GitLab

https://swiftshader.googlesource.com/SwiftShader.git/+log/a2749f39f47d..9b156615197d


git log a2749f39f47d..9b156615197d --date=short --no-merges --format='%ad %ae %s'
2019-03-14 bclayton@google.com SpirvShader: Split instructions into blocks.
2019-03-14 capn@google.com Fix allocation of low-alignment memory


Created with:
  gclient setdep -r src/third_party/swiftshader@9b156615197d

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

BUG=chromium:b/128527271
TBR=swiftshader-team+autoroll@chromium.org

Change-Id: I5385d7628f5985a5085977d08495951d0c8569a8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525034Reviewed-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@{#641048}

Change-Id: I9c0c299f8aa576ea34f357ac88c2c79c0a50fb48
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1315669
Commit-Queue: Christian Biesinger <cbiesinger@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641047}

Our previous implementation would destroy the connection to the mDNS
responder service host when the media capture permission is granted.
As a result, any mDNS candidate created before the grant would become
unresolvable. This issue is particularly evident when we have multiple
peer connections of the same origin, where a newly created peer
connection would temporarily toggle the permission to blocked before
restoring the granted status if the user did give the permission. This
could cause a peer connection created earlier and in the process of ICE
candidate gathering to perform unnecessary mDNS obfuscation. After this
change, the connection to the mDNS responder service host keeps alive,
but just becomes unavailable for registration of new names when we have
permissions granted.

Bug: 930339, 937992
Change-Id: I8ccdf0e6b578bc9e327337c6fbd102cdc542d15c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1522687
Commit-Queue: Qingsi Wang <qingsi@chromium.org>
Reviewed-by: Sergey Ulanov <sergeyu@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641046}

Some benchmarks such as rendering.mobile produce more data than the chromeperf
dashboard can handle at once.
This CL passes the new --max_bytes flag to add_reserved_diagnostics to batch
histograms into multiple files, then uploads them one at a time.

Initial CL: https://chromium-review.googlesource.com/c/1478085
Revert: https://chromium-review.googlesource.com/c/chromium/src/+/1486679
Reason for revert:
File "...\upload_results_to_perf_dashboard.py", line 61, in _GenerateDashboardJSONs
    with open(options.output_json_file, 'w') as f:
TypeError: coercing to Unicode: need string or buffer, file found

Patchset 1 is the initial CL.
Patchset 5 simplifies the initial approach and adds a test.

Bug: chromium:909961
Change-Id: Ia66cff24531333ecea386785ccbae04e7386f633
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1515649
Commit-Queue: Ben Hayden <benjhayden@chromium.org>
Reviewed-by: Caleb Rouleau <crouleau@chromium.org>
Reviewed-by: Simon Hatch <simonhatch@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641045}

if the block element is at selection end.

This patch adds a logic in blink::StyledMarkupSerializer<Strategy>::CreateMarkup
and blink::StyledMarkupTraverser<Strategy>::Traverse so that the empty block
element that only contains br element can be appended to
blink::StyledMarkupAccumulator if the block element is at the end of selection
during copy. The logic helps in creating the right markup to put onto clipboard
during serialization. For example, <i>some text</i><p><br></p>, select all and
copy, clipboard content should be look like:
<!--startFrament--><i>some text</i></font><p><br></p><!--EndFragment-->

Bug: 856324
Change-Id: Idd55bde473d4e6b3d7bf06b62167dd227ae65672
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1495371
Commit-Queue: Siye Liu <siliu@microsoft.com>
Reviewed-by: Yoshifumi Inoue <yosin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641044}

We simplified the location bar text color logic to be mostly based on
the primary color at
https://chromium-review.googlesource.com/c/chromium/src/+/1489416.

This becomes an issue for the search widget because the primary color
was transparent all the time.

Bug: 940897
Change-Id: I16db7596d9a0db0ee277a106b5b890be2045e391
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525286Reviewed-by: Yusuf Ozuysal <yusufo@chromium.org>
Commit-Queue: Becky Zhou <huayinz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641043}

https://chromium.googlesource.com/external/github.com/google/eDistantObject.git/+log/0146bd6e1f05..57586e721389

$ git log 0146bd6e1..57586e721 --date=short --no-merges --format='%ad %ae %s'
2019-03-14 haowoo Move the block from the stack to the heap before assigning it to the weak pointer.

Created with:
  roll-dep src/ios/third_party/edo/src

BUG=941680

Change-Id: Id616069d5894a535839547169669d49b02648e50
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524653Reviewed-by: Justin Cohen <justincohen@chromium.org>
Commit-Queue: Rohit Rao <rohitrao@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641042}

This change removes InkDropContainerView usage and overrides of:

* AddInkDropLayer
* RemoveInkDropLayer
* CreateInkDrop
* CreateInkDropRipple
* CreateInkDropHighlight
* CreateInkDropMask

... and replaces them with a HighlightPath already in use by the
FocusRing.

Bug: None
Change-Id: Id347f19d32f8b38443092eedd6604c71c5d9e6f1
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525462
Commit-Queue: Peter Boström <pbos@chromium.org>
Reviewed-by: Collin Baker <collinbaker@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641041}

Several member variables of WebViewRequestContextGetter may attempt to
access net::NetLog during destruction, but since the latter gets
destructed *after* the former, it can easily lead to dangling pointer
issues.

The fix is to just ensure those member variables are destroyed before
net::NetLog so they can still access it safely.

Alternatively, we can also just reorder the member variable list.

Change-Id: I768f235fc8f62835fdcab603b254a87620ab046b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525446
Commit-Queue: John Wu <jzw@chromium.org>
Reviewed-by: Hiroshi Ichikawa <ichikawa@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641040}

last one was reverted.

This time, a separated test binary is used to avoid dependency clashes
between protobuf libs.

Bug: 935711
Change-Id: I0789418f3a2c6c959845a1926f21832dcd0c389d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1515206
Commit-Queue: G. Silva <gfaus@chromium.org>
Reviewed-by: Peter Kasting <pkasting@chromium.org>
Reviewed-by: Michael Giuffrida <michaelpg@chromium.org>
Reviewed-by: Dirk Pranke <dpranke@chromium.org>
Reviewed-by: Henrique Grandinetti <hgrandinetti@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641039}

We were missing some goog.provide directives for generated interface
types. This also removes usage of mojo.internal.exportModule when
we're generating goog.provide, since only one of the two is necessary.

Bug: 939680
Change-Id: I1f813fe9cebb3cac6b68c2e86cff650fcf51ed21
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524331Reviewed-by: calamity <calamity@chromium.org>
Commit-Queue: Ken Rockot <rockot@google.com>
Cr-Commit-Position: refs/heads/master@{#641038}

A provider host on creation is always added into
ServiceWorkerContextCore and owned there, later it'll destroy either
when ServiceWorkerContextCore destroys or the Mojo connection it's bound
breaks.
This CL refactors code to strengthen the logic above.

BUG=931092

Change-Id: I8feaebec0e1c539ad5e24bde3af7518e2f81aa17
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1522559
Commit-Queue: Leon Han <leon.han@intel.com>
Reviewed-by: Matt Falkenhagen <falken@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641037}

+ bar background to grey 900 + 6% white in dark mode
+ promo background to grey 100 in light, grey 900 to dark
+ icon colors for close button and arrow button

Bug: 934580
Change-Id: I35ed4db9ba156c92223e4e2aa9f48c22e1e2a6e9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1521770Reviewed-by: Theresa <twellington@chromium.org>
Reviewed-by: Matthew Jones <mdjones@chromium.org>
Commit-Queue: Becky Zhou <huayinz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641036}

Previously the browser-assigned provider id was a negative int going
down from -2, to differentiate from the renderer-assigned provider id
which has non-negative value.
Now the renderer-assigned provider id has gone away with
https://crrev.com/c/1520688 and we just have all provider ids created
in the browser process, so we can make it be non-negative to be more
like of an ID.

BUG=931092

Change-Id: Icdc9933cc04e13c823e0a733d963f1f4411abf6e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1522557
Commit-Queue: Leon Han <leon.han@intel.com>
Auto-Submit: Leon Han <leon.han@intel.com>
Reviewed-by: Matt Falkenhagen <falken@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641035}

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: If46e19cad054575896ac34ed44f96efc4460c58f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525348Reviewed-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@{#641034}

This adds support for loading/re-indexing multiple rulesets in
FileSequenceHelper.

No behavior changes are made in this CL. This is in preparation of adding
support for dynamic rules and possibly multiple static rulesets.

BUG=930961

Change-Id: I13c99a3e5a89f7615317a04e08ecd5fc52464a45
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1519548
Commit-Queue: Karan Bhatia <karandeepb@chromium.org>
Auto-Submit: Karan Bhatia <karandeepb@chromium.org>
Reviewed-by: Jay Civelli <jcivelli@chromium.org>
Reviewed-by: Devlin <rdevlin.cronin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641033}

The Open button in the quickview panel was changed from an icon
to text to match the main FilesApp panel.

This change removes the tooltip, since it's redundant.

Bug: 941956
Change-Id: I1cd38e8787a56fe9a8647a5793314dbfcb1866fd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525248
Commit-Queue: Luciano Pacheco <lucmult@chromium.org>
Reviewed-by: Luciano Pacheco <lucmult@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641032}

This reverts commit fb08feb5.

Reason for revert: caused a crash (http://crbug.com/939667)

Original change's description:
> app_list: Set AppList as parent of uninstall dialog
> 
> Currently, when the uninstall dialog is opened from the Launcher, the Launcher
> is dismissed, which is confusing to the user. By parenting the dialog to the
> Launcher, the UX is much better.
> 
> Because we cannot directly access Ash windows anymore, this fix requires us to
> explicitly set the dialog's parent window via MUS properties.
> 
> TEST=manually verified + new browser tests
> BUG=814619
> 
> Change-Id: Ic25cb376191aaba8048f95fa5d0032b0328b8bc8
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1496447
> Reviewed-by: Devlin <rdevlin.cronin@chromium.org>
> Reviewed-by: Xiyuan Xia <xiyuan@chromium.org>
> Commit-Queue: Tim Song <tengs@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#638016}

TBR=xiyuan@chromium.org,tengs@chromium.org,rdevlin.cronin@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug: 814619
Change-Id: I7dd1c47d2edfc19fddd4302ae0b7fb1e42d835ac
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525273Reviewed-by: Vladislav Kaznacheev <kaznacheev@chromium.org>
Commit-Queue: Vladislav Kaznacheev <kaznacheev@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641031}

Make HomeScreenController responsible for determining home screen
visibility instead of AppListControllerImpl, including updates based on
window dragging and wallpaper preview.

The app list controller now implements
HomeScreenDelegate::ShowHomeScreen() to show the full-screen app list
window.

Also adds IsHomeScreenAvailable() which should be used to determine
whether to use the home screen instead of checking tablet mode directly.

Bug: 937549
Change-Id: If864614e7dcc4a818dc5368cf5e68e9475afacb1
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1504847
Commit-Queue: Michael Giuffrida <michaelpg@chromium.org>
Reviewed-by: Xiyuan Xia <xiyuan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641030}

Fix directory tree context menu for removable partitions when they're
grouped with a common root, where the "format" option wasn't shown.

Directory tree has 2 context menus root and non-root menus, when
partitions are grouped it's using a non-root menu which didn't have the
"format" command. This CL add the "format" command to non-root menu and
fix the condition to only display for removable root.

Bug: 941174
Change-Id: I27ff6f8e1813f84d76341c6abd4376a0743c14d8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1522508Reviewed-by: Stuart Langley <slangley@chromium.org>
Commit-Queue: Luciano Pacheco <lucmult@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641029}

Followup to http://crrev.com/c/1444011, which corrected remote sessions for
slimnav, but not for legacy nav.

Bug: 942139
Change-Id: I824bd4f7b45fac99d97ef78f53d84563086775b1
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525054Reviewed-by: Eugene But <eugenebut@chromium.org>
Commit-Queue: Justin Cohen <justincohen@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641028}

Based on https://crrev.com/c/1524907 by Hector Carmona <hcarmona@chromium.org>

Bug: 941882
Change-Id: Iea5cd2c0bd22bb57eafa6a4d507c93b045e50c31
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525196Reviewed-by: Hector Carmona <hcarmona@chromium.org>
Commit-Queue: Hector Carmona <hcarmona@chromium.org>
Commit-Queue: Dan Beam <dbeam@chromium.org>
Auto-Submit: Dan Beam <dbeam@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641027}

BUG=714018,940130
TBR=sky@chromium.org

Change-Id: I330f1a83003019ce1af35cf554fb7116e3da5bd7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525271Reviewed-by: Robert Liao <robliao@chromium.org>
Commit-Queue: Robert Liao <robliao@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641026}

Bug: 942241
Change-Id: I39e2bf63413a0bf2b646d465a9f694f411b0b0fb
Tbr: akihiroota@chromium.org,dtseng@chromium.org
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525006Reviewed-by: John Budorick <jbudorick@chromium.org>
Commit-Queue: John Budorick <jbudorick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641025}

Before this CL, the video capture service, on ChromeOS, would run in a
separate utility process unless the HAL v3 service was being used.

This CL changes this so that on ChromeOS the video capture service always
runs in the browser process. This is to mitigate an issue with screen
rotation events only being available in the browser process.

Bug: 891961
Change-Id: I4764f7cf131fff819e9e0f726d45bb86af09ed9f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1506604
Commit-Queue: Christian Fremerey <chfremer@chromium.org>
Reviewed-by: John Abd-El-Malek <jam@chromium.org>
Reviewed-by: Ricky Liang <jcliang@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641024}

This is a reland of 5b33ce1b

Original change's description:
> Update Crashpad to 1bb8ca4059d8bcc933478e02d7869dda7daa0263
> 
> 48ee086ca4c4 fuchsia: Update to run crashpad_test as package rather than
>              from /system
> 2271f00fe076 Add missing BUILD.gn for third_party/apple_cf
> 3332ae354672 compat: Provide <elf.h> and <mach-o/loader.h> for everyone
> 708367f5badf POSIX: ScopedMmap length refinement
> 5a21d038c9c9 Work around libcrashpad_handler_trampoline.so ASan trunk
>              clang problem
> c8fdad4fe58f Update DEPS for buildtools and GYP
> 2a47ea4663dd Obtain the Fuchsia SDK from the correct (updated) location
> 36f9051cafca Enable clang -Wextra-semi in POSIX/Fuchsia clang builds
> 348022def8d5 Fuchsia: Proper Fuchsia-on-macOS cross build support
> aa160f658175 [fuchsia] Update header include
> 1bb8ca4059d8 Silence some warnings
> 
> Change-Id: Ic40aa09f581c5b1339ac7d7c0a50a55f2d15ea99
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524084
> Commit-Queue: Scott Graham <scottmg@chromium.org>
> Commit-Queue: Mark Mentovai <mark@chromium.org>
> Auto-Submit: Scott Graham <scottmg@chromium.org>
> Reviewed-by: Mark Mentovai <mark@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#640817}

Change-Id: If37cdebfafda6ae1526c665f9362a86b5a59ed68
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525281
Commit-Queue: Scott Graham <scottmg@chromium.org>
Commit-Queue: Wez <wez@chromium.org>
Auto-Submit: Scott Graham <scottmg@chromium.org>
Reviewed-by: Wez <wez@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641023}

Because tab thumbnailing logic will be using standard tab size as part
of its calculation of how large of a thumbnail to create (to be used in
tab hover cards) the core properties of a tab need to be available in
c/b/ui rather than just c/b/ui/views. Otherwise, an illegal dependency
is created.

The updated thumbnailing logic will be added in a later CL.

Bug: 928954
Change-Id: Ie51ef3057828a0baad670fd1459df3cb6784651e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1519322
Commit-Queue: Dana Fried <dfried@chromium.org>
Reviewed-by: Peter Kasting <pkasting@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641022}

Due to limitations in how gclient works, we currently need to
duplicate entries in //DEPS and //buildtools/DEPS, and it is
possible for them to not be kept in sync. We don't want that.

This CL adds a presubmit check so that if you modify the relevant
lines in one file you are required to modify them in another.

If we ever fix the linked bug so that we don't need to duplicate
the content, we can remove these presubmit checks.

This CL also gets rid of deps_revisions.gni in //buildtools,
and switches to passing the libcxx_revision value from gclient
to GN via gclient_gn_args.

In order for this to work correctly, anyone using both
//buildtools and //build in GN will need to *also* export the
variable from their DEPS files.

Bug: 941824, 930836
Change-Id: I438c9966c497737cd4984a14cdb0c8993ffcad9b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1522466
Commit-Queue: Dirk Pranke <dpranke@chromium.org>
Reviewed-by: Thomas Anderson <thomasanderson@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641021}

OnFilePickerDestroy() needs to also destroy the pointer it receives.

BUG=714018,940130

Change-Id: Ib0402575a7115b0e8b125e1131b61238561e95f8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524879
Auto-Submit: Robert Liao <robliao@chromium.org>
Reviewed-by: Thomas Anderson <thomasanderson@chromium.org>
Commit-Queue: Thomas Anderson <thomasanderson@chromium.org>
Commit-Queue: Robert Liao <robliao@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641020}

Technically, if we violated this assertion, we'd do an invalid cast before
calling a member function, leading to undefined behavior.

This also switches to using the generic View* base type in cases where the
subtype is not needed, which allows using standard Views accessors more.

Bug: none
Change-Id: I40b51579273b0b1f3c64d0cf15beeaed4708957e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524540
Commit-Queue: Peter Kasting <pkasting@chromium.org>
Commit-Queue: Weiyong Yao <braveyao@chromium.org>
Auto-Submit: Peter Kasting <pkasting@chromium.org>
Reviewed-by: Weiyong Yao <braveyao@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641019}

This reverts commit 48e48116.

Reason for revert: To unblock libassistant roll

Original change's description:
> [protobuf] remove deprecated deps parameter
> 
> I replaced all appearance of deps with proto_deps or link_deps.
> 
> Bug: 938011
> Change-Id: I95b8d6363358ecd8c350f095baaa10b12a968e3e
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1506915
> Auto-Submit: Takuto Ikuta <tikuta@chromium.org>
> Commit-Queue: Peter Kasting <pkasting@chromium.org>
> Reviewed-by: Peter Kasting <pkasting@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#640063}

TBR=pkasting@chromium.org,xyzzyz@chromium.org,tikuta@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug: 938011
Change-Id: Idccc769b22a11b46a8db421c6d62e09ce012b67d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524744Reviewed-by: Takuto Ikuta <tikuta@chromium.org>
Commit-Queue: Takuto Ikuta <tikuta@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641018}

Don't send metrics for unrelated crashes, only for GWP-ASan crashes and
error reports. The unrelated crashes are uninteresting and obscure the
other (smaller) numbers of reports that I am interested in.

Bug: 896019
Change-Id: Ic205f5fb0798b489889c2fa639fc2464f066bb6b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1518755
Commit-Queue: Vlad Tsyrklevich <vtsyrklevich@chromium.org>
Auto-Submit: Vlad Tsyrklevich <vtsyrklevich@chromium.org>
Reviewed-by: Vitaly Buka <vitalybuka@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641017}

This is a speculative fix.

We create a KVO for -[NSRunningApplication isTerminated] on the IO
thread, where the NSRunningApplication is created. In the observer, we
post a task to the UI thread, and we remove the KVO in that callback.
This removal causes a crash claiming that no such observer exists.
Note that the observer is also called with the initial value, because
of the NSKeyValueObservingOptionInitial argument.

I can see two potential problems here:
1. This code assumes that the KVO is only called once.
    - This is a boolean property that will only ever to NO->YES, so this
      would seem a safe assumption.
    - If this were not true, then we would call removeObserver twice,
      which is what we're seeing occur. We would also use-after-free the
      TerminationObserver.
2. The callback on the UI thread could potentially be made before the
   call to addObserver on the IO thread is complete.
    - This would call removeObserver from within addObserver, which is
      not allowed.

Fix the two problems by:
1a. In onTerminated, early-out if |callback_| has already been issued.
1b. Use a scoped_nsobject for the task posted to the UI thread.
2.  Call both addObserver and removeObserver on the UI thread (so that
    they won't become reentrant).
X.  Only create the observer if -[NSRunningApplication isTerminated] is
    NO. This is not needed to fix any bugs, but limits the cases in
    which we create a TerminationObserver.

Bug: 941909
Change-Id: Id3b36900ac0c46a7ec48d4c7d2e3600cc5c61ef8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524340Reviewed-by: Avi Drissman <avi@chromium.org>
Commit-Queue: ccameron <ccameron@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641016}

Summary of changes available at:
https://chromium.googlesource.com/v8/v8/+log/7677b798..3f6dd8f7

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;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel;luci.chromium.try:android_optional_gpu_tests_rel

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

Change-Id: I4c3a802251af5656fa7f6ac064265dd0708a99f7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524916Reviewed-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@{#641015}

Should fix a flaky testLongPressGestureFollowedByTapDoesntSelect test
which seems to work fine on M phones, and hopefully later versions
of Android too.

BUG=837998

Change-Id: I4233b6eda177ff1bde9449581f32b8029d4b41bd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524092
Auto-Submit: Donn Denman <donnd@chromium.org>
Commit-Queue: Matt Falkenhagen <falken@chromium.org>
Reviewed-by: Matt Falkenhagen <falken@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641014}

The old JS-based bindings for extension messaging had a handful of
native helpers, which are now obsolete with the full native
implementation. Remove the JSRendererMessagingService and ExtensionPort
classes entirely, and remove the vast majority of MessagingBindings.
MessagingBindings::BindToGC() is used outside of messaging, and so needs
to remain (but we should rename the class or move the method in a
follow-up).

Bug: 938998
Change-Id: Ia697f4e19b8ac03ef87a62d9b8385cced8f08fbd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1523981Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Commit-Queue: Devlin <rdevlin.cronin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641013}

This CL:
1. Moves the ScopedVisibility tracker to chrome/browser
2. Adds it to the PageLoadMetricsObserverDelegate
3. Updates the http/s engagement observer to use the new delegate
   method.
4. Updates ScopedVisibilityTracker::GetForegroundDuration to be
   properly const.

This CL is in preparation for another page load metrics observer
that needs this kind of capability. It shouldn't have any behavior
change.

Bug: 937298
Change-Id: I685b05e191ccc04915eb5be7685332d800fda278
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1525182Reviewed-by: Avi Drissman <avi@chromium.org>
Commit-Queue: Charlie Harrison <csharrison@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641012}

Change-Id: I73c538d0ebf1e4b1aa71e91f32d8102e1fd0c435
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1510696Reviewed-by: Yury Khmel <khmel@chromium.org>
Commit-Queue: Nigel Tao <nigeltao@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641011}

This is important if first attempt to upload CL has failed for some reason, e.g.
presubmit failure, and on the second try we only need to execute git-cl-upload.

R=perezju@chromium.org

Bug: 895891
Change-Id: If55e2143365ae4b1135e0d65bdaf2fc5a4edcb4d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1520679
Commit-Queue: Sergiy Belozorov <sergiyb@chromium.org>
Reviewed-by: Juan Antonio Navarro Pérez <perezju@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641010}

Bitrate caps were set only during OnStatsDelivered(), which is
connected to the ICE connection state, not the signalling (SDP) state.

This CL sets the global (PeerConnection) bitrates after
SetRemoteDescription() is called. And it sets the Sender bitrates when
the (Video)Sender is created.

Since the relay status is (probably) not known at those times, the
bitrates are also set in OnStatsDelivered() if the relay status has
changed.

By setting the min/max caps sooner, this may improve BWE performance.

Bug: 941217
Change-Id: Ic623eb78544d22b1ebc94dcef906fd85e4d289b5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1524690Reviewed-by: Jamie Walch <jamiewalch@chromium.org>
Reviewed-by: Seth Hampson <shampson@chromium.org>
Commit-Queue: Lambros Lambrou <lambroslambrou@chromium.org>
Cr-Commit-Position: refs/heads/master@{#641009}