GitLab

use new onDrawPicture variant, as the previous is now deprecated and is no longer called by the base-class.

TBR=fmalita@chromium.org

Review URL: https://codereview.chromium.org/455213002

git-svn-id: svn://svn.chromium.org/blink/trunk@179891 bbb929c8-8fbe-4397-9dbb-9b2b20218538

- Merge multiple |namespace blink {}| blocks
- Remove unnecessary blink:: prefixes
- Make blank lines consistent

BUG=

Review URL: https://codereview.chromium.org/455883003

git-svn-id: svn://svn.chromium.org/blink/trunk@179889 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This innocuous-looking call was changed in
https://codereview.chromium.org/255563004 to produce huge data URLs for
canvas elements.  This caused a performance regression in Android
WebView which needs to report URLs of the last tapped element to the
embedding app.  Switch absoluteImageURL() back to its previous behavior
of ignoring canvas elements, and introduce a new method
absoluteImageURLIncludingCanvasDataURL() that's only used in
WebViewImpl::saveImageAt and copyImageAt.

NOTRY=true
BUG=395859

Review URL: https://codereview.chromium.org/457013002

git-svn-id: svn://svn.chromium.org/blink/trunk@179888 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The 'border-image' has 'border-image-source', 'border-image-slice',
'border-image-width', 'border-image-outset' and 'border-image-repeat'.
http://www.w3.org/TR/css3-background/
m_imageSlice is for 'border-image-slice' and m_borderSlice is for
'border-image-width'.
It causes confusion between m_borderSlice and m_imageSlice.
This patch changes m_borderSlice to m_borderWidth along with
property name.

BUG=
R=tonyg@chromium.org

Review URL: https://codereview.chromium.org/456753002

git-svn-id: svn://svn.chromium.org/blink/trunk@179887 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Add some performance testing for paint invalidations (https://codereview.chromium.org/450813002/)

Reason for revert:
Timing out on perf bots. One example:
http://chromegw/i/chromium.perf/builders/Linux%20Perf%20%281%29/builds/2268/steps/blink_perf/logs/stdio

Original issue's description:
> Add some performance testing for paint invalidations
> 
> After moving paint invalidations after compositing updates, we
> lost most of the performance coverage we had as most performance
> tests trigger layout without generating a frame.
> 
> The new cases are DRT specific as we use
> Internals.forceCompositingUpdate. This is required as rAF
> will cap the tests to the sync frequency.
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=179881

TBR=dsinclair@chromium.org,esprehn@chromium.org,leviw@chromium.org,xianxianzhu@chromium.org,jchaffraix@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/459663002

git-svn-id: svn://svn.chromium.org/blink/trunk@179886 bbb929c8-8fbe-4397-9dbb-9b2b20218538

skia is changing the signature for onDrawPicture, so this CL adds the new one, and removes OVERRIDE from the old.

BUG=

Review URL: https://codereview.chromium.org/454993002

git-svn-id: svn://svn.chromium.org/blink/trunk@179885 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The try except block in _rename_results_folder() is throwing an UnboundLocalError.
Because the return statement in except block is inside if statement.
If the condition is false the function doesnot return which throws an error so the return statement must be outside the if statement.

Review URL: https://codereview.chromium.org/454853002

git-svn-id: svn://svn.chromium.org/blink/trunk@179884 bbb929c8-8fbe-4397-9dbb-9b2b20218538

All this class did was prevent callers from null pointer checking the
object, but all hasCompositedLayerMapping() does is null check it so it's
not clear what value this is adding, lets remove it.

Review URL: https://codereview.chromium.org/457473002

git-svn-id: svn://svn.chromium.org/blink/trunk@179883 bbb929c8-8fbe-4397-9dbb-9b2b20218538

After moving paint invalidations after compositing updates, we
lost most of the performance coverage we had as most performance
tests trigger layout without generating a frame.

The new cases are DRT specific as we use
Internals.forceCompositingUpdate. This is required as rAF
will cap the tests to the sync frequency.

Review URL: https://codereview.chromium.org/450813002

git-svn-id: svn://svn.chromium.org/blink/trunk@179881 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The EME specification now describes the format of the license
request provided to the application via the message attribute of
the message event as being JSON. This change allows the browser
tests to work once the new message format is implemented.

BUG=358271
TEST=EME layout tests pass with and without the JSON change

Review URL: https://codereview.chromium.org/429033003

git-svn-id: svn://svn.chromium.org/blink/trunk@179880 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Also report INFINITY if number of samples is 1. This is another way to interpret the confidence interval of 1 data point.

NOTRY=True
BUG=401806
TEST=tools/perf/run_benchmark smoothness.fast_path.key_silk_cases;  # Open tools/perf/results.html and ensure no NaNs.

Review URL: https://codereview.chromium.org/457873002

git-svn-id: svn://svn.chromium.org/blink/trunk@179878 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This logs errors when the underlying commands fail, and tries
to recover a bit more cleanly.

TBR=eseidel@chromium.org, ojan@chromium.org, wangxianzhu@chromium.org
BUG=397678

Review URL: https://codereview.chromium.org/417123006

git-svn-id: svn://svn.chromium.org/blink/trunk@179877 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The interface for getExcludedIntervals was designed when shape-inside
was still a thing, and it was possible for a line in a shape to have
multiple segments. shape-outside can only have one segment per line, so
this refactors the code to simplify and reflect that reality.

Review URL: https://codereview.chromium.org/454123002

git-svn-id: svn://svn.chromium.org/blink/trunk@179876 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This code was attempting to take the fast scrolling path for RenderViews that
were not composited. Unfortunately, there's no way for the compositor to scroll
RenderViews that are not composited. Instead, we should bail out and run the
slow scrolling path.

This CL also removes some dead code from this function and updates the comments
to more accurately reflect reality.

R=jamesr@chromium.org
BUG=401878

Review URL: https://codereview.chromium.org/457773002

git-svn-id: svn://svn.chromium.org/blink/trunk@179875 bbb929c8-8fbe-4397-9dbb-9b2b20218538

No one should be reading the hasVisibleDescendant state while its dirty. This
CL adds an ASSERT to prevent that from happening and fixes the callers that are
reading dirty state today.

R=skobes@chromium.org

Review URL: https://codereview.chromium.org/455963002

git-svn-id: svn://svn.chromium.org/blink/trunk@179874 bbb929c8-8fbe-4397-9dbb-9b2b20218538

And fix compilation breakage from r179850.

TBR=haraken,oilpan-reviews
BUG=340522
NOTRY=true

Review URL: https://codereview.chromium.org/451153002

git-svn-id: svn://svn.chromium.org/blink/trunk@179873 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=haraken,oilpan-reviews
BUG=399178
NOTRY=true

Review URL: https://codereview.chromium.org/453713004

git-svn-id: svn://svn.chromium.org/blink/trunk@179872 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This object is the proxy for the ServiceWorker cache; it's created
entirely through the CacheStorage interface.

R=falken@chromium.org,jkarlin@chromium.org
BUG=399178

Review URL: https://codereview.chromium.org/434543002

git-svn-id: svn://svn.chromium.org/blink/trunk@179871 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=179857

BUG=394004
TBR=wangxianzhu@chromium.org

Review URL: https://codereview.chromium.org/452923007

git-svn-id: svn://svn.chromium.org/blink/trunk@179870 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Send QUIT_LOG to LogToServer thread immediately on exception
to avoid blocking.

We still need to press ctrl-c twice to quit the whole rebaseline-o-matic
script. This is intended because we want the first ctrl-c to break the
current rebaseline, and the second to break the outer loop. This also
ensures that the loop continues on any temporary exception raised during
a rebaseline.

Review URL: https://codereview.chromium.org/454933003

git-svn-id: svn://svn.chromium.org/blink/trunk@179869 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Bust the MemoryCache to ensure script requests reach the browser-side and get added to and retrieved from the ServiceWorker's script cache.

BUG=387589

Review URL: https://codereview.chromium.org/419693002

git-svn-id: svn://svn.chromium.org/blink/trunk@179868 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In practice, it is never set to true, since the only way for it to be set true is based on a previous DocumentWriter being set to true.

BUG=

Review URL: https://codereview.chromium.org/444113002

git-svn-id: svn://svn.chromium.org/blink/trunk@179867 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The pixel expectations existed when the tests were pixel tests but now the
tests are text-only tests.

TBR=dpranke@chromium.org

Review URL: https://codereview.chromium.org/449273003

git-svn-id: svn://svn.chromium.org/blink/trunk@179866 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We shouldn't clear dirty bits by walking up the tree. Instead, we should mark
dirty bits up the tree and clean the dirty bits by walking down the tree.

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/452653002

git-svn-id: svn://svn.chromium.org/blink/trunk@179865 bbb929c8-8fbe-4397-9dbb-9b2b20218538

SVGPathSegRole is removed since it was not used at all.

BUG=401943

Review URL: https://codereview.chromium.org/456763002

git-svn-id: svn://svn.chromium.org/blink/trunk@179864 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Before these change, a missing hand-cursor at the touched node
highlighted the nearest ancestor node in DOM that has such a cursor.
This is not expected since the touched node may be visually detached
from the ancestor.

BUG=322323

Review URL: https://codereview.chromium.org/402603005

git-svn-id: svn://svn.chromium.org/blink/trunk@179863 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The pixel expectations existed when the tests were pixel tests but now the
tests are text-only tests.

CL was generated with the following python script:
files = subprocess.Popen(['find', '.', '-type', 'f', '-not', '-name', '*-expected.*'], stdout=subprocess.PIPE).stdout.readlines()
for f in files:
  f = f.strip()
  c = open(f).read()
  if (c.find('text-based-repaint.js') != -1 and c.find('runRepaintAndPixelTest') == -1) or c.find('testRunner.dumpAsText()') != -1:
    test_name = re.sub('\.[a-z]*$', '', f)
    os.system('rm ' + test_name + '-expected.png' + ' platform/*/' + test_name + '-expected.png')

False positives are restored manually according to trybot results.

R=dpranke@chromium.org

Review URL: https://codereview.chromium.org/453733002

git-svn-id: svn://svn.chromium.org/blink/trunk@179862 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Flexbox with flex-direction: row/row-reverse and with
overflow:auto on flex item shows incorrect flexbox height.
The flexbox seems to use the height from before
adding the horizontal scrollbar.
    
This bug is occuring because we are not taking into account
the horizontal scrollbar height of flex-item while computing
the childCrossAxisMarginBoxExtent. This patch fixes this bug
by adding scrollbarLogicalHeight while calculating childCrossAxisMarginBoxExtent.


BUG=327446

Review URL: https://codereview.chromium.org/439063002

git-svn-id: svn://svn.chromium.org/blink/trunk@179861 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It's possible for some calls (e.g. WebViewImpl::layout) to try to
initialize compositing after shutdown.

This just moves an early out earlier so that it does not assert when
this happens.

This is a small band-aid fix to stop this assert from happening, rather
than the larger fix of trying to stop all the cases where this happen or
refactoring WebViewImpl to not have this awkwardness int he first place.

R=chrishtr@chromium.org
BUG=none

Review URL: https://codereview.chromium.org/453523002

git-svn-id: svn://svn.chromium.org/blink/trunk@179860 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The pixel expectations existed when the tests were pixel tests but now the
tests are text-only tests.

TBR=dpranke@chromium.org

Review URL: https://codereview.chromium.org/457443002

git-svn-id: svn://svn.chromium.org/blink/trunk@179859 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The previous version had an infinite loop bug in RenderObject when there are two nested frames.

Previously, the *actual* paint invalidation container, meaning the enclosing
compositing layer / root RenderView, could already have been across a frame boundary. The logic to do this correctly was done via special code in RenderView.

Instead, generalize the existing mechanisms to find a paint invalidation container
and map rects to repaint container coordinate space to cross frame boundaries.
This simplifies the code, and also causes paint invalidation rects to always be stored
in the coordinate space of their graphics layer backing. The latter is important if we want
to use these rects for determining which parts of a graphics layer need to be painted.

BUG=401156

Review URL: https://codereview.chromium.org/455883002

git-svn-id: svn://svn.chromium.org/blink/trunk@179858 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Use setShouldDoFullPaintInvalidation() for repaint-only style change
to avoid paintInvalidationForWholeRenderer() on repaint-only style
change.

BUG=394004
TEST=existing layout tests. Some previous duplicated repaints combined.

Review URL: https://codereview.chromium.org/443123003

git-svn-id: svn://svn.chromium.org/blink/trunk@179857 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Argument conversion, particularly variadic argument conversion, goes
through toNativeArray<>(), which means that this is needed to support
idl for arguments of type "any...".

R=haraken@chromium.org
BUG=None

Review URL: https://codereview.chromium.org/440243003

git-svn-id: svn://svn.chromium.org/blink/trunk@179856 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Chromium's //content/browser is not allowed to import header files that are more
than simple definitions and so cannot include WebTextInfo to get flags from it.
Move them to another, existing, "simple" header so they can be used.

BUG=303883

Review URL: https://codereview.chromium.org/450373002

git-svn-id: svn://svn.chromium.org/blink/trunk@179855 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This converts the treestatus JS file into a model HTML file.

BUG=400392
NOTRY=true

Review URL: https://codereview.chromium.org/443243002

git-svn-id: svn://svn.chromium.org/blink/trunk@179854 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We are planning to return ReadableStream as "stream" response type. In preparation
to do so, this CL makes XHR moves exsisting "stream" response type to "legacystream".

BUG=401396

Review URL: https://codereview.chromium.org/449803002

git-svn-id: svn://svn.chromium.org/blink/trunk@179853 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Uninline HTMLCollection / LiveNodeList's item() and length() member functions
as these are big. Those call directly methods on CollectionIndexCache and all
of them are inlined.

This reduces the stripped release binary size by ~24Kb. I measured the impact
of this change on BrowserMark 2.1's DOM Search subtest (which tests this code
extensively) and it did not seem to regress:
- Before: 5785.4 +/- 0.8%
- After: 5835.4 +/- 1%

R=adamk@chromium.org

Review URL: https://codereview.chromium.org/450273003

git-svn-id: svn://svn.chromium.org/blink/trunk@179852 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch removes some deadcode from the inspector:
1) GraphicsContextSnapshot had an unused isCertainlyOpaque bool.
2) PageDebuggerAgent had unused preprocessEventListener and preprocess.
3) The continueWithPolicyDownload instrumentation was unused.
4) Three unused function declarations have been removed from InspectorDOMAgent.

No new tests as there is no change in behavior.

R=caseq

Review URL: https://codereview.chromium.org/447553002

git-svn-id: svn://svn.chromium.org/blink/trunk@179851 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=vsevik@chromium.org,aandrey@chromium.org

Review URL: https://codereview.chromium.org/376213002

git-svn-id: svn://svn.chromium.org/blink/trunk@179850 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Instead, choose to promote if the canvas element is accelerated.

R=abarth
BUG=364908

Review URL: https://codereview.chromium.org/456733003

git-svn-id: svn://svn.chromium.org/blink/trunk@179849 bbb929c8-8fbe-4397-9dbb-9b2b20218538