GitLab

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/058f01e..e228ba3

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

Review URL: https://codereview.chromium.org/693493005

Cr-Commit-Position: refs/heads/master@{#302163}

This CL links the it2me host to the Chrome binary on ChromeOS
behind a flag.
The following changes are made to the it2me host so that it can be run in the browser process.
1. Initializes SSL server sockets and specific CPU media features on ChromeOS startup.
2. Fixes a crash in it2me shutdown by making It2meHost owns the ChromotingHostContext.
3. Replace the blocking shutdown wait on PolicyWatcher with a callback.
   Implements policy_watcher on ChromeOS using policy services.
4. Re-use existing threads, url request context getters and policy service on ChromeOS.
5. Fixed a incorrect DCHECK regarding the color format of the frames captured on ChromeOS.

BUG=334087

Committed: https://crrev.com/54dde6f02d121ff745e66b57205583087ff720ec
Cr-Commit-Position: refs/heads/master@{#302034}

Review URL: https://codereview.chromium.org/639233002

Cr-Commit-Position: refs/heads/master@{#302162}

BUG=338884

Review URL: https://codereview.chromium.org/687833002

Cr-Commit-Position: refs/heads/master@{#302161}

Use NSWindowDidChangeOcclusionStateNotification (new in 10.9
Mavericks) to detect when the application is hidden, the window is
minimized or obscured by another window.

Send the WasHidden to the WebContentsImpl, because this will
take into account tab capture.

When RenderWidgetHostViewMac::WasHidden is called, the previous
behavior was to destroy the BrowserCompositorViewMac. This has
the undesirable effect of creating a white flash when a hidden
window is re-exposed.

To prevent this, only destroy the BrowserCompositorViewMac when
the the RenderWidgetHostViewMac is removed from all NSWindows.
Manage the state of BrowserCompositorViewMac explicitly as being
Active (previously whenever it existed), Suspended (the new state
where it is kept around, but the RWH and DFH are hidden), and
Destroyed (previously wherever it did not exist).

Move some 10.7 APIs to base because they belong there, not in
content.

BUG=310374

Review URL: https://codereview.chromium.org/684343002

Cr-Commit-Position: refs/heads/master@{#302160}

- add a protobuf for WiFi credentials
- add a ModelType for syncing WiFi credentials
- add UI for controlling whether or not WiFi credentials are synced
- add new data type to testserver

While there:
- update comment in sync_setup_overlay.js
- fix ordering of arguments to EXPECT_EQ in CheckBool
  (sync_setup_handler_unittest)
- move GetSelectableTypeNameMap from sync_setup_handler.cc to model_type.cc

BUG=chromium:422045
TEST=ProfileSync, SyncSetup, SyncModel, ModelType, NigoriUtil,
     ProtoEnumConversions, ProtoValueConversions,
     SyncEncryptionHandlerImpl

Review URL: https://codereview.chromium.org/674633002

Cr-Commit-Position: refs/heads/master@{#302159}

Consumers of this class (powerd) might want to know if it actually can
wake the system from suspend so that they can try alternatives if it
doesn't have this ability.

BUG=chrome-os-partner:31970
TBR=derat

Review URL: https://codereview.chromium.org/685503006

Cr-Commit-Position: refs/heads/master@{#302158}

There are now files in the primary locations for flac and trace-viewer.

R=brettw@chromium.org
BUG=

Review URL: https://codereview.chromium.org/690613004

Cr-Commit-Position: refs/heads/master@{#302157}

Disable data reduction proxy by deleting user level proxy prefs only if
the proxy is known to be a data_reduction_proxy. Remove duplicate
implementation in off_the_record_profile_impl.cc.

BUG=426220

Review URL: https://codereview.chromium.org/669163003

Cr-Commit-Position: refs/heads/master@{#302156}

BUG=411543

Review URL: https://codereview.chromium.org/691123002

Cr-Commit-Position: refs/heads/master@{#302155}

We should not assume that the underlying decoders
keep their state after a Seek.
So for example, the SPS/PPS should be added again.

BUG=140371

Review URL: https://codereview.chromium.org/683573002

Cr-Commit-Position: refs/heads/master@{#302154}

TBR=vsevik
BUG=

Review URL: https://codereview.chromium.org/690553004

Cr-Commit-Position: refs/heads/master@{#302153}

BUG=412150

Review URL: https://codereview.chromium.org/691883002

Cr-Commit-Position: refs/heads/master@{#302152}

Revert of Windows video capture: Remove duplicated code from GetDeviceSupportedFormats* (patchset #3 id:200001 of https://codereview.chromium.org/558503003/)

Reason for revert:
See https://code.google.com/p/chromium/issues/detail?id=428958

Original issue's description:
> Windows video capture: Remove duplicated code from GetDeviceSupportedFormats*
>
> The code for GetDeviceSupportedFormatsMediaFoundation and GetDeviceSupportedFormatsDirectShow are already implemented in video_capture_device_mf_win.cc and video_capture_device_win.cc respectively. This CL refactors the implementation so that video_capture_device_factory_win.cc can make use of that code.
>
> Committed: https://crrev.com/b6dae80f8f10b8af1e3026de9f5af4f3128e812a
> Cr-Commit-Position: refs/heads/master@{#301891}

TBR=mcasas@chromium.org,perkj@chromium.org,tommi@chromium.org,magjed@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/691143002

Cr-Commit-Position: refs/heads/master@{#302151}

we can determine whether the data_reduction_proxy was used or not for
PLT UMA.

BUG=374264

Review URL: https://codereview.chromium.org/686313002

Cr-Commit-Position: refs/heads/master@{#302150}

It's now up to the caller to avoid calls to
::Begin/FinishDrawingFrame when it makes sense to do so.

BUG=428071
R=danakj

Review URL: https://codereview.chromium.org/678113004

Cr-Commit-Position: refs/heads/master@{#302149}

Currently, touch selection quick menu buttons have the fixed size of
63x38. In some languages, texts need more space; so those buttons should
be expanded to fit the whole text.

BUG=417254

Review URL: https://codereview.chromium.org/693583002

Cr-Commit-Position: refs/heads/master@{#302148}

Previous instrumentations showed that TemplateURLService::ChangeToLoadedState
alone is responsible for 2.74 janks per hour in UI thread in Win (8.15 on Mac). I need to instrument
the code inside it to find out which part causes jank.

This is a mechanical change that adds instrumentation required to locate the
source of jankiness (i.e. a long-running fragment of code executed as a part of
the task that causes jank) in the code. See the bug for details on what kind of
jank we are after.
A number of similar CLs were landed, and none of them caused issues. They've
helped to find and fix janky code. The code of the instrumentation is highly
optimized and is not expected to affect performance. The code simply creates a
diagnostic task which is identical to ones created by PostTask or IPC message
handlers.

BUG=422460

Review URL: https://codereview.chromium.org/684333002

Cr-Commit-Position: refs/heads/master@{#302147}

Some of the functions are direct wrapper to WebContents methods. In this
patch, taken care of removing those dependencies and added extra validation to prevent run time crashes.

BUG=398263

Review URL: https://codereview.chromium.org/639703004

Cr-Commit-Position: refs/heads/master@{#302146}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/6500760..058f01e

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

Review URL: https://codereview.chromium.org/680833004

Cr-Commit-Position: refs/heads/master@{#302145}

TBR=isherman@chromium.org
BUG=428892

Review URL: https://codereview.chromium.org/682223005

Cr-Commit-Position: refs/heads/master@{#302144}

Revert of Allow POSIX callers to specify a new file's mode. (patchset #1 id:1 of https://codereview.chromium.org/689713006/)

Reason for revert:
This seems to have broken the Linux MSan test bot.

[ RUN      ] FileUtilTest.WriteFileWithMode
../../base/files/file_util_unittest.cc:2164: Failure
Value of: status.st_mode & 0777
  Actual: 384
Expected: mode
Which is: 420
[  FAILED  ] FileUtilTest.WriteFileWithMode (1 ms)

Original issue's description:
> Allow POSIX callers to specify a new file's mode.
>
> Retain the tight 0640 mode for most existing callers, but honor the user's
> umask for downloaded files.
>
> This is a re-land of https://codereview.chromium.org/639253004/.
>
> BUG=409416, 362887
> TBR=asanka,mdempsky,rvargas
>
> Committed: https://crrev.com/931e35070097af9837f79ff19f8668e45b92ed3a
> Cr-Commit-Position: refs/heads/master@{#302113}

TBR=palmer@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=409416, 362887

Review URL: https://codereview.chromium.org/693493004

Cr-Commit-Position: refs/heads/master@{#302143}

Original CL: Review URL: https://codereview.chromium.org/686503002

BUG=425996
TBR=oshima, sadrul, tbarzic

Review URL: https://codereview.chromium.org/690943003

Cr-Commit-Position: refs/heads/master@{#302142}

BUG=417090

Review URL: https://codereview.chromium.org/627573002

Cr-Commit-Position: refs/heads/master@{#302141}

Method AccountTracker::StartFetchingUserInfo is executed in browser's UI thread, and it causes jank.
In Canary 40.0.2201.2 on Win, 19.9 jph is caused by AccountTracker::StartFetchingUserInfo,
which average execution time in UI thread is 227 ms. On Mac, we have 17.7 jph and 71.1ms avg. I need to instrument
the code inside it to find out which part causes jank.

This is a mechanical change that adds instrumentation required to locate the
source of jankiness (i.e. a long-running fragment of code executed as a part of
the task that causes jank) in the code. See the bug for details on what kind of
jank we are after.
A number of similar CLs were landed, and none of them caused issues. They've
helped to find and fix janky code. The code of the instrumentation is highly
optimized and is not expected to affect performance. The code simply creates a
diagnostic task which is identical to ones created by PostTask or IPC message
handlers. The task gets created only in developer build and in Canary channel.

BUG=422460

Review URL: https://codereview.chromium.org/690073002

Cr-Commit-Position: refs/heads/master@{#302140}

BUG=

Review URL: https://codereview.chromium.org/681103002

Cr-Commit-Position: refs/heads/master@{#302139}

Previously, when generate_v14_compatible_resources.py encountered an
invalid XML file, it showed an error message that didn't tell which
file it was processing. E.g.:

  Traceback (most recent call last):
    ...
    File "/usr/lib/python2.7/xml/dom/expatbuilder.py", line 207, in parseFile
      parser.Parse(buffer, 0)
  ExpatError: unbound prefix: line 6, column 0

Now it also prints the name of the file that couldn't be parsed. E.g.

  Traceback (most recent call last):
    ...
    File "/usr/lib/python2.7/xml/dom/expatbuilder.py", line 207, in parseFile
      parser.Parse(buffer, 0)
  ExpatError: unbound prefix: line 6, column 0
  Failed to parse XML file: chrome/android/res/layout/my_layout.xml

Review URL: https://codereview.chromium.org/661433005

Cr-Commit-Position: refs/heads/master@{#302138}

BUG=423229

Review URL: https://codereview.chromium.org/686803003

Cr-Commit-Position: refs/heads/master@{#302137}

Push Notifications not only support Ask already, it is the default value, so this should match that reality.

BUG=425698

Review URL: https://codereview.chromium.org/689993002

Cr-Commit-Position: refs/heads/master@{#302136}

In Ozone the touch events are received in device coordinates. This
change consolidates the different transformations into one function and
makes it use the touchscreen size when computing the transformation for
Ozone and the framebuffer size for X11.

BUG=425258
TESTS=Manually verified that X11 in single, mirror, extended and docked
mode work with external touchscreen.

Review URL: https://codereview.chromium.org/688183002

Cr-Commit-Position: refs/heads/master@{#302135}

TBR=yurys
BUG=

Review URL: https://codereview.chromium.org/686643003

Cr-Commit-Position: refs/heads/master@{#302134}

This reverts commit ea73a785.

TBR=pstew@chromium.org
BUG=

Review URL: https://codereview.chromium.org/655833007

Cr-Commit-Position: refs/heads/master@{#302133}

Namely, -Winconsistent-missing-override. No behavior change.

BUG=428099

Review URL: https://codereview.chromium.org/687453006

Cr-Commit-Position: refs/heads/master@{#302132}

BUG=421422

Review URL: https://codereview.chromium.org/685403004

Cr-Commit-Position: refs/heads/master@{#302131}

Revert of Add Search Service in Enhanced Bookmark Bridge (patchset #6 id:100001 of https://codereview.chromium.org/637323005/)

Reason for revert:
Broke compilation in http://build.chromium.org/p/chromium.gpu/builders/Android%20Debug%20%28Nexus%207%29/builds/14002

Original issue's description:
> Add Search Service in Enhanced Bookmark Bridge
>
> A Search client is created during initialization of
> EnhancedBookmarkBridge. And when the bridge destructs, search service is
> also destroyed.
>
> BUG=415774
>
> Committed: https://crrev.com/e52346f194e9f2d097b4f6d3fbfa359eca7c92df
> Cr-Commit-Position: refs/heads/master@{#302107}

TBR=tedchoc@chromium.org,danduong@chromium.org,kkimlabs@chromium.org,noyau@chromium.org,lpromero@chromium.org,ianwen@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=415774

Review URL: https://codereview.chromium.org/688883002

Cr-Commit-Position: refs/heads/master@{#302130}

This is needed to make more enums work with GlEnumToString, particularly
CONTEXT_LOST_KHR.

BUG=428332

Review URL: https://codereview.chromium.org/690753005

Cr-Commit-Position: refs/heads/master@{#302129}

Revert of Wait 60s instead of 30s for browser to launch. (patchset #1 id:1 of https://codereview.chromium.org/691683002/)

Reason for revert:
Didn't help -- stuck subprocesses are still showing up.

Original issue's description:
> Wait 60s instead of 30s for browser to launch.
>
> There's a guess that the browser may be launching extremely slowly on
> some bots and that when the browser main process is killed it's
> leaking sub-processes that are in the middle of being spawned.
> Increase the timeout to try to confirm this hypothesis.
>
> BUG=424024
> R=tonyg@chromium.org
>
> Committed: https://crrev.com/7ef2c0036f70bdcc59dd7c8f23c8b7c5a3f114d9
> Cr-Commit-Position: refs/heads/master@{#302087}

TBR=tonyg@chromium.org,nednguyen@google.com
NOTREECHECKS=true
NOTRY=true
BUG=424024

Review URL: https://codereview.chromium.org/667733004

Cr-Commit-Position: refs/heads/master@{#302128}

* It was broken after the git switch (it used SVN internally)
* Generate Pepper indexes to
  c-api-{channel} instead of pepper_{channel}/c

This relies on using Doxygen 1.7.6.1, which was installed by default on Lucid. Doxygen 1.8 does not work, and maybe never will.

You can set the DOXYGEN environment variable to point to a good version of doxygen.

No trybots because it is a documentation generation only change. I tested locally to ensure this wouldn't break stuff.

BUG=none
TEST=local
NOTRY=true
R=sbc@chromium.org

Review URL: https://codereview.chromium.org/591683002

Cr-Commit-Position: refs/heads/master@{#302127}

As discovered by japhet@, if we navigate cross-site in a browser test using the '/cross-site/' redirector syntax, the initial URL is relative to the embedded_test_server, which happens to also be the SiteInstance of the top-level page. This causes an extra process swap, which is unnecessary and can create unexpected effects in the test. I've added GetURL(hostname, path) method to EmbeddedTestServer to give us the proper URL directly without following a redirect path and this CL updates the tests to use this new method.

BUG=418236

Review URL: https://codereview.chromium.org/688363002

Cr-Commit-Position: refs/heads/master@{#302126}

Review URL: https://codereview.chromium.org/690713002

Cr-Commit-Position: refs/heads/master@{#302125}

BUG=425601

Review URL: https://codereview.chromium.org/660563004

Cr-Commit-Position: refs/heads/master@{#302124}