GitLab

BUG=none

Review-Url: https://codereview.chromium.org/2779293002
Cr-Commit-Position: refs/heads/master@{#460545}

ElementTraversal::firstWithin is the same as doing
::next(rootNode, &rootNode) we can just do that here. We can also use
isOpenOrV0() to remove some verbose checks, and finally we don't need to
check isShadowHost (which contains a bunch of duplciate checks) and can
just null check ::shadow() which is equivalent.

BUG=703900

Review-Url: https://codereview.chromium.org/2782483004
Cr-Commit-Position: refs/heads/master@{#460544}

Promote DC layers to DirectComposition SwapChains. Currently only videos
are supported, so an ID3D11VideoProcessor is used to copy the video data
from the DXGI image's texture into the YUY2 swapchain backbuffer.

BUG=678800
NOTRY=true
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2749023011
Cr-Commit-Position: refs/heads/master@{#460543}

Reland of Add EnabledStateObserver to BackgroundTracingManager (patchset #1 id:1 of https://codereview.chromium.org/2785663002/ )

Reason for revert:
|config_| can be null. So, do not dereference it.

Original issue's description:
> Revert of Add EnabledStateObserver to BackgroundTracingManager (patchset #3 id:130001 of https://codereview.chromium.org/2769963005/ )
>
> Reason for revert:
> This CL causes tests failures (browser_side_navigation_content_browsertests, content_browsertests, site_per_process_content_browsertests).
>
> https://findit-for-me.appspot.com/waterfall/build-failure?url=https://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Tests%20%28dbg%29%281%29/builds/24171
>
> Original issue's description:
> > Add EnabledStateObserver to BackgroundTracingManager
> >
> > This CL removes the memory-infra related code into memory-tracing
> > observer. Also the tests implement the observer instead of setting
> > callbacks.
> >
> > BUG=700245
> >
> > Review-Url: https://codereview.chromium.org/2769963005
> > Cr-Commit-Position: refs/heads/master@{#460293}
> > Committed: https://chromium.googlesource.com/chromium/src/+/f64ac5cb273e3abab3564f1a5997b3da164ab9ff
>
> TBR=oysteine@chromium.org,alexmos@chromium.org,ssid@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=700245
>
> Review-Url: https://codereview.chromium.org/2785663002
> Cr-Commit-Position: refs/heads/master@{#460343}
> Committed: https://chromium.googlesource.com/chromium/src/+/88d07c763d35621da2c62b9f18822109bbd7a261

TBR=alexmos@chromium.org,kolos@chromium.org
BUG=700245

Review-Url: https://codereview.chromium.org/2782933002
Cr-Commit-Position: refs/heads/master@{#460542}

Does not compile for flat_map<base::StringPiece, base::TimeDelta>
without this fix.

BUG=none
TEST=none

Review-Url: https://codereview.chromium.org/2783823002
Cr-Commit-Position: refs/heads/master@{#460541}

BUG=608780
NOTRY=true

Review-Url: https://codereview.chromium.org/2779113007
Cr-Commit-Position: refs/heads/master@{#460540}

As requested in 2766093002 after that CL landed.

BUG=None
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2777343005
Cr-Commit-Position: refs/heads/master@{#460539}

Having these enabled can use excessive power sometimes (depending on
the layout of content above and the damage rect), so disable them for now.

BUG=678800
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel

Review-Url: https://codereview.chromium.org/2780073002
Cr-Commit-Position: refs/heads/master@{#460538}

It's a valid case that screen capture continues when Chrome is put to
background on Android(while camera capture should be paused.).
Here a query interface is added in MediaStreamDispatch to return all
the non-screen capture devices on the current Tab. Then RenderViewImpl
can suspend/resume every non-screen capture device only when
OnPageWasHidden/Shown() is called.
And the SuspendDevices() is removed from VideoCaptureManagerImpl to
simplify the logic.

Also because each MediaStreamDispatch corresponds one single Tab, this
means now RenderViewImpl can pause/resume the devices related to the
current Tab only. Which is also a fix to crbug/593192.

BUG=487935, 593192

Review-Url: https://codereview.chromium.org/2763743002
Cr-Commit-Position: refs/heads/master@{#460537}

GoToLineDialog and QuickOpen are spun off from OpenResourceDialog.
QuickOpen switches between GoToLineDialog and OpenResourceDialog depending on
whether the query starts with ':'.
QuickOpen loads delegates through the runtime system.

BUG=662081

Review-Url: https://codereview.chromium.org/2679483002
Cr-Commit-Position: refs/heads/master@{#460536}

BUG=671420

Review-Url: https://codereview.chromium.org/2758803003
Cr-Commit-Position: refs/heads/master@{#460535}

ComputedStyle::getRoundedInnerBorderFor() could generate negative
content boxes for boxes with no content and sub pixel borders that
round up (eg. 1.5px-1.99px).  Negative content boxes are illegal per
spec, and causes BoxBorderPainter::paintBorderFastPath() to bail,
and performance to be substantially degraded.

This issue would be evident on the paint-offset-changes perftest when
device pixel ratio was set to eg. 1.5.  The issue was introduced in
934becac

BUG=692955

Review-Url: https://codereview.chromium.org/2782153002
Cr-Commit-Position: refs/heads/master@{#460534}

BUG=704977

Review-Url: https://codereview.chromium.org/2775773003
Cr-Commit-Position: refs/heads/master@{#460533}

Bug: 706118
Change-Id: I7c61b567ade902a2a2c68ce783096d1c9e77a681
Reviewed-on: https://chromium-review.googlesource.com/461353Reviewed-by: Jeff Carpenter <jeffcarp@chromium.org>
Commit-Queue: Quinten Yearsley <qyearsley@chromium.org>
Cr-Commit-Position: refs/heads/master@{#460532}

1. As part of program shutdown, VideoCaptureManager::StopCaptureForClient() is
   called. This posts a task to destroy the capture device to the "device
   thread", which eventually leads to this destructor being called (on the
   "device thread").
2. While things on the "device thread" are running asynchronously,
   BrowserMainLoop is being destroyed. It's destructor waits for the "device
   thread" (audio thread) to complete.
3. |io_task_runner_| does not accept the posted task anymore.
4. |event.Wait()| will wait indefinitely causing the shutdown to hang.

This deadlock occurred on build bot linux_android_rel_ng in CL
https://codereview.chromium.org/2772963002/. It reproduces locally on a Nexus 5
device running KitKat.

In production, class GpuJpegDecodeAcceleratorHost is currenlty only used on
Chrome OS, but it might happen there as well, since the situation does not
appear to be platform-dependent.

BUG=706186
TEST=
  content_browsertests --gtest_filter="VideoCaptureBrowserTest.*"
  content_unittests --gtest_filter="*Video*"
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2786503002
Cr-Commit-Position: refs/heads/master@{#460531}

Marking move constructors "noexcept" allows std::vector to use the move
constructors in container resizes. Otherwise it will copy which is not
expected by most callers.
http://en.cppreference.com/w/cpp/language/move_constructor

This also replaces some technically invalid characters with HTML entities.

Discussion:
https://groups.google.com/a/chromium.org/d/msg/cxx/ze4WJFg7RvU/Qhbb6DTrBQAJ

Review-Url: https://codereview.chromium.org/2782973002
Cr-Commit-Position: refs/heads/master@{#460530}

And disable it for all desktop & android chrome until the issue is fixed.

This is a reland of crrev.com/459971.

BUG=chromedriver:833

Review-Url: https://codereview.chromium.org/1882603003
Cr-Commit-Position: refs/heads/master@{#459971}
Committed: https://chromium.googlesource.com/chromium/src/+/1512539632abf713b55f5a16fe7b709fb1f4a57d

patch from issue 1882603003 at patchset 40001 (http://crrev.com/1882603003#ps40001)

Review-Url: https://codereview.chromium.org/2779243003
Cr-Commit-Position: refs/heads/master@{#460529}

Currently is was shown right on the window border and thus invisible.

Review-Url: https://codereview.chromium.org/2784843003
Cr-Commit-Position: refs/heads/master@{#460528}

Remove SpecialPopupRow.

BUG=686220

Review-Url: https://codereview.chromium.org/2780743002
Cr-Commit-Position: refs/heads/master@{#460527}

We accidentally didn't handle the case where the iterator was evicted
and invalid. This is now much more explicit and tested.

R=pwnall,cmumford
BUG=705837

Review-Url: https://codereview.chromium.org/2786463002
Cr-Commit-Position: refs/heads/master@{#460526}

EventFilters are currently built by the LayoutManager and all gestures
are routed through it to the active layout. This indirection is
unnecessary since only the StackLayout requires gesture handling. This
also ends up exposing public APIs on the Layouts making it difficult to
follow the path for events from different sources (CompositorView vs
the Toolbar).

This change pushes EventFilters deeper into the Layout implementations,
so the Layouts directly consume touch events and generate gestures if
necessary. Layout still has an API for swipe gestures coming from the
toolbar.

BUG=

Review-Url: https://codereview.chromium.org/2774443002
Cr-Original-Commit-Position: refs/heads/master@{#459980}
Committed: https://chromium.googlesource.com/chromium/src/+/ef25f0df9d626028752de7d945d89abe15d39dfb
Review-Url: https://codereview.chromium.org/2774443002
Cr-Commit-Position: refs/heads/master@{#460525}

BUG=689479

Review-Url: https://codereview.chromium.org/2784673004
Cr-Commit-Position: refs/heads/master@{#460524}

Review-Url: https://codereview.chromium.org/2778943002
Cr-Commit-Position: refs/heads/master@{#460523}

BUG=367113

Review-Url: https://codereview.chromium.org/2648683003
Cr-Commit-Position: refs/heads/master@{#460522}

Revert of Disable SPInvalidation by default (patchset #2 id:20001 of https://codereview.chromium.org/2772833003/ )

Reason for revert:
Enable SPInvalidation on M59.

Original issue's description:
> Disable SPInvalidation by default
>
> We decided to delay launch until M59.
>
> CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
>
> Review-Url: https://codereview.chromium.org/2752673002
> Cr-Commit-Position: refs/heads/master@{#456857}
> TBR=nasko@chromium.org (content_features.cc changes SlimmingPaintInvalidation disabled by default)
>
> Review-Url: https://codereview.chromium.org/2772833003
> Cr-Commit-Position: refs/heads/master@{#459201}
> Committed: https://chromium.googlesource.com/chromium/src/+/a038d334a0e1baae632e583dbc023f3383e8bf6f

TBR=chrishtr@chromium.org,pdr@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.

Review-Url: https://codereview.chromium.org/2782943003
Cr-Commit-Position: refs/heads/master@{#460521}

Previously, if an extension requesting automation has more than one page it opens in its extension process (e.g. background.html and panel.html, in ChromeVox), we instantiated two AutomationInternalCustomBindings (one for each page).

This fix makes it so we only ever make "active" the instance that is associated with the extension's background page.

Note that as a result, any calls to chrome.automation from non-background page contexts will result in an no-op (e.g. getDesktop will not trigger the callback).

In the case that the extension has no background page, we revert to the previous behavior of having one instance per page.

TEST=open the ChromeVox Panel, and close it again. Press caps lock. Verify we only process one event and only output one alert.
BUG=691165
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2782753002
Cr-Commit-Position: refs/heads/master@{#460520}

Revert of [scheduler] Split suspendable tq from unthrottled tq. (patchset #4 id:60001 of https://codereview.chromium.org/2754673002/ )

Reason for revert:
Causing failures in perf tests. Reverting to confirm it is the culprit.

crbug.com/706108
crbug.com/706104

Original issue's description:
> [scheduler] Add suspendable task queue.
>
> Add suspendable task queue to avoid running (and cancelling) callbacks
> when execution context is suspended.
>
> R=skyostil@chromium.org,alexclarke@chromium.org,haraken@chromium.org
> CC=nhiroki@chromium.org,tzik@chromium.org
> BUG=700792
>
> Review-Url: https://codereview.chromium.org/2754673002
> Cr-Commit-Position: refs/heads/master@{#459787}
> Committed: https://chromium.googlesource.com/chromium/src/+/12e16766adec5bb1e5cdd73d3cd8f5016b59ce18

TBR=alexclarke@chromium.org,haraken@chromium.org,skyostil@chromium.org,altimin@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=700792

Review-Url: https://codereview.chromium.org/2780233002
Cr-Commit-Position: refs/heads/master@{#460519}

Review-Url: https://codereview.chromium.org/2770083004
Cr-Commit-Position: refs/heads/master@{#460518}

Ensure the floor calculation is taken into account when running the
layout test. The calculation matches what is happening when e.timeStamp
is being returned.

BUG=706494

Review-Url: https://codereview.chromium.org/2784913002
Cr-Commit-Position: refs/heads/master@{#460517}

When latching is disabled, instead of setting a timer with zero timeout
to send the scrollEnd event, we send the event right away.
This change has happened to prevent unittests from unintentionally
simulating the latching behavior when it's disabled.

BUG=706175

Review-Url: https://codereview.chromium.org/2782733002
Cr-Commit-Position: refs/heads/master@{#460516}

BUG=internal b/36563551

Review-Url: https://codereview.chromium.org/2786493004
Cr-Commit-Position: refs/heads/master@{#460515}

It seems that there are some abusive extensions that act as if they were
other extensions with fake names. Instead of recording only the extensions
names in feedback reports, we should also record the ID and the version.

BUG=699266
TEST=manually

Review-Url: https://codereview.chromium.org/2772403004
Cr-Commit-Position: refs/heads/master@{#460514}

This patch adds the placeholder for NGColumnLayoutAlgorithm class
and refactors NG(Inline|Block|Column)LayoutAlgorithm classes
moving some common logic to NGLayoutAlgorithm base class.

BUG=635619

Review-Url: https://codereview.chromium.org/2782603003
Cr-Commit-Position: refs/heads/master@{#460513}

This patch enables persistence2.0 experiment by default, which results
in the following changes:
- automapping will be enabled by default for all workspace users
  (automapping design doc: https://goo.gl/lqYrql)
- a new UI for workspaces will be enabled by default for all workspace
  users

BUG=277885
R=dgozman

Review-Url: https://codereview.chromium.org/2744553003
Cr-Commit-Position: refs/heads/master@{#460512}

As per crrev.com/2283253002#msg37 this patch updates one of the
conditional branches BrowserNonClientFrameView::DoesIntersectRect to
return true if a large hit-test rectangle overlaps both the tabstrip and
the client area.

Review-Url: https://codereview.chromium.org/2775083002
Cr-Commit-Position: refs/heads/master@{#460511}

Because we now need to maintain state outside of the normal SessionWindow
struct, this patch introduces a SyncedSessionWindow type. It wraps a
normal SessionWindow, while allowing tab sync to track additional window
information.

As part of this patch, I've also cleaned up the unused variation ids.
Moving forward, nothing sync specific should need to be in
SessionWindow. We'll probably need to add a wrapper
for SessionTab in the future as well.

The patch itself does not have any behavioral impact, but paves the way for
consuming sync window type information.

BUG=639009
TBR=bauerb, skuhne, thestig

Review-Url: https://codereview.chromium.org/2499023004
Cr-Commit-Position: refs/heads/master@{#460510}

A major part of android's ChildProcessLauncher is implemented in Java
so having an easy access to the launcher thread is important.

Refactor JavaHandlerThread in base so that that java part can be created
first, and then connected with native peer after consutrction. This is
needed because the launcher thread can be used before native library is
loaded.

Add LauncherThread in content which is a simple wrapper around
JavaHandlerThread. Then for android, override the launcher thread
message loop. Note this means the launcher thread will no longer be
joined on shutdown, but this is not a problem in practice since android
never does clean shutdowns.

Convert two cases of random background threads in ChildProcessLauncher
to use LauncherThread. There are more in the future.

BUG=701442

Review-Url: https://codereview.chromium.org/2774363003
Cr-Commit-Position: refs/heads/master@{#460509}

This CL adds a timer to WebApkIconHasher. Consequently, classes that
invoke WebApkIconHasher::DownloadAndComputeMurmur2Hash no longer need
to maintain a timer dedicated to WebApkIconHasher.

This CL also simplifies the error checking and handling logic.
WebApkIconHasher now checks for URL validity itself, and would return
an empty hash string for any errors encountered.

BUG=649771

Review-Url: https://codereview.chromium.org/2771793003
Cr-Commit-Position: refs/heads/master@{#460508}

even for zero byte files.

URLRequestExtensionJob depends on this behavior, so add a zero-byte
test to url_request_file_job_unittest. Also update other relevant tests
in that file to check for DoneReading.

This is merely a test addition/modification for crbug.com/703892

BUG=703892
Test=None, internal change only.

Review-Url: https://codereview.chromium.org/2776873002
Cr-Commit-Position: refs/heads/master@{#460507}

Added a concrete example in comments on how to call Chrome OS Component API.

BUG=690521
TEST=None

Review-Url: https://codereview.chromium.org/2762393002
Cr-Commit-Position: refs/heads/master@{#460506}