GitLab

Verify that a very large shape-margin value combined with a very small
shape-outside image defines a shape that matches the shape-outside
element's margin-box.

This is a merge of http://trac.webkit.org/changeset/166019 by
Hans Muller <hmuller@adobe.com>.

BUG=354917

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=169768

Review URL: https://codereview.chromium.org/208423003

git-svn-id: svn://svn.chromium.org/blink/trunk@169877 bbb929c8-8fbe-4397-9dbb-9b2b20218538

git-svn-id: svn://svn.chromium.org/blink/trunk@169876 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Update image data upon changing the img.crossorigin attribute. (https://codereview.chromium.org/201143002/)

Reason for revert:
The re-issuing of another image load upon setting 'crossorigin' has been reported to cause a site problem (a Google internal site/service.)

Until diagnosed and understood fully, let's back this one out.

Original issue's description:
> Update image data upon changing the img.crossorigin attribute.
> 
> The spec tells us that setting the crossorigin attribute on an img
> element should cause the "update the image data" steps to be invoked,
> 
>  http://www.whatwg.org/specs/web-apps/current-work/#update-the-image-data
> 
> Do so here, along with updating the underlying test's expectation.
> 
> R=
> BUG=341047
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=169382

TBR=adamk@chromium.org,abarth@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=341047

Review URL: https://codereview.chromium.org/210403002

git-svn-id: svn://svn.chromium.org/blink/trunk@169874 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://www.whatwg.org/specs/web-apps/current-work/#attr-script-type

  The type attribute gives the language of the script or format of the
  data. If the attribute is present, its value must be a valid MIME
  type. The charset parameter must not be specified. The default,
  which is used if the attribute is absent, is "text/javascript"."

Blink is less strict than other browsers. Removing the optional type
attribute enables tests to be run in other browsers, notably Firefox.

BUG=308992
NOTRY=true

Review URL: https://codereview.chromium.org/209333007

git-svn-id: svn://svn.chromium.org/blink/trunk@169872 bbb929c8-8fbe-4397-9dbb-9b2b20218538

These map to the eventnames 'beginEvent', 'endEvent' and 'repeatEvent'
respectively.

Spec:
https://svgwg.org/svg2-draft/animate.html#InterfaceSVGAnimationElement

BUG=338288
NOTRY=true

Review URL: https://codereview.chromium.org/201673003

git-svn-id: svn://svn.chromium.org/blink/trunk@169871 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We were early'ing out, but not calling cache->setScrollParent. The result: the
cache thought it was dirty, but it wasn't. This CL fixes that.

BUG=TBD

Review URL: https://codereview.chromium.org/209663007

git-svn-id: svn://svn.chromium.org/blink/trunk@169870 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Mark the following layout test as flaky on windows debug builds:
  virtual/threaded/animations/animation-hit-test-transform.html
  virtual/threaded/transitions/blendmode-transitions.html
  virtual/threaded/transitions/cubic-bezier-overflow-svg-length.html
  virtual/threaded/animations/compositor-start-event-timing.html

BUG=248938
TBR=ajuma@chromium.org

Review URL: https://codereview.chromium.org/208393018

git-svn-id: svn://svn.chromium.org/blink/trunk@169869 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove paintSkiaText wrapper function that is no longer needed now that we always have SkTypeface.

R=bungeman@chromium.org

Review URL: https://codereview.chromium.org/209113002

git-svn-id: svn://svn.chromium.org/blink/trunk@169868 bbb929c8-8fbe-4397-9dbb-9b2b20218538

These files do not appear to be referenced in the build and can therefore be
deleted.

R=adamk@chromium.org

Review URL: https://codereview.chromium.org/199793020

git-svn-id: svn://svn.chromium.org/blink/trunk@169867 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove has3DContent, layerHas3DContent
The last vestiges of these functions were removed with
https://src.chromium.org/viewvc/blink?revision=167695&view=revision
when Abarth removed isSoftwareRenderable.

Remove computeFixedViewportConstraints, computeStickyViewportConstraints
Jamesr removed the last callers in
https://src.chromium.org/viewvc/blink?view=rev&revision=147653

Remove clearMappingForAllRenderLayers, clearMappingForRenderLayerIncludingDescendants
These functions were added (under a different name) by Simon Fraser in
http://trac.webkit.org/changeset/90484. Jamesr removed the caller
in https://codereview.chromium.org/13818029.

Remove setCompositingParent, removeCompositedChildren
This code predates written history and possibly humankind.

Remove updateLayerTreeGeometry
Abarth removed the implementation but left the function definition in
https://src.chromium.org/viewvc/blink?view=rev&revision=168244

Remove updateGraphicsLayersMappedToRenderLayer
Abarth removed the implementation but left the function definition in
https://src.chromium.org/viewvc/blink?view=rev&revision=167697

Remove hasAnyAdditionalCompositedLayers
Shawn Singh removed the implementation but left the function definition
in https://src.chromium.org/viewvc/blink?view=rev&revision=163720

Review URL: https://codereview.chromium.org/209703002

git-svn-id: svn://svn.chromium.org/blink/trunk@169866 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Before the change 'WTF::copyChars' contained an algorithm that was used
instead of 'memcpy' if the number of characters had been less than 20.

This however was unnecessary as 'memcpy' produces more efficient asm code
than the home-brewed algorithm did as it uses 'movq' (instead of 'movl')
doing the copying within the less number of asm commands (tested with
GCC 4.6.3).

So, this patch removes the home-brewed algorithm from 'WTF::copyChars'
and makes it always rely on 'memcpy'.

A simple testing programm doing the following code

    for (int i = 10000000; --i;) {
        // Re-create 'src' to prevent from optimizing cycle iterations out.
        char* src = new char[15];
        copyChars(dest, src, 15);
        copyChars(dest, src, 14);
        delete[] src;
    }

takes:

0m0.324s before the change
0m0.235s after the change (using 'memcpy').

Review URL: https://codereview.chromium.org/210013002

git-svn-id: svn://svn.chromium.org/blink/trunk@169865 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=353991
NOTRY=true

Review URL: https://codereview.chromium.org/208393016

git-svn-id: svn://svn.chromium.org/blink/trunk@169864 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=304341

Review URL: https://codereview.chromium.org/208923002

git-svn-id: svn://svn.chromium.org/blink/trunk@169863 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Mark the virtual/threaded/animations/3d/transform-perspective.html test
as timing out on all debug bots as per flakiness dashboard data.

BUG=248938
TBR=ajuma@chromium.org

Review URL: https://codereview.chromium.org/200123004

git-svn-id: svn://svn.chromium.org/blink/trunk@169862 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch updates User-Agent for Samsung Galaxy S4.

BUG=354920
TBR=pfeldman, apavlov

Review URL: https://codereview.chromium.org/210043002

git-svn-id: svn://svn.chromium.org/blink/trunk@169861 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The user can search styles by property name, value, or rule selector.
Computed style properties are filtered, so that only matching ones are visible.

R=eustas, pfeldman, vsevik
BUG=278852

Review URL: https://codereview.chromium.org/129793013

git-svn-id: svn://svn.chromium.org/blink/trunk@169860 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The new interface is called createWebAudioSourceFromMediaStreamTrack(), it is used to create one WebAudioSource for each CreateMediaStreamSource() call. By doing this, we avoid calling into the same source provider when there are multiple WebAudio audioContexts.

Note that the code in this is not hooked up to anything. We need to implement the interface in Chrome then hook the code up in AudioContext.cpp

NOTRY=true

BUG=354468

Review URL: https://codereview.chromium.org/204293007

git-svn-id: svn://svn.chromium.org/blink/trunk@169859 bbb929c8-8fbe-4397-9dbb-9b2b20218538

consolidates that into a separate file and also updates the generated gesture
sequences to be valid.

BUG=352164

Review URL: https://codereview.chromium.org/197923006

git-svn-id: svn://svn.chromium.org/blink/trunk@169858 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/210053002

git-svn-id: svn://svn.chromium.org/blink/trunk@169857 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Changing the allocation of the initaial SkSurface in
Canvas2DLayerBridge so that it may be taken from the GrContext
texture pool. This should improve performance for apps that
rely heavily on transient canvas objects.

BUG=351798

Review URL: https://codereview.chromium.org/196833006

git-svn-id: svn://svn.chromium.org/blink/trunk@169856 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/196133031

git-svn-id: svn://svn.chromium.org/blink/trunk@169855 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Getting scopes info from V8 debugger takes significant time.
Sometimes we don't need the scopes at all, thus we can save on this expensive work.
For async stacks we will be getting "fast" scope chains: up to the Local scope only (coming soon).

R=yurys

Review URL: https://codereview.chromium.org/208263015

git-svn-id: svn://svn.chromium.org/blink/trunk@169854 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The issue that required this workaround was fixed in
https://codereview.chromium.org/204983002/.

BUG=328156

Review URL: https://codereview.chromium.org/209963003

git-svn-id: svn://svn.chromium.org/blink/trunk@169853 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=pfeldman, vsevik
BUG=353916

Review URL: https://codereview.chromium.org/198713006

git-svn-id: svn://svn.chromium.org/blink/trunk@169851 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Added NULL check.

BUG=354958
TEST=Check there is no crashes when the minimized test case is passed to
content_shell

Review URL: https://codereview.chromium.org/201573008

git-svn-id: svn://svn.chromium.org/blink/trunk@169849 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=None

Review URL: https://codereview.chromium.org/209813003

git-svn-id: svn://svn.chromium.org/blink/trunk@169848 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=haraken@chromium.org
BUG=

Review URL: https://codereview.chromium.org/208313010

git-svn-id: svn://svn.chromium.org/blink/trunk@169847 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=354405
R=ager@chromium.org

Review URL: https://codereview.chromium.org/209843002

git-svn-id: svn://svn.chromium.org/blink/trunk@169846 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The interface describes common part of nodes as well as containment and retainment edges. It used to iterate over the elements and serialize them.

BUG=None

Review URL: https://codereview.chromium.org/209653003

git-svn-id: svn://svn.chromium.org/blink/trunk@169845 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This will distinguish between setting the value on a newly created Attr
object from Document.createAttribute() and setting the value on an Attr
object already attached to an Element, e.g. one from Element.attributes.

If the first case is common but the second is rare it won't be possible
to make value readonly, but it could still be possible to drop the
Element association, which is the point of making it readonly.

BUG=none

Review URL: https://codereview.chromium.org/208333014

git-svn-id: svn://svn.chromium.org/blink/trunk@169844 bbb929c8-8fbe-4397-9dbb-9b2b20218538

While distributing the extra rowspan height in rows, result of multiplication
is overflowing in integer and whole method is getting failed due to this. So
typecasted a value to 'long long' and final result of the calculation always
fits in integer range.

BUG=334652
R=jchaffraix@chromium.org

Review URL: https://codereview.chromium.org/131103017

git-svn-id: svn://svn.chromium.org/blink/trunk@169843 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=pfeldman, vsevik
BUG=354853

Review URL: https://codereview.chromium.org/208223002

git-svn-id: svn://svn.chromium.org/blink/trunk@169842 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The overload resolution needed for the two FormData.append() methods can
be handled directly by generated code, so drop the custom code we had
in place for this API.

Along with it, handle the conversion of append()'s DOMString arguments
per WebIDL (as it is now), removing the special treatment of null and
undefined arguments. This also aligns with how others (Gecko, Trident)
handle these arguments.

R=haraken@chromium.org
BUG=

Review URL: https://codereview.chromium.org/209933002

git-svn-id: svn://svn.chromium.org/blink/trunk@169841 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=305842
TBR=pfeldman

Review URL: https://codereview.chromium.org/209423003

git-svn-id: svn://svn.chromium.org/blink/trunk@169840 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Follow up on crrev.com/206793006 and also call wrap() in the code generated
for a particular constructor overloading when [Custom=Wrap] is in effect
for one or more constructors.

R=nbarth@chromium.org,haraken@chromium.org
BUG=338804

Review URL: https://codereview.chromium.org/209893002

git-svn-id: svn://svn.chromium.org/blink/trunk@169839 bbb929c8-8fbe-4397-9dbb-9b2b20218538

ElementStyleResources are only allocated as a part object of another stack allocated object so we don't need
to use persistent heap collections as the entries will be marked conservatively via the stack.

R=ager@chromium.org, erik.corry@gmail.com, haraken@chromium.org, oilpan-reviews@chromium.org, tkent@chromium.org, vegorov@chromium.org, zerny@chromium.org
BUG=341815

Review URL: https://codereview.chromium.org/208053004

git-svn-id: svn://svn.chromium.org/blink/trunk@169838 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=355489
TBR=philipj

Review URL: https://codereview.chromium.org/209913002

git-svn-id: svn://svn.chromium.org/blink/trunk@169837 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch fixes CSSParser which failed to properly report last
parsed chunk.

BUG=329825

Review URL: https://codereview.chromium.org/208593002

git-svn-id: svn://svn.chromium.org/blink/trunk@169836 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Adds support for constructors when there's a custom wrap.

This is split off as 2/2 bindings-only change from:
Implement ImageData constructors.
https://codereview.chromium.org/196343032/

Follow-up to:
Use wrapper_configuration consistently in bindings
https://codereview.chromium.org/199633014/

R=haraken
BUG=338804

Review URL: https://codereview.chromium.org/206793006

git-svn-id: svn://svn.chromium.org/blink/trunk@169835 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=248938
TBR=eseidel

Review URL: https://codereview.chromium.org/204983014

git-svn-id: svn://svn.chromium.org/blink/trunk@169834 bbb929c8-8fbe-4397-9dbb-9b2b20218538