https://codereview.chromium.org/1051343002 adds a dependency
on Chromium's third_party/junit into base/ which affects our
Android tests that uses that code.
The precompiled JUnit 4.11 JAR file that is only by the
libjingle_peerconnection_java_unittest target on Linux has been
moved to third_party/junit-jar, since it collided with the expected
path for the JUnit dependency mentioned above.
It had to be kept since the Chromium JUnit is only possible to build
when OS==android.
This CL also brings in Mockito and Robolectric, which should be
useful for our Android tests.
Other relevant changes:
* src/buildtools: 3b302fe..15308f4
* src/third_party/libjpeg_turbo: 034e9a9..9e9058b
* src/third_party/libyuv: 32ad6e0..01db3d1
Details: a12e1e1..0cb2549/DEPS
Clang version was not updated in this roll.
BUG=4499
R=phoglund@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/48239004
Cr-Commit-Position: refs/heads/master@{#9113}
When Chromium decides to move conflicting dependencies
to the _bad_scm directory it is possible to end up with
problems when that directory gets added to the local index.
Ignoring it should avoid such problems.
R=hbos@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/45969004
Cr-Commit-Position: refs/heads/master@{#8988}
We used to symlink the .gn file from Chromium but it's now
replaced by our own copy (needed for recent GN changes in
https://codereview.chromium.org/988563002).
Relevant changes:
* src/third_party/boringssl/src: e2e1326..40acdae
* src/third_party/icu: 46be516..10834e8
* src/third_party/nss: bb4e75a..d1edb68
* src/tools/gyp: d174d75..2889664
The entries for
* src/third_party/jsoncpp/source/include
* src/third_party/jsoncpp/source/src/lib_json
are removed and replaced by:
* src/third_party/jsoncpp/source @ab1e40f
(which doesn't matter for us since we symlink third_party/jsoncpp
Details: 5333e14..d8f8dc8/DEPS
Clang version was not updated in this roll.
R=henrika@webrtc.org, pbos@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/48919004
Cr-Commit-Position: refs/heads/master@{#8959}
This enables OpenSSL by default for Windows, see
8e72e1d..271c6cc/build/common.gypi
which required libjingle_tests.gyp to be updated since the
targets in third_party/nss/nss.gyp was moved into a condition in
https://codereview.chromium.org/694643002.
New Android dependencies are required due to being introduced in
build/android/pylib/remote/device/remote_device_test_run.py
of 5c49978f09
This should also fix Android test execution that started failing after
https://codereview.chromium.org/815213002 was submitted, since
it's based on e2a338fac9
Relevant other changes:
* src/buildtools: 535aff2..23a4e2f
* src/third_party/android_tools: 4f723e2..8fe116f
* src/third_party/boringssl/src: 00505ec..306e520
* src/third_party/icu: 53ecf0f..51c1a4c
* src/third_party/libvpx: 9fbec81..d3f3dce
* src/tools/swarming_client: 1d4965c..119b084
Details: 8e72e1d..271c6cc/DEPS
Clang version updated 218707:223108:
8e72e1d..271c6cc/tools/clang/scripts/update.sh
Due to this, we had to disable deadlock detection for TSan
due to a bug in Clang (see webrtc:
BUG=4106
R=pbos@webrtc.org, pthatcher@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/36459004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@8003 4adac7df-926f-26a2-2b94-8c16560cd09d
This changes sync_chromium.py to download a full Chromium
checkout instead of one with no history. It has been noticed
that the download of the no-history checkout is very slow, even
when on high-speed internet connections, due to current limitations
in the Git backend serving these clones.
Switching to a full checkout is faster, but requires more bandwidth
and disk space.
To keep the old behavior, users must set the CHROMIUM_NO_HISTORY
environment variable to 1.
Using a full checkout also enables the use of the Chromium
infrastructure teams' Git cache functionality, that speeds up
the initial download and also heavily reduces the traffic when
setting up multiple checkouts on the same machine.
This is not enabled by default, but is supported if the user is
setting the cache_dir variable in his checkout's .gclient file to
point at a directory on local disk.
BUG=3882
TESTED=
* Ran gclient sync and verified chromium/src now contained a Git
repo with full history.
* Tested rolling chromium_revision in DEPS forward + sync.
* Tested rolling it back again + sync.
* Tested with an existing no-history checkout:
CHROMIUM_NO_HISTORY=1 gclient sync
No change was performed.
* Tested with a .gclient that had cache_dir configured.
* Verified error message is displayed when .gclient has cache_dir
configured and CHROMIUM_NO_HISTORY=1.
R=iannucci@chromium.org
Review URL: https://webrtc-codereview.appspot.com/22869004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@7506 4adac7df-926f-26a2-2b94-8c16560cd09d
Add third_party/instrumented_libraries to setup_links.py
Add tools/msan/blacklist.txt which is the default location used
by MSan.
These changes are prerequisites to be able to use MSan with WebRTC.
To use it, one must also run:
sudo third_party/instrumented_libraries/install-build-deps.sh
to get the instrumented libraries installed (requires
/etc/apt/sources.list to be setup with deb-src entries).
NOTICE: Compilation is not yet working, but with this we can setup
a FYI bot to work with.
BUG=chromium:416871
TESTED=gclient sync + generate projects using:
GYP_DEFINES='msan=1 use_instrumented_libraries=1 instrumented_libraries_jobs=20' webrtc/build/gyp_webrtc
Built successfully in Release and ran a couple of tests (some crashed, some passed).
R=pbos@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/25649004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@7352 4adac7df-926f-26a2-2b94-8c16560cd09d
Libvpx now supports GN and this CL turns on compiling it.
I also introduced an executable target 'webrtc_tests'
that depends on all in WeBRTC + tests in order to get a full
linking step executed (since we've seen link problems for GN
when rolling WebRTC into Chromium).
I also converted a few test targets and made a GN file for
third_party/gflags.
BUG=3441
TESTED=Trybots + full Chromium build with a symlinked src/third_party/webrtc
dir to a workspace wit this CL applied.
R=brettw@chromium.orgTBR=niklas.enbom@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/25569004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@7344 4adac7df-926f-26a2-2b94-8c16560cd09d
Restructure how the Android APK tests are compiled now
that we have a Chromium checkout available (since r6938).
This removes the need of several hacks that were needed when
building these targets from inside a Chromium checkout.
By creating a symlink to Chromium's base we can compile the required
targets. This also removes the need of the previously precompiled
binaries we keep in /deps/tools/android at Google code.
All the user needs to do is to add the target_os = ["android"]
entry to his .gclient as described at
https://code.google.com/p/chromium/wiki/AndroidBuildInstructions
Before committing this CL, the Android APK buildbots will need
to be updated.
This also solves http://crbug.com/402594 since the apply_svn_patch.py
usage will be similar to the other standalone bots.
It also solves http://crbug.com/399297
BUG=chromium:399297, chromium:402594
TESTED=Locally compiled all APK targets by running:
GYP_DEFINES="OS=android include_tests=1 enable_tracing=1" gclient runhooks
ninja -C out/Release
checkdeps
R=henrike@webrtc.org, tommi@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/22149004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@7014 4adac7df-926f-26a2-2b94-8c16560cd09d
WebRTC standalone shares a lot of dependencies and build
tools with Chromium. To make the build work, many of the
paths of a Chromium checkout is now emulated by creating
symlinks to files and directories.
All DEPS entries that previously used Var("chromium_trunk")
to reference a Chromium checkout or From("chromium_deps"..)
to reference the Chromium DEPS file are now removed and
replaced by symlink entries in setup_links.py.
The script also handles cleanup of the legacy
Subversion-based dependencies that's needed for the
transition.
Windows: One Windows-specific important change is that
gclient sync|runhooks must now be run from a shell
with Administrator privileges in order to be able to create
symlinks. This also means that Windows XP is no longer
supported.
To transition a previously created checkout:
Run "python setup_links.py --force" to cleanup the old
SVN-based dependencies that have been synced by gclient sync.
For Buildbots, the --force flag is automatically enabled for
their syncs.
BUG=2863, chromium:339647
TEST=Manual testing on Linux, Mac and Windows.
R=andrew@webrtc.org, iannucci@chromium.org, phoglund@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/18379005
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6938 4adac7df-926f-26a2-2b94-8c16560cd09d
This adds copying of Chromium's third_party/BUILD.gn
to acommondate libyuv's BUILD.gn that imports the 'jpeg'
config from that file.
BUG=3441
TEST=trybots + local compile passing with:
gn gen out/Default --args="build_with_chromium=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false build_libyuv=false" && ninja -C out/Default
R=niklas.enbom@webrtc.org, pbos@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/14839004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6595 4adac7df-926f-26a2-2b94-8c16560cd09d
After rolling chromium_revision in r6516 it seems
TSan v2 turned on deadlock detection by default.
This caused a collection of tests to start failing.
This CL suppresses these failures awaiting further
investigation.
TBR=pbos@webrtc.org
BUG=3509
TEST=Tests passing local execution on Linux using the
reproduction steps in the bug.
Review URL: https://webrtc-codereview.appspot.com/18609004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6519 4adac7df-926f-26a2-2b94-8c16560cd09d
Rolling to this new Chromium revision required us to introduce
a sanitizer_options similar to the one in Chromium's base
(see https://code.google.com/p/chromium/codesearch#chromium/src/base/base.gyp&l=977
and https://codereview.chromium.org/238123003) in order
to get the same defaults for ASan and LSan. Without it
compilation will break since LeakSanitizer (LSan) is enabled by
default in Clang r209387 that is pulled with this roll.
I setup so that we pull in the sanitizer_options.cc and
tsan_suppressions.cc files using DEPS, so we don't have to maintain
them separately for now. We can still use our own TSan suppressions.txt
file as we do today with no changes needed.
This roll also brings in http://crrev.com/276676 so we can enable
GN build for WebRTC.
Overview of changes in Chrome DEPS:
$ svn diff http://src.chromium.org/chrome/trunk/src/DEPS -r 272489:277350
which can be compared with the output of:
$ svn cat http://webrtc.googlecode.com/svn/trunk/DEPS | grep chromium_deps | sed 's/^ *//' | sort | uniq
in a WebRTC checkout, gives the following relevant changes:
* third_party/android_tools 6fc0e1:c6e658
* third_party/libjpeg_turbo 263594:272637
* third_party/libyuv 1000:1007
* third_party/nss 271760:277057
* tools/gyp 1921:1927
* tools/swarming_client ae8085:aea506
The following also shows that Clang is upgraded from r206824 to r209387:
$ svn diff http://src.chromium.org/chrome/trunk/src/tools/clang/scripts/update.sh -r 272489:277350
BUG=3441
TEST=Trybots are not passing since after the recipe switch, SVN-based try jobs doesn't seem to support auto-detecting that a sync is needed if there's a DEPS change.
R=andrew@webrtc.org, pbos@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/12729004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6516 4adac7df-926f-26a2-2b94-8c16560cd09d
This CL makes it possible to build the 'webrtc_base'
target using GN.
The majority of our GYP stuff in webrtc/build/common.gypi has been
translated into the configs of webrtc/BUILD.gn.
The webrtc/base/base.gyp file is translated into webrtc/base/BUILD.gn.
This CL depends on https://codereview.chromium.org/322373002/ for the
jsoncpp BUILD.gn file and the ssl config.
To build inside Chromium, https://codereview.chromium.org/321313006/
needs to be landed first.
BUG=webrtc:3441
TEST=
Successful compilation of WebRTC as standalone:
gn gen out/Default --args="build_with_chromium=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false is_clang=true" && ninja -C out/Default
I also ran:
gn gen out/Default --args="build_with_chromium=false have_dbus_glib=true"
but it fails to compile: something is probably wrong with with pkg-config for that.
For Chromium, I symlinked src/third_party/webrtc to the webrtc subfolder of the
WebRTC checkout and applied the following patches:
https://codereview.chromium.org/322373002 (for jsoncpp and ssl config)
https://codereview.chromium.org/321313006 (enable building WebRTC)
Then I built successfully using:
gn gen out/Default && ninja -C out/Default webrtc_base
R=brettw@chromium.orgTBR=niklas.enbom@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/17669004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6461 4adac7df-926f-26a2-2b94-8c16560cd09d
Unfortunately needs to introduce yet another workaround
script for the Visual Studio toolchain download.
This will resolve the failures with our Dr Memory Full bot
(see https://code.google.com/p/chromium/issues/detail?id=366637#c2
for details). Long term, I'm considering a better approach
than using the added gclient solution pointing at
svn://svn-mirror.golo.chromium.org/chrome/trunk/deps/third_party/drmemory/drmemory.DEPS
i.e. add an entry that we roll separately in our DEPS file
instead. However, the Dr Memory team assured that changes
in their reporting format like this are rare.
Thanks fischman@ for the video_render.gypi fix!
Thanks kma@ for the transform_neon.S fix even if it turned out
not to be needed right now (probably will come back).
BUG=chromium:366637
TEST=git try -t compile
R=tommi@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/13369007
git-svn-id: http://webrtc.googlecode.com/svn/trunk@6010 4adac7df-926f-26a2-2b94-8c16560cd09d
Overview of changes in Chrome DEPS:
$ svn diff http://src.chromium.org/chrome/trunk/src/DEPS -r 249215:255773
which can be compared with the output of:
$ grep chromium_deps DEPS
in a WebRTC checkout, gives the following relevant changes:
* third_party/icu 246118:249466
* third_party/libyuv 978:979
* third_party/libjpeg_turbo 239595:251747
* third_party/libsrtp 214783:250757
* third_party/nss 246067:254867
* tools/clang-format 198831:202065
* tools/gyp 1846:1860
Among a variety of updated DEPS, this enables us to use
the new automatic download of Chromium's stripped down
Visual Studio 2013 toolchain on Windows.
For Windows, Visual Studio 2013 is also the default compiler
in Chrome. This CL sets the GYP_MSVS_VERSION to 2010 unless
otherwise specified. Doing that we can first fix our 2013 problems
before we move over to having 2013 by default.
The plan is to build 2013 at the WebRTC FYI waterfall at
http://build.chromium.org/p/client.webrtc.fyi/waterfall
to ensure we can support VS2013 before the switch.
I realized we can sync Chromium's find_depot_tools.py script
into it's own folder and just alter the PYTHONPATH for the
gyp_webrtc script. That way there's no need to have the dummy
module in webrtc/build anymore. The real script is also needed
for the logic that handles checking VS2013 and downloading it if
not found.
BUG=chromium:340973
TEST=All trybots passing runhooks and compile step.
R=tommi@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/9299004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5667 4adac7df-926f-26a2-2b94-8c16560cd09d
With help from hinoka@, we're now using a more efficient approach
to download only the files that have changed from Google Storge.
When uploading new resource files, use
upload_to_google_storage.py --bucket chromium-webrtc-resources ./filename
which of course requires gsutil authentication setup.
NOTICE: Before deploying this, svn:ignore should be removed for
the resources folder, or the bots will run into problems with a
non-versioned file being found in the checkout during sync (as
this CL adds resources to version control).
All developers will also need to be informed to wipe their local
resources dir to avoid getting an error during checkout due to the
already existing non-versioned resources directory.
BUG=2294
TEST=locally running gclient runhooks
R=andrew@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/2095004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@5076 4adac7df-926f-26a2-2b94-8c16560cd09d
This is a re-land attempt of http://review.webrtc.org/1673004/
It now includes a build/isolate.gypi in WebRTC that includes the same
file as the one that would be included when WebRTC is used in a Chromium
checkout. It is needed since it is not possible to use variables in GYP's
includes sections.
Implemented according to the instructions at
http://www.chromium.org/developers/testing/isolated-testing
Workflow has been like this:
1. create _run GYP target
2. create a stripped down .isolate file
3. export GYP_DEFINES="$GYP_DEFINES test_isolation_mode=check"
4. runhooks
5. compile
6. test if the test would run (i.e. find it's dependencies) without
actually executing it:
tools/swarm_client/isolate.py run --isolated out/Release/testname.isolated
7. If failing, run the fix_test_cases.py script like this:
tools/swarm_client/googletest/fix_test_cases.py --isolated out/Release/testname.isolated
All tests that run on the bots for WebRTC has got _run target
and .isolate file created.
"Normal tests" that run fine on any machine:
* audio_decoder_unittests
* common_audio_unittests
* common_video_unittests
* metrics_unittests
* modules_tests
* modules_unittests
* neteq_unittests
* system_wrappers_unittests
* test_support_unittests
* tools_unittests
* video_engine_core_unittests
* voice_engine_unittests
Tests that requires bare-metal and audio/video devices:
* audio_device_tests
* video_capture_tests
I also added the isolate boilerplate code for the following
tests that are not yet pure gtest binaries (which means they
cannot run isolated yet):
* video_render_tests
* vie_auto_test
* voe_auto_test
TEST=running isolate.py as described above. WebRTC trybots passing. Created a Chromium checkout with third_party/webrtc ToT and this patch applied, passing the runhooks step.
BUG=1916
R=henrike@webrtc.org, tommi@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/2056004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@4590 4adac7df-926f-26a2-2b94-8c16560cd09d
As this breaks the FYI bots in
http://build.chromium.org/p/chromium.webrtc.fyi/waterfall
due to different path to isolate.gypi (which cannot easily
be resolved due to limitations in GYP)
> Isolate GYP target and .isolate files for tests
>
> Implemented according to the instructions at
> http://www.chromium.org/developers/testing/isolated-testing
>
> Workflow has been like this:
> 1. create _run GYP target
> 2. create a stripped down .isolate file
> 3. export GYP_DEFINES="$GYP_DEFINES test_isolation_mode=check"
> 4. runhooks
> 5. compile
> 6. test if the test would run (i.e. find it's dependencies) without
> actually executing it:
> tools/swarm_client/isolate.py run --isolated out/Release/testname.isolated
> 7. If failing, run the fix_test_cases.py script like this:
> tools/swarm_client/fix_test_cases.py --isolated out/Release/testname.isolated
>
> All tests that run on the bots for WebRTC has got _run target
> and .isolate file created.
>
> "Normal tests" that run fine on any machine:
> * audio_decoder_unittests
> * common_audio_unittests
> * common_video_unittests
> * metrics_unittests
> * modules_integrationtests
> * modules_unittests
> * neteq_unittests
> * system_wrappers_unittests
> * test_support_unittests
> * tools_unittests
> * video_engine_core_unittests
> * voice_engine_unittests
>
> Tests that requires bare-metal and audio/video devices:
> * audio_device_integrationtests
> * video_capture_integrationtests
>
> I also added the isolate boilerplate code for the following
> tests that are not yet pure gtest binaries (which means they
> cannot run isolated yet):
> * video_render_integrationtests
> * vie_auto_test
> * voe_auto_test
>
> TEST=running isolate.py as described above.
> BUG=1916
> R=tommi@webrtc.org
>
> Review URL: https://webrtc-codereview.appspot.com/1673004TBR=kjellander@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/2040004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@4548 4adac7df-926f-26a2-2b94-8c16560cd09d
Implemented according to the instructions at
http://www.chromium.org/developers/testing/isolated-testing
Workflow has been like this:
1. create _run GYP target
2. create a stripped down .isolate file
3. export GYP_DEFINES="$GYP_DEFINES test_isolation_mode=check"
4. runhooks
5. compile
6. test if the test would run (i.e. find it's dependencies) without
actually executing it:
tools/swarm_client/isolate.py run --isolated out/Release/testname.isolated
7. If failing, run the fix_test_cases.py script like this:
tools/swarm_client/fix_test_cases.py --isolated out/Release/testname.isolated
All tests that run on the bots for WebRTC has got _run target
and .isolate file created.
"Normal tests" that run fine on any machine:
* audio_decoder_unittests
* common_audio_unittests
* common_video_unittests
* metrics_unittests
* modules_integrationtests
* modules_unittests
* neteq_unittests
* system_wrappers_unittests
* test_support_unittests
* tools_unittests
* video_engine_core_unittests
* voice_engine_unittests
Tests that requires bare-metal and audio/video devices:
* audio_device_integrationtests
* video_capture_integrationtests
I also added the isolate boilerplate code for the following
tests that are not yet pure gtest binaries (which means they
cannot run isolated yet):
* video_render_integrationtests
* vie_auto_test
* voe_auto_test
TEST=running isolate.py as described above.
BUG=1916
R=tommi@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/1673004
git-svn-id: http://webrtc.googlecode.com/svn/trunk@4547 4adac7df-926f-26a2-2b94-8c16560cd09d
Things included in this CL:
- updated READMEs to provide an exact/reproable set of steps for getting the app
running.
- gyp changes to build the iOS AppRTCDemo sample app using gyp+ninja instead of
the hand-crafted Xcode project (which has never worked in its checked-in
form), including a gyp action to sign the sample app for deployment to an iOS
device (the app can also be used in the simulator)
- deleted the busted hand-crafted Xcode project for the sample app
- updated the sample app to match the PeerConnection API that ended up landing
(in a surprising twist of fate, the API landed quite a bit later than the
sample app and this is the first time the CR-time changes in the API are
reflected in the sample app)
- updated the sample app to reflect apprtc.appspot.com HTML/JS changes (equiv to
the AppRTCClient.java changes in http://s10/47299162)
- picked up the iossim DEPS to enable launching the sample app in the simulator
from the command-line.
- renamed some files to match capitalization of the classes they contain (Ice ->
ICE) per ObjC naming guidelines.
- ran the files involved in this CL through clang-format to deal with xcode
formatting craxy.
BUG=2106
RISK=P2
TESTED=unittest builds with ninja and passes on OS=mac; sample app builds with ninja and runs on simulator and device, though no audio flows from simulator/device (will fix in a follow-up CL)
R=andrew@webrtc.org, justincohen@google.com, wu@webrtc.org
Review URL: https://webrtc-codereview.appspot.com/1874005
git-svn-id: http://webrtc.googlecode.com/svn/trunk@4466 4adac7df-926f-26a2-2b94-8c16560cd09d
