GitLab

This is a follow-up fix for https://codereview.chromium.org/462423003/.
It's inconsistent that workerGlobalScopeDestroyed is called while
workerGlobalScopeStarted is not called. In fact, workerGlobalScopeDestroyed
is a signal to let the WorkerMessagingProxy know that the worker thread
is terminated, it should be renamed to workerThreadTerminated.

BUG=403179

Review URL: https://codereview.chromium.org/468883002

git-svn-id: svn://svn.chromium.org/blink/trunk@180254 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This allows for a better code structure and code reuse by alternative
devtools frontends.

R=pfeldman@chromium.org, pfeldman, vsevik

Review URL: https://codereview.chromium.org/471433004

git-svn-id: svn://svn.chromium.org/blink/trunk@180253 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=pfeldman@chromium.org, pfeldman, vsevik
BUG=403087

Review URL: https://codereview.chromium.org/475713002

git-svn-id: svn://svn.chromium.org/blink/trunk@180252 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Before this patch, register() always returns new registration object
even if the same registration already exists. This patch makes
register() return the existing object if it's available.

Chromium-side change: https://codereview.chromium.org/470843002/

NOTE: ServiceWorkerRegistration is still disabled and this change does
not affect the existing code/behavior.


BUG=396400
TEST=https://codereview.chromium.org/466723002/
NOTRY=true

Review URL: https://codereview.chromium.org/473743002

git-svn-id: svn://svn.chromium.org/blink/trunk@180251 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=apavlov
NOTRY=true

Review URL: https://codereview.chromium.org/474783003

git-svn-id: svn://svn.chromium.org/blink/trunk@180250 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Use |element| for the variable (italic in spec) and add missing word:
https://www.w3.org/Bugs/Public/show_bug.cgi?id=26580

NOTRY=true

Review URL: https://codereview.chromium.org/473803002

git-svn-id: svn://svn.chromium.org/blink/trunk@180249 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=180241

BUG=350404
TBR=h.joshi@samsung.com

Review URL: https://codereview.chromium.org/476633003

git-svn-id: svn://svn.chromium.org/blink/trunk@180248 bbb929c8-8fbe-4397-9dbb-9b2b20218538

 - Remove blink:: prefix
 - Remove *using namespace* usage

BUG=

Review URL: https://codereview.chromium.org/471503002

git-svn-id: svn://svn.chromium.org/blink/trunk@180247 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=403430
R=eustas@chromium.org, lushnikov@chromium.org

Review URL: https://codereview.chromium.org/468043004

git-svn-id: svn://svn.chromium.org/blink/trunk@180245 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/464273002

git-svn-id: svn://svn.chromium.org/blink/trunk@180244 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This exemption allowed the fullscreen element stack to get into a state
not otherwise possible, namely with elements in any order as opposed to
in tree order. It also seemed possible to enter fullscreen in a sibling
iframe, leading to a (cross-document) set of fullscreen elements forming
a tree instead of a simple chain, a case without any test coverage.

Removing the exemption means that a webkitfullscreenerror event will be
fired if one is already in fullscreen and requests fullscreen for an
element other than a descendant of the current fullscreen element.

The use counter was intended to tell whether any content depends on this
or not, but data will not be available for a long time yet. In order to
be affected one needs to, in fullscreen, have a currently visible button
request fullscreen for an element currently not visible, which seems
far-fetched, so site compat is probably not a problem.

The use counter for webkitRequestFullScreen() provides an upper bound:
http://www.chromestatus.com/metrics/feature/timeline/popularity/177

BUG=383813

Review URL: https://codereview.chromium.org/458223003

git-svn-id: svn://svn.chromium.org/blink/trunk@180243 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=haraken
BUG=
NOTRY=true

Review URL: https://codereview.chromium.org/471823002

git-svn-id: svn://svn.chromium.org/blink/trunk@180242 bbb929c8-8fbe-4397-9dbb-9b2b20218538

ThirdWidth, QuarterWidth and will take complex path
while rendering.

BUG=350404

Review URL: https://codereview.chromium.org/263363005

git-svn-id: svn://svn.chromium.org/blink/trunk@180241 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=180228

BUG=351981
TBR=yosin@chromium.org

Review URL: https://codereview.chromium.org/477563002

git-svn-id: svn://svn.chromium.org/blink/trunk@180240 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=398742

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180013

Review URL: https://codereview.chromium.org/451603002

git-svn-id: svn://svn.chromium.org/blink/trunk@180239 bbb929c8-8fbe-4397-9dbb-9b2b20218538

SVGElementRareData is including more than it needs to, this patch reduces it.

Review URL: https://codereview.chromium.org/468553004

git-svn-id: svn://svn.chromium.org/blink/trunk@180238 bbb929c8-8fbe-4397-9dbb-9b2b20218538

At the moment alerts are being stored both at
http://sheriff-o-matic.appspot.com/alerts and
http://auto-sheriff.appspot.com/data . auto-sheriff will be turned
down, so stop reading alerts from there.

BUG=401358
NOTRY=true

Review URL: https://codereview.chromium.org/476593002

git-svn-id: svn://svn.chromium.org/blink/trunk@180237 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is for consistency with ScriptForbiddenScope,
which is often used with EventDispatchForbiddenScope.

Review URL: https://codereview.chromium.org/446063007

git-svn-id: svn://svn.chromium.org/blink/trunk@180236 bbb929c8-8fbe-4397-9dbb-9b2b20218538

There's not really a useful difference between unittests and integration
tests in webkitpy these days, so this removes the integrationtest concept
and renames the files for consistency w/ everything else.

This is being done in preparation for replacing test-webkitpy with the
generic python test runner (by removing nonstandard things from test-webkitpy).

This also removes the custom TestLoader I wrote since we don't need it any more.

R=eseidel@chromium.org
BUG=402172

Review URL: https://codereview.chromium.org/454223003

git-svn-id: svn://svn.chromium.org/blink/trunk@180235 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch removes unnecessary blink:: prefixes.
It removes unnecesssary *using* keyword lines.

BUG=

Review URL: https://codereview.chromium.org/462353003

git-svn-id: svn://svn.chromium.org/blink/trunk@180234 bbb929c8-8fbe-4397-9dbb-9b2b20218538

AudioNode::breakConnectionWithLock() doesn't need to check 'm_normalRefCount > 1' in non-oilpan builds

This check was introduced in https://codereview.chromium.org/386403002/.
I agree that it's OK to have the check, but the check is not needed.
If we have the check, we can avoid the overhead of calling disableOutputsIfNecessary()
when the AudioNode is about to die. However, having the check causes an issue
when rewriting the reference counting system (see full CL: https://codereview.chromium.org/438293003/),
I remove the check in this CL in preparation for the rewrite.

BUG=340522

Review URL: https://codereview.chromium.org/466763002

git-svn-id: svn://svn.chromium.org/blink/trunk@180233 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently, in oilpan builds, removeAutomaticPullNode() is called in AudioNode::dispose().
In non-oilpan builds, removeAutomaticPullNode() is called in AudioContext::markForDeletion().
We don't want to have different call paths between oilpan builds and non-oilpan builds.

Given that removeAutomaticPullNode() should be called when the AudioNode is destructed
(but before the associated AudioContext is gone), removeAutomaticPullNode() should always
be called in AudioNode::dispose().

BUG=340522

Review URL: https://codereview.chromium.org/468473002

git-svn-id: svn://svn.chromium.org/blink/trunk@180232 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The minutes field was updated to be zero-padded, but the test wasn't, so the
test would fail at certain times of day. This updates that test to always pass.

NOTRY=true

Review URL: https://codereview.chromium.org/473433002

git-svn-id: svn://svn.chromium.org/blink/trunk@180231 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Their absence tripped stronger asserts in multivm's refactoring of ScriptWrappable.

Review URL: https://codereview.chromium.org/469923002

git-svn-id: svn://svn.chromium.org/blink/trunk@180230 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Before this patch, Blink clears selection when focus is changed. This patch
changes to clear selection for text field only. Thus, we keep selection for
content editable.

This patch also updates below tests to follow new behavior introduced by this
patch.

- fast/dom/blur-contenteditable.html
 Render tree dump has a caret.
- fast/layers/scroll-rect-to-visible.html
 Render tree dump has a caret.
- fast/events/selectionchange-user-initiated.html
 One "selectionchange" event from clear selection for content editable has
 been gone.
- /fast/events/selectstart-prevent-selectall.html
 same as above.
- fast/forms/focus-selection-input.html
 One "selectionchange" event from clear selection for caret on LABEL element
 by test step 5 been gone.
- fast/forms/focus-selection-textarea.html
 same as above

BUG=351981
TEST=LayoutTests/editing/selection/keep-selection-after-set-focus.html

Review URL: https://codereview.chromium.org/424973003

git-svn-id: svn://svn.chromium.org/blink/trunk@180228 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Move matching Element traversal functions from LiveNodeListBase to
ElementTraversal and generalized them a bit so that they can be reused in the
rest of the code base.

LiveNodeListBase exists so that code can be shared between LiveNodeList and
HTMLCollection. However, the ElementTraversal class is a better place to share
DOM Tree traversal utility functions. It is quite common in the code base that
we want to traverse Elements that match a specific rule so this will reduce
code duplication once the rest of the code is ported to these new
ElementTraversal functions.

R=adamk@chromium.org, esprehn@chromium.org

Review URL: https://codereview.chromium.org/468183002

git-svn-id: svn://svn.chromium.org/blink/trunk@180227 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In a recent change, I broke the logic for determining whether fixed position
elements should trigger main thread scrolling. This CL fixes the logic to
trigger main thread scrolling if we have fixed position elements that lack
composited layers.

R=vollick@chromium.org
BUG=402359

Review URL: https://codereview.chromium.org/465333004

git-svn-id: svn://svn.chromium.org/blink/trunk@180226 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We were correctly squashing fixed position layers together, but we weren't
setting the correct position constraint on the squashing layer. Instead, we
were setting the constraint on the main graphics layer directly, which meant
the compositor incorrectly scrolled the squashed fixed position layer instead
of keeping it stuck to the viewport.

R=vollick@chromium.org
BUG=402819

Review URL: https://codereview.chromium.org/475553002

git-svn-id: svn://svn.chromium.org/blink/trunk@180225 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/468443002

git-svn-id: svn://svn.chromium.org/blink/trunk@180224 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL causes us to bail out of the fast scrolling path if we have slow
repainting objects (i.e., background-attachment: fixed). We've done this for a
long time, by my recent CL removed this behavior and we appear to lack test
coverage for it.

R=vollick@chromium.org
BUG=402747

Review URL: https://codereview.chromium.org/474683002

git-svn-id: svn://svn.chromium.org/blink/trunk@180223 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When there were 2000 failures recently the resulting alerts JSON was
~2.5MB. This is too big to fit in memcache's 1MB limit [1].

However the data compresses readily (sizes in K; level 0 is the
uncompressed size):

>>> import zlib
>>> uncompressed = open('data.json', 'r').read()
>>> for level in range(0,10):
...   compressed = zlib.compress(uncompressed, level)
...   print level, len(compressed) / 1024.0
...
0 2525.62207031
1 129.721679688
2 121.662109375
3 120.194335938
4 117.762695312
5 111.912109375
6 105.26953125
7 104.64453125
8 101.791015625
9 101.745117188

[1] https://developers.google.com/appengine/docs/python/memcache/#Python_Limits

TEST=PYTHONPATH=/usr/local/google_appengine Tools/GardeningServer/tests.py
BUG=401358
R=eseidel@chromium.org, ojan@chromium.org

Review URL: https://codereview.chromium.org/469983002

git-svn-id: svn://svn.chromium.org/blink/trunk@180222 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of LeakDetector: Add an expectation for resolve-after-window-close.html (patchset #1 of https://codereview.chromium.org/460213002/)

Reason for revert:
Fixed by http://src.chromium.org/viewvc/blink?view=revision&revision=180139

Original issue's description:
> LeakDetector: Add an expectation for resolve-after-window-close.html
> 
> r179771 fixed the crash bug but it didn't fix the leak bug.
> 
> BUG=400012
> TEST=n/a
> NOTRY=true
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180028

TBR=hajimehoshi@chromium.org,kouhei@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=400012

Review URL: https://codereview.chromium.org/475593003

git-svn-id: svn://svn.chromium.org/blink/trunk@180221 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Cannot print PDFs

In r180112, I removed some seemly dead code, but that code actually existed to
make some functions on PrintContext virtual. Instead of hacking in virtuals via
a subclass in the web layer, we can just make the functions virtual in the base
class.

I've also added OVERRIDE annotations to catch this sort of error at compile
time. These annotations required me to fix the return type of one of these
functions to match. Finally, I've added a test, which appears to be the first
unit test to exercise this codepath.

BUG=403467

Review URL: https://codereview.chromium.org/476463002

git-svn-id: svn://svn.chromium.org/blink/trunk@180220 bbb929c8-8fbe-4397-9dbb-9b2b20218538

- Remove unnecessary blink:: prefixes
- Make blank lines consistent

Review URL: https://codereview.chromium.org/463123003

git-svn-id: svn://svn.chromium.org/blink/trunk@180219 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Cleanup namespace usage in Source/core/platform/[A-Z]*.h

- Merge multiple |namespace blink {}| block
- Remove unnecessary blink:: prefixes
- Make blank lines consistent

BUG=None

Review URL: https://codereview.chromium.org/467193002

git-svn-id: svn://svn.chromium.org/blink/trunk@180218 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Previously test-webkitpy had a way to annotate tests that could not
be run in parallel with other tests. This was useful in rare occasions,
but such tests usually are flaky and slow, so they were gradually removed
and/or reworked. Ultimately this feature wasn't really being used and
was just a source of needless complexity, so I'm killing it.

R=eseidel@chromium.org
BUG=402172

Review URL: https://codereview.chromium.org/458643003

git-svn-id: svn://svn.chromium.org/blink/trunk@180217 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch adds the flag 'inherited' to CSSProperties.in, replacing the
existing switch statement in CSSProperty.cpp. The primary use for
isInheritedProperty appears to be as part of the MatchedPropertiesCache
optimisation.

I've added the inherited flag to overflow-wrap and paint-order, which
are inherited in the code and specs[1][2] but were not returning true
in the switch. As far as I can tell, this has no difference aside from
making us still use the MatchedPropertiesCache if these are explicitly
set to inherit. I also added a fixme indicating that the property
resize shouldn't inherit by default.

[1] http://dev.w3.org/csswg/css-text/#overflow-wrap-property
[2] https://svgwg.org/svg2-draft/painting.html#PaintOrder

BUG=396992

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180051

Review URL: https://codereview.chromium.org/450103005

git-svn-id: svn://svn.chromium.org/blink/trunk@180216 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The patch makes use of an intermediate render buffer at an alternate
resolution for rasterizing scaled canvases in a spec compliant way.

BUG=392600
TEST=layout tests with zoomed canvases (there's lots) with --enabled display-list-2d-canvas

Review URL: https://codereview.chromium.org/467353002

git-svn-id: svn://svn.chromium.org/blink/trunk@180215 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It seems in certain cases font fallback returns a fontconfig interface
id which can then later not be reused by Skia. This patch fixes an
incorrect fallback to using the family in that case and returns a nullptr
instead, which fixes the assertion failure. However, we need to
investigate why the Skia typeface creation with the returned
fciId is unsuccessful, filed crbug.com/403348 for this remaining question.

BUG=273261
R=eae,eseidel

Review URL: https://codereview.chromium.org/470493003

git-svn-id: svn://svn.chromium.org/blink/trunk@180214 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch adds the flag 'animatable' to CSSProperties.in, replacing the
existing switch statement in CSSAnimations.cpp. Setting the flag
indicates that the property can be animated by CSS animations and
transitions.

The added code will generate a function in CSSPropertyMetadata, which
will eventually also support functions like isInheritedProperty and
isValidFirstLetterStyleProperty. This means that the individual lines in
CSSProperties.in may start getting a bit long, although a centralized
specification of our supported CSS properties arguably helps developers
add new CSS properties, by reducing the lines they need to add and
adding documentation on what changes need to be made.

BUG=396992

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180039

Review URL: https://codereview.chromium.org/454423003

git-svn-id: svn://svn.chromium.org/blink/trunk@180213 bbb929c8-8fbe-4397-9dbb-9b2b20218538