GitLab

When an SVG has either intrinsic width or height but not both, and a
intrinsic ratio, the missing intrinsic width or height can be
calculated. Doing so enables us to return correct intrinsic width and
height for SVG images in more cases when the image has no renderer.

Also, this patch removes the unspecified use of the viewBox attribute
as the source of intrinsic size. A small number of tests depended on
this and has been updated. The reasoning to make this change is
because we don't use the viewBox for intrinsic size in any other
context than SVGImage. And we only use it until the renderer has a
containerSize() anyway.

http://www.whatwg.org/specs/web-apps/current-work/multipage/edits.html#dom-img-width
http://www.whatwg.org/specs/web-apps/current-work/multipage/edits.html#dom-img-height

BUG=376895

Review URL: https://codereview.chromium.org/308643002

git-svn-id: svn://svn.chromium.org/blink/trunk@175784 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Should check whether document.settings() is available or not when updating GenericFontFamilySettings.

Looking at crash reports - WebCore::StyleEngine::updateGenericFontFamilySettings, document().settings() is probably null. In the case, we should not update GenericFontSettings.

BUG=376525
TEST=It is very hard to reproduce crashes reported in crbug.com/376525. Will check # of crash reports to see whether the bug is really fixed or not.

Review URL: https://codereview.chromium.org/304033005

git-svn-id: svn://svn.chromium.org/blink/trunk@175783 bbb929c8-8fbe-4397-9dbb-9b2b20218538

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

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=174956

Review URL: https://codereview.chromium.org/267323004

git-svn-id: svn://svn.chromium.org/blink/trunk@175782 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Also add test for frame model. Note there are still discrepancies wrt
how time per category is computed for timeline records -- these will
be the subject of a separate patch.

BUG=361045

Review URL: https://codereview.chromium.org/320523004

git-svn-id: svn://svn.chromium.org/blink/trunk@175781 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/319603007

Review URL: https://codereview.chromium.org/319603007

git-svn-id: svn://svn.chromium.org/blink/trunk@175780 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch fixes a long-standing editing bug that a Range object returned by
Selection.getRangeAt() may not be equivalent to the Range used to set selection
with Selection.collapse().

FrameSelection.setSelectedRange sets a Range without canonicalization:
https://src.chromium.org/viewvc/blink?revision=169100&view=revision
Hence this patch adopts it.

Following is the implementation detail.
DOMSelection.cpp:
- When selection.collapse(node, offset) is called, we create a caret Range corresponding to node and offset and pass it to FrameSelection.setSelectedRange.

FrameSelection.cpp:
- If new Selection is collapsed, use old isDirectional. This is following the current behavior.

LayoutTests:
editing/selection/no-range-canonicalization.html
- The main test added to current setRange test.

fast/css/counters/counter-before-content-not-incremented.html
- Some tests expect the collapse method to canonicalize Selection so we need to set a exact DOM position.

editing/editing.js
- Some tests expect the collapse method to canonicalize Selection at start so I canonicalize that as the old Blink collapse implementation.

editing/selection/DOMSelection-DocumentType.html
- The current HTML spec says that if Range is set DocumentType node, we should throw type exception. 

editing/selection/selection-invalid-offset.html
- We should throw exception when the method is called with a offset larger than length of a node.
- Remove a redundant case.


BUG=346613

Review URL: https://codereview.chromium.org/225303002

git-svn-id: svn://svn.chromium.org/blink/trunk@175779 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The workaround seems specific to JSC, and thus not needed in Blink.

BUG=None

Review URL: https://codereview.chromium.org/293983020

git-svn-id: svn://svn.chromium.org/blink/trunk@175778 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The following chromium commits fixed chromium code to use  new
public web methods which take creation context and isolate:

WebBlob
- https://src.chromium.org/viewvc/chrome?view=rev&revision=275420

WebDOMError
- https://src.chromium.org/viewvc/chrome?view=rev&revision=275396

WebDOMFileSystem:
- https://src.chromium.org/viewvc/chrome?view=rev&revision=275291

BUG=369451

Review URL: https://codereview.chromium.org/326573002

git-svn-id: svn://svn.chromium.org/blink/trunk@175777 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We don't have to hold the FormData instance to be used by createRequest()
as a class member. It's instantiated right before createRequest() call.
We can just allocate it on stack.

BUG=none
R=ricea,tkent

Review URL: https://codereview.chromium.org/314283005

git-svn-id: svn://svn.chromium.org/blink/trunk@175776 bbb929c8-8fbe-4397-9dbb-9b2b20218538

To apply XSLStyleSheet, we only need the first one. We don't need
to have all XSL stylesheets if the XML document has more than two
XSL stylesheets. So modified not to add XSLStyleSheet to
StyleSheetCandidateNodes. Instead, modified to make StyleEngine have
the first XSLStyleSheet.

BUG=367689
TEST=No new test, because this patch is just for refactoring.

Review URL: https://codereview.chromium.org/257363002

git-svn-id: svn://svn.chromium.org/blink/trunk@175775 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Before this patch, |Heap::findGCInfo()|'s iteration on
AttachedThreadStateSet was not guarded. This triggered "modification
while iteration" assert when |Heap::findGCInfo()| was called from
non-safepoint.

This patch exposes |ThreadState::threadAttachMutex()| and
|Heap::findGCInfo()| will acquire lock on it when we are not in safepoint.

Unfortunately there are no trybot which has an ENABLE(GC_TRACING) flag.
NOTRY=true
TEST=content_shell --dump-render-tree http/tests/workers/terminate-during-sync-operation-websql.html wont crash.
BUG=380474

Review URL: https://codereview.chromium.org/311263008

git-svn-id: svn://svn.chromium.org/blink/trunk@175774 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Gardening: mark the test as flaky.

BUG=338794, 192172
TBR=dcheng@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/318343002

git-svn-id: svn://svn.chromium.org/blink/trunk@175773 bbb929c8-8fbe-4397-9dbb-9b2b20218538

There seems to be some cases where a SVGSVGElement wrapper is gc-ed while the SVGSVGElement is alive.
However, |m_transform| and other properties held by SVGViewSpec manage its lifetime via the SVGSVGElement v8 wrapper.

This patch avoids the situation by keeping the context SVGSVGElement wrapper alive from SVGViewSpec wrapper.

BUG=379998

Review URL: https://codereview.chromium.org/315893002

git-svn-id: svn://svn.chromium.org/blink/trunk@175772 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Scheme has been compared in case-sensitive manner. However, according to spec, scheme must be compared in an ASCII case-insensitive manner by user agents.

Merge from WebKit : http://trac.webkit.org/changeset/169670

BUG=none

Review URL: https://codereview.chromium.org/323493003

git-svn-id: svn://svn.chromium.org/blink/trunk@175771 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=380075

Review URL: https://codereview.chromium.org/312253002

git-svn-id: svn://svn.chromium.org/blink/trunk@175770 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=340522

Review URL: https://codereview.chromium.org/319863003

git-svn-id: svn://svn.chromium.org/blink/trunk@175769 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=isherman@chromium.org, tkent@chromium.org
BUG=380603

Review URL: https://codereview.chromium.org/319443003

git-svn-id: svn://svn.chromium.org/blink/trunk@175768 bbb929c8-8fbe-4397-9dbb-9b2b20218538

loadRequest() was calling load() before
http://trac.webkit.org/changeset/129462 but no longer calling it. The
same about revalidateResource(). Give them new names describing what's
done in them.

Also, this CL makes some style fix in them.

BUG=none
R=japhet

Review URL: https://codereview.chromium.org/318003002

git-svn-id: svn://svn.chromium.org/blink/trunk@175767 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL fixed the failure.
https://codereview.chromium.org/319083003/

BUG=347109
R=ricea

Review URL: https://codereview.chromium.org/319873004

git-svn-id: svn://svn.chromium.org/blink/trunk@175766 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=none

Review URL: https://codereview.chromium.org/316123003

git-svn-id: svn://svn.chromium.org/blink/trunk@175764 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Objects in XMLDocumentParser::m_currentNodeStack are correctly traces in
Oilpan. We don't need ref/deref.

Also, this CL replace one instance of Vector<Element*> with
HeapVector<Member<Elemenet>>.

BUG=357163

Review URL: https://codereview.chromium.org/320773002

git-svn-id: svn://svn.chromium.org/blink/trunk@175763 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=357163

Review URL: https://codereview.chromium.org/320233002

git-svn-id: svn://svn.chromium.org/blink/trunk@175762 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

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175621

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

git-svn-id: svn://svn.chromium.org/blink/trunk@175761 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL folds m_compositingLayersNeedRebuild into m_updateType, which lets us
remove a bunch of parallel bookkeeping.

BUG=332248

Review URL: https://codereview.chromium.org/320103005

git-svn-id: svn://svn.chromium.org/blink/trunk@175760 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=dpranke@chromium.org
BUG=369385

Review URL: https://codereview.chromium.org/319423002

git-svn-id: svn://svn.chromium.org/blink/trunk@175759 bbb929c8-8fbe-4397-9dbb-9b2b20218538

While re-reading the code, I noticed that some of ImageCandidate's member variables
are not initialized as they should.
This CL fixes that.

BUG=

Review URL: https://codereview.chromium.org/311053011

git-svn-id: svn://svn.chromium.org/blink/trunk@175758 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch schedules console viewport refresh on console view resize
event.

BUG=381547

Review URL: https://codereview.chromium.org/322723002

git-svn-id: svn://svn.chromium.org/blink/trunk@175757 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch adds test based on the regression issue with styles sidebar
pane not showing styles for ::contet selectors.

BUG=380414

Review URL: https://codereview.chromium.org/316003002

git-svn-id: svn://svn.chromium.org/blink/trunk@175756 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Reason for revert:
Causes several browser tests to fail on ChromiumOS (dbg):
http://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Tests%20%28dbg%29%281%29/builds/30876

Original issue's description:
> Remove scoped styles.
> 
> Removes <style scoped>.
> 
> This was discussed and approved with LGTMs in blink-dev:
> https://groups.google.com/a/chromium.org/d/msg/blink-dev/R1x18ZLS5qQ/Bjuh_cENhlQJ
> 
> Chrome status dashboard entry:
> http://www.chromestatus.com/features/5374137958662144
> 
> Note:
> this is based on esprehn's original cl:
> https://codereview.chromium.org/214693002/
> 
> BUG=379096
> TEST=pass all layout tests
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175555
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175720

TBR=esprehn@chromium.org,tasak@google.com,kochi@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=379096

Review URL: https://codereview.chromium.org/324593003

git-svn-id: svn://svn.chromium.org/blink/trunk@175754 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Rather than using a different update type, we can just mark the root as needing
its properties updated, which will cause us to update the entire tree. This CL
alone isn't a huge win, but it will let us remove
RenderLayerCompositor::m_compositingLayersNeedRebuild in the next CL.

Also, move the call to enableCompositingModeIfNeeded into the new
RenderLayerCompositor::didLayout function. This makes the sequence look a bit
silly (because we mark the dirty bit only to clear it immediately), but that's
what the code did before (albeit in a more roundabout way).

BUG=332248

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175738

Review URL: https://codereview.chromium.org/323653002

git-svn-id: svn://svn.chromium.org/blink/trunk@175752 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Reason for revert:
Triggering buffer overflow checks on Linux ASAN:
http://build.chromium.org/p/chromium.memory/buildstatus?builder=Linux%20ASan%20Tests%20%28sandboxed%29&number=2348

Original issue's description:
> Initial ServiceWorker Cache API polyfill.
> 
> A basic Cache API. No serialization of any kind, no testing, and no
> add() method, as it depends on Fetch().
> 
> R=falken@chromium.org,jkarlin@chromium.org
> TBR=jochen@chromium.org
> BUG=374822
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175749

TBR=falken@chromium.org,jkarlin@chromium.org,jsbell@chromium.org,jochen@chromium.org,gavinp@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=374822

Review URL: https://codereview.chromium.org/319383003

git-svn-id: svn://svn.chromium.org/blink/trunk@175751 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This reverts commit ebab27e4fe728206d3be9a54fd66ca99de12a298.

Revert "Oilpan: fix build after r175732."

This reverts commit 1838d888958380a6e51de4d677665bd204bd5134.

Breaks plugins/document-open.html

TBR=esprehn@chromium.org

Review URL: https://codereview.chromium.org/318263002

git-svn-id: svn://svn.chromium.org/blink/trunk@175750 bbb929c8-8fbe-4397-9dbb-9b2b20218538

A basic Cache API. No serialization of any kind, no testing, and no
add() method, as it depends on Fetch().

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

Review URL: https://codereview.chromium.org/314133002

git-svn-id: svn://svn.chromium.org/blink/trunk@175749 bbb929c8-8fbe-4397-9dbb-9b2b20218538

By convention, each virtual test suite should have a README.txt documenting
what directory it mirrors and what command line args are needed, so that
you can tell what an "empty" directory under LayoutTests/ is for.

This patch modifies lint-test-expectations to check that those READMEs
exist, and will help avoid the possibility that we'll delete whole directories
thinking they are unneeded and introduce weird errors when TestExpectations
still refer to nonexistent directories or files.

R=ojan@chromium.org
BUG=369385

Review URL: https://codereview.chromium.org/267803007

git-svn-id: svn://svn.chromium.org/blink/trunk@175748 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Revert of Oilpan: Prepare to make ExecutionContext GarbageCollectedMixin. (https://codereview.chromium.org/319333002/)

Reason for revert:
This was an incorrect revert--this patch was not responsible for the crash.

Original issue's description:
> Revert of Oilpan: Prepare to make ExecutionContext GarbageCollectedMixin. (https://codereview.chromium.org/318023002/)
> 
> Reason for revert:
> plugins/document-open.html is crashing on debug layout bots
> 
> Original issue's description:
> > Oilpan: Prepare to make ExecutionContext GarbageCollectedMixin.
> > 
> > We creates closures with ExecutionContexts in LocalFileSystem. The
> > ExecutionContext needs to be a CrossThreadPersistent data member of
> > closure. This CL adds ParamStorageTraits specialized to garbage-collected
> > classes.
> > 
> > This is a preparation to remove RefCounted from Node.
> > 
> > BUG=357163
> > 
> > Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175734
> 
> TBR=oilpan-reviews@chromium.org,haraken@chromium.org,tkent@chromium.org
> NOTREECHECKS=true
> NOTRY=true
> BUG=357163
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175741

TBR=oilpan-reviews@chromium.org,haraken@chromium.org,tkent@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=357163

Review URL: https://codereview.chromium.org/319373002

git-svn-id: svn://svn.chromium.org/blink/trunk@175747 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Timeline uses time in milliseconds while trace event use microseconds. Do the conversion earlier.

BUG=361045

Review URL: https://codereview.chromium.org/318093002

git-svn-id: svn://svn.chromium.org/blink/trunk@175746 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Restore equality test back to being between Node* and a RawPtr<Node>,
for which there is an implicit conversion from a RawPtr<Node> to a
Node*. i.e., a RawPtr<Node> cannot be implicitly converted to a Node&
(and vice versa.)

TBR=haraken@chromium.org,oilpan-reviews@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/325573002

git-svn-id: svn://svn.chromium.org/blink/trunk@175745 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/317733008

git-svn-id: svn://svn.chromium.org/blink/trunk@175743 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Reason for revert:
Triggers ASSERT when running browser_test CastStreamingApiTestWithPixelOutput.RtpStreamError

Original issue's description:
> Delete CompositingUpdateAfterLayout
> 
> Rather than using a different update type, we can just mark the root as needing
> its properties updated, which will cause us to update the entire tree. This CL
> alone isn't a huge win, but it will let us remove
> RenderLayerCompositor::m_compositingLayersNeedRebuild in the next CL.
> 
> Also, move the call to enableCompositingModeIfNeeded into the new
> RenderLayerCompositor::didLayout function. This makes the sequence look a bit
> silly (because we mark the dirty bit only to clear it immediately), but that's
> what the code did before (albeit in a more roundabout way).
> 
> BUG=332248
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175738

TBR=esprehn@chromium.org,abarth@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=332248

Review URL: https://codereview.chromium.org/321663002

git-svn-id: svn://svn.chromium.org/blink/trunk@175742 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Oilpan: Prepare to make ExecutionContext GarbageCollectedMixin. (https://codereview.chromium.org/318023002/)

Reason for revert:
plugins/document-open.html is crashing on debug layout bots

Original issue's description:
> Oilpan: Prepare to make ExecutionContext GarbageCollectedMixin.
> 
> We creates closures with ExecutionContexts in LocalFileSystem. The
> ExecutionContext needs to be a CrossThreadPersistent data member of
> closure. This CL adds ParamStorageTraits specialized to garbage-collected
> classes.
> 
> This is a preparation to remove RefCounted from Node.
> 
> BUG=357163
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175734

TBR=oilpan-reviews@chromium.org,haraken@chromium.org,tkent@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=357163

Review URL: https://codereview.chromium.org/319333002

git-svn-id: svn://svn.chromium.org/blink/trunk@175741 bbb929c8-8fbe-4397-9dbb-9b2b20218538