GitLab

It is valid to do this:

linearRampToValueAtTime(0, 1);
setValueCurveAtTime(curve, 1, 1);

Previously, this caused an error because the code thought the two
methods overlapped.

BUG=528423
TEST=audioparam-setValueCurve-exceptions.html updated with more tests

Review URL: https://codereview.chromium.org/1328963003

git-svn-id: svn://svn.chromium.org/blink/trunk@201832 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=528186

Review URL: https://codereview.chromium.org/1313223005

git-svn-id: svn://svn.chromium.org/blink/trunk@201831 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=201813

BUG=524134
TBR=wangxianzhu@chromium.org

Review URL: https://codereview.chromium.org/1317353008 .

git-svn-id: svn://svn.chromium.org/blink/trunk@201828 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch fixes a long-standing bug where crashing reftests would not
show some crash logs on the layout tests results page.

single_test_runner.py has an optimization to not run reftests when
the base test crashes. This optimization accidentally failed to write
crash logs (done through write_test_result(...)) as well. This patch
fixes this bug and adds a small test.

This is a reland of https://codereview.chromium.org/1312693002 which
was rolled out due to crashes on Win10 (see http://crbug.com/524248#50).
This patch has two potential fixes for the crashes seen before:
1) test results are only written for crash tests, not timeouts.
2) The non-null expected_driver_output is used in the call for
write_test_result instead of reference_output which is null and
caused asserts.

Review URL: https://codereview.chromium.org/1313533004

git-svn-id: svn://svn.chromium.org/blink/trunk@201827 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=527138
R=benjhayden@chromium.org,dsinclair@chromium.org

Review URL: https://codereview.chromium.org/1305043007

git-svn-id: svn://svn.chromium.org/blink/trunk@201825 bbb929c8-8fbe-4397-9dbb-9b2b20218538

LayoutBlocks delay updating their scrolling information until the
uppermost flex box has finished layout.  However, if the flex item
has overflow:auto, this means that the scrollbar width is not
included in the flex item's logical height during flexing.  This
leads to spurious scrollbars in the flex direction.

The original reason for delaying the update of scrolling information
was to fix a bug where scroll positions would be lost when a flexbox
changed:

https://bugs.webkit.org/show_bug.cgi?id=97706

For better or worse, this also turned out to be a performance
optimization (albeit at the expense of correctness).  Updating the
scroll information can cause a block to run layout again, if the
existence of scrollbars has changed for the block.  If a block in a
nested hierarchy of flexboxes runs layout N times during flexing, it
would then potentially run one additional layout at the end of
flexing, for a total of N+1.  However, if the the optimization is
removed (i.e., scroll info updating for the flex item is not delayed, but
run immediately at the end of layout, potentially triggering a second
layout if scrollbars changed), then the potential maximum number of
layouts during flexing is N*2.

This patch fixes the correctness issue while preserving the
optimization: when a flex item runs layout, and its scrollbars have
changed, it immediately adds the size of the scrollbars to its
logical width/height, but defers the second layout until flexing is
finished.  The implementation splits the existing code from
DeprecatedPaintLayerScrollableArea::updateAfterLayout() into the
methods updateScrollDimensions, which is called immediately at the
end of layout; and finalizeScrollDimensions, which is called at the
end of flexing.  When layout happens outside the context of flexing,
the two methods are called in immediate sequence, preserving existing
behavior.

BUG=512229
R=ojan@chromium.org,cbiesinger@chromium.org,leviw@chromium.org

Review URL: https://codereview.chromium.org/1295933003

git-svn-id: svn://svn.chromium.org/blink/trunk@201824 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This test didn't need to setTimeout, so lets speed it up by 10+ms.

TBR=estaab

Review URL: https://codereview.chromium.org/1312903016

git-svn-id: svn://svn.chromium.org/blink/trunk@201823 bbb929c8-8fbe-4397-9dbb-9b2b20218538

command buffer side CL: https://codereview.chromium.org/1319453010/

BUG=527249
TEST=conformance2/state/gl-object-get-calls.html related parts pass
R=kbr@chromium.org,bajones@chromium.org

Review URL: https://codereview.chromium.org/1303343009

git-svn-id: svn://svn.chromium.org/blink/trunk@201822 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Suppress failures on fixed-img-src-change-after-scroll.html and background-image-paint-invalidation.html

BUG=528431
TBR=wangxianzhu
NOTRY=true

Review URL: https://codereview.chromium.org/1331483002

git-svn-id: svn://svn.chromium.org/blink/trunk@201821 bbb929c8-8fbe-4397-9dbb-9b2b20218538

For all display items for one object, if we check for interect rect
intersection and early return, we must use the same rule. Otherwise,
we may generate display item A but not display item B in some cases;
and once we generate display item B again, we'll mistakenly generate
CachedDisplayItem for it because we think the object is validly cached.

Table cell may generate more than two display items for backgrounds,
one drawing backgrounds from containers, one drawing its own background.
Previously they use different early return rules.

Now let them use the same rule (BlockPainter::isVisibleInPaintRect()).

Also add DisplayItem::TableCellBackgroundFromContainers because the
display items are issued by TableSectionPainter which doesn't know if
TableCellPainter::paint() would use the phase for other display items.

BUG=522338
TEST=DisplayItemListPaintTest.TableCellBackbroundInterestRect

Review URL: https://codereview.chromium.org/1312493007

git-svn-id: svn://svn.chromium.org/blink/trunk@201820 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch marks fast/repaint/fixed-in-page-scale.html as timing out on
windows. For more information, see the bug.

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Frepaint%2Ffixed-in-page-scale.html&showExpectations=true

TBR=bokan
BUG=528421
NOTRY=true

Review URL: https://codereview.chromium.org/1301423007

git-svn-id: svn://svn.chromium.org/blink/trunk@201819 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=521663
TEST=this is a test only change
r=japhet@chromium.org

Review URL: https://codereview.chromium.org/1302333003

git-svn-id: svn://svn.chromium.org/blink/trunk@201818 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This may be helpful when we call showDebugData() during the merge
algorithm.

Review URL: https://codereview.chromium.org/1323263002

git-svn-id: svn://svn.chromium.org/blink/trunk@201817 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Added class level comments to DeprecatedPaintLayerClipper
and DeprecatedPaintLayerStackingNode.

While touching DeprecatedPaintLayerStackingNode, added a comment
about NormalFlowChildren.

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=201805

Review URL: https://codereview.chromium.org/1327483003

git-svn-id: svn://svn.chromium.org/blink/trunk@201815 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This change adds some documentation about how we structure <table>
internally. In particular, it explains LayoutTable::m_column and
LayoutTableSection::m_grid  and the role of LayoutTable as a grand
coordinator for determining the table structure.

It's still missing some pieces but it's a good start and will be
completed as I clear my thoughts on the code.

Review URL: https://codereview.chromium.org/1319453009

git-svn-id: svn://svn.chromium.org/blink/trunk@201814 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This ensures test coverage of repaint tests for spv2.

paint/invalidation/spv2/resources/convert.sh is a semi-automatic tool to
convert an text-based-repaint test to a spv2 paint invalidation test.

See https://codereview.chromium.org/1324763002/ for details about spv2 paint
invalidation test.

Removed fast/repaint and compositing/repaint as spv2 virtual suites.
Run paint/invalidation/spv2 as spv2 physical suite.

BUG=524236,524134,527242

Review URL: https://codereview.chromium.org/1302183007

git-svn-id: svn://svn.chromium.org/blink/trunk@201813 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove obsolete WebThread::Task class.

Patch 1: https://codereview.chromium.org/1325073002
Patch 2: https://codereview.chromium.org/1308183005
Patch 3: https://codereview.chromium.org/1303153005
Patch 4: https://codereview.chromium.org/1309423004

BUG=510398, 497761

Review URL: https://codereview.chromium.org/1310253009

git-svn-id: svn://svn.chromium.org/blink/trunk@201812 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In BMPImageDecoder, never call SharedBuffer::data(), which potentially
copies all the data. Instead, do the following:

- Read directly from the SharedBuffer when it is safe to do so. e.g.
  when only one byte is needed. We already know there is enough data,
  and getSomeData will return at least a byte if there is enough data.
- Use a FastSharedBufferReader when we may be on a segment boundary.

As a result, we make fewer copies when decoding BMPs. We never call
SharedBuffer::data(), which copies up to the entire encoded data
received so far. Occasionally we will make small copies when a two or
four byte number is on a segment boundary.

Add a setter to FastSharedBufferReader to update its SharedBuffer, so the
BMPImageReader can have one as a member, which it will reuse across calls.

Remove BIG_ENDIAN, and eliminate another memcpy that was there to support
it.

Using my modified version of ImageDecodeBench (see crrev.com/1145373003),
I see a slight speedup as compared to the old version which relied on
SharedBuffer::data(). (My data can be found here: [1].) This matches my
expectations, since the new version is copying less data.

Using a version of ImageDecodeBench with a premerged buffer (see
crrev.com/1318713005), I still see a speed improvement, although it is
very small. (See sheet 2 of the above spreadsheet).

[1] https://docs.google.com/a/google.com/spreadsheets/d/1GN2FDytKvkdZzV-WF4IEQaYQF65wa4bRxG67bNgpjAE/edit?usp=sharing

BUG=467772

Review URL: https://codereview.chromium.org/1259083003

git-svn-id: svn://svn.chromium.org/blink/trunk@201811 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch suppresses a timeout on [1] likely due to [2]. See:
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=http%2Ftests%2Fnavigation%2Fscrollstate-after-http-equiv-refresh-fragment-identifier.html&testType=layout-tests

[1] http/tests/navigation/scrollstate-after-http-equiv-refresh-fragment-identifier.html
[2] https://src.chromium.org/viewvc/blink?view=rev&revision=201780

NOTRY=true
BUG=528352
TBR=alexclarke

Review URL: https://codereview.chromium.org/1326193002

git-svn-id: svn://svn.chromium.org/blink/trunk@201810 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Add documentation to 2 DeprecatedPaintLayer objects (patchset #2 id:20001 of https://codereview.chromium.org/1327483003/ )

Reason for revert:
Speculative revert: Failed build job.

I don't know why a comment would cause a breakage but the build output shows: 

In file included from ../../third_party/WebKit/Source/core/paint/DeprecatedPaintLayer.h:50:0,
                 from ../../third_party/WebKit/Source/core/layout/svg/LayoutSVGModelObject.cpp:39:
../../third_party/WebKit/Source/core/paint/DeprecatedPaintLayerClipper.h:135:1:error: multi-line comment [-Werror=comment]
 //                /    \
 ^

Original issue's description:
> Add documentation to 2 DeprecatedPaintLayer objects
> 
> Added class level comments to DeprecatedPaintLayerClipper
> and DeprecatedPaintLayerStackingNode.
> 
> While touching DeprecatedPaintLayerStackingNode, added a comment
> about NormalFlowChildren.
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=201805

TBR=eae@chromium.org,mstensho@opera.com,wkorman@chromium.org,jchaffraix@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/1312203004

git-svn-id: svn://svn.chromium.org/blink/trunk@201809 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Switch blink repo to use chromium-status.appspot.com (patchset #1 id:1 of https://codereview.chromium.org/1315323007/ )

Reason for revert:
This moves chromium.webkit off blink tab to chromium tab on sheriff-o-matic, which is undesired (https://code.google.com/p/chromium/issues/detail?id=528346)

Original issue's description:
> Switch blink repo to use chromium-status.appspot.com
> 
> See https://groups.google.com/a/chromium.org/d/msg/blink-dev/S-P3N0kdkMM/EypTXImtAwAJ
> for context.
> 
> R=akuegel@chromium.org, akuegel
> TBR=jochen
> 
> BUG=431478
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=201792

TBR=akuegel@chromium.org,tandrii@chromium.org,jochen@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=431478, 528346

Review URL: https://codereview.chromium.org/1308163009

git-svn-id: svn://svn.chromium.org/blink/trunk@201808 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=jochen@chromium.org, kbr@chromium.org
BUG=345227

Review URL: https://codereview.chromium.org/1325453007

git-svn-id: svn://svn.chromium.org/blink/trunk@201806 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Added class level comments to DeprecatedPaintLayerClipper
and DeprecatedPaintLayerStackingNode.

While touching DeprecatedPaintLayerStackingNode, added a comment
about NormalFlowChildren.

Review URL: https://codereview.chromium.org/1327483003

git-svn-id: svn://svn.chromium.org/blink/trunk@201805 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch fixes a bug where a background would be considered obscured
during loading but would not get reexamined for the background
obscuration bit after the image loaded. Other loading codepaths
(e.g., <images>) had this case covered but background images were
missing this check.

BUG=422272

Review URL: https://codereview.chromium.org/1312943008

git-svn-id: svn://svn.chromium.org/blink/trunk@201804 bbb929c8-8fbe-4397-9dbb-9b2b20218538

See discussion in bug.
BUG=526104

Review URL: https://codereview.chromium.org/1310153004

git-svn-id: svn://svn.chromium.org/blink/trunk@201803 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Use of the base namespace in Blink is not currently recommended. Fortunately
WTF::OwnPtr is a drop-in replacement here.

Review URL: https://codereview.chromium.org/1310743006

git-svn-id: svn://svn.chromium.org/blink/trunk@201802 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=pdr

Review URL: https://codereview.chromium.org/1316863011

git-svn-id: svn://svn.chromium.org/blink/trunk@201800 bbb929c8-8fbe-4397-9dbb-9b2b20218538

By storing parsed meta viewport, we can apply it without reload when emulating mobile device.

BUG=none

Review URL: https://codereview.chromium.org/1328553003

git-svn-id: svn://svn.chromium.org/blink/trunk@201799 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We were only testing the default value for align-items (stretch),
but not other values (like we do for justify-items).

Modified the test to check the non-stretch behavior too.

BUG=510061
TEST=fast/css-grid-layout/min-width-height-auto.html

Review URL: https://codereview.chromium.org/1310603008

git-svn-id: svn://svn.chromium.org/blink/trunk@201797 bbb929c8-8fbe-4397-9dbb-9b2b20218538

For the reason described in the bug, using a Member for SVGPropertyTearOffBase::m_contextElement
causes a cycle. To break the cycle, this CL changes it to a raw pointer.
The raw pointer is safe because the context element is guaranteed to be kept alive
by a V8 wrapper of the SVG tear off object.

BUG=528275

Review URL: https://codereview.chromium.org/1330673002

git-svn-id: svn://svn.chromium.org/blink/trunk@201796 bbb929c8-8fbe-4397-9dbb-9b2b20218538

according to RFC 7230.

In this CL:
- We don't show messages for Fetch API.
- We don't actually update header checks.
Also, this CL is preparation for actually updating
header value checks to RFC 7230.

These issue is a merge from:
https://bugs.webkit.org/show_bug.cgi?id=128593.

BUG=455099

Review URL: https://codereview.chromium.org/1018903002

git-svn-id: svn://svn.chromium.org/blink/trunk@201795 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Allocate this object separately and not as a part of a LocalFrame. The
main motivation for this is to play better with its use of a
CancellableTaskFactory following r201780.

R=haraken
BUG=

Review URL: https://codereview.chromium.org/1330653003

git-svn-id: svn://svn.chromium.org/blink/trunk@201794 bbb929c8-8fbe-4397-9dbb-9b2b20218538

See https://groups.google.com/a/chromium.org/d/msg/blink-dev/S-P3N0kdkMM/EypTXImtAwAJ
for context.

R=akuegel@chromium.org, akuegel
TBR=jochen

BUG=431478

Review URL: https://codereview.chromium.org/1315323007 .

git-svn-id: svn://svn.chromium.org/blink/trunk@201792 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Add partitionsOutOfMemoryUsingXXX to know Blink memory usage from crash reports when PartitionAlloc hits OOM.

To avoid linker optimization, copied alias.cc and alias.h from //base/debug to WTF.

BUG=

Review URL: https://codereview.chromium.org/1315113006

git-svn-id: svn://svn.chromium.org/blink/trunk@201791 bbb929c8-8fbe-4397-9dbb-9b2b20218538

They failed and a suspected culprit was reverted,
but they passed before the revert landed,
and the change was relanded and now the tests are
failing again. Disabling and filing a bug.

BUG=528186
TBR=alexmos

Review URL: https://codereview.chromium.org/1321233004

git-svn-id: svn://svn.chromium.org/blink/trunk@201790 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Make classes and structures in web fast-allocated. (patchset #1 id:1 of https://codereview.chromium.org/1328773002/ )

Reason for revert:
Revert to see whether SitePerProcessBrowserTest.SurfaceHitTestTest flakiness is caused by this commit or not.


Original issue's description:
> Make classes and structures in web fast-allocated.
> 
> - Added STATIC_ONLY, STACK_ALLOCATED, DISALLOW_ALLOCATION, or ALLOW_ONLY_INLINE_ALLOCATION
>   instead of WTF_MAKE_FAST_ALLOCATED(_WILL_BE_REMOVED) if possible.
> 
> BUG=523249
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=201771

TBR=haraken@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=523249

Review URL: https://codereview.chromium.org/1326693003

git-svn-id: svn://svn.chromium.org/blink/trunk@201789 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=528198
TBR=horo
NOTRY=true

Review URL: https://codereview.chromium.org/1307233007

git-svn-id: svn://svn.chromium.org/blink/trunk@201788 bbb929c8-8fbe-4397-9dbb-9b2b20218538

If a CancellableTaskFactory is used by an Oilpan heap object, the
closure that the factory works with/over, cannot embed a reference
back to that object by way of an off-heap Persistent<> (WTF::Closure
is not on the heap.) If it does, such a reference will keep the heap
object alive, without it ever being released. Memory leaks are very likely.

This is too easy a slip-up to make with the current CancellableTaskFactory
constructor, so rephrase the constructor so as to make leaks no longer (easily)
possible. For Oilpan heap objects baked into the closure, the persistent reference
now held will be weak.

At the same time, take the opportunity to have this object no longer be a part object,
but a separate (off-heap) allocation. This lets us drop the ad-hoc ASan unpoisoning
support that was previously needed if CancellableTaskFactory was a part object
of an Oilpan heap object. Less magic.

R=haraken
BUG=

Review URL: https://codereview.chromium.org/1312843009

git-svn-id: svn://svn.chromium.org/blink/trunk@201787 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Just for consistency. This won't affect practical performance.

BUG=

Review URL: https://codereview.chromium.org/1317033008

git-svn-id: svn://svn.chromium.org/blink/trunk@201786 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch moves |VisibleSelection::ChangeObserver| out from |VisibleSelection|
class as |VisibleSelectionChangeObserver| as preparation of introducing
composed tree version of |VisibleSelection| since it doesn't depend on
DOM tree traversal and composed tree traversal.

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/1309413006

git-svn-id: svn://svn.chromium.org/blink/trunk@201785 bbb929c8-8fbe-4397-9dbb-9b2b20218538