Timeline
Apr 21, 2014:
- 10:44 PM Changeset in webkit [167650] by
-
- 4 edits in trunk
Source/WebCore: [Mac] Follow-up: After a horizontal overflow scroll with a mechanical wheel mouse, vertical scrolling no longer works
https://bugs.webkit.org/show_bug.cgi?id=131960
<rdar://problem/16142100>
Unreviewed build fix.
- page/EventHandler.cpp:
(WebCore::EventHandler::handleWheelEvent): The test to clear the
last scrolled node was too broad. It should only follow this path
if both the phase and momentum phase are 'None'.
Tools: Unreviewed build fix.
- DumpRenderTree/mac/EventSendingController.mm:
(-[EventSendingController mouseScrollByX:andY:withWheel:andMomentumPhases:]):
Correct a typo that prevented tests from working properly.
- 9:26 PM Changeset in webkit [167649] by
-
- 2 edits in trunk/Source/JavaScriptCore
DFG::clobberize() should honestly admit that profiler and debugger nodes are effectful
https://bugs.webkit.org/show_bug.cgi?id=131971
<rdar://problem/16676511>
Reviewed by Mark Lam.
- dfg/DFGClobberize.h:
(JSC::DFG::clobberize):
- 8:59 PM Changeset in webkit [167648] by
-
- 3 edits2 adds in trunk
[Mac] After a horizontal overflow scroll with a mechanical wheel mouse, vertical scrolling no longer works
https://bugs.webkit.org/show_bug.cgi?id=131960
<rdar://problem/16142100>
Reviewed by Darin Adler.
Source/WebCore:
Test: platform/mac/fast/scrolling/scroll-div-no-latching.html
- page/EventHandler.cpp:
(WebCore::EventHandler::handleWheelEvent): When not latching, always clear m_latchedWheelEventElement and m_previousWheelScrolledElement.
LayoutTests:
- platform/mac/fast/scrolling/scroll-div-no-latching-expected.txt: Added.
- platform/mac/fast/scrolling/scroll-div-no-latching.html: Added.
- 8:26 PM Changeset in webkit [167647] by
-
- 2 edits in trunk/Source/WebCore
WebProfilerAgent::disable is calling the wrong superclass method
https://bugs.webkit.org/show_bug.cgi?id=131968
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2014-04-21
Reviewed by Simon Fraser.
- inspector/WebProfilerAgent.cpp:
(WebCore::WebProfilerAgent::disable):
- 7:58 PM Changeset in webkit [167646] by
-
- 5 edits1 add in trunk/Source/JavaScriptCore
Switch statements that skip the baseline JIT should work
https://bugs.webkit.org/show_bug.cgi?id=131965
Reviewed by Mark Hahnenberg.
- bytecode/JumpTable.h:
(JSC::SimpleJumpTable::ensureCTITable):
- dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::emitSwitchIntJump):
- jit/JITOpcodes.cpp:
(JSC::JIT::emit_op_switch_imm):
(JSC::JIT::emit_op_switch_char):
- jit/JITOpcodes32_64.cpp:
(JSC::JIT::emit_op_switch_imm):
(JSC::JIT::emit_op_switch_char):
- tests/stress/inline-llint-with-switch.js: Added.
(foo):
(bar):
(test):
- 7:16 PM Changeset in webkit [167645] by
-
- 3 edits in trunk/Source/WebKit2
[wk2] Don't send intrinsic content size updates with dirty layout
https://bugs.webkit.org/show_bug.cgi?id=131964
<rdar://problem/14292718>
Reviewed by Simon Fraser.
- WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.h:
- WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::TiledCoreAnimationDrawingArea):
(WebKit::TiledCoreAnimationDrawingArea::mainFrameContentSizeChanged):
(WebKit::TiledCoreAnimationDrawingArea::updateIntrinsicContentSizeIfNeeded):
(WebKit::TiledCoreAnimationDrawingArea::flushLayers):
(WebKit::TiledCoreAnimationDrawingArea::updateGeometry):
(WebKit::TiledCoreAnimationDrawingArea::updateIntrinsicContentSizeTimerFired): Deleted.
Instead of using a timer (which can fire while layout is dirty) to coalesce
intrinsic content size changes, send them only from flushLayers, just after updating layout.
This ensures that we never get intermediate intrinsic content sizes that won't be painted.
- 7:15 PM Changeset in webkit [167644] by
-
- 2 edits in trunk/Source/WebKit2
Remove a unnecessary forward-declared class from TiledCoreAnimationDrawingArea.
Rubber-stamped by Simon Fraser.
- WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.h:
- 6:56 PM Changeset in webkit [167643] by
-
- 3 edits in trunk/Tools
[GTK] Should have the ability to output composited contents during layout testing
https://bugs.webkit.org/show_bug.cgi?id=131864
Reviewed by Philippe Normand.
- WebKitTestRunner/cairo/TestInvocationCairo.cpp:
(WTR::TestInvocation::dumpPixelsAndCompareWithExpected): Add a path that paints the widget contents
to an image surface. We run the main loop a bit to match the implementation from WebKit1 and EFL.
- WebKitTestRunner/gtk/PlatformWebViewGtk.cpp:
(WTR::PlatformWebView::windowSnapshotImage): Use the new UI-process path on GTK+.
- 6:48 PM Changeset in webkit [167642] by
-
- 2 edits in trunk/Source/WebCore
Followup from http://trac.webkit.org/changeset/167624
Reviewed by Darin Adler.
Fixing incorrect use of ? : syntax.
- editing/VisibleUnits.cpp:
(WebCore::characterBeforePosition):
- 6:37 PM Changeset in webkit [167641] by
-
- 6 edits in trunk/Source/JavaScriptCore
Arguments objects shouldn't need a destructor
https://bugs.webkit.org/show_bug.cgi?id=131899
Reviewed by Oliver Hunt.
This patch rids Arguments objects of their destructors. It does this by
switching their backing stores to use CopiedSpace rather than malloc memory.
- dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::emitAllocateArguments): Fix the code emitted for inline
Arguments allocation so that it only emits an extra write for strict mode code rather
than unconditionally.
- heap/CopyToken.h: New CopyTokens for the two different types of Arguments backing stores.
- runtime/Arguments.cpp:
(JSC::Arguments::visitChildren): We need to tell the collector to copy the back stores now.
(JSC::Arguments::copyBackingStore): Do the actual copying of the backing stores.
(JSC::Arguments::deletePropertyByIndex): Update all the accesses to SlowArgumentData and m_registerArray.
(JSC::Arguments::deleteProperty):
(JSC::Arguments::defineOwnProperty):
(JSC::Arguments::allocateRegisterArray):
(JSC::Arguments::tearOff):
(JSC::Arguments::destroy): Deleted. We don't need the destructor any more.
- runtime/Arguments.h:
(JSC::Arguments::registerArraySizeInBytes):
(JSC::Arguments::SlowArgumentData::SlowArgumentData): Switch SlowArgumentData to being allocated
in CopiedSpace. Now the SlowArgumentData and its backing store are a single contiguous CopiedSpace
allocation.
(JSC::Arguments::SlowArgumentData::slowArguments):
(JSC::Arguments::SlowArgumentData::bytecodeToMachineCaptureOffset):
(JSC::Arguments::SlowArgumentData::setBytecodeToMachineCaptureOffset):
(JSC::Arguments::SlowArgumentData::sizeForNumArguments):
(JSC::Arguments::Arguments):
(JSC::Arguments::allocateSlowArguments):
(JSC::Arguments::tryDeleteArgument):
(JSC::Arguments::isDeletedArgument):
(JSC::Arguments::isArgument):
(JSC::Arguments::argument):
(JSC::Arguments::finishCreation):
- runtime/SymbolTable.h:
- 6:34 PM Changeset in webkit [167640] by
-
- 3 edits in trunk/Source/WebCore
Move CoordinatedGraphics related codes out of CMakeLists.txt
https://bugs.webkit.org/show_bug.cgi?id=131931
Reviewed by Gyuyoung Kim.
Move CoordinatedGraphics related files to PlatformEfl.cmake because they are not common files.
- CMakeLists.txt:
- PlatformEfl.cmake:
- 5:57 PM Changeset in webkit [167639] by
-
- 2 edits in trunk/Source/WebCore
[MSE][Mac] AVSampleBufferDisplayLayer and AudioRenderer is overretained.
https://bugs.webkit.org/show_bug.cgi?id=131963
Reviewed by Jon Honeycutt.
Adopt the results of -[[class alloc] init] rather than retaining them.
- platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:
(WebCore::SourceBufferPrivateAVFObjC::hasVideo):
(WebCore::SourceBufferPrivateAVFObjC::trackDidChangeEnabled):
- 5:23 PM Changeset in webkit [167638] by
-
- 2 edits in trunk/Source/WebKit2
Don't try to use a null connection()->xpcConnection()
https://bugs.webkit.org/show_bug.cgi?id=131961
Reviewed by Dan Bernstein.
If running with -WebKit2UseXPCServiceForWebProcess NO, connection()->xpcConnection()
is null and xpc_connection_get_pid() would crash, so null-check the xpcConnection.
- UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::didFinishLaunching):
- 5:15 PM Changeset in webkit [167637] by
-
- 3 edits2 adds in trunk
REGRESSION (r166784): Gradient at background of iCloud login page doesn’t go all the way to the bottom
https://bugs.webkit.org/show_bug.cgi?id=131924
Computing tile dimension for contain/cover requires higher precision than what LayoutUnit has. Switching to floats.
Reviewed by Simon Fraser.
Source/WebCore:
Test: fast/backgrounds/hidpi-background-image-contain-cover-scale-needs-more-precision.html
- rendering/RenderBoxModelObject.cpp:
(WebCore::RenderBoxModelObject::calculateFillTileSize):
LayoutTests:
- fast/backgrounds/hidpi-background-image-contain-cover-scale-needs-more-precision-expected.html: Added.
- fast/backgrounds/hidpi-background-image-contain-cover-scale-needs-more-precision.html: Added.
- 5:14 PM Changeset in webkit [167636] by
-
- 2 edits in trunk/Source/WebCore
Build fix for Mountain Lion.
- platform/mac/SerializedPlatformRepresentationMac.mm:
(WebCore::SerializedPlatformRepresentationMac::deserialize):
- 5:10 PM Changeset in webkit [167635] by
-
- 15 edits in trunk/Source
Source/WebCore: WebCore part of <rdar://problem/16631323> Page::setDefersLoading doesn’t do anything when using the Network process
https://bugs.webkit.org/show_bug.cgi?id=131939
Reviewed by David Kilzer.
- WebCore.exp.in: Exported WebCore::ResourceHandle::setDefersLoading.
- loader/ResourceLoadScheduler.cpp:
(WebCore::ResourceLoadScheduler::setDefersLoading): Added a no-op implementation.
- loader/ResourceLoadScheduler.h: Declared new virtual member function setDefersLoading.
- loader/ResourceLoader.cpp:
(WebCore::ResourceLoader::setDefersLoading): Added a call to the above function.
Source/WebKit2: WebKit part of <rdar://problem/16631323> Page::setDefersLoading doesn’t do anything when using the Network process
https://bugs.webkit.org/show_bug.cgi?id=131939
Reviewed by David Kilzer.
- NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::setDefersLoading): Added. Forwards the message to
the NetworkResourceLoader.
- NetworkProcess/NetworkConnectionToWebProcess.h:
- NetworkProcess/NetworkConnectionToWebProcess.messages.in: Added SetDefersLoading message.
- NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::NetworkResourceLoader): Initialize new member variable
m_defersLoading from the new field in the loader parameters.
(WebKit::NetworkResourceLoader::start): Added. Like WebCore::ResourceLoader, if loading is
deferred, keep the request in m_deferredRequest and bail out.
(WebKit::NetworkResourceLoader::setDefersLoading): Added. Similar to
WebCore::ResourceLoader::setDefersLoading.
- NetworkProcess/NetworkResourceLoader.h:
(WebKit::NetworkResourceLoader::defersLoading):
- Shared/Network/NetworkResourceLoadParameters.cpp:
(WebKit::NetworkResourceLoadParameters::NetworkResourceLoadParameters): Initialize new field
defersLoading to false.
(WebKit::NetworkResourceLoadParameters::encode): Encode new field.
(WebKit::NetworkResourceLoadParameters::decode): Decode new field.
- Shared/Network/NetworkResourceLoadParameters.h: Added defersLoading field.
- WebProcess/Network/WebResourceLoadScheduler.cpp:
(WebKit::WebResourceLoadScheduler::scheduleLoad): Populate defersLoading field in the load
parameters.
(WebKit::WebResourceLoadScheduler::setDefersLoading): Send a message to the Network process.
- WebProcess/Network/WebResourceLoadScheduler.h:
- 5:05 PM Changeset in webkit [167634] by
-
- 2 edits in trunk/Source/WebCore
[Mac] Difficulty gesture scrolling vertically with trackpad after scrolling horizontally
https://bugs.webkit.org/show_bug.cgi?id=131959
<rdar://problem/16654523>
Reviewed by Simon Fraser.
- page/mac/EventHandlerMac.mm:
(WebCore::deltaIsPredominantlyVertical): Added.
(WebCore::scrolledToEdgeInDominantDirection): Only consider current mouse wheel event. We don't care about
overall history when deciding if we are bumping against the edge of a scrollable region. Short-circuit if
the element style indicates that overflow is hidden, since this means there is no scroll possible in that
direction.
(WebCore::EventHandler::platformPrepareForWheelEvents): Update for new signature.
- 4:51 PM Changeset in webkit [167633] by
-
- 3 edits4 adds in trunk/Source/WebKit2
Errors passed to delegate methods should have recovery attempters
https://bugs.webkit.org/show_bug.cgi?id=131957
<rdar://problem/16642539>
Reviewed by Tim Horton.
- UIProcess/API/Cocoa/_WKErrorRecoveryAttempting.h: Added.
- UIProcess/API/Cocoa/_WKErrorRecoveryAttempting.mm: Added.
- UIProcess/Cocoa/NavigationState.mm:
(WebKit::createErrorWithRecoveryAttempter):
(WebKit::NavigationState::LoaderClient::didFailProvisionalLoadWithErrorForFrame):
(WebKit::NavigationState::LoaderClient::didFailLoadWithErrorForFrame):
- UIProcess/Cocoa/WKReloadFrameErrorRecoveryAttempter.h: Added.
- UIProcess/Cocoa/WKReloadFrameErrorRecoveryAttempter.mm: Added.
(-[WKReloadFrameErrorRecoveryAttempter initWithWebView:frameHandle:urlString:]):
(-[WKReloadFrameErrorRecoveryAttempter attemptRecovery]):
- WebKit2.xcodeproj/project.pbxproj:
- 4:18 PM Changeset in webkit [167632] by
-
- 45 edits17 adds in trunk
[Mac] implement WebKitDataCue
https://bugs.webkit.org/show_bug.cgi?id=131799
Reviewed by Dean Jackson.
Source/JavaScriptCore:
- Configurations/FeatureDefines.xcconfig: Define ENABLE_DATACUE_VALUE.
Source/WebCore:
Tests: http/tests/media/track-in-band-hls-metadata.html
media/track/track-datacue-value.html
- Configurations/FeatureDefines.xcconfig: Define ENABLE_DATACUE_VALUE.
- DerivedSources.make: Add ENABLE_DATACUE_VALUE to HTML_FLAGS when appropriate.
- CMakeLists.txt: Add JSDataCueCustom.cpp.
- bindings/js/JSBindingsAllInOne.cpp:
- WebCore.xcodeproj/project.pbxproj: Add new files.
- bindings/js/JSDataCueCustom.cpp: Added.
(WebCore::JSDataCue::value):
(WebCore::JSDataCue::setValue):
(WebCore::JSDataCueConstructor::constructJSDataCue): Custom constructor.
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::updateActiveTextTrackCues): Drive-by fixes: don't schedule timeupdate
events when paused, don't call sort on an Vector that can't be sorted.
- html/track/DataCue.cpp:
(WebCore::DataCue::DataCue): Initialize m_type.
(WebCore::DataCue::~DataCue): Unprotect the JSValue if necessary.
(WebCore::DataCue::data): Ask the platform value for data if non-null.
(WebCore::DataCue::setData): Clear m_platformValue and m_value.
(WebCore::DataCue::isEqual): New.
(WebCore::DataCue::value): Return a JSValue from the platform value, or the value passed
to the constructor/set by script.
(WebCore::DataCue::setValue): Set m_value.
- html/track/DataCue.h:
- html/track/DataCue.idl:
- html/track/InbandDataTextTrack.cpp:
(WebCore::InbandDataTextTrack::addDataCue): Don't add the same cue more than once.
(WebCore::InbandDataTextTrack::updateDataCue): Update a cue's duration.
(WebCore::InbandDataTextTrack::removeDataCue): Remove an incomplete cue.
(WebCore::InbandDataTextTrack::removeCue): Remove a cue from the incomplete cue map if necessary.
- html/track/InbandDataTextTrack.h:
- html/track/InbandGenericTextTrack.cpp:
(WebCore::InbandGenericTextTrack::addGenericCue): CueMatchRules is now in TextTrackCue instead
of VTTCue.
- html/track/InbandTextTrack.h:
- html/track/InbandWebVTTTextTrack.cpp:
(WebCore::InbandWebVTTTextTrack::newCuesParsed): Ditto.
- html/track/TextTrack.cpp:
(WebCore::TextTrack::hasCue): Ditto.
- html/track/TextTrack.h:
- html/track/TextTrackCue.cpp:
(WebCore::TextTrackCue::isEqual): New, test base class equality.
- html/track/TextTrackCue.h:
- html/track/TextTrackCueGeneric.cpp:
(WebCore::TextTrackCueGeneric::isEqual): Call TextTrackCue::isEqual first.
- html/track/TextTrackCueGeneric.h:
- html/track/VTTCue.cpp:
(WebCore::VTTCue::isEqual): Call TextTrackCue::isEqual first.
- html/track/VTTCue.h:
- platform/SerializedPlatformRepresentation.h: Added.
(WebCore::SerializedPlatformRepresentation::~SerializedPlatformRepresentation):
(WebCore::SerializedPlatformRepresentation::SerializedPlatformRepresentation):
- platform/graphics/InbandTextTrackPrivateClient.h: Add methods for DataCue with SerializedPlatformRepresentation.
- platform/graphics/avfoundation/InbandMetadataTextTrackPrivateAVF.cpp: Added.
(WebCore::InbandMetadataTextTrackPrivateAVF::create):
(WebCore::InbandMetadataTextTrackPrivateAVF::InbandMetadataTextTrackPrivateAVF):
(WebCore::InbandMetadataTextTrackPrivateAVF::~InbandMetadataTextTrackPrivateAVF):
(WebCore::InbandMetadataTextTrackPrivateAVF::addDataCue):
(WebCore::InbandMetadataTextTrackPrivateAVF::updatePendingCueEndTimes):
(WebCore::InbandMetadataTextTrackPrivateAVF::flushPartialCues):
- platform/graphics/avfoundation/InbandMetadataTextTrackPrivateAVF.h: Added.
- platform/graphics/avfoundation/MediaPlayerPrivateAVFoundation.cpp:
(WebCore::MediaPlayerPrivateAVFoundation::seekWithTolerance): currentTrack -> currentTextTrack.
(WebCore::MediaPlayerPrivateAVFoundation::seekCompleted): Ditto.
(WebCore::MediaPlayerPrivateAVFoundation::configureInbandTracks): Ditto.
- platform/graphics/avfoundation/MediaPlayerPrivateAVFoundation.h:
(WebCore::MediaPlayerPrivateAVFoundation::setCurrentTextTrack):
(WebCore::MediaPlayerPrivateAVFoundation::setCurrentTrack): Deleted.
Renamed currentTrack and setCurrentTrack to currentTextTrack and setCurrentTextTrack.
- platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
(WebCore::AVFWrapper::currentTextTrack):
(WebCore::MediaPlayerPrivateAVFoundationCF::setCurrentTextTrack):
(WebCore::MediaPlayerPrivateAVFoundationCF::currentTextTrack):
(WebCore::AVFWrapper::setCurrentTextTrack):
(WebCore::AVFWrapper::AVFWrapper):
(WebCore::AVFWrapper::processCue):
(WebCore::AVFWrapper::currentTrack): Deleted.
(WebCore::MediaPlayerPrivateAVFoundationCF::setCurrentTrack): Deleted.
(WebCore::MediaPlayerPrivateAVFoundationCF::currentTrack): Deleted.
(WebCore::AVFWrapper::setCurrentTrack): Deleted.
- platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.h:
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.h:
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::MediaPlayerPrivateAVFoundationObjC):
(WebCore::MediaPlayerPrivateAVFoundationObjC::seekToTime): currentTrack -> currentTextTrack
(WebCore::MediaPlayerPrivateAVFoundationObjC::tracksChanged): currentTrack -> currentTextTrack.
(WebCore::MediaPlayerPrivateAVFoundationObjC::processMetadataTrack): New.
(WebCore::MediaPlayerPrivateAVFoundationObjC::processCue): m_currentTrack -> m_currentTextTrack.
(WebCore::MediaPlayerPrivateAVFoundationObjC::flushCues): Ditto.
(WebCore::MediaPlayerPrivateAVFoundationObjC::setCurrentTextTrack): Renamed from setCurrentTextTrack.
(WebCore::metadataType): Map an AVFoundation metadata key space to a metadata cue type.
(WebCore::MediaPlayerPrivateAVFoundationObjC::metadataDidArrive): Process new metadata.
(-[WebCoreAVFMovieObserver observeValueForKeyPath:ofObject:change:context:]):
(WebCore::MediaPlayerPrivateAVFoundationObjC::setCurrentTrack): Deleted.
Create a JSValue representation from an AVMetadataItem.
- platform/mac/SerializedPlatformRepresentationMac.h: Added.
(WebCore::SerializedPlatformRepresentationMac::platformType):
(WebCore::SerializedPlatformRepresentationMac::nativeValue):
- platform/mac/SerializedPlatformRepresentationMac.mm: Added.
(WebCore::SerializedPlatformRepresentationMac::SerializedPlatformRepresentationMac):
(WebCore::SerializedPlatformRepresentationMac::~SerializedPlatformRepresentationMac):
(WebCore::SerializedPlatformRepresentationMac::create):
(WebCore::SerializedPlatformRepresentationMac::data):
(WebCore::SerializedPlatformRepresentationMac::deserialize):
(WebCore::SerializedPlatformRepresentationMac::isEqual):
(WebCore::toSerializedPlatformRepresentationMac):
(WebCore::jsValueWithValueInContext):
(WebCore::jsValueWithDataInContext):
(WebCore::jsValueWithArrayInContext):
(WebCore::jsValueWithDictionaryInContext):
(WebCore::jsValueWithAVMetadataItemInContext):
Source/WebKit/mac:
- Configurations/FeatureDefines.xcconfig: Define ENABLE_DATACUE_VALUE.
Source/WebKit2:
- Configurations/FeatureDefines.xcconfig: Define ENABLE_DATACUE_VALUE.
Source/WTF:
- wtf/FeatureDefines.h: Define ENABLE_DATACUE_VALUE.
LayoutTests:
- http/tests/media/resources/hls: Added.
- http/tests/media/resources/hls/metadata: Added.
- http/tests/media/resources/hls/metadata/fileSequence0.ts: Added.
- http/tests/media/resources/hls/metadata/fileSequence1.ts: Added.
- http/tests/media/resources/hls/metadata/fileSequence2.ts: Added.
- http/tests/media/resources/hls/metadata/fileSequence3.ts: Added.
- http/tests/media/resources/hls/metadata/prog_index.m3u8: Added.
- http/tests/media/track-in-band-hls-metadata-expected.txt: Added.
- http/tests/media/track-in-band-hls-metadata.html: Added.
- media/track/track-datacue-value-expected.txt: Added.
- media/track/track-datacue-value.html: Added.
- platform/efl/TestExpectations: Skip the new tests.
- platform/gtk/TestExpectations: Ditto.
- platform/mac/js/dom/global-constructors-attributes-expected.txt: Update.
- platform/mac-mountainlion/js/dom/global-constructors-attributes-expected.txt: Update.
- platform/mac/TestExpectations: Skip DataCue test on all Mac versions. Skip HLS test on
Mountain Lion.
- platform/win/TestExpectations: Skip the new tests.
- 3:25 PM Changeset in webkit [167631] by
-
- 33 edits in trunk
Add Element.matches, the standard name for webkitMatchesSelector
https://bugs.webkit.org/show_bug.cgi?id=131922
Reviewed by Andreas Kling.
Source/WebCore:
This patch just renames webkitMatchesSelector() to matches().
- bindings/scripts/CodeGeneratorObjC.pm:
(GenerateImplementation):
The Objective-C binding generator only supported ImplementedAs for getters and setters.
This adds support for the option so that Element.matches can be generated for Obj-C.
- dom/Element.cpp:
(WebCore::Element::matches):
(WebCore::Element::webkitMatchesSelector): Deleted.
- dom/Element.h:
- dom/Element.idl:
- inspector/InspectorCSSAgent.cpp:
(WebCore::InspectorCSSAgent::buildArrayForMatchedRuleList):
LayoutTests:
Update the tests to verify both Element.matches and Element.webkitMatchesSelector
- fast/dom/SelectorAPI/attrname-case-insensitive-expected.txt:
- fast/dom/SelectorAPI/attrname-case-insensitive.html:
- fast/dom/SelectorAPI/attrname-case-sensitive-expected.txt:
- fast/dom/SelectorAPI/attrname-case-sensitive.xhtml:
- fast/dom/SelectorAPI/caseID-almost-strict-expected.txt:
- fast/dom/SelectorAPI/caseID-almost-strict.html:
- fast/dom/SelectorAPI/caseID-expected.txt:
- fast/dom/SelectorAPI/caseID-strict-expected.txt:
- fast/dom/SelectorAPI/caseID-strict.html:
- fast/dom/SelectorAPI/caseID.html:
- fast/dom/SelectorAPI/caseTag-expected.txt:
- fast/dom/SelectorAPI/caseTag.html:
- fast/dom/SelectorAPI/caseTagX-expected.txt:
- fast/dom/SelectorAPI/caseTagX.xhtml:
- fast/dom/SelectorAPI/detached-element-expected.txt:
- fast/dom/SelectorAPI/not-supported-namespace-in-selector-expected.txt:
- fast/dom/SelectorAPI/not-supported-namespace-in-selector.html:
- fast/dom/SelectorAPI/script-tests/detached-element.js:
- fast/dom/SelectorAPI/script-tests/undefined-null-stringify.js:
- fast/dom/SelectorAPI/script-tests/viewless-document.js:
- fast/dom/SelectorAPI/undefined-null-stringify-expected.txt:
- fast/dom/SelectorAPI/viewless-document-expected.txt:
- fast/forms/radio/radio-live-validation-style-expected.txt:
- fast/forms/radio/radio-live-validation-style.html:
- fast/harness/results.html:
- 3:22 PM Changeset in webkit [167630] by
-
- 7 edits in trunk/Source
topContentInset does not play well with fullscreen elements
https://bugs.webkit.org/show_bug.cgi?id=131955
-and corresponding-
<rdar://problem/16651925>
Reviewed by Sam Weinig.
Source/WebCore:
Calling setNeedsLayout() is not sufficient when the topContentInset has changed
dynamically. We need to perform the layout right away and update the scrollbars.
This works completely when the inset changes due to entering/exiting fullscreen,
but I left a FIXME behind because it is possible to change the inset dynamically
just through the WK2 API and end up in a situation where everything looks right
except for the scrollbars.
- page/FrameView.cpp:
(WebCore::FrameView::topContentInsetDidChange):
- page/FrameView.h:
- page/Page.cpp:
(WebCore::Page::setTopContentInset):
Source/WebKit2:
If you take an element (such as a <video>) fullscreen when you have a
topContentInset set, there will be an inset-sized gap at the top of the fullscreen
video.
Save and restore the topContentInset whenever the scroll position is saved and
restored.
- WebProcess/FullScreen/WebFullScreenManager.cpp:
(WebKit::WebFullScreenManager::WebFullScreenManager):
(WebKit::WebFullScreenManager::saveScrollPosition):
(WebKit::WebFullScreenManager::restoreScrollPosition):
- WebProcess/FullScreen/WebFullScreenManager.h:
- 2:49 PM Changeset in webkit [167629] by
-
- 5 edits in tags/Safari-538.30.2/Source
Versioning.
- 2:45 PM Changeset in webkit [167628] by
-
- 3 edits1 add in trunk
Empty RenderInline objects should not be line break objects.
https://bugs.webkit.org/show_bug.cgi?id=131861
<rdar://problem/15663617>
Reviewed by David Hyatt.
.:
- ManualTests/empty-inline-as-line-break-position.html: Added.
This issue wouldn't reproduce in WebKitTestRunner.
Source/WebCore:
- rendering/line/BreakingContextInlineHeaders.h:
(WebCore::BreakingContext::canBreakAtThisPosition):
Don't break at this position if this is an empty RenderInline.
- 2:44 PM Changeset in webkit [167627] by
-
- 2 edits in trunk/Tools
Attempt to fix the tests.
- TestWebKitAPI/Tests/WebKit2Cocoa/Navigation.mm:
(-[DidFailProvisionalNavigationDelegate webView:decidePolicyForNavigationAction:decisionHandler:]):
Add policy handler to override the new default of denying loads that NSURLConnection cannot handle.
- 2:36 PM Changeset in webkit [167626] by
-
- 1 copy in tags/Safari-538.30.2
New tag.
- 2:33 PM Changeset in webkit [167625] by
-
- 2 edits in trunk/Source/WebKit2
Don't use ProcessAssertion on simulator
https://bugs.webkit.org/show_bug.cgi?id=131950
Rubber stamped by Enrica Casucci
- UIProcess/Launcher/mac/ProcessLauncherMac.mm:
(WebKit::connectToService):
- UIProcess/ios/ProcessAssertion.h:
- 2:14 PM Changeset in webkit [167624] by
-
- 10 edits in trunk/Source
[iOS WebKit2] support replacements for misspelled words.
https://bugs.webkit.org/show_bug.cgi?id=131827
<rdar://problem/16319657>
Reviewed by Darin Adler.
Source/WebCore:
Adds utility function to return the character before the current selection
and converts the nbsp to the space character.
- WebCore.exp.in:
- editing/VisibleUnits.cpp:
(WebCore::characterBeforePosition):
- editing/VisibleUnits.h:
Source/WebKit2:
This is the second a final piece to support replacements.
It adds some entrypoints used by the keyboard code to perform
replacement when reaching the edge of a word using backspace.
I've modified the behavior of replaceSelectedText to work both
with caret or range selections.
- Shared/EditorState.cpp:
(WebKit::EditorState::encode):
(WebKit::EditorState::decode):
- Shared/EditorState.h:
(WebKit::EditorState::EditorState):
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView isReplaceAllowed]):
(-[WKContentView _characterBeforeCaretSelection]):
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::editorState):
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::replaceSelectedText):
- 2:06 PM Changeset in webkit [167623] by
-
- 2 edits in trunk/Source/WebCore
[iOS] enable plug-in replacement
https://bugs.webkit.org/show_bug.cgi?id=131947
Reviewed by Jer Noble.
- bindings/generic/RuntimeEnabledFeatures.cpp:
(WebCore::RuntimeEnabledFeatures::RuntimeEnabledFeatures): initialize m_isPluginReplacementEnabled
to true on iOS.
- 2:00 PM Changeset in webkit [167622] by
-
- 2 edits in trunk/Source/WebKit2
Don't use ProcessAssertion on simulator
https://bugs.webkit.org/show_bug.cgi?id=131950
Rubber stamped by Benjamin Poulain
- UIProcess/ios/ProcessAssertion.mm:
(WebKit::ProcessAssertion::ProcessAssertion):
(WebKit::ProcessAssertion::setState):
- 1:44 PM Changeset in webkit [167621] by
-
- 2 edits in trunk/Source/WebKit2
Another build fix.
- Shared/EntryPointUtilities/mac/XPCService/XPCServiceEntryPoint.mm:
(WebKit::XPCServiceInitializerDelegate::checkEntitlements):
- 1:44 PM Changeset in webkit [167620] by
-
- 12 edits4 deletes in trunk
[CSS Blending] Remove support for non-separable blend modes from mix-blend-mode
https://bugs.webkit.org/show_bug.cgi?id=131824
Reviewed by Dean Jackson.
Source/WebCore:
Removed support for non-separable blend modes from the CSS parser.
Covered by existing tests
- css/CSSParser.cpp:
(WebCore::isValidKeywordPropertyAndValue):
- platform/graphics/ca/mac/PlatformCAFiltersMac.mm:
(PlatformCAFilters::setBlendingFiltersOnLayer):
LayoutTests:
Removed or updated tests that used non-separable blend modes in combination
with mix-blend-mode.
- css3/compositing/blend-mode-property-expected.txt:
- css3/compositing/blend-mode-property-parsing-expected.txt:
- css3/compositing/blend-mode-simple-composited.html:
- css3/compositing/blend-mode-simple.html:
- css3/compositing/script-tests/blend-mode-property-parsing.js:
- css3/compositing/script-tests/blend-mode-property.js:
- css3/compositing/svg-blend-color.html: Removed.
- css3/compositing/svg-blend-hue.html: Removed.
- css3/compositing/svg-blend-luminosity.html: Removed.
- css3/compositing/svg-blend-saturation.html: Removed.
- platform/mac/css3/compositing/blend-mode-simple-composited-expected.txt:
- platform/mac/css3/compositing/blend-mode-simple-expected.txt:
- 1:39 PM Changeset in webkit [167619] by
-
- 2 edits in trunk/LayoutTests
[New Multicolumn] vertical-rl/unsplittable-inline-block misrenders the column rule.
https://bugs.webkit.org/show_bug.cgi?id=131818
Reviewed by Anders Carlsson.
Patch the test to be column-fill:auto, since the other versions
of the test all had that specified already.
- fast/multicol/vertical-rl/unsplittable-inline-block.html:
- 1:35 PM Changeset in webkit [167618] by
-
- 2 edits in trunk/Source/WebKit2
Build fix.
- Shared/EntryPointUtilities/mac/XPCService/XPCServiceEntryPoint.h:
(WebKit::XPCServiceInitializer):
- 1:20 PM Changeset in webkit [167617] by
-
- 5 edits2 adds in trunk
[New Multicolumn] Column set drawing under horizontal scrollbar.
https://bugs.webkit.org/show_bug.cgi?id=131812.
Reviewed by Sam Weinig.
Source/WebCore:
Added fast/multicol/newmulticol/compare-with-old-impl/LeftToRight-tb.html.
- rendering/RenderBlockFlow.cpp:
(WebCore::RenderBlockFlow::layoutBlock):
(WebCore::RenderBlockFlow::checkForPaginationLogicalHeightChange):
- rendering/RenderBlockFlow.h:
Pass in relayoutChildren as a reference boolean to checkForPaginationLogicalHeightChange.
The new columns need to be able to set it to true if the amount of available column
height ever changes, so that the RenderFlowThread and the sets get a layout call.
- rendering/style/RenderStyle.cpp:
(WebCore::RenderStyle::setColumnStylesFromPaginationMode):
Make sure that the pagination mode is set up with an auto column-fill specified. You
don't ever want to balance when you set a pagination mode on the RenderView.
LayoutTests:
- fast/multicol/newmulticol/compare-with-old-impl/LeftToRight-tb-expected.html: Added.
- fast/multicol/newmulticol/compare-with-old-impl/LeftToRight-tb.html: Added.
- 1:16 PM Changeset in webkit [167616] by
-
- 13 edits in trunk/Source
[iOS WebKit2] Implement CSS viewport units
https://bugs.webkit.org/show_bug.cgi?id=131863
<rdar://problem/16279088>
Reviewed by Darin Adler.
- UIProcess/API/Cocoa/WKWebView.mm:
(setMinimumLayoutSizeForMinimalUI):
(-[WKWebView _minimumLayoutSizeOverrideForMinimalUI]):
(-[WKWebView _setMinimumLayoutSizeOverrideForMinimalUI:]):
Add _setMinimumLayoutSizeOverrideForMinimalUI:, which pushes said size
down from the client to WebPageIOS, below.
- UIProcess/API/Cocoa/WKWebViewPrivate.h:
- UIProcess/WebPageProxy.h:
- UIProcess/ios/WebPageProxyIOS.mm:
(WebKit::WebPageProxy::setMinimumLayoutSizeForMinimalUI):
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/WebPage.messages.in:
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::setMinimumLayoutSizeForMinimalUI):
(WebKit::WebPage::viewportConfigurationChanged):
Update FrameView's viewport size based on the minimum layout size for minimal UI (if we have it)
or the normal minimum layout size, scaled to the page's initial scale.
- WebCore.exp.in:
- page/FrameView.cpp:
(WebCore::FrameView::setViewportSize):
- page/FrameView.h:
(WebCore::FrameView::viewportSize):
Add setViewportSize() and viewportSize().
If the viewport size changes, we have to do style recalc.
If we have an overridden viewport size, use it.
- rendering/RenderView.cpp:
(WebCore::RenderView::viewportSize):
- rendering/RenderView.h:
Get the size from FrameView.
- 1:14 PM Changeset in webkit [167615] by
-
- 2 edits in trunk/Source/WebKit2
[iOS][WK2] Fix the dynamic viewport update when the viewport's max-scale conflict with the transition
https://bugs.webkit.org/show_bug.cgi?id=131884
Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-04-21
Reviewed by Darin Adler.
<rdar://problem/16667291>
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _beginAnimatedResizeWithUpdates:]):
It is fairly common to have a viewport configuration with min-scale=1 and max-scale=1.
Since we were only changing the minimumScale for the transition, the target scale was affected
by max-scale. This was breaking the transition from portrait to landscale.
- 1:10 PM Changeset in webkit [167614] by
-
- 4 edits in trunk/Source/WebKit2
[iOS][WK2] Make dynamic viewport size update content aware
https://bugs.webkit.org/show_bug.cgi?id=131874
Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-04-21
Reviewed by Tim Horton.
When possible, adjust the scroll position based on the content on dynamic viewport resize.
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::WebPage):
(WebKit::WebPage::scalePage):
(WebKit::WebPage::pageDidScroll):
- WebProcess/WebPage/WebPage.h:
The scrolling heuristic are not invertible, especially the content heuristic. To have the right
behavior when doing resize without changing the page, we save the old scroll position and restore
it when the content size and scale is restored.
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::dynamicViewportSizeUpdate):
On dynamic update, start by finding what node is at the center of the screen. After the layout, put that
node back in the center.
(WebKit::WebPage::updateVisibleContentRects):
- 12:53 PM Changeset in webkit [167613] by
-
- 5 edits in trunk
Source/JavaScriptCore: Unreviewed test gardening, run the repeat-out-of-bounds tests again.
- tests/stress/float32-repeat-out-of-bounds.js:
- tests/stress/int8-repeat-out-of-bounds.js:
Tools: Unreviewed test gardening, make no-eager runs run everything except eager (as opposed to
being the same as a "quick run", which is meant for genuinely slow tests).
- Scripts/run-jsc-stress-tests:
- 12:51 PM Changeset in webkit [167612] by
-
- 6 edits2 adds in trunk/Source/JavaScriptCore
OSR exit should know about Int52 and Double constants
https://bugs.webkit.org/show_bug.cgi?id=131945
Reviewed by Oliver Hunt.
The DFG OSR exit machinery's ignorance would lead to some constants becoming
jsUndefined() after OSR exit.
The FTL OSR exit machinery's ignorance just meant that we would sometimes use a
stackmap constant rather than baking the constant into the OSRExit data structure.
So, not a big deal, but worth fixing.
Also added some helpful hacks to jsc.cpp for testing such OSR exit pathologies.
- dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::handleIntrinsic):
- dfg/DFGMinifiedNode.h:
(JSC::DFG::belongsInMinifiedGraph):
(JSC::DFG::MinifiedNode::hasConstantNumber):
- ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::tryToSetConstantExitArgument):
- jsc.cpp:
(GlobalObject::finishCreation):
(functionOtherFalse):
(functionUndefined):
- runtime/Intrinsic.h:
- tests/stress/fold-to-double-constant-then-exit.js: Added.
(foo):
- tests/stress/fold-to-int52-constant-then-exit.js: Added.
(foo):
- 12:51 PM Changeset in webkit [167611] by
-
- 2 edits in trunk/Source/WebKit2
Need default WKNavigationDelegate behavior
https://bugs.webkit.org/show_bug.cgi?id=131944
<rdar://problem/15949822>
Reviewed by Dan Bernstein.
Port the WebKit1 default delegate behavior over.
- UIProcess/Cocoa/NavigationState.mm:
(WebKit::NavigationState::PolicyClient::decidePolicyForNavigationAction):
(WebKit::NavigationState::PolicyClient::decidePolicyForResponse):
- 12:49 PM Changeset in webkit [167610] by
-
- 5 edits in trunk/Source/WebKit2
[Cocoa] Expose WebPage::setDefersLoading
https://bugs.webkit.org/show_bug.cgi?id=131943
Reviewed by Anders Carlsson, but he doesn’t like this.
- WebProcess/InjectedBundle/API/mac/WKWebProcessPlugInBrowserContextController.mm:
(-[WKWebProcessPlugInBrowserContextController _defersLoading]): Added this accessor.
(-[WKWebProcessPlugInBrowserContextController _setDefersLoading:]): Ditto.
- WebProcess/InjectedBundle/API/mac/WKWebProcessPlugInBrowserContextControllerPrivate.h:
Declared new property _defersLoading.
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::defersLoading): Added this getter.
- WebProcess/WebPage/WebPage.h:
- 12:48 PM Changeset in webkit [167609] by
-
- 9 edits in trunk/Source
Remove unused WebKitSystemInterface functions
https://bugs.webkit.org/show_bug.cgi?id=131938
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2014-04-21
Reviewed by Alexey Proskuryakov.
Source/WebCore:
- WebCore.exp.in:
- platform/ios/WebCoreSystemInterfaceIOS.mm:
- platform/mac/WebCoreSystemInterface.h:
- platform/mac/WebCoreSystemInterface.mm:
Source/WebKit/mac:
- WebCoreSupport/WebSystemInterface.mm:
(InitWebCoreSystemInterface):
Source/WebKit2:
- WebProcess/WebCoreSupport/mac/WebSystemInterface.mm:
(InitWebCoreSystemInterface):
- 12:41 PM Changeset in webkit [167608] by
-
- 2 edits in trunk/Source/JavaScriptCore
Provide feedback when we encounter an unrecognied node in the FTL backend.
Rubber stamped by Alexey Proskuryakov.
- ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::compileNode):
- 12:37 PM Changeset in webkit [167607] by
-
- 4 edits in trunk/LayoutTests
Update results of fast/images/image-controls-basic.html
- platform/mac/TestExpectations:
- platform/mac/fast/images/image-controls-basic-expected.png:
- platform/mac/fast/images/image-controls-basic-expected.txt:
- 12:17 PM Changeset in webkit [167606] by
-
- 1 edit2 adds in trunk/LayoutTests
[New Multicolumn] Add test case for shrink-to-column-height now that
it passes in the new impl.
https://bugs.webkit.org/show_bug.cgi?id=131814
Reviewed by Dean Jackson.
- fast/multicol/newmulticol/compare-with-old-impl/shrink-to-column-height-for-pagination-expected.html: Added.
- fast/multicol/newmulticol/compare-with-old-impl/shrink-to-column-height-for-pagination.html: Added.
- 12:17 PM Changeset in webkit [167605] by
-
- 6 edits in trunk/Source
Move the JSString cache from DOMWrapperWorld to VM.
<https://webkit.org/b/131940>
Source/JavaScriptCore:
Reviewed by Geoff Garen.
- runtime/VM.h:
Source/WebCore:
Since there's no need for JSStrings to be world-specific, this patch
moves the string cache to JSC::VM. This makes jsStringWithCache()
a lot faster since it no longer has to jump through twenty-eleven
hoops to find the DOMWrapperWorld.
Reviewed by Geoff Garen.
- bindings/js/DOMWrapperWorld.cpp:
(WebCore::DOMWrapperWorld::clearWrappers):
- bindings/js/DOMWrapperWorld.h:
- bindings/js/JSDOMBinding.cpp:
(WebCore::jsStringWithCache):
- 12:15 PM Changeset in webkit [167604] by
-
- 13 edits1 add in trunk/LayoutTests
[GTK] Unreviewed GTK gardening
Update gtk/TestExpectations and rebaseline some gtk test results.
Patch by Carlos Alberto Lopez Perez <clopez@igalia.com> on 2014-04-21
- platform/gtk/TestExpectations: Mark fast/repaint/hidpi-wrong-repaint-rect-when-parent-has-noncompositing-transform.html
as failure and fast/dom/Window/mozilla-focus-blur.html as flaky.
- platform/gtk/fast/multicol/newmulticol/client-rects-expected.txt: Rebaseline after r167335.
- platform/gtk/fast/repaint/reflection-redraw-expected.txt: Rebaseline after r167568.
- platform/gtk/fast/table/dynamic-caption-add-remove-before-child-expected.txt: Rebaseline after r167568.
- platform/gtk/fast/table/multiple-captions-display-expected.txt: Rebaseline after r167568.
- platform/gtk/tables/mozilla/marvin/body_col-expected.txt: Rebaseline after r167568.
- platform/gtk/tables/mozilla/marvin/x_th_valign_baseline-expected.txt: Rebaseline after r167568.
- platform/gtk/tables/mozilla/other/body_col-expected.txt: Rebaseline after r167568.
- platform/gtk/tables/mozilla_expected_failures/bugs/bug10140-expected.txt: Rebaseline after r167568.
- platform/gtk/tables/mozilla_expected_failures/bugs/bug10216-expected.txt: Rebaseline after r167568.
- platform/gtk/tables/mozilla_expected_failures/core/captions3-expected.txt: Rebaseline after r167568.
- platform/gtk/tables/mozilla_expected_failures/other/test4-expected.txt: Rebaseline after r167568.
- platform/gtk/fast/repaint/spanner-with-margin-expected.txt: Added. Rebaseline after r167439.
- 12:10 PM Changeset in webkit [167603] by
-
- 4 edits in trunk/Source/WebKit2
Check the com.apple.security.network.client entitlement for all processes
https://bugs.webkit.org/show_bug.cgi?id=131942
<rdar://problem/12354188>
Reviewed by Dan Bernstein.
- NetworkProcess/EntryPoint/mac/XPCService/NetworkServiceEntryPoint.mm:
(WebKit::NetworkServiceInitializerDelegate::NetworkServiceInitializerDelegate):
- Shared/EntryPointUtilities/mac/XPCService/XPCServiceEntryPoint.h:
- Shared/EntryPointUtilities/mac/XPCService/XPCServiceEntryPoint.mm:
(WebKit::XPCServiceInitializerDelegate::checkEntitlements):
- 12:05 PM Changeset in webkit [167602] by
-
- 4 edits2 adds in trunk
Source/WebCore: [New Multicolumn] Column balancing is slow on float-multicol.html
https://bugs.webkit.org/show_bug.cgi?id=131801
Reviewed by Enrica Casucci.
Added fast/multicol/tall-float.html
- rendering/RenderBlock.cpp:
(WebCore::RenderBlock::expandsToEncloseOverhangingFloats):
Make RenderFlowThreads expand to encompass floats. Also from Morten's patch for
Blink (#361551).
- rendering/RenderBlockFlow.cpp:
(WebCore::RenderBlockFlow::computeOverflow):
Fix a bug where the new multi-column code was adding in overflow for floats
when it should not. This was resulting in the creation of scrollbars when none
should exist. This is my own addition.
(WebCore::RenderBlockFlow::addOverhangingFloats):
Make RenderFlowThreads and RenderRegions ignore overhanging float additions. This
comes from Morten's patch for Blink (#361551).
LayoutTests: http/tests/cache/subresource-failover-to-network.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=131936
Patch by Alexey Proskuryakov <ap@apple.com> on 2014-04-21
Reviewed by Brady Eidson.
The test was racy, because it was unintentionally queuing TestRunner operations
twice. Also, it failed to clean up its cookie.
Rewrote to not use TestRunner queuing, making the test work in browser.
- http/tests/cache/resources/subresource-failover-to-network.cgi:
- http/tests/cache/subresource-failover-to-network.html:
- 11:55 AM Changeset in webkit [167601] by
-
- 3 edits in trunk/LayoutTests
http/tests/cache/subresource-failover-to-network.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=131936
Reviewed by Brady Eidson.
The test was racy, because it was unintentionally queuing TestRunner operations
twice. Also, it failed to clean up its cookie.
Rewrote to not use TestRunner queuing, making the test work in browser.
- http/tests/cache/resources/subresource-failover-to-network.cgi:
- http/tests/cache/subresource-failover-to-network.html:
- 11:43 AM Changeset in webkit [167600] by
-
- 13 edits1 add in trunk/Source/JavaScriptCore
Take block execution count estimates into account when voting double
https://bugs.webkit.org/show_bug.cgi?id=131906
Reviewed by Geoffrey Garen.
This was a drama in three acts.
Act I: Slurp in BasicBlock::executionCount and use it as a weight when counting the
number of uses of a variable that want double or non-double. Easy as pie. This
gave me a huge speed-up on FloatMM and a huge slow-down on basically everything
else.
Act II: Realize that there were some programs where our previous double voting was
just on the edge of disaster and making it more precise tipped it over. In
particular, if you had an integer variable that would infrequently be used in a
computation that resulted in a variable that was frequently used as an array index,
the outer infrequentness would be the thing we'd use in the vote. So, an array
index would become double. We fix this by reviving global backwards propagation
and introducing the concept of ReallyWantsInt, which is used just for array
indices. Any variable transitively flagged as ReallyWantsInt will never be forced
double. We need that flag to be separate from UsedAsInt, since UsedAsInt needs to
be set in bitops for RageConversion but using it for double forcing is too much.
Basically, it's cheaper to have to convert a double to an int for a bitop than it
is to convert a double to an int for an array index; also a variable being used as
an array index is a much stronger hint that it ought to be an int. This recovered
performance on everything except programs that used FTL OSR entry.
Act III: Realize that OSR entrypoint creation creates blocks that have NaN execution
count, which then completely pollutes the weighting - essentially all votes go
NaN. Fix this with some surgical defenses. Basically, any client of execution
counts should allow for them to be NaN and shouldn't completely fall off a cliff
when it happens.
This is awesome. 75% speed-up on FloatMM. 11% speed-up on audio-dft. This leads to
7% speed-up on AsmBench and 2% speed-up on Kraken.
- CMakeLists.txt:
- JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
- JavaScriptCore.xcodeproj/project.pbxproj:
- dfg/DFGBackwardsPropagationPhase.cpp:
(JSC::DFG::BackwardsPropagationPhase::run):
(JSC::DFG::BackwardsPropagationPhase::propagate):
- dfg/DFGGraph.cpp:
(JSC::DFG::Graph::dumpBlockHeader):
- dfg/DFGGraph.h:
(JSC::DFG::Graph::voteNode):
(JSC::DFG::Graph::voteChildren):
- dfg/DFGNodeFlags.cpp:
(JSC::DFG::dumpNodeFlags):
- dfg/DFGNodeFlags.h:
- dfg/DFGOSREntrypointCreationPhase.cpp:
(JSC::DFG::OSREntrypointCreationPhase::run):
- dfg/DFGPlan.cpp:
(JSC::DFG::Plan::compileInThreadImpl):
- dfg/DFGPredictionPropagationPhase.cpp:
(JSC::DFG::PredictionPropagationPhase::doDoubleVoting):
(JSC::DFG::PredictionPropagationPhase::doRoundOfDoubleVoting):
- dfg/DFGVariableAccessData.cpp: Added.
(JSC::DFG::VariableAccessData::VariableAccessData):
(JSC::DFG::VariableAccessData::mergeIsCaptured):
(JSC::DFG::VariableAccessData::mergeShouldNeverUnbox):
(JSC::DFG::VariableAccessData::predict):
(JSC::DFG::VariableAccessData::mergeArgumentAwarePrediction):
(JSC::DFG::VariableAccessData::shouldUseDoubleFormatAccordingToVote):
(JSC::DFG::VariableAccessData::tallyVotesForShouldUseDoubleFormat):
(JSC::DFG::VariableAccessData::mergeDoubleFormatState):
(JSC::DFG::VariableAccessData::makePredictionForDoubleFormat):
(JSC::DFG::VariableAccessData::flushFormat):
- dfg/DFGVariableAccessData.h:
(JSC::DFG::VariableAccessData::vote):
(JSC::DFG::VariableAccessData::VariableAccessData): Deleted.
(JSC::DFG::VariableAccessData::mergeIsCaptured): Deleted.
(JSC::DFG::VariableAccessData::mergeShouldNeverUnbox): Deleted.
(JSC::DFG::VariableAccessData::predict): Deleted.
(JSC::DFG::VariableAccessData::mergeArgumentAwarePrediction): Deleted.
(JSC::DFG::VariableAccessData::shouldUseDoubleFormatAccordingToVote): Deleted.
(JSC::DFG::VariableAccessData::tallyVotesForShouldUseDoubleFormat): Deleted.
(JSC::DFG::VariableAccessData::mergeDoubleFormatState): Deleted.
(JSC::DFG::VariableAccessData::makePredictionForDoubleFormat): Deleted.
(JSC::DFG::VariableAccessData::flushFormat): Deleted.
- 11:16 AM WebKitGTK/KeepingTheTreeGreen edited by
- (diff)
- 11:10 AM Changeset in webkit [167599] by
-
- 3 edits1 add in trunk/Source/JavaScriptCore
REGRESSION(r167591): ARM64 and ARM traditional builds broken
https://bugs.webkit.org/show_bug.cgi?id=131935
Reviewed by Mark Hahnenberg.
Added store8(TrustedImm32, MacroAssembler::Address) to the ARM traditional and ARM64
macro assemblers. Added a new test for the original patch.
- assembler/MacroAssemblerARM.h:
(JSC::MacroAssemblerARM::store8):
- assembler/MacroAssemblerARM64.h:
(JSC::MacroAssemblerARM64::store8):
- tests/stress/dfg-create-arguments-inline-alloc.js: New test.
- 11:08 AM Changeset in webkit [167598] by
-
- 8 edits in trunk/Source/WebCore
REGRESSION (r160908): Safari doesn't draw rotated images properly first time
https://bugs.webkit.org/show_bug.cgi?id=131930
Reviewed by Anders Carlsson.
With render tree creation sufficiently lazy, we had to put some explicit
calls to updateStyleIfNeeded into the ImageDocument class.
- dom/DOMImplementation.cpp:
(WebCore::DOMImplementation::createDocument): Pass a Frame&.
- html/ImageDocument.cpp:
(WebCore::ImageDocument::imageSize): Added this helper function. This updates
style so we can use the renderer to compute an image size taking things like
orientation into account, and also passes in the zoom factor.
(WebCore::ImageDocument::updateDuringParsing): Moved the body of the
ImageDocumentParser::appendBytes function into this function. Since it's a
member of ImageDocument, it has better access to appropriate private members,
and we don't have to expose as many functions to the public. Also changed to
not use the frame so much. Also merged in the code that creates the document
instead of having it be a side effect in the cachedImage helper function; all
the other call sites were checking for null before calling cachedImage.
(WebCore::ImageDocument::finishedParsing): Moved the body of the
ImageDocumentParser::finish function here. Added a call to updateStyleIfNeeded
before using the renderer to get the image size.
(WebCore::ImageDocumentParser::document): Made this return a reference. It's
just a convenience function used in the two functions below.
(WebCore::ImageDocumentParser::appendBytes): Just call updateDuringParsing.
(WebCore::ImageDocumentParser::finish): Just call finishedParsing.
(WebCore::ImageDocument::ImageDocument): Take a reference for the argument,
use nullptr, and also initialize m_shouldShrinkImage directly instead of using
a shouldShrinkToFit() helper function to do it.
(WebCore::ImageDocument::createDocumentStructure): Removed unneeded check of
null for the frame, since the caller does it. Removed some unneeded
IGNORE_EXCEPTON and a stray semicolon. Use m_shouldShrinkImage instead of
calling shouldShrinkToFit since at this early point they will be identical.
(WebCore::ImageDocument::scale): Make non-const so we can call the new
imageSize helper. Use Document::view instead of Frame::view. Improved the
conversion to floating point so it doesn't involve a static_cast and an extra
conversion from int to LayoutUnit.
(WebCore::ImageDocument::resizeImageToFit): Ditto.
(WebCore::ImageDocument::imageClicked): Ditto.
(WebCore::ImageDocument::imageUpdated): Ditto. Also use m_shouldShrinkImage
ibnstead of shouldShrinkToFit.
(WebCore::ImageDocument::restoreImageSize): Ditto.
(WebCore::ImageDocument::imageFitsInWindow): Ditto.
(WebCore::ImageDocument::windowSizeChanged): Ditto.
(WebCore::ImageDocument::cachedImage): Deleted. Moved logic into updateDuringParsing.
(WebCore::ImageDocument::shouldShrinkToFit): Deleted. Moved logic into constructor.
(WebCore::ImageEventListener::handleEvent): Updated for changes above.
(WebCore::ImageEventListener::operator==): Simplified. No reason this has to check
whether both listeners are in the same document.
(WebCore::ImageDocumentElement::didMoveToNewDocument): Use nullptr.
- html/ImageDocument.h: Deleted some unused things, updated for changes above,
and made some memer functions private.
- loader/FrameLoaderClient.h: Deleted the allowScriptFromSource, allowPlugins,
allowImage, and allowDisplayingInsecureContent functions. These were only used in
Chromium and are now unneeded complexity. We could bring them back if anyone wants them.
- loader/MixedContentChecker.cpp:
(WebCore::MixedContentChecker::canDisplayInsecureContent): Removed call to
allowDisplayingInsecureContent, just use the setting directly.
(WebCore::MixedContentChecker::canRunInsecureContent): Removed call to
allowRunningInsecureContent, just use the setting directly.
- loader/SubframeLoader.cpp:
(WebCore::SubframeLoader::allowPlugins): Removed call to allowPlugins, just
use the arePluginsEnabled setting directly.
- loader/cache/CachedResourceLoader.cpp:
(WebCore::CachedResourceLoader::canRequest): Removed call to allowScriptFromSource, just
use the isScriptEnabled setting directly.
(WebCore::CachedResourceLoader::clientDefersImage): Removed call to allowImage.
- 11:07 AM Changeset in webkit [167597] by
-
- 7 edits2 adds in trunk
[New Multicolumn] Pagination mode messed up with non-inline axis and reversed direction.
https://bugs.webkit.org/show_bug.cgi?id=131811
Reviewed by Dean Jackson.
Source/WebCore:
Added fast/multicol/newmulticol/compare-with-old-impl/BottomToTop-tb.html
With block axis pagination mode, it is possible to set a column height that is not the same
as the available fill height for a block. The new multi-column code had the assumption that
the column height was the same as the amount of fill room you had available. This is not
the case.
To correct the issue, I added a member variable to RenderMultiColumnSet that stores the
available column height as a separate variable from the computed column height. This allows
the pagination API to specify a different column height that is not the same as the view's
content height.
Even though it isn't involved in the solution, I also patched pageOrViewLogicalHeight on
RenderView to work with the new column code as well.
To address the layout test failures (that caused the previous rollout), I made sure to
initialize m_availableHeight to 0 when m_computedColumnHeight also gets reset to 0.
The assertion is not something I could reproduce on any machine, but I can see the problem.
I patched Page's pageCount method to not have column code directly in Page.cpp,
and to make a new pageCount() method on RenderView that Page calls
into. This method is now patched to handle the new column code as well as the old. I have
no real way of testing this method though, since I can't reproduce the assertion that the
bots were experiencing.
- page/Page.cpp:
(WebCore::Page::pageCount):
- rendering/RenderMultiColumnSet.cpp:
(WebCore::RenderMultiColumnSet::RenderMultiColumnSet):
(WebCore::RenderMultiColumnSet::setAndConstrainColumnHeight):
(WebCore::RenderMultiColumnSet::prepareForLayout):
(WebCore::RenderMultiColumnSet::computeLogicalHeight):
- rendering/RenderMultiColumnSet.h:
- rendering/RenderView.cpp:
(WebCore::RenderView::pageOrViewLogicalHeight):
(WebCore::RenderView::pageCount):
- rendering/RenderView.h:
LayoutTests:
- fast/multicol/newmulticol/compare-with-old-impl/BottomToTop-tb-expected.html: Added.
- fast/multicol/newmulticol/compare-with-old-impl/BottomToTop-tb.html: Added.
- 11:06 AM Changeset in webkit [167596] by
-
- 6 edits1 delete in trunk/Source/WebCore
[Media] Clean up localized strings in controls
https://bugs.webkit.org/show_bug.cgi?id=131857
<rdar://problem/16663005>
Reviewed by Eric Carlson.
There is no need for separate localization files
on OS X and iOS. Move the small number of iOS-specific
localizations into a shared file.
We also don't need a function to return a static table.
Ultimately I think we want a top-level WebKit object that
hold the table, but for now use a global variable
called UIStringTable. Since we're in an isolated world, we
won't clash with anything.
I noticed a few cases were we grab a localized string and
then instantly run a replacement on it. This should be
supported directly in the UIString() accessor.
- English.lproj/mediaControlsLocalizedStrings.js: Add iOS strings. Just
use a static table.
(mediaControlsLocalizedStrings): Deleted.
- English.lproj/mediaControlsLocalizedStringsiOS.js: Removed.
- Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.extend): Fix braces.
(Controller.prototype.UIString): Use the string table, and allow replacement.
(Controller.prototype.createControls): Use new UIString replacement.
- Modules/mediacontrols/mediaControlsiOS.js:
(ControllerIOS.prototype.updateWirelessPlaybackStatus): Use new UIString.
(ControllerIOS.prototype.UIString): Deleted. Use the parent class version.
- WebCore.xcodeproj/project.pbxproj: Remove file from Resources.
- rendering/RenderThemeIOS.mm:
(WebCore::RenderThemeIOS::mediaControlsScript): No need to include the iOS
specific localization file.
- 10:31 AM Changeset in webkit [167595] by
-
- 2 edits in trunk/Source/WebCore
[Win] Unreviewed build fix.
- inspector/TimelineRecordFactory.cpp: Add missing include.
- 10:11 AM Changeset in webkit [167594] by
-
- 26 edits in trunk/Source
Move cross-port Source/WebCore/page/ code to std::unique_ptr
https://bugs.webkit.org/show_bug.cgi?id=130534
Reviewed by Darin Adler.
Source/WebCore:
Replace uses of OwnPtr and PassOwnPtr in cross-port code under Source/WebCore/page/ with
std::unique_ptr. Unnecessary OwnPtr.h and PassOwnPtr.h header inclusions are removed as well.
- dom/Document.cpp:
(WebCore::Document::initSecurityContext):
- dom/SecurityContext.cpp:
(WebCore::SecurityContext::setContentSecurityPolicy):
- dom/SecurityContext.h:
- page/CaptionUserPreferences.h:
- page/ContentSecurityPolicy.cpp:
(WebCore::CSPDirectiveList::create):
(WebCore::CSPDirectiveList::checkEvalAndReportViolation):
(WebCore::CSPDirectiveList::checkInlineAndReportViolation):
(WebCore::CSPDirectiveList::checkSourceAndReportViolation):
(WebCore::CSPDirectiveList::setCSPDirective):
(WebCore::ContentSecurityPolicy::didReceiveHeader):
- page/ContentSecurityPolicy.h:
- page/DOMSecurityPolicy.h:
- page/DOMWindow.cpp:
(WebCore::PostMessageTimer::fired):
(WebCore::DOMWindow::postMessageTimerFired):
- page/DOMWindow.h:
- page/EventHandler.cpp:
(WebCore::EventHandler::EventHandler):
- page/EventHandler.h:
- page/Frame.cpp:
(WebCore::Frame::Frame):
(WebCore::createRegExpForLabels):
(WebCore::Frame::searchForLabelsAboveCell):
(WebCore::Frame::searchForLabelsBeforeElement):
(WebCore::matchLabelsAgainstString):
(WebCore::Frame::setTiledBackingStoreEnabled):
- page/Frame.h:
- page/FrameView.cpp:
(WebCore::FrameView::addEmbeddedObjectToUpdate):
(WebCore::FrameView::addSlowRepaintObject):
(WebCore::FrameView::addViewportConstrainedObject):
(WebCore::FrameView::addScrollableArea):
- page/FrameView.h:
- page/Page.cpp:
(WebCore::Page::Page):
- page/Page.h:
- page/PageConsole.h:
- page/PageThrottler.h:
- page/SecurityPolicy.cpp:
(WebCore::SecurityPolicy::addOriginAccessWhitelistEntry):
- page/animation/AnimationController.h:
- page/animation/CSSPropertyAnimation.cpp:
(WebCore::PropertyWrapperVisitedAffectedColor::PropertyWrapperVisitedAffectedColor):
(WebCore::CSSPropertyAnimationWrapperMap::instance):
(WebCore::CSSPropertyAnimationWrapperMap::CSSPropertyAnimationWrapperMap):
- page/mac/PageMac.cpp:
(WebCore::Page::addSchedulePair):
- workers/WorkerGlobalScope.cpp:
(WebCore::WorkerGlobalScope::applyContentSecurityPolicyFromString):
Source/WebKit/mac:
- WebView/WebHTMLRepresentation.mm:
(searchForLabelsBeforeElement): Pass a RegExp reference to Frame::searchForLabelsAboveCell().
- 9:56 AM Changeset in webkit [167593] by
-
- 2 edits in trunk/LayoutTests
REGRESSION (r167530): ASSERT(m_selfTime <= m_totalTime) on tests that run after certain Inspector tests
https://bugs.webkit.org/show_bug.cgi?id=131919
- TestExpectations: Skipping some Inspector tests, hopefully these are the ones that
cause trouble later.
- 8:52 AM Changeset in webkit [167592] by
-
- 7 edits in trunk
Add HashSet::takeAny
https://bugs.webkit.org/show_bug.cgi?id=131928
Reviewed by Benjamin Poulain.
Source/WebCore:
- dom/Document.cpp:
(WebCore::Document::takeAnyMediaCanStartListener): Use HashSet::takeAny.
- dom/ScriptExecutionContext.cpp:
(WebCore::takeAny): Deleted.
(WebCore::ScriptExecutionContext::~ScriptExecutionContext): Use HashSet::takeAny.
Source/WTF:
- wtf/HashSet.h: Added an overload of take that takes an iterator,
and used it to implement both the existing take and new takeAny functions.
Tools:
- TestWebKitAPI/Tests/WTF/HashSet.cpp: Added a test for takeAny.
- 8:11 AM Changeset in webkit [167591] by
-
- 6 edits in trunk/Source/JavaScriptCore
Inline allocate Arguments objects in the DFG
https://bugs.webkit.org/show_bug.cgi?id=131897
Reviewed by Geoffrey Garen.
Many libraries/frameworks depend on the arguments object for overloaded API entry points.
This is the first step to making Arguments fast(er). We'll duplicate the logic in Arguments::create
for now and take the slow path for complicated cases like slow arguments, tearing off for strict mode, etc.
- dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::emitAllocateArguments):
- dfg/DFGSpeculativeJIT.h:
(JSC::DFG::SpeculativeJIT::emitAllocateDestructibleObject):
- dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
- dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
- runtime/Arguments.h:
(JSC::Arguments::offsetOfActivation):
(JSC::Arguments::offsetOfOverrodeLength):
(JSC::Arguments::offsetOfIsStrictMode):
(JSC::Arguments::offsetOfRegisterArray):
(JSC::Arguments::offsetOfCallee):
(JSC::Arguments::allocationSize):
- 4:52 AM WebKitGTK/KeepingTheTreeGreen edited by
- (diff)
- 3:39 AM Changeset in webkit [167590] by
-
- 4 edits in trunk/Source/WebCore
iOS and non-iOS RemoteCommandListener::create() functions should use std::make_unique<>()
https://bugs.webkit.org/show_bug.cgi?id=131916
Reviewed by Darin Adler.
A follow-up to r167197. Make RemoteCommandListener ctor public and use std::make_unique<>()
in both versions of RemoteCommandListener::create() functions (iOS and non-iOS).
- platform/RemoteCommandListener.cpp:
(WebCore::RemoteCommandListener::create):
- platform/RemoteCommandListener.h:
(WebCore::RemoteCommandListener::RemoteCommandListener):
- platform/ios/RemoteCommandListenerIOS.mm:
(WebCore::RemoteCommandListener::create):
- 3:15 AM Changeset in webkit [167589] by
-
- 2 edits in trunk/Source/WebCore
Micro-optimize the way we hand NodeLists to JSC.
<https://webkit.org/b/131932>
Use HashMap::fastAdd() when returning cached node lists and collections.
10.9% progression on Bindings/get-elements-by-tag-name.html
Reviewed by Antti Koivisto.
- dom/NodeRareData.h:
(WebCore::NodeListsNodeData::addCacheWithAtomicName):
(WebCore::NodeListsNodeData::addCacheWithName):
(WebCore::NodeListsNodeData::addCacheWithQualifiedName):
(WebCore::NodeListsNodeData::addCachedCollection):
- 12:25 AM Changeset in webkit [167588] by
-
- 32 edits in trunk
Unreviewed, rolling out r167584.
https://bugs.webkit.org/show_bug.cgi?id=131929
Broke Objective-C bindings test (Requested by ap on #webkit).
Reverted changeset:
"Add Element.matches, the standard name for
webkitMatchesSelector"
https://bugs.webkit.org/show_bug.cgi?id=131922
http://trac.webkit.org/changeset/167584
Apr 20, 2014:
- 11:42 PM Changeset in webkit [167587] by
-
- 2 edits in trunk/Source/WebKit
Blind Windows build fix. Hope I guessed the 64-bit mangled form right.
- WebKit.vcxproj/WebKitExportGenerator/WebKitExports.def.in:
- 11:35 PM Changeset in webkit [167586] by
-
- 5 edits9 deletes in trunk/Source/WebKit2
[Cocoa] Remove interfaces in the To Be Removed groups
https://bugs.webkit.org/show_bug.cgi?id=131925
Reviewed by Darin Adler.
- Shared/API/Cocoa/WKRemoteObjectInterface.h: Removed.
- Shared/API/Cocoa/WKRemoteObjectInterface.mm: Removed.
- Shared/API/Cocoa/WKRemoteObjectRegistry.h: Removed.
- Shared/API/Cocoa/WKRemoteObjectRegistry.mm: Removed.
- UIProcess/API/Cocoa/WKProcessPool.mm:
- UIProcess/API/Cocoa/WKProcessPoolConfiguration.h: Removed.
- UIProcess/API/Cocoa/WKProcessPoolConfiguration.mm: Removed.
- UIProcess/API/Cocoa/WKProcessPoolConfigurationPrivate.h: Removed.
- UIProcess/API/Cocoa/WKProcessPoolToBeRemoved.mm: Removed.
- UIProcess/API/Cocoa/_WKProcessPoolConfiguration.mm:
- WebKit2.xcodeproj/project.pbxproj:
- WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInBrowserContextControllerToBeRemoved.mm: Removed.
- WebProcess/InjectedBundle/API/mac/WKWebProcessPlugInBrowserContextControllerPrivate.h:
- 11:26 PM Changeset in webkit [167585] by
-
- 4 edits2 adds in trunk
Compile the :root pseudo class and fix a related issue with :nth-child()
https://bugs.webkit.org/show_bug.cgi?id=131926
Reviewed by Andreas Kling.
Source/WebCore:
Add the :root pseudo class. This is another trivial selector, we just need to compare
the element pointer with the documentElement.
I discovered some issues with :nth-child(n) through the layout tests for ":root".
When the pseudo class nth-child could match anything, no code was generated. That decision
was taken when generating the fragments.
The specification of :nth-child() has two tests: the parent test and the counter test.
Since some fragments would not generate any code for :nth-child(n), they would succeed on the root,
which is incorrect since the root should fail the parent test.
This was fixed by moving the filtering of non-counting :nth-child() after we generate the parent
check.
We still don't generate any counter test unless required.
Test: fast/selectors/nth-child-on-root.html
- cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::addPseudoClassType):
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatching):
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsNthChild):
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsRoot):
- dom/Document.h:
(WebCore::Document::documentElementMemoryOffset):
LayoutTests:
Add more test coverage that would have caught the bug with :nth-child(n).
- fast/selectors/nth-child-on-root-expected.txt: Added.
- fast/selectors/nth-child-on-root.html: Added.
- 11:08 PM Changeset in webkit [167584] by
-
- 32 edits in trunk
Add Element.matches, the standard name for webkitMatchesSelector
https://bugs.webkit.org/show_bug.cgi?id=131922
Reviewed by Andreas Kling.
Source/WebCore:
This patch just renames webkitMatchesSelector() to matches().
- bindings/scripts/CodeGeneratorObjC.pm:
(GenerateImplementation):
The Objective-C binding generator only supported ImplementedAs for getters and setters.
This adds support for the option so that Element.matches can be generated for Obj-C.
- dom/Element.cpp:
(WebCore::Element::matches):
(WebCore::Element::webkitMatchesSelector): Deleted.
- dom/Element.h:
- dom/Element.idl:
- inspector/InspectorCSSAgent.cpp:
(WebCore::InspectorCSSAgent::buildArrayForMatchedRuleList):
LayoutTests:
Update the tests to verify both Element.matches and Element.webkitMatchesSelector
- fast/dom/SelectorAPI/attrname-case-insensitive-expected.txt:
- fast/dom/SelectorAPI/attrname-case-insensitive.html:
- fast/dom/SelectorAPI/attrname-case-sensitive-expected.txt:
- fast/dom/SelectorAPI/attrname-case-sensitive.xhtml:
- fast/dom/SelectorAPI/caseID-almost-strict-expected.txt:
- fast/dom/SelectorAPI/caseID-almost-strict.html:
- fast/dom/SelectorAPI/caseID-expected.txt:
- fast/dom/SelectorAPI/caseID-strict-expected.txt:
- fast/dom/SelectorAPI/caseID-strict.html:
- fast/dom/SelectorAPI/caseID.html:
- fast/dom/SelectorAPI/caseTag-expected.txt:
- fast/dom/SelectorAPI/caseTag.html:
- fast/dom/SelectorAPI/caseTagX-expected.txt:
- fast/dom/SelectorAPI/caseTagX.xhtml:
- fast/dom/SelectorAPI/detached-element-expected.txt:
- fast/dom/SelectorAPI/not-supported-namespace-in-selector-expected.txt:
- fast/dom/SelectorAPI/not-supported-namespace-in-selector.html:
- fast/dom/SelectorAPI/script-tests/detached-element.js:
- fast/dom/SelectorAPI/script-tests/undefined-null-stringify.js:
- fast/dom/SelectorAPI/script-tests/viewless-document.js:
- fast/dom/SelectorAPI/undefined-null-stringify-expected.txt:
- fast/dom/SelectorAPI/viewless-document-expected.txt:
- fast/forms/radio/radio-live-validation-style-expected.txt:
- fast/forms/radio/radio-live-validation-style.html:
- fast/harness/results.html:
- 11:02 PM Changeset in webkit [167583] by
-
- 5 edits in trunk/LayoutTests
Expando properties on attribute nodes disappear
https://bugs.webkit.org/show_bug.cgi?id=88045
- TestExpectations:
- platform/efl/TestExpectations:
- platform/mac/TestExpectations:
- platform/win/TestExpectations:
Consolidated test expectations, because all ports are equally affected. Some
platforms accidentally start to pass this test sometimes, but there is no real fix yet.
- 10:21 PM Changeset in webkit [167582] by
-
- 3 edits2 adds in trunk
Subpixel rendering: RenderLayer's size is set using enclosingRect() which can result in cruft.
https://bugs.webkit.org/show_bug.cgi?id=131468
Reviewed by Darin Adler.
Layers marked opaque may not paint edge pixels, when RenderLayer bounds are non-integral and
get device pixel snapped.
For details see this: https://trac.webkit.org/r159463
Source/WebCore:
Test: compositing/hidpi-non-simple-compositing-layer-with-fractional-size-and-background.html
- rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::updateGraphicsLayerGeometry):
LayoutTests:
- compositing/hidpi-non-simple-compositing-layer-with-fractional-size-and-background-expected.html: Added.
- compositing/hidpi-non-simple-compositing-layer-with-fractional-size-and-background.html: Added.
- 9:43 PM Changeset in webkit [167581] by
-
- 1 edit in trunk/Source/WebCore/ChangeLog
Fix typo in change log
- 9:40 PM Changeset in webkit [167580] by
-
- 3 edits2 adds in trunk
Simple layout can get confused by coordinate overflow
https://bugs.webkit.org/show_bug.cgi?id=131890
rdar://problem/15558510
Reviewed by Andreas Kling.
Source/WebCore:
Test: fast/css/simple-layout-overflow.html
- rendering/SimpleLineLayoutResolver.h:
(WebCore::SimpleLineLayout::RunResolver::rangeForRect):
If lastLine is smaller than firstList, just ignore it.
LayoutTests:
- fast/css/simple-layout-overflow-expected.txt: Added.
- fast/css/simple-layout-overflow.html: Added.
- 9:39 PM Changeset in webkit [167579] by
-
- 11 edits in trunk/Source/WebCore
ScriptExecutionContext::stopActiveDOMObjects iterates a hash map that can change during iteration (for multiple reasons, including GC)
https://bugs.webkit.org/show_bug.cgi?id=52719
Reviewed by Alexey Proskuryakov.
At least two specific ways this can happen:
1) XMLHttpRequest::stop can trigger a JavaScript garbage collection.
2) NotificationCenter::stop can delete the last references to notifications;
those notifications are also active DOM objects.
Besides fixing the iteration in that function, did some other fixes for the
ScriptExecutionContext class, including some coding style changes. Many uses
of nullptr instead of 0, without listing each function separately below.
- Modules/webdatabase/DatabaseContext.cpp:
(WebCore::DatabaseContext::contextDestroyed): Call through to the base class
version of contextDestroyed rather than repeating what it does (with a large
comment that doesn't acknowledge the base class alread does it).
- Modules/webdatabase/DatabaseContext.h: Removed some unneeded includes.
Wrote out "private" explicitly for deriving from ActiveDOMObject. Made the
ActiveDOMObject function overrides private, and marked them override and final.
- dom/ActiveDOMObject.h: Updated comments. Replaced suspendIfNeededCalled with
assertSuspendIfNeededWasCalled, which has an empty inline version in the header.
Renamed m_suspendIfNeededCalled to m_suspendIfNeededWasCalled.
- dom/ActiveDOMObject.cpp:
(WebCore::ActiveDOMObject::ActiveDOMObject): Pass a reference instead of a pointer.
(WebCore::ActiveDOMObject::~ActiveDOMObject): Ditto.
(WebCore::ActiveDOMObject::suspendIfNeeded): Ditto.
- dom/ContextDestructionObserver.cpp:
(WebCore::ContextDestructionObserver::observeContext): Pass a reference instead of a pointer.
- dom/MessagePort.cpp:
(WebCore::MessagePort::MessagePort): Pass a reference instead of a pointer.
(WebCore::MessagePort::~MessagePort): Ditto.
(WebCore::MessagePort::disentangle): Ditto.
- dom/ScriptExecutionContext.cpp:
(WebCore::ScriptExecutionContext::ScriptExecutionContext): Updated flags used
for assertions so they are conditional and updated their names.
(WebCore::takeAny): Added. Helper function that we can consider for HashSet in
the future; makes loop below easier to read.
(WebCore::checkConsistency): Added. Assertions that were done multiple places below,
and should not be written over and over again.
(WebCore::ScriptExecutionContext::~ScriptExecutionContext): Changed to use C++11
for loops and the takeAny function above.
(WebCore::ScriptExecutionContext::dispatchMessagePortEvents): Ditto.
(WebCore::ScriptExecutionContext::createdMessagePort): Changed to take a reference
for clarity and so it doesn't have to do an assert the pointer is non-null.
(WebCore::ScriptExecutionContext::destroyedMessagePort): Ditto.
(WebCore::ScriptExecutionContext::canSuspendActiveDOMObjects): Changed to use
C++11 for loop and reworded comment and redid assertions.
(WebCore::ScriptExecutionContext::suspendActiveDOMObjects): Ditto.
(WebCore::ScriptExecutionContext::resumeActiveDOMObjects): Ditto.
(WebCore::ScriptExecutionContext::stopActiveDOMObjects): Changed to support
removal of an active DOM object during the stop function. Included new comments
to clarify what the rules are.
(WebCore::ScriptExecutionContext::suspendActiveDOMObjectIfNeeded): Changed to take
a reference for clarity and so it doesn't have to assert a pointer is non-null.
(WebCore::ScriptExecutionContext::didCreateActiveDOMObject): Ditto. Also changed to
use RELEASE_ASSERT instead of CRASH.
(WebCore::ScriptExecutionContext::willDestroyActiveDOMObject): Ditto.
(WebCore::ScriptExecutionContext::didCreateDestructionObserver): Ditto.
(WebCore::ScriptExecutionContext::willDestroyDestructionObserver): Ditto.
(WebCore::ScriptExecutionContext::closeMessagePorts): Moved the body of this
function into its one call site, ScriptExecutionContext::stopActiveDOMObjects,
since it's simple enough when written as a C++11 for loop.
(WebCore::ScriptExecutionContext::hasPendingActivity): Added. This function was
already exported for workers, and implementing it outside this class required
exposing the private HashSet members; more sensible to implement it here and
simply make it public in WorkerGlobalScope.
- dom/ScriptExecutionContext.h: Removed unnecessary includes and forward declarations.
Removed a long-ago-fixed FIXME. Changed various functions to take references instead of
pointers. Added a protected hasPendingActivity function, deleted the closeMessagePorts
function, deleted the ActiveDOMObjectsSet typedef, made the assertion flags be
!ASSERT_DISABLED only, and deleted the messagePorts and activeDOMObjects functions.
- workers/WorkerGlobalScope.cpp:
(WebCore::WorkerGlobalScope::hasPendingActivity): Deleted. This is now implemented
in the base class.
- workers/WorkerGlobalScope.h: Make hasPendingActivity function from the base class
public instead of declaring it in this class.
- 9:28 PM Changeset in webkit [167578] by
-
- 2 edits in trunk/Source/WebCore
[Mac] Unable to select 'Off' or 'Auto' from track menu when tracks consist of unsupported track types
https://bugs.webkit.org/show_bug.cgi?id=131908
<rdar://problem/15999882>
Reviewed by Eric Carlson.
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::setSelectedTextTrack): When the user selects 'Off', make sure the currently 'showing'
track is switched to 'disabled'.
- 9:19 PM Changeset in webkit [167577] by
-
- 9 edits in trunk/Source
Speed up jsStringWithCache() through WeakGCMap inlining.
<https://webkit.org/b/131923>
Source/JavaScriptCore:
Always inline WeakGCMap::add() but move the slow garbage collecting
path out-of-line.
Reviewed by Darin Adler.
- runtime/WeakGCMap.h:
(JSC::WeakGCMap::add):
(JSC::WeakGCMap::gcMap):
Source/WebCore:
Inline the common path of WeakGCMap::add() in jsStringWithCache().
26% progression on Bindings/id-getter.html
Reviewed by Darin Adler.
- WebCore.exp.in:
- bindings/js/JSDOMBinding.h:
- bindings/js/JSDOMBinding.cpp:
(WebCore::jsStringWithCache):
Move jsStringWithCache() out of line since we're now blowing up
its size quite a bit.
Source/WTF:
Add HashMap::fastAdd(), which is the same as add() except we'll tell
the compiler to aggressively inline it.
Reviewed by Darin Adler.
- wtf/HashMap.h:
- wtf/HashTable.h:
- 3:45 PM Changeset in webkit [167576] by
-
- 2 edits in trunk/Tools
Fixed a test timing out after r167572.
- TestWebKitAPI/Tests/WebKit2Cocoa/Navigation.mm:
(-[NavigationDelegate webView:didFinishNavigation:]): Renamed from
-webView:didFinishLoadingNavigation:.
- 2:53 PM Changeset in webkit [167575] by
-
- 2 edits in trunk/Source/WebCore
Fix the build after r167574
- cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatching):
- 2:28 PM Changeset in webkit [167574] by
-
- 3 edits in trunk/Source/WebCore
Compile the pseudo class :target
https://bugs.webkit.org/show_bug.cgi?id=131904
Reviewed by Andreas Kling.
The pseudo class :target is trivial, it just compare the current element with
Document::cssTarget().
- cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::addPseudoType):
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatching):
(WebCore::SelectorCompiler::getDocument):
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementAttributeValueExactMatching):
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsTarget):
- dom/Document.h:
(WebCore::Document::cssTargetMemoryOffset):
- 1:54 PM Changeset in webkit [167573] by
-
- 4 edits in trunk/Source/WebKit2
[Cocoa] Crash when setting a bundle parameter to nil
https://bugs.webkit.org/show_bug.cgi?id=131917
Reviewed by Sam Weinig.
- UIProcess/API/Cocoa/WKProcessPool.mm:
(-[WKProcessPool _setObject:forBundleParameter:]): If the object is nil, use
-[NSMutableDictionary removeObjectForKey:] rather than -setObject:forKey:.
- WebProcess/InjectedBundle/API/Cocoa/WKWebProcessBundleParameters.mm:
(-[WKWebProcessBundleParameters setParameter:forKey:]): Ditto.
- WebProcess/InjectedBundle/mac/InjectedBundleMac.mm:
(WebKit::InjectedBundle::setBundleParameter): Unrelatedly, create m_bundleParameters if
needed, as it may have not been created on intialization if no bundle parameters were set
at the time.
- 1:53 PM Changeset in webkit [167572] by
-
- 3 edits in trunk/Source/WebKit2
[Cocoa] Remove unused -webView:didFinishLoadingNavigation: delegate method
https://bugs.webkit.org/show_bug.cgi?id=131914
Reviewed by Sam Weinig.
- UIProcess/Cocoa/NavigationState.h: Removed flag from m_navigationDelegateMethods struct.
- UIProcess/Cocoa/NavigationState.mm:
(WebKit::NavigationState::setNavigationDelegate): Removed check for whether the method is
implemented.
(WebKit::NavigationState::LoaderClient::didFinishLoadForFrame): Stopped calling the method
if implemented.
- 1:43 PM Changeset in webkit [167571] by
-
- 19 edits in trunk/Source/WebCore
Rename the CSSSelector PseudoType to PseudoClassType
https://bugs.webkit.org/show_bug.cgi?id=131907
Reviewed by Andreas Kling.
Pseudo Elements and Page Pseudo Classes have been moved out of PseudoType in previous patches.
The remaining values in the PseudoType enumeration are the pseudo classes.
This patch is the final clean up, PseudoType is renamed to PseudoClassType.
- css/CSSGrammar.y.in:
- css/CSSParserValues.cpp:
(WebCore::CSSParserSelector::parsePseudoClassAndCompatibilityElementSelector):
- css/CSSParserValues.h:
(WebCore::CSSParserSelector::pseudoClassType):
(WebCore::CSSParserSelector::pseudoType): Deleted.
- css/CSSSelector.cpp:
(WebCore::CSSSelector::specificityForOneSelector):
(WebCore::appendPseudoClassFunctionTail):
(WebCore::CSSSelector::selectorText):
- css/CSSSelector.h:
(WebCore::CSSSelector::pseudoClassType):
(WebCore::CSSSelector::pseudoElementType):
(WebCore::CSSSelector::pagePseudoClassType):
(WebCore::pseudoClassIsRelativeToSiblings):
(WebCore::CSSSelector::isSiblingSelector):
(WebCore::CSSSelector::CSSSelector):
(WebCore::CSSSelector::pseudoType): Deleted.
- css/RuleSet.cpp:
(WebCore::RuleSet::findBestRuleSetAndAdd):
- css/SelectorChecker.cpp:
(WebCore::SelectorChecker::checkOne):
(WebCore::SelectorChecker::checkScrollbarPseudoClass):
(WebCore::SelectorChecker::determineLinkMatchType):
(WebCore::SelectorChecker::matchesFocusPseudoClass):
- css/SelectorChecker.h:
(WebCore::SelectorChecker::isCommonPseudoClassSelector):
- css/SelectorCheckerFastPath.cpp:
(WebCore::SelectorCheckerFastPath::commonPseudoClassSelectorMatches):
- css/SelectorPseudoClassAndCompatibilityElementMap.in:
- css/SelectorPseudoTypeMap.h:
- css/StyleResolver.cpp:
(WebCore::StyleResolver::styleForElement):
- css/makeSelectorPseudoClassAndCompatibilityElementMap.py:
(enumerablePseudoType):
- cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::addPseudoClassType):
(WebCore::SelectorCompiler::SelectorCodeGenerator::SelectorCodeGenerator):
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatching):
(WebCore::SelectorCompiler::addPseudoType): Deleted.
- inspector/InspectorCSSAgent.cpp:
(WebCore::computePseudoClassMask):
(WebCore::InspectorCSSAgent::forcePseudoState):
- inspector/InspectorCSSAgent.h:
- inspector/InspectorInstrumentation.cpp:
(WebCore::InspectorInstrumentation::forcePseudoStateImpl):
- inspector/InspectorInstrumentation.h:
(WebCore::InspectorInstrumentation::forcePseudoState):
- 12:33 PM Changeset in webkit [167570] by
-
- 10 edits in trunk/Source/bmalloc
bmalloc: Segregate pages by objects size
https://bugs.webkit.org/show_bug.cgi?id=131909
Reviewed by Andreas Kling.
2% reduction in memory-at-end on the Membuster memory_warning benchmarks.
- bmalloc/Allocator.cpp:
(bmalloc::Allocator::allocateSlowCase):
- bmalloc/Allocator.h:
(bmalloc::Allocator::allocateFastCase):
(bmalloc::Allocator::smallAllocatorFor): Use the new shared helper
function for size class calculation.
- bmalloc/Deallocator.cpp:
(bmalloc::Deallocator::Deallocator):
(bmalloc::Deallocator::scavenge):
(bmalloc::Deallocator::deallocateSmallLine):
(bmalloc::Deallocator::allocateSmallLine):
- bmalloc/Deallocator.h: Keep a cache for every size class, since the
cache can't be shared anymore.
- bmalloc/Heap.cpp:
(bmalloc::Heap::allocateSmallLineSlowCase):
- bmalloc/Heap.h:
(bmalloc::Heap::deallocateSmallLine): Ditto.
(bmalloc::Heap::allocateSmallLine): Check size class in addition to
page refcount when allocating a line because we might have deallocated
the page and the recycled it for another size class.
(bmalloc::Heap::deallocateMediumLine):
(bmalloc::Heap::allocateMediumLine):
- bmalloc/Line.h:
(bmalloc::Line::refCount):
- bmalloc/Page.h:
(bmalloc::Page::refCount):
(bmalloc::Page::smallSizeClass):
(bmalloc::Page::setSmallSizeClass):
(bmalloc::Page<Traits>::refCount): Deleted.
- bmalloc/Sizes.h:
(bmalloc::Sizes::smallSizeClassFor): New shared API for computing
an index into an array from a size.
- 12:02 PM Changeset in webkit [167569] by
-
- 4 edits4 adds in trunk
Crashes in HTMLFormElement::submit.
https://bugs.webkit.org/show_bug.cgi?id=131910
<rdar://problem/15661790>
Source/WebCore:
Based on a patch by Kent Tamura.
Reviewed by Anders Carlsson.
Tests: fast/forms/form-submission-crash-2.html
fast/forms/form-submission-crash.html
Code that executes arbitrary JS needs to protect objects that it uses afterwards.
- html/HTMLFormElement.cpp:
(WebCore::HTMLFormElement::prepareForSubmission):
(WebCore::HTMLFormElement::submit):
- loader/FrameLoader.cpp:
(WebCore::FrameLoader::submitForm):
LayoutTests:
Reviewed by Anders Carlsson.
- fast/forms/form-submission-crash-2-expected.txt: Added.
- fast/forms/form-submission-crash-2.html: Added.
- fast/forms/form-submission-crash-expected.txt: Added.
- fast/forms/form-submission-crash.html: Added.
- 10:43 AM Changeset in webkit [167568] by
-
- 23 edits in trunk
Text bounding box computation for simple line layout is wrong
https://bugs.webkit.org/show_bug.cgi?id=131912
Reviewed by Andreas Kling.
Source/WebCore:
Top-left is currently the first line top-left which is not always correct.
- WebCore.exp.in:
- rendering/RenderText.cpp:
(WebCore::RenderText::firstRunLocation):
(WebCore::RenderText::firstRunOrigin): Deleted.
(WebCore::RenderText::firstRunX): Deleted.
(WebCore::RenderText::firstRunY): Deleted.
Keep just one accessor and rename it.
Encapsulate the line box and simple line versions.
- rendering/RenderText.h:
- rendering/RenderTextLineBoxes.cpp:
(WebCore::RenderTextLineBoxes::firstRunLocation):
Line box version.
- rendering/RenderTextLineBoxes.h:
- rendering/RenderTreeAsText.cpp:
Simplify RenderText dumping.
(WebCore::RenderTreeAsText::writeRenderObject):
- rendering/SimpleLineLayoutFunctions.cpp:
(WebCore::SimpleLineLayout::computeTextBoundingBox):
Return the correct x position.
(WebCore::SimpleLineLayout::computeTextFirstRunLocation):
Simple line version.
- rendering/SimpleLineLayoutFunctions.h:
- rendering/svg/SVGRenderTreeAsText.cpp:
(WebCore::writeSVGInlineText):
LayoutTests:
Some dumped RenderText sizes change in table related test. These are progressions,
the new results match the contained lines. There are no visual changes.
- platform/mac/fast/table/dynamic-caption-add-remove-before-child-expected.txt:
- platform/mac/fast/table/multiple-captions-display-expected.txt:
- platform/mac/tables/mozilla/marvin/body_col-expected.txt:
- platform/mac/tables/mozilla/marvin/x_th_valign_baseline-expected.txt:
- platform/mac/tables/mozilla/other/body_col-expected.txt:
- platform/mac/tables/mozilla_expected_failures/bugs/bug10140-expected.txt:
- platform/mac/tables/mozilla_expected_failures/bugs/bug10216-expected.txt:
- platform/mac/tables/mozilla_expected_failures/core/captions3-expected.txt:
- platform/mac/tables/mozilla_expected_failures/other/test4-expected.txt:
- 10:40 AM Changeset in webkit [167567] by
-
- 4 edits in trunk/Tools
Unreviewed, rolling out r167510.
https://bugs.webkit.org/show_bug.cgi?id=131915
broke GTK Debug bot (Requested by philn on #webkit).
Reverted changeset:
"[GTK] Add llvmpipe (Mesa) to the JHBuild moduleset and force
it when running layout tests"
https://bugs.webkit.org/show_bug.cgi?id=131472
http://trac.webkit.org/changeset/167510
- 6:58 AM Changeset in webkit [167566] by
-
- 4 edits in trunk/Source/JavaScriptCore
JavaScriptCore: ARM build fix after r167094.
https://bugs.webkit.org/show_bug.cgi?id=131612
Patch by László Langó <llango.u-szeged@partner.samsung.com> on 2014-04-20
Reviewed by Michael Saboff.
After r167094 there are many build errors on ARM like these:
/tmp/ccgtHRno.s:370: Error: invalid constant (425a) after fixup
/tmp/ccgtHRno.s:374: Error: invalid constant (426e) after fixup
/tmp/ccgtHRno.s:378: Error: invalid constant (4282) after fixup
/tmp/ccgtHRno.s:382: Error: invalid constant (4296) after fixup
Problem is caused by the wrong generated assembly like:
"\tmov r2, (" LOCAL_LABEL_STRING(llint_op_strcat) " - " LOCAL_LABEL_STRING(relativePCBase) ")\n" /home/webkit/WebKit/Source/JavaScriptCore/llint/LowLevelInterpreter.asm:741
movcan only move 8 bit immediate, but not every constant fit into 8 bit. Clang converts
the mov to a single movw or a movw and a movt, depending on the immediate, but binutils doesn't.
Add a new ARM specific offline assembler instruction (mvlbl) for the following llint_entry
use case: move rn, (label1-label2) which is translated to movw and movt.
- llint/LowLevelInterpreter.asm:
- offlineasm/arm.rb:
- offlineasm/instructions.rb:
- 6:46 AM Changeset in webkit [167565] by
-
- 2 edits in trunk/Source/JavaScriptCore
[ARM] Unreviewed build fix after r167336.
- assembler/MacroAssemblerARM.h:
(JSC::MacroAssemblerARM::branchAdd32):
- 4:45 AM Changeset in webkit [167564] by
-
- 6 edits3 deletes in trunk
Unreviewed, rolling out r167501.
https://bugs.webkit.org/show_bug.cgi?id=131913
It broke DYEBench (Requested by mhahnenberg on #webkit).
Reverted changeset:
"Deleting properties poisons objects"
https://bugs.webkit.org/show_bug.cgi?id=131551
http://trac.webkit.org/changeset/167501
