GitLab

BUG=

Review URL: https://codereview.chromium.org/429723006

git-svn-id: svn://svn.chromium.org/blink/trunk@179193 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The current SVGAnimation usecounter [1] is skewed by Modernizr [2] and
doesn't count the other animation types such as <animateMotion>. This
patch adds a new usecounter for SMIL animation elements that are in
the document. I manually confirmed this does not fire on the
Modernizr test page [3] whereas the old SVGAnimation counter does.

[1] http://www.chromestatus.com/metrics/feature/timeline/popularity/90
[2] https://github.com/Modernizr/Modernizr/blob/v2.8.3/modernizr.js#L832-L835
[3] http://modernizr.github.io/Modernizr/test/

BUG=258298

Review URL: https://codereview.chromium.org/424273003

git-svn-id: svn://svn.chromium.org/blink/trunk@179191 bbb929c8-8fbe-4397-9dbb-9b2b20218538

After this CL, CodeGeneratorV8 generates not only V8 bindings, but also
DOM implementation classes for IDL dictionaries.

- For an IDL file which defines a dictionary, the code generator
  generates bindings code (V8Foo.{cpp,h}) and DOM impl code
  (Foo.{cpp,h}).

- Added Jinja templates for DOM impl generation. The context for these
  templates is created by v8_dictionary.dictionary_impl_context().

- v8_types.cpp_types() takes a new argument called
  |used_as_return_type|. This flags is used by member_impl_context().
  Also, |used_in_cpp_sequence| argument is renamed to |used_as_member|.

IDL dictionaries are still not be able to use in core/modules. We need
to generate impl classes in the right place.
(e.g. core/dom/Foo.idl -> gen/blink/core/dom/Foo.{cpp,h})

Also, following types are not supported yet as dictionary members:
- union types.
- enumeration types.
- composite types (arrays, sequences and dictionaries) of which element
  type is an interface.

BUG=321462

Review URL: https://codereview.chromium.org/420763002

git-svn-id: svn://svn.chromium.org/blink/trunk@179190 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is a follow-up CL for r179158.
Applied tkent-san's comment in https://codereview.chromium.org/427803003/#msg4.

Review URL: https://codereview.chromium.org/423293002

git-svn-id: svn://svn.chromium.org/blink/trunk@179189 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is a followup to [1] which dropped the actual removal of
TextAutosizer{.h,.cpp}.

[1] https://src.chromium.org/viewvc/blink?view=rev&revision=179181

TBR=skobes
BUG=385380

Review URL: https://codereview.chromium.org/427953003

git-svn-id: svn://svn.chromium.org/blink/trunk@179188 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Reason for revert:
The change broke several Chromium OS interactive_ui_tests:

ExtensionPointerLockAccessFail
ExtensionPointerLockAccessPass
EscapingMouseLock
MouseLockSilentAfterTargetUnlock
TestTabDoesntExitMouseLockOnSubFrameNavigation
TestTabExitsMouseLockOnGoBack
TestTabExitsMouseLockOnNavigation
PointerLock_PointerLockLostWithFocus

that were using the prefixed API (""Uncaught TypeError: undefined is not a function").

See
http://build.chromium.org/p/chromium.webkit/builders/Linux%20ChromiumOS%20Tests%20%282%29/builds/735/steps/interactive_ui_tests/logs/stdio

Original issue's description:
> Remove prefixed Pointer Lock API.
> 
> The Pointer Lock API http://www.w3.org/TR/pointerlock has already been
> implemented without prefix, this change removes the 'webkit' prefixed
> API endpoints.
> 
> Intent to Deprecate and Remove:
> https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/tN8fWRtG0MI
> 
> BUG=398457
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=179172

TBR=abarth@chromium.org,scheib@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=398457

Review URL: https://codereview.chromium.org/424213007

git-svn-id: svn://svn.chromium.org/blink/trunk@179186 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove the WTF_USE_HARFBUZZ define and the remaining use of it as it is
it only used in files compiled on platforms where it is guaranteed to be
true.

Review URL: https://codereview.chromium.org/416273004

git-svn-id: svn://svn.chromium.org/blink/trunk@179185 bbb929c8-8fbe-4397-9dbb-9b2b20218538

m_autoTextsizer is gone in r179181. This CL removes the leftover.

TBR=pdr@chromium.org

Review URL: https://codereview.chromium.org/423293003

git-svn-id: svn://svn.chromium.org/blink/trunk@179184 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Intent to Deprecate and Remove:
https://groups.google.com/a/chromium.org/d/msg/blink-dev/h_GoezQ4MQ4/Q5abFkhA76sJ

BUG=398422

Review URL: https://codereview.chromium.org/425633002

git-svn-id: svn://svn.chromium.org/blink/trunk@179183 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This allows the test to pass on linux32 after the clang switch.
The test passed on 32bit mac already with the same compiler – this turned
out to be because clang is allowed to treat sqrtf() as a builtin on mac, but not
on linux (due to glibc's math routines setting errno). See the bug for details.

BUG=392764

Review URL: https://codereview.chromium.org/426633005

git-svn-id: svn://svn.chromium.org/blink/trunk@179182 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch removes the old text autosizer. Long live the fast text
autosizer!*


*which we should rename to TextAutosizer in a followup patch.

BUG=385380

Review URL: https://codereview.chromium.org/346603005

git-svn-id: svn://svn.chromium.org/blink/trunk@179181 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Use tighter typing in editing/ to increase code readability. This CL focuses on
InsertParagraphSeparatorCommand / InsertTextCommand.

R=leviw@chromium.org, yosin@chromium.org

Review URL: https://codereview.chromium.org/423003003

git-svn-id: svn://svn.chromium.org/blink/trunk@179180 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=398336
R=loislo, yurys

Review URL: https://codereview.chromium.org/430503002

git-svn-id: svn://svn.chromium.org/blink/trunk@179179 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Prepratory cleanup for making recursive frame detach work for all
frames. FrameDestructionObserver can only observe a LocalFrame.
Similarly willDetachFrameHost() and detachFromFrameHost() can only be
called on LocalFrames, since these methods are only called by
FrameLoader.

BUG=none

Review URL: https://codereview.chromium.org/427673008

git-svn-id: svn://svn.chromium.org/blink/trunk@179178 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Have the BidiRunList destructor delete all runs and change all
BidiRunList call sites, except RenderBlockLineLayout, to rely on the
destructor instead of explicit calls to deleteRuns.

In RenderBlockFlow::layoutRunsAndFloatsInRange deleteRuns is called
early, and explicitly, in two places. Removing these calls would likely
involve splitting the run list of changing the life time of it.

BUG=396963
R=dglazkov@chromium.org

Review URL: https://codereview.chromium.org/421463002

git-svn-id: svn://svn.chromium.org/blink/trunk@179177 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently, location.reload() always triggers a get. Change it to go through
the FrameLoaer::reload() path, like a user-initiated reload.

This aligns our behavior with firefox and IE.

BUG=6429
TEST=http/tests/navigation/location-reload-after-post.php

Review URL: https://codereview.chromium.org/410223003

git-svn-id: svn://svn.chromium.org/blink/trunk@179176 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch inserts frame barriers at the end of script execution tasks
that draw to a canvas. This is to prevent the accumulation of a multi-
frame backlog of draw commands in the SkPicture used by the display
list canvas implementation.

BUG=397516
TEST=RecordingImageBufferSurface unit tests

Review URL: https://codereview.chromium.org/429643002

git-svn-id: svn://svn.chromium.org/blink/trunk@179175 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Use tighter typing in editing/ to increase code readability.
This CL focuses on the create*Element() functions in htmlediting.

R=leviw@chromium.org, yosin@chromium.org

Review URL: https://codereview.chromium.org/422693002

git-svn-id: svn://svn.chromium.org/blink/trunk@179173 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The Pointer Lock API http://www.w3.org/TR/pointerlock has already been
implemented without prefix, this change removes the 'webkit' prefixed
API endpoints.

Intent to Deprecate and Remove:
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/tN8fWRtG0MI

BUG=398457

Review URL: https://codereview.chromium.org/424093006

git-svn-id: svn://svn.chromium.org/blink/trunk@179172 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The result of m_blocksForFingerprint.get() can be a null pointer, so don't
dereference it (that invokes undefined behaviour). Instead, just return the pointer.

Recent versions of Clang will optimize based on references always pointing
to valid objects. In the previous code below:

  BlockSet* roots = m_fingerprintMapper.getTentativeClusterRoots(fingerprint);
  if (!roots || roots->size() < 2 || !roots->contains(block))
      return 0;

the "!roots" null check would be optimized away, since "roots" is the address
of a reference, and must therefore point to a valid object.

BUG=38149
TEST=fast/text-autosizing tests don't crash with new Clang

Review URL: https://codereview.chromium.org/418253003

git-svn-id: svn://svn.chromium.org/blink/trunk@179171 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Use tighter typing in editing/ to increase code readability. This CL focuses
on FormatBlockCommand / FrameSelection.

R=leviw@chromium.org

Review URL: https://codereview.chromium.org/420103002

git-svn-id: svn://svn.chromium.org/blink/trunk@179170 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Use tighter typing in editing to increase code readability and avoid
unnecessary type checks. This CL focuses on markup.

Review URL: https://codereview.chromium.org/428823002

git-svn-id: svn://svn.chromium.org/blink/trunk@179169 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=vsevik

Review URL: https://codereview.chromium.org/429593006

git-svn-id: svn://svn.chromium.org/blink/trunk@179168 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=none
R=rbyers@chromium.org

Review URL: https://codereview.chromium.org/424153002

git-svn-id: svn://svn.chromium.org/blink/trunk@179167 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Drop slower Node::hasTagName(QualifiedName) overload. It is only used by
SiblingTraversalStrategies at this point. The objective is to encourage
future callers to use the faster alternatives:
- HTMLElement::hasTagName(HTMLQualifiedName) / SVGElement::hasTagName(SVGQualifiedName)
- Node::hasTagName(HTMLQualifiedName) / Node::hasTagName(SVGQualifiedName)
- Element::hasTagName(QualifiedName)

R=abarth@chromium.org

Review URL: https://codereview.chromium.org/415823002

git-svn-id: svn://svn.chromium.org/blink/trunk@179165 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Make disabled regex text lines line-through styled.

BUG=267592
R=vsevik

Review URL: https://codereview.chromium.org/423153002

git-svn-id: svn://svn.chromium.org/blink/trunk@179164 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove ContainerNode overload for NodeTraversal::previous() as it adds no
benefit compared to the one taking a Node in argument.

The ElementTraversal::previous() implementation calls NodeTraversal::previous()
which does not have a ContainerNode overload. The NodeTraversal::previous()
implementation calls previousSibling() and parentNode() on the argument so
having a ContainerNode overload for it wouldn't help as those operations are
fast on pure Nodes.

R=esprehn@chromium.org, adamk@chromium.org

Review URL: https://codereview.chromium.org/425033002

git-svn-id: svn://svn.chromium.org/blink/trunk@179163 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Believe changes related to this introduced flakiness again in the
browser. See Issue 390497.

> Drop the 'WebURLRequest::TargetType' enum.
> 
> We no longer use this enum in Chromium; everything has been ported over
> to using '::RequestContext' and '::FrameType', so we can safely remove
> both the enum and the translation machinery in '::setTargetType'.
> 
> BUG=390497
> 
> Review URL: https://codereview.chromium.org/417483002

TBR=mkwst@chromium.org

Review URL: https://codereview.chromium.org/422343003

git-svn-id: svn://svn.chromium.org/blink/trunk@179162 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We've seen recent examples of servers that replace an invalid set of
of high-bytes with a literal questionmark.  We are already excluding
the high bytes from consideration, so we do the same with the
questionmark to ensure a match should this happen.

To test this, we hack up our "server script" to replace an arbitrary high
byte with a questionmark. This is sufficient for testing although it may
not match any real server.

BUG=395351

Review URL: https://codereview.chromium.org/421623003

git-svn-id: svn://svn.chromium.org/blink/trunk@179161 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Update AssertMatchingEnums.cpp now that everything is in blink namespace.
This CL defines all assertions inside the blink namespace as well so that we
don't have have to specify blink:: explicitly in the assertion macro. Public
names cannot be identical to core blink ones anyway or they would clash now
that they are in the same namespace.

As a side effect, the same COMPILE_ASSERT_MATCHING_*() macros could now be
reused for enumerations that are not in the blink namespace (e.g. in the WTF
one) since the macros no longer specify blink:: explicitly.

This CL also removes a few outdated references to WebKit / WebCore.

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

Review URL: https://codereview.chromium.org/416313006

git-svn-id: svn://svn.chromium.org/blink/trunk@179160 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The namespace migration is complete.

TBR=eseidel@chromium.org

Review URL: https://codereview.chromium.org/429593004

git-svn-id: svn://svn.chromium.org/blink/trunk@179159 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It looks like r179102 increased sizeof(CSSValue) only in Windows builds with Oilpan enabled.
I don't know why. At the moment, this CL fixes the size expectation and fixed the compile error.

TBR=yukishiino@chromium.org

Review URL: https://codereview.chromium.org/427803003

git-svn-id: svn://svn.chromium.org/blink/trunk@179158 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=loislo, yurys

Review URL: https://codereview.chromium.org/427803002

git-svn-id: svn://svn.chromium.org/blink/trunk@179157 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/423433004

git-svn-id: svn://svn.chromium.org/blink/trunk@179156 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The API calls are per Isolate, and therefore should be on the isolate.
Make it so.

BUG=397026
R=haraken@chromium.org

Review URL: https://codereview.chromium.org/422293004

git-svn-id: svn://svn.chromium.org/blink/trunk@179155 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=pfeldman

Review URL: https://codereview.chromium.org/422313002

git-svn-id: svn://svn.chromium.org/blink/trunk@179154 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Review URL: https://codereview.chromium.org/393223003

BUG=394210

Review URL: https://codereview.chromium.org/393223003

git-svn-id: svn://svn.chromium.org/blink/trunk@179153 bbb929c8-8fbe-4397-9dbb-9b2b20218538

To reduce number of active tr/td hashed css selectors.

BUG=316092

Review URL: https://codereview.chromium.org/422293002

git-svn-id: svn://svn.chromium.org/blink/trunk@179152 bbb929c8-8fbe-4397-9dbb-9b2b20218538

I don't see any reason to convert the internal pointer of type void* to Node* and convert it back to void*.

BUG=235436

Review URL: https://codereview.chromium.org/422233003

git-svn-id: svn://svn.chromium.org/blink/trunk@179151 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We used to pass a typed set of quads plus the config that specified
color for each type of quad. This is not flexible enough to cover
the needs of highlighting of different types of render objects.
This switches to a more flexible interface passing an array of paths
with fill & outline colors specified for each path.

BUG=

Review URL: https://codereview.chromium.org/430463002

git-svn-id: svn://svn.chromium.org/blink/trunk@179150 bbb929c8-8fbe-4397-9dbb-9b2b20218538