GitLab

https://skia.googlesource.com/skia.git/+log/edf8a762190e..bcd8637772e3

$ git log edf8a7621..bcd863777 --date=short --no-merges --format='%ad %ae %s'
2017-05-22 brianosman Revert "Revert "fix ubsan warning""
2017-05-22 brianosman Revert "fix ubsan warning"
2017-05-22 brianosman Revert "We can mask load and store with just AVX."
2017-05-22 brianosman Revert "add tilers against 1"
2017-05-22 brianosman Revert "mark SkRasterPipelineBlitter final"
2017-05-22 brianosman Revert "fix ubsan warning"
2017-05-22 bsalomon Remove kZeroCopy_GrSurfaceFlag
2017-05-22 msarett Blacklist memory instensive tests on 32-bit bots (part 2) Bug: skia:
2017-05-22 mtklein fix ubsan warning

Created with:
  roll-dep src/third_party/skia


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

If the roll is causing failures, see:
http://www.chromium.org/developers/tree-sheriffs/sheriff-details-chromium#TOC-Failures-due-to-DEPS-rolls


CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.android:android_optional_gpu_tests_rel
TBR=brianosman@chromium.org

Change-Id: I9be699fe612c0027e3194d170126858b41467708
Reviewed-on: https://chromium-review.googlesource.com/510943Reviewed-by: Skia Deps Roller <skia-deps-roller@chromium.org>
Commit-Queue: Skia Deps Roller <skia-deps-roller@chromium.org>
Cr-Commit-Position: refs/heads/master@{#473720}

This change extends the peeking behavior when there are multiple
infobars attempting to show simultaneously. When a new infobar is
added, it will peek above the existing ones giving a brief view of
it's contents and bringing more attention to it.

BUG=721389

Review-Url: https://codereview.chromium.org/2846663002
Cr-Commit-Position: refs/heads/master@{#473719}

This only fails when no X11 headers are present on the system and

  use_ozone = true
  ozone_platform_wayland = true
  ozone_platform_x11 = false
  use_sysroot = false

BUG=295089

Review-Url: https://codereview.chromium.org/2892343002
Cr-Commit-Position: refs/heads/master@{#473718}

Functional changes here are:
 - Treating an extension-origin "about:blank" iframe as extension,
   for the purposes of the web_accessible_resources ancestor
   check.
 - Treating nested URLs as non-web_accessible_resources for
   the purposes of subframe navigation (this fixes 661324)

These should both be minor, and are incidental.

Pure refactoring here is:
 - Preferring url::Origin() checks (which leads to the functional
   changes above).
 - Renaming of variables for clarity.
 - Hoisting the extension lookup to the top of the function, and
   eliminating !extension and !registry checks after dealing
   with the lookup result.
 - A more efficient lookup of the parent frame, in a way
   that shouldn't require an "this is safe" apology comment.
 - Reflow control logic in anticipation of adding new BLOCK
   cases for platform apps (in a follow-on CL).

BUG=661324
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/2830893002
Cr-Commit-Position: refs/heads/master@{#473717}

It's better to leave this test app mainly for the manual testing
purposes and for the whitelist tests, but place the code for the
automated tests in another test app. The reason for that is the
complexity of updating it: besides changes in the source tree, each
updated version has to be published on the Web Store and then the CRX
file has to be fetched from it.

Therefore this CL makes this decision explicit by reflecting it in the
file names, readme file and the constant names.

Also an OWNERS file is introduced for it, in order to make the
ownership of this code clearer.

Bug: chromium:626343
Change-Id: Ic644b55379c45fa2da79db2df2e65c802778a2cc
Reviewed-on: https://chromium-review.googlesource.com/509512Reviewed-by: Achuith Bhandarkar <achuith@chromium.org>
Commit-Queue: Maksim Ivanov <emaxx@chromium.org>
Cr-Commit-Position: refs/heads/master@{#473716}

Also remove the function completely where it makes the code simpler:
- FrameCaptionButton
- ExitWarningWidgetDelegateView
- UserCardView

BUG=none

Review-Url: https://codereview.chromium.org/2901503003
Cr-Commit-Position: refs/heads/master@{#473715}

https://pdfium.googlesource.com/pdfium.git/+log/671630e52cda..ed75ee25f40a

$ git log 671630e52..ed75ee25f --date=short --no-merges --format='%ad %ae %s'
2017-05-22 tsepez CFX_ByteStringC: avoid taking unreferenceable vector::data() ptr.

Created with:
  roll-dep src/third_party/pdfium


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

If the roll is causing failures, see:
http://www.chromium.org/developers/tree-sheriffs/sheriff-details-chromium#TOC-Failures-due-to-DEPS-rolls


TBR=dsinclair@chromium.org

Change-Id: I4dc5ce447a5f84ed238fb6fd7642d1c017bc53a0
Reviewed-on: https://chromium-review.googlesource.com/511162
Reviewed-by: <pdfium-deps-roller@chromium.org>
Commit-Queue: <pdfium-deps-roller@chromium.org>
Cr-Commit-Position: refs/heads/master@{#473714}

BUG=689758

Change-Id: Iab841c65ab1a5acdd81b7e3e41015b16edd7ef30
Reviewed-on: https://chromium-review.googlesource.com/511302Reviewed-by: Jay Civelli <jcivelli@chromium.org>
Commit-Queue: Bo Liu <boliu@chromium.org>
Cr-Commit-Position: refs/heads/master@{#473713}

It appears that SkSurface is not getting created in some particular
cases.  It's not clear why this is happening, so add a bunch of
CHECKs that replicate Skia code to try to diagnose from the callsite
which of these conditions are the culprit.

Bug: 721744
Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel
Change-Id: Ida76afe7086756ba52276d1212860ebaf7a2f7c7
Reviewed-on: https://chromium-review.googlesource.com/511182Reviewed-by: ccameron chromium <ccameron@chromium.org>
Commit-Queue: enne <enne@chromium.org>
Cr-Commit-Position: refs/heads/master@{#473712}

BUG=724692

Review-Url: https://codereview.chromium.org/2896053002
Cr-Commit-Position: refs/heads/master@{#473711}

BUG=706699

Review-Url: https://codereview.chromium.org/2805593002
Cr-Commit-Position: refs/heads/master@{#473710}

It's left over from the early days of mustash/mash and we don't need it
anymore.

I left one reference because it'll be easier to clean up when WmShelf
gets converted.

Pure refactor, no behavior changes.

BUG=722552
TEST=ash_unittests, manual tests of chrome stylus tools

Review-Url: https://codereview.chromium.org/2895843002
Cr-Commit-Position: refs/heads/master@{#473709}

Changing from:
To get your settings on all your devices, sign in to Chrome.

to:
Sign in to Chrome to get your bookmarks, passwords, and more on all your devices.

BUG=723547

Review-Url: https://codereview.chromium.org/2893353003
Cr-Commit-Position: refs/heads/master@{#473708}

This CL fixes some issues in FlingTracker:

* The velocity from the OS is in pixel/second so it needs to be converted into
  pixel/ms.
* The displacement equation is missing the time constant multiplier.
* It is safer to decide the fling threshold by calculating the maximum
  displacement rather than just the time it takes.
* Just check whether the time elapsed is greater than the fling duration rather
  than calculating the speed every time the object moves.

Review-Url: https://codereview.chromium.org/2888423003
Cr-Commit-Position: refs/heads/master@{#473707}

BUG=724943

Review-Url: https://codereview.chromium.org/2893343002
Cr-Commit-Position: refs/heads/master@{#473706}

Follow-up work
- Trigger the right Feedback activity
- Control frequency of feedback via Finch

BUG=706438

Review-Url: https://codereview.chromium.org/2887383002
Cr-Commit-Position: refs/heads/master@{#473705}

This CL adds the following to the OIB (Page Info):
 - The menu only has two items (no default), with custom strings
 - The selected text uses user managed strings instead of default string
   (e.g. "Block" instead of "Block (default)"

BUG=689487

Review-Url: https://codereview.chromium.org/2884813003
Cr-Commit-Position: refs/heads/master@{#473704}

SequencedWorkerPool is being deprecated in favor of TaskScheduler.

BUG=667892
R=nick@chromium.org

Review-Url: https://codereview.chromium.org/2883503002
Cr-Commit-Position: refs/heads/master@{#473703}

address_input_strings.grd and the translation were moved to
third_party/libaddressinput/chromium/ in:
https://codereview.chromium.org/2886173002/

BUG=672458

Review-Url: https://codereview.chromium.org/2894673006
Cr-Commit-Position: refs/heads/master@{#473702}

XMLHTTPRequest::response_document_ gets incorrectly cleared because
Document::well_formed_ is incorrectly false. This is because
well_formed_ is initialized in Document::ImplicitClose().
XMLHttpRequest used to call ImplicitClose() directly, but I changed
it to call Document::CheckCompleted() (which may then call
ImplicitClose) in
https://chromium.googlesource.com/chromium/src/+/716a136332fbd8efc12a3892b6ae2a6c0be94ba7

I had thought that CheckCompleted() would always call ImplicitClose()
for an XHR response document, but it's possible for CheckCompleted()
to early-exit because the document is not complete if a style element
has blocked load completion and won't un-block it until a timer is
fired.

There's no reason to wait until ImplicitClose() to set the well-formed
bit, though. Document::FinishedParsing() is sufficient, and setting it
there removes the need for XMLHttpRequest to manually call
CheckCompleted().

BUG=721225
TEST=http/tests/xmlhttprequest/style-tag-in-responseXML.html

Review-Url: https://codereview.chromium.org/2885633002
Cr-Commit-Position: refs/heads/master@{#473701}

BUG=717501

Review-Url: https://codereview.chromium.org/2857143002
Cr-Commit-Position: refs/heads/master@{#473700}

Add SheetContentStateChangeObserver, which emits events specialised
for the visibility of a BottomSheetContent, and use it to record
user actions for the Suggestions bottom sheet

BUG=693593,717704

Review-Url: https://codereview.chromium.org/2862893002
Cr-Commit-Position: refs/heads/master@{#473699}

Because BeaconSeeds can take quite a bit of space, we also add an option in
RemoteDeviceLoader whether or not to load the BeaconSeeds.

Review-Url: https://codereview.chromium.org/2859053003
Cr-Original-Commit-Position: refs/heads/master@{#473489}
Committed: https://chromium.googlesource.com/chromium/src/+/c3636c2887c896911de99d208889b2582cf5e56b
Review-Url: https://codereview.chromium.org/2859053003
Cr-Commit-Position: refs/heads/master@{#473698}

This forces us to no longer assume that resetting the state of the mocks
happens synchronously at the end of the test.

Change-Id: I765f806439383b294a3b869b9b5df8b3fd192721
Reviewed-on: https://chromium-review.googlesource.com/510968
Commit-Queue: Reilly Grant <reillyg@chromium.org>
Reviewed-by: Vincent Scheib <scheib@chromium.org>
Cr-Commit-Position: refs/heads/master@{#473697}

Previously, when native heap profiling was enabled, all allocations in a process
were gated by a global lock. This proves to be a significant performance hit on
macOS. Profiling shows that 50-75% of all wall time is spent waiting on this
lock.

This CL introduces a new class ShardedAllocationRegister to handle sharding for
desktop platforms. This stores allocation/backtrace information across 64
different AllocationRegister instances. In addition, the sizes of the fixed size
hash maps were changed. The number of allocation buckets was reduced by a factor
of 8, and the number of backtrace buckets reduced by a factor of 16.

The new class ShardedAllocationRegister is thread-safe, and its consumers no
longer need to acquire a lock when using the container. Each consumer still
needs to know whether heap profiling is enabled. This CL uses a subtle::Atomic32
along with Acquire_Load and Release_Store to determine this state. Using a bool
to record this state and a base::Lock to synchronize reading/writing from the
bool has performance almost as bad as the global lock.

BUG=724651

Review-Url: https://codereview.chromium.org/2890363003
Cr-Commit-Position: refs/heads/master@{#473696}

This is a better place given that it is used by both FrameView and the root
PaintLayerScrollableArea.

SetCanHaveScrollbars remains on FrameView, but with a comment to clarify that it
is RLS-aware.

BUG=724309
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2900513002
Cr-Commit-Position: refs/heads/master@{#473695}

SequencedWorkerPool is being deprecated in favor of TaskScheduler.

BUG=667892
R=droger@chromium.org

Review-Url: https://codereview.chromium.org/2878943002
Cr-Commit-Position: refs/heads/master@{#473694}

SequencedWorkerPool is being deprecated in favor of TaskScheduler.

BUG=667892
R=droger@chromium.org

Review-Url: https://codereview.chromium.org/2881863002
Cr-Commit-Position: refs/heads/master@{#473693}

Replaces the g_desktops list with g_desktop, a single object. The script was originally written to handle a number of sessions at once, which were stored in g_desktops; however, we determined this feature wasn't needed and thus removed it to make the code simpler.

Review-Url: https://codereview.chromium.org/2891913003
Cr-Commit-Position: refs/heads/master@{#473692}

SequencedWorkerPool is being deprecated in favor of TaskScheduler.

BUG=667892
R=droger@chromium.org

Review-Url: https://codereview.chromium.org/2877783006
Cr-Commit-Position: refs/heads/master@{#473691}

Remove the task_runner from AudioDebugFileWriter's constructor and make it
use the new TaskScheduler API.

TaskRunners shouldn't be passed through several components. Instead, the
component that uses a TaskRunner should be the one that creates it.

The main change of this CL is in media/audio/audio_debug_file_writer.h, the
rest is mostly plumbing.

TBR=jam@chromium.org
BUG=689520
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.android:android_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2885173002
Cr-Commit-Position: refs/heads/master@{#473690}

BUG=531673

Review-Url: https://codereview.chromium.org/2894223002
Cr-Commit-Position: refs/heads/master@{#473689}

https://chromium.googlesource.com/webm/libvpx.git/+log/31cb852a908e..b3bf91bdc602

$ git log 31cb852a9..b3bf91bdc --date=short --no-merges --format='%ad %ae %s'
2017-05-21 marpan vp9: Adjustments to cyclic refresh for high motion.
2017-05-21 marpan vp9: Speed >= 8: Modify condition for low-resoln.
2017-05-18 marpan vp9: SVC: Modify condition to allow for copy partition.
2017-05-15 marpan vp9: Make copy partition work for SVC and dynamic resize.
2016-02-03 paulwilkins Changes to modified error.
2017-05-17 marpan vp9: Modify ChangingDropFrameThresh unittest.
2017-05-16 linfengz Update partial idct testing code
2017-05-12 johannkoenig neon fdct: 4x4 implementation
2017-04-05 paulwilkins Experiment. Store first pass errors as per MB values.
2017-05-16 linfengz Add transpose_32bit_4x4() and rename transpose_4x4() for vpx_dsp/x86
2017-05-12 johannkoenig neon 4 byte helper functions

Created with:
  roll-dep src/third_party/libvpx/source/libvpx
R=johannkoenig@google.com

Review-Url: https://codereview.chromium.org/2892403004
Cr-Commit-Position: refs/heads/master@{#473688}

This is a speculative fix, but seems to mostly eliminate
black screen flashes during VR presentation when the
"WebVR experimental rendering optimization" flag is enabled.
(It's off by default.)

Since this flush can happen in parallel with the previous
frame's rendering, move it ahead of the "wait for previous
render to finish" wait.

BUG=723962

Review-Url: https://codereview.chromium.org/2891033002
Cr-Commit-Position: refs/heads/master@{#473687}

Some code was still hanging around from a deprecated field trial.
This CL cleans it up.

BUG=

Review-Url: https://codereview.chromium.org/2895923003
Cr-Commit-Position: refs/heads/master@{#473686}

Disable flaky ExtensionUninstallDialogViewBrowserTest.EnsureCWSReportAbusePageIsActiveTabAfterUninstall on windows

BUG=725197
TBR=rdevlin.cronin@chromium.org

Review-Url: https://codereview.chromium.org/2898763003
Cr-Commit-Position: refs/heads/master@{#473685}

ImageLoader works like:
(1) Image loading is started (AddObserver()),
(2) Image loading is finished and the image element's load event is
    scheduled (ImageNotifyFinished()), and then
(3) The load event is fired.

Previously, has_pending_load_event_ was true from (1) to (3).
This CL makes has_pending_load_event_ true only from (2) to (3), i.e.
only while there is a pending task on EventSender.

Because "image_ && !image_complete_" is true from (1) to (2),
The previous "has_pending_load_event_" is equivalent to the current
"(image_ && !image_complete_) || has_pending_load_event_".
This CL, however, leaves most of "has_pending_load_event_" usage
unchanged, because in such cases they are equivalent, given the
assumptions around image_ and image_complete_, some of which are
confirmed by [1].

This is preparation for [2] that replaces has_pending_load_event_ with
a reference to a task that corresponds to the pending load event.

[1] https://codereview.chromium.org/2859383002/
[2] https://codereview.chromium.org/2863763003/

BUG=624697, 719759

Review-Url: https://codereview.chromium.org/2859093003
Cr-Commit-Position: refs/heads/master@{#473684}

Talked with Jon and there are no mocks for what the menu should look like, so I did enough to make the menu work on iPad but that is it. Going to wait until we gets some screens for design to finish the menu but for now this will allow us to continue work.

R=Yuweih@chromium.org

Review-Url: https://codereview.chromium.org/2885103002
Cr-Commit-Position: refs/heads/master@{#473683}

BUG=722553
TEST=covered by tests
R=msw@chromium.org

Review-Url: https://codereview.chromium.org/2897993003
Cr-Commit-Position: refs/heads/master@{#473682}

Based on updated design doc, Compose Email settings should always be
shown even if there is only one Mail client app installed. The settings
screen should show all but dim the unavailable ones.

BUG=711511
TEST=Delete one or both of Mail and Gmail app and verify the Settings dialog. Check VoiceOver usage as well.

Review-Url: https://codereview.chromium.org/2889183005
Cr-Commit-Position: refs/heads/master@{#473681}