GitLab

This patch gets rid of a |Position::deprecatedEditingOffset()| by replacing
them to |computeEditingOffset()|, since |deprecatedEditingOffset()| is just
calling a |computeEditingOffset()| to simplify code for improving code health.

BUG=513114
TEST=n/a; no behavior changes

Review URL: https://codereview.chromium.org/1300813002

git-svn-id: svn://svn.chromium.org/blink/trunk@200779 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=521538

Review URL: https://codereview.chromium.org/1301743002

git-svn-id: svn://svn.chromium.org/blink/trunk@200778 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/1301813002

git-svn-id: svn://svn.chromium.org/blink/trunk@200777 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=508984
TBR=jsbell@chromium.org

Review URL: https://codereview.chromium.org/1289023004 .

git-svn-id: svn://svn.chromium.org/blink/trunk@200776 bbb929c8-8fbe-4397-9dbb-9b2b20218538

As found while investigating crbug.com/503445, crash by assertion hit
may induce other innocent tests to crash (to be precise, look
like a crash) when an used object is collected by the garbage
collector and hit an assertion.

I've run independent tests repeatedly to identify which tests
causes similar failures and assigned bug for each.

They crash anyway, so they must be fixed, but to prevent other
innocent tests from crashing, mark them as "Skip" rather than
"Crash".

BUG=503445, 505364, 520616, 521907, 521913, 521914, 521918, 521929, 521936, 521937, 521941

Review URL: https://codereview.chromium.org/1298893002

git-svn-id: svn://svn.chromium.org/blink/trunk@200775 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The change was supposed to be made in three-sided patch:
[1] (blink) https://codereview.chromium.org/1262943002/
[2] (chromium) https://codereview.chromium.org/1254023006/
[3] (blink) https://codereview.chromium.org/1262603004/

But the change [1] needed to duplicate one header file (rather than
having a forwarding include) and a changes made between [1] and [2]
wasn't copied to the dup'ed file.

This change's making the same change made in the old location
to the new location.

I also added a FIXME comment so that we could keep the files in sync
until we land [2].

BUG=453032

Review URL: https://codereview.chromium.org/1294803005

git-svn-id: svn://svn.chromium.org/blink/trunk@200774 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Also, update bug numbers for other crashes in virtual/gpu/fast/canvas/.

BUG=454267,522315
TBR=leviw@chromium.org

Review URL: https://codereview.chromium.org/1295413003 .

git-svn-id: svn://svn.chromium.org/blink/trunk@200773 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This test was relying on animation events to function. This change
updates the test to use requestAnimationFrame to be more deterministic
and avoid timing out when running under threaded debug.

BUG=469683

Review URL: https://codereview.chromium.org/1292893005

git-svn-id: svn://svn.chromium.org/blink/trunk@200772 bbb929c8-8fbe-4397-9dbb-9b2b20218538

If the source canvas is not animated, no need to exit deferred rendering
because there is no risk of memory bloat due to snapshotting.

Fixed reson for original revert: m_snapshotState was not initialized

BUG=521001, 521732
NOTRY=true

Review URL: https://codereview.chromium.org/1289423003

git-svn-id: svn://svn.chromium.org/blink/trunk@200771 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Relayout during paint asserts.  There were a few cases where
the layout was actually dirty during paint, causing the
clientHeight() to relayout and crash.

Instead, we now cache the last known layout size in the media
element, and use it to compute things during paint.  It would be
much nicer if we could use CSS for this, but it doesn't seem to
be possible for the media shadow controls.  They all paint as
background images.

This may also be related to crbug.com/519495 .

BUG=520914

Review URL: https://codereview.chromium.org/1296333002

git-svn-id: svn://svn.chromium.org/blink/trunk@200770 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/1290893005

git-svn-id: svn://svn.chromium.org/blink/trunk@200769 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We only compare the 2nd attempt of a run to the tests expected results to
see if it was very flaky. Instead we should see if any of the runs had
an unexpected result. That way, any test that gets marked flaky will
stop getting printed out by this script as soon as the bots have cycled once.

Also, use TestExpectations.result_was_expected to determine if a result was
actually expected. The previous "in" check didn't work for cases like Text,
which are a subset of Failure instead of verbatim include in TestExpectations
as Text.

Finally, just use the length of the results entry to decide if something
is very flaky vs regular flaky. If there are <=2 entries, then it's not
very flaky. That catches the cases of a test that reliably fails as well
as the case of a test that fails the first attempt and runs as expected
on the second run (not necessarily passing on the second attempt, just
matching the listing in TestExpectations for that run).

R=joelo@chromium.org

Review URL: https://codereview.chromium.org/1289163002 .

git-svn-id: svn://svn.chromium.org/blink/trunk@200768 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Previously almost every layout test crashed with
--enable-slimming-paint-v2 when advancing document lifecycle state from
CompositingForSlimmingPaintV2Clean to InStyleRecalc or InPreLayout.

Review URL: https://codereview.chromium.org/1297943005

git-svn-id: svn://svn.chromium.org/blink/trunk@200767 bbb929c8-8fbe-4397-9dbb-9b2b20218538

. FocusController::setFocusedElement() assumes the focusedFrame() is
  local.
. ChromeClientImpl assumes webview client is non-null.

BUG=522260
TEST=none
R=japhet@chromium.org

Review URL: https://codereview.chromium.org/1298803003

git-svn-id: svn://svn.chromium.org/blink/trunk@200766 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The upstream implementation of promise_test() in web-platform-test has
changed to run the tests sequentially, removing the need for the
sequential_promise_test() helper. Tests now just use promise_test()
directly.

BUG=451338
R=ortuno@chromium.org

Review URL: https://codereview.chromium.org/1295373003

git-svn-id: svn://svn.chromium.org/blink/trunk@200765 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This lays the groundwork for doing transform and clip tree construction in one
place. It also cleans up the part of the interface which is exposed in the Blink
API, and separates range records from the tree itself.

BUG=470248

Review URL: https://codereview.chromium.org/1296963002

git-svn-id: svn://svn.chromium.org/blink/trunk@200764 bbb929c8-8fbe-4397-9dbb-9b2b20218538

LocalFrame::paintIntoDragImage() translates the content to be origin-
aligned, but leaves the picture bounds unchanged => the picture content
lies outside the declared picture cull rect.

We need to also shift the picture bounds, to keep them aligned with its
content.

R=junov@chromium.org,mtklein@chromium.org
BUG=522143

Review URL: https://codereview.chromium.org/1303453003

git-svn-id: svn://svn.chromium.org/blink/trunk@200763 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=521225

Review URL: https://codereview.chromium.org/1298283002

git-svn-id: svn://svn.chromium.org/blink/trunk@200762 bbb929c8-8fbe-4397-9dbb-9b2b20218538

There's no real reason to do all this work in multiple passes over the Vector
of Documents. This change does alter the web exposed behavior slightly since
now given two same origin iframes that are siblings in the frame tree:

   A
  / \
 B   C

when B runs if it looks at C's scroll position during a scroll animation and
SMIL animation state it may see slightly different results, especially if the
B frame actually started the animation in the C frame. This is already very
racing and unpredictable in browsers since it's an obscure edge case, so I'm
not too worried it'll break content.

This will allow easily skipping iframes that are still loading when processing
the blink pipeline.

BUG=521692

Review URL: https://codereview.chromium.org/1298253003

git-svn-id: svn://svn.chromium.org/blink/trunk@200761 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=521215
R=caseq

Review URL: https://codereview.chromium.org/1293273002

git-svn-id: svn://svn.chromium.org/blink/trunk@200760 bbb929c8-8fbe-4397-9dbb-9b2b20218538

used a RawPtrWillBeMember instead because Oilpan can deal with
these ownership cycles


BUG=382542

Review URL: https://codereview.chromium.org/1290683003

git-svn-id: svn://svn.chromium.org/blink/trunk@200759 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The value attribute should not be set for non-native text fields, according to HTML to Platform A11y Spec.

E.g. <div contenteditable<Some text.</div>. The node with role=SECTION should not have the value="Some text.".
BUG=491027
R=dmazzoni@chromium.org

Review URL: https://codereview.chromium.org/1289103002

git-svn-id: svn://svn.chromium.org/blink/trunk@200758 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=521200

Review URL: https://codereview.chromium.org/1300743002

git-svn-id: svn://svn.chromium.org/blink/trunk@200757 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=200742

BUG=455939
TBR=leviw@chromium.org

Review URL: https://codereview.chromium.org/1288273004 .

git-svn-id: svn://svn.chromium.org/blink/trunk@200756 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=520702

Review URL: https://codereview.chromium.org/1286423003

git-svn-id: svn://svn.chromium.org/blink/trunk@200755 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The test cases initially ran in parallel, but the web-platform-tests'
promise_test() switched to running cases serially for predictability,
which made these tests with many cases and lots of setup time out.

* Split up the match() and matchAll() tests into separate files
* Do the pre-population of the caches (within each test) in parallel

BUG=508984

Review URL: https://codereview.chromium.org/1295573002

git-svn-id: svn://svn.chromium.org/blink/trunk@200754 bbb929c8-8fbe-4397-9dbb-9b2b20218538

There are frames that should be merged together, but they are not
because some of them contain callUID (those coming from CPU profiler)
and those that don't contain it (those coming from trace event stacks).

Do not use callUID in the frame id calculation.

BUG=521473

Review URL: https://codereview.chromium.org/1294183002

git-svn-id: svn://svn.chromium.org/blink/trunk@200751 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is a reland of https://codereview.chromium.org/1290073003/

Reproduced crash locally, https://codereview.chromium.org/1302543002/ fixes it. So  after that lands land this again.

TBR=jsbell@chromium.org
BUG=502214

Review URL: https://codereview.chromium.org/1299903002

git-svn-id: svn://svn.chromium.org/blink/trunk@200749 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Add Unicode Variation Selector support on Linux official builds (patchset #1 id:1 of https://codereview.chromium.org/1283923002/ )

Reason for revert:
This CL prevents building official Chrome images on Linux.

../../third_party/WebKit/Source/platform/fonts/shaping/HarfBuzzFace.cpp:176:13: error: use of undeclared identifier 'hb_ot_font_set_funcs'

The proposed workaround doesn't work, so reverting this CL.

Original issue's description:
> Add Unicode Variation Selector support on Linux official builds
> 
> Most of the work to support Unicode Variation Selector was done in a
> separate CL[1], but it was not enabled for Linux due to crbug.com/462689.
> 
> This patch enables the Unicode Variation Selector support on Linux
> official builds.
> 
> The CL to enable on all Linux[2] was reverted because chromeos=1 is still
> bound to the harfbuzz in the system and thus unable to use
> hb_ot_font_set_funcs() on Trusty.
> 
> [1] https://codereview.chromium.org/1149693002/
> [2] https://codereview.chromium.org/1209043010/
> 
> BUG=383580
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=200427

TBR=eae@chromium.org,drott@chromium.org,kojii@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=383580

Review URL: https://codereview.chromium.org/1292583006

git-svn-id: svn://svn.chromium.org/blink/trunk@200747 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=200705

BUG=521801
TBR=fmalita@chromium.org

Review URL: https://codereview.chromium.org/1290223006 .

git-svn-id: svn://svn.chromium.org/blink/trunk@200746 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=521457
TBR=dgozman
NOTRY=true

Review URL: https://codereview.chromium.org/1301603003

git-svn-id: svn://svn.chromium.org/blink/trunk@200745 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Specifically, Vary headers should not prevent us from reusing an image that
already has been used in the current document.

BUG=518733,520376
TEST=ResourceFetcherTest.VaryImage
NOPRESUBMIT=true

Review URL: https://codereview.chromium.org/1287073007

git-svn-id: svn://svn.chromium.org/blink/trunk@200744 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The call to layout() could invalidate some of the null-checks we did previously.
Instead layout first, then do all the null checks.

BUG=521479

Review URL: https://codereview.chromium.org/1291343008

git-svn-id: svn://svn.chromium.org/blink/trunk@200743 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=csharrison
BUG=455939

Review URL: https://codereview.chromium.org/1300853003

git-svn-id: svn://svn.chromium.org/blink/trunk@200742 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Since CSSParserTokenRange represents a stream, calling something
like  size on it makes no sense. So change the logic in these
places to not call size by checking for end of the value list.

BUG=499780

Review URL: https://codereview.chromium.org/1225403010

git-svn-id: svn://svn.chromium.org/blink/trunk@200741 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Also, merge the data url path with the SubstitueData path, as
they do very similar work in emulating a cache hit.

BUG=521112

Review URL: https://codereview.chromium.org/1291883002

git-svn-id: svn://svn.chromium.org/blink/trunk@200738 bbb929c8-8fbe-4397-9dbb-9b2b20218538

A constant series of compositor updates would force the serviceOnNextFrame
to be kept being put off. The wakeAfter was always called with a fixed
duration so it would just keep getting cancelled and rescheduled.

Since the serviceOnNextFrame ends up calling setNeedsLayout inside the
scheduleAnimation callback; the LayoutTree would get updated in that call.
The hit testing problem identified in the bug was related to the fact that
the LayoutTree's layer transforms weren't updated correctly until it was
told it needed to do a Layout.

The fix is to check the remaining duration vs the new scheduled duration
and avoid the schedule if the new duration is longer.


BUG=513833

Review URL: https://codereview.chromium.org/1287013003

git-svn-id: svn://svn.chromium.org/blink/trunk@200737 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Previously for the following case:
 current display items:
    A B C D
 new display items:
    C B A D

A and B in current display items will be skipped and indexed when adding
C, then the index will be used when adding B and A.

When adding D, currentIt points to A, and B and C will be iterated
again.

Add nextItemToIndexIt variable to memorize the place to index, to avoid
repeated iteration.

BTW
- Change DisplayItem::ignoreInDisplayList() to (!)DisplayItem::isValid()
- Let currentIt point to the next item of the item just being copied so
  that we can synchronize sooner in some cases.

Review URL: https://codereview.chromium.org/1287863003

git-svn-id: svn://svn.chromium.org/blink/trunk@200736 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Message ports were using the source frame's execution context
rather than the targets when ports were transferred from a frame
to another. This seems to be leaky and also creates difficulties
when implementing message port support for android webview
and OOPIF. Use the target instead.

This is patch 1 of 3. The next CL will move chromium code to
use newly introduced method and the third CL will delete
the old methods.

BUG=473258

Review URL: https://codereview.chromium.org/1289033003

git-svn-id: svn://svn.chromium.org/blink/trunk@200735 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Confirmed that it is possible to construct a page that will produce a
negative maxHeight, as seen in:

fast/block/line-layout/negative-max-height.html

which produces maxHeight of -5.

Review URL: https://codereview.chromium.org/1296383005

git-svn-id: svn://svn.chromium.org/blink/trunk@200734 bbb929c8-8fbe-4397-9dbb-9b2b20218538