Skip to content

GitLab

T
tangled
Switch branch/tag
  1. 24 Aug, 2018 24 commits
  3. 23 Aug, 2018 16 commits
    • Eric Karl's avatar
      Android OOP-D: Tear down display when going invisible · 7c6485d8
      Eric Karl authored 
      When Android goes invisible in OOP-D, it wasn't tearing down the
display, which can lead to GL issues as we continue to use GL
after the window (used to create the GL surface) is destroyed.

In order to tear down the display for Viz, we need to invalidate
our root frame sink ID. This change refactors things so that we
always invalidate the root frame sink ID on going invisible, and
re-register it on becoming visible. This allows both viz/non-viz
to share the same logic.

As registering/unregistering isn't doing much in non-viz case,
this doesn't add significant overhead there.

Bug: 863049
Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel
Change-Id: I1589e402185fd9e2cdb007d3d8cd739f303ad48a
Reviewed-on: https://chromium-review.googlesource.com/1184376Reviewed-by: default avatarKhushal <khushalsagar@chromium.org>
Reviewed-by: default avatarTom Sepez <tsepez@chromium.org>
Commit-Queue: Eric Karl <ericrk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585664}
      7c6485d8
    • Tom Anderson's avatar
      GTK4: Fix crash when drawing certain frame buttons · 7004b5ed
      Tom Anderson authored 
      BUG=876586
R=sky

Change-Id: Iac4f45a4aec1a905b75332ed7b365d1a2fce1d87
Reviewed-on: https://chromium-review.googlesource.com/1187223
Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
Reviewed-by: default avatarScott Violet <sky@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585663}
      7004b5ed
    • Paul Miller's avatar
      Rename android_webview::GetChannel · f7542b65
      Paul Miller authored 
      Commit 0c0fc42d moved
AwVariationsServiceClient's "channel == unknown ? stable : channel"
logic into a new android_webview::GetChannel. But this is confusing
because the new GetChannel retains the variations-specific comment, and
the name "android_webview::GetChannel" might sound like all of WebView
should be using this sanitized channel, when UMA still needs the
un-sanitized channel.

Move the variations-specific comment back into AwVariationsServiceClient
and rename GetChannel to GetChannelOrStable.

BUG=758290

Change-Id: I4590aa22b73424e14e47444c3fdf3abc151bf441
Reviewed-on: https://chromium-review.googlesource.com/1187271Reviewed-by: default avatarBo <boliu@chromium.org>
Commit-Queue: Paul Miller <paulmiller@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585662}
      f7542b65
    • Aga Wronska's avatar
      Allow to customize day start for activity reporting aggreegation in · abae2fe8
      Aga Wronska authored 
      DeviceStatusCollector.

DeviceStatusCollector takes day start as an argument. For enterprise
reporting it is set to midnight as it was before. For consumer/child
reporting it is read from UsageTimeLimits policy.

If the day start changes data are aggregated into new day buckets
whithout affecting pre-existing buckets. It means that previously
stored data will not be mapped to the new day start buckets.

Bug: 877178
Test: Run tests in device_status_collector_browsertests.
      Manually simulate status reports for enterprise and consumer.


Change-Id: I60fe3a9f22149aefbfe81c0a48ddbf6ed29ce1fd
Reviewed-on: https://chromium-review.googlesource.com/1184397
Commit-Queue: Aga Wronska <agawronska@chromium.org>
Reviewed-by: default avatarRahul Chaturvedi <rkc@chromium.org>
Reviewed-by: default avatarMaksim Ivanov <emaxx@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585661}
      abae2fe8
    • akaba's avatar
      Cleanup surface_aggregator_unittest.cc and surface_synchronization_unittest.cc · 6bc32445
      akaba authored 
      The unit tests should start using SurfaceRanges instead of two SurfaceIds for
simplicity.

Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel
Change-Id: If6d41f07d9f4735399e990aca9c4a8e93449324a
Bug: 876731
Reviewed-on: https://chromium-review.googlesource.com/1184226
Commit-Queue: Fady Samuel <fsamuel@chromium.org>
Reviewed-by: default avatarFady Samuel <fsamuel@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585660}
      6bc32445
    • Scott Violet's avatar
      chromeos: fixs cursor lookup for WebContents · b9741588
      Scott Violet authored 
      Only use the cursor at an embed point if the embedder isn't intercepting
events. I'm assuming that if the embedder is intercepting events, then it's
setting the cursor. This is the case for WebContents.

BUG=877099
TEST=covered by test

Change-Id: I04304a2ae8e668f5a7506378868c2a53eac7de0a
Reviewed-on: https://chromium-review.googlesource.com/1187409
Commit-Queue: Scott Violet <sky@chromium.org>
Reviewed-by: default avatarEvan Stade <estade@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585659}
      b9741588
    • Jun Cai's avatar
      Network Service: Convert port_forwarding_controller.cc to mojo host resolver · 8e3e6c83
      Jun Cai authored 
      Bug: 874651
Change-Id: I4762526cc8fc4dbf18512b8ab7fbeeda8434336d
Reviewed-on: https://chromium-review.googlesource.com/1184306Reviewed-by: default avatarAndrey Kosyakov <caseq@chromium.org>
Reviewed-by: default avatarMatt Menke <mmenke@chromium.org>
Commit-Queue: Jun Cai <juncai@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585658}
      8e3e6c83
    • Kurt Horimoto's avatar
      [iOS] Adds TestFullscreenController functionality. · f2729620
      Kurt Horimoto authored 
      This class now supports:
- enabling/disabling a FullscreenModel,
- sending broadcast values,
- and notifying observers of service shutdown.

Bug: none
Cq-Include-Trybots: luci.chromium.try:ios-simulator-full-configs;master.tryserver.chromium.mac:ios-simulator-cronet
Change-Id: I6d0b6964599aee1606dd992577ddc43d4ce43a4d
Reviewed-on: https://chromium-review.googlesource.com/1185977
Commit-Queue: Kurt Horimoto <kkhorimoto@chromium.org>
Reviewed-by: default avatarEugene But <eugenebut@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585657}
      f2729620
    • Lan Wei's avatar
      Add wheel scroll options "scroll_by_page" to GpuBenchmarking · 7f617841
      Lan Wei authored 
      In GpuBenchmarking, for mouse wheel and touchpad scroll, we want to add
"scroll_by_page" scroll option to GpuBenchmarking::SmoothScrollBy.

Bug: 871970
Change-Id: If73cc2bb41a77499cfc6215b5d8fde09f33f6caa
Reviewed-on: https://chromium-review.googlesource.com/1174993
Commit-Queue: Lan Wei <lanwei@chromium.org>
Reviewed-by: default avatarNasko Oskov <nasko@chromium.org>
Reviewed-by: default avatarSadrul Chowdhury <sadrul@chromium.org>
Reviewed-by: default avatarSahel Sharify <sahel@chromium.org>
Reviewed-by: default avatarDave Tapuska <dtapuska@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585656}
      7f617841
    • Thomas Guilbert's avatar
      Implement FlingingRenderer::GetMediaTime · 8a6d66b9
      Thomas Guilbert authored 
      This CL uses RemoteMediaPlayer.getApproximateStreamPosition to implement
FlingingRenderer::GetMediaTime().

NOTE: GMS core ignores stream position changes in the MediaStatus, when
we have a pending seek. Calling both Seek() and Play() back to back in
FlingingRenderer::StartPlayingFrom() sometimes causes the stream
position update to be lost. This causes getApproximateStreamPosition to
report a time that is out of sync with the cast device. We already use
the autoplay flag when loading the RemoteMediaPlayer, so removing the
call to MediaController::Play() doest not cause any problems.

Bug: 830871
Change-Id: Ia5a6384cf11a77a6aa49e9669f015f7835c5a775
Reviewed-on: https://chromium-review.googlesource.com/1180511Reviewed-by: default avatarZhiqiang Zhang <zqzhang@chromium.org>
Commit-Queue: Thomas Guilbert <tguilbert@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585655}
      8a6d66b9
    • akaba's avatar
      Surface should clear it's observed sinks on deletion · b2998c3d
      akaba authored 
      This CL will force a surface to clear it's observed sinks when it gets
garbage collected. This should fix the memory leak problem.

Bug: 876884
Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel
Change-Id: I957fa212053649bf095925f9e7a7da7cdc218829
Reviewed-on: https://chromium-review.googlesource.com/1185874
Commit-Queue: Fady Samuel <fsamuel@chromium.org>
Reviewed-by: default avatarFady Samuel <fsamuel@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585654}
      b2998c3d
    • Christian Fremerey's avatar
      Reland [Image Capture, Android] Fix various threading issues · 5a9e7de8
      Christian Fremerey authored 
      This is a reland of
https://chromium-review.googlesource.com/c/chromium/src/+/1185262

Patch Set 1 is the same as the reverted CL % resolved merge conflicts.
Patch Set 3 adds a fix for the issue that caused the revert.

Issue Analysis: There was an (existing) race condition where video frames
could still get delivered from the camera after the device had been stopped
and the native object destroyed. The CL made this condition more likely
to happen.

--------------

Original CL description:

This CL simplifies the threading of class VideoCaptureCamera2.java
and by doing so resolves potential issues caused by concurrent
access to member variables and Android video API calls.

See https://bugs.chromium.org/p/chromium/issues/detail?id=857530 for
details on issues.

The simplified model is to (still) have the constructor and public API
calls happen on a native thread, and to use a single dedicated thread
owned by the class instance to post to, do work on, and call back into
the native code.

This CL is part of a series, see Design Doc at
https://docs.google.com/document/d/1h1kva4VR1gaV3HVXaSYZFY41icfaB58j-WnHmJdyqc8/edit?usp=sharing

Bug: 857530
Change-Id: I7e914fead264990fdd644c32b9ba6e681fa39dec
Reviewed-on: https://chromium-review.googlesource.com/1187324Reviewed-by: default avatarEmircan Uysaler <emircan@chromium.org>
Commit-Queue: Christian Fremerey <chfremer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585653}
      5a9e7de8
    • Wez's avatar
      Check that TaskScheduler is not leaked between tests, or test-cases. · a2a4e75d
      Wez authored 
      - Move some APIs off of base::TestSuite, that we unnecessarily public.
- Rename the various internal TestEventListeners to express purpose.
- Add CheckForLeakedGlobals check, and DisableCheckForLeakedGlobals API.
  The latter is called by the content::ContentTestSuiteBase, since the
  various browser- and ui-tests running under that suite base run more
  like the browser itself, and so are expected to let global singletons
  leak.

Bug: 875486
Change-Id: If749dfe8530abc4a5c13a840bbf8258d48c377c5
Reviewed-on: https://chromium-review.googlesource.com/1182612Reviewed-by: default avatarScott Violet <sky@chromium.org>
Reviewed-by: default avatarGabriel Charette <gab@chromium.org>
Commit-Queue: Wez <wez@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585652}
      a2a4e75d
    • Sophie Chang's avatar
      Fix slightly off-center data_reduction_big.xml asset · fe778c04
      Sophie Chang authored 
      bshealy@ says this is ok in comment 21 of the attached bug.

Bug: 867129
Change-Id: I00ea0620f556c33e8741cd565f62ec4bbd65c4a6
Reviewed-on: https://chromium-review.googlesource.com/1187516Reviewed-by: default avatarTheresa <twellington@chromium.org>
Commit-Queue: Sophie Chang <sophiechang@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585651}
      fe778c04
    • Johann's avatar
      Roll src/third_party/libaom/source/libaom/ 7a76b645a..10df8d1b5 (88 commits) · da69f3ae
      Johann authored 
      https://aomedia.googlesource.com/aom.git/+log/7a76b645a08c..10df8d1b5861

$ git log 7a76b645a..10df8d1b5 --date=short --no-merges --format='%ad %ae %s'
2018-08-21 slavarnway Delete loopfilter_avx2.c
2018-08-22 remya.prakasan Highbd Warp module modified for alpha=0 and beta=0
2018-08-22 sanampudi.venkatarao Add eob logic for highbd 64x64 inv_txfm
2018-08-17 binpengsmail Early skip of FILTER_INTRA_MODE search
2018-08-21 angiebird Fix aomenc build err when coeff range check is on
2018-08-21 urvang Add a macro to reduce border extension.
2018-08-21 urvang Avoid redefining 2 macros.
2018-08-21 debargha Clamp to prevent rate overflow
2018-08-17 ravi.chaudhary Add unittest for encoder row based multi-threading
2018-08-16 binpengsmail Skip uv's tx search if y's cost is too large
2018-08-21 deepa.kg Account for mode cost for chroma RD early exit
2018-08-20 binpengsmail Fix speed feature skip_intra_in_interframe
2018-08-20 binpengsmail Add config flag for two_pass_partition_search
2018-08-16 tomfinegan Restore external frame buffer tests.
2018-08-20 johann.koenig convolve neon: hide unused functions
2018-08-20 johann.koenig av1 inv txfm: remove unused function
2018-08-20 sarahparker Remove outdated assert
2018-08-15 binpengsmail Speedup inter_txb_hash's tree generation
2018-08-20 debargha Prevent overflow in model_rd_from_sse
2018-08-20 johann.koenig av1 inv txfm: remove unused function
(...)

Created with:
  roll-dep src/third_party/libaom/source/libaom

Change-Id: I0524db5a3bd1232afb968624ecdff031bf060ef3
Reviewed-on: https://chromium-review.googlesource.com/1187240Reviewed-by: default avatarTom Finegan <tomfinegan@chromium.org>
Commit-Queue: Johann Koenig <johannkoenig@google.com>
Cr-Commit-Position: refs/heads/master@{#585650}
      da69f3ae
    • Evan Stade's avatar
      OopAsh: improve immersive mode reveal widget. · b2270914
      Evan Stade authored 
      1. Make sure located events go to the top container when it's revealed.
This is accomplished with an EventRewriter which clears the target
window, which had been provided by the window service, from the event.
Without this, clicks on the top container are targeted to the
web contents window because the Window Service doesn't know about the
top container's window. In the long term, we may be able to remove this
EventRewriter with a more general solution (e.g. improving how the window
service targets the event in the first place).

2. Instead of mirroring the top container's layers, make the widget
completely transparent except for the ash-provided window controls.
Mirroring doesn't update the reveal widget in response to changes in
the top container's layers, such as when a button is hovered and an ink
drop is added.

3. Allow the reveal widget to accept events so that it's possible to
interact with the window controls. Resize the reveal widget to be just
big enough for the window controls so that clicks anywhere else will
go straight through to the underlying top container view.

4. Make the browser-side non client frame view paint the frame bg. This
is necessary because the reveal widget (i.e. window controls) is now on
top of the container view and has no client-side rendering (what the
mirroring view used to provide).

Future issues to address:
- the cursor is not updated (e.g. when hovering the omnibox) presumably
  because the cursor is coming from the window that the WS thinks is
  underneath the pointer, i.e. the web contents.

Alternatives considered:
- update the mirrored layers every time the layer tree changes. This
  doesn't seem to be easy to accomplish or likely to be very performant.
- move the top container into the reveal widget. This somewhat works
  with a few tweaks here and there to code that assumes the top container
  is always in the browser widget, but ultimately it seems that making
  the reveal widget focusable causes difficulties with window activation
  and focus traversal.

Bug: 640365
Change-Id: I03956c07354e20dadc2919f9a0402ce1f7b4a672
Reviewed-on: https://chromium-review.googlesource.com/1171547
Commit-Queue: Evan Stade <estade@chromium.org>
Reviewed-by: default avatarScott Violet <sky@chromium.org>
Cr-Commit-Position: refs/heads/master@{#585649}
      b2270914