GitLab

Without this CL, a SQLResultSet object in SQLStatementBackend can be constructed
in a database thread, moved to main/worker thread, and referred by JavaScript
object in the main/worker thread.

We can't move object ownership over threads in Oilpan. So, SQLStatementBackend
should construct a SQLResultSet object in the SQLStatementBackend constructor,
which runs in main/worker thread. SQLResultSet should have a validity flag to
represent nullness in the code without this CL.

Note:
* Change the return type of AbstractSQLStatementBackend::sqlResultSet from
  PassRefPtr<SQLResultSet> to SQLResultSet* because the function doesn't release
  the ownership of the SQLResultSet object.

* Change the order of data members of SQLResultSet to pack better.

BUG=347902

Review URL: https://codereview.chromium.org/207453006

git-svn-id: svn://svn.chromium.org/blink/trunk@169929 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/206313004

git-svn-id: svn://svn.chromium.org/blink/trunk@169928 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Traverse the selectorList() for :host pseudo, extract features, and create
class invalidation sets like we do for features and classes on the top level.

R=esprehn,ojan
BUG=335247,339729

Review URL: https://codereview.chromium.org/209333005

git-svn-id: svn://svn.chromium.org/blink/trunk@169927 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=353921

Review URL: https://codereview.chromium.org/206073005

git-svn-id: svn://svn.chromium.org/blink/trunk@169926 bbb929c8-8fbe-4397-9dbb-9b2b20218538

RESOLUTION: to change draw SystemFocusRing to drawFocusIfNeeded.

public discussion:
  http://www.w3.org/Bugs/Public/show_bug.cgi?id=23980
  http://www.w3.org/WAI/PF/HTML/wiki/Summary_of_2014-01-13_Meeting
  http://www.w3.org/2014/01/13-html-a11y-minutes.html#item02

other browsers:
  It has already been agreed or changed.
  http://bugzilla.mozilla.org/show_bug.cgi?id=959820 (Firefox)

BUG=354960

Review URL: https://codereview.chromium.org/208003003

git-svn-id: svn://svn.chromium.org/blink/trunk@169925 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Intent to Deprecate:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/nNsZDn_utUc/dUtD3ulsFMkJ

BUG=354921

Review URL: https://codereview.chromium.org/208263005

git-svn-id: svn://svn.chromium.org/blink/trunk@169924 bbb929c8-8fbe-4397-9dbb-9b2b20218538

EventPath::parent() is only used in adjusting the result of HitTest in some places.
For example, if a text node is clicked, we have to adjust the result to the *reasonable* parent element of the text node.

In adjusting, we should use NodeRenderingTraversal::parent() instead of EventPath::parent().
That means the adjusted result is no longer shadow roots or insertion points.
The result would be the parent element of the text node in the composed tree.
That should make sense because neither shadow roots nor insertion points are rendered.

BUG=354366

Review URL: https://codereview.chromium.org/210753002

git-svn-id: svn://svn.chromium.org/blink/trunk@169923 bbb929c8-8fbe-4397-9dbb-9b2b20218538

RawPtr is a transition type that will be removed and replace with a
normal pointer. Normal pointers are not default initialized and therefore
RawPtr shouldn't be either. Instead, we introduce a zap value in debug
mode to make it easier to find uninitialized RawPtrs during the transition.

R=erik.corry@gmail.com, oilpan-reviews@chromium.org, tkent@chromium.org, zerny@chromium.org

Review URL: https://codereview.chromium.org/209433004

git-svn-id: svn://svn.chromium.org/blink/trunk@169922 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Bump up the limit from 10 StepIns to 20.
Add a test on framework code.

BUG=267592
R=pfeldman, yurys

Review URL: https://codereview.chromium.org/210243002

git-svn-id: svn://svn.chromium.org/blink/trunk@169921 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We can use add() to avoid doing two hash lookups in the common case where
there's no callback queue yet (ex. setting a new attribute) and we can
also return a reference.

Review URL: https://codereview.chromium.org/206623006

git-svn-id: svn://svn.chromium.org/blink/trunk@169920 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL moves Interpolation's hierarchy and InterpolationRecord to oilpan's heap. I'll move other Interpolation objects in a follow-up CL.

BUG=341032
NOTRY=true

Review URL: https://codereview.chromium.org/210783002

git-svn-id: svn://svn.chromium.org/blink/trunk@169919 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is a prerequisite for deprecating the prefixed function, because
otherwise the forms tests get a deprecation message. This internal use
of webkitRequestAnimationFrame may also have inflated the use counter
data for it.

BUG=354921
R=abarth@chromium.org, keishi@chromium.org, tkent@chromium.org

Review URL: https://codereview.chromium.org/209003012

git-svn-id: svn://svn.chromium.org/blink/trunk@169918 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/209683002

git-svn-id: svn://svn.chromium.org/blink/trunk@169917 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Promises are now implemented tested in V8.
I delete Promises tests except for tests testing ScriptPromise or
ScriptPromiseResolver.

BUG=352597

Review URL: https://codereview.chromium.org/209443009

git-svn-id: svn://svn.chromium.org/blink/trunk@169916 bbb929c8-8fbe-4397-9dbb-9b2b20218538

NOTRY=true

Review URL: https://codereview.chromium.org/209413007

git-svn-id: svn://svn.chromium.org/blink/trunk@169915 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is as per the the coding guidelines for blink, which disallows
importing names from the STL library through the "using" declarations.

Review URL: https://codereview.chromium.org/199793011

git-svn-id: svn://svn.chromium.org/blink/trunk@169914 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The function draws a focus ring around the current default path or the given path if the given element is focused.
Currently, the focus ring is drawn incorrectly because it is larger than the given path region.
To solve this problem, we should add focus ring width to dirty rect.

BUG=352976

Review URL: https://codereview.chromium.org/200283003

git-svn-id: svn://svn.chromium.org/blink/trunk@169913 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Runtime ShadowDOM flag was already set to true at r169201. This change shouldn't have any effects.

CSSSelector still uses this flag. I'll clean up CSSSelector in another patch.

BUG=336121

Review URL: https://codereview.chromium.org/206033004

git-svn-id: svn://svn.chromium.org/blink/trunk@169912 bbb929c8-8fbe-4397-9dbb-9b2b20218538

If you delete the documentElement then that also means all <style> and
<link> elements must have been removed from the document so the
StyleResolver should already be reset or cleared and we don't need to
call clearStyleResolver() from inside Document::childrenChanged.

Review URL: https://codereview.chromium.org/209383015

git-svn-id: svn://svn.chromium.org/blink/trunk@169911 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This removes 6 FIXMEs, and removes duplicate calls of info.GetIsolate().

This updates the license block to the new form (since this is *generated* code),
and makes it automatically use the correct module name.
I didn't make the year update automatically, as otherwise r-b-t will break
every new year and this detail doesn't seem worth the hassle.

I also moved one long comment into a template comment, so it's not in the
generated code. This seems clearer (it's in the template), and reduces
duplication and noise.

Overall these reduce the size of the generated code and tests
by about 6% (generated code from 17 MB from 16 MB)!
Not a big deal, but makes them a bit faster to compile and easier to read.

R=haraken
BUG=345503

Review URL: https://codereview.chromium.org/210663003

git-svn-id: svn://svn.chromium.org/blink/trunk@169910 bbb929c8-8fbe-4397-9dbb-9b2b20218538

To conform the latest W3C spec, add sysexEnabled attribute.
http://www.w3.org/TR/webmidi/#idl-def-MIDIAccess

MIDIAccess originally has a method sysExEnabled().
To avoid having two functions for the same operation, this change
renames it to sysexEnabled(). This change affects other caller
implementations.

To rename everything in blink, it modifies blink API.
So this change is also the first step of blink API change.

Here are a series of changes.
1. this change
2. https://codereview.chromium.org/210003002/
3. https://codereview.chromium.org/208423016/

BUG=349538
TEST=LayoutTests/webmidi

Review URL: https://codereview.chromium.org/208243014

git-svn-id: svn://svn.chromium.org/blink/trunk@169909 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL removes a histgram from GraphicsLayerUpdater::rebuildTree. This
histogram accounts for 9.2% of the time spent in
GraphicsLayerUpdater::rebuildTree and 1.6% of the total compositing update in
Polymer's calculator.

R=ojan@chromium.org

Review URL: https://codereview.chromium.org/210093005

git-svn-id: svn://svn.chromium.org/blink/trunk@169908 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Perl cruft: Python compiler assumes a header file exists.
This removes a special case and simplifies CG logic by having stub header
files if needed.

R=haraken
BUG=341748

Review URL: https://codereview.chromium.org/208393014

git-svn-id: svn://svn.chromium.org/blink/trunk@169906 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is another attempt to investigate crashes in the wild.
This change deploys more RELEASE_ASSERT() to see if/where the
wrongly-killed Resource objects appear.

BUG=352043
TEST=none
R=eseidel, japhet@chromium.org, kouhei

Review URL: https://codereview.chromium.org/210293005

git-svn-id: svn://svn.chromium.org/blink/trunk@169905 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/208943005

git-svn-id: svn://svn.chromium.org/blink/trunk@169903 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Delete WebSocketHandle::connect(..., const WebString&, ...), which is replaced by
WebSocketHandle::connect(..., const WebSerializedOrigin&, ...).

BUG=339373

Review URL: https://codereview.chromium.org/206143003

git-svn-id: svn://svn.chromium.org/blink/trunk@169902 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Replace hide() in ExternalPopupMenu:show() with
m_webExternalPopupMenu->close(). It means we remove
PopupMenuClient::popupDidHide() call.

We should not call it because the client changes its state to "no popup" state
by the popupDidHide call, and it won't call PopupMenu::hide() even if it's
necessary.

BUG=30550,145658

Review URL: https://codereview.chromium.org/208243012

git-svn-id: svn://svn.chromium.org/blink/trunk@169901 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Linear easing was being supplied instead of the timing function specified by the user.

Example for testing: http://jsbin.com/sihif/1/quiet

Cubic bezier keyframe easings were being calculated without knowledge of the iteration duration. This led to inconsistency in the accuracy being requested.

We now always use the iteration duration when choosing the accuracy for timing function evaluation.

Without this, the transitions layout tests opacity-transform-transitions-inside-iframe.html and
cubic-bezier-overflow-length.html would have needed rebaselining.

BUG=352919

Review URL: https://codereview.chromium.org/196413030

git-svn-id: svn://svn.chromium.org/blink/trunk@169897 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/205803002

git-svn-id: svn://svn.chromium.org/blink/trunk@169896 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=335345
R=eseidel@chromium.org

Review URL: https://codereview.chromium.org/206283004

git-svn-id: svn://svn.chromium.org/blink/trunk@169895 bbb929c8-8fbe-4397-9dbb-9b2b20218538

* Splits up large tests into smaller files.
* Extracts common key data to its own file
* Reduces use of shouldRejectPromiseWithNull()

BUG=245025

Review URL: https://codereview.chromium.org/206483010

git-svn-id: svn://svn.chromium.org/blink/trunk@169894 bbb929c8-8fbe-4397-9dbb-9b2b20218538

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

Review URL: https://codereview.chromium.org/196243010

git-svn-id: svn://svn.chromium.org/blink/trunk@169893 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Attempt 2.

Notry because we don't have an XP trybot.
NOTRY=TRUE
BUG=354849

Review URL: https://codereview.chromium.org/196173017

git-svn-id: svn://svn.chromium.org/blink/trunk@169892 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The last consumer of PostAttachCallbacks is layer updating, this patch
switches away from the generic "call me after recalc style" system that
was PostAttachCallbacks to a very specialized system that only marks
elements as needing a style recalc for the purposes of creating and
destroying layers.

Review URL: https://codereview.chromium.org/143983007

git-svn-id: svn://svn.chromium.org/blink/trunk@169891 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http/tests/serviceworker/registration.html
* was waiting on CR side to land

virtual/serviceworker/http/tests/serviceworker/registration.html
* was crashy but seems okay now (?)

virtual/serviceworker/http/tests/serviceworker/indexeddb.html
virtual/serviceworker/http/tests/serviceworker/postmessage.html
* were waiting on CR side to land

BUG=350103,352703,285976
TBR=alecflett

Review URL: https://codereview.chromium.org/210473003

git-svn-id: svn://svn.chromium.org/blink/trunk@169890 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When the root renderer is fully collapsed with a background
image, the solid color background painting path would assume
that the background image would fill the viewport, but the
background image code would, correctly, only use the actual
height of the root and therefor not fill anything. The
result is the viewport isn't filled at all.

This is a regression from my early patch (r166582) that avoids
implicitly painting the viewport background when the root
itself will do so as an optimization to avoid rasterizing the entire
viewport twice. Catching the case in the solid color background
painting path. Behavior matches FireFox and Safari.

BUG=349936

Review URL: https://codereview.chromium.org/208393004

git-svn-id: svn://svn.chromium.org/blink/trunk@169889 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove the SkiaFontWin.cpp and SkiaFontWin.h files and all references
to them as the paintSkiaText method provided by SkiaFontWin.h is no
longer used now that the Uniscribe code path has been removed.

R=schenney@chromium.org
BUG=313423

Review URL: https://codereview.chromium.org/209383011

git-svn-id: svn://svn.chromium.org/blink/trunk@169888 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This attribute is already present on Element.idl. This change avoids extra bindings code
being written.

Review URL: https://codereview.chromium.org/204983019

git-svn-id: svn://svn.chromium.org/blink/trunk@169887 bbb929c8-8fbe-4397-9dbb-9b2b20218538

With Repaint-After-Layout we would repaint the SVG Root element a lot more
then needed. This ports the check for when to repaint the SVG Root element
into the LayoutRectRecorder to better match what the LayoutRepainting was doing
originally.

BUG=320139

Review URL: https://codereview.chromium.org/208103002

git-svn-id: svn://svn.chromium.org/blink/trunk@169886 bbb929c8-8fbe-4397-9dbb-9b2b20218538

git-svn-id: svn://svn.chromium.org/blink/trunk@169880 bbb929c8-8fbe-4397-9dbb-9b2b20218538