Timeline

Sep 3, 2013:

11:26 PM Changeset in webkit [155023] by fpizlo@apple.com

• 54 edits
  15 adds in trunk

The DFG should be able to tier-up and OSR enter into the FTL
​https://bugs.webkit.org/show_bug.cgi?id=112838

Source/JavaScriptCore:

Reviewed by Mark Hahnenberg.

This adds the ability for the DFG to tier-up into the FTL. This works in both
of the expected tier-up modes:

Replacement: frequently called functions eventually have their entrypoint
replaced with one that goes into FTL-compiled code. Note, this will be a
slow-down for now since we don't yet have LLVM calling convention integration.

OSR entry: code stuck in hot loops gets OSR'd into the FTL from the DFG.

This means that if the DFG detects that a function is an FTL candidate, it
inserts execution counting code similar to the kind that the baseline JIT
would use. If you trip on a loop count in a loop header that is an OSR
candidate (it's not an inlined loop), we do OSR; otherwise we do replacement.
OSR almost always also implies future replacement.

OSR entry into the FTL is really cool. It uses a specialized FTL compile of
the code, where early in the DFG pipeline we replace the original root block
with an OSR entrypoint block that jumps to the pre-header of the hot loop.
The OSR entrypoint loads all live state at the loop pre-header using loads
from a scratch buffer, which gets populated by the runtime's OSR entry
preparation code (FTL::prepareOSREntry()). This approach appears to work well
with all of our subsequent optimizations, including prediction propagation,
CFA, and LICM. LLVM seems happy with it, too. Best of all, it works naturally
with concurrent compilation: when we hit the tier-up trigger we spawn a
compilation plan at the bytecode index from which we triggered; once the
compilation finishes the next trigger will try to enter, at that bytecode
index. If it can't - for example because the code has moved on to another
loop - then we just try again. Loops that get hot enough for OSR entry (about
25,000 iterations) will probably still be running when a concurrent compile
finishes, so this doesn't appear to be a big problem.

This immediately gives us a 70% speed-up on imaging-gaussian-blur. We could
get a bigger speed-up by adding some more intelligence and tweaking LLVM to
compile code faster. Those things will happen eventually but this is a good
start. Probably this code will see more tuning as we get more coverage in the
FTL JIT, but I'll worry about that in future patches.

• CMakeLists.txt:
• GNUmakefile.list.am:
• JavaScriptCore.xcodeproj/project.pbxproj:
• Target.pri:
• bytecode/CodeBlock.cpp:

(JSC::CodeBlock::CodeBlock):
(JSC::CodeBlock::hasOptimizedReplacement):
(JSC::CodeBlock::setOptimizationThresholdBasedOnCompilationResult):

• bytecode/CodeBlock.h:
• dfg/DFGAbstractInterpreterInlines.h:

(JSC::DFG::::executeEffects):

• dfg/DFGByteCodeParser.cpp:

(JSC::DFG::ByteCodeParser::parseBlock):
(JSC::DFG::ByteCodeParser::parse):

• dfg/DFGCFGSimplificationPhase.cpp:

(JSC::DFG::CFGSimplificationPhase::run):

• dfg/DFGClobberize.h:

(JSC::DFG::clobberize):

• dfg/DFGDriver.cpp:

(JSC::DFG::compileImpl):
(JSC::DFG::compile):

• dfg/DFGDriver.h:
• dfg/DFGFixupPhase.cpp:

(JSC::DFG::FixupPhase::fixupNode):

• dfg/DFGGraph.cpp:

(JSC::DFG::Graph::dump):
(JSC::DFG::Graph::killBlockAndItsContents):
(JSC::DFG::Graph::killUnreachableBlocks):

• dfg/DFGGraph.h:
• dfg/DFGInPlaceAbstractState.cpp:

(JSC::DFG::InPlaceAbstractState::initialize):

• dfg/DFGJITCode.cpp:

(JSC::DFG::JITCode::reconstruct):
(JSC::DFG::JITCode::checkIfOptimizationThresholdReached):
(JSC::DFG::JITCode::optimizeNextInvocation):
(JSC::DFG::JITCode::dontOptimizeAnytimeSoon):
(JSC::DFG::JITCode::optimizeAfterWarmUp):
(JSC::DFG::JITCode::optimizeSoon):
(JSC::DFG::JITCode::forceOptimizationSlowPathConcurrently):
(JSC::DFG::JITCode::setOptimizationThresholdBasedOnCompilationResult):

• dfg/DFGJITCode.h:
• dfg/DFGJITFinalizer.cpp:

(JSC::DFG::JITFinalizer::finalize):
(JSC::DFG::JITFinalizer::finalizeFunction):
(JSC::DFG::JITFinalizer::finalizeCommon):

• dfg/DFGLoopPreHeaderCreationPhase.cpp:

(JSC::DFG::createPreHeader):
(JSC::DFG::LoopPreHeaderCreationPhase::run):

• dfg/DFGLoopPreHeaderCreationPhase.h:
• dfg/DFGNode.h:

(JSC::DFG::Node::hasUnlinkedLocal):
(JSC::DFG::Node::unlinkedLocal):

• dfg/DFGNodeType.h:
• dfg/DFGOSREntry.cpp:

(JSC::DFG::prepareOSREntry):

• dfg/DFGOSREntrypointCreationPhase.cpp: Added.

(JSC::DFG::OSREntrypointCreationPhase::OSREntrypointCreationPhase):
(JSC::DFG::OSREntrypointCreationPhase::run):
(JSC::DFG::performOSREntrypointCreation):

• dfg/DFGOSREntrypointCreationPhase.h: Added.
• dfg/DFGOperations.cpp:
• dfg/DFGOperations.h:
• dfg/DFGPlan.cpp:

(JSC::DFG::Plan::Plan):
(JSC::DFG::Plan::compileInThread):
(JSC::DFG::Plan::compileInThreadImpl):

• dfg/DFGPlan.h:
• dfg/DFGPredictionInjectionPhase.cpp:

(JSC::DFG::PredictionInjectionPhase::run):

• dfg/DFGPredictionPropagationPhase.cpp:

(JSC::DFG::PredictionPropagationPhase::propagate):

• dfg/DFGSafeToExecute.h:

(JSC::DFG::safeToExecute):

• dfg/DFGSpeculativeJIT32_64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

• dfg/DFGSpeculativeJIT64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

• dfg/DFGTierUpCheckInjectionPhase.cpp: Added.

(JSC::DFG::TierUpCheckInjectionPhase::TierUpCheckInjectionPhase):
(JSC::DFG::TierUpCheckInjectionPhase::run):
(JSC::DFG::performTierUpCheckInjection):

• dfg/DFGTierUpCheckInjectionPhase.h: Added.
• dfg/DFGToFTLDeferredCompilationCallback.cpp: Added.

(JSC::DFG::ToFTLDeferredCompilationCallback::ToFTLDeferredCompilationCallback):
(JSC::DFG::ToFTLDeferredCompilationCallback::~ToFTLDeferredCompilationCallback):
(JSC::DFG::ToFTLDeferredCompilationCallback::create):
(JSC::DFG::ToFTLDeferredCompilationCallback::compilationDidBecomeReadyAsynchronously):
(JSC::DFG::ToFTLDeferredCompilationCallback::compilationDidComplete):

• dfg/DFGToFTLDeferredCompilationCallback.h: Added.
• dfg/DFGToFTLForOSREntryDeferredCompilationCallback.cpp: Added.

(JSC::DFG::ToFTLForOSREntryDeferredCompilationCallback::ToFTLForOSREntryDeferredCompilationCallback):
(JSC::DFG::ToFTLForOSREntryDeferredCompilationCallback::~ToFTLForOSREntryDeferredCompilationCallback):
(JSC::DFG::ToFTLForOSREntryDeferredCompilationCallback::create):
(JSC::DFG::ToFTLForOSREntryDeferredCompilationCallback::compilationDidBecomeReadyAsynchronously):
(JSC::DFG::ToFTLForOSREntryDeferredCompilationCallback::compilationDidComplete):

• dfg/DFGToFTLForOSREntryDeferredCompilationCallback.h: Added.
• dfg/DFGWorklist.cpp:

(JSC::DFG::globalWorklist):

• dfg/DFGWorklist.h:
• ftl/FTLCapabilities.cpp:

(JSC::FTL::canCompile):

• ftl/FTLCapabilities.h:
• ftl/FTLForOSREntryJITCode.cpp: Added.

(JSC::FTL::ForOSREntryJITCode::ForOSREntryJITCode):
(JSC::FTL::ForOSREntryJITCode::~ForOSREntryJITCode):
(JSC::FTL::ForOSREntryJITCode::ftlForOSREntry):
(JSC::FTL::ForOSREntryJITCode::initializeEntryBuffer):

• ftl/FTLForOSREntryJITCode.h: Added.

(JSC::FTL::ForOSREntryJITCode::entryBuffer):
(JSC::FTL::ForOSREntryJITCode::setBytecodeIndex):
(JSC::FTL::ForOSREntryJITCode::bytecodeIndex):
(JSC::FTL::ForOSREntryJITCode::countEntryFailure):
(JSC::FTL::ForOSREntryJITCode::entryFailureCount):

• ftl/FTLJITFinalizer.cpp:

(JSC::FTL::JITFinalizer::finalizeFunction):

• ftl/FTLLink.cpp:

(JSC::FTL::link):

• ftl/FTLLowerDFGToLLVM.cpp:

(JSC::FTL::LowerDFGToLLVM::compileBlock):
(JSC::FTL::LowerDFGToLLVM::compileNode):
(JSC::FTL::LowerDFGToLLVM::compileExtractOSREntryLocal):
(JSC::FTL::LowerDFGToLLVM::compileGetLocal):
(JSC::FTL::LowerDFGToLLVM::addWeakReference):

• ftl/FTLOSREntry.cpp: Added.

(JSC::FTL::prepareOSREntry):

• ftl/FTLOSREntry.h: Added.
• ftl/FTLOutput.h:

(JSC::FTL::Output::crashNonTerminal):
(JSC::FTL::Output::crash):

• ftl/FTLState.cpp:

(JSC::FTL::State::State):

• interpreter/Register.h:

(JSC::Register::unboxedDouble):

• jit/JIT.cpp:

(JSC::JIT::emitEnterOptimizationCheck):

• jit/JITCode.cpp:

(JSC::JITCode::ftlForOSREntry):

• jit/JITCode.h:
• jit/JITStubs.cpp:

(JSC::DEFINE_STUB_FUNCTION):

• runtime/Executable.cpp:

(JSC::ScriptExecutable::newReplacementCodeBlockFor):

• runtime/Options.h:
• runtime/VM.cpp:

(JSC::VM::ensureWorklist):

• runtime/VM.h:

LayoutTests:

Reviewed by Mark Hahnenberg.

Fix marsaglia to check the result instead of printing, and add a second
version that relies on OSR entry.

• fast/js/regress/marsaglia-osr-entry-expected.txt: Added.
• fast/js/regress/marsaglia-osr-entry.html: Added.
• fast/js/regress/script-tests/marsaglia-osr-entry.js: Added.

(marsaglia):

• fast/js/regress/script-tests/marsaglia.js:

11:13 PM Changeset in webkit [155022] by Chris Fleizach

• 6 edits
  2 adds in trunk

AX: REGRESSION: @title is exposed as AXDescription when label label from contents already exists.
​https://bugs.webkit.org/show_bug.cgi?id=120550

Reviewed by Mario Sanchez Prada.

Source/WebCore:

Resolve a FIXME from the accessible name computation refactoring so that alternative text for links do not
show up in the title field and do not duplicate naming when a title tag is used.

Effectively, this means that links no longer use AXTitle for alternative text. They use AXDescription
like all other elements.

Test: platform/mac/accessibility/link-with-title.html

• accessibility/mac/WebAccessibilityObjectWrapperMac.mm:

(-[WebAccessibilityObjectWrapper accessibilityTitle]):
(-[WebAccessibilityObjectWrapper accessibilityDescription]):

LayoutTests:

• accessibility/image-map1.html:
• platform/mac/accessibility/document-links-expected.txt:
• platform/mac/accessibility/image-map1-expected.txt:
• platform/mac/accessibility/link-with-title-expected.txt: Added.
• platform/mac/accessibility/link-with-title.html: Added.

10:48 PM Changeset in webkit [155021] by fpizlo@apple.com

• 4 edits in trunk/Source

CodeBlock memory cost reporting should be rationalized
​https://bugs.webkit.org/show_bug.cgi?id=120615

Source/JavaScriptCore:

Reviewed by Darin Adler.

Report the size of the instruction stream, and then remind the GC that we're
using memory when we trace.

This is a slight slow-down on some JSBench tests because it makes us GC a
bit more frequently. But I think it's well worth it; if we really want those
tests to GC less frequently then we can achieve that through other kinds of
tuning. It's better that the GC knows that CodeBlocks do in fact use memory;
what it does with that information is a somewhat orthogonal question.

• bytecode/CodeBlock.cpp:

(JSC::CodeBlock::CodeBlock):
(JSC::CodeBlock::visitAggregate):

Source/WTF:

Reviewed by Darin Adler.

• wtf/RefCountedArray.h:

(WTF::RefCountedArray::refCount):

10:06 PM Changeset in webkit [155020] by enrica@apple.com

• 3 edits in trunk/Source/WTF

Follow up to ​http://trac.webkit.org/changeset/155014

Reviewed by Alexey Proskuryakov.

In the r155014 I renamed hasLineBreakingPropertyComplexContext
to requiresComplexContextForWordBreaking but forgot to
make the same change in UnicodeWchar.h.

• wtf/unicode/wchar/UnicodeWchar.cpp:

(WTF::Unicode::requiresComplexContextForWordBreaking):

• wtf/unicode/wchar/UnicodeWchar.h:

9:53 PM Changeset in webkit [155019] by Darin Adler

• 4 edits in trunk/Source/WebCore

Change type of Document::doctype back to a raw pointer
​https://bugs.webkit.org/show_bug.cgi?id=120617

Reviewed by Andreas Kling.

• dom/Document.cpp:

(WebCore::Document::doctype): Return a raw pointer.
(WebCore::Document::childrenChanged): Use the raw pointer.
Also added a FIXME about this code that is probably in the wrong place.

• dom/Document.h: More of the same.

• editing/markup.cpp:

(WebCore::documentTypeString): Get rid of local variable entirely,
since null is already handled right by createMarkup, and also remove
the call to get since doctype is just a raw pointer.

9:49 PM Changeset in webkit [155018] by akling@apple.com

• 2 edits in trunk/Source/WebCore

ASSERTION FAILED: frame().view() == this closing a page with SVG or video
<​https://webkit.org/b/120645>

Reviewed by Antti Koivisto.

Have RenderSVGResourceContainer check if the document is being destroyed before
triggering any repaints. This replaces the previous check for a null RenderView
which meant basically the same thing.

We could add more and better assertions to catch unnecessary work during tree
teardown, but let's do that separately.

• rendering/svg/RenderSVGResourceContainer.cpp:

(WebCore::RenderSVGResourceContainer::markClientForInvalidation):

9:27 PM Changeset in webkit [155017] by Darin Adler

• 2 edits in trunk/Source/WebKit2

REGRESSION (r154967) window resize is very choppy
​https://bugs.webkit.org/show_bug.cgi?id=120653

Reviewed by Andreas Kling.

Andreas Kling spotted the bad change.

• Platform/CoreIPC/Connection.cpp:

(CoreIPC::Connection::waitForMessage): Roll out this incorrect change.
The code here is not the same as a call to take.

8:34 PM Changeset in webkit [155016] by gyuyoung.kim@samsung.com

• 2 edits in trunk/Source/WebCore

Fix uninitialized build warning in make_names.pl
​https://bugs.webkit.org/show_bug.cgi?id=120658

Reviewed by Andreas Kling.

No new tests, no behavior change.

• dom/make_names.pl:

(printTypeChecks): Fixed a build warning since r154965.

7:46 PM Changeset in webkit [155015] by Darin Adler

• 2 edits in trunk/Source/WebCore

Fix backwards branch in ~Node from r154967
​https://bugs.webkit.org/show_bug.cgi?id=120659

Reviewed by Andreas Kling.

• dom/Node.cpp:

(WebCore::Node::~Node): Fix backwards branch. The old code ran the code only
when we did not remove the item from the table. I removed a ! from this expression
after review; bad idea.

6:13 PM Changeset in webkit [155014] by enrica@apple.com

• 5 edits
  2 adds in trunk

Can't select Katakana word by double-clicking.
<rdar://problem/14654926>

Reviewed by Alexey Proskuryakov and Ryosuke Niwa.

Source/WebCore:

For some languages, like Japanese we need
to use more context for word breaking.

New test: editing/selection/doubleclick-japanese-text.html

• platform/text/TextBoundaries.h:

(WebCore::requiresContextForWordBoundary):

Source/WTF:

For some languages, like Japanese we need
to use more context for word breaking.
I've renamed the function to better reflect its use
and remove the unused hasLineBreakingPropertyComplexContextOrIdeographic.

• wtf/unicode/icu/UnicodeIcu.h:

(WTF::Unicode::requiresComplexContextForWordBreaking):

LayoutTests:

Added new test for this scenario.

• editing/selection/doubleclick-japanese-text-expected.txt: Added.
• editing/selection/doubleclick-japanese-text.html: Added.

5:26 PM Changeset in webkit [155013] by mark.lam@apple.com

• 16 edits in trunk/Source

Converting StackIterator to a callback interface.
​https://bugs.webkit.org/show_bug.cgi?id=120564.

Reviewed by Filip Pizlo.

Source/JavaScriptCore:

• API/JSContextRef.cpp:

(BacktraceFunctor::BacktraceFunctor):
(BacktraceFunctor::operator()):
(JSContextCreateBacktrace):

• interpreter/CallFrame.cpp:
• interpreter/CallFrame.h:
• interpreter/Interpreter.cpp:

(JSC::DumpRegisterFunctor::DumpRegisterFunctor):
(JSC::DumpRegisterFunctor::operator()):
(JSC::Interpreter::dumpRegisters):
(JSC::unwindCallFrame):
(JSC::GetStackTraceFunctor::GetStackTraceFunctor):
(JSC::GetStackTraceFunctor::operator()):
(JSC::Interpreter::getStackTrace):
(JSC::Interpreter::stackTraceAsString):
(JSC::UnwindFunctor::UnwindFunctor):
(JSC::UnwindFunctor::operator()):
(JSC::Interpreter::unwind):

• interpreter/Interpreter.h:
• interpreter/StackIterator.cpp:

(JSC::StackIterator::numberOfFrames):
(JSC::StackIterator::gotoFrameAtIndex):
(JSC::StackIterator::gotoNextFrameWithFilter):
(JSC::StackIterator::resetIterator):
(JSC::StackIterator::Frame::print):
(debugPrintCallFrame):
(DebugPrintStackFunctor::operator()):
(debugPrintStack): Added for debugging convenience.

• interpreter/StackIterator.h:

(JSC::StackIterator::Frame::index):
(JSC::StackIterator::iterate):

• jsc.cpp:

(FunctionJSCStackFunctor::FunctionJSCStackFunctor):
(FunctionJSCStackFunctor::operator()):
(functionJSCStack):

• profiler/ProfileGenerator.cpp:

(JSC::AddParentForConsoleStartFunctor::AddParentForConsoleStartFunctor):
(JSC::AddParentForConsoleStartFunctor::foundParent):
(JSC::AddParentForConsoleStartFunctor::operator()):
(JSC::ProfileGenerator::addParentForConsoleStart):

• runtime/JSFunction.cpp:

(JSC::RetrieveArgumentsFunctor::RetrieveArgumentsFunctor):
(JSC::RetrieveArgumentsFunctor::result):
(JSC::RetrieveArgumentsFunctor::operator()):
(JSC::retrieveArguments):
(JSC::RetrieveCallerFunctionFunctor::RetrieveCallerFunctionFunctor):
(JSC::RetrieveCallerFunctionFunctor::result):
(JSC::RetrieveCallerFunctionFunctor::operator()):
(JSC::retrieveCallerFunction):

• runtime/JSGlobalObjectFunctions.cpp:

(JSC::GlobalFuncProtoGetterFunctor::GlobalFuncProtoGetterFunctor):
(JSC::GlobalFuncProtoGetterFunctor::result):
(JSC::GlobalFuncProtoGetterFunctor::operator()):
(JSC::globalFuncProtoGetter):
(JSC::GlobalFuncProtoSetterFunctor::GlobalFuncProtoSetterFunctor):
(JSC::GlobalFuncProtoSetterFunctor::allowsAccess):
(JSC::GlobalFuncProtoSetterFunctor::operator()):
(JSC::globalFuncProtoSetter):

• runtime/ObjectConstructor.cpp:

(JSC::ObjectConstructorGetPrototypeOfFunctor::ObjectConstructorGetPrototypeOfFunctor):
(JSC::ObjectConstructorGetPrototypeOfFunctor::result):
(JSC::ObjectConstructorGetPrototypeOfFunctor::operator()):
(JSC::objectConstructorGetPrototypeOf):

Source/WebCore:

No new tests.

• bindings/js/JSXMLHttpRequestCustom.cpp:

(WebCore::SendFunctor::SendFunctor):
(WebCore::SendFunctor::hasViableFrame):
(WebCore::SendFunctor::operator()):
(WebCore::JSXMLHttpRequest::send):

• bindings/js/ScriptCallStackFactory.cpp:

(WebCore::CreateScriptCallStackFunctor::CreateScriptCallStackFunctor):
(WebCore::CreateScriptCallStackFunctor::operator()):
(WebCore::createScriptCallStack):
(WebCore::CreateScriptCallStackForConsoleFunctor::CreateScriptCallStackForConsoleFunctor):
(WebCore::CreateScriptCallStackForConsoleFunctor::operator()):

5:16 PM Changeset in webkit [155012] by benjamin@webkit.org

• 2 edits in trunk/LayoutTests

Try unskipping compositing/images/positioned-image-content-rect.html

Unreviewed.

• platform/mac/TestExpectations:

The test compositing/images/positioned-image-content-rect.html seems to pass reliably
on the bots and locally.
Try to unskip it.

5:04 PM Changeset in webkit [155011] by Lucas Forschler

• 5 edits in branches/safari-537.60-branch/Source

Versioning.

4:57 PM Changeset in webkit [155010] by Lucas Forschler

• 1 copy in tags/Safari-537.60.1

New Tag.

4:56 PM Changeset in webkit [155009] by benjamin@webkit.org

• 2 edits in trunk/LayoutTests

The test inspector/geolocation-success.html is unreliable
​https://bugs.webkit.org/show_bug.cgi?id=120655

Reviewed by Alexey Proskuryakov.

• inspector/geolocation-success.html:

The test was expecting everything would be done in the page context
when InspectorTest.evaluateInPage invoke the callback.

This is not the case, geolocation APIs are asynchronous.
The callbacks printLocation() and printError() may or may not be executed.

To fix this, the execution of each step is changed to depends on the completion
of the geolocation callback.

4:21 PM Changeset in webkit [155008] by oliver@apple.com

• 10 edits
  3 adds in trunk

Support structured clone of Map and Set
​https://bugs.webkit.org/show_bug.cgi?id=120654

Reviewed by Simon Fraser.

Source/JavaScriptCore:

Make xcode copy the required headers, and add appropriate export attributes

• JavaScriptCore.xcodeproj/project.pbxproj:
• runtime/JSMap.h:
• runtime/JSSet.h:
• runtime/MapData.h:

Source/WebCore:

Add support for cloning Map and Set. Fairly self explanatory change.
Needed to add Forwarding headers for the JSMap, JSSet and MapData classes.

• ForwardingHeaders/runtime/JSMap.h: Added.
• ForwardingHeaders/runtime/JSSet.h: Added.
• ForwardingHeaders/runtime/MapData.h: Added.
• bindings/js/SerializedScriptValue.cpp:

(WebCore::CloneSerializer::isMap):
(WebCore::CloneSerializer::isSet):
(WebCore::CloneSerializer::startSet):
(WebCore::CloneSerializer::startMap):
(WebCore::CloneSerializer::serialize):
(WebCore::CloneDeserializer::consumeMapDataTerminationIfPossible):
(WebCore::CloneDeserializer::deserialize):

LayoutTests:

Tests!

• fast/dom/Window/script-tests/postmessage-clone.js:

(set new):
(set add.set add):

4:04 PM Changeset in webkit [155007] by Lucas Forschler

• 1 edit in branches/safari-537.60-branch/Source/WebKit/win/WebFrame.cpp

Windows build fix.

3:36 PM Changeset in webkit [155006] by Lucas Forschler

• 5 edits in branches/safari-537-branch/Source

Versioning.

3:34 PM Changeset in webkit [155005] by Lucas Forschler

• 1 copy in tags/Safari-537.67

New Tag.

3:31 PM Changeset in webkit [155004] by Lucas Forschler

• 1 edit in branches/safari-537.60-branch/Source/WebKit/win/WebView.cpp

Windows build fix after 155001.

3:20 PM Changeset in webkit [155003] by Lucas Forschler

• 3 edits in branches/safari-537.60-branch/Source/WebKit/win

Merged r154764. <rdar://problem/14879688>

3:14 PM Changeset in webkit [155002] by betravis@adobe.com

• 6 edits
  2 adds in trunk

[CSS Shapes] Shape's content gets extra left offset when left-border is positive on the content box
​https://bugs.webkit.org/show_bug.cgi?id=117573

Reviewed by David Hyatt.

Source/WebCore:

Nested blocks need to take into account their offset from the shape-inside container.
The new code calculates the offset from the shape-inside container, then applies the
offset to the computed segments. The line must be moved down by the offset's height,
and each segment must be moved left by the offset's width.

Test: fast/shapes/shape-inside/shape-inside-offset-block-children.html

• rendering/RenderBlock.cpp:

(WebCore::RenderBlock::relayoutShapeDescendantIfMoved): Relayout a block child if its
new logical left would cause it to rest at a new position within a shape container.
(WebCore::RenderBlock::logicalOffsetFromShapeAncestorContainer): Calculate the logical
offset form a shape inside ancestor container.
(WebCore::RenderBlock::layoutBlockChild): Call relayoutShapeDescendantIfMoved with the
new position offset.

• rendering/RenderBlock.h:
• rendering/RenderBlockLineLayout.cpp:

(WebCore::RenderBlock::updateShapeAndSegmentsForCurrentLine): Use layout offset, rather
than just vertical offset.
(WebCore::RenderBlock::updateShapeAndSegmentsForCurrentLineInFlowThread): Ditto.
(WebCore::RenderBlock::layoutRunsAndFloatsInRange): Ditto.

• rendering/shapes/ShapeInsideInfo.h:

(WebCore::ShapeInsideInfo::computeSegmentsForLine): Shift segments logically left when
there is an inline offset.

LayoutTests:

Test that nested children with padding correctly apply an ancestor's shape-inside
across different writing modes.

• fast/shapes/shape-inside/shape-inside-offset-block-children-expected.html: Added.
• fast/shapes/shape-inside/shape-inside-offset-block-children.html: Added.

3:08 PM Changeset in webkit [155001] by Lucas Forschler

• 3 edits in branches/safari-537.60-branch/Source/WebKit/win

Merged r154759. <rdar://problem/14879679>

2:47 PM Changeset in webkit [155000] by ap@apple.com

• 2 edits in trunk/Tools

[Mac] WebKitTestRunner still beeps sometimes
​https://bugs.webkit.org/show_bug.cgi?id=120652

Reviewed by Tim Horton.

In bug 107251, we disabled beeping in WebProcess, but some of the beeps happen in
UI process (notably, AppKit beeps when handling a key equivalent returns NO).

• WebKitTestRunner/mac/TestControllerMac.mm: (WTR::TestController::platformInitialize): Use the same SPI that we use in DRT and in WebProcess to disable beeping.

2:36 PM Changeset in webkit [154999] by benjamin@webkit.org

• 2 edits in trunk/Websites/webkit.org

Fix the recommended testing platform on the website

Rubberstamped by Enrica Casucci.

• quality/testing.html:

1:16 PM Changeset in webkit [154998] by Antoine Quint

• 9 edits in trunk

Web Inspector: exceptions triggered from console evaluation do not pause the debugger
​https://bugs.webkit.org/show_bug.cgi?id=120460

Source/WebCore:

Reviewed by Timothy Hatcher.

• inspector/InjectedScriptSource.js:

Explicitly set a sourceURL such that the frontend may identify injected script when
processing call frames in order to hide such code from the debugger.

Source/WebInspectorUI:

We used to preclude any debugging from errors stemming from code evaluated in the console
as we would always set the doNotPauseOnExceptionsAndMuteConsole parameter to "false" when
calling JavaScriptLogViewController._evaluateInInspectedWindow(). However, it is desirable
to allow debugging code ran from the console.

We now allow debugging in such a scenario and we filter out call frames coming from the
Web Inspector injected script as well as the call frame for the console prompt such that
we only pause in the debugger in case the exception is in code under the console prompt
and not the console code prompt itself.

Additionally, to prevent stepping out to call frames we may have filtered out, we disable
the "step out" button in cases where there are no further frames in the frontend to go out to.

Reviewed by Timothy Hatcher.

• UserInterface/DebuggerManager.js:

(WebInspector.DebuggerManager.prototype.debuggerDidPause):
Filter out call frames that have a URL coming from Web Inspector injected script by looking
for a URL starting with the "WebInspector" prefix. If we determine that there are no call
frames left after filtering, we resume code evaluation such that we only pause in the debugger
when the exception is in code evluated under the console prompt.

• UserInterface/DebuggerSidebarPanel.js:

(WebInspector.DebuggerSidebarPanel):
(WebInspector.DebuggerSidebarPanel.prototype._debuggerDidPause):
(WebInspector.DebuggerSidebarPanel.prototype._debuggerActiveCallFrameDidChange):
Monitor any change to the active call frame such that we may tie the state of the
"step out" button to the availability of a call frame to step out to in the filtered
list set on the DebuggerManager.

• UserInterface/JavaScriptLogViewController.js:

(WebInspector.JavaScriptLogViewController.prototype.consolePromptTextCommitted):
Set the doNotPauseOnExceptionsAndMuteConsole to "false" when calling _evaluateInInspectedWindow()
in order to allow pausing on exceptions coming from code evalued in the console. Also, explicitly
set a sourceURL for the script to evaluate such that we may identify its origin when filtering
call frames stemming from inspector code.

• UserInterface/ResourceSidebarPanel.js:

(WebInspector.ResourceSidebarPanel.prototype._scriptWasAdded):
Filter out any script resource starting with the Web Inspector-specific "WebInspector" prefix
so that injected script does not show up.

LayoutTests:

Reviewed by Timothy Hatcher.

• platform/mac/inspector/console/command-line-api-expected.txt:

Take into account the addition of a sourceURL to inspector/InjectedScriptSource.js.

12:30 PM Changeset in webkit [154997] by akling@apple.com

• 6 edits in trunk/Source

Support Vector<Ref<T>>.
<​https://webkit.org/b/120637>

Reviewed by Antti Koivisto.

Source/WebCore:

Use Vector<Ref<T>> internally in Page.

• page/Page.cpp:

(WebCore::networkStateChanged):
(WebCore::Page::refreshPlugins):

Clean up these functions and use Vector<Ref<Frame>> to store pointers to frames
since we know they are not going to be null.

(WebCore::Page::pluginViews):

Changed this to return a Vector<Ref<PluginView>> by value instead of passing a
writable vector as an argument. Clean up loops with 'auto'.

(WebCore::Page::storageBlockingStateChanged):
(WebCore::Page::privateBrowsingStateChanged):

Tweaked for pluginViews() returning a Vector now.

(WebCore::Page::setVisibilityState):

Store Documents that need a visibilitychange event in a Vector<Ref<Document>>.

Source/WTF:

Add a Ref(const T&) constructor to enable Vector<Ref<T>>. This looks a bit awkward
but is necessary for Vector::append(const T&) to find a constructor.

An alternative would be to add something like std::vector::emplace_back, but I can't
think of a good name for that, and it'd be nice if append() would "just work."

Also add operator->(). I initially excluded this because I felt it made for
unsafe-looking code. Things quickly got out of hand with .get() everywhere though.

IMO -> looks OK as long as it's only used on the first link in a dereference chain,
as that variable and its type will be "in context."

• wtf/Ref.h:

(WTF::Ref::Ref):
(WTF::Ref::~Ref):
(WTF::Ref::operator->):
(WTF::Ref::get):

• wtf/VectorTraits.h:

Add simple traits for Ref<T> so it can be moved around freely by Vector.

12:14 PM Changeset in webkit [154996] by svillar@igalia.com

• 10 edits in trunk

[CSS Grid Layout] Add parsing for named grid lines
​https://bugs.webkit.org/show_bug.cgi?id=119540

Reviewed by Andreas Kling.

From Blink r150381,r150587 by <​jchaffraix@chromium.org>

Source/WebCore:

Adds parsing support for named grid lines at <grid-line> level,
i.e., inside grid-{row|column}-{start|end}. This change covers
only the parsing, layout changes coming in a follow up patch.

• css/CSSComputedStyleDeclaration.cpp:

(WebCore::valueForGridPosition):

• css/CSSParser.cpp:

(WebCore::CSSParser::parseIntegerOrStringFromGridPosition):
(WebCore::CSSParser::parseGridPosition):

• css/CSSParser.h:
• css/StyleResolver.cpp:

(WebCore::createGridPosition):

• rendering/RenderGrid.cpp:

(WebCore::RenderGrid::resolveGridPositionFromStyle):

• rendering/style/GridPosition.h:

(WebCore::GridPosition::setExplicitPosition):
(WebCore::GridPosition::setSpanPosition):
(WebCore::GridPosition::integerPosition):
(WebCore::GridPosition::namedGridLine):

LayoutTests:

Added several new test cases which include different types of
named grid lines, this means including the names at different
positions, multiple names per line or mixing names with keywords
like 'span'.

• fast/css-grid-layout/grid-item-column-row-get-set-expected.txt:
• fast/css-grid-layout/grid-item-column-row-get-set.html:

12:00 PM Changeset in webkit [154995] by ap@apple.com

• 3 edits
  1 add in trunk

[Mac] Hyphenation respects regional format settings language instead of primary language
​https://bugs.webkit.org/show_bug.cgi?id=120641

Reviewed by Dan Bernstein.

Fixes hyphenation tests on my machine with non-English regional format settings.

• platform/text/cf/HyphenationCF.cpp: (createValueForNullKey): Use primary UI language for hyphenation, not regional settings language.

12:00 PM Changeset in webkit [154994] by commit-queue@webkit.org

• 3 edits in trunk/Source/WebCore

[GTK][EFL] include missing localized strings for subtitle auto track
​https://bugs.webkit.org/show_bug.cgi?id=120629

those methods are necessary to show the "Auto" track on webkitgtk

Patch by Danilo Cesar Lemes de Paula <​danilo.cesar@collabora.co.uk> on 2013-09-03
Reviewed by Gustavo Noronha Silva.

• platform/efl/LocalizedStringsEfl.cpp:

(WebCore::textTrackAutomaticMenuItemText):

• platform/gtk/LocalizedStringsGtk.cpp:

(WebCore::textTrackAutomaticMenuItemText):

11:51 AM Changeset in webkit [154993] by dbates@webkit.org

• 9 edits
  2 adds in trunk

Require layout when -webkit-overflow-scrolling changes
​https://bugs.webkit.org/show_bug.cgi?id=120535

Reviewed by Darin Adler.

Source/WebCore:

Test: fast/repaint/overflow-scroll-touch-repaint.html

We want to require a layout when the value of -webkit-overflow-scrolling changes
since -webkit-overflow-scrolling creates a stacking context.

• rendering/style/RenderStyle.cpp:

(WebCore::RenderStyle::changeRequiresLayout):

LayoutTests:

Add a test to ensure we perform a layout and repaint when the
value of -webkit-overflow-scrolling changes on an element.

At the time of writing, the iOS port enables ACCELERATED_OVERFLOW_SCROLLING.

• fast/repaint/overflow-scroll-touch-repaint-expected.txt: Added.
• fast/repaint/overflow-scroll-touch-repaint.html: Added.
• platform/efl/TestExpectations: Skip test fast/repaint/overflow-scroll-touch-repaint.html

since this platform doesn't enable ACCELERATED_OVERFLOW_SCROLLING.

• platform/gtk/TestExpectations: Ditto.
• platform/mac/TestExpectations: Ditto.
• platform/qt/TestExpectations: Ditto.
• platform/win/TestExpectations: Ditto.
• platform/wincairo/TestExpectations: Ditto.

11:45 AM Changeset in webkit [154992] by rniwa@webkit.org

• 10 edits
  9 adds in trunk

Support the "json" responseType and JSON response entity in XHR
​https://bugs.webkit.org/show_bug.cgi?id=73648

Reviewed by Oliver Hunt.

Source/JavaScriptCore:

Based on the patch written by Jarred Nicholls.

Add JSC::JSONParse. This function will be used in XMLHttpRequest.response of type 'json'.

• JavaScriptCore.xcodeproj/project.pbxproj:
• runtime/JSONObject.cpp:

(JSC::JSONParse):

• runtime/JSONObject.h:

Source/WebCore:

Based on the patch written by Jarred Nicholls.

Implement 'json' type for XMLHttpRequest.response. We cache the result on JSC side as a cached attribute
unlike other response types like 'document' and 'blob' for which the parsed response object is cached
in XMLHttpRequest itself. In the long run, we should do the same for other types of response types.

Also refactored the various code to share the code.

Tests: fast/xmlhttprequest/xmlhttprequest-responsetype-json-invalid.html

fast/xmlhttprequest/xmlhttprequest-responsetype-json-utf16.html
fast/xmlhttprequest/xmlhttprequest-responsetype-json-valid.html

• ForwardingHeaders/runtime/JSONObject.h: Added.

• bindings/js/JSXMLHttpRequestCustom.cpp:

(WebCore::JSXMLHttpRequest::visitChildren):
(WebCore::JSXMLHttpRequest::response): Use JSONParse to parse the response text and cache the result.
Call didCacheResponseJSON to set the cache status and clear the original response buffer.

• xml/XMLHttpRequest.cpp:

(WebCore::XMLHttpRequest::XMLHttpRequest): Added m_responseCacheIsValid to invalidate the cache of
a json response.
(WebCore::XMLHttpRequest::responseText):
(WebCore::XMLHttpRequest::didCacheResponseJSON): Added; Updates m_responseCacheIsValid and clears the
response buffer to save memory.
(WebCore::XMLHttpRequest::responseXML):
(WebCore::XMLHttpRequest::setResponseType):
(WebCore::XMLHttpRequest::responseType):
(WebCore::XMLHttpRequest::clearResponseBuffers):
(WebCore::XMLHttpRequest::didReceiveData):

• xml/XMLHttpRequest.h:

(WebCore::XMLHttpRequest::doneWithoutErrors): Extracted from responseXML.
(WebCore::XMLHttpRequest::responseTextIgnoringResponseType): Extracted from responseText.
(WebCore::XMLHttpRequest::responseCacheIsValid): Added.
(WebCore::XMLHttpRequest::shouldDecodeResponse): Extracted from didReceiveData.
Also modified to decode when the response type is ResponseTypeJSON.

• xml/XMLHttpRequest.idl: Added CachedAttribute IDL extention on response property. This cache is

used when the response type is 'json'.

LayoutTests:

Add regression tests for XMLHttpRequest.response of type 'json'.

Two of these tests (valid & invalid) come from Jarred Nicholls's original patch.

• fast/xmlhttprequest/resources/xmlhttprequest-responsetype-json-utf-16.json: Added.
• fast/xmlhttprequest/resources/xmlhttprequest-responsetype-json.json: Added.
• fast/xmlhttprequest/xmlhttprequest-responsetype-json-invalid-expected.txt: Added.
• fast/xmlhttprequest/xmlhttprequest-responsetype-json-invalid.html: Added.
• fast/xmlhttprequest/xmlhttprequest-responsetype-json-utf16-expected.txt: Added.
• fast/xmlhttprequest/xmlhttprequest-responsetype-json-utf16.html: Added.
• fast/xmlhttprequest/xmlhttprequest-responsetype-json-valid-expected.txt: Added.
• fast/xmlhttprequest/xmlhttprequest-responsetype-json-valid.html: Added.

11:30 AM Changeset in webkit [154991] by commit-queue@webkit.org

• 3 edits
  2 deletes in trunk

Unreviewed, rolling out r154881.
​http://trac.webkit.org/changeset/154881
​https://bugs.webkit.org/show_bug.cgi?id=120643

Crashes on macworld.com (Requested by kling on #webkit).

Source/WebCore:

• dom/Element.cpp:

(WebCore::Element::setAttributeInternal):

LayoutTests:

• fast/dom/Element/setAttributeNode-for-existing-attribute-expected.txt: Removed.
• fast/dom/Element/setAttributeNode-for-existing-attribute.html: Removed.

10:33 AM Changeset in webkit [154990] by mikhail.pozdnyakov@intel.com

• 2 edits in trunk/Source/WTF

Check WTF::VectorFiller template argument type size in compile time
​https://bugs.webkit.org/show_bug.cgi?id=120631

Reviewed by Darin Adler.

The template argument's type size in WTF::VectorFiller 'memset' specialization
should be checked during compilation rather than in runtime.

• wtf/Vector.h:

10:08 AM Changeset in webkit [154989] by commit-queue@webkit.org

• 3 edits in trunk/Source/WebKit2

[GTK] gdk threads deprecated functions calls should be refactored
​https://bugs.webkit.org/show_bug.cgi?id=120070

Patch by Anton Obzhirov <Anton Obzhirov> on 2013-09-03
Reviewed by Mario Sanchez Prada.

Removed deprecated functions gdk_threads_leave()/gdk_threads_enter() functions since
there is no more checks for threads lock in GTK 3.6.

• UIProcess/API/gtk/WebKitWebView.cpp:

(webkitWebViewRunAsModal):

• UIProcess/gtk/WebPopupMenuProxyGtk.cpp:

(WebKit::WebPopupMenuProxyGtk::showPopupMenu):

9:51 AM Changeset in webkit [154988] by commit-queue@webkit.org

• 3 edits in trunk/Source/WebCore

[GStreamer] Don't set state to NULL until element is destroyed
​https://bugs.webkit.org/show_bug.cgi?id=117354

Patch by Andre Moreira Magalhaes <Andre Moreira Magalhaes> on 2013-09-03
Reviewed by Philippe Normand.

Don't set playbin to NULL until it is going to be destroyed or if we stay
for too long on the READY state. Instead only set the state to READY as this
allows much faster state changes to PAUSED/PLAYING again. playbin internally
caches some state that is destroyed when setting it to NULL.
This state is independent of the URI and it is even possible to change the
URI in READY state.

To avoid having resources (e.g. audio devices) open indefinitely,
when setting the state to READY we create a timeout and if the timeout
is reached we reset the pipeline state to NULL to free resources.

Also now all state changes use the changePipelineState method instead of setting
the playbin state directly with gst_element_set_state, so we have a better control
of when we are requesting state changes.

• platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:

(WebCore::mediaPlayerPrivateReadyStateTimeoutCallback):
(WebCore::MediaPlayerPrivateGStreamer::MediaPlayerPrivateGStreamer):
(WebCore::MediaPlayerPrivateGStreamer::~MediaPlayerPrivateGStreamer):
(WebCore::MediaPlayerPrivateGStreamer::commitLoad):
(WebCore::MediaPlayerPrivateGStreamer::changePipelineState):
(WebCore::MediaPlayerPrivateGStreamer::setRate):
(WebCore::MediaPlayerPrivateGStreamer::handlePluginInstallerResult):

• platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.h:

9:51 AM Changeset in webkit [154987] by commit-queue@webkit.org

• 4 edits in trunk/Source/WebKit2

[GTK] [WK2] TestContextMenu default-menu fails
​https://bugs.webkit.org/show_bug.cgi?id=120459

Patch by Brian Holt <​brian.holt@samsung.com> on 2013-09-03
Reviewed by Gustavo Noronha Silva.

Add context menu items for downloading media elements.

• UIProcess/API/gtk/WebKitContextMenuActions.cpp:

(webkitContextMenuActionGetActionTag):
(webkitContextMenuActionGetForContextMenuItem):
(webkitContextMenuActionGetLabel):

• UIProcess/API/gtk/WebKitContextMenuActions.h:
• UIProcess/API/gtk/tests/TestContextMenu.cpp:

9:39 AM Changeset in webkit [154986] by fpizlo@apple.com

• 18 edits
  2 adds
  2 deletes in trunk/Source/JavaScriptCore

CodeBlock::jettison() should be implicit
​https://bugs.webkit.org/show_bug.cgi?id=120567

Reviewed by Oliver Hunt.

This is a risky change from a performance standpoint, but I believe it's
necessary. This makes all CodeBlocks get swept by GC. Nobody but the GC
can delete CodeBlocks because the GC always holds a reference to them.
Once a CodeBlock reaches just one reference (i.e. the one from the GC)
then the GC will free it only if it's not on the stack.

This allows me to get rid of the jettisoning logic. We need this for FTL
tier-up. Well; we don't need it, but it will help prevent a lot of bugs.
Previously, if you wanted to to replace one code block with another, you
had to remember to tell the GC that the previous code block is
"jettisoned". We would need to do this when tiering up from DFG to FTL
and when dealing with DFG-to-FTL OSR entry code blocks. There are a lot
of permutations here - tiering up to the FTL, OSR entering into the FTL,
deciding that an OSR entry code block is not relevant anymore - just to
name a few. In each of these cases we'd have to jettison the previous
code block. It smells like a huge source of future bugs.

So I made jettisoning implicit by making the GC always watch out for a
CodeBlock being owned solely by the GC.

This change is performance neutral.

• CMakeLists.txt:
• GNUmakefile.list.am:
• JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
• JavaScriptCore.xcodeproj/project.pbxproj:
• Target.pri:
• bytecode/CodeBlock.cpp:

(JSC::CodeBlock::CodeBlock):
(JSC::CodeBlock::~CodeBlock):
(JSC::CodeBlock::visitAggregate):
(JSC::CodeBlock::jettison):

• bytecode/CodeBlock.h:

(JSC::CodeBlock::setJITCode):
(JSC::CodeBlock::shouldImmediatelyAssumeLivenessDuringScan):
(JSC::CodeBlockSet::mark):

• dfg/DFGCommonData.h:

(JSC::DFG::CommonData::CommonData):

• heap/CodeBlockSet.cpp: Added.

(JSC::CodeBlockSet::CodeBlockSet):
(JSC::CodeBlockSet::~CodeBlockSet):
(JSC::CodeBlockSet::add):
(JSC::CodeBlockSet::clearMarks):
(JSC::CodeBlockSet::deleteUnmarkedAndUnreferenced):
(JSC::CodeBlockSet::traceMarked):

• heap/CodeBlockSet.h: Added.
• heap/ConservativeRoots.cpp:

(JSC::ConservativeRoots::add):

• heap/ConservativeRoots.h:
• heap/DFGCodeBlocks.cpp: Removed.
• heap/DFGCodeBlocks.h: Removed.
• heap/Heap.cpp:

(JSC::Heap::markRoots):
(JSC::Heap::deleteAllCompiledCode):
(JSC::Heap::deleteUnmarkedCompiledCode):

• heap/Heap.h:
• interpreter/JSStack.cpp:

(JSC::JSStack::gatherConservativeRoots):

• interpreter/JSStack.h:
• runtime/Executable.cpp:

(JSC::ScriptExecutable::installCode):

• runtime/Executable.h:
• runtime/VM.h:

8:32 AM Changeset in webkit [154985] by allan.jensen@digia.com

• 2 edits in trunk/Source/WebKit/qt

[Qt] Images scaled poorly on composited canvas
​https://bugs.webkit.org/show_bug.cgi?id=120632

Reviewed by Jocelyn Turcotte.

Explicitly set a imageInterpolationQuality on the TextureMapper, because
InterpolationDefault may be interpreted differently by nested GraphicsContexts.

• WebCoreSupport/TextureMapperLayerClientQt.cpp:

(TextureMapperLayerClientQt::renderCompositedLayers):

7:50 AM Changeset in webkit [154984] by commit-queue@webkit.org

• 2 edits in trunk/Source/WebCore

[WinCairo] Unneeded code in method GlyphPage::fill().
​https://bugs.webkit.org/show_bug.cgi?id=120634

Patch by ​peavo@outlook.com <​peavo@outlook.com> on 2013-09-03
Reviewed by Andreas Kling.

• platform/graphics/win/GlyphPageTreeNodeCairoWin.cpp:

(WebCore::GlyphPage::fill): Remove unneeded call to GetTextMetrics() function.

7:47 AM Changeset in webkit [154983] by allan.jensen@digia.com

• 2 edits in trunk/Source/WebKit/qt

[Qt] Tiled-backing store not clipped to frame or visible rect
​https://bugs.webkit.org/show_bug.cgi?id=120606

Reviewed by Jocelyn Turcotte.

Clip painting from the tiled-backing store to the frame rect.

• WebCoreSupport/QWebFrameAdapter.cpp:

(QWebFrameAdapter::renderFromTiledBackingStore):

7:43 AM Changeset in webkit [154982] by mihnea@adobe.com

• 3 edits
  4 adds in trunk

[CSSRegions] Pseudo-elements as regions should not be exposed to JS
​https://bugs.webkit.org/show_bug.cgi?id=120633

Reviewed by Andreas Kling.

Source/WebCore:

Until we properly implement the Region interface (​http://dev.w3.org/csswg/css-regions/#the-region-interface)
for pseudo-elements, we should not return these as regions in JS.

Tests: fast/regions/get-regions-by-content-pseudo.html

fast/regions/webkit-named-flow-get-regions-pseudo.html

• dom/WebKitNamedFlow.cpp:

(WebCore::WebKitNamedFlow::firstEmptyRegionIndex): Skip pseudo-elements as regions here too,
otherwise we may get an index that cannot be used with getRegions().
(WebCore::WebKitNamedFlow::getRegionsByContent):
(WebCore::WebKitNamedFlow::getRegions):

LayoutTests:

Add tests with pseudo-elements as regions checking the following API:
WebKitNamedFlow::getRegions(), WebKitNamedFlow::getRegionsByContent()
Because we do not return the pseudo-elements as regions in the region list,
i modified WebKitNamedFlow::firstEmptyRegionIndex to skip these regions too.

• fast/regions/get-regions-by-content-pseudo-expected.txt: Added.
• fast/regions/get-regions-by-content-pseudo.html: Added.
• fast/regions/webkit-named-flow-get-regions-pseudo-expected.txt: Added.
• fast/regions/webkit-named-flow-get-regions-pseudo.html: Added.

7:20 AM Changeset in webkit [154981] by zandobersek@gmail.com

• 2 edits in trunk/Source/WebCore

REGRESSION(r154967): http appcache tests crashing on WK1
​https://bugs.webkit.org/show_bug.cgi?id=120620

Reviewed by Andreas Kling.

• loader/appcache/ApplicationCacheGroup.cpp:

(WebCore::ApplicationCacheGroup::cacheDestroyed): Reintroduce pre-r154967 behavior that returned early in
this method if the passed-in ApplicationCache object was not found in the ApplicationCacheGroup's HashSet
of all the caches. This is now done by checking that the HashSet<T>::remove(T) returns true (meaning the
object was found in the HashSet and removed from it) in addition to that HashSet being subsequently empty
before the method moves on to destroying its ApplicationCacheGroup instance.

6:36 AM Changeset in webkit [154980] by kadam@inf.u-szeged.hu

• 3 edits in trunk/LayoutTests

[Qt] Unreviewed gardening. Skip some failing tests.

• platform/qt-wk1/TestExpectations:
• platform/qt/TestExpectations:

6:25 AM Changeset in webkit [154979] by commit-queue@webkit.org

• 2 edits in trunk/LayoutTests

[EFL] accessibility/aria-describedby-on-input.html is failing
​https://bugs.webkit.org/show_bug.cgi?id=112027

Unreviewed EFL gardening.

accessibility/aria-describedby-on-input.html passes after r154976.

Patch by Krzysztof Czech <​k.czech@samsung.com> on 2013-09-03

• platform/efl-wk2/TestExpectations:

6:25 AM Changeset in webkit [154978] by Patrick Gansterer

• 2 edits in trunk

[CMake] Fix detection of x86_64 platform with MSVC
​https://bugs.webkit.org/show_bug.cgi?id=116662

Reviewed by Gyuyoung Kim.

Use ${MSVC_CXX_ARCHITECTURE_ID} instead of ${CMAKE_SYSTEM_PROCESSOR}, since
the later one just resolves to the host processor on Windows.

• CMakeLists.txt:

6:16 AM Changeset in webkit [154977] by commit-queue@webkit.org

• 3 edits
  3 adds in trunk

[gstreamer] Disable HTTP request "Accept-Encoding:" header field on gstreamer source element to avoid receiving the wrong size when retrieving data
​https://bugs.webkit.org/show_bug.cgi?id=115354

Patch by Andre Moreira Magalhaes <Andre Moreira Magalhaes> on 2013-09-03
Reviewed by Philippe Normand.

Source/WebCore:

Also disable Accept-Encoding on ResourceRequest::toSoupMessage accordingly.

• platform/network/soup/ResourceRequestSoup.cpp:

(WebCore::ResourceRequest::toSoupMessage):
Call ResourceRequest::updateSoupMessage from ResourceRequest::toSoupMessage so that the
Accept-Encoding header is also respected.

LayoutTests:

Add test to check that video requests will send no "Accept-Encoding" header.

• http/tests/media/video-accept-encoding-expected.txt: Added.
• http/tests/media/video-accept-encoding.cgi: Added.
• http/tests/media/video-accept-encoding.html: Added.

4:29 AM Changeset in webkit [154976] by commit-queue@webkit.org

• 15 edits
  2 moves in trunk

Source/WebCore: [AX][ATK] Added support for sort and help attributes.
​https://bugs.webkit.org/show_bug.cgi?id=120456

Patch by Krzysztof Czech <​k.czech@samsung.com> on 2013-09-03
Reviewed by Chris Fleizach.

Added support for aria-sort and aria-help attributes.

Test: accessibility/aria-sort.html

• accessibility/atk/WebKitAccessibleWrapperAtk.cpp:

(webkitAccessibleGetAttributes):

Tools: [AX][ATK] Added support for sort and help attributes
​https://bugs.webkit.org/show_bug.cgi?id=120456

Patch by Krzysztof Czech <​k.czech@samsung.com> on 2013-09-03
Reviewed by Chris Fleizach.

Added missing implementation to AccessibilityUIElement::helpText and support for
aria-sort attribute.

• DumpRenderTree/atk/AccessibilityUIElementAtk.cpp:

(coreAttributeToAtkAttribute):

• WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:

(WTR::coreAttributeToAtkAttribute):
(WTR::AccessibilityUIElement::helpText):

LayoutTests: [AX][ATK] Added support for sort and help attributes
​https://bugs.webkit.org/show_bug.cgi?id=120456

Patch by Krzysztof Czech <​k.czech@samsung.com> on 2013-09-03
Reviewed by Chris Fleizach.

Sharing aria-sort.html specific mac test with efl and gtk.
Changing specific expectations of some accessibility tests.

• accessibility/aria-sort-expected.txt: Renamed from LayoutTests/platform/mac/accessibility/aria-sort-expected.txt.
• accessibility/aria-sort.html: Renamed from LayoutTests/platform/mac/accessibility/aria-sort.html.
• platform/efl-wk2/TestExpectations:
• platform/efl-wk2/accessibility/image-link-expected.txt:
• platform/efl-wk2/accessibility/image-map2-expected.txt:
• platform/efl-wk2/accessibility/table-cell-spans-expected.txt:
• platform/efl-wk2/accessibility/table-cells-expected.txt:
• platform/gtk/accessibility/image-link-expected.txt:
• platform/gtk/accessibility/image-map2-expected.txt:
• platform/gtk/accessibility/table-cell-spans-expected.txt:
• platform/gtk/accessibility/table-cells-expected.txt:

3:10 AM Changeset in webkit [154975] by allan.jensen@digia.com

• 4 edits
  3 deletes in trunk/Source/WebCore

[Qt] Remove dead code for QtXmlPatterns
​https://bugs.webkit.org/show_bug.cgi?id=120624

Reviewed by Simon Hausmann.

Remove code supporting XSLT using QtXmlPatterns which has been
dead for some time.

• Target.pri:
• WebCore.pri:
• dom/TransformSourceQt.cpp: Removed.
• xml/XSLStyleSheetQt.cpp: Removed.
• xml/XSLTProcessorQt.cpp: Removed.
• xml/parser/XMLDocumentParserQt.cpp:

(WebCore::XMLDocumentParser::doEnd):
(WebCore::XMLDocumentParser::parseProcessingInstruction):

1:54 AM Changeset in webkit [154974] by allan.jensen@digia.com

• 5 edits in trunk

[Qt][WK1] PageVisibility tests are flaky
​https://bugs.webkit.org/show_bug.cgi?id=120418

Reviewed by Jocelyn Turcotte.

Source/WebKit/qt:

• WebCoreSupport/DumpRenderTreeSupportQt.cpp:

(DumpRenderTreeSupportQt::resetPageVisibility):

• WebCoreSupport/DumpRenderTreeSupportQt.h:

Tools:

Implement resetPageVisibility so we can reset visibility without
emiting visibility state change events.

• DumpRenderTree/qt/TestRunnerQt.cpp:

(TestRunner::resetPageVisibility):

1:21 AM Changeset in webkit [154973] by mihnea@adobe.com

• 12 edits
  20 adds in trunk

[CSS Regions] position: fixed is computed relative to the first region, not the viewport
​https://bugs.webkit.org/show_bug.cgi?id=111176

Reviewed by David Hyatt.

Source/WebCore:

Fixed positioned elements inside a named flow should be positioned and sized relative to the viewport,
not on the first region, as described in the spec: ​http://dev.w3.org/csswg/css-regions/#the-flow-into-property.
While the flow thread will still act as containing block for the fixed positioned elements, the painting and hit
testing for the fixed positioned elements is done by RenderView. The layers for the fixed positioned elements
are collected by the flow thread.

Tests: fast/regions/element-in-named-flow-absolute-from-fixed.html

fast/regions/element-in-named-flow-fixed-from-absolute.html
fast/regions/element-inflow-fixed-from-outflow-static.html
fast/regions/element-outflow-static-from-inflow-fixed.html
fast/regions/fixed-element-transformed-parent.html
fast/regions/fixed-in-named-flow-scroll.html
fast/regions/fixed-inside-fixed-in-named-flow.html
fast/regions/fixed-inside-named-flow-zIndex.html
fast/regions/fixed-pos-elem-in-namedflow-noregions.html
fast/regions/fixed-pos-region-in-nested-flow.html

• rendering/FlowThreadController.cpp:

(WebCore::FlowThreadController::collectFixedPositionedLayers):
Return the list of layers for the fixed positioned elements with named flows as containing blocks.
Used in RenderLayer:: paintFixedLayersInNamedFlows and RenderLayer:: hitTestFixedLayersInNamedFlows.

• rendering/FlowThreadController.h:
• rendering/RenderBox.cpp:

(WebCore::RenderBox::mapLocalToContainer):
(WebCore::RenderBox::containingBlockLogicalWidthForPositioned):
(WebCore::RenderBox::containingBlockLogicalHeightForPositioned):
For fixed positioned elements, width and height are given by the view instead of the first region.

• rendering/RenderLayer.cpp:

(WebCore::accumulateOffsetTowardsAncestor):
Modified for the fixed positioned elements inside named flows with the named flows as containing block
to take into account the view scroll.
(WebCore::compareZIndex):
Moved upwards because it is used in RenderLayer::paintFixedLayersInNamedFlows.
(WebCore::RenderLayer::paintFixedLayersInNamedFlows):
Paint the list of fixed layers directly from the RenderView instead of painting them through regions -> named flows.
(WebCore::RenderLayer::paintLayerContents):
(WebCore::RenderLayer::paintList):
(WebCore::RenderLayer::hitTestFixedLayersInNamedFlows):
Hit test the layers for the fix positioned elements inside named flows from the RenderView layer
instead of the region -> named flow layer.
(WebCore::RenderLayer::hitTestLayer):
(WebCore::RenderLayer::calculateRects):

• rendering/RenderLayer.h:
• rendering/RenderLayerCompositor.cpp:

(WebCore::RenderLayerCompositor::updateLayerTreeGeometry):
We do not support yet the accelerated compositing for elements in named flows,
so bail out early here. We need to revisit fixed elements once we finish accelerated compositing for elements in named flows.

• rendering/RenderObject.cpp:

(WebCore::RenderObject::fixedPositionedWithNamedFlowContainingBlock):
(WebCore::hasFixedPosInNamedFlowContainingBlock):
(WebCore::RenderObject::containerForRepaint):
Changed to take into account that RenderView should be the repaintContainer for the
fixed positioned elements with named flow as the containing block.

• rendering/RenderObject.h:

LayoutTests:

Added new tests and changed the existing ones that were relying on incorrect positioning
for fixed elements.

• fast/regions/element-in-named-flow-absolute-from-fixed-expected.txt: Added.
• fast/regions/element-in-named-flow-absolute-from-fixed.html: Added.
• fast/regions/element-in-named-flow-fixed-from-absolute-expected.txt: Added.
• fast/regions/element-in-named-flow-fixed-from-absolute.html: Added.
• fast/regions/element-inflow-fixed-from-outflow-static-expected.txt: Added.
• fast/regions/element-inflow-fixed-from-outflow-static.html: Added.
• fast/regions/element-outflow-static-from-inflow-fixed-expected.txt: Added.
• fast/regions/element-outflow-static-from-inflow-fixed.html: Added.
• fast/regions/fixed-element-transformed-parent-expected.txt: Added.
• fast/regions/fixed-element-transformed-parent.html: Added.
• fast/regions/fixed-in-named-flow-scroll-expected.txt: Added.
• fast/regions/fixed-in-named-flow-scroll.html: Added.
• fast/regions/fixed-inside-fixed-in-named-flow-expected.html: Added.
• fast/regions/fixed-inside-fixed-in-named-flow.html: Added.
• fast/regions/fixed-inside-named-flow-zIndex-expected.html: Added.
• fast/regions/fixed-inside-named-flow-zIndex.html: Added.
• fast/regions/fixed-pos-elem-in-namedflow-noregions-expected.html: Added.
• fast/regions/fixed-pos-elem-in-namedflow-noregions.html: Added.
• fast/regions/fixed-pos-elem-in-region-expected.html:
• fast/regions/fixed-pos-elem-in-region.html:
• fast/regions/fixed-pos-region-in-nested-flow-expected.html: Added.
• fast/regions/fixed-pos-region-in-nested-flow.html: Added.

1:15 AM Changeset in webkit [154972] by ChangSeok Oh

• 2 edits in trunk/Tools

[GTK] libsoup upversion to fix a gstreamer issue, bug115354
​https://bugs.webkit.org/show_bug.cgi?id=120613

Reviewed by Philippe Normand.

Up version of libsoup to 2.43.90. But the exact version which we'll use is not 2.43.90.
To fix bug115354, we need the Andre's patch for libsoup, but the lastest release
does not contain it. ​https://bugzilla.gnome.org/show_bug.cgi?id=706338
For the reason, we'll use libsoup git repo directly for a while until the next
libsoup release.

• gtk/jhbuild.modules:

12:29 AM Changeset in webkit [154971] by anilsson@rim.com

• 7 edits in trunk/Source/WebCore

[BlackBerry] Remove LayerData::LayerProgram
​https://bugs.webkit.org/show_bug.cgi?id=120601

Reviewed by Anders Carlsson.

JIRA 490672
All layer contents are RGBA now, so there's no need to support BGRA any
more and the LayerData::LayerProgram enum can be removed. Related dead
code was removed.

No new tests, no change in behavior.

• platform/graphics/blackberry/EGLImageLayerWebKitThread.cpp:

(WebCore::EGLImageLayerWebKitThread::EGLImageLayerWebKitThread):

• platform/graphics/blackberry/LayerData.h:

(WebCore::LayerData::LayerData):

• platform/graphics/blackberry/LayerRenderer.cpp:

(WebCore::LayerRenderer::compositeLayersRecursive):
(WebCore::LayerRenderer::createProgram):

• platform/graphics/blackberry/LayerRenderer.h:
• platform/graphics/blackberry/LayerWebKitThread.h:
• platform/graphics/blackberry/PluginLayerWebKitThread.cpp:

(WebCore::PluginLayerWebKitThread::setPluginView):

12:16 AM Changeset in webkit [154970] by calvaris@igalia.com

• 5 edits in trunk/Source/WebCore

[GStreamer] Video player sets system volume to 100%
​https://bugs.webkit.org/show_bug.cgi?id=118974

Reviewed by Philippe Normand.

In order to preserve the system volume we need to keep track of
the volume being initialized in the HTMLMediaElement and then just
setting the volume to the sink when initializing the pipeline if
that volume was changed before.

• html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::HTMLMediaElement): Initialized
attribute to false.
(WebCore::HTMLMediaElement::setVolume): Set the attribute to true
when volume is changed.
(WebCore::HTMLMediaElement::updateVolume): Set the volume only if
volume was initialized.
(WebCore::HTMLMediaElement::mediaPlayerPlatformVolumeConfigurationRequired):
Platform volume configuration is required only if volume was not
initialized before.

• html/HTMLMediaElement.h: Added attribute and interface method.
• platform/graphics/MediaPlayer.h:

(WebCore::MediaPlayerClient::mediaPlayerPlatformVolumeConfigurationRequired):
Declared and added default implementation for the interface method.
(WebCore::MediaPlayer::platformVolumeConfigurationRequired):
Asked the client, meaning the HTMLMediaElement if the platform
volume configuration is required.

• platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:

(WebCore::mediaPlayerPrivateVolumeChangedCallback): Added log.
(WebCore::MediaPlayerPrivateGStreamerBase::setVolume): Added log.
(WebCore::MediaPlayerPrivateGStreamerBase::setStreamVolumeElement):
Set the volume only if not platform volume is required and added log.

Sep 2, 2013:

11:06 PM Changeset in webkit [154969] by Lucas Forschler

• 6 edits
  24 copies in branches/safari-537-branch

Merged r154785. <rdar://problem/14664586>

11:02 PM Changeset in webkit [154968] by Darin Adler

• 2 edits in trunk/Source/WebCore

• inspector/InspectorProfilerAgent.cpp:

(WebCore::InspectorProfilerAgent::removeProfile): Fix braces here; a review
comment I forgot to address in my last check-in.

11:00 PM Changeset in webkit [154967] by Darin Adler

• 39 edits in trunk/Source

Cut down on double hashing and code needlessly using hash table iterators
​https://bugs.webkit.org/show_bug.cgi?id=120611

Reviewed by Andreas Kling.

Source/WebCore:

Some of these changes are primarily code cleanup, but others could provide
a small code size and speed improvement by avoiding extra hashing.

• Modules/geolocation/Geolocation.cpp:

(WebCore::Geolocation::Watchers::find): Use get instead of find.
(WebCore::Geolocation::Watchers::remove): Use take instead of find.
(WebCore::Geolocation::makeCachedPositionCallbacks): Use the return
value from remove to avoid hashing twice.

• Modules/webaudio/AudioContext.cpp:

(WebCore::AudioContext::addAutomaticPullNode): Use the return value from
add to avoid hashing twice.
(WebCore::AudioContext::removeAutomaticPullNode): Use the return value
from remove to avoid hashing twice.

• Modules/webaudio/AudioNodeInput.cpp:

(WebCore::AudioNodeInput::connect): Use the return value from add to avoid
hashing twice.
(WebCore::AudioNodeInput::disconnect): Use the return value from remove
to avoid hashing twice.

• Modules/webaudio/AudioParam.cpp:

(WebCore::AudioParam::connect): Use the return value from add to avoid
hashing twice.
(WebCore::AudioParam::disconnect): Use the return value from remove to
avoid hashing twice.

• bridge/NP_jsobject.cpp:

(ObjectMap::remove): Use remove instead of find/remove.

• dom/Node.cpp:

(WebCore::Node::~Node): Use the return value from remove instead of
find/remove.

• inspector/InspectorProfilerAgent.cpp:

(WebCore::InspectorProfilerAgent::removeProfile): Remove needless
calls to contains.

• loader/DocumentLoader.cpp:

(WebCore::DocumentLoader::removeSubresourceLoader): Use the return
value from remove instead of find/remove.

• loader/ResourceLoadScheduler.cpp:

(WebCore::ResourceLoadScheduler::HostInformation::remove): Use the
return value from remove to avoid hashing twice.

• loader/appcache/ApplicationCacheGroup.cpp:

(WebCore::ApplicationCacheGroup::disassociateDocumentLoader): Use
remove instead of find/remove.
(WebCore::ApplicationCacheGroup::cacheDestroyed): Removed a needless
call to contains to avoid hashing twice. It's fine to do the check
for an empty hash table unconditionally.

• page/DOMWindow.cpp:

(WebCore::addUnloadEventListener): Eliminated a local variable for clarity.
(WebCore::removeUnloadEventListener): Ditto. Also use remove instead
of find/remove.
(WebCore::removeAllUnloadEventListeners): Ditto. Also use removeAll instead
of find/removeAll.
(WebCore::addBeforeUnloadEventListener): Ditto.
(WebCore::removeBeforeUnloadEventListener): Ditto.
(WebCore::removeAllBeforeUnloadEventListeners): Ditto.

• page/FrameView.cpp:

(WebCore::FrameView::removeViewportConstrainedObject): Use the return
value from remove to avoid hashing twice.
(WebCore::FrameView::removeScrollableArea): Use the return value from
remove instead of find/remove.
(WebCore::FrameView::containsScrollableArea): Use && instead of an if
statement in a way that is idiomatic for this kind of function.

• page/Page.cpp:

(WebCore::Page::addRelevantRepaintedObject): Use the return value from
remove instead of find/remove.

• page/PageGroup.cpp:

(WebCore::PageGroup::removeUserScriptsFromWorld): Use remove instead
of find/remove.
(WebCore::PageGroup::removeUserStyleSheetsFromWorld): Use the return
value from remove instead of find/remove.

• page/PerformanceUserTiming.cpp:

(WebCore::clearPeformanceEntries): Removed a needless call to contains.

• platform/graphics/DisplayRefreshMonitor.cpp:

(WebCore::DisplayRefreshMonitor::removeClient): Use the return value
from remove instead of find/remove.
(WebCore::DisplayRefreshMonitorManager::displayDidRefresh): Use remove
instead of find/remove.

• platform/graphics/blackberry/LayerRenderer.cpp:

(WebCore::LayerRenderer::removeLayer): Use the return value from remove
instead of find/remove.

• platform/win/WindowMessageBroadcaster.cpp:

(WebCore::WindowMessageBroadcaster::removeListener): Use remove instead
of find/remove. It's fine to do the check for an empty hash table unconditionally.

• plugins/PluginDatabase.cpp:

(WebCore::PluginDatabase::removeDisabledPluginFile): Use the return value
from remove instead of find/remove.

• rendering/style/StyleCustomFilterProgramCache.cpp:

(WebCore::StyleCustomFilterProgramCache::lookup): Use get instead of find.
(WebCore::StyleCustomFilterProgramCache::add): Use contains instead of find
in an assertion.
(WebCore::StyleCustomFilterProgramCache::remove): Use remove instead of
find/remove.

• svg/SVGCursorElement.cpp:

(WebCore::SVGCursorElement::removeClient): Use the return value from remove
instead of find/remove.

• svg/SVGDocumentExtensions.cpp:

(WebCore::SVGDocumentExtensions::removeResource): Removed an unneeded call
to contains.
(WebCore::SVGDocumentExtensions::removeAllTargetReferencesForElement): Use
remove instead of find/remove. It's fine to do the check for an empty hash
table unconditionally.
(WebCore::SVGDocumentExtensions::removeAllElementReferencesForTarget): Use
remove instead of find/remove. Also removed unhelpful assertions. One is
already done by HashMap, and the other is just checking a basic invariant
of every HashMap that doesn't need to be checked.

• svg/graphics/SVGImageCache.cpp:

(WebCore::SVGImageCache::removeClientFromCache): Removed an unneeded call
to contains.

• svg/properties/SVGAnimatedProperty.cpp:

(WebCore::SVGAnimatedProperty::~SVGAnimatedProperty): Use the version of
remove that takes an iterator rather than the one that takes a key, so
we don't need to redo the hashing.

Source/WebKit2:

• Platform/CoreIPC/Connection.cpp:

(CoreIPC::Connection::waitForMessage): Use take instead of find/remove.

• UIProcess/WebPreferences.cpp:

(WebKit::WebPreferences::removePageGroup): Use the return value from remove
instead of find/remove.

• WebProcess/Geolocation/GeolocationPermissionRequestManager.cpp:

(WebKit::GeolocationPermissionRequestManager::cancelRequestForGeolocation):
(WebKit::GeolocationPermissionRequestManager::didReceiveGeolocationPermissionDecision):
Use take instead of find/remove.

• WebProcess/Plugins/Netscape/NetscapePlugin.cpp:

(WebKit::NetscapePlugin::frameDidFinishLoading): Use take instead of find/remove.
(WebKit::NetscapePlugin::frameDidFail): Use take instead of find/remove.

• WebProcess/WebPage/WebBackForwardListProxy.cpp:

(WebKit::WebBackForwardListProxy::removeItem): Use take instead of find/remove.

• WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::didFinishCheckingText): Use take instead of get/remove so we
hash only once.
(WebKit::WebPage::didCancelCheckingText): Ditto.
(WebKit::WebPage::stopExtendingIncrementalRenderingSuppression): Use the return
value from remove instead of contains/remove so we hash only once.

Source/WTF:

Double hashing is common in code that needs to combine a remove with some
action to only be done if the code is removed. The only way to avoid it is
to write code using find and a hash table iterator. To help with this, add
a boolean return value to remove functions to indicate if anything was removed.

Double hashing also happens in code that does a get followed by a remove.
The take function is helpful in this case. To help with this, add a takeFirst
funciton to ListHashSet.

• wtf/HashCountedSet.h:

(WTF::HashCountedSet::removeAll): Added a boolean return value, analogous to the one
that the HashCountedSet::remove function already has.

• wtf/HashMap.h:

(WTF::HashMap::remove): Added a boolean return value, true if something was removed.

• wtf/HashSet.h:

(WTF::HashSet::remove): Ditto.

• wtf/RefPtrHashMap.h:

(WTF::RefPtrHashMap::remove): Ditto.

• wtf/ListHashSet.h:

(WTF::ListHashSet::takeFirst): Added.
(WTF::ListHashSet::takeLast): Added.
(WTF::ListHashSet::remove): Added a boolean return value, true if something was removed.

• wtf/WTFThreadData.h:

(JSC::IdentifierTable::remove): Use the new remove return value to get rid of most of
the code in this function.

1:30 PM Changeset in webkit [154966] by ddkilzer@apple.com

• 2 edits in trunk/Source/WTF

Remove duplicate entries found by Xcode in WTF project

Platform.h was duplicated in r111778 after being added in
r111504.

A dangling reference to Ref.h was added in r154962.

• WTF.xcodeproj/project.pbxproj: Remove duplicate entries for

Platform.h and Ref.h.

1:10 PM Changeset in webkit [154965] by akling@apple.com

• 41 edits in trunk/Source/WebCore

Generate isFooElement() functions from tagname data.
<​https://webkit.org/b/120584>

Reviewed by Antti Koivisto.

Add a "generateTypeChecks" attribute that can be used in HTMLTagNames.in & friends.
If present, isFooElement() style helpers will be added to HTMLElementTypeChecks.h.
This also outputs an isElementOfType<T> check for the Element iterators.

Removed all the hand-written isFooElement() functions that only checked tag name.

• html/HTMLTagNames.in:
• svg/svgtags.in:

Added "generateTypeChecks" attribute as appropriate.

• GNUmakefile.am:
• GNUmakefile.list.am:
• WebCore.vcxproj/WebCore.vcxproj:
• WebCore.xcodeproj/project.pbxproj:

Added to build systems based on how HTMLNames.h was done.
We're just outputting an additional header file in the generated code directory
so I suspect most ports will just pick this up automagically.

• dom/make_names.pl:

(defaultTagPropertyHash):
(printLicenseHeader):
(printTypeChecks):
(printTypeChecksHeaderFile):

Generate a separate file for each namespace with isFooElement() helpers for
elements with "generateTypeChecks" attribute set.

12:41 PM Changeset in webkit [154964] by akling@apple.com

• 1 edit
  1 add in trunk/Source/WTF

Actually add Ref.h

11:55 AM Changeset in webkit [154963] by Darin Adler

• 21 edits in trunk

[Mac] No need for HardAutorelease, which is same as CFBridgingRelease
​https://bugs.webkit.org/show_bug.cgi?id=120569

Reviewed by Andy Estes.

Source/JavaScriptCore:

• API/JSValue.mm:

(valueToString): Use CFBridgingRelease.

Source/WebCore:

• accessibility/mac/WebAccessibilityObjectWrapperMac.mm:

(AXTextMarkerRange):
(AXTextMarkerRangeStart):
(AXTextMarkerRangeEnd):
(textMarkerForVisiblePosition):
Use CFBridgingRelease.

• platform/mac/KURLMac.mm:

(WebCore::KURL::operator NSURL *): Use CFBridgingRelease.
(WebCore::KURL::createCFURL): Get rid of needless local variable.

• platform/mac/WebCoreNSURLExtras.mm:

(WebCore::mapHostNameWithRange):
(WebCore::URLWithData):
(WebCore::userVisibleString):

• platform/text/mac/StringImplMac.mm:

(WTF::StringImpl::operator NSString *):
Use CFBridgingRelease.

Source/WebKit/mac:

• Misc/WebNSFileManagerExtras.mm:

(-[NSFileManager _webkit_startupVolumeName]): Removed some unneeded locals.
Got rid of the pointless ref/leakRef/HardAutorelease dance, and replaced it
with a [[x copy] autorelease].

• Misc/WebNSURLExtras.mm:

(-[NSURL _web_URLWithLowercasedScheme]): Use CFBridgingRelease, and got rid
of unneeded type casts.

• Plugins/WebBasePluginPackage.mm:

(+[WebBasePluginPackage preferredLocalizationName]): Use CFBridgingRelease.

• WebView/WebPDFRepresentation.mm:

(-[WebPDFRepresentation convertPostScriptDataSourceToPDF:]): Ditto.

• WebView/WebView.mm:

(+[WebView _setCacheModel:]): Use CFBridgingRelease and got rid of unneeded
type cast.

Source/WebKit2:

• Platform/mac/StringUtilities.mm:

(WebKit::nsStringFromWebCoreString): Use CFBridgingRelease. Also
changed condition to be a little cleaner and use a constant string for empty
strings as well as null strings.

• UIProcess/API/mac/WKBrowsingContextController.mm:

(autoreleased): Switched from autorelease to CFBridgingRelease for strings,
which eliminates a type cast and makes this work under GC, although I don't
think we should compile WebKit2 for GC.

• WebProcess/WebPage/mac/WKAccessibilityWebPageObject.mm:

(-[WKAccessibilityWebPageObject accessibilityAttributeValue:forParameter:]):
Use CFBridgingRelease.

Source/WTF:

• wtf/ObjcRuntimeExtras.h: Added a FIXME about miscapitalization of ObjC.

Deleted HardAutorelease.
(wtfObjcMsgSend): Dropped the use of abbreviations in local class and argument names.
(wtfCallIMP): Ditto.

Tools:

• DumpRenderTree/mac/DumpRenderTree.mm:

(dump): Use CFBridgingRelease.

11:50 AM Changeset in webkit [154962] by akling@apple.com

• 68 edits in trunk/Source

Ref: A smart pointer for the reference age.
<​https://webkit.org/b/120570>

Reviewed by Antti Koivisto.

Source/WebCore:

Use Ref<T> for various stack guards where null checking isn't needed.

Source/WTF:

Add a very simple simple Ref<T> smart pointer class that is never null.
It's initialized by passing a T& to the constructor and cannot be assigned to.

operator-> is not overloaded, to prevent unsafe-looking code.
The value is extracted by "T& get()", since C++ does not let you override operator.()

• wtf/Ref.h:

10:05 AM Changeset in webkit [154961] by akling@apple.com

• 6 edits in trunk/Source/WebCore

Simplify DocumentType handling.
<​https://webkit.org/b/120529>

Reviewed by Antti Koivisto.

Removed the insertedInto()/removedFrom() handlers from DocumentType.

Document no longer keeps a pointer to its doctype node, it was only used for the
document.doctype DOM API, which now just looks through the list of (<=2) children.

The ENABLE(LEGACY_VIEWPORT_ADAPTION) hunk from Document::setDocType() was moved
into Document::childrenChanged().

We no longer clear the style resolver on doctype insertion/removal since it
doesn't actually affect style anyway.

Also made doctype() return a PassRefPtr<DocumentType> instead of a raw pointer.

• dom/Document.cpp:

(WebCore::Document::dispose):
(WebCore::Document::doctype):
(WebCore::Document::childrenChanged):

• dom/Document.h:
• dom/DocumentType.cpp:
• dom/DocumentType.h:
• editing/markup.cpp:

(WebCore::documentTypeString):

8:44 AM Changeset in webkit [154960] by commit-queue@webkit.org

• 10 edits in trunk

<​https://webkit.org/b/98350> [GTK] accessibility/aria-invalid.html times out

Patch by Anton Obzhirov <Anton Obzhirov> on 2013-09-02
Reviewed by Mario Sanchez Prada.

Source/WebCore:

The patch exposes aria-invalid attribute to ATK.

• accessibility/atk/AXObjectCacheAtk.cpp:

(WebCore::AXObjectCache::postPlatformNotification):
Added emitting state-change signal for aria-invalid event.

• accessibility/atk/WebKitAccessibleWrapperAtk.cpp:

(webkitAccessibleGetAttributes):
Added aria-invalid attribute.

Tools:

Added few mappings in DumpRenderTree and WebKitTestRunner for aria-invalid in order to get the tests run properly.

• DumpRenderTree/atk/AccessibilityCallbacksAtk.cpp:

(axObjectEventListener):
Added mapping for invalid-entry event parameter.

• DumpRenderTree/atk/AccessibilityUIElementAtk.cpp:

(coreAttributeToAtkAttribute):
Added mapping to aria-invalid.
(AccessibilityUIElement::stringAttributeValue):

• WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:

(WTR::coreAttributeToAtkAttribute):
Added mapping to aria-invalid.

LayoutTests:

Unskipped accessibility/aria-invalid.html which is passing now.

• platform/gtk/TestExpectations: Removed passing test.
• platform/gtk-wk2/TestExpectations: Added test timing out in WK2 only.

8:43 AM Changeset in webkit [154959] by Csaba Osztrogonác

• 2 edits in trunk/Source/WebKit2

REGRESSION(r154909): caused many crashes on Qt WK2, EFL WK2
​https://bugs.webkit.org/show_bug.cgi?id=120600

Reviewed by Andreas Kling.

• Shared/CoordinatedGraphics/CoordinatedGraphicsArgumentCoders.cpp:

(CoreIPC::::decode): keyTime should be double here too.

8:22 AM Changeset in webkit [154958] by akling@apple.com

• 41 edits in trunk/Source/WebCore

Unreviewed, rolling out r154955.
​http://trac.webkit.org/changeset/154955
​https://bugs.webkit.org/show_bug.cgi?id=120605

broke xcode4 build :| (Requested by kling on #webkit).

Patch by Commit Queue <​commit-queue@webkit.org> on 2013-09-02

• GNUmakefile.am:
• GNUmakefile.list.am:
• WebCore.vcxproj/WebCore.vcxproj:
• WebCore.xcodeproj/project.pbxproj:
• dom/make_names.pl:

(defaultTagPropertyHash):
(printLicenseHeader):

• html/HTMLAnchorElement.h:

(WebCore::isHTMLAnchorElement):
(WebCore::HTMLAnchorElement):

• html/HTMLAreaElement.h:

(WebCore::isHTMLAreaElement):
(WebCore::HTMLAreaElement):

• html/HTMLAudioElement.h:

(WebCore::isHTMLAudioElement):

• html/HTMLBaseElement.h:

(WebCore::isHTMLBaseElement):
(WebCore::HTMLBaseElement):

• html/HTMLCanvasElement.h:

(WebCore::isHTMLCanvasElement):

• html/HTMLElement.h:
• html/HTMLFieldSetElement.h:

(WebCore::isHTMLFieldSetElement):
(WebCore::HTMLFieldSetElement):

• html/HTMLFormElement.h:

(WebCore::isHTMLFormElement):

• html/HTMLFrameSetElement.h:

(WebCore::isHTMLFrameSetElement):
(WebCore::HTMLFrameSetElement):

• html/HTMLImageElement.h:

(WebCore::isHTMLImageElement):

• html/HTMLInputElement.h:

(WebCore::isHTMLInputElement):

• html/HTMLLabelElement.h:

(WebCore::isHTMLLabelElement):
(WebCore::HTMLLabelElement):

• html/HTMLLegendElement.h:

(WebCore::isHTMLLegendElement):
(WebCore::HTMLLegendElement):

• html/HTMLMapElement.h:

(WebCore::isHTMLMapElement):

• html/HTMLMeterElement.h:

(WebCore::isHTMLMeterElement):

• html/HTMLOptGroupElement.h:

(WebCore::isHTMLOptGroupElement):

• html/HTMLOptionElement.h:

(WebCore::isHTMLOptionElement):

• html/HTMLParamElement.h:

(WebCore::isHTMLParamElement):
(WebCore::HTMLParamElement):

• html/HTMLProgressElement.h:

(WebCore::isHTMLProgressElement):

• html/HTMLScriptElement.h:

(WebCore::isHTMLScriptElement):

• html/HTMLSourceElement.h:

(WebCore::isHTMLSourceElement):
(WebCore::HTMLSourceElement):

• html/HTMLStyleElement.h:

(WebCore::isHTMLStyleElement):
(WebCore::HTMLStyleElement):

• html/HTMLTableElement.h:

(WebCore::isHTMLTableElement):

• html/HTMLTableRowElement.h:

(WebCore::isHTMLTableRowElement):
(WebCore::HTMLTableRowElement):

• html/HTMLTagNames.in:
• html/HTMLTextAreaElement.h:

(WebCore::isHTMLTextAreaElement):

• html/HTMLTitleElement.h:

(WebCore::isHTMLTitleElement):
(WebCore::HTMLTitleElement):

• html/HTMLTrackElement.h:

(WebCore::isHTMLTrackElement):
(WebCore::HTMLTrackElement):

• svg/SVGElement.h:
• svg/SVGFontElement.h:

(WebCore::isSVGFontElement):

• svg/SVGFontFaceElement.h:

(WebCore::isSVGFontFaceElement):
(WebCore::SVGFontFaceElement):

• svg/SVGForeignObjectElement.h:

(WebCore::isSVGForeignObjectElement):
(WebCore::SVGForeignObjectElement):

• svg/SVGImageElement.h:

(WebCore::isSVGImageElement):

• svg/SVGScriptElement.h:

(WebCore::isSVGScriptElement):

• svg/svgtags.in:

8:17 AM Changeset in webkit [154957] by Antti Koivisto

• 68 edits in trunk/Source/WebCore

Clean up ContainerNode::childrenChanged
​https://bugs.webkit.org/show_bug.cgi?id=120599

Reviewed by Andreas Kling.

• Make childrenChanged take a single struct argument instead of a long list of arguments.
• Use enum instead of childCountDelta. It was always -1, 0, 1 or the total number of children (in case of removing them all).
• Remove use of Node*, give the change range as Elements.
• Related cleanups.

• dom/Attr.cpp:

(WebCore::Attr::childrenChanged):

• dom/Attr.h:
• dom/CharacterData.cpp:

(WebCore::CharacterData::parserAppendData):
(WebCore::CharacterData::dispatchModifiedEvent):

• dom/ContainerNode.cpp:

(WebCore::ContainerNode::insertBefore):
(WebCore::ContainerNode::notifyChildInserted):
(WebCore::ContainerNode::notifyChildRemoved):

Add private helpers for setting up the struct.

(WebCore::ContainerNode::parserInsertBefore):
(WebCore::ContainerNode::replaceChild):
(WebCore::ContainerNode::removeChild):
(WebCore::ContainerNode::parserRemoveChild):
(WebCore::ContainerNode::removeChildren):
(WebCore::ContainerNode::appendChild):
(WebCore::ContainerNode::parserAppendChild):
(WebCore::ContainerNode::childrenChanged):
(WebCore::ContainerNode::updateTreeAfterInsertion):

• dom/ContainerNode.h:
• dom/Document.cpp:

(WebCore::Document::childrenChanged):

• dom/Document.h:
• dom/Element.cpp:

(WebCore::checkForSiblingStyleChanges):

Clean up and simplify. Since we now get element range automatically we don't need to compute it.

(WebCore::Element::childrenChanged):
(WebCore::Element::finishParsingChildren):

• dom/Element.h:
• dom/ShadowRoot.cpp:

(WebCore::ShadowRoot::childrenChanged):

• dom/ShadowRoot.h:
• html/HTMLElement.cpp:

(WebCore::HTMLElement::childrenChanged):
(WebCore::HTMLElement::adjustDirectionalityIfNeededAfterChildrenChanged):

Try to keep the existing behavior. This code needs more cleanup to be sane. It shouldn't operate on Nodes
as it only really cares about Elements.

• html/HTMLElement.h:
• html/HTMLFieldSetElement.cpp:

(WebCore::HTMLFieldSetElement::childrenChanged):

• html/HTMLFieldSetElement.h:
• html/HTMLObjectElement.cpp:

(WebCore::HTMLObjectElement::childrenChanged):

• html/HTMLObjectElement.h:
• html/HTMLOptGroupElement.cpp:

(WebCore::HTMLOptGroupElement::childrenChanged):

• html/HTMLOptGroupElement.h:
• html/HTMLOptionElement.cpp:

(WebCore::HTMLOptionElement::childrenChanged):

• html/HTMLOptionElement.h:
• html/HTMLOutputElement.cpp:

(WebCore::HTMLOutputElement::childrenChanged):

• html/HTMLOutputElement.h:
• html/HTMLScriptElement.cpp:

(WebCore::HTMLScriptElement::childrenChanged):

• html/HTMLScriptElement.h:
• html/HTMLSelectElement.cpp:

(WebCore::HTMLSelectElement::childrenChanged):

• html/HTMLSelectElement.h:
• html/HTMLStyleElement.cpp:

(WebCore::HTMLStyleElement::childrenChanged):

• html/HTMLStyleElement.h:
• html/HTMLTextAreaElement.cpp:

(WebCore::HTMLTextAreaElement::childrenChanged):

• html/HTMLTextAreaElement.h:
• html/HTMLTitleElement.cpp:

(WebCore::HTMLTitleElement::childrenChanged):

• html/HTMLTitleElement.h:
• html/shadow/InsertionPoint.cpp:

(WebCore::InsertionPoint::childrenChanged):

• html/shadow/InsertionPoint.h:
• svg/SVGClipPathElement.cpp:

(WebCore::SVGClipPathElement::childrenChanged):

• svg/SVGClipPathElement.h:
• svg/SVGElement.cpp:

(WebCore::SVGElement::childrenChanged):

• svg/SVGElement.h:
• svg/SVGFELightElement.cpp:

(WebCore::SVGFELightElement::childrenChanged):

• svg/SVGFELightElement.h:
• svg/SVGFilterElement.cpp:

(WebCore::SVGFilterElement::childrenChanged):

• svg/SVGFilterElement.h:
• svg/SVGFilterPrimitiveStandardAttributes.cpp:

(WebCore::SVGFilterPrimitiveStandardAttributes::childrenChanged):

• svg/SVGFilterPrimitiveStandardAttributes.h:
• svg/SVGFontFaceElement.cpp:

(WebCore::SVGFontFaceElement::childrenChanged):

• svg/SVGFontFaceElement.h:
• svg/SVGFontFaceFormatElement.cpp:

(WebCore::SVGFontFaceFormatElement::childrenChanged):

• svg/SVGFontFaceFormatElement.h:
• svg/SVGFontFaceSrcElement.cpp:

(WebCore::SVGFontFaceSrcElement::childrenChanged):

• svg/SVGFontFaceSrcElement.h:
• svg/SVGFontFaceUriElement.cpp:

(WebCore::SVGFontFaceUriElement::childrenChanged):

• svg/SVGFontFaceUriElement.h:
• svg/SVGGradientElement.cpp:

(WebCore::SVGGradientElement::childrenChanged):

• svg/SVGGradientElement.h:
• svg/SVGMarkerElement.cpp:

(WebCore::SVGMarkerElement::childrenChanged):

• svg/SVGMarkerElement.h:
• svg/SVGMaskElement.cpp:

(WebCore::SVGMaskElement::childrenChanged):

• svg/SVGMaskElement.h:
• svg/SVGPatternElement.cpp:

(WebCore::SVGPatternElement::childrenChanged):

• svg/SVGPatternElement.h:
• svg/SVGScriptElement.cpp:

(WebCore::SVGScriptElement::childrenChanged):

• svg/SVGScriptElement.h:
• svg/SVGStyleElement.cpp:

(WebCore::SVGStyleElement::childrenChanged):

• svg/SVGStyleElement.h:
• svg/SVGTitleElement.cpp:

(WebCore::SVGTitleElement::childrenChanged):

• svg/SVGTitleElement.h:

8:15 AM Changeset in webkit [154956] by kadam@inf.u-szeged.hu

• 2 edits in trunk/LayoutTests

[Qt] Unreviewed gardening. Skip some failing tests.

• platform/qt/TestExpectations:

8:12 AM Changeset in webkit [154955] by akling@apple.com

• 41 edits in trunk/Source/WebCore

Generate isFooElement() functions from tagname data.
<​https://webkit.org/b/120584>

Reviewed by Antti Koivisto.

Add a "generateTypeChecks" attribute that can be used in HTMLTagNames.in & friends.
If present, isFooElement() style helpers will be added to HTMLElementTypeChecks.h.
This also outputs an isElementOfType<T> check for the Element iterators.

Removed all the hand-written isFooElement() functions that only checked tag name.

• html/HTMLTagNames.in:
• svg/svgtags.in:

Added "generateTypeChecks" attribute as appropriate.

• GNUmakefile.am:
• GNUmakefile.list.am:
• WebCore.vcxproj/WebCore.vcxproj:
• WebCore.xcodeproj/project.pbxproj:

Added to build systems based on how HTMLNames.h was done.
We're just outputting an additional header file in the generated code directory
so I suspect most ports will just pick this up automagically.

• dom/make_names.pl:

(defaultTagPropertyHash):
(printLicenseHeader):
(printTypeChecks):
(printTypeChecksHeaderFile):

Generate a separate file for each namespace with isFooElement() helpers for
elements with "generateTypeChecks" attribute set.

8:03 AM Changeset in webkit [154954] by krit@webkit.org

• 6 edits
  4 adds in trunk

Use edgeMode=duplicate for blurring on filter() function
​https://bugs.webkit.org/show_bug.cgi?id=120590

Reviewed by Antti Koivisto.

Source/WebCore:

Filters on the CSS Image function filter() are not allowed to extend the
dimension of the input image. This causes weird results on blurring an image,
where the fading on the edges is clipped at the half of the fading.
We shouldn't fade edges at all and use the edgeMode=duplicate instead.
This will duplicate the pixel value on the nearest edge of the input image
instead of taking transparent black and results in nice blurred images with
sharp edges.

Spec: ​http://dev.w3.org/fxtf/filters/#blurEquivalent

Test: fast/filter-image/filter-image-blur.html

• css/CSSFilterImageValue.cpp: Pass consumer information to the renderer.

(WebCore::CSSFilterImageValue::image):

• rendering/FilterEffectRenderer.cpp: Set edgeMode for feGaussianBlur to

'duplicate' or 'none' depending on the consumer.

(WebCore::FilterEffectRenderer::build):

• rendering/FilterEffectRenderer.h: Add enumeration to differ between the

different consumers of the renderer.

• rendering/RenderLayer.cpp:

(WebCore::RenderLayer::updateOrRemoveFilterEffectRenderer): Pass consumer

information to the renderer.

LayoutTests:

Added test to check that filter(<image>, blur(<value>)) takes
the edgeMode 'duplicate' instead of none.

• fast/filter-image/filter-image-blur-expected.html: Added.
• fast/filter-image/filter-image-blur.html: Added.
• fast/filter-image/resources/svg-blur.svg: Added.
• fast/filter-image/resources/svg-noblur.svg: Added.

7:40 AM Changeset in webkit [154953] by commit-queue@webkit.org

• 2 edits in trunk/Source/WebKit/gtk

[ATK] Leak: Leaks in testatk.c
​https://bugs.webkit.org/show_bug.cgi?id=118675

Patch by Brian Holt <​brian.holt@samsung.com> on 2013-09-02
Reviewed by Mario Sanchez Prada.

Fixed memory leaks by matching ref calls with unrefs.

• tests/testatk.c:

(testWebkitAtkCaretOffsets):
(testWebkitAtkCaretOffsetsAndExtranousWhiteSpaces):
(testWebkitAtkGetTextAtOffset):
(testWebkitAtkGetTextAtOffsetNewlines):
(testWebkitAtkGetTextAtOffsetTextarea):
(testWebkitAtkGetTextAtOffsetTextInput):
(testWebkitAtkGetTextInParagraphAndBodySimple):
(testWebkitAtkGetTextInParagraphAndBodyModerate):
(testWebkitAtkTextAttributes):
(testWebkitAtkTextSelections):
(testWebkitAtkListsOfItems):

7:37 AM Changeset in webkit [154952] by zarvai@inf.u-szeged.hu

• 2 edits in trunk/LayoutTests

[Qt] Unreviewed gardening.
​https://bugs.webkit.org/show_bug.cgi?id=120595

Patch by Gabor Abraham <​abrhm@inf.u-szeged.hu> on 2013-09-02

• platform/qt/TestExpectations: Skipping failing xss-DENIED tests.

5:48 AM Changeset in webkit [154951] by commit-queue@webkit.org

• 2 edits in trunk/Tools

Save md5 correctly when jhbuildPath doesn't exist yet
​https://bugs.webkit.org/show_bug.cgi?id=120548

Patch by Nick Diego Yamane <​nick.yamane@openbossa.org> on 2013-09-02
Reviewed by Gustavo Noronha Silva.

After r152605, Md5 for jhbuild files are saved before the update
process, this prevents the script to restart update from scratch
when initial checkouts fail. However it causes an issue when builddir
(or builddir/Dependencies) doesn't exist yet. In that case the
saveJhbuildMd5 function fails to create md5 files.
This patch adds a checking for the jhbuildPath and creates it if
necessary before trying to open the md5 files.

• Scripts/update-webkit-libs-jhbuild:

(saveJhbuildMd5):

5:41 AM Changeset in webkit [154950] by mario@webkit.org

• 2 edits in trunk/Tools

Unreviewed. Move myself to the reviewers list.

• Scripts/webkitpy/common/config/contributors.json: