GitLab

This method is no longer being called, and can be removed.

BUG=250049
TEST=encrypted media layout and browser tests pass

Review URL: https://codereview.chromium.org/197333006

git-svn-id: svn://svn.chromium.org/blink/trunk@169999 bbb929c8-8fbe-4397-9dbb-9b2b20218538

EventInterfaces.in should depend on event IDL files.
It currently doesn't, which means it fails to rebuild correctly.
This fixes it.

TBR=haraken
BUG=341748

Review URL: https://codereview.chromium.org/209723002

git-svn-id: svn://svn.chromium.org/blink/trunk@169998 bbb929c8-8fbe-4397-9dbb-9b2b20218538

All page-level state changes now go through updatePageInfo, which marks blocks
as needing layout if their multipliers might need updating.

This simplifies the flow and ensures the page updates correctly on orientation
changes and movements of the text scaling slider in accessibility settings.

BUG=353309,347384

Review URL: https://codereview.chromium.org/209353003

git-svn-id: svn://svn.chromium.org/blink/trunk@169997 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Web SQL Database is an obsolete specification.
We already count Web SQL Database usage in the DOMWindow context.

BUG=

Review URL: https://codereview.chromium.org/207823004

git-svn-id: svn://svn.chromium.org/blink/trunk@169996 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=278596

Review URL: https://codereview.chromium.org/208743007

git-svn-id: svn://svn.chromium.org/blink/trunk@169995 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This feature makes the insertedInto tree walk much more complicated and
is the source of security bugs. We've had it disabled behind a flag for a
while now and not seen complaints or bugs from users.

BUG=333318

Review URL: https://codereview.chromium.org/205523003

git-svn-id: svn://svn.chromium.org/blink/trunk@169994 bbb929c8-8fbe-4397-9dbb-9b2b20218538

There's no evidence that this fast path actually matters, the only difference 
seems to be the number of if and switch statements. If it does we should look 
into how we can make the regular SelectorChecker faster instead of having two modes.

Review URL: https://codereview.chromium.org/208423010

git-svn-id: svn://svn.chromium.org/blink/trunk@169993 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL moves the compositorDrivenAcceleratedScrollingEnabled check out of
scrollParent so that it's clearer in the callers that this code only does
something when this setting is enabled.

In compositorDrivenAcceleratedScrollingEnabled, this lets us call
compositorDrivenAcceleratedScrollingEnabled once instead of twice, which is
faster. In a future CL, I'll cache compositorDrivenAcceleratedScrollingEnabled
as a compositing trigger, which will make this code actually go fast.

This CL also reorders a bunch of methods in CompositingReasonFinder.cpp to make
it clearer which of these functions are called during style recalc and which
are call unconditionally on all RenderLayers.

Review URL: https://codereview.chromium.org/205343009

git-svn-id: svn://svn.chromium.org/blink/trunk@169992 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently the WebCrypto API only has support for one-shot digests. This CL
provides the necessary API scaffolding for methods to generate digests one chunk
at a time.

BUG=349514

Review URL: https://codereview.chromium.org/195983024

git-svn-id: svn://svn.chromium.org/blink/trunk@169991 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Also rename StyleInvalidationAnalysis to StyleSheetInvalidationAnalysis, to distinguish from StyleInvalidator.

BUG=
R=ojan@chromium.org, rune@opera.com

Review URL: https://codereview.chromium.org/199633020

git-svn-id: svn://svn.chromium.org/blink/trunk@169990 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Without this CL, a SQLError object can be constructed in a database thread,
moved to main/worker thread, and referred by JavaScript object in the
main/worker thread.

We can't move object ownership over threads in Oilpan. So, we should not
construct SQLError objects in a database thread. Unlike the SQLResultSet case
[1], SQLError is created by multiple classes and it's hard to make sure they
create SQLError in a context thread.

So, this CL introduce SQLErrorData, which is an immutable data container. We can
construct a SQLErrorData object in a database thread, and move it to a context
thread, then construct a SQLError with the SQLErrorData object.

Note:
* Change the return type of AbstractSQLStatementBackend::sqlError from
PassRefPtr<SQLError> to SQLErrorData* because the function doesn't release the
ownership of the SQLErrorData object.

[1] https://src.chromium.org/viewvc/blink?revision=169929&view=revision

BUG=347902

Review URL: https://codereview.chromium.org/210833005

git-svn-id: svn://svn.chromium.org/blink/trunk@169989 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=169714

BUG=354688
TBR=abarth@chromium.org

Review URL: https://codereview.chromium.org/211573006

git-svn-id: svn://svn.chromium.org/blink/trunk@169988 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=169756

BUG=344497,352628
TBR=fmalita@chromium.org

Review URL: https://codereview.chromium.org/199323010

git-svn-id: svn://svn.chromium.org/blink/trunk@169987 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=
NOTRY=true

Review URL: https://codereview.chromium.org/209043006

git-svn-id: svn://svn.chromium.org/blink/trunk@169986 bbb929c8-8fbe-4397-9dbb-9b2b20218538

All touch based state is set via gesture events so there is no need for these tests.

BUG=325930

Review URL: https://codereview.chromium.org/211423002

git-svn-id: svn://svn.chromium.org/blink/trunk@169985 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently in RenderBox::styleWillChange we will issue a view()->repaint() if the
node is the HTML or body element and the diff >= StyleDifferenceRepaint. For
repaint-after-layout, if this is StyleDifferenceLayout, we will do the layout
and end up issuing the same invalidation again.

This CL makes the view()->repaint() call conditional on either
  1) repaint-after-layout is disabled
  2) this is not a StyleDifferenceLayout.

BUG=320139

Review URL: https://codereview.chromium.org/211613002

git-svn-id: svn://svn.chromium.org/blink/trunk@169984 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The argument for these functions have been replaced by HTMLFormControlElement. Since changes at
both chromium and blink sides are landed, it's safe to remove these functions.

R=isherman@chromium.org, tkent@chromium.org
BUG=332557

Review URL: https://codereview.chromium.org/203263004

git-svn-id: svn://svn.chromium.org/blink/trunk@169983 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Previously, no Blink platform was initialized, leading to a segfault
when a trace event occurred in NativeImageSkia::draw.

This was fixed by making the platform unit test runner provide a
minimal platform by default. The gyp target was adjusted to match the
way blink_heap_unittests, for example, sets up its environment.

This seems like a reasonable compromise between having a null platform
and having a more complete platform implemented on the Chromium side,
like the web unit tests have.

BUG=353585

Review URL: https://codereview.chromium.org/206473009

git-svn-id: svn://svn.chromium.org/blink/trunk@169982 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=345027
R=japhet@chromium.org

Review URL: https://codereview.chromium.org/211903002

git-svn-id: svn://svn.chromium.org/blink/trunk@169981 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The spec requires a minimum range of 22.05-96 kHz. (http://webaudio.github.io/web-audio-api/#widl-AudioContext-createBuffer-AudioBuffer-unsigned-long-numberOfChannels-unsigned-long-length-float-sampleRate)

Add two new tests to verify that 3 kHz and 192 kHz are allowed and
update expected test results for allowed range.

BUG=344375

Review URL: https://codereview.chromium.org/201673004

git-svn-id: svn://svn.chromium.org/blink/trunk@169980 bbb929c8-8fbe-4397-9dbb-9b2b20218538

As it turns out, we only needed to patch the Chrome side.

> Prevent web content from forging File entries in drag and drop.
> 
> There are two separate bugs that this and the corresponding Chrome patch
> aim to address:
> - On Linux, files and URLs are transferred in the same MIME type, so
>   it's impossible to tell if a filename was set by a trusted source or
>   forged by web content.
> - DownloadURL triggers the download of potentially cross-origin content.
>   On some platforms, such as Windows, the resulting download is treated
>   as a file drag by Chrome, allowing web content to read cross origin
>   content.
> 
> In order to prevent web content from doing this, drags initiated by a
> renderer will be marked as tainted. When tainted drags are over web
> content, Blink will only allow the resulting filename to be used for
> navigation, with Chrome enforcing this with the sandbox policy.
> 
> Unfortunately, this does break some potentially interesting use cases
> like being able to drag an attachment from Gmail to a file input, but
> those will have to be separately addressed, if possible.
> 
> BUG=346135
> R=abarth@chromium.org, tony@chromium.org
> 
> Review URL: https://codereview.chromium.org/193803002

TBR=dcheng@chromium.org

Review URL: https://codereview.chromium.org/211853002

git-svn-id: svn://svn.chromium.org/blink/trunk@169979 bbb929c8-8fbe-4397-9dbb-9b2b20218538

V8 now hosts the end-of-microtask work queue. This allows consistent
scheduling of microtask work regardless of where it is implemented (v8 vs blink).

R=adamk@chromium.org

Review URL: https://codereview.chromium.org/167683003

git-svn-id: svn://svn.chromium.org/blink/trunk@169978 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/211203003

git-svn-id: svn://svn.chromium.org/blink/trunk@169977 bbb929c8-8fbe-4397-9dbb-9b2b20218538

[Mac] Rely on science, not chance, to get the proper definition of the macro NSGEOMETRY_TYPES_SAME_AS_CGGEOMETRY_TYPES.

NSGEOMETRY_TYPES_SAME_AS_CGGEOMETRY_TYPES is used in
{Int,Float}{Point,Size,Rect}.h to determine whether to provide distinct
conversion constructors and operators for Foundation NS{Point,Size,Rect} types
in addition to the CoreGraphics CG{Point,Size,Rect} types. When this macro is
defined (as in 64-bit Mac OS X), the NS and CG types are identical and no
additional overloads are needed. When not defined (as in 32-bit Mac OS X), the
types are distinct and additional overrides are needed for the NS types.

This macro is set (or not) by <Foundation/NSGeometry.h>. It is an error to
consider this macro without first #importing this header or its umbrella,
<Foundation/Foundation.h> or <Cocoa/Cocoa.h>. It is also an error to #import
these headers from non-Objective-C code, but in that case, it is impossible to
correctly use the NS types. The net effect is that the overloads for the
distinct NS types will available in Objective-C translation units when the
types are in fact distinct from the CG ones, and not available in
non-Objective-C translation units where they wouldn't be useful anyway, even
if the types are distinct. Where the types are not distinct, the conversion
constructor and operator for CG types is able to handle NS ones as well.

This backs out a previous hack fix for this problem, Blink r169826.

BUG=355576
TEST=compile, both 32-bit and 64-bit

Review URL: https://codereview.chromium.org/208333020

git-svn-id: svn://svn.chromium.org/blink/trunk@169976 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Line height is cached in the RenderBlock; when FTA updates a multiplier it needs to force recomputation of the line height.

(The old autosizer avoids this issue by using setStyle instead of setStyleInternal.  This triggers style recalc, which invalidates the line height.)

BUG=353309,354032

Review URL: https://codereview.chromium.org/209853007

git-svn-id: svn://svn.chromium.org/blink/trunk@169975 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=242405

Review URL: https://codereview.chromium.org/188643003

git-svn-id: svn://svn.chromium.org/blink/trunk@169974 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=354356

Review URL: https://codereview.chromium.org/211313003

git-svn-id: svn://svn.chromium.org/blink/trunk@169973 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=162827

Review URL: https://codereview.chromium.org/204653002

git-svn-id: svn://svn.chromium.org/blink/trunk@169972 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This ASSERT should become useless once the LayoutState code is fixed. Remove
the ASSERT to make normal development easier.

BUG=320139

Review URL: https://codereview.chromium.org/211033009

git-svn-id: svn://svn.chromium.org/blink/trunk@169971 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This was regressed in http://src.chromium.org/viewvc/blink?revision=161843&view=revision,
when we started propagating sandbox flags to new windows via FrameLoader::forceSandboxFlags
even if the new window was caused by clicking a link with target="_blank".

BUG=353253
TEST=http/tests/navigation/new-window-sandboxed-iframe.html

Review URL: https://codereview.chromium.org/208853004

git-svn-id: svn://svn.chromium.org/blink/trunk@169970 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is in order to help determine which of these need to be kept in the
DOM spec, since it looks like some of them will be needed:
https://www.w3.org/Bugs/Public/show_bug.cgi?id=21986

BUG=none
NOTRY=true

Review URL: https://codereview.chromium.org/211363003

git-svn-id: svn://svn.chromium.org/blink/trunk@169969 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The old methods are temporarily left on WebView until Chrome switches.

Also remove WebView::isSelectionEditable since it's not called.

BUG=304341

Review URL: https://codereview.chromium.org/201773005

git-svn-id: svn://svn.chromium.org/blink/trunk@169968 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This trailing "/*" (and so forth) do not contribute very much to the signal, and not using them allows us to
catch some interesting corner cases involving server-side page post-processors.

BUG=355479

Review URL: https://codereview.chromium.org/210033008

git-svn-id: svn://svn.chromium.org/blink/trunk@169967 bbb929c8-8fbe-4397-9dbb-9b2b20218538

All of the layout tests are updated to use unprefixed AudioContext and
OfflineAudioContext. Conversion was done using a script. All tests pass.

However, we still want to be able to run these tests on older browsers,
so where possible, we add a compatibility layer (in resources/compatibility.js) to
support running selected tests in browsers with prefixed AudioContext. This was
tested manually using Chrome 34.0.1847.60 beta on Linux.

BUG=

Review URL: https://codereview.chromium.org/208943004

git-svn-id: svn://svn.chromium.org/blink/trunk@169966 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=272416
R=pfeldman

Review URL: https://codereview.chromium.org/202113007

git-svn-id: svn://svn.chromium.org/blink/trunk@169965 bbb929c8-8fbe-4397-9dbb-9b2b20218538

NOTRY=true

Review URL: https://codereview.chromium.org/210713005

git-svn-id: svn://svn.chromium.org/blink/trunk@169964 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Now each View knows about it's minimum size and containers use it to layout.
Therefore split view constraints have moved from split view to the child views.

Root view now clips the content when available window size is less than minimum size of content.

BUG=none

Review URL: https://codereview.chromium.org/197823010

git-svn-id: svn://svn.chromium.org/blink/trunk@169963 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The test occasionally crashes by touching null HTMLImportsController:m_master in a timer function.
This change cancel the timer when m_master is cleared, and add an ASSERT() to make sure that
|m_master| is available when timer is fired.

BUG=355640
TEST=imports-custom-element-abort.html
R=dominicc

Review URL: https://codereview.chromium.org/209333012

git-svn-id: svn://svn.chromium.org/blink/trunk@169962 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=353309

Review URL: https://codereview.chromium.org/208393008

git-svn-id: svn://svn.chromium.org/blink/trunk@169961 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Shadow DOM styling spec link:
http://drafts.csswg.org/css-scoping/#deep-combinator

BUG=348891
TEST=no new tests because just renaming.

Review URL: https://codereview.chromium.org/210813002

git-svn-id: svn://svn.chromium.org/blink/trunk@169960 bbb929c8-8fbe-4397-9dbb-9b2b20218538