GitLab

BUG=403876

Review URL: https://codereview.chromium.org/485563003

git-svn-id: svn://svn.chromium.org/blink/trunk@180476 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=
NOTRY=true

Review URL: https://codereview.chromium.org/473553003

git-svn-id: svn://svn.chromium.org/blink/trunk@180475 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We are doing paint invalidations after compositing updates
so nothing blocks enabling it anymore.

BUG=360286

Review URL: https://codereview.chromium.org/479563006

git-svn-id: svn://svn.chromium.org/blink/trunk@180474 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is patch 3 of 4.

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

Review URL: https://codereview.chromium.org/479693003

git-svn-id: svn://svn.chromium.org/blink/trunk@180473 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=403289

Review URL: https://codereview.chromium.org/485643002

git-svn-id: svn://svn.chromium.org/blink/trunk@180472 bbb929c8-8fbe-4397-9dbb-9b2b20218538

RenderBox's methods for computing logical height would use
frameView()->visibleWidth() and visibleHeight() once they
reach the RenderView. These methods return the frame's
dimensions divided by the content scale factor (i.e. pinch-zoom).
This was causing problems in the linked bug since the method
would remove scrollbars before applying scale, causing the removed
scrollbar area to scale with the page. This CL changes the
occurances of visibleWidth/Height to unscaledVisibleContentSize().

BUG=362337

Review URL: https://codereview.chromium.org/462243002

git-svn-id: svn://svn.chromium.org/blink/trunk@180471 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Change HarfBuzzShaper to shape text across HarfBuzzRuns. This allows
for better handling of zero width joiners and is the first step towards
supporting shaping across elements.

Must land after change 167703008 as without that change it will cache
incorrect information due to the text string value no longer being
sufficient as shaping now includes contextual information from adjacent
runs.

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

Review URL: https://codereview.chromium.org/478823003

git-svn-id: svn://svn.chromium.org/blink/trunk@180470 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=339798
R=vsevik, apavlov

Review URL: https://codereview.chromium.org/485503002

git-svn-id: svn://svn.chromium.org/blink/trunk@180469 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=404375
R=vsevik@chromium.org

Review URL: https://codereview.chromium.org/486573003

git-svn-id: svn://svn.chromium.org/blink/trunk@180468 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The uglifier tokenizer, which is used under the hood in javascript
formatter, is not clever enough to figure out that the "return" word
in "a.return" is not a "return" statement, but identifier.
This patch works around this issue via verifying the token value after
regex to be a valid identifier.

BUG=218470
R=vsevik, apavlov

Review URL: https://codereview.chromium.org/479033002

git-svn-id: svn://svn.chromium.org/blink/trunk@180467 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove the CompositedScrollingForFrames setting and change any callers
to use PreferCompositingToLCDText instead.

R=abarth
BUG=365851

Review URL: https://codereview.chromium.org/480803002

git-svn-id: svn://svn.chromium.org/blink/trunk@180466 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The property 'clip' is (with some modifications) compatible with the
apply_auto-properties.

Changes in this commit:

 * Use apply_auto + converter instead of custom building functions.
 * StyleVisualData::hasClip -> hasAutoClip (meaning reversed).
 * Remove unnecessary RenderStyle::setClip*-functions.

Review URL: https://codereview.chromium.org/462133002

git-svn-id: svn://svn.chromium.org/blink/trunk@180465 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=none

Review URL: https://codereview.chromium.org/481913002

git-svn-id: svn://svn.chromium.org/blink/trunk@180464 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Ids, attributes, tag names, and custom pseudo elements are handled as
invalidation set features and not skippable. Should be unnecessary to check
for them in isSkippableComponentForInvalidation.

R=chrishtr@chromium.org, esprehn@chromium.org

Review URL: https://codereview.chromium.org/450293002

git-svn-id: svn://svn.chromium.org/blink/trunk@180463 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The patch adds missing inputs to the script_formatter_worker module
target.

R=apavlov

Review URL: https://codereview.chromium.org/481933002

git-svn-id: svn://svn.chromium.org/blink/trunk@180462 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Add missing implementation of pushAllRegisters for MIPS64.

BUG=400684

Review URL: https://codereview.chromium.org/476143002

git-svn-id: svn://svn.chromium.org/blink/trunk@180461 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://fullscreen.spec.whatwg.org/#fullscreen-element-ready-check
https://www.w3.org/Bugs/Public/show_bug.cgi?id=26568

BUG=403741

Review URL: https://codereview.chromium.org/485443002

git-svn-id: svn://svn.chromium.org/blink/trunk@180460 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Enumeration types are effectively strings, and should thus be handled the
same, meaning they don't need a v8::TryCatch object.

This changes generated code for a few methods that take at least one
enumeration type argument, and don't otherwise need a v8::TryCatch, by
eliminating the unnecessary v8::TryCatch locals.

Also refactor slightly: pass in the IdlOperation object and check whether
the return type is Promise inside argument_needs_try_catch(), instead
of passing in that fact as an argument. This keeps the logic localized in
argument_needs_try_catch().

R=haraken@chromium.org
BUG=

Review URL: https://codereview.chromium.org/482713002

git-svn-id: svn://svn.chromium.org/blink/trunk@180459 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/393123007

Review URL: https://codereview.chromium.org/393123007

git-svn-id: svn://svn.chromium.org/blink/trunk@180458 bbb929c8-8fbe-4397-9dbb-9b2b20218538

(1) This CL basically does the following two things:

- Drop "WillBe" types from webaudio/.

- Replace the hand-written reference counting system with RefCountedGarbageCollected.

(2) Given that AudioNode inherits from EventTarget, which is still on-heap
in non-oilpan builds, we need to make AudioNode RefCountedGarbageCollected.

(3) Ideally the following collections should use strong Members,
but we cannot do that because these collections are touched by audio threads
(which are not registered to oilpan). Thus we use raw pointers
with GC_PLUGIN_IGNORE.

- AudioNodeInput::m_disabledOutputs
- AudioSummingJunction::m_outputs
- AudioSummingJunction::m_renderingOutputs
- AudioContext::m_finishedNodes
- AudioContext::m_referencedNodes
- AudioContext::m_nodesMarkForDeletion
- AudioContext::m_nodesToDelete
- AudioContext::m_dirtySummingJunctions
- AudioContext::m_dirtyAudioNodeOutputs
- AudioContext::m_automaticPullNodes
- AudioContext::m_renderingAutomaticPullNodes
- AudioContext::m_deferredBreakConnectionList
- AudioContext::m_deferredFinishDerefList

(4) Ideally AudioDSPKernel should be moved to the heap and
AudioDSPKernel::m_kernelProcessor should be a Member.
However, we cannot do that because AudioDSPKernel can be allocated
by audio threads. Thus we leave AudioDSPKernel::m_kernelProcessor
as a raw pointer. This raw pointer is guaranteed to be safe.

(5) This CL fixes leaks of webaudio tests in LeakExpectations.


BUG=340522

Review URL: https://codereview.chromium.org/438293003

git-svn-id: svn://svn.chromium.org/blink/trunk@180457 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=eustas, lushnikov
BUG=404555

Review URL: https://codereview.chromium.org/486633002

git-svn-id: svn://svn.chromium.org/blink/trunk@180456 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of XMLHttpRequest::overrideMimeType should throw InvalidStateError (patchset #4 of https://codereview.chromium.org/478223002/)

Reason for revert:
It looks like this broke a number of Chromium-side browser tests: See http://build.chromium.org/p/chromium.webkit/builders/Linux%20Tests%20%28dbg%29/builds/3608 for example. http://test-results.appspot.com/dashboards/flakiness_dashboard.html#group=%40ToT%20Blink&testType=browser_tests&tests=ExtensionBrowserTest.RSSParseFeedInvalidFeed1,ExtensionBrowserTest.RSSParseFeedInvalidFeed2,ExtensionBrowserTest.RSSParseFeedInvalidFeed3,ExtensionBrowserTest.RSSParseFeedInvalidFeed4,ExtensionBrowserTest.RSSParseFeedValidFeed0,ExtensionBrowserTest.RSSParseFeedValidFeed2,ExtensionBrowserTest.RSSParseFeedValidFeed3,ExtensionBrowserTest.RSSParseFeedValidFeed4,ExtensionBrowserTest.RSSParseFeedValidFeed5,ExtensionBrowserTest.RSSParseFeedValidFeed6,ExtensionBrowserTest.RSSParseFeedValidFeedNoLinks for all the broken tests.

Original issue's description:
> XMLHttpRequest::overrideMimeType should throw InvalidStateError
> 
> Before this CL, XMLHttpRequest accepted |overrideMimeType| regardless
> of its state. This is not conformant to the current spec:
> http://www.w3.org/TR/XMLHttpRequest2/#the-overridemimetype-method
> 
> This CL changes it to throw InvalidStateError when the state
> is LOADING or DONE, which aligns to the spec.
> 
> BUG=402375
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180440

TBR=tyoshino@chromium.org,abarth@chromium.org,haraken@chromium.org,kouhei@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=402375

Review URL: https://codereview.chromium.org/486703002

git-svn-id: svn://svn.chromium.org/blink/trunk@180455 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Let generated bindings code handle the type checking of Blob arguments.

R=tzik@chromium.org
BUG=

Review URL: https://codereview.chromium.org/474353004

git-svn-id: svn://svn.chromium.org/blink/trunk@180454 bbb929c8-8fbe-4397-9dbb-9b2b20218538

There was a left-over condition in the TextPrompt which has never being
satisfied. As a result, the WebInspector.TextPrompt.Events.ItemAccepted
event has never being fired.

BUG=398224

Review URL: https://codereview.chromium.org/477593005

git-svn-id: svn://svn.chromium.org/blink/trunk@180453 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The _module.js files are currently used only at build-time in the
release mode and at run-time when starting workers.
We have all the required module info in module.json files,
and workers can be loaded in the debug mode through a blob containing
concatenated scripts for the worker, loaded via sync XHRs.
This patch also gets rid of common/modules.js: module descriptors
are now inlined into Runtime.js.

R=pfeldman, vsevik
BUG=391566

Review URL: https://codereview.chromium.org/472903003

git-svn-id: svn://svn.chromium.org/blink/trunk@180452 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In preparation to switching to tracing backend the test is changed to use higher level abstractions.

BUG=398785
R=pfeldman@chromium.org

Review URL: https://codereview.chromium.org/480013002

git-svn-id: svn://svn.chromium.org/blink/trunk@180451 bbb929c8-8fbe-4397-9dbb-9b2b20218538

If a FileReader is abort()ed, a task is scheduled to perform the actual
abort operation. Should the containing Document's ActiveDOMObjects be
stopped before that task gets to run, the FileReader would already
advance to a DONE state and be at risk from being GCed before the abort
task gets to run. If so, it would then access a dead object.

The provided test elicits an assert that shows up the problem, but
doesn't trigger the GC and subsequent access of the dead object.

R=kouhei@chromium.org, tzik@chromium.org
BUG=404513

Review URL: https://codereview.chromium.org/478263002

git-svn-id: svn://svn.chromium.org/blink/trunk@180450 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Rendering gradient consumes considerable amount of time.

BUG=316092

Review URL: https://codereview.chromium.org/440853002

git-svn-id: svn://svn.chromium.org/blink/trunk@180449 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The patch fixes inspector/audits/audits-panel-noimages-functional.html
and inspector/audits/audits-panel-functional.html

BUG=364088, 357079
R=apavlov

Review URL: https://codereview.chromium.org/483683002

git-svn-id: svn://svn.chromium.org/blink/trunk@180448 bbb929c8-8fbe-4397-9dbb-9b2b20218538

guidelines and improved performance in WebViewImpl.cpp a bit.

Review URL: https://codereview.chromium.org/479803002

git-svn-id: svn://svn.chromium.org/blink/trunk@180447 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=398840

Review URL: https://codereview.chromium.org/483663002

git-svn-id: svn://svn.chromium.org/blink/trunk@180446 bbb929c8-8fbe-4397-9dbb-9b2b20218538

 - Removes unnecessary blink:: prefixes.
 - Removes *using namespace* usage.

BUG=

Review URL: https://codereview.chromium.org/468083003

git-svn-id: svn://svn.chromium.org/blink/trunk@180445 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://fullscreen.spec.whatwg.org/#dom-element-requestfullscreen

This spec change was made in order to simplify the fix for mis-nested
fullscreen in iframes:
https://www.w3.org/Bugs/Public/show_bug.cgi?id=26568 (comments 9-13)

The requestFullscreen() implementation is not in sync with the spec, but
an early return is equivalent until the "run the remaining steps
asynchronously" bit is implemented.

In order to be affected by this change, one would have to request
fullscreen for the same element twice and do something meaningful in the
fullscreenerror event, which is likely rare.

Note that exitFullscreen() already does nothing if the fullscreen
element stack is empty, so there is some symmetry to this.

TEST=LayoutTests/fullscreen/api/element-request-fullscreen-top.html
TEST=LayoutTests/fullscreen/api/element-request-fullscreen-non-top.html

BUG=403741

Review URL: https://codereview.chromium.org/482543002

git-svn-id: svn://svn.chromium.org/blink/trunk@180444 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The method is always defined, no need for additional checks in inspector tests.

BUG=None

Review URL: https://codereview.chromium.org/486543003

git-svn-id: svn://svn.chromium.org/blink/trunk@180443 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The PresentationConsoleMessage management is moved from UISourceCode
into PresentationConsoleMessageHelper.

R=eustas, vsevik

Review URL: https://codereview.chromium.org/472353002

git-svn-id: svn://svn.chromium.org/blink/trunk@180442 bbb929c8-8fbe-4397-9dbb-9b2b20218538

WorkerGlobalScopeProxy is dedicated worker specific, while this interface could be reused for
shared/service workers as well.

BUG=404355
R=pfeldman, yurys, sergeyv

Review URL: https://codereview.chromium.org/477313002

git-svn-id: svn://svn.chromium.org/blink/trunk@180441 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Before this CL, XMLHttpRequest accepted |overrideMimeType| regardless
of its state. This is not conformant to the current spec:
http://www.w3.org/TR/XMLHttpRequest2/#the-overridemimetype-method

This CL changes it to throw InvalidStateError when the state
is LOADING or DONE, which aligns to the spec.

BUG=402375

Review URL: https://codereview.chromium.org/478223002

git-svn-id: svn://svn.chromium.org/blink/trunk@180440 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It's equivalent to webkitCurrentFullScreenElement != null.

BUG=383813

Review URL: https://codereview.chromium.org/476353002

git-svn-id: svn://svn.chromium.org/blink/trunk@180439 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We are going to switch to tracing backend for the timeline and there won't be TimelineAgent so the tests are changed to use higher level abstractions.

BUG=398785

Review URL: https://codereview.chromium.org/480003002

git-svn-id: svn://svn.chromium.org/blink/trunk@180438 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch adds support for the 'converter' attribute for a group of
semi-custom CSS properties handled by the 'apply_auto' macro. This gets
rid of the redundant 'compute_length=true/false' argument, plus the code
related to that.

It also makes it easier to add other things to the apply_auto-list which
can't be converted with computeLength or a primitive value mapping.

Review URL: https://codereview.chromium.org/454133004

git-svn-id: svn://svn.chromium.org/blink/trunk@180437 bbb929c8-8fbe-4397-9dbb-9b2b20218538