GitLab

Instead of driving the blend mode computation from
updateDescendantDependentFlags, this CL drives the blend mode computation from
CompositingInputsUpdater. Using this approach lets us remove a bunch of
bookkeeping because CompositingInputsUpdater does all the bookkeeping for us.

Because CompositingInputsUpdater happens during the compositing update, we need
to add some code to the compositing update to run the CompositingInputsUpdater
even when compositing isn't available (e.g., SVGImage and InspectorOverlay).

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/466193002

git-svn-id: svn://svn.chromium.org/blink/trunk@180133 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=180119

TBR=abarth@chromium.org

Review URL: https://codereview.chromium.org/465173002

git-svn-id: svn://svn.chromium.org/blink/trunk@180132 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It seems this one is no longer needed, just
the one in HTMLCollection is used.

Review URL: https://codereview.chromium.org/461213002

git-svn-id: svn://svn.chromium.org/blink/trunk@180131 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Previous implementation had three CoreInitializer methods: initEventNames,
initEventTargetNames, and initBindings, which were never called, and
ModulesInitializer contained identically named methods which were a
superset of the base class implementation. This change removes duplication
and separates core and modules initialization for a cleaner implementation.

This change doesn't fix 393728, but is a necessary refactoring for the
actual fix.

BUG=393728

Review URL: https://codereview.chromium.org/464043002

git-svn-id: svn://svn.chromium.org/blink/trunk@180130 bbb929c8-8fbe-4397-9dbb-9b2b20218538

A recent CL to Promises changed when JavaScript is run when a promise
is resolved. In the past the JS ran immediately -- now it is run
using a microtask. For EME code, the "message" event is generated right
after creating the session. It relies on the JS running and binding
an event handler. With the recent CL, it is possible for the event to
be generated before the handler is bound, so the event is lost. This
change delays posting the "message" event if there are no event handlers
registered.

BUG=402766,403121
TEST=Disabled EME browser_test passes

Review URL: https://codereview.chromium.org/465053002

git-svn-id: svn://svn.chromium.org/blink/trunk@180128 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It seems these are not used (anymore).

Review URL: https://codereview.chromium.org/463173003

git-svn-id: svn://svn.chromium.org/blink/trunk@180127 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch removes unnecessary blink:: prefixes.
Also it merges *namespace blink{}* blocks.
Also correcting blink presubmit errors in the changed files.

BUG=

Review URL: https://codereview.chromium.org/467573002

git-svn-id: svn://svn.chromium.org/blink/trunk@180126 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This converts the remaining ui test suites, besides the two that are currently failing (ct-failure-card, ct-failure-stream).

BUG=400387
R=ojan@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/464163002

git-svn-id: svn://svn.chromium.org/blink/trunk@180125 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Previously, it was possible to read stale data from
hasSelfPaintingLayerDescendant. After this CL, we clean out the dirty bit
automatically when you read this state. In order to make this work, this CL
separates hasSelfPaintingLayerDescendant from updateDescendantDependentFlags.

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/468633002

git-svn-id: svn://svn.chromium.org/blink/trunk@180124 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Avoid importing net.html in the top level and instead have the files that
use it importing it themselves.

NOTRY=true

Review URL: https://codereview.chromium.org/467013002

git-svn-id: svn://svn.chromium.org/blink/trunk@180123 bbb929c8-8fbe-4397-9dbb-9b2b20218538

A recent CL (Blink revision 179922) changed the way JavaScript is run
after resolving a promise (used to run immediately, now run by a
microtask later). encrypted-media-lifetime-mediakeysession-reference.html
started failing as it expected the PromiseResolver object to be around
and counted. Fixing the test as the PromiseResolver object is gone by the
time the JavaScript executes after resolving a promise.

BUG=374772,378245
TEST=failing test passes

Review URL: https://codereview.chromium.org/465833003

git-svn-id: svn://svn.chromium.org/blink/trunk@180121 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We don't want to pass strings into the browser process for conversion
into URL objects; we should validate input when creating data from
JavaScript.

This patch adjusts {Local,Federated}Credential objects to throw exceptions
in their constructors if URL validation fails, and changes both Web*Credential
and Platform*Credential objects to accept WebURL/KURL objects rather
than strings.

BUG=400674

Review URL: https://codereview.chromium.org/464783003

git-svn-id: svn://svn.chromium.org/blink/trunk@180120 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In some obscure code paths (e.g., printing, drawing drag images), we would
paint without calling updateLayoutAndStyleForPainting, which means we skip some
important steps in the DocumentLifecycle.

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/457973003

git-svn-id: svn://svn.chromium.org/blink/trunk@180119 bbb929c8-8fbe-4397-9dbb-9b2b20218538

described here: 
http://lists.w3.org/Archives/Public/public-css-testsuite/2014May/0000.html

Review URL: https://codereview.chromium.org/311673002

git-svn-id: svn://svn.chromium.org/blink/trunk@180118 bbb929c8-8fbe-4397-9dbb-9b2b20218538

I broke sheriff-o-matic by not updating the link to scripts/net.js when
I transitioned it to lib/net.html. Updating fixes the UI.

NOTRY=true

Review URL: https://codereview.chromium.org/465093002

git-svn-id: svn://svn.chromium.org/blink/trunk@180116 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=180095

BUG=401954
TBR=adamk@chromium.org

Review URL: https://codereview.chromium.org/462263002

git-svn-id: svn://svn.chromium.org/blink/trunk@180115 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The InspectorOverlay does a bunch of manual poking and prodding of the layout
state machine. Instead, InspectorOverlay should just use
updateLayoutAndStyleForPainting like the regular engine does.

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

Review URL: https://codereview.chromium.org/465033002

git-svn-id: svn://svn.chromium.org/blink/trunk@180114 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This code is dead. This CL also removes a bunch of other related dead code.

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/465983003

git-svn-id: svn://svn.chromium.org/blink/trunk@180112 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL removes the "isolatesBlending" string from render tree dumps. When we
compute render tree dumps, this state is not always up-to-date, which means
we're getting stale information in the render tree dumps. We don't include
other descendant-dependent flags in render tree dumps for this reason.

This CL costs us some test coverage, but the appropriate way to regain that
test coverage to use reference tests to ensure that the pixels we draw are
correctly blended rather than hard-coding internal state from the engine into
test results.

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/466943002

git-svn-id: svn://svn.chromium.org/blink/trunk@180111 bbb929c8-8fbe-4397-9dbb-9b2b20218538

To keep us abreast of all the changes going on in sheriff-o-matic
code and keep us from stepping on each others' toes, I'm adding
a watchlist and putting all the folks currently actively working
on it on the list.

NOTRY=true

Review URL: https://codereview.chromium.org/461233002

git-svn-id: svn://svn.chromium.org/blink/trunk@180110 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The spec tells us that if an element has a title attribute, its value
should be returned as its tooltip/advisory information:

  http://www.whatwg.org/specs/web-apps/current-work/#the-title-attribute

Align with how other browsers behave for <foo title> and return the
empty string rather than consult the parent element.

R=
BUG=399255

Review URL: https://codereview.chromium.org/462073003

git-svn-id: svn://svn.chromium.org/blink/trunk@180109 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We shouldn't ever ASSERT a condition and then handle its negation. If we can
get here with layout dirty, we should crash so we can fix those cases.

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/466063002

git-svn-id: svn://svn.chromium.org/blink/trunk@180108 bbb929c8-8fbe-4397-9dbb-9b2b20218538

m_dirtySourceRect is duplicating the paint invalidation logic
by storing the exact same information.

On top of that, FilterEffectRendererHelper::prepareFilterEffect
already gets the invalidation rect from |dirtyRect|.

This change is covered by css3/filters/filter-repaint-*.html
that were added with m_dirtySourceRect in
https://bugs.webkit.org/show_bug.cgi?id=80323.

Review URL: https://codereview.chromium.org/453033005

git-svn-id: svn://svn.chromium.org/blink/trunk@180107 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Cleanup namespace usage in Source/core/modules/[battery/* to indexeddb/*]

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

BUG=None

Review URL: https://codereview.chromium.org/466813003

git-svn-id: svn://svn.chromium.org/blink/trunk@180106 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Cleanup namespace usage in Source/core/modules/[mediasource/* to websockets/*]

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

BUG=None

Review URL: https://codereview.chromium.org/461163002

git-svn-id: svn://svn.chromium.org/blink/trunk@180105 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Bug fixed on the Chromium side, just adding a layout test as well
for full coverage.

BUG=395472
R=cmumford@chromium.org

Review URL: https://codereview.chromium.org/462153002

git-svn-id: svn://svn.chromium.org/blink/trunk@180103 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Let RenderObject::canUpdateSelectionOnRootLineBoxes() return false
if the renderer is in an orphaned render tree, to prevent
RenderSelectionInfo from calling selectionRectForPaintInvalidation()
which would cause the crash.

BUG=396596
TEST=editing/selection/move-node-in-selection-crash.html

Review URL: https://codereview.chromium.org/452293004

git-svn-id: svn://svn.chromium.org/blink/trunk@180102 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Moving them to lib, wrapping them in script tags, and converting
the test to Mocha.

BUG=400388
NOTRY=true

Review URL: https://codereview.chromium.org/464003002

git-svn-id: svn://svn.chromium.org/blink/trunk@180101 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This fixes an issue where "last updated" data would be displayed before date
was populated, showing some weird data. I tried to use "template if" to
restrict when this would be painted, but couldn't get that to work, so I
resorted to using styling instead.

NOTRY=true
BUG=402661

Review URL: https://codereview.chromium.org/466783002

git-svn-id: svn://svn.chromium.org/blink/trunk@180100 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Address type conversion subtlety over RawPtr<>s. We have at hand:

 Handle<Value> toV8(PassRefPtrWillBeRawPtr<NodeList>, ...);

 template<NodeType>class StaticNodeTypeList : public NodeList {
   public:
       static PassRefPtrWillBeRawPtr<StaticNodeTypeList> adopt(..);
       ...
   }

 typedef StaticNodeTypeList<Element> StaticElementList;

and the call

 toV8(StaticElementList::adopt(namedItems), ....);

For the toV8() call to be successfully resolved, the adopt() result
must be convertible to PassRefPtrWillBeRawPtr<NodeList>. For the non-Oilpan
case, we do have the

   PassRefPtr(const PassRefPtr<U>, EnsurePtrConvertibleArgDecl(U, T))

constructor at hand, which will convert the PassRefPtr<StaticElementList>
into PassRefPtr<NodeList>. We do not have the corresponding type-convertible
constructor over RawPtr<>, as providing that runs into problems when
using RawPtr<> as part of the PassOwnPtrWillBeRawPtr transition type.

Address/step-around-the-issue by being explicit about the conversion.

TBR=haraken,oilpan-reviews
BUG=
NOTRY=true

Review URL: https://codereview.chromium.org/451303003

git-svn-id: svn://svn.chromium.org/blink/trunk@180096 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The currently checked-in result was incorrectly put there by
https://src.chromium.org/viewvc/blink?revision=180085&view=revision, while
the actual results are correct.

BUG=401954
TBR=bungeman@chromium.org,fdegans@chromium.org

Review URL: https://codereview.chromium.org/467593003

git-svn-id: svn://svn.chromium.org/blink/trunk@180095 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Also, adding cmumford@ to modules/indexeddb OWNERS

TBR=alecflett@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/466833002

git-svn-id: svn://svn.chromium.org/blink/trunk@180094 bbb929c8-8fbe-4397-9dbb-9b2b20218538

NOTRY=true

Review URL: https://codereview.chromium.org/458353005

git-svn-id: svn://svn.chromium.org/blink/trunk@180091 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This reverts https://src.chromium.org/viewvc/blink?revision=180034&view=revision
and its two followups:
https://src.chromium.org/viewvc/blink?revision=180039&view=revision
https://src.chromium.org/viewvc/blink?revision=180051&view=revision

r180034 caused lots of browser_test failures in debug mode, all with the
following ASSERT:

ASSERTION FAILED: RuntimeCSSEnabled::isCSSPropertyEnabled(propertyID) || isInternalProperty(propertyID)
../../third_party/WebKit/Source/core/css/StylePropertySet.cpp(558) : int blink::MutableStylePropertySet::findPropertyIndex(blink::CSSPropertyID) const

Most of the failures I saw were in the PrintPreviewWebUITest suite. The followup
patches had to be reverted to make the revert of r180034 clean.

BUG=396992
TBR=timloh@chromium.org

Review URL: https://codereview.chromium.org/468543002

git-svn-id: svn://svn.chromium.org/blink/trunk@180090 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Apparently, Failure does not include ImageOnlyFailure.


git-svn-id: svn://svn.chromium.org/blink/trunk@180089 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=340522

Review URL: https://codereview.chromium.org/422273002

git-svn-id: svn://svn.chromium.org/blink/trunk@180088 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The original changes to the test expectations have already been reverted.
See https://codereview.chromium.org/466823002/


git-svn-id: svn://svn.chromium.org/blink/trunk@180087 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Rebaseline font tests for Android after r288168 (https://codereview.chromium.org/453923002/)

Reason for revert:
Failure in blink_tests
The old files now seem correct.
http://build.chromium.org/p/chromium.webkit/builders/WebKit%20Android%20%28Nexus4%29/builds/18077

Original issue's description:
> Rebaseline font tests for Android after r288168
> 
> - svg/overflow/overflow-on-outermost-svg-element-in-xhtml-defaults.xhtml
> - svg/W3C-SVG-1.1-SE/coords-dom-02-f.svg
> - svg/as-background-image/svg-width-100p-as-background.html
> - css1/font_properties/font_variant.html
> - virtual/antialiasedtext/fast/text/chromium-linux-fontconfig-renderstyle.html
> - fast/text/chromium-linux-fontconfig-renderstyle.html
> 
> 
> TBR=mvanouwerkerk@chromium.org,zerny@chromium.org
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=179820

TBR=mvanouwerkerk@chromium.org,zerny@chromium.org,haraken@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/466823002

git-svn-id: svn://svn.chromium.org/blink/trunk@180085 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Merge NamedNodesCollection and StaticNodeList classes into a single
StaticNodeTypeList templated class as these are identical besides the type
of Node they hold. NamedNodesCollection was holding Elements so it is now
named StaticElementList which is a typedef to StaticNodeTypeList<Element>.
StaticNodeList holds Node and is now a typedef to StaticNodeTypeList<Node>.

This avoids code duplication and allows for tighter typing when using a
StaticNodeList. As a side effect, memory used by StaticElementList (formely
NamedNodesCollection is now reported to V8 as StaticNodeList code was doing
so).

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

Review URL: https://codereview.chromium.org/465483002

git-svn-id: svn://svn.chromium.org/blink/trunk@180084 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Sync: `make update`

Run the tests: `make test`

R=ojan@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/462973002

git-svn-id: svn://svn.chromium.org/blink/trunk@180082 bbb929c8-8fbe-4397-9dbb-9b2b20218538