GitLab

BUG=

Review URL: https://codereview.chromium.org/710743005

git-svn-id: svn://svn.chromium.org/blink/trunk@185048 bbb929c8-8fbe-4397-9dbb-9b2b20218538

All recorded events are processed at once. It doesn't make sense to process them one by one now that we use tracing backend and there are no live updates while the profile is being recorded.

BUG=361045

Review URL: https://codereview.chromium.org/710113002

git-svn-id: svn://svn.chromium.org/blink/trunk@185047 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Since union container types have setters rather than assignment operators,
restructure the default value handling to generate a C++ expression that
sets the value instead of the value as a C++ literal expression.

And as a special case, since union containers are null initially, if the
default value is null, generate an empty expression then.

This also fixes a minor bug in the union handling: enum types were not
supported. They should be treated as string types, including conflicting
with any other string type in the union.

BUG=240176

Review URL: https://codereview.chromium.org/713683003

git-svn-id: svn://svn.chromium.org/blink/trunk@185046 bbb929c8-8fbe-4397-9dbb-9b2b20218538

So the problem is that in
StyleBuilderFunctions::applyValueCSSPropertyGridTemplateAreas we were assuming
that we were always getting a fresh RenderStyle, something that is not true,
because we might have overlapping declarations (multipled matching media
queries, user styles, etc).

As that function might be called multiple times, we need to properly regenerate
the list of named grid lines each time the function is called because the
NamedGridLinesMap returned by RenderStyle contains both the old explicit named
grid lines and the old implicit named grid lines.

BUG=427481

Review URL: https://codereview.chromium.org/680063005

git-svn-id: svn://svn.chromium.org/blink/trunk@185045 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Updates include:
   * Test that fragment part of URLs are ignored.
   * Add Cache.match tests for scenarios that were only being tested
     using Cache.matchAll. Currently all Cache.matchAll tests are
     failing due to the method not being implemented.
   * Add tests for URLs that contain embedded credentials.

BUG=374802,428970
R=jsbell, bkelly@mozilla.com

Review URL: https://codereview.chromium.org/704043002

git-svn-id: svn://svn.chromium.org/blink/trunk@185044 bbb929c8-8fbe-4397-9dbb-9b2b20218538

According to https://xhr.spec.whatwg.org/, sync XHRs on the main thread
have detrimental effects on the end user's experience and shall be removed.

R=vsevik

Review URL: https://codereview.chromium.org/685203003

git-svn-id: svn://svn.chromium.org/blink/trunk@185043 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Test is no longer reported as timing out on Windows bots.

R=tyoshino@chromium.org
BUG=409535
NOTRY=true

Review URL: https://codereview.chromium.org/711013002

git-svn-id: svn://svn.chromium.org/blink/trunk@185042 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We can drop a boolean from Document by adding a value to the ReferrerPolicy
enum to distinguish "default" from an explicitly-set "no referrer when
downgrade".

This patch also marginally increases the detail of the metrics we're
gathering ReferrerPolicy resets.

[1/2] Chromium: https://codereview.chromium.org/717503002/
[2/2] Blink: This patch.

BUG=431694

Review URL: https://codereview.chromium.org/698973008

git-svn-id: svn://svn.chromium.org/blink/trunk@185041 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=yhirano
BUG=431625

Review URL: https://codereview.chromium.org/713033002

git-svn-id: svn://svn.chromium.org/blink/trunk@185040 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=185017

BUG=405803
TBR=noel@chromium.org

Review URL: https://codereview.chromium.org/717523003

git-svn-id: svn://svn.chromium.org/blink/trunk@185039 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=185007

BUG=418521
TBR=chrishtr@chromium.org

Review URL: https://codereview.chromium.org/710753005

git-svn-id: svn://svn.chromium.org/blink/trunk@185038 bbb929c8-8fbe-4397-9dbb-9b2b20218538

For some reason the generated scrollbar pseudo element count was different
on Mac, although the test passed on the Mac bot when landing. I've removed
the test for :single-button/:double-button as the test coverage by the
other pseudo classes should be enough.

R=leviw@chromium.org
BUG=427953

Review URL: https://codereview.chromium.org/687943002

git-svn-id: svn://svn.chromium.org/blink/trunk@185036 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Now that Timeline uses tracing as its back-end it is time to remove old front-end implementation.

BUG=428700,344238,298188,359961,371276

Review URL: https://codereview.chromium.org/709423002

git-svn-id: svn://svn.chromium.org/blink/trunk@185035 bbb929c8-8fbe-4397-9dbb-9b2b20218538

JSONArray can contain thread-unsafe Strings and thus isn't safe to be contained in TracedValue.
However, TracedValue::setArray takes JSONArray as input.

This CL removes setArray and adds ScriptCallStack::toTracedValue() that uses other thread-safe methods in TracedValue instead.

BUG=427893

Review URL: https://codereview.chromium.org/688623002

git-svn-id: svn://svn.chromium.org/blink/trunk@185034 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Simply passing the RenderStyle(s) to computeLengthDouble, and letting it decide
how to convert units is ugly, since the conversion factors depends on context.
(Notably font-size, where parent specified font size is to be used rather than
current computed font size).

The same applies to zoom, where we normally want to use the effective zoom
of some RenderStyle, and in other cases some alternative specific zoom.

To make matters worse, the same RenderStyle from which we produce font unit
factors, and (maybe) effective zoom, is also marked as having viewport units (if
encountered), even though the RenderStyle we want to mark is not always the same
as the one we want to produce factors from.

This patch reworks CSSToLengthConversionData to contain more specific coversion
data, instead of just forwarding RenderStyles into dark places.

CSSToLengthConversionData now consists of:

 * CSSToLengthFontSizes which knows which factors to use for em/rem/ex/ch.
   The units that only depend on FontDescription (em, rem) are always computed,
   and the units which depend on FontMetrics (ex, ch) are computed on demand.
   This is to avoid loading of fonts in cases where its metrics weren't really
   needed after all.
 * CSSToLengthViewportSize just contains the size of the viewport, and exists
   only to improve the structure of CSSToLengthConversionData. (Less params).
 * Zoom. This is the zoom used in all cases. Effective/non-effective is now
   handled by the outside.
 * (Private) RenderStyle. Only to mark as having viewport units.

This is better, because:

 * The special situation of computing font-size is now handled in the converter
   for font-size, not computeLengthDouble, which really shouldn't have to care.
 * The RenderStyle we're passing is only used for viewport unit marking, so we
   can actually pass the correct one for font-size.
 * The ... extravagant ... constructors of CSSToLengthConversionData are now
   under control.

Notes:

 * StyleResolver now calls inheritFrom on styles before setting the style on the
   StyleResolverState. This ensures that correct font unit factors are computed
   (as the Font is inheritable).
 * An extra updateFont is needed if we inheritFrom a cached style.
 * StyleResolver::defaultStyleForElement does not need to use a
   StyleResolverState anymore. A FontBuilder will suffice.

Review URL: https://codereview.chromium.org/705783002

git-svn-id: svn://svn.chromium.org/blink/trunk@185033 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=425586

Review URL: https://codereview.chromium.org/693893002

git-svn-id: svn://svn.chromium.org/blink/trunk@185032 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=None
TEST=none; No behavior changes.

Review URL: https://codereview.chromium.org/712923002

git-svn-id: svn://svn.chromium.org/blink/trunk@185031 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=235436

Review URL: https://codereview.chromium.org/704193004

git-svn-id: svn://svn.chromium.org/blink/trunk@185030 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=427868
R=yurys, vsevik

Review URL: https://codereview.chromium.org/694753002

git-svn-id: svn://svn.chromium.org/blink/trunk@185029 bbb929c8-8fbe-4397-9dbb-9b2b20218538

mmap can return MAP_FAILED in some unlucky cases (e.g., a specified address was too large etc).
This CL adds a logic to handle a case where mmap returns MAP_FAILED.
Specifically, this CL tries mmap at most 10 times until mmap succeeds.

BUG=429911

Review URL: https://codereview.chromium.org/682993007

git-svn-id: svn://svn.chromium.org/blink/trunk@185028 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL makes DOMDataStore receive the v8::Isolate from the caller
(creator of DOMDataStore), which should better know what isolate is
to be used.

BUG=235436

Review URL: https://codereview.chromium.org/707213002

git-svn-id: svn://svn.chromium.org/blink/trunk@185027 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL corrects the counter values in the cases of allocation failure.

Review URL: https://codereview.chromium.org/710963002

git-svn-id: svn://svn.chromium.org/blink/trunk@185026 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=358074

Review URL: https://codereview.chromium.org/653283006

git-svn-id: svn://svn.chromium.org/blink/trunk@185025 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When the main thread passes a pointer of GarbageCollected object to a
HRTFDatabase thread (via postTask), the HRTFDatabase thread creates
a CrossThreadPersistent handle for the pointer. However, this is not
allowed because the HRTFDatabase thread is not attached to Oilpan.

We should attach the HRTFDatabase thread to Oilpan.

This CL removes MutexLocker from HRTFDatabase's destructor since we don't need to clear m_htrfDatabase in the destructor.

BUG=427303

Review URL: https://codereview.chromium.org/685153005

git-svn-id: svn://svn.chromium.org/blink/trunk@185024 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In the near future, Window.idl will be implemented as DOMWindow, not
LocalDOMWindow. Some of the custom Window bindings try to call
UsageCounter methods using Document as the reference, but DOMWindows
do not have a Document if they are remote.

BUG=425623

Review URL: https://codereview.chromium.org/707333003

git-svn-id: svn://svn.chromium.org/blink/trunk@185023 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Replaces ArrayBuffer with DOMArrayBuffer in core/ and modules/.

This CL is retry of http://crrev.com/708983002
Fixes Windows (64-bit) build.

BUG=235436

Review URL: https://codereview.chromium.org/711763002

git-svn-id: svn://svn.chromium.org/blink/trunk@185022 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=431643
TBR=senorblanco@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/710013002

git-svn-id: svn://svn.chromium.org/blink/trunk@185021 bbb929c8-8fbe-4397-9dbb-9b2b20218538

These tests consistently pass (see bug), and the bug assigned
to them was closed long ago.

TBR=dpranke@chromium.org
BUG=237270
NOTRY=true

Review URL: https://codereview.chromium.org/716483002

git-svn-id: svn://svn.chromium.org/blink/trunk@185020 bbb929c8-8fbe-4397-9dbb-9b2b20218538

These forward references are not needed anymore. After
this patch, core/rendering and core/painting has no unneeded
forward references any more.

This patch also move some forward declarations into correct
header files.

Review URL: https://codereview.chromium.org/713553002

git-svn-id: svn://svn.chromium.org/blink/trunk@185019 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Revert of Fix RemoteFrame detachment to signal its client. (patchset #1 id:1 of https://codereview.chromium.org/699053004/)

Reason for revert:
Reverting the revert. This turned out not to be the cause of https://code.google.com/p/chromium/issues/detail?id=431608

Original issue's description:
> Revert of Fix RemoteFrame detachment to signal its client. (patchset #1 id:1 of https://codereview.chromium.org/712653002/)
> 
> Reason for revert:
> Speculative revert to attempt to fix https://code.google.com/p/chromium/issues/detail?id=431608
> 
> Original issue's description:
> > Fix RemoteFrame detachment to signal its client.
> > 
> > We were seeing crashes in checkFrameCountConsistency() because during
> > RemoteFrame detachment the client was not getting signalled to modify
> > the frame tree, resulting in inconsistent state. This CL corrects that
> > problem.
> > 
> > BUG=430972
> > 
> > Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=184987
> 
> TBR=japhet@chromium.org,dcheng@chromium.org,kenrb@chromium.org
> NOTREECHECKS=true
> NOTRY=true
> BUG=430972
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=185015

TBR=japhet@chromium.org,dcheng@chromium.org,kenrb@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=430972

Review URL: https://codereview.chromium.org/709363002

git-svn-id: svn://svn.chromium.org/blink/trunk@185018 bbb929c8-8fbe-4397-9dbb-9b2b20218538

After issue 409435 fixes, both text and image baselines should be
better optimized across blink ports.

BUG=409435, 405803
TBR=dstockwell@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/691943003

git-svn-id: svn://svn.chromium.org/blink/trunk@185017 bbb929c8-8fbe-4397-9dbb-9b2b20218538

select() API is already supported, test cover scenarios for email, number, date, datetime-local, color and file.

While support for selection API does not exists.

BUG=415391
R=tkent, keishi

Review URL: https://codereview.chromium.org/707853002

git-svn-id: svn://svn.chromium.org/blink/trunk@185016 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Fix RemoteFrame detachment to signal its client. (patchset #1 id:1 of https://codereview.chromium.org/712653002/)

Reason for revert:
Speculative revert to attempt to fix https://code.google.com/p/chromium/issues/detail?id=431608

Original issue's description:
> Fix RemoteFrame detachment to signal its client.
> 
> We were seeing crashes in checkFrameCountConsistency() because during
> RemoteFrame detachment the client was not getting signalled to modify
> the frame tree, resulting in inconsistent state. This CL corrects that
> problem.
> 
> BUG=430972
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=184987

TBR=japhet@chromium.org,dcheng@chromium.org,kenrb@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=430972

Review URL: https://codereview.chromium.org/699053004

git-svn-id: svn://svn.chromium.org/blink/trunk@185015 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is in preparation for changing Window.idl ImplementedAs to
DOMWindow. These should be the remainder of the methods needed by V8
bindings.

BUG=425623

Review URL: https://codereview.chromium.org/710873003

git-svn-id: svn://svn.chromium.org/blink/trunk@185013 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The reason of that is a difference in URL for TempStorageSharedWorker between the host and the device front-ends
If DevTools window for a host page was opened before DevTools window for a device then the second DevTools
window will fail to run the shared worker and vice versa.

BUG=430794

Review URL: https://codereview.chromium.org/712593002

git-svn-id: svn://svn.chromium.org/blink/trunk@185012 bbb929c8-8fbe-4397-9dbb-9b2b20218538

original search implementation had no idea about flame chart and did search only for data grid nodes.

Also there was a single search box for both types, cpu and heap.

Two separate search view were introduced in the patch.
One for cpu view, the other for heap view.

New code for search in flame chart was written.

Original code for search in cpu data grid was moved to CPUProfileDataGrid.

HeapSnapshot search implementation was slightly tweaked.

BUG=425928

Review URL: https://codereview.chromium.org/696703003

git-svn-id: svn://svn.chromium.org/blink/trunk@185011 bbb929c8-8fbe-4397-9dbb-9b2b20218538

As getPropertyCSSValue is no longer web-exposed we shouldn't require it
in tests so we can delete the interface.

BUG=428595

Review URL: https://codereview.chromium.org/705373004

git-svn-id: svn://svn.chromium.org/blink/trunk@185010 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Intent to Remove:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/bka_hQuzAu0/tjD-GhJ6zDAJ

Attribute::setPrefix() was unused and showed up in grep.

BUG=431552

Review URL: https://codereview.chromium.org/688333002

git-svn-id: svn://svn.chromium.org/blink/trunk@185009 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This will be necessary once Push events support event.waitUntil to keep the
Service Worker alive until the Promise is resolved
(https://github.com/w3c/push-api/issues/85).

Depends on Chromium https://codereview.chromium.org/708913003
Required by Chromium https://codereview.chromium.org/707243003

BUG=430888, 401438

Review URL: https://codereview.chromium.org/709103002

git-svn-id: svn://svn.chromium.org/blink/trunk@185008 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The referenced bug is caused by a composited descendant that has an opacity
animation not having the correct background painted for its ancestor that it needs
to blend with.

However, HasOwnBackingButPaintsIntoAncestor is also just an optimization to reduce
the number of composited layers that draw content, in cases where we think we can
get away with it. (e.g. only existing to pass along preserves-3d state).

In exchange for these optimizations, the feature introduces significant complexity,
and has caused a number of bugs. Let's just kill it.

BUG=418521

Review URL: https://codereview.chromium.org/708213002

git-svn-id: svn://svn.chromium.org/blink/trunk@185007 bbb929c8-8fbe-4397-9dbb-9b2b20218538