GitLab

We capture sk picture from PictureLayer, which happens before we had
a chance to set the image decode controller during commit. The fix here
is not to use PlaybackToCanvas, but rather use other functions to
accomplish the same goal. PlaybackToCanvas should only be called after
commit.

R=enne
BUG=581107
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1633203003

Cr-Commit-Position: refs/heads/master@{#371655}

Calling Layer::set_property_tree_sequence_number() does not set the
flag that a commit is needed, which means it can quickly get out of
date for layers that are not updated for other reasons when the Layer
is (de)serialized.

This CL makes the LayerTreeHost forcefully set the sequence number
to the same value as the current PropertyTrees.sequence_number, since
all valid in the hierarchy is valid at that point.

BUG=577352, 538710
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1584653004

Cr-Commit-Position: refs/heads/master@{#371654}

BUG=536669

Review URL: https://codereview.chromium.org/1632333002

Cr-Commit-Position: refs/heads/master@{#371653}

BUG=574938
TEST=PaintLayerPainterTest.PaintPhaseOutline
TEST=fast/multicol/span/outline.html

Review URL: https://codereview.chromium.org/1584493002

Cr-Commit-Position: refs/heads/master@{#371652}

BUG=523278

Review URL: https://codereview.chromium.org/1609133002

Cr-Commit-Position: refs/heads/master@{#371651}

The tool is run in parallel many times, so they may clobber each
other writing to the rewrite-sym.txt file. So add a rewrite-sym.lock
file that each instance grabs when writing.

Also attempts a windows implementation which is untested.

R=dcheng
BUG=578344

Review URL: https://codereview.chromium.org/1616223002

Cr-Commit-Position: refs/heads/master@{#371650}

original cl:
https://codereview.chromium.org/1633713002/
revert:
https://codereview.chromium.org/1639603002/
reason:
broken paths

BUG=514857

Review URL: https://codereview.chromium.org/1632973002

Cr-Commit-Position: refs/heads/master@{#371649}

The CVDisplayLink is tied to a ScopedTypeRef data member of DisplayLinkMac.
However, there may be other references. Since there's no guarantee that
~DisplayLinkMac() will invoke the last CVDisplayLinkRelease(..) and clear
the callback automatically, clear it explicitly in ~DisplayLinkMac.

BUG=564780

Review URL: https://codereview.chromium.org/1626163002

Cr-Commit-Position: refs/heads/master@{#371648}

A recent commit broke the ability for the Google Cast and Media Router
extensions to start tab capture for mirroring.  The root cause was due
to the new behavior where the presence of deprecated constraints now
causes the getUserMedia() API to error-out (instead of silently ignoring
the constraints).

The short-term fix here is to simply filter out all goog* constraints
since they have no effect on tab capture.

BUG=579729
TBR=hta@chromium.org

Review URL: https://codereview.chromium.org/1631243003

Cr-Commit-Position: refs/heads/master@{#371647}

BUG=581519
TBR=groby@chromium.org

Review URL: https://codereview.chromium.org/1634973003

Cr-Commit-Position: refs/heads/master@{#371646}

Cross-process frames are possible in OOPIF-based webview.

BUG=578884

Review URL: https://codereview.chromium.org/1639013002

Cr-Commit-Position: refs/heads/master@{#371645}

Pass to the renderer a list of texture targets, one per GpuMemoryBuffer
format, via a command line argument.
In this way GMBVFP can pick whatever GMB format it needs to use and bind
to the correct texture target.
This is similar to what we do for content (kContentImageTextureTarget).

kContentImageTextureTarget can't be used since the buffer usage could
differ from the buffer usage in GMBVFP.

Once all GMB formats supports persistent usage and we can consolidate
GPU_READ_CPU_READ_WRITE_PERSISTENT and GPU_READ_CPU_READ_WRITE this
code can go away and we can use just one command line switch.

BUG=524582

Review URL: https://codereview.chromium.org/1605423002

Cr-Commit-Position: refs/heads/master@{#371644}

BUG=525873

Review URL: https://codereview.chromium.org/1638583005

Cr-Commit-Position: refs/heads/master@{#371643}

BUG=525873

Review URL: https://codereview.chromium.org/1638523004

Cr-Commit-Position: refs/heads/master@{#371642}

This still does not include cronet or webview.

BUG=580129

Review URL: https://codereview.chromium.org/1636003003

Cr-Commit-Position: refs/heads/master@{#371641}

*PERF SHERRIF: this CL intentionally breaks dummy_benchmark.noisy_benchmark_1
for intergration testing purpose.
Please coordinate with prasadv@ before your revert this.

BUG=537676, 581465

Review URL: https://codereview.chromium.org/1439453002

Cr-Commit-Position: refs/heads/master@{#371640}

The objects keeping their own WebURLLoaders (and thus being
WebURLLoaderClients) must promptly let go of their ownership upon
becoming garbage. Not doing so risks the embedder calling the client
while it is in a sweepable state.

R=
BUG=568173

Review URL: https://codereview.chromium.org/1635113002

Cr-Commit-Position: refs/heads/master@{#371639}

BUG=525873

Review URL: https://codereview.chromium.org/1639763002

Cr-Commit-Position: refs/heads/master@{#371638}

This allows us to use tracing agents that require more time to start.
The first use case of this will be the BattOr tracing agent, which
requires several round trips over USB before it can know whether
StartTracing was successful.

Review URL: https://codereview.chromium.org/1614063005

Cr-Commit-Position: refs/heads/master@{#371637}

BUG=525873

Review URL: https://codereview.chromium.org/1639753002

Cr-Commit-Position: refs/heads/master@{#371636}

Previously, when navigating from one page with 3+ infobars to another
page with infobars, the animation would show only one or two infobars
hiding then one or two infobars appearing, even though all the existing
infobars should disappear, then all the new infobars should appear. To
the user, it appeared as if some of the new infobars had been there all
along (peeking up in the back), which is confusing and wrong.

This strange behavior was caused by the logic for deciding when to show
a hiding or appearing animation. This logic didn't associate a
particular infobar with a particular view; it just ensured that the
number of views matched the number of infobars. So, if infobars were
removed and then new infobars were quickly added, there would be fewer
animations than expected.

The new code explicitly associates an infobar with each view, so if all
the infobars are removed and then several new ones added, the animations
will reflect that correctly.

BUG=396223

Review URL: https://codereview.chromium.org/1634083002

Cr-Commit-Position: refs/heads/master@{#371635}

When a user switches tabs while the infobar container is showing,
its state is not saved and it is possible for events in the currently
active tab to reset reader mode's knowledge of the infobars. This
change introduces an event "onInfobarContainerAttachedToWindow" that
notifies observers when the infobar container is attached and if
it needs to show any infobars.

BUG=579036

Review URL: https://codereview.chromium.org/1639673002

Cr-Commit-Position: refs/heads/master@{#371634}

This will allow removing simple_service from the NaCl side.

BUG=302078
TEST=build

Review URL: https://codereview.chromium.org/1631393002

Cr-Commit-Position: refs/heads/master@{#371633}

Looks like the target_rect param to BindFramebufferToTexture is now
unused. And it looks like UseScopedTexture is entirely unused.
Remove them both.

BUG=
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Review URL: https://codereview.chromium.org/1633283002

Cr-Commit-Position: refs/heads/master@{#371632}

MultiWindowUtils#isLecacyMultiWindow() will check for old multi-window
mode (e.g. Samsung's multi-window mode), and #isMultiWindow() will check
for Android N multi-window.

BUG=570741

Review URL: https://codereview.chromium.org/1633763003

Cr-Commit-Position: refs/heads/master@{#371631}

Also:

 - Move RecordTimeSinceLastStartup() and RecordStartupCount() as implementation
   details of RecordBrowserMainMessageLoopStart().

 - Deprecate "Startup.BrowserMessageLoopStartHardFaultCount.FirstRun" histogram
   (obsolete and also replaced by the more generic
    "Startup.BrowserMessageLoopStartHardFaultCount.1")

 - Fix "Startup.BrowserMessageLoopStartHardFaultCount" to use 1 instead of 0 as
   minimum since 0 is implicit in UMA.

 - Fix "Startup.Temperature" to use STARTUP_TEMPERATURE_MAX instead of
   STARTUP_TEMPERATURE_COUNT as the |max_boundary| as this is what the
   UMA_HISTOGRAM_ENUMERATION macro expects FWICT.

BUG=580211

Review URL: https://codereview.chromium.org/1637493002

Cr-Commit-Position: refs/heads/master@{#371630}

This adds a menu item to remove the MR component action from the toolbar.  It
sets a pref so this preference is respected across sessions.

The user can restore the action by reinstalling the Cast extension from the Web
Store.  (They may have to reload the Web Store page immediately after
installation to do so.)

BUG=576362
TESTED=Unit test, manually tested

Review URL: https://codereview.chromium.org/1612203002

Cr-Commit-Position: refs/heads/master@{#371629}

BUG=None
TEST=None

Review URL: https://codereview.chromium.org/1639463002

Cr-Commit-Position: refs/heads/master@{#371628}

BUG=295792
TEST=webgl2_conformance
R=bajones@chromium.org

Review URL: https://codereview.chromium.org/1625643003

Cr-Commit-Position: refs/heads/master@{#371627}

TBR=dgozman
BUG=432444

Review URL: https://codereview.chromium.org/1636113003

Cr-Commit-Position: refs/heads/master@{#371626}

Update the code handling switching out the various host-list-related
views to handle the three views (loading, no hosts, and host chooser)
directly instead of setting the no-hosts view as the chooser's empty
view. The two-state empty view functionality doesn't appear to be
designed to be used along side a third modal loading state.

This fixes an issue where both the loading spinner and the "There's
nothing to connect to" message could appear simultaneously under certain
conditions.

BUG=566177

Review URL: https://codereview.chromium.org/1608093002

Cr-Commit-Position: refs/heads/master@{#371625}

BUG=NONE

Review URL: https://codereview.chromium.org/1560503002

Cr-Commit-Position: refs/heads/master@{#371624}

BUG=580262,562189

Committed: https://crrev.com/5db5d9e5815c64ba4bc93fda27115c542df35316
Cr-Commit-Position: refs/heads/master@{#371573}

Review URL: https://codereview.chromium.org/1618973002

Cr-Commit-Position: refs/heads/master@{#371623}

In SQLite 3.10.2, the Unix VFS no longer defines unixCurrentTime(),
instead relying on unixCurrentTimeInt64().  The websql vfs wrapper
wraps at the v1 level, so expected to see xCurrentTime().

Backports http://www.sqlite.org/src/info/c11c85fdb6514cae

"Reenable the xCurrentTime and xGetLastError methods on the
unix VFSes as long as SQLITE_OMIT_DEPRECATED is not defined."

BUG=580948
TBR=michaeln@chromium.org

Review URL: https://codereview.chromium.org/1638943002

Cr-Commit-Position: refs/heads/master@{#371622}

NOTRY=true

Review URL: https://codereview.chromium.org/1635173002

Cr-Commit-Position: refs/heads/master@{#371621}

1. AutomationUtil.findNodeUntil isn't entirely clear due to the use of a callback that wants to keep track of the previous node being visited. This has since become wrong because of the changes made to AutomationTreeWalker. Instead, let's just call AutomationUtil.findNextNode which should gibve us the proper linearization. Take pairs in order and apply the predicate.

2. wrapping by line didn't quite work as expected because of the asymmetric nature of the linearization. We need to get to the end when wrapping from the beginning to the end and start applying our predicate. Note that this is subject to some weird corner cases depending on how we write the predicates (e.g. ones that check for children).

3. Ensure we consider invisible nodes to be leaves. They should also be ignored when encountered (this is already done).

TEST=navigate around using ChromeVox Next on the shelf. Observe desirable behaviors.

Review URL: https://codereview.chromium.org/1622803002

Cr-Commit-Position: refs/heads/master@{#371620}

https://chromium.googlesource.com/skia.git/+log/ed0bf1b43e19..727b7d27afd0

$ git log ed0bf1b43..727b7d27a --date=short --no-merges --format='%ad %ae %s'
2016-01-26 robertphillips Add gpu backend for SkPixelXorXfermode GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1631873002

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

Review URL: https://codereview.chromium.org/1631403002

Cr-Commit-Position: refs/heads/master@{#371619}

ScopedConstants.TEN_TOO has been removed from mojom.

BUG=None

Review URL: https://codereview.chromium.org/1635783003

Cr-Commit-Position: refs/heads/master@{#371618}

SetLocaleString took a string by value and used std::move on it.

After a discussion on chromium-dev [1], pasing a const ref was recognised as
the preferred style in Chromium, so this CL changes it.

[1]
https://groups.google.com/a/chromium.org/forum/#!searchin/chromium-dev/string$20const$20ref/chromium-dev/ZoAfW7GXUVI/fCYjkg2JBAAJ

BUG=571610
R=isherman@chromium.org

Review URL: https://codereview.chromium.org/1635813003

Cr-Commit-Position: refs/heads/master@{#371617}

Make changes necessary to handle the screen size change when a tab enters
and exits fullscreen, and enable Contextual Search for non-video fullscreen.

BUG=515151

Review URL: https://codereview.chromium.org/1589023003

Cr-Commit-Position: refs/heads/master@{#371616}