GitLab

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/0c6b223..ed24ca8

TBR=eustas@chromium.org
BUG=

Review URL: https://codereview.chromium.org/611993003

Cr-Commit-Position: refs/heads/master@{#297306}

After we show and activate a popup window from a user gesture, we should
always focus the web contents.

BUG=411322

Review URL: https://codereview.chromium.org/599763003

Cr-Commit-Position: refs/heads/master@{#297305}

For example "cr_cronet.py test -f *Quic*" will run just Quic tests.

BUG=

Review URL: https://codereview.chromium.org/601863002

Cr-Commit-Position: refs/heads/master@{#297304}

This patch ensures that we always activate when we become invisible.
The reason being is that when we come back to being visible again,
we don't need to rasterize both active and pending trees. We can
just rasterize what used to be the pending tree.

Since we set RequiresHighResToDraw on becoming visible, this also
ensures that we don't checkerboard on becoming visible.

BUG=410000, 417598
R=danakj, enne, brianderson

Review URL: https://codereview.chromium.org/605823002

Cr-Commit-Position: refs/heads/master@{#297303}

Also refactors some common code into a method in MetricsService.

BUG=400357

Review URL: https://codereview.chromium.org/612883003

Cr-Commit-Position: refs/heads/master@{#297302}

I moved all tests to Win Tests x64 (1) in preparation for removing 2 & 3.

BUG=414808
R=maruel@chromium.org

Review URL: https://codereview.chromium.org/608363004

Cr-Commit-Position: refs/heads/master@{#297301}

When the download shelf is closed, it's still in the view hierarchy with height
of 0. To make it not appear in the AX hierarchy, we hide the view after the
close animation is finished.

BUG=417698

Review URL: https://codereview.chromium.org/615663002

Cr-Commit-Position: refs/heads/master@{#297300}

App_shell doesn't have a ComponentExtensionResourceManager implementation, so we need to check if it's available before using it.

BUG=352293

Review URL: https://codereview.chromium.org/615703002

Cr-Commit-Position: refs/heads/master@{#297299}

Tests added in this CL:
TestLoadAbortChromeExtensionURLWrongPartition
TestLoadAbortIllegalChromeURL
TestLoadAbortIllegalFileURL
TestLoadAbortIllegalJavaScriptURL
TestLoadAbortInvalidNavigation
TestLoadAbortNonWebSafeScheme
TestLoadProgressEvent

BUG=352293

Review URL: https://codereview.chromium.org/601553005

Cr-Commit-Position: refs/heads/master@{#297298}

BUG=none
TEST=none
R=keybuk@chromium.org

Review URL: https://codereview.chromium.org/580643002

Cr-Commit-Position: refs/heads/master@{#297297}

BUG=

Review URL: https://codereview.chromium.org/611013003

Cr-Commit-Position: refs/heads/master@{#297296}

BUG=224786

Review URL: https://codereview.chromium.org/605243004

Cr-Commit-Position: refs/heads/master@{#297295}

BUG=418767
TBR=asargent@chromium.org

Review URL: https://codereview.chromium.org/613843004

Cr-Commit-Position: refs/heads/master@{#297294}

Also tracks the number of active frames in a SiteInstance, rather than
the number of views.  These changes are necessary for tracking the
pending swap-out state of frames.

This involves a lot of test cleanup as well.

BUG=418264
TEST=Subframes can swap out.  (No visible behavior change.)

Review URL: https://codereview.chromium.org/606113005

Cr-Commit-Position: refs/heads/master@{#297293}

Use chrome.test.sendMessage instead of chrome.test.fail so we
short-circuit instead of timeout on javascript failures where we're
waiting for a go signal.

BUG=388958
R=armansito@chromium.org

Review URL: https://codereview.chromium.org/593163002

Cr-Commit-Position: refs/heads/master@{#297292}

Linux. Linux generates an X error when it receives a request for 0x0 bounds

BUG=416455
TEST=Manual, see bug

Review URL: https://codereview.chromium.org/610263002

Cr-Commit-Position: refs/heads/master@{#297291}

BUG=416722
R=gdk@chromium.org

Review URL: https://codereview.chromium.org/604613002

Cr-Commit-Position: refs/heads/master@{#297290}

BUG=

Review URL: https://codereview.chromium.org/580593002

Cr-Commit-Position: refs/heads/master@{#297289}

BUG=384876
R=bauerb@chromium.org, markusheintz@chromium.org, vabr@chromium.org, rlp@chromium.org
TEST=no functional changes

Review URL: https://codereview.chromium.org/596613002

Cr-Commit-Position: refs/heads/master@{#297288}

On some lower-end platforms, 64-bit integer division is
emulated or is significantly slower than 32-bit integer
division even if not emulated.

When converting ns to us in TimeTicks::Now(), we don't
need the range of a 64-bit integer because we are starting
with a 32-bit integer and making it smaller. There is no
need to convert it to 64-bits before the division.

BUG=none

Review URL: https://codereview.chromium.org/600013003

Cr-Commit-Position: refs/heads/master@{#297287}

BUG=402671
TEST=None

Review URL: https://codereview.chromium.org/590263002

Cr-Commit-Position: refs/heads/master@{#297286}

This gets rid of raw close() call usign base::ScopedFD.
Ownership of FDs become clearer.

This is a preparation for kiling base::FileDescriptor.

R=jam@chromium.org, agl@chromium.org
BUG=415294

Review URL: https://codereview.chromium.org/602193004

Cr-Commit-Position: refs/heads/master@{#297285}

The previous behaviour was to hide it.

Also, update some icon resources.

Note: this does not yet work well with easy signin.

BUG=406234

Review URL: https://codereview.chromium.org/608693006

Cr-Commit-Position: refs/heads/master@{#297284}

BUG=418258
TBR=dgozman@chromium.org

Review URL: https://codereview.chromium.org/616633003

Cr-Commit-Position: refs/heads/master@{#297283}

Cleanup patch to address below mentioned issue -

[*] Change const char* foo to const char foo[]

BUG=416468

Review URL: https://codereview.chromium.org/587403002

Cr-Commit-Position: refs/heads/master@{#297282}

r296874 regressed initial tooltip sizing on Aura.
(Widget::SetContentsView sizes the view to the widget)

Reorder the View sizing to after the Widget init.

BUG=418456
TEST=The first Linux Aura tooltip is sized correctly.
R=sky@chromium.org

Review URL: https://codereview.chromium.org/613753003

Cr-Commit-Position: refs/heads/master@{#297281}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/c91b5c8..0c6b223

TBR=eustas@chromium.org
BUG=

Review URL: https://codereview.chromium.org/617453003

Cr-Commit-Position: refs/heads/master@{#297280}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/656f2cd..c5e15a1

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

Review URL: https://codereview.chromium.org/618463002

Cr-Commit-Position: refs/heads/master@{#297279}

BUG=418748
TBR=asanka@chromium.org,

Review URL: https://codereview.chromium.org/613853002

Cr-Commit-Position: refs/heads/master@{#297278}

BUG=418684

Review URL: https://codereview.chromium.org/615613002

Cr-Commit-Position: refs/heads/master@{#297277}

This is no longer used after removing in-process GPU service support
for CHROMIUM_image.

BUG=418553

Review URL: https://codereview.chromium.org/604833005

Cr-Commit-Position: refs/heads/master@{#297276}

Seems to happen in Debug mode. Turn these off until we can find a fix.

BUG=417824,416594

Review URL: https://codereview.chromium.org/604983002

Cr-Commit-Position: refs/heads/master@{#297275}

After crrev.com/295851 the PipeMessagingChannel verifies that incoming
messages are dictionaries and then passes them as base::Value to
OnMessage() and then OnMessage() implementation static_cast<> them to
base::DictionaryValue. This is fragile, and also it shouldn't be
PipeMessagingChannel's responsibility to verify format of the messages
it reads. Moved the check to OnMessage() implementations.

Review URL: https://codereview.chromium.org/615543004

Cr-Commit-Position: refs/heads/master@{#297274}

Revert of [Android] Disable enhanced bookmark opt-in flag. (patchset #1 id:1 of https://codereview.chromium.org/610443002/)

Reason for revert:
M39 branched, so re-enable the flag on trunk.

Original issue's description:
> [Android] Disable enhanced bookmark opt-in flag.
>
> We won't launch this feature for M39, so disable public opt-in flag.
>
> BUG=416664
>
> Committed: https://crrev.com/5da932e693169900a4a4be346446788aa1944eeb
> Cr-Commit-Position: refs/heads/master@{#296825}

TBR=yfriedman@google.com,tedchoc@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=416664

Review URL: https://codereview.chromium.org/612203002

Cr-Commit-Position: refs/heads/master@{#297273}

BUG=404771

Review URL: https://codereview.chromium.org/595213003

Cr-Commit-Position: refs/heads/master@{#297272}

If a previous mojo_shell run exited uncleanly, there may
be a unix domain socket sitting at the path that
IncomingConnectionListener wants to own and listen at.
This change destroys it, if possible, and creates a new
one.

BUG=407782
TEST=new unit test, run mojo_shell multiple times in a row with --enable-external-applications

Review URL: https://codereview.chromium.org/611833005

Cr-Commit-Position: refs/heads/master@{#297271}

Pure HTML/CSS changes. No tests affected.

BUG=413337,416228

Review URL: https://codereview.chromium.org/612823005

Cr-Commit-Position: refs/heads/master@{#297270}

Save the client base JID for authentication in case it differs from the email (for accounts non-Google email). Keep a separate email field for use in user-facing UIs, logs, etc. Using the existing host_owner field for the base JID guarantees the correct *functional* behavior (i.e. authentication-wise) if the new config is used with an older host (the only difference is aesthetic - in that older host some UI/logs may show an uglier address instead of the user email).

BUG=224742

Review URL: https://codereview.chromium.org/595063005

Cr-Commit-Position: refs/heads/master@{#297269}

clang's auto-vectorized C version performs better according to the
Chrome Performance Dashboard.  Searching back through the logs, this
occurred when we switched over to clang by default.

We could try to microoptimize further, but it's less of a maintenance
burden to just let the compiler do its thing!

The main reason the clang version is faster is it does 2x 128bit
operations per loop. Simply copying these optimization yields ~97%
similar performance, but  the SIMD code a bit gnarlier. As such I
choose to simply use the C variant when clang is present.

BUG=none
TEST=none

Review URL: https://codereview.chromium.org/599693002

Cr-Commit-Position: refs/heads/master@{#297268}

Remove page_set argument from benchmark.CreateExpectations to limit the access to page_set (user_story)

BUG=418278

Review URL: https://codereview.chromium.org/614783002

Cr-Commit-Position: refs/heads/master@{#297267}