GitLab

BUG=none
R=haraken@chromium.org

Review URL: https://codereview.chromium.org/311053004

git-svn-id: svn://svn.chromium.org/blink/trunk@175884 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It was added as a part of this patch
https://codereview.chromium.org/13937017

The patch was reverted here
https://chromiumcodereview.appspot.com/23072015

Review URL: https://codereview.chromium.org/322103003

git-svn-id: svn://svn.chromium.org/blink/trunk@175883 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This will allow debugger to pause on "load", "error" and etc. event breakpoints only
when they happen on XHR event targets.

BUG=381470
R=yurys, pfeldman@chromium.org

Review URL: https://codereview.chromium.org/320933003

git-svn-id: svn://svn.chromium.org/blink/trunk@175882 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Just moving things around, no actualy logic change, though dropped unused
FrameAdded event.

BUG=361045

Review URL: https://codereview.chromium.org/320953002

git-svn-id: svn://svn.chromium.org/blink/trunk@175881 bbb929c8-8fbe-4397-9dbb-9b2b20218538

|m_elementDependencies| had raw ptr to |SVGElement|s.
This was to avoid reference cycles, but we can convert this to strong references in oilpan.

If the whole document is dead, rebuild code is not called. This means that
rebuildElementReference is not called for referencing element, but the
referencing element must be in the same document and is going to die anyway.
If the whole document isn't dead (== the part of the document tree is going
away), the unregister/rebuild code is guaranteed to be called from removedFrom,
and rebuildElementReference is guaranteed to be triggered from there.

BUG=357163, 370834

Review URL: https://codereview.chromium.org/327473002

git-svn-id: svn://svn.chromium.org/blink/trunk@175880 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=

Review URL: https://codereview.chromium.org/322793003

git-svn-id: svn://svn.chromium.org/blink/trunk@175879 bbb929c8-8fbe-4397-9dbb-9b2b20218538

These are accidentally added by r175288.


BUG=n/a
TEST=n/a

Review URL: https://codereview.chromium.org/320423006

git-svn-id: svn://svn.chromium.org/blink/trunk@175878 bbb929c8-8fbe-4397-9dbb-9b2b20218538

and reduce binary footprint by 5% ~25kb of 587k (clang x64 release)

BUG=382569
R=yurys@chromium.org

Review URL: https://codereview.chromium.org/325833002

git-svn-id: svn://svn.chromium.org/blink/trunk@175877 bbb929c8-8fbe-4397-9dbb-9b2b20218538

- We need to add Persistent<Element> in ImageLoader in order to
  follow the current behavior by manual ref/deref.
  Also, the manual ref/deref is replaced with RefPtr<Element>.

- ImageLoader was the last off-heap user of EventSender<T>. Now all of
  EventSender<T> users are on-heap. We can trace EventSender::m_dispatchSoonList
  and m_dispatchingList.

- HTMLImageElement and SVGImageElement had ImageLoader as part objects. This CL
  makes them OwnPtr / Member because using on-heap objects as part objects are
  dangerous.

- Add HTMLImageElement::create and SVGImageElement::create.

BUG=357163
R=haraken@chromium.org

Review URL: https://codereview.chromium.org/320253002

git-svn-id: svn://svn.chromium.org/blink/trunk@175876 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Go the whole hog and use (non-const) bare pointers for the conversion
constructor and assignment operator.

R=
BUG=340522

Review URL: https://codereview.chromium.org/329463004

git-svn-id: svn://svn.chromium.org/blink/trunk@175875 bbb929c8-8fbe-4397-9dbb-9b2b20218538

According to this GCC bug/patch, a user-specified constructor is
required for the emulated C++11 nullptr_t class if we want to be able
to create const nullptr_t objects without an explicit initializer,
a la: "const std::nullptr_t nullptr;"

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=42844

This fixes a build failure in GCC 4.5.

BUG=374431

Review URL: https://codereview.chromium.org/322913002

git-svn-id: svn://svn.chromium.org/blink/trunk@175874 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=337618
TEST=fast/dom/shadow/getComputedStyle-with-distribution.html

Review URL: https://codereview.chromium.org/240293007

git-svn-id: svn://svn.chromium.org/blink/trunk@175873 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=dmazzoni@chromium.org,haraken@chromium.org,tommi@chromium.org,tkent@chromium.org
BUG=340522

Review URL: https://codereview.chromium.org/315133004

git-svn-id: svn://svn.chromium.org/blink/trunk@175872 bbb929c8-8fbe-4397-9dbb-9b2b20218538

No one is using a value different from the FormDataList object's m_encoding.

BUG=none

Review URL: https://codereview.chromium.org/329513002

git-svn-id: svn://svn.chromium.org/blink/trunk@175871 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Debugger is always enabled at the moment when we run the test so we don't need to check that and dump the message.
Also we don't need to disable it manually at the end because it might hide an error as it happens with debugger-step-into-custom-element-callbacks.html

BUG=

Review URL: https://codereview.chromium.org/316403002

git-svn-id: svn://svn.chromium.org/blink/trunk@175870 bbb929c8-8fbe-4397-9dbb-9b2b20218538

When upgrading ICU to 52, I removed the stringprep data for IDNA 2003 because Chrome/Blink switched to UTS46/IDNA 2008. I overlooked that Blink email form handling still uses IDNA 2003 API that fails without stringprep data for IDNA 2003. 

This CL changes Blink to use UTS46 API. 

BUG=132145,376587
TEST=LayoutTest: fast/forms/email-idn-conversion.html

Review URL: https://codereview.chromium.org/326453002

git-svn-id: svn://svn.chromium.org/blink/trunk@175869 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Followup to:
IDL build: factor out global objects action into a GYP template
https://codereview.chromium.org/316173008/

R=haraken

Review URL: https://codereview.chromium.org/319933002

git-svn-id: svn://svn.chromium.org/blink/trunk@175868 bbb929c8-8fbe-4397-9dbb-9b2b20218538

R=haraken@chromium.org,tzik@chromium.org,tkent@chromium.org
BUG=340522

Review URL: https://codereview.chromium.org/314333002

git-svn-id: svn://svn.chromium.org/blink/trunk@175867 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=378942

Review URL: https://codereview.chromium.org/316153002

git-svn-id: svn://svn.chromium.org/blink/trunk@175866 bbb929c8-8fbe-4397-9dbb-9b2b20218538

|SVGDocumentExtensions::m_relativeLengthSVGRoots| is made a HashSet<Member<SVGSVGELement> > in oilpan.
This is valid in oilpan, because document and <svg> elements are guaranteed to die at the same time. When a <svg> element is removed from the document, |SVGSVGElement::removedFrom| would remove itself from the set.

BUG=357163

Review URL: https://codereview.chromium.org/323803002

git-svn-id: svn://svn.chromium.org/blink/trunk@175865 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=none
R=dpranke@chromium.org, dpranke@chomium.org

Review URL: https://codereview.chromium.org/323873005

git-svn-id: svn://svn.chromium.org/blink/trunk@175864 bbb929c8-8fbe-4397-9dbb-9b2b20218538

It's called only by its subclasses.

BUG=none
R=haraken

Review URL: https://codereview.chromium.org/324003003

git-svn-id: svn://svn.chromium.org/blink/trunk@175863 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Revert of Add an ASSERT about cross-world wrapper leakage into ScriptValue (https://codereview.chromium.org/314953005/)

Reason for revert:
Causes many browser_tests to fail:
ActiveScriptControllerBrowserTest.*
ExtensionApiTest.ContentScript*
etc

Original issue's description:
> Add an ASSERT about cross-world wrapper leakage into ScriptValue
> 
> - This CL adds an ASSERT to ScriptValue to verify that ScriptValue::v8Value() is used in the same world that created the ScriptValue.
> 
> - To insert the check, we need to get a current world in ScriptValue::v8Value(). For that goal, this CL makes sure that we're in some context when calling ScriptValue::v8Value().
> 
> BUG=341032
> 
> Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=175837

TBR=dcarney@chromium.org,adamk@chromium.org,jochen@chromium.org,tasak@google.com,abarth@chromium.org,haraken@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=341032

Review URL: https://codereview.chromium.org/326853002

git-svn-id: svn://svn.chromium.org/blink/trunk@175862 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Content size notification was only used for the Mac overlay navigation code, which no longer needs it as of Chromium r274906.

Remove it.

BUG=321437
TEST=everything still compiles and works.

Review URL: https://codereview.chromium.org/322793004

git-svn-id: svn://svn.chromium.org/blink/trunk@175861 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Step 1 of renames preceding the fix for family based fallback font selection.

BUG=382411
R=eae,behdad

Review URL: https://codereview.chromium.org/326763002

git-svn-id: svn://svn.chromium.org/blink/trunk@175858 bbb929c8-8fbe-4397-9dbb-9b2b20218538

There are two fixes here:
1) Previously frame scrollbars never had a chance to handle an event if an element or event handler swallowed the event. So, for example, if a mousedown event listener was added to the window and the handler called preventDefault(), then the frame scrollbar would never receive mouse clicks. This change always gives frame scrollbars an opportunity to handle the event. This would also affect elements (such as plugins) that lie underneath overlay scrollbars and swallowed events.

2) Previously if an overlay scrollbar was above a plugin, clicking on the scrollbar could trigger mouse capture on the plugin, which would interfere with the scrollbar receiving events. This change prevents mouse capture on a plugin from starting if the event was initiated above a scrollbar.

BUG=369898,358248

Review URL: https://codereview.chromium.org/296003011

git-svn-id: svn://svn.chromium.org/blink/trunk@175857 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This patch changes Position::isCandidate() and Position::isCandidate() not to
return true for SVG elements except for rendered text, since we don't implement
SVG editor yet.

Note: Rendered text in SVG documents are represented by |Text| class and their
renderer is |RenderSVGInlineText| derived from |RenderText| class.

BUG=380043
TEST=LayoutTests/editing/execCommand/format-block-with-uneditable-crash-2.html

Review URL: https://codereview.chromium.org/314923003

git-svn-id: svn://svn.chromium.org/blink/trunk@175855 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Currently ServiceWorkers are non-null, however when we implement the .waiting property, the browser will "set" a null ServiceWorker to clear the property. Make Blink ready to unwrap one of those.

BUG=379012

Review URL: https://codereview.chromium.org/323523002

git-svn-id: svn://svn.chromium.org/blink/trunk@175853 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This CL splits ScrollView::updateScrollbars() into 3 independent helper
functions, namely:

computeScrollbarExistence() computes whether scrollbars are needed given
the current ScrollView state.

adjustScrollbarExistence() creates/removes scrollbars and invoke appropriate
callbacks to update layout (if applicable).

updateScrollbarGeometry() positions the scrollbars and pushes the scroll
extents.

Also we no longer do update scrollbar passes in recursion. Instead we do
it in a loop and early exits if updateScrollbars() is re-entered.

No tests as there should be no behavior difference.

Review URL: https://codereview.chromium.org/314583008

git-svn-id: svn://svn.chromium.org/blink/trunk@175852 bbb929c8-8fbe-4397-9dbb-9b2b20218538

We're going to change element factories at least twice.
 1. Replace adoptRefWillBeRefCountedGarbageCollected with adoptRefWillBeNoop
 2-1. Replace PassRefPtrWillBeRawPtr<T> with T* and remove adoptRefWillBeNoop if
   we switch to Oilpan successfully.
 2-2. Replace PassRefPtrWillBeRawPtr<T> with PassRefPtr<T>, and
   adoptRefWillBeNoop with adoptRef if we decide not to switch to Oilpan.

This CL makes these changes easier.

Also, this CL split existing DEFINE_NODE_FACTORY into DECLARE_NODE_FACTORY and
DEFINE_NODE_FACTORY to avoid inline definition in headers.
http://www.chromium.org/developers/coding-style/cpp-dos-and-donts#TOC-Stop-inlining-code-in-headers

BUG=357163

Review URL: https://codereview.chromium.org/327633004

git-svn-id: svn://svn.chromium.org/blink/trunk@175851 bbb929c8-8fbe-4397-9dbb-9b2b20218538

First (simple) step in factoring out redundant GYP actions into templates.
Also moves the 'write_file_only_if_changed' flag to scripts, as that's where
it's used, which allows more simplification.

R=haraken

Review URL: https://codereview.chromium.org/316173008

git-svn-id: svn://svn.chromium.org/blink/trunk@175848 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175832

BUG=381138
TBR=leviw@chromium.org

Review URL: https://codereview.chromium.org/322693003

git-svn-id: svn://svn.chromium.org/blink/trunk@175847 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is by no means a good implementation of Fetch, however it is sufficient
to let us start proceeding with the cache API (and add).

R=falken@chromium.org,jkarlin@chromium.org
TBR=jochen@chromium.org
BUG=381179

Review URL: https://codereview.chromium.org/311343003

git-svn-id: svn://svn.chromium.org/blink/trunk@175846 bbb929c8-8fbe-4397-9dbb-9b2b20218538

http://src.chromium.org/viewvc/blink?view=revision&revision=175827

BUG=372837
TBR=chrishtr@chromium.org

Review URL: https://codereview.chromium.org/328553003

git-svn-id: svn://svn.chromium.org/blink/trunk@175845 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Reduce the size of Node by 8 bytes on 64bit with gcc and clang compilers
(from 72 bytes to 64 bytes) by altering the order of its parent classes. The
TreeShared<Node> parent class now comes last because it contains only 1 member
(in release) which is an int and the first member of Node is a uint32_t. This
way no padding is added by gcc/clang and these two members will use the same
word in memory.

MSVC likely won't get any benefit from this because it makes sure each class
in the hierarchy is memory-aligned (see "Compiler being used" section in [1]).
There should be no size impact for MSVC.

[1] http://www.cprogramming.com/tutorial/size_of_class_object.html

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/326793002

git-svn-id: svn://svn.chromium.org/blink/trunk@175844 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Fix RenderObject::repaintRectangle in situations when the render object is not a layer and is squashed.

Previously it was recording a repaint rect in the coordinate space of the enclosing *squashed* layer,
not the coordinate space of the *squashing layer* that the squashed layer paints into.
BUG=382129

Review URL: https://codereview.chromium.org/328533003

git-svn-id: svn://svn.chromium.org/blink/trunk@175843 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This replaces calls to RenderLayer::hasDirectReasonsForCompositing in
RenderObject::adjustStyleDifference with calls to
RenderLayer::styleDeterminedCompositingReasons.

This also removes hasDirectReasonsForCompositing altogether, as it
no longer has any callers.

Review URL: https://codereview.chromium.org/327593002

git-svn-id: svn://svn.chromium.org/blink/trunk@175842 bbb929c8-8fbe-4397-9dbb-9b2b20218538

BUG=368814
TBR=jww

Review URL: https://codereview.chromium.org/320403003

git-svn-id: svn://svn.chromium.org/blink/trunk@175841 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove impossible conditions from SelectorChecker::checkOne(). The code is in
an if "(selector.m_match == CSSSelector::PseudoClass)" check scope so it is
impossible for "selector.m_match == CSSSelector::Tag" to ever be true in
the two conditions that are removed.

This is based the following WebKit revision by benjamin@webkit.org:
http://trac.webkit.org/changeset/156380

R=esprehn@chromium.org

Review URL: https://codereview.chromium.org/328443003

git-svn-id: svn://svn.chromium.org/blink/trunk@175839 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is a follow on patch / fix after r175488, where I started
forcing a "crash" if we detected output from the sanitizer after a
test. We were not setting the process name or pid, which caused
problems later when writing out the crash logs.

TBR=ojan@chromium.org
BUG=336971

Review URL: https://codereview.chromium.org/325993002

git-svn-id: svn://svn.chromium.org/blink/trunk@175838 bbb929c8-8fbe-4397-9dbb-9b2b20218538