GitLab

TBR=esprehn@chromium.org

Review URL: https://codereview.chromium.org/185273002

git-svn-id: svn://svn.chromium.org/blink/trunk@168248 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Prefer this helper function when throwing TypeErrors.

R=
BUG=

Review URL: https://codereview.chromium.org/183253003

git-svn-id: svn://svn.chromium.org/blink/trunk@168247 bbb929c8-8fbe-4397-9dbb-9b2b20218538

indexeddb is available in Worker contexts, including this event.

R=
BUG=340522

Review URL: https://codereview.chromium.org/185323003

git-svn-id: svn://svn.chromium.org/blink/trunk@168246 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Rationale
1. GraphicsLayer deals with background-color layer as a contents layer, so a
background-color layer is in trouble in terms of the correct order with negative
z-order children.
2. CompositedLayerMapping has the big hack to decide the given layer can be
background-color layer. Unfortunately, the big hack is incomplete and invasive.
3. Hybrid Accelerated Rasterization will cover this optimization in more generic way.
4. background-color layer covers very limited cases. A layer-promoted box
object must not have children, not have decoration, only have a background-color
property.

BUG=114658, 340691
TEST=compositing/background-color/background-color-drawn-over-child.html,
compositing/*

Review URL: https://codereview.chromium.org/178013003

git-svn-id: svn://svn.chromium.org/blink/trunk@168245 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL moves some code relating to walking the RenderLayer tree and updating
GraphicsLayers out of RenderLayerCompositor. This code didn't interact with any
of the state in RenderLayerCompositor, so it's easy to untangle.

I've also taken the opportunity to remove a pretty gross hack using static
variables to communicate data across the recursive tree walk. Now we just use
normal member variables in the stack object that manages the tree walk.

This CL shouldn't change any behavior.

Review URL: https://codereview.chromium.org/185293002

git-svn-id: svn://svn.chromium.org/blink/trunk@168244 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Use SkLayerDrawLooper::Builder to construct SkLayerDrawLooper. (https://codereview.chromium.org/177473003/)

Reason for revert:
crash in SkCanvas::getTotalMatrix (crbug.com/348122)
Reverting this to see if it fixes it, otherwise we can reland.

Original issue's description:
> Use SkLayerDrawLooper::Builder to construct SkLayerDrawLooper.
> 
> The addLayer() methods of SkLayerDrawLooper will soon be removed. This CL
> updates all uses of those methods with the corresponding methods in
> SkLayerDrawLooper::Builder.
> 
> BUG=skia:2141
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=168041

TBR=schenney@chromium.org,jbroman@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=skia:2141

Review URL: https://codereview.chromium.org/178113003

git-svn-id: svn://svn.chromium.org/blink/trunk@168243 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The only reason we had a HashMap in the overlap map was to handle a recursive
treewalk in the case where we need to be composited because of our children. It
turns out the HashMap was wrong in the case where those children got composited
into another non-root layer. In that case, they would already be in the overlap
map and would be excluded in the recursive walk.

This CL removes the HashMap and instead just unites the bounds of all the
children during the normal recusive walk. If we need to composite ourself
because of our decendants, we use the united rect instead of recursively adding
all the descendant rects individually. This is both faster (3% on the
toggle_drawer compositing update) and more correct (see test).

R=enne@chromium.org

Review URL: https://codereview.chromium.org/178313004

git-svn-id: svn://svn.chromium.org/blink/trunk@168242 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Two problems with the current srcset parsing algorithm are fixed here:
1. When a set of descriptors contained only invalid descriptors, the resource wasn't ignored.
2. When more than 1 resource shared a single DPR value, the last one was taken into account, rather than the first one.

I've also added test and fixed current ones to prevent this from regressing in the future.

BUG=347998

Review URL: https://codereview.chromium.org/184523003

git-svn-id: svn://svn.chromium.org/blink/trunk@168241 bbb929c8-8fbe-4397-9dbb-9b2b20218538

CloseEvent and XMLHttpRequestProgressEvent are also exposed inside Workers.

R=
BUG=340522

Review URL: https://codereview.chromium.org/178193014

git-svn-id: svn://svn.chromium.org/blink/trunk@168240 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=pkasting@chromium.org,dpranke@chromium.org,pdr@chromium.org
BUG=309197

Review URL: https://codereview.chromium.org/183803005

git-svn-id: svn://svn.chromium.org/blink/trunk@168239 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=335247

Review URL: https://codereview.chromium.org/179873018

git-svn-id: svn://svn.chromium.org/blink/trunk@168238 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It always created like InlineBox(*this) so it always has a renderer.
Also removes NULL checks and asserts which are not longer required.

BUG=

Review URL: https://codereview.chromium.org/184023003

git-svn-id: svn://svn.chromium.org/blink/trunk@168237 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove call to ResourceLoadPriorityOptimizer from RenderBlock Destructor. This isn't necessary since it's already called from the RenderObject destructor.

BUG=

Review URL: https://codereview.chromium.org/148523010

git-svn-id: svn://svn.chromium.org/blink/trunk@168236 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Calling clippedByAncestor is not free, we should avoid calling it unless
the compositingReasons() say we need to.

R=ojan@chromium.org,vollick@chromium.org,abarth@chromium.org

Review URL: https://codereview.chromium.org/183263003

git-svn-id: svn://svn.chromium.org/blink/trunk@168234 bbb929c8-8fbe-4397-9dbb-9b2b20218538

bot more likely to succeed at rebaselining.

BUG=none
TEST=none
TBR=ojan

Review URL: https://codereview.chromium.org/185173002

git-svn-id: svn://svn.chromium.org/blink/trunk@168233 bbb929c8-8fbe-4397-9dbb-9b2b20218538

These tests should continue to run correctly without rebaselines.

BUG=345027

Review URL: https://codereview.chromium.org/183703003

git-svn-id: svn://svn.chromium.org/blink/trunk@168232 bbb929c8-8fbe-4397-9dbb-9b2b20218538

I incorrectly removed some these workaround functions for VS2013 at
https://codereview.chromium.org/27009003 . The other non-workaround
functions (that just didn't exist previously) seem to be OK, but the
behaviour around +-Inf is still different than we expect for JS and
XPath.

R=jamesr@chromium.org
BUG=347870

Review URL: https://codereview.chromium.org/184103006

git-svn-id: svn://svn.chromium.org/blink/trunk@168231 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The member was no longer big enough to store all the 
enumerators so the setter could fail. The net result was
that if you tried to set the reason to 4, it *always*
looked like this reason changed. I was seeing a full
tree build on every mouse move!

R=abarth@chromium.org
BUG=None

Review URL: https://codereview.chromium.org/177653010

git-svn-id: svn://svn.chromium.org/blink/trunk@168181 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch updates the computed position values used for blending to be a horizontal
left offset and a vertical top offset. When positions include a center, bottom, or
right position offset, it is converted to the appropriate top/left coordinate as
a calc expression. Serialized values still use the original bottom/right directions
when present, and also omit the top/left keywords where possible.

A test is added to shape-outside.html since it had no test for interpolation of extended positions.

Patch by Bear Travis.

BUG=308251

Review URL: https://codereview.chromium.org/144373002

git-svn-id: svn://svn.chromium.org/blink/trunk@168180 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Because we always compile SVG there is no need anymore to have this in a separate file. Also now we end up with only one isSystemColor.

Review URL: https://codereview.chromium.org/184693003

git-svn-id: svn://svn.chromium.org/blink/trunk@168179 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=337919


git-svn-id: svn://svn.chromium.org/blink/trunk@168178 bbb929c8-8fbe-4397-9dbb-9b2b20218538

with Windows-specific native behaviors from ScrollbarThemeWin.cpp:

* Snap the scrollbar back to its drag origin when dragged too far off the track.
  This code matches the old code we had for years and also matches my Win 7
  native scrollbars pixel-for-pixel.

* Don't jump the scroll thumb to the mouse cursor on a middle-click, only on a
  shift-click.  Middle-click must be a GTK-ism.

This is the Blink side of the changes to re-add this behavior.

BUG=337919
TEST=Drag a scroll thumb to somewhere in the track, then without releasing, move the mouse far away from the track.  The scroll thumb should snap back to the scroll origin.
R=abarth@chromium.org

Review URL: https://codereview.chromium.org/151483008

git-svn-id: svn://svn.chromium.org/blink/trunk@168177 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=249427
TBR=

Review URL: https://codereview.chromium.org/184113008

git-svn-id: svn://svn.chromium.org/blink/trunk@168176 bbb929c8-8fbe-4397-9dbb-9b2b20218538

targeted style recalc.

This simplifies the code a bit and unifies the code paths.

BUG=335247
R=esprehn@chromium.org, ojan@chromium.org

Review URL: https://codereview.chromium.org/183583002

git-svn-id: svn://svn.chromium.org/blink/trunk@168175 bbb929c8-8fbe-4397-9dbb-9b2b20218538

- standard font family should go Android path on Android;
- for non-generic font families, familyNameFromSettings should return
  an empty AtomicString.

BUG=347121
TEST=platform/android/fast/text/international/non-generic-font-fallback.html
R=dglazkov@chromium.org

Review URL: https://codereview.chromium.org/184703003

git-svn-id: svn://svn.chromium.org/blink/trunk@168173 bbb929c8-8fbe-4397-9dbb-9b2b20218538

SpeechSynthesis.cpp incorrectly assumed that calling
m_platformSpeechSynthesizer->cancel() would immediately call
didFinishSpeaking or speakingErrorOccurred, which would null out
m_currentSpeechUtterance. This assumption was true in WebKit/Mac, but
Chromium's platform implementation is asynchronous, so that call may
come later.

Fix the issue and simplify the logic by getting rid of the raw pointer
to the current utterance altogether. Now the RefPtr at the front of the
utterance queue is the current utterance, and the platform implementation
is allowed to fire events on utterances that are no longer in the queue.

BUG=344881
R=abarth@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=168092

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=168169

Review URL: https://codereview.chromium.org/180553004

git-svn-id: svn://svn.chromium.org/blink/trunk@168171 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Wrong merge base when committing!

> Fix use-after-free of m_currentSpeechUtterance.
> 
> SpeechSynthesis.cpp incorrectly assumed that calling
> m_platformSpeechSynthesizer->cancel() would immediately call
> didFinishSpeaking or speakingErrorOccurred, which would null out
> m_currentSpeechUtterance. This assumption was true in WebKit/Mac, but
> Chromium's platform implementation is asynchronous, so that call may
> come later.
> 
> Fix the issue and simplify the logic by getting rid of the raw pointer
> to the current utterance altogether. Now the RefPtr at the front of the
> utterance queue is the current utterance, and the platform implementation
> is allowed to fire events on utterances that are no longer in the queue.
> 
> BUG=344881
> R=abarth@chromium.org
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=168092
> 
> Review URL: https://codereview.chromium.org/180553004

TBR=dmazzoni@google.com

Review URL: https://codereview.chromium.org/185093002

git-svn-id: svn://svn.chromium.org/blink/trunk@168170 bbb929c8-8fbe-4397-9dbb-9b2b20218538

SpeechSynthesis.cpp incorrectly assumed that calling
m_platformSpeechSynthesizer->cancel() would immediately call
didFinishSpeaking or speakingErrorOccurred, which would null out
m_currentSpeechUtterance. This assumption was true in WebKit/Mac, but
Chromium's platform implementation is asynchronous, so that call may
come later.

Fix the issue and simplify the logic by getting rid of the raw pointer
to the current utterance altogether. Now the RefPtr at the front of the
utterance queue is the current utterance, and the platform implementation
is allowed to fire events on utterances that are no longer in the queue.

BUG=344881
R=abarth@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=168092

Review URL: https://codereview.chromium.org/180553004

git-svn-id: svn://svn.chromium.org/blink/trunk@168169 bbb929c8-8fbe-4397-9dbb-9b2b20218538

At the time that landed, these were already marked as needing rebaselines for a
different bug, and the rebaselining for that bug crossed in-flight, leaving
these with incorrect baselines.

BUG=345609
TEST=none
TBR=tkent

Review URL: https://codereview.chromium.org/184043012

git-svn-id: svn://svn.chromium.org/blink/trunk@168168 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=248938
TEST=none
TBR=apavlov

Review URL: https://codereview.chromium.org/185053003

git-svn-id: svn://svn.chromium.org/blink/trunk@168167 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Caused various iframe accessibility tests to fail across platforms.

> Add an accessibility role for the iframe element.
> 
> We can add a layout test after adding TestRunner support on the Chromium side,
> or we could just test it in Chromium.
> 
> BUG=347464
> 
> Review URL: https://codereview.chromium.org/183183005

TBR=dmazzoni@chromium.org

Review URL: https://codereview.chromium.org/185003006

git-svn-id: svn://svn.chromium.org/blink/trunk@168166 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=304341

Review URL: https://codereview.chromium.org/183693004

git-svn-id: svn://svn.chromium.org/blink/trunk@168165 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is the second part of a three-sided MediaSource 
Chromium/Blink+roll Blink/Chromium set of changes to enable 
JavaScript web apps to continue to round-trip 
SourceBuffer.timestampOffset set/get if SourceBuffer append 
operations do not update the timestamp offset, and to 
retrieve any updated timestamp offset as can occur during 
upcoming "sequence" append mode compliant coded frame 
processing. 

R=acolwell@chromium.org,jamesr@chromium.org
BUG=249422
TEST=No local linux regressions of http/tests/media/media-source/ layout tests

Review URL: https://codereview.chromium.org/178763006

git-svn-id: svn://svn.chromium.org/blink/trunk@168163 bbb929c8-8fbe-4397-9dbb-9b2b20218538

A rarely used part of the IndexedDB spec was overlooked - like most
other events, script can call the constructor with a dict to populate
the fields.

This required some finessing: the fields were stored as 'IDBAny'
but our automagical code gen wiring for EventInit prefers ScriptValue
or SerializedScriptValue. I went with the latter, and as an aside
eliminate a call to v8::Isolate::GetCurrent() by SSV - serializing
simple values such as null or numbers can be done with no isolate.

BUG=345881
R=dgrogan,haraken

Review URL: https://codereview.chromium.org/177633006

git-svn-id: svn://svn.chromium.org/blink/trunk@168162 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Fix regression introduced in
https://codereview.chromium.org/54173005. That CL prevented the user
from calling start() before setting a buffer source for the
AudioBufferSourceNode.  Previously, this was allowed and when the
buffer was set it would get played. This CL brings back that behavior.

The WebAudio spec is silent on this issue, but this was raised on the
W3C audio list
(http://lists.w3.org/Archives/Public/public-audio/2014JanMar/0059.html)
and there was no apparent opposition to keeping the old behavior.

This also allows http://chromium.googlecode.com/svn/trunk/samples/audio/wavetable-synth.html to work again since it calls start with null buffer which is set later.

BUG=334237

Review URL: https://codereview.chromium.org/161553002

git-svn-id: svn://svn.chromium.org/blink/trunk@168161 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The root cause is CompositeEditCommand::moveParagraphWithClones() passing
two positions |start| and |end| which do not follow the document order,
i.e. in some situations |start| is located after |end| because of
the difference in affinity.

This patch fixes this crash by normalizing |end| to |start| in such situations.
It also adds an ASSERT that checks the relationship between |start| and |end|.

BUG=345005

Review URL: https://codereview.chromium.org/182533004

git-svn-id: svn://svn.chromium.org/blink/trunk@168160 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=168130

BUG=345609
TBR=tkent@chromium.org

Review URL: https://codereview.chromium.org/184553007

git-svn-id: svn://svn.chromium.org/blink/trunk@168159 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Micro cleanup of awkward nested 'if statement' in FrameView::layout().

BUG=

Review URL: https://codereview.chromium.org/177003021

git-svn-id: svn://svn.chromium.org/blink/trunk@168158 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Important since pointers been replaced with references.

BUG=

Review URL: https://codereview.chromium.org/184673002

git-svn-id: svn://svn.chromium.org/blink/trunk@168157 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We can add a layout test after adding TestRunner support on the Chromium side,
or we could just test it in Chromium.

BUG=347464

Review URL: https://codereview.chromium.org/183183005

git-svn-id: svn://svn.chromium.org/blink/trunk@168156 bbb929c8-8fbe-4397-9dbb-9b2b20218538