GitLab

At insertion points, we schedule style recalcs for inserted nodes when we
have a StyleRecalcChange at the insertion point's parent. One of the reasons
is that :host-context selectors which affect inserted nodes may rely on
matching nodes which are not ascendants of the inserted nodes in the dom
used for selector matching.

We did use a LocalStyleChange for the inserted nodes, but these selectors
may affect any node in the inserted node's subtree. Hence, we start using
SubtreeStyleChange instead.

R=esprehn@chromium.org,dglazkov@chromium.org
BUG=381020

Review URL: https://codereview.chromium.org/317953002

git-svn-id: svn://svn.chromium.org/blink/trunk@175636 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/315293003

git-svn-id: svn://svn.chromium.org/blink/trunk@175635 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This change introduces a UI that appears when the last roll
is getting old. The UI is not shown by default to avoid
visual clutter and its visual attribute matches the urgency
(currently 2 levels: warning / urgent).

Review URL: https://codereview.chromium.org/316033002

git-svn-id: svn://svn.chromium.org/blink/trunk@175634 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It doesn't need to get host echoed back. Just use the simple echo
handler.

Some unnecessary div elements are removed.

BUG=none

Review URL: https://codereview.chromium.org/317113002

git-svn-id: svn://svn.chromium.org/blink/trunk@175633 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Simplify CSSSelector::isAttributeSelector() by not checking each Match type
explicitly and reordering the Match enum instead so that the attribute
selector match type are contiguous and at the end of the enum. This way, the
check can become (m_match >= FirstAttributeSelectorMatch).

This optimization is similar to the one we already use in CollectionType.h
for collection types. isAttributeSelector() is called from
SelectorChecker::checkOne() which is hot code.

I see a ~2% progression on Dromaeo's cssquery-jquery:
http://dromaeo.com/?id=222514,222518

Review URL: https://codereview.chromium.org/315043005

git-svn-id: svn://svn.chromium.org/blink/trunk@175632 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Moves "constructors on modules global objects from core interfaces" to modules.
I.e., moves ServiceWorkerGlobalScopeCore.idl from
gen/blink/core to gen/blink/modules

bindings/core/generated.gyp no longer references modules!

R=haraken
BUG=358074

Review URL: https://codereview.chromium.org/319593006

git-svn-id: svn://svn.chromium.org/blink/trunk@175631 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently parseHTTPHeaders returns (the number of consumed bytes - x) where
 - x = 0 if it consumes an empty line
 - x = 1 otherwise
This CL make parseHTTPHeaders always return the number of consumed bytes.

BUG=NONE
R=tyoshino

Review URL: https://codereview.chromium.org/314083002

git-svn-id: svn://svn.chromium.org/blink/trunk@175630 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL implements a |hardwareConcurrency| property on the
|navigator| object that returns the number of CPU cores the
machine has.

Intent-to-ship:
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/xwl0ab20hVc

R=esprehn@chromium.org
BUG=369629

Review URL: https://codereview.chromium.org/262723002

git-svn-id: svn://svn.chromium.org/blink/trunk@175629 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175599

BUG=381242
TBR=wangxianzhu@chromium.org

Review URL: https://codereview.chromium.org/321503002

git-svn-id: svn://svn.chromium.org/blink/trunk@175628 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=380990

Review URL: https://codereview.chromium.org/315163002

git-svn-id: svn://svn.chromium.org/blink/trunk@175627 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Followup to:
Cleanup v8 usage in modules
https://codereview.chromium.org/319653002/

Review URL: https://codereview.chromium.org/314063002

git-svn-id: svn://svn.chromium.org/blink/trunk@175626 bbb929c8-8fbe-4397-9dbb-9b2b20218538

all compositing layers and their geometry are updated.

BUG=370664

Review URL: https://codereview.chromium.org/311813005

git-svn-id: svn://svn.chromium.org/blink/trunk@175625 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Reason for revert:
Broke the build.

Original issue's description:
> Implement enableFastMobileScrolling web api.
> 
> The effort needs to be done in 3 steps:
> 1. Create the runtime flag interface in blink;
> 2. Enable the flag for Android from chromium even it does not do anything in blink;
> 3. Replace compile time guard with runtime flag in blink and remove ENABLE_FAST_MOBILE_SCROLLING in gyp file.
> 
> This is the 1st set of changes.
> 
> BUG=381391
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175621

TBR=tkent@chromium.org,adamk@chromium.org,c.shu@samsung.com
NOTREECHECKS=true
NOTRY=true
BUG=381391

Review URL: https://codereview.chromium.org/317313002

git-svn-id: svn://svn.chromium.org/blink/trunk@175622 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The effort needs to be done in 3 steps:
1. Create the runtime flag interface in blink;
2. Enable the flag for Android from chromium even it does not do anything in blink;
3. Replace compile time guard with runtime flag in blink and remove ENABLE_FAST_MOBILE_SCROLLING in gyp file.

This is the 1st set of changes.

BUG=381391

Review URL: https://codereview.chromium.org/321473002

git-svn-id: svn://svn.chromium.org/blink/trunk@175621 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This method has been renamed already, so it's not a strict revert, but
it restores the original behavior.

BUG=none
TEST=browser_tests --gtest_filter=WebViewTest*Auto*
TBR=ojan@chromium.org

Review URL: https://codereview.chromium.org/316373005

git-svn-id: svn://svn.chromium.org/blink/trunk@175620 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This method is all sorts of badness. It leads to strange chicken-egg
bugs because people don't think that you will need to update layer
state after the layout has completed. 

FrameView::updateLayoutAndStyleIfNeededRecursive needs to do a second
style recalc and layout just for this and in theory it should probably
do a while loop there to get this right.

Now that we've removed all the non-SVG filter callers of this function,
rename it to something scary so that people won't add new callers.

Review URL: https://codereview.chromium.org/319613004

git-svn-id: svn://svn.chromium.org/blink/trunk@175618 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=380167
R=japhet@chromium.org

Review URL: https://codereview.chromium.org/316313002

git-svn-id: svn://svn.chromium.org/blink/trunk@175617 bbb929c8-8fbe-4397-9dbb-9b2b20218538

whitespace-length.html is timing out consistently on Webkit XP:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#group=%40ToT%20Blink&tests=fast/svg/whitespace-length.html

BUG=377503,339899
TBR=ed@opera.com

Review URL: https://codereview.chromium.org/319883002

git-svn-id: svn://svn.chromium.org/blink/trunk@175616 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This timed out on the waterfall and is failing on unrelated tryjobs runs
as well.

TBR=jsbell@chromium.org
BUG=380910

Review URL: https://codereview.chromium.org/316383002

git-svn-id: svn://svn.chromium.org/blink/trunk@175615 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=eae@chromium.org

Review URL: https://codereview.chromium.org/316373002

git-svn-id: svn://svn.chromium.org/blink/trunk@175614 bbb929c8-8fbe-4397-9dbb-9b2b20218538

https://src.chromium.org/viewvc/blink?view=rev&revision=175416

BUG=none
TBR=peter@chromium.org

Review URL: https://codereview.chromium.org/315423002

git-svn-id: svn://svn.chromium.org/blink/trunk@175613 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175388

BUG=371640
TBR=dsinclair@chromium.org

Review URL: https://codereview.chromium.org/314283004

git-svn-id: svn://svn.chromium.org/blink/trunk@175612 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Rename ElementData's attribute related methods for clarity:
* length() -> attributeCount()
* isEmpty() -> hasAttributes()
* getAttributeItem(name) -> findAttributeByName(name)
* attributeItem(index) -> attributeAt(index)
* getAttributeItemIndex(name) -> findAttributeIndexByName(name)
* addAttribute(name, value) -> appendAttribute(name, value)
* removeAttribute(index) -> removeAttributeAt(index)

This is inpired by the following WebKit revision by benjamin@webkit.org:
http://trac.webkit.org/changeset/162394

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

Review URL: https://codereview.chromium.org/313973003

git-svn-id: svn://svn.chromium.org/blink/trunk@175611 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Oilpan: Prepare to move AbstractSQLStatement, SQLStatement, and SQLCallbackWrapper<T> to Oilpan heap.

SQLCallbackWrapper has code to call ExecutionContext::deref() in the origin
thread. This is unnecessary in Oilpan because implementation classes of
ExecutionContext are already on-heap and they are always destructed in their
origin threads.

This is a prepration to make ExecutionContext GarbageCollected.

BUG=357163

Review URL: https://codereview.chromium.org/311273005

git-svn-id: svn://svn.chromium.org/blink/trunk@175610 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Keep a traced Member reference to the Element that's recorded as
the Document's CSS target. When that Element is removed from the
tree, the Member will be explicitly cleared.

R=
BUG=357163

Review URL: https://codereview.chromium.org/315333002

git-svn-id: svn://svn.chromium.org/blink/trunk@175609 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Since rAc was adopted into the standard, we revisited this and decided
to do it a different way. This "revert" does keep the minor change
about removing the extra param to
WebAutofillClient::didRequestAutocomplete.

BUG=354897

Review URL: https://codereview.chromium.org/293853002

git-svn-id: svn://svn.chromium.org/blink/trunk@175608 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=381364
TBR=peter@chromium.org

Review URL: https://codereview.chromium.org/322443003

git-svn-id: svn://svn.chromium.org/blink/trunk@175607 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This allows the file to be compiled with C++11.

BUG=

Review URL: https://codereview.chromium.org/316113002

git-svn-id: svn://svn.chromium.org/blink/trunk@175606 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=381471
TBR=peter@chromium.org

Review URL: https://codereview.chromium.org/317283002

git-svn-id: svn://svn.chromium.org/blink/trunk@175605 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=381464
TBR=peter@chromium.org

Review URL: https://codereview.chromium.org/314113004

git-svn-id: svn://svn.chromium.org/blink/trunk@175604 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This fixes cases where a node that previously had a >1 multiplier ceases to have direct text children.  This can happen if its text content is removed in script, or if its 'display' style is changed from inline to block.  (There are probably other ways for it to happen also.)

Previously such a node would have a "stale" multiplier, since the autosizer only updated multipliers on text nodes and their immediate parents.  In theory the stale multiplier should not have affected the rendering, but:

- it is observable through the lineHeight property, and
- due to http://crbug.com/380903 the stale multiplier on an ancestor might be used to render text, even when the text node has the correct multiplier

BUG=378959

Review URL: https://codereview.chromium.org/319013003

git-svn-id: svn://svn.chromium.org/blink/trunk@175603 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=dominicc
NOTRY=true

Review URL: https://codereview.chromium.org/319093002

git-svn-id: svn://svn.chromium.org/blink/trunk@175602 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Not having this can cause problems in the test harnass, where some of the
FrameLoaderClients do substantially more work than they do in a non-test
environment.

BUG=357283

Review URL: https://codereview.chromium.org/317513002

git-svn-id: svn://svn.chromium.org/blink/trunk@175601 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The IDB spec defines that transactions are created 'active' and become
'inactive' when control returns to the event loop. The way this was
implemented, transactions created within microtasks (e.g. Promise
callbacks) were not deactivated.

Flip the order of microtasks and deactivation, and add tests.

BUG=380910
R=adamk@chromium.org

Review URL: https://codereview.chromium.org/321433002

git-svn-id: svn://svn.chromium.org/blink/trunk@175600 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Some async layout tests use waitUntilDone() and notifyDone() by
themselves, without setting window.testIsAsync to true.

This may be either unnecessary or incorrect, depending on which of the
test's notifyDone() and text-based-repaint.js's notifyDone() executes
first. If the latter executes first, the test may finish before it's
done causing what the test want to test not fully executed.

BUG=381242

Review URL: https://codereview.chromium.org/316193003

git-svn-id: svn://svn.chromium.org/blink/trunk@175599 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=381451
TBR=peter@chromium.org

Review URL: https://codereview.chromium.org/321493002

git-svn-id: svn://svn.chromium.org/blink/trunk@175598 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This was added in http://src.chromium.org/viewvc/blink?view=revision&revision=80584.
It didn't have tests, so I'm not sure how to test this.
But plugins always have a RenderLayer, so there should be no
need to schedule a layer update anymore.

Review URL: https://codereview.chromium.org/318773003

git-svn-id: svn://svn.chromium.org/blink/trunk@175597 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=falken@chromium.org,jkarlin@chromium.org
BUG=374822

Review URL: https://codereview.chromium.org/311233007

git-svn-id: svn://svn.chromium.org/blink/trunk@175596 bbb929c8-8fbe-4397-9dbb-9b2b20218538

needed to be updated. Long press on text will no longer show a context menu.

BUG=381152

Review URL: https://codereview.chromium.org/320453002

git-svn-id: svn://svn.chromium.org/blink/trunk@175595 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=dcheng@chromium.org

Review URL: https://codereview.chromium.org/319813003

git-svn-id: svn://svn.chromium.org/blink/trunk@175594 bbb929c8-8fbe-4397-9dbb-9b2b20218538