GitLab

BUG=412251
R=pdr@chromium.org
TEST=./Tools/Scripts/run-webkit-tests fast/events/click-*

Review URL: https://codereview.chromium.org/562653002

git-svn-id: svn://svn.chromium.org/blink/trunk@181844 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=ager@chromium.org,haraken@chromium.org
BUG=

Review URL: https://codereview.chromium.org/559383002

git-svn-id: svn://svn.chromium.org/blink/trunk@181843 bbb929c8-8fbe-4397-9dbb-9b2b20218538

All scripts currently load at a medium priority and compete with each
other for resources. Async scripts do not block the parser and should be
loaded at a lower priority than those that do.

This patch lowers the priority of scripts with async or defer attributes
and scripts that were injected dynamically to load at a "low" priority
which is below blocking scripts but above (non-visible) images.

Visible images are loaded at the same priority but with a higher
intra-priority based on the screen area they cover so visible images
can preempt async scripts once layout is done (a good thing).

We want to keep scripts at a higher priority than (non-visible) images
for a couple of reasons:

1 - Scripts may cause other resource loads while images do not
(generally). There is potential to push out the onload time of the page
if a script with lots of follow-on content is loaded at the end after
images.

2 - By the time the code executes to inject the script the preload
parser may have already discovered the bulk of the images on the page.

Test results: https://docs.google.com/spreadsheet/ccc?key=0As3TLupYw2RedC1RUHlMRHhFY3RoVHEtc2FhbXQ1aWc&usp=sharing

Slight improvements (~1%) to onload, render and DOM Content Loaded.
1-2% improvement on Speed Index.

BUG=408229

Review URL: https://codereview.chromium.org/514893002

git-svn-id: svn://svn.chromium.org/blink/trunk@181842 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Some compile failures on Sheriff-O-Matic link to the wrong page today, causing
"No results to display" to be shown under the information for a host, instead
of stdio for that compile failure.

This change checks if the step is "compile" and always displays stdio in that
case.

I haven't added a test case for this; I'd like to clean up the routines in this
class to make it easier to test. I tested this manually while waiting for a
compile failure to happen that didn't display well.

NOTRY=true
BUG=404927

Review URL: https://codereview.chromium.org/549313005

git-svn-id: svn://svn.chromium.org/blink/trunk@181841 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Since we are supposed to end up with an AtomicString, don't rely on implicit conversion, but provide it as static local variable while using ConstructFromLiteral to avoid using strlen.

Review URL: https://codereview.chromium.org/567653002

git-svn-id: svn://svn.chromium.org/blink/trunk@181840 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Revert of CSP: Convert CSPSource constructor bools into a enum. (patchset #1 id:1 of https://codereview.chromium.org/565723002/)

Reason for revert:
Not a culprit.

Original issue's description:
> Revert of CSP: Convert CSPSource constructor bools into a enum. (patchset #1 id:1 of https://codereview.chromium.org/568433002/)
> 
> Reason for revert:
> Your patch broke quite a number of webkit tests:
> http://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.8%20%28retina%29/builds/19050
> http://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.7/builds/31462
> 
> Original issue's description:
> > CSP: Convert CSPSource constructor bools into a enum.
> > 
> > Rather than passing the wildcard disposition of CSPSource's hosts and ports
> > as booleans, we should have an enum that allows us to figure out hat we
> > mean at the callsite. This patch adds CSPSource::WildcardDisposition in
> > order to make this clear.
> > 
> > Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=181821
> 
> TBR=jochen@chromium.org,mkwst@chromium.org
> NOTREECHECKS=true
> NOTRY=true
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=181833

TBR=jochen@chromium.org,mkwst@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/563863002

git-svn-id: svn://svn.chromium.org/blink/trunk@181838 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Split out CSSParser public API (patchset #3 id:40001 of https://codereview.chromium.org/566483002/)

Reason for revert:
Caused compiling errors.

Original issue's description:
> Split out CSSParser public API
> 
> This patch separates the public API of BisonCSSParser from the hundreds
> of public methods it exposes. The plan is to build a new parser based on
> the css-syntax specification, which will eventually complete replace the
> bison-based parser. The functions in this class will do the run-time
> switching between the bison parser and the new (unwritten) parser.
> 
> For the functions here which are only ever used with a newly instantiated
> BisonCSSParser, I've moved instantiation to inside the CSSParser so that
> the functions can be static.
> 
> This is an updated version of a patch by eseidel from 8 months ago:
> https://codereview.chromium.org/112933010/
> 
> BUG=330389
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=181825

TBR=eseidel@chromium.org,esprehn@chromium.org,alancutter@chromium.org,timloh@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=330389

Review URL: https://codereview.chromium.org/566703002

git-svn-id: svn://svn.chromium.org/blink/trunk@181837 bbb929c8-8fbe-4397-9dbb-9b2b20218538

And switch LinkHighlight over to using DisplayList.

The GraphicsContext change is needed for cases when we wish to record a DisplayList without
an existing GraphicsContext. The LinkHighlight case demonstrates that.

R=pdr@chromium.org,chrishtr@chromium.org,esprehn@chromium.org
BUG=410019

Review URL: https://codereview.chromium.org/541913002

git-svn-id: svn://svn.chromium.org/blink/trunk@181835 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This field is no longer needed or in use by embedders, and can be
safely removed.

This is part 4 of a 4-sided patch simplifying WebTouchEvent:
1) https://codereview.chromium.org/557863002/
2) https://codereview.chromium.org/558773002/
3) https://codereview.chromium.org/559693003/
4) This change

BUG=358132

Review URL: https://codereview.chromium.org/560493003

git-svn-id: svn://svn.chromium.org/blink/trunk@181834 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of CSP: Convert CSPSource constructor bools into a enum. (patchset #1 id:1 of https://codereview.chromium.org/568433002/)

Reason for revert:
Your patch broke quite a number of webkit tests:
http://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.8%20%28retina%29/builds/19050
http://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac10.7/builds/31462

Original issue's description:
> CSP: Convert CSPSource constructor bools into a enum.
> 
> Rather than passing the wildcard disposition of CSPSource's hosts and ports
> as booleans, we should have an enum that allows us to figure out hat we
> mean at the callsite. This patch adds CSPSource::WildcardDisposition in
> order to make this clear.
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=181821

TBR=jochen@chromium.org,mkwst@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/565723002

git-svn-id: svn://svn.chromium.org/blink/trunk@181833 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This also moves the basic V8 initialization to gin. The main advantage
is that we now initialize V8 first and then start creating isolates. It
also removes some code duplication.

BUG=none
R=haraken@chromium.org,andrewhayden@chromium.org

Review URL: https://codereview.chromium.org/559363002

git-svn-id: svn://svn.chromium.org/blink/trunk@181832 bbb929c8-8fbe-4397-9dbb-9b2b20218538

I verified manually that the new files differ from the original expected
results by a few bits. I did not investigate the underlying cause for
the difference, but I suspect some changes in ffmpeg is causing these
differences.  Android was using the linux files for the expected
results so changes in ffmpeg can cause Android results to differ since
WebAudio uses Android's MediaCodec framework to decode audio files.

BUG=407368
TBR=dalecurtis@chromium.org

Review URL: https://codereview.chromium.org/501213002

git-svn-id: svn://svn.chromium.org/blink/trunk@181831 bbb929c8-8fbe-4397-9dbb-9b2b20218538

blink scheduler still needs to know about them because we
want it to enter low latency mode when the user is
interacting with the device.  To facilitate that this patch
adds an didReceiveInputEvent API function.

BUG=391005

Review URL: https://codereview.chromium.org/554693004

git-svn-id: svn://svn.chromium.org/blink/trunk@181830 bbb929c8-8fbe-4397-9dbb-9b2b20218538

WebTouchEvents passed to plugins no longer use the changedTouches or
targetTouches properties, rebaseline appropriately.

TBR=loislo@chromium.org,jianli@chromium.org
BUG=413098

Review URL: https://codereview.chromium.org/563643003

git-svn-id: svn://svn.chromium.org/blink/trunk@181828 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=yosin@chromium.org
BUG=388681

Review URL: https://codereview.chromium.org/560653003

git-svn-id: svn://svn.chromium.org/blink/trunk@181827 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This follows the lates spec update: https://www.w3.org/Bugs/Public/show_bug.cgi?id=23170

R=dglazkov@chromium.org
BUG=408272
TEST=import-defaultView.html, import-script-prototypes.html

Review URL: https://codereview.chromium.org/560183002

git-svn-id: svn://svn.chromium.org/blink/trunk@181826 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch separates the public API of BisonCSSParser from the hundreds
of public methods it exposes. The plan is to build a new parser based on
the css-syntax specification, which will eventually complete replace the
bison-based parser. The functions in this class will do the run-time
switching between the bison parser and the new (unwritten) parser.

For the functions here which are only ever used with a newly instantiated
BisonCSSParser, I've moved instantiation to inside the CSSParser so that
the functions can be static.

This is an updated version of a patch by eseidel from 8 months ago:
https://codereview.chromium.org/112933010/

BUG=330389

Review URL: https://codereview.chromium.org/566483002

git-svn-id: svn://svn.chromium.org/blink/trunk@181825 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This reverts commit 44cef15dcdbf3b78339c3678fe533f3a97e59870.

The patch were reverted due to problem with uninitialized memory.

BUG=413229
TBR= jianli@chromium.org,  dominik.rottsches@intel.com

Review URL: https://codereview.chromium.org/557823003

git-svn-id: svn://svn.chromium.org/blink/trunk@181824 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This allows MediaStreamTrack::getSources to go through UserMediaClient.

To land after this:
Chrome CL part 1: https://codereview.chromium.org/562643003/
Blink CL part 2: https://codereview.chromium.org/559423002
Chrome CL part 2 (TBD)

BUG=406094

Review URL: https://codereview.chromium.org/560843002

git-svn-id: svn://svn.chromium.org/blink/trunk@181822 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Rather than passing the wildcard disposition of CSPSource's hosts and ports
as booleans, we should have an enum that allows us to figure out hat we
mean at the callsite. This patch adds CSPSource::WildcardDisposition in
order to make this clear.

Review URL: https://codereview.chromium.org/568433002

git-svn-id: svn://svn.chromium.org/blink/trunk@181821 bbb929c8-8fbe-4397-9dbb-9b2b20218538

DevTools uses TracingStartedInWorker trace event to figure out which worker threads correspond to the inspected page. The events were written on the worker thread which is racy. Current implementation always writes the event on the main thread. Since Tracing.start/end commands are handled on the main thread this eliminates racy behavior.

BUG=401895,405429

Review URL: https://codereview.chromium.org/552323002

git-svn-id: svn://svn.chromium.org/blink/trunk@181820 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=394734

Review URL: https://codereview.chromium.org/555433002

git-svn-id: svn://svn.chromium.org/blink/trunk@181819 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In Normal mode (the default), there's virtually no 
prioritization of compositor events, and
shouldYieldForHighPriorityWork always returns false. In 
CompositorPriority mode, there *is* prioritization of 
compositor tasks and shouldYieldForHighPriorityWork will 
return false if there are pending high priority events.

Currently there is no way for client code to set the policy
but that will be introduced in a follow up CL.

Also adds some disabled by default tracing in
"blink.scheduler" to track the number of pending high
priority tasks and which policy mode the scheduler is in.

BUG=391005,411520

Review URL: https://codereview.chromium.org/559973003

git-svn-id: svn://svn.chromium.org/blink/trunk@181818 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch begins using the static version of the mixed content checker
which was introduced in https://codereview.chromium.org/537983002/. It
also takes the opportunity to rework the error message to use the new
type information we've now passed into the checker.

This patch only touches the ResourceFetcher usage; future patches will
clean up the other points in the code where we grab a MixedContentChecker
from the loader, and eventually remove the object creation entirely.
This will make the site isolation folks happy.

BUG=411184

Review URL: https://codereview.chromium.org/561153002

git-svn-id: svn://svn.chromium.org/blink/trunk@181817 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently, AudioBuffer is supported 3.0KHz to 192.0 KHz.
However, offlineAudioContext is not supported lower sample rate than 44.0KHz.

http://webaudio.github.io/web-audio-api/#the-audiobuffersourcenode-interface

“Describes the sample-rate of the linear PCM audio data
in the buffer in sample-frames per second.
An implementation must support sample-rates
in at least the range 22050 to 96000”

Changed supported sampleRate range 3KHz to 192KHz and fftsize for convolution.

BUG=394009

Review URL: https://codereview.chromium.org/375383002

git-svn-id: svn://svn.chromium.org/blink/trunk@181816 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch moves CSP generation out of FrameLoader and into
DocumentLoader in order to ensure that the 'frame-ancestors' directive
can properly block loading a document. The current implementation checks
too late in the loading process.

The new implementation works, with the caveat that it breaks both
console messages and violation reports. Since we have no
ExecutionContext, we have nothing to dump console messages into. This
work will be covered in https://crbug.com/412725

BUG=411889,411600,357462

Review URL: https://codereview.chromium.org/563473002

git-svn-id: svn://svn.chromium.org/blink/trunk@181814 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=24182
TBR= jianli@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/562123003

git-svn-id: svn://svn.chromium.org/blink/trunk@181813 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Like we do for all integer types, use helper functions from V8Binding.h
to convert from v8::Value to native types. These helper functions throw
all exceptions via their ExceptionState& argument, which means we don't
need to initialize a v8::TryCatch in the caller.

The helpers initialize a v8::TryCatch if they call in to V8 to convert
a value, but they do not do this if the value is already of the correct
type.

Review URL: https://codereview.chromium.org/567503002

git-svn-id: svn://svn.chromium.org/blink/trunk@181812 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch finishes the process of allowing policy parsing outside of
an ExecutionContext, and adjusts Document::initContentSecurityPolicy
to accept a ContentSecurityPolicy object rather than a set of headers.
If such an object is provided, the document will take ownership of it
and use it going forward. If no such object is provided, an empty
policy will be created and stored.

A future patch will move parsing out of FrameLoader::didBeginDocument
into DocumentLoader::responseReceived so that we can begin doing some
checks before a document is created.

BUG=411889

Review URL: https://codereview.chromium.org/559503002

git-svn-id: svn://svn.chromium.org/blink/trunk@181811 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Only DatabaseManager::closeDatabasesImmediately could be called in non-main
threads. However, it just calls
DatabaseTracker::closeDatabasesImmediately(). This CL removes
DatabaseManager::closeDatabasesImmediately, and only the main thread uses
DatabaseManager.

So, we can remove DatabaseManager::m_contextMapLock.  We still need m_contextMap
because the main thread can have multiple Documents.

Oilpan: We can use PersistentHeapHashMap, and remove some |#if ENABLE(OILPAN)|.

BUG=347902,397429
TEST=No new tests because of no behavior changes.

Review URL: https://codereview.chromium.org/567453002

git-svn-id: svn://svn.chromium.org/blink/trunk@181810 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL merges
 - AbstractSQLStatement into SQLStatement
 - AbstractSQLStatementBackend into SQLStatementBackend
 - AbstractSQLTransaction into SQLTransaction
 - AbstractSQLTransactionBackend into SQLTransactionBackend
 - DatabaseBase into Database.

'Sync' versions of their subclasses were removed. These abstract classes are
unnecessary.

BUG=397429
TEST=none; No behavior changes.

Review URL: https://codereview.chromium.org/561143002

git-svn-id: svn://svn.chromium.org/blink/trunk@181809 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch adjusts DrawingBuffer's behavior when a context is lost in
two ways:

1. Always tell DrawingBuffer whether it is visible or not regardless of
   whether the WebGL context is lost. This is needed because visibility
   notifications may arrive from the browser while the context is lost
   and the compositor may also attempt to composite the DrawingBuffer
   during this time.

2. Don't check the current visibility in
   DrawingBuffer::prepareMailbox() if the DrawingBuffer is being
   destroyed. This is because the DrawingBuffer may become hidden right
   before the context is forcibly lost (using WEBGL_lose_context), and
   after that point WebGLRenderingContextBase no longer has a reference
   to the DrawingBuffer to update its visibility state.

BUG=411372
TEST=Added in https://codereview.chromium.org/560023002

Review URL: https://codereview.chromium.org/556303002

git-svn-id: svn://svn.chromium.org/blink/trunk@181808 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Removed unused threading/os code from tool/main.py and
tool/mocktool.py.

BUG=None

Review URL: https://codereview.chromium.org/563533002

git-svn-id: svn://svn.chromium.org/blink/trunk@181807 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This enables and tweaks ready.html to follow up the implementation changes.

1) Blink: https://codereview.chromium.org/476043002/
2) Chromium: https://codereview.chromium.org/477593007/
3) Blink: THIS PATCH


BUG=399533
TEST=run_webkit_tests.py --debug http/tests/serviceworker/

Review URL: https://codereview.chromium.org/532653002

git-svn-id: svn://svn.chromium.org/blink/trunk@181806 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Schedule an invalidation set for :empty on an element when the pseudo state
changes instead of a SubtreeStyleChange.

R=chrishtr@chromium.org,esprehn@chromium.org
BUG=412999

Review URL: https://codereview.chromium.org/565493002

git-svn-id: svn://svn.chromium.org/blink/trunk@181805 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We're currently performing mixed content checks based on Resource::Type,
which isn't granular enough to support the behavior specified in the
mixed content specification (hence the introduction of the
MixedContentBlockingTreatment enum).

This patch introduces a new static mixed content checking method based
on WebURLRequest::RequestContext, which will allow us to remove most
of the complexity of MixedContentChecking in future patches.

BUG=411184

Review URL: https://codereview.chromium.org/537983002

git-svn-id: svn://svn.chromium.org/blink/trunk@181804 bbb929c8-8fbe-4397-9dbb-9b2b20218538

ExecutionContextClient was introduced in https://codereview.chromium.org/27311002. Now it isn't used.

R=vsevik@chromium.org

Review URL: https://codereview.chromium.org/559553002

git-svn-id: svn://svn.chromium.org/blink/trunk@181803 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Since Blink 175446
https://src.chromium.org/viewvc/blink?revision=175446&view=revision
we're loading data into a file. The downloadToFile flag is determined in
createRequest() i.e. inside xhr.send() call. But responseType is allowed
to be changed even after xhr.send() call.

Accumulate data into m_binaryResponseBuilder and create a Blob from it
anyway if m_downloadingToFile is not set.

BUG=412704

Review URL: https://codereview.chromium.org/562563003

git-svn-id: svn://svn.chromium.org/blink/trunk@181802 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Left-over from the removal of diffNeedsRecompositeLayer [1].

[1] https://codereview.chromium.org/360103008

R=abarth@chromium.org

Review URL: https://codereview.chromium.org/565583002

git-svn-id: svn://svn.chromium.org/blink/trunk@181801 bbb929c8-8fbe-4397-9dbb-9b2b20218538

After r164062 percentages of indefinite sizes were correctly
computed to "auto". The problem is that we were not doing it when
the percentage was inside the minmax() function. In those cases
it should compute to min-content for the min track sizing
function or to max-content for the max track sizing function.

Added a couple of tests, one to verify that we do not assert with
percentages inside minmax() and another one to verify that we
properly compute the percentage of indefinite size in various
situations.

BUG=401477,409085

Review URL: https://codereview.chromium.org/450093004

git-svn-id: svn://svn.chromium.org/blink/trunk@181800 bbb929c8-8fbe-4397-9dbb-9b2b20218538