GitLab

BUG=560769

Review URL: https://codereview.chromium.org/1477133002

Cr-Commit-Position: refs/heads/master@{#362936}

The xdg_shell allows clients to turn a wl_surface into a "real window"
which can be dragged, resized, stacked, and moved around by the
user. This interface is suited towards traditional desktop
environments and is what most wayland clients use today.

This patch adds minimal support for this interface. Requests such
as move and resize will be added by follow up patches.

BUG=549781

Review URL: https://codereview.chromium.org/1411403007

Cr-Commit-Position: refs/heads/master@{#362935}

It's unclear what the purpose of this test is, but it's creating busy
work for folks working on blink features, so I'll disable this while
someone figures this out.

BUG=564765
TBR=timvolodine@chromium.org

Review URL: https://codereview.chromium.org/1495793003

Cr-Commit-Position: refs/heads/master@{#362934}

BUG=555325

Review URL: https://codereview.chromium.org/1444493002

Cr-Commit-Position: refs/heads/master@{#362933}

Reland of Ship promise rejection events (patchset #1 id:1 of https://codereview.chromium.org/1498703002/ )

Reason for revert:
i'll disable the test instead

Original issue's description:
> Revert of Ship promise rejection events (patchset #2 id:20001 of https://codereview.chromium.org/1496453002/ )
>
> Reason for revert:
> This is breaking SystemWebViewShellLayoutTests on Android Webview L. See crbug/564765.
>
> Original issue's description:
> > Ship promise rejection events
> >
> > I2S: https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/AmChp_uVpXU
> >
> > BUG=495801
> > R=philipj@opera.com
> >
> > Committed: https://crrev.com/98bd85306f3a170fa465c3f5a754ac2519a36be1
> > Cr-Commit-Position: refs/heads/master@{#362695}
>
> TBR=philipj@opera.com,jochen@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=495801
>
> Committed: https://crrev.com/f3e241833fad43bec928922e949fa8cd2f4d84ff
> Cr-Commit-Position: refs/heads/master@{#362891}

TBR=philipj@opera.com,ianwen@chromium.org,khushalsagar@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=495801

Review URL: https://codereview.chromium.org/1495723003

Cr-Commit-Position: refs/heads/master@{#362932}

Any unregistration is (already) taken care of by Oilpan.

R=haraken
BUG=

Review URL: https://codereview.chromium.org/1485353006

Cr-Commit-Position: refs/heads/master@{#362931}

The application ID identifies the general class of applications to which
the surface belongs. Chromium can use this to group multiple surfaces
together, or to determine how to launch a new application.

BUG=549781
TEST=exo_unittest --gtest_filter=ShellSurfaceTest.SetApplicationId

Review URL: https://codereview.chromium.org/1491843002

Cr-Commit-Position: refs/heads/master@{#362930}

Use FetchFormDataConsumerHandle in Response::create to simplify the code.

This CL also includes some minor refactoring.
 - Rename "responseInit" with "init" for the simplicity
 - Replace 0 with nullptr
 - Wrap comments in 80 cols

BUG=564479

Review URL: https://codereview.chromium.org/1490243002

Cr-Commit-Position: refs/heads/master@{#362929}

This restriction was lifted in
https://codereview.chromium.org/1257823006. But ES3 still has this
restriction. Move this check after format validation before data
validation due to format is used for data validation.

BUG=295792
TEST=deqp/functional/gles3/negativetextureapi.html

Review URL: https://codereview.chromium.org/1489253003

Cr-Commit-Position: refs/heads/master@{#362928}

Use the chain tool to compute how a given color profile transforms white
input to XYZ PCS (profile connection space).

No change in behavior, no new tests. Tested with an iMac P3 profile: the
returned XYZ result is exactly D50 white: [0.9642, 1.0000, 0.8249].

BUG=564355

Review URL: https://codereview.chromium.org/1494473003

Cr-Commit-Position: refs/heads/master@{#362927}

These methods are not actually used. Methods of identical names/types
exist in ExclusiveAccessContext. BrowserView inherits both BrowserWindow
and ExclusiveAccessContext and double-overrides this method. It is only
virtual-dispatched through ExclusiveAccessContext, so the BrowserWindow
virtual methods do not need to exist.

Updated tests that were depending on this old interface.

BUG=352425

Review URL: https://codereview.chromium.org/1480183002

Cr-Commit-Position: refs/heads/master@{#362926}

This is necessary to pass 'gn check' on Windows because
components/nacl/loader/nacl_main_platform_delegate_win.cc uses
the sandbox/win/src/sandbox.h header.  The direct dependency is
not strictly necessary on Mac because its implementation of
NaClMainPlatformDelegate::EnableSandbox uses the wrapper
interface content::InitializeSandbox, nor on Linux because the
zygote handles it.  But it doesn't hurt.

BUG= 512869
R=bradnelson@chromium.org, mseaborn@chromium.org

Review URL: https://codereview.chromium.org/1498673002

Cr-Commit-Position: refs/heads/master@{#362925}

Much of the functionality in CompositorWorkerManager
was related to shared compositor thread state. This
patch moves that to CompositorWorkerSharedState, an
implementation detail of CompositorWorkerThread.

The shared state also now uses the real compositor
thread.

The existing compositor worker layout tests needed
to be moved to a virtual test suite since they now
depend on having a compositor thread.

BUG=430155

Committed: https://crrev.com/d710cbc079b51b819349203ee408f36dc056055e
Cr-Commit-Position: refs/heads/master@{#362590}

Review URL: https://codereview.chromium.org/1449953002

Cr-Commit-Position: refs/heads/master@{#362924}

Makes shell_util more modular and breaks it up a bit. RegistryEntry is a
generally useful API.

BUG=419972

Review URL: https://codereview.chromium.org/623903002

Cr-Commit-Position: refs/heads/master@{#362923}

I'm hopeful this will work because then it can simplify the scoring code.

The experiment for the HUP-like-mode in HQP seemed promising
https://code.google.com/p/chromium/issues/detail?id=490812
(i.e., neutral).  Let's see if we can disable HUP if that mode is enabled.

We do lose some features.  For instance, HUP culls
redirects.  HQP does not.  This means that if HUP
manages to find three good sites post-redirect-culling,
HQP may not be able to because it may have some
equivalent (redirect) URLs laying around.

Note that HUP still uses its database implicitly.
This experiment will not test whether we can remove
the database entirely.  The HUP still uses its database
for identify typed hosts or shorter matches for instance.

BUG=

Review URL: https://codereview.chromium.org/1498473003

Cr-Commit-Position: refs/heads/master@{#362922}

aura::Windows are needed for proper event-handling.

BUG=549781
TEST=exo_unittest --gtest_filter=*SurfaceTest.*

Review URL: https://codereview.chromium.org/1490223002

Cr-Commit-Position: refs/heads/master@{#362921}

BUG=none
TEST=none

Review URL: https://codereview.chromium.org/1496843002

Cr-Commit-Position: refs/heads/master@{#362920}

This layout test tests createPattern operates synchronously with an
animated image. The test was removed because it was using SVG animation
which was flacky. This CL changes the SVG animation to CSS animation.

BUG=316730

Review URL: https://codereview.chromium.org/1493883002

Cr-Commit-Position: refs/heads/master@{#362919}

BUG=NONE
NOTRY=true

Review URL: https://codereview.chromium.org/1495613003

Cr-Commit-Position: refs/heads/master@{#362918}

Cleanup <image> type handling in two ways:
- to handle list-style-image, border-image-source and
webkit-mask-box-image-source simply reuse parseFillImage.
- to make isGeneratedImageValue more reusable use a CSSValueID parameter.

BUG=499780

Review URL: https://codereview.chromium.org/1498553002

Cr-Commit-Position: refs/heads/master@{#362917}

This is a modified version of bzanotti@'s CL:
https://codereview.chromium.org/1496573002/

We now clear the access token anytime the sync engine shuts down.

BUG=563416

Review URL: https://codereview.chromium.org/1498663002

Cr-Commit-Position: refs/heads/master@{#362916}

. An app can request an AcceleratorRegistrar from the window manager, and
  install some accelerators. The app must install an AcceleratorHandler to
  process these accelerators.
. The AcceleratorRegistrar implementation in the window manager makes sure
  that the accelerator ids from different apps do not conflict with each
  other.
. When an AcceleratorHandler is disconnected, the accelerators its supposed
  to handle are automatically unregistered from mus.

BUG=548363

Review URL: https://codereview.chromium.org/1488713002

Cr-Commit-Position: refs/heads/master@{#362915}

We don't need to keep them because parseAttribute knows their old values.

BUG=561251

Review URL: https://codereview.chromium.org/1490973002

Cr-Commit-Position: refs/heads/master@{#362914}

This is a defensive change - it seems unlikely that CSS Variables
can persist beyond the style sheets that create them, but it's also
easy and cheap to ensure that resolved CSS variables pin their
backing strings.

It's worth noting that resolved CSS Variables can *only* reference
either their own string, or strings from other *already resolved*
variables.

BUG=465126

Review URL: https://codereview.chromium.org/1498713002

Cr-Commit-Position: refs/heads/master@{#362913}

This CL simplifies the code.  This CL has no behavior changes.

BUG=561251

Review URL: https://codereview.chromium.org/1496773002

Cr-Commit-Position: refs/heads/master@{#362912}

In general, we want to enable external animation system for blink here: https://codereview.chromium.org/1308053006/

Make sure that HtmlViewer always uses the new animation system for CC and Blink (cc::AnimationHost).

Eventually, we will delete LayerSettings everywhere.

BUG=394777

Review URL: https://codereview.chromium.org/1481343002

Cr-Commit-Position: refs/heads/master@{#362911}

It proves to be really hard to change the first content users see in
bookmark manager on both phones and tablets. This is because EBManager
does not initialize itself the same as other native pages.

This CL proposes a different approach about initializing the bookmark
manager. In EBManager's constructor, it will always initialize itself in
the loading mode. And after the constructor returns, EBPage and
EBActivity call updateForUrl() with an initial url that is determined by
some complicated logics in EBUtils. This procedure aligns with
NativePageFactory's logic.

To achieve this goal, several changes have been made:
1. Loading state no longer carries a url. Instead mInitialUrl is created
to catch that info.
2. Detecting bookmark model loaded state has been made simple by
introducing doAfterBookmarkModelLoaded in BookmarksBridge.
3. Now both EBActivity and EBPage need to call EBmanager#updateForUrl.
4. EBPage no longer calls tab#loadUrl if this url change is originally
initiated by tab itself.

BUG=554595

Review URL: https://codereview.chromium.org/1440623004

Cr-Commit-Position: refs/heads/master@{#362910}

The tests started failing consistently for Lollipop 64 bit Tester since it started running.

TBR=bajones@chromium.org

BUG=550292

Review URL: https://codereview.chromium.org/1499563002

Cr-Commit-Position: refs/heads/master@{#362909}

BUG=565071

NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1493553006

Cr-Commit-Position: refs/heads/master@{#362908}

Reland of Tests: Simplify SequencedWorkerPoolOwner, call Shutdown on destructor. (patchset #1 id:1 of https://codereview.chromium.org/1496493004/ )

Reason for revert:
Relanding, no recovery in

https://build.chromium.org/p/chromium.mac/builders/iOS_Simulator_%28dbg%29/builds/32878

Original issue's description:
> Revert of Tests: Simplify SequencedWorkerPoolOwner, call Shutdown on destructor. (patchset #4 id:60001 of https://codereview.chromium.org/1417353006/ )
>
> Reason for revert:
> Suspected for persistent failures on ios_net_unittests
>
> - RoundTripTestCookieStore/CookieStoreTest/0.TestNonDottedAndTLD
> - CookieStoreIOS/CookieStoreTest/0.TestNonDottedAndTLD
>
> Failing since https://build.chromium.org/p/chromium.mac/builders/iOS_Simulator_%28dbg%29/builds/32876
>
> Errors like
>
> ../../net/cookies/cookie_store_unittest.h:536: Failure
> Value of: this->SetCookie(cs.get(), url, "a=1; domain=com")
>   Actual: false
> Expected: true
>
> (It's a tricky one! Sorry if this is not it)
>
> Original issue's description:
> > Tests: Simplify SequencedWorkerPoolOwner, call Shutdown on destructor.
> >
> > Also ports remaining tests using raw SWPs to use SWPOwner.
> >
> > BUG=450228
> >
> > Committed: https://crrev.com/fc939726c283e34112d9bc845a39460410fe9cd9
> > Cr-Commit-Position: refs/heads/master@{#362805}
>
> TBR=phajdan.jr@chromium.org,jam@chromium.org,brettw@chromium.org,tommycli@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=450228
>
> Committed: https://crrev.com/d7365d9aaa55e88d9ecca49eb2158b62d1190f09
> Cr-Commit-Position: refs/heads/master@{#362890}

TBR=phajdan.jr@chromium.org,jam@chromium.org,brettw@chromium.org,tommycli@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=450228

Review URL: https://codereview.chromium.org/1500503002

Cr-Commit-Position: refs/heads/master@{#362907}

This patch has the tests use a temporary download directory so that the download directory won't be full with downloaded files and the flakiness on trybots should be fixed.

And this patch also re-enable the tests except SimultaneousIncognitoAndNormalDownloads, which was flakiness due to downloaded files in the download directory.

This fix had been committed once (crrev.com/f01a43dd) but reverted (crrev.com/c96cd8bc) due to SimultaneousIncognitoAndNormalDownloads failure. But this failure is caused by an other reason. I'll fix it and re-enable separately.

BUG=560329
TEST=browser_tests pass both on trybots and locally
TBR=asanka@chromium.org, tapted@chromium.org

Review URL: https://codereview.chromium.org/1490233006

Cr-Commit-Position: refs/heads/master@{#362906}

Revert of mojo-apptests: Fix detecting failure when no test ran. (patchset #2 id:20001 of https://codereview.chromium.org/1499463002/ )

Reason for revert:
Speculative revert as cq is failing in an unusual way when running mojo_apptests:
'module' object has no attribute 'should_start_xvfb'

from https://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_rel_ng/builds/149295/steps/mojo_apptests%20%28without%20patch%29/logs/stdio

Original issue's description:
> mojo-apptests: Fix detecting failure when no test ran.
>
> BUG=none
> R=msw@chromium.org
>
> Committed: https://crrev.com/c36c6282445549860ff359b3a9bc1515a8566e1f
> Cr-Commit-Position: refs/heads/master@{#362825}

TBR=msw@chromium.org,sadrul@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=none

Review URL: https://codereview.chromium.org/1495783002

Cr-Commit-Position: refs/heads/master@{#362905}

https://chromium.googlesource.com/chromium/src/+/0d93c1a79

BUG=474535
TBR=dcheng@chromium.org

Review URL: https://codereview.chromium.org/1494053002 .

Cr-Commit-Position: refs/heads/master@{#362904}

https://chromium.googlesource.com/chromium/src/+/7b3f9b2b6

BUG=315277
TBR=tkent@chromium.org

Review URL: https://codereview.chromium.org/1498733002 .

Cr-Commit-Position: refs/heads/master@{#362903}

R=piman@chromium.org
BUG=559357,552075

Review URL: https://codereview.chromium.org/1472483002

Cr-Commit-Position: refs/heads/master@{#362902}

The test was flaky on OSX, but it has been failing consistently since around v8 4.9.75 roll.

BUG=520199
TBR=aboxhall@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/1494873002

Cr-Commit-Position: refs/heads/master@{#362901}

Following the new visual spec, volume-related UI components are updated as follows:
- Volume slider is always shown on the right-bottom corner. It used to be shown as pop-up.
- Clicking volume button toggles mute status, rather than opening volume slider.
- minimum width is expanded to fit time slider and volume slider in a line.

BUG=488229,554448
TEST=manually

Review URL: https://codereview.chromium.org/1491543002

Cr-Commit-Position: refs/heads/master@{#362900}

Cr-Commit-Position: refs/heads/master@{#362899}

|change_acked_callback_| isn't set or used by anything. So remove it.

BUG=none
R=sky@chromium.org

Review URL: https://codereview.chromium.org/1492183002

Cr-Commit-Position: refs/heads/master@{#362898}

This breaks the former BadClockBlockingPage into three parts:

1. security_interstitials::BadClockUI handles strings and commands.

2. security_interstitials::ControllerClient abstracts the functionality
   provided by the embedder to act on commands. ChromeControllerClient
   implements this functionality for Chrome. BadClockUI has a
   ControllerClient to ask to do its bidding.

3. BadClockBlockingPage implements InterstitialPageDelegate and has a
   BadClockUI, which it asks to handle strings, commands, and everything
   else not platform-specific.

(BadClockBlockingPage has a BadClockUI has a ControllerClient.)

Other embedders will:
- make their own analogs to the BadClockBlockingPage
- implement their own PlatformControllerClient that inherits from
  ControllerClient
- share the BadClockUI class

I also added security_interstitials::common_string_utils to hold string-
manipulation methods that I expect to reuse in the future for other
classes that are componentized.

BUG=488673

Review URL: https://codereview.chromium.org/1481213003

Cr-Commit-Position: refs/heads/master@{#362897}