GitLab

Add @RetryOnFailure to
IncognitoStorageLeakageTest#testStorageDoesNotLeakFromActivityToActivity

Although there are talk about removing the RetyOnFailure flag (see
crbug.com/797002) for now this should be slightly less expensive than
re-running the entire test suite.

TBR: tedchoch@chromium.org
Bug: 1087796
Change-Id: I2f76509aa7f5aecf9f7554cc3d825149055b428d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222727
Commit-Queue: Thomas Guilbert <tguilbert@chromium.org>
Reviewed-by: Thomas Guilbert <tguilbert@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773362}

These histograms are still in use.

Bug: 1072542
Change-Id: I7524ba1fc7f233ca84171ffdb4c11c58c767a8ab
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222630Reviewed-by: Alexander Timin <altimin@chromium.org>
Commit-Queue: Katie Dillon <kdillon@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773361}

Same thing that crrev.com/c/2219679 did, but for tests invoked via
host-side cmd (eg: telemetry) rather than device-side cmd (eg: gtests).

Bug: 1087407
Change-Id: I514a01583bbfea6a4a600d738587d8663ae0adec
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222901Reviewed-by: Brian Sheedy <bsheedy@chromium.org>
Commit-Queue: Ben Pastene <bpastene@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773360}

Summary of changes available at:
https://chromium.googlesource.com/v8/v8/+log/7693b668..772196e5

Please follow these instructions for assigning/CC'ing issues:
https://v8.dev/docs/triage-issues

Please close rolling in case of a roll revert:
https://v8-roll.appspot.com/
This only works with a Google account.

CQ_INCLUDE_TRYBOTS=luci.chromium.try:linux-blink-rel
CQ_INCLUDE_TRYBOTS=luci.chromium.try:linux_optional_gpu_tests_rel
CQ_INCLUDE_TRYBOTS=luci.chromium.try:mac_optional_gpu_tests_rel
CQ_INCLUDE_TRYBOTS=luci.chromium.try:win_optional_gpu_tests_rel
CQ_INCLUDE_TRYBOTS=luci.chromium.try:android_optional_gpu_tests_rel

TBR=hablich@chromium.org,vahl@chromium.org,v8-waterfall-sheriff@grotations.appspotmail.com

Change-Id: I13bec7b00a9e05d3e8cba4fb72c105b90d82a9c6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222207Reviewed-by: v8-ci-autoroll-builder <v8-ci-autoroll-builder@chops-service-accounts.iam.gserviceaccount.com>
Commit-Queue: v8-ci-autoroll-builder <v8-ci-autoroll-builder@chops-service-accounts.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#773359}

Instead of hardcoding a top and bottom border in the graph, add a border
around each node's bounding area in boundingForce. This will keep nodes
away from the lane markers as well as the graph edges.

R=siggi

Bug: 1086231
Change-Id: I58b52649aec0e921de5a85d8bd191591bf36f13b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222529Reviewed-by: Sigurður Ásgeirsson <siggi@chromium.org>
Commit-Queue: Sigurður Ásgeirsson <siggi@chromium.org>
Auto-Submit: Joe Mason <joenotcharles@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773358}

In-memory representations of the two types of graphs were made
serializable/deserializable to and from JSON. A new module
`serialization` has been created to help with this process.

The end result is one .txt file containing the JSON dump of both the
class and package dependency graphs - more rationale on this format is
included in the docstring of serialization functions.

The performance of the de/serialization is very acceptable (within
a second on my machine). The filesize of the output JSON file is ~7MB,
which I also find acceptable. There are rooms for optimization in the
JSON structure if the filesize becomes an issue in the future.

Bug: 1081889
Change-Id: I03061372e0ed94aed62de58f41fb3f363f543b3f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2219817
Commit-Queue: James Long <yjlong@google.com>
Reviewed-by: Henrique Nakashima <hnakashima@chromium.org>
Reviewed-by: Mohamed Heikal <mheikal@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773357}

This CL changes SetWakeUpRate(double wake_ups_per_second) to
SetWakeInterval(base::TimeDelta interval) to avoid conversions
between rate and interval.

This CL is a no-op.

Bug: 1075553
Change-Id: I1b72f6b56f364c24c73a14eb8f1c3aea8ca31bc3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2219524Reviewed-by: Alexander Timin <altimin@chromium.org>
Commit-Queue: François Doray <fdoray@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773356}

https://chromium.googlesource.com/openscreen.git/+log/b4e53be8a002..cfe3be0ee4c5

2020-05-29 rwkeane@google.com Code Coverage: Add new buildbot mixin

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/openscreen-chromium
Please CC jophba@chromium.org on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

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

Bug: None
Tbr: jophba@chromium.org
Change-Id: I40c444d57a8ab9729262d99c327c7a66f52c077a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222712Reviewed-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@{#773355}

This CL is suspending the execution of delayed task while the
system is suspended. The ThreadControllerPowerMonitor is
used to notify the ThreadController when the power is back to
normal.

The delayed tasks can no longer be executed between suspend and
resume notifications. They are postponed after the resume notification.

On Windows, while going to suspend mode or waking up, the computer has
bottleneck on clock API (system-wide lock) and has a lot of disk IO
contention. It is not the right moment to wake up all the pending
delayed tasks at once. It is better waiting for the resume notification
while the computer is in a running state.

Bug: 1074332
Change-Id: Icda16b76fc87a20fee22997baf7eeb5e5d02eb09
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2186200
Commit-Queue: Etienne Bergeron <etienneb@chromium.org>
Reviewed-by: François Doray <fdoray@chromium.org>
Reviewed-by: Sami Kyöstilä <skyostil@chromium.org>
Reviewed-by: Gabriel Charette <gab@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773354}

To allow construction with multiple unwinder. This is necessary on android
to provide native and chrome unwinders and will be used in a follow in
stack_sampling_profiler_unittest.cc
https://chromium-review.googlesource.com/c/chromium/src/+/2201658

Bug: 989102
Change-Id: I05ef97403668d82dc7e6e4e9505be95e941174dd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2211862
Commit-Queue: Etienne Pierre-Doray <etiennep@chromium.org>
Reviewed-by: Mike Wittman <wittman@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773353}

WebLayer doesn't use ApplicationStatus. In order for WebLayer to use
AccessibilityUtil this splits AccessibilityUtil into two parts. The
core part dealing with determining if accessibility is enabled remains
in AccessibilityUtil. The application state remains in a subclass called
ChromeAccessibilityUtil.

This code is easier to split when the there is an actual instance
created. So, rather than a bunch of static methods, you have to
create an instance of AccessibilityUtil. ChromeAccessibilityUtil takes
care of this. Doing this means all chrome code now has to call a static
function to get the instance. This necessitated a bunch of changes.

There is one other interesting change here. Previously
ApplicationStatus.ActivityStateListener, when RESUME was called, would
reset state so that the next call to isAccessibilityEnabled() would
trigger recalculating the value. If the value changed, this would *not*
notify observers. This seemed like a bug, so I made the RESUME case
immediately calculate the value and notify observers if necessary. Please
let me know if you think the old code was correct.

BUG=1085210
TEST=none

Change-Id: Id4e58da4f96fb9d218e59d3e62a94f0085827bb7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2210358
Commit-Queue: Scott Violet <sky@chromium.org>
Reviewed-by: Jinsuk Kim <jinsukkim@chromium.org>
Reviewed-by: Theresa  <twellington@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773352}

R=thestig
CC=​thakis,maskray@google.com

Bug: 1086636, 1084359
Change-Id: Ibc5f5da51278d90d7f45c4fdadff07f62f6dfd51
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2219606
Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
Auto-Submit: Thomas Anderson <thomasanderson@chromium.org>
Reviewed-by: Lei Zhang <thestig@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773351}

When there's an Expect-CT failure, a new URLRequest will be created to
report the failure, and that request needs to have a NetworkIsolationKey
to prevent cross-site information leakage.

Bug: 969893, 1082280
Change-Id: I5daa400fed87ce23f7a9b34121888f91b9249895
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2219091
Commit-Queue: Matt Menke <mmenke@chromium.org>
Reviewed-by: Emily Stark <estark@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773350}

Change-Id: I20c0ea03bf3fd21a260dd9f1fc67e66b6d8c7c44
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2216546Reviewed-by: Renjie Tang <renjietang@chromium.org>
Reviewed-by: Nick Harper <nharper@chromium.org>
Commit-Queue: Nick Harper <nharper@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773349}

Change-Id: I2f05f6eaa5ea7bcbaa27e0a8f4472affb42afa5e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2168768
Commit-Queue: Piotr Bialecki <bialpio@chromium.org>
Reviewed-by: Alexander Cooper <alcooper@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773348}

If a history navigation targets a subframe, but the subframe has been
removed by its parent, we currently fall back to reload the whole
page.  This CL tries to modify this behavior to instead do a no-op
same-document navigation (to the same item sequence number) to the
main frame.  This approach will update the current history entry to
the target one, but it avoids a reload and allows the main frame and
other frames to continue to function.  The CL tries to preserve the
main frame reload behavior in the cases where that's still needed,
including various kinds of reloads, session restore, and crash
recovery, by refactoring FindFramesToNavigate() to explicitly queue a
different-document load in those cases.

We've also attempted ignoring the navigation completely, rather than
doing a no-op same-document navigation.  This had the problem of not
notifying observers about the updated history entry, meaning that
things like greying out of back/forward buttons didn't work properly,
and history length/offset information wasn't updated properly in the
renderer processes.  Doing a redundant same-document navigation avoids
these problems, as it goes through a complete navigation flow.

There was one unwanted side effect of this approach, which is that the
main frame would redundantly fire popstate events when the ISNs aren't
changing.  The spec [1] argues that popstate events should be fired
whenever the "entry" changes, but "entry" in that context probably
means the FrameNavigationEntry and not NavigationEntry.  Although the
NavigationEntry changes here, the spec probably expects
FrameNavigationEntry to not change (due to how joint session history
entries are supposed to work), even though it does change in our
current implementation (see https://crbug.com/373041).  Other browsers
also don't trigger a redundant popstate in this case.  So this CL also
modifies DocumentLoader to avoid triggering redundant popstate events
when the ISN doesn't change.

This is a less risky option to solve this problem.  It has some
drawbacks -- if there are multiple history entries targeting a removed
subframe, the user might have to keep clicking back/forward until
finding an entry that still works.  In the future, we could consider
improving this to automatically walk past all history entries
corresponding to the affected subframe.

[1] https://html.spec.whatwg.org/multipage/browsing-the-web.html#history-traversal

Bug: 705550
Change-Id: I11fca7a8f1062af347442922f49b2dfe50970d13
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2181973Reviewed-by: Nate Chapin <japhet@chromium.org>
Reviewed-by: Charlie Reis <creis@chromium.org>
Commit-Queue: Alex Moshchuk <alexmos@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773347}

This CL creates a whole new flag for the On-Content variant of
Contextual Web ZeroSuggest.

Bug: 1087700
Change-Id: I4ccb748421470e2ccd3508d69fc317e12c27ff9e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2219597Reviewed-by: Justin Donnelly <jdonnelly@chromium.org>
Commit-Queue: Tommy Li <tommycli@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773346}

It then passes the NetworkIsolationKey on to the ExpectCTReporter. Also
plumb NIK through the other TransportSecurityState methods that call
it.  Only one caller of TransportSecurityState is updated to pass in
a non-empty NIK.  The other 3 callers will be addressed in followup CLs.

Bug: 969893, 1082280
Change-Id: I2982c06288e6ff9ec2dd863b0ecbcf01f46b2791
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2217871Reviewed-by: Emily Stark <estark@chromium.org>
Commit-Queue: Matt Menke <mmenke@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773345}

This avoid having to type 1 << (numbits_ - 1) in several places which
can make code harder to read and can be typo prone. This function can
also be used within another template.

Change-Id: I46af3ab832bb58a6565702a391db028b2a69807f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222526
Commit-Queue: Oliver Li <olivierli@chromium.org>
Reviewed-by: François Doray <fdoray@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773344}

This ensures the most visited tiles as up-to-date. It's also what the
local NTP does.

Bug: 1087564
Change-Id: I724b99763952894617400d2733ce55992e3933dd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222635
Auto-Submit: Tibor Goldschwendt <tiborg@chromium.org>
Reviewed-by: Moe Ahmadi <mahmadi@chromium.org>
Commit-Queue: Tibor Goldschwendt <tiborg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773343}

SuperSize-archive calls _AddUnattributedSectionSymbols() to create
symbols (in .other section) to account for sections that are not covered
by symbols. However, previously this was done only ELF files inside APK
and not for stand-alone ELF files. Also, the accounting was off,
resulting in Supersize-console
  Print(size_info)
producing file sizes that are off. This CL fixes the above problems.
Details:
* Add models.ClassifySections() to classify sections into one of three
  groups:
  (1) Unsummed sections: These don't add to binary size. E.g., .bss.
  (2) Explicitly summed sections: These add to binary size, and have
      symbols that are extracted by the "usual method" via .map file
      parsing, nm, etc. E.g., .text.
  (3) Implicitly summed sections: These add to binary size, but needs
      to be extracted via _AddUnattributedSectionSymbols(), and counted
      as part of the .other section. E.g., .dynsym.
  * The function is used by archive.py and describe.py, to eliminate
    previous size mismatches.
  * Add BaseSizeInfo.ClassifySection() to wrap this.
* archive.py: Change main flow so that _AddUnattributedSectionSymbols()
  get called when an ELF file is the main input (even without APK).
* describe.py: When describing section sizes, add note to make it clear
  that a section is counted in the .other section.

This CL leads to extensive .golden file changes.

Bug: 1084580
Change-Id: If663eb9acce89b32f857a0056b6f5ec39db35d8e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2217861Reviewed-by: Samuel Huang <huangs@chromium.org>
Reviewed-by: Andrew Grieve <agrieve@chromium.org>
Commit-Queue: Samuel Huang <huangs@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773342}

https://skia.googlesource.com/skia.git/+log/df546db4d8f7..cc7ec24ca824

2020-05-29 egdaniel@google.com Move backend surface vk getters and setters back out from behind compile flag.
2020-05-29 bungeman@google.com Revert "Hide ICU C++ API from Skia users."
2020-05-29 bungeman@google.com Create CTTypesetter without Options.
2020-05-29 herb@google.com tease apart GrTextBlob and SubRun
2020-05-29 bungeman@google.com Hide ICU C++ API from Skia users.
2020-05-29 herb@google.com Rename fUniqueID -> fMessageBoxID
2020-05-29 csmartdalton@google.com Wire up mouse and keyboard events in CanvasKit viewer
2020-05-29 egdaniel@google.com Add GrBackendMutableState object to handle shared texture state.
2020-05-29 michaelludwig@google.com Rename fIH variable to fInvH
2020-05-29 csmartdalton@google.com Merge PathText samples into one, with different modes
2020-05-29 nigeltao@google.com Fix markdown typo in the style guide
2020-05-29 senorblanco@chromium.org Modify SurfaceTest to only test readbacks from supported surfaces.
2020-05-29 brianosman@google.com Minor cleanup in the sample_matrix GMs
2020-05-29 robertphillips@google.com Update DDL test harness to better match OOP-R
2020-05-29 nifong@google.com Record code coverage in canvaskit tests, increase coverage
2020-05-29 herb@google.com Move hasW to SubRun

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/skia-autoroll
Please CC westont@google.com on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

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

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
Bug: None
Tbr: westont@google.com
Change-Id: I5ced804b0b88c5c04788f5e10a879779ef17990f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222710Reviewed-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@{#773341}

CL:2035212 was reverted because we encountered some issue with the
guest user, and VP8 on Intel device. Now those issues are solved.
It's time to enable ChromeosVideoDecoder by default again.

Bug: 984482
Test: ninja media/base
Change-Id: I7ad037186fa4f62a0f7d14025e5f4590c7caad53
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2214764
Commit-Queue: Andres Calderon Jaramillo <andrescj@chromium.org>
Reviewed-by: Becca Hughes <beccahughes@chromium.org>
Reviewed-by: Andres Calderon Jaramillo <andrescj@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773340}

This CL enables possibility to isolate blink implementation of
screen enumeration from the rest of the stack.

It provides the possibility to create mock displays at test conveniences
by rerouting the GetDisplay call to the JS implementation in the test.

Existing test has been modified to use this test framework.

Bug: 1075656
Change-Id: I1a38426deaccff14036f00806ef7dfc878bcd4ce
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2214538
Commit-Queue: Arnaud Mandy <arnaud.mandy@intel.com>
Reviewed-by: Michael Wasserman <msw@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773339}

Previously there were tests that play audio and tests that play video,
but no tests were playing a file with both audio and video. Also video
was not played to the end. Added PlayVideo test that plays
bear-vp9-opus.webm to the end. Also updated SetBlockMediaLoading tests
to share play_video.html and reduce amount of duplicated code.

Bug: b/153914562
Change-Id: Iba492ec252f692f7aa3f93c319d858a91bb38278
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2219404
Auto-Submit: Sergey Ulanov <sergeyu@chromium.org>
Reviewed-by: Kevin Marshall <kmarshall@chromium.org>
Commit-Queue: Kevin Marshall <kmarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773338}

This patch removes logic that would prevent the bottom sheet from
showing when contextual search was visible. The code in question was
relevant when the bottom sheet could trigger without a user action.
Since this is no longer the case, the code is unneeded. This block of
code should have been removed with https://crrev.com/c/1808180, which
contained the rest of this logic.

Bug: 100227
Change-Id: I6c1896c9c8907ec01542568162291e85aa2892cd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2204298Reviewed-by: Theresa  <twellington@chromium.org>
Commit-Queue: Matthew Jones <mdjones@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773337}

This CL provides an extra layer of protection to top domains with
top 3 most visited TLDs. If a potential target embedding domain embeds
a hostname with another valid TLD instead of the domain's actual TLD,
the heuristic will flag it. For example, the heuristic will trigger on
google.info-foo.com.

This CL also flags domains that embed a target without the target's
separators (e.g. googlecom-evil.com).

Bug: 1087636
Change-Id: I14205e5b9830146d8aecfd216082d1213e5622bd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2215377
Commit-Queue: Behnood Momenzadeh <behnoodm@google.com>
Reviewed-by: Joe DeBlasio <jdeblasio@chromium.org>
Reviewed-by: Mustafa Emre Acer <meacer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773336}

Starting in M83, WebXr uses the permissions api. As such, the custom
consent flow that it initially shipped with is no longer needed and can
now be removed.

Since the change to permissions is launched via finch for M83 (and not
enabled by default in code until 84), the feature itself (and the finch
configuration) will be left until M86; though it will not change any
behavior. The feature is not exposed in about_flags.

Bug: 1064384
Change-Id: I7fda1e70bc57feb9977bef7bcd07b47cb93f38ea
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2219161
Commit-Queue: Alexander Cooper <alcooper@chromium.org>
Reviewed-by: Klaus Weidner <klausw@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773335}

Add variants of the gfx::ColorSpace::GetRangeAdjustMatrix() and
gfx::ColorTransform::NewColorTransform() methods that take bit depths as
input parameters, so that the range adjustment for high bit depths can
be performed accurately.

Use the new gfx::ColorTransform::NewColorTransform() method in
AVIFImageDecoder.

Tested:
gfx_unittests --gtest_filter=*ColorSpace*
blink_platform_unittests --gtest_filter=*AVIF*

Change-Id: Ib36ba6c05d5d7da3a4d0715ad5880336c0b1693c
Bug: 1087192
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2213127
Commit-Queue: Wan-Teh Chang <wtc@google.com>
Reviewed-by: ccameron <ccameron@chromium.org>
Reviewed-by: Peter Kasting <pkasting@chromium.org>
Reviewed-by: Dale Curtis <dalecurtis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773334}

This is a reland of d18bad64

This changes evaluateInPagePromise() to be evaluateInPageAsync() instead
as the latter waits on the promise returned from eval()'ing the code
before resolving the promise from the TestRunner.

It also puts back the promise returned at the end of the eval()'d code
in order for evaluateInPageAsync() to wait on it.

Original change's description:
> Fix the http/tests/devtools/console-cross-origin-iframe-logging.js test
>
> This test wants to verify that an error is seen when using a frame in
> another origin. It was using a different site, where the frame's
> contents is not visible at all when site-isolated. Instead, put the
> frame in another origin in the same site so that the main frame can see
> the contents of the iframe.
>
> R=caseq@chromium.org
>
> Bug: 623268
> Change-Id: Ib4ab03229f1033fc01717e561eb35b3c6d047e2c
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2218373
> Commit-Queue: danakj <danakj@chromium.org>
> Reviewed-by: Andrey Kosyakov <caseq@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#772769}

Bug: 623268
Change-Id: I48bdb6ed1b24665c1b2f03a84454d7d4fe2c0360
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2220468
Commit-Queue: danakj <danakj@chromium.org>
Reviewed-by: Andrey Kosyakov <caseq@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773333}

The failure is not site-isolation related. So moving to the correct
part of TestExpectations and changing the bug number.

TBR=yangguo@chromium.org
NOTRY=true

Bug: 1046784, 901502
Change-Id: I93f8f1250c00a43889ae3112fa89759884e7766c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2221429Reviewed-by: danakj <danakj@chromium.org>
Commit-Queue: danakj <danakj@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773332}

This CL modifies the content description of the buttons in tab strip
based on different features so that the difference can be reflected in
TalkBack.

Bug: 1085574
Change-Id: I26f7360f5900c1446b63a0edfeab1d7faa4595c6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2216364
Commit-Queue: Wei-Yin Chen (陳威尹) <wychen@chromium.org>
Reviewed-by: Wei-Yin Chen (陳威尹) <wychen@chromium.org>
Auto-Submit: Yue Zhang <yuezhanggg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773331}

This flag will control whether the nickname can be uploaded to Payments
or not, in case the pay.google.com change needs to launch separately.

Bug: 1082013
Change-Id: Icd9905fc4b211c16b7280d4cdb02f20ea9aef38e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2219600
Commit-Queue: Siyu An <siyua@chromium.org>
Commit-Queue: Jared Saul <jsaul@google.com>
Auto-Submit: Siyu An <siyua@chromium.org>
Reviewed-by: Jared Saul <jsaul@google.com>
Cr-Commit-Position: refs/heads/master@{#773330}

Disable MediaHistoryStoreFeedsTest.* on TSAN builders. These tests are
flaky on TSAN, likely due to a race condition with ScopedFeatureList.

TBR=beccahughes@chromium.org

Bug: 1087974
Change-Id: Id215412e6918a91b390f8fbd099adc1e4a0cc043
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222554Reviewed-by: Nina Satragno <nsatragno@chromium.org>
Commit-Queue: Nina Satragno <nsatragno@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773329}

https://chromium.googlesource.com/angle/angle.git/+log/98496cab72c7..3e1e10875278

2020-05-29 lehoangq@gmail.com Metal: Use 2d array for caching index conversion's pipeline state.
2020-05-29 amaiorano@google.com Fix CopyTextureTestES3.ES3FloatFormats for VK
2020-05-29 msisov@igalia.com Rename DisplayOzone to DisplayGbm.
2020-05-29 ianelliott@google.com Create a pre-rotation test for dFdx() & dFdy()
2020-05-29 cnorthrop@google.com Capture/Replay: Update MEC for Subway Surfer

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/angle-chromium-autoroll
Please CC ynovikov@google.com on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

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

Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win-asan;luci.chromium.try:win_optional_gpu_tests_rel;luci.chromium.try:linux-swangle-try-x64;luci.chromium.try:win-swangle-try-x86
Bug: chromium:1084458
Tbr: ynovikov@google.com
Change-Id: Ifddbf2b2eb511d8b3a913269b1bf636f5e9ad314
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222708Reviewed-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@{#773328}

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/perfetto-trace-processor-win-chromium
Please CC perfetto-bugs@google.com on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

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

Tbr: perfetto-bugs@google.com
Change-Id: I97194ae462156597638309e4b3160e3e9fc5e9d4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222632Reviewed-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@{#773327}

WindowState handles the kImmersiveIsActive property updates by updating
the shelf visibility state, which also hides the shelf if the shelf has
auto-hidden visibility state. The kImmersiveIsActive property has
started to be set repeatedly, even when the property value does not
change. The constant shelf visibility updates were thus preventing users
from dragging the shelf up in the immersive fullscreen, as the shelf
auto-hide visibility (i.e. the actual shelf visibility when it's in
auto-hidden state) set by dragging the shelf up was constantly getting
invalidated.

This CL updates WindowState logic to only update the shelf visibility
if the kImmersiveIsActive property actually changes, so the shelf
auto-hide visibility gets invalidated only if the property changes.
The same logic is applied for kHideShelfWhenFullscreenKey property
updates.

BUG=1084854
TEST=Play a video in the Youtube Android app, and set the video to play
     in fullscreen (to start immersive fullscreen). Drag the shelf up
     from the hidden state. Verify the shelf stays visible. Tapping on
     the app window should hide the shelf again.

Change-Id: Ica96b896e33d0c2dc876d87f302a8b68a8624409
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2212968
Commit-Queue: Toni Baržić <tbarzic@chromium.org>
Reviewed-by: Mitsuru Oshima <oshima@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773326}

Bug: 1084591
Tbr: geofflang@chromium.org
Change-Id: I159e7f85df1feb654a45de6a91252ec767f07f2f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2222549Reviewed-by: Kenneth Russell <kbr@chromium.org>
Commit-Queue: Kenneth Russell <kbr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773325}

This is a reland of 002510a0

This changes evaluateInPagePromise() to be evaluateInPageAsync() instead
as the latter waits on the promise returned from eval()'ing the code
before resolving the promise from the TestRunner.

Original change's description:
> Fix the http/tests/devtools/inspect-iframe-from-different-domain.js test
>
> This test wants to verify devtools inspect works with an iframe in
> another domain. It was putting the iframe in another site, which causes
> it to become detached from the local root in site isolation. Instead,
> put both the main frame and iframe into the example.test domain, with
> the iframe in a subdomain.
>
> This test was originally added for webkit bug 31587 which is found at
> https://bugs.webkit.org/show_bug.cgi?id=31587.
>
> It said:
> > If you try to inspect elements located inside an IFRAME that was
> > loaded from a different domain, not satisfying the same origin
> > policy, Web Inpector in recent WebKit nightly builds cannot show
> > style information. Works in Safari 4.0.4.
>
> R=yangguo@chromium.org
>
> Bug: 623268
> Change-Id: I842bf835e49ee79e60e423fd36a5d3cedb8a8cfa
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2218765
> Commit-Queue: danakj <danakj@chromium.org>
> Reviewed-by: Andrey Kosyakov <caseq@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#772743}

Bug: 623268
Change-Id: I05b8d89d51ec29d536677f9894937531592a249e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2220748Reviewed-by: Andrey Kosyakov <caseq@chromium.org>
Commit-Queue: danakj <danakj@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773324}

If a local frame gets a user activation, we set a bit on FrameTreeNode,
indicating that it has once received a user gesture. That bit is then
used to set an /active/ user gesture in RemoteFrames created for the
FrameTreeNode thereafter.

This manifests as flaky web tests that see a user gesture in OOP
iframes that they were not expecting - because they see an active user
gesture in the remote main frame that was created and consumed in the
previous test.

This would manifest in the wild as an OOPIF gaining a user gesture
that it can use to autoplay or whatnot when there is no active user
gesture.

R=avi@chromium.org, beccahughes@chromium.org, dcheng@chromium.org

Bug: 872952
Change-Id: Ia6f76ab33f4906e5ed60c553a5e9ea07ad75b439
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2220475
Commit-Queue: Daniel Cheng <dcheng@chromium.org>
Auto-Submit: danakj <danakj@chromium.org>
Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Reviewed-by: Becca Hughes <beccahughes@chromium.org>
Reviewed-by: Mustaq Ahmed <mustaq@chromium.org>
Cr-Commit-Position: refs/heads/master@{#773323}