GitLab

This is the third step in a sequence of patches to refactor
Blink accessibility enums to be able to use them in Chromium.

This step finishes renaming the public interface. Subsequent
changes will continue to rename and refactor Blink
accessibility code to be more consistent as well.

Step 1 (Blink): https://chromiumcodereview.appspot.com/22331005
Step 2 (Chromium): https://codereview.chromium.org/23651003

BUG=269034

Review URL: https://chromiumcodereview.appspot.com/23726007

git-svn-id: svn://svn.chromium.org/blink/trunk@157398 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is a subset of previously reviewed CL,
https://codereview.chromium.org/23531010/ that had to be reverted due to
unknown reasons for a performance regression. Hence, relanding in pieces.

R=abarth@chromium.org, tsepez@chromium.org

Review URL: https://codereview.chromium.org/23536035

git-svn-id: svn://svn.chromium.org/blink/trunk@157397 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When removing the old positioning behavior, the code that makes a float
with shape-outside gain a layer was left behind. This patch removes
that, so now floats with shape-outside are painted just like floats
without shape-outside.  Also remove the isFloatingWithShapeOutside
helper method because its use is now somewhat redundant, and the check
it does is now only needed in one place.

BUG=286448

Review URL: https://chromiumcodereview.appspot.com/23866007

git-svn-id: svn://svn.chromium.org/blink/trunk@157396 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch adds a use counter for animateColor.

BUG=286197

Review URL: https://chromiumcodereview.appspot.com/23857007

git-svn-id: svn://svn.chromium.org/blink/trunk@157395 bbb929c8-8fbe-4397-9dbb-9b2b20218538

I intend to review changes related to input events, scrolling/scaling,
viewport meta tag, fixed layout size, and Android-specific features
(I'll delegate review of other areas to other OWNERS).

NOTRY=true
BUG=

Review URL: https://chromiumcodereview.appspot.com/23731009

git-svn-id: svn://svn.chromium.org/blink/trunk@157394 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Measurements notes in the bug note that this is a reasonable
memory saving. Perf seems unimpacted by the smaller default.

BUG=285442
R=abarth@chromium.org

Review URL: https://codereview.chromium.org/23453040

git-svn-id: svn://svn.chromium.org/blink/trunk@157393 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When normalizing a block with anonymous blocks for first-letter,
before content, and contained text, a collapsing anonymous block
cascade is triggered that attempts to collapse the contained
text's anonymous block within its destruction method. To avoid
this, adding logic to bail early if we're already destroying
the child anonymous block we're trying to collapse.

While in the function, doing a little cleanup to make it more
obvious that it only operates on RenderBlocks, not RenderBoxes,
and updating some of the comments to hopefully be more useful.

BUG=282088

Review URL: https://chromiumcodereview.appspot.com/23463021

git-svn-id: svn://svn.chromium.org/blink/trunk@157392 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=leviw@chromium.org
BUG=285572

Review URL: https://codereview.chromium.org/23541008

git-svn-id: svn://svn.chromium.org/blink/trunk@157389 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently SerializedScriptValue's Writer is resizing its StringBuffer to the
exact size it needs for every object it writes into the buffer. Instead, it
should grow the buffer exponentially to avoid calling realloc so many times.

This CL fixes a regression from Blink@155225. I've also removed a call to
isolatedCopy which is unnecessary.

R=jsbell@chromium.org
BUG=266837

Review URL: https://chromiumcodereview.appspot.com/24043003

git-svn-id: svn://svn.chromium.org/blink/trunk@157388 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Amendment to CL: https://codereview.chromium.org/23483024

BUG=

Review URL: https://chromiumcodereview.appspot.com/23551010

git-svn-id: svn://svn.chromium.org/blink/trunk@157387 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This fixes Chromium test AwSettingsTest.testUseWideViewportLayoutWidth

BUG=285995

Review URL: https://chromiumcodereview.appspot.com/23459006

git-svn-id: svn://svn.chromium.org/blink/trunk@157386 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=ojan@chromium.org, leviw@chromium.org
BUG=285572

Review URL: https://codereview.chromium.org/23694024

git-svn-id: svn://svn.chromium.org/blink/trunk@157385 bbb929c8-8fbe-4397-9dbb-9b2b20218538

for synthesizing callbacks for memory cache hits.

BUG=none

Review URL: https://chromiumcodereview.appspot.com/24009002

git-svn-id: svn://svn.chromium.org/blink/trunk@157384 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This reverts commit ed002c01e93c3d9ee90fdf16d7461b149176c650.

Landing this as a speculative fix for the three perf regressions from
the listed bug.

TBR=ojan
BUG=286376

Review URL: https://chromiumcodereview.appspot.com/23701014

git-svn-id: svn://svn.chromium.org/blink/trunk@157383 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Until (1) a transaction ends or (2) a cursor hits the end of its
range, an IDBRequest holds on to an IDBCursor as its result. Per spec,
calling continue() or advance() on the cursor re-uses the same
IDBRequest, requiring a reference cycle.

Previously, the cycle was broken explicitly on either of those two
conditions, but until that time a cursor-request pair would "leak",
holding on to potentially large script value results.

This patch makes both classes RefCountedBase::deref() and check
to see if they have a partner object and both refcounts are 1. If
so, the cycle is broken.

Special case cruft for condition #1 is removed to simplify the
code - just rely on GC to reclaim the objects if necessary.

R=alecflett@chromium.org,dgrogan@chromium.org
BUG=225860

Review URL: https://chromiumcodereview.appspot.com/23653024

git-svn-id: svn://svn.chromium.org/blink/trunk@157382 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Turns out FrameView::isScrollable can sometimes crash. Reverting this change
while I try and figure out why.

This reverts commit 1a3edc542e31626a9b46ae2b18eb227c7031f3d6.

TBR=jamesr@chromium.org
BUG=280679

Review URL: https://chromiumcodereview.appspot.com/23477042

git-svn-id: svn://svn.chromium.org/blink/trunk@157381 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=261277

Review URL: https://chromiumcodereview.appspot.com/23711005

git-svn-id: svn://svn.chromium.org/blink/trunk@157379 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The last piece of cleanup.

Depends on chromium-side change (with ifdef):
https://codereview.chromium.org/23762002/

BUG=257349

Review URL: https://chromiumcodereview.appspot.com/23704004

git-svn-id: svn://svn.chromium.org/blink/trunk@157378 bbb929c8-8fbe-4397-9dbb-9b2b20218538

In rendering benchmarks we want to only capture trace events that are
required for the benchmark. console.time is used frequently, but everything
else from the 'webkit' category is not. Putting console.time into a separate
category 'webkit.console' will reduce overhead and make trace buffer overflows
less likely.

R=jamesr@chromium.org, nduca@chromium.org, tonyg@chromium.org
BUG=264308

Review URL: https://chromiumcodereview.appspot.com/23848006

git-svn-id: svn://svn.chromium.org/blink/trunk@157377 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch addresses half of crbug.com/225513 by grabbing the currently
active DOMWrapperWorld when creating ErrorEvent objects, and checking
that we're still in the same world when triggering `window.onerror`
handlers.

The next step will be to adjust event dispatching to ensure that 'error'
event listeners only trigger in the same world as the exception. I'll address
that in a separate patch; this patch adds several FAILing test results to make
that gap clear.

BUG=225513

Review URL: https://chromiumcodereview.appspot.com/21071003

git-svn-id: svn://svn.chromium.org/blink/trunk@157376 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This was regressed in r155659.

BUG=279919
TEST=manually tested with repro in bug

Review URL: https://chromiumcodereview.appspot.com/23640006

git-svn-id: svn://svn.chromium.org/blink/trunk@157375 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Have CharacterData constructor take a Document reference in argument instead
of a pointer as it can never be NULL. CharacterData subclasses (Comment,
Text), have been updated to use references as well.

BUG=281400

Review URL: https://chromiumcodereview.appspot.com/23754005

git-svn-id: svn://svn.chromium.org/blink/trunk@157373 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://chromiumcodereview.appspot.com/23861007

git-svn-id: svn://svn.chromium.org/blink/trunk@157372 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Autoscroll, as well as other user-driven scroll actions,
has to respect the scrollability styled into the web page.
More specifically, if a html or body tags are styled with
overflow:hidden, autoscroll should not scroll the containing document.

In order to fix this, patch hardens RenderBox::canAutoscroll as
following: previously, ::canAutoscroll was relying only on
::canBeScrolledAndHasScrollableArea to determine the scrollability
of #document node, which was unconditionally returned as 'true'.
Patch extends ::canAutoscroll to handle the #document case for
main and inner frames, and now it asks through ::isScrollable if
the corresponding document's FrameView is actually user-scrollable.

Note, that the patch changes ::canAutoscroll to cover the non-mainFrame
case now.

Autoscroll'ing the mainframe's document is hard with the current
EventSender machinary. Because of that, patch adds an iframe's document test case.
Test added in fast/events/autoscroll-in-overflow-hidden-html.html

Corresponding WebKit commit: <http://trac.webkit.org/changeset/154722>

BUG=chromium:1701

Review URL: https://chromiumcodereview.appspot.com/23450017

git-svn-id: svn://svn.chromium.org/blink/trunk@157371 bbb929c8-8fbe-4397-9dbb-9b2b20218538

DevTools: Add a fallback for console message callframe linkifying when scriptId could not be resolved.

No test is added as I was only able to reproduce it in DevTools on DevTools.

R=pfeldman

Review URL: https://chromiumcodereview.appspot.com/23956003

git-svn-id: svn://svn.chromium.org/blink/trunk@157370 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=152212

Review URL: https://chromiumcodereview.appspot.com/23135009

git-svn-id: svn://svn.chromium.org/blink/trunk@157369 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Table was created as inline renderer and another table block renderer wrapper is
created for table's content. So it was displayed as block element and all its style
properties also associated with inline renderer and was not applied to table.

Creating table inline-table renderer wrapper for table's content if parent
is inline except inline-block parent.

R=eseidel@chromium.org
BUG=53693

Review URL: https://chromiumcodereview.appspot.com/23628004

git-svn-id: svn://svn.chromium.org/blink/trunk@157368 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The parsing algorithm for the meta content attribute reports an error
for presense of ';'. It is currently using viewport error reporting, though
it is not tied to the viewport meta tag, but meta tags in general.

Review URL: https://chromiumcodereview.appspot.com/24025002

git-svn-id: svn://svn.chromium.org/blink/trunk@157367 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Introduce a shouldOverrideLegacyViewport method and use it instead
of comparing types manually.

Make the setViewportArguments method handle both legacy and @viewport,
instead of just the latter.

Now the code is a bit more self documenting, so reword some of the
comments.

Review URL: https://chromiumcodereview.appspot.com/23809007

git-svn-id: svn://svn.chromium.org/blink/trunk@157366 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=286287
TBR=

Review URL: https://codereview.chromium.org/23598007

git-svn-id: svn://svn.chromium.org/blink/trunk@157365 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Every table-section's rows height is calculating separately. So border spacing
had applied to bottom of the first section and top of the second section.

Now, Only Top most section would apply border spacing on top of it's section.

R=eseidel@chromium.org,jchaffraix@chromium.org
BUG=29502

Review URL: https://chromiumcodereview.appspot.com/23548017

git-svn-id: svn://svn.chromium.org/blink/trunk@157364 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We do not need them for many reasons:
- they are not used
- they are not even defined
- they could not work even if they were defined (after r157328)

Review URL: https://chromiumcodereview.appspot.com/23527003

git-svn-id: svn://svn.chromium.org/blink/trunk@157363 bbb929c8-8fbe-4397-9dbb-9b2b20218538

With high resolution sampling self/total times can be tenths of millisecond but FlameChart uses 1tick == 1ms scale when displaying time for hovered entry. This patch changes it to show tenths of ms.

BUG=None
R=loislo@chromium.org

Review URL: https://codereview.chromium.org/23924003

git-svn-id: svn://svn.chromium.org/blink/trunk@157362 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=278002

Review URL: https://chromiumcodereview.appspot.com/23290002

git-svn-id: svn://svn.chromium.org/blink/trunk@157361 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The problem happened because SimpleClassVectorTraits was used as a vector trait for WTF::String class, hence 'memcmp' was used for string vector's comparison, which is apparently wrong.
This CL corrects vector trait for WTF::String and also introduces a unit test to verify that the problem is not happening any more.

Review URL: https://chromiumcodereview.appspot.com/23819029

git-svn-id: svn://svn.chromium.org/blink/trunk@157360 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This reverts commit 4933abfbd2cdc88443cd0c57b8a90dd27ce48ec2.

BUG=284998

Review URL: https://chromiumcodereview.appspot.com/23818003

git-svn-id: svn://svn.chromium.org/blink/trunk@157359 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Headers like 'content-type' should be ignored for 304 responses, even if
they are delivered as 'Content-Type', or 'CoNtEnT-TyPe', etc.

BUG=246875

Review URL: https://chromiumcodereview.appspot.com/23591029

git-svn-id: svn://svn.chromium.org/blink/trunk@157358 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We are planning to refine XHR with responseType='blob'.
To test that the Inspector works with the new implementation,
we add the ability of specifying responseType to InspectorTest.makeXHR.

BUG=269055

Review URL: https://chromiumcodereview.appspot.com/23464028

git-svn-id: svn://svn.chromium.org/blink/trunk@157357 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is a subset of previously reviewed CL,
https://codereview.chromium.org/23531010/ that had to be reverted due to
unknown reasons for a performance regression. Hence, relanding in pieces.

R=abarth@chromium.org, tsepez@chromium.org

Review URL: https://codereview.chromium.org/23775006

git-svn-id: svn://svn.chromium.org/blink/trunk@157356 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch rename the enum and makes editing trigger a recalcStyle using the new enum value
DeprecatedRecalcStyleImmediatlelyForEditing. This will be slower, but editing causes so many
recalc styles and does so much work already I don't think this is going to be an issue.

This means we now lazy attach everywhere.

BUG=276689

Review URL: https://chromiumcodereview.appspot.com/23874007

git-svn-id: svn://svn.chromium.org/blink/trunk@157355 bbb929c8-8fbe-4397-9dbb-9b2b20218538