GitLab

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/fa85120..543f3bc

TBR=pilgrim@chromium.org,timloh@chromium.org

Review URL: https://codereview.chromium.org/878013002

Cr-Commit-Position: refs/heads/master@{#313218}

BUG=450908

Review URL: https://codereview.chromium.org/869863002

Cr-Commit-Position: refs/heads/master@{#313217}

I was confused about how this worked when writing
http://crrev.com/864993002 and none of the existing
examples clarified it. Is there a utility method that will
just always append to the path? like "a.com/b" and "c"
makes "a.com/b/c"?

Review URL: https://codereview.chromium.org/870413005

Cr-Commit-Position: refs/heads/master@{#313216}

Currently, the PDF extension isn't permitted access to paths to network
mounts (e.g. file://network_share/path.pdf). A permission of <all_urls>
allows access to such URLs.

BUG=452319

Review URL: https://codereview.chromium.org/880773002

Cr-Commit-Position: refs/heads/master@{#313215}

Revert "Remove hardware check that enables always-on hotwording."

This reverts commit e014345e.

Review URL: https://codereview.chromium.org/838753004

Cr-Commit-Position: refs/heads/master@{#313214}

Previously JingleThreadWrapper was initialized in the remoting PNaCl
plugin such that JingleThreadWrapper::current() was always returning
NULL. JingleThreadWrapper::current() is used only to implement
rtc::Thread::Send() and until recently Send() wasn't used in the parts
of webrtc used in remoting. Now rtc::StunRequest uses Send(), which
crashes the plugin.

Added JingleThreadWrapper::WrapTaskRunner() that properly initializes
current() and is used in the remoting plugin now.

Also JingleThreadWrapper is no longer passes SocketServer to the
parent as it is no longer necessary.

BUG=451643

Review URL: https://codereview.chromium.org/876433003

Cr-Commit-Position: refs/heads/master@{#313213}

This may cause the draw callback to be run earlier than before if the Surface would later be added to a Display, but prevents it ensures that the callback runs if the Surface wouldn't be drawn at all

BUG=440269

Review URL: https://codereview.chromium.org/792383004

Cr-Commit-Position: refs/heads/master@{#313212}

R=lcwu@chromium.org,erickung@chromium.org
BUG=None

Review URL: https://codereview.chromium.org/882563002

Cr-Commit-Position: refs/heads/master@{#313211}

--enable-plugin-power-saver flag previously overrode anything in the
content settings flag. This breaks for BLOCK setting and unauthorized
plugins, as the browser side needs to authorize the plugin first.

This patch restricts the flag to only impact users on ALLOW. This makes
sense to me, as the flag should not decrease the security of users
who have explicitly chosen ASK or BLOCK.

It will also dodge the bug.

BUG=448171,403800

Review URL: https://codereview.chromium.org/876803005

Cr-Commit-Position: refs/heads/master@{#313210}

R=maruel@chromium.org, yoz@chromium.org
TBR=jam@chromium.org
BUG=98637

Committed: https://crrev.com/d1caf97fe8db47e5c76043bccbf4dd4a42265831
Cr-Commit-Position: refs/heads/master@{#312993}

Review URL: https://codereview.chromium.org/869783004

Cr-Commit-Position: refs/heads/master@{#313209}

BUG=451932

Review URL: https://codereview.chromium.org/883503003

Cr-Commit-Position: refs/heads/master@{#313208}

The original CL was reverted because it caused non-deterministic failures in
browser tests that used the file tab-restore-visibility.html. The test file had
an edge case that caused thrashing and prevented the page from loading. This
has since been fixed.

> BrowserWindowController always layer-backs the contentView of the window. If
> Chrome were linked against OSX 10.9, this would also cause the root view to be
> layer backed. Since Chrome is linked against OSX 10.6, the root view by default
> is not layer backed.
>
> BrowserWindowController is going to have a custom AppKit fullscreen animation,
> which requires that the root view is layer backed. While it is possible to
> implement a custom AppKit fullscreen animation as early as OSX 10.7, the
> effects of layer backing the root view on OSX 10.7 are murky. It seems much
> safer to just layer back the root view on OSX 10.9+, and only enable the custom
> AppKit fullscreen animation on OSX 10.9+.
>
> As far as I know, this change has no detrimental side effects. There is no
> spike in memory usage
> (https://code.google.com/p/chromium/issues/detail?id=396264#c11), presumably
> because NSThemeFrame implements -wantsUpdateLayer on OSX 10.9+.
>
> BUG=413009, 414527
> Committed: https://crrev.com/8dec34f8f3504352e4eef95b2aaa36c561aea930
> Cr-Commit-Position: refs/heads/master@{#311191}

BUG=413009, 414527
TBR=rsesek@chromium.org

Review URL: https://codereview.chromium.org/878673006

Cr-Commit-Position: refs/heads/master@{#313207}

Failing on the bots. See:
http://build.chromium.org/p/chromium.perf/builders/Mac%2010.9%20Perf%20%285%29/builds/3254

BUG=452257
TBR=manisharora@google.com

Review URL: https://codereview.chromium.org/879763003

Cr-Commit-Position: refs/heads/master@{#313206}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/1d124c3..2d80dd2

CQ_EXTRA_TRYBOTS=tryserver.blink:linux_blink_rel,linux_blink_dbg
TBR=fmalita@google.com

Review URL: https://codereview.chromium.org/875293002

Cr-Commit-Position: refs/heads/master@{#313205}

This is part of a series moving IO on evdev devices to a fast thread,
since UI thread is too slow for the cursor.

DeviceDataManager is a UI thread object, so move final transformation
of touches into EventFactoryEvdev prior to dispatch into UI. This
will avoid sharing state across threads.

BUG=449710
TEST=boot link_freon & move cursor

Review URL: https://codereview.chromium.org/874713002

Cr-Commit-Position: refs/heads/master@{#313204}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/02cd199..fa85120

TBR=pilgrim@chromium.org,timloh@chromium.org

Review URL: https://codereview.chromium.org/877973002

Cr-Commit-Position: refs/heads/master@{#313203}

To remove ResourceEntry completely, tests for syncFileSystem need to be modified.
I'll do it by a separate CL.

BUG=357038
TBR=rogerta@chromium.org
TBRing for removal of unused files from *.gyp and *.gn under google_api
TEST=build all targets

Review URL: https://codereview.chromium.org/861133002

Cr-Commit-Position: refs/heads/master@{#313202}

BUG=429053
TEST=gpu_unittests
R=sievers@chromium.org

Review URL: https://codereview.chromium.org/867123008

Cr-Commit-Position: refs/heads/master@{#313201}

Summary of changes available at:
https://chromium.googlesource.com/external/google-breakpad/src/+log/58aca2d..398c70a

398c70a Fix a source of memory corruption.

BUG=449214

Review URL: https://codereview.chromium.org/877943002

Cr-Commit-Position: refs/heads/master@{#313200}

  StateMachine state(scheduler_settings);
  state.SetCanStart();
  state.UpdateState(state.NextAction());
  state.CreateAndInitializeOutputSurfaceWithActivatedCommit();
  state.SetVisible(true);
  state.SetCanDraw(true);

BUG=443821

Review URL: https://codereview.chromium.org/875613002

Cr-Commit-Position: refs/heads/master@{#313199}

BUG=110020

Committed: https://crrev.com/50567ebbd5ca2ff79d87382addc11d25658776d2
Cr-Commit-Position: refs/heads/master@{#312076}

Review URL: https://codereview.chromium.org/840493002

Cr-Commit-Position: refs/heads/master@{#313198}

Modify SetAlternateProtocol() behavior with respect to logging and histograms, by
ignoring alternate protocol being shadowed by |g_forced_alternate_protocol| or
being inactive due to too high probability threshold.

Also clean up HasAlternateProtocol() and GetAlternateProtocol().

BUG=392575

Review URL: https://codereview.chromium.org/858433004

Cr-Commit-Position: refs/heads/master@{#313197}

SkTypeface is moving to use SkStreamAsset instead of SkStream. The
implementations are already assuming this, and all users are in
practice already creating the correct types. This change simply
clarifies existing practice and documents actual restrictions.

The Skia side change will be https://codereview.chromium.org/869763002/

Review URL: https://codereview.chromium.org/870133002

Cr-Commit-Position: refs/heads/master@{#313196}

See: https://codereview.chromium.org/870693002

R=erickung@chromium.org,dalecurtis@chromium.org
BUG=450764

Review URL: https://codereview.chromium.org/869283003

Cr-Commit-Position: refs/heads/master@{#313195}

BUG=152422

Review URL: https://codereview.chromium.org/880733002

Cr-Commit-Position: refs/heads/master@{#313194}

There are two reasons for this change:

1. The current code path doesn't work correctly if the media stream and the output devices clocks don't match up *exactly*. Underruns/lipsync issues can occur.

2. This allows for audio to be explicitly buffered for some period of time, which I plan to utilize in the cast_streaming receiver code.

Currently this code path is not often used, since it only happens when you get a media stream from getUserMedia and plug it into a media player without going through webrtc.

Review URL: https://codereview.chromium.org/856843002

Cr-Commit-Position: refs/heads/master@{#313193}

The libgcc_eh.a file is no longer in the translator tarball
possibly after NaCl revision:

1fceb45: (dschuff) Wrap PNaCl sandboxed translator build in toolchain_build recipe

Cr-Commit-Position: refs/heads/master@{#312891}, fixed
this for windows builds, but not linux / mac.

Review URL: https://codereview.chromium.org/879803002

Cr-Commit-Position: refs/heads/master@{#313192}

This file was unused since r263451 and it looks like it was accidentally
not deleted.

Review URL: https://codereview.chromium.org/836753005

Cr-Commit-Position: refs/heads/master@{#313191}

effect of the prerenderer not working right.

Convert the CHECK into a conditional for now and leave an explanation of why it
can fail.

BUG=452299
TEST=none
TBR=vitalybuka

Review URL: https://codereview.chromium.org/876203002

Cr-Commit-Position: refs/heads/master@{#313190}

The ChromeOS system UI (rendered through HarfBuzz) now
relies on gfx::RenderText::CreateSkiaTypeface(), and
skia actually returns exactly same typeface for
'Noto Sans UI' and 'ui-sans' in the current configuration.
ui-sans is simply redundant.

BUG=444894
R=jshin@chromium.org

Review URL: https://codereview.chromium.org/823003002

Cr-Commit-Position: refs/heads/master@{#313189}

The CL [1] tightened the error check too much and revealed
an existing inconsistency and resulted a production crash.
This CL makes a workaround to that crash, turning a PCHECK()
to DPCHECK().

[1] https://crrev.com/ce44fef5fd60dd2be5c587d4b084bdcd36adcee4

BUG=449233,448245
R=agl@chromium.org,jam@chromium.org

Review URL: https://codereview.chromium.org/882543002

Cr-Commit-Position: refs/heads/master@{#313188}

This constant is no longer used in internal repository.

TBR=tedchoc@chromium.org
BUG=436083

Review URL: https://codereview.chromium.org/876013004

Cr-Commit-Position: refs/heads/master@{#313187}

BUG=451654

Review URL: https://codereview.chromium.org/873403004

Cr-Commit-Position: refs/heads/master@{#313186}

This separates out all the service worker specific code from the generic
connection handling. As well as exposing some API in content/public so
code outside of content/ can handle connections.

This is the first step in making it possible to have chrome extensions,
or other code outside of content/ be the endpoint of a connection.

This doesn't change any functionality, it just moves code around.

BUG=426458

Review URL: https://codereview.chromium.org/861373002

Cr-Commit-Position: refs/heads/master@{#313185}

As we addressed certain amount of the performance problem,
it's time to give it another try.

R=viettrungluu@chromium.org
BUG=377980

Review URL: https://codereview.chromium.org/857483004

Cr-Commit-Position: refs/heads/master@{#313184}

Fix a TODO I added a while back.  Instead of calling SetMinContentsScale
and SetTileGridSize directly, pass these as parameters to constructor of the
PicturePile.

Committed: https://crrev.com/6d1f2981f3204852a9189d0b6e7ad0b18659d68d
Cr-Commit-Position: refs/heads/master@{#309867}

Review URL: https://codereview.chromium.org/811953005

Cr-Commit-Position: refs/heads/master@{#313183}

Cherry-pick potential divide-by-zero fix
https://codereview.chromium.org/877783004

This includes remnants of an attempt to turn on AVX2. That failed and the
critical parts have been reverted. However, the attempt revealed some necessary
changes which have been kept for the next attempt. These constitute a no-op for
this particular roll.

Enable AVX2:
https://codereview.chromium.org/863143002

Fix GN build:
https://codereview.chromium.org/865893004
https://codereview.chromium.org/817293008
https://codereview.chromium.org/866273003

Fix cpu detection:
https://codereview.chromium.org/864393002

Disable AVX2:
https://codereview.chromium.org/880513003

R=tomfinegan@chromium.org
R=yaowu@chromium.org
BUG=396572

Review URL: https://codereview.chromium.org/883503002

Cr-Commit-Position: refs/heads/master@{#313182}

Gnubby was added a a Capbility in https://codereview.chromium.org/787033003,
but since it is not negotiated between host and client, that's not the right
way to make it configurable.

BUG=448131

Review URL: https://codereview.chromium.org/878573006

Cr-Commit-Position: refs/heads/master@{#313181}

The flags were disabled to suppress a build error in ffmpeg.
The error should be fixed now:
https://crrev.com/a458f95e5718918a70506be2a33ea3be55e59024

BUG=450771, 371125
TEST=emerge-mipsel-o32-generic chromeos-chrome
cbuildbot --remote mipsel-o32-generic-tot-chromium-pfq-informational

Review URL: https://codereview.chromium.org/878603005

Cr-Commit-Position: refs/heads/master@{#313180}

The idea of this change is to fire Win AX events
based on the AX tree changing, rather than based
on Blink firing the right event and mapping it to a
Win event.

Store the Win-specific AX attributes (ia role,
ia state, name, value, etc.) in a struct, and when a
node is updated, compare the old value of each attribute
to the new value and fire an appropriate event for each
one (name changed, state changed, etc.).

Handle hypertext_ similarly, and properly diff the old and new hypertext in get_oldText and get_newText.

This enables all of the "event" tests that were
previously added to pass now, and it makes live regions
work correctly in NVDA and JAWS.

BUG=447962

Committed: https://crrev.com/eb691a960bef3c2aefa38627baa7cbdde480b81e
Cr-Commit-Position: refs/heads/master@{#313052}

Review URL: https://codereview.chromium.org/848653002

Cr-Commit-Position: refs/heads/master@{#313179}