GitLab

AndroidVideoDecodeAccelerator, which runs in the GPU process, needs a
way to get a java Surface from the browser process. GpuSurfaceLookup and
SurfaceTextureManagerImpl already provide a way to get an ANativeWindow
but not the corresponding java Surface. This new method returns the java
surface.

BUG=533630

Review URL: https://codereview.chromium.org/1567193002

Cr-Commit-Position: refs/heads/master@{#368478}

Note that this CL didn't do the upgrade for unpack, although the
foundation is there.

BUG=563714
TEST=conformance2/reading/read-pixels-pack-parameters.html
R=kbr@chromium.org,bajones@chromium.org

Review URL: https://codereview.chromium.org/1566283003

Cr-Commit-Position: refs/heads/master@{#368477}

Chrome used to have EAC3/DD+ support (crbug.com/215773),
but it was removed last year (crbug.com/334126). This CL
partially restores the old code and adds AC3 support.
This will allow AC3/EAC3 audio demuxing and can be used
for AC3/EAC3 pass-through support.

BUG=448878

Review URL: https://codereview.chromium.org/812643005

Cr-Commit-Position: refs/heads/master@{#368476}

With this change, RenderFrameImpl::createMediaPlayer() will select between WMPI and WMPA on Android. Specifically, WMPI will be used unless:
  - --disable-accelerated-video-decode is set.
  - MediaCodec is not supported (because the API level is too low or the device is blacklisted).
  - The URL indicates that the video is HLS.

Additionally, --disable-accelerated-video-decode is now automatically set whenever any of these flags are set:
  --disable-gpu
  --single-process
  --in-process-gpu

Since in these modes VDAs do not currently work (and these flags are not passed to the renderer).

BUG=516765

Review URL: https://codereview.chromium.org/1560953002

Cr-Commit-Position: refs/heads/master@{#368475}

BUG=chromium:411923
TEST=validate_format, pretty_print

Review URL: https://codereview.chromium.org/1576543002

Cr-Commit-Position: refs/heads/master@{#368474}

Removing following three indirect methods which can be directly
implemented within blink platform:
- readFromFile
- enterRunLoop
- exitRunLoop

BUG=571332
TBR=danakj

Review URL: https://codereview.chromium.org/1568433005

Cr-Commit-Position: refs/heads/master@{#368473}

WebMediaPlayerMS owns WebMediaPlayerMSCompositor, but they live on two
different threads, and then there is a chance that when WebMediaPlayerMS
is destructed on the main thread, there are still tasks regarding
WebMediaPlayerMSCompositor on the compositor thread, and finally crash
will be observed.

Pass weak pointer to the PostTask call, then when the object is
destroyed, the task will be voided.

BUG=572398

Review URL: https://codereview.chromium.org/1565893002

Cr-Commit-Position: refs/heads/master@{#368472}

The original motivation for this CL was to add an overload
to the isApiEnabled() method, without an error message.
Like ExecutionContext.isSecureContext(), this allows
callers to check for enabled experiments, but not pass
an error message parameter, which would be ignored.

In adding the overload, it was determined that all callers should
be using the ExperimentalFeatures generated class
(with the exception of unit tests). As a result, the
Experiments class now just provides a private
implementation of the enabled check. Callers use the
ExperimentalFeatures class for the integrated check of
runtime-enabled + appropriate API key.

The motivation for this change was this CL to add the
[ExperimentEnabled] IDL attribute:
https://codereview.chromium.org/1531443003

BUG=551609

Review URL: https://codereview.chromium.org/1541983003

Cr-Commit-Position: refs/heads/master@{#368471}

Handling themes (e.g. dark custom themes) is a TODO, but the asset icons don't handle them either so this is not a regression.

BUG=520623, 575031
TBR=oshima@chromium.org

Review URL: https://codereview.chromium.org/1561923005

Cr-Commit-Position: refs/heads/master@{#368470}

Also: unchecked errors are now also reported to the extension's console.
Unchecked errors were silenced before because the internal disconnect
handler accessed chrome.runtime.lastError.

BUG=439780
TEST=browser_tests --gtest_filter=ExtensionApiTest.MessagingCrash

Review URL: https://codereview.chromium.org/1567423002

Cr-Commit-Position: refs/heads/master@{#368469}

Add an OutputSurface::ApplyExternalStencil method.
Android WebView implements this method on the render thread
and applies the stencil parameters read in functor, through
the command buffer client.

BUG=574570
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1564703005

Cr-Commit-Position: refs/heads/master@{#368468}

read_wrapper() and write_wrapper() are no longer used now that the
PNaCl code doesn't use SRPC.

Open() had the side effect of checking the validity of the
handle/descriptor before, so preserve that, but rename it to
CheckValidity() to clarify what it does.

BUG=302078
TEST=e.g. NaClBrowserTestPnacl.PPAPICore (tests PNaCl translation)

Review URL: https://codereview.chromium.org/1575573002

Cr-Commit-Position: refs/heads/master@{#368467}

There are many crashes in device_forwarder on bots, e.g.
https://build.chromium.org/p/tryserver.chromium.android/builders/linux_android_rel_ng/builds/5713/steps/stack_tool_for_tombstones/logs/stdio/text

The original method was incorrect because
ServiceDelegate::DeleteControllerOnInternalThread()
may execute after ServiceDelegate::~ServiceDelegate().

Review URL: https://codereview.chromium.org/1571643003

Cr-Commit-Position: refs/heads/master@{#368466}

Review URL: https://codereview.chromium.org/1514633006

Cr-Commit-Position: refs/heads/master@{#368465}

Without this the client and server may be out of sync.

BUG=none
TEST=covered by tests
R=ben@chromium.org

Review URL: https://codereview.chromium.org/1568263003

Cr-Commit-Position: refs/heads/master@{#368464}

Mainly by getting rid of excessive painting.

Changes:

1. We don't need to repaint progress every 30ms in the common case (known
file size) because we will only paint something different if the percent
downloaded has actually changed. When that changes, the item receives
OnDownloadUpdated(), so there's no purpose of a repeating timer on top of
that.

2. We don't need to repaint the entire shelf every time one item on it
updates. I don't know why that was ever necessary, but I can no longer see
a reason for it.

Effects:

1. On a release build on my local machine, the GPU process drops from ~10%
per active download to ~1% per active download (almost indistinguishable from
steady state given my unscientific testing methodology).

2. The browser process still uses a high amount of CPU, but most of that is
in handling the actual download, not displaying the UI. Closing the shelf has
nearly no effect on browser process and gpu process utilization.

3. Indeterminate progress DLs still use a lot of CPU. TODO(estade): fix this.

Side notes:

1. Even without this change, the material shelf is a bit faster than the
non-material shelf, but not by an order of magnitude.

BUG=569354

Review URL: https://codereview.chromium.org/1538773002

Cr-Commit-Position: refs/heads/master@{#368463}

    effective in chrome.tabs.query.

BUG=572545
Test=Manually using ./preview.py

Review URL: https://codereview.chromium.org/1576433002

Cr-Commit-Position: refs/heads/master@{#368462}

This allows for iOS to only ship one icuctl.dat file for the main bundle
and for the today extension.

BUG=

Review URL: https://codereview.chromium.org/1568363002

Cr-Commit-Position: refs/heads/master@{#368461}

When moving activation (e.g. either because of window-switch accelerator,
or when the active window is destroyed/hidden), make sure the previous
active window is activated first, instead of the oldest active window.
However, when explicitly cycling through the active windows (e.g. by using
the window-cycle accelerator), make sure it cycles through all activatable
windows, instead of just the two topmost windows.

BUG=548422

Review URL: https://codereview.chromium.org/1560063003

Cr-Commit-Position: refs/heads/master@{#368460}

SPDY/2 code path has been unused for a long time, this CL actually removes it.

Note that there have been two comparisons between an enum SpdyMajorVersion value
and an interger literal: one in SpdyFramer::WriteHeaderBlockToZ(), and one in
spdy_http_utils.cc::SpdyHeadersToHttpResponse().  Changing enum SpdyMajorVersion
member values lead to errors.  However, removing SPDY/2 made these comparisons
mute (they were both |version < SPDY3|), so they are removed instead of fixed.

This CL lands server change 111503966 by alyssar.

BUG=431299, 488484

Review URL: https://codereview.chromium.org/1561203003

Cr-Commit-Position: refs/heads/master@{#368459}

All major features are implemented, a few TODO's, and a few tests.

Design doc:

https://docs.google.com/document/d/1hLK0eyT5_BzyNS4OkjsnoqqFQDYCbKfyBinj94OnLiQ/edit

BUG=540528

Review URL: https://codereview.chromium.org/1449623002

Cr-Commit-Position: refs/heads/master@{#368458}

The offline pages team would like to add unit tests, but nobody on our team
is in the OWNERS heirarchy for these files.  I think it is reasonable for
us to add an owners file with our team.

BUG=

Review URL: https://codereview.chromium.org/1573553003

Cr-Commit-Position: refs/heads/master@{#368457}

* Reorder to group related items together.  Add group comments.
* Rename {NUM,CAPS,SCROLL}_LOCK_DOWN to ..._ON, in hopes of avoiding confusion
  as to whether the event is about one of these keys actually being pressed down
  versus just conveying the current state of the various locks.
* Rename EF_EXTENDED to EF_IS_EXTENDED_KEY in hopes of greater clarity.  Move it
  to the KeyEvent-specific enum since it's non-sensical for non-KeyEvents.
* Expose MouseEvent::button_flags() since a couple callers outside the class
  wanted it.
* Move Event::IsRepeat() to KeyEvent (to match the enum placement) and rename it
  unix_hacker()-style.  (There are a ton of other functions in event.h that
  should be named that way, but there's only so much drudgery I can stand.)
* Add some missing bits: missing static_asserts in the Mojo code that didn't
  check all enum values were equal, a missing SCROLL_LOCK_ON check in the exo
  code that would have no functional effect but looked unduly suspicious.
* Attempt to reorder code to either check/use these enum values in the same
  order they're declared, or, in cases like code mapping GTK enum values to
  these values, in the order of the GTK enum values (and similar).

BUG=none
TEST=none

Review URL: https://codereview.chromium.org/1559163002

Cr-Commit-Position: refs/heads/master@{#368456}

Removed unnecessary permission checks that were requiring a permission
even on M where they are not needed

Updated the WebViewShell manifest to properly support negotiate auth

BUG=533513

Review URL: https://codereview.chromium.org/1508833003

Cr-Commit-Position: refs/heads/master@{#368455}

BUG=571500

Review URL: https://codereview.chromium.org/1573573003

Cr-Commit-Position: refs/heads/master@{#368454}

https://chromium.googlesource.com/skia.git/+log/0e101667d604..d9c0037211e1

$ git log 0e101667d..d9c003721 --date=short --no-merges --format='%ad %ae %s'
2016-01-08 mtklein Clean up order of arguments to d,s[,aa].
2016-01-08 mtklein Revert of Add a class representing texture swizzle. (patchset #6 id:100001 of https://codereview.chromium.org/1567733005/ )
2016-01-08 mtklein Clean up SkXfermode_opts.h
2016-01-08 bsalomon Add a class representing texture swizzle.
2016-01-08 lsalzman fix opacity check in SkBitmapDevice::Create

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

Review URL: https://codereview.chromium.org/1576523002

Cr-Commit-Position: refs/heads/master@{#368453}

Summary of changes available at:
https://chromium.googlesource.com/v8/v8/+log/925b381b..91b2915b

Please follow these instructions for assigning/CC'ing issues:
https://github.com/v8/v8/wiki/Triaging%20issues

Please close rolling in case of a roll revert:
https://v8-roll.appspot.com/
This only works with a Google account.

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

Review URL: https://codereview.chromium.org/1571903003

Cr-Commit-Position: refs/heads/master@{#368452}

When a prerender completes in the background, the tab title is not
correctly set when the page is shown. This was actually never done right
in the past and was only not visible because there were an overflow of
signals about pageLoadStart. After
https://codereview.chromium.org/1381003004, the observer behavior was
fied, which surfaced the prerender issue. This is not visible in omnibox
prerendering since the suggestion selection mechanism already updates
the url.

BUG=554236
TEST=CustomActivityTest#testPrerenderingCorrectUrl

Review URL: https://codereview.chromium.org/1563423002

Cr-Commit-Position: refs/heads/master@{#368451}

When the new --use-direct-composition command-line flag is specified the GPU process will create a new child window to present into. This will be necessary when using DirectComposition because DirectComposition can only render into windows created by the same process.

The sandbox prevents the GPU process from directly doing SetParent of its child window to the browser window, so a new IPC is added to allow that.

BUG=524838,545203

Review URL: https://codereview.chromium.org/1538803004

Cr-Commit-Position: refs/heads/master@{#368450}

Simple Cache previously recorded regular reads and writes in the NetLog
but not sparse reads and writes. This CL fixes that.

BUG=

Review URL: https://codereview.chromium.org/1423063007

Cr-Commit-Position: refs/heads/master@{#368449}

The selection handle positioning logic wasn't written with vertical
writing modes in mind and they end up attached to odd edges of
the selection rect (see screenshots on bug).  Position and
orient them more appropriately.

BUG=557529

Review URL: https://codereview.chromium.org/1531473004

Cr-Commit-Position: refs/heads/master@{#368448}

https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+log/fd0080e3837f..57af89e1614f

$ git log fd0080e38..57af89e16 --date=short --no-merges --format='%ad %ae %s'

CQ_INCLUDE_TRYBOTS=
TBR=catapult-sheriff@chromium.org

Review URL: https://codereview.chromium.org/1571903002

Cr-Commit-Position: refs/heads/master@{#368447}

BUG=571709

Review URL: https://codereview.chromium.org/1570663003

Cr-Commit-Position: refs/heads/master@{#368446}

This patch also fixes a bug when the tool renamed expressions involving
nested name specifiers: given A::B::C, it incorrectly replaced `A` when
it should have replaced `C`.

BUG=none

Review URL: https://codereview.chromium.org/1557243002

Cr-Commit-Position: refs/heads/master@{#368445}

This lets us remove const_cast<SkPicture*> from callers (i.e. Blink) which
hold SkPicture through const pointers.

CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1564233005

Cr-Commit-Position: refs/heads/master@{#368444}

BUG=574951
TEST=On OOBE or login screen, turn on screen magnifier, then click on
blank parts of page.  Magnification shouldn't move to upper left of
screen.

Review URL: https://codereview.chromium.org/1564973002

Cr-Commit-Position: refs/heads/master@{#368443}

Allow response headers up to 256 kB to be processed when speaking HTTP/2
(increased from earlier 32 kB limit).  Instead of allocating a 256 kB buffer at
once, start with a 8 kB buffer and grow exponentially to limit memory usage.

Also increase string literal (header key and value) limit to 256 kB (from
earlier 16 kB limit).

BUG=535629

Review URL: https://codereview.chromium.org/1566703002

Cr-Commit-Position: refs/heads/master@{#368442}

This patch fixes a subtle bug that could occur if a sync point client
is waiting without a corresponding order number. This can occur for any
wait that is not done within the usual command buffer processing loop.
An example is the command buffer stub OnSignalSyncToken() function
which issues a wait directly on the IO thread.

Previously, a similar concept was introduced in SyncPointClientWaiter
which was used to wait out of order when a Gpu Memory Buffer was
destroyed. Because of the need for this functionality in the regular
SyncPointClient, I have folded the SyncPointClientWaiter functions into
SyncPointClient under WaitOutOfOrder() & WaitOutOfOrderNonThreadSafe().

Because of how subtle this bug is, further state tracking is now done
so we can test whether or not an order number is processing or not.
Using that test, DCHECKs have been added so Wait()/WaitNonThreadSafe
can only be called when processing an order number, and the the out of
order variants can only be called when not processing an order number.

BUG=514815
TEST=Added unit tests to test SignalSyncToken().

Review URL: https://codereview.chromium.org/1568563002

Cr-Commit-Position: refs/heads/master@{#368441}

 * Remove the NaClSrpcModuleInit() call.  Since we no longer use SRPC,
   we don't need to initialise it.

 * NaClSubprocess: Remove SrpcClient field and constructor argument,
   since the SrpcClient object is no longer used.

BUG=302078
TEST=e.g. NaClBrowserTestPnacl.PPAPICore (tests PNaCl translation)

Review URL: https://codereview.chromium.org/1569283002

Cr-Commit-Position: refs/heads/master@{#368440}

These targets have headers included in //content/renderer, but which fail
`gn check` because //content/renderer doesn't have a direct dependency on
them.

This resolves 16 errors in `gn check out/Default //content/renderer/*`.

Review URL: https://codereview.chromium.org/1568613006

Cr-Commit-Position: refs/heads/master@{#368439}