Move generic isPrint and columnWidth implementations to a separate header/source to allow using both generic and system-dependent versions on win32.

Summary:
This is needed so we can use generic columnWidthUTF8 in clang-format on
win32 simultaneously with a separate system-dependent implementations of
isPrint/columnWidth in TextDiagnostic.cpp to avoid attempts to print Unicode
characters using narrow-character interfaces (which is not supported on Windows,
and we'll have to figure out how to handle this).

Reviewers: jordan_rose

Reviewed By: jordan_rose

CC: llvm-commits, klimek

Differential Revision: http://llvm-reviews.chandlerc.com/D1559

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@189952 91177308-0d34-0410-b5e6-96231b3b80d8

5 files changed, 41 insertions, 115 deletions

diff --git a/lib/Support/CMakeLists.txt b/lib/Support/CMakeLists.txt
index 5823836..3aecf3f 100644
--- a/lib/Support/CMakeLists.txt
+++ b/lib/Support/CMakeLists.txt
@@ -54,6 +54,7 @@ add_llvm_library(LLVMSupport
   ToolOutputFile.cpp
   Triple.cpp
   Twine.cpp
+  Unicode.cpp
   YAMLParser.cpp
   YAMLTraits.cpp
   raw_os_ostream.cpp
diff --git a/lib/Support/Locale.cpp b/lib/Support/Locale.cpp
index 17b9b6c..35ddf7f 100644
--- a/lib/Support/Locale.cpp
+++ b/lib/Support/Locale.cpp
@@ -1,10 +1,31 @@
 #include "llvm/Support/Locale.h"
-#include "llvm/Config/config.h"
+#include "llvm/Support/Unicode.h"
 
-#ifdef __APPLE__
-#include "LocaleXlocale.inc"
-#elif LLVM_ON_WIN32
-#include "LocaleWindows.inc"
+namespace llvm {
+namespace sys {
+namespace locale {
+
+int columnWidth(StringRef Text) {
+#if LLVM_ON_WIN32
+  return Text.size();
 #else
-#include "LocaleGeneric.inc"
+  return llvm::sys::unicode::columnWidthUTF8(Text);
 #endif
+}
+
+bool isPrint(int UCS) {
+#if LLVM_ON_WIN32
+  // Restrict characters that we'll try to print to the the lower part of ASCII
+  // except for the control characters (0x20 - 0x7E). In general one can not
+  // reliably output code points U+0080 and higher using narrow character C/C++
+  // output functions in Windows, because the meaning of the upper 128 codes is
+  // determined by the active code page in the console.
+  return ' ' <= UCS && UCS <= '~';
+#else
+  return llvm::sys::unicode::isPrintable(UCS);
+#endif
+}
+
+} // namespace locale
+} // namespace sys
+} // namespace llvm
diff --git a/lib/Support/LocaleWindows.inc b/lib/Support/LocaleWindows.inc
deleted file mode 100644
index 28e429c..0000000
--- a/lib/Support/LocaleWindows.inc
+++ /dev/null
@@ -1,15 +0,0 @@
-namespace llvm {
-namespace sys {
-namespace locale {
-
-int columnWidth(StringRef s) {
-    return s.size();
-}
-
-bool isPrint(int c) {
-    return ' ' <= c && c <= '~';
-}
-
-}
-}
-}
diff --git a/lib/Support/LocaleXlocale.inc b/lib/Support/LocaleXlocale.inc
deleted file mode 100644
index 389fe3d..0000000
--- a/lib/Support/LocaleXlocale.inc
+++ /dev/null
@@ -1,61 +0,0 @@
-#include "llvm/ADT/SmallString.h"
-#include "llvm/ADT/SmallVector.h"
-#include "llvm/Support/ManagedStatic.h"
-#include <cassert>
-#include <xlocale.h>
-
-
-namespace {
-  struct locale_holder {
-    locale_holder()
-    : l(newlocale(LC_CTYPE_MASK,"en_US.UTF-8",LC_GLOBAL_LOCALE))
-    {
-      assert(NULL!=l);
-    }
-    ~locale_holder() {
-      freelocale(l);
-    }
-
-    int mbswidth(llvm::SmallString<16> s) const {
-       // this implementation assumes no '\0' in s
-      assert(s.size()==strlen(s.c_str()));
-
-      size_t size = mbstowcs_l(NULL,s.c_str(),0,l);
-      assert(size!=(size_t)-1);
-      if (size==0)
-        return 0;
-      llvm::SmallVector<wchar_t,200> ws(size);
-      size = mbstowcs_l(&ws[0],s.c_str(),ws.size(),l);
-      assert(ws.size()==size);
-      return wcswidth_l(&ws[0],ws.size(),l);
-    }
-
-    int isprint(int c) const {
-      return iswprint_l(c,l);
-    }
-
-  private:
-
-    locale_t l;
-  };
-
-  llvm::ManagedStatic<locale_holder> l;
-}
-
-namespace llvm {
-namespace sys {
-namespace locale {
-
-int columnWidth(StringRef s) {
-  int width = l->mbswidth(s);
-  assert(width>=0);
-  return width;
-}
-
-bool isPrint(int c) {
-  return l->isprint(c);
-}
-
-}
-}
-}
diff --git a/lib/Support/LocaleGeneric.inc b/lib/Support/Unicode.cpp
index 9fb8953..b719bd8 100644
--- a/lib/Support/LocaleGeneric.inc
+++ b/lib/Support/Unicode.cpp
@@ -1,4 +1,4 @@
-//===- llvm/Support/LocaleGeneric.inc - Locale-dependent stuff  -*- C++ -*-===//
+//===- llvm/Support/Unicode.cpp - Unicode character properties  -*- C++ -*-===//
 //
 //                     The LLVM Compiler Infrastructure
 //
@@ -7,41 +7,20 @@
 //
 //===----------------------------------------------------------------------===//
 //
-// This file implements llvm::sys::locale::columnWidth and
-// llvm::sys::locale::isPrint functions for UTF-8 locales.
+// This file implements functions that allow querying certain properties of
+// Unicode characters.
 //
 //===----------------------------------------------------------------------===//
 
-#include "llvm/ADT/ArrayRef.h"
+#include "llvm/Support/Unicode.h"
 #include "llvm/Support/ConvertUTF.h"
-#include "llvm/Support/raw_ostream.h"
 #include "llvm/Support/UnicodeCharRanges.h"
 
 namespace llvm {
 namespace sys {
-namespace locale {
+namespace unicode {
 
-enum ColumnWidthErrors {
-  ErrorInvalidUTF8 = -2,
-  ErrorNonPrintableCharacter = -1
-};
-
-/// Determines if a character is likely to be displayed correctly on the
-/// terminal. Exact implementation would have to depend on the specific
-/// terminal, so we define the semantic that should be suitable for generic case
-/// of a terminal capable to output Unicode characters.
-/// All characters from the Unicode codepoint range are considered printable
-/// except for:
-///   * C0 and C1 control character ranges;
-///   * default ignorable code points as per 5.21 of
-///     http://www.unicode.org/versions/Unicode6.2.0/UnicodeStandard-6.2.pdf
-///     except for U+00AD SOFT HYPHEN, as it's actually displayed on most
-///     terminals;
-///   * format characters (category = Cf);
-///   * surrogates (category = Cs);
-///   * unassigned characters (category = Cn).
-/// \return true if the character is considered printable.
-bool isPrint(int UCS) {
+bool isPrintable(int UCS) {
   // Sorted list of non-overlapping intervals of code points that are not
   // supposed to be printable.
   static const UnicodeCharRange NonPrintableRanges[] = {
@@ -241,13 +220,13 @@ bool isPrint(int UCS) {
 /// with a generic Unicode-capable terminal.
 /// \return Character width:
 ///   * ErrorNonPrintableCharacter (-1) for non-printable characters (as
-///     identified by isPrint);
+///     identified by isPrintable);
 ///   * 0 for non-spacing and enclosing combining marks;
 ///   * 2 for CJK characters excluding halfwidth forms;
 ///   * 1 for all remaining characters.
 static inline int charWidth(int UCS)
 {
-  if (!isPrint(UCS))
+  if (!isPrintable(UCS))
     return ErrorNonPrintableCharacter;
 
   // Sorted list of non-spacing and enclosing combining mark intervals as
@@ -361,7 +340,7 @@ static inline int charWidth(int UCS)
   return 1;
 }
 
-int columnWidth(StringRef Text) {
+int columnWidthUTF8(StringRef Text) {
   unsigned ColumnWidth = 0;
   unsigned Length;
   for (size_t i = 0, e = Text.size(); i < e; i += Length) {
@@ -382,6 +361,7 @@ int columnWidth(StringRef Text) {
   return ColumnWidth;
 }
 
-}
-}
-}
+} // namespace unicode
+} // namespace sys
+} // namespace llvm
+