GitLab

Followup to CL 486193002 to remove default arguments from template types,
which is only allowed in c++11 and later (and we haven't officially switched
to c++11 yet afaik).

Review URL: https://codereview.chromium.org/559733004

git-svn-id: svn://svn.chromium.org/blink/trunk@181719 bbb929c8-8fbe-4397-9dbb-9b2b20218538

If a garbage collected class is polymorphic (has virtual methods) it
needs to have a vtable from its left-most base class. Otherwise
pointers to the polymorphic subclasses are not guaranteed to point
to the beginning of the object.

This should have been caught by the gc plugin. We will investigate
that separately.

R=oilpan-reviews@chromium.org, sigbjornf@chromium.org, yhirano@chromium.org, zerny@chromium.org
BUG=411742

Review URL: https://codereview.chromium.org/562573002

git-svn-id: svn://svn.chromium.org/blink/trunk@181718 bbb929c8-8fbe-4397-9dbb-9b2b20218538

As all objects deriving from DOMWindowProperty are on the heap, make
it a GC mixin and have it trace its Window reference. The registered
DOMWindowProperty instances on a LocalDOMWindow live at least as long
it.

The benefit is that we can remove its destructor. Deriving objects can
as a consequence remove their empty finalizers.

R=ager@chromium.org, haraken@chromium.org
BUG=340522

Review URL: https://codereview.chromium.org/552733003

git-svn-id: svn://svn.chromium.org/blink/trunk@181717 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=396372,410229

Review URL: https://codereview.chromium.org/555063002

git-svn-id: svn://svn.chromium.org/blink/trunk@181716 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch does two things:

* The single parse-time UseCounter is now out of the parsing flow, and
  in the side-effects method.

* Parse-time ConsoleMessages generated in ContentSecurityPolicy and
  CSPDirectiveList are now stored on ContentSecurityPolicy and spun out
  to the console in the side-effects method.

BUG=411889

Review URL: https://codereview.chromium.org/554303002

git-svn-id: svn://svn.chromium.org/blink/trunk@181715 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove all the embedded resources from blink and migrate to using grd generated
pak files.

BUG=312586

R=haraken@chromium.org, jochen@chromium.org

Part 2: https://codereview.chromium.org/561653002
Part 3: https://codereview.chromium.org/560763002

Review URL: https://codereview.chromium.org/558933002

git-svn-id: svn://svn.chromium.org/blink/trunk@181714 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=181703

BUG=410706
TBR=h.joshi@samsung.com

Review URL: https://codereview.chromium.org/559923002

git-svn-id: svn://svn.chromium.org/blink/trunk@181713 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Point w3c/web-platform-tests/custom-elements/ failures to
appropriate bug.

Review URL: https://codereview.chromium.org/547383002

git-svn-id: svn://svn.chromium.org/blink/trunk@181712 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Method DatabaseContext::contextDestroyed always calls from ExecutionContext destructor. In the case when the context is WorkerGlobalScope - from WorkerGlobalScope destructor. This unregister call does nothing in this situation.
And it looks dangerous when we try to use the methods of the object in destructor. E.g. if we make the method ExecutionContext::isWorkerGlobalScope virtual the program will crash in DatabaseContext::contextDestroyed.

R=tkent@chromium.org,vsevik@chromium.org

Review URL: https://codereview.chromium.org/551373003

git-svn-id: svn://svn.chromium.org/blink/trunk@181711 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is a follow-up of Blink r181385.

BUG=402536

Review URL: https://codereview.chromium.org/553353002

git-svn-id: svn://svn.chromium.org/blink/trunk@181710 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In those the conversion functions that take an ExceptionState& argument,
always throw via the ExceptionState object, and never directly to V8.
Various existing callers assume this, and fail to detect exceptions that
were thrown directly to V8.

This misbehavior is a regression from
  https://codereview.chromium.org/313033002/
which changed the definition of the TONATIVE_DEFAULT_EXCEPTIONSTATE()
macro, but these functions were somewhat tricky to use even before then,
since exceptions were sometimes thrown to V8 and sometimes not, while
always stored in the ExceptionState object.

Review URL: https://codereview.chromium.org/559553003

git-svn-id: svn://svn.chromium.org/blink/trunk@181709 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When a node to reveal exists in the current profile show it right there
rather than switching to another profile.

Review URL: https://codereview.chromium.org/553063002

git-svn-id: svn://svn.chromium.org/blink/trunk@181708 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Reenable data URL layout tests which were temporarily skipped until Chromium side change gets landed

- Enables the temporarily skipped tests
- Change the blob test case to expect failure. This is because it turned
  out that we cannot use the optimization for data URL loading when XHR
  is set to load data as Blob.
- Delete the unnecessary expectation file

Depends on
- https://codereview.chromium.org/480413007/
- https://codereview.chromium.org/294193002/

BUG=308768

Review URL: https://codereview.chromium.org/492183002

git-svn-id: svn://svn.chromium.org/blink/trunk@181707 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Linux sometimes means android too.

R=brettw@chromium.org

Review URL: https://codereview.chromium.org/556143002

git-svn-id: svn://svn.chromium.org/blink/trunk@181706 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The spec[1] says the EventSource constructor should take EventSourceInit
dictionary as an argument. This CL doesn't change behavior. Tests under
http/tests/eventsource cover this change.

[1] http://www.w3.org/TR/eventsource/#eventsource

BUG=403150

Review URL: https://codereview.chromium.org/561533002

git-svn-id: svn://svn.chromium.org/blink/trunk@181705 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=396372

Review URL: https://codereview.chromium.org/556453004

git-svn-id: svn://svn.chromium.org/blink/trunk@181704 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Allowing drawText to return width of what was drawn, 
so that we don't have to make call to width to 
re-measure in Canvas drawing.

R=eae@chromium.org
BUG=410706
TEST=TEST=fast/canvas/canvas-drawtext-width-return.html

Review URL: https://codereview.chromium.org/519373002

git-svn-id: svn://svn.chromium.org/blink/trunk@181703 bbb929c8-8fbe-4397-9dbb-9b2b20218538

to oilpan's heap

- MediaStreamSource, MediaStreamComponent and MediaStreamDescriptor must be moved
to oilpan's heap in one go because their lifetime are tightly coupled.

- Removed m_descriptor->setClient(0) from MediaStream's destructor. This is OK
because MediaStreamDescriptor has a strong Member back to the Client
(i.e., MediaStream).

- Removed m_component->source()->removeObserver(this) from MediaStreamTrack's
destructor. This is OK because this CL made the observers weak
(i.e., MediaStreamSource::m_observers is a hash set of weak members to
MediaStreamTrack objects).

- Removed WebMediaStreamTrack::ExtraData::m_owner because it's unused.

- Introduced MediaStreamComponentDisposer to delay the destruction of
MediaStreamComponent::m_extraData. The ExtraData is exposed to the web
and some Chromium objects inherit from the ExtraData. If we clear the
m_extraData in the MediaStreamComponent's destructor, those Chromium-side
objects are also destructed in the MediaStreamComponent's destructor.
This is problematic because the destructors of the Chromium-side classes
can touch other on-heap objects in the Blink side. To avoid the issue,
we need to delay the destruction of the ExtraData to thread-specific
weak processing. The disposer pattern actually makes the deletion of
the extra data happen earlier and not later. The disposer makes sure
that the extra data is destructed in weak processing which is run before
sweeping and therefore all the objects are still alive and can be touched.


- The same problem arises for MediaStreamDescriptor and MediaStreamSource.
This CL adds MediaStreamDescriptorDisposer and MediaStreamSourceDisposer.

- This CL depends on https://codereview.chromium.org/543603003/ and https://codereview.chromium.org/549153002.

BUG=340522

Review URL: https://codereview.chromium.org/552653005

git-svn-id: svn://svn.chromium.org/blink/trunk@181702 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In getPositionAndOffset, previous block was used when space is entered for dir=auto. While for dir=auto text using previous block gives wrong caret position, which was leading to caret offset not calculated.  In case of RTL, previous inline box position is used.

Since auto scenario, combines RTL and LTR text. In renderText, only render block direction is used. New condition has been added in RenderText to test if dir=auto, then use inlineBlock bidiLevel to decide about caret position rendering.

R=leviw, eae
BUG=296847
TEST=Direction auto tests covering move caret position to the right when
LTR text is entered and adding RTL text to move caret to the left.

Review URL: https://codereview.chromium.org/541823003

git-svn-id: svn://svn.chromium.org/blink/trunk@181701 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Add another category of failure cards that display bots that have
been running or offline over a threshold of time. This is special-
cased for chromium.webkit & blink to start.

Also removing ct-revision-details since it's effectively obsoleted
by this change.

BUG=399957
NOTRY=true

Review URL: https://codereview.chromium.org/555263004

git-svn-id: svn://svn.chromium.org/blink/trunk@181700 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Don't compute all the keys for a loop that we almost
always early out from. Also, the early out was wrong because
it was just returning from the anonymous function.

NOTRY=true

Review URL: https://codereview.chromium.org/557173002

git-svn-id: svn://svn.chromium.org/blink/trunk@181699 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=396366

Review URL: https://codereview.chromium.org/562493002

git-svn-id: svn://svn.chromium.org/blink/trunk@181698 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Removing showValidationMessage, as code on the chromium side has landed.

R=tkent@chromium.org
BUG=91638

Review URL: https://codereview.chromium.org/557753002

git-svn-id: svn://svn.chromium.org/blink/trunk@181697 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This will guard the next shipping subset of the Web Animations API.

BUG=396369

Review URL: https://codereview.chromium.org/538503002

git-svn-id: svn://svn.chromium.org/blink/trunk@181696 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Previously we supported two conventions for naming virtual test suites in
the layout tests. The original convention was to map a subpath onto a
different subpath, e.g. "gpu/compositedscrolling/overflow" would map onto
the tests in "compositing/overflow". Later we started to just use a prefix
plus matching the base subpath exactly, so the prior example would instead be
refered to as "gpu/compositing/overflow".

There were two virtual test suites remaining that used the original syntax.
This patch renames those suites and removes the code that supported the syntax.

gpu/compositedscrolling/overflow -> prefer_composited_to_lcd_text/compositing/overflow
gpu/compositedscrolling/scrollbars -> prefer_composited_to_lcd_text/scrollbars

R=enne@chromium.org, ojan@chromium.org
BUG=410210

Review URL: https://codereview.chromium.org/560563005

git-svn-id: svn://svn.chromium.org/blink/trunk@181695 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Since these V8DOMActivityLogger's methods can be called from core/
(not via bindings/), it makes sense to allocate a new HandleScope
in these methods.

BUG=411445
NOTRY=true

Review URL: https://codereview.chromium.org/559813002

git-svn-id: svn://svn.chromium.org/blink/trunk@181694 bbb929c8-8fbe-4397-9dbb-9b2b20218538

AudioNodeOutput::disable() can be reentered and causes the following crash.

ASSERTION FAILED: m_outputs.contains(&output)
../../third_party/WebKit/Source/modules/webaudio/AudioNodeInput.cpp(98) : void blink::AudioNodeInput::disable(blink::AudioNodeOutput &)
Received signal 11 SEGV_MAPERR 0000fbadbeef

To avoid AudioNodeOutput::disable() from getting reentered, this CL moves the place
of flipping m_isEnabled.

The same issue is in AudioNodeOutput::enable() and this CL addresses the issue as well.

BUG=411989
TEST=None, it's hard to create a simple test html that reproduces the bug.

Review URL: https://codereview.chromium.org/558523002

git-svn-id: svn://svn.chromium.org/blink/trunk@181693 bbb929c8-8fbe-4397-9dbb-9b2b20218538

ServiceWorkerRegistration::unregister returns false when the registration is already unregistered. (1/3)

1. [Blink] this patch
2. [Chromium] https://codereview.chromium.org/554573002/
3. [Blink] https://codereview.chromium.org/550093002/

BUG=390894
TEST=run_webkit_tests.py http/tests/serviceworker/unregister.html

Review URL: https://codereview.chromium.org/549093002

git-svn-id: svn://svn.chromium.org/blink/trunk@181692 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=412570
TBR=wangxianzhu@chromium.org

Review URL: https://codereview.chromium.org/557963002

git-svn-id: svn://svn.chromium.org/blink/trunk@181691 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is an experimental utility based on autopep8 and lib2to3. 

It is capable of reformatting to either the Blink or Chromium style guides (the difference between the two is that Blink uses 132c lines, unix_hacker naming for functions and methods, a 4-space indent, and prefers single-quoted string literals, whereas Chromium uses 80c lines, InitialCaps for functions and methods, a 2-space indent, and double-quoted literals).

Note that format-webkitpy does *not* convert method names between the two conventions, since that is a potentially unsafe thing to do. It is probably possible to write a lib2to3 fixer that does do this relatively safely, but that was more work than I felt like doing for an initial go.

As part of this, we add autopep8 into webkitpy third_party (lib2to3 is part of the standard lib) and update pep8, which hadn't been updated for years.

This patch also adds some simple helper methods to Executive() to do parallel map calls using multiprocessing (and to mock them out properly) so that client code doesn't need to be aware of the hoop-jumping you need to do since hosts aren't picklable. This part of the code is still a little crufty.

BUG=

Review URL: https://codereview.chromium.org/546613003

git-svn-id: svn://svn.chromium.org/blink/trunk@181690 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It is a concept that doesn't make sense for a remote frame (whether the load
event has fired), and all callsites are using WebLocalFrames

BUG=

Review URL: https://codereview.chromium.org/555953002

git-svn-id: svn://svn.chromium.org/blink/trunk@181689 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Limit the noSubpixelForSmallSizeFont functionality that disables
subpixel text for older fonts at sizes below 16px to low-DPI devices.
At or above a 150% DPI subpixel text is enabled for all fonts assuming
anti-aliasing is enabled.

As the font code isn't aware of the DPI or device pixel ratio it needs
to be supplied by Chrome though WebFontRendering.

Chrome side change is in https://codereview.chromium.org/561503002/

R=cpu@chromium.org, dglazkov@chromium.org
BUG=405445

Review URL: https://codereview.chromium.org/553163002

git-svn-id: svn://svn.chromium.org/blink/trunk@181688 bbb929c8-8fbe-4397-9dbb-9b2b20218538

dcheck_always_on=1 has the same effect now.

BUG=411008

Review URL: https://codereview.chromium.org/556683003

git-svn-id: svn://svn.chromium.org/blink/trunk@181687 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=181676

TBR=esprehn@chromium.org

Review URL: https://codereview.chromium.org/559783002

git-svn-id: svn://svn.chromium.org/blink/trunk@181686 bbb929c8-8fbe-4397-9dbb-9b2b20218538

https://codereview.chromium.org/511253003/ replaced applyScrollAndScale
with applyViewportDeltas. This method can now be removed from the Blink
API.

BUG=333143

Review URL: https://codereview.chromium.org/557103002

git-svn-id: svn://svn.chromium.org/blink/trunk@181685 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch enables background html parser for documents with no frames,
which are XMLHttpRequest response documents. HTML imports and
inspector/DOMPatchSupport will still to use the synchronous mode as the
current code expect the change to happen synchronously.

Before this change, XMLHttpRequest immediately set its state to DONE after
receiving the last data. After this change, XMLHttpRequest will wait for
|Document::finishedParsing| before changing its state to DONE if it has
responseType == "document". This is needed because |DocumentParser::finish()|
do not synchronously give the complete document, but we have to wait until all
tokens from the background parser is processed.

BUG=409461

Review URL: https://codereview.chromium.org/521363002

git-svn-id: svn://svn.chromium.org/blink/trunk@181684 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Removed an unused CompositionUnderline constructor.

Review URL: https://codereview.chromium.org/550943002

git-svn-id: svn://svn.chromium.org/blink/trunk@181682 bbb929c8-8fbe-4397-9dbb-9b2b20218538

to keep figcation infomation on accessibility tree.

BUG=None

Review URL: https://codereview.chromium.org/553033002

git-svn-id: svn://svn.chromium.org/blink/trunk@181681 bbb929c8-8fbe-4397-9dbb-9b2b20218538

All of these measure stable features except CSSPropertyWebkitCursorVisibility which
is actually removed.

Review URL: https://codereview.chromium.org/558473002

git-svn-id: svn://svn.chromium.org/blink/trunk@181680 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=411864
TEST=blink unit test Canvas2DLayerBridgeTest.testPrepareMailboxAndLoseResource

Review URL: https://codereview.chromium.org/552423002

git-svn-id: svn://svn.chromium.org/blink/trunk@181679 bbb929c8-8fbe-4397-9dbb-9b2b20218538