GitLab

This CL broke worker-reload-repeated.html in oilpan bots.

> Oilpan: Move ThreadableWebSocketChannelClientWrapper to Oilpan's heap
> 
> This is a preparation for moving WebSocketChannelClient to the heap.
> 
> This CL adds a template specialization to CrossThreadCopier.h so that ThreadableWebSocketChannelClientWrapper* can be passed to createCallbackTask().
> 
> BUG=340522
> 
> Review URL: https://codereview.chromium.org/267323004

TBR=haraken@chromium.org

Review URL: https://codereview.chromium.org/309503013

git-svn-id: svn://svn.chromium.org/blink/trunk@175242 bbb929c8-8fbe-4397-9dbb-9b2b20218538

- Basically, this CL replaces V8AbstractEventListener::m_world with V8AbstractEventListener::m_scriptState.

- This CL uses the ScriptState in a lot of places about event listeners and cleans up the code (for example, this CL removes an empty creation context from toV8()).

- The complicated part is that V8LazyEventListener does not know the ScriptState when it's created. Thus, V8AbstractEventListener::m_scriptState can be null if the listener is a V8LazyEventListener until the V8LazyEventListener is called back from Blink (i.e., either of handleEvent(), callListenerFunction() and getEventListener() is called).

- This CL changes web-exposed behavior. In a case where the frame that registered an onload event handler (say Frame1) is different from the frame that triggers the onload event (Frame2), before this CL, the onload event handler is invoked in the context of Frame2. After this CL, the onload event handler is not triggered if Frame1 != Frame2. This behavior is not speced but aligns with Firefox. See comment #17 and #19 for more details.

BUG=357144

Review URL: https://codereview.chromium.org/293053007

git-svn-id: svn://svn.chromium.org/blink/trunk@175241 bbb929c8-8fbe-4397-9dbb-9b2b20218538

> The reason for reverting is: broke interactive_ui_tests:
> 
> Referrer
> SearchReusesInstantTab
> TypedSearchURLDoesntReuseInstantTab.
I fixed these tests in https://src.chromium.org/viewvc/chrome?view=rev&revision=273996 .

Original Description:

Initializing |v8::Context| is a heavy operation, and should be avoided if necessary. This is meant to be delayed until first time |toV8Context| is called, but |ScriptController::updateDocument| was forcing creation.

This patch fixes the if branch so that it would not create a |v8::Context| if it does not exist.

BUG=368548, 368555

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=173044

Review URL: https://codereview.chromium.org/263583002

git-svn-id: svn://svn.chromium.org/blink/trunk@175240 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This belongs on the HTML elements, not on the renderer which is torn
down when the fallback content is attached.

BUG=379594

Review URL: https://codereview.chromium.org/301353005

git-svn-id: svn://svn.chromium.org/blink/trunk@175239 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Automatically generated using:
webkit-patch update-flaky-tests

R=eseidel
BUG=

Review URL: https://codereview.chromium.org/306293002

git-svn-id: svn://svn.chromium.org/blink/trunk@175238 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Spec: https://slightlyoff.github.io/ServiceWorker/spec/service_worker/#navigator-service-worker-controller

BUG=363967

Review URL: https://codereview.chromium.org/301103003

git-svn-id: svn://svn.chromium.org/blink/trunk@175237 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL contains a number of minor cleanups to RenderLayerStackingNode. The
most interesting change is merging ancestorStackingContextNode and
ancestorStackingNode, which do exactly the same thing.

Review URL: https://codereview.chromium.org/301233008

git-svn-id: svn://svn.chromium.org/blink/trunk@175236 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Unreviewed rebaseline of virtual/gpu/compositedscrolling/overflow/overflow-positioning.html

TBR=ojan@chromium.org

Review URL: https://codereview.chromium.org/309753002

git-svn-id: svn://svn.chromium.org/blink/trunk@175235 bbb929c8-8fbe-4397-9dbb-9b2b20218538

- write-while-waiting.html was fixed by r175127.
- Tighten terminate-during-sync-operation-*.html expectations

BUG=357427
TBR=sigbjornf@opera.com

Review URL: https://codereview.chromium.org/308673008

git-svn-id: svn://svn.chromium.org/blink/trunk@175234 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove unnecessary Persistent reference (to an object contained within
another Persistent) + arrange for manual GCing of resources to avoid
test objects being reported as not destructed and leaking with
Oilpan enabled.

R=
BUG=

Review URL: https://codereview.chromium.org/306273002

git-svn-id: svn://svn.chromium.org/blink/trunk@175233 bbb929c8-8fbe-4397-9dbb-9b2b20218538

For :hover, :active, :focus, and :drag, childrenAffectedBy really means
that descendants, or siblings, or elements in the siblings' subtrees may be
affected. Methods and constants are renamed to reflect that.

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/308683007

git-svn-id: svn://svn.chromium.org/blink/trunk@175232 bbb929c8-8fbe-4397-9dbb-9b2b20218538

No one reads this state so we can remove it.

Review URL: https://codereview.chromium.org/306123008

git-svn-id: svn://svn.chromium.org/blink/trunk@175231 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Right now when it exceeds the appengine storage limit for a single
entry it 500s. Instead, just create a new entry. This isn't totally
ideal. Ideally we'd create a new entry but link it to the old
one so we can show it correctly in the UI, but that's too much work.

NOTRY=true
BUG=378490

Review URL: https://codereview.chromium.org/304303003

git-svn-id: svn://svn.chromium.org/blink/trunk@175230 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=326139
TBR=mkwst@chromium.org

Review URL: https://codereview.chromium.org/305323002

git-svn-id: svn://svn.chromium.org/blink/trunk@175229 bbb929c8-8fbe-4397-9dbb-9b2b20218538

I accidentally already landed part of this support
in an earlier change, but this makes --upload actually
work.  You can see an example change here:
https://codereview.chromium.org/308743011/
(which again accidentally includes this change).

This also adds a hack-and-slash version of sheriff-calendar
code which I and Eric Boren originally wrote for the AutoRollBot
but that lives in a different repository:
https://chromium.googlesource.com/chromium/tools/build/+/master/scripts/tools/blink_roller/auto_roll.py

This combined with a cron job should make
it easy for gardeners to keep FlakyTests up to date.

Looking at 308743011, we clearly need to address some
of this marauding flakiness as FlakyTests should be
a very stable file.  Right now it is changing by the hour.

BUG=242366

Review URL: https://codereview.chromium.org/307183002

git-svn-id: svn://svn.chromium.org/blink/trunk@175228 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Windows XP specific rebaselines for r175038 as the XP > Win7 fallback is
no longer valid for text tests.

TBR=dglazkov@chromium.org

Review URL: https://codereview.chromium.org/308653008

git-svn-id: svn://svn.chromium.org/blink/trunk@175227 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The showSecurityErrorMessage wasn't called by anything. So I killed it.

R=morrita
BUG=

Review URL: https://codereview.chromium.org/305643009

git-svn-id: svn://svn.chromium.org/blink/trunk@175226 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Prior to this CL, we computed RenderLayer::isUnclippedDescendant in a separate
O(n^2) walk of the out-of-flow RenderLayers. On workloads like famo.us, this
walk is quite expensive because everything is out-of-flow positioned, and
O(n^2) is large.

After this CL, we compute this information as we walk down the tree in the
CompositingPropertyUpdater, which means we can compute the information in
linear time and we can entirely remove the separate tree walk. Moving this
computation also lets us rip out a large amount of machinery that existed to
prune the extra tree walk that this CL removes.

Now that we compute this information with the CompositingPropertyUpdater,
we compute it even when composited overflow scrolling is off.  This confused
some tests that used Internals to poke this state.  I've added a workaround to
keep theses tests working identically.  I'll remove the work around and update
the tests in the next CL.

Review URL: https://codereview.chromium.org/309743002

git-svn-id: svn://svn.chromium.org/blink/trunk@175225 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The notion of "importedDocument" only exists at the LinkImport level
(because its used, HTMLLinkElement has to deal with both importer
and the importee), so we can just call the imported documents simply
"documents" inside of the imports plumbing.

No functional changes, plumbotron.

R=morrita
BUG=

Review URL: https://codereview.chromium.org/298413004

git-svn-id: svn://svn.chromium.org/blink/trunk@175224 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We can just test root-ness by the absence of parent.

No behavior changes, just plumbotron.

R=morrita
BUG=

Review URL: https://codereview.chromium.org/302433016

git-svn-id: svn://svn.chromium.org/blink/trunk@175223 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175220

BUG=331251
TBR=robhogan@gmail.com

Review URL: https://codereview.chromium.org/306253002

git-svn-id: svn://svn.chromium.org/blink/trunk@175222 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When a block element becomes a float or out of flow we should strip
any anonymous blocks wrapping its inline siblings.

BUG=331251

Review URL: https://codereview.chromium.org/253313005

git-svn-id: svn://svn.chromium.org/blink/trunk@175220 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This does not change behavior because Chrome's service worker
dispatcher always mints new WebServiceWorker instances, and
WebServiceWorker::proxy returns null.

BUG=361907

Review URL: https://codereview.chromium.org/306133002

git-svn-id: svn://svn.chromium.org/blink/trunk@175219 bbb929c8-8fbe-4397-9dbb-9b2b20218538

CSS 2.1 (http://www.w3.org/TR/CSS2/visudet.html#leading) says: "Empty inline 
elements generate empty inline boxes, but these boxes still have margins, padding, 
borders and a line height, and thus influence [line-height calculation] just 
like elements with content."

We weren't creating lineboxes for empty inline elements when they occurred
between a single trailing space and a single leading space - this is because
our linebox creation was based on the too-narrow premise that we were already
ignoring spaces. However, if an empty inline element comes after a trailing 
space but before one or more leading spaces then it will be ignored as part 
of the collapsed space between the trailing space and the next non-collapsible 
space - so we should force the creation of a linebox in such cases.

This allows us to pass Hixie's empty inline test: 
http://www.hixie.ch/tests/evil/mixed/emptyinline.html

BUG=328939

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175205

Review URL: https://codereview.chromium.org/300853007

git-svn-id: svn://svn.chromium.org/blink/trunk@175218 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This warning is disabled in the GYP build but was missed.

Review URL: https://codereview.chromium.org/305303002

git-svn-id: svn://svn.chromium.org/blink/trunk@175217 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Un-inline KURL's default constructor to avoid bloating. KURL has non-POD data
members and its constructor is thus non-trivial.

This reduces the release binary size by ~17Kb.

R=eseidel@chromium.org

Review URL: https://codereview.chromium.org/307193002

git-svn-id: svn://svn.chromium.org/blink/trunk@175216 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Un-inline Node's constructor to avoid bloating. Node's constructor is not
trivial and Node is used in inheritance.

This reduces the release binary size by ~4Kb. I did not see any impact on the
Dromaeo score:
http://dromaeo.com/?id=222265,222267

R=eseidel@chromium.org

Review URL: https://codereview.chromium.org/309613006

git-svn-id: svn://svn.chromium.org/blink/trunk@175215 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Since it owns the HTMLImportTreeRoot, it can just ask it for the
master document. Additionally, remove indirection plumbing that's
just straightforward queries of master document.

No changes in behavior, refactoring.

R=morrita
BUG=

Review URL: https://codereview.chromium.org/305743003

git-svn-id: svn://svn.chromium.org/blink/trunk@175214 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Un-inline Element's constructor to avoid bloating, as advised by Chromium
coding style:
http://www.chromium.org/developers/coding-style/cpp-dos-and-donts#TOC-Stop-inlining-constructors-and-destructors.

The Element class is used in inheritence and has non-POD data members. The
Element constructor is thus non-trivial.

This reduces release binary size by ~12Kb. I saw no impact on the Dromaeo
score:
http://dromaeo.com/?id=222265,222266

R=eseidel@chromium.org

Review URL: https://codereview.chromium.org/309593004

git-svn-id: svn://svn.chromium.org/blink/trunk@175213 bbb929c8-8fbe-4397-9dbb-9b2b20218538

needed anymore after
https://codereview.chromium.org/268223003

BUG=313438

Review URL: https://codereview.chromium.org/309553009

git-svn-id: svn://svn.chromium.org/blink/trunk@175212 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It seems a lot of code moved out of this class, so remove various includes and forward references.

Review URL: https://codereview.chromium.org/302093006

git-svn-id: svn://svn.chromium.org/blink/trunk@175211 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Windows XP specific rebaselines for r175038 as the XP > Win7 fallback is
no longer valid for text tests.

TBR=dglazkov@chromium.org

Review URL: https://codereview.chromium.org/301193007

git-svn-id: svn://svn.chromium.org/blink/trunk@175210 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Windows XP specific rebaselines for r175038 as the XP > Win7 fallback is
no longer valid for text tests.

TBR=dglazkov@chromium.org

Review URL: https://codereview.chromium.org/308123005

git-svn-id: svn://svn.chromium.org/blink/trunk@175209 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=pdr@chromium.org

Review URL: https://codereview.chromium.org/306083006

git-svn-id: svn://svn.chromium.org/blink/trunk@175208 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Reason for revert:
This changed the subpixel positioning of glyphs on lines with first-letter renderers on Mac builds. See https://storage.googleapis.com/chromium-layout-test-archives/WebKit_Mac10_6/28423/layout-test-results/results.html as an example.

I think the new results are benign - but would like to confirm with leviw and reland as NeedsRebaseline if we're both happy with them.


Original issue's description:
> Empty inline elements always get a line box.
> 
> CSS 2.1 (http://www.w3.org/TR/CSS2/visudet.html#leading) says: "Empty inline 
> elements generate empty inline boxes, but these boxes still have margins, padding, 
> borders and a line height, and thus influence [line-height calculation] just 
> like elements with content."
> 
> We weren't creating lineboxes for empty inline elements when they occurred
> between a single trailing space and a single leading space - this is because
> our linebox creation was based on the too-narrow premise that we were already
> ignoring spaces. However, if an empty inline element comes after a trailing 
> space but before one or more leading spaces then it will be ignored as part 
> of the collapsed space between the trailing space and the next non-collapsible 
> space - so we should force the creation of a linebox in such cases.
> 
> This allows us to pass Hixie's empty inline test: 
> http://www.hixie.ch/tests/evil/mixed/emptyinline.html
> 
> BUG=328939
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175205

TBR=leviw@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=328939

Review URL: https://codereview.chromium.org/302233002

git-svn-id: svn://svn.chromium.org/blink/trunk@175207 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL also moves CompositingReasonOverflowScrollingTouch inside
hasOverflowScrollTrigger because needsCompositedScrolling cannot return true if
we don't have the overflow scroll trigger.

Review URL: https://codereview.chromium.org/309593003

git-svn-id: svn://svn.chromium.org/blink/trunk@175206 bbb929c8-8fbe-4397-9dbb-9b2b20218538

CSS 2.1 (http://www.w3.org/TR/CSS2/visudet.html#leading) says: "Empty inline 
elements generate empty inline boxes, but these boxes still have margins, padding, 
borders and a line height, and thus influence [line-height calculation] just 
like elements with content."

We weren't creating lineboxes for empty inline elements when they occurred
between a single trailing space and a single leading space - this is because
our linebox creation was based on the too-narrow premise that we were already
ignoring spaces. However, if an empty inline element comes after a trailing 
space but before one or more leading spaces then it will be ignored as part 
of the collapsed space between the trailing space and the next non-collapsible 
space - so we should force the creation of a linebox in such cases.

This allows us to pass Hixie's empty inline test: 
http://www.hixie.ch/tests/evil/mixed/emptyinline.html

BUG=328939

Review URL: https://codereview.chromium.org/300853007

git-svn-id: svn://svn.chromium.org/blink/trunk@175205 bbb929c8-8fbe-4397-9dbb-9b2b20218538

No one reads this state.

Review URL: https://codereview.chromium.org/310543002

git-svn-id: svn://svn.chromium.org/blink/trunk@175204 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Synchronise releasing / default activating this feature with the other platforms.

BUG=378195
R=eae,dglazkov,eseidel

Review URL: https://codereview.chromium.org/304723002

git-svn-id: svn://svn.chromium.org/blink/trunk@175203 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Intent to Deprecate and Remove:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/Tu9pL7gMPyo/f-eYf_AkpQMJ

The conclusion of the thread was to deprecate only at this time.

Review URL: https://codereview.chromium.org/301403002

git-svn-id: svn://svn.chromium.org/blink/trunk@175202 bbb929c8-8fbe-4397-9dbb-9b2b20218538