GitLab

We set the audible state every time after the pipeline
has successfully decoded the frame: to false if prerolling
and true otherwise, taking the volume into account.

Also, we set the state to false if the flow is going to be
 interrupted because of a coded error, abort or Pause.

As well, we set audible state to false when the starvation
happens and there is no data in the MediaDecoderJob queue,
we believe this indicates the demuxer freeze.

BUG=414810

Review URL: https://codereview.chromium.org/1008093002

Cr-Commit-Position: refs/heads/master@{#322102}

We use clang with Wextra which seems to be more reliable and report less false positives. We no longer have an intent to turn on Wextra for gcc on posix-like systems any more (except MacOS).

Review URL: https://codereview.chromium.org/1029953003

Cr-Commit-Position: refs/heads/master@{#322101}

With a few webgl aquarium demo windows open and a 1080p video playing
at the same time, you can notice the video dropping a lot frames. From
time to time being stuck on a same frame for a second or more.

This is due to the VideoRendererImpl dropping frames. It takes too
much time for the compositor to paint its content, keeping the GPU
busy and delaying decoded video frames too long. Decoded frames arrive
too late and are never presented and we stick to an old frame.

This changes the behavior of VideoRendererImpl to a best effort
strategy where we might start displaying images past their display
time, but at least we're not stuck on a same image for too long.

BUG=chrome-os-partner:37786
TEST=run the test as explained on the bug

Review URL: https://codereview.chromium.org/1005303004

Cr-Commit-Position: refs/heads/master@{#322100}

Update scanning to scan files in small batches (well batches of 1)...this fixes an issue where scanning appeared to be frozen for long periods of time.

This change also sets the stage for making Scans cancelable...which we'll want to do when a user changes directory...and even allow them to explicitly cancel a scan from the details panel.

TEST=MediaScannerTest

TBR=hirono  // for mock_media_scanner.js

Review URL: https://codereview.chromium.org/1038433002

Cr-Commit-Position: refs/heads/master@{#322099}

BUG=446327

Review URL: https://codereview.chromium.org/1026553005

Cr-Commit-Position: refs/heads/master@{#322098}

ICU 54 fixed this issue. (see http://bugs.icu-project.org/trac/ticket/10295 )

BUG=270037
TEST=TSAN build of unit_tests --gtest_filter=*ShorterUR*

Review URL: https://codereview.chromium.org/1024463010

Cr-Commit-Position: refs/heads/master@{#322097}

Needed for exposing graphics API to shared library replacement.

BUG=

Review URL: https://codereview.chromium.org/1036503003

Cr-Commit-Position: refs/heads/master@{#322096}

BUG=469868

Review URL: https://codereview.chromium.org/1027393002

Cr-Commit-Position: refs/heads/master@{#322095}

This CL adds MOJO_SIGNAL_PEER_CLOSED for the MessagePipeReader's
waiting bits. Without this, PRECONDITION_FAILED error can be notified
at the end of the channel lifetime, and the browser process misinterprets
it as a renderer crash, which results sadface screen.

This rarely happens and is reproduced only with some underpowered CPU
with excessive pressure. So we cannot have any reliable test.

BUG=466814
R=viettrungluu@chromium.org

Review URL: https://codereview.chromium.org/1023043002

Cr-Commit-Position: refs/heads/master@{#322094}

The button is still used for layout, the size of the user name text
remains resized according to the button, but this way we don't have
to override the Paint() function which helps us make Paint() no longer
virtual.

R=pkotwicz, skuhne,sky
BUG=466426

Review URL: https://codereview.chromium.org/1037443003

Cr-Commit-Position: refs/heads/master@{#322093}

It was disabled due to crbug.com/466814 that is now fixed.
It's time to turn it on again.

R=viettrungluu@chromium.org
BUG=466814,377980

Review URL: https://codereview.chromium.org/1021393002

Cr-Commit-Position: refs/heads/master@{#322092}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/3be346d..0ba31e2

TBR=alph@chromium.org,jchaffraix@chromium.org

Review URL: https://codereview.chromium.org/1035513004

Cr-Commit-Position: refs/heads/master@{#322091}

BUG=463125

Review URL: https://codereview.chromium.org/1032513003

Cr-Commit-Position: refs/heads/master@{#322090}

BUG=470229

Review URL: https://codereview.chromium.org/1033773002

Cr-Commit-Position: refs/heads/master@{#322089}

Only takes effect in --site-per-process mode.  Just the first phase of
adding frame entries; won't handle nested iframes or history navigations
yet.

BUG=236848
TEST=New unit and browser tests pass in --site-per-process mode.

Review URL: https://codereview.chromium.org/1029893002

Cr-Commit-Position: refs/heads/master@{#322088}

Remove references to android_webview_build now that we no longer support
that build configuration. Fold all conditions into the general
OS==android sections.

BUG=440793

Review URL: https://codereview.chromium.org/1010653003

Cr-Commit-Position: refs/heads/master@{#322087}

Enums used to simply be strings, but now we expose them (as string constants)
on the api object as part of the extension bindings. As such, it's appropriate
to have the externs generate an enum definition for them.

BUG=469920

Review URL: https://codereview.chromium.org/1015033003

Cr-Commit-Position: refs/heads/master@{#322086}

BUG=462584

Review URL: https://codereview.chromium.org/1028243002

Cr-Commit-Position: refs/heads/master@{#322085}

In this case Segoe UI and its linked fonts are already in the fallback list.
This is an attempt to fix the perf regression caused on the page_cycler.intl_hi_ru bot caused by the change to
remove duplicate fonts from the fallback list.

BUG=469700

Review URL: https://codereview.chromium.org/999173008

Cr-Commit-Position: refs/heads/master@{#322084}

R=brettw@chromium.org, erg@chromium.org, jamesr@chromium.org
TBR=jamesr@chromium.org

Review URL: https://codereview.chromium.org/1029113005

Cr-Commit-Position: refs/heads/master@{#322083}

This makes the BoundedLabel override View::OnPaint, allowing us to make
View::Paint non-virtual.

The functional difference is that Paint() calls PaintCommon() which:
 - Early outs if not visible
 - Paints any children

The BoundedLabel is used by NotificationView, but no children are ever
added to it and it's not exposed outside the class. So it's safe to
convert it to use OnPaint() and remove the visible early out.

R=dewittj@chromium.org, mukai@chromium.org, stevenjb@chromium.org
BUG=466426

Review URL: https://codereview.chromium.org/1035553002

Cr-Commit-Position: refs/heads/master@{#322082}

also, only cover the controls and not the instructions
also various tweaks to colors and font sizes and layout as per latest mocks.

BUG=none

Review URL: https://codereview.chromium.org/1008233004

Cr-Commit-Position: refs/heads/master@{#322081}

This upstreams all the injected JS for the web layer (but not any of
the packaging bits, which will come later)

BUG=464810

Review URL: https://codereview.chromium.org/1029983002

Cr-Commit-Position: refs/heads/master@{#322080}

This CL also have some renames and code changes related to state.

BUG=463671

Review URL: https://codereview.chromium.org/1026323002

Cr-Commit-Position: refs/heads/master@{#322079}

This will enable building the ANGLE perf tests on Windows perf bot
builders.

BUG=angleproject:744
BUG=468852
R=sievers@chromium.org, thakis@chromium.org

Review URL: https://codereview.chromium.org/1021763002

Cr-Commit-Position: refs/heads/master@{#322078}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/6b6978f..3be346d

TBR=alph@chromium.org,jchaffraix@chromium.org

Review URL: https://codereview.chromium.org/1024553011

Cr-Commit-Position: refs/heads/master@{#322077}

In order to avoid future regressions.

BUG=469235
TEST=interactive_ui_tests --gtest_filter=AcceleratorInteractiveUITest.*

Review URL: https://codereview.chromium.org/1026513005

Cr-Commit-Position: refs/heads/master@{#322076}

This logic was simplified and made more robust with
http://crrev.com/848003006, but there is still a report that this is still a
problem.

I am suspicious of the popover detection logic. The popover could belong
to any window, and could be in the middle of fading in or out. Change the
logic to a hopefully more robust on-demand detection using
[NSApplication targetForAction:@selector(performClose:].

BUG=47134

Review URL: https://codereview.chromium.org/1025793002

Cr-Commit-Position: refs/heads/master@{#322075}

1.6.0c provides the ability for clients to manually provide an app version when instantiating the analytics service.

BUG=469258

Review URL: https://codereview.chromium.org/1036443004

Cr-Commit-Position: refs/heads/master@{#322074}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/46bffd6..48dc6dc

CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
TBR=hcm@google.com

Review URL: https://codereview.chromium.org/1011383008

Cr-Commit-Position: refs/heads/master@{#322073}

There are assumptions that the IPC message limit is 128MB
so ChannelMojo has to allow messages in upto the message size.
Such large messages happen when IDB reads/writes huge records.

BUG=464151
R=agl@chromium.org

Review URL: https://codereview.chromium.org/1029703002

Cr-Commit-Position: refs/heads/master@{#322072}

The answers in suggest doesn't put a space between, for example
+ the degrees Fahrenheit and the day of the week
+ the stock price and the stock delta (+/- value)

This CL puts a space between those values, which is more readable for the user.

BUG=

Review URL: https://codereview.chromium.org/1023993003

Cr-Commit-Position: refs/heads/master@{#322071}

Gate out ICU initialization for OS_NACL,
as it calls other code that is not built for nacl.
Undetected in gyp build since we always build the nacl irt as release
with gc-sections.

Stop building the ppb_var_deprecated_proxy as it's only used in the
trusted build for flash.
Undetected in gyp build since we always build the nacl irt as release
with gc-sections.

BUG=https://code.google.com/p/chromium/issues/detail?id=462791
TEST=trybots
R=dmichael@chromium.org,dpranke@chromium.org

Review URL: https://codereview.chromium.org/1034553002

Cr-Commit-Position: refs/heads/master@{#322070}

Design doc: https://docs.google.com/a/chromium.org/document/d/1PCQltNDdyZyuPHUdqYIjeObRrVOhf8WTr6BLfuTQ4ew/edit?pli=1#

BUG=468093
NOPRESUBMIT=true

Review URL: https://codereview.chromium.org/1015623005

Cr-Commit-Position: refs/heads/master@{#322069}

The HTML encoding spec unified Big5 and Big5-HKSCS and we're not
distinguishing them any more.

The spec also has IBM866 that is not in the menu, yet.

BUG=None
TEST=manual:
1. load a page in ibm866 and check the encoding in More Tools|Encoding.
2. load pages labelled as in Big5 and Big5-HKSCS and check the encoding.
TEST=browser_tests --gtest_filter=*ncodin*
TEST=unit_tests --gtest_filter EncodingMenu*

Review URL: https://codereview.chromium.org/1024713002

Cr-Commit-Position: refs/heads/master@{#322068}

BUG=439651

Review URL: https://codereview.chromium.org/1031943002

Cr-Commit-Position: refs/heads/master@{#322067}

I extended the current CaptureAPIType exposure on chrome://media-internals page to work on ANDROID with the enum types defined below.

    enum CaptureApiType {
      API1,
      API2_LEGACY,
      API2_FULL,
      API2_LIMITED,
      TANGO,
      API_TYPE_UNKNOWN
    };

Enums are sourced from media/video/capture/video_capture_device.h file and auto-generated on the Android Java side by using java_cpp_enum scripts. They correspond to the same int values, and this is how it is communicated to cpp side on media/video/capture/android/video_capture_device_factory_android.cc.

                        +------------+
                        |VideoCapture|
                        +--+-------+-+
                           |       |
          +----------------+-+   +-+-----------------+
          |VideoCaptureCamera|   |VideoCaptureCamera2|
          +---+------------+-+   +-------------------+
              |            |
+-------------+-----+  ----+-------------+
|VideoCaptureAndroid|  |VideoCaptureTango|
+-------------------+  +-----------------+

On Java side, camera class structure is as above. CaptureApiType is defined by static getCaptureApiType() methods in each child.

Note that media/video/capture/android/imageformat_list.h is deleted as it is currently not used. Instead, enums defined on media/video/capture/android/video_capture_device_factory_android.h were autogenerated on Java side using java_cpp_enum scripts.

BUG=458743

TEST=Built with gyp and deployed:
- on Nexus 7, shows API2_LEGACY
- on Nexus 5, shows API2_LIMITED for front camera and API2_FULL for back camera
I added Android tests to MediaInternalsVideoCaptureDeviceTest.* and tested on Nexus 7.
Looked for the adding unittests on video_capture_device_unittest but ANDROID tests are currently disabled with a TODO specified[0]. I will investigate further, but let me know if you have suggestions.

[0] https://code.google.com/p/chromium/codesearch#chromium/src/media/video/capture/video_capture_device_unittest.cc&l=42

Review URL: https://codereview.chromium.org/1014943005

Cr-Commit-Position: refs/heads/master@{#322066}

BUG=465884

Review URL: https://codereview.chromium.org/1016113002

Cr-Commit-Position: refs/heads/master@{#322065}

This provides a list of new browser to run perf try jobs on tryserver.chromium.perf:
try-all: For tryjobs on all bisect bots (android, mac, win, winx64, linux).
trybot-all-android: tryjobs on all android bisect bots.
trybot-all-win: tryjobs on all windows bisect bots (win, win x64).
trybot-all-mac: tryjobs on all mac bisect bots.
trybot-all-linux: tryjobs on all linux bisect bots.

BUG=467124

Review URL: https://codereview.chromium.org/1020103006

Cr-Commit-Position: refs/heads/master@{#322064}

Revert of Decrease the lag when switching between different categories in the Cros wallpaper selector. (patchset #5 id:140001 of https://codereview.chromium.org/1028513003/)

Reason for revert:
I suspect this is causing failures in the cros_trunk official builder:
http://master.chrome.corp.google.com:8011/builders/cros%20trunk/builds/31816

See https://code.google.com/p/chromium/issues/detail?id=470237

Original issue's description:
> Decrease the lag when switching between different categories in the Cros wallpaper selector.
>
> In the CL, we maintain a thumbnail list so that we don't need to fetch the thumbnail images from file system
> each time we switch between different categories. It might take some time at the first time, but will be
> instantly ready after that.
>
> And we also just preload grid items' thumbnail images in the current viewport, so that we don't experience noticeable
> lag when switching to a category at the first time.
>
> BUG=454932
> TEST=tested on devices (pixel & peppy)
>
> Committed: https://crrev.com/bb033ee318fc12de0b9e136c99acd25950b08e8f
> Cr-Commit-Position: refs/heads/master@{#322015}

TBR=bshe@chromium.org,xdai@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=454932

Review URL: https://codereview.chromium.org/1030973002

Cr-Commit-Position: refs/heads/master@{#322063}