GitLab

Robolectric allows us to run tests that use Android code in the host
JVM. This means faster and more stable tests since we don't have to
deal with a device. This CL has the build file changes to build the
unit test suite, and two example tests that have been converted to
run with Robolectric.

BUG=448030

Review URL: https://codereview.chromium.org/994283006

Cr-Commit-Position: refs/heads/master@{#322512}

Summary of changes available at:
https://chromium.googlesource.com/v8/v8/+log/e4dfe3b9..b08c7327

Please follow these instructions for assigning/CC'ing issues:
https://code.google.com/p/v8-wiki/wiki/TriagingIssues

TBR=hablich@chromium.org,machenbach@chromium.org,yangguo@chromium.org,vogelheim@chromium.org

Review URL: https://codereview.chromium.org/1037143002

Cr-Commit-Position: refs/heads/master@{#322511}

currently this flag is not used, but we'll soon flip it to off for desktop linux.

BUG=468045

Review URL: https://codereview.chromium.org/1012223002

Cr-Commit-Position: refs/heads/master@{#322510}

If someone suspends their computer and resumes more than a minute
later, the browser io thread may tick before the audio thread can
update the last tick time ==> crash :(

The solution is to disable hang detection during suspend and only
reenable it once resume completes; at which point we can update
the tick count. If we're actually hung, we'll crash after the next
interval elapses.

BUG=471046
TEST=none

Review URL: https://codereview.chromium.org/1036933003

Cr-Commit-Position: refs/heads/master@{#322509}

BUG=

Review URL: https://codereview.chromium.org/1036093002

Cr-Commit-Position: refs/heads/master@{#322508}

Add targets for epoll and chromium versions of the QUIC
client and server.

Review URL: https://codereview.chromium.org/1036023002

Cr-Commit-Position: refs/heads/master@{#322507}

This is temporary while we are adding runtime detection.

BUG=458645
TEST=4k and 1080p playback

Review URL: https://codereview.chromium.org/1016283003

Cr-Commit-Position: refs/heads/master@{#322506}

Default was recently changed to reject in:
https://codereview.chromium.org/955383003/

R=lcwu@chromium.org
BUG=internal b/19954002

Review URL: https://codereview.chromium.org/1041513002

Cr-Commit-Position: refs/heads/master@{#322505}

For extern comments, line wrapping sometimes has indentation. For instance:
/**
 * @param {FooBar} This is some foobar
 *     that has the comment wrapped.
 */
We should support this.  Add an optional parameter to the Code Comment method
to allow for comment indentation.

BUG=461039

Review URL: https://codereview.chromium.org/1010603007

Cr-Commit-Position: refs/heads/master@{#322504}

We redesigned the GpuMemoryBuffer interface to handle multiple buffers. These changes started with the purpose of making use of new EGL_LINUX_DMA_BUF_EXT capabilities[0] and followed by an attempt to change the high level API [1].

[0] https://code.google.com/p/chromium/issues/detail?id=439520
[1] https://codereview.chromium.org/962723002/

The future steps are:
- Add new GPUMemoryBuffer format gfx::GpuMemoryBuffer::YUV_420
  StrideInBytes() should take index as input
- Implement multi buffer support on GpuMemoryBufferFactoryOzoneNativeBuffer
  Change BufferToPixmapMap as std::map<std::pair<uint32_t, uint32_t>, ScopedVector<NativePixmap> > so that we can create multiple pixmaps(file descriptors) for single GpuMemoryBufferId
- Add support for multiple pixmaps in GLImageLinuxDMABuffer
- Implement support for GpuMemoryBufferImplSharedMemory
  Change GpuMemoryBufferHandle to contain std::vector<base::SharedMemoryHandle>
- Add support for GLImageSharedMemory and GLImageMemory
- Look for possible use cases of multiple buffers on Android, Mac, and Win platforms

The new functions added are as below:

class GFX_EXPORT GpuMemoryBuffer {
  // Maps each plane of the buffer into the client's address space so it can be
  // written to by the CPU. A pointer to plane K is stored at index K-1 of the
  // |data| array. This call may block, for instance if the GPU needs to finish
  // accessing the buffer or if CPU caches need to be synchronized. Returns
  // false on failure.
  virtual bool Map(void** data) = 0;

  // Fills the stride in bytes for the each plane of the buffer. The stride of
  // plane K is stored at index K-1 of the |stride| array.
  virtual void GetStride(uint32* stride) const = 0;
};

The classes effected by this change is as below(except the test implementations). Currently none of them supports multiple buffers as the format doesn't exist. They are going to be implemented per platform in future CLs.

                                +---------------+
                                |GpuMemoryBuffer|
                                +------+--------+
                                       |
                              +--------+----------+
                              |GpuMemoryBufferImpl|
                              ++----+-----+------++
                               |    |     |      |
                               |    |     |      |
+------------------------------++   |     |   +--+-------------------------+
|GpuMemoryBufferImplSharedMemory|   |     |   |GpuMemoryBufferImplIOSurface|
+-------------------------------+   |     |   +----------------------------+
                                    |     |
+-----------------------------------++  +-+-------------------------------+
|GpuMemoryBufferImplOzoneNatiVeBuffer|  |GpuMemoryBufferImplSurfaceTexture|
+------------------------------------+  +---------------------------------+
|ui::GpuMemoryBufferFactory          |
|OzoneNatiVeBuffer                   |
|          INITALIZES                |
| ui::GLImageOzoneNatiVePixmapDmaBuf |
| OR                                 |
| ui::GLImageOzoneNatiVePixmap       |
+------------------------------------+

BUG=439520
TEST= GpuMemoryBufferTest.* in gl_tests and GpuMemoryBufferImplTest.* in content_unittests.

Review URL: https://codereview.chromium.org/1024113003

Cr-Commit-Position: refs/heads/master@{#322503}

Currently, when the zoom level is set to a particular value, the
viewport's fitting type is unchanged. This can cause the zoom to be
manually set to one value, but be overridden if the viewport changes
size as it will automatically re-apply the zoom for the fitting type.
This CL fixes this by setting the fitting type to none when the zoom is
set manually.

BUG=470427

Review URL: https://codereview.chromium.org/1038573003

Cr-Commit-Position: refs/heads/master@{#322502}

The number of SCTs required for certificates with a validity period of over 14 months (but under 15) was incorrect. Fixed by not counting partial months.

BUG=470169

Review URL: https://codereview.chromium.org/1032093002

Cr-Commit-Position: refs/heads/master@{#322501}

BUG=466848

Review URL: https://codereview.chromium.org/1038143002

Cr-Commit-Position: refs/heads/master@{#322500}

Now that os and cpu_arch are no longer needed or used anywhere.

R=brettw@chromium.org, rockot@chromium.org
BUG=344767
CQ_EXTRA_TRYBOTS=tryserver.chromium.linux:android_chromium_gn_compile_dbg,android_chromium_gn_compile_rel;tryserver.chromium.win:win8_chromium_gn_rel,win8_chromium_gn_dbg;tryserver.chromium.mac:mac_chromium_gn_rel,mac_chromium_gn_dbg

Review URL: https://codereview.chromium.org/1017273002

Cr-Commit-Position: refs/heads/master@{#322499}

R=vmiura@chromium.org
BUG=471049, 467202

Review URL: https://codereview.chromium.org/1033093002

Cr-Commit-Position: refs/heads/master@{#322498}

Increase the maximum allowed remaining lifetime for the
receiver's self-signed cert to 4 days.

This is being increased from 2 days because errors are
being seen in the field because of sender clocks being off.

In the future, we might find a way to warn the user that
their clock might be off, but for now just changing the
tolerance seems the right approach.

BUG=467736

R=mfoltz@chromium.org

Review URL: https://codereview.chromium.org/1011133005

Cr-Commit-Position: refs/heads/master@{#322497}

So these tasks have been queueing with different path names than
those used by the template processor. The latter uses "Unix name"
equivalents of the API features, while the former was not.

For example, "apps/appviewTag" has been caching as such but the
template looks for "apps/appview_tag". It's unclear when this
incongruity appeared but it explains several "my-docs-arent-updating"
bugs that have appeared in the past several weeks.

This changes the refresh servlet to queue API refresh tasks using
API features in unix_name form, resolving the incongruity.

BUG=471019
TBR=kalman@chromium.org

Review URL: https://codereview.chromium.org/1038993002

Cr-Commit-Position: refs/heads/master@{#322496}

When an app window closes, we skip restoring the Chrome main menu if
there are other windows to minimize the number of times the main menu is
changed. This doesn't work if the other windows are on different spaces.
This ensures the Chrome menu is restored if there are no other windows
on the active space.

BUG=467323

Review URL: https://codereview.chromium.org/1024363002

Cr-Commit-Position: refs/heads/master@{#322495}

GLSurface::Destroy() could be called from outside the GLSurface
implementation. Since it is also called in NativeViewGLSurfaceGLX's
destructor, calling Destroy() twice will cause a BadWindow error.

BUG=470004

Review URL: https://codereview.chromium.org/1034173004

Cr-Commit-Position: refs/heads/master@{#322494}

BUG=466091

Review URL: https://codereview.chromium.org/1019543002

Cr-Commit-Position: refs/heads/master@{#322493}

https://chromium.googlesource.com/angle/angle/+log/3cd438d1..1ea584c56

BUG=470695,461274
R=vitalybuka@chromium.org

Review URL: https://codereview.chromium.org/1032323005

Cr-Commit-Position: refs/heads/master@{#322492}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/e0b19d4..fd3a91e

CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
TBR=hcm@google.com

Review URL: https://codereview.chromium.org/1034253002

Cr-Commit-Position: refs/heads/master@{#322491}

Disable expiration date popups, CVC text input field, and storage
checkbox once a permanent error message is shown.

BUG=448572

Review URL: https://codereview.chromium.org/1038043003

Cr-Commit-Position: refs/heads/master@{#322490}

Refactor a little to share some ui->skia converters.

BUG=468509
TEST=set HKLM\Software[\Wow6432Node]\Microsoft\Avalon.Graphics\DISPLAY1\PixelStructure to 1 or 2, zoom in and confirm RGB or BGR fringes respectively.

Review URL: https://codereview.chromium.org/1021643002

Cr-Commit-Position: refs/heads/master@{#322489}

Build VP9 postproc/mfqe for WebRTC
https://codereview.chromium.org/1033873004

R=marpan@google.com, tomfinegan@chromium.org

Review URL: https://codereview.chromium.org/1009963004

Cr-Commit-Position: refs/heads/master@{#322488}

OnPaint() is not part of the recursion, it won't have access to the
necessary pieces to do recursion.

If you need to do painting while inside the recursion, this should
happen in PaintChildren(). So move the contents of OnPaint() to
PaintChildren().

R=estade@chromium.org
BUG=466426

Review URL: https://codereview.chromium.org/1029263011

Cr-Commit-Position: refs/heads/master@{#322487}

This can only fail if a CDM declares support for a container that Chrome does not, which would be a bug in CDM registration.

BUG=374751

Review URL: https://codereview.chromium.org/1035923004

Cr-Commit-Position: refs/heads/master@{#322486}

This only affects PlzNavigate. When a same-site navigation should reuse the current WebUI, in RFHM::GetFrameHostForNavigation the (shadowing) method-local |should_reuse_web_ui_| would be set to true and the speculative WebUI would (correctly) not be created. But later on as the instance-level |should_reuse_web_ui_| defaults to false, RFHM::CommitPending would try to get the |speculative_web_ui_| which would not be set and *could* cause a crash. As of now the crash doesn't happen because CommitPending is not called when the current RFH is kept for the navigation.

BUG=376094

Review URL: https://codereview.chromium.org/1012863004

Cr-Commit-Position: refs/heads/master@{#322485}

This was used as part of an upstream transition and is no longer used.

BUG=428882

Review URL: https://codereview.chromium.org/1026333005

Cr-Commit-Position: refs/heads/master@{#322484}

BUG=chromium:470719
TEST=build successfully
Signed-off-by: Simon Que <sque@chromium.org>

Review URL: https://codereview.chromium.org/1036953002

Cr-Commit-Position: refs/heads/master@{#322483}

BUG=

Review URL: https://codereview.chromium.org/1010803014

Cr-Commit-Position: refs/heads/master@{#322482}

BUG=470660

Review URL: https://codereview.chromium.org/1036853004

Cr-Commit-Position: refs/heads/master@{#322481}

Tweaks to the Data Saver promo currently shown on Chrome
Android to improve the messaging and the call-to-action.
Adds a clickable link to the settings page and makes the
enable button, renamed to "Turn on data saver," directly
enable the proxy.

BUG=455267

Review URL: https://codereview.chromium.org/1011363006

Cr-Commit-Position: refs/heads/master@{#322480}

The return line being outside the synchronized block allowed the cached
state to be reset to null before it was calculated.

BUG=470966

Review URL: https://codereview.chromium.org/1031413005

Cr-Commit-Position: refs/heads/master@{#322479}

$ gn gen out-gn
$ gn check out-gn //components/nacl/renderer/plugin/*

BUG=367595,376000
TEST=see above
R=bradnelson@google.com,brettw@chromium.org
TBR=bradnelson@chromium.org

Review URL: https://codereview.chromium.org/1034193003

Cr-Commit-Position: refs/heads/master@{#322478}

Expose <webview>.contextMenus.onShow
chrome EventTarget like object so one can "addListener" that can
preventDefault() to cancel bringing up <webview> context menu.

BUG=465733
Test=Open a chrome app with <webview>. E.g.
Load a <webview> in a chrome app, e.g.
the browser sample app:
https://github.com/GoogleChrome/chrome-app-samples/tree/master/samples/webview-samples/browser

Right click on the <webview>, context menu is expected to be shown.

Now open the app's inspector: from chrome://inspect, switch to "Apps" then select to inspect "Browser sample".

Register a listener to disable context menu:

document.querySelector('webview').contextMenus.onShow.addListener(function(e) {
  e.preventDefault();
});

Right click on the <webview> again, context menu should not show up.

Review URL: https://codereview.chromium.org/1026383003

Cr-Commit-Position: refs/heads/master@{#322477}

This CL added read_mtu into the signature of
BluetoothAudioSink::Observer::BluetoothAudioSinkDataAvailable and
modified corresponding tests.

BUG=441581
TEST=device_unittests --gtest_filter=*AudioSink*

Review URL: https://codereview.chromium.org/1034893002

Cr-Commit-Position: refs/heads/master@{#322476}

BUG=470231

Review URL: https://codereview.chromium.org/1031413004

Cr-Commit-Position: refs/heads/master@{#322475}

This adds histogram values for extension permission messages that
are missing.

BUG=None
TEST=None

Review URL: https://codereview.chromium.org/1037053003

Cr-Commit-Position: refs/heads/master@{#322474}

The latter is just a backwards-compatibility redirect to the former,
and I'm hoping to get rid of it.

Review URL: https://codereview.chromium.org/1036653004

Cr-Commit-Position: refs/heads/master@{#322473}