of bytes? (only width_mmx case) (near line 2335)
+ *  - rewrite all MMX interlacing code so it's aligned with beginning
+ *     of the row buffer, not the end (see 19991007 for details)
+ *  - add error messages to any remaining bogus default cases
+ *  - enable pixel_depth == 8 cases in png_read_filter_row()? (test speed)
+ *  - try =r, etc., as reg constraints?  (would gcc use 64-bit ones on x86-64?)
+ *  - need full, non-graphical, CRC-based test suite...  maybe autogenerate
+ *     random data of various height/width/depth, compute CRCs, write (C
+ *     funcs), read (asm/MMX), recompute CRCs, and compare?
+ *  - write true x86-64 version using 128-bit "media instructions", %xmm0-15,
+ *     and extra general-purpose registers
+ */
+
+#if defined(__GNUC__)
+
+#define PNG_INTERNAL
+#include "png.h"
+
+
+/* for some inexplicable reason, gcc 3.3.5 on OpenBSD (and elsewhere?) does
+ * *not* define __PIC__ when the -fPIC option is used, so we have to rely on
+ * makefiles and whatnot to define the PIC macro explicitly */
+#if defined(PIC) && !defined(__PIC__)   // (this can/should move to pngconf.h)
+#  define __PIC__
+#endif
+
+#if defined(PNG_ASSEMBLER_CODE_SUPPORTED) && defined(PNG_USE_PNGGCCRD)
+
+/* if you want/need full thread-safety on x86-64 even when linking statically,
+ * comment out the "&& defined(__PIC__)" part here: */
+#if defined(__x86_64__) && defined(__PIC__)
+#  define PNG_x86_64_USE_GOTPCREL            // GOTPCREL => full thread-safety
+#  define PNG_CLOBBER_x86_64_REGS_SUPPORTED  // works as of gcc 3.4.3 ...
+#endif
+
+int PNGAPI png_mmx_support(void);
+
+#if defined(PNG_USE_LOCAL_ARRAYS)
+static PNG_CONST int FARDATA png_pass_start[7] = {0, 4, 0, 2, 0, 1, 0};
+static PNG_CONST int FARDATA png_pass_inc[7]   = {8, 8, 4, 4, 2, 2, 1};
+static PNG_CONST int FARDATA png_pass_width[7] = {8, 4, 4, 2, 2, 1, 1};
+#endif
+
+/* djgpp, Win32, Cygwin, and OS2 add their own underscores to global variables,
+ * so define them without: */
+#if defined(__DJGPP__) || defined(WIN32) || defined(__CYGWIN__) || \
+    defined(__OS2__)
+#  define _mmx_supported  mmx_supported
+#  define _mask8_0        mask8_0
+#  define _mask16_1       mask16_1
+#  define _mask16_0       mask16_0
+#  define _mask24_2       mask24_2
+#  define _mask24_1       mask24_1
+#  define _mask24_0       mask24_0
+#  define _mask32_3       mask32_3
+#  define _mask32_2       mask32_2
+#  define _mask32_1       mask32_1
+#  define _mask32_0       mask32_0
+#  define _mask48_5       mask48_5
+#  define _mask48_4       mask48_4
+#  define _mask48_3       mask48_3
+#  define _mask48_2       mask48_2
+#  define _mask48_1       mask48_1
+#  define _mask48_0       mask48_0
+#  define _amask5_3_0     amask5_3_0
+#  define _amask7_1_0     amask7_1_0
+#  define _LBCarryMask    LBCarryMask
+#  define _HBClearMask    HBClearMask
+#  define _amask0_8_0     amask0_8_0
+#  define _amask6_2_0     amask6_2_0
+#  define _amask4_4_0     amask4_4_0
+#  define _amask0_2_6     amask0_2_6
+#  define _amask2_3_3     amask2_3_3
+#  define _amask4_2_2     amask4_2_2
+#  if defined(PNG_THREAD_UNSAFE_OK)
+#    define _patemp       patemp
+#    define _pbtemp       pbtemp
+#    define _pctemp       pctemp
+#  endif
+#endif // djgpp, Win32, Cygwin, OS2
+
+
+/* These constants are used in the inlined MMX assembly code. */
+
+typedef unsigned long long  ull;
+
+#if defined(PNG_x86_64_USE_GOTPCREL)
+static PNG_CONST struct {
+    //ull _mask_array[26];
+
+    // png_combine_row() constants:
+    ull _mask8_0;
+    ull _mask16_0, _mask16_1;
+    ull _mask24_0, _mask24_1, _mask24_2;
+    ull _mask32_0, _mask32_1, _mask32_2, _mask32_3;
+    ull _mask48_0, _mask48_1, _mask48_2, _mask48_3, _mask48_4, _mask48_5;
+
+    // png_do_read_interlace() constants:
+    ull _amask5_3_0, _amask7_1_0;  // was _const4 and _const6, respectively
+
+    // png_read_filter_row_mmx_avg() constants (also uses _amask5_3_0):
+    ull _LBCarryMask, _HBClearMask;
+    ull _amask0_8_0, _amask6_2_0;  // was ActiveMask for bpp 4/6 and 2 cases
+
+    // png_read_filter_row_mmx_paeth() constants (also uses _amask5_3_0):
+    ull _amask4_4_0, _amask0_2_6;  // was ActiveMask{,End} for bpp 6/4/8 and 3
+
+    // png_read_filter_row_mmx_sub() constants:
+    ull _amask2_3_3, _amask4_2_2;  // was ActiveMask for bpp 3 and 2 cases
+
+} _c64 __attribute__((used, aligned(8))) = {
+
+    // png_combine_row() constants:
+    0x0102040810204080LL, // _mask8_0      offset 0
+
+    0x1010202040408080LL, // _mask16_0     offset 8
+    0x0101020204040808LL, // _mask16_1     offset 16
+
+    0x2020404040808080LL, // _mask24_0     offset 24
+    0x0408080810101020LL, // _mask24_1     offset 32
+    0x0101010202020404LL, // _mask24_2     offset 40
+
+    0x4040404080808080LL, // _mask32_0     offset 48
+    0x1010101020202020LL, // _mask32_1     offset 56
+    0x0404040408080808LL, // _mask32_2     offset 64
+    0x0101010102020202LL, // _mask32_3     offset 72
+
+    0x4040808080808080LL, // _mask48_0     offset 80
+    0x2020202040404040LL, // _mask48_1     offset 88
+    0x1010101010102020LL, // _mask48_2     offset 96
+    0x0404080808080808LL, // _mask48_3     offset 104
+    0x0202020204040404LL, // _mask48_4     offset 112
+    0x0101010101010202LL, // _mask48_5     offset 120
+
+    // png_do_read_interlace() constants:
+    0x0000000000FFFFFFLL, // _amask5_3_0   offset 128  (bpp 3, avg/paeth) const4
+    0x00000000000000FFLL, // _amask7_1_0   offset 136                     const6
+
+    // png_read_filter_row_mmx_avg() constants:
+    0x0101010101010101LL, // _LBCarryMask  offset 144
+    0x7F7F7F7F7F7F7F7FLL, // _HBClearMask  offset 152
+    0xFFFFFFFFFFFFFFFFLL, // _amask0_8_0   offset 160  (bpp 4/6, avg)
+    0x000000000000FFFFLL, // _amask6_2_0   offset 168  (bpp 2,   avg)
+
+    // png_read_filter_row_mmx_paeth() constants:
+    0x00000000FFFFFFFFLL, // _amask4_4_0   offset 176  (bpp 6/4/8, paeth)
+    0xFFFF000000000000LL, // _amask0_2_6   offset 184  (bpp 3, paeth)   A.M.End
+
+    // png_read_filter_row_mmx_sub() constants:
+    0x0000FFFFFF000000LL, // _amask2_3_3   offset 192  (bpp 3, sub)
+    0x00000000FFFF0000LL, // _amask4_2_2   offset 200  (bpp 2, sub)
+
+};
+
+#define MASK8_0        "(%%rbp)"
+#define MASK16_0       "8(%%rbp)"
+#define MASK16_1       "16(%%rbp)"
+#define MASK24_0       "24(%%rbp)"
+#define MASK24_1       "32(%%rbp)"
+#define MASK24_2       "40(%%rbp)"
+#define MASK32_0       "48(%%rbp)"
+#define MASK32_1       "56(%%rbp)"
+#define MASK32_2       "64(%%rbp)"
+#define MASK32_3       "72(%%rbp)"
+#define MASK48_0       "80(%%rbp)"
+#define MASK48_1       "88(%%rbp)"
+#define MASK48_2       "96(%%rbp)"
+#define MASK48_3       "104(%%rbp)"
+#define MASK48_4       "112(%%rbp)"
+#define MASK48_5       "120(%%rbp)"
+#define AMASK5_3_0     "128(%%rbp)"
+#define AMASK7_1_0     "136(%%rbp)"
+#define LB_CARRY_MASK  "144(%%rbp)"
+#define HB_CLEAR_MASK  "152(%%rbp)"
+#define AMASK0_8_0     "160(%%rbp)"
+#define AMASK6_2_0     "168(%%rbp)"
+#define AMASK4_4_0     "176(%%rbp)"
+#define AMASK0_2_6     "184(%%rbp)"
+#define AMASK2_3_3     "192(%%rbp)"
+#define AMASK4_2_2     "200(%%rbp)"
+
+#else // !PNG_x86_64_USE_GOTPCREL
+
+static PNG_CONST ull _mask8_0  __attribute__((used, aligned(8))) = 0x0102040810204080LL;
+
+static PNG_CONST ull _mask16_1 __attribute__((used, aligned(8))) = 0x0101020204040808LL;
+static PNG_CONST ull _mask16_0 __attribute__((used, aligned(8))) = 0x1010202040408080LL;
+
+static PNG_CONST ull _mask24_2 __attribute__((used, aligned(8))) = 0x0101010202020404LL;
+static PNG_CONST ull _mask24_1 __attribute__((used, aligned(8))) = 0x0408080810101020LL;
+static PNG_CONST ull _mask24_0 __attribute__((used, aligned(8))) = 0x2020404040808080LL;
+
+static PNG_CONST ull _mask32_3 __attribute__((used, aligned(8))) = 0x0101010102020202LL;
+static PNG_CONST ull _mask32_2 __attribute__((used, aligned(8))) = 0x0404040408080808LL;
+static PNG_CONST ull _mask32_1 __attribute__((used, aligned(8))) = 0x1010101020202020LL;
+static PNG_CONST ull _mask32_0 __attribute__((used, aligned(8))) = 0x4040404080808080LL;
+
+static PNG_CONST ull _mask48_5 __attribute__((used, aligned(8))) = 0x0101010101010202LL;
+static PNG_CONST ull _mask48_4 __attribute__((used, aligned(8))) = 0x0202020204040404LL;
+static PNG_CONST ull _mask48_3 __attribute__((used, aligned(8))) = 0x0404080808080808LL;
+static PNG_CONST ull _mask48_2 __attribute__((used, aligned(8))) = 0x1010101010102020LL;
+static PNG_CONST ull _mask48_1 __attribute__((used, aligned(8))) = 0x2020202040404040LL;
+static PNG_CONST ull _mask48_0 __attribute__((used, aligned(8))) = 0x4040808080808080LL;
+
+// png_do_read_interlace() constants:
+static PNG_CONST ull _amask5_3_0  __attribute__((aligned(8))) = 0x0000000000FFFFFFLL;  // was _const4
+static PNG_CONST ull _amask7_1_0  __attribute__((aligned(8))) = 0x00000000000000FFLL;  // was _const6
+
+// png_read_filter_row_mmx_avg() constants:
+static PNG_CONST ull _LBCarryMask __attribute__((used, aligned(8))) = 0x0101010101010101LL;
+static PNG_CONST ull _HBClearMask __attribute__((used, aligned(8))) = 0x7f7f7f7f7f7f7f7fLL;
+static PNG_CONST ull _amask0_8_0  __attribute__((used, aligned(8))) = 0xFFFFFFFFFFFFFFFFLL;
+static PNG_CONST ull _amask6_2_0  __attribute__((used, aligned(8))) = 0x000000000000FFFFLL;
+
+// png_read_filter_row_mmx_paeth() constants:
+static PNG_CONST ull _amask4_4_0  __attribute__((used, aligned(8))) = 0x00000000FFFFFFFFLL;
+static PNG_CONST ull _amask0_2_6  __attribute__((used, aligned(8))) = 0xFFFF000000000000LL;
+
+// png_read_filter_row_mmx_sub() constants:
+static PNG_CONST ull _amask2_3_3  __attribute__((used, aligned(8))) = 0x0000FFFFFF000000LL;
+static PNG_CONST ull _amask4_2_2  __attribute__((used, aligned(8))) = 0x00000000FFFF0000LL;
+
+#define MASK8_0        "_mask8_0"
+#define MASK16_0       "_mask16_0"
+#define MASK16_1       "_mask16_1"
+#define MASK24_0       "_mask24_0"
+#define MASK24_1       "_mask24_1"
+#define MASK24_2       "_mask24_2"
+#define MASK32_0       "_mask32_0"
+#define MASK32_1       "_mask32_1"
+#define MASK32_2       "_mask32_2"
+#define MASK32_3       "_mask32_3"
+#define MASK48_0       "_mask48_0"
+#define MASK48_1       "_mask48_1"
+#define MASK48_2       "_mask48_2"
+#define MASK48_3       "_mask48_3"
+#define MASK48_4       "_mask48_4"
+#define MASK48_5       "_mask48_5"
+#define AMASK5_3_0     "_amask5_3_0"
+#define AMASK7_1_0     "_amask7_1_0"
+#define LB_CARRY_MASK  "_LBCarryMask"
+#define HB_CLEAR_MASK  "_HBClearMask"
+#define AMASK0_8_0     "_amask0_8_0"
+#define AMASK6_2_0     "_amask6_2_0"
+#define AMASK4_4_0     "_amask4_4_0"
+#define AMASK0_2_6     "_amask0_2_6"
+#define AMASK2_3_3     "_amask2_3_3"
+#define AMASK4_2_2     "_amask4_2_2"
+
+#endif // ?PNG_x86_64_USE_GOTPCREL
+
+
+#if defined(PNG_HAVE_MMX_READ_FILTER_ROW) || defined(PNG_HAVE_MMX_COMBINE_ROW)
+
+// this block is specific to png_read_filter_row_mmx_paeth() except for
+// LOAD_GOT_rbp and RESTORE_rbp, which are also used in png_combine_row()
+#if defined(PNG_x86_64_USE_GOTPCREL)
+#  define pa_TEMP                "%%r11d"
+#  define pb_TEMP                "%%r12d"
+#  define pc_TEMP                "%%r13d"
+#  if defined(PNG_CLOBBER_x86_64_REGS_SUPPORTED)  // works as of gcc 3.4.3 ...
+#    define SAVE_r11_r12_r13
+#    define RESTORE_r11_r12_r13
+#    define _CLOBBER_r11_r12_r13 ,"%r11", "%r12", "%r13"
+#    define CLOBBER_r11_r12_r13  "%r11", "%r12", "%r13"
+#  else // !PNG_CLOBBER_x86_64_REGS_SUPPORTED
+#    define SAVE_r11_r12_r13     "pushq %%r11  \n\t" \
+                                 "pushq %%r12  \n\t" \
+                                 "pushq %%r13  \n\t"  // "normally 0-extended"
+#    define RESTORE_r11_r12_r13  "popq  %%r13  \n\t" \
+                                 "popq  %%r12  \n\t" \
+                                 "popq  %%r11  \n\t"
+#    define _CLOBBER_r11_r12_r13
+#    define CLOBBER_r11_r12_r13
+#  endif
+#  define LOAD_GOT_rbp           "pushq %%rbp                        \n\t" \
+                                 "movq  _c64@GOTPCREL(%%rip), %%rbp  \n\t"
+#  define RESTORE_rbp            "popq  %%rbp                        \n\t"
+#else // 32-bit and/or non-PIC
+#  if defined(PNG_THREAD_UNSAFE_OK)
+     // These variables are used in png_read_filter_row_mmx_paeth() and would be
+     //   local variables if not for gcc-inline-assembly addressing limitations
+     //   (some apparently related to ELF format, others to CPU type).
+     //
+     // WARNING: Their presence defeats the thread-safety of libpng.
+     static int                     _patemp  __attribute__((used));
+     static int                     _pbtemp  __attribute__((used));
+     static int                     _pctemp  __attribute__((used));
+#    define pa_TEMP                "_patemp"
+#    define pb_TEMP                "_pbtemp"  // temp variables for
+#    define pc_TEMP                "_pctemp"  //  Paeth routine
+#    define SAVE_r11_r12_r13
+#    define RESTORE_r11_r12_r13
+#    define _CLOBBER_r11_r12_r13   // not using regs => not clobbering
+#    define CLOBBER_r11_r12_r13
+#  endif // PNG_THREAD_UNSAFE_OK
+#  define LOAD_GOT_rbp
+#  define RESTORE_rbp
+#endif
+
+#if defined(__x86_64__)
+#  define SAVE_ebp
+#  define RESTORE_ebp
+#  define _CLOBBER_ebp         ,"%ebp"
+#  define CLOBBER_ebp          "%ebp"
+#  define SAVE_FullLength      "movl %%eax, %%r15d  \n\t"
+#  define RESTORE_FullLength   "movl %%r15d, "     // may go into eax or ecx
+#  if defined(PNG_CLOBBER_x86_64_REGS_SUPPORTED)   // works as of gcc 3.4.3 ...
+#    define SAVE_r15
+#    define RESTORE_r15
+#    define _CLOBBER_r15       ,"%r15"
+#  else
+#    define SAVE_r15           "pushq %%r15  \n\t"
+#    define RESTORE_r15        "popq  %%r15  \n\t"
+#    define _CLOBBER_r15
+#  endif
+#  define PBP                  "%%rbp"             // regs used for 64-bit
+#  define PAX                  "%%rax"             //  pointers or in
+#  define PBX                  "%%rbx"             //  combination with
+#  define PCX                  "%%rcx"             //  64-bit pointer-regs
+#  define PDX                  "%%rdx"             //  (base/index pairs,
+#  define PSI                  "%%rsi"             //  add/sub/mov pairs)
+#  define CLEAR_BOTTOM_3_BITS  "and  $0xfffffffffffffff8, "
+#else
+#  define SAVE_ebp             "pushl %%ebp \n\t"  // clobber list doesn't work
+#  define RESTORE_ebp          "popl  %%ebp \n\t"  //  for %ebp on 32-bit; not
+#  define _CLOBBER_ebp                             //  clear why not
+#  define CLOBBER_ebp
+#  define SAVE_FullLength      "pushl %%eax \n\t"
+#  define RESTORE_FullLength   "popl "             // eax (avg) or ecx (paeth)
+#  define SAVE_r15
+#  define RESTORE_r15
+#  define _CLOBBER_r15
+#  define PBP                  "%%ebp"             // regs used for or in
+#  define PAX                  "%%eax"             //  combination with
+#  define PBX                  "%%ebx"             //  "normal," 32-bit
+#  define PCX                  "%%ecx"             //  pointers
+#  define PDX                  "%%edx"
+#  define PSI                  "%%esi"
+#  define CLEAR_BOTTOM_3_BITS  "and  $0xfffffff8, "
+#endif
+
+// CLOB_COMMA_ebx_ebp:  need comma ONLY if both CLOBBER_ebp and CLOBBER_GOT_ebx
+//                      have values, i.e., only if __x86_64__ AND !__PIC__
+#if defined(__x86_64__) && !defined(__PIC__)
+#  define CLOB_COMMA_ebx_ebp    , // clobbering both ebp and ebx => need comma
+#else
+#  define CLOB_COMMA_ebx_ebp
+#endif
+
+// CLOB_COMMA_ebX_r1X:  need comma UNLESS both CLOBBER_ebp and CLOBBER_GOT_ebx
+//                   are empty OR CLOBBER_r11_r12_r13 is empty--i.e., NO comma
+//                   if (!__x86_64__ AND __PIC__) OR !(PNG_x86_64_USE_GOTPCREL
+//                   AND PNG_CLOBBER_x86_64_REGS_SUPPORTED)   (double sigh...)
+#if (!defined(__x86_64__) && defined(__PIC__)) || \
+    !defined(PNG_x86_64_USE_GOTPCREL) || \
+    !defined(PNG_CLOBBER_x86_64_REGS_SUPPORTED)
+#  define CLOB_COMMA_ebX_r1X
+#else
+#  define CLOB_COMMA_ebX_r1X    , // clobbering (ebp OR ebx) AND r11_r12_r13
+#endif
+
+// CLOB_COLON_ebx_ebp:  need colon unless CLOBBER_ebp and CLOBBER_GOT_ebx are
+//                      BOTH empty--i.e., NO colon if (!__x86_64__ AND __PIC__)
+// CLOB_COLON_ebx_ebp_r1X:  if, in addition, CLOBBER_r11_r12_r13 is empty, then
+//                          no colon for Paeth blocks, either--i.e., NO colon
+//                          if !(PNG_x86_64_USE_GOTPCREL AND
+//                               PNG_CLOBBER_x86_64_REGS_SUPPORTED)
+#if (!defined(__x86_64__) && defined(__PIC__))
+#  define CLOB_COLON_ebx_ebp
+#  if !(defined(PNG_x86_64_USE_GOTPCREL) && \
+        defined(PNG_CLOBBER_x86_64_REGS_SUPPORTED))
+#    define CLOB_COLON_ebx_ebp_r1X
+#  else
+#    define CLOB_COLON_ebx_ebp_r1X  : // clobbering ebp OR ebx OR r11_r12_r13
+#  endif
+#else
+#  define CLOB_COLON_ebx_ebp        : // clobbering ebp OR ebx
+#  define CLOB_COLON_ebx_ebp_r1X    : // clobbering ebp OR ebx OR r11_r12_r13
+#endif
+
+#endif // PNG_HAVE_MMX_READ_FILTER_ROW
+
+#if defined(__PIC__)  // macros to save, restore index to Global Offset Table
+#  if defined(__x86_64__)
+#    define SAVE_GOT_ebx     "pushq %%rbx \n\t"
+#    define RESTORE_GOT_ebx  "popq  %%rbx \n\t"
+#  else
+#    define SAVE_GOT_ebx     "pushl %%ebx \n\t"
+#    define RESTORE_GOT_ebx  "popl  %%ebx \n\t"
+#  endif
+#  define _CLOBBER_GOT_ebx   // explicitly saved, restored => not clobbered
+#  define CLOBBER_GOT_ebx
+#else
+#  define SAVE_GOT_ebx
+#  define RESTORE_GOT_ebx
+#  define _CLOBBER_GOT_ebx   ,"%ebx"
+#  define CLOBBER_GOT_ebx    "%ebx"
+#endif
+
+#if defined(PNG_HAVE_MMX_COMBINE_ROW) || defined(PNG_HAVE_MMX_READ_INTERLACE)
+#  define BPP2  2
+#  define BPP3  3  // bytes per pixel (a.k.a. pixel_bytes)
+#  define BPP4  4  // (defined only to help avoid cut-and-paste errors)
+#  define BPP6  6
+#  define BPP8  8
+#endif
+
+
+
+static int _mmx_supported = 2; // 0: no MMX; 1: MMX supported; 2: not tested
+
+/*===========================================================================*/
+/*                                                                           */
+/*                      P N G _ M M X _ S U P P O R T                        */
+/*                                                                           */
+/*===========================================================================*/
+
+// GRR NOTES:  (1) the following code assumes 386 or better (pushfl/popfl)
+//             (2) all instructions compile with gcc 2.7.2.3 and later
+//           x (3) the function is moved down here to prevent gcc from
+//           x      inlining it in multiple places and then barfing be-
+//           x      cause the ".NOT_SUPPORTED" label is multiply defined
+//                  [need to retest with gcc 2.7.2.3]
+
+// GRR 20070524:  This declaration apparently is compatible with but supersedes
+//   the one in png.h; in any case, the generated object file is slightly
+//   smaller.  It is unnecessary with gcc 4.1.2, but gcc 2.x apparently
+//   replicated the ".NOT_SUPPORTED" label in each location the function was
+//   inlined, leading to compilation errors due to the "multiply defined"
+//   label.  Old workaround was to leave the function at the end of this
+//   file; new one (still testing) is to use a gcc-specific function attribute
+//   to prevent local inlining.
+int PNGAPI
+png_mmx_support(void) __attribute__((noinline));
+
+int PNGAPI
+png_mmx_support(void)
+{
+#if defined(PNG_MMX_CODE_SUPPORTED)  // superfluous, but what the heck
+    int result;
+    __asm__ __volatile__ (
+#if defined(__x86_64__)
+        "pushq %%rbx          \n\t"  // rbx gets clobbered by CPUID instruction
+        "pushq %%rcx          \n\t"  // so does rcx...
+        "pushq %%rdx          \n\t"  // ...and rdx (but rcx & rdx safe on Linux)
+        "pushfq               \n\t"  // save Eflag to stack
+        "popq %%rax           \n\t"  // get Eflag from stack into rax
+        "movq %%rax, %%rcx    \n\t"  // make another copy of Eflag in rcx
+        "xorl $0x200000, %%eax \n\t" // toggle ID bit in Eflag (i.e., bit 21)
+        "pushq %%rax          \n\t"  // save modified Eflag back to stack
+        "popfq                \n\t"  // restore modified value to Eflag reg
+        "pushfq               \n\t"  // save Eflag to stack
+        "popq %%rax           \n\t"  // get Eflag from stack
+        "pushq %%rcx          \n\t"  // save original Eflag to stack
+        "popfq                \n\t"  // restore original Eflag
+#else
+        "pushl %%ebx          \n\t"  // ebx gets clobbered by CPUID instruction
+        "pushl %%ecx          \n\t"  // so does ecx...
+        "pushl %%edx          \n\t"  // ...and edx (but ecx & edx safe on Linux)
+        "pushfl               \n\t"  // save Eflag to stack
+        "popl %%eax           \n\t"  // get Eflag from stack into eax
+        "movl %%eax, %%ecx    \n\t"  // make another copy of Eflag in ecx
+        "xorl $0x200000, %%eax \n\t" // toggle ID bit in Eflag (i.e., bit 21)
+        "pushl %%eax          \n\t"  // save modified Eflag back to stack
+        "popfl                \n\t"  // restore modified value to Eflag reg
+        "pushfl               \n\t"  // save Eflag to stack
+        "popl %%eax           \n\t"  // get Eflag from stack
+        "pushl %%ecx          \n\t"  // save original Eflag to stack
+        "popfl                \n\t"  // restore original Eflag
+#endif
+        "xorl %%ecx, %%eax    \n\t"  // compare new Eflag with original Eflag
+        "jz 0f                \n\t"  // if same, CPUID instr. is not supported
+
+        "xorl %%eax, %%eax    \n\t"  // set eax to zero
+//      ".byte  0x0f, 0xa2    \n\t"  // CPUID instruction (two-byte opcode)
+        "cpuid                \n\t"  // get the CPU identification info
+        "cmpl $1, %%eax       \n\t"  // make sure eax return non-zero value
+        "jl 0f                \n\t"  // if eax is zero, MMX is not supported
+
+        "xorl %%eax, %%eax    \n\t"  // set eax to zero and...
+        "incl %%eax           \n\t"  // ...increment eax to 1.  This pair is
+                                     // faster than the instruction "mov eax, 1"
+        "cpuid                \n\t"  // get the CPU identification info again
+        "andl $0x800000, %%edx \n\t" // mask out all bits but MMX bit (23)
+        "cmpl $0, %%edx       \n\t"  // 0 = MMX not supported
+        "jz 0f                \n\t"  // non-zero = yes, MMX IS supported
+
+        "movl $1, %%eax       \n\t"  // set return value to 1
+        "jmp  1f              \n\t"  // DONE:  have MMX support
+
+    "0:                       \n\t"  // .NOT_SUPPORTED: target label for jump instructions
+        "movl $0, %%eax       \n\t"  // set return value to 0
+    "1:                       \n\t"  // .RETURN: target label for jump instructions
+#if defined(__x86_64__)
+        "popq %%rdx           \n\t"  // restore rdx
+        "popq %%rcx           \n\t"  // restore rcx
+        "popq %%rbx           \n\t"  // restore rbx
+#else
+        "popl %%edx           \n\t"  // restore edx
+        "popl %%ecx           \n\t"  // restore ecx
+        "popl %%ebx           \n\t"  // restore ebx
+#endif
+
+//      "ret                  \n\t"  // DONE:  no MMX support
+                                     // (fall through to standard C "ret")
+
+        : "=a" (result)              // output list
+
+        :                            // any variables used on input (none)
+
+                                     // no clobber list
+//      , "%ebx", "%ecx", "%edx"     // GRR:  we handle these manually
+//      , "memory"   // if write to a variable gcc thought was in a reg
+//      , "cc"       // "condition codes" (flag bits)
+    );
+    _mmx_supported = result;
+#else
+    _mmx_supported = 0;
+#endif /* PNG_MMX_CODE_SUPPORTED */
+
+    return _mmx_supported;
+}
+
+
+/*===========================================================================*/
+/*                                                                           */
+/*                       P N G _ C O M B I N E _ R O W                       */
+/*                                                                           */
+/*===========================================================================*/
+
+#if defined(PNG_HAVE_MMX_COMBINE_ROW)
+
+/* Combines the row recently read in with the previous row.
+   This routine takes care of alpha and transparency if requested.
+   This routine also handles the two methods of progressive display
+   of interlaced images, depending on the mask value.
+   The mask value describes which pixels are to be combined with
+   the row.  The pattern always repeats every 8 pixels, so just 8
+   bits are needed.  A one indicates the pixel is to be combined; a
+   zero indicates the pixel is to be skipped.  This is in addition
+   to any alpha or transparency value associated with the pixel.
+   If you want all pixels to be combined, pass 0xff (255) in mask. */
+
+/* Use this routine for the x86 platform - it uses a faster MMX routine
+   if the machine supports MMX. */
+
+void /* PRIVATE */
+png_combine_row(png_structp png_ptr, png_bytep row, int mask)
+{
+   int dummy_value_a;    // fix 'forbidden register spilled' error
+   int dummy_value_c;
+   int dummy_value_d;
+   png_bytep dummy_value_S;
+   png_bytep dummy_value_D;
+
+   png_debug(1, "in png_combine_row (pnggccrd.c)\n");
+
+   if (_mmx_supported == 2) {
+#if !defined(PNG_1_0_X)
+       /* this should have happened in png_init_mmx_flags() already */
+       png_warning(png_ptr, "asm_flags may not have been initialized");
+#endif
+       png_mmx_support();
+   }
+
+   if (mask == 0xff)
+   {
+      png_debug(2,"mask == 0xff:  doing single png_memcpy()\n");
+      png_memcpy(row, png_ptr->row_buf + 1,
+       (png_size_t)PNG_ROWBYTES(png_ptr->row_info.pixel_depth,png_ptr->width));
+   }
+   else   /* (png_combine_row() is never called with mask == 0) */
+   {
+      switch (png_ptr->row_info.pixel_depth)
+      {
+         case 24:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+
+#if !defined(PNG_1_0_X)
+            if (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+#else
+            if (_mmx_supported)
+#endif
+            {
+               png_uint_32 len;
+               int diff;
+
+               srcptr = png_ptr->row_buf + 1;
+               dstptr = row;
+               len  = png_ptr->width & ~7;          // reduce to multiple of 8
+               diff = (int) (png_ptr->width & 7);   // amount lost
+
+               __asm__ __volatile__ (
+                  "not       %%edx            \n\t" // mask => unmask
+                  "movd      %%edx, %%mm7     \n\t" // load bit pattern
+                  "not       %%edx            \n\t" // unmask => mask for later
+                  "psubb     %%mm6, %%mm6     \n\t" // zero mm6
+                  "punpcklbw %%mm7, %%mm7     \n\t"
+                  "punpcklwd %%mm7, %%mm7     \n\t"
+                  "punpckldq %%mm7, %%mm7     \n\t" // fill reg with 8 masks
+
+                  LOAD_GOT_rbp
+                  "movq   " MASK24_0 ", %%mm0 \n\t" // _mask24_0 -> mm0
+                  "movq   " MASK24_1 ", %%mm1 \n\t" // _mask24_1 -> mm1
+                  "movq   " MASK24_2 ", %%mm2 \n\t" // _mask24_2 -> mm2
+                  RESTORE_rbp
+
+                  "pand      %%mm7, %%mm0     \n\t"
+                  "pand      %%mm7, %%mm1     \n\t"
+                  "pand      %%mm7, %%mm2     \n\t"
+
+                  "pcmpeqb   %%mm6, %%mm0     \n\t"
+                  "pcmpeqb   %%mm6, %%mm1     \n\t"
+                  "pcmpeqb   %%mm6, %%mm2     \n\t"
+
+// preload        "movl      len, %%ecx       \n\t" // load length of line
+// preload        "movl      srcptr, %3       \n\t" // load source
+// preload        "movl      dstptr, %4       \n\t" // load dest
+
+                  "cmpl      $0, %%ecx        \n\t"
+                  "jz        mainloop24end    \n\t"
+
+                "mainloop24:                  \n\t"
+                  "movq      (%3), %%mm4      \n\t"
+                  "pand      %%mm0, %%mm4     \n\t"
+                  "movq      %%mm0, %%mm6     \n\t"
+                  "movq      (%4), %%mm7      \n\t"
+                  "pandn     %%mm7, %%mm6     \n\t"
+                  "por       %%mm6, %%mm4     \n\t"
+                  "movq      %%mm4, (%4)      \n\t"
+
+                  "movq      8(%3), %%mm5     \n\t"
+                  "pand      %%mm1, %%mm5     \n\t"
+                  "movq      %%mm1, %%mm7     \n\t"
+                  "movq      8(%4), %%mm6     \n\t"
+                  "pandn     %%mm6, %%mm7     \n\t"
+                  "por       %%mm7, %%mm5     \n\t"
+                  "movq      %%mm5, 8(%4)     \n\t"
+
+                  "movq      16(%3), %%mm6    \n\t"
+                  "pand      %%mm2, %%mm6     \n\t"
+                  "movq      %%mm2, %%mm4     \n\t"
+                  "movq      16(%4), %%mm7    \n\t"
+                  "pandn     %%mm7, %%mm4     \n\t"
+                  "por       %%mm4, %%mm6     \n\t"
+                  "movq      %%mm6, 16(%4)    \n\t"
+
+                  "add       $24, %3          \n\t" // inc by 24 bytes processed
+                  "add       $24, %4          \n\t"
+                  "subl      $8, %%ecx        \n\t" // dec by 8 pixels processed
+
+                  "ja        mainloop24       \n\t"
+
+                "mainloop24end:               \n\t"
+// preload        "movl      diff, %%ecx      \n\t" // (diff is in eax)
+                  "movl      %%eax, %%ecx     \n\t"
+                  "cmpl      $0, %%ecx        \n\t"
+                  "jz        end24            \n\t"
+// preload        "movl      mask, %%edx      \n\t"
+                  "sall      $24, %%edx       \n\t" // make low byte, high byte
+
+                "secondloop24:                \n\t"
+                  "sall      %%edx            \n\t" // move high bit to CF
+                  "jnc       skip24           \n\t" // if CF = 0
+                  "movw      (%3), %%ax       \n\t"
+                  "movw      %%ax, (%4)       \n\t"
+                  "xorl      %%eax, %%eax     \n\t"
+                  "movb      2(%3), %%al      \n\t"
+                  "movb      %%al, 2(%4)      \n\t"
+
+                "skip24:                      \n\t"
+                  "add       $3, %3           \n\t"
+                  "add       $3, %4           \n\t"
+                  "decl      %%ecx            \n\t"
+                  "jnz       secondloop24     \n\t"
+
+                "end24:                       \n\t"
+                  "EMMS                       \n\t" // DONE
+
+                  : "=a" (dummy_value_a),           // output regs (dummy)
+                    "=d" (dummy_value_d),
+                    "=c" (dummy_value_c),
+                    "=S" (dummy_value_S),
+                    "=D" (dummy_value_D)
+
+                  : "0" (diff),        // eax       // input regs
+                    "1" (mask),        // edx
+                    "2" (len),         // ecx
+// was (unmask)     "b"    RESERVED    // ebx       // Global Offset Table idx
+                    "3" (srcptr),      // esi/rsi
+                    "4" (dstptr)       // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                  : "%mm0", "%mm1", "%mm2"          // clobber list
+                  , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+               );
+            }
+            else /* not _mmx_supported - use modified C routine */
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = BPP3 * png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = BPP3 * png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = BPP3 * png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7;  /* reduce to mult. of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = BPP3 * len;   /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff*BPP3;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+            } /* end of else (_mmx_supported) */
+
+            break;
+         }       /* end 24 bpp */
+
+         // formerly claimed to be most common case (combining 32-bit RGBA),
+         // but almost certainly less common than 24-bit RGB case
+         case 32:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+
+#if !defined(PNG_1_0_X)
+            if (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+#else
+            if (_mmx_supported)
+#endif
+            {
+               png_uint_32 len;
+               int diff;
+
+               srcptr = png_ptr->row_buf + 1;
+               dstptr = row;
+               len  = png_ptr->width & ~7;          // reduce to multiple of 8
+               diff = (int) (png_ptr->width & 7);   // amount lost
+
+               __asm__ __volatile__ (
+                  "not       %%edx            \n\t" // mask => unmask
+                  "movd      %%edx, %%mm7     \n\t" // load bit pattern
+                  "not       %%edx            \n\t" // unmask => mask for later
+                  "psubb     %%mm6, %%mm6     \n\t" // zero mm6
+                  "punpcklbw %%mm7, %%mm7     \n\t"
+                  "punpcklwd %%mm7, %%mm7     \n\t"
+                  "punpckldq %%mm7, %%mm7     \n\t" // fill reg with 8 masks
+
+                  LOAD_GOT_rbp
+                  "movq   " MASK32_0 ", %%mm0 \n\t" // _mask32_0
+                  "movq   " MASK32_1 ", %%mm1 \n\t" // _mask32_1
+                  "movq   " MASK32_2 ", %%mm2 \n\t" // _mask32_2
+                  "movq   " MASK32_3 ", %%mm3 \n\t" // _mask32_3
+                  RESTORE_rbp
+
+                  "pand      %%mm7, %%mm0     \n\t"
+                  "pand      %%mm7, %%mm1     \n\t"
+                  "pand      %%mm7, %%mm2     \n\t"
+                  "pand      %%mm7, %%mm3     \n\t"
+
+                  "pcmpeqb   %%mm6, %%mm0     \n\t"
+                  "pcmpeqb   %%mm6, %%mm1     \n\t"
+                  "pcmpeqb   %%mm6, %%mm2     \n\t"
+                  "pcmpeqb   %%mm6, %%mm3     \n\t"
+
+// preload        "movl      len, %%ecx       \n\t" // load length of line
+// preload        "movl      srcptr, %3       \n\t" // load source
+// preload        "movl      dstptr, %4       \n\t" // load dest
+
+                  "cmpl      $0, %%ecx        \n\t" // lcr
+                  "jz        mainloop32end    \n\t"
+
+                "mainloop32:                  \n\t"
+                  "movq      (%3), %%mm4      \n\t"
+                  "pand      %%mm0, %%mm4     \n\t"
+                  "movq      %%mm0, %%mm6     \n\t"
+                  "movq      (%4), %%mm7      \n\t"
+                  "pandn     %%mm7, %%mm6     \n\t"
+                  "por       %%mm6, %%mm4     \n\t"
+                  "movq      %%mm4, (%4)      \n\t"
+
+                  "movq      8(%3), %%mm5     \n\t"
+                  "pand      %%mm1, %%mm5     \n\t"
+                  "movq      %%mm1, %%mm7     \n\t"
+                  "movq      8(%4), %%mm6     \n\t"
+                  "pandn     %%mm6, %%mm7     \n\t"
+                  "por       %%mm7, %%mm5     \n\t"
+                  "movq      %%mm5, 8(%4)     \n\t"
+
+                  "movq      16(%3), %%mm6    \n\t"
+                  "pand      %%mm2, %%mm6     \n\t"
+                  "movq      %%mm2, %%mm4     \n\t"
+                  "movq      16(%4), %%mm7    \n\t"
+                  "pandn     %%mm7, %%mm4     \n\t"
+                  "por       %%mm4, %%mm6     \n\t"
+                  "movq      %%mm6, 16(%4)    \n\t"
+
+                  "movq      24(%3), %%mm7    \n\t"
+                  "pand      %%mm3, %%mm7     \n\t"
+                  "movq      %%mm3, %%mm5     \n\t"
+                  "movq      24(%4), %%mm4    \n\t"
+                  "pandn     %%mm4, %%mm5     \n\t"
+                  "por       %%mm5, %%mm7     \n\t"
+                  "movq      %%mm7, 24(%4)    \n\t"
+
+                  "add       $32, %3          \n\t" // inc by 32 bytes processed
+                  "add       $32, %4          \n\t"
+                  "subl      $8, %%ecx        \n\t" // dec by 8 pixels processed
+                  "ja        mainloop32       \n\t"
+
+                "mainloop32end:               \n\t"
+// preload        "movl      diff, %%ecx      \n\t" // (diff is in eax)
+                  "movl      %%eax, %%ecx     \n\t"
+                  "cmpl      $0, %%ecx        \n\t"
+                  "jz        end32            \n\t"
+// preload        "movl      mask, %%edx      \n\t"
+                  "sall      $24, %%edx       \n\t" // low byte => high byte
+
+                "secondloop32:                \n\t"
+                  "sall      %%edx            \n\t" // move high bit to CF
+                  "jnc       skip32           \n\t" // if CF = 0
+                  "movl      (%3), %%eax      \n\t"
+                  "movl      %%eax, (%4)      \n\t"
+
+                "skip32:                      \n\t"
+                  "add       $4, %3           \n\t"
+                  "add       $4, %4           \n\t"
+                  "decl      %%ecx            \n\t"
+                  "jnz       secondloop32     \n\t"
+
+                "end32:                       \n\t"
+                  "EMMS                       \n\t" // DONE
+
+                  : "=a" (dummy_value_a),           // output regs (dummy)
+                    "=d" (dummy_value_d),
+                    "=c" (dummy_value_c),
+                    "=S" (dummy_value_S),
+                    "=D" (dummy_value_D)
+
+                  : "0" (diff),        // eax       // input regs
+                    "1" (mask),        // edx
+                    "2" (len),         // ecx
+// was (unmask)     "b"    RESERVED    // ebx       // Global Offset Table idx
+                    "3" (srcptr),      // esi/rsi
+                    "4" (dstptr)       // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                  : "%mm0", "%mm1", "%mm2", "%mm3"  // clobber list
+                  , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+               );
+            }
+            else /* not _mmx_supported - use modified C routine */
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = BPP4 * png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = BPP4 * png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = BPP4 * png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7;  /* reduce to mult. of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = BPP4 * len;   /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff*BPP4;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+            } /* end of else (_mmx_supported) */
+
+            break;
+         }       /* end 32 bpp */
+
+         case 8:        /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+
+#if !defined(PNG_1_0_X)
+            if (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+#else
+            if (_mmx_supported)
+#endif
+            {
+               png_uint_32 len;
+               int diff;
+
+               srcptr = png_ptr->row_buf + 1;
+               dstptr = row;
+               len  = png_ptr->width & ~7;          // reduce to multiple of 8
+               diff = (int) (png_ptr->width & 7);   // amount lost
+
+               __asm__ __volatile__ (
+                  "not       %%edx            \n\t" // mask => unmask
+                  "movd      %%edx, %%mm7     \n\t" // load bit pattern
+                  "not       %%edx            \n\t" // unmask => mask for later
+                  "psubb     %%mm6, %%mm6     \n\t" // zero mm6
+                  "punpcklbw %%mm7, %%mm7     \n\t"
+                  "punpcklwd %%mm7, %%mm7     \n\t"
+                  "punpckldq %%mm7, %%mm7     \n\t" // fill reg with 8 masks
+
+                  LOAD_GOT_rbp
+                  "movq   " MASK8_0 ", %%mm0  \n\t" // _mask8_0 -> mm0
+                  RESTORE_rbp
+
+                  "pand      %%mm7, %%mm0     \n\t" // nonzero if keep byte
+                  "pcmpeqb   %%mm6, %%mm0     \n\t" // zeros->1s, v versa
+
+// preload        "movl      len, %%ecx       \n\t" // load length of line
+// preload        "movl      srcptr, %3       \n\t" // load source
+// preload        "movl      dstptr, %4       \n\t" // load dest
+
+                  "cmpl      $0, %%ecx        \n\t" // len == 0 ?
+                  "je        mainloop8end     \n\t"
+
+                "mainloop8:                   \n\t"
+                  "movq      (%3), %%mm4      \n\t" // *srcptr
+                  "pand      %%mm0, %%mm4     \n\t"
+                  "movq      %%mm0, %%mm6     \n\t"
+                  "pandn     (%4), %%mm6      \n\t" // *dstptr
+                  "por       %%mm6, %%mm4     \n\t"
+                  "movq      %%mm4, (%4)      \n\t"
+                  "add       $8, %3           \n\t" // inc by 8 bytes processed
+                  "add       $8, %4           \n\t"
+                  "subl      $8, %%ecx        \n\t" // dec by 8 pixels processed
+                  "ja        mainloop8        \n\t"
+
+                "mainloop8end:                \n\t"
+// preload        "movl      diff, %%ecx      \n\t" // (diff is in eax)
+                  "movl      %%eax, %%ecx     \n\t"
+                  "cmpl      $0, %%ecx        \n\t"
+                  "jz        end8             \n\t"
+// preload        "movl      mask, %%edx      \n\t"
+                  "sall      $24, %%edx       \n\t" // make low byte, high byte
+
+                "secondloop8:                 \n\t"
+                  "sall      %%edx            \n\t" // move high bit to CF
+                  "jnc       skip8            \n\t" // if CF = 0
+                  "movb      (%3), %%al       \n\t"
+                  "movb      %%al, (%4)       \n\t"
+
+                "skip8:                       \n\t"
+                  "inc       %3               \n\t"
+                  "inc       %4               \n\t"
+                  "decl      %%ecx            \n\t"
+                  "jnz       secondloop8      \n\t"
+
+                "end8:                        \n\t"
+                  "EMMS                       \n\t" // DONE
+
+                  : "=a" (dummy_value_a),           // output regs (dummy)
+                    "=d" (dummy_value_d),
+                    "=c" (dummy_value_c),
+                    "=S" (dummy_value_S),
+                    "=D" (dummy_value_D)
+
+                  : "0" (diff),        // eax       // input regs
+                    "1" (mask),        // edx
+                    "2" (len),         // ecx
+// was (unmask)     "b"    RESERVED    // ebx       // Global Offset Table idx
+                    "3" (srcptr),      // esi/rsi
+                    "4" (dstptr)       // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                  : "%mm0", "%mm4", "%mm6", "%mm7"  // clobber list
+#endif
+               );
+            }
+            else /* not _mmx_supported - use modified C routine */
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7;  /* reduce to mult. of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = len;  /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff /* *BPP1 */ ;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+
+            } /* end of else (_mmx_supported) */
+
+            break;
+         }       /* end 8 bpp */
+
+         case 1:        /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int s_inc, s_start, s_end;
+            int m;
+            int shift;
+            png_uint_32 i;
+
+            sp = png_ptr->row_buf + 1;
+            dp = row;
+            m = 0x80;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+               s_start = 0;
+               s_end = 7;
+               s_inc = 1;
+            }
+            else
+#endif
+            {
+               s_start = 7;
+               s_end = 0;
+               s_inc = -1;
+            }
+
+            shift = s_start;
+
+            for (i = 0; i < png_ptr->width; i++)
+            {
+               if (m & mask)
+               {
+                  int value;
+
+                  value = (*sp >> shift) & 0x1;
+                  *dp &= (png_byte)((0x7f7f >> (7 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }       /* end 1 bpp */
+
+         case 2:        /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int s_start, s_end, s_inc;
+            int m;
+            int shift;
+            png_uint_32 i;
+            int value;
+
+            sp = png_ptr->row_buf + 1;
+            dp = row;
+            m = 0x80;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+               s_start = 0;
+               s_end = 6;
+               s_inc = 2;
+            }
+            else
+#endif
+            {
+               s_start = 6;
+               s_end = 0;
+               s_inc = -2;
+            }
+
+            shift = s_start;
+
+            for (i = 0; i < png_ptr->width; i++)
+            {
+               if (m & mask)
+               {
+                  value = (*sp >> shift) & 0x3;
+                  *dp &= (png_byte)((0x3f3f >> (6 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }       /* end 2 bpp */
+
+         case 4:        /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int s_start, s_end, s_inc;
+            int m;
+            int shift;
+            png_uint_32 i;
+            int value;
+
+            sp = png_ptr->row_buf + 1;
+            dp = row;
+            m = 0x80;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+               s_start = 0;
+               s_end = 4;
+               s_inc = 4;
+            }
+            else
+#endif
+            {
+               s_start = 4;
+               s_end = 0;
+               s_inc = -4;
+            }
+
+            shift = s_start;
+
+            for (i = 0; i < png_ptr->width; i++)
+            {
+               if (m & mask)
+               {
+                  value = (*sp >> shift) & 0xf;
+                  *dp &= (png_byte)((0xf0f >> (4 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }       /* end 4 bpp */
+
+         case 16:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+
+#if !defined(PNG_1_0_X)
+            if (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+#else
+            if (_mmx_supported)
+#endif
+            {
+               png_uint_32 len;
+               int diff;
+
+               srcptr = png_ptr->row_buf + 1;
+               dstptr = row;
+               len  = png_ptr->width & ~7;          // reduce to multiple of 8
+               diff = (int) (png_ptr->width & 7);   // amount lost
+
+               __asm__ __volatile__ (
+                  "not       %%edx            \n\t" // mask => unmask
+                  "movd      %%edx, %%mm7     \n\t" // load bit pattern
+                  "not       %%edx            \n\t" // unmask => mask for later
+                  "psubb     %%mm6, %%mm6     \n\t" // zero mm6
+                  "punpcklbw %%mm7, %%mm7     \n\t"
+                  "punpcklwd %%mm7, %%mm7     \n\t"
+                  "punpckldq %%mm7, %%mm7     \n\t" // fill reg with 8 masks
+
+                  LOAD_GOT_rbp
+                  "movq   " MASK16_0 ", %%mm0 \n\t" // _mask16_0 -> mm0
+                  "movq   " MASK16_1 ", %%mm1 \n\t" // _mask16_1 -> mm1
+                  RESTORE_rbp
+
+                  "pand      %%mm7, %%mm0     \n\t"
+                  "pand      %%mm7, %%mm1     \n\t"
+
+                  "pcmpeqb   %%mm6, %%mm0     \n\t"
+                  "pcmpeqb   %%mm6, %%mm1     \n\t"
+
+// preload        "movl      len, %%ecx       \n\t" // load length of line
+// preload        "movl      srcptr, %3       \n\t" // load source
+// preload        "movl      dstptr, %4       \n\t" // load dest
+
+                  "cmpl      $0, %%ecx        \n\t"
+                  "jz        mainloop16end    \n\t"
+
+                "mainloop16:                  \n\t"
+                  "movq      (%3), %%mm4      \n\t"
+                  "pand      %%mm0, %%mm4     \n\t"
+                  "movq      %%mm0, %%mm6     \n\t"
+                  "movq      (%4), %%mm7      \n\t"
+                  "pandn     %%mm7, %%mm6     \n\t"
+                  "por       %%mm6, %%mm4     \n\t"
+                  "movq      %%mm4, (%4)      \n\t"
+
+                  "movq      8(%3), %%mm5     \n\t"
+                  "pand      %%mm1, %%mm5     \n\t"
+                  "movq      %%mm1, %%mm7     \n\t"
+                  "movq      8(%4), %%mm6     \n\t"
+                  "pandn     %%mm6, %%mm7     \n\t"
+                  "por       %%mm7, %%mm5     \n\t"
+                  "movq      %%mm5, 8(%4)     \n\t"
+
+                  "add       $16, %3          \n\t" // inc by 16 bytes processed
+                  "add       $16, %4          \n\t"
+                  "subl      $8, %%ecx        \n\t" // dec by 8 pixels processed
+                  "ja        mainloop16       \n\t"
+
+                "mainloop16end:               \n\t"
+// preload        "movl      diff, %%ecx      \n\t" // (diff is in eax)
+                  "movl      %%eax, %%ecx     \n\t"
+                  "cmpl      $0, %%ecx        \n\t"
+                  "jz        end16            \n\t"
+// preload        "movl      mask, %%edx      \n\t"
+                  "sall      $24, %%edx       \n\t" // make low byte, high byte
+
+                "secondloop16:                \n\t"
+                  "sall      %%edx            \n\t" // move high bit to CF
+                  "jnc       skip16           \n\t" // if CF = 0
+                  "movw      (%3), %%ax       \n\t"
+                  "movw      %%ax, (%4)       \n\t"
+
+                "skip16:                      \n\t"
+                  "add       $2, %3           \n\t"
+                  "add       $2, %4           \n\t"
+                  "decl      %%ecx            \n\t"
+                  "jnz       secondloop16     \n\t"
+
+                "end16:                       \n\t"
+                  "EMMS                       \n\t" // DONE
+
+                  : "=a" (dummy_value_a),           // output regs (dummy)
+                    "=d" (dummy_value_d),
+                    "=c" (dummy_value_c),
+                    "=S" (dummy_value_S),
+                    "=D" (dummy_value_D)
+
+                  : "0" (diff),        // eax       // input regs
+                    "1" (mask),        // edx
+                    "2" (len),         // ecx
+// was (unmask)     "b"    RESERVED    // ebx       // Global Offset Table idx
+                    "3" (srcptr),      // esi/rsi
+                    "4" (dstptr)       // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                  : "%mm0", "%mm1", "%mm4"          // clobber list
+                  , "%mm5", "%mm6", "%mm7"
+#endif
+               );
+            }
+            else /* not _mmx_supported - use modified C routine */
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = BPP2 * png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = BPP2 * png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = BPP2 * png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7;  /* reduce to mult. of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = BPP2 * len;   /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff*BPP2;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+            } /* end of else (_mmx_supported) */
+
+            break;
+         }       /* end 16 bpp */
+
+         case 48:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+
+#if !defined(PNG_1_0_X)
+            if (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+#else
+            if (_mmx_supported)
+#endif
+            {
+               png_uint_32 len;
+               int diff;
+
+               srcptr = png_ptr->row_buf + 1;
+               dstptr = row;
+               len  = png_ptr->width & ~7;          // reduce to multiple of 8
+               diff = (int) (png_ptr->width & 7);   // amount lost
+
+               __asm__ __volatile__ (
+                  "not       %%edx            \n\t" // mask => unmask
+                  "movd      %%edx, %%mm7     \n\t" // load bit pattern
+                  "not       %%edx            \n\t" // unmask => mask for later
+                  "psubb     %%mm6, %%mm6     \n\t" // zero mm6
+                  "punpcklbw %%mm7, %%mm7     \n\t"
+                  "punpcklwd %%mm7, %%mm7     \n\t"
+                  "punpckldq %%mm7, %%mm7     \n\t" // fill reg with 8 masks
+
+                  LOAD_GOT_rbp
+                  "movq   " MASK48_0 ", %%mm0 \n\t" // _mask48_0 -> mm0
+                  "movq   " MASK48_1 ", %%mm1 \n\t" // _mask48_1 -> mm1
+                  "movq   " MASK48_2 ", %%mm2 \n\t" // _mask48_2 -> mm2
+                  "movq   " MASK48_3 ", %%mm3 \n\t" // _mask48_3 -> mm3
+                  "movq   " MASK48_4 ", %%mm4 \n\t" // _mask48_4 -> mm4
+                  "movq   " MASK48_5 ", %%mm5 \n\t" // _mask48_5 -> mm5
+                  RESTORE_rbp
+
+                  "pand      %%mm7, %%mm0     \n\t"
+                  "pand      %%mm7, %%mm1     \n\t"
+                  "pand      %%mm7, %%mm2     \n\t"
+                  "pand      %%mm7, %%mm3     \n\t"
+                  "pand      %%mm7, %%mm4     \n\t"
+                  "pand      %%mm7, %%mm5     \n\t"
+
+                  "pcmpeqb   %%mm6, %%mm0     \n\t"
+                  "pcmpeqb   %%mm6, %%mm1     \n\t"
+                  "pcmpeqb   %%mm6, %%mm2     \n\t"
+                  "pcmpeqb   %%mm6, %%mm3     \n\t"
+                  "pcmpeqb   %%mm6, %%mm4     \n\t"
+                  "pcmpeqb   %%mm6, %%mm5     \n\t"
+
+// preload        "movl      len, %%ecx       \n\t" // load length of line
+// preload        "movl      srcptr, %3       \n\t" // load source
+// preload        "movl      dstptr, %4       \n\t" // load dest
+
+                  "cmpl      $0, %%ecx        \n\t"
+                  "jz        mainloop48end    \n\t"
+
+                "mainloop48:                  \n\t"
+                  "movq      (%3), %%mm7      \n\t"
+                  "pand      %%mm0, %%mm7     \n\t"
+                  "movq      %%mm0, %%mm6     \n\t"
+                  "pandn     (%4), %%mm6      \n\t"
+                  "por       %%mm6, %%mm7     \n\t"
+                  "movq      %%mm7, (%4)      \n\t"
+
+                  "movq      8(%3), %%mm6     \n\t"
+                  "pand      %%mm1, %%mm6     \n\t"
+                  "movq      %%mm1, %%mm7     \n\t"
+                  "pandn     8(%4), %%mm7     \n\t"
+                  "por       %%mm7, %%mm6     \n\t"
+                  "movq      %%mm6, 8(%4)     \n\t"
+
+                  "movq      16(%3), %%mm6    \n\t"
+                  "pand      %%mm2, %%mm6     \n\t"
+                  "movq      %%mm2, %%mm7     \n\t"
+                  "pandn     16(%4), %%mm7    \n\t"
+                  "por       %%mm7, %%mm6     \n\t"
+                  "movq      %%mm6, 16(%4)    \n\t"
+
+                  "movq      24(%3), %%mm7    \n\t"
+                  "pand      %%mm3, %%mm7     \n\t"
+                  "movq      %%mm3, %%mm6     \n\t"
+                  "pandn     24(%4), %%mm6    \n\t"
+                  "por       %%mm6, %%mm7     \n\t"
+                  "movq      %%mm7, 24(%4)    \n\t"
+
+                  "movq      32(%3), %%mm6    \n\t"
+                  "pand      %%mm4, %%mm6     \n\t"
+                  "movq      %%mm4, %%mm7     \n\t"
+                  "pandn     32(%4), %%mm7    \n\t"
+                  "por       %%mm7, %%mm6     \n\t"
+                  "movq      %%mm6, 32(%4)    \n\t"
+
+                  "movq      40(%3), %%mm7    \n\t"
+                  "pand      %%mm5, %%mm7     \n\t"
+                  "movq      %%mm5, %%mm6     \n\t"
+                  "pandn     40(%4), %%mm6    \n\t"
+                  "por       %%mm6, %%mm7     \n\t"
+                  "movq      %%mm7, 40(%4)    \n\t"
+
+                  "add       $48, %3          \n\t" // inc by 48 bytes processed
+                  "add       $48, %4          \n\t"
+                  "subl      $8, %%ecx        \n\t" // dec by 8 pixels processed
+
+                  "ja        mainloop48       \n\t"
+
+                "mainloop48end:               \n\t"
+// preload        "movl      diff, %%ecx      \n\t" // (diff is in eax)
+                  "movl      %%eax, %%ecx     \n\t"
+                  "cmpl      $0, %%ecx        \n\t"
+                  "jz        end48            \n\t"
+// preload        "movl      mask, %%edx      \n\t"
+                  "sall      $24, %%edx       \n\t" // make low byte, high byte
+
+                "secondloop48:                \n\t"
+                  "sall      %%edx            \n\t" // move high bit to CF
+                  "jnc       skip48           \n\t" // if CF = 0
+                  "movl      (%3), %%eax      \n\t"
+                  "movl      %%eax, (%4)      \n\t"
+                  "movw      4(%3), %%ax      \n\t" // GR-P bugfix 20070717
+                  "movw      %%ax, 4(%4)      \n\t" // GR-P bugfix 20070717
+
+                "skip48:                      \n\t"
+                  "add       $6, %3           \n\t" // GR-P bugfix 20070717
+                  "add       $6, %4           \n\t" // GR-P bugfix 20070717
+                  "decl      %%ecx            \n\t"
+                  "jnz       secondloop48     \n\t"
+
+                "end48:                       \n\t"
+                  "EMMS                       \n\t" // DONE
+
+                  : "=a" (dummy_value_a),           // output regs (dummy)
+                    "=d" (dummy_value_d),
+                    "=c" (dummy_value_c),
+                    "=S" (dummy_value_S),
+                    "=D" (dummy_value_D)
+
+                  : "0" (diff),        // eax       // input regs
+                    "1" (mask),        // edx
+                    "2" (len),         // ecx
+// was (unmask)     "b"    RESERVED    // ebx       // Global Offset Table idx
+                    "3" (srcptr),      // esi/rsi
+                    "4" (dstptr)       // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                  : "%mm0", "%mm1", "%mm2", "%mm3"  // clobber list
+                  , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+               );
+            }
+            else /* not _mmx_supported - use modified C routine */
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = BPP6 * png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = BPP6 * png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = BPP6 * png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7;  /* reduce to mult. of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = BPP6 * len;   /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff*BPP6;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+            } /* end of else (_mmx_supported) */
+
+            break;
+         }       /* end 48 bpp */
+
+         case 64:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+            register png_uint_32 i;
+            png_uint_32 initial_val = BPP8 * png_pass_start[png_ptr->pass];
+              /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+            register int stride = BPP8 * png_pass_inc[png_ptr->pass];
+              /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+            register int rep_bytes = BPP8 * png_pass_width[png_ptr->pass];
+              /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+            png_uint_32 len = png_ptr->width &~7;  /* reduce to mult. of 8 */
+            int diff = (int) (png_ptr->width & 7); /* amount lost */
+            register png_uint_32 final_val = BPP8 * len;   /* GRR bugfix */
+
+            srcptr = png_ptr->row_buf + 1 + initial_val;
+            dstptr = row + initial_val;
+
+            for (i = initial_val; i < final_val; i += stride)
+            {
+               png_memcpy(dstptr, srcptr, rep_bytes);
+               srcptr += stride;
+               dstptr += stride;
+            }
+            if (diff)  /* number of leftover pixels:  3 for pngtest */
+            {
+               final_val += diff*BPP8;
+               for (; i < final_val; i += stride)
+               {
+                  if (rep_bytes > (int)(final_val-i))
+                     rep_bytes = (int)(final_val-i);
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+            }
+
+            break;
+         }       /* end 64 bpp */
+
+         default: /* png_ptr->row_info.pixel_depth != 1,2,4,8,16,24,32,48,64 */
+         {
+            // ERROR:  SHOULD NEVER BE REACHED
+#if defined(PNG_DEBUG)
+            png_debug(1, "Internal libpng logic error (GCC "
+              "png_combine_row() pixel_depth)\n");
+#endif
+            break;
+         }
+      } /* end switch (png_ptr->row_info.pixel_depth) */
+
+   } /* end if (non-trivial mask) */
+
+} /* end png_combine_row() */
+
+#endif /* PNG_HAVE_MMX_COMBINE_ROW */
+
+
+
+
+/*===========================================================================*/
+/*                                                                           */
+/*                 P N G _ D O _ R E A D _ I N T E R L A C E                 */
+/*                                                                           */
+/*===========================================================================*/
+
+#if defined(PNG_READ_INTERLACING_SUPPORTED)
+#if defined(PNG_HAVE_MMX_READ_INTERLACE)
+
+/* png_do_read_interlace() is called after any 16-bit to 8-bit conversion
+ * has taken place.  [GRR: what other steps come before and/or after?]
+ */
+
+void /* PRIVATE */
+png_do_read_interlace(png_structp png_ptr)
+{
+   png_row_infop row_info = &(png_ptr->row_info);
+   png_bytep row = png_ptr->row_buf + 1;
+   int pass = png_ptr->pass;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+   png_uint_32 transformations = png_ptr->transformations;
+#endif
+
+   png_debug(1, "in png_do_read_interlace (pnggccrd.c)\n");
+
+   if (_mmx_supported == 2) {
+#if !defined(PNG_1_0_X)
+       /* this should have happened in png_init_mmx_flags() already */
+       png_warning(png_ptr, "asm_flags may not have been initialized");
+#endif
+       png_mmx_support();
+   }
+
+   if (row != NULL && row_info != NULL)
+   {
+      png_uint_32 final_width;
+
+      final_width = row_info->width * png_pass_inc[pass];
+
+      switch (row_info->pixel_depth)
+      {
+         case 1:
+         {
+            png_bytep sp, dp;
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_byte v;
+            png_uint_32 i;
+            int j;
+
+            sp = row + (png_size_t)((row_info->width - 1) >> 3);
+            dp = row + (png_size_t)((final_width - 1) >> 3);
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (int)((row_info->width + 7) & 7);
+               dshift = (int)((final_width + 7) & 7);
+               s_start = 7;
+               s_end = 0;
+               s_inc = -1;
+            }
+            else
+#endif
+            {
+               sshift = 7 - (int)((row_info->width + 7) & 7);
+               dshift = 7 - (int)((final_width + 7) & 7);
+               s_start = 0;
+               s_end = 7;
+               s_inc = 1;
+            }
+
+            for (i = row_info->width; i; i--)
+            {
+               v = (png_byte)((*sp >> sshift) & 0x1);
+               for (j = 0; j < png_pass_inc[pass]; j++)
+               {
+                  *dp &= (png_byte)((0x7f7f >> (7 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+
+         case 2:
+         {
+            png_bytep sp, dp;
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_uint_32 i;
+
+            sp = row + (png_size_t)((row_info->width - 1) >> 2);
+            dp = row + (png_size_t)((final_width - 1) >> 2);
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (png_size_t)(((row_info->width + 3) & 3) << 1);
+               dshift = (png_size_t)(((final_width + 3) & 3) << 1);
+               s_start = 6;
+               s_end = 0;
+               s_inc = -2;
+            }
+            else
+#endif
+            {
+               sshift = (png_size_t)((3 - ((row_info->width + 3) & 3)) << 1);
+               dshift = (png_size_t)((3 - ((final_width + 3) & 3)) << 1);
+               s_start = 0;
+               s_end = 6;
+               s_inc = 2;
+            }
+
+            for (i = row_info->width; i; i--)
+            {
+               png_byte v;
+               int j;
+
+               v = (png_byte)((*sp >> sshift) & 0x3);
+               for (j = 0; j < png_pass_inc[pass]; j++)
+               {
+                  *dp &= (png_byte)((0x3f3f >> (6 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+
+         case 4:
+         {
+            png_bytep sp, dp;
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_uint_32 i;
+
+            sp = row + (png_size_t)((row_info->width - 1) >> 1);
+            dp = row + (png_size_t)((final_width - 1) >> 1);
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (png_size_t)(((row_info->width + 1) & 1) << 2);
+               dshift = (png_size_t)(((final_width + 1) & 1) << 2);
+               s_start = 4;
+               s_end = 0;
+               s_inc = -4;
+            }
+            else
+#endif
+            {
+               sshift = (png_size_t)((1 - ((row_info->width + 1) & 1)) << 2);
+               dshift = (png_size_t)((1 - ((final_width + 1) & 1)) << 2);
+               s_start = 0;
+               s_end = 4;
+               s_inc = 4;
+            }
+
+            for (i = row_info->width; i; i--)
+            {
+               png_byte v;
+               int j;
+
+               v = (png_byte)((*sp >> sshift) & 0xf);
+               for (j = 0; j < png_pass_inc[pass]; j++)
+               {
+                  *dp &= (png_byte)((0xf0f >> (4 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+
+       /*====================================================================*/
+
+         default: /* 8-bit or larger (this is where the routine is modified) */
+         {
+            png_bytep sptr, dp;
+            png_uint_32 i;
+            png_size_t pixel_bytes;
+            int width = (int)row_info->width;
+
+            pixel_bytes = (row_info->pixel_depth >> 3);
+
+            /* point sptr at the last pixel in the pre-expanded row: */
+            sptr = row + (width - 1) * pixel_bytes;
+
+            /* point dp at the last pixel position in the expanded row: */
+            dp = row + (final_width - 1) * pixel_bytes;
+
+            /* New code by Nirav Chhatrapati - Intel Corporation */
+
+#if !defined(PNG_1_0_X)
+            if (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_INTERLACE)
+#else
+            if (_mmx_supported)
+#endif
+            {
+               int dummy_value_c;        // fix 'forbidden register spilled'
+               png_bytep dummy_value_S;
+               png_bytep dummy_value_D;
+               png_bytep dummy_value_a;
+               png_bytep dummy_value_d;
+
+               //--------------------------------------------------------------
+               if (pixel_bytes == BPP3)
+               {
+                  if (((pass == 4) || (pass == 5)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1) - 8;   // GRR:  huh?
+                     if (width_mmx < 0)
+                         width_mmx = 0;
+                     width -= width_mmx;        // 8 or 9 pix, 24 or 27 bytes
+                     if (width_mmx)
+                     {
+                        // png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1};
+                        // sptr points at last pixel in pre-expanded row
+                        // dp points at last pixel position in expanded row
+                        __asm__ __volatile__ (
+                           "sub  $3, %1             \n\t"
+                           "sub  $9, %2             \n\t"
+                                        // (png_pass_inc[pass] + 1)*pixel_bytes
+
+                        ".loop3_pass4:              \n\t"
+                           "movq (%1), %%mm0        \n\t" // x x 5 4 3 2 1 0
+                           "movq %%mm0, %%mm1       \n\t" // x x 5 4 3 2 1 0
+                           "movq %%mm0, %%mm2       \n\t" // x x 5 4 3 2 1 0
+                           "psllq $24, %%mm0        \n\t" // 4 3 2 1 0 z z z
+                           "pand (%3), %%mm1        \n\t" // z z z z z 2 1 0
+                           "psrlq $24, %%mm2        \n\t" // z z z x x 5 4 3
+                           "por %%mm1, %%mm0        \n\t" // 4 3 2 1 0 2 1 0
+                           "movq %%mm2, %%mm3       \n\t" // z z z x x 5 4 3
+                           "psllq $8, %%mm2         \n\t" // z z x x 5 4 3 z
+                           "movq %%mm0, (%2)        \n\t"
+                           "psrlq $16, %%mm3        \n\t" // z z z z z x x 5
+                           "pand (%4), %%mm3        \n\t" // z z z z z z z 5
+                           "por %%mm3, %%mm2        \n\t" // z z x x 5 4 3 5
+                           "sub  $6, %1             \n\t"
+                           "movd %%mm2, 8(%2)       \n\t"
+                           "sub  $12, %2            \n\t"
+                           "subl $2, %%ecx          \n\t"
+                           "jnz .loop3_pass4        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D),
+                             "=a" (dummy_value_a),
+                             "=d" (dummy_value_d)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp),            // edi/rdi
+#if defined(PNG_x86_64_USE_GOTPCREL)     // formerly _const4 and _const6:
+                             "3" (&_c64._amask5_3_0), // (0x0000000000FFFFFFLL)
+                             "4" (&_c64._amask7_1_0)  // (0x00000000000000FFLL)
+#else
+                             "3" (&_amask5_3_0),  // eax (0x0000000000FFFFFFLL)
+                             "4" (&_amask7_1_0)   // edx (0x00000000000000FFLL)
+#endif
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0", "%mm1"               // clobber list
+                           , "%mm2", "%mm3"
+#endif
+                        );
+                     }
+
+                     sptr -= width_mmx*BPP3;
+                     dp -= width_mmx*2*BPP3;
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+
+                        png_memcpy(v, sptr, BPP3);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           png_memcpy(dp, v, BPP3);
+                           dp -= BPP3;
+                        }
+                        sptr -= BPP3;
+                     }
+                  }
+                  else if (((pass == 2) || (pass == 3)) && width)
+                  {
+                     __asm__ __volatile__ (
+                        "sub  $9, %2             \n\t"
+                                     // (png_pass_inc[pass] - 1)*pixel_bytes
+
+                     ".loop3_pass2:              \n\t"
+                        "movd (%1), %%mm0        \n\t" // x x x x x 2 1 0
+                        "pand (%3), %%mm0        \n\t" // z z z z z 2 1 0
+                        "movq %%mm0, %%mm1       \n\t" // z z z z z 2 1 0
+                        "psllq $16, %%mm0        \n\t" // z z z 2 1 0 z z
+                        "movq %%mm0, %%mm2       \n\t" // z z z 2 1 0 z z
+                        "psllq $24, %%mm0        \n\t" // 2 1 0 z z z z z
+                        "psrlq $8, %%mm1         \n\t" // z z z z z z 2 1
+                        "por %%mm2, %%mm0        \n\t" // 2 1 0 2 1 0 z z
+                        "por %%mm1, %%mm0        \n\t" // 2 1 0 2 1 0 2 1
+                        "movq %%mm0, 4(%2)       \n\t"
+                        "psrlq $16, %%mm0        \n\t" // z z 2 1 0 2 1 0
+                        "sub  $3, %1             \n\t"
+                        "movd %%mm0, (%2)        \n\t"
+                        "sub  $12, %2            \n\t"
+                        "decl %%ecx              \n\t"
+                        "jnz .loop3_pass2        \n\t"
+                        "EMMS                    \n\t" // DONE
+
+                        : "=c" (dummy_value_c),        // output regs (dummy)
+                          "=S" (dummy_value_S),
+                          "=D" (dummy_value_D),
+                          "=a" (dummy_value_a)
+
+                        : "0" (width),         // ecx  // input regs
+                          "1" (sptr),          // esi/rsi
+                          "2" (dp),            // edi/rdi
+#if defined(PNG_x86_64_USE_GOTPCREL)           // formerly _const4:
+                          "3" (&_c64._amask5_3_0)  // (0x0000000000FFFFFFLL)
+#else
+                          "3" (&_amask5_3_0)   // eax (0x0000000000FFFFFFLL)
+#endif
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+                        : "%mm0", "%mm1", "%mm2"       // clobber list
+#endif
+                     );
+                  }
+                  else if (width)  // && ((pass == 0) || (pass == 1))
+                  {
+                     __asm__ __volatile__ (
+                        "sub  $21, %2            \n\t"
+                                     // (png_pass_inc[pass] - 1)*pixel_bytes
+
+                     ".loop3_pass0:              \n\t"
+                        "movd (%1), %%mm0        \n\t" // x x x x x 2 1 0
+                        "pand (%3), %%mm0        \n\t" // z z z z z 2 1 0
+                        "movq %%mm0, %%mm1       \n\t" // z z z z z 2 1 0
+                        "psllq $16, %%mm0        \n\t" // z z z 2 1 0 z z
+                        "movq %%mm0, %%mm2       \n\t" // z z z 2 1 0 z z
+                        "psllq $24, %%mm0        \n\t" // 2 1 0 z z z z z
+                        "psrlq $8, %%mm1         \n\t" // z z z z z z 2 1
+                        "por %%mm2, %%mm0        \n\t" // 2 1 0 2 1 0 z z
+                        "por %%mm1, %%mm0        \n\t" // 2 1 0 2 1 0 2 1
+                        "movq %%mm0, %%mm3       \n\t" // 2 1 0 2 1 0 2 1
+                        "psllq $16, %%mm0        \n\t" // 0 2 1 0 2 1 z z
+                        "movq %%mm3, %%mm4       \n\t" // 2 1 0 2 1 0 2 1
+                        "punpckhdq %%mm0, %%mm3  \n\t" // 0 2 1 0 2 1 0 2
+                        "movq %%mm4, 16(%2)      \n\t"
+                        "psrlq $32, %%mm0        \n\t" // z z z z 0 2 1 0
+                        "movq %%mm3, 8(%2)       \n\t"
+                        "punpckldq %%mm4, %%mm0  \n\t" // 1 0 2 1 0 2 1 0
+                        "sub  $3, %1             \n\t"
+                        "movq %%mm0, (%2)        \n\t"
+                        "sub  $24, %2            \n\t"
+                        "decl %%ecx              \n\t"
+                        "jnz .loop3_pass0        \n\t"
+                        "EMMS                    \n\t" // DONE
+
+                        : "=c" (dummy_value_c),        // output regs (dummy)
+                          "=S" (dummy_value_S),
+                          "=D" (dummy_value_D),
+                          "=a" (dummy_value_a)
+
+                        : "0" (width),         // ecx  // input regs
+                          "1" (sptr),          // esi/rsi
+                          "2" (dp),            // edi/rdi
+#if defined(PNG_x86_64_USE_GOTPCREL)           // formerly _const4:
+                          "3" (&_c64._amask5_3_0)  // (0x0000000000FFFFFFLL)
+#else
+                          "3" (&_amask5_3_0)   // eax (0x0000000000FFFFFFLL)
+#endif
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                        : "%mm0", "%mm1", "%mm2"       // clobber list
+                        , "%mm3", "%mm4"
+#endif
+                     );
+                  }
+               } /* end of pixel_bytes == 3 */
+
+               //--------------------------------------------------------------
+               else if (pixel_bytes == BPP4)
+               {
+                  if (((pass == 4) || (pass == 5)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1) ;
+                     width -= width_mmx;        // 0,1 pixels => 0,4 bytes
+                     if (width_mmx)
+                     {
+                        __asm__ __volatile__ (
+                           "sub  $4, %1             \n\t"
+                           "sub  $12, %2            \n\t"
+
+                        ".loop4_pass4:              \n\t"
+                           "movq (%1), %%mm0        \n\t" // 7 6 5 4 3 2 1 0
+                           "movq %%mm0, %%mm1       \n\t" // 7 6 5 4 3 2 1 0
+                           "punpckldq %%mm0, %%mm0  \n\t" // 3 2 1 0 3 2 1 0
+                           "punpckhdq %%mm1, %%mm1  \n\t" // 7 6 5 4 7 6 5 4
+                           "movq %%mm0, (%2)        \n\t"
+                           "sub  $8, %1             \n\t"
+                           "movq %%mm1, 8(%2)       \n\t"
+                           "sub  $16, %2            \n\t"
+                           "subl $2, %%ecx          \n\t"
+                           "jnz .loop4_pass4        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp)             // edi/rdi
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0", "%mm1"               // clobber list
+#endif
+                        );
+                     }
+
+                     sptr -= (width_mmx*BPP4 - BPP4); // sign fixed
+                     dp -= (width_mmx*2*BPP4 - BPP4); // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= BPP4;
+                        png_memcpy(v, sptr, BPP4);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= BPP4;
+                           png_memcpy(dp, v, BPP4);
+                        }
+                     }
+                  }
+                  else if (((pass == 2) || (pass == 3)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1);
+                     width -= width_mmx;        // 0,1 pixels => 0,4 bytes
+                     if (width_mmx)
+                     {
+                        __asm__ __volatile__ (
+                           "sub  $4, %1             \n\t"
+                           "sub  $28, %2            \n\t"
+
+                        ".loop4_pass2:              \n\t"
+                           "movq (%1), %%mm0        \n\t" // 7 6 5 4 3 2 1 0
+                           "movq %%mm0, %%mm1       \n\t" // 7 6 5 4 3 2 1 0
+                           "punpckldq %%mm0, %%mm0  \n\t" // 3 2 1 0 3 2 1 0
+                           "punpckhdq %%mm1, %%mm1  \n\t" // 7 6 5 4 7 6 5 4
+                           "movq %%mm0, (%2)        \n\t"
+                           "movq %%mm0, 8(%2)       \n\t"
+                           "movq %%mm1, 16(%2)      \n\t"
+                           "movq %%mm1, 24(%2)      \n\t"
+                           "sub  $8, %1             \n\t"
+                           "sub  $32, %2            \n\t"
+                           "subl $2, %%ecx          \n\t"
+                           "jnz .loop4_pass2        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp)             // edi/rdi
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0", "%mm1"               // clobber list
+#endif
+                        );
+                     }
+
+                     sptr -= (width_mmx*4 - 4); // sign fixed
+                     dp -= (width_mmx*16 - 4);  // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 4;
+                        png_memcpy(v, sptr, 4);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 4;
+                           png_memcpy(dp, v, 4);
+                        }
+                     }
+                  }
+                  else if (width)  // && ((pass == 0) || (pass == 1))
+                  {
+                     int width_mmx = ((width >> 1) << 1);
+                     width -= width_mmx;        // 0,1 pixels => 0,4 bytes
+                     if (width_mmx)
+                     {
+                        __asm__ __volatile__ (
+                           "sub  $4, %1             \n\t"
+                           "sub  $60, %2            \n\t"
+
+                        ".loop4_pass0:              \n\t"
+                           "movq (%1), %%mm0        \n\t" // 7 6 5 4 3 2 1 0
+                           "movq %%mm0, %%mm1       \n\t" // 7 6 5 4 3 2 1 0
+                           "punpckldq %%mm0, %%mm0  \n\t" // 3 2 1 0 3 2 1 0
+                           "punpckhdq %%mm1, %%mm1  \n\t" // 7 6 5 4 7 6 5 4
+                           "movq %%mm0, (%2)        \n\t"
+                           "movq %%mm0, 8(%2)       \n\t"
+                           "movq %%mm0, 16(%2)      \n\t"
+                           "movq %%mm0, 24(%2)      \n\t"
+                           "movq %%mm1, 32(%2)      \n\t"
+                           "movq %%mm1, 40(%2)      \n\t"
+                           "movq %%mm1, 48(%2)      \n\t"
+                           "sub  $8, %1             \n\t"
+                           "movq %%mm1, 56(%2)      \n\t"
+                           "sub  $64, %2            \n\t"
+                           "subl $2, %%ecx          \n\t"
+                           "jnz .loop4_pass0        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp)             // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0", "%mm1"               // clobber list
+#endif
+                        );
+                     }
+
+                     sptr -= (width_mmx*4 - 4); // sign fixed
+                     dp -= (width_mmx*32 - 4);  // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 4;
+                        png_memcpy(v, sptr, 4);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 4;
+                           png_memcpy(dp, v, 4);
+                        }
+                     }
+                  }
+               } /* end of pixel_bytes == 4 */
+
+               //--------------------------------------------------------------
+               else if (pixel_bytes == 1)
+               {
+                  if (((pass == 4) || (pass == 5)) && width)
+                  {
+                     int width_mmx = ((width >> 3) << 3);
+                     width -= width_mmx;        // 0-3 pixels => 0-3 bytes
+                     if (width_mmx)
+                     {
+                        __asm__ __volatile__ (
+                           "sub  $7, %1             \n\t"
+                           "sub  $15, %2            \n\t"
+
+                        ".loop1_pass4:              \n\t"
+                           "movq (%1), %%mm0        \n\t" // 7 6 5 4 3 2 1 0
+                           "movq %%mm0, %%mm1       \n\t" // 7 6 5 4 3 2 1 0
+                           "punpcklbw %%mm0, %%mm0  \n\t" // 3 3 2 2 1 1 0 0
+                           "punpckhbw %%mm1, %%mm1  \n\t" // 7 7 6 6 5 5 4 4
+                           "movq %%mm1, 8(%2)       \n\t"
+                           "sub  $8, %1             \n\t"
+                           "movq %%mm0, (%2)        \n\t"
+                           "sub  $16, %2            \n\t"
+                           "subl $8, %%ecx          \n\t"
+                           "jnz .loop1_pass4        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp)             // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0", "%mm1"               // clobber list
+#endif
+                        );
+                     }
+
+                     sptr -= width_mmx;
+                     dp -= width_mmx*2;
+                     for (i = width; i; i--)
+                     {
+                        int j;
+
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           *dp-- = *sptr;
+                        }
+                        --sptr;
+                     }
+                  }
+                  else if (((pass == 2) || (pass == 3)) && width)
+                  {
+                     int width_mmx = ((width >> 2) << 2);
+                     width -= width_mmx;        // 0-3 pixels => 0-3 bytes
+                     if (width_mmx)
+                     {
+                        __asm__ __volatile__ (
+                           "sub  $3, %1             \n\t"
+                           "sub  $15, %2            \n\t"
+
+                        ".loop1_pass2:              \n\t"
+                           "movd (%1), %%mm0        \n\t" // x x x x 3 2 1 0
+                           "punpcklbw %%mm0, %%mm0  \n\t" // 3 3 2 2 1 1 0 0
+                           "movq %%mm0, %%mm1       \n\t" // 3 3 2 2 1 1 0 0
+                           "punpcklwd %%mm0, %%mm0  \n\t" // 1 1 1 1 0 0 0 0
+                           "punpckhwd %%mm1, %%mm1  \n\t" // 3 3 3 3 2 2 2 2
+                           "movq %%mm0, (%2)        \n\t"
+                           "sub  $4, %1             \n\t"
+                           "movq %%mm1, 8(%2)       \n\t"
+                           "sub  $16, %2            \n\t"
+                           "subl $4, %%ecx          \n\t"
+                           "jnz .loop1_pass2        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp)             // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0", "%mm1"               // clobber list
+#endif
+                        );
+                     }
+
+                     sptr -= width_mmx;
+                     dp -= width_mmx*4;
+                     for (i = width; i; i--)
+                     {
+                        int j;
+
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           *dp-- = *sptr;
+                        }
+                        --sptr;
+                     }
+                  }
+                  else if (width)  // && ((pass == 0) || (pass == 1))
+                  {
+                     int width_mmx = ((width >> 2) << 2);
+                     width -= width_mmx;        // 0-3 pixels => 0-3 bytes
+                     if (width_mmx)
+                     {
+                        __asm__ __volatile__ (
+                           "sub  $3, %1             \n\t"
+                           "sub  $31, %2            \n\t"
+
+                        ".loop1_pass0:              \n\t"
+                           "movd (%1), %%mm0        \n\t" // x x x x 3 2 1 0
+                           "movq %%mm0, %%mm1       \n\t" // x x x x 3 2 1 0
+                           "punpcklbw %%mm0, %%mm0  \n\t" // 3 3 2 2 1 1 0 0
+                           "movq %%mm0, %%mm2       \n\t" // 3 3 2 2 1 1 0 0
+                           "punpcklwd %%mm0, %%mm0  \n\t" // 1 1 1 1 0 0 0 0
+                           "movq %%mm0, %%mm3       \n\t" // 1 1 1 1 0 0 0 0
+                           "punpckldq %%mm0, %%mm0  \n\t" // 0 0 0 0 0 0 0 0
+                           "punpckhdq %%mm3, %%mm3  \n\t" // 1 1 1 1 1 1 1 1
+                           "movq %%mm0, (%2)        \n\t"
+                           "punpckhwd %%mm2, %%mm2  \n\t" // 3 3 3 3 2 2 2 2
+                           "movq %%mm3, 8(%2)       \n\t"
+                           "movq %%mm2, %%mm4       \n\t" // 3 3 3 3 2 2 2 2
+                           "punpckldq %%mm2, %%mm2  \n\t" // 2 2 2 2 2 2 2 2
+                           "punpckhdq %%mm4, %%mm4  \n\t" // 3 3 3 3 3 3 3 3
+                           "movq %%mm2, 16(%2)      \n\t"
+                           "sub  $4, %1             \n\t"
+                           "movq %%mm4, 24(%2)      \n\t"
+                           "sub  $32, %2            \n\t"
+                           "subl $4, %%ecx          \n\t"
+                           "jnz .loop1_pass0        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp)             // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0", "%mm1", "%mm2"       // clobber list
+                           , "%mm3", "%mm4"
+#endif
+                        );
+                     }
+
+                     sptr -= width_mmx;
+                     dp -= width_mmx*8;
+                     for (i = width; i; i--)
+                     {
+                        int j;
+
+                       /* I simplified this part in version 1.0.4e
+                        * here and in several other instances where
+                        * pixel_bytes == 1  -- GR-P
+                        *
+                        * Original code:
+                        *
+                        * png_byte v[8];
+                        * png_memcpy(v, sptr, pixel_bytes);
+                        * for (j = 0; j < png_pass_inc[pass]; j++)
+                        * {
+                        *    png_memcpy(dp, v, pixel_bytes);
+                        *    dp -= pixel_bytes;
+                        * }
+                        * sptr -= pixel_bytes;
+                        *
+                        * Replacement code is in the next three lines:
+                        */
+
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           *dp-- = *sptr;
+                        }
+                        --sptr;
+                     }
+                  }
+               } /* end of pixel_bytes == 1 */
+
+               //--------------------------------------------------------------
+               else if (pixel_bytes == BPP2)
+               {
+                  if (((pass == 4) || (pass == 5)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1) ;
+                     width -= width_mmx;        // 0,1 pixels => 0,2 bytes
+                     if (width_mmx)
+                     {
+                        __asm__ __volatile__ (
+                           "sub  $2, %1             \n\t"
+                           "sub  $6, %2             \n\t"
+
+                        ".loop2_pass4:              \n\t"
+                           "movd (%1), %%mm0        \n\t" // x x x x 3 2 1 0
+                           "punpcklwd %%mm0, %%mm0  \n\t" // 3 2 3 2 1 0 1 0
+                           "sub  $4, %1             \n\t"
+                           "movq %%mm0, (%2)        \n\t"
+                           "sub  $8, %2             \n\t"
+                           "subl $2, %%ecx          \n\t"
+                           "jnz .loop2_pass4        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp)             // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0"                       // clobber list
+#endif
+                        );
+                     }
+
+                     sptr -= (width_mmx*BPP2 - BPP2); // sign fixed
+                     dp -= (width_mmx*2*BPP2 - BPP2); // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= BPP2;
+                        png_memcpy(v, sptr, BPP2);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= BPP2;
+                           png_memcpy(dp, v, BPP2);
+                        }
+                     }
+                  }
+                  else if (((pass == 2) || (pass == 3)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1) ;
+                     width -= width_mmx;        // 0,1 pixels => 0,2 bytes
+                     if (width_mmx)
+                     {
+                        __asm__ __volatile__ (
+                           "sub  $2, %1             \n\t"
+                           "sub  $14, %2            \n\t"
+
+                        ".loop2_pass2:              \n\t"
+                           "movd (%1), %%mm0        \n\t" // x x x x 3 2 1 0
+                           "punpcklwd %%mm0, %%mm0  \n\t" // 3 2 3 2 1 0 1 0
+                           "movq %%mm0, %%mm1       \n\t" // 3 2 3 2 1 0 1 0
+                           "punpckldq %%mm0, %%mm0  \n\t" // 1 0 1 0 1 0 1 0
+                           "punpckhdq %%mm1, %%mm1  \n\t" // 3 2 3 2 3 2 3 2
+                           "movq %%mm0, (%2)        \n\t"
+                           "sub  $4, %1             \n\t"
+                           "movq %%mm1, 8(%2)       \n\t"
+                           "sub  $16, %2            \n\t"
+                           "subl $2, %%ecx          \n\t"
+                           "jnz .loop2_pass2        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp)             // edi/rdi
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0", "%mm1"               // clobber list
+#endif
+                        );
+                     }
+
+                     sptr -= (width_mmx*2 - 2); // sign fixed
+                     dp -= (width_mmx*8 - 2);   // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 2;
+                        png_memcpy(v, sptr, 2);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 2;
+                           png_memcpy(dp, v, 2);
+                        }
+                     }
+                  }
+                  else if (width)  // && ((pass == 0) || (pass == 1))
+                  {
+                     int width_mmx = ((width >> 1) << 1);
+                     width -= width_mmx;        // 0,1 pixels => 0,2 bytes
+                     if (width_mmx)
+                     {
+                        __asm__ __volatile__ (
+                           "sub  $2, %1             \n\t"
+                           "sub  $30, %2            \n\t"
+
+                        ".loop2_pass0:              \n\t"
+                           "movd (%1), %%mm0        \n\t" // x x x x 3 2 1 0
+                           "punpcklwd %%mm0, %%mm0  \n\t" // 3 2 3 2 1 0 1 0
+                           "movq %%mm0, %%mm1       \n\t" // 3 2 3 2 1 0 1 0
+                           "punpckldq %%mm0, %%mm0  \n\t" // 1 0 1 0 1 0 1 0
+                           "punpckhdq %%mm1, %%mm1  \n\t" // 3 2 3 2 3 2 3 2
+                           "movq %%mm0, (%2)        \n\t"
+                           "movq %%mm0, 8(%2)       \n\t"
+                           "movq %%mm1, 16(%2)      \n\t"
+                           "sub  $4, %1             \n\t"
+                           "movq %%mm1, 24(%2)      \n\t"
+                           "sub  $32, %2            \n\t"
+                           "subl $2, %%ecx          \n\t"
+                           "jnz .loop2_pass0        \n\t"
+                           "EMMS                    \n\t" // DONE
+
+                           : "=c" (dummy_value_c),        // output regs (dummy)
+                             "=S" (dummy_value_S),
+                             "=D" (dummy_value_D)
+
+                           : "0" (width_mmx),     // ecx  // input regs
+                             "1" (sptr),          // esi/rsi
+                             "2" (dp)             // edi/rdi
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+                           : "%mm0", "%mm1"               // clobber list
+#endif
+                        );
+                     }
+
+                     sptr -= (width_mmx*2 - 2); // sign fixed
+                     dp -= (width_mmx*16 - 2);  // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 2;
+                        png_memcpy(v, sptr, 2);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 2;
+                           png_memcpy(dp, v, 2);
+                        }
+                     }
+                  }
+               } /* end of pixel_bytes == 2 */
+
+               //--------------------------------------------------------------
+               else if (pixel_bytes == BPP8)
+               {
+// GRR TEST:  should work, but needs testing (special 64-bit version of rpng2?)
+                  // GRR NOTE:  no need to combine passes here!
+                  if (((pass == 4) || (pass == 5)) && width)
+                  {
+                     // source is 8-byte RRGGBBAA
+                     // dest is 16-byte RRGGBBAA RRGGBBAA
+                     __asm__ __volatile__ (
+                        "sub  $8, %2             \n\t" // start of last block
+
+                     ".loop8_pass4:              \n\t"
+                        "movq (%1), %%mm0        \n\t" // 7 6 5 4 3 2 1 0
+                        "movq %%mm0, (%2)        \n\t"
+                        "sub  $8, %1             \n\t"
+                        "movq %%mm0, 8(%2)       \n\t"
+                        "sub  $16, %2            \n\t"
+                        "decl %%ecx              \n\t"
+                        "jnz .loop8_pass4        \n\t"
+                        "EMMS                    \n\t" // DONE
+
+                        : "=c" (dummy_value_c),        // output regs (dummy)
+                          "=S" (dummy_value_S),
+                          "=D" (dummy_value_D)
+
+                        : "0" (width),         // ecx  // input regs
+                          "1" (sptr),          // esi/rsi
+                          "2" (dp)             // edi/rdi
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+                        : "%mm0"                       // clobber list
+#endif
+                     );
+                  }
+                  else if (((pass == 2) || (pass == 3)) && width)
+                  {
+                     // source is 8-byte RRGGBBAA
+                     // dest is 32-byte RRGGBBAA RRGGBBAA RRGGBBAA RRGGBBAA
+                     // (recall that expansion is _in place_:  sptr and dp
+                     //  both point at locations within same row buffer)
+                     __asm__ __volatile__ (
+                        "sub  $24, %2            \n\t" // start of last block
+
+                     ".loop8_pass2:              \n\t"
+                        "movq (%1), %%mm0        \n\t" // 7 6 5 4 3 2 1 0
+                        "movq %%mm0, (%2)        \n\t"
+                        "movq %%mm0, 8(%2)       \n\t"
+                        "movq %%mm0, 16(%2)      \n\t"
+                        "sub  $8, %1             \n\t"
+                        "movq %%mm0, 24(%2)      \n\t"
+                        "sub  $32, %2            \n\t"
+                        "decl %%ecx              \n\t"
+                        "jnz .loop8_pass2        \n\t"
+                        "EMMS                    \n\t" // DONE
+
+                        : "=c" (dummy_value_c),        // output regs (dummy)
+                          "=S" (dummy_value_S),
+                          "=D" (dummy_value_D)
+
+                        : "0" (width),         // ecx  // input regs
+                          "1" (sptr),          // esi/rsi
+                          "2" (dp)             // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                        : "%mm0"                       // clobber list
+#endif
+                     );
+                  }
+                  else if (width)  // && ((pass == 0) || (pass == 1))
+                  {
+                     // source is 8-byte RRGGBBAA
+                     // dest is 64-byte RRGGBBAA RRGGBBAA RRGGBBAA RRGGBBAA ...
+                     __asm__ __volatile__ (
+                        "sub  $56, %2            \n\t" // start of last block
+
+                     ".loop8_pass0:              \n\t"
+                        "movq (%1), %%mm0        \n\t" // 7 6 5 4 3 2 1 0
+                        "movq %%mm0, (%2)        \n\t"
+                        "movq %%mm0, 8(%2)       \n\t"
+                        "movq %%mm0, 16(%2)      \n\t"
+                        "movq %%mm0, 24(%2)      \n\t"
+                        "movq %%mm0, 32(%2)      \n\t"
+                        "movq %%mm0, 40(%2)      \n\t"
+                        "movq %%mm0, 48(%2)      \n\t"
+                        "sub  $8, %1             \n\t"
+                        "movq %%mm0, 56(%2)      \n\t"
+                        "sub  $64, %2            \n\t"
+                        "decl %%ecx              \n\t"
+                        "jnz .loop8_pass0        \n\t"
+                        "EMMS                    \n\t" // DONE
+
+                        : "=c" (dummy_value_c),        // output regs (dummy)
+                          "=S" (dummy_value_S),
+                          "=D" (dummy_value_D)
+
+                        : "0" (width),         // ecx  // input regs
+                          "1" (sptr),          // esi/rsi
+                          "2" (dp)             // edi/rdi
+
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+                        : "%mm0"                       // clobber list
+#endif
+                     );
+                  }
+               } /* end of pixel_bytes == 8 */
+
+               //--------------------------------------------------------------
+               else if (pixel_bytes == BPP6)   // why no MMX for this case?
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, BPP6);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, BPP6);
+                        dp -= BPP6;
+                     }
+                     sptr -= BPP6;
+                  }
+               } /* end of pixel_bytes == 6 */
+
+               //--------------------------------------------------------------
+               else
+               {
+                  // ERROR:  SHOULD NEVER BE REACHED
+#if defined(PNG_DEBUG)
+                  png_debug(1, "Internal libpng logic error (GCC "
+                    "png_do_read_interlace() _mmx_supported)\n");
+#endif
+               }
+
+            } // end of _mmx_supported ========================================
+
+            else /* MMX not supported:  use modified C code - takes advantage
+                  *   of inlining of png_memcpy for a constant */
+            {
+               if (pixel_bytes == BPP3)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, BPP3);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, BPP3);
+                        dp -= BPP3;
+                     }
+                     sptr -= BPP3;
+                  }
+               }
+               else if (pixel_bytes == BPP4)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, BPP4);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+#if defined(PNG_DEBUG) && defined(PNG_1_0_X)  // row_buf_size gone in 1.2.x
+                        if (dp < row || dp+3 > row+png_ptr->row_buf_size)
+                        {
+                           printf("dp out of bounds: row=%10p, dp=%10p, "
+                             "rp=%10p\n", row, dp, row+png_ptr->row_buf_size);
+                           printf("row_buf_size=%lu\n", png_ptr->row_buf_size);
+                        }
+#endif
+                        png_memcpy(dp, v, BPP4);
+                        dp -= BPP4;
+                     }
+                     sptr -= BPP4;
+                  }
+               }
+               else if (pixel_bytes == 1)
+               {
+                  for (i = width; i; i--)
+                  {
+                     int j;
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        *dp-- = *sptr;
+                     }
+                     --sptr;
+                  }
+               }
+               else if (pixel_bytes == BPP2)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, BPP2);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, BPP2);
+                        dp -= BPP2;
+                     }
+                     sptr -= BPP2;
+                  }
+               }
+               else if (pixel_bytes == BPP6)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, BPP6);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, BPP6);
+                        dp -= BPP6;
+                     }
+                     sptr -= BPP6;
+                  }
+               }
+               else if (pixel_bytes == BPP8)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, BPP8);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, BPP8);
+                        dp -= BPP8;
+                     }
+                     sptr -= BPP8;
+                  }
+               }
+               else
+               {
+                  // ERROR:  SHOULD NEVER BE REACHED
+#if defined(PNG_DEBUG)
+                  png_debug(1, "Internal libpng logic error (GCC "
+                    "png_do_read_interlace() !_mmx_supported)\n");
+#endif
+               }
+
+            } /* end if (MMX not supported) */
+            break;
+         } /* end default (8-bit or larger) */
+      } /* end switch (row_info->pixel_depth) */
+
+      row_info->width = final_width;
+
+      row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,final_width);
+   }
+
+} /* end png_do_read_interlace() */
+
+#endif /* PNG_HAVE_MMX_READ_INTERLACE */
+#endif /* PNG_READ_INTERLACING_SUPPORTED */
+
+
+
+#if defined(PNG_HAVE_MMX_READ_FILTER_ROW)
+#if defined(PNG_MMX_READ_FILTER_AVG_SUPPORTED)
+
+//===========================================================================//
+//                                                                           //
+//           P N G _ R E A D _ F I L T E R _ R O W _ M M X _ A V G           //
+//                                                                           //
+//===========================================================================//
+
+// Optimized code for PNG Average filter decoder
+
+static void /* PRIVATE */
+png_read_filter_row_mmx_avg(png_row_infop row_info, png_bytep row,
+                            png_bytep prev_row)
+{
+   unsigned FullLength, MMXLength;  // png_uint_32 is actually 64-bit on x86-64
+   int bpp;
+   int dummy_value_a;
+   int dummy_value_c;   // fix 'forbidden register 2 (cx) was spilled' error
+   int dummy_value_d;
+   png_bytep dummy_value_S;
+   png_bytep dummy_value_D;
+   int diff; //     __attribute__((used));
+
+   bpp = (row_info->pixel_depth + 7) >> 3;  // calc number of bytes per pixel
+   FullLength = row_info->rowbytes;         // number of bytes to filter
+
+   __asm__ __volatile__ (
+   "avg_top:                       \n\t"
+      SAVE_GOT_ebx
+      SAVE_r15
+      SAVE_ebp
+      // initialize address pointers and offset
+//pre "movl row, %5                \n\t" // edi/rdi:  ptr to Avg(x)
+      "xorl %%ebx, %%ebx           \n\t" // ebx:  x
+//pre "movl prev_row, %4           \n\t" // esi/rsi:  ptr to Prior(x)
+      "mov  %5, " PDX "            \n\t" // copy of row ptr...
+//pre "subl bpp, " PDX "           \n\t" // (bpp is preloaded into ecx)
+      "sub  " PCX "," PDX "        \n\t" // edx/rdx:  ptr to Raw(x-bpp)
+//pre "movl FullLength, %%eax      \n\t" // bring in via eax...
+      SAVE_FullLength                    // ...but store for later use
+      "xorl %%eax, %%eax           \n\t"
+
+      // Compute the Raw value for the first bpp bytes
+      //    Raw(x) = Avg(x) + (Prior(x)/2)
+   "avg_rlp:                       \n\t"
+      "movb (%4," PBX ",), %%al    \n\t" // load al with Prior(x)
+      "incl %%ebx                  \n\t"
+      "shrb %%al                   \n\t" // divide by 2
+      "addb -1(%5," PBX ",), %%al  \n\t" // add Avg(x); -1 to offset inc ebx
+//pre "cmpl bpp, %%ebx             \n\t" // (bpp is preloaded into ecx)
+      "cmpl %%ecx, %%ebx           \n\t"
+      "movb %%al, -1(%5," PBX ",)  \n\t" // write Raw(x); -1 to offset inc ebx
+      "jb avg_rlp                  \n\t" // mov does not affect flags
+
+      // get # of bytes to alignment (32-bit mask _would_ be good enough
+      // [computing delta], but 32-bit ops are zero-extended on 64-bit, argh)
+      // (if swapped edx and ebp, could do 8-bit or 16-bit mask...FIXME?)
+      "mov  %5, " PBP "            \n\t" // take start of row
+      "add  " PBX "," PBP "        \n\t" // add bpp
+      "add  $0xf, " PBP "          \n\t" // add 7+8 to incr past alignment bdry
+//    "andl $0xfffffff8, %%ebp     \n\t" // mask to alignment boundary (32-bit!)
+      CLEAR_BOTTOM_3_BITS  PBP    "\n\t" // mask to alignment boundary
+      "sub  %5, " PBP "            \n\t" // subtract row ptr again => ebp =
+      "jz avg_go                   \n\t" //  target value of ebx at alignment
+
+      "xorl %%ecx, %%ecx           \n\t"
+
+      // fix alignment
+      // Compute the Raw value for the bytes up to the alignment boundary
+      //    Raw(x) = Avg(x) + ((Raw(x-bpp) + Prior(x))/2)
+   "avg_lp1:                       \n\t"
+      "xorl %%eax, %%eax           \n\t"
+      "movb (%4," PBX ",), %%cl    \n\t" // load cl with Prior(x)
+      "movb (" PDX "," PBX ",), %%al \n\t" // load al with Raw(x-bpp)
+      "addw %%cx, %%ax             \n\t"
+      "incl %%ebx                  \n\t"
+      "shrw %%ax                   \n\t" // divide by 2
+      "addb -1(%5," PBX ",), %%al  \n\t" // add Avg(x); -1 to offset inc ebx
+      "cmpl %%ebp, %%ebx           \n\t" // check if at alignment boundary
+      "movb %%al, -1(%5," PBX ",)  \n\t" // write Raw(x); -1 to offset inc ebx
+      "jb avg_lp1                  \n\t" // repeat until at alignment boundary
+
+   "avg_go:                        \n\t"
+      RESTORE_FullLength "%%eax    \n\t" // FullLength -> eax
+      "movl %%eax, %%ecx           \n\t" // copy -> ecx
+      "subl %%ebx, %%eax           \n\t" // subtract alignment fix
+      "andl $0x00000007, %%eax     \n\t" // calc bytes over mult of 8
+      "subl %%eax, %%ecx           \n\t" // sub over-bytes from original length
+//out "movl %%ecx, MMXLength       \n\t"
+      "movl %%ebp, %%eax           \n\t" // ebp = diff, but no reg constraint(?)
+      RESTORE_ebp                        //  (could swap ebp and edx functions)
+      RESTORE_r15
+      RESTORE_GOT_ebx
+
+// "There is no way for you to specify that an input operand is modified
+// without also specifying it as an output operand."  [makes sense]
+
+// "Unless an output operand has the `&' constraint modifier, GCC may
+// allocate it in the same register as an unrelated input operand, on the
+// assumption the inputs are consumed before the outputs are produced."
+// [trying to _force_ this]
+
+// "`='   Means that this operand is write-only for this instruction:
+//        the previous value is discarded and replaced by output data."
+//        [operand == variable name, presumably]
+
+      // output regs
+      // these are operands 0-1 (originally 0-3):
+      : "=c" (MMXLength),      // %0 -> %0
+        "=a" (diff)            // %3 -> %1
+//      "=S" (dummy_value_S),  // %1 -> GONE
+//      "=D" (dummy_value_D),  // %2 -> GONE
+
+      // input regs
+      // these are operands 2-5 (originally 4-7); two of their constraints say
+      // they must go in same places as operands 0-1 (originally 0-3) above:
+      : "0" (bpp),         // %4 -> %2 ecx
+        "1" (FullLength),  // %7 -> %3 eax
+        "S" (prev_row),    // %5 -> %4 esi/rsi
+        "D" (row)          // %6 -> %5 edi/rdi
+
+      : "%edx"                           // clobber list
+        _CLOBBER_r15
+        _CLOBBER_ebp
+        _CLOBBER_GOT_ebx
+   );
+
+   // now do the math for the rest of the row
+   switch (bpp)
+   {
+      case 3:
+      {
+//       _ShiftBpp = 24;    // == 3 * 8
+//       _ShiftRem = 40;    // == 64 - 24
+
+         __asm__ __volatile__ (
+            // re-init address pointers and offset
+            LOAD_GOT_rbp
+            "movq " AMASK5_3_0 ", %%mm7    \n\t" // _amask5_3_0 -> mm7
+// preload  "movl  diff, %%ecx             \n\t" // ecx:  x = offset to
+                                                 //  alignment boundary
+            "movq " LB_CARRY_MASK ", %%mm5 \n\t" // [interleave for parallel.?]
+// preload  "movl  row, %1                 \n\t" // edi:  Avg(x)
+            "movq " HB_CLEAR_MASK ", %%mm4 \n\t" // _HBClearMask -> mm4
+// preload  "movl  prev_row, %0            \n\t" // esi:  Prior(x)
+            RESTORE_rbp
+
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq  -8(%1," PCX ",), %%mm2 \n\t"// load previous aligned 8 bytes
+                                               // (correct pos. in loop below)
+         "avg_3lp:                        \n\t"
+            "movq  (%1," PCX ",), %%mm0   \n\t" // load mm0 with Avg(x)
+            "movq  %%mm5, %%mm3           \n\t"
+            "psrlq $40, %%mm2             \n\t" // correct position Raw(x-bpp)
+                                                // data
+            "movq  (%0," PCX ",), %%mm1   \n\t" // load mm1 with Prior(x)
+            "movq  %%mm7, %%mm6           \n\t"
+            "pand  %%mm1, %%mm3           \n\t" // get lsb for each prevrow byte
+            "psrlq $1, %%mm1              \n\t" // divide prev_row bytes by 2
+            "pand  %%mm4, %%mm1           \n\t" // clear invalid bit 7 of each
+                                                // byte
+            "paddb %%mm1, %%mm0           \n\t" // add (Prev_row/2) to Avg for
+                                                // each byte
+            // add 1st active group (Raw(x-bpp)/2) to average with LBCarry
+            "movq  %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                                // LBCarrys
+            "pand  %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                                // where both lsb's were == 1
+                                                // (valid only for active group)
+            "psrlq $1, %%mm2              \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2           \n\t" // clear invalid bit 7 of each
+                                                // byte
+            "paddb %%mm1, %%mm2           \n\t" // add LBCarrys to Raw(x-bpp)/2
+                                                // for each byte
+            "pand  %%mm6, %%mm2           \n\t" // leave only Active Group 1
+                                                // bytes to add to Avg
+            "paddb %%mm2, %%mm0           \n\t" // add (Raw/2) + LBCarrys to
+                                                // Avg for each Active byte
+            // add 2nd active group (Raw(x-bpp)/2) to average with _LBCarry
+            "psllq $24, %%mm6             \n\t" // shift the mm6 mask to cover
+                                                // bytes 3-5
+            "movq  %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "psllq $24, %%mm2             \n\t" // shift data to pos. correctly
+            "movq  %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                                // LBCarrys
+            "pand  %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                                // where both lsb's were == 1
+                                                // (valid only for active group)
+            "psrlq $1, %%mm2              \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2           \n\t" // clear invalid bit 7 of each
+                                                // byte
+            "paddb %%mm1, %%mm2           \n\t" // add LBCarrys to Raw(x-bpp)/2
+                                                // for each byte
+            "pand  %%mm6, %%mm2           \n\t" // leave only Active Group 2
+                                                // bytes to add to Avg
+            "paddb %%mm2, %%mm0           \n\t" // add (Raw/2) + LBCarrys to
+                                                // Avg for each Active byte
+
+            // add 3rd active group (Raw(x-bpp)/2) to average with _LBCarry
+            "psllq $24, %%mm6             \n\t" // shift mm6 mask to cover last
+                                                // two bytes
+            "movq  %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "psllq $24, %%mm2             \n\t" // shift data to pos. correctly
+                              // Data need be shifted only once here to
+                              // get the correct x-bpp offset.
+            "movq  %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                                // LBCarrys
+            "pand  %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                                // where both
+                              // lsb's were == 1 (only valid for active group)
+            "psrlq $1, %%mm2              \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2           \n\t" // clear invalid bit 7 of each
+                                                // byte
+            "paddb %%mm1, %%mm2           \n\t" // add LBCarrys to Raw(x-bpp)/2
+                                                // for each byte
+            "pand  %%mm6, %%mm2           \n\t" // leave only Active Group 2
+                                                // bytes to add to Avg
+            "addl  $8, %%ecx              \n\t"
+            "paddb %%mm2, %%mm0           \n\t" // add (Raw/2) + LBCarrys to
+                                                // Avg for each Active byte
+            // now ready to write back to memory
+            "movq  %%mm0, -8(%1," PCX ",) \n\t"
+            // move updated Raw(x) to use as Raw(x-bpp) for next loop
+            "cmpl  %%eax, %%ecx           \n\t" // MMXLength
+            "movq  %%mm0, %%mm2           \n\t" // mov updated Raw(x) to mm2
+            "jb avg_3lp                   \n\t"
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),    // esi/rsi    // input regs
+              "1" (row),         // edi/rdi
+              "2" (diff),        // ecx
+              "3" (MMXLength)    // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm2", "%mm3"   // clobber list
+            , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+         );
+      }
+      break;  // end 3 bpp
+
+      case 4:   // formerly shared with 6 bpp case via _ShiftBpp and _ShiftRem,
+      {         // but loop uses all 8 MMX regs, and psrlq/psllq require 64-bit
+                // mem (PIC/.so problems), MMX reg (none left), or immediate
+//       _ShiftBpp = bpp << 3;        // 32 (psllq)
+//       _ShiftRem = 64 - _ShiftBpp;  // 32 (psrlq)
+
+         __asm__ __volatile__ (
+            LOAD_GOT_rbp
+            "movq " HB_CLEAR_MASK ", %%mm4 \n\t" // _HBClearMask -> mm4
+            "movq " LB_CARRY_MASK ", %%mm5 \n\t" // _LBCarryMask -> mm5
+            // re-init address pointers and offset
+// preload  "movl  diff, %%ecx             \n\t" // ecx:  x = offset to
+                                                 // alignment boundary
+            "movq " AMASK0_8_0 ", %%mm7    \n\t" // _amask0_8_0 -> mm7
+            RESTORE_rbp
+
+            // ... and clear all bytes except for 1st active group
+// preload  "movl  row, %1               \n\t" // edi:  Avg(x)
+            "psrlq $32, %%mm7            \n\t" // was _ShiftRem
+// preload  "movl  prev_row, %0          \n\t" // esi:  Prior(x)
+            "movq  %%mm7, %%mm6          \n\t"
+            "psllq $32, %%mm6            \n\t" // mask for 2nd active group
+
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PCX ",), %%mm2 \n\t" // load previous aligned 8 bytes
+                                             // (we correct pos. in loop below)
+         "avg_4lp:                       \n\t"
+            "movq (%1," PCX ",), %%mm0   \n\t"
+            "psrlq $32, %%mm2            \n\t" // shift data to pos. correctly
+            "movq (%0," PCX ",), %%mm1   \n\t"
+            // add (Prev_row/2) to average
+            "movq %%mm5, %%mm3           \n\t"
+            "pand %%mm1, %%mm3           \n\t" // get lsb for each prev_row byte
+            "psrlq $1, %%mm1             \n\t" // divide prev_row bytes by 2
+            "pand  %%mm4, %%mm1          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm0          \n\t" // add (Prev_row/2) to Avg for
+                                               // each byte
+            // add 1st active group (Raw(x-bpp)/2) to average with _LBCarry
+            "movq %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                               // LBCarrys
+            "pand %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                               // where both
+                              // lsb's were == 1 (only valid for active group)
+            "psrlq $1, %%mm2             \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm2          \n\t" // add LBCarrys to (Raw(x-bpp)/2)
+                                               // for each byte
+            "pand %%mm7, %%mm2           \n\t" // leave only Active Group 1
+                                               // bytes to add to Avg
+            "paddb %%mm2, %%mm0          \n\t" // add (Raw/2) + LBCarrys to Avg
+                                               // for each Active
+                              // byte
+            // add 2nd active group (Raw(x-bpp)/2) to average with _LBCarry
+            "movq %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "psllq $32, %%mm2            \n\t" // shift data to pos. correctly
+            "addl $8, %%ecx              \n\t"
+            "movq %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                               // LBCarrys
+            "pand %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                               // where both
+                              // lsb's were == 1 (only valid for active group)
+            "psrlq $1, %%mm2             \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm2          \n\t" // add LBCarrys to (Raw(x-bpp)/2)
+                                               // for each byte
+            "pand %%mm6, %%mm2           \n\t" // leave only Active Group 2
+                                               // bytes to add to Avg
+            "paddb %%mm2, %%mm0          \n\t" // add (Raw/2) + LBCarrys to
+                                               // Avg for each Active byte
+            "cmpl %%eax, %%ecx           \n\t" // MMXLength
+            // now ready to write back to memory
+            "movq %%mm0, -8(%1," PCX ",) \n\t"
+            // prep Raw(x-bpp) for next loop
+            "movq %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "jb avg_4lp                  \n\t"
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),    // esi/rsi    // input regs
+              "1" (row),         // edi/rdi
+              "2" (diff),        // ecx
+              "3" (MMXLength)    // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm2", "%mm3"   // clobber list
+            , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+         );
+      }
+      break;  // end 4 bpp
+
+      case 1:
+      {
+         __asm__ __volatile__ (
+            // re-init address pointers and offset
+// preload  "movl diff, %%ecx            \n\t" // ecx: x = offset to align. bdry
+// preload  "movl row, %1                \n\t" // edi/rdi:  Avg(x)
+// preload  "movl FullLength, %%eax      \n\t"
+            "cmpl %%eax, %%ecx           \n\t" // test if offset at end of array
+            "jnb avg_1end                \n\t"
+
+            SAVE_ebp
+
+            // do Avg decode for remaining bytes
+// preload  "movl prev_row, %0           \n\t" // esi/rsi:  Prior(x)
+            "mov  %1, " PBP "            \n\t" // copy of row pointer...
+            "dec  " PBP "                \n\t" // ebp/rbp:  Raw(x-bpp)
+            "xorl %%edx, %%edx           \n\t" // zero edx before using dl & dx
+                                               //  in loop below
+            SAVE_GOT_ebx
+
+         "avg_1lp:                       \n\t"
+            // Raw(x) = Avg(x) + ((Raw(x-bpp) + Prior(x))/2)
+            "xorl %%ebx, %%ebx           \n\t"
+            "movb (%0," PCX ",), %%dl    \n\t" // load dl with Prior(x)
+            "movb (" PBP "," PCX ",), %%bl \n\t" // load bl with Raw(x-bpp)
+            "addw %%dx, %%bx             \n\t"
+            "incl %%ecx                  \n\t"
+            "shrw %%bx                   \n\t" // divide by 2
+            "addb -1(%1," PCX ",), %%bl  \n\t" // add Avg(x); -1 to offset
+                                               // inc ecx
+            "cmpl %%eax, %%ecx           \n\t" // check if at end of array
+            "movb %%bl, -1(%1," PCX ",)  \n\t" // write back Raw(x);
+                         // mov does not affect flags; -1 to offset inc ecx
+            "jb avg_1lp                  \n\t"
+
+            RESTORE_GOT_ebx
+            RESTORE_ebp
+
+         "avg_1end:                      \n\t"
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),    // esi/rsi    // input regs
+              "1" (row),         // edi/rdi
+              "2" (diff),        // ecx
+              "3" (FullLength)   // eax
+
+            : "%edx"                           // clobber list
+              _CLOBBER_GOT_ebx
+              _CLOBBER_ebp
+         );
+      }
+      return;  // end 1 bpp
+
+      case 2:
+      {
+//       _ShiftBpp = 16;   // == 2 * 8
+//       _ShiftRem = 48;   // == 64 - _ShiftBpp
+
+         __asm__ __volatile__ (
+            LOAD_GOT_rbp
+            // load (former) _ActiveMask
+            "movq " AMASK6_2_0 ", %%mm7    \n\t" // _amask6_2_0 -> mm7
+            // re-init address pointers and offset
+// preload  "movl  diff, %%ecx             \n\t" // ecx:  x = offset to
+                                                 // alignment boundary
+            "movq " LB_CARRY_MASK ", %%mm5 \n\t" // _LBCarryMask -> mm5
+// preload  "movl  row, %1                 \n\t" // edi:  Avg(x)
+            "movq " HB_CLEAR_MASK ", %%mm4 \n\t" // _HBClearMask -> mm4
+// preload  "movl  prev_row, %0            \n\t" // esi:  Prior(x)
+            RESTORE_rbp
+
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PCX ",), %%mm2 \n\t" // load previous aligned 8 bytes
+                                             // (we correct pos. in loop below)
+         "avg_2lp:                       \n\t"
+            "movq (%1," PCX ",), %%mm0   \n\t"
+            "psrlq $48, %%mm2            \n\t" // shift data to pos. correctly
+            "movq (%0," PCX ",), %%mm1   \n\t" //  (GRR BUGFIX:  was psllq)
+            // add (Prev_row/2) to average
+            "movq %%mm5, %%mm3           \n\t"
+            "pand %%mm1, %%mm3           \n\t" // get lsb for each prev_row byte
+            "psrlq $1, %%mm1             \n\t" // divide prev_row bytes by 2
+            "pand  %%mm4, %%mm1          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "movq %%mm7, %%mm6           \n\t"
+            "paddb %%mm1, %%mm0          \n\t" // add (Prev_row/2) to Avg for
+                                               // each byte
+
+            // add 1st active group (Raw(x-bpp)/2) to average with _LBCarry
+            "movq %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                               // LBCarrys
+            "pand %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                               // where both
+                                               // lsb's were == 1 (only valid
+                                               // for active group)
+            "psrlq $1, %%mm2             \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm2          \n\t" // add LBCarrys to (Raw(x-bpp)/2)
+                                               // for each byte
+            "pand %%mm6, %%mm2           \n\t" // leave only Active Group 1
+                                               // bytes to add to Avg
+            "paddb %%mm2, %%mm0          \n\t" // add (Raw/2) + LBCarrys to Avg
+                                               // for each Active byte
+
+            // add 2nd active group (Raw(x-bpp)/2) to average with _LBCarry
+            "psllq $16, %%mm6            \n\t" // shift the mm6 mask to cover
+                                               // bytes 2 & 3
+            "movq %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "psllq $16, %%mm2            \n\t" // shift data to pos. correctly
+            "movq %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                               // LBCarrys
+            "pand %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                               // where both
+                                               // lsb's were == 1 (only valid
+                                               // for active group)
+            "psrlq $1, %%mm2             \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm2          \n\t" // add LBCarrys to (Raw(x-bpp)/2)
+                                               // for each byte
+            "pand %%mm6, %%mm2           \n\t" // leave only Active Group 2
+                                               // bytes to add to Avg
+            "paddb %%mm2, %%mm0          \n\t" // add (Raw/2) + LBCarrys to
+                                               // Avg for each Active byte
+
+            // add 3rd active group (Raw(x-bpp)/2) to average with _LBCarry
+            "psllq $16, %%mm6            \n\t" // shift the mm6 mask to cover
+                                               // bytes 4 & 5
+            "movq %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "psllq $16, %%mm2            \n\t" // shift data to pos. correctly
+            "movq %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                               // LBCarrys
+            "pand %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                               // where both lsb's were == 1
+                                               // (only valid for active group)
+            "psrlq $1, %%mm2             \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm2          \n\t" // add LBCarrys to (Raw(x-bpp)/2)
+                                               // for each byte
+            "pand %%mm6, %%mm2           \n\t" // leave only Active Group 2
+                                               // bytes to add to Avg
+            "paddb %%mm2, %%mm0          \n\t" // add (Raw/2) + LBCarrys to
+                                               // Avg for each Active byte
+
+            // add 4th active group (Raw(x-bpp)/2) to average with _LBCarry
+            "psllq $16, %%mm6            \n\t" // shift the mm6 mask to cover
+                                               // bytes 6 & 7
+            "movq %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "psllq $16, %%mm2            \n\t" // shift data to pos. correctly
+            "addl $8, %%ecx              \n\t"
+            "movq %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                               // LBCarrys
+            "pand %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                               // where both
+                                               // lsb's were == 1 (only valid
+                                               // for active group)
+            "psrlq $1, %%mm2             \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm2          \n\t" // add LBCarrys to (Raw(x-bpp)/2)
+                                               // for each byte
+            "pand %%mm6, %%mm2           \n\t" // leave only Active Group 2
+                                               // bytes to add to Avg
+            "paddb %%mm2, %%mm0          \n\t" // add (Raw/2) + LBCarrys to
+                                               // Avg for each Active byte
+            "cmpl %%eax, %%ecx           \n\t" // MMXLength
+            // now ready to write back to memory
+            "movq %%mm0, -8(%1," PCX ",) \n\t"
+            // prep Raw(x-bpp) for next loop
+            "movq %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "jb avg_2lp                  \n\t"
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),    // esi/rsi    // input regs
+              "1" (row),         // edi/rdi
+              "2" (diff),        // ecx
+              "3" (MMXLength)    // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm2", "%mm3"   // clobber list
+            , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+         );
+      }
+      break;  // end 2 bpp
+
+      case 6:   // formerly shared with 4 bpp case (see comments there)
+      {
+//       _ShiftBpp = bpp << 3;        // 48 (psllq)
+//       _ShiftRem = 64 - _ShiftBpp;  // 16 (psrlq)
+
+         __asm__ __volatile__ (
+            LOAD_GOT_rbp
+            "movq " HB_CLEAR_MASK ", %%mm4 \n\t" // _HBClearMask -> mm4
+            "movq " LB_CARRY_MASK ", %%mm5 \n\t" // _LBCarryMask -> mm5
+            // re-init address pointers and offset
+// preload  "movl  diff, %%ecx             \n\t" // ecx:  x = offset to
+                                                 // alignment boundary
+            "movq " AMASK0_8_0 ", %%mm7    \n\t" // _amask0_8_0 -> mm7
+            RESTORE_rbp
+
+            // ... and clear all bytes except for 1st active group
+// preload  "movl  row, %1               \n\t" // edi:  Avg(x)
+            "psrlq $16, %%mm7            \n\t"
+// preload  "movl  prev_row, %0          \n\t" // esi:  Prior(x)
+            "movq  %%mm7, %%mm6          \n\t"
+            "psllq $48, %%mm6            \n\t" // mask for 2nd active group
+
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PCX ",), %%mm2 \n\t" // load previous aligned 8 bytes
+                                             // (we correct pos. in loop below)
+         "avg_6lp:                       \n\t"
+            "movq (%1," PCX ",), %%mm0   \n\t"
+            "psrlq $16, %%mm2            \n\t" // shift data to pos. correctly
+            "movq (%0," PCX ",), %%mm1   \n\t"
+            // add (Prev_row/2) to average
+            "movq %%mm5, %%mm3           \n\t"
+            "pand %%mm1, %%mm3           \n\t" // get lsb for each prev_row byte
+            "psrlq $1, %%mm1             \n\t" // divide prev_row bytes by 2
+            "pand  %%mm4, %%mm1          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm0          \n\t" // add (Prev_row/2) to Avg for
+                                               // each byte
+            // add 1st active group (Raw(x-bpp)/2) to average with _LBCarry
+            "movq %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                               // LBCarrys
+            "pand %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                               // where both
+                              // lsb's were == 1 (only valid for active group)
+            "psrlq $1, %%mm2             \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm2          \n\t" // add LBCarrys to (Raw(x-bpp)/2)
+                                               // for each byte
+            "pand %%mm7, %%mm2           \n\t" // leave only Active Group 1
+                                               // bytes to add to Avg
+            "paddb %%mm2, %%mm0          \n\t" // add (Raw/2) + LBCarrys to Avg
+                                               // for each Active
+                              // byte
+            // add 2nd active group (Raw(x-bpp)/2) to average with _LBCarry
+            "movq %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "psllq $48, %%mm2            \n\t" // shift data to pos. correctly
+            "addl $8, %%ecx              \n\t"
+            "movq %%mm3, %%mm1           \n\t" // now use mm1 for getting
+                                               // LBCarrys
+            "pand %%mm2, %%mm1           \n\t" // get LBCarrys for each byte
+                                               // where both
+                              // lsb's were == 1 (only valid for active group)
+            "psrlq $1, %%mm2             \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm2          \n\t" // clear invalid bit 7 of each
+                                               // byte
+            "paddb %%mm1, %%mm2          \n\t" // add LBCarrys to (Raw(x-bpp)/2)
+                                               // for each byte
+            "pand %%mm6, %%mm2           \n\t" // leave only Active Group 2
+                                               // bytes to add to Avg
+            "paddb %%mm2, %%mm0          \n\t" // add (Raw/2) + LBCarrys to
+                                               // Avg for each Active byte
+            "cmpl %%eax, %%ecx           \n\t" // MMXLength
+            // now ready to write back to memory
+            "movq %%mm0, -8(%1," PCX ",) \n\t"
+            // prep Raw(x-bpp) for next loop
+            "movq %%mm0, %%mm2           \n\t" // mov updated Raws to mm2
+            "jb avg_6lp                  \n\t"
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),    // esi/rsi    // input regs
+              "1" (row),         // edi/rdi
+              "2" (diff),        // ecx
+              "3" (MMXLength)    // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm2", "%mm3"   // clobber list
+            , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+         );
+      }
+      break;  // end 6 bpp
+
+      case 8:
+      {
+         __asm__ __volatile__ (
+            // re-init address pointers and offset
+// preload  "movl  diff, %%ecx             \n\t" // ecx:  x = offset to
+                                                 // alignment boundary
+            LOAD_GOT_rbp
+            "movq " LB_CARRY_MASK ", %%mm5 \n\t" // [interleave for parallel.?]
+// preload  "movl  row, %1                 \n\t" // edi:  Avg(x)
+            "movq " HB_CLEAR_MASK ", %%mm4 \n\t" // _HBClearMask -> mm4
+// preload  "movl  prev_row, %0            \n\t" // esi:  Prior(x)
+            RESTORE_rbp
+
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PCX ",), %%mm2 \n\t" // load previous aligned 8 bytes
+                                      // (NO NEED to correct pos. in loop below)
+
+         "avg_8lp:                       \n\t"
+            "movq (%1," PCX ",), %%mm0   \n\t"
+            "movq %%mm5, %%mm3           \n\t"
+            "movq (%0," PCX ",), %%mm1   \n\t"
+            "addl $8, %%ecx              \n\t"
+            "pand %%mm1, %%mm3           \n\t" // get lsb for each prev_row byte
+            "psrlq $1, %%mm1             \n\t" // divide prev_row bytes by 2
+            "pand %%mm2, %%mm3           \n\t" // get LBCarrys for each byte
+                                               //  where both lsb's were == 1
+            "psrlq $1, %%mm2             \n\t" // divide raw bytes by 2
+            "pand  %%mm4, %%mm1          \n\t" // clear invalid bit 7, each byte
+            "paddb %%mm3, %%mm0          \n\t" // add LBCarrys to Avg, each byte
+            "pand  %%mm4, %%mm2          \n\t" // clear invalid bit 7, each byte
+            "paddb %%mm1, %%mm0          \n\t" // add (Prev_row/2) to Avg, each
+            "paddb %%mm2, %%mm0          \n\t" // add (Raw/2) to Avg for each
+            "cmpl %%eax, %%ecx           \n\t" // MMXLength
+            "movq %%mm0, -8(%1," PCX ",) \n\t"
+            "movq %%mm0, %%mm2           \n\t" // reuse as Raw(x-bpp)
+            "jb avg_8lp                  \n\t"
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),    // esi/rsi    // input regs
+              "1" (row),         // edi/rdi
+              "2" (diff),        // ecx
+              "3" (MMXLength)    // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm2"           // clobber list
+            , "%mm3", "%mm4", "%mm5"
+#endif
+         );
+      }
+      break;  // end 8 bpp
+
+      default:                // bpp != 1,2,3,4,6,8:  doesn't exist
+      {
+         // ERROR:  SHOULD NEVER BE REACHED
+#if defined(PNG_DEBUG)
+         png_debug(1, "Internal libpng logic error (GCC "
+           "png_read_filter_row_mmx_avg())\n");
+#endif
+      }
+      break;
+
+   } // end switch (bpp)
+
+   __asm__ __volatile__ (
+      // MMX acceleration complete; now do clean-up
+      // check if any remaining bytes left to decode
+//pre "movl FullLength, %%edx      \n\t"
+//pre "movl MMXLength, %%eax       \n\t" // eax:  x == offset bytes after MMX
+//pre "movl row, %2                \n\t" // edi:  Avg(x)
+      "cmpl %%edx, %%eax           \n\t" // test if offset at end of array
+      "jnb avg_end                 \n\t"
+
+      SAVE_ebp
+
+      // do Avg decode for remaining bytes
+//pre "movl prev_row, %1           \n\t" // esi:  Prior(x)
+      "mov  %2, " PBP "            \n\t" // copy of row pointer...
+//pre "subl bpp, " PBP "           \n\t" // (bpp is preloaded into ecx)
+      "sub  " PCX "," PBP "        \n\t" // ebp:  Raw(x-bpp)
+      "xorl %%ecx, %%ecx           \n\t" // zero ecx before using cl & cx below
+
+      SAVE_GOT_ebx
+
+   "avg_lp2:                       \n\t"
+      // Raw(x) = Avg(x) + ((Raw(x-bpp) + Prior(x))/2)
+      "xorl %%ebx, %%ebx           \n\t"
+      "movb (%1," PAX ",), %%cl    \n\t" // load cl with Prior(x)
+      "movb (" PBP "," PAX ",), %%bl \n\t" // load bl with Raw(x-bpp)
+      "addw %%cx, %%bx             \n\t"
+      "incl %%eax                  \n\t"
+      "shrw %%bx                   \n\t" // divide by 2
+      "addb -1(%2," PAX ",), %%bl  \n\t" // add Avg(x); -1 to offset inc eax
+      "cmpl %%edx, %%eax           \n\t" // check if at end of array
+      "movb %%bl, -1(%2," PAX ",)  \n\t" // write back Raw(x) [mov does not
+      "jb avg_lp2                  \n\t" //  affect flags; -1 to offset inc eax]
+
+      RESTORE_GOT_ebx
+      RESTORE_ebp
+
+   "avg_end:                       \n\t"
+      "EMMS                        \n\t" // end MMX; prep for poss. FP instrs.
+
+      : "=c" (dummy_value_c),            // output regs (dummy)
+        "=S" (dummy_value_S),
+        "=D" (dummy_value_D),
+        "=a" (dummy_value_a),
+        "=d" (dummy_value_d)
+
+      : "0" (bpp),         // ecx        // input regs
+        "1" (prev_row),    // esi/rsi
+        "2" (row),         // edi/rdi
+        "3" (MMXLength),   // eax
+        "4" (FullLength)   // edx
+
+      CLOB_COLON_ebx_ebp                 // clobber list
+        CLOBBER_GOT_ebx
+        CLOB_COMMA_ebx_ebp
+        CLOBBER_ebp
+   );
+
+} /* end png_read_filter_row_mmx_avg() */
+
+#endif /* PNG_MMX_READ_FILTER_AVG_SUPPORTED */
+
+
+
+#if defined(PNG_MMX_READ_FILTER_PAETH_SUPPORTED)
+#if defined(PNG_x86_64_USE_GOTPCREL) || defined(PNG_THREAD_UNSAFE_OK)
+
+//===========================================================================//
+//                                                                           //
+//         P N G _ R E A D _ F I L T E R _ R O W _ M M X _ P A E T H         //
+//                                                                           //
+//===========================================================================//
+
+// Optimized code for PNG Paeth filter decoder
+
+static void /* PRIVATE */
+png_read_filter_row_mmx_paeth(png_row_infop row_info, png_bytep row,
+                              png_bytep prev_row)
+{
+   unsigned FullLength, MMXLength;  // png_uint_32 is actually 64-bit on x86-64
+   int bpp;
+   int dummy_value_a;
+   int dummy_value_c;   // fix 'forbidden register 2 (cx) was spilled' error
+   int dummy_value_d;
+   png_charp dummy_value_S;
+   png_charp dummy_value_D;
+   int diff; //     __attribute__((used));
+
+   bpp = (row_info->pixel_depth + 7) >> 3;  // calc number of bytes per pixel
+   FullLength = row_info->rowbytes;         // number of bytes to filter
+
+   __asm__ __volatile__ (
+      SAVE_GOT_ebx
+      SAVE_r15
+      SAVE_ebp
+//pre "movl row, %2                \n\t" // edi/rdi
+      "xorl %%ebx, %%ebx           \n\t" // ebx:  x offset
+//pre "movl prev_row, %1           \n\t" // esi/rsi
+      "xorl %%edx, %%edx           \n\t" // edx:  x-bpp offset
+//pre "movl FullLength, %%eax      \n\t" // bring in via eax...
+      SAVE_FullLength                    // ...but store for later use
+      "xorl %%eax, %%eax           \n\t"
+
+      // Compute the Raw value for the first bpp bytes
+      // Note: the formula works out to be always
+      //   Paeth(x) = Raw(x) + Prior(x)      where x < bpp
+   "paeth_rlp:                     \n\t"
+      "movb (%2," PBX ",), %%al    \n\t"
+      "addb (%1," PBX ",), %%al    \n\t"
+      "incl %%ebx                  \n\t"
+//pre "cmpl bpp, %%ebx             \n\t" (bpp is preloaded into ecx)
+      "cmpl %%ecx, %%ebx           \n\t"
+      "movb %%al, -1(%2," PBX ",)  \n\t"
+      "jb paeth_rlp                \n\t"
+
+      // get # of bytes to alignment (note:  computing _delta_ of two pointers,
+      // so hereafter %%ebp is sufficient even on 64-bit)
+      "mov  %2, " PBP "            \n\t" // take start of row
+      "add  " PBX "," PBP "        \n\t" // add bpp
+      "add  $0xf, " PBP "          \n\t" // add 7+8 to incr past alignment bdry
+//    "andl $0xfffffff8, %%ebp     \n\t" // mask to alignment boundary (32-bit!)
+      CLEAR_BOTTOM_3_BITS  PBP    "\n\t" // mask to alignment boundary
+      "sub  %2, " PBP "            \n\t" // subtract row ptr again => ebp =
+      "jz paeth_go                 \n\t" //  target value of ebx at alignment
+
+      "xorl %%ecx, %%ecx           \n\t"
+
+      SAVE_r11_r12_r13
+
+      // fix alignment
+   "paeth_lp1:                     \n\t"
+      "xorl %%eax, %%eax           \n\t"
+      // pav = p - a = (a + b - c) - a = b - c
+      "movb (%1," PBX ",), %%al    \n\t" // load Prior(x) into al
+      "movb (%1," PDX ",), %%cl    \n\t" // load Prior(x-bpp) into cl
+      "subl %%ecx, %%eax           \n\t" // subtract Prior(x-bpp)
+      "movl %%eax, " pa_TEMP "     \n\t" // Save pav for later use
+      "xorl %%eax, %%eax           \n\t"
+      // pbv = p - b = (a + b - c) - b = a - c
+      "movb (%2," PDX ",), %%al    \n\t" // load Raw(x-bpp) into al
+      "subl %%ecx, %%eax           \n\t" // subtract Prior(x-bpp)
+      "movl %%eax, %%ecx           \n\t"
+      // pcv = p - c = (a + b - c) - c = (a - c) + (b - c) = pav + pbv
+      "addl " pa_TEMP ", %%eax     \n\t" // pcv = pav + pbv
+      // pc = abs(pcv)
+      "testl $0x80000000, %%eax    \n\t"
+      "jz paeth_pca                \n\t"
+      "negl %%eax                  \n\t" // reverse sign of neg values
+
+   "paeth_pca:                     \n\t"
+      "movl %%eax, " pc_TEMP "     \n\t" // save pc for later use
+      // pb = abs(pbv)
+      "testl $0x80000000, %%ecx    \n\t"
+      "jz paeth_pba                \n\t"
+      "negl %%ecx                  \n\t" // reverse sign of neg values
+
+   "paeth_pba:                     \n\t"
+      "movl %%ecx, " pb_TEMP "     \n\t" // save pb for later use
+      // pa = abs(pav)
+      "movl " pa_TEMP ", %%eax     \n\t"
+      "testl $0x80000000, %%eax    \n\t"
+      "jz paeth_paa                \n\t"
+      "negl %%eax                  \n\t" // reverse sign of neg values
+
+   "paeth_paa:                     \n\t"
+      "movl %%eax, " pa_TEMP "     \n\t" // save pa for later use
+      // test if pa <= pb
+      "cmpl %%ecx, %%eax           \n\t"
+      "jna paeth_abb               \n\t"
+      // pa > pb; now test if pb <= pc
+      "cmpl " pc_TEMP ", %%ecx     \n\t"
+      "jna paeth_bbc               \n\t"
+      // pb > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+      "movb (%1," PDX ",), %%cl    \n\t" // load Prior(x-bpp) into cl
+      "jmp paeth_paeth             \n\t"
+
+   "paeth_bbc:                     \n\t"
+      // pb <= pc; Raw(x) = Paeth(x) + Prior(x)
+      "movb (%1," PBX ",), %%cl    \n\t" // load Prior(x) into cl
+      "jmp paeth_paeth             \n\t"
+
+   "paeth_abb:                     \n\t"
+      // pa <= pb; now test if pa <= pc
+      "cmpl " pc_TEMP ", %%eax     \n\t"
+      "jna paeth_abc               \n\t"
+      // pa > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+      "movb (%1," PDX ",), %%cl    \n\t" // load Prior(x-bpp) into cl
+      "jmp paeth_paeth             \n\t"
+
+   "paeth_abc:                     \n\t"
+      // pa <= pc; Raw(x) = Paeth(x) + Raw(x-bpp)
+      "movb (%2," PDX ",), %%cl    \n\t" // load Raw(x-bpp) into cl
+
+   "paeth_paeth:                   \n\t"
+      "incl %%ebx                  \n\t"
+      "incl %%edx                  \n\t"
+      // Raw(x) = (Paeth(x) + Paeth_Predictor( a, b, c )) mod 256
+      "addb %%cl, -1(%2," PBX ",)  \n\t"
+      "cmpl %%ebp, %%ebx           \n\t"
+      "jb paeth_lp1                \n\t"
+
+      RESTORE_r11_r12_r13
+
+   "paeth_go:                      \n\t"
+      RESTORE_FullLength "%%ecx    \n\t" // FullLength -> ecx
+      "movl %%ecx, %%eax           \n\t"
+      "subl %%ebx, %%eax           \n\t" // subtract alignment fix
+      "andl $0x00000007, %%eax     \n\t" // calc bytes over mult of 8
+      "subl %%eax, %%ecx           \n\t" // drop over bytes from original length
+//out "movl %%ecx, MMXLength       \n\t"
+      "movl %%ebp, %%eax           \n\t" // ebp = diff, but no reg constraint(?)
+      RESTORE_ebp                        //  (could swap ebp and edx functions)
+      RESTORE_r15
+      RESTORE_GOT_ebx
+
+      : "=c" (MMXLength),                // output regs
+        "=S" (dummy_value_S),
+        "=D" (dummy_value_D),
+        "=a" (diff)
+
+      : "0" (bpp),         // ecx        // input regs
+        "1" (prev_row),    // esi/rsi
+        "2" (row),         // edi/rdi
+        "3" (FullLength)   // eax
+
+      : "%edx"                           // clobber list
+        _CLOBBER_r11_r12_r13
+        _CLOBBER_r15
+        _CLOBBER_ebp
+        _CLOBBER_GOT_ebx
+   );
+
+   // now do the math for the rest of the row
+   switch (bpp)
+   {
+      case 3:
+      {
+//       _ShiftBpp = 24;    // == bpp * 8
+//       _ShiftRem = 40;    // == 64 - _ShiftBpp
+
+         __asm__ __volatile__ (
+            LOAD_GOT_rbp
+// preload  "movl diff, %%ecx            \n\t"
+// preload  "movl row, %1                \n\t" // edi/rdi
+// preload  "movl prev_row, %0           \n\t" // esi/rsi
+            "pxor %%mm0, %%mm0           \n\t"
+
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PCX ",), %%mm1 \n\t"
+         "paeth_3lp:                     \n\t"
+            "psrlq $40, %%mm1            \n\t" // shift last 3 bytes to 1st
+                                               // 3 bytes
+            "movq (%0," PCX ",), %%mm2   \n\t" // load b=Prior(x)
+            "punpcklbw %%mm0, %%mm1      \n\t" // unpack High bytes of a
+            "movq -8(%0," PCX ",), %%mm3 \n\t" // prep c=Prior(x-bpp) bytes
+            "punpcklbw %%mm0, %%mm2      \n\t" // unpack High bytes of b
+            "psrlq $40, %%mm3            \n\t" // shift last 3 bytes to 1st
+                                               // 3 bytes
+            // pav = p - a = (a + b - c) - a = b - c
+            "movq %%mm2, %%mm4           \n\t"
+            "punpcklbw %%mm0, %%mm3      \n\t" // unpack High bytes of c
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movq %%mm1, %%mm5           \n\t"
+            "psubw %%mm3, %%mm4          \n\t"
+            "pxor %%mm7, %%mm7           \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            "movq %%mm4, %%mm6           \n\t"
+            "psubw %%mm3, %%mm5          \n\t"
+
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            "pcmpgtw %%mm4, %%mm0        \n\t" // create mask pav bytes < 0
+            "paddw %%mm5, %%mm6          \n\t"
+            "pand %%mm4, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "pcmpgtw %%mm5, %%mm7        \n\t" // create mask pbv bytes < 0
+            "psubw %%mm0, %%mm4          \n\t"
+            "pand %%mm5, %%mm7           \n\t" // only pbv bytes < 0 in mm0
+            "psubw %%mm0, %%mm4          \n\t"
+            "psubw %%mm7, %%mm5          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "pcmpgtw %%mm6, %%mm0        \n\t" // create mask pcv bytes < 0
+            "pand %%mm6, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm7, %%mm5          \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            //  test pa <= pb
+            "movq %%mm4, %%mm7           \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            "pcmpgtw %%mm5, %%mm7        \n\t" // pa > pb?
+            "movq %%mm7, %%mm0           \n\t"
+            // use mm7 mask to merge pa & pb
+            "pand %%mm7, %%mm5           \n\t"
+            // use mm0 mask copy to merge a & b
+            "pand %%mm0, %%mm2           \n\t"
+            "pandn %%mm4, %%mm7          \n\t"
+            "pandn %%mm1, %%mm0          \n\t"
+            "paddw %%mm5, %%mm7          \n\t"
+            "paddw %%mm2, %%mm0          \n\t"
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            "pcmpgtw %%mm6, %%mm7        \n\t" // pab > pc?
+            "pxor %%mm1, %%mm1           \n\t"
+            "pand %%mm7, %%mm3           \n\t"
+            "pandn %%mm0, %%mm7          \n\t"
+            "paddw %%mm3, %%mm7          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "packuswb %%mm1, %%mm7       \n\t"
+            "movq (%0," PCX ",), %%mm3   \n\t" // load c=Prior(x-bpp)
+            "pand " AMASK5_3_0 ", %%mm7  \n\t" // _amask5_3_0 (was _ActiveMask)
+            "movq %%mm3, %%mm2           \n\t" // load b=Prior(x) step 1
+            "paddb (%1," PCX ",), %%mm7  \n\t" // add Paeth predictor + Raw(x)
+            "punpcklbw %%mm0, %%mm3      \n\t" // unpack High bytes of c
+            "movq %%mm7, (%1," PCX ",)   \n\t" // write back updated value
+            "movq %%mm7, %%mm1           \n\t" // now mm1 will be used as
+                                               // Raw(x-bpp)
+            // now do Paeth for 2nd set of bytes (3-5)
+            "psrlq $24, %%mm2            \n\t" // load b=Prior(x) step 2
+            "punpcklbw %%mm0, %%mm1      \n\t" // unpack High bytes of a
+            "pxor %%mm7, %%mm7           \n\t"
+            "punpcklbw %%mm0, %%mm2      \n\t" // unpack High bytes of b
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movq %%mm1, %%mm5           \n\t"
+            // pav = p - a = (a + b - c) - a = b - c
+            "movq %%mm2, %%mm4           \n\t"
+            "psubw %%mm3, %%mm5          \n\t"
+            "psubw %%mm3, %%mm4          \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) =
+            //       pav + pbv = pbv + pav
+            "movq %%mm5, %%mm6           \n\t"
+            "paddw %%mm4, %%mm6          \n\t"
+
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            "pcmpgtw %%mm5, %%mm0        \n\t" // create mask pbv bytes < 0
+            "pcmpgtw %%mm4, %%mm7        \n\t" // create mask pav bytes < 0
+            "pand %%mm5, %%mm0           \n\t" // only pbv bytes < 0 in mm0
+            "pand %%mm4, %%mm7           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm0, %%mm5          \n\t"
+            "psubw %%mm7, %%mm4          \n\t"
+            "psubw %%mm0, %%mm5          \n\t"
+            "psubw %%mm7, %%mm4          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "pcmpgtw %%mm6, %%mm0        \n\t" // create mask pcv bytes < 0
+            "pand %%mm6, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm0, %%mm6          \n\t"
+            //  test pa <= pb
+            "movq %%mm4, %%mm7           \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            "pcmpgtw %%mm5, %%mm7        \n\t" // pa > pb?
+            "movq %%mm7, %%mm0           \n\t"
+            // use mm7 mask to merge pa & pb
+            "pand %%mm7, %%mm5           \n\t"
+            // use mm0 mask copy to merge a & b
+            "pand %%mm0, %%mm2           \n\t"
+            "pandn %%mm4, %%mm7          \n\t"
+            "pandn %%mm1, %%mm0          \n\t"
+            "paddw %%mm5, %%mm7          \n\t"
+            "paddw %%mm2, %%mm0          \n\t"
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            "pcmpgtw %%mm6, %%mm7        \n\t" // pab > pc?
+            "movq (%0," PCX ",), %%mm2   \n\t" // load b=Prior(x)
+            "pand %%mm7, %%mm3           \n\t"
+            "pandn %%mm0, %%mm7          \n\t"
+            "pxor %%mm1, %%mm1           \n\t"
+            "paddw %%mm3, %%mm7          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "packuswb %%mm1, %%mm7       \n\t"
+            "movq %%mm2, %%mm3           \n\t" // load c=Prior(x-bpp) step 1
+            "pand " AMASK5_3_0 ", %%mm7  \n\t" // _amask5_3_0 (was _ActiveMask)
+            "punpckhbw %%mm0, %%mm2      \n\t" // unpack High bytes of b
+            "psllq $24, %%mm7            \n\t" // shift bytes to 2nd group of
+                                               // 3 bytes
+             // pav = p - a = (a + b - c) - a = b - c
+            "movq %%mm2, %%mm4           \n\t"
+            "paddb (%1," PCX ",), %%mm7  \n\t" // add Paeth predictor + Raw(x)
+            "psllq $24, %%mm3            \n\t" // load c=Prior(x-bpp) step 2
+            "movq %%mm7, (%1," PCX ",)   \n\t" // write back updated value
+            "movq %%mm7, %%mm1           \n\t"
+            "punpckhbw %%mm0, %%mm3      \n\t" // unpack High bytes of c
+            "psllq $24, %%mm1            \n\t" // shift bytes (was _ShiftBpp)
+                                    // now mm1 will be used as Raw(x-bpp)
+            // now do Paeth for 3rd, and final, set of bytes (6-7)
+            "pxor %%mm7, %%mm7           \n\t"
+            "punpckhbw %%mm0, %%mm1      \n\t" // unpack High bytes of a
+            "psubw %%mm3, %%mm4          \n\t"
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movq %%mm1, %%mm5           \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            "movq %%mm4, %%mm6           \n\t"
+            "psubw %%mm3, %%mm5          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "paddw %%mm5, %%mm6          \n\t"
+
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            "pcmpgtw %%mm4, %%mm0        \n\t" // create mask pav bytes < 0
+            "pcmpgtw %%mm5, %%mm7        \n\t" // create mask pbv bytes < 0
+            "pand %%mm4, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "pand %%mm5, %%mm7           \n\t" // only pbv bytes < 0 in mm0
+            "psubw %%mm0, %%mm4          \n\t"
+            "psubw %%mm7, %%mm5          \n\t"
+            "psubw %%mm0, %%mm4          \n\t"
+            "psubw %%mm7, %%mm5          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "pcmpgtw %%mm6, %%mm0        \n\t" // create mask pcv bytes < 0
+            "pand %%mm6, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm0, %%mm6          \n\t"
+            //  test pa <= pb
+            "movq %%mm4, %%mm7           \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            "pcmpgtw %%mm5, %%mm7        \n\t" // pa > pb?
+            "movq %%mm7, %%mm0           \n\t"
+            // use mm0 mask copy to merge a & b
+            "pand %%mm0, %%mm2           \n\t"
+            // use mm7 mask to merge pa & pb
+            "pand %%mm7, %%mm5           \n\t"
+            "pandn %%mm1, %%mm0          \n\t"
+            "pandn %%mm4, %%mm7          \n\t"
+            "paddw %%mm2, %%mm0          \n\t"
+            "paddw %%mm5, %%mm7          \n\t"
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            "pcmpgtw %%mm6, %%mm7        \n\t" // pab > pc?
+            "pand %%mm7, %%mm3           \n\t"
+            "pandn %%mm0, %%mm7          \n\t"
+            "paddw %%mm3, %%mm7          \n\t"
+            "pxor %%mm1, %%mm1           \n\t"
+            "packuswb %%mm7, %%mm1       \n\t"
+            // step ecx to next set of 8 bytes and repeat loop til done
+            "addl $8, %%ecx              \n\t"
+            "pand " AMASK0_2_6 ", %%mm1  \n\t" // _amask0_2_6 (_ActiveMaskEnd)
+            "paddb -8(%1," PCX ",), %%mm1 \n\t" // add Paeth predictor + Raw(x)
+            "cmpl %%eax, %%ecx           \n\t" // MMXLength
+            "pxor %%mm0, %%mm0           \n\t" // pxor does not affect flags
+            "movq %%mm1, -8(%1," PCX ",) \n\t" // write back updated value
+                                 // mm1 will be used as Raw(x-bpp) next loop
+                           // mm3 ready to be used as Prior(x-bpp) next loop
+            "jb paeth_3lp                \n\t"
+            RESTORE_rbp
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),  // esi/rsi      // input regs
+              "1" (row),       // edi/rdi
+              "2" (diff),      // ecx
+              "3" (MMXLength)  // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm2", "%mm3"   // clobber list
+            , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+         );
+      }
+      break;  // end 3 bpp
+
+      case 4:
+      {
+         __asm__ __volatile__ (
+// preload  "movl diff, %%ecx            \n\t"
+// preload  "movl row, %1                \n\t" // edi/rdi
+// preload  "movl prev_row, %0           \n\t" // esi/rsi
+            "pxor %%mm0, %%mm0           \n\t"
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PCX ",), %%mm1 \n\t" // only time should need to read
+                                               //  a=Raw(x-bpp) bytes
+         "paeth_4lp:                     \n\t"
+            // do first set of 4 bytes
+            "movq -8(%0," PCX ",), %%mm3 \n\t" // read c=Prior(x-bpp) bytes
+            "punpckhbw %%mm0, %%mm1      \n\t" // unpack Low bytes of a
+            "movq (%0," PCX ",), %%mm2   \n\t" // load b=Prior(x)
+            "punpcklbw %%mm0, %%mm2      \n\t" // unpack High bytes of b
+            // pav = p - a = (a + b - c) - a = b - c
+            "movq %%mm2, %%mm4           \n\t"
+            "punpckhbw %%mm0, %%mm3      \n\t" // unpack High bytes of c
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movq %%mm1, %%mm5           \n\t"
+            "psubw %%mm3, %%mm4          \n\t"
+            "pxor %%mm7, %%mm7           \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            "movq %%mm4, %%mm6           \n\t"
+            "psubw %%mm3, %%mm5          \n\t"
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            "pcmpgtw %%mm4, %%mm0        \n\t" // create mask pav bytes < 0
+            "paddw %%mm5, %%mm6          \n\t"
+            "pand %%mm4, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "pcmpgtw %%mm5, %%mm7        \n\t" // create mask pbv bytes < 0
+            "psubw %%mm0, %%mm4          \n\t"
+            "pand %%mm5, %%mm7           \n\t" // only pbv bytes < 0 in mm0
+            "psubw %%mm0, %%mm4          \n\t"
+            "psubw %%mm7, %%mm5          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "pcmpgtw %%mm6, %%mm0        \n\t" // create mask pcv bytes < 0
+            "pand %%mm6, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm7, %%mm5          \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            //  test pa <= pb
+            "movq %%mm4, %%mm7           \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            "pcmpgtw %%mm5, %%mm7        \n\t" // pa > pb?
+            "movq %%mm7, %%mm0           \n\t"
+            // use mm7 mask to merge pa & pb
+            "pand %%mm7, %%mm5           \n\t"
+            // use mm0 mask copy to merge a & b
+            "pand %%mm0, %%mm2           \n\t"
+            "pandn %%mm4, %%mm7          \n\t"
+            "pandn %%mm1, %%mm0          \n\t"
+            "paddw %%mm5, %%mm7          \n\t"
+            "paddw %%mm2, %%mm0          \n\t"
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            "pcmpgtw %%mm6, %%mm7        \n\t" // pab > pc?
+            "pxor %%mm1, %%mm1           \n\t"
+            "pand %%mm7, %%mm3           \n\t"
+            "pandn %%mm0, %%mm7          \n\t"
+            "paddw %%mm3, %%mm7          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "packuswb %%mm1, %%mm7       \n\t"
+            "movq (%0," PCX ",), %%mm3   \n\t" // load c=Prior(x-bpp)
+            LOAD_GOT_rbp
+            "pand " AMASK4_4_0 ", %%mm7  \n\t" // _amask4_4_0 (was _ActiveMask)
+            RESTORE_rbp
+            "movq %%mm3, %%mm2           \n\t" // load b=Prior(x) step 1
+            "paddb (%1," PCX ",), %%mm7  \n\t" // add Paeth predictor + Raw(x)
+            "punpcklbw %%mm0, %%mm3      \n\t" // unpack High bytes of c
+            "movq %%mm7, (%1," PCX ",)   \n\t" // write back updated value
+            "movq %%mm7, %%mm1           \n\t" // now mm1 will be used as
+                                               // Raw(x-bpp)
+            // do second set of 4 bytes
+            "punpckhbw %%mm0, %%mm2      \n\t" // unpack Low bytes of b
+            "punpcklbw %%mm0, %%mm1      \n\t" // unpack Low bytes of a
+            // pav = p - a = (a + b - c) - a = b - c
+            "movq %%mm2, %%mm4           \n\t"
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movq %%mm1, %%mm5           \n\t"
+            "psubw %%mm3, %%mm4          \n\t"
+            "pxor %%mm7, %%mm7           \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            "movq %%mm4, %%mm6           \n\t"
+            "psubw %%mm3, %%mm5          \n\t"
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            "pcmpgtw %%mm4, %%mm0        \n\t" // create mask pav bytes < 0
+            "paddw %%mm5, %%mm6          \n\t"
+            "pand %%mm4, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "pcmpgtw %%mm5, %%mm7        \n\t" // create mask pbv bytes < 0
+            "psubw %%mm0, %%mm4          \n\t"
+            "pand %%mm5, %%mm7           \n\t" // only pbv bytes < 0 in mm0
+            "psubw %%mm0, %%mm4          \n\t"
+            "psubw %%mm7, %%mm5          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "pcmpgtw %%mm6, %%mm0        \n\t" // create mask pcv bytes < 0
+            "pand %%mm6, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm7, %%mm5          \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            //  test pa <= pb
+            "movq %%mm4, %%mm7           \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            "pcmpgtw %%mm5, %%mm7        \n\t" // pa > pb?
+            "movq %%mm7, %%mm0           \n\t"
+            // use mm7 mask to merge pa & pb
+            "pand %%mm7, %%mm5           \n\t"
+            // use mm0 mask copy to merge a & b
+            "pand %%mm0, %%mm2           \n\t"
+            "pandn %%mm4, %%mm7          \n\t"
+            "pandn %%mm1, %%mm0          \n\t"
+            "paddw %%mm5, %%mm7          \n\t"
+            "paddw %%mm2, %%mm0          \n\t"
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            "pcmpgtw %%mm6, %%mm7        \n\t" // pab > pc?
+            "pxor %%mm1, %%mm1           \n\t"
+            "pand %%mm7, %%mm3           \n\t"
+            "pandn %%mm0, %%mm7          \n\t"
+            "pxor %%mm1, %%mm1           \n\t"
+            "paddw %%mm3, %%mm7          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            // step ecx to next set of 8 bytes and repeat loop til done
+            "addl $8, %%ecx              \n\t"
+            "packuswb %%mm7, %%mm1       \n\t"
+            "paddb -8(%1," PCX ",), %%mm1 \n\t" // add predictor with Raw(x)
+            "cmpl %%eax, %%ecx           \n\t" // MMXLength
+            "movq %%mm1, -8(%1," PCX ",) \n\t" // write back updated value
+                                 // mm1 will be used as Raw(x-bpp) next loop
+            "jb paeth_4lp                \n\t"
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),  // esi/rsi      // input regs
+              "1" (row),       // edi/rdi
+              "2" (diff),      // ecx
+              "3" (MMXLength)  // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm2", "%mm3"   // clobber list
+            , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+         );
+      }
+      break;  // end 4 bpp
+
+      case 1:
+      case 2:
+      {
+         __asm__ __volatile__ (
+// preload  "movl diff, %%eax            \n\t" // eax: x = offset to align. bdry
+// preload  "movl FullLength, %%edx      \n\t"
+            "cmpl %%edx, %%eax           \n\t"
+            "jnb paeth_dend              \n\t"
+
+            SAVE_ebp
+
+// preload  "movl row, %2                \n\t" // edi/rdi
+            // do Paeth decode for remaining bytes
+// preload  "movl prev_row, %1           \n\t" // esi/rsi
+            "movl %%eax, %%ebp           \n\t"
+// preload  "subl bpp, %%ebp             \n\t" // (bpp is preloaded into ecx)
+            "subl %%ecx, %%ebp           \n\t" // ebp = eax - bpp
+            "xorl %%ecx, %%ecx           \n\t" // zero ecx before using cl & cx
+
+            SAVE_GOT_ebx
+            SAVE_r11_r12_r13
+
+         "paeth_dlp:                     \n\t"
+            "xorl %%ebx, %%ebx           \n\t"
+            // pav = p - a = (a + b - c) - a = b - c
+            "movb (%1," PAX ",), %%bl    \n\t" // load Prior(x) into bl
+            "movb (%1," PBP ",), %%cl    \n\t" // load Prior(x-bpp) into cl
+            "subl %%ecx, %%ebx           \n\t" // subtract Prior(x-bpp)
+            "movl %%ebx, " pa_TEMP "     \n\t" // Save pav for later use
+            "xorl %%ebx, %%ebx           \n\t"
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movb (%2," PBP ",), %%bl    \n\t" // load Raw(x-bpp) into bl
+            "subl %%ecx, %%ebx           \n\t" // subtract Prior(x-bpp)
+            "movl %%ebx, %%ecx           \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            "addl " pa_TEMP ", %%ebx     \n\t" // pcv = pav + pbv
+            // pc = abs(pcv)
+            "testl $0x80000000, %%ebx    \n\t"
+            "jz paeth_dpca               \n\t"
+            "negl %%ebx                  \n\t" // reverse sign of neg values
+
+         "paeth_dpca:                    \n\t"
+            "movl %%ebx, " pc_TEMP "     \n\t" // save pc for later use
+            // pb = abs(pbv)
+            "testl $0x80000000, %%ecx    \n\t"
+            "jz paeth_dpba               \n\t"
+            "negl %%ecx                  \n\t" // reverse sign of neg values
+
+         "paeth_dpba:                    \n\t"
+            "movl %%ecx, " pb_TEMP "     \n\t" // save pb for later use
+            // pa = abs(pav)
+            "movl " pa_TEMP ", %%ebx     \n\t"
+            "testl $0x80000000, %%ebx    \n\t"
+            "jz paeth_dpaa               \n\t"
+            "negl %%ebx                  \n\t" // reverse sign of neg values
+
+         "paeth_dpaa:                    \n\t"
+            "movl %%ebx, " pa_TEMP "     \n\t" // save pa for later use
+            // test if pa <= pb
+            "cmpl %%ecx, %%ebx           \n\t"
+            "jna paeth_dabb              \n\t"
+            // pa > pb; now test if pb <= pc
+            "cmpl " pc_TEMP ", %%ecx     \n\t"
+            "jna paeth_dbbc              \n\t"
+            // pb > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+            "movb (%1," PBP ",), %%cl    \n\t" // load Prior(x-bpp) into cl
+            "jmp paeth_dpaeth            \n\t"
+
+         "paeth_dbbc:                    \n\t"
+            // pb <= pc; Raw(x) = Paeth(x) + Prior(x)
+            "movb (%1," PAX ",), %%cl    \n\t" // load Prior(x) into cl
+            "jmp paeth_dpaeth            \n\t"
+
+         "paeth_dabb:                    \n\t"
+            // pa <= pb; now test if pa <= pc
+            "cmpl " pc_TEMP ", %%ebx     \n\t"
+            "jna paeth_dabc              \n\t"
+            // pa > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+            "movb (%1," PBP ",), %%cl   \n\t" // load Prior(x-bpp) into cl
+            "jmp paeth_dpaeth            \n\t"
+
+         "paeth_dabc:                    \n\t"
+            // pa <= pc; Raw(x) = Paeth(x) + Raw(x-bpp)
+            "movb (%2," PBP ",), %%cl    \n\t" // load Raw(x-bpp) into cl
+
+         "paeth_dpaeth:                  \n\t"
+            "incl %%eax                  \n\t"
+            "incl %%ebp                  \n\t"
+            // Raw(x) = (Paeth(x) + Paeth_Predictor( a, b, c )) mod 256
+            "addb %%cl, -1(%2," PAX ",)  \n\t"
+            "cmpl %%edx, %%eax           \n\t" // check against FullLength
+            "jb paeth_dlp                \n\t"
+
+            RESTORE_r11_r12_r13
+            RESTORE_GOT_ebx
+            RESTORE_ebp
+
+         "paeth_dend:                    \n\t"
+
+            : "=c" (dummy_value_c),            // output regs (dummy)
+              "=S" (dummy_value_S),
+              "=D" (dummy_value_D),
+              "=a" (dummy_value_a),
+              "=d" (dummy_value_d)
+
+            : "0" (bpp),         // ecx        // input regs
+              "1" (prev_row),    // esi/rsi
+              "2" (row),         // edi/rdi
+              "3" (diff),        // eax
+              "4" (FullLength)   // edx
+
+            CLOB_COLON_ebx_ebp_r1X             // clobber list
+              CLOBBER_GOT_ebx
+              CLOB_COMMA_ebx_ebp
+              CLOBBER_ebp
+              CLOB_COMMA_ebX_r1X
+              CLOBBER_r11_r12_r13
+         );
+      }
+      return; // end 1 or 2 bpp (no need to go further with this one)
+
+      case 6:
+      {
+//       _ActiveMask2 = 0xffffffff00000000LL;  // NOT USED ("_amask_0_4_4")
+//       _ShiftBpp = 48;       // bpp << 3 == bpp * 8
+//       _ShiftRem = 16;       // 64 - _ShiftBpp
+
+         __asm__ __volatile__ (
+// preload  "movl diff, %%ecx            \n\t"
+// preload  "movl row, %1                \n\t" // edi/rdi
+// preload  "movl prev_row, %0           \n\t" // esi/rsi
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PCX ",), %%mm1 \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+
+         "paeth_6lp:                     \n\t"
+            // must shift to position Raw(x-bpp) data
+            "psrlq $16, %%mm1            \n\t" // was _ShiftRem
+            // do first set of 4 bytes
+            "movq -8(%0," PCX ",), %%mm3 \n\t" // read c=Prior(x-bpp) bytes
+            "punpcklbw %%mm0, %%mm1      \n\t" // unpack Low bytes of a
+            "movq (%0," PCX ",), %%mm2   \n\t" // load b=Prior(x)
+            "punpcklbw %%mm0, %%mm2      \n\t" // unpack Low bytes of b
+            // must shift to position Prior(x-bpp) data
+            "psrlq $16, %%mm3            \n\t" // was _ShiftRem
+            // pav = p - a = (a + b - c) - a = b - c
+            "movq %%mm2, %%mm4           \n\t"
+            "punpcklbw %%mm0, %%mm3      \n\t" // unpack Low bytes of c
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movq %%mm1, %%mm5           \n\t"
+            "psubw %%mm3, %%mm4          \n\t"
+            "pxor %%mm7, %%mm7           \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            "movq %%mm4, %%mm6           \n\t"
+            "psubw %%mm3, %%mm5          \n\t"
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            "pcmpgtw %%mm4, %%mm0        \n\t" // create mask pav bytes < 0
+            "paddw %%mm5, %%mm6          \n\t"
+            "pand %%mm4, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "pcmpgtw %%mm5, %%mm7        \n\t" // create mask pbv bytes < 0
+            "psubw %%mm0, %%mm4          \n\t"
+            "pand %%mm5, %%mm7           \n\t" // only pbv bytes < 0 in mm0
+            "psubw %%mm0, %%mm4          \n\t"
+            "psubw %%mm7, %%mm5          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "pcmpgtw %%mm6, %%mm0        \n\t" // create mask pcv bytes < 0
+            "pand %%mm6, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm7, %%mm5          \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            //  test pa <= pb
+            "movq %%mm4, %%mm7           \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            "pcmpgtw %%mm5, %%mm7        \n\t" // pa > pb?
+            "movq %%mm7, %%mm0           \n\t"
+            // use mm7 mask to merge pa & pb
+            "pand %%mm7, %%mm5           \n\t"
+            // use mm0 mask copy to merge a & b
+            "pand %%mm0, %%mm2           \n\t"
+            "pandn %%mm4, %%mm7          \n\t"
+            "pandn %%mm1, %%mm0          \n\t"
+            "paddw %%mm5, %%mm7          \n\t"
+            "paddw %%mm2, %%mm0          \n\t"
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            "pcmpgtw %%mm6, %%mm7        \n\t" // pab > pc?
+            "pxor %%mm1, %%mm1           \n\t"
+            "pand %%mm7, %%mm3           \n\t"
+            "pandn %%mm0, %%mm7          \n\t"
+            "paddw %%mm3, %%mm7          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "packuswb %%mm1, %%mm7       \n\t"
+            "movq -8(%0," PCX ",), %%mm3 \n\t" // load c=Prior(x-bpp)
+            LOAD_GOT_rbp
+            "pand " AMASK4_4_0 ", %%mm7  \n\t" // _amask4_4_0 (was _ActiveMask)
+            RESTORE_rbp
+            "psrlq $16, %%mm3            \n\t"
+            "movq (%0," PCX ",), %%mm2   \n\t" // load b=Prior(x) step 1
+            "paddb (%1," PCX ",), %%mm7  \n\t" // add Paeth predictor + Raw(x)
+            "movq %%mm2, %%mm6           \n\t"
+            "movq %%mm7, (%1," PCX ",)   \n\t" // write back updated value
+            "movq -8(%1," PCX ",), %%mm1 \n\t"
+            "psllq $48, %%mm6            \n\t" // bpp * 8 = bits per pixel
+            "movq %%mm7, %%mm5           \n\t"
+            "psrlq $16, %%mm1            \n\t" // 64 - (bpp * 8) = remainder
+            "por %%mm6, %%mm3            \n\t"
+            "psllq $48, %%mm5            \n\t" // was _ShiftBpp
+            "punpckhbw %%mm0, %%mm3      \n\t" // unpack High bytes of c
+            "por %%mm5, %%mm1            \n\t"
+            // do second set of 4 bytes
+            "punpckhbw %%mm0, %%mm2      \n\t" // unpack High bytes of b
+            "punpckhbw %%mm0, %%mm1      \n\t" // unpack High bytes of a
+            // pav = p - a = (a + b - c) - a = b - c
+            "movq %%mm2, %%mm4           \n\t"
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movq %%mm1, %%mm5           \n\t"
+            "psubw %%mm3, %%mm4          \n\t"
+            "pxor %%mm7, %%mm7           \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            "movq %%mm4, %%mm6           \n\t"
+            "psubw %%mm3, %%mm5          \n\t"
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            "pcmpgtw %%mm4, %%mm0        \n\t" // create mask pav bytes < 0
+            "paddw %%mm5, %%mm6          \n\t"
+            "pand %%mm4, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "pcmpgtw %%mm5, %%mm7        \n\t" // create mask pbv bytes < 0
+            "psubw %%mm0, %%mm4          \n\t"
+            "pand %%mm5, %%mm7           \n\t" // only pbv bytes < 0 in mm0
+            "psubw %%mm0, %%mm4          \n\t"
+            "psubw %%mm7, %%mm5          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "pcmpgtw %%mm6, %%mm0        \n\t" // create mask pcv bytes < 0
+            "pand %%mm6, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm7, %%mm5          \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            //  test pa <= pb
+            "movq %%mm4, %%mm7           \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            "pcmpgtw %%mm5, %%mm7        \n\t" // pa > pb?
+            "movq %%mm7, %%mm0           \n\t"
+            // use mm7 mask to merge pa & pb
+            "pand %%mm7, %%mm5           \n\t"
+            // use mm0 mask copy to merge a & b
+            "pand %%mm0, %%mm2           \n\t"
+            "pandn %%mm4, %%mm7          \n\t"
+            "pandn %%mm1, %%mm0          \n\t"
+            "paddw %%mm5, %%mm7          \n\t"
+            "paddw %%mm2, %%mm0          \n\t"
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            "pcmpgtw %%mm6, %%mm7        \n\t" // pab > pc?
+            "pxor %%mm1, %%mm1           \n\t"
+            "pand %%mm7, %%mm3           \n\t"
+            "pandn %%mm0, %%mm7          \n\t"
+            "pxor %%mm1, %%mm1           \n\t"
+            "paddw %%mm3, %%mm7          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            // step ecx to next set of 8 bytes and repeat loop til done
+            "addl $8, %%ecx              \n\t"
+            "packuswb %%mm7, %%mm1       \n\t"
+            "paddb -8(%1," PCX ",), %%mm1 \n\t" // add Paeth predictor + Raw(x)
+            "cmpl %%eax, %%ecx           \n\t" // MMXLength
+            "movq %%mm1, -8(%1," PCX ",) \n\t" // write back updated value
+                                 // mm1 will be used as Raw(x-bpp) next loop
+            "jb paeth_6lp                \n\t"
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),  // esi/rsi      // input regs
+              "1" (row),       // edi/rdi
+              "2" (diff),      // ecx
+              "3" (MMXLength)  // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm2", "%mm3"   // clobber list
+            , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+         );
+      }
+      break;  // end 6 bpp
+
+      case 8:                          // bpp == 8
+      {
+         __asm__ __volatile__ (
+// preload  "movl diff, %%ecx            \n\t"
+// preload  "movl row, %1                \n\t" // edi/rdi
+// preload  "movl prev_row, %0           \n\t" // esi/rsi
+            "pxor %%mm0, %%mm0           \n\t"
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PCX ",), %%mm1 \n\t" // only time should need to read
+                                               //  a=Raw(x-bpp) bytes
+         "paeth_8lp:                     \n\t"
+            // do first set of 4 bytes
+            "movq -8(%0," PCX ",), %%mm3 \n\t" // read c=Prior(x-bpp) bytes
+            "punpcklbw %%mm0, %%mm1      \n\t" // unpack Low bytes of a
+            "movq (%0," PCX ",), %%mm2   \n\t" // load b=Prior(x)
+            "punpcklbw %%mm0, %%mm2      \n\t" // unpack Low bytes of b
+            // pav = p - a = (a + b - c) - a = b - c
+            "movq %%mm2, %%mm4           \n\t"
+            "punpcklbw %%mm0, %%mm3      \n\t" // unpack Low bytes of c
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movq %%mm1, %%mm5           \n\t"
+            "psubw %%mm3, %%mm4          \n\t"
+            "pxor %%mm7, %%mm7           \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            "movq %%mm4, %%mm6           \n\t"
+            "psubw %%mm3, %%mm5          \n\t"
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            "pcmpgtw %%mm4, %%mm0        \n\t" // create mask pav bytes < 0
+            "paddw %%mm5, %%mm6          \n\t"
+            "pand %%mm4, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "pcmpgtw %%mm5, %%mm7        \n\t" // create mask pbv bytes < 0
+            "psubw %%mm0, %%mm4          \n\t"
+            "pand %%mm5, %%mm7           \n\t" // only pbv bytes < 0 in mm0
+            "psubw %%mm0, %%mm4          \n\t"
+            "psubw %%mm7, %%mm5          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "pcmpgtw %%mm6, %%mm0        \n\t" // create mask pcv bytes < 0
+            "pand %%mm6, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm7, %%mm5          \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            //  test pa <= pb
+            "movq %%mm4, %%mm7           \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            "pcmpgtw %%mm5, %%mm7        \n\t" // pa > pb?
+            "movq %%mm7, %%mm0           \n\t"
+            // use mm7 mask to merge pa & pb
+            "pand %%mm7, %%mm5           \n\t"
+            // use mm0 mask copy to merge a & b
+            "pand %%mm0, %%mm2           \n\t"
+            "pandn %%mm4, %%mm7          \n\t"
+            "pandn %%mm1, %%mm0          \n\t"
+            "paddw %%mm5, %%mm7          \n\t"
+            "paddw %%mm2, %%mm0          \n\t"
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            "pcmpgtw %%mm6, %%mm7        \n\t" // pab > pc?
+            "pxor %%mm1, %%mm1           \n\t"
+            "pand %%mm7, %%mm3           \n\t"
+            "pandn %%mm0, %%mm7          \n\t"
+            "paddw %%mm3, %%mm7          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "packuswb %%mm1, %%mm7       \n\t"
+            "movq -8(%0," PCX ",), %%mm3 \n\t" // read c=Prior(x-bpp) bytes
+            LOAD_GOT_rbp
+            "pand " AMASK4_4_0 ", %%mm7  \n\t" // _amask4_4_0 (was _ActiveMask)
+            RESTORE_rbp
+            "movq (%0," PCX ",), %%mm2   \n\t" // load b=Prior(x)
+            "paddb (%1," PCX ",), %%mm7  \n\t" // add Paeth predictor + Raw(x)
+            "punpckhbw %%mm0, %%mm3      \n\t" // unpack High bytes of c
+            "movq %%mm7, (%1," PCX ",)   \n\t" // write back updated value
+            "movq -8(%1," PCX ",), %%mm1 \n\t" // read a=Raw(x-bpp) bytes
+
+            // do second set of 4 bytes
+            "punpckhbw %%mm0, %%mm2      \n\t" // unpack High bytes of b
+            "punpckhbw %%mm0, %%mm1      \n\t" // unpack High bytes of a
+            // pav = p - a = (a + b - c) - a = b - c
+            "movq %%mm2, %%mm4           \n\t"
+            // pbv = p - b = (a + b - c) - b = a - c
+            "movq %%mm1, %%mm5           \n\t"
+            "psubw %%mm3, %%mm4          \n\t"
+            "pxor %%mm7, %%mm7           \n\t"
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            "movq %%mm4, %%mm6           \n\t"
+            "psubw %%mm3, %%mm5          \n\t"
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            "pcmpgtw %%mm4, %%mm0        \n\t" // create mask pav bytes < 0
+            "paddw %%mm5, %%mm6          \n\t"
+            "pand %%mm4, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "pcmpgtw %%mm5, %%mm7        \n\t" // create mask pbv bytes < 0
+            "psubw %%mm0, %%mm4          \n\t"
+            "pand %%mm5, %%mm7           \n\t" // only pbv bytes < 0 in mm0
+            "psubw %%mm0, %%mm4          \n\t"
+            "psubw %%mm7, %%mm5          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            "pcmpgtw %%mm6, %%mm0        \n\t" // create mask pcv bytes < 0
+            "pand %%mm6, %%mm0           \n\t" // only pav bytes < 0 in mm7
+            "psubw %%mm7, %%mm5          \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            //  test pa <= pb
+            "movq %%mm4, %%mm7           \n\t"
+            "psubw %%mm0, %%mm6          \n\t"
+            "pcmpgtw %%mm5, %%mm7        \n\t" // pa > pb?
+            "movq %%mm7, %%mm0           \n\t"
+            // use mm7 mask to merge pa & pb
+            "pand %%mm7, %%mm5           \n\t"
+            // use mm0 mask copy to merge a & b
+            "pand %%mm0, %%mm2           \n\t"
+            "pandn %%mm4, %%mm7          \n\t"
+            "pandn %%mm1, %%mm0          \n\t"
+            "paddw %%mm5, %%mm7          \n\t"
+            "paddw %%mm2, %%mm0          \n\t"
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            "pcmpgtw %%mm6, %%mm7        \n\t" // pab > pc?
+            "pxor %%mm1, %%mm1           \n\t"
+            "pand %%mm7, %%mm3           \n\t"
+            "pandn %%mm0, %%mm7          \n\t"
+            "pxor %%mm1, %%mm1           \n\t"
+            "paddw %%mm3, %%mm7          \n\t"
+            "pxor %%mm0, %%mm0           \n\t"
+            // step ecx to next set of 8 bytes and repeat loop til done
+            "addl $8, %%ecx              \n\t"
+            "packuswb %%mm7, %%mm1       \n\t"
+            "paddb -8(%1," PCX ",), %%mm1 \n\t" // add Paeth predictor + Raw(x)
+            "cmpl %%eax, %%ecx           \n\t" // MMXLength
+            "movq %%mm1, -8(%1," PCX ",) \n\t" // write back updated value
+                                 // mm1 will be used as Raw(x-bpp) next loop
+            "jb paeth_8lp                \n\t"
+
+            : "=S" (dummy_value_S),            // output regs (dummy)
+              "=D" (dummy_value_D),
+              "=c" (dummy_value_c),
+              "=a" (dummy_value_a)
+
+            : "0" (prev_row),  // esi/rsi      // input regs
+              "1" (row),       // edi/rdi
+              "2" (diff),      // ecx
+              "3" (MMXLength)  // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm2", "%mm3"   // clobber list
+            , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+         );
+      }
+      break;  // end 8 bpp
+
+      default:                // bpp != 1,2,3,4,6,8:  doesn't exist
+      {
+         // ERROR:  SHOULD NEVER BE REACHED
+#if defined(PNG_DEBUG)
+         png_debug(1, "Internal libpng logic error (GCC "
+           "png_read_filter_row_mmx_paeth())\n");
+#endif
+      }
+      break;
+
+   } // end switch (bpp)
+
+   __asm__ __volatile__ (
+      // MMX acceleration complete; now do clean-up
+      // check if any remaining bytes left to decode
+//pre "movl FullLength, %%edx      \n\t"
+//pre "movl MMXLength, %%eax       \n\t"
+      "cmpl %%edx, %%eax           \n\t"
+      "jnb paeth_end               \n\t"
+
+      SAVE_ebp
+
+//pre "movl row, %2                \n\t" // edi/rdi
+//pre "movl prev_row, %1           \n\t" // esi/rsi
+      // do Paeth decode for remaining bytes
+      "movl %%eax, %%ebp           \n\t"
+//pre "subl bpp, %%ebp             \n\t" // (bpp is preloaded into ecx)
+      "subl %%ecx, %%ebp           \n\t" // ebp = eax - bpp
+      "xorl %%ecx, %%ecx           \n\t" // zero ecx before using cl & cx below
+
+      SAVE_GOT_ebx
+      SAVE_r11_r12_r13
+
+   "paeth_lp2:                     \n\t"
+      "xorl %%ebx, %%ebx           \n\t"
+      // pav = p - a = (a + b - c) - a = b - c
+      "movb (%1," PAX ",), %%bl    \n\t" // load Prior(x) into bl
+      "movb (%1," PBP ",), %%cl    \n\t" // load Prior(x-bpp) into cl
+      "subl %%ecx, %%ebx           \n\t" // subtract Prior(x-bpp)
+      "movl %%ebx, " pa_TEMP "     \n\t" // Save pav for later use
+      "xorl %%ebx, %%ebx           \n\t"
+      // pbv = p - b = (a + b - c) - b = a - c
+      "movb (%2," PBP ",), %%bl    \n\t" // load Raw(x-bpp) into bl
+      "subl %%ecx, %%ebx           \n\t" // subtract Prior(x-bpp)
+      "movl %%ebx, %%ecx           \n\t"
+      // pcv = p - c = (a + b - c) - c = (a - c) + (b - c) = pav + pbv
+      "addl " pa_TEMP ", %%ebx     \n\t" // pcv = pav + pbv
+      // pc = abs(pcv)
+      "testl $0x80000000, %%ebx    \n\t"
+      "jz paeth_pca2               \n\t"
+      "negl %%ebx                  \n\t" // reverse sign of neg values
+
+   "paeth_pca2:                    \n\t"
+      "movl %%ebx, " pc_TEMP "     \n\t" // save pc for later use
+      // pb = abs(pbv)
+      "testl $0x80000000, %%ecx    \n\t"
+      "jz paeth_pba2               \n\t"
+      "negl %%ecx                  \n\t" // reverse sign of neg values
+
+   "paeth_pba2:                    \n\t"
+      "movl %%ecx, " pb_TEMP "     \n\t" // save pb for later use
+      // pa = abs(pav)
+      "movl " pa_TEMP ", %%ebx     \n\t"
+      "testl $0x80000000, %%ebx    \n\t"
+      "jz paeth_paa2               \n\t"
+      "negl %%ebx                  \n\t" // reverse sign of neg values
+
+   "paeth_paa2:                    \n\t"
+      "movl %%ebx, " pa_TEMP "     \n\t" // save pa for later use
+      // test if pa <= pb
+      "cmpl %%ecx, %%ebx           \n\t"
+      "jna paeth_abb2              \n\t"
+      // pa > pb; now test if pb <= pc
+      "cmpl " pc_TEMP ", %%ecx     \n\t"
+      "jna paeth_bbc2              \n\t"
+      // pb > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+      "movb (%1," PBP ",), %%cl    \n\t" // load Prior(x-bpp) into cl
+      "jmp paeth_paeth2            \n\t"
+
+   "paeth_bbc2:                    \n\t"
+      // pb <= pc; Raw(x) = Paeth(x) + Prior(x)
+      "movb (%1," PAX ",), %%cl    \n\t" // load Prior(x) into cl
+      "jmp paeth_paeth2            \n\t"
+
+   "paeth_abb2:                    \n\t"
+      // pa <= pb; now test if pa <= pc
+      "cmpl " pc_TEMP ", %%ebx     \n\t"
+      "jna paeth_abc2              \n\t"
+      // pa > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+      "movb (%1," PBP ",), %%cl    \n\t" // load Prior(x-bpp) into cl
+      "jmp paeth_paeth2            \n\t"
+
+   "paeth_abc2:                    \n\t"
+      // pa <= pc; Raw(x) = Paeth(x) + Raw(x-bpp)
+      "movb (%2," PBP ",), %%cl    \n\t" // load Raw(x-bpp) into cl
+
+   "paeth_paeth2:                  \n\t"
+      "incl %%eax                  \n\t"
+      "incl %%ebp                  \n\t"
+      // Raw(x) = (Paeth(x) + Paeth_Predictor( a, b, c )) mod 256
+      "addb %%cl, -1(%2," PAX ",)  \n\t"
+      "cmpl %%edx, %%eax           \n\t" // check against FullLength
+      "jb paeth_lp2                \n\t"
+
+      RESTORE_r11_r12_r13
+      RESTORE_GOT_ebx
+      RESTORE_ebp
+
+   "paeth_end:                     \n\t"
+      "EMMS                        \n\t" // end MMX; prep for poss. FP instrs.
+
+      : "=c" (dummy_value_c),            // output regs (dummy)
+        "=S" (dummy_value_S),
+        "=D" (dummy_value_D),
+        "=a" (dummy_value_a),
+        "=d" (dummy_value_d)
+
+      : "0" (bpp),         // ecx        // input regs
+        "1" (prev_row),    // esi/rsi
+        "2" (row),         // edi/rdi
+        "3" (MMXLength),   // eax
+        "4" (FullLength)   // edx
+
+      CLOB_COLON_ebx_ebp_r1X             // clobber list
+        CLOBBER_GOT_ebx
+        CLOB_COMMA_ebx_ebp
+        CLOBBER_ebp
+        CLOB_COMMA_ebX_r1X
+        CLOBBER_r11_r12_r13
+   );
+
+} /* end png_read_filter_row_mmx_paeth() */
+
+#endif // PNG_x86_64_USE_GOTPCREL || PNG_THREAD_UNSAFE_OK
+#endif /* PNG_MMX_READ_FILTER_PAETH_SUPPORTED */
+
+
+
+
+#if defined(PNG_MMX_READ_FILTER_SUB_SUPPORTED)
+
+//===========================================================================//
+//                                                                           //
+//           P N G _ R E A D _ F I L T E R _ R O W _ M M X _ S U B           //
+//                                                                           //
+//===========================================================================//
+
+// Optimized code for PNG Sub filter decoder
+
+static void /* PRIVATE */
+png_read_filter_row_mmx_sub(png_row_infop row_info, png_bytep row)
+{
+   unsigned FullLength, MMXLength;  // png_uint_32 is actually 64-bit on x86-64
+   int bpp;
+   int dummy_value_a;
+   int dummy_value_c;
+   int dummy_value_d;
+   png_bytep dummy_value_D;
+   int diff; //     __attribute__((used));
+
+   bpp = (row_info->pixel_depth + 7) >> 3;  // calc number of bytes per pixel
+   FullLength = row_info->rowbytes - bpp;   // number of bytes to filter
+     // (why do we subtract off bpp?  not so in avg or paeth...)
+
+   __asm__ __volatile__ (
+      SAVE_r15
+      SAVE_ebp
+//pre "movl row, %1                \n\t" // edi/rdi
+      "mov  %1, " PSI "            \n\t" // lp = row
+//pre "movl bpp, %%ecx             \n\t"
+      "add  " PCX ", %1            \n\t" // rp = row + bpp
+//pre "movl FullLength, %%eax      \n\t" // bring in via eax...
+      SAVE_FullLength                    // ...but store for later use
+
+      "xorl %%eax, %%eax           \n\t"
+
+      // get # of bytes to alignment (note:  computing _delta_ of two pointers,
+      // so hereafter %%ebp is sufficient even on 64-bit)
+      "mov  %1, " PBP "            \n\t" // take start of row
+      "add  $0xf, " PBP "          \n\t" // add 7+8 to incr past alignment bdry
+//    "andl $0xfffffff8, %%ebp     \n\t" // mask to alignment boundary (32-bit!)
+      CLEAR_BOTTOM_3_BITS  PBP    "\n\t" // mask to alignment boundary
+      "sub  %1, " PBP "            \n\t" // subtract row ptr again => ebp =
+      "jz sub_go                   \n\t" //  target value of eax at alignment
+
+   "sub_lp1:                       \n\t" // fix alignment
+      "movb (" PSI "," PAX ",), %%cl \n\t"
+      "addb %%cl, (%1," PAX ",)    \n\t"
+      "incl %%eax                  \n\t"
+      "cmpl %%ebp, %%eax           \n\t"
+      "jb sub_lp1                  \n\t"
+
+   "sub_go:                        \n\t"
+      RESTORE_FullLength "%%ecx    \n\t" // FullLength -> ecx
+      "movl %%ecx, %%edx           \n\t"
+      "subl %%eax, %%edx           \n\t" // subtract alignment fix
+      "andl $0x00000007, %%edx     \n\t" // calc bytes over mult of 8
+      "subl %%edx, %%ecx           \n\t" // drop over bytes from length
+//out "movl %%ecx, MMXLength       \n\t"
+      "movl %%ebp, %%eax           \n\t" // ebp = diff, but no reg constraint(?)
+      RESTORE_ebp                        //  (could swap ebp and ecx functions,
+      RESTORE_r15                        //  but %%cl issues...)
+
+      : "=c" (MMXLength),       // 0     // output regs
+        "=D" (dummy_value_D),   // 1
+        "=a" (diff)             // 2
+
+      : "0" (bpp),              // ecx   // input regs
+        "1" (row),              // edi
+        "2" (FullLength)        // eax
+
+      : "%esi", "%edx"                   // clobber list
+        _CLOBBER_r15
+        _CLOBBER_ebp
+   );
+
+   // now do the math for the rest of the row
+   switch (bpp)
+   {
+      case 3:
+      {
+//       _ShiftBpp = 24;       // == 3 * 8
+//       _ShiftRem  = 40;      // == 64 - 24
+
+         __asm__ __volatile__ (
+// preload  "mov  row, %1                 \n\t" // edi/rdi
+            LOAD_GOT_rbp
+            // load (former) _ActiveMask for 2nd active byte group
+            "movq " AMASK2_3_3 ", %%mm7   \n\t" // _amask2_3_3
+            RESTORE_rbp
+
+// notused  "mov  %1, " PSI "             \n\t" // lp = row
+// preload  "movl bpp, %%ecx              \n\t"
+            "add  " PCX ", %1             \n\t" // rp = row + bpp
+            "movq %%mm7, %%mm6            \n\t"
+// preload  "movl diff, %%edx             \n\t"
+            "psllq $24, %%mm6             \n\t" // move mask in mm6 to cover
+                                                //  3rd active byte group
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PDX ",), %%mm1  \n\t"
+
+         "sub_3lp:                        \n\t" // shift data for adding first
+            "psrlq $40, %%mm1             \n\t" //  bpp bytes (no need for mask;
+                                                //  shift clears inactive bytes)
+            // add 1st active group
+            "movq (%1," PDX ",), %%mm0    \n\t"
+            "paddb %%mm1, %%mm0           \n\t"
+
+            // add 2nd active group
+            "movq %%mm0, %%mm1            \n\t" // mov updated Raws to mm1
+            "psllq $24, %%mm1             \n\t" // shift data to pos. correctly
+            "pand %%mm7, %%mm1            \n\t" // mask to use 2nd active group
+            "paddb %%mm1, %%mm0           \n\t"
+
+            // add 3rd active group
+            "movq %%mm0, %%mm1            \n\t" // mov updated Raws to mm1
+            "psllq $24, %%mm1             \n\t" // shift data to pos. correctly
+            "pand %%mm6, %%mm1            \n\t" // mask to use 3rd active group
+            "addl $8, %%edx               \n\t"
+            "paddb %%mm1, %%mm0           \n\t"
+
+            "cmpl %%eax, %%edx            \n\t" // MMXLength
+            "movq %%mm0, -8(%1," PDX ",)  \n\t" // write updated Raws to array
+            "movq %%mm0, %%mm1            \n\t" // prep 1st add at top of loop
+            "jb sub_3lp                   \n\t"
+
+            : "=c" (dummy_value_c),   // 0      // output regs (dummy)
+              "=D" (dummy_value_D),   // 1
+              "=d" (dummy_value_d),   // 2
+              "=a" (dummy_value_a)    // 3
+
+            : "0" (bpp),              // ecx    // input regs
+              "1" (row),              // edi
+              "2" (diff),             // edx
+              "3" (MMXLength)         // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm6", "%mm7"    // clobber list
+#endif
+         );
+      }
+      break;  // end 3 bpp
+
+      case 4:   // formerly shared with 6 bpp case via _ShiftBpp and _ShiftRem,
+      {         // but 64-bit PIC/.so problems (could still share, moving vars
+                // into unused MMX regs via ecx/edx, but kludgy)
+//       _ShiftBpp = bpp << 3;        // 32 (psllq)
+//       _ShiftRem = 64 - _ShiftBpp;  // 32 (psrlq)
+
+         __asm__ __volatile__ (
+// preload  "mov  row, %1                 \n\t" // edi/rdi
+// preload  "movl diff, %%edx             \n\t"
+// notused  "mov  %1, " PSI "             \n\t" // lp = row
+// preload  "movl bpp, %%ecx              \n\t"
+            "add  " PCX ", %1             \n\t" // rp = row + bpp
+
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PDX ",), %%mm1  \n\t"
+
+         "sub_4lp:                        \n\t" // shift data for adding first
+            "psrlq $32, %%mm1             \n\t" //  bpp bytes (no need for mask;
+                                                //  shift clears inactive bytes)
+            "movq (%1," PDX ",), %%mm0    \n\t"
+            "paddb %%mm1, %%mm0           \n\t"
+
+            // add 2nd active group
+            "movq %%mm0, %%mm1            \n\t" // mov updated Raws to mm1
+            "psllq $32, %%mm1             \n\t" // shift data to pos. correctly
+            "addl $8, %%edx               \n\t"
+            "paddb %%mm1, %%mm0           \n\t"
+
+            "cmpl %%eax, %%edx            \n\t" // MMXLength
+            "movq %%mm0, -8(%1," PDX ",)  \n\t" // write updated Raws to array
+            "movq %%mm0, %%mm1            \n\t" // prep 1st add at top of loop
+            "jb sub_4lp                   \n\t"
+
+            : "=c" (dummy_value_c),   // 0      // output regs (dummy)
+              "=D" (dummy_value_D),   // 1
+              "=d" (dummy_value_d),   // 2
+              "=a" (dummy_value_a)    // 3
+
+            : "0" (bpp),              // ecx    // input regs
+              "1" (row),              // edi
+              "2" (diff),             // edx
+              "3" (MMXLength)         // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1"                    // clobber list
+#endif
+         );
+      }
+      break;  // end 4 bpp
+
+      case 1:
+      {
+         __asm__ __volatile__ (
+// preload  "movl diff, %%edx              \n\t"
+// preload  "mov  row, %1                  \n\t" // edi/rdi
+// preload  "cmpl FullLength, %%edx        \n\t"
+            "cmpl %%eax, %%edx             \n\t"
+            "jnb sub_1end                  \n\t"
+            "mov  %1, " PSI "              \n\t" // lp = row
+// irrel.   "xorl %%ecx, %%ecx             \n\t" // (actually bug with preload)
+// preload  "movl bpp, %%ecx               \n\t"
+            "add  " PCX ", %1              \n\t" // rp = row + bpp
+
+         "sub_1lp:                         \n\t"
+            "movb (" PSI "," PDX ",), %%cl \n\t"
+            "addb %%cl, (%1," PDX ",)      \n\t"
+            "incl %%edx                    \n\t"
+            "cmpl %%eax, %%edx             \n\t" // compare with FullLength
+            "jb sub_1lp                    \n\t"
+
+         "sub_1end:                        \n\t"
+
+            : "=c" (dummy_value_c),   // 0      // output regs (dummy)
+              "=D" (dummy_value_D),   // 1
+              "=d" (dummy_value_d),   // 2
+              "=a" (dummy_value_a)    // 3
+
+            : "0" (bpp),              // ecx    // input regs
+              "1" (row),              // edi
+              "2" (diff),             // edx
+              "3" (FullLength)        // eax
+
+            : "%esi"                            // clobber list
+         );
+      }
+      return;  // end 1 bpp (bypassing cleanup block!)
+
+      case 2:
+      {
+//       _ShiftBpp = 16;       // == 2 * 8
+//       _ShiftRem = 48;       // == 64 - 16
+
+         __asm__ __volatile__ (
+            LOAD_GOT_rbp
+            // load (former) _ActiveMask for 2nd active byte group
+            "movq " AMASK4_2_2 ", %%mm7   \n\t" // _amask4_2_2
+            RESTORE_rbp
+// preload  "movl diff, %%edx             \n\t"
+            "movq %%mm7, %%mm6            \n\t"
+// preload  "mov  row, %1                 \n\t" // edi/rdi
+            "psllq $16, %%mm6             \n\t" // move mask in mm6 to cover
+                                                //  3rd active byte group
+// notused  "mov  %1, " PSI "             \n\t" // lp = row
+            "movq %%mm6, %%mm5            \n\t"
+// preload  "movl bpp, %%ecx              \n\t"
+            "add  " PCX ", %1             \n\t" // rp = row + bpp
+            "psllq $16, %%mm5             \n\t" // move mask in mm5 to cover
+                                                //  4th active byte group
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PDX ",), %%mm1  \n\t"
+
+         "sub_2lp:                        \n\t" // shift data for adding first
+            "psrlq $48, %%mm1             \n\t" //  bpp bytes (no need for mask;
+                                                //  shift clears inactive bytes)
+            // add 1st active group
+            "movq (%1," PDX ",), %%mm0    \n\t"
+            "paddb %%mm1, %%mm0           \n\t"
+
+            // add 2nd active group
+            "movq %%mm0, %%mm1            \n\t" // mov updated Raws to mm1
+            "psllq $16, %%mm1             \n\t" // shift data to pos. correctly
+            "pand %%mm7, %%mm1            \n\t" // mask to use 2nd active group
+            "paddb %%mm1, %%mm0           \n\t"
+
+            // add 3rd active group
+            "movq %%mm0, %%mm1            \n\t" // mov updated Raws to mm1
+            "psllq $16, %%mm1             \n\t" // shift data to pos. correctly
+            "pand %%mm6, %%mm1            \n\t" // mask to use 3rd active group
+            "paddb %%mm1, %%mm0           \n\t"
+
+            // add 4th active group
+            "movq %%mm0, %%mm1            \n\t" // mov updated Raws to mm1
+            "psllq $16, %%mm1             \n\t" // shift data to pos. correctly
+            "pand %%mm5, %%mm1            \n\t" // mask to use 4th active group
+            "addl $8, %%edx               \n\t"
+            "paddb %%mm1, %%mm0           \n\t"
+            "cmpl %%eax, %%edx            \n\t" // MMXLength
+            "movq %%mm0, -8(%1," PDX ",)  \n\t" // write updated Raws to array
+            "movq %%mm0, %%mm1            \n\t" // prep 1st add at top of loop
+            "jb sub_2lp                   \n\t"
+
+            : "=c" (dummy_value_c),   // 0      // output regs (dummy)
+              "=D" (dummy_value_D),   // 1
+              "=d" (dummy_value_d),   // 2
+              "=a" (dummy_value_a)    // 3
+
+            : "0" (bpp),              // ecx    // input regs
+              "1" (row),              // edi
+              "2" (diff),             // edx
+              "3" (MMXLength)         // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1", "%mm5", "%mm6"    // clobber list
+            , "%mm7"
+#endif
+         );
+      }
+      break;  // end 2 bpp
+
+      case 6:   // formerly shared with 4 bpp case (see comments there)
+      {
+//       _ShiftBpp = bpp << 3;        // 48 (psllq)
+//       _ShiftRem = 64 - _ShiftBpp;  // 16 (psrlq)
+
+         __asm__ __volatile__ (
+// preload  "mov  row, %1                 \n\t" // edi/rdi
+// preload  "movl diff, %%edx             \n\t"
+// notused  "mov  %1, " PSI "             \n\t" // lp = row
+// preload  "movl bpp, %%ecx              \n\t"
+            "add  " PCX ", %1             \n\t" // rp = row + bpp
+
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PDX ",), %%mm1  \n\t"
+
+         "sub_6lp:                        \n\t" // shift data for adding first
+            "psrlq $16, %%mm1             \n\t" //  bpp bytes (no need for mask;
+                                                //  shift clears inactive bytes)
+            "movq (%1," PDX ",), %%mm0    \n\t"
+            "paddb %%mm1, %%mm0           \n\t"
+
+            // add 2nd active group
+            "movq %%mm0, %%mm1            \n\t" // mov updated Raws to mm1
+            "psllq $48, %%mm1             \n\t" // shift data to pos. correctly
+            "addl $8, %%edx               \n\t"
+            "paddb %%mm1, %%mm0           \n\t"
+
+            "cmpl %%eax, %%edx            \n\t" // MMXLength
+            "movq %%mm0, -8(%1," PDX ",)  \n\t" // write updated Raws to array
+            "movq %%mm0, %%mm1            \n\t" // prep 1st add at top of loop
+            "jb sub_6lp                   \n\t"
+
+            : "=c" (dummy_value_c),   // 0      // output regs (dummy)
+              "=D" (dummy_value_D),   // 1
+              "=d" (dummy_value_d),   // 2
+              "=a" (dummy_value_a)    // 3
+
+            : "0" (bpp),              // ecx    // input regs
+              "1" (row),              // edi
+              "2" (diff),             // edx
+              "3" (MMXLength)         // eax
+
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            : "%mm0", "%mm1"                    // clobber list
+#endif
+         );
+      }
+      break;  // end 6 bpp
+
+      case 8:
+      {
+         __asm__ __volatile__ (
+// preload  "mov  row, %1                 \n\t" // edi/rdi
+// preload  "movl diff, %%edx             \n\t"
+// notused  "mov  %1, " PSI "             \n\t" // lp = row
+// preload  "movl bpp, %%ecx              \n\t"
+            "add  " PCX ", %1             \n\t" // rp = row + bpp
+// preload  "movl MMXLength, %%eax        \n\t"
+
+            // prime the pump:  load the first Raw(x-bpp) data set
+            "movq -8(%1," PDX ",), %%mm7  \n\t"
+            "movl %%eax, %%esi            \n\t" // copy of MMXLength -> esi
+            "andl $0x0000003f, %%esi      \n\t" // calc bytes over mult of 64
+
+         "sub_8lp:                        \n\t"
+            "movq (%1," PDX ",), %%mm0    \n\t" // load Sub(x) for 1st 8 bytes
+            "paddb %%mm7, %%mm0           \n\t"
+            "movq 8(%1," PDX ",), %%mm1   \n\t" // load Sub(x) for 2nd 8 bytes
+            "movq %%mm0, (%1," PDX ",)    \n\t" // write Raw(x) for 1st 8 bytes
+
+            // Now mm0 will be used as Raw(x-bpp) for the 2nd group of 8 bytes.
+            // This will be repeated for each group of 8 bytes with the 8th
+            // group being used as the Raw(x-bpp) for the 1st group of the
+            // next loop.
+
+            "paddb %%mm0, %%mm1           \n\t"
+            "movq 16(%1," PDX ",), %%mm2  \n\t" // load Sub(x) for 3rd 8 bytes
+            "movq %%mm1, 8(%1," PDX ",)   \n\t" // write Raw(x) for 2nd 8 bytes
+            "paddb %%mm1, %%mm2           \n\t"
+            "movq 24(%1," PDX ",), %%mm3  \n\t" // load Sub(x) for 4th 8 bytes
+            "movq %%mm2, 16(%1," PDX ",)  \n\t" // write Raw(x) for 3rd 8 bytes
+            "paddb %%mm2, %%mm3           \n\t"
+            "movq 32(%1," PDX ",), %%mm4  \n\t" // load Sub(x) for 5th 8 bytes
+            "movq %%mm3, 24(%1," PDX ",)  \n\t" // write Raw(x) for 4th 8 bytes
+            "paddb %%mm3, %%mm4           \n\t"
+            "movq 40(%1," PDX ",), %%mm5  \n\t" // load Sub(x) for 6th 8 bytes
+            "movq %%mm4, 32(%1," PDX ",)  \n\t" // write Raw(x) for 5th 8 bytes
+            "paddb %%mm4, %%mm5           \n\t"
+            "movq 48(%1," PDX ",), %%mm6  \n\t" // load Sub(x) for 7th 8 bytes
+            "movq %%mm5, 40(%1," PDX ",)  \n\t" // write Raw(x) for 6th 8 bytes
+            "paddb %%mm5, %%mm6           \n\t"
+            "movq 56(%1," PDX ",), %%mm7  \n\t" // load Sub(x) for 8th 8 bytes
+            "movq %%mm6, 48(%1," PDX ",)  \n\t" // write Raw(x) for 7th 8 bytes
+            "addl $64, %%edx              \n\t"
+            "paddb %%mm6, %%mm7           \n\t"
+            "cmpl %%esi, %%edx            \n\t" // cmp to bytes over mult of 64
+            "movq %%mm7, -8(%1," PDX ",)  \n\t" // write Raw(x) for 8th 8 bytes
+            "jb sub_8lp                   \n\t"
+
+            "cmpl %%eax, %%edx            \n\t" // compare to MMXLength
+            "jnb sub_8lt8                 \n\t"
+
+         "sub_8lpA:                       \n\t"
+            "movq (%1," PDX ",), %%mm0    \n\t"
+            "addl $8, %%edx               \n\t"
+            "paddb %%mm7, %%mm0           \n\t"
+            "cmpl %%eax, %%edx            \n\t" // compare to MMXLength
+            "movq %%mm0, -8(%1," PDX ",)  \n\t" // -8 to offset early addl edx
+            "movq %%mm0, %%mm7            \n\t" // move calculated Raw(x) data
+            "jb sub_8lpA                  \n\t" //  to mm7 to be new Raw(x-bpp)
+                                                //  for next loop
+         "sub_8lt8:                       \n\t"
+
+            : "=c" (dummy_value_c),   // 0      // output regs (dummy)
+              "=D" (dummy_value_D),   // 1
+              "=d" (dummy_value_d),   // 2
+              "=a" (dummy_value_a)    // 3
+
+            : "0" (bpp),              // ecx    // input regs
+              "1" (row),              // edi
+              "2" (diff),             // edx
+              "3" (MMXLength)         // eax
+
+            : "%esi"                            // clobber list
+#if defined(CLOBBER_MMX_REGS_SUPPORTED)
+            , "%mm0", "%mm1", "%mm2", "%mm3"
+            , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+         );
+      }
+      break;  // end 8 bpp
+
+      default:                // bpp != 1,2,3,4,6,8:  doesn't exist
+      {
+         // ERROR:  SHOULD NEVER BE REACHED
+#if defined(PNG_DEBUG)
+         png_debug(1, "Internal libpng logic error (GCC "
+           "png_read_filter_row_mmx_sub())\n");
+#endif
+      }
+      break;
+
+   } // end switch (bpp)
+
+   __asm__ __volatile__ (
+//pre "movl MMXLength, %%eax         \n\t"
+//pre "mov  row, %1                  \n\t" // edi/rdi
+//pre "cmpl FullLength, %%eax        \n\t"
+      "cmpl %%edx, %%eax             \n\t"
+      "jnb sub_end                   \n\t"
+
+      "mov  %1, " PSI "              \n\t" // lp = row
+//pre "movl bpp, %%ecx               \n\t"
+      "add  " PCX ", %1              \n\t" // rp = row + bpp
+      "xorl %%ecx, %%ecx             \n\t"
+
+   "sub_lp2:                         \n\t"
+      "movb (" PSI "," PAX ",), %%cl \n\t"
+      "addb %%cl, (%1," PAX ",)      \n\t"
+      "incl %%eax                    \n\t"
+      "cmpl %%edx, %%eax             \n\t" // FullLength
+      "jb sub_lp2                    \n\t"
+
+   "sub_end:                         \n\t"
+      "EMMS                          \n\t" // end MMX instructions
+
+      : "=c" (dummy_value_c),   // 0      // output regs (dummy)
+        "=D" (dummy_value_D),   // 1
+        "=a" (dummy_value_a),   // 2
+        "=d" (dummy_value_d)    // 3
+
+      : "0" (bpp),              // ecx    // input regs
+        "1" (row),              // edi
+        "2" (MMXLength),        // eax
+        "3" (FullLength)        // edx
+
+      : "%esi"                            // clobber list
+   );
+
+} // end of png_read_filter_row_mmx_sub()
+
+#endif /* PNG_MMX_READ_FILTER_SUB_SUPPORTED */
+
+
+
+
+#if defined(PNG_MMX_READ_FILTER_UP_SUPPORTED)
+
+//===========================================================================//
+//                                                                           //
+//            P N G _ R E A D _ F I L T E R _ R O W _ M M X _ U P            //
+//                                                                           //
+//===========================================================================//
+
+// Optimized code for PNG Up filter decoder
+
+static void /* PRIVATE */
+png_read_filter_row_mmx_up(png_row_infop row_info, png_bytep row,
+                           png_bytep prev_row)
+{
+   unsigned len;        // png_uint_32 is actually 64-bit on x86-64
+   int dummy_value_d;   // fix 'forbidden register 3 (dx) was spilled' error
+   png_bytep dummy_value_S;
+   png_bytep dummy_value_D;
+
+   len = row_info->rowbytes;              // number of bytes to filter
+
+   __asm__ __volatile__ (
+      SAVE_GOT_ebx
+//pre "mov  prev_row, %1           \n\t" // esi/rsi
+//pre "movl row, %2                \n\t" // edi/rdi
+
+      "xorl %%ebx, %%ebx           \n\t"
+      "xorl %%eax, %%eax           \n\t"
+
+      // get # of bytes to alignment (note:  computing _delta_ of two pointers,
+      // so hereafter %%ecx is sufficient even on 64-bit)
+      "mov  %2, " PCX "            \n\t" // take start of row
+      "add  $0x7, " PCX "          \n\t" // add 7 to incr past alignment bdry
+//    "andl $0xfffffff8, %%ecx     \n\t" // mask to alignment boundary (32-bit!)
+      CLEAR_BOTTOM_3_BITS  PCX    "\n\t" // mask to alignment boundary
+      "sub  %2, " PCX "            \n\t" // subtract row ptr again => ebp =
+      "jz up_go                    \n\t" //  target value of ecx at alignment
+
+   "up_lp1:                        \n\t" // fix alignment
+      "movb (%2," PBX ",), %%al    \n\t"
+      "addb (%1," PBX ",), %%al    \n\t"
+      "incl %%ebx                  \n\t"
+      "cmpl %%ecx, %%ebx           \n\t"
+      "movb %%al, -1(%2," PBX ",)  \n\t" // mov does not affect flags; -1 to
+      "jb up_lp1                   \n\t" //  offset incl ebx
+
+   "up_go:                         \n\t"
+//pre "movl len, %%edx             \n\t"
+      "movl %%edx, %%ecx           \n\t"
+      "subl %%ebx, %%edx           \n\t" // subtract alignment fix
+      "andl $0x0000003f, %%edx     \n\t" // calc bytes over mult of 64
+      "subl %%edx, %%ecx           \n\t" // sub over-bytes from original length
+
+      // unrolled loop - use all MMX registers and interleave to reduce
+      // number of branch instructions (loops) and reduce partial stalls
+   "up_loop:                       \n\t"
+      "movq (%1," PBX ",), %%mm1   \n\t"
+      "movq (%2," PBX ",), %%mm0   \n\t"
+      "movq 8(%1," PBX ",), %%mm3  \n\t"
+      "paddb %%mm1, %%mm0          \n\t"
+      "movq 8(%2," PBX ",), %%mm2  \n\t"
+      "movq %%mm0, (%2," PBX ",)   \n\t"
+      "paddb %%mm3, %%mm2          \n\t"
+      "movq 16(%1," PBX ",), %%mm5 \n\t"
+      "movq %%mm2, 8(%2," PBX ",)  \n\t"
+      "movq 16(%2," PBX ",), %%mm4 \n\t"
+      "movq 24(%1," PBX ",), %%mm7 \n\t"
+      "paddb %%mm5, %%mm4          \n\t"
+      "movq 24(%2," PBX ",), %%mm6 \n\t"
+      "movq %%mm4, 16(%2," PBX ",) \n\t"
+      "paddb %%mm7, %%mm6          \n\t"
+      "movq 32(%1," PBX ",), %%mm1 \n\t"
+      "movq %%mm6, 24(%2," PBX ",) \n\t"
+      "movq 32(%2," PBX ",), %%mm0 \n\t"
+      "movq 40(%1," PBX ",), %%mm3 \n\t"
+      "paddb %%mm1, %%mm0          \n\t"
+      "movq 40(%2," PBX ",), %%mm2 \n\t"
+      "movq %%mm0, 32(%2," PBX ",) \n\t"
+      "paddb %%mm3, %%mm2          \n\t"
+      "movq 48(%1," PBX ",), %%mm5 \n\t"
+      "movq %%mm2, 40(%2," PBX ",) \n\t"
+      "movq 48(%2," PBX ",), %%mm4 \n\t"
+      "movq 56(%1," PBX ",), %%mm7 \n\t"
+      "paddb %%mm5, %%mm4          \n\t"
+      "movq 56(%2," PBX ",), %%mm6 \n\t"
+      "movq %%mm4, 48(%2," PBX ",) \n\t"
+      "addl $64, %%ebx             \n\t"
+      "paddb %%mm7, %%mm6          \n\t"
+      "cmpl %%ecx, %%ebx           \n\t"
+      "movq %%mm6, -8(%2," PBX ",) \n\t" // (+56)movq does not affect flags;
+      "jb up_loop                  \n\t" //  -8 to offset addl ebx
+
+      "cmpl $0, %%edx              \n\t" // test for bytes over mult of 64
+      "jz up_end                   \n\t"
+
+      "cmpl $8, %%edx              \n\t" // test for less than 8 bytes
+      "jb up_lt8                   \n\t" //  [added by lcreeve at netins.net]
+
+      "addl %%edx, %%ecx           \n\t"
+      "andl $0x00000007, %%edx     \n\t" // calc bytes over mult of 8
+      "subl %%edx, %%ecx           \n\t" // drop over-bytes from length
+      "jz up_lt8                   \n\t"
+
+   "up_lpA:                        \n\t" // use MMX regs to update 8 bytes sim.
+      "movq (%1," PBX ",), %%mm1   \n\t"
+      "movq (%2," PBX ",), %%mm0   \n\t"
+      "addl $8, %%ebx              \n\t"
+      "paddb %%mm1, %%mm0          \n\t"
+      "cmpl %%ecx, %%ebx           \n\t"
+      "movq %%mm0, -8(%2," PBX ",) \n\t" // movq does not affect flags; -8 to
+      "jb up_lpA                   \n\t" //  offset add ebx
+      "cmpl $0, %%edx              \n\t" // test for bytes over mult of 8
+      "jz up_end                   \n\t"
+
+   "up_lt8:                        \n\t"
+      "xorl %%eax, %%eax           \n\t"
+      "addl %%edx, %%ecx           \n\t" // move over byte count into counter
+
+   "up_lp2:                        \n\t" // use x86 regs for remaining bytes
+      "movb (%2," PBX ",), %%al    \n\t"
+      "addb (%1," PBX ",), %%al    \n\t"
+      "incl %%ebx                  \n\t"
+      "cmpl %%ecx, %%ebx           \n\t"
+      "movb %%al, -1(%2," PBX ",)  \n\t" // mov does not affect flags; -1 to
+      "jb up_lp2                   \n\t" //  offset inc ebx
+
+   "up_end:                        \n\t"
+      "EMMS                        \n\t" // conversion of filtered row complete
+      RESTORE_GOT_ebx
+
+      : "=d" (dummy_value_d),   // 0     // output regs (dummy)
+        "=S" (dummy_value_S),   // 1
+        "=D" (dummy_value_D)    // 2
+
+      : "0" (len),              // edx   // input regs
+        "1" (prev_row),         // esi
+        "2" (row)               // edi
+
+      : "%eax", "%ecx"                   // clobber list (no input regs!)
+        _CLOBBER_GOT_ebx
+#if defined(PNG_CLOBBER_MMX_REGS_SUPPORTED)
+      , "%mm0", "%mm1", "%mm2", "%mm3"
+      , "%mm4", "%mm5", "%mm6", "%mm7"
+#endif
+   );
+
+} // end of png_read_filter_row_mmx_up()
+
+#endif /* PNG_MMX_READ_FILTER_UP_SUPPORTED */
+
+
+
+
+/*===========================================================================*/
+/*                                                                           */
+/*                   P N G _ R E A D _ F I L T E R _ R O W                   */
+/*                                                                           */
+/*===========================================================================*/
+
+/* Optimized png_read_filter_row routines */
+
+void /* PRIVATE */
+png_read_filter_row(png_structp png_ptr, png_row_infop row_info, png_bytep
+   row, png_bytep prev_row, int filter)
+{
+#if defined(PNG_DEBUG)
+   char filtname[10];
+#endif
+
+   if (_mmx_supported == 2) {
+#if !defined(PNG_1_0_X)
+       /* this should have happened in png_init_mmx_flags() already */
+       png_warning(png_ptr, "asm_flags may not have been initialized");
+#endif
+       png_mmx_support();
+   }
+
+#if defined(PNG_DEBUG)
+   png_debug(1, "in png_read_filter_row (pnggccrd.c)\n");
+   switch (filter)
+   {
+      case 0:
+         png_snprintf(filtname, 10, "none");
+         break;
+
+      case 1:
+         png_snprintf(filtname, 10, "sub-%s",
+#ifdef PNG_MMX_READ_FILTER_SUB_SUPPORTED
+#if !defined(PNG_1_0_X)
+           ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_SUB) &&
+            (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+            (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+           _mmx_supported
+#endif
+           ? "MMX" :
+#endif
+           "C");
+         break;
+
+      case 2:
+         png_snprintf(filtname, 10, "up-%s",
+#ifdef PNG_MMX_READ_FILTER_UP_SUPPORTED
+#if !defined(PNG_1_0_X)
+           ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_UP) &&
+            (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+            (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+           _mmx_supported
+#endif
+           ? "MMX" :
+#endif
+           "C");
+         break;
+
+      case 3:
+         png_snprintf(filtname, 10, "avg-%s",
+#ifdef PNG_MMX_READ_FILTER_AVG_SUPPORTED
+#if !defined(PNG_1_0_X)
+           ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_AVG) &&
+            (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+            (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+           _mmx_supported
+#endif
+           ? "MMX" :
+#endif
+           "C");
+         break;
+
+      case 4:
+         png_snprintf(filtname, 10, "paeth-%s",
+#ifdef PNG_MMX_READ_FILTER_PAETH_SUPPORTED
+#if defined(PNG_x86_64_USE_GOTPCREL) || defined(PNG_THREAD_UNSAFE_OK)
+#if !defined(PNG_1_0_X)
+           ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_PAETH) &&
+            (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+            (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+           _mmx_supported
+#endif
+           ? "MMX" :
+#endif /* PNG_x86_64_USE_GOTPCREL || PNG_THREAD_UNSAFE_OK */
+#endif
+           "C");
+         break;
+
+      default:
+         png_snprintf(filtname, 10, "unknown");
+         break;
+   }
+   png_debug2(2, "row_number=%ld, %s, ", png_ptr->row_number, filtname);
+   //png_debug1(0, "png_ptr=%10p, ", png_ptr);
+   //png_debug1(0, "asm_flags=0x%08lx, ", png_ptr->asm_flags);
+   png_debug1(0, "row=%10p, ", row);
+   png_debug2(0, "pixdepth=%d, bytes=%d, ", (int)row_info->pixel_depth,
+      (int)((row_info->pixel_depth + 7) >> 3));
+   png_debug1(0, "rowbytes=%ld\n", row_info->rowbytes);
+#endif /* PNG_DEBUG */
+
+   switch (filter)
+   {
+      case PNG_FILTER_VALUE_NONE:
+         break;
+
+      case PNG_FILTER_VALUE_SUB:
+#ifdef PNG_MMX_READ_FILTER_SUB_SUPPORTED
+#if !defined(PNG_1_0_X)
+         if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_SUB) &&
+             (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+             (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+         if (_mmx_supported)
+#endif
+         {
+            png_read_filter_row_mmx_sub(row_info, row);
+         }
+         else
+#endif
+         {
+            png_uint_32 i;
+            png_uint_32 istop = row_info->rowbytes;
+            png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+            png_bytep rp = row + bpp;
+            png_bytep lp = row;
+
+            for (i = bpp; i < istop; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) + (int)(*lp++)) & 0xff);
+               rp++;
+            }
+         }  /* end !UseMMX_sub */
+         break;
+
+      case PNG_FILTER_VALUE_UP:
+#ifdef PNG_MMX_READ_FILTER_UP_SUPPORTED
+#if !defined(PNG_1_0_X)
+         if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_UP) &&
+             (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+             (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+         if (_mmx_supported)
+#endif
+         {
+            png_read_filter_row_mmx_up(row_info, row, prev_row);
+         }
+          else
+#endif
+         {
+            png_uint_32 i;
+            png_uint_32 istop = row_info->rowbytes;
+            png_bytep rp = row;
+            png_bytep pp = prev_row;
+
+            for (i = 0; i < istop; ++i)
+            {
+               *rp = (png_byte)(((int)(*rp) + (int)(*pp++)) & 0xff);
+               rp++;
+            }
+         }  /* end !UseMMX_up */
+         break;
+
+      case PNG_FILTER_VALUE_AVG:
+#ifdef PNG_MMX_READ_FILTER_AVG_SUPPORTED
+#if !defined(PNG_1_0_X)
+         if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_AVG) &&
+             (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+             (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+         if (_mmx_supported)
+#endif
+         {
+            png_read_filter_row_mmx_avg(row_info, row, prev_row);
+         }
+         else
+#endif
+         {
+            png_uint_32 i;
+            png_bytep rp = row;
+            png_bytep pp = prev_row;
+            png_bytep lp = row;
+            png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+            png_uint_32 istop = row_info->rowbytes - bpp;
+
+            for (i = 0; i < bpp; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) +
+                  ((int)(*pp++) >> 1)) & 0xff);
+               rp++;
+            }
+
+            for (i = 0; i < istop; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) +
+                  ((int)(*pp++ + *lp++) >> 1)) & 0xff);
+               rp++;
+            }
+         }  /* end !UseMMX_avg */
+         break;
+
+      case PNG_FILTER_VALUE_PAETH:
+#ifdef PNG_MMX_READ_FILTER_PAETH_SUPPORTED
+#if defined(PNG_x86_64_USE_GOTPCREL) || defined(PNG_THREAD_UNSAFE_OK)
+#if !defined(PNG_1_0_X)
+         if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_PAETH) &&
+             (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+             (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+         if (_mmx_supported)
+#endif
+         {
+            png_read_filter_row_mmx_paeth(row_info, row, prev_row);
+         }
+         else
+#endif /* PNG_x86_64_USE_GOTPCREL || PNG_THREAD_UNSAFE_OK */
+#endif
+         {
+            png_uint_32 i;
+            png_bytep rp = row;
+            png_bytep pp = prev_row;
+            png_bytep lp = row;
+            png_bytep cp = prev_row;
+            png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+            png_uint_32 istop = row_info->rowbytes - bpp;
+
+            for (i = 0; i < bpp; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) + (int)(*pp++)) & 0xff);
+               rp++;
+            }
+
+            for (i = 0; i < istop; i++)   /* use leftover rp,pp */
+            {
+               int a, b, c, pa, pb, pc, p;
+
+               a = *lp++;
+               b = *pp++;
+               c = *cp++;
+
+               p = b - c;
+               pc = a - c;
+
+#if defined(PNG_USE_ABS)
+               pa = abs(p);
+               pb = abs(pc);
+               pc = abs(p + pc);
+#else
+               pa = p < 0 ? -p : p;
+               pb = pc < 0 ? -pc : pc;
+               pc = (p + pc) < 0 ? -(p + pc) : p + pc;
+#endif
+
+               /*
+                  if (pa <= pb && pa <= pc)
+                     p = a;
+                  else if (pb <= pc)
+                     p = b;
+                  else
+                     p = c;
+                */
+
+               p = (pa <= pb && pa <= pc) ? a : (pb <= pc) ? b : c;
+
+               *rp = (png_byte)(((int)(*rp) + p) & 0xff);
+               rp++;
+            }
+         }  /* end !UseMMX_paeth */
+         break;
+
+      default:
+         png_warning(png_ptr, "Ignoring bad row-filter type");
+         *row=0;
+         break;
+   }
+}
+
+#endif /* PNG_HAVE_MMX_READ_FILTER_ROW */
+
+
+#endif /* PNG_ASSEMBLER_CODE_SUPPORTED && PNG_USE_PNGGCCRD */
+#endif /* __GNUC__ */
diff --git a/distrib/libpng-1.2.19/pngget.c b/distrib/libpng-1.2.19/pngget.c
new file mode 100644
index 0000000..75d2ca0
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngget.c
@@ -0,0 +1,962 @@
+
+/* pngget.c - retrieval of values from info struct
+ *
+ * Last changed in libpng 1.2.15 January 5, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED)
+
+png_uint_32 PNGAPI
+png_get_valid(png_structp png_ptr, png_infop info_ptr, png_uint_32 flag)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+      return(info_ptr->valid & flag);
+   else
+      return(0);
+}
+
+png_uint_32 PNGAPI
+png_get_rowbytes(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+      return(info_ptr->rowbytes);
+   else
+      return(0);
+}
+
+#if defined(PNG_INFO_IMAGE_SUPPORTED)
+png_bytepp PNGAPI
+png_get_rows(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+      return(info_ptr->row_pointers);
+   else
+      return(0);
+}
+#endif
+
+#ifdef PNG_EASY_ACCESS_SUPPORTED
+/* easy access to info, added in libpng-0.99 */
+png_uint_32 PNGAPI
+png_get_image_width(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+   {
+      return info_ptr->width;
+   }
+   return (0);
+}
+
+png_uint_32 PNGAPI
+png_get_image_height(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+   {
+      return info_ptr->height;
+   }
+   return (0);
+}
+
+png_byte PNGAPI
+png_get_bit_depth(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+   {
+      return info_ptr->bit_depth;
+   }
+   return (0);
+}
+
+png_byte PNGAPI
+png_get_color_type(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+   {
+      return info_ptr->color_type;
+   }
+   return (0);
+}
+
+png_byte PNGAPI
+png_get_filter_type(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+   {
+      return info_ptr->filter_type;
+   }
+   return (0);
+}
+
+png_byte PNGAPI
+png_get_interlace_type(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+   {
+      return info_ptr->interlace_type;
+   }
+   return (0);
+}
+
+png_byte PNGAPI
+png_get_compression_type(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+   {
+      return info_ptr->compression_type;
+   }
+   return (0);
+}
+
+png_uint_32 PNGAPI
+png_get_x_pixels_per_meter(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+#if defined(PNG_pHYs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_pHYs)
+   {
+      png_debug1(1, "in %s retrieval function\n", "png_get_x_pixels_per_meter");
+      if(info_ptr->phys_unit_type != PNG_RESOLUTION_METER)
+          return (0);
+      else return (info_ptr->x_pixels_per_unit);
+   }
+#else
+   return (0);
+#endif
+   return (0);
+}
+
+png_uint_32 PNGAPI
+png_get_y_pixels_per_meter(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+#if defined(PNG_pHYs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_pHYs)
+   {
+      png_debug1(1, "in %s retrieval function\n", "png_get_y_pixels_per_meter");
+      if(info_ptr->phys_unit_type != PNG_RESOLUTION_METER)
+          return (0);
+      else return (info_ptr->y_pixels_per_unit);
+   }
+#else
+   return (0);
+#endif
+   return (0);
+}
+
+png_uint_32 PNGAPI
+png_get_pixels_per_meter(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+#if defined(PNG_pHYs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_pHYs)
+   {
+      png_debug1(1, "in %s retrieval function\n", "png_get_pixels_per_meter");
+      if(info_ptr->phys_unit_type != PNG_RESOLUTION_METER ||
+         info_ptr->x_pixels_per_unit != info_ptr->y_pixels_per_unit)
+          return (0);
+      else return (info_ptr->x_pixels_per_unit);
+   }
+#else
+   return (0);
+#endif
+   return (0);
+}
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+float PNGAPI
+png_get_pixel_aspect_ratio(png_structp png_ptr, png_infop info_ptr)
+   {
+   if (png_ptr != NULL && info_ptr != NULL)
+#if defined(PNG_pHYs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_pHYs)
+   {
+      png_debug1(1, "in %s retrieval function\n", "png_get_aspect_ratio");
+      if (info_ptr->x_pixels_per_unit == 0)
+         return ((float)0.0);
+      else
+         return ((float)((float)info_ptr->y_pixels_per_unit
+            /(float)info_ptr->x_pixels_per_unit));
+   }
+#else
+   return (0.0);
+#endif
+   return ((float)0.0);
+}
+#endif
+
+png_int_32 PNGAPI
+png_get_x_offset_microns(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+#if defined(PNG_oFFs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_oFFs)
+   {
+      png_debug1(1, "in %s retrieval function\n", "png_get_x_offset_microns");
+      if(info_ptr->offset_unit_type != PNG_OFFSET_MICROMETER)
+          return (0);
+      else return (info_ptr->x_offset);
+   }
+#else
+   return (0);
+#endif
+   return (0);
+}
+
+png_int_32 PNGAPI
+png_get_y_offset_microns(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+#if defined(PNG_oFFs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_oFFs)
+   {
+      png_debug1(1, "in %s retrieval function\n", "png_get_y_offset_microns");
+      if(info_ptr->offset_unit_type != PNG_OFFSET_MICROMETER)
+          return (0);
+      else return (info_ptr->y_offset);
+   }
+#else
+   return (0);
+#endif
+   return (0);
+}
+
+png_int_32 PNGAPI
+png_get_x_offset_pixels(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+#if defined(PNG_oFFs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_oFFs)
+   {
+      png_debug1(1, "in %s retrieval function\n", "png_get_x_offset_microns");
+      if(info_ptr->offset_unit_type != PNG_OFFSET_PIXEL)
+          return (0);
+      else return (info_ptr->x_offset);
+   }
+#else
+   return (0);
+#endif
+   return (0);
+}
+
+png_int_32 PNGAPI
+png_get_y_offset_pixels(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+#if defined(PNG_oFFs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_oFFs)
+   {
+      png_debug1(1, "in %s retrieval function\n", "png_get_y_offset_microns");
+      if(info_ptr->offset_unit_type != PNG_OFFSET_PIXEL)
+          return (0);
+      else return (info_ptr->y_offset);
+   }
+#else
+   return (0);
+#endif
+   return (0);
+}
+
+#if defined(PNG_INCH_CONVERSIONS) && defined(PNG_FLOATING_POINT_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_pixels_per_inch(png_structp png_ptr, png_infop info_ptr)
+{
+   return ((png_uint_32)((float)png_get_pixels_per_meter(png_ptr, info_ptr)
+     *.0254 +.5));
+}
+
+png_uint_32 PNGAPI
+png_get_x_pixels_per_inch(png_structp png_ptr, png_infop info_ptr)
+{
+   return ((png_uint_32)((float)png_get_x_pixels_per_meter(png_ptr, info_ptr)
+     *.0254 +.5));
+}
+
+png_uint_32 PNGAPI
+png_get_y_pixels_per_inch(png_structp png_ptr, png_infop info_ptr)
+{
+   return ((png_uint_32)((float)png_get_y_pixels_per_meter(png_ptr, info_ptr)
+     *.0254 +.5));
+}
+
+float PNGAPI
+png_get_x_offset_inches(png_structp png_ptr, png_infop info_ptr)
+{
+   return ((float)png_get_x_offset_microns(png_ptr, info_ptr)
+     *.00003937);
+}
+
+float PNGAPI
+png_get_y_offset_inches(png_structp png_ptr, png_infop info_ptr)
+{
+   return ((float)png_get_y_offset_microns(png_ptr, info_ptr)
+     *.00003937);
+}
+
+#if defined(PNG_pHYs_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_pHYs_dpi(png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 *res_x, png_uint_32 *res_y, int *unit_type)
+{
+   png_uint_32 retval = 0;
+
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_pHYs))
+   {
+      png_debug1(1, "in %s retrieval function\n", "pHYs");
+      if (res_x != NULL)
+      {
+         *res_x = info_ptr->x_pixels_per_unit;
+         retval |= PNG_INFO_pHYs;
+      }
+      if (res_y != NULL)
+      {
+         *res_y = info_ptr->y_pixels_per_unit;
+         retval |= PNG_INFO_pHYs;
+      }
+      if (unit_type != NULL)
+      {
+         *unit_type = (int)info_ptr->phys_unit_type;
+         retval |= PNG_INFO_pHYs;
+         if(*unit_type == 1)
+         {
+            if (res_x != NULL) *res_x = (png_uint_32)(*res_x * .0254 + .50);
+            if (res_y != NULL) *res_y = (png_uint_32)(*res_y * .0254 + .50);
+         }
+      }
+   }
+   return (retval);
+}
+#endif /* PNG_pHYs_SUPPORTED */
+#endif  /* PNG_INCH_CONVERSIONS && PNG_FLOATING_POINT_SUPPORTED */
+
+/* png_get_channels really belongs in here, too, but it's been around longer */
+
+#endif  /* PNG_EASY_ACCESS_SUPPORTED */
+
+png_byte PNGAPI
+png_get_channels(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+      return(info_ptr->channels);
+   else
+      return (0);
+}
+
+png_bytep PNGAPI
+png_get_signature(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr != NULL && info_ptr != NULL)
+      return(info_ptr->signature);
+   else
+      return (NULL);
+}
+
+#if defined(PNG_bKGD_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_bKGD(png_structp png_ptr, png_infop info_ptr,
+   png_color_16p *background)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_bKGD)
+      && background != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "bKGD");
+      *background = &(info_ptr->background);
+      return (PNG_INFO_bKGD);
+   }
+   return (0);
+}
+#endif
+
+#if defined(PNG_cHRM_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+png_uint_32 PNGAPI
+png_get_cHRM(png_structp png_ptr, png_infop info_ptr,
+   double *white_x, double *white_y, double *red_x, double *red_y,
+   double *green_x, double *green_y, double *blue_x, double *blue_y)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_cHRM))
+   {
+      png_debug1(1, "in %s retrieval function\n", "cHRM");
+      if (white_x != NULL)
+         *white_x = (double)info_ptr->x_white;
+      if (white_y != NULL)
+         *white_y = (double)info_ptr->y_white;
+      if (red_x != NULL)
+         *red_x = (double)info_ptr->x_red;
+      if (red_y != NULL)
+         *red_y = (double)info_ptr->y_red;
+      if (green_x != NULL)
+         *green_x = (double)info_ptr->x_green;
+      if (green_y != NULL)
+         *green_y = (double)info_ptr->y_green;
+      if (blue_x != NULL)
+         *blue_x = (double)info_ptr->x_blue;
+      if (blue_y != NULL)
+         *blue_y = (double)info_ptr->y_blue;
+      return (PNG_INFO_cHRM);
+   }
+   return (0);
+}
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+png_uint_32 PNGAPI
+png_get_cHRM_fixed(png_structp png_ptr, png_infop info_ptr,
+   png_fixed_point *white_x, png_fixed_point *white_y, png_fixed_point *red_x,
+   png_fixed_point *red_y, png_fixed_point *green_x, png_fixed_point *green_y,
+   png_fixed_point *blue_x, png_fixed_point *blue_y)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_cHRM))
+   {
+      png_debug1(1, "in %s retrieval function\n", "cHRM");
+      if (white_x != NULL)
+         *white_x = info_ptr->int_x_white;
+      if (white_y != NULL)
+         *white_y = info_ptr->int_y_white;
+      if (red_x != NULL)
+         *red_x = info_ptr->int_x_red;
+      if (red_y != NULL)
+         *red_y = info_ptr->int_y_red;
+      if (green_x != NULL)
+         *green_x = info_ptr->int_x_green;
+      if (green_y != NULL)
+         *green_y = info_ptr->int_y_green;
+      if (blue_x != NULL)
+         *blue_x = info_ptr->int_x_blue;
+      if (blue_y != NULL)
+         *blue_y = info_ptr->int_y_blue;
+      return (PNG_INFO_cHRM);
+   }
+   return (0);
+}
+#endif
+#endif
+
+#if defined(PNG_gAMA_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+png_uint_32 PNGAPI
+png_get_gAMA(png_structp png_ptr, png_infop info_ptr, double *file_gamma)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_gAMA)
+      && file_gamma != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "gAMA");
+      *file_gamma = (double)info_ptr->gamma;
+      return (PNG_INFO_gAMA);
+   }
+   return (0);
+}
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+png_uint_32 PNGAPI
+png_get_gAMA_fixed(png_structp png_ptr, png_infop info_ptr,
+    png_fixed_point *int_file_gamma)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_gAMA)
+      && int_file_gamma != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "gAMA");
+      *int_file_gamma = info_ptr->int_gamma;
+      return (PNG_INFO_gAMA);
+   }
+   return (0);
+}
+#endif
+#endif
+
+#if defined(PNG_sRGB_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_sRGB(png_structp png_ptr, png_infop info_ptr, int *file_srgb_intent)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_sRGB)
+      && file_srgb_intent != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "sRGB");
+      *file_srgb_intent = (int)info_ptr->srgb_intent;
+      return (PNG_INFO_sRGB);
+   }
+   return (0);
+}
+#endif
+
+#if defined(PNG_iCCP_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_iCCP(png_structp png_ptr, png_infop info_ptr,
+             png_charpp name, int *compression_type,
+             png_charpp profile, png_uint_32 *proflen)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_iCCP)
+      && name != NULL && profile != NULL && proflen != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "iCCP");
+      *name = info_ptr->iccp_name;
+      *profile = info_ptr->iccp_profile;
+      /* compression_type is a dummy so the API won't have to change
+         if we introduce multiple compression types later. */
+      *proflen = (int)info_ptr->iccp_proflen;
+      *compression_type = (int)info_ptr->iccp_compression;
+      return (PNG_INFO_iCCP);
+   }
+   return (0);
+}
+#endif
+
+#if defined(PNG_sPLT_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_sPLT(png_structp png_ptr, png_infop info_ptr,
+             png_sPLT_tpp spalettes)
+{
+   if (png_ptr != NULL && info_ptr != NULL && spalettes != NULL)
+   {
+     *spalettes = info_ptr->splt_palettes;
+     return ((png_uint_32)info_ptr->splt_palettes_num);
+   }
+   return (0);
+}
+#endif
+
+#if defined(PNG_hIST_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_hIST(png_structp png_ptr, png_infop info_ptr, png_uint_16p *hist)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_hIST)
+      && hist != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "hIST");
+      *hist = info_ptr->hist;
+      return (PNG_INFO_hIST);
+   }
+   return (0);
+}
+#endif
+
+png_uint_32 PNGAPI
+png_get_IHDR(png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 *width, png_uint_32 *height, int *bit_depth,
+   int *color_type, int *interlace_type, int *compression_type,
+   int *filter_type)
+
+{
+   if (png_ptr != NULL && info_ptr != NULL && width != NULL && height != NULL &&
+      bit_depth != NULL && color_type != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "IHDR");
+      *width = info_ptr->width;
+      *height = info_ptr->height;
+      *bit_depth = info_ptr->bit_depth;
+      if (info_ptr->bit_depth < 1 || info_ptr->bit_depth > 16)
+        png_error(png_ptr, "Invalid bit depth");
+      *color_type = info_ptr->color_type;
+      if (info_ptr->color_type > 6)
+        png_error(png_ptr, "Invalid color type");
+      if (compression_type != NULL)
+         *compression_type = info_ptr->compression_type;
+      if (filter_type != NULL)
+         *filter_type = info_ptr->filter_type;
+      if (interlace_type != NULL)
+         *interlace_type = info_ptr->interlace_type;
+
+      /* check for potential overflow of rowbytes */
+      if (*width == 0 || *width > PNG_UINT_31_MAX)
+        png_error(png_ptr, "Invalid image width");
+      if (*height == 0 || *height > PNG_UINT_31_MAX)
+        png_error(png_ptr, "Invalid image height");
+      if (info_ptr->width > (PNG_UINT_32_MAX
+                 >> 3)      /* 8-byte RGBA pixels */
+                 - 64       /* bigrowbuf hack */
+                 - 1        /* filter byte */
+                 - 7*8      /* rounding of width to multiple of 8 pixels */
+                 - 8)       /* extra max_pixel_depth pad */
+      {
+         png_warning(png_ptr,
+            "Width too large for libpng to process image data.");
+      }
+      return (1);
+   }
+   return (0);
+}
+
+#if defined(PNG_oFFs_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_oFFs(png_structp png_ptr, png_infop info_ptr,
+   png_int_32 *offset_x, png_int_32 *offset_y, int *unit_type)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_oFFs)
+      && offset_x != NULL && offset_y != NULL && unit_type != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "oFFs");
+      *offset_x = info_ptr->x_offset;
+      *offset_y = info_ptr->y_offset;
+      *unit_type = (int)info_ptr->offset_unit_type;
+      return (PNG_INFO_oFFs);
+   }
+   return (0);
+}
+#endif
+
+#if defined(PNG_pCAL_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_pCAL(png_structp png_ptr, png_infop info_ptr,
+   png_charp *purpose, png_int_32 *X0, png_int_32 *X1, int *type, int *nparams,
+   png_charp *units, png_charpp *params)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_pCAL)
+      && purpose != NULL && X0 != NULL && X1 != NULL && type != NULL &&
+      nparams != NULL && units != NULL && params != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "pCAL");
+      *purpose = info_ptr->pcal_purpose;
+      *X0 = info_ptr->pcal_X0;
+      *X1 = info_ptr->pcal_X1;
+      *type = (int)info_ptr->pcal_type;
+      *nparams = (int)info_ptr->pcal_nparams;
+      *units = info_ptr->pcal_units;
+      *params = info_ptr->pcal_params;
+      return (PNG_INFO_pCAL);
+   }
+   return (0);
+}
+#endif
+
+#if defined(PNG_sCAL_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+png_uint_32 PNGAPI
+png_get_sCAL(png_structp png_ptr, png_infop info_ptr,
+             int *unit, double *width, double *height)
+{
+    if (png_ptr != NULL && info_ptr != NULL &&
+       (info_ptr->valid & PNG_INFO_sCAL))
+    {
+        *unit = info_ptr->scal_unit;
+        *width = info_ptr->scal_pixel_width;
+        *height = info_ptr->scal_pixel_height;
+        return (PNG_INFO_sCAL);
+    }
+    return(0);
+}
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+png_uint_32 PNGAPI
+png_get_sCAL_s(png_structp png_ptr, png_infop info_ptr,
+             int *unit, png_charpp width, png_charpp height)
+{
+    if (png_ptr != NULL && info_ptr != NULL &&
+       (info_ptr->valid & PNG_INFO_sCAL))
+    {
+        *unit = info_ptr->scal_unit;
+        *width = info_ptr->scal_s_width;
+        *height = info_ptr->scal_s_height;
+        return (PNG_INFO_sCAL);
+    }
+    return(0);
+}
+#endif
+#endif
+#endif
+
+#if defined(PNG_pHYs_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_pHYs(png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 *res_x, png_uint_32 *res_y, int *unit_type)
+{
+   png_uint_32 retval = 0;
+
+   if (png_ptr != NULL && info_ptr != NULL &&
+      (info_ptr->valid & PNG_INFO_pHYs))
+   {
+      png_debug1(1, "in %s retrieval function\n", "pHYs");
+      if (res_x != NULL)
+      {
+         *res_x = info_ptr->x_pixels_per_unit;
+         retval |= PNG_INFO_pHYs;
+      }
+      if (res_y != NULL)
+      {
+         *res_y = info_ptr->y_pixels_per_unit;
+         retval |= PNG_INFO_pHYs;
+      }
+      if (unit_type != NULL)
+      {
+         *unit_type = (int)info_ptr->phys_unit_type;
+         retval |= PNG_INFO_pHYs;
+      }
+   }
+   return (retval);
+}
+#endif
+
+png_uint_32 PNGAPI
+png_get_PLTE(png_structp png_ptr, png_infop info_ptr, png_colorp *palette,
+   int *num_palette)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_PLTE)
+       && palette != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "PLTE");
+      *palette = info_ptr->palette;
+      *num_palette = info_ptr->num_palette;
+      png_debug1(3, "num_palette = %d\n", *num_palette);
+      return (PNG_INFO_PLTE);
+   }
+   return (0);
+}
+
+#if defined(PNG_sBIT_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_sBIT(png_structp png_ptr, png_infop info_ptr, png_color_8p *sig_bit)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_sBIT)
+      && sig_bit != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "sBIT");
+      *sig_bit = &(info_ptr->sig_bit);
+      return (PNG_INFO_sBIT);
+   }
+   return (0);
+}
+#endif
+
+#if defined(PNG_TEXT_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_text(png_structp png_ptr, png_infop info_ptr, png_textp *text_ptr,
+   int *num_text)
+{
+   if (png_ptr != NULL && info_ptr != NULL && info_ptr->num_text > 0)
+   {
+      png_debug1(1, "in %s retrieval function\n",
+         (png_ptr->chunk_name[0] == '\0' ? "text"
+             : (png_const_charp)png_ptr->chunk_name));
+      if (text_ptr != NULL)
+         *text_ptr = info_ptr->text;
+      if (num_text != NULL)
+         *num_text = info_ptr->num_text;
+      return ((png_uint_32)info_ptr->num_text);
+   }
+   if (num_text != NULL)
+     *num_text = 0;
+   return(0);
+}
+#endif
+
+#if defined(PNG_tIME_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_tIME(png_structp png_ptr, png_infop info_ptr, png_timep *mod_time)
+{
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_tIME)
+       && mod_time != NULL)
+   {
+      png_debug1(1, "in %s retrieval function\n", "tIME");
+      *mod_time = &(info_ptr->mod_time);
+      return (PNG_INFO_tIME);
+   }
+   return (0);
+}
+#endif
+
+#if defined(PNG_tRNS_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_tRNS(png_structp png_ptr, png_infop info_ptr,
+   png_bytep *trans, int *num_trans, png_color_16p *trans_values)
+{
+   png_uint_32 retval = 0;
+   if (png_ptr != NULL && info_ptr != NULL && (info_ptr->valid & PNG_INFO_tRNS))
+   {
+      png_debug1(1, "in %s retrieval function\n", "tRNS");
+      if (info_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      {
+          if (trans != NULL)
+          {
+             *trans = info_ptr->trans;
+             retval |= PNG_INFO_tRNS;
+          }
+          if (trans_values != NULL)
+             *trans_values = &(info_ptr->trans_values);
+      }
+      else /* if (info_ptr->color_type != PNG_COLOR_TYPE_PALETTE) */
+      {
+          if (trans_values != NULL)
+          {
+             *trans_values = &(info_ptr->trans_values);
+             retval |= PNG_INFO_tRNS;
+          }
+          if(trans != NULL)
+             *trans = NULL;
+      }
+      if(num_trans != NULL)
+      {
+         *num_trans = info_ptr->num_trans;
+         retval |= PNG_INFO_tRNS;
+      }
+   }
+   return (retval);
+}
+#endif
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+png_uint_32 PNGAPI
+png_get_unknown_chunks(png_structp png_ptr, png_infop info_ptr,
+             png_unknown_chunkpp unknowns)
+{
+   if (png_ptr != NULL && info_ptr != NULL && unknowns != NULL)
+   {
+     *unknowns = info_ptr->unknown_chunks;
+     return ((png_uint_32)info_ptr->unknown_chunks_num);
+   }
+   return (0);
+}
+#endif
+
+#if defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+png_byte PNGAPI
+png_get_rgb_to_gray_status (png_structp png_ptr)
+{
+   return (png_byte)(png_ptr? png_ptr->rgb_to_gray_status : 0);
+}
+#endif
+
+#if defined(PNG_USER_CHUNKS_SUPPORTED)
+png_voidp PNGAPI
+png_get_user_chunk_ptr(png_structp png_ptr)
+{
+   return (png_ptr? png_ptr->user_chunk_ptr : NULL);
+}
+#endif
+
+#ifdef PNG_WRITE_SUPPORTED
+png_uint_32 PNGAPI
+png_get_compression_buffer_size(png_structp png_ptr)
+{
+   return (png_uint_32)(png_ptr? png_ptr->zbuf_size : 0L);
+}
+#endif
+
+#ifdef PNG_ASSEMBLER_CODE_SUPPORTED
+#ifndef PNG_1_0_X
+/* this function was added to libpng 1.2.0 and should exist by default */
+png_uint_32 PNGAPI
+png_get_asm_flags (png_structp png_ptr)
+{
+#ifdef PNG_MMX_CODE_SUPPORTED
+    return (png_uint_32)(png_ptr? png_ptr->asm_flags : 0L);
+#else
+    return (png_ptr? 0L: 0L);
+#endif
+}
+
+/* this function was added to libpng 1.2.0 and should exist by default */
+png_uint_32 PNGAPI
+png_get_asm_flagmask (int flag_select)
+{
+#ifdef PNG_MMX_CODE_SUPPORTED
+    png_uint_32 settable_asm_flags = 0;
+
+    if (flag_select & PNG_SELECT_READ)
+        settable_asm_flags |=
+          PNG_ASM_FLAG_MMX_READ_COMBINE_ROW  |
+          PNG_ASM_FLAG_MMX_READ_INTERLACE    |
+          PNG_ASM_FLAG_MMX_READ_FILTER_SUB   |
+          PNG_ASM_FLAG_MMX_READ_FILTER_UP    |
+          PNG_ASM_FLAG_MMX_READ_FILTER_AVG   |
+          PNG_ASM_FLAG_MMX_READ_FILTER_PAETH ;
+          /* no non-MMX flags yet */
+
+#if 0
+    /* GRR:  no write-flags yet, either, but someday... */
+    if (flag_select & PNG_SELECT_WRITE)
+        settable_asm_flags |=
+          PNG_ASM_FLAG_MMX_WRITE_ [whatever] ;
+#endif /* 0 */
+
+    return settable_asm_flags;  /* _theoretically_ settable capabilities only */
+#else
+    return (0L);
+#endif /* PNG_MMX_CODE_SUPPORTED */
+}
+
+
+    /* GRR:  could add this:   && defined(PNG_MMX_CODE_SUPPORTED) */
+/* this function was added to libpng 1.2.0 */
+png_uint_32 PNGAPI
+png_get_mmx_flagmask (int flag_select, int *compilerID)
+{
+#if defined(PNG_MMX_CODE_SUPPORTED)
+    png_uint_32 settable_mmx_flags = 0;
+
+    if (flag_select & PNG_SELECT_READ)
+        settable_mmx_flags |=
+          PNG_ASM_FLAG_MMX_READ_COMBINE_ROW  |
+          PNG_ASM_FLAG_MMX_READ_INTERLACE    |
+          PNG_ASM_FLAG_MMX_READ_FILTER_SUB   |
+          PNG_ASM_FLAG_MMX_READ_FILTER_UP    |
+          PNG_ASM_FLAG_MMX_READ_FILTER_AVG   |
+          PNG_ASM_FLAG_MMX_READ_FILTER_PAETH ;
+#if 0
+    /* GRR:  no MMX write support yet, but someday... */
+    if (flag_select & PNG_SELECT_WRITE)
+        settable_mmx_flags |=
+          PNG_ASM_FLAG_MMX_WRITE_ [whatever] ;
+#endif /* 0 */
+
+    if (compilerID != NULL) {
+#ifdef PNG_USE_PNGVCRD
+        *compilerID = 1;    /* MSVC */
+#else
+#ifdef PNG_USE_PNGGCCRD
+        *compilerID = 2;    /* gcc/gas */
+#else
+        *compilerID = -1;   /* unknown (i.e., no asm/MMX code compiled) */
+#endif
+#endif
+    }
+
+    return settable_mmx_flags;  /* _theoretically_ settable capabilities only */
+#else
+    return (0L);
+#endif /* ?PNG_MMX_CODE_SUPPORTED */
+}
+
+/* this function was added to libpng 1.2.0 */
+png_byte PNGAPI
+png_get_mmx_bitdepth_threshold (png_structp png_ptr)
+{
+#if defined(PNG_MMX_CODE_SUPPORTED)
+    return (png_byte)(png_ptr? png_ptr->mmx_bitdepth_threshold : 0);
+#else
+    return (png_ptr? 0: 0);
+#endif /* ?PNG_MMX_CODE_SUPPORTED */
+}
+
+/* this function was added to libpng 1.2.0 */
+png_uint_32 PNGAPI
+png_get_mmx_rowbytes_threshold (png_structp png_ptr)
+{
+#if defined(PNG_MMX_CODE_SUPPORTED)
+    return (png_uint_32)(png_ptr? png_ptr->mmx_rowbytes_threshold : 0L);
+#else
+    return (png_ptr? 0L: 0L);
+#endif /* ?PNG_MMX_CODE_SUPPORTED */
+}
+#endif /* ?PNG_1_0_X */
+#endif /* ?PNG_ASSEMBLER_CODE_SUPPORTED */
+
+#ifdef PNG_SET_USER_LIMITS_SUPPORTED
+/* these functions were added to libpng 1.2.6 */
+png_uint_32 PNGAPI
+png_get_user_width_max (png_structp png_ptr)
+{
+    return (png_ptr? png_ptr->user_width_max : 0);
+}
+png_uint_32 PNGAPI
+png_get_user_height_max (png_structp png_ptr)
+{
+    return (png_ptr? png_ptr->user_height_max : 0);
+}
+#endif /* ?PNG_SET_USER_LIMITS_SUPPORTED */
+ 
+
+#endif /* PNG_READ_SUPPORTED || PNG_WRITE_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngmem.c b/distrib/libpng-1.2.19/pngmem.c
new file mode 100644
index 0000000..248060f
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngmem.c
@@ -0,0 +1,608 @@
+
+/* pngmem.c - stub functions for memory allocation
+ *
+ * Last changed in libpng 1.2.13 November 13, 2006
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2006 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ *
+ * This file provides a location for all memory allocation.  Users who
+ * need special memory handling are expected to supply replacement
+ * functions for png_malloc() and png_free(), and to use
+ * png_create_read_struct_2() and png_create_write_struct_2() to
+ * identify the replacement functions.
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED)
+
+/* Borland DOS special memory handler */
+#if defined(__TURBOC__) && !defined(_Windows) && !defined(__FLAT__)
+/* if you change this, be sure to change the one in png.h also */
+
+/* Allocate memory for a png_struct.  The malloc and memset can be replaced
+   by a single call to calloc() if this is thought to improve performance. */
+png_voidp /* PRIVATE */
+png_create_struct(int type)
+{
+#ifdef PNG_USER_MEM_SUPPORTED
+   return (png_create_struct_2(type, png_malloc_ptr_NULL, png_voidp_NULL));
+}
+
+/* Alternate version of png_create_struct, for use with user-defined malloc. */
+png_voidp /* PRIVATE */
+png_create_struct_2(int type, png_malloc_ptr malloc_fn, png_voidp mem_ptr)
+{
+#endif /* PNG_USER_MEM_SUPPORTED */
+   png_size_t size;
+   png_voidp struct_ptr;
+
+   if (type == PNG_STRUCT_INFO)
+     size = png_sizeof(png_info);
+   else if (type == PNG_STRUCT_PNG)
+     size = png_sizeof(png_struct);
+   else
+     return (png_get_copyright(NULL));
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   if(malloc_fn != NULL)
+   {
+      png_struct dummy_struct;
+      png_structp png_ptr = &dummy_struct;
+      png_ptr->mem_ptr=mem_ptr;
+      struct_ptr = (*(malloc_fn))(png_ptr, (png_uint_32)size);
+   }
+   else
+#endif /* PNG_USER_MEM_SUPPORTED */
+      struct_ptr = (png_voidp)farmalloc(size);
+   if (struct_ptr != NULL)
+      png_memset(struct_ptr, 0, size);
+   return (struct_ptr);
+}
+
+/* Free memory allocated by a png_create_struct() call */
+void /* PRIVATE */
+png_destroy_struct(png_voidp struct_ptr)
+{
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_destroy_struct_2(struct_ptr, png_free_ptr_NULL, png_voidp_NULL);
+}
+
+/* Free memory allocated by a png_create_struct() call */
+void /* PRIVATE */
+png_destroy_struct_2(png_voidp struct_ptr, png_free_ptr free_fn,
+    png_voidp mem_ptr)
+{
+#endif
+   if (struct_ptr != NULL)
+   {
+#ifdef PNG_USER_MEM_SUPPORTED
+      if(free_fn != NULL)
+      {
+         png_struct dummy_struct;
+         png_structp png_ptr = &dummy_struct;
+         png_ptr->mem_ptr=mem_ptr;
+         (*(free_fn))(png_ptr, struct_ptr);
+         return;
+      }
+#endif /* PNG_USER_MEM_SUPPORTED */
+      farfree (struct_ptr);
+   }
+}
+
+/* Allocate memory.  For reasonable files, size should never exceed
+ * 64K.  However, zlib may allocate more then 64K if you don't tell
+ * it not to.  See zconf.h and png.h for more information. zlib does
+ * need to allocate exactly 64K, so whatever you call here must
+ * have the ability to do that.
+ *
+ * Borland seems to have a problem in DOS mode for exactly 64K.
+ * It gives you a segment with an offset of 8 (perhaps to store its
+ * memory stuff).  zlib doesn't like this at all, so we have to
+ * detect and deal with it.  This code should not be needed in
+ * Windows or OS/2 modes, and only in 16 bit mode.  This code has
+ * been updated by Alexander Lehmann for version 0.89 to waste less
+ * memory.
+ *
+ * Note that we can't use png_size_t for the "size" declaration,
+ * since on some systems a png_size_t is a 16-bit quantity, and as a
+ * result, we would be truncating potentially larger memory requests
+ * (which should cause a fatal error) and introducing major problems.
+ */
+
+png_voidp PNGAPI
+png_malloc(png_structp png_ptr, png_uint_32 size)
+{
+   png_voidp ret;
+
+   if (png_ptr == NULL || size == 0)
+      return (NULL);
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   if(png_ptr->malloc_fn != NULL)
+       ret = ((png_voidp)(*(png_ptr->malloc_fn))(png_ptr, (png_size_t)size));
+   else
+       ret = (png_malloc_default(png_ptr, size));
+   if (ret == NULL && (png_ptr->flags&PNG_FLAG_MALLOC_NULL_MEM_OK) == 0)
+       png_error(png_ptr, "Out of memory!");
+   return (ret);
+}
+
+png_voidp PNGAPI
+png_malloc_default(png_structp png_ptr, png_uint_32 size)
+{
+   png_voidp ret;
+#endif /* PNG_USER_MEM_SUPPORTED */
+
+   if (png_ptr == NULL || size == 0)
+      return (NULL);
+
+#ifdef PNG_MAX_MALLOC_64K
+   if (size > (png_uint_32)65536L)
+   {
+      png_warning(png_ptr, "Cannot Allocate > 64K");
+      ret = NULL;
+   }
+   else
+#endif
+
+   if (size != (size_t)size)
+     ret = NULL;
+   else if (size == (png_uint_32)65536L)
+   {
+      if (png_ptr->offset_table == NULL)
+      {
+         /* try to see if we need to do any of this fancy stuff */
+         ret = farmalloc(size);
+         if (ret == NULL || ((png_size_t)ret & 0xffff))
+         {
+            int num_blocks;
+            png_uint_32 total_size;
+            png_bytep table;
+            int i;
+            png_byte huge * hptr;
+
+            if (ret != NULL)
+            {
+               farfree(ret);
+               ret = NULL;
+            }
+
+            if(png_ptr->zlib_window_bits > 14)
+               num_blocks = (int)(1 << (png_ptr->zlib_window_bits - 14));
+            else
+               num_blocks = 1;
+            if (png_ptr->zlib_mem_level >= 7)
+               num_blocks += (int)(1 << (png_ptr->zlib_mem_level - 7));
+            else
+               num_blocks++;
+
+            total_size = ((png_uint_32)65536L) * (png_uint_32)num_blocks+16;
+
+            table = farmalloc(total_size);
+
+            if (table == NULL)
+            {
+#ifndef PNG_USER_MEM_SUPPORTED
+               if ((png_ptr->flags&PNG_FLAG_MALLOC_NULL_MEM_OK) == 0)
+                  png_error(png_ptr, "Out Of Memory."); /* Note "O" and "M" */
+               else
+                  png_warning(png_ptr, "Out Of Memory.");
+#endif
+               return (NULL);
+            }
+
+            if ((png_size_t)table & 0xfff0)
+            {
+#ifndef PNG_USER_MEM_SUPPORTED
+               if ((png_ptr->flags&PNG_FLAG_MALLOC_NULL_MEM_OK) == 0)
+                  png_error(png_ptr,
+                    "Farmalloc didn't return normalized pointer");
+               else
+                  png_warning(png_ptr,
+                    "Farmalloc didn't return normalized pointer");
+#endif
+               return (NULL);
+            }
+
+            png_ptr->offset_table = table;
+            png_ptr->offset_table_ptr = farmalloc(num_blocks *
+               png_sizeof (png_bytep));
+
+            if (png_ptr->offset_table_ptr == NULL)
+            {
+#ifndef PNG_USER_MEM_SUPPORTED
+               if ((png_ptr->flags&PNG_FLAG_MALLOC_NULL_MEM_OK) == 0)
+                  png_error(png_ptr, "Out Of memory."); /* Note "O" and "M" */
+               else
+                  png_warning(png_ptr, "Out Of memory.");
+#endif
+               return (NULL);
+            }
+
+            hptr = (png_byte huge *)table;
+            if ((png_size_t)hptr & 0xf)
+            {
+               hptr = (png_byte huge *)((long)(hptr) & 0xfffffff0L);
+               hptr = hptr + 16L;  /* "hptr += 16L" fails on Turbo C++ 3.0 */
+            }
+            for (i = 0; i < num_blocks; i++)
+            {
+               png_ptr->offset_table_ptr[i] = (png_bytep)hptr;
+               hptr = hptr + (png_uint_32)65536L;  /* "+=" fails on TC++3.0 */
+            }
+
+            png_ptr->offset_table_number = num_blocks;
+            png_ptr->offset_table_count = 0;
+            png_ptr->offset_table_count_free = 0;
+         }
+      }
+
+      if (png_ptr->offset_table_count >= png_ptr->offset_table_number)
+      {
+#ifndef PNG_USER_MEM_SUPPORTED
+         if ((png_ptr->flags&PNG_FLAG_MALLOC_NULL_MEM_OK) == 0)
+            png_error(png_ptr, "Out of Memory."); /* Note "o" and "M" */
+         else
+            png_warning(png_ptr, "Out of Memory.");
+#endif
+         return (NULL);
+      }
+
+      ret = png_ptr->offset_table_ptr[png_ptr->offset_table_count++];
+   }
+   else
+      ret = farmalloc(size);
+
+#ifndef PNG_USER_MEM_SUPPORTED
+   if (ret == NULL)
+   {
+      if ((png_ptr->flags&PNG_FLAG_MALLOC_NULL_MEM_OK) == 0)
+         png_error(png_ptr, "Out of memory."); /* Note "o" and "m" */
+      else
+         png_warning(png_ptr, "Out of memory."); /* Note "o" and "m" */
+   }
+#endif
+
+   return (ret);
+}
+
+/* free a pointer allocated by png_malloc().  In the default
+   configuration, png_ptr is not used, but is passed in case it
+   is needed.  If ptr is NULL, return without taking any action. */
+void PNGAPI
+png_free(png_structp png_ptr, png_voidp ptr)
+{
+   if (png_ptr == NULL || ptr == NULL)
+      return;
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   if (png_ptr->free_fn != NULL)
+   {
+      (*(png_ptr->free_fn))(png_ptr, ptr);
+      return;
+   }
+   else png_free_default(png_ptr, ptr);
+}
+
+void PNGAPI
+png_free_default(png_structp png_ptr, png_voidp ptr)
+{
+#endif /* PNG_USER_MEM_SUPPORTED */
+
+   if(png_ptr == NULL) return;
+
+   if (png_ptr->offset_table != NULL)
+   {
+      int i;
+
+      for (i = 0; i < png_ptr->offset_table_count; i++)
+      {
+         if (ptr == png_ptr->offset_table_ptr[i])
+         {
+            ptr = NULL;
+            png_ptr->offset_table_count_free++;
+            break;
+         }
+      }
+      if (png_ptr->offset_table_count_free == png_ptr->offset_table_count)
+      {
+         farfree(png_ptr->offset_table);
+         farfree(png_ptr->offset_table_ptr);
+         png_ptr->offset_table = NULL;
+         png_ptr->offset_table_ptr = NULL;
+      }
+   }
+
+   if (ptr != NULL)
+   {
+      farfree(ptr);
+   }
+}
+
+#else /* Not the Borland DOS special memory handler */
+
+/* Allocate memory for a png_struct or a png_info.  The malloc and
+   memset can be replaced by a single call to calloc() if this is thought
+   to improve performance noticably. */
+png_voidp /* PRIVATE */
+png_create_struct(int type)
+{
+#ifdef PNG_USER_MEM_SUPPORTED
+   return (png_create_struct_2(type, png_malloc_ptr_NULL, png_voidp_NULL));
+}
+
+/* Allocate memory for a png_struct or a png_info.  The malloc and
+   memset can be replaced by a single call to calloc() if this is thought
+   to improve performance noticably. */
+png_voidp /* PRIVATE */
+png_create_struct_2(int type, png_malloc_ptr malloc_fn, png_voidp mem_ptr)
+{
+#endif /* PNG_USER_MEM_SUPPORTED */
+   png_size_t size;
+   png_voidp struct_ptr;
+
+   if (type == PNG_STRUCT_INFO)
+      size = png_sizeof(png_info);
+   else if (type == PNG_STRUCT_PNG)
+      size = png_sizeof(png_struct);
+   else
+      return (NULL);
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   if(malloc_fn != NULL)
+   {
+      png_struct dummy_struct;
+      png_structp png_ptr = &dummy_struct;
+      png_ptr->mem_ptr=mem_ptr;
+      struct_ptr = (*(malloc_fn))(png_ptr, size);
+      if (struct_ptr != NULL)
+         png_memset(struct_ptr, 0, size);
+      return (struct_ptr);
+   }
+#endif /* PNG_USER_MEM_SUPPORTED */
+
+#if defined(__TURBOC__) && !defined(__FLAT__)
+   struct_ptr = (png_voidp)farmalloc(size);
+#else
+# if defined(_MSC_VER) && defined(MAXSEG_64K)
+   struct_ptr = (png_voidp)halloc(size,1);
+# else
+   struct_ptr = (png_voidp)malloc(size);
+# endif
+#endif
+   if (struct_ptr != NULL)
+      png_memset(struct_ptr, 0, size);
+
+   return (struct_ptr);
+}
+
+
+/* Free memory allocated by a png_create_struct() call */
+void /* PRIVATE */
+png_destroy_struct(png_voidp struct_ptr)
+{
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_destroy_struct_2(struct_ptr, png_free_ptr_NULL, png_voidp_NULL);
+}
+
+/* Free memory allocated by a png_create_struct() call */
+void /* PRIVATE */
+png_destroy_struct_2(png_voidp struct_ptr, png_free_ptr free_fn,
+    png_voidp mem_ptr)
+{
+#endif /* PNG_USER_MEM_SUPPORTED */
+   if (struct_ptr != NULL)
+   {
+#ifdef PNG_USER_MEM_SUPPORTED
+      if(free_fn != NULL)
+      {
+         png_struct dummy_struct;
+         png_structp png_ptr = &dummy_struct;
+         png_ptr->mem_ptr=mem_ptr;
+         (*(free_fn))(png_ptr, struct_ptr);
+         return;
+      }
+#endif /* PNG_USER_MEM_SUPPORTED */
+#if defined(__TURBOC__) && !defined(__FLAT__)
+      farfree(struct_ptr);
+#else
+# if defined(_MSC_VER) && defined(MAXSEG_64K)
+      hfree(struct_ptr);
+# else
+      free(struct_ptr);
+# endif
+#endif
+   }
+}
+
+/* Allocate memory.  For reasonable files, size should never exceed
+   64K.  However, zlib may allocate more then 64K if you don't tell
+   it not to.  See zconf.h and png.h for more information.  zlib does
+   need to allocate exactly 64K, so whatever you call here must
+   have the ability to do that. */
+
+png_voidp PNGAPI
+png_malloc(png_structp png_ptr, png_uint_32 size)
+{
+   png_voidp ret;
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   if (png_ptr == NULL || size == 0)
+      return (NULL);
+
+   if(png_ptr->malloc_fn != NULL)
+       ret = ((png_voidp)(*(png_ptr->malloc_fn))(png_ptr, (png_size_t)size));
+   else
+       ret = (png_malloc_default(png_ptr, size));
+   if (ret == NULL && (png_ptr->flags&PNG_FLAG_MALLOC_NULL_MEM_OK) == 0)
+       png_error(png_ptr, "Out of Memory!");
+   return (ret);
+}
+
+png_voidp PNGAPI
+png_malloc_default(png_structp png_ptr, png_uint_32 size)
+{
+   png_voidp ret;
+#endif /* PNG_USER_MEM_SUPPORTED */
+
+   if (png_ptr == NULL || size == 0)
+      return (NULL);
+
+#ifdef PNG_MAX_MALLOC_64K
+   if (size > (png_uint_32)65536L)
+   {
+#ifndef PNG_USER_MEM_SUPPORTED
+      if(png_ptr->flags&PNG_FLAG_MALLOC_NULL_MEM_OK) == 0)
+         png_error(png_ptr, "Cannot Allocate > 64K");
+      else
+#endif
+         return NULL;
+   }
+#endif
+
+ /* Check for overflow */
+#if defined(__TURBOC__) && !defined(__FLAT__)
+ if (size != (unsigned long)size)
+   ret = NULL;
+ else
+   ret = farmalloc(size);
+#else
+# if defined(_MSC_VER) && defined(MAXSEG_64K)
+ if (size != (unsigned long)size)
+   ret = NULL;
+ else
+   ret = halloc(size, 1);
+# else
+ if (size != (size_t)size)
+   ret = NULL;
+ else
+   ret = malloc((size_t)size);
+# endif
+#endif
+
+#ifndef PNG_USER_MEM_SUPPORTED
+   if (ret == NULL && (png_ptr->flags&PNG_FLAG_MALLOC_NULL_MEM_OK) == 0)
+      png_error(png_ptr, "Out of Memory");
+#endif
+
+   return (ret);
+}
+
+/* Free a pointer allocated by png_malloc().  If ptr is NULL, return
+   without taking any action. */
+void PNGAPI
+png_free(png_structp png_ptr, png_voidp ptr)
+{
+   if (png_ptr == NULL || ptr == NULL)
+      return;
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   if (png_ptr->free_fn != NULL)
+   {
+      (*(png_ptr->free_fn))(png_ptr, ptr);
+      return;
+   }
+   else png_free_default(png_ptr, ptr);
+}
+void PNGAPI
+png_free_default(png_structp png_ptr, png_voidp ptr)
+{
+   if (png_ptr == NULL || ptr == NULL)
+      return;
+
+#endif /* PNG_USER_MEM_SUPPORTED */
+
+#if defined(__TURBOC__) && !defined(__FLAT__)
+   farfree(ptr);
+#else
+# if defined(_MSC_VER) && defined(MAXSEG_64K)
+   hfree(ptr);
+# else
+   free(ptr);
+# endif
+#endif
+}
+
+#endif /* Not Borland DOS special memory handler */
+
+#if defined(PNG_1_0_X)
+#  define png_malloc_warn png_malloc
+#else
+/* This function was added at libpng version 1.2.3.  The png_malloc_warn()
+ * function will set up png_malloc() to issue a png_warning and return NULL
+ * instead of issuing a png_error, if it fails to allocate the requested
+ * memory.
+ */
+png_voidp PNGAPI
+png_malloc_warn(png_structp png_ptr, png_uint_32 size)
+{
+   png_voidp ptr;
+   png_uint_32 save_flags;
+   if(png_ptr == NULL) return (NULL);
+
+   save_flags=png_ptr->flags;
+   png_ptr->flags|=PNG_FLAG_MALLOC_NULL_MEM_OK;
+   ptr = (png_voidp)png_malloc((png_structp)png_ptr, size);
+   png_ptr->flags=save_flags;
+   return(ptr);
+}
+#endif
+
+png_voidp PNGAPI
+png_memcpy_check (png_structp png_ptr, png_voidp s1, png_voidp s2,
+   png_uint_32 length)
+{
+   png_size_t size;
+
+   size = (png_size_t)length;
+   if ((png_uint_32)size != length)
+      png_error(png_ptr,"Overflow in png_memcpy_check.");
+
+   return(png_memcpy (s1, s2, size));
+}
+
+png_voidp PNGAPI
+png_memset_check (png_structp png_ptr, png_voidp s1, int value,
+   png_uint_32 length)
+{
+   png_size_t size;
+
+   size = (png_size_t)length;
+   if ((png_uint_32)size != length)
+      png_error(png_ptr,"Overflow in png_memset_check.");
+
+   return (png_memset (s1, value, size));
+
+}
+
+#ifdef PNG_USER_MEM_SUPPORTED
+/* This function is called when the application wants to use another method
+ * of allocating and freeing memory.
+ */
+void PNGAPI
+png_set_mem_fn(png_structp png_ptr, png_voidp mem_ptr, png_malloc_ptr
+  malloc_fn, png_free_ptr free_fn)
+{
+   if(png_ptr != NULL) {
+   png_ptr->mem_ptr = mem_ptr;
+   png_ptr->malloc_fn = malloc_fn;
+   png_ptr->free_fn = free_fn;
+   }
+}
+
+/* This function returns a pointer to the mem_ptr associated with the user
+ * functions.  The application should free any memory associated with this
+ * pointer before png_write_destroy and png_read_destroy are called.
+ */
+png_voidp PNGAPI
+png_get_mem_ptr(png_structp png_ptr)
+{
+   if(png_ptr == NULL) return (NULL);
+   return ((png_voidp)png_ptr->mem_ptr);
+}
+#endif /* PNG_USER_MEM_SUPPORTED */
+#endif /* PNG_READ_SUPPORTED || PNG_WRITE_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngpread.c b/distrib/libpng-1.2.19/pngpread.c
new file mode 100644
index 0000000..7d29e98
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngpread.c
@@ -0,0 +1,1585 @@
+
+/* pngpread.c - read a png file in push mode
+ *
+ * Last changed in libpng 1.2.19 August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#ifdef PNG_PROGRESSIVE_READ_SUPPORTED
+
+/* push model modes */
+#define PNG_READ_SIG_MODE   0
+#define PNG_READ_CHUNK_MODE 1
+#define PNG_READ_IDAT_MODE  2
+#define PNG_SKIP_MODE       3
+#define PNG_READ_tEXt_MODE  4
+#define PNG_READ_zTXt_MODE  5
+#define PNG_READ_DONE_MODE  6
+#define PNG_READ_iTXt_MODE  7
+#define PNG_ERROR_MODE      8
+
+void PNGAPI
+png_process_data(png_structp png_ptr, png_infop info_ptr,
+   png_bytep buffer, png_size_t buffer_size)
+{
+   if(png_ptr == NULL) return;
+   png_push_restore_buffer(png_ptr, buffer, buffer_size);
+
+   while (png_ptr->buffer_size)
+   {
+      png_process_some_data(png_ptr, info_ptr);
+   }
+}
+
+/* What we do with the incoming data depends on what we were previously
+ * doing before we ran out of data...
+ */
+void /* PRIVATE */
+png_process_some_data(png_structp png_ptr, png_infop info_ptr)
+{
+   if(png_ptr == NULL) return;
+   switch (png_ptr->process_mode)
+   {
+      case PNG_READ_SIG_MODE:
+      {
+         png_push_read_sig(png_ptr, info_ptr);
+         break;
+      }
+      case PNG_READ_CHUNK_MODE:
+      {
+         png_push_read_chunk(png_ptr, info_ptr);
+         break;
+      }
+      case PNG_READ_IDAT_MODE:
+      {
+         png_push_read_IDAT(png_ptr);
+         break;
+      }
+#if defined(PNG_READ_tEXt_SUPPORTED)
+      case PNG_READ_tEXt_MODE:
+      {
+         png_push_read_tEXt(png_ptr, info_ptr);
+         break;
+      }
+#endif
+#if defined(PNG_READ_zTXt_SUPPORTED)
+      case PNG_READ_zTXt_MODE:
+      {
+         png_push_read_zTXt(png_ptr, info_ptr);
+         break;
+      }
+#endif
+#if defined(PNG_READ_iTXt_SUPPORTED)
+      case PNG_READ_iTXt_MODE:
+      {
+         png_push_read_iTXt(png_ptr, info_ptr);
+         break;
+      }
+#endif
+      case PNG_SKIP_MODE:
+      {
+         png_push_crc_finish(png_ptr);
+         break;
+      }
+      default:
+      {
+         png_ptr->buffer_size = 0;
+         break;
+      }
+   }
+}
+
+/* Read any remaining signature bytes from the stream and compare them with
+ * the correct PNG signature.  It is possible that this routine is called
+ * with bytes already read from the signature, either because they have been
+ * checked by the calling application, or because of multiple calls to this
+ * routine.
+ */
+void /* PRIVATE */
+png_push_read_sig(png_structp png_ptr, png_infop info_ptr)
+{
+   png_size_t num_checked = png_ptr->sig_bytes,
+             num_to_check = 8 - num_checked;
+
+   if (png_ptr->buffer_size < num_to_check)
+   {
+      num_to_check = png_ptr->buffer_size;
+   }
+
+   png_push_fill_buffer(png_ptr, &(info_ptr->signature[num_checked]),
+      num_to_check);
+   png_ptr->sig_bytes = (png_byte)(png_ptr->sig_bytes+num_to_check);
+
+   if (png_sig_cmp(info_ptr->signature, num_checked, num_to_check))
+   {
+      if (num_checked < 4 &&
+          png_sig_cmp(info_ptr->signature, num_checked, num_to_check - 4))
+         png_error(png_ptr, "Not a PNG file");
+      else
+         png_error(png_ptr, "PNG file corrupted by ASCII conversion");
+   }
+   else
+   {
+      if (png_ptr->sig_bytes >= 8)
+      {
+         png_ptr->process_mode = PNG_READ_CHUNK_MODE;
+      }
+   }
+}
+
+void /* PRIVATE */
+png_push_read_chunk(png_structp png_ptr, png_infop info_ptr)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+      PNG_CONST PNG_IHDR;
+      PNG_CONST PNG_IDAT;
+      PNG_CONST PNG_IEND;
+      PNG_CONST PNG_PLTE;
+#if defined(PNG_READ_bKGD_SUPPORTED)
+      PNG_CONST PNG_bKGD;
+#endif
+#if defined(PNG_READ_cHRM_SUPPORTED)
+      PNG_CONST PNG_cHRM;
+#endif
+#if defined(PNG_READ_gAMA_SUPPORTED)
+      PNG_CONST PNG_gAMA;
+#endif
+#if defined(PNG_READ_hIST_SUPPORTED)
+      PNG_CONST PNG_hIST;
+#endif
+#if defined(PNG_READ_iCCP_SUPPORTED)
+      PNG_CONST PNG_iCCP;
+#endif
+#if defined(PNG_READ_iTXt_SUPPORTED)
+      PNG_CONST PNG_iTXt;
+#endif
+#if defined(PNG_READ_oFFs_SUPPORTED)
+      PNG_CONST PNG_oFFs;
+#endif
+#if defined(PNG_READ_pCAL_SUPPORTED)
+      PNG_CONST PNG_pCAL;
+#endif
+#if defined(PNG_READ_pHYs_SUPPORTED)
+      PNG_CONST PNG_pHYs;
+#endif
+#if defined(PNG_READ_sBIT_SUPPORTED)
+      PNG_CONST PNG_sBIT;
+#endif
+#if defined(PNG_READ_sCAL_SUPPORTED)
+      PNG_CONST PNG_sCAL;
+#endif
+#if defined(PNG_READ_sRGB_SUPPORTED)
+      PNG_CONST PNG_sRGB;
+#endif
+#if defined(PNG_READ_sPLT_SUPPORTED)
+      PNG_CONST PNG_sPLT;
+#endif
+#if defined(PNG_READ_tEXt_SUPPORTED)
+      PNG_CONST PNG_tEXt;
+#endif
+#if defined(PNG_READ_tIME_SUPPORTED)
+      PNG_CONST PNG_tIME;
+#endif
+#if defined(PNG_READ_tRNS_SUPPORTED)
+      PNG_CONST PNG_tRNS;
+#endif
+#if defined(PNG_READ_zTXt_SUPPORTED)
+      PNG_CONST PNG_zTXt;
+#endif
+#endif /* PNG_USE_LOCAL_ARRAYS */
+   /* First we make sure we have enough data for the 4 byte chunk name
+    * and the 4 byte chunk length before proceeding with decoding the
+    * chunk data.  To fully decode each of these chunks, we also make
+    * sure we have enough data in the buffer for the 4 byte CRC at the
+    * end of every chunk (except IDAT, which is handled separately).
+    */
+   if (!(png_ptr->mode & PNG_HAVE_CHUNK_HEADER))
+   {
+      png_byte chunk_length[4];
+
+      if (png_ptr->buffer_size < 8)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+
+      png_push_fill_buffer(png_ptr, chunk_length, 4);
+      png_ptr->push_length = png_get_uint_31(png_ptr,chunk_length);
+      png_reset_crc(png_ptr);
+      png_crc_read(png_ptr, png_ptr->chunk_name, 4);
+      png_ptr->mode |= PNG_HAVE_CHUNK_HEADER;
+   }
+
+   if (!png_memcmp(png_ptr->chunk_name, (png_bytep)png_IDAT, 4))
+     if(png_ptr->mode & PNG_AFTER_IDAT)
+        png_ptr->mode |= PNG_HAVE_CHUNK_AFTER_IDAT;
+
+   if (!png_memcmp(png_ptr->chunk_name, png_IHDR, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_IHDR(png_ptr, info_ptr, png_ptr->push_length);
+   }
+   else if (!png_memcmp(png_ptr->chunk_name, png_IEND, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_IEND(png_ptr, info_ptr, png_ptr->push_length);
+
+      png_ptr->process_mode = PNG_READ_DONE_MODE;
+      png_push_have_end(png_ptr, info_ptr);
+   }
+#ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED
+   else if (png_handle_as_unknown(png_ptr, png_ptr->chunk_name))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      if (!png_memcmp(png_ptr->chunk_name, png_IDAT, 4))
+         png_ptr->mode |= PNG_HAVE_IDAT;
+      png_handle_unknown(png_ptr, info_ptr, png_ptr->push_length);
+      if (!png_memcmp(png_ptr->chunk_name, png_PLTE, 4))
+         png_ptr->mode |= PNG_HAVE_PLTE;
+      else if (!png_memcmp(png_ptr->chunk_name, png_IDAT, 4))
+      {
+         if (!(png_ptr->mode & PNG_HAVE_IHDR))
+            png_error(png_ptr, "Missing IHDR before IDAT");
+         else if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE &&
+                  !(png_ptr->mode & PNG_HAVE_PLTE))
+            png_error(png_ptr, "Missing PLTE before IDAT");
+      }
+   }
+#endif
+   else if (!png_memcmp(png_ptr->chunk_name, png_PLTE, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_PLTE(png_ptr, info_ptr, png_ptr->push_length);
+   }
+   else if (!png_memcmp(png_ptr->chunk_name, (png_bytep)png_IDAT, 4))
+   {
+      /* If we reach an IDAT chunk, this means we have read all of the
+       * header chunks, and we can start reading the image (or if this
+       * is called after the image has been read - we have an error).
+       */
+     if (!(png_ptr->mode & PNG_HAVE_IHDR))
+       png_error(png_ptr, "Missing IHDR before IDAT");
+     else if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE &&
+         !(png_ptr->mode & PNG_HAVE_PLTE))
+       png_error(png_ptr, "Missing PLTE before IDAT");
+
+      if (png_ptr->mode & PNG_HAVE_IDAT)
+      {
+         if (!(png_ptr->mode & PNG_HAVE_CHUNK_AFTER_IDAT))
+           if (png_ptr->push_length == 0)
+              return;
+
+         if (png_ptr->mode & PNG_AFTER_IDAT)
+            png_error(png_ptr, "Too many IDAT's found");
+      }
+
+      png_ptr->idat_size = png_ptr->push_length;
+      png_ptr->mode |= PNG_HAVE_IDAT;
+      png_ptr->process_mode = PNG_READ_IDAT_MODE;
+      png_push_have_info(png_ptr, info_ptr);
+      png_ptr->zstream.avail_out = (uInt)png_ptr->irowbytes;
+      png_ptr->zstream.next_out = png_ptr->row_buf;
+      return;
+   }
+#if defined(PNG_READ_gAMA_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_gAMA, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_gAMA(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_sBIT_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_sBIT, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_sBIT(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_cHRM_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_cHRM, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_cHRM(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_sRGB_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_sRGB, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_sRGB(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_iCCP_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_iCCP, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_iCCP(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_sPLT_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_sPLT, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_sPLT(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_tRNS_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_tRNS, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_tRNS(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_bKGD_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_bKGD, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_bKGD(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_hIST_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_hIST, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_hIST(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_pHYs_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_pHYs, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_pHYs(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_oFFs_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_oFFs, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_oFFs(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_pCAL_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_pCAL, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_pCAL(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_sCAL_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_sCAL, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_sCAL(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_tIME_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_tIME, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_handle_tIME(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_tEXt_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_tEXt, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_push_handle_tEXt(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_zTXt_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_zTXt, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_push_handle_zTXt(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+#if defined(PNG_READ_iTXt_SUPPORTED)
+   else if (!png_memcmp(png_ptr->chunk_name, png_iTXt, 4))
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_push_handle_iTXt(png_ptr, info_ptr, png_ptr->push_length);
+   }
+#endif
+   else
+   {
+      if (png_ptr->push_length + 4 > png_ptr->buffer_size)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+      png_push_handle_unknown(png_ptr, info_ptr, png_ptr->push_length);
+   }
+
+   png_ptr->mode &= ~PNG_HAVE_CHUNK_HEADER;
+}
+
+void /* PRIVATE */
+png_push_crc_skip(png_structp png_ptr, png_uint_32 skip)
+{
+   png_ptr->process_mode = PNG_SKIP_MODE;
+   png_ptr->skip_length = skip;
+}
+
+void /* PRIVATE */
+png_push_crc_finish(png_structp png_ptr)
+{
+   if (png_ptr->skip_length && png_ptr->save_buffer_size)
+   {
+      png_size_t save_size;
+
+      if (png_ptr->skip_length < (png_uint_32)png_ptr->save_buffer_size)
+         save_size = (png_size_t)png_ptr->skip_length;
+      else
+         save_size = png_ptr->save_buffer_size;
+
+      png_calculate_crc(png_ptr, png_ptr->save_buffer_ptr, save_size);
+
+      png_ptr->skip_length -= save_size;
+      png_ptr->buffer_size -= save_size;
+      png_ptr->save_buffer_size -= save_size;
+      png_ptr->save_buffer_ptr += save_size;
+   }
+   if (png_ptr->skip_length && png_ptr->current_buffer_size)
+   {
+      png_size_t save_size;
+
+      if (png_ptr->skip_length < (png_uint_32)png_ptr->current_buffer_size)
+         save_size = (png_size_t)png_ptr->skip_length;
+      else
+         save_size = png_ptr->current_buffer_size;
+
+      png_calculate_crc(png_ptr, png_ptr->current_buffer_ptr, save_size);
+
+      png_ptr->skip_length -= save_size;
+      png_ptr->buffer_size -= save_size;
+      png_ptr->current_buffer_size -= save_size;
+      png_ptr->current_buffer_ptr += save_size;
+   }
+   if (!png_ptr->skip_length)
+   {
+      if (png_ptr->buffer_size < 4)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+
+      png_crc_finish(png_ptr, 0);
+      png_ptr->process_mode = PNG_READ_CHUNK_MODE;
+   }
+}
+
+void PNGAPI
+png_push_fill_buffer(png_structp png_ptr, png_bytep buffer, png_size_t length)
+{
+   png_bytep ptr;
+
+   if(png_ptr == NULL) return;
+   ptr = buffer;
+   if (png_ptr->save_buffer_size)
+   {
+      png_size_t save_size;
+
+      if (length < png_ptr->save_buffer_size)
+         save_size = length;
+      else
+         save_size = png_ptr->save_buffer_size;
+
+      png_memcpy(ptr, png_ptr->save_buffer_ptr, save_size);
+      length -= save_size;
+      ptr += save_size;
+      png_ptr->buffer_size -= save_size;
+      png_ptr->save_buffer_size -= save_size;
+      png_ptr->save_buffer_ptr += save_size;
+   }
+   if (length && png_ptr->current_buffer_size)
+   {
+      png_size_t save_size;
+
+      if (length < png_ptr->current_buffer_size)
+         save_size = length;
+      else
+         save_size = png_ptr->current_buffer_size;
+
+      png_memcpy(ptr, png_ptr->current_buffer_ptr, save_size);
+      png_ptr->buffer_size -= save_size;
+      png_ptr->current_buffer_size -= save_size;
+      png_ptr->current_buffer_ptr += save_size;
+   }
+}
+
+void /* PRIVATE */
+png_push_save_buffer(png_structp png_ptr)
+{
+   if (png_ptr->save_buffer_size)
+   {
+      if (png_ptr->save_buffer_ptr != png_ptr->save_buffer)
+      {
+         png_size_t i,istop;
+         png_bytep sp;
+         png_bytep dp;
+
+         istop = png_ptr->save_buffer_size;
+         for (i = 0, sp = png_ptr->save_buffer_ptr, dp = png_ptr->save_buffer;
+            i < istop; i++, sp++, dp++)
+         {
+            *dp = *sp;
+         }
+      }
+   }
+   if (png_ptr->save_buffer_size + png_ptr->current_buffer_size >
+      png_ptr->save_buffer_max)
+   {
+      png_size_t new_max;
+      png_bytep old_buffer;
+
+      if (png_ptr->save_buffer_size > PNG_SIZE_MAX -
+         (png_ptr->current_buffer_size + 256))
+      {
+        png_error(png_ptr, "Potential overflow of save_buffer");
+      }
+      new_max = png_ptr->save_buffer_size + png_ptr->current_buffer_size + 256;
+      old_buffer = png_ptr->save_buffer;
+      png_ptr->save_buffer = (png_bytep)png_malloc(png_ptr,
+         (png_uint_32)new_max);
+      png_memcpy(png_ptr->save_buffer, old_buffer, png_ptr->save_buffer_size);
+      png_free(png_ptr, old_buffer);
+      png_ptr->save_buffer_max = new_max;
+   }
+   if (png_ptr->current_buffer_size)
+   {
+      png_memcpy(png_ptr->save_buffer + png_ptr->save_buffer_size,
+         png_ptr->current_buffer_ptr, png_ptr->current_buffer_size);
+      png_ptr->save_buffer_size += png_ptr->current_buffer_size;
+      png_ptr->current_buffer_size = 0;
+   }
+   png_ptr->save_buffer_ptr = png_ptr->save_buffer;
+   png_ptr->buffer_size = 0;
+}
+
+void /* PRIVATE */
+png_push_restore_buffer(png_structp png_ptr, png_bytep buffer,
+   png_size_t buffer_length)
+{
+   png_ptr->current_buffer = buffer;
+   png_ptr->current_buffer_size = buffer_length;
+   png_ptr->buffer_size = buffer_length + png_ptr->save_buffer_size;
+   png_ptr->current_buffer_ptr = png_ptr->current_buffer;
+}
+
+void /* PRIVATE */
+png_push_read_IDAT(png_structp png_ptr)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_CONST PNG_IDAT;
+#endif
+   if (!(png_ptr->mode & PNG_HAVE_CHUNK_HEADER))
+   {
+      png_byte chunk_length[4];
+
+      if (png_ptr->buffer_size < 8)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+
+      png_push_fill_buffer(png_ptr, chunk_length, 4);
+      png_ptr->push_length = png_get_uint_31(png_ptr,chunk_length);
+      png_reset_crc(png_ptr);
+      png_crc_read(png_ptr, png_ptr->chunk_name, 4);
+      png_ptr->mode |= PNG_HAVE_CHUNK_HEADER;
+
+      if (png_memcmp(png_ptr->chunk_name, (png_bytep)png_IDAT, 4))
+      {
+         png_ptr->process_mode = PNG_READ_CHUNK_MODE;
+         if (!(png_ptr->flags & PNG_FLAG_ZLIB_FINISHED))
+            png_error(png_ptr, "Not enough compressed data");
+         return;
+      }
+
+      png_ptr->idat_size = png_ptr->push_length;
+   }
+   if (png_ptr->idat_size && png_ptr->save_buffer_size)
+   {
+      png_size_t save_size;
+
+      if (png_ptr->idat_size < (png_uint_32)png_ptr->save_buffer_size)
+      {
+         save_size = (png_size_t)png_ptr->idat_size;
+         /* check for overflow */
+         if((png_uint_32)save_size != png_ptr->idat_size)
+            png_error(png_ptr, "save_size overflowed in pngpread");
+      }
+      else
+         save_size = png_ptr->save_buffer_size;
+
+      png_calculate_crc(png_ptr, png_ptr->save_buffer_ptr, save_size);
+      if (!(png_ptr->flags & PNG_FLAG_ZLIB_FINISHED))
+         png_process_IDAT_data(png_ptr, png_ptr->save_buffer_ptr, save_size);
+      png_ptr->idat_size -= save_size;
+      png_ptr->buffer_size -= save_size;
+      png_ptr->save_buffer_size -= save_size;
+      png_ptr->save_buffer_ptr += save_size;
+   }
+   if (png_ptr->idat_size && png_ptr->current_buffer_size)
+   {
+      png_size_t save_size;
+
+      if (png_ptr->idat_size < (png_uint_32)png_ptr->current_buffer_size)
+      {
+         save_size = (png_size_t)png_ptr->idat_size;
+         /* check for overflow */
+         if((png_uint_32)save_size != png_ptr->idat_size)
+            png_error(png_ptr, "save_size overflowed in pngpread");
+      }
+      else
+         save_size = png_ptr->current_buffer_size;
+
+      png_calculate_crc(png_ptr, png_ptr->current_buffer_ptr, save_size);
+      if (!(png_ptr->flags & PNG_FLAG_ZLIB_FINISHED))
+        png_process_IDAT_data(png_ptr, png_ptr->current_buffer_ptr, save_size);
+
+      png_ptr->idat_size -= save_size;
+      png_ptr->buffer_size -= save_size;
+      png_ptr->current_buffer_size -= save_size;
+      png_ptr->current_buffer_ptr += save_size;
+   }
+   if (!png_ptr->idat_size)
+   {
+      if (png_ptr->buffer_size < 4)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+
+      png_crc_finish(png_ptr, 0);
+      png_ptr->mode &= ~PNG_HAVE_CHUNK_HEADER;
+      png_ptr->mode |= PNG_AFTER_IDAT;
+   }
+}
+
+void /* PRIVATE */
+png_process_IDAT_data(png_structp png_ptr, png_bytep buffer,
+   png_size_t buffer_length)
+{
+   int ret;
+
+   if ((png_ptr->flags & PNG_FLAG_ZLIB_FINISHED) && buffer_length)
+      png_error(png_ptr, "Extra compression data");
+
+   png_ptr->zstream.next_in = buffer;
+   png_ptr->zstream.avail_in = (uInt)buffer_length;
+   for(;;)
+   {
+      ret = inflate(&png_ptr->zstream, Z_PARTIAL_FLUSH);
+      if (ret != Z_OK)
+      {
+         if (ret == Z_STREAM_END)
+         {
+            if (png_ptr->zstream.avail_in)
+               png_error(png_ptr, "Extra compressed data");
+            if (!(png_ptr->zstream.avail_out))
+            {
+               png_push_process_row(png_ptr);
+            }
+
+            png_ptr->mode |= PNG_AFTER_IDAT;
+            png_ptr->flags |= PNG_FLAG_ZLIB_FINISHED;
+            break;
+         }
+         else if (ret == Z_BUF_ERROR)
+            break;
+         else
+            png_error(png_ptr, "Decompression Error");
+      }
+      if (!(png_ptr->zstream.avail_out))
+      {
+         if ((
+#if defined(PNG_READ_INTERLACING_SUPPORTED)
+             png_ptr->interlaced && png_ptr->pass > 6) ||
+             (!png_ptr->interlaced &&
+#endif
+             png_ptr->row_number == png_ptr->num_rows))
+         {
+           if (png_ptr->zstream.avail_in)
+             png_warning(png_ptr, "Too much data in IDAT chunks");
+           png_ptr->flags |= PNG_FLAG_ZLIB_FINISHED;
+           break;
+         }
+         png_push_process_row(png_ptr);
+         png_ptr->zstream.avail_out = (uInt)png_ptr->irowbytes;
+         png_ptr->zstream.next_out = png_ptr->row_buf;
+      }
+      else
+         break;
+   }
+}
+
+void /* PRIVATE */
+png_push_process_row(png_structp png_ptr)
+{
+   png_ptr->row_info.color_type = png_ptr->color_type;
+   png_ptr->row_info.width = png_ptr->iwidth;
+   png_ptr->row_info.channels = png_ptr->channels;
+   png_ptr->row_info.bit_depth = png_ptr->bit_depth;
+   png_ptr->row_info.pixel_depth = png_ptr->pixel_depth;
+
+   png_ptr->row_info.rowbytes = PNG_ROWBYTES(png_ptr->row_info.pixel_depth,
+       png_ptr->row_info.width);
+
+   png_read_filter_row(png_ptr, &(png_ptr->row_info),
+      png_ptr->row_buf + 1, png_ptr->prev_row + 1,
+      (int)(png_ptr->row_buf[0]));
+
+   png_memcpy_check(png_ptr, png_ptr->prev_row, png_ptr->row_buf,
+      png_ptr->rowbytes + 1);
+
+   if (png_ptr->transformations || (png_ptr->flags&PNG_FLAG_STRIP_ALPHA))
+      png_do_read_transformations(png_ptr);
+
+#if defined(PNG_READ_INTERLACING_SUPPORTED)
+   /* blow up interlaced rows to full size */
+   if (png_ptr->interlaced && (png_ptr->transformations & PNG_INTERLACE))
+   {
+      if (png_ptr->pass < 6)
+/*       old interface (pre-1.0.9):
+         png_do_read_interlace(&(png_ptr->row_info),
+            png_ptr->row_buf + 1, png_ptr->pass, png_ptr->transformations);
+ */
+         png_do_read_interlace(png_ptr);
+
+    switch (png_ptr->pass)
+    {
+         case 0:
+         {
+            int i;
+            for (i = 0; i < 8 && png_ptr->pass == 0; i++)
+            {
+               png_push_have_row(png_ptr, png_ptr->row_buf + 1);
+               png_read_push_finish_row(png_ptr); /* updates png_ptr->pass */
+            }
+            if (png_ptr->pass == 2) /* pass 1 might be empty */
+            {
+               for (i = 0; i < 4 && png_ptr->pass == 2; i++)
+               {
+                  png_push_have_row(png_ptr, png_bytep_NULL);
+                  png_read_push_finish_row(png_ptr);
+               }
+            }
+            if (png_ptr->pass == 4 && png_ptr->height <= 4)
+            {
+               for (i = 0; i < 2 && png_ptr->pass == 4; i++)
+               {
+                  png_push_have_row(png_ptr, png_bytep_NULL);
+                  png_read_push_finish_row(png_ptr);
+               }
+            }
+            if (png_ptr->pass == 6 && png_ptr->height <= 4)
+            {
+                png_push_have_row(png_ptr, png_bytep_NULL);
+                png_read_push_finish_row(png_ptr);
+            }
+            break;
+         }
+         case 1:
+         {
+            int i;
+            for (i = 0; i < 8 && png_ptr->pass == 1; i++)
+            {
+               png_push_have_row(png_ptr, png_ptr->row_buf + 1);
+               png_read_push_finish_row(png_ptr);
+            }
+            if (png_ptr->pass == 2) /* skip top 4 generated rows */
+            {
+               for (i = 0; i < 4 && png_ptr->pass == 2; i++)
+               {
+                  png_push_have_row(png_ptr, png_bytep_NULL);
+                  png_read_push_finish_row(png_ptr);
+               }
+            }
+            break;
+         }
+         case 2:
+         {
+            int i;
+            for (i = 0; i < 4 && png_ptr->pass == 2; i++)
+            {
+               png_push_have_row(png_ptr, png_ptr->row_buf + 1);
+               png_read_push_finish_row(png_ptr);
+            }
+            for (i = 0; i < 4 && png_ptr->pass == 2; i++)
+            {
+               png_push_have_row(png_ptr, png_bytep_NULL);
+               png_read_push_finish_row(png_ptr);
+            }
+            if (png_ptr->pass == 4) /* pass 3 might be empty */
+            {
+               for (i = 0; i < 2 && png_ptr->pass == 4; i++)
+               {
+                  png_push_have_row(png_ptr, png_bytep_NULL);
+                  png_read_push_finish_row(png_ptr);
+               }
+            }
+            break;
+         }
+         case 3:
+         {
+            int i;
+            for (i = 0; i < 4 && png_ptr->pass == 3; i++)
+            {
+               png_push_have_row(png_ptr, png_ptr->row_buf + 1);
+               png_read_push_finish_row(png_ptr);
+            }
+            if (png_ptr->pass == 4) /* skip top two generated rows */
+            {
+               for (i = 0; i < 2 && png_ptr->pass == 4; i++)
+               {
+                  png_push_have_row(png_ptr, png_bytep_NULL);
+                  png_read_push_finish_row(png_ptr);
+               }
+            }
+            break;
+         }
+         case 4:
+         {
+            int i;
+            for (i = 0; i < 2 && png_ptr->pass == 4; i++)
+            {
+               png_push_have_row(png_ptr, png_ptr->row_buf + 1);
+               png_read_push_finish_row(png_ptr);
+            }
+            for (i = 0; i < 2 && png_ptr->pass == 4; i++)
+            {
+               png_push_have_row(png_ptr, png_bytep_NULL);
+               png_read_push_finish_row(png_ptr);
+            }
+            if (png_ptr->pass == 6) /* pass 5 might be empty */
+            {
+               png_push_have_row(png_ptr, png_bytep_NULL);
+               png_read_push_finish_row(png_ptr);
+            }
+            break;
+         }
+         case 5:
+         {
+            int i;
+            for (i = 0; i < 2 && png_ptr->pass == 5; i++)
+            {
+               png_push_have_row(png_ptr, png_ptr->row_buf + 1);
+               png_read_push_finish_row(png_ptr);
+            }
+            if (png_ptr->pass == 6) /* skip top generated row */
+            {
+               png_push_have_row(png_ptr, png_bytep_NULL);
+               png_read_push_finish_row(png_ptr);
+            }
+            break;
+         }
+         case 6:
+         {
+            png_push_have_row(png_ptr, png_ptr->row_buf + 1);
+            png_read_push_finish_row(png_ptr);
+            if (png_ptr->pass != 6)
+               break;
+            png_push_have_row(png_ptr, png_bytep_NULL);
+            png_read_push_finish_row(png_ptr);
+         }
+      }
+   }
+   else
+#endif
+   {
+      png_push_have_row(png_ptr, png_ptr->row_buf + 1);
+      png_read_push_finish_row(png_ptr);
+   }
+}
+
+void /* PRIVATE */
+png_read_push_finish_row(png_structp png_ptr)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   /* arrays to facilitate easy interlacing - use pass (0 - 6) as index */
+
+   /* start of interlace block */
+   PNG_CONST int FARDATA png_pass_start[] = {0, 4, 0, 2, 0, 1, 0};
+
+   /* offset to next interlace block */
+   PNG_CONST int FARDATA png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1};
+
+   /* start of interlace block in the y direction */
+   PNG_CONST int FARDATA png_pass_ystart[] = {0, 0, 4, 0, 2, 0, 1};
+
+   /* offset to next interlace block in the y direction */
+   PNG_CONST int FARDATA png_pass_yinc[] = {8, 8, 8, 4, 4, 2, 2};
+
+   /* Width of interlace block.  This is not currently used - if you need
+    * it, uncomment it here and in png.h
+   PNG_CONST int FARDATA png_pass_width[] = {8, 4, 4, 2, 2, 1, 1};
+   */
+
+   /* Height of interlace block.  This is not currently used - if you need
+    * it, uncomment it here and in png.h
+   PNG_CONST int FARDATA png_pass_height[] = {8, 8, 4, 4, 2, 2, 1};
+   */
+#endif
+
+   png_ptr->row_number++;
+   if (png_ptr->row_number < png_ptr->num_rows)
+      return;
+
+   if (png_ptr->interlaced)
+   {
+      png_ptr->row_number = 0;
+      png_memset_check(png_ptr, png_ptr->prev_row, 0,
+         png_ptr->rowbytes + 1);
+      do
+      {
+         png_ptr->pass++;
+         if ((png_ptr->pass == 1 && png_ptr->width < 5) ||
+             (png_ptr->pass == 3 && png_ptr->width < 3) ||
+             (png_ptr->pass == 5 && png_ptr->width < 2))
+           png_ptr->pass++;
+
+         if (png_ptr->pass > 7)
+            png_ptr->pass--;
+         if (png_ptr->pass >= 7)
+            break;
+
+         png_ptr->iwidth = (png_ptr->width +
+            png_pass_inc[png_ptr->pass] - 1 -
+            png_pass_start[png_ptr->pass]) /
+            png_pass_inc[png_ptr->pass];
+
+         png_ptr->irowbytes = PNG_ROWBYTES(png_ptr->pixel_depth,
+            png_ptr->iwidth) + 1;
+
+         if (png_ptr->transformations & PNG_INTERLACE)
+            break;
+
+         png_ptr->num_rows = (png_ptr->height +
+            png_pass_yinc[png_ptr->pass] - 1 -
+            png_pass_ystart[png_ptr->pass]) /
+            png_pass_yinc[png_ptr->pass];
+
+      } while (png_ptr->iwidth == 0 || png_ptr->num_rows == 0);
+   }
+}
+
+#if defined(PNG_READ_tEXt_SUPPORTED)
+void /* PRIVATE */
+png_push_handle_tEXt(png_structp png_ptr, png_infop info_ptr, png_uint_32
+   length)
+{
+   if (!(png_ptr->mode & PNG_HAVE_IHDR) || (png_ptr->mode & PNG_HAVE_IEND))
+      {
+         png_error(png_ptr, "Out of place tEXt");
+         info_ptr = info_ptr; /* to quiet some compiler warnings */
+      }
+
+#ifdef PNG_MAX_MALLOC_64K
+   png_ptr->skip_length = 0;  /* This may not be necessary */
+
+   if (length > (png_uint_32)65535L) /* Can't hold entire string in memory */
+   {
+      png_warning(png_ptr, "tEXt chunk too large to fit in memory");
+      png_ptr->skip_length = length - (png_uint_32)65535L;
+      length = (png_uint_32)65535L;
+   }
+#endif
+
+   png_ptr->current_text = (png_charp)png_malloc(png_ptr,
+         (png_uint_32)(length+1));
+   png_ptr->current_text[length] = '\0';
+   png_ptr->current_text_ptr = png_ptr->current_text;
+   png_ptr->current_text_size = (png_size_t)length;
+   png_ptr->current_text_left = (png_size_t)length;
+   png_ptr->process_mode = PNG_READ_tEXt_MODE;
+}
+
+void /* PRIVATE */
+png_push_read_tEXt(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr->buffer_size && png_ptr->current_text_left)
+   {
+      png_size_t text_size;
+
+      if (png_ptr->buffer_size < png_ptr->current_text_left)
+         text_size = png_ptr->buffer_size;
+      else
+         text_size = png_ptr->current_text_left;
+      png_crc_read(png_ptr, (png_bytep)png_ptr->current_text_ptr, text_size);
+      png_ptr->current_text_left -= text_size;
+      png_ptr->current_text_ptr += text_size;
+   }
+   if (!(png_ptr->current_text_left))
+   {
+      png_textp text_ptr;
+      png_charp text;
+      png_charp key;
+      int ret;
+
+      if (png_ptr->buffer_size < 4)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+
+      png_push_crc_finish(png_ptr);
+
+#if defined(PNG_MAX_MALLOC_64K)
+      if (png_ptr->skip_length)
+         return;
+#endif
+
+      key = png_ptr->current_text;
+
+      for (text = key; *text; text++)
+         /* empty loop */ ;
+
+      if (text != key + png_ptr->current_text_size)
+         text++;
+
+      text_ptr = (png_textp)png_malloc(png_ptr,
+         (png_uint_32)png_sizeof(png_text));
+      text_ptr->compression = PNG_TEXT_COMPRESSION_NONE;
+      text_ptr->key = key;
+#ifdef PNG_iTXt_SUPPORTED
+      text_ptr->lang = NULL;
+      text_ptr->lang_key = NULL;
+#endif
+      text_ptr->text = text;
+
+      ret = png_set_text_2(png_ptr, info_ptr, text_ptr, 1);
+
+      png_free(png_ptr, key);
+      png_free(png_ptr, text_ptr);
+      png_ptr->current_text = NULL;
+
+      if (ret)
+        png_warning(png_ptr, "Insufficient memory to store text chunk.");
+   }
+}
+#endif
+
+#if defined(PNG_READ_zTXt_SUPPORTED)
+void /* PRIVATE */
+png_push_handle_zTXt(png_structp png_ptr, png_infop info_ptr, png_uint_32
+   length)
+{
+   if (!(png_ptr->mode & PNG_HAVE_IHDR) || (png_ptr->mode & PNG_HAVE_IEND))
+      {
+         png_error(png_ptr, "Out of place zTXt");
+         info_ptr = info_ptr; /* to quiet some compiler warnings */
+      }
+
+#ifdef PNG_MAX_MALLOC_64K
+   /* We can't handle zTXt chunks > 64K, since we don't have enough space
+    * to be able to store the uncompressed data.  Actually, the threshold
+    * is probably around 32K, but it isn't as definite as 64K is.
+    */
+   if (length > (png_uint_32)65535L)
+   {
+      png_warning(png_ptr, "zTXt chunk too large to fit in memory");
+      png_push_crc_skip(png_ptr, length);
+      return;
+   }
+#endif
+
+   png_ptr->current_text = (png_charp)png_malloc(png_ptr,
+       (png_uint_32)(length+1));
+   png_ptr->current_text[length] = '\0';
+   png_ptr->current_text_ptr = png_ptr->current_text;
+   png_ptr->current_text_size = (png_size_t)length;
+   png_ptr->current_text_left = (png_size_t)length;
+   png_ptr->process_mode = PNG_READ_zTXt_MODE;
+}
+
+void /* PRIVATE */
+png_push_read_zTXt(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr->buffer_size && png_ptr->current_text_left)
+   {
+      png_size_t text_size;
+
+      if (png_ptr->buffer_size < (png_uint_32)png_ptr->current_text_left)
+         text_size = png_ptr->buffer_size;
+      else
+         text_size = png_ptr->current_text_left;
+      png_crc_read(png_ptr, (png_bytep)png_ptr->current_text_ptr, text_size);
+      png_ptr->current_text_left -= text_size;
+      png_ptr->current_text_ptr += text_size;
+   }
+   if (!(png_ptr->current_text_left))
+   {
+      png_textp text_ptr;
+      png_charp text;
+      png_charp key;
+      int ret;
+      png_size_t text_size, key_size;
+
+      if (png_ptr->buffer_size < 4)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+
+      png_push_crc_finish(png_ptr);
+
+      key = png_ptr->current_text;
+
+      for (text = key; *text; text++)
+         /* empty loop */ ;
+
+      /* zTXt can't have zero text */
+      if (text == key + png_ptr->current_text_size)
+      {
+         png_ptr->current_text = NULL;
+         png_free(png_ptr, key);
+         return;
+      }
+
+      text++;
+
+      if (*text != PNG_TEXT_COMPRESSION_zTXt) /* check compression byte */
+      {
+         png_ptr->current_text = NULL;
+         png_free(png_ptr, key);
+         return;
+      }
+
+      text++;
+
+      png_ptr->zstream.next_in = (png_bytep )text;
+      png_ptr->zstream.avail_in = (uInt)(png_ptr->current_text_size -
+         (text - key));
+      png_ptr->zstream.next_out = png_ptr->zbuf;
+      png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+
+      key_size = text - key;
+      text_size = 0;
+      text = NULL;
+      ret = Z_STREAM_END;
+
+      while (png_ptr->zstream.avail_in)
+      {
+         ret = inflate(&png_ptr->zstream, Z_PARTIAL_FLUSH);
+         if (ret != Z_OK && ret != Z_STREAM_END)
+         {
+            inflateReset(&png_ptr->zstream);
+            png_ptr->zstream.avail_in = 0;
+            png_ptr->current_text = NULL;
+            png_free(png_ptr, key);
+            png_free(png_ptr, text);
+            return;
+         }
+         if (!(png_ptr->zstream.avail_out) || ret == Z_STREAM_END)
+         {
+            if (text == NULL)
+            {
+               text = (png_charp)png_malloc(png_ptr,
+                  (png_uint_32)(png_ptr->zbuf_size - png_ptr->zstream.avail_out
+                     + key_size + 1));
+               png_memcpy(text + key_size, png_ptr->zbuf,
+                  png_ptr->zbuf_size - png_ptr->zstream.avail_out);
+               png_memcpy(text, key, key_size);
+               text_size = key_size + png_ptr->zbuf_size -
+                  png_ptr->zstream.avail_out;
+               *(text + text_size) = '\0';
+            }
+            else
+            {
+               png_charp tmp;
+
+               tmp = text;
+               text = (png_charp)png_malloc(png_ptr, text_size +
+                  (png_uint_32)(png_ptr->zbuf_size - png_ptr->zstream.avail_out
+                   + 1));
+               png_memcpy(text, tmp, text_size);
+               png_free(png_ptr, tmp);
+               png_memcpy(text + text_size, png_ptr->zbuf,
+                  png_ptr->zbuf_size - png_ptr->zstream.avail_out);
+               text_size += png_ptr->zbuf_size - png_ptr->zstream.avail_out;
+               *(text + text_size) = '\0';
+            }
+            if (ret != Z_STREAM_END)
+            {
+               png_ptr->zstream.next_out = png_ptr->zbuf;
+               png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+            }
+         }
+         else
+         {
+            break;
+         }
+
+         if (ret == Z_STREAM_END)
+            break;
+      }
+
+      inflateReset(&png_ptr->zstream);
+      png_ptr->zstream.avail_in = 0;
+
+      if (ret != Z_STREAM_END)
+      {
+         png_ptr->current_text = NULL;
+         png_free(png_ptr, key);
+         png_free(png_ptr, text);
+         return;
+      }
+
+      png_ptr->current_text = NULL;
+      png_free(png_ptr, key);
+      key = text;
+      text += key_size;
+
+      text_ptr = (png_textp)png_malloc(png_ptr,
+          (png_uint_32)png_sizeof(png_text));
+      text_ptr->compression = PNG_TEXT_COMPRESSION_zTXt;
+      text_ptr->key = key;
+#ifdef PNG_iTXt_SUPPORTED
+      text_ptr->lang = NULL;
+      text_ptr->lang_key = NULL;
+#endif
+      text_ptr->text = text;
+
+      ret = png_set_text_2(png_ptr, info_ptr, text_ptr, 1);
+
+      png_free(png_ptr, key);
+      png_free(png_ptr, text_ptr);
+
+      if (ret)
+        png_warning(png_ptr, "Insufficient memory to store text chunk.");
+   }
+}
+#endif
+
+#if defined(PNG_READ_iTXt_SUPPORTED)
+void /* PRIVATE */
+png_push_handle_iTXt(png_structp png_ptr, png_infop info_ptr, png_uint_32
+   length)
+{
+   if (!(png_ptr->mode & PNG_HAVE_IHDR) || (png_ptr->mode & PNG_HAVE_IEND))
+      {
+         png_error(png_ptr, "Out of place iTXt");
+         info_ptr = info_ptr; /* to quiet some compiler warnings */
+      }
+
+#ifdef PNG_MAX_MALLOC_64K
+   png_ptr->skip_length = 0;  /* This may not be necessary */
+
+   if (length > (png_uint_32)65535L) /* Can't hold entire string in memory */
+   {
+      png_warning(png_ptr, "iTXt chunk too large to fit in memory");
+      png_ptr->skip_length = length - (png_uint_32)65535L;
+      length = (png_uint_32)65535L;
+   }
+#endif
+
+   png_ptr->current_text = (png_charp)png_malloc(png_ptr,
+         (png_uint_32)(length+1));
+   png_ptr->current_text[length] = '\0';
+   png_ptr->current_text_ptr = png_ptr->current_text;
+   png_ptr->current_text_size = (png_size_t)length;
+   png_ptr->current_text_left = (png_size_t)length;
+   png_ptr->process_mode = PNG_READ_iTXt_MODE;
+}
+
+void /* PRIVATE */
+png_push_read_iTXt(png_structp png_ptr, png_infop info_ptr)
+{
+
+   if (png_ptr->buffer_size && png_ptr->current_text_left)
+   {
+      png_size_t text_size;
+
+      if (png_ptr->buffer_size < png_ptr->current_text_left)
+         text_size = png_ptr->buffer_size;
+      else
+         text_size = png_ptr->current_text_left;
+      png_crc_read(png_ptr, (png_bytep)png_ptr->current_text_ptr, text_size);
+      png_ptr->current_text_left -= text_size;
+      png_ptr->current_text_ptr += text_size;
+   }
+   if (!(png_ptr->current_text_left))
+   {
+      png_textp text_ptr;
+      png_charp key;
+      int comp_flag;
+      png_charp lang;
+      png_charp lang_key;
+      png_charp text;
+      int ret;
+
+      if (png_ptr->buffer_size < 4)
+      {
+         png_push_save_buffer(png_ptr);
+         return;
+      }
+
+      png_push_crc_finish(png_ptr);
+
+#if defined(PNG_MAX_MALLOC_64K)
+      if (png_ptr->skip_length)
+         return;
+#endif
+
+      key = png_ptr->current_text;
+
+      for (lang = key; *lang; lang++)
+         /* empty loop */ ;
+
+      if (lang != key + png_ptr->current_text_size)
+         lang++;
+
+      comp_flag = *lang++;
+      lang++;     /* skip comp_type, always zero */
+
+      for (lang_key = lang; *lang_key; lang_key++)
+         /* empty loop */ ;
+      lang_key++;        /* skip NUL separator */
+
+      for (text = lang_key; *text; text++)
+         /* empty loop */ ;
+
+      if (text != key + png_ptr->current_text_size)
+         text++;
+
+      text_ptr = (png_textp)png_malloc(png_ptr,
+         (png_uint_32)png_sizeof(png_text));
+      text_ptr->compression = comp_flag + 2;
+      text_ptr->key = key;
+      text_ptr->lang = lang;
+      text_ptr->lang_key = lang_key;
+      text_ptr->text = text;
+      text_ptr->text_length = 0;
+      text_ptr->itxt_length = png_strlen(text);
+
+      ret = png_set_text_2(png_ptr, info_ptr, text_ptr, 1);
+
+      png_ptr->current_text = NULL;
+
+      png_free(png_ptr, text_ptr);
+      if (ret)
+        png_warning(png_ptr, "Insufficient memory to store iTXt chunk.");
+   }
+}
+#endif
+
+/* This function is called when we haven't found a handler for this
+ * chunk.  If there isn't a problem with the chunk itself (ie a bad chunk
+ * name or a critical chunk), the chunk is (currently) silently ignored.
+ */
+void /* PRIVATE */
+png_push_handle_unknown(png_structp png_ptr, png_infop info_ptr, png_uint_32
+   length)
+{
+   png_uint_32 skip=0;
+   png_check_chunk_name(png_ptr, png_ptr->chunk_name);
+
+   if (!(png_ptr->chunk_name[0] & 0x20))
+   {
+#if defined(PNG_READ_UNKNOWN_CHUNKS_SUPPORTED)
+      if(png_handle_as_unknown(png_ptr, png_ptr->chunk_name) !=
+           PNG_HANDLE_CHUNK_ALWAYS
+#if defined(PNG_READ_USER_CHUNKS_SUPPORTED)
+           && png_ptr->read_user_chunk_fn == NULL
+#endif
+         )
+#endif
+         png_chunk_error(png_ptr, "unknown critical chunk");
+
+      info_ptr = info_ptr; /* to quiet some compiler warnings */
+   }
+
+#if defined(PNG_READ_UNKNOWN_CHUNKS_SUPPORTED)
+   if (png_ptr->flags & PNG_FLAG_KEEP_UNKNOWN_CHUNKS)
+   {
+#ifdef PNG_MAX_MALLOC_64K
+       if (length > (png_uint_32)65535L)
+       {
+           png_warning(png_ptr, "unknown chunk too large to fit in memory");
+           skip = length - (png_uint_32)65535L;
+           length = (png_uint_32)65535L;
+       }
+#endif
+       png_strncpy((png_charp)png_ptr->unknown_chunk.name,
+	 (png_charp)png_ptr->chunk_name,
+         png_sizeof((png_charp)png_ptr->chunk_name));
+       png_ptr->unknown_chunk.data = (png_bytep)png_malloc(png_ptr, length);
+       png_ptr->unknown_chunk.size = (png_size_t)length;
+       png_crc_read(png_ptr, (png_bytep)png_ptr->unknown_chunk.data, length);
+#if defined(PNG_READ_USER_CHUNKS_SUPPORTED)
+       if(png_ptr->read_user_chunk_fn != NULL)
+       {
+          /* callback to user unknown chunk handler */
+          int ret;
+          ret = (*(png_ptr->read_user_chunk_fn))
+            (png_ptr, &png_ptr->unknown_chunk);
+          if (ret < 0)
+             png_chunk_error(png_ptr, "error in user chunk");
+          if (ret == 0)
+          {
+             if (!(png_ptr->chunk_name[0] & 0x20))
+                if(png_handle_as_unknown(png_ptr, png_ptr->chunk_name) !=
+                     PNG_HANDLE_CHUNK_ALWAYS)
+                   png_chunk_error(png_ptr, "unknown critical chunk");
+                png_set_unknown_chunks(png_ptr, info_ptr,
+                   &png_ptr->unknown_chunk, 1);
+          }
+       }
+#else
+       png_set_unknown_chunks(png_ptr, info_ptr, &png_ptr->unknown_chunk, 1);
+#endif
+       png_free(png_ptr, png_ptr->unknown_chunk.data);
+       png_ptr->unknown_chunk.data = NULL;
+   }
+   else
+#endif
+      skip=length;
+   png_push_crc_skip(png_ptr, skip);
+}
+
+void /* PRIVATE */
+png_push_have_info(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr->info_fn != NULL)
+      (*(png_ptr->info_fn))(png_ptr, info_ptr);
+}
+
+void /* PRIVATE */
+png_push_have_end(png_structp png_ptr, png_infop info_ptr)
+{
+   if (png_ptr->end_fn != NULL)
+      (*(png_ptr->end_fn))(png_ptr, info_ptr);
+}
+
+void /* PRIVATE */
+png_push_have_row(png_structp png_ptr, png_bytep row)
+{
+   if (png_ptr->row_fn != NULL)
+      (*(png_ptr->row_fn))(png_ptr, row, png_ptr->row_number,
+         (int)png_ptr->pass);
+}
+
+void PNGAPI
+png_progressive_combine_row (png_structp png_ptr,
+   png_bytep old_row, png_bytep new_row)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_CONST int FARDATA png_pass_dsp_mask[7] =
+      {0xff, 0x0f, 0xff, 0x33, 0xff, 0x55, 0xff};
+#endif
+   if(png_ptr == NULL) return;
+   if (new_row != NULL)    /* new_row must == png_ptr->row_buf here. */
+      png_combine_row(png_ptr, old_row, png_pass_dsp_mask[png_ptr->pass]);
+}
+
+void PNGAPI
+png_set_progressive_read_fn(png_structp png_ptr, png_voidp progressive_ptr,
+   png_progressive_info_ptr info_fn, png_progressive_row_ptr row_fn,
+   png_progressive_end_ptr end_fn)
+{
+   if(png_ptr == NULL) return;
+   png_ptr->info_fn = info_fn;
+   png_ptr->row_fn = row_fn;
+   png_ptr->end_fn = end_fn;
+
+   png_set_read_fn(png_ptr, progressive_ptr, png_push_fill_buffer);
+}
+
+png_voidp PNGAPI
+png_get_progressive_ptr(png_structp png_ptr)
+{
+   if(png_ptr == NULL) return (NULL);
+   return png_ptr->io_ptr;
+}
+#endif /* PNG_PROGRESSIVE_READ_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngread.c b/distrib/libpng-1.2.19/pngread.c
new file mode 100644
index 0000000..2e561e8
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngread.c
@@ -0,0 +1,1478 @@
+
+/* pngread.c - read a PNG file
+ *
+ * Last changed in libpng 1.2.19 August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ *
+ * This file contains routines that an application calls directly to
+ * read a PNG file or stream.
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_READ_SUPPORTED)
+
+/* Create a PNG structure for reading, and allocate any memory needed. */
+png_structp PNGAPI
+png_create_read_struct(png_const_charp user_png_ver, png_voidp error_ptr,
+   png_error_ptr error_fn, png_error_ptr warn_fn)
+{
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   return (png_create_read_struct_2(user_png_ver, error_ptr, error_fn,
+      warn_fn, png_voidp_NULL, png_malloc_ptr_NULL, png_free_ptr_NULL));
+}
+
+/* Alternate create PNG structure for reading, and allocate any memory needed. */
+png_structp PNGAPI
+png_create_read_struct_2(png_const_charp user_png_ver, png_voidp error_ptr,
+   png_error_ptr error_fn, png_error_ptr warn_fn, png_voidp mem_ptr,
+   png_malloc_ptr malloc_fn, png_free_ptr free_fn)
+{
+#endif /* PNG_USER_MEM_SUPPORTED */
+
+   png_structp png_ptr;
+
+#ifdef PNG_SETJMP_SUPPORTED
+#ifdef USE_FAR_KEYWORD
+   jmp_buf jmpbuf;
+#endif
+#endif
+
+   int i;
+
+   png_debug(1, "in png_create_read_struct\n");
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_ptr = (png_structp)png_create_struct_2(PNG_STRUCT_PNG,
+      (png_malloc_ptr)malloc_fn, (png_voidp)mem_ptr);
+#else
+   png_ptr = (png_structp)png_create_struct(PNG_STRUCT_PNG);
+#endif
+   if (png_ptr == NULL)
+      return (NULL);
+
+#if !defined(PNG_1_0_X)
+#ifdef PNG_MMX_CODE_SUPPORTED
+   png_init_mmx_flags(png_ptr);   /* 1.2.0 addition */
+#endif
+#endif /* PNG_1_0_X */
+
+   /* added at libpng-1.2.6 */
+#ifdef PNG_SET_USER_LIMITS_SUPPORTED
+   png_ptr->user_width_max=PNG_USER_WIDTH_MAX;
+   png_ptr->user_height_max=PNG_USER_HEIGHT_MAX;
+#endif
+
+#ifdef PNG_SETJMP_SUPPORTED
+#ifdef USE_FAR_KEYWORD
+   if (setjmp(jmpbuf))
+#else
+   if (setjmp(png_ptr->jmpbuf))
+#endif
+   {
+      png_free(png_ptr, png_ptr->zbuf);
+      png_ptr->zbuf=NULL;
+#ifdef PNG_USER_MEM_SUPPORTED
+      png_destroy_struct_2((png_voidp)png_ptr,
+         (png_free_ptr)free_fn, (png_voidp)mem_ptr);
+#else
+      png_destroy_struct((png_voidp)png_ptr);
+#endif
+      return (NULL);
+   }
+#ifdef USE_FAR_KEYWORD
+   png_memcpy(png_ptr->jmpbuf,jmpbuf,png_sizeof(jmp_buf));
+#endif
+#endif
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_set_mem_fn(png_ptr, mem_ptr, malloc_fn, free_fn);
+#endif
+
+   png_set_error_fn(png_ptr, error_ptr, error_fn, warn_fn);
+
+   i=0;
+   do
+   {
+     if(user_png_ver[i] != png_libpng_ver[i])
+        png_ptr->flags |= PNG_FLAG_LIBRARY_MISMATCH;
+   } while (png_libpng_ver[i++]);
+
+   if (png_ptr->flags & PNG_FLAG_LIBRARY_MISMATCH)
+   {
+     /* Libpng 0.90 and later are binary incompatible with libpng 0.89, so
+      * we must recompile any applications that use any older library version.
+      * For versions after libpng 1.0, we will be compatible, so we need
+      * only check the first digit.
+      */
+     if (user_png_ver == NULL || user_png_ver[0] != png_libpng_ver[0] ||
+         (user_png_ver[0] == '1' && user_png_ver[2] != png_libpng_ver[2]) ||
+         (user_png_ver[0] == '0' && user_png_ver[2] < '9'))
+     {
+#if !defined(PNG_NO_STDIO) && !defined(_WIN32_WCE)
+        char msg[80];
+        if (user_png_ver)
+        {
+          png_snprintf(msg, 80,
+             "Application was compiled with png.h from libpng-%.20s",
+             user_png_ver);
+          png_warning(png_ptr, msg);
+        }
+        png_snprintf(msg, 80,
+             "Application  is  running with png.c from libpng-%.20s",
+           png_libpng_ver);
+        png_warning(png_ptr, msg);
+#endif
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+        png_ptr->flags=0;
+#endif
+        png_error(png_ptr,
+           "Incompatible libpng version in application and library");
+     }
+   }
+
+   /* initialize zbuf - compression buffer */
+   png_ptr->zbuf_size = PNG_ZBUF_SIZE;
+   png_ptr->zbuf = (png_bytep)png_malloc(png_ptr,
+     (png_uint_32)png_ptr->zbuf_size);
+   png_ptr->zstream.zalloc = png_zalloc;
+   png_ptr->zstream.zfree = png_zfree;
+   png_ptr->zstream.opaque = (voidpf)png_ptr;
+
+   switch (inflateInit(&png_ptr->zstream))
+   {
+     case Z_OK: /* Do nothing */ break;
+     case Z_MEM_ERROR:
+     case Z_STREAM_ERROR: png_error(png_ptr, "zlib memory error"); break;
+     case Z_VERSION_ERROR: png_error(png_ptr, "zlib version error"); break;
+     default: png_error(png_ptr, "Unknown zlib error");
+   }
+
+   png_ptr->zstream.next_out = png_ptr->zbuf;
+   png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+
+   png_set_read_fn(png_ptr, png_voidp_NULL, png_rw_ptr_NULL);
+
+#ifdef PNG_SETJMP_SUPPORTED
+/* Applications that neglect to set up their own setjmp() and then encounter
+   a png_error() will longjmp here.  Since the jmpbuf is then meaningless we
+   abort instead of returning. */
+#ifdef USE_FAR_KEYWORD
+   if (setjmp(jmpbuf))
+      PNG_ABORT();
+   png_memcpy(png_ptr->jmpbuf,jmpbuf,png_sizeof(jmp_buf));
+#else
+   if (setjmp(png_ptr->jmpbuf))
+      PNG_ABORT();
+#endif
+#endif
+   return (png_ptr);
+}
+
+#if defined(PNG_1_0_X) || defined(PNG_1_2_X)
+/* Initialize PNG structure for reading, and allocate any memory needed.
+   This interface is deprecated in favour of the png_create_read_struct(),
+   and it will disappear as of libpng-1.3.0. */
+#undef png_read_init
+void PNGAPI
+png_read_init(png_structp png_ptr)
+{
+   /* We only come here via pre-1.0.7-compiled applications */
+   png_read_init_2(png_ptr, "1.0.6 or earlier", 0, 0);
+}
+
+void PNGAPI
+png_read_init_2(png_structp png_ptr, png_const_charp user_png_ver,
+   png_size_t png_struct_size, png_size_t png_info_size)
+{
+   /* We only come here via pre-1.0.12-compiled applications */
+   if(png_ptr == NULL) return;
+#if !defined(PNG_NO_STDIO) && !defined(_WIN32_WCE)
+   if(png_sizeof(png_struct) > png_struct_size ||
+      png_sizeof(png_info) > png_info_size)
+   {
+      char msg[80];
+      png_ptr->warning_fn=NULL;
+      if (user_png_ver)
+      {
+        png_snprintf(msg, 80,
+           "Application was compiled with png.h from libpng-%.20s",
+           user_png_ver);
+        png_warning(png_ptr, msg);
+      }
+      png_snprintf(msg, 80,
+         "Application  is  running with png.c from libpng-%.20s",
+         png_libpng_ver);
+      png_warning(png_ptr, msg);
+   }
+#endif
+   if(png_sizeof(png_struct) > png_struct_size)
+     {
+       png_ptr->error_fn=NULL;
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+       png_ptr->flags=0;
+#endif
+       png_error(png_ptr,
+       "The png struct allocated by the application for reading is too small.");
+     }
+   if(png_sizeof(png_info) > png_info_size)
+     {
+       png_ptr->error_fn=NULL;
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+       png_ptr->flags=0;
+#endif
+       png_error(png_ptr,
+         "The info struct allocated by application for reading is too small.");
+     }
+   png_read_init_3(&png_ptr, user_png_ver, png_struct_size);
+}
+#endif /* PNG_1_0_X || PNG_1_2_X */
+
+void PNGAPI
+png_read_init_3(png_structpp ptr_ptr, png_const_charp user_png_ver,
+   png_size_t png_struct_size)
+{
+#ifdef PNG_SETJMP_SUPPORTED
+   jmp_buf tmp_jmp;  /* to save current jump buffer */
+#endif
+
+   int i=0;
+
+   png_structp png_ptr=*ptr_ptr;
+
+   if(png_ptr == NULL) return;
+
+   do
+   {
+     if(user_png_ver[i] != png_libpng_ver[i])
+     {
+#ifdef PNG_LEGACY_SUPPORTED
+       png_ptr->flags |= PNG_FLAG_LIBRARY_MISMATCH;
+#else
+       png_ptr->warning_fn=NULL;
+       png_warning(png_ptr,
+        "Application uses deprecated png_read_init() and should be recompiled.");
+       break;
+#endif
+     }
+   } while (png_libpng_ver[i++]);
+
+   png_debug(1, "in png_read_init_3\n");
+
+#ifdef PNG_SETJMP_SUPPORTED
+   /* save jump buffer and error functions */
+   png_memcpy(tmp_jmp, png_ptr->jmpbuf, png_sizeof (jmp_buf));
+#endif
+
+   if(png_sizeof(png_struct) > png_struct_size)
+     {
+       png_destroy_struct(png_ptr);
+       *ptr_ptr = (png_structp)png_create_struct(PNG_STRUCT_PNG);
+       png_ptr = *ptr_ptr;
+     }
+
+   /* reset all variables to 0 */
+   png_memset(png_ptr, 0, png_sizeof (png_struct));
+
+#ifdef PNG_SETJMP_SUPPORTED
+   /* restore jump buffer */
+   png_memcpy(png_ptr->jmpbuf, tmp_jmp, png_sizeof (jmp_buf));
+#endif
+
+   /* added at libpng-1.2.6 */
+#ifdef PNG_SET_USER_LIMITS_SUPPORTED
+   png_ptr->user_width_max=PNG_USER_WIDTH_MAX;
+   png_ptr->user_height_max=PNG_USER_HEIGHT_MAX;
+#endif
+
+   /* initialize zbuf - compression buffer */
+   png_ptr->zbuf_size = PNG_ZBUF_SIZE;
+   png_ptr->zbuf = (png_bytep)png_malloc(png_ptr,
+     (png_uint_32)png_ptr->zbuf_size);
+   png_ptr->zstream.zalloc = png_zalloc;
+   png_ptr->zstream.zfree = png_zfree;
+   png_ptr->zstream.opaque = (voidpf)png_ptr;
+
+   switch (inflateInit(&png_ptr->zstream))
+   {
+     case Z_OK: /* Do nothing */ break;
+     case Z_MEM_ERROR:
+     case Z_STREAM_ERROR: png_error(png_ptr, "zlib memory"); break;
+     case Z_VERSION_ERROR: png_error(png_ptr, "zlib version"); break;
+     default: png_error(png_ptr, "Unknown zlib error");
+   }
+
+   png_ptr->zstream.next_out = png_ptr->zbuf;
+   png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+
+   png_set_read_fn(png_ptr, png_voidp_NULL, png_rw_ptr_NULL);
+}
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* Read the information before the actual image data.  This has been
+ * changed in v0.90 to allow reading a file that already has the magic
+ * bytes read from the stream.  You can tell libpng how many bytes have
+ * been read from the beginning of the stream (up to the maximum of 8)
+ * via png_set_sig_bytes(), and we will only check the remaining bytes
+ * here.  The application can then have access to the signature bytes we
+ * read if it is determined that this isn't a valid PNG file.
+ */
+void PNGAPI
+png_read_info(png_structp png_ptr, png_infop info_ptr)
+{
+   if(png_ptr == NULL) return;
+   png_debug(1, "in png_read_info\n");
+   /* If we haven't checked all of the PNG signature bytes, do so now. */
+   if (png_ptr->sig_bytes < 8)
+   {
+      png_size_t num_checked = png_ptr->sig_bytes,
+                 num_to_check = 8 - num_checked;
+
+      png_read_data(png_ptr, &(info_ptr->signature[num_checked]), num_to_check);
+      png_ptr->sig_bytes = 8;
+
+      if (png_sig_cmp(info_ptr->signature, num_checked, num_to_check))
+      {
+         if (num_checked < 4 &&
+             png_sig_cmp(info_ptr->signature, num_checked, num_to_check - 4))
+            png_error(png_ptr, "Not a PNG file");
+         else
+            png_error(png_ptr, "PNG file corrupted by ASCII conversion");
+      }
+      if (num_checked < 3)
+         png_ptr->mode |= PNG_HAVE_PNG_SIGNATURE;
+   }
+
+   for(;;)
+   {
+#ifdef PNG_USE_LOCAL_ARRAYS
+      PNG_CONST PNG_IHDR;
+      PNG_CONST PNG_IDAT;
+      PNG_CONST PNG_IEND;
+      PNG_CONST PNG_PLTE;
+#if defined(PNG_READ_bKGD_SUPPORTED)
+      PNG_CONST PNG_bKGD;
+#endif
+#if defined(PNG_READ_cHRM_SUPPORTED)
+      PNG_CONST PNG_cHRM;
+#endif
+#if defined(PNG_READ_gAMA_SUPPORTED)
+      PNG_CONST PNG_gAMA;
+#endif
+#if defined(PNG_READ_hIST_SUPPORTED)
+      PNG_CONST PNG_hIST;
+#endif
+#if defined(PNG_READ_iCCP_SUPPORTED)
+      PNG_CONST PNG_iCCP;
+#endif
+#if defined(PNG_READ_iTXt_SUPPORTED)
+      PNG_CONST PNG_iTXt;
+#endif
+#if defined(PNG_READ_oFFs_SUPPORTED)
+      PNG_CONST PNG_oFFs;
+#endif
+#if defined(PNG_READ_pCAL_SUPPORTED)
+      PNG_CONST PNG_pCAL;
+#endif
+#if defined(PNG_READ_pHYs_SUPPORTED)
+      PNG_CONST PNG_pHYs;
+#endif
+#if defined(PNG_READ_sBIT_SUPPORTED)
+      PNG_CONST PNG_sBIT;
+#endif
+#if defined(PNG_READ_sCAL_SUPPORTED)
+      PNG_CONST PNG_sCAL;
+#endif
+#if defined(PNG_READ_sPLT_SUPPORTED)
+      PNG_CONST PNG_sPLT;
+#endif
+#if defined(PNG_READ_sRGB_SUPPORTED)
+      PNG_CONST PNG_sRGB;
+#endif
+#if defined(PNG_READ_tEXt_SUPPORTED)
+      PNG_CONST PNG_tEXt;
+#endif
+#if defined(PNG_READ_tIME_SUPPORTED)
+      PNG_CONST PNG_tIME;
+#endif
+#if defined(PNG_READ_tRNS_SUPPORTED)
+      PNG_CONST PNG_tRNS;
+#endif
+#if defined(PNG_READ_zTXt_SUPPORTED)
+      PNG_CONST PNG_zTXt;
+#endif
+#endif /* PNG_USE_LOCAL_ARRAYS */
+      png_byte chunk_length[4];
+      png_uint_32 length;
+
+      png_read_data(png_ptr, chunk_length, 4);
+      length = png_get_uint_31(png_ptr,chunk_length);
+
+      png_reset_crc(png_ptr);
+      png_crc_read(png_ptr, png_ptr->chunk_name, 4);
+
+      png_debug2(0, "Reading %s chunk, length=%lu.\n", png_ptr->chunk_name,
+         length);
+
+      /* This should be a binary subdivision search or a hash for
+       * matching the chunk name rather than a linear search.
+       */
+      if (!png_memcmp(png_ptr->chunk_name, (png_bytep)png_IDAT, 4))
+        if(png_ptr->mode & PNG_AFTER_IDAT)
+          png_ptr->mode |= PNG_HAVE_CHUNK_AFTER_IDAT;
+
+      if (!png_memcmp(png_ptr->chunk_name, png_IHDR, 4))
+         png_handle_IHDR(png_ptr, info_ptr, length);
+      else if (!png_memcmp(png_ptr->chunk_name, png_IEND, 4))
+         png_handle_IEND(png_ptr, info_ptr, length);
+#ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED
+      else if (png_handle_as_unknown(png_ptr, png_ptr->chunk_name))
+      {
+         if (!png_memcmp(png_ptr->chunk_name, png_IDAT, 4))
+            png_ptr->mode |= PNG_HAVE_IDAT;
+         png_handle_unknown(png_ptr, info_ptr, length);
+         if (!png_memcmp(png_ptr->chunk_name, png_PLTE, 4))
+            png_ptr->mode |= PNG_HAVE_PLTE;
+         else if (!png_memcmp(png_ptr->chunk_name, png_IDAT, 4))
+         {
+            if (!(png_ptr->mode & PNG_HAVE_IHDR))
+               png_error(png_ptr, "Missing IHDR before IDAT");
+            else if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE &&
+                     !(png_ptr->mode & PNG_HAVE_PLTE))
+               png_error(png_ptr, "Missing PLTE before IDAT");
+            break;
+         }
+      }
+#endif
+      else if (!png_memcmp(png_ptr->chunk_name, png_PLTE, 4))
+         png_handle_PLTE(png_ptr, info_ptr, length);
+      else if (!png_memcmp(png_ptr->chunk_name, png_IDAT, 4))
+      {
+         if (!(png_ptr->mode & PNG_HAVE_IHDR))
+            png_error(png_ptr, "Missing IHDR before IDAT");
+         else if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE &&
+                  !(png_ptr->mode & PNG_HAVE_PLTE))
+            png_error(png_ptr, "Missing PLTE before IDAT");
+
+         png_ptr->idat_size = length;
+         png_ptr->mode |= PNG_HAVE_IDAT;
+         break;
+      }
+#if defined(PNG_READ_bKGD_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_bKGD, 4))
+         png_handle_bKGD(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_cHRM_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_cHRM, 4))
+         png_handle_cHRM(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_gAMA_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_gAMA, 4))
+         png_handle_gAMA(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_hIST_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_hIST, 4))
+         png_handle_hIST(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_oFFs_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_oFFs, 4))
+         png_handle_oFFs(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_pCAL_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_pCAL, 4))
+         png_handle_pCAL(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_sCAL_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_sCAL, 4))
+         png_handle_sCAL(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_pHYs_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_pHYs, 4))
+         png_handle_pHYs(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_sBIT_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_sBIT, 4))
+         png_handle_sBIT(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_sRGB_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_sRGB, 4))
+         png_handle_sRGB(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_iCCP_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_iCCP, 4))
+         png_handle_iCCP(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_sPLT_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_sPLT, 4))
+         png_handle_sPLT(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_tEXt_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_tEXt, 4))
+         png_handle_tEXt(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_tIME_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_tIME, 4))
+         png_handle_tIME(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_tRNS_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_tRNS, 4))
+         png_handle_tRNS(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_zTXt_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_zTXt, 4))
+         png_handle_zTXt(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_iTXt_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_iTXt, 4))
+         png_handle_iTXt(png_ptr, info_ptr, length);
+#endif
+      else
+         png_handle_unknown(png_ptr, info_ptr, length);
+   }
+}
+#endif /* PNG_NO_SEQUENTIAL_READ_SUPPORTED */
+
+/* optional call to update the users info_ptr structure */
+void PNGAPI
+png_read_update_info(png_structp png_ptr, png_infop info_ptr)
+{
+   png_debug(1, "in png_read_update_info\n");
+   if(png_ptr == NULL) return;
+   if (!(png_ptr->flags & PNG_FLAG_ROW_INIT))
+      png_read_start_row(png_ptr);
+   else
+      png_warning(png_ptr,
+      "Ignoring extra png_read_update_info() call; row buffer not reallocated");
+   png_read_transform_info(png_ptr, info_ptr);
+}
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* Initialize palette, background, etc, after transformations
+ * are set, but before any reading takes place.  This allows
+ * the user to obtain a gamma-corrected palette, for example.
+ * If the user doesn't call this, we will do it ourselves.
+ */
+void PNGAPI
+png_start_read_image(png_structp png_ptr)
+{
+   png_debug(1, "in png_start_read_image\n");
+   if(png_ptr == NULL) return;
+   if (!(png_ptr->flags & PNG_FLAG_ROW_INIT))
+      png_read_start_row(png_ptr);
+}
+#endif /* PNG_NO_SEQUENTIAL_READ_SUPPORTED */
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+void PNGAPI
+png_read_row(png_structp png_ptr, png_bytep row, png_bytep dsp_row)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_CONST PNG_IDAT;
+   PNG_CONST int png_pass_dsp_mask[7] = {0xff, 0x0f, 0xff, 0x33, 0xff, 0x55,
+     0xff};
+   PNG_CONST int png_pass_mask[7] = {0x80, 0x08, 0x88, 0x22, 0xaa, 0x55, 0xff};
+#endif
+   int ret;
+   if(png_ptr == NULL) return;
+   png_debug2(1, "in png_read_row (row %lu, pass %d)\n",
+      png_ptr->row_number, png_ptr->pass);
+   if (!(png_ptr->flags & PNG_FLAG_ROW_INIT))
+      png_read_start_row(png_ptr);
+   if (png_ptr->row_number == 0 && png_ptr->pass == 0)
+   {
+   /* check for transforms that have been set but were defined out */
+#if defined(PNG_WRITE_INVERT_SUPPORTED) && !defined(PNG_READ_INVERT_SUPPORTED)
+   if (png_ptr->transformations & PNG_INVERT_MONO)
+      png_warning(png_ptr, "PNG_READ_INVERT_SUPPORTED is not defined.");
+#endif
+#if defined(PNG_WRITE_FILLER_SUPPORTED) && !defined(PNG_READ_FILLER_SUPPORTED)
+   if (png_ptr->transformations & PNG_FILLER)
+      png_warning(png_ptr, "PNG_READ_FILLER_SUPPORTED is not defined.");
+#endif
+#if defined(PNG_WRITE_PACKSWAP_SUPPORTED) && !defined(PNG_READ_PACKSWAP_SUPPORTED)
+   if (png_ptr->transformations & PNG_PACKSWAP)
+      png_warning(png_ptr, "PNG_READ_PACKSWAP_SUPPORTED is not defined.");
+#endif
+#if defined(PNG_WRITE_PACK_SUPPORTED) && !defined(PNG_READ_PACK_SUPPORTED)
+   if (png_ptr->transformations & PNG_PACK)
+      png_warning(png_ptr, "PNG_READ_PACK_SUPPORTED is not defined.");
+#endif
+#if defined(PNG_WRITE_SHIFT_SUPPORTED) && !defined(PNG_READ_SHIFT_SUPPORTED)
+   if (png_ptr->transformations & PNG_SHIFT)
+      png_warning(png_ptr, "PNG_READ_SHIFT_SUPPORTED is not defined.");
+#endif
+#if defined(PNG_WRITE_BGR_SUPPORTED) && !defined(PNG_READ_BGR_SUPPORTED)
+   if (png_ptr->transformations & PNG_BGR)
+      png_warning(png_ptr, "PNG_READ_BGR_SUPPORTED is not defined.");
+#endif
+#if defined(PNG_WRITE_SWAP_SUPPORTED) && !defined(PNG_READ_SWAP_SUPPORTED)
+   if (png_ptr->transformations & PNG_SWAP_BYTES)
+      png_warning(png_ptr, "PNG_READ_SWAP_SUPPORTED is not defined.");
+#endif
+   }
+
+#if defined(PNG_READ_INTERLACING_SUPPORTED)
+   /* if interlaced and we do not need a new row, combine row and return */
+   if (png_ptr->interlaced && (png_ptr->transformations & PNG_INTERLACE))
+   {
+      switch (png_ptr->pass)
+      {
+         case 0:
+            if (png_ptr->row_number & 0x07)
+            {
+               if (dsp_row != NULL)
+                  png_combine_row(png_ptr, dsp_row,
+                     png_pass_dsp_mask[png_ptr->pass]);
+               png_read_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 1:
+            if ((png_ptr->row_number & 0x07) || png_ptr->width < 5)
+            {
+               if (dsp_row != NULL)
+                  png_combine_row(png_ptr, dsp_row,
+                     png_pass_dsp_mask[png_ptr->pass]);
+               png_read_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 2:
+            if ((png_ptr->row_number & 0x07) != 4)
+            {
+               if (dsp_row != NULL && (png_ptr->row_number & 4))
+                  png_combine_row(png_ptr, dsp_row,
+                     png_pass_dsp_mask[png_ptr->pass]);
+               png_read_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 3:
+            if ((png_ptr->row_number & 3) || png_ptr->width < 3)
+            {
+               if (dsp_row != NULL)
+                  png_combine_row(png_ptr, dsp_row,
+                     png_pass_dsp_mask[png_ptr->pass]);
+               png_read_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 4:
+            if ((png_ptr->row_number & 3) != 2)
+            {
+               if (dsp_row != NULL && (png_ptr->row_number & 2))
+                  png_combine_row(png_ptr, dsp_row,
+                     png_pass_dsp_mask[png_ptr->pass]);
+               png_read_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 5:
+            if ((png_ptr->row_number & 1) || png_ptr->width < 2)
+            {
+               if (dsp_row != NULL)
+                  png_combine_row(png_ptr, dsp_row,
+                     png_pass_dsp_mask[png_ptr->pass]);
+               png_read_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 6:
+            if (!(png_ptr->row_number & 1))
+            {
+               png_read_finish_row(png_ptr);
+               return;
+            }
+            break;
+      }
+   }
+#endif
+
+   if (!(png_ptr->mode & PNG_HAVE_IDAT))
+      png_error(png_ptr, "Invalid attempt to read row data");
+
+   png_ptr->zstream.next_out = png_ptr->row_buf;
+   png_ptr->zstream.avail_out = (uInt)png_ptr->irowbytes;
+   do
+   {
+      if (!(png_ptr->zstream.avail_in))
+      {
+         while (!png_ptr->idat_size)
+         {
+            png_byte chunk_length[4];
+
+            png_crc_finish(png_ptr, 0);
+
+            png_read_data(png_ptr, chunk_length, 4);
+            png_ptr->idat_size = png_get_uint_31(png_ptr,chunk_length);
+
+            png_reset_crc(png_ptr);
+            png_crc_read(png_ptr, png_ptr->chunk_name, 4);
+            if (png_memcmp(png_ptr->chunk_name, png_IDAT, 4))
+               png_error(png_ptr, "Not enough image data");
+         }
+         png_ptr->zstream.avail_in = (uInt)png_ptr->zbuf_size;
+         png_ptr->zstream.next_in = png_ptr->zbuf;
+         if (png_ptr->zbuf_size > png_ptr->idat_size)
+            png_ptr->zstream.avail_in = (uInt)png_ptr->idat_size;
+         png_crc_read(png_ptr, png_ptr->zbuf,
+            (png_size_t)png_ptr->zstream.avail_in);
+         png_ptr->idat_size -= png_ptr->zstream.avail_in;
+      }
+      ret = inflate(&png_ptr->zstream, Z_PARTIAL_FLUSH);
+      if (ret == Z_STREAM_END)
+      {
+         if (png_ptr->zstream.avail_out || png_ptr->zstream.avail_in ||
+            png_ptr->idat_size)
+            png_error(png_ptr, "Extra compressed data");
+         png_ptr->mode |= PNG_AFTER_IDAT;
+         png_ptr->flags |= PNG_FLAG_ZLIB_FINISHED;
+         break;
+      }
+      if (ret != Z_OK)
+         png_error(png_ptr, png_ptr->zstream.msg ? png_ptr->zstream.msg :
+                   "Decompression error");
+
+   } while (png_ptr->zstream.avail_out);
+
+   png_ptr->row_info.color_type = png_ptr->color_type;
+   png_ptr->row_info.width = png_ptr->iwidth;
+   png_ptr->row_info.channels = png_ptr->channels;
+   png_ptr->row_info.bit_depth = png_ptr->bit_depth;
+   png_ptr->row_info.pixel_depth = png_ptr->pixel_depth;
+   png_ptr->row_info.rowbytes = PNG_ROWBYTES(png_ptr->row_info.pixel_depth,
+       png_ptr->row_info.width);
+
+   if(png_ptr->row_buf[0])
+   png_read_filter_row(png_ptr, &(png_ptr->row_info),
+      png_ptr->row_buf + 1, png_ptr->prev_row + 1,
+      (int)(png_ptr->row_buf[0]));
+
+   png_memcpy_check(png_ptr, png_ptr->prev_row, png_ptr->row_buf,
+      png_ptr->rowbytes + 1);
+
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+   if((png_ptr->mng_features_permitted & PNG_FLAG_MNG_FILTER_64) &&
+      (png_ptr->filter_type == PNG_INTRAPIXEL_DIFFERENCING))
+   {
+      /* Intrapixel differencing */
+      png_do_read_intrapixel(&(png_ptr->row_info), png_ptr->row_buf + 1);
+   }
+#endif
+
+
+   if (png_ptr->transformations || (png_ptr->flags&PNG_FLAG_STRIP_ALPHA))
+      png_do_read_transformations(png_ptr);
+
+#if defined(PNG_READ_INTERLACING_SUPPORTED)
+   /* blow up interlaced rows to full size */
+   if (png_ptr->interlaced &&
+      (png_ptr->transformations & PNG_INTERLACE))
+   {
+      if (png_ptr->pass < 6)
+/*       old interface (pre-1.0.9):
+         png_do_read_interlace(&(png_ptr->row_info),
+            png_ptr->row_buf + 1, png_ptr->pass, png_ptr->transformations);
+ */
+         png_do_read_interlace(png_ptr);
+
+      if (dsp_row != NULL)
+         png_combine_row(png_ptr, dsp_row,
+            png_pass_dsp_mask[png_ptr->pass]);
+      if (row != NULL)
+         png_combine_row(png_ptr, row,
+            png_pass_mask[png_ptr->pass]);
+   }
+   else
+#endif
+   {
+      if (row != NULL)
+         png_combine_row(png_ptr, row, 0xff);
+      if (dsp_row != NULL)
+         png_combine_row(png_ptr, dsp_row, 0xff);
+   }
+   png_read_finish_row(png_ptr);
+
+   if (png_ptr->read_row_fn != NULL)
+      (*(png_ptr->read_row_fn))(png_ptr, png_ptr->row_number, png_ptr->pass);
+}
+#endif /* PNG_NO_SEQUENTIAL_READ_SUPPORTED */
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* Read one or more rows of image data.  If the image is interlaced,
+ * and png_set_interlace_handling() has been called, the rows need to
+ * contain the contents of the rows from the previous pass.  If the
+ * image has alpha or transparency, and png_handle_alpha()[*] has been
+ * called, the rows contents must be initialized to the contents of the
+ * screen.
+ *
+ * "row" holds the actual image, and pixels are placed in it
+ * as they arrive.  If the image is displayed after each pass, it will
+ * appear to "sparkle" in.  "display_row" can be used to display a
+ * "chunky" progressive image, with finer detail added as it becomes
+ * available.  If you do not want this "chunky" display, you may pass
+ * NULL for display_row.  If you do not want the sparkle display, and
+ * you have not called png_handle_alpha(), you may pass NULL for rows.
+ * If you have called png_handle_alpha(), and the image has either an
+ * alpha channel or a transparency chunk, you must provide a buffer for
+ * rows.  In this case, you do not have to provide a display_row buffer
+ * also, but you may.  If the image is not interlaced, or if you have
+ * not called png_set_interlace_handling(), the display_row buffer will
+ * be ignored, so pass NULL to it.
+ *
+ * [*] png_handle_alpha() does not exist yet, as of this version of libpng
+ */
+
+void PNGAPI
+png_read_rows(png_structp png_ptr, png_bytepp row,
+   png_bytepp display_row, png_uint_32 num_rows)
+{
+   png_uint_32 i;
+   png_bytepp rp;
+   png_bytepp dp;
+
+   png_debug(1, "in png_read_rows\n");
+   if(png_ptr == NULL) return;
+   rp = row;
+   dp = display_row;
+   if (rp != NULL && dp != NULL)
+      for (i = 0; i < num_rows; i++)
+      {
+         png_bytep rptr = *rp++;
+         png_bytep dptr = *dp++;
+
+         png_read_row(png_ptr, rptr, dptr);
+      }
+   else if(rp != NULL)
+      for (i = 0; i < num_rows; i++)
+      {
+         png_bytep rptr = *rp;
+         png_read_row(png_ptr, rptr, png_bytep_NULL);
+         rp++;
+      }
+   else if(dp != NULL)
+      for (i = 0; i < num_rows; i++)
+      {
+         png_bytep dptr = *dp;
+         png_read_row(png_ptr, png_bytep_NULL, dptr);
+         dp++;
+      }
+}
+#endif /* PNG_NO_SEQUENTIAL_READ_SUPPORTED */
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* Read the entire image.  If the image has an alpha channel or a tRNS
+ * chunk, and you have called png_handle_alpha()[*], you will need to
+ * initialize the image to the current image that PNG will be overlaying.
+ * We set the num_rows again here, in case it was incorrectly set in
+ * png_read_start_row() by a call to png_read_update_info() or
+ * png_start_read_image() if png_set_interlace_handling() wasn't called
+ * prior to either of these functions like it should have been.  You can
+ * only call this function once.  If you desire to have an image for
+ * each pass of a interlaced image, use png_read_rows() instead.
+ *
+ * [*] png_handle_alpha() does not exist yet, as of this version of libpng
+ */
+void PNGAPI
+png_read_image(png_structp png_ptr, png_bytepp image)
+{
+   png_uint_32 i,image_height;
+   int pass, j;
+   png_bytepp rp;
+
+   png_debug(1, "in png_read_image\n");
+   if(png_ptr == NULL) return;
+
+#ifdef PNG_READ_INTERLACING_SUPPORTED
+   pass = png_set_interlace_handling(png_ptr);
+#else
+   if (png_ptr->interlaced)
+      png_error(png_ptr,
+        "Cannot read interlaced image -- interlace handler disabled.");
+   pass = 1;
+#endif
+
+
+   image_height=png_ptr->height;
+   png_ptr->num_rows = image_height; /* Make sure this is set correctly */
+
+   for (j = 0; j < pass; j++)
+   {
+      rp = image;
+      for (i = 0; i < image_height; i++)
+      {
+         png_read_row(png_ptr, *rp, png_bytep_NULL);
+         rp++;
+      }
+   }
+}
+#endif /* PNG_NO_SEQUENTIAL_READ_SUPPORTED */
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+/* Read the end of the PNG file.  Will not read past the end of the
+ * file, will verify the end is accurate, and will read any comments
+ * or time information at the end of the file, if info is not NULL.
+ */
+void PNGAPI
+png_read_end(png_structp png_ptr, png_infop info_ptr)
+{
+   png_byte chunk_length[4];
+   png_uint_32 length;
+
+   png_debug(1, "in png_read_end\n");
+   if(png_ptr == NULL) return;
+   png_crc_finish(png_ptr, 0); /* Finish off CRC from last IDAT chunk */
+
+   do
+   {
+#ifdef PNG_USE_LOCAL_ARRAYS
+      PNG_CONST PNG_IHDR;
+      PNG_CONST PNG_IDAT;
+      PNG_CONST PNG_IEND;
+      PNG_CONST PNG_PLTE;
+#if defined(PNG_READ_bKGD_SUPPORTED)
+      PNG_CONST PNG_bKGD;
+#endif
+#if defined(PNG_READ_cHRM_SUPPORTED)
+      PNG_CONST PNG_cHRM;
+#endif
+#if defined(PNG_READ_gAMA_SUPPORTED)
+      PNG_CONST PNG_gAMA;
+#endif
+#if defined(PNG_READ_hIST_SUPPORTED)
+      PNG_CONST PNG_hIST;
+#endif
+#if defined(PNG_READ_iCCP_SUPPORTED)
+      PNG_CONST PNG_iCCP;
+#endif
+#if defined(PNG_READ_iTXt_SUPPORTED)
+      PNG_CONST PNG_iTXt;
+#endif
+#if defined(PNG_READ_oFFs_SUPPORTED)
+      PNG_CONST PNG_oFFs;
+#endif
+#if defined(PNG_READ_pCAL_SUPPORTED)
+      PNG_CONST PNG_pCAL;
+#endif
+#if defined(PNG_READ_pHYs_SUPPORTED)
+      PNG_CONST PNG_pHYs;
+#endif
+#if defined(PNG_READ_sBIT_SUPPORTED)
+      PNG_CONST PNG_sBIT;
+#endif
+#if defined(PNG_READ_sCAL_SUPPORTED)
+      PNG_CONST PNG_sCAL;
+#endif
+#if defined(PNG_READ_sPLT_SUPPORTED)
+      PNG_CONST PNG_sPLT;
+#endif
+#if defined(PNG_READ_sRGB_SUPPORTED)
+      PNG_CONST PNG_sRGB;
+#endif
+#if defined(PNG_READ_tEXt_SUPPORTED)
+      PNG_CONST PNG_tEXt;
+#endif
+#if defined(PNG_READ_tIME_SUPPORTED)
+      PNG_CONST PNG_tIME;
+#endif
+#if defined(PNG_READ_tRNS_SUPPORTED)
+      PNG_CONST PNG_tRNS;
+#endif
+#if defined(PNG_READ_zTXt_SUPPORTED)
+      PNG_CONST PNG_zTXt;
+#endif
+#endif /* PNG_USE_LOCAL_ARRAYS */
+
+      png_read_data(png_ptr, chunk_length, 4);
+      length = png_get_uint_31(png_ptr,chunk_length);
+
+      png_reset_crc(png_ptr);
+      png_crc_read(png_ptr, png_ptr->chunk_name, 4);
+
+      png_debug1(0, "Reading %s chunk.\n", png_ptr->chunk_name);
+
+      if (!png_memcmp(png_ptr->chunk_name, png_IHDR, 4))
+         png_handle_IHDR(png_ptr, info_ptr, length);
+      else if (!png_memcmp(png_ptr->chunk_name, png_IEND, 4))
+         png_handle_IEND(png_ptr, info_ptr, length);
+#ifdef PNG_HANDLE_AS_UNKNOWN_SUPPORTED
+      else if (png_handle_as_unknown(png_ptr, png_ptr->chunk_name))
+      {
+         if (!png_memcmp(png_ptr->chunk_name, png_IDAT, 4))
+         {
+            if ((length > 0) || (png_ptr->mode & PNG_HAVE_CHUNK_AFTER_IDAT))
+               png_error(png_ptr, "Too many IDAT's found");
+         }
+         png_handle_unknown(png_ptr, info_ptr, length);
+         if (!png_memcmp(png_ptr->chunk_name, png_PLTE, 4))
+            png_ptr->mode |= PNG_HAVE_PLTE;
+      }
+#endif
+      else if (!png_memcmp(png_ptr->chunk_name, png_IDAT, 4))
+      {
+         /* Zero length IDATs are legal after the last IDAT has been
+          * read, but not after other chunks have been read.
+          */
+         if ((length > 0) || (png_ptr->mode & PNG_HAVE_CHUNK_AFTER_IDAT))
+            png_error(png_ptr, "Too many IDAT's found");
+         png_crc_finish(png_ptr, length);
+      }
+      else if (!png_memcmp(png_ptr->chunk_name, png_PLTE, 4))
+         png_handle_PLTE(png_ptr, info_ptr, length);
+#if defined(PNG_READ_bKGD_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_bKGD, 4))
+         png_handle_bKGD(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_cHRM_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_cHRM, 4))
+         png_handle_cHRM(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_gAMA_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_gAMA, 4))
+         png_handle_gAMA(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_hIST_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_hIST, 4))
+         png_handle_hIST(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_oFFs_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_oFFs, 4))
+         png_handle_oFFs(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_pCAL_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_pCAL, 4))
+         png_handle_pCAL(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_sCAL_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_sCAL, 4))
+         png_handle_sCAL(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_pHYs_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_pHYs, 4))
+         png_handle_pHYs(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_sBIT_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_sBIT, 4))
+         png_handle_sBIT(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_sRGB_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_sRGB, 4))
+         png_handle_sRGB(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_iCCP_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_iCCP, 4))
+         png_handle_iCCP(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_sPLT_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_sPLT, 4))
+         png_handle_sPLT(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_tEXt_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_tEXt, 4))
+         png_handle_tEXt(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_tIME_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_tIME, 4))
+         png_handle_tIME(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_tRNS_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_tRNS, 4))
+         png_handle_tRNS(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_zTXt_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_zTXt, 4))
+         png_handle_zTXt(png_ptr, info_ptr, length);
+#endif
+#if defined(PNG_READ_iTXt_SUPPORTED)
+      else if (!png_memcmp(png_ptr->chunk_name, png_iTXt, 4))
+         png_handle_iTXt(png_ptr, info_ptr, length);
+#endif
+      else
+         png_handle_unknown(png_ptr, info_ptr, length);
+   } while (!(png_ptr->mode & PNG_HAVE_IEND));
+}
+#endif /* PNG_NO_SEQUENTIAL_READ_SUPPORTED */
+
+/* free all memory used by the read */
+void PNGAPI
+png_destroy_read_struct(png_structpp png_ptr_ptr, png_infopp info_ptr_ptr,
+   png_infopp end_info_ptr_ptr)
+{
+   png_structp png_ptr = NULL;
+   png_infop info_ptr = NULL, end_info_ptr = NULL;
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_free_ptr free_fn;
+   png_voidp mem_ptr;
+#endif
+
+   png_debug(1, "in png_destroy_read_struct\n");
+   if (png_ptr_ptr != NULL)
+      png_ptr = *png_ptr_ptr;
+
+   if (info_ptr_ptr != NULL)
+      info_ptr = *info_ptr_ptr;
+
+   if (end_info_ptr_ptr != NULL)
+      end_info_ptr = *end_info_ptr_ptr;
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   free_fn = png_ptr->free_fn;
+   mem_ptr = png_ptr->mem_ptr;
+#endif
+
+   png_read_destroy(png_ptr, info_ptr, end_info_ptr);
+
+   if (info_ptr != NULL)
+   {
+#if defined(PNG_TEXT_SUPPORTED)
+      png_free_data(png_ptr, info_ptr, PNG_FREE_TEXT, -1);
+#endif
+
+#ifdef PNG_USER_MEM_SUPPORTED
+      png_destroy_struct_2((png_voidp)info_ptr, (png_free_ptr)free_fn,
+          (png_voidp)mem_ptr);
+#else
+      png_destroy_struct((png_voidp)info_ptr);
+#endif
+      *info_ptr_ptr = NULL;
+   }
+
+   if (end_info_ptr != NULL)
+   {
+#if defined(PNG_READ_TEXT_SUPPORTED)
+      png_free_data(png_ptr, end_info_ptr, PNG_FREE_TEXT, -1);
+#endif
+#ifdef PNG_USER_MEM_SUPPORTED
+      png_destroy_struct_2((png_voidp)end_info_ptr, (png_free_ptr)free_fn,
+         (png_voidp)mem_ptr);
+#else
+      png_destroy_struct((png_voidp)end_info_ptr);
+#endif
+      *end_info_ptr_ptr = NULL;
+   }
+
+   if (png_ptr != NULL)
+   {
+#ifdef PNG_USER_MEM_SUPPORTED
+      png_destroy_struct_2((png_voidp)png_ptr, (png_free_ptr)free_fn,
+          (png_voidp)mem_ptr);
+#else
+      png_destroy_struct((png_voidp)png_ptr);
+#endif
+      *png_ptr_ptr = NULL;
+   }
+}
+
+/* free all memory used by the read (old method) */
+void /* PRIVATE */
+png_read_destroy(png_structp png_ptr, png_infop info_ptr, png_infop end_info_ptr)
+{
+#ifdef PNG_SETJMP_SUPPORTED
+   jmp_buf tmp_jmp;
+#endif
+   png_error_ptr error_fn;
+   png_error_ptr warning_fn;
+   png_voidp error_ptr;
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_free_ptr free_fn;
+#endif
+
+   png_debug(1, "in png_read_destroy\n");
+   if (info_ptr != NULL)
+      png_info_destroy(png_ptr, info_ptr);
+
+   if (end_info_ptr != NULL)
+      png_info_destroy(png_ptr, end_info_ptr);
+
+   png_free(png_ptr, png_ptr->zbuf);
+   png_free(png_ptr, png_ptr->big_row_buf);
+   png_free(png_ptr, png_ptr->prev_row);
+#if defined(PNG_READ_DITHER_SUPPORTED)
+   png_free(png_ptr, png_ptr->palette_lookup);
+   png_free(png_ptr, png_ptr->dither_index);
+#endif
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+   png_free(png_ptr, png_ptr->gamma_table);
+#endif
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+   png_free(png_ptr, png_ptr->gamma_from_1);
+   png_free(png_ptr, png_ptr->gamma_to_1);
+#endif
+#ifdef PNG_FREE_ME_SUPPORTED
+   if (png_ptr->free_me & PNG_FREE_PLTE)
+      png_zfree(png_ptr, png_ptr->palette);
+   png_ptr->free_me &= ~PNG_FREE_PLTE;
+#else
+   if (png_ptr->flags & PNG_FLAG_FREE_PLTE)
+      png_zfree(png_ptr, png_ptr->palette);
+   png_ptr->flags &= ~PNG_FLAG_FREE_PLTE;
+#endif
+#if defined(PNG_tRNS_SUPPORTED) || \
+    defined(PNG_READ_EXPAND_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+#ifdef PNG_FREE_ME_SUPPORTED
+   if (png_ptr->free_me & PNG_FREE_TRNS)
+      png_free(png_ptr, png_ptr->trans);
+   png_ptr->free_me &= ~PNG_FREE_TRNS;
+#else
+   if (png_ptr->flags & PNG_FLAG_FREE_TRNS)
+      png_free(png_ptr, png_ptr->trans);
+   png_ptr->flags &= ~PNG_FLAG_FREE_TRNS;
+#endif
+#endif
+#if defined(PNG_READ_hIST_SUPPORTED)
+#ifdef PNG_FREE_ME_SUPPORTED
+   if (png_ptr->free_me & PNG_FREE_HIST)
+      png_free(png_ptr, png_ptr->hist);
+   png_ptr->free_me &= ~PNG_FREE_HIST;
+#else
+   if (png_ptr->flags & PNG_FLAG_FREE_HIST)
+      png_free(png_ptr, png_ptr->hist);
+   png_ptr->flags &= ~PNG_FLAG_FREE_HIST;
+#endif
+#endif
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+   if (png_ptr->gamma_16_table != NULL)
+   {
+      int i;
+      int istop = (1 << (8 - png_ptr->gamma_shift));
+      for (i = 0; i < istop; i++)
+      {
+         png_free(png_ptr, png_ptr->gamma_16_table[i]);
+      }
+   png_free(png_ptr, png_ptr->gamma_16_table);
+   }
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+   if (png_ptr->gamma_16_from_1 != NULL)
+   {
+      int i;
+      int istop = (1 << (8 - png_ptr->gamma_shift));
+      for (i = 0; i < istop; i++)
+      {
+         png_free(png_ptr, png_ptr->gamma_16_from_1[i]);
+      }
+   png_free(png_ptr, png_ptr->gamma_16_from_1);
+   }
+   if (png_ptr->gamma_16_to_1 != NULL)
+   {
+      int i;
+      int istop = (1 << (8 - png_ptr->gamma_shift));
+      for (i = 0; i < istop; i++)
+      {
+         png_free(png_ptr, png_ptr->gamma_16_to_1[i]);
+      }
+   png_free(png_ptr, png_ptr->gamma_16_to_1);
+   }
+#endif
+#endif
+#if defined(PNG_TIME_RFC1123_SUPPORTED)
+   png_free(png_ptr, png_ptr->time_buffer);
+#endif
+
+   inflateEnd(&png_ptr->zstream);
+#ifdef PNG_PROGRESSIVE_READ_SUPPORTED
+   png_free(png_ptr, png_ptr->save_buffer);
+#endif
+
+#ifdef PNG_PROGRESSIVE_READ_SUPPORTED
+#ifdef PNG_TEXT_SUPPORTED
+   png_free(png_ptr, png_ptr->current_text);
+#endif /* PNG_TEXT_SUPPORTED */
+#endif /* PNG_PROGRESSIVE_READ_SUPPORTED */
+
+   /* Save the important info out of the png_struct, in case it is
+    * being used again.
+    */
+#ifdef PNG_SETJMP_SUPPORTED
+   png_memcpy(tmp_jmp, png_ptr->jmpbuf, png_sizeof (jmp_buf));
+#endif
+
+   error_fn = png_ptr->error_fn;
+   warning_fn = png_ptr->warning_fn;
+   error_ptr = png_ptr->error_ptr;
+#ifdef PNG_USER_MEM_SUPPORTED
+   free_fn = png_ptr->free_fn;
+#endif
+
+   png_memset(png_ptr, 0, png_sizeof (png_struct));
+
+   png_ptr->error_fn = error_fn;
+   png_ptr->warning_fn = warning_fn;
+   png_ptr->error_ptr = error_ptr;
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_ptr->free_fn = free_fn;
+#endif
+
+#ifdef PNG_SETJMP_SUPPORTED
+   png_memcpy(png_ptr->jmpbuf, tmp_jmp, png_sizeof (jmp_buf));
+#endif
+
+}
+
+void PNGAPI
+png_set_read_status_fn(png_structp png_ptr, png_read_status_ptr read_row_fn)
+{
+   if(png_ptr == NULL) return;
+   png_ptr->read_row_fn = read_row_fn;
+}
+
+
+#ifndef PNG_NO_SEQUENTIAL_READ_SUPPORTED
+#if defined(PNG_INFO_IMAGE_SUPPORTED)
+void PNGAPI
+png_read_png(png_structp png_ptr, png_infop info_ptr,
+                           int transforms,
+                           voidp params)
+{
+   int row;
+
+   if(png_ptr == NULL) return;
+#if defined(PNG_READ_INVERT_ALPHA_SUPPORTED)
+   /* invert the alpha channel from opacity to transparency
+    */
+   if (transforms & PNG_TRANSFORM_INVERT_ALPHA)
+       png_set_invert_alpha(png_ptr);
+#endif
+
+   /* png_read_info() gives us all of the information from the
+    * PNG file before the first IDAT (image data chunk).
+    */
+   png_read_info(png_ptr, info_ptr);
+   if (info_ptr->height > PNG_UINT_32_MAX/png_sizeof(png_bytep))
+      png_error(png_ptr,"Image is too high to process with png_read_png()");
+
+   /* -------------- image transformations start here ------------------- */
+
+#if defined(PNG_READ_16_TO_8_SUPPORTED)
+   /* tell libpng to strip 16 bit/color files down to 8 bits per color
+    */
+   if (transforms & PNG_TRANSFORM_STRIP_16)
+       png_set_strip_16(png_ptr);
+#endif
+
+#if defined(PNG_READ_STRIP_ALPHA_SUPPORTED)
+   /* Strip alpha bytes from the input data without combining with
+    * the background (not recommended).
+    */
+   if (transforms & PNG_TRANSFORM_STRIP_ALPHA)
+       png_set_strip_alpha(png_ptr);
+#endif
+
+#if defined(PNG_READ_PACK_SUPPORTED) && !defined(PNG_READ_EXPAND_SUPPORTED)
+   /* Extract multiple pixels with bit depths of 1, 2, or 4 from a single
+    * byte into separate bytes (useful for paletted and grayscale images).
+    */
+   if (transforms & PNG_TRANSFORM_PACKING)
+       png_set_packing(png_ptr);
+#endif
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+   /* Change the order of packed pixels to least significant bit first
+    * (not useful if you are using png_set_packing).
+    */
+   if (transforms & PNG_TRANSFORM_PACKSWAP)
+       png_set_packswap(png_ptr);
+#endif
+
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+   /* Expand paletted colors into true RGB triplets
+    * Expand grayscale images to full 8 bits from 1, 2, or 4 bits/pixel
+    * Expand paletted or RGB images with transparency to full alpha
+    * channels so the data will be available as RGBA quartets.
+    */
+   if (transforms & PNG_TRANSFORM_EXPAND)
+       if ((png_ptr->bit_depth < 8) ||
+           (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE) ||
+           (png_get_valid(png_ptr, info_ptr, PNG_INFO_tRNS)))
+         png_set_expand(png_ptr);
+#endif
+
+   /* We don't handle background color or gamma transformation or dithering.
+    */
+
+#if defined(PNG_READ_INVERT_SUPPORTED)
+   /* invert monochrome files to have 0 as white and 1 as black
+    */
+   if (transforms & PNG_TRANSFORM_INVERT_MONO)
+       png_set_invert_mono(png_ptr);
+#endif
+
+#if defined(PNG_READ_SHIFT_SUPPORTED)
+   /* If you want to shift the pixel values from the range [0,255] or
+    * [0,65535] to the original [0,7] or [0,31], or whatever range the
+    * colors were originally in:
+    */
+   if ((transforms & PNG_TRANSFORM_SHIFT)
+       && png_get_valid(png_ptr, info_ptr, PNG_INFO_sBIT))
+   {
+      png_color_8p sig_bit;
+
+      png_get_sBIT(png_ptr, info_ptr, &sig_bit);
+      png_set_shift(png_ptr, sig_bit);
+   }
+#endif
+
+#if defined(PNG_READ_BGR_SUPPORTED)
+   /* flip the RGB pixels to BGR (or RGBA to BGRA)
+    */
+   if (transforms & PNG_TRANSFORM_BGR)
+       png_set_bgr(png_ptr);
+#endif
+
+#if defined(PNG_READ_SWAP_ALPHA_SUPPORTED)
+   /* swap the RGBA or GA data to ARGB or AG (or BGRA to ABGR)
+    */
+   if (transforms & PNG_TRANSFORM_SWAP_ALPHA)
+       png_set_swap_alpha(png_ptr);
+#endif
+
+#if defined(PNG_READ_SWAP_SUPPORTED)
+   /* swap bytes of 16 bit files to least significant byte first
+    */
+   if (transforms & PNG_TRANSFORM_SWAP_ENDIAN)
+       png_set_swap(png_ptr);
+#endif
+
+   /* We don't handle adding filler bytes */
+
+   /* Optional call to gamma correct and add the background to the palette
+    * and update info structure.  REQUIRED if you are expecting libpng to
+    * update the palette for you (i.e., you selected such a transform above).
+    */
+   png_read_update_info(png_ptr, info_ptr);
+
+   /* -------------- image transformations end here ------------------- */
+
+#ifdef PNG_FREE_ME_SUPPORTED
+   png_free_data(png_ptr, info_ptr, PNG_FREE_ROWS, 0);
+#endif
+   if(info_ptr->row_pointers == NULL)
+   {
+      info_ptr->row_pointers = (png_bytepp)png_malloc(png_ptr,
+         info_ptr->height * png_sizeof(png_bytep));
+#ifdef PNG_FREE_ME_SUPPORTED
+      info_ptr->free_me |= PNG_FREE_ROWS;
+#endif
+      for (row = 0; row < (int)info_ptr->height; row++)
+      {
+         info_ptr->row_pointers[row] = (png_bytep)png_malloc(png_ptr,
+            png_get_rowbytes(png_ptr, info_ptr));
+      }
+   }
+
+   png_read_image(png_ptr, info_ptr->row_pointers);
+   info_ptr->valid |= PNG_INFO_IDAT;
+
+   /* read rest of file, and get additional chunks in info_ptr - REQUIRED */
+   png_read_end(png_ptr, info_ptr);
+
+   transforms = transforms; /* quiet compiler warnings */
+   params = params;
+
+}
+#endif /* PNG_INFO_IMAGE_SUPPORTED */
+#endif /* PNG_NO_SEQUENTIAL_READ_SUPPORTED */
+#endif /* PNG_READ_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngrio.c b/distrib/libpng-1.2.19/pngrio.c
new file mode 100644
index 0000000..7d2522f
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngrio.c
@@ -0,0 +1,167 @@
+
+/* pngrio.c - functions for data input
+ *
+ * Last changed in libpng 1.2.13 November 13, 2006
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2006 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ *
+ * This file provides a location for all input.  Users who need
+ * special handling are expected to write a function that has the same
+ * arguments as this and performs a similar function, but that possibly
+ * has a different input method.  Note that you shouldn't change this
+ * function, but rather write a replacement function and then make
+ * libpng use it at run time with png_set_read_fn(...).
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_READ_SUPPORTED)
+
+/* Read the data from whatever input you are using.  The default routine
+   reads from a file pointer.  Note that this routine sometimes gets called
+   with very small lengths, so you should implement some kind of simple
+   buffering if you are using unbuffered reads.  This should never be asked
+   to read more then 64K on a 16 bit machine. */
+void /* PRIVATE */
+png_read_data(png_structp png_ptr, png_bytep data, png_size_t length)
+{
+   png_debug1(4,"reading %d bytes\n", (int)length);
+   if (png_ptr->read_data_fn != NULL)
+      (*(png_ptr->read_data_fn))(png_ptr, data, length);
+   else
+      png_error(png_ptr, "Call to NULL read function");
+}
+
+#if !defined(PNG_NO_STDIO)
+/* This is the function that does the actual reading of data.  If you are
+   not reading from a standard C stream, you should create a replacement
+   read_data function and use it at run time with png_set_read_fn(), rather
+   than changing the library. */
+#ifndef USE_FAR_KEYWORD
+void PNGAPI
+png_default_read_data(png_structp png_ptr, png_bytep data, png_size_t length)
+{
+   png_size_t check;
+
+   if(png_ptr == NULL) return;
+   /* fread() returns 0 on error, so it is OK to store this in a png_size_t
+    * instead of an int, which is what fread() actually returns.
+    */
+#if defined(_WIN32_WCE)
+   if ( !ReadFile((HANDLE)(png_ptr->io_ptr), data, length, &check, NULL) )
+      check = 0;
+#else
+   check = (png_size_t)fread(data, (png_size_t)1, length,
+      (png_FILE_p)png_ptr->io_ptr);
+#endif
+
+   if (check != length)
+      png_error(png_ptr, "Read Error");
+}
+#else
+/* this is the model-independent version. Since the standard I/O library
+   can't handle far buffers in the medium and small models, we have to copy
+   the data.
+*/
+
+#define NEAR_BUF_SIZE 1024
+#define MIN(a,b) (a <= b ? a : b)
+
+static void PNGAPI
+png_default_read_data(png_structp png_ptr, png_bytep data, png_size_t length)
+{
+   int check;
+   png_byte *n_data;
+   png_FILE_p io_ptr;
+
+   if(png_ptr == NULL) return;
+   /* Check if data really is near. If so, use usual code. */
+   n_data = (png_byte *)CVT_PTR_NOCHECK(data);
+   io_ptr = (png_FILE_p)CVT_PTR(png_ptr->io_ptr);
+   if ((png_bytep)n_data == data)
+   {
+#if defined(_WIN32_WCE)
+      if ( !ReadFile((HANDLE)(png_ptr->io_ptr), data, length, &check, NULL) )
+         check = 0;
+#else
+      check = fread(n_data, 1, length, io_ptr);
+#endif
+   }
+   else
+   {
+      png_byte buf[NEAR_BUF_SIZE];
+      png_size_t read, remaining, err;
+      check = 0;
+      remaining = length;
+      do
+      {
+         read = MIN(NEAR_BUF_SIZE, remaining);
+#if defined(_WIN32_WCE)
+         if ( !ReadFile((HANDLE)(io_ptr), buf, read, &err, NULL) )
+            err = 0;
+#else
+         err = fread(buf, (png_size_t)1, read, io_ptr);
+#endif
+         png_memcpy(data, buf, read); /* copy far buffer to near buffer */
+         if(err != read)
+            break;
+         else
+            check += err;
+         data += read;
+         remaining -= read;
+      }
+      while (remaining != 0);
+   }
+   if ((png_uint_32)check != (png_uint_32)length)
+      png_error(png_ptr, "read Error");
+}
+#endif
+#endif
+
+/* This function allows the application to supply a new input function
+   for libpng if standard C streams aren't being used.
+
+   This function takes as its arguments:
+   png_ptr      - pointer to a png input data structure
+   io_ptr       - pointer to user supplied structure containing info about
+                  the input functions.  May be NULL.
+   read_data_fn - pointer to a new input function that takes as its
+                  arguments a pointer to a png_struct, a pointer to
+                  a location where input data can be stored, and a 32-bit
+                  unsigned int that is the number of bytes to be read.
+                  To exit and output any fatal error messages the new write
+                  function should call png_error(png_ptr, "Error msg"). */
+void PNGAPI
+png_set_read_fn(png_structp png_ptr, png_voidp io_ptr,
+   png_rw_ptr read_data_fn)
+{
+   if(png_ptr == NULL) return;
+   png_ptr->io_ptr = io_ptr;
+
+#if !defined(PNG_NO_STDIO)
+   if (read_data_fn != NULL)
+      png_ptr->read_data_fn = read_data_fn;
+   else
+      png_ptr->read_data_fn = png_default_read_data;
+#else
+   png_ptr->read_data_fn = read_data_fn;
+#endif
+
+   /* It is an error to write to a read device */
+   if (png_ptr->write_data_fn != NULL)
+   {
+      png_ptr->write_data_fn = NULL;
+      png_warning(png_ptr,
+         "It's an error to set both read_data_fn and write_data_fn in the ");
+      png_warning(png_ptr,
+         "same structure.  Resetting write_data_fn to NULL.");
+   }
+
+#if defined(PNG_WRITE_FLUSH_SUPPORTED)
+   png_ptr->output_flush_fn = NULL;
+#endif
+}
+#endif /* PNG_READ_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngrtran.c b/distrib/libpng-1.2.19/pngrtran.c
new file mode 100644
index 0000000..3f04051
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngrtran.c
@@ -0,0 +1,4284 @@
+
+/* pngrtran.c - transforms the data in a row for PNG readers
+ *
+ * Last changed in libpng 1.2.19 August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ *
+ * This file contains functions optionally called by an application
+ * in order to tell libpng how to handle data when reading a PNG.
+ * Transformations that are used in both reading and writing are
+ * in pngtrans.c.
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_READ_SUPPORTED)
+
+/* Set the action on getting a CRC error for an ancillary or critical chunk. */
+void PNGAPI
+png_set_crc_action(png_structp png_ptr, int crit_action, int ancil_action)
+{
+   png_debug(1, "in png_set_crc_action\n");
+   /* Tell libpng how we react to CRC errors in critical chunks */
+   if(png_ptr == NULL) return;
+   switch (crit_action)
+   {
+      case PNG_CRC_NO_CHANGE:                        /* leave setting as is */
+         break;
+      case PNG_CRC_WARN_USE:                               /* warn/use data */
+         png_ptr->flags &= ~PNG_FLAG_CRC_CRITICAL_MASK;
+         png_ptr->flags |= PNG_FLAG_CRC_CRITICAL_USE;
+         break;
+      case PNG_CRC_QUIET_USE:                             /* quiet/use data */
+         png_ptr->flags &= ~PNG_FLAG_CRC_CRITICAL_MASK;
+         png_ptr->flags |= PNG_FLAG_CRC_CRITICAL_USE |
+                           PNG_FLAG_CRC_CRITICAL_IGNORE;
+         break;
+      case PNG_CRC_WARN_DISCARD:    /* not a valid action for critical data */
+         png_warning(png_ptr, "Can't discard critical data on CRC error.");
+      case PNG_CRC_ERROR_QUIT:                                /* error/quit */
+      case PNG_CRC_DEFAULT:
+      default:
+         png_ptr->flags &= ~PNG_FLAG_CRC_CRITICAL_MASK;
+         break;
+   }
+
+   switch (ancil_action)
+   {
+      case PNG_CRC_NO_CHANGE:                       /* leave setting as is */
+         break;
+      case PNG_CRC_WARN_USE:                              /* warn/use data */
+         png_ptr->flags &= ~PNG_FLAG_CRC_ANCILLARY_MASK;
+         png_ptr->flags |= PNG_FLAG_CRC_ANCILLARY_USE;
+         break;
+      case PNG_CRC_QUIET_USE:                            /* quiet/use data */
+         png_ptr->flags &= ~PNG_FLAG_CRC_ANCILLARY_MASK;
+         png_ptr->flags |= PNG_FLAG_CRC_ANCILLARY_USE |
+                           PNG_FLAG_CRC_ANCILLARY_NOWARN;
+         break;
+      case PNG_CRC_ERROR_QUIT:                               /* error/quit */
+         png_ptr->flags &= ~PNG_FLAG_CRC_ANCILLARY_MASK;
+         png_ptr->flags |= PNG_FLAG_CRC_ANCILLARY_NOWARN;
+         break;
+      case PNG_CRC_WARN_DISCARD:                      /* warn/discard data */
+      case PNG_CRC_DEFAULT:
+      default:
+         png_ptr->flags &= ~PNG_FLAG_CRC_ANCILLARY_MASK;
+         break;
+   }
+}
+
+#if defined(PNG_READ_BACKGROUND_SUPPORTED) && \
+    defined(PNG_FLOATING_POINT_SUPPORTED)
+/* handle alpha and tRNS via a background color */
+void PNGAPI
+png_set_background(png_structp png_ptr,
+   png_color_16p background_color, int background_gamma_code,
+   int need_expand, double background_gamma)
+{
+   png_debug(1, "in png_set_background\n");
+   if(png_ptr == NULL) return;
+   if (background_gamma_code == PNG_BACKGROUND_GAMMA_UNKNOWN)
+   {
+      png_warning(png_ptr, "Application must supply a known background gamma");
+      return;
+   }
+
+   png_ptr->transformations |= PNG_BACKGROUND;
+   png_memcpy(&(png_ptr->background), background_color,
+      png_sizeof(png_color_16));
+   png_ptr->background_gamma = (float)background_gamma;
+   png_ptr->background_gamma_type = (png_byte)(background_gamma_code);
+   png_ptr->transformations |= (need_expand ? PNG_BACKGROUND_EXPAND : 0);
+}
+#endif
+
+#if defined(PNG_READ_16_TO_8_SUPPORTED)
+/* strip 16 bit depth files to 8 bit depth */
+void PNGAPI
+png_set_strip_16(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_strip_16\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= PNG_16_TO_8;
+}
+#endif
+
+#if defined(PNG_READ_STRIP_ALPHA_SUPPORTED)
+void PNGAPI
+png_set_strip_alpha(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_strip_alpha\n");
+   if(png_ptr == NULL) return;
+   png_ptr->flags |= PNG_FLAG_STRIP_ALPHA;
+}
+#endif
+
+#if defined(PNG_READ_DITHER_SUPPORTED)
+/* Dither file to 8 bit.  Supply a palette, the current number
+ * of elements in the palette, the maximum number of elements
+ * allowed, and a histogram if possible.  If the current number
+ * of colors is greater then the maximum number, the palette will be
+ * modified to fit in the maximum number.  "full_dither" indicates
+ * whether we need a dithering cube set up for RGB images, or if we
+ * simply are reducing the number of colors in a paletted image.
+ */
+
+typedef struct png_dsort_struct
+{
+   struct png_dsort_struct FAR * next;
+   png_byte left;
+   png_byte right;
+} png_dsort;
+typedef png_dsort FAR *       png_dsortp;
+typedef png_dsort FAR * FAR * png_dsortpp;
+
+void PNGAPI
+png_set_dither(png_structp png_ptr, png_colorp palette,
+   int num_palette, int maximum_colors, png_uint_16p histogram,
+   int full_dither)
+{
+   png_debug(1, "in png_set_dither\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= PNG_DITHER;
+
+   if (!full_dither)
+   {
+      int i;
+
+      png_ptr->dither_index = (png_bytep)png_malloc(png_ptr,
+         (png_uint_32)(num_palette * png_sizeof (png_byte)));
+      for (i = 0; i < num_palette; i++)
+         png_ptr->dither_index[i] = (png_byte)i;
+   }
+
+   if (num_palette > maximum_colors)
+   {
+      if (histogram != NULL)
+      {
+         /* This is easy enough, just throw out the least used colors.
+            Perhaps not the best solution, but good enough. */
+
+         int i;
+
+         /* initialize an array to sort colors */
+         png_ptr->dither_sort = (png_bytep)png_malloc(png_ptr,
+            (png_uint_32)(num_palette * png_sizeof (png_byte)));
+
+         /* initialize the dither_sort array */
+         for (i = 0; i < num_palette; i++)
+            png_ptr->dither_sort[i] = (png_byte)i;
+
+         /* Find the least used palette entries by starting a
+            bubble sort, and running it until we have sorted
+            out enough colors.  Note that we don't care about
+            sorting all the colors, just finding which are
+            least used. */
+
+         for (i = num_palette - 1; i >= maximum_colors; i--)
+         {
+            int done; /* to stop early if the list is pre-sorted */
+            int j;
+
+            done = 1;
+            for (j = 0; j < i; j++)
+            {
+               if (histogram[png_ptr->dither_sort[j]]
+                   < histogram[png_ptr->dither_sort[j + 1]])
+               {
+                  png_byte t;
+
+                  t = png_ptr->dither_sort[j];
+                  png_ptr->dither_sort[j] = png_ptr->dither_sort[j + 1];
+                  png_ptr->dither_sort[j + 1] = t;
+                  done = 0;
+               }
+            }
+            if (done)
+               break;
+         }
+
+         /* swap the palette around, and set up a table, if necessary */
+         if (full_dither)
+         {
+            int j = num_palette;
+
+            /* put all the useful colors within the max, but don't
+               move the others */
+            for (i = 0; i < maximum_colors; i++)
+            {
+               if ((int)png_ptr->dither_sort[i] >= maximum_colors)
+               {
+                  do
+                     j--;
+                  while ((int)png_ptr->dither_sort[j] >= maximum_colors);
+                  palette[i] = palette[j];
+               }
+            }
+         }
+         else
+         {
+            int j = num_palette;
+
+            /* move all the used colors inside the max limit, and
+               develop a translation table */
+            for (i = 0; i < maximum_colors; i++)
+            {
+               /* only move the colors we need to */
+               if ((int)png_ptr->dither_sort[i] >= maximum_colors)
+               {
+                  png_color tmp_color;
+
+                  do
+                     j--;
+                  while ((int)png_ptr->dither_sort[j] >= maximum_colors);
+
+                  tmp_color = palette[j];
+                  palette[j] = palette[i];
+                  palette[i] = tmp_color;
+                  /* indicate where the color went */
+                  png_ptr->dither_index[j] = (png_byte)i;
+                  png_ptr->dither_index[i] = (png_byte)j;
+               }
+            }
+
+            /* find closest color for those colors we are not using */
+            for (i = 0; i < num_palette; i++)
+            {
+               if ((int)png_ptr->dither_index[i] >= maximum_colors)
+               {
+                  int min_d, k, min_k, d_index;
+
+                  /* find the closest color to one we threw out */
+                  d_index = png_ptr->dither_index[i];
+                  min_d = PNG_COLOR_DIST(palette[d_index], palette[0]);
+                  for (k = 1, min_k = 0; k < maximum_colors; k++)
+                  {
+                     int d;
+
+                     d = PNG_COLOR_DIST(palette[d_index], palette[k]);
+
+                     if (d < min_d)
+                     {
+                        min_d = d;
+                        min_k = k;
+                     }
+                  }
+                  /* point to closest color */
+                  png_ptr->dither_index[i] = (png_byte)min_k;
+               }
+            }
+         }
+         png_free(png_ptr, png_ptr->dither_sort);
+         png_ptr->dither_sort=NULL;
+      }
+      else
+      {
+         /* This is much harder to do simply (and quickly).  Perhaps
+            we need to go through a median cut routine, but those
+            don't always behave themselves with only a few colors
+            as input.  So we will just find the closest two colors,
+            and throw out one of them (chosen somewhat randomly).
+            [We don't understand this at all, so if someone wants to
+             work on improving it, be our guest - AED, GRP]
+            */
+         int i;
+         int max_d;
+         int num_new_palette;
+         png_dsortp t;
+         png_dsortpp hash;
+
+         t=NULL;
+
+         /* initialize palette index arrays */
+         png_ptr->index_to_palette = (png_bytep)png_malloc(png_ptr,
+            (png_uint_32)(num_palette * png_sizeof (png_byte)));
+         png_ptr->palette_to_index = (png_bytep)png_malloc(png_ptr,
+            (png_uint_32)(num_palette * png_sizeof (png_byte)));
+
+         /* initialize the sort array */
+         for (i = 0; i < num_palette; i++)
+         {
+            png_ptr->index_to_palette[i] = (png_byte)i;
+            png_ptr->palette_to_index[i] = (png_byte)i;
+         }
+
+         hash = (png_dsortpp)png_malloc(png_ptr, (png_uint_32)(769 *
+            png_sizeof (png_dsortp)));
+         for (i = 0; i < 769; i++)
+            hash[i] = NULL;
+/*         png_memset(hash, 0, 769 * png_sizeof (png_dsortp)); */
+
+         num_new_palette = num_palette;
+
+         /* initial wild guess at how far apart the farthest pixel
+            pair we will be eliminating will be.  Larger
+            numbers mean more areas will be allocated, Smaller
+            numbers run the risk of not saving enough data, and
+            having to do this all over again.
+
+            I have not done extensive checking on this number.
+            */
+         max_d = 96;
+
+         while (num_new_palette > maximum_colors)
+         {
+            for (i = 0; i < num_new_palette - 1; i++)
+            {
+               int j;
+
+               for (j = i + 1; j < num_new_palette; j++)
+               {
+                  int d;
+
+                  d = PNG_COLOR_DIST(palette[i], palette[j]);
+
+                  if (d <= max_d)
+                  {
+
+                     t = (png_dsortp)png_malloc_warn(png_ptr,
+                         (png_uint_32)(png_sizeof(png_dsort)));
+                     if (t == NULL)
+                         break;
+                     t->next = hash[d];
+                     t->left = (png_byte)i;
+                     t->right = (png_byte)j;
+                     hash[d] = t;
+                  }
+               }
+               if (t == NULL)
+                  break;
+            }
+
+            if (t != NULL)
+            for (i = 0; i <= max_d; i++)
+            {
+               if (hash[i] != NULL)
+               {
+                  png_dsortp p;
+
+                  for (p = hash[i]; p; p = p->next)
+                  {
+                     if ((int)png_ptr->index_to_palette[p->left]
+                        < num_new_palette &&
+                        (int)png_ptr->index_to_palette[p->right]
+                        < num_new_palette)
+                     {
+                        int j, next_j;
+
+                        if (num_new_palette & 0x01)
+                        {
+                           j = p->left;
+                           next_j = p->right;
+                        }
+                        else
+                        {
+                           j = p->right;
+                           next_j = p->left;
+                        }
+
+                        num_new_palette--;
+                        palette[png_ptr->index_to_palette[j]]
+                          = palette[num_new_palette];
+                        if (!full_dither)
+                        {
+                           int k;
+
+                           for (k = 0; k < num_palette; k++)
+                           {
+                              if (png_ptr->dither_index[k] ==
+                                 png_ptr->index_to_palette[j])
+                                 png_ptr->dither_index[k] =
+                                    png_ptr->index_to_palette[next_j];
+                              if ((int)png_ptr->dither_index[k] ==
+                                 num_new_palette)
+                                 png_ptr->dither_index[k] =
+                                    png_ptr->index_to_palette[j];
+                           }
+                        }
+
+                        png_ptr->index_to_palette[png_ptr->palette_to_index
+                           [num_new_palette]] = png_ptr->index_to_palette[j];
+                        png_ptr->palette_to_index[png_ptr->index_to_palette[j]]
+                           = png_ptr->palette_to_index[num_new_palette];
+
+                        png_ptr->index_to_palette[j] = (png_byte)num_new_palette;
+                        png_ptr->palette_to_index[num_new_palette] = (png_byte)j;
+                     }
+                     if (num_new_palette <= maximum_colors)
+                        break;
+                  }
+                  if (num_new_palette <= maximum_colors)
+                     break;
+               }
+            }
+
+            for (i = 0; i < 769; i++)
+            {
+               if (hash[i] != NULL)
+               {
+                  png_dsortp p = hash[i];
+                  while (p)
+                  {
+                     t = p->next;
+                     png_free(png_ptr, p);
+                     p = t;
+                  }
+               }
+               hash[i] = 0;
+            }
+            max_d += 96;
+         }
+         png_free(png_ptr, hash);
+         png_free(png_ptr, png_ptr->palette_to_index);
+         png_free(png_ptr, png_ptr->index_to_palette);
+         png_ptr->palette_to_index=NULL;
+         png_ptr->index_to_palette=NULL;
+      }
+      num_palette = maximum_colors;
+   }
+   if (png_ptr->palette == NULL)
+   {
+      png_ptr->palette = palette;
+   }
+   png_ptr->num_palette = (png_uint_16)num_palette;
+
+   if (full_dither)
+   {
+      int i;
+      png_bytep distance;
+      int total_bits = PNG_DITHER_RED_BITS + PNG_DITHER_GREEN_BITS +
+         PNG_DITHER_BLUE_BITS;
+      int num_red = (1 << PNG_DITHER_RED_BITS);
+      int num_green = (1 << PNG_DITHER_GREEN_BITS);
+      int num_blue = (1 << PNG_DITHER_BLUE_BITS);
+      png_size_t num_entries = ((png_size_t)1 << total_bits);
+
+      png_ptr->palette_lookup = (png_bytep )png_malloc(png_ptr,
+         (png_uint_32)(num_entries * png_sizeof (png_byte)));
+
+      png_memset(png_ptr->palette_lookup, 0, num_entries *
+         png_sizeof (png_byte));
+
+      distance = (png_bytep)png_malloc(png_ptr, (png_uint_32)(num_entries *
+         png_sizeof(png_byte)));
+
+      png_memset(distance, 0xff, num_entries * png_sizeof(png_byte));
+
+      for (i = 0; i < num_palette; i++)
+      {
+         int ir, ig, ib;
+         int r = (palette[i].red >> (8 - PNG_DITHER_RED_BITS));
+         int g = (palette[i].green >> (8 - PNG_DITHER_GREEN_BITS));
+         int b = (palette[i].blue >> (8 - PNG_DITHER_BLUE_BITS));
+
+         for (ir = 0; ir < num_red; ir++)
+         {
+            /* int dr = abs(ir - r); */
+            int dr = ((ir > r) ? ir - r : r - ir);
+            int index_r = (ir << (PNG_DITHER_BLUE_BITS + PNG_DITHER_GREEN_BITS));
+
+            for (ig = 0; ig < num_green; ig++)
+            {
+               /* int dg = abs(ig - g); */
+               int dg = ((ig > g) ? ig - g : g - ig);
+               int dt = dr + dg;
+               int dm = ((dr > dg) ? dr : dg);
+               int index_g = index_r | (ig << PNG_DITHER_BLUE_BITS);
+
+               for (ib = 0; ib < num_blue; ib++)
+               {
+                  int d_index = index_g | ib;
+                  /* int db = abs(ib - b); */
+                  int db = ((ib > b) ? ib - b : b - ib);
+                  int dmax = ((dm > db) ? dm : db);
+                  int d = dmax + dt + db;
+
+                  if (d < (int)distance[d_index])
+                  {
+                     distance[d_index] = (png_byte)d;
+                     png_ptr->palette_lookup[d_index] = (png_byte)i;
+                  }
+               }
+            }
+         }
+      }
+
+      png_free(png_ptr, distance);
+   }
+}
+#endif
+
+#if defined(PNG_READ_GAMMA_SUPPORTED) && defined(PNG_FLOATING_POINT_SUPPORTED)
+/* Transform the image from the file_gamma to the screen_gamma.  We
+ * only do transformations on images where the file_gamma and screen_gamma
+ * are not close reciprocals, otherwise it slows things down slightly, and
+ * also needlessly introduces small errors.
+ *
+ * We will turn off gamma transformation later if no semitransparent entries
+ * are present in the tRNS array for palette images.  We can't do it here
+ * because we don't necessarily have the tRNS chunk yet.
+ */
+void PNGAPI
+png_set_gamma(png_structp png_ptr, double scrn_gamma, double file_gamma)
+{
+   png_debug(1, "in png_set_gamma\n");
+   if(png_ptr == NULL) return;
+   if ((fabs(scrn_gamma * file_gamma - 1.0) > PNG_GAMMA_THRESHOLD) ||
+       (png_ptr->color_type & PNG_COLOR_MASK_ALPHA) ||
+       (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE))
+     png_ptr->transformations |= PNG_GAMMA;
+   png_ptr->gamma = (float)file_gamma;
+   png_ptr->screen_gamma = (float)scrn_gamma;
+}
+#endif
+
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+/* Expand paletted images to RGB, expand grayscale images of
+ * less than 8-bit depth to 8-bit depth, and expand tRNS chunks
+ * to alpha channels.
+ */
+void PNGAPI
+png_set_expand(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_expand\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= (PNG_EXPAND | PNG_EXPAND_tRNS);
+#ifdef PNG_WARN_UNINITIALIZED_ROW
+   png_ptr->flags &= !(PNG_FLAG_ROW_INIT);
+#endif
+}
+
+/* GRR 19990627:  the following three functions currently are identical
+ *  to png_set_expand().  However, it is entirely reasonable that someone
+ *  might wish to expand an indexed image to RGB but *not* expand a single,
+ *  fully transparent palette entry to a full alpha channel--perhaps instead
+ *  convert tRNS to the grayscale/RGB format (16-bit RGB value), or replace
+ *  the transparent color with a particular RGB value, or drop tRNS entirely.
+ *  IOW, a future version of the library may make the transformations flag
+ *  a bit more fine-grained, with separate bits for each of these three
+ *  functions.
+ *
+ *  More to the point, these functions make it obvious what libpng will be
+ *  doing, whereas "expand" can (and does) mean any number of things.
+ *
+ *  GRP 20060307: In libpng-1.4.0, png_set_gray_1_2_4_to_8() was modified
+ *  to expand only the sample depth but not to expand the tRNS to alpha.
+ */
+
+/* Expand paletted images to RGB. */
+void PNGAPI
+png_set_palette_to_rgb(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_palette_to_rgb\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= (PNG_EXPAND | PNG_EXPAND_tRNS);
+#ifdef PNG_WARN_UNINITIALIZED_ROW
+   png_ptr->flags &= !(PNG_FLAG_ROW_INIT);
+#endif
+}
+
+#if !defined(PNG_1_0_X)
+/* Expand grayscale images of less than 8-bit depth to 8 bits. */
+void PNGAPI
+png_set_expand_gray_1_2_4_to_8(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_expand_gray_1_2_4_to_8\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= PNG_EXPAND;
+#ifdef PNG_WARN_UNINITIALIZED_ROW
+   png_ptr->flags &= !(PNG_FLAG_ROW_INIT);
+#endif
+}
+#endif
+
+#if defined(PNG_1_0_X) || defined(PNG_1_2_X)
+/* Expand grayscale images of less than 8-bit depth to 8 bits. */
+/* Deprecated as of libpng-1.2.9 */
+void PNGAPI
+png_set_gray_1_2_4_to_8(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_gray_1_2_4_to_8\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= (PNG_EXPAND | PNG_EXPAND_tRNS);
+}
+#endif
+
+
+/* Expand tRNS chunks to alpha channels. */
+void PNGAPI
+png_set_tRNS_to_alpha(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_tRNS_to_alpha\n");
+   png_ptr->transformations |= (PNG_EXPAND | PNG_EXPAND_tRNS);
+#ifdef PNG_WARN_UNINITIALIZED_ROW
+   png_ptr->flags &= !(PNG_FLAG_ROW_INIT);
+#endif
+}
+#endif /* defined(PNG_READ_EXPAND_SUPPORTED) */
+
+#if defined(PNG_READ_GRAY_TO_RGB_SUPPORTED)
+void PNGAPI
+png_set_gray_to_rgb(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_gray_to_rgb\n");
+   png_ptr->transformations |= PNG_GRAY_TO_RGB;
+#ifdef PNG_WARN_UNINITIALIZED_ROW
+   png_ptr->flags &= !(PNG_FLAG_ROW_INIT);
+#endif
+}
+#endif
+
+#if defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+#if defined(PNG_FLOATING_POINT_SUPPORTED)
+/* Convert a RGB image to a grayscale of the same width.  This allows us,
+ * for example, to convert a 24 bpp RGB image into an 8 bpp grayscale image.
+ */
+
+void PNGAPI
+png_set_rgb_to_gray(png_structp png_ptr, int error_action, double red,
+   double green)
+{
+      int red_fixed = (int)((float)red*100000.0 + 0.5);
+      int green_fixed = (int)((float)green*100000.0 + 0.5);
+      if(png_ptr == NULL) return;
+      png_set_rgb_to_gray_fixed(png_ptr, error_action, red_fixed, green_fixed);
+}
+#endif
+
+void PNGAPI
+png_set_rgb_to_gray_fixed(png_structp png_ptr, int error_action,
+   png_fixed_point red, png_fixed_point green)
+{
+   png_debug(1, "in png_set_rgb_to_gray\n");
+   if(png_ptr == NULL) return;
+   switch(error_action)
+   {
+      case 1: png_ptr->transformations |= PNG_RGB_TO_GRAY;
+              break;
+      case 2: png_ptr->transformations |= PNG_RGB_TO_GRAY_WARN;
+              break;
+      case 3: png_ptr->transformations |= PNG_RGB_TO_GRAY_ERR;
+   }
+   if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+      png_ptr->transformations |= PNG_EXPAND;
+#else
+   {
+      png_warning(png_ptr, "Cannot do RGB_TO_GRAY without EXPAND_SUPPORTED.");
+      png_ptr->transformations &= ~PNG_RGB_TO_GRAY;
+   }
+#endif
+   {
+      png_uint_16 red_int, green_int;
+      if(red < 0 || green < 0)
+      {
+         red_int   =  6968; /* .212671 * 32768 + .5 */
+         green_int = 23434; /* .715160 * 32768 + .5 */
+      }
+      else if(red + green < 100000L)
+      {
+        red_int = (png_uint_16)(((png_uint_32)red*32768L)/100000L);
+        green_int = (png_uint_16)(((png_uint_32)green*32768L)/100000L);
+      }
+      else
+      {
+         png_warning(png_ptr, "ignoring out of range rgb_to_gray coefficients");
+         red_int   =  6968;
+         green_int = 23434;
+      }
+      png_ptr->rgb_to_gray_red_coeff   = red_int;
+      png_ptr->rgb_to_gray_green_coeff = green_int;
+      png_ptr->rgb_to_gray_blue_coeff  = (png_uint_16)(32768-red_int-green_int);
+   }
+}
+#endif
+
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_LEGACY_SUPPORTED)
+void PNGAPI
+png_set_read_user_transform_fn(png_structp png_ptr, png_user_transform_ptr
+   read_user_transform_fn)
+{
+   png_debug(1, "in png_set_read_user_transform_fn\n");
+   if(png_ptr == NULL) return;
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED)
+   png_ptr->transformations |= PNG_USER_TRANSFORM;
+   png_ptr->read_user_transform_fn = read_user_transform_fn;
+#endif
+#ifdef PNG_LEGACY_SUPPORTED
+   if(read_user_transform_fn)
+      png_warning(png_ptr,
+        "This version of libpng does not support user transforms");
+#endif
+}
+#endif
+
+/* Initialize everything needed for the read.  This includes modifying
+ * the palette.
+ */
+void /* PRIVATE */
+png_init_read_transformations(png_structp png_ptr)
+{
+   png_debug(1, "in png_init_read_transformations\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if(png_ptr != NULL)
+#endif
+  {
+#if defined(PNG_READ_BACKGROUND_SUPPORTED) || defined(PNG_READ_SHIFT_SUPPORTED) \
+ || defined(PNG_READ_GAMMA_SUPPORTED)
+   int color_type = png_ptr->color_type;
+#endif
+
+#if defined(PNG_READ_EXPAND_SUPPORTED) && defined(PNG_READ_BACKGROUND_SUPPORTED)
+
+#if defined(PNG_READ_GRAY_TO_RGB_SUPPORTED)
+   /* Detect gray background and attempt to enable optimization
+    * for gray --> RGB case */
+   /* Note:  if PNG_BACKGROUND_EXPAND is set and color_type is either RGB or
+    * RGB_ALPHA (in which case need_expand is superfluous anyway), the
+    * background color might actually be gray yet not be flagged as such.
+    * This is not a problem for the current code, which uses
+    * PNG_BACKGROUND_IS_GRAY only to decide when to do the
+    * png_do_gray_to_rgb() transformation.
+    */
+   if ((png_ptr->transformations & PNG_BACKGROUND_EXPAND) &&
+       !(color_type & PNG_COLOR_MASK_COLOR))
+   {
+          png_ptr->mode |= PNG_BACKGROUND_IS_GRAY;
+   } else if ((png_ptr->transformations & PNG_BACKGROUND) &&
+              !(png_ptr->transformations & PNG_BACKGROUND_EXPAND) &&
+              (png_ptr->transformations & PNG_GRAY_TO_RGB) &&
+              png_ptr->background.red == png_ptr->background.green &&
+              png_ptr->background.red == png_ptr->background.blue)
+   {
+          png_ptr->mode |= PNG_BACKGROUND_IS_GRAY;
+          png_ptr->background.gray = png_ptr->background.red;
+   }
+#endif
+
+   if ((png_ptr->transformations & PNG_BACKGROUND_EXPAND) &&
+       (png_ptr->transformations & PNG_EXPAND))
+   {
+      if (!(color_type & PNG_COLOR_MASK_COLOR))  /* i.e., GRAY or GRAY_ALPHA */
+      {
+         /* expand background and tRNS chunks */
+         switch (png_ptr->bit_depth)
+         {
+            case 1:
+               png_ptr->background.gray *= (png_uint_16)0xff;
+               png_ptr->background.red = png_ptr->background.green
+                 =  png_ptr->background.blue = png_ptr->background.gray;
+               if (!(png_ptr->transformations & PNG_EXPAND_tRNS))
+               {
+                 png_ptr->trans_values.gray *= (png_uint_16)0xff;
+                 png_ptr->trans_values.red = png_ptr->trans_values.green
+                   = png_ptr->trans_values.blue = png_ptr->trans_values.gray;
+               }
+               break;
+            case 2:
+               png_ptr->background.gray *= (png_uint_16)0x55;
+               png_ptr->background.red = png_ptr->background.green
+                 = png_ptr->background.blue = png_ptr->background.gray;
+               if (!(png_ptr->transformations & PNG_EXPAND_tRNS))
+               {
+                 png_ptr->trans_values.gray *= (png_uint_16)0x55;
+                 png_ptr->trans_values.red = png_ptr->trans_values.green
+                   = png_ptr->trans_values.blue = png_ptr->trans_values.gray;
+               }
+               break;
+            case 4:
+               png_ptr->background.gray *= (png_uint_16)0x11;
+               png_ptr->background.red = png_ptr->background.green
+                 = png_ptr->background.blue = png_ptr->background.gray;
+               if (!(png_ptr->transformations & PNG_EXPAND_tRNS))
+               {
+                 png_ptr->trans_values.gray *= (png_uint_16)0x11;
+                 png_ptr->trans_values.red = png_ptr->trans_values.green
+                   = png_ptr->trans_values.blue = png_ptr->trans_values.gray;
+               }
+               break;
+            case 8:
+            case 16:
+               png_ptr->background.red = png_ptr->background.green
+                 = png_ptr->background.blue = png_ptr->background.gray;
+               break;
+         }
+      }
+      else if (color_type == PNG_COLOR_TYPE_PALETTE)
+      {
+         png_ptr->background.red   =
+            png_ptr->palette[png_ptr->background.index].red;
+         png_ptr->background.green =
+            png_ptr->palette[png_ptr->background.index].green;
+         png_ptr->background.blue  =
+            png_ptr->palette[png_ptr->background.index].blue;
+
+#if defined(PNG_READ_INVERT_ALPHA_SUPPORTED)
+        if (png_ptr->transformations & PNG_INVERT_ALPHA)
+        {
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+           if (!(png_ptr->transformations & PNG_EXPAND_tRNS))
+#endif
+           {
+           /* invert the alpha channel (in tRNS) unless the pixels are
+              going to be expanded, in which case leave it for later */
+              int i,istop;
+              istop=(int)png_ptr->num_trans;
+              for (i=0; i<istop; i++)
+                 png_ptr->trans[i] = (png_byte)(255 - png_ptr->trans[i]);
+           }
+        }
+#endif
+
+      }
+   }
+#endif
+
+#if defined(PNG_READ_BACKGROUND_SUPPORTED) && defined(PNG_READ_GAMMA_SUPPORTED)
+   png_ptr->background_1 = png_ptr->background;
+#endif
+#if defined(PNG_READ_GAMMA_SUPPORTED) && defined(PNG_FLOATING_POINT_SUPPORTED)
+
+   if ((color_type == PNG_COLOR_TYPE_PALETTE && png_ptr->num_trans != 0)
+       && (fabs(png_ptr->screen_gamma * png_ptr->gamma - 1.0)
+         < PNG_GAMMA_THRESHOLD))
+   {
+    int i,k;
+    k=0;
+    for (i=0; i<png_ptr->num_trans; i++)
+    {
+      if (png_ptr->trans[i] != 0 && png_ptr->trans[i] != 0xff)
+        k=1; /* partial transparency is present */
+    }
+    if (k == 0)
+      png_ptr->transformations &= (~PNG_GAMMA);
+   }
+
+   if ((png_ptr->transformations & (PNG_GAMMA | PNG_RGB_TO_GRAY)) &&
+        png_ptr->gamma != 0.0)
+   {
+      png_build_gamma_table(png_ptr);
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+      if (png_ptr->transformations & PNG_BACKGROUND)
+      {
+         if (color_type == PNG_COLOR_TYPE_PALETTE)
+         {
+           /* could skip if no transparency and
+           */
+            png_color back, back_1;
+            png_colorp palette = png_ptr->palette;
+            int num_palette = png_ptr->num_palette;
+            int i;
+            if (png_ptr->background_gamma_type == PNG_BACKGROUND_GAMMA_FILE)
+            {
+               back.red = png_ptr->gamma_table[png_ptr->background.red];
+               back.green = png_ptr->gamma_table[png_ptr->background.green];
+               back.blue = png_ptr->gamma_table[png_ptr->background.blue];
+
+               back_1.red = png_ptr->gamma_to_1[png_ptr->background.red];
+               back_1.green = png_ptr->gamma_to_1[png_ptr->background.green];
+               back_1.blue = png_ptr->gamma_to_1[png_ptr->background.blue];
+            }
+            else
+            {
+               double g, gs;
+
+               switch (png_ptr->background_gamma_type)
+               {
+                  case PNG_BACKGROUND_GAMMA_SCREEN:
+                     g = (png_ptr->screen_gamma);
+                     gs = 1.0;
+                     break;
+                  case PNG_BACKGROUND_GAMMA_FILE:
+                     g = 1.0 / (png_ptr->gamma);
+                     gs = 1.0 / (png_ptr->gamma * png_ptr->screen_gamma);
+                     break;
+                  case PNG_BACKGROUND_GAMMA_UNIQUE:
+                     g = 1.0 / (png_ptr->background_gamma);
+                     gs = 1.0 / (png_ptr->background_gamma *
+                                 png_ptr->screen_gamma);
+                     break;
+                  default:
+                     g = 1.0;    /* back_1 */
+                     gs = 1.0;   /* back */
+               }
+
+               if ( fabs(gs - 1.0) < PNG_GAMMA_THRESHOLD)
+               {
+                  back.red   = (png_byte)png_ptr->background.red;
+                  back.green = (png_byte)png_ptr->background.green;
+                  back.blue  = (png_byte)png_ptr->background.blue;
+               }
+               else
+               {
+                  back.red = (png_byte)(pow(
+                     (double)png_ptr->background.red/255, gs) * 255.0 + .5);
+                  back.green = (png_byte)(pow(
+                     (double)png_ptr->background.green/255, gs) * 255.0 + .5);
+                  back.blue = (png_byte)(pow(
+                     (double)png_ptr->background.blue/255, gs) * 255.0 + .5);
+               }
+
+               back_1.red = (png_byte)(pow(
+                  (double)png_ptr->background.red/255, g) * 255.0 + .5);
+               back_1.green = (png_byte)(pow(
+                  (double)png_ptr->background.green/255, g) * 255.0 + .5);
+               back_1.blue = (png_byte)(pow(
+                  (double)png_ptr->background.blue/255, g) * 255.0 + .5);
+            }
+            for (i = 0; i < num_palette; i++)
+            {
+               if (i < (int)png_ptr->num_trans && png_ptr->trans[i] != 0xff)
+               {
+                  if (png_ptr->trans[i] == 0)
+                  {
+                     palette[i] = back;
+                  }
+                  else /* if (png_ptr->trans[i] != 0xff) */
+                  {
+                     png_byte v, w;
+
+                     v = png_ptr->gamma_to_1[palette[i].red];
+                     png_composite(w, v, png_ptr->trans[i], back_1.red);
+                     palette[i].red = png_ptr->gamma_from_1[w];
+
+                     v = png_ptr->gamma_to_1[palette[i].green];
+                     png_composite(w, v, png_ptr->trans[i], back_1.green);
+                     palette[i].green = png_ptr->gamma_from_1[w];
+
+                     v = png_ptr->gamma_to_1[palette[i].blue];
+                     png_composite(w, v, png_ptr->trans[i], back_1.blue);
+                     palette[i].blue = png_ptr->gamma_from_1[w];
+                  }
+               }
+               else
+               {
+                  palette[i].red = png_ptr->gamma_table[palette[i].red];
+                  palette[i].green = png_ptr->gamma_table[palette[i].green];
+                  palette[i].blue = png_ptr->gamma_table[palette[i].blue];
+               }
+            }
+         }
+         /* if (png_ptr->background_gamma_type!=PNG_BACKGROUND_GAMMA_UNKNOWN) */
+         else
+         /* color_type != PNG_COLOR_TYPE_PALETTE */
+         {
+            double m = (double)(((png_uint_32)1 << png_ptr->bit_depth) - 1);
+            double g = 1.0;
+            double gs = 1.0;
+
+            switch (png_ptr->background_gamma_type)
+            {
+               case PNG_BACKGROUND_GAMMA_SCREEN:
+                  g = (png_ptr->screen_gamma);
+                  gs = 1.0;
+                  break;
+               case PNG_BACKGROUND_GAMMA_FILE:
+                  g = 1.0 / (png_ptr->gamma);
+                  gs = 1.0 / (png_ptr->gamma * png_ptr->screen_gamma);
+                  break;
+               case PNG_BACKGROUND_GAMMA_UNIQUE:
+                  g = 1.0 / (png_ptr->background_gamma);
+                  gs = 1.0 / (png_ptr->background_gamma *
+                     png_ptr->screen_gamma);
+                  break;
+            }
+
+            png_ptr->background_1.gray = (png_uint_16)(pow(
+               (double)png_ptr->background.gray / m, g) * m + .5);
+            png_ptr->background.gray = (png_uint_16)(pow(
+               (double)png_ptr->background.gray / m, gs) * m + .5);
+
+            if ((png_ptr->background.red != png_ptr->background.green) ||
+                (png_ptr->background.red != png_ptr->background.blue) ||
+                (png_ptr->background.red != png_ptr->background.gray))
+            {
+               /* RGB or RGBA with color background */
+               png_ptr->background_1.red = (png_uint_16)(pow(
+                  (double)png_ptr->background.red / m, g) * m + .5);
+               png_ptr->background_1.green = (png_uint_16)(pow(
+                  (double)png_ptr->background.green / m, g) * m + .5);
+               png_ptr->background_1.blue = (png_uint_16)(pow(
+                  (double)png_ptr->background.blue / m, g) * m + .5);
+               png_ptr->background.red = (png_uint_16)(pow(
+                  (double)png_ptr->background.red / m, gs) * m + .5);
+               png_ptr->background.green = (png_uint_16)(pow(
+                  (double)png_ptr->background.green / m, gs) * m + .5);
+               png_ptr->background.blue = (png_uint_16)(pow(
+                  (double)png_ptr->background.blue / m, gs) * m + .5);
+            }
+            else
+            {
+               /* GRAY, GRAY ALPHA, RGB, or RGBA with gray background */
+               png_ptr->background_1.red = png_ptr->background_1.green
+                 = png_ptr->background_1.blue = png_ptr->background_1.gray;
+               png_ptr->background.red = png_ptr->background.green
+                 = png_ptr->background.blue = png_ptr->background.gray;
+            }
+         }
+      }
+      else
+      /* transformation does not include PNG_BACKGROUND */
+#endif /* PNG_READ_BACKGROUND_SUPPORTED */
+      if (color_type == PNG_COLOR_TYPE_PALETTE)
+      {
+         png_colorp palette = png_ptr->palette;
+         int num_palette = png_ptr->num_palette;
+         int i;
+
+         for (i = 0; i < num_palette; i++)
+         {
+            palette[i].red = png_ptr->gamma_table[palette[i].red];
+            palette[i].green = png_ptr->gamma_table[palette[i].green];
+            palette[i].blue = png_ptr->gamma_table[palette[i].blue];
+         }
+      }
+   }
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+   else
+#endif
+#endif /* PNG_READ_GAMMA_SUPPORTED && PNG_FLOATING_POINT_SUPPORTED */
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+   /* No GAMMA transformation */
+   if ((png_ptr->transformations & PNG_BACKGROUND) &&
+       (color_type == PNG_COLOR_TYPE_PALETTE))
+   {
+      int i;
+      int istop = (int)png_ptr->num_trans;
+      png_color back;
+      png_colorp palette = png_ptr->palette;
+
+      back.red   = (png_byte)png_ptr->background.red;
+      back.green = (png_byte)png_ptr->background.green;
+      back.blue  = (png_byte)png_ptr->background.blue;
+
+      for (i = 0; i < istop; i++)
+      {
+         if (png_ptr->trans[i] == 0)
+         {
+            palette[i] = back;
+         }
+         else if (png_ptr->trans[i] != 0xff)
+         {
+            /* The png_composite() macro is defined in png.h */
+            png_composite(palette[i].red, palette[i].red,
+               png_ptr->trans[i], back.red);
+            png_composite(palette[i].green, palette[i].green,
+               png_ptr->trans[i], back.green);
+            png_composite(palette[i].blue, palette[i].blue,
+               png_ptr->trans[i], back.blue);
+         }
+      }
+   }
+#endif /* PNG_READ_BACKGROUND_SUPPORTED */
+
+#if defined(PNG_READ_SHIFT_SUPPORTED)
+   if ((png_ptr->transformations & PNG_SHIFT) &&
+      (color_type == PNG_COLOR_TYPE_PALETTE))
+   {
+      png_uint_16 i;
+      png_uint_16 istop = png_ptr->num_palette;
+      int sr = 8 - png_ptr->sig_bit.red;
+      int sg = 8 - png_ptr->sig_bit.green;
+      int sb = 8 - png_ptr->sig_bit.blue;
+
+      if (sr < 0 || sr > 8)
+         sr = 0;
+      if (sg < 0 || sg > 8)
+         sg = 0;
+      if (sb < 0 || sb > 8)
+         sb = 0;
+      for (i = 0; i < istop; i++)
+      {
+         png_ptr->palette[i].red >>= sr;
+         png_ptr->palette[i].green >>= sg;
+         png_ptr->palette[i].blue >>= sb;
+      }
+   }
+#endif  /* PNG_READ_SHIFT_SUPPORTED */
+ }
+#if !defined(PNG_READ_GAMMA_SUPPORTED) && !defined(PNG_READ_SHIFT_SUPPORTED) \
+ && !defined(PNG_READ_BACKGROUND_SUPPORTED)
+   if(png_ptr)
+      return;
+#endif
+}
+
+/* Modify the info structure to reflect the transformations.  The
+ * info should be updated so a PNG file could be written with it,
+ * assuming the transformations result in valid PNG data.
+ */
+void /* PRIVATE */
+png_read_transform_info(png_structp png_ptr, png_infop info_ptr)
+{
+   png_debug(1, "in png_read_transform_info\n");
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+   if (png_ptr->transformations & PNG_EXPAND)
+   {
+      if (info_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      {
+         if (png_ptr->num_trans && (png_ptr->transformations & PNG_EXPAND_tRNS))
+            info_ptr->color_type = PNG_COLOR_TYPE_RGB_ALPHA;
+         else
+            info_ptr->color_type = PNG_COLOR_TYPE_RGB;
+         info_ptr->bit_depth = 8;
+         info_ptr->num_trans = 0;
+      }
+      else
+      {
+         if (png_ptr->num_trans)
+         {
+            if (png_ptr->transformations & PNG_EXPAND_tRNS)
+              info_ptr->color_type |= PNG_COLOR_MASK_ALPHA;
+            else
+              info_ptr->color_type |= PNG_COLOR_MASK_COLOR;
+         }
+         if (info_ptr->bit_depth < 8)
+            info_ptr->bit_depth = 8;
+         info_ptr->num_trans = 0;
+      }
+   }
+#endif
+
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+   if (png_ptr->transformations & PNG_BACKGROUND)
+   {
+      info_ptr->color_type &= ~PNG_COLOR_MASK_ALPHA;
+      info_ptr->num_trans = 0;
+      info_ptr->background = png_ptr->background;
+   }
+#endif
+
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+   if (png_ptr->transformations & PNG_GAMMA)
+   {
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+      info_ptr->gamma = png_ptr->gamma;
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+      info_ptr->int_gamma = png_ptr->int_gamma;
+#endif
+   }
+#endif
+
+#if defined(PNG_READ_16_TO_8_SUPPORTED)
+   if ((png_ptr->transformations & PNG_16_TO_8) && (info_ptr->bit_depth == 16))
+      info_ptr->bit_depth = 8;
+#endif
+
+#if defined(PNG_READ_GRAY_TO_RGB_SUPPORTED)
+   if (png_ptr->transformations & PNG_GRAY_TO_RGB)
+      info_ptr->color_type |= PNG_COLOR_MASK_COLOR;
+#endif
+
+#if defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+   if (png_ptr->transformations & PNG_RGB_TO_GRAY)
+      info_ptr->color_type &= ~PNG_COLOR_MASK_COLOR;
+#endif
+
+#if defined(PNG_READ_DITHER_SUPPORTED)
+   if (png_ptr->transformations & PNG_DITHER)
+   {
+      if (((info_ptr->color_type == PNG_COLOR_TYPE_RGB) ||
+         (info_ptr->color_type == PNG_COLOR_TYPE_RGB_ALPHA)) &&
+         png_ptr->palette_lookup && info_ptr->bit_depth == 8)
+      {
+         info_ptr->color_type = PNG_COLOR_TYPE_PALETTE;
+      }
+   }
+#endif
+
+#if defined(PNG_READ_PACK_SUPPORTED)
+   if ((png_ptr->transformations & PNG_PACK) && (info_ptr->bit_depth < 8))
+      info_ptr->bit_depth = 8;
+#endif
+
+   if (info_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      info_ptr->channels = 1;
+   else if (info_ptr->color_type & PNG_COLOR_MASK_COLOR)
+      info_ptr->channels = 3;
+   else
+      info_ptr->channels = 1;
+
+#if defined(PNG_READ_STRIP_ALPHA_SUPPORTED)
+   if (png_ptr->flags & PNG_FLAG_STRIP_ALPHA)
+      info_ptr->color_type &= ~PNG_COLOR_MASK_ALPHA;
+#endif
+
+   if (info_ptr->color_type & PNG_COLOR_MASK_ALPHA)
+      info_ptr->channels++;
+
+#if defined(PNG_READ_FILLER_SUPPORTED)
+   /* STRIP_ALPHA and FILLER allowed:  MASK_ALPHA bit stripped above */
+   if ((png_ptr->transformations & PNG_FILLER) &&
+       ((info_ptr->color_type == PNG_COLOR_TYPE_RGB) ||
+       (info_ptr->color_type == PNG_COLOR_TYPE_GRAY)))
+   {
+      info_ptr->channels++;
+      /* if adding a true alpha channel not just filler */
+#if !defined(PNG_1_0_X)
+      if (png_ptr->transformations & PNG_ADD_ALPHA)
+        info_ptr->color_type |= PNG_COLOR_MASK_ALPHA;
+#endif
+   }
+#endif
+
+#if defined(PNG_USER_TRANSFORM_PTR_SUPPORTED) && \
+defined(PNG_READ_USER_TRANSFORM_SUPPORTED)
+   if(png_ptr->transformations & PNG_USER_TRANSFORM)
+     {
+       if(info_ptr->bit_depth < png_ptr->user_transform_depth)
+         info_ptr->bit_depth = png_ptr->user_transform_depth;
+       if(info_ptr->channels < png_ptr->user_transform_channels)
+         info_ptr->channels = png_ptr->user_transform_channels;
+     }
+#endif
+
+   info_ptr->pixel_depth = (png_byte)(info_ptr->channels *
+      info_ptr->bit_depth);
+
+   info_ptr->rowbytes = PNG_ROWBYTES(info_ptr->pixel_depth,info_ptr->width);
+
+#if !defined(PNG_READ_EXPAND_SUPPORTED)
+   if(png_ptr)
+      return;
+#endif
+}
+
+/* Transform the row.  The order of transformations is significant,
+ * and is very touchy.  If you add a transformation, take care to
+ * decide how it fits in with the other transformations here.
+ */
+void /* PRIVATE */
+png_do_read_transformations(png_structp png_ptr)
+{
+   png_debug(1, "in png_do_read_transformations\n");
+   if (png_ptr->row_buf == NULL)
+   {
+#if !defined(PNG_NO_STDIO) && !defined(_WIN32_WCE)
+      char msg[50];
+
+      png_snprintf2(msg, 50,
+         "NULL row buffer for row %ld, pass %d", png_ptr->row_number,
+         png_ptr->pass);
+      png_error(png_ptr, msg);
+#else
+      png_error(png_ptr, "NULL row buffer");
+#endif
+   }
+#ifdef PNG_WARN_UNINITIALIZED_ROW
+   if (!(png_ptr->flags & PNG_FLAG_ROW_INIT))
+      /* Application has failed to call either png_read_start_image()
+       * or png_read_update_info() after setting transforms that expand
+       * pixels.  This check added to libpng-1.2.19 */
+#if (PNG_WARN_UNINITIALIZED_ROW==1)
+      png_error(png_ptr, "Uninitialized row");
+#else
+      png_warning(png_ptr, "Uninitialized row");
+#endif
+#endif
+
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+   if (png_ptr->transformations & PNG_EXPAND)
+   {
+      if (png_ptr->row_info.color_type == PNG_COLOR_TYPE_PALETTE)
+      {
+         png_do_expand_palette(&(png_ptr->row_info), png_ptr->row_buf + 1,
+            png_ptr->palette, png_ptr->trans, png_ptr->num_trans);
+      }
+      else
+      {
+         if (png_ptr->num_trans &&
+             (png_ptr->transformations & PNG_EXPAND_tRNS))
+            png_do_expand(&(png_ptr->row_info), png_ptr->row_buf + 1,
+               &(png_ptr->trans_values));
+         else
+            png_do_expand(&(png_ptr->row_info), png_ptr->row_buf + 1,
+               NULL);
+      }
+   }
+#endif
+
+#if defined(PNG_READ_STRIP_ALPHA_SUPPORTED)
+   if (png_ptr->flags & PNG_FLAG_STRIP_ALPHA)
+      png_do_strip_filler(&(png_ptr->row_info), png_ptr->row_buf + 1,
+         PNG_FLAG_FILLER_AFTER | (png_ptr->flags & PNG_FLAG_STRIP_ALPHA));
+#endif
+
+#if defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+   if (png_ptr->transformations & PNG_RGB_TO_GRAY)
+   {
+      int rgb_error =
+         png_do_rgb_to_gray(png_ptr, &(png_ptr->row_info), png_ptr->row_buf + 1);
+      if(rgb_error)
+      {
+         png_ptr->rgb_to_gray_status=1;
+         if((png_ptr->transformations & PNG_RGB_TO_GRAY) == 
+             PNG_RGB_TO_GRAY_WARN)
+            png_warning(png_ptr, "png_do_rgb_to_gray found nongray pixel");
+         if((png_ptr->transformations & PNG_RGB_TO_GRAY) ==
+             PNG_RGB_TO_GRAY_ERR)
+            png_error(png_ptr, "png_do_rgb_to_gray found nongray pixel");
+      }
+   }
+#endif
+
+/*
+From Andreas Dilger e-mail to png-implement, 26 March 1998:
+
+  In most cases, the "simple transparency" should be done prior to doing
+  gray-to-RGB, or you will have to test 3x as many bytes to check if a
+  pixel is transparent.  You would also need to make sure that the
+  transparency information is upgraded to RGB.
+
+  To summarize, the current flow is:
+  - Gray + simple transparency -> compare 1 or 2 gray bytes and composite
+                                  with background "in place" if transparent,
+                                  convert to RGB if necessary
+  - Gray + alpha -> composite with gray background and remove alpha bytes,
+                                  convert to RGB if necessary
+
+  To support RGB backgrounds for gray images we need:
+  - Gray + simple transparency -> convert to RGB + simple transparency, compare
+                                  3 or 6 bytes and composite with background
+                                  "in place" if transparent (3x compare/pixel
+                                  compared to doing composite with gray bkgrnd)
+  - Gray + alpha -> convert to RGB + alpha, composite with background and
+                                  remove alpha bytes (3x float operations/pixel
+                                  compared with composite on gray background)
+
+  Greg's change will do this.  The reason it wasn't done before is for
+  performance, as this increases the per-pixel operations.  If we would check
+  in advance if the background was gray or RGB, and position the gray-to-RGB
+  transform appropriately, then it would save a lot of work/time.
+ */
+
+#if defined(PNG_READ_GRAY_TO_RGB_SUPPORTED)
+   /* if gray -> RGB, do so now only if background is non-gray; else do later
+    * for performance reasons */
+   if ((png_ptr->transformations & PNG_GRAY_TO_RGB) &&
+       !(png_ptr->mode & PNG_BACKGROUND_IS_GRAY))
+      png_do_gray_to_rgb(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+   if ((png_ptr->transformations & PNG_BACKGROUND) &&
+      ((png_ptr->num_trans != 0 ) ||
+      (png_ptr->color_type & PNG_COLOR_MASK_ALPHA)))
+      png_do_background(&(png_ptr->row_info), png_ptr->row_buf + 1,
+         &(png_ptr->trans_values), &(png_ptr->background)
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+         , &(png_ptr->background_1),
+         png_ptr->gamma_table, png_ptr->gamma_from_1,
+         png_ptr->gamma_to_1, png_ptr->gamma_16_table,
+         png_ptr->gamma_16_from_1, png_ptr->gamma_16_to_1,
+         png_ptr->gamma_shift
+#endif
+);
+#endif
+
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+   if ((png_ptr->transformations & PNG_GAMMA) &&
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+      !((png_ptr->transformations & PNG_BACKGROUND) &&
+      ((png_ptr->num_trans != 0) ||
+      (png_ptr->color_type & PNG_COLOR_MASK_ALPHA))) &&
+#endif
+      (png_ptr->color_type != PNG_COLOR_TYPE_PALETTE))
+      png_do_gamma(&(png_ptr->row_info), png_ptr->row_buf + 1,
+         png_ptr->gamma_table, png_ptr->gamma_16_table,
+         png_ptr->gamma_shift);
+#endif
+
+#if defined(PNG_READ_16_TO_8_SUPPORTED)
+   if (png_ptr->transformations & PNG_16_TO_8)
+      png_do_chop(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_DITHER_SUPPORTED)
+   if (png_ptr->transformations & PNG_DITHER)
+   {
+      png_do_dither((png_row_infop)&(png_ptr->row_info), png_ptr->row_buf + 1,
+         png_ptr->palette_lookup, png_ptr->dither_index);
+      if(png_ptr->row_info.rowbytes == (png_uint_32)0)
+         png_error(png_ptr, "png_do_dither returned rowbytes=0");
+   }
+#endif
+
+#if defined(PNG_READ_INVERT_SUPPORTED)
+   if (png_ptr->transformations & PNG_INVERT_MONO)
+      png_do_invert(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_SHIFT_SUPPORTED)
+   if (png_ptr->transformations & PNG_SHIFT)
+      png_do_unshift(&(png_ptr->row_info), png_ptr->row_buf + 1,
+         &(png_ptr->shift));
+#endif
+
+#if defined(PNG_READ_PACK_SUPPORTED)
+   if (png_ptr->transformations & PNG_PACK)
+      png_do_unpack(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_BGR_SUPPORTED)
+   if (png_ptr->transformations & PNG_BGR)
+      png_do_bgr(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+   if (png_ptr->transformations & PNG_PACKSWAP)
+      png_do_packswap(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_GRAY_TO_RGB_SUPPORTED)
+   /* if gray -> RGB, do so now only if we did not do so above */
+   if ((png_ptr->transformations & PNG_GRAY_TO_RGB) &&
+       (png_ptr->mode & PNG_BACKGROUND_IS_GRAY))
+      png_do_gray_to_rgb(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_FILLER_SUPPORTED)
+   if (png_ptr->transformations & PNG_FILLER)
+      png_do_read_filler(&(png_ptr->row_info), png_ptr->row_buf + 1,
+         (png_uint_32)png_ptr->filler, png_ptr->flags);
+#endif
+
+#if defined(PNG_READ_INVERT_ALPHA_SUPPORTED)
+   if (png_ptr->transformations & PNG_INVERT_ALPHA)
+      png_do_read_invert_alpha(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_SWAP_ALPHA_SUPPORTED)
+   if (png_ptr->transformations & PNG_SWAP_ALPHA)
+      png_do_read_swap_alpha(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_SWAP_SUPPORTED)
+   if (png_ptr->transformations & PNG_SWAP_BYTES)
+      png_do_swap(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED)
+   if (png_ptr->transformations & PNG_USER_TRANSFORM)
+    {
+      if(png_ptr->read_user_transform_fn != NULL)
+        (*(png_ptr->read_user_transform_fn)) /* user read transform function */
+          (png_ptr,                    /* png_ptr */
+           &(png_ptr->row_info),       /* row_info:     */
+             /*  png_uint_32 width;          width of row */
+             /*  png_uint_32 rowbytes;       number of bytes in row */
+             /*  png_byte color_type;        color type of pixels */
+             /*  png_byte bit_depth;         bit depth of samples */
+             /*  png_byte channels;          number of channels (1-4) */
+             /*  png_byte pixel_depth;       bits per pixel (depth*channels) */
+           png_ptr->row_buf + 1);      /* start of pixel data for row */
+#if defined(PNG_USER_TRANSFORM_PTR_SUPPORTED)
+      if(png_ptr->user_transform_depth)
+         png_ptr->row_info.bit_depth = png_ptr->user_transform_depth;
+      if(png_ptr->user_transform_channels)
+         png_ptr->row_info.channels = png_ptr->user_transform_channels;
+#endif
+      png_ptr->row_info.pixel_depth = (png_byte)(png_ptr->row_info.bit_depth *
+         png_ptr->row_info.channels);
+      png_ptr->row_info.rowbytes = PNG_ROWBYTES(png_ptr->row_info.pixel_depth,
+         png_ptr->row_info.width);
+   }
+#endif
+
+}
+
+#if defined(PNG_READ_PACK_SUPPORTED)
+/* Unpack pixels of 1, 2, or 4 bits per pixel into 1 byte per pixel,
+ * without changing the actual values.  Thus, if you had a row with
+ * a bit depth of 1, you would end up with bytes that only contained
+ * the numbers 0 or 1.  If you would rather they contain 0 and 255, use
+ * png_do_shift() after this.
+ */
+void /* PRIVATE */
+png_do_unpack(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_unpack\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL && row_info->bit_depth < 8)
+#else
+   if (row_info->bit_depth < 8)
+#endif
+   {
+      png_uint_32 i;
+      png_uint_32 row_width=row_info->width;
+
+      switch (row_info->bit_depth)
+      {
+         case 1:
+         {
+            png_bytep sp = row + (png_size_t)((row_width - 1) >> 3);
+            png_bytep dp = row + (png_size_t)row_width - 1;
+            png_uint_32 shift = 7 - (int)((row_width + 7) & 0x07);
+            for (i = 0; i < row_width; i++)
+            {
+               *dp = (png_byte)((*sp >> shift) & 0x01);
+               if (shift == 7)
+               {
+                  shift = 0;
+                  sp--;
+               }
+               else
+                  shift++;
+
+               dp--;
+            }
+            break;
+         }
+         case 2:
+         {
+
+            png_bytep sp = row + (png_size_t)((row_width - 1) >> 2);
+            png_bytep dp = row + (png_size_t)row_width - 1;
+            png_uint_32 shift = (int)((3 - ((row_width + 3) & 0x03)) << 1);
+            for (i = 0; i < row_width; i++)
+            {
+               *dp = (png_byte)((*sp >> shift) & 0x03);
+               if (shift == 6)
+               {
+                  shift = 0;
+                  sp--;
+               }
+               else
+                  shift += 2;
+
+               dp--;
+            }
+            break;
+         }
+         case 4:
+         {
+            png_bytep sp = row + (png_size_t)((row_width - 1) >> 1);
+            png_bytep dp = row + (png_size_t)row_width - 1;
+            png_uint_32 shift = (int)((1 - ((row_width + 1) & 0x01)) << 2);
+            for (i = 0; i < row_width; i++)
+            {
+               *dp = (png_byte)((*sp >> shift) & 0x0f);
+               if (shift == 4)
+               {
+                  shift = 0;
+                  sp--;
+               }
+               else
+                  shift = 4;
+
+               dp--;
+            }
+            break;
+         }
+      }
+      row_info->bit_depth = 8;
+      row_info->pixel_depth = (png_byte)(8 * row_info->channels);
+      row_info->rowbytes = row_width * row_info->channels;
+   }
+}
+#endif
+
+#if defined(PNG_READ_SHIFT_SUPPORTED)
+/* Reverse the effects of png_do_shift.  This routine merely shifts the
+ * pixels back to their significant bits values.  Thus, if you have
+ * a row of bit depth 8, but only 5 are significant, this will shift
+ * the values back to 0 through 31.
+ */
+void /* PRIVATE */
+png_do_unshift(png_row_infop row_info, png_bytep row, png_color_8p sig_bits)
+{
+   png_debug(1, "in png_do_unshift\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL && sig_bits != NULL &&
+#endif
+       row_info->color_type != PNG_COLOR_TYPE_PALETTE)
+   {
+      int shift[4];
+      int channels = 0;
+      int c;
+      png_uint_16 value = 0;
+      png_uint_32 row_width = row_info->width;
+
+      if (row_info->color_type & PNG_COLOR_MASK_COLOR)
+      {
+         shift[channels++] = row_info->bit_depth - sig_bits->red;
+         shift[channels++] = row_info->bit_depth - sig_bits->green;
+         shift[channels++] = row_info->bit_depth - sig_bits->blue;
+      }
+      else
+      {
+         shift[channels++] = row_info->bit_depth - sig_bits->gray;
+      }
+      if (row_info->color_type & PNG_COLOR_MASK_ALPHA)
+      {
+         shift[channels++] = row_info->bit_depth - sig_bits->alpha;
+      }
+
+      for (c = 0; c < channels; c++)
+      {
+         if (shift[c] <= 0)
+            shift[c] = 0;
+         else
+            value = 1;
+      }
+
+      if (!value)
+         return;
+
+      switch (row_info->bit_depth)
+      {
+         case 2:
+         {
+            png_bytep bp;
+            png_uint_32 i;
+            png_uint_32 istop = row_info->rowbytes;
+
+            for (bp = row, i = 0; i < istop; i++)
+            {
+               *bp >>= 1;
+               *bp++ &= 0x55;
+            }
+            break;
+         }
+         case 4:
+         {
+            png_bytep bp = row;
+            png_uint_32 i;
+            png_uint_32 istop = row_info->rowbytes;
+            png_byte mask = (png_byte)((((int)0xf0 >> shift[0]) & (int)0xf0) |
+               (png_byte)((int)0xf >> shift[0]));
+
+            for (i = 0; i < istop; i++)
+            {
+               *bp >>= shift[0];
+               *bp++ &= mask;
+            }
+            break;
+         }
+         case 8:
+         {
+            png_bytep bp = row;
+            png_uint_32 i;
+            png_uint_32 istop = row_width * channels;
+
+            for (i = 0; i < istop; i++)
+            {
+               *bp++ >>= shift[i%channels];
+            }
+            break;
+         }
+         case 16:
+         {
+            png_bytep bp = row;
+            png_uint_32 i;
+            png_uint_32 istop = channels * row_width;
+
+            for (i = 0; i < istop; i++)
+            {
+               value = (png_uint_16)((*bp << 8) + *(bp + 1));
+               value >>= shift[i%channels];
+               *bp++ = (png_byte)(value >> 8);
+               *bp++ = (png_byte)(value & 0xff);
+            }
+            break;
+         }
+      }
+   }
+}
+#endif
+
+#if defined(PNG_READ_16_TO_8_SUPPORTED)
+/* chop rows of bit depth 16 down to 8 */
+void /* PRIVATE */
+png_do_chop(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_chop\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL && row_info->bit_depth == 16)
+#else
+   if (row_info->bit_depth == 16)
+#endif
+   {
+      png_bytep sp = row;
+      png_bytep dp = row;
+      png_uint_32 i;
+      png_uint_32 istop = row_info->width * row_info->channels;
+
+      for (i = 0; i<istop; i++, sp += 2, dp++)
+      {
+#if defined(PNG_READ_16_TO_8_ACCURATE_SCALE_SUPPORTED)
+      /* This does a more accurate scaling of the 16-bit color
+       * value, rather than a simple low-byte truncation.
+       *
+       * What the ideal calculation should be:
+       *   *dp = (((((png_uint_32)(*sp) << 8) |
+       *          (png_uint_32)(*(sp + 1))) * 255 + 127) / (png_uint_32)65535L;
+       *
+       * GRR: no, I think this is what it really should be:
+       *   *dp = (((((png_uint_32)(*sp) << 8) |
+       *           (png_uint_32)(*(sp + 1))) + 128L) / (png_uint_32)257L;
+       *
+       * GRR: here's the exact calculation with shifts:
+       *   temp = (((png_uint_32)(*sp) << 8) | (png_uint_32)(*(sp + 1))) + 128L;
+       *   *dp = (temp - (temp >> 8)) >> 8;
+       *
+       * Approximate calculation with shift/add instead of multiply/divide:
+       *   *dp = ((((png_uint_32)(*sp) << 8) |
+       *          (png_uint_32)((int)(*(sp + 1)) - *sp)) + 128) >> 8;
+       *
+       * What we actually do to avoid extra shifting and conversion:
+       */
+
+         *dp = *sp + ((((int)(*(sp + 1)) - *sp) > 128) ? 1 : 0);
+#else
+       /* Simply discard the low order byte */
+         *dp = *sp;
+#endif
+      }
+      row_info->bit_depth = 8;
+      row_info->pixel_depth = (png_byte)(8 * row_info->channels);
+      row_info->rowbytes = row_info->width * row_info->channels;
+   }
+}
+#endif
+
+#if defined(PNG_READ_SWAP_ALPHA_SUPPORTED)
+void /* PRIVATE */
+png_do_read_swap_alpha(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_read_swap_alpha\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL)
+#endif
+   {
+      png_uint_32 row_width = row_info->width;
+      if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+      {
+         /* This converts from RGBA to ARGB */
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp = row + row_info->rowbytes;
+            png_bytep dp = sp;
+            png_byte save;
+            png_uint_32 i;
+
+            for (i = 0; i < row_width; i++)
+            {
+               save = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = save;
+            }
+         }
+         /* This converts from RRGGBBAA to AARRGGBB */
+         else
+         {
+            png_bytep sp = row + row_info->rowbytes;
+            png_bytep dp = sp;
+            png_byte save[2];
+            png_uint_32 i;
+
+            for (i = 0; i < row_width; i++)
+            {
+               save[0] = *(--sp);
+               save[1] = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = save[0];
+               *(--dp) = save[1];
+            }
+         }
+      }
+      else if (row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
+      {
+         /* This converts from GA to AG */
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp = row + row_info->rowbytes;
+            png_bytep dp = sp;
+            png_byte save;
+            png_uint_32 i;
+
+            for (i = 0; i < row_width; i++)
+            {
+               save = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = save;
+            }
+         }
+         /* This converts from GGAA to AAGG */
+         else
+         {
+            png_bytep sp = row + row_info->rowbytes;
+            png_bytep dp = sp;
+            png_byte save[2];
+            png_uint_32 i;
+
+            for (i = 0; i < row_width; i++)
+            {
+               save[0] = *(--sp);
+               save[1] = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = save[0];
+               *(--dp) = save[1];
+            }
+         }
+      }
+   }
+}
+#endif
+
+#if defined(PNG_READ_INVERT_ALPHA_SUPPORTED)
+void /* PRIVATE */
+png_do_read_invert_alpha(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_read_invert_alpha\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL)
+#endif
+   {
+      png_uint_32 row_width = row_info->width;
+      if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+      {
+         /* This inverts the alpha channel in RGBA */
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp = row + row_info->rowbytes;
+            png_bytep dp = sp;
+            png_uint_32 i;
+
+            for (i = 0; i < row_width; i++)
+            {
+               *(--dp) = (png_byte)(255 - *(--sp));
+
+/*             This does nothing:
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               We can replace it with:
+*/
+               sp-=3;
+               dp=sp;
+            }
+         }
+         /* This inverts the alpha channel in RRGGBBAA */
+         else
+         {
+            png_bytep sp = row + row_info->rowbytes;
+            png_bytep dp = sp;
+            png_uint_32 i;
+
+            for (i = 0; i < row_width; i++)
+            {
+               *(--dp) = (png_byte)(255 - *(--sp));
+               *(--dp) = (png_byte)(255 - *(--sp));
+
+/*             This does nothing:
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               We can replace it with:
+*/
+               sp-=6;
+               dp=sp;
+            }
+         }
+      }
+      else if (row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
+      {
+         /* This inverts the alpha channel in GA */
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp = row + row_info->rowbytes;
+            png_bytep dp = sp;
+            png_uint_32 i;
+
+            for (i = 0; i < row_width; i++)
+            {
+               *(--dp) = (png_byte)(255 - *(--sp));
+               *(--dp) = *(--sp);
+            }
+         }
+         /* This inverts the alpha channel in GGAA */
+         else
+         {
+            png_bytep sp  = row + row_info->rowbytes;
+            png_bytep dp = sp;
+            png_uint_32 i;
+
+            for (i = 0; i < row_width; i++)
+            {
+               *(--dp) = (png_byte)(255 - *(--sp));
+               *(--dp) = (png_byte)(255 - *(--sp));
+/*
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+*/
+               sp-=2;
+               dp=sp;
+            }
+         }
+      }
+   }
+}
+#endif
+
+#if defined(PNG_READ_FILLER_SUPPORTED)
+/* Add filler channel if we have RGB color */
+void /* PRIVATE */
+png_do_read_filler(png_row_infop row_info, png_bytep row,
+   png_uint_32 filler, png_uint_32 flags)
+{
+   png_uint_32 i;
+   png_uint_32 row_width = row_info->width;
+
+   png_byte hi_filler = (png_byte)((filler>>8) & 0xff);
+   png_byte lo_filler = (png_byte)(filler & 0xff);
+
+   png_debug(1, "in png_do_read_filler\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL  && row_info != NULL &&
+#endif
+       row_info->color_type == PNG_COLOR_TYPE_GRAY)
+   {
+      if(row_info->bit_depth == 8)
+      {
+         /* This changes the data from G to GX */
+         if (flags & PNG_FLAG_FILLER_AFTER)
+         {
+            png_bytep sp = row + (png_size_t)row_width;
+            png_bytep dp =  sp + (png_size_t)row_width;
+            for (i = 1; i < row_width; i++)
+            {
+               *(--dp) = lo_filler;
+               *(--dp) = *(--sp);
+            }
+            *(--dp) = lo_filler;
+            row_info->channels = 2;
+            row_info->pixel_depth = 16;
+            row_info->rowbytes = row_width * 2;
+         }
+      /* This changes the data from G to XG */
+         else
+         {
+            png_bytep sp = row + (png_size_t)row_width;
+            png_bytep dp = sp  + (png_size_t)row_width;
+            for (i = 0; i < row_width; i++)
+            {
+               *(--dp) = *(--sp);
+               *(--dp) = lo_filler;
+            }
+            row_info->channels = 2;
+            row_info->pixel_depth = 16;
+            row_info->rowbytes = row_width * 2;
+         }
+      }
+      else if(row_info->bit_depth == 16)
+      {
+         /* This changes the data from GG to GGXX */
+         if (flags & PNG_FLAG_FILLER_AFTER)
+         {
+            png_bytep sp = row + (png_size_t)row_width * 2;
+            png_bytep dp = sp  + (png_size_t)row_width * 2;
+            for (i = 1; i < row_width; i++)
+            {
+               *(--dp) = hi_filler;
+               *(--dp) = lo_filler;
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+            }
+            *(--dp) = hi_filler;
+            *(--dp) = lo_filler;
+            row_info->channels = 2;
+            row_info->pixel_depth = 32;
+            row_info->rowbytes = row_width * 4;
+         }
+         /* This changes the data from GG to XXGG */
+         else
+         {
+            png_bytep sp = row + (png_size_t)row_width * 2;
+            png_bytep dp = sp  + (png_size_t)row_width * 2;
+            for (i = 0; i < row_width; i++)
+            {
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = hi_filler;
+               *(--dp) = lo_filler;
+            }
+            row_info->channels = 2;
+            row_info->pixel_depth = 32;
+            row_info->rowbytes = row_width * 4;
+         }
+      }
+   } /* COLOR_TYPE == GRAY */
+   else if (row_info->color_type == PNG_COLOR_TYPE_RGB)
+   {
+      if(row_info->bit_depth == 8)
+      {
+         /* This changes the data from RGB to RGBX */
+         if (flags & PNG_FLAG_FILLER_AFTER)
+         {
+            png_bytep sp = row + (png_size_t)row_width * 3;
+            png_bytep dp = sp  + (png_size_t)row_width;
+            for (i = 1; i < row_width; i++)
+            {
+               *(--dp) = lo_filler;
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+            }
+            *(--dp) = lo_filler;
+            row_info->channels = 4;
+            row_info->pixel_depth = 32;
+            row_info->rowbytes = row_width * 4;
+         }
+      /* This changes the data from RGB to XRGB */
+         else
+         {
+            png_bytep sp = row + (png_size_t)row_width * 3;
+            png_bytep dp = sp + (png_size_t)row_width;
+            for (i = 0; i < row_width; i++)
+            {
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = lo_filler;
+            }
+            row_info->channels = 4;
+            row_info->pixel_depth = 32;
+            row_info->rowbytes = row_width * 4;
+         }
+      }
+      else if(row_info->bit_depth == 16)
+      {
+         /* This changes the data from RRGGBB to RRGGBBXX */
+         if (flags & PNG_FLAG_FILLER_AFTER)
+         {
+            png_bytep sp = row + (png_size_t)row_width * 6;
+            png_bytep dp = sp  + (png_size_t)row_width * 2;
+            for (i = 1; i < row_width; i++)
+            {
+               *(--dp) = hi_filler;
+               *(--dp) = lo_filler;
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+            }
+            *(--dp) = hi_filler;
+            *(--dp) = lo_filler;
+            row_info->channels = 4;
+            row_info->pixel_depth = 64;
+            row_info->rowbytes = row_width * 8;
+         }
+         /* This changes the data from RRGGBB to XXRRGGBB */
+         else
+         {
+            png_bytep sp = row + (png_size_t)row_width * 6;
+            png_bytep dp = sp  + (png_size_t)row_width * 2;
+            for (i = 0; i < row_width; i++)
+            {
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = *(--sp);
+               *(--dp) = hi_filler;
+               *(--dp) = lo_filler;
+            }
+            row_info->channels = 4;
+            row_info->pixel_depth = 64;
+            row_info->rowbytes = row_width * 8;
+         }
+      }
+   } /* COLOR_TYPE == RGB */
+}
+#endif
+
+#if defined(PNG_READ_GRAY_TO_RGB_SUPPORTED)
+/* expand grayscale files to RGB, with or without alpha */
+void /* PRIVATE */
+png_do_gray_to_rgb(png_row_infop row_info, png_bytep row)
+{
+   png_uint_32 i;
+   png_uint_32 row_width = row_info->width;
+
+   png_debug(1, "in png_do_gray_to_rgb\n");
+   if (row_info->bit_depth >= 8 &&
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+      !(row_info->color_type & PNG_COLOR_MASK_COLOR))
+   {
+      if (row_info->color_type == PNG_COLOR_TYPE_GRAY)
+      {
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp = row + (png_size_t)row_width - 1;
+            png_bytep dp = sp  + (png_size_t)row_width * 2;
+            for (i = 0; i < row_width; i++)
+            {
+               *(dp--) = *sp;
+               *(dp--) = *sp;
+               *(dp--) = *(sp--);
+            }
+         }
+         else
+         {
+            png_bytep sp = row + (png_size_t)row_width * 2 - 1;
+            png_bytep dp = sp  + (png_size_t)row_width * 4;
+            for (i = 0; i < row_width; i++)
+            {
+               *(dp--) = *sp;
+               *(dp--) = *(sp - 1);
+               *(dp--) = *sp;
+               *(dp--) = *(sp - 1);
+               *(dp--) = *(sp--);
+               *(dp--) = *(sp--);
+            }
+         }
+      }
+      else if (row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
+      {
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp = row + (png_size_t)row_width * 2 - 1;
+            png_bytep dp = sp  + (png_size_t)row_width * 2;
+            for (i = 0; i < row_width; i++)
+            {
+               *(dp--) = *(sp--);
+               *(dp--) = *sp;
+               *(dp--) = *sp;
+               *(dp--) = *(sp--);
+            }
+         }
+         else
+         {
+            png_bytep sp = row + (png_size_t)row_width * 4 - 1;
+            png_bytep dp = sp  + (png_size_t)row_width * 4;
+            for (i = 0; i < row_width; i++)
+            {
+               *(dp--) = *(sp--);
+               *(dp--) = *(sp--);
+               *(dp--) = *sp;
+               *(dp--) = *(sp - 1);
+               *(dp--) = *sp;
+               *(dp--) = *(sp - 1);
+               *(dp--) = *(sp--);
+               *(dp--) = *(sp--);
+            }
+         }
+      }
+      row_info->channels += (png_byte)2;
+      row_info->color_type |= PNG_COLOR_MASK_COLOR;
+      row_info->pixel_depth = (png_byte)(row_info->channels *
+         row_info->bit_depth);
+      row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,row_width);
+   }
+}
+#endif
+
+#if defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+/* reduce RGB files to grayscale, with or without alpha
+ * using the equation given in Poynton's ColorFAQ at
+ * <http://www.inforamp.net/~poynton/>
+ * Copyright (c) 1998-01-04 Charles Poynton poynton at inforamp.net
+ *
+ *     Y = 0.212671 * R + 0.715160 * G + 0.072169 * B
+ *
+ *  We approximate this with
+ *
+ *     Y = 0.21268 * R    + 0.7151 * G    + 0.07217 * B
+ *
+ *  which can be expressed with integers as
+ *
+ *     Y = (6969 * R + 23434 * G + 2365 * B)/32768
+ *
+ *  The calculation is to be done in a linear colorspace.
+ *
+ *  Other integer coefficents can be used via png_set_rgb_to_gray().
+ */
+int /* PRIVATE */
+png_do_rgb_to_gray(png_structp png_ptr, png_row_infop row_info, png_bytep row)
+
+{
+   png_uint_32 i;
+
+   png_uint_32 row_width = row_info->width;
+   int rgb_error = 0;
+
+   png_debug(1, "in png_do_rgb_to_gray\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+      (row_info->color_type & PNG_COLOR_MASK_COLOR))
+   {
+      png_uint_32 rc = png_ptr->rgb_to_gray_red_coeff;
+      png_uint_32 gc = png_ptr->rgb_to_gray_green_coeff;
+      png_uint_32 bc = png_ptr->rgb_to_gray_blue_coeff;
+
+      if (row_info->color_type == PNG_COLOR_TYPE_RGB)
+      {
+         if (row_info->bit_depth == 8)
+         {
+#if defined(PNG_READ_GAMMA_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+            if (png_ptr->gamma_from_1 != NULL && png_ptr->gamma_to_1 != NULL)
+            {
+               png_bytep sp = row;
+               png_bytep dp = row;
+
+               for (i = 0; i < row_width; i++)
+               {
+                  png_byte red   = png_ptr->gamma_to_1[*(sp++)];
+                  png_byte green = png_ptr->gamma_to_1[*(sp++)];
+                  png_byte blue  = png_ptr->gamma_to_1[*(sp++)];
+                  if(red != green || red != blue)
+                  {
+                     rgb_error |= 1;
+                     *(dp++) = png_ptr->gamma_from_1[
+                       (rc*red+gc*green+bc*blue)>>15];
+                  }
+                  else
+                     *(dp++) = *(sp-1);
+               }
+            }
+            else
+#endif
+            {
+               png_bytep sp = row;
+               png_bytep dp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_byte red   = *(sp++);
+                  png_byte green = *(sp++);
+                  png_byte blue  = *(sp++);
+                  if(red != green || red != blue)
+                  {
+                     rgb_error |= 1;
+                     *(dp++) = (png_byte)((rc*red+gc*green+bc*blue)>>15);
+                  }
+                  else
+                     *(dp++) = *(sp-1);
+               }
+            }
+         }
+
+         else /* RGB bit_depth == 16 */
+         {
+#if defined(PNG_READ_GAMMA_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+            if (png_ptr->gamma_16_to_1 != NULL &&
+                png_ptr->gamma_16_from_1 != NULL)
+            {
+               png_bytep sp = row;
+               png_bytep dp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_uint_16 red, green, blue, w;
+
+                  red   = (png_uint_16)(((*(sp))<<8) | *(sp+1)); sp+=2;
+                  green = (png_uint_16)(((*(sp))<<8) | *(sp+1)); sp+=2;
+                  blue  = (png_uint_16)(((*(sp))<<8) | *(sp+1)); sp+=2;
+
+                  if(red == green && red == blue)
+                     w = red;
+                  else
+                  {
+                     png_uint_16 red_1   = png_ptr->gamma_16_to_1[(red&0xff) >>
+                                  png_ptr->gamma_shift][red>>8];
+                     png_uint_16 green_1 = png_ptr->gamma_16_to_1[(green&0xff) >>
+                                  png_ptr->gamma_shift][green>>8];
+                     png_uint_16 blue_1  = png_ptr->gamma_16_to_1[(blue&0xff) >>
+                                  png_ptr->gamma_shift][blue>>8];
+                     png_uint_16 gray16  = (png_uint_16)((rc*red_1 + gc*green_1
+                                  + bc*blue_1)>>15);
+                     w = png_ptr->gamma_16_from_1[(gray16&0xff) >>
+                         png_ptr->gamma_shift][gray16 >> 8];
+                     rgb_error |= 1;
+                  }
+
+                  *(dp++) = (png_byte)((w>>8) & 0xff);
+                  *(dp++) = (png_byte)(w & 0xff);
+               }
+            }
+            else
+#endif
+            {
+               png_bytep sp = row;
+               png_bytep dp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_uint_16 red, green, blue, gray16;
+
+                  red   = (png_uint_16)(((*(sp))<<8) | *(sp+1)); sp+=2;
+                  green = (png_uint_16)(((*(sp))<<8) | *(sp+1)); sp+=2;
+                  blue  = (png_uint_16)(((*(sp))<<8) | *(sp+1)); sp+=2;
+
+                  if(red != green || red != blue)
+                     rgb_error |= 1;
+                  gray16  = (png_uint_16)((rc*red + gc*green + bc*blue)>>15);
+                  *(dp++) = (png_byte)((gray16>>8) & 0xff);
+                  *(dp++) = (png_byte)(gray16 & 0xff);
+               }
+            }
+         }
+      }
+      if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+      {
+         if (row_info->bit_depth == 8)
+         {
+#if defined(PNG_READ_GAMMA_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+            if (png_ptr->gamma_from_1 != NULL && png_ptr->gamma_to_1 != NULL)
+            {
+               png_bytep sp = row;
+               png_bytep dp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_byte red   = png_ptr->gamma_to_1[*(sp++)];
+                  png_byte green = png_ptr->gamma_to_1[*(sp++)];
+                  png_byte blue  = png_ptr->gamma_to_1[*(sp++)];
+                  if(red != green || red != blue)
+                     rgb_error |= 1;
+                  *(dp++) =  png_ptr->gamma_from_1
+                             [(rc*red + gc*green + bc*blue)>>15];
+                  *(dp++) = *(sp++);  /* alpha */
+               }
+            }
+            else
+#endif
+            {
+               png_bytep sp = row;
+               png_bytep dp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_byte red   = *(sp++);
+                  png_byte green = *(sp++);
+                  png_byte blue  = *(sp++);
+                  if(red != green || red != blue)
+                     rgb_error |= 1;
+                  *(dp++) =  (png_byte)((rc*red + gc*green + bc*blue)>>15);
+                  *(dp++) = *(sp++);  /* alpha */
+               }
+            }
+         }
+         else /* RGBA bit_depth == 16 */
+         {
+#if defined(PNG_READ_GAMMA_SUPPORTED) || defined(PNG_READ_BACKGROUND_SUPPORTED)
+            if (png_ptr->gamma_16_to_1 != NULL &&
+                png_ptr->gamma_16_from_1 != NULL)
+            {
+               png_bytep sp = row;
+               png_bytep dp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_uint_16 red, green, blue, w;
+
+                  red   = (png_uint_16)(((*(sp))<<8) | *(sp+1)); sp+=2;
+                  green = (png_uint_16)(((*(sp))<<8) | *(sp+1)); sp+=2;
+                  blue  = (png_uint_16)(((*(sp))<<8) | *(sp+1)); sp+=2;
+
+                  if(red == green && red == blue)
+                     w = red;
+                  else
+                  {
+                     png_uint_16 red_1   = png_ptr->gamma_16_to_1[(red&0xff) >>
+                                  png_ptr->gamma_shift][red>>8];
+                     png_uint_16 green_1 = png_ptr->gamma_16_to_1[(green&0xff) >>
+                                  png_ptr->gamma_shift][green>>8];
+                     png_uint_16 blue_1  = png_ptr->gamma_16_to_1[(blue&0xff) >>
+                                  png_ptr->gamma_shift][blue>>8];
+                     png_uint_16 gray16  = (png_uint_16)((rc * red_1
+                                  + gc * green_1 + bc * blue_1)>>15);
+                     w = png_ptr->gamma_16_from_1[(gray16&0xff) >>
+                         png_ptr->gamma_shift][gray16 >> 8];
+                     rgb_error |= 1;
+                  }
+
+                  *(dp++) = (png_byte)((w>>8) & 0xff);
+                  *(dp++) = (png_byte)(w & 0xff);
+                  *(dp++) = *(sp++);  /* alpha */
+                  *(dp++) = *(sp++);
+               }
+            }
+            else
+#endif
+            {
+               png_bytep sp = row;
+               png_bytep dp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_uint_16 red, green, blue, gray16;
+                  red   = (png_uint_16)((*(sp)<<8) | *(sp+1)); sp+=2;
+                  green = (png_uint_16)((*(sp)<<8) | *(sp+1)); sp+=2;
+                  blue  = (png_uint_16)((*(sp)<<8) | *(sp+1)); sp+=2;
+                  if(red != green || red != blue)
+                     rgb_error |= 1;
+                  gray16  = (png_uint_16)((rc*red + gc*green + bc*blue)>>15);
+                  *(dp++) = (png_byte)((gray16>>8) & 0xff);
+                  *(dp++) = (png_byte)(gray16 & 0xff);
+                  *(dp++) = *(sp++);  /* alpha */
+                  *(dp++) = *(sp++);
+               }
+            }
+         }
+      }
+   row_info->channels -= (png_byte)2;
+      row_info->color_type &= ~PNG_COLOR_MASK_COLOR;
+      row_info->pixel_depth = (png_byte)(row_info->channels *
+         row_info->bit_depth);
+      row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,row_width);
+   }
+   return rgb_error;
+}
+#endif
+
+/* Build a grayscale palette.  Palette is assumed to be 1 << bit_depth
+ * large of png_color.  This lets grayscale images be treated as
+ * paletted.  Most useful for gamma correction and simplification
+ * of code.
+ */
+void PNGAPI
+png_build_grayscale_palette(int bit_depth, png_colorp palette)
+{
+   int num_palette;
+   int color_inc;
+   int i;
+   int v;
+
+   png_debug(1, "in png_do_build_grayscale_palette\n");
+   if (palette == NULL)
+      return;
+
+   switch (bit_depth)
+   {
+      case 1:
+         num_palette = 2;
+         color_inc = 0xff;
+         break;
+      case 2:
+         num_palette = 4;
+         color_inc = 0x55;
+         break;
+      case 4:
+         num_palette = 16;
+         color_inc = 0x11;
+         break;
+      case 8:
+         num_palette = 256;
+         color_inc = 1;
+         break;
+      default:
+         num_palette = 0;
+         color_inc = 0;
+         break;
+   }
+
+   for (i = 0, v = 0; i < num_palette; i++, v += color_inc)
+   {
+      palette[i].red = (png_byte)v;
+      palette[i].green = (png_byte)v;
+      palette[i].blue = (png_byte)v;
+   }
+}
+
+/* This function is currently unused.  Do we really need it? */
+#if defined(PNG_READ_DITHER_SUPPORTED) && defined(PNG_CORRECT_PALETTE_SUPPORTED)
+void /* PRIVATE */
+png_correct_palette(png_structp png_ptr, png_colorp palette,
+   int num_palette)
+{
+   png_debug(1, "in png_correct_palette\n");
+#if defined(PNG_READ_BACKGROUND_SUPPORTED) && \
+    defined(PNG_READ_GAMMA_SUPPORTED) && defined(PNG_FLOATING_POINT_SUPPORTED)
+   if (png_ptr->transformations & (PNG_GAMMA | PNG_BACKGROUND))
+   {
+      png_color back, back_1;
+
+      if (png_ptr->background_gamma_type == PNG_BACKGROUND_GAMMA_FILE)
+      {
+         back.red = png_ptr->gamma_table[png_ptr->background.red];
+         back.green = png_ptr->gamma_table[png_ptr->background.green];
+         back.blue = png_ptr->gamma_table[png_ptr->background.blue];
+
+         back_1.red = png_ptr->gamma_to_1[png_ptr->background.red];
+         back_1.green = png_ptr->gamma_to_1[png_ptr->background.green];
+         back_1.blue = png_ptr->gamma_to_1[png_ptr->background.blue];
+      }
+      else
+      {
+         double g;
+
+         g = 1.0 / (png_ptr->background_gamma * png_ptr->screen_gamma);
+
+         if (png_ptr->background_gamma_type == PNG_BACKGROUND_GAMMA_SCREEN ||
+             fabs(g - 1.0) < PNG_GAMMA_THRESHOLD)
+         {
+            back.red = png_ptr->background.red;
+            back.green = png_ptr->background.green;
+            back.blue = png_ptr->background.blue;
+         }
+         else
+         {
+            back.red =
+               (png_byte)(pow((double)png_ptr->background.red/255, g) *
+                255.0 + 0.5);
+            back.green =
+               (png_byte)(pow((double)png_ptr->background.green/255, g) *
+                255.0 + 0.5);
+            back.blue =
+               (png_byte)(pow((double)png_ptr->background.blue/255, g) *
+                255.0 + 0.5);
+         }
+
+         g = 1.0 / png_ptr->background_gamma;
+
+         back_1.red =
+            (png_byte)(pow((double)png_ptr->background.red/255, g) *
+             255.0 + 0.5);
+         back_1.green =
+            (png_byte)(pow((double)png_ptr->background.green/255, g) *
+             255.0 + 0.5);
+         back_1.blue =
+            (png_byte)(pow((double)png_ptr->background.blue/255, g) *
+             255.0 + 0.5);
+      }
+
+      if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      {
+         png_uint_32 i;
+
+         for (i = 0; i < (png_uint_32)num_palette; i++)
+         {
+            if (i < png_ptr->num_trans && png_ptr->trans[i] == 0)
+            {
+               palette[i] = back;
+            }
+            else if (i < png_ptr->num_trans && png_ptr->trans[i] != 0xff)
+            {
+               png_byte v, w;
+
+               v = png_ptr->gamma_to_1[png_ptr->palette[i].red];
+               png_composite(w, v, png_ptr->trans[i], back_1.red);
+               palette[i].red = png_ptr->gamma_from_1[w];
+
+               v = png_ptr->gamma_to_1[png_ptr->palette[i].green];
+               png_composite(w, v, png_ptr->trans[i], back_1.green);
+               palette[i].green = png_ptr->gamma_from_1[w];
+
+               v = png_ptr->gamma_to_1[png_ptr->palette[i].blue];
+               png_composite(w, v, png_ptr->trans[i], back_1.blue);
+               palette[i].blue = png_ptr->gamma_from_1[w];
+            }
+            else
+            {
+               palette[i].red = png_ptr->gamma_table[palette[i].red];
+               palette[i].green = png_ptr->gamma_table[palette[i].green];
+               palette[i].blue = png_ptr->gamma_table[palette[i].blue];
+            }
+         }
+      }
+      else
+      {
+         int i;
+
+         for (i = 0; i < num_palette; i++)
+         {
+            if (palette[i].red == (png_byte)png_ptr->trans_values.gray)
+            {
+               palette[i] = back;
+            }
+            else
+            {
+               palette[i].red = png_ptr->gamma_table[palette[i].red];
+               palette[i].green = png_ptr->gamma_table[palette[i].green];
+               palette[i].blue = png_ptr->gamma_table[palette[i].blue];
+            }
+         }
+      }
+   }
+   else
+#endif
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+   if (png_ptr->transformations & PNG_GAMMA)
+   {
+      int i;
+
+      for (i = 0; i < num_palette; i++)
+      {
+         palette[i].red = png_ptr->gamma_table[palette[i].red];
+         palette[i].green = png_ptr->gamma_table[palette[i].green];
+         palette[i].blue = png_ptr->gamma_table[palette[i].blue];
+      }
+   }
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+   else
+#endif
+#endif
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+   if (png_ptr->transformations & PNG_BACKGROUND)
+   {
+      if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      {
+         png_color back;
+
+         back.red   = (png_byte)png_ptr->background.red;
+         back.green = (png_byte)png_ptr->background.green;
+         back.blue  = (png_byte)png_ptr->background.blue;
+
+         for (i = 0; i < (int)png_ptr->num_trans; i++)
+         {
+            if (png_ptr->trans[i] == 0)
+            {
+               palette[i].red = back.red;
+               palette[i].green = back.green;
+               palette[i].blue = back.blue;
+            }
+            else if (png_ptr->trans[i] != 0xff)
+            {
+               png_composite(palette[i].red, png_ptr->palette[i].red,
+                  png_ptr->trans[i], back.red);
+               png_composite(palette[i].green, png_ptr->palette[i].green,
+                  png_ptr->trans[i], back.green);
+               png_composite(palette[i].blue, png_ptr->palette[i].blue,
+                  png_ptr->trans[i], back.blue);
+            }
+         }
+      }
+      else /* assume grayscale palette (what else could it be?) */
+      {
+         int i;
+
+         for (i = 0; i < num_palette; i++)
+         {
+            if (i == (png_byte)png_ptr->trans_values.gray)
+            {
+               palette[i].red = (png_byte)png_ptr->background.red;
+               palette[i].green = (png_byte)png_ptr->background.green;
+               palette[i].blue = (png_byte)png_ptr->background.blue;
+            }
+         }
+      }
+   }
+#endif
+}
+#endif
+
+#if defined(PNG_READ_BACKGROUND_SUPPORTED)
+/* Replace any alpha or transparency with the supplied background color.
+ * "background" is already in the screen gamma, while "background_1" is
+ * at a gamma of 1.0.  Paletted files have already been taken care of.
+ */
+void /* PRIVATE */
+png_do_background(png_row_infop row_info, png_bytep row,
+   png_color_16p trans_values, png_color_16p background
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+   , png_color_16p background_1,
+   png_bytep gamma_table, png_bytep gamma_from_1, png_bytep gamma_to_1,
+   png_uint_16pp gamma_16, png_uint_16pp gamma_16_from_1,
+   png_uint_16pp gamma_16_to_1, int gamma_shift
+#endif
+   )
+{
+   png_bytep sp, dp;
+   png_uint_32 i;
+   png_uint_32 row_width=row_info->width;
+   int shift;
+
+   png_debug(1, "in png_do_background\n");
+   if (background != NULL &&
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+      (!(row_info->color_type & PNG_COLOR_MASK_ALPHA) ||
+      (row_info->color_type != PNG_COLOR_TYPE_PALETTE && trans_values)))
+   {
+      switch (row_info->color_type)
+      {
+         case PNG_COLOR_TYPE_GRAY:
+         {
+            switch (row_info->bit_depth)
+            {
+               case 1:
+               {
+                  sp = row;
+                  shift = 7;
+                  for (i = 0; i < row_width; i++)
+                  {
+                     if ((png_uint_16)((*sp >> shift) & 0x01)
+                        == trans_values->gray)
+                     {
+                        *sp &= (png_byte)((0x7f7f >> (7 - shift)) & 0xff);
+                        *sp |= (png_byte)(background->gray << shift);
+                     }
+                     if (!shift)
+                     {
+                        shift = 7;
+                        sp++;
+                     }
+                     else
+                        shift--;
+                  }
+                  break;
+               }
+               case 2:
+               {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+                  if (gamma_table != NULL)
+                  {
+                     sp = row;
+                     shift = 6;
+                     for (i = 0; i < row_width; i++)
+                     {
+                        if ((png_uint_16)((*sp >> shift) & 0x03)
+                            == trans_values->gray)
+                        {
+                           *sp &= (png_byte)((0x3f3f >> (6 - shift)) & 0xff);
+                           *sp |= (png_byte)(background->gray << shift);
+                        }
+                        else
+                        {
+                           png_byte p = (png_byte)((*sp >> shift) & 0x03);
+                           png_byte g = (png_byte)((gamma_table [p | (p << 2) |
+                               (p << 4) | (p << 6)] >> 6) & 0x03);
+                           *sp &= (png_byte)((0x3f3f >> (6 - shift)) & 0xff);
+                           *sp |= (png_byte)(g << shift);
+                        }
+                        if (!shift)
+                        {
+                           shift = 6;
+                           sp++;
+                        }
+                        else
+                           shift -= 2;
+                     }
+                  }
+                  else
+#endif
+                  {
+                     sp = row;
+                     shift = 6;
+                     for (i = 0; i < row_width; i++)
+                     {
+                        if ((png_uint_16)((*sp >> shift) & 0x03)
+                            == trans_values->gray)
+                        {
+                           *sp &= (png_byte)((0x3f3f >> (6 - shift)) & 0xff);
+                           *sp |= (png_byte)(background->gray << shift);
+                        }
+                        if (!shift)
+                        {
+                           shift = 6;
+                           sp++;
+                        }
+                        else
+                           shift -= 2;
+                     }
+                  }
+                  break;
+               }
+               case 4:
+               {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+                  if (gamma_table != NULL)
+                  {
+                     sp = row;
+                     shift = 4;
+                     for (i = 0; i < row_width; i++)
+                     {
+                        if ((png_uint_16)((*sp >> shift) & 0x0f)
+                            == trans_values->gray)
+                        {
+                           *sp &= (png_byte)((0xf0f >> (4 - shift)) & 0xff);
+                           *sp |= (png_byte)(background->gray << shift);
+                        }
+                        else
+                        {
+                           png_byte p = (png_byte)((*sp >> shift) & 0x0f);
+                           png_byte g = (png_byte)((gamma_table[p |
+                             (p << 4)] >> 4) & 0x0f);
+                           *sp &= (png_byte)((0xf0f >> (4 - shift)) & 0xff);
+                           *sp |= (png_byte)(g << shift);
+                        }
+                        if (!shift)
+                        {
+                           shift = 4;
+                           sp++;
+                        }
+                        else
+                           shift -= 4;
+                     }
+                  }
+                  else
+#endif
+                  {
+                     sp = row;
+                     shift = 4;
+                     for (i = 0; i < row_width; i++)
+                     {
+                        if ((png_uint_16)((*sp >> shift) & 0x0f)
+                            == trans_values->gray)
+                        {
+                           *sp &= (png_byte)((0xf0f >> (4 - shift)) & 0xff);
+                           *sp |= (png_byte)(background->gray << shift);
+                        }
+                        if (!shift)
+                        {
+                           shift = 4;
+                           sp++;
+                        }
+                        else
+                           shift -= 4;
+                     }
+                  }
+                  break;
+               }
+               case 8:
+               {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+                  if (gamma_table != NULL)
+                  {
+                     sp = row;
+                     for (i = 0; i < row_width; i++, sp++)
+                     {
+                        if (*sp == trans_values->gray)
+                        {
+                           *sp = (png_byte)background->gray;
+                        }
+                        else
+                        {
+                           *sp = gamma_table[*sp];
+                        }
+                     }
+                  }
+                  else
+#endif
+                  {
+                     sp = row;
+                     for (i = 0; i < row_width; i++, sp++)
+                     {
+                        if (*sp == trans_values->gray)
+                        {
+                           *sp = (png_byte)background->gray;
+                        }
+                     }
+                  }
+                  break;
+               }
+               case 16:
+               {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+                  if (gamma_16 != NULL)
+                  {
+                     sp = row;
+                     for (i = 0; i < row_width; i++, sp += 2)
+                     {
+                        png_uint_16 v;
+
+                        v = (png_uint_16)(((*sp) << 8) + *(sp + 1));
+                        if (v == trans_values->gray)
+                        {
+                           /* background is already in screen gamma */
+                           *sp = (png_byte)((background->gray >> 8) & 0xff);
+                           *(sp + 1) = (png_byte)(background->gray & 0xff);
+                        }
+                        else
+                        {
+                           v = gamma_16[*(sp + 1) >> gamma_shift][*sp];
+                           *sp = (png_byte)((v >> 8) & 0xff);
+                           *(sp + 1) = (png_byte)(v & 0xff);
+                        }
+                     }
+                  }
+                  else
+#endif
+                  {
+                     sp = row;
+                     for (i = 0; i < row_width; i++, sp += 2)
+                     {
+                        png_uint_16 v;
+
+                        v = (png_uint_16)(((*sp) << 8) + *(sp + 1));
+                        if (v == trans_values->gray)
+                        {
+                           *sp = (png_byte)((background->gray >> 8) & 0xff);
+                           *(sp + 1) = (png_byte)(background->gray & 0xff);
+                        }
+                     }
+                  }
+                  break;
+               }
+            }
+            break;
+         }
+         case PNG_COLOR_TYPE_RGB:
+         {
+            if (row_info->bit_depth == 8)
+            {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+               if (gamma_table != NULL)
+               {
+                  sp = row;
+                  for (i = 0; i < row_width; i++, sp += 3)
+                  {
+                     if (*sp == trans_values->red &&
+                        *(sp + 1) == trans_values->green &&
+                        *(sp + 2) == trans_values->blue)
+                     {
+                        *sp = (png_byte)background->red;
+                        *(sp + 1) = (png_byte)background->green;
+                        *(sp + 2) = (png_byte)background->blue;
+                     }
+                     else
+                     {
+                        *sp = gamma_table[*sp];
+                        *(sp + 1) = gamma_table[*(sp + 1)];
+                        *(sp + 2) = gamma_table[*(sp + 2)];
+                     }
+                  }
+               }
+               else
+#endif
+               {
+                  sp = row;
+                  for (i = 0; i < row_width; i++, sp += 3)
+                  {
+                     if (*sp == trans_values->red &&
+                        *(sp + 1) == trans_values->green &&
+                        *(sp + 2) == trans_values->blue)
+                     {
+                        *sp = (png_byte)background->red;
+                        *(sp + 1) = (png_byte)background->green;
+                        *(sp + 2) = (png_byte)background->blue;
+                     }
+                  }
+               }
+            }
+            else /* if (row_info->bit_depth == 16) */
+            {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+               if (gamma_16 != NULL)
+               {
+                  sp = row;
+                  for (i = 0; i < row_width; i++, sp += 6)
+                  {
+                     png_uint_16 r = (png_uint_16)(((*sp) << 8) + *(sp + 1));
+                     png_uint_16 g = (png_uint_16)(((*(sp+2)) << 8) + *(sp+3));
+                     png_uint_16 b = (png_uint_16)(((*(sp+4)) << 8) + *(sp+5));
+                     if (r == trans_values->red && g == trans_values->green &&
+                        b == trans_values->blue)
+                     {
+                        /* background is already in screen gamma */
+                        *sp = (png_byte)((background->red >> 8) & 0xff);
+                        *(sp + 1) = (png_byte)(background->red & 0xff);
+                        *(sp + 2) = (png_byte)((background->green >> 8) & 0xff);
+                        *(sp + 3) = (png_byte)(background->green & 0xff);
+                        *(sp + 4) = (png_byte)((background->blue >> 8) & 0xff);
+                        *(sp + 5) = (png_byte)(background->blue & 0xff);
+                     }
+                     else
+                     {
+                        png_uint_16 v = gamma_16[*(sp + 1) >> gamma_shift][*sp];
+                        *sp = (png_byte)((v >> 8) & 0xff);
+                        *(sp + 1) = (png_byte)(v & 0xff);
+                        v = gamma_16[*(sp + 3) >> gamma_shift][*(sp + 2)];
+                        *(sp + 2) = (png_byte)((v >> 8) & 0xff);
+                        *(sp + 3) = (png_byte)(v & 0xff);
+                        v = gamma_16[*(sp + 5) >> gamma_shift][*(sp + 4)];
+                        *(sp + 4) = (png_byte)((v >> 8) & 0xff);
+                        *(sp + 5) = (png_byte)(v & 0xff);
+                     }
+                  }
+               }
+               else
+#endif
+               {
+                  sp = row;
+                  for (i = 0; i < row_width; i++, sp += 6)
+                  {
+                     png_uint_16 r = (png_uint_16)(((*sp) << 8) + *(sp+1));
+                     png_uint_16 g = (png_uint_16)(((*(sp+2)) << 8) + *(sp+3));
+                     png_uint_16 b = (png_uint_16)(((*(sp+4)) << 8) + *(sp+5));
+
+                     if (r == trans_values->red && g == trans_values->green &&
+                        b == trans_values->blue)
+                     {
+                        *sp = (png_byte)((background->red >> 8) & 0xff);
+                        *(sp + 1) = (png_byte)(background->red & 0xff);
+                        *(sp + 2) = (png_byte)((background->green >> 8) & 0xff);
+                        *(sp + 3) = (png_byte)(background->green & 0xff);
+                        *(sp + 4) = (png_byte)((background->blue >> 8) & 0xff);
+                        *(sp + 5) = (png_byte)(background->blue & 0xff);
+                     }
+                  }
+               }
+            }
+            break;
+         }
+         case PNG_COLOR_TYPE_GRAY_ALPHA:
+         {
+            if (row_info->bit_depth == 8)
+            {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+               if (gamma_to_1 != NULL && gamma_from_1 != NULL &&
+                   gamma_table != NULL)
+               {
+                  sp = row;
+                  dp = row;
+                  for (i = 0; i < row_width; i++, sp += 2, dp++)
+                  {
+                     png_uint_16 a = *(sp + 1);
+
+                     if (a == 0xff)
+                     {
+                        *dp = gamma_table[*sp];
+                     }
+                     else if (a == 0)
+                     {
+                        /* background is already in screen gamma */
+                        *dp = (png_byte)background->gray;
+                     }
+                     else
+                     {
+                        png_byte v, w;
+
+                        v = gamma_to_1[*sp];
+                        png_composite(w, v, a, background_1->gray);
+                        *dp = gamma_from_1[w];
+                     }
+                  }
+               }
+               else
+#endif
+               {
+                  sp = row;
+                  dp = row;
+                  for (i = 0; i < row_width; i++, sp += 2, dp++)
+                  {
+                     png_byte a = *(sp + 1);
+
+                     if (a == 0xff)
+                     {
+                        *dp = *sp;
+                     }
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+                     else if (a == 0)
+                     {
+                        *dp = (png_byte)background->gray;
+                     }
+                     else
+                     {
+                        png_composite(*dp, *sp, a, background_1->gray);
+                     }
+#else
+                     *dp = (png_byte)background->gray;
+#endif
+                  }
+               }
+            }
+            else /* if (png_ptr->bit_depth == 16) */
+            {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+               if (gamma_16 != NULL && gamma_16_from_1 != NULL &&
+                   gamma_16_to_1 != NULL)
+               {
+                  sp = row;
+                  dp = row;
+                  for (i = 0; i < row_width; i++, sp += 4, dp += 2)
+                  {
+                     png_uint_16 a = (png_uint_16)(((*(sp+2)) << 8) + *(sp+3));
+
+                     if (a == (png_uint_16)0xffff)
+                     {
+                        png_uint_16 v;
+
+                        v = gamma_16[*(sp + 1) >> gamma_shift][*sp];
+                        *dp = (png_byte)((v >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(v & 0xff);
+                     }
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+                     else if (a == 0)
+#else
+                     else
+#endif
+                     {
+                        /* background is already in screen gamma */
+                        *dp = (png_byte)((background->gray >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(background->gray & 0xff);
+                     }
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+                     else
+                     {
+                        png_uint_16 g, v, w;
+
+                        g = gamma_16_to_1[*(sp + 1) >> gamma_shift][*sp];
+                        png_composite_16(v, g, a, background_1->gray);
+                        w = gamma_16_from_1[(v&0xff) >> gamma_shift][v >> 8];
+                        *dp = (png_byte)((w >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(w & 0xff);
+                     }
+#endif
+                  }
+               }
+               else
+#endif
+               {
+                  sp = row;
+                  dp = row;
+                  for (i = 0; i < row_width; i++, sp += 4, dp += 2)
+                  {
+                     png_uint_16 a = (png_uint_16)(((*(sp+2)) << 8) + *(sp+3));
+                     if (a == (png_uint_16)0xffff)
+                     {
+                        png_memcpy(dp, sp, 2);
+                     }
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+                     else if (a == 0)
+#else
+                     else
+#endif
+                     {
+                        *dp = (png_byte)((background->gray >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(background->gray & 0xff);
+                     }
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+                     else
+                     {
+                        png_uint_16 g, v;
+
+                        g = (png_uint_16)(((*sp) << 8) + *(sp + 1));
+                        png_composite_16(v, g, a, background_1->gray);
+                        *dp = (png_byte)((v >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(v & 0xff);
+                     }
+#endif
+                  }
+               }
+            }
+            break;
+         }
+         case PNG_COLOR_TYPE_RGB_ALPHA:
+         {
+            if (row_info->bit_depth == 8)
+            {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+               if (gamma_to_1 != NULL && gamma_from_1 != NULL &&
+                   gamma_table != NULL)
+               {
+                  sp = row;
+                  dp = row;
+                  for (i = 0; i < row_width; i++, sp += 4, dp += 3)
+                  {
+                     png_byte a = *(sp + 3);
+
+                     if (a == 0xff)
+                     {
+                        *dp = gamma_table[*sp];
+                        *(dp + 1) = gamma_table[*(sp + 1)];
+                        *(dp + 2) = gamma_table[*(sp + 2)];
+                     }
+                     else if (a == 0)
+                     {
+                        /* background is already in screen gamma */
+                        *dp = (png_byte)background->red;
+                        *(dp + 1) = (png_byte)background->green;
+                        *(dp + 2) = (png_byte)background->blue;
+                     }
+                     else
+                     {
+                        png_byte v, w;
+
+                        v = gamma_to_1[*sp];
+                        png_composite(w, v, a, background_1->red);
+                        *dp = gamma_from_1[w];
+                        v = gamma_to_1[*(sp + 1)];
+                        png_composite(w, v, a, background_1->green);
+                        *(dp + 1) = gamma_from_1[w];
+                        v = gamma_to_1[*(sp + 2)];
+                        png_composite(w, v, a, background_1->blue);
+                        *(dp + 2) = gamma_from_1[w];
+                     }
+                  }
+               }
+               else
+#endif
+               {
+                  sp = row;
+                  dp = row;
+                  for (i = 0; i < row_width; i++, sp += 4, dp += 3)
+                  {
+                     png_byte a = *(sp + 3);
+
+                     if (a == 0xff)
+                     {
+                        *dp = *sp;
+                        *(dp + 1) = *(sp + 1);
+                        *(dp + 2) = *(sp + 2);
+                     }
+                     else if (a == 0)
+                     {
+                        *dp = (png_byte)background->red;
+                        *(dp + 1) = (png_byte)background->green;
+                        *(dp + 2) = (png_byte)background->blue;
+                     }
+                     else
+                     {
+                        png_composite(*dp, *sp, a, background->red);
+                        png_composite(*(dp + 1), *(sp + 1), a,
+                           background->green);
+                        png_composite(*(dp + 2), *(sp + 2), a,
+                           background->blue);
+                     }
+                  }
+               }
+            }
+            else /* if (row_info->bit_depth == 16) */
+            {
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+               if (gamma_16 != NULL && gamma_16_from_1 != NULL &&
+                   gamma_16_to_1 != NULL)
+               {
+                  sp = row;
+                  dp = row;
+                  for (i = 0; i < row_width; i++, sp += 8, dp += 6)
+                  {
+                     png_uint_16 a = (png_uint_16)(((png_uint_16)(*(sp + 6))
+                         << 8) + (png_uint_16)(*(sp + 7)));
+                     if (a == (png_uint_16)0xffff)
+                     {
+                        png_uint_16 v;
+
+                        v = gamma_16[*(sp + 1) >> gamma_shift][*sp];
+                        *dp = (png_byte)((v >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(v & 0xff);
+                        v = gamma_16[*(sp + 3) >> gamma_shift][*(sp + 2)];
+                        *(dp + 2) = (png_byte)((v >> 8) & 0xff);
+                        *(dp + 3) = (png_byte)(v & 0xff);
+                        v = gamma_16[*(sp + 5) >> gamma_shift][*(sp + 4)];
+                        *(dp + 4) = (png_byte)((v >> 8) & 0xff);
+                        *(dp + 5) = (png_byte)(v & 0xff);
+                     }
+                     else if (a == 0)
+                     {
+                        /* background is already in screen gamma */
+                        *dp = (png_byte)((background->red >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(background->red & 0xff);
+                        *(dp + 2) = (png_byte)((background->green >> 8) & 0xff);
+                        *(dp + 3) = (png_byte)(background->green & 0xff);
+                        *(dp + 4) = (png_byte)((background->blue >> 8) & 0xff);
+                        *(dp + 5) = (png_byte)(background->blue & 0xff);
+                     }
+                     else
+                     {
+                        png_uint_16 v, w, x;
+
+                        v = gamma_16_to_1[*(sp + 1) >> gamma_shift][*sp];
+                        png_composite_16(w, v, a, background_1->red);
+                        x = gamma_16_from_1[((w&0xff) >> gamma_shift)][w >> 8];
+                        *dp = (png_byte)((x >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(x & 0xff);
+                        v = gamma_16_to_1[*(sp + 3) >> gamma_shift][*(sp + 2)];
+                        png_composite_16(w, v, a, background_1->green);
+                        x = gamma_16_from_1[((w&0xff) >> gamma_shift)][w >> 8];
+                        *(dp + 2) = (png_byte)((x >> 8) & 0xff);
+                        *(dp + 3) = (png_byte)(x & 0xff);
+                        v = gamma_16_to_1[*(sp + 5) >> gamma_shift][*(sp + 4)];
+                        png_composite_16(w, v, a, background_1->blue);
+                        x = gamma_16_from_1[(w & 0xff) >> gamma_shift][w >> 8];
+                        *(dp + 4) = (png_byte)((x >> 8) & 0xff);
+                        *(dp + 5) = (png_byte)(x & 0xff);
+                     }
+                  }
+               }
+               else
+#endif
+               {
+                  sp = row;
+                  dp = row;
+                  for (i = 0; i < row_width; i++, sp += 8, dp += 6)
+                  {
+                     png_uint_16 a = (png_uint_16)(((png_uint_16)(*(sp + 6))
+                        << 8) + (png_uint_16)(*(sp + 7)));
+                     if (a == (png_uint_16)0xffff)
+                     {
+                        png_memcpy(dp, sp, 6);
+                     }
+                     else if (a == 0)
+                     {
+                        *dp = (png_byte)((background->red >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(background->red & 0xff);
+                        *(dp + 2) = (png_byte)((background->green >> 8) & 0xff);
+                        *(dp + 3) = (png_byte)(background->green & 0xff);
+                        *(dp + 4) = (png_byte)((background->blue >> 8) & 0xff);
+                        *(dp + 5) = (png_byte)(background->blue & 0xff);
+                     }
+                     else
+                     {
+                        png_uint_16 v;
+
+                        png_uint_16 r = (png_uint_16)(((*sp) << 8) + *(sp + 1));
+                        png_uint_16 g = (png_uint_16)(((*(sp + 2)) << 8)
+                            + *(sp + 3));
+                        png_uint_16 b = (png_uint_16)(((*(sp + 4)) << 8)
+                            + *(sp + 5));
+
+                        png_composite_16(v, r, a, background->red);
+                        *dp = (png_byte)((v >> 8) & 0xff);
+                        *(dp + 1) = (png_byte)(v & 0xff);
+                        png_composite_16(v, g, a, background->green);
+                        *(dp + 2) = (png_byte)((v >> 8) & 0xff);
+                        *(dp + 3) = (png_byte)(v & 0xff);
+                        png_composite_16(v, b, a, background->blue);
+                        *(dp + 4) = (png_byte)((v >> 8) & 0xff);
+                        *(dp + 5) = (png_byte)(v & 0xff);
+                     }
+                  }
+               }
+            }
+            break;
+         }
+      }
+
+      if (row_info->color_type & PNG_COLOR_MASK_ALPHA)
+      {
+         row_info->color_type &= ~PNG_COLOR_MASK_ALPHA;
+         row_info->channels--;
+         row_info->pixel_depth = (png_byte)(row_info->channels *
+            row_info->bit_depth);
+         row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,row_width);
+      }
+   }
+}
+#endif
+
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+/* Gamma correct the image, avoiding the alpha channel.  Make sure
+ * you do this after you deal with the transparency issue on grayscale
+ * or RGB images. If your bit depth is 8, use gamma_table, if it
+ * is 16, use gamma_16_table and gamma_shift.  Build these with
+ * build_gamma_table().
+ */
+void /* PRIVATE */
+png_do_gamma(png_row_infop row_info, png_bytep row,
+   png_bytep gamma_table, png_uint_16pp gamma_16_table,
+   int gamma_shift)
+{
+   png_bytep sp;
+   png_uint_32 i;
+   png_uint_32 row_width=row_info->width;
+
+   png_debug(1, "in png_do_gamma\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+       ((row_info->bit_depth <= 8 && gamma_table != NULL) ||
+        (row_info->bit_depth == 16 && gamma_16_table != NULL)))
+   {
+      switch (row_info->color_type)
+      {
+         case PNG_COLOR_TYPE_RGB:
+         {
+            if (row_info->bit_depth == 8)
+            {
+               sp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  *sp = gamma_table[*sp];
+                  sp++;
+                  *sp = gamma_table[*sp];
+                  sp++;
+                  *sp = gamma_table[*sp];
+                  sp++;
+               }
+            }
+            else /* if (row_info->bit_depth == 16) */
+            {
+               sp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_uint_16 v;
+
+                  v = gamma_16_table[*(sp + 1) >> gamma_shift][*sp];
+                  *sp = (png_byte)((v >> 8) & 0xff);
+                  *(sp + 1) = (png_byte)(v & 0xff);
+                  sp += 2;
+                  v = gamma_16_table[*(sp + 1) >> gamma_shift][*sp];
+                  *sp = (png_byte)((v >> 8) & 0xff);
+                  *(sp + 1) = (png_byte)(v & 0xff);
+                  sp += 2;
+                  v = gamma_16_table[*(sp + 1) >> gamma_shift][*sp];
+                  *sp = (png_byte)((v >> 8) & 0xff);
+                  *(sp + 1) = (png_byte)(v & 0xff);
+                  sp += 2;
+               }
+            }
+            break;
+         }
+         case PNG_COLOR_TYPE_RGB_ALPHA:
+         {
+            if (row_info->bit_depth == 8)
+            {
+               sp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  *sp = gamma_table[*sp];
+                  sp++;
+                  *sp = gamma_table[*sp];
+                  sp++;
+                  *sp = gamma_table[*sp];
+                  sp++;
+                  sp++;
+               }
+            }
+            else /* if (row_info->bit_depth == 16) */
+            {
+               sp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_uint_16 v = gamma_16_table[*(sp + 1) >> gamma_shift][*sp];
+                  *sp = (png_byte)((v >> 8) & 0xff);
+                  *(sp + 1) = (png_byte)(v & 0xff);
+                  sp += 2;
+                  v = gamma_16_table[*(sp + 1) >> gamma_shift][*sp];
+                  *sp = (png_byte)((v >> 8) & 0xff);
+                  *(sp + 1) = (png_byte)(v & 0xff);
+                  sp += 2;
+                  v = gamma_16_table[*(sp + 1) >> gamma_shift][*sp];
+                  *sp = (png_byte)((v >> 8) & 0xff);
+                  *(sp + 1) = (png_byte)(v & 0xff);
+                  sp += 4;
+               }
+            }
+            break;
+         }
+         case PNG_COLOR_TYPE_GRAY_ALPHA:
+         {
+            if (row_info->bit_depth == 8)
+            {
+               sp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  *sp = gamma_table[*sp];
+                  sp += 2;
+               }
+            }
+            else /* if (row_info->bit_depth == 16) */
+            {
+               sp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_uint_16 v = gamma_16_table[*(sp + 1) >> gamma_shift][*sp];
+                  *sp = (png_byte)((v >> 8) & 0xff);
+                  *(sp + 1) = (png_byte)(v & 0xff);
+                  sp += 4;
+               }
+            }
+            break;
+         }
+         case PNG_COLOR_TYPE_GRAY:
+         {
+            if (row_info->bit_depth == 2)
+            {
+               sp = row;
+               for (i = 0; i < row_width; i += 4)
+               {
+                  int a = *sp & 0xc0;
+                  int b = *sp & 0x30;
+                  int c = *sp & 0x0c;
+                  int d = *sp & 0x03;
+
+                  *sp = (png_byte)(
+                        ((((int)gamma_table[a|(a>>2)|(a>>4)|(a>>6)])   ) & 0xc0)|
+                        ((((int)gamma_table[(b<<2)|b|(b>>2)|(b>>4)])>>2) & 0x30)|
+                        ((((int)gamma_table[(c<<4)|(c<<2)|c|(c>>2)])>>4) & 0x0c)|
+                        ((((int)gamma_table[(d<<6)|(d<<4)|(d<<2)|d])>>6) ));
+                  sp++;
+               }
+            }
+            if (row_info->bit_depth == 4)
+            {
+               sp = row;
+               for (i = 0; i < row_width; i += 2)
+               {
+                  int msb = *sp & 0xf0;
+                  int lsb = *sp & 0x0f;
+
+                  *sp = (png_byte)((((int)gamma_table[msb | (msb >> 4)]) & 0xf0)
+                          | (((int)gamma_table[(lsb << 4) | lsb]) >> 4));
+                  sp++;
+               }
+            }
+            else if (row_info->bit_depth == 8)
+            {
+               sp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  *sp = gamma_table[*sp];
+                  sp++;
+               }
+            }
+            else if (row_info->bit_depth == 16)
+            {
+               sp = row;
+               for (i = 0; i < row_width; i++)
+               {
+                  png_uint_16 v = gamma_16_table[*(sp + 1) >> gamma_shift][*sp];
+                  *sp = (png_byte)((v >> 8) & 0xff);
+                  *(sp + 1) = (png_byte)(v & 0xff);
+                  sp += 2;
+               }
+            }
+            break;
+         }
+      }
+   }
+}
+#endif
+
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+/* Expands a palette row to an RGB or RGBA row depending
+ * upon whether you supply trans and num_trans.
+ */
+void /* PRIVATE */
+png_do_expand_palette(png_row_infop row_info, png_bytep row,
+   png_colorp palette, png_bytep trans, int num_trans)
+{
+   int shift, value;
+   png_bytep sp, dp;
+   png_uint_32 i;
+   png_uint_32 row_width=row_info->width;
+
+   png_debug(1, "in png_do_expand_palette\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+       row_info->color_type == PNG_COLOR_TYPE_PALETTE)
+   {
+      if (row_info->bit_depth < 8)
+      {
+         switch (row_info->bit_depth)
+         {
+            case 1:
+            {
+               sp = row + (png_size_t)((row_width - 1) >> 3);
+               dp = row + (png_size_t)row_width - 1;
+               shift = 7 - (int)((row_width + 7) & 0x07);
+               for (i = 0; i < row_width; i++)
+               {
+                  if ((*sp >> shift) & 0x01)
+                     *dp = 1;
+                  else
+                     *dp = 0;
+                  if (shift == 7)
+                  {
+                     shift = 0;
+                     sp--;
+                  }
+                  else
+                     shift++;
+
+                  dp--;
+               }
+               break;
+            }
+            case 2:
+            {
+               sp = row + (png_size_t)((row_width - 1) >> 2);
+               dp = row + (png_size_t)row_width - 1;
+               shift = (int)((3 - ((row_width + 3) & 0x03)) << 1);
+               for (i = 0; i < row_width; i++)
+               {
+                  value = (*sp >> shift) & 0x03;
+                  *dp = (png_byte)value;
+                  if (shift == 6)
+                  {
+                     shift = 0;
+                     sp--;
+                  }
+                  else
+                     shift += 2;
+
+                  dp--;
+               }
+               break;
+            }
+            case 4:
+            {
+               sp = row + (png_size_t)((row_width - 1) >> 1);
+               dp = row + (png_size_t)row_width - 1;
+               shift = (int)((row_width & 0x01) << 2);
+               for (i = 0; i < row_width; i++)
+               {
+                  value = (*sp >> shift) & 0x0f;
+                  *dp = (png_byte)value;
+                  if (shift == 4)
+                  {
+                     shift = 0;
+                     sp--;
+                  }
+                  else
+                     shift += 4;
+
+                  dp--;
+               }
+               break;
+            }
+         }
+         row_info->bit_depth = 8;
+         row_info->pixel_depth = 8;
+         row_info->rowbytes = row_width;
+      }
+      switch (row_info->bit_depth)
+      {
+         case 8:
+         {
+            if (trans != NULL)
+            {
+               sp = row + (png_size_t)row_width - 1;
+               dp = row + (png_size_t)(row_width << 2) - 1;
+
+               for (i = 0; i < row_width; i++)
+               {
+                  if ((int)(*sp) >= num_trans)
+                     *dp-- = 0xff;
+                  else
+                     *dp-- = trans[*sp];
+                  *dp-- = palette[*sp].blue;
+                  *dp-- = palette[*sp].green;
+                  *dp-- = palette[*sp].red;
+                  sp--;
+               }
+               row_info->bit_depth = 8;
+               row_info->pixel_depth = 32;
+               row_info->rowbytes = row_width * 4;
+               row_info->color_type = 6;
+               row_info->channels = 4;
+            }
+            else
+            {
+               sp = row + (png_size_t)row_width - 1;
+               dp = row + (png_size_t)(row_width * 3) - 1;
+
+               for (i = 0; i < row_width; i++)
+               {
+                  *dp-- = palette[*sp].blue;
+                  *dp-- = palette[*sp].green;
+                  *dp-- = palette[*sp].red;
+                  sp--;
+               }
+               row_info->bit_depth = 8;
+               row_info->pixel_depth = 24;
+               row_info->rowbytes = row_width * 3;
+               row_info->color_type = 2;
+               row_info->channels = 3;
+            }
+            break;
+         }
+      }
+   }
+}
+
+/* If the bit depth < 8, it is expanded to 8.  Also, if the already
+ * expanded transparency value is supplied, an alpha channel is built.
+ */
+void /* PRIVATE */
+png_do_expand(png_row_infop row_info, png_bytep row,
+   png_color_16p trans_value)
+{
+   int shift, value;
+   png_bytep sp, dp;
+   png_uint_32 i;
+   png_uint_32 row_width=row_info->width;
+
+   png_debug(1, "in png_do_expand\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL)
+#endif
+   {
+      if (row_info->color_type == PNG_COLOR_TYPE_GRAY)
+      {
+         png_uint_16 gray = (png_uint_16)(trans_value ? trans_value->gray : 0);
+
+         if (row_info->bit_depth < 8)
+         {
+            switch (row_info->bit_depth)
+            {
+               case 1:
+               {
+                  gray = (png_uint_16)((gray&0x01)*0xff);
+                  sp = row + (png_size_t)((row_width - 1) >> 3);
+                  dp = row + (png_size_t)row_width - 1;
+                  shift = 7 - (int)((row_width + 7) & 0x07);
+                  for (i = 0; i < row_width; i++)
+                  {
+                     if ((*sp >> shift) & 0x01)
+                        *dp = 0xff;
+                     else
+                        *dp = 0;
+                     if (shift == 7)
+                     {
+                        shift = 0;
+                        sp--;
+                     }
+                     else
+                        shift++;
+
+                     dp--;
+                  }
+                  break;
+               }
+               case 2:
+               {
+                  gray = (png_uint_16)((gray&0x03)*0x55);
+                  sp = row + (png_size_t)((row_width - 1) >> 2);
+                  dp = row + (png_size_t)row_width - 1;
+                  shift = (int)((3 - ((row_width + 3) & 0x03)) << 1);
+                  for (i = 0; i < row_width; i++)
+                  {
+                     value = (*sp >> shift) & 0x03;
+                     *dp = (png_byte)(value | (value << 2) | (value << 4) |
+                        (value << 6));
+                     if (shift == 6)
+                     {
+                        shift = 0;
+                        sp--;
+                     }
+                     else
+                        shift += 2;
+
+                     dp--;
+                  }
+                  break;
+               }
+               case 4:
+               {
+                  gray = (png_uint_16)((gray&0x0f)*0x11);
+                  sp = row + (png_size_t)((row_width - 1) >> 1);
+                  dp = row + (png_size_t)row_width - 1;
+                  shift = (int)((1 - ((row_width + 1) & 0x01)) << 2);
+                  for (i = 0; i < row_width; i++)
+                  {
+                     value = (*sp >> shift) & 0x0f;
+                     *dp = (png_byte)(value | (value << 4));
+                     if (shift == 4)
+                     {
+                        shift = 0;
+                        sp--;
+                     }
+                     else
+                        shift = 4;
+
+                     dp--;
+                  }
+                  break;
+               }
+            }
+            row_info->bit_depth = 8;
+            row_info->pixel_depth = 8;
+            row_info->rowbytes = row_width;
+         }
+
+         if (trans_value != NULL)
+         {
+            if (row_info->bit_depth == 8)
+            {
+               gray = gray & 0xff;
+               sp = row + (png_size_t)row_width - 1;
+               dp = row + (png_size_t)(row_width << 1) - 1;
+               for (i = 0; i < row_width; i++)
+               {
+                  if (*sp == gray)
+                     *dp-- = 0;
+                  else
+                     *dp-- = 0xff;
+                  *dp-- = *sp--;
+               }
+            }
+            else if (row_info->bit_depth == 16)
+            {
+               png_byte gray_high = (gray >> 8) & 0xff;
+               png_byte gray_low = gray & 0xff;
+               sp = row + row_info->rowbytes - 1;
+               dp = row + (row_info->rowbytes << 1) - 1;
+               for (i = 0; i < row_width; i++)
+               {
+                  if (*(sp-1) == gray_high && *(sp) == gray_low) 
+                  {
+                     *dp-- = 0;
+                     *dp-- = 0;
+                  }
+                  else
+                  {
+                     *dp-- = 0xff;
+                     *dp-- = 0xff;
+                  }
+                  *dp-- = *sp--;
+                  *dp-- = *sp--;
+               }
+            }
+            row_info->color_type = PNG_COLOR_TYPE_GRAY_ALPHA;
+            row_info->channels = 2;
+            row_info->pixel_depth = (png_byte)(row_info->bit_depth << 1);
+            row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,
+               row_width);
+         }
+      }
+      else if (row_info->color_type == PNG_COLOR_TYPE_RGB && trans_value)
+      {
+         if (row_info->bit_depth == 8)
+         {
+            png_byte red = trans_value->red & 0xff;
+            png_byte green = trans_value->green & 0xff;
+            png_byte blue = trans_value->blue & 0xff;
+            sp = row + (png_size_t)row_info->rowbytes - 1;
+            dp = row + (png_size_t)(row_width << 2) - 1;
+            for (i = 0; i < row_width; i++)
+            {
+               if (*(sp - 2) == red && *(sp - 1) == green && *(sp) == blue)
+                  *dp-- = 0;
+               else
+                  *dp-- = 0xff;
+               *dp-- = *sp--;
+               *dp-- = *sp--;
+               *dp-- = *sp--;
+            }
+         }
+         else if (row_info->bit_depth == 16)
+         {
+            png_byte red_high = (trans_value->red > 8) & 0xff;
+            png_byte green_high = (trans_value->green > 8) & 0xff;
+            png_byte blue_high = (trans_value->blue > 8) & 0xff;
+            png_byte red_low = trans_value->red & 0xff;
+            png_byte green_low = trans_value->green & 0xff;
+            png_byte blue_low = trans_value->blue & 0xff;
+            sp = row + row_info->rowbytes - 1;
+            dp = row + (png_size_t)(row_width << 3) - 1;
+            for (i = 0; i < row_width; i++)
+            {
+               if (*(sp - 5) == red_high &&
+                  *(sp - 4) == red_low &&
+                  *(sp - 3) == green_high &&
+                  *(sp - 2) == green_low &&
+                  *(sp - 1) == blue_high &&
+                  *(sp    ) == blue_low)
+               {
+                  *dp-- = 0;
+                  *dp-- = 0;
+               }
+               else
+               {
+                  *dp-- = 0xff;
+                  *dp-- = 0xff;
+               }
+               *dp-- = *sp--;
+               *dp-- = *sp--;
+               *dp-- = *sp--;
+               *dp-- = *sp--;
+               *dp-- = *sp--;
+               *dp-- = *sp--;
+            }
+         }
+         row_info->color_type = PNG_COLOR_TYPE_RGB_ALPHA;
+         row_info->channels = 4;
+         row_info->pixel_depth = (png_byte)(row_info->bit_depth << 2);
+         row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,row_width);
+      }
+   }
+}
+#endif
+
+#if defined(PNG_READ_DITHER_SUPPORTED)
+void /* PRIVATE */
+png_do_dither(png_row_infop row_info, png_bytep row,
+    png_bytep palette_lookup, png_bytep dither_lookup)
+{
+   png_bytep sp, dp;
+   png_uint_32 i;
+   png_uint_32 row_width=row_info->width;
+
+   png_debug(1, "in png_do_dither\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL)
+#endif
+   {
+      if (row_info->color_type == PNG_COLOR_TYPE_RGB &&
+         palette_lookup && row_info->bit_depth == 8)
+      {
+         int r, g, b, p;
+         sp = row;
+         dp = row;
+         for (i = 0; i < row_width; i++)
+         {
+            r = *sp++;
+            g = *sp++;
+            b = *sp++;
+
+            /* this looks real messy, but the compiler will reduce
+               it down to a reasonable formula.  For example, with
+               5 bits per color, we get:
+               p = (((r >> 3) & 0x1f) << 10) |
+                  (((g >> 3) & 0x1f) << 5) |
+                  ((b >> 3) & 0x1f);
+               */
+            p = (((r >> (8 - PNG_DITHER_RED_BITS)) &
+               ((1 << PNG_DITHER_RED_BITS) - 1)) <<
+               (PNG_DITHER_GREEN_BITS + PNG_DITHER_BLUE_BITS)) |
+               (((g >> (8 - PNG_DITHER_GREEN_BITS)) &
+               ((1 << PNG_DITHER_GREEN_BITS) - 1)) <<
+               (PNG_DITHER_BLUE_BITS)) |
+               ((b >> (8 - PNG_DITHER_BLUE_BITS)) &
+               ((1 << PNG_DITHER_BLUE_BITS) - 1));
+
+            *dp++ = palette_lookup[p];
+         }
+         row_info->color_type = PNG_COLOR_TYPE_PALETTE;
+         row_info->channels = 1;
+         row_info->pixel_depth = row_info->bit_depth;
+         row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,row_width);
+      }
+      else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA &&
+         palette_lookup != NULL && row_info->bit_depth == 8)
+      {
+         int r, g, b, p;
+         sp = row;
+         dp = row;
+         for (i = 0; i < row_width; i++)
+         {
+            r = *sp++;
+            g = *sp++;
+            b = *sp++;
+            sp++;
+
+            p = (((r >> (8 - PNG_DITHER_RED_BITS)) &
+               ((1 << PNG_DITHER_RED_BITS) - 1)) <<
+               (PNG_DITHER_GREEN_BITS + PNG_DITHER_BLUE_BITS)) |
+               (((g >> (8 - PNG_DITHER_GREEN_BITS)) &
+               ((1 << PNG_DITHER_GREEN_BITS) - 1)) <<
+               (PNG_DITHER_BLUE_BITS)) |
+               ((b >> (8 - PNG_DITHER_BLUE_BITS)) &
+               ((1 << PNG_DITHER_BLUE_BITS) - 1));
+
+            *dp++ = palette_lookup[p];
+         }
+         row_info->color_type = PNG_COLOR_TYPE_PALETTE;
+         row_info->channels = 1;
+         row_info->pixel_depth = row_info->bit_depth;
+         row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,row_width);
+      }
+      else if (row_info->color_type == PNG_COLOR_TYPE_PALETTE &&
+         dither_lookup && row_info->bit_depth == 8)
+      {
+         sp = row;
+         for (i = 0; i < row_width; i++, sp++)
+         {
+            *sp = dither_lookup[*sp];
+         }
+      }
+   }
+}
+#endif
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+#if defined(PNG_READ_GAMMA_SUPPORTED)
+static PNG_CONST int png_gamma_shift[] =
+   {0x10, 0x21, 0x42, 0x84, 0x110, 0x248, 0x550, 0xff0, 0x00};
+
+/* We build the 8- or 16-bit gamma tables here.  Note that for 16-bit
+ * tables, we don't make a full table if we are reducing to 8-bit in
+ * the future.  Note also how the gamma_16 tables are segmented so that
+ * we don't need to allocate > 64K chunks for a full 16-bit table.
+ */
+void /* PRIVATE */
+png_build_gamma_table(png_structp png_ptr)
+{
+  png_debug(1, "in png_build_gamma_table\n");
+
+  if (png_ptr->bit_depth <= 8)
+  {
+     int i;
+     double g;
+
+     if (png_ptr->screen_gamma > .000001)
+        g = 1.0 / (png_ptr->gamma * png_ptr->screen_gamma);
+     else
+        g = 1.0;
+
+     png_ptr->gamma_table = (png_bytep)png_malloc(png_ptr,
+        (png_uint_32)256);
+
+     for (i = 0; i < 256; i++)
+     {
+        png_ptr->gamma_table[i] = (png_byte)(pow((double)i / 255.0,
+           g) * 255.0 + .5);
+     }
+
+#if defined(PNG_READ_BACKGROUND_SUPPORTED) || \
+   defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+     if (png_ptr->transformations & ((PNG_BACKGROUND) | PNG_RGB_TO_GRAY))
+     {
+
+        g = 1.0 / (png_ptr->gamma);
+
+        png_ptr->gamma_to_1 = (png_bytep)png_malloc(png_ptr,
+           (png_uint_32)256);
+
+        for (i = 0; i < 256; i++)
+        {
+           png_ptr->gamma_to_1[i] = (png_byte)(pow((double)i / 255.0,
+              g) * 255.0 + .5);
+        }
+
+
+        png_ptr->gamma_from_1 = (png_bytep)png_malloc(png_ptr,
+           (png_uint_32)256);
+
+        if(png_ptr->screen_gamma > 0.000001)
+           g = 1.0 / png_ptr->screen_gamma;
+        else
+           g = png_ptr->gamma;   /* probably doing rgb_to_gray */
+
+        for (i = 0; i < 256; i++)
+        {
+           png_ptr->gamma_from_1[i] = (png_byte)(pow((double)i / 255.0,
+              g) * 255.0 + .5);
+
+        }
+     }
+#endif /* PNG_READ_BACKGROUND_SUPPORTED || PNG_RGB_TO_GRAY_SUPPORTED */
+  }
+  else
+  {
+     double g;
+     int i, j, shift, num;
+     int sig_bit;
+     png_uint_32 ig;
+
+     if (png_ptr->color_type & PNG_COLOR_MASK_COLOR)
+     {
+        sig_bit = (int)png_ptr->sig_bit.red;
+        if ((int)png_ptr->sig_bit.green > sig_bit)
+           sig_bit = png_ptr->sig_bit.green;
+        if ((int)png_ptr->sig_bit.blue > sig_bit)
+           sig_bit = png_ptr->sig_bit.blue;
+     }
+     else
+     {
+        sig_bit = (int)png_ptr->sig_bit.gray;
+     }
+
+     if (sig_bit > 0)
+        shift = 16 - sig_bit;
+     else
+        shift = 0;
+
+     if (png_ptr->transformations & PNG_16_TO_8)
+     {
+        if (shift < (16 - PNG_MAX_GAMMA_8))
+           shift = (16 - PNG_MAX_GAMMA_8);
+     }
+
+     if (shift > 8)
+        shift = 8;
+     if (shift < 0)
+        shift = 0;
+
+     png_ptr->gamma_shift = (png_byte)shift;
+
+     num = (1 << (8 - shift));
+
+     if (png_ptr->screen_gamma > .000001)
+        g = 1.0 / (png_ptr->gamma * png_ptr->screen_gamma);
+     else
+        g = 1.0;
+
+     png_ptr->gamma_16_table = (png_uint_16pp)png_malloc(png_ptr,
+        (png_uint_32)(num * png_sizeof (png_uint_16p)));
+
+     if (png_ptr->transformations & (PNG_16_TO_8 | PNG_BACKGROUND))
+     {
+        double fin, fout;
+        png_uint_32 last, max;
+
+        for (i = 0; i < num; i++)
+        {
+           png_ptr->gamma_16_table[i] = (png_uint_16p)png_malloc(png_ptr,
+              (png_uint_32)(256 * png_sizeof (png_uint_16)));
+        }
+
+        g = 1.0 / g;
+        last = 0;
+        for (i = 0; i < 256; i++)
+        {
+           fout = ((double)i + 0.5) / 256.0;
+           fin = pow(fout, g);
+           max = (png_uint_32)(fin * (double)((png_uint_32)num << 8));
+           while (last <= max)
+           {
+              png_ptr->gamma_16_table[(int)(last & (0xff >> shift))]
+                 [(int)(last >> (8 - shift))] = (png_uint_16)(
+                 (png_uint_16)i | ((png_uint_16)i << 8));
+              last++;
+           }
+        }
+        while (last < ((png_uint_32)num << 8))
+        {
+           png_ptr->gamma_16_table[(int)(last & (0xff >> shift))]
+              [(int)(last >> (8 - shift))] = (png_uint_16)65535L;
+           last++;
+        }
+     }
+     else
+     {
+        for (i = 0; i < num; i++)
+        {
+           png_ptr->gamma_16_table[i] = (png_uint_16p)png_malloc(png_ptr,
+              (png_uint_32)(256 * png_sizeof (png_uint_16)));
+
+           ig = (((png_uint_32)i * (png_uint_32)png_gamma_shift[shift]) >> 4);
+           for (j = 0; j < 256; j++)
+           {
+              png_ptr->gamma_16_table[i][j] =
+                 (png_uint_16)(pow((double)(ig + ((png_uint_32)j << 8)) /
+                    65535.0, g) * 65535.0 + .5);
+           }
+        }
+     }
+
+#if defined(PNG_READ_BACKGROUND_SUPPORTED) || \
+   defined(PNG_READ_RGB_TO_GRAY_SUPPORTED)
+     if (png_ptr->transformations & (PNG_BACKGROUND | PNG_RGB_TO_GRAY))
+     {
+
+        g = 1.0 / (png_ptr->gamma);
+
+        png_ptr->gamma_16_to_1 = (png_uint_16pp)png_malloc(png_ptr,
+           (png_uint_32)(num * png_sizeof (png_uint_16p )));
+
+        for (i = 0; i < num; i++)
+        {
+           png_ptr->gamma_16_to_1[i] = (png_uint_16p)png_malloc(png_ptr,
+              (png_uint_32)(256 * png_sizeof (png_uint_16)));
+
+           ig = (((png_uint_32)i *
+              (png_uint_32)png_gamma_shift[shift]) >> 4);
+           for (j = 0; j < 256; j++)
+           {
+              png_ptr->gamma_16_to_1[i][j] =
+                 (png_uint_16)(pow((double)(ig + ((png_uint_32)j << 8)) /
+                    65535.0, g) * 65535.0 + .5);
+           }
+        }
+
+        if(png_ptr->screen_gamma > 0.000001)
+           g = 1.0 / png_ptr->screen_gamma;
+        else
+           g = png_ptr->gamma;   /* probably doing rgb_to_gray */
+
+        png_ptr->gamma_16_from_1 = (png_uint_16pp)png_malloc(png_ptr,
+           (png_uint_32)(num * png_sizeof (png_uint_16p)));
+
+        for (i = 0; i < num; i++)
+        {
+           png_ptr->gamma_16_from_1[i] = (png_uint_16p)png_malloc(png_ptr,
+              (png_uint_32)(256 * png_sizeof (png_uint_16)));
+
+           ig = (((png_uint_32)i *
+              (png_uint_32)png_gamma_shift[shift]) >> 4);
+           for (j = 0; j < 256; j++)
+           {
+              png_ptr->gamma_16_from_1[i][j] =
+                 (png_uint_16)(pow((double)(ig + ((png_uint_32)j << 8)) /
+                    65535.0, g) * 65535.0 + .5);
+           }
+        }
+     }
+#endif /* PNG_READ_BACKGROUND_SUPPORTED || PNG_RGB_TO_GRAY_SUPPORTED */
+  }
+}
+#endif
+/* To do: install integer version of png_build_gamma_table here */
+#endif
+
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+/* undoes intrapixel differencing  */
+void /* PRIVATE */
+png_do_read_intrapixel(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_read_intrapixel\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+       (row_info->color_type & PNG_COLOR_MASK_COLOR))
+   {
+      int bytes_per_pixel;
+      png_uint_32 row_width = row_info->width;
+      if (row_info->bit_depth == 8)
+      {
+         png_bytep rp;
+         png_uint_32 i;
+
+         if (row_info->color_type == PNG_COLOR_TYPE_RGB)
+            bytes_per_pixel = 3;
+         else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+            bytes_per_pixel = 4;
+         else
+            return;
+
+         for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
+         {
+            *(rp) = (png_byte)((256 + *rp + *(rp+1))&0xff);
+            *(rp+2) = (png_byte)((256 + *(rp+2) + *(rp+1))&0xff);
+         }
+      }
+      else if (row_info->bit_depth == 16)
+      {
+         png_bytep rp;
+         png_uint_32 i;
+
+         if (row_info->color_type == PNG_COLOR_TYPE_RGB)
+            bytes_per_pixel = 6;
+         else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+            bytes_per_pixel = 8;
+         else
+            return;
+
+         for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
+         {
+            png_uint_32 s0   = (*(rp  ) << 8) | *(rp+1);
+            png_uint_32 s1   = (*(rp+2) << 8) | *(rp+3);
+            png_uint_32 s2   = (*(rp+4) << 8) | *(rp+5);
+            png_uint_32 red  = (png_uint_32)((s0+s1+65536L) & 0xffffL);
+            png_uint_32 blue = (png_uint_32)((s2+s1+65536L) & 0xffffL);
+            *(rp  ) = (png_byte)((red >> 8) & 0xff);
+            *(rp+1) = (png_byte)(red & 0xff);
+            *(rp+4) = (png_byte)((blue >> 8) & 0xff);
+            *(rp+5) = (png_byte)(blue & 0xff);
+         }
+      }
+   }
+}
+#endif /* PNG_MNG_FEATURES_SUPPORTED */
+#endif /* PNG_READ_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngrutil.c b/distrib/libpng-1.2.19/pngrutil.c
new file mode 100644
index 0000000..86e924c
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngrutil.c
@@ -0,0 +1,4189 @@
+
+/* pngrutil.c - utilities to read a PNG file
+ *
+ * Last changed in libpng 1.2.19 August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ *
+ * This file contains routines that are only called from within
+ * libpng itself during the course of reading an image.
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_READ_SUPPORTED)
+
+#if defined(_WIN32_WCE) && (_WIN32_WCE<0x500)
+#  define WIN32_WCE_OLD
+#endif
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+#  if defined(WIN32_WCE_OLD)
+/* strtod() function is not supported on WindowsCE */
+__inline double png_strtod(png_structp png_ptr, PNG_CONST char *nptr, char **endptr)
+{
+   double result = 0;
+   int len;
+   wchar_t *str, *end;
+
+   len = MultiByteToWideChar(CP_ACP, 0, nptr, -1, NULL, 0);
+   str = (wchar_t *)png_malloc(png_ptr, len * sizeof(wchar_t));
+   if ( NULL != str )
+   {
+      MultiByteToWideChar(CP_ACP, 0, nptr, -1, str, len);
+      result = wcstod(str, &end);
+      len = WideCharToMultiByte(CP_ACP, 0, end, -1, NULL, 0, NULL, NULL);
+      *endptr = (char *)nptr + (png_strlen(nptr) - len + 1);
+      png_free(png_ptr, str);
+   }
+   return result;
+}
+#  else
+#    define png_strtod(p,a,b) strtod(a,b)
+#  endif
+#endif
+
+png_uint_32 PNGAPI
+png_get_uint_31(png_structp png_ptr, png_bytep buf)
+{
+   png_uint_32 i = png_get_uint_32(buf);
+   if (i > PNG_UINT_31_MAX)
+     png_error(png_ptr, "PNG unsigned integer out of range.");
+   return (i);
+}
+#ifndef PNG_READ_BIG_ENDIAN_SUPPORTED
+/* Grab an unsigned 32-bit integer from a buffer in big-endian format. */
+png_uint_32 PNGAPI
+png_get_uint_32(png_bytep buf)
+{
+   png_uint_32 i = ((png_uint_32)(*buf) << 24) +
+      ((png_uint_32)(*(buf + 1)) << 16) +
+      ((png_uint_32)(*(buf + 2)) << 8) +
+      (png_uint_32)(*(buf + 3));
+
+   return (i);
+}
+
+/* Grab a signed 32-bit integer from a buffer in big-endian format.  The
+ * data is stored in the PNG file in two's complement format, and it is
+ * assumed that the machine format for signed integers is the same. */
+png_int_32 PNGAPI
+png_get_int_32(png_bytep buf)
+{
+   png_int_32 i = ((png_int_32)(*buf) << 24) +
+      ((png_int_32)(*(buf + 1)) << 16) +
+      ((png_int_32)(*(buf + 2)) << 8) +
+      (png_int_32)(*(buf + 3));
+
+   return (i);
+}
+
+/* Grab an unsigned 16-bit integer from a buffer in big-endian format. */
+png_uint_16 PNGAPI
+png_get_uint_16(png_bytep buf)
+{
+   png_uint_16 i = (png_uint_16)(((png_uint_16)(*buf) << 8) +
+      (png_uint_16)(*(buf + 1)));
+
+   return (i);
+}
+#endif /* PNG_READ_BIG_ENDIAN_SUPPORTED */
+
+/* Read data, and (optionally) run it through the CRC. */
+void /* PRIVATE */
+png_crc_read(png_structp png_ptr, png_bytep buf, png_size_t length)
+{
+   if(png_ptr == NULL) return;
+   png_read_data(png_ptr, buf, length);
+   png_calculate_crc(png_ptr, buf, length);
+}
+
+/* Optionally skip data and then check the CRC.  Depending on whether we
+   are reading a ancillary or critical chunk, and how the program has set
+   things up, we may calculate the CRC on the data and print a message.
+   Returns '1' if there was a CRC error, '0' otherwise. */
+int /* PRIVATE */
+png_crc_finish(png_structp png_ptr, png_uint_32 skip)
+{
+   png_size_t i;
+   png_size_t istop = png_ptr->zbuf_size;
+
+   for (i = (png_size_t)skip; i > istop; i -= istop)
+   {
+      png_crc_read(png_ptr, png_ptr->zbuf, png_ptr->zbuf_size);
+   }
+   if (i)
+   {
+      png_crc_read(png_ptr, png_ptr->zbuf, i);
+   }
+
+   if (png_crc_error(png_ptr))
+   {
+      if (((png_ptr->chunk_name[0] & 0x20) &&                /* Ancillary */
+           !(png_ptr->flags & PNG_FLAG_CRC_ANCILLARY_NOWARN)) ||
+          (!(png_ptr->chunk_name[0] & 0x20) &&             /* Critical  */
+          (png_ptr->flags & PNG_FLAG_CRC_CRITICAL_USE)))
+      {
+         png_chunk_warning(png_ptr, "CRC error");
+      }
+      else
+      {
+         png_chunk_error(png_ptr, "CRC error");
+      }
+      return (1);
+   }
+
+   return (0);
+}
+
+/* Compare the CRC stored in the PNG file with that calculated by libpng from
+   the data it has read thus far. */
+int /* PRIVATE */
+png_crc_error(png_structp png_ptr)
+{
+   png_byte crc_bytes[4];
+   png_uint_32 crc;
+   int need_crc = 1;
+
+   if (png_ptr->chunk_name[0] & 0x20)                     /* ancillary */
+   {
+      if ((png_ptr->flags & PNG_FLAG_CRC_ANCILLARY_MASK) ==
+          (PNG_FLAG_CRC_ANCILLARY_USE | PNG_FLAG_CRC_ANCILLARY_NOWARN))
+         need_crc = 0;
+   }
+   else                                                    /* critical */
+   {
+      if (png_ptr->flags & PNG_FLAG_CRC_CRITICAL_IGNORE)
+         need_crc = 0;
+   }
+
+   png_read_data(png_ptr, crc_bytes, 4);
+
+   if (need_crc)
+   {
+      crc = png_get_uint_32(crc_bytes);
+      return ((int)(crc != png_ptr->crc));
+   }
+   else
+      return (0);
+}
+
+#if defined(PNG_READ_zTXt_SUPPORTED) || defined(PNG_READ_iTXt_SUPPORTED) || \
+    defined(PNG_READ_iCCP_SUPPORTED)
+/*
+ * Decompress trailing data in a chunk.  The assumption is that chunkdata
+ * points at an allocated area holding the contents of a chunk with a
+ * trailing compressed part.  What we get back is an allocated area
+ * holding the original prefix part and an uncompressed version of the
+ * trailing part (the malloc area passed in is freed).
+ */
+png_charp /* PRIVATE */
+png_decompress_chunk(png_structp png_ptr, int comp_type,
+                              png_charp chunkdata, png_size_t chunklength,
+                              png_size_t prefix_size, png_size_t *newlength)
+{
+   static PNG_CONST char msg[] = "Error decoding compressed text";
+   png_charp text;
+   png_size_t text_size;
+
+   if (comp_type == PNG_COMPRESSION_TYPE_BASE)
+   {
+      int ret = Z_OK;
+      png_ptr->zstream.next_in = (png_bytep)(chunkdata + prefix_size);
+      png_ptr->zstream.avail_in = (uInt)(chunklength - prefix_size);
+      png_ptr->zstream.next_out = png_ptr->zbuf;
+      png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+
+      text_size = 0;
+      text = NULL;
+
+      while (png_ptr->zstream.avail_in)
+      {
+         ret = inflate(&png_ptr->zstream, Z_PARTIAL_FLUSH);
+         if (ret != Z_OK && ret != Z_STREAM_END)
+         {
+            if (png_ptr->zstream.msg != NULL)
+               png_warning(png_ptr, png_ptr->zstream.msg);
+            else
+               png_warning(png_ptr, msg);
+            inflateReset(&png_ptr->zstream);
+            png_ptr->zstream.avail_in = 0;
+
+            if (text ==  NULL)
+            {
+               text_size = prefix_size + png_sizeof(msg) + 1;
+               text = (png_charp)png_malloc_warn(png_ptr, text_size);
+               if (text ==  NULL)
+                 {
+                    png_free(png_ptr,chunkdata);
+                    png_error(png_ptr,"Not enough memory to decompress chunk");
+                 }
+               png_memcpy(text, chunkdata, prefix_size);
+            }
+
+            text[text_size - 1] = 0x00;
+
+            /* Copy what we can of the error message into the text chunk */
+            text_size = (png_size_t)(chunklength - (text - chunkdata) - 1);
+            text_size = png_sizeof(msg) > text_size ? text_size :
+               png_sizeof(msg);
+            png_memcpy(text + prefix_size, msg, text_size + 1);
+            break;
+         }
+         if (!png_ptr->zstream.avail_out || ret == Z_STREAM_END)
+         {
+            if (text == NULL)
+            {
+               text_size = prefix_size +
+                   png_ptr->zbuf_size - png_ptr->zstream.avail_out;
+               text = (png_charp)png_malloc_warn(png_ptr, text_size + 1);
+               if (text ==  NULL)
+                 {
+                    png_free(png_ptr,chunkdata);
+                    png_error(png_ptr,"Not enough memory to decompress chunk.");
+                 }
+               png_memcpy(text + prefix_size, png_ptr->zbuf,
+                    text_size - prefix_size);
+               png_memcpy(text, chunkdata, prefix_size);
+               *(text + text_size) = 0x00;
+            }
+            else
+            {
+               png_charp tmp;
+
+               tmp = text;
+               text = (png_charp)png_malloc_warn(png_ptr,
+                  (png_uint_32)(text_size +
+                  png_ptr->zbuf_size - png_ptr->zstream.avail_out + 1));
+               if (text == NULL)
+               {
+                  png_free(png_ptr, tmp);
+                  png_free(png_ptr, chunkdata);
+                  png_error(png_ptr,"Not enough memory to decompress chunk..");
+               }
+               png_memcpy(text, tmp, text_size);
+               png_free(png_ptr, tmp);
+               png_memcpy(text + text_size, png_ptr->zbuf,
+                  (png_ptr->zbuf_size - png_ptr->zstream.avail_out));
+               text_size += png_ptr->zbuf_size - png_ptr->zstream.avail_out;
+               *(text + text_size) = 0x00;
+            }
+            if (ret == Z_STREAM_END)
+               break;
+            else
+            {
+               png_ptr->zstream.next_out = png_ptr->zbuf;
+               png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+            }
+         }
+      }
+      if (ret != Z_STREAM_END)
+      {
+#if !defined(PNG_NO_STDIO) && !defined(_WIN32_WCE)
+         char umsg[52];
+
+         if (ret == Z_BUF_ERROR)
+            png_snprintf(umsg, 52,
+                "Buffer error in compressed datastream in %s chunk",
+                png_ptr->chunk_name);
+         else if (ret == Z_DATA_ERROR)
+            png_snprintf(umsg, 52,
+                "Data error in compressed datastream in %s chunk",
+                png_ptr->chunk_name);
+         else
+            png_snprintf(umsg, 52,
+                "Incomplete compressed datastream in %s chunk",
+                png_ptr->chunk_name);
+         png_warning(png_ptr, umsg);
+#else
+         png_warning(png_ptr,
+            "Incomplete compressed datastream in chunk other than IDAT");
+#endif
+         text_size=prefix_size;
+         if (text ==  NULL)
+         {
+            text = (png_charp)png_malloc_warn(png_ptr, text_size+1);
+            if (text == NULL)
+              {
+                png_free(png_ptr, chunkdata);
+                png_error(png_ptr,"Not enough memory for text.");
+              }
+            png_memcpy(text, chunkdata, prefix_size);
+         }
+         *(text + text_size) = 0x00;
+      }
+
+      inflateReset(&png_ptr->zstream);
+      png_ptr->zstream.avail_in = 0;
+
+      png_free(png_ptr, chunkdata);
+      chunkdata = text;
+      *newlength=text_size;
+   }
+   else /* if (comp_type != PNG_COMPRESSION_TYPE_BASE) */
+   {
+#if !defined(PNG_NO_STDIO) && !defined(_WIN32_WCE)
+      char umsg[50];
+
+      png_snprintf(umsg, 50,
+         "Unknown zTXt compression type %d", comp_type);
+      png_warning(png_ptr, umsg);
+#else
+      png_warning(png_ptr, "Unknown zTXt compression type");
+#endif
+
+      *(chunkdata + prefix_size) = 0x00;
+      *newlength=prefix_size;
+   }
+
+   return chunkdata;
+}
+#endif
+
+/* read and check the IDHR chunk */
+void /* PRIVATE */
+png_handle_IHDR(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_byte buf[13];
+   png_uint_32 width, height;
+   int bit_depth, color_type, compression_type, filter_type;
+   int interlace_type;
+
+   png_debug(1, "in png_handle_IHDR\n");
+
+   if (png_ptr->mode & PNG_HAVE_IHDR)
+      png_error(png_ptr, "Out of place IHDR");
+
+   /* check the length */
+   if (length != 13)
+      png_error(png_ptr, "Invalid IHDR chunk");
+
+   png_ptr->mode |= PNG_HAVE_IHDR;
+
+   png_crc_read(png_ptr, buf, 13);
+   png_crc_finish(png_ptr, 0);
+
+   width = png_get_uint_31(png_ptr, buf);
+   height = png_get_uint_31(png_ptr, buf + 4);
+   bit_depth = buf[8];
+   color_type = buf[9];
+   compression_type = buf[10];
+   filter_type = buf[11];
+   interlace_type = buf[12];
+
+   /* set internal variables */
+   png_ptr->width = width;
+   png_ptr->height = height;
+   png_ptr->bit_depth = (png_byte)bit_depth;
+   png_ptr->interlaced = (png_byte)interlace_type;
+   png_ptr->color_type = (png_byte)color_type;
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+   png_ptr->filter_type = (png_byte)filter_type;
+#endif
+   png_ptr->compression_type = (png_byte)compression_type;
+
+   /* find number of channels */
+   switch (png_ptr->color_type)
+   {
+      case PNG_COLOR_TYPE_GRAY:
+      case PNG_COLOR_TYPE_PALETTE:
+         png_ptr->channels = 1;
+         break;
+      case PNG_COLOR_TYPE_RGB:
+         png_ptr->channels = 3;
+         break;
+      case PNG_COLOR_TYPE_GRAY_ALPHA:
+         png_ptr->channels = 2;
+         break;
+      case PNG_COLOR_TYPE_RGB_ALPHA:
+         png_ptr->channels = 4;
+         break;
+   }
+
+   /* set up other useful info */
+   png_ptr->pixel_depth = (png_byte)(png_ptr->bit_depth *
+   png_ptr->channels);
+   png_ptr->rowbytes = PNG_ROWBYTES(png_ptr->pixel_depth,png_ptr->width);
+   png_debug1(3,"bit_depth = %d\n", png_ptr->bit_depth);
+   png_debug1(3,"channels = %d\n", png_ptr->channels);
+   png_debug1(3,"rowbytes = %lu\n", png_ptr->rowbytes);
+   png_set_IHDR(png_ptr, info_ptr, width, height, bit_depth,
+      color_type, interlace_type, compression_type, filter_type);
+}
+
+/* read and check the palette */
+void /* PRIVATE */
+png_handle_PLTE(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_color palette[PNG_MAX_PALETTE_LENGTH];
+   int num, i;
+#ifndef PNG_NO_POINTER_INDEXING
+   png_colorp pal_ptr;
+#endif
+
+   png_debug(1, "in png_handle_PLTE\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before PLTE");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid PLTE after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (png_ptr->mode & PNG_HAVE_PLTE)
+      png_error(png_ptr, "Duplicate PLTE chunk");
+
+   png_ptr->mode |= PNG_HAVE_PLTE;
+
+   if (!(png_ptr->color_type&PNG_COLOR_MASK_COLOR))
+   {
+      png_warning(png_ptr,
+        "Ignoring PLTE chunk in grayscale PNG");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+#if !defined(PNG_READ_OPT_PLTE_SUPPORTED)
+   if (png_ptr->color_type != PNG_COLOR_TYPE_PALETTE)
+   {
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+#endif
+
+   if (length > 3*PNG_MAX_PALETTE_LENGTH || length % 3)
+   {
+      if (png_ptr->color_type != PNG_COLOR_TYPE_PALETTE)
+      {
+         png_warning(png_ptr, "Invalid palette chunk");
+         png_crc_finish(png_ptr, length);
+         return;
+      }
+      else
+      {
+         png_error(png_ptr, "Invalid palette chunk");
+      }
+   }
+
+   num = (int)length / 3;
+
+#ifndef PNG_NO_POINTER_INDEXING
+   for (i = 0, pal_ptr = palette; i < num; i++, pal_ptr++)
+   {
+      png_byte buf[3];
+
+      png_crc_read(png_ptr, buf, 3);
+      pal_ptr->red = buf[0];
+      pal_ptr->green = buf[1];
+      pal_ptr->blue = buf[2];
+   }
+#else
+   for (i = 0; i < num; i++)
+   {
+      png_byte buf[3];
+
+      png_crc_read(png_ptr, buf, 3);
+      /* don't depend upon png_color being any order */
+      palette[i].red = buf[0];
+      palette[i].green = buf[1];
+      palette[i].blue = buf[2];
+   }
+#endif
+
+   /* If we actually NEED the PLTE chunk (ie for a paletted image), we do
+      whatever the normal CRC configuration tells us.  However, if we
+      have an RGB image, the PLTE can be considered ancillary, so
+      we will act as though it is. */
+#if !defined(PNG_READ_OPT_PLTE_SUPPORTED)
+   if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+#endif
+   {
+      png_crc_finish(png_ptr, 0);
+   }
+#if !defined(PNG_READ_OPT_PLTE_SUPPORTED)
+   else if (png_crc_error(png_ptr))  /* Only if we have a CRC error */
+   {
+      /* If we don't want to use the data from an ancillary chunk,
+         we have two options: an error abort, or a warning and we
+         ignore the data in this chunk (which should be OK, since
+         it's considered ancillary for a RGB or RGBA image). */
+      if (!(png_ptr->flags & PNG_FLAG_CRC_ANCILLARY_USE))
+      {
+         if (png_ptr->flags & PNG_FLAG_CRC_ANCILLARY_NOWARN)
+         {
+            png_chunk_error(png_ptr, "CRC error");
+         }
+         else
+         {
+            png_chunk_warning(png_ptr, "CRC error");
+            return;
+         }
+      }
+      /* Otherwise, we (optionally) emit a warning and use the chunk. */
+      else if (!(png_ptr->flags & PNG_FLAG_CRC_ANCILLARY_NOWARN))
+      {
+         png_chunk_warning(png_ptr, "CRC error");
+      }
+   }
+#endif
+
+   png_set_PLTE(png_ptr, info_ptr, palette, num);
+
+#if defined(PNG_READ_tRNS_SUPPORTED)
+   if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+   {
+      if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_tRNS))
+      {
+         if (png_ptr->num_trans > (png_uint_16)num)
+         {
+            png_warning(png_ptr, "Truncating incorrect tRNS chunk length");
+            png_ptr->num_trans = (png_uint_16)num;
+         }
+         if (info_ptr->num_trans > (png_uint_16)num)
+         {
+            png_warning(png_ptr, "Truncating incorrect info tRNS chunk length");
+            info_ptr->num_trans = (png_uint_16)num;
+         }
+      }
+   }
+#endif
+
+}
+
+void /* PRIVATE */
+png_handle_IEND(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_debug(1, "in png_handle_IEND\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR) || !(png_ptr->mode & PNG_HAVE_IDAT))
+   {
+      png_error(png_ptr, "No image in file");
+   }
+
+   png_ptr->mode |= (PNG_AFTER_IDAT | PNG_HAVE_IEND);
+
+   if (length != 0)
+   {
+      png_warning(png_ptr, "Incorrect IEND chunk length");
+   }
+   png_crc_finish(png_ptr, length);
+
+   info_ptr =info_ptr; /* quiet compiler warnings about unused info_ptr */
+}
+
+#if defined(PNG_READ_gAMA_SUPPORTED)
+void /* PRIVATE */
+png_handle_gAMA(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_fixed_point igamma;
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   float file_gamma;
+#endif
+   png_byte buf[4];
+
+   png_debug(1, "in png_handle_gAMA\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before gAMA");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid gAMA after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (png_ptr->mode & PNG_HAVE_PLTE)
+      /* Should be an error, but we can cope with it */
+      png_warning(png_ptr, "Out of place gAMA chunk");
+
+   if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_gAMA)
+#if defined(PNG_READ_sRGB_SUPPORTED)
+      && !(info_ptr->valid & PNG_INFO_sRGB)
+#endif
+      )
+   {
+      png_warning(png_ptr, "Duplicate gAMA chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (length != 4)
+   {
+      png_warning(png_ptr, "Incorrect gAMA chunk length");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_crc_read(png_ptr, buf, 4);
+   if (png_crc_finish(png_ptr, 0))
+      return;
+
+   igamma = (png_fixed_point)png_get_uint_32(buf);
+   /* check for zero gamma */
+   if (igamma == 0)
+      {
+         png_warning(png_ptr,
+           "Ignoring gAMA chunk with gamma=0");
+         return;
+      }
+
+#if defined(PNG_READ_sRGB_SUPPORTED)
+   if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_sRGB))
+      if (PNG_OUT_OF_RANGE(igamma, 45500L, 500))
+      {
+         png_warning(png_ptr,
+           "Ignoring incorrect gAMA value when sRGB is also present");
+#ifndef PNG_NO_CONSOLE_IO
+         fprintf(stderr, "gamma = (%d/100000)\n", (int)igamma);
+#endif
+         return;
+      }
+#endif /* PNG_READ_sRGB_SUPPORTED */
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   file_gamma = (float)igamma / (float)100000.0;
+#  ifdef PNG_READ_GAMMA_SUPPORTED
+     png_ptr->gamma = file_gamma;
+#  endif
+     png_set_gAMA(png_ptr, info_ptr, file_gamma);
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   png_set_gAMA_fixed(png_ptr, info_ptr, igamma);
+#endif
+}
+#endif
+
+#if defined(PNG_READ_sBIT_SUPPORTED)
+void /* PRIVATE */
+png_handle_sBIT(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_size_t truelen;
+   png_byte buf[4];
+
+   png_debug(1, "in png_handle_sBIT\n");
+
+   buf[0] = buf[1] = buf[2] = buf[3] = 0;
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before sBIT");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid sBIT after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (png_ptr->mode & PNG_HAVE_PLTE)
+   {
+      /* Should be an error, but we can cope with it */
+      png_warning(png_ptr, "Out of place sBIT chunk");
+   }
+   if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_sBIT))
+   {
+      png_warning(png_ptr, "Duplicate sBIT chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      truelen = 3;
+   else
+      truelen = (png_size_t)png_ptr->channels;
+
+   if (length != truelen || length > 4)
+   {
+      png_warning(png_ptr, "Incorrect sBIT chunk length");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_crc_read(png_ptr, buf, truelen);
+   if (png_crc_finish(png_ptr, 0))
+      return;
+
+   if (png_ptr->color_type & PNG_COLOR_MASK_COLOR)
+   {
+      png_ptr->sig_bit.red = buf[0];
+      png_ptr->sig_bit.green = buf[1];
+      png_ptr->sig_bit.blue = buf[2];
+      png_ptr->sig_bit.alpha = buf[3];
+   }
+   else
+   {
+      png_ptr->sig_bit.gray = buf[0];
+      png_ptr->sig_bit.red = buf[0];
+      png_ptr->sig_bit.green = buf[0];
+      png_ptr->sig_bit.blue = buf[0];
+      png_ptr->sig_bit.alpha = buf[1];
+   }
+   png_set_sBIT(png_ptr, info_ptr, &(png_ptr->sig_bit));
+}
+#endif
+
+#if defined(PNG_READ_cHRM_SUPPORTED)
+void /* PRIVATE */
+png_handle_cHRM(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_byte buf[4];
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   float white_x, white_y, red_x, red_y, green_x, green_y, blue_x, blue_y;
+#endif
+   png_fixed_point int_x_white, int_y_white, int_x_red, int_y_red, int_x_green,
+      int_y_green, int_x_blue, int_y_blue;
+
+   png_uint_32 uint_x, uint_y;
+
+   png_debug(1, "in png_handle_cHRM\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before cHRM");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid cHRM after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (png_ptr->mode & PNG_HAVE_PLTE)
+      /* Should be an error, but we can cope with it */
+      png_warning(png_ptr, "Missing PLTE before cHRM");
+
+   if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_cHRM)
+#if defined(PNG_READ_sRGB_SUPPORTED)
+      && !(info_ptr->valid & PNG_INFO_sRGB)
+#endif
+      )
+   {
+      png_warning(png_ptr, "Duplicate cHRM chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (length != 32)
+   {
+      png_warning(png_ptr, "Incorrect cHRM chunk length");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_crc_read(png_ptr, buf, 4);
+   uint_x = png_get_uint_32(buf);
+
+   png_crc_read(png_ptr, buf, 4);
+   uint_y = png_get_uint_32(buf);
+
+   if (uint_x > 80000L || uint_y > 80000L ||
+      uint_x + uint_y > 100000L)
+   {
+      png_warning(png_ptr, "Invalid cHRM white point");
+      png_crc_finish(png_ptr, 24);
+      return;
+   }
+   int_x_white = (png_fixed_point)uint_x;
+   int_y_white = (png_fixed_point)uint_y;
+
+   png_crc_read(png_ptr, buf, 4);
+   uint_x = png_get_uint_32(buf);
+
+   png_crc_read(png_ptr, buf, 4);
+   uint_y = png_get_uint_32(buf);
+
+   if (uint_x + uint_y > 100000L)
+   {
+      png_warning(png_ptr, "Invalid cHRM red point");
+      png_crc_finish(png_ptr, 16);
+      return;
+   }
+   int_x_red = (png_fixed_point)uint_x;
+   int_y_red = (png_fixed_point)uint_y;
+
+   png_crc_read(png_ptr, buf, 4);
+   uint_x = png_get_uint_32(buf);
+
+   png_crc_read(png_ptr, buf, 4);
+   uint_y = png_get_uint_32(buf);
+
+   if (uint_x + uint_y > 100000L)
+   {
+      png_warning(png_ptr, "Invalid cHRM green point");
+      png_crc_finish(png_ptr, 8);
+      return;
+   }
+   int_x_green = (png_fixed_point)uint_x;
+   int_y_green = (png_fixed_point)uint_y;
+
+   png_crc_read(png_ptr, buf, 4);
+   uint_x = png_get_uint_32(buf);
+
+   png_crc_read(png_ptr, buf, 4);
+   uint_y = png_get_uint_32(buf);
+
+   if (uint_x + uint_y > 100000L)
+   {
+      png_warning(png_ptr, "Invalid cHRM blue point");
+      png_crc_finish(png_ptr, 0);
+      return;
+   }
+   int_x_blue = (png_fixed_point)uint_x;
+   int_y_blue = (png_fixed_point)uint_y;
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   white_x = (float)int_x_white / (float)100000.0;
+   white_y = (float)int_y_white / (float)100000.0;
+   red_x   = (float)int_x_red   / (float)100000.0;
+   red_y   = (float)int_y_red   / (float)100000.0;
+   green_x = (float)int_x_green / (float)100000.0;
+   green_y = (float)int_y_green / (float)100000.0;
+   blue_x  = (float)int_x_blue  / (float)100000.0;
+   blue_y  = (float)int_y_blue  / (float)100000.0;
+#endif
+
+#if defined(PNG_READ_sRGB_SUPPORTED)
+   if ((info_ptr != NULL) && (info_ptr->valid & PNG_INFO_sRGB))
+      {
+      if (PNG_OUT_OF_RANGE(int_x_white, 31270,  1000) ||
+          PNG_OUT_OF_RANGE(int_y_white, 32900,  1000) ||
+          PNG_OUT_OF_RANGE(int_x_red,   64000L, 1000) ||
+          PNG_OUT_OF_RANGE(int_y_red,   33000,  1000) ||
+          PNG_OUT_OF_RANGE(int_x_green, 30000,  1000) ||
+          PNG_OUT_OF_RANGE(int_y_green, 60000L, 1000) ||
+          PNG_OUT_OF_RANGE(int_x_blue,  15000,  1000) ||
+          PNG_OUT_OF_RANGE(int_y_blue,   6000,  1000))
+         {
+            png_warning(png_ptr,
+              "Ignoring incorrect cHRM value when sRGB is also present");
+#ifndef PNG_NO_CONSOLE_IO
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+            fprintf(stderr,"wx=%f, wy=%f, rx=%f, ry=%f\n",
+               white_x, white_y, red_x, red_y);
+            fprintf(stderr,"gx=%f, gy=%f, bx=%f, by=%f\n",
+               green_x, green_y, blue_x, blue_y);
+#else
+            fprintf(stderr,"wx=%ld, wy=%ld, rx=%ld, ry=%ld\n",
+               int_x_white, int_y_white, int_x_red, int_y_red);
+            fprintf(stderr,"gx=%ld, gy=%ld, bx=%ld, by=%ld\n",
+               int_x_green, int_y_green, int_x_blue, int_y_blue);
+#endif
+#endif /* PNG_NO_CONSOLE_IO */
+         }
+         png_crc_finish(png_ptr, 0);
+         return;
+      }
+#endif /* PNG_READ_sRGB_SUPPORTED */
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   png_set_cHRM(png_ptr, info_ptr,
+      white_x, white_y, red_x, red_y, green_x, green_y, blue_x, blue_y);
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   png_set_cHRM_fixed(png_ptr, info_ptr,
+      int_x_white, int_y_white, int_x_red, int_y_red, int_x_green,
+      int_y_green, int_x_blue, int_y_blue);
+#endif
+   if (png_crc_finish(png_ptr, 0))
+      return;
+}
+#endif
+
+#if defined(PNG_READ_sRGB_SUPPORTED)
+void /* PRIVATE */
+png_handle_sRGB(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   int intent;
+   png_byte buf[1];
+
+   png_debug(1, "in png_handle_sRGB\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before sRGB");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid sRGB after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (png_ptr->mode & PNG_HAVE_PLTE)
+      /* Should be an error, but we can cope with it */
+      png_warning(png_ptr, "Out of place sRGB chunk");
+
+   if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_sRGB))
+   {
+      png_warning(png_ptr, "Duplicate sRGB chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (length != 1)
+   {
+      png_warning(png_ptr, "Incorrect sRGB chunk length");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_crc_read(png_ptr, buf, 1);
+   if (png_crc_finish(png_ptr, 0))
+      return;
+
+   intent = buf[0];
+   /* check for bad intent */
+   if (intent >= PNG_sRGB_INTENT_LAST)
+   {
+      png_warning(png_ptr, "Unknown sRGB intent");
+      return;
+   }
+
+#if defined(PNG_READ_gAMA_SUPPORTED) && defined(PNG_READ_GAMMA_SUPPORTED)
+   if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_gAMA))
+   {
+   png_fixed_point igamma;
+#ifdef PNG_FIXED_POINT_SUPPORTED
+      igamma=info_ptr->int_gamma;
+#else
+#  ifdef PNG_FLOATING_POINT_SUPPORTED
+      igamma=(png_fixed_point)(info_ptr->gamma * 100000.);
+#  endif
+#endif
+      if (PNG_OUT_OF_RANGE(igamma, 45500L, 500))
+      {
+         png_warning(png_ptr,
+           "Ignoring incorrect gAMA value when sRGB is also present");
+#ifndef PNG_NO_CONSOLE_IO
+#  ifdef PNG_FIXED_POINT_SUPPORTED
+         fprintf(stderr,"incorrect gamma=(%d/100000)\n",(int)png_ptr->int_gamma);
+#  else
+#    ifdef PNG_FLOATING_POINT_SUPPORTED
+         fprintf(stderr,"incorrect gamma=%f\n",png_ptr->gamma);
+#    endif
+#  endif
+#endif
+      }
+   }
+#endif /* PNG_READ_gAMA_SUPPORTED */
+
+#ifdef PNG_READ_cHRM_SUPPORTED
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_cHRM))
+      if (PNG_OUT_OF_RANGE(info_ptr->int_x_white, 31270,  1000) ||
+          PNG_OUT_OF_RANGE(info_ptr->int_y_white, 32900,  1000) ||
+          PNG_OUT_OF_RANGE(info_ptr->int_x_red,   64000L, 1000) ||
+          PNG_OUT_OF_RANGE(info_ptr->int_y_red,   33000,  1000) ||
+          PNG_OUT_OF_RANGE(info_ptr->int_x_green, 30000,  1000) ||
+          PNG_OUT_OF_RANGE(info_ptr->int_y_green, 60000L, 1000) ||
+          PNG_OUT_OF_RANGE(info_ptr->int_x_blue,  15000,  1000) ||
+          PNG_OUT_OF_RANGE(info_ptr->int_y_blue,   6000,  1000))
+         {
+            png_warning(png_ptr,
+              "Ignoring incorrect cHRM value when sRGB is also present");
+         }
+#endif /* PNG_FIXED_POINT_SUPPORTED */
+#endif /* PNG_READ_cHRM_SUPPORTED */
+
+   png_set_sRGB_gAMA_and_cHRM(png_ptr, info_ptr, intent);
+}
+#endif /* PNG_READ_sRGB_SUPPORTED */
+
+#if defined(PNG_READ_iCCP_SUPPORTED)
+void /* PRIVATE */
+png_handle_iCCP(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+/* Note: this does not properly handle chunks that are > 64K under DOS */
+{
+   png_charp chunkdata;
+   png_byte compression_type;
+   png_bytep pC;
+   png_charp profile;
+   png_uint_32 skip = 0;
+   png_uint_32 profile_size, profile_length;
+   png_size_t slength, prefix_length, data_length;
+
+   png_debug(1, "in png_handle_iCCP\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before iCCP");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid iCCP after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (png_ptr->mode & PNG_HAVE_PLTE)
+      /* Should be an error, but we can cope with it */
+      png_warning(png_ptr, "Out of place iCCP chunk");
+
+   if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_iCCP))
+   {
+      png_warning(png_ptr, "Duplicate iCCP chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+#ifdef PNG_MAX_MALLOC_64K
+   if (length > (png_uint_32)65535L)
+   {
+      png_warning(png_ptr, "iCCP chunk too large to fit in memory");
+      skip = length - (png_uint_32)65535L;
+      length = (png_uint_32)65535L;
+   }
+#endif
+
+   chunkdata = (png_charp)png_malloc(png_ptr, length + 1);
+   slength = (png_size_t)length;
+   png_crc_read(png_ptr, (png_bytep)chunkdata, slength);
+
+   if (png_crc_finish(png_ptr, skip))
+   {
+      png_free(png_ptr, chunkdata);
+      return;
+   }
+
+   chunkdata[slength] = 0x00;
+
+   for (profile = chunkdata; *profile; profile++)
+      /* empty loop to find end of name */ ;
+
+   ++profile;
+
+   /* there should be at least one zero (the compression type byte)
+      following the separator, and we should be on it  */
+   if ( profile >= chunkdata + slength - 1)
+   {
+      png_free(png_ptr, chunkdata);
+      png_warning(png_ptr, "Malformed iCCP chunk");
+      return;
+   }
+
+   /* compression_type should always be zero */
+   compression_type = *profile++;
+   if (compression_type)
+   {
+      png_warning(png_ptr, "Ignoring nonzero compression type in iCCP chunk");
+      compression_type=0x00;  /* Reset it to zero (libpng-1.0.6 through 1.0.8
+                                 wrote nonzero) */
+   }
+
+   prefix_length = profile - chunkdata;
+   chunkdata = png_decompress_chunk(png_ptr, compression_type, chunkdata,
+                                    slength, prefix_length, &data_length);
+
+   profile_length = data_length - prefix_length;
+
+   if ( prefix_length > data_length || profile_length < 4)
+   {
+      png_free(png_ptr, chunkdata);
+      png_warning(png_ptr, "Profile size field missing from iCCP chunk");
+      return;
+   }
+
+   /* Check the profile_size recorded in the first 32 bits of the ICC profile */
+   pC = (png_bytep)(chunkdata+prefix_length);
+   profile_size = ((*(pC  ))<<24) |
+                  ((*(pC+1))<<16) |
+                  ((*(pC+2))<< 8) |
+                  ((*(pC+3))    );
+
+   if(profile_size < profile_length)
+      profile_length = profile_size;
+
+   if(profile_size > profile_length)
+   {
+      png_free(png_ptr, chunkdata);
+      png_warning(png_ptr, "Ignoring truncated iCCP profile.");
+      return;
+   }
+
+   png_set_iCCP(png_ptr, info_ptr, chunkdata, compression_type,
+                chunkdata + prefix_length, profile_length);
+   png_free(png_ptr, chunkdata);
+}
+#endif /* PNG_READ_iCCP_SUPPORTED */
+
+#if defined(PNG_READ_sPLT_SUPPORTED)
+void /* PRIVATE */
+png_handle_sPLT(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+/* Note: this does not properly handle chunks that are > 64K under DOS */
+{
+   png_bytep chunkdata;
+   png_bytep entry_start;
+   png_sPLT_t new_palette;
+#ifdef PNG_NO_POINTER_INDEXING
+   png_sPLT_entryp pp;
+#endif
+   int data_length, entry_size, i;
+   png_uint_32 skip = 0;
+   png_size_t slength;
+
+   png_debug(1, "in png_handle_sPLT\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before sPLT");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid sPLT after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+#ifdef PNG_MAX_MALLOC_64K
+   if (length > (png_uint_32)65535L)
+   {
+      png_warning(png_ptr, "sPLT chunk too large to fit in memory");
+      skip = length - (png_uint_32)65535L;
+      length = (png_uint_32)65535L;
+   }
+#endif
+
+   chunkdata = (png_bytep)png_malloc(png_ptr, length + 1);
+   slength = (png_size_t)length;
+   png_crc_read(png_ptr, (png_bytep)chunkdata, slength);
+
+   if (png_crc_finish(png_ptr, skip))
+   {
+      png_free(png_ptr, chunkdata);
+      return;
+   }
+
+   chunkdata[slength] = 0x00;
+
+   for (entry_start = chunkdata; *entry_start; entry_start++)
+      /* empty loop to find end of name */ ;
+   ++entry_start;
+
+   /* a sample depth should follow the separator, and we should be on it  */
+   if (entry_start > chunkdata + slength - 2)
+   {
+      png_free(png_ptr, chunkdata);
+      png_warning(png_ptr, "malformed sPLT chunk");
+      return;
+   }
+
+   new_palette.depth = *entry_start++;
+   entry_size = (new_palette.depth == 8 ? 6 : 10);
+   data_length = (slength - (entry_start - chunkdata));
+
+   /* integrity-check the data length */
+   if (data_length % entry_size)
+   {
+      png_free(png_ptr, chunkdata);
+      png_warning(png_ptr, "sPLT chunk has bad length");
+      return;
+   }
+
+   new_palette.nentries = (png_int_32) ( data_length / entry_size);
+   if ((png_uint_32) new_palette.nentries > (png_uint_32) (PNG_SIZE_MAX /
+       png_sizeof(png_sPLT_entry)))
+   {
+       png_warning(png_ptr, "sPLT chunk too long");
+       return;
+   }
+   new_palette.entries = (png_sPLT_entryp)png_malloc_warn(
+       png_ptr, new_palette.nentries * png_sizeof(png_sPLT_entry));
+   if (new_palette.entries == NULL)
+   {
+       png_warning(png_ptr, "sPLT chunk requires too much memory");
+       return;
+   }
+
+#ifndef PNG_NO_POINTER_INDEXING
+   for (i = 0; i < new_palette.nentries; i++)
+   {
+      png_sPLT_entryp pp = new_palette.entries + i;
+
+      if (new_palette.depth == 8)
+      {
+          pp->red = *entry_start++;
+          pp->green = *entry_start++;
+          pp->blue = *entry_start++;
+          pp->alpha = *entry_start++;
+      }
+      else
+      {
+          pp->red   = png_get_uint_16(entry_start); entry_start += 2;
+          pp->green = png_get_uint_16(entry_start); entry_start += 2;
+          pp->blue  = png_get_uint_16(entry_start); entry_start += 2;
+          pp->alpha = png_get_uint_16(entry_start); entry_start += 2;
+      }
+      pp->frequency = png_get_uint_16(entry_start); entry_start += 2;
+   }
+#else
+   pp = new_palette.entries;
+   for (i = 0; i < new_palette.nentries; i++)
+   {
+
+      if (new_palette.depth == 8)
+      {
+          pp[i].red   = *entry_start++;
+          pp[i].green = *entry_start++;
+          pp[i].blue  = *entry_start++;
+          pp[i].alpha = *entry_start++;
+      }
+      else
+      {
+          pp[i].red   = png_get_uint_16(entry_start); entry_start += 2;
+          pp[i].green = png_get_uint_16(entry_start); entry_start += 2;
+          pp[i].blue  = png_get_uint_16(entry_start); entry_start += 2;
+          pp[i].alpha = png_get_uint_16(entry_start); entry_start += 2;
+      }
+      pp->frequency = png_get_uint_16(entry_start); entry_start += 2;
+   }
+#endif
+
+   /* discard all chunk data except the name and stash that */
+   new_palette.name = (png_charp)chunkdata;
+
+   png_set_sPLT(png_ptr, info_ptr, &new_palette, 1);
+
+   png_free(png_ptr, chunkdata);
+   png_free(png_ptr, new_palette.entries);
+}
+#endif /* PNG_READ_sPLT_SUPPORTED */
+
+#if defined(PNG_READ_tRNS_SUPPORTED)
+void /* PRIVATE */
+png_handle_tRNS(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_byte readbuf[PNG_MAX_PALETTE_LENGTH];
+   int bit_mask;
+
+   png_debug(1, "in png_handle_tRNS\n");
+
+   /* For non-indexed color, mask off any bits in the tRNS value that
+    * exceed the bit depth.  Some creators were writing extra bits there.
+    * This is not needed for indexed color. */
+   bit_mask = (1 << png_ptr->bit_depth) - 1;
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before tRNS");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid tRNS after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_tRNS))
+   {
+      png_warning(png_ptr, "Duplicate tRNS chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (png_ptr->color_type == PNG_COLOR_TYPE_GRAY)
+   {
+      png_byte buf[2];
+
+      if (length != 2)
+      {
+         png_warning(png_ptr, "Incorrect tRNS chunk length");
+         png_crc_finish(png_ptr, length);
+         return;
+      }
+
+      png_crc_read(png_ptr, buf, 2);
+      png_ptr->num_trans = 1;
+      png_ptr->trans_values.gray = png_get_uint_16(buf) & bit_mask;
+   }
+   else if (png_ptr->color_type == PNG_COLOR_TYPE_RGB)
+   {
+      png_byte buf[6];
+
+      if (length != 6)
+      {
+         png_warning(png_ptr, "Incorrect tRNS chunk length");
+         png_crc_finish(png_ptr, length);
+         return;
+      }
+      png_crc_read(png_ptr, buf, (png_size_t)length);
+      png_ptr->num_trans = 1;
+      png_ptr->trans_values.red = png_get_uint_16(buf) & bit_mask;
+      png_ptr->trans_values.green = png_get_uint_16(buf + 2) & bit_mask;
+      png_ptr->trans_values.blue = png_get_uint_16(buf + 4) & bit_mask;
+   }
+   else if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+   {
+      if (!(png_ptr->mode & PNG_HAVE_PLTE))
+      {
+         /* Should be an error, but we can cope with it. */
+         png_warning(png_ptr, "Missing PLTE before tRNS");
+      }
+      if (length > (png_uint_32)png_ptr->num_palette ||
+          length > PNG_MAX_PALETTE_LENGTH)
+      {
+         png_warning(png_ptr, "Incorrect tRNS chunk length");
+         png_crc_finish(png_ptr, length);
+         return;
+      }
+      if (length == 0)
+      {
+         png_warning(png_ptr, "Zero length tRNS chunk");
+         png_crc_finish(png_ptr, length);
+         return;
+      }
+      png_crc_read(png_ptr, readbuf, (png_size_t)length);
+      png_ptr->num_trans = (png_uint_16)length;
+   }
+   else
+   {
+      png_warning(png_ptr, "tRNS chunk not allowed with alpha channel");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (png_crc_finish(png_ptr, 0))
+   {
+      png_ptr->num_trans = 0;
+      return;
+   }
+
+   png_set_tRNS(png_ptr, info_ptr, readbuf, png_ptr->num_trans,
+      &(png_ptr->trans_values));
+}
+#endif
+
+#if defined(PNG_READ_bKGD_SUPPORTED)
+void /* PRIVATE */
+png_handle_bKGD(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_size_t truelen;
+   png_byte buf[6];
+
+   png_debug(1, "in png_handle_bKGD\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before bKGD");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid bKGD after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE &&
+            !(png_ptr->mode & PNG_HAVE_PLTE))
+   {
+      png_warning(png_ptr, "Missing PLTE before bKGD");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_bKGD))
+   {
+      png_warning(png_ptr, "Duplicate bKGD chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      truelen = 1;
+   else if (png_ptr->color_type & PNG_COLOR_MASK_COLOR)
+      truelen = 6;
+   else
+      truelen = 2;
+
+   if (length != truelen)
+   {
+      png_warning(png_ptr, "Incorrect bKGD chunk length");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_crc_read(png_ptr, buf, truelen);
+   if (png_crc_finish(png_ptr, 0))
+      return;
+
+   /* We convert the index value into RGB components so that we can allow
+    * arbitrary RGB values for background when we have transparency, and
+    * so it is easy to determine the RGB values of the background color
+    * from the info_ptr struct. */
+   if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+   {
+      png_ptr->background.index = buf[0];
+      if(info_ptr->num_palette)
+      {
+          if(buf[0] > info_ptr->num_palette)
+          {
+             png_warning(png_ptr, "Incorrect bKGD chunk index value");
+             return;
+          }
+          png_ptr->background.red =
+             (png_uint_16)png_ptr->palette[buf[0]].red;
+          png_ptr->background.green =
+             (png_uint_16)png_ptr->palette[buf[0]].green;
+          png_ptr->background.blue =
+             (png_uint_16)png_ptr->palette[buf[0]].blue;
+      }
+   }
+   else if (!(png_ptr->color_type & PNG_COLOR_MASK_COLOR)) /* GRAY */
+   {
+      png_ptr->background.red =
+      png_ptr->background.green =
+      png_ptr->background.blue =
+      png_ptr->background.gray = png_get_uint_16(buf);
+   }
+   else
+   {
+      png_ptr->background.red = png_get_uint_16(buf);
+      png_ptr->background.green = png_get_uint_16(buf + 2);
+      png_ptr->background.blue = png_get_uint_16(buf + 4);
+   }
+
+   png_set_bKGD(png_ptr, info_ptr, &(png_ptr->background));
+}
+#endif
+
+#if defined(PNG_READ_hIST_SUPPORTED)
+void /* PRIVATE */
+png_handle_hIST(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   unsigned int num, i;
+   png_uint_16 readbuf[PNG_MAX_PALETTE_LENGTH];
+
+   png_debug(1, "in png_handle_hIST\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before hIST");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid hIST after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (!(png_ptr->mode & PNG_HAVE_PLTE))
+   {
+      png_warning(png_ptr, "Missing PLTE before hIST");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_hIST))
+   {
+      png_warning(png_ptr, "Duplicate hIST chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   num = length / 2 ;
+   if (num != (unsigned int) png_ptr->num_palette || num >
+      (unsigned int) PNG_MAX_PALETTE_LENGTH)
+   {
+      png_warning(png_ptr, "Incorrect hIST chunk length");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   for (i = 0; i < num; i++)
+   {
+      png_byte buf[2];
+
+      png_crc_read(png_ptr, buf, 2);
+      readbuf[i] = png_get_uint_16(buf);
+   }
+
+   if (png_crc_finish(png_ptr, 0))
+      return;
+
+   png_set_hIST(png_ptr, info_ptr, readbuf);
+}
+#endif
+
+#if defined(PNG_READ_pHYs_SUPPORTED)
+void /* PRIVATE */
+png_handle_pHYs(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_byte buf[9];
+   png_uint_32 res_x, res_y;
+   int unit_type;
+
+   png_debug(1, "in png_handle_pHYs\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before pHYs");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid pHYs after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_pHYs))
+   {
+      png_warning(png_ptr, "Duplicate pHYs chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (length != 9)
+   {
+      png_warning(png_ptr, "Incorrect pHYs chunk length");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_crc_read(png_ptr, buf, 9);
+   if (png_crc_finish(png_ptr, 0))
+      return;
+
+   res_x = png_get_uint_32(buf);
+   res_y = png_get_uint_32(buf + 4);
+   unit_type = buf[8];
+   png_set_pHYs(png_ptr, info_ptr, res_x, res_y, unit_type);
+}
+#endif
+
+#if defined(PNG_READ_oFFs_SUPPORTED)
+void /* PRIVATE */
+png_handle_oFFs(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_byte buf[9];
+   png_int_32 offset_x, offset_y;
+   int unit_type;
+
+   png_debug(1, "in png_handle_oFFs\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before oFFs");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid oFFs after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_oFFs))
+   {
+      png_warning(png_ptr, "Duplicate oFFs chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (length != 9)
+   {
+      png_warning(png_ptr, "Incorrect oFFs chunk length");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_crc_read(png_ptr, buf, 9);
+   if (png_crc_finish(png_ptr, 0))
+      return;
+
+   offset_x = png_get_int_32(buf);
+   offset_y = png_get_int_32(buf + 4);
+   unit_type = buf[8];
+   png_set_oFFs(png_ptr, info_ptr, offset_x, offset_y, unit_type);
+}
+#endif
+
+#if defined(PNG_READ_pCAL_SUPPORTED)
+/* read the pCAL chunk (described in the PNG Extensions document) */
+void /* PRIVATE */
+png_handle_pCAL(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_charp purpose;
+   png_int_32 X0, X1;
+   png_byte type, nparams;
+   png_charp buf, units, endptr;
+   png_charpp params;
+   png_size_t slength;
+   int i;
+
+   png_debug(1, "in png_handle_pCAL\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before pCAL");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid pCAL after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_pCAL))
+   {
+      png_warning(png_ptr, "Duplicate pCAL chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_debug1(2, "Allocating and reading pCAL chunk data (%lu bytes)\n",
+      length + 1);
+   purpose = (png_charp)png_malloc_warn(png_ptr, length + 1);
+   if (purpose == NULL)
+     {
+       png_warning(png_ptr, "No memory for pCAL purpose.");
+       return;
+     }
+   slength = (png_size_t)length;
+   png_crc_read(png_ptr, (png_bytep)purpose, slength);
+
+   if (png_crc_finish(png_ptr, 0))
+   {
+      png_free(png_ptr, purpose);
+      return;
+   }
+
+   purpose[slength] = 0x00; /* null terminate the last string */
+
+   png_debug(3, "Finding end of pCAL purpose string\n");
+   for (buf = purpose; *buf; buf++)
+      /* empty loop */ ;
+
+   endptr = purpose + slength;
+
+   /* We need to have at least 12 bytes after the purpose string
+      in order to get the parameter information. */
+   if (endptr <= buf + 12)
+   {
+      png_warning(png_ptr, "Invalid pCAL data");
+      png_free(png_ptr, purpose);
+      return;
+   }
+
+   png_debug(3, "Reading pCAL X0, X1, type, nparams, and units\n");
+   X0 = png_get_int_32((png_bytep)buf+1);
+   X1 = png_get_int_32((png_bytep)buf+5);
+   type = buf[9];
+   nparams = buf[10];
+   units = buf + 11;
+
+   png_debug(3, "Checking pCAL equation type and number of parameters\n");
+   /* Check that we have the right number of parameters for known
+      equation types. */
+   if ((type == PNG_EQUATION_LINEAR && nparams != 2) ||
+       (type == PNG_EQUATION_BASE_E && nparams != 3) ||
+       (type == PNG_EQUATION_ARBITRARY && nparams != 3) ||
+       (type == PNG_EQUATION_HYPERBOLIC && nparams != 4))
+   {
+      png_warning(png_ptr, "Invalid pCAL parameters for equation type");
+      png_free(png_ptr, purpose);
+      return;
+   }
+   else if (type >= PNG_EQUATION_LAST)
+   {
+      png_warning(png_ptr, "Unrecognized equation type for pCAL chunk");
+   }
+
+   for (buf = units; *buf; buf++)
+      /* Empty loop to move past the units string. */ ;
+
+   png_debug(3, "Allocating pCAL parameters array\n");
+   params = (png_charpp)png_malloc_warn(png_ptr, (png_uint_32)(nparams
+      *png_sizeof(png_charp))) ;
+   if (params == NULL)
+     {
+       png_free(png_ptr, purpose);
+       png_warning(png_ptr, "No memory for pCAL params.");
+       return;
+     }
+
+   /* Get pointers to the start of each parameter string. */
+   for (i = 0; i < (int)nparams; i++)
+   {
+      buf++; /* Skip the null string terminator from previous parameter. */
+
+      png_debug1(3, "Reading pCAL parameter %d\n", i);
+      for (params[i] = buf; *buf != 0x00 && buf <= endptr; buf++)
+         /* Empty loop to move past each parameter string */ ;
+
+      /* Make sure we haven't run out of data yet */
+      if (buf > endptr)
+      {
+         png_warning(png_ptr, "Invalid pCAL data");
+         png_free(png_ptr, purpose);
+         png_free(png_ptr, params);
+         return;
+      }
+   }
+
+   png_set_pCAL(png_ptr, info_ptr, purpose, X0, X1, type, nparams,
+      units, params);
+
+   png_free(png_ptr, purpose);
+   png_free(png_ptr, params);
+}
+#endif
+
+#if defined(PNG_READ_sCAL_SUPPORTED)
+/* read the sCAL chunk */
+void /* PRIVATE */
+png_handle_sCAL(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_charp buffer, ep;
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   double width, height;
+   png_charp vp;
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   png_charp swidth, sheight;
+#endif
+#endif
+   png_size_t slength;
+
+   png_debug(1, "in png_handle_sCAL\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before sCAL");
+   else if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+      png_warning(png_ptr, "Invalid sCAL after IDAT");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+   else if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_sCAL))
+   {
+      png_warning(png_ptr, "Duplicate sCAL chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_debug1(2, "Allocating and reading sCAL chunk data (%lu bytes)\n",
+      length + 1);
+   buffer = (png_charp)png_malloc_warn(png_ptr, length + 1);
+   if (buffer == NULL)
+     {
+       png_warning(png_ptr, "Out of memory while processing sCAL chunk");
+       return;
+     }
+   slength = (png_size_t)length;
+   png_crc_read(png_ptr, (png_bytep)buffer, slength);
+
+   if (png_crc_finish(png_ptr, 0))
+   {
+      png_free(png_ptr, buffer);
+      return;
+   }
+
+   buffer[slength] = 0x00; /* null terminate the last string */
+
+   ep = buffer + 1;        /* skip unit byte */
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   width = png_strtod(png_ptr, ep, &vp);
+   if (*vp)
+   {
+       png_warning(png_ptr, "malformed width string in sCAL chunk");
+       return;
+   }
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   swidth = (png_charp)png_malloc_warn(png_ptr, png_strlen(ep) + 1);
+   if (swidth == NULL)
+     {
+       png_warning(png_ptr, "Out of memory while processing sCAL chunk width");
+       return;
+     }
+   png_memcpy(swidth, ep, (png_size_t)png_strlen(ep));
+#endif
+#endif
+
+   for (ep = buffer; *ep; ep++)
+      /* empty loop */ ;
+   ep++;
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   height = png_strtod(png_ptr, ep, &vp);
+   if (*vp)
+   {
+       png_warning(png_ptr, "malformed height string in sCAL chunk");
+       return;
+   }
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   sheight = (png_charp)png_malloc_warn(png_ptr, png_strlen(ep) + 1);
+   if (swidth == NULL)
+     {
+       png_warning(png_ptr, "Out of memory while processing sCAL chunk height");
+       return;
+     }
+   png_memcpy(sheight, ep, (png_size_t)png_strlen(ep));
+#endif
+#endif
+
+   if (buffer + slength < ep
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+      || width <= 0. || height <= 0.
+#endif
+      )
+   {
+      png_warning(png_ptr, "Invalid sCAL data");
+      png_free(png_ptr, buffer);
+#if defined(PNG_FIXED_POINT_SUPPORTED) && !defined(PNG_FLOATING_POINT_SUPPORTED)
+      png_free(png_ptr, swidth);
+      png_free(png_ptr, sheight);
+#endif
+      return;
+   }
+
+
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   png_set_sCAL(png_ptr, info_ptr, buffer[0], width, height);
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   png_set_sCAL_s(png_ptr, info_ptr, buffer[0], swidth, sheight);
+#endif
+#endif
+
+   png_free(png_ptr, buffer);
+#if defined(PNG_FIXED_POINT_SUPPORTED) && !defined(PNG_FLOATING_POINT_SUPPORTED)
+   png_free(png_ptr, swidth);
+   png_free(png_ptr, sheight);
+#endif
+}
+#endif
+
+#if defined(PNG_READ_tIME_SUPPORTED)
+void /* PRIVATE */
+png_handle_tIME(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_byte buf[7];
+   png_time mod_time;
+
+   png_debug(1, "in png_handle_tIME\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Out of place tIME chunk");
+   else if (info_ptr != NULL && (info_ptr->valid & PNG_INFO_tIME))
+   {
+      png_warning(png_ptr, "Duplicate tIME chunk");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   if (png_ptr->mode & PNG_HAVE_IDAT)
+      png_ptr->mode |= PNG_AFTER_IDAT;
+
+   if (length != 7)
+   {
+      png_warning(png_ptr, "Incorrect tIME chunk length");
+      png_crc_finish(png_ptr, length);
+      return;
+   }
+
+   png_crc_read(png_ptr, buf, 7);
+   if (png_crc_finish(png_ptr, 0))
+      return;
+
+   mod_time.second = buf[6];
+   mod_time.minute = buf[5];
+   mod_time.hour = buf[4];
+   mod_time.day = buf[3];
+   mod_time.month = buf[2];
+   mod_time.year = png_get_uint_16(buf);
+
+   png_set_tIME(png_ptr, info_ptr, &mod_time);
+}
+#endif
+
+#if defined(PNG_READ_tEXt_SUPPORTED)
+/* Note: this does not properly handle chunks that are > 64K under DOS */
+void /* PRIVATE */
+png_handle_tEXt(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_textp text_ptr;
+   png_charp key;
+   png_charp text;
+   png_uint_32 skip = 0;
+   png_size_t slength;
+   int ret;
+
+   png_debug(1, "in png_handle_tEXt\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before tEXt");
+
+   if (png_ptr->mode & PNG_HAVE_IDAT)
+      png_ptr->mode |= PNG_AFTER_IDAT;
+
+#ifdef PNG_MAX_MALLOC_64K
+   if (length > (png_uint_32)65535L)
+   {
+      png_warning(png_ptr, "tEXt chunk too large to fit in memory");
+      skip = length - (png_uint_32)65535L;
+      length = (png_uint_32)65535L;
+   }
+#endif
+
+   key = (png_charp)png_malloc_warn(png_ptr, length + 1);
+   if (key == NULL)
+   {
+     png_warning(png_ptr, "No memory to process text chunk.");
+     return;
+   }
+   slength = (png_size_t)length;
+   png_crc_read(png_ptr, (png_bytep)key, slength);
+
+   if (png_crc_finish(png_ptr, skip))
+   {
+      png_free(png_ptr, key);
+      return;
+   }
+
+   key[slength] = 0x00;
+
+   for (text = key; *text; text++)
+      /* empty loop to find end of key */ ;
+
+   if (text != key + slength)
+      text++;
+
+   text_ptr = (png_textp)png_malloc_warn(png_ptr,
+      (png_uint_32)png_sizeof(png_text));
+   if (text_ptr == NULL)
+   {
+     png_warning(png_ptr, "Not enough memory to process text chunk.");
+     png_free(png_ptr, key);
+     return;
+   }
+   text_ptr->compression = PNG_TEXT_COMPRESSION_NONE;
+   text_ptr->key = key;
+#ifdef PNG_iTXt_SUPPORTED
+   text_ptr->lang = NULL;
+   text_ptr->lang_key = NULL;
+   text_ptr->itxt_length = 0;
+#endif
+   text_ptr->text = text;
+   text_ptr->text_length = png_strlen(text);
+
+   ret=png_set_text_2(png_ptr, info_ptr, text_ptr, 1);
+
+   png_free(png_ptr, key);
+   png_free(png_ptr, text_ptr);
+   if (ret)
+     png_warning(png_ptr, "Insufficient memory to process text chunk.");
+}
+#endif
+
+#if defined(PNG_READ_zTXt_SUPPORTED)
+/* note: this does not correctly handle chunks that are > 64K under DOS */
+void /* PRIVATE */
+png_handle_zTXt(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_textp text_ptr;
+   png_charp chunkdata;
+   png_charp text;
+   int comp_type;
+   int ret;
+   png_size_t slength, prefix_len, data_len;
+
+   png_debug(1, "in png_handle_zTXt\n");
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before zTXt");
+
+   if (png_ptr->mode & PNG_HAVE_IDAT)
+      png_ptr->mode |= PNG_AFTER_IDAT;
+
+#ifdef PNG_MAX_MALLOC_64K
+   /* We will no doubt have problems with chunks even half this size, but
+      there is no hard and fast rule to tell us where to stop. */
+   if (length > (png_uint_32)65535L)
+   {
+     png_warning(png_ptr,"zTXt chunk too large to fit in memory");
+     png_crc_finish(png_ptr, length);
+     return;
+   }
+#endif
+
+   chunkdata = (png_charp)png_malloc_warn(png_ptr, length + 1);
+   if (chunkdata == NULL)
+   {
+     png_warning(png_ptr,"Out of memory processing zTXt chunk.");
+     return;
+   }
+   slength = (png_size_t)length;
+   png_crc_read(png_ptr, (png_bytep)chunkdata, slength);
+   if (png_crc_finish(png_ptr, 0))
+   {
+      png_free(png_ptr, chunkdata);
+      return;
+   }
+
+   chunkdata[slength] = 0x00;
+
+   for (text = chunkdata; *text; text++)
+      /* empty loop */ ;
+
+   /* zTXt must have some text after the chunkdataword */
+   if (text == chunkdata + slength - 1)
+   {
+      png_warning(png_ptr, "Truncated zTXt chunk");
+      png_free(png_ptr, chunkdata);
+      return;
+   }
+   else
+   {
+       comp_type = *(++text);
+       if (comp_type != PNG_TEXT_COMPRESSION_zTXt)
+       {
+          png_warning(png_ptr, "Unknown compression type in zTXt chunk");
+          comp_type = PNG_TEXT_COMPRESSION_zTXt;
+       }
+       text++;        /* skip the compression_method byte */
+   }
+   prefix_len = text - chunkdata;
+
+   chunkdata = (png_charp)png_decompress_chunk(png_ptr, comp_type, chunkdata,
+                                    (png_size_t)length, prefix_len, &data_len);
+
+   text_ptr = (png_textp)png_malloc_warn(png_ptr,
+     (png_uint_32)png_sizeof(png_text));
+   if (text_ptr == NULL)
+   {
+     png_warning(png_ptr,"Not enough memory to process zTXt chunk.");
+     png_free(png_ptr, chunkdata);
+     return;
+   }
+   text_ptr->compression = comp_type;
+   text_ptr->key = chunkdata;
+#ifdef PNG_iTXt_SUPPORTED
+   text_ptr->lang = NULL;
+   text_ptr->lang_key = NULL;
+   text_ptr->itxt_length = 0;
+#endif
+   text_ptr->text = chunkdata + prefix_len;
+   text_ptr->text_length = data_len;
+
+   ret=png_set_text_2(png_ptr, info_ptr, text_ptr, 1);
+
+   png_free(png_ptr, text_ptr);
+   png_free(png_ptr, chunkdata);
+   if (ret)
+     png_error(png_ptr, "Insufficient memory to store zTXt chunk.");
+}
+#endif
+
+#if defined(PNG_READ_iTXt_SUPPORTED)
+/* note: this does not correctly handle chunks that are > 64K under DOS */
+void /* PRIVATE */
+png_handle_iTXt(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_textp text_ptr;
+   png_charp chunkdata;
+   png_charp key, lang, text, lang_key;
+   int comp_flag;
+   int comp_type = 0;
+   int ret;
+   png_size_t slength, prefix_len, data_len;
+
+   png_debug(1, "in png_handle_iTXt\n");
+
+   if (!(png_ptr->mode & PNG_HAVE_IHDR))
+      png_error(png_ptr, "Missing IHDR before iTXt");
+
+   if (png_ptr->mode & PNG_HAVE_IDAT)
+      png_ptr->mode |= PNG_AFTER_IDAT;
+
+#ifdef PNG_MAX_MALLOC_64K
+   /* We will no doubt have problems with chunks even half this size, but
+      there is no hard and fast rule to tell us where to stop. */
+   if (length > (png_uint_32)65535L)
+   {
+     png_warning(png_ptr,"iTXt chunk too large to fit in memory");
+     png_crc_finish(png_ptr, length);
+     return;
+   }
+#endif
+
+   chunkdata = (png_charp)png_malloc_warn(png_ptr, length + 1);
+   if (chunkdata == NULL)
+   {
+     png_warning(png_ptr, "No memory to process iTXt chunk.");
+     return;
+   }
+   slength = (png_size_t)length;
+   png_crc_read(png_ptr, (png_bytep)chunkdata, slength);
+   if (png_crc_finish(png_ptr, 0))
+   {
+      png_free(png_ptr, chunkdata);
+      return;
+   }
+
+   chunkdata[slength] = 0x00;
+
+   for (lang = chunkdata; *lang; lang++)
+      /* empty loop */ ;
+   lang++;        /* skip NUL separator */
+
+   /* iTXt must have a language tag (possibly empty), two compression bytes,
+      translated keyword (possibly empty), and possibly some text after the
+      keyword */
+
+   if (lang >= chunkdata + slength - 3)
+   {
+      png_warning(png_ptr, "Truncated iTXt chunk");
+      png_free(png_ptr, chunkdata);
+      return;
+   }
+   else
+   {
+       comp_flag = *lang++;
+       comp_type = *lang++;
+   }
+
+   for (lang_key = lang; *lang_key; lang_key++)
+      /* empty loop */ ;
+   lang_key++;        /* skip NUL separator */
+
+   for (text = lang_key; *text; text++)
+      /* empty loop */ ;
+   text++;        /* skip NUL separator */
+   if (text >= chunkdata + slength)
+   {
+      png_warning(png_ptr, "Malformed iTXt chunk");
+      png_free(png_ptr, chunkdata);
+      return;
+   }
+
+   prefix_len = text - chunkdata;
+
+   key=chunkdata;
+   if (comp_flag)
+       chunkdata = png_decompress_chunk(png_ptr, comp_type, chunkdata,
+          (size_t)length, prefix_len, &data_len);
+   else
+       data_len=png_strlen(chunkdata + prefix_len);
+   text_ptr = (png_textp)png_malloc_warn(png_ptr,
+      (png_uint_32)png_sizeof(png_text));
+   if (text_ptr == NULL)
+   {
+     png_warning(png_ptr,"Not enough memory to process iTXt chunk.");
+     png_free(png_ptr, chunkdata);
+     return;
+   }
+   text_ptr->compression = (int)comp_flag + 1;
+   text_ptr->lang_key = chunkdata+(lang_key-key);
+   text_ptr->lang = chunkdata+(lang-key);
+   text_ptr->itxt_length = data_len;
+   text_ptr->text_length = 0;
+   text_ptr->key = chunkdata;
+   text_ptr->text = chunkdata + prefix_len;
+
+   ret=png_set_text_2(png_ptr, info_ptr, text_ptr, 1);
+
+   png_free(png_ptr, text_ptr);
+   png_free(png_ptr, chunkdata);
+   if (ret)
+     png_error(png_ptr, "Insufficient memory to store iTXt chunk.");
+}
+#endif
+
+/* This function is called when we haven't found a handler for a
+   chunk.  If there isn't a problem with the chunk itself (ie bad
+   chunk name, CRC, or a critical chunk), the chunk is silently ignored
+   -- unless the PNG_FLAG_UNKNOWN_CHUNKS_SUPPORTED flag is on in which
+   case it will be saved away to be written out later. */
+void /* PRIVATE */
+png_handle_unknown(png_structp png_ptr, png_infop info_ptr, png_uint_32 length)
+{
+   png_uint_32 skip = 0;
+
+   png_debug(1, "in png_handle_unknown\n");
+
+   if (png_ptr->mode & PNG_HAVE_IDAT)
+   {
+#ifdef PNG_USE_LOCAL_ARRAYS
+      PNG_CONST PNG_IDAT;
+#endif
+      if (png_memcmp(png_ptr->chunk_name, png_IDAT, 4))  /* not an IDAT */
+         png_ptr->mode |= PNG_AFTER_IDAT;
+   }
+
+   png_check_chunk_name(png_ptr, png_ptr->chunk_name);
+
+   if (!(png_ptr->chunk_name[0] & 0x20))
+   {
+#if defined(PNG_READ_UNKNOWN_CHUNKS_SUPPORTED)
+      if(png_handle_as_unknown(png_ptr, png_ptr->chunk_name) !=
+           PNG_HANDLE_CHUNK_ALWAYS
+#if defined(PNG_READ_USER_CHUNKS_SUPPORTED)
+           && png_ptr->read_user_chunk_fn == NULL
+#endif
+        )
+#endif
+          png_chunk_error(png_ptr, "unknown critical chunk");
+   }
+
+#if defined(PNG_READ_UNKNOWN_CHUNKS_SUPPORTED)
+   if ((png_ptr->flags & PNG_FLAG_KEEP_UNKNOWN_CHUNKS) ||
+       (png_ptr->read_user_chunk_fn != NULL))
+   {
+#ifdef PNG_MAX_MALLOC_64K
+       if (length > (png_uint_32)65535L)
+       {
+           png_warning(png_ptr, "unknown chunk too large to fit in memory");
+           skip = length - (png_uint_32)65535L;
+           length = (png_uint_32)65535L;
+       }
+#endif
+       png_strncpy((png_charp)png_ptr->unknown_chunk.name,
+	 (png_charp)png_ptr->chunk_name,
+         png_sizeof((png_charp)png_ptr->chunk_name));
+       png_ptr->unknown_chunk.data = (png_bytep)png_malloc(png_ptr, length);
+       png_ptr->unknown_chunk.size = (png_size_t)length;
+       png_crc_read(png_ptr, (png_bytep)png_ptr->unknown_chunk.data, length);
+#if defined(PNG_READ_USER_CHUNKS_SUPPORTED)
+       if(png_ptr->read_user_chunk_fn != NULL)
+       {
+          /* callback to user unknown chunk handler */
+          int ret;
+          ret = (*(png_ptr->read_user_chunk_fn))
+            (png_ptr, &png_ptr->unknown_chunk);
+          if (ret < 0)
+             png_chunk_error(png_ptr, "error in user chunk");
+          if (ret == 0)
+          {
+             if (!(png_ptr->chunk_name[0] & 0x20))
+                if(png_handle_as_unknown(png_ptr, png_ptr->chunk_name) !=
+                     PNG_HANDLE_CHUNK_ALWAYS)
+                   png_chunk_error(png_ptr, "unknown critical chunk");
+             png_set_unknown_chunks(png_ptr, info_ptr,
+               &png_ptr->unknown_chunk, 1);
+          }
+       }
+#else
+       png_set_unknown_chunks(png_ptr, info_ptr, &png_ptr->unknown_chunk, 1);
+#endif
+       png_free(png_ptr, png_ptr->unknown_chunk.data);
+       png_ptr->unknown_chunk.data = NULL;
+   }
+   else
+#endif
+      skip = length;
+
+   png_crc_finish(png_ptr, skip);
+
+#if !defined(PNG_READ_USER_CHUNKS_SUPPORTED)
+   info_ptr = info_ptr; /* quiet compiler warnings about unused info_ptr */
+#endif
+}
+
+/* This function is called to verify that a chunk name is valid.
+   This function can't have the "critical chunk check" incorporated
+   into it, since in the future we will need to be able to call user
+   functions to handle unknown critical chunks after we check that
+   the chunk name itself is valid. */
+
+#define isnonalpha(c) ((c) < 65 || (c) > 122 || ((c) > 90 && (c) < 97))
+
+void /* PRIVATE */
+png_check_chunk_name(png_structp png_ptr, png_bytep chunk_name)
+{
+   png_debug(1, "in png_check_chunk_name\n");
+   if (isnonalpha(chunk_name[0]) || isnonalpha(chunk_name[1]) ||
+       isnonalpha(chunk_name[2]) || isnonalpha(chunk_name[3]))
+   {
+      png_chunk_error(png_ptr, "invalid chunk type");
+   }
+}
+
+/* Combines the row recently read in with the existing pixels in the
+   row.  This routine takes care of alpha and transparency if requested.
+   This routine also handles the two methods of progressive display
+   of interlaced images, depending on the mask value.
+   The mask value describes which pixels are to be combined with
+   the row.  The pattern always repeats every 8 pixels, so just 8
+   bits are needed.  A one indicates the pixel is to be combined,
+   a zero indicates the pixel is to be skipped.  This is in addition
+   to any alpha or transparency value associated with the pixel.  If
+   you want all pixels to be combined, pass 0xff (255) in mask.  */
+
+/* Optimized C version of utilities to read a PNG file
+ *
+ * Based on code contributed by Nirav Chhatrapati, Intel Corp., 1998.
+ * Interface to libpng contributed by Gilles Vollant, 1999.
+ * GNU C port by Greg Roelofs, 1999-2001.
+ *
+ */
+
+#if defined(PNG_OPTIMIZED_CODE_SUPPORTED)
+#if !defined(PNG_HAVE_MMX_COMBINE_ROW)
+
+/*===========================================================================*/
+/*                                                                           */
+/*                       P N G _ C O M B I N E _ R O W                       */
+/*                                                                           */
+/*===========================================================================*/
+
+
+#define BPP2  2
+#define BPP3  3 /* bytes per pixel (a.k.a. pixel_bytes) */
+#define BPP4  4
+#define BPP6  6 /* (defined only to help avoid cut-and-paste errors) */
+#define BPP8  8
+
+/* Combines the row recently read in with the previous row.
+   This routine takes care of alpha and transparency if requested.
+   This routine also handles the two methods of progressive display
+   of interlaced images, depending on the mask value.
+   The mask value describes which pixels are to be combined with
+   the row.  The pattern always repeats every 8 pixels, so just 8
+   bits are needed.  A one indicates the pixel is to be combined; a
+   zero indicates the pixel is to be skipped.  This is in addition
+   to any alpha or transparency value associated with the pixel.
+   If you want all pixels to be combined, pass 0xff (255) in mask. */
+
+/* Use this routine for the x86 platform - it uses a faster MMX routine
+   if the machine supports MMX. */
+
+void /* PRIVATE */
+png_combine_row(png_structp png_ptr, png_bytep row, int mask)
+{
+
+#if defined(PNG_USE_LOCAL_ARRAYS)
+static PNG_CONST int FARDATA png_pass_start[7] = {0, 4, 0, 2, 0, 1, 0};
+static PNG_CONST int FARDATA png_pass_inc[7]   = {8, 8, 4, 4, 2, 2, 1};
+static PNG_CONST int FARDATA png_pass_width[7] = {8, 4, 4, 2, 2, 1, 1};
+#endif
+
+   png_debug(1, "in png_combine_row (pngrutil.c OPTIMIZED)\n");
+
+   if (mask == 0xff)
+   {
+      png_debug(2,"mask == 0xff:  doing single png_memcpy()\n");
+      png_memcpy(row, png_ptr->row_buf + 1,
+       (png_size_t)PNG_ROWBYTES(png_ptr->row_info.pixel_depth,png_ptr->width));
+   }
+   else   /* (png_combine_row() is never called with mask == 0) */
+   {
+      switch (png_ptr->row_info.pixel_depth)
+      {
+         /* most common case:  combining 24-bit RGB */
+         case 24:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = BPP3 * png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = BPP3 * png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = BPP3 * png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7; /* reduce to mult of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = BPP3 * len;   /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff*BPP3;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+            } /* end of else (_mmx_supported) */
+
+            break;
+         }       /* end 24 bpp */
+
+         case 32:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = BPP4 * png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = BPP4 * png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = BPP4 * png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7; /* reduce to mult of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = BPP4 * len;   /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff*BPP4;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+            }
+
+            break;
+         }       /* end 32 bpp */
+
+         case 8:        /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7; /* reduce to mult of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = len;  /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff /* *BPP1 */ ;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+            }
+
+            break;
+         }       /* end 8 bpp */
+
+         case 1:        /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int s_inc, s_start, s_end;
+            int m;
+            int shift;
+            png_uint_32 i;
+
+            sp = png_ptr->row_buf + 1;
+            dp = row;
+            m = 0x80;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+                s_start = 0;
+                s_end = 7;
+                s_inc = 1;
+            }
+            else
+#endif
+            {
+                s_start = 7;
+                s_end = 0;
+                s_inc = -1;
+            }
+
+            shift = s_start;
+
+            for (i = 0; i < png_ptr->width; i++)
+            {
+               if (m & mask)
+               {
+                  int value;
+
+                  value = (*sp >> shift) & 0x1;
+                  *dp &= (png_byte)((0x7f7f >> (7 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }       /* end 1 bpp */
+
+         case 2:        /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int s_start, s_end, s_inc;
+            int m;
+            int shift;
+            png_uint_32 i;
+            int value;
+
+            sp = png_ptr->row_buf + 1;
+            dp = row;
+            m = 0x80;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+               s_start = 0;
+               s_end = 6;
+               s_inc = 2;
+            }
+            else
+#endif
+            {
+               s_start = 6;
+               s_end = 0;
+               s_inc = -2;
+            }
+
+            shift = s_start;
+
+            for (i = 0; i < png_ptr->width; i++)
+            {
+               if (m & mask)
+               {
+                  value = (*sp >> shift) & 0x3;
+                  *dp &= (png_byte)((0x3f3f >> (6 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }       /* end 2 bpp */
+
+         case 4:        /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int s_start, s_end, s_inc;
+            int m;
+            int shift;
+            png_uint_32 i;
+            int value;
+
+            sp = png_ptr->row_buf + 1;
+            dp = row;
+            m = 0x80;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+               s_start = 0;
+               s_end = 4;
+               s_inc = 4;
+            }
+            else
+#endif
+            {
+               s_start = 4;
+               s_end = 0;
+               s_inc = -4;
+            }
+            shift = s_start;
+
+            for (i = 0; i < png_ptr->width; i++)
+            {
+               if (m & mask)
+               {
+                  value = (*sp >> shift) & 0xf;
+                  *dp &= (png_byte)((0xf0f >> (4 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }       /* end 4 bpp */
+
+         case 16:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = BPP2 * png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = BPP2 * png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = BPP2 * png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7;  /* reduce to mult of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = BPP2 * len;   /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff*BPP2;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+            } /* end of else (_mmx_supported) */
+
+            break;
+         }       /* end 16 bpp */
+
+
+
+         case 48:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+            {
+               register png_uint_32 i;
+               png_uint_32 initial_val = BPP6 * png_pass_start[png_ptr->pass];
+                 /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+               register int stride = BPP6 * png_pass_inc[png_ptr->pass];
+                 /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+               register int rep_bytes = BPP6 * png_pass_width[png_ptr->pass];
+                 /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+               png_uint_32 len = png_ptr->width &~7; /* reduce to mult of 8 */
+               int diff = (int) (png_ptr->width & 7); /* amount lost */
+               register png_uint_32 final_val = BPP6 * len;   /* GRR bugfix */
+
+               srcptr = png_ptr->row_buf + 1 + initial_val;
+               dstptr = row + initial_val;
+
+               for (i = initial_val; i < final_val; i += stride)
+               {
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+               if (diff)  /* number of leftover pixels:  3 for pngtest */
+               {
+                  final_val += diff*BPP6;
+                  for (; i < final_val; i += stride)
+                  {
+                     if (rep_bytes > (int)(final_val-i))
+                        rep_bytes = (int)(final_val-i);
+                     png_memcpy(dstptr, srcptr, rep_bytes);
+                     srcptr += stride;
+                     dstptr += stride;
+                  }
+               }
+            }
+            break;
+         }       /* end 48 bpp */
+
+         case 64:       /* png_ptr->row_info.pixel_depth */
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+            register png_uint_32 i;
+            png_uint_32 initial_val = BPP8 * png_pass_start[png_ptr->pass];
+              /* png.c:  png_pass_start[] = {0, 4, 0, 2, 0, 1, 0}; */
+            register int stride = BPP8 * png_pass_inc[png_ptr->pass];
+              /* png.c:  png_pass_inc[] = {8, 8, 4, 4, 2, 2, 1}; */
+            register int rep_bytes = BPP8 * png_pass_width[png_ptr->pass];
+              /* png.c:  png_pass_width[] = {8, 4, 4, 2, 2, 1, 1}; */
+            png_uint_32 len = png_ptr->width &~7;  /* reduce to mult of 8 */
+            int diff = (int) (png_ptr->width & 7); /* amount lost */
+            register png_uint_32 final_val = BPP8 * len;   /* GRR bugfix */
+
+            srcptr = png_ptr->row_buf + 1 + initial_val;
+            dstptr = row + initial_val;
+
+            for (i = initial_val; i < final_val; i += stride)
+            {
+               png_memcpy(dstptr, srcptr, rep_bytes);
+               srcptr += stride;
+               dstptr += stride;
+            }
+            if (diff)  /* number of leftover pixels:  3 for pngtest */
+            {
+               final_val += diff*BPP8;
+               for (; i < final_val; i += stride)
+               {
+                  if (rep_bytes > (int)(final_val-i))
+                     rep_bytes = (int)(final_val-i);
+                  png_memcpy(dstptr, srcptr, rep_bytes);
+                  srcptr += stride;
+                  dstptr += stride;
+               }
+            }
+
+            break;
+         }       /* end 64 bpp */
+
+         default: /* png_ptr->row_info.pixel_depth != 1,2,4,8,16,24,32,48,64 */
+         {
+            /* this should never happen */
+            png_warning(png_ptr, "Invalid row_info.pixel_depth in pngrutil");
+            break;
+         }
+      } /* end switch (png_ptr->row_info.pixel_depth) */
+
+   } /* end if (non-trivial mask) */
+
+} /* end png_combine_row() */
+#endif /* PNG_HAVE_MMX_COMBINE_ROW */
+
+
+
+/*===========================================================================*/
+/*                                                                           */
+/*                 P N G _ D O _ R E A D _ I N T E R L A C E                 */
+/*                                                                           */
+/*===========================================================================*/
+
+#if defined(PNG_READ_INTERLACING_SUPPORTED)
+#if !defined(PNG_HAVE_MMX_READ_INTERLACE)
+
+/* png_do_read_interlace() is called after any 16-bit to 8-bit conversion
+ * has taken place.  [GRR: what other steps come before and/or after?]
+ */
+
+void /* PRIVATE */
+png_do_read_interlace(png_structp png_ptr)
+{
+#if defined(PNG_USE_LOCAL_ARRAYS)
+static PNG_CONST int FARDATA png_pass_inc[7]   = {8, 8, 4, 4, 2, 2, 1};
+#endif
+   png_row_infop row_info = &(png_ptr->row_info);
+   png_bytep row = png_ptr->row_buf + 1;
+   int pass = png_ptr->pass;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+   png_uint_32 transformations = png_ptr->transformations;
+#endif
+   png_debug(1,"in png_do_read_interlace (pngrutil.c OPTIMIZED)\n");
+
+   if (row != NULL && row_info != NULL)
+   {
+      png_uint_32 final_width;
+
+      final_width = row_info->width * png_pass_inc[pass];
+
+      switch (row_info->pixel_depth)
+      {
+         case 1:
+         {
+            png_bytep sp, dp;
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_byte v;
+            png_uint_32 i;
+            int j;
+
+            sp = row + (png_size_t)((row_info->width - 1) >> 3);
+            dp = row + (png_size_t)((final_width - 1) >> 3);
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (int)((row_info->width + 7) & 7);
+               dshift = (int)((final_width + 7) & 7);
+               s_start = 7;
+               s_end = 0;
+               s_inc = -1;
+            }
+            else
+#endif
+            {
+               sshift = 7 - (int)((row_info->width + 7) & 7);
+               dshift = 7 - (int)((final_width + 7) & 7);
+               s_start = 0;
+               s_end = 7;
+               s_inc = 1;
+            }
+
+            for (i = row_info->width; i; i--)
+            {
+               v = (png_byte)((*sp >> sshift) & 0x1);
+               for (j = 0; j < png_pass_inc[pass]; j++)
+               {
+                  *dp &= (png_byte)((0x7f7f >> (7 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+
+         case 2:
+         {
+            png_bytep sp, dp;
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_uint_32 i;
+
+            sp = row + (png_size_t)((row_info->width - 1) >> 2);
+            dp = row + (png_size_t)((final_width - 1) >> 2);
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (png_size_t)(((row_info->width + 3) & 3) << 1);
+               dshift = (png_size_t)(((final_width + 3) & 3) << 1);
+               s_start = 6;
+               s_end = 0;
+               s_inc = -2;
+            }
+            else
+#endif
+            {
+               sshift = (png_size_t)((3 - ((row_info->width + 3) & 3)) << 1);
+               dshift = (png_size_t)((3 - ((final_width + 3) & 3)) << 1);
+               s_start = 0;
+               s_end = 6;
+               s_inc = 2;
+            }
+
+            for (i = row_info->width; i; i--)
+            {
+               png_byte v;
+               int j;
+
+               v = (png_byte)((*sp >> sshift) & 0x3);
+               for (j = 0; j < png_pass_inc[pass]; j++)
+               {
+                  *dp &= (png_byte)((0x3f3f >> (6 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+
+         case 4:
+         {
+            png_bytep sp, dp;
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_uint_32 i;
+
+            sp = row + (png_size_t)((row_info->width - 1) >> 1);
+            dp = row + (png_size_t)((final_width - 1) >> 1);
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (png_size_t)(((row_info->width + 1) & 1) << 2);
+               dshift = (png_size_t)(((final_width + 1) & 1) << 2);
+               s_start = 4;
+               s_end = 0;
+               s_inc = -4;
+            }
+            else
+#endif
+            {
+               sshift = (png_size_t)((1 - ((row_info->width + 1) & 1)) << 2);
+               dshift = (png_size_t)((1 - ((final_width + 1) & 1)) << 2);
+               s_start = 0;
+               s_end = 4;
+               s_inc = 4;
+            }
+
+            for (i = row_info->width; i; i--)
+            {
+               png_byte v;
+               int j;
+
+               v = (png_byte)((*sp >> sshift) & 0xf);
+               for (j = 0; j < png_pass_inc[pass]; j++)
+               {
+                  *dp &= (png_byte)((0xf0f >> (4 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+
+       /*====================================================================*/
+
+         default: /* 8-bit or larger (this is where the routine is modified) */
+         {
+            png_bytep sptr, dp;
+            png_uint_32 i;
+            png_size_t pixel_bytes;
+            int width = (int)row_info->width;
+
+            pixel_bytes = (row_info->pixel_depth >> 3);
+
+            /* point sptr at the last pixel in the pre-expanded row: */
+            sptr = row + (width - 1) * pixel_bytes;
+
+            /* point dp at the last pixel position in the expanded row: */
+            dp = row + (final_width - 1) * pixel_bytes;
+
+            /* MMX not supported:  use modified C code - takes advantage
+             *   of inlining of png_memcpy for a constant */
+            /* GRR 19991007:  does it?  or should pixel_bytes in each
+             *   block be replaced with immediate value (e.g., 1)? */
+            /* GRR 19991017:  replaced with constants in each case */
+            {
+               if (pixel_bytes == 1)
+               {
+                  for (i = width; i; i--)
+                  {
+                     int j;
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        *dp-- = *sptr;
+                     }
+                     --sptr;
+                  }
+               }
+               else if (pixel_bytes == 3)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, 3);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, 3);
+                        dp -= 3;
+                     }
+                     sptr -= 3;
+                  }
+               }
+               else if (pixel_bytes == 2)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, 2);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, 2);
+                        dp -= 2;
+                     }
+                     sptr -= 2;
+                  }
+               }
+               else if (pixel_bytes == 4)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, 4);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+#if defined(PNG_DEBUG) && defined(PNG_1_0_X)
+                        if (dp < row || dp+3 > row+png_ptr->row_buf_size)
+                        {
+                           printf("dp out of bounds: row=%d, dp=%d, rp=%d\n",
+                             row, dp, row+png_ptr->row_buf_size);
+                           printf("row_buf=%d\n",png_ptr->row_buf_size);
+                        }
+#endif
+                        png_memcpy(dp, v, 4);
+                        dp -= 4;
+                     }
+                     sptr -= 4;
+                  }
+               }
+               else if (pixel_bytes == 6)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, 6);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, 6);
+                        dp -= 6;
+                     }
+                     sptr -= 6;
+                  }
+               }
+               else if (pixel_bytes == 8)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, 8);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, 8);
+                        dp -= 8;
+                     }
+                     sptr -= 8;
+                  }
+               }
+               else     /* GRR:  should never be reached */
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, pixel_bytes);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, pixel_bytes);
+                        dp -= pixel_bytes;
+                     }
+                     sptr -= pixel_bytes;
+                  }
+               }
+
+            }
+            break;
+         }
+      } /* end switch (row_info->pixel_depth) */
+
+      row_info->width = final_width;
+
+      row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,final_width);
+   }
+
+} /* end png_do_read_interlace() */
+
+#endif /* PNG_HAVE_MMX_READ_INTERLACE */
+#endif /* PNG_READ_INTERLACING_SUPPORTED */
+
+
+
+#if !defined(PNG_HAVE_MMX_READ_FILTER_ROW)
+/*===========================================================================*/
+/*                                                                           */
+/*                   P N G _ R E A D _ F I L T E R _ R O W                   */
+/*                                                                           */
+/*===========================================================================*/
+
+
+/* Optimized png_read_filter_row routines */
+
+void /* PRIVATE */
+png_read_filter_row(png_structp png_ptr, png_row_infop row_info, png_bytep
+   row, png_bytep prev_row, int filter)
+{
+#if defined(PNG_DEBUG)
+   char filnm[10];
+#endif
+
+
+#if defined(PNG_DEBUG)
+   png_debug(1, "in png_read_filter_row (pngrutil.c OPTIMIZED)\n");
+   switch (filter)
+   {
+      case 0:
+         png_snprintf(filnm, 10, "none");
+         break;
+
+      case 1:
+         png_snprintf(filnm, 10, "sub-%s",
+             "x86");
+         break;
+
+      case 2:
+         png_snprintf(filnm, 10, "up-%s",
+             "x86");
+         break;
+
+      case 3:
+         png_snprintf(filnm, 10, "avg-%s",
+             "x86");
+         break;
+
+      case 4:
+         png_snprintf(filnm, 10, "Paeth-%s",
+             "x86");
+         break;
+
+      default:
+         png_snprintf(filnm, 10, "unknown");
+         break;
+   }
+   png_debug2(0, "row_number=%5ld, %10s, ", png_ptr->row_number, filnm);
+   png_debug1(0, "row=0x%08lx, ", (unsigned long)row);
+   png_debug2(0, "pixdepth=%2d, bytes=%d, ", (int)row_info->pixel_depth,
+      (int)((row_info->pixel_depth + 7) >> 3));
+   png_debug1(0,"rowbytes=%8ld\n", row_info->rowbytes);
+#endif /* PNG_DEBUG */
+
+   switch (filter)
+   {
+      case PNG_FILTER_VALUE_NONE:
+         break;
+
+      case PNG_FILTER_VALUE_SUB:
+         {
+            png_uint_32 i;
+            png_uint_32 istop = row_info->rowbytes;
+            png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+            png_bytep rp = row + bpp;
+            png_bytep lp = row;
+
+            for (i = bpp; i < istop; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) + (int)(*lp++)) & 0xff);
+               rp++;
+            }
+         }
+         break;
+
+      case PNG_FILTER_VALUE_UP:
+         {
+            png_uint_32 i;
+            png_uint_32 istop = row_info->rowbytes;
+            png_bytep rp = row;
+            png_bytep pp = prev_row;
+
+            for (i = 0; i < istop; ++i)
+            {
+               *rp = (png_byte)(((int)(*rp) + (int)(*pp++)) & 0xff);
+               rp++;
+            }
+         }
+         break;
+
+      case PNG_FILTER_VALUE_AVG:
+         {
+            png_uint_32 i;
+            png_bytep rp = row;
+            png_bytep pp = prev_row;
+            png_bytep lp = row;
+            png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+            png_uint_32 istop = row_info->rowbytes - bpp;
+
+            for (i = 0; i < bpp; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) +
+                  ((int)(*pp++) >> 1)) & 0xff);
+               rp++;
+            }
+
+            for (i = 0; i < istop; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) +
+                  ((int)(*pp++ + *lp++) >> 1)) & 0xff);
+               rp++;
+            }
+         }
+         break;
+
+      case PNG_FILTER_VALUE_PAETH:
+         {
+            png_uint_32 i;
+            png_bytep rp = row;
+            png_bytep pp = prev_row;
+            png_bytep lp = row;
+            png_bytep cp = prev_row;
+            png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+            png_uint_32 istop = row_info->rowbytes - bpp;
+
+            for (i = 0; i < bpp; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) + (int)(*pp++)) & 0xff);
+               rp++;
+            }
+
+            for (i = 0; i < istop; i++)   /* use leftover rp,pp */
+            {
+               int a, b, c, pa, pb, pc, p;
+
+               a = *lp++;
+               b = *pp++;
+               c = *cp++;
+
+               p = b - c;
+               pc = a - c;
+
+#if defined(PNG_USE_ABS)
+               pa = abs(p);
+               pb = abs(pc);
+               pc = abs(p + pc);
+#else
+               pa = p < 0 ? -p : p;
+               pb = pc < 0 ? -pc : pc;
+               pc = (p + pc) < 0 ? -(p + pc) : p + pc;
+#endif
+
+               /*
+                  if (pa <= pb && pa <= pc)
+                     p = a;
+                  else if (pb <= pc)
+                     p = b;
+                  else
+                     p = c;
+                */
+
+               p = (pa <= pb && pa <= pc) ? a : (pb <= pc) ? b : c;
+
+               *rp = (png_byte)(((int)(*rp) + p) & 0xff);
+               rp++;
+            }
+         }
+         break;
+
+      default:
+         png_warning(png_ptr, "Ignoring bad row-filter type");
+         *row=0;
+         break;
+   }
+}
+
+#endif /* PNG_HAVE_MMX_READ_FILTER_ROW */
+#endif /* PNG_OPTIMIZED_CODE_SUPPORTED */
+
+#if !defined(PNG_USE_PNGGCCRD) && !defined(PNG_USE_PNGVCRD)
+#if !defined(PNG_OPTIMIZED_CODE_SUPPORTED)
+/* Use the unoptimized original C code.  This might be removed from a future
+ * version of libpng if testing proves it to be worthless. */
+void /* PRIVATE */
+png_combine_row(png_structp png_ptr, png_bytep row, int mask)
+{
+   png_debug(1,"in png_combine_row NOT OPTIMIZED\n");
+   if (mask == 0xff)
+   {
+      png_memcpy(row, png_ptr->row_buf + 1,
+         PNG_ROWBYTES(png_ptr->row_info.pixel_depth, png_ptr->width));
+   }
+   else
+   {
+      switch (png_ptr->row_info.pixel_depth)
+      {
+         case 1:
+         {
+            png_bytep sp = png_ptr->row_buf + 1;
+            png_bytep dp = row;
+            int s_inc, s_start, s_end;
+            int m = 0x80;
+            int shift;
+            png_uint_32 i;
+            png_uint_32 row_width = png_ptr->width;
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+                s_start = 0;
+                s_end = 7;
+                s_inc = 1;
+            }
+            else
+#endif
+            {
+                s_start = 7;
+                s_end = 0;
+                s_inc = -1;
+            }
+
+            shift = s_start;
+
+            for (i = 0; i < row_width; i++)
+            {
+               if (m & mask)
+               {
+                  int value;
+
+                  value = (*sp >> shift) & 0x01;
+                  *dp &= (png_byte)((0x7f7f >> (7 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }
+         case 2:
+         {
+            png_bytep sp = png_ptr->row_buf + 1;
+            png_bytep dp = row;
+            int s_start, s_end, s_inc;
+            int m = 0x80;
+            int shift;
+            png_uint_32 i;
+            png_uint_32 row_width = png_ptr->width;
+            int value;
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+               s_start = 0;
+               s_end = 6;
+               s_inc = 2;
+            }
+            else
+#endif
+            {
+               s_start = 6;
+               s_end = 0;
+               s_inc = -2;
+            }
+
+            shift = s_start;
+
+            for (i = 0; i < row_width; i++)
+            {
+               if (m & mask)
+               {
+                  value = (*sp >> shift) & 0x03;
+                  *dp &= (png_byte)((0x3f3f >> (6 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }
+         case 4:
+         {
+            png_bytep sp = png_ptr->row_buf + 1;
+            png_bytep dp = row;
+            int s_start, s_end, s_inc;
+            int m = 0x80;
+            int shift;
+            png_uint_32 i;
+            png_uint_32 row_width = png_ptr->width;
+            int value;
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+               s_start = 0;
+               s_end = 4;
+               s_inc = 4;
+            }
+            else
+#endif
+            {
+               s_start = 4;
+               s_end = 0;
+               s_inc = -4;
+            }
+            shift = s_start;
+
+            for (i = 0; i < row_width; i++)
+            {
+               if (m & mask)
+               {
+                  value = (*sp >> shift) & 0xf;
+                  *dp &= (png_byte)((0xf0f >> (4 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }
+         default:
+         {
+            png_bytep sp = png_ptr->row_buf + 1;
+            png_bytep dp = row;
+            png_size_t pixel_bytes = (png_ptr->row_info.pixel_depth >> 3);
+            png_uint_32 i;
+            png_uint_32 row_width = png_ptr->width;
+            png_byte m = 0x80;
+
+
+            for (i = 0; i < row_width; i++)
+            {
+               if (m & mask)
+               {
+                  png_memcpy(dp, sp, pixel_bytes);
+               }
+
+               sp += pixel_bytes;
+               dp += pixel_bytes;
+
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }
+      }
+   }
+}
+
+#ifdef PNG_READ_INTERLACING_SUPPORTED
+/* OLD pre-1.0.9 interface:
+void png_do_read_interlace(png_row_infop row_info, png_bytep row, int pass,
+   png_uint_32 transformations)
+ */
+void /* PRIVATE */
+png_do_read_interlace(png_structp png_ptr)
+{
+   png_row_infop row_info = &(png_ptr->row_info);
+   png_bytep row = png_ptr->row_buf + 1;
+   int pass = png_ptr->pass;
+   png_uint_32 transformations = png_ptr->transformations;
+#ifdef PNG_USE_LOCAL_ARRAYS
+   /* arrays to facilitate easy interlacing - use pass (0 - 6) as index */
+   /* offset to next interlace block */
+   PNG_CONST int png_pass_inc[7] = {8, 8, 4, 4, 2, 2, 1};
+#endif
+
+   png_debug(1,"in png_do_read_interlace (pngrutil.c NOT OPTIMIZED)\n");
+   if (row != NULL && row_info != NULL)
+   {
+      png_uint_32 final_width;
+
+      final_width = row_info->width * png_pass_inc[pass];
+
+      switch (row_info->pixel_depth)
+      {
+         case 1:
+         {
+            png_bytep sp = row + (png_size_t)((row_info->width - 1) >> 3);
+            png_bytep dp = row + (png_size_t)((final_width - 1) >> 3);
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            int jstop = png_pass_inc[pass];
+            png_byte v;
+            png_uint_32 i;
+            int j;
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+                sshift = (int)((row_info->width + 7) & 0x07);
+                dshift = (int)((final_width + 7) & 0x07);
+                s_start = 7;
+                s_end = 0;
+                s_inc = -1;
+            }
+            else
+#endif
+            {
+                sshift = 7 - (int)((row_info->width + 7) & 0x07);
+                dshift = 7 - (int)((final_width + 7) & 0x07);
+                s_start = 0;
+                s_end = 7;
+                s_inc = 1;
+            }
+
+            for (i = 0; i < row_info->width; i++)
+            {
+               v = (png_byte)((*sp >> sshift) & 0x01);
+               for (j = 0; j < jstop; j++)
+               {
+                  *dp &= (png_byte)((0x7f7f >> (7 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+         case 2:
+         {
+            png_bytep sp = row + (png_uint_32)((row_info->width - 1) >> 2);
+            png_bytep dp = row + (png_uint_32)((final_width - 1) >> 2);
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            int jstop = png_pass_inc[pass];
+            png_uint_32 i;
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (int)(((row_info->width + 3) & 0x03) << 1);
+               dshift = (int)(((final_width + 3) & 0x03) << 1);
+               s_start = 6;
+               s_end = 0;
+               s_inc = -2;
+            }
+            else
+#endif
+            {
+               sshift = (int)((3 - ((row_info->width + 3) & 0x03)) << 1);
+               dshift = (int)((3 - ((final_width + 3) & 0x03)) << 1);
+               s_start = 0;
+               s_end = 6;
+               s_inc = 2;
+            }
+
+            for (i = 0; i < row_info->width; i++)
+            {
+               png_byte v;
+               int j;
+
+               v = (png_byte)((*sp >> sshift) & 0x03);
+               for (j = 0; j < jstop; j++)
+               {
+                  *dp &= (png_byte)((0x3f3f >> (6 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+         case 4:
+         {
+            png_bytep sp = row + (png_size_t)((row_info->width - 1) >> 1);
+            png_bytep dp = row + (png_size_t)((final_width - 1) >> 1);
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_uint_32 i;
+            int jstop = png_pass_inc[pass];
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (int)(((row_info->width + 1) & 0x01) << 2);
+               dshift = (int)(((final_width + 1) & 0x01) << 2);
+               s_start = 4;
+               s_end = 0;
+               s_inc = -4;
+            }
+            else
+#endif
+            {
+               sshift = (int)((1 - ((row_info->width + 1) & 0x01)) << 2);
+               dshift = (int)((1 - ((final_width + 1) & 0x01)) << 2);
+               s_start = 0;
+               s_end = 4;
+               s_inc = 4;
+            }
+
+            for (i = 0; i < row_info->width; i++)
+            {
+               png_byte v = (png_byte)((*sp >> sshift) & 0xf);
+               int j;
+
+               for (j = 0; j < jstop; j++)
+               {
+                  *dp &= (png_byte)((0xf0f >> (4 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+         default:
+         {
+            png_size_t pixel_bytes = (row_info->pixel_depth >> 3);
+            png_bytep sp = row + (png_size_t)(row_info->width - 1) * pixel_bytes;
+            png_bytep dp = row + (png_size_t)(final_width - 1) * pixel_bytes;
+
+            int jstop = png_pass_inc[pass];
+            png_uint_32 i;
+
+            for (i = 0; i < row_info->width; i++)
+            {
+               png_byte v[8];
+               int j;
+
+               png_memcpy(v, sp, pixel_bytes);
+               for (j = 0; j < jstop; j++)
+               {
+                  png_memcpy(dp, v, pixel_bytes);
+                  dp -= pixel_bytes;
+               }
+               sp -= pixel_bytes;
+            }
+            break;
+         }
+      }
+      row_info->width = final_width;
+      row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,final_width);
+   }
+#if !defined(PNG_READ_PACKSWAP_SUPPORTED)
+   transformations = transformations; /* silence compiler warning */
+#endif
+}
+#endif /* PNG_READ_INTERLACING_SUPPORTED */
+
+void /* PRIVATE */
+png_read_filter_row(png_structp png_ptr, png_row_infop row_info, png_bytep row,
+   png_bytep prev_row, int filter)
+{
+   png_debug(1, "in png_read_filter_row (NOT OPTIMIZED)\n");
+   png_debug2(2,"row = %lu, filter = %d\n", png_ptr->row_number, filter);
+   switch (filter)
+   {
+      case PNG_FILTER_VALUE_NONE:
+         break;
+      case PNG_FILTER_VALUE_SUB:
+      {
+         png_uint_32 i;
+         png_uint_32 istop = row_info->rowbytes;
+         png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+         png_bytep rp = row + bpp;
+         png_bytep lp = row;
+
+         for (i = bpp; i < istop; i++)
+         {
+            *rp = (png_byte)(((int)(*rp) + (int)(*lp++)) & 0xff);
+            rp++;
+         }
+         break;
+      }
+      case PNG_FILTER_VALUE_UP:
+      {
+         png_uint_32 i;
+         png_uint_32 istop = row_info->rowbytes;
+         png_bytep rp = row;
+         png_bytep pp = prev_row;
+
+         for (i = 0; i < istop; i++)
+         {
+            *rp = (png_byte)(((int)(*rp) + (int)(*pp++)) & 0xff);
+            rp++;
+         }
+         break;
+      }
+      case PNG_FILTER_VALUE_AVG:
+      {
+         png_uint_32 i;
+         png_bytep rp = row;
+         png_bytep pp = prev_row;
+         png_bytep lp = row;
+         png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+         png_uint_32 istop = row_info->rowbytes - bpp;
+
+         for (i = 0; i < bpp; i++)
+         {
+            *rp = (png_byte)(((int)(*rp) +
+               ((int)(*pp++) / 2 )) & 0xff);
+            rp++;
+         }
+
+         for (i = 0; i < istop; i++)
+         {
+            *rp = (png_byte)(((int)(*rp) +
+               (int)(*pp++ + *lp++) / 2 ) & 0xff);
+            rp++;
+         }
+         break;
+      }
+      case PNG_FILTER_VALUE_PAETH:
+      {
+         png_uint_32 i;
+         png_bytep rp = row;
+         png_bytep pp = prev_row;
+         png_bytep lp = row;
+         png_bytep cp = prev_row;
+         png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+         png_uint_32 istop=row_info->rowbytes - bpp;
+
+         for (i = 0; i < bpp; i++)
+         {
+            *rp = (png_byte)(((int)(*rp) + (int)(*pp++)) & 0xff);
+            rp++;
+         }
+
+         for (i = 0; i < istop; i++)   /* use leftover rp,pp */
+         {
+            int a, b, c, pa, pb, pc, p;
+
+            a = *lp++;
+            b = *pp++;
+            c = *cp++;
+
+            p = b - c;
+            pc = a - c;
+
+#ifdef PNG_USE_ABS
+            pa = abs(p);
+            pb = abs(pc);
+            pc = abs(p + pc);
+#else
+            pa = p < 0 ? -p : p;
+            pb = pc < 0 ? -pc : pc;
+            pc = (p + pc) < 0 ? -(p + pc) : p + pc;
+#endif
+
+            /*
+               if (pa <= pb && pa <= pc)
+                  p = a;
+               else if (pb <= pc)
+                  p = b;
+               else
+                  p = c;
+             */
+
+            p = (pa <= pb && pa <=pc) ? a : (pb <= pc) ? b : c;
+
+            *rp = (png_byte)(((int)(*rp) + p) & 0xff);
+            rp++;
+         }
+         break;
+      }
+      default:
+         png_warning(png_ptr, "Ignoring bad adaptive filter type");
+         *row=0;
+         break;
+   }
+}
+#endif /* !PNG_OPTIMIZED_CODE_SUPPORTED */
+#endif /* !PNG_USE_PNGGCCRD && !PNG_USE_PNGVCRD */
+
+void /* PRIVATE */
+png_read_finish_row(png_structp png_ptr)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   /* arrays to facilitate easy interlacing - use pass (0 - 6) as index */
+
+   /* start of interlace block */
+   PNG_CONST int png_pass_start[7] = {0, 4, 0, 2, 0, 1, 0};
+
+   /* offset to next interlace block */
+   PNG_CONST int png_pass_inc[7] = {8, 8, 4, 4, 2, 2, 1};
+
+   /* start of interlace block in the y direction */
+   PNG_CONST int png_pass_ystart[7] = {0, 0, 4, 0, 2, 0, 1};
+
+   /* offset to next interlace block in the y direction */
+   PNG_CONST int png_pass_yinc[7] = {8, 8, 8, 4, 4, 2, 2};
+#endif
+
+   png_debug(1, "in png_read_finish_row\n");
+   png_ptr->row_number++;
+   if (png_ptr->row_number < png_ptr->num_rows)
+      return;
+
+   if (png_ptr->interlaced)
+   {
+      png_ptr->row_number = 0;
+      png_memset_check(png_ptr, png_ptr->prev_row, 0,
+         png_ptr->rowbytes + 1);
+      do
+      {
+         png_ptr->pass++;
+         if (png_ptr->pass >= 7)
+            break;
+         png_ptr->iwidth = (png_ptr->width +
+            png_pass_inc[png_ptr->pass] - 1 -
+            png_pass_start[png_ptr->pass]) /
+            png_pass_inc[png_ptr->pass];
+
+         png_ptr->irowbytes = PNG_ROWBYTES(png_ptr->pixel_depth,
+            png_ptr->iwidth) + 1;
+
+         if (!(png_ptr->transformations & PNG_INTERLACE))
+         {
+            png_ptr->num_rows = (png_ptr->height +
+               png_pass_yinc[png_ptr->pass] - 1 -
+               png_pass_ystart[png_ptr->pass]) /
+               png_pass_yinc[png_ptr->pass];
+            if (!(png_ptr->num_rows))
+               continue;
+         }
+         else  /* if (png_ptr->transformations & PNG_INTERLACE) */
+            break;
+      } while (png_ptr->iwidth == 0);
+
+      if (png_ptr->pass < 7)
+         return;
+   }
+
+   if (!(png_ptr->flags & PNG_FLAG_ZLIB_FINISHED))
+   {
+#ifdef PNG_USE_LOCAL_ARRAYS
+      PNG_CONST PNG_IDAT;
+#endif
+      char extra;
+      int ret;
+
+      png_ptr->zstream.next_out = (Byte *)&extra;
+      png_ptr->zstream.avail_out = (uInt)1;
+      for(;;)
+      {
+         if (!(png_ptr->zstream.avail_in))
+         {
+            while (!png_ptr->idat_size)
+            {
+               png_byte chunk_length[4];
+
+               png_crc_finish(png_ptr, 0);
+
+               png_read_data(png_ptr, chunk_length, 4);
+               png_ptr->idat_size = png_get_uint_31(png_ptr, chunk_length);
+               png_reset_crc(png_ptr);
+               png_crc_read(png_ptr, png_ptr->chunk_name, 4);
+               if (png_memcmp(png_ptr->chunk_name, (png_bytep)png_IDAT, 4))
+                  png_error(png_ptr, "Not enough image data");
+
+            }
+            png_ptr->zstream.avail_in = (uInt)png_ptr->zbuf_size;
+            png_ptr->zstream.next_in = png_ptr->zbuf;
+            if (png_ptr->zbuf_size > png_ptr->idat_size)
+               png_ptr->zstream.avail_in = (uInt)png_ptr->idat_size;
+            png_crc_read(png_ptr, png_ptr->zbuf, png_ptr->zstream.avail_in);
+            png_ptr->idat_size -= png_ptr->zstream.avail_in;
+         }
+         ret = inflate(&png_ptr->zstream, Z_PARTIAL_FLUSH);
+         if (ret == Z_STREAM_END)
+         {
+            if (!(png_ptr->zstream.avail_out) || png_ptr->zstream.avail_in ||
+               png_ptr->idat_size)
+               png_warning(png_ptr, "Extra compressed data");
+            png_ptr->mode |= PNG_AFTER_IDAT;
+            png_ptr->flags |= PNG_FLAG_ZLIB_FINISHED;
+            break;
+         }
+         if (ret != Z_OK)
+            png_error(png_ptr, png_ptr->zstream.msg ? png_ptr->zstream.msg :
+                      "Decompression Error");
+
+         if (!(png_ptr->zstream.avail_out))
+         {
+            png_warning(png_ptr, "Extra compressed data.");
+            png_ptr->mode |= PNG_AFTER_IDAT;
+            png_ptr->flags |= PNG_FLAG_ZLIB_FINISHED;
+            break;
+         }
+
+      }
+      png_ptr->zstream.avail_out = 0;
+   }
+
+   if (png_ptr->idat_size || png_ptr->zstream.avail_in)
+      png_warning(png_ptr, "Extra compression data");
+
+   inflateReset(&png_ptr->zstream);
+
+   png_ptr->mode |= PNG_AFTER_IDAT;
+}
+
+void /* PRIVATE */
+png_read_start_row(png_structp png_ptr)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   /* arrays to facilitate easy interlacing - use pass (0 - 6) as index */
+
+   /* start of interlace block */
+   PNG_CONST int png_pass_start[7] = {0, 4, 0, 2, 0, 1, 0};
+
+   /* offset to next interlace block */
+   PNG_CONST int png_pass_inc[7] = {8, 8, 4, 4, 2, 2, 1};
+
+   /* start of interlace block in the y direction */
+   PNG_CONST int png_pass_ystart[7] = {0, 0, 4, 0, 2, 0, 1};
+
+   /* offset to next interlace block in the y direction */
+   PNG_CONST int png_pass_yinc[7] = {8, 8, 8, 4, 4, 2, 2};
+#endif
+
+   int max_pixel_depth;
+   png_uint_32 row_bytes;
+
+   png_debug(1, "in png_read_start_row\n");
+   png_ptr->zstream.avail_in = 0;
+   png_init_read_transformations(png_ptr);
+   if (png_ptr->interlaced)
+   {
+      if (!(png_ptr->transformations & PNG_INTERLACE))
+         png_ptr->num_rows = (png_ptr->height + png_pass_yinc[0] - 1 -
+            png_pass_ystart[0]) / png_pass_yinc[0];
+      else
+         png_ptr->num_rows = png_ptr->height;
+
+      png_ptr->iwidth = (png_ptr->width +
+         png_pass_inc[png_ptr->pass] - 1 -
+         png_pass_start[png_ptr->pass]) /
+         png_pass_inc[png_ptr->pass];
+
+         row_bytes = PNG_ROWBYTES(png_ptr->pixel_depth,png_ptr->iwidth) + 1;
+
+         png_ptr->irowbytes = (png_size_t)row_bytes;
+         if((png_uint_32)png_ptr->irowbytes != row_bytes)
+            png_error(png_ptr, "Rowbytes overflow in png_read_start_row");
+   }
+   else
+   {
+      png_ptr->num_rows = png_ptr->height;
+      png_ptr->iwidth = png_ptr->width;
+      png_ptr->irowbytes = png_ptr->rowbytes + 1;
+   }
+   max_pixel_depth = png_ptr->pixel_depth;
+
+#if defined(PNG_READ_PACK_SUPPORTED)
+   if ((png_ptr->transformations & PNG_PACK) && png_ptr->bit_depth < 8)
+      max_pixel_depth = 8;
+#endif
+
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+   if (png_ptr->transformations & PNG_EXPAND)
+   {
+      if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      {
+         if (png_ptr->num_trans)
+            max_pixel_depth = 32;
+         else
+            max_pixel_depth = 24;
+      }
+      else if (png_ptr->color_type == PNG_COLOR_TYPE_GRAY)
+      {
+         if (max_pixel_depth < 8)
+            max_pixel_depth = 8;
+         if (png_ptr->num_trans)
+            max_pixel_depth *= 2;
+      }
+      else if (png_ptr->color_type == PNG_COLOR_TYPE_RGB)
+      {
+         if (png_ptr->num_trans)
+         {
+            max_pixel_depth *= 4;
+            max_pixel_depth /= 3;
+         }
+      }
+   }
+#endif
+
+#if defined(PNG_READ_FILLER_SUPPORTED)
+   if (png_ptr->transformations & (PNG_FILLER))
+   {
+      if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+         max_pixel_depth = 32;
+      else if (png_ptr->color_type == PNG_COLOR_TYPE_GRAY)
+      {
+         if (max_pixel_depth <= 8)
+            max_pixel_depth = 16;
+         else
+            max_pixel_depth = 32;
+      }
+      else if (png_ptr->color_type == PNG_COLOR_TYPE_RGB)
+      {
+         if (max_pixel_depth <= 32)
+            max_pixel_depth = 32;
+         else
+            max_pixel_depth = 64;
+      }
+   }
+#endif
+
+#if defined(PNG_READ_GRAY_TO_RGB_SUPPORTED)
+   if (png_ptr->transformations & PNG_GRAY_TO_RGB)
+   {
+      if (
+#if defined(PNG_READ_EXPAND_SUPPORTED)
+        (png_ptr->num_trans && (png_ptr->transformations & PNG_EXPAND)) ||
+#endif
+#if defined(PNG_READ_FILLER_SUPPORTED)
+        (png_ptr->transformations & (PNG_FILLER)) ||
+#endif
+        png_ptr->color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
+      {
+         if (max_pixel_depth <= 16)
+            max_pixel_depth = 32;
+         else
+            max_pixel_depth = 64;
+      }
+      else
+      {
+         if (max_pixel_depth <= 8)
+           {
+             if (png_ptr->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+               max_pixel_depth = 32;
+             else
+               max_pixel_depth = 24;
+           }
+         else if (png_ptr->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+            max_pixel_depth = 64;
+         else
+            max_pixel_depth = 48;
+      }
+   }
+#endif
+
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED) && \
+defined(PNG_USER_TRANSFORM_PTR_SUPPORTED)
+   if(png_ptr->transformations & PNG_USER_TRANSFORM)
+     {
+       int user_pixel_depth=png_ptr->user_transform_depth*
+         png_ptr->user_transform_channels;
+       if(user_pixel_depth > max_pixel_depth)
+         max_pixel_depth=user_pixel_depth;
+     }
+#endif
+
+   /* align the width on the next larger 8 pixels.  Mainly used
+      for interlacing */
+   row_bytes = ((png_ptr->width + 7) & ~((png_uint_32)7));
+   /* calculate the maximum bytes needed, adding a byte and a pixel
+      for safety's sake */
+   row_bytes = PNG_ROWBYTES(max_pixel_depth,row_bytes) +
+      1 + ((max_pixel_depth + 7) >> 3);
+#ifdef PNG_MAX_MALLOC_64K
+   if (row_bytes > (png_uint_32)65536L)
+      png_error(png_ptr, "This image requires a row greater than 64KB");
+#endif
+   png_ptr->big_row_buf = (png_bytep)png_malloc(png_ptr, row_bytes+64);
+   png_ptr->row_buf = png_ptr->big_row_buf+32;
+#if defined(PNG_DEBUG) && defined(PNG_USE_PNGGCCRD) && defined(PNG_1_0_X)
+   png_ptr->row_buf_size = row_bytes;
+#endif
+
+#ifdef PNG_MAX_MALLOC_64K
+   if ((png_uint_32)png_ptr->rowbytes + 1 > (png_uint_32)65536L)
+      png_error(png_ptr, "This image requires a row greater than 64KB");
+#endif
+   if ((png_uint_32)png_ptr->rowbytes > (png_uint_32)(PNG_SIZE_MAX - 1))
+      png_error(png_ptr, "Row has too many bytes to allocate in memory.");
+   png_ptr->prev_row = (png_bytep)png_malloc(png_ptr, (png_uint_32)(
+      png_ptr->rowbytes + 1));
+
+   png_memset_check(png_ptr, png_ptr->prev_row, 0, png_ptr->rowbytes + 1);
+
+   png_debug1(3, "width = %lu,\n", png_ptr->width);
+   png_debug1(3, "height = %lu,\n", png_ptr->height);
+   png_debug1(3, "iwidth = %lu,\n", png_ptr->iwidth);
+   png_debug1(3, "num_rows = %lu\n", png_ptr->num_rows);
+   png_debug1(3, "rowbytes = %lu,\n", png_ptr->rowbytes);
+   png_debug1(3, "irowbytes = %lu,\n", png_ptr->irowbytes);
+
+   png_ptr->flags |= PNG_FLAG_ROW_INIT;
+}
+#endif /* PNG_READ_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngset.c b/distrib/libpng-1.2.19/pngset.c
new file mode 100644
index 0000000..8d1cbfc
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngset.c
@@ -0,0 +1,1284 @@
+
+/* pngset.c - storage of image information into info struct
+ *
+ * Last changed in libpng 1.2.17 May 15, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ *
+ * The functions here are used during reads to store data from the file
+ * into the info struct, and during writes to store application data
+ * into the info struct for writing into the file.  This abstracts the
+ * info struct and allows us to change the structure in the future.
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED)
+
+#if defined(PNG_bKGD_SUPPORTED)
+void PNGAPI
+png_set_bKGD(png_structp png_ptr, png_infop info_ptr, png_color_16p background)
+{
+   png_debug1(1, "in %s storage function\n", "bKGD");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   png_memcpy(&(info_ptr->background), background, png_sizeof(png_color_16));
+   info_ptr->valid |= PNG_INFO_bKGD;
+}
+#endif
+
+#if defined(PNG_cHRM_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+void PNGAPI
+png_set_cHRM(png_structp png_ptr, png_infop info_ptr,
+   double white_x, double white_y, double red_x, double red_y,
+   double green_x, double green_y, double blue_x, double blue_y)
+{
+   png_debug1(1, "in %s storage function\n", "cHRM");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   if (white_x < 0.0 || white_y < 0.0 ||
+         red_x < 0.0 ||   red_y < 0.0 ||
+       green_x < 0.0 || green_y < 0.0 ||
+        blue_x < 0.0 ||  blue_y < 0.0)
+   {
+      png_warning(png_ptr,
+        "Ignoring attempt to set negative chromaticity value");
+      return;
+   }
+   if (white_x > 21474.83 || white_y > 21474.83 ||
+         red_x > 21474.83 ||   red_y > 21474.83 ||
+       green_x > 21474.83 || green_y > 21474.83 ||
+        blue_x > 21474.83 ||  blue_y > 21474.83)
+   {
+      png_warning(png_ptr,
+        "Ignoring attempt to set chromaticity value exceeding 21474.83");
+      return;
+   }
+
+   info_ptr->x_white = (float)white_x;
+   info_ptr->y_white = (float)white_y;
+   info_ptr->x_red   = (float)red_x;
+   info_ptr->y_red   = (float)red_y;
+   info_ptr->x_green = (float)green_x;
+   info_ptr->y_green = (float)green_y;
+   info_ptr->x_blue  = (float)blue_x;
+   info_ptr->y_blue  = (float)blue_y;
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   info_ptr->int_x_white = (png_fixed_point)(white_x*100000.+0.5);
+   info_ptr->int_y_white = (png_fixed_point)(white_y*100000.+0.5);
+   info_ptr->int_x_red   = (png_fixed_point)(  red_x*100000.+0.5);
+   info_ptr->int_y_red   = (png_fixed_point)(  red_y*100000.+0.5);
+   info_ptr->int_x_green = (png_fixed_point)(green_x*100000.+0.5);
+   info_ptr->int_y_green = (png_fixed_point)(green_y*100000.+0.5);
+   info_ptr->int_x_blue  = (png_fixed_point)( blue_x*100000.+0.5);
+   info_ptr->int_y_blue  = (png_fixed_point)( blue_y*100000.+0.5);
+#endif
+   info_ptr->valid |= PNG_INFO_cHRM;
+}
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+void PNGAPI
+png_set_cHRM_fixed(png_structp png_ptr, png_infop info_ptr,
+   png_fixed_point white_x, png_fixed_point white_y, png_fixed_point red_x,
+   png_fixed_point red_y, png_fixed_point green_x, png_fixed_point green_y,
+   png_fixed_point blue_x, png_fixed_point blue_y)
+{
+   png_debug1(1, "in %s storage function\n", "cHRM");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   if (white_x < 0 || white_y < 0 ||
+         red_x < 0 ||   red_y < 0 ||
+       green_x < 0 || green_y < 0 ||
+        blue_x < 0 ||  blue_y < 0)
+   {
+      png_warning(png_ptr,
+        "Ignoring attempt to set negative chromaticity value");
+      return;
+   }
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   if (white_x > (double) PNG_UINT_31_MAX ||
+       white_y > (double) PNG_UINT_31_MAX ||
+         red_x > (double) PNG_UINT_31_MAX ||
+         red_y > (double) PNG_UINT_31_MAX ||
+       green_x > (double) PNG_UINT_31_MAX ||
+       green_y > (double) PNG_UINT_31_MAX ||
+        blue_x > (double) PNG_UINT_31_MAX ||
+        blue_y > (double) PNG_UINT_31_MAX)
+#else
+   if (white_x > (png_fixed_point) PNG_UINT_31_MAX/100000L ||
+       white_y > (png_fixed_point) PNG_UINT_31_MAX/100000L ||
+         red_x > (png_fixed_point) PNG_UINT_31_MAX/100000L ||
+         red_y > (png_fixed_point) PNG_UINT_31_MAX/100000L ||
+       green_x > (png_fixed_point) PNG_UINT_31_MAX/100000L ||
+       green_y > (png_fixed_point) PNG_UINT_31_MAX/100000L ||
+        blue_x > (png_fixed_point) PNG_UINT_31_MAX/100000L ||
+        blue_y > (png_fixed_point) PNG_UINT_31_MAX/100000L)
+#endif
+   {
+      png_warning(png_ptr,
+        "Ignoring attempt to set chromaticity value exceeding 21474.83");
+      return;
+   }
+   info_ptr->int_x_white = white_x;
+   info_ptr->int_y_white = white_y;
+   info_ptr->int_x_red   = red_x;
+   info_ptr->int_y_red   = red_y;
+   info_ptr->int_x_green = green_x;
+   info_ptr->int_y_green = green_y;
+   info_ptr->int_x_blue  = blue_x;
+   info_ptr->int_y_blue  = blue_y;
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   info_ptr->x_white = (float)(white_x/100000.);
+   info_ptr->y_white = (float)(white_y/100000.);
+   info_ptr->x_red   = (float)(  red_x/100000.);
+   info_ptr->y_red   = (float)(  red_y/100000.);
+   info_ptr->x_green = (float)(green_x/100000.);
+   info_ptr->y_green = (float)(green_y/100000.);
+   info_ptr->x_blue  = (float)( blue_x/100000.);
+   info_ptr->y_blue  = (float)( blue_y/100000.);
+#endif
+   info_ptr->valid |= PNG_INFO_cHRM;
+}
+#endif
+#endif
+
+#if defined(PNG_gAMA_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+void PNGAPI
+png_set_gAMA(png_structp png_ptr, png_infop info_ptr, double file_gamma)
+{
+   double gamma;
+   png_debug1(1, "in %s storage function\n", "gAMA");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   /* Check for overflow */
+   if (file_gamma > 21474.83)
+   {
+      png_warning(png_ptr, "Limiting gamma to 21474.83");
+      gamma=21474.83;
+   }
+   else
+      gamma=file_gamma;
+   info_ptr->gamma = (float)gamma;
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   info_ptr->int_gamma = (int)(gamma*100000.+.5);
+#endif
+   info_ptr->valid |= PNG_INFO_gAMA;
+   if(gamma == 0.0)
+      png_warning(png_ptr, "Setting gamma=0");
+}
+#endif
+void PNGAPI
+png_set_gAMA_fixed(png_structp png_ptr, png_infop info_ptr, png_fixed_point
+   int_gamma)
+{
+   png_fixed_point gamma;
+
+   png_debug1(1, "in %s storage function\n", "gAMA");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   if (int_gamma > (png_fixed_point) PNG_UINT_31_MAX)
+   {
+     png_warning(png_ptr, "Limiting gamma to 21474.83");
+     gamma=PNG_UINT_31_MAX;
+   }
+   else
+   {
+     if (int_gamma < 0)
+     {
+       png_warning(png_ptr, "Setting negative gamma to zero");
+       gamma=0;
+     }
+     else
+       gamma=int_gamma;
+   }
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   info_ptr->gamma = (float)(gamma/100000.);
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   info_ptr->int_gamma = gamma;
+#endif
+   info_ptr->valid |= PNG_INFO_gAMA;
+   if(gamma == 0)
+      png_warning(png_ptr, "Setting gamma=0");
+}
+#endif
+
+#if defined(PNG_hIST_SUPPORTED)
+void PNGAPI
+png_set_hIST(png_structp png_ptr, png_infop info_ptr, png_uint_16p hist)
+{
+   int i;
+
+   png_debug1(1, "in %s storage function\n", "hIST");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+   if (info_ptr->num_palette <= 0 || info_ptr->num_palette
+       > PNG_MAX_PALETTE_LENGTH)
+   {
+       png_warning(png_ptr,
+          "Invalid palette size, hIST allocation skipped.");
+       return;
+   }
+
+#ifdef PNG_FREE_ME_SUPPORTED
+   png_free_data(png_ptr, info_ptr, PNG_FREE_HIST, 0);
+#endif
+   /* Changed from info->num_palette to PNG_MAX_PALETTE_LENGTH in version
+      1.2.1 */
+   png_ptr->hist = (png_uint_16p)png_malloc_warn(png_ptr,
+      (png_uint_32)(PNG_MAX_PALETTE_LENGTH * png_sizeof (png_uint_16)));
+   if (png_ptr->hist == NULL)
+     {
+       png_warning(png_ptr, "Insufficient memory for hIST chunk data.");
+       return;
+     }
+
+   for (i = 0; i < info_ptr->num_palette; i++)
+       png_ptr->hist[i] = hist[i];
+   info_ptr->hist = png_ptr->hist;
+   info_ptr->valid |= PNG_INFO_hIST;
+
+#ifdef PNG_FREE_ME_SUPPORTED
+   info_ptr->free_me |= PNG_FREE_HIST;
+#else
+   png_ptr->flags |= PNG_FLAG_FREE_HIST;
+#endif
+}
+#endif
+
+void PNGAPI
+png_set_IHDR(png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 width, png_uint_32 height, int bit_depth,
+   int color_type, int interlace_type, int compression_type,
+   int filter_type)
+{
+   png_debug1(1, "in %s storage function\n", "IHDR");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   /* check for width and height valid values */
+   if (width == 0 || height == 0)
+      png_error(png_ptr, "Image width or height is zero in IHDR");
+#ifdef PNG_SET_USER_LIMITS_SUPPORTED
+   if (width > png_ptr->user_width_max || height > png_ptr->user_height_max)
+      png_error(png_ptr, "image size exceeds user limits in IHDR");
+#else
+   if (width > PNG_USER_WIDTH_MAX || height > PNG_USER_HEIGHT_MAX)
+      png_error(png_ptr, "image size exceeds user limits in IHDR");
+#endif
+   if (width > PNG_UINT_31_MAX || height > PNG_UINT_31_MAX)
+      png_error(png_ptr, "Invalid image size in IHDR");
+   if ( width > (PNG_UINT_32_MAX
+                 >> 3)      /* 8-byte RGBA pixels */
+                 - 64       /* bigrowbuf hack */
+                 - 1        /* filter byte */
+                 - 7*8      /* rounding of width to multiple of 8 pixels */
+                 - 8)       /* extra max_pixel_depth pad */
+      png_warning(png_ptr, "Width is too large for libpng to process pixels");
+
+   /* check other values */
+   if (bit_depth != 1 && bit_depth != 2 && bit_depth != 4 &&
+      bit_depth != 8 && bit_depth != 16)
+      png_error(png_ptr, "Invalid bit depth in IHDR");
+
+   if (color_type < 0 || color_type == 1 ||
+      color_type == 5 || color_type > 6)
+      png_error(png_ptr, "Invalid color type in IHDR");
+
+   if (((color_type == PNG_COLOR_TYPE_PALETTE) && bit_depth > 8) ||
+       ((color_type == PNG_COLOR_TYPE_RGB ||
+         color_type == PNG_COLOR_TYPE_GRAY_ALPHA ||
+         color_type == PNG_COLOR_TYPE_RGB_ALPHA) && bit_depth < 8))
+      png_error(png_ptr, "Invalid color type/bit depth combination in IHDR");
+
+   if (interlace_type >= PNG_INTERLACE_LAST)
+      png_error(png_ptr, "Unknown interlace method in IHDR");
+
+   if (compression_type != PNG_COMPRESSION_TYPE_BASE)
+      png_error(png_ptr, "Unknown compression method in IHDR");
+
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+   /* Accept filter_method 64 (intrapixel differencing) only if
+    * 1. Libpng was compiled with PNG_MNG_FEATURES_SUPPORTED and
+    * 2. Libpng did not read a PNG signature (this filter_method is only
+    *    used in PNG datastreams that are embedded in MNG datastreams) and
+    * 3. The application called png_permit_mng_features with a mask that
+    *    included PNG_FLAG_MNG_FILTER_64 and
+    * 4. The filter_method is 64 and
+    * 5. The color_type is RGB or RGBA
+    */
+   if((png_ptr->mode&PNG_HAVE_PNG_SIGNATURE)&&png_ptr->mng_features_permitted)
+      png_warning(png_ptr,"MNG features are not allowed in a PNG datastream");
+   if(filter_type != PNG_FILTER_TYPE_BASE)
+   {
+     if(!((png_ptr->mng_features_permitted & PNG_FLAG_MNG_FILTER_64) &&
+        (filter_type == PNG_INTRAPIXEL_DIFFERENCING) &&
+        ((png_ptr->mode&PNG_HAVE_PNG_SIGNATURE) == 0) &&
+        (color_type == PNG_COLOR_TYPE_RGB ||
+         color_type == PNG_COLOR_TYPE_RGB_ALPHA)))
+        png_error(png_ptr, "Unknown filter method in IHDR");
+     if(png_ptr->mode&PNG_HAVE_PNG_SIGNATURE)
+        png_warning(png_ptr, "Invalid filter method in IHDR");
+   }
+#else
+   if(filter_type != PNG_FILTER_TYPE_BASE)
+      png_error(png_ptr, "Unknown filter method in IHDR");
+#endif
+
+   info_ptr->width = width;
+   info_ptr->height = height;
+   info_ptr->bit_depth = (png_byte)bit_depth;
+   info_ptr->color_type =(png_byte) color_type;
+   info_ptr->compression_type = (png_byte)compression_type;
+   info_ptr->filter_type = (png_byte)filter_type;
+   info_ptr->interlace_type = (png_byte)interlace_type;
+   if (info_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      info_ptr->channels = 1;
+   else if (info_ptr->color_type & PNG_COLOR_MASK_COLOR)
+      info_ptr->channels = 3;
+   else
+      info_ptr->channels = 1;
+   if (info_ptr->color_type & PNG_COLOR_MASK_ALPHA)
+      info_ptr->channels++;
+   info_ptr->pixel_depth = (png_byte)(info_ptr->channels * info_ptr->bit_depth);
+
+   /* check for potential overflow */
+   if (width > (PNG_UINT_32_MAX
+                 >> 3)      /* 8-byte RGBA pixels */
+                 - 64       /* bigrowbuf hack */
+                 - 1        /* filter byte */
+                 - 7*8      /* rounding of width to multiple of 8 pixels */
+                 - 8)       /* extra max_pixel_depth pad */
+      info_ptr->rowbytes = (png_size_t)0;
+   else
+      info_ptr->rowbytes = PNG_ROWBYTES(info_ptr->pixel_depth,width);
+}
+
+#if defined(PNG_oFFs_SUPPORTED)
+void PNGAPI
+png_set_oFFs(png_structp png_ptr, png_infop info_ptr,
+   png_int_32 offset_x, png_int_32 offset_y, int unit_type)
+{
+   png_debug1(1, "in %s storage function\n", "oFFs");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   info_ptr->x_offset = offset_x;
+   info_ptr->y_offset = offset_y;
+   info_ptr->offset_unit_type = (png_byte)unit_type;
+   info_ptr->valid |= PNG_INFO_oFFs;
+}
+#endif
+
+#if defined(PNG_pCAL_SUPPORTED)
+void PNGAPI
+png_set_pCAL(png_structp png_ptr, png_infop info_ptr,
+   png_charp purpose, png_int_32 X0, png_int_32 X1, int type, int nparams,
+   png_charp units, png_charpp params)
+{
+   png_uint_32 length;
+   int i;
+
+   png_debug1(1, "in %s storage function\n", "pCAL");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   length = png_strlen(purpose) + 1;
+   png_debug1(3, "allocating purpose for info (%lu bytes)\n", length);
+   info_ptr->pcal_purpose = (png_charp)png_malloc_warn(png_ptr, length);
+   if (info_ptr->pcal_purpose == NULL)
+     {
+       png_warning(png_ptr, "Insufficient memory for pCAL purpose.");
+       return;
+     }
+   png_memcpy(info_ptr->pcal_purpose, purpose, (png_size_t)length);
+
+   png_debug(3, "storing X0, X1, type, and nparams in info\n");
+   info_ptr->pcal_X0 = X0;
+   info_ptr->pcal_X1 = X1;
+   info_ptr->pcal_type = (png_byte)type;
+   info_ptr->pcal_nparams = (png_byte)nparams;
+
+   length = png_strlen(units) + 1;
+   png_debug1(3, "allocating units for info (%lu bytes)\n", length);
+   info_ptr->pcal_units = (png_charp)png_malloc_warn(png_ptr, length);
+   if (info_ptr->pcal_units == NULL)
+     {
+       png_warning(png_ptr, "Insufficient memory for pCAL units.");
+       return;
+     }
+   png_memcpy(info_ptr->pcal_units, units, (png_size_t)length);
+
+   info_ptr->pcal_params = (png_charpp)png_malloc_warn(png_ptr,
+      (png_uint_32)((nparams + 1) * png_sizeof(png_charp)));
+   if (info_ptr->pcal_params == NULL)
+     {
+       png_warning(png_ptr, "Insufficient memory for pCAL params.");
+       return;
+     }
+
+   info_ptr->pcal_params[nparams] = NULL;
+
+   for (i = 0; i < nparams; i++)
+   {
+      length = png_strlen(params[i]) + 1;
+      png_debug2(3, "allocating parameter %d for info (%lu bytes)\n", i, length);
+      info_ptr->pcal_params[i] = (png_charp)png_malloc_warn(png_ptr, length);
+      if (info_ptr->pcal_params[i] == NULL)
+        {
+          png_warning(png_ptr, "Insufficient memory for pCAL parameter.");
+          return;
+        }
+      png_memcpy(info_ptr->pcal_params[i], params[i], (png_size_t)length);
+   }
+
+   info_ptr->valid |= PNG_INFO_pCAL;
+#ifdef PNG_FREE_ME_SUPPORTED
+   info_ptr->free_me |= PNG_FREE_PCAL;
+#endif
+}
+#endif
+
+#if defined(PNG_READ_sCAL_SUPPORTED) || defined(PNG_WRITE_sCAL_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+void PNGAPI
+png_set_sCAL(png_structp png_ptr, png_infop info_ptr,
+             int unit, double width, double height)
+{
+   png_debug1(1, "in %s storage function\n", "sCAL");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   info_ptr->scal_unit = (png_byte)unit;
+   info_ptr->scal_pixel_width = width;
+   info_ptr->scal_pixel_height = height;
+
+   info_ptr->valid |= PNG_INFO_sCAL;
+}
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+void PNGAPI
+png_set_sCAL_s(png_structp png_ptr, png_infop info_ptr,
+             int unit, png_charp swidth, png_charp sheight)
+{
+   png_uint_32 length;
+
+   png_debug1(1, "in %s storage function\n", "sCAL");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   info_ptr->scal_unit = (png_byte)unit;
+
+   length = png_strlen(swidth) + 1;
+   png_debug1(3, "allocating unit for info (%d bytes)\n", length);
+   info_ptr->scal_s_width = (png_charp)png_malloc_warn(png_ptr, length);
+   if (info_ptr->scal_s_width == NULL)
+   {
+      png_warning(png_ptr,
+       "Memory allocation failed while processing sCAL.");
+   }
+   png_memcpy(info_ptr->scal_s_width, swidth, (png_size_t)length);
+
+   length = png_strlen(sheight) + 1;
+   png_debug1(3, "allocating unit for info (%d bytes)\n", length);
+   info_ptr->scal_s_height = (png_charp)png_malloc_warn(png_ptr, length);
+   if (info_ptr->scal_s_height == NULL)
+   {
+      png_free (png_ptr, info_ptr->scal_s_width);
+      png_warning(png_ptr,
+       "Memory allocation failed while processing sCAL.");
+   }
+   png_memcpy(info_ptr->scal_s_height, sheight, (png_size_t)length);
+
+   info_ptr->valid |= PNG_INFO_sCAL;
+#ifdef PNG_FREE_ME_SUPPORTED
+   info_ptr->free_me |= PNG_FREE_SCAL;
+#endif
+}
+#endif
+#endif
+#endif
+
+#if defined(PNG_pHYs_SUPPORTED)
+void PNGAPI
+png_set_pHYs(png_structp png_ptr, png_infop info_ptr,
+   png_uint_32 res_x, png_uint_32 res_y, int unit_type)
+{
+   png_debug1(1, "in %s storage function\n", "pHYs");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   info_ptr->x_pixels_per_unit = res_x;
+   info_ptr->y_pixels_per_unit = res_y;
+   info_ptr->phys_unit_type = (png_byte)unit_type;
+   info_ptr->valid |= PNG_INFO_pHYs;
+}
+#endif
+
+void PNGAPI
+png_set_PLTE(png_structp png_ptr, png_infop info_ptr,
+   png_colorp palette, int num_palette)
+{
+
+   png_debug1(1, "in %s storage function\n", "PLTE");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   if (num_palette < 0 || num_palette > PNG_MAX_PALETTE_LENGTH)
+     {
+       if (info_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+         png_error(png_ptr, "Invalid palette length");
+       else
+       {
+         png_warning(png_ptr, "Invalid palette length");
+         return;
+       }
+     }
+
+   /*
+    * It may not actually be necessary to set png_ptr->palette here;
+    * we do it for backward compatibility with the way the png_handle_tRNS
+    * function used to do the allocation.
+    */
+#ifdef PNG_FREE_ME_SUPPORTED
+   png_free_data(png_ptr, info_ptr, PNG_FREE_PLTE, 0);
+#endif
+
+   /* Changed in libpng-1.2.1 to allocate PNG_MAX_PALETTE_LENGTH instead
+      of num_palette entries,
+      in case of an invalid PNG file that has too-large sample values. */
+   png_ptr->palette = (png_colorp)png_malloc(png_ptr,
+      PNG_MAX_PALETTE_LENGTH * png_sizeof(png_color));
+   png_memset(png_ptr->palette, 0, PNG_MAX_PALETTE_LENGTH *
+      png_sizeof(png_color));
+   png_memcpy(png_ptr->palette, palette, num_palette * png_sizeof (png_color));
+   info_ptr->palette = png_ptr->palette;
+   info_ptr->num_palette = png_ptr->num_palette = (png_uint_16)num_palette;
+
+#ifdef PNG_FREE_ME_SUPPORTED
+   info_ptr->free_me |= PNG_FREE_PLTE;
+#else
+   png_ptr->flags |= PNG_FLAG_FREE_PLTE;
+#endif
+
+   info_ptr->valid |= PNG_INFO_PLTE;
+}
+
+#if defined(PNG_sBIT_SUPPORTED)
+void PNGAPI
+png_set_sBIT(png_structp png_ptr, png_infop info_ptr,
+   png_color_8p sig_bit)
+{
+   png_debug1(1, "in %s storage function\n", "sBIT");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   png_memcpy(&(info_ptr->sig_bit), sig_bit, png_sizeof (png_color_8));
+   info_ptr->valid |= PNG_INFO_sBIT;
+}
+#endif
+
+#if defined(PNG_sRGB_SUPPORTED)
+void PNGAPI
+png_set_sRGB(png_structp png_ptr, png_infop info_ptr, int intent)
+{
+   png_debug1(1, "in %s storage function\n", "sRGB");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   info_ptr->srgb_intent = (png_byte)intent;
+   info_ptr->valid |= PNG_INFO_sRGB;
+}
+
+void PNGAPI
+png_set_sRGB_gAMA_and_cHRM(png_structp png_ptr, png_infop info_ptr,
+   int intent)
+{
+#if defined(PNG_gAMA_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   float file_gamma;
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   png_fixed_point int_file_gamma;
+#endif
+#endif
+#if defined(PNG_cHRM_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   float white_x, white_y, red_x, red_y, green_x, green_y, blue_x, blue_y;
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   png_fixed_point int_white_x, int_white_y, int_red_x, int_red_y, int_green_x,
+      int_green_y, int_blue_x, int_blue_y;
+#endif
+#endif
+   png_debug1(1, "in %s storage function\n", "sRGB_gAMA_and_cHRM");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   png_set_sRGB(png_ptr, info_ptr, intent);
+
+#if defined(PNG_gAMA_SUPPORTED)
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   file_gamma = (float).45455;
+   png_set_gAMA(png_ptr, info_ptr, file_gamma);
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   int_file_gamma = 45455L;
+   png_set_gAMA_fixed(png_ptr, info_ptr, int_file_gamma);
+#endif
+#endif
+
+#if defined(PNG_cHRM_SUPPORTED)
+#ifdef PNG_FIXED_POINT_SUPPORTED
+   int_white_x = 31270L;
+   int_white_y = 32900L;
+   int_red_x   = 64000L;
+   int_red_y   = 33000L;
+   int_green_x = 30000L;
+   int_green_y = 60000L;
+   int_blue_x  = 15000L;
+   int_blue_y  =  6000L;
+
+   png_set_cHRM_fixed(png_ptr, info_ptr,
+      int_white_x, int_white_y, int_red_x, int_red_y, int_green_x, int_green_y,
+      int_blue_x, int_blue_y);
+#endif
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+   white_x = (float).3127;
+   white_y = (float).3290;
+   red_x   = (float).64;
+   red_y   = (float).33;
+   green_x = (float).30;
+   green_y = (float).60;
+   blue_x  = (float).15;
+   blue_y  = (float).06;
+
+   png_set_cHRM(png_ptr, info_ptr,
+      white_x, white_y, red_x, red_y, green_x, green_y, blue_x, blue_y);
+#endif
+#endif
+}
+#endif
+
+
+#if defined(PNG_iCCP_SUPPORTED)
+void PNGAPI
+png_set_iCCP(png_structp png_ptr, png_infop info_ptr,
+             png_charp name, int compression_type,
+             png_charp profile, png_uint_32 proflen)
+{
+   png_charp new_iccp_name;
+   png_charp new_iccp_profile;
+
+   png_debug1(1, "in %s storage function\n", "iCCP");
+   if (png_ptr == NULL || info_ptr == NULL || name == NULL || profile == NULL)
+      return;
+
+   new_iccp_name = (png_charp)png_malloc_warn(png_ptr, png_strlen(name)+1);
+   if (new_iccp_name == NULL)
+   {
+      png_warning(png_ptr, "Insufficient memory to process iCCP chunk.");
+      return;
+   }
+   png_strncpy(new_iccp_name, name, png_sizeof(new_iccp_name));
+   new_iccp_profile = (png_charp)png_malloc_warn(png_ptr, proflen);
+   if (new_iccp_profile == NULL)
+   {
+      png_free (png_ptr, new_iccp_name);
+      png_warning(png_ptr, "Insufficient memory to process iCCP profile.");
+      return;
+   }
+   png_memcpy(new_iccp_profile, profile, (png_size_t)proflen);
+
+   png_free_data(png_ptr, info_ptr, PNG_FREE_ICCP, 0);
+
+   info_ptr->iccp_proflen = proflen;
+   info_ptr->iccp_name = new_iccp_name;
+   info_ptr->iccp_profile = new_iccp_profile;
+   /* Compression is always zero but is here so the API and info structure
+    * does not have to change if we introduce multiple compression types */
+   info_ptr->iccp_compression = (png_byte)compression_type;
+#ifdef PNG_FREE_ME_SUPPORTED
+   info_ptr->free_me |= PNG_FREE_ICCP;
+#endif
+   info_ptr->valid |= PNG_INFO_iCCP;
+}
+#endif
+
+#if defined(PNG_TEXT_SUPPORTED)
+void PNGAPI
+png_set_text(png_structp png_ptr, png_infop info_ptr, png_textp text_ptr,
+   int num_text)
+{
+   int ret;
+   ret=png_set_text_2(png_ptr, info_ptr, text_ptr, num_text);
+   if (ret)
+     png_error(png_ptr, "Insufficient memory to store text");
+}
+
+int /* PRIVATE */
+png_set_text_2(png_structp png_ptr, png_infop info_ptr, png_textp text_ptr,
+   int num_text)
+{
+   int i;
+
+   png_debug1(1, "in %s storage function\n", (png_ptr->chunk_name[0] == '\0' ?
+      "text" : (png_const_charp)png_ptr->chunk_name));
+
+   if (png_ptr == NULL || info_ptr == NULL || num_text == 0)
+      return(0);
+
+   /* Make sure we have enough space in the "text" array in info_struct
+    * to hold all of the incoming text_ptr objects.
+    */
+   if (info_ptr->num_text + num_text > info_ptr->max_text)
+   {
+      if (info_ptr->text != NULL)
+      {
+         png_textp old_text;
+         int old_max;
+
+         old_max = info_ptr->max_text;
+         info_ptr->max_text = info_ptr->num_text + num_text + 8;
+         old_text = info_ptr->text;
+         info_ptr->text = (png_textp)png_malloc_warn(png_ptr,
+            (png_uint_32)(info_ptr->max_text * png_sizeof (png_text)));
+         if (info_ptr->text == NULL)
+           {
+             png_free(png_ptr, old_text);
+             return(1);
+           }
+         png_memcpy(info_ptr->text, old_text, (png_size_t)(old_max *
+            png_sizeof(png_text)));
+         png_free(png_ptr, old_text);
+      }
+      else
+      {
+         info_ptr->max_text = num_text + 8;
+         info_ptr->num_text = 0;
+         info_ptr->text = (png_textp)png_malloc_warn(png_ptr,
+            (png_uint_32)(info_ptr->max_text * png_sizeof (png_text)));
+         if (info_ptr->text == NULL)
+           return(1);
+#ifdef PNG_FREE_ME_SUPPORTED
+         info_ptr->free_me |= PNG_FREE_TEXT;
+#endif
+      }
+      png_debug1(3, "allocated %d entries for info_ptr->text\n",
+         info_ptr->max_text);
+   }
+   for (i = 0; i < num_text; i++)
+   {
+      png_size_t text_length,key_len;
+      png_size_t lang_len,lang_key_len;
+      png_textp textp = &(info_ptr->text[info_ptr->num_text]);
+
+      if (text_ptr[i].key == NULL)
+          continue;
+
+      key_len = png_strlen(text_ptr[i].key);
+
+      if(text_ptr[i].compression <= 0)
+      {
+        lang_len = 0;
+        lang_key_len = 0;
+      }
+      else
+#ifdef PNG_iTXt_SUPPORTED
+      {
+        /* set iTXt data */
+        if (text_ptr[i].lang != NULL)
+          lang_len = png_strlen(text_ptr[i].lang);
+        else
+          lang_len = 0;
+        if (text_ptr[i].lang_key != NULL)
+          lang_key_len = png_strlen(text_ptr[i].lang_key);
+        else
+          lang_key_len = 0;
+      }
+#else
+      {
+        png_warning(png_ptr, "iTXt chunk not supported.");
+        continue;
+      }
+#endif
+
+      if (text_ptr[i].text == NULL || text_ptr[i].text[0] == '\0')
+      {
+         text_length = 0;
+#ifdef PNG_iTXt_SUPPORTED
+         if(text_ptr[i].compression > 0)
+            textp->compression = PNG_ITXT_COMPRESSION_NONE;
+         else
+#endif
+            textp->compression = PNG_TEXT_COMPRESSION_NONE;
+      }
+      else
+      {
+         text_length = png_strlen(text_ptr[i].text);
+         textp->compression = text_ptr[i].compression;
+      }
+
+      textp->key = (png_charp)png_malloc_warn(png_ptr,
+         (png_uint_32)(key_len + text_length + lang_len + lang_key_len + 4));
+      if (textp->key == NULL)
+        return(1);
+      png_debug2(2, "Allocated %lu bytes at %x in png_set_text\n",
+         (png_uint_32)(key_len + lang_len + lang_key_len + text_length + 4),
+         (int)textp->key);
+
+      png_memcpy(textp->key, text_ptr[i].key,
+         (png_size_t)(key_len));
+      *(textp->key+key_len) = '\0';
+#ifdef PNG_iTXt_SUPPORTED
+      if (text_ptr[i].compression > 0)
+      {
+         textp->lang=textp->key + key_len + 1;
+         png_memcpy(textp->lang, text_ptr[i].lang, lang_len);
+         *(textp->lang+lang_len) = '\0';
+         textp->lang_key=textp->lang + lang_len + 1;
+         png_memcpy(textp->lang_key, text_ptr[i].lang_key, lang_key_len);
+         *(textp->lang_key+lang_key_len) = '\0';
+         textp->text=textp->lang_key + lang_key_len + 1;
+      }
+      else
+#endif
+      {
+#ifdef PNG_iTXt_SUPPORTED
+         textp->lang=NULL;
+         textp->lang_key=NULL;
+#endif
+         textp->text=textp->key + key_len + 1;
+      }
+      if(text_length)
+         png_memcpy(textp->text, text_ptr[i].text,
+            (png_size_t)(text_length));
+      *(textp->text+text_length) = '\0';
+
+#ifdef PNG_iTXt_SUPPORTED
+      if(textp->compression > 0)
+      {
+         textp->text_length = 0;
+         textp->itxt_length = text_length;
+      }
+      else
+#endif
+      {
+         textp->text_length = text_length;
+#ifdef PNG_iTXt_SUPPORTED
+         textp->itxt_length = 0;
+#endif
+      }
+      info_ptr->num_text++;
+      png_debug1(3, "transferred text chunk %d\n", info_ptr->num_text);
+   }
+   return(0);
+}
+#endif
+
+#if defined(PNG_tIME_SUPPORTED)
+void PNGAPI
+png_set_tIME(png_structp png_ptr, png_infop info_ptr, png_timep mod_time)
+{
+   png_debug1(1, "in %s storage function\n", "tIME");
+   if (png_ptr == NULL || info_ptr == NULL ||
+       (png_ptr->mode & PNG_WROTE_tIME))
+      return;
+
+   png_memcpy(&(info_ptr->mod_time), mod_time, png_sizeof (png_time));
+   info_ptr->valid |= PNG_INFO_tIME;
+}
+#endif
+
+#if defined(PNG_tRNS_SUPPORTED)
+void PNGAPI
+png_set_tRNS(png_structp png_ptr, png_infop info_ptr,
+   png_bytep trans, int num_trans, png_color_16p trans_values)
+{
+   png_debug1(1, "in %s storage function\n", "tRNS");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   if (trans != NULL)
+   {
+       /*
+        * It may not actually be necessary to set png_ptr->trans here;
+        * we do it for backward compatibility with the way the png_handle_tRNS
+        * function used to do the allocation.
+        */
+#ifdef PNG_FREE_ME_SUPPORTED
+       png_free_data(png_ptr, info_ptr, PNG_FREE_TRNS, 0);
+#endif
+       /* Changed from num_trans to PNG_MAX_PALETTE_LENGTH in version 1.2.1 */
+       png_ptr->trans = info_ptr->trans = (png_bytep)png_malloc(png_ptr,
+           (png_uint_32)PNG_MAX_PALETTE_LENGTH);
+       if (num_trans <= PNG_MAX_PALETTE_LENGTH)
+         png_memcpy(info_ptr->trans, trans, (png_size_t)num_trans);
+#ifdef PNG_FREE_ME_SUPPORTED
+       info_ptr->free_me |= PNG_FREE_TRNS;
+#else
+       png_ptr->flags |= PNG_FLAG_FREE_TRNS;
+#endif
+   }
+
+   if (trans_values != NULL)
+   {
+      png_memcpy(&(info_ptr->trans_values), trans_values,
+         png_sizeof(png_color_16));
+      if (num_trans == 0)
+        num_trans = 1;
+   }
+   info_ptr->num_trans = (png_uint_16)num_trans;
+   info_ptr->valid |= PNG_INFO_tRNS;
+}
+#endif
+
+#if defined(PNG_sPLT_SUPPORTED)
+void PNGAPI
+png_set_sPLT(png_structp png_ptr,
+             png_infop info_ptr, png_sPLT_tp entries, int nentries)
+{
+    png_sPLT_tp np;
+    int i;
+
+    if (png_ptr == NULL || info_ptr == NULL)
+       return;
+
+    np = (png_sPLT_tp)png_malloc_warn(png_ptr,
+        (info_ptr->splt_palettes_num + nentries) * png_sizeof(png_sPLT_t));
+    if (np == NULL)
+    {
+      png_warning(png_ptr, "No memory for sPLT palettes.");
+      return;
+    }
+
+    png_memcpy(np, info_ptr->splt_palettes,
+           info_ptr->splt_palettes_num * png_sizeof(png_sPLT_t));
+    png_free(png_ptr, info_ptr->splt_palettes);
+    info_ptr->splt_palettes=NULL;
+
+    for (i = 0; i < nentries; i++)
+    {
+        png_sPLT_tp to = np + info_ptr->splt_palettes_num + i;
+        png_sPLT_tp from = entries + i;
+
+        to->name = (png_charp)png_malloc_warn(png_ptr,
+          png_strlen(from->name) + 1);
+        if (to->name == NULL)
+        {
+           png_warning(png_ptr,
+             "Out of memory while processing sPLT chunk");
+        }
+        /* TODO: use png_malloc_warn */
+        png_strncpy(to->name, from->name, png_strlen(from->name));
+        to->entries = (png_sPLT_entryp)png_malloc_warn(png_ptr,
+            from->nentries * png_sizeof(png_sPLT_entry));
+        /* TODO: use png_malloc_warn */
+        png_memcpy(to->entries, from->entries,
+            from->nentries * png_sizeof(png_sPLT_entry));
+        if (to->entries == NULL)
+        {
+           png_warning(png_ptr,
+             "Out of memory while processing sPLT chunk");
+           png_free(png_ptr,to->name);
+           to->name = NULL;
+        }
+        to->nentries = from->nentries;
+        to->depth = from->depth;
+    }
+
+    info_ptr->splt_palettes = np;
+    info_ptr->splt_palettes_num += nentries;
+    info_ptr->valid |= PNG_INFO_sPLT;
+#ifdef PNG_FREE_ME_SUPPORTED
+    info_ptr->free_me |= PNG_FREE_SPLT;
+#endif
+}
+#endif /* PNG_sPLT_SUPPORTED */
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+void PNGAPI
+png_set_unknown_chunks(png_structp png_ptr,
+   png_infop info_ptr, png_unknown_chunkp unknowns, int num_unknowns)
+{
+    png_unknown_chunkp np;
+    int i;
+
+    if (png_ptr == NULL || info_ptr == NULL || num_unknowns == 0)
+        return;
+
+    np = (png_unknown_chunkp)png_malloc_warn(png_ptr,
+        (info_ptr->unknown_chunks_num + num_unknowns) *
+        png_sizeof(png_unknown_chunk));
+    if (np == NULL)
+    {
+       png_warning(png_ptr,
+          "Out of memory while processing unknown chunk.");
+       return;
+    }
+
+    png_memcpy(np, info_ptr->unknown_chunks,
+           info_ptr->unknown_chunks_num * png_sizeof(png_unknown_chunk));
+    png_free(png_ptr, info_ptr->unknown_chunks);
+    info_ptr->unknown_chunks=NULL;
+
+    for (i = 0; i < num_unknowns; i++)
+    {
+        png_unknown_chunkp to = np + info_ptr->unknown_chunks_num + i;
+        png_unknown_chunkp from = unknowns + i;
+
+        png_strncpy((png_charp)to->name, (png_charp)from->name, 5);
+        to->data = (png_bytep)png_malloc_warn(png_ptr, from->size);
+        if (to->data == NULL)
+        {
+           png_warning(png_ptr,
+              "Out of memory while processing unknown chunk.");
+        }
+        else
+        {
+           png_memcpy(to->data, from->data, from->size);
+           to->size = from->size;
+
+           /* note our location in the read or write sequence */
+           to->location = (png_byte)(png_ptr->mode & 0xff);
+        }
+    }
+
+    info_ptr->unknown_chunks = np;
+    info_ptr->unknown_chunks_num += num_unknowns;
+#ifdef PNG_FREE_ME_SUPPORTED
+    info_ptr->free_me |= PNG_FREE_UNKN;
+#endif
+}
+void PNGAPI
+png_set_unknown_chunk_location(png_structp png_ptr, png_infop info_ptr,
+   int chunk, int location)
+{
+   if(png_ptr != NULL && info_ptr != NULL && chunk >= 0 && chunk <
+         (int)info_ptr->unknown_chunks_num)
+      info_ptr->unknown_chunks[chunk].location = (png_byte)location;
+}
+#endif
+
+#if defined(PNG_1_0_X) || defined(PNG_1_2_X)
+#if defined(PNG_READ_EMPTY_PLTE_SUPPORTED) || \
+    defined(PNG_WRITE_EMPTY_PLTE_SUPPORTED)
+void PNGAPI
+png_permit_empty_plte (png_structp png_ptr, int empty_plte_permitted)
+{
+   /* This function is deprecated in favor of png_permit_mng_features()
+      and will be removed from libpng-1.3.0 */
+   png_debug(1, "in png_permit_empty_plte, DEPRECATED.\n");
+   if (png_ptr == NULL)
+      return;
+   png_ptr->mng_features_permitted = (png_byte)
+     ((png_ptr->mng_features_permitted & (~(PNG_FLAG_MNG_EMPTY_PLTE))) |
+     ((empty_plte_permitted & PNG_FLAG_MNG_EMPTY_PLTE)));
+}
+#endif
+#endif
+
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+png_uint_32 PNGAPI
+png_permit_mng_features (png_structp png_ptr, png_uint_32 mng_features)
+{
+   png_debug(1, "in png_permit_mng_features\n");
+   if (png_ptr == NULL)
+      return (png_uint_32)0;
+   png_ptr->mng_features_permitted =
+     (png_byte)(mng_features & PNG_ALL_MNG_FEATURES);
+   return (png_uint_32)png_ptr->mng_features_permitted;
+}
+#endif
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+void PNGAPI
+png_set_keep_unknown_chunks(png_structp png_ptr, int keep, png_bytep
+   chunk_list, int num_chunks)
+{
+    png_bytep new_list, p;
+    int i, old_num_chunks;
+    if (png_ptr == NULL)
+       return;
+    if (num_chunks == 0)
+    {
+      if(keep == PNG_HANDLE_CHUNK_ALWAYS || keep == PNG_HANDLE_CHUNK_IF_SAFE)
+        png_ptr->flags |= PNG_FLAG_KEEP_UNKNOWN_CHUNKS;
+      else
+        png_ptr->flags &= ~PNG_FLAG_KEEP_UNKNOWN_CHUNKS;
+
+      if(keep == PNG_HANDLE_CHUNK_ALWAYS)
+        png_ptr->flags |= PNG_FLAG_KEEP_UNSAFE_CHUNKS;
+      else
+        png_ptr->flags &= ~PNG_FLAG_KEEP_UNSAFE_CHUNKS;
+      return;
+    }
+    if (chunk_list == NULL)
+      return;
+    old_num_chunks=png_ptr->num_chunk_list;
+    new_list=(png_bytep)png_malloc(png_ptr,
+       (png_uint_32)(5*(num_chunks+old_num_chunks)));
+    if(png_ptr->chunk_list != NULL)
+    {
+       png_memcpy(new_list, png_ptr->chunk_list,
+          (png_size_t)(5*old_num_chunks));
+       png_free(png_ptr, png_ptr->chunk_list);
+       png_ptr->chunk_list=NULL;
+    }
+    png_memcpy(new_list+5*old_num_chunks, chunk_list,
+       (png_size_t)(5*num_chunks));
+    for (p=new_list+5*old_num_chunks+4, i=0; i<num_chunks; i++, p+=5)
+       *p=(png_byte)keep;
+    png_ptr->num_chunk_list=old_num_chunks+num_chunks;
+    png_ptr->chunk_list=new_list;
+#ifdef PNG_FREE_ME_SUPPORTED
+    png_ptr->free_me |= PNG_FREE_LIST;
+#endif
+}
+#endif
+
+#if defined(PNG_READ_USER_CHUNKS_SUPPORTED)
+void PNGAPI
+png_set_read_user_chunk_fn(png_structp png_ptr, png_voidp user_chunk_ptr,
+   png_user_chunk_ptr read_user_chunk_fn)
+{
+   png_debug(1, "in png_set_read_user_chunk_fn\n");
+   if (png_ptr == NULL)
+      return;
+   png_ptr->read_user_chunk_fn = read_user_chunk_fn;
+   png_ptr->user_chunk_ptr = user_chunk_ptr;
+}
+#endif
+
+#if defined(PNG_INFO_IMAGE_SUPPORTED)
+void PNGAPI
+png_set_rows(png_structp png_ptr, png_infop info_ptr, png_bytepp row_pointers)
+{
+   png_debug1(1, "in %s storage function\n", "rows");
+
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   if(info_ptr->row_pointers && (info_ptr->row_pointers != row_pointers))
+      png_free_data(png_ptr, info_ptr, PNG_FREE_ROWS, 0);
+   info_ptr->row_pointers = row_pointers;
+   if(row_pointers)
+      info_ptr->valid |= PNG_INFO_IDAT;
+}
+#endif
+
+#ifdef PNG_WRITE_SUPPORTED
+void PNGAPI
+png_set_compression_buffer_size(png_structp png_ptr, png_uint_32 size)
+{
+    if (png_ptr == NULL)
+       return;
+    if(png_ptr->zbuf)
+       png_free(png_ptr, png_ptr->zbuf);
+    png_ptr->zbuf_size = (png_size_t)size;
+    png_ptr->zbuf = (png_bytep)png_malloc(png_ptr, size);
+    png_ptr->zstream.next_out = png_ptr->zbuf;
+    png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+}
+#endif
+
+void PNGAPI
+png_set_invalid(png_structp png_ptr, png_infop info_ptr, int mask)
+{
+   if (png_ptr && info_ptr)
+      info_ptr->valid &= ~(mask);
+}
+
+
+#ifndef PNG_1_0_X
+#ifdef PNG_ASSEMBLER_CODE_SUPPORTED
+/* this function was added to libpng 1.2.0 and should always exist by default */
+void PNGAPI
+png_set_asm_flags (png_structp png_ptr, png_uint_32 asm_flags)
+{
+#ifdef PNG_MMX_CODE_SUPPORTED
+    png_uint_32 settable_asm_flags;
+    png_uint_32 settable_mmx_flags;
+#endif
+    if (png_ptr == NULL)
+       return;
+#ifdef PNG_MMX_CODE_SUPPORTED
+
+    settable_mmx_flags =
+#ifdef PNG_HAVE_MMX_COMBINE_ROW
+                         PNG_ASM_FLAG_MMX_READ_COMBINE_ROW  |
+#endif
+#ifdef PNG_HAVE_MMX_READ_INTERLACE
+                         PNG_ASM_FLAG_MMX_READ_INTERLACE    |
+#endif
+#ifdef PNG_HAVE_MMX_READ_FILTER_ROW
+                         PNG_ASM_FLAG_MMX_READ_FILTER_SUB   |
+                         PNG_ASM_FLAG_MMX_READ_FILTER_UP    |
+                         PNG_ASM_FLAG_MMX_READ_FILTER_AVG   |
+                         PNG_ASM_FLAG_MMX_READ_FILTER_PAETH |
+#endif
+                         0;
+
+    /* could be some non-MMX ones in the future, but not currently: */
+    settable_asm_flags = settable_mmx_flags;
+
+    if (!(png_ptr->asm_flags & PNG_ASM_FLAG_MMX_SUPPORT_COMPILED) ||
+        !(png_ptr->asm_flags & PNG_ASM_FLAG_MMX_SUPPORT_IN_CPU))
+    {
+        /* clear all MMX flags if MMX isn't supported */
+        settable_asm_flags &= ~settable_mmx_flags;
+        png_ptr->asm_flags &= ~settable_mmx_flags;
+    }
+
+    /* we're replacing the settable bits with those passed in by the user,
+     * so first zero them out of the master copy, then bitwise-OR in the
+     * allowed subset that was requested */
+
+    png_ptr->asm_flags &= ~settable_asm_flags;               /* zero them */
+    png_ptr->asm_flags |= (asm_flags & settable_asm_flags);  /* set them */
+#endif /* ?PNG_MMX_CODE_SUPPORTED */
+}
+
+/* this function was added to libpng 1.2.0 */
+void PNGAPI
+png_set_mmx_thresholds (png_structp png_ptr,
+                        png_byte mmx_bitdepth_threshold,
+                        png_uint_32 mmx_rowbytes_threshold)
+{
+    if (png_ptr == NULL)
+       return;
+#ifdef PNG_MMX_CODE_SUPPORTED
+    png_ptr->mmx_bitdepth_threshold = mmx_bitdepth_threshold;
+    png_ptr->mmx_rowbytes_threshold = mmx_rowbytes_threshold;
+#endif /* ?PNG_MMX_CODE_SUPPORTED */
+}
+#endif /* ?PNG_ASSEMBLER_CODE_SUPPORTED */
+
+#ifdef PNG_SET_USER_LIMITS_SUPPORTED
+/* this function was added to libpng 1.2.6 */
+void PNGAPI
+png_set_user_limits (png_structp png_ptr, png_uint_32 user_width_max,
+    png_uint_32 user_height_max)
+{
+    /* Images with dimensions larger than these limits will be
+     * rejected by png_set_IHDR().  To accept any PNG datastream
+     * regardless of dimensions, set both limits to 0x7ffffffL.
+     */
+    if(png_ptr == NULL) return;
+    png_ptr->user_width_max = user_width_max;
+    png_ptr->user_height_max = user_height_max;
+}
+#endif /* ?PNG_SET_USER_LIMITS_SUPPORTED */
+
+#endif /* ?PNG_1_0_X */
+#endif /* PNG_READ_SUPPORTED || PNG_WRITE_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngtrans.c b/distrib/libpng-1.2.19/pngtrans.c
new file mode 100644
index 0000000..1640095
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngtrans.c
@@ -0,0 +1,662 @@
+
+/* pngtrans.c - transforms the data in a row (used by both readers and writers)
+ *
+ * Last changed in libpng 1.2.17 May 15, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_READ_SUPPORTED) || defined(PNG_WRITE_SUPPORTED)
+#if defined(PNG_READ_BGR_SUPPORTED) || defined(PNG_WRITE_BGR_SUPPORTED)
+/* turn on BGR-to-RGB mapping */
+void PNGAPI
+png_set_bgr(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_bgr\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= PNG_BGR;
+}
+#endif
+
+#if defined(PNG_READ_SWAP_SUPPORTED) || defined(PNG_WRITE_SWAP_SUPPORTED)
+/* turn on 16 bit byte swapping */
+void PNGAPI
+png_set_swap(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_swap\n");
+   if(png_ptr == NULL) return;
+   if (png_ptr->bit_depth == 16)
+      png_ptr->transformations |= PNG_SWAP_BYTES;
+}
+#endif
+
+#if defined(PNG_READ_PACK_SUPPORTED) || defined(PNG_WRITE_PACK_SUPPORTED)
+/* turn on pixel packing */
+void PNGAPI
+png_set_packing(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_packing\n");
+   if(png_ptr == NULL) return;
+   if (png_ptr->bit_depth < 8)
+   {
+      png_ptr->transformations |= PNG_PACK;
+      png_ptr->usr_bit_depth = 8;
+   }
+}
+#endif
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)||defined(PNG_WRITE_PACKSWAP_SUPPORTED)
+/* turn on packed pixel swapping */
+void PNGAPI
+png_set_packswap(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_packswap\n");
+   if(png_ptr == NULL) return;
+   if (png_ptr->bit_depth < 8)
+      png_ptr->transformations |= PNG_PACKSWAP;
+}
+#endif
+
+#if defined(PNG_READ_SHIFT_SUPPORTED) || defined(PNG_WRITE_SHIFT_SUPPORTED)
+void PNGAPI
+png_set_shift(png_structp png_ptr, png_color_8p true_bits)
+{
+   png_debug(1, "in png_set_shift\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= PNG_SHIFT;
+   png_ptr->shift = *true_bits;
+}
+#endif
+
+#if defined(PNG_READ_INTERLACING_SUPPORTED) || \
+    defined(PNG_WRITE_INTERLACING_SUPPORTED)
+int PNGAPI
+png_set_interlace_handling(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_interlace handling\n");
+   if (png_ptr && png_ptr->interlaced)
+   {
+      png_ptr->transformations |= PNG_INTERLACE;
+      return (7);
+   }
+
+   return (1);
+}
+#endif
+
+#if defined(PNG_READ_FILLER_SUPPORTED) || defined(PNG_WRITE_FILLER_SUPPORTED)
+/* Add a filler byte on read, or remove a filler or alpha byte on write.
+ * The filler type has changed in v0.95 to allow future 2-byte fillers
+ * for 48-bit input data, as well as to avoid problems with some compilers
+ * that don't like bytes as parameters.
+ */
+void PNGAPI
+png_set_filler(png_structp png_ptr, png_uint_32 filler, int filler_loc)
+{
+   png_debug(1, "in png_set_filler\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= PNG_FILLER;
+   png_ptr->filler = (png_byte)filler;
+   if (filler_loc == PNG_FILLER_AFTER)
+      png_ptr->flags |= PNG_FLAG_FILLER_AFTER;
+   else
+      png_ptr->flags &= ~PNG_FLAG_FILLER_AFTER;
+
+   /* This should probably go in the "do_read_filler" routine.
+    * I attempted to do that in libpng-1.0.1a but that caused problems
+    * so I restored it in libpng-1.0.2a
+   */
+
+   if (png_ptr->color_type == PNG_COLOR_TYPE_RGB)
+   {
+      png_ptr->usr_channels = 4;
+   }
+
+   /* Also I added this in libpng-1.0.2a (what happens when we expand
+    * a less-than-8-bit grayscale to GA? */
+
+   if (png_ptr->color_type == PNG_COLOR_TYPE_GRAY && png_ptr->bit_depth >= 8)
+   {
+      png_ptr->usr_channels = 2;
+   }
+}
+
+#if !defined(PNG_1_0_X)
+/* Added to libpng-1.2.7 */
+void PNGAPI
+png_set_add_alpha(png_structp png_ptr, png_uint_32 filler, int filler_loc)
+{
+   png_debug(1, "in png_set_add_alpha\n");
+   if(png_ptr == NULL) return;
+   png_set_filler(png_ptr, filler, filler_loc);
+   png_ptr->transformations |= PNG_ADD_ALPHA;
+}
+#endif
+
+#endif
+
+#if defined(PNG_READ_SWAP_ALPHA_SUPPORTED) || \
+    defined(PNG_WRITE_SWAP_ALPHA_SUPPORTED)
+void PNGAPI
+png_set_swap_alpha(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_swap_alpha\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= PNG_SWAP_ALPHA;
+}
+#endif
+
+#if defined(PNG_READ_INVERT_ALPHA_SUPPORTED) || \
+    defined(PNG_WRITE_INVERT_ALPHA_SUPPORTED)
+void PNGAPI
+png_set_invert_alpha(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_invert_alpha\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= PNG_INVERT_ALPHA;
+}
+#endif
+
+#if defined(PNG_READ_INVERT_SUPPORTED) || defined(PNG_WRITE_INVERT_SUPPORTED)
+void PNGAPI
+png_set_invert_mono(png_structp png_ptr)
+{
+   png_debug(1, "in png_set_invert_mono\n");
+   if(png_ptr == NULL) return;
+   png_ptr->transformations |= PNG_INVERT_MONO;
+}
+
+/* invert monochrome grayscale data */
+void /* PRIVATE */
+png_do_invert(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_invert\n");
+  /* This test removed from libpng version 1.0.13 and 1.2.0:
+   *   if (row_info->bit_depth == 1 &&
+   */
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row == NULL || row_info == NULL)
+     return;
+#endif
+   if (row_info->color_type == PNG_COLOR_TYPE_GRAY)
+   {
+      png_bytep rp = row;
+      png_uint_32 i;
+      png_uint_32 istop = row_info->rowbytes;
+
+      for (i = 0; i < istop; i++)
+      {
+         *rp = (png_byte)(~(*rp));
+         rp++;
+      }
+   }
+   else if (row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA &&
+      row_info->bit_depth == 8)
+   {
+      png_bytep rp = row;
+      png_uint_32 i;
+      png_uint_32 istop = row_info->rowbytes;
+
+      for (i = 0; i < istop; i+=2)
+      {
+         *rp = (png_byte)(~(*rp));
+         rp+=2;
+      }
+   }
+   else if (row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA &&
+      row_info->bit_depth == 16)
+   {
+      png_bytep rp = row;
+      png_uint_32 i;
+      png_uint_32 istop = row_info->rowbytes;
+
+      for (i = 0; i < istop; i+=4)
+      {
+         *rp = (png_byte)(~(*rp));
+         *(rp+1) = (png_byte)(~(*(rp+1)));
+         rp+=4;
+      }
+   }
+}
+#endif
+
+#if defined(PNG_READ_SWAP_SUPPORTED) || defined(PNG_WRITE_SWAP_SUPPORTED)
+/* swaps byte order on 16 bit depth images */
+void /* PRIVATE */
+png_do_swap(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_swap\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+       row_info->bit_depth == 16)
+   {
+      png_bytep rp = row;
+      png_uint_32 i;
+      png_uint_32 istop= row_info->width * row_info->channels;
+
+      for (i = 0; i < istop; i++, rp += 2)
+      {
+         png_byte t = *rp;
+         *rp = *(rp + 1);
+         *(rp + 1) = t;
+      }
+   }
+}
+#endif
+
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)||defined(PNG_WRITE_PACKSWAP_SUPPORTED)
+static PNG_CONST png_byte onebppswaptable[256] = {
+   0x00, 0x80, 0x40, 0xC0, 0x20, 0xA0, 0x60, 0xE0,
+   0x10, 0x90, 0x50, 0xD0, 0x30, 0xB0, 0x70, 0xF0,
+   0x08, 0x88, 0x48, 0xC8, 0x28, 0xA8, 0x68, 0xE8,
+   0x18, 0x98, 0x58, 0xD8, 0x38, 0xB8, 0x78, 0xF8,
+   0x04, 0x84, 0x44, 0xC4, 0x24, 0xA4, 0x64, 0xE4,
+   0x14, 0x94, 0x54, 0xD4, 0x34, 0xB4, 0x74, 0xF4,
+   0x0C, 0x8C, 0x4C, 0xCC, 0x2C, 0xAC, 0x6C, 0xEC,
+   0x1C, 0x9C, 0x5C, 0xDC, 0x3C, 0xBC, 0x7C, 0xFC,
+   0x02, 0x82, 0x42, 0xC2, 0x22, 0xA2, 0x62, 0xE2,
+   0x12, 0x92, 0x52, 0xD2, 0x32, 0xB2, 0x72, 0xF2,
+   0x0A, 0x8A, 0x4A, 0xCA, 0x2A, 0xAA, 0x6A, 0xEA,
+   0x1A, 0x9A, 0x5A, 0xDA, 0x3A, 0xBA, 0x7A, 0xFA,
+   0x06, 0x86, 0x46, 0xC6, 0x26, 0xA6, 0x66, 0xE6,
+   0x16, 0x96, 0x56, 0xD6, 0x36, 0xB6, 0x76, 0xF6,
+   0x0E, 0x8E, 0x4E, 0xCE, 0x2E, 0xAE, 0x6E, 0xEE,
+   0x1E, 0x9E, 0x5E, 0xDE, 0x3E, 0xBE, 0x7E, 0xFE,
+   0x01, 0x81, 0x41, 0xC1, 0x21, 0xA1, 0x61, 0xE1,
+   0x11, 0x91, 0x51, 0xD1, 0x31, 0xB1, 0x71, 0xF1,
+   0x09, 0x89, 0x49, 0xC9, 0x29, 0xA9, 0x69, 0xE9,
+   0x19, 0x99, 0x59, 0xD9, 0x39, 0xB9, 0x79, 0xF9,
+   0x05, 0x85, 0x45, 0xC5, 0x25, 0xA5, 0x65, 0xE5,
+   0x15, 0x95, 0x55, 0xD5, 0x35, 0xB5, 0x75, 0xF5,
+   0x0D, 0x8D, 0x4D, 0xCD, 0x2D, 0xAD, 0x6D, 0xED,
+   0x1D, 0x9D, 0x5D, 0xDD, 0x3D, 0xBD, 0x7D, 0xFD,
+   0x03, 0x83, 0x43, 0xC3, 0x23, 0xA3, 0x63, 0xE3,
+   0x13, 0x93, 0x53, 0xD3, 0x33, 0xB3, 0x73, 0xF3,
+   0x0B, 0x8B, 0x4B, 0xCB, 0x2B, 0xAB, 0x6B, 0xEB,
+   0x1B, 0x9B, 0x5B, 0xDB, 0x3B, 0xBB, 0x7B, 0xFB,
+   0x07, 0x87, 0x47, 0xC7, 0x27, 0xA7, 0x67, 0xE7,
+   0x17, 0x97, 0x57, 0xD7, 0x37, 0xB7, 0x77, 0xF7,
+   0x0F, 0x8F, 0x4F, 0xCF, 0x2F, 0xAF, 0x6F, 0xEF,
+   0x1F, 0x9F, 0x5F, 0xDF, 0x3F, 0xBF, 0x7F, 0xFF
+};
+
+static PNG_CONST png_byte twobppswaptable[256] = {
+   0x00, 0x40, 0x80, 0xC0, 0x10, 0x50, 0x90, 0xD0,
+   0x20, 0x60, 0xA0, 0xE0, 0x30, 0x70, 0xB0, 0xF0,
+   0x04, 0x44, 0x84, 0xC4, 0x14, 0x54, 0x94, 0xD4,
+   0x24, 0x64, 0xA4, 0xE4, 0x34, 0x74, 0xB4, 0xF4,
+   0x08, 0x48, 0x88, 0xC8, 0x18, 0x58, 0x98, 0xD8,
+   0x28, 0x68, 0xA8, 0xE8, 0x38, 0x78, 0xB8, 0xF8,
+   0x0C, 0x4C, 0x8C, 0xCC, 0x1C, 0x5C, 0x9C, 0xDC,
+   0x2C, 0x6C, 0xAC, 0xEC, 0x3C, 0x7C, 0xBC, 0xFC,
+   0x01, 0x41, 0x81, 0xC1, 0x11, 0x51, 0x91, 0xD1,
+   0x21, 0x61, 0xA1, 0xE1, 0x31, 0x71, 0xB1, 0xF1,
+   0x05, 0x45, 0x85, 0xC5, 0x15, 0x55, 0x95, 0xD5,
+   0x25, 0x65, 0xA5, 0xE5, 0x35, 0x75, 0xB5, 0xF5,
+   0x09, 0x49, 0x89, 0xC9, 0x19, 0x59, 0x99, 0xD9,
+   0x29, 0x69, 0xA9, 0xE9, 0x39, 0x79, 0xB9, 0xF9,
+   0x0D, 0x4D, 0x8D, 0xCD, 0x1D, 0x5D, 0x9D, 0xDD,
+   0x2D, 0x6D, 0xAD, 0xED, 0x3D, 0x7D, 0xBD, 0xFD,
+   0x02, 0x42, 0x82, 0xC2, 0x12, 0x52, 0x92, 0xD2,
+   0x22, 0x62, 0xA2, 0xE2, 0x32, 0x72, 0xB2, 0xF2,
+   0x06, 0x46, 0x86, 0xC6, 0x16, 0x56, 0x96, 0xD6,
+   0x26, 0x66, 0xA6, 0xE6, 0x36, 0x76, 0xB6, 0xF6,
+   0x0A, 0x4A, 0x8A, 0xCA, 0x1A, 0x5A, 0x9A, 0xDA,
+   0x2A, 0x6A, 0xAA, 0xEA, 0x3A, 0x7A, 0xBA, 0xFA,
+   0x0E, 0x4E, 0x8E, 0xCE, 0x1E, 0x5E, 0x9E, 0xDE,
+   0x2E, 0x6E, 0xAE, 0xEE, 0x3E, 0x7E, 0xBE, 0xFE,
+   0x03, 0x43, 0x83, 0xC3, 0x13, 0x53, 0x93, 0xD3,
+   0x23, 0x63, 0xA3, 0xE3, 0x33, 0x73, 0xB3, 0xF3,
+   0x07, 0x47, 0x87, 0xC7, 0x17, 0x57, 0x97, 0xD7,
+   0x27, 0x67, 0xA7, 0xE7, 0x37, 0x77, 0xB7, 0xF7,
+   0x0B, 0x4B, 0x8B, 0xCB, 0x1B, 0x5B, 0x9B, 0xDB,
+   0x2B, 0x6B, 0xAB, 0xEB, 0x3B, 0x7B, 0xBB, 0xFB,
+   0x0F, 0x4F, 0x8F, 0xCF, 0x1F, 0x5F, 0x9F, 0xDF,
+   0x2F, 0x6F, 0xAF, 0xEF, 0x3F, 0x7F, 0xBF, 0xFF
+};
+
+static PNG_CONST png_byte fourbppswaptable[256] = {
+   0x00, 0x10, 0x20, 0x30, 0x40, 0x50, 0x60, 0x70,
+   0x80, 0x90, 0xA0, 0xB0, 0xC0, 0xD0, 0xE0, 0xF0,
+   0x01, 0x11, 0x21, 0x31, 0x41, 0x51, 0x61, 0x71,
+   0x81, 0x91, 0xA1, 0xB1, 0xC1, 0xD1, 0xE1, 0xF1,
+   0x02, 0x12, 0x22, 0x32, 0x42, 0x52, 0x62, 0x72,
+   0x82, 0x92, 0xA2, 0xB2, 0xC2, 0xD2, 0xE2, 0xF2,
+   0x03, 0x13, 0x23, 0x33, 0x43, 0x53, 0x63, 0x73,
+   0x83, 0x93, 0xA3, 0xB3, 0xC3, 0xD3, 0xE3, 0xF3,
+   0x04, 0x14, 0x24, 0x34, 0x44, 0x54, 0x64, 0x74,
+   0x84, 0x94, 0xA4, 0xB4, 0xC4, 0xD4, 0xE4, 0xF4,
+   0x05, 0x15, 0x25, 0x35, 0x45, 0x55, 0x65, 0x75,
+   0x85, 0x95, 0xA5, 0xB5, 0xC5, 0xD5, 0xE5, 0xF5,
+   0x06, 0x16, 0x26, 0x36, 0x46, 0x56, 0x66, 0x76,
+   0x86, 0x96, 0xA6, 0xB6, 0xC6, 0xD6, 0xE6, 0xF6,
+   0x07, 0x17, 0x27, 0x37, 0x47, 0x57, 0x67, 0x77,
+   0x87, 0x97, 0xA7, 0xB7, 0xC7, 0xD7, 0xE7, 0xF7,
+   0x08, 0x18, 0x28, 0x38, 0x48, 0x58, 0x68, 0x78,
+   0x88, 0x98, 0xA8, 0xB8, 0xC8, 0xD8, 0xE8, 0xF8,
+   0x09, 0x19, 0x29, 0x39, 0x49, 0x59, 0x69, 0x79,
+   0x89, 0x99, 0xA9, 0xB9, 0xC9, 0xD9, 0xE9, 0xF9,
+   0x0A, 0x1A, 0x2A, 0x3A, 0x4A, 0x5A, 0x6A, 0x7A,
+   0x8A, 0x9A, 0xAA, 0xBA, 0xCA, 0xDA, 0xEA, 0xFA,
+   0x0B, 0x1B, 0x2B, 0x3B, 0x4B, 0x5B, 0x6B, 0x7B,
+   0x8B, 0x9B, 0xAB, 0xBB, 0xCB, 0xDB, 0xEB, 0xFB,
+   0x0C, 0x1C, 0x2C, 0x3C, 0x4C, 0x5C, 0x6C, 0x7C,
+   0x8C, 0x9C, 0xAC, 0xBC, 0xCC, 0xDC, 0xEC, 0xFC,
+   0x0D, 0x1D, 0x2D, 0x3D, 0x4D, 0x5D, 0x6D, 0x7D,
+   0x8D, 0x9D, 0xAD, 0xBD, 0xCD, 0xDD, 0xED, 0xFD,
+   0x0E, 0x1E, 0x2E, 0x3E, 0x4E, 0x5E, 0x6E, 0x7E,
+   0x8E, 0x9E, 0xAE, 0xBE, 0xCE, 0xDE, 0xEE, 0xFE,
+   0x0F, 0x1F, 0x2F, 0x3F, 0x4F, 0x5F, 0x6F, 0x7F,
+   0x8F, 0x9F, 0xAF, 0xBF, 0xCF, 0xDF, 0xEF, 0xFF
+};
+
+/* swaps pixel packing order within bytes */
+void /* PRIVATE */
+png_do_packswap(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_packswap\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+       row_info->bit_depth < 8)
+   {
+      png_bytep rp, end, table;
+
+      end = row + row_info->rowbytes;
+
+      if (row_info->bit_depth == 1)
+         table = (png_bytep)onebppswaptable;
+      else if (row_info->bit_depth == 2)
+         table = (png_bytep)twobppswaptable;
+      else if (row_info->bit_depth == 4)
+         table = (png_bytep)fourbppswaptable;
+      else
+         return;
+
+      for (rp = row; rp < end; rp++)
+         *rp = table[*rp];
+   }
+}
+#endif /* PNG_READ_PACKSWAP_SUPPORTED or PNG_WRITE_PACKSWAP_SUPPORTED */
+
+#if defined(PNG_WRITE_FILLER_SUPPORTED) || \
+    defined(PNG_READ_STRIP_ALPHA_SUPPORTED)
+/* remove filler or alpha byte(s) */
+void /* PRIVATE */
+png_do_strip_filler(png_row_infop row_info, png_bytep row, png_uint_32 flags)
+{
+   png_debug(1, "in png_do_strip_filler\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL)
+#endif
+   {
+      png_bytep sp=row;
+      png_bytep dp=row;
+      png_uint_32 row_width=row_info->width;
+      png_uint_32 i;
+
+      if ((row_info->color_type == PNG_COLOR_TYPE_RGB ||
+         (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA &&
+         (flags & PNG_FLAG_STRIP_ALPHA))) &&
+         row_info->channels == 4)
+      {
+         if (row_info->bit_depth == 8)
+         {
+            /* This converts from RGBX or RGBA to RGB */
+            if (flags & PNG_FLAG_FILLER_AFTER)
+            {
+               dp+=3; sp+=4;
+               for (i = 1; i < row_width; i++)
+               {
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  sp++;
+               }
+            }
+            /* This converts from XRGB or ARGB to RGB */
+            else
+            {
+               for (i = 0; i < row_width; i++)
+               {
+                  sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+               }
+            }
+            row_info->pixel_depth = 24;
+            row_info->rowbytes = row_width * 3;
+         }
+         else /* if (row_info->bit_depth == 16) */
+         {
+            if (flags & PNG_FLAG_FILLER_AFTER)
+            {
+               /* This converts from RRGGBBXX or RRGGBBAA to RRGGBB */
+               sp += 8; dp += 6;
+               for (i = 1; i < row_width; i++)
+               {
+                  /* This could be (although png_memcpy is probably slower):
+                  png_memcpy(dp, sp, 6);
+                  sp += 8;
+                  dp += 6;
+                  */
+
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  sp += 2;
+               }
+            }
+            else
+            {
+               /* This converts from XXRRGGBB or AARRGGBB to RRGGBB */
+               for (i = 0; i < row_width; i++)
+               {
+                  /* This could be (although png_memcpy is probably slower):
+                  png_memcpy(dp, sp, 6);
+                  sp += 8;
+                  dp += 6;
+                  */
+
+                  sp+=2;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+               }
+            }
+            row_info->pixel_depth = 48;
+            row_info->rowbytes = row_width * 6;
+         }
+         row_info->channels = 3;
+      }
+      else if ((row_info->color_type == PNG_COLOR_TYPE_GRAY ||
+         (row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA &&
+         (flags & PNG_FLAG_STRIP_ALPHA))) &&
+          row_info->channels == 2)
+      {
+         if (row_info->bit_depth == 8)
+         {
+            /* This converts from GX or GA to G */
+            if (flags & PNG_FLAG_FILLER_AFTER)
+            {
+               for (i = 0; i < row_width; i++)
+               {
+                  *dp++ = *sp++;
+                  sp++;
+               }
+            }
+            /* This converts from XG or AG to G */
+            else
+            {
+               for (i = 0; i < row_width; i++)
+               {
+                  sp++;
+                  *dp++ = *sp++;
+               }
+            }
+            row_info->pixel_depth = 8;
+            row_info->rowbytes = row_width;
+         }
+         else /* if (row_info->bit_depth == 16) */
+         {
+            if (flags & PNG_FLAG_FILLER_AFTER)
+            {
+               /* This converts from GGXX or GGAA to GG */
+               sp += 4; dp += 2;
+               for (i = 1; i < row_width; i++)
+               {
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+                  sp += 2;
+               }
+            }
+            else
+            {
+               /* This converts from XXGG or AAGG to GG */
+               for (i = 0; i < row_width; i++)
+               {
+                  sp += 2;
+                  *dp++ = *sp++;
+                  *dp++ = *sp++;
+               }
+            }
+            row_info->pixel_depth = 16;
+            row_info->rowbytes = row_width * 2;
+         }
+         row_info->channels = 1;
+      }
+      if (flags & PNG_FLAG_STRIP_ALPHA)
+        row_info->color_type &= ~PNG_COLOR_MASK_ALPHA;
+   }
+}
+#endif
+
+#if defined(PNG_READ_BGR_SUPPORTED) || defined(PNG_WRITE_BGR_SUPPORTED)
+/* swaps red and blue bytes within a pixel */
+void /* PRIVATE */
+png_do_bgr(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_bgr\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+       (row_info->color_type & PNG_COLOR_MASK_COLOR))
+   {
+      png_uint_32 row_width = row_info->width;
+      if (row_info->bit_depth == 8)
+      {
+         if (row_info->color_type == PNG_COLOR_TYPE_RGB)
+         {
+            png_bytep rp;
+            png_uint_32 i;
+
+            for (i = 0, rp = row; i < row_width; i++, rp += 3)
+            {
+               png_byte save = *rp;
+               *rp = *(rp + 2);
+               *(rp + 2) = save;
+            }
+         }
+         else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+         {
+            png_bytep rp;
+            png_uint_32 i;
+
+            for (i = 0, rp = row; i < row_width; i++, rp += 4)
+            {
+               png_byte save = *rp;
+               *rp = *(rp + 2);
+               *(rp + 2) = save;
+            }
+         }
+      }
+      else if (row_info->bit_depth == 16)
+      {
+         if (row_info->color_type == PNG_COLOR_TYPE_RGB)
+         {
+            png_bytep rp;
+            png_uint_32 i;
+
+            for (i = 0, rp = row; i < row_width; i++, rp += 6)
+            {
+               png_byte save = *rp;
+               *rp = *(rp + 4);
+               *(rp + 4) = save;
+               save = *(rp + 1);
+               *(rp + 1) = *(rp + 5);
+               *(rp + 5) = save;
+            }
+         }
+         else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+         {
+            png_bytep rp;
+            png_uint_32 i;
+
+            for (i = 0, rp = row; i < row_width; i++, rp += 8)
+            {
+               png_byte save = *rp;
+               *rp = *(rp + 4);
+               *(rp + 4) = save;
+               save = *(rp + 1);
+               *(rp + 1) = *(rp + 5);
+               *(rp + 5) = save;
+            }
+         }
+      }
+   }
+}
+#endif /* PNG_READ_BGR_SUPPORTED or PNG_WRITE_BGR_SUPPORTED */
+
+#if defined(PNG_READ_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED) || \
+    defined(PNG_LEGACY_SUPPORTED)
+void PNGAPI
+png_set_user_transform_info(png_structp png_ptr, png_voidp
+   user_transform_ptr, int user_transform_depth, int user_transform_channels)
+{
+   png_debug(1, "in png_set_user_transform_info\n");
+   if(png_ptr == NULL) return;
+#if defined(PNG_USER_TRANSFORM_PTR_SUPPORTED)
+   png_ptr->user_transform_ptr = user_transform_ptr;
+   png_ptr->user_transform_depth = (png_byte)user_transform_depth;
+   png_ptr->user_transform_channels = (png_byte)user_transform_channels;
+#else
+   if(user_transform_ptr || user_transform_depth || user_transform_channels)
+      png_warning(png_ptr,
+        "This version of libpng does not support user transform info");
+#endif
+}
+#endif
+
+/* This function returns a pointer to the user_transform_ptr associated with
+ * the user transform functions.  The application should free any memory
+ * associated with this pointer before png_write_destroy and png_read_destroy
+ * are called.
+ */
+png_voidp PNGAPI
+png_get_user_transform_ptr(png_structp png_ptr)
+{
+#if defined(PNG_USER_TRANSFORM_PTR_SUPPORTED)
+   if (png_ptr == NULL) return (NULL);
+   return ((png_voidp)png_ptr->user_transform_ptr);
+#else
+   return (NULL);
+#endif
+}
+#endif /* PNG_READ_SUPPORTED || PNG_WRITE_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngvcrd.c b/distrib/libpng-1.2.19/pngvcrd.c
new file mode 100644
index 0000000..34d42c9
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngvcrd.c
@@ -0,0 +1,3922 @@
+
+/* pngvcrd.c - mixed C/assembler version of utilities to read a PNG file
+ *
+ * For Intel x86 CPU and Microsoft Visual C++ compiler
+ *
+ * Last changed in libpng 1.2.19 August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * Copyright (c) 1998, Intel Corporation
+ *
+ * Contributed by Nirav Chhatrapati, Intel Corporation, 1998
+ * Interface to libpng contributed by Gilles Vollant, 1999
+ *
+ *
+ * In png_do_read_interlace() in libpng versions 1.0.3a through 1.0.4d,
+ * a sign error in the post-MMX cleanup code for each pixel_depth resulted
+ * in bad pixels at the beginning of some rows of some images, and also
+ * (due to out-of-range memory reads and writes) caused heap corruption
+ * when compiled with MSVC 6.0.  The error was fixed in version 1.0.4e.
+ *
+ * [png_read_filter_row_mmx_avg() bpp == 2 bugfix, GRR 20000916]
+ *
+ * [runtime MMX configuration, GRR 20010102]
+ *
+ * [Copy 6 bytes per pixel, not 4, and use stride of 6, not 4, in the
+ *  second loop of interlace processing of 48-bit pixels, GR-P 20070717]
+ *
+ * [move instances of uAll union into local, except for two constant
+ * instances, GR-P 20070805]
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+
+#if defined(PNG_MMX_CODE_SUPPORTED) && defined(PNG_USE_PNGVCRD)
+
+
+static int mmx_supported=2;
+
+int PNGAPI
+png_mmx_support(void)
+{
+  int mmx_supported_local = 0;
+  _asm {
+    push ebx          //CPUID will trash these
+    push ecx
+    push edx
+
+    pushfd            //Save Eflag to stack
+    pop eax           //Get Eflag from stack into eax
+    mov ecx, eax      //Make another copy of Eflag in ecx
+    xor eax, 0x200000 //Toggle ID bit in Eflag [i.e. bit(21)]
+    push eax          //Save modified Eflag back to stack
+
+    popfd             //Restored modified value back to Eflag reg
+    pushfd            //Save Eflag to stack
+    pop eax           //Get Eflag from stack
+    push ecx          // save original Eflag to stack
+    popfd             // restore original Eflag
+    xor eax, ecx      //Compare the new Eflag with the original Eflag
+    jz NOT_SUPPORTED  //If the same, CPUID instruction is not supported,
+                      //skip following instructions and jump to
+                      //NOT_SUPPORTED label
+
+    xor eax, eax      //Set eax to zero
+
+    _asm _emit 0x0f   //CPUID instruction  (two bytes opcode)
+    _asm _emit 0xa2
+
+    cmp eax, 1        //make sure eax return non-zero value
+    jl NOT_SUPPORTED  //If eax is zero, mmx not supported
+
+    xor eax, eax      //set eax to zero
+    inc eax           //Now increment eax to 1.  This instruction is
+                      //faster than the instruction "mov eax, 1"
+
+    _asm _emit 0x0f   //CPUID instruction
+    _asm _emit 0xa2
+
+    and edx, 0x00800000  //mask out all bits but mmx bit(24)
+    cmp edx, 0        // 0 = mmx not supported
+    jz  NOT_SUPPORTED // non-zero = Yes, mmx IS supported
+
+    mov  mmx_supported_local, 1  //set return value to 1
+
+NOT_SUPPORTED:
+    mov  eax, mmx_supported_local  //move return value to eax
+    pop edx          //CPUID trashed these
+    pop ecx
+    pop ebx
+  }
+
+  //mmx_supported_local=0; // test code for force don't support MMX
+  //printf("MMX : %u (1=MMX supported)\n",mmx_supported_local);
+
+  mmx_supported = mmx_supported_local;
+  return mmx_supported_local;
+}
+
+/* Combines the row recently read in with the previous row.
+   This routine takes care of alpha and transparency if requested.
+   This routine also handles the two methods of progressive display
+   of interlaced images, depending on the mask value.
+   The mask value describes which pixels are to be combined with
+   the row.  The pattern always repeats every 8 pixels, so just 8
+   bits are needed.  A one indicates the pixel is to be combined; a
+   zero indicates the pixel is to be skipped.  This is in addition
+   to any alpha or transparency value associated with the pixel.  If
+   you want all pixels to be combined, pass 0xff (255) in mask.  */
+
+/* Use this routine for x86 platform - uses faster MMX routine if machine
+   supports MMX */
+
+void /* PRIVATE */
+png_combine_row(png_structp png_ptr, png_bytep row, int mask)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_CONST int png_pass_inc[7] = {8, 8, 4, 4, 2, 2, 1};
+#endif
+
+   png_debug(1,"in png_combine_row_asm\n");
+
+   if (mmx_supported == 2) {
+#if !defined(PNG_1_0_X)
+       /* this should have happened in png_init_mmx_flags() already */
+       png_warning(png_ptr, "asm_flags may not have been initialized");
+#endif
+       png_mmx_support();
+   }
+
+   if (mask == 0xff)
+   {
+      png_memcpy(row, png_ptr->row_buf + 1,
+       (png_size_t)PNG_ROWBYTES(png_ptr->row_info.pixel_depth,
+       png_ptr->width));
+   }
+   /* GRR:  add "else if (mask == 0)" case?
+    *       or does png_combine_row() not even get called in that case? */
+   else
+   {
+      switch (png_ptr->row_info.pixel_depth)
+      {
+         case 24:
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+            png_uint_32 len;
+            int unmask, diff;
+
+            __int64 mask2=0x0101010202020404,  //24bpp
+                    mask1=0x0408080810101020,
+                    mask0=0x2020404040808080;
+
+            srcptr = png_ptr->row_buf + 1;
+            dstptr = row;
+
+            unmask = ~mask;
+            len     = (png_ptr->width)&~7;
+            diff = (png_ptr->width)&7;
+
+#if !defined(PNG_1_0_X)
+            if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+                /* && mmx_supported */ )
+#else
+            if (mmx_supported)
+#endif
+            {
+               _asm
+               {
+                  movd       mm7, unmask       //load bit pattern
+                  psubb      mm6,mm6           //zero mm6
+                  punpcklbw  mm7,mm7
+                  punpcklwd  mm7,mm7
+                  punpckldq  mm7,mm7           //fill register with 8 masks
+
+                  movq       mm0,mask0
+                  movq       mm1,mask1
+                  movq       mm2,mask2
+
+                  pand       mm0,mm7
+                  pand       mm1,mm7
+                  pand       mm2,mm7
+
+                  pcmpeqb    mm0,mm6
+                  pcmpeqb    mm1,mm6
+                  pcmpeqb    mm2,mm6
+
+                  mov        ecx,len           //load length of line
+                  mov        esi,srcptr        //load source
+                  mov        ebx,dstptr        //load dest
+                  cmp        ecx,0
+                  jz         mainloop24end
+
+mainloop24:
+                  movq       mm4,[esi]
+                  pand       mm4,mm0
+                  movq       mm6,mm0
+                  movq       mm7,[ebx]
+                  pandn      mm6,mm7
+                  por        mm4,mm6
+                  movq       [ebx],mm4
+
+
+                  movq       mm5,[esi+8]
+                  pand       mm5,mm1
+                  movq       mm7,mm1
+                  movq       mm6,[ebx+8]
+                  pandn      mm7,mm6
+                  por        mm5,mm7
+                  movq       [ebx+8],mm5
+
+                  movq       mm6,[esi+16]
+                  pand       mm6,mm2
+                  movq       mm4,mm2
+                  movq       mm7,[ebx+16]
+                  pandn      mm4,mm7
+                  por        mm6,mm4
+                  movq       [ebx+16],mm6
+
+                  add        esi,24            //inc by 24 bytes processed
+                  add        ebx,24
+                  sub        ecx,8             //dec by 8 pixels processed
+
+                  ja         mainloop24
+
+mainloop24end:
+                  mov        ecx,diff
+                  cmp        ecx,0
+                  jz         end24
+
+                  mov        edx,mask
+                  sal        edx,24            //make low byte the high byte
+secondloop24:
+                  sal        edx,1             //move high bit to CF
+                  jnc        skip24            //if CF = 0
+                  mov        ax,[esi]
+                  mov        [ebx],ax
+                  xor        eax,eax
+                  mov        al,[esi+2]
+                  mov        [ebx+2],al
+skip24:
+                  add        esi,3
+                  add        ebx,3
+
+                  dec        ecx
+                  jnz        secondloop24
+
+end24:
+                  emms
+               }
+            }
+            else /* mmx not supported - use modified C routine */
+            {
+               register unsigned int incr1, initial_val, final_val;
+               png_size_t pixel_bytes;
+               png_uint_32 i;
+               register int disp = png_pass_inc[png_ptr->pass];
+               int offset_table[7] = {0, 4, 0, 2, 0, 1, 0};
+
+               pixel_bytes = (png_ptr->row_info.pixel_depth >> 3);
+               srcptr = png_ptr->row_buf + 1 + offset_table[png_ptr->pass]*
+                  pixel_bytes;
+               dstptr = row + offset_table[png_ptr->pass]*pixel_bytes;
+               initial_val = offset_table[png_ptr->pass]*pixel_bytes;
+               final_val = png_ptr->width*pixel_bytes;
+               incr1 = (disp)*pixel_bytes;
+               for (i = initial_val; i < final_val; i += incr1)
+               {
+                  png_memcpy(dstptr, srcptr, pixel_bytes);
+                  srcptr += incr1;
+                  dstptr += incr1;
+               }
+            } /* end of else */
+
+            break;
+         }       // end 24 bpp
+
+         case 32:
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+            png_uint_32 len;
+            int unmask, diff;
+
+            __int64 mask3=0x0101010102020202,  //32bpp
+                    mask2=0x0404040408080808,
+                    mask1=0x1010101020202020,
+                    mask0=0x4040404080808080;
+
+            srcptr = png_ptr->row_buf + 1;
+            dstptr = row;
+
+            unmask = ~mask;
+            len     = (png_ptr->width)&~7;
+            diff = (png_ptr->width)&7;
+
+#if !defined(PNG_1_0_X)
+            if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+                /* && mmx_supported */ )
+#else
+            if (mmx_supported)
+#endif
+            {
+               _asm
+               {
+                  movd       mm7, unmask       //load bit pattern
+                  psubb      mm6,mm6           //zero mm6
+                  punpcklbw  mm7,mm7
+                  punpcklwd  mm7,mm7
+                  punpckldq  mm7,mm7           //fill register with 8 masks
+
+                  movq       mm0,mask0
+                  movq       mm1,mask1
+                  movq       mm2,mask2
+                  movq       mm3,mask3
+
+                  pand       mm0,mm7
+                  pand       mm1,mm7
+                  pand       mm2,mm7
+                  pand       mm3,mm7
+
+                  pcmpeqb    mm0,mm6
+                  pcmpeqb    mm1,mm6
+                  pcmpeqb    mm2,mm6
+                  pcmpeqb    mm3,mm6
+
+                  mov        ecx,len           //load length of line
+                  mov        esi,srcptr        //load source
+                  mov        ebx,dstptr        //load dest
+
+                  cmp        ecx,0             //lcr
+                  jz         mainloop32end
+
+mainloop32:
+                  movq       mm4,[esi]
+                  pand       mm4,mm0
+                  movq       mm6,mm0
+                  movq       mm7,[ebx]
+                  pandn      mm6,mm7
+                  por        mm4,mm6
+                  movq       [ebx],mm4
+
+                  movq       mm5,[esi+8]
+                  pand       mm5,mm1
+                  movq       mm7,mm1
+                  movq       mm6,[ebx+8]
+                  pandn      mm7,mm6
+                  por        mm5,mm7
+                  movq       [ebx+8],mm5
+
+                  movq       mm6,[esi+16]
+                  pand       mm6,mm2
+                  movq       mm4,mm2
+                  movq       mm7,[ebx+16]
+                  pandn      mm4,mm7
+                  por        mm6,mm4
+                  movq       [ebx+16],mm6
+
+                  movq       mm7,[esi+24]
+                  pand       mm7,mm3
+                  movq       mm5,mm3
+                  movq       mm4,[ebx+24]
+                  pandn      mm5,mm4
+                  por        mm7,mm5
+                  movq       [ebx+24],mm7
+
+                  add        esi,32            //inc by 32 bytes processed
+                  add        ebx,32
+                  sub        ecx,8             //dec by 8 pixels processed
+
+                  ja         mainloop32
+
+mainloop32end:
+                  mov        ecx,diff
+                  cmp        ecx,0
+                  jz         end32
+
+                  mov        edx,mask
+                  sal        edx,24            //make low byte the high byte
+secondloop32:
+                  sal        edx,1             //move high bit to CF
+                  jnc        skip32            //if CF = 0
+                  mov        eax,[esi]
+                  mov        [ebx],eax
+skip32:
+                  add        esi,4
+                  add        ebx,4
+
+                  dec        ecx
+                  jnz        secondloop32
+
+end32:
+                  emms
+               }
+            }
+            else /* mmx _not supported - Use modified C routine */
+            {
+               register unsigned int incr1, initial_val, final_val;
+               png_size_t pixel_bytes;
+               png_uint_32 i;
+               register int disp = png_pass_inc[png_ptr->pass];
+               int offset_table[7] = {0, 4, 0, 2, 0, 1, 0};
+
+               pixel_bytes = (png_ptr->row_info.pixel_depth >> 3);
+               srcptr = png_ptr->row_buf + 1 + offset_table[png_ptr->pass]*
+                  pixel_bytes;
+               dstptr = row + offset_table[png_ptr->pass]*pixel_bytes;
+               initial_val = offset_table[png_ptr->pass]*pixel_bytes;
+               final_val = png_ptr->width*pixel_bytes;
+               incr1 = (disp)*pixel_bytes;
+               for (i = initial_val; i < final_val; i += incr1)
+               {
+                  png_memcpy(dstptr, srcptr, pixel_bytes);
+                  srcptr += incr1;
+                  dstptr += incr1;
+               }
+            } /* end of else */
+
+            break;
+         }       // end 32 bpp
+
+         case 8:
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+            png_uint_32 len;
+            int m;
+            int diff, unmask;
+
+            __int64 mask0=0x0102040810204080;
+
+#if !defined(PNG_1_0_X)
+            if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+                /* && mmx_supported */ )
+#else
+            if (mmx_supported)
+#endif
+            {
+               srcptr = png_ptr->row_buf + 1;
+               dstptr = row;
+               m = 0x80;
+               unmask = ~mask;
+               len  = png_ptr->width &~7;  //reduce to multiple of 8
+               diff = png_ptr->width & 7;  //amount lost
+
+               _asm
+               {
+                  movd       mm7, unmask   //load bit pattern
+                  psubb      mm6,mm6       //zero mm6
+                  punpcklbw  mm7,mm7
+                  punpcklwd  mm7,mm7
+                  punpckldq  mm7,mm7       //fill register with 8 masks
+
+                  movq       mm0,mask0
+
+                  pand       mm0,mm7       //nonzero if keep byte
+                  pcmpeqb    mm0,mm6       //zeros->1s, v versa
+
+                  mov        ecx,len       //load length of line (pixels)
+                  mov        esi,srcptr    //load source
+                  mov        ebx,dstptr    //load dest
+                  cmp        ecx,0         //lcr
+                  je         mainloop8end
+
+mainloop8:
+                  movq       mm4,[esi]
+                  pand       mm4,mm0
+                  movq       mm6,mm0
+                  pandn      mm6,[ebx]
+                  por        mm4,mm6
+                  movq       [ebx],mm4
+
+                  add        esi,8         //inc by 8 bytes processed
+                  add        ebx,8
+                  sub        ecx,8         //dec by 8 pixels processed
+
+                  ja         mainloop8
+mainloop8end:
+
+                  mov        ecx,diff
+                  cmp        ecx,0
+                  jz         end8
+
+                  mov        edx,mask
+                  sal        edx,24        //make low byte the high byte
+
+secondloop8:
+                  sal        edx,1         //move high bit to CF
+                  jnc        skip8         //if CF = 0
+                  mov        al,[esi]
+                  mov        [ebx],al
+skip8:
+                  inc        esi
+                  inc        ebx
+
+                  dec        ecx
+                  jnz        secondloop8
+end8:
+                  emms
+               }
+            }
+            else /* mmx not supported - use modified C routine */
+            {
+               register unsigned int incr1, initial_val, final_val;
+               png_size_t pixel_bytes;
+               png_uint_32 i;
+               register int disp = png_pass_inc[png_ptr->pass];
+               int offset_table[7] = {0, 4, 0, 2, 0, 1, 0};
+
+               pixel_bytes = (png_ptr->row_info.pixel_depth >> 3);
+               srcptr = png_ptr->row_buf + 1 + offset_table[png_ptr->pass]*
+                  pixel_bytes;
+               dstptr = row + offset_table[png_ptr->pass]*pixel_bytes;
+               initial_val = offset_table[png_ptr->pass]*pixel_bytes;
+               final_val = png_ptr->width*pixel_bytes;
+               incr1 = (disp)*pixel_bytes;
+               for (i = initial_val; i < final_val; i += incr1)
+               {
+                  png_memcpy(dstptr, srcptr, pixel_bytes);
+                  srcptr += incr1;
+                  dstptr += incr1;
+               }
+            } /* end of else */
+
+            break;
+         }       // end 8 bpp
+
+         case 1:
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int s_inc, s_start, s_end;
+            int m;
+            int shift;
+            png_uint_32 i;
+
+            sp = png_ptr->row_buf + 1;
+            dp = row;
+            m = 0x80;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+                s_start = 0;
+                s_end = 7;
+                s_inc = 1;
+            }
+            else
+#endif
+            {
+                s_start = 7;
+                s_end = 0;
+                s_inc = -1;
+            }
+
+            shift = s_start;
+
+            for (i = 0; i < png_ptr->width; i++)
+            {
+               if (m & mask)
+               {
+                  int value;
+
+                  value = (*sp >> shift) & 0x1;
+                  *dp &= (png_byte)((0x7f7f >> (7 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }
+
+         case 2:
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int s_start, s_end, s_inc;
+            int m;
+            int shift;
+            png_uint_32 i;
+            int value;
+
+            sp = png_ptr->row_buf + 1;
+            dp = row;
+            m = 0x80;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+               s_start = 0;
+               s_end = 6;
+               s_inc = 2;
+            }
+            else
+#endif
+            {
+               s_start = 6;
+               s_end = 0;
+               s_inc = -2;
+            }
+
+            shift = s_start;
+
+            for (i = 0; i < png_ptr->width; i++)
+            {
+               if (m & mask)
+               {
+                  value = (*sp >> shift) & 0x3;
+                  *dp &= (png_byte)((0x3f3f >> (6 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }
+
+         case 4:
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int s_start, s_end, s_inc;
+            int m;
+            int shift;
+            png_uint_32 i;
+            int value;
+
+            sp = png_ptr->row_buf + 1;
+            dp = row;
+            m = 0x80;
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (png_ptr->transformations & PNG_PACKSWAP)
+            {
+               s_start = 0;
+               s_end = 4;
+               s_inc = 4;
+            }
+            else
+#endif
+            {
+               s_start = 4;
+               s_end = 0;
+               s_inc = -4;
+            }
+            shift = s_start;
+
+            for (i = 0; i < png_ptr->width; i++)
+            {
+               if (m & mask)
+               {
+                  value = (*sp >> shift) & 0xf;
+                  *dp &= (png_byte)((0xf0f >> (4 - shift)) & 0xff);
+                  *dp |= (png_byte)(value << shift);
+               }
+
+               if (shift == s_end)
+               {
+                  shift = s_start;
+                  sp++;
+                  dp++;
+               }
+               else
+                  shift += s_inc;
+               if (m == 1)
+                  m = 0x80;
+               else
+                  m >>= 1;
+            }
+            break;
+         }
+
+         case 16:
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+            png_uint_32 len;
+            int unmask, diff;
+            __int64 mask1=0x0101020204040808,
+                    mask0=0x1010202040408080;
+
+#if !defined(PNG_1_0_X)
+            if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+                /* && mmx_supported */ )
+#else
+            if (mmx_supported)
+#endif
+            {
+               srcptr = png_ptr->row_buf + 1;
+               dstptr = row;
+
+               unmask = ~mask;
+               len     = (png_ptr->width)&~7;
+               diff = (png_ptr->width)&7;
+               _asm
+               {
+                  movd       mm7, unmask       //load bit pattern
+                  psubb      mm6,mm6           //zero mm6
+                  punpcklbw  mm7,mm7
+                  punpcklwd  mm7,mm7
+                  punpckldq  mm7,mm7           //fill register with 8 masks
+
+                  movq       mm0,mask0
+                  movq       mm1,mask1
+
+                  pand       mm0,mm7
+                  pand       mm1,mm7
+
+                  pcmpeqb    mm0,mm6
+                  pcmpeqb    mm1,mm6
+
+                  mov        ecx,len           //load length of line
+                  mov        esi,srcptr        //load source
+                  mov        ebx,dstptr        //load dest
+                  cmp        ecx,0             //lcr
+                  jz         mainloop16end
+
+mainloop16:
+                  movq       mm4,[esi]
+                  pand       mm4,mm0
+                  movq       mm6,mm0
+                  movq       mm7,[ebx]
+                  pandn      mm6,mm7
+                  por        mm4,mm6
+                  movq       [ebx],mm4
+
+                  movq       mm5,[esi+8]
+                  pand       mm5,mm1
+                  movq       mm7,mm1
+                  movq       mm6,[ebx+8]
+                  pandn      mm7,mm6
+                  por        mm5,mm7
+                  movq       [ebx+8],mm5
+
+                  add        esi,16            //inc by 16 bytes processed
+                  add        ebx,16
+                  sub        ecx,8             //dec by 8 pixels processed
+
+                  ja         mainloop16
+
+mainloop16end:
+                  mov        ecx,diff
+                  cmp        ecx,0
+                  jz         end16
+
+                  mov        edx,mask
+                  sal        edx,24            //make low byte the high byte
+secondloop16:
+                  sal        edx,1             //move high bit to CF
+                  jnc        skip16            //if CF = 0
+                  mov        ax,[esi]
+                  mov        [ebx],ax
+skip16:
+                  add        esi,2
+                  add        ebx,2
+
+                  dec        ecx
+                  jnz        secondloop16
+end16:
+                  emms
+               }
+            }
+            else /* mmx not supported - use modified C routine */
+            {
+               register unsigned int incr1, initial_val, final_val;
+               png_size_t pixel_bytes;
+               png_uint_32 i;
+               register int disp = png_pass_inc[png_ptr->pass];
+               int offset_table[7] = {0, 4, 0, 2, 0, 1, 0};
+
+               pixel_bytes = (png_ptr->row_info.pixel_depth >> 3);
+               srcptr = png_ptr->row_buf + 1 + offset_table[png_ptr->pass]*
+                  pixel_bytes;
+               dstptr = row + offset_table[png_ptr->pass]*pixel_bytes;
+               initial_val = offset_table[png_ptr->pass]*pixel_bytes;
+               final_val = png_ptr->width*pixel_bytes;
+               incr1 = (disp)*pixel_bytes;
+               for (i = initial_val; i < final_val; i += incr1)
+               {
+                  png_memcpy(dstptr, srcptr, pixel_bytes);
+                  srcptr += incr1;
+                  dstptr += incr1;
+               }
+            } /* end of else */
+
+            break;
+         }       // end 16 bpp
+
+         case 48:
+         {
+            png_bytep srcptr;
+            png_bytep dstptr;
+            png_uint_32 len;
+            int unmask, diff;
+
+            __int64 mask5=0x0101010101010202,
+                    mask4=0x0202020204040404,
+                    mask3=0x0404080808080808,
+                    mask2=0x1010101010102020,
+                    mask1=0x2020202040404040,
+                    mask0=0x4040808080808080;
+
+#if !defined(PNG_1_0_X)
+            if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_COMBINE_ROW)
+                /* && mmx_supported */ )
+#else
+            if (mmx_supported)
+#endif
+            {
+               srcptr = png_ptr->row_buf + 1;
+               dstptr = row;
+
+               unmask = ~mask;
+               len     = (png_ptr->width)&~7;
+               diff = (png_ptr->width)&7;
+               _asm
+               {
+                  movd       mm7, unmask       //load bit pattern
+                  psubb      mm6,mm6           //zero mm6
+                  punpcklbw  mm7,mm7
+                  punpcklwd  mm7,mm7
+                  punpckldq  mm7,mm7           //fill register with 8 masks
+
+                  movq       mm0,mask0
+                  movq       mm1,mask1
+                  movq       mm2,mask2
+                  movq       mm3,mask3
+                  movq       mm4,mask4
+                  movq       mm5,mask5
+
+                  pand       mm0,mm7
+                  pand       mm1,mm7
+                  pand       mm2,mm7
+                  pand       mm3,mm7
+                  pand       mm4,mm7
+                  pand       mm5,mm7
+
+                  pcmpeqb    mm0,mm6
+                  pcmpeqb    mm1,mm6
+                  pcmpeqb    mm2,mm6
+                  pcmpeqb    mm3,mm6
+                  pcmpeqb    mm4,mm6
+                  pcmpeqb    mm5,mm6
+
+                  mov        ecx,len           //load length of line
+                  mov        esi,srcptr        //load source
+                  mov        ebx,dstptr        //load dest
+
+                  cmp        ecx,0
+                  jz         mainloop48end
+
+mainloop48:
+                  movq       mm7,[esi]
+                  pand       mm7,mm0
+                  movq       mm6,mm0
+                  pandn      mm6,[ebx]
+                  por        mm7,mm6
+                  movq       [ebx],mm7
+
+                  movq       mm6,[esi+8]
+                  pand       mm6,mm1
+                  movq       mm7,mm1
+                  pandn      mm7,[ebx+8]
+                  por        mm6,mm7
+                  movq       [ebx+8],mm6
+
+                  movq       mm6,[esi+16]
+                  pand       mm6,mm2
+                  movq       mm7,mm2
+                  pandn      mm7,[ebx+16]
+                  por        mm6,mm7
+                  movq       [ebx+16],mm6
+
+                  movq       mm7,[esi+24]
+                  pand       mm7,mm3
+                  movq       mm6,mm3
+                  pandn      mm6,[ebx+24]
+                  por        mm7,mm6
+                  movq       [ebx+24],mm7
+
+                  movq       mm6,[esi+32]
+                  pand       mm6,mm4
+                  movq       mm7,mm4
+                  pandn      mm7,[ebx+32]
+                  por        mm6,mm7
+                  movq       [ebx+32],mm6
+
+                  movq       mm7,[esi+40]
+                  pand       mm7,mm5
+                  movq       mm6,mm5
+                  pandn      mm6,[ebx+40]
+                  por        mm7,mm6
+                  movq       [ebx+40],mm7
+
+                  add        esi,48            //inc by 32 bytes processed
+                  add        ebx,48
+                  sub        ecx,8             //dec by 8 pixels processed
+
+                  ja         mainloop48
+mainloop48end:
+
+                  mov        ecx,diff
+                  cmp        ecx,0
+                  jz         end48
+
+                  mov        edx,mask
+                  sal        edx,24            //make low byte the high byte
+
+secondloop48:
+                  sal        edx,1             //move high bit to CF
+                  jnc        skip48            //if CF = 0
+                  mov        eax,[esi]
+                  mov        [ebx],eax
+                  mov        ax,[esi+4]       // These 2 lines added 20070717
+                  mov        [ebx+4],ax       // Glenn R-P
+skip48:
+                  add        esi,6            // Changed 4 to 6 on these 2
+                  add        ebx,6            // lines.  Glenn R-P 20070717
+
+                  dec        ecx
+                  jnz        secondloop48
+
+end48:
+                  emms
+               }
+            }
+            else /* mmx _not supported - Use modified C routine */
+            {
+               register unsigned int incr1, initial_val, final_val;
+               png_size_t pixel_bytes;
+               png_uint_32 i;
+               register int disp = png_pass_inc[png_ptr->pass];
+               int offset_table[7] = {0, 4, 0, 2, 0, 1, 0};
+
+               pixel_bytes = (png_ptr->row_info.pixel_depth >> 3);
+               srcptr = png_ptr->row_buf + 1 + offset_table[png_ptr->pass]*
+                  pixel_bytes;
+               dstptr = row + offset_table[png_ptr->pass]*pixel_bytes;
+               initial_val = offset_table[png_ptr->pass]*pixel_bytes;
+               final_val = png_ptr->width*pixel_bytes;
+               incr1 = (disp)*pixel_bytes;
+               for (i = initial_val; i < final_val; i += incr1)
+               {
+                  png_memcpy(dstptr, srcptr, pixel_bytes);
+                  srcptr += incr1;
+                  dstptr += incr1;
+               }
+            } /* end of else */
+
+            break;
+         }       // end 48 bpp
+
+         default:
+         {
+            png_bytep sptr;
+            png_bytep dp;
+            png_size_t pixel_bytes;
+            int offset_table[7] = {0, 4, 0, 2, 0, 1, 0};
+            unsigned int i;
+            register int disp = png_pass_inc[png_ptr->pass];  // get the offset
+            register unsigned int incr1, initial_val, final_val;
+
+            pixel_bytes = (png_ptr->row_info.pixel_depth >> 3);
+            sptr = png_ptr->row_buf + 1 + offset_table[png_ptr->pass]*
+               pixel_bytes;
+            dp = row + offset_table[png_ptr->pass]*pixel_bytes;
+            initial_val = offset_table[png_ptr->pass]*pixel_bytes;
+            final_val = png_ptr->width*pixel_bytes;
+            incr1 = (disp)*pixel_bytes;
+            for (i = initial_val; i < final_val; i += incr1)
+            {
+               png_memcpy(dp, sptr, pixel_bytes);
+               sptr += incr1;
+               dp += incr1;
+            }
+            break;
+         }
+      } /* end switch (png_ptr->row_info.pixel_depth) */
+   } /* end if (non-trivial mask) */
+
+} /* end png_combine_row() */
+
+
+#if defined(PNG_READ_INTERLACING_SUPPORTED)
+
+void /* PRIVATE */
+png_do_read_interlace(png_structp png_ptr)
+{
+   png_row_infop row_info = &(png_ptr->row_info);
+   png_bytep row = png_ptr->row_buf + 1;
+   int pass = png_ptr->pass;
+   png_uint_32 transformations = png_ptr->transformations;
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_CONST int png_pass_inc[7] = {8, 8, 4, 4, 2, 2, 1};
+#endif
+
+   png_debug(1,"in png_do_read_interlace\n");
+
+   if (mmx_supported == 2) {
+#if !defined(PNG_1_0_X)
+       /* this should have happened in png_init_mmx_flags() already */
+       png_warning(png_ptr, "asm_flags may not have been initialized");
+#endif
+       png_mmx_support();
+   }
+
+   if (row != NULL && row_info != NULL)
+   {
+      png_uint_32 final_width;
+
+      final_width = row_info->width * png_pass_inc[pass];
+
+      switch (row_info->pixel_depth)
+      {
+         case 1:
+         {
+            png_bytep sp, dp;
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_byte v;
+            png_uint_32 i;
+            int j;
+
+            sp = row + (png_size_t)((row_info->width - 1) >> 3);
+            dp = row + (png_size_t)((final_width - 1) >> 3);
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (int)((row_info->width + 7) & 7);
+               dshift = (int)((final_width + 7) & 7);
+               s_start = 7;
+               s_end = 0;
+               s_inc = -1;
+            }
+            else
+#endif
+            {
+               sshift = 7 - (int)((row_info->width + 7) & 7);
+               dshift = 7 - (int)((final_width + 7) & 7);
+               s_start = 0;
+               s_end = 7;
+               s_inc = 1;
+            }
+
+            for (i = row_info->width; i; i--)
+            {
+               v = (png_byte)((*sp >> sshift) & 0x1);
+               for (j = 0; j < png_pass_inc[pass]; j++)
+               {
+                  *dp &= (png_byte)((0x7f7f >> (7 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+
+         case 2:
+         {
+            png_bytep sp, dp;
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_uint_32 i;
+
+            sp = row + (png_size_t)((row_info->width - 1) >> 2);
+            dp = row + (png_size_t)((final_width - 1) >> 2);
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (png_size_t)(((row_info->width + 3) & 3) << 1);
+               dshift = (png_size_t)(((final_width + 3) & 3) << 1);
+               s_start = 6;
+               s_end = 0;
+               s_inc = -2;
+            }
+            else
+#endif
+            {
+               sshift = (png_size_t)((3 - ((row_info->width + 3) & 3)) << 1);
+               dshift = (png_size_t)((3 - ((final_width + 3) & 3)) << 1);
+               s_start = 0;
+               s_end = 6;
+               s_inc = 2;
+            }
+
+            for (i = row_info->width; i; i--)
+            {
+               png_byte v;
+               int j;
+
+               v = (png_byte)((*sp >> sshift) & 0x3);
+               for (j = 0; j < png_pass_inc[pass]; j++)
+               {
+                  *dp &= (png_byte)((0x3f3f >> (6 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+
+         case 4:
+         {
+            png_bytep sp, dp;
+            int sshift, dshift;
+            int s_start, s_end, s_inc;
+            png_uint_32 i;
+
+            sp = row + (png_size_t)((row_info->width - 1) >> 1);
+            dp = row + (png_size_t)((final_width - 1) >> 1);
+#if defined(PNG_READ_PACKSWAP_SUPPORTED)
+            if (transformations & PNG_PACKSWAP)
+            {
+               sshift = (png_size_t)(((row_info->width + 1) & 1) << 2);
+               dshift = (png_size_t)(((final_width + 1) & 1) << 2);
+               s_start = 4;
+               s_end = 0;
+               s_inc = -4;
+            }
+            else
+#endif
+            {
+               sshift = (png_size_t)((1 - ((row_info->width + 1) & 1)) << 2);
+               dshift = (png_size_t)((1 - ((final_width + 1) & 1)) << 2);
+               s_start = 0;
+               s_end = 4;
+               s_inc = 4;
+            }
+
+            for (i = row_info->width; i; i--)
+            {
+               png_byte v;
+               int j;
+
+               v = (png_byte)((*sp >> sshift) & 0xf);
+               for (j = 0; j < png_pass_inc[pass]; j++)
+               {
+                  *dp &= (png_byte)((0xf0f >> (4 - dshift)) & 0xff);
+                  *dp |= (png_byte)(v << dshift);
+                  if (dshift == s_end)
+                  {
+                     dshift = s_start;
+                     dp--;
+                  }
+                  else
+                     dshift += s_inc;
+               }
+               if (sshift == s_end)
+               {
+                  sshift = s_start;
+                  sp--;
+               }
+               else
+                  sshift += s_inc;
+            }
+            break;
+         }
+
+         default:         // This is the place where the routine is modified
+         {
+            __int64 const4 = 0x0000000000FFFFFF;
+            // __int64 const5 = 0x000000FFFFFF0000;  // unused...
+            __int64 const6 = 0x00000000000000FF;
+            png_bytep sptr, dp;
+            png_uint_32 i;
+            png_size_t pixel_bytes;
+            int width = row_info->width;
+
+            pixel_bytes = (row_info->pixel_depth >> 3);
+
+            sptr = row + (width - 1) * pixel_bytes;
+            dp = row + (final_width - 1) * pixel_bytes;
+            // New code by Nirav Chhatrapati - Intel Corporation
+            // sign fix by GRR
+            // NOTE:  there is NO MMX code for 48-bit and 64-bit images
+
+            // use MMX routine if machine supports it
+#if !defined(PNG_1_0_X)
+            if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_INTERLACE)
+                /* && mmx_supported */ )
+#else
+            if (mmx_supported)
+#endif
+            {
+               if (pixel_bytes == 3)
+               {
+                  if (((pass == 4) || (pass == 5)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1) - 8;
+                     if (width_mmx < 0)
+                         width_mmx = 0;
+                     width -= width_mmx;        // 8 or 9 pix, 24 or 27 bytes
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub esi, 3
+                           sub edi, 9
+loop_pass4:
+                           movq mm0, [esi]     ; X X v2 v1 v0 v5 v4 v3
+                           movq mm7, mm0       ; X X v2 v1 v0 v5 v4 v3
+                           movq mm6, mm0       ; X X v2 v1 v0 v5 v4 v3
+                           psllq mm0, 24       ; v1 v0 v5 v4 v3 0 0 0
+                           pand mm7, const4    ; 0 0 0 0 0 v5 v4 v3
+                           psrlq mm6, 24       ; 0 0 0 X X v2 v1 v0
+                           por mm0, mm7        ; v1 v0 v5 v4 v3 v5 v4 v3
+                           movq mm5, mm6       ; 0 0 0 X X v2 v1 v0
+                           psllq mm6, 8        ; 0 0 X X v2 v1 v0 0
+                           movq [edi], mm0     ; move quad to memory
+                           psrlq mm5, 16       ; 0 0 0 0 0 X X v2
+                           pand mm5, const6    ; 0 0 0 0 0 0 0 v2
+                           por mm6, mm5        ; 0 0 X X v2 v1 v0 v2
+                           movd [edi+8], mm6   ; move double to memory
+                           sub esi, 6
+                           sub edi, 12
+                           sub ecx, 2
+                           jnz loop_pass4
+                           EMMS
+                        }
+                     }
+
+                     sptr -= width_mmx*3;
+                     dp -= width_mmx*6;
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+
+                        png_memcpy(v, sptr, 3);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           png_memcpy(dp, v, 3);
+                           dp -= 3;
+                        }
+                        sptr -= 3;
+                     }
+                  }
+                  else if (((pass == 2) || (pass == 3)) && width)
+                  {
+                     _asm
+                     {
+                        mov esi, sptr
+                        mov edi, dp
+                        mov ecx, width
+                        sub edi, 9   // (png_pass_inc[pass] - 1)*pixel_bytes
+loop_pass2:
+                        movd mm0, [esi]     ; X X X X X v2 v1 v0
+                        pand mm0, const4    ; 0 0 0 0 0 v2 v1 v0
+                        movq mm1, mm0       ; 0 0 0 0 0 v2 v1 v0
+                        psllq mm0, 16       ; 0 0 0 v2 v1 v0 0 0
+                        movq mm2, mm0       ; 0 0 0 v2 v1 v0 0 0
+                        psllq mm0, 24       ; v2 v1 v0 0 0 0 0 0
+                        psrlq mm1, 8        ; 0 0 0 0 0 0 v2 v1
+                        por mm0, mm2        ; v2 v1 v0 v2 v1 v0 0 0
+                        por mm0, mm1        ; v2 v1 v0 v2 v1 v0 v2 v1
+                        movq [edi+4], mm0   ; move to memory
+                        psrlq mm0, 16       ; 0 0 v2 v1 v0 v2 v1 v0
+                        movd [edi], mm0     ; move to memory
+                        sub esi, 3
+                        sub edi, 12
+                        dec ecx
+                        jnz loop_pass2
+                        EMMS
+                     }
+                  }
+                  else if (width) /* && ((pass == 0) || (pass == 1))) */
+                  {
+                     _asm
+                     {
+                        mov esi, sptr
+                        mov edi, dp
+                        mov ecx, width
+                        sub edi, 21   // (png_pass_inc[pass] - 1)*pixel_bytes
+loop_pass0:
+                        movd mm0, [esi]     ; X X X X X v2 v1 v0
+                        pand mm0, const4    ; 0 0 0 0 0 v2 v1 v0
+                        movq mm1, mm0       ; 0 0 0 0 0 v2 v1 v0
+                        psllq mm0, 16       ; 0 0 0 v2 v1 v0 0 0
+                        movq mm2, mm0       ; 0 0 0 v2 v1 v0 0 0
+                        psllq mm0, 24       ; v2 v1 v0 0 0 0 0 0
+                        psrlq mm1, 8        ; 0 0 0 0 0 0 v2 v1
+                        por mm0, mm2        ; v2 v1 v0 v2 v1 v0 0 0
+                        por mm0, mm1        ; v2 v1 v0 v2 v1 v0 v2 v1
+                        movq mm3, mm0       ; v2 v1 v0 v2 v1 v0 v2 v1
+                        psllq mm0, 16       ; v0 v2 v1 v0 v2 v1 0 0
+                        movq mm4, mm3       ; v2 v1 v0 v2 v1 v0 v2 v1
+                        punpckhdq mm3, mm0  ; v0 v2 v1 v0 v2 v1 v0 v2
+                        movq [edi+16] , mm4
+                        psrlq mm0, 32       ; 0 0 0 0 v0 v2 v1 v0
+                        movq [edi+8] , mm3
+                        punpckldq mm0, mm4  ; v1 v0 v2 v1 v0 v2 v1 v0
+                        sub esi, 3
+                        movq [edi], mm0
+                        sub edi, 24
+                        //sub esi, 3
+                        dec ecx
+                        jnz loop_pass0
+                        EMMS
+                     }
+                  }
+               } /* end of pixel_bytes == 3 */
+
+               else if (pixel_bytes == 1)
+               {
+                  if (((pass == 4) || (pass == 5)) && width)
+                  {
+                     int width_mmx = ((width >> 3) << 3);
+                     width -= width_mmx;
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub edi, 15
+                           sub esi, 7
+loop1_pass4:
+                           movq mm0, [esi]     ; v0 v1 v2 v3 v4 v5 v6 v7
+                           movq mm1, mm0       ; v0 v1 v2 v3 v4 v5 v6 v7
+                           punpcklbw mm0, mm0  ; v4 v4 v5 v5 v6 v6 v7 v7
+                           //movq mm1, mm0     ; v0 v0 v1 v1 v2 v2 v3 v3
+                           punpckhbw mm1, mm1  ;v0 v0 v1 v1 v2 v2 v3 v3
+                           movq [edi+8], mm1   ; move to memory v0 v1 v2 and v3
+                           sub esi, 8
+                           movq [edi], mm0     ; move to memory v4 v5 v6 and v7
+                           //sub esi, 4
+                           sub edi, 16
+                           sub ecx, 8
+                           jnz loop1_pass4
+                           EMMS
+                        }
+                     }
+
+                     sptr -= width_mmx;
+                     dp -= width_mmx*2;
+                     for (i = width; i; i--)
+                     {
+                        int j;
+
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           *dp-- = *sptr;
+                        }
+                        sptr --;
+                     }
+                  }
+                  else if (((pass == 2) || (pass == 3)) && width)
+                  {
+                     int width_mmx = ((width >> 2) << 2);
+                     width -= width_mmx;
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub edi, 15
+                           sub esi, 3
+loop1_pass2:
+                           movd mm0, [esi]     ; X X X X v0 v1 v2 v3
+                           punpcklbw mm0, mm0  ; v0 v0 v1 v1 v2 v2 v3 v3
+                           movq mm1, mm0       ; v0 v0 v1 v1 v2 v2 v3 v3
+                           punpcklwd mm0, mm0  ; v2 v2 v2 v2 v3 v3 v3 v3
+                           punpckhwd mm1, mm1  ; v0 v0 v0 v0 v1 v1 v1 v1
+                           movq [edi], mm0     ; move to memory v2 and v3
+                           sub esi, 4
+                           movq [edi+8], mm1   ; move to memory v1     and v0
+                           sub edi, 16
+                           sub ecx, 4
+                           jnz loop1_pass2
+                           EMMS
+                        }
+                     }
+
+                     sptr -= width_mmx;
+                     dp -= width_mmx*4;
+                     for (i = width; i; i--)
+                     {
+                        int j;
+
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           *dp-- = *sptr;
+                        }
+                        sptr --;
+                     }
+                  }
+                  else if (width) /* && ((pass == 0) || (pass == 1))) */
+                  {
+                     int width_mmx = ((width >> 2) << 2);
+                     width -= width_mmx;
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub edi, 31
+                           sub esi, 3
+loop1_pass0:
+                           movd mm0, [esi]     ; X X X X v0 v1 v2 v3
+                           movq mm1, mm0       ; X X X X v0 v1 v2 v3
+                           punpcklbw mm0, mm0  ; v0 v0 v1 v1 v2 v2 v3 v3
+                           movq mm2, mm0       ; v0 v0 v1 v1 v2 v2 v3 v3
+                           punpcklwd mm0, mm0  ; v2 v2 v2 v2 v3 v3 v3 v3
+                           movq mm3, mm0       ; v2 v2 v2 v2 v3 v3 v3 v3
+                           punpckldq mm0, mm0  ; v3 v3 v3 v3 v3 v3 v3 v3
+                           punpckhdq mm3, mm3  ; v2 v2 v2 v2 v2 v2 v2 v2
+                           movq [edi], mm0     ; move to memory v3
+                           punpckhwd mm2, mm2  ; v0 v0 v0 v0 v1 v1 v1 v1
+                           movq [edi+8], mm3   ; move to memory v2
+                           movq mm4, mm2       ; v0 v0 v0 v0 v1 v1 v1 v1
+                           punpckldq mm2, mm2  ; v1 v1 v1 v1 v1 v1 v1 v1
+                           punpckhdq mm4, mm4  ; v0 v0 v0 v0 v0 v0 v0 v0
+                           movq [edi+16], mm2  ; move to memory v1
+                           movq [edi+24], mm4  ; move to memory v0
+                           sub esi, 4
+                           sub edi, 32
+                           sub ecx, 4
+                           jnz loop1_pass0
+                           EMMS
+                        }
+                     }
+
+                     sptr -= width_mmx;
+                     dp -= width_mmx*8;
+                     for (i = width; i; i--)
+                     {
+                        int j;
+
+                       /* I simplified this part in version 1.0.4e
+                        * here and in several other instances where
+                        * pixel_bytes == 1  -- GR-P
+                        *
+                        * Original code:
+                        *
+                        * png_byte v[8];
+                        * png_memcpy(v, sptr, pixel_bytes);
+                        * for (j = 0; j < png_pass_inc[pass]; j++)
+                        * {
+                        *    png_memcpy(dp, v, pixel_bytes);
+                        *    dp -= pixel_bytes;
+                        * }
+                        * sptr -= pixel_bytes;
+                        *
+                        * Replacement code is in the next three lines:
+                        */
+
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                           *dp-- = *sptr;
+                        sptr--;
+                     }
+                  }
+               } /* end of pixel_bytes == 1 */
+
+               else if (pixel_bytes == 2)
+               {
+                  if (((pass == 4) || (pass == 5)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1) ;
+                     width -= width_mmx;
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub esi, 2
+                           sub edi, 6
+loop2_pass4:
+                           movd mm0, [esi]        ; X X X X v1 v0 v3 v2
+                           punpcklwd mm0, mm0     ; v1 v0 v1 v0 v3 v2 v3 v2
+                           sub esi, 4
+                           movq [edi], mm0
+                           sub edi, 8
+                           sub ecx, 2
+                           jnz loop2_pass4
+                           EMMS
+                        }
+                     }
+
+                     sptr -= (width_mmx*2 - 2);            // sign fixed
+                     dp -= (width_mmx*4 - 2);            // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 2;
+                        png_memcpy(v, sptr, 2);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 2;
+                           png_memcpy(dp, v, 2);
+                        }
+                     }
+                  }
+                  else if (((pass == 2) || (pass == 3)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1) ;
+                     width -= width_mmx;
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub esi, 2
+                           sub edi, 14
+loop2_pass2:
+                           movd mm0, [esi]        ; X X X X v1 v0 v3 v2
+                           punpcklwd mm0, mm0     ; v1 v0 v1 v0 v3 v2 v3 v2
+                           movq mm1, mm0          ; v1 v0 v1 v0 v3 v2 v3 v2
+                           punpckldq mm0, mm0     ; v3 v2 v3 v2 v3 v2 v3 v2
+                           punpckhdq mm1, mm1     ; v1 v0 v1 v0 v1 v0 v1 v0
+                           movq [edi], mm0
+                           sub esi, 4
+                           movq [edi + 8], mm1
+                           //sub esi, 4
+                           sub edi, 16
+                           sub ecx, 2
+                           jnz loop2_pass2
+                           EMMS
+                        }
+                     }
+
+                     sptr -= (width_mmx*2 - 2);            // sign fixed
+                     dp -= (width_mmx*8 - 2);            // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 2;
+                        png_memcpy(v, sptr, 2);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 2;
+                           png_memcpy(dp, v, 2);
+                        }
+                     }
+                  }
+                  else if (width) /* && ((pass == 0) || (pass == 1))) */
+                  {
+                     int width_mmx = ((width >> 1) << 1);
+                     width -= width_mmx;
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub esi, 2
+                           sub edi, 30
+loop2_pass0:
+                           movd mm0, [esi]        ; X X X X v1 v0 v3 v2
+                           punpcklwd mm0, mm0     ; v1 v0 v1 v0 v3 v2 v3 v2
+                           movq mm1, mm0          ; v1 v0 v1 v0 v3 v2 v3 v2
+                           punpckldq mm0, mm0     ; v3 v2 v3 v2 v3 v2 v3 v2
+                           punpckhdq mm1, mm1     ; v1 v0 v1 v0 v1 v0 v1 v0
+                           movq [edi], mm0
+                           movq [edi + 8], mm0
+                           movq [edi + 16], mm1
+                           movq [edi + 24], mm1
+                           sub esi, 4
+                           sub edi, 32
+                           sub ecx, 2
+                           jnz loop2_pass0
+                           EMMS
+                        }
+                     }
+
+                     sptr -= (width_mmx*2 - 2);            // sign fixed
+                     dp -= (width_mmx*16 - 2);            // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 2;
+                        png_memcpy(v, sptr, 2);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 2;
+                           png_memcpy(dp, v, 2);
+                        }
+                     }
+                  }
+               } /* end of pixel_bytes == 2 */
+
+               else if (pixel_bytes == 4)
+               {
+                  if (((pass == 4) || (pass == 5)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1) ;
+                     width -= width_mmx;
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub esi, 4
+                           sub edi, 12
+loop4_pass4:
+                           movq mm0, [esi]      ; v3 v2 v1 v0 v7 v6 v5 v4
+                           movq mm1, mm0        ; v3 v2 v1 v0 v7 v6 v5 v4
+                           punpckldq mm0, mm0   ; v7 v6 v5 v4 v7 v6 v5 v4
+                           punpckhdq mm1, mm1   ; v3 v2 v1 v0 v3 v2 v1 v0
+                           movq [edi], mm0
+                           sub esi, 8
+                           movq [edi + 8], mm1
+                           sub edi, 16
+                           sub ecx, 2
+                           jnz loop4_pass4
+                           EMMS
+                        }
+                     }
+
+                     sptr -= (width_mmx*4 - 4);          // sign fixed
+                     dp -= (width_mmx*8 - 4);            // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 4;
+                        png_memcpy(v, sptr, 4);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 4;
+                           png_memcpy(dp, v, 4);
+                        }
+                     }
+                  }
+                  else if (((pass == 2) || (pass == 3)) && width)
+                  {
+                     int width_mmx = ((width >> 1) << 1) ;
+                     width -= width_mmx;
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub esi, 4
+                           sub edi, 28
+loop4_pass2:
+                           movq mm0, [esi]      ; v3 v2 v1 v0 v7 v6 v5 v4
+                           movq mm1, mm0        ; v3 v2 v1 v0 v7 v6 v5 v4
+                           punpckldq mm0, mm0   ; v7 v6 v5 v4 v7 v6 v5 v4
+                           punpckhdq mm1, mm1   ; v3 v2 v1 v0 v3 v2 v1 v0
+                           movq [edi], mm0
+                           movq [edi + 8], mm0
+                           movq [edi+16], mm1
+                           movq [edi + 24], mm1
+                           sub esi, 8
+                           sub edi, 32
+                           sub ecx, 2
+                           jnz loop4_pass2
+                           EMMS
+                        }
+                     }
+
+                     sptr -= (width_mmx*4 - 4);            // sign fixed
+                     dp -= (width_mmx*16 - 4);            // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 4;
+                        png_memcpy(v, sptr, 4);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 4;
+                           png_memcpy(dp, v, 4);
+                        }
+                     }
+                  }
+                  else if (width) /* && ((pass == 0) || (pass == 1))) */
+                  {
+                     int width_mmx = ((width >> 1) << 1) ;
+                     width -= width_mmx;
+                     if (width_mmx)
+                     {
+                        _asm
+                        {
+                           mov esi, sptr
+                           mov edi, dp
+                           mov ecx, width_mmx
+                           sub esi, 4
+                           sub edi, 60
+loop4_pass0:
+                           movq mm0, [esi]        ; v3 v2 v1 v0 v7 v6 v5 v4
+                           movq mm1, mm0          ; v3 v2 v1 v0 v7 v6 v5 v4
+                           punpckldq mm0, mm0     ; v7 v6 v5 v4 v7 v6 v5 v4
+                           punpckhdq mm1, mm1     ; v3 v2 v1 v0 v3 v2 v1 v0
+                           movq [edi], mm0
+                           movq [edi + 8], mm0
+                           movq [edi + 16], mm0
+                           movq [edi + 24], mm0
+                           movq [edi+32], mm1
+                           movq [edi + 40], mm1
+                           movq [edi+ 48], mm1
+                           sub esi, 8
+                           movq [edi + 56], mm1
+                           sub edi, 64
+                           sub ecx, 2
+                           jnz loop4_pass0
+                           EMMS
+                        }
+                     }
+
+                     sptr -= (width_mmx*4 - 4);            // sign fixed
+                     dp -= (width_mmx*32 - 4);            // sign fixed
+                     for (i = width; i; i--)
+                     {
+                        png_byte v[8];
+                        int j;
+                        sptr -= 4;
+                        png_memcpy(v, sptr, 4);
+                        for (j = 0; j < png_pass_inc[pass]; j++)
+                        {
+                           dp -= 4;
+                           png_memcpy(dp, v, 4);
+                        }
+                     }
+                  }
+
+               } /* end of pixel_bytes == 4 */
+
+               else if (pixel_bytes == 6)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, 6);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, 6);
+                        dp -= 6;
+                     }
+                     sptr -= 6;
+                  }
+               } /* end of pixel_bytes == 6 */
+
+               else
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, pixel_bytes);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, pixel_bytes);
+                        dp -= pixel_bytes;
+                     }
+                     sptr-= pixel_bytes;
+                  }
+               }
+            } /* end of mmx_supported */
+
+            else /* MMX not supported:  use modified C code - takes advantage
+                  * of inlining of memcpy for a constant */
+            {
+               if (pixel_bytes == 1)
+               {
+                  for (i = width; i; i--)
+                  {
+                     int j;
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                        *dp-- = *sptr;
+                     sptr--;
+                  }
+               }
+               else if (pixel_bytes == 3)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, pixel_bytes);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, pixel_bytes);
+                        dp -= pixel_bytes;
+                     }
+                     sptr -= pixel_bytes;
+                  }
+               }
+               else if (pixel_bytes == 2)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, pixel_bytes);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, pixel_bytes);
+                        dp -= pixel_bytes;
+                     }
+                     sptr -= pixel_bytes;
+                  }
+               }
+               else if (pixel_bytes == 4)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, pixel_bytes);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, pixel_bytes);
+                        dp -= pixel_bytes;
+                     }
+                     sptr -= pixel_bytes;
+                  }
+               }
+               else if (pixel_bytes == 6)
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, pixel_bytes);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, pixel_bytes);
+                        dp -= pixel_bytes;
+                     }
+                     sptr -= pixel_bytes;
+                  }
+               }
+               else
+               {
+                  for (i = width; i; i--)
+                  {
+                     png_byte v[8];
+                     int j;
+                     png_memcpy(v, sptr, pixel_bytes);
+                     for (j = 0; j < png_pass_inc[pass]; j++)
+                     {
+                        png_memcpy(dp, v, pixel_bytes);
+                        dp -= pixel_bytes;
+                     }
+                     sptr -= pixel_bytes;
+                  }
+               }
+
+            } /* end of MMX not supported */
+            break;
+         }
+      } /* end switch (row_info->pixel_depth) */
+
+      row_info->width = final_width;
+
+      row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,final_width);
+   }
+
+}
+
+#endif /* PNG_READ_INTERLACING_SUPPORTED */
+
+
+// These global constants are declared
+// here to ensure alignment on 8-byte boundaries.
+  union uAll {
+     __int64 use;
+     double  double_align;
+     long long long_long_align;
+  } ;
+  static PNG_CONST union uAll LBCarryMask = {0x0101010101010101},
+                              HBClearMask = {0x7f7f7f7f7f7f7f7f};
+
+// Optimized code for PNG Average filter decoder
+void /* PRIVATE */
+png_read_filter_row_mmx_avg(png_row_infop row_info, png_bytep row
+                            , png_bytep prev_row)
+{
+  // These variables are declared
+  // here to ensure alignment on 8-byte boundaries.
+  union uAll ActiveMask, ShiftBpp, ShiftRem;
+
+   int bpp;
+   png_uint_32 FullLength;
+   png_uint_32 MMXLength;
+   //png_uint_32 len;
+   int diff;
+
+   bpp = (row_info->pixel_depth + 7) >> 3; // Get # bytes per pixel
+   FullLength  = row_info->rowbytes; // # of bytes to filter
+   _asm {
+         // Init address pointers and offset
+         mov edi, row          // edi ==> Avg(x)
+         xor ebx, ebx          // ebx ==> x
+         mov edx, edi
+         mov esi, prev_row           // esi ==> Prior(x)
+         sub edx, bpp          // edx ==> Raw(x-bpp)
+
+         xor eax, eax
+         // Compute the Raw value for the first bpp bytes
+         //    Raw(x) = Avg(x) + (Prior(x)/2)
+davgrlp:
+         mov al, [esi + ebx]   // Load al with Prior(x)
+         inc ebx
+         shr al, 1             // divide by 2
+         add al, [edi+ebx-1]   // Add Avg(x); -1 to offset inc ebx
+         cmp ebx, bpp
+         mov [edi+ebx-1], al    // Write back Raw(x);
+                            // mov does not affect flags; -1 to offset inc ebx
+         jb davgrlp
+         // get # of bytes to alignment
+         mov diff, edi         // take start of row
+         add diff, ebx         // add bpp
+         add diff, 0xf         // add 7 + 8 to incr past alignment boundary
+         and diff, 0xfffffff8  // mask to alignment boundary
+         sub diff, edi         // subtract from start ==> value ebx at alignment
+         jz davggo
+         // fix alignment
+         // Compute the Raw value for the bytes upto the alignment boundary
+         //    Raw(x) = Avg(x) + ((Raw(x-bpp) + Prior(x))/2)
+         xor ecx, ecx
+davglp1:
+         xor eax, eax
+         mov cl, [esi + ebx]        // load cl with Prior(x)
+         mov al, [edx + ebx]  // load al with Raw(x-bpp)
+         add ax, cx
+         inc ebx
+         shr ax, 1            // divide by 2
+         add al, [edi+ebx-1]  // Add Avg(x); -1 to offset inc ebx
+         cmp ebx, diff              // Check if at alignment boundary
+         mov [edi+ebx-1], al        // Write back Raw(x);
+                            // mov does not affect flags; -1 to offset inc ebx
+         jb davglp1               // Repeat until at alignment boundary
+davggo:
+         mov eax, FullLength
+         mov ecx, eax
+         sub eax, ebx          // subtract alignment fix
+         and eax, 0x00000007   // calc bytes over mult of 8
+         sub ecx, eax          // drop over bytes from original length
+         mov MMXLength, ecx
+   } // end _asm block
+   // Now do the math for the rest of the row
+   switch ( bpp )
+   {
+      case 3:
+      {
+         ActiveMask.use  = 0x0000000000ffffff;
+         ShiftBpp.use = 24;    // == 3 * 8
+         ShiftRem.use = 40;    // == 64 - 24
+         _asm {
+            // Re-init address pointers and offset
+            movq mm7, ActiveMask
+            mov ebx, diff      // ebx ==> x = offset to alignment boundary
+            movq mm5, LBCarryMask
+            mov edi, row       // edi ==> Avg(x)
+            movq mm4, HBClearMask
+            mov esi, prev_row        // esi ==> Prior(x)
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm2, [edi + ebx - 8]  // Load previous aligned 8 bytes
+                               // (we correct position in loop below)
+davg3lp:
+            movq mm0, [edi + ebx]      // Load mm0 with Avg(x)
+            // Add (Prev_row/2) to Average
+            movq mm3, mm5
+            psrlq mm2, ShiftRem      // Correct position Raw(x-bpp) data
+            movq mm1, [esi + ebx]    // Load mm1 with Prior(x)
+            movq mm6, mm7
+            pand mm3, mm1      // get lsb for each prev_row byte
+            psrlq mm1, 1       // divide prev_row bytes by 2
+            pand  mm1, mm4     // clear invalid bit 7 of each byte
+            paddb mm0, mm1     // add (Prev_row/2) to Avg for each byte
+            // Add 1st active group (Raw(x-bpp)/2) to Average with LBCarry
+            movq mm1, mm3      // now use mm1 for getting LBCarrys
+            pand mm1, mm2      // get LBCarrys for each byte where both
+                               // lsb's were == 1 (Only valid for active group)
+            psrlq mm2, 1       // divide raw bytes by 2
+            pand  mm2, mm4     // clear invalid bit 7 of each byte
+            paddb mm2, mm1     // add LBCarrys to (Raw(x-bpp)/2) for each byte
+            pand mm2, mm6      // Leave only Active Group 1 bytes to add to Avg
+            paddb mm0, mm2     // add (Raw/2) + LBCarrys to Avg for each Active
+                               //  byte
+            // Add 2nd active group (Raw(x-bpp)/2) to Average with LBCarry
+            psllq mm6, ShiftBpp  // shift the mm6 mask to cover bytes 3-5
+            movq mm2, mm0        // mov updated Raws to mm2
+            psllq mm2, ShiftBpp  // shift data to position correctly
+            movq mm1, mm3        // now use mm1 for getting LBCarrys
+            pand mm1, mm2      // get LBCarrys for each byte where both
+                               // lsb's were == 1 (Only valid for active group)
+            psrlq mm2, 1       // divide raw bytes by 2
+            pand  mm2, mm4     // clear invalid bit 7 of each byte
+            paddb mm2, mm1     // add LBCarrys to (Raw(x-bpp)/2) for each byte
+            pand mm2, mm6      // Leave only Active Group 2 bytes to add to Avg
+            paddb mm0, mm2     // add (Raw/2) + LBCarrys to Avg for each Active
+                               //  byte
+
+            // Add 3rd active group (Raw(x-bpp)/2) to Average with LBCarry
+            psllq mm6, ShiftBpp  // shift the mm6 mask to cover the last two
+                                 // bytes
+            movq mm2, mm0        // mov updated Raws to mm2
+            psllq mm2, ShiftBpp  // shift data to position correctly
+                              // Data only needs to be shifted once here to
+                              // get the correct x-bpp offset.
+            movq mm1, mm3     // now use mm1 for getting LBCarrys
+            pand mm1, mm2     // get LBCarrys for each byte where both
+                              // lsb's were == 1 (Only valid for active group)
+            psrlq mm2, 1      // divide raw bytes by 2
+            pand  mm2, mm4    // clear invalid bit 7 of each byte
+            paddb mm2, mm1    // add LBCarrys to (Raw(x-bpp)/2) for each byte
+            pand mm2, mm6     // Leave only Active Group 2 bytes to add to Avg
+            add ebx, 8
+            paddb mm0, mm2    // add (Raw/2) + LBCarrys to Avg for each Active
+                              // byte
+
+            // Now ready to write back to memory
+            movq [edi + ebx - 8], mm0
+            // Move updated Raw(x) to use as Raw(x-bpp) for next loop
+            cmp ebx, MMXLength
+            movq mm2, mm0     // mov updated Raw(x) to mm2
+            jb davg3lp
+         } // end _asm block
+      }
+      break;
+
+      case 6:
+      case 4:
+      case 7:
+      case 5:
+      {
+         ActiveMask.use  = 0xffffffffffffffff;  // use shift below to clear
+                                                // appropriate inactive bytes
+         ShiftBpp.use = bpp << 3;
+         ShiftRem.use = 64 - ShiftBpp.use;
+         _asm {
+            movq mm4, HBClearMask
+            // Re-init address pointers and offset
+            mov ebx, diff       // ebx ==> x = offset to alignment boundary
+            // Load ActiveMask and clear all bytes except for 1st active group
+            movq mm7, ActiveMask
+            mov edi, row         // edi ==> Avg(x)
+            psrlq mm7, ShiftRem
+            mov esi, prev_row    // esi ==> Prior(x)
+            movq mm6, mm7
+            movq mm5, LBCarryMask
+            psllq mm6, ShiftBpp  // Create mask for 2nd active group
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm2, [edi + ebx - 8]  // Load previous aligned 8 bytes
+                                 // (we correct position in loop below)
+davg4lp:
+            movq mm0, [edi + ebx]
+            psrlq mm2, ShiftRem  // shift data to position correctly
+            movq mm1, [esi + ebx]
+            // Add (Prev_row/2) to Average
+            movq mm3, mm5
+            pand mm3, mm1     // get lsb for each prev_row byte
+            psrlq mm1, 1      // divide prev_row bytes by 2
+            pand  mm1, mm4    // clear invalid bit 7 of each byte
+            paddb mm0, mm1    // add (Prev_row/2) to Avg for each byte
+            // Add 1st active group (Raw(x-bpp)/2) to Average with LBCarry
+            movq mm1, mm3     // now use mm1 for getting LBCarrys
+            pand mm1, mm2     // get LBCarrys for each byte where both
+                              // lsb's were == 1 (Only valid for active group)
+            psrlq mm2, 1      // divide raw bytes by 2
+            pand  mm2, mm4    // clear invalid bit 7 of each byte
+            paddb mm2, mm1    // add LBCarrys to (Raw(x-bpp)/2) for each byte
+            pand mm2, mm7     // Leave only Active Group 1 bytes to add to Avg
+            paddb mm0, mm2    // add (Raw/2) + LBCarrys to Avg for each Active
+                              // byte
+            // Add 2nd active group (Raw(x-bpp)/2) to Average with LBCarry
+            movq mm2, mm0     // mov updated Raws to mm2
+            psllq mm2, ShiftBpp // shift data to position correctly
+            add ebx, 8
+            movq mm1, mm3     // now use mm1 for getting LBCarrys
+            pand mm1, mm2     // get LBCarrys for each byte where both
+                              // lsb's were == 1 (Only valid for active group)
+            psrlq mm2, 1      // divide raw bytes by 2
+            pand  mm2, mm4    // clear invalid bit 7 of each byte
+            paddb mm2, mm1    // add LBCarrys to (Raw(x-bpp)/2) for each byte
+            pand mm2, mm6     // Leave only Active Group 2 bytes to add to Avg
+            paddb mm0, mm2    // add (Raw/2) + LBCarrys to Avg for each Active
+                              // byte
+            cmp ebx, MMXLength
+            // Now ready to write back to memory
+            movq [edi + ebx - 8], mm0
+            // Prep Raw(x-bpp) for next loop
+            movq mm2, mm0     // mov updated Raws to mm2
+            jb davg4lp
+         } // end _asm block
+      }
+      break;
+      case 2:
+      {
+         ActiveMask.use  = 0x000000000000ffff;
+         ShiftBpp.use = 16;   // == 2 * 8     [BUGFIX]
+         ShiftRem.use = 48;   // == 64 - 16   [BUGFIX]
+         _asm {
+            // Load ActiveMask
+            movq mm7, ActiveMask
+            // Re-init address pointers and offset
+            mov ebx, diff     // ebx ==> x = offset to alignment boundary
+            movq mm5, LBCarryMask
+            mov edi, row      // edi ==> Avg(x)
+            movq mm4, HBClearMask
+            mov esi, prev_row  // esi ==> Prior(x)
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm2, [edi + ebx - 8]  // Load previous aligned 8 bytes
+                              // (we correct position in loop below)
+davg2lp:
+            movq mm0, [edi + ebx]
+            psrlq mm2, ShiftRem  // shift data to position correctly   [BUGFIX]
+            movq mm1, [esi + ebx]
+            // Add (Prev_row/2) to Average
+            movq mm3, mm5
+            pand mm3, mm1     // get lsb for each prev_row byte
+            psrlq mm1, 1      // divide prev_row bytes by 2
+            pand  mm1, mm4    // clear invalid bit 7 of each byte
+            movq mm6, mm7
+            paddb mm0, mm1    // add (Prev_row/2) to Avg for each byte
+            // Add 1st active group (Raw(x-bpp)/2) to Average with LBCarry
+            movq mm1, mm3     // now use mm1 for getting LBCarrys
+            pand mm1, mm2     // get LBCarrys for each byte where both
+                              // lsb's were == 1 (Only valid for active group)
+            psrlq mm2, 1      // divide raw bytes by 2
+            pand  mm2, mm4    // clear invalid bit 7 of each byte
+            paddb mm2, mm1    // add LBCarrys to (Raw(x-bpp)/2) for each byte
+            pand mm2, mm6     // Leave only Active Group 1 bytes to add to Avg
+            paddb mm0, mm2 // add (Raw/2) + LBCarrys to Avg for each Active byte
+            // Add 2nd active group (Raw(x-bpp)/2) to Average with LBCarry
+            psllq mm6, ShiftBpp // shift the mm6 mask to cover bytes 2 & 3
+            movq mm2, mm0       // mov updated Raws to mm2
+            psllq mm2, ShiftBpp // shift data to position correctly
+            movq mm1, mm3       // now use mm1 for getting LBCarrys
+            pand mm1, mm2       // get LBCarrys for each byte where both
+                                // lsb's were == 1 (Only valid for active group)
+            psrlq mm2, 1        // divide raw bytes by 2
+            pand  mm2, mm4      // clear invalid bit 7 of each byte
+            paddb mm2, mm1      // add LBCarrys to (Raw(x-bpp)/2) for each byte
+            pand mm2, mm6       // Leave only Active Group 2 bytes to add to Avg
+            paddb mm0, mm2 // add (Raw/2) + LBCarrys to Avg for each Active byte
+
+            // Add rdd active group (Raw(x-bpp)/2) to Average with LBCarry
+            psllq mm6, ShiftBpp // shift the mm6 mask to cover bytes 4 & 5
+            movq mm2, mm0       // mov updated Raws to mm2
+            psllq mm2, ShiftBpp // shift data to position correctly
+                                // Data only needs to be shifted once here to
+                                // get the correct x-bpp offset.
+            movq mm1, mm3       // now use mm1 for getting LBCarrys
+            pand mm1, mm2       // get LBCarrys for each byte where both
+                                // lsb's were == 1 (Only valid for active group)
+            psrlq mm2, 1        // divide raw bytes by 2
+            pand  mm2, mm4      // clear invalid bit 7 of each byte
+            paddb mm2, mm1      // add LBCarrys to (Raw(x-bpp)/2) for each byte
+            pand mm2, mm6       // Leave only Active Group 2 bytes to add to Avg
+            paddb mm0, mm2 // add (Raw/2) + LBCarrys to Avg for each Active byte
+
+            // Add 4th active group (Raw(x-bpp)/2) to Average with LBCarry
+            psllq mm6, ShiftBpp  // shift the mm6 mask to cover bytes 6 & 7
+            movq mm2, mm0        // mov updated Raws to mm2
+            psllq mm2, ShiftBpp  // shift data to position correctly
+                                 // Data only needs to be shifted once here to
+                                 // get the correct x-bpp offset.
+            add ebx, 8
+            movq mm1, mm3    // now use mm1 for getting LBCarrys
+            pand mm1, mm2    // get LBCarrys for each byte where both
+                             // lsb's were == 1 (Only valid for active group)
+            psrlq mm2, 1     // divide raw bytes by 2
+            pand  mm2, mm4   // clear invalid bit 7 of each byte
+            paddb mm2, mm1   // add LBCarrys to (Raw(x-bpp)/2) for each byte
+            pand mm2, mm6    // Leave only Active Group 2 bytes to add to Avg
+            paddb mm0, mm2 // add (Raw/2) + LBCarrys to Avg for each Active byte
+
+            cmp ebx, MMXLength
+            // Now ready to write back to memory
+            movq [edi + ebx - 8], mm0
+            // Prep Raw(x-bpp) for next loop
+            movq mm2, mm0    // mov updated Raws to mm2
+            jb davg2lp
+        } // end _asm block
+      }
+      break;
+
+      case 1:                 // bpp == 1
+      {
+         _asm {
+            // Re-init address pointers and offset
+            mov ebx, diff     // ebx ==> x = offset to alignment boundary
+            mov edi, row      // edi ==> Avg(x)
+            cmp ebx, FullLength  // Test if offset at end of array
+            jnb davg1end
+            // Do Paeth decode for remaining bytes
+            mov esi, prev_row    // esi ==> Prior(x)
+            mov edx, edi
+            xor ecx, ecx         // zero ecx before using cl & cx in loop below
+            sub edx, bpp         // edx ==> Raw(x-bpp)
+davg1lp:
+            // Raw(x) = Avg(x) + ((Raw(x-bpp) + Prior(x))/2)
+            xor eax, eax
+            mov cl, [esi + ebx]  // load cl with Prior(x)
+            mov al, [edx + ebx]  // load al with Raw(x-bpp)
+            add ax, cx
+            inc ebx
+            shr ax, 1            // divide by 2
+            add al, [edi+ebx-1]  // Add Avg(x); -1 to offset inc ebx
+            cmp ebx, FullLength  // Check if at end of array
+            mov [edi+ebx-1], al  // Write back Raw(x);
+                         // mov does not affect flags; -1 to offset inc ebx
+            jb davg1lp
+davg1end:
+         } // end _asm block
+      }
+      return;
+
+      case 8:             // bpp == 8
+      {
+         _asm {
+            // Re-init address pointers and offset
+            mov ebx, diff           // ebx ==> x = offset to alignment boundary
+            movq mm5, LBCarryMask
+            mov edi, row            // edi ==> Avg(x)
+            movq mm4, HBClearMask
+            mov esi, prev_row       // esi ==> Prior(x)
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm2, [edi + ebx - 8]  // Load previous aligned 8 bytes
+                                // (NO NEED to correct position in loop below)
+davg8lp:
+            movq mm0, [edi + ebx]
+            movq mm3, mm5
+            movq mm1, [esi + ebx]
+            add ebx, 8
+            pand mm3, mm1       // get lsb for each prev_row byte
+            psrlq mm1, 1        // divide prev_row bytes by 2
+            pand mm3, mm2       // get LBCarrys for each byte where both
+                                // lsb's were == 1
+            psrlq mm2, 1        // divide raw bytes by 2
+            pand  mm1, mm4      // clear invalid bit 7 of each byte
+            paddb mm0, mm3      // add LBCarrys to Avg for each byte
+            pand  mm2, mm4      // clear invalid bit 7 of each byte
+            paddb mm0, mm1      // add (Prev_row/2) to Avg for each byte
+            paddb mm0, mm2      // add (Raw/2) to Avg for each byte
+            cmp ebx, MMXLength
+            movq [edi + ebx - 8], mm0
+            movq mm2, mm0       // reuse as Raw(x-bpp)
+            jb davg8lp
+        } // end _asm block
+      }
+      break;
+      default:                  // bpp greater than 8
+      {
+        _asm {
+            movq mm5, LBCarryMask
+            // Re-init address pointers and offset
+            mov ebx, diff       // ebx ==> x = offset to alignment boundary
+            mov edi, row        // edi ==> Avg(x)
+            movq mm4, HBClearMask
+            mov edx, edi
+            mov esi, prev_row   // esi ==> Prior(x)
+            sub edx, bpp        // edx ==> Raw(x-bpp)
+davgAlp:
+            movq mm0, [edi + ebx]
+            movq mm3, mm5
+            movq mm1, [esi + ebx]
+            pand mm3, mm1       // get lsb for each prev_row byte
+            movq mm2, [edx + ebx]
+            psrlq mm1, 1        // divide prev_row bytes by 2
+            pand mm3, mm2       // get LBCarrys for each byte where both
+                                // lsb's were == 1
+            psrlq mm2, 1        // divide raw bytes by 2
+            pand  mm1, mm4      // clear invalid bit 7 of each byte
+            paddb mm0, mm3      // add LBCarrys to Avg for each byte
+            pand  mm2, mm4      // clear invalid bit 7 of each byte
+            paddb mm0, mm1      // add (Prev_row/2) to Avg for each byte
+            add ebx, 8
+            paddb mm0, mm2      // add (Raw/2) to Avg for each byte
+            cmp ebx, MMXLength
+            movq [edi + ebx - 8], mm0
+            jb davgAlp
+        } // end _asm block
+      }
+      break;
+   }                         // end switch ( bpp )
+
+   _asm {
+         // MMX acceleration complete now do clean-up
+         // Check if any remaining bytes left to decode
+         mov ebx, MMXLength    // ebx ==> x = offset bytes remaining after MMX
+         mov edi, row          // edi ==> Avg(x)
+         cmp ebx, FullLength   // Test if offset at end of array
+         jnb davgend
+         // Do Paeth decode for remaining bytes
+         mov esi, prev_row     // esi ==> Prior(x)
+         mov edx, edi
+         xor ecx, ecx          // zero ecx before using cl & cx in loop below
+         sub edx, bpp          // edx ==> Raw(x-bpp)
+davglp2:
+         // Raw(x) = Avg(x) + ((Raw(x-bpp) + Prior(x))/2)
+         xor eax, eax
+         mov cl, [esi + ebx]   // load cl with Prior(x)
+         mov al, [edx + ebx]   // load al with Raw(x-bpp)
+         add ax, cx
+         inc ebx
+         shr ax, 1              // divide by 2
+         add al, [edi+ebx-1]    // Add Avg(x); -1 to offset inc ebx
+         cmp ebx, FullLength    // Check if at end of array
+         mov [edi+ebx-1], al    // Write back Raw(x);
+                          // mov does not affect flags; -1 to offset inc ebx
+         jb davglp2
+davgend:
+         emms             // End MMX instructions; prep for possible FP instrs.
+   } // end _asm block
+}
+
+// Optimized code for PNG Paeth filter decoder
+void /* PRIVATE */
+png_read_filter_row_mmx_paeth(png_row_infop row_info, png_bytep row,
+                              png_bytep prev_row)
+{
+  // These variables are declared
+  // here to ensure alignment on 8-byte boundaries.
+  union uAll  ActiveMask, ActiveMask2, ActiveMaskEnd, ShiftBpp, ShiftRem;
+
+   png_uint_32 FullLength;
+   png_uint_32 MMXLength;
+   //png_uint_32 len;
+   int bpp;
+   int diff;
+   //int ptemp;
+   int patemp, pbtemp, pctemp;
+
+   bpp = (row_info->pixel_depth + 7) >> 3; // Get # bytes per pixel
+   FullLength  = row_info->rowbytes; // # of bytes to filter
+   _asm
+   {
+         xor ebx, ebx        // ebx ==> x offset
+         mov edi, row
+         xor edx, edx        // edx ==> x-bpp offset
+         mov esi, prev_row
+         xor eax, eax
+
+         // Compute the Raw value for the first bpp bytes
+         // Note: the formula works out to be always
+         //   Paeth(x) = Raw(x) + Prior(x)      where x < bpp
+dpthrlp:
+         mov al, [edi + ebx]
+         add al, [esi + ebx]
+         inc ebx
+         cmp ebx, bpp
+         mov [edi + ebx - 1], al
+         jb dpthrlp
+         // get # of bytes to alignment
+         mov diff, edi         // take start of row
+         add diff, ebx         // add bpp
+         xor ecx, ecx
+         add diff, 0xf         // add 7 + 8 to incr past alignment boundary
+         and diff, 0xfffffff8  // mask to alignment boundary
+         sub diff, edi         // subtract from start ==> value ebx at alignment
+         jz dpthgo
+         // fix alignment
+dpthlp1:
+         xor eax, eax
+         // pav = p - a = (a + b - c) - a = b - c
+         mov al, [esi + ebx]   // load Prior(x) into al
+         mov cl, [esi + edx]   // load Prior(x-bpp) into cl
+         sub eax, ecx          // subtract Prior(x-bpp)
+         mov patemp, eax       // Save pav for later use
+         xor eax, eax
+         // pbv = p - b = (a + b - c) - b = a - c
+         mov al, [edi + edx]   // load Raw(x-bpp) into al
+         sub eax, ecx          // subtract Prior(x-bpp)
+         mov ecx, eax
+         // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+         add eax, patemp       // pcv = pav + pbv
+         // pc = abs(pcv)
+         test eax, 0x80000000
+         jz dpthpca
+         neg eax               // reverse sign of neg values
+dpthpca:
+         mov pctemp, eax       // save pc for later use
+         // pb = abs(pbv)
+         test ecx, 0x80000000
+         jz dpthpba
+         neg ecx               // reverse sign of neg values
+dpthpba:
+         mov pbtemp, ecx       // save pb for later use
+         // pa = abs(pav)
+         mov eax, patemp
+         test eax, 0x80000000
+         jz dpthpaa
+         neg eax               // reverse sign of neg values
+dpthpaa:
+         mov patemp, eax       // save pa for later use
+         // test if pa <= pb
+         cmp eax, ecx
+         jna dpthabb
+         // pa > pb; now test if pb <= pc
+         cmp ecx, pctemp
+         jna dpthbbc
+         // pb > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+         mov cl, [esi + edx]  // load Prior(x-bpp) into cl
+         jmp dpthpaeth
+dpthbbc:
+         // pb <= pc; Raw(x) = Paeth(x) + Prior(x)
+         mov cl, [esi + ebx]   // load Prior(x) into cl
+         jmp dpthpaeth
+dpthabb:
+         // pa <= pb; now test if pa <= pc
+         cmp eax, pctemp
+         jna dpthabc
+         // pa > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+         mov cl, [esi + edx]  // load Prior(x-bpp) into cl
+         jmp dpthpaeth
+dpthabc:
+         // pa <= pc; Raw(x) = Paeth(x) + Raw(x-bpp)
+         mov cl, [edi + edx]  // load Raw(x-bpp) into cl
+dpthpaeth:
+         inc ebx
+         inc edx
+         // Raw(x) = (Paeth(x) + Paeth_Predictor( a, b, c )) mod 256
+         add [edi + ebx - 1], cl
+         cmp ebx, diff
+         jb dpthlp1
+dpthgo:
+         mov ecx, FullLength
+         mov eax, ecx
+         sub eax, ebx          // subtract alignment fix
+         and eax, 0x00000007   // calc bytes over mult of 8
+         sub ecx, eax          // drop over bytes from original length
+         mov MMXLength, ecx
+   } // end _asm block
+   // Now do the math for the rest of the row
+   switch ( bpp )
+   {
+      case 3:
+      {
+         ActiveMask.use = 0x0000000000ffffff;
+         ActiveMaskEnd.use = 0xffff000000000000;
+         ShiftBpp.use = 24;    // == bpp(3) * 8
+         ShiftRem.use = 40;    // == 64 - 24
+         _asm
+         {
+            mov ebx, diff
+            mov edi, row
+            mov esi, prev_row
+            pxor mm0, mm0
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm1, [edi+ebx-8]
+dpth3lp:
+            psrlq mm1, ShiftRem     // shift last 3 bytes to 1st 3 bytes
+            movq mm2, [esi + ebx]   // load b=Prior(x)
+            punpcklbw mm1, mm0      // Unpack High bytes of a
+            movq mm3, [esi+ebx-8]   // Prep c=Prior(x-bpp) bytes
+            punpcklbw mm2, mm0      // Unpack High bytes of b
+            psrlq mm3, ShiftRem     // shift last 3 bytes to 1st 3 bytes
+            // pav = p - a = (a + b - c) - a = b - c
+            movq mm4, mm2
+            punpcklbw mm3, mm0      // Unpack High bytes of c
+            // pbv = p - b = (a + b - c) - b = a - c
+            movq mm5, mm1
+            psubw mm4, mm3
+            pxor mm7, mm7
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            movq mm6, mm4
+            psubw mm5, mm3
+
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            pcmpgtw mm0, mm4    // Create mask pav bytes < 0
+            paddw mm6, mm5
+            pand mm0, mm4       // Only pav bytes < 0 in mm7
+            pcmpgtw mm7, mm5    // Create mask pbv bytes < 0
+            psubw mm4, mm0
+            pand mm7, mm5       // Only pbv bytes < 0 in mm0
+            psubw mm4, mm0
+            psubw mm5, mm7
+            pxor mm0, mm0
+            pcmpgtw mm0, mm6    // Create mask pcv bytes < 0
+            pand mm0, mm6       // Only pav bytes < 0 in mm7
+            psubw mm5, mm7
+            psubw mm6, mm0
+            //  test pa <= pb
+            movq mm7, mm4
+            psubw mm6, mm0
+            pcmpgtw mm7, mm5    // pa > pb?
+            movq mm0, mm7
+            // use mm7 mask to merge pa & pb
+            pand mm5, mm7
+            // use mm0 mask copy to merge a & b
+            pand mm2, mm0
+            pandn mm7, mm4
+            pandn mm0, mm1
+            paddw mm7, mm5
+            paddw mm0, mm2
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            pcmpgtw mm7, mm6       // pab > pc?
+            pxor mm1, mm1
+            pand mm3, mm7
+            pandn mm7, mm0
+            paddw mm7, mm3
+            pxor mm0, mm0
+            packuswb mm7, mm1
+            movq mm3, [esi + ebx]   // load c=Prior(x-bpp)
+            pand mm7, ActiveMask
+            movq mm2, mm3           // load b=Prior(x) step 1
+            paddb mm7, [edi + ebx]  // add Paeth predictor with Raw(x)
+            punpcklbw mm3, mm0      // Unpack High bytes of c
+            movq [edi + ebx], mm7   // write back updated value
+            movq mm1, mm7           // Now mm1 will be used as Raw(x-bpp)
+            // Now do Paeth for 2nd set of bytes (3-5)
+            psrlq mm2, ShiftBpp     // load b=Prior(x) step 2
+            punpcklbw mm1, mm0      // Unpack High bytes of a
+            pxor mm7, mm7
+            punpcklbw mm2, mm0      // Unpack High bytes of b
+            // pbv = p - b = (a + b - c) - b = a - c
+            movq mm5, mm1
+            // pav = p - a = (a + b - c) - a = b - c
+            movq mm4, mm2
+            psubw mm5, mm3
+            psubw mm4, mm3
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) =
+            //       pav + pbv = pbv + pav
+            movq mm6, mm5
+            paddw mm6, mm4
+
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            pcmpgtw mm0, mm5       // Create mask pbv bytes < 0
+            pcmpgtw mm7, mm4       // Create mask pav bytes < 0
+            pand mm0, mm5          // Only pbv bytes < 0 in mm0
+            pand mm7, mm4          // Only pav bytes < 0 in mm7
+            psubw mm5, mm0
+            psubw mm4, mm7
+            psubw mm5, mm0
+            psubw mm4, mm7
+            pxor mm0, mm0
+            pcmpgtw mm0, mm6       // Create mask pcv bytes < 0
+            pand mm0, mm6          // Only pav bytes < 0 in mm7
+            psubw mm6, mm0
+            //  test pa <= pb
+            movq mm7, mm4
+            psubw mm6, mm0
+            pcmpgtw mm7, mm5       // pa > pb?
+            movq mm0, mm7
+            // use mm7 mask to merge pa & pb
+            pand mm5, mm7
+            // use mm0 mask copy to merge a & b
+            pand mm2, mm0
+            pandn mm7, mm4
+            pandn mm0, mm1
+            paddw mm7, mm5
+            paddw mm0, mm2
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            pcmpgtw mm7, mm6       // pab > pc?
+            movq mm2, [esi + ebx]  // load b=Prior(x)
+            pand mm3, mm7
+            pandn mm7, mm0
+            pxor mm1, mm1
+            paddw mm7, mm3
+            pxor mm0, mm0
+            packuswb mm7, mm1
+            movq mm3, mm2           // load c=Prior(x-bpp) step 1
+            pand mm7, ActiveMask
+            punpckhbw mm2, mm0      // Unpack High bytes of b
+            psllq mm7, ShiftBpp     // Shift bytes to 2nd group of 3 bytes
+             // pav = p - a = (a + b - c) - a = b - c
+            movq mm4, mm2
+            paddb mm7, [edi + ebx]  // add Paeth predictor with Raw(x)
+            psllq mm3, ShiftBpp     // load c=Prior(x-bpp) step 2
+            movq [edi + ebx], mm7   // write back updated value
+            movq mm1, mm7
+            punpckhbw mm3, mm0      // Unpack High bytes of c
+            psllq mm1, ShiftBpp     // Shift bytes
+                                    // Now mm1 will be used as Raw(x-bpp)
+            // Now do Paeth for 3rd, and final, set of bytes (6-7)
+            pxor mm7, mm7
+            punpckhbw mm1, mm0      // Unpack High bytes of a
+            psubw mm4, mm3
+            // pbv = p - b = (a + b - c) - b = a - c
+            movq mm5, mm1
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            movq mm6, mm4
+            psubw mm5, mm3
+            pxor mm0, mm0
+            paddw mm6, mm5
+
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            pcmpgtw mm0, mm4    // Create mask pav bytes < 0
+            pcmpgtw mm7, mm5    // Create mask pbv bytes < 0
+            pand mm0, mm4       // Only pav bytes < 0 in mm7
+            pand mm7, mm5       // Only pbv bytes < 0 in mm0
+            psubw mm4, mm0
+            psubw mm5, mm7
+            psubw mm4, mm0
+            psubw mm5, mm7
+            pxor mm0, mm0
+            pcmpgtw mm0, mm6    // Create mask pcv bytes < 0
+            pand mm0, mm6       // Only pav bytes < 0 in mm7
+            psubw mm6, mm0
+            //  test pa <= pb
+            movq mm7, mm4
+            psubw mm6, mm0
+            pcmpgtw mm7, mm5    // pa > pb?
+            movq mm0, mm7
+            // use mm0 mask copy to merge a & b
+            pand mm2, mm0
+            // use mm7 mask to merge pa & pb
+            pand mm5, mm7
+            pandn mm0, mm1
+            pandn mm7, mm4
+            paddw mm0, mm2
+            paddw mm7, mm5
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            pcmpgtw mm7, mm6    // pab > pc?
+            pand mm3, mm7
+            pandn mm7, mm0
+            paddw mm7, mm3
+            pxor mm1, mm1
+            packuswb mm1, mm7
+            // Step ebx to next set of 8 bytes and repeat loop til done
+            add ebx, 8
+            pand mm1, ActiveMaskEnd
+            paddb mm1, [edi + ebx - 8] // add Paeth predictor with Raw(x)
+
+            cmp ebx, MMXLength
+            pxor mm0, mm0              // pxor does not affect flags
+            movq [edi + ebx - 8], mm1  // write back updated value
+                                 // mm1 will be used as Raw(x-bpp) next loop
+                           // mm3 ready to be used as Prior(x-bpp) next loop
+            jb dpth3lp
+         } // end _asm block
+      }
+      break;
+
+      case 6:
+      case 7:
+      case 5:
+      {
+         ActiveMask.use  = 0x00000000ffffffff;
+         ActiveMask2.use = 0xffffffff00000000;
+         ShiftBpp.use = bpp << 3;    // == bpp * 8
+         ShiftRem.use = 64 - ShiftBpp.use;
+         _asm
+         {
+            mov ebx, diff
+            mov edi, row
+            mov esi, prev_row
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm1, [edi+ebx-8]
+            pxor mm0, mm0
+dpth6lp:
+            // Must shift to position Raw(x-bpp) data
+            psrlq mm1, ShiftRem
+            // Do first set of 4 bytes
+            movq mm3, [esi+ebx-8]      // read c=Prior(x-bpp) bytes
+            punpcklbw mm1, mm0      // Unpack Low bytes of a
+            movq mm2, [esi + ebx]   // load b=Prior(x)
+            punpcklbw mm2, mm0      // Unpack Low bytes of b
+            // Must shift to position Prior(x-bpp) data
+            psrlq mm3, ShiftRem
+            // pav = p - a = (a + b - c) - a = b - c
+            movq mm4, mm2
+            punpcklbw mm3, mm0      // Unpack Low bytes of c
+            // pbv = p - b = (a + b - c) - b = a - c
+            movq mm5, mm1
+            psubw mm4, mm3
+            pxor mm7, mm7
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            movq mm6, mm4
+            psubw mm5, mm3
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            pcmpgtw mm0, mm4    // Create mask pav bytes < 0
+            paddw mm6, mm5
+            pand mm0, mm4       // Only pav bytes < 0 in mm7
+            pcmpgtw mm7, mm5    // Create mask pbv bytes < 0
+            psubw mm4, mm0
+            pand mm7, mm5       // Only pbv bytes < 0 in mm0
+            psubw mm4, mm0
+            psubw mm5, mm7
+            pxor mm0, mm0
+            pcmpgtw mm0, mm6    // Create mask pcv bytes < 0
+            pand mm0, mm6       // Only pav bytes < 0 in mm7
+            psubw mm5, mm7
+            psubw mm6, mm0
+            //  test pa <= pb
+            movq mm7, mm4
+            psubw mm6, mm0
+            pcmpgtw mm7, mm5    // pa > pb?
+            movq mm0, mm7
+            // use mm7 mask to merge pa & pb
+            pand mm5, mm7
+            // use mm0 mask copy to merge a & b
+            pand mm2, mm0
+            pandn mm7, mm4
+            pandn mm0, mm1
+            paddw mm7, mm5
+            paddw mm0, mm2
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            pcmpgtw mm7, mm6    // pab > pc?
+            pxor mm1, mm1
+            pand mm3, mm7
+            pandn mm7, mm0
+            paddw mm7, mm3
+            pxor mm0, mm0
+            packuswb mm7, mm1
+            movq mm3, [esi + ebx - 8]  // load c=Prior(x-bpp)
+            pand mm7, ActiveMask
+            psrlq mm3, ShiftRem
+            movq mm2, [esi + ebx]      // load b=Prior(x) step 1
+            paddb mm7, [edi + ebx]     // add Paeth predictor with Raw(x)
+            movq mm6, mm2
+            movq [edi + ebx], mm7      // write back updated value
+            movq mm1, [edi+ebx-8]
+            psllq mm6, ShiftBpp
+            movq mm5, mm7
+            psrlq mm1, ShiftRem
+            por mm3, mm6
+            psllq mm5, ShiftBpp
+            punpckhbw mm3, mm0         // Unpack High bytes of c
+            por mm1, mm5
+            // Do second set of 4 bytes
+            punpckhbw mm2, mm0         // Unpack High bytes of b
+            punpckhbw mm1, mm0         // Unpack High bytes of a
+            // pav = p - a = (a + b - c) - a = b - c
+            movq mm4, mm2
+            // pbv = p - b = (a + b - c) - b = a - c
+            movq mm5, mm1
+            psubw mm4, mm3
+            pxor mm7, mm7
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            movq mm6, mm4
+            psubw mm5, mm3
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            pcmpgtw mm0, mm4       // Create mask pav bytes < 0
+            paddw mm6, mm5
+            pand mm0, mm4          // Only pav bytes < 0 in mm7
+            pcmpgtw mm7, mm5       // Create mask pbv bytes < 0
+            psubw mm4, mm0
+            pand mm7, mm5          // Only pbv bytes < 0 in mm0
+            psubw mm4, mm0
+            psubw mm5, mm7
+            pxor mm0, mm0
+            pcmpgtw mm0, mm6       // Create mask pcv bytes < 0
+            pand mm0, mm6          // Only pav bytes < 0 in mm7
+            psubw mm5, mm7
+            psubw mm6, mm0
+            //  test pa <= pb
+            movq mm7, mm4
+            psubw mm6, mm0
+            pcmpgtw mm7, mm5       // pa > pb?
+            movq mm0, mm7
+            // use mm7 mask to merge pa & pb
+            pand mm5, mm7
+            // use mm0 mask copy to merge a & b
+            pand mm2, mm0
+            pandn mm7, mm4
+            pandn mm0, mm1
+            paddw mm7, mm5
+            paddw mm0, mm2
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            pcmpgtw mm7, mm6           // pab > pc?
+            pxor mm1, mm1
+            pand mm3, mm7
+            pandn mm7, mm0
+            pxor mm1, mm1
+            paddw mm7, mm3
+            pxor mm0, mm0
+            // Step ex to next set of 8 bytes and repeat loop til done
+            add ebx, 8
+            packuswb mm1, mm7
+            paddb mm1, [edi + ebx - 8]     // add Paeth predictor with Raw(x)
+            cmp ebx, MMXLength
+            movq [edi + ebx - 8], mm1      // write back updated value
+                                // mm1 will be used as Raw(x-bpp) next loop
+            jb dpth6lp
+         } // end _asm block
+      }
+      break;
+
+      case 4:
+      {
+         ActiveMask.use  = 0x00000000ffffffff;
+         _asm {
+            mov ebx, diff
+            mov edi, row
+            mov esi, prev_row
+            pxor mm0, mm0
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm1, [edi+ebx-8]    // Only time should need to read
+                                     //  a=Raw(x-bpp) bytes
+dpth4lp:
+            // Do first set of 4 bytes
+            movq mm3, [esi+ebx-8]    // read c=Prior(x-bpp) bytes
+            punpckhbw mm1, mm0       // Unpack Low bytes of a
+            movq mm2, [esi + ebx]    // load b=Prior(x)
+            punpcklbw mm2, mm0       // Unpack High bytes of b
+            // pav = p - a = (a + b - c) - a = b - c
+            movq mm4, mm2
+            punpckhbw mm3, mm0       // Unpack High bytes of c
+            // pbv = p - b = (a + b - c) - b = a - c
+            movq mm5, mm1
+            psubw mm4, mm3
+            pxor mm7, mm7
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            movq mm6, mm4
+            psubw mm5, mm3
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            pcmpgtw mm0, mm4       // Create mask pav bytes < 0
+            paddw mm6, mm5
+            pand mm0, mm4          // Only pav bytes < 0 in mm7
+            pcmpgtw mm7, mm5       // Create mask pbv bytes < 0
+            psubw mm4, mm0
+            pand mm7, mm5          // Only pbv bytes < 0 in mm0
+            psubw mm4, mm0
+            psubw mm5, mm7
+            pxor mm0, mm0
+            pcmpgtw mm0, mm6       // Create mask pcv bytes < 0
+            pand mm0, mm6          // Only pav bytes < 0 in mm7
+            psubw mm5, mm7
+            psubw mm6, mm0
+            //  test pa <= pb
+            movq mm7, mm4
+            psubw mm6, mm0
+            pcmpgtw mm7, mm5       // pa > pb?
+            movq mm0, mm7
+            // use mm7 mask to merge pa & pb
+            pand mm5, mm7
+            // use mm0 mask copy to merge a & b
+            pand mm2, mm0
+            pandn mm7, mm4
+            pandn mm0, mm1
+            paddw mm7, mm5
+            paddw mm0, mm2
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            pcmpgtw mm7, mm6       // pab > pc?
+            pxor mm1, mm1
+            pand mm3, mm7
+            pandn mm7, mm0
+            paddw mm7, mm3
+            pxor mm0, mm0
+            packuswb mm7, mm1
+            movq mm3, [esi + ebx]      // load c=Prior(x-bpp)
+            pand mm7, ActiveMask
+            movq mm2, mm3              // load b=Prior(x) step 1
+            paddb mm7, [edi + ebx]     // add Paeth predictor with Raw(x)
+            punpcklbw mm3, mm0         // Unpack High bytes of c
+            movq [edi + ebx], mm7      // write back updated value
+            movq mm1, mm7              // Now mm1 will be used as Raw(x-bpp)
+            // Do second set of 4 bytes
+            punpckhbw mm2, mm0         // Unpack Low bytes of b
+            punpcklbw mm1, mm0         // Unpack Low bytes of a
+            // pav = p - a = (a + b - c) - a = b - c
+            movq mm4, mm2
+            // pbv = p - b = (a + b - c) - b = a - c
+            movq mm5, mm1
+            psubw mm4, mm3
+            pxor mm7, mm7
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            movq mm6, mm4
+            psubw mm5, mm3
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            pcmpgtw mm0, mm4       // Create mask pav bytes < 0
+            paddw mm6, mm5
+            pand mm0, mm4          // Only pav bytes < 0 in mm7
+            pcmpgtw mm7, mm5       // Create mask pbv bytes < 0
+            psubw mm4, mm0
+            pand mm7, mm5          // Only pbv bytes < 0 in mm0
+            psubw mm4, mm0
+            psubw mm5, mm7
+            pxor mm0, mm0
+            pcmpgtw mm0, mm6       // Create mask pcv bytes < 0
+            pand mm0, mm6          // Only pav bytes < 0 in mm7
+            psubw mm5, mm7
+            psubw mm6, mm0
+            //  test pa <= pb
+            movq mm7, mm4
+            psubw mm6, mm0
+            pcmpgtw mm7, mm5       // pa > pb?
+            movq mm0, mm7
+            // use mm7 mask to merge pa & pb
+            pand mm5, mm7
+            // use mm0 mask copy to merge a & b
+            pand mm2, mm0
+            pandn mm7, mm4
+            pandn mm0, mm1
+            paddw mm7, mm5
+            paddw mm0, mm2
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            pcmpgtw mm7, mm6       // pab > pc?
+            pxor mm1, mm1
+            pand mm3, mm7
+            pandn mm7, mm0
+            pxor mm1, mm1
+            paddw mm7, mm3
+            pxor mm0, mm0
+            // Step ex to next set of 8 bytes and repeat loop til done
+            add ebx, 8
+            packuswb mm1, mm7
+            paddb mm1, [edi + ebx - 8]     // add Paeth predictor with Raw(x)
+            cmp ebx, MMXLength
+            movq [edi + ebx - 8], mm1      // write back updated value
+                                // mm1 will be used as Raw(x-bpp) next loop
+            jb dpth4lp
+         } // end _asm block
+      }
+      break;
+      case 8:                          // bpp == 8
+      {
+         ActiveMask.use  = 0x00000000ffffffff;
+         _asm {
+            mov ebx, diff
+            mov edi, row
+            mov esi, prev_row
+            pxor mm0, mm0
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm1, [edi+ebx-8]      // Only time should need to read
+                                       //  a=Raw(x-bpp) bytes
+dpth8lp:
+            // Do first set of 4 bytes
+            movq mm3, [esi+ebx-8]      // read c=Prior(x-bpp) bytes
+            punpcklbw mm1, mm0         // Unpack Low bytes of a
+            movq mm2, [esi + ebx]      // load b=Prior(x)
+            punpcklbw mm2, mm0         // Unpack Low bytes of b
+            // pav = p - a = (a + b - c) - a = b - c
+            movq mm4, mm2
+            punpcklbw mm3, mm0         // Unpack Low bytes of c
+            // pbv = p - b = (a + b - c) - b = a - c
+            movq mm5, mm1
+            psubw mm4, mm3
+            pxor mm7, mm7
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            movq mm6, mm4
+            psubw mm5, mm3
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            pcmpgtw mm0, mm4       // Create mask pav bytes < 0
+            paddw mm6, mm5
+            pand mm0, mm4          // Only pav bytes < 0 in mm7
+            pcmpgtw mm7, mm5       // Create mask pbv bytes < 0
+            psubw mm4, mm0
+            pand mm7, mm5          // Only pbv bytes < 0 in mm0
+            psubw mm4, mm0
+            psubw mm5, mm7
+            pxor mm0, mm0
+            pcmpgtw mm0, mm6       // Create mask pcv bytes < 0
+            pand mm0, mm6          // Only pav bytes < 0 in mm7
+            psubw mm5, mm7
+            psubw mm6, mm0
+            //  test pa <= pb
+            movq mm7, mm4
+            psubw mm6, mm0
+            pcmpgtw mm7, mm5       // pa > pb?
+            movq mm0, mm7
+            // use mm7 mask to merge pa & pb
+            pand mm5, mm7
+            // use mm0 mask copy to merge a & b
+            pand mm2, mm0
+            pandn mm7, mm4
+            pandn mm0, mm1
+            paddw mm7, mm5
+            paddw mm0, mm2
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            pcmpgtw mm7, mm6       // pab > pc?
+            pxor mm1, mm1
+            pand mm3, mm7
+            pandn mm7, mm0
+            paddw mm7, mm3
+            pxor mm0, mm0
+            packuswb mm7, mm1
+            movq mm3, [esi+ebx-8]    // read c=Prior(x-bpp) bytes
+            pand mm7, ActiveMask
+            movq mm2, [esi + ebx]    // load b=Prior(x)
+            paddb mm7, [edi + ebx]   // add Paeth predictor with Raw(x)
+            punpckhbw mm3, mm0       // Unpack High bytes of c
+            movq [edi + ebx], mm7    // write back updated value
+            movq mm1, [edi+ebx-8]    // read a=Raw(x-bpp) bytes
+
+            // Do second set of 4 bytes
+            punpckhbw mm2, mm0       // Unpack High bytes of b
+            punpckhbw mm1, mm0       // Unpack High bytes of a
+            // pav = p - a = (a + b - c) - a = b - c
+            movq mm4, mm2
+            // pbv = p - b = (a + b - c) - b = a - c
+            movq mm5, mm1
+            psubw mm4, mm3
+            pxor mm7, mm7
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            movq mm6, mm4
+            psubw mm5, mm3
+            // pa = abs(p-a) = abs(pav)
+            // pb = abs(p-b) = abs(pbv)
+            // pc = abs(p-c) = abs(pcv)
+            pcmpgtw mm0, mm4       // Create mask pav bytes < 0
+            paddw mm6, mm5
+            pand mm0, mm4          // Only pav bytes < 0 in mm7
+            pcmpgtw mm7, mm5       // Create mask pbv bytes < 0
+            psubw mm4, mm0
+            pand mm7, mm5          // Only pbv bytes < 0 in mm0
+            psubw mm4, mm0
+            psubw mm5, mm7
+            pxor mm0, mm0
+            pcmpgtw mm0, mm6       // Create mask pcv bytes < 0
+            pand mm0, mm6          // Only pav bytes < 0 in mm7
+            psubw mm5, mm7
+            psubw mm6, mm0
+            //  test pa <= pb
+            movq mm7, mm4
+            psubw mm6, mm0
+            pcmpgtw mm7, mm5       // pa > pb?
+            movq mm0, mm7
+            // use mm7 mask to merge pa & pb
+            pand mm5, mm7
+            // use mm0 mask copy to merge a & b
+            pand mm2, mm0
+            pandn mm7, mm4
+            pandn mm0, mm1
+            paddw mm7, mm5
+            paddw mm0, mm2
+            //  test  ((pa <= pb)? pa:pb) <= pc
+            pcmpgtw mm7, mm6       // pab > pc?
+            pxor mm1, mm1
+            pand mm3, mm7
+            pandn mm7, mm0
+            pxor mm1, mm1
+            paddw mm7, mm3
+            pxor mm0, mm0
+            // Step ex to next set of 8 bytes and repeat loop til done
+            add ebx, 8
+            packuswb mm1, mm7
+            paddb mm1, [edi + ebx - 8]     // add Paeth predictor with Raw(x)
+            cmp ebx, MMXLength
+            movq [edi + ebx - 8], mm1      // write back updated value
+                            // mm1 will be used as Raw(x-bpp) next loop
+            jb dpth8lp
+         } // end _asm block
+      }
+      break;
+
+      case 1:                // bpp = 1
+      case 2:                // bpp = 2
+      default:               // bpp > 8
+      {
+         _asm {
+            mov ebx, diff
+            cmp ebx, FullLength
+            jnb dpthdend
+            mov edi, row
+            mov esi, prev_row
+            // Do Paeth decode for remaining bytes
+            mov edx, ebx
+            xor ecx, ecx        // zero ecx before using cl & cx in loop below
+            sub edx, bpp        // Set edx = ebx - bpp
+dpthdlp:
+            xor eax, eax
+            // pav = p - a = (a + b - c) - a = b - c
+            mov al, [esi + ebx]        // load Prior(x) into al
+            mov cl, [esi + edx]        // load Prior(x-bpp) into cl
+            sub eax, ecx                 // subtract Prior(x-bpp)
+            mov patemp, eax                 // Save pav for later use
+            xor eax, eax
+            // pbv = p - b = (a + b - c) - b = a - c
+            mov al, [edi + edx]        // load Raw(x-bpp) into al
+            sub eax, ecx                 // subtract Prior(x-bpp)
+            mov ecx, eax
+            // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+            add eax, patemp                 // pcv = pav + pbv
+            // pc = abs(pcv)
+            test eax, 0x80000000
+            jz dpthdpca
+            neg eax                     // reverse sign of neg values
+dpthdpca:
+            mov pctemp, eax             // save pc for later use
+            // pb = abs(pbv)
+            test ecx, 0x80000000
+            jz dpthdpba
+            neg ecx                     // reverse sign of neg values
+dpthdpba:
+            mov pbtemp, ecx             // save pb for later use
+            // pa = abs(pav)
+            mov eax, patemp
+            test eax, 0x80000000
+            jz dpthdpaa
+            neg eax                     // reverse sign of neg values
+dpthdpaa:
+            mov patemp, eax             // save pa for later use
+            // test if pa <= pb
+            cmp eax, ecx
+            jna dpthdabb
+            // pa > pb; now test if pb <= pc
+            cmp ecx, pctemp
+            jna dpthdbbc
+            // pb > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+            mov cl, [esi + edx]  // load Prior(x-bpp) into cl
+            jmp dpthdpaeth
+dpthdbbc:
+            // pb <= pc; Raw(x) = Paeth(x) + Prior(x)
+            mov cl, [esi + ebx]        // load Prior(x) into cl
+            jmp dpthdpaeth
+dpthdabb:
+            // pa <= pb; now test if pa <= pc
+            cmp eax, pctemp
+            jna dpthdabc
+            // pa > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+            mov cl, [esi + edx]  // load Prior(x-bpp) into cl
+            jmp dpthdpaeth
+dpthdabc:
+            // pa <= pc; Raw(x) = Paeth(x) + Raw(x-bpp)
+            mov cl, [edi + edx]  // load Raw(x-bpp) into cl
+dpthdpaeth:
+            inc ebx
+            inc edx
+            // Raw(x) = (Paeth(x) + Paeth_Predictor( a, b, c )) mod 256
+            add [edi + ebx - 1], cl
+            cmp ebx, FullLength
+            jb dpthdlp
+dpthdend:
+         } // end _asm block
+      }
+      return;                   // No need to go further with this one
+   }                         // end switch ( bpp )
+   _asm
+   {
+         // MMX acceleration complete now do clean-up
+         // Check if any remaining bytes left to decode
+         mov ebx, MMXLength
+         cmp ebx, FullLength
+         jnb dpthend
+         mov edi, row
+         mov esi, prev_row
+         // Do Paeth decode for remaining bytes
+         mov edx, ebx
+         xor ecx, ecx         // zero ecx before using cl & cx in loop below
+         sub edx, bpp         // Set edx = ebx - bpp
+dpthlp2:
+         xor eax, eax
+         // pav = p - a = (a + b - c) - a = b - c
+         mov al, [esi + ebx]  // load Prior(x) into al
+         mov cl, [esi + edx]  // load Prior(x-bpp) into cl
+         sub eax, ecx         // subtract Prior(x-bpp)
+         mov patemp, eax      // Save pav for later use
+         xor eax, eax
+         // pbv = p - b = (a + b - c) - b = a - c
+         mov al, [edi + edx]  // load Raw(x-bpp) into al
+         sub eax, ecx         // subtract Prior(x-bpp)
+         mov ecx, eax
+         // pcv = p - c = (a + b - c) -c = (a - c) + (b - c) = pav + pbv
+         add eax, patemp      // pcv = pav + pbv
+         // pc = abs(pcv)
+         test eax, 0x80000000
+         jz dpthpca2
+         neg eax              // reverse sign of neg values
+dpthpca2:
+         mov pctemp, eax      // save pc for later use
+         // pb = abs(pbv)
+         test ecx, 0x80000000
+         jz dpthpba2
+         neg ecx              // reverse sign of neg values
+dpthpba2:
+         mov pbtemp, ecx      // save pb for later use
+         // pa = abs(pav)
+         mov eax, patemp
+         test eax, 0x80000000
+         jz dpthpaa2
+         neg eax              // reverse sign of neg values
+dpthpaa2:
+         mov patemp, eax      // save pa for later use
+         // test if pa <= pb
+         cmp eax, ecx
+         jna dpthabb2
+         // pa > pb; now test if pb <= pc
+         cmp ecx, pctemp
+         jna dpthbbc2
+         // pb > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+         mov cl, [esi + edx]  // load Prior(x-bpp) into cl
+         jmp dpthpaeth2
+dpthbbc2:
+         // pb <= pc; Raw(x) = Paeth(x) + Prior(x)
+         mov cl, [esi + ebx]        // load Prior(x) into cl
+         jmp dpthpaeth2
+dpthabb2:
+         // pa <= pb; now test if pa <= pc
+         cmp eax, pctemp
+         jna dpthabc2
+         // pa > pc; Raw(x) = Paeth(x) + Prior(x-bpp)
+         mov cl, [esi + edx]  // load Prior(x-bpp) into cl
+         jmp dpthpaeth2
+dpthabc2:
+         // pa <= pc; Raw(x) = Paeth(x) + Raw(x-bpp)
+         mov cl, [edi + edx]  // load Raw(x-bpp) into cl
+dpthpaeth2:
+         inc ebx
+         inc edx
+         // Raw(x) = (Paeth(x) + Paeth_Predictor( a, b, c )) mod 256
+         add [edi + ebx - 1], cl
+         cmp ebx, FullLength
+         jb dpthlp2
+dpthend:
+         emms             // End MMX instructions; prep for possible FP instrs.
+   } // end _asm block
+}
+
+// Optimized code for PNG Sub filter decoder
+void /* PRIVATE */
+png_read_filter_row_mmx_sub(png_row_infop row_info, png_bytep row)
+{
+  // These variables are declared
+  // here to ensure alignment on 8-byte boundaries.
+  union uAll ActiveMask, ShiftBpp, ShiftRem;
+
+   //int test;
+   int bpp;
+   png_uint_32 FullLength;
+   png_uint_32 MMXLength;
+   int diff;
+
+   bpp = (row_info->pixel_depth + 7) >> 3; // Get # bytes per pixel
+   FullLength  = row_info->rowbytes - bpp; // # of bytes to filter
+   _asm {
+        mov edi, row
+        mov esi, edi               // lp = row
+        add edi, bpp               // rp = row + bpp
+        xor eax, eax
+        // get # of bytes to alignment
+        mov diff, edi               // take start of row
+        add diff, 0xf               // add 7 + 8 to incr past
+                                        // alignment boundary
+        xor ebx, ebx
+        and diff, 0xfffffff8        // mask to alignment boundary
+        sub diff, edi               // subtract from start ==> value
+                                        //  ebx at alignment
+        jz dsubgo
+        // fix alignment
+dsublp1:
+        mov al, [esi+ebx]
+        add [edi+ebx], al
+        inc ebx
+        cmp ebx, diff
+        jb dsublp1
+dsubgo:
+        mov ecx, FullLength
+        mov edx, ecx
+        sub edx, ebx                  // subtract alignment fix
+        and edx, 0x00000007           // calc bytes over mult of 8
+        sub ecx, edx                  // drop over bytes from length
+        mov MMXLength, ecx
+   } // end _asm block
+
+   // Now do the math for the rest of the row
+   switch ( bpp )
+   {
+        case 3:
+        {
+         ActiveMask.use  = 0x0000ffffff000000;
+         ShiftBpp.use = 24;       // == 3 * 8
+         ShiftRem.use  = 40;      // == 64 - 24
+         _asm {
+            mov edi, row
+            movq mm7, ActiveMask  // Load ActiveMask for 2nd active byte group
+            mov esi, edi              // lp = row
+            add edi, bpp          // rp = row + bpp
+            movq mm6, mm7
+            mov ebx, diff
+            psllq mm6, ShiftBpp   // Move mask in mm6 to cover 3rd active
+                                  // byte group
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm1, [edi+ebx-8]
+dsub3lp:
+            psrlq mm1, ShiftRem   // Shift data for adding 1st bpp bytes
+                          // no need for mask; shift clears inactive bytes
+            // Add 1st active group
+            movq mm0, [edi+ebx]
+            paddb mm0, mm1
+            // Add 2nd active group
+            movq mm1, mm0         // mov updated Raws to mm1
+            psllq mm1, ShiftBpp   // shift data to position correctly
+            pand mm1, mm7         // mask to use only 2nd active group
+            paddb mm0, mm1
+            // Add 3rd active group
+            movq mm1, mm0         // mov updated Raws to mm1
+            psllq mm1, ShiftBpp   // shift data to position correctly
+            pand mm1, mm6         // mask to use only 3rd active group
+            add ebx, 8
+            paddb mm0, mm1
+            cmp ebx, MMXLength
+            movq [edi+ebx-8], mm0     // Write updated Raws back to array
+            // Prep for doing 1st add at top of loop
+            movq mm1, mm0
+            jb dsub3lp
+         } // end _asm block
+      }
+      break;
+
+      case 1:
+      {
+         // Placed here just in case this is a duplicate of the
+         // non-MMX code for the SUB filter in png_read_filter_row below
+         //
+         //         png_bytep rp;
+         //         png_bytep lp;
+         //         png_uint_32 i;
+         //         bpp = (row_info->pixel_depth + 7) >> 3;
+         //         for (i = (png_uint_32)bpp, rp = row + bpp, lp = row;
+         //            i < row_info->rowbytes; i++, rp++, lp++)
+         //      {
+         //            *rp = (png_byte)(((int)(*rp) + (int)(*lp)) & 0xff);
+         //      }
+         _asm {
+            mov ebx, diff
+            mov edi, row
+            cmp ebx, FullLength
+            jnb dsub1end
+            mov esi, edi          // lp = row
+            xor eax, eax
+            add edi, bpp      // rp = row + bpp
+dsub1lp:
+            mov al, [esi+ebx]
+            add [edi+ebx], al
+            inc ebx
+            cmp ebx, FullLength
+            jb dsub1lp
+dsub1end:
+         } // end _asm block
+      }
+      return;
+
+      case 6:
+      case 7:
+      case 4:
+      case 5:
+      {
+         ShiftBpp.use = bpp << 3;
+         ShiftRem.use = 64 - ShiftBpp.use;
+         _asm {
+            mov edi, row
+            mov ebx, diff
+            mov esi, edi               // lp = row
+            add edi, bpp           // rp = row + bpp
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm1, [edi+ebx-8]
+dsub4lp:
+            psrlq mm1, ShiftRem // Shift data for adding 1st bpp bytes
+                          // no need for mask; shift clears inactive bytes
+            movq mm0, [edi+ebx]
+            paddb mm0, mm1
+            // Add 2nd active group
+            movq mm1, mm0          // mov updated Raws to mm1
+            psllq mm1, ShiftBpp    // shift data to position correctly
+                                   // there is no need for any mask
+                                   // since shift clears inactive bits/bytes
+            add ebx, 8
+            paddb mm0, mm1
+            cmp ebx, MMXLength
+            movq [edi+ebx-8], mm0
+            movq mm1, mm0          // Prep for doing 1st add at top of loop
+            jb dsub4lp
+         } // end _asm block
+      }
+      break;
+
+      case 2:
+      {
+         ActiveMask.use  = 0x00000000ffff0000;
+         ShiftBpp.use = 16;       // == 2 * 8
+         ShiftRem.use = 48;       // == 64 - 16
+         _asm {
+            movq mm7, ActiveMask  // Load ActiveMask for 2nd active byte group
+            mov ebx, diff
+            movq mm6, mm7
+            mov edi, row
+            psllq mm6, ShiftBpp     // Move mask in mm6 to cover 3rd active
+                                    //  byte group
+            mov esi, edi            // lp = row
+            movq mm5, mm6
+            add edi, bpp            // rp = row + bpp
+            psllq mm5, ShiftBpp     // Move mask in mm5 to cover 4th active
+                                    //  byte group
+            // PRIME the pump (load the first Raw(x-bpp) data set
+            movq mm1, [edi+ebx-8]
+dsub2lp:
+            // Add 1st active group
+            psrlq mm1, ShiftRem     // Shift data for adding 1st bpp bytes
+                                    // no need for mask; shift clears inactive
+                                    //  bytes
+            movq mm0, [edi+ebx]
+            paddb mm0, mm1
+            // Add 2nd active group
+            movq mm1, mm0           // mov updated Raws to mm1
+            psllq mm1, ShiftBpp     // shift data to position correctly
+            pand mm1, mm7           // mask to use only 2nd active group
+            paddb mm0, mm1
+            // Add 3rd active group
+            movq mm1, mm0           // mov updated Raws to mm1
+            psllq mm1, ShiftBpp     // shift data to position correctly
+            pand mm1, mm6           // mask to use only 3rd active group
+            paddb mm0, mm1
+            // Add 4th active group
+            movq mm1, mm0           // mov updated Raws to mm1
+            psllq mm1, ShiftBpp     // shift data to position correctly
+            pand mm1, mm5           // mask to use only 4th active group
+            add ebx, 8
+            paddb mm0, mm1
+            cmp ebx, MMXLength
+            movq [edi+ebx-8], mm0   // Write updated Raws back to array
+            movq mm1, mm0           // Prep for doing 1st add at top of loop
+            jb dsub2lp
+         } // end _asm block
+      }
+      break;
+      case 8:
+      {
+         _asm {
+            mov edi, row
+            mov ebx, diff
+            mov esi, edi            // lp = row
+            add edi, bpp            // rp = row + bpp
+            mov ecx, MMXLength
+            movq mm7, [edi+ebx-8]   // PRIME the pump (load the first
+                                    // Raw(x-bpp) data set
+            and ecx, 0x0000003f     // calc bytes over mult of 64
+dsub8lp:
+            movq mm0, [edi+ebx]     // Load Sub(x) for 1st 8 bytes
+            paddb mm0, mm7
+            movq mm1, [edi+ebx+8]   // Load Sub(x) for 2nd 8 bytes
+            movq [edi+ebx], mm0    // Write Raw(x) for 1st 8 bytes
+                                   // Now mm0 will be used as Raw(x-bpp) for
+                                   // the 2nd group of 8 bytes.  This will be
+                                   // repeated for each group of 8 bytes with
+                                   // the 8th group being used as the Raw(x-bpp)
+                                   // for the 1st group of the next loop.
+            paddb mm1, mm0
+            movq mm2, [edi+ebx+16]  // Load Sub(x) for 3rd 8 bytes
+            movq [edi+ebx+8], mm1   // Write Raw(x) for 2nd 8 bytes
+            paddb mm2, mm1
+            movq mm3, [edi+ebx+24]  // Load Sub(x) for 4th 8 bytes
+            movq [edi+ebx+16], mm2  // Write Raw(x) for 3rd 8 bytes
+            paddb mm3, mm2
+            movq mm4, [edi+ebx+32]  // Load Sub(x) for 5th 8 bytes
+            movq [edi+ebx+24], mm3  // Write Raw(x) for 4th 8 bytes
+            paddb mm4, mm3
+            movq mm5, [edi+ebx+40]  // Load Sub(x) for 6th 8 bytes
+            movq [edi+ebx+32], mm4  // Write Raw(x) for 5th 8 bytes
+            paddb mm5, mm4
+            movq mm6, [edi+ebx+48]  // Load Sub(x) for 7th 8 bytes
+            movq [edi+ebx+40], mm5  // Write Raw(x) for 6th 8 bytes
+            paddb mm6, mm5
+            movq mm7, [edi+ebx+56]  // Load Sub(x) for 8th 8 bytes
+            movq [edi+ebx+48], mm6  // Write Raw(x) for 7th 8 bytes
+            add ebx, 64
+            paddb mm7, mm6
+            cmp ebx, ecx
+            movq [edi+ebx-8], mm7   // Write Raw(x) for 8th 8 bytes
+            jb dsub8lp
+            cmp ebx, MMXLength
+            jnb dsub8lt8
+dsub8lpA:
+            movq mm0, [edi+ebx]
+            add ebx, 8
+            paddb mm0, mm7
+            cmp ebx, MMXLength
+            movq [edi+ebx-8], mm0   // use -8 to offset early add to ebx
+            movq mm7, mm0           // Move calculated Raw(x) data to mm1 to
+                                    // be the new Raw(x-bpp) for the next loop
+            jb dsub8lpA
+dsub8lt8:
+         } // end _asm block
+      }
+      break;
+
+      default:                // bpp greater than 8 bytes
+      {
+         _asm {
+            mov ebx, diff
+            mov edi, row
+            mov esi, edi           // lp = row
+            add edi, bpp           // rp = row + bpp
+dsubAlp:
+            movq mm0, [edi+ebx]
+            movq mm1, [esi+ebx]
+            add ebx, 8
+            paddb mm0, mm1
+            cmp ebx, MMXLength
+            movq [edi+ebx-8], mm0  // mov does not affect flags; -8 to offset
+                                   //  add ebx
+            jb dsubAlp
+         } // end _asm block
+      }
+      break;
+
+   } // end switch ( bpp )
+
+   _asm {
+        mov ebx, MMXLength
+        mov edi, row
+        cmp ebx, FullLength
+        jnb dsubend
+        mov esi, edi               // lp = row
+        xor eax, eax
+        add edi, bpp               // rp = row + bpp
+dsublp2:
+        mov al, [esi+ebx]
+        add [edi+ebx], al
+        inc ebx
+        cmp ebx, FullLength
+        jb dsublp2
+dsubend:
+        emms             // End MMX instructions; prep for possible FP instrs.
+   } // end _asm block
+}
+
+// Optimized code for PNG Up filter decoder
+void /* PRIVATE */
+png_read_filter_row_mmx_up(png_row_infop row_info, png_bytep row,
+   png_bytep prev_row)
+{
+   png_uint_32 len;
+   len  = row_info->rowbytes;       // # of bytes to filter
+   _asm {
+      mov edi, row
+      // get # of bytes to alignment
+      mov ecx, edi
+      xor ebx, ebx
+      add ecx, 0x7
+      xor eax, eax
+      and ecx, 0xfffffff8
+      mov esi, prev_row
+      sub ecx, edi
+      jz dupgo
+      // fix alignment
+duplp1:
+      mov al, [edi+ebx]
+      add al, [esi+ebx]
+      inc ebx
+      cmp ebx, ecx
+      mov [edi + ebx-1], al  // mov does not affect flags; -1 to offset inc ebx
+      jb duplp1
+dupgo:
+      mov ecx, len
+      mov edx, ecx
+      sub edx, ebx                  // subtract alignment fix
+      and edx, 0x0000003f           // calc bytes over mult of 64
+      sub ecx, edx                  // drop over bytes from length
+      // Unrolled loop - use all MMX registers and interleave to reduce
+      // number of branch instructions (loops) and reduce partial stalls
+duploop:
+      movq mm1, [esi+ebx]
+      movq mm0, [edi+ebx]
+      movq mm3, [esi+ebx+8]
+      paddb mm0, mm1
+      movq mm2, [edi+ebx+8]
+      movq [edi+ebx], mm0
+      paddb mm2, mm3
+      movq mm5, [esi+ebx+16]
+      movq [edi+ebx+8], mm2
+      movq mm4, [edi+ebx+16]
+      movq mm7, [esi+ebx+24]
+      paddb mm4, mm5
+      movq mm6, [edi+ebx+24]
+      movq [edi+ebx+16], mm4
+      paddb mm6, mm7
+      movq mm1, [esi+ebx+32]
+      movq [edi+ebx+24], mm6
+      movq mm0, [edi+ebx+32]
+      movq mm3, [esi+ebx+40]
+      paddb mm0, mm1
+      movq mm2, [edi+ebx+40]
+      movq [edi+ebx+32], mm0
+      paddb mm2, mm3
+      movq mm5, [esi+ebx+48]
+      movq [edi+ebx+40], mm2
+      movq mm4, [edi+ebx+48]
+      movq mm7, [esi+ebx+56]
+      paddb mm4, mm5
+      movq mm6, [edi+ebx+56]
+      movq [edi+ebx+48], mm4
+      add ebx, 64
+      paddb mm6, mm7
+      cmp ebx, ecx
+      movq [edi+ebx-8], mm6 // (+56)movq does not affect flags;
+                                     // -8 to offset add ebx
+      jb duploop
+
+      cmp edx, 0                     // Test for bytes over mult of 64
+      jz dupend
+
+
+      // 2 lines added by lcreeve at netins.net
+      // (mail 11 Jul 98 in png-implement list)
+      cmp edx, 8 //test for less than 8 bytes
+      jb duplt8
+
+
+      add ecx, edx
+      and edx, 0x00000007           // calc bytes over mult of 8
+      sub ecx, edx                  // drop over bytes from length
+      jz duplt8
+      // Loop using MMX registers mm0 & mm1 to update 8 bytes simultaneously
+duplpA:
+      movq mm1, [esi+ebx]
+      movq mm0, [edi+ebx]
+      add ebx, 8
+      paddb mm0, mm1
+      cmp ebx, ecx
+      movq [edi+ebx-8], mm0 // movq does not affect flags; -8 to offset add ebx
+      jb duplpA
+      cmp edx, 0            // Test for bytes over mult of 8
+      jz dupend
+duplt8:
+      xor eax, eax
+      add ecx, edx          // move over byte count into counter
+      // Loop using x86 registers to update remaining bytes
+duplp2:
+      mov al, [edi + ebx]
+      add al, [esi + ebx]
+      inc ebx
+      cmp ebx, ecx
+      mov [edi + ebx-1], al // mov does not affect flags; -1 to offset inc ebx
+      jb duplp2
+dupend:
+      // Conversion of filtered row completed
+      emms          // End MMX instructions; prep for possible FP instrs.
+   } // end _asm block
+}
+
+
+// Optimized png_read_filter_row routines
+void /* PRIVATE */
+png_read_filter_row(png_structp png_ptr, png_row_infop row_info, png_bytep
+   row, png_bytep prev_row, int filter)
+{
+#ifdef PNG_DEBUG
+   char filnm[10];
+#endif
+
+   if (mmx_supported == 2) {
+#if !defined(PNG_1_0_X)
+       /* this should have happened in png_init_mmx_flags() already */
+       png_warning(png_ptr, "asm_flags may not have been initialized");
+#endif
+       png_mmx_support();
+   }
+
+#ifdef PNG_DEBUG
+   png_debug(1, "in png_read_filter_row\n");
+   switch (filter)
+   {
+      case 0: png_snprintf(filnm, 10, "none");
+         break;
+#if !defined(PNG_1_0_X)
+      case 1: png_snprintf(filnm, 10, "sub-%s",
+        (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_SUB)? "MMX" : "x86");
+         break;
+      case 2: png_snprintf(filnm, 10, "up-%s",
+        (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_UP)? "MMX" : "x86");
+         break;
+      case 3: png_snprintf(filnm, 10, "avg-%s",
+        (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_AVG)? "MMX" : "x86");
+         break;
+      case 4: png_snprintf(filnm, 10, "Paeth-%s",
+        (png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_PAETH)? "MMX":"x86");
+         break;
+#else
+      case 1: png_snprintf(filnm, 10, "sub");
+         break;
+      case 2: png_snprintf(filnm, 10, "up");
+         break;
+      case 3: png_snprintf(filnm, 10, "avg");
+         break;
+      case 4: png_snprintf(filnm, 10, "Paeth");
+         break;
+#endif
+      default: png_snprintf(filnm, 10, "unknw");
+         break;
+   }
+   png_debug2(0,"row=%5d, %s, ", png_ptr->row_number, filnm);
+   png_debug2(0, "pd=%2d, b=%d, ", (int)row_info->pixel_depth,
+      (int)((row_info->pixel_depth + 7) >> 3));
+   png_debug1(0,"len=%8d, ", row_info->rowbytes);
+#endif /* PNG_DEBUG */
+
+   switch (filter)
+   {
+      case PNG_FILTER_VALUE_NONE:
+         break;
+
+      case PNG_FILTER_VALUE_SUB:
+      {
+#if !defined(PNG_1_0_X)
+         if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_SUB) &&
+             (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+             (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+         if (mmx_supported)
+#endif
+         {
+            png_read_filter_row_mmx_sub(row_info, row);
+         }
+         else
+         {
+            png_uint_32 i;
+            png_uint_32 istop = row_info->rowbytes;
+            png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+            png_bytep rp = row + bpp;
+            png_bytep lp = row;
+
+            for (i = bpp; i < istop; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) + (int)(*lp++)) & 0xff);
+               rp++;
+            }
+         }
+         break;
+      }
+
+      case PNG_FILTER_VALUE_UP:
+      {
+#if !defined(PNG_1_0_X)
+         if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_UP) &&
+             (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+             (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+         if (mmx_supported)
+#endif
+         {
+            png_read_filter_row_mmx_up(row_info, row, prev_row);
+         }
+         else
+         {
+            png_uint_32 i;
+            png_uint_32 istop = row_info->rowbytes;
+            png_bytep rp = row;
+            png_bytep pp = prev_row;
+
+            for (i = 0; i < istop; ++i)
+            {
+               *rp = (png_byte)(((int)(*rp) + (int)(*pp++)) & 0xff);
+               rp++;
+            }
+         }
+         break;
+      }
+
+      case PNG_FILTER_VALUE_AVG:
+      {
+#if !defined(PNG_1_0_X)
+         if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_AVG) &&
+             (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+             (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+         if (mmx_supported)
+#endif
+         {
+            png_read_filter_row_mmx_avg(row_info, row, prev_row);
+         }
+         else
+         {
+            png_uint_32 i;
+            png_bytep rp = row;
+            png_bytep pp = prev_row;
+            png_bytep lp = row;
+            png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+            png_uint_32 istop = row_info->rowbytes - bpp;
+
+            for (i = 0; i < bpp; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) +
+                  ((int)(*pp++) >> 1)) & 0xff);
+               rp++;
+            }
+
+            for (i = 0; i < istop; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) +
+                  ((int)(*pp++ + *lp++) >> 1)) & 0xff);
+               rp++;
+            }
+         }
+         break;
+      }
+
+      case PNG_FILTER_VALUE_PAETH:
+      {
+#if !defined(PNG_1_0_X)
+         if ((png_ptr->asm_flags & PNG_ASM_FLAG_MMX_READ_FILTER_PAETH) &&
+             (row_info->pixel_depth >= png_ptr->mmx_bitdepth_threshold) &&
+             (row_info->rowbytes >= png_ptr->mmx_rowbytes_threshold))
+#else
+         if (mmx_supported)
+#endif
+         {
+            png_read_filter_row_mmx_paeth(row_info, row, prev_row);
+         }
+         else
+         {
+            png_uint_32 i;
+            png_bytep rp = row;
+            png_bytep pp = prev_row;
+            png_bytep lp = row;
+            png_bytep cp = prev_row;
+            png_uint_32 bpp = (row_info->pixel_depth + 7) >> 3;
+            png_uint_32 istop=row_info->rowbytes - bpp;
+
+            for (i = 0; i < bpp; i++)
+            {
+               *rp = (png_byte)(((int)(*rp) + (int)(*pp++)) & 0xff);
+               rp++;
+            }
+
+            for (i = 0; i < istop; i++)   // use leftover rp,pp
+            {
+               int a, b, c, pa, pb, pc, p;
+
+               a = *lp++;
+               b = *pp++;
+               c = *cp++;
+
+               p = b - c;
+               pc = a - c;
+
+#ifdef PNG_USE_ABS
+               pa = abs(p);
+               pb = abs(pc);
+               pc = abs(p + pc);
+#else
+               pa = p < 0 ? -p : p;
+               pb = pc < 0 ? -pc : pc;
+               pc = (p + pc) < 0 ? -(p + pc) : p + pc;
+#endif
+
+               /*
+                  if (pa <= pb && pa <= pc)
+                     p = a;
+                  else if (pb <= pc)
+                     p = b;
+                  else
+                     p = c;
+                */
+
+               p = (pa <= pb && pa <=pc) ? a : (pb <= pc) ? b : c;
+
+               *rp = (png_byte)(((int)(*rp) + p) & 0xff);
+               rp++;
+            }
+         }
+         break;
+      }
+
+      default:
+         png_warning(png_ptr, "Ignoring bad row filter type");
+         *row=0;
+         break;
+   }
+}
+
+#endif /* PNG_MMX_CODE_SUPPORTED && PNG_USE_PNGVCRD */
diff --git a/distrib/libpng-1.2.19/pngwio.c b/distrib/libpng-1.2.19/pngwio.c
new file mode 100644
index 0000000..371a4fa
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngwio.c
@@ -0,0 +1,234 @@
+
+/* pngwio.c - functions for data output
+ *
+ * Last changed in libpng 1.2.13 November 13, 2006
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2006 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ *
+ * This file provides a location for all output.  Users who need
+ * special handling are expected to write functions that have the same
+ * arguments as these and perform similar functions, but that possibly
+ * use different output methods.  Note that you shouldn't change these
+ * functions, but rather write replacement functions and then change
+ * them at run time with png_set_write_fn(...).
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+#ifdef PNG_WRITE_SUPPORTED
+
+/* Write the data to whatever output you are using.  The default routine
+   writes to a file pointer.  Note that this routine sometimes gets called
+   with very small lengths, so you should implement some kind of simple
+   buffering if you are using unbuffered writes.  This should never be asked
+   to write more than 64K on a 16 bit machine.  */
+
+void /* PRIVATE */
+png_write_data(png_structp png_ptr, png_bytep data, png_size_t length)
+{
+   if (png_ptr->write_data_fn != NULL )
+      (*(png_ptr->write_data_fn))(png_ptr, data, length);
+   else
+      png_error(png_ptr, "Call to NULL write function");
+}
+
+#if !defined(PNG_NO_STDIO)
+/* This is the function that does the actual writing of data.  If you are
+   not writing to a standard C stream, you should create a replacement
+   write_data function and use it at run time with png_set_write_fn(), rather
+   than changing the library. */
+#ifndef USE_FAR_KEYWORD
+void PNGAPI
+png_default_write_data(png_structp png_ptr, png_bytep data, png_size_t length)
+{
+   png_uint_32 check;
+
+   if(png_ptr == NULL) return;
+#if defined(_WIN32_WCE)
+   if ( !WriteFile((HANDLE)(png_ptr->io_ptr), data, length, &check, NULL) )
+      check = 0;
+#else
+   check = fwrite(data, 1, length, (png_FILE_p)(png_ptr->io_ptr));
+#endif
+   if (check != length)
+      png_error(png_ptr, "Write Error");
+}
+#else
+/* this is the model-independent version. Since the standard I/O library
+   can't handle far buffers in the medium and small models, we have to copy
+   the data.
+*/
+
+#define NEAR_BUF_SIZE 1024
+#define MIN(a,b) (a <= b ? a : b)
+
+void PNGAPI
+png_default_write_data(png_structp png_ptr, png_bytep data, png_size_t length)
+{
+   png_uint_32 check;
+   png_byte *near_data;  /* Needs to be "png_byte *" instead of "png_bytep" */
+   png_FILE_p io_ptr;
+
+   if(png_ptr == NULL) return;
+   /* Check if data really is near. If so, use usual code. */
+   near_data = (png_byte *)CVT_PTR_NOCHECK(data);
+   io_ptr = (png_FILE_p)CVT_PTR(png_ptr->io_ptr);
+   if ((png_bytep)near_data == data)
+   {
+#if defined(_WIN32_WCE)
+      if ( !WriteFile(io_ptr, near_data, length, &check, NULL) )
+         check = 0;
+#else
+      check = fwrite(near_data, 1, length, io_ptr);
+#endif
+   }
+   else
+   {
+      png_byte buf[NEAR_BUF_SIZE];
+      png_size_t written, remaining, err;
+      check = 0;
+      remaining = length;
+      do
+      {
+         written = MIN(NEAR_BUF_SIZE, remaining);
+         png_memcpy(buf, data, written); /* copy far buffer to near buffer */
+#if defined(_WIN32_WCE)
+         if ( !WriteFile(io_ptr, buf, written, &err, NULL) )
+            err = 0;
+#else
+         err = fwrite(buf, 1, written, io_ptr);
+#endif
+         if (err != written)
+            break;
+         else
+            check += err;
+         data += written;
+         remaining -= written;
+      }
+      while (remaining != 0);
+   }
+   if (check != length)
+      png_error(png_ptr, "Write Error");
+}
+
+#endif
+#endif
+
+/* This function is called to output any data pending writing (normally
+   to disk).  After png_flush is called, there should be no data pending
+   writing in any buffers. */
+#if defined(PNG_WRITE_FLUSH_SUPPORTED)
+void /* PRIVATE */
+png_flush(png_structp png_ptr)
+{
+   if (png_ptr->output_flush_fn != NULL)
+      (*(png_ptr->output_flush_fn))(png_ptr);
+}
+
+#if !defined(PNG_NO_STDIO)
+void PNGAPI
+png_default_flush(png_structp png_ptr)
+{
+#if !defined(_WIN32_WCE)
+   png_FILE_p io_ptr;
+#endif
+   if(png_ptr == NULL) return;
+#if !defined(_WIN32_WCE)
+   io_ptr = (png_FILE_p)CVT_PTR((png_ptr->io_ptr));
+   if (io_ptr != NULL)
+      fflush(io_ptr);
+#endif
+}
+#endif
+#endif
+
+/* This function allows the application to supply new output functions for
+   libpng if standard C streams aren't being used.
+
+   This function takes as its arguments:
+   png_ptr       - pointer to a png output data structure
+   io_ptr        - pointer to user supplied structure containing info about
+                   the output functions.  May be NULL.
+   write_data_fn - pointer to a new output function that takes as its
+                   arguments a pointer to a png_struct, a pointer to
+                   data to be written, and a 32-bit unsigned int that is
+                   the number of bytes to be written.  The new write
+                   function should call png_error(png_ptr, "Error msg")
+                   to exit and output any fatal error messages.
+   flush_data_fn - pointer to a new flush function that takes as its
+                   arguments a pointer to a png_struct.  After a call to
+                   the flush function, there should be no data in any buffers
+                   or pending transmission.  If the output method doesn't do
+                   any buffering of ouput, a function prototype must still be
+                   supplied although it doesn't have to do anything.  If
+                   PNG_WRITE_FLUSH_SUPPORTED is not defined at libpng compile
+                   time, output_flush_fn will be ignored, although it must be
+                   supplied for compatibility. */
+void PNGAPI
+png_set_write_fn(png_structp png_ptr, png_voidp io_ptr,
+   png_rw_ptr write_data_fn, png_flush_ptr output_flush_fn)
+{
+   if(png_ptr == NULL) return;
+   png_ptr->io_ptr = io_ptr;
+
+#if !defined(PNG_NO_STDIO)
+   if (write_data_fn != NULL)
+      png_ptr->write_data_fn = write_data_fn;
+   else
+      png_ptr->write_data_fn = png_default_write_data;
+#else
+   png_ptr->write_data_fn = write_data_fn;
+#endif
+
+#if defined(PNG_WRITE_FLUSH_SUPPORTED)
+#if !defined(PNG_NO_STDIO)
+   if (output_flush_fn != NULL)
+      png_ptr->output_flush_fn = output_flush_fn;
+   else
+      png_ptr->output_flush_fn = png_default_flush;
+#else
+   png_ptr->output_flush_fn = output_flush_fn;
+#endif
+#endif /* PNG_WRITE_FLUSH_SUPPORTED */
+
+   /* It is an error to read while writing a png file */
+   if (png_ptr->read_data_fn != NULL)
+   {
+      png_ptr->read_data_fn = NULL;
+      png_warning(png_ptr,
+         "Attempted to set both read_data_fn and write_data_fn in");
+      png_warning(png_ptr,
+         "the same structure.  Resetting read_data_fn to NULL.");
+   }
+}
+
+#if defined(USE_FAR_KEYWORD)
+#if defined(_MSC_VER)
+void *png_far_to_near(png_structp png_ptr,png_voidp ptr, int check)
+{
+   void *near_ptr;
+   void FAR *far_ptr;
+   FP_OFF(near_ptr) = FP_OFF(ptr);
+   far_ptr = (void FAR *)near_ptr;
+   if(check != 0)
+      if(FP_SEG(ptr) != FP_SEG(far_ptr))
+         png_error(png_ptr,"segment lost in conversion");
+   return(near_ptr);
+}
+#  else
+void *png_far_to_near(png_structp png_ptr,png_voidp ptr, int check)
+{
+   void *near_ptr;
+   void FAR *far_ptr;
+   near_ptr = (void FAR *)ptr;
+   far_ptr = (void FAR *)near_ptr;
+   if(check != 0)
+      if(far_ptr != ptr)
+         png_error(png_ptr,"segment lost in conversion");
+   return(near_ptr);
+}
+#   endif
+#   endif
+#endif /* PNG_WRITE_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngwrite.c b/distrib/libpng-1.2.19/pngwrite.c
new file mode 100644
index 0000000..8d5b98a
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngwrite.c
@@ -0,0 +1,1530 @@
+
+/* pngwrite.c - general routines to write a PNG file
+ *
+ * Last changed in libpng 1.2.15 January 5, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ */
+
+/* get internal access to png.h */
+#define PNG_INTERNAL
+#include "png.h"
+#ifdef PNG_WRITE_SUPPORTED
+
+/* Writes all the PNG information.  This is the suggested way to use the
+ * library.  If you have a new chunk to add, make a function to write it,
+ * and put it in the correct location here.  If you want the chunk written
+ * after the image data, put it in png_write_end().  I strongly encourage
+ * you to supply a PNG_INFO_ flag, and check info_ptr->valid before writing
+ * the chunk, as that will keep the code from breaking if you want to just
+ * write a plain PNG file.  If you have long comments, I suggest writing
+ * them in png_write_end(), and compressing them.
+ */
+void PNGAPI
+png_write_info_before_PLTE(png_structp png_ptr, png_infop info_ptr)
+{
+   png_debug(1, "in png_write_info_before_PLTE\n");
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+   if (!(png_ptr->mode & PNG_WROTE_INFO_BEFORE_PLTE))
+   {
+   png_write_sig(png_ptr); /* write PNG signature */
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+   if((png_ptr->mode&PNG_HAVE_PNG_SIGNATURE)&&(png_ptr->mng_features_permitted))
+   {
+      png_warning(png_ptr,"MNG features are not allowed in a PNG datastream");
+      png_ptr->mng_features_permitted=0;
+   }
+#endif
+   /* write IHDR information. */
+   png_write_IHDR(png_ptr, info_ptr->width, info_ptr->height,
+      info_ptr->bit_depth, info_ptr->color_type, info_ptr->compression_type,
+      info_ptr->filter_type,
+#if defined(PNG_WRITE_INTERLACING_SUPPORTED)
+      info_ptr->interlace_type);
+#else
+      0);
+#endif
+   /* the rest of these check to see if the valid field has the appropriate
+      flag set, and if it does, writes the chunk. */
+#if defined(PNG_WRITE_gAMA_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_gAMA)
+   {
+#  ifdef PNG_FLOATING_POINT_SUPPORTED
+      png_write_gAMA(png_ptr, info_ptr->gamma);
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+      png_write_gAMA_fixed(png_ptr, info_ptr->int_gamma);
+#  endif
+#endif
+   }
+#endif
+#if defined(PNG_WRITE_sRGB_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_sRGB)
+      png_write_sRGB(png_ptr, (int)info_ptr->srgb_intent);
+#endif
+#if defined(PNG_WRITE_iCCP_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_iCCP)
+      png_write_iCCP(png_ptr, info_ptr->iccp_name, PNG_COMPRESSION_TYPE_BASE,
+                     info_ptr->iccp_profile, (int)info_ptr->iccp_proflen);
+#endif
+#if defined(PNG_WRITE_sBIT_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_sBIT)
+      png_write_sBIT(png_ptr, &(info_ptr->sig_bit), info_ptr->color_type);
+#endif
+#if defined(PNG_WRITE_cHRM_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_cHRM)
+   {
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+      png_write_cHRM(png_ptr,
+         info_ptr->x_white, info_ptr->y_white,
+         info_ptr->x_red, info_ptr->y_red,
+         info_ptr->x_green, info_ptr->y_green,
+         info_ptr->x_blue, info_ptr->y_blue);
+#else
+#  ifdef PNG_FIXED_POINT_SUPPORTED
+      png_write_cHRM_fixed(png_ptr,
+         info_ptr->int_x_white, info_ptr->int_y_white,
+         info_ptr->int_x_red, info_ptr->int_y_red,
+         info_ptr->int_x_green, info_ptr->int_y_green,
+         info_ptr->int_x_blue, info_ptr->int_y_blue);
+#  endif
+#endif
+   }
+#endif
+#if defined(PNG_WRITE_UNKNOWN_CHUNKS_SUPPORTED)
+   if (info_ptr->unknown_chunks_num)
+   {
+       png_unknown_chunk *up;
+
+       png_debug(5, "writing extra chunks\n");
+
+       for (up = info_ptr->unknown_chunks;
+            up < info_ptr->unknown_chunks + info_ptr->unknown_chunks_num;
+            up++)
+       {
+         int keep=png_handle_as_unknown(png_ptr, up->name);
+         if (keep != PNG_HANDLE_CHUNK_NEVER &&
+            up->location && !(up->location & PNG_HAVE_PLTE) &&
+            !(up->location & PNG_HAVE_IDAT) &&
+            ((up->name[3] & 0x20) || keep == PNG_HANDLE_CHUNK_ALWAYS ||
+            (png_ptr->flags & PNG_FLAG_KEEP_UNSAFE_CHUNKS)))
+         {
+            png_write_chunk(png_ptr, up->name, up->data, up->size);
+         }
+       }
+   }
+#endif
+      png_ptr->mode |= PNG_WROTE_INFO_BEFORE_PLTE;
+   }
+}
+
+void PNGAPI
+png_write_info(png_structp png_ptr, png_infop info_ptr)
+{
+#if defined(PNG_WRITE_TEXT_SUPPORTED) || defined(PNG_WRITE_sPLT_SUPPORTED)
+   int i;
+#endif
+
+   png_debug(1, "in png_write_info\n");
+
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+
+   png_write_info_before_PLTE(png_ptr, info_ptr);
+
+   if (info_ptr->valid & PNG_INFO_PLTE)
+      png_write_PLTE(png_ptr, info_ptr->palette,
+         (png_uint_32)info_ptr->num_palette);
+   else if (info_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+      png_error(png_ptr, "Valid palette required for paletted images");
+
+#if defined(PNG_WRITE_tRNS_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_tRNS)
+      {
+#if defined(PNG_WRITE_INVERT_ALPHA_SUPPORTED)
+         /* invert the alpha channel (in tRNS) */
+         if ((png_ptr->transformations & PNG_INVERT_ALPHA) &&
+            info_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+         {
+            int j;
+            for (j=0; j<(int)info_ptr->num_trans; j++)
+               info_ptr->trans[j] = (png_byte)(255 - info_ptr->trans[j]);
+         }
+#endif
+      png_write_tRNS(png_ptr, info_ptr->trans, &(info_ptr->trans_values),
+         info_ptr->num_trans, info_ptr->color_type);
+      }
+#endif
+#if defined(PNG_WRITE_bKGD_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_bKGD)
+      png_write_bKGD(png_ptr, &(info_ptr->background), info_ptr->color_type);
+#endif
+#if defined(PNG_WRITE_hIST_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_hIST)
+      png_write_hIST(png_ptr, info_ptr->hist, info_ptr->num_palette);
+#endif
+#if defined(PNG_WRITE_oFFs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_oFFs)
+      png_write_oFFs(png_ptr, info_ptr->x_offset, info_ptr->y_offset,
+         info_ptr->offset_unit_type);
+#endif
+#if defined(PNG_WRITE_pCAL_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_pCAL)
+      png_write_pCAL(png_ptr, info_ptr->pcal_purpose, info_ptr->pcal_X0,
+         info_ptr->pcal_X1, info_ptr->pcal_type, info_ptr->pcal_nparams,
+         info_ptr->pcal_units, info_ptr->pcal_params);
+#endif
+#if defined(PNG_WRITE_sCAL_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_sCAL)
+#if defined(PNG_FLOATING_POINT_SUPPORTED) && !defined(PNG_NO_STDIO)
+      png_write_sCAL(png_ptr, (int)info_ptr->scal_unit,
+          info_ptr->scal_pixel_width, info_ptr->scal_pixel_height);
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+      png_write_sCAL_s(png_ptr, (int)info_ptr->scal_unit,
+          info_ptr->scal_s_width, info_ptr->scal_s_height);
+#else
+      png_warning(png_ptr,
+          "png_write_sCAL not supported; sCAL chunk not written.");
+#endif
+#endif
+#endif
+#if defined(PNG_WRITE_pHYs_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_pHYs)
+      png_write_pHYs(png_ptr, info_ptr->x_pixels_per_unit,
+         info_ptr->y_pixels_per_unit, info_ptr->phys_unit_type);
+#endif
+#if defined(PNG_WRITE_tIME_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_tIME)
+   {
+      png_write_tIME(png_ptr, &(info_ptr->mod_time));
+      png_ptr->mode |= PNG_WROTE_tIME;
+   }
+#endif
+#if defined(PNG_WRITE_sPLT_SUPPORTED)
+   if (info_ptr->valid & PNG_INFO_sPLT)
+     for (i = 0; i < (int)info_ptr->splt_palettes_num; i++)
+       png_write_sPLT(png_ptr, info_ptr->splt_palettes + i);
+#endif
+#if defined(PNG_WRITE_TEXT_SUPPORTED)
+   /* Check to see if we need to write text chunks */
+   for (i = 0; i < info_ptr->num_text; i++)
+   {
+      png_debug2(2, "Writing header text chunk %d, type %d\n", i,
+         info_ptr->text[i].compression);
+      /* an internationalized chunk? */
+      if (info_ptr->text[i].compression > 0)
+      {
+#if defined(PNG_WRITE_iTXt_SUPPORTED)
+          /* write international chunk */
+          png_write_iTXt(png_ptr,
+                         info_ptr->text[i].compression,
+                         info_ptr->text[i].key,
+                         info_ptr->text[i].lang,
+                         info_ptr->text[i].lang_key,
+                         info_ptr->text[i].text);
+#else
+          png_warning(png_ptr, "Unable to write international text");
+#endif
+          /* Mark this chunk as written */
+          info_ptr->text[i].compression = PNG_TEXT_COMPRESSION_NONE_WR;
+      }
+      /* If we want a compressed text chunk */
+      else if (info_ptr->text[i].compression == PNG_TEXT_COMPRESSION_zTXt)
+      {
+#if defined(PNG_WRITE_zTXt_SUPPORTED)
+         /* write compressed chunk */
+         png_write_zTXt(png_ptr, info_ptr->text[i].key,
+            info_ptr->text[i].text, 0,
+            info_ptr->text[i].compression);
+#else
+         png_warning(png_ptr, "Unable to write compressed text");
+#endif
+         /* Mark this chunk as written */
+         info_ptr->text[i].compression = PNG_TEXT_COMPRESSION_zTXt_WR;
+      }
+      else if (info_ptr->text[i].compression == PNG_TEXT_COMPRESSION_NONE)
+      {
+#if defined(PNG_WRITE_tEXt_SUPPORTED)
+         /* write uncompressed chunk */
+         png_write_tEXt(png_ptr, info_ptr->text[i].key,
+                         info_ptr->text[i].text,
+                         0);
+#else
+         png_warning(png_ptr, "Unable to write uncompressed text");
+#endif
+         /* Mark this chunk as written */
+         info_ptr->text[i].compression = PNG_TEXT_COMPRESSION_NONE_WR;
+      }
+   }
+#endif
+#if defined(PNG_WRITE_UNKNOWN_CHUNKS_SUPPORTED)
+   if (info_ptr->unknown_chunks_num)
+   {
+       png_unknown_chunk *up;
+
+       png_debug(5, "writing extra chunks\n");
+
+       for (up = info_ptr->unknown_chunks;
+            up < info_ptr->unknown_chunks + info_ptr->unknown_chunks_num;
+            up++)
+       {
+         int keep=png_handle_as_unknown(png_ptr, up->name);
+         if (keep != PNG_HANDLE_CHUNK_NEVER &&
+            up->location && (up->location & PNG_HAVE_PLTE) &&
+            !(up->location & PNG_HAVE_IDAT) &&
+            ((up->name[3] & 0x20) || keep == PNG_HANDLE_CHUNK_ALWAYS ||
+            (png_ptr->flags & PNG_FLAG_KEEP_UNSAFE_CHUNKS)))
+         {
+            png_write_chunk(png_ptr, up->name, up->data, up->size);
+         }
+       }
+   }
+#endif
+}
+
+/* Writes the end of the PNG file.  If you don't want to write comments or
+ * time information, you can pass NULL for info.  If you already wrote these
+ * in png_write_info(), do not write them again here.  If you have long
+ * comments, I suggest writing them here, and compressing them.
+ */
+void PNGAPI
+png_write_end(png_structp png_ptr, png_infop info_ptr)
+{
+   png_debug(1, "in png_write_end\n");
+   if (png_ptr == NULL)
+      return;
+   if (!(png_ptr->mode & PNG_HAVE_IDAT))
+      png_error(png_ptr, "No IDATs written into file");
+
+   /* see if user wants us to write information chunks */
+   if (info_ptr != NULL)
+   {
+#if defined(PNG_WRITE_TEXT_SUPPORTED)
+      int i; /* local index variable */
+#endif
+#if defined(PNG_WRITE_tIME_SUPPORTED)
+      /* check to see if user has supplied a time chunk */
+      if ((info_ptr->valid & PNG_INFO_tIME) &&
+         !(png_ptr->mode & PNG_WROTE_tIME))
+         png_write_tIME(png_ptr, &(info_ptr->mod_time));
+#endif
+#if defined(PNG_WRITE_TEXT_SUPPORTED)
+      /* loop through comment chunks */
+      for (i = 0; i < info_ptr->num_text; i++)
+      {
+         png_debug2(2, "Writing trailer text chunk %d, type %d\n", i,
+            info_ptr->text[i].compression);
+         /* an internationalized chunk? */
+         if (info_ptr->text[i].compression > 0)
+         {
+#if defined(PNG_WRITE_iTXt_SUPPORTED)
+             /* write international chunk */
+             png_write_iTXt(png_ptr,
+                         info_ptr->text[i].compression,
+                         info_ptr->text[i].key,
+                         info_ptr->text[i].lang,
+                         info_ptr->text[i].lang_key,
+                         info_ptr->text[i].text);
+#else
+             png_warning(png_ptr, "Unable to write international text");
+#endif
+             /* Mark this chunk as written */
+             info_ptr->text[i].compression = PNG_TEXT_COMPRESSION_NONE_WR;
+         }
+         else if (info_ptr->text[i].compression >= PNG_TEXT_COMPRESSION_zTXt)
+         {
+#if defined(PNG_WRITE_zTXt_SUPPORTED)
+            /* write compressed chunk */
+            png_write_zTXt(png_ptr, info_ptr->text[i].key,
+               info_ptr->text[i].text, 0,
+               info_ptr->text[i].compression);
+#else
+            png_warning(png_ptr, "Unable to write compressed text");
+#endif
+            /* Mark this chunk as written */
+            info_ptr->text[i].compression = PNG_TEXT_COMPRESSION_zTXt_WR;
+         }
+         else if (info_ptr->text[i].compression == PNG_TEXT_COMPRESSION_NONE)
+         {
+#if defined(PNG_WRITE_tEXt_SUPPORTED)
+            /* write uncompressed chunk */
+            png_write_tEXt(png_ptr, info_ptr->text[i].key,
+               info_ptr->text[i].text, 0);
+#else
+            png_warning(png_ptr, "Unable to write uncompressed text");
+#endif
+
+            /* Mark this chunk as written */
+            info_ptr->text[i].compression = PNG_TEXT_COMPRESSION_NONE_WR;
+         }
+      }
+#endif
+#if defined(PNG_WRITE_UNKNOWN_CHUNKS_SUPPORTED)
+   if (info_ptr->unknown_chunks_num)
+   {
+       png_unknown_chunk *up;
+
+       png_debug(5, "writing extra chunks\n");
+
+       for (up = info_ptr->unknown_chunks;
+            up < info_ptr->unknown_chunks + info_ptr->unknown_chunks_num;
+            up++)
+       {
+         int keep=png_handle_as_unknown(png_ptr, up->name);
+         if (keep != PNG_HANDLE_CHUNK_NEVER &&
+            up->location && (up->location & PNG_AFTER_IDAT) &&
+            ((up->name[3] & 0x20) || keep == PNG_HANDLE_CHUNK_ALWAYS ||
+            (png_ptr->flags & PNG_FLAG_KEEP_UNSAFE_CHUNKS)))
+         {
+            png_write_chunk(png_ptr, up->name, up->data, up->size);
+         }
+       }
+   }
+#endif
+   }
+
+   png_ptr->mode |= PNG_AFTER_IDAT;
+
+   /* write end of PNG file */
+   png_write_IEND(png_ptr);
+}
+
+#if defined(PNG_WRITE_tIME_SUPPORTED)
+#if !defined(_WIN32_WCE)
+/* "time.h" functions are not supported on WindowsCE */
+void PNGAPI
+png_convert_from_struct_tm(png_timep ptime, struct tm FAR * ttime)
+{
+   png_debug(1, "in png_convert_from_struct_tm\n");
+   ptime->year = (png_uint_16)(1900 + ttime->tm_year);
+   ptime->month = (png_byte)(ttime->tm_mon + 1);
+   ptime->day = (png_byte)ttime->tm_mday;
+   ptime->hour = (png_byte)ttime->tm_hour;
+   ptime->minute = (png_byte)ttime->tm_min;
+   ptime->second = (png_byte)ttime->tm_sec;
+}
+
+void PNGAPI
+png_convert_from_time_t(png_timep ptime, time_t ttime)
+{
+   struct tm *tbuf;
+
+   png_debug(1, "in png_convert_from_time_t\n");
+   tbuf = gmtime(&ttime);
+   png_convert_from_struct_tm(ptime, tbuf);
+}
+#endif
+#endif
+
+/* Initialize png_ptr structure, and allocate any memory needed */
+png_structp PNGAPI
+png_create_write_struct(png_const_charp user_png_ver, png_voidp error_ptr,
+   png_error_ptr error_fn, png_error_ptr warn_fn)
+{
+#ifdef PNG_USER_MEM_SUPPORTED
+   return (png_create_write_struct_2(user_png_ver, error_ptr, error_fn,
+      warn_fn, png_voidp_NULL, png_malloc_ptr_NULL, png_free_ptr_NULL));
+}
+
+/* Alternate initialize png_ptr structure, and allocate any memory needed */
+png_structp PNGAPI
+png_create_write_struct_2(png_const_charp user_png_ver, png_voidp error_ptr,
+   png_error_ptr error_fn, png_error_ptr warn_fn, png_voidp mem_ptr,
+   png_malloc_ptr malloc_fn, png_free_ptr free_fn)
+{
+#endif /* PNG_USER_MEM_SUPPORTED */
+   png_structp png_ptr;
+#ifdef PNG_SETJMP_SUPPORTED
+#ifdef USE_FAR_KEYWORD
+   jmp_buf jmpbuf;
+#endif
+#endif
+   int i;
+   png_debug(1, "in png_create_write_struct\n");
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_ptr = (png_structp)png_create_struct_2(PNG_STRUCT_PNG,
+      (png_malloc_ptr)malloc_fn, (png_voidp)mem_ptr);
+#else
+   png_ptr = (png_structp)png_create_struct(PNG_STRUCT_PNG);
+#endif /* PNG_USER_MEM_SUPPORTED */
+   if (png_ptr == NULL)
+      return (NULL);
+
+#if !defined(PNG_1_0_X)
+#ifdef PNG_ASSEMBLER_CODE_SUPPORTED
+#ifdef PNG_MMX_CODE_SUPPORTED
+   png_init_mmx_flags(png_ptr);   /* 1.2.0 addition */
+#endif
+#endif
+#endif /* PNG_1_0_X */
+
+   /* added at libpng-1.2.6 */
+#ifdef PNG_SET_USER_LIMITS_SUPPORTED
+   png_ptr->user_width_max=PNG_USER_WIDTH_MAX;
+   png_ptr->user_height_max=PNG_USER_HEIGHT_MAX;
+#endif
+
+#ifdef PNG_SETJMP_SUPPORTED
+#ifdef USE_FAR_KEYWORD
+   if (setjmp(jmpbuf))
+#else
+   if (setjmp(png_ptr->jmpbuf))
+#endif
+   {
+      png_free(png_ptr, png_ptr->zbuf);
+      png_ptr->zbuf=NULL;
+      png_destroy_struct(png_ptr);
+      return (NULL);
+   }
+#ifdef USE_FAR_KEYWORD
+   png_memcpy(png_ptr->jmpbuf,jmpbuf,png_sizeof(jmp_buf));
+#endif
+#endif
+
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_set_mem_fn(png_ptr, mem_ptr, malloc_fn, free_fn);
+#endif /* PNG_USER_MEM_SUPPORTED */
+   png_set_error_fn(png_ptr, error_ptr, error_fn, warn_fn);
+
+   i=0;
+   do
+   {
+     if(user_png_ver[i] != png_libpng_ver[i])
+        png_ptr->flags |= PNG_FLAG_LIBRARY_MISMATCH;
+   } while (png_libpng_ver[i++]);
+
+   if (png_ptr->flags & PNG_FLAG_LIBRARY_MISMATCH)
+   {
+     /* Libpng 0.90 and later are binary incompatible with libpng 0.89, so
+      * we must recompile any applications that use any older library version.
+      * For versions after libpng 1.0, we will be compatible, so we need
+      * only check the first digit.
+      */
+     if (user_png_ver == NULL || user_png_ver[0] != png_libpng_ver[0] ||
+         (user_png_ver[0] == '1' && user_png_ver[2] != png_libpng_ver[2]) ||
+         (user_png_ver[0] == '0' && user_png_ver[2] < '9'))
+     {
+#if !defined(PNG_NO_STDIO) && !defined(_WIN32_WCE)
+        char msg[80];
+        if (user_png_ver)
+        {
+          png_snprintf(msg, 80,
+             "Application was compiled with png.h from libpng-%.20s",
+             user_png_ver);
+          png_warning(png_ptr, msg);
+        }
+        png_snprintf(msg, 80,
+           "Application  is  running with png.c from libpng-%.20s",
+           png_libpng_ver);
+        png_warning(png_ptr, msg);
+#endif
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+        png_ptr->flags=0;
+#endif
+        png_error(png_ptr,
+           "Incompatible libpng version in application and library");
+     }
+   }
+
+   /* initialize zbuf - compression buffer */
+   png_ptr->zbuf_size = PNG_ZBUF_SIZE;
+   png_ptr->zbuf = (png_bytep)png_malloc(png_ptr,
+      (png_uint_32)png_ptr->zbuf_size);
+
+   png_set_write_fn(png_ptr, png_voidp_NULL, png_rw_ptr_NULL,
+      png_flush_ptr_NULL);
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+   png_set_filter_heuristics(png_ptr, PNG_FILTER_HEURISTIC_DEFAULT,
+      1, png_doublep_NULL, png_doublep_NULL);
+#endif
+
+#ifdef PNG_SETJMP_SUPPORTED
+/* Applications that neglect to set up their own setjmp() and then encounter
+   a png_error() will longjmp here.  Since the jmpbuf is then meaningless we
+   abort instead of returning. */
+#ifdef USE_FAR_KEYWORD
+   if (setjmp(jmpbuf))
+      PNG_ABORT();
+   png_memcpy(png_ptr->jmpbuf,jmpbuf,png_sizeof(jmp_buf));
+#else
+   if (setjmp(png_ptr->jmpbuf))
+      PNG_ABORT();
+#endif
+#endif
+   return (png_ptr);
+}
+
+/* Initialize png_ptr structure, and allocate any memory needed */
+#if defined(PNG_1_0_X) || defined(PNG_1_2_X)
+/* Deprecated. */
+#undef png_write_init
+void PNGAPI
+png_write_init(png_structp png_ptr)
+{
+   /* We only come here via pre-1.0.7-compiled applications */
+   png_write_init_2(png_ptr, "1.0.6 or earlier", 0, 0);
+}
+
+void PNGAPI
+png_write_init_2(png_structp png_ptr, png_const_charp user_png_ver,
+   png_size_t png_struct_size, png_size_t png_info_size)
+{
+   /* We only come here via pre-1.0.12-compiled applications */
+   if(png_ptr == NULL) return;
+#if !defined(PNG_NO_STDIO) && !defined(_WIN32_WCE)
+   if(png_sizeof(png_struct) > png_struct_size ||
+      png_sizeof(png_info) > png_info_size)
+   {
+      char msg[80];
+      png_ptr->warning_fn=NULL;
+      if (user_png_ver)
+      {
+        png_snprintf(msg, 80,
+           "Application was compiled with png.h from libpng-%.20s",
+           user_png_ver);
+        png_warning(png_ptr, msg);
+      }
+      png_snprintf(msg, 80,
+         "Application  is  running with png.c from libpng-%.20s",
+         png_libpng_ver);
+      png_warning(png_ptr, msg);
+   }
+#endif
+   if(png_sizeof(png_struct) > png_struct_size)
+     {
+       png_ptr->error_fn=NULL;
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+       png_ptr->flags=0;
+#endif
+       png_error(png_ptr,
+       "The png struct allocated by the application for writing is too small.");
+     }
+   if(png_sizeof(png_info) > png_info_size)
+     {
+       png_ptr->error_fn=NULL;
+#ifdef PNG_ERROR_NUMBERS_SUPPORTED
+       png_ptr->flags=0;
+#endif
+       png_error(png_ptr,
+       "The info struct allocated by the application for writing is too small.");
+     }
+   png_write_init_3(&png_ptr, user_png_ver, png_struct_size);
+}
+#endif /* PNG_1_0_X || PNG_1_2_X */
+
+
+void PNGAPI
+png_write_init_3(png_structpp ptr_ptr, png_const_charp user_png_ver,
+   png_size_t png_struct_size)
+{
+   png_structp png_ptr=*ptr_ptr;
+#ifdef PNG_SETJMP_SUPPORTED
+   jmp_buf tmp_jmp; /* to save current jump buffer */
+#endif
+
+   int i = 0;
+
+   if (png_ptr == NULL)
+      return;
+
+   do
+   {
+     if (user_png_ver[i] != png_libpng_ver[i])
+     {
+#ifdef PNG_LEGACY_SUPPORTED
+       png_ptr->flags |= PNG_FLAG_LIBRARY_MISMATCH;
+#else
+       png_ptr->warning_fn=NULL;
+       png_warning(png_ptr,
+     "Application uses deprecated png_write_init() and should be recompiled.");
+       break;
+#endif
+     }
+   } while (png_libpng_ver[i++]);
+
+   png_debug(1, "in png_write_init_3\n");
+
+#ifdef PNG_SETJMP_SUPPORTED
+   /* save jump buffer and error functions */
+   png_memcpy(tmp_jmp, png_ptr->jmpbuf, png_sizeof (jmp_buf));
+#endif
+
+   if (png_sizeof(png_struct) > png_struct_size)
+     {
+       png_destroy_struct(png_ptr);
+       png_ptr = (png_structp)png_create_struct(PNG_STRUCT_PNG);
+       *ptr_ptr = png_ptr;
+     }
+
+   /* reset all variables to 0 */
+   png_memset(png_ptr, 0, png_sizeof (png_struct));
+
+   /* added at libpng-1.2.6 */
+#ifdef PNG_SET_USER_LIMITS_SUPPORTED
+   png_ptr->user_width_max=PNG_USER_WIDTH_MAX;
+   png_ptr->user_height_max=PNG_USER_HEIGHT_MAX;
+#endif
+
+#if !defined(PNG_1_0_X)
+#ifdef PNG_ASSEMBLER_CODE_SUPPORTED
+#ifdef PNG_MMX_CODE_SUPPORTED
+   png_init_mmx_flags(png_ptr);   /* 1.2.0 addition */
+#endif
+#endif
+#endif /* PNG_1_0_X */
+
+#ifdef PNG_SETJMP_SUPPORTED
+   /* restore jump buffer */
+   png_memcpy(png_ptr->jmpbuf, tmp_jmp, png_sizeof (jmp_buf));
+#endif
+
+   png_set_write_fn(png_ptr, png_voidp_NULL, png_rw_ptr_NULL,
+      png_flush_ptr_NULL);
+
+   /* initialize zbuf - compression buffer */
+   png_ptr->zbuf_size = PNG_ZBUF_SIZE;
+   png_ptr->zbuf = (png_bytep)png_malloc(png_ptr,
+      (png_uint_32)png_ptr->zbuf_size);
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+   png_set_filter_heuristics(png_ptr, PNG_FILTER_HEURISTIC_DEFAULT,
+      1, png_doublep_NULL, png_doublep_NULL);
+#endif
+}
+
+/* Write a few rows of image data.  If the image is interlaced,
+ * either you will have to write the 7 sub images, or, if you
+ * have called png_set_interlace_handling(), you will have to
+ * "write" the image seven times.
+ */
+void PNGAPI
+png_write_rows(png_structp png_ptr, png_bytepp row,
+   png_uint_32 num_rows)
+{
+   png_uint_32 i; /* row counter */
+   png_bytepp rp; /* row pointer */
+
+   png_debug(1, "in png_write_rows\n");
+
+   if (png_ptr == NULL)
+      return;
+
+   /* loop through the rows */
+   for (i = 0, rp = row; i < num_rows; i++, rp++)
+   {
+      png_write_row(png_ptr, *rp);
+   }
+}
+
+/* Write the image.  You only need to call this function once, even
+ * if you are writing an interlaced image.
+ */
+void PNGAPI
+png_write_image(png_structp png_ptr, png_bytepp image)
+{
+   png_uint_32 i; /* row index */
+   int pass, num_pass; /* pass variables */
+   png_bytepp rp; /* points to current row */
+
+   if (png_ptr == NULL)
+      return;
+
+   png_debug(1, "in png_write_image\n");
+#if defined(PNG_WRITE_INTERLACING_SUPPORTED)
+   /* intialize interlace handling.  If image is not interlaced,
+      this will set pass to 1 */
+   num_pass = png_set_interlace_handling(png_ptr);
+#else
+   num_pass = 1;
+#endif
+   /* loop through passes */
+   for (pass = 0; pass < num_pass; pass++)
+   {
+      /* loop through image */
+      for (i = 0, rp = image; i < png_ptr->height; i++, rp++)
+      {
+         png_write_row(png_ptr, *rp);
+      }
+   }
+}
+
+/* called by user to write a row of image data */
+void PNGAPI
+png_write_row(png_structp png_ptr, png_bytep row)
+{
+   if (png_ptr == NULL)
+      return;
+   png_debug2(1, "in png_write_row (row %ld, pass %d)\n",
+      png_ptr->row_number, png_ptr->pass);
+
+   /* initialize transformations and other stuff if first time */
+   if (png_ptr->row_number == 0 && png_ptr->pass == 0)
+   {
+   /* make sure we wrote the header info */
+   if (!(png_ptr->mode & PNG_WROTE_INFO_BEFORE_PLTE))
+      png_error(png_ptr,
+         "png_write_info was never called before png_write_row.");
+
+   /* check for transforms that have been set but were defined out */
+#if !defined(PNG_WRITE_INVERT_SUPPORTED) && defined(PNG_READ_INVERT_SUPPORTED)
+   if (png_ptr->transformations & PNG_INVERT_MONO)
+      png_warning(png_ptr, "PNG_WRITE_INVERT_SUPPORTED is not defined.");
+#endif
+#if !defined(PNG_WRITE_FILLER_SUPPORTED) && defined(PNG_READ_FILLER_SUPPORTED)
+   if (png_ptr->transformations & PNG_FILLER)
+      png_warning(png_ptr, "PNG_WRITE_FILLER_SUPPORTED is not defined.");
+#endif
+#if !defined(PNG_WRITE_PACKSWAP_SUPPORTED) && defined(PNG_READ_PACKSWAP_SUPPORTED)
+   if (png_ptr->transformations & PNG_PACKSWAP)
+      png_warning(png_ptr, "PNG_WRITE_PACKSWAP_SUPPORTED is not defined.");
+#endif
+#if !defined(PNG_WRITE_PACK_SUPPORTED) && defined(PNG_READ_PACK_SUPPORTED)
+   if (png_ptr->transformations & PNG_PACK)
+      png_warning(png_ptr, "PNG_WRITE_PACK_SUPPORTED is not defined.");
+#endif
+#if !defined(PNG_WRITE_SHIFT_SUPPORTED) && defined(PNG_READ_SHIFT_SUPPORTED)
+   if (png_ptr->transformations & PNG_SHIFT)
+      png_warning(png_ptr, "PNG_WRITE_SHIFT_SUPPORTED is not defined.");
+#endif
+#if !defined(PNG_WRITE_BGR_SUPPORTED) && defined(PNG_READ_BGR_SUPPORTED)
+   if (png_ptr->transformations & PNG_BGR)
+      png_warning(png_ptr, "PNG_WRITE_BGR_SUPPORTED is not defined.");
+#endif
+#if !defined(PNG_WRITE_SWAP_SUPPORTED) && defined(PNG_READ_SWAP_SUPPORTED)
+   if (png_ptr->transformations & PNG_SWAP_BYTES)
+      png_warning(png_ptr, "PNG_WRITE_SWAP_SUPPORTED is not defined.");
+#endif
+
+      png_write_start_row(png_ptr);
+   }
+
+#if defined(PNG_WRITE_INTERLACING_SUPPORTED)
+   /* if interlaced and not interested in row, return */
+   if (png_ptr->interlaced && (png_ptr->transformations & PNG_INTERLACE))
+   {
+      switch (png_ptr->pass)
+      {
+         case 0:
+            if (png_ptr->row_number & 0x07)
+            {
+               png_write_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 1:
+            if ((png_ptr->row_number & 0x07) || png_ptr->width < 5)
+            {
+               png_write_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 2:
+            if ((png_ptr->row_number & 0x07) != 4)
+            {
+               png_write_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 3:
+            if ((png_ptr->row_number & 0x03) || png_ptr->width < 3)
+            {
+               png_write_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 4:
+            if ((png_ptr->row_number & 0x03) != 2)
+            {
+               png_write_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 5:
+            if ((png_ptr->row_number & 0x01) || png_ptr->width < 2)
+            {
+               png_write_finish_row(png_ptr);
+               return;
+            }
+            break;
+         case 6:
+            if (!(png_ptr->row_number & 0x01))
+            {
+               png_write_finish_row(png_ptr);
+               return;
+            }
+            break;
+      }
+   }
+#endif
+
+   /* set up row info for transformations */
+   png_ptr->row_info.color_type = png_ptr->color_type;
+   png_ptr->row_info.width = png_ptr->usr_width;
+   png_ptr->row_info.channels = png_ptr->usr_channels;
+   png_ptr->row_info.bit_depth = png_ptr->usr_bit_depth;
+   png_ptr->row_info.pixel_depth = (png_byte)(png_ptr->row_info.bit_depth *
+      png_ptr->row_info.channels);
+
+   png_ptr->row_info.rowbytes = PNG_ROWBYTES(png_ptr->row_info.pixel_depth,
+      png_ptr->row_info.width);
+
+   png_debug1(3, "row_info->color_type = %d\n", png_ptr->row_info.color_type);
+   png_debug1(3, "row_info->width = %lu\n", png_ptr->row_info.width);
+   png_debug1(3, "row_info->channels = %d\n", png_ptr->row_info.channels);
+   png_debug1(3, "row_info->bit_depth = %d\n", png_ptr->row_info.bit_depth);
+   png_debug1(3, "row_info->pixel_depth = %d\n", png_ptr->row_info.pixel_depth);
+   png_debug1(3, "row_info->rowbytes = %lu\n", png_ptr->row_info.rowbytes);
+
+   /* Copy user's row into buffer, leaving room for filter byte. */
+   png_memcpy_check(png_ptr, png_ptr->row_buf + 1, row,
+      png_ptr->row_info.rowbytes);
+
+#if defined(PNG_WRITE_INTERLACING_SUPPORTED)
+   /* handle interlacing */
+   if (png_ptr->interlaced && png_ptr->pass < 6 &&
+      (png_ptr->transformations & PNG_INTERLACE))
+   {
+      png_do_write_interlace(&(png_ptr->row_info),
+         png_ptr->row_buf + 1, png_ptr->pass);
+      /* this should always get caught above, but still ... */
+      if (!(png_ptr->row_info.width))
+      {
+         png_write_finish_row(png_ptr);
+         return;
+      }
+   }
+#endif
+
+   /* handle other transformations */
+   if (png_ptr->transformations)
+      png_do_write_transformations(png_ptr);
+
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+   /* Write filter_method 64 (intrapixel differencing) only if
+    * 1. Libpng was compiled with PNG_MNG_FEATURES_SUPPORTED and
+    * 2. Libpng did not write a PNG signature (this filter_method is only
+    *    used in PNG datastreams that are embedded in MNG datastreams) and
+    * 3. The application called png_permit_mng_features with a mask that
+    *    included PNG_FLAG_MNG_FILTER_64 and
+    * 4. The filter_method is 64 and
+    * 5. The color_type is RGB or RGBA
+    */
+   if((png_ptr->mng_features_permitted & PNG_FLAG_MNG_FILTER_64) &&
+      (png_ptr->filter_type == PNG_INTRAPIXEL_DIFFERENCING))
+   {
+      /* Intrapixel differencing */
+      png_do_write_intrapixel(&(png_ptr->row_info), png_ptr->row_buf + 1);
+   }
+#endif
+
+   /* Find a filter if necessary, filter the row and write it out. */
+   png_write_find_filter(png_ptr, &(png_ptr->row_info));
+
+   if (png_ptr->write_row_fn != NULL)
+      (*(png_ptr->write_row_fn))(png_ptr, png_ptr->row_number, png_ptr->pass);
+}
+
+#if defined(PNG_WRITE_FLUSH_SUPPORTED)
+/* Set the automatic flush interval or 0 to turn flushing off */
+void PNGAPI
+png_set_flush(png_structp png_ptr, int nrows)
+{
+   png_debug(1, "in png_set_flush\n");
+   if (png_ptr == NULL)
+      return;
+   png_ptr->flush_dist = (nrows < 0 ? 0 : nrows);
+}
+
+/* flush the current output buffers now */
+void PNGAPI
+png_write_flush(png_structp png_ptr)
+{
+   int wrote_IDAT;
+
+   png_debug(1, "in png_write_flush\n");
+   if (png_ptr == NULL)
+      return;
+   /* We have already written out all of the data */
+   if (png_ptr->row_number >= png_ptr->num_rows)
+     return;
+
+   do
+   {
+      int ret;
+
+      /* compress the data */
+      ret = deflate(&png_ptr->zstream, Z_SYNC_FLUSH);
+      wrote_IDAT = 0;
+
+      /* check for compression errors */
+      if (ret != Z_OK)
+      {
+         if (png_ptr->zstream.msg != NULL)
+            png_error(png_ptr, png_ptr->zstream.msg);
+         else
+            png_error(png_ptr, "zlib error");
+      }
+
+      if (!(png_ptr->zstream.avail_out))
+      {
+         /* write the IDAT and reset the zlib output buffer */
+         png_write_IDAT(png_ptr, png_ptr->zbuf,
+                        png_ptr->zbuf_size);
+         png_ptr->zstream.next_out = png_ptr->zbuf;
+         png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+         wrote_IDAT = 1;
+      }
+   } while(wrote_IDAT == 1);
+
+   /* If there is any data left to be output, write it into a new IDAT */
+   if (png_ptr->zbuf_size != png_ptr->zstream.avail_out)
+   {
+      /* write the IDAT and reset the zlib output buffer */
+      png_write_IDAT(png_ptr, png_ptr->zbuf,
+                     png_ptr->zbuf_size - png_ptr->zstream.avail_out);
+      png_ptr->zstream.next_out = png_ptr->zbuf;
+      png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+   }
+   png_ptr->flush_rows = 0;
+   png_flush(png_ptr);
+}
+#endif /* PNG_WRITE_FLUSH_SUPPORTED */
+
+/* free all memory used by the write */
+void PNGAPI
+png_destroy_write_struct(png_structpp png_ptr_ptr, png_infopp info_ptr_ptr)
+{
+   png_structp png_ptr = NULL;
+   png_infop info_ptr = NULL;
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_free_ptr free_fn = NULL;
+   png_voidp mem_ptr = NULL;
+#endif
+
+   png_debug(1, "in png_destroy_write_struct\n");
+   if (png_ptr_ptr != NULL)
+   {
+      png_ptr = *png_ptr_ptr;
+#ifdef PNG_USER_MEM_SUPPORTED
+      free_fn = png_ptr->free_fn;
+      mem_ptr = png_ptr->mem_ptr;
+#endif
+   }
+
+   if (info_ptr_ptr != NULL)
+      info_ptr = *info_ptr_ptr;
+
+   if (info_ptr != NULL)
+   {
+      png_free_data(png_ptr, info_ptr, PNG_FREE_ALL, -1);
+
+#if defined(PNG_UNKNOWN_CHUNKS_SUPPORTED)
+      if (png_ptr->num_chunk_list)
+      {
+         png_free(png_ptr, png_ptr->chunk_list);
+         png_ptr->chunk_list=NULL;
+         png_ptr->num_chunk_list=0;
+      }
+#endif
+
+#ifdef PNG_USER_MEM_SUPPORTED
+      png_destroy_struct_2((png_voidp)info_ptr, (png_free_ptr)free_fn,
+         (png_voidp)mem_ptr);
+#else
+      png_destroy_struct((png_voidp)info_ptr);
+#endif
+      *info_ptr_ptr = NULL;
+   }
+
+   if (png_ptr != NULL)
+   {
+      png_write_destroy(png_ptr);
+#ifdef PNG_USER_MEM_SUPPORTED
+      png_destroy_struct_2((png_voidp)png_ptr, (png_free_ptr)free_fn,
+         (png_voidp)mem_ptr);
+#else
+      png_destroy_struct((png_voidp)png_ptr);
+#endif
+      *png_ptr_ptr = NULL;
+   }
+}
+
+
+/* Free any memory used in png_ptr struct (old method) */
+void /* PRIVATE */
+png_write_destroy(png_structp png_ptr)
+{
+#ifdef PNG_SETJMP_SUPPORTED
+   jmp_buf tmp_jmp; /* save jump buffer */
+#endif
+   png_error_ptr error_fn;
+   png_error_ptr warning_fn;
+   png_voidp error_ptr;
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_free_ptr free_fn;
+#endif
+
+   png_debug(1, "in png_write_destroy\n");
+   /* free any memory zlib uses */
+   deflateEnd(&png_ptr->zstream);
+
+   /* free our memory.  png_free checks NULL for us. */
+   png_free(png_ptr, png_ptr->zbuf);
+   png_free(png_ptr, png_ptr->row_buf);
+   png_free(png_ptr, png_ptr->prev_row);
+   png_free(png_ptr, png_ptr->sub_row);
+   png_free(png_ptr, png_ptr->up_row);
+   png_free(png_ptr, png_ptr->avg_row);
+   png_free(png_ptr, png_ptr->paeth_row);
+
+#if defined(PNG_TIME_RFC1123_SUPPORTED)
+   png_free(png_ptr, png_ptr->time_buffer);
+#endif
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+   png_free(png_ptr, png_ptr->prev_filters);
+   png_free(png_ptr, png_ptr->filter_weights);
+   png_free(png_ptr, png_ptr->inv_filter_weights);
+   png_free(png_ptr, png_ptr->filter_costs);
+   png_free(png_ptr, png_ptr->inv_filter_costs);
+#endif
+
+#ifdef PNG_SETJMP_SUPPORTED
+   /* reset structure */
+   png_memcpy(tmp_jmp, png_ptr->jmpbuf, png_sizeof (jmp_buf));
+#endif
+
+   error_fn = png_ptr->error_fn;
+   warning_fn = png_ptr->warning_fn;
+   error_ptr = png_ptr->error_ptr;
+#ifdef PNG_USER_MEM_SUPPORTED
+   free_fn = png_ptr->free_fn;
+#endif
+
+   png_memset(png_ptr, 0, png_sizeof (png_struct));
+
+   png_ptr->error_fn = error_fn;
+   png_ptr->warning_fn = warning_fn;
+   png_ptr->error_ptr = error_ptr;
+#ifdef PNG_USER_MEM_SUPPORTED
+   png_ptr->free_fn = free_fn;
+#endif
+
+#ifdef PNG_SETJMP_SUPPORTED
+   png_memcpy(png_ptr->jmpbuf, tmp_jmp, png_sizeof (jmp_buf));
+#endif
+}
+
+/* Allow the application to select one or more row filters to use. */
+void PNGAPI
+png_set_filter(png_structp png_ptr, int method, int filters)
+{
+   png_debug(1, "in png_set_filter\n");
+   if (png_ptr == NULL)
+      return;
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+   if((png_ptr->mng_features_permitted & PNG_FLAG_MNG_FILTER_64) &&
+      (method == PNG_INTRAPIXEL_DIFFERENCING))
+         method = PNG_FILTER_TYPE_BASE;
+#endif
+   if (method == PNG_FILTER_TYPE_BASE)
+   {
+      switch (filters & (PNG_ALL_FILTERS | 0x07))
+      {
+#ifndef PNG_NO_WRITE_FILTER
+         case 5:
+         case 6:
+         case 7: png_warning(png_ptr, "Unknown row filter for method 0");
+#endif /* PNG_NO_WRITE_FILTER */
+         case PNG_FILTER_VALUE_NONE:
+              png_ptr->do_filter=PNG_FILTER_NONE; break;
+#ifndef PNG_NO_WRITE_FILTER
+         case PNG_FILTER_VALUE_SUB:
+              png_ptr->do_filter=PNG_FILTER_SUB; break;
+         case PNG_FILTER_VALUE_UP:
+              png_ptr->do_filter=PNG_FILTER_UP; break;
+         case PNG_FILTER_VALUE_AVG:
+              png_ptr->do_filter=PNG_FILTER_AVG; break;
+         case PNG_FILTER_VALUE_PAETH:
+              png_ptr->do_filter=PNG_FILTER_PAETH; break;
+         default: png_ptr->do_filter = (png_byte)filters; break;
+#else
+         default: png_warning(png_ptr, "Unknown row filter for method 0");
+#endif /* PNG_NO_WRITE_FILTER */
+      }
+
+      /* If we have allocated the row_buf, this means we have already started
+       * with the image and we should have allocated all of the filter buffers
+       * that have been selected.  If prev_row isn't already allocated, then
+       * it is too late to start using the filters that need it, since we
+       * will be missing the data in the previous row.  If an application
+       * wants to start and stop using particular filters during compression,
+       * it should start out with all of the filters, and then add and
+       * remove them after the start of compression.
+       */
+      if (png_ptr->row_buf != NULL)
+      {
+#ifndef PNG_NO_WRITE_FILTER
+         if ((png_ptr->do_filter & PNG_FILTER_SUB) && png_ptr->sub_row == NULL)
+         {
+            png_ptr->sub_row = (png_bytep)png_malloc(png_ptr,
+              (png_ptr->rowbytes + 1));
+            png_ptr->sub_row[0] = PNG_FILTER_VALUE_SUB;
+         }
+
+         if ((png_ptr->do_filter & PNG_FILTER_UP) && png_ptr->up_row == NULL)
+         {
+            if (png_ptr->prev_row == NULL)
+            {
+               png_warning(png_ptr, "Can't add Up filter after starting");
+               png_ptr->do_filter &= ~PNG_FILTER_UP;
+            }
+            else
+            {
+               png_ptr->up_row = (png_bytep)png_malloc(png_ptr,
+                  (png_ptr->rowbytes + 1));
+               png_ptr->up_row[0] = PNG_FILTER_VALUE_UP;
+            }
+         }
+
+         if ((png_ptr->do_filter & PNG_FILTER_AVG) && png_ptr->avg_row == NULL)
+         {
+            if (png_ptr->prev_row == NULL)
+            {
+               png_warning(png_ptr, "Can't add Average filter after starting");
+               png_ptr->do_filter &= ~PNG_FILTER_AVG;
+            }
+            else
+            {
+               png_ptr->avg_row = (png_bytep)png_malloc(png_ptr,
+                  (png_ptr->rowbytes + 1));
+               png_ptr->avg_row[0] = PNG_FILTER_VALUE_AVG;
+            }
+         }
+
+         if ((png_ptr->do_filter & PNG_FILTER_PAETH) &&
+             png_ptr->paeth_row == NULL)
+         {
+            if (png_ptr->prev_row == NULL)
+            {
+               png_warning(png_ptr, "Can't add Paeth filter after starting");
+               png_ptr->do_filter &= (png_byte)(~PNG_FILTER_PAETH);
+            }
+            else
+            {
+               png_ptr->paeth_row = (png_bytep)png_malloc(png_ptr,
+                  (png_ptr->rowbytes + 1));
+               png_ptr->paeth_row[0] = PNG_FILTER_VALUE_PAETH;
+            }
+         }
+
+         if (png_ptr->do_filter == PNG_NO_FILTERS)
+#endif /* PNG_NO_WRITE_FILTER */
+            png_ptr->do_filter = PNG_FILTER_NONE;
+      }
+   }
+   else
+      png_error(png_ptr, "Unknown custom filter method");
+}
+
+/* This allows us to influence the way in which libpng chooses the "best"
+ * filter for the current scanline.  While the "minimum-sum-of-absolute-
+ * differences metric is relatively fast and effective, there is some
+ * question as to whether it can be improved upon by trying to keep the
+ * filtered data going to zlib more consistent, hopefully resulting in
+ * better compression.
+ */
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)      /* GRR 970116 */
+void PNGAPI
+png_set_filter_heuristics(png_structp png_ptr, int heuristic_method,
+   int num_weights, png_doublep filter_weights,
+   png_doublep filter_costs)
+{
+   int i;
+
+   png_debug(1, "in png_set_filter_heuristics\n");
+   if (png_ptr == NULL)
+      return;
+   if (heuristic_method >= PNG_FILTER_HEURISTIC_LAST)
+   {
+      png_warning(png_ptr, "Unknown filter heuristic method");
+      return;
+   }
+
+   if (heuristic_method == PNG_FILTER_HEURISTIC_DEFAULT)
+   {
+      heuristic_method = PNG_FILTER_HEURISTIC_UNWEIGHTED;
+   }
+
+   if (num_weights < 0 || filter_weights == NULL ||
+      heuristic_method == PNG_FILTER_HEURISTIC_UNWEIGHTED)
+   {
+      num_weights = 0;
+   }
+
+   png_ptr->num_prev_filters = (png_byte)num_weights;
+   png_ptr->heuristic_method = (png_byte)heuristic_method;
+
+   if (num_weights > 0)
+   {
+      if (png_ptr->prev_filters == NULL)
+      {
+         png_ptr->prev_filters = (png_bytep)png_malloc(png_ptr,
+            (png_uint_32)(png_sizeof(png_byte) * num_weights));
+
+         /* To make sure that the weighting starts out fairly */
+         for (i = 0; i < num_weights; i++)
+         {
+            png_ptr->prev_filters[i] = 255;
+         }
+      }
+
+      if (png_ptr->filter_weights == NULL)
+      {
+         png_ptr->filter_weights = (png_uint_16p)png_malloc(png_ptr,
+            (png_uint_32)(png_sizeof(png_uint_16) * num_weights));
+
+         png_ptr->inv_filter_weights = (png_uint_16p)png_malloc(png_ptr,
+            (png_uint_32)(png_sizeof(png_uint_16) * num_weights));
+         for (i = 0; i < num_weights; i++)
+         {
+            png_ptr->inv_filter_weights[i] =
+            png_ptr->filter_weights[i] = PNG_WEIGHT_FACTOR;
+         }
+      }
+
+      for (i = 0; i < num_weights; i++)
+      {
+         if (filter_weights[i] < 0.0)
+         {
+            png_ptr->inv_filter_weights[i] =
+            png_ptr->filter_weights[i] = PNG_WEIGHT_FACTOR;
+         }
+         else
+         {
+            png_ptr->inv_filter_weights[i] =
+               (png_uint_16)((double)PNG_WEIGHT_FACTOR*filter_weights[i]+0.5);
+            png_ptr->filter_weights[i] =
+               (png_uint_16)((double)PNG_WEIGHT_FACTOR/filter_weights[i]+0.5);
+         }
+      }
+   }
+
+   /* If, in the future, there are other filter methods, this would
+    * need to be based on png_ptr->filter.
+    */
+   if (png_ptr->filter_costs == NULL)
+   {
+      png_ptr->filter_costs = (png_uint_16p)png_malloc(png_ptr,
+         (png_uint_32)(png_sizeof(png_uint_16) * PNG_FILTER_VALUE_LAST));
+
+      png_ptr->inv_filter_costs = (png_uint_16p)png_malloc(png_ptr,
+         (png_uint_32)(png_sizeof(png_uint_16) * PNG_FILTER_VALUE_LAST));
+
+      for (i = 0; i < PNG_FILTER_VALUE_LAST; i++)
+      {
+         png_ptr->inv_filter_costs[i] =
+         png_ptr->filter_costs[i] = PNG_COST_FACTOR;
+      }
+   }
+
+   /* Here is where we set the relative costs of the different filters.  We
+    * should take the desired compression level into account when setting
+    * the costs, so that Paeth, for instance, has a high relative cost at low
+    * compression levels, while it has a lower relative cost at higher
+    * compression settings.  The filter types are in order of increasing
+    * relative cost, so it would be possible to do this with an algorithm.
+    */
+   for (i = 0; i < PNG_FILTER_VALUE_LAST; i++)
+   {
+      if (filter_costs == NULL || filter_costs[i] < 0.0)
+      {
+         png_ptr->inv_filter_costs[i] =
+         png_ptr->filter_costs[i] = PNG_COST_FACTOR;
+      }
+      else if (filter_costs[i] >= 1.0)
+      {
+         png_ptr->inv_filter_costs[i] =
+            (png_uint_16)((double)PNG_COST_FACTOR / filter_costs[i] + 0.5);
+         png_ptr->filter_costs[i] =
+            (png_uint_16)((double)PNG_COST_FACTOR * filter_costs[i] + 0.5);
+      }
+   }
+}
+#endif /* PNG_WRITE_WEIGHTED_FILTER_SUPPORTED */
+
+void PNGAPI
+png_set_compression_level(png_structp png_ptr, int level)
+{
+   png_debug(1, "in png_set_compression_level\n");
+   if (png_ptr == NULL)
+      return;
+   png_ptr->flags |= PNG_FLAG_ZLIB_CUSTOM_LEVEL;
+   png_ptr->zlib_level = level;
+}
+
+void PNGAPI
+png_set_compression_mem_level(png_structp png_ptr, int mem_level)
+{
+   png_debug(1, "in png_set_compression_mem_level\n");
+   if (png_ptr == NULL)
+      return;
+   png_ptr->flags |= PNG_FLAG_ZLIB_CUSTOM_MEM_LEVEL;
+   png_ptr->zlib_mem_level = mem_level;
+}
+
+void PNGAPI
+png_set_compression_strategy(png_structp png_ptr, int strategy)
+{
+   png_debug(1, "in png_set_compression_strategy\n");
+   if (png_ptr == NULL)
+      return;
+   png_ptr->flags |= PNG_FLAG_ZLIB_CUSTOM_STRATEGY;
+   png_ptr->zlib_strategy = strategy;
+}
+
+void PNGAPI
+png_set_compression_window_bits(png_structp png_ptr, int window_bits)
+{
+   if (png_ptr == NULL)
+      return;
+   if (window_bits > 15)
+      png_warning(png_ptr, "Only compression windows <= 32k supported by PNG");
+   else if (window_bits < 8)
+      png_warning(png_ptr, "Only compression windows >= 256 supported by PNG");
+#ifndef WBITS_8_OK
+   /* avoid libpng bug with 256-byte windows */
+   if (window_bits == 8)
+     {
+       png_warning(png_ptr, "Compression window is being reset to 512");
+       window_bits=9;
+     }
+#endif
+   png_ptr->flags |= PNG_FLAG_ZLIB_CUSTOM_WINDOW_BITS;
+   png_ptr->zlib_window_bits = window_bits;
+}
+
+void PNGAPI
+png_set_compression_method(png_structp png_ptr, int method)
+{
+   png_debug(1, "in png_set_compression_method\n");
+   if (png_ptr == NULL)
+      return;
+   if (method != 8)
+      png_warning(png_ptr, "Only compression method 8 is supported by PNG");
+   png_ptr->flags |= PNG_FLAG_ZLIB_CUSTOM_METHOD;
+   png_ptr->zlib_method = method;
+}
+
+void PNGAPI
+png_set_write_status_fn(png_structp png_ptr, png_write_status_ptr write_row_fn)
+{
+   if (png_ptr == NULL)
+      return;
+   png_ptr->write_row_fn = write_row_fn;
+}
+
+#if defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED)
+void PNGAPI
+png_set_write_user_transform_fn(png_structp png_ptr, png_user_transform_ptr
+   write_user_transform_fn)
+{
+   png_debug(1, "in png_set_write_user_transform_fn\n");
+   if (png_ptr == NULL)
+      return;
+   png_ptr->transformations |= PNG_USER_TRANSFORM;
+   png_ptr->write_user_transform_fn = write_user_transform_fn;
+}
+#endif
+
+
+#if defined(PNG_INFO_IMAGE_SUPPORTED)
+void PNGAPI
+png_write_png(png_structp png_ptr, png_infop info_ptr,
+              int transforms, voidp params)
+{
+   if (png_ptr == NULL || info_ptr == NULL)
+      return;
+#if defined(PNG_WRITE_INVERT_ALPHA_SUPPORTED)
+   /* invert the alpha channel from opacity to transparency */
+   if (transforms & PNG_TRANSFORM_INVERT_ALPHA)
+       png_set_invert_alpha(png_ptr);
+#endif
+
+   /* Write the file header information. */
+   png_write_info(png_ptr, info_ptr);
+
+   /* ------ these transformations don't touch the info structure ------- */
+
+#if defined(PNG_WRITE_INVERT_SUPPORTED)
+   /* invert monochrome pixels */
+   if (transforms & PNG_TRANSFORM_INVERT_MONO)
+       png_set_invert_mono(png_ptr);
+#endif
+
+#if defined(PNG_WRITE_SHIFT_SUPPORTED)
+   /* Shift the pixels up to a legal bit depth and fill in
+    * as appropriate to correctly scale the image.
+    */
+   if ((transforms & PNG_TRANSFORM_SHIFT)
+               && (info_ptr->valid & PNG_INFO_sBIT))
+       png_set_shift(png_ptr, &info_ptr->sig_bit);
+#endif
+
+#if defined(PNG_WRITE_PACK_SUPPORTED)
+   /* pack pixels into bytes */
+   if (transforms & PNG_TRANSFORM_PACKING)
+       png_set_packing(png_ptr);
+#endif
+
+#if defined(PNG_WRITE_SWAP_ALPHA_SUPPORTED)
+   /* swap location of alpha bytes from ARGB to RGBA */
+   if (transforms & PNG_TRANSFORM_SWAP_ALPHA)
+       png_set_swap_alpha(png_ptr);
+#endif
+
+#if defined(PNG_WRITE_FILLER_SUPPORTED)
+   /* Get rid of filler (OR ALPHA) bytes, pack XRGB/RGBX/ARGB/RGBA into
+    * RGB (4 channels -> 3 channels). The second parameter is not used.
+    */
+   if (transforms & PNG_TRANSFORM_STRIP_FILLER)
+       png_set_filler(png_ptr, 0, PNG_FILLER_BEFORE);
+#endif
+
+#if defined(PNG_WRITE_BGR_SUPPORTED)
+   /* flip BGR pixels to RGB */
+   if (transforms & PNG_TRANSFORM_BGR)
+       png_set_bgr(png_ptr);
+#endif
+
+#if defined(PNG_WRITE_SWAP_SUPPORTED)
+   /* swap bytes of 16-bit files to most significant byte first */
+   if (transforms & PNG_TRANSFORM_SWAP_ENDIAN)
+       png_set_swap(png_ptr);
+#endif
+
+#if defined(PNG_WRITE_PACKSWAP_SUPPORTED)
+   /* swap bits of 1, 2, 4 bit packed pixel formats */
+   if (transforms & PNG_TRANSFORM_PACKSWAP)
+       png_set_packswap(png_ptr);
+#endif
+
+   /* ----------------------- end of transformations ------------------- */
+
+   /* write the bits */
+   if (info_ptr->valid & PNG_INFO_IDAT)
+       png_write_image(png_ptr, info_ptr->row_pointers);
+
+   /* It is REQUIRED to call this to finish writing the rest of the file */
+   png_write_end(png_ptr, info_ptr);
+
+   transforms = transforms; /* quiet compiler warnings */
+   params = params;
+}
+#endif
+#endif /* PNG_WRITE_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngwtran.c b/distrib/libpng-1.2.19/pngwtran.c
new file mode 100644
index 0000000..0372fe6
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngwtran.c
@@ -0,0 +1,572 @@
+
+/* pngwtran.c - transforms the data in a row for PNG writers
+ *
+ * Last changed in libpng 1.2.9 April 14, 2006
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2006 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+#ifdef PNG_WRITE_SUPPORTED
+
+/* Transform the data according to the user's wishes.  The order of
+ * transformations is significant.
+ */
+void /* PRIVATE */
+png_do_write_transformations(png_structp png_ptr)
+{
+   png_debug(1, "in png_do_write_transformations\n");
+
+   if (png_ptr == NULL)
+      return;
+
+#if defined(PNG_WRITE_USER_TRANSFORM_SUPPORTED)
+   if (png_ptr->transformations & PNG_USER_TRANSFORM)
+      if(png_ptr->write_user_transform_fn != NULL)
+        (*(png_ptr->write_user_transform_fn)) /* user write transform function */
+          (png_ptr,                    /* png_ptr */
+           &(png_ptr->row_info),       /* row_info:     */
+             /*  png_uint_32 width;          width of row */
+             /*  png_uint_32 rowbytes;       number of bytes in row */
+             /*  png_byte color_type;        color type of pixels */
+             /*  png_byte bit_depth;         bit depth of samples */
+             /*  png_byte channels;          number of channels (1-4) */
+             /*  png_byte pixel_depth;       bits per pixel (depth*channels) */
+           png_ptr->row_buf + 1);      /* start of pixel data for row */
+#endif
+#if defined(PNG_WRITE_FILLER_SUPPORTED)
+   if (png_ptr->transformations & PNG_FILLER)
+      png_do_strip_filler(&(png_ptr->row_info), png_ptr->row_buf + 1,
+         png_ptr->flags);
+#endif
+#if defined(PNG_WRITE_PACKSWAP_SUPPORTED)
+   if (png_ptr->transformations & PNG_PACKSWAP)
+      png_do_packswap(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+#if defined(PNG_WRITE_PACK_SUPPORTED)
+   if (png_ptr->transformations & PNG_PACK)
+      png_do_pack(&(png_ptr->row_info), png_ptr->row_buf + 1,
+         (png_uint_32)png_ptr->bit_depth);
+#endif
+#if defined(PNG_WRITE_SWAP_SUPPORTED)
+   if (png_ptr->transformations & PNG_SWAP_BYTES)
+      png_do_swap(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+#if defined(PNG_WRITE_SHIFT_SUPPORTED)
+   if (png_ptr->transformations & PNG_SHIFT)
+      png_do_shift(&(png_ptr->row_info), png_ptr->row_buf + 1,
+         &(png_ptr->shift));
+#endif
+#if defined(PNG_WRITE_SWAP_ALPHA_SUPPORTED)
+   if (png_ptr->transformations & PNG_SWAP_ALPHA)
+      png_do_write_swap_alpha(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+#if defined(PNG_WRITE_INVERT_ALPHA_SUPPORTED)
+   if (png_ptr->transformations & PNG_INVERT_ALPHA)
+      png_do_write_invert_alpha(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+#if defined(PNG_WRITE_BGR_SUPPORTED)
+   if (png_ptr->transformations & PNG_BGR)
+      png_do_bgr(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+#if defined(PNG_WRITE_INVERT_SUPPORTED)
+   if (png_ptr->transformations & PNG_INVERT_MONO)
+      png_do_invert(&(png_ptr->row_info), png_ptr->row_buf + 1);
+#endif
+}
+
+#if defined(PNG_WRITE_PACK_SUPPORTED)
+/* Pack pixels into bytes.  Pass the true bit depth in bit_depth.  The
+ * row_info bit depth should be 8 (one pixel per byte).  The channels
+ * should be 1 (this only happens on grayscale and paletted images).
+ */
+void /* PRIVATE */
+png_do_pack(png_row_infop row_info, png_bytep row, png_uint_32 bit_depth)
+{
+   png_debug(1, "in png_do_pack\n");
+   if (row_info->bit_depth == 8 &&
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+      row_info->channels == 1)
+   {
+      switch ((int)bit_depth)
+      {
+         case 1:
+         {
+            png_bytep sp, dp;
+            int mask, v;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            sp = row;
+            dp = row;
+            mask = 0x80;
+            v = 0;
+
+            for (i = 0; i < row_width; i++)
+            {
+               if (*sp != 0)
+                  v |= mask;
+               sp++;
+               if (mask > 1)
+                  mask >>= 1;
+               else
+               {
+                  mask = 0x80;
+                  *dp = (png_byte)v;
+                  dp++;
+                  v = 0;
+               }
+            }
+            if (mask != 0x80)
+               *dp = (png_byte)v;
+            break;
+         }
+         case 2:
+         {
+            png_bytep sp, dp;
+            int shift, v;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            sp = row;
+            dp = row;
+            shift = 6;
+            v = 0;
+            for (i = 0; i < row_width; i++)
+            {
+               png_byte value;
+
+               value = (png_byte)(*sp & 0x03);
+               v |= (value << shift);
+               if (shift == 0)
+               {
+                  shift = 6;
+                  *dp = (png_byte)v;
+                  dp++;
+                  v = 0;
+               }
+               else
+                  shift -= 2;
+               sp++;
+            }
+            if (shift != 6)
+               *dp = (png_byte)v;
+            break;
+         }
+         case 4:
+         {
+            png_bytep sp, dp;
+            int shift, v;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            sp = row;
+            dp = row;
+            shift = 4;
+            v = 0;
+            for (i = 0; i < row_width; i++)
+            {
+               png_byte value;
+
+               value = (png_byte)(*sp & 0x0f);
+               v |= (value << shift);
+
+               if (shift == 0)
+               {
+                  shift = 4;
+                  *dp = (png_byte)v;
+                  dp++;
+                  v = 0;
+               }
+               else
+                  shift -= 4;
+
+               sp++;
+            }
+            if (shift != 4)
+               *dp = (png_byte)v;
+            break;
+         }
+      }
+      row_info->bit_depth = (png_byte)bit_depth;
+      row_info->pixel_depth = (png_byte)(bit_depth * row_info->channels);
+      row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,
+         row_info->width);
+   }
+}
+#endif
+
+#if defined(PNG_WRITE_SHIFT_SUPPORTED)
+/* Shift pixel values to take advantage of whole range.  Pass the
+ * true number of bits in bit_depth.  The row should be packed
+ * according to row_info->bit_depth.  Thus, if you had a row of
+ * bit depth 4, but the pixels only had values from 0 to 7, you
+ * would pass 3 as bit_depth, and this routine would translate the
+ * data to 0 to 15.
+ */
+void /* PRIVATE */
+png_do_shift(png_row_infop row_info, png_bytep row, png_color_8p bit_depth)
+{
+   png_debug(1, "in png_do_shift\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL &&
+#else
+   if (
+#endif
+      row_info->color_type != PNG_COLOR_TYPE_PALETTE)
+   {
+      int shift_start[4], shift_dec[4];
+      int channels = 0;
+
+      if (row_info->color_type & PNG_COLOR_MASK_COLOR)
+      {
+         shift_start[channels] = row_info->bit_depth - bit_depth->red;
+         shift_dec[channels] = bit_depth->red;
+         channels++;
+         shift_start[channels] = row_info->bit_depth - bit_depth->green;
+         shift_dec[channels] = bit_depth->green;
+         channels++;
+         shift_start[channels] = row_info->bit_depth - bit_depth->blue;
+         shift_dec[channels] = bit_depth->blue;
+         channels++;
+      }
+      else
+      {
+         shift_start[channels] = row_info->bit_depth - bit_depth->gray;
+         shift_dec[channels] = bit_depth->gray;
+         channels++;
+      }
+      if (row_info->color_type & PNG_COLOR_MASK_ALPHA)
+      {
+         shift_start[channels] = row_info->bit_depth - bit_depth->alpha;
+         shift_dec[channels] = bit_depth->alpha;
+         channels++;
+      }
+
+      /* with low row depths, could only be grayscale, so one channel */
+      if (row_info->bit_depth < 8)
+      {
+         png_bytep bp = row;
+         png_uint_32 i;
+         png_byte mask;
+         png_uint_32 row_bytes = row_info->rowbytes;
+
+         if (bit_depth->gray == 1 && row_info->bit_depth == 2)
+            mask = 0x55;
+         else if (row_info->bit_depth == 4 && bit_depth->gray == 3)
+            mask = 0x11;
+         else
+            mask = 0xff;
+
+         for (i = 0; i < row_bytes; i++, bp++)
+         {
+            png_uint_16 v;
+            int j;
+
+            v = *bp;
+            *bp = 0;
+            for (j = shift_start[0]; j > -shift_dec[0]; j -= shift_dec[0])
+            {
+               if (j > 0)
+                  *bp |= (png_byte)((v << j) & 0xff);
+               else
+                  *bp |= (png_byte)((v >> (-j)) & mask);
+            }
+         }
+      }
+      else if (row_info->bit_depth == 8)
+      {
+         png_bytep bp = row;
+         png_uint_32 i;
+         png_uint_32 istop = channels * row_info->width;
+
+         for (i = 0; i < istop; i++, bp++)
+         {
+
+            png_uint_16 v;
+            int j;
+            int c = (int)(i%channels);
+
+            v = *bp;
+            *bp = 0;
+            for (j = shift_start[c]; j > -shift_dec[c]; j -= shift_dec[c])
+            {
+               if (j > 0)
+                  *bp |= (png_byte)((v << j) & 0xff);
+               else
+                  *bp |= (png_byte)((v >> (-j)) & 0xff);
+            }
+         }
+      }
+      else
+      {
+         png_bytep bp;
+         png_uint_32 i;
+         png_uint_32 istop = channels * row_info->width;
+
+         for (bp = row, i = 0; i < istop; i++)
+         {
+            int c = (int)(i%channels);
+            png_uint_16 value, v;
+            int j;
+
+            v = (png_uint_16)(((png_uint_16)(*bp) << 8) + *(bp + 1));
+            value = 0;
+            for (j = shift_start[c]; j > -shift_dec[c]; j -= shift_dec[c])
+            {
+               if (j > 0)
+                  value |= (png_uint_16)((v << j) & (png_uint_16)0xffff);
+               else
+                  value |= (png_uint_16)((v >> (-j)) & (png_uint_16)0xffff);
+            }
+            *bp++ = (png_byte)(value >> 8);
+            *bp++ = (png_byte)(value & 0xff);
+         }
+      }
+   }
+}
+#endif
+
+#if defined(PNG_WRITE_SWAP_ALPHA_SUPPORTED)
+void /* PRIVATE */
+png_do_write_swap_alpha(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_write_swap_alpha\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL)
+#endif
+   {
+      if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+      {
+         /* This converts from ARGB to RGBA */
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp, dp;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+            for (i = 0, sp = dp = row; i < row_width; i++)
+            {
+               png_byte save = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = save;
+            }
+         }
+         /* This converts from AARRGGBB to RRGGBBAA */
+         else
+         {
+            png_bytep sp, dp;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            for (i = 0, sp = dp = row; i < row_width; i++)
+            {
+               png_byte save[2];
+               save[0] = *(sp++);
+               save[1] = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = save[0];
+               *(dp++) = save[1];
+            }
+         }
+      }
+      else if (row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
+      {
+         /* This converts from AG to GA */
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp, dp;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            for (i = 0, sp = dp = row; i < row_width; i++)
+            {
+               png_byte save = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = save;
+            }
+         }
+         /* This converts from AAGG to GGAA */
+         else
+         {
+            png_bytep sp, dp;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            for (i = 0, sp = dp = row; i < row_width; i++)
+            {
+               png_byte save[2];
+               save[0] = *(sp++);
+               save[1] = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = save[0];
+               *(dp++) = save[1];
+            }
+         }
+      }
+   }
+}
+#endif
+
+#if defined(PNG_WRITE_INVERT_ALPHA_SUPPORTED)
+void /* PRIVATE */
+png_do_write_invert_alpha(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_write_invert_alpha\n");
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL)
+#endif
+   {
+      if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+      {
+         /* This inverts the alpha channel in RGBA */
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp, dp;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+            for (i = 0, sp = dp = row; i < row_width; i++)
+            {
+               /* does nothing
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               */
+               sp+=3; dp = sp;
+               *(dp++) = (png_byte)(255 - *(sp++));
+            }
+         }
+         /* This inverts the alpha channel in RRGGBBAA */
+         else
+         {
+            png_bytep sp, dp;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            for (i = 0, sp = dp = row; i < row_width; i++)
+            {
+               /* does nothing
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               */
+               sp+=6; dp = sp;
+               *(dp++) = (png_byte)(255 - *(sp++));
+               *(dp++) = (png_byte)(255 - *(sp++));
+            }
+         }
+      }
+      else if (row_info->color_type == PNG_COLOR_TYPE_GRAY_ALPHA)
+      {
+         /* This inverts the alpha channel in GA */
+         if (row_info->bit_depth == 8)
+         {
+            png_bytep sp, dp;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            for (i = 0, sp = dp = row; i < row_width; i++)
+            {
+               *(dp++) = *(sp++);
+               *(dp++) = (png_byte)(255 - *(sp++));
+            }
+         }
+         /* This inverts the alpha channel in GGAA */
+         else
+         {
+            png_bytep sp, dp;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            for (i = 0, sp = dp = row; i < row_width; i++)
+            {
+               /* does nothing
+               *(dp++) = *(sp++);
+               *(dp++) = *(sp++);
+               */
+               sp+=2; dp = sp;
+               *(dp++) = (png_byte)(255 - *(sp++));
+               *(dp++) = (png_byte)(255 - *(sp++));
+            }
+         }
+      }
+   }
+}
+#endif
+
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+/* undoes intrapixel differencing  */
+void /* PRIVATE */
+png_do_write_intrapixel(png_row_infop row_info, png_bytep row)
+{
+   png_debug(1, "in png_do_write_intrapixel\n");
+   if (
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+       row != NULL && row_info != NULL &&
+#endif
+       (row_info->color_type & PNG_COLOR_MASK_COLOR))
+   {
+      int bytes_per_pixel;
+      png_uint_32 row_width = row_info->width;
+      if (row_info->bit_depth == 8)
+      {
+         png_bytep rp;
+         png_uint_32 i;
+
+         if (row_info->color_type == PNG_COLOR_TYPE_RGB)
+            bytes_per_pixel = 3;
+         else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+            bytes_per_pixel = 4;
+         else
+            return;
+
+         for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
+         {
+            *(rp)   = (png_byte)((*rp     - *(rp+1))&0xff);
+            *(rp+2) = (png_byte)((*(rp+2) - *(rp+1))&0xff);
+         }
+      }
+      else if (row_info->bit_depth == 16)
+      {
+         png_bytep rp;
+         png_uint_32 i;
+
+         if (row_info->color_type == PNG_COLOR_TYPE_RGB)
+            bytes_per_pixel = 6;
+         else if (row_info->color_type == PNG_COLOR_TYPE_RGB_ALPHA)
+            bytes_per_pixel = 8;
+         else
+            return;
+
+         for (i = 0, rp = row; i < row_width; i++, rp += bytes_per_pixel)
+         {
+            png_uint_32 s0   = (*(rp  ) << 8) | *(rp+1);
+            png_uint_32 s1   = (*(rp+2) << 8) | *(rp+3);
+            png_uint_32 s2   = (*(rp+4) << 8) | *(rp+5);
+            png_uint_32 red  = (png_uint_32)((s0-s1) & 0xffffL);
+            png_uint_32 blue = (png_uint_32)((s2-s1) & 0xffffL);
+            *(rp  ) = (png_byte)((red >> 8) & 0xff);
+            *(rp+1) = (png_byte)(red & 0xff);
+            *(rp+4) = (png_byte)((blue >> 8) & 0xff);
+            *(rp+5) = (png_byte)(blue & 0xff);
+         }
+      }
+   }
+}
+#endif /* PNG_MNG_FEATURES_SUPPORTED */
+#endif /* PNG_WRITE_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/pngwutil.c b/distrib/libpng-1.2.19/pngwutil.c
new file mode 100644
index 0000000..849ee7d
--- /dev/null
+++ b/distrib/libpng-1.2.19/pngwutil.c
@@ -0,0 +1,2782 @@
+
+/* pngwutil.c - utilities to write a PNG file
+ *
+ * Last changed in libpng 1.2.19 August 18, 2007
+ * For conditions of distribution and use, see copyright notice in png.h
+ * Copyright (c) 1998-2007 Glenn Randers-Pehrson
+ * (Version 0.96 Copyright (c) 1996, 1997 Andreas Dilger)
+ * (Version 0.88 Copyright (c) 1995, 1996 Guy Eric Schalnat, Group 42, Inc.)
+ */
+
+#define PNG_INTERNAL
+#include "png.h"
+#ifdef PNG_WRITE_SUPPORTED
+
+/* Place a 32-bit number into a buffer in PNG byte order.  We work
+ * with unsigned numbers for convenience, although one supported
+ * ancillary chunk uses signed (two's complement) numbers.
+ */
+void PNGAPI
+png_save_uint_32(png_bytep buf, png_uint_32 i)
+{
+   buf[0] = (png_byte)((i >> 24) & 0xff);
+   buf[1] = (png_byte)((i >> 16) & 0xff);
+   buf[2] = (png_byte)((i >> 8) & 0xff);
+   buf[3] = (png_byte)(i & 0xff);
+}
+
+/* The png_save_int_32 function assumes integers are stored in two's
+ * complement format.  If this isn't the case, then this routine needs to
+ * be modified to write data in two's complement format.
+ */
+void PNGAPI
+png_save_int_32(png_bytep buf, png_int_32 i)
+{
+   buf[0] = (png_byte)((i >> 24) & 0xff);
+   buf[1] = (png_byte)((i >> 16) & 0xff);
+   buf[2] = (png_byte)((i >> 8) & 0xff);
+   buf[3] = (png_byte)(i & 0xff);
+}
+
+/* Place a 16-bit number into a buffer in PNG byte order.
+ * The parameter is declared unsigned int, not png_uint_16,
+ * just to avoid potential problems on pre-ANSI C compilers.
+ */
+void PNGAPI
+png_save_uint_16(png_bytep buf, unsigned int i)
+{
+   buf[0] = (png_byte)((i >> 8) & 0xff);
+   buf[1] = (png_byte)(i & 0xff);
+}
+
+/* Write a PNG chunk all at once.  The type is an array of ASCII characters
+ * representing the chunk name.  The array must be at least 4 bytes in
+ * length, and does not need to be null terminated.  To be safe, pass the
+ * pre-defined chunk names here, and if you need a new one, define it
+ * where the others are defined.  The length is the length of the data.
+ * All the data must be present.  If that is not possible, use the
+ * png_write_chunk_start(), png_write_chunk_data(), and png_write_chunk_end()
+ * functions instead.
+ */
+void PNGAPI
+png_write_chunk(png_structp png_ptr, png_bytep chunk_name,
+   png_bytep data, png_size_t length)
+{
+   if(png_ptr == NULL) return;
+   png_write_chunk_start(png_ptr, chunk_name, (png_uint_32)length);
+   png_write_chunk_data(png_ptr, data, length);
+   png_write_chunk_end(png_ptr);
+}
+
+/* Write the start of a PNG chunk.  The type is the chunk type.
+ * The total_length is the sum of the lengths of all the data you will be
+ * passing in png_write_chunk_data().
+ */
+void PNGAPI
+png_write_chunk_start(png_structp png_ptr, png_bytep chunk_name,
+   png_uint_32 length)
+{
+   png_byte buf[4];
+   png_debug2(0, "Writing %s chunk (%lu bytes)\n", chunk_name, length);
+   if(png_ptr == NULL) return;
+
+   /* write the length */
+   png_save_uint_32(buf, length);
+   png_write_data(png_ptr, buf, (png_size_t)4);
+
+   /* write the chunk name */
+   png_write_data(png_ptr, chunk_name, (png_size_t)4);
+   /* reset the crc and run it over the chunk name */
+   png_reset_crc(png_ptr);
+   png_calculate_crc(png_ptr, chunk_name, (png_size_t)4);
+}
+
+/* Write the data of a PNG chunk started with png_write_chunk_start().
+ * Note that multiple calls to this function are allowed, and that the
+ * sum of the lengths from these calls *must* add up to the total_length
+ * given to png_write_chunk_start().
+ */
+void PNGAPI
+png_write_chunk_data(png_structp png_ptr, png_bytep data, png_size_t length)
+{
+   /* write the data, and run the CRC over it */
+   if(png_ptr == NULL) return;
+   if (data != NULL && length > 0)
+   {
+      png_calculate_crc(png_ptr, data, length);
+      png_write_data(png_ptr, data, length);
+   }
+}
+
+/* Finish a chunk started with png_write_chunk_start(). */
+void PNGAPI
+png_write_chunk_end(png_structp png_ptr)
+{
+   png_byte buf[4];
+
+   if(png_ptr == NULL) return;
+
+   /* write the crc */
+   png_save_uint_32(buf, png_ptr->crc);
+
+   png_write_data(png_ptr, buf, (png_size_t)4);
+}
+
+/* Simple function to write the signature.  If we have already written
+ * the magic bytes of the signature, or more likely, the PNG stream is
+ * being embedded into another stream and doesn't need its own signature,
+ * we should call png_set_sig_bytes() to tell libpng how many of the
+ * bytes have already been written.
+ */
+void /* PRIVATE */
+png_write_sig(png_structp png_ptr)
+{
+   png_byte png_signature[8] = {137, 80, 78, 71, 13, 10, 26, 10};
+   /* write the rest of the 8 byte signature */
+   png_write_data(png_ptr, &png_signature[png_ptr->sig_bytes],
+      (png_size_t)8 - png_ptr->sig_bytes);
+   if(png_ptr->sig_bytes < 3)
+      png_ptr->mode |= PNG_HAVE_PNG_SIGNATURE;
+}
+
+#if defined(PNG_WRITE_TEXT_SUPPORTED) || defined(PNG_WRITE_iCCP_SUPPORTED)
+/*
+ * This pair of functions encapsulates the operation of (a) compressing a
+ * text string, and (b) issuing it later as a series of chunk data writes.
+ * The compression_state structure is shared context for these functions
+ * set up by the caller in order to make the whole mess thread-safe.
+ */
+
+typedef struct
+{
+    char *input;   /* the uncompressed input data */
+    int input_len;   /* its length */
+    int num_output_ptr; /* number of output pointers used */
+    int max_output_ptr; /* size of output_ptr */
+    png_charpp output_ptr; /* array of pointers to output */
+} compression_state;
+
+/* compress given text into storage in the png_ptr structure */
+static int /* PRIVATE */
+png_text_compress(png_structp png_ptr,
+        png_charp text, png_size_t text_len, int compression,
+        compression_state *comp)
+{
+   int ret;
+
+   comp->num_output_ptr = 0;
+   comp->max_output_ptr = 0;
+   comp->output_ptr = NULL;
+   comp->input = NULL;
+   comp->input_len = 0;
+
+   /* we may just want to pass the text right through */
+   if (compression == PNG_TEXT_COMPRESSION_NONE)
+   {
+       comp->input = text;
+       comp->input_len = text_len;
+       return((int)text_len);
+   }
+
+   if (compression >= PNG_TEXT_COMPRESSION_LAST)
+   {
+#if !defined(PNG_NO_STDIO) && !defined(_WIN32_WCE)
+      char msg[50];
+      png_snprintf(msg, 50, "Unknown compression type %d", compression);
+      png_warning(png_ptr, msg);
+#else
+      png_warning(png_ptr, "Unknown compression type");
+#endif
+   }
+
+   /* We can't write the chunk until we find out how much data we have,
+    * which means we need to run the compressor first and save the
+    * output.  This shouldn't be a problem, as the vast majority of
+    * comments should be reasonable, but we will set up an array of
+    * malloc'd pointers to be sure.
+    *
+    * If we knew the application was well behaved, we could simplify this
+    * greatly by assuming we can always malloc an output buffer large
+    * enough to hold the compressed text ((1001 * text_len / 1000) + 12)
+    * and malloc this directly.  The only time this would be a bad idea is
+    * if we can't malloc more than 64K and we have 64K of random input
+    * data, or if the input string is incredibly large (although this
+    * wouldn't cause a failure, just a slowdown due to swapping).
+    */
+
+   /* set up the compression buffers */
+   png_ptr->zstream.avail_in = (uInt)text_len;
+   png_ptr->zstream.next_in = (Bytef *)text;
+   png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+   png_ptr->zstream.next_out = (Bytef *)png_ptr->zbuf;
+
+   /* this is the same compression loop as in png_write_row() */
+   do
+   {
+      /* compress the data */
+      ret = deflate(&png_ptr->zstream, Z_NO_FLUSH);
+      if (ret != Z_OK)
+      {
+         /* error */
+         if (png_ptr->zstream.msg != NULL)
+            png_error(png_ptr, png_ptr->zstream.msg);
+         else
+            png_error(png_ptr, "zlib error");
+      }
+      /* check to see if we need more room */
+      if (!(png_ptr->zstream.avail_out))
+      {
+         /* make sure the output array has room */
+         if (comp->num_output_ptr >= comp->max_output_ptr)
+         {
+            int old_max;
+
+            old_max = comp->max_output_ptr;
+            comp->max_output_ptr = comp->num_output_ptr + 4;
+            if (comp->output_ptr != NULL)
+            {
+               png_charpp old_ptr;
+
+               old_ptr = comp->output_ptr;
+               comp->output_ptr = (png_charpp)png_malloc(png_ptr,
+                  (png_uint_32)(comp->max_output_ptr *
+                  png_sizeof (png_charpp)));
+               png_memcpy(comp->output_ptr, old_ptr, old_max
+                  * png_sizeof (png_charp));
+               png_free(png_ptr, old_ptr);
+            }
+            else
+               comp->output_ptr = (png_charpp)png_malloc(png_ptr,
+                  (png_uint_32)(comp->max_output_ptr *
+                  png_sizeof (png_charp)));
+         }
+
+         /* save the data */
+         comp->output_ptr[comp->num_output_ptr] = (png_charp)png_malloc(png_ptr,
+            (png_uint_32)png_ptr->zbuf_size);
+         png_memcpy(comp->output_ptr[comp->num_output_ptr], png_ptr->zbuf,
+            png_ptr->zbuf_size);
+         comp->num_output_ptr++;
+
+         /* and reset the buffer */
+         png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+         png_ptr->zstream.next_out = png_ptr->zbuf;
+      }
+   /* continue until we don't have any more to compress */
+   } while (png_ptr->zstream.avail_in);
+
+   /* finish the compression */
+   do
+   {
+      /* tell zlib we are finished */
+      ret = deflate(&png_ptr->zstream, Z_FINISH);
+
+      if (ret == Z_OK)
+      {
+         /* check to see if we need more room */
+         if (!(png_ptr->zstream.avail_out))
+         {
+            /* check to make sure our output array has room */
+            if (comp->num_output_ptr >= comp->max_output_ptr)
+            {
+               int old_max;
+
+               old_max = comp->max_output_ptr;
+               comp->max_output_ptr = comp->num_output_ptr + 4;
+               if (comp->output_ptr != NULL)
+               {
+                  png_charpp old_ptr;
+
+                  old_ptr = comp->output_ptr;
+                  /* This could be optimized to realloc() */
+                  comp->output_ptr = (png_charpp)png_malloc(png_ptr,
+                     (png_uint_32)(comp->max_output_ptr *
+                     png_sizeof (png_charpp)));
+                  png_memcpy(comp->output_ptr, old_ptr,
+                     old_max * png_sizeof (png_charp));
+                  png_free(png_ptr, old_ptr);
+               }
+               else
+                  comp->output_ptr = (png_charpp)png_malloc(png_ptr,
+                     (png_uint_32)(comp->max_output_ptr *
+                     png_sizeof (png_charp)));
+            }
+
+            /* save off the data */
+            comp->output_ptr[comp->num_output_ptr] =
+               (png_charp)png_malloc(png_ptr, (png_uint_32)png_ptr->zbuf_size);
+            png_memcpy(comp->output_ptr[comp->num_output_ptr], png_ptr->zbuf,
+               png_ptr->zbuf_size);
+            comp->num_output_ptr++;
+
+            /* and reset the buffer pointers */
+            png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+            png_ptr->zstream.next_out = png_ptr->zbuf;
+         }
+      }
+      else if (ret != Z_STREAM_END)
+      {
+         /* we got an error */
+         if (png_ptr->zstream.msg != NULL)
+            png_error(png_ptr, png_ptr->zstream.msg);
+         else
+            png_error(png_ptr, "zlib error");
+      }
+   } while (ret != Z_STREAM_END);
+
+   /* text length is number of buffers plus last buffer */
+   text_len = png_ptr->zbuf_size * comp->num_output_ptr;
+   if (png_ptr->zstream.avail_out < png_ptr->zbuf_size)
+      text_len += png_ptr->zbuf_size - (png_size_t)png_ptr->zstream.avail_out;
+
+   return((int)text_len);
+}
+
+/* ship the compressed text out via chunk writes */
+static void /* PRIVATE */
+png_write_compressed_data_out(png_structp png_ptr, compression_state *comp)
+{
+   int i;
+
+   /* handle the no-compression case */
+   if (comp->input)
+   {
+       png_write_chunk_data(png_ptr, (png_bytep)comp->input,
+                            (png_size_t)comp->input_len);
+       return;
+   }
+
+   /* write saved output buffers, if any */
+   for (i = 0; i < comp->num_output_ptr; i++)
+   {
+      png_write_chunk_data(png_ptr,(png_bytep)comp->output_ptr[i],
+         png_ptr->zbuf_size);
+      png_free(png_ptr, comp->output_ptr[i]);
+      comp->output_ptr[i]=NULL;
+   }
+   if (comp->max_output_ptr != 0)
+      png_free(png_ptr, comp->output_ptr);
+      comp->output_ptr=NULL;
+   /* write anything left in zbuf */
+   if (png_ptr->zstream.avail_out < (png_uint_32)png_ptr->zbuf_size)
+      png_write_chunk_data(png_ptr, png_ptr->zbuf,
+         png_ptr->zbuf_size - png_ptr->zstream.avail_out);
+
+   /* reset zlib for another zTXt/iTXt or image data */
+   deflateReset(&png_ptr->zstream);
+   png_ptr->zstream.data_type = Z_BINARY;
+}
+#endif
+
+/* Write the IHDR chunk, and update the png_struct with the necessary
+ * information.  Note that the rest of this code depends upon this
+ * information being correct.
+ */
+void /* PRIVATE */
+png_write_IHDR(png_structp png_ptr, png_uint_32 width, png_uint_32 height,
+   int bit_depth, int color_type, int compression_type, int filter_type,
+   int interlace_type)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_IHDR;
+#endif
+   png_byte buf[13]; /* buffer to store the IHDR info */
+
+   png_debug(1, "in png_write_IHDR\n");
+   /* Check that we have valid input data from the application info */
+   switch (color_type)
+   {
+      case PNG_COLOR_TYPE_GRAY:
+         switch (bit_depth)
+         {
+            case 1:
+            case 2:
+            case 4:
+            case 8:
+            case 16: png_ptr->channels = 1; break;
+            default: png_error(png_ptr,"Invalid bit depth for grayscale image");
+         }
+         break;
+      case PNG_COLOR_TYPE_RGB:
+         if (bit_depth != 8 && bit_depth != 16)
+            png_error(png_ptr, "Invalid bit depth for RGB image");
+         png_ptr->channels = 3;
+         break;
+      case PNG_COLOR_TYPE_PALETTE:
+         switch (bit_depth)
+         {
+            case 1:
+            case 2:
+            case 4:
+            case 8: png_ptr->channels = 1; break;
+            default: png_error(png_ptr, "Invalid bit depth for paletted image");
+         }
+         break;
+      case PNG_COLOR_TYPE_GRAY_ALPHA:
+         if (bit_depth != 8 && bit_depth != 16)
+            png_error(png_ptr, "Invalid bit depth for grayscale+alpha image");
+         png_ptr->channels = 2;
+         break;
+      case PNG_COLOR_TYPE_RGB_ALPHA:
+         if (bit_depth != 8 && bit_depth != 16)
+            png_error(png_ptr, "Invalid bit depth for RGBA image");
+         png_ptr->channels = 4;
+         break;
+      default:
+         png_error(png_ptr, "Invalid image color type specified");
+   }
+
+   if (compression_type != PNG_COMPRESSION_TYPE_BASE)
+   {
+      png_warning(png_ptr, "Invalid compression type specified");
+      compression_type = PNG_COMPRESSION_TYPE_BASE;
+   }
+
+   /* Write filter_method 64 (intrapixel differencing) only if
+    * 1. Libpng was compiled with PNG_MNG_FEATURES_SUPPORTED and
+    * 2. Libpng did not write a PNG signature (this filter_method is only
+    *    used in PNG datastreams that are embedded in MNG datastreams) and
+    * 3. The application called png_permit_mng_features with a mask that
+    *    included PNG_FLAG_MNG_FILTER_64 and
+    * 4. The filter_method is 64 and
+    * 5. The color_type is RGB or RGBA
+    */
+   if (
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+      !((png_ptr->mng_features_permitted & PNG_FLAG_MNG_FILTER_64) &&
+      ((png_ptr->mode&PNG_HAVE_PNG_SIGNATURE) == 0) &&
+      (color_type == PNG_COLOR_TYPE_RGB ||
+       color_type == PNG_COLOR_TYPE_RGB_ALPHA) &&
+      (filter_type == PNG_INTRAPIXEL_DIFFERENCING)) &&
+#endif
+      filter_type != PNG_FILTER_TYPE_BASE)
+   {
+      png_warning(png_ptr, "Invalid filter type specified");
+      filter_type = PNG_FILTER_TYPE_BASE;
+   }
+
+#ifdef PNG_WRITE_INTERLACING_SUPPORTED
+   if (interlace_type != PNG_INTERLACE_NONE &&
+      interlace_type != PNG_INTERLACE_ADAM7)
+   {
+      png_warning(png_ptr, "Invalid interlace type specified");
+      interlace_type = PNG_INTERLACE_ADAM7;
+   }
+#else
+   interlace_type=PNG_INTERLACE_NONE;
+#endif
+
+   /* save off the relevent information */
+   png_ptr->bit_depth = (png_byte)bit_depth;
+   png_ptr->color_type = (png_byte)color_type;
+   png_ptr->interlaced = (png_byte)interlace_type;
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+   png_ptr->filter_type = (png_byte)filter_type;
+#endif
+   png_ptr->compression_type = (png_byte)compression_type;
+   png_ptr->width = width;
+   png_ptr->height = height;
+
+   png_ptr->pixel_depth = (png_byte)(bit_depth * png_ptr->channels);
+   png_ptr->rowbytes = PNG_ROWBYTES(png_ptr->pixel_depth, width);
+   /* set the usr info, so any transformations can modify it */
+   png_ptr->usr_width = png_ptr->width;
+   png_ptr->usr_bit_depth = png_ptr->bit_depth;
+   png_ptr->usr_channels = png_ptr->channels;
+
+   /* pack the header information into the buffer */
+   png_save_uint_32(buf, width);
+   png_save_uint_32(buf + 4, height);
+   buf[8] = (png_byte)bit_depth;
+   buf[9] = (png_byte)color_type;
+   buf[10] = (png_byte)compression_type;
+   buf[11] = (png_byte)filter_type;
+   buf[12] = (png_byte)interlace_type;
+
+   /* write the chunk */
+   png_write_chunk(png_ptr, (png_bytep)png_IHDR, buf, (png_size_t)13);
+
+   /* initialize zlib with PNG info */
+   png_ptr->zstream.zalloc = png_zalloc;
+   png_ptr->zstream.zfree = png_zfree;
+   png_ptr->zstream.opaque = (voidpf)png_ptr;
+   if (!(png_ptr->do_filter))
+   {
+      if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE ||
+         png_ptr->bit_depth < 8)
+         png_ptr->do_filter = PNG_FILTER_NONE;
+      else
+         png_ptr->do_filter = PNG_ALL_FILTERS;
+   }
+   if (!(png_ptr->flags & PNG_FLAG_ZLIB_CUSTOM_STRATEGY))
+   {
+      if (png_ptr->do_filter != PNG_FILTER_NONE)
+         png_ptr->zlib_strategy = Z_FILTERED;
+      else
+         png_ptr->zlib_strategy = Z_DEFAULT_STRATEGY;
+   }
+   if (!(png_ptr->flags & PNG_FLAG_ZLIB_CUSTOM_LEVEL))
+      png_ptr->zlib_level = Z_DEFAULT_COMPRESSION;
+   if (!(png_ptr->flags & PNG_FLAG_ZLIB_CUSTOM_MEM_LEVEL))
+      png_ptr->zlib_mem_level = 8;
+   if (!(png_ptr->flags & PNG_FLAG_ZLIB_CUSTOM_WINDOW_BITS))
+      png_ptr->zlib_window_bits = 15;
+   if (!(png_ptr->flags & PNG_FLAG_ZLIB_CUSTOM_METHOD))
+      png_ptr->zlib_method = 8;
+   if (deflateInit2(&png_ptr->zstream, png_ptr->zlib_level,
+      png_ptr->zlib_method, png_ptr->zlib_window_bits,
+      png_ptr->zlib_mem_level, png_ptr->zlib_strategy) != Z_OK)
+       png_error(png_ptr, "zlib failed to initialize compressor");
+   png_ptr->zstream.next_out = png_ptr->zbuf;
+   png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+   /* libpng is not interested in zstream.data_type */
+   /* set it to a predefined value, to avoid its evaluation inside zlib */
+   png_ptr->zstream.data_type = Z_BINARY;
+
+   png_ptr->mode = PNG_HAVE_IHDR;
+}
+
+/* write the palette.  We are careful not to trust png_color to be in the
+ * correct order for PNG, so people can redefine it to any convenient
+ * structure.
+ */
+void /* PRIVATE */
+png_write_PLTE(png_structp png_ptr, png_colorp palette, png_uint_32 num_pal)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_PLTE;
+#endif
+   png_uint_32 i;
+   png_colorp pal_ptr;
+   png_byte buf[3];
+
+   png_debug(1, "in png_write_PLTE\n");
+   if ((
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+        !(png_ptr->mng_features_permitted & PNG_FLAG_MNG_EMPTY_PLTE) &&
+#endif
+        num_pal == 0) || num_pal > 256)
+   {
+     if (png_ptr->color_type == PNG_COLOR_TYPE_PALETTE)
+     {
+        png_error(png_ptr, "Invalid number of colors in palette");
+     }
+     else
+     {
+        png_warning(png_ptr, "Invalid number of colors in palette");
+        return;
+     }
+   }
+
+   if (!(png_ptr->color_type&PNG_COLOR_MASK_COLOR))
+   {
+      png_warning(png_ptr,
+        "Ignoring request to write a PLTE chunk in grayscale PNG");
+      return;
+   }
+
+   png_ptr->num_palette = (png_uint_16)num_pal;
+   png_debug1(3, "num_palette = %d\n", png_ptr->num_palette);
+
+   png_write_chunk_start(png_ptr, png_PLTE, num_pal * 3);
+#ifndef PNG_NO_POINTER_INDEXING
+   for (i = 0, pal_ptr = palette; i < num_pal; i++, pal_ptr++)
+   {
+      buf[0] = pal_ptr->red;
+      buf[1] = pal_ptr->green;
+      buf[2] = pal_ptr->blue;
+      png_write_chunk_data(png_ptr, buf, (png_size_t)3);
+   }
+#else
+   /* This is a little slower but some buggy compilers need to do this instead */
+   pal_ptr=palette;
+   for (i = 0; i < num_pal; i++)
+   {
+      buf[0] = pal_ptr[i].red;
+      buf[1] = pal_ptr[i].green;
+      buf[2] = pal_ptr[i].blue;
+      png_write_chunk_data(png_ptr, buf, (png_size_t)3);
+   }
+#endif
+   png_write_chunk_end(png_ptr);
+   png_ptr->mode |= PNG_HAVE_PLTE;
+}
+
+/* write an IDAT chunk */
+void /* PRIVATE */
+png_write_IDAT(png_structp png_ptr, png_bytep data, png_size_t length)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_IDAT;
+#endif
+   png_debug(1, "in png_write_IDAT\n");
+
+   /* Optimize the CMF field in the zlib stream. */
+   /* This hack of the zlib stream is compliant to the stream specification. */
+   if (!(png_ptr->mode & PNG_HAVE_IDAT) &&
+       png_ptr->compression_type == PNG_COMPRESSION_TYPE_BASE)
+   {
+      unsigned int z_cmf = data[0];  /* zlib compression method and flags */
+      if ((z_cmf & 0x0f) == 8 && (z_cmf & 0xf0) <= 0x70)
+      {
+         /* Avoid memory underflows and multiplication overflows. */
+         /* The conditions below are practically always satisfied;
+            however, they still must be checked. */
+         if (length >= 2 &&
+             png_ptr->height < 16384 && png_ptr->width < 16384)
+         {
+            png_uint_32 uncompressed_idat_size = png_ptr->height *
+               ((png_ptr->width *
+               png_ptr->channels * png_ptr->bit_depth + 15) >> 3);
+            unsigned int z_cinfo = z_cmf >> 4;
+            unsigned int half_z_window_size = 1 << (z_cinfo + 7);
+            while (uncompressed_idat_size <= half_z_window_size &&
+                   half_z_window_size >= 256)
+            {
+               z_cinfo--;
+               half_z_window_size >>= 1;
+            }
+            z_cmf = (z_cmf & 0x0f) | (z_cinfo << 4);
+            if (data[0] != (png_byte)z_cmf)
+            {
+               data[0] = (png_byte)z_cmf;
+               data[1] &= 0xe0;
+               data[1] += (png_byte)(0x1f - ((z_cmf << 8) + data[1]) % 0x1f);
+            }
+         }
+      }
+      else
+         png_error(png_ptr,
+            "Invalid zlib compression method or flags in IDAT");
+   }
+
+   png_write_chunk(png_ptr, (png_bytep)png_IDAT, data, length);
+   png_ptr->mode |= PNG_HAVE_IDAT;
+}
+
+/* write an IEND chunk */
+void /* PRIVATE */
+png_write_IEND(png_structp png_ptr)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_IEND;
+#endif
+   png_debug(1, "in png_write_IEND\n");
+   png_write_chunk(png_ptr, (png_bytep)png_IEND, png_bytep_NULL,
+     (png_size_t)0);
+   png_ptr->mode |= PNG_HAVE_IEND;
+}
+
+#if defined(PNG_WRITE_gAMA_SUPPORTED)
+/* write a gAMA chunk */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+void /* PRIVATE */
+png_write_gAMA(png_structp png_ptr, double file_gamma)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_gAMA;
+#endif
+   png_uint_32 igamma;
+   png_byte buf[4];
+
+   png_debug(1, "in png_write_gAMA\n");
+   /* file_gamma is saved in 1/100,000ths */
+   igamma = (png_uint_32)(file_gamma * 100000.0 + 0.5);
+   png_save_uint_32(buf, igamma);
+   png_write_chunk(png_ptr, (png_bytep)png_gAMA, buf, (png_size_t)4);
+}
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+void /* PRIVATE */
+png_write_gAMA_fixed(png_structp png_ptr, png_fixed_point file_gamma)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_gAMA;
+#endif
+   png_byte buf[4];
+
+   png_debug(1, "in png_write_gAMA\n");
+   /* file_gamma is saved in 1/100,000ths */
+   png_save_uint_32(buf, (png_uint_32)file_gamma);
+   png_write_chunk(png_ptr, (png_bytep)png_gAMA, buf, (png_size_t)4);
+}
+#endif
+#endif
+
+#if defined(PNG_WRITE_sRGB_SUPPORTED)
+/* write a sRGB chunk */
+void /* PRIVATE */
+png_write_sRGB(png_structp png_ptr, int srgb_intent)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_sRGB;
+#endif
+   png_byte buf[1];
+
+   png_debug(1, "in png_write_sRGB\n");
+   if(srgb_intent >= PNG_sRGB_INTENT_LAST)
+         png_warning(png_ptr,
+            "Invalid sRGB rendering intent specified");
+   buf[0]=(png_byte)srgb_intent;
+   png_write_chunk(png_ptr, (png_bytep)png_sRGB, buf, (png_size_t)1);
+}
+#endif
+
+#if defined(PNG_WRITE_iCCP_SUPPORTED)
+/* write an iCCP chunk */
+void /* PRIVATE */
+png_write_iCCP(png_structp png_ptr, png_charp name, int compression_type,
+   png_charp profile, int profile_len)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_iCCP;
+#endif
+   png_size_t name_len;
+   png_charp new_name;
+   compression_state comp;
+   int embedded_profile_len = 0;
+
+   png_debug(1, "in png_write_iCCP\n");
+
+   comp.num_output_ptr = 0;
+   comp.max_output_ptr = 0;
+   comp.output_ptr = NULL;
+   comp.input = NULL;
+   comp.input_len = 0;
+
+   if (name == NULL || (name_len = png_check_keyword(png_ptr, name,
+      &new_name)) == 0)
+   {
+      png_warning(png_ptr, "Empty keyword in iCCP chunk");
+      return;
+   }
+
+   if (compression_type != PNG_COMPRESSION_TYPE_BASE)
+      png_warning(png_ptr, "Unknown compression type in iCCP chunk");
+
+   if (profile == NULL)
+      profile_len = 0;
+
+   if (profile_len > 3)
+      embedded_profile_len =
+          ((*( (png_bytep)profile  ))<<24) |
+          ((*( (png_bytep)profile+1))<<16) |
+          ((*( (png_bytep)profile+2))<< 8) |
+          ((*( (png_bytep)profile+3))    );
+
+   if (profile_len < embedded_profile_len)
+     {
+        png_warning(png_ptr,
+          "Embedded profile length too large in iCCP chunk");
+        return;
+     }
+
+   if (profile_len > embedded_profile_len)
+     {
+        png_warning(png_ptr,
+          "Truncating profile to actual length in iCCP chunk");
+        profile_len = embedded_profile_len;
+     }
+
+   if (profile_len)
+       profile_len = png_text_compress(png_ptr, profile, (png_size_t)profile_len,
+          PNG_COMPRESSION_TYPE_BASE, &comp);
+
+   /* make sure we include the NULL after the name and the compression type */
+   png_write_chunk_start(png_ptr, png_iCCP,
+          (png_uint_32)name_len+profile_len+2);
+   new_name[name_len+1]=0x00;
+   png_write_chunk_data(png_ptr, (png_bytep)new_name, name_len + 2);
+
+   if (profile_len)
+      png_write_compressed_data_out(png_ptr, &comp);
+
+   png_write_chunk_end(png_ptr);
+   png_free(png_ptr, new_name);
+}
+#endif
+
+#if defined(PNG_WRITE_sPLT_SUPPORTED)
+/* write a sPLT chunk */
+void /* PRIVATE */
+png_write_sPLT(png_structp png_ptr, png_sPLT_tp spalette)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_sPLT;
+#endif
+   png_size_t name_len;
+   png_charp new_name;
+   png_byte entrybuf[10];
+   int entry_size = (spalette->depth == 8 ? 6 : 10);
+   int palette_size = entry_size * spalette->nentries;
+   png_sPLT_entryp ep;
+#ifdef PNG_NO_POINTER_INDEXING
+   int i;
+#endif
+
+   png_debug(1, "in png_write_sPLT\n");
+   if (spalette->name == NULL || (name_len = png_check_keyword(png_ptr,
+      spalette->name, &new_name))==0)
+   {
+      png_warning(png_ptr, "Empty keyword in sPLT chunk");
+      return;
+   }
+
+   /* make sure we include the NULL after the name */
+   png_write_chunk_start(png_ptr, png_sPLT,
+          (png_uint_32)(name_len + 2 + palette_size));
+   png_write_chunk_data(png_ptr, (png_bytep)new_name, name_len + 1);
+   png_write_chunk_data(png_ptr, (png_bytep)&spalette->depth, 1);
+
+   /* loop through each palette entry, writing appropriately */
+#ifndef PNG_NO_POINTER_INDEXING
+   for (ep = spalette->entries; ep<spalette->entries+spalette->nentries; ep++)
+   {
+       if (spalette->depth == 8)
+       {
+           entrybuf[0] = (png_byte)ep->red;
+           entrybuf[1] = (png_byte)ep->green;
+           entrybuf[2] = (png_byte)ep->blue;
+           entrybuf[3] = (png_byte)ep->alpha;
+           png_save_uint_16(entrybuf + 4, ep->frequency);
+       }
+       else
+       {
+           png_save_uint_16(entrybuf + 0, ep->red);
+           png_save_uint_16(entrybuf + 2, ep->green);
+           png_save_uint_16(entrybuf + 4, ep->blue);
+           png_save_uint_16(entrybuf + 6, ep->alpha);
+           png_save_uint_16(entrybuf + 8, ep->frequency);
+       }
+       png_write_chunk_data(png_ptr, entrybuf, (png_size_t)entry_size);
+   }
+#else
+   ep=spalette->entries;
+   for (i=0; i>spalette->nentries; i++)
+   {
+       if (spalette->depth == 8)
+       {
+           entrybuf[0] = (png_byte)ep[i].red;
+           entrybuf[1] = (png_byte)ep[i].green;
+           entrybuf[2] = (png_byte)ep[i].blue;
+           entrybuf[3] = (png_byte)ep[i].alpha;
+           png_save_uint_16(entrybuf + 4, ep[i].frequency);
+       }
+       else
+       {
+           png_save_uint_16(entrybuf + 0, ep[i].red);
+           png_save_uint_16(entrybuf + 2, ep[i].green);
+           png_save_uint_16(entrybuf + 4, ep[i].blue);
+           png_save_uint_16(entrybuf + 6, ep[i].alpha);
+           png_save_uint_16(entrybuf + 8, ep[i].frequency);
+       }
+       png_write_chunk_data(png_ptr, entrybuf, entry_size);
+   }
+#endif
+
+   png_write_chunk_end(png_ptr);
+   png_free(png_ptr, new_name);
+}
+#endif
+
+#if defined(PNG_WRITE_sBIT_SUPPORTED)
+/* write the sBIT chunk */
+void /* PRIVATE */
+png_write_sBIT(png_structp png_ptr, png_color_8p sbit, int color_type)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_sBIT;
+#endif
+   png_byte buf[4];
+   png_size_t size;
+
+   png_debug(1, "in png_write_sBIT\n");
+   /* make sure we don't depend upon the order of PNG_COLOR_8 */
+   if (color_type & PNG_COLOR_MASK_COLOR)
+   {
+      png_byte maxbits;
+
+      maxbits = (png_byte)(color_type==PNG_COLOR_TYPE_PALETTE ? 8 :
+                png_ptr->usr_bit_depth);
+      if (sbit->red == 0 || sbit->red > maxbits ||
+          sbit->green == 0 || sbit->green > maxbits ||
+          sbit->blue == 0 || sbit->blue > maxbits)
+      {
+         png_warning(png_ptr, "Invalid sBIT depth specified");
+         return;
+      }
+      buf[0] = sbit->red;
+      buf[1] = sbit->green;
+      buf[2] = sbit->blue;
+      size = 3;
+   }
+   else
+   {
+      if (sbit->gray == 0 || sbit->gray > png_ptr->usr_bit_depth)
+      {
+         png_warning(png_ptr, "Invalid sBIT depth specified");
+         return;
+      }
+      buf[0] = sbit->gray;
+      size = 1;
+   }
+
+   if (color_type & PNG_COLOR_MASK_ALPHA)
+   {
+      if (sbit->alpha == 0 || sbit->alpha > png_ptr->usr_bit_depth)
+      {
+         png_warning(png_ptr, "Invalid sBIT depth specified");
+         return;
+      }
+      buf[size++] = sbit->alpha;
+   }
+
+   png_write_chunk(png_ptr, (png_bytep)png_sBIT, buf, size);
+}
+#endif
+
+#if defined(PNG_WRITE_cHRM_SUPPORTED)
+/* write the cHRM chunk */
+#ifdef PNG_FLOATING_POINT_SUPPORTED
+void /* PRIVATE */
+png_write_cHRM(png_structp png_ptr, double white_x, double white_y,
+   double red_x, double red_y, double green_x, double green_y,
+   double blue_x, double blue_y)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_cHRM;
+#endif
+   png_byte buf[32];
+   png_uint_32 itemp;
+
+   png_debug(1, "in png_write_cHRM\n");
+   /* each value is saved in 1/100,000ths */
+   if (white_x < 0 || white_x > 0.8 || white_y < 0 || white_y > 0.8 ||
+       white_x + white_y > 1.0)
+   {
+      png_warning(png_ptr, "Invalid cHRM white point specified");
+#if !defined(PNG_NO_CONSOLE_IO)
+      fprintf(stderr,"white_x=%f, white_y=%f\n",white_x, white_y);
+#endif
+      return;
+   }
+   itemp = (png_uint_32)(white_x * 100000.0 + 0.5);
+   png_save_uint_32(buf, itemp);
+   itemp = (png_uint_32)(white_y * 100000.0 + 0.5);
+   png_save_uint_32(buf + 4, itemp);
+
+   if (red_x < 0 ||  red_y < 0 || red_x + red_y > 1.0)
+   {
+      png_warning(png_ptr, "Invalid cHRM red point specified");
+      return;
+   }
+   itemp = (png_uint_32)(red_x * 100000.0 + 0.5);
+   png_save_uint_32(buf + 8, itemp);
+   itemp = (png_uint_32)(red_y * 100000.0 + 0.5);
+   png_save_uint_32(buf + 12, itemp);
+
+   if (green_x < 0 || green_y < 0 || green_x + green_y > 1.0)
+   {
+      png_warning(png_ptr, "Invalid cHRM green point specified");
+      return;
+   }
+   itemp = (png_uint_32)(green_x * 100000.0 + 0.5);
+   png_save_uint_32(buf + 16, itemp);
+   itemp = (png_uint_32)(green_y * 100000.0 + 0.5);
+   png_save_uint_32(buf + 20, itemp);
+
+   if (blue_x < 0 || blue_y < 0 || blue_x + blue_y > 1.0)
+   {
+      png_warning(png_ptr, "Invalid cHRM blue point specified");
+      return;
+   }
+   itemp = (png_uint_32)(blue_x * 100000.0 + 0.5);
+   png_save_uint_32(buf + 24, itemp);
+   itemp = (png_uint_32)(blue_y * 100000.0 + 0.5);
+   png_save_uint_32(buf + 28, itemp);
+
+   png_write_chunk(png_ptr, (png_bytep)png_cHRM, buf, (png_size_t)32);
+}
+#endif
+#ifdef PNG_FIXED_POINT_SUPPORTED
+void /* PRIVATE */
+png_write_cHRM_fixed(png_structp png_ptr, png_fixed_point white_x,
+   png_fixed_point white_y, png_fixed_point red_x, png_fixed_point red_y,
+   png_fixed_point green_x, png_fixed_point green_y, png_fixed_point blue_x,
+   png_fixed_point blue_y)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_cHRM;
+#endif
+   png_byte buf[32];
+
+   png_debug(1, "in png_write_cHRM\n");
+   /* each value is saved in 1/100,000ths */
+   if (white_x > 80000L || white_y > 80000L || white_x + white_y > 100000L)
+   {
+      png_warning(png_ptr, "Invalid fixed cHRM white point specified");
+#if !defined(PNG_NO_CONSOLE_IO)
+      fprintf(stderr,"white_x=%ld, white_y=%ld\n",white_x, white_y);
+#endif
+      return;
+   }
+   png_save_uint_32(buf, (png_uint_32)white_x);
+   png_save_uint_32(buf + 4, (png_uint_32)white_y);
+
+   if (red_x + red_y > 100000L)
+   {
+      png_warning(png_ptr, "Invalid cHRM fixed red point specified");
+      return;
+   }
+   png_save_uint_32(buf + 8, (png_uint_32)red_x);
+   png_save_uint_32(buf + 12, (png_uint_32)red_y);
+
+   if (green_x + green_y > 100000L)
+   {
+      png_warning(png_ptr, "Invalid fixed cHRM green point specified");
+      return;
+   }
+   png_save_uint_32(buf + 16, (png_uint_32)green_x);
+   png_save_uint_32(buf + 20, (png_uint_32)green_y);
+
+   if (blue_x + blue_y > 100000L)
+   {
+      png_warning(png_ptr, "Invalid fixed cHRM blue point specified");
+      return;
+   }
+   png_save_uint_32(buf + 24, (png_uint_32)blue_x);
+   png_save_uint_32(buf + 28, (png_uint_32)blue_y);
+
+   png_write_chunk(png_ptr, (png_bytep)png_cHRM, buf, (png_size_t)32);
+}
+#endif
+#endif
+
+#if defined(PNG_WRITE_tRNS_SUPPORTED)
+/* write the tRNS chunk */
+void /* PRIVATE */
+png_write_tRNS(png_structp png_ptr, png_bytep trans, png_color_16p tran,
+   int num_trans, int color_type)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_tRNS;
+#endif
+   png_byte buf[6];
+
+   png_debug(1, "in png_write_tRNS\n");
+   if (color_type == PNG_COLOR_TYPE_PALETTE)
+   {
+      if (num_trans <= 0 || num_trans > (int)png_ptr->num_palette)
+      {
+         png_warning(png_ptr,"Invalid number of transparent colors specified");
+         return;
+      }
+      /* write the chunk out as it is */
+      png_write_chunk(png_ptr, (png_bytep)png_tRNS, trans, (png_size_t)num_trans);
+   }
+   else if (color_type == PNG_COLOR_TYPE_GRAY)
+   {
+      /* one 16 bit value */
+      if(tran->gray >= (1 << png_ptr->bit_depth))
+      {
+         png_warning(png_ptr,
+           "Ignoring attempt to write tRNS chunk out-of-range for bit_depth");
+         return;
+      }
+      png_save_uint_16(buf, tran->gray);
+      png_write_chunk(png_ptr, (png_bytep)png_tRNS, buf, (png_size_t)2);
+   }
+   else if (color_type == PNG_COLOR_TYPE_RGB)
+   {
+      /* three 16 bit values */
+      png_save_uint_16(buf, tran->red);
+      png_save_uint_16(buf + 2, tran->green);
+      png_save_uint_16(buf + 4, tran->blue);
+      if(png_ptr->bit_depth == 8 && (buf[0] | buf[2] | buf[4]))
+         {
+            png_warning(png_ptr,
+              "Ignoring attempt to write 16-bit tRNS chunk when bit_depth is 8");
+            return;
+         }
+      png_write_chunk(png_ptr, (png_bytep)png_tRNS, buf, (png_size_t)6);
+   }
+   else
+   {
+      png_warning(png_ptr, "Can't write tRNS with an alpha channel");
+   }
+}
+#endif
+
+#if defined(PNG_WRITE_bKGD_SUPPORTED)
+/* write the background chunk */
+void /* PRIVATE */
+png_write_bKGD(png_structp png_ptr, png_color_16p back, int color_type)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_bKGD;
+#endif
+   png_byte buf[6];
+
+   png_debug(1, "in png_write_bKGD\n");
+   if (color_type == PNG_COLOR_TYPE_PALETTE)
+   {
+      if (
+#if defined(PNG_MNG_FEATURES_SUPPORTED)
+          (png_ptr->num_palette ||
+          (!(png_ptr->mng_features_permitted & PNG_FLAG_MNG_EMPTY_PLTE))) &&
+#endif
+         back->index > png_ptr->num_palette)
+      {
+         png_warning(png_ptr, "Invalid background palette index");
+         return;
+      }
+      buf[0] = back->index;
+      png_write_chunk(png_ptr, (png_bytep)png_bKGD, buf, (png_size_t)1);
+   }
+   else if (color_type & PNG_COLOR_MASK_COLOR)
+   {
+      png_save_uint_16(buf, back->red);
+      png_save_uint_16(buf + 2, back->green);
+      png_save_uint_16(buf + 4, back->blue);
+      if(png_ptr->bit_depth == 8 && (buf[0] | buf[2] | buf[4]))
+         {
+            png_warning(png_ptr,
+              "Ignoring attempt to write 16-bit bKGD chunk when bit_depth is 8");
+            return;
+         }
+      png_write_chunk(png_ptr, (png_bytep)png_bKGD, buf, (png_size_t)6);
+   }
+   else
+   {
+      if(back->gray >= (1 << png_ptr->bit_depth))
+      {
+         png_warning(png_ptr,
+           "Ignoring attempt to write bKGD chunk out-of-range for bit_depth");
+         return;
+      }
+      png_save_uint_16(buf, back->gray);
+      png_write_chunk(png_ptr, (png_bytep)png_bKGD, buf, (png_size_t)2);
+   }
+}
+#endif
+
+#if defined(PNG_WRITE_hIST_SUPPORTED)
+/* write the histogram */
+void /* PRIVATE */
+png_write_hIST(png_structp png_ptr, png_uint_16p hist, int num_hist)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_hIST;
+#endif
+   int i;
+   png_byte buf[3];
+
+   png_debug(1, "in png_write_hIST\n");
+   if (num_hist > (int)png_ptr->num_palette)
+   {
+      png_debug2(3, "num_hist = %d, num_palette = %d\n", num_hist,
+         png_ptr->num_palette);
+      png_warning(png_ptr, "Invalid number of histogram entries specified");
+      return;
+   }
+
+   png_write_chunk_start(png_ptr, png_hIST, (png_uint_32)(num_hist * 2));
+   for (i = 0; i < num_hist; i++)
+   {
+      png_save_uint_16(buf, hist[i]);
+      png_write_chunk_data(png_ptr, buf, (png_size_t)2);
+   }
+   png_write_chunk_end(png_ptr);
+}
+#endif
+
+#if defined(PNG_WRITE_TEXT_SUPPORTED) || defined(PNG_WRITE_pCAL_SUPPORTED) || \
+    defined(PNG_WRITE_iCCP_SUPPORTED) || defined(PNG_WRITE_sPLT_SUPPORTED)
+/* Check that the tEXt or zTXt keyword is valid per PNG 1.0 specification,
+ * and if invalid, correct the keyword rather than discarding the entire
+ * chunk.  The PNG 1.0 specification requires keywords 1-79 characters in
+ * length, forbids leading or trailing whitespace, multiple internal spaces,
+ * and the non-break space (0x80) from ISO 8859-1.  Returns keyword length.
+ *
+ * The new_key is allocated to hold the corrected keyword and must be freed
+ * by the calling routine.  This avoids problems with trying to write to
+ * static keywords without having to have duplicate copies of the strings.
+ */
+png_size_t /* PRIVATE */
+png_check_keyword(png_structp png_ptr, png_charp key, png_charpp new_key)
+{
+   png_size_t key_len;
+   png_charp kp, dp;
+   int kflag;
+   int kwarn=0;
+
+   png_debug(1, "in png_check_keyword\n");
+   *new_key = NULL;
+
+   if (key == NULL || (key_len = png_strlen(key)) == 0)
+   {
+      png_warning(png_ptr, "zero length keyword");
+      return ((png_size_t)0);
+   }
+
+   png_debug1(2, "Keyword to be checked is '%s'\n", key);
+
+   *new_key = (png_charp)png_malloc_warn(png_ptr, (png_uint_32)(key_len + 2));
+   if (*new_key == NULL)
+   {
+      png_warning(png_ptr, "Out of memory while procesing keyword");
+      return ((png_size_t)0);
+   }
+
+   /* Replace non-printing characters with a blank and print a warning */
+   for (kp = key, dp = *new_key; *kp != '\0'; kp++, dp++)
+   {
+      if ((png_byte)*kp < 0x20 ||
+         ((png_byte)*kp > 0x7E && (png_byte)*kp < 0xA1))
+      {
+#if !defined(PNG_NO_STDIO) && !defined(_WIN32_WCE)
+         char msg[40];
+
+         png_snprintf(msg, 40,
+           "invalid keyword character 0x%02X", (png_byte)*kp);
+         png_warning(png_ptr, msg);
+#else
+         png_warning(png_ptr, "invalid character in keyword");
+#endif
+         *dp = ' ';
+      }
+      else
+      {
+         *dp = *kp;
+      }
+   }
+   *dp = '\0';
+
+   /* Remove any trailing white space. */
+   kp = *new_key + key_len - 1;
+   if (*kp == ' ')
+   {
+      png_warning(png_ptr, "trailing spaces removed from keyword");
+
+      while (*kp == ' ')
+      {
+        *(kp--) = '\0';
+        key_len--;
+      }
+   }
+
+   /* Remove any leading white space. */
+   kp = *new_key;
+   if (*kp == ' ')
+   {
+      png_warning(png_ptr, "leading spaces removed from keyword");
+
+      while (*kp == ' ')
+      {
+        kp++;
+        key_len--;
+      }
+   }
+
+   png_debug1(2, "Checking for multiple internal spaces in '%s'\n", kp);
+
+   /* Remove multiple internal spaces. */
+   for (kflag = 0, dp = *new_key; *kp != '\0'; kp++)
+   {
+      if (*kp == ' ' && kflag == 0)
+      {
+         *(dp++) = *kp;
+         kflag = 1;
+      }
+      else if (*kp == ' ')
+      {
+         key_len--;
+         kwarn=1;
+      }
+      else
+      {
+         *(dp++) = *kp;
+         kflag = 0;
+      }
+   }
+   *dp = '\0';
+   if(kwarn)
+      png_warning(png_ptr, "extra interior spaces removed from keyword");
+
+   if (key_len == 0)
+   {
+      png_free(png_ptr, *new_key);
+      *new_key=NULL;
+      png_warning(png_ptr, "Zero length keyword");
+   }
+
+   if (key_len > 79)
+   {
+      png_warning(png_ptr, "keyword length must be 1 - 79 characters");
+      new_key[79] = '\0';
+      key_len = 79;
+   }
+
+   return (key_len);
+}
+#endif
+
+#if defined(PNG_WRITE_tEXt_SUPPORTED)
+/* write a tEXt chunk */
+void /* PRIVATE */
+png_write_tEXt(png_structp png_ptr, png_charp key, png_charp text,
+   png_size_t text_len)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_tEXt;
+#endif
+   png_size_t key_len;
+   png_charp new_key;
+
+   png_debug(1, "in png_write_tEXt\n");
+   if (key == NULL || (key_len = png_check_keyword(png_ptr, key, &new_key))==0)
+   {
+      png_warning(png_ptr, "Empty keyword in tEXt chunk");
+      return;
+   }
+
+   if (text == NULL || *text == '\0')
+      text_len = 0;
+   else
+      text_len = png_strlen(text);
+
+   /* make sure we include the 0 after the key */
+   png_write_chunk_start(png_ptr, (png_bytep)png_tEXt, (png_uint_32)key_len+text_len+1);
+   /*
+    * We leave it to the application to meet PNG-1.0 requirements on the
+    * contents of the text.  PNG-1.0 through PNG-1.2 discourage the use of
+    * any non-Latin-1 characters except for NEWLINE.  ISO PNG will forbid them.
+    * The NUL character is forbidden by PNG-1.0 through PNG-1.2 and ISO PNG.
+    */
+   png_write_chunk_data(png_ptr, (png_bytep)new_key, key_len + 1);
+   if (text_len)
+      png_write_chunk_data(png_ptr, (png_bytep)text, text_len);
+
+   png_write_chunk_end(png_ptr);
+   png_free(png_ptr, new_key);
+}
+#endif
+
+#if defined(PNG_WRITE_zTXt_SUPPORTED)
+/* write a compressed text chunk */
+void /* PRIVATE */
+png_write_zTXt(png_structp png_ptr, png_charp key, png_charp text,
+   png_size_t text_len, int compression)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_zTXt;
+#endif
+   png_size_t key_len;
+   char buf[1];
+   png_charp new_key;
+   compression_state comp;
+
+   png_debug(1, "in png_write_zTXt\n");
+
+   comp.num_output_ptr = 0;
+   comp.max_output_ptr = 0;
+   comp.output_ptr = NULL;
+   comp.input = NULL;
+   comp.input_len = 0;
+
+   if (key == NULL || (key_len = png_check_keyword(png_ptr, key, &new_key))==0)
+   {
+      png_warning(png_ptr, "Empty keyword in zTXt chunk");
+      return;
+   }
+
+   if (text == NULL || *text == '\0' || compression==PNG_TEXT_COMPRESSION_NONE)
+   {
+      png_write_tEXt(png_ptr, new_key, text, (png_size_t)0);
+      png_free(png_ptr, new_key);
+      return;
+   }
+
+   text_len = png_strlen(text);
+
+   /* compute the compressed data; do it now for the length */
+   text_len = png_text_compress(png_ptr, text, text_len, compression,
+       &comp);
+
+   /* write start of chunk */
+   png_write_chunk_start(png_ptr, (png_bytep)png_zTXt, (png_uint_32)
+      (key_len+text_len+2));
+   /* write key */
+   png_write_chunk_data(png_ptr, (png_bytep)new_key, key_len + 1);
+   png_free(png_ptr, new_key);
+
+   buf[0] = (png_byte)compression;
+   /* write compression */
+   png_write_chunk_data(png_ptr, (png_bytep)buf, (png_size_t)1);
+   /* write the compressed data */
+   png_write_compressed_data_out(png_ptr, &comp);
+
+   /* close the chunk */
+   png_write_chunk_end(png_ptr);
+}
+#endif
+
+#if defined(PNG_WRITE_iTXt_SUPPORTED)
+/* write an iTXt chunk */
+void /* PRIVATE */
+png_write_iTXt(png_structp png_ptr, int compression, png_charp key,
+    png_charp lang, png_charp lang_key, png_charp text)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_iTXt;
+#endif
+   png_size_t lang_len, key_len, lang_key_len, text_len;
+   png_charp new_lang, new_key;
+   png_byte cbuf[2];
+   compression_state comp;
+
+   png_debug(1, "in png_write_iTXt\n");
+
+   comp.num_output_ptr = 0;
+   comp.max_output_ptr = 0;
+   comp.output_ptr = NULL;
+   comp.input = NULL;
+
+   if (key == NULL || (key_len = png_check_keyword(png_ptr, key, &new_key))==0)
+   {
+      png_warning(png_ptr, "Empty keyword in iTXt chunk");
+      return;
+   }
+   if (lang == NULL || (lang_len = png_check_keyword(png_ptr, lang, &new_lang))==0)
+   {
+      png_warning(png_ptr, "Empty language field in iTXt chunk");
+      new_lang = NULL;
+      lang_len = 0;
+   }
+
+   if (lang_key == NULL)
+     lang_key_len = 0;
+   else
+     lang_key_len = png_strlen(lang_key);
+
+   if (text == NULL)
+      text_len = 0;
+   else
+     text_len = png_strlen(text);
+
+   /* compute the compressed data; do it now for the length */
+   text_len = png_text_compress(png_ptr, text, text_len, compression-2,
+      &comp);
+
+
+   /* make sure we include the compression flag, the compression byte,
+    * and the NULs after the key, lang, and lang_key parts */
+
+   png_write_chunk_start(png_ptr, (png_bytep)png_iTXt,
+          (png_uint_32)(
+        5 /* comp byte, comp flag, terminators for key, lang and lang_key */
+        + key_len
+        + lang_len
+        + lang_key_len
+        + text_len));
+
+   /*
+    * We leave it to the application to meet PNG-1.0 requirements on the
+    * contents of the text.  PNG-1.0 through PNG-1.2 discourage the use of
+    * any non-Latin-1 characters except for NEWLINE.  ISO PNG will forbid them.
+    * The NUL character is forbidden by PNG-1.0 through PNG-1.2 and ISO PNG.
+    */
+   png_write_chunk_data(png_ptr, (png_bytep)new_key, key_len + 1);
+
+   /* set the compression flag */
+   if (compression == PNG_ITXT_COMPRESSION_NONE || \
+       compression == PNG_TEXT_COMPRESSION_NONE)
+       cbuf[0] = 0;
+   else /* compression == PNG_ITXT_COMPRESSION_zTXt */
+       cbuf[0] = 1;
+   /* set the compression method */
+   cbuf[1] = 0;
+   png_write_chunk_data(png_ptr, cbuf, 2);
+
+   cbuf[0] = 0;
+   png_write_chunk_data(png_ptr, (new_lang ? (png_bytep)new_lang : cbuf), lang_len + 1);
+   png_write_chunk_data(png_ptr, (lang_key ? (png_bytep)lang_key : cbuf), lang_key_len + 1);
+   png_write_compressed_data_out(png_ptr, &comp);
+
+   png_write_chunk_end(png_ptr);
+   png_free(png_ptr, new_key);
+   if (new_lang)
+     png_free(png_ptr, new_lang);
+}
+#endif
+
+#if defined(PNG_WRITE_oFFs_SUPPORTED)
+/* write the oFFs chunk */
+void /* PRIVATE */
+png_write_oFFs(png_structp png_ptr, png_int_32 x_offset, png_int_32 y_offset,
+   int unit_type)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_oFFs;
+#endif
+   png_byte buf[9];
+
+   png_debug(1, "in png_write_oFFs\n");
+   if (unit_type >= PNG_OFFSET_LAST)
+      png_warning(png_ptr, "Unrecognized unit type for oFFs chunk");
+
+   png_save_int_32(buf, x_offset);
+   png_save_int_32(buf + 4, y_offset);
+   buf[8] = (png_byte)unit_type;
+
+   png_write_chunk(png_ptr, png_oFFs, buf, (png_size_t)9);
+}
+#endif
+#if defined(PNG_WRITE_pCAL_SUPPORTED)
+/* write the pCAL chunk (described in the PNG extensions document) */
+void /* PRIVATE */
+png_write_pCAL(png_structp png_ptr, png_charp purpose, png_int_32 X0,
+   png_int_32 X1, int type, int nparams, png_charp units, png_charpp params)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_pCAL;
+#endif
+   png_size_t purpose_len, units_len, total_len;
+   png_uint_32p params_len;
+   png_byte buf[10];
+   png_charp new_purpose;
+   int i;
+
+   png_debug1(1, "in png_write_pCAL (%d parameters)\n", nparams);
+   if (type >= PNG_EQUATION_LAST)
+      png_warning(png_ptr, "Unrecognized equation type for pCAL chunk");
+
+   purpose_len = png_check_keyword(png_ptr, purpose, &new_purpose) + 1;
+   png_debug1(3, "pCAL purpose length = %d\n", (int)purpose_len);
+   units_len = png_strlen(units) + (nparams == 0 ? 0 : 1);
+   png_debug1(3, "pCAL units length = %d\n", (int)units_len);
+   total_len = purpose_len + units_len + 10;
+
+   params_len = (png_uint_32p)png_malloc(png_ptr, (png_uint_32)(nparams
+      *png_sizeof(png_uint_32)));
+
+   /* Find the length of each parameter, making sure we don't count the
+      null terminator for the last parameter. */
+   for (i = 0; i < nparams; i++)
+   {
+      params_len[i] = png_strlen(params[i]) + (i == nparams - 1 ? 0 : 1);
+      png_debug2(3, "pCAL parameter %d length = %lu\n", i, params_len[i]);
+      total_len += (png_size_t)params_len[i];
+   }
+
+   png_debug1(3, "pCAL total length = %d\n", (int)total_len);
+   png_write_chunk_start(png_ptr, png_pCAL, (png_uint_32)total_len);
+   png_write_chunk_data(png_ptr, (png_bytep)new_purpose, purpose_len);
+   png_save_int_32(buf, X0);
+   png_save_int_32(buf + 4, X1);
+   buf[8] = (png_byte)type;
+   buf[9] = (png_byte)nparams;
+   png_write_chunk_data(png_ptr, buf, (png_size_t)10);
+   png_write_chunk_data(png_ptr, (png_bytep)units, (png_size_t)units_len);
+
+   png_free(png_ptr, new_purpose);
+
+   for (i = 0; i < nparams; i++)
+   {
+      png_write_chunk_data(png_ptr, (png_bytep)params[i],
+         (png_size_t)params_len[i]);
+   }
+
+   png_free(png_ptr, params_len);
+   png_write_chunk_end(png_ptr);
+}
+#endif
+
+#if defined(PNG_WRITE_sCAL_SUPPORTED)
+/* write the sCAL chunk */
+#if defined(PNG_FLOATING_POINT_SUPPORTED) && !defined(PNG_NO_STDIO)
+void /* PRIVATE */
+png_write_sCAL(png_structp png_ptr, int unit, double width, double height)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_sCAL;
+#endif
+   char buf[64];
+   png_size_t total_len;
+
+   png_debug(1, "in png_write_sCAL\n");
+
+   buf[0] = (char)unit;
+#if defined(_WIN32_WCE)
+/* sprintf() function is not supported on WindowsCE */
+   {
+      wchar_t wc_buf[32];
+      size_t wc_len;
+      swprintf(wc_buf, TEXT("%12.12e"), width);
+      wc_len = wcslen(wc_buf);
+      WideCharToMultiByte(CP_ACP, 0, wc_buf, -1, buf + 1, wc_len, NULL, NULL);
+      total_len = wc_len + 2;
+      swprintf(wc_buf, TEXT("%12.12e"), height);
+      wc_len = wcslen(wc_buf);
+      WideCharToMultiByte(CP_ACP, 0, wc_buf, -1, buf + total_len, wc_len,
+         NULL, NULL);
+      total_len += wc_len;
+   }
+#else
+   png_snprintf(buf + 1, 63, "%12.12e", width);
+   total_len = 1 + png_strlen(buf + 1) + 1;
+   png_snprintf(buf + total_len, 64-total_len, "%12.12e", height);
+   total_len += png_strlen(buf + total_len);
+#endif
+
+   png_debug1(3, "sCAL total length = %u\n", (unsigned int)total_len);
+   png_write_chunk(png_ptr, png_sCAL, (png_bytep)buf, total_len);
+}
+#else
+#ifdef PNG_FIXED_POINT_SUPPORTED
+void /* PRIVATE */
+png_write_sCAL_s(png_structp png_ptr, int unit, png_charp width,
+   png_charp height)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_sCAL;
+#endif
+   png_byte buf[64];
+   png_size_t wlen, hlen, total_len;
+
+   png_debug(1, "in png_write_sCAL_s\n");
+
+   wlen = png_strlen(width);
+   hlen = png_strlen(height);
+   total_len = wlen + hlen + 2;
+   if (total_len > 64)
+   {
+      png_warning(png_ptr, "Can't write sCAL (buffer too small)");
+      return;
+   }
+
+   buf[0] = (png_byte)unit;
+   png_memcpy(buf + 1, width, wlen + 1);      /* append the '\0' here */
+   png_memcpy(buf + wlen + 2, height, hlen);  /* do NOT append the '\0' here */
+
+   png_debug1(3, "sCAL total length = %u\n", (unsigned int)total_len);
+   png_write_chunk(png_ptr, png_sCAL, buf, total_len);
+}
+#endif
+#endif
+#endif
+
+#if defined(PNG_WRITE_pHYs_SUPPORTED)
+/* write the pHYs chunk */
+void /* PRIVATE */
+png_write_pHYs(png_structp png_ptr, png_uint_32 x_pixels_per_unit,
+   png_uint_32 y_pixels_per_unit,
+   int unit_type)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_pHYs;
+#endif
+   png_byte buf[9];
+
+   png_debug(1, "in png_write_pHYs\n");
+   if (unit_type >= PNG_RESOLUTION_LAST)
+      png_warning(png_ptr, "Unrecognized unit type for pHYs chunk");
+
+   png_save_uint_32(buf, x_pixels_per_unit);
+   png_save_uint_32(buf + 4, y_pixels_per_unit);
+   buf[8] = (png_byte)unit_type;
+
+   png_write_chunk(png_ptr, png_pHYs, buf, (png_size_t)9);
+}
+#endif
+
+#if defined(PNG_WRITE_tIME_SUPPORTED)
+/* Write the tIME chunk.  Use either png_convert_from_struct_tm()
+ * or png_convert_from_time_t(), or fill in the structure yourself.
+ */
+void /* PRIVATE */
+png_write_tIME(png_structp png_ptr, png_timep mod_time)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   PNG_tIME;
+#endif
+   png_byte buf[7];
+
+   png_debug(1, "in png_write_tIME\n");
+   if (mod_time->month  > 12 || mod_time->month  < 1 ||
+       mod_time->day    > 31 || mod_time->day    < 1 ||
+       mod_time->hour   > 23 || mod_time->second > 60)
+   {
+      png_warning(png_ptr, "Invalid time specified for tIME chunk");
+      return;
+   }
+
+   png_save_uint_16(buf, mod_time->year);
+   buf[2] = mod_time->month;
+   buf[3] = mod_time->day;
+   buf[4] = mod_time->hour;
+   buf[5] = mod_time->minute;
+   buf[6] = mod_time->second;
+
+   png_write_chunk(png_ptr, png_tIME, buf, (png_size_t)7);
+}
+#endif
+
+/* initializes the row writing capability of libpng */
+void /* PRIVATE */
+png_write_start_row(png_structp png_ptr)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   /* arrays to facilitate easy interlacing - use pass (0 - 6) as index */
+
+   /* start of interlace block */
+   int png_pass_start[7] = {0, 4, 0, 2, 0, 1, 0};
+
+   /* offset to next interlace block */
+   int png_pass_inc[7] = {8, 8, 4, 4, 2, 2, 1};
+
+   /* start of interlace block in the y direction */
+   int png_pass_ystart[7] = {0, 0, 4, 0, 2, 0, 1};
+
+   /* offset to next interlace block in the y direction */
+   int png_pass_yinc[7] = {8, 8, 8, 4, 4, 2, 2};
+#endif
+
+   png_size_t buf_size;
+
+   png_debug(1, "in png_write_start_row\n");
+   buf_size = (png_size_t)(PNG_ROWBYTES(
+      png_ptr->usr_channels*png_ptr->usr_bit_depth,png_ptr->width)+1);
+
+   /* set up row buffer */
+   png_ptr->row_buf = (png_bytep)png_malloc(png_ptr, (png_uint_32)buf_size);
+   png_ptr->row_buf[0] = PNG_FILTER_VALUE_NONE;
+
+#ifndef PNG_NO_WRITE_FILTERING
+   /* set up filtering buffer, if using this filter */
+   if (png_ptr->do_filter & PNG_FILTER_SUB)
+   {
+      png_ptr->sub_row = (png_bytep)png_malloc(png_ptr,
+         (png_ptr->rowbytes + 1));
+      png_ptr->sub_row[0] = PNG_FILTER_VALUE_SUB;
+   }
+
+   /* We only need to keep the previous row if we are using one of these. */
+   if (png_ptr->do_filter & (PNG_FILTER_AVG | PNG_FILTER_UP | PNG_FILTER_PAETH))
+   {
+     /* set up previous row buffer */
+      png_ptr->prev_row = (png_bytep)png_malloc(png_ptr, (png_uint_32)buf_size);
+      png_memset(png_ptr->prev_row, 0, buf_size);
+
+      if (png_ptr->do_filter & PNG_FILTER_UP)
+      {
+         png_ptr->up_row = (png_bytep)png_malloc(png_ptr,
+            (png_ptr->rowbytes + 1));
+         png_ptr->up_row[0] = PNG_FILTER_VALUE_UP;
+      }
+
+      if (png_ptr->do_filter & PNG_FILTER_AVG)
+      {
+         png_ptr->avg_row = (png_bytep)png_malloc(png_ptr,
+            (png_ptr->rowbytes + 1));
+         png_ptr->avg_row[0] = PNG_FILTER_VALUE_AVG;
+      }
+
+      if (png_ptr->do_filter & PNG_FILTER_PAETH)
+      {
+         png_ptr->paeth_row = (png_bytep)png_malloc(png_ptr,
+            (png_ptr->rowbytes + 1));
+         png_ptr->paeth_row[0] = PNG_FILTER_VALUE_PAETH;
+      }
+#endif /* PNG_NO_WRITE_FILTERING */
+   }
+
+#ifdef PNG_WRITE_INTERLACING_SUPPORTED
+   /* if interlaced, we need to set up width and height of pass */
+   if (png_ptr->interlaced)
+   {
+      if (!(png_ptr->transformations & PNG_INTERLACE))
+      {
+         png_ptr->num_rows = (png_ptr->height + png_pass_yinc[0] - 1 -
+            png_pass_ystart[0]) / png_pass_yinc[0];
+         png_ptr->usr_width = (png_ptr->width + png_pass_inc[0] - 1 -
+            png_pass_start[0]) / png_pass_inc[0];
+      }
+      else
+      {
+         png_ptr->num_rows = png_ptr->height;
+         png_ptr->usr_width = png_ptr->width;
+      }
+   }
+   else
+#endif
+   {
+      png_ptr->num_rows = png_ptr->height;
+      png_ptr->usr_width = png_ptr->width;
+   }
+   png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+   png_ptr->zstream.next_out = png_ptr->zbuf;
+}
+
+/* Internal use only.  Called when finished processing a row of data. */
+void /* PRIVATE */
+png_write_finish_row(png_structp png_ptr)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   /* arrays to facilitate easy interlacing - use pass (0 - 6) as index */
+
+   /* start of interlace block */
+   int png_pass_start[7] = {0, 4, 0, 2, 0, 1, 0};
+
+   /* offset to next interlace block */
+   int png_pass_inc[7] = {8, 8, 4, 4, 2, 2, 1};
+
+   /* start of interlace block in the y direction */
+   int png_pass_ystart[7] = {0, 0, 4, 0, 2, 0, 1};
+
+   /* offset to next interlace block in the y direction */
+   int png_pass_yinc[7] = {8, 8, 8, 4, 4, 2, 2};
+#endif
+
+   int ret;
+
+   png_debug(1, "in png_write_finish_row\n");
+   /* next row */
+   png_ptr->row_number++;
+
+   /* see if we are done */
+   if (png_ptr->row_number < png_ptr->num_rows)
+      return;
+
+#ifdef PNG_WRITE_INTERLACING_SUPPORTED
+   /* if interlaced, go to next pass */
+   if (png_ptr->interlaced)
+   {
+      png_ptr->row_number = 0;
+      if (png_ptr->transformations & PNG_INTERLACE)
+      {
+         png_ptr->pass++;
+      }
+      else
+      {
+         /* loop until we find a non-zero width or height pass */
+         do
+         {
+            png_ptr->pass++;
+            if (png_ptr->pass >= 7)
+               break;
+            png_ptr->usr_width = (png_ptr->width +
+               png_pass_inc[png_ptr->pass] - 1 -
+               png_pass_start[png_ptr->pass]) /
+               png_pass_inc[png_ptr->pass];
+            png_ptr->num_rows = (png_ptr->height +
+               png_pass_yinc[png_ptr->pass] - 1 -
+               png_pass_ystart[png_ptr->pass]) /
+               png_pass_yinc[png_ptr->pass];
+            if (png_ptr->transformations & PNG_INTERLACE)
+               break;
+         } while (png_ptr->usr_width == 0 || png_ptr->num_rows == 0);
+
+      }
+
+      /* reset the row above the image for the next pass */
+      if (png_ptr->pass < 7)
+      {
+         if (png_ptr->prev_row != NULL)
+            png_memset(png_ptr->prev_row, 0,
+               (png_size_t)(PNG_ROWBYTES(png_ptr->usr_channels*
+               png_ptr->usr_bit_depth,png_ptr->width))+1);
+         return;
+      }
+   }
+#endif
+
+   /* if we get here, we've just written the last row, so we need
+      to flush the compressor */
+   do
+   {
+      /* tell the compressor we are done */
+      ret = deflate(&png_ptr->zstream, Z_FINISH);
+      /* check for an error */
+      if (ret == Z_OK)
+      {
+         /* check to see if we need more room */
+         if (!(png_ptr->zstream.avail_out))
+         {
+            png_write_IDAT(png_ptr, png_ptr->zbuf, png_ptr->zbuf_size);
+            png_ptr->zstream.next_out = png_ptr->zbuf;
+            png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+         }
+      }
+      else if (ret != Z_STREAM_END)
+      {
+         if (png_ptr->zstream.msg != NULL)
+            png_error(png_ptr, png_ptr->zstream.msg);
+         else
+            png_error(png_ptr, "zlib error");
+      }
+   } while (ret != Z_STREAM_END);
+
+   /* write any extra space */
+   if (png_ptr->zstream.avail_out < png_ptr->zbuf_size)
+   {
+      png_write_IDAT(png_ptr, png_ptr->zbuf, png_ptr->zbuf_size -
+         png_ptr->zstream.avail_out);
+   }
+
+   deflateReset(&png_ptr->zstream);
+   png_ptr->zstream.data_type = Z_BINARY;
+}
+
+#if defined(PNG_WRITE_INTERLACING_SUPPORTED)
+/* Pick out the correct pixels for the interlace pass.
+ * The basic idea here is to go through the row with a source
+ * pointer and a destination pointer (sp and dp), and copy the
+ * correct pixels for the pass.  As the row gets compacted,
+ * sp will always be >= dp, so we should never overwrite anything.
+ * See the default: case for the easiest code to understand.
+ */
+void /* PRIVATE */
+png_do_write_interlace(png_row_infop row_info, png_bytep row, int pass)
+{
+#ifdef PNG_USE_LOCAL_ARRAYS
+   /* arrays to facilitate easy interlacing - use pass (0 - 6) as index */
+
+   /* start of interlace block */
+   int png_pass_start[7] = {0, 4, 0, 2, 0, 1, 0};
+
+   /* offset to next interlace block */
+   int png_pass_inc[7] = {8, 8, 4, 4, 2, 2, 1};
+#endif
+
+   png_debug(1, "in png_do_write_interlace\n");
+   /* we don't have to do anything on the last pass (6) */
+#if defined(PNG_USELESS_TESTS_SUPPORTED)
+   if (row != NULL && row_info != NULL && pass < 6)
+#else
+   if (pass < 6)
+#endif
+   {
+      /* each pixel depth is handled separately */
+      switch (row_info->pixel_depth)
+      {
+         case 1:
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int shift;
+            int d;
+            int value;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            dp = row;
+            d = 0;
+            shift = 7;
+            for (i = png_pass_start[pass]; i < row_width;
+               i += png_pass_inc[pass])
+            {
+               sp = row + (png_size_t)(i >> 3);
+               value = (int)(*sp >> (7 - (int)(i & 0x07))) & 0x01;
+               d |= (value << shift);
+
+               if (shift == 0)
+               {
+                  shift = 7;
+                  *dp++ = (png_byte)d;
+                  d = 0;
+               }
+               else
+                  shift--;
+
+            }
+            if (shift != 7)
+               *dp = (png_byte)d;
+            break;
+         }
+         case 2:
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int shift;
+            int d;
+            int value;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            dp = row;
+            shift = 6;
+            d = 0;
+            for (i = png_pass_start[pass]; i < row_width;
+               i += png_pass_inc[pass])
+            {
+               sp = row + (png_size_t)(i >> 2);
+               value = (*sp >> ((3 - (int)(i & 0x03)) << 1)) & 0x03;
+               d |= (value << shift);
+
+               if (shift == 0)
+               {
+                  shift = 6;
+                  *dp++ = (png_byte)d;
+                  d = 0;
+               }
+               else
+                  shift -= 2;
+            }
+            if (shift != 6)
+                   *dp = (png_byte)d;
+            break;
+         }
+         case 4:
+         {
+            png_bytep sp;
+            png_bytep dp;
+            int shift;
+            int d;
+            int value;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+
+            dp = row;
+            shift = 4;
+            d = 0;
+            for (i = png_pass_start[pass]; i < row_width;
+               i += png_pass_inc[pass])
+            {
+               sp = row + (png_size_t)(i >> 1);
+               value = (*sp >> ((1 - (int)(i & 0x01)) << 2)) & 0x0f;
+               d |= (value << shift);
+
+               if (shift == 0)
+               {
+                  shift = 4;
+                  *dp++ = (png_byte)d;
+                  d = 0;
+               }
+               else
+                  shift -= 4;
+            }
+            if (shift != 4)
+               *dp = (png_byte)d;
+            break;
+         }
+         default:
+         {
+            png_bytep sp;
+            png_bytep dp;
+            png_uint_32 i;
+            png_uint_32 row_width = row_info->width;
+            png_size_t pixel_bytes;
+
+            /* start at the beginning */
+            dp = row;
+            /* find out how many bytes each pixel takes up */
+            pixel_bytes = (row_info->pixel_depth >> 3);
+            /* loop through the row, only looking at the pixels that
+               matter */
+            for (i = png_pass_start[pass]; i < row_width;
+               i += png_pass_inc[pass])
+            {
+               /* find out where the original pixel is */
+               sp = row + (png_size_t)i * pixel_bytes;
+               /* move the pixel */
+               if (dp != sp)
+                  png_memcpy(dp, sp, pixel_bytes);
+               /* next pixel */
+               dp += pixel_bytes;
+            }
+            break;
+         }
+      }
+      /* set new row width */
+      row_info->width = (row_info->width +
+         png_pass_inc[pass] - 1 -
+         png_pass_start[pass]) /
+         png_pass_inc[pass];
+         row_info->rowbytes = PNG_ROWBYTES(row_info->pixel_depth,
+            row_info->width);
+   }
+}
+#endif
+
+/* This filters the row, chooses which filter to use, if it has not already
+ * been specified by the application, and then writes the row out with the
+ * chosen filter.
+ */
+#define PNG_MAXSUM (((png_uint_32)(-1)) >> 1)
+#define PNG_HISHIFT 10
+#define PNG_LOMASK ((png_uint_32)0xffffL)
+#define PNG_HIMASK ((png_uint_32)(~PNG_LOMASK >> PNG_HISHIFT))
+void /* PRIVATE */
+png_write_find_filter(png_structp png_ptr, png_row_infop row_info)
+{
+   png_bytep prev_row, best_row, row_buf;
+   png_uint_32 mins, bpp;
+   png_byte filter_to_do = png_ptr->do_filter;
+   png_uint_32 row_bytes = row_info->rowbytes;
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+   int num_p_filters = (int)png_ptr->num_prev_filters;
+#endif
+
+   png_debug(1, "in png_write_find_filter\n");
+   /* find out how many bytes offset each pixel is */
+   bpp = (row_info->pixel_depth + 7) >> 3;
+
+   prev_row = png_ptr->prev_row;
+   best_row = row_buf = png_ptr->row_buf;
+#ifndef PNG_NO_WRITE_FILTER
+   mins = PNG_MAXSUM;
+
+   /* The prediction method we use is to find which method provides the
+    * smallest value when summing the absolute values of the distances
+    * from zero, using anything >= 128 as negative numbers.  This is known
+    * as the "minimum sum of absolute differences" heuristic.  Other
+    * heuristics are the "weighted minimum sum of absolute differences"
+    * (experimental and can in theory improve compression), and the "zlib
+    * predictive" method (not implemented yet), which does test compressions
+    * of lines using different filter methods, and then chooses the
+    * (series of) filter(s) that give minimum compressed data size (VERY
+    * computationally expensive).
+    *
+    * GRR 980525:  consider also
+    *   (1) minimum sum of absolute differences from running average (i.e.,
+    *       keep running sum of non-absolute differences & count of bytes)
+    *       [track dispersion, too?  restart average if dispersion too large?]
+    *  (1b) minimum sum of absolute differences from sliding average, probably
+    *       with window size <= deflate window (usually 32K)
+    *   (2) minimum sum of squared differences from zero or running average
+    *       (i.e., ~ root-mean-square approach)
+    */
+
+
+   /* We don't need to test the 'no filter' case if this is the only filter
+    * that has been chosen, as it doesn't actually do anything to the data.
+    */
+   if ((filter_to_do & PNG_FILTER_NONE) &&
+       filter_to_do != PNG_FILTER_NONE)
+   {
+      png_bytep rp;
+      png_uint_32 sum = 0;
+      png_uint_32 i;
+      int v;
+
+      for (i = 0, rp = row_buf + 1; i < row_bytes; i++, rp++)
+      {
+         v = *rp;
+         sum += (v < 128) ? v : 256 - v;
+      }
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+      if (png_ptr->heuristic_method == PNG_FILTER_HEURISTIC_WEIGHTED)
+      {
+         png_uint_32 sumhi, sumlo;
+         int j;
+         sumlo = sum & PNG_LOMASK;
+         sumhi = (sum >> PNG_HISHIFT) & PNG_HIMASK; /* Gives us some footroom */
+
+         /* Reduce the sum if we match any of the previous rows */
+         for (j = 0; j < num_p_filters; j++)
+         {
+            if (png_ptr->prev_filters[j] == PNG_FILTER_VALUE_NONE)
+            {
+               sumlo = (sumlo * png_ptr->filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+               sumhi = (sumhi * png_ptr->filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+            }
+         }
+
+         /* Factor in the cost of this filter (this is here for completeness,
+          * but it makes no sense to have a "cost" for the NONE filter, as
+          * it has the minimum possible computational cost - none).
+          */
+         sumlo = (sumlo * png_ptr->filter_costs[PNG_FILTER_VALUE_NONE]) >>
+            PNG_COST_SHIFT;
+         sumhi = (sumhi * png_ptr->filter_costs[PNG_FILTER_VALUE_NONE]) >>
+            PNG_COST_SHIFT;
+
+         if (sumhi > PNG_HIMASK)
+            sum = PNG_MAXSUM;
+         else
+            sum = (sumhi << PNG_HISHIFT) + sumlo;
+      }
+#endif
+      mins = sum;
+   }
+
+   /* sub filter */
+   if (filter_to_do == PNG_FILTER_SUB)
+   /* it's the only filter so no testing is needed */
+   {
+      png_bytep rp, lp, dp;
+      png_uint_32 i;
+      for (i = 0, rp = row_buf + 1, dp = png_ptr->sub_row + 1; i < bpp;
+           i++, rp++, dp++)
+      {
+         *dp = *rp;
+      }
+      for (lp = row_buf + 1; i < row_bytes;
+         i++, rp++, lp++, dp++)
+      {
+         *dp = (png_byte)(((int)*rp - (int)*lp) & 0xff);
+      }
+      best_row = png_ptr->sub_row;
+   }
+
+   else if (filter_to_do & PNG_FILTER_SUB)
+   {
+      png_bytep rp, dp, lp;
+      png_uint_32 sum = 0, lmins = mins;
+      png_uint_32 i;
+      int v;
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+      /* We temporarily increase the "minimum sum" by the factor we
+       * would reduce the sum of this filter, so that we can do the
+       * early exit comparison without scaling the sum each time.
+       */
+      if (png_ptr->heuristic_method == PNG_FILTER_HEURISTIC_WEIGHTED)
+      {
+         int j;
+         png_uint_32 lmhi, lmlo;
+         lmlo = lmins & PNG_LOMASK;
+         lmhi = (lmins >> PNG_HISHIFT) & PNG_HIMASK;
+
+         for (j = 0; j < num_p_filters; j++)
+         {
+            if (png_ptr->prev_filters[j] == PNG_FILTER_VALUE_SUB)
+            {
+               lmlo = (lmlo * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+               lmhi = (lmhi * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+            }
+         }
+
+         lmlo = (lmlo * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_SUB]) >>
+            PNG_COST_SHIFT;
+         lmhi = (lmhi * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_SUB]) >>
+            PNG_COST_SHIFT;
+
+         if (lmhi > PNG_HIMASK)
+            lmins = PNG_MAXSUM;
+         else
+            lmins = (lmhi << PNG_HISHIFT) + lmlo;
+      }
+#endif
+
+      for (i = 0, rp = row_buf + 1, dp = png_ptr->sub_row + 1; i < bpp;
+           i++, rp++, dp++)
+      {
+         v = *dp = *rp;
+
+         sum += (v < 128) ? v : 256 - v;
+      }
+      for (lp = row_buf + 1; i < row_bytes;
+         i++, rp++, lp++, dp++)
+      {
+         v = *dp = (png_byte)(((int)*rp - (int)*lp) & 0xff);
+
+         sum += (v < 128) ? v : 256 - v;
+
+         if (sum > lmins)  /* We are already worse, don't continue. */
+            break;
+      }
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+      if (png_ptr->heuristic_method == PNG_FILTER_HEURISTIC_WEIGHTED)
+      {
+         int j;
+         png_uint_32 sumhi, sumlo;
+         sumlo = sum & PNG_LOMASK;
+         sumhi = (sum >> PNG_HISHIFT) & PNG_HIMASK;
+
+         for (j = 0; j < num_p_filters; j++)
+         {
+            if (png_ptr->prev_filters[j] == PNG_FILTER_VALUE_SUB)
+            {
+               sumlo = (sumlo * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+               sumhi = (sumhi * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+            }
+         }
+
+         sumlo = (sumlo * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_SUB]) >>
+            PNG_COST_SHIFT;
+         sumhi = (sumhi * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_SUB]) >>
+            PNG_COST_SHIFT;
+
+         if (sumhi > PNG_HIMASK)
+            sum = PNG_MAXSUM;
+         else
+            sum = (sumhi << PNG_HISHIFT) + sumlo;
+      }
+#endif
+
+      if (sum < mins)
+      {
+         mins = sum;
+         best_row = png_ptr->sub_row;
+      }
+   }
+
+   /* up filter */
+   if (filter_to_do == PNG_FILTER_UP)
+   {
+      png_bytep rp, dp, pp;
+      png_uint_32 i;
+
+      for (i = 0, rp = row_buf + 1, dp = png_ptr->up_row + 1,
+           pp = prev_row + 1; i < row_bytes;
+           i++, rp++, pp++, dp++)
+      {
+         *dp = (png_byte)(((int)*rp - (int)*pp) & 0xff);
+      }
+      best_row = png_ptr->up_row;
+   }
+
+   else if (filter_to_do & PNG_FILTER_UP)
+   {
+      png_bytep rp, dp, pp;
+      png_uint_32 sum = 0, lmins = mins;
+      png_uint_32 i;
+      int v;
+
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+      if (png_ptr->heuristic_method == PNG_FILTER_HEURISTIC_WEIGHTED)
+      {
+         int j;
+         png_uint_32 lmhi, lmlo;
+         lmlo = lmins & PNG_LOMASK;
+         lmhi = (lmins >> PNG_HISHIFT) & PNG_HIMASK;
+
+         for (j = 0; j < num_p_filters; j++)
+         {
+            if (png_ptr->prev_filters[j] == PNG_FILTER_VALUE_UP)
+            {
+               lmlo = (lmlo * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+               lmhi = (lmhi * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+            }
+         }
+
+         lmlo = (lmlo * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_UP]) >>
+            PNG_COST_SHIFT;
+         lmhi = (lmhi * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_UP]) >>
+            PNG_COST_SHIFT;
+
+         if (lmhi > PNG_HIMASK)
+            lmins = PNG_MAXSUM;
+         else
+            lmins = (lmhi << PNG_HISHIFT) + lmlo;
+      }
+#endif
+
+      for (i = 0, rp = row_buf + 1, dp = png_ptr->up_row + 1,
+           pp = prev_row + 1; i < row_bytes; i++)
+      {
+         v = *dp++ = (png_byte)(((int)*rp++ - (int)*pp++) & 0xff);
+
+         sum += (v < 128) ? v : 256 - v;
+
+         if (sum > lmins)  /* We are already worse, don't continue. */
+            break;
+      }
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+      if (png_ptr->heuristic_method == PNG_FILTER_HEURISTIC_WEIGHTED)
+      {
+         int j;
+         png_uint_32 sumhi, sumlo;
+         sumlo = sum & PNG_LOMASK;
+         sumhi = (sum >> PNG_HISHIFT) & PNG_HIMASK;
+
+         for (j = 0; j < num_p_filters; j++)
+         {
+            if (png_ptr->prev_filters[j] == PNG_FILTER_VALUE_UP)
+            {
+               sumlo = (sumlo * png_ptr->filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+               sumhi = (sumhi * png_ptr->filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+            }
+         }
+
+         sumlo = (sumlo * png_ptr->filter_costs[PNG_FILTER_VALUE_UP]) >>
+            PNG_COST_SHIFT;
+         sumhi = (sumhi * png_ptr->filter_costs[PNG_FILTER_VALUE_UP]) >>
+            PNG_COST_SHIFT;
+
+         if (sumhi > PNG_HIMASK)
+            sum = PNG_MAXSUM;
+         else
+            sum = (sumhi << PNG_HISHIFT) + sumlo;
+      }
+#endif
+
+      if (sum < mins)
+      {
+         mins = sum;
+         best_row = png_ptr->up_row;
+      }
+   }
+
+   /* avg filter */
+   if (filter_to_do == PNG_FILTER_AVG)
+   {
+      png_bytep rp, dp, pp, lp;
+      png_uint_32 i;
+      for (i = 0, rp = row_buf + 1, dp = png_ptr->avg_row + 1,
+           pp = prev_row + 1; i < bpp; i++)
+      {
+         *dp++ = (png_byte)(((int)*rp++ - ((int)*pp++ / 2)) & 0xff);
+      }
+      for (lp = row_buf + 1; i < row_bytes; i++)
+      {
+         *dp++ = (png_byte)(((int)*rp++ - (((int)*pp++ + (int)*lp++) / 2))
+                 & 0xff);
+      }
+      best_row = png_ptr->avg_row;
+   }
+
+   else if (filter_to_do & PNG_FILTER_AVG)
+   {
+      png_bytep rp, dp, pp, lp;
+      png_uint_32 sum = 0, lmins = mins;
+      png_uint_32 i;
+      int v;
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+      if (png_ptr->heuristic_method == PNG_FILTER_HEURISTIC_WEIGHTED)
+      {
+         int j;
+         png_uint_32 lmhi, lmlo;
+         lmlo = lmins & PNG_LOMASK;
+         lmhi = (lmins >> PNG_HISHIFT) & PNG_HIMASK;
+
+         for (j = 0; j < num_p_filters; j++)
+         {
+            if (png_ptr->prev_filters[j] == PNG_FILTER_VALUE_AVG)
+            {
+               lmlo = (lmlo * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+               lmhi = (lmhi * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+            }
+         }
+
+         lmlo = (lmlo * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_AVG]) >>
+            PNG_COST_SHIFT;
+         lmhi = (lmhi * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_AVG]) >>
+            PNG_COST_SHIFT;
+
+         if (lmhi > PNG_HIMASK)
+            lmins = PNG_MAXSUM;
+         else
+            lmins = (lmhi << PNG_HISHIFT) + lmlo;
+      }
+#endif
+
+      for (i = 0, rp = row_buf + 1, dp = png_ptr->avg_row + 1,
+           pp = prev_row + 1; i < bpp; i++)
+      {
+         v = *dp++ = (png_byte)(((int)*rp++ - ((int)*pp++ / 2)) & 0xff);
+
+         sum += (v < 128) ? v : 256 - v;
+      }
+      for (lp = row_buf + 1; i < row_bytes; i++)
+      {
+         v = *dp++ =
+          (png_byte)(((int)*rp++ - (((int)*pp++ + (int)*lp++) / 2)) & 0xff);
+
+         sum += (v < 128) ? v : 256 - v;
+
+         if (sum > lmins)  /* We are already worse, don't continue. */
+            break;
+      }
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+      if (png_ptr->heuristic_method == PNG_FILTER_HEURISTIC_WEIGHTED)
+      {
+         int j;
+         png_uint_32 sumhi, sumlo;
+         sumlo = sum & PNG_LOMASK;
+         sumhi = (sum >> PNG_HISHIFT) & PNG_HIMASK;
+
+         for (j = 0; j < num_p_filters; j++)
+         {
+            if (png_ptr->prev_filters[j] == PNG_FILTER_VALUE_NONE)
+            {
+               sumlo = (sumlo * png_ptr->filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+               sumhi = (sumhi * png_ptr->filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+            }
+         }
+
+         sumlo = (sumlo * png_ptr->filter_costs[PNG_FILTER_VALUE_AVG]) >>
+            PNG_COST_SHIFT;
+         sumhi = (sumhi * png_ptr->filter_costs[PNG_FILTER_VALUE_AVG]) >>
+            PNG_COST_SHIFT;
+
+         if (sumhi > PNG_HIMASK)
+            sum = PNG_MAXSUM;
+         else
+            sum = (sumhi << PNG_HISHIFT) + sumlo;
+      }
+#endif
+
+      if (sum < mins)
+      {
+         mins = sum;
+         best_row = png_ptr->avg_row;
+      }
+   }
+
+   /* Paeth filter */
+   if (filter_to_do == PNG_FILTER_PAETH)
+   {
+      png_bytep rp, dp, pp, cp, lp;
+      png_uint_32 i;
+      for (i = 0, rp = row_buf + 1, dp = png_ptr->paeth_row + 1,
+           pp = prev_row + 1; i < bpp; i++)
+      {
+         *dp++ = (png_byte)(((int)*rp++ - (int)*pp++) & 0xff);
+      }
+
+      for (lp = row_buf + 1, cp = prev_row + 1; i < row_bytes; i++)
+      {
+         int a, b, c, pa, pb, pc, p;
+
+         b = *pp++;
+         c = *cp++;
+         a = *lp++;
+
+         p = b - c;
+         pc = a - c;
+
+#ifdef PNG_USE_ABS
+         pa = abs(p);
+         pb = abs(pc);
+         pc = abs(p + pc);
+#else
+         pa = p < 0 ? -p : p;
+         pb = pc < 0 ? -pc : pc;
+         pc = (p + pc) < 0 ? -(p + pc) : p + pc;
+#endif
+
+         p = (pa <= pb && pa <=pc) ? a : (pb <= pc) ? b : c;
+
+         *dp++ = (png_byte)(((int)*rp++ - p) & 0xff);
+      }
+      best_row = png_ptr->paeth_row;
+   }
+
+   else if (filter_to_do & PNG_FILTER_PAETH)
+   {
+      png_bytep rp, dp, pp, cp, lp;
+      png_uint_32 sum = 0, lmins = mins;
+      png_uint_32 i;
+      int v;
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+      if (png_ptr->heuristic_method == PNG_FILTER_HEURISTIC_WEIGHTED)
+      {
+         int j;
+         png_uint_32 lmhi, lmlo;
+         lmlo = lmins & PNG_LOMASK;
+         lmhi = (lmins >> PNG_HISHIFT) & PNG_HIMASK;
+
+         for (j = 0; j < num_p_filters; j++)
+         {
+            if (png_ptr->prev_filters[j] == PNG_FILTER_VALUE_PAETH)
+            {
+               lmlo = (lmlo * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+               lmhi = (lmhi * png_ptr->inv_filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+            }
+         }
+
+         lmlo = (lmlo * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_PAETH]) >>
+            PNG_COST_SHIFT;
+         lmhi = (lmhi * png_ptr->inv_filter_costs[PNG_FILTER_VALUE_PAETH]) >>
+            PNG_COST_SHIFT;
+
+         if (lmhi > PNG_HIMASK)
+            lmins = PNG_MAXSUM;
+         else
+            lmins = (lmhi << PNG_HISHIFT) + lmlo;
+      }
+#endif
+
+      for (i = 0, rp = row_buf + 1, dp = png_ptr->paeth_row + 1,
+           pp = prev_row + 1; i < bpp; i++)
+      {
+         v = *dp++ = (png_byte)(((int)*rp++ - (int)*pp++) & 0xff);
+
+         sum += (v < 128) ? v : 256 - v;
+      }
+
+      for (lp = row_buf + 1, cp = prev_row + 1; i < row_bytes; i++)
+      {
+         int a, b, c, pa, pb, pc, p;
+
+         b = *pp++;
+         c = *cp++;
+         a = *lp++;
+
+#ifndef PNG_SLOW_PAETH
+         p = b - c;
+         pc = a - c;
+#ifdef PNG_USE_ABS
+         pa = abs(p);
+         pb = abs(pc);
+         pc = abs(p + pc);
+#else
+         pa = p < 0 ? -p : p;
+         pb = pc < 0 ? -pc : pc;
+         pc = (p + pc) < 0 ? -(p + pc) : p + pc;
+#endif
+         p = (pa <= pb && pa <=pc) ? a : (pb <= pc) ? b : c;
+#else /* PNG_SLOW_PAETH */
+         p = a + b - c;
+         pa = abs(p - a);
+         pb = abs(p - b);
+         pc = abs(p - c);
+         if (pa <= pb && pa <= pc)
+            p = a;
+         else if (pb <= pc)
+            p = b;
+         else
+            p = c;
+#endif /* PNG_SLOW_PAETH */
+
+         v = *dp++ = (png_byte)(((int)*rp++ - p) & 0xff);
+
+         sum += (v < 128) ? v : 256 - v;
+
+         if (sum > lmins)  /* We are already worse, don't continue. */
+            break;
+      }
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+      if (png_ptr->heuristic_method == PNG_FILTER_HEURISTIC_WEIGHTED)
+      {
+         int j;
+         png_uint_32 sumhi, sumlo;
+         sumlo = sum & PNG_LOMASK;
+         sumhi = (sum >> PNG_HISHIFT) & PNG_HIMASK;
+
+         for (j = 0; j < num_p_filters; j++)
+         {
+            if (png_ptr->prev_filters[j] == PNG_FILTER_VALUE_PAETH)
+            {
+               sumlo = (sumlo * png_ptr->filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+               sumhi = (sumhi * png_ptr->filter_weights[j]) >>
+                  PNG_WEIGHT_SHIFT;
+            }
+         }
+
+         sumlo = (sumlo * png_ptr->filter_costs[PNG_FILTER_VALUE_PAETH]) >>
+            PNG_COST_SHIFT;
+         sumhi = (sumhi * png_ptr->filter_costs[PNG_FILTER_VALUE_PAETH]) >>
+            PNG_COST_SHIFT;
+
+         if (sumhi > PNG_HIMASK)
+            sum = PNG_MAXSUM;
+         else
+            sum = (sumhi << PNG_HISHIFT) + sumlo;
+      }
+#endif
+
+      if (sum < mins)
+      {
+         best_row = png_ptr->paeth_row;
+      }
+   }
+#endif /* PNG_NO_WRITE_FILTER */
+   /* Do the actual writing of the filtered row data from the chosen filter. */
+
+   png_write_filtered_row(png_ptr, best_row);
+
+#if defined(PNG_WRITE_WEIGHTED_FILTER_SUPPORTED)
+   /* Save the type of filter we picked this time for future calculations */
+   if (png_ptr->num_prev_filters > 0)
+   {
+      int j;
+      for (j = 1; j < num_p_filters; j++)
+      {
+         png_ptr->prev_filters[j] = png_ptr->prev_filters[j - 1];
+      }
+      png_ptr->prev_filters[j] = best_row[0];
+   }
+#endif
+}
+
+
+/* Do the actual writing of a previously filtered row. */
+void /* PRIVATE */
+png_write_filtered_row(png_structp png_ptr, png_bytep filtered_row)
+{
+   png_debug(1, "in png_write_filtered_row\n");
+   png_debug1(2, "filter = %d\n", filtered_row[0]);
+   /* set up the zlib input buffer */
+
+   png_ptr->zstream.next_in = filtered_row;
+   png_ptr->zstream.avail_in = (uInt)png_ptr->row_info.rowbytes + 1;
+   /* repeat until we have compressed all the data */
+   do
+   {
+      int ret; /* return of zlib */
+
+      /* compress the data */
+      ret = deflate(&png_ptr->zstream, Z_NO_FLUSH);
+      /* check for compression errors */
+      if (ret != Z_OK)
+      {
+         if (png_ptr->zstream.msg != NULL)
+            png_error(png_ptr, png_ptr->zstream.msg);
+         else
+            png_error(png_ptr, "zlib error");
+      }
+
+      /* see if it is time to write another IDAT */
+      if (!(png_ptr->zstream.avail_out))
+      {
+         /* write the IDAT and reset the zlib output buffer */
+         png_write_IDAT(png_ptr, png_ptr->zbuf, png_ptr->zbuf_size);
+         png_ptr->zstream.next_out = png_ptr->zbuf;
+         png_ptr->zstream.avail_out = (uInt)png_ptr->zbuf_size;
+      }
+   /* repeat until all data has been compressed */
+   } while (png_ptr->zstream.avail_in);
+
+   /* swap the current and previous rows */
+   if (png_ptr->prev_row != NULL)
+   {
+      png_bytep tptr;
+
+      tptr = png_ptr->prev_row;
+      png_ptr->prev_row = png_ptr->row_buf;
+      png_ptr->row_buf = tptr;
+   }
+
+   /* finish row - updates counters and flushes zlib if last row */
+   png_write_finish_row(png_ptr);
+
+#if defined(PNG_WRITE_FLUSH_SUPPORTED)
+   png_ptr->flush_rows++;
+
+   if (png_ptr->flush_dist > 0 &&
+       png_ptr->flush_rows >= png_ptr->flush_dist)
+   {
+      png_write_flush(png_ptr);
+   }
+#endif
+}
+#endif /* PNG_WRITE_SUPPORTED */
diff --git a/distrib/libpng-1.2.19/sources.make b/distrib/libpng-1.2.19/sources.make
new file mode 100644
index 0000000..d8088c7
--- /dev/null
+++ b/distrib/libpng-1.2.19/sources.make
@@ -0,0 +1,14 @@
+# this file is included by various Makefiles and defines the set of sources used by our version of LibPng
+#
+LIBPNG_SOURCES := png.c pngerror.c pngget.c pngmem.c pngpread.c pngread.c \
+                  pngrio.c pngrtran.c pngrutil.c pngset.c pngtrans.c pngvcrd.c pngwio.c \
+                  pngwrite.c pngwtran.c pngwutil.c
+
+ifeq ($(HOST_OS),darwin)
+    LIBPNG_CFLAGS += -DPNG_NO_MMX_CODE
+else
+    LIBPNG_SOURCES += pnggccrd.c
+endif
+
+LIBPNG_SOURCES := $(LIBPNG_SOURCES:%=$(LIBPNG_DIR)/%)
+
diff --git a/distrib/make-distrib.sh b/distrib/make-distrib.sh
new file mode 100755
index 0000000..537bc2c
--- /dev/null
+++ b/distrib/make-distrib.sh
@@ -0,0 +1,112 @@
+#!/bin/bash
+#
+# this script is used to build a source distribution package for the Android emulator
+# the package includes:
+#  - the sources of our patched SDL library
+#  - the sources of our patched QEMU emulator
+#  - appropriate scripts to rebuild the emulator binary
+#
+
+# create temporary directory
+TMPROOT=/tmp/android-package
+DATE=$(date +%Y%m%d)
+PACKAGE=android-emulator-$DATE
+TMPDIR=$TMPROOT/$PACKAGE
+if ! ( rm -rf $TMPROOT && mkdir -p $TMPDIR ) then
+    echo "could not create temporary directory $TMPDIR"
+    exit 3
+fi
+
+locate_qemu_viewpath ()
+{
+    viewpath=$(p4 files $0 | sed -e "s/\(.*\)#.*/\\1/g")
+    # assumes that this program is in the 'distrib' directory of the QEMU sources
+    echo $(dirname $(dirname $viewpath))
+}
+
+locate_depot_files ()
+{
+    root=$(p4 where $1) || (
+        echo "you need to map $1 into your workspace to build an emulator source release package"
+        exit 3
+    )
+    root=$(echo $root | cut -d" " -f3 | sed -e "s%/\.\.\.%%")
+    echo $root
+}
+
+locate_source_files ()
+{
+    files=$(p4 files $1/... | grep -v "delete change" | sed -e "s/\(.*\)#.*/\\1/g")
+    files=$(echo $files | sed -e "s%$1/%%g")
+    echo $files
+}
+
+# locate SDL root directory in client workspace
+if [ -z "$SDLROOT" ] ; then
+    SDLROOT=$(locate_depot_files //toolchain/sdl/...)
+    echo "SDLROOT is $SDLROOT"
+fi
+
+if [ ! -x "$SDLROOT" ] ; then
+    if [ -z "$TOP" ] ; then
+        echo "please define the TOP variable"
+        exit 3
+    fi
+    echo "unable to find $SDLROOT as the SDL root directory"
+    echo "please define SDLROOT to point to the correct location"
+    exit 3
+fi
+
+# locate QEMU root directory
+if  [ -z "$QEMUROOT" ] ; then
+    QEMUVIEW=$(locate_qemu_viewpath)
+    echo "QEMUVIEW is $QEMUVIEW"
+    QEMUROOT=$(locate_depot_files $QEMUVIEW/...)
+    echo "QEMUROOT is $QEMUROOT"
+fi
+
+if [ ! -x "$QEMUROOT" ] ; then
+    if [ -z "$TOP" ] ; then
+        echo "please define the TOP variable"
+        exit 3
+    fi
+    echo "unable to find $QEMUROOT as the QEMU root directory"
+    echo "please define QEMUROOT to point to the correct location"
+    exit 3
+fi
+
+copy_source_files ()
+{
+  DSTDIR=$1
+  SRCDIR=$2
+  files=$(locate_source_files $3)
+  mkdir $DSTDIR && for f in $files; do
+    mkdir -p $(dirname $DSTDIR/$f);
+    cp $SRCDIR/$f $DSTDIR/$f
+  done
+}
+
+# copy and cleanup the SDL sources
+echo "copying SDL sources"
+SDLDIR=$TMPDIR/sdl
+copy_source_files $SDLDIR $SDLROOT //toolchain/sdl
+
+# copy and cleanup the QEMU sources
+echo "copying QEMU sources"
+QEMUDIR=$TMPDIR/qemu
+copy_source_files $QEMUDIR $QEMUROOT $QEMUVIEW
+
+echo "copying control scripts"
+cp -f $QEMUDIR/Makefile.qemu $QEMUDIR/Makefile
+rm -f $QEMUDIR/Makefile.android
+cp $QEMUDIR/distrib/build-emulator.sh $TMPDIR/build-emulator.sh
+cp $QEMUDIR/distrib/README $TMPDIR/README
+
+echo "packaging release into a tarball"
+cd $TMPROOT
+tar cjf $PACKAGE.tar.bz2 $PACKAGE
+
+echo "cleaning up"
+rm -rf $TMPDIR
+
+echo "please grab $TMPROOT/$PACKAGE.tar.bz2"
diff --git a/distrib/update-audio.sh b/distrib/update-audio.sh
new file mode 100755
index 0000000..8afd3c4
--- /dev/null
+++ b/distrib/update-audio.sh
@@ -0,0 +1,85 @@
+#!/bin/bash
+#
+# this script is used to update the prebuilt libqemu-audio.a file in the Android source tree
+# we use a prebuilt package because we don't want to force the installation of the ALSA / EsounD / Whatever
+# development packages on every developer machine, or every build server.
+#
+
+# assumes this script is located in the 'distrib' sub-directory
+cd `dirname $0`
+cd ..
+
+locate_depot_files ()
+{
+    root=$(p4 where $1) || (
+        echo "you need to map $1 into your workspace to build an emulator source release package"
+        exit 3
+    )
+    root=$(echo $root | cut -d" " -f3 | sed -e "s%/\.\.\.%%")
+    echo $root
+}
+
+# find the prebuilt directory
+OS=`uname -s`
+EXE=""
+case "$OS" in
+    Darwin)
+        CPU=`uname -p`
+        if [ "$CPU" == "i386" ] ; then
+            OS=darwin-x86
+        else
+            OS=darwin-ppc
+        fi
+        ;;
+    *_NT-5.1)
+        OS=windows
+        EXE=.exe
+        ;;
+esac
+
+PREBUILT=$(locate_depot_files //device/prebuilt/$OS)
+
+# find the GNU Make program
+is_gnu_make ()
+{
+    version=$($1 -v | grep GNU)
+    if test -n "$version"; then
+        echo "$1"
+    else
+        echo ""
+    fi
+}
+
+if test -z "$GNUMAKE"; then
+    GNUMAKE=`which make` && GNUMAKE=$(is_gnu_make $GNUMAKE)
+fi
+
+if test -z "$GNUMAKE"; then
+    GNUMAKE=`which gmake` && GNUMAKE=$(is_gnu_make $GNUMAKE)
+fi
+
+if test -z "$GNUMAKE"; then
+    echo "could not find GNU Make on this machine. please define GNUMAKE to point to it"
+    exit 3
+fi
+
+TEST=$(is_gnu_make $GNUMAKE)
+if test -z "$TEST"; then
+    echo "it seems that '$GNUMAKE' is not a working GNU Make binary. please check the definition of GNUMAKE"
+    exit 3
+fi
+
+# ensure we have a recent audio library built
+#
+#echo "GNUMAKE is $GNUMAKE"
+source=arm-softmmu/libqemu-audio.a
+$GNUMAKE $source || (echo "could not build the audio library. Aborting" && exit 1)
+
+# now do a p4 edit, a copy and ask for submission
+#
+TARGET=$PREBUILT/qemu/libqemu-audio.a
+
+p4 edit $TARGET || (echo "could not p4 edit $TARGET" && exit 3)
+cp -f $source $TARGET
+echo "please do: p4 submit $TARGET"
+
diff --git a/distrib/zlib-1.2.3/Makefile b/distrib/zlib-1.2.3/Makefile
new file mode 100644
index 0000000..9cf80c9
--- /dev/null
+++ b/distrib/zlib-1.2.3/Makefile
@@ -0,0 +1,15 @@
+# Makefile used to compile zlib statically
+#
+ZLIB_LIB    := $(SRC_PATH)/libz.a
+ZLIB_CFLAGS := -I$(ZLIB_DIR)
+
+include $(ZLIB_DIR)/sources.make
+ZLIB_OBJS := $(ZLIB_SOURCES:%.c=%.o)
+
+$(ZLIB_LIB): $(ZLIB_OBJS)
+	ar ru $(ZLIB_LIB) $(ZLIB_OBJS)
+
+$(ZLIB_OBJS): CFLAGS += $(ZLIB_CFLAGS)
+
+clean-zlib:
+	rm -f $(ZLIB_OBJS) $(zlib_lib)
diff --git a/distrib/zlib-1.2.3/adler32.c b/distrib/zlib-1.2.3/adler32.c
new file mode 100644
index 0000000..007ba26
--- /dev/null
+++ b/distrib/zlib-1.2.3/adler32.c
@@ -0,0 +1,149 @@
+/* adler32.c -- compute the Adler-32 checksum of a data stream
+ * Copyright (C) 1995-2004 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* @(#) $Id$ */
+
+#define ZLIB_INTERNAL
+#include "zlib.h"
+
+#define BASE 65521UL    /* largest prime smaller than 65536 */
+#define NMAX 5552
+/* NMAX is the largest n such that 255n(n+1)/2 + (n+1)(BASE-1) <= 2^32-1 */
+
+#define DO1(buf,i)  {adler += (buf)[i]; sum2 += adler;}
+#define DO2(buf,i)  DO1(buf,i); DO1(buf,i+1);
+#define DO4(buf,i)  DO2(buf,i); DO2(buf,i+2);
+#define DO8(buf,i)  DO4(buf,i); DO4(buf,i+4);
+#define DO16(buf)   DO8(buf,0); DO8(buf,8);
+
+/* use NO_DIVIDE if your processor does not do division in hardware */
+#ifdef NO_DIVIDE
+#  define MOD(a) \
+    do { \
+        if (a >= (BASE << 16)) a -= (BASE << 16); \
+        if (a >= (BASE << 15)) a -= (BASE << 15); \
+        if (a >= (BASE << 14)) a -= (BASE << 14); \
+        if (a >= (BASE << 13)) a -= (BASE << 13); \
+        if (a >= (BASE << 12)) a -= (BASE << 12); \
+        if (a >= (BASE << 11)) a -= (BASE << 11); \
+        if (a >= (BASE << 10)) a -= (BASE << 10); \
+        if (a >= (BASE << 9)) a -= (BASE << 9); \
+        if (a >= (BASE << 8)) a -= (BASE << 8); \
+        if (a >= (BASE << 7)) a -= (BASE << 7); \
+        if (a >= (BASE << 6)) a -= (BASE << 6); \
+        if (a >= (BASE << 5)) a -= (BASE << 5); \
+        if (a >= (BASE << 4)) a -= (BASE << 4); \
+        if (a >= (BASE << 3)) a -= (BASE << 3); \
+        if (a >= (BASE << 2)) a -= (BASE << 2); \
+        if (a >= (BASE << 1)) a -= (BASE << 1); \
+        if (a >= BASE) a -= BASE; \
+    } while (0)
+#  define MOD4(a) \
+    do { \
+        if (a >= (BASE << 4)) a -= (BASE << 4); \
+        if (a >= (BASE << 3)) a -= (BASE << 3); \
+        if (a >= (BASE << 2)) a -= (BASE << 2); \
+        if (a >= (BASE << 1)) a -= (BASE << 1); \
+        if (a >= BASE) a -= BASE; \
+    } while (0)
+#else
+#  define MOD(a) a %= BASE
+#  define MOD4(a) a %= BASE
+#endif
+
+/* ========================================================================= */
+uLong ZEXPORT adler32(adler, buf, len)
+    uLong adler;
+    const Bytef *buf;
+    uInt len;
+{
+    unsigned long sum2;
+    unsigned n;
+
+    /* split Adler-32 into component sums */
+    sum2 = (adler >> 16) & 0xffff;
+    adler &= 0xffff;
+
+    /* in case user likes doing a byte at a time, keep it fast */
+    if (len == 1) {
+        adler += buf[0];
+        if (adler >= BASE)
+            adler -= BASE;
+        sum2 += adler;
+        if (sum2 >= BASE)
+            sum2 -= BASE;
+        return adler | (sum2 << 16);
+    }
+
+    /* initial Adler-32 value (deferred check for len == 1 speed) */
+    if (buf == Z_NULL)
+        return 1L;
+
+    /* in case short lengths are provided, keep it somewhat fast */
+    if (len < 16) {
+        while (len--) {
+            adler += *buf++;
+            sum2 += adler;
+        }
+        if (adler >= BASE)
+            adler -= BASE;
+        MOD4(sum2);             /* only added so many BASE's */
+        return adler | (sum2 << 16);
+    }
+
+    /* do length NMAX blocks -- requires just one modulo operation */
+    while (len >= NMAX) {
+        len -= NMAX;
+        n = NMAX / 16;          /* NMAX is divisible by 16 */
+        do {
+            DO16(buf);          /* 16 sums unrolled */
+            buf += 16;
+        } while (--n);
+        MOD(adler);
+        MOD(sum2);
+    }
+
+    /* do remaining bytes (less than NMAX, still just one modulo) */
+    if (len) {                  /* avoid modulos if none remaining */
+        while (len >= 16) {
+            len -= 16;
+            DO16(buf);
+            buf += 16;
+        }
+        while (len--) {
+            adler += *buf++;
+            sum2 += adler;
+        }
+        MOD(adler);
+        MOD(sum2);
+    }
+
+    /* return recombined sums */
+    return adler | (sum2 << 16);
+}
+
+/* ========================================================================= */
+uLong ZEXPORT adler32_combine(adler1, adler2, len2)
+    uLong adler1;
+    uLong adler2;
+    z_off_t len2;
+{
+    unsigned long sum1;
+    unsigned long sum2;
+    unsigned rem;
+
+    /* the derivation of this formula is left as an exercise for the reader */
+    rem = (unsigned)(len2 % BASE);
+    sum1 = adler1 & 0xffff;
+    sum2 = rem * sum1;
+    MOD(sum2);
+    sum1 += (adler2 & 0xffff) + BASE - 1;
+    sum2 += ((adler1 >> 16) & 0xffff) + ((adler2 >> 16) & 0xffff) + BASE - rem;
+    if (sum1 > BASE) sum1 -= BASE;
+    if (sum1 > BASE) sum1 -= BASE;
+    if (sum2 > (BASE << 1)) sum2 -= (BASE << 1);
+    if (sum2 > BASE) sum2 -= BASE;
+    return sum1 | (sum2 << 16);
+}
diff --git a/distrib/zlib-1.2.3/compress.c b/distrib/zlib-1.2.3/compress.c
new file mode 100644
index 0000000..df04f01
--- /dev/null
+++ b/distrib/zlib-1.2.3/compress.c
@@ -0,0 +1,79 @@
+/* compress.c -- compress a memory buffer
+ * Copyright (C) 1995-2003 Jean-loup Gailly.
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* @(#) $Id$ */
+
+#define ZLIB_INTERNAL
+#include "zlib.h"
+
+/* ===========================================================================
+     Compresses the source buffer into the destination buffer. The level
+   parameter has the same meaning as in deflateInit.  sourceLen is the byte
+   length of the source buffer. Upon entry, destLen is the total size of the
+   destination buffer, which must be at least 0.1% larger than sourceLen plus
+   12 bytes. Upon exit, destLen is the actual size of the compressed buffer.
+
+     compress2 returns Z_OK if success, Z_MEM_ERROR if there was not enough
+   memory, Z_BUF_ERROR if there was not enough room in the output buffer,
+   Z_STREAM_ERROR if the level parameter is invalid.
+*/
+int ZEXPORT compress2 (dest, destLen, source, sourceLen, level)
+    Bytef *dest;
+    uLongf *destLen;
+    const Bytef *source;
+    uLong sourceLen;
+    int level;
+{
+    z_stream stream;
+    int err;
+
+    stream.next_in = (Bytef*)source;
+    stream.avail_in = (uInt)sourceLen;
+#ifdef MAXSEG_64K
+    /* Check for source > 64K on 16-bit machine: */
+    if ((uLong)stream.avail_in != sourceLen) return Z_BUF_ERROR;
+#endif
+    stream.next_out = dest;
+    stream.avail_out = (uInt)*destLen;
+    if ((uLong)stream.avail_out != *destLen) return Z_BUF_ERROR;
+
+    stream.zalloc = (alloc_func)0;
+    stream.zfree = (free_func)0;
+    stream.opaque = (voidpf)0;
+
+    err = deflateInit(&stream, level);
+    if (err != Z_OK) return err;
+
+    err = deflate(&stream, Z_FINISH);
+    if (err != Z_STREAM_END) {
+        deflateEnd(&stream);
+        return err == Z_OK ? Z_BUF_ERROR : err;
+    }
+    *destLen = stream.total_out;
+
+    err = deflateEnd(&stream);
+    return err;
+}
+
+/* ===========================================================================
+ */
+int ZEXPORT compress (dest, destLen, source, sourceLen)
+    Bytef *dest;
+    uLongf *destLen;
+    const Bytef *source;
+    uLong sourceLen;
+{
+    return compress2(dest, destLen, source, sourceLen, Z_DEFAULT_COMPRESSION);
+}
+
+/* ===========================================================================
+     If the default memLevel or windowBits for deflateInit() is changed, then
+   this function needs to be updated.
+ */
+uLong ZEXPORT compressBound (sourceLen)
+    uLong sourceLen;
+{
+    return sourceLen + (sourceLen >> 12) + (sourceLen >> 14) + 11;
+}
diff --git a/distrib/zlib-1.2.3/configure b/distrib/zlib-1.2.3/configure
new file mode 100755
index 0000000..d7ffdc3
--- /dev/null
+++ b/distrib/zlib-1.2.3/configure
@@ -0,0 +1,459 @@
+#!/bin/sh
+# configure script for zlib. This script is needed only if
+# you wish to build a shared library and your system supports them,
+# of if you need special compiler, flags or install directory.
+# Otherwise, you can just use directly "make test; make install"
+#
+# To create a shared library, use "configure --shared"; by default a static
+# library is created. If the primitive shared library support provided here
+# does not work, use ftp://prep.ai.mit.edu/pub/gnu/libtool-*.tar.gz
+#
+# To impose specific compiler or flags or install directory, use for example:
+#    prefix=$HOME CC=cc CFLAGS="-O4" ./configure
+# or for csh/tcsh users:
+#    (setenv prefix $HOME; setenv CC cc; setenv CFLAGS "-O4"; ./configure)
+# LDSHARED is the command to be used to create a shared library
+
+# Incorrect settings of CC or CFLAGS may prevent creating a shared library.
+# If you have problems, try without defining CC and CFLAGS before reporting
+# an error.
+
+LIBS=libz.a
+LDFLAGS="-L. ${LIBS}"
+VER=`sed -n -e '/VERSION "/s/.*"\(.*\)".*/\1/p' < zlib.h`
+VER2=`sed -n -e '/VERSION "/s/.*"\([0-9]*\\.[0-9]*\)\\..*/\1/p' < zlib.h`
+VER1=`sed -n -e '/VERSION "/s/.*"\([0-9]*\)\\..*/\1/p' < zlib.h`
+AR=${AR-"ar rc"}
+RANLIB=${RANLIB-"ranlib"}
+prefix=${prefix-/usr/local}
+exec_prefix=${exec_prefix-'${prefix}'}
+libdir=${libdir-'${exec_prefix}/lib'}
+includedir=${includedir-'${prefix}/include'}
+mandir=${mandir-'${prefix}/share/man'}
+shared_ext='.so'
+shared=0
+gcc=0
+old_cc="$CC"
+old_cflags="$CFLAGS"
+
+while test $# -ge 1
+do
+case "$1" in
+    -h* | --h*)
+      echo 'usage:'
+      echo '  configure [--shared] [--prefix=PREFIX]  [--exec_prefix=EXPREFIX]'
+      echo '     [--libdir=LIBDIR] [--includedir=INCLUDEDIR]'
+        exit 0;;
+    -p*=* | --p*=*) prefix=`echo $1 | sed 's/[-a-z_]*=//'`; shift;;
+    -e*=* | --e*=*) exec_prefix=`echo $1 | sed 's/[-a-z_]*=//'`; shift;;
+    -l*=* | --libdir=*) libdir=`echo $1 | sed 's/[-a-z_]*=//'`; shift;;
+    -i*=* | --includedir=*) includedir=`echo $1 | sed 's/[-a-z_]*=//'`;shift;;
+    -p* | --p*) prefix="$2"; shift; shift;;
+    -e* | --e*) exec_prefix="$2"; shift; shift;;
+    -l* | --l*) libdir="$2"; shift; shift;;
+    -i* | --i*) includedir="$2"; shift; shift;;
+    -s* | --s*) shared=1; shift;;
+    *) echo "unknown option: $1"; echo "$0 --help for help"; exit 1;;
+    esac
+done
+
+test=ztest$$
+cat > $test.c <<EOF
+extern int getchar();
+int hello() {return getchar();}
+EOF
+
+test -z "$CC" && echo Checking for gcc...
+cc=${CC-gcc}
+cflags=${CFLAGS-"-O3"}
+# to force the asm version use: CFLAGS="-O3 -DASMV" ./configure
+case "$cc" in
+  *gcc*) gcc=1;;
+esac
+
+if test "$gcc" -eq 1 && ($cc -c $cflags $test.c) 2>/dev/null; then
+  CC="$cc"
+  SFLAGS=${CFLAGS-"-fPIC -O3"}
+  CFLAGS="$cflags"
+  case `(uname -s || echo unknown) 2>/dev/null` in
+  Linux | linux | GNU | GNU/*) LDSHARED=${LDSHARED-"$cc -shared -Wl,-soname,libz.so.1"};;
+  CYGWIN* | Cygwin* | cygwin* | OS/2* )
+             EXE='.exe';;
+  QNX*)  # This is for QNX6. I suppose that the QNX rule below is for QNX2,QNX4
+         # (alain.bonnefoy@icbt.com)
+                 LDSHARED=${LDSHARED-"$cc -shared -Wl,-hlibz.so.1"};;
+  HP-UX*)
+         LDSHARED=${LDSHARED-"$cc -shared $SFLAGS"}
+         case `(uname -m || echo unknown) 2>/dev/null` in
+         ia64)
+                 shared_ext='.so'
+                 SHAREDLIB='libz.so';;
+         *)
+                 shared_ext='.sl'
+                 SHAREDLIB='libz.sl';;
+         esac;;
+  Darwin*)   shared_ext='.dylib'
+             SHAREDLIB=libz$shared_ext
+             SHAREDLIBV=libz.$VER$shared_ext
+             SHAREDLIBM=libz.$VER1$shared_ext
+             LDSHARED=${LDSHARED-"$cc -dynamiclib -install_name $libdir/$SHAREDLIBM -compatibility_version $VER1 -current_version $VER"};;
+  *)             LDSHARED=${LDSHARED-"$cc -shared"};;
+  esac
+else
+  # find system name and corresponding cc options
+  CC=${CC-cc}
+  case `(uname -sr || echo unknown) 2>/dev/null` in
+  HP-UX*)    SFLAGS=${CFLAGS-"-O +z"}
+             CFLAGS=${CFLAGS-"-O"}
+#            LDSHARED=${LDSHARED-"ld -b +vnocompatwarnings"}
+             LDSHARED=${LDSHARED-"ld -b"}
+         case `(uname -m || echo unknown) 2>/dev/null` in
+         ia64)
+             shared_ext='.so'
+             SHAREDLIB='libz.so';;
+         *)
+             shared_ext='.sl'
+             SHAREDLIB='libz.sl';;
+         esac;;
+  IRIX*)     SFLAGS=${CFLAGS-"-ansi -O2 -rpath ."}
+             CFLAGS=${CFLAGS-"-ansi -O2"}
+             LDSHARED=${LDSHARED-"cc -shared"};;
+  OSF1\ V4*) SFLAGS=${CFLAGS-"-O -std1"}
+             CFLAGS=${CFLAGS-"-O -std1"}
+             LDSHARED=${LDSHARED-"cc -shared  -Wl,-soname,libz.so -Wl,-msym -Wl,-rpath,$(libdir) -Wl,-set_version,${VER}:1.0"};;
+  OSF1*)     SFLAGS=${CFLAGS-"-O -std1"}
+             CFLAGS=${CFLAGS-"-O -std1"}
+             LDSHARED=${LDSHARED-"cc -shared"};;
+  QNX*)      SFLAGS=${CFLAGS-"-4 -O"}
+             CFLAGS=${CFLAGS-"-4 -O"}
+             LDSHARED=${LDSHARED-"cc"}
+             RANLIB=${RANLIB-"true"}
+             AR="cc -A";;
+  SCO_SV\ 3.2*) SFLAGS=${CFLAGS-"-O3 -dy -KPIC "}
+             CFLAGS=${CFLAGS-"-O3"}
+             LDSHARED=${LDSHARED-"cc -dy -KPIC -G"};;
+  SunOS\ 5*) SFLAGS=${CFLAGS-"-fast -xcg89 -KPIC -R."}
+             CFLAGS=${CFLAGS-"-fast -xcg89"}
+             LDSHARED=${LDSHARED-"cc -G"};;
+  SunOS\ 4*) SFLAGS=${CFLAGS-"-O2 -PIC"}
+             CFLAGS=${CFLAGS-"-O2"}
+             LDSHARED=${LDSHARED-"ld"};;
+  SunStudio\ 9*) SFLAGS=${CFLAGS-"-DUSE_MMAP -fast -xcode=pic32 -xtarget=ultra3 -xarch=v9b"}
+             CFLAGS=${CFLAGS-"-DUSE_MMAP -fast -xtarget=ultra3 -xarch=v9b"}
+             LDSHARED=${LDSHARED-"cc -xarch=v9b"};;
+  UNIX_System_V\ 4.2.0)
+             SFLAGS=${CFLAGS-"-KPIC -O"}
+             CFLAGS=${CFLAGS-"-O"}
+             LDSHARED=${LDSHARED-"cc -G"};;
+  UNIX_SV\ 4.2MP)
+             SFLAGS=${CFLAGS-"-Kconform_pic -O"}
+             CFLAGS=${CFLAGS-"-O"}
+             LDSHARED=${LDSHARED-"cc -G"};;
+  OpenUNIX\ 5)
+             SFLAGS=${CFLAGS-"-KPIC -O"}
+             CFLAGS=${CFLAGS-"-O"}
+             LDSHARED=${LDSHARED-"cc -G"};;
+  AIX*)  # Courtesy of dbakker@arrayasolutions.com
+             SFLAGS=${CFLAGS-"-O -qmaxmem=8192"}
+             CFLAGS=${CFLAGS-"-O -qmaxmem=8192"}
+             LDSHARED=${LDSHARED-"xlc -G"};;
+  # send working options for other systems to support@gzip.org
+  *)         SFLAGS=${CFLAGS-"-O"}
+             CFLAGS=${CFLAGS-"-O"}
+             LDSHARED=${LDSHARED-"cc -shared"};;
+  esac
+fi
+
+SHAREDLIB=${SHAREDLIB-"libz$shared_ext"}
+SHAREDLIBV=${SHAREDLIBV-"libz$shared_ext.$VER"}
+SHAREDLIBM=${SHAREDLIBM-"libz$shared_ext.$VER1"}
+
+if test $shared -eq 1; then
+  echo Checking for shared library support...
+  # we must test in two steps (cc then ld), required at least on SunOS 4.x
+  if test "`($CC -c $SFLAGS $test.c) 2>&1`" = "" &&
+     test "`($LDSHARED -o $test$shared_ext $test.o) 2>&1`" = ""; then
+    CFLAGS="$SFLAGS"
+    LIBS="$SHAREDLIBV"
+    echo Building shared library $SHAREDLIBV with $CC.
+  elif test -z "$old_cc" -a -z "$old_cflags"; then
+    echo No shared library support.
+    shared=0;
+  else
+    echo 'No shared library support; try without defining CC and CFLAGS'
+    shared=0;
+  fi
+fi
+if test $shared -eq 0; then
+  LDSHARED="$CC"
+  echo Building static library $LIBS version $VER with $CC.
+else
+  LDFLAGS="-L. ${SHAREDLIBV}"
+fi
+
+cat > $test.c <<EOF
+#include <unistd.h>
+int main() { return 0; }
+EOF
+if test "`($CC -c $CFLAGS $test.c) 2>&1`" = ""; then
+  sed < zconf.in.h "/HAVE_UNISTD_H/s%0%1%" > zconf.h
+  echo "Checking for unistd.h... Yes."
+else
+  cp -p zconf.in.h zconf.h
+  echo "Checking for unistd.h... No."
+fi
+
+cat > $test.c <<EOF
+#include <stdio.h>
+#include <stdarg.h>
+#include "zconf.h"
+
+int main()
+{
+#ifndef STDC
+  choke me
+#endif
+
+  return 0;
+}
+EOF
+
+if test "`($CC -c $CFLAGS $test.c) 2>&1`" = ""; then
+  echo "Checking whether to use vs[n]printf() or s[n]printf()... using vs[n]printf()"
+
+  cat > $test.c <<EOF
+#include <stdio.h>
+#include <stdarg.h>
+
+int mytest(char *fmt, ...)
+{
+  char buf[20];
+  va_list ap;
+
+  va_start(ap, fmt);
+  vsnprintf(buf, sizeof(buf), fmt, ap);
+  va_end(ap);
+  return 0;
+}
+
+int main()
+{
+  return (mytest("Hello%d\n", 1));
+}
+EOF
+
+  if test "`($CC $CFLAGS -o $test $test.c) 2>&1`" = ""; then
+    echo "Checking for vsnprintf() in stdio.h... Yes."
+
+    cat >$test.c <<EOF
+#include <stdio.h>
+#include <stdarg.h>
+
+int mytest(char *fmt, ...)
+{
+  int n;
+  char buf[20];
+  va_list ap;
+
+  va_start(ap, fmt);
+  n = vsnprintf(buf, sizeof(buf), fmt, ap);
+  va_end(ap);
+  return n;
+}
+
+int main()
+{
+  return (mytest("Hello%d\n", 1));
+}
+EOF
+
+    if test "`($CC -c $CFLAGS $test.c) 2>&1`" = ""; then
+      echo "Checking for return value of vsnprintf()... Yes."
+    else
+      CFLAGS="$CFLAGS -DHAS_vsnprintf_void"
+      echo "Checking for return value of vsnprintf()... No."
+      echo "  WARNING: apparently vsnprintf() does not return a value. zlib"
+      echo "  can build but will be open to possible string-format security"
+      echo "  vulnerabilities."
+    fi
+  else
+    CFLAGS="$CFLAGS -DNO_vsnprintf"
+    echo "Checking for vsnprintf() in stdio.h... No."
+    echo "  WARNING: vsnprintf() not found, falling back to vsprintf(). zlib"
+    echo "  can build but will be open to possible buffer-overflow security"
+    echo "  vulnerabilities."
+
+    cat >$test.c <<EOF
+#include <stdio.h>
+#include <stdarg.h>
+
+int mytest(char *fmt, ...)
+{
+  int n;
+  char buf[20];
+  va_list ap;
+
+  va_start(ap, fmt);
+  n = vsprintf(buf, fmt, ap);
+  va_end(ap);
+  return n;
+}
+
+int main()
+{
+  return (mytest("Hello%d\n", 1));
+}
+EOF
+
+    if test "`($CC -c $CFLAGS $test.c) 2>&1`" = ""; then
+      echo "Checking for return value of vsprintf()... Yes."
+    else
+      CFLAGS="$CFLAGS -DHAS_vsprintf_void"
+      echo "Checking for return value of vsprintf()... No."
+      echo "  WARNING: apparently vsprintf() does not return a value. zlib"
+      echo "  can build but will be open to possible string-format security"
+      echo "  vulnerabilities."
+    fi
+  fi
+else
+  echo "Checking whether to use vs[n]printf() or s[n]printf()... using s[n]printf()"
+
+  cat >$test.c <<EOF
+#include <stdio.h>
+
+int mytest()
+{
+  char buf[20];
+
+  snprintf(buf, sizeof(buf), "%s", "foo");
+  return 0;
+}
+
+int main()
+{
+  return (mytest());
+}
+EOF
+
+  if test "`($CC $CFLAGS -o $test $test.c) 2>&1`" = ""; then
+    echo "Checking for snprintf() in stdio.h... Yes."
+
+    cat >$test.c <<EOF
+#include <stdio.h>
+
+int mytest()
+{
+  char buf[20];
+
+  return snprintf(buf, sizeof(buf), "%s", "foo");
+}
+
+int main()
+{
+  return (mytest());
+}
+EOF
+
+    if test "`($CC -c $CFLAGS $test.c) 2>&1`" = ""; then
+      echo "Checking for return value of snprintf()... Yes."
+    else
+      CFLAGS="$CFLAGS -DHAS_snprintf_void"
+      echo "Checking for return value of snprintf()... No."
+      echo "  WARNING: apparently snprintf() does not return a value. zlib"
+      echo "  can build but will be open to possible string-format security"
+      echo "  vulnerabilities."
+    fi
+  else
+    CFLAGS="$CFLAGS -DNO_snprintf"
+    echo "Checking for snprintf() in stdio.h... No."
+    echo "  WARNING: snprintf() not found, falling back to sprintf(). zlib"
+    echo "  can build but will be open to possible buffer-overflow security"
+    echo "  vulnerabilities."
+
+    cat >$test.c <<EOF
+#include <stdio.h>
+
+int mytest()
+{
+  char buf[20];
+
+  return sprintf(buf, "%s", "foo");
+}
+
+int main()
+{
+  return (mytest());
+}
+EOF
+
+    if test "`($CC -c $CFLAGS $test.c) 2>&1`" = ""; then
+      echo "Checking for return value of sprintf()... Yes."
+    else
+      CFLAGS="$CFLAGS -DHAS_sprintf_void"
+      echo "Checking for return value of sprintf()... No."
+      echo "  WARNING: apparently sprintf() does not return a value. zlib"
+      echo "  can build but will be open to possible string-format security"
+      echo "  vulnerabilities."
+    fi
+  fi
+fi
+
+cat >$test.c <<EOF
+#include <errno.h>
+int main() { return 0; }
+EOF
+if test "`($CC -c $CFLAGS $test.c) 2>&1`" = ""; then
+  echo "Checking for errno.h... Yes."
+else
+  echo "Checking for errno.h... No."
+  CFLAGS="$CFLAGS -DNO_ERRNO_H"
+fi
+
+cat > $test.c <<EOF
+#include <sys/types.h>
+#include <sys/mman.h>
+#include <sys/stat.h>
+caddr_t hello() {
+  return mmap((caddr_t)0, (off_t)0, PROT_READ, MAP_SHARED, 0, (off_t)0);
+}
+EOF
+if test "`($CC -c $CFLAGS $test.c) 2>&1`" = ""; then
+  CFLAGS="$CFLAGS -DUSE_MMAP"
+  echo Checking for mmap support... Yes.
+else
+  echo Checking for mmap support... No.
+fi
+
+CPP=${CPP-"$CC -E"}
+case $CFLAGS in
+  *ASMV*)
+    if test "`nm $test.o | grep _hello`" = ""; then
+      CPP="$CPP -DNO_UNDERLINE"
+      echo Checking for underline in external names... No.
+    else
+      echo Checking for underline in external names... Yes.
+    fi;;
+esac
+
+rm -f $test.[co] $test $test$shared_ext
+
+# udpate Makefile
+sed < Makefile.in "
+/^CC *=/s#=.*#=$CC#
+/^CFLAGS *=/s#=.*#=$CFLAGS#
+/^CPP *=/s#=.*#=$CPP#
+/^LDSHARED *=/s#=.*#=$LDSHARED#
+/^LIBS *=/s#=.*#=$LIBS#
+/^SHAREDLIB *=/s#=.*#=$SHAREDLIB#
+/^SHAREDLIBV *=/s#=.*#=$SHAREDLIBV#
+/^SHAREDLIBM *=/s#=.*#=$SHAREDLIBM#
+/^AR *=/s#=.*#=$AR#
+/^RANLIB *=/s#=.*#=$RANLIB#
+/^EXE *=/s#=.*#=$EXE#
+/^prefix *=/s#=.*#=$prefix#
+/^exec_prefix *=/s#=.*#=$exec_prefix#
+/^libdir *=/s#=.*#=$libdir#
+/^includedir *=/s#=.*#=$includedir#
+/^mandir *=/s#=.*#=$mandir#
+/^LDFLAGS *=/s#=.*#=$LDFLAGS#
+" > Makefile
diff --git a/distrib/zlib-1.2.3/crc32.c b/distrib/zlib-1.2.3/crc32.c
new file mode 100644
index 0000000..f658a9e
--- /dev/null
+++ b/distrib/zlib-1.2.3/crc32.c
@@ -0,0 +1,423 @@
+/* crc32.c -- compute the CRC-32 of a data stream
+ * Copyright (C) 1995-2005 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ *
+ * Thanks to Rodney Brown <rbrown64@csc.com.au> for his contribution of faster
+ * CRC methods: exclusive-oring 32 bits of data at a time, and pre-computing
+ * tables for updating the shift register in one step with three exclusive-ors
+ * instead of four steps with four exclusive-ors.  This results in about a
+ * factor of two increase in speed on a Power PC G4 (PPC7455) using gcc -O3.
+ */
+
+/* @(#) $Id$ */
+
+/*
+  Note on the use of DYNAMIC_CRC_TABLE: there is no mutex or semaphore
+  protection on the static variables used to control the first-use generation
+  of the crc tables.  Therefore, if you #define DYNAMIC_CRC_TABLE, you should
+  first call get_crc_table() to initialize the tables before allowing more than
+  one thread to use crc32().
+ */
+
+#ifdef MAKECRCH
+#  include <stdio.h>
+#  ifndef DYNAMIC_CRC_TABLE
+#    define DYNAMIC_CRC_TABLE
+#  endif /* !DYNAMIC_CRC_TABLE */
+#endif /* MAKECRCH */
+
+#include "zutil.h"      /* for STDC and FAR definitions */
+
+#define local static
+
+/* Find a four-byte integer type for crc32_little() and crc32_big(). */
+#ifndef NOBYFOUR
+#  ifdef STDC           /* need ANSI C limits.h to determine sizes */
+#    include <limits.h>
+#    define BYFOUR
+#    if (UINT_MAX == 0xffffffffUL)
+       typedef unsigned int u4;
+#    else
+#      if (ULONG_MAX == 0xffffffffUL)
+         typedef unsigned long u4;
+#      else
+#        if (USHRT_MAX == 0xffffffffUL)
+           typedef unsigned short u4;
+#        else
+#          undef BYFOUR     /* can't find a four-byte integer type! */
+#        endif
+#      endif
+#    endif
+#  endif /* STDC */
+#endif /* !NOBYFOUR */
+
+/* Definitions for doing the crc four data bytes at a time. */
+#ifdef BYFOUR
+#  define REV(w) (((w)>>24)+(((w)>>8)&0xff00)+ \
+                (((w)&0xff00)<<8)+(((w)&0xff)<<24))
+   local unsigned long crc32_little OF((unsigned long,
+                        const unsigned char FAR *, unsigned));
+   local unsigned long crc32_big OF((unsigned long,
+                        const unsigned char FAR *, unsigned));
+#  define TBLS 8
+#else
+#  define TBLS 1
+#endif /* BYFOUR */
+
+/* Local functions for crc concatenation */
+local unsigned long gf2_matrix_times OF((unsigned long *mat,
+                                         unsigned long vec));
+local void gf2_matrix_square OF((unsigned long *square, unsigned long *mat));
+
+#ifdef DYNAMIC_CRC_TABLE
+
+local volatile int crc_table_empty = 1;
+local unsigned long FAR crc_table[TBLS][256];
+local void make_crc_table OF((void));
+#ifdef MAKECRCH
+   local void write_table OF((FILE *, const unsigned long FAR *));
+#endif /* MAKECRCH */
+/*
+  Generate tables for a byte-wise 32-bit CRC calculation on the polynomial:
+  x^32+x^26+x^23+x^22+x^16+x^12+x^11+x^10+x^8+x^7+x^5+x^4+x^2+x+1.
+
+  Polynomials over GF(2) are represented in binary, one bit per coefficient,
+  with the lowest powers in the most significant bit.  Then adding polynomials
+  is just exclusive-or, and multiplying a polynomial by x is a right shift by
+  one.  If we call the above polynomial p, and represent a byte as the
+  polynomial q, also with the lowest power in the most significant bit (so the
+  byte 0xb1 is the polynomial x^7+x^3+x+1), then the CRC is (q*x^32) mod p,
+  where a mod b means the remainder after dividing a by b.
+
+  This calculation is done using the shift-register method of multiplying and
+  taking the remainder.  The register is initialized to zero, and for each
+  incoming bit, x^32 is added mod p to the register if the bit is a one (where
+  x^32 mod p is p+x^32 = x^26+...+1), and the register is multiplied mod p by
+  x (which is shifting right by one and adding x^32 mod p if the bit shifted
+  out is a one).  We start with the highest power (least significant bit) of
+  q and repeat for all eight bits of q.
+
+  The first table is simply the CRC of all possible eight bit values.  This is
+  all the information needed to generate CRCs on data a byte at a time for all
+  combinations of CRC register values and incoming bytes.  The remaining tables
+  allow for word-at-a-time CRC calculation for both big-endian and little-
+  endian machines, where a word is four bytes.
+*/
+local void make_crc_table()
+{
+    unsigned long c;
+    int n, k;
+    unsigned long poly;                 /* polynomial exclusive-or pattern */
+    /* terms of polynomial defining this crc (except x^32): */
+    static volatile int first = 1;      /* flag to limit concurrent making */
+    static const unsigned char p[] = {0,1,2,4,5,7,8,10,11,12,16,22,23,26};
+
+    /* See if another task is already doing this (not thread-safe, but better
+       than nothing -- significantly reduces duration of vulnerability in
+       case the advice about DYNAMIC_CRC_TABLE is ignored) */
+    if (first) {
+        first = 0;
+
+        /* make exclusive-or pattern from polynomial (0xedb88320UL) */
+        poly = 0UL;
+        for (n = 0; n < sizeof(p)/sizeof(unsigned char); n++)
+            poly |= 1UL << (31 - p[n]);
+
+        /* generate a crc for every 8-bit value */
+        for (n = 0; n < 256; n++) {
+            c = (unsigned long)n;
+            for (k = 0; k < 8; k++)
+                c = c & 1 ? poly ^ (c >> 1) : c >> 1;
+            crc_table[0][n] = c;
+        }
+
+#ifdef BYFOUR
+        /* generate crc for each value followed by one, two, and three zeros,
+           and then the byte reversal of those as well as the first table */
+        for (n = 0; n < 256; n++) {
+            c = crc_table[0][n];
+            crc_table[4][n] = REV(c);
+            for (k = 1; k < 4; k++) {
+                c = crc_table[0][c & 0xff] ^ (c >> 8);
+                crc_table[k][n] = c;
+                crc_table[k + 4][n] = REV(c);
+            }
+        }
+#endif /* BYFOUR */
+
+        crc_table_empty = 0;
+    }
+    else {      /* not first */
+        /* wait for the other guy to finish (not efficient, but rare) */
+        while (crc_table_empty)
+            ;
+    }
+
+#ifdef MAKECRCH
+    /* write out CRC tables to crc32.h */
+    {
+        FILE *out;
+
+        out = fopen("crc32.h", "w");
+        if (out == NULL) return;
+        fprintf(out, "/* crc32.h -- tables for rapid CRC calculation\n");
+        fprintf(out, " * Generated automatically by crc32.c\n */\n\n");
+        fprintf(out, "local const unsigned long FAR ");
+        fprintf(out, "crc_table[TBLS][256] =\n{\n  {\n");
+        write_table(out, crc_table[0]);
+#  ifdef BYFOUR
+        fprintf(out, "#ifdef BYFOUR\n");
+        for (k = 1; k < 8; k++) {
+            fprintf(out, "  },\n  {\n");
+            write_table(out, crc_table[k]);
+        }
+        fprintf(out, "#endif\n");
+#  endif /* BYFOUR */
+        fprintf(out, "  }\n};\n");
+        fclose(out);
+    }
+#endif /* MAKECRCH */
+}
+
+#ifdef MAKECRCH
+local void write_table(out, table)
+    FILE *out;
+    const unsigned long FAR *table;
+{
+    int n;
+
+    for (n = 0; n < 256; n++)
+        fprintf(out, "%s0x%08lxUL%s", n % 5 ? "" : "    ", table[n],
+                n == 255 ? "\n" : (n % 5 == 4 ? ",\n" : ", "));
+}
+#endif /* MAKECRCH */
+
+#else /* !DYNAMIC_CRC_TABLE */
+/* ========================================================================
+ * Tables of CRC-32s of all single-byte values, made by make_crc_table().
+ */
+#include "crc32.h"
+#endif /* DYNAMIC_CRC_TABLE */
+
+/* =========================================================================
+ * This function can be used by asm versions of crc32()
+ */
+const unsigned long FAR * ZEXPORT get_crc_table()
+{
+#ifdef DYNAMIC_CRC_TABLE
+    if (crc_table_empty)
+        make_crc_table();
+#endif /* DYNAMIC_CRC_TABLE */
+    return (const unsigned long FAR *)crc_table;
+}
+
+/* ========================================================================= */
+#define DO1 crc = crc_table[0][((int)crc ^ (*buf++)) & 0xff] ^ (crc >> 8)
+#define DO8 DO1; DO1; DO1; DO1; DO1; DO1; DO1; DO1
+
+/* ========================================================================= */
+unsigned long ZEXPORT crc32(crc, buf, len)
+    unsigned long crc;
+    const unsigned char FAR *buf;
+    unsigned len;
+{
+    if (buf == Z_NULL) return 0UL;
+
+#ifdef DYNAMIC_CRC_TABLE
+    if (crc_table_empty)
+        make_crc_table();
+#endif /* DYNAMIC_CRC_TABLE */
+
+#ifdef BYFOUR
+    if (sizeof(void *) == sizeof(ptrdiff_t)) {
+        u4 endian;
+
+        endian = 1;
+        if (*((unsigned char *)(&endian)))
+            return crc32_little(crc, buf, len);
+        else
+            return crc32_big(crc, buf, len);
+    }
+#endif /* BYFOUR */
+    crc = crc ^ 0xffffffffUL;
+    while (len >= 8) {
+        DO8;
+        len -= 8;
+    }
+    if (len) do {
+        DO1;
+    } while (--len);
+    return crc ^ 0xffffffffUL;
+}
+
+#ifdef BYFOUR
+
+/* ========================================================================= */
+#define DOLIT4 c ^= *buf4++; \
+        c = crc_table[3][c & 0xff] ^ crc_table[2][(c >> 8) & 0xff] ^ \
+            crc_table[1][(c >> 16) & 0xff] ^ crc_table[0][c >> 24]
+#define DOLIT32 DOLIT4; DOLIT4; DOLIT4; DOLIT4; DOLIT4; DOLIT4; DOLIT4; DOLIT4
+
+/* ========================================================================= */
+local unsigned long crc32_little(crc, buf, len)
+    unsigned long crc;
+    const unsigned char FAR *buf;
+    unsigned len;
+{
+    register u4 c;
+    register const u4 FAR *buf4;
+
+    c = (u4)crc;
+    c = ~c;
+    while (len && ((ptrdiff_t)buf & 3)) {
+        c = crc_table[0][(c ^ *buf++) & 0xff] ^ (c >> 8);
+        len--;
+    }
+
+    buf4 = (const u4 FAR *)(const void FAR *)buf;
+    while (len >= 32) {
+        DOLIT32;
+        len -= 32;
+    }
+    while (len >= 4) {
+        DOLIT4;
+        len -= 4;
+    }
+    buf = (const unsigned char FAR *)buf4;
+
+    if (len) do {
+        c = crc_table[0][(c ^ *buf++) & 0xff] ^ (c >> 8);
+    } while (--len);
+    c = ~c;
+    return (unsigned long)c;
+}
+
+/* ========================================================================= */
+#define DOBIG4 c ^= *++buf4; \
+        c = crc_table[4][c & 0xff] ^ crc_table[5][(c >> 8) & 0xff] ^ \
+            crc_table[6][(c >> 16) & 0xff] ^ crc_table[7][c >> 24]
+#define DOBIG32 DOBIG4; DOBIG4; DOBIG4; DOBIG4; DOBIG4; DOBIG4; DOBIG4; DOBIG4
+
+/* ========================================================================= */
+local unsigned long crc32_big(crc, buf, len)
+    unsigned long crc;
+    const unsigned char FAR *buf;
+    unsigned len;
+{
+    register u4 c;
+    register const u4 FAR *buf4;
+
+    c = REV((u4)crc);
+    c = ~c;
+    while (len && ((ptrdiff_t)buf & 3)) {
+        c = crc_table[4][(c >> 24) ^ *buf++] ^ (c << 8);
+        len--;
+    }
+
+    buf4 = (const u4 FAR *)(const void FAR *)buf;
+    buf4--;
+    while (len >= 32) {
+        DOBIG32;
+        len -= 32;
+    }
+    while (len >= 4) {
+        DOBIG4;
+        len -= 4;
+    }
+    buf4++;
+    buf = (const unsigned char FAR *)buf4;
+
+    if (len) do {
+        c = crc_table[4][(c >> 24) ^ *buf++] ^ (c << 8);
+    } while (--len);
+    c = ~c;
+    return (unsigned long)(REV(c));
+}
+
+#endif /* BYFOUR */
+
+#define GF2_DIM 32      /* dimension of GF(2) vectors (length of CRC) */
+
+/* ========================================================================= */
+local unsigned long gf2_matrix_times(mat, vec)
+    unsigned long *mat;
+    unsigned long vec;
+{
+    unsigned long sum;
+
+    sum = 0;
+    while (vec) {
+        if (vec & 1)
+            sum ^= *mat;
+        vec >>= 1;
+        mat++;
+    }
+    return sum;
+}
+
+/* ========================================================================= */
+local void gf2_matrix_square(square, mat)
+    unsigned long *square;
+    unsigned long *mat;
+{
+    int n;
+
+    for (n = 0; n < GF2_DIM; n++)
+        square[n] = gf2_matrix_times(mat, mat[n]);
+}
+
+/* ========================================================================= */
+uLong ZEXPORT crc32_combine(crc1, crc2, len2)
+    uLong crc1;
+    uLong crc2;
+    z_off_t len2;
+{
+    int n;
+    unsigned long row;
+    unsigned long even[GF2_DIM];    /* even-power-of-two zeros operator */
+    unsigned long odd[GF2_DIM];     /* odd-power-of-two zeros operator */
+
+    /* degenerate case */
+    if (len2 == 0)
+        return crc1;
+
+    /* put operator for one zero bit in odd */
+    odd[0] = 0xedb88320L;           /* CRC-32 polynomial */
+    row = 1;
+    for (n = 1; n < GF2_DIM; n++) {
+        odd[n] = row;
+        row <<= 1;
+    }
+
+    /* put operator for two zero bits in even */
+    gf2_matrix_square(even, odd);
+
+    /* put operator for four zero bits in odd */
+    gf2_matrix_square(odd, even);
+
+    /* apply len2 zeros to crc1 (first square will put the operator for one
+       zero byte, eight zero bits, in even) */
+    do {
+        /* apply zeros operator for this bit of len2 */
+        gf2_matrix_square(even, odd);
+        if (len2 & 1)
+            crc1 = gf2_matrix_times(even, crc1);
+        len2 >>= 1;
+
+        /* if no more bits set, then done */
+        if (len2 == 0)
+            break;
+
+        /* another iteration of the loop with odd and even swapped */
+        gf2_matrix_square(odd, even);
+        if (len2 & 1)
+            crc1 = gf2_matrix_times(odd, crc1);
+        len2 >>= 1;
+
+        /* if no more bits set, then done */
+    } while (len2 != 0);
+
+    /* return combined crc */
+    crc1 ^= crc2;
+    return crc1;
+}
diff --git a/distrib/zlib-1.2.3/crc32.h b/distrib/zlib-1.2.3/crc32.h
new file mode 100644
index 0000000..8053b61
--- /dev/null
+++ b/distrib/zlib-1.2.3/crc32.h
@@ -0,0 +1,441 @@
+/* crc32.h -- tables for rapid CRC calculation
+ * Generated automatically by crc32.c
+ */
+
+local const unsigned long FAR crc_table[TBLS][256] =
+{
+  {
+    0x00000000UL, 0x77073096UL, 0xee0e612cUL, 0x990951baUL, 0x076dc419UL,
+    0x706af48fUL, 0xe963a535UL, 0x9e6495a3UL, 0x0edb8832UL, 0x79dcb8a4UL,
+    0xe0d5e91eUL, 0x97d2d988UL, 0x09b64c2bUL, 0x7eb17cbdUL, 0xe7b82d07UL,
+    0x90bf1d91UL, 0x1db71064UL, 0x6ab020f2UL, 0xf3b97148UL, 0x84be41deUL,
+    0x1adad47dUL, 0x6ddde4ebUL, 0xf4d4b551UL, 0x83d385c7UL, 0x136c9856UL,
+    0x646ba8c0UL, 0xfd62f97aUL, 0x8a65c9ecUL, 0x14015c4fUL, 0x63066cd9UL,
+    0xfa0f3d63UL, 0x8d080df5UL, 0x3b6e20c8UL, 0x4c69105eUL, 0xd56041e4UL,
+    0xa2677172UL, 0x3c03e4d1UL, 0x4b04d447UL, 0xd20d85fdUL, 0xa50ab56bUL,
+    0x35b5a8faUL, 0x42b2986cUL, 0xdbbbc9d6UL, 0xacbcf940UL, 0x32d86ce3UL,
+    0x45df5c75UL, 0xdcd60dcfUL, 0xabd13d59UL, 0x26d930acUL, 0x51de003aUL,
+    0xc8d75180UL, 0xbfd06116UL, 0x21b4f4b5UL, 0x56b3c423UL, 0xcfba9599UL,
+    0xb8bda50fUL, 0x2802b89eUL, 0x5f058808UL, 0xc60cd9b2UL, 0xb10be924UL,
+    0x2f6f7c87UL, 0x58684c11UL, 0xc1611dabUL, 0xb6662d3dUL, 0x76dc4190UL,
+    0x01db7106UL, 0x98d220bcUL, 0xefd5102aUL, 0x71b18589UL, 0x06b6b51fUL,
+    0x9fbfe4a5UL, 0xe8b8d433UL, 0x7807c9a2UL, 0x0f00f934UL, 0x9609a88eUL,
+    0xe10e9818UL, 0x7f6a0dbbUL, 0x086d3d2dUL, 0x91646c97UL, 0xe6635c01UL,
+    0x6b6b51f4UL, 0x1c6c6162UL, 0x856530d8UL, 0xf262004eUL, 0x6c0695edUL,
+    0x1b01a57bUL, 0x8208f4c1UL, 0xf50fc457UL, 0x65b0d9c6UL, 0x12b7e950UL,
+    0x8bbeb8eaUL, 0xfcb9887cUL, 0x62dd1ddfUL, 0x15da2d49UL, 0x8cd37cf3UL,
+    0xfbd44c65UL, 0x4db26158UL, 0x3ab551ceUL, 0xa3bc0074UL, 0xd4bb30e2UL,
+    0x4adfa541UL, 0x3dd895d7UL, 0xa4d1c46dUL, 0xd3d6f4fbUL, 0x4369e96aUL,
+    0x346ed9fcUL, 0xad678846UL, 0xda60b8d0UL, 0x44042d73UL, 0x33031de5UL,
+    0xaa0a4c5fUL, 0xdd0d7cc9UL, 0x5005713cUL, 0x270241aaUL, 0xbe0b1010UL,
+    0xc90c2086UL, 0x5768b525UL, 0x206f85b3UL, 0xb966d409UL, 0xce61e49fUL,
+    0x5edef90eUL, 0x29d9c998UL, 0xb0d09822UL, 0xc7d7a8b4UL, 0x59b33d17UL,
+    0x2eb40d81UL, 0xb7bd5c3bUL, 0xc0ba6cadUL, 0xedb88320UL, 0x9abfb3b6UL,
+    0x03b6e20cUL, 0x74b1d29aUL, 0xead54739UL, 0x9dd277afUL, 0x04db2615UL,
+    0x73dc1683UL, 0xe3630b12UL, 0x94643b84UL, 0x0d6d6a3eUL, 0x7a6a5aa8UL,
+    0xe40ecf0bUL, 0x9309ff9dUL, 0x0a00ae27UL, 0x7d079eb1UL, 0xf00f9344UL,
+    0x8708a3d2UL, 0x1e01f268UL, 0x6906c2feUL, 0xf762575dUL, 0x806567cbUL,
+    0x196c3671UL, 0x6e6b06e7UL, 0xfed41b76UL, 0x89d32be0UL, 0x10da7a5aUL,
+    0x67dd4accUL, 0xf9b9df6fUL, 0x8ebeeff9UL, 0x17b7be43UL, 0x60b08ed5UL,
+    0xd6d6a3e8UL, 0xa1d1937eUL, 0x38d8c2c4UL, 0x4fdff252UL, 0xd1bb67f1UL,
+    0xa6bc5767UL, 0x3fb506ddUL, 0x48b2364bUL, 0xd80d2bdaUL, 0xaf0a1b4cUL,
+    0x36034af6UL, 0x41047a60UL, 0xdf60efc3UL, 0xa867df55UL, 0x316e8eefUL,
+    0x4669be79UL, 0xcb61b38cUL, 0xbc66831aUL, 0x256fd2a0UL, 0x5268e236UL,
+    0xcc0c7795UL, 0xbb0b4703UL, 0x220216b9UL, 0x5505262fUL, 0xc5ba3bbeUL,
+    0xb2bd0b28UL, 0x2bb45a92UL, 0x5cb36a04UL, 0xc2d7ffa7UL, 0xb5d0cf31UL,
+    0x2cd99e8bUL, 0x5bdeae1dUL, 0x9b64c2b0UL, 0xec63f226UL, 0x756aa39cUL,
+    0x026d930aUL, 0x9c0906a9UL, 0xeb0e363fUL, 0x72076785UL, 0x05005713UL,
+    0x95bf4a82UL, 0xe2b87a14UL, 0x7bb12baeUL, 0x0cb61b38UL, 0x92d28e9bUL,
+    0xe5d5be0dUL, 0x7cdcefb7UL, 0x0bdbdf21UL, 0x86d3d2d4UL, 0xf1d4e242UL,
+    0x68ddb3f8UL, 0x1fda836eUL, 0x81be16cdUL, 0xf6b9265bUL, 0x6fb077e1UL,
+    0x18b74777UL, 0x88085ae6UL, 0xff0f6a70UL, 0x66063bcaUL, 0x11010b5cUL,
+    0x8f659effUL, 0xf862ae69UL, 0x616bffd3UL, 0x166ccf45UL, 0xa00ae278UL,
+    0xd70dd2eeUL, 0x4e048354UL, 0x3903b3c2UL, 0xa7672661UL, 0xd06016f7UL,
+    0x4969474dUL, 0x3e6e77dbUL, 0xaed16a4aUL, 0xd9d65adcUL, 0x40df0b66UL,
+    0x37d83bf0UL, 0xa9bcae53UL, 0xdebb9ec5UL, 0x47b2cf7fUL, 0x30b5ffe9UL,
+    0xbdbdf21cUL, 0xcabac28aUL, 0x53b39330UL, 0x24b4a3a6UL, 0xbad03605UL,
+    0xcdd70693UL, 0x54de5729UL, 0x23d967bfUL, 0xb3667a2eUL, 0xc4614ab8UL,
+    0x5d681b02UL, 0x2a6f2b94UL, 0xb40bbe37UL, 0xc30c8ea1UL, 0x5a05df1bUL,
+    0x2d02ef8dUL
+#ifdef BYFOUR
+  },
+  {
+    0x00000000UL, 0x191b3141UL, 0x32366282UL, 0x2b2d53c3UL, 0x646cc504UL,
+    0x7d77f445UL, 0x565aa786UL, 0x4f4196c7UL, 0xc8d98a08UL, 0xd1c2bb49UL,
+    0xfaefe88aUL, 0xe3f4d9cbUL, 0xacb54f0cUL, 0xb5ae7e4dUL, 0x9e832d8eUL,
+    0x87981ccfUL, 0x4ac21251UL, 0x53d92310UL, 0x78f470d3UL, 0x61ef4192UL,
+    0x2eaed755UL, 0x37b5e614UL, 0x1c98b5d7UL, 0x05838496UL, 0x821b9859UL,
+    0x9b00a918UL, 0xb02dfadbUL, 0xa936cb9aUL, 0xe6775d5dUL, 0xff6c6c1cUL,
+    0xd4413fdfUL, 0xcd5a0e9eUL, 0x958424a2UL, 0x8c9f15e3UL, 0xa7b24620UL,
+    0xbea97761UL, 0xf1e8e1a6UL, 0xe8f3d0e7UL, 0xc3de8324UL, 0xdac5b265UL,
+    0x5d5daeaaUL, 0x44469febUL, 0x6f6bcc28UL, 0x7670fd69UL, 0x39316baeUL,
+    0x202a5aefUL, 0x0b07092cUL, 0x121c386dUL, 0xdf4636f3UL, 0xc65d07b2UL,
+    0xed705471UL, 0xf46b6530UL, 0xbb2af3f7UL, 0xa231c2b6UL, 0x891c9175UL,
+    0x9007a034UL, 0x179fbcfbUL, 0x0e848dbaUL, 0x25a9de79UL, 0x3cb2ef38UL,
+    0x73f379ffUL, 0x6ae848beUL, 0x41c51b7dUL, 0x58de2a3cUL, 0xf0794f05UL,
+    0xe9627e44UL, 0xc24f2d87UL, 0xdb541cc6UL, 0x94158a01UL, 0x8d0ebb40UL,
+    0xa623e883UL, 0xbf38d9c2UL, 0x38a0c50dUL, 0x21bbf44cUL, 0x0a96a78fUL,
+    0x138d96ceUL, 0x5ccc0009UL, 0x45d73148UL, 0x6efa628bUL, 0x77e153caUL,
+    0xbabb5d54UL, 0xa3a06c15UL, 0x888d3fd6UL, 0x91960e97UL, 0xded79850UL,
+    0xc7cca911UL, 0xece1fad2UL, 0xf5facb93UL, 0x7262d75cUL, 0x6b79e61dUL,
+    0x4054b5deUL, 0x594f849fUL, 0x160e1258UL, 0x0f152319UL, 0x243870daUL,
+    0x3d23419bUL, 0x65fd6ba7UL, 0x7ce65ae6UL, 0x57cb0925UL, 0x4ed03864UL,
+    0x0191aea3UL, 0x188a9fe2UL, 0x33a7cc21UL, 0x2abcfd60UL, 0xad24e1afUL,
+    0xb43fd0eeUL, 0x9f12832dUL, 0x8609b26cUL, 0xc94824abUL, 0xd05315eaUL,
+    0xfb7e4629UL, 0xe2657768UL, 0x2f3f79f6UL, 0x362448b7UL, 0x1d091b74UL,
+    0x04122a35UL, 0x4b53bcf2UL, 0x52488db3UL, 0x7965de70UL, 0x607eef31UL,
+    0xe7e6f3feUL, 0xfefdc2bfUL, 0xd5d0917cUL, 0xcccba03dUL, 0x838a36faUL,
+    0x9a9107bbUL, 0xb1bc5478UL, 0xa8a76539UL, 0x3b83984bUL, 0x2298a90aUL,
+    0x09b5fac9UL, 0x10aecb88UL, 0x5fef5d4fUL, 0x46f46c0eUL, 0x6dd93fcdUL,
+    0x74c20e8cUL, 0xf35a1243UL, 0xea412302UL, 0xc16c70c1UL, 0xd8774180UL,
+    0x9736d747UL, 0x8e2de606UL, 0xa500b5c5UL, 0xbc1b8484UL, 0x71418a1aUL,
+    0x685abb5bUL, 0x4377e898UL, 0x5a6cd9d9UL, 0x152d4f1eUL, 0x0c367e5fUL,
+    0x271b2d9cUL, 0x3e001cddUL, 0xb9980012UL, 0xa0833153UL, 0x8bae6290UL,
+    0x92b553d1UL, 0xddf4c516UL, 0xc4eff457UL, 0xefc2a794UL, 0xf6d996d5UL,
+    0xae07bce9UL, 0xb71c8da8UL, 0x9c31de6bUL, 0x852aef2aUL, 0xca6b79edUL,
+    0xd37048acUL, 0xf85d1b6fUL, 0xe1462a2eUL, 0x66de36e1UL, 0x7fc507a0UL,
+    0x54e85463UL, 0x4df36522UL, 0x02b2f3e5UL, 0x1ba9c2a4UL, 0x30849167UL,
+    0x299fa026UL, 0xe4c5aeb8UL, 0xfdde9ff9UL, 0xd6f3cc3aUL, 0xcfe8fd7bUL,
+    0x80a96bbcUL, 0x99b25afdUL, 0xb29f093eUL, 0xab84387fUL, 0x2c1c24b0UL,
+    0x350715f1UL, 0x1e2a4632UL, 0x07317773UL, 0x4870e1b4UL, 0x516bd0f5UL,
+    0x7a468336UL, 0x635db277UL, 0xcbfad74eUL, 0xd2e1e60fUL, 0xf9ccb5ccUL,
+    0xe0d7848dUL, 0xaf96124aUL, 0xb68d230bUL, 0x9da070c8UL, 0x84bb4189UL,
+    0x03235d46UL, 0x1a386c07UL, 0x31153fc4UL, 0x280e0e85UL, 0x674f9842UL,
+    0x7e54a903UL, 0x5579fac0UL, 0x4c62cb81UL, 0x8138c51fUL, 0x9823f45eUL,
+    0xb30ea79dUL, 0xaa1596dcUL, 0xe554001bUL, 0xfc4f315aUL, 0xd7626299UL,
+    0xce7953d8UL, 0x49e14f17UL, 0x50fa7e56UL, 0x7bd72d95UL, 0x62cc1cd4UL,
+    0x2d8d8a13UL, 0x3496bb52UL, 0x1fbbe891UL, 0x06a0d9d0UL, 0x5e7ef3ecUL,
+    0x4765c2adUL, 0x6c48916eUL, 0x7553a02fUL, 0x3a1236e8UL, 0x230907a9UL,
+    0x0824546aUL, 0x113f652bUL, 0x96a779e4UL, 0x8fbc48a5UL, 0xa4911b66UL,
+    0xbd8a2a27UL, 0xf2cbbce0UL, 0xebd08da1UL, 0xc0fdde62UL, 0xd9e6ef23UL,
+    0x14bce1bdUL, 0x0da7d0fcUL, 0x268a833fUL, 0x3f91b27eUL, 0x70d024b9UL,
+    0x69cb15f8UL, 0x42e6463bUL, 0x5bfd777aUL, 0xdc656bb5UL, 0xc57e5af4UL,
+    0xee530937UL, 0xf7483876UL, 0xb809aeb1UL, 0xa1129ff0UL, 0x8a3fcc33UL,
+    0x9324fd72UL
+  },
+  {
+    0x00000000UL, 0x01c26a37UL, 0x0384d46eUL, 0x0246be59UL, 0x0709a8dcUL,
+    0x06cbc2ebUL, 0x048d7cb2UL, 0x054f1685UL, 0x0e1351b8UL, 0x0fd13b8fUL,
+    0x0d9785d6UL, 0x0c55efe1UL, 0x091af964UL, 0x08d89353UL, 0x0a9e2d0aUL,
+    0x0b5c473dUL, 0x1c26a370UL, 0x1de4c947UL, 0x1fa2771eUL, 0x1e601d29UL,
+    0x1b2f0bacUL, 0x1aed619bUL, 0x18abdfc2UL, 0x1969b5f5UL, 0x1235f2c8UL,
+    0x13f798ffUL, 0x11b126a6UL, 0x10734c91UL, 0x153c5a14UL, 0x14fe3023UL,
+    0x16b88e7aUL, 0x177ae44dUL, 0x384d46e0UL, 0x398f2cd7UL, 0x3bc9928eUL,
+    0x3a0bf8b9UL, 0x3f44ee3cUL, 0x3e86840bUL, 0x3cc03a52UL, 0x3d025065UL,
+    0x365e1758UL, 0x379c7d6fUL, 0x35dac336UL, 0x3418a901UL, 0x3157bf84UL,
+    0x3095d5b3UL, 0x32d36beaUL, 0x331101ddUL, 0x246be590UL, 0x25a98fa7UL,
+    0x27ef31feUL, 0x262d5bc9UL, 0x23624d4cUL, 0x22a0277bUL, 0x20e69922UL,
+    0x2124f315UL, 0x2a78b428UL, 0x2bbade1fUL, 0x29fc6046UL, 0x283e0a71UL,
+    0x2d711cf4UL, 0x2cb376c3UL, 0x2ef5c89aUL, 0x2f37a2adUL, 0x709a8dc0UL,
+    0x7158e7f7UL, 0x731e59aeUL, 0x72dc3399UL, 0x7793251cUL, 0x76514f2bUL,
+    0x7417f172UL, 0x75d59b45UL, 0x7e89dc78UL, 0x7f4bb64fUL, 0x7d0d0816UL,
+    0x7ccf6221UL, 0x798074a4UL, 0x78421e93UL, 0x7a04a0caUL, 0x7bc6cafdUL,
+    0x6cbc2eb0UL, 0x6d7e4487UL, 0x6f38fadeUL, 0x6efa90e9UL, 0x6bb5866cUL,
+    0x6a77ec5bUL, 0x68315202UL, 0x69f33835UL, 0x62af7f08UL, 0x636d153fUL,
+    0x612bab66UL, 0x60e9c151UL, 0x65a6d7d4UL, 0x6464bde3UL, 0x662203baUL,
+    0x67e0698dUL, 0x48d7cb20UL, 0x4915a117UL, 0x4b531f4eUL, 0x4a917579UL,
+    0x4fde63fcUL, 0x4e1c09cbUL, 0x4c5ab792UL, 0x4d98dda5UL, 0x46c49a98UL,
+    0x4706f0afUL, 0x45404ef6UL, 0x448224c1UL, 0x41cd3244UL, 0x400f5873UL,
+    0x4249e62aUL, 0x438b8c1dUL, 0x54f16850UL, 0x55330267UL, 0x5775bc3eUL,
+    0x56b7d609UL, 0x53f8c08cUL, 0x523aaabbUL, 0x507c14e2UL, 0x51be7ed5UL,
+    0x5ae239e8UL, 0x5b2053dfUL, 0x5966ed86UL, 0x58a487b1UL, 0x5deb9134UL,
+    0x5c29fb03UL, 0x5e6f455aUL, 0x5fad2f6dUL, 0xe1351b80UL, 0xe0f771b7UL,
+    0xe2b1cfeeUL, 0xe373a5d9UL, 0xe63cb35cUL, 0xe7fed96bUL, 0xe5b86732UL,
+    0xe47a0d05UL, 0xef264a38UL, 0xeee4200fUL, 0xeca29e56UL, 0xed60f461UL,
+    0xe82fe2e4UL, 0xe9ed88d3UL, 0xebab368aUL, 0xea695cbdUL, 0xfd13b8f0UL,
+    0xfcd1d2c7UL, 0xfe976c9eUL, 0xff5506a9UL, 0xfa1a102cUL, 0xfbd87a1bUL,
+    0xf99ec442UL, 0xf85cae75UL, 0xf300e948UL, 0xf2c2837fUL, 0xf0843d26UL,
+    0xf1465711UL, 0xf4094194UL, 0xf5cb2ba3UL, 0xf78d95faUL, 0xf64fffcdUL,
+    0xd9785d60UL, 0xd8ba3757UL, 0xdafc890eUL, 0xdb3ee339UL, 0xde71f5bcUL,
+    0xdfb39f8bUL, 0xddf521d2UL, 0xdc374be5UL, 0xd76b0cd8UL, 0xd6a966efUL,
+    0xd4efd8b6UL, 0xd52db281UL, 0xd062a404UL, 0xd1a0ce33UL, 0xd3e6706aUL,
+    0xd2241a5dUL, 0xc55efe10UL, 0xc49c9427UL, 0xc6da2a7eUL, 0xc7184049UL,
+    0xc25756ccUL, 0xc3953cfbUL, 0xc1d382a2UL, 0xc011e895UL, 0xcb4dafa8UL,
+    0xca8fc59fUL, 0xc8c97bc6UL, 0xc90b11f1UL, 0xcc440774UL, 0xcd866d43UL,
+    0xcfc0d31aUL, 0xce02b92dUL, 0x91af9640UL, 0x906dfc77UL, 0x922b422eUL,
+    0x93e92819UL, 0x96a63e9cUL, 0x976454abUL, 0x9522eaf2UL, 0x94e080c5UL,
+    0x9fbcc7f8UL, 0x9e7eadcfUL, 0x9c381396UL, 0x9dfa79a1UL, 0x98b56f24UL,
+    0x99770513UL, 0x9b31bb4aUL, 0x9af3d17dUL, 0x8d893530UL, 0x8c4b5f07UL,
+    0x8e0de15eUL, 0x8fcf8b69UL, 0x8a809decUL, 0x8b42f7dbUL, 0x89044982UL,
+    0x88c623b5UL, 0x839a6488UL, 0x82580ebfUL, 0x801eb0e6UL, 0x81dcdad1UL,
+    0x8493cc54UL, 0x8551a663UL, 0x8717183aUL, 0x86d5720dUL, 0xa9e2d0a0UL,
+    0xa820ba97UL, 0xaa6604ceUL, 0xaba46ef9UL, 0xaeeb787cUL, 0xaf29124bUL,
+    0xad6fac12UL, 0xacadc625UL, 0xa7f18118UL, 0xa633eb2fUL, 0xa4755576UL,
+    0xa5b73f41UL, 0xa0f829c4UL, 0xa13a43f3UL, 0xa37cfdaaUL, 0xa2be979dUL,
+    0xb5c473d0UL, 0xb40619e7UL, 0xb640a7beUL, 0xb782cd89UL, 0xb2cddb0cUL,
+    0xb30fb13bUL, 0xb1490f62UL, 0xb08b6555UL, 0xbbd72268UL, 0xba15485fUL,
+    0xb853f606UL, 0xb9919c31UL, 0xbcde8ab4UL, 0xbd1ce083UL, 0xbf5a5edaUL,
+    0xbe9834edUL
+  },
+  {
+    0x00000000UL, 0xb8bc6765UL, 0xaa09c88bUL, 0x12b5afeeUL, 0x8f629757UL,
+    0x37def032UL, 0x256b5fdcUL, 0x9dd738b9UL, 0xc5b428efUL, 0x7d084f8aUL,
+    0x6fbde064UL, 0xd7018701UL, 0x4ad6bfb8UL, 0xf26ad8ddUL, 0xe0df7733UL,
+    0x58631056UL, 0x5019579fUL, 0xe8a530faUL, 0xfa109f14UL, 0x42acf871UL,
+    0xdf7bc0c8UL, 0x67c7a7adUL, 0x75720843UL, 0xcdce6f26UL, 0x95ad7f70UL,
+    0x2d111815UL, 0x3fa4b7fbUL, 0x8718d09eUL, 0x1acfe827UL, 0xa2738f42UL,
+    0xb0c620acUL, 0x087a47c9UL, 0xa032af3eUL, 0x188ec85bUL, 0x0a3b67b5UL,
+    0xb28700d0UL, 0x2f503869UL, 0x97ec5f0cUL, 0x8559f0e2UL, 0x3de59787UL,
+    0x658687d1UL, 0xdd3ae0b4UL, 0xcf8f4f5aUL, 0x7733283fUL, 0xeae41086UL,
+    0x525877e3UL, 0x40edd80dUL, 0xf851bf68UL, 0xf02bf8a1UL, 0x48979fc4UL,
+    0x5a22302aUL, 0xe29e574fUL, 0x7f496ff6UL, 0xc7f50893UL, 0xd540a77dUL,
+    0x6dfcc018UL, 0x359fd04eUL, 0x8d23b72bUL, 0x9f9618c5UL, 0x272a7fa0UL,
+    0xbafd4719UL, 0x0241207cUL, 0x10f48f92UL, 0xa848e8f7UL, 0x9b14583dUL,
+    0x23a83f58UL, 0x311d90b6UL, 0x89a1f7d3UL, 0x1476cf6aUL, 0xaccaa80fUL,
+    0xbe7f07e1UL, 0x06c36084UL, 0x5ea070d2UL, 0xe61c17b7UL, 0xf4a9b859UL,
+    0x4c15df3cUL, 0xd1c2e785UL, 0x697e80e0UL, 0x7bcb2f0eUL, 0xc377486bUL,
+    0xcb0d0fa2UL, 0x73b168c7UL, 0x6104c729UL, 0xd9b8a04cUL, 0x446f98f5UL,
+    0xfcd3ff90UL, 0xee66507eUL, 0x56da371bUL, 0x0eb9274dUL, 0xb6054028UL,
+    0xa4b0efc6UL, 0x1c0c88a3UL, 0x81dbb01aUL, 0x3967d77fUL, 0x2bd27891UL,
+    0x936e1ff4UL, 0x3b26f703UL, 0x839a9066UL, 0x912f3f88UL, 0x299358edUL,
+    0xb4446054UL, 0x0cf80731UL, 0x1e4da8dfUL, 0xa6f1cfbaUL, 0xfe92dfecUL,
+    0x462eb889UL, 0x549b1767UL, 0xec277002UL, 0x71f048bbUL, 0xc94c2fdeUL,
+    0xdbf98030UL, 0x6345e755UL, 0x6b3fa09cUL, 0xd383c7f9UL, 0xc1366817UL,
+    0x798a0f72UL, 0xe45d37cbUL, 0x5ce150aeUL, 0x4e54ff40UL, 0xf6e89825UL,
+    0xae8b8873UL, 0x1637ef16UL, 0x048240f8UL, 0xbc3e279dUL, 0x21e91f24UL,
+    0x99557841UL, 0x8be0d7afUL, 0x335cb0caUL, 0xed59b63bUL, 0x55e5d15eUL,
+    0x47507eb0UL, 0xffec19d5UL, 0x623b216cUL, 0xda874609UL, 0xc832e9e7UL,
+    0x708e8e82UL, 0x28ed9ed4UL, 0x9051f9b1UL, 0x82e4565fUL, 0x3a58313aUL,
+    0xa78f0983UL, 0x1f336ee6UL, 0x0d86c108UL, 0xb53aa66dUL, 0xbd40e1a4UL,
+    0x05fc86c1UL, 0x1749292fUL, 0xaff54e4aUL, 0x322276f3UL, 0x8a9e1196UL,
+    0x982bbe78UL, 0x2097d91dUL, 0x78f4c94bUL, 0xc048ae2eUL, 0xd2fd01c0UL,
+    0x6a4166a5UL, 0xf7965e1cUL, 0x4f2a3979UL, 0x5d9f9697UL, 0xe523f1f2UL,
+    0x4d6b1905UL, 0xf5d77e60UL, 0xe762d18eUL, 0x5fdeb6ebUL, 0xc2098e52UL,
+    0x7ab5e937UL, 0x680046d9UL, 0xd0bc21bcUL, 0x88df31eaUL, 0x3063568fUL,
+    0x22d6f961UL, 0x9a6a9e04UL, 0x07bda6bdUL, 0xbf01c1d8UL, 0xadb46e36UL,
+    0x15080953UL, 0x1d724e9aUL, 0xa5ce29ffUL, 0xb77b8611UL, 0x0fc7e174UL,
+    0x9210d9cdUL, 0x2aacbea8UL, 0x38191146UL, 0x80a57623UL, 0xd8c66675UL,
+    0x607a0110UL, 0x72cfaefeUL, 0xca73c99bUL, 0x57a4f122UL, 0xef189647UL,
+    0xfdad39a9UL, 0x45115eccUL, 0x764dee06UL, 0xcef18963UL, 0xdc44268dUL,
+    0x64f841e8UL, 0xf92f7951UL, 0x41931e34UL, 0x5326b1daUL, 0xeb9ad6bfUL,
+    0xb3f9c6e9UL, 0x0b45a18cUL, 0x19f00e62UL, 0xa14c6907UL, 0x3c9b51beUL,
+    0x842736dbUL, 0x96929935UL, 0x2e2efe50UL, 0x2654b999UL, 0x9ee8defcUL,
+    0x8c5d7112UL, 0x34e11677UL, 0xa9362eceUL, 0x118a49abUL, 0x033fe645UL,
+    0xbb838120UL, 0xe3e09176UL, 0x5b5cf613UL, 0x49e959fdUL, 0xf1553e98UL,
+    0x6c820621UL, 0xd43e6144UL, 0xc68bceaaUL, 0x7e37a9cfUL, 0xd67f4138UL,
+    0x6ec3265dUL, 0x7c7689b3UL, 0xc4caeed6UL, 0x591dd66fUL, 0xe1a1b10aUL,
+    0xf3141ee4UL, 0x4ba87981UL, 0x13cb69d7UL, 0xab770eb2UL, 0xb9c2a15cUL,
+    0x017ec639UL, 0x9ca9fe80UL, 0x241599e5UL, 0x36a0360bUL, 0x8e1c516eUL,
+    0x866616a7UL, 0x3eda71c2UL, 0x2c6fde2cUL, 0x94d3b949UL, 0x090481f0UL,
+    0xb1b8e695UL, 0xa30d497bUL, 0x1bb12e1eUL, 0x43d23e48UL, 0xfb6e592dUL,
+    0xe9dbf6c3UL, 0x516791a6UL, 0xccb0a91fUL, 0x740cce7aUL, 0x66b96194UL,
+    0xde0506f1UL
+  },
+  {
+    0x00000000UL, 0x96300777UL, 0x2c610eeeUL, 0xba510999UL, 0x19c46d07UL,
+    0x8ff46a70UL, 0x35a563e9UL, 0xa395649eUL, 0x3288db0eUL, 0xa4b8dc79UL,
+    0x1ee9d5e0UL, 0x88d9d297UL, 0x2b4cb609UL, 0xbd7cb17eUL, 0x072db8e7UL,
+    0x911dbf90UL, 0x6410b71dUL, 0xf220b06aUL, 0x4871b9f3UL, 0xde41be84UL,
+    0x7dd4da1aUL, 0xebe4dd6dUL, 0x51b5d4f4UL, 0xc785d383UL, 0x56986c13UL,
+    0xc0a86b64UL, 0x7af962fdUL, 0xecc9658aUL, 0x4f5c0114UL, 0xd96c0663UL,
+    0x633d0ffaUL, 0xf50d088dUL, 0xc8206e3bUL, 0x5e10694cUL, 0xe44160d5UL,
+    0x727167a2UL, 0xd1e4033cUL, 0x47d4044bUL, 0xfd850dd2UL, 0x6bb50aa5UL,
+    0xfaa8b535UL, 0x6c98b242UL, 0xd6c9bbdbUL, 0x40f9bcacUL, 0xe36cd832UL,
+    0x755cdf45UL, 0xcf0dd6dcUL, 0x593dd1abUL, 0xac30d926UL, 0x3a00de51UL,
+    0x8051d7c8UL, 0x1661d0bfUL, 0xb5f4b421UL, 0x23c4b356UL, 0x9995bacfUL,
+    0x0fa5bdb8UL, 0x9eb80228UL, 0x0888055fUL, 0xb2d90cc6UL, 0x24e90bb1UL,
+    0x877c6f2fUL, 0x114c6858UL, 0xab1d61c1UL, 0x3d2d66b6UL, 0x9041dc76UL,
+    0x0671db01UL, 0xbc20d298UL, 0x2a10d5efUL, 0x8985b171UL, 0x1fb5b606UL,
+    0xa5e4bf9fUL, 0x33d4b8e8UL, 0xa2c90778UL, 0x34f9000fUL, 0x8ea80996UL,
+    0x18980ee1UL, 0xbb0d6a7fUL, 0x2d3d6d08UL, 0x976c6491UL, 0x015c63e6UL,
+    0xf4516b6bUL, 0x62616c1cUL, 0xd8306585UL, 0x4e0062f2UL, 0xed95066cUL,
+    0x7ba5011bUL, 0xc1f40882UL, 0x57c40ff5UL, 0xc6d9b065UL, 0x50e9b712UL,
+    0xeab8be8bUL, 0x7c88b9fcUL, 0xdf1ddd62UL, 0x492dda15UL, 0xf37cd38cUL,
+    0x654cd4fbUL, 0x5861b24dUL, 0xce51b53aUL, 0x7400bca3UL, 0xe230bbd4UL,
+    0x41a5df4aUL, 0xd795d83dUL, 0x6dc4d1a4UL, 0xfbf4d6d3UL, 0x6ae96943UL,
+    0xfcd96e34UL, 0x468867adUL, 0xd0b860daUL, 0x732d0444UL, 0xe51d0333UL,
+    0x5f4c0aaaUL, 0xc97c0dddUL, 0x3c710550UL, 0xaa410227UL, 0x10100bbeUL,
+    0x86200cc9UL, 0x25b56857UL, 0xb3856f20UL, 0x09d466b9UL, 0x9fe461ceUL,
+    0x0ef9de5eUL, 0x98c9d929UL, 0x2298d0b0UL, 0xb4a8d7c7UL, 0x173db359UL,
+    0x810db42eUL, 0x3b5cbdb7UL, 0xad6cbac0UL, 0x2083b8edUL, 0xb6b3bf9aUL,
+    0x0ce2b603UL, 0x9ad2b174UL, 0x3947d5eaUL, 0xaf77d29dUL, 0x1526db04UL,
+    0x8316dc73UL, 0x120b63e3UL, 0x843b6494UL, 0x3e6a6d0dUL, 0xa85a6a7aUL,
+    0x0bcf0ee4UL, 0x9dff0993UL, 0x27ae000aUL, 0xb19e077dUL, 0x44930ff0UL,
+    0xd2a30887UL, 0x68f2011eUL, 0xfec20669UL, 0x5d5762f7UL, 0xcb676580UL,
+    0x71366c19UL, 0xe7066b6eUL, 0x761bd4feUL, 0xe02bd389UL, 0x5a7ada10UL,
+    0xcc4add67UL, 0x6fdfb9f9UL, 0xf9efbe8eUL, 0x43beb717UL, 0xd58eb060UL,
+    0xe8a3d6d6UL, 0x7e93d1a1UL, 0xc4c2d838UL, 0x52f2df4fUL, 0xf167bbd1UL,
+    0x6757bca6UL, 0xdd06b53fUL, 0x4b36b248UL, 0xda2b0dd8UL, 0x4c1b0aafUL,
+    0xf64a0336UL, 0x607a0441UL, 0xc3ef60dfUL, 0x55df67a8UL, 0xef8e6e31UL,
+    0x79be6946UL, 0x8cb361cbUL, 0x1a8366bcUL, 0xa0d26f25UL, 0x36e26852UL,
+    0x95770cccUL, 0x03470bbbUL, 0xb9160222UL, 0x2f260555UL, 0xbe3bbac5UL,
+    0x280bbdb2UL, 0x925ab42bUL, 0x046ab35cUL, 0xa7ffd7c2UL, 0x31cfd0b5UL,
+    0x8b9ed92cUL, 0x1daede5bUL, 0xb0c2649bUL, 0x26f263ecUL, 0x9ca36a75UL,
+    0x0a936d02UL, 0xa906099cUL, 0x3f360eebUL, 0x85670772UL, 0x13570005UL,
+    0x824abf95UL, 0x147ab8e2UL, 0xae2bb17bUL, 0x381bb60cUL, 0x9b8ed292UL,
+    0x0dbed5e5UL, 0xb7efdc7cUL, 0x21dfdb0bUL, 0xd4d2d386UL, 0x42e2d4f1UL,
+    0xf8b3dd68UL, 0x6e83da1fUL, 0xcd16be81UL, 0x5b26b9f6UL, 0xe177b06fUL,
+    0x7747b718UL, 0xe65a0888UL, 0x706a0fffUL, 0xca3b0666UL, 0x5c0b0111UL,
+    0xff9e658fUL, 0x69ae62f8UL, 0xd3ff6b61UL, 0x45cf6c16UL, 0x78e20aa0UL,
+    0xeed20dd7UL, 0x5483044eUL, 0xc2b30339UL, 0x612667a7UL, 0xf71660d0UL,
+    0x4d476949UL, 0xdb776e3eUL, 0x4a6ad1aeUL, 0xdc5ad6d9UL, 0x660bdf40UL,
+    0xf03bd837UL, 0x53aebca9UL, 0xc59ebbdeUL, 0x7fcfb247UL, 0xe9ffb530UL,
+    0x1cf2bdbdUL, 0x8ac2bacaUL, 0x3093b353UL, 0xa6a3b424UL, 0x0536d0baUL,
+    0x9306d7cdUL, 0x2957de54UL, 0xbf67d923UL, 0x2e7a66b3UL, 0xb84a61c4UL,
+    0x021b685dUL, 0x942b6f2aUL, 0x37be0bb4UL, 0xa18e0cc3UL, 0x1bdf055aUL,
+    0x8def022dUL
+  },
+  {
+    0x00000000UL, 0x41311b19UL, 0x82623632UL, 0xc3532d2bUL, 0x04c56c64UL,
+    0x45f4777dUL, 0x86a75a56UL, 0xc796414fUL, 0x088ad9c8UL, 0x49bbc2d1UL,
+    0x8ae8effaUL, 0xcbd9f4e3UL, 0x0c4fb5acUL, 0x4d7eaeb5UL, 0x8e2d839eUL,
+    0xcf1c9887UL, 0x5112c24aUL, 0x1023d953UL, 0xd370f478UL, 0x9241ef61UL,
+    0x55d7ae2eUL, 0x14e6b537UL, 0xd7b5981cUL, 0x96848305UL, 0x59981b82UL,
+    0x18a9009bUL, 0xdbfa2db0UL, 0x9acb36a9UL, 0x5d5d77e6UL, 0x1c6c6cffUL,
+    0xdf3f41d4UL, 0x9e0e5acdUL, 0xa2248495UL, 0xe3159f8cUL, 0x2046b2a7UL,
+    0x6177a9beUL, 0xa6e1e8f1UL, 0xe7d0f3e8UL, 0x2483dec3UL, 0x65b2c5daUL,
+    0xaaae5d5dUL, 0xeb9f4644UL, 0x28cc6b6fUL, 0x69fd7076UL, 0xae6b3139UL,
+    0xef5a2a20UL, 0x2c09070bUL, 0x6d381c12UL, 0xf33646dfUL, 0xb2075dc6UL,
+    0x715470edUL, 0x30656bf4UL, 0xf7f32abbUL, 0xb6c231a2UL, 0x75911c89UL,
+    0x34a00790UL, 0xfbbc9f17UL, 0xba8d840eUL, 0x79dea925UL, 0x38efb23cUL,
+    0xff79f373UL, 0xbe48e86aUL, 0x7d1bc541UL, 0x3c2ade58UL, 0x054f79f0UL,
+    0x447e62e9UL, 0x872d4fc2UL, 0xc61c54dbUL, 0x018a1594UL, 0x40bb0e8dUL,
+    0x83e823a6UL, 0xc2d938bfUL, 0x0dc5a038UL, 0x4cf4bb21UL, 0x8fa7960aUL,
+    0xce968d13UL, 0x0900cc5cUL, 0x4831d745UL, 0x8b62fa6eUL, 0xca53e177UL,
+    0x545dbbbaUL, 0x156ca0a3UL, 0xd63f8d88UL, 0x970e9691UL, 0x5098d7deUL,
+    0x11a9ccc7UL, 0xd2fae1ecUL, 0x93cbfaf5UL, 0x5cd76272UL, 0x1de6796bUL,
+    0xdeb55440UL, 0x9f844f59UL, 0x58120e16UL, 0x1923150fUL, 0xda703824UL,
+    0x9b41233dUL, 0xa76bfd65UL, 0xe65ae67cUL, 0x2509cb57UL, 0x6438d04eUL,
+    0xa3ae9101UL, 0xe29f8a18UL, 0x21cca733UL, 0x60fdbc2aUL, 0xafe124adUL,
+    0xeed03fb4UL, 0x2d83129fUL, 0x6cb20986UL, 0xab2448c9UL, 0xea1553d0UL,
+    0x29467efbUL, 0x687765e2UL, 0xf6793f2fUL, 0xb7482436UL, 0x741b091dUL,
+    0x352a1204UL, 0xf2bc534bUL, 0xb38d4852UL, 0x70de6579UL, 0x31ef7e60UL,
+    0xfef3e6e7UL, 0xbfc2fdfeUL, 0x7c91d0d5UL, 0x3da0cbccUL, 0xfa368a83UL,
+    0xbb07919aUL, 0x7854bcb1UL, 0x3965a7a8UL, 0x4b98833bUL, 0x0aa99822UL,
+    0xc9fab509UL, 0x88cbae10UL, 0x4f5def5fUL, 0x0e6cf446UL, 0xcd3fd96dUL,
+    0x8c0ec274UL, 0x43125af3UL, 0x022341eaUL, 0xc1706cc1UL, 0x804177d8UL,
+    0x47d73697UL, 0x06e62d8eUL, 0xc5b500a5UL, 0x84841bbcUL, 0x1a8a4171UL,
+    0x5bbb5a68UL, 0x98e87743UL, 0xd9d96c5aUL, 0x1e4f2d15UL, 0x5f7e360cUL,
+    0x9c2d1b27UL, 0xdd1c003eUL, 0x120098b9UL, 0x533183a0UL, 0x9062ae8bUL,
+    0xd153b592UL, 0x16c5f4ddUL, 0x57f4efc4UL, 0x94a7c2efUL, 0xd596d9f6UL,
+    0xe9bc07aeUL, 0xa88d1cb7UL, 0x6bde319cUL, 0x2aef2a85UL, 0xed796bcaUL,
+    0xac4870d3UL, 0x6f1b5df8UL, 0x2e2a46e1UL, 0xe136de66UL, 0xa007c57fUL,
+    0x6354e854UL, 0x2265f34dUL, 0xe5f3b202UL, 0xa4c2a91bUL, 0x67918430UL,
+    0x26a09f29UL, 0xb8aec5e4UL, 0xf99fdefdUL, 0x3accf3d6UL, 0x7bfde8cfUL,
+    0xbc6ba980UL, 0xfd5ab299UL, 0x3e099fb2UL, 0x7f3884abUL, 0xb0241c2cUL,
+    0xf1150735UL, 0x32462a1eUL, 0x73773107UL, 0xb4e17048UL, 0xf5d06b51UL,
+    0x3683467aUL, 0x77b25d63UL, 0x4ed7facbUL, 0x0fe6e1d2UL, 0xccb5ccf9UL,
+    0x8d84d7e0UL, 0x4a1296afUL, 0x0b238db6UL, 0xc870a09dUL, 0x8941bb84UL,
+    0x465d2303UL, 0x076c381aUL, 0xc43f1531UL, 0x850e0e28UL, 0x42984f67UL,
+    0x03a9547eUL, 0xc0fa7955UL, 0x81cb624cUL, 0x1fc53881UL, 0x5ef42398UL,
+    0x9da70eb3UL, 0xdc9615aaUL, 0x1b0054e5UL, 0x5a314ffcUL, 0x996262d7UL,
+    0xd85379ceUL, 0x174fe149UL, 0x567efa50UL, 0x952dd77bUL, 0xd41ccc62UL,
+    0x138a8d2dUL, 0x52bb9634UL, 0x91e8bb1fUL, 0xd0d9a006UL, 0xecf37e5eUL,
+    0xadc26547UL, 0x6e91486cUL, 0x2fa05375UL, 0xe836123aUL, 0xa9070923UL,
+    0x6a542408UL, 0x2b653f11UL, 0xe479a796UL, 0xa548bc8fUL, 0x661b91a4UL,
+    0x272a8abdUL, 0xe0bccbf2UL, 0xa18dd0ebUL, 0x62defdc0UL, 0x23efe6d9UL,
+    0xbde1bc14UL, 0xfcd0a70dUL, 0x3f838a26UL, 0x7eb2913fUL, 0xb924d070UL,
+    0xf815cb69UL, 0x3b46e642UL, 0x7a77fd5bUL, 0xb56b65dcUL, 0xf45a7ec5UL,
+    0x370953eeUL, 0x763848f7UL, 0xb1ae09b8UL, 0xf09f12a1UL, 0x33cc3f8aUL,
+    0x72fd2493UL
+  },
+  {
+    0x00000000UL, 0x376ac201UL, 0x6ed48403UL, 0x59be4602UL, 0xdca80907UL,
+    0xebc2cb06UL, 0xb27c8d04UL, 0x85164f05UL, 0xb851130eUL, 0x8f3bd10fUL,
+    0xd685970dUL, 0xe1ef550cUL, 0x64f91a09UL, 0x5393d808UL, 0x0a2d9e0aUL,
+    0x3d475c0bUL, 0x70a3261cUL, 0x47c9e41dUL, 0x1e77a21fUL, 0x291d601eUL,
+    0xac0b2f1bUL, 0x9b61ed1aUL, 0xc2dfab18UL, 0xf5b56919UL, 0xc8f23512UL,
+    0xff98f713UL, 0xa626b111UL, 0x914c7310UL, 0x145a3c15UL, 0x2330fe14UL,
+    0x7a8eb816UL, 0x4de47a17UL, 0xe0464d38UL, 0xd72c8f39UL, 0x8e92c93bUL,
+    0xb9f80b3aUL, 0x3cee443fUL, 0x0b84863eUL, 0x523ac03cUL, 0x6550023dUL,
+    0x58175e36UL, 0x6f7d9c37UL, 0x36c3da35UL, 0x01a91834UL, 0x84bf5731UL,
+    0xb3d59530UL, 0xea6bd332UL, 0xdd011133UL, 0x90e56b24UL, 0xa78fa925UL,
+    0xfe31ef27UL, 0xc95b2d26UL, 0x4c4d6223UL, 0x7b27a022UL, 0x2299e620UL,
+    0x15f32421UL, 0x28b4782aUL, 0x1fdeba2bUL, 0x4660fc29UL, 0x710a3e28UL,
+    0xf41c712dUL, 0xc376b32cUL, 0x9ac8f52eUL, 0xada2372fUL, 0xc08d9a70UL,
+    0xf7e75871UL, 0xae591e73UL, 0x9933dc72UL, 0x1c259377UL, 0x2b4f5176UL,
+    0x72f11774UL, 0x459bd575UL, 0x78dc897eUL, 0x4fb64b7fUL, 0x16080d7dUL,
+    0x2162cf7cUL, 0xa4748079UL, 0x931e4278UL, 0xcaa0047aUL, 0xfdcac67bUL,
+    0xb02ebc6cUL, 0x87447e6dUL, 0xdefa386fUL, 0xe990fa6eUL, 0x6c86b56bUL,
+    0x5bec776aUL, 0x02523168UL, 0x3538f369UL, 0x087faf62UL, 0x3f156d63UL,
+    0x66ab2b61UL, 0x51c1e960UL, 0xd4d7a665UL, 0xe3bd6464UL, 0xba032266UL,
+    0x8d69e067UL, 0x20cbd748UL, 0x17a11549UL, 0x4e1f534bUL, 0x7975914aUL,
+    0xfc63de4fUL, 0xcb091c4eUL, 0x92b75a4cUL, 0xa5dd984dUL, 0x989ac446UL,
+    0xaff00647UL, 0xf64e4045UL, 0xc1248244UL, 0x4432cd41UL, 0x73580f40UL,
+    0x2ae64942UL, 0x1d8c8b43UL, 0x5068f154UL, 0x67023355UL, 0x3ebc7557UL,
+    0x09d6b756UL, 0x8cc0f853UL, 0xbbaa3a52UL, 0xe2147c50UL, 0xd57ebe51UL,
+    0xe839e25aUL, 0xdf53205bUL, 0x86ed6659UL, 0xb187a458UL, 0x3491eb5dUL,
+    0x03fb295cUL, 0x5a456f5eUL, 0x6d2fad5fUL, 0x801b35e1UL, 0xb771f7e0UL,
+    0xeecfb1e2UL, 0xd9a573e3UL, 0x5cb33ce6UL, 0x6bd9fee7UL, 0x3267b8e5UL,
+    0x050d7ae4UL, 0x384a26efUL, 0x0f20e4eeUL, 0x569ea2ecUL, 0x61f460edUL,
+    0xe4e22fe8UL, 0xd388ede9UL, 0x8a36abebUL, 0xbd5c69eaUL, 0xf0b813fdUL,
+    0xc7d2d1fcUL, 0x9e6c97feUL, 0xa90655ffUL, 0x2c101afaUL, 0x1b7ad8fbUL,
+    0x42c49ef9UL, 0x75ae5cf8UL, 0x48e900f3UL, 0x7f83c2f2UL, 0x263d84f0UL,
+    0x115746f1UL, 0x944109f4UL, 0xa32bcbf5UL, 0xfa958df7UL, 0xcdff4ff6UL,
+    0x605d78d9UL, 0x5737bad8UL, 0x0e89fcdaUL, 0x39e33edbUL, 0xbcf571deUL,
+    0x8b9fb3dfUL, 0xd221f5ddUL, 0xe54b37dcUL, 0xd80c6bd7UL, 0xef66a9d6UL,
+    0xb6d8efd4UL, 0x81b22dd5UL, 0x04a462d0UL, 0x33cea0d1UL, 0x6a70e6d3UL,
+    0x5d1a24d2UL, 0x10fe5ec5UL, 0x27949cc4UL, 0x7e2adac6UL, 0x494018c7UL,
+    0xcc5657c2UL, 0xfb3c95c3UL, 0xa282d3c1UL, 0x95e811c0UL, 0xa8af4dcbUL,
+    0x9fc58fcaUL, 0xc67bc9c8UL, 0xf1110bc9UL, 0x740744ccUL, 0x436d86cdUL,
+    0x1ad3c0cfUL, 0x2db902ceUL, 0x4096af91UL, 0x77fc6d90UL, 0x2e422b92UL,
+    0x1928e993UL, 0x9c3ea696UL, 0xab546497UL, 0xf2ea2295UL, 0xc580e094UL,
+    0xf8c7bc9fUL, 0xcfad7e9eUL, 0x9613389cUL, 0xa179fa9dUL, 0x246fb598UL,
+    0x13057799UL, 0x4abb319bUL, 0x7dd1f39aUL, 0x3035898dUL, 0x075f4b8cUL,
+    0x5ee10d8eUL, 0x698bcf8fUL, 0xec9d808aUL, 0xdbf7428bUL, 0x82490489UL,
+    0xb523c688UL, 0x88649a83UL, 0xbf0e5882UL, 0xe6b01e80UL, 0xd1dadc81UL,
+    0x54cc9384UL, 0x63a65185UL, 0x3a181787UL, 0x0d72d586UL, 0xa0d0e2a9UL,
+    0x97ba20a8UL, 0xce0466aaUL, 0xf96ea4abUL, 0x7c78ebaeUL, 0x4b1229afUL,
+    0x12ac6fadUL, 0x25c6adacUL, 0x1881f1a7UL, 0x2feb33a6UL, 0x765575a4UL,
+    0x413fb7a5UL, 0xc429f8a0UL, 0xf3433aa1UL, 0xaafd7ca3UL, 0x9d97bea2UL,
+    0xd073c4b5UL, 0xe71906b4UL, 0xbea740b6UL, 0x89cd82b7UL, 0x0cdbcdb2UL,
+    0x3bb10fb3UL, 0x620f49b1UL, 0x55658bb0UL, 0x6822d7bbUL, 0x5f4815baUL,
+    0x06f653b8UL, 0x319c91b9UL, 0xb48adebcUL, 0x83e01cbdUL, 0xda5e5abfUL,
+    0xed3498beUL
+  },
+  {
+    0x00000000UL, 0x6567bcb8UL, 0x8bc809aaUL, 0xeeafb512UL, 0x5797628fUL,
+    0x32f0de37UL, 0xdc5f6b25UL, 0xb938d79dUL, 0xef28b4c5UL, 0x8a4f087dUL,
+    0x64e0bd6fUL, 0x018701d7UL, 0xb8bfd64aUL, 0xddd86af2UL, 0x3377dfe0UL,
+    0x56106358UL, 0x9f571950UL, 0xfa30a5e8UL, 0x149f10faUL, 0x71f8ac42UL,
+    0xc8c07bdfUL, 0xada7c767UL, 0x43087275UL, 0x266fcecdUL, 0x707fad95UL,
+    0x1518112dUL, 0xfbb7a43fUL, 0x9ed01887UL, 0x27e8cf1aUL, 0x428f73a2UL,
+    0xac20c6b0UL, 0xc9477a08UL, 0x3eaf32a0UL, 0x5bc88e18UL, 0xb5673b0aUL,
+    0xd00087b2UL, 0x6938502fUL, 0x0c5fec97UL, 0xe2f05985UL, 0x8797e53dUL,
+    0xd1878665UL, 0xb4e03addUL, 0x5a4f8fcfUL, 0x3f283377UL, 0x8610e4eaUL,
+    0xe3775852UL, 0x0dd8ed40UL, 0x68bf51f8UL, 0xa1f82bf0UL, 0xc49f9748UL,
+    0x2a30225aUL, 0x4f579ee2UL, 0xf66f497fUL, 0x9308f5c7UL, 0x7da740d5UL,
+    0x18c0fc6dUL, 0x4ed09f35UL, 0x2bb7238dUL, 0xc518969fUL, 0xa07f2a27UL,
+    0x1947fdbaUL, 0x7c204102UL, 0x928ff410UL, 0xf7e848a8UL, 0x3d58149bUL,
+    0x583fa823UL, 0xb6901d31UL, 0xd3f7a189UL, 0x6acf7614UL, 0x0fa8caacUL,
+    0xe1077fbeUL, 0x8460c306UL, 0xd270a05eUL, 0xb7171ce6UL, 0x59b8a9f4UL,
+    0x3cdf154cUL, 0x85e7c2d1UL, 0xe0807e69UL, 0x0e2fcb7bUL, 0x6b4877c3UL,
+    0xa20f0dcbUL, 0xc768b173UL, 0x29c70461UL, 0x4ca0b8d9UL, 0xf5986f44UL,
+    0x90ffd3fcUL, 0x7e5066eeUL, 0x1b37da56UL, 0x4d27b90eUL, 0x284005b6UL,
+    0xc6efb0a4UL, 0xa3880c1cUL, 0x1ab0db81UL, 0x7fd76739UL, 0x9178d22bUL,
+    0xf41f6e93UL, 0x03f7263bUL, 0x66909a83UL, 0x883f2f91UL, 0xed589329UL,
+    0x546044b4UL, 0x3107f80cUL, 0xdfa84d1eUL, 0xbacff1a6UL, 0xecdf92feUL,
+    0x89b82e46UL, 0x67179b54UL, 0x027027ecUL, 0xbb48f071UL, 0xde2f4cc9UL,
+    0x3080f9dbUL, 0x55e74563UL, 0x9ca03f6bUL, 0xf9c783d3UL, 0x176836c1UL,
+    0x720f8a79UL, 0xcb375de4UL, 0xae50e15cUL, 0x40ff544eUL, 0x2598e8f6UL,
+    0x73888baeUL, 0x16ef3716UL, 0xf8408204UL, 0x9d273ebcUL, 0x241fe921UL,
+    0x41785599UL, 0xafd7e08bUL, 0xcab05c33UL, 0x3bb659edUL, 0x5ed1e555UL,
+    0xb07e5047UL, 0xd519ecffUL, 0x6c213b62UL, 0x094687daUL, 0xe7e932c8UL,
+    0x828e8e70UL, 0xd49eed28UL, 0xb1f95190UL, 0x5f56e482UL, 0x3a31583aUL,
+    0x83098fa7UL, 0xe66e331fUL, 0x08c1860dUL, 0x6da63ab5UL, 0xa4e140bdUL,
+    0xc186fc05UL, 0x2f294917UL, 0x4a4ef5afUL, 0xf3762232UL, 0x96119e8aUL,
+    0x78be2b98UL, 0x1dd99720UL, 0x4bc9f478UL, 0x2eae48c0UL, 0xc001fdd2UL,
+    0xa566416aUL, 0x1c5e96f7UL, 0x79392a4fUL, 0x97969f5dUL, 0xf2f123e5UL,
+    0x05196b4dUL, 0x607ed7f5UL, 0x8ed162e7UL, 0xebb6de5fUL, 0x528e09c2UL,
+    0x37e9b57aUL, 0xd9460068UL, 0xbc21bcd0UL, 0xea31df88UL, 0x8f566330UL,
+    0x61f9d622UL, 0x049e6a9aUL, 0xbda6bd07UL, 0xd8c101bfUL, 0x366eb4adUL,
+    0x53090815UL, 0x9a4e721dUL, 0xff29cea5UL, 0x11867bb7UL, 0x74e1c70fUL,
+    0xcdd91092UL, 0xa8beac2aUL, 0x46111938UL, 0x2376a580UL, 0x7566c6d8UL,
+    0x10017a60UL, 0xfeaecf72UL, 0x9bc973caUL, 0x22f1a457UL, 0x479618efUL,
+    0xa939adfdUL, 0xcc5e1145UL, 0x06ee4d76UL, 0x6389f1ceUL, 0x8d2644dcUL,
+    0xe841f864UL, 0x51792ff9UL, 0x341e9341UL, 0xdab12653UL, 0xbfd69aebUL,
+    0xe9c6f9b3UL, 0x8ca1450bUL, 0x620ef019UL, 0x07694ca1UL, 0xbe519b3cUL,
+    0xdb362784UL, 0x35999296UL, 0x50fe2e2eUL, 0x99b95426UL, 0xfcdee89eUL,
+    0x12715d8cUL, 0x7716e134UL, 0xce2e36a9UL, 0xab498a11UL, 0x45e63f03UL,
+    0x208183bbUL, 0x7691e0e3UL, 0x13f65c5bUL, 0xfd59e949UL, 0x983e55f1UL,
+    0x2106826cUL, 0x44613ed4UL, 0xaace8bc6UL, 0xcfa9377eUL, 0x38417fd6UL,
+    0x5d26c36eUL, 0xb389767cUL, 0xd6eecac4UL, 0x6fd61d59UL, 0x0ab1a1e1UL,
+    0xe41e14f3UL, 0x8179a84bUL, 0xd769cb13UL, 0xb20e77abUL, 0x5ca1c2b9UL,
+    0x39c67e01UL, 0x80fea99cUL, 0xe5991524UL, 0x0b36a036UL, 0x6e511c8eUL,
+    0xa7166686UL, 0xc271da3eUL, 0x2cde6f2cUL, 0x49b9d394UL, 0xf0810409UL,
+    0x95e6b8b1UL, 0x7b490da3UL, 0x1e2eb11bUL, 0x483ed243UL, 0x2d596efbUL,
+    0xc3f6dbe9UL, 0xa6916751UL, 0x1fa9b0ccUL, 0x7ace0c74UL, 0x9461b966UL,
+    0xf10605deUL
+#endif
+  }
+};
diff --git a/distrib/zlib-1.2.3/deflate.c b/distrib/zlib-1.2.3/deflate.c
new file mode 100644
index 0000000..29ce1f6
--- /dev/null
+++ b/distrib/zlib-1.2.3/deflate.c
@@ -0,0 +1,1736 @@
+/* deflate.c -- compress data using the deflation algorithm
+ * Copyright (C) 1995-2005 Jean-loup Gailly.
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/*
+ *  ALGORITHM
+ *
+ *      The "deflation" process depends on being able to identify portions
+ *      of the input text which are identical to earlier input (within a
+ *      sliding window trailing behind the input currently being processed).
+ *
+ *      The most straightforward technique turns out to be the fastest for
+ *      most input files: try all possible matches and select the longest.
+ *      The key feature of this algorithm is that insertions into the string
+ *      dictionary are very simple and thus fast, and deletions are avoided
+ *      completely. Insertions are performed at each input character, whereas
+ *      string matches are performed only when the previous match ends. So it
+ *      is preferable to spend more time in matches to allow very fast string
+ *      insertions and avoid deletions. The matching algorithm for small
+ *      strings is inspired from that of Rabin & Karp. A brute force approach
+ *      is used to find longer strings when a small match has been found.
+ *      A similar algorithm is used in comic (by Jan-Mark Wams) and freeze
+ *      (by Leonid Broukhis).
+ *         A previous version of this file used a more sophisticated algorithm
+ *      (by Fiala and Greene) which is guaranteed to run in linear amortized
+ *      time, but has a larger average cost, uses more memory and is patented.
+ *      However the F&G algorithm may be faster for some highly redundant
+ *      files if the parameter max_chain_length (described below) is too large.
+ *
+ *  ACKNOWLEDGEMENTS
+ *
+ *      The idea of lazy evaluation of matches is due to Jan-Mark Wams, and
+ *      I found it in 'freeze' written by Leonid Broukhis.
+ *      Thanks to many people for bug reports and testing.
+ *
+ *  REFERENCES
+ *
+ *      Deutsch, L.P.,"DEFLATE Compressed Data Format Specification".
+ *      Available in http://www.ietf.org/rfc/rfc1951.txt
+ *
+ *      A description of the Rabin and Karp algorithm is given in the book
+ *         "Algorithms" by R. Sedgewick, Addison-Wesley, p252.
+ *
+ *      Fiala,E.R., and Greene,D.H.
+ *         Data Compression with Finite Windows, Comm.ACM, 32,4 (1989) 490-595
+ *
+ */
+
+/* @(#) $Id$ */
+
+#include "deflate.h"
+
+const char deflate_copyright[] =
+   " deflate 1.2.3 Copyright 1995-2005 Jean-loup Gailly ";
+/*
+  If you use the zlib library in a product, an acknowledgment is welcome
+  in the documentation of your product. If for some reason you cannot
+  include such an acknowledgment, I would appreciate that you keep this
+  copyright string in the executable of your product.
+ */
+
+/* ===========================================================================
+ *  Function prototypes.
+ */
+typedef enum {
+    need_more,      /* block not completed, need more input or more output */
+    block_done,     /* block flush performed */
+    finish_started, /* finish started, need only more output at next deflate */
+    finish_done     /* finish done, accept no more input or output */
+} block_state;
+
+typedef block_state (*compress_func) OF((deflate_state *s, int flush));
+/* Compression function. Returns the block state after the call. */
+
+local void fill_window    OF((deflate_state *s));
+local block_state deflate_stored OF((deflate_state *s, int flush));
+local block_state deflate_fast   OF((deflate_state *s, int flush));
+#ifndef FASTEST
+local block_state deflate_slow   OF((deflate_state *s, int flush));
+#endif
+local void lm_init        OF((deflate_state *s));
+local void putShortMSB    OF((deflate_state *s, uInt b));
+local void flush_pending  OF((z_streamp strm));
+local int read_buf        OF((z_streamp strm, Bytef *buf, unsigned size));
+#ifndef FASTEST
+#ifdef ASMV
+      void match_init OF((void)); /* asm code initialization */
+      uInt longest_match  OF((deflate_state *s, IPos cur_match));
+#else
+local uInt longest_match  OF((deflate_state *s, IPos cur_match));
+#endif
+#endif
+local uInt longest_match_fast OF((deflate_state *s, IPos cur_match));
+
+#ifdef DEBUG
+local  void check_match OF((deflate_state *s, IPos start, IPos match,
+                            int length));
+#endif
+
+/* ===========================================================================
+ * Local data
+ */
+
+#define NIL 0
+/* Tail of hash chains */
+
+#ifndef TOO_FAR
+#  define TOO_FAR 4096
+#endif
+/* Matches of length 3 are discarded if their distance exceeds TOO_FAR */
+
+#define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
+/* Minimum amount of lookahead, except at the end of the input file.
+ * See deflate.c for comments about the MIN_MATCH+1.
+ */
+
+/* Values for max_lazy_match, good_match and max_chain_length, depending on
+ * the desired pack level (0..9). The values given below have been tuned to
+ * exclude worst case performance for pathological files. Better values may be
+ * found for specific files.
+ */
+typedef struct config_s {
+   ush good_length; /* reduce lazy search above this match length */
+   ush max_lazy;    /* do not perform lazy search above this match length */
+   ush nice_length; /* quit search above this match length */
+   ush max_chain;
+   compress_func func;
+} config;
+
+#ifdef FASTEST
+local const config configuration_table[2] = {
+/*      good lazy nice chain */
+/* 0 */ {0,    0,  0,    0, deflate_stored},  /* store only */
+/* 1 */ {4,    4,  8,    4, deflate_fast}}; /* max speed, no lazy matches */
+#else
+local const config configuration_table[10] = {
+/*      good lazy nice chain */
+/* 0 */ {0,    0,  0,    0, deflate_stored},  /* store only */
+/* 1 */ {4,    4,  8,    4, deflate_fast}, /* max speed, no lazy matches */
+/* 2 */ {4,    5, 16,    8, deflate_fast},
+/* 3 */ {4,    6, 32,   32, deflate_fast},
+
+/* 4 */ {4,    4, 16,   16, deflate_slow},  /* lazy matches */
+/* 5 */ {8,   16, 32,   32, deflate_slow},
+/* 6 */ {8,   16, 128, 128, deflate_slow},
+/* 7 */ {8,   32, 128, 256, deflate_slow},
+/* 8 */ {32, 128, 258, 1024, deflate_slow},
+/* 9 */ {32, 258, 258, 4096, deflate_slow}}; /* max compression */
+#endif
+
+/* Note: the deflate() code requires max_lazy >= MIN_MATCH and max_chain >= 4
+ * For deflate_fast() (levels <= 3) good is ignored and lazy has a different
+ * meaning.
+ */
+
+#define EQUAL 0
+/* result of memcmp for equal strings */
+
+#ifndef NO_DUMMY_DECL
+struct static_tree_desc_s {int dummy;}; /* for buggy compilers */
+#endif
+
+/* ===========================================================================
+ * Update a hash value with the given input byte
+ * IN  assertion: all calls to to UPDATE_HASH are made with consecutive
+ *    input characters, so that a running hash key can be computed from the
+ *    previous key instead of complete recalculation each time.
+ */
+#define UPDATE_HASH(s,h,c) (h = (((h)<<s->hash_shift) ^ (c)) & s->hash_mask)
+
+
+/* ===========================================================================
+ * Insert string str in the dictionary and set match_head to the previous head
+ * of the hash chain (the most recent string with same hash key). Return
+ * the previous length of the hash chain.
+ * If this file is compiled with -DFASTEST, the compression level is forced
+ * to 1, and no hash chains are maintained.
+ * IN  assertion: all calls to to INSERT_STRING are made with consecutive
+ *    input characters and the first MIN_MATCH bytes of str are valid
+ *    (except for the last MIN_MATCH-1 bytes of the input file).
+ */
+#ifdef FASTEST
+#define INSERT_STRING(s, str, match_head) \
+   (UPDATE_HASH(s, s->ins_h, s->window[(str) + (MIN_MATCH-1)]), \
+    match_head = s->head[s->ins_h], \
+    s->head[s->ins_h] = (Pos)(str))
+#else
+#define INSERT_STRING(s, str, match_head) \
+   (UPDATE_HASH(s, s->ins_h, s->window[(str) + (MIN_MATCH-1)]), \
+    match_head = s->prev[(str) & s->w_mask] = s->head[s->ins_h], \
+    s->head[s->ins_h] = (Pos)(str))
+#endif
+
+/* ===========================================================================
+ * Initialize the hash table (avoiding 64K overflow for 16 bit systems).
+ * prev[] will be initialized on the fly.
+ */
+#define CLEAR_HASH(s) \
+    s->head[s->hash_size-1] = NIL; \
+    zmemzero((Bytef *)s->head, (unsigned)(s->hash_size-1)*sizeof(*s->head));
+
+/* ========================================================================= */
+int ZEXPORT deflateInit_(strm, level, version, stream_size)
+    z_streamp strm;
+    int level;
+    const char *version;
+    int stream_size;
+{
+    return deflateInit2_(strm, level, Z_DEFLATED, MAX_WBITS, DEF_MEM_LEVEL,
+                         Z_DEFAULT_STRATEGY, version, stream_size);
+    /* To do: ignore strm->next_in if we use it as window */
+}
+
+/* ========================================================================= */
+int ZEXPORT deflateInit2_(strm, level, method, windowBits, memLevel, strategy,
+                  version, stream_size)
+    z_streamp strm;
+    int  level;
+    int  method;
+    int  windowBits;
+    int  memLevel;
+    int  strategy;
+    const char *version;
+    int stream_size;
+{
+    deflate_state *s;
+    int wrap = 1;
+    static const char my_version[] = ZLIB_VERSION;
+
+    ushf *overlay;
+    /* We overlay pending_buf and d_buf+l_buf. This works since the average
+     * output size for (length,distance) codes is <= 24 bits.
+     */
+
+    if (version == Z_NULL || version[0] != my_version[0] ||
+        stream_size != sizeof(z_stream)) {
+        return Z_VERSION_ERROR;
+    }
+    if (strm == Z_NULL) return Z_STREAM_ERROR;
+
+    strm->msg = Z_NULL;
+    if (strm->zalloc == (alloc_func)0) {
+        strm->zalloc = zcalloc;
+        strm->opaque = (voidpf)0;
+    }
+    if (strm->zfree == (free_func)0) strm->zfree = zcfree;
+
+#ifdef FASTEST
+    if (level != 0) level = 1;
+#else
+    if (level == Z_DEFAULT_COMPRESSION) level = 6;
+#endif
+
+    if (windowBits < 0) { /* suppress zlib wrapper */
+        wrap = 0;
+        windowBits = -windowBits;
+    }
+#ifdef GZIP
+    else if (windowBits > 15) {
+        wrap = 2;       /* write gzip wrapper instead */
+        windowBits -= 16;
+    }
+#endif
+    if (memLevel < 1 || memLevel > MAX_MEM_LEVEL || method != Z_DEFLATED ||
+        windowBits < 8 || windowBits > 15 || level < 0 || level > 9 ||
+        strategy < 0 || strategy > Z_FIXED) {
+        return Z_STREAM_ERROR;
+    }
+    if (windowBits == 8) windowBits = 9;  /* until 256-byte window bug fixed */
+    s = (deflate_state *) ZALLOC(strm, 1, sizeof(deflate_state));
+    if (s == Z_NULL) return Z_MEM_ERROR;
+    strm->state = (struct internal_state FAR *)s;
+    s->strm = strm;
+
+    s->wrap = wrap;
+    s->gzhead = Z_NULL;
+    s->w_bits = windowBits;
+    s->w_size = 1 << s->w_bits;
+    s->w_mask = s->w_size - 1;
+
+    s->hash_bits = memLevel + 7;
+    s->hash_size = 1 << s->hash_bits;
+    s->hash_mask = s->hash_size - 1;
+    s->hash_shift =  ((s->hash_bits+MIN_MATCH-1)/MIN_MATCH);
+
+    s->window = (Bytef *) ZALLOC(strm, s->w_size, 2*sizeof(Byte));
+    s->prev   = (Posf *)  ZALLOC(strm, s->w_size, sizeof(Pos));
+    s->head   = (Posf *)  ZALLOC(strm, s->hash_size, sizeof(Pos));
+
+    s->lit_bufsize = 1 << (memLevel + 6); /* 16K elements by default */
+
+    overlay = (ushf *) ZALLOC(strm, s->lit_bufsize, sizeof(ush)+2);
+    s->pending_buf = (uchf *) overlay;
+    s->pending_buf_size = (ulg)s->lit_bufsize * (sizeof(ush)+2L);
+
+    if (s->window == Z_NULL || s->prev == Z_NULL || s->head == Z_NULL ||
+        s->pending_buf == Z_NULL) {
+        s->status = FINISH_STATE;
+        strm->msg = (char*)ERR_MSG(Z_MEM_ERROR);
+        deflateEnd (strm);
+        return Z_MEM_ERROR;
+    }
+    s->d_buf = overlay + s->lit_bufsize/sizeof(ush);
+    s->l_buf = s->pending_buf + (1+sizeof(ush))*s->lit_bufsize;
+
+    s->level = level;
+    s->strategy = strategy;
+    s->method = (Byte)method;
+
+    return deflateReset(strm);
+}
+
+/* ========================================================================= */
+int ZEXPORT deflateSetDictionary (strm, dictionary, dictLength)
+    z_streamp strm;
+    const Bytef *dictionary;
+    uInt  dictLength;
+{
+    deflate_state *s;
+    uInt length = dictLength;
+    uInt n;
+    IPos hash_head = 0;
+
+    if (strm == Z_NULL || strm->state == Z_NULL || dictionary == Z_NULL ||
+        strm->state->wrap == 2 ||
+        (strm->state->wrap == 1 && strm->state->status != INIT_STATE))
+        return Z_STREAM_ERROR;
+
+    s = strm->state;
+    if (s->wrap)
+        strm->adler = adler32(strm->adler, dictionary, dictLength);
+
+    if (length < MIN_MATCH) return Z_OK;
+    if (length > MAX_DIST(s)) {
+        length = MAX_DIST(s);
+        dictionary += dictLength - length; /* use the tail of the dictionary */
+    }
+    zmemcpy(s->window, dictionary, length);
+    s->strstart = length;
+    s->block_start = (long)length;
+
+    /* Insert all strings in the hash table (except for the last two bytes).
+     * s->lookahead stays null, so s->ins_h will be recomputed at the next
+     * call of fill_window.
+     */
+    s->ins_h = s->window[0];
+    UPDATE_HASH(s, s->ins_h, s->window[1]);
+    for (n = 0; n <= length - MIN_MATCH; n++) {
+        INSERT_STRING(s, n, hash_head);
+    }
+    if (hash_head) hash_head = 0;  /* to make compiler happy */
+    return Z_OK;
+}
+
+/* ========================================================================= */
+int ZEXPORT deflateReset (strm)
+    z_streamp strm;
+{
+    deflate_state *s;
+
+    if (strm == Z_NULL || strm->state == Z_NULL ||
+        strm->zalloc == (alloc_func)0 || strm->zfree == (free_func)0) {
+        return Z_STREAM_ERROR;
+    }
+
+    strm->total_in = strm->total_out = 0;
+    strm->msg = Z_NULL; /* use zfree if we ever allocate msg dynamically */
+    strm->data_type = Z_UNKNOWN;
+
+    s = (deflate_state *)strm->state;
+    s->pending = 0;
+    s->pending_out = s->pending_buf;
+
+    if (s->wrap < 0) {
+        s->wrap = -s->wrap; /* was made negative by deflate(..., Z_FINISH); */
+    }
+    s->status = s->wrap ? INIT_STATE : BUSY_STATE;
+    strm->adler =
+#ifdef GZIP
+        s->wrap == 2 ? crc32(0L, Z_NULL, 0) :
+#endif
+        adler32(0L, Z_NULL, 0);
+    s->last_flush = Z_NO_FLUSH;
+
+    _tr_init(s);
+    lm_init(s);
+
+    return Z_OK;
+}
+
+/* ========================================================================= */
+int ZEXPORT deflateSetHeader (strm, head)
+    z_streamp strm;
+    gz_headerp head;
+{
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    if (strm->state->wrap != 2) return Z_STREAM_ERROR;
+    strm->state->gzhead = head;
+    return Z_OK;
+}
+
+/* ========================================================================= */
+int ZEXPORT deflatePrime (strm, bits, value)
+    z_streamp strm;
+    int bits;
+    int value;
+{
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    strm->state->bi_valid = bits;
+    strm->state->bi_buf = (ush)(value & ((1 << bits) - 1));
+    return Z_OK;
+}
+
+/* ========================================================================= */
+int ZEXPORT deflateParams(strm, level, strategy)
+    z_streamp strm;
+    int level;
+    int strategy;
+{
+    deflate_state *s;
+    compress_func func;
+    int err = Z_OK;
+
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    s = strm->state;
+
+#ifdef FASTEST
+    if (level != 0) level = 1;
+#else
+    if (level == Z_DEFAULT_COMPRESSION) level = 6;
+#endif
+    if (level < 0 || level > 9 || strategy < 0 || strategy > Z_FIXED) {
+        return Z_STREAM_ERROR;
+    }
+    func = configuration_table[s->level].func;
+
+    if (func != configuration_table[level].func && strm->total_in != 0) {
+        /* Flush the last buffer: */
+        err = deflate(strm, Z_PARTIAL_FLUSH);
+    }
+    if (s->level != level) {
+        s->level = level;
+        s->max_lazy_match   = configuration_table[level].max_lazy;
+        s->good_match       = configuration_table[level].good_length;
+        s->nice_match       = configuration_table[level].nice_length;
+        s->max_chain_length = configuration_table[level].max_chain;
+    }
+    s->strategy = strategy;
+    return err;
+}
+
+/* ========================================================================= */
+int ZEXPORT deflateTune(strm, good_length, max_lazy, nice_length, max_chain)
+    z_streamp strm;
+    int good_length;
+    int max_lazy;
+    int nice_length;
+    int max_chain;
+{
+    deflate_state *s;
+
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    s = strm->state;
+    s->good_match = good_length;
+    s->max_lazy_match = max_lazy;
+    s->nice_match = nice_length;
+    s->max_chain_length = max_chain;
+    return Z_OK;
+}
+
+/* =========================================================================
+ * For the default windowBits of 15 and memLevel of 8, this function returns
+ * a close to exact, as well as small, upper bound on the compressed size.
+ * They are coded as constants here for a reason--if the #define's are
+ * changed, then this function needs to be changed as well.  The return
+ * value for 15 and 8 only works for those exact settings.
+ *
+ * For any setting other than those defaults for windowBits and memLevel,
+ * the value returned is a conservative worst case for the maximum expansion
+ * resulting from using fixed blocks instead of stored blocks, which deflate
+ * can emit on compressed data for some combinations of the parameters.
+ *
+ * This function could be more sophisticated to provide closer upper bounds
+ * for every combination of windowBits and memLevel, as well as wrap.
+ * But even the conservative upper bound of about 14% expansion does not
+ * seem onerous for output buffer allocation.
+ */
+uLong ZEXPORT deflateBound(strm, sourceLen)
+    z_streamp strm;
+    uLong sourceLen;
+{
+    deflate_state *s;
+    uLong destLen;
+
+    /* conservative upper bound */
+    destLen = sourceLen +
+              ((sourceLen + 7) >> 3) + ((sourceLen + 63) >> 6) + 11;
+
+    /* if can't get parameters, return conservative bound */
+    if (strm == Z_NULL || strm->state == Z_NULL)
+        return destLen;
+
+    /* if not default parameters, return conservative bound */
+    s = strm->state;
+    if (s->w_bits != 15 || s->hash_bits != 8 + 7)
+        return destLen;
+
+    /* default settings: return tight bound for that case */
+    return compressBound(sourceLen);
+}
+
+/* =========================================================================
+ * Put a short in the pending buffer. The 16-bit value is put in MSB order.
+ * IN assertion: the stream state is correct and there is enough room in
+ * pending_buf.
+ */
+local void putShortMSB (s, b)
+    deflate_state *s;
+    uInt b;
+{
+    put_byte(s, (Byte)(b >> 8));
+    put_byte(s, (Byte)(b & 0xff));
+}
+
+/* =========================================================================
+ * Flush as much pending output as possible. All deflate() output goes
+ * through this function so some applications may wish to modify it
+ * to avoid allocating a large strm->next_out buffer and copying into it.
+ * (See also read_buf()).
+ */
+local void flush_pending(strm)
+    z_streamp strm;
+{
+    unsigned len = strm->state->pending;
+
+    if (len > strm->avail_out) len = strm->avail_out;
+    if (len == 0) return;
+
+    zmemcpy(strm->next_out, strm->state->pending_out, len);
+    strm->next_out  += len;
+    strm->state->pending_out  += len;
+    strm->total_out += len;
+    strm->avail_out  -= len;
+    strm->state->pending -= len;
+    if (strm->state->pending == 0) {
+        strm->state->pending_out = strm->state->pending_buf;
+    }
+}
+
+/* ========================================================================= */
+int ZEXPORT deflate (strm, flush)
+    z_streamp strm;
+    int flush;
+{
+    int old_flush; /* value of flush param for previous deflate call */
+    deflate_state *s;
+
+    if (strm == Z_NULL || strm->state == Z_NULL ||
+        flush > Z_FINISH || flush < 0) {
+        return Z_STREAM_ERROR;
+    }
+    s = strm->state;
+
+    if (strm->next_out == Z_NULL ||
+        (strm->next_in == Z_NULL && strm->avail_in != 0) ||
+        (s->status == FINISH_STATE && flush != Z_FINISH)) {
+        ERR_RETURN(strm, Z_STREAM_ERROR);
+    }
+    if (strm->avail_out == 0) ERR_RETURN(strm, Z_BUF_ERROR);
+
+    s->strm = strm; /* just in case */
+    old_flush = s->last_flush;
+    s->last_flush = flush;
+
+    /* Write the header */
+    if (s->status == INIT_STATE) {
+#ifdef GZIP
+        if (s->wrap == 2) {
+            strm->adler = crc32(0L, Z_NULL, 0);
+            put_byte(s, 31);
+            put_byte(s, 139);
+            put_byte(s, 8);
+            if (s->gzhead == NULL) {
+                put_byte(s, 0);
+                put_byte(s, 0);
+                put_byte(s, 0);
+                put_byte(s, 0);
+                put_byte(s, 0);
+                put_byte(s, s->level == 9 ? 2 :
+                            (s->strategy >= Z_HUFFMAN_ONLY || s->level < 2 ?
+                             4 : 0));
+                put_byte(s, OS_CODE);
+                s->status = BUSY_STATE;
+            }
+            else {
+                put_byte(s, (s->gzhead->text ? 1 : 0) +
+                            (s->gzhead->hcrc ? 2 : 0) +
+                            (s->gzhead->extra == Z_NULL ? 0 : 4) +
+                            (s->gzhead->name == Z_NULL ? 0 : 8) +
+                            (s->gzhead->comment == Z_NULL ? 0 : 16)
+                        );
+                put_byte(s, (Byte)(s->gzhead->time & 0xff));
+                put_byte(s, (Byte)((s->gzhead->time >> 8) & 0xff));
+                put_byte(s, (Byte)((s->gzhead->time >> 16) & 0xff));
+                put_byte(s, (Byte)((s->gzhead->time >> 24) & 0xff));
+                put_byte(s, s->level == 9 ? 2 :
+                            (s->strategy >= Z_HUFFMAN_ONLY || s->level < 2 ?
+                             4 : 0));
+                put_byte(s, s->gzhead->os & 0xff);
+                if (s->gzhead->extra != NULL) {
+                    put_byte(s, s->gzhead->extra_len & 0xff);
+                    put_byte(s, (s->gzhead->extra_len >> 8) & 0xff);
+                }
+                if (s->gzhead->hcrc)
+                    strm->adler = crc32(strm->adler, s->pending_buf,
+                                        s->pending);
+                s->gzindex = 0;
+                s->status = EXTRA_STATE;
+            }
+        }
+        else
+#endif
+        {
+            uInt header = (Z_DEFLATED + ((s->w_bits-8)<<4)) << 8;
+            uInt level_flags;
+
+            if (s->strategy >= Z_HUFFMAN_ONLY || s->level < 2)
+                level_flags = 0;
+            else if (s->level < 6)
+                level_flags = 1;
+            else if (s->level == 6)
+                level_flags = 2;
+            else
+                level_flags = 3;
+            header |= (level_flags << 6);
+            if (s->strstart != 0) header |= PRESET_DICT;
+            header += 31 - (header % 31);
+
+            s->status = BUSY_STATE;
+            putShortMSB(s, header);
+
+            /* Save the adler32 of the preset dictionary: */
+            if (s->strstart != 0) {
+                putShortMSB(s, (uInt)(strm->adler >> 16));
+                putShortMSB(s, (uInt)(strm->adler & 0xffff));
+            }
+            strm->adler = adler32(0L, Z_NULL, 0);
+        }
+    }
+#ifdef GZIP
+    if (s->status == EXTRA_STATE) {
+        if (s->gzhead->extra != NULL) {
+            uInt beg = s->pending;  /* start of bytes to update crc */
+
+            while (s->gzindex < (s->gzhead->extra_len & 0xffff)) {
+                if (s->pending == s->pending_buf_size) {
+                    if (s->gzhead->hcrc && s->pending > beg)
+                        strm->adler = crc32(strm->adler, s->pending_buf + beg,
+                                            s->pending - beg);
+                    flush_pending(strm);
+                    beg = s->pending;
+                    if (s->pending == s->pending_buf_size)
+                        break;
+                }
+                put_byte(s, s->gzhead->extra[s->gzindex]);
+                s->gzindex++;
+            }
+            if (s->gzhead->hcrc && s->pending > beg)
+                strm->adler = crc32(strm->adler, s->pending_buf + beg,
+                                    s->pending - beg);
+            if (s->gzindex == s->gzhead->extra_len) {
+                s->gzindex = 0;
+                s->status = NAME_STATE;
+            }
+        }
+        else
+            s->status = NAME_STATE;
+    }
+    if (s->status == NAME_STATE) {
+        if (s->gzhead->name != NULL) {
+            uInt beg = s->pending;  /* start of bytes to update crc */
+            int val;
+
+            do {
+                if (s->pending == s->pending_buf_size) {
+                    if (s->gzhead->hcrc && s->pending > beg)
+                        strm->adler = crc32(strm->adler, s->pending_buf + beg,
+                                            s->pending - beg);
+                    flush_pending(strm);
+                    beg = s->pending;
+                    if (s->pending == s->pending_buf_size) {
+                        val = 1;
+                        break;
+                    }
+                }
+                val = s->gzhead->name[s->gzindex++];
+                put_byte(s, val);
+            } while (val != 0);
+            if (s->gzhead->hcrc && s->pending > beg)
+                strm->adler = crc32(strm->adler, s->pending_buf + beg,
+                                    s->pending - beg);
+            if (val == 0) {
+                s->gzindex = 0;
+                s->status = COMMENT_STATE;
+            }
+        }
+        else
+            s->status = COMMENT_STATE;
+    }
+    if (s->status == COMMENT_STATE) {
+        if (s->gzhead->comment != NULL) {
+            uInt beg = s->pending;  /* start of bytes to update crc */
+            int val;
+
+            do {
+                if (s->pending == s->pending_buf_size) {
+                    if (s->gzhead->hcrc && s->pending > beg)
+                        strm->adler = crc32(strm->adler, s->pending_buf + beg,
+                                            s->pending - beg);
+                    flush_pending(strm);
+                    beg = s->pending;
+                    if (s->pending == s->pending_buf_size) {
+                        val = 1;
+                        break;
+                    }
+                }
+                val = s->gzhead->comment[s->gzindex++];
+                put_byte(s, val);
+            } while (val != 0);
+            if (s->gzhead->hcrc && s->pending > beg)
+                strm->adler = crc32(strm->adler, s->pending_buf + beg,
+                                    s->pending - beg);
+            if (val == 0)
+                s->status = HCRC_STATE;
+        }
+        else
+            s->status = HCRC_STATE;
+    }
+    if (s->status == HCRC_STATE) {
+        if (s->gzhead->hcrc) {
+            if (s->pending + 2 > s->pending_buf_size)
+                flush_pending(strm);
+            if (s->pending + 2 <= s->pending_buf_size) {
+                put_byte(s, (Byte)(strm->adler & 0xff));
+                put_byte(s, (Byte)((strm->adler >> 8) & 0xff));
+                strm->adler = crc32(0L, Z_NULL, 0);
+                s->status = BUSY_STATE;
+            }
+        }
+        else
+            s->status = BUSY_STATE;
+    }
+#endif
+
+    /* Flush as much pending output as possible */
+    if (s->pending != 0) {
+        flush_pending(strm);
+        if (strm->avail_out == 0) {
+            /* Since avail_out is 0, deflate will be called again with
+             * more output space, but possibly with both pending and
+             * avail_in equal to zero. There won't be anything to do,
+             * but this is not an error situation so make sure we
+             * return OK instead of BUF_ERROR at next call of deflate:
+             */
+            s->last_flush = -1;
+            return Z_OK;
+        }
+
+    /* Make sure there is something to do and avoid duplicate consecutive
+     * flushes. For repeated and useless calls with Z_FINISH, we keep
+     * returning Z_STREAM_END instead of Z_BUF_ERROR.
+     */
+    } else if (strm->avail_in == 0 && flush <= old_flush &&
+               flush != Z_FINISH) {
+        ERR_RETURN(strm, Z_BUF_ERROR);
+    }
+
+    /* User must not provide more input after the first FINISH: */
+    if (s->status == FINISH_STATE && strm->avail_in != 0) {
+        ERR_RETURN(strm, Z_BUF_ERROR);
+    }
+
+    /* Start a new block or continue the current one.
+     */
+    if (strm->avail_in != 0 || s->lookahead != 0 ||
+        (flush != Z_NO_FLUSH && s->status != FINISH_STATE)) {
+        block_state bstate;
+
+        bstate = (*(configuration_table[s->level].func))(s, flush);
+
+        if (bstate == finish_started || bstate == finish_done) {
+            s->status = FINISH_STATE;
+        }
+        if (bstate == need_more || bstate == finish_started) {
+            if (strm->avail_out == 0) {
+                s->last_flush = -1; /* avoid BUF_ERROR next call, see above */
+            }
+            return Z_OK;
+            /* If flush != Z_NO_FLUSH && avail_out == 0, the next call
+             * of deflate should use the same flush parameter to make sure
+             * that the flush is complete. So we don't have to output an
+             * empty block here, this will be done at next call. This also
+             * ensures that for a very small output buffer, we emit at most
+             * one empty block.
+             */
+        }
+        if (bstate == block_done) {
+            if (flush == Z_PARTIAL_FLUSH) {
+                _tr_align(s);
+            } else { /* FULL_FLUSH or SYNC_FLUSH */
+                _tr_stored_block(s, (char*)0, 0L, 0);
+                /* For a full flush, this empty block will be recognized
+                 * as a special marker by inflate_sync().
+                 */
+                if (flush == Z_FULL_FLUSH) {
+                    CLEAR_HASH(s);             /* forget history */
+                }
+            }
+            flush_pending(strm);
+            if (strm->avail_out == 0) {
+              s->last_flush = -1; /* avoid BUF_ERROR at next call, see above */
+              return Z_OK;
+            }
+        }
+    }
+    Assert(strm->avail_out > 0, "bug2");
+
+    if (flush != Z_FINISH) return Z_OK;
+    if (s->wrap <= 0) return Z_STREAM_END;
+
+    /* Write the trailer */
+#ifdef GZIP
+    if (s->wrap == 2) {
+        put_byte(s, (Byte)(strm->adler & 0xff));
+        put_byte(s, (Byte)((strm->adler >> 8) & 0xff));
+        put_byte(s, (Byte)((strm->adler >> 16) & 0xff));
+        put_byte(s, (Byte)((strm->adler >> 24) & 0xff));
+        put_byte(s, (Byte)(strm->total_in & 0xff));
+        put_byte(s, (Byte)((strm->total_in >> 8) & 0xff));
+        put_byte(s, (Byte)((strm->total_in >> 16) & 0xff));
+        put_byte(s, (Byte)((strm->total_in >> 24) & 0xff));
+    }
+    else
+#endif
+    {
+        putShortMSB(s, (uInt)(strm->adler >> 16));
+        putShortMSB(s, (uInt)(strm->adler & 0xffff));
+    }
+    flush_pending(strm);
+    /* If avail_out is zero, the application will call deflate again
+     * to flush the rest.
+     */
+    if (s->wrap > 0) s->wrap = -s->wrap; /* write the trailer only once! */
+    return s->pending != 0 ? Z_OK : Z_STREAM_END;
+}
+
+/* ========================================================================= */
+int ZEXPORT deflateEnd (strm)
+    z_streamp strm;
+{
+    int status;
+
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+
+    status = strm->state->status;
+    if (status != INIT_STATE &&
+        status != EXTRA_STATE &&
+        status != NAME_STATE &&
+        status != COMMENT_STATE &&
+        status != HCRC_STATE &&
+        status != BUSY_STATE &&
+        status != FINISH_STATE) {
+      return Z_STREAM_ERROR;
+    }
+
+    /* Deallocate in reverse order of allocations: */
+    TRY_FREE(strm, strm->state->pending_buf);
+    TRY_FREE(strm, strm->state->head);
+    TRY_FREE(strm, strm->state->prev);
+    TRY_FREE(strm, strm->state->window);
+
+    ZFREE(strm, strm->state);
+    strm->state = Z_NULL;
+
+    return status == BUSY_STATE ? Z_DATA_ERROR : Z_OK;
+}
+
+/* =========================================================================
+ * Copy the source state to the destination state.
+ * To simplify the source, this is not supported for 16-bit MSDOS (which
+ * doesn't have enough memory anyway to duplicate compression states).
+ */
+int ZEXPORT deflateCopy (dest, source)
+    z_streamp dest;
+    z_streamp source;
+{
+#ifdef MAXSEG_64K
+    return Z_STREAM_ERROR;
+#else
+    deflate_state *ds;
+    deflate_state *ss;
+    ushf *overlay;
+
+
+    if (source == Z_NULL || dest == Z_NULL || source->state == Z_NULL) {
+        return Z_STREAM_ERROR;
+    }
+
+    ss = source->state;
+
+    zmemcpy(dest, source, sizeof(z_stream));
+
+    ds = (deflate_state *) ZALLOC(dest, 1, sizeof(deflate_state));
+    if (ds == Z_NULL) return Z_MEM_ERROR;
+    dest->state = (struct internal_state FAR *) ds;
+    zmemcpy(ds, ss, sizeof(deflate_state));
+    ds->strm = dest;
+
+    ds->window = (Bytef *) ZALLOC(dest, ds->w_size, 2*sizeof(Byte));
+    ds->prev   = (Posf *)  ZALLOC(dest, ds->w_size, sizeof(Pos));
+    ds->head   = (Posf *)  ZALLOC(dest, ds->hash_size, sizeof(Pos));
+    overlay = (ushf *) ZALLOC(dest, ds->lit_bufsize, sizeof(ush)+2);
+    ds->pending_buf = (uchf *) overlay;
+
+    if (ds->window == Z_NULL || ds->prev == Z_NULL || ds->head == Z_NULL ||
+        ds->pending_buf == Z_NULL) {
+        deflateEnd (dest);
+        return Z_MEM_ERROR;
+    }
+    /* following zmemcpy do not work for 16-bit MSDOS */
+    zmemcpy(ds->window, ss->window, ds->w_size * 2 * sizeof(Byte));
+    zmemcpy(ds->prev, ss->prev, ds->w_size * sizeof(Pos));
+    zmemcpy(ds->head, ss->head, ds->hash_size * sizeof(Pos));
+    zmemcpy(ds->pending_buf, ss->pending_buf, (uInt)ds->pending_buf_size);
+
+    ds->pending_out = ds->pending_buf + (ss->pending_out - ss->pending_buf);
+    ds->d_buf = overlay + ds->lit_bufsize/sizeof(ush);
+    ds->l_buf = ds->pending_buf + (1+sizeof(ush))*ds->lit_bufsize;
+
+    ds->l_desc.dyn_tree = ds->dyn_ltree;
+    ds->d_desc.dyn_tree = ds->dyn_dtree;
+    ds->bl_desc.dyn_tree = ds->bl_tree;
+
+    return Z_OK;
+#endif /* MAXSEG_64K */
+}
+
+/* ===========================================================================
+ * Read a new buffer from the current input stream, update the adler32
+ * and total number of bytes read.  All deflate() input goes through
+ * this function so some applications may wish to modify it to avoid
+ * allocating a large strm->next_in buffer and copying from it.
+ * (See also flush_pending()).
+ */
+local int read_buf(strm, buf, size)
+    z_streamp strm;
+    Bytef *buf;
+    unsigned size;
+{
+    unsigned len = strm->avail_in;
+
+    if (len > size) len = size;
+    if (len == 0) return 0;
+
+    strm->avail_in  -= len;
+
+    if (strm->state->wrap == 1) {
+        strm->adler = adler32(strm->adler, strm->next_in, len);
+    }
+#ifdef GZIP
+    else if (strm->state->wrap == 2) {
+        strm->adler = crc32(strm->adler, strm->next_in, len);
+    }
+#endif
+    zmemcpy(buf, strm->next_in, len);
+    strm->next_in  += len;
+    strm->total_in += len;
+
+    return (int)len;
+}
+
+/* ===========================================================================
+ * Initialize the "longest match" routines for a new zlib stream
+ */
+local void lm_init (s)
+    deflate_state *s;
+{
+    s->window_size = (ulg)2L*s->w_size;
+
+    CLEAR_HASH(s);
+
+    /* Set the default configuration parameters:
+     */
+    s->max_lazy_match   = configuration_table[s->level].max_lazy;
+    s->good_match       = configuration_table[s->level].good_length;
+    s->nice_match       = configuration_table[s->level].nice_length;
+    s->max_chain_length = configuration_table[s->level].max_chain;
+
+    s->strstart = 0;
+    s->block_start = 0L;
+    s->lookahead = 0;
+    s->match_length = s->prev_length = MIN_MATCH-1;
+    s->match_available = 0;
+    s->ins_h = 0;
+#ifndef FASTEST
+#ifdef ASMV
+    match_init(); /* initialize the asm code */
+#endif
+#endif
+}
+
+#ifndef FASTEST
+/* ===========================================================================
+ * Set match_start to the longest match starting at the given string and
+ * return its length. Matches shorter or equal to prev_length are discarded,
+ * in which case the result is equal to prev_length and match_start is
+ * garbage.
+ * IN assertions: cur_match is the head of the hash chain for the current
+ *   string (strstart) and its distance is <= MAX_DIST, and prev_length >= 1
+ * OUT assertion: the match length is not greater than s->lookahead.
+ */
+#ifndef ASMV
+/* For 80x86 and 680x0, an optimized version will be provided in match.asm or
+ * match.S. The code will be functionally equivalent.
+ */
+local uInt longest_match(s, cur_match)
+    deflate_state *s;
+    IPos cur_match;                             /* current match */
+{
+    unsigned chain_length = s->max_chain_length;/* max hash chain length */
+    register Bytef *scan = s->window + s->strstart; /* current string */
+    register Bytef *match;                       /* matched string */
+    register int len;                           /* length of current match */
+    int best_len = s->prev_length;              /* best match length so far */
+    int nice_match = s->nice_match;             /* stop if match long enough */
+    IPos limit = s->strstart > (IPos)MAX_DIST(s) ?
+        s->strstart - (IPos)MAX_DIST(s) : NIL;
+    /* Stop when cur_match becomes <= limit. To simplify the code,
+     * we prevent matches with the string of window index 0.
+     */
+    Posf *prev = s->prev;
+    uInt wmask = s->w_mask;
+
+#ifdef UNALIGNED_OK
+    /* Compare two bytes at a time. Note: this is not always beneficial.
+     * Try with and without -DUNALIGNED_OK to check.
+     */
+    register Bytef *strend = s->window + s->strstart + MAX_MATCH - 1;
+    register ush scan_start = *(ushf*)scan;
+    register ush scan_end   = *(ushf*)(scan+best_len-1);
+#else
+    register Bytef *strend = s->window + s->strstart + MAX_MATCH;
+    register Byte scan_end1  = scan[best_len-1];
+    register Byte scan_end   = scan[best_len];
+#endif
+
+    /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16.
+     * It is easy to get rid of this optimization if necessary.
+     */
+    Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever");
+
+    /* Do not waste too much time if we already have a good match: */
+    if (s->prev_length >= s->good_match) {
+        chain_length >>= 2;
+    }
+    /* Do not look for matches beyond the end of the input. This is necessary
+     * to make deflate deterministic.
+     */
+    if ((uInt)nice_match > s->lookahead) nice_match = s->lookahead;
+
+    Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead");
+
+    do {
+        Assert(cur_match < s->strstart, "no future");
+        match = s->window + cur_match;
+
+        /* Skip to next match if the match length cannot increase
+         * or if the match length is less than 2.  Note that the checks below
+         * for insufficient lookahead only occur occasionally for performance
+         * reasons.  Therefore uninitialized memory will be accessed, and
+         * conditional jumps will be made that depend on those values.
+         * However the length of the match is limited to the lookahead, so
+         * the output of deflate is not affected by the uninitialized values.
+         */
+#if (defined(UNALIGNED_OK) && MAX_MATCH == 258)
+        /* This code assumes sizeof(unsigned short) == 2. Do not use
+         * UNALIGNED_OK if your compiler uses a different size.
+         */
+        if (*(ushf*)(match+best_len-1) != scan_end ||
+            *(ushf*)match != scan_start) continue;
+
+        /* It is not necessary to compare scan[2] and match[2] since they are
+         * always equal when the other bytes match, given that the hash keys
+         * are equal and that HASH_BITS >= 8. Compare 2 bytes at a time at
+         * strstart+3, +5, ... up to strstart+257. We check for insufficient
+         * lookahead only every 4th comparison; the 128th check will be made
+         * at strstart+257. If MAX_MATCH-2 is not a multiple of 8, it is
+         * necessary to put more guard bytes at the end of the window, or
+         * to check more often for insufficient lookahead.
+         */
+        Assert(scan[2] == match[2], "scan[2]?");
+        scan++, match++;
+        do {
+        } while (*(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
+                 *(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
+                 *(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
+                 *(ushf*)(scan+=2) == *(ushf*)(match+=2) &&
+                 scan < strend);
+        /* The funny "do {}" generates better code on most compilers */
+
+        /* Here, scan <= window+strstart+257 */
+        Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan");
+        if (*scan == *match) scan++;
+
+        len = (MAX_MATCH - 1) - (int)(strend-scan);
+        scan = strend - (MAX_MATCH-1);
+
+#else /* UNALIGNED_OK */
+
+        if (match[best_len]   != scan_end  ||
+            match[best_len-1] != scan_end1 ||
+            *match            != *scan     ||
+            *++match          != scan[1])      continue;
+
+        /* The check at best_len-1 can be removed because it will be made
+         * again later. (This heuristic is not always a win.)
+         * It is not necessary to compare scan[2] and match[2] since they
+         * are always equal when the other bytes match, given that
+         * the hash keys are equal and that HASH_BITS >= 8.
+         */
+        scan += 2, match++;
+        Assert(*scan == *match, "match[2]?");
+
+        /* We check for insufficient lookahead only every 8th comparison;
+         * the 256th check will be made at strstart+258.
+         */
+        do {
+        } while (*++scan == *++match && *++scan == *++match &&
+                 *++scan == *++match && *++scan == *++match &&
+                 *++scan == *++match && *++scan == *++match &&
+                 *++scan == *++match && *++scan == *++match &&
+                 scan < strend);
+
+        Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan");
+
+        len = MAX_MATCH - (int)(strend - scan);
+        scan = strend - MAX_MATCH;
+
+#endif /* UNALIGNED_OK */
+
+        if (len > best_len) {
+            s->match_start = cur_match;
+            best_len = len;
+            if (len >= nice_match) break;
+#ifdef UNALIGNED_OK
+            scan_end = *(ushf*)(scan+best_len-1);
+#else
+            scan_end1  = scan[best_len-1];
+            scan_end   = scan[best_len];
+#endif
+        }
+    } while ((cur_match = prev[cur_match & wmask]) > limit
+             && --chain_length != 0);
+
+    if ((uInt)best_len <= s->lookahead) return (uInt)best_len;
+    return s->lookahead;
+}
+#endif /* ASMV */
+#endif /* FASTEST */
+
+/* ---------------------------------------------------------------------------
+ * Optimized version for level == 1 or strategy == Z_RLE only
+ */
+local uInt longest_match_fast(s, cur_match)
+    deflate_state *s;
+    IPos cur_match;                             /* current match */
+{
+    register Bytef *scan = s->window + s->strstart; /* current string */
+    register Bytef *match;                       /* matched string */
+    register int len;                           /* length of current match */
+    register Bytef *strend = s->window + s->strstart + MAX_MATCH;
+
+    /* The code is optimized for HASH_BITS >= 8 and MAX_MATCH-2 multiple of 16.
+     * It is easy to get rid of this optimization if necessary.
+     */
+    Assert(s->hash_bits >= 8 && MAX_MATCH == 258, "Code too clever");
+
+    Assert((ulg)s->strstart <= s->window_size-MIN_LOOKAHEAD, "need lookahead");
+
+    Assert(cur_match < s->strstart, "no future");
+
+    match = s->window + cur_match;
+
+    /* Return failure if the match length is less than 2:
+     */
+    if (match[0] != scan[0] || match[1] != scan[1]) return MIN_MATCH-1;
+
+    /* The check at best_len-1 can be removed because it will be made
+     * again later. (This heuristic is not always a win.)
+     * It is not necessary to compare scan[2] and match[2] since they
+     * are always equal when the other bytes match, given that
+     * the hash keys are equal and that HASH_BITS >= 8.
+     */
+    scan += 2, match += 2;
+    Assert(*scan == *match, "match[2]?");
+
+    /* We check for insufficient lookahead only every 8th comparison;
+     * the 256th check will be made at strstart+258.
+     */
+    do {
+    } while (*++scan == *++match && *++scan == *++match &&
+             *++scan == *++match && *++scan == *++match &&
+             *++scan == *++match && *++scan == *++match &&
+             *++scan == *++match && *++scan == *++match &&
+             scan < strend);
+
+    Assert(scan <= s->window+(unsigned)(s->window_size-1), "wild scan");
+
+    len = MAX_MATCH - (int)(strend - scan);
+
+    if (len < MIN_MATCH) return MIN_MATCH - 1;
+
+    s->match_start = cur_match;
+    return (uInt)len <= s->lookahead ? (uInt)len : s->lookahead;
+}
+
+#ifdef DEBUG
+/* ===========================================================================
+ * Check that the match at match_start is indeed a match.
+ */
+local void check_match(s, start, match, length)
+    deflate_state *s;
+    IPos start, match;
+    int length;
+{
+    /* check that the match is indeed a match */
+    if (zmemcmp(s->window + match,
+                s->window + start, length) != EQUAL) {
+        fprintf(stderr, " start %u, match %u, length %d\n",
+                start, match, length);
+        do {
+            fprintf(stderr, "%c%c", s->window[match++], s->window[start++]);
+        } while (--length != 0);
+        z_error("invalid match");
+    }
+    if (z_verbose > 1) {
+        fprintf(stderr,"\\[%d,%d]", start-match, length);
+        do { putc(s->window[start++], stderr); } while (--length != 0);
+    }
+}
+#else
+#  define check_match(s, start, match, length)
+#endif /* DEBUG */
+
+/* ===========================================================================
+ * Fill the window when the lookahead becomes insufficient.
+ * Updates strstart and lookahead.
+ *
+ * IN assertion: lookahead < MIN_LOOKAHEAD
+ * OUT assertions: strstart <= window_size-MIN_LOOKAHEAD
+ *    At least one byte has been read, or avail_in == 0; reads are
+ *    performed for at least two bytes (required for the zip translate_eol
+ *    option -- not supported here).
+ */
+local void fill_window(s)
+    deflate_state *s;
+{
+    register unsigned n, m;
+    register Posf *p;
+    unsigned more;    /* Amount of free space at the end of the window. */
+    uInt wsize = s->w_size;
+
+    do {
+        more = (unsigned)(s->window_size -(ulg)s->lookahead -(ulg)s->strstart);
+
+        /* Deal with !@#$% 64K limit: */
+        if (sizeof(int) <= 2) {
+            if (more == 0 && s->strstart == 0 && s->lookahead == 0) {
+                more = wsize;
+
+            } else if (more == (unsigned)(-1)) {
+                /* Very unlikely, but possible on 16 bit machine if
+                 * strstart == 0 && lookahead == 1 (input done a byte at time)
+                 */
+                more--;
+            }
+        }
+
+        /* If the window is almost full and there is insufficient lookahead,
+         * move the upper half to the lower one to make room in the upper half.
+         */
+        if (s->strstart >= wsize+MAX_DIST(s)) {
+
+            zmemcpy(s->window, s->window+wsize, (unsigned)wsize);
+            s->match_start -= wsize;
+            s->strstart    -= wsize; /* we now have strstart >= MAX_DIST */
+            s->block_start -= (long) wsize;
+
+            /* Slide the hash table (could be avoided with 32 bit values
+               at the expense of memory usage). We slide even when level == 0
+               to keep the hash table consistent if we switch back to level > 0
+               later. (Using level 0 permanently is not an optimal usage of
+               zlib, so we don't care about this pathological case.)
+             */
+            /* %%% avoid this when Z_RLE */
+            n = s->hash_size;
+            p = &s->head[n];
+            do {
+                m = *--p;
+                *p = (Pos)(m >= wsize ? m-wsize : NIL);
+            } while (--n);
+
+            n = wsize;
+#ifndef FASTEST
+            p = &s->prev[n];
+            do {
+                m = *--p;
+                *p = (Pos)(m >= wsize ? m-wsize : NIL);
+                /* If n is not on any hash chain, prev[n] is garbage but
+                 * its value will never be used.
+                 */
+            } while (--n);
+#endif
+            more += wsize;
+        }
+        if (s->strm->avail_in == 0) return;
+
+        /* If there was no sliding:
+         *    strstart <= WSIZE+MAX_DIST-1 && lookahead <= MIN_LOOKAHEAD - 1 &&
+         *    more == window_size - lookahead - strstart
+         * => more >= window_size - (MIN_LOOKAHEAD-1 + WSIZE + MAX_DIST-1)
+         * => more >= window_size - 2*WSIZE + 2
+         * In the BIG_MEM or MMAP case (not yet supported),
+         *   window_size == input_size + MIN_LOOKAHEAD  &&
+         *   strstart + s->lookahead <= input_size => more >= MIN_LOOKAHEAD.
+         * Otherwise, window_size == 2*WSIZE so more >= 2.
+         * If there was sliding, more >= WSIZE. So in all cases, more >= 2.
+         */
+        Assert(more >= 2, "more < 2");
+
+        n = read_buf(s->strm, s->window + s->strstart + s->lookahead, more);
+        s->lookahead += n;
+
+        /* Initialize the hash value now that we have some input: */
+        if (s->lookahead >= MIN_MATCH) {
+            s->ins_h = s->window[s->strstart];
+            UPDATE_HASH(s, s->ins_h, s->window[s->strstart+1]);
+#if MIN_MATCH != 3
+            Call UPDATE_HASH() MIN_MATCH-3 more times
+#endif
+        }
+        /* If the whole input has less than MIN_MATCH bytes, ins_h is garbage,
+         * but this is not important since only literal bytes will be emitted.
+         */
+
+    } while (s->lookahead < MIN_LOOKAHEAD && s->strm->avail_in != 0);
+}
+
+/* ===========================================================================
+ * Flush the current block, with given end-of-file flag.
+ * IN assertion: strstart is set to the end of the current match.
+ */
+#define FLUSH_BLOCK_ONLY(s, eof) { \
+   _tr_flush_block(s, (s->block_start >= 0L ? \
+                   (charf *)&s->window[(unsigned)s->block_start] : \
+                   (charf *)Z_NULL), \
+                (ulg)((long)s->strstart - s->block_start), \
+                (eof)); \
+   s->block_start = s->strstart; \
+   flush_pending(s->strm); \
+   Tracev((stderr,"[FLUSH]")); \
+}
+
+/* Same but force premature exit if necessary. */
+#define FLUSH_BLOCK(s, eof) { \
+   FLUSH_BLOCK_ONLY(s, eof); \
+   if (s->strm->avail_out == 0) return (eof) ? finish_started : need_more; \
+}
+
+/* ===========================================================================
+ * Copy without compression as much as possible from the input stream, return
+ * the current block state.
+ * This function does not insert new strings in the dictionary since
+ * uncompressible data is probably not useful. This function is used
+ * only for the level=0 compression option.
+ * NOTE: this function should be optimized to avoid extra copying from
+ * window to pending_buf.
+ */
+local block_state deflate_stored(s, flush)
+    deflate_state *s;
+    int flush;
+{
+    /* Stored blocks are limited to 0xffff bytes, pending_buf is limited
+     * to pending_buf_size, and each stored block has a 5 byte header:
+     */
+    ulg max_block_size = 0xffff;
+    ulg max_start;
+
+    if (max_block_size > s->pending_buf_size - 5) {
+        max_block_size = s->pending_buf_size - 5;
+    }
+
+    /* Copy as much as possible from input to output: */
+    for (;;) {
+        /* Fill the window as much as possible: */
+        if (s->lookahead <= 1) {
+
+            Assert(s->strstart < s->w_size+MAX_DIST(s) ||
+                   s->block_start >= (long)s->w_size, "slide too late");
+
+            fill_window(s);
+            if (s->lookahead == 0 && flush == Z_NO_FLUSH) return need_more;
+
+            if (s->lookahead == 0) break; /* flush the current block */
+        }
+        Assert(s->block_start >= 0L, "block gone");
+
+        s->strstart += s->lookahead;
+        s->lookahead = 0;
+
+        /* Emit a stored block if pending_buf will be full: */
+        max_start = s->block_start + max_block_size;
+        if (s->strstart == 0 || (ulg)s->strstart >= max_start) {
+            /* strstart == 0 is possible when wraparound on 16-bit machine */
+            s->lookahead = (uInt)(s->strstart - max_start);
+            s->strstart = (uInt)max_start;
+            FLUSH_BLOCK(s, 0);
+        }
+        /* Flush if we may have to slide, otherwise block_start may become
+         * negative and the data will be gone:
+         */
+        if (s->strstart - (uInt)s->block_start >= MAX_DIST(s)) {
+            FLUSH_BLOCK(s, 0);
+        }
+    }
+    FLUSH_BLOCK(s, flush == Z_FINISH);
+    return flush == Z_FINISH ? finish_done : block_done;
+}
+
+/* ===========================================================================
+ * Compress as much as possible from the input stream, return the current
+ * block state.
+ * This function does not perform lazy evaluation of matches and inserts
+ * new strings in the dictionary only for unmatched strings or for short
+ * matches. It is used only for the fast compression options.
+ */
+local block_state deflate_fast(s, flush)
+    deflate_state *s;
+    int flush;
+{
+    IPos hash_head = NIL; /* head of the hash chain */
+    int bflush;           /* set if current block must be flushed */
+
+    for (;;) {
+        /* Make sure that we always have enough lookahead, except
+         * at the end of the input file. We need MAX_MATCH bytes
+         * for the next match, plus MIN_MATCH bytes to insert the
+         * string following the next match.
+         */
+        if (s->lookahead < MIN_LOOKAHEAD) {
+            fill_window(s);
+            if (s->lookahead < MIN_LOOKAHEAD && flush == Z_NO_FLUSH) {
+                return need_more;
+            }
+            if (s->lookahead == 0) break; /* flush the current block */
+        }
+
+        /* Insert the string window[strstart .. strstart+2] in the
+         * dictionary, and set hash_head to the head of the hash chain:
+         */
+        if (s->lookahead >= MIN_MATCH) {
+            INSERT_STRING(s, s->strstart, hash_head);
+        }
+
+        /* Find the longest match, discarding those <= prev_length.
+         * At this point we have always match_length < MIN_MATCH
+         */
+        if (hash_head != NIL && s->strstart - hash_head <= MAX_DIST(s)) {
+            /* To simplify the code, we prevent matches with the string
+             * of window index 0 (in particular we have to avoid a match
+             * of the string with itself at the start of the input file).
+             */
+#ifdef FASTEST
+            if ((s->strategy != Z_HUFFMAN_ONLY && s->strategy != Z_RLE) ||
+                (s->strategy == Z_RLE && s->strstart - hash_head == 1)) {
+                s->match_length = longest_match_fast (s, hash_head);
+            }
+#else
+            if (s->strategy != Z_HUFFMAN_ONLY && s->strategy != Z_RLE) {
+                s->match_length = longest_match (s, hash_head);
+            } else if (s->strategy == Z_RLE && s->strstart - hash_head == 1) {
+                s->match_length = longest_match_fast (s, hash_head);
+            }
+#endif
+            /* longest_match() or longest_match_fast() sets match_start */
+        }
+        if (s->match_length >= MIN_MATCH) {
+            check_match(s, s->strstart, s->match_start, s->match_length);
+
+            _tr_tally_dist(s, s->strstart - s->match_start,
+                           s->match_length - MIN_MATCH, bflush);
+
+            s->lookahead -= s->match_length;
+
+            /* Insert new strings in the hash table only if the match length
+             * is not too large. This saves time but degrades compression.
+             */
+#ifndef FASTEST
+            if (s->match_length <= s->max_insert_length &&
+                s->lookahead >= MIN_MATCH) {
+                s->match_length--; /* string at strstart already in table */
+                do {
+                    s->strstart++;
+                    INSERT_STRING(s, s->strstart, hash_head);
+                    /* strstart never exceeds WSIZE-MAX_MATCH, so there are
+                     * always MIN_MATCH bytes ahead.
+                     */
+                } while (--s->match_length != 0);
+                s->strstart++;
+            } else
+#endif
+            {
+                s->strstart += s->match_length;
+                s->match_length = 0;
+                s->ins_h = s->window[s->strstart];
+                UPDATE_HASH(s, s->ins_h, s->window[s->strstart+1]);
+#if MIN_MATCH != 3
+                Call UPDATE_HASH() MIN_MATCH-3 more times
+#endif
+                /* If lookahead < MIN_MATCH, ins_h is garbage, but it does not
+                 * matter since it will be recomputed at next deflate call.
+                 */
+            }
+        } else {
+            /* No match, output a literal byte */
+            Tracevv((stderr,"%c", s->window[s->strstart]));
+            _tr_tally_lit (s, s->window[s->strstart], bflush);
+            s->lookahead--;
+            s->strstart++;
+        }
+        if (bflush) FLUSH_BLOCK(s, 0);
+    }
+    FLUSH_BLOCK(s, flush == Z_FINISH);
+    return flush == Z_FINISH ? finish_done : block_done;
+}
+
+#ifndef FASTEST
+/* ===========================================================================
+ * Same as above, but achieves better compression. We use a lazy
+ * evaluation for matches: a match is finally adopted only if there is
+ * no better match at the next window position.
+ */
+local block_state deflate_slow(s, flush)
+    deflate_state *s;
+    int flush;
+{
+    IPos hash_head = NIL;    /* head of hash chain */
+    int bflush;              /* set if current block must be flushed */
+
+    /* Process the input block. */
+    for (;;) {
+        /* Make sure that we always have enough lookahead, except
+         * at the end of the input file. We need MAX_MATCH bytes
+         * for the next match, plus MIN_MATCH bytes to insert the
+         * string following the next match.
+         */
+        if (s->lookahead < MIN_LOOKAHEAD) {
+            fill_window(s);
+            if (s->lookahead < MIN_LOOKAHEAD && flush == Z_NO_FLUSH) {
+                return need_more;
+            }
+            if (s->lookahead == 0) break; /* flush the current block */
+        }
+
+        /* Insert the string window[strstart .. strstart+2] in the
+         * dictionary, and set hash_head to the head of the hash chain:
+         */
+        if (s->lookahead >= MIN_MATCH) {
+            INSERT_STRING(s, s->strstart, hash_head);
+        }
+
+        /* Find the longest match, discarding those <= prev_length.
+         */
+        s->prev_length = s->match_length, s->prev_match = s->match_start;
+        s->match_length = MIN_MATCH-1;
+
+        if (hash_head != NIL && s->prev_length < s->max_lazy_match &&
+            s->strstart - hash_head <= MAX_DIST(s)) {
+            /* To simplify the code, we prevent matches with the string
+             * of window index 0 (in particular we have to avoid a match
+             * of the string with itself at the start of the input file).
+             */
+            if (s->strategy != Z_HUFFMAN_ONLY && s->strategy != Z_RLE) {
+                s->match_length = longest_match (s, hash_head);
+            } else if (s->strategy == Z_RLE && s->strstart - hash_head == 1) {
+                s->match_length = longest_match_fast (s, hash_head);
+            }
+            /* longest_match() or longest_match_fast() sets match_start */
+
+            if (s->match_length <= 5 && (s->strategy == Z_FILTERED
+#if TOO_FAR <= 32767
+                || (s->match_length == MIN_MATCH &&
+                    s->strstart - s->match_start > TOO_FAR)
+#endif
+                )) {
+
+                /* If prev_match is also MIN_MATCH, match_start is garbage
+                 * but we will ignore the current match anyway.
+                 */
+                s->match_length = MIN_MATCH-1;
+            }
+        }
+        /* If there was a match at the previous step and the current
+         * match is not better, output the previous match:
+         */
+        if (s->prev_length >= MIN_MATCH && s->match_length <= s->prev_length) {
+            uInt max_insert = s->strstart + s->lookahead - MIN_MATCH;
+            /* Do not insert strings in hash table beyond this. */
+
+            check_match(s, s->strstart-1, s->prev_match, s->prev_length);
+
+            _tr_tally_dist(s, s->strstart -1 - s->prev_match,
+                           s->prev_length - MIN_MATCH, bflush);
+
+            /* Insert in hash table all strings up to the end of the match.
+             * strstart-1 and strstart are already inserted. If there is not
+             * enough lookahead, the last two strings are not inserted in
+             * the hash table.
+             */
+            s->lookahead -= s->prev_length-1;
+            s->prev_length -= 2;
+            do {
+                if (++s->strstart <= max_insert) {
+                    INSERT_STRING(s, s->strstart, hash_head);
+                }
+            } while (--s->prev_length != 0);
+            s->match_available = 0;
+            s->match_length = MIN_MATCH-1;
+            s->strstart++;
+
+            if (bflush) FLUSH_BLOCK(s, 0);
+
+        } else if (s->match_available) {
+            /* If there was no match at the previous position, output a
+             * single literal. If there was a match but the current match
+             * is longer, truncate the previous match to a single literal.
+             */
+            Tracevv((stderr,"%c", s->window[s->strstart-1]));
+            _tr_tally_lit(s, s->window[s->strstart-1], bflush);
+            if (bflush) {
+                FLUSH_BLOCK_ONLY(s, 0);
+            }
+            s->strstart++;
+            s->lookahead--;
+            if (s->strm->avail_out == 0) return need_more;
+        } else {
+            /* There is no previous match to compare with, wait for
+             * the next step to decide.
+             */
+            s->match_available = 1;
+            s->strstart++;
+            s->lookahead--;
+        }
+    }
+    Assert (flush != Z_NO_FLUSH, "no flush?");
+    if (s->match_available) {
+        Tracevv((stderr,"%c", s->window[s->strstart-1]));
+        _tr_tally_lit(s, s->window[s->strstart-1], bflush);
+        s->match_available = 0;
+    }
+    FLUSH_BLOCK(s, flush == Z_FINISH);
+    return flush == Z_FINISH ? finish_done : block_done;
+}
+#endif /* FASTEST */
+
+#if 0
+/* ===========================================================================
+ * For Z_RLE, simply look for runs of bytes, generate matches only of distance
+ * one.  Do not maintain a hash table.  (It will be regenerated if this run of
+ * deflate switches away from Z_RLE.)
+ */
+local block_state deflate_rle(s, flush)
+    deflate_state *s;
+    int flush;
+{
+    int bflush;         /* set if current block must be flushed */
+    uInt run;           /* length of run */
+    uInt max;           /* maximum length of run */
+    uInt prev;          /* byte at distance one to match */
+    Bytef *scan;        /* scan for end of run */
+
+    for (;;) {
+        /* Make sure that we always have enough lookahead, except
+         * at the end of the input file. We need MAX_MATCH bytes
+         * for the longest encodable run.
+         */
+        if (s->lookahead < MAX_MATCH) {
+            fill_window(s);
+            if (s->lookahead < MAX_MATCH && flush == Z_NO_FLUSH) {
+                return need_more;
+            }
+            if (s->lookahead == 0) break; /* flush the current block */
+        }
+
+        /* See how many times the previous byte repeats */
+        run = 0;
+        if (s->strstart > 0) {      /* if there is a previous byte, that is */
+            max = s->lookahead < MAX_MATCH ? s->lookahead : MAX_MATCH;
+            scan = s->window + s->strstart - 1;
+            prev = *scan++;
+            do {
+                if (*scan++ != prev)
+                    break;
+            } while (++run < max);
+        }
+
+        /* Emit match if have run of MIN_MATCH or longer, else emit literal */
+        if (run >= MIN_MATCH) {
+            check_match(s, s->strstart, s->strstart - 1, run);
+            _tr_tally_dist(s, 1, run - MIN_MATCH, bflush);
+            s->lookahead -= run;
+            s->strstart += run;
+        } else {
+            /* No match, output a literal byte */
+            Tracevv((stderr,"%c", s->window[s->strstart]));
+            _tr_tally_lit (s, s->window[s->strstart], bflush);
+            s->lookahead--;
+            s->strstart++;
+        }
+        if (bflush) FLUSH_BLOCK(s, 0);
+    }
+    FLUSH_BLOCK(s, flush == Z_FINISH);
+    return flush == Z_FINISH ? finish_done : block_done;
+}
+#endif
diff --git a/distrib/zlib-1.2.3/deflate.h b/distrib/zlib-1.2.3/deflate.h
new file mode 100644
index 0000000..05a5ab3
--- /dev/null
+++ b/distrib/zlib-1.2.3/deflate.h
@@ -0,0 +1,331 @@
+/* deflate.h -- internal compression state
+ * Copyright (C) 1995-2004 Jean-loup Gailly
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* WARNING: this file should *not* be used by applications. It is
+   part of the implementation of the compression library and is
+   subject to change. Applications should only use zlib.h.
+ */
+
+/* @(#) $Id$ */
+
+#ifndef DEFLATE_H
+#define DEFLATE_H
+
+#include "zutil.h"
+
+/* define NO_GZIP when compiling if you want to disable gzip header and
+   trailer creation by deflate().  NO_GZIP would be used to avoid linking in
+   the crc code when it is not needed.  For shared libraries, gzip encoding
+   should be left enabled. */
+#ifndef NO_GZIP
+#  define GZIP
+#endif
+
+/* ===========================================================================
+ * Internal compression state.
+ */
+
+#define LENGTH_CODES 29
+/* number of length codes, not counting the special END_BLOCK code */
+
+#define LITERALS  256
+/* number of literal bytes 0..255 */
+
+#define L_CODES (LITERALS+1+LENGTH_CODES)
+/* number of Literal or Length codes, including the END_BLOCK code */
+
+#define D_CODES   30
+/* number of distance codes */
+
+#define BL_CODES  19
+/* number of codes used to transfer the bit lengths */
+
+#define HEAP_SIZE (2*L_CODES+1)
+/* maximum heap size */
+
+#define MAX_BITS 15
+/* All codes must not exceed MAX_BITS bits */
+
+#define INIT_STATE    42
+#define EXTRA_STATE   69
+#define NAME_STATE    73
+#define COMMENT_STATE 91
+#define HCRC_STATE   103
+#define BUSY_STATE   113
+#define FINISH_STATE 666
+/* Stream status */
+
+
+/* Data structure describing a single value and its code string. */
+typedef struct ct_data_s {
+    union {
+        ush  freq;       /* frequency count */
+        ush  code;       /* bit string */
+    } fc;
+    union {
+        ush  dad;        /* father node in Huffman tree */
+        ush  len;        /* length of bit string */
+    } dl;
+} FAR ct_data;
+
+#define Freq fc.freq
+#define Code fc.code
+#define Dad  dl.dad
+#define Len  dl.len
+
+typedef struct static_tree_desc_s  static_tree_desc;
+
+typedef struct tree_desc_s {
+    ct_data *dyn_tree;           /* the dynamic tree */
+    int     max_code;            /* largest code with non zero frequency */
+    static_tree_desc *stat_desc; /* the corresponding static tree */
+} FAR tree_desc;
+
+typedef ush Pos;
+typedef Pos FAR Posf;
+typedef unsigned IPos;
+
+/* A Pos is an index in the character window. We use short instead of int to
+ * save space in the various tables. IPos is used only for parameter passing.
+ */
+
+typedef struct internal_state {
+    z_streamp strm;      /* pointer back to this zlib stream */
+    int   status;        /* as the name implies */
+    Bytef *pending_buf;  /* output still pending */
+    ulg   pending_buf_size; /* size of pending_buf */
+    Bytef *pending_out;  /* next pending byte to output to the stream */
+    uInt   pending;      /* nb of bytes in the pending buffer */
+    int   wrap;          /* bit 0 true for zlib, bit 1 true for gzip */
+    gz_headerp  gzhead;  /* gzip header information to write */
+    uInt   gzindex;      /* where in extra, name, or comment */
+    Byte  method;        /* STORED (for zip only) or DEFLATED */
+    int   last_flush;    /* value of flush param for previous deflate call */
+
+                /* used by deflate.c: */
+
+    uInt  w_size;        /* LZ77 window size (32K by default) */
+    uInt  w_bits;        /* log2(w_size)  (8..16) */
+    uInt  w_mask;        /* w_size - 1 */
+
+    Bytef *window;
+    /* Sliding window. Input bytes are read into the second half of the window,
+     * and move to the first half later to keep a dictionary of at least wSize
+     * bytes. With this organization, matches are limited to a distance of
+     * wSize-MAX_MATCH bytes, but this ensures that IO is always
+     * performed with a length multiple of the block size. Also, it limits
+     * the window size to 64K, which is quite useful on MSDOS.
+     * To do: use the user input buffer as sliding window.
+     */
+
+    ulg window_size;
+    /* Actual size of window: 2*wSize, except when the user input buffer
+     * is directly used as sliding window.
+     */
+
+    Posf *prev;
+    /* Link to older string with same hash index. To limit the size of this
+     * array to 64K, this link is maintained only for the last 32K strings.
+     * An index in this array is thus a window index modulo 32K.
+     */
+
+    Posf *head; /* Heads of the hash chains or NIL. */
+
+    uInt  ins_h;          /* hash index of string to be inserted */
+    uInt  hash_size;      /* number of elements in hash table */
+    uInt  hash_bits;      /* log2(hash_size) */
+    uInt  hash_mask;      /* hash_size-1 */
+
+    uInt  hash_shift;
+    /* Number of bits by which ins_h must be shifted at each input
+     * step. It must be such that after MIN_MATCH steps, the oldest
+     * byte no longer takes part in the hash key, that is:
+     *   hash_shift * MIN_MATCH >= hash_bits
+     */
+
+    long block_start;
+    /* Window position at the beginning of the current output block. Gets
+     * negative when the window is moved backwards.
+     */
+
+    uInt match_length;           /* length of best match */
+    IPos prev_match;             /* previous match */
+    int match_available;         /* set if previous match exists */
+    uInt strstart;               /* start of string to insert */
+    uInt match_start;            /* start of matching string */
+    uInt lookahead;              /* number of valid bytes ahead in window */
+
+    uInt prev_length;
+    /* Length of the best match at previous step. Matches not greater than this
+     * are discarded. This is used in the lazy match evaluation.
+     */
+
+    uInt max_chain_length;
+    /* To speed up deflation, hash chains are never searched beyond this
+     * length.  A higher limit improves compression ratio but degrades the
+     * speed.
+     */
+
+    uInt max_lazy_match;
+    /* Attempt to find a better match only when the current match is strictly
+     * smaller than this value. This mechanism is used only for compression
+     * levels >= 4.
+     */
+#   define max_insert_length  max_lazy_match
+    /* Insert new strings in the hash table only if the match length is not
+     * greater than this length. This saves time but degrades compression.
+     * max_insert_length is used only for compression levels <= 3.
+     */
+
+    int level;    /* compression level (1..9) */
+    int strategy; /* favor or force Huffman coding*/
+
+    uInt good_match;
+    /* Use a faster search when the previous match is longer than this */
+
+    int nice_match; /* Stop searching when current match exceeds this */
+
+                /* used by trees.c: */
+    /* Didn't use ct_data typedef below to supress compiler warning */
+    struct ct_data_s dyn_ltree[HEAP_SIZE];   /* literal and length tree */
+    struct ct_data_s dyn_dtree[2*D_CODES+1]; /* distance tree */
+    struct ct_data_s bl_tree[2*BL_CODES+1];  /* Huffman tree for bit lengths */
+
+    struct tree_desc_s l_desc;               /* desc. for literal tree */
+    struct tree_desc_s d_desc;               /* desc. for distance tree */
+    struct tree_desc_s bl_desc;              /* desc. for bit length tree */
+
+    ush bl_count[MAX_BITS+1];
+    /* number of codes at each bit length for an optimal tree */
+
+    int heap[2*L_CODES+1];      /* heap used to build the Huffman trees */
+    int heap_len;               /* number of elements in the heap */
+    int heap_max;               /* element of largest frequency */
+    /* The sons of heap[n] are heap[2*n] and heap[2*n+1]. heap[0] is not used.
+     * The same heap array is used to build all trees.
+     */
+
+    uch depth[2*L_CODES+1];
+    /* Depth of each subtree used as tie breaker for trees of equal frequency
+     */
+
+    uchf *l_buf;          /* buffer for literals or lengths */
+
+    uInt  lit_bufsize;
+    /* Size of match buffer for literals/lengths.  There are 4 reasons for
+     * limiting lit_bufsize to 64K:
+     *   - frequencies can be kept in 16 bit counters
+     *   - if compression is not successful for the first block, all input
+     *     data is still in the window so we can still emit a stored block even
+     *     when input comes from standard input.  (This can also be done for
+     *     all blocks if lit_bufsize is not greater than 32K.)
+     *   - if compression is not successful for a file smaller than 64K, we can
+     *     even emit a stored file instead of a stored block (saving 5 bytes).
+     *     This is applicable only for zip (not gzip or zlib).
+     *   - creating new Huffman trees less frequently may not provide fast
+     *     adaptation to changes in the input data statistics. (Take for
+     *     example a binary file with poorly compressible code followed by
+     *     a highly compressible string table.) Smaller buffer sizes give
+     *     fast adaptation but have of course the overhead of transmitting
+     *     trees more frequently.
+     *   - I can't count above 4
+     */
+
+    uInt last_lit;      /* running index in l_buf */
+
+    ushf *d_buf;
+    /* Buffer for distances. To simplify the code, d_buf and l_buf have
+     * the same number of elements. To use different lengths, an extra flag
+     * array would be necessary.
+     */
+
+    ulg opt_len;        /* bit length of current block with optimal trees */
+    ulg static_len;     /* bit length of current block with static trees */
+    uInt matches;       /* number of string matches in current block */
+    int last_eob_len;   /* bit length of EOB code for last block */
+
+#ifdef DEBUG
+    ulg compressed_len; /* total bit length of compressed file mod 2^32 */
+    ulg bits_sent;      /* bit length of compressed data sent mod 2^32 */
+#endif
+
+    ush bi_buf;
+    /* Output buffer. bits are inserted starting at the bottom (least
+     * significant bits).
+     */
+    int bi_valid;
+    /* Number of valid bits in bi_buf.  All bits above the last valid bit
+     * are always zero.
+     */
+
+} FAR deflate_state;
+
+/* Output a byte on the stream.
+ * IN assertion: there is enough room in pending_buf.
+ */
+#define put_byte(s, c) {s->pending_buf[s->pending++] = (c);}
+
+
+#define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
+/* Minimum amount of lookahead, except at the end of the input file.
+ * See deflate.c for comments about the MIN_MATCH+1.
+ */
+
+#define MAX_DIST(s)  ((s)->w_size-MIN_LOOKAHEAD)
+/* In order to simplify the code, particularly on 16 bit machines, match
+ * distances are limited to MAX_DIST instead of WSIZE.
+ */
+
+        /* in trees.c */
+void _tr_init         OF((deflate_state *s));
+int  _tr_tally        OF((deflate_state *s, unsigned dist, unsigned lc));
+void _tr_flush_block  OF((deflate_state *s, charf *buf, ulg stored_len,
+                          int eof));
+void _tr_align        OF((deflate_state *s));
+void _tr_stored_block OF((deflate_state *s, charf *buf, ulg stored_len,
+                          int eof));
+
+#define d_code(dist) \
+   ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
+/* Mapping from a distance to a distance code. dist is the distance - 1 and
+ * must not have side effects. _dist_code[256] and _dist_code[257] are never
+ * used.
+ */
+
+#ifndef DEBUG
+/* Inline versions of _tr_tally for speed: */
+
+#if defined(GEN_TREES_H) || !defined(STDC)
+  extern uch _length_code[];
+  extern uch _dist_code[];
+#else
+  extern const uch _length_code[];
+  extern const uch _dist_code[];
+#endif
+
+# define _tr_tally_lit(s, c, flush) \
+  { uch cc = (c); \
+    s->d_buf[s->last_lit] = 0; \
+    s->l_buf[s->last_lit++] = cc; \
+    s->dyn_ltree[cc].Freq++; \
+    flush = (s->last_lit == s->lit_bufsize-1); \
+   }
+# define _tr_tally_dist(s, distance, length, flush) \
+  { uch len = (length); \
+    ush dist = (distance); \
+    s->d_buf[s->last_lit] = dist; \
+    s->l_buf[s->last_lit++] = len; \
+    dist--; \
+    s->dyn_ltree[_length_code[len]+LITERALS+1].Freq++; \
+    s->dyn_dtree[d_code(dist)].Freq++; \
+    flush = (s->last_lit == s->lit_bufsize-1); \
+  }
+#else
+# define _tr_tally_lit(s, c, flush) flush = _tr_tally(s, 0, c)
+# define _tr_tally_dist(s, distance, length, flush) \
+              flush = _tr_tally(s, distance, length)
+#endif
+
+#endif /* DEFLATE_H */
diff --git a/distrib/zlib-1.2.3/gzio.c b/distrib/zlib-1.2.3/gzio.c
new file mode 100644
index 0000000..7e90f49
--- /dev/null
+++ b/distrib/zlib-1.2.3/gzio.c
@@ -0,0 +1,1026 @@
+/* gzio.c -- IO on .gz files
+ * Copyright (C) 1995-2005 Jean-loup Gailly.
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ *
+ * Compile this file with -DNO_GZCOMPRESS to avoid the compression code.
+ */
+
+/* @(#) $Id$ */
+
+#include <stdio.h>
+
+#include "zutil.h"
+
+#ifdef NO_DEFLATE       /* for compatibility with old definition */
+#  define NO_GZCOMPRESS
+#endif
+
+#ifndef NO_DUMMY_DECL
+struct internal_state {int dummy;}; /* for buggy compilers */
+#endif
+
+#ifndef Z_BUFSIZE
+#  ifdef MAXSEG_64K
+#    define Z_BUFSIZE 4096 /* minimize memory usage for 16-bit DOS */
+#  else
+#    define Z_BUFSIZE 16384
+#  endif
+#endif
+#ifndef Z_PRINTF_BUFSIZE
+#  define Z_PRINTF_BUFSIZE 4096
+#endif
+
+#ifdef __MVS__
+#  pragma map (fdopen , "\174\174FDOPEN")
+   FILE *fdopen(int, const char *);
+#endif
+
+#ifndef STDC
+extern voidp  malloc OF((uInt size));
+extern void   free   OF((voidpf ptr));
+#endif
+
+#define ALLOC(size) malloc(size)
+#define TRYFREE(p) {if (p) free(p);}
+
+static int const gz_magic[2] = {0x1f, 0x8b}; /* gzip magic header */
+
+/* gzip flag byte */
+#define ASCII_FLAG   0x01 /* bit 0 set: file probably ascii text */
+#define HEAD_CRC     0x02 /* bit 1 set: header CRC present */
+#define EXTRA_FIELD  0x04 /* bit 2 set: extra field present */
+#define ORIG_NAME    0x08 /* bit 3 set: original file name present */
+#define COMMENT      0x10 /* bit 4 set: file comment present */
+#define RESERVED     0xE0 /* bits 5..7: reserved */
+
+typedef struct gz_stream {
+    z_stream stream;
+    int      z_err;   /* error code for last stream operation */
+    int      z_eof;   /* set if end of input file */
+    FILE     *file;   /* .gz file */
+    Byte     *inbuf;  /* input buffer */
+    Byte     *outbuf; /* output buffer */
+    uLong    crc;     /* crc32 of uncompressed data */
+    char     *msg;    /* error message */
+    char     *path;   /* path name for debugging only */
+    int      transparent; /* 1 if input file is not a .gz file */
+    char     mode;    /* 'w' or 'r' */
+    z_off_t  start;   /* start of compressed data in file (header skipped) */
+    z_off_t  in;      /* bytes into deflate or inflate */
+    z_off_t  out;     /* bytes out of deflate or inflate */
+    int      back;    /* one character push-back */
+    int      last;    /* true if push-back is last character */
+} gz_stream;
+
+
+local gzFile gz_open      OF((const char *path, const char *mode, int  fd));
+local int do_flush        OF((gzFile file, int flush));
+local int    get_byte     OF((gz_stream *s));
+local void   check_header OF((gz_stream *s));
+local int    destroy      OF((gz_stream *s));
+local void   putLong      OF((FILE *file, uLong x));
+local uLong  getLong      OF((gz_stream *s));
+
+/* ===========================================================================
+     Opens a gzip (.gz) file for reading or writing. The mode parameter
+   is as in fopen ("rb" or "wb"). The file is given either by file descriptor
+   or path name (if fd == -1).
+     gz_open returns NULL if the file could not be opened or if there was
+   insufficient memory to allocate the (de)compression state; errno
+   can be checked to distinguish the two cases (if errno is zero, the
+   zlib error is Z_MEM_ERROR).
+*/
+local gzFile gz_open (path, mode, fd)
+    const char *path;
+    const char *mode;
+    int  fd;
+{
+    int err;
+    int level = Z_DEFAULT_COMPRESSION; /* compression level */
+    int strategy = Z_DEFAULT_STRATEGY; /* compression strategy */
+    char *p = (char*)mode;
+    gz_stream *s;
+    char fmode[80]; /* copy of mode, without the compression level */
+    char *m = fmode;
+
+    if (!path || !mode) return Z_NULL;
+
+    s = (gz_stream *)ALLOC(sizeof(gz_stream));
+    if (!s) return Z_NULL;
+
+    s->stream.zalloc = (alloc_func)0;
+    s->stream.zfree = (free_func)0;
+    s->stream.opaque = (voidpf)0;
+    s->stream.next_in = s->inbuf = Z_NULL;
+    s->stream.next_out = s->outbuf = Z_NULL;
+    s->stream.avail_in = s->stream.avail_out = 0;
+    s->file = NULL;
+    s->z_err = Z_OK;
+    s->z_eof = 0;
+    s->in = 0;
+    s->out = 0;
+    s->back = EOF;
+    s->crc = crc32(0L, Z_NULL, 0);
+    s->msg = NULL;
+    s->transparent = 0;
+
+    s->path = (char*)ALLOC(strlen(path)+1);
+    if (s->path == NULL) {
+        return destroy(s), (gzFile)Z_NULL;
+    }
+    strcpy(s->path, path); /* do this early for debugging */
+
+    s->mode = '\0';
+    do {
+        if (*p == 'r') s->mode = 'r';
+        if (*p == 'w' || *p == 'a') s->mode = 'w';
+        if (*p >= '0' && *p <= '9') {
+            level = *p - '0';
+        } else if (*p == 'f') {
+          strategy = Z_FILTERED;
+        } else if (*p == 'h') {
+          strategy = Z_HUFFMAN_ONLY;
+        } else if (*p == 'R') {
+          strategy = Z_RLE;
+        } else {
+            *m++ = *p; /* copy the mode */
+        }
+    } while (*p++ && m != fmode + sizeof(fmode));
+    if (s->mode == '\0') return destroy(s), (gzFile)Z_NULL;
+
+    if (s->mode == 'w') {
+#ifdef NO_GZCOMPRESS
+        err = Z_STREAM_ERROR;
+#else
+        err = deflateInit2(&(s->stream), level,
+                           Z_DEFLATED, -MAX_WBITS, DEF_MEM_LEVEL, strategy);
+        /* windowBits is passed < 0 to suppress zlib header */
+
+        s->stream.next_out = s->outbuf = (Byte*)ALLOC(Z_BUFSIZE);
+#endif
+        if (err != Z_OK || s->outbuf == Z_NULL) {
+            return destroy(s), (gzFile)Z_NULL;
+        }
+    } else {
+        s->stream.next_in  = s->inbuf = (Byte*)ALLOC(Z_BUFSIZE);
+
+        err = inflateInit2(&(s->stream), -MAX_WBITS);
+        /* windowBits is passed < 0 to tell that there is no zlib header.
+         * Note that in this case inflate *requires* an extra "dummy" byte
+         * after the compressed stream in order to complete decompression and
+         * return Z_STREAM_END. Here the gzip CRC32 ensures that 4 bytes are
+         * present after the compressed stream.
+         */
+        if (err != Z_OK || s->inbuf == Z_NULL) {
+            return destroy(s), (gzFile)Z_NULL;
+        }
+    }
+    s->stream.avail_out = Z_BUFSIZE;
+
+    errno = 0;
+    s->file = fd < 0 ? F_OPEN(path, fmode) : (FILE*)fdopen(fd, fmode);
+
+    if (s->file == NULL) {
+        return destroy(s), (gzFile)Z_NULL;
+    }
+    if (s->mode == 'w') {
+        /* Write a very simple .gz header:
+         */
+        fprintf(s->file, "%c%c%c%c%c%c%c%c%c%c", gz_magic[0], gz_magic[1],
+             Z_DEFLATED, 0 /*flags*/, 0,0,0,0 /*time*/, 0 /*xflags*/, OS_CODE);
+        s->start = 10L;
+        /* We use 10L instead of ftell(s->file) to because ftell causes an
+         * fflush on some systems. This version of the library doesn't use
+         * start anyway in write mode, so this initialization is not
+         * necessary.
+         */
+    } else {
+        check_header(s); /* skip the .gz header */
+        s->start = ftell(s->file) - s->stream.avail_in;
+    }
+
+    return (gzFile)s;
+}
+
+/* ===========================================================================
+     Opens a gzip (.gz) file for reading or writing.
+*/
+gzFile ZEXPORT gzopen (path, mode)
+    const char *path;
+    const char *mode;
+{
+    return gz_open (path, mode, -1);
+}
+
+/* ===========================================================================
+     Associate a gzFile with the file descriptor fd. fd is not dup'ed here
+   to mimic the behavio(u)r of fdopen.
+*/
+gzFile ZEXPORT gzdopen (fd, mode)
+    int fd;
+    const char *mode;
+{
+    char name[46];      /* allow for up to 128-bit integers */
+
+    if (fd < 0) return (gzFile)Z_NULL;
+    sprintf(name, "<fd:%d>", fd); /* for debugging */
+
+    return gz_open (name, mode, fd);
+}
+
+/* ===========================================================================
+ * Update the compression level and strategy
+ */
+int ZEXPORT gzsetparams (file, level, strategy)
+    gzFile file;
+    int level;
+    int strategy;
+{
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL || s->mode != 'w') return Z_STREAM_ERROR;
+
+    /* Make room to allow flushing */
+    if (s->stream.avail_out == 0) {
+
+        s->stream.next_out = s->outbuf;
+        if (fwrite(s->outbuf, 1, Z_BUFSIZE, s->file) != Z_BUFSIZE) {
+            s->z_err = Z_ERRNO;
+        }
+        s->stream.avail_out = Z_BUFSIZE;
+    }
+
+    return deflateParams (&(s->stream), level, strategy);
+}
+
+/* ===========================================================================
+     Read a byte from a gz_stream; update next_in and avail_in. Return EOF
+   for end of file.
+   IN assertion: the stream s has been sucessfully opened for reading.
+*/
+local int get_byte(s)
+    gz_stream *s;
+{
+    if (s->z_eof) return EOF;
+    if (s->stream.avail_in == 0) {
+        errno = 0;
+        s->stream.avail_in = (uInt)fread(s->inbuf, 1, Z_BUFSIZE, s->file);
+        if (s->stream.avail_in == 0) {
+            s->z_eof = 1;
+            if (ferror(s->file)) s->z_err = Z_ERRNO;
+            return EOF;
+        }
+        s->stream.next_in = s->inbuf;
+    }
+    s->stream.avail_in--;
+    return *(s->stream.next_in)++;
+}
+
+/* ===========================================================================
+      Check the gzip header of a gz_stream opened for reading. Set the stream
+    mode to transparent if the gzip magic header is not present; set s->err
+    to Z_DATA_ERROR if the magic header is present but the rest of the header
+    is incorrect.
+    IN assertion: the stream s has already been created sucessfully;
+       s->stream.avail_in is zero for the first time, but may be non-zero
+       for concatenated .gz files.
+*/
+local void check_header(s)
+    gz_stream *s;
+{
+    int method; /* method byte */
+    int flags;  /* flags byte */
+    uInt len;
+    int c;
+
+    /* Assure two bytes in the buffer so we can peek ahead -- handle case
+       where first byte of header is at the end of the buffer after the last
+       gzip segment */
+    len = s->stream.avail_in;
+    if (len < 2) {
+        if (len) s->inbuf[0] = s->stream.next_in[0];
+        errno = 0;
+        len = (uInt)fread(s->inbuf + len, 1, Z_BUFSIZE >> len, s->file);
+        if (len == 0 && ferror(s->file)) s->z_err = Z_ERRNO;
+        s->stream.avail_in += len;
+        s->stream.next_in = s->inbuf;
+        if (s->stream.avail_in < 2) {
+            s->transparent = s->stream.avail_in;
+            return;
+        }
+    }
+
+    /* Peek ahead to check the gzip magic header */
+    if (s->stream.next_in[0] != gz_magic[0] ||
+        s->stream.next_in[1] != gz_magic[1]) {
+        s->transparent = 1;
+        return;
+    }
+    s->stream.avail_in -= 2;
+    s->stream.next_in += 2;
+
+    /* Check the rest of the gzip header */
+    method = get_byte(s);
+    flags = get_byte(s);
+    if (method != Z_DEFLATED || (flags & RESERVED) != 0) {
+        s->z_err = Z_DATA_ERROR;
+        return;
+    }
+
+    /* Discard time, xflags and OS code: */
+    for (len = 0; len < 6; len++) (void)get_byte(s);
+
+    if ((flags & EXTRA_FIELD) != 0) { /* skip the extra field */
+        len  =  (uInt)get_byte(s);
+        len += ((uInt)get_byte(s))<<8;
+        /* len is garbage if EOF but the loop below will quit anyway */
+        while (len-- != 0 && get_byte(s) != EOF) ;
+    }
+    if ((flags & ORIG_NAME) != 0) { /* skip the original file name */
+        while ((c = get_byte(s)) != 0 && c != EOF) ;
+    }
+    if ((flags & COMMENT) != 0) {   /* skip the .gz file comment */
+        while ((c = get_byte(s)) != 0 && c != EOF) ;
+    }
+    if ((flags & HEAD_CRC) != 0) {  /* skip the header crc */
+        for (len = 0; len < 2; len++) (void)get_byte(s);
+    }
+    s->z_err = s->z_eof ? Z_DATA_ERROR : Z_OK;
+}
+
+ /* ===========================================================================
+ * Cleanup then free the given gz_stream. Return a zlib error code.
+   Try freeing in the reverse order of allocations.
+ */
+local int destroy (s)
+    gz_stream *s;
+{
+    int err = Z_OK;
+
+    if (!s) return Z_STREAM_ERROR;
+
+    TRYFREE(s->msg);
+
+    if (s->stream.state != NULL) {
+        if (s->mode == 'w') {
+#ifdef NO_GZCOMPRESS
+            err = Z_STREAM_ERROR;
+#else
+            err = deflateEnd(&(s->stream));
+#endif
+        } else if (s->mode == 'r') {
+            err = inflateEnd(&(s->stream));
+        }
+    }
+    if (s->file != NULL && fclose(s->file)) {
+#ifdef ESPIPE
+        if (errno != ESPIPE) /* fclose is broken for pipes in HP/UX */
+#endif
+            err = Z_ERRNO;
+    }
+    if (s->z_err < 0) err = s->z_err;
+
+    TRYFREE(s->inbuf);
+    TRYFREE(s->outbuf);
+    TRYFREE(s->path);
+    TRYFREE(s);
+    return err;
+}
+
+/* ===========================================================================
+     Reads the given number of uncompressed bytes from the compressed file.
+   gzread returns the number of bytes actually read (0 for end of file).
+*/
+int ZEXPORT gzread (file, buf, len)
+    gzFile file;
+    voidp buf;
+    unsigned len;
+{
+    gz_stream *s = (gz_stream*)file;
+    Bytef *start = (Bytef*)buf; /* starting point for crc computation */
+    Byte  *next_out; /* == stream.next_out but not forced far (for MSDOS) */
+
+    if (s == NULL || s->mode != 'r') return Z_STREAM_ERROR;
+
+    if (s->z_err == Z_DATA_ERROR || s->z_err == Z_ERRNO) return -1;
+    if (s->z_err == Z_STREAM_END) return 0;  /* EOF */
+
+    next_out = (Byte*)buf;
+    s->stream.next_out = (Bytef*)buf;
+    s->stream.avail_out = len;
+
+    if (s->stream.avail_out && s->back != EOF) {
+        *next_out++ = s->back;
+        s->stream.next_out++;
+        s->stream.avail_out--;
+        s->back = EOF;
+        s->out++;
+        start++;
+        if (s->last) {
+            s->z_err = Z_STREAM_END;
+            return 1;
+        }
+    }
+
+    while (s->stream.avail_out != 0) {
+
+        if (s->transparent) {
+            /* Copy first the lookahead bytes: */
+            uInt n = s->stream.avail_in;
+            if (n > s->stream.avail_out) n = s->stream.avail_out;
+            if (n > 0) {
+                zmemcpy(s->stream.next_out, s->stream.next_in, n);
+                next_out += n;
+                s->stream.next_out = next_out;
+                s->stream.next_in   += n;
+                s->stream.avail_out -= n;
+                s->stream.avail_in  -= n;
+            }
+            if (s->stream.avail_out > 0) {
+                s->stream.avail_out -=
+                    (uInt)fread(next_out, 1, s->stream.avail_out, s->file);
+            }
+            len -= s->stream.avail_out;
+            s->in  += len;
+            s->out += len;
+            if (len == 0) s->z_eof = 1;
+            return (int)len;
+        }
+        if (s->stream.avail_in == 0 && !s->z_eof) {
+
+            errno = 0;
+            s->stream.avail_in = (uInt)fread(s->inbuf, 1, Z_BUFSIZE, s->file);
+            if (s->stream.avail_in == 0) {
+                s->z_eof = 1;
+                if (ferror(s->file)) {
+                    s->z_err = Z_ERRNO;
+                    break;
+                }
+            }
+            s->stream.next_in = s->inbuf;
+        }
+        s->in += s->stream.avail_in;
+        s->out += s->stream.avail_out;
+        s->z_err = inflate(&(s->stream), Z_NO_FLUSH);
+        s->in -= s->stream.avail_in;
+        s->out -= s->stream.avail_out;
+
+        if (s->z_err == Z_STREAM_END) {
+            /* Check CRC and original size */
+            s->crc = crc32(s->crc, start, (uInt)(s->stream.next_out - start));
+            start = s->stream.next_out;
+
+            if (getLong(s) != s->crc) {
+                s->z_err = Z_DATA_ERROR;
+            } else {
+                (void)getLong(s);
+                /* The uncompressed length returned by above getlong() may be
+                 * different from s->out in case of concatenated .gz files.
+                 * Check for such files:
+                 */
+                check_header(s);
+                if (s->z_err == Z_OK) {
+                    inflateReset(&(s->stream));
+                    s->crc = crc32(0L, Z_NULL, 0);
+                }
+            }
+        }
+        if (s->z_err != Z_OK || s->z_eof) break;
+    }
+    s->crc = crc32(s->crc, start, (uInt)(s->stream.next_out - start));
+
+    if (len == s->stream.avail_out &&
+        (s->z_err == Z_DATA_ERROR || s->z_err == Z_ERRNO))
+        return -1;
+    return (int)(len - s->stream.avail_out);
+}
+
+
+/* ===========================================================================
+      Reads one byte from the compressed file. gzgetc returns this byte
+   or -1 in case of end of file or error.
+*/
+int ZEXPORT gzgetc(file)
+    gzFile file;
+{
+    unsigned char c;
+
+    return gzread(file, &c, 1) == 1 ? c : -1;
+}
+
+
+/* ===========================================================================
+      Push one byte back onto the stream.
+*/
+int ZEXPORT gzungetc(c, file)
+    int c;
+    gzFile file;
+{
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL || s->mode != 'r' || c == EOF || s->back != EOF) return EOF;
+    s->back = c;
+    s->out--;
+    s->last = (s->z_err == Z_STREAM_END);
+    if (s->last) s->z_err = Z_OK;
+    s->z_eof = 0;
+    return c;
+}
+
+
+/* ===========================================================================
+      Reads bytes from the compressed file until len-1 characters are
+   read, or a newline character is read and transferred to buf, or an
+   end-of-file condition is encountered.  The string is then terminated
+   with a null character.
+      gzgets returns buf, or Z_NULL in case of error.
+
+      The current implementation is not optimized at all.
+*/
+char * ZEXPORT gzgets(file, buf, len)
+    gzFile file;
+    char *buf;
+    int len;
+{
+    char *b = buf;
+    if (buf == Z_NULL || len <= 0) return Z_NULL;
+
+    while (--len > 0 && gzread(file, buf, 1) == 1 && *buf++ != '\n') ;
+    *buf = '\0';
+    return b == buf && len > 0 ? Z_NULL : b;
+}
+
+
+#ifndef NO_GZCOMPRESS
+/* ===========================================================================
+     Writes the given number of uncompressed bytes into the compressed file.
+   gzwrite returns the number of bytes actually written (0 in case of error).
+*/
+int ZEXPORT gzwrite (file, buf, len)
+    gzFile file;
+    voidpc buf;
+    unsigned len;
+{
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL || s->mode != 'w') return Z_STREAM_ERROR;
+
+    s->stream.next_in = (Bytef*)buf;
+    s->stream.avail_in = len;
+
+    while (s->stream.avail_in != 0) {
+
+        if (s->stream.avail_out == 0) {
+
+            s->stream.next_out = s->outbuf;
+            if (fwrite(s->outbuf, 1, Z_BUFSIZE, s->file) != Z_BUFSIZE) {
+                s->z_err = Z_ERRNO;
+                break;
+            }
+            s->stream.avail_out = Z_BUFSIZE;
+        }
+        s->in += s->stream.avail_in;
+        s->out += s->stream.avail_out;
+        s->z_err = deflate(&(s->stream), Z_NO_FLUSH);
+        s->in -= s->stream.avail_in;
+        s->out -= s->stream.avail_out;
+        if (s->z_err != Z_OK) break;
+    }
+    s->crc = crc32(s->crc, (const Bytef *)buf, len);
+
+    return (int)(len - s->stream.avail_in);
+}
+
+
+/* ===========================================================================
+     Converts, formats, and writes the args to the compressed file under
+   control of the format string, as in fprintf. gzprintf returns the number of
+   uncompressed bytes actually written (0 in case of error).
+*/
+#ifdef STDC
+#include <stdarg.h>
+
+int ZEXPORTVA gzprintf (gzFile file, const char *format, /* args */ ...)
+{
+    char buf[Z_PRINTF_BUFSIZE];
+    va_list va;
+    int len;
+
+    buf[sizeof(buf) - 1] = 0;
+    va_start(va, format);
+#ifdef NO_vsnprintf
+#  ifdef HAS_vsprintf_void
+    (void)vsprintf(buf, format, va);
+    va_end(va);
+    for (len = 0; len < sizeof(buf); len++)
+        if (buf[len] == 0) break;
+#  else
+    len = vsprintf(buf, format, va);
+    va_end(va);
+#  endif
+#else
+#  ifdef HAS_vsnprintf_void
+    (void)vsnprintf(buf, sizeof(buf), format, va);
+    va_end(va);
+    len = strlen(buf);
+#  else
+    len = vsnprintf(buf, sizeof(buf), format, va);
+    va_end(va);
+#  endif
+#endif
+    if (len <= 0 || len >= (int)sizeof(buf) || buf[sizeof(buf) - 1] != 0)
+        return 0;
+    return gzwrite(file, buf, (unsigned)len);
+}
+#else /* not ANSI C */
+
+int ZEXPORTVA gzprintf (file, format, a1, a2, a3, a4, a5, a6, a7, a8, a9, a10,
+                       a11, a12, a13, a14, a15, a16, a17, a18, a19, a20)
+    gzFile file;
+    const char *format;
+    int a1, a2, a3, a4, a5, a6, a7, a8, a9, a10,
+        a11, a12, a13, a14, a15, a16, a17, a18, a19, a20;
+{
+    char buf[Z_PRINTF_BUFSIZE];
+    int len;
+
+    buf[sizeof(buf) - 1] = 0;
+#ifdef NO_snprintf
+#  ifdef HAS_sprintf_void
+    sprintf(buf, format, a1, a2, a3, a4, a5, a6, a7, a8,
+            a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20);
+    for (len = 0; len < sizeof(buf); len++)
+        if (buf[len] == 0) break;
+#  else
+    len = sprintf(buf, format, a1, a2, a3, a4, a5, a6, a7, a8,
+                a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20);
+#  endif
+#else
+#  ifdef HAS_snprintf_void
+    snprintf(buf, sizeof(buf), format, a1, a2, a3, a4, a5, a6, a7, a8,
+             a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20);
+    len = strlen(buf);
+#  else
+    len = snprintf(buf, sizeof(buf), format, a1, a2, a3, a4, a5, a6, a7, a8,
+                 a9, a10, a11, a12, a13, a14, a15, a16, a17, a18, a19, a20);
+#  endif
+#endif
+    if (len <= 0 || len >= sizeof(buf) || buf[sizeof(buf) - 1] != 0)
+        return 0;
+    return gzwrite(file, buf, len);
+}
+#endif
+
+/* ===========================================================================
+      Writes c, converted to an unsigned char, into the compressed file.
+   gzputc returns the value that was written, or -1 in case of error.
+*/
+int ZEXPORT gzputc(file, c)
+    gzFile file;
+    int c;
+{
+    unsigned char cc = (unsigned char) c; /* required for big endian systems */
+
+    return gzwrite(file, &cc, 1) == 1 ? (int)cc : -1;
+}
+
+
+/* ===========================================================================
+      Writes the given null-terminated string to the compressed file, excluding
+   the terminating null character.
+      gzputs returns the number of characters written, or -1 in case of error.
+*/
+int ZEXPORT gzputs(file, s)
+    gzFile file;
+    const char *s;
+{
+    return gzwrite(file, (char*)s, (unsigned)strlen(s));
+}
+
+
+/* ===========================================================================
+     Flushes all pending output into the compressed file. The parameter
+   flush is as in the deflate() function.
+*/
+local int do_flush (file, flush)
+    gzFile file;
+    int flush;
+{
+    uInt len;
+    int done = 0;
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL || s->mode != 'w') return Z_STREAM_ERROR;
+
+    s->stream.avail_in = 0; /* should be zero already anyway */
+
+    for (;;) {
+        len = Z_BUFSIZE - s->stream.avail_out;
+
+        if (len != 0) {
+            if ((uInt)fwrite(s->outbuf, 1, len, s->file) != len) {
+                s->z_err = Z_ERRNO;
+                return Z_ERRNO;
+            }
+            s->stream.next_out = s->outbuf;
+            s->stream.avail_out = Z_BUFSIZE;
+        }
+        if (done) break;
+        s->out += s->stream.avail_out;
+        s->z_err = deflate(&(s->stream), flush);
+        s->out -= s->stream.avail_out;
+
+        /* Ignore the second of two consecutive flushes: */
+        if (len == 0 && s->z_err == Z_BUF_ERROR) s->z_err = Z_OK;
+
+        /* deflate has finished flushing only when it hasn't used up
+         * all the available space in the output buffer:
+         */
+        done = (s->stream.avail_out != 0 || s->z_err == Z_STREAM_END);
+
+        if (s->z_err != Z_OK && s->z_err != Z_STREAM_END) break;
+    }
+    return  s->z_err == Z_STREAM_END ? Z_OK : s->z_err;
+}
+
+int ZEXPORT gzflush (file, flush)
+     gzFile file;
+     int flush;
+{
+    gz_stream *s = (gz_stream*)file;
+    int err = do_flush (file, flush);
+
+    if (err) return err;
+    fflush(s->file);
+    return  s->z_err == Z_STREAM_END ? Z_OK : s->z_err;
+}
+#endif /* NO_GZCOMPRESS */
+
+/* ===========================================================================
+      Sets the starting position for the next gzread or gzwrite on the given
+   compressed file. The offset represents a number of bytes in the
+      gzseek returns the resulting offset location as measured in bytes from
+   the beginning of the uncompressed stream, or -1 in case of error.
+      SEEK_END is not implemented, returns error.
+      In this version of the library, gzseek can be extremely slow.
+*/
+z_off_t ZEXPORT gzseek (file, offset, whence)
+    gzFile file;
+    z_off_t offset;
+    int whence;
+{
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL || whence == SEEK_END ||
+        s->z_err == Z_ERRNO || s->z_err == Z_DATA_ERROR) {
+        return -1L;
+    }
+
+    if (s->mode == 'w') {
+#ifdef NO_GZCOMPRESS
+        return -1L;
+#else
+        if (whence == SEEK_SET) {
+            offset -= s->in;
+        }
+        if (offset < 0) return -1L;
+
+        /* At this point, offset is the number of zero bytes to write. */
+        if (s->inbuf == Z_NULL) {
+            s->inbuf = (Byte*)ALLOC(Z_BUFSIZE); /* for seeking */
+            if (s->inbuf == Z_NULL) return -1L;
+            zmemzero(s->inbuf, Z_BUFSIZE);
+        }
+        while (offset > 0)  {
+            uInt size = Z_BUFSIZE;
+            if (offset < Z_BUFSIZE) size = (uInt)offset;
+
+            size = gzwrite(file, s->inbuf, size);
+            if (size == 0) return -1L;
+
+            offset -= size;
+        }
+        return s->in;
+#endif
+    }
+    /* Rest of function is for reading only */
+
+    /* compute absolute position */
+    if (whence == SEEK_CUR) {
+        offset += s->out;
+    }
+    if (offset < 0) return -1L;
+
+    if (s->transparent) {
+        /* map to fseek */
+        s->back = EOF;
+        s->stream.avail_in = 0;
+        s->stream.next_in = s->inbuf;
+        if (fseek(s->file, offset, SEEK_SET) < 0) return -1L;
+
+        s->in = s->out = offset;
+        return offset;
+    }
+
+    /* For a negative seek, rewind and use positive seek */
+    if (offset >= s->out) {
+        offset -= s->out;
+    } else if (gzrewind(file) < 0) {
+        return -1L;
+    }
+    /* offset is now the number of bytes to skip. */
+
+    if (offset != 0 && s->outbuf == Z_NULL) {
+        s->outbuf = (Byte*)ALLOC(Z_BUFSIZE);
+        if (s->outbuf == Z_NULL) return -1L;
+    }
+    if (offset && s->back != EOF) {
+        s->back = EOF;
+        s->out++;
+        offset--;
+        if (s->last) s->z_err = Z_STREAM_END;
+    }
+    while (offset > 0)  {
+        int size = Z_BUFSIZE;
+        if (offset < Z_BUFSIZE) size = (int)offset;
+
+        size = gzread(file, s->outbuf, (uInt)size);
+        if (size <= 0) return -1L;
+        offset -= size;
+    }
+    return s->out;
+}
+
+/* ===========================================================================
+     Rewinds input file.
+*/
+int ZEXPORT gzrewind (file)
+    gzFile file;
+{
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL || s->mode != 'r') return -1;
+
+    s->z_err = Z_OK;
+    s->z_eof = 0;
+    s->back = EOF;
+    s->stream.avail_in = 0;
+    s->stream.next_in = s->inbuf;
+    s->crc = crc32(0L, Z_NULL, 0);
+    if (!s->transparent) (void)inflateReset(&s->stream);
+    s->in = 0;
+    s->out = 0;
+    return fseek(s->file, s->start, SEEK_SET);
+}
+
+/* ===========================================================================
+     Returns the starting position for the next gzread or gzwrite on the
+   given compressed file. This position represents a number of bytes in the
+   uncompressed data stream.
+*/
+z_off_t ZEXPORT gztell (file)
+    gzFile file;
+{
+    return gzseek(file, 0L, SEEK_CUR);
+}
+
+/* ===========================================================================
+     Returns 1 when EOF has previously been detected reading the given
+   input stream, otherwise zero.
+*/
+int ZEXPORT gzeof (file)
+    gzFile file;
+{
+    gz_stream *s = (gz_stream*)file;
+
+    /* With concatenated compressed files that can have embedded
+     * crc trailers, z_eof is no longer the only/best indicator of EOF
+     * on a gz_stream. Handle end-of-stream error explicitly here.
+     */
+    if (s == NULL || s->mode != 'r') return 0;
+    if (s->z_eof) return 1;
+    return s->z_err == Z_STREAM_END;
+}
+
+/* ===========================================================================
+     Returns 1 if reading and doing so transparently, otherwise zero.
+*/
+int ZEXPORT gzdirect (file)
+    gzFile file;
+{
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL || s->mode != 'r') return 0;
+    return s->transparent;
+}
+
+/* ===========================================================================
+   Outputs a long in LSB order to the given file
+*/
+local void putLong (file, x)
+    FILE *file;
+    uLong x;
+{
+    int n;
+    for (n = 0; n < 4; n++) {
+        fputc((int)(x & 0xff), file);
+        x >>= 8;
+    }
+}
+
+/* ===========================================================================
+   Reads a long in LSB order from the given gz_stream. Sets z_err in case
+   of error.
+*/
+local uLong getLong (s)
+    gz_stream *s;
+{
+    uLong x = (uLong)get_byte(s);
+    int c;
+
+    x += ((uLong)get_byte(s))<<8;
+    x += ((uLong)get_byte(s))<<16;
+    c = get_byte(s);
+    if (c == EOF) s->z_err = Z_DATA_ERROR;
+    x += ((uLong)c)<<24;
+    return x;
+}
+
+/* ===========================================================================
+     Flushes all pending output if necessary, closes the compressed file
+   and deallocates all the (de)compression state.
+*/
+int ZEXPORT gzclose (file)
+    gzFile file;
+{
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL) return Z_STREAM_ERROR;
+
+    if (s->mode == 'w') {
+#ifdef NO_GZCOMPRESS
+        return Z_STREAM_ERROR;
+#else
+        if (do_flush (file, Z_FINISH) != Z_OK)
+            return destroy((gz_stream*)file);
+
+        putLong (s->file, s->crc);
+        putLong (s->file, (uLong)(s->in & 0xffffffff));
+#endif
+    }
+    return destroy((gz_stream*)file);
+}
+
+#ifdef STDC
+#  define zstrerror(errnum) strerror(errnum)
+#else
+#  define zstrerror(errnum) ""
+#endif
+
+/* ===========================================================================
+     Returns the error message for the last error which occurred on the
+   given compressed file. errnum is set to zlib error number. If an
+   error occurred in the file system and not in the compression library,
+   errnum is set to Z_ERRNO and the application may consult errno
+   to get the exact error code.
+*/
+const char * ZEXPORT gzerror (file, errnum)
+    gzFile file;
+    int *errnum;
+{
+    char *m;
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL) {
+        *errnum = Z_STREAM_ERROR;
+        return (const char*)ERR_MSG(Z_STREAM_ERROR);
+    }
+    *errnum = s->z_err;
+    if (*errnum == Z_OK) return (const char*)"";
+
+    m = (char*)(*errnum == Z_ERRNO ? zstrerror(errno) : s->stream.msg);
+
+    if (m == NULL || *m == '\0') m = (char*)ERR_MSG(s->z_err);
+
+    TRYFREE(s->msg);
+    s->msg = (char*)ALLOC(strlen(s->path) + strlen(m) + 3);
+    if (s->msg == Z_NULL) return (const char*)ERR_MSG(Z_MEM_ERROR);
+    strcpy(s->msg, s->path);
+    strcat(s->msg, ": ");
+    strcat(s->msg, m);
+    return (const char*)s->msg;
+}
+
+/* ===========================================================================
+     Clear the error and end-of-file flags, and do the same for the real file.
+*/
+void ZEXPORT gzclearerr (file)
+    gzFile file;
+{
+    gz_stream *s = (gz_stream*)file;
+
+    if (s == NULL) return;
+    if (s->z_err != Z_STREAM_END) s->z_err = Z_OK;
+    s->z_eof = 0;
+    clearerr(s->file);
+}
diff --git a/distrib/zlib-1.2.3/infback.c b/distrib/zlib-1.2.3/infback.c
new file mode 100644
index 0000000..455dbc9
--- /dev/null
+++ b/distrib/zlib-1.2.3/infback.c
@@ -0,0 +1,623 @@
+/* infback.c -- inflate using a call-back interface
+ * Copyright (C) 1995-2005 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/*
+   This code is largely copied from inflate.c.  Normally either infback.o or
+   inflate.o would be linked into an application--not both.  The interface
+   with inffast.c is retained so that optimized assembler-coded versions of
+   inflate_fast() can be used with either inflate.c or infback.c.
+ */
+
+#include "zutil.h"
+#include "inftrees.h"
+#include "inflate.h"
+#include "inffast.h"
+
+/* function prototypes */
+local void fixedtables OF((struct inflate_state FAR *state));
+
+/*
+   strm provides memory allocation functions in zalloc and zfree, or
+   Z_NULL to use the library memory allocation functions.
+
+   windowBits is in the range 8..15, and window is a user-supplied
+   window and output buffer that is 2**windowBits bytes.
+ */
+int ZEXPORT inflateBackInit_(strm, windowBits, window, version, stream_size)
+z_streamp strm;
+int windowBits;
+unsigned char FAR *window;
+const char *version;
+int stream_size;
+{
+    struct inflate_state FAR *state;
+
+    if (version == Z_NULL || version[0] != ZLIB_VERSION[0] ||
+        stream_size != (int)(sizeof(z_stream)))
+        return Z_VERSION_ERROR;
+    if (strm == Z_NULL || window == Z_NULL ||
+        windowBits < 8 || windowBits > 15)
+        return Z_STREAM_ERROR;
+    strm->msg = Z_NULL;                 /* in case we return an error */
+    if (strm->zalloc == (alloc_func)0) {
+        strm->zalloc = zcalloc;
+        strm->opaque = (voidpf)0;
+    }
+    if (strm->zfree == (free_func)0) strm->zfree = zcfree;
+    state = (struct inflate_state FAR *)ZALLOC(strm, 1,
+                                               sizeof(struct inflate_state));
+    if (state == Z_NULL) return Z_MEM_ERROR;
+    Tracev((stderr, "inflate: allocated\n"));
+    strm->state = (struct internal_state FAR *)state;
+    state->dmax = 32768U;
+    state->wbits = windowBits;
+    state->wsize = 1U << windowBits;
+    state->window = window;
+    state->write = 0;
+    state->whave = 0;
+    return Z_OK;
+}
+
+/*
+   Return state with length and distance decoding tables and index sizes set to
+   fixed code decoding.  Normally this returns fixed tables from inffixed.h.
+   If BUILDFIXED is defined, then instead this routine builds the tables the
+   first time it's called, and returns those tables the first time and
+   thereafter.  This reduces the size of the code by about 2K bytes, in
+   exchange for a little execution time.  However, BUILDFIXED should not be
+   used for threaded applications, since the rewriting of the tables and virgin
+   may not be thread-safe.
+ */
+local void fixedtables(state)
+struct inflate_state FAR *state;
+{
+#ifdef BUILDFIXED
+    static int virgin = 1;
+    static code *lenfix, *distfix;
+    static code fixed[544];
+
+    /* build fixed huffman tables if first call (may not be thread safe) */
+    if (virgin) {
+        unsigned sym, bits;
+        static code *next;
+
+        /* literal/length table */
+        sym = 0;
+        while (sym < 144) state->lens[sym++] = 8;
+        while (sym < 256) state->lens[sym++] = 9;
+        while (sym < 280) state->lens[sym++] = 7;
+        while (sym < 288) state->lens[sym++] = 8;
+        next = fixed;
+        lenfix = next;
+        bits = 9;
+        inflate_table(LENS, state->lens, 288, &(next), &(bits), state->work);
+
+        /* distance table */
+        sym = 0;
+        while (sym < 32) state->lens[sym++] = 5;
+        distfix = next;
+        bits = 5;
+        inflate_table(DISTS, state->lens, 32, &(next), &(bits), state->work);
+
+        /* do this just once */
+        virgin = 0;
+    }
+#else /* !BUILDFIXED */
+#   include "inffixed.h"
+#endif /* BUILDFIXED */
+    state->lencode = lenfix;
+    state->lenbits = 9;
+    state->distcode = distfix;
+    state->distbits = 5;
+}
+
+/* Macros for inflateBack(): */
+
+/* Load returned state from inflate_fast() */
+#define LOAD() \
+    do { \
+        put = strm->next_out; \
+        left = strm->avail_out; \
+        next = strm->next_in; \
+        have = strm->avail_in; \
+        hold = state->hold; \
+        bits = state->bits; \
+    } while (0)
+
+/* Set state from registers for inflate_fast() */
+#define RESTORE() \
+    do { \
+        strm->next_out = put; \
+        strm->avail_out = left; \
+        strm->next_in = next; \
+        strm->avail_in = have; \
+        state->hold = hold; \
+        state->bits = bits; \
+    } while (0)
+
+/* Clear the input bit accumulator */
+#define INITBITS() \
+    do { \
+        hold = 0; \
+        bits = 0; \
+    } while (0)
+
+/* Assure that some input is available.  If input is requested, but denied,
+   then return a Z_BUF_ERROR from inflateBack(). */
+#define PULL() \
+    do { \
+        if (have == 0) { \
+            have = in(in_desc, &next); \
+            if (have == 0) { \
+                next = Z_NULL; \
+                ret = Z_BUF_ERROR; \
+                goto inf_leave; \
+            } \
+        } \
+    } while (0)
+
+/* Get a byte of input into the bit accumulator, or return from inflateBack()
+   with an error if there is no input available. */
+#define PULLBYTE() \
+    do { \
+        PULL(); \
+        have--; \
+        hold += (unsigned long)(*next++) << bits; \
+        bits += 8; \
+    } while (0)
+
+/* Assure that there are at least n bits in the bit accumulator.  If there is
+   not enough available input to do that, then return from inflateBack() with
+   an error. */
+#define NEEDBITS(n) \
+    do { \
+        while (bits < (unsigned)(n)) \
+            PULLBYTE(); \
+    } while (0)
+
+/* Return the low n bits of the bit accumulator (n < 16) */
+#define BITS(n) \
+    ((unsigned)hold & ((1U << (n)) - 1))
+
+/* Remove n bits from the bit accumulator */
+#define DROPBITS(n) \
+    do { \
+        hold >>= (n); \
+        bits -= (unsigned)(n); \
+    } while (0)
+
+/* Remove zero to seven bits as needed to go to a byte boundary */
+#define BYTEBITS() \
+    do { \
+        hold >>= bits & 7; \
+        bits -= bits & 7; \
+    } while (0)
+
+/* Assure that some output space is available, by writing out the window
+   if it's full.  If the write fails, return from inflateBack() with a
+   Z_BUF_ERROR. */
+#define ROOM() \
+    do { \
+        if (left == 0) { \
+            put = state->window; \
+            left = state->wsize; \
+            state->whave = left; \
+            if (out(out_desc, put, left)) { \
+                ret = Z_BUF_ERROR; \
+                goto inf_leave; \
+            } \
+        } \
+    } while (0)
+
+/*
+   strm provides the memory allocation functions and window buffer on input,
+   and provides information on the unused input on return.  For Z_DATA_ERROR
+   returns, strm will also provide an error message.
+
+   in() and out() are the call-back input and output functions.  When
+   inflateBack() needs more input, it calls in().  When inflateBack() has
+   filled the window with output, or when it completes with data in the
+   window, it calls out() to write out the data.  The application must not
+   change the provided input until in() is called again or inflateBack()
+   returns.  The application must not change the window/output buffer until
+   inflateBack() returns.
+
+   in() and out() are called with a descriptor parameter provided in the
+   inflateBack() call.  This parameter can be a structure that provides the
+   information required to do the read or write, as well as accumulated
+   information on the input and output such as totals and check values.
+
+   in() should return zero on failure.  out() should return non-zero on
+   failure.  If either in() or out() fails, than inflateBack() returns a
+   Z_BUF_ERROR.  strm->next_in can be checked for Z_NULL to see whether it
+   was in() or out() that caused in the error.  Otherwise,  inflateBack()
+   returns Z_STREAM_END on success, Z_DATA_ERROR for an deflate format
+   error, or Z_MEM_ERROR if it could not allocate memory for the state.
+   inflateBack() can also return Z_STREAM_ERROR if the input parameters
+   are not correct, i.e. strm is Z_NULL or the state was not initialized.
+ */
+int ZEXPORT inflateBack(strm, in, in_desc, out, out_desc)
+z_streamp strm;
+in_func in;
+void FAR *in_desc;
+out_func out;
+void FAR *out_desc;
+{
+    struct inflate_state FAR *state;
+    unsigned char FAR *next;    /* next input */
+    unsigned char FAR *put;     /* next output */
+    unsigned have, left;        /* available input and output */
+    unsigned long hold;         /* bit buffer */
+    unsigned bits;              /* bits in bit buffer */
+    unsigned copy;              /* number of stored or match bytes to copy */
+    unsigned char FAR *from;    /* where to copy match bytes from */
+    code this;                  /* current decoding table entry */
+    code last;                  /* parent table entry */
+    unsigned len;               /* length to copy for repeats, bits to drop */
+    int ret;                    /* return code */
+    static const unsigned short order[19] = /* permutation of code lengths */
+        {16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15};
+
+    /* Check that the strm exists and that the state was initialized */
+    if (strm == Z_NULL || strm->state == Z_NULL)
+        return Z_STREAM_ERROR;
+    state = (struct inflate_state FAR *)strm->state;
+
+    /* Reset the state */
+    strm->msg = Z_NULL;
+    state->mode = TYPE;
+    state->last = 0;
+    state->whave = 0;
+    next = strm->next_in;
+    have = next != Z_NULL ? strm->avail_in : 0;
+    hold = 0;
+    bits = 0;
+    put = state->window;
+    left = state->wsize;
+
+    /* Inflate until end of block marked as last */
+    for (;;)
+        switch (state->mode) {
+        case TYPE:
+            /* determine and dispatch block type */
+            if (state->last) {
+                BYTEBITS();
+                state->mode = DONE;
+                break;
+            }
+            NEEDBITS(3);
+            state->last = BITS(1);
+            DROPBITS(1);
+            switch (BITS(2)) {
+            case 0:                             /* stored block */
+                Tracev((stderr, "inflate:     stored block%s\n",
+                        state->last ? " (last)" : ""));
+                state->mode = STORED;
+                break;
+            case 1:                             /* fixed block */
+                fixedtables(state);
+                Tracev((stderr, "inflate:     fixed codes block%s\n",
+                        state->last ? " (last)" : ""));
+                state->mode = LEN;              /* decode codes */
+                break;
+            case 2:                             /* dynamic block */
+                Tracev((stderr, "inflate:     dynamic codes block%s\n",
+                        state->last ? " (last)" : ""));
+                state->mode = TABLE;
+                break;
+            case 3:
+                strm->msg = (char *)"invalid block type";
+                state->mode = BAD;
+            }
+            DROPBITS(2);
+            break;
+
+        case STORED:
+            /* get and verify stored block length */
+            BYTEBITS();                         /* go to byte boundary */
+            NEEDBITS(32);
+            if ((hold & 0xffff) != ((hold >> 16) ^ 0xffff)) {
+                strm->msg = (char *)"invalid stored block lengths";
+                state->mode = BAD;
+                break;
+            }
+            state->length = (unsigned)hold & 0xffff;
+            Tracev((stderr, "inflate:       stored length %u\n",
+                    state->length));
+            INITBITS();
+
+            /* copy stored block from input to output */
+            while (state->length != 0) {
+                copy = state->length;
+                PULL();
+                ROOM();
+                if (copy > have) copy = have;
+                if (copy > left) copy = left;
+                zmemcpy(put, next, copy);
+                have -= copy;
+                next += copy;
+                left -= copy;
+                put += copy;
+                state->length -= copy;
+            }
+            Tracev((stderr, "inflate:       stored end\n"));
+            state->mode = TYPE;
+            break;
+
+        case TABLE:
+            /* get dynamic table entries descriptor */
+            NEEDBITS(14);
+            state->nlen = BITS(5) + 257;
+            DROPBITS(5);
+            state->ndist = BITS(5) + 1;
+            DROPBITS(5);
+            state->ncode = BITS(4) + 4;
+            DROPBITS(4);
+#ifndef PKZIP_BUG_WORKAROUND
+            if (state->nlen > 286 || state->ndist > 30) {
+                strm->msg = (char *)"too many length or distance symbols";
+                state->mode = BAD;
+                break;
+            }
+#endif
+            Tracev((stderr, "inflate:       table sizes ok\n"));
+
+            /* get code length code lengths (not a typo) */
+            state->have = 0;
+            while (state->have < state->ncode) {
+                NEEDBITS(3);
+                state->lens[order[state->have++]] = (unsigned short)BITS(3);
+                DROPBITS(3);
+            }
+            while (state->have < 19)
+                state->lens[order[state->have++]] = 0;
+            state->next = state->codes;
+            state->lencode = (code const FAR *)(state->next);
+            state->lenbits = 7;
+            ret = inflate_table(CODES, state->lens, 19, &(state->next),
+                                &(state->lenbits), state->work);
+            if (ret) {
+                strm->msg = (char *)"invalid code lengths set";
+                state->mode = BAD;
+                break;
+            }
+            Tracev((stderr, "inflate:       code lengths ok\n"));
+
+            /* get length and distance code code lengths */
+            state->have = 0;
+            while (state->have < state->nlen + state->ndist) {
+                for (;;) {
+                    this = state->lencode[BITS(state->lenbits)];
+                    if ((unsigned)(this.bits) <= bits) break;
+                    PULLBYTE();
+                }
+                if (this.val < 16) {
+                    NEEDBITS(this.bits);
+                    DROPBITS(this.bits);
+                    state->lens[state->have++] = this.val;
+                }
+                else {
+                    if (this.val == 16) {
+                        NEEDBITS(this.bits + 2);
+                        DROPBITS(this.bits);
+                        if (state->have == 0) {
+                            strm->msg = (char *)"invalid bit length repeat";
+                            state->mode = BAD;
+                            break;
+                        }
+                        len = (unsigned)(state->lens[state->have - 1]);
+                        copy = 3 + BITS(2);
+                        DROPBITS(2);
+                    }
+                    else if (this.val == 17) {
+                        NEEDBITS(this.bits + 3);
+                        DROPBITS(this.bits);
+                        len = 0;
+                        copy = 3 + BITS(3);
+                        DROPBITS(3);
+                    }
+                    else {
+                        NEEDBITS(this.bits + 7);
+                        DROPBITS(this.bits);
+                        len = 0;
+                        copy = 11 + BITS(7);
+                        DROPBITS(7);
+                    }
+                    if (state->have + copy > state->nlen + state->ndist) {
+                        strm->msg = (char *)"invalid bit length repeat";
+                        state->mode = BAD;
+                        break;
+                    }
+                    while (copy--)
+                        state->lens[state->have++] = (unsigned short)len;
+                }
+            }
+
+            /* handle error breaks in while */
+            if (state->mode == BAD) break;
+
+            /* build code tables */
+            state->next = state->codes;
+            state->lencode = (code const FAR *)(state->next);
+            state->lenbits = 9;
+            ret = inflate_table(LENS, state->lens, state->nlen, &(state->next),
+                                &(state->lenbits), state->work);
+            if (ret) {
+                strm->msg = (char *)"invalid literal/lengths set";
+                state->mode = BAD;
+                break;
+            }
+            state->distcode = (code const FAR *)(state->next);
+            state->distbits = 6;
+            ret = inflate_table(DISTS, state->lens + state->nlen, state->ndist,
+                            &(state->next), &(state->distbits), state->work);
+            if (ret) {
+                strm->msg = (char *)"invalid distances set";
+                state->mode = BAD;
+                break;
+            }
+            Tracev((stderr, "inflate:       codes ok\n"));
+            state->mode = LEN;
+
+        case LEN:
+            /* use inflate_fast() if we have enough input and output */
+            if (have >= 6 && left >= 258) {
+                RESTORE();
+                if (state->whave < state->wsize)
+                    state->whave = state->wsize - left;
+                inflate_fast(strm, state->wsize);
+                LOAD();
+                break;
+            }
+
+            /* get a literal, length, or end-of-block code */
+            for (;;) {
+                this = state->lencode[BITS(state->lenbits)];
+                if ((unsigned)(this.bits) <= bits) break;
+                PULLBYTE();
+            }
+            if (this.op && (this.op & 0xf0) == 0) {
+                last = this;
+                for (;;) {
+                    this = state->lencode[last.val +
+                            (BITS(last.bits + last.op) >> last.bits)];
+                    if ((unsigned)(last.bits + this.bits) <= bits) break;
+                    PULLBYTE();
+                }
+                DROPBITS(last.bits);
+            }
+            DROPBITS(this.bits);
+            state->length = (unsigned)this.val;
+
+            /* process literal */
+            if (this.op == 0) {
+                Tracevv((stderr, this.val >= 0x20 && this.val < 0x7f ?
+                        "inflate:         literal '%c'\n" :
+                        "inflate:         literal 0x%02x\n", this.val));
+                ROOM();
+                *put++ = (unsigned char)(state->length);
+                left--;
+                state->mode = LEN;
+                break;
+            }
+
+            /* process end of block */
+            if (this.op & 32) {
+                Tracevv((stderr, "inflate:         end of block\n"));
+                state->mode = TYPE;
+                break;
+            }
+
+            /* invalid code */
+            if (this.op & 64) {
+                strm->msg = (char *)"invalid literal/length code";
+                state->mode = BAD;
+                break;
+            }
+
+            /* length code -- get extra bits, if any */
+            state->extra = (unsigned)(this.op) & 15;
+            if (state->extra != 0) {
+                NEEDBITS(state->extra);
+                state->length += BITS(state->extra);
+                DROPBITS(state->extra);
+            }
+            Tracevv((stderr, "inflate:         length %u\n", state->length));
+
+            /* get distance code */
+            for (;;) {
+                this = state->distcode[BITS(state->distbits)];
+                if ((unsigned)(this.bits) <= bits) break;
+                PULLBYTE();
+            }
+            if ((this.op & 0xf0) == 0) {
+                last = this;
+                for (;;) {
+                    this = state->distcode[last.val +
+                            (BITS(last.bits + last.op) >> last.bits)];
+                    if ((unsigned)(last.bits + this.bits) <= bits) break;
+                    PULLBYTE();
+                }
+                DROPBITS(last.bits);
+            }
+            DROPBITS(this.bits);
+            if (this.op & 64) {
+                strm->msg = (char *)"invalid distance code";
+                state->mode = BAD;
+                break;
+            }
+            state->offset = (unsigned)this.val;
+
+            /* get distance extra bits, if any */
+            state->extra = (unsigned)(this.op) & 15;
+            if (state->extra != 0) {
+                NEEDBITS(state->extra);
+                state->offset += BITS(state->extra);
+                DROPBITS(state->extra);
+            }
+            if (state->offset > state->wsize - (state->whave < state->wsize ?
+                                                left : 0)) {
+                strm->msg = (char *)"invalid distance too far back";
+                state->mode = BAD;
+                break;
+            }
+            Tracevv((stderr, "inflate:         distance %u\n", state->offset));
+
+            /* copy match from window to output */
+            do {
+                ROOM();
+                copy = state->wsize - state->offset;
+                if (copy < left) {
+                    from = put + copy;
+                    copy = left - copy;
+                }
+                else {
+                    from = put - state->offset;
+                    copy = left;
+                }
+                if (copy > state->length) copy = state->length;
+                state->length -= copy;
+                left -= copy;
+                do {
+                    *put++ = *from++;
+                } while (--copy);
+            } while (state->length != 0);
+            break;
+
+        case DONE:
+            /* inflate stream terminated properly -- write leftover output */
+            ret = Z_STREAM_END;
+            if (left < state->wsize) {
+                if (out(out_desc, state->window, state->wsize - left))
+                    ret = Z_BUF_ERROR;
+            }
+            goto inf_leave;
+
+        case BAD:
+            ret = Z_DATA_ERROR;
+            goto inf_leave;
+
+        default:                /* can't happen, but makes compilers happy */
+            ret = Z_STREAM_ERROR;
+            goto inf_leave;
+        }
+
+    /* Return unused input */
+  inf_leave:
+    strm->next_in = next;
+    strm->avail_in = have;
+    return ret;
+}
+
+int ZEXPORT inflateBackEnd(strm)
+z_streamp strm;
+{
+    if (strm == Z_NULL || strm->state == Z_NULL || strm->zfree == (free_func)0)
+        return Z_STREAM_ERROR;
+    ZFREE(strm, strm->state);
+    strm->state = Z_NULL;
+    Tracev((stderr, "inflate: end\n"));
+    return Z_OK;
+}
diff --git a/distrib/zlib-1.2.3/inffast.c b/distrib/zlib-1.2.3/inffast.c
new file mode 100644
index 0000000..bbee92e
--- /dev/null
+++ b/distrib/zlib-1.2.3/inffast.c
@@ -0,0 +1,318 @@
+/* inffast.c -- fast decoding
+ * Copyright (C) 1995-2004 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+#include "zutil.h"
+#include "inftrees.h"
+#include "inflate.h"
+#include "inffast.h"
+
+#ifndef ASMINF
+
+/* Allow machine dependent optimization for post-increment or pre-increment.
+   Based on testing to date,
+   Pre-increment preferred for:
+   - PowerPC G3 (Adler)
+   - MIPS R5000 (Randers-Pehrson)
+   Post-increment preferred for:
+   - none
+   No measurable difference:
+   - Pentium III (Anderson)
+   - M68060 (Nikl)
+ */
+#ifdef POSTINC
+#  define OFF 0
+#  define PUP(a) *(a)++
+#else
+#  define OFF 1
+#  define PUP(a) *++(a)
+#endif
+
+/*
+   Decode literal, length, and distance codes and write out the resulting
+   literal and match bytes until either not enough input or output is
+   available, an end-of-block is encountered, or a data error is encountered.
+   When large enough input and output buffers are supplied to inflate(), for
+   example, a 16K input buffer and a 64K output buffer, more than 95% of the
+   inflate execution time is spent in this routine.
+
+   Entry assumptions:
+
+        state->mode == LEN
+        strm->avail_in >= 6
+        strm->avail_out >= 258
+        start >= strm->avail_out
+        state->bits < 8
+
+   On return, state->mode is one of:
+
+        LEN -- ran out of enough output space or enough available input
+        TYPE -- reached end of block code, inflate() to interpret next block
+        BAD -- error in block data
+
+   Notes:
+
+    - The maximum input bits used by a length/distance pair is 15 bits for the
+      length code, 5 bits for the length extra, 15 bits for the distance code,
+      and 13 bits for the distance extra.  This totals 48 bits, or six bytes.
+      Therefore if strm->avail_in >= 6, then there is enough input to avoid
+      checking for available input while decoding.
+
+    - The maximum bytes that a single length/distance pair can output is 258
+      bytes, which is the maximum length that can be coded.  inflate_fast()
+      requires strm->avail_out >= 258 for each loop to avoid checking for
+      output space.
+ */
+void inflate_fast(strm, start)
+z_streamp strm;
+unsigned start;         /* inflate()'s starting value for strm->avail_out */
+{
+    struct inflate_state FAR *state;
+    unsigned char FAR *in;      /* local strm->next_in */
+    unsigned char FAR *last;    /* while in < last, enough input available */
+    unsigned char FAR *out;     /* local strm->next_out */
+    unsigned char FAR *beg;     /* inflate()'s initial strm->next_out */
+    unsigned char FAR *end;     /* while out < end, enough space available */
+#ifdef INFLATE_STRICT
+    unsigned dmax;              /* maximum distance from zlib header */
+#endif
+    unsigned wsize;             /* window size or zero if not using window */
+    unsigned whave;             /* valid bytes in the window */
+    unsigned write;             /* window write index */
+    unsigned char FAR *window;  /* allocated sliding window, if wsize != 0 */
+    unsigned long hold;         /* local strm->hold */
+    unsigned bits;              /* local strm->bits */
+    code const FAR *lcode;      /* local strm->lencode */
+    code const FAR *dcode;      /* local strm->distcode */
+    unsigned lmask;             /* mask for first level of length codes */
+    unsigned dmask;             /* mask for first level of distance codes */
+    code this;                  /* retrieved table entry */
+    unsigned op;                /* code bits, operation, extra bits, or */
+                                /*  window position, window bytes to copy */
+    unsigned len;               /* match length, unused bytes */
+    unsigned dist;              /* match distance */
+    unsigned char FAR *from;    /* where to copy match from */
+
+    /* copy state to local variables */
+    state = (struct inflate_state FAR *)strm->state;
+    in = strm->next_in - OFF;
+    last = in + (strm->avail_in - 5);
+    out = strm->next_out - OFF;
+    beg = out - (start - strm->avail_out);
+    end = out + (strm->avail_out - 257);
+#ifdef INFLATE_STRICT
+    dmax = state->dmax;
+#endif
+    wsize = state->wsize;
+    whave = state->whave;
+    write = state->write;
+    window = state->window;
+    hold = state->hold;
+    bits = state->bits;
+    lcode = state->lencode;
+    dcode = state->distcode;
+    lmask = (1U << state->lenbits) - 1;
+    dmask = (1U << state->distbits) - 1;
+
+    /* decode literals and length/distances until end-of-block or not enough
+       input data or output space */
+    do {
+        if (bits < 15) {
+            hold += (unsigned long)(PUP(in)) << bits;
+            bits += 8;
+            hold += (unsigned long)(PUP(in)) << bits;
+            bits += 8;
+        }
+        this = lcode[hold & lmask];
+      dolen:
+        op = (unsigned)(this.bits);
+        hold >>= op;
+        bits -= op;
+        op = (unsigned)(this.op);
+        if (op == 0) {                          /* literal */
+            Tracevv((stderr, this.val >= 0x20 && this.val < 0x7f ?
+                    "inflate:         literal '%c'\n" :
+                    "inflate:         literal 0x%02x\n", this.val));
+            PUP(out) = (unsigned char)(this.val);
+        }
+        else if (op & 16) {                     /* length base */
+            len = (unsigned)(this.val);
+            op &= 15;                           /* number of extra bits */
+            if (op) {
+                if (bits < op) {
+                    hold += (unsigned long)(PUP(in)) << bits;
+                    bits += 8;
+                }
+                len += (unsigned)hold & ((1U << op) - 1);
+                hold >>= op;
+                bits -= op;
+            }
+            Tracevv((stderr, "inflate:         length %u\n", len));
+            if (bits < 15) {
+                hold += (unsigned long)(PUP(in)) << bits;
+                bits += 8;
+                hold += (unsigned long)(PUP(in)) << bits;
+                bits += 8;
+            }
+            this = dcode[hold & dmask];
+          dodist:
+            op = (unsigned)(this.bits);
+            hold >>= op;
+            bits -= op;
+            op = (unsigned)(this.op);
+            if (op & 16) {                      /* distance base */
+                dist = (unsigned)(this.val);
+                op &= 15;                       /* number of extra bits */
+                if (bits < op) {
+                    hold += (unsigned long)(PUP(in)) << bits;
+                    bits += 8;
+                    if (bits < op) {
+                        hold += (unsigned long)(PUP(in)) << bits;
+                        bits += 8;
+                    }
+                }
+                dist += (unsigned)hold & ((1U << op) - 1);
+#ifdef INFLATE_STRICT
+                if (dist > dmax) {
+                    strm->msg = (char *)"invalid distance too far back";
+                    state->mode = BAD;
+                    break;
+                }
+#endif
+                hold >>= op;
+                bits -= op;
+                Tracevv((stderr, "inflate:         distance %u\n", dist));
+                op = (unsigned)(out - beg);     /* max distance in output */
+                if (dist > op) {                /* see if copy from window */
+                    op = dist - op;             /* distance back in window */
+                    if (op > whave) {
+                        strm->msg = (char *)"invalid distance too far back";
+                        state->mode = BAD;
+                        break;
+                    }
+                    from = window - OFF;
+                    if (write == 0) {           /* very common case */
+                        from += wsize - op;
+                        if (op < len) {         /* some from window */
+                            len -= op;
+                            do {
+                                PUP(out) = PUP(from);
+                            } while (--op);
+                            from = out - dist;  /* rest from output */
+                        }
+                    }
+                    else if (write < op) {      /* wrap around window */
+                        from += wsize + write - op;
+                        op -= write;
+                        if (op < len) {         /* some from end of window */
+                            len -= op;
+                            do {
+                                PUP(out) = PUP(from);
+                            } while (--op);
+                            from = window - OFF;
+                            if (write < len) {  /* some from start of window */
+                                op = write;
+                                len -= op;
+                                do {
+                                    PUP(out) = PUP(from);
+                                } while (--op);
+                                from = out - dist;      /* rest from output */
+                            }
+                        }
+                    }
+                    else {                      /* contiguous in window */
+                        from += write - op;
+                        if (op < len) {         /* some from window */
+                            len -= op;
+                            do {
+                                PUP(out) = PUP(from);
+                            } while (--op);
+                            from = out - dist;  /* rest from output */
+                        }
+                    }
+                    while (len > 2) {
+                        PUP(out) = PUP(from);
+                        PUP(out) = PUP(from);
+                        PUP(out) = PUP(from);
+                        len -= 3;
+                    }
+                    if (len) {
+                        PUP(out) = PUP(from);
+                        if (len > 1)
+                            PUP(out) = PUP(from);
+                    }
+                }
+                else {
+                    from = out - dist;          /* copy direct from output */
+                    do {                        /* minimum length is three */
+                        PUP(out) = PUP(from);
+                        PUP(out) = PUP(from);
+                        PUP(out) = PUP(from);
+                        len -= 3;
+                    } while (len > 2);
+                    if (len) {
+                        PUP(out) = PUP(from);
+                        if (len > 1)
+                            PUP(out) = PUP(from);
+                    }
+                }
+            }
+            else if ((op & 64) == 0) {          /* 2nd level distance code */
+                this = dcode[this.val + (hold & ((1U << op) - 1))];
+                goto dodist;
+            }
+            else {
+                strm->msg = (char *)"invalid distance code";
+                state->mode = BAD;
+                break;
+            }
+        }
+        else if ((op & 64) == 0) {              /* 2nd level length code */
+            this = lcode[this.val + (hold & ((1U << op) - 1))];
+            goto dolen;
+        }
+        else if (op & 32) {                     /* end-of-block */
+            Tracevv((stderr, "inflate:         end of block\n"));
+            state->mode = TYPE;
+            break;
+        }
+        else {
+            strm->msg = (char *)"invalid literal/length code";
+            state->mode = BAD;
+            break;
+        }
+    } while (in < last && out < end);
+
+    /* return unused bytes (on entry, bits < 8, so in won't go too far back) */
+    len = bits >> 3;
+    in -= len;
+    bits -= len << 3;
+    hold &= (1U << bits) - 1;
+
+    /* update state and return */
+    strm->next_in = in + OFF;
+    strm->next_out = out + OFF;
+    strm->avail_in = (unsigned)(in < last ? 5 + (last - in) : 5 - (in - last));
+    strm->avail_out = (unsigned)(out < end ?
+                                 257 + (end - out) : 257 - (out - end));
+    state->hold = hold;
+    state->bits = bits;
+    return;
+}
+
+/*
+   inflate_fast() speedups that turned out slower (on a PowerPC G3 750CXe):
+   - Using bit fields for code structure
+   - Different op definition to avoid & for extra bits (do & for table bits)
+   - Three separate decoding do-loops for direct, window, and write == 0
+   - Special case for distance > 1 copies to do overlapped load and store copy
+   - Explicit branch predictions (based on measured branch probabilities)
+   - Deferring match copy and interspersed it with decoding subsequent codes
+   - Swapping literal/length else
+   - Swapping window/direct else
+   - Larger unrolled copy loops (three is about right)
+   - Moving len -= 3 statement into middle of loop
+ */
+
+#endif /* !ASMINF */
diff --git a/distrib/zlib-1.2.3/inffast.h b/distrib/zlib-1.2.3/inffast.h
new file mode 100644
index 0000000..1e88d2d
--- /dev/null
+++ b/distrib/zlib-1.2.3/inffast.h
@@ -0,0 +1,11 @@
+/* inffast.h -- header to use inffast.c
+ * Copyright (C) 1995-2003 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* WARNING: this file should *not* be used by applications. It is
+   part of the implementation of the compression library and is
+   subject to change. Applications should only use zlib.h.
+ */
+
+void inflate_fast OF((z_streamp strm, unsigned start));
diff --git a/distrib/zlib-1.2.3/inffixed.h b/distrib/zlib-1.2.3/inffixed.h
new file mode 100644
index 0000000..75ed4b5
--- /dev/null
+++ b/distrib/zlib-1.2.3/inffixed.h
@@ -0,0 +1,94 @@
+    /* inffixed.h -- table for decoding fixed codes
+     * Generated automatically by makefixed().
+     */
+
+    /* WARNING: this file should *not* be used by applications. It
+       is part of the implementation of the compression library and
+       is subject to change. Applications should only use zlib.h.
+     */
+
+    static const code lenfix[512] = {
+        {96,7,0},{0,8,80},{0,8,16},{20,8,115},{18,7,31},{0,8,112},{0,8,48},
+        {0,9,192},{16,7,10},{0,8,96},{0,8,32},{0,9,160},{0,8,0},{0,8,128},
+        {0,8,64},{0,9,224},{16,7,6},{0,8,88},{0,8,24},{0,9,144},{19,7,59},
+        {0,8,120},{0,8,56},{0,9,208},{17,7,17},{0,8,104},{0,8,40},{0,9,176},
+        {0,8,8},{0,8,136},{0,8,72},{0,9,240},{16,7,4},{0,8,84},{0,8,20},
+        {21,8,227},{19,7,43},{0,8,116},{0,8,52},{0,9,200},{17,7,13},{0,8,100},
+        {0,8,36},{0,9,168},{0,8,4},{0,8,132},{0,8,68},{0,9,232},{16,7,8},
+        {0,8,92},{0,8,28},{0,9,152},{20,7,83},{0,8,124},{0,8,60},{0,9,216},
+        {18,7,23},{0,8,108},{0,8,44},{0,9,184},{0,8,12},{0,8,140},{0,8,76},
+        {0,9,248},{16,7,3},{0,8,82},{0,8,18},{21,8,163},{19,7,35},{0,8,114},
+        {0,8,50},{0,9,196},{17,7,11},{0,8,98},{0,8,34},{0,9,164},{0,8,2},
+        {0,8,130},{0,8,66},{0,9,228},{16,7,7},{0,8,90},{0,8,26},{0,9,148},
+        {20,7,67},{0,8,122},{0,8,58},{0,9,212},{18,7,19},{0,8,106},{0,8,42},
+        {0,9,180},{0,8,10},{0,8,138},{0,8,74},{0,9,244},{16,7,5},{0,8,86},
+        {0,8,22},{64,8,0},{19,7,51},{0,8,118},{0,8,54},{0,9,204},{17,7,15},
+        {0,8,102},{0,8,38},{0,9,172},{0,8,6},{0,8,134},{0,8,70},{0,9,236},
+        {16,7,9},{0,8,94},{0,8,30},{0,9,156},{20,7,99},{0,8,126},{0,8,62},
+        {0,9,220},{18,7,27},{0,8,110},{0,8,46},{0,9,188},{0,8,14},{0,8,142},
+        {0,8,78},{0,9,252},{96,7,0},{0,8,81},{0,8,17},{21,8,131},{18,7,31},
+        {0,8,113},{0,8,49},{0,9,194},{16,7,10},{0,8,97},{0,8,33},{0,9,162},
+        {0,8,1},{0,8,129},{0,8,65},{0,9,226},{16,7,6},{0,8,89},{0,8,25},
+        {0,9,146},{19,7,59},{0,8,121},{0,8,57},{0,9,210},{17,7,17},{0,8,105},
+        {0,8,41},{0,9,178},{0,8,9},{0,8,137},{0,8,73},{0,9,242},{16,7,4},
+        {0,8,85},{0,8,21},{16,8,258},{19,7,43},{0,8,117},{0,8,53},{0,9,202},
+        {17,7,13},{0,8,101},{0,8,37},{0,9,170},{0,8,5},{0,8,133},{0,8,69},
+        {0,9,234},{16,7,8},{0,8,93},{0,8,29},{0,9,154},{20,7,83},{0,8,125},
+        {0,8,61},{0,9,218},{18,7,23},{0,8,109},{0,8,45},{0,9,186},{0,8,13},
+        {0,8,141},{0,8,77},{0,9,250},{16,7,3},{0,8,83},{0,8,19},{21,8,195},
+        {19,7,35},{0,8,115},{0,8,51},{0,9,198},{17,7,11},{0,8,99},{0,8,35},
+        {0,9,166},{0,8,3},{0,8,131},{0,8,67},{0,9,230},{16,7,7},{0,8,91},
+        {0,8,27},{0,9,150},{20,7,67},{0,8,123},{0,8,59},{0,9,214},{18,7,19},
+        {0,8,107},{0,8,43},{0,9,182},{0,8,11},{0,8,139},{0,8,75},{0,9,246},
+        {16,7,5},{0,8,87},{0,8,23},{64,8,0},{19,7,51},{0,8,119},{0,8,55},
+        {0,9,206},{17,7,15},{0,8,103},{0,8,39},{0,9,174},{0,8,7},{0,8,135},
+        {0,8,71},{0,9,238},{16,7,9},{0,8,95},{0,8,31},{0,9,158},{20,7,99},
+        {0,8,127},{0,8,63},{0,9,222},{18,7,27},{0,8,111},{0,8,47},{0,9,190},
+        {0,8,15},{0,8,143},{0,8,79},{0,9,254},{96,7,0},{0,8,80},{0,8,16},
+        {20,8,115},{18,7,31},{0,8,112},{0,8,48},{0,9,193},{16,7,10},{0,8,96},
+        {0,8,32},{0,9,161},{0,8,0},{0,8,128},{0,8,64},{0,9,225},{16,7,6},
+        {0,8,88},{0,8,24},{0,9,145},{19,7,59},{0,8,120},{0,8,56},{0,9,209},
+        {17,7,17},{0,8,104},{0,8,40},{0,9,177},{0,8,8},{0,8,136},{0,8,72},
+        {0,9,241},{16,7,4},{0,8,84},{0,8,20},{21,8,227},{19,7,43},{0,8,116},
+        {0,8,52},{0,9,201},{17,7,13},{0,8,100},{0,8,36},{0,9,169},{0,8,4},
+        {0,8,132},{0,8,68},{0,9,233},{16,7,8},{0,8,92},{0,8,28},{0,9,153},
+        {20,7,83},{0,8,124},{0,8,60},{0,9,217},{18,7,23},{0,8,108},{0,8,44},
+        {0,9,185},{0,8,12},{0,8,140},{0,8,76},{0,9,249},{16,7,3},{0,8,82},
+        {0,8,18},{21,8,163},{19,7,35},{0,8,114},{0,8,50},{0,9,197},{17,7,11},
+        {0,8,98},{0,8,34},{0,9,165},{0,8,2},{0,8,130},{0,8,66},{0,9,229},
+        {16,7,7},{0,8,90},{0,8,26},{0,9,149},{20,7,67},{0,8,122},{0,8,58},
+        {0,9,213},{18,7,19},{0,8,106},{0,8,42},{0,9,181},{0,8,10},{0,8,138},
+        {0,8,74},{0,9,245},{16,7,5},{0,8,86},{0,8,22},{64,8,0},{19,7,51},
+        {0,8,118},{0,8,54},{0,9,205},{17,7,15},{0,8,102},{0,8,38},{0,9,173},
+        {0,8,6},{0,8,134},{0,8,70},{0,9,237},{16,7,9},{0,8,94},{0,8,30},
+        {0,9,157},{20,7,99},{0,8,126},{0,8,62},{0,9,221},{18,7,27},{0,8,110},
+        {0,8,46},{0,9,189},{0,8,14},{0,8,142},{0,8,78},{0,9,253},{96,7,0},
+        {0,8,81},{0,8,17},{21,8,131},{18,7,31},{0,8,113},{0,8,49},{0,9,195},
+        {16,7,10},{0,8,97},{0,8,33},{0,9,163},{0,8,1},{0,8,129},{0,8,65},
+        {0,9,227},{16,7,6},{0,8,89},{0,8,25},{0,9,147},{19,7,59},{0,8,121},
+        {0,8,57},{0,9,211},{17,7,17},{0,8,105},{0,8,41},{0,9,179},{0,8,9},
+        {0,8,137},{0,8,73},{0,9,243},{16,7,4},{0,8,85},{0,8,21},{16,8,258},
+        {19,7,43},{0,8,117},{0,8,53},{0,9,203},{17,7,13},{0,8,101},{0,8,37},
+        {0,9,171},{0,8,5},{0,8,133},{0,8,69},{0,9,235},{16,7,8},{0,8,93},
+        {0,8,29},{0,9,155},{20,7,83},{0,8,125},{0,8,61},{0,9,219},{18,7,23},
+        {0,8,109},{0,8,45},{0,9,187},{0,8,13},{0,8,141},{0,8,77},{0,9,251},
+        {16,7,3},{0,8,83},{0,8,19},{21,8,195},{19,7,35},{0,8,115},{0,8,51},
+        {0,9,199},{17,7,11},{0,8,99},{0,8,35},{0,9,167},{0,8,3},{0,8,131},
+        {0,8,67},{0,9,231},{16,7,7},{0,8,91},{0,8,27},{0,9,151},{20,7,67},
+        {0,8,123},{0,8,59},{0,9,215},{18,7,19},{0,8,107},{0,8,43},{0,9,183},
+        {0,8,11},{0,8,139},{0,8,75},{0,9,247},{16,7,5},{0,8,87},{0,8,23},
+        {64,8,0},{19,7,51},{0,8,119},{0,8,55},{0,9,207},{17,7,15},{0,8,103},
+        {0,8,39},{0,9,175},{0,8,7},{0,8,135},{0,8,71},{0,9,239},{16,7,9},
+        {0,8,95},{0,8,31},{0,9,159},{20,7,99},{0,8,127},{0,8,63},{0,9,223},
+        {18,7,27},{0,8,111},{0,8,47},{0,9,191},{0,8,15},{0,8,143},{0,8,79},
+        {0,9,255}
+    };
+
+    static const code distfix[32] = {
+        {16,5,1},{23,5,257},{19,5,17},{27,5,4097},{17,5,5},{25,5,1025},
+        {21,5,65},{29,5,16385},{16,5,3},{24,5,513},{20,5,33},{28,5,8193},
+        {18,5,9},{26,5,2049},{22,5,129},{64,5,0},{16,5,2},{23,5,385},
+        {19,5,25},{27,5,6145},{17,5,7},{25,5,1537},{21,5,97},{29,5,24577},
+        {16,5,4},{24,5,769},{20,5,49},{28,5,12289},{18,5,13},{26,5,3073},
+        {22,5,193},{64,5,0}
+    };
diff --git a/distrib/zlib-1.2.3/inflate.c b/distrib/zlib-1.2.3/inflate.c
new file mode 100644
index 0000000..792fdee
--- /dev/null
+++ b/distrib/zlib-1.2.3/inflate.c
@@ -0,0 +1,1368 @@
+/* inflate.c -- zlib decompression
+ * Copyright (C) 1995-2005 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/*
+ * Change history:
+ *
+ * 1.2.beta0    24 Nov 2002
+ * - First version -- complete rewrite of inflate to simplify code, avoid
+ *   creation of window when not needed, minimize use of window when it is
+ *   needed, make inffast.c even faster, implement gzip decoding, and to
+ *   improve code readability and style over the previous zlib inflate code
+ *
+ * 1.2.beta1    25 Nov 2002
+ * - Use pointers for available input and output checking in inffast.c
+ * - Remove input and output counters in inffast.c
+ * - Change inffast.c entry and loop from avail_in >= 7 to >= 6
+ * - Remove unnecessary second byte pull from length extra in inffast.c
+ * - Unroll direct copy to three copies per loop in inffast.c
+ *
+ * 1.2.beta2    4 Dec 2002
+ * - Change external routine names to reduce potential conflicts
+ * - Correct filename to inffixed.h for fixed tables in inflate.c
+ * - Make hbuf[] unsigned char to match parameter type in inflate.c
+ * - Change strm->next_out[-state->offset] to *(strm->next_out - state->offset)
+ *   to avoid negation problem on Alphas (64 bit) in inflate.c
+ *
+ * 1.2.beta3    22 Dec 2002
+ * - Add comments on state->bits assertion in inffast.c
+ * - Add comments on op field in inftrees.h
+ * - Fix bug in reuse of allocated window after inflateReset()
+ * - Remove bit fields--back to byte structure for speed
+ * - Remove distance extra == 0 check in inflate_fast()--only helps for lengths
+ * - Change post-increments to pre-increments in inflate_fast(), PPC biased?
+ * - Add compile time option, POSTINC, to use post-increments instead (Intel?)
+ * - Make MATCH copy in inflate() much faster for when inflate_fast() not used
+ * - Use local copies of stream next and avail values, as well as local bit
+ *   buffer and bit count in inflate()--for speed when inflate_fast() not used
+ *
+ * 1.2.beta4    1 Jan 2003
+ * - Split ptr - 257 statements in inflate_table() to avoid compiler warnings
+ * - Move a comment on output buffer sizes from inffast.c to inflate.c
+ * - Add comments in inffast.c to introduce the inflate_fast() routine
+ * - Rearrange window copies in inflate_fast() for speed and simplification
+ * - Unroll last copy for window match in inflate_fast()
+ * - Use local copies of window variables in inflate_fast() for speed
+ * - Pull out common write == 0 case for speed in inflate_fast()
+ * - Make op and len in inflate_fast() unsigned for consistency
+ * - Add FAR to lcode and dcode declarations in inflate_fast()
+ * - Simplified bad distance check in inflate_fast()
+ * - Added inflateBackInit(), inflateBack(), and inflateBackEnd() in new
+ *   source file infback.c to provide a call-back interface to inflate for
+ *   programs like gzip and unzip -- uses window as output buffer to avoid
+ *   window copying
+ *
+ * 1.2.beta5    1 Jan 2003
+ * - Improved inflateBack() interface to allow the caller to provide initial
+ *   input in strm.
+ * - Fixed stored blocks bug in inflateBack()
+ *
+ * 1.2.beta6    4 Jan 2003
+ * - Added comments in inffast.c on effectiveness of POSTINC
+ * - Typecasting all around to reduce compiler warnings
+ * - Changed loops from while (1) or do {} while (1) to for (;;), again to
+ *   make compilers happy
+ * - Changed type of window in inflateBackInit() to unsigned char *
+ *
+ * 1.2.beta7    27 Jan 2003
+ * - Changed many types to unsigned or unsigned short to avoid warnings
+ * - Added inflateCopy() function
+ *
+ * 1.2.0        9 Mar 2003
+ * - Changed inflateBack() interface to provide separate opaque descriptors
+ *   for the in() and out() functions
+ * - Changed inflateBack() argument and in_func typedef to swap the length
+ *   and buffer address return values for the input function
+ * - Check next_in and next_out for Z_NULL on entry to inflate()
+ *
+ * The history for versions after 1.2.0 are in ChangeLog in zlib distribution.
+ */
+
+#include "zutil.h"
+#include "inftrees.h"
+#include "inflate.h"
+#include "inffast.h"
+
+#ifdef MAKEFIXED
+#  ifndef BUILDFIXED
+#    define BUILDFIXED
+#  endif
+#endif
+
+/* function prototypes */
+local void fixedtables OF((struct inflate_state FAR *state));
+local int updatewindow OF((z_streamp strm, unsigned out));
+#ifdef BUILDFIXED
+   void makefixed OF((void));
+#endif
+local unsigned syncsearch OF((unsigned FAR *have, unsigned char FAR *buf,
+                              unsigned len));
+
+int ZEXPORT inflateReset(strm)
+z_streamp strm;
+{
+    struct inflate_state FAR *state;
+
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    state = (struct inflate_state FAR *)strm->state;
+    strm->total_in = strm->total_out = state->total = 0;
+    strm->msg = Z_NULL;
+    strm->adler = 1;        /* to support ill-conceived Java test suite */
+    state->mode = HEAD;
+    state->last = 0;
+    state->havedict = 0;
+    state->dmax = 32768U;
+    state->head = Z_NULL;
+    state->wsize = 0;
+    state->whave = 0;
+    state->write = 0;
+    state->hold = 0;
+    state->bits = 0;
+    state->lencode = state->distcode = state->next = state->codes;
+    Tracev((stderr, "inflate: reset\n"));
+    return Z_OK;
+}
+
+int ZEXPORT inflatePrime(strm, bits, value)
+z_streamp strm;
+int bits;
+int value;
+{
+    struct inflate_state FAR *state;
+
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    state = (struct inflate_state FAR *)strm->state;
+    if (bits > 16 || state->bits + bits > 32) return Z_STREAM_ERROR;
+    value &= (1L << bits) - 1;
+    state->hold += value << state->bits;
+    state->bits += bits;
+    return Z_OK;
+}
+
+int ZEXPORT inflateInit2_(strm, windowBits, version, stream_size)
+z_streamp strm;
+int windowBits;
+const char *version;
+int stream_size;
+{
+    struct inflate_state FAR *state;
+
+    if (version == Z_NULL || version[0] != ZLIB_VERSION[0] ||
+        stream_size != (int)(sizeof(z_stream)))
+        return Z_VERSION_ERROR;
+    if (strm == Z_NULL) return Z_STREAM_ERROR;
+    strm->msg = Z_NULL;                 /* in case we return an error */
+    if (strm->zalloc == (alloc_func)0) {
+        strm->zalloc = zcalloc;
+        strm->opaque = (voidpf)0;
+    }
+    if (strm->zfree == (free_func)0) strm->zfree = zcfree;
+    state = (struct inflate_state FAR *)
+            ZALLOC(strm, 1, sizeof(struct inflate_state));
+    if (state == Z_NULL) return Z_MEM_ERROR;
+    Tracev((stderr, "inflate: allocated\n"));
+    strm->state = (struct internal_state FAR *)state;
+    if (windowBits < 0) {
+        state->wrap = 0;
+        windowBits = -windowBits;
+    }
+    else {
+        state->wrap = (windowBits >> 4) + 1;
+#ifdef GUNZIP
+        if (windowBits < 48) windowBits &= 15;
+#endif
+    }
+    if (windowBits < 8 || windowBits > 15) {
+        ZFREE(strm, state);
+        strm->state = Z_NULL;
+        return Z_STREAM_ERROR;
+    }
+    state->wbits = (unsigned)windowBits;
+    state->window = Z_NULL;
+    return inflateReset(strm);
+}
+
+int ZEXPORT inflateInit_(strm, version, stream_size)
+z_streamp strm;
+const char *version;
+int stream_size;
+{
+    return inflateInit2_(strm, DEF_WBITS, version, stream_size);
+}
+
+/*
+   Return state with length and distance decoding tables and index sizes set to
+   fixed code decoding.  Normally this returns fixed tables from inffixed.h.
+   If BUILDFIXED is defined, then instead this routine builds the tables the
+   first time it's called, and returns those tables the first time and
+   thereafter.  This reduces the size of the code by about 2K bytes, in
+   exchange for a little execution time.  However, BUILDFIXED should not be
+   used for threaded applications, since the rewriting of the tables and virgin
+   may not be thread-safe.
+ */
+local void fixedtables(state)
+struct inflate_state FAR *state;
+{
+#ifdef BUILDFIXED
+    static int virgin = 1;
+    static code *lenfix, *distfix;
+    static code fixed[544];
+
+    /* build fixed huffman tables if first call (may not be thread safe) */
+    if (virgin) {
+        unsigned sym, bits;
+        static code *next;
+
+        /* literal/length table */
+        sym = 0;
+        while (sym < 144) state->lens[sym++] = 8;
+        while (sym < 256) state->lens[sym++] = 9;
+        while (sym < 280) state->lens[sym++] = 7;
+        while (sym < 288) state->lens[sym++] = 8;
+        next = fixed;
+        lenfix = next;
+        bits = 9;
+        inflate_table(LENS, state->lens, 288, &(next), &(bits), state->work);
+
+        /* distance table */
+        sym = 0;
+        while (sym < 32) state->lens[sym++] = 5;
+        distfix = next;
+        bits = 5;
+        inflate_table(DISTS, state->lens, 32, &(next), &(bits), state->work);
+
+        /* do this just once */
+        virgin = 0;
+    }
+#else /* !BUILDFIXED */
+#   include "inffixed.h"
+#endif /* BUILDFIXED */
+    state->lencode = lenfix;
+    state->lenbits = 9;
+    state->distcode = distfix;
+    state->distbits = 5;
+}
+
+#ifdef MAKEFIXED
+#include <stdio.h>
+
+/*
+   Write out the inffixed.h that is #include'd above.  Defining MAKEFIXED also
+   defines BUILDFIXED, so the tables are built on the fly.  makefixed() writes
+   those tables to stdout, which would be piped to inffixed.h.  A small program
+   can simply call makefixed to do this:
+
+    void makefixed(void);
+
+    int main(void)
+    {
+        makefixed();
+        return 0;
+    }
+
+   Then that can be linked with zlib built with MAKEFIXED defined and run:
+
+    a.out > inffixed.h
+ */
+void makefixed()
+{
+    unsigned low, size;
+    struct inflate_state state;
+
+    fixedtables(&state);
+    puts("    /* inffixed.h -- table for decoding fixed codes");
+    puts("     * Generated automatically by makefixed().");
+    puts("     */");
+    puts("");
+    puts("    /* WARNING: this file should *not* be used by applications.");
+    puts("       It is part of the implementation of this library and is");
+    puts("       subject to change. Applications should only use zlib.h.");
+    puts("     */");
+    puts("");
+    size = 1U << 9;
+    printf("    static const code lenfix[%u] = {", size);
+    low = 0;
+    for (;;) {
+        if ((low % 7) == 0) printf("\n        ");
+        printf("{%u,%u,%d}", state.lencode[low].op, state.lencode[low].bits,
+               state.lencode[low].val);
+        if (++low == size) break;
+        putchar(',');
+    }
+    puts("\n    };");
+    size = 1U << 5;
+    printf("\n    static const code distfix[%u] = {", size);
+    low = 0;
+    for (;;) {
+        if ((low % 6) == 0) printf("\n        ");
+        printf("{%u,%u,%d}", state.distcode[low].op, state.distcode[low].bits,
+               state.distcode[low].val);
+        if (++low == size) break;
+        putchar(',');
+    }
+    puts("\n    };");
+}
+#endif /* MAKEFIXED */
+
+/*
+   Update the window with the last wsize (normally 32K) bytes written before
+   returning.  If window does not exist yet, create it.  This is only called
+   when a window is already in use, or when output has been written during this
+   inflate call, but the end of the deflate stream has not been reached yet.
+   It is also called to create a window for dictionary data when a dictionary
+   is loaded.
+
+   Providing output buffers larger than 32K to inflate() should provide a speed
+   advantage, since only the last 32K of output is copied to the sliding window
+   upon return from inflate(), and since all distances after the first 32K of
+   output will fall in the output data, making match copies simpler and faster.
+   The advantage may be dependent on the size of the processor's data caches.
+ */
+local int updatewindow(strm, out)
+z_streamp strm;
+unsigned out;
+{
+    struct inflate_state FAR *state;
+    unsigned copy, dist;
+
+    state = (struct inflate_state FAR *)strm->state;
+
+    /* if it hasn't been done already, allocate space for the window */
+    if (state->window == Z_NULL) {
+        state->window = (unsigned char FAR *)
+                        ZALLOC(strm, 1U << state->wbits,
+                               sizeof(unsigned char));
+        if (state->window == Z_NULL) return 1;
+    }
+
+    /* if window not in use yet, initialize */
+    if (state->wsize == 0) {
+        state->wsize = 1U << state->wbits;
+        state->write = 0;
+        state->whave = 0;
+    }
+
+    /* copy state->wsize or less output bytes into the circular window */
+    copy = out - strm->avail_out;
+    if (copy >= state->wsize) {
+        zmemcpy(state->window, strm->next_out - state->wsize, state->wsize);
+        state->write = 0;
+        state->whave = state->wsize;
+    }
+    else {
+        dist = state->wsize - state->write;
+        if (dist > copy) dist = copy;
+        zmemcpy(state->window + state->write, strm->next_out - copy, dist);
+        copy -= dist;
+        if (copy) {
+            zmemcpy(state->window, strm->next_out - copy, copy);
+            state->write = copy;
+            state->whave = state->wsize;
+        }
+        else {
+            state->write += dist;
+            if (state->write == state->wsize) state->write = 0;
+            if (state->whave < state->wsize) state->whave += dist;
+        }
+    }
+    return 0;
+}
+
+/* Macros for inflate(): */
+
+/* check function to use adler32() for zlib or crc32() for gzip */
+#ifdef GUNZIP
+#  define UPDATE(check, buf, len) \
+    (state->flags ? crc32(check, buf, len) : adler32(check, buf, len))
+#else
+#  define UPDATE(check, buf, len) adler32(check, buf, len)
+#endif
+
+/* check macros for header crc */
+#ifdef GUNZIP
+#  define CRC2(check, word) \
+    do { \
+        hbuf[0] = (unsigned char)(word); \
+        hbuf[1] = (unsigned char)((word) >> 8); \
+        check = crc32(check, hbuf, 2); \
+    } while (0)
+
+#  define CRC4(check, word) \
+    do { \
+        hbuf[0] = (unsigned char)(word); \
+        hbuf[1] = (unsigned char)((word) >> 8); \
+        hbuf[2] = (unsigned char)((word) >> 16); \
+        hbuf[3] = (unsigned char)((word) >> 24); \
+        check = crc32(check, hbuf, 4); \
+    } while (0)
+#endif
+
+/* Load registers with state in inflate() for speed */
+#define LOAD() \
+    do { \
+        put = strm->next_out; \
+        left = strm->avail_out; \
+        next = strm->next_in; \
+        have = strm->avail_in; \
+        hold = state->hold; \
+        bits = state->bits; \
+    } while (0)
+
+/* Restore state from registers in inflate() */
+#define RESTORE() \
+    do { \
+        strm->next_out = put; \
+        strm->avail_out = left; \
+        strm->next_in = next; \
+        strm->avail_in = have; \
+        state->hold = hold; \
+        state->bits = bits; \
+    } while (0)
+
+/* Clear the input bit accumulator */
+#define INITBITS() \
+    do { \
+        hold = 0; \
+        bits = 0; \
+    } while (0)
+
+/* Get a byte of input into the bit accumulator, or return from inflate()
+   if there is no input available. */
+#define PULLBYTE() \
+    do { \
+        if (have == 0) goto inf_leave; \
+        have--; \
+        hold += (unsigned long)(*next++) << bits; \
+        bits += 8; \
+    } while (0)
+
+/* Assure that there are at least n bits in the bit accumulator.  If there is
+   not enough available input to do that, then return from inflate(). */
+#define NEEDBITS(n) \
+    do { \
+        while (bits < (unsigned)(n)) \
+            PULLBYTE(); \
+    } while (0)
+
+/* Return the low n bits of the bit accumulator (n < 16) */
+#define BITS(n) \
+    ((unsigned)hold & ((1U << (n)) - 1))
+
+/* Remove n bits from the bit accumulator */
+#define DROPBITS(n) \
+    do { \
+        hold >>= (n); \
+        bits -= (unsigned)(n); \
+    } while (0)
+
+/* Remove zero to seven bits as needed to go to a byte boundary */
+#define BYTEBITS() \
+    do { \
+        hold >>= bits & 7; \
+        bits -= bits & 7; \
+    } while (0)
+
+/* Reverse the bytes in a 32-bit value */
+#define REVERSE(q) \
+    ((((q) >> 24) & 0xff) + (((q) >> 8) & 0xff00) + \
+     (((q) & 0xff00) << 8) + (((q) & 0xff) << 24))
+
+/*
+   inflate() uses a state machine to process as much input data and generate as
+   much output data as possible before returning.  The state machine is
+   structured roughly as follows:
+
+    for (;;) switch (state) {
+    ...
+    case STATEn:
+        if (not enough input data or output space to make progress)
+            return;
+        ... make progress ...
+        state = STATEm;
+        break;
+    ...
+    }
+
+   so when inflate() is called again, the same case is attempted again, and
+   if the appropriate resources are provided, the machine proceeds to the
+   next state.  The NEEDBITS() macro is usually the way the state evaluates
+   whether it can proceed or should return.  NEEDBITS() does the return if
+   the requested bits are not available.  The typical use of the BITS macros
+   is:
+
+        NEEDBITS(n);
+        ... do something with BITS(n) ...
+        DROPBITS(n);
+
+   where NEEDBITS(n) either returns from inflate() if there isn't enough
+   input left to load n bits into the accumulator, or it continues.  BITS(n)
+   gives the low n bits in the accumulator.  When done, DROPBITS(n) drops
+   the low n bits off the accumulator.  INITBITS() clears the accumulator
+   and sets the number of available bits to zero.  BYTEBITS() discards just
+   enough bits to put the accumulator on a byte boundary.  After BYTEBITS()
+   and a NEEDBITS(8), then BITS(8) would return the next byte in the stream.
+
+   NEEDBITS(n) uses PULLBYTE() to get an available byte of input, or to return
+   if there is no input available.  The decoding of variable length codes uses
+   PULLBYTE() directly in order to pull just enough bytes to decode the next
+   code, and no more.
+
+   Some states loop until they get enough input, making sure that enough
+   state information is maintained to continue the loop where it left off
+   if NEEDBITS() returns in the loop.  For example, want, need, and keep
+   would all have to actually be part of the saved state in case NEEDBITS()
+   returns:
+
+    case STATEw:
+        while (want < need) {
+            NEEDBITS(n);
+            keep[want++] = BITS(n);
+            DROPBITS(n);
+        }
+        state = STATEx;
+    case STATEx:
+
+   As shown above, if the next state is also the next case, then the break
+   is omitted.
+
+   A state may also return if there is not enough output space available to
+   complete that state.  Those states are copying stored data, writing a
+   literal byte, and copying a matching string.
+
+   When returning, a "goto inf_leave" is used to update the total counters,
+   update the check value, and determine whether any progress has been made
+   during that inflate() call in order to return the proper return code.
+   Progress is defined as a change in either strm->avail_in or strm->avail_out.
+   When there is a window, goto inf_leave will update the window with the last
+   output written.  If a goto inf_leave occurs in the middle of decompression
+   and there is no window currently, goto inf_leave will create one and copy
+   output to the window for the next call of inflate().
+
+   In this implementation, the flush parameter of inflate() only affects the
+   return code (per zlib.h).  inflate() always writes as much as possible to
+   strm->next_out, given the space available and the provided input--the effect
+   documented in zlib.h of Z_SYNC_FLUSH.  Furthermore, inflate() always defers
+   the allocation of and copying into a sliding window until necessary, which
+   provides the effect documented in zlib.h for Z_FINISH when the entire input
+   stream available.  So the only thing the flush parameter actually does is:
+   when flush is set to Z_FINISH, inflate() cannot return Z_OK.  Instead it
+   will return Z_BUF_ERROR if it has not reached the end of the stream.
+ */
+
+int ZEXPORT inflate(strm, flush)
+z_streamp strm;
+int flush;
+{
+    struct inflate_state FAR *state;
+    unsigned char FAR *next;    /* next input */
+    unsigned char FAR *put;     /* next output */
+    unsigned have, left;        /* available input and output */
+    unsigned long hold;         /* bit buffer */
+    unsigned bits;              /* bits in bit buffer */
+    unsigned in, out;           /* save starting available input and output */
+    unsigned copy;              /* number of stored or match bytes to copy */
+    unsigned char FAR *from;    /* where to copy match bytes from */
+    code this;                  /* current decoding table entry */
+    code last;                  /* parent table entry */
+    unsigned len;               /* length to copy for repeats, bits to drop */
+    int ret;                    /* return code */
+#ifdef GUNZIP
+    unsigned char hbuf[4];      /* buffer for gzip header crc calculation */
+#endif
+    static const unsigned short order[19] = /* permutation of code lengths */
+        {16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15};
+
+    if (strm == Z_NULL || strm->state == Z_NULL || strm->next_out == Z_NULL ||
+        (strm->next_in == Z_NULL && strm->avail_in != 0))
+        return Z_STREAM_ERROR;
+
+    state = (struct inflate_state FAR *)strm->state;
+    if (state->mode == TYPE) state->mode = TYPEDO;      /* skip check */
+    LOAD();
+    in = have;
+    out = left;
+    ret = Z_OK;
+    for (;;)
+        switch (state->mode) {
+        case HEAD:
+            if (state->wrap == 0) {
+                state->mode = TYPEDO;
+                break;
+            }
+            NEEDBITS(16);
+#ifdef GUNZIP
+            if ((state->wrap & 2) && hold == 0x8b1f) {  /* gzip header */
+                state->check = crc32(0L, Z_NULL, 0);
+                CRC2(state->check, hold);
+                INITBITS();
+                state->mode = FLAGS;
+                break;
+            }
+            state->flags = 0;           /* expect zlib header */
+            if (state->head != Z_NULL)
+                state->head->done = -1;
+            if (!(state->wrap & 1) ||   /* check if zlib header allowed */
+#else
+            if (
+#endif
+                ((BITS(8) << 8) + (hold >> 8)) % 31) {
+                strm->msg = (char *)"incorrect header check";
+                state->mode = BAD;
+                break;
+            }
+            if (BITS(4) != Z_DEFLATED) {
+                strm->msg = (char *)"unknown compression method";
+                state->mode = BAD;
+                break;
+            }
+            DROPBITS(4);
+            len = BITS(4) + 8;
+            if (len > state->wbits) {
+                strm->msg = (char *)"invalid window size";
+                state->mode = BAD;
+                break;
+            }
+            state->dmax = 1U << len;
+            Tracev((stderr, "inflate:   zlib header ok\n"));
+            strm->adler = state->check = adler32(0L, Z_NULL, 0);
+            state->mode = hold & 0x200 ? DICTID : TYPE;
+            INITBITS();
+            break;
+#ifdef GUNZIP
+        case FLAGS:
+            NEEDBITS(16);
+            state->flags = (int)(hold);
+            if ((state->flags & 0xff) != Z_DEFLATED) {
+                strm->msg = (char *)"unknown compression method";
+                state->mode = BAD;
+                break;
+            }
+            if (state->flags & 0xe000) {
+                strm->msg = (char *)"unknown header flags set";
+                state->mode = BAD;
+                break;
+            }
+            if (state->head != Z_NULL)
+                state->head->text = (int)((hold >> 8) & 1);
+            if (state->flags & 0x0200) CRC2(state->check, hold);
+            INITBITS();
+            state->mode = TIME;
+        case TIME:
+            NEEDBITS(32);
+            if (state->head != Z_NULL)
+                state->head->time = hold;
+            if (state->flags & 0x0200) CRC4(state->check, hold);
+            INITBITS();
+            state->mode = OS;
+        case OS:
+            NEEDBITS(16);
+            if (state->head != Z_NULL) {
+                state->head->xflags = (int)(hold & 0xff);
+                state->head->os = (int)(hold >> 8);
+            }
+            if (state->flags & 0x0200) CRC2(state->check, hold);
+            INITBITS();
+            state->mode = EXLEN;
+        case EXLEN:
+            if (state->flags & 0x0400) {
+                NEEDBITS(16);
+                state->length = (unsigned)(hold);
+                if (state->head != Z_NULL)
+                    state->head->extra_len = (unsigned)hold;
+                if (state->flags & 0x0200) CRC2(state->check, hold);
+                INITBITS();
+            }
+            else if (state->head != Z_NULL)
+                state->head->extra = Z_NULL;
+            state->mode = EXTRA;
+        case EXTRA:
+            if (state->flags & 0x0400) {
+                copy = state->length;
+                if (copy > have) copy = have;
+                if (copy) {
+                    if (state->head != Z_NULL &&
+                        state->head->extra != Z_NULL) {
+                        len = state->head->extra_len - state->length;
+                        zmemcpy(state->head->extra + len, next,
+                                len + copy > state->head->extra_max ?
+                                state->head->extra_max - len : copy);
+                    }
+                    if (state->flags & 0x0200)
+                        state->check = crc32(state->check, next, copy);
+                    have -= copy;
+                    next += copy;
+                    state->length -= copy;
+                }
+                if (state->length) goto inf_leave;
+            }
+            state->length = 0;
+            state->mode = NAME;
+        case NAME:
+            if (state->flags & 0x0800) {
+                if (have == 0) goto inf_leave;
+                copy = 0;
+                do {
+                    len = (unsigned)(next[copy++]);
+                    if (state->head != Z_NULL &&
+                            state->head->name != Z_NULL &&
+                            state->length < state->head->name_max)
+                        state->head->name[state->length++] = len;
+                } while (len && copy < have);
+                if (state->flags & 0x0200)
+                    state->check = crc32(state->check, next, copy);
+                have -= copy;
+                next += copy;
+                if (len) goto inf_leave;
+            }
+            else if (state->head != Z_NULL)
+                state->head->name = Z_NULL;
+            state->length = 0;
+            state->mode = COMMENT;
+        case COMMENT:
+            if (state->flags & 0x1000) {
+                if (have == 0) goto inf_leave;
+                copy = 0;
+                do {
+                    len = (unsigned)(next[copy++]);
+                    if (state->head != Z_NULL &&
+                            state->head->comment != Z_NULL &&
+                            state->length < state->head->comm_max)
+                        state->head->comment[state->length++] = len;
+                } while (len && copy < have);
+                if (state->flags & 0x0200)
+                    state->check = crc32(state->check, next, copy);
+                have -= copy;
+                next += copy;
+                if (len) goto inf_leave;
+            }
+            else if (state->head != Z_NULL)
+                state->head->comment = Z_NULL;
+            state->mode = HCRC;
+        case HCRC:
+            if (state->flags & 0x0200) {
+                NEEDBITS(16);
+                if (hold != (state->check & 0xffff)) {
+                    strm->msg = (char *)"header crc mismatch";
+                    state->mode = BAD;
+                    break;
+                }
+                INITBITS();
+            }
+            if (state->head != Z_NULL) {
+                state->head->hcrc = (int)((state->flags >> 9) & 1);
+                state->head->done = 1;
+            }
+            strm->adler = state->check = crc32(0L, Z_NULL, 0);
+            state->mode = TYPE;
+            break;
+#endif
+        case DICTID:
+            NEEDBITS(32);
+            strm->adler = state->check = REVERSE(hold);
+            INITBITS();
+            state->mode = DICT;
+        case DICT:
+            if (state->havedict == 0) {
+                RESTORE();
+                return Z_NEED_DICT;
+            }
+            strm->adler = state->check = adler32(0L, Z_NULL, 0);
+            state->mode = TYPE;
+        case TYPE:
+            if (flush == Z_BLOCK) goto inf_leave;
+        case TYPEDO:
+            if (state->last) {
+                BYTEBITS();
+                state->mode = CHECK;
+                break;
+            }
+            NEEDBITS(3);
+            state->last = BITS(1);
+            DROPBITS(1);
+            switch (BITS(2)) {
+            case 0:                             /* stored block */
+                Tracev((stderr, "inflate:     stored block%s\n",
+                        state->last ? " (last)" : ""));
+                state->mode = STORED;
+                break;
+            case 1:                             /* fixed block */
+                fixedtables(state);
+                Tracev((stderr, "inflate:     fixed codes block%s\n",
+                        state->last ? " (last)" : ""));
+                state->mode = LEN;              /* decode codes */
+                break;
+            case 2:                             /* dynamic block */
+                Tracev((stderr, "inflate:     dynamic codes block%s\n",
+                        state->last ? " (last)" : ""));
+                state->mode = TABLE;
+                break;
+            case 3:
+                strm->msg = (char *)"invalid block type";
+                state->mode = BAD;
+            }
+            DROPBITS(2);
+            break;
+        case STORED:
+            BYTEBITS();                         /* go to byte boundary */
+            NEEDBITS(32);
+            if ((hold & 0xffff) != ((hold >> 16) ^ 0xffff)) {
+                strm->msg = (char *)"invalid stored block lengths";
+                state->mode = BAD;
+                break;
+            }
+            state->length = (unsigned)hold & 0xffff;
+            Tracev((stderr, "inflate:       stored length %u\n",
+                    state->length));
+            INITBITS();
+            state->mode = COPY;
+        case COPY:
+            copy = state->length;
+            if (copy) {
+                if (copy > have) copy = have;
+                if (copy > left) copy = left;
+                if (copy == 0) goto inf_leave;
+                zmemcpy(put, next, copy);
+                have -= copy;
+                next += copy;
+                left -= copy;
+                put += copy;
+                state->length -= copy;
+                break;
+            }
+            Tracev((stderr, "inflate:       stored end\n"));
+            state->mode = TYPE;
+            break;
+        case TABLE:
+            NEEDBITS(14);
+            state->nlen = BITS(5) + 257;
+            DROPBITS(5);
+            state->ndist = BITS(5) + 1;
+            DROPBITS(5);
+            state->ncode = BITS(4) + 4;
+            DROPBITS(4);
+#ifndef PKZIP_BUG_WORKAROUND
+            if (state->nlen > 286 || state->ndist > 30) {
+                strm->msg = (char *)"too many length or distance symbols";
+                state->mode = BAD;
+                break;
+            }
+#endif
+            Tracev((stderr, "inflate:       table sizes ok\n"));
+            state->have = 0;
+            state->mode = LENLENS;
+        case LENLENS:
+            while (state->have < state->ncode) {
+                NEEDBITS(3);
+                state->lens[order[state->have++]] = (unsigned short)BITS(3);
+                DROPBITS(3);
+            }
+            while (state->have < 19)
+                state->lens[order[state->have++]] = 0;
+            state->next = state->codes;
+            state->lencode = (code const FAR *)(state->next);
+            state->lenbits = 7;
+            ret = inflate_table(CODES, state->lens, 19, &(state->next),
+                                &(state->lenbits), state->work);
+            if (ret) {
+                strm->msg = (char *)"invalid code lengths set";
+                state->mode = BAD;
+                break;
+            }
+            Tracev((stderr, "inflate:       code lengths ok\n"));
+            state->have = 0;
+            state->mode = CODELENS;
+        case CODELENS:
+            while (state->have < state->nlen + state->ndist) {
+                for (;;) {
+                    this = state->lencode[BITS(state->lenbits)];
+                    if ((unsigned)(this.bits) <= bits) break;
+                    PULLBYTE();
+                }
+                if (this.val < 16) {
+                    NEEDBITS(this.bits);
+                    DROPBITS(this.bits);
+                    state->lens[state->have++] = this.val;
+                }
+                else {
+                    if (this.val == 16) {
+                        NEEDBITS(this.bits + 2);
+                        DROPBITS(this.bits);
+                        if (state->have == 0) {
+                            strm->msg = (char *)"invalid bit length repeat";
+                            state->mode = BAD;
+                            break;
+                        }
+                        len = state->lens[state->have - 1];
+                        copy = 3 + BITS(2);
+                        DROPBITS(2);
+                    }
+                    else if (this.val == 17) {
+                        NEEDBITS(this.bits + 3);
+                        DROPBITS(this.bits);
+                        len = 0;
+                        copy = 3 + BITS(3);
+                        DROPBITS(3);
+                    }
+                    else {
+                        NEEDBITS(this.bits + 7);
+                        DROPBITS(this.bits);
+                        len = 0;
+                        copy = 11 + BITS(7);
+                        DROPBITS(7);
+                    }
+                    if (state->have + copy > state->nlen + state->ndist) {
+                        strm->msg = (char *)"invalid bit length repeat";
+                        state->mode = BAD;
+                        break;
+                    }
+                    while (copy--)
+                        state->lens[state->have++] = (unsigned short)len;
+                }
+            }
+
+            /* handle error breaks in while */
+            if (state->mode == BAD) break;
+
+            /* build code tables */
+            state->next = state->codes;
+            state->lencode = (code const FAR *)(state->next);
+            state->lenbits = 9;
+            ret = inflate_table(LENS, state->lens, state->nlen, &(state->next),
+                                &(state->lenbits), state->work);
+            if (ret) {
+                strm->msg = (char *)"invalid literal/lengths set";
+                state->mode = BAD;
+                break;
+            }
+            state->distcode = (code const FAR *)(state->next);
+            state->distbits = 6;
+            ret = inflate_table(DISTS, state->lens + state->nlen, state->ndist,
+                            &(state->next), &(state->distbits), state->work);
+            if (ret) {
+                strm->msg = (char *)"invalid distances set";
+                state->mode = BAD;
+                break;
+            }
+            Tracev((stderr, "inflate:       codes ok\n"));
+            state->mode = LEN;
+        case LEN:
+            if (have >= 6 && left >= 258) {
+                RESTORE();
+                inflate_fast(strm, out);
+                LOAD();
+                break;
+            }
+            for (;;) {
+                this = state->lencode[BITS(state->lenbits)];
+                if ((unsigned)(this.bits) <= bits) break;
+                PULLBYTE();
+            }
+            if (this.op && (this.op & 0xf0) == 0) {
+                last = this;
+                for (;;) {
+                    this = state->lencode[last.val +
+                            (BITS(last.bits + last.op) >> last.bits)];
+                    if ((unsigned)(last.bits + this.bits) <= bits) break;
+                    PULLBYTE();
+                }
+                DROPBITS(last.bits);
+            }
+            DROPBITS(this.bits);
+            state->length = (unsigned)this.val;
+            if ((int)(this.op) == 0) {
+                Tracevv((stderr, this.val >= 0x20 && this.val < 0x7f ?
+                        "inflate:         literal '%c'\n" :
+                        "inflate:         literal 0x%02x\n", this.val));
+                state->mode = LIT;
+                break;
+            }
+            if (this.op & 32) {
+                Tracevv((stderr, "inflate:         end of block\n"));
+                state->mode = TYPE;
+                break;
+            }
+            if (this.op & 64) {
+                strm->msg = (char *)"invalid literal/length code";
+                state->mode = BAD;
+                break;
+            }
+            state->extra = (unsigned)(this.op) & 15;
+            state->mode = LENEXT;
+        case LENEXT:
+            if (state->extra) {
+                NEEDBITS(state->extra);
+                state->length += BITS(state->extra);
+                DROPBITS(state->extra);
+            }
+            Tracevv((stderr, "inflate:         length %u\n", state->length));
+            state->mode = DIST;
+        case DIST:
+            for (;;) {
+                this = state->distcode[BITS(state->distbits)];
+                if ((unsigned)(this.bits) <= bits) break;
+                PULLBYTE();
+            }
+            if ((this.op & 0xf0) == 0) {
+                last = this;
+                for (;;) {
+                    this = state->distcode[last.val +
+                            (BITS(last.bits + last.op) >> last.bits)];
+                    if ((unsigned)(last.bits + this.bits) <= bits) break;
+                    PULLBYTE();
+                }
+                DROPBITS(last.bits);
+            }
+            DROPBITS(this.bits);
+            if (this.op & 64) {
+                strm->msg = (char *)"invalid distance code";
+                state->mode = BAD;
+                break;
+            }
+            state->offset = (unsigned)this.val;
+            state->extra = (unsigned)(this.op) & 15;
+            state->mode = DISTEXT;
+        case DISTEXT:
+            if (state->extra) {
+                NEEDBITS(state->extra);
+                state->offset += BITS(state->extra);
+                DROPBITS(state->extra);
+            }
+#ifdef INFLATE_STRICT
+            if (state->offset > state->dmax) {
+                strm->msg = (char *)"invalid distance too far back";
+                state->mode = BAD;
+                break;
+            }
+#endif
+            if (state->offset > state->whave + out - left) {
+                strm->msg = (char *)"invalid distance too far back";
+                state->mode = BAD;
+                break;
+            }
+            Tracevv((stderr, "inflate:         distance %u\n", state->offset));
+            state->mode = MATCH;
+        case MATCH:
+            if (left == 0) goto inf_leave;
+            copy = out - left;
+            if (state->offset > copy) {         /* copy from window */
+                copy = state->offset - copy;
+                if (copy > state->write) {
+                    copy -= state->write;
+                    from = state->window + (state->wsize - copy);
+                }
+                else
+                    from = state->window + (state->write - copy);
+                if (copy > state->length) copy = state->length;
+            }
+            else {                              /* copy from output */
+                from = put - state->offset;
+                copy = state->length;
+            }
+            if (copy > left) copy = left;
+            left -= copy;
+            state->length -= copy;
+            do {
+                *put++ = *from++;
+            } while (--copy);
+            if (state->length == 0) state->mode = LEN;
+            break;
+        case LIT:
+            if (left == 0) goto inf_leave;
+            *put++ = (unsigned char)(state->length);
+            left--;
+            state->mode = LEN;
+            break;
+        case CHECK:
+            if (state->wrap) {
+                NEEDBITS(32);
+                out -= left;
+                strm->total_out += out;
+                state->total += out;
+                if (out)
+                    strm->adler = state->check =
+                        UPDATE(state->check, put - out, out);
+                out = left;
+                if ((
+#ifdef GUNZIP
+                     state->flags ? hold :
+#endif
+                     REVERSE(hold)) != state->check) {
+                    strm->msg = (char *)"incorrect data check";
+                    state->mode = BAD;
+                    break;
+                }
+                INITBITS();
+                Tracev((stderr, "inflate:   check matches trailer\n"));
+            }
+#ifdef GUNZIP
+            state->mode = LENGTH;
+        case LENGTH:
+            if (state->wrap && state->flags) {
+                NEEDBITS(32);
+                if (hold != (state->total & 0xffffffffUL)) {
+                    strm->msg = (char *)"incorrect length check";
+                    state->mode = BAD;
+                    break;
+                }
+                INITBITS();
+                Tracev((stderr, "inflate:   length matches trailer\n"));
+            }
+#endif
+            state->mode = DONE;
+        case DONE:
+            ret = Z_STREAM_END;
+            goto inf_leave;
+        case BAD:
+            ret = Z_DATA_ERROR;
+            goto inf_leave;
+        case MEM:
+            return Z_MEM_ERROR;
+        case SYNC:
+        default:
+            return Z_STREAM_ERROR;
+        }
+
+    /*
+       Return from inflate(), updating the total counts and the check value.
+       If there was no progress during the inflate() call, return a buffer
+       error.  Call updatewindow() to create and/or update the window state.
+       Note: a memory error from inflate() is non-recoverable.
+     */
+  inf_leave:
+    RESTORE();
+    if (state->wsize || (state->mode < CHECK && out != strm->avail_out))
+        if (updatewindow(strm, out)) {
+            state->mode = MEM;
+            return Z_MEM_ERROR;
+        }
+    in -= strm->avail_in;
+    out -= strm->avail_out;
+    strm->total_in += in;
+    strm->total_out += out;
+    state->total += out;
+    if (state->wrap && out)
+        strm->adler = state->check =
+            UPDATE(state->check, strm->next_out - out, out);
+    strm->data_type = state->bits + (state->last ? 64 : 0) +
+                      (state->mode == TYPE ? 128 : 0);
+    if (((in == 0 && out == 0) || flush == Z_FINISH) && ret == Z_OK)
+        ret = Z_BUF_ERROR;
+    return ret;
+}
+
+int ZEXPORT inflateEnd(strm)
+z_streamp strm;
+{
+    struct inflate_state FAR *state;
+    if (strm == Z_NULL || strm->state == Z_NULL || strm->zfree == (free_func)0)
+        return Z_STREAM_ERROR;
+    state = (struct inflate_state FAR *)strm->state;
+    if (state->window != Z_NULL) ZFREE(strm, state->window);
+    ZFREE(strm, strm->state);
+    strm->state = Z_NULL;
+    Tracev((stderr, "inflate: end\n"));
+    return Z_OK;
+}
+
+int ZEXPORT inflateSetDictionary(strm, dictionary, dictLength)
+z_streamp strm;
+const Bytef *dictionary;
+uInt dictLength;
+{
+    struct inflate_state FAR *state;
+    unsigned long id;
+
+    /* check state */
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    state = (struct inflate_state FAR *)strm->state;
+    if (state->wrap != 0 && state->mode != DICT)
+        return Z_STREAM_ERROR;
+
+    /* check for correct dictionary id */
+    if (state->mode == DICT) {
+        id = adler32(0L, Z_NULL, 0);
+        id = adler32(id, dictionary, dictLength);
+        if (id != state->check)
+            return Z_DATA_ERROR;
+    }
+
+    /* copy dictionary to window */
+    if (updatewindow(strm, strm->avail_out)) {
+        state->mode = MEM;
+        return Z_MEM_ERROR;
+    }
+    if (dictLength > state->wsize) {
+        zmemcpy(state->window, dictionary + dictLength - state->wsize,
+                state->wsize);
+        state->whave = state->wsize;
+    }
+    else {
+        zmemcpy(state->window + state->wsize - dictLength, dictionary,
+                dictLength);
+        state->whave = dictLength;
+    }
+    state->havedict = 1;
+    Tracev((stderr, "inflate:   dictionary set\n"));
+    return Z_OK;
+}
+
+int ZEXPORT inflateGetHeader(strm, head)
+z_streamp strm;
+gz_headerp head;
+{
+    struct inflate_state FAR *state;
+
+    /* check state */
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    state = (struct inflate_state FAR *)strm->state;
+    if ((state->wrap & 2) == 0) return Z_STREAM_ERROR;
+
+    /* save header structure */
+    state->head = head;
+    head->done = 0;
+    return Z_OK;
+}
+
+/*
+   Search buf[0..len-1] for the pattern: 0, 0, 0xff, 0xff.  Return when found
+   or when out of input.  When called, *have is the number of pattern bytes
+   found in order so far, in 0..3.  On return *have is updated to the new
+   state.  If on return *have equals four, then the pattern was found and the
+   return value is how many bytes were read including the last byte of the
+   pattern.  If *have is less than four, then the pattern has not been found
+   yet and the return value is len.  In the latter case, syncsearch() can be
+   called again with more data and the *have state.  *have is initialized to
+   zero for the first call.
+ */
+local unsigned syncsearch(have, buf, len)
+unsigned FAR *have;
+unsigned char FAR *buf;
+unsigned len;
+{
+    unsigned got;
+    unsigned next;
+
+    got = *have;
+    next = 0;
+    while (next < len && got < 4) {
+        if ((int)(buf[next]) == (got < 2 ? 0 : 0xff))
+            got++;
+        else if (buf[next])
+            got = 0;
+        else
+            got = 4 - got;
+        next++;
+    }
+    *have = got;
+    return next;
+}
+
+int ZEXPORT inflateSync(strm)
+z_streamp strm;
+{
+    unsigned len;               /* number of bytes to look at or looked at */
+    unsigned long in, out;      /* temporary to save total_in and total_out */
+    unsigned char buf[4];       /* to restore bit buffer to byte string */
+    struct inflate_state FAR *state;
+
+    /* check parameters */
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    state = (struct inflate_state FAR *)strm->state;
+    if (strm->avail_in == 0 && state->bits < 8) return Z_BUF_ERROR;
+
+    /* if first time, start search in bit buffer */
+    if (state->mode != SYNC) {
+        state->mode = SYNC;
+        state->hold <<= state->bits & 7;
+        state->bits -= state->bits & 7;
+        len = 0;
+        while (state->bits >= 8) {
+            buf[len++] = (unsigned char)(state->hold);
+            state->hold >>= 8;
+            state->bits -= 8;
+        }
+        state->have = 0;
+        syncsearch(&(state->have), buf, len);
+    }
+
+    /* search available input */
+    len = syncsearch(&(state->have), strm->next_in, strm->avail_in);
+    strm->avail_in -= len;
+    strm->next_in += len;
+    strm->total_in += len;
+
+    /* return no joy or set up to restart inflate() on a new block */
+    if (state->have != 4) return Z_DATA_ERROR;
+    in = strm->total_in;  out = strm->total_out;
+    inflateReset(strm);
+    strm->total_in = in;  strm->total_out = out;
+    state->mode = TYPE;
+    return Z_OK;
+}
+
+/*
+   Returns true if inflate is currently at the end of a block generated by
+   Z_SYNC_FLUSH or Z_FULL_FLUSH. This function is used by one PPP
+   implementation to provide an additional safety check. PPP uses
+   Z_SYNC_FLUSH but removes the length bytes of the resulting empty stored
+   block. When decompressing, PPP checks that at the end of input packet,
+   inflate is waiting for these length bytes.
+ */
+int ZEXPORT inflateSyncPoint(strm)
+z_streamp strm;
+{
+    struct inflate_state FAR *state;
+
+    if (strm == Z_NULL || strm->state == Z_NULL) return Z_STREAM_ERROR;
+    state = (struct inflate_state FAR *)strm->state;
+    return state->mode == STORED && state->bits == 0;
+}
+
+int ZEXPORT inflateCopy(dest, source)
+z_streamp dest;
+z_streamp source;
+{
+    struct inflate_state FAR *state;
+    struct inflate_state FAR *copy;
+    unsigned char FAR *window;
+    unsigned wsize;
+
+    /* check input */
+    if (dest == Z_NULL || source == Z_NULL || source->state == Z_NULL ||
+        source->zalloc == (alloc_func)0 || source->zfree == (free_func)0)
+        return Z_STREAM_ERROR;
+    state = (struct inflate_state FAR *)source->state;
+
+    /* allocate space */
+    copy = (struct inflate_state FAR *)
+           ZALLOC(source, 1, sizeof(struct inflate_state));
+    if (copy == Z_NULL) return Z_MEM_ERROR;
+    window = Z_NULL;
+    if (state->window != Z_NULL) {
+        window = (unsigned char FAR *)
+                 ZALLOC(source, 1U << state->wbits, sizeof(unsigned char));
+        if (window == Z_NULL) {
+            ZFREE(source, copy);
+            return Z_MEM_ERROR;
+        }
+    }
+
+    /* copy state */
+    zmemcpy(dest, source, sizeof(z_stream));
+    zmemcpy(copy, state, sizeof(struct inflate_state));
+    if (state->lencode >= state->codes &&
+        state->lencode <= state->codes + ENOUGH - 1) {
+        copy->lencode = copy->codes + (state->lencode - state->codes);
+        copy->distcode = copy->codes + (state->distcode - state->codes);
+    }
+    copy->next = copy->codes + (state->next - state->codes);
+    if (window != Z_NULL) {
+        wsize = 1U << state->wbits;
+        zmemcpy(window, state->window, wsize);
+    }
+    copy->window = window;
+    dest->state = (struct internal_state FAR *)copy;
+    return Z_OK;
+}
diff --git a/distrib/zlib-1.2.3/inflate.h b/distrib/zlib-1.2.3/inflate.h
new file mode 100644
index 0000000..07bd3e7
--- /dev/null
+++ b/distrib/zlib-1.2.3/inflate.h
@@ -0,0 +1,115 @@
+/* inflate.h -- internal inflate state definition
+ * Copyright (C) 1995-2004 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* WARNING: this file should *not* be used by applications. It is
+   part of the implementation of the compression library and is
+   subject to change. Applications should only use zlib.h.
+ */
+
+/* define NO_GZIP when compiling if you want to disable gzip header and
+   trailer decoding by inflate().  NO_GZIP would be used to avoid linking in
+   the crc code when it is not needed.  For shared libraries, gzip decoding
+   should be left enabled. */
+#ifndef NO_GZIP
+#  define GUNZIP
+#endif
+
+/* Possible inflate modes between inflate() calls */
+typedef enum {
+    HEAD,       /* i: waiting for magic header */
+    FLAGS,      /* i: waiting for method and flags (gzip) */
+    TIME,       /* i: waiting for modification time (gzip) */
+    OS,         /* i: waiting for extra flags and operating system (gzip) */
+    EXLEN,      /* i: waiting for extra length (gzip) */
+    EXTRA,      /* i: waiting for extra bytes (gzip) */
+    NAME,       /* i: waiting for end of file name (gzip) */
+    COMMENT,    /* i: waiting for end of comment (gzip) */
+    HCRC,       /* i: waiting for header crc (gzip) */
+    DICTID,     /* i: waiting for dictionary check value */
+    DICT,       /* waiting for inflateSetDictionary() call */
+        TYPE,       /* i: waiting for type bits, including last-flag bit */
+        TYPEDO,     /* i: same, but skip check to exit inflate on new block */
+        STORED,     /* i: waiting for stored size (length and complement) */
+        COPY,       /* i/o: waiting for input or output to copy stored block */
+        TABLE,      /* i: waiting for dynamic block table lengths */
+        LENLENS,    /* i: waiting for code length code lengths */
+        CODELENS,   /* i: waiting for length/lit and distance code lengths */
+            LEN,        /* i: waiting for length/lit code */
+            LENEXT,     /* i: waiting for length extra bits */
+            DIST,       /* i: waiting for distance code */
+            DISTEXT,    /* i: waiting for distance extra bits */
+            MATCH,      /* o: waiting for output space to copy string */
+            LIT,        /* o: waiting for output space to write literal */
+    CHECK,      /* i: waiting for 32-bit check value */
+    LENGTH,     /* i: waiting for 32-bit length (gzip) */
+    DONE,       /* finished check, done -- remain here until reset */
+    BAD,        /* got a data error -- remain here until reset */
+    MEM,        /* got an inflate() memory error -- remain here until reset */
+    SYNC        /* looking for synchronization bytes to restart inflate() */
+} inflate_mode;
+
+/*
+    State transitions between above modes -
+
+    (most modes can go to the BAD or MEM mode -- not shown for clarity)
+
+    Process header:
+        HEAD -> (gzip) or (zlib)
+        (gzip) -> FLAGS -> TIME -> OS -> EXLEN -> EXTRA -> NAME
+        NAME -> COMMENT -> HCRC -> TYPE
+        (zlib) -> DICTID or TYPE
+        DICTID -> DICT -> TYPE
+    Read deflate blocks:
+            TYPE -> STORED or TABLE or LEN or CHECK
+            STORED -> COPY -> TYPE
+            TABLE -> LENLENS -> CODELENS -> LEN
+    Read deflate codes:
+                LEN -> LENEXT or LIT or TYPE
+                LENEXT -> DIST -> DISTEXT -> MATCH -> LEN
+                LIT -> LEN
+    Process trailer:
+        CHECK -> LENGTH -> DONE
+ */
+
+/* state maintained between inflate() calls.  Approximately 7K bytes. */
+struct inflate_state {
+    inflate_mode mode;          /* current inflate mode */
+    int last;                   /* true if processing last block */
+    int wrap;                   /* bit 0 true for zlib, bit 1 true for gzip */
+    int havedict;               /* true if dictionary provided */
+    int flags;                  /* gzip header method and flags (0 if zlib) */
+    unsigned dmax;              /* zlib header max distance (INFLATE_STRICT) */
+    unsigned long check;        /* protected copy of check value */
+    unsigned long total;        /* protected copy of output count */
+    gz_headerp head;            /* where to save gzip header information */
+        /* sliding window */
+    unsigned wbits;             /* log base 2 of requested window size */
+    unsigned wsize;             /* window size or zero if not using window */
+    unsigned whave;             /* valid bytes in the window */
+    unsigned write;             /* window write index */
+    unsigned char FAR *window;  /* allocated sliding window, if needed */
+        /* bit accumulator */
+    unsigned long hold;         /* input bit accumulator */
+    unsigned bits;              /* number of bits in "in" */
+        /* for string and stored block copying */
+    unsigned length;            /* literal or length of data to copy */
+    unsigned offset;            /* distance back to copy string from */
+        /* for table and code decoding */
+    unsigned extra;             /* extra bits needed */
+        /* fixed and dynamic code tables */
+    code const FAR *lencode;    /* starting table for length/literal codes */
+    code const FAR *distcode;   /* starting table for distance codes */
+    unsigned lenbits;           /* index bits for lencode */
+    unsigned distbits;          /* index bits for distcode */
+        /* dynamic table building */
+    unsigned ncode;             /* number of code length code lengths */
+    unsigned nlen;              /* number of length code lengths */
+    unsigned ndist;             /* number of distance code lengths */
+    unsigned have;              /* number of code lengths in lens[] */
+    code FAR *next;             /* next available space in codes[] */
+    unsigned short lens[320];   /* temporary storage for code lengths */
+    unsigned short work[288];   /* work area for code table building */
+    code codes[ENOUGH];         /* space for code tables */
+};
diff --git a/distrib/zlib-1.2.3/inftrees.c b/distrib/zlib-1.2.3/inftrees.c
new file mode 100644
index 0000000..8a9c13f
--- /dev/null
+++ b/distrib/zlib-1.2.3/inftrees.c
@@ -0,0 +1,329 @@
+/* inftrees.c -- generate Huffman trees for efficient decoding
+ * Copyright (C) 1995-2005 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+#include "zutil.h"
+#include "inftrees.h"
+
+#define MAXBITS 15
+
+const char inflate_copyright[] =
+   " inflate 1.2.3 Copyright 1995-2005 Mark Adler ";
+/*
+  If you use the zlib library in a product, an acknowledgment is welcome
+  in the documentation of your product. If for some reason you cannot
+  include such an acknowledgment, I would appreciate that you keep this
+  copyright string in the executable of your product.
+ */
+
+/*
+   Build a set of tables to decode the provided canonical Huffman code.
+   The code lengths are lens[0..codes-1].  The result starts at *table,
+   whose indices are 0..2^bits-1.  work is a writable array of at least
+   lens shorts, which is used as a work area.  type is the type of code
+   to be generated, CODES, LENS, or DISTS.  On return, zero is success,
+   -1 is an invalid code, and +1 means that ENOUGH isn't enough.  table
+   on return points to the next available entry's address.  bits is the
+   requested root table index bits, and on return it is the actual root
+   table index bits.  It will differ if the request is greater than the
+   longest code or if it is less than the shortest code.
+ */
+int inflate_table(type, lens, codes, table, bits, work)
+codetype type;
+unsigned short FAR *lens;
+unsigned codes;
+code FAR * FAR *table;
+unsigned FAR *bits;
+unsigned short FAR *work;
+{
+    unsigned len;               /* a code's length in bits */
+    unsigned sym;               /* index of code symbols */
+    unsigned min, max;          /* minimum and maximum code lengths */
+    unsigned root;              /* number of index bits for root table */
+    unsigned curr;              /* number of index bits for current table */
+    unsigned drop;              /* code bits to drop for sub-table */
+    int left;                   /* number of prefix codes available */
+    unsigned used;              /* code entries in table used */
+    unsigned huff;              /* Huffman code */
+    unsigned incr;              /* for incrementing code, index */
+    unsigned fill;              /* index for replicating entries */
+    unsigned low;               /* low bits for current root entry */
+    unsigned mask;              /* mask for low root bits */
+    code this;                  /* table entry for duplication */
+    code FAR *next;             /* next available space in table */
+    const unsigned short FAR *base;     /* base value table to use */
+    const unsigned short FAR *extra;    /* extra bits table to use */
+    int end;                    /* use base and extra for symbol > end */
+    unsigned short count[MAXBITS+1];    /* number of codes of each length */
+    unsigned short offs[MAXBITS+1];     /* offsets in table for each length */
+    static const unsigned short lbase[31] = { /* Length codes 257..285 base */
+        3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
+        35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0};
+    static const unsigned short lext[31] = { /* Length codes 257..285 extra */
+        16, 16, 16, 16, 16, 16, 16, 16, 17, 17, 17, 17, 18, 18, 18, 18,
+        19, 19, 19, 19, 20, 20, 20, 20, 21, 21, 21, 21, 16, 201, 196};
+    static const unsigned short dbase[32] = { /* Distance codes 0..29 base */
+        1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
+        257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
+        8193, 12289, 16385, 24577, 0, 0};
+    static const unsigned short dext[32] = { /* Distance codes 0..29 extra */
+        16, 16, 16, 16, 17, 17, 18, 18, 19, 19, 20, 20, 21, 21, 22, 22,
+        23, 23, 24, 24, 25, 25, 26, 26, 27, 27,
+        28, 28, 29, 29, 64, 64};
+
+    /*
+       Process a set of code lengths to create a canonical Huffman code.  The
+       code lengths are lens[0..codes-1].  Each length corresponds to the
+       symbols 0..codes-1.  The Huffman code is generated by first sorting the
+       symbols by length from short to long, and retaining the symbol order
+       for codes with equal lengths.  Then the code starts with all zero bits
+       for the first code of the shortest length, and the codes are integer
+       increments for the same length, and zeros are appended as the length
+       increases.  For the deflate format, these bits are stored backwards
+       from their more natural integer increment ordering, and so when the
+       decoding tables are built in the large loop below, the integer codes
+       are incremented backwards.
+
+       This routine assumes, but does not check, that all of the entries in
+       lens[] are in the range 0..MAXBITS.  The caller must assure this.
+       1..MAXBITS is interpreted as that code length.  zero means that that
+       symbol does not occur in this code.
+
+       The codes are sorted by computing a count of codes for each length,
+       creating from that a table of starting indices for each length in the
+       sorted table, and then entering the symbols in order in the sorted
+       table.  The sorted table is work[], with that space being provided by
+       the caller.
+
+       The length counts are used for other purposes as well, i.e. finding
+       the minimum and maximum length codes, determining if there are any
+       codes at all, checking for a valid set of lengths, and looking ahead
+       at length counts to determine sub-table sizes when building the
+       decoding tables.
+     */
+
+    /* accumulate lengths for codes (assumes lens[] all in 0..MAXBITS) */
+    for (len = 0; len <= MAXBITS; len++)
+        count[len] = 0;
+    for (sym = 0; sym < codes; sym++)
+        count[lens[sym]]++;
+
+    /* bound code lengths, force root to be within code lengths */
+    root = *bits;
+    for (max = MAXBITS; max >= 1; max--)
+        if (count[max] != 0) break;
+    if (root > max) root = max;
+    if (max == 0) {                     /* no symbols to code at all */
+        this.op = (unsigned char)64;    /* invalid code marker */
+        this.bits = (unsigned char)1;
+        this.val = (unsigned short)0;
+        *(*table)++ = this;             /* make a table to force an error */
+        *(*table)++ = this;
+        *bits = 1;
+        return 0;     /* no symbols, but wait for decoding to report error */
+    }
+    for (min = 1; min <= MAXBITS; min++)
+        if (count[min] != 0) break;
+    if (root < min) root = min;
+
+    /* check for an over-subscribed or incomplete set of lengths */
+    left = 1;
+    for (len = 1; len <= MAXBITS; len++) {
+        left <<= 1;
+        left -= count[len];
+        if (left < 0) return -1;        /* over-subscribed */
+    }
+    if (left > 0 && (type == CODES || max != 1))
+        return -1;                      /* incomplete set */
+
+    /* generate offsets into symbol table for each length for sorting */
+    offs[1] = 0;
+    for (len = 1; len < MAXBITS; len++)
+        offs[len + 1] = offs[len] + count[len];
+
+    /* sort symbols by length, by symbol order within each length */
+    for (sym = 0; sym < codes; sym++)
+        if (lens[sym] != 0) work[offs[lens[sym]]++] = (unsigned short)sym;
+
+    /*
+       Create and fill in decoding tables.  In this loop, the table being
+       filled is at next and has curr index bits.  The code being used is huff
+       with length len.  That code is converted to an index by dropping drop
+       bits off of the bottom.  For codes where len is less than drop + curr,
+       those top drop + curr - len bits are incremented through all values to
+       fill the table with replicated entries.
+
+       root is the number of index bits for the root table.  When len exceeds
+       root, sub-tables are created pointed to by the root entry with an index
+       of the low root bits of huff.  This is saved in low to check for when a
+       new sub-table should be started.  drop is zero when the root table is
+       being filled, and drop is root when sub-tables are being filled.
+
+       When a new sub-table is needed, it is necessary to look ahead in the
+       code lengths to determine what size sub-table is needed.  The length
+       counts are used for this, and so count[] is decremented as codes are
+       entered in the tables.
+
+       used keeps track of how many table entries have been allocated from the
+       provided *table space.  It is checked when a LENS table is being made
+       against the space in *table, ENOUGH, minus the maximum space needed by
+       the worst case distance code, MAXD.  This should never happen, but the
+       sufficiency of ENOUGH has not been proven exhaustively, hence the check.
+       This assumes that when type == LENS, bits == 9.
+
+       sym increments through all symbols, and the loop terminates when
+       all codes of length max, i.e. all codes, have been processed.  This
+       routine permits incomplete codes, so another loop after this one fills
+       in the rest of the decoding tables with invalid code markers.
+     */
+
+    /* set up for code type */
+    switch (type) {
+    case CODES:
+        base = extra = work;    /* dummy value--not used */
+        end = 19;
+        break;
+    case LENS:
+        base = lbase;
+        base -= 257;
+        extra = lext;
+        extra -= 257;
+        end = 256;
+        break;
+    default:            /* DISTS */
+        base = dbase;
+        extra = dext;
+        end = -1;
+    }
+
+    /* initialize state for loop */
+    huff = 0;                   /* starting code */
+    sym = 0;                    /* starting code symbol */
+    len = min;                  /* starting code length */
+    next = *table;              /* current table to fill in */
+    curr = root;                /* current table index bits */
+    drop = 0;                   /* current bits to drop from code for index */
+    low = (unsigned)(-1);       /* trigger new sub-table when len > root */
+    used = 1U << root;          /* use root table entries */
+    mask = used - 1;            /* mask for comparing low */
+
+    /* check available table space */
+    if (type == LENS && used >= ENOUGH - MAXD)
+        return 1;
+
+    /* process all codes and make table entries */
+    for (;;) {
+        /* create table entry */
+        this.bits = (unsigned char)(len - drop);
+        if ((int)(work[sym]) < end) {
+            this.op = (unsigned char)0;
+            this.val = work[sym];
+        }
+        else if ((int)(work[sym]) > end) {
+            this.op = (unsigned char)(extra[work[sym]]);
+            this.val = base[work[sym]];
+        }
+        else {
+            this.op = (unsigned char)(32 + 64);         /* end of block */
+            this.val = 0;
+        }
+
+        /* replicate for those indices with low len bits equal to huff */
+        incr = 1U << (len - drop);
+        fill = 1U << curr;
+        min = fill;                 /* save offset to next table */
+        do {
+            fill -= incr;
+            next[(huff >> drop) + fill] = this;
+        } while (fill != 0);
+
+        /* backwards increment the len-bit code huff */
+        incr = 1U << (len - 1);
+        while (huff & incr)
+            incr >>= 1;
+        if (incr != 0) {
+            huff &= incr - 1;
+            huff += incr;
+        }
+        else
+            huff = 0;
+
+        /* go to next symbol, update count, len */
+        sym++;
+        if (--(count[len]) == 0) {
+            if (len == max) break;
+            len = lens[work[sym]];
+        }
+
+        /* create new sub-table if needed */
+        if (len > root && (huff & mask) != low) {
+            /* if first time, transition to sub-tables */
+            if (drop == 0)
+                drop = root;
+
+            /* increment past last table */
+            next += min;            /* here min is 1 << curr */
+
+            /* determine length of next table */
+            curr = len - drop;
+            left = (int)(1 << curr);
+            while (curr + drop < max) {
+                left -= count[curr + drop];
+                if (left <= 0) break;
+                curr++;
+                left <<= 1;
+            }
+
+            /* check for enough space */
+            used += 1U << curr;
+            if (type == LENS && used >= ENOUGH - MAXD)
+                return 1;
+
+            /* point entry in root table to sub-table */
+            low = huff & mask;
+            (*table)[low].op = (unsigned char)curr;
+            (*table)[low].bits = (unsigned char)root;
+            (*table)[low].val = (unsigned short)(next - *table);
+        }
+    }
+
+    /*
+       Fill in rest of table for incomplete codes.  This loop is similar to the
+       loop above in incrementing huff for table indices.  It is assumed that
+       len is equal to curr + drop, so there is no loop needed to increment
+       through high index bits.  When the current sub-table is filled, the loop
+       drops back to the root table to fill in any remaining entries there.
+     */
+    this.op = (unsigned char)64;                /* invalid code marker */
+    this.bits = (unsigned char)(len - drop);
+    this.val = (unsigned short)0;
+    while (huff != 0) {
+        /* when done with sub-table, drop back to root table */
+        if (drop != 0 && (huff & mask) != low) {
+            drop = 0;
+            len = root;
+            next = *table;
+            this.bits = (unsigned char)len;
+        }
+
+        /* put invalid code marker in table */
+        next[huff >> drop] = this;
+
+        /* backwards increment the len-bit code huff */
+        incr = 1U << (len - 1);
+        while (huff & incr)
+            incr >>= 1;
+        if (incr != 0) {
+            huff &= incr - 1;
+            huff += incr;
+        }
+        else
+            huff = 0;
+    }
+
+    /* set return parameters */
+    *table += used;
+    *bits = root;
+    return 0;
+}
diff --git a/distrib/zlib-1.2.3/inftrees.h b/distrib/zlib-1.2.3/inftrees.h
new file mode 100644
index 0000000..b1104c8
--- /dev/null
+++ b/distrib/zlib-1.2.3/inftrees.h
@@ -0,0 +1,55 @@
+/* inftrees.h -- header to use inftrees.c
+ * Copyright (C) 1995-2005 Mark Adler
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* WARNING: this file should *not* be used by applications. It is
+   part of the implementation of the compression library and is
+   subject to change. Applications should only use zlib.h.
+ */
+
+/* Structure for decoding tables.  Each entry provides either the
+   information needed to do the operation requested by the code that
+   indexed that table entry, or it provides a pointer to another
+   table that indexes more bits of the code.  op indicates whether
+   the entry is a pointer to another table, a literal, a length or
+   distance, an end-of-block, or an invalid code.  For a table
+   pointer, the low four bits of op is the number of index bits of
+   that table.  For a length or distance, the low four bits of op
+   is the number of extra bits to get after the code.  bits is
+   the number of bits in this code or part of the code to drop off
+   of the bit buffer.  val is the actual byte to output in the case
+   of a literal, the base length or distance, or the offset from
+   the current table to the next table.  Each entry is four bytes. */
+typedef struct {
+    unsigned char op;           /* operation, extra bits, table bits */
+    unsigned char bits;         /* bits in this part of the code */
+    unsigned short val;         /* offset in table or code value */
+} code;
+
+/* op values as set by inflate_table():
+    00000000 - literal
+    0000tttt - table link, tttt != 0 is the number of table index bits
+    0001eeee - length or distance, eeee is the number of extra bits
+    01100000 - end of block
+    01000000 - invalid code
+ */
+
+/* Maximum size of dynamic tree.  The maximum found in a long but non-
+   exhaustive search was 1444 code structures (852 for length/literals
+   and 592 for distances, the latter actually the result of an
+   exhaustive search).  The true maximum is not known, but the value
+   below is more than safe. */
+#define ENOUGH 2048
+#define MAXD 592
+
+/* Type of code to build for inftable() */
+typedef enum {
+    CODES,
+    LENS,
+    DISTS
+} codetype;
+
+extern int inflate_table OF((codetype type, unsigned short FAR *lens,
+                             unsigned codes, code FAR * FAR *table,
+                             unsigned FAR *bits, unsigned short FAR *work));
diff --git a/distrib/zlib-1.2.3/sources.make b/distrib/zlib-1.2.3/sources.make
new file mode 100644
index 0000000..c15fd78
--- /dev/null
+++ b/distrib/zlib-1.2.3/sources.make
@@ -0,0 +1,4 @@
+# this is included by various Makefiles
+ZLIB_SOURCES := adler32.c compress.c crc32.c deflate.c gzio.c infback.c inffast.c inflate.c inftrees.c trees.c uncompr.c zutil.c
+ZLIB_SOURCES := $(ZLIB_SOURCES:%=$(ZLIB_DIR)/%)
+
diff --git a/distrib/zlib-1.2.3/trees.c b/distrib/zlib-1.2.3/trees.c
new file mode 100644
index 0000000..395e4e1
--- /dev/null
+++ b/distrib/zlib-1.2.3/trees.c
@@ -0,0 +1,1219 @@
+/* trees.c -- output deflated data using Huffman coding
+ * Copyright (C) 1995-2005 Jean-loup Gailly
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/*
+ *  ALGORITHM
+ *
+ *      The "deflation" process uses several Huffman trees. The more
+ *      common source values are represented by shorter bit sequences.
+ *
+ *      Each code tree is stored in a compressed form which is itself
+ * a Huffman encoding of the lengths of all the code strings (in
+ * ascending order by source values).  The actual code strings are
+ * reconstructed from the lengths in the inflate process, as described
+ * in the deflate specification.
+ *
+ *  REFERENCES
+ *
+ *      Deutsch, L.P.,"'Deflate' Compressed Data Format Specification".
+ *      Available in ftp.uu.net:/pub/archiving/zip/doc/deflate-1.1.doc
+ *
+ *      Storer, James A.
+ *          Data Compression:  Methods and Theory, pp. 49-50.
+ *          Computer Science Press, 1988.  ISBN 0-7167-8156-5.
+ *
+ *      Sedgewick, R.
+ *          Algorithms, p290.
+ *          Addison-Wesley, 1983. ISBN 0-201-06672-6.
+ */
+
+/* @(#) $Id$ */
+
+/* #define GEN_TREES_H */
+
+#include "deflate.h"
+
+#ifdef DEBUG
+#  include <ctype.h>
+#endif
+
+/* ===========================================================================
+ * Constants
+ */
+
+#define MAX_BL_BITS 7
+/* Bit length codes must not exceed MAX_BL_BITS bits */
+
+#define END_BLOCK 256
+/* end of block literal code */
+
+#define REP_3_6      16
+/* repeat previous bit length 3-6 times (2 bits of repeat count) */
+
+#define REPZ_3_10    17
+/* repeat a zero length 3-10 times  (3 bits of repeat count) */
+
+#define REPZ_11_138  18
+/* repeat a zero length 11-138 times  (7 bits of repeat count) */
+
+local const int extra_lbits[LENGTH_CODES] /* extra bits for each length code */
+   = {0,0,0,0,0,0,0,0,1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,5,0};
+
+local const int extra_dbits[D_CODES] /* extra bits for each distance code */
+   = {0,0,0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9,10,10,11,11,12,12,13,13};
+
+local const int extra_blbits[BL_CODES]/* extra bits for each bit length code */
+   = {0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,3,7};
+
+local const uch bl_order[BL_CODES]
+   = {16,17,18,0,8,7,9,6,10,5,11,4,12,3,13,2,14,1,15};
+/* The lengths of the bit length codes are sent in order of decreasing
+ * probability, to avoid transmitting the lengths for unused bit length codes.
+ */
+
+#define Buf_size (8 * 2*sizeof(char))
+/* Number of bits used within bi_buf. (bi_buf might be implemented on
+ * more than 16 bits on some systems.)
+ */
+
+/* ===========================================================================
+ * Local data. These are initialized only once.
+ */
+
+#define DIST_CODE_LEN  512 /* see definition of array dist_code below */
+
+#if defined(GEN_TREES_H) || !defined(STDC)
+/* non ANSI compilers may not accept trees.h */
+
+local ct_data static_ltree[L_CODES+2];
+/* The static literal tree. Since the bit lengths are imposed, there is no
+ * need for the L_CODES extra codes used during heap construction. However
+ * The codes 286 and 287 are needed to build a canonical tree (see _tr_init
+ * below).
+ */
+
+local ct_data static_dtree[D_CODES];
+/* The static distance tree. (Actually a trivial tree since all codes use
+ * 5 bits.)
+ */
+
+uch _dist_code[DIST_CODE_LEN];
+/* Distance codes. The first 256 values correspond to the distances
+ * 3 .. 258, the last 256 values correspond to the top 8 bits of
+ * the 15 bit distances.
+ */
+
+uch _length_code[MAX_MATCH-MIN_MATCH+1];
+/* length code for each normalized match length (0 == MIN_MATCH) */
+
+local int base_length[LENGTH_CODES];
+/* First normalized length for each code (0 = MIN_MATCH) */
+
+local int base_dist[D_CODES];
+/* First normalized distance for each code (0 = distance of 1) */
+
+#else
+#  include "trees.h"
+#endif /* GEN_TREES_H */
+
+struct static_tree_desc_s {
+    const ct_data *static_tree;  /* static tree or NULL */
+    const intf *extra_bits;      /* extra bits for each code or NULL */
+    int     extra_base;          /* base index for extra_bits */
+    int     elems;               /* max number of elements in the tree */
+    int     max_length;          /* max bit length for the codes */
+};
+
+local static_tree_desc  static_l_desc =
+{static_ltree, extra_lbits, LITERALS+1, L_CODES, MAX_BITS};
+
+local static_tree_desc  static_d_desc =
+{static_dtree, extra_dbits, 0,          D_CODES, MAX_BITS};
+
+local static_tree_desc  static_bl_desc =
+{(const ct_data *)0, extra_blbits, 0,   BL_CODES, MAX_BL_BITS};
+
+/* ===========================================================================
+ * Local (static) routines in this file.
+ */
+
+local void tr_static_init OF((void));
+local void init_block     OF((deflate_state *s));
+local void pqdownheap     OF((deflate_state *s, ct_data *tree, int k));
+local void gen_bitlen     OF((deflate_state *s, tree_desc *desc));
+local void gen_codes      OF((ct_data *tree, int max_code, ushf *bl_count));
+local void build_tree     OF((deflate_state *s, tree_desc *desc));
+local void scan_tree      OF((deflate_state *s, ct_data *tree, int max_code));
+local void send_tree      OF((deflate_state *s, ct_data *tree, int max_code));
+local int  build_bl_tree  OF((deflate_state *s));
+local void send_all_trees OF((deflate_state *s, int lcodes, int dcodes,
+                              int blcodes));
+local void compress_block OF((deflate_state *s, ct_data *ltree,
+                              ct_data *dtree));
+local void set_data_type  OF((deflate_state *s));
+local unsigned bi_reverse OF((unsigned value, int length));
+local void bi_windup      OF((deflate_state *s));
+local void bi_flush       OF((deflate_state *s));
+local void copy_block     OF((deflate_state *s, charf *buf, unsigned len,
+                              int header));
+
+#ifdef GEN_TREES_H
+local void gen_trees_header OF((void));
+#endif
+
+#ifndef DEBUG
+#  define send_code(s, c, tree) send_bits(s, tree[c].Code, tree[c].Len)
+   /* Send a code of the given tree. c and tree must not have side effects */
+
+#else /* DEBUG */
+#  define send_code(s, c, tree) \
+     { if (z_verbose>2) fprintf(stderr,"\ncd %3d ",(c)); \
+       send_bits(s, tree[c].Code, tree[c].Len); }
+#endif
+
+/* ===========================================================================
+ * Output a short LSB first on the stream.
+ * IN assertion: there is enough room in pendingBuf.
+ */
+#define put_short(s, w) { \
+    put_byte(s, (uch)((w) & 0xff)); \
+    put_byte(s, (uch)((ush)(w) >> 8)); \
+}
+
+/* ===========================================================================
+ * Send a value on a given number of bits.
+ * IN assertion: length <= 16 and value fits in length bits.
+ */
+#ifdef DEBUG
+local void send_bits      OF((deflate_state *s, int value, int length));
+
+local void send_bits(s, value, length)
+    deflate_state *s;
+    int value;  /* value to send */
+    int length; /* number of bits */
+{
+    Tracevv((stderr," l %2d v %4x ", length, value));
+    Assert(length > 0 && length <= 15, "invalid length");
+    s->bits_sent += (ulg)length;
+
+    /* If not enough room in bi_buf, use (valid) bits from bi_buf and
+     * (16 - bi_valid) bits from value, leaving (width - (16-bi_valid))
+     * unused bits in value.
+     */
+    if (s->bi_valid > (int)Buf_size - length) {
+        s->bi_buf |= (value << s->bi_valid);
+        put_short(s, s->bi_buf);
+        s->bi_buf = (ush)value >> (Buf_size - s->bi_valid);
+        s->bi_valid += length - Buf_size;
+    } else {
+        s->bi_buf |= value << s->bi_valid;
+        s->bi_valid += length;
+    }
+}
+#else /* !DEBUG */
+
+#define send_bits(s, value, length) \
+{ int len = length;\
+  if (s->bi_valid > (int)Buf_size - len) {\
+    int val = value;\
+    s->bi_buf |= (val << s->bi_valid);\
+    put_short(s, s->bi_buf);\
+    s->bi_buf = (ush)val >> (Buf_size - s->bi_valid);\
+    s->bi_valid += len - Buf_size;\
+  } else {\
+    s->bi_buf |= (value) << s->bi_valid;\
+    s->bi_valid += len;\
+  }\
+}
+#endif /* DEBUG */
+
+
+/* the arguments must not have side effects */
+
+/* ===========================================================================
+ * Initialize the various 'constant' tables.
+ */
+local void tr_static_init()
+{
+#if defined(GEN_TREES_H) || !defined(STDC)
+    static int static_init_done = 0;
+    int n;        /* iterates over tree elements */
+    int bits;     /* bit counter */
+    int length;   /* length value */
+    int code;     /* code value */
+    int dist;     /* distance index */
+    ush bl_count[MAX_BITS+1];
+    /* number of codes at each bit length for an optimal tree */
+
+    if (static_init_done) return;
+
+    /* For some embedded targets, global variables are not initialized: */
+    static_l_desc.static_tree = static_ltree;
+    static_l_desc.extra_bits = extra_lbits;
+    static_d_desc.static_tree = static_dtree;
+    static_d_desc.extra_bits = extra_dbits;
+    static_bl_desc.extra_bits = extra_blbits;
+
+    /* Initialize the mapping length (0..255) -> length code (0..28) */
+    length = 0;
+    for (code = 0; code < LENGTH_CODES-1; code++) {
+        base_length[code] = length;
+        for (n = 0; n < (1<<extra_lbits[code]); n++) {
+            _length_code[length++] = (uch)code;
+        }
+    }
+    Assert (length == 256, "tr_static_init: length != 256");
+    /* Note that the length 255 (match length 258) can be represented
+     * in two different ways: code 284 + 5 bits or code 285, so we
+     * overwrite length_code[255] to use the best encoding:
+     */
+    _length_code[length-1] = (uch)code;
+
+    /* Initialize the mapping dist (0..32K) -> dist code (0..29) */
+    dist = 0;
+    for (code = 0 ; code < 16; code++) {
+        base_dist[code] = dist;
+        for (n = 0; n < (1<<extra_dbits[code]); n++) {
+            _dist_code[dist++] = (uch)code;
+        }
+    }
+    Assert (dist == 256, "tr_static_init: dist != 256");
+    dist >>= 7; /* from now on, all distances are divided by 128 */
+    for ( ; code < D_CODES; code++) {
+        base_dist[code] = dist << 7;
+        for (n = 0; n < (1<<(extra_dbits[code]-7)); n++) {
+            _dist_code[256 + dist++] = (uch)code;
+        }
+    }
+    Assert (dist == 256, "tr_static_init: 256+dist != 512");
+
+    /* Construct the codes of the static literal tree */
+    for (bits = 0; bits <= MAX_BITS; bits++) bl_count[bits] = 0;
+    n = 0;
+    while (n <= 143) static_ltree[n++].Len = 8, bl_count[8]++;
+    while (n <= 255) static_ltree[n++].Len = 9, bl_count[9]++;
+    while (n <= 279) static_ltree[n++].Len = 7, bl_count[7]++;
+    while (n <= 287) static_ltree[n++].Len = 8, bl_count[8]++;
+    /* Codes 286 and 287 do not exist, but we must include them in the
+     * tree construction to get a canonical Huffman tree (longest code
+     * all ones)
+     */
+    gen_codes((ct_data *)static_ltree, L_CODES+1, bl_count);
+
+    /* The static distance tree is trivial: */
+    for (n = 0; n < D_CODES; n++) {
+        static_dtree[n].Len = 5;
+        static_dtree[n].Code = bi_reverse((unsigned)n, 5);
+    }
+    static_init_done = 1;
+
+#  ifdef GEN_TREES_H
+    gen_trees_header();
+#  endif
+#endif /* defined(GEN_TREES_H) || !defined(STDC) */
+}
+
+/* ===========================================================================
+ * Genererate the file trees.h describing the static trees.
+ */
+#ifdef GEN_TREES_H
+#  ifndef DEBUG
+#    include <stdio.h>
+#  endif
+
+#  define SEPARATOR(i, last, width) \
+      ((i) == (last)? "\n};\n\n" :    \
+       ((i) % (width) == (width)-1 ? ",\n" : ", "))
+
+void gen_trees_header()
+{
+    FILE *header = fopen("trees.h", "w");
+    int i;
+
+    Assert (header != NULL, "Can't open trees.h");
+    fprintf(header,
+            "/* header created automatically with -DGEN_TREES_H */\n\n");
+
+    fprintf(header, "local const ct_data static_ltree[L_CODES+2] = {\n");
+    for (i = 0; i < L_CODES+2; i++) {
+        fprintf(header, "{{%3u},{%3u}}%s", static_ltree[i].Code,
+                static_ltree[i].Len, SEPARATOR(i, L_CODES+1, 5));
+    }
+
+    fprintf(header, "local const ct_data static_dtree[D_CODES] = {\n");
+    for (i = 0; i < D_CODES; i++) {
+        fprintf(header, "{{%2u},{%2u}}%s", static_dtree[i].Code,
+                static_dtree[i].Len, SEPARATOR(i, D_CODES-1, 5));
+    }
+
+    fprintf(header, "const uch _dist_code[DIST_CODE_LEN] = {\n");
+    for (i = 0; i < DIST_CODE_LEN; i++) {
+        fprintf(header, "%2u%s", _dist_code[i],
+                SEPARATOR(i, DIST_CODE_LEN-1, 20));
+    }
+
+    fprintf(header, "const uch _length_code[MAX_MATCH-MIN_MATCH+1]= {\n");
+    for (i = 0; i < MAX_MATCH-MIN_MATCH+1; i++) {
+        fprintf(header, "%2u%s", _length_code[i],
+                SEPARATOR(i, MAX_MATCH-MIN_MATCH, 20));
+    }
+
+    fprintf(header, "local const int base_length[LENGTH_CODES] = {\n");
+    for (i = 0; i < LENGTH_CODES; i++) {
+        fprintf(header, "%1u%s", base_length[i],
+                SEPARATOR(i, LENGTH_CODES-1, 20));
+    }
+
+    fprintf(header, "local const int base_dist[D_CODES] = {\n");
+    for (i = 0; i < D_CODES; i++) {
+        fprintf(header, "%5u%s", base_dist[i],
+                SEPARATOR(i, D_CODES-1, 10));
+    }
+
+    fclose(header);
+}
+#endif /* GEN_TREES_H */
+
+/* ===========================================================================
+ * Initialize the tree data structures for a new zlib stream.
+ */
+void _tr_init(s)
+    deflate_state *s;
+{
+    tr_static_init();
+
+    s->l_desc.dyn_tree = s->dyn_ltree;
+    s->l_desc.stat_desc = &static_l_desc;
+
+    s->d_desc.dyn_tree = s->dyn_dtree;
+    s->d_desc.stat_desc = &static_d_desc;
+
+    s->bl_desc.dyn_tree = s->bl_tree;
+    s->bl_desc.stat_desc = &static_bl_desc;
+
+    s->bi_buf = 0;
+    s->bi_valid = 0;
+    s->last_eob_len = 8; /* enough lookahead for inflate */
+#ifdef DEBUG
+    s->compressed_len = 0L;
+    s->bits_sent = 0L;
+#endif
+
+    /* Initialize the first block of the first file: */
+    init_block(s);
+}
+
+/* ===========================================================================
+ * Initialize a new block.
+ */
+local void init_block(s)
+    deflate_state *s;
+{
+    int n; /* iterates over tree elements */
+
+    /* Initialize the trees. */
+    for (n = 0; n < L_CODES;  n++) s->dyn_ltree[n].Freq = 0;
+    for (n = 0; n < D_CODES;  n++) s->dyn_dtree[n].Freq = 0;
+    for (n = 0; n < BL_CODES; n++) s->bl_tree[n].Freq = 0;
+
+    s->dyn_ltree[END_BLOCK].Freq = 1;
+    s->opt_len = s->static_len = 0L;
+    s->last_lit = s->matches = 0;
+}
+
+#define SMALLEST 1
+/* Index within the heap array of least frequent node in the Huffman tree */
+
+
+/* ===========================================================================
+ * Remove the smallest element from the heap and recreate the heap with
+ * one less element. Updates heap and heap_len.
+ */
+#define pqremove(s, tree, top) \
+{\
+    top = s->heap[SMALLEST]; \
+    s->heap[SMALLEST] = s->heap[s->heap_len--]; \
+    pqdownheap(s, tree, SMALLEST); \
+}
+
+/* ===========================================================================
+ * Compares to subtrees, using the tree depth as tie breaker when
+ * the subtrees have equal frequency. This minimizes the worst case length.
+ */
+#define smaller(tree, n, m, depth) \
+   (tree[n].Freq < tree[m].Freq || \
+   (tree[n].Freq == tree[m].Freq && depth[n] <= depth[m]))
+
+/* ===========================================================================
+ * Restore the heap property by moving down the tree starting at node k,
+ * exchanging a node with the smallest of its two sons if necessary, stopping
+ * when the heap property is re-established (each father smaller than its
+ * two sons).
+ */
+local void pqdownheap(s, tree, k)
+    deflate_state *s;
+    ct_data *tree;  /* the tree to restore */
+    int k;               /* node to move down */
+{
+    int v = s->heap[k];
+    int j = k << 1;  /* left son of k */
+    while (j <= s->heap_len) {
+        /* Set j to the smallest of the two sons: */
+        if (j < s->heap_len &&
+            smaller(tree, s->heap[j+1], s->heap[j], s->depth)) {
+            j++;
+        }
+        /* Exit if v is smaller than both sons */
+        if (smaller(tree, v, s->heap[j], s->depth)) break;
+
+        /* Exchange v with the smallest son */
+        s->heap[k] = s->heap[j];  k = j;
+
+        /* And continue down the tree, setting j to the left son of k */
+        j <<= 1;
+    }
+    s->heap[k] = v;
+}
+
+/* ===========================================================================
+ * Compute the optimal bit lengths for a tree and update the total bit length
+ * for the current block.
+ * IN assertion: the fields freq and dad are set, heap[heap_max] and
+ *    above are the tree nodes sorted by increasing frequency.
+ * OUT assertions: the field len is set to the optimal bit length, the
+ *     array bl_count contains the frequencies for each bit length.
+ *     The length opt_len is updated; static_len is also updated if stree is
+ *     not null.
+ */
+local void gen_bitlen(s, desc)
+    deflate_state *s;
+    tree_desc *desc;    /* the tree descriptor */
+{
+    ct_data *tree        = desc->dyn_tree;
+    int max_code         = desc->max_code;
+    const ct_data *stree = desc->stat_desc->static_tree;
+    const intf *extra    = desc->stat_desc->extra_bits;
+    int base             = desc->stat_desc->extra_base;
+    int max_length       = desc->stat_desc->max_length;
+    int h;              /* heap index */
+    int n, m;           /* iterate over the tree elements */
+    int bits;           /* bit length */
+    int xbits;          /* extra bits */
+    ush f;              /* frequency */
+    int overflow = 0;   /* number of elements with bit length too large */
+
+    for (bits = 0; bits <= MAX_BITS; bits++) s->bl_count[bits] = 0;
+
+    /* In a first pass, compute the optimal bit lengths (which may
+     * overflow in the case of the bit length tree).
+     */
+    tree[s->heap[s->heap_max]].Len = 0; /* root of the heap */
+
+    for (h = s->heap_max+1; h < HEAP_SIZE; h++) {
+        n = s->heap[h];
+        bits = tree[tree[n].Dad].Len + 1;
+        if (bits > max_length) bits = max_length, overflow++;
+        tree[n].Len = (ush)bits;
+        /* We overwrite tree[n].Dad which is no longer needed */
+
+        if (n > max_code) continue; /* not a leaf node */
+
+        s->bl_count[bits]++;
+        xbits = 0;
+        if (n >= base) xbits = extra[n-base];
+        f = tree[n].Freq;
+        s->opt_len += (ulg)f * (bits + xbits);
+        if (stree) s->static_len += (ulg)f * (stree[n].Len + xbits);
+    }
+    if (overflow == 0) return;
+
+    Trace((stderr,"\nbit length overflow\n"));
+    /* This happens for example on obj2 and pic of the Calgary corpus */
+
+    /* Find the first bit length which could increase: */
+    do {
+        bits = max_length-1;
+        while (s->bl_count[bits] == 0) bits--;
+        s->bl_count[bits]--;      /* move one leaf down the tree */
+        s->bl_count[bits+1] += 2; /* move one overflow item as its brother */
+        s->bl_count[max_length]--;
+        /* The brother of the overflow item also moves one step up,
+         * but this does not affect bl_count[max_length]
+         */
+        overflow -= 2;
+    } while (overflow > 0);
+
+    /* Now recompute all bit lengths, scanning in increasing frequency.
+     * h is still equal to HEAP_SIZE. (It is simpler to reconstruct all
+     * lengths instead of fixing only the wrong ones. This idea is taken
+     * from 'ar' written by Haruhiko Okumura.)
+     */
+    for (bits = max_length; bits != 0; bits--) {
+        n = s->bl_count[bits];
+        while (n != 0) {
+            m = s->heap[--h];
+            if (m > max_code) continue;
+            if ((unsigned) tree[m].Len != (unsigned) bits) {
+                Trace((stderr,"code %d bits %d->%d\n", m, tree[m].Len, bits));
+                s->opt_len += ((long)bits - (long)tree[m].Len)
+                              *(long)tree[m].Freq;
+                tree[m].Len = (ush)bits;
+            }
+            n--;
+        }
+    }
+}
+
+/* ===========================================================================
+ * Generate the codes for a given tree and bit counts (which need not be
+ * optimal).
+ * IN assertion: the array bl_count contains the bit length statistics for
+ * the given tree and the field len is set for all tree elements.
+ * OUT assertion: the field code is set for all tree elements of non
+ *     zero code length.
+ */
+local void gen_codes (tree, max_code, bl_count)
+    ct_data *tree;             /* the tree to decorate */
+    int max_code;              /* largest code with non zero frequency */
+    ushf *bl_count;            /* number of codes at each bit length */
+{
+    ush next_code[MAX_BITS+1]; /* next code value for each bit length */
+    ush code = 0;              /* running code value */
+    int bits;                  /* bit index */
+    int n;                     /* code index */
+
+    /* The distribution counts are first used to generate the code values
+     * without bit reversal.
+     */
+    for (bits = 1; bits <= MAX_BITS; bits++) {
+        next_code[bits] = code = (code + bl_count[bits-1]) << 1;
+    }
+    /* Check that the bit counts in bl_count are consistent. The last code
+     * must be all ones.
+     */
+    Assert (code + bl_count[MAX_BITS]-1 == (1<<MAX_BITS)-1,
+            "inconsistent bit counts");
+    Tracev((stderr,"\ngen_codes: max_code %d ", max_code));
+
+    for (n = 0;  n <= max_code; n++) {
+        int len = tree[n].Len;
+        if (len == 0) continue;
+        /* Now reverse the bits */
+        tree[n].Code = bi_reverse(next_code[len]++, len);
+
+        Tracecv(tree != static_ltree, (stderr,"\nn %3d %c l %2d c %4x (%x) ",
+             n, (isgraph(n) ? n : ' '), len, tree[n].Code, next_code[len]-1));
+    }
+}
+
+/* ===========================================================================
+ * Construct one Huffman tree and assigns the code bit strings and lengths.
+ * Update the total bit length for the current block.
+ * IN assertion: the field freq is set for all tree elements.
+ * OUT assertions: the fields len and code are set to the optimal bit length
+ *     and corresponding code. The length opt_len is updated; static_len is
+ *     also updated if stree is not null. The field max_code is set.
+ */
+local void build_tree(s, desc)
+    deflate_state *s;
+    tree_desc *desc; /* the tree descriptor */
+{
+    ct_data *tree         = desc->dyn_tree;
+    const ct_data *stree  = desc->stat_desc->static_tree;
+    int elems             = desc->stat_desc->elems;
+    int n, m;          /* iterate over heap elements */
+    int max_code = -1; /* largest code with non zero frequency */
+    int node;          /* new node being created */
+
+    /* Construct the initial heap, with least frequent element in
+     * heap[SMALLEST]. The sons of heap[n] are heap[2*n] and heap[2*n+1].
+     * heap[0] is not used.
+     */
+    s->heap_len = 0, s->heap_max = HEAP_SIZE;
+
+    for (n = 0; n < elems; n++) {
+        if (tree[n].Freq != 0) {
+            s->heap[++(s->heap_len)] = max_code = n;
+            s->depth[n] = 0;
+        } else {
+            tree[n].Len = 0;
+        }
+    }
+
+    /* The pkzip format requires that at least one distance code exists,
+     * and that at least one bit should be sent even if there is only one
+     * possible code. So to avoid special checks later on we force at least
+     * two codes of non zero frequency.
+     */
+    while (s->heap_len < 2) {
+        node = s->heap[++(s->heap_len)] = (max_code < 2 ? ++max_code : 0);
+        tree[node].Freq = 1;
+        s->depth[node] = 0;
+        s->opt_len--; if (stree) s->static_len -= stree[node].Len;
+        /* node is 0 or 1 so it does not have extra bits */
+    }
+    desc->max_code = max_code;
+
+    /* The elements heap[heap_len/2+1 .. heap_len] are leaves of the tree,
+     * establish sub-heaps of increasing lengths:
+     */
+    for (n = s->heap_len/2; n >= 1; n--) pqdownheap(s, tree, n);
+
+    /* Construct the Huffman tree by repeatedly combining the least two
+     * frequent nodes.
+     */
+    node = elems;              /* next internal node of the tree */
+    do {
+        pqremove(s, tree, n);  /* n = node of least frequency */
+        m = s->heap[SMALLEST]; /* m = node of next least frequency */
+
+        s->heap[--(s->heap_max)] = n; /* keep the nodes sorted by frequency */
+        s->heap[--(s->heap_max)] = m;
+
+        /* Create a new node father of n and m */
+        tree[node].Freq = tree[n].Freq + tree[m].Freq;
+        s->depth[node] = (uch)((s->depth[n] >= s->depth[m] ?
+                                s->depth[n] : s->depth[m]) + 1);
+        tree[n].Dad = tree[m].Dad = (ush)node;
+#ifdef DUMP_BL_TREE
+        if (tree == s->bl_tree) {
+            fprintf(stderr,"\nnode %d(%d), sons %d(%d) %d(%d)",
+                    node, tree[node].Freq, n, tree[n].Freq, m, tree[m].Freq);
+        }
+#endif
+        /* and insert the new node in the heap */
+        s->heap[SMALLEST] = node++;
+        pqdownheap(s, tree, SMALLEST);
+
+    } while (s->heap_len >= 2);
+
+    s->heap[--(s->heap_max)] = s->heap[SMALLEST];
+
+    /* At this point, the fields freq and dad are set. We can now
+     * generate the bit lengths.
+     */
+    gen_bitlen(s, (tree_desc *)desc);
+
+    /* The field len is now set, we can generate the bit codes */
+    gen_codes ((ct_data *)tree, max_code, s->bl_count);
+}
+
+/* ===========================================================================
+ * Scan a literal or distance tree to determine the frequencies of the codes
+ * in the bit length tree.
+ */
+local void scan_tree (s, tree, max_code)
+    deflate_state *s;
+    ct_data *tree;   /* the tree to be scanned */
+    int max_code;    /* and its largest code of non zero frequency */
+{
+    int n;                     /* iterates over all tree elements */
+    int prevlen = -1;          /* last emitted length */
+    int curlen;                /* length of current code */
+    int nextlen = tree[0].Len; /* length of next code */
+    int count = 0;             /* repeat count of the current code */
+    int max_count = 7;         /* max repeat count */
+    int min_count = 4;         /* min repeat count */
+
+    if (nextlen == 0) max_count = 138, min_count = 3;
+    tree[max_code+1].Len = (ush)0xffff; /* guard */
+
+    for (n = 0; n <= max_code; n++) {
+        curlen = nextlen; nextlen = tree[n+1].Len;
+        if (++count < max_count && curlen == nextlen) {
+            continue;
+        } else if (count < min_count) {
+            s->bl_tree[curlen].Freq += count;
+        } else if (curlen != 0) {
+            if (curlen != prevlen) s->bl_tree[curlen].Freq++;
+            s->bl_tree[REP_3_6].Freq++;
+        } else if (count <= 10) {
+            s->bl_tree[REPZ_3_10].Freq++;
+        } else {
+            s->bl_tree[REPZ_11_138].Freq++;
+        }
+        count = 0; prevlen = curlen;
+        if (nextlen == 0) {
+            max_count = 138, min_count = 3;
+        } else if (curlen == nextlen) {
+            max_count = 6, min_count = 3;
+        } else {
+            max_count = 7, min_count = 4;
+        }
+    }
+}
+
+/* ===========================================================================
+ * Send a literal or distance tree in compressed form, using the codes in
+ * bl_tree.
+ */
+local void send_tree (s, tree, max_code)
+    deflate_state *s;
+    ct_data *tree; /* the tree to be scanned */
+    int max_code;       /* and its largest code of non zero frequency */
+{
+    int n;                     /* iterates over all tree elements */
+    int prevlen = -1;          /* last emitted length */
+    int curlen;                /* length of current code */
+    int nextlen = tree[0].Len; /* length of next code */
+    int count = 0;             /* repeat count of the current code */
+    int max_count = 7;         /* max repeat count */
+    int min_count = 4;         /* min repeat count */
+
+    /* tree[max_code+1].Len = -1; */  /* guard already set */
+    if (nextlen == 0) max_count = 138, min_count = 3;
+
+    for (n = 0; n <= max_code; n++) {
+        curlen = nextlen; nextlen = tree[n+1].Len;
+        if (++count < max_count && curlen == nextlen) {
+            continue;
+        } else if (count < min_count) {
+            do { send_code(s, curlen, s->bl_tree); } while (--count != 0);
+
+        } else if (curlen != 0) {
+            if (curlen != prevlen) {
+                send_code(s, curlen, s->bl_tree); count--;
+            }
+            Assert(count >= 3 && count <= 6, " 3_6?");
+            send_code(s, REP_3_6, s->bl_tree); send_bits(s, count-3, 2);
+
+        } else if (count <= 10) {
+            send_code(s, REPZ_3_10, s->bl_tree); send_bits(s, count-3, 3);
+
+        } else {
+            send_code(s, REPZ_11_138, s->bl_tree); send_bits(s, count-11, 7);
+        }
+        count = 0; prevlen = curlen;
+        if (nextlen == 0) {
+            max_count = 138, min_count = 3;
+        } else if (curlen == nextlen) {
+            max_count = 6, min_count = 3;
+        } else {
+            max_count = 7, min_count = 4;
+        }
+    }
+}
+
+/* ===========================================================================
+ * Construct the Huffman tree for the bit lengths and return the index in
+ * bl_order of the last bit length code to send.
+ */
+local int build_bl_tree(s)
+    deflate_state *s;
+{
+    int max_blindex;  /* index of last bit length code of non zero freq */
+
+    /* Determine the bit length frequencies for literal and distance trees */
+    scan_tree(s, (ct_data *)s->dyn_ltree, s->l_desc.max_code);
+    scan_tree(s, (ct_data *)s->dyn_dtree, s->d_desc.max_code);
+
+    /* Build the bit length tree: */
+    build_tree(s, (tree_desc *)(&(s->bl_desc)));
+    /* opt_len now includes the length of the tree representations, except
+     * the lengths of the bit lengths codes and the 5+5+4 bits for the counts.
+     */
+
+    /* Determine the number of bit length codes to send. The pkzip format
+     * requires that at least 4 bit length codes be sent. (appnote.txt says
+     * 3 but the actual value used is 4.)
+     */
+    for (max_blindex = BL_CODES-1; max_blindex >= 3; max_blindex--) {
+        if (s->bl_tree[bl_order[max_blindex]].Len != 0) break;
+    }
+    /* Update opt_len to include the bit length tree and counts */
+    s->opt_len += 3*(max_blindex+1) + 5+5+4;
+    Tracev((stderr, "\ndyn trees: dyn %ld, stat %ld",
+            s->opt_len, s->static_len));
+
+    return max_blindex;
+}
+
+/* ===========================================================================
+ * Send the header for a block using dynamic Huffman trees: the counts, the
+ * lengths of the bit length codes, the literal tree and the distance tree.
+ * IN assertion: lcodes >= 257, dcodes >= 1, blcodes >= 4.
+ */
+local void send_all_trees(s, lcodes, dcodes, blcodes)
+    deflate_state *s;
+    int lcodes, dcodes, blcodes; /* number of codes for each tree */
+{
+    int rank;                    /* index in bl_order */
+
+    Assert (lcodes >= 257 && dcodes >= 1 && blcodes >= 4, "not enough codes");
+    Assert (lcodes <= L_CODES && dcodes <= D_CODES && blcodes <= BL_CODES,
+            "too many codes");
+    Tracev((stderr, "\nbl counts: "));
+    send_bits(s, lcodes-257, 5); /* not +255 as stated in appnote.txt */
+    send_bits(s, dcodes-1,   5);
+    send_bits(s, blcodes-4,  4); /* not -3 as stated in appnote.txt */
+    for (rank = 0; rank < blcodes; rank++) {
+        Tracev((stderr, "\nbl code %2d ", bl_order[rank]));
+        send_bits(s, s->bl_tree[bl_order[rank]].Len, 3);
+    }
+    Tracev((stderr, "\nbl tree: sent %ld", s->bits_sent));
+
+    send_tree(s, (ct_data *)s->dyn_ltree, lcodes-1); /* literal tree */
+    Tracev((stderr, "\nlit tree: sent %ld", s->bits_sent));
+
+    send_tree(s, (ct_data *)s->dyn_dtree, dcodes-1); /* distance tree */
+    Tracev((stderr, "\ndist tree: sent %ld", s->bits_sent));
+}
+
+/* ===========================================================================
+ * Send a stored block
+ */
+void _tr_stored_block(s, buf, stored_len, eof)
+    deflate_state *s;
+    charf *buf;       /* input block */
+    ulg stored_len;   /* length of input block */
+    int eof;          /* true if this is the last block for a file */
+{
+    send_bits(s, (STORED_BLOCK<<1)+eof, 3);  /* send block type */
+#ifdef DEBUG
+    s->compressed_len = (s->compressed_len + 3 + 7) & (ulg)~7L;
+    s->compressed_len += (stored_len + 4) << 3;
+#endif
+    copy_block(s, buf, (unsigned)stored_len, 1); /* with header */
+}
+
+/* ===========================================================================
+ * Send one empty static block to give enough lookahead for inflate.
+ * This takes 10 bits, of which 7 may remain in the bit buffer.
+ * The current inflate code requires 9 bits of lookahead. If the
+ * last two codes for the previous block (real code plus EOB) were coded
+ * on 5 bits or less, inflate may have only 5+3 bits of lookahead to decode
+ * the last real code. In this case we send two empty static blocks instead
+ * of one. (There are no problems if the previous block is stored or fixed.)
+ * To simplify the code, we assume the worst case of last real code encoded
+ * on one bit only.
+ */
+void _tr_align(s)
+    deflate_state *s;
+{
+    send_bits(s, STATIC_TREES<<1, 3);
+    send_code(s, END_BLOCK, static_ltree);
+#ifdef DEBUG
+    s->compressed_len += 10L; /* 3 for block type, 7 for EOB */
+#endif
+    bi_flush(s);
+    /* Of the 10 bits for the empty block, we have already sent
+     * (10 - bi_valid) bits. The lookahead for the last real code (before
+     * the EOB of the previous block) was thus at least one plus the length
+     * of the EOB plus what we have just sent of the empty static block.
+     */
+    if (1 + s->last_eob_len + 10 - s->bi_valid < 9) {
+        send_bits(s, STATIC_TREES<<1, 3);
+        send_code(s, END_BLOCK, static_ltree);
+#ifdef DEBUG
+        s->compressed_len += 10L;
+#endif
+        bi_flush(s);
+    }
+    s->last_eob_len = 7;
+}
+
+/* ===========================================================================
+ * Determine the best encoding for the current block: dynamic trees, static
+ * trees or store, and output the encoded block to the zip file.
+ */
+void _tr_flush_block(s, buf, stored_len, eof)
+    deflate_state *s;
+    charf *buf;       /* input block, or NULL if too old */
+    ulg stored_len;   /* length of input block */
+    int eof;          /* true if this is the last block for a file */
+{
+    ulg opt_lenb, static_lenb; /* opt_len and static_len in bytes */
+    int max_blindex = 0;  /* index of last bit length code of non zero freq */
+
+    /* Build the Huffman trees unless a stored block is forced */
+    if (s->level > 0) {
+
+        /* Check if the file is binary or text */
+        if (stored_len > 0 && s->strm->data_type == Z_UNKNOWN)
+            set_data_type(s);
+
+        /* Construct the literal and distance trees */
+        build_tree(s, (tree_desc *)(&(s->l_desc)));
+        Tracev((stderr, "\nlit data: dyn %ld, stat %ld", s->opt_len,
+                s->static_len));
+
+        build_tree(s, (tree_desc *)(&(s->d_desc)));
+        Tracev((stderr, "\ndist data: dyn %ld, stat %ld", s->opt_len,
+                s->static_len));
+        /* At this point, opt_len and static_len are the total bit lengths of
+         * the compressed block data, excluding the tree representations.
+         */
+
+        /* Build the bit length tree for the above two trees, and get the index
+         * in bl_order of the last bit length code to send.
+         */
+        max_blindex = build_bl_tree(s);
+
+        /* Determine the best encoding. Compute the block lengths in bytes. */
+        opt_lenb = (s->opt_len+3+7)>>3;
+        static_lenb = (s->static_len+3+7)>>3;
+
+        Tracev((stderr, "\nopt %lu(%lu) stat %lu(%lu) stored %lu lit %u ",
+                opt_lenb, s->opt_len, static_lenb, s->static_len, stored_len,
+                s->last_lit));
+
+        if (static_lenb <= opt_lenb) opt_lenb = static_lenb;
+
+    } else {
+        Assert(buf != (char*)0, "lost buf");
+        opt_lenb = static_lenb = stored_len + 5; /* force a stored block */
+    }
+
+#ifdef FORCE_STORED
+    if (buf != (char*)0) { /* force stored block */
+#else
+    if (stored_len+4 <= opt_lenb && buf != (char*)0) {
+                       /* 4: two words for the lengths */
+#endif
+        /* The test buf != NULL is only necessary if LIT_BUFSIZE > WSIZE.
+         * Otherwise we can't have processed more than WSIZE input bytes since
+         * the last block flush, because compression would have been
+         * successful. If LIT_BUFSIZE <= WSIZE, it is never too late to
+         * transform a block into a stored block.
+         */
+        _tr_stored_block(s, buf, stored_len, eof);
+
+#ifdef FORCE_STATIC
+    } else if (static_lenb >= 0) { /* force static trees */
+#else
+    } else if (s->strategy == Z_FIXED || static_lenb == opt_lenb) {
+#endif
+        send_bits(s, (STATIC_TREES<<1)+eof, 3);
+        compress_block(s, (ct_data *)static_ltree, (ct_data *)static_dtree);
+#ifdef DEBUG
+        s->compressed_len += 3 + s->static_len;
+#endif
+    } else {
+        send_bits(s, (DYN_TREES<<1)+eof, 3);
+        send_all_trees(s, s->l_desc.max_code+1, s->d_desc.max_code+1,
+                       max_blindex+1);
+        compress_block(s, (ct_data *)s->dyn_ltree, (ct_data *)s->dyn_dtree);
+#ifdef DEBUG
+        s->compressed_len += 3 + s->opt_len;
+#endif
+    }
+    Assert (s->compressed_len == s->bits_sent, "bad compressed size");
+    /* The above check is made mod 2^32, for files larger than 512 MB
+     * and uLong implemented on 32 bits.
+     */
+    init_block(s);
+
+    if (eof) {
+        bi_windup(s);
+#ifdef DEBUG
+        s->compressed_len += 7;  /* align on byte boundary */
+#endif
+    }
+    Tracev((stderr,"\ncomprlen %lu(%lu) ", s->compressed_len>>3,
+           s->compressed_len-7*eof));
+}
+
+/* ===========================================================================
+ * Save the match info and tally the frequency counts. Return true if
+ * the current block must be flushed.
+ */
+int _tr_tally (s, dist, lc)
+    deflate_state *s;
+    unsigned dist;  /* distance of matched string */
+    unsigned lc;    /* match length-MIN_MATCH or unmatched char (if dist==0) */
+{
+    s->d_buf[s->last_lit] = (ush)dist;
+    s->l_buf[s->last_lit++] = (uch)lc;
+    if (dist == 0) {
+        /* lc is the unmatched char */
+        s->dyn_ltree[lc].Freq++;
+    } else {
+        s->matches++;
+        /* Here, lc is the match length - MIN_MATCH */
+        dist--;             /* dist = match distance - 1 */
+        Assert((ush)dist < (ush)MAX_DIST(s) &&
+               (ush)lc <= (ush)(MAX_MATCH-MIN_MATCH) &&
+               (ush)d_code(dist) < (ush)D_CODES,  "_tr_tally: bad match");
+
+        s->dyn_ltree[_length_code[lc]+LITERALS+1].Freq++;
+        s->dyn_dtree[d_code(dist)].Freq++;
+    }
+
+#ifdef TRUNCATE_BLOCK
+    /* Try to guess if it is profitable to stop the current block here */
+    if ((s->last_lit & 0x1fff) == 0 && s->level > 2) {
+        /* Compute an upper bound for the compressed length */
+        ulg out_length = (ulg)s->last_lit*8L;
+        ulg in_length = (ulg)((long)s->strstart - s->block_start);
+        int dcode;
+        for (dcode = 0; dcode < D_CODES; dcode++) {
+            out_length += (ulg)s->dyn_dtree[dcode].Freq *
+                (5L+extra_dbits[dcode]);
+        }
+        out_length >>= 3;
+        Tracev((stderr,"\nlast_lit %u, in %ld, out ~%ld(%ld%%) ",
+               s->last_lit, in_length, out_length,
+               100L - out_length*100L/in_length));
+        if (s->matches < s->last_lit/2 && out_length < in_length/2) return 1;
+    }
+#endif
+    return (s->last_lit == s->lit_bufsize-1);
+    /* We avoid equality with lit_bufsize because of wraparound at 64K
+     * on 16 bit machines and because stored blocks are restricted to
+     * 64K-1 bytes.
+     */
+}
+
+/* ===========================================================================
+ * Send the block data compressed using the given Huffman trees
+ */
+local void compress_block(s, ltree, dtree)
+    deflate_state *s;
+    ct_data *ltree; /* literal tree */
+    ct_data *dtree; /* distance tree */
+{
+    unsigned dist;      /* distance of matched string */
+    int lc;             /* match length or unmatched char (if dist == 0) */
+    unsigned lx = 0;    /* running index in l_buf */
+    unsigned code;      /* the code to send */
+    int extra;          /* number of extra bits to send */
+
+    if (s->last_lit != 0) do {
+        dist = s->d_buf[lx];
+        lc = s->l_buf[lx++];
+        if (dist == 0) {
+            send_code(s, lc, ltree); /* send a literal byte */
+            Tracecv(isgraph(lc), (stderr," '%c' ", lc));
+        } else {
+            /* Here, lc is the match length - MIN_MATCH */
+            code = _length_code[lc];
+            send_code(s, code+LITERALS+1, ltree); /* send the length code */
+            extra = extra_lbits[code];
+            if (extra != 0) {
+                lc -= base_length[code];
+                send_bits(s, lc, extra);       /* send the extra length bits */
+            }
+            dist--; /* dist is now the match distance - 1 */
+            code = d_code(dist);
+            Assert (code < D_CODES, "bad d_code");
+
+            send_code(s, code, dtree);       /* send the distance code */
+            extra = extra_dbits[code];
+            if (extra != 0) {
+                dist -= base_dist[code];
+                send_bits(s, dist, extra);   /* send the extra distance bits */
+            }
+        } /* literal or match pair ? */
+
+        /* Check that the overlay between pending_buf and d_buf+l_buf is ok: */
+        Assert((uInt)(s->pending) < s->lit_bufsize + 2*lx,
+               "pendingBuf overflow");
+
+    } while (lx < s->last_lit);
+
+    send_code(s, END_BLOCK, ltree);
+    s->last_eob_len = ltree[END_BLOCK].Len;
+}
+
+/* ===========================================================================
+ * Set the data type to BINARY or TEXT, using a crude approximation:
+ * set it to Z_TEXT if all symbols are either printable characters (33 to 255)
+ * or white spaces (9 to 13, or 32); or set it to Z_BINARY otherwise.
+ * IN assertion: the fields Freq of dyn_ltree are set.
+ */
+local void set_data_type(s)
+    deflate_state *s;
+{
+    int n;
+
+    for (n = 0; n < 9; n++)
+        if (s->dyn_ltree[n].Freq != 0)
+            break;
+    if (n == 9)
+        for (n = 14; n < 32; n++)
+            if (s->dyn_ltree[n].Freq != 0)
+                break;
+    s->strm->data_type = (n == 32) ? Z_TEXT : Z_BINARY;
+}
+
+/* ===========================================================================
+ * Reverse the first len bits of a code, using straightforward code (a faster
+ * method would use a table)
+ * IN assertion: 1 <= len <= 15
+ */
+local unsigned bi_reverse(code, len)
+    unsigned code; /* the value to invert */
+    int len;       /* its bit length */
+{
+    register unsigned res = 0;
+    do {
+        res |= code & 1;
+        code >>= 1, res <<= 1;
+    } while (--len > 0);
+    return res >> 1;
+}
+
+/* ===========================================================================
+ * Flush the bit buffer, keeping at most 7 bits in it.
+ */
+local void bi_flush(s)
+    deflate_state *s;
+{
+    if (s->bi_valid == 16) {
+        put_short(s, s->bi_buf);
+        s->bi_buf = 0;
+        s->bi_valid = 0;
+    } else if (s->bi_valid >= 8) {
+        put_byte(s, (Byte)s->bi_buf);
+        s->bi_buf >>= 8;
+        s->bi_valid -= 8;
+    }
+}
+
+/* ===========================================================================
+ * Flush the bit buffer and align the output on a byte boundary
+ */
+local void bi_windup(s)
+    deflate_state *s;
+{
+    if (s->bi_valid > 8) {
+        put_short(s, s->bi_buf);
+    } else if (s->bi_valid > 0) {
+        put_byte(s, (Byte)s->bi_buf);
+    }
+    s->bi_buf = 0;
+    s->bi_valid = 0;
+#ifdef DEBUG
+    s->bits_sent = (s->bits_sent+7) & ~7;
+#endif
+}
+
+/* ===========================================================================
+ * Copy a stored block, storing first the length and its
+ * one's complement if requested.
+ */
+local void copy_block(s, buf, len, header)
+    deflate_state *s;
+    charf    *buf;    /* the input data */
+    unsigned len;     /* its length */
+    int      header;  /* true if block header must be written */
+{
+    bi_windup(s);        /* align on byte boundary */
+    s->last_eob_len = 8; /* enough lookahead for inflate */
+
+    if (header) {
+        put_short(s, (ush)len);
+        put_short(s, (ush)~len);
+#ifdef DEBUG
+        s->bits_sent += 2*16;
+#endif
+    }
+#ifdef DEBUG
+    s->bits_sent += (ulg)len<<3;
+#endif
+    while (len--) {
+        put_byte(s, *buf++);
+    }
+}
diff --git a/distrib/zlib-1.2.3/trees.h b/distrib/zlib-1.2.3/trees.h
new file mode 100644
index 0000000..72facf9
--- /dev/null
+++ b/distrib/zlib-1.2.3/trees.h
@@ -0,0 +1,128 @@
+/* header created automatically with -DGEN_TREES_H */
+
+local const ct_data static_ltree[L_CODES+2] = {
+{{ 12},{  8}}, {{140},{  8}}, {{ 76},{  8}}, {{204},{  8}}, {{ 44},{  8}},
+{{172},{  8}}, {{108},{  8}}, {{236},{  8}}, {{ 28},{  8}}, {{156},{  8}},
+{{ 92},{  8}}, {{220},{  8}}, {{ 60},{  8}}, {{188},{  8}}, {{124},{  8}},
+{{252},{  8}}, {{  2},{  8}}, {{130},{  8}}, {{ 66},{  8}}, {{194},{  8}},
+{{ 34},{  8}}, {{162},{  8}}, {{ 98},{  8}}, {{226},{  8}}, {{ 18},{  8}},
+{{146},{  8}}, {{ 82},{  8}}, {{210},{  8}}, {{ 50},{  8}}, {{178},{  8}},
+{{114},{  8}}, {{242},{  8}}, {{ 10},{  8}}, {{138},{  8}}, {{ 74},{  8}},
+{{202},{  8}}, {{ 42},{  8}}, {{170},{  8}}, {{106},{  8}}, {{234},{  8}},
+{{ 26},{  8}}, {{154},{  8}}, {{ 90},{  8}}, {{218},{  8}}, {{ 58},{  8}},
+{{186},{  8}}, {{122},{  8}}, {{250},{  8}}, {{  6},{  8}}, {{134},{  8}},
+{{ 70},{  8}}, {{198},{  8}}, {{ 38},{  8}}, {{166},{  8}}, {{102},{  8}},
+{{230},{  8}}, {{ 22},{  8}}, {{150},{  8}}, {{ 86},{  8}}, {{214},{  8}},
+{{ 54},{  8}}, {{182},{  8}}, {{118},{  8}}, {{246},{  8}}, {{ 14},{  8}},
+{{142},{  8}}, {{ 78},{  8}}, {{206},{  8}}, {{ 46},{  8}}, {{174},{  8}},
+{{110},{  8}}, {{238},{  8}}, {{ 30},{  8}}, {{158},{  8}}, {{ 94},{  8}},
+{{222},{  8}}, {{ 62},{  8}}, {{190},{  8}}, {{126},{  8}}, {{254},{  8}},
+{{  1},{  8}}, {{129},{  8}}, {{ 65},{  8}}, {{193},{  8}}, {{ 33},{  8}},
+{{161},{  8}}, {{ 97},{  8}}, {{225},{  8}}, {{ 17},{  8}}, {{145},{  8}},
+{{ 81},{  8}}, {{209},{  8}}, {{ 49},{  8}}, {{177},{  8}}, {{113},{  8}},
+{{241},{  8}}, {{  9},{  8}}, {{137},{  8}}, {{ 73},{  8}}, {{201},{  8}},
+{{ 41},{  8}}, {{169},{  8}}, {{105},{  8}}, {{233},{  8}}, {{ 25},{  8}},
+{{153},{  8}}, {{ 89},{  8}}, {{217},{  8}}, {{ 57},{  8}}, {{185},{  8}},
+{{121},{  8}}, {{249},{  8}}, {{  5},{  8}}, {{133},{  8}}, {{ 69},{  8}},
+{{197},{  8}}, {{ 37},{  8}}, {{165},{  8}}, {{101},{  8}}, {{229},{  8}},
+{{ 21},{  8}}, {{149},{  8}}, {{ 85},{  8}}, {{213},{  8}}, {{ 53},{  8}},
+{{181},{  8}}, {{117},{  8}}, {{245},{  8}}, {{ 13},{  8}}, {{141},{  8}},
+{{ 77},{  8}}, {{205},{  8}}, {{ 45},{  8}}, {{173},{  8}}, {{109},{  8}},
+{{237},{  8}}, {{ 29},{  8}}, {{157},{  8}}, {{ 93},{  8}}, {{221},{  8}},
+{{ 61},{  8}}, {{189},{  8}}, {{125},{  8}}, {{253},{  8}}, {{ 19},{  9}},
+{{275},{  9}}, {{147},{  9}}, {{403},{  9}}, {{ 83},{  9}}, {{339},{  9}},
+{{211},{  9}}, {{467},{  9}}, {{ 51},{  9}}, {{307},{  9}}, {{179},{  9}},
+{{435},{  9}}, {{115},{  9}}, {{371},{  9}}, {{243},{  9}}, {{499},{  9}},
+{{ 11},{  9}}, {{267},{  9}}, {{139},{  9}}, {{395},{  9}}, {{ 75},{  9}},
+{{331},{  9}}, {{203},{  9}}, {{459},{  9}}, {{ 43},{  9}}, {{299},{  9}},
+{{171},{  9}}, {{427},{  9}}, {{107},{  9}}, {{363},{  9}}, {{235},{  9}},
+{{491},{  9}}, {{ 27},{  9}}, {{283},{  9}}, {{155},{  9}}, {{411},{  9}},
+{{ 91},{  9}}, {{347},{  9}}, {{219},{  9}}, {{475},{  9}}, {{ 59},{  9}},
+{{315},{  9}}, {{187},{  9}}, {{443},{  9}}, {{123},{  9}}, {{379},{  9}},
+{{251},{  9}}, {{507},{  9}}, {{  7},{  9}}, {{263},{  9}}, {{135},{  9}},
+{{391},{  9}}, {{ 71},{  9}}, {{327},{  9}}, {{199},{  9}}, {{455},{  9}},
+{{ 39},{  9}}, {{295},{  9}}, {{167},{  9}}, {{423},{  9}}, {{103},{  9}},
+{{359},{  9}}, {{231},{  9}}, {{487},{  9}}, {{ 23},{  9}}, {{279},{  9}},
+{{151},{  9}}, {{407},{  9}}, {{ 87},{  9}}, {{343},{  9}}, {{215},{  9}},
+{{471},{  9}}, {{ 55},{  9}}, {{311},{  9}}, {{183},{  9}}, {{439},{  9}},
+{{119},{  9}}, {{375},{  9}}, {{247},{  9}}, {{503},{  9}}, {{ 15},{  9}},
+{{271},{  9}}, {{143},{  9}}, {{399},{  9}}, {{ 79},{  9}}, {{335},{  9}},
+{{207},{  9}}, {{463},{  9}}, {{ 47},{  9}}, {{303},{  9}}, {{175},{  9}},
+{{431},{  9}}, {{111},{  9}}, {{367},{  9}}, {{239},{  9}}, {{495},{  9}},
+{{ 31},{  9}}, {{287},{  9}}, {{159},{  9}}, {{415},{  9}}, {{ 95},{  9}},
+{{351},{  9}}, {{223},{  9}}, {{479},{  9}}, {{ 63},{  9}}, {{319},{  9}},
+{{191},{  9}}, {{447},{  9}}, {{127},{  9}}, {{383},{  9}}, {{255},{  9}},
+{{511},{  9}}, {{  0},{  7}}, {{ 64},{  7}}, {{ 32},{  7}}, {{ 96},{  7}},
+{{ 16},{  7}}, {{ 80},{  7}}, {{ 48},{  7}}, {{112},{  7}}, {{  8},{  7}},
+{{ 72},{  7}}, {{ 40},{  7}}, {{104},{  7}}, {{ 24},{  7}}, {{ 88},{  7}},
+{{ 56},{  7}}, {{120},{  7}}, {{  4},{  7}}, {{ 68},{  7}}, {{ 36},{  7}},
+{{100},{  7}}, {{ 20},{  7}}, {{ 84},{  7}}, {{ 52},{  7}}, {{116},{  7}},
+{{  3},{  8}}, {{131},{  8}}, {{ 67},{  8}}, {{195},{  8}}, {{ 35},{  8}},
+{{163},{  8}}, {{ 99},{  8}}, {{227},{  8}}
+};
+
+local const ct_data static_dtree[D_CODES] = {
+{{ 0},{ 5}}, {{16},{ 5}}, {{ 8},{ 5}}, {{24},{ 5}}, {{ 4},{ 5}},
+{{20},{ 5}}, {{12},{ 5}}, {{28},{ 5}}, {{ 2},{ 5}}, {{18},{ 5}},
+{{10},{ 5}}, {{26},{ 5}}, {{ 6},{ 5}}, {{22},{ 5}}, {{14},{ 5}},
+{{30},{ 5}}, {{ 1},{ 5}}, {{17},{ 5}}, {{ 9},{ 5}}, {{25},{ 5}},
+{{ 5},{ 5}}, {{21},{ 5}}, {{13},{ 5}}, {{29},{ 5}}, {{ 3},{ 5}},
+{{19},{ 5}}, {{11},{ 5}}, {{27},{ 5}}, {{ 7},{ 5}}, {{23},{ 5}}
+};
+
+const uch _dist_code[DIST_CODE_LEN] = {
+ 0,  1,  2,  3,  4,  4,  5,  5,  6,  6,  6,  6,  7,  7,  7,  7,  8,  8,  8,  8,
+ 8,  8,  8,  8,  9,  9,  9,  9,  9,  9,  9,  9, 10, 10, 10, 10, 10, 10, 10, 10,
+10, 10, 10, 10, 10, 10, 10, 10, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11,
+11, 11, 11, 11, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12,
+12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12, 13, 13, 13, 13,
+13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13, 13,
+13, 13, 13, 13, 13, 13, 13, 13, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14,
+14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14,
+14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14,
+14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 14, 15, 15, 15, 15, 15, 15, 15, 15,
+15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15,
+15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15,
+15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15, 15,  0,  0, 16, 17,
+18, 18, 19, 19, 20, 20, 20, 20, 21, 21, 21, 21, 22, 22, 22, 22, 22, 22, 22, 22,
+23, 23, 23, 23, 23, 23, 23, 23, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24,
+24, 24, 24, 24, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25,
+26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26,
+26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 27, 27, 27, 27, 27, 27, 27, 27,
+27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27,
+27, 27, 27, 27, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28,
+28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28,
+28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28,
+28, 28, 28, 28, 28, 28, 28, 28, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29,
+29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29,
+29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29,
+29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29
+};
+
+const uch _length_code[MAX_MATCH-MIN_MATCH+1]= {
+ 0,  1,  2,  3,  4,  5,  6,  7,  8,  8,  9,  9, 10, 10, 11, 11, 12, 12, 12, 12,
+13, 13, 13, 13, 14, 14, 14, 14, 15, 15, 15, 15, 16, 16, 16, 16, 16, 16, 16, 16,
+17, 17, 17, 17, 17, 17, 17, 17, 18, 18, 18, 18, 18, 18, 18, 18, 19, 19, 19, 19,
+19, 19, 19, 19, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20,
+21, 21, 21, 21, 21, 21, 21, 21, 21, 21, 21, 21, 21, 21, 21, 21, 22, 22, 22, 22,
+22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 22, 23, 23, 23, 23, 23, 23, 23, 23,
+23, 23, 23, 23, 23, 23, 23, 23, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24,
+24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 24,
+25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25,
+25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 25, 26, 26, 26, 26, 26, 26, 26, 26,
+26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26, 26,
+26, 26, 26, 26, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27,
+27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 27, 28
+};
+
+local const int base_length[LENGTH_CODES] = {
+0, 1, 2, 3, 4, 5, 6, 7, 8, 10, 12, 14, 16, 20, 24, 28, 32, 40, 48, 56,
+64, 80, 96, 112, 128, 160, 192, 224, 0
+};
+
+local const int base_dist[D_CODES] = {
+    0,     1,     2,     3,     4,     6,     8,    12,    16,    24,
+   32,    48,    64,    96,   128,   192,   256,   384,   512,   768,
+ 1024,  1536,  2048,  3072,  4096,  6144,  8192, 12288, 16384, 24576
+};
+
diff --git a/distrib/zlib-1.2.3/uncompr.c b/distrib/zlib-1.2.3/uncompr.c
new file mode 100644
index 0000000..b59e3d0
--- /dev/null
+++ b/distrib/zlib-1.2.3/uncompr.c
@@ -0,0 +1,61 @@
+/* uncompr.c -- decompress a memory buffer
+ * Copyright (C) 1995-2003 Jean-loup Gailly.
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* @(#) $Id$ */
+
+#define ZLIB_INTERNAL
+#include "zlib.h"
+
+/* ===========================================================================
+     Decompresses the source buffer into the destination buffer.  sourceLen is
+   the byte length of the source buffer. Upon entry, destLen is the total
+   size of the destination buffer, which must be large enough to hold the
+   entire uncompressed data. (The size of the uncompressed data must have
+   been saved previously by the compressor and transmitted to the decompressor
+   by some mechanism outside the scope of this compression library.)
+   Upon exit, destLen is the actual size of the compressed buffer.
+     This function can be used to decompress a whole file at once if the
+   input file is mmap'ed.
+
+     uncompress returns Z_OK if success, Z_MEM_ERROR if there was not
+   enough memory, Z_BUF_ERROR if there was not enough room in the output
+   buffer, or Z_DATA_ERROR if the input data was corrupted.
+*/
+int ZEXPORT uncompress (dest, destLen, source, sourceLen)
+    Bytef *dest;
+    uLongf *destLen;
+    const Bytef *source;
+    uLong sourceLen;
+{
+    z_stream stream;
+    int err;
+
+    stream.next_in = (Bytef*)source;
+    stream.avail_in = (uInt)sourceLen;
+    /* Check for source > 64K on 16-bit machine: */
+    if ((uLong)stream.avail_in != sourceLen) return Z_BUF_ERROR;
+
+    stream.next_out = dest;
+    stream.avail_out = (uInt)*destLen;
+    if ((uLong)stream.avail_out != *destLen) return Z_BUF_ERROR;
+
+    stream.zalloc = (alloc_func)0;
+    stream.zfree = (free_func)0;
+
+    err = inflateInit(&stream);
+    if (err != Z_OK) return err;
+
+    err = inflate(&stream, Z_FINISH);
+    if (err != Z_STREAM_END) {
+        inflateEnd(&stream);
+        if (err == Z_NEED_DICT || (err == Z_BUF_ERROR && stream.avail_in == 0))
+            return Z_DATA_ERROR;
+        return err;
+    }
+    *destLen = stream.total_out;
+
+    err = inflateEnd(&stream);
+    return err;
+}
diff --git a/distrib/zlib-1.2.3/zconf.h b/distrib/zlib-1.2.3/zconf.h
new file mode 100644
index 0000000..03a9431
--- /dev/null
+++ b/distrib/zlib-1.2.3/zconf.h
@@ -0,0 +1,332 @@
+/* zconf.h -- configuration of the zlib compression library
+ * Copyright (C) 1995-2005 Jean-loup Gailly.
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* @(#) $Id$ */
+
+#ifndef ZCONF_H
+#define ZCONF_H
+
+/*
+ * If you *really* need a unique prefix for all types and library functions,
+ * compile with -DZ_PREFIX. The "standard" zlib should be compiled without it.
+ */
+#ifdef Z_PREFIX
+#  define deflateInit_          z_deflateInit_
+#  define deflate               z_deflate
+#  define deflateEnd            z_deflateEnd
+#  define inflateInit_          z_inflateInit_
+#  define inflate               z_inflate
+#  define inflateEnd            z_inflateEnd
+#  define deflateInit2_         z_deflateInit2_
+#  define deflateSetDictionary  z_deflateSetDictionary
+#  define deflateCopy           z_deflateCopy
+#  define deflateReset          z_deflateReset
+#  define deflateParams         z_deflateParams
+#  define deflateBound          z_deflateBound
+#  define deflatePrime          z_deflatePrime
+#  define inflateInit2_         z_inflateInit2_
+#  define inflateSetDictionary  z_inflateSetDictionary
+#  define inflateSync           z_inflateSync
+#  define inflateSyncPoint      z_inflateSyncPoint
+#  define inflateCopy           z_inflateCopy
+#  define inflateReset          z_inflateReset
+#  define inflateBack           z_inflateBack
+#  define inflateBackEnd        z_inflateBackEnd
+#  define compress              z_compress
+#  define compress2             z_compress2
+#  define compressBound         z_compressBound
+#  define uncompress            z_uncompress
+#  define adler32               z_adler32
+#  define crc32                 z_crc32
+#  define get_crc_table         z_get_crc_table
+#  define zError                z_zError
+
+#  define alloc_func            z_alloc_func
+#  define free_func             z_free_func
+#  define in_func               z_in_func
+#  define out_func              z_out_func
+#  define Byte                  z_Byte
+#  define uInt                  z_uInt
+#  define uLong                 z_uLong
+#  define Bytef                 z_Bytef
+#  define charf                 z_charf
+#  define intf                  z_intf
+#  define uIntf                 z_uIntf
+#  define uLongf                z_uLongf
+#  define voidpf                z_voidpf
+#  define voidp                 z_voidp
+#endif
+
+#if defined(__MSDOS__) && !defined(MSDOS)
+#  define MSDOS
+#endif
+#if (defined(OS_2) || defined(__OS2__)) && !defined(OS2)
+#  define OS2
+#endif
+#if defined(_WINDOWS) && !defined(WINDOWS)
+#  define WINDOWS
+#endif
+#if defined(_WIN32) || defined(_WIN32_WCE) || defined(__WIN32__)
+#  ifndef WIN32
+#    define WIN32
+#  endif
+#endif
+#if (defined(MSDOS) || defined(OS2) || defined(WINDOWS)) && !defined(WIN32)
+#  if !defined(__GNUC__) && !defined(__FLAT__) && !defined(__386__)
+#    ifndef SYS16BIT
+#      define SYS16BIT
+#    endif
+#  endif
+#endif
+
+/*
+ * Compile with -DMAXSEG_64K if the alloc function cannot allocate more
+ * than 64k bytes at a time (needed on systems with 16-bit int).
+ */
+#ifdef SYS16BIT
+#  define MAXSEG_64K
+#endif
+#ifdef MSDOS
+#  define UNALIGNED_OK
+#endif
+
+#ifdef __STDC_VERSION__
+#  ifndef STDC
+#    define STDC
+#  endif
+#  if __STDC_VERSION__ >= 199901L
+#    ifndef STDC99
+#      define STDC99
+#    endif
+#  endif
+#endif
+#if !defined(STDC) && (defined(__STDC__) || defined(__cplusplus))
+#  define STDC
+#endif
+#if !defined(STDC) && (defined(__GNUC__) || defined(__BORLANDC__))
+#  define STDC
+#endif
+#if !defined(STDC) && (defined(MSDOS) || defined(WINDOWS) || defined(WIN32))
+#  define STDC
+#endif
+#if !defined(STDC) && (defined(OS2) || defined(__HOS_AIX__))
+#  define STDC
+#endif
+
+#if defined(__OS400__) && !defined(STDC)    /* iSeries (formerly AS/400). */
+#  define STDC
+#endif
+
+#ifndef STDC
+#  ifndef const /* cannot use !defined(STDC) && !defined(const) on Mac */
+#    define const       /* note: need a more gentle solution here */
+#  endif
+#endif
+
+/* Some Mac compilers merge all .h files incorrectly: */
+#if defined(__MWERKS__)||defined(applec)||defined(THINK_C)||defined(__SC__)
+#  define NO_DUMMY_DECL
+#endif
+
+/* Maximum value for memLevel in deflateInit2 */
+#ifndef MAX_MEM_LEVEL
+#  ifdef MAXSEG_64K
+#    define MAX_MEM_LEVEL 8
+#  else
+#    define MAX_MEM_LEVEL 9
+#  endif
+#endif
+
+/* Maximum value for windowBits in deflateInit2 and inflateInit2.
+ * WARNING: reducing MAX_WBITS makes minigzip unable to extract .gz files
+ * created by gzip. (Files created by minigzip can still be extracted by
+ * gzip.)
+ */
+#ifndef MAX_WBITS
+#  define MAX_WBITS   15 /* 32K LZ77 window */
+#endif
+
+/* The memory requirements for deflate are (in bytes):
+            (1 << (windowBits+2)) +  (1 << (memLevel+9))
+ that is: 128K for windowBits=15  +  128K for memLevel = 8  (default values)
+ plus a few kilobytes for small objects. For example, if you want to reduce
+ the default memory requirements from 256K to 128K, compile with
+     make CFLAGS="-O -DMAX_WBITS=14 -DMAX_MEM_LEVEL=7"
+ Of course this will generally degrade compression (there's no free lunch).
+
+   The memory requirements for inflate are (in bytes) 1 << windowBits
+ that is, 32K for windowBits=15 (default value) plus a few kilobytes
+ for small objects.
+*/
+
+                        /* Type declarations */
+
+#ifndef OF /* function prototypes */
+#  ifdef STDC
+#    define OF(args)  args
+#  else
+#    define OF(args)  ()
+#  endif
+#endif
+
+/* The following definitions for FAR are needed only for MSDOS mixed
+ * model programming (small or medium model with some far allocations).
+ * This was tested only with MSC; for other MSDOS compilers you may have
+ * to define NO_MEMCPY in zutil.h.  If you don't need the mixed model,
+ * just define FAR to be empty.
+ */
+#ifdef SYS16BIT
+#  if defined(M_I86SM) || defined(M_I86MM)
+     /* MSC small or medium model */
+#    define SMALL_MEDIUM
+#    ifdef _MSC_VER
+#      define FAR _far
+#    else
+#      define FAR far
+#    endif
+#  endif
+#  if (defined(__SMALL__) || defined(__MEDIUM__))
+     /* Turbo C small or medium model */
+#    define SMALL_MEDIUM
+#    ifdef __BORLANDC__
+#      define FAR _far
+#    else
+#      define FAR far
+#    endif
+#  endif
+#endif
+
+#if defined(WINDOWS) || defined(WIN32)
+   /* If building or using zlib as a DLL, define ZLIB_DLL.
+    * This is not mandatory, but it offers a little performance increase.
+    */
+#  ifdef ZLIB_DLL
+#    if defined(WIN32) && (!defined(__BORLANDC__) || (__BORLANDC__ >= 0x500))
+#      ifdef ZLIB_INTERNAL
+#        define ZEXTERN extern __declspec(dllexport)
+#      else
+#        define ZEXTERN extern __declspec(dllimport)
+#      endif
+#    endif
+#  endif  /* ZLIB_DLL */
+   /* If building or using zlib with the WINAPI/WINAPIV calling convention,
+    * define ZLIB_WINAPI.
+    * Caution: the standard ZLIB1.DLL is NOT compiled using ZLIB_WINAPI.
+    */
+#  ifdef ZLIB_WINAPI
+#    ifdef FAR
+#      undef FAR
+#    endif
+#    include <windows.h>
+     /* No need for _export, use ZLIB.DEF instead. */
+     /* For complete Windows compatibility, use WINAPI, not __stdcall. */
+#    define ZEXPORT WINAPI
+#    ifdef WIN32
+#      define ZEXPORTVA WINAPIV
+#    else
+#      define ZEXPORTVA FAR CDECL
+#    endif
+#  endif
+#endif
+
+#if defined (__BEOS__)
+#  ifdef ZLIB_DLL
+#    ifdef ZLIB_INTERNAL
+#      define ZEXPORT   __declspec(dllexport)
+#      define ZEXPORTVA __declspec(dllexport)
+#    else
+#      define ZEXPORT   __declspec(dllimport)
+#      define ZEXPORTVA __declspec(dllimport)
+#    endif
+#  endif
+#endif
+
+#ifndef ZEXTERN
+#  define ZEXTERN extern
+#endif
+#ifndef ZEXPORT
+#  define ZEXPORT
+#endif
+#ifndef ZEXPORTVA
+#  define ZEXPORTVA
+#endif
+
+#ifndef FAR
+#  define FAR
+#endif
+
+#if !defined(__MACTYPES__)
+typedef unsigned char  Byte;  /* 8 bits */
+#endif
+typedef unsigned int   uInt;  /* 16 bits or more */
+typedef unsigned long  uLong; /* 32 bits or more */
+
+#ifdef SMALL_MEDIUM
+   /* Borland C/C++ and some old MSC versions ignore FAR inside typedef */
+#  define Bytef Byte FAR
+#else
+   typedef Byte  FAR Bytef;
+#endif
+typedef char  FAR charf;
+typedef int   FAR intf;
+typedef uInt  FAR uIntf;
+typedef uLong FAR uLongf;
+
+#ifdef STDC
+   typedef void const *voidpc;
+   typedef void FAR   *voidpf;
+   typedef void       *voidp;
+#else
+   typedef Byte const *voidpc;
+   typedef Byte FAR   *voidpf;
+   typedef Byte       *voidp;
+#endif
+
+#if 0           /* HAVE_UNISTD_H -- this line is updated by ./configure */
+#  include <sys/types.h> /* for off_t */
+#  include <unistd.h>    /* for SEEK_* and off_t */
+#  ifdef VMS
+#    include <unixio.h>   /* for off_t */
+#  endif
+#  define z_off_t off_t
+#endif
+#ifndef SEEK_SET
+#  define SEEK_SET        0       /* Seek from beginning of file.  */
+#  define SEEK_CUR        1       /* Seek from current position.  */
+#  define SEEK_END        2       /* Set file pointer to EOF plus "offset" */
+#endif
+#ifndef z_off_t
+#  define z_off_t long
+#endif
+
+#if defined(__OS400__)
+#  define NO_vsnprintf
+#endif
+
+#if defined(__MVS__)
+#  define NO_vsnprintf
+#  ifdef FAR
+#    undef FAR
+#  endif
+#endif
+
+/* MVS linker does not support external names larger than 8 bytes */
+#if defined(__MVS__)
+#   pragma map(deflateInit_,"DEIN")
+#   pragma map(deflateInit2_,"DEIN2")
+#   pragma map(deflateEnd,"DEEND")
+#   pragma map(deflateBound,"DEBND")
+#   pragma map(inflateInit_,"ININ")
+#   pragma map(inflateInit2_,"ININ2")
+#   pragma map(inflateEnd,"INEND")
+#   pragma map(inflateSync,"INSY")
+#   pragma map(inflateSetDictionary,"INSEDI")
+#   pragma map(compressBound,"CMBND")
+#   pragma map(inflate_table,"INTABL")
+#   pragma map(inflate_fast,"INFA")
+#   pragma map(inflate_copyright,"INCOPY")
+#endif
+
+#endif /* ZCONF_H */
diff --git a/distrib/zlib-1.2.3/zlib.h b/distrib/zlib-1.2.3/zlib.h
new file mode 100644
index 0000000..0228179
--- /dev/null
+++ b/distrib/zlib-1.2.3/zlib.h
@@ -0,0 +1,1357 @@
+/* zlib.h -- interface of the 'zlib' general purpose compression library
+  version 1.2.3, July 18th, 2005
+
+  Copyright (C) 1995-2005 Jean-loup Gailly and Mark Adler
+
+  This software is provided 'as-is', without any express or implied
+  warranty.  In no event will the authors be held liable for any damages
+  arising from the use of this software.
+
+  Permission is granted to anyone to use this software for any purpose,
+  including commercial applications, and to alter it and redistribute it
+  freely, subject to the following restrictions:
+
+  1. The origin of this software must not be misrepresented; you must not
+     claim that you wrote the original software. If you use this software
+     in a product, an acknowledgment in the product documentation would be
+     appreciated but is not required.
+  2. Altered source versions must be plainly marked as such, and must not be
+     misrepresented as being the original software.
+  3. This notice may not be removed or altered from any source distribution.
+
+  Jean-loup Gailly        Mark Adler
+  jloup@gzip.org          madler@alumni.caltech.edu
+
+
+  The data format used by the zlib library is described by RFCs (Request for
+  Comments) 1950 to 1952 in the files http://www.ietf.org/rfc/rfc1950.txt
+  (zlib format), rfc1951.txt (deflate format) and rfc1952.txt (gzip format).
+*/
+
+#ifndef ZLIB_H
+#define ZLIB_H
+
+#include "zconf.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#define ZLIB_VERSION "1.2.3"
+#define ZLIB_VERNUM 0x1230
+
+/*
+     The 'zlib' compression library provides in-memory compression and
+  decompression functions, including integrity checks of the uncompressed
+  data.  This version of the library supports only one compression method
+  (deflation) but other algorithms will be added later and will have the same
+  stream interface.
+
+     Compression can be done in a single step if the buffers are large
+  enough (for example if an input file is mmap'ed), or can be done by
+  repeated calls of the compression function.  In the latter case, the
+  application must provide more input and/or consume the output
+  (providing more output space) before each call.
+
+     The compressed data format used by default by the in-memory functions is
+  the zlib format, which is a zlib wrapper documented in RFC 1950, wrapped
+  around a deflate stream, which is itself documented in RFC 1951.
+
+     The library also supports reading and writing files in gzip (.gz) format
+  with an interface similar to that of stdio using the functions that start
+  with "gz".  The gzip format is different from the zlib format.  gzip is a
+  gzip wrapper, documented in RFC 1952, wrapped around a deflate stream.
+
+     This library can optionally read and write gzip streams in memory as well.
+
+     The zlib format was designed to be compact and fast for use in memory
+  and on communications channels.  The gzip format was designed for single-
+  file compression on file systems, has a larger header than zlib to maintain
+  directory information, and uses a different, slower check method than zlib.
+
+     The library does not install any signal handler. The decoder checks
+  the consistency of the compressed data, so the library should never
+  crash even in case of corrupted input.
+*/
+
+typedef voidpf (*alloc_func) OF((voidpf opaque, uInt items, uInt size));
+typedef void   (*free_func)  OF((voidpf opaque, voidpf address));
+
+struct internal_state;
+
+typedef struct z_stream_s {
+    Bytef    *next_in;  /* next input byte */
+    uInt     avail_in;  /* number of bytes available at next_in */
+    uLong    total_in;  /* total nb of input bytes read so far */
+
+    Bytef    *next_out; /* next output byte should be put there */
+    uInt     avail_out; /* remaining free space at next_out */
+    uLong    total_out; /* total nb of bytes output so far */
+
+    char     *msg;      /* last error message, NULL if no error */
+    struct internal_state FAR *state; /* not visible by applications */
+
+    alloc_func zalloc;  /* used to allocate the internal state */
+    free_func  zfree;   /* used to free the internal state */
+    voidpf     opaque;  /* private data object passed to zalloc and zfree */
+
+    int     data_type;  /* best guess about the data type: binary or text */
+    uLong   adler;      /* adler32 value of the uncompressed data */
+    uLong   reserved;   /* reserved for future use */
+} z_stream;
+
+typedef z_stream FAR *z_streamp;
+
+/*
+     gzip header information passed to and from zlib routines.  See RFC 1952
+  for more details on the meanings of these fields.
+*/
+typedef struct gz_header_s {
+    int     text;       /* true if compressed data believed to be text */
+    uLong   time;       /* modification time */
+    int     xflags;     /* extra flags (not used when writing a gzip file) */
+    int     os;         /* operating system */
+    Bytef   *extra;     /* pointer to extra field or Z_NULL if none */
+    uInt    extra_len;  /* extra field length (valid if extra != Z_NULL) */
+    uInt    extra_max;  /* space at extra (only when reading header) */
+    Bytef   *name;      /* pointer to zero-terminated file name or Z_NULL */
+    uInt    name_max;   /* space at name (only when reading header) */
+    Bytef   *comment;   /* pointer to zero-terminated comment or Z_NULL */
+    uInt    comm_max;   /* space at comment (only when reading header) */
+    int     hcrc;       /* true if there was or will be a header crc */
+    int     done;       /* true when done reading gzip header (not used
+                           when writing a gzip file) */
+} gz_header;
+
+typedef gz_header FAR *gz_headerp;
+
+/*
+   The application must update next_in and avail_in when avail_in has
+   dropped to zero. It must update next_out and avail_out when avail_out
+   has dropped to zero. The application must initialize zalloc, zfree and
+   opaque before calling the init function. All other fields are set by the
+   compression library and must not be updated by the application.
+
+   The opaque value provided by the application will be passed as the first
+   parameter for calls of zalloc and zfree. This can be useful for custom
+   memory management. The compression library attaches no meaning to the
+   opaque value.
+
+   zalloc must return Z_NULL if there is not enough memory for the object.
+   If zlib is used in a multi-threaded application, zalloc and zfree must be
+   thread safe.
+
+   On 16-bit systems, the functions zalloc and zfree must be able to allocate
+   exactly 65536 bytes, but will not be required to allocate more than this
+   if the symbol MAXSEG_64K is defined (see zconf.h). WARNING: On MSDOS,
+   pointers returned by zalloc for objects of exactly 65536 bytes *must*
+   have their offset normalized to zero. The default allocation function
+   provided by this library ensures this (see zutil.c). To reduce memory
+   requirements and avoid any allocation of 64K objects, at the expense of
+   compression ratio, compile the library with -DMAX_WBITS=14 (see zconf.h).
+
+   The fields total_in and total_out can be used for statistics or
+   progress reports. After compression, total_in holds the total size of
+   the uncompressed data and may be saved for use in the decompressor
+   (particularly if the decompressor wants to decompress everything in
+   a single step).
+*/
+
+                        /* constants */
+
+#define Z_NO_FLUSH      0
+#define Z_PARTIAL_FLUSH 1 /* will be removed, use Z_SYNC_FLUSH instead */
+#define Z_SYNC_FLUSH    2
+#define Z_FULL_FLUSH    3
+#define Z_FINISH        4
+#define Z_BLOCK         5
+/* Allowed flush values; see deflate() and inflate() below for details */
+
+#define Z_OK            0
+#define Z_STREAM_END    1
+#define Z_NEED_DICT     2
+#define Z_ERRNO        (-1)
+#define Z_STREAM_ERROR (-2)
+#define Z_DATA_ERROR   (-3)
+#define Z_MEM_ERROR    (-4)
+#define Z_BUF_ERROR    (-5)
+#define Z_VERSION_ERROR (-6)
+/* Return codes for the compression/decompression functions. Negative
+ * values are errors, positive values are used for special but normal events.
+ */
+
+#define Z_NO_COMPRESSION         0
+#define Z_BEST_SPEED             1
+#define Z_BEST_COMPRESSION       9
+#define Z_DEFAULT_COMPRESSION  (-1)
+/* compression levels */
+
+#define Z_FILTERED            1
+#define Z_HUFFMAN_ONLY        2
+#define Z_RLE                 3
+#define Z_FIXED               4
+#define Z_DEFAULT_STRATEGY    0
+/* compression strategy; see deflateInit2() below for details */
+
+#define Z_BINARY   0
+#define Z_TEXT     1
+#define Z_ASCII    Z_TEXT   /* for compatibility with 1.2.2 and earlier */
+#define Z_UNKNOWN  2
+/* Possible values of the data_type field (though see inflate()) */
+
+#define Z_DEFLATED   8
+/* The deflate compression method (the only one supported in this version) */
+
+#define Z_NULL  0  /* for initializing zalloc, zfree, opaque */
+
+#define zlib_version zlibVersion()
+/* for compatibility with versions < 1.0.2 */
+
+                        /* basic functions */
+
+ZEXTERN const char * ZEXPORT zlibVersion OF((void));
+/* The application can compare zlibVersion and ZLIB_VERSION for consistency.
+   If the first character differs, the library code actually used is
+   not compatible with the zlib.h header file used by the application.
+   This check is automatically made by deflateInit and inflateInit.
+ */
+
+/*
+ZEXTERN int ZEXPORT deflateInit OF((z_streamp strm, int level));
+
+     Initializes the internal stream state for compression. The fields
+   zalloc, zfree and opaque must be initialized before by the caller.
+   If zalloc and zfree are set to Z_NULL, deflateInit updates them to
+   use default allocation functions.
+
+     The compression level must be Z_DEFAULT_COMPRESSION, or between 0 and 9:
+   1 gives best speed, 9 gives best compression, 0 gives no compression at
+   all (the input data is simply copied a block at a time).
+   Z_DEFAULT_COMPRESSION requests a default compromise between speed and
+   compression (currently equivalent to level 6).
+
+     deflateInit returns Z_OK if success, Z_MEM_ERROR if there was not
+   enough memory, Z_STREAM_ERROR if level is not a valid compression level,
+   Z_VERSION_ERROR if the zlib library version (zlib_version) is incompatible
+   with the version assumed by the caller (ZLIB_VERSION).
+   msg is set to null if there is no error message.  deflateInit does not
+   perform any compression: this will be done by deflate().
+*/
+
+
+ZEXTERN int ZEXPORT deflate OF((z_streamp strm, int flush));
+/*
+    deflate compresses as much data as possible, and stops when the input
+  buffer becomes empty or the output buffer becomes full. It may introduce some
+  output latency (reading input without producing any output) except when
+  forced to flush.
+
+    The detailed semantics are as follows. deflate performs one or both of the
+  following actions:
+
+  - Compress more input starting at next_in and update next_in and avail_in
+    accordingly. If not all input can be processed (because there is not
+    enough room in the output buffer), next_in and avail_in are updated and
+    processing will resume at this point for the next call of deflate().
+
+  - Provide more output starting at next_out and update next_out and avail_out
+    accordingly. This action is forced if the parameter flush is non zero.
+    Forcing flush frequently degrades the compression ratio, so this parameter
+    should be set only when necessary (in interactive applications).
+    Some output may be provided even if flush is not set.
+
+  Before the call of deflate(), the application should ensure that at least
+  one of the actions is possible, by providing more input and/or consuming
+  more output, and updating avail_in or avail_out accordingly; avail_out
+  should never be zero before the call. The application can consume the
+  compressed output when it wants, for example when the output buffer is full
+  (avail_out == 0), or after each call of deflate(). If deflate returns Z_OK
+  and with zero avail_out, it must be called again after making room in the
+  output buffer because there might be more output pending.
+
+    Normally the parameter flush is set to Z_NO_FLUSH, which allows deflate to
+  decide how much data to accumualte before producing output, in order to
+  maximize compression.
+
+    If the parameter flush is set to Z_SYNC_FLUSH, all pending output is
+  flushed to the output buffer and the output is aligned on a byte boundary, so
+  that the decompressor can get all input data available so far. (In particular
+  avail_in is zero after the call if enough output space has been provided
+  before the call.)  Flushing may degrade compression for some compression
+  algorithms and so it should be used only when necessary.
+
+    If flush is set to Z_FULL_FLUSH, all output is flushed as with
+  Z_SYNC_FLUSH, and the compression state is reset so that decompression can
+  restart from this point if previous compressed data has been damaged or if
+  random access is desired. Using Z_FULL_FLUSH too often can seriously degrade
+  compression.
+
+    If deflate returns with avail_out == 0, this function must be called again
+  with the same value of the flush parameter and more output space (updated
+  avail_out), until the flush is complete (deflate returns with non-zero
+  avail_out). In the case of a Z_FULL_FLUSH or Z_SYNC_FLUSH, make sure that
+  avail_out is greater than six to avoid repeated flush markers due to
+  avail_out == 0 on return.
+
+    If the parameter flush is set to Z_FINISH, pending input is processed,
+  pending output is flushed and deflate returns with Z_STREAM_END if there
+  was enough output space; if deflate returns with Z_OK, this function must be
+  called again with Z_FINISH and more output space (updated avail_out) but no
+  more input data, until it returns with Z_STREAM_END or an error. After
+  deflate has returned Z_STREAM_END, the only possible operations on the
+  stream are deflateReset or deflateEnd.
+
+    Z_FINISH can be used immediately after deflateInit if all the compression
+  is to be done in a single step. In this case, avail_out must be at least
+  the value returned by deflateBound (see below). If deflate does not return
+  Z_STREAM_END, then it must be called again as described above.
+
+    deflate() sets strm->adler to the adler32 checksum of all input read
+  so far (that is, total_in bytes).
+
+    deflate() may update strm->data_type if it can make a good guess about
+  the input data type (Z_BINARY or Z_TEXT). In doubt, the data is considered
+  binary. This field is only for information purposes and does not affect
+  the compression algorithm in any manner.
+
+    deflate() returns Z_OK if some progress has been made (more input
+  processed or more output produced), Z_STREAM_END if all input has been
+  consumed and all output has been produced (only when flush is set to
+  Z_FINISH), Z_STREAM_ERROR if the stream state was inconsistent (for example
+  if next_in or next_out was NULL), Z_BUF_ERROR if no progress is possible
+  (for example avail_in or avail_out was zero). Note that Z_BUF_ERROR is not
+  fatal, and deflate() can be called again with more input and more output
+  space to continue compressing.
+*/
+
+
+ZEXTERN int ZEXPORT deflateEnd OF((z_streamp strm));
+/*
+     All dynamically allocated data structures for this stream are freed.
+   This function discards any unprocessed input and does not flush any
+   pending output.
+
+     deflateEnd returns Z_OK if success, Z_STREAM_ERROR if the
+   stream state was inconsistent, Z_DATA_ERROR if the stream was freed
+   prematurely (some input or output was discarded). In the error case,
+   msg may be set but then points to a static string (which must not be
+   deallocated).
+*/
+
+
+/*
+ZEXTERN int ZEXPORT inflateInit OF((z_streamp strm));
+
+     Initializes the internal stream state for decompression. The fields
+   next_in, avail_in, zalloc, zfree and opaque must be initialized before by
+   the caller. If next_in is not Z_NULL and avail_in is large enough (the exact
+   value depends on the compression method), inflateInit determines the
+   compression method from the zlib header and allocates all data structures
+   accordingly; otherwise the allocation will be deferred to the first call of
+   inflate.  If zalloc and zfree are set to Z_NULL, inflateInit updates them to
+   use default allocation functions.
+
+     inflateInit returns Z_OK if success, Z_MEM_ERROR if there was not enough
+   memory, Z_VERSION_ERROR if the zlib library version is incompatible with the
+   version assumed by the caller.  msg is set to null if there is no error
+   message. inflateInit does not perform any decompression apart from reading
+   the zlib header if present: this will be done by inflate().  (So next_in and
+   avail_in may be modified, but next_out and avail_out are unchanged.)
+*/
+
+
+ZEXTERN int ZEXPORT inflate OF((z_streamp strm, int flush));
+/*
+    inflate decompresses as much data as possible, and stops when the input
+  buffer becomes empty or the output buffer becomes full. It may introduce
+  some output latency (reading input without producing any output) except when
+  forced to flush.
+
+  The detailed semantics are as follows. inflate performs one or both of the
+  following actions:
+
+  - Decompress more input starting at next_in and update next_in and avail_in
+    accordingly. If not all input can be processed (because there is not
+    enough room in the output buffer), next_in is updated and processing
+    will resume at this point for the next call of inflate().
+
+  - Provide more output starting at next_out and update next_out and avail_out
+    accordingly.  inflate() provides as much output as possible, until there
+    is no more input data or no more space in the output buffer (see below
+    about the flush parameter).
+
+  Before the call of inflate(), the application should ensure that at least
+  one of the actions is possible, by providing more input and/or consuming
+  more output, and updating the next_* and avail_* values accordingly.
+  The application can consume the uncompressed output when it wants, for
+  example when the output buffer is full (avail_out == 0), or after each
+  call of inflate(). If inflate returns Z_OK and with zero avail_out, it
+  must be called again after making room in the output buffer because there
+  might be more output pending.
+
+    The flush parameter of inflate() can be Z_NO_FLUSH, Z_SYNC_FLUSH,
+  Z_FINISH, or Z_BLOCK. Z_SYNC_FLUSH requests that inflate() flush as much
+  output as possible to the output buffer. Z_BLOCK requests that inflate() stop
+  if and when it gets to the next deflate block boundary. When decoding the
+  zlib or gzip format, this will cause inflate() to return immediately after
+  the header and before the first block. When doing a raw inflate, inflate()
+  will go ahead and process the first block, and will return when it gets to
+  the end of that block, or when it runs out of data.
+
+    The Z_BLOCK option assists in appending to or combining deflate streams.
+  Also to assist in this, on return inflate() will set strm->data_type to the
+  number of unused bits in the last byte taken from strm->next_in, plus 64
+  if inflate() is currently decoding the last block in the deflate stream,
+  plus 128 if inflate() returned immediately after decoding an end-of-block
+  code or decoding the complete header up to just before the first byte of the
+  deflate stream. The end-of-block will not be indicated until all of the
+  uncompressed data from that block has been written to strm->next_out.  The
+  number of unused bits may in general be greater than seven, except when
+  bit 7 of data_type is set, in which case the number of unused bits will be
+  less than eight.
+
+    inflate() should normally be called until it returns Z_STREAM_END or an
+  error. However if all decompression is to be performed in a single step
+  (a single call of inflate), the parameter flush should be set to
+  Z_FINISH. In this case all pending input is processed and all pending
+  output is flushed; avail_out must be large enough to hold all the
+  uncompressed data. (The size of the uncompressed data may have been saved
+  by the compressor for this purpose.) The next operation on this stream must
+  be inflateEnd to deallocate the decompression state. The use of Z_FINISH
+  is never required, but can be used to inform inflate that a faster approach
+  may be used for the single inflate() call.
+
+     In this implementation, inflate() always flushes as much output as
+  possible to the output buffer, and always uses the faster approach on the
+  first call. So the only effect of the flush parameter in this implementation
+  is on the return value of inflate(), as noted below, or when it returns early
+  because Z_BLOCK is used.
+
+     If a preset dictionary is needed after this call (see inflateSetDictionary
+  below), inflate sets strm->adler to the adler32 checksum of the dictionary
+  chosen by the compressor and returns Z_NEED_DICT; otherwise it sets
+  strm->adler to the adler32 checksum of all output produced so far (that is,
+  total_out bytes) and returns Z_OK, Z_STREAM_END or an error code as described
+  below. At the end of the stream, inflate() checks that its computed adler32
+  checksum is equal to that saved by the compressor and returns Z_STREAM_END
+  only if the checksum is correct.
+
+    inflate() will decompress and check either zlib-wrapped or gzip-wrapped
+  deflate data.  The header type is detected automatically.  Any information
+  contained in the gzip header is not retained, so applications that need that
+  information should instead use raw inflate, see inflateInit2() below, or
+  inflateBack() and perform their own processing of the gzip header and
+  trailer.
+
+    inflate() returns Z_OK if some progress has been made (more input processed
+  or more output produced), Z_STREAM_END if the end of the compressed data has
+  been reached and all uncompressed output has been produced, Z_NEED_DICT if a
+  preset dictionary is needed at this point, Z_DATA_ERROR if the input data was
+  corrupted (input stream not conforming to the zlib format or incorrect check
+  value), Z_STREAM_ERROR if the stream structure was inconsistent (for example
+  if next_in or next_out was NULL), Z_MEM_ERROR if there was not enough memory,
+  Z_BUF_ERROR if no progress is possible or if there was not enough room in the
+  output buffer when Z_FINISH is used. Note that Z_BUF_ERROR is not fatal, and
+  inflate() can be called again with more input and more output space to
+  continue decompressing. If Z_DATA_ERROR is returned, the application may then
+  call inflateSync() to look for a good compression block if a partial recovery
+  of the data is desired.
+*/
+
+
+ZEXTERN int ZEXPORT inflateEnd OF((z_streamp strm));
+/*
+     All dynamically allocated data structures for this stream are freed.
+   This function discards any unprocessed input and does not flush any
+   pending output.
+
+     inflateEnd returns Z_OK if success, Z_STREAM_ERROR if the stream state
+   was inconsistent. In the error case, msg may be set but then points to a
+   static string (which must not be deallocated).
+*/
+
+                        /* Advanced functions */
+
+/*
+    The following functions are needed only in some special applications.
+*/
+
+/*
+ZEXTERN int ZEXPORT deflateInit2 OF((z_streamp strm,
+                                     int  level,
+                                     int  method,
+                                     int  windowBits,
+                                     int  memLevel,
+                                     int  strategy));
+
+     This is another version of deflateInit with more compression options. The
+   fields next_in, zalloc, zfree and opaque must be initialized before by
+   the caller.
+
+     The method parameter is the compression method. It must be Z_DEFLATED in
+   this version of the library.
+
+     The windowBits parameter is the base two logarithm of the window size
+   (the size of the history buffer). It should be in the range 8..15 for this
+   version of the library. Larger values of this parameter result in better
+   compression at the expense of memory usage. The default value is 15 if
+   deflateInit is used instead.
+
+     windowBits can also be -8..-15 for raw deflate. In this case, -windowBits
+   determines the window size. deflate() will then generate raw deflate data
+   with no zlib header or trailer, and will not compute an adler32 check value.
+
+     windowBits can also be greater than 15 for optional gzip encoding. Add
+   16 to windowBits to write a simple gzip header and trailer around the
+   compressed data instead of a zlib wrapper. The gzip header will have no
+   file name, no extra data, no comment, no modification time (set to zero),
+   no header crc, and the operating system will be set to 255 (unknown).  If a
+   gzip stream is being written, strm->adler is a crc32 instead of an adler32.
+
+     The memLevel parameter specifies how much memory should be allocated
+   for the internal compression state. memLevel=1 uses minimum memory but
+   is slow and reduces compression ratio; memLevel=9 uses maximum memory
+   for optimal speed. The default value is 8. See zconf.h for total memory
+   usage as a function of windowBits and memLevel.
+
+     The strategy parameter is used to tune the compression algorithm. Use the
+   value Z_DEFAULT_STRATEGY for normal data, Z_FILTERED for data produced by a
+   filter (or predictor), Z_HUFFMAN_ONLY to force Huffman encoding only (no
+   string match), or Z_RLE to limit match distances to one (run-length
+   encoding). Filtered data consists mostly of small values with a somewhat
+   random distribution. In this case, the compression algorithm is tuned to
+   compress them better. The effect of Z_FILTERED is to force more Huffman
+   coding and less string matching; it is somewhat intermediate between
+   Z_DEFAULT and Z_HUFFMAN_ONLY. Z_RLE is designed to be almost as fast as
+   Z_HUFFMAN_ONLY, but give better compression for PNG image data. The strategy
+   parameter only affects the compression ratio but not the correctness of the
+   compressed output even if it is not set appropriately.  Z_FIXED prevents the
+   use of dynamic Huffman codes, allowing for a simpler decoder for special
+   applications.
+
+      deflateInit2 returns Z_OK if success, Z_MEM_ERROR if there was not enough
+   memory, Z_STREAM_ERROR if a parameter is invalid (such as an invalid
+   method). msg is set to null if there is no error message.  deflateInit2 does
+   not perform any compression: this will be done by deflate().
+*/
+
+ZEXTERN int ZEXPORT deflateSetDictionary OF((z_streamp strm,
+                                             const Bytef *dictionary,
+                                             uInt  dictLength));
+/*
+     Initializes the compression dictionary from the given byte sequence
+   without producing any compressed output. This function must be called
+   immediately after deflateInit, deflateInit2 or deflateReset, before any
+   call of deflate. The compressor and decompressor must use exactly the same
+   dictionary (see inflateSetDictionary).
+
+     The dictionary should consist of strings (byte sequences) that are likely
+   to be encountered later in the data to be compressed, with the most commonly
+   used strings preferably put towards the end of the dictionary. Using a
+   dictionary is most useful when the data to be compressed is short and can be
+   predicted with good accuracy; the data can then be compressed better than
+   with the default empty dictionary.
+
+     Depending on the size of the compression data structures selected by
+   deflateInit or deflateInit2, a part of the dictionary may in effect be
+   discarded, for example if the dictionary is larger than the window size in
+   deflate or deflate2. Thus the strings most likely to be useful should be
+   put at the end of the dictionary, not at the front. In addition, the
+   current implementation of deflate will use at most the window size minus
+   262 bytes of the provided dictionary.
+
+     Upon return of this function, strm->adler is set to the adler32 value
+   of the dictionary; the decompressor may later use this value to determine
+   which dictionary has been used by the compressor. (The adler32 value
+   applies to the whole dictionary even if only a subset of the dictionary is
+   actually used by the compressor.) If a raw deflate was requested, then the
+   adler32 value is not computed and strm->adler is not set.
+
+     deflateSetDictionary returns Z_OK if success, or Z_STREAM_ERROR if a
+   parameter is invalid (such as NULL dictionary) or the stream state is
+   inconsistent (for example if deflate has already been called for this stream
+   or if the compression method is bsort). deflateSetDictionary does not
+   perform any compression: this will be done by deflate().
+*/
+
+ZEXTERN int ZEXPORT deflateCopy OF((z_streamp dest,
+                                    z_streamp source));
+/*
+     Sets the destination stream as a complete copy of the source stream.
+
+     This function can be useful when several compression strategies will be
+   tried, for example when there are several ways of pre-processing the input
+   data with a filter. The streams that will be discarded should then be freed
+   by calling deflateEnd.  Note that deflateCopy duplicates the internal
+   compression state which can be quite large, so this strategy is slow and
+   can consume lots of memory.
+
+     deflateCopy returns Z_OK if success, Z_MEM_ERROR if there was not
+   enough memory, Z_STREAM_ERROR if the source stream state was inconsistent
+   (such as zalloc being NULL). msg is left unchanged in both source and
+   destination.
+*/
+
+ZEXTERN int ZEXPORT deflateReset OF((z_streamp strm));
+/*
+     This function is equivalent to deflateEnd followed by deflateInit,
+   but does not free and reallocate all the internal compression state.
+   The stream will keep the same compression level and any other attributes
+   that may have been set by deflateInit2.
+
+      deflateReset returns Z_OK if success, or Z_STREAM_ERROR if the source
+   stream state was inconsistent (such as zalloc or state being NULL).
+*/
+
+ZEXTERN int ZEXPORT deflateParams OF((z_streamp strm,
+                                      int level,
+                                      int strategy));
+/*
+     Dynamically update the compression level and compression strategy.  The
+   interpretation of level and strategy is as in deflateInit2.  This can be
+   used to switch between compression and straight copy of the input data, or
+   to switch to a different kind of input data requiring a different
+   strategy. If the compression level is changed, the input available so far
+   is compressed with the old level (and may be flushed); the new level will
+   take effect only at the next call of deflate().
+
+     Before the call of deflateParams, the stream state must be set as for
+   a call of deflate(), since the currently available input may have to
+   be compressed and flushed. In particular, strm->avail_out must be non-zero.
+
+     deflateParams returns Z_OK if success, Z_STREAM_ERROR if the source
+   stream state was inconsistent or if a parameter was invalid, Z_BUF_ERROR
+   if strm->avail_out was zero.
+*/
+
+ZEXTERN int ZEXPORT deflateTune OF((z_streamp strm,
+                                    int good_length,
+                                    int max_lazy,
+                                    int nice_length,
+                                    int max_chain));
+/*
+     Fine tune deflate's internal compression parameters.  This should only be
+   used by someone who understands the algorithm used by zlib's deflate for
+   searching for the best matching string, and even then only by the most
+   fanatic optimizer trying to squeeze out the last compressed bit for their
+   specific input data.  Read the deflate.c source code for the meaning of the
+   max_lazy, good_length, nice_length, and max_chain parameters.
+
+     deflateTune() can be called after deflateInit() or deflateInit2(), and
+   returns Z_OK on success, or Z_STREAM_ERROR for an invalid deflate stream.
+ */
+
+ZEXTERN uLong ZEXPORT deflateBound OF((z_streamp strm,
+                                       uLong sourceLen));
+/*
+     deflateBound() returns an upper bound on the compressed size after
+   deflation of sourceLen bytes.  It must be called after deflateInit()
+   or deflateInit2().  This would be used to allocate an output buffer
+   for deflation in a single pass, and so would be called before deflate().
+*/
+
+ZEXTERN int ZEXPORT deflatePrime OF((z_streamp strm,
+                                     int bits,
+                                     int value));
+/*
+     deflatePrime() inserts bits in the deflate output stream.  The intent
+  is that this function is used to start off the deflate output with the
+  bits leftover from a previous deflate stream when appending to it.  As such,
+  this function can only be used for raw deflate, and must be used before the
+  first deflate() call after a deflateInit2() or deflateReset().  bits must be
+  less than or equal to 16, and that many of the least significant bits of
+  value will be inserted in the output.
+
+      deflatePrime returns Z_OK if success, or Z_STREAM_ERROR if the source
+   stream state was inconsistent.
+*/
+
+ZEXTERN int ZEXPORT deflateSetHeader OF((z_streamp strm,
+                                         gz_headerp head));
+/*
+      deflateSetHeader() provides gzip header information for when a gzip
+   stream is requested by deflateInit2().  deflateSetHeader() may be called
+   after deflateInit2() or deflateReset() and before the first call of
+   deflate().  The text, time, os, extra field, name, and comment information
+   in the provided gz_header structure are written to the gzip header (xflag is
+   ignored -- the extra flags are set according to the compression level).  The
+   caller must assure that, if not Z_NULL, name and comment are terminated with
+   a zero byte, and that if extra is not Z_NULL, that extra_len bytes are
+   available there.  If hcrc is true, a gzip header crc is included.  Note that
+   the current versions of the command-line version of gzip (up through version
+   1.3.x) do not support header crc's, and will report that it is a "multi-part
+   gzip file" and give up.
+
+      If deflateSetHeader is not used, the default gzip header has text false,
+   the time set to zero, and os set to 255, with no extra, name, or comment
+   fields.  The gzip header is returned to the default state by deflateReset().
+
+      deflateSetHeader returns Z_OK if success, or Z_STREAM_ERROR if the source
+   stream state was inconsistent.
+*/
+
+/*
+ZEXTERN int ZEXPORT inflateInit2 OF((z_streamp strm,
+                                     int  windowBits));
+
+     This is another version of inflateInit with an extra parameter. The
+   fields next_in, avail_in, zalloc, zfree and opaque must be initialized
+   before by the caller.
+
+     The windowBits parameter is the base two logarithm of the maximum window
+   size (the size of the history buffer).  It should be in the range 8..15 for
+   this version of the library. The default value is 15 if inflateInit is used
+   instead. windowBits must be greater than or equal to the windowBits value
+   provided to deflateInit2() while compressing, or it must be equal to 15 if
+   deflateInit2() was not used. If a compressed stream with a larger window
+   size is given as input, inflate() will return with the error code
+   Z_DATA_ERROR instead of trying to allocate a larger window.
+
+     windowBits can also be -8..-15 for raw inflate. In this case, -windowBits
+   determines the window size. inflate() will then process raw deflate data,
+   not looking for a zlib or gzip header, not generating a check value, and not
+   looking for any check values for comparison at the end of the stream. This
+   is for use with other formats that use the deflate compressed data format
+   such as zip.  Those formats provide their own check values. If a custom
+   format is developed using the raw deflate format for compressed data, it is
+   recommended that a check value such as an adler32 or a crc32 be applied to
+   the uncompressed data as is done in the zlib, gzip, and zip formats.  For
+   most applications, the zlib format should be used as is. Note that comments
+   above on the use in deflateInit2() applies to the magnitude of windowBits.
+
+     windowBits can also be greater than 15 for optional gzip decoding. Add
+   32 to windowBits to enable zlib and gzip decoding with automatic header
+   detection, or add 16 to decode only the gzip format (the zlib format will
+   return a Z_DATA_ERROR).  If a gzip stream is being decoded, strm->adler is
+   a crc32 instead of an adler32.
+
+     inflateInit2 returns Z_OK if success, Z_MEM_ERROR if there was not enough
+   memory, Z_STREAM_ERROR if a parameter is invalid (such as a null strm). msg
+   is set to null if there is no error message.  inflateInit2 does not perform
+   any decompression apart from reading the zlib header if present: this will
+   be done by inflate(). (So next_in and avail_in may be modified, but next_out
+   and avail_out are unchanged.)
+*/
+
+ZEXTERN int ZEXPORT inflateSetDictionary OF((z_streamp strm,
+                                             const Bytef *dictionary,
+                                             uInt  dictLength));
+/*
+     Initializes the decompression dictionary from the given uncompressed byte
+   sequence. This function must be called immediately after a call of inflate,
+   if that call returned Z_NEED_DICT. The dictionary chosen by the compressor
+   can be determined from the adler32 value returned by that call of inflate.
+   The compressor and decompressor must use exactly the same dictionary (see
+   deflateSetDictionary).  For raw inflate, this function can be called
+   immediately after inflateInit2() or inflateReset() and before any call of
+   inflate() to set the dictionary.  The application must insure that the
+   dictionary that was used for compression is provided.
+
+     inflateSetDictionary returns Z_OK if success, Z_STREAM_ERROR if a
+   parameter is invalid (such as NULL dictionary) or the stream state is
+   inconsistent, Z_DATA_ERROR if the given dictionary doesn't match the
+   expected one (incorrect adler32 value). inflateSetDictionary does not
+   perform any decompression: this will be done by subsequent calls of
+   inflate().
+*/
+
+ZEXTERN int ZEXPORT inflateSync OF((z_streamp strm));
+/*
+    Skips invalid compressed data until a full flush point (see above the
+  description of deflate with Z_FULL_FLUSH) can be found, or until all
+  available input is skipped. No output is provided.
+
+    inflateSync returns Z_OK if a full flush point has been found, Z_BUF_ERROR
+  if no more input was provided, Z_DATA_ERROR if no flush point has been found,
+  or Z_STREAM_ERROR if the stream structure was inconsistent. In the success
+  case, the application may save the current current value of total_in which
+  indicates where valid compressed data was found. In the error case, the
+  application may repeatedly call inflateSync, providing more input each time,
+  until success or end of the input data.
+*/
+
+ZEXTERN int ZEXPORT inflateCopy OF((z_streamp dest,
+                                    z_streamp source));
+/*
+     Sets the destination stream as a complete copy of the source stream.
+
+     This function can be useful when randomly accessing a large stream.  The
+   first pass through the stream can periodically record the inflate state,
+   allowing restarting inflate at those points when randomly accessing the
+   stream.
+
+     inflateCopy returns Z_OK if success, Z_MEM_ERROR if there was not
+   enough memory, Z_STREAM_ERROR if the source stream state was inconsistent
+   (such as zalloc being NULL). msg is left unchanged in both source and
+   destination.
+*/
+
+ZEXTERN int ZEXPORT inflateReset OF((z_streamp strm));
+/*
+     This function is equivalent to inflateEnd followed by inflateInit,
+   but does not free and reallocate all the internal decompression state.
+   The stream will keep attributes that may have been set by inflateInit2.
+
+      inflateReset returns Z_OK if success, or Z_STREAM_ERROR if the source
+   stream state was inconsistent (such as zalloc or state being NULL).
+*/
+
+ZEXTERN int ZEXPORT inflatePrime OF((z_streamp strm,
+                                     int bits,
+                                     int value));
+/*
+     This function inserts bits in the inflate input stream.  The intent is
+  that this function is used to start inflating at a bit position in the
+  middle of a byte.  The provided bits will be used before any bytes are used
+  from next_in.  This function should only be used with raw inflate, and
+  should be used before the first inflate() call after inflateInit2() or
+  inflateReset().  bits must be less than or equal to 16, and that many of the
+  least significant bits of value will be inserted in the input.
+
+      inflatePrime returns Z_OK if success, or Z_STREAM_ERROR if the source
+   stream state was inconsistent.
+*/
+
+ZEXTERN int ZEXPORT inflateGetHeader OF((z_streamp strm,
+                                         gz_headerp head));
+/*
+      inflateGetHeader() requests that gzip header information be stored in the
+   provided gz_header structure.  inflateGetHeader() may be called after
+   inflateInit2() or inflateReset(), and before the first call of inflate().
+   As inflate() processes the gzip stream, head->done is zero until the header
+   is completed, at which time head->done is set to one.  If a zlib stream is
+   being decoded, then head->done is set to -1 to indicate that there will be
+   no gzip header information forthcoming.  Note that Z_BLOCK can be used to
+   force inflate() to return immediately after header processing is complete
+   and before any actual data is decompressed.
+
+      The text, time, xflags, and os fields are filled in with the gzip header
+   contents.  hcrc is set to true if there is a header CRC.  (The header CRC
+   was valid if done is set to one.)  If extra is not Z_NULL, then extra_max
+   contains the maximum number of bytes to write to extra.  Once done is true,
+   extra_len contains the actual extra field length, and extra contains the
+   extra field, or that field truncated if extra_max is less than extra_len.
+   If name is not Z_NULL, then up to name_max characters are written there,
+   terminated with a zero unless the length is greater than name_max.  If
+   comment is not Z_NULL, then up to comm_max characters are written there,
+   terminated with a zero unless the length is greater than comm_max.  When
+   any of extra, name, or comment are not Z_NULL and the respective field is
+   not present in the header, then that field is set to Z_NULL to signal its
+   absence.  This allows the use of deflateSetHeader() with the returned
+   structure to duplicate the header.  However if those fields are set to
+   allocated memory, then the application will need to save those pointers
+   elsewhere so that they can be eventually freed.
+
+      If inflateGetHeader is not used, then the header information is simply
+   discarded.  The header is always checked for validity, including the header
+   CRC if present.  inflateReset() will reset the process to discard the header
+   information.  The application would need to call inflateGetHeader() again to
+   retrieve the header from the next gzip stream.
+
+      inflateGetHeader returns Z_OK if success, or Z_STREAM_ERROR if the source
+   stream state was inconsistent.
+*/
+
+/*
+ZEXTERN int ZEXPORT inflateBackInit OF((z_streamp strm, int windowBits,
+                                        unsigned char FAR *window));
+
+     Initialize the internal stream state for decompression using inflateBack()
+   calls.  The fields zalloc, zfree and opaque in strm must be initialized
+   before the call.  If zalloc and zfree are Z_NULL, then the default library-
+   derived memory allocation routines are used.  windowBits is the base two
+   logarithm of the window size, in the range 8..15.  window is a caller
+   supplied buffer of that size.  Except for special applications where it is
+   assured that deflate was used with small window sizes, windowBits must be 15
+   and a 32K byte window must be supplied to be able to decompress general
+   deflate streams.
+
+     See inflateBack() for the usage of these routines.
+
+     inflateBackInit will return Z_OK on success, Z_STREAM_ERROR if any of
+   the paramaters are invalid, Z_MEM_ERROR if the internal state could not
+   be allocated, or Z_VERSION_ERROR if the version of the library does not
+   match the version of the header file.
+*/
+
+typedef unsigned (*in_func) OF((void FAR *, unsigned char FAR * FAR *));
+typedef int (*out_func) OF((void FAR *, unsigned char FAR *, unsigned));
+
+ZEXTERN int ZEXPORT inflateBack OF((z_streamp strm,
+                                    in_func in, void FAR *in_desc,
+                                    out_func out, void FAR *out_desc));
+/*
+     inflateBack() does a raw inflate with a single call using a call-back
+   interface for input and output.  This is more efficient than inflate() for
+   file i/o applications in that it avoids copying between the output and the
+   sliding window by simply making the window itself the output buffer.  This
+   function trusts the application to not change the output buffer passed by
+   the output function, at least until inflateBack() returns.
+
+     inflateBackInit() must be called first to allocate the internal state
+   and to initialize the state with the user-provided window buffer.
+   inflateBack() may then be used multiple times to inflate a complete, raw
+   deflate stream with each call.  inflateBackEnd() is then called to free
+   the allocated state.
+
+     A raw deflate stream is one with no zlib or gzip header or trailer.
+   This routine would normally be used in a utility that reads zip or gzip
+   files and writes out uncompressed files.  The utility would decode the
+   header and process the trailer on its own, hence this routine expects
+   only the raw deflate stream to decompress.  This is different from the
+   normal behavior of inflate(), which expects either a zlib or gzip header and
+   trailer around the deflate stream.
+
+     inflateBack() uses two subroutines supplied by the caller that are then
+   called by inflateBack() for input and output.  inflateBack() calls those
+   routines until it reads a complete deflate stream and writes out all of the
+   uncompressed data, or until it encounters an error.  The function's
+   parameters and return types are defined above in the in_func and out_func
+   typedefs.  inflateBack() will call in(in_desc, &buf) which should return the
+   number of bytes of provided input, and a pointer to that input in buf.  If
+   there is no input available, in() must return zero--buf is ignored in that
+   case--and inflateBack() will return a buffer error.  inflateBack() will call
+   out(out_desc, buf, len) to write the uncompressed data buf[0..len-1].  out()
+   should return zero on success, or non-zero on failure.  If out() returns
+   non-zero, inflateBack() will return with an error.  Neither in() nor out()
+   are permitted to change the contents of the window provided to
+   inflateBackInit(), which is also the buffer that out() uses to write from.
+   The length written by out() will be at most the window size.  Any non-zero
+   amount of input may be provided by in().
+
+     For convenience, inflateBack() can be provided input on the first call by
+   setting strm->next_in and strm->avail_in.  If that input is exhausted, then
+   in() will be called.  Therefore strm->next_in must be initialized before
+   calling inflateBack().  If strm->next_in is Z_NULL, then in() will be called
+   immediately for input.  If strm->next_in is not Z_NULL, then strm->avail_in
+   must also be initialized, and then if strm->avail_in is not zero, input will
+   initially be taken from strm->next_in[0 .. strm->avail_in - 1].
+
+     The in_desc and out_desc parameters of inflateBack() is passed as the
+   first parameter of in() and out() respectively when they are called.  These
+   descriptors can be optionally used to pass any information that the caller-
+   supplied in() and out() functions need to do their job.
+
+     On return, inflateBack() will set strm->next_in and strm->avail_in to
+   pass back any unused input that was provided by the last in() call.  The
+   return values of inflateBack() can be Z_STREAM_END on success, Z_BUF_ERROR
+   if in() or out() returned an error, Z_DATA_ERROR if there was a format
+   error in the deflate stream (in which case strm->msg is set to indicate the
+   nature of the error), or Z_STREAM_ERROR if the stream was not properly
+   initialized.  In the case of Z_BUF_ERROR, an input or output error can be
+   distinguished using strm->next_in which will be Z_NULL only if in() returned
+   an error.  If strm->next is not Z_NULL, then the Z_BUF_ERROR was due to
+   out() returning non-zero.  (in() will always be called before out(), so
+   strm->next_in is assured to be defined if out() returns non-zero.)  Note
+   that inflateBack() cannot return Z_OK.
+*/
+
+ZEXTERN int ZEXPORT inflateBackEnd OF((z_streamp strm));
+/*
+     All memory allocated by inflateBackInit() is freed.
+
+     inflateBackEnd() returns Z_OK on success, or Z_STREAM_ERROR if the stream
+   state was inconsistent.
+*/
+
+ZEXTERN uLong ZEXPORT zlibCompileFlags OF((void));
+/* Return flags indicating compile-time options.
+
+    Type sizes, two bits each, 00 = 16 bits, 01 = 32, 10 = 64, 11 = other:
+     1.0: size of uInt
+     3.2: size of uLong
+     5.4: size of voidpf (pointer)
+     7.6: size of z_off_t
+
+    Compiler, assembler, and debug options:
+     8: DEBUG
+     9: ASMV or ASMINF -- use ASM code
+     10: ZLIB_WINAPI -- exported functions use the WINAPI calling convention
+     11: 0 (reserved)
+
+    One-time table building (smaller code, but not thread-safe if true):
+     12: BUILDFIXED -- build static block decoding tables when needed
+     13: DYNAMIC_CRC_TABLE -- build CRC calculation tables when needed
+     14,15: 0 (reserved)
+
+    Library content (indicates missing functionality):
+     16: NO_GZCOMPRESS -- gz* functions cannot compress (to avoid linking
+                          deflate code when not needed)
+     17: NO_GZIP -- deflate can't write gzip streams, and inflate can't detect
+                    and decode gzip streams (to avoid linking crc code)
+     18-19: 0 (reserved)
+
+    Operation variations (changes in library functionality):
+     20: PKZIP_BUG_WORKAROUND -- slightly more permissive inflate
+     21: FASTEST -- deflate algorithm with only one, lowest compression level
+     22,23: 0 (reserved)
+
+    The sprintf variant used by gzprintf (zero is best):
+     24: 0 = vs*, 1 = s* -- 1 means limited to 20 arguments after the format
+     25: 0 = *nprintf, 1 = *printf -- 1 means gzprintf() not secure!
+     26: 0 = returns value, 1 = void -- 1 means inferred string length returned
+
+    Remainder:
+     27-31: 0 (reserved)
+ */
+
+
+                        /* utility functions */
+
+/*
+     The following utility functions are implemented on top of the
+   basic stream-oriented functions. To simplify the interface, some
+   default options are assumed (compression level and memory usage,
+   standard memory allocation functions). The source code of these
+   utility functions can easily be modified if you need special options.
+*/
+
+ZEXTERN int ZEXPORT compress OF((Bytef *dest,   uLongf *destLen,
+                                 const Bytef *source, uLong sourceLen));
+/*
+     Compresses the source buffer into the destination buffer.  sourceLen is
+   the byte length of the source buffer. Upon entry, destLen is the total
+   size of the destination buffer, which must be at least the value returned
+   by compressBound(sourceLen). Upon exit, destLen is the actual size of the
+   compressed buffer.
+     This function can be used to compress a whole file at once if the
+   input file is mmap'ed.
+     compress returns Z_OK if success, Z_MEM_ERROR if there was not
+   enough memory, Z_BUF_ERROR if there was not enough room in the output
+   buffer.
+*/
+
+ZEXTERN int ZEXPORT compress2 OF((Bytef *dest,   uLongf *destLen,
+                                  const Bytef *source, uLong sourceLen,
+                                  int level));
+/*
+     Compresses the source buffer into the destination buffer. The level
+   parameter has the same meaning as in deflateInit.  sourceLen is the byte
+   length of the source buffer. Upon entry, destLen is the total size of the
+   destination buffer, which must be at least the value returned by
+   compressBound(sourceLen). Upon exit, destLen is the actual size of the
+   compressed buffer.
+
+     compress2 returns Z_OK if success, Z_MEM_ERROR if there was not enough
+   memory, Z_BUF_ERROR if there was not enough room in the output buffer,
+   Z_STREAM_ERROR if the level parameter is invalid.
+*/
+
+ZEXTERN uLong ZEXPORT compressBound OF((uLong sourceLen));
+/*
+     compressBound() returns an upper bound on the compressed size after
+   compress() or compress2() on sourceLen bytes.  It would be used before
+   a compress() or compress2() call to allocate the destination buffer.
+*/
+
+ZEXTERN int ZEXPORT uncompress OF((Bytef *dest,   uLongf *destLen,
+                                   const Bytef *source, uLong sourceLen));
+/*
+     Decompresses the source buffer into the destination buffer.  sourceLen is
+   the byte length of the source buffer. Upon entry, destLen is the total
+   size of the destination buffer, which must be large enough to hold the
+   entire uncompressed data. (The size of the uncompressed data must have
+   been saved previously by the compressor and transmitted to the decompressor
+   by some mechanism outside the scope of this compression library.)
+   Upon exit, destLen is the actual size of the compressed buffer.
+     This function can be used to decompress a whole file at once if the
+   input file is mmap'ed.
+
+     uncompress returns Z_OK if success, Z_MEM_ERROR if there was not
+   enough memory, Z_BUF_ERROR if there was not enough room in the output
+   buffer, or Z_DATA_ERROR if the input data was corrupted or incomplete.
+*/
+
+
+typedef voidp gzFile;
+
+ZEXTERN gzFile ZEXPORT gzopen  OF((const char *path, const char *mode));
+/*
+     Opens a gzip (.gz) file for reading or writing. The mode parameter
+   is as in fopen ("rb" or "wb") but can also include a compression level
+   ("wb9") or a strategy: 'f' for filtered data as in "wb6f", 'h' for
+   Huffman only compression as in "wb1h", or 'R' for run-length encoding
+   as in "wb1R". (See the description of deflateInit2 for more information
+   about the strategy parameter.)
+
+     gzopen can be used to read a file which is not in gzip format; in this
+   case gzread will directly read from the file without decompression.
+
+     gzopen returns NULL if the file could not be opened or if there was
+   insufficient memory to allocate the (de)compression state; errno
+   can be checked to distinguish the two cases (if errno is zero, the
+   zlib error is Z_MEM_ERROR).  */
+
+ZEXTERN gzFile ZEXPORT gzdopen  OF((int fd, const char *mode));
+/*
+     gzdopen() associates a gzFile with the file descriptor fd.  File
+   descriptors are obtained from calls like open, dup, creat, pipe or
+   fileno (in the file has been previously opened with fopen).
+   The mode parameter is as in gzopen.
+     The next call of gzclose on the returned gzFile will also close the
+   file descriptor fd, just like fclose(fdopen(fd), mode) closes the file
+   descriptor fd. If you want to keep fd open, use gzdopen(dup(fd), mode).
+     gzdopen returns NULL if there was insufficient memory to allocate
+   the (de)compression state.
+*/
+
+ZEXTERN int ZEXPORT gzsetparams OF((gzFile file, int level, int strategy));
+/*
+     Dynamically update the compression level or strategy. See the description
+   of deflateInit2 for the meaning of these parameters.
+     gzsetparams returns Z_OK if success, or Z_STREAM_ERROR if the file was not
+   opened for writing.
+*/
+
+ZEXTERN int ZEXPORT    gzread  OF((gzFile file, voidp buf, unsigned len));
+/*
+     Reads the given number of uncompressed bytes from the compressed file.
+   If the input file was not in gzip format, gzread copies the given number
+   of bytes into the buffer.
+     gzread returns the number of uncompressed bytes actually read (0 for
+   end of file, -1 for error). */
+
+ZEXTERN int ZEXPORT    gzwrite OF((gzFile file,
+                                   voidpc buf, unsigned len));
+/*
+     Writes the given number of uncompressed bytes into the compressed file.
+   gzwrite returns the number of uncompressed bytes actually written
+   (0 in case of error).
+*/
+
+ZEXTERN int ZEXPORTVA   gzprintf OF((gzFile file, const char *format, ...));
+/*
+     Converts, formats, and writes the args to the compressed file under
+   control of the format string, as in fprintf. gzprintf returns the number of
+   uncompressed bytes actually written (0 in case of error).  The number of
+   uncompressed bytes written is limited to 4095. The caller should assure that
+   this limit is not exceeded. If it is exceeded, then gzprintf() will return
+   return an error (0) with nothing written. In this case, there may also be a
+   buffer overflow with unpredictable consequences, which is possible only if
+   zlib was compiled with the insecure functions sprintf() or vsprintf()
+   because the secure snprintf() or vsnprintf() functions were not available.
+*/
+
+ZEXTERN int ZEXPORT gzputs OF((gzFile file, const char *s));
+/*
+      Writes the given null-terminated string to the compressed file, excluding
+   the terminating null character.
+      gzputs returns the number of characters written, or -1 in case of error.
+*/
+
+ZEXTERN char * ZEXPORT gzgets OF((gzFile file, char *buf, int len));
+/*
+      Reads bytes from the compressed file until len-1 characters are read, or
+   a newline character is read and transferred to buf, or an end-of-file
+   condition is encountered.  The string is then terminated with a null
+   character.
+      gzgets returns buf, or Z_NULL in case of error.
+*/
+
+ZEXTERN int ZEXPORT    gzputc OF((gzFile file, int c));
+/*
+      Writes c, converted to an unsigned char, into the compressed file.
+   gzputc returns the value that was written, or -1 in case of error.
+*/
+
+ZEXTERN int ZEXPORT    gzgetc OF((gzFile file));
+/*
+      Reads one byte from the compressed file. gzgetc returns this byte
+   or -1 in case of end of file or error.
+*/
+
+ZEXTERN int ZEXPORT    gzungetc OF((int c, gzFile file));
+/*
+      Push one character back onto the stream to be read again later.
+   Only one character of push-back is allowed.  gzungetc() returns the
+   character pushed, or -1 on failure.  gzungetc() will fail if a
+   character has been pushed but not read yet, or if c is -1. The pushed
+   character will be discarded if the stream is repositioned with gzseek()
+   or gzrewind().
+*/
+
+ZEXTERN int ZEXPORT    gzflush OF((gzFile file, int flush));
+/*
+     Flushes all pending output into the compressed file. The parameter
+   flush is as in the deflate() function. The return value is the zlib
+   error number (see function gzerror below). gzflush returns Z_OK if
+   the flush parameter is Z_FINISH and all output could be flushed.
+     gzflush should be called only when strictly necessary because it can
+   degrade compression.
+*/
+
+ZEXTERN z_off_t ZEXPORT    gzseek OF((gzFile file,
+                                      z_off_t offset, int whence));
+/*
+      Sets the starting position for the next gzread or gzwrite on the
+   given compressed file. The offset represents a number of bytes in the
+   uncompressed data stream. The whence parameter is defined as in lseek(2);
+   the value SEEK_END is not supported.
+     If the file is opened for reading, this function is emulated but can be
+   extremely slow. If the file is opened for writing, only forward seeks are
+   supported; gzseek then compresses a sequence of zeroes up to the new
+   starting position.
+
+      gzseek returns the resulting offset location as measured in bytes from
+   the beginning of the uncompressed stream, or -1 in case of error, in
+   particular if the file is opened for writing and the new starting position
+   would be before the current position.
+*/
+
+ZEXTERN int ZEXPORT    gzrewind OF((gzFile file));
+/*
+     Rewinds the given file. This function is supported only for reading.
+
+   gzrewind(file) is equivalent to (int)gzseek(file, 0L, SEEK_SET)
+*/
+
+ZEXTERN z_off_t ZEXPORT    gztell OF((gzFile file));
+/*
+     Returns the starting position for the next gzread or gzwrite on the
+   given compressed file. This position represents a number of bytes in the
+   uncompressed data stream.
+
+   gztell(file) is equivalent to gzseek(file, 0L, SEEK_CUR)
+*/
+
+ZEXTERN int ZEXPORT gzeof OF((gzFile file));
+/*
+     Returns 1 when EOF has previously been detected reading the given
+   input stream, otherwise zero.
+*/
+
+ZEXTERN int ZEXPORT gzdirect OF((gzFile file));
+/*
+     Returns 1 if file is being read directly without decompression, otherwise
+   zero.
+*/
+
+ZEXTERN int ZEXPORT    gzclose OF((gzFile file));
+/*
+     Flushes all pending output if necessary, closes the compressed file
+   and deallocates all the (de)compression state. The return value is the zlib
+   error number (see function gzerror below).
+*/
+
+ZEXTERN const char * ZEXPORT gzerror OF((gzFile file, int *errnum));
+/*
+     Returns the error message for the last error which occurred on the
+   given compressed file. errnum is set to zlib error number. If an
+   error occurred in the file system and not in the compression library,
+   errnum is set to Z_ERRNO and the application may consult errno
+   to get the exact error code.
+*/
+
+ZEXTERN void ZEXPORT gzclearerr OF((gzFile file));
+/*
+     Clears the error and end-of-file flags for file. This is analogous to the
+   clearerr() function in stdio. This is useful for continuing to read a gzip
+   file that is being written concurrently.
+*/
+
+                        /* checksum functions */
+
+/*
+     These functions are not related to compression but are exported
+   anyway because they might be useful in applications using the
+   compression library.
+*/
+
+ZEXTERN uLong ZEXPORT adler32 OF((uLong adler, const Bytef *buf, uInt len));
+/*
+     Update a running Adler-32 checksum with the bytes buf[0..len-1] and
+   return the updated checksum. If buf is NULL, this function returns
+   the required initial value for the checksum.
+   An Adler-32 checksum is almost as reliable as a CRC32 but can be computed
+   much faster. Usage example:
+
+     uLong adler = adler32(0L, Z_NULL, 0);
+
+     while (read_buffer(buffer, length) != EOF) {
+       adler = adler32(adler, buffer, length);
+     }
+     if (adler != original_adler) error();
+*/
+
+ZEXTERN uLong ZEXPORT adler32_combine OF((uLong adler1, uLong adler2,
+                                          z_off_t len2));
+/*
+     Combine two Adler-32 checksums into one.  For two sequences of bytes, seq1
+   and seq2 with lengths len1 and len2, Adler-32 checksums were calculated for
+   each, adler1 and adler2.  adler32_combine() returns the Adler-32 checksum of
+   seq1 and seq2 concatenated, requiring only adler1, adler2, and len2.
+*/
+
+ZEXTERN uLong ZEXPORT crc32   OF((uLong crc, const Bytef *buf, uInt len));
+/*
+     Update a running CRC-32 with the bytes buf[0..len-1] and return the
+   updated CRC-32. If buf is NULL, this function returns the required initial
+   value for the for the crc. Pre- and post-conditioning (one's complement) is
+   performed within this function so it shouldn't be done by the application.
+   Usage example:
+
+     uLong crc = crc32(0L, Z_NULL, 0);
+
+     while (read_buffer(buffer, length) != EOF) {
+       crc = crc32(crc, buffer, length);
+     }
+     if (crc != original_crc) error();
+*/
+
+ZEXTERN uLong ZEXPORT crc32_combine OF((uLong crc1, uLong crc2, z_off_t len2));
+
+/*
+     Combine two CRC-32 check values into one.  For two sequences of bytes,
+   seq1 and seq2 with lengths len1 and len2, CRC-32 check values were
+   calculated for each, crc1 and crc2.  crc32_combine() returns the CRC-32
+   check value of seq1 and seq2 concatenated, requiring only crc1, crc2, and
+   len2.
+*/
+
+
+                        /* various hacks, don't look :) */
+
+/* deflateInit and inflateInit are macros to allow checking the zlib version
+ * and the compiler's view of z_stream:
+ */
+ZEXTERN int ZEXPORT deflateInit_ OF((z_streamp strm, int level,
+                                     const char *version, int stream_size));
+ZEXTERN int ZEXPORT inflateInit_ OF((z_streamp strm,
+                                     const char *version, int stream_size));
+ZEXTERN int ZEXPORT deflateInit2_ OF((z_streamp strm, int  level, int  method,
+                                      int windowBits, int memLevel,
+                                      int strategy, const char *version,
+                                      int stream_size));
+ZEXTERN int ZEXPORT inflateInit2_ OF((z_streamp strm, int  windowBits,
+                                      const char *version, int stream_size));
+ZEXTERN int ZEXPORT inflateBackInit_ OF((z_streamp strm, int windowBits,
+                                         unsigned char FAR *window,
+                                         const char *version,
+                                         int stream_size));
+#define deflateInit(strm, level) \
+        deflateInit_((strm), (level),       ZLIB_VERSION, sizeof(z_stream))
+#define inflateInit(strm) \
+        inflateInit_((strm),                ZLIB_VERSION, sizeof(z_stream))
+#define deflateInit2(strm, level, method, windowBits, memLevel, strategy) \
+        deflateInit2_((strm),(level),(method),(windowBits),(memLevel),\
+                      (strategy),           ZLIB_VERSION, sizeof(z_stream))
+#define inflateInit2(strm, windowBits) \
+        inflateInit2_((strm), (windowBits), ZLIB_VERSION, sizeof(z_stream))
+#define inflateBackInit(strm, windowBits, window) \
+        inflateBackInit_((strm), (windowBits), (window), \
+        ZLIB_VERSION, sizeof(z_stream))
+
+
+#if !defined(ZUTIL_H) && !defined(NO_DUMMY_DECL)
+    struct internal_state {int dummy;}; /* hack for buggy compilers */
+#endif
+
+ZEXTERN const char   * ZEXPORT zError           OF((int));
+ZEXTERN int            ZEXPORT inflateSyncPoint OF((z_streamp z));
+ZEXTERN const uLongf * ZEXPORT get_crc_table    OF((void));
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* ZLIB_H */
diff --git a/distrib/zlib-1.2.3/zutil.c b/distrib/zlib-1.2.3/zutil.c
new file mode 100644
index 0000000..d55f594
--- /dev/null
+++ b/distrib/zlib-1.2.3/zutil.c
@@ -0,0 +1,318 @@
+/* zutil.c -- target dependent utility functions for the compression library
+ * Copyright (C) 1995-2005 Jean-loup Gailly.
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* @(#) $Id$ */
+
+#include "zutil.h"
+
+#ifndef NO_DUMMY_DECL
+struct internal_state      {int dummy;}; /* for buggy compilers */
+#endif
+
+const char * const z_errmsg[10] = {
+"need dictionary",     /* Z_NEED_DICT       2  */
+"stream end",          /* Z_STREAM_END      1  */
+"",                    /* Z_OK              0  */
+"file error",          /* Z_ERRNO         (-1) */
+"stream error",        /* Z_STREAM_ERROR  (-2) */
+"data error",          /* Z_DATA_ERROR    (-3) */
+"insufficient memory", /* Z_MEM_ERROR     (-4) */
+"buffer error",        /* Z_BUF_ERROR     (-5) */
+"incompatible version",/* Z_VERSION_ERROR (-6) */
+""};
+
+
+const char * ZEXPORT zlibVersion()
+{
+    return ZLIB_VERSION;
+}
+
+uLong ZEXPORT zlibCompileFlags()
+{
+    uLong flags;
+
+    flags = 0;
+    switch (sizeof(uInt)) {
+    case 2:     break;
+    case 4:     flags += 1;     break;
+    case 8:     flags += 2;     break;
+    default:    flags += 3;
+    }
+    switch (sizeof(uLong)) {
+    case 2:     break;
+    case 4:     flags += 1 << 2;        break;
+    case 8:     flags += 2 << 2;        break;
+    default:    flags += 3 << 2;
+    }
+    switch (sizeof(voidpf)) {
+    case 2:     break;
+    case 4:     flags += 1 << 4;        break;
+    case 8:     flags += 2 << 4;        break;
+    default:    flags += 3 << 4;
+    }
+    switch (sizeof(z_off_t)) {
+    case 2:     break;
+    case 4:     flags += 1 << 6;        break;
+    case 8:     flags += 2 << 6;        break;
+    default:    flags += 3 << 6;
+    }
+#ifdef DEBUG
+    flags += 1 << 8;
+#endif
+#if defined(ASMV) || defined(ASMINF)
+    flags += 1 << 9;
+#endif
+#ifdef ZLIB_WINAPI
+    flags += 1 << 10;
+#endif
+#ifdef BUILDFIXED
+    flags += 1 << 12;
+#endif
+#ifdef DYNAMIC_CRC_TABLE
+    flags += 1 << 13;
+#endif
+#ifdef NO_GZCOMPRESS
+    flags += 1L << 16;
+#endif
+#ifdef NO_GZIP
+    flags += 1L << 17;
+#endif
+#ifdef PKZIP_BUG_WORKAROUND
+    flags += 1L << 20;
+#endif
+#ifdef FASTEST
+    flags += 1L << 21;
+#endif
+#ifdef STDC
+#  ifdef NO_vsnprintf
+        flags += 1L << 25;
+#    ifdef HAS_vsprintf_void
+        flags += 1L << 26;
+#    endif
+#  else
+#    ifdef HAS_vsnprintf_void
+        flags += 1L << 26;
+#    endif
+#  endif
+#else
+        flags += 1L << 24;
+#  ifdef NO_snprintf
+        flags += 1L << 25;
+#    ifdef HAS_sprintf_void
+        flags += 1L << 26;
+#    endif
+#  else
+#    ifdef HAS_snprintf_void
+        flags += 1L << 26;
+#    endif
+#  endif
+#endif
+    return flags;
+}
+
+#ifdef DEBUG
+
+#  ifndef verbose
+#    define verbose 0
+#  endif
+int z_verbose = verbose;
+
+void z_error (m)
+    char *m;
+{
+    fprintf(stderr, "%s\n", m);
+    exit(1);
+}
+#endif
+
+/* exported to allow conversion of error code to string for compress() and
+ * uncompress()
+ */
+const char * ZEXPORT zError(err)
+    int err;
+{
+    return ERR_MSG(err);
+}
+
+#if defined(_WIN32_WCE)
+    /* The Microsoft C Run-Time Library for Windows CE doesn't have
+     * errno.  We define it as a global variable to simplify porting.
+     * Its value is always 0 and should not be used.
+     */
+    int errno = 0;
+#endif
+
+#ifndef HAVE_MEMCPY
+
+void zmemcpy(dest, source, len)
+    Bytef* dest;
+    const Bytef* source;
+    uInt  len;
+{
+    if (len == 0) return;
+    do {
+        *dest++ = *source++; /* ??? to be unrolled */
+    } while (--len != 0);
+}
+
+int zmemcmp(s1, s2, len)
+    const Bytef* s1;
+    const Bytef* s2;
+    uInt  len;
+{
+    uInt j;
+
+    for (j = 0; j < len; j++) {
+        if (s1[j] != s2[j]) return 2*(s1[j] > s2[j])-1;
+    }
+    return 0;
+}
+
+void zmemzero(dest, len)
+    Bytef* dest;
+    uInt  len;
+{
+    if (len == 0) return;
+    do {
+        *dest++ = 0;  /* ??? to be unrolled */
+    } while (--len != 0);
+}
+#endif
+
+
+#ifdef SYS16BIT
+
+#ifdef __TURBOC__
+/* Turbo C in 16-bit mode */
+
+#  define MY_ZCALLOC
+
+/* Turbo C malloc() does not allow dynamic allocation of 64K bytes
+ * and farmalloc(64K) returns a pointer with an offset of 8, so we
+ * must fix the pointer. Warning: the pointer must be put back to its
+ * original form in order to free it, use zcfree().
+ */
+
+#define MAX_PTR 10
+/* 10*64K = 640K */
+
+local int next_ptr = 0;
+
+typedef struct ptr_table_s {
+    voidpf org_ptr;
+    voidpf new_ptr;
+} ptr_table;
+
+local ptr_table table[MAX_PTR];
+/* This table is used to remember the original form of pointers
+ * to large buffers (64K). Such pointers are normalized with a zero offset.
+ * Since MSDOS is not a preemptive multitasking OS, this table is not
+ * protected from concurrent access. This hack doesn't work anyway on
+ * a protected system like OS/2. Use Microsoft C instead.
+ */
+
+voidpf zcalloc (voidpf opaque, unsigned items, unsigned size)
+{
+    voidpf buf = opaque; /* just to make some compilers happy */
+    ulg bsize = (ulg)items*size;
+
+    /* If we allocate less than 65520 bytes, we assume that farmalloc
+     * will return a usable pointer which doesn't have to be normalized.
+     */
+    if (bsize < 65520L) {
+        buf = farmalloc(bsize);
+        if (*(ush*)&buf != 0) return buf;
+    } else {
+        buf = farmalloc(bsize + 16L);
+    }
+    if (buf == NULL || next_ptr >= MAX_PTR) return NULL;
+    table[next_ptr].org_ptr = buf;
+
+    /* Normalize the pointer to seg:0 */
+    *((ush*)&buf+1) += ((ush)((uch*)buf-0) + 15) >> 4;
+    *(ush*)&buf = 0;
+    table[next_ptr++].new_ptr = buf;
+    return buf;
+}
+
+void  zcfree (voidpf opaque, voidpf ptr)
+{
+    int n;
+    if (*(ush*)&ptr != 0) { /* object < 64K */
+        farfree(ptr);
+        return;
+    }
+    /* Find the original pointer */
+    for (n = 0; n < next_ptr; n++) {
+        if (ptr != table[n].new_ptr) continue;
+
+        farfree(table[n].org_ptr);
+        while (++n < next_ptr) {
+            table[n-1] = table[n];
+        }
+        next_ptr--;
+        return;
+    }
+    ptr = opaque; /* just to make some compilers happy */
+    Assert(0, "zcfree: ptr not found");
+}
+
+#endif /* __TURBOC__ */
+
+
+#ifdef M_I86
+/* Microsoft C in 16-bit mode */
+
+#  define MY_ZCALLOC
+
+#if (!defined(_MSC_VER) || (_MSC_VER <= 600))
+#  define _halloc  halloc
+#  define _hfree   hfree
+#endif
+
+voidpf zcalloc (voidpf opaque, unsigned items, unsigned size)
+{
+    if (opaque) opaque = 0; /* to make compiler happy */
+    return _halloc((long)items, size);
+}
+
+void  zcfree (voidpf opaque, voidpf ptr)
+{
+    if (opaque) opaque = 0; /* to make compiler happy */
+    _hfree(ptr);
+}
+
+#endif /* M_I86 */
+
+#endif /* SYS16BIT */
+
+
+#ifndef MY_ZCALLOC /* Any system without a special alloc function */
+
+#ifndef STDC
+extern voidp  malloc OF((uInt size));
+extern voidp  calloc OF((uInt items, uInt size));
+extern void   free   OF((voidpf ptr));
+#endif
+
+voidpf zcalloc (opaque, items, size)
+    voidpf opaque;
+    unsigned items;
+    unsigned size;
+{
+    if (opaque) items += size - size; /* make compiler happy */
+    return sizeof(uInt) > 2 ? (voidpf)malloc(items * size) :
+                              (voidpf)calloc(items, size);
+}
+
+void  zcfree (opaque, ptr)
+    voidpf opaque;
+    voidpf ptr;
+{
+    free(ptr);
+    if (opaque) return; /* make compiler happy */
+}
+
+#endif /* MY_ZCALLOC */
diff --git a/distrib/zlib-1.2.3/zutil.h b/distrib/zlib-1.2.3/zutil.h
new file mode 100644
index 0000000..b7d5eff
--- /dev/null
+++ b/distrib/zlib-1.2.3/zutil.h
@@ -0,0 +1,269 @@
+/* zutil.h -- internal interface and configuration of the compression library
+ * Copyright (C) 1995-2005 Jean-loup Gailly.
+ * For conditions of distribution and use, see copyright notice in zlib.h
+ */
+
+/* WARNING: this file should *not* be used by applications. It is
+   part of the implementation of the compression library and is
+   subject to change. Applications should only use zlib.h.
+ */
+
+/* @(#) $Id$ */
+
+#ifndef ZUTIL_H
+#define ZUTIL_H
+
+#define ZLIB_INTERNAL
+#include "zlib.h"
+
+#ifdef STDC
+#  ifndef _WIN32_WCE
+#    include <stddef.h>
+#  endif
+#  include <string.h>
+#  include <stdlib.h>
+#endif
+#ifdef NO_ERRNO_H
+#   ifdef _WIN32_WCE
+      /* The Microsoft C Run-Time Library for Windows CE doesn't have
+       * errno.  We define it as a global variable to simplify porting.
+       * Its value is always 0 and should not be used.  We rename it to
+       * avoid conflict with other libraries that use the same workaround.
+       */
+#     define errno z_errno
+#   endif
+    extern int errno;
+#else
+#  ifndef _WIN32_WCE
+#    include <errno.h>
+#  endif
+#endif
+
+#ifndef local
+#  define local static
+#endif
+/* compile with -Dlocal if your debugger can't find static symbols */
+
+typedef unsigned char  uch;
+typedef uch FAR uchf;
+typedef unsigned short ush;
+typedef ush FAR ushf;
+typedef unsigned long  ulg;
+
+extern const char * const z_errmsg[10]; /* indexed by 2-zlib_error */
+/* (size given to avoid silly warnings with Visual C++) */
+
+#define ERR_MSG(err) z_errmsg[Z_NEED_DICT-(err)]
+
+#define ERR_RETURN(strm,err) \
+  return (strm->msg = (char*)ERR_MSG(err), (err))
+/* To be used only when the state is known to be valid */
+
+        /* common constants */
+
+#ifndef DEF_WBITS
+#  define DEF_WBITS MAX_WBITS
+#endif
+/* default windowBits for decompression. MAX_WBITS is for compression only */
+
+#if MAX_MEM_LEVEL >= 8
+#  define DEF_MEM_LEVEL 8
+#else
+#  define DEF_MEM_LEVEL  MAX_MEM_LEVEL
+#endif
+/* default memLevel */
+
+#define STORED_BLOCK 0
+#define STATIC_TREES 1
+#define DYN_TREES    2
+/* The three kinds of block type */
+
+#define MIN_MATCH  3
+#define MAX_MATCH  258
+/* The minimum and maximum match lengths */
+
+#define PRESET_DICT 0x20 /* preset dictionary flag in zlib header */
+
+        /* target dependencies */
+
+#if defined(MSDOS) || (defined(WINDOWS) && !defined(WIN32))
+#  define OS_CODE  0x00
+#  if defined(__TURBOC__) || defined(__BORLANDC__)
+#    if(__STDC__ == 1) && (defined(__LARGE__) || defined(__COMPACT__))
+       /* Allow compilation with ANSI keywords only enabled */
+       void _Cdecl farfree( void *block );
+       void *_Cdecl farmalloc( unsigned long nbytes );
+#    else
+#      include <alloc.h>
+#    endif
+#  else /* MSC or DJGPP */
+#    include <malloc.h>
+#  endif
+#endif
+
+#ifdef AMIGA
+#  define OS_CODE  0x01
+#endif
+
+#if defined(VAXC) || defined(VMS)
+#  define OS_CODE  0x02
+#  define F_OPEN(name, mode) \
+     fopen((name), (mode), "mbc=60", "ctx=stm", "rfm=fix", "mrs=512")
+#endif
+
+#if defined(ATARI) || defined(atarist)
+#  define OS_CODE  0x05
+#endif
+
+#ifdef OS2
+#  define OS_CODE  0x06
+#  ifdef M_I86
+     #include <malloc.h>
+#  endif
+#endif
+
+#if defined(MACOS) || defined(TARGET_OS_MAC)
+#  define OS_CODE  0x07
+#  if defined(__MWERKS__) && __dest_os != __be_os && __dest_os != __win32_os
+#    include <unix.h> /* for fdopen */
+#  else
+#    ifndef fdopen
+#      define fdopen(fd,mode) NULL /* No fdopen() */
+#    endif
+#  endif
+#endif
+
+#ifdef TOPS20
+#  define OS_CODE  0x0a
+#endif
+
+#ifdef WIN32
+#  ifndef __CYGWIN__  /* Cygwin is Unix, not Win32 */
+#    define OS_CODE  0x0b
+#  endif
+#endif
+
+#ifdef __50SERIES /* Prime/PRIMOS */
+#  define OS_CODE  0x0f
+#endif
+
+#if defined(_BEOS_) || defined(RISCOS)
+#  define fdopen(fd,mode) NULL /* No fdopen() */
+#endif
+
+#if (defined(_MSC_VER) && (_MSC_VER > 600))
+#  if defined(_WIN32_WCE)
+#    define fdopen(fd,mode) NULL /* No fdopen() */
+#    ifndef _PTRDIFF_T_DEFINED
+       typedef int ptrdiff_t;
+#      define _PTRDIFF_T_DEFINED
+#    endif
+#  else
+#    define fdopen(fd,type)  _fdopen(fd,type)
+#  endif
+#endif
+
+        /* common defaults */
+
+#ifndef OS_CODE
+#  define OS_CODE  0x03  /* assume Unix */
+#endif
+
+#ifndef F_OPEN
+#  define F_OPEN(name, mode) fopen((name), (mode))
+#endif
+
+         /* functions */
+
+#if defined(STDC99) || (defined(__TURBOC__) && __TURBOC__ >= 0x550)
+#  ifndef HAVE_VSNPRINTF
+#    define HAVE_VSNPRINTF
+#  endif
+#endif
+#if defined(__CYGWIN__)
+#  ifndef HAVE_VSNPRINTF
+#    define HAVE_VSNPRINTF
+#  endif
+#endif
+#ifndef HAVE_VSNPRINTF
+#  ifdef MSDOS
+     /* vsnprintf may exist on some MS-DOS compilers (DJGPP?),
+        but for now we just assume it doesn't. */
+#    define NO_vsnprintf
+#  endif
+#  ifdef __TURBOC__
+#    define NO_vsnprintf
+#  endif
+#  ifdef WIN32
+     /* In Win32, vsnprintf is available as the "non-ANSI" _vsnprintf. */
+#    if !defined(vsnprintf) && !defined(NO_vsnprintf)
+#      define vsnprintf _vsnprintf
+#    endif
+#  endif
+#  ifdef __SASC
+#    define NO_vsnprintf
+#  endif
+#endif
+#ifdef VMS
+#  define NO_vsnprintf
+#endif
+
+#if defined(pyr)
+#  define NO_MEMCPY
+#endif
+#if defined(SMALL_MEDIUM) && !defined(_MSC_VER) && !defined(__SC__)
+ /* Use our own functions for small and medium model with MSC <= 5.0.
+  * You may have to use the same strategy for Borland C (untested).
+  * The __SC__ check is for Symantec.
+  */
+#  define NO_MEMCPY
+#endif
+#if defined(STDC) && !defined(HAVE_MEMCPY) && !defined(NO_MEMCPY)
+#  define HAVE_MEMCPY
+#endif
+#ifdef HAVE_MEMCPY
+#  ifdef SMALL_MEDIUM /* MSDOS small or medium model */
+#    define zmemcpy _fmemcpy
+#    define zmemcmp _fmemcmp
+#    define zmemzero(dest, len) _fmemset(dest, 0, len)
+#  else
+#    define zmemcpy memcpy
+#    define zmemcmp memcmp
+#    define zmemzero(dest, len) memset(dest, 0, len)
+#  endif
+#else
+   extern void zmemcpy  OF((Bytef* dest, const Bytef* source, uInt len));
+   extern int  zmemcmp  OF((const Bytef* s1, const Bytef* s2, uInt len));
+   extern void zmemzero OF((Bytef* dest, uInt len));
+#endif
+
+/* Diagnostic functions */
+#ifdef DEBUG
+#  include <stdio.h>
+   extern int z_verbose;
+   extern void z_error    OF((char *m));
+#  define Assert(cond,msg) {if(!(cond)) z_error(msg);}
+#  define Trace(x) {if (z_verbose>=0) fprintf x ;}
+#  define Tracev(x) {if (z_verbose>0) fprintf x ;}
+#  define Tracevv(x) {if (z_verbose>1) fprintf x ;}
+#  define Tracec(c,x) {if (z_verbose>0 && (c)) fprintf x ;}
+#  define Tracecv(c,x) {if (z_verbose>1 && (c)) fprintf x ;}
+#else
+#  define Assert(cond,msg)
+#  define Trace(x)
+#  define Tracev(x)
+#  define Tracevv(x)
+#  define Tracec(c,x)
+#  define Tracecv(c,x)
+#endif
+
+
+voidpf zcalloc OF((voidpf opaque, unsigned items, unsigned size));
+void   zcfree  OF((voidpf opaque, voidpf ptr));
+
+#define ZALLOC(strm, items, size) \
+           (*((strm)->zalloc))((strm)->opaque, (items), (size))
+#define ZFREE(strm, addr)  (*((strm)->zfree))((strm)->opaque, (voidpf)(addr))
+#define TRY_FREE(s, p) {if (p) ZFREE(s, p);}
+
+#endif /* ZUTIL_H */