GitLab

This CL replaces the gray outline of the navigation bubble with a
shadow. Also fixes a bug on Kitkat where the bubble appears in dark
gray which was a side effect from RippleBackgroundHelper.

Bug: 979499, 988341
Change-Id: Ibea69d60834d916f99ec7add54e1d4ae7484c5e5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1722581Reviewed-by: John Budorick <jbudorick@chromium.org>
Reviewed-by: Theresa  <twellington@chromium.org>
Commit-Queue: Jinsuk Kim <jinsukkim@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683489}

This reverts commit 536bae08.

Reason for revert: unit_tests failure fixed by crrev.com/c/1732110

Original change's description:
> Revert "Flip the AppServiceAsh feature default to enabled"
> 
> This reverts commit 1b08fa8e.
> 
> Reason for revert: suspect causing unit_tests failure on linux-chromeos-dbg.
> https://ci.chromium.org/p/chromium/builders/ci/linux-chromeos-dbg/13673
> deterministic failures:
> ArcAppLauncherForDefaultAppTest.AppIconUpdated/0
> ArcAppLauncherForDefaultAppTest.AppIconUpdated/1
> ...
> E.g.
> https://logs.chromium.org/logs/chromium/buildbucket/cr-buildbucket.appspot.com/8906357365557383728/+/steps/unit_tests/0/logs/Deterministic_failure:_ArcAppLauncherForDefaultAppTest.AppIconUpdated__x2f_0__status_CRASH_/0
> [ RUN      ] ArcAppLauncherForDefaultAppTest.AppIconUpdated/0
> Received signal 11 <unknown> 000000000000
> #9 0x7f40f3cfd40b base::ThreadCheckerImpl::CalledOnValidThread()
> #10 0x5611fbac0319 arc::internal::ConnectionNotifier::RemoveObserver()
> #11 0x5611f8ea55e9 arc::ConnectionHolder<>::RemoveObserver()
> #12 0x5611ff697745 apps::ArcApps::~ArcApps()
> #13 0x5611ff69786c apps::ArcApps::~ArcApps()
> ...
> #22 0x7f40e9d8e740 KeyedServiceFactory::Disassociate()
> #23 0x7f40e9d8e834 KeyedServiceFactory::ContextDestroyed()
> #24 0x7f40de6a3503 BrowserContextKeyedServiceFactory::BrowserContextDestroyed()
> #25 0x7f40de6a3677 BrowserContextKeyedServiceFactory::ContextDestroyed()
> #26 0x7f40e9d8b145 DependencyManager::DestroyFactoriesInOrder()
> #27 0x7f40e9d8b1f5 DependencyManager::PerformInterlockedTwoPhaseShutdown()
> #28 0x5611fe3b3fc3 TestingProfile::~TestingProfile()
> #29 0x5611fe3b43cc TestingProfile::~TestingProfile()
> ...
> 
> Original change's description:
> > Flip the AppServiceAsh feature default to enabled
> > 
> > BUG=826982
> > BUG=989789
> > 
> > Change-Id: I49bd3793bf9fcfc89726fd2f7df690421996e0fd
> > Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1728411
> > Reviewed-by: Dominick Ng <dominickn@chromium.org>
> > Commit-Queue: Nigel Tao <nigeltao@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#683064}
> 
> TBR=dominickn@chromium.org,nigeltao@chromium.org
> 
> Change-Id: Iec931954750b6a58d2274c6479507e99efd2467a
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Bug: 826982, 989789
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731091
> Reviewed-by: Takashi Sakamoto <tasak@google.com>
> Commit-Queue: Takashi Sakamoto <tasak@google.com>
> Cr-Commit-Position: refs/heads/master@{#683108}

TBR=tasak@google.com,dominickn@chromium.org,nigeltao@chromium.org

Change-Id: I0f8e2a335011012df144a02cf75fe943f6196171
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 826982, 989789
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731373Reviewed-by: Nigel Tao <nigeltao@chromium.org>
Reviewed-by: Dominick Ng <dominickn@chromium.org>
Commit-Queue: Nigel Tao <nigeltao@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683488}

This (PS1) passes all existing tests as-is but
ParkableStringTest.ReportTotalUnparkingTime.

This failure happens because samples of Now() within the same task
no longer have any delta under MOCK_TIME regardless of how CPU expensive
the operation actually is.

That test was using a state of the art approach, base::ElapsedTimer,
MOCK_TIME, base::HistogramTester, etc.

But base::ElaspedTimer is neutered by the change to MOCK_TIME in this CL.

There are two potential solutions:
  1) Have ElapsedTimer play a role in auto-advancing mock-time.
  2) Mock timings returned by ElpasedTimer.

(1) is undesirable because it would make it possible for delayed tasks
to run as part of a basic RunUntilIdle() which takes control of delayed
tasks away from the ScopedTaskEnvironment owner and is highly undesired.

(2) is more explicit but also easier to control. This CL does that by
adding base::ScopedMockElapsedTimersForTest.

This actually makes ParkableStringTest.ReportTotalUnparkingTime cleaner
because it can now test for a specific timing instead of just
"has a non-zero timing".

Bug: 905412
Change-Id: I91fab300b4005fd7802ce837733665e77766d4b2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1707870
Auto-Submit: Gabriel Charette <gab@chromium.org>
Commit-Queue: Kentaro Hara <haraken@chromium.org>
Reviewed-by: Kentaro Hara <haraken@chromium.org>
Reviewed-by: kylechar <kylechar@chromium.org>
Reviewed-by: Benoit L <lizeb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683487}

https://skia.googlesource.com/skia.git/+log/39dda0fe6aa0..0e04cdf70fe2

git log 39dda0fe6aa0..0e04cdf70fe2 --date=short --no-merges --format='%ad %ae %s'
2019-08-02 skia-autoroll@skia-public.iam.gserviceaccount.com Roll third_party/externals/swiftshader 27e2873765fa..f63c4e51101a (3 commits)

Created with:
  gclient setdep -r src/third_party/skia@0e04cdf70fe2

The AutoRoll server is located here: https://autoroll.skia.org/r/skia-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


CQ_INCLUDE_TRYBOTS=luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux-blink-rel;luci.chromium.try:linux-chromeos-compile-dbg;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel
TBR=benjaminwagner@google.com

Bug: chromium:None
Change-Id: I3b4376d8693c7bb183efaae8fb97a83df3bd1d1d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732654Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#683486}

There were some DCHECK(IsMainThread()) in
ServiceWorkerGlobalScopeProxy. When we move service worker
startup off the main thread, these DCHECKs will start
failing. Use |parent_task_runner_| instead.

Bug: 692909
Change-Id: I92c788184b9b3753bf6eb2eec5393471e70e439e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732669Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Commit-Queue: Kenichi Ishibashi <bashi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683485}

This CL adds metrics to record the types of schemes that users click
on and then how they proceed, whether that's opening an app with
persistence, without persistence, or clicking out of the dialog.

BUG=984840

Change-Id: Ie6de5f6c31812edeb403138f7bac947ce087337b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1718600
Commit-Queue: Melissa Zhang <melzhang@chromium.org>
Reviewed-by: David Jacobo <djacobo@chromium.org>
Reviewed-by: Dominick Ng <dominickn@chromium.org>
Reviewed-by: Ilya Sherman <isherman@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683484}

Move CustomTabSessionImpl to its own file. Also, factor out code that
can be used by other classes that create and manage an ARC Custom Tab.

BUG=b:64863368
TEST=Deploy Chrome to DUT and verify ARC Custom Tabs work correctly
R=djacobo@chromium.org, hashimoto@chromium.org, oshima@chromium.org

Change-Id: Ib11a233aa04ac52cfe73abaa9ecbb145791aa263
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1728446Reviewed-by: David Jacobo <djacobo@chromium.org>
Reviewed-by: Mitsuru Oshima (Slow - traveling) <oshima@chromium.org>
Reviewed-by: Ryo Hashimoto <hashimoto@chromium.org>
Commit-Queue: Jesse Schettler <jschettler@chromium.org>
Auto-Submit: Jesse Schettler <jschettler@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683483}

This CL adds the default positions of the Settings and OS Settings PWAs.

Bug: 989833
Change-Id: I32a5400b87a810d4e06137c57d44ec1b0360427f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1728512
Auto-Submit: calamity <calamity@chromium.org>
Reviewed-by: Ben Wells <benwells@chromium.org>
Reviewed-by: Giovanni Ortuño Urquidi <ortuno@chromium.org>
Commit-Queue: calamity <calamity@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683482}

This change renames TryToShowTouchToFill To ShouldSuppressKeyboard in
WebLocalFrame and WebAutofillClient. This is because Touch To Fill is a
password manager feature, which Blink doesn't need to know about.
ShouldSuppressKeyboard describes the actual behavior better.

Bug: 957532
Change-Id: I866145d0b63bb2242ab1c575299a95aac7009e44
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1728650Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Reviewed-by: Dominic Battré <battre@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Commit-Queue: Jan Wilken Dörrie <jdoerrie@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683481}

https://chromium.googlesource.com/chromiumos/chromite.git/+log/367a491d5bc1..9532972a16bb

git log 367a491d5bc1..9532972a16bb --date=short --no-merges --format='%ad %ae %s'
2019-08-02 briannorris@chromium.org repo_manifest: update manifest-format doc link

Created with:
  gclient setdep -r src/third_party/chromite@9532972a16bb

The AutoRoll server is located here: https://autoroll.skia.org/r/chromite-chromium-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


CQ_INCLUDE_TRYBOTS=luci.chromium.try:chromeos-kevin-rel
TBR=chrome-os-gardeners@google.com

Bug: None
Change-Id: I6bdfda7a997b6b523f1a6ee0e4ed0f92a2c1721b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732810Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#683480}

blink::MemoryCache can be accessed only from the main thread.

Before this CL WorkerGlobalScope used ParentExecutionContextTaskRunners
to post a task to remove an item from MemoryCache. This worked because
the task runners were always bound to the main thread. However, when
we start a service worker on the IO thread, there would be no
"parent execution context" and thus no
ParentExecutionContextTaskRunners. Using task runners which are
bound to the IO thread violates the assumption MemoryCache has.

This CL changes to use the main thread task runner instead of
ParentExecutionContextTaskRunners. Thread::GetTaskRunner() is
a deprecated method but it's hard not to use because alternatives
are bound to execution contexts and there is no parent execution
context for service workers.

Bug: 692909
Change-Id: Ieb9b97bcc06398b323c2a78d8bbc0755533580a1
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731832Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Commit-Queue: Kenichi Ishibashi <bashi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683479}

MemoryPressureListenerRegistry is a garbage collected object and was
instantiated lazily. This could be a problem when we start a service worker
on the IO thread because blink::scheduler::WorkerThread tries to get the
instance of the registry during worker start up. Since the IO thread runs
without Oilpan and thus can't create garbage collected objects,
MakeGarbageCollected() would fail when the instance hasn't been
instantiated.

This CL ensures the instantiation happens on the main thread. This moves
the instantiation earlier than before but it wouldn't affect overall
performance because it just creates a garbage collected object and
instantiation eventually happens as blink loads resources.

Bug: 988335
Change-Id: I3c2fa3f2be5bd1de913bd916b761548fbff4dda0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732111Reviewed-by: Kentaro Hara <haraken@chromium.org>
Commit-Queue: Kenichi Ishibashi <bashi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683478}

In order to deal with mojo message ordering uncertainty between
WebSocketHandshakeClient and WebSocketClient, this CL has
WebSocketHandleImpl::binding_as_client_ binds when
OnConnectionEstablished is called.

mojom.WebSocketClient.OnFailChannel is replaced with
InterfacePtr::ResetWithReason because |client_| may not be available
when network::WebSocket::WebSocketEventHandler::OnFailChannel is called.

Bug: 989406, 967524
Change-Id: Ic14357133a85317a07271a3615a21ff3e61fe37d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1728917Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Reviewed-by: Ken Rockot <rockot@google.com>
Reviewed-by: Adam Rice <ricea@chromium.org>
Reviewed-by: Yoichi Osato <yoichio@chromium.org>
Commit-Queue: Yutaka Hirano <yhirano@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683477}

4 consecutive bools + 1 enum + an unsigned overflow the 8-byte alignment
boundary by 1. Packing these bools into a bitfield saves 8 bytes which reduces
the size of TransformPaintPropertyNode from 96B to 88B on a 64-bit release
build.

Bug: 969531
Change-Id: I30879972bfd097162e92d79299f167dc2e57ea3d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1728036
Commit-Queue: Bartek Nowierski <bartekn@chromium.org>
Auto-Submit: Bartek Nowierski <bartekn@chromium.org>
Reviewed-by: Philip Rogers <pdr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683476}

It is incorrect to cache the value of ArcAppListPrefs::Get(profile). The
value returned can change over the lifetime of the profile.

Also move observer clean-up from the ArcApps destructor to its
KeyedService::Shutdown method, since the ArcApps keyed service depends
on the AALP keyed service.

This fixes the ArcAppLauncherForDefaultAppTest.* unit_tests in debug
mode, when combined with --enable-features=AppServiceAsh. The tests
passed in release mode, both before and after this commit.

In the unit_tests, ArcAppModelBuilderRecreate::StartArc calls
ArcAppListPrefsFactory::GetInstance()->RecreateServiceInstanceForTesting,
which destroys the existing ArcAppListPrefs and creates a new one.
Meanwhile, ArcApps code continues to follow the old pointer.

In the release build, there's the (un)happy coincidence that the new
AALP's pointer value is exactly the same as the old one: the memory is
re-used and that 'dangling' pointer still points to a valid AALP object.
In the debug build, malloc/free behavior is presumably different (the
pointer value is certainly different), and the latent bug surfaces.

BUG=826982

Change-Id: Ie77fb2ab8c65662d66faff8947901414d9cb1892
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732110
Commit-Queue: Nigel Tao <nigeltao@chromium.org>
Reviewed-by: Dominick Ng <dominickn@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683475}

These are no longer used after http://crrev.com/c/1731919

Bug: 824840
Change-Id: I7163336861c227fb7fa7aab9fe297898d92067c7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731180
Commit-Queue: Clark DuVall <cduvall@chromium.org>
Reviewed-by: John Abd-El-Malek <jam@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683474}

This change adds back method accessibilityAttributeNames that was
removed as part of a migration to NSAccessibility API. Despite being
deprecated, this is still called internally by AppKit, and the need for
this method was found because keyboard focus and VO focus were not in
sync anymore while navigating through the browser frame.

This issue is due to setAccessibilityFocused not being called because
accessibilityAttributeNames is the method in charge of returning the
attributes supported and one of them is NSAccessibilityFocusedAttribute.

The unit tests for setAccessibilityFocused are already handled in
ui\views\widget\ax_native_widget_mac_unittest.mm.

Original change: http://crrev.com/c/1508580

Bug: 990014

Change-Id: I505be15afae809690c5cea7599af96a7ce6ca691
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1728043
Commit-Queue: Kevin Babbitt <kbabbitt@microsoft.com>
Reviewed-by: Nektarios Paisios <nektar@chromium.org>
Reviewed-by: Elly Fong-Jones <ellyjones@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683473}

TBR=xhwang@chromium.org

Bug: None
Test: None
Change-Id: I215889e5ab8c25577c04e164a657c3b6f027a5f3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731638
Commit-Queue: Hirokazu Honda <hiroh@chromium.org>
Reviewed-by: Pawel Osciak <posciak@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683472}

https://chromium.googlesource.com/chromiumos/chromite.git/+log/3bd52c2334a0..367a491d5bc1

git log 3bd52c2334a0..367a491d5bc1 --date=short --no-merges --format='%ad %ae %s'
2019-08-02 shaochuan@chromium.org chromite: Temporarily move betty-arcnext-pi-android-pfq to experimental

Created with:
  gclient setdep -r src/third_party/chromite@367a491d5bc1

The AutoRoll server is located here: https://autoroll.skia.org/r/chromite-chromium-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


CQ_INCLUDE_TRYBOTS=luci.chromium.try:chromeos-kevin-rel
TBR=chrome-os-gardeners@google.com

Bug: None
Change-Id: Id13ee4344073abcb452263d6ee46f960feb3d3d7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732631Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#683471}

Currently, ImagePaintTimingDetector and TextPaintTimingDetector each
has their own paint-time callback. Two callbacks are unnecessary since
if they are registered from the same (animated) frame, they would get
the same paint-time. Therefore, they should be combined into one
callback.

Combining the callbacks also creates a chance for LCPCalculator to
hook to "either text records or image records have been changed",
i.e., the end of the combined callback. This hook would enable us to
remove the intermediate state between two callbacks, and thus fixed
the out-of-sync issue in crbug.com/982307.

Bug: 982307

Change-Id: Ic9bc3076d163b471d91e6414da70bd6d7af0fd6d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1730517
Commit-Queue: Liquan (Max) Gu <maxlg@chromium.org>
Reviewed-by: Steve Kobes <skobes@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683470}

This reverts commit b0caaa35.

Reason for revert: Suspect introduces flake (https://crbug.com/990225).

LayoutSubtreeRootList::RandomRoot is crashing in about 50% of builds since this landed.

Original change's description:
> Call PaintLayer::UpdateLayerPositionsAfterLayout only from layout roots.
> 
> Previously, it always called it from the LayoutView, regardless of what
> layout changed.
> 
> This will be more efficient in typical cases of non-root layouts. It
> may be more inefficient in the following cases
> a. Two layout roots are present, of which one is an ancestor of the other
> (in this case, the new code will update the overlapping layout tree
> portions twice).
> b. Cases when LocalFrameview::UpdateLayout recurses, due to NeedsLayout
> being true after the first one.
> 
> If we consider these cases sufficiently rare, then this CL is ok as-is.
> Handling (a) and (b) is possible but is harder to implement.
> 
> Bug: 462719,970224
> Change-Id: Ie65dc73a70bb9362dfa20ff9ad4070e3945d3c56
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1728717
> Reviewed-by: Stefan Zager <szager@chromium.org>
> Commit-Queue: Chris Harrelson <chrishtr@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#683409}

TBR=szager@chromium.org,chrishtr@chromium.org,eae@chromium.org

Change-Id: I50a9d9861dae814d0c564e072df8a1c66a9b5898
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 462719, 970224, 990225
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732809Reviewed-by: Matt Giuca <mgiuca@chromium.org>
Commit-Queue: Matt Giuca <mgiuca@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683469}

Delegate CanReparentAppTabToWindow/ReparentAppTabToWindow to
WebAppUiManager since this is UI-related business.

In next CLs:
We will get rid of chrome/browser/extensions/bookmark_app_extension_util.cc
completely. That's the last bookmark_app* prefixed file in extensions/.

TBR=ortuno@chromium.org

Bug: 901226
Change-Id: I1fb9658d0671d9ca9509280625911371aed95fd1
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732670Reviewed-by: Alexey Baskakov <loyso@chromium.org>
Reviewed-by: Eric Willigers <ericwilligers@chromium.org>
Commit-Queue: Alexey Baskakov <loyso@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683468}

This CL relands the following change:
https://chromium-review.googlesource.com/c/chromium/src/+/1718064

Which was reverted by this other change:
https://chromium-review.googlesource.com/c/chromium/src/+/1728130

Reason for revert: suspect causing content_browsertests failure on Windows bots:
E.g. https://ci.chromium.org/p/chromium/builders/ci/Win10%20Tests%20x64/39015

Fixing issues with `ComPtr` being incorrectly handled as return values.

Original change's description:
> Add trailing whitespace and punctuation to UIA Move(TextUnit_Word)
>
> 1. Changing the behavior of `ITextRangeProvider` methods `Move` and
>    `ExpandToEnclosingUnit` with `TextUnit_Word`.
>
>    Previously we relied on both, word start and end offsets, computed
>    from Blink, but such approach didn't matched every AT's expectations.
>
>    Updated logic makes use of start offsets only, this forces movement
>    by word to include whitespace and punctuation characters before the
>    next word start; the only special case being the last word, whose end
>    offset will be located after the last character on the document.
>
> 2. Removing `AXPosition`'s methods `AsLeafTextPositionBeforeCharacter`
>    and `AsLeafTextPositionAfterCharacter` in favor of refactoring
>    `AsPositionBeforeCharacter` and `AsPositionAfterCharacter`.
>
>    This is a follow-up optimization since the former methods called
>    their `AsLeafTextPosition` equivalents, but reverted the result if
>    the position was not adjusted to another anchor.
>
> 3. Adding a new browser test to cover some AT's expectations.
>
>    Rebaselining `AXPlatformNodeTextRangeProviderWin` unit and browser
>    tests to account for the new word movement behavior.
>
> Bug: 928948
> Change-Id: Iab689fde2a2ea973e86f7cdb10c1c22cf1af1704
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1718064
> Commit-Queue: Ethan Jimenez <ethavar@microsoft.com>
> Reviewed-by: Kurt Catti-Schmidt <kschmi@microsoft.com>
> Reviewed-by: Nektarios Paisios <nektar@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#682564}

Bug: 928948
Change-Id: I62022f54fb236d3ffe638b6ef17648891da4aef9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731750Reviewed-by: Kurt Catti-Schmidt <kschmi@microsoft.com>
Reviewed-by: Nektarios Paisios <nektar@chromium.org>
Commit-Queue: Kevin Babbitt <kbabbitt@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#683467}

https://skia.googlesource.com/skia.git/+log/c711a8649301..39dda0fe6aa0

git log c711a8649301..39dda0fe6aa0 --date=short --no-merges --format='%ad %ae %s'
2019-08-02 skia-autoroll@skia-public.iam.gserviceaccount.com Roll third_party/externals/angle2 1ebe6014389e..1dbbc7b3e232 (5 commits)
2019-08-01 recipe-mega-autoroller@chops-service-accounts.iam.gserviceaccount.com Roll recipe dependencies (trivial).
2019-08-01 fmalita@chromium.org Reland: SkColorfilters::HSLAMatrix
2019-08-01 mtklein@google.com bigger channel buffers in fm_bot
2019-08-01 bsalomon@google.com Renove GrPixelConfig from GrGLGpu::uploadTexData
2019-08-01 mtklein@google.com sketch SkVMBlitter
2019-08-01 reed@google.com remove unneeded bitmapprovider abstraction
2019-08-01 mtklein@google.com Revert "let color filters tell us their alpha format"
2019-08-01 mtklein@google.com Revert "SkColorfilters::HSLAMatrix"
2019-08-01 bsalomon@google.com Make read/write/transfer pixels functions take surface color type
2019-08-01 mtklein@google.com make presubmit failures clearer
2019-08-01 robertphillips@google.com Add a means of printing out the GrBackendFormat
2019-08-01 nifong@google.com Add a kind of asset for MSKPs so the skpbench bot can read them.
2019-08-01 robertphillips@google.com Make GrGLCaps::getYUVAColorTypeFromBackendFormat also map R8 to GrColorType::kGray_8
2019-08-01 bsalomon@google.com Use GrConvertPixels in GrOpFlushState::doUpload
2019-08-01 stani@google.com Allow to rebind GL texture if AHB content has changed
2019-08-01 fmalita@chromium.org SkColorfilters::HSLAMatrix
2019-08-01 reed@google.com make bicubic decision logic shareable
2019-08-01 egdaniel@google.com Remove use of isConfigRenderable.
2019-08-01 mtklein@google.com let color filters tell us their alpha format

Created with:
  gclient setdep -r src/third_party/skia@39dda0fe6aa0

The AutoRoll server is located here: https://autoroll.skia.org/r/skia-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


CQ_INCLUDE_TRYBOTS=luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux-blink-rel;luci.chromium.try:linux-chromeos-compile-dbg;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel
TBR=benjaminwagner@google.com

Bug: chromium:None,chromium:b/138674291
Change-Id: Idcd7aab9fc79b8c0e6c1f479880356b0ec794363
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732630Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#683466}

So far we were forcibly setting the number of output planes to 1, but
the Kukui decoder uses a multiplanar format.

The V4L2 driver will update the number of planes if it is incorrect, so
use this information when attempting to set the format to get the
correct number of planes. This is what the v4l2vda already does.

Bug: b:132589320
Test: vdaunittest and vdatests pass on Minnie.

Change-Id: Idb4170875adf1df2fdbd1cbe65c233937aed0702
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1730732
Commit-Queue: Alexandre Courbot <acourbot@chromium.org>
Reviewed-by: Chih-Yu Huang <akahuang@chromium.org>
Reviewed-by: Hirokazu Honda <hiroh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683465}

Change-Id: Iecab9924c9035887e946ffa994fcb757b00d054e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731785
Auto-Submit: Eric Stevenson <estevenson@chromium.org>
Reviewed-by: Samuel Huang <huangs@chromium.org>
Commit-Queue: Samuel Huang <huangs@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683464}

This reverts commit b302b4c7.

Reason for revert: Introduces test flake (https://crbug.com/990207)

Original change's description:
> [Background Sync] Show updated delay in Devtools.
> 
> We sometimes change the actual delay after which a periodicsync event is fired
> based on other registrations for the origin. Update what we show in Devtools so
> it corresponds to the actual delay_until set on the registration.
> 
> Bug: 989524
> Change-Id: I9ba06f9c5f65d93775b0a9bf9d934f1d9b5b5ce6
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1730921
> Reviewed-by: Rayan Kanso <rayankans@chromium.org>
> Commit-Queue: Mugdha Lakhani <nator@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#683298}

TBR=nator@chromium.org,rayankans@chromium.org

Change-Id: If6c1b97d518011b5960e4a095ad6f2a67b3907f3
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 989524, 990207
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731338Reviewed-by: Matt Giuca <mgiuca@chromium.org>
Commit-Queue: Matt Giuca <mgiuca@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683463}

The AutoRoll server is located here: https://autoroll.skia.org/r/fuchsia-sdk-chromium-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.

CQ_INCLUDE_TRYBOTS=luci.chromium.try:fuchsia-arm64-cast;luci.chromium.try:fuchsia-x64-cast
TBR=cr-fuchsia+bot@chromium.org

Change-Id: I288e91ac918d3bba5e6538b57360a4fba19f5c4a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732629Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#683462}

Bug: 990026
Change-Id: Iacc19b3449ac927e2c98ad8a16b4b5f6c793e302
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731912Reviewed-by: John Abd-El-Malek <jam@chromium.org>
Commit-Queue: Henrique Nakashima <hnakashima@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683461}

The Launch Service is used to open a new application
window.

This is a rebase of
https://chromium-review.googlesource.com/c/chromium/src/+/1723794
that now also updates the OpenApplication call that was recently added in
https://chromium-review.googlesource.com/c/chromium/src/+/1674709
(Without that call updated, linux-chromeos-chrome faied to compile during a roll.)

TBR=avi@chromium.org

Bug: 966288, 989897
Change-Id: I3a51b6ca61a26508e1bad38a3df41170e98811e6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731144
Commit-Queue: Eric Willigers <ericwilligers@chromium.org>
Reviewed-by: Alexey Baskakov <loyso@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683460}

# This is an automated release commit.
# Do not revert without consulting chrome-pmo@google.com.
NOAUTOREVERT=true
TBR=kariah@chromium.org

Change-Id: I99c611eb02d9ddca7c70b3f4fc6fb7fb78c3e9c0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1732729Reviewed-by: Chrome Release Bot (LUCI) <chrome-official-brancher@chops-service-accounts.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#683459}

Next CL will remove references in content/

Bug: 824840
Change-Id: Ia002662d91d397b26cd2877dd4da30b6343d2e80
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731919Reviewed-by: John Abd-El-Malek <jam@chromium.org>
Commit-Queue: Clark DuVall <cduvall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683458}

This reverts commit 55bc919a.

Reason for revert: Suspected cause of https://crbug.com/990192

Since this changes the ownership of RenderWidget and the newly flaky
test (which started failing on MSAN 1 build after this landed) has a
use-after-free in RenderWidget, I am suspecting this CL.

Original change's description:
> Make browser process own RenderView.
> 
> Historically, RenderView and RenderWidget were 1:1, and their lifetimes were
> entwined.
> 1) RenderViewHost would create the RenderView.
> 2) RenderView would create a RenderWidget and pass ownership of itself to the
> RenderWidget.
> 3) RenderViewHost's destructor would destroy the RenderWidget, thus destroying
> the RenderView
> 
> We want the lifetime of RenderView and RenderWidget to be decoupled. The first
> step of this is making RenderView explicitly owned by the browser. This means
> that instead of (3), RenderViewHost's destructor will destroy the RenderView,
> which will in turn destroy the RenderWidget.
> 
> One subtlety is that prior to this CL, RenderWidget was always destroyed
> asynchronously. The original reason for supporting this -- dealing with
> re-entrancy from nested message loops -- is no longer applicable. The IPC that
> destroys RenderWidget is asynchronous, which means it can never be called from a
> re-entrant context. However, it is possible for a RenderWidget associated with a
> child-frame to be synchronously destroyed by JS. This can be re-entrant. This CL
> updates destruction of RenderWidget to be synchronous when called from IPC.
> 
> Bug: 987731
> Change-Id: If4b319fab19d02c5495ba14e5cc929f441ca4d2e
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1717456
> Commit-Queue: Avi Drissman <avi@chromium.org>
> Reviewed-by: Avi Drissman <avi@chromium.org>
> Reviewed-by: Daniel Cheng <dcheng@chromium.org>
> Auto-Submit: Erik Chen <erikchen@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#683247}

TBR=avi@chromium.org,dcheng@chromium.org,erikchen@chromium.org

Change-Id: I762230d0df57c6bcd75af80c1aa231d7b4d2183e
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 987731, 990192
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731991Reviewed-by: Matt Giuca <mgiuca@chromium.org>
Commit-Queue: Matt Giuca <mgiuca@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683457}

R=dpapad@chromium.org
BUG=953962

Change-Id: I1b74794585cb6a53ad5f5278a4b4262c9019f280
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1730295
Commit-Queue: Dan Beam <dbeam@chromium.org>
Reviewed-by: Demetrios Papadopoulos <dpapad@chromium.org>
Auto-Submit: Dan Beam <dbeam@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683456}

BUG=988321

Change-Id: I5cf9fff36c7cdb9aaf777172f3169035e0a05d3a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1730413Reviewed-by: Nigel Tao <nigeltao@chromium.org>
Commit-Queue: Nancy Wang <nancylingwang@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683455}

This is CL is generated by a script. Data is pulled from
review conducted in Q2 2019. See https://tinyurl.com/yxau5sc4
for context.

NOPRESUBMIT=true
Set due to authors in OWNERS files having issues with LGTM. They will be fixing this separately.

Bug: 977050
Change-Id: I997300c3bf79c789d5270ea3c0c47197b2d045b5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1725159Reviewed-by: Eric Foo <efoo@chromium.org>
Commit-Queue: Eric Foo <efoo@chromium.org>
Auto-Submit: Eric Foo <efoo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683454}

Bug: 934009
Change-Id: I396ea73e2d39bea85a93e07368cac6226680d146
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1731784
Auto-Submit: John Abd-El-Malek <jam@chromium.org>
Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Commit-Queue: John Abd-El-Malek <jam@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683453}

Some ozone platforms (e.g. Wayland) require mojo communication while
others either don't use mojo or use it only if OzoneDrmMojo feature is
enabled. |requires_mojo| shows whether a platform requires mojo or not.

|requires_mojo| is used in some places before OzonePlatform is
initialized.

Part of this usage is for populating InitParams used for OzonePlatform
initialization. These can easily be removed as each type of
OzonePlatform knows whether it requires mojo or not when it is being
initialized. There is no need to pass this value to the initialization
codes.

Other cases use this and other info to check whether mojo is actually
used or not. None of these cases really need to happen before
initialization and can be moved to after initialization.

This change is part of a larger effort to remove the need for
OzonePlatform instance before it is initialized.

This CL also cleans up some code in OzonePlatform implementations,
mostly using std::make_unique<> instead of new operator.

BUG=958387

Change-Id: Ie44c6e441389e1f119f6458bc4e1c15bdc1d83e2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1604112Reviewed-by: Robert Kroeger <rjkroege@chromium.org>
Reviewed-by: Jonathan Backer <backer@chromium.org>
Reviewed-by: Scott Violet <sky@chromium.org>
Commit-Queue: Mohsen Izadi <mohsen@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683452}

This CL renames mojo to mojom in //fuchsia for consistency
with the rest of the chromium.

fuchsia/mojo => fuchsia/mojom

Bug: 806965
Change-Id: I1f36512ddb418a5a644c546f88fc51ca47cf1a67
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1722168Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Reviewed-by: Ken Rockot <rockot@google.com>
Reviewed-by: Sergey Ulanov <sergeyu@chromium.org>
Commit-Queue: Miyoung Shin <myid.shin@igalia.com>
Cr-Commit-Position: refs/heads/master@{#683451}

This effectively reverts commit
7e8de286.

Reason for revert: Testing whether this change was responsible for
the latency regression seen in crbug.com/959048.

This isn't a true revert, as that CL no longer reverts cleanly, but this
restores our previous behavior.

Bug: 935630, 959048
Change-Id: Iefb4bd614b5802cfa24a8cb459e62842cfa6d4e8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1698629
Auto-Submit: Eric Karl <ericrk@chromium.org>
Reviewed-by: Sadrul Chowdhury <sadrul@chromium.org>
Commit-Queue: Eric Karl <ericrk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#683450}