GitLab

RenderFrames could be reused for same-origin navigations. This causes
the hints from previous navigations to be used in next navigation
until the new hints fetch finishes and updates the hints. This CL clears
the hints.

Unfortunately this causes race conditions between the image fetch and
the hints update, which causes images to not redirected to compressed
versions. So the image fetches are delayed a bit to circumvent this.

Bug: 1051283
Change-Id: Ie9e5351e68c0081ab9dbfe23faed2c3c5e8e4d42
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2056207
Commit-Queue: rajendrant <rajendrant@chromium.org>
Reviewed-by: Michael Crouse <mcrouse@chromium.org>
Reviewed-by: Tarun Bansal <tbansal@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742326}

TBR=govind@chromium.org

Change-Id: I7acc0ab00daa28f53f2c2f1e0eda99d6356f43bc
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2063249Reviewed-by: Ben Mason <benmason@chromium.org>
Reviewed-by: Krishna Govind <govind@chromium.org>
Commit-Queue: Krishna Govind <govind@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742325}

Change-Id: I6831f6b34bfa4891a17bc7dbb1145b436f6e5213
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2062690
Commit-Queue: Garrett Beaty <gbeaty@chromium.org>
Auto-Submit: Garrett Beaty <gbeaty@chromium.org>
Reviewed-by: Stephen Martinis <martiniss@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742324}

These code path that isn't activated anymore. Pepper will always be
composited so remove the IsAcceleratedCompositingActive method from
WebWidget.

The is_accelerated_compsiting_active_ bool was never set on the
WebFrameWidgetImpl as well.

Change-Id: I52e561863c2258694282d3aee9737f91f8fd282b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2062549Reviewed-by: Bill Budge <bbudge@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Reviewed-by: danakj <danakj@chromium.org>
Commit-Queue: Dave Tapuska <dtapuska@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742323}

This CL combines several block for the initial layout pass into a single
block.  Additionally, this CL updates the layout of the primary toolbar
layout so that it is behind the tab strip on iPad and in back of the
typing shield on iPhone.  This more explicitly describes the semantic
purpose of the previous |bottomView| logic.

Bug: none
Change-Id: I272882c4958afd2cdc25694a508bfe9119cb7db4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2045212
Commit-Queue: Kurt Horimoto <kkhorimoto@chromium.org>
Reviewed-by: Sergio Collazos <sczs@chromium.org>
Auto-Submit: Kurt Horimoto <kkhorimoto@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742322}

This reverts commit 388f1d99.

Reason for revert:

Findit (https://goo.gl/kROfz5) identified CL at revision 742315 as the
culprit for failures in the build cycles as shown on:
https://analysis.chromium.org/waterfall/culprit?key=ag9zfmZpbmRpdC1mb3ItbWVyRAsSDVdmU3VzcGVjdGVkQ0wiMWNocm9taXVtLzM4OGYxZDk5NzYyYzAzM2U0NzBhY2U0N2QxOGY5ZWU0NGU4NzRkN2QM

Sample Failed Build: https://ci.chromium.org/b/8888080110640214784

Sample Failed Step: compile

Original change's description:
> Reland "Enable ATK and Glib on Linux/Ozone"
> 
> This is a reland of 16c153e6
> 
> Original change's description:
> > Enable ATK and Glib on Linux/Ozone
> >
> > BUG=988094,1014934
> > R=sky
> > CC=nickdiego,msisov,adunaev
> >
> > Change-Id: I05b4bae5ccce53053ad4ae8269db849c2db9e569
> > Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2056570
> > Reviewed-by: Sadrul Chowdhury <sadrul@chromium.org>
> > Reviewed-by: Nick Yamane <nickdiego@igalia.com>
> > Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#742233}
> 
> TBR=sadrul,jamiewalch
> 
> Bug: 988094, 1014934
> Change-Id: Ide8b89767eddbdd6e6751db58c92d64e2d756fec
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2062911
> Reviewed-by: Thomas Anderson <thomasanderson@chromium.org>
> Reviewed-by: Jamie Walch <jamiewalch@chromium.org>
> Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#742315}


Change-Id: I06b8d6a675b1e862fb073345694639c0da024e0a
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 988094, 1014934
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2063257
Cr-Commit-Position: refs/heads/master@{#742321}

This was added in [1] to deal with the login web page. I believe
this is no longer relevant.

[1] https://chromiumcodereview.appspot.com/10454017

Bug: None
Test: login/oobe screen works as is.
Change-Id: Ib6275fe7a4c934797631e00c1c4a29ed650b5502
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2062672Reviewed-by: James Cook <jamescook@chromium.org>
Commit-Queue: Mitsuru Oshima <oshima@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742320}

This is a reland of 11357e7b

Original change's description:
> Proof of concept for XPC service in the updater.
> 
> Continues https://chromium-review.googlesource.com/c/chromium/src/+/2045058
> 
> Rebased. Moved code to the new chrome/updater/server/mac/server.mm::LaunchServer, addressed more feedback.
> 
> Need to create a plist for the xpc service. It is not clear yet what executatble will be used.
> 
> Bug: 1048653
> Change-Id: Iadbbd7004afa39fcc902f2ec68297845c2f94571
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2055284
> Commit-Queue: Mila Green <milagreen@chromium.org>
> Reviewed-by: Joshua Pawlicki <waffles@chromium.org>
> Reviewed-by: Sorin Jianu <sorin@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#742204}

Bug: 1048653
Change-Id: Ie6dbf13727627c9a448a1044bf059a82f33745cc
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2062872Reviewed-by: Robert Sesek <rsesek@chromium.org>
Reviewed-by: Joshua Pawlicki <waffles@chromium.org>
Commit-Queue: Mila Green <milagreen@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742319}

https://chromium.googlesource.com/angle/angle.git/+log/f87fac56d22f..4b7bac78a9ea

git log f87fac56d22f..4b7bac78a9ea --date=short --first-parent --format='%ad %ae %s'
2020-02-18 jmadill@chromium.org Vulkan: Throttle when way ahead of device.
2020-02-18 jmadill@chromium.org Vulkan: Command graph linearization for SemaphoreVk.
2020-02-18 lexa.knyazev@gmail.com Move sampleAlphaToCoverage out of blendState

Created with:
  gclient setdep -r src/third_party/angle@4b7bac78a9ea

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 cwallez@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/+/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
Bug: None
Tbr: cwallez@google.com
Change-Id: I9d353c87c0d620af2f276f806f7e70eb40d50191
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2062799Reviewed-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@{#742318}

InstallServiceWorkItem is something that is useful to have outside Chrome Setup. So this work item is for moving it to installer/util.

Bug: 1051723
Change-Id: I442e9de86394099c9c67483020915d609c6a51f7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2053083
Commit-Queue: S. Ganesh <ganesh@chromium.org>
Reviewed-by: Gabriel Charette <gab@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742317}

Bug: 982818
Change-Id: I49058c6b533e0f7355020b3c9bf7644d25bf5e56
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2052597Reviewed-by: Ted Choc <tedchoc@chromium.org>
Reviewed-by: Patrick Noland <pnoland@chromium.org>
Commit-Queue: Ender <ender@google.com>
Cr-Commit-Position: refs/heads/master@{#742316}

This is a reland of 16c153e6

Original change's description:
> Enable ATK and Glib on Linux/Ozone
>
> BUG=988094,1014934
> R=sky
> CC=nickdiego,msisov,adunaev
>
> Change-Id: I05b4bae5ccce53053ad4ae8269db849c2db9e569
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2056570
> Reviewed-by: Sadrul Chowdhury <sadrul@chromium.org>
> Reviewed-by: Nick Yamane <nickdiego@igalia.com>
> Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#742233}

TBR=sadrul,jamiewalch

Bug: 988094, 1014934
Change-Id: Ide8b89767eddbdd6e6751db58c92d64e2d756fec
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2062911Reviewed-by: Thomas Anderson <thomasanderson@chromium.org>
Reviewed-by: Jamie Walch <jamiewalch@chromium.org>
Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742315}

Using wpt-import in Chromium d97232ca.
With Chromium commits locally applied on WPT:
ec83d167 "Update performance.measureMemory to the latest proposal"
d5cd6b98 "SharedWorker: Assign unique names to SharedWorkers to avoid unintentional matching"


Note to sheriffs: This CL imports external tests and adds
expectations for those tests; if this CL is large and causes
a few new failures, please fix the failures by adding new
lines to TestExpectations rather than reverting. See:
https://chromium.googlesource.com/chromium/src/+/master/docs/testing/web_platform_tests.md

NOAUTOREVERT=true
TBR=lpz

No-Export: true
Change-Id: I05b667e8d26fe1ee172e129b72424d29ec3a3fd0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2063109Reviewed-by: WPT Autoroller <wpt-autoroller@chops-service-accounts.iam.gserviceaccount.com>
Commit-Queue: WPT Autoroller <wpt-autoroller@chops-service-accounts.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#742314}

views::ImageButton flips canvas by default.

Bug: b:149256457
Change-Id: I4399508956fb2c3a98c4edf2fbaa82df321178b4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2056941Reviewed-by: Xiaohui Chen <xiaohuic@chromium.org>
Reviewed-by: Xiyuan Xia <xiyuan@chromium.org>
Commit-Queue: David Black <dmblack@google.com>
Cr-Commit-Position: refs/heads/master@{#742313}

This is a reland of [1], which was accidentally removed by [2].  The
only difference is the gethostname() check is removed to address [3].

> For remote connections, SHM doesn't work, but the MIT-SHM extension
> may still be available, so we need to do an explicit check before
> trying to use SHM.

[1] https://chromium.googlesource.com/chromium/src/+/633c011511bae4ff193ac007e3e8ba64287c1d89
[2] https://chromium.googlesource.com/chromium/src/+/12a86d61adf2520c177ed51aec6583f490e29f7d
[3] https://bugs.chromium.org/p/chromium/issues/detail?id=1035803#c12

R=nickdiego
BUG=1035803

Change-Id: I5b707de48a8c2991aaf73c347b29ef813a9c28b4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2062900
Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
Commit-Queue: Nick Yamane <nickdiego@igalia.com>
Auto-Submit: Thomas Anderson <thomasanderson@chromium.org>
Reviewed-by: Nick Yamane <nickdiego@igalia.com>
Cr-Commit-Position: refs/heads/master@{#742312}

Right now, MenuScrollViewContainer claims to be an AXMenuBar, which
produces the following puzzling AX tree:

  AXGroup (menu root)
    AXMenuBar (scroll container)
      AXMenu (actual menu view)
        AXMenuItem ...

An AXMenuBar is something like the top system menu bar, at least on
macOS, not the "contents of a menu" or anything similar to that. After
this change, the AX tree is instead:

  AXGroup (menu root)
    AXMenu (actual menu view)
      AXMenuItem ...

which makes much more sense.

Bug: 954701
Change-Id: I00215874d75d9ee4870acd0e2101ef280dc2a199
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1962587
Commit-Queue: Elly Fong-Jones <ellyjones@chromium.org>
Reviewed-by: Peter Boström <pbos@chromium.org>
Reviewed-by: Dominic Mazzoni <dmazzoni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742311}

The recommendation is to use IFACEMETHODIMP and STDMETHODIMP in the
declaration.

See discussion thread at
https://groups.google.com/a/chromium.org/g/cxx/c/h4bdRqbduNY/

Change-Id: I810cfdec17917d7b758f5a3f050a8dfc70aedcf9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2051200Reviewed-by: Daniel Cheng <dcheng@chromium.org>
Reviewed-by: Peter Kasting <pkasting@chromium.org>
Reviewed-by: Robert Liao <robliao@chromium.org>
Commit-Queue: Xiaohan Wang <xhwang@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742310}

Display the deep scan request and response on chrome://safe-browsing for
debugging purposes

Screenshot: https://screenshot.googleplex.com/3Yk21T8p379.png
Bug: 1021945
Change-Id: I52c6a882069f014993fb50902d3c604c8c84b7d2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2055991Reviewed-by: Bettina Dea <bdea@chromium.org>
Commit-Queue: Daniel Rubery <drubery@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742309}

We originally disallowed same-document scroll to text navigations as a
security mitigation. We now have a better idea of the security concerns,
which are mitigated by ensuring the frame is not script accessible by
another frame, in which case it is safe to do a same-document scroll to
text navigation.

Even for same-document scroll to text navigations, we still require it
to be a browser navigation (e.g. same-document bookmark clicked).

Bug: 1041065 1023640
Change-Id: I2924ae9fc4bb01db4b75f184cc1fb70fae487b1c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1999000
Commit-Queue: Nick Burris <nburris@chromium.org>
Reviewed-by: Camille Lamy <clamy@chromium.org>
Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Reviewed-by: David Bokan <bokan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742308}

In r718454 we removed a magic constant from ScrollIntoView that
considered elements horizontally onscreen by more than 32px to be
fully-visible. This made it so that elements partially offscreen
horizontally would now be scrolled fully into view, just like in the
vertical axis. This was helpful as it makes ScrollIntoView more
predictable and interoperable with Gecko.

However, it caused issues for commonly occurring patterns. DevTools in
particular has several of these. It's common to have lists and similar
controls that are scrollable only vertically but may have some elements
that overflow (and are clipped) horizontally. In these cases, navigating
focus through the list should cause scrolling vertically but not
horizontally.

WebKit has this behavior as a result of the magic constant mentioned
above (from focus as well as scrollIntoView). Gecko has similar behavior
but only for focus and in both axes. More details in [1]

This CL makes it so that scroll into view coming from a focus() will
avoid scrolling partially visible elements in the horizontal axis. This
is close to what Gecko does but we still scroll in the vertical
direction. We keep the vertical axis because it's existing behavior and
likely desirable (if the last element in such a list isn't fully in view
it won't be able to be brought into view since there's no items to focus
below it).

We also rewrote the scroll-element-into-view web_test to make it more
modern and easier to manually test. In the process, the expectations are
updated since partially onscreen elements no longer scroll horizontally.
Prior to r718454, this test passed because the element was partially on
screen by less than the 32px threshold amount.

[1] https://github.com/w3c/csswg-drafts/issues/4778

Bug: 916631,1036817,1050868,1050979
Change-Id: Iff058123e368a6659a313ab3bb1a568187da1e23
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2056752Reviewed-by: Stefan Zager <szager@chromium.org>
Commit-Queue: David Bokan <bokan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742307}

This CL fixes the crash generated from
blink::scheduler::`anonymous namespace'::TaskTypesFromFieldTrialParam.

Refactors the code for InitializeTaskTypeQueueTraitsMap, lot of the code used
for testing the feature kThrottleAndFreezeTaskTypes is not being used any more.

Bug: 1049800
Change-Id: I8237981149b7518084d480cb79da047b876b8c3d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2057265
Commit-Queue: Sreeja Kamishetty <sreejakshetty@chromium.org>
Reviewed-by: Alexander Timin <altimin@chromium.org>
Reviewed-by: Scott Haseley <shaseley@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742306}

A CL has been landed to fix software decode video path on vulkan -
https://chromium-review.googlesource.com/c/chromium/src/+/2045130.

So enabling the media tests now which are passing on pixel2 skia
vulkan bot.

TBR=ynovikov@chromium.org

Bug: 985387
Change-Id: Iac61d31ccb12e35586e238e6c7780011385d86e9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2057540Reviewed-by: vikas soni <vikassoni@chromium.org>
Commit-Queue: vikas soni <vikassoni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742305}

Adding default page break too early will cause the default page break
gets deleted immediately.

Now move the installation of default page break after ChromeOS settings
app is installed.

Bug: 1031410
Test: Manual

Change-Id: If8d6522993518ccc109ac0bf368cdd077a97f301
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2056578Reviewed-by: Xiyuan Xia <xiyuan@chromium.org>
Commit-Queue: Long Cheng <lgcheng@google.com>
Cr-Commit-Position: refs/heads/master@{#742304}

When returning to the NTP after native is initialized, the native init
observer registration has no effect. This adds a check to for this
and triggers the colorful mic behavior immediately if native is already
initialized.

Bug: 1052073
Change-Id: I229ddebb16ddeb418219275874552e54fb64646e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2055389
Commit-Queue: Brandon Wylie <wylieb@chromium.org>
Reviewed-by: Ted Choc <tedchoc@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742303}

This patch allows users to try again when an authenticator with internal
user verification reports a failure to read their biometrics on
GetAssertion requests.

Summary of changes:
* Rename pin::RetriesRequest to pin::PinRetriesRequest and add
  pin::UvRetriesRequest.
* Have GetAssertionRequestHandler handle CTAP2_ERR_INVALID_PIN, get uv
  retries, and signal to the dialog that the user needs to retry.
* Add a sheet with a new fingerprint icon asking the user to try again
  and informing them of the number of retries.
* Pull up the error label code from AuthenticatorClientPinEntrySheetView
  into AuthenticatorRequestSheetView and simplify it by having it be
  sent by a method on the model rather than a delegate callback.

TBR'ing grt@ as per chrome/app/OWNERS

Tbr: grt@chromium.org
Bug: 1042779
Change-Id: I9d5db9da93a7119c3c2f3d005dbcc474045e586f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2005369
Commit-Queue: Nina Satragno <nsatragno@chromium.org>
Reviewed-by: Evan Stade <estade@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Reviewed-by: Martin Kreichgauer <martinkr@google.com>
Cr-Commit-Position: refs/heads/master@{#742302}

This avoids an interesting bug in Windows as per the code comment.

BUG=933975
R=brucedawson@chromium.org

Change-Id: I11c18e9706520d1c52238914dfac87039b45138b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2058815
Auto-Submit: Greg Thompson <grt@chromium.org>
Reviewed-by: Bruce Dawson <brucedawson@chromium.org>
Commit-Queue: Greg Thompson <grt@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742301}

When the color conversion pass is added for the first time, ensure that
there is full damage for the intermediate pass. If this is not
performed, then there will be uninitialized sections of the screen.

Similarly, add full damage when the color conversion pass is removed.
This turns out not to be completely necessary, because DirectRenderer
will call Reshape, and add full damage there. But there's no harm in
adding the damage in SurfaceAggregator, and it is easier to follow.

Add tests.

Bug: 1037954
Change-Id: I8272034547d2744f3ebeabf854ca78770034163b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2060661Reviewed-by: Sunny Sachanandani <sunnyps@chromium.org>
Commit-Queue: ccameron <ccameron@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742300}

Some tests in the virtual suite threaded-prefer-compositing will fail
when the feature is turned on by default since they run with smooth
scrolling enabled. The easy fix here would have been to simply add the
arg "--disable-smooth-scrolling" to the test suite but that will not
work since some tests under fast/scroll-behavior rely on smooth scroll
being enabled.
Additionally, disabling animated scrolling by using javascript only
works on the main thread (see bug for more info) so that isn't an
option either. This CL fixes the issue by allowing for adequate
time to finish animating by using waitForAnimationEndTimeBased.

Bug: 1038542
Change-Id: Ia5081a504cd99b3f10105f8ac20ad77ab966f063
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1986899
Commit-Queue: Rahul Arakeri <arakeri@microsoft.com>
Reviewed-by: David Bokan <bokan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742299}

Ensures that BrowserAccessibilityManager and its subclasses stop observing changes to AXTree upon destruction

Inspired by the seemingly unnecessary need to clear all observers in the AXTree's destructor that I noticed while reviewing
https://crrev.com/a37b994a406b8a900f60a0b2314d50c932411297
This patch ensures that:
1. BrowserAccessibilityManager and its subclasses stop observing all tree changes upon their destruction.
2. Subclasses of BrowserAccessibilityManager are disallowed from replacing the observed AXTree or its associated AXEventGenerator.

R=aleventhal@chromium.org, dmazzoni@chromium.org

Change-Id: Ib76612f4fabc095db942756bd876e9bc8da077d4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2058003
Commit-Queue: Nektarios Paisios <nektar@chromium.org>
Reviewed-by: Aaron Leventhal <aleventhal@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742298}

This fixes TabletModeTransitionHistogramsSnappedWindows test with
kHideShelfControlsInTabletMode enabled. The test verifies that snapped
windows do not animate when transitioning in and out of tablet mode.

The bounds animation was getting prevented using
WindowState::ScopedBoundsChangeAnimation in
TabletModeWindowState::LeaveTabletMode(), though the bounds animation
override was not being taken into account when updating bounds due to
display work area changes (which becomes relevant if display work area
when leaving tablet mode is different from display works area when
entering tablet mode).

What changes with the kHideShelfConrtolsInTabletMode feature enabled is
the timing of work area (shelf size) changes when entering tablet mode -
change in the shelf button visibility causes the shelf config update when
entering/exiting tablet mode, which was not previously being triggered
for the display bounds in the test (though, it probably should, as
the actual shelf bounds do change during tablet mode transition).
When entering tablet mode, shelf config is updated during
OnTabletModeStarting, which happens before tablet mode window state gets
detached for tablet mode transition, and thus, it won't match the display
work area when tablet mode exits.
This means that on tablet mode exit DefaultState::AttachState will
detect display work area change, and send the associated VMEvent to the
test window state, which will update the bounds with animation.

(With the feature disabled, work area changes happen after test window's
tablet mode state is set up, so display work areas when normal window
state is attached and detached match, avoiding extra bounds update).
Note that this happens only if one of the display dimensions is < 600
dip (i.e. not in general case, where work area gets updated in
OnTabletModeStarting due to a shelf config update).

Fix this by respecting the IMMEDIATE bounds change animation override
when animating window bounds for display metrics changes.

BUG=1051714

Change-Id: I88454f79bd88e06ab47677336b59650dac5a5864
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2057866
Commit-Queue: Toni Baržić <tbarzic@chromium.org>
Reviewed-by: Ahmed Fakhry <afakhry@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742297}

BUG=1023913

Change-Id: I6c581846479edd6f7938ab039bef38f3c38eb251
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2036492Reviewed-by: Elly Fong-Jones <ellyjones@chromium.org>
Commit-Queue: Avi Drissman <avi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742296}

This CL rewrites @support parsing, with a more obvious one-to-one
correspondence with the productions listed in the spec.

This is because I want to try to add support for the "unknown" value
logic [1], in which e.g. "not unknown(1)" becomes "false", and not
"true" like it does today. In order to do that, the parser need to
return two results so I can use-count when they differ: one which treats
"unknown" as false (like we do today), and one which treats it as
"unknown" until the top-level (i.e. per spec).

[1] https://drafts.csswg.org/mediaqueries-4/#typedef-general-enclosed
    See note on Kleene 3-valued logic.

Change-Id: I709805a303495ee3b3d41c25c914d2ac8496ce8b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2054143
Commit-Queue: Anders Hartvoll Ruud <andruud@chromium.org>
Reviewed-by: Xiaocheng Hu <xiaochengh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742295}

Bug: 983071
Test: `./grit/test_suite_all.py` passes
Change-Id: Idf9c94a82e9bcd692b80b280295893c169281313
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2045334Reviewed-by: Robert Flack <flackr@chromium.org>
Commit-Queue: Mike Frysinger <vapier@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742294}

Most of the functions in ContextFactoryPrivate are keyed on
ui::Compositor* to modify something for a specific display compositor.
The implementation forwards these calls to the corresponding
viz::mojom::DisplayPrivate. Instead, give Compositor a pointer to the
DisplayPrivate and make the calls directly.

Make VizProcessTransportFactory a ContextFactoryPrivate implementation
so that HostContextFactoryPrivate can be deleted.

For InProcessTransportFactory, move these calls into the
PerCompositorData object, which implements DisplayPrivate.

Bug: 1042259
Change-Id: I64f93e2024e2db74bd04da0ff4e4a2f7c82409d0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2050892
Commit-Queue: Sean Gilhuly <sgilhuly@chromium.org>
Reviewed-by: Scott Violet <sky@chromium.org>
Reviewed-by: kylechar <kylechar@chromium.org>
Reviewed-by: Johannes Henkel <johannes@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742293}

Root cause for the test failure was fixed in:
https://crrev.com/c/1865529

Change-Id: I513c8948803be3e432a4331920e5454b717f9ad7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2059048Reviewed-by: David Bokan <bokan@chromium.org>
Commit-Queue: Rahul Arakeri <arakeri@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#742292}

This CL adds boilerplate for initializing a Trust Tokens store in the
network service. The control flow is almost exactly the same as that for
initializing a cookie store: if NetworkContext gets a path in its
params, it constructs a disk-backed store; if it doesn't, it constructs
a store in memory.

by the 50K per engineer-week rule of thumb, not be cost-effective to
refactor away. The remainder is split quite widely among methods with a
median cost of perhaps 400 bytes each. I looked at the optimization
advice in the docs and did not find that a substantial amount of the
binary size increase corresponded obviously to any of the main patterns
mentioned in the docs.

Binary-Size: Out of the 18.5K increase, 5K is proto files which would,
Bug: 1042962
Change-Id: I3fe48d9cad1b36bb9424b5d2f1d50b5942e407df
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2044651Reviewed-by: Matt Menke <mmenke@chromium.org>
Reviewed-by: Robert Sesek <rsesek@chromium.org>
Reviewed-by: Vasilii Sukhanov <vasilii@chromium.org>
Commit-Queue: David Van Cleve <davidvc@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742291}

https://swiftshader.googlesource.com/SwiftShader.git/+log/2883de942283..8def9063c4fd

git log 2883de942283..8def9063c4fd --date=short --first-parent --format='%ad %ae %s'
2020-02-18 cwallez@chromium.org Only call updateBorders on ranges that can be cubemaps.

Created with:
  gclient setdep -r src/third_party/swiftshader@8def9063c4fd

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/swiftshader-chromium-autoroll
Please CC swiftshader-team+autoroll@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/+/master/autoroll/README.md

Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_chromium_msan_rel_ng;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: swiftshader-team+autoroll@google.com
Change-Id: Ibb598cd9bec0d32f7946c37ab908e2a016d4fe58
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2062728Reviewed-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@{#742290}

Clean up feature flag of tab switcher long-press menu.

Bug: 1046558
Change-Id: Iad0f443caf2c4e3dbc29f2eb30f608e50496dc21
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2057757Reviewed-by: Theresa  <twellington@chromium.org>
Commit-Queue: Lijin Shen <lazzzis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742289}

Make sure that home screen is not visible while split view is active
(similar to overview mode), to prevent home UI flashing behind split
view when one side transitions to/from overview.

BUG=1052502

Change-Id: Iffd215143463b41f1c4079b7ac0c7cf148f1fad5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2060665Reviewed-by: Alex Newcomer <newcomer@chromium.org>
Reviewed-by: Xiaoqian Dai <xdai@chromium.org>
Commit-Queue: Toni Baržić <tbarzic@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742288}

The address filling study was removed in favor of directly setting the
respective parameters in a multi-arm experiment. This change should
therefore not have any effect at all.

Bug: None
Change-Id: I9536bae8441032285f6103749b6c62ff0603b702
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2056803
Commit-Queue: Ilya Sherman <isherman@chromium.org>
Reviewed-by: Ilya Sherman <isherman@chromium.org>
Auto-Submit: Friedrich [CET] <fhorschig@chromium.org>
Cr-Commit-Position: refs/heads/master@{#742287}