GitLab

This removes the acceleratedCompositingForFixedRootBackground setting
and makes the WebSetting setter a no-op. Any code that checked for
this setting is replaced with a check for
preferCompositingToLCDTextEnabled instead.

R=abarth, vollick@chromium.org
BUG=365851

Review URL: https://codereview.chromium.org/479893002

git-svn-id: svn://svn.chromium.org/blink/trunk@180431 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=402387

Review URL: https://codereview.chromium.org/483603003

git-svn-id: svn://svn.chromium.org/blink/trunk@180430 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=402536

Review URL: https://codereview.chromium.org/486563002

git-svn-id: svn://svn.chromium.org/blink/trunk@180429 bbb929c8-8fbe-4397-9dbb-9b2b20218538

1. This stack copying limit is a performance sanity check;
2. ASan/LSan use more space on the stack, and performance is not an issue for LSan/ASan. So, we allow more stack copying to avoid test case failed on these builds.

R=ager@chromium.org, oilpan-reviews@chromium.org
BUG=403273

Review URL: https://codereview.chromium.org/474203002

git-svn-id: svn://svn.chromium.org/blink/trunk@180428 bbb929c8-8fbe-4397-9dbb-9b2b20218538

- Merge multiple |namespace blink| blocks
- Remove unnecessary blink:: prefix
- Other minor style fixes.

BUG=

Review URL: https://codereview.chromium.org/474183002

git-svn-id: svn://svn.chromium.org/blink/trunk@180427 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Another step toward removing the deprecated ServiceWorkerContainer.unregister.
service_worker_unregister_and_register is the biggest consumer of it in our
tests.

Now to unregister, the function first registers at that scope to get a
registration, then calls unregister on it. This only works with registrations
that cleanly install, which is the most common use case.

BUG=404064

Review URL: https://codereview.chromium.org/472293002

git-svn-id: svn://svn.chromium.org/blink/trunk@180426 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL removes |DocumentParser::startParsing()|.
This was only used from XMLDocumentParser, but it was a no-op.
The method initializes |m_state| to |ParsingState|, however the state should
already be in |ParsingState| from c-tor initializer.

BUG=404107

Review URL: https://codereview.chromium.org/471423002

git-svn-id: svn://svn.chromium.org/blink/trunk@180425 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Oilpan: Move more code of RenderObject destructors to destroy(). (patchset #1 of https://codereview.chromium.org/459353002/)

Reason for revert:
Caused crashes in ImageQualityController::highQualityRepaintTimerFired().


Original issue's description:
> Oilpan: Move more code of RenderObject destructors to destroy().
> 
> Like [1] and [2]. We should do it in general.
> 
> [1] http://src.chromium.org/viewvc/blink?view=revision&revision=180023
> [2] http://src.chromium.org/viewvc/blink?view=revision&revision=180029
> 
> BUG=398342
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180046

TBR=oilpan-reviews@chromium.org,haraken@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=398342

Review URL: https://codereview.chromium.org/478923005

git-svn-id: svn://svn.chromium.org/blink/trunk@180424 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Given that we clear the m_referencedNodes in AudioContext::uninitialize(),
the m_referencedNodes should be empty in both oilpan builds and non-oilpan builds.

BUG=340522

Review URL: https://codereview.chromium.org/475373002

git-svn-id: svn://svn.chromium.org/blink/trunk@180423 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The methods multiply a scale transformation on the current matrix.

The specification:
  http://dev.w3.org/fxtf/geometry/

Intent to Implement:
  https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/V_bJNtOg0oM

BUG=388780

Review URL: https://codereview.chromium.org/450533006

git-svn-id: svn://svn.chromium.org/blink/trunk@180421 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=yurys

Review URL: https://codereview.chromium.org/470543002

git-svn-id: svn://svn.chromium.org/blink/trunk@180420 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This rolls codemirror to 2b30a049fd, which is slightly
later than 4.4.1.

Layout test change: In codemirror version 4.4.1 the viewport update logic has changed and it no longer works if text editor offsetHeight is 0px. So creation of editor instance
for tests now setups default editor height of 100px.

R=apavlov
BUG=404120

Review URL: https://codereview.chromium.org/475393002

git-svn-id: svn://svn.chromium.org/blink/trunk@180419 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The methods multiply a translation transformation on the current matrix.

The specification:
  http://dev.w3.org/fxtf/geometry/

Intent to Implement:
  https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/V_bJNtOg0oM

BUG=388780

Review URL: https://codereview.chromium.org/446803002

git-svn-id: svn://svn.chromium.org/blink/trunk@180418 bbb929c8-8fbe-4397-9dbb-9b2b20218538

ASSERTION FAILED: !extraRowSpanningHeight in WebCore::RenderTableSection::distributeRowSpanHeightToRows

Spanning rows have only percent rows and total percent is less than 100 So only
percent portion of extraRowSpnnaingHeight is consumed in percent spanning rows and
remaining is left to distribute in non-spanning rows. But non-percent rows height
is zero So assertion is failing.

Now, whole extraRowSpanningHeight is distributed to percent spanning rows if non-percent
spanning rows height is zero.

R=jchaffraix@chromium.org
BUG=366579

Review URL: https://codereview.chromium.org/398013005

git-svn-id: svn://svn.chromium.org/blink/trunk@180417 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is patch 1 of 4.

TBR=phoglund@chromium.org, jsbell@chromium.org, cmumford@chromium.org
BUG=404303, 404305

Review URL: https://codereview.chromium.org/478863002

git-svn-id: svn://svn.chromium.org/blink/trunk@180416 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Allow Elements who are not styled but have class set to be retrieved using getElementsByClassName.

Matches FF behavior.

https://bugs.webkit.org/show_bug.cgi?id=94718

Review URL: https://codereview.chromium.org/458663003

git-svn-id: svn://svn.chromium.org/blink/trunk@180414 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Add support to new-line character handling in error and warning console
message texts which are shown in a popover in editor.

BUG=404380
R=vsevik

Review URL: https://codereview.chromium.org/478963002

git-svn-id: svn://svn.chromium.org/blink/trunk@180413 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=none

Review URL: https://codereview.chromium.org/476083002

git-svn-id: svn://svn.chromium.org/blink/trunk@180412 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The patch adds tooltips with actual pixel values of serifs, representing
CSS breakpoints on CSS ruler.

BUG=391240
R=pfeldman, dgozman

Review URL: https://codereview.chromium.org/478953002

git-svn-id: svn://svn.chromium.org/blink/trunk@180411 bbb929c8-8fbe-4397-9dbb-9b2b20218538

There's no need to repaint fixed position layers that are squashed. Necessarily
they must be squashed onto other fixed position layers, which means they don't
move during scrolling.

R=vollick@chromium.org

Review URL: https://codereview.chromium.org/478823002

git-svn-id: svn://svn.chromium.org/blink/trunk@180410 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The event should report DOM.RGBA color instead of string.

R=pfeldman@chromium.org, pfeldman

Review URL: https://codereview.chromium.org/478943002

git-svn-id: svn://svn.chromium.org/blink/trunk@180409 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove the HarfBuzzRun cache as using the run text value as the cache
isn't good enough when shaping across runs (support for which is coming
in a separate change) or when text rendering settings have changed (CSS
font-feature-settings).
Early testing indicates that the performance implications of removing
the cache are negligible as the overhead of maintaining the cache
negates most of the savings.

R=behdad@chromium.org,leviw@chromium.org
BUG=345401
TEST=PerformanceTests/Layout/hindi-line-layout.html

Review URL: https://codereview.chromium.org/167703008

git-svn-id: svn://svn.chromium.org/blink/trunk@180408 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently "https://" is accepted and treated like "https:". This behavior has never been part of any standard.

The syntax is specified in http://www.w3.org/TR/CSP11/#source-list-syntax

host-source       = [ scheme-part "://" ] host-part [ port-part ] [ path-part ]
host-part         = "*" / [ "*." ] 1*host-char *( "." 1*host-char )

As you can see, the host-part is NOT optional.

BUG=404295

Review URL: https://codereview.chromium.org/470723006

git-svn-id: svn://svn.chromium.org/blink/trunk@180407 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/471863002

git-svn-id: svn://svn.chromium.org/blink/trunk@180406 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Since we now use non-null to represent union member presence, either by
wrapping with Nullable<> or using the implicit null value (for pointers
and strings,) we must make sure the locals we pass to the implementation
are all properly initialized to null.

For strings and RefPtr<>/OwnPtr<>, this requires no special care due to
default constructors taking care of it. For RawPtr<> (when used in
practice, i.e. mostly in Oilpan builds) we need to initialize explicitly.

This fixes Oilpan crashes introduced by this CL:
  https://codereview.chromium.org/466323002/

Review URL: https://codereview.chromium.org/474323002

git-svn-id: svn://svn.chromium.org/blink/trunk@180405 bbb929c8-8fbe-4397-9dbb-9b2b20218538

More webkitpy cleanup.

webkitunittest isn't needed once we required Python 2.7.

mock.Mock() was really only being used in one file (and didn't need to be used
there, either), and I'm not generally fan of the generic Mock approach
it uses (preferring concrete mock objects or fakes).

R=abarth@chromium.org

Review URL: https://codereview.chromium.org/467133003

git-svn-id: svn://svn.chromium.org/blink/trunk@180404 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=404324
R=ojan@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/479723003

git-svn-id: svn://svn.chromium.org/blink/trunk@180402 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=ojan@chromium.org
NOTRY=true
BUG=399734

Review URL: https://codereview.chromium.org/476903003

git-svn-id: svn://svn.chromium.org/blink/trunk@180401 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Use tighter typing for Document::ownerElement() return value. It returns an
HTMLFrameOwnerElement* but a lot of callers assign it to an Element* variable.

This gets rid of a few unnecessary type checks and avoids an unnecessary
isHTMLElement() check in isHTMLFrameElementBase() in several cases.

R=adamk@chromium.org, rob.buis@samsung.com

Review URL: https://codereview.chromium.org/476223003

git-svn-id: svn://svn.chromium.org/blink/trunk@180400 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=388780

Review URL: https://codereview.chromium.org/460713002

git-svn-id: svn://svn.chromium.org/blink/trunk@180399 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Use Traversal<>::firstAncestor() API more in the code base to simplify the
code and increase readability.

R=adamk@chromium.org

Review URL: https://codereview.chromium.org/476273004

git-svn-id: svn://svn.chromium.org/blink/trunk@180398 bbb929c8-8fbe-4397-9dbb-9b2b20218538

webkitpy.common.editdistance, webkitpy.common.lru_cache, and
webkitpy.common.thread were no longer being used.

webkitpy.common.newstringio could be (and should've been) replaced by
io.StringIO() or io.BytesIO() as appropriate to indicate if you were
dealing with text or bytes.

R=abarth@chromium.org

Review URL: https://codereview.chromium.org/479633002

git-svn-id: svn://svn.chromium.org/blink/trunk@180397 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/472423002

git-svn-id: svn://svn.chromium.org/blink/trunk@180396 bbb929c8-8fbe-4397-9dbb-9b2b20218538

No one uses this value. We can remove it.

R=skobes@chromium.org

Review URL: https://codereview.chromium.org/480623002

git-svn-id: svn://svn.chromium.org/blink/trunk@180395 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Turns out we don't have any test coverage for this case.

R=leviw@chromium.org
BUG=403734

Review URL: https://codereview.chromium.org/474373002

git-svn-id: svn://svn.chromium.org/blink/trunk@180394 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This isn't awesome, but it was the best icon I could find and it
at least will be more clear since it's not using something that's
overloaded for opening menus.

BUG=404049
NOTRY=true

Review URL: https://codereview.chromium.org/475323002

git-svn-id: svn://svn.chromium.org/blink/trunk@180393 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This adds a preferCompositingToLCDText setting, and makes it be set by the
setAcceleratedCompositingForFixedPositionEnabled() WebSettings method so
that behaviour is not changed. Next, Chromium will be changed to call
setPreferCompositingToLCDTextEnabled instead so we can remove this method
from settings.

Later, other accelerated compositing trigger settings will be merged with
preferCompositingToLCDText.

R=abarth
BUG=365851

Review URL: https://codereview.chromium.org/476273003

git-svn-id: svn://svn.chromium.org/blink/trunk@180392 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This moves the expectations for the new w3c repos upstream
from test_expectations_w3c.txt.

This is patch 5 of 6 in the roll.

TBR=phoglund@chromium.org
BUG=404022

Review URL: https://codereview.chromium.org/469993005

git-svn-id: svn://svn.chromium.org/blink/trunk@180391 bbb929c8-8fbe-4397-9dbb-9b2b20218538

For from/to/by animations, the length of keyTimes and keyPoints
attribute if are specified must be equal and must have at least
2 values.

R=fs@opera.com,dschulze@chromium.org
BUG=377050

Review URL: https://codereview.chromium.org/453243003

git-svn-id: svn://svn.chromium.org/blink/trunk@180390 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When doing a tap, if the synthetic mousedown lands on some node, but
changes the DOM such that there's no node at that point afterwards
(eg. a frame scrollbar), then we'll pass a null node pointer to
commonAncestor and crash.

In addition to adding a test for this case (the mousedown case in
gesture-tap-frame-scrollbar) I added tests for a number of similar
cases (null hit test and removed node) that are already working
correctly but could be subject to similar bugs.

Also corrected a minor typo in a related test
(gesture-tap-click-common-ancestor) - doesn't change behavior, just
better reflects the intention of the test (the ancestor node itself
should not be hittable).

BUG=402338

Review URL: https://codereview.chromium.org/466143005

git-svn-id: svn://svn.chromium.org/blink/trunk@180389 bbb929c8-8fbe-4397-9dbb-9b2b20218538