external/webrtc.git
2 hours agoRemove deprecated RTCVideoRenderer constructor. master
tkchin@webrtc.org [Tue, 2 Sep 2014 20:50:00 +0000 (20:50 +0000)]
Remove deprecated RTCVideoRenderer constructor.

Removes -[RTCVideoRenderer initWithView]. Also, fix potential issue where we hold on to a video frame longer than the lifetime of its associated track.

BUG=3341
R=glaznev@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/16099004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7032 4adac7df-926f-26a2-2b94-8c16560cd09d

4 hours agoRemove the checks.h dependence on logging.h in a standalone build.
andrew@webrtc.org [Tue, 2 Sep 2014 19:00:45 +0000 (19:00 +0000)]
Remove the checks.h dependence on logging.h in a standalone build.

logging.h apparently drags in a lot of undesirable dependencies. It was
only required for the trivial LogMessageVoidify; simply add an
identical FatalMessageVoidify instead.

Keep the include in a Chromium build to still have the override
mechanism use Chromium's macros.

Bonus: Add the missing DCHECK_GT (noticed by bercic).

R=kwiberg@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/17259004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7031 4adac7df-926f-26a2-2b94-8c16560cd09d

4 hours agoFix race in Voice Engine's Channel where it accesses RemoteNtpTimeEstimator from...
stefan@webrtc.org [Tue, 2 Sep 2014 18:58:24 +0000 (18:58 +0000)]
Fix race in Voice Engine's Channel where it accesses RemoteNtpTimeEstimator from both the audio playback thread and the network thread without locking.

BUG=3681
R=pbos@webrtc.org, xians@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/24379004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7030 4adac7df-926f-26a2-2b94-8c16560cd09d

7 hours agoRemove WebRtcVideoEngine::default_codec_format().
pbos@webrtc.org [Tue, 2 Sep 2014 16:33:09 +0000 (16:33 +0000)]
Remove WebRtcVideoEngine::default_codec_format().

R=pthatcher@webrtc.org
BUG=1788

Review URL: https://webrtc-codereview.appspot.com/24399004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7029 4adac7df-926f-26a2-2b94-8c16560cd09d

7 hours agoRemove files from talk/PRESUBMIT.py.
pbos@webrtc.org [Tue, 2 Sep 2014 16:17:36 +0000 (16:17 +0000)]
Remove files from talk/PRESUBMIT.py.

BUG=
R=pthatcher@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/23429005

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7028 4adac7df-926f-26a2-2b94-8c16560cd09d

7 hours agoCreate a copy of talk/xmllite under webrtc/xmllite.
henrike@webrtc.org [Tue, 2 Sep 2014 15:41:12 +0000 (15:41 +0000)]
Create a copy of talk/xmllite under webrtc/xmllite.

BUG=3379
R=andrew@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22249004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7027 4adac7df-926f-26a2-2b94-8c16560cd09d

8 hours agoDisable video_engine_tests and webrtc_perf_tests on Android.
kjellander@webrtc.org [Tue, 2 Sep 2014 15:13:55 +0000 (15:13 +0000)]
Disable video_engine_tests and webrtc_perf_tests on Android.

BUG=3770
TESTED=Running the tests locally on an Android device.
R=phoglund@webrtc.org
TBR=henrik.lundin@webrtc.org, pbos@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14299004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7026 4adac7df-926f-26a2-2b94-8c16560cd09d

10 hours agoDivide-by-zero problem in NetEq's Normal::Process fixed
henrik.lundin@webrtc.org [Tue, 2 Sep 2014 13:22:11 +0000 (13:22 +0000)]
Divide-by-zero problem in NetEq's Normal::Process fixed

Adding a couple of tests that tries to trigger a certain divide-by-zero
issue. The tests triggered the issue, but this CL also includes a fix
for this.

BUG=3761
R=tina.legrand@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22269004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7025 4adac7df-926f-26a2-2b94-8c16560cd09d

10 hours agoRemove retired android_apk[_rel] trybots from PRESUBMIT.py
kjellander@webrtc.org [Tue, 2 Sep 2014 13:05:58 +0000 (13:05 +0000)]
Remove retired android_apk[_rel] trybots from PRESUBMIT.py

BUG=webrtc:3741
TBR=niklas.enbom@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14289004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7024 4adac7df-926f-26a2-2b94-8c16560cd09d

11 hours agoDisable video_capture_tests for Android.
kjellander@webrtc.org [Tue, 2 Sep 2014 12:37:50 +0000 (12:37 +0000)]
Disable video_capture_tests for Android.

BUG=3768
TESTED=Passing the steps in webrtc:3768
TBR=glaznev@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/18309004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7023 4adac7df-926f-26a2-2b94-8c16560cd09d

12 hours agoGN: Update webrtc/base to recent GYP changes.
kjellander@webrtc.org [Tue, 2 Sep 2014 11:22:06 +0000 (11:22 +0000)]
GN: Update webrtc/base to recent GYP changes.

Update the webrtc/base/BUILD.gn file to reflect
webrtc/base/base.gyp changes between r6438 and r7011.

BUG=3441
TESTED= Trybots + compilation with a standalone WebRTC checkout:
gn gen out/Default --args="build_with_chromium=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false is_debug=true" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false os=\"android\" cpu_arch=\"arm\" is_clang=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false os=\"android\" cpu_arch=\"arm\" arm_version=7 is_clang=false" && ninja -C out/Default

Compilation of Chromium's 'all' target with src/third_party/webrtc
symlinked to the WebRTC checkout with this CL applied, both
with the default GN settings and using
--args="is_debug=false os=\"android\" cpu_arch=\"arm\""

R=brettw@chromium.org

Review URL: https://webrtc-codereview.appspot.com/13359004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7022 4adac7df-926f-26a2-2b94-8c16560cd09d

12 hours agoRTCBot is a framework that allows to write tests where logic runs on a single
andresp@webrtc.org [Tue, 2 Sep 2014 10:52:54 +0000 (10:52 +0000)]
RTCBot is a framework that allows to write tests where logic runs on a single
host that controls multiple endpoints ("bots"). Thus allowing to create more
complex scenarios that would otherwise require non-trival signalling between
multiple parties.

R=houssainy@google.com, phoglund@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22239004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7021 4adac7df-926f-26a2-2b94-8c16560cd09d

14 hours agoUpdate checkedeps.py rules in DEPS.
kjellander@webrtc.org [Tue, 2 Sep 2014 09:39:35 +0000 (09:39 +0000)]
Update checkedeps.py rules in DEPS.

Add allow-rules as well in addition to the
disallow-rule in r7014.

BUG=
TBR=tommi@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14279005

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7020 4adac7df-926f-26a2-2b94-8c16560cd09d

15 hours agoRemove build_with_chromium==1 conditions for Android
kjellander@webrtc.org [Tue, 2 Sep 2014 08:40:39 +0000 (08:40 +0000)]
Remove build_with_chromium==1 conditions for Android

Most of these changes were done in r7014, but a few targets
were missed. This should make these tests run better
(but they might still be failing due to webrtc:3764).

BUG=webrtc:3741
TESTED=Local compilation using:
GYP_DEFINES="OS=android component=static_library fastbuild=1 target_arch=arm" webrtc/build/gyp_webrtc
ninja -C out/Debug

R=phoglund@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/24369004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7019 4adac7df-926f-26a2-2b94-8c16560cd09d

15 hours agoUnpacking aecdumps generates wav files
aluebs@webrtc.org [Tue, 2 Sep 2014 07:51:51 +0000 (07:51 +0000)]
Unpacking aecdumps generates wav files

BUG=webrtc:3359
R=andrew@webrtc.org, bjornv@webrtc.org, kwiberg@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14239004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7018 4adac7df-926f-26a2-2b94-8c16560cd09d

32 hours agoFix audio_decoder_unittests.isolate
kjellander@webrtc.org [Mon, 1 Sep 2014 15:06:14 +0000 (15:06 +0000)]
Fix audio_decoder_unittests.isolate

In r6427 all .isolate files except
audio_decooder_unittests.isolate was updated to use the
<(DEPTH) variable instead of relative paths.
This started breaking the Android bots after committing
r7014.

BUG=3741
TBR=phoglund@webrtc.org,

Review URL: https://webrtc-codereview.appspot.com/23409004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7017 4adac7df-926f-26a2-2b94-8c16560cd09d

33 hours agoAdding more codecs to the AcmSenderBitExactness
henrik.lundin@webrtc.org [Mon, 1 Sep 2014 14:19:00 +0000 (14:19 +0000)]
Adding more codecs to the AcmSenderBitExactness

New tests include iSAC-swb, PCM16b (8, 16, 32 kHz; mono and stereo),
PCM A/u (mono and stereo), iLBC, G.722 (mono and stereo), and Opus.

Also adding checks on number of output channels.

BUG=3521
R=kwiberg@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/15319004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7016 4adac7df-926f-26a2-2b94-8c16560cd09d

35 hours agoRoll chromium_revision 681cc8e..f0a439d (r292217:r292861)
kjellander@webrtc.org [Mon, 1 Sep 2014 11:41:56 +0000 (11:41 +0000)]
Roll chromium_revision 681cc8e..f0a439d (r292217:r292861)

Mainly to pick up https://codereview.chromium.org/500423004/
that enables us to build the Android APK tests from
a standalone checkout.

Other changes:
* tools/swarming_client to e7d8b988423ff1966d64db3ef7ca766296f9b0c1
* third_party/boringssl to 6c7aed048ca0a335e02dfee10976c5dc8620783e
* third_party/icu 527ea2dd86afa2751a85d1cc4695f9e2e2d18022 (r291706)
* third_party/libjpeg_turbo to 2ed5319 (r291725)
* third_party/libvpx 563c46b:982d147 (r291661:r291730)
* third_party/nss to 90c5f9a8b8980fe60165813f578bbeb4fe20b18d

Trybot failures at Android trybots are expected, since
they're currently in a bad state since they in the middle
of being reconfigured, partially pending this CL.

BUG=webrtc:3741
R=tommi@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14269004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7015 4adac7df-926f-26a2-2b94-8c16560cd09d

36 hours agoAndroid APK tests built from a normal WebRTC checkout.
kjellander@webrtc.org [Mon, 1 Sep 2014 11:06:37 +0000 (11:06 +0000)]
Android APK tests built from a normal WebRTC checkout.

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

43 hours agoGN: Audio device module
kjellander@webrtc.org [Mon, 1 Sep 2014 04:24:11 +0000 (04:24 +0000)]
GN: Audio device module

The GN files are based upon the GYP files as of r7009.

BUG=3441
TESTED=Passing builds with:
gn gen out/Default --args="build_with_chromium=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false is_debug=true" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false os=\"android\" cpu_arch=\"arm\" is_clang=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false os=\"android\" cpu_arch=\"arm\" arm_version=7 is_clang=false" && ninja -C out/Default

Compilation of Chromium's 'all' target with src/third_party/webrtc
symlinked to the WebRTC checkout with this CL applied, both
with the default GN settings and using
--args="is_debug=false os=\"android\" cpu_arch=\"arm\""

R=brettw@chromium.org

Review URL: https://webrtc-codereview.appspot.com/14259004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7013 4adac7df-926f-26a2-2b94-8c16560cd09d

2 days agoGN: Implement voice engine, common audio, audio coding and audio processing
kjellander@webrtc.org [Sun, 31 Aug 2014 20:32:53 +0000 (20:32 +0000)]
GN: Implement voice engine, common audio, audio coding and audio processing

NOTICE: Assembly offsets generation for audio processing will
not be ported to GN and the process of removing them is tracked
in https://code.google.com/p/webrtc/issues/detail?id=3580.

The GN files are based upon the GYP files as of r7009.

BUG=3441
TESTED=Passing builds with:
gn gen out/Default --args="build_with_chromium=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false is_debug=true" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false aec_debug_dump=true" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false aec_untrusted_delay_for_testing=true" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false prefer_fixed_point=true" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false os=\"android\" cpu_arch=\"arm\" is_clang=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false os=\"android\" cpu_arch=\"arm\" arm_version=7 is_clang=false" && ninja -C out/Default

I don't know how to setup the mips toolchain to test the following, but it's out of scope for the GN effort for now:
gn gen out/Default --args="build_with_chromium=false cpu_arch=\"mipsel\" mips_dsp_rev=0" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false cpu_arch=\"mipsel\" mips_dsp_rev=1" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false cpu_arch=\"mipsel\" mips_dsp_rev=2" && ninja -C out/Default

Compilation of Chromium's 'all' target with src/third_party/webrtc
symlinked to the WebRTC checkout with this CL applied, both
with the default GN settings and using
--args="is_debug=false os=\"android\" cpu_arch=\"arm\""

R=andrew@webrtc.org, brettw@chromium.org

Review URL: https://webrtc-codereview.appspot.com/15999004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7012 4adac7df-926f-26a2-2b94-8c16560cd09d

4 days agoGN: Fix webrtc/video/BUILD.gn for Chromium build.
kjellander@webrtc.org [Fri, 29 Aug 2014 21:39:35 +0000 (21:39 +0000)]
GN: Fix webrtc/video/BUILD.gn for Chromium build.

A mistake was made in https://review.webrtc.org/18709004/
so it doesn't build in Chromium. Adding a config to get
the root folder included in the include path solves it.

BUG=3441
TESTED=Local compilation of Chromium's all target with
src/third_party/webrtc linked to the WebRTC checkout with
this CL applied.
TBR=brettw@chromium.org

Review URL: https://webrtc-codereview.appspot.com/19169004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7011 4adac7df-926f-26a2-2b94-8c16560cd09d

4 days agoMIPS optimizations for AEC audio processing module
andrew@webrtc.org [Fri, 29 Aug 2014 17:51:28 +0000 (17:51 +0000)]
MIPS optimizations for AEC audio processing module

Added new optimizations for MIPS that were removed in r6797.
For more information about this see https://code.google.com/p/webrtc/source/detail?r=6797

R=andrew@webrtc.org, djordje.pesut@imgtec.com

Review URL: https://webrtc-codereview.appspot.com/15259004

Patch from Ljubomir Papuga <ljubomir.papuga@mips.com>.

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7010 4adac7df-926f-26a2-2b94-8c16560cd09d

4 days agoAdd LTO support for Android Chromium.
andrew@webrtc.org [Fri, 29 Aug 2014 17:41:13 +0000 (17:41 +0000)]
Add LTO support for Android Chromium.

This is to add support for a Link-Time Optimizations experiment in Android Chromium. As it is disabled by default, it won't change anything for most configurations.
BUG=chromium:407544
R=andrew@webrtc.org, tina.legrand@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21299004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7009 4adac7df-926f-26a2-2b94-8c16560cd09d

4 days agoAllow same src and dst in InputAudioFile::DuplicateInterleaved
henrik.lundin@webrtc.org [Fri, 29 Aug 2014 07:26:40 +0000 (07:26 +0000)]
Allow same src and dst in InputAudioFile::DuplicateInterleaved

This change allows the input and output to the static method
InputAudioFile::DuplicateInterleaved to be the same array. That is,
in-place manipulation is now possible. A unit test is also added.

R=tina.legrand@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/20239004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7008 4adac7df-926f-26a2-2b94-8c16560cd09d

4 days agowin: Replace custom assert() macro with regular assert.h
thakis@chromium.org [Fri, 29 Aug 2014 03:00:15 +0000 (03:00 +0000)]
win: Replace custom assert() macro with regular assert.h

The current code got added in libjingle r103; I don't see a good reason for it.
Things still build with plain old assert.h.

The custom assert was wrong: __debugbreak() is documented to return void,
so doing `cond ? true : __debugbreak()` shouldn't build (and it doesn't in
clang-cl). It's possible to make it build by writing
`cond ? true : (__debugbreak(), true)`, but just using the regular header
seems like a much better fix.

BUG=chromium:82385
Review URL: https://webrtc-codereview.appspot.com/19139004/

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7007 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoAdd jiayl to talk OWNERS.
jiayl@webrtc.org [Thu, 28 Aug 2014 23:24:36 +0000 (23:24 +0000)]
Add jiayl to talk OWNERS.

BUG=
R=juberti@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/23379004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7006 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoWhen the peerconnection creates the offer with a constraint to disable the audio...
jiayl@webrtc.org [Thu, 28 Aug 2014 22:21:34 +0000 (22:21 +0000)]
When the peerconnection creates the offer with a constraint to disable the audio offering, stats will not get properly updated.

  constraints . SetMandatoryReceiveAudio (false);

The problem is that webrtc::GetTrackIdBySsrc returns false if audio is not available. However it should continue and check for the video track.

BUG=webrtc:3755
R=jiayl@webrtc.org, juberti@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22479004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7005 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoPrecompile out our standalone CHECK macros in a Chromium build.
andrew@webrtc.org [Thu, 28 Aug 2014 19:00:15 +0000 (19:00 +0000)]
Precompile out our standalone CHECK macros in a Chromium build.

As documented, the use of overrides/webrtc/base/logging.h in a Chromium
build reuslts in redefined macro errors. Fortunately, Chromium's macros
can be used as drop-in replacements for the standalone versions.

TBR=henrike

Review URL: https://webrtc-codereview.appspot.com/17239004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7004 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoAdd CHECK and friends from Chromium.
andrew@webrtc.org [Thu, 28 Aug 2014 16:28:26 +0000 (16:28 +0000)]
Add CHECK and friends from Chromium.

Replace FATAL_ERROR_IF with the more familiar (to Chromium developers)
CHECK and DCHECK. The full Chromium implementation is fairly elaborate
but I copied enough to get us most of the benefits. I believe the main
missing component is a more advanced stack dump. For this bit I relied
on the V8 implementation.

There are a few minor modifications from the Chromium original:
- The FatalMessage class is specialized for logging fatal error
messages and aborting. Chromium uses the general LogMessage class,
which we could consider moving towards in the future.
- NOTIMPLEMENTED() and NOTREACHED() have been removed, partly because
I don't want to rely on our logging.h until base/ and system_wrappers/
are consolidated.
- FATAL() replaces LOG(FATAL).

Minor modifications from V8's stack dump:
- If parsing of a stack trace symbol fails, just print the unparsed
symbol. (I noticed this happened on Mac.)
- Use __GLIBCXX__ and __UCLIBC__. This is from examining the backtrace
use in Chromium.

UNREACHABLE() has been removed because its behavior is different than
Chromium's NOTREACHED(), which is bound to cause confusion. The few uses
were replaced with FATAL(), matching the previous behavior.

Add a NO_RETURN macro, allowing us to remove unreachable return
statements following a CHECK/FATAL.

TESTED=the addition of dummy CHECK, DCHECK, CHECK_EQ and FATAL did the
did the right things. Stack traces work on Mac, but I don't get symbols
on Linux.

R=henrik.lundin@webrtc.org, kwiberg@webrtc.org, tommi@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22449004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7003 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoSpecify an ECDH group for ECDHE.
jiayl@webrtc.org [Thu, 28 Aug 2014 16:14:38 +0000 (16:14 +0000)]
Specify an ECDH group for ECDHE.

By default, OpenSSL cannot negotiate ECDHE cipher suites as a server because it
doesn't know what curve to use.

BUG=chromium:406458
TEST=Download Firefox nightly build from 2014-08-12.
  https://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2014-08-12-mozilla-central-debug/
  Point Firefox to https://apprtc.appspot.com
  Point Chrome on Android to the URL Firefox redirects to (it'll say ?r=NUMBERS at the end)
  After tapping through various permissions prompts on either side, the call goes through.

R=agl@chromium.org, henrike@webrtc.org, jiayl@webrtc.org, juberti@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/18269004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7002 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoAdd talk owners to migrated talk folders
henrike@webrtc.org [Thu, 28 Aug 2014 16:03:58 +0000 (16:03 +0000)]
Add talk owners to migrated talk folders

BUG=N/A
R=niklas.enbom@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/15309004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7001 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoAdd 60 fps video support
niklas.enbom@webrtc.org [Thu, 28 Aug 2014 14:57:46 +0000 (14:57 +0000)]
Add 60 fps video support

R=henrike@webrtc.org, mflodman@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22209004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@7000 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoGN: Implement video_engine, video_capture and video_render.
kjellander@webrtc.org [Thu, 28 Aug 2014 13:51:08 +0000 (13:51 +0000)]
GN: Implement video_engine, video_capture and video_render.

Also add more from common.gypi to webrtc.gni.

These GN configs are based on GYP files in r6997.

BUG=3441
TEST=Trybots and local compile using:
gn gen out/Default --args="build_with_chromium=false" && ninja -C out/Default
gn gen out/Default --args="build_with_chromium=false is_debug=true" && ninja -C out/Default

Passed compile from a Chromium checkout with src/third_party/webrtc linked to the webrtc/ dir of a checkout with this patch applied.

R=brettw@chromium.org, glaznev@webrtc.org, mflodman@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/18709004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6999 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agocommon_audio: Removed macro WEBRTC_SPL_DIV
bjornv@webrtc.org [Thu, 28 Aug 2014 12:57:32 +0000 (12:57 +0000)]
common_audio: Removed macro WEBRTC_SPL_DIV

The macro has no built-in divide by zero check. The only thing that is done is casting to int32_t.
In addition a bug was discovered where it was supposed to do a division with rounding, but instead did a division with truncation + addition by 2. This is corrected in this CL.

BUG=3348,3353
TESTED=locally on Linux
R=kwiberg@webrtc.org, tina.legrand@webrtc.org, turaj@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/19129004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6998 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days ago(Auto)update libjingle 74235596-> 74297316
buildbot@webrtc.org [Thu, 28 Aug 2014 10:52:44 +0000 (10:52 +0000)]
(Auto)update libjingle 74235596-> 74297316

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6997 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoFix the different samples per channel in aecdump
aluebs@webrtc.org [Thu, 28 Aug 2014 10:43:09 +0000 (10:43 +0000)]
Fix the different samples per channel in aecdump

BUG=webrtc:3359
R=bjornv@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/15299004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6996 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoDisable VideoAdapterTest.BlackOutput on DrMemory.
pbos@webrtc.org [Thu, 28 Aug 2014 09:55:34 +0000 (09:55 +0000)]
Disable VideoAdapterTest.BlackOutput on DrMemory.

Reports uninitialized-memory reads that seem to originate from when the
frame is copied. The test passes if we remove CPU optimizations from
libyuv, disabling test until we figure out whether it's an unsupported
instruction in DrMemory, bug in libyuv or bug in the test.

BUG=3754
R=aluebs@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/19149004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6995 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoAdd unit tests to rtcp_receiver_test.
asapersson@webrtc.org [Thu, 28 Aug 2014 07:35:06 +0000 (07:35 +0000)]
Add unit tests to rtcp_receiver_test.

R=mflodman@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21989004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6994 4adac7df-926f-26a2-2b94-8c16560cd09d

5 days agoRoll chromium_revision b1748b:681cc8
marpan@webrtc.org [Thu, 28 Aug 2014 02:32:45 +0000 (02:32 +0000)]
Roll chromium_revision b1748b:681cc8

Pick the libvpx roll: https://codereview.chromium.org/513593002

BUG=3747
R=andrew@webrtc.org
TBR=ajm@google.com

Review URL: https://webrtc-codereview.appspot.com/14229004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6993 4adac7df-926f-26a2-2b94-8c16560cd09d

6 days agoRe-enable all VideoAdapterTests on DrMemory.
pbos@webrtc.org [Wed, 27 Aug 2014 18:41:58 +0000 (18:41 +0000)]
Re-enable all VideoAdapterTests on DrMemory.

These bugs should've been resolved as of r6991.

BUG=3655,3671
R=henrike@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22529004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6992 4adac7df-926f-26a2-2b94-8c16560cd09d

6 days agoFix data races during VideoAdapterTest tear-down.
pbos@webrtc.org [Wed, 27 Aug 2014 18:16:13 +0000 (18:16 +0000)]
Fix data races during VideoAdapterTest tear-down.

Explicitly disconnect the VideoCapturer to avoid frames being
delivered during listener destruction. This manifested only on DrMemory
Full on Windows which I was able to repro locally.

BUG=3671
R=henrike@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/15289004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6991 4adac7df-926f-26a2-2b94-8c16560cd09d

6 days ago(Auto)update libjingle 74202294-> 74230205
buildbot@webrtc.org [Wed, 27 Aug 2014 17:21:19 +0000 (17:21 +0000)]
(Auto)update libjingle 74202294-> 74230205

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6990 4adac7df-926f-26a2-2b94-8c16560cd09d

6 days agoMove end of namespace inside #ifdef
henrik.lundin@webrtc.org [Wed, 27 Aug 2014 10:17:22 +0000 (10:17 +0000)]
Move end of namespace inside #ifdef

The code did not compile unless WEBRTC_ANDROID was defined.

R=henrike@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21319004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6989 4adac7df-926f-26a2-2b94-8c16560cd09d

6 days agoExpose setPayloadType on the rtp_sender. Thus allowing other users of this module
andresp@webrtc.org [Wed, 27 Aug 2014 09:39:43 +0000 (09:39 +0000)]
Expose setPayloadType on the rtp_sender. Thus allowing other users of this module
to set the payload type to be used without having to call SendOutgoingData.

BUG=3694
R=asapersson@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/18289004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6988 4adac7df-926f-26a2-2b94-8c16560cd09d

6 days ago- Make local constant non-static.
solenberg@webrtc.org [Wed, 27 Aug 2014 08:52:17 +0000 (08:52 +0000)]
- Make local constant non-static.
- Remove spammy log line.

BUG=
R=henrike@webrtc.org, kjellander@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21339004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6987 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoCreate a copy of talk/sound under webrtc/sound.
henrike@webrtc.org [Tue, 26 Aug 2014 22:04:04 +0000 (22:04 +0000)]
Create a copy of talk/sound under webrtc/sound.

BUG=3379
R=andrew@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22379004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6986 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoimplement handling ALTERNATE-SERVER response from turn protocol as
guoweis@webrtc.org [Tue, 26 Aug 2014 21:37:49 +0000 (21:37 +0000)]
implement handling ALTERNATE-SERVER response from turn protocol as
specified in RFC 5766, also created 2 test cases for both the normal
redirection case as well as when a pingpong situation happens, the
allocation should fail

BUG=1986 TURN ALTERNATE-SERVER support
R=juberti@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21249004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6985 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoAvoid syncing unnecessary Chromium deps for WebRTC.
kjellander@webrtc.org [Tue, 26 Aug 2014 19:22:03 +0000 (19:22 +0000)]
Avoid syncing unnecessary Chromium deps for WebRTC.

This should save several gigabytes of traffic and disk space.

On Linux this is about 2.6 GB:
346M src/chrome/tools/test/reference_build
340M src/native_client
170M src/third_party/ffmpeg
1.5G src/third_party/WebKit
196M src/v8

BUG=2863
TESTED=Removed the directories locally, ran a sync and verified they didn't reappear (or fail because of platform-specific ones).
R=iannucci@chromium.org, niklas.enbom@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22189004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6984 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days ago(Auto)update libjingle 74132319-> 74133664
buildbot@webrtc.org [Tue, 26 Aug 2014 15:50:23 +0000 (15:50 +0000)]
(Auto)update libjingle 74132319-> 74133664

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6983 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days ago(Auto)update libjingle 74128148-> 74132319
buildbot@webrtc.org [Tue, 26 Aug 2014 15:24:54 +0000 (15:24 +0000)]
(Auto)update libjingle 74128148-> 74132319

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6982 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoDisable EndToEndTest.RestartingSendStreamPreservesRtpState in video_engine_tests...
aluebs@webrtc.org [Tue, 26 Aug 2014 14:22:51 +0000 (14:22 +0000)]
Disable EndToEndTest.RestartingSendStreamPreservesRtpState in video_engine_tests because it is flaky

BUG=webrtc:3745
R=pbos@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21349004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6981 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoFix Win64 compile of videoadapter_unittest.cc.
pbos@webrtc.org [Tue, 26 Aug 2014 12:46:57 +0000 (12:46 +0000)]
Fix Win64 compile of videoadapter_unittest.cc.

Missed an typecast in videoadapter_unittest.cc in r6979 due to
tryservers being clogged and me waiting for a windows, linux, mac and
tsanv2 bot to finish was not enough. Committing fix straight away to
un-break tree.

TBR=tommi@webrtc.org
BUG=3671

Review URL: https://webrtc-codereview.appspot.com/18279004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6980 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoFix data races in VideoAdapterTest.
pbos@webrtc.org [Tue, 26 Aug 2014 12:33:18 +0000 (12:33 +0000)]
Fix data races in VideoAdapterTest.

Adressing clear races between the test thread and capturer thread shown
as heap-use-after-free in vpx_codec_destroy in
WebRtcVideoMediaChannelTest.SetSend (way later in the rest run).

When capturing a frame the test copied it to a separate frame that would
then be read by the test without synchronization, if the test didn't
manage to examine the frame in between captures the adapted frame would
be overwritten by the following frame during accesses to it.

The actual races are suppressed by race:webrtc/base/messagequeue.cc and
race:webrtc/base/thread.cc. These fixes reduce the suppression count
locally from around 3000 to 30 for VideoAdapterTest.*.

Also removing tsan suppressions for talk/base as it's been moved to
webrtc/base.

R=tommi@webrtc.org
BUG=3671

Review URL: https://webrtc-codereview.appspot.com/22169004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6979 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoUpdating svn:ignore entries
kjellander@webrtc.org [Tue, 26 Aug 2014 11:22:54 +0000 (11:22 +0000)]
Updating svn:ignore entries

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6978 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoRemove test constructor in WebRtcVideoEngine2.
pbos@webrtc.org [Tue, 26 Aug 2014 11:08:06 +0000 (11:08 +0000)]
Remove test constructor in WebRtcVideoEngine2.

Removes the need for ::Construct().

BUG=1788
R=pthatcher@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/15279004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6977 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoRefactoring common_audio/signal_processing: Remove macro WEBRTC_SPL_UDIV
bjornv@webrtc.org [Tue, 26 Aug 2014 10:25:10 +0000 (10:25 +0000)]
Refactoring common_audio/signal_processing: Remove macro WEBRTC_SPL_UDIV

This macro is a direct use of the division operator without checking for division by zero. Hence, it is dangerous to use.
This CL replaces the macro with '/' at place.

BUG=3348,3353
TESTED=locally on linux and trybots
R=kwiberg@webrtc.org, tina.legrand@webrtc.org, turaj@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14169004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6976 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agocommon_audio: Re-enable WebRtcSpl_AddSatW32() and WebRtcSpl_SubSatW32() optimizations...
bjornv@webrtc.org [Tue, 26 Aug 2014 09:36:25 +0000 (09:36 +0000)]
common_audio: Re-enable WebRtcSpl_AddSatW32() and WebRtcSpl_SubSatW32() optimizations on armv7

According to the issue, common_audio_unittests failed on armv7. It currently pass, so we should turn it on again. There is no print out in the issue, so the cause of failure is unknown.

BUG=740
TESTED=locally on N7
R=tina.legrand@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14199004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6975 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoRemove Android.mk build files.
pbos@webrtc.org [Tue, 26 Aug 2014 08:48:51 +0000 (08:48 +0000)]
Remove Android.mk build files.

These files are generally not maintained and break, some contain files
that don't exist anymore and do not build anymore. If we need to add
some of these back we should really set up a bot for them.

R=andrew@webrtc.org, glaznev@webrtc.org, henrike@webrtc.org
BUG=

Review URL: https://webrtc-codereview.appspot.com/15249004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6974 4adac7df-926f-26a2-2b94-8c16560cd09d

7 days agoRemove former team members from OWNERS and WATCHLISTS
kjellander@webrtc.org [Tue, 26 Aug 2014 06:12:08 +0000 (06:12 +0000)]
Remove former team members from OWNERS and WATCHLISTS

Remove the following (CCed) former team members from all
OWNERS files and the WATCHLISTS file:
* fischman@
* leozwang@
* mikhal@
* pwestin@
* wu@

BUG=
R=henrike@webrtc.org, niklas.enbom@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22509004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6973 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days ago(Auto)update libjingle 74064646-> 74072040
buildbot@webrtc.org [Mon, 25 Aug 2014 21:10:18 +0000 (21:10 +0000)]
(Auto)update libjingle 74064646-> 74072040

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6972 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoMove constant so it is not stripped out for TSAN bots.
kjellander@webrtc.org [Mon, 25 Aug 2014 19:46:26 +0000 (19:46 +0000)]
Move constant so it is not stripped out for TSAN bots.

BUG=
R=henrike@webrtc.org, kjellander@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22179004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6971 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days ago(Auto)update libjingle 74039473-> 74044292
buildbot@webrtc.org [Mon, 25 Aug 2014 16:07:12 +0000 (16:07 +0000)]
(Auto)update libjingle 74039473-> 74044292

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6970 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoUpdate root OWNERS file
kjellander@webrtc.org [Mon, 25 Aug 2014 14:41:41 +0000 (14:41 +0000)]
Update root OWNERS file

Add kjellander to owner for the new way of
syncing Chromium deps.
Remove redundant webrtc_examples.gyp entry.
Convert the file from Win to Unix line endings.

BUG=
R=tommi@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/19119004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6969 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoAs expected, r6569 (https://code.google.com/p/webrtc/source/detail?r=6965) caused...
solenberg@webrtc.org [Mon, 25 Aug 2014 14:35:40 +0000 (14:35 +0000)]
As expected, r6569 (https://code.google.com/p/webrtc/source/detail?r=6965) caused memcheck bots to complain. Adding expections for that, in line with outher peerconnection tests.

Also, caused some issues with other peerconnection_unittest tests, so changed the design of those.

BUG=
R=kjellander@webrtc.org, perkj@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22019004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6968 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoRoll chromium_revision 289723:291647
kjellander@webrtc.org [Mon, 25 Aug 2014 14:16:32 +0000 (14:16 +0000)]
Roll chromium_revision 289723:291647

To pick up recent fixes after the Chromium Git switch.

Relevant changes pulled in by this roll:
* r291168 refactor sanitizer_options (we can now remove some hacks)
* r291647 roll of nss.gyp (its paths work with how we build for iOS).

BUG=2863,3731
R=iannucci@chromium.org, tommi@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22489004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6967 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoGN: Disable Chromium clang plugins for standalone build.
kjellander@webrtc.org [Mon, 25 Aug 2014 14:15:35 +0000 (14:15 +0000)]
GN: Disable Chromium clang plugins for standalone build.

Now that WebRTC has rolled the chromium_revision past
http://crrev.com/284372 in r6784, clang has become the
default compiler. Since WebRTC standalone code doesn't
yet compile the Chromium Clang plugins enabled, this CL
disables them for the parts of the code that doesn't yet pass
compilation with them enabled.

The buildbots are using Goma which is not yet switched
over to Clang by default. That's why they're not red yet.

BUG=163
TEST=Passing compile locally on Linux using:
gn gen out/Debug --args="build_with_chromium=false is_debug=true" && ninja
-C out/Debug
gn gen out/Release --args="build_with_chromium=false is_debug=false" && ninja
-C out/Release
gn gen out/Default --args="build_with_chromium=false os=\"android\" cpu_arch=\"arm\" arm_version=7" && ninja -C out/Default

R=brettw@chromium.org

Review URL: https://webrtc-codereview.appspot.com/16279004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6966 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days ago(Auto)update libjingle 73927775-> 74032598
buildbot@webrtc.org [Mon, 25 Aug 2014 12:11:58 +0000 (12:11 +0000)]
(Auto)update libjingle 73927775-> 74032598

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6965 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoRefactoring common_audio: Replace trivial multiplication macro
bjornv@webrtc.org [Mon, 25 Aug 2014 11:42:42 +0000 (11:42 +0000)]
Refactoring common_audio: Replace trivial multiplication macro

This multiplication macro literally use the '*' operator, so there is no need for it.

BUG=3348,3353
TESTED=locally on linux and trybots
R=kwiberg@webrtc.org, tina.legrand@webrtc.org, turaj@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22109004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6964 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoRe-landing r6961
bjornv@webrtc.org [Mon, 25 Aug 2014 11:19:05 +0000 (11:19 +0000)]
Re-landing r6961

common_audio/signal_processing: Remove macro WEBRTC_SPL_MEMCPY_W8

This macro is nothing but memcpy() and further used at one single place in webrtc, so it makes no sense to keep it. Replaced the operation where it is used.

BUG=3348,3353
TESTED=locally on linux
TBR=aluebs@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/18259004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6963 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoRevert 6961 "common_audio/signal_processing: Remove macro WEBRTC..."
bjornv@webrtc.org [Mon, 25 Aug 2014 10:32:22 +0000 (10:32 +0000)]
Revert 6961 "common_audio/signal_processing: Remove macro WEBRTC..."

> common_audio/signal_processing: Remove macro WEBRTC_SPL_MEMCPY_W8
>
> This macro is nothing but memcpy() and further used at one single place in webrtc, so it makes no sense to keep it. Replaced the operation where it is used.
>
> BUG=3348,3353
> TESTED=locally on linux and trybots
> R=kwiberg@webrtc.org, tina.legrand@webrtc.org
>
> Review URL: https://webrtc-codereview.appspot.com/16359004

TBR=bjornv@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22499004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6962 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agocommon_audio/signal_processing: Remove macro WEBRTC_SPL_MEMCPY_W8
bjornv@webrtc.org [Mon, 25 Aug 2014 10:23:22 +0000 (10:23 +0000)]
common_audio/signal_processing: Remove macro WEBRTC_SPL_MEMCPY_W8

This macro is nothing but memcpy() and further used at one single place in webrtc, so it makes no sense to keep it. Replaced the operation where it is used.

BUG=3348,3353
TESTED=locally on linux and trybots
R=kwiberg@webrtc.org, tina.legrand@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/16359004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6961 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoRefactoring common_audio/signal_processing: Replaces trivial macros
bjornv@webrtc.org [Mon, 25 Aug 2014 07:44:52 +0000 (07:44 +0000)]
Refactoring common_audio/signal_processing: Replaces trivial macros

The macros WEBRTC_SPL_ADD_SAT_W16 and WEBRTC_SPL_ADD_SAT_W32 make direct use of the corresponding functions WebRtcSpl_AddSatW16() and WebRtcSpl_AddSatW32().
This CL replaces these macros in the code.

BUG=3348,3353
TESTED=locally on linux and trybots
R=kwiberg@webrtc.org, tina.legrand@webrtc.org, turaj@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21199004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6960 4adac7df-926f-26a2-2b94-8c16560cd09d

8 days agoFix WEBRTC_AEC_DEBUG_DUMP (broken by int16->float conversion)
kwiberg@webrtc.org [Mon, 25 Aug 2014 06:26:04 +0000 (06:26 +0000)]
Fix WEBRTC_AEC_DEBUG_DUMP (broken by int16->float conversion)

And in the process, make it dump WAV files instead of raw PCM.

R=andrew@webrtc.org, bjornv@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/19089004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6959 4adac7df-926f-26a2-2b94-8c16560cd09d

11 days ago(Auto)update libjingle 73927658-> 73927775
buildbot@webrtc.org [Fri, 22 Aug 2014 22:27:04 +0000 (22:27 +0000)]
(Auto)update libjingle 73927658-> 73927775

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6958 4adac7df-926f-26a2-2b94-8c16560cd09d

11 days ago(Auto)update libjingle 73891518-> 73927658
buildbot@webrtc.org [Fri, 22 Aug 2014 22:24:54 +0000 (22:24 +0000)]
(Auto)update libjingle 73891518-> 73927658

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6957 4adac7df-926f-26a2-2b94-8c16560cd09d

11 days ago(Auto)update libjingle 73794259-> 73891518
buildbot@webrtc.org [Fri, 22 Aug 2014 14:08:15 +0000 (14:08 +0000)]
(Auto)update libjingle 73794259-> 73891518

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6955 4adac7df-926f-26a2-2b94-8c16560cd09d

11 days agoRemove static initializer in WebRtcVideoEngine2.
pbos@webrtc.org [Fri, 22 Aug 2014 10:36:23 +0000 (10:36 +0000)]
Remove static initializer in WebRtcVideoEngine2.

Blocks import into chromium.

R=tommi@webrtc.org
BUG=

Review URL: https://webrtc-codereview.appspot.com/18249004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6954 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoIncrement sync_chromium.py version to force re-sync
kjellander@webrtc.org [Thu, 21 Aug 2014 19:51:06 +0000 (19:51 +0000)]
Increment sync_chromium.py version to force re-sync

This should make the remaining red Windows bots cycle green.
Currently, some of them are in a bad state for the Chromium
checkout.

BUG=webrtc:2863
TBR=niklas.enbom@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/18229004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6953 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoMake the last_sync_chromium file a bit more comprehensive.
iannucci@chromium.org [Thu, 21 Aug 2014 15:48:23 +0000 (15:48 +0000)]
Make the last_sync_chromium file a bit more comprehensive.

Adds a SCRIPT_VERSION and the target_os_list to the flag file content. The
script version is so that we can arbitrarially make all slaves/devs re-sync (in
case we change the implementation but don't want to roll chromium), and the
target_os_list is so that devs who change the target_os_list in their .gclient
file don't mysteriously fail to get the new deps.

R=kjellander@webrtc.org, agable@chromium.org, szager@chromium.org
BUG=2863, chromium:339647

Review URL: https://webrtc-codereview.appspot.com/17189004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6952 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoLanding issue 15189004
niklas.enbom@webrtc.org [Thu, 21 Aug 2014 14:49:28 +0000 (14:49 +0000)]
Landing issue 15189004

TBR=henrike@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/14189004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6951 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoMaking sure muc members get recorded.
phoglund@webrtc.org [Thu, 21 Aug 2014 09:53:28 +0000 (09:53 +0000)]
Making sure muc members get recorded.

This is an upstream of a change I made; will describe in a separate
email thread.

Essentially, the members map wasn't getting populated in the callclient
example, so it was always empty. Now it will be populated correctly.

R=henrike@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/13189004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6950 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoAdd send-side bit-exactness test for AudioCoding Module
henrik.lundin@webrtc.org [Thu, 21 Aug 2014 08:59:14 +0000 (08:59 +0000)]
Add send-side bit-exactness test for AudioCoding Module

This test verifies bit exactness for the send-side of ACM. The test
setup is a chain of three different test classes:

test::AcmSendTest -> AcmSenderBitExactness -> test::AcmReceiveTest

The receiver side is driving the test by requesting new packets from
AcmSenderBitExactness::NextPacket(). This method, in turn, asks for the
packet from test::AcmSendTest::NextPacket, which inserts audio from the
input file until one packet is produced. (The input file loops
indefinitely.)  Before passing the packet to the receiver, the
AcmSenderBitExactness class verifies the packet header and updates a
payload checksum with the new payload. The decoded output from the
receiver is also verified with a (separate) checksum.

The current CL only adds tests for 30 ms and 60 ms iSAC. More codecs
will be added in coming changes.

BUG=3521
R=kwiberg@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/20179004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6949 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoUse a deterministic input in NetEqBgnTest
henrik.lundin@webrtc.org [Thu, 21 Aug 2014 08:27:44 +0000 (08:27 +0000)]
Use a deterministic input in NetEqBgnTest

This test has been failing every now and then. This is likely due to the
random input that was used. With this change, the input is now read from
an audio file, making it identical on each run.

The encoding is moved to inside the main test loop, so that new data is
added with each packet. (Before this change, the same payload was added
over and over again; only the RTP header was updated.)

BUG=3715
R=turaj@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/19079004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6948 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoRefactoring common_audio/signal_processing: Remove unused macro WEBRTC_SPL_MUL_32_32_...
bjornv@webrtc.org [Thu, 21 Aug 2014 06:13:57 +0000 (06:13 +0000)]
Refactoring common_audio/signal_processing: Remove unused macro WEBRTC_SPL_MUL_32_32_RSFT32BI

The WEBRTC_SPL_MUL_32_32_RSFT32BI macro was removed in r6169, since it was unused. This CL removes the arm and mips optimizations of it.

BUG=3348, 3353
TESTED=locally and trybots
TBR=tina.legrand@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/17149004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6947 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoFix clang -Wformat warnings.
thakis@chromium.org [Thu, 21 Aug 2014 02:23:30 +0000 (02:23 +0000)]
Fix clang -Wformat warnings.

An unsigned int was passed through %lu instead of %u (harmless on 32bit).
More seriously, a wide string was passed through %s, which means only the
first byte in the string got printed (since the 2nd byte is likely 0 in
UCS-2). Use %ls to include the whole string, even though it might not be
renderable in the target 8bit buffer.

BUG=chromium:82385
R=henrike@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22409004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6946 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoConvert nsx_core_neon.S to unified syntax.
thakis@chromium.org [Thu, 21 Aug 2014 02:23:26 +0000 (02:23 +0000)]
Convert nsx_core_neon.S to unified syntax.

That way, it builds with both gcc and clang's integrated assembler.
No intentional behavior change.

BUG=chromium:124610
R=andrew@webrtc.org, johannkoenig@google.com

Review URL: https://webrtc-codereview.appspot.com/15199004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6945 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoUse --gclientfile instead of --spec, because windows is THE WORST.
iannucci@chromium.org [Thu, 21 Aug 2014 02:14:11 +0000 (02:14 +0000)]
Use --gclientfile instead of --spec, because windows is THE WORST.

--spec contains newlines, which are interpreted as actual newlines in the
command line, which causes gclient to fall apart at the seams.

TBR=agable@chromium.org, kjellander@webrtc.org, szager@chromium.org
BUG=2863, chromium:339647

Review URL: https://webrtc-codereview.appspot.com/22429004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6944 4adac7df-926f-26a2-2b94-8c16560cd09d

12 days agoMake sync_chromium use the git-cache when on the bots.
iannucci@chromium.org [Wed, 20 Aug 2014 23:53:59 +0000 (23:53 +0000)]
Make sync_chromium use the git-cache when on the bots.

This should help bootstrapping speed, as well as allow better clobbering
support.

R=agable@chromium.org
TBR=agable@chromium.org, kjellander@webrtc.org, szager@chromium.org
BUG=2863, chromium:339647

Review URL: https://webrtc-codereview.appspot.com/17179004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6943 4adac7df-926f-26a2-2b94-8c16560cd09d

13 days agoBump WebRTC version number. Starting now, we will be setting WebRTC major version...
tnakamura@webrtc.org [Wed, 20 Aug 2014 22:44:18 +0000 (22:44 +0000)]
Bump WebRTC version number. Starting now, we will be setting WebRTC major version numbers to align with Chrome.

R=niklas.enbom@webrtc.org
TBR=niklas.embom

Review URL: https://webrtc-codereview.appspot.com/15219004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6942 4adac7df-926f-26a2-2b94-8c16560cd09d

13 days agoIncrease verbosity for gclient sync of Chromium
kjellander@webrtc.org [Wed, 20 Aug 2014 14:25:43 +0000 (14:25 +0000)]
Increase verbosity for gclient sync of Chromium

In r6939 the --verbose flag was passed to the problematic
(approx 2.2GB large) gclient sync of Chromium's src.git repo.
However the bots are still hitting killed sync jobs due to
lack of output. This is a speculative attempt to provoke
even more logging, in order to trigger buffer flushing for
the buildbot execution.

BUG=2863, chromium:339647
TEST=Ran gclient runhooks locally with CHROME_HEADLESS=1 set.
TBR=phoglund@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21269004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6940 4adac7df-926f-26a2-2b94-8c16560cd09d

13 days agoPass --verbose to gclient sync of Chromium
kjellander@webrtc.org [Wed, 20 Aug 2014 13:04:49 +0000 (13:04 +0000)]
Pass --verbose to gclient sync of Chromium

In r6938 the switch to using Chromium's Git repo was
deployed. However this fails on the bots since their timeout
for steps without output is 1200 seconds, which is not enough
to checkout the large Chromium Git repo.
Adding --verbose will print more output, thus getting a longer
timeout that should be enough for the runhooks step to complete.

BUG=2863, chromium:339647
TEST=None
TBR=phoglund@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/15209004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6939 4adac7df-926f-26a2-2b94-8c16560cd09d

13 days agoMake WebRTC work with Chromium Git checkouts
kjellander@webrtc.org [Wed, 20 Aug 2014 12:10:11 +0000 (12:10 +0000)]
Make WebRTC work with Chromium Git checkouts

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

13 days agoAdd TSAN suppression for heap-use-after-free in libvpx
henrik.lundin@webrtc.org [Wed, 20 Aug 2014 11:07:29 +0000 (11:07 +0000)]
Add TSAN suppression for heap-use-after-free in libvpx

BUG=3671
R=kjellander@webrtc.org, pbos@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/18199004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6937 4adac7df-926f-26a2-2b94-8c16560cd09d

13 days agoRemove DEPS reference to third_party/clang_format
kjellander@webrtc.org [Wed, 20 Aug 2014 10:47:47 +0000 (10:47 +0000)]
Remove DEPS reference to third_party/clang_format

Clang format has moved into Chromium's src/buildtools
and the last traces from third_party/clang_format were
removed in http://crrev.com/285030.

This removes it from the WebRTC checkouts as well (it is
now an tree of empty directories).

Our DEPS entry for removing the old binaries from pre-move
into src/buildtools was removed in
https://code.google.com/p/webrtc/source/detail?r=6788

BUG=
R=henrika@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/21259004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6936 4adac7df-926f-26a2-2b94-8c16560cd09d

13 days agoRefactoring common_audio: Remove macro WEBRTC_SPL_MEMMOVE_W16
bjornv@webrtc.org [Wed, 20 Aug 2014 10:09:34 +0000 (10:09 +0000)]
Refactoring common_audio: Remove macro WEBRTC_SPL_MEMMOVE_W16

Yet another macro that utilizes a function directly.

BUG=3348,3353
TESTED=locally on linux and trybots
R=kwiberg@webrtc.org, turaj@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/18159004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6935 4adac7df-926f-26a2-2b94-8c16560cd09d

13 days agoDisable two tests in TurnPortTest
henrik.lundin@webrtc.org [Wed, 20 Aug 2014 09:47:58 +0000 (09:47 +0000)]
Disable two tests in TurnPortTest

The tests are flaky.

BUG=3720
TBR=jiayl@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/17159004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6934 4adac7df-926f-26a2-2b94-8c16560cd09d

13 days ago(Auto)update libjingle 73627179-> 73695227
buildbot@webrtc.org [Wed, 20 Aug 2014 07:49:30 +0000 (07:49 +0000)]
(Auto)update libjingle 73627179-> 73695227

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6933 4adac7df-926f-26a2-2b94-8c16560cd09d

13 days agoNew utility class for easy debug dumping to WAV files
kwiberg@webrtc.org [Wed, 20 Aug 2014 07:42:46 +0000 (07:42 +0000)]
New utility class for easy debug dumping to WAV files

There are currently a number of places in the code where we dump audio
data in various stages of processing for debug purposes. Currently
these all write raw, uncompressed PCM files, which isn't supported by
the most common audio players, and requires the user to supply
metadata such as sample rate, sample size and endianness, etc.

This patch adds a simple class that makes it easy to write WAV files
instead. WAV files still contain the same uncompressed PCM data, but
they have a small header that contains all the requisite metadata, and
are supported by virtually all audio players.

Since some of the debug code that will be writing WAV files is written
in plain C, a C API is included as well.

R=andrew@webrtc.org, bjornv@webrtc.org, henrike@webrtc.org, tommi@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/16809004

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6932 4adac7df-926f-26a2-2b94-8c16560cd09d

2 weeks agoMinor bug fix and cosmetic changes in AEC MIPS optimizations.
andrew@webrtc.org [Tue, 19 Aug 2014 15:42:50 +0000 (15:42 +0000)]
Minor bug fix and cosmetic changes in AEC MIPS optimizations.

Minor bug fix in WebRtcAec_FilterAdaptation_mips, which did not manifest with
gcc 4.7.2, but it did with version 4.9.0. While there, also made some cosmetic
changes to comply with Chromium coding style.

R=andrew@webrtc.org

Review URL: https://webrtc-codereview.appspot.com/22399004

Patch from Ljubomir Papuga <lpapuga@mips.com>.

git-svn-id: http://webrtc.googlecode.com/svn/trunk@6931 4adac7df-926f-26a2-2b94-8c16560cd09d