GitLab

Review URL: https://codereview.chromium.org/333073005

git-svn-id: svn://svn.chromium.org/blink/trunk@176496 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Drive-by: removed formatter for counters field which is not in the protocol anymore.

BUG=361045
R=loislo@chromium.org

Review URL: https://codereview.chromium.org/343613004

git-svn-id: svn://svn.chromium.org/blink/trunk@176495 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Required for instantiating the fallback fonts from the sandbox since
this instantiation is not allowed by filename, but only through
Skia's SkFontConfigInterface, where paths are previously kept track
of in the Linux sandbox support process.
Preparation for a Chromium side change which fills fciId.

BUG=285303
R=abarth,eae

Review URL: https://codereview.chromium.org/340073002

git-svn-id: svn://svn.chromium.org/blink/trunk@176494 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=361045
R=loislo@chromium.org

Review URL: https://codereview.chromium.org/339493006

git-svn-id: svn://svn.chromium.org/blink/trunk@176493 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Border-image and its friend mask-box-image are shorthands that
expand out into 4 css properties[1]. When setting the initial values
for these we would blindly create a new NinePieceBorderImage which
is large and contains non-trivial destructors. This patch first
checks whether the current value is already equal to the initial value
before setting.

[1] https://developer.mozilla.org/en-US/docs/Web/CSS/border-image

Review URL: https://codereview.chromium.org/340703004

git-svn-id: svn://svn.chromium.org/blink/trunk@176492 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch renames abbreviated name "p" to "editablePosition" in {first,last}EditableVisiblePosition{After,Before}PositionInRoot for improving readability and ease of maintenance.

This patch is a preparation of introducing |lastEditablePositionBeforePositionInRoot.

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

Review URL: https://codereview.chromium.org/342553008

git-svn-id: svn://svn.chromium.org/blink/trunk@176491 bbb929c8-8fbe-4397-9dbb-9b2b20218538

HarfBuzzRunCache is a global cache with entries holding |Font| which indirectly hold strong ref to |SVGFontFace|.
In future, the leak detector should clear the cache before counting number of |Document|s.

These "leaks" disappear if following line is commented out:
WebKit/Source/platform/fonts/harfbuzz/HarfBuzzShaper.cpp
861         runCache.insert(key, new CachedShapingResults(harfBuzzBuffer.get(), m_font, props.direction, localeString));

BUG=366455
TBR=haraken
NOTRY=true

Review URL: https://codereview.chromium.org/342153002

git-svn-id: svn://svn.chromium.org/blink/trunk@176490 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The spec mandates that transition-property (and thus the transition
shorthand) are able to support arbitrary unknown properties, for
backwards and forwards compatibility. This patch adds support for such
unknown properties.

Note that this makes the shorthand strings like "ease" or "ease ease-in"
ambiguous. In this patch we take the first case to refer to a timing
function, and in the second case the first token as a timing function
and the second token as the property name.

This change progresses the w3c transitions test transition-property-002,
although the first two cases of the test are incorrect.

BUG=304020,375098

Review URL: https://codereview.chromium.org/298043007

git-svn-id: svn://svn.chromium.org/blink/trunk@176489 bbb929c8-8fbe-4397-9dbb-9b2b20218538

https://codereview.chromium.org/200923002/diff/380001/Source/core/loader/ImageLoader.cpp
moved a document.isActive() check elsewhere, leading to us doing more work
for non-active documents. This caused a big perf regression for the bloat test.

This patch restores that check to fix the regression.

BUG=372971
R=japhet@chromium.org
CC=yoav@yoav.ws

Review URL: https://codereview.chromium.org/343513003

git-svn-id: svn://svn.chromium.org/blink/trunk@176488 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The WebThreadedDataReceiver will now check if the parserthread is still alive before returning it to the ThreadedDataProvider

This fixes a crash when a renderer is shut down while a request is still alive, due to the ResourceDispatcher being shut down after Blink is.

R=abarth
BUG=386383

Review URL: https://codereview.chromium.org/345663002

git-svn-id: svn://svn.chromium.org/blink/trunk@176487 bbb929c8-8fbe-4397-9dbb-9b2b20218538

There's no reason this would ever be called on remote frames.

BUG=386340

Review URL: https://codereview.chromium.org/340213002

git-svn-id: svn://svn.chromium.org/blink/trunk@176486 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch removes a style recalc that occurred every time
element.animate() was being called regardless of the input.
This is no longer needed now that we are able to defer keyframe
resolution the style recalc using DeferredLegacyStyleInterpolation.

Review URL: https://codereview.chromium.org/343563002

git-svn-id: svn://svn.chromium.org/blink/trunk@176485 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Unnecessary delay is removed

Use js-test.js instead of manual logging, asserts

BUG=27569

Review URL: https://codereview.chromium.org/344533006

git-svn-id: svn://svn.chromium.org/blink/trunk@176484 bbb929c8-8fbe-4397-9dbb-9b2b20218538

'innerText' is super-confusing because we have Element::innerText and
Element::setInnerText.

BUG=
TEST=none; no behavior changes.

Review URL: https://codereview.chromium.org/340273003

git-svn-id: svn://svn.chromium.org/blink/trunk@176483 bbb929c8-8fbe-4397-9dbb-9b2b20218538

r175792 did a good job on refactoring this for the most parts, but it didn't do
so well by leaving behind the old classes only for NavigatorGamepad. This CL
trasitions NavigatorGamepad to the new base classes and and removes the code
duplication.
Also removeEventListener fixed not to assume that it was the last listener.

Review URL: https://codereview.chromium.org/336693004

git-svn-id: svn://svn.chromium.org/blink/trunk@176482 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=373690

Review URL: https://codereview.chromium.org/342683002

git-svn-id: svn://svn.chromium.org/blink/trunk@176481 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=381876

TBR=eseidel

Review URL: https://codereview.chromium.org/343733002

git-svn-id: svn://svn.chromium.org/blink/trunk@176479 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The semicolon-separated case must come first to prevent it from
being handled in the ordinary manner when the string starts with
javascript:, since it need not obey the normal termination rules
when it is first split by semicolons.

BUG=384077

Review URL: https://codereview.chromium.org/346623003

git-svn-id: svn://svn.chromium.org/blink/trunk@176478 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Copy/Paste&Find/Replace of r176370 - same thing for ByteString.

Not strictly needed by any existing IDLs, but trivial code change,
and no code size difference until used.

R=horo@chromium.org,haraken@chromium.org
BUG=379009

Review URL: https://codereview.chromium.org/344693002

git-svn-id: svn://svn.chromium.org/blink/trunk@176477 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=tkent

Review URL: https://codereview.chromium.org/342553007

git-svn-id: svn://svn.chromium.org/blink/trunk@176476 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Rename {first,last}EditablePosition{After,Before}PositionInRoot to denote them returned VisiblePosion

This patch renames {first,last}EditablePosition{After,Before}PositionInRoot to
{first,last}EditableVisiblePosition{After,Before}PositionInRoot for having
lastEditblePositionBeforePositionInRoot() which returns Position instead of
VisiblePosition.

This patch is a preparation of fixing issue 339187 by
http://crrev.com/340713003

BUG=n/a
TEST=n/a

Review URL: https://codereview.chromium.org/340013002

git-svn-id: svn://svn.chromium.org/blink/trunk@176475 bbb929c8-8fbe-4397-9dbb-9b2b20218538

/shadow-deep/ -> /deep/
and one minor comment correction.

BUG=none
TEST=none

Review URL: https://codereview.chromium.org/334833002

git-svn-id: svn://svn.chromium.org/blink/trunk@176474 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL adds histogram statistics for the following
web components-related CSS selectors.

- :unresolved pseudo-class (for custom element)
- ::shadow pseudo element
- ::content pseudo element
- :host pseudo class
- :host-context() pseudo class
- /deep/ combinator

BUG=383749
TEST=open http://jsbin.com/bopezaso/2/edit, reload several times and see if FeatureObserver in chrome://histograms reflects it (number 465-470)

Review URL: https://codereview.chromium.org/339563004

git-svn-id: svn://svn.chromium.org/blink/trunk@176473 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It's a little opaque to say "if (fontDescription.variant())".

Review URL: https://codereview.chromium.org/343813002

git-svn-id: svn://svn.chromium.org/blink/trunk@176472 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=373917
TBR=abarth

Review URL: https://codereview.chromium.org/345593004

git-svn-id: svn://svn.chromium.org/blink/trunk@176471 bbb929c8-8fbe-4397-9dbb-9b2b20218538

TBR=haraken
NOTRY=true
BUG=364377

Review URL: https://codereview.chromium.org/342883002

git-svn-id: svn://svn.chromium.org/blink/trunk@176470 bbb929c8-8fbe-4397-9dbb-9b2b20218538

MSVC treats their in-class initializations as definitions, but other
compilers don't. This updates the #ifdef to handle the Clang on Windows
case.

BUG=82385
TEST=build blink_web.dll with Clang on Windows in Release mode

Review URL: https://codereview.chromium.org/330823006

git-svn-id: svn://svn.chromium.org/blink/trunk@176469 bbb929c8-8fbe-4397-9dbb-9b2b20218538

On non-Windows, PLATFORM_EXPORT expands to a visibility attribute,
which is inherited by nested classes. On Windows, it's expanded to
a dllexport/dllimport attribute, which is not inherited.

When using Clang on Windows, MediaStreamComponent's dtor got
inlined, which destroys m_sourceProvider, whose destructor got inlined,
and tried to reference provideInput for the vtable, but failed to link
since it wasn't in the right dll. Exporting the inner class fixes this.

BUG=82385
TEST=build blink_web.dll with Clang on Windows in Release mode

Review URL: https://codereview.chromium.org/343763002

git-svn-id: svn://svn.chromium.org/blink/trunk@176468 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=385313
TBR=bokan@chromium.org
NOTRY=True

Review URL: https://codereview.chromium.org/338373006

git-svn-id: svn://svn.chromium.org/blink/trunk@176467 bbb929c8-8fbe-4397-9dbb-9b2b20218538

and reconfiguring the squashing GraphicsLayer.

Previously we were issuing them only after.

BUG=385103

Review URL: https://codereview.chromium.org/345653002

git-svn-id: svn://svn.chromium.org/blink/trunk@176466 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Skia's save() SaveFlags bits are going the way of the dodo. At this
point all callers have been updated to always pass kMatrixClip, but
SkCanvas subclasses also need to use the new willSave() API.

(depends on https://codereview.chromium.org/338833003/ rolling)

BUG=skia:2297
R=reed@google.com,senorblanco@chromium.org,malch@chromium.org

Review URL: https://codereview.chromium.org/338863002

git-svn-id: svn://svn.chromium.org/blink/trunk@176465 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This code is unused.

R=pfeldman@chromium.org

Review URL: https://codereview.chromium.org/333153003

git-svn-id: svn://svn.chromium.org/blink/trunk@176464 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Before this patch, font load was noticed to CSSFontSelector through the
following path:

 RemoteFontFaceSource
 -> CSSFontFace
 -> CSSSegmentedFontFace
 -> CSSFontSelector

However, StyleEngine clears FontFaceCache upon stylesheet changes and
reconstruct it lazily, it is possible that CSSSegmentedFontFace is
absent when a load is completed. As a result, loaded font is not
rendered.

This patch fixes the problem by changing the notification path as follows:
 RemoteFontFaceSource -> FontLoader -> CSSFontSelector.

BUG=378238
TEST=fast/css/font-face-css-change-while-loading.html

Review URL: https://codereview.chromium.org/321403003

git-svn-id: svn://svn.chromium.org/blink/trunk@176463 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The embedders should now use WebScreenOrientationClient instead.

BUG=None

Review URL: https://codereview.chromium.org/339343003

git-svn-id: svn://svn.chromium.org/blink/trunk@176462 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Inline styles that are added to a page from an isolated world should bypass the
main world's CSP, much like how scripts do. As an example, this is important for
extensions to make sure they can bypass the page's CSP when they inject a style
tag into the page.

R=mkwst@chromium.org
BUG=385246

Review URL: https://codereview.chromium.org/341443003

git-svn-id: svn://svn.chromium.org/blink/trunk@176461 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/344693003

git-svn-id: svn://svn.chromium.org/blink/trunk@176460 bbb929c8-8fbe-4397-9dbb-9b2b20218538

the RenderFrame and WebFrame has 1:1 mapping now

Review URL: https://codereview.chromium.org/339133004

git-svn-id: svn://svn.chromium.org/blink/trunk@176459 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/337113005

git-svn-id: svn://svn.chromium.org/blink/trunk@176458 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Correct the squashing origin w.r.t. the transformed ancestor to take into account rounding to integer
offsets.

Review URL: https://codereview.chromium.org/345613004

git-svn-id: svn://svn.chromium.org/blink/trunk@176457 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Prior to this patch, a replaced element with intrinsic ratio but with
no intrinsic width nor height and with 'width', 'height' both computed
values of 'auto', would have preferred width and preferred minimal
width that depended on the available width of the container.

After this patch, such a replaced element will have zero preferred
width and zero preferred minimal width.

Example:

<!doctype html>
<div style="width: 200px">
  <svg viewBox="0 0 1 1" style="background: blue">
  </svg>
<div>

<svg> becomes a 200x200 blue rectangle, intrinsic ratio (1) comes from
viewBox attribute. Preferred minimal width and preferred width would
be 200px prior to patch.

Changing the width of <div> would require recalculating preferred
widths on <svg>, something we rather not do, and does not currently
do. (See RenderReplaced::needsPreferredWidthsRecalculation()). This
means preferred minimal widths was left with stale values after
changes to the container.

The new behavior matches Firefox but is not yet in spec, possibly
because the sizing of such replaced element is undefined in CSS 2.1,
even though there is a suggestion, followed by Blink, in the spec.

BUG=382922

Review URL: https://codereview.chromium.org/325183002

git-svn-id: svn://svn.chromium.org/blink/trunk@176456 bbb929c8-8fbe-4397-9dbb-9b2b20218538