https://bugs.webkit.org/show_bug.cgi?id=107601

Reviewed by Beth Dakin.

Source/WebCore:

Test: fullscreen/video-cursor-auto-hide.html

To facilitate automatically hiding the mouse cursor after a certain amount of idle mouse
time, add a new CSS property '-webkit-cursor-visibility', and handle that value inside of
EventHandler. When that value is present, the EventHandler will start a timer which, when
fired, will set the view's cursor to the None cursor. Subsequent mouse moves will either
cancel the timer, or reset the auto-hide timer.

Add a new CSS property, '-webkit-cursor-visibility':
* css/CSSComputedStyleDeclaration.cpp:
(WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):
* css/CSSParser.cpp:
(WebCore::CSSParser::parseValue):
* css/CSSPrimitiveValueMappings.h:
(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
(WebCore::CSSPrimitiveValue::operator ECursorVisibility):
* css/CSSProperty.cpp:
(WebCore::CSSProperty::isInheritedProperty):
* css/CSSPropertyNames.in:
* css/CSSValueKeywords.in:
* css/StyleBuilder.cpp:
(WebCore::StyleBuilder::StyleBuilder):
* css/StyleResolver.cpp:
(WebCore::StyleResolver::applyProperty):

Add a rule for video descendants of full screen elements to auto-hide the cursor on hover:
* css/fullscreen.css:
(:-webkit-full-screen video:hover):

Add a new timer, m_autoHideCursorTimer, to control resetting the view's cursor to None after
enough mouse idle time has passed.
* page/EventHandler.cpp:
(WebCore::EventHandler::EventHandler):
(WebCore::EventHandler::~EventHandler):
(WebCore::EventHandler::clear):
(WebCore::EventHandler::selectCursor):
(WebCore::EventHandler::startAutoHideCursorTimer):
(WebCore::EventHandler::cancelAutoHideCursorTimer):
(WebCore::EventHandler::autoHideCursorTimerFired):
* page/EventHandler.h:

Make FrameView::isActive() a public function (like it is in its superclass).
* page/FrameView.h:
(FrameView):

Move the definition for timeWithoutMouseMovementBeforeHidingControls from page->theme()
to page->settings() and allow that setting to be overridden in InternalSettings for
testing purposes:
* page/Settings.cpp:
(WebCore::Settings::Settings):
* page/Settings.h:
(WebCore::Settings::setTimeWithoutMouseMovementBeforeHidingControls):
(WebCore::Settings::timeWithoutMouseMovementBeforeHidingControls):
* rendering/RenderTheme.h:
* testing/InternalSettings.cpp:
(WebCore::InternalSettings::Backup::Backup):
(WebCore::InternalSettings::Backup::restoreTo):
(WebCore::InternalSettings::setTimeWithoutMouseMovementBeforeHidingControls):
* testing/InternalSettings.h:
(InternalSettings):
* testing/InternalSettings.idl:
* html/shadow/MediaControls.cpp:
(WebCore::MediaControls::startHideFullscreenControlsTimer):

Add support to FeatureObserver for CURSOR_VISIBILITY:
* page/FeatureObserver.h:

Enable the feature flag:
* Configurations/FeatureDefines.xcconfig:

Source/WebKit:

Enable the CURSOR_VISIBILITY feature.

* WebKit.vcxproj/FeatureDefines.props:

Source/WebKit/mac:

Enable the CURSOR_VISIBILITY feature.

* Configurations/FeatureDefines.xcconfig:

Source/WebKit2:

Enable the CURSOR_VISIBILITY feature.

* Configurations/FeatureDefines.xcconfig:

Source/WTF:

Enable the CURSOR_VISIBILITY feature.

* wtf/FeatureDefines.h:

LayoutTests:

* fullscreen/video-cursor-auto-hide-expected.txt: Added.
* fullscreen/video-cursor-auto-hide.html: Added.

Add an expected failing result in platforms which do not enable CURSOR_VISIBILITY.
* platform/chromium/TestExpectations:
* platform/efl/TestExpectations:
* platform/gtk/TestExpectations:
* platform/qt/TestExpectations:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@145003 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111633

Patch by Sudarsana Nagineni <sudarsana.nagineni@intel.com> on 2013-03-06
Reviewed by Martin Robinson.

Added missing return in InspectorClient::inspectorFilesPath().

* WebCoreSupport/InspectorClientGtk.cpp:
(WebKit::InspectorClient::inspectorFilesPath):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@145002 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Full screen from an iframe with a z-index style will allow siblings with higher z-index to overlay full screen content.
https://bugs.webkit.org/show_bug.cgi?id=111637

Reviewed by Simon Fraser.

Source/WebCore:

No new tests; updated fullscreen/full-screen-iframe-zIndex.html to expose bug.

Pages with explicit z-index: styles on their iframes will override the default z-index: UA style when
elements within that iframe enter full screen mode. Mark the UA style as !important so as to not allow
iframe siblings with higher z-index styles to overlay the full screen element.

* css/fullscreen.css:
(:-webkit-full-screen):

LayoutTests:

* fullscreen/full-screen-iframe-zIndex.html:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@145001 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111497

Reviewed by Geoffrey Garen.

This reverts the get/put_scoped_var part of the great non-local
variable resolution refactoring.  This still leaves all the lazy
variable resolution logic as it's necessary for global property
resolution, and i don't want to make the patch bigger than it
already is.

* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpBytecode):
(JSC::CodeBlock::CodeBlock):
* bytecode/CodeBlock.h:
(CodeBlock):
* bytecode/Opcode.h:
(JSC):
(JSC::padOpcodeName):
* bytecode/UnlinkedCodeBlock.cpp:
(JSC::generateFunctionCodeBlock):
(JSC::UnlinkedFunctionExecutable::codeBlockFor):
(JSC::UnlinkedCodeBlock::UnlinkedCodeBlock):
* bytecode/UnlinkedCodeBlock.h:
(JSC):
(UnlinkedFunctionExecutable):
(UnlinkedCodeBlock):
(JSC::UnlinkedCodeBlock::usesGlobalObject):
(JSC::UnlinkedCodeBlock::setGlobalObjectRegister):
(JSC::UnlinkedCodeBlock::globalObjectRegister):
* bytecompiler/BytecodeGenerator.cpp:
(JSC::ResolveResult::checkValidity):
(JSC::BytecodeGenerator::BytecodeGenerator):
(JSC::BytecodeGenerator::emitLoadGlobalObject):
(JSC):
(JSC::BytecodeGenerator::resolve):
(JSC::BytecodeGenerator::resolveConstDecl):
(JSC::BytecodeGenerator::emitResolve):
(JSC::BytecodeGenerator::emitResolveBase):
(JSC::BytecodeGenerator::emitResolveBaseForPut):
(JSC::BytecodeGenerator::emitResolveWithBaseForPut):
(JSC::BytecodeGenerator::emitResolveWithThis):
(JSC::BytecodeGenerator::emitGetStaticVar):
(JSC::BytecodeGenerator::emitPutStaticVar):
* bytecompiler/BytecodeGenerator.h:
(JSC::ResolveResult::lexicalResolve):
(JSC::ResolveResult::isStatic):
(JSC::ResolveResult::depth):
(JSC::ResolveResult::index):
(ResolveResult):
(JSC::ResolveResult::ResolveResult):
(BytecodeGenerator):
* bytecompiler/NodesCodegen.cpp:
(JSC::ResolveNode::isPure):
(JSC::FunctionCallResolveNode::emitBytecode):
(JSC::PostfixNode::emitResolve):
(JSC::TypeOfResolveNode::emitBytecode):
(JSC::PrefixNode::emitResolve):
(JSC::ReadModifyResolveNode::emitBytecode):
(JSC::AssignResolveNode::emitBytecode):
(JSC::ConstDeclNode::emitCodeSingle):
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::parseBlock):
* dfg/DFGCapabilities.cpp:
(JSC::DFG::debugFail):
* dfg/DFGCapabilities.h:
(JSC::DFG::canCompileOpcode):
(JSC::DFG::canInlineOpcode):
* jit/JIT.cpp:
(JSC::JIT::privateCompileMainPass):
* jit/JIT.h:
(JIT):
* jit/JITPropertyAccess.cpp:
(JSC::JIT::emit_op_get_scoped_var):
(JSC):
(JSC::JIT::emit_op_put_scoped_var):
* jit/JITPropertyAccess32_64.cpp:
(JSC::JIT::emit_op_get_scoped_var):
(JSC):
(JSC::JIT::emit_op_put_scoped_var):
* llint/LowLevelInterpreter32_64.asm:
* llint/LowLevelInterpreter64.asm:
* runtime/CodeCache.cpp:
(JSC::CodeCache::getCodeBlock):
(JSC::CodeCache::getProgramCodeBlock):
(JSC::CodeCache::getEvalCodeBlock):
* runtime/CodeCache.h:
(JSC):
(CodeCache):
* runtime/Executable.cpp:
(JSC::EvalExecutable::compileInternal):
(JSC::FunctionExecutable::produceCodeBlockFor):
* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::createEvalCodeBlock):
* runtime/JSGlobalObject.h:
(JSGlobalObject):
* runtime/Options.cpp:
(JSC::Options::initialize):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@145000 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111639

Reviewed by Levi Weintraub.

Converted the test. Also renamed it to line-feed-between-br-and-b-should-not-reorder-pasted-content.html
to reflect the semantics of the test, and also added more description per radar bug.

* editing/pasteboard/line-feed-between-br-and-b-should-not-reorder-pasted-content-expected.txt: Added.
* editing/pasteboard/line-feed-between-br-and-b-should-not-reorder-pasted-content.html: Copied from
LayoutTests/editing/pasteboard/paste-4035648-fix.html.
* editing/pasteboard/paste-4035648-fix.html: Removed.
* platform/chromium-linux/editing/pasteboard/paste-4035648-fix-expected.png: Removed.
* platform/chromium-mac-lion/editing/pasteboard/paste-4035648-fix-expected.png: Removed.
* platform/chromium-mac/editing/pasteboard/paste-4035648-fix-expected.png: Removed.
* platform/chromium-win/editing/pasteboard/paste-4035648-fix-expected.png: Removed.
* platform/chromium-win/editing/pasteboard/paste-4035648-fix-expected.txt: Removed.
* platform/chromium/editing/pasteboard/paste-4035648-fix-expected.txt: Removed.
* platform/gtk/editing/pasteboard/paste-4035648-fix-expected.png: Removed.
* platform/gtk/editing/pasteboard/paste-4035648-fix-expected.txt: Removed.
* platform/mac/editing/pasteboard/paste-4035648-fix-expected.png: Removed.
* platform/mac/editing/pasteboard/paste-4035648-fix-expected.txt: Removed.
* platform/qt/editing/pasteboard/paste-4035648-fix-expected.png: Removed.
* platform/qt/editing/pasteboard/paste-4035648-fix-expected.txt: Removed.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144999 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* platform/chromium/TestExpectations: fast/dom/Window/window-special-properties.html

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144998 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* EWSTools/start-queue-win.sh:
* Scripts/webkitpy/tool/commands/earlywarningsystem.py:
(WinEWS):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144997 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=110546

Unreviewed change to test expectations while the above bug is open.

* platform/mac/TestExpectations:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144996 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111228

Reviewed by Ryosuke Niwa.

Source/WebCore:

When a list is wrapped in a presentational inline like a b tag, we'd create markup that included
everything up to the b tag from moveParagraphs when intending to only move the contents of the
list item. This could make it impossible to remove content from a list and trigger loss of
editable text.

Test: editing/execCommand/insert-remove-block-list-inside-presentational-inline.html

* editing/EditingStyle.cpp:
(WebCore::elementMatchesAndPropertyIsNotInInlineStyleDecl): Ensure there's an inline style
before calling propertyExistsInStyle.
(WebCore::HTMLElementEquivalent::propertyExistsInStyle): Removing null check for style.
All callers ensure this value isn't null.
* editing/markup.cpp:
(WebCore::highestAncestorToWrapMarkup): Avoid going over RenderBlocks when finding the highest
presentational ancestor to avoid leaving the bounds of the original paragraph.

LayoutTests:

* editing/deleting/pruning-after-merge-1-expected.txt:
* editing/execCommand/insert-remove-block-list-inside-presentational-inline-expected.txt: Added.
* editing/execCommand/insert-remove-block-list-inside-presentational-inline.html: Added.
* editing/pasteboard/paste-and-sanitize-expected.txt:
* editing/pasteboard/paste-and-sanitize.html:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144995 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111382

Reviewed by Adam Barth.

.:

* ManualTests/mutation-observer-leaks-nodes.html: Added.

Source/WebCore:

Two-phase approach to implicit references: after grouping objects
together, add an implicit reference between each registered node's
group and the MutationObserver's group (which includes wrappers from
all worlds).

Also changed many uses of v8::Value to v8::Object where we know we're
dealing with Object and the V8 API expects them.

Test: ManualTests/mutation-observer-leaks-nodes.html

* bindings/v8/V8GCController.cpp:
(WebCore::ImplicitConnection::ImplicitConnection):
(WebCore::ImplicitConnection::wrapper):
(ImplicitConnection):
(WebCore::ImplicitReference::ImplicitReference): Wrapper class holding a parent who should have an implicit reference to a child.
(ImplicitReference):
(WebCore::operator<): Needed for std::sort() call to avoid the overhead of using a HashMap
(WebCore::WrapperGrouper::addObjectWrapperToGroup):
(WebCore::WrapperGrouper::addNodeWrapperToGroup):
(WebCore::WrapperGrouper::addImplicitReference):
(WrapperGrouper):
(WebCore::WrapperGrouper::apply):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144994 268f45cc-cd09-0410-ab3c-d52691b4dbfc

        
I think we want the assertion that I removed.

* dfg/DFGAbstractState.cpp:
(JSC::DFG::AbstractState::merge):
(JSC::DFG::AbstractState::mergeVariableBetweenBlocks):
* dfg/DFGAbstractState.h:
(AbstractState):



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144991 268f45cc-cd09-0410-ab3c-d52691b4dbfc

such as RenderThemeGtk2.cpp, in debug mode. RenderThemeGtk2.cpp tries to allow usage
of deprecated functions by undefining GTK_DISABLE_DEPRECATED, but it ended up being
redefined because autotoolsconfig.h was included again by headers that came after
config.h.

Reviewed by Martin Robinson.

* Source/autotools/SetupWebKitFeatures.m4: add checks to ensure the
autotoolsconfig.h header is only included once.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144990 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111619

Reviewed by Mark Hahnenberg.
        
This method is the one place where we still do some minimal amount of liveness pruning, but the style with
which it is written is awkward, and it makes an assertion about variablesAtTail that will be invalidated
by https://bugs.webkit.org/show_bug.cgi?id=111539.

* dfg/DFGAbstractState.cpp:
(JSC::DFG::AbstractState::merge):
(JSC::DFG::AbstractState::mergeVariableBetweenBlocks):
* dfg/DFGAbstractState.h:
(AbstractState):



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144989 268f45cc-cd09-0410-ab3c-d52691b4dbfc

extraInfo arguments are of type const String& instead of const char*.
This change gets rid of some of the memory ownership issues that arise
with char* pointers.
https://bugs.webkit.org/show_bug.cgi?id=111483

Patch by Ankur Taly <ataly@google.com> on 2013-03-06
Reviewed by Adam Barth.

There are no new tests as there is no change in behavior.

* bindings/v8/V8DOMActivityLogger.h:
(WebCore::V8DOMActivityLogger::log):

Source/WebKit/chromium: Modify log method in WebDOMActivityLogger so that the apiName and
extraInfo arguments are of type const WebString& instead of const char*.
This change gets rid of some of the memory ownership issues that arise
with char* pointers.
https://bugs.webkit.org/show_bug.cgi?id=111483

Patch by Ankur Taly <ataly@google.com> on 2013-03-06
Reviewed by Adam Barth.

There are no new tests as there is no change in behavior.

* public/WebDOMActivityLogger.h: Added property svn:eol-style.
(WebKit::WebDOMActivityLogger::log):
* src/WebDOMActivityLogger.cpp: Added property svn:eol-style.
(WebKit::DOMActivityLoggerContainer::log):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144988 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* platform/chromium/TestExpectations: compositing/rtl/..

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144987 268f45cc-cd09-0410-ab3c-d52691b4dbfc

"James Robinson" <jamesr@chromium.org> via sheriffbot.

Patch by Sheriff Bot <webkit.review.bot@gmail.com> on 2013-03-06

* DEPS:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144986 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* platform/win/TestExpectations:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144985 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* platform/chromium/TestExpectations: svg/css/font-face-crash.html

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144984 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK][WK2] Implement WebInspector::localizedStringsURL() to return the file URL of localizedStrings.js
https://bugs.webkit.org/show_bug.cgi?id=111448

Patch by Sudarsana Nagineni <sudarsana.nagineni@intel.com> on 2013-03-06
Reviewed by Alexey Proskuryakov.

Implement WebInspector::localizedStringsURL() method to return the
file URL of the localizedStrings.js.

This prevents printing warnings about "Localized string not found" in
the console.

* WebProcess/WebPage/gtk/WebInspectorGtk.cpp:
(WebKit::WebInspector::localizedStringsURL):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144982 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* platform/chromium/TestExpectations: svg/custom/foreign-object-skew.svg

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144981 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[chromium] |m_gestureScrollOnImplThread| is not reset to false following the end of a fling on the fast path
https://bugs.webkit.org/show_bug.cgi?id=111390

Reviewed by James Robinson.

If there is a fast path gesture scroll which turns into a fling, the member
|m_gestureScrollOnImplThread| is not reset to false once the fling ends. As a
result, GestureScrollUpdate events belonging to a subsequent scroll are always
handled on the fast path, even if they should have been handled on the slow
path instead.

* src/WebCompositorInputHandlerImpl.cpp:
(WebKit::WebCompositorInputHandlerImpl::cancelCurrentFling):
* src/WebCompositorInputHandlerImpl.h:
(WebCompositorInputHandlerImpl):
(WebKit::WebCompositorInputHandlerImpl::isGestureScrollOnImplThread):
* tests/WebCompositorInputHandlerImplTest.cpp:
(WebKit::TEST_F):
(WebKit):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144980 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111618

Reviewed by Rob Buis.

PR 303712
If we are using a RTL language, we should ensure that the selection handles are angled the right way.
Further, when expanding the selection, the start and end handles should be appropriately set so that
expansion in the natural direction is possible.

Internally reviewed by Mike Fenton.

* WebKitSupport/SelectionHandler.cpp:
(BlackBerry::WebKit::SelectionHandler::SelectionHandler):
(BlackBerry::WebKit::SelectionHandler::selectionPositionChanged):
* WebKitSupport/SelectionHandler.h:
(SelectionHandler):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144977 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111588

Reviewed by Stephen White.

No new tests, as existing tests in fast/writing-mode exercise this code.

e.g. japanese-rl-text-with-broken-font.html
     vertical-subst-font-vert-no-dflt.html

* platform/graphics/harfbuzz/FontPlatformDataHarfBuzz.cpp:
(WebCore::FontPlatformData::emSizeInFontUnits):
(WebCore::FontPlatformData::openTypeTable):
* platform/graphics/harfbuzz/HarfBuzzFaceSkia.cpp:
(WebCore::harfBuzzSkiaGetTable):
(WebCore::HarfBuzzFace::createFace):
* platform/graphics/skia/SimpleFontDataSkia.cpp:
(WebCore::SimpleFontData::platformInit):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144976 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=110608

Patch by Sudarsana Nagineni <sudarsana.nagineni@intel.com> on 2013-03-06
Reviewed by Kenneth Rohde Christiansen.

Test was expecting ContentsSizeChanged signal from ViewClientEfl, but
WebView was not calling client callback if useFixedLayout() is true.

This patch prevent early return if useFixedLayout() is true and call
client callback didChangeContentsSize(), which will emit the signal.

* UIProcess/efl/ViewClientEfl.cpp:
(WebKit::ViewClientEfl::didChangeContentsSize):
* UIProcess/efl/WebView.cpp:
(WebKit::WebView::didChangeContentsSize):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144975 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111603

Patch by Victor Costan <costan@gmail.com> on 2013-03-06
Reviewed by Alexey Proskuryakov.

No new tests. This is cleanup.

* platform/network/FormDataBuilder.cpp:
(WebCore::appendQuotedString): use proper types (e.g, size_t instead of
unsigned long)

LayoutTests: Cleanup in multipart FormData tests.
https://bugs.webkit.org/show_bug.cgi?id=111603

Patch by Victor Costan <costan@gmail.com> on 2013-03-06
Reviewed by Alexey Proskuryakov.

* http/tests/local/formdata/resources/send-form-data-common.js:
(sendFormData): JavaScript style cleanup.
(testSendingFormData): JavaScript style, bugfix in checking if
eventSender.beginDragWithFiles needs to be called.
* http/tests/local/formdata/send-form-data-with-filename-expected.txt: Better test description.
* http/tests/local/formdata/send-form-data-with-filename.html: Better test description.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144974 268f45cc-cd09-0410-ab3c-d52691b4dbfc

DFG should not run full CSE after the optimization fixpoint, since it really just wants store elimination
https://bugs.webkit.org/show_bug.cgi?id=111536

Reviewed by Oliver Hunt and Mark Hahnenberg.
        
The fixpoint will do aggressive load elimination and pure CSE. There's no need to do it after the fixpoint.
On the other hand, the fixpoint does not profit from doing store elimination (except for SetLocal/Flush).
Previously we had CSE do both, and had it avoid doing some store elimination during the fixpoint by querying
the fixpoint state. This changes CSE to be templated on mode - either NormalCSE or StoreElimination - so
that we explicitly put it into one of those modes depending on where we call it from. The goal is to reduce
time spent doing load elimination after the fixpoint, since that is just wasted cycles.

* dfg/DFGCSEPhase.cpp:
(JSC::DFG::CSEPhase::CSEPhase):
(JSC::DFG::CSEPhase::run):
(JSC::DFG::CSEPhase::performNodeCSE):
(JSC::DFG::CSEPhase::performBlockCSE):
(JSC::DFG::performCSE):
(DFG):
(JSC::DFG::performStoreElimination):
* dfg/DFGCSEPhase.h:
(DFG):
* dfg/DFGDriver.cpp:
(JSC::DFG::compile):



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144973 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111610

Reviewed by Eric Seidel.

This fixes a 27% regression in dromaeo_jslibattrjquery on the threaded HTML parser vs. the main thread parser.

No new tests because no new functionality.

* html/parser/CompactHTMLToken.cpp:
(WebCore::CompactHTMLToken::CompactHTMLToken):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144972 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111353

Reviewed by Anders Carlsson.

Add PROT_NONE pages at the beginning and end of every
mmap call made by fastmalloc.

* wtf/TCSystemAlloc.cpp:
(TryMmap):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144971 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111508

Patch by James Robinson <jamesr@chromium.org> on 2013-03-06
Reviewed by Dirk Pranke.

Some tests in webkit_unit_tests, for example WebPageSerializerTest.HTMLNodes, depend on localizable resources
from a pak file. Since unit tests initialize for testing with the same webkit_support functions as layout tests
they end up using DumpRenderTree.pak, but this dependency isn't listed explicitly in the gyp so if you manage to
run webkit_unit_tests without building DumpRenderTree first these tests just crash.

This adds a dependency from webkit_unit_tests to DumpRenderTree_resources so the pak is always built. It's a bit
weird for a gyp file in Source/WebKit/chromium/ to reach into Tools/DumpRenderTree, but I can't think of
anything better.

* WebKitUnitTests.gyp:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144970 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=82866

Patch by Manuel Rego Casasnovas <rego@igalia.com> on 2013-03-06
Reviewed by Martin Robinson.

Source/WebKit2:

Test was failing because of new media controls needs more space to be
painted than just 10x10 pixels. This is similar to r144070.

* UIProcess/API/gtk/tests/TestWebKitWebView.cpp:
(testWebViewMouseTarget): Modified HTML in order to give more space to the media controls.

Tools:

* Scripts/run-gtk-tests:
(TestRunner): Unflag WebKit2APITests/WebKitWebView/mouse-target test as
it is passing now.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144969 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111608

Reviewed by Dirk Pranke.

Remove logTime.  It was never accurate, and the dev tools
have better ways to measure this if anyone cares to look
into dashboard perf.

* TestResultServer/static-dashboards/dashboard_base.js:
* TestResultServer/static-dashboards/flakiness_dashboard.js:
(processTestRunsForBuilder):
(appendHTML):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144968 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111509

Reviewed by Dirk Pranke.

Add DumpRenderTree.pak to DumpRenderTree.app again.

* DumpRenderTree/DumpRenderTree.gyp/DumpRenderTree.gyp:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144966 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add a flaky crash test expectation to platform/mac/fast/speechsynthesis/speech-synthesis-cancel.html
on Mac per 111613.

* platform/mac/TestExpectations:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144965 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* platform/chromium/TestExpectations: animations/animation-delay-changed.html

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144964 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111592

Reviewed by Ryosuke Niwa.

Speculative fix for a flaky test. The test sometimes fails on slower machines (mostly debug).
Re-check the results in a loop for up to 200ms to see if the :hover pseudo style applies.

Also force a layout after changing the style, which should also trigger the :hover style.

* fast/css/hover-update.html:
* platform/chromium/TestExpectations:
* platform/efl/TestExpectations:
* platform/gtk/TestExpectations:
* platform/mac/TestExpectations:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144963 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* platform/chromium/TestExpectations:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144962 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111137

Reviewed by Stephen White.

Refactor GIFImageReaderReader with the following changes:
+ Separate GIFLZWContext for decoding states.
+ Replace unsigned char* with Vector<unsigned char>

There is no change in code behavior and just refactoring.

No new tests. This is covered by existing GIFImageReaderTest.
I also did a local testing on a 50k image corpus and showed no regression.

* platform/image-decoders/gif/GIFImageDecoder.cpp:
(WebCore::GIFImageDecoder::haveDecodedRow):
* platform/image-decoders/gif/GIFImageDecoder.h:
(GIFImageDecoder):
* platform/image-decoders/gif/GIFImageReader.cpp:
(GIFImageReader::outputRow):
(GIFImageReader::doLZW):
(GIFImageReader::decodeInternal):
(GIFImageReader::prepareLZWContext):
* platform/image-decoders/gif/GIFImageReader.h:
(GIFFrameContext):
(GIFFrameContext::GIFFrameContext):
(GIFFrameContext::~GIFFrameContext):
(GIFLZWContext):
(GIFLZWContext::GIFLZWContext):
(GIFImageReader):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144961 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111547
<rdar://problem/13355801>

Reviewed by Simon Fraser.

* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::updatePreferences):
(WebKit::TiledCoreAnimationDrawingArea::createPageOverlayLayer):
Keep the page overlay layer's debug border and repaint counter settings updated.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144960 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111557

Reviewed by Geoff Garen.

* page/scrolling/mac/ScrollingCoordinatorMac.mm:
(WebCore::ScrollingCoordinatorMac::commitTreeState):
* platform/graphics/TiledBacking.h:
* platform/graphics/ca/mac/TileCache.mm:
(WebCore::TileCache::updateTileCoverageMap):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144959 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=111467

Reviewed by Darin Adler.

Responding to review comments by Darin Adler.

* NetworkProcess/mac/NetworkProcessMac.mm:
(WebKit::NetworkProcess::platformInitializeNetworkProcess):
* WebProcess/mac/WebProcessMac.mm:
(WebKit::WebProcess::platformInitializeWebProcess): Tighten up this code
by reducing extraneous local variables.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@144958 268f45cc-cd09-0410-ab3c-d52691b4dbfc