GitLab

TBR=achuith@chromium.org
BUG=646965

Review-Url: https://codereview.chromium.org/2341833003
Cr-Commit-Position: refs/heads/master@{#418661}

Component builds are, by definition, not optimized for maximum
performance. So incremental linking - which trades a little bit of
performance for much faster build times - is appropriate in release
component builds, as well as in debug builds.

This change enables incremental linking in release component builds.
This also requires turning off /OPT:ICF and /PROFILE in component
builds because they are incompatible with /incremental, but they are
performance related flags so they are also not important for component
builds.

blink_core.dll cannot be incrementally linked in debug x64 component
builds, but it *can* be incrementally linked in release x64 component
builds (and debug/release x86 component builds), and this change
enables incremental linking of blink_core.dll in all supported
configurations.

This drops the chrome rebuild time after touching
third_party\WebKit\Source\core\html\AutoplayExperimentHelper.cpp from
~260 seconds to 9 seconds. The chrome rebuild time after touching
base\win\registry.cc drops from ~200 seconds to less than 2 seconds,
with 21 of 24 link steps skipped entirely.

This also fixes some /INCREMENTAL /OPT:ICF mismatch warnings.

BUG=608801,621236

Review-Url: https://codereview.chromium.org/2331373006
Cr-Commit-Position: refs/heads/master@{#418660}

There is a crash which happens when AutoplayUmaHelper receives unload
event from window, but m_element is already garbage collected. This CL
fixes this crash by making Element as a Member of AutoplayerUmaHelper.

BUG=646890
R=avayvod@chromium.org,mlamouri@chromium.org

Review-Url: https://codereview.chromium.org/2341783002
Cr-Commit-Position: refs/heads/master@{#418659}

https://chromium.googlesource.com/external/khronosgroup/webgl.git/+log/6a1c97a..582169b

BUG=634525

TEST=bots

CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.android:android_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2335113002
Cr-Commit-Position: refs/heads/master@{#418658}

The change is desirable because:

- Ref-counting doesn't result in memory safety here, because
  ExternalMemoryAllocator wraps a raw pointer to short-lived
  memory.  Ref-counting cannot extend the lifetime of the
  memory this pointer points to.

- No consumers of SkBitmap::Allocator currently use ref-counting.
  If SkBitmap::Allocator was designed today, it probably wouldn't
  use SkRefCnt as a base class.

- Blink-to-Chromium style conversion will rename |ref| to |Ref| in WTF,
  but won't do it outside of Blink/WTF (e.g. in Skia), so RefPtr should
  only be used with Blink/WTF types going forward (i.e. Skia types
  should be wrapped in smart pointers provided by Skia).

More discussion is in an older code review:
https://codereview.chromium.org/1880993003/#msg10

BUG=641014

Review-Url: https://codereview.chromium.org/2326473005
Cr-Commit-Position: refs/heads/master@{#418657}

BUG=513459

Review-Url: https://codereview.chromium.org/2340603002
Cr-Commit-Position: refs/heads/master@{#418656}

BUG=619037

Review-Url: https://codereview.chromium.org/2341773003
Cr-Commit-Position: refs/heads/master@{#418655}

This removes all .gyp files not called "compiled_resources" in the main Chromium repository.
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.win:win10_chromium_x64_rel_ng

Review-Url: https://codereview.chromium.org/2332843004
Cr-Commit-Position: refs/heads/master@{#418654}

New code running in the browser process should always use
ui::XScopedEventSelector instead of XSelectInput.

BUG=634084

Review-Url: https://codereview.chromium.org/2316373003
Cr-Commit-Position: refs/heads/master@{#418653}

https://pdfium.googlesource.com/pdfium.git/+log/1f2489023f2b..41d6bbe3d413

$ git log 1f2489023..41d6bbe3d --date=short --no-merges --format='%ad %ae %s'
2016-09-14 npm Delete unused parameter nWordStyle from GetCharWidth()

TBR=dsinclair@chromium.org

Review-Url: https://codereview.chromium.org/2340103002
Cr-Commit-Position: refs/heads/master@{#418652}

When draining AVDA for destroy or reset, DequeueOutput will discard
any decoded frames from the codec.  Previously, it would still
refuse to check for available output unless a free picture buffer
was available to hold it.

This could prevent a flush for destroy or reset from completing.

In the case of destruction, this would also prevent the codec from
being released.  This affected only VP8 codecs, which require a
drain on some platforms.

BUG=642948

Review-Url: https://codereview.chromium.org/2332253004
Cr-Commit-Position: refs/heads/master@{#418651}

This patch updates PaintArtifactCompositor to create cc's scroll tree
from blink's scroll tree.

PropertyTreeManager::updateScrollOffset is an interesting part of this
patch and is where we resolve the differences between cc's trees and
blink's trees. For example, cc's transform node currently stores
scroll offset in a separate vector, and cc's scroll tree maintains a
separate map of scroll offsets for handling synchronization.

A followup patch will add main thread scrolling reasons. Until then,
a couple simple tests have been added verifying the cc scroll tree
structure. The existing layout tests currently exercise the new scroll
tree code end-to-end.

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
BUG=644514

Review-Url: https://codereview.chromium.org/2334403002
Cr-Commit-Position: refs/heads/master@{#418650}

This is a follow-up patch of https://codereview.chromium.org/2332843002/

Import wpt@de70de04726dedd0981c644669911096854929b6

Using update-w3c-deps in Chromium
8e715596.

Review-Url: https://codereview.chromium.org/2342603002
Cr-Commit-Position: refs/heads/master@{#418649}

https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+log/8c5c37517378..8484270f6d2c

$ git log 8c5c37517..8484270f6 --date=short --no-merges --format='%ad %ae %s'
2016-09-14 benjhayden [value-set-table] Move Show all to the right.
2016-09-14 petrcermak [tracing] Add step-into event to <tr-ui-b-table>
2016-09-14 nednguyen Fix running statistics's mean implementation

TBR=catapult-sheriff@chromium.org

Review-Url: https://codereview.chromium.org/2340733004
Cr-Commit-Position: refs/heads/master@{#418648}

Currently, while the user is hiding the top controls, the compositor expands the
main frame's clipping layer so that we reveal more content. This has to happen
because we don't actually resize the anything in the Layer/Layout trees in
Blink until after the top controls are fully shown or hidden.

When an element is set as the root scroller, we want to provide the same
effect. However, since there's still masking layers above the root scroller, we
can't see the expanded clip.  Instead of expanding all the layers above the
root scroller, this CL removes the masksToBounds bit on all layers above the
root scroller. This means parent boxes with hidden overflow will not clip the
root scroller but this is acceptable as a property of the root scroller.

Also, GraphicsLayer setMasksToBoundaries kept a bit on GraphicsLayer but this was
never read so I removed it.

BUG=505516
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2289833002
Cr-Commit-Position: refs/heads/master@{#418647}

BUG=625270

Review-Url: https://codereview.chromium.org/2278793002
Cr-Commit-Position: refs/heads/master@{#418646}

This CL adds the following metrics:
1. Usage of each pen palette option
2. Usage of each pen palette option on auto-open
3. Length of time spent in each mode
4. count of Mode cancellation

BUG=643798

Review-Url: https://codereview.chromium.org/2308823002
Cr-Commit-Position: refs/heads/master@{#418645}

This regressed in r417444. Previously (https://codereview.chromium.org/2327433002/diff/250001/chrome/browser/certificate_viewer.cc) we would check that the certificate id sent by the renderer was valid (say because renderer died). Races could also happen now, in that the page might navigate right after user presses certificate viewer button and the last committed entry wouldn't have a certificate.

BUG=646882

Review-Url: https://codereview.chromium.org/2337783005
Cr-Commit-Position: refs/heads/master@{#418644}

Makes Instance ownership and lifetime management easier to
follow. Root instances are now explicitly owned by the
ServiceManager and non-root instances are explicitly owned by
their parent Instance.

OnInstanceError always destroys the Instance, either by
removing it from the root instance set or delegating to the parent
Instance. This avoids the need for children to re-enter their
parent on destruction.

Notification logic has been extracted from OnInstanceError and
moved to a new OnInstanceStopped, which is always called by
each Instance's destructor. This allows parents to delete their
own children without calling through to OnInstanceError, while
still allowing the ServiceManager to know about the destruction
and dispatch notifications accordingly.

InstanceErrorType is removed: if an Instance becomes
unreachable it's simply removed from the identity map
via a new OnInstanceUnreachable call.

BUG=640488

Review-Url: https://codereview.chromium.org/2341853002
Cr-Commit-Position: refs/heads/master@{#418643}

BUG=619055
TBR=tedchoc@chromium.org

Review-Url: https://codereview.chromium.org/2337973005
Cr-Commit-Position: refs/heads/master@{#418642}

…and remove the duplicate divider drawing in DownloadShelfView.

BUG=589943

Review-Url: https://codereview.chromium.org/2336463002
Cr-Commit-Position: refs/heads/master@{#418641}

Reason for revert:
The original fix is opt-in, which causes unit test binaries to miss symbols. I will
reimplement an opt-out for binaries that we do not want to whole-archive libs.

Original issue description:
>> All chromecast executable include all libstdc++/libgcc symbols by default
>> due to b/25566835. Added a config for --whole-archive for only ones that
>> depend on shared libs. ~700kb saving per executable.

BUG=internal b/31417380

This reverts commit 92947f23.

Review-Url: https://codereview.chromium.org/2334163004
Cr-Commit-Position: refs/heads/master@{#418640}

This reverts commit a7ce77dd. Monitoring tells
us severely regressed isolate performance.

TBR=jam@chromium.org
BUG=644410

Review-Url: https://codereview.chromium.org/2344563003
Cr-Commit-Position: refs/heads/master@{#418639}

This is to make EnsurePtrConvertibleArgDecl(T, const T) to fail
for https://codereview.chromium.org/2220543002.

This CL also adds unit tests for EnsurePtrConvertibleArgDecl.

BUG=

Review-Url: https://codereview.chromium.org/2220443004
Cr-Commit-Position: refs/heads/master@{#418638}

With this patch, the passthrough command buffer is capable of rendering chrome
and basic WebGL content.

BUG=602688
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2317363005
Cr-Commit-Position: refs/heads/master@{#418637}

BUG= internal b/31488922

Review-Url: https://codereview.chromium.org/2340013002
Cr-Commit-Position: refs/heads/master@{#418636}

Can be enabled in about://flags

BUG=638198

Review-Url: https://codereview.chromium.org/2323563002
Cr-Commit-Position: refs/heads/master@{#418635}

This CL removes all calls to XSelectInput, or XCreateWindow using
CWEventMask, from the browser process that uses the main X
display (ie, the one returned by gfx::GetXDisplay()).  Window event
selection will instead by managed by XWindowEventManager.

BUG=634084

Review-Url: https://codereview.chromium.org/2319933002
Cr-Commit-Position: refs/heads/master@{#418634}

- Add a full integration test that spawns both an engine and client in
  the same process and validates that the rendering pipeline produces
  the correct thing.
- Also add Linux support for BlimpContents and exposed a
  BlimpContentsView, which is required to gain access to the right
  classes to validate the rendering path.

Marking this as NOPRESUBMIT because there are some ScopedAllowIO calls during test setup and client tear down.  Please feel free to revert or ping me on any landing issues.

BUG=644262
NOPRESUBMIT=true

Review-Url: https://codereview.chromium.org/2320923002
Cr-Commit-Position: refs/heads/master@{#418633}

TBR=nednguyen@chromium.org
BUG=646925

Review-Url: https://codereview.chromium.org/2340033002
Cr-Commit-Position: refs/heads/master@{#418632}

This patch enables caching of spellcheck results to improve performance
and address issues mentioned below.

In its current state i.e. without caching spellcheck on Android and Mac
has a number of issues e.g.:
- selection (click) of words or movement of the caret can result
in unnecessary calls to the platform spellchecker.
- typing of non-word characters in the end results in unnecessary
calls to the platform spellchecker.
- modifying text does not always result in invocation of spellchecker,
hence not showing the misspelled words.

BUG=583616, 629609

Review-Url: https://codereview.chromium.org/2317113004
Cr-Commit-Position: refs/heads/master@{#418631}

BUG=

Review-Url: https://codereview.chromium.org/2344583002
Cr-Commit-Position: refs/heads/master@{#418630}

As the mouse cursor moves across OOPIF boundaries, elements in
other processes that have certain mouse event handlers might require
notification in order to fire those handlers. An example is MouseLeave,
which is not currently fired when the MouseMove is directed to a
parent frame.

This CL tracks the last RenderWidgetHostView to receive a mouse
event and sends appropriate events to other renderers when that changes.

BUG=632035

Review-Url: https://codereview.chromium.org/2229463004
Cr-Commit-Position: refs/heads/master@{#418629}

so that the function takes content::WebContents* instead of
content::NavigationHandle*. That allows us to use the bubble
view class outside the throttle.

BUG=b:29079617, b:31406784
TEST=try

Review-Url: https://codereview.chromium.org/2335313003
Cr-Commit-Position: refs/heads/master@{#418628}

BUG=644268
TEST=none

Review-Url: https://codereview.chromium.org/2339813004
Cr-Commit-Position: refs/heads/master@{#418627}

https://pdfium.googlesource.com/pdfium.git/+log/8b1ed1614492..1f2489023f2b

$ git log 8b1ed1614..1f2489023 --date=short --no-merges --format='%ad %ae %s'
2016-09-14 dsinclair Remove FFI_ from CPDFSDK_Environment method names

TBR=dsinclair@chromium.org

Review-Url: https://codereview.chromium.org/2343673002
Cr-Commit-Position: refs/heads/master@{#418626}

Nothing outside of MostVisitedSites needs to use it.

Review-Url: https://codereview.chromium.org/2337893003
Cr-Commit-Position: refs/heads/master@{#418625}

BUG=619055
TBR=tedchoc@chromium.org

Review-Url: https://codereview.chromium.org/2336803005
Cr-Commit-Position: refs/heads/master@{#418624}

Cronet API classes must be in org.chromium.net package and
Cronet implementation classes must not be in org.chromium.net package.

BUG=629299

Review-Url: https://codereview.chromium.org/2320413002
Cr-Commit-Position: refs/heads/master@{#418623}

[Offline Pages] Gates immediate triggering (non scheduled) of background loading by whether low end device.
Makes unit tests conditional on this check - ran unit tests on both svelte and regular devices locally.

BUG=646888

Review-Url: https://codereview.chromium.org/2338313002
Cr-Commit-Position: refs/heads/master@{#418622}