GitLab

This also gets rid of a bug in clampScrollOffset that sometimes caused it to use a maximum offset that was one pixel too small, so a few tests need to be rebaselined.

BUG=492871
R=skobes@chromium.org

Review URL: https://codereview.chromium.org/1223173006

git-svn-id: svn://svn.chromium.org/blink/trunk@201983 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The old implementation was "allocate or die".  It's not good for
ImageData, which may try to allocate a huge memory block.  This CL
makes ImageData gracefully returns nullptr optionally raising an
exception in case of out of memory.

BUG=528898

Review URL: https://codereview.chromium.org/1328343002

git-svn-id: svn://svn.chromium.org/blink/trunk@201982 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Follows the same approach as BMPImageDecoder: Use a
FastSharedBufferReader to avoid consolidation.

Add an optional offset to PNGImageDecoder for ICOImageDecoder, so that
PNGImageDecoder can read a PNG subimage in an ICO without copying the
data. This eliminates another call to SharedBuffer::data(), as well as
a copy to create the separate SharedBuffer for PNG.

BUG=467772
BUG=528625

Review URL: https://codereview.chromium.org/1316203008

git-svn-id: svn://svn.chromium.org/blink/trunk@201981 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=529776

Review URL: https://codereview.chromium.org/1330163005

git-svn-id: svn://svn.chromium.org/blink/trunk@201980 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch introduces composed tree version of |startOfDocument()| by
templatizing original |startOfDocument()| to work both DOM tree position and
composed tree position version for introducing composed tree version of
|VisibleSelection| class.

This patch is a preparation of making selection to handle granularity for web
component, http://crrev.com/1277863002

BUG=513568
TEST=n/a; no behavior changes

Review URL: https://codereview.chromium.org/1324123003

git-svn-id: svn://svn.chromium.org/blink/trunk@201979 bbb929c8-8fbe-4397-9dbb-9b2b20218538

These APIs are intended to support spatial scheduling and task attribution.

Chromium implementation: https://codereview.chromium.org/1314903007/

BUG=510398

Review URL: https://codereview.chromium.org/1316233010

git-svn-id: svn://svn.chromium.org/blink/trunk@201978 bbb929c8-8fbe-4397-9dbb-9b2b20218538

V8 posts idle tasks explicitly.

BUG=chromium:490559

Review URL: https://codereview.chromium.org/1320073005

git-svn-id: svn://svn.chromium.org/blink/trunk@201977 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=529753
TBR=rbyers@chromium.org,eae@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/1312633007

git-svn-id: svn://svn.chromium.org/blink/trunk@201976 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Don't do base level clamp for generateMipmap. Generate error message for
generateMipmap if base level texture is not specified.

BUG=295792

Review URL: https://codereview.chromium.org/1306343007

git-svn-id: svn://svn.chromium.org/blink/trunk@201975 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We weren't properly performing the bypass check for inline style
attributes, nor were we correctly bypassing CSP for inline event
handlers. Of course, extensions shouldn't use either of these, but
they do. Because of course they do.

BUG=524356
R=jochen@chromium.org

Review URL: https://codereview.chromium.org/1318153009

git-svn-id: svn://svn.chromium.org/blink/trunk@201974 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The two users can be trivially replaced with queries on the layout tree
instead.

Review URL: https://codereview.chromium.org/1317563010

git-svn-id: svn://svn.chromium.org/blink/trunk@201973 bbb929c8-8fbe-4397-9dbb-9b2b20218538

- Fix virtual vs resident size reporting in the "size" column:
  Prior to this change the PartitionAlloc dumper was reporting the
  mmap-ed (virtual address spave) size in the "size" column.
  This is inconsistent with the rest of the dumpers, which strive
  to report an estimation of actually resident memory in the "size"
  column.
  This change moves the PA resident size into "size".   
- Remove the "thread_1234" sub-node: turns out that there can be at
  most one PartitionAlloc instance per-process, even in single-process
  mode.
- Improve reporting of allocated_objects:
  - Each bucket gets a column "allocated_objects_size", previously 
    called "active_size" (confusing), which reports the size of bytes
    requested by PA clients.
  - The per-bucket "allocated_objects" sub-rows are removed: they just
    added visual clutter and didn't add any extra information than an
    extra column.
  - A global allocated_objects node is introduced. This will be required
    by subsequent CLs to properly account memory reporter by other pieces
    of blink. This will be to avoid that WebCache memory that is 
    allocated via PA gets double-counted in the effective_size.
- Improve other column names for sake of clarity:
   - committed -> virtual_committed
   - num_{active,full,...} -> {active,full,...}_pages. The previous name
     didn't make it clear that they count pages and not objects.

BUG=488472

Review URL: https://codereview.chromium.org/1312843010

git-svn-id: svn://svn.chromium.org/blink/trunk@201972 bbb929c8-8fbe-4397-9dbb-9b2b20218538

See spec change: https://github.com/w3c/presentation-api/issues/184

BUG=529051

Review URL: https://codereview.chromium.org/1311003004

git-svn-id: svn://svn.chromium.org/blink/trunk@201971 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch expands two version of |endOfLine()| which changes behavior with
|mode| parameter into callers, one parameter version of |endOfLine()| and
|logicalEndOfLine()| to simplify implementation as a preparation for
templatizing |endOfLine()| and |logicalEndOfLine()| to introduce composed tree
version.

This patch is a preparation of making selection to handle granularity for web
component, http://crrev.com/1277863002

BUG=513568
TEST=n/a; no behavior changes

Review URL: https://codereview.chromium.org/1326403003

git-svn-id: svn://svn.chromium.org/blink/trunk@201970 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We should never get to a point where a `chrome-extension:` resource
shows up in the stack trace for a blocked resource. That said, we're
apparently hitting that scenario somewhat often.

This patch at least ensures that our brokenness isn't spamming reporting
endpoints, and revealing the extensions that users have installed to the
site owners.

BUG=524356

Review URL: https://codereview.chromium.org/1302363004

git-svn-id: svn://svn.chromium.org/blink/trunk@201969 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=eae@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/1326183005

git-svn-id: svn://svn.chromium.org/blink/trunk@201968 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch introduces composed tree version of |rightPositionOf()|
by templatizing original |rightPositionOf()| to work both DOM tree
position and composed tree position version for introducing composed tree
version of |VisibleSelection| class.

This patch also templaizes |rightVisuallyDistinctCandidate()| used by
|rightPositionOf()|.

This patch is a preparation of making selection to handle granularity for web
component, http://crrev.com/1277863002

BUG=513568
TEST=n/a; no behavior changes

Review URL: https://codereview.chromium.org/1311803004

git-svn-id: svn://svn.chromium.org/blink/trunk@201967 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Fold FormDataList::getEntry() and getAll() into DOMFormData.  So, no one uses
FormDataList::Entry.
This CL has no behavior changes.

BUG=528840

Review URL: https://codereview.chromium.org/1312333005

git-svn-id: svn://svn.chromium.org/blink/trunk@201966 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch introduces |ComposedTreeTraversal::previousPostOrder()| which is
composed tree version of |NodeTraversal::previousPostOrder()| as a preparation
of implementing composed tree version of |startOfParagraph()|.

This patch is a preparation of making selection to handle granularity for web
component, http://crrev.com/1277863002

BUG=513568
TEST=n/a; no behavior changes

Review URL: https://codereview.chromium.org/1308063008

git-svn-id: svn://svn.chromium.org/blink/trunk@201965 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We changed the behavior of 'Document::firstPartyForCookies' to return an
empty URL in the case where any URL in the ancestor chain doesn't match
the current document's URL. Previously, we'd simply return the top-level
document's URL.

This means that the content-settings bypass check in
'ContentSettingsObserver::IsWhitelistedForContentSettings' sees an empty
URL as opposed to a 'chrome-extension://' URL for cases in which an
extension loads a resource, so content settings are applied as per usual.
This breaks things fairly badly for folks who have turned on third-party
cookie blocking.

In these patches, we introduce a new scheme registry for those schemes
which ought to override Blink's concept of "first-party" when they're
loaded into the top-level browsing context.

Patch 1 (Blink): [This patch]
Patch 2 (Chromium): https://codereview.chromium.org/1332563006

BUG=527963
R=jochen@chromium.org

Review URL: https://codereview.chromium.org/1305253012

git-svn-id: svn://svn.chromium.org/blink/trunk@201964 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch introduces composed tree version of |previousPositionOf()|
by templatizing original |previousPositionOf()| to work both DOM tree
position and composed tree position version for introducing composed tree
version of |VisibleSelection| class.

This patch also templatizes |skipToStartOfEditingBoundary()| used by
|previousPositionOf()|.

This patch is a preparation of making selection to handle granularity for web
component, http://crrev.com/1277863002

BUG=513568
TEST=n/a; no behavior changes

Review URL: https://codereview.chromium.org/1312473017

git-svn-id: svn://svn.chromium.org/blink/trunk@201963 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This fixes an assertion failure in ~PickerIndicatorElement.

We need to do rebaseline of many tests because these tests opened pickers
without focus.  openPicker() force to focus on <input> elements now.

We confirmed Win10 failures were trivial.  They were due to month
format change and minor font glitch.

BUG=529530

Review URL: https://codereview.chromium.org/1311013012

git-svn-id: svn://svn.chromium.org/blink/trunk@201962 bbb929c8-8fbe-4397-9dbb-9b2b20218538

6 tests have super-minor differences in pixel color. Probably due to the
roll in https://codereview.chromium.org/1315893005

TBR=dalecurtis@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/1316663006

git-svn-id: svn://svn.chromium.org/blink/trunk@201961 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Add a FOUC painting test. (patchset #6 id:100001 of https://codereview.chromium.org/1329553004/ )

Reason for revert:
This patch is failing on debug bots, all platforms. For example:

* https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20%28dbg%29/builds/3244

* https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.7%20%28dbg%29/builds/24575

Original issue's description:
> Add a FOUC painting test.
> 
> This expands the Sim(ulation) testing framework to support a very rudimentary
> compositor that can run a simple BeginMainFrame which runs the WebView's
> beginFrame() and layout() steps followed by a simplified paint step which paints
> the main layer of every composited layer mapping. This is a very rough
> approximation of what the real compositor would do, but is enough to test the
> code for avoiding FOUC in DeprecatedPaintLayerPainter's shouldSuppressPaintingLayer
> and BlockPainter::paintContents. It's also enough to test the paint invalidation
> logic in Document::styleResolverChanged for when we painted by skipping painting
> the actual content because of the previously mentioned FOUC avoidance logic.
> 
> I also went and added a bunch of comments to the Sim* classes to explain what
> they do and how to use the testing framework. A future patch may add a README
> so we can use this framework to write future pipeline tests.
> 
> BUG=521692
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=201944

TBR=dglazkov@chromium.org,pdr@chromium.org,esprehn@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=521692

Review URL: https://codereview.chromium.org/1316673009

git-svn-id: svn://svn.chromium.org/blink/trunk@201960 bbb929c8-8fbe-4397-9dbb-9b2b20218538

As specified in [1], the 'frame-ancestors' CSP directive should take
control of the access checks when loading a document. In particular,
the 'x-frame-options' header should be ignored if a 'frame-ancestors'
directive is present and enforced.

[1]: https://w3c.github.io/webappsec/specs/content-security-policy/#frame-ancestors-and-frame-options

BUG=510423
R=estark@chromium.org

Review URL: https://codereview.chromium.org/1326823003

git-svn-id: svn://svn.chromium.org/blink/trunk@201959 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch introduces composed tree version of |endOfDocument()| by
templatizing original |endOfDocument()| to work both DOM tree position and
composed tree position version for introducing composed tree version of
|VisibleSelection| class.

This patch is a preparation of making selection to handle granularity for web
component, http://crrev.com/1277863002

BUG=513568
TEST=n/a; no behavior changes

Review URL: https://codereview.chromium.org/1326683006

git-svn-id: svn://svn.chromium.org/blink/trunk@201958 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Fold FormDataList::getEntry(index) into DOMFormData without creating
FormDataList::Entry.
 - Add Item::file() to share code between itemsToEntry() and
   DOMFormDataIterationSource::next().
 - Add Item::isString() and isFile() for compatibility with Entry.
 - Add DOMFormData::decode() to decode CStrings in Item.

This CL has no behavior changes.

BUG=528840

Review URL: https://codereview.chromium.org/1334513002

git-svn-id: svn://svn.chromium.org/blink/trunk@201957 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Because CSSParserTokens depends on PartitionAlloc.

BUG=501171

Review URL: https://codereview.chromium.org/1333583003

git-svn-id: svn://svn.chromium.org/blink/trunk@201956 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch introduces composed tree version of
|isEndOfEditableOrNonEditableContent()| by templatizing original
|isEndOfEditableOrNonEditableContent()| to work both DOM tree position and
composed tree position version for introducing composed tree version of
|VisibleSelection| class.

This patch is a preparation of making selection to handle granularity for web
component, http://crrev.com/1277863002

BUG=513568
TEST=n/a; no behavior changes

Review URL: https://codereview.chromium.org/1333583002

git-svn-id: svn://svn.chromium.org/blink/trunk@201955 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch introduces composed tree version of |leftPositionOf()|
by templatizing original |leftPositionOf()| to work both DOM tree
position and composed tree position version for introducing composed tree
version of |VisibleSelection| class.

This patch also introduces |createVisiblePosition()| for
|PositionInComposdTreeWithAffinity| and templaizes
|honorEditingBoundaryAtOrAfter()| and |leftVisuallyDistinctCandidate()| used
by |leftPositionOf()|.

This patch is a preparation of making selection to handle granularity for web
component, http://crrev.com/1277863002

BUG=513568
TEST=n/a; no behavior changes

Review URL: https://codereview.chromium.org/1326313002

git-svn-id: svn://svn.chromium.org/blink/trunk@201954 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The XHR specification defines 'FormData' interface [1], but the blink::FormData
is not its implementation.  It's confusing.
A later patch will rename DOMFormData to FormData.

This CL has no behavior changes.

[1] https://xhr.spec.whatwg.org/#interface-formdata

BUG=528840

Review URL: https://codereview.chromium.org/1311923004

git-svn-id: svn://svn.chromium.org/blink/trunk@201953 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is a speculative fix for crbug.com/515899.  If a JavaScript dialog is
opened while a SELECT popup is shown, and the SELECT popup is hidden before the
dialog is hidden, window focus is not correctly restored.  So, we guess delaying
change events solves the problem.

BUG=515899

Review URL: https://codereview.chromium.org/1321993005

git-svn-id: svn://svn.chromium.org/blink/trunk@201952 bbb929c8-8fbe-4397-9dbb-9b2b20218538

- Move FormDataList::deleteEntry() to DOMFormData
- Fold FormDataList::hasEntry() into DOMFormData::has().

This CL has no behavior changes.

BUG=528840

Review URL: https://codereview.chromium.org/1308183006

git-svn-id: svn://svn.chromium.org/blink/trunk@201951 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=201938

BUG=521124
TBR=eae@chromium.org

Review URL: https://codereview.chromium.org/1307203008 .

git-svn-id: svn://svn.chromium.org/blink/trunk@201950 bbb929c8-8fbe-4397-9dbb-9b2b20218538

See https://code.google.com/p/chromium/issues/detail?id=452745#c17 for details.

BUG=452745

Review URL: https://codereview.chromium.org/1300083002

git-svn-id: svn://svn.chromium.org/blink/trunk@201949 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=none

Review URL: https://codereview.chromium.org/1328273003

git-svn-id: svn://svn.chromium.org/blink/trunk@201947 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=201930

BUG=237050
TBR=eae@chromium.org

Review URL: https://codereview.chromium.org/1314943011 .

git-svn-id: svn://svn.chromium.org/blink/trunk@201946 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We create ScrollingCoordinator for pages with non-local frames, but it's noop in functionality because of guards:

if (!m_page->mainFrame()->isLocalFrame())
  return;

Let's align the behavior with the rest of ScrollingCoordinator code. Otherwise it crashes browser tests.

BUG=394777

Review URL: https://codereview.chromium.org/1304213003

git-svn-id: svn://svn.chromium.org/blink/trunk@201945 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This expands the Sim(ulation) testing framework to support a very rudimentary
compositor that can run a simple BeginMainFrame which runs the WebView's
beginFrame() and layout() steps followed by a simplified paint step which paints
the main layer of every composited layer mapping. This is a very rough
approximation of what the real compositor would do, but is enough to test the
code for avoiding FOUC in DeprecatedPaintLayerPainter's shouldSuppressPaintingLayer
and BlockPainter::paintContents. It's also enough to test the paint invalidation
logic in Document::styleResolverChanged for when we painted by skipping painting
the actual content because of the previously mentioned FOUC avoidance logic.

I also went and added a bunch of comments to the Sim* classes to explain what
they do and how to use the testing framework. A future patch may add a README
so we can use this framework to write future pipeline tests.

BUG=521692

Review URL: https://codereview.chromium.org/1329553004

git-svn-id: svn://svn.chromium.org/blink/trunk@201944 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=474759

Review URL: https://codereview.chromium.org/1328293002

git-svn-id: svn://svn.chromium.org/blink/trunk@201943 bbb929c8-8fbe-4397-9dbb-9b2b20218538