GitLab

BUG=327641

Review URL: https://codereview.chromium.org/306123005

git-svn-id: svn://svn.chromium.org/blink/trunk@175195 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175189

BUG=370664
TBR=chrishtr@chromium.org

Review URL: https://codereview.chromium.org/304283009

git-svn-id: svn://svn.chromium.org/blink/trunk@175194 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Allow sub-css-pixel font-sizes and thus smooth font scaling which in
turn allows for stable layout when zooming and drastically improves
readability and font rendering quality on high dpi devices.

This change enabled the experimental subpixel font scaling flag by
default across all supported platforms.

Entry in Chromium Dashboard
  http://www.chromestatus.com/features/6259163214643200

OWP launch tracking bug
  https://code.google.com/p/chromium/issues/detail?id=164167

Intent to ship
  https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/4-f2i7OEWeg

R=eseidel@chromium.org, dglazkov@chromium.org
BUG=164167

Review URL: https://codereview.chromium.org/307993009

git-svn-id: svn://svn.chromium.org/blink/trunk@175192 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is by far the slowest part of doing rebaselines.
We can do everything but the SCM commands in parallel.
Like we do for the other commands, save off the SCM commands
and do them all at the end.

Review URL: https://codereview.chromium.org/308793004

git-svn-id: svn://svn.chromium.org/blink/trunk@175191 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Windows XP specific rebaselines for r175038 as the XP > Win7 fallback is
no longer valid for text tests.

TBR=dglazkov@chromium.org

Review URL: https://codereview.chromium.org/309503009

git-svn-id: svn://svn.chromium.org/blink/trunk@175190 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This fixes a mistake in crrev.com/301843002

Includes a test.

BUG=370664

Review URL: https://codereview.chromium.org/306093002

git-svn-id: svn://svn.chromium.org/blink/trunk@175189 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175124

BUG=372502
TBR=h.joshi@samsung.com

Review URL: https://codereview.chromium.org/302223002

git-svn-id: svn://svn.chromium.org/blink/trunk@175188 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=174989

BUG=376306
TBR=wangxianzhu@chromium.org

Review URL: https://codereview.chromium.org/307163002

git-svn-id: svn://svn.chromium.org/blink/trunk@175187 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175010

BUG=314704
TBR=dtrebbien@gmail.com

Review URL: https://codereview.chromium.org/304413003

git-svn-id: svn://svn.chromium.org/blink/trunk@175186 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=174988

BUG=369358
TBR=wangxianzhu@chromium.org

Review URL: https://codereview.chromium.org/306223002

git-svn-id: svn://svn.chromium.org/blink/trunk@175185 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175147

BUG=377610
TBR=fs@opera.com

Review URL: https://codereview.chromium.org/309553006

git-svn-id: svn://svn.chromium.org/blink/trunk@175184 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Previously we use localToContainerPoint((0,0)) to calculate the offset
of a child when drawing focus rings. If the child is rotated, (0, 0)
won't be its top-left corner in container‘s coordinations.

Now use localToContainerQuad(childLocalFocusRingRect).enclosingBoundingBox().

Also updated RenderObject::absoluteFocusRingQuads() to be
transform-aware.

BUG=374168
TEST=transforms/transform-focus-ring.html

Review URL: https://codereview.chromium.org/288013005

git-svn-id: svn://svn.chromium.org/blink/trunk@175183 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Based on dsinclair@'s original patch https://codereview.chromium.org/174073002/

For the new world of text based repaint tests it can be hard
to visualize how the repaint rects appear on the screen. This
is a first pass at visualizing the difference between the old
and new repaint rects when one of the text based repaint tests
fail.

Review URL: https://codereview.chromium.org/303223008

git-svn-id: svn://svn.chromium.org/blink/trunk@175182 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=174983

BUG=164167
TBR=eae@chromium.org

Review URL: https://codereview.chromium.org/308913003

git-svn-id: svn://svn.chromium.org/blink/trunk@175181 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=174992

BUG=164167
TBR=eae@chromium.org

Review URL: https://codereview.chromium.org/304353011

git-svn-id: svn://svn.chromium.org/blink/trunk@175180 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=376506

Review URL: https://codereview.chromium.org/305723002

git-svn-id: svn://svn.chromium.org/blink/trunk@175179 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Windows XP specific rebaselines for r175038 as the XP > Win7 fallback is
no longer valid for text tests.

TBR=dglazkov@chromium.org

Review URL: https://codereview.chromium.org/303163007

git-svn-id: svn://svn.chromium.org/blink/trunk@175178 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Replaced getShapeImageAndRect() in ShapeOutsideInfo.cpp with
ShapeOutsideInfo::createShapeForImage(). The original function
failed to protect the PassRefPtr return value returned by StyleImage::image(). 

BUG=378469

Review URL: https://codereview.chromium.org/305173003

git-svn-id: svn://svn.chromium.org/blink/trunk@175176 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL can't solve all issues when the source position approaches the listener position because when the locations are equal, we get a singular point in the equations.

This just makes it continuous as the source approaches the listener. If the source then goes past, there will be a unavoidable discontinuity.

TEST=Listen to http://jsfiddle.net/UxwGP/10. When the slider position is 0, there should be no sudden change in pitch.

TEST=Listen to http://jsfiddle.net/qWHAt/34/ using mono source and HRTF panner.  Initially, the pitch is very low. Move the slider just to the left or right of center and the pitch goes up. As you move the slider through center slowly, the pitch drops drastically at center.  With this CL, this the pitch does not start out low, and as you move the slider around, the pitch never drops low at center. There is a discontinuity moving from left to right, but this is because the source was moving away from the listener and is not moving towards the listener.

BUG=141739

Review URL: https://codereview.chromium.org/301733002

git-svn-id: svn://svn.chromium.org/blink/trunk@175175 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This should fix an uncommon crash during history traversal.

I believe the steps to get into this state are:
1. Load a page with a slow-loading iframe. The iframe load must begin before the main frame's load event fires.
2. While the iframe is still in the provisional load state, attempt a same-document history navigation in the child frame.
3. The child frame's provisional load is cancelled by the history navigation in FrameLoader::loadInSameDocument, which in turn causes the parent frame's load event to fire synchronously.
4. The parent frame's onload event handler detaches the iframe.
5. No checks are performed after cancelling the provisional load in FrameLoader::loadInSameDocument, leading to a null deref and crash.

We should be able to prevent a crash in this case by checking whether the frame is still attached after cancelling the provisional load in FrameLoader::loadInSameDocument.

BUG=374391

Review URL: https://codereview.chromium.org/303133004

git-svn-id: svn://svn.chromium.org/blink/trunk@175174 bbb929c8-8fbe-4397-9dbb-9b2b20218538

MSVC doesn't set __SSE__ or__SSE2__ defines:

http://msdn.microsoft.com/en-us/library/b0084kay.aspx

BUG=349320
TEST=compiles

Review URL: https://codereview.chromium.org/307963005

git-svn-id: svn://svn.chromium.org/blink/trunk@175173 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Combined RenderStyle::needsFullRepaintOnContainer(Width|Height)Change
and local function mustRepaintFillLayers() in RenderObject.cpp into
RenderObject::mustRepaintFillLayersOn(Width|Height)Change.

Added missing logics of the original mustRepaintFillLayers():
- repaint if background-repeat is not repeat and no-repeat;
- repaint if background-position is not left or top origin;
- don't repaint if background-position is of fixed length;

TEST=fast/repaint/background-resize-width.html
TEST=fast/repaint/background-resize-height.html

Review URL: https://codereview.chromium.org/306483011

git-svn-id: svn://svn.chromium.org/blink/trunk@175172 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When the only things that changed are not tree boundary crossing rules we
can skip invalidation in descendant ShadowRoots. This means that
if you have lots of nested components with duplicate class names then using
a descendant selector in an enclosing component won't invalidate style in
nested ShadowRoots.

I also packed the booleans on DescendantInvalidationSet since we keep growing
it with new flags.

In the future we can make this tree boundary crossing logic even smarter so that
if you have both tree boundary and not tree boundary crossing invalidation
sets on the same subtree we can skip running the non-tree boundary crossing
rules when recursing into a ShadowRoot.

I also took this as an opportunity to clean up the addFeaturesToInvalidationSets
logic and merged the new treeBoundaryCrossing flag and the wholeSubtreeInvalid
flag into the InvalidationSetFeatures.

Review URL: https://codereview.chromium.org/286903024

git-svn-id: svn://svn.chromium.org/blink/trunk@175171 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=174930

BUG=373240
TBR=dominik.rottsches@intel.com

Review URL: https://codereview.chromium.org/309613004

git-svn-id: svn://svn.chromium.org/blink/trunk@175170 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Make importNode work when the XML document contains a CDATA section
by avoiding calling createCDATASection since createCDATASection
throws an exception for HTML Documents.

The new behavior matches Firefox.

BUG=340059

Review URL: https://codereview.chromium.org/167653002

git-svn-id: svn://svn.chromium.org/blink/trunk@175169 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This will allow us to remove SK_SUPPORT_LEGACY_IMAGEGENERATORAPI

BUG=

Review URL: https://codereview.chromium.org/304353008

git-svn-id: svn://svn.chromium.org/blink/trunk@175168 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Right now it's unique per view, but as part of converting the content side code to use frames instead of views in https://codereview.chromium.org/303503008/ we need this to be unique per view. That can change back in the future once all the permission-related code in the browser (PermissionRequestID in particular) is converted.

BUG=304341

Review URL: https://codereview.chromium.org/301183005

git-svn-id: svn://svn.chromium.org/blink/trunk@175167 bbb929c8-8fbe-4397-9dbb-9b2b20218538

With this change Android will stop using the old discardable memory path
and will use Skia's implementation of discardable memory and caching.

Tested on Android with ChromeShell.

BUG=169282

Review URL: https://codereview.chromium.org/303033004

git-svn-id: svn://svn.chromium.org/blink/trunk@175166 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When running a repaint test in browser, sometimes we may want to see
the page contents before and after repaintTest(). The origianl 100ms
delay seems too short to show the page contents before repaintTest().

This may be more useful when we have repaint overlay feature when
showing layout test results (https://codereview.chromium.org/303223008/).

Review URL: https://codereview.chromium.org/304273005

git-svn-id: svn://svn.chromium.org/blink/trunk@175165 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch removes the semi-transparent color used int the test, which
was irrelevant to the feature being testes and was causing the test
to fail due to rounding errors in compositing.

BUG=368008

Review URL: https://codereview.chromium.org/302083006

git-svn-id: svn://svn.chromium.org/blink/trunk@175164 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The mapToContainer method is calculating the value in relation
to the container, but when we verify the result we calculate
the localToAbsolute value. This doesn't take the container
into account so we get the wrong values.

BUG=371640

Review URL: https://codereview.chromium.org/310483002

git-svn-id: svn://svn.chromium.org/blink/trunk@175163 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Since the lifetimes of HTMLImportTreeNode and all of its
descendants are inextricably linked, it makes sense for
all ownership responsibilities to be transfered from
HTMLImportsController to HTMLImportTreeNode.

No functional changes, refactoring.

R=morrita
BUG=

Review URL: https://codereview.chromium.org/305723004

git-svn-id: svn://svn.chromium.org/blink/trunk@175162 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Spec: http://www.whatwg.org/specs/web-apps/current-work/multipage/timers.html#dom-navigator-languages

BUG=365123

Review URL: https://codereview.chromium.org/288393002

git-svn-id: svn://svn.chromium.org/blink/trunk@175161 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Create the compositor in WebPopupMenuImpl::initialize because it's always going to be used.

BUG=378029

Review URL: https://codereview.chromium.org/304273002

git-svn-id: svn://svn.chromium.org/blink/trunk@175160 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Avoid redundant initialization of GraphicsContextState. (https://codereview.chromium.org/303613002/)

Reason for revert:
This change is the cause of a
leak in PhishingDOMFeatureExtractorTest.ScriptAndImageFeatures
under LSAN. See
http://build.chromium.org/p/chromium.memory/builders/Linux%20ASan%20LSan%20Tests%20%281%29/builds/2642/steps/browser_tests/logs/ScriptAndImageFeatures

Original issue's description:
> Avoid redundant initialization of GraphicsContextState.
> 
> When creating a new GraphicsContextState the constructor initializes the state
> with various default values. In GraphicsContext::realizePaintState we immediately
> overwrite these values after creation. This patch adds a copy constructor to
> GraphicsContextState to avoid that.
> 
> Even though this seems like a clear win, the benefits are not visible when
> profiling. Looks like the overhead of the redundant initialization is
> not very large. Still seems worth fixing though.
> 
> BUG=377687
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=174985

TBR=schenney@chromium.org,fmalita@chromium.org,dominikg@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=377687

Review URL: https://codereview.chromium.org/303403006

git-svn-id: svn://svn.chromium.org/blink/trunk@175159 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Revert of Prevent <canvas> destructor from causing nested garbage collection (https://codereview.chromium.org/307013006/)

Reason for revert:
This isn't the right CL, was off-by-one.

Original issue's description:
> Revert of Prevent <canvas> destructor from causing nested garbage collection (https://codereview.chromium.org/303803002/)
> 
> Reason for revert:
> This change is the cause of a
> leak in PhishingDOMFeatureExtractorTest.ScriptAndImageFeatures
> under LSAN. See
> http://build.chromium.org/p/chromium.memory/builders/Linux%20ASan%20LSan%20Tests%20%281%29/builds/2642/steps/browser_tests/logs/ScriptAndImageFeatures
> 
> Original issue's description:
> > Prevent <canvas> destructor from causing nested garbage collection
> > 
> > In the destruction of HTMLCanvasElement, we call
> > HTMLCanvasElement::existingDrawingContext(), whose behavior was changed
> > incorrectly in r162330 and was triggering the allocation of an ImageBuffer
> > during tear-down.  This is not only wasteful, but it can trigger a
> > garbage collection because the creation of an ImageBuffer will increase
> > the amount of externallay allocated memory that is declared to V8.
> > Triggering a GC during the finalization of a garbage collected object
> > is bad.
> > 
> > BUG=378289
> > 
> > Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=174984
> 
> TBR=senorblanco@chromium.org,junov@chromium.org
> NOTREECHECKS=true
> NOTRY=true
> BUG=378289
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175157

TBR=senorblanco@chromium.org,junov@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=378289

Review URL: https://codereview.chromium.org/301243009

git-svn-id: svn://svn.chromium.org/blink/trunk@175158 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Prevent <canvas> destructor from causing nested garbage collection (https://codereview.chromium.org/303803002/)

Reason for revert:
This change is the cause of a
leak in PhishingDOMFeatureExtractorTest.ScriptAndImageFeatures
under LSAN. See
http://build.chromium.org/p/chromium.memory/builders/Linux%20ASan%20LSan%20Tests%20%281%29/builds/2642/steps/browser_tests/logs/ScriptAndImageFeatures

Original issue's description:
> Prevent <canvas> destructor from causing nested garbage collection
>
> In the destruction of HTMLCanvasElement, we call
> HTMLCanvasElement::existingDrawingContext(), whose behavior was changed
> incorrectly in r162330 and was triggering the allocation of an ImageBuffer
> during tear-down.  This is not only wasteful, but it can trigger a
> garbage collection because the creation of an ImageBuffer will increase
> the amount of externallay allocated memory that is declared to V8.
> Triggering a GC during the finalization of a garbage collected object
> is bad.
>
> BUG=378289
>
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=174984

TBR=senorblanco@chromium.org,junov@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=378289

Review URL: https://codereview.chromium.org/307013006

git-svn-id: svn://svn.chromium.org/blink/trunk@175157 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=174930

BUG=373240
TBR=dominik.rottsches@intel.com

Review URL: https://codereview.chromium.org/308643005

git-svn-id: svn://svn.chromium.org/blink/trunk@175155 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The browser side code of the new WebSocket stack needs to
know the ID of the render frame where the corresponding
WebSocketBridge is created in order to handle certificate errors.

This CL just ports the code used by the old WebSocket
implementation to new one.

This CL also contains clean up on the old stack.

Chromium side change: https://codereview.chromium.org/308513004/

BUG=364361

Review URL: https://codereview.chromium.org/304113002

git-svn-id: svn://svn.chromium.org/blink/trunk@175154 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch simplifies parsing CSS animations and transitions slightly by
making us only return CSSValueLists from the parser, including when we
only have a single value in the list. This was probably done differently
in the past (with single values not being wrapped in a CSSValueList) as
premature optimisation, which adds needless complexity to the code. Note
this changes the foo.style.getPropertyCSSValue output slightly.

I've also made BisonCSSParser::parseAnimationTimingFunctionValue return
a single (non-valuelist-wrapped) CSSValue which isn't initial or inherit
to better match the intention of the callers.

R=alancutter,dstockwell

Review URL: https://codereview.chromium.org/293113007

git-svn-id: svn://svn.chromium.org/blink/trunk@175153 bbb929c8-8fbe-4397-9dbb-9b2b20218538