GitLab

The resize event was added to HTMLMediaElement in http://crbug.com/328140

Add a trailing newline to some files to work around http://crbug.com/307023

Also sync with the list of properties in the latest spec revision.

BUG=330953

Review URL: https://codereview.chromium.org/136593002

git-svn-id: svn://svn.chromium.org/blink/trunk@165057 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This was used to notify the inspector only of events originating from
the FrameLoader. Instead, do this explicitly

BUG=none
R=dcarney@chromium.org

Review URL: https://codereview.chromium.org/137983002

git-svn-id: svn://svn.chromium.org/blink/trunk@165056 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=149069
TEST=no functional changes
R=kbr@chromium.org

Review URL: https://codereview.chromium.org/135893002

git-svn-id: svn://svn.chromium.org/blink/trunk@165055 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Like WebMediaStreamSource having access to the owner of the ExtraData
object simplifies the chromium code significantly.
Also taking the opportunity to clean up the ExtraData code in
MediaStreamSource after the last of the non-standard code added for
WebAudio has been refactored away.

Review URL: https://codereview.chromium.org/133733004

git-svn-id: svn://svn.chromium.org/blink/trunk@165054 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL replaces SVGAnimatedBoolean class with NewSVGProperty implementation.

For details, see design doc: https://docs.google.com/document/d/1bg7CUyUszqdwmENY3JX6_PoQD6uHRCNcRPJMlC4qlkw/edit?usp=sharing

BUG=308818

Review URL: https://codereview.chromium.org/131253002

git-svn-id: svn://svn.chromium.org/blink/trunk@165053 bbb929c8-8fbe-4397-9dbb-9b2b20218538

updateWidgetPositions() can blow away the RenderView by running script or
calling into plugins. This patch moves it from RenderView to FrameView
since having this method on RenderView which might destroy itself is not
safe. It also switches to using normal RefPtr instead of manually managing
the refcount and finally adds RefPtr to callers of updateWidgetPositions()
to avoid use-after-frees.

There's one final call inside RenderLayerScrollableArea::setScrollOffset
which is not safe but is difficult to mitigate since we're way down a
callstack by the time this call is made which can destroy the render tree
and the RenderLayerScrollableArea. This patch adds a RELEASE_ASSERT to
kill the renderer in case we get into a sitaution where this happens.
In the future we should detangle this concept entirely so such an
ASSERT isn't needed and so that the render tree can never destroy itself
from the inside.

It's not clear how to write a test for this since you need to get us to
go into the scrolling code with a dirty tree or have a plugin that does
something nefarious.

BUG=322891

Review URL: https://codereview.chromium.org/132913002

git-svn-id: svn://svn.chromium.org/blink/trunk@165052 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/102003006

git-svn-id: svn://svn.chromium.org/blink/trunk@165051 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is cleanup CL, as part of implementing partial interfaces in the
Python IDL compiler.
Primarily it moves overall [RuntimeEnabled] extended attributes (and
[Conditional] and [PerContextEnabled]) to the partial interface itself,
rather than individual members, so:

[
　　RuntimeEnabled=FeatureName,
] partial interface Foo {
　　void a();
　　void b();
};

instead of:

partial interface Foo {
　　[RuntimeEnabled=FeatureName] void a();
　　[RuntimeEnabled=FeatureName] void b();
};

This is for consistency: these extended attributes can be applied
to the partial interface itself, which means they are applied to
all members, so we can switch it on or off with one line.

Currently some IDL files put this on the partial interface, while
others put it on all members; this makes it consistent and clarifies intent.

(We could remove all of these, simplifying merging a bit, at the cost
of putting these on all the individual member lines.)

Docs:
https://sites.google.com/a/chromium.org/dev/blink/webidl/blink-idl-extended-attributes#TOC-Partial-interfaces

No changes to generated code.

R=haraken

Review URL: https://codereview.chromium.org/130943004

git-svn-id: svn://svn.chromium.org/blink/trunk@165049 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Near as I can tell updateFromElement() can no longer cause the selection to
change so lets stop using PostAttachCallbacks. This leaves only
scheduleLayerUpdate and autofocus using post attach callbacks. A future patch
will move autofocus off it and then we can make PostAttachCallbacks specific
to the layer updating system (and hopefully remove that usage someday too).

Review URL: https://codereview.chromium.org/131153002

git-svn-id: svn://svn.chromium.org/blink/trunk@165048 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Update heap/ classes to use OVERRIDE / FINAL when needed. Those are
useful for error checking or performance when the compiler supports it.

These were found using the chromium style clang plugin.

R=haraken
BUG=333316

Review URL: https://codereview.chromium.org/137803004

git-svn-id: svn://svn.chromium.org/blink/trunk@165047 bbb929c8-8fbe-4397-9dbb-9b2b20218538

After expanding an accessor in the preview, the value may not be lossless
(e.g. a long truncated string), thus assume such preview not lossless as well.

BUG=261470
R=pfeldman

Review URL: https://codereview.chromium.org/137883002

git-svn-id: svn://svn.chromium.org/blink/trunk@165046 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Update more svg classes to use OVERRIDE / FINAL when needed. Those are
useful for error checking or performance when the compiler supports it.

These were found using the chromium style clang plugin.

R=pdr, schenney@chromium.org
BUG=333316

Review URL: https://codereview.chromium.org/137063005

git-svn-id: svn://svn.chromium.org/blink/trunk@165045 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The new WebSocket implementation receives HTTP headers on
the opening handshake from the browser process.
The headers may be duplicated, e.g. there are two "Foo" headers.
In order to show duplicated headers in devtools, this CL merges them
during storing them to HTTPHeaderMap.
This CL changes the existing WebSocket behavior as well.

R=tyoshino
BUG=310405

Review URL: https://codereview.chromium.org/130863002

git-svn-id: svn://svn.chromium.org/blink/trunk@165044 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Align with the spec and Gecko for the enumerated attribute
'autocomplete' over form elements. That is, map both missing and
invalid content attribute values to the "on" keyword/string.

R=
BUG=333229

Review URL: https://codereview.chromium.org/133703002

git-svn-id: svn://svn.chromium.org/blink/trunk@165043 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Update layout test case to use runtime flag.

Depends on 131913003 & 132203007

BUG=181136, 331337

Review URL: https://codereview.chromium.org/137143002

git-svn-id: svn://svn.chromium.org/blink/trunk@165042 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Update more bindings classes to use OVERRIDE / FINAL when needed. Those are
useful for error checking or performance when the compiler supports it.

These were found using the chromium style clang plugin.

R=haraken
BUG=333316

Review URL: https://codereview.chromium.org/137603004

git-svn-id: svn://svn.chromium.org/blink/trunk@165041 bbb929c8-8fbe-4397-9dbb-9b2b20218538

As per spec change: http://html5.org/r/8343

Now, document.body is also inert. Previously events that
"skipped" an inert node were fired on document.body; now,
they are fired on document.

Test: modal-dialog-ancestor-is-inert.html
Also, other tests cover some of this behavior and have
expectations updated.

BUG=329407

Review URL: https://codereview.chromium.org/133393002

git-svn-id: svn://svn.chromium.org/blink/trunk@165040 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Typed arrays all have custom bindings, so code path that checks if an
interface is a typed array (IsTypedArray($interfaceName)) is dead code.

This removes it to avoid being misleading (and clarify rewrite);
no changes to generated code.

Obviously we want to remove these custom bindings, but the existing dead
code isn't helpful.

R=haraken

Review URL: https://codereview.chromium.org/129793007

git-svn-id: svn://svn.chromium.org/blink/trunk@165039 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This partially reverts r162663
(https://codereview.chromium.org/85763003).

I believe the real fix is https://codereview.chromium.org/103583007/, but
that solution is causing layout test failures that need
to be debugged. Reverting this should get us working in
the interim and the fix is not at all risky.

TBR=jamesr@chromium.org
BUG=303920

Review URL: https://codereview.chromium.org/135853004

git-svn-id: svn://svn.chromium.org/blink/trunk@165038 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch moves around the checks inside RenderStyle::diff so that as
much a possible the checks for a given structure of data are all grouped
and then guards those checks with a pointer compare on the struct when
possible. This was already done for some of the structs but not all of
them.

I also noticed that the logic for table display types had not been updated
for flex box so I removed the >= TABLE check and replaced it with a
LAST_TABLE_DISPLAY and FIRST_TABLE_DISPLAY guard cutting out a bunch of
if checks for flex boxes.

Similarly we were deep comparing the SVGRenderStyle because we were using
the overloaded operator== on the DataRef<SVGRenderStyle> before calling
into ::diff() to do the real comparison which is a waste. I changed to a
pointer comparison which will reduce the number of duplicate checks for
SVG elements.

While this is definitely a set of micro-optimizations it should reduce
the number of checks for a typical RenderStyle::diff which is very hot
code in all our benchmarks.

Review URL: https://codereview.chromium.org/136693009

git-svn-id: svn://svn.chromium.org/blink/trunk@165037 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Two important simplifications:
1) Remove the foul union. It isn't necessary if we don't try and put pages on
the active and free lists at the same time. Instead, we can simply tag free
pages as free and mop them up when we next walk the active pages list.
2) Remove the flags field. It's strictly a duplication of information we have
available in other fields. And we hate duplication because it will get out of
sync and cause bugs.
Also, the freed up field can likely be profitably used to avoid having to
re-introduce a union later.

BUG=332282

Review URL: https://codereview.chromium.org/136333002

git-svn-id: svn://svn.chromium.org/blink/trunk@165036 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=165024

BUG=289764
TBR=bokan@chromium.org

Review URL: https://codereview.chromium.org/137813003

git-svn-id: svn://svn.chromium.org/blink/trunk@165035 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Get rid of custom bindings for V8NodeList by generating visitDOMWrapper() and
adding an ownerNode() method to the NodeList class. This method calls
LiveNodeList::ownerNode() if the NodeList is live, and returns null otherwise.

There is no web-exposed change in behavior.

R=haraken
BUG=None

Review URL: https://codereview.chromium.org/137713002

git-svn-id: svn://svn.chromium.org/blink/trunk@165034 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Rename getType() to type() in ArrayBufferView to follow Blink coding style.

R=eseidel, tkent
BUG=None

Review URL: https://codereview.chromium.org/135573002

git-svn-id: svn://svn.chromium.org/blink/trunk@165033 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This allows garden-o-matic to handle ASAN rebaselines.

BUG=333903
R=dpranke@chromium.org

Review URL: https://codereview.chromium.org/137063002

git-svn-id: svn://svn.chromium.org/blink/trunk@165032 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This makes display:flex match display:block. Unfortunately,
our display:block behavior does not match Gecko or IE8+ though.

Review URL: https://codereview.chromium.org/134603002

git-svn-id: svn://svn.chromium.org/blink/trunk@165031 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This aligns the timing of start events under Web Animations with the legacy
implementation. Prior to this patch Web Animations would dispatch the start
event for a compositor animation before the start time was known.

This patch moves the skipping of redundant timing updates from Player to
TimedItem allowing updateInheritedTime to be aware of when a Player gains
a start time.

This opens up opportunities to further simplify updateInheritedTime, but they
have not been attempted in this patch to ease merging this fix.

BUG=333059

Review URL: https://codereview.chromium.org/133153003

git-svn-id: svn://svn.chromium.org/blink/trunk@165030 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This change broke Twitter the first time: http://crbug.com/330261

Twitter was fixed on Dec 20, 2013, so it should be safe to try again.

BUG=325922

Review URL: https://codereview.chromium.org/130193003

git-svn-id: svn://svn.chromium.org/blink/trunk@165028 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The only supported value for the <keygen/> element's keytype attribute
is the keyword for the RSA state ("rsa".) Impose the required
normalization of the IDL attribute when getting its value.

R=
BUG=321862

Review URL: https://codereview.chromium.org/123833005

git-svn-id: svn://svn.chromium.org/blink/trunk@165027 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove tests for unused features:
* sequence<SerializedScriptValue>
* [Immutable] on non-SVG attributes
* readonly constructor attributes
* static constructor attributes

Now Python passes TestTypedefs.idl!

TBR=haraken
NOTRY=true

Review URL: https://codereview.chromium.org/137583006

git-svn-id: svn://svn.chromium.org/blink/trunk@165026 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This allows completeURL and baseURL work correctly when called on a ShadowRoot.
They should fall back to the ShadowRoot's Document's baseURL unless the ShadowRoot
itself has a baseURL specified.

Since no code in Blink yet calls either on a ShadowRoot, there's no
test I can write that triggers this, but once we start calling
treeScope().completeURL() this will be well-tested.

R=esprehn@chromium.org
BUG=311682

Review URL: https://codereview.chromium.org/135643004

git-svn-id: svn://svn.chromium.org/blink/trunk@165025 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Link highlights are now generated by walking down the render tree and
accumulating quads, rather than taking one bounding quad.

The real issue was with RenderInlines; absoluteQuads would return the
quads for the line box but only as calculated from line-height and font
metrics. This meant that replaced elements like <img> would overflow out of
the highlight. With this approach we get a highlight that hopefully
matches the actual hit testing region.

BUG=289764

Review URL: https://codereview.chromium.org/129643002

git-svn-id: svn://svn.chromium.org/blink/trunk@165024 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The current prefixed event names used (e.g. webkitkeymessage, etc.) don't match
the event names in the latest EME spec, so this change updates them.  Also add
new event 'close', which wasn't specified in earlier versions of the spec.

BUG=224786
TEST=content_shell tests updated with new event names (and pass)

Review URL: https://codereview.chromium.org/124253003

git-svn-id: svn://svn.chromium.org/blink/trunk@165023 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Other case, caught while checking test cases.
(Most [RaisesException] methods are void, but not all.)

BUG=239771
R=haraken
NOTRY=true

Review URL: https://codereview.chromium.org/137593005

git-svn-id: svn://svn.chromium.org/blink/trunk@165022 bbb929c8-8fbe-4397-9dbb-9b2b20218538

RenderStyle::diff would always return StyleDifferenceLayout when adding
or removing an opacity even though layout is only required if doing so
would change if the layer is a stacking context.

See:
http://www.w3.org/TR/css3-color/#transparency
http://www.w3.org/TR/CSS21/zindex.html

BUG=334033

Review URL: https://codereview.chromium.org/136903004

git-svn-id: svn://svn.chromium.org/blink/trunk@165021 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Also includes some misc related refactors that were missed on previous CLs

BUG=326382

Review URL: https://codereview.chromium.org/127493002

git-svn-id: svn://svn.chromium.org/blink/trunk@165020 bbb929c8-8fbe-4397-9dbb-9b2b20218538

As a step to use toSVGFooElement(), this cl generate toSVGFooElement() for mentioned SVG Element.
This will help to detect bad type cast.

BUG=309516

Review URL: https://codereview.chromium.org/133203003

git-svn-id: svn://svn.chromium.org/blink/trunk@165019 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It is only used once and can be replaced by using isSVG().

Review URL: https://codereview.chromium.org/137303004

git-svn-id: svn://svn.chromium.org/blink/trunk@165018 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Some fixes to IDL test files:
* Move SVG tests to separate TestSVG.idl, as we will not need these in future
(so we can just delete this test case when SVG rewrite complete)

Remove as unused in real IDL files:
* callback interface arguments in constructors
* nullable array

Style fix to TestTypedefs:
* typedefs at *top* of file

BUG=239771
TBR=haraken
NOTRY=true

Review URL: https://codereview.chromium.org/137773002

git-svn-id: svn://svn.chromium.org/blink/trunk@165017 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The default value of the shape-inside and shape-outside properties has
changed from 'auto' to 'none'. This patch changes the CSS parsing code
and fixes up the tests.

BUG=322165

Review URL: https://codereview.chromium.org/112103012

git-svn-id: svn://svn.chromium.org/blink/trunk@165016 bbb929c8-8fbe-4397-9dbb-9b2b20218538