| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
| |
Make the dex target use a different property to read
the input folder for classes to dexify.
By default the new property has the same value but can
be overriden if a post-compile target modify (obfuscate)
code into a different location.
Merged from master into tools_r7
Change-Id: I1fe2fd2a2678d8c4354702250cea1481f74dd7cf
|
|
|
|
|
|
|
| |
android.bat: properly encode file using CRLF (not Unix mode).
find_java.bat: properly check for Java in the 64-bit ProgramFiles path.
Change-Id: If4a106f144b5b890a62fc5543b35df85ab5ab06c
|
|
|
|
|
|
|
| |
Original commit msg:
Fix missing dependency in the sdklib manifest + improved error reporting.
Change-Id: I778a839cc280db8bcc7a4a0b11378119de9aed57
|
|
|
|
|
|
|
|
|
|
|
| |
The location of the debug keystore was validated (to make sure
it was not a directory or an unreadable file), but it also threw
an exception if the file didn't exist, preventing the creation
from happening.
This is integrated from master.
Change-Id: I4025bb7282c1c4ea5dd56d6c2497c9c670ee5579
|
|
|
|
| |
Change-Id: I1da6089cabe4a00567f2b7d1a7e5d2847f5799d4
|
|
|
|
|
|
|
|
|
| |
device mode is when the qualifier values must match what a device
would report. For instance nodpi should not be selectable.
Cherry-picked from master to tools_r7
Change-Id: I22cbd5b3522ae17d4d7f61e95ca20deeea825a79
|
|
|
|
|
|
|
|
|
|
| |
If the workspace has broken projects, Sdk.getProjectState() can return
null for them, causing an NPE in the IProjectChooserFilter.
I had fixed NonLibraryProjectOnlyFilter but apparently, not
LibraryProjectOnlyFilter.
Change-Id: I9a85ce61885888e0107acf2f9cb5e3491ed8f1fb
|
|
|
|
| |
Change-Id: I215862930f2336b59e6d733bff4ed9dffcca58d8
|
|
|
|
|
|
|
|
|
| |
It turns out it's a bad idea as the user will get a warning
when editing the files in the linked folder. Eclipse documentation
says that Team plug-ins should handle the linked folder and not
add them to the repository.
Change-Id: I1f439cf2007a730a4d857ff98283311b1cf36afd
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Improved library list added to the bundle by comparing
the old and new library list when something changed.
This is the only way to ensure that all indirect libraries
are added removed.
Since it means we now have a list of added and a list
of removed libraries, the link and unlink methods are
merged into a single updateLink method that does it all.
This also means that there's only one type of action
bundle, so the class/fields name have been updated to
reflect the new way it's done.
Change-Id: I808a8f1536171ebebca3a9bb5370e9dc0b6433d5
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
- AdbCommandRejectedException is thrown when adb doesn't respond
to the command with OKAY. This used to throw a normal IOException
but it can be useful to throw a different type. The message
of the exception is the error string returned by adb.
- ShellCommandUnresponsiveException is the new "timeout" exception
for output received by shell command running on devices. This
makes the distinction between timeout when talking to adb and
issue with shell command not outputting anything. Also made the
javadoc for the IDevice.executeShellCommand clearer to what the
"timeout" (renamed to maxTimeToOutputResponse) does.
Also added a better timeout to the IDevice methods to install/uninstall
apps as the default 5sec timeout was likely to be too low.
Current default value is 2min.
Change-Id: I4ecb9498926295a4e801e71b33df5d611e8120b8
|
|\ \ \
| |/ /
|/| /
| |/ |
clean-up/refresh issue in the builders"
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
issue in the builders
One of the fix is that the gen folder is no longer deleted on clean (or full build)
which should reduce errors where eclipse complain the gen folder is missing
even though it's present.
Also move ProjectState into SDK as it's really only managed
by the SDK and it allows us to move its constructor to package private.
Change-Id: I321ac23f756652e7082297d2ef560a654e97cf67
|
| |
| |
| |
| |
| |
| | |
Also handle TimeoutException in ADT test runner.
Change-Id: I585c2818f1c77ec80222bad282ae10b3ee8f5e1c
|
| |
| |
| |
| | |
Change-Id: I7e96c253aa07b124bceab99b6f5ce11e17f7b950
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The UI to edit the target and library changed the ProjectState
directly, causing havoc on the filelistener which *must* be the
entry point to reload modification into ProjectState.
This patch makes the UI work on a copy of the properties which
are then saved separately from the ProjectState (whose save()
method disappeared).
A next patch will enforce the ProjectProperties of the ProjectState
to be read only so that this does not happen again (forcing
to clone it before modifying/saving it).
This patch also fixes the reloading/relinking of the libraries upon
modification of the default.prop but more work is needed here.
Change-Id: Ie6a23111242005eb91b7533b506c029ba602f0f3
|
|\ \
| |/
|/| |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The issue is that sometimes phones will stop responding to shell
commands launched through adb.
One of these commands is executed in DeviceMonitor#queryNewDeviceForInfo
However this is done from a loop that acquires a lock that is used in
other places, including building. This basically renders eclipse
useless as Eclipse will refuse to do anything until the build is
finished.
The fix is twofold:
First, move the call to queryNewDeviceForInfo outside of the
synchronized block.
This fixes the deadlock issue but this will not fix the problem
that the device monitor will still be stuck with a shell command
that never returns.
The second part of the fix is to add a timeout to shell command
execution. Additionnaly, this patch contains a lot of clean up
of the adb API in ddmlib, especially around exceptions thrown
during timeout.
Change-Id: Ice8ef787c825e0e7b535ff0bb939bf6f25e3d7e4
|
|\ \ |
|
| | |
| | |
| | |
| | | |
Change-Id: Ibb2085fbbb59216f73da376b8283d2ccc1652c08
|
|/ /
| |
| |
| | |
Change-Id: I293b049e73e4776d969706b28ae7533ed581dfcd
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The rendering requires a looper, but never actually uses it.
Some views (like scrollView) make use of messages during rendering
putting messages into the looper message queue which was never
read or emptied.
In the case of the scrollview, the message actually contains
a reference to the scrollview, which would then leak (with all
its children views)
The fix is to delete the looper that was created. This must be
done by reflection as there's no public access to it, and this
must be done from ADT so that all versions of layoutlib get
the fix.
Change-Id: I998ad0ec17e77e36a42d77f8ab888917a9ff6441
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Added proper support for qualifiers that have fake values
like "__" for language/region or "none" for dock mode. Those
are used by ADT internally but do not represent valid values, so
some UI (like the config creator) must handle them.
They didn't do it before which led to some weird UI behavior such
as a language or region qualifier with "__" value.
Added some missing icons.
Clean-up some qualifiers by moving similar code into a base class
for enum based qualifiers
Removed dependency on the current IAndroidTarget to get the
folder segment corresponding to a qualifier as it's not needed
anymore now that targets use recent versions of aapt (and will
move to using the most recent one soon).
Change-Id: I29c7bb5a7783db1696f53d0e38f46f64671e5e0d
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This is due to link updates on closed projects.
Change-Id: I64eda43cbe574efc6e76f2a83490b5cb2585dfc4
|
|/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Those devices represent one of each res/density combo for officially
supported screens.
See http://developer.android.com/guide/practices/screens_support.html#range
Change-Id: I6b8348ae1ab96b6b1f27fa2afd655ff6f3866dd8
|
|/ /
| |
| |
| |
| |
| |
| | |
Team plug-ins should already ignore linked resources, but it seems
that some don't so marking them as derived is a safer bet.
Change-Id: Id1c82ae8a5aafabd39829034aa805cd3339f0c1a
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
SDK Bug: 2814467
The fix is a workaround around Eclipse issue
https://bugs.eclipse.org/bugs/show_bug.cgi?id=318108
Change-Id: I5684db244d5663ffa47407bd356a5bd233578612
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | | |
Change-Id: I2f2af2a0ba6087b99a7d0b9ed3acea62a48991f4
|
|\ \ \ \
| |/ / /
|/| / /
| |/ / |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This CL adds a LayoutCanvasViewer, a JFace viewer wrapping the
LayoutCanvas control. This allows the canvas to participate in the
workbench "site" selection.
To summarize:
- The workbench site selection service can be seen as "centralized"
service that registers selection providers and selection listeners.
- The editor part and the ouline are selection providers.
- The editor part, the outline and the property sheet are listener
which all listen to each others indirectly.
- Hilarity ensues.
I tried to add enough javadoc in the classes to explain what's
going on, so please tell me if more is needed. (Editor part also
has a link to the one web page article that brings some sense to
this stuff. I recommend reading the web page first.)
Change-Id: Ief83f9fe2fc1cb5c0c1fa9ae174a58c8daf17ac4
|
| | |
| | |
| | |
| | | |
Change-Id: Ie6c37e7273ef5ed6c006170f32af396f0b4ef98d
|
|\ \ \
| | | |
| | | |
| | | | |
libraries."
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This is simply enabling the ability to add libraries to a library project
in the UI.
Also updated the change log.
Change-Id: Ibaf67ce8c2af10ff890038a72de9f176bfae8592
|
|\ \ \ \
| | | | |
| | | | |
| | | | | |
.gitignore file."
|
| | |_|/
| |/| |
| | | |
| | | |
| | | |
| | | | |
file.
Change-Id: I3fc0fb9195b661b1b6732128cfd95e518ffb93b1
|
|\ \ \ \
| |/ / /
|/| / /
| |/ / |
|
| |/
| |
| |
| |
| |
| |
| |
| |
| | |
Major change in the workflow linking projects and library together.
This is now done through a single job with a queue of action to do.
This ensure that each new opened project is processed one at a time
and not in parallel which would generate problems.
Change-Id: I2a05ada293a21faba65bb639092f77ff7e5ffb2d
|
|\ \
| |/
|/|
| | |
button."
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Since there's no way to figure out automatically which project is valid
for a given running app, we read the name of the project through an
environment variable.
NOTE: The goal of this new feature is for people working on the
platform but who still want to use the DDMS plugin and some of its
features.
Change-Id: I6f124c0413dab5c9f1fb240a3b311ba7234c4378
|
|\ \
| |/
|/| |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
For some reason the locale used by the ConfigState in the
ConfigurationComposite had null members. Failed to reproduce
it. One clue might be there was a non-selected GLE2 open
as an editor tab but the Java workspace was not the main
visible one. It's possible some event fired before the
composite controls got created.
Change-Id: I2c8b6c86842f3c772c2a7ec113941eb8fc01b9e4
|
|\ \ |
|
| | |
| | |
| | |
| | | |
Change-Id: Ief8261327f7917d158fc8ad4dd4e4c3d322bbce2
|