Skip to content

GitLab

W
webkit
Switch branch/tag
  1. 06 Apr, 2013 14 commits
  3. 05 Apr, 2013 26 commits
    • kangil.han@samsung.com's avatar
      Prefer prefix ++/-- operators for non-primitive types · d97b4ea1
      kangil.han@samsung.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=114033

Reviewed by Alexey Proskuryakov.

Post ++/-- creates a copy of current value and it is not necessary, so use prefix instead.

* bindings/js/Dictionary.cpp:
(WebCore::Dictionary::getOwnPropertiesAsStringHashMap):
(WebCore::Dictionary::getOwnPropertyNames):
* bindings/js/ScriptCallStackFactory.cpp:
(WebCore::createScriptCallStack):
* dom/ContainerNode.cpp:
(WebCore::willRemoveChildren):
* dom/Range.cpp:
(WebCore::Range::processAncestorsAndTheirSiblings):
* loader/FrameLoader.cpp:
(WebCore::FrameLoader::detachChildren):
* platform/graphics/gpu/LoopBlinnPathProcessor.cpp:
(WebCore):
(WebCore::LoopBlinnPathProcessor::subdivideCurvesSlow):
* rendering/InlineTextBox.cpp:
(WebCore::InlineTextBox::paintDocumentMarkers):
* xml/XPathFunctions.cpp:
(WebCore::XPath::Function::setArguments):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147832 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d97b4ea1
    • hmuller@adobe.com's avatar
      [CSS Exclusions] Add support for the simple case of shape-margin polygonal shape-outside · 2a8e9f3b
      hmuller@adobe.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=113726

Reviewed by Dirk Schulze.

Source/WebCore:

Add support for computing the shape-outside shape-margin boundary for a polygon.
This change is similar to the one added for padding polygons, see bug 112592.
The algorithm used for computing the margin boundary is is provisional. It works
correctly for convex polygons and will work correctly for non self-intersecting
polygons whose margin boundary is the same shape as the original polygon.

Tests: fast/exclusions/shape-outside-floats/shape-outside-floats-diamond-margin-polygon.html
       fast/exclusions/shape-outside-floats/shape-outside-floats-left-margin-polygon.html
       fast/exclusions/shape-outside-floats/shape-outside-floats-right-margin-polygon.html

* rendering/ExclusionPolygon.cpp:
(WebCore::appendArc): Generalized this method a little, to deal with margin and padding connecting arcs.
(WebCore::computeShapePaddingBounds): Pass the new appendArc() parameter.
(WebCore::computeShapeMarginBounds): Compute the margin boundary for a polygon, per the constraints summarized above.
(WebCore::ExclusionPolygon::getExcludedIntervals): Use the margin polygon for this computation.

LayoutTests:

Verify that shape-outside layout works correctly on floats, when shape-margin is specifed.

* fast/exclusions/shape-outside-floats/shape-outside-floats-diamond-margin-polygon-expected.txt: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-diamond-margin-polygon.html: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-left-margin-polygon-expected.html: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-left-margin-polygon.html: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-right-margin-polygon-expected.html: Added.
* fast/exclusions/shape-outside-floats/shape-outside-floats-right-margin-polygon.html: Added.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147831 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      2a8e9f3b
    • benjamin@webkit.org's avatar
      Modernize fast/events/autoscroll.html · 191b1d5e
      benjamin@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=112508

Patch by Benjamin Poulain <bpoulain@apple.com> on 2013-04-05
Reviewed by Jer Noble.

The test fast/events/autoscroll.html was introduced in
https://bugs.webkit.org/show_bug.cgi?id=12331 (r19102) to verify the page
does not scroll when the text is selected.

It seems reasonable to test the use case with a text dump instead of a pixel test.
This should reduces the maintenance needed to keep the test useful.

* fast/events/autoscroll.html: Removed.
* fast/events/no-scroll-on-input-text-selection-expected.txt: Added.
* fast/events/no-scroll-on-input-text-selection.html: Added.
* platform/efl/fast/events/autoscroll-expected.png: Removed.
* platform/efl/fast/events/autoscroll-expected.txt: Removed.
* platform/gtk/fast/events/autoscroll-expected.png: Removed.
* platform/gtk/fast/events/autoscroll-expected.txt: Removed.
* platform/mac-wk2/TestExpectations:
* platform/mac/fast/events/autoscroll-expected.png: Removed.
* platform/mac/fast/events/autoscroll-expected.txt: Removed.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147830 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      191b1d5e
    • aestes@apple.com's avatar
      Returning NULL from willSendRequest should cancel a load from the memory cache · 0c8e191b
      aestes@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=114075

Reviewed by Darin Adler.

Source/WebCore:

When a resource is loaded from the memory cache and the client does not
implement the didLoadResourceFromMemoryCache delegate method, WebKit
synthesizes the typical sequence of resource load callbacks. One of
these is willSendRequest, which gives the client the opportunity to
modify the request. We should respect these modifications.

Handling any arbitrary modification is difficult given where in the
loading process we check the memory cache (see <http://webkit.org/b/113251>),
but we can handle the common case where the client cancels the load by
returning a NULL request.

Test: fast/loader/willsendrequest-returns-null-for-memory-cache-load.html

* loader/FrameLoader.cpp:
(WebCore::FrameLoader::willLoadMediaElementURL): Passed request to sendRemainingDelegateMessages.
(WebCore::FrameLoader::commitProvisionalLoad): Ditto.
(WebCore::FrameLoader::loadResourceSynchronously): Ditto.
(WebCore::FrameLoader::loadedResourceFromMemoryCache): Added an out
parameter to pass back a request potentially modified by
requestFromDelegate, which is also passed to sendRemainingDelegateMessages.
* loader/FrameLoader.h:
* loader/ResourceLoadNotifier.cpp:
(WebCore::ResourceLoadNotifier::dispatchDidFailLoading): Added a
function to dispatch didFailLoading and to call the correct InspectorInstrumentation.
(WebCore::ResourceLoadNotifier::sendRemainingDelegateMessages): If the
request is NULL, call dispatchDidFailLoading and return. This matches
the delegate callback sequence of a non memory cache load that is
cancelled by willSendRequest.
* loader/ResourceLoadNotifier.h:
* loader/cache/CachedResourceLoader.cpp:
(WebCore::CachedResourceLoader::shouldContinueAfterNotifyingLoadedFromMemoryCache):
Return false if the memory cache load is cancelled by the client.
(WebCore::CachedResourceLoader::requestResource): Return early if
shouldContinueAfterNotifyingLoadedFromMemoryCache is false.

LayoutTests:

* fast/loader/resources/cached-image.html: Added.
* fast/loader/willsendrequest-returns-null-for-memory-cache-load-expected.txt: Added.
* fast/loader/willsendrequest-returns-null-for-memory-cache-load.html: Added.
* platform/wk2/TestExpectations: Expect the test to fail in WKTR due to <http://webkit.org/b/114074>.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147829 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      0c8e191b
    • benjamin@webkit.org's avatar
      Second attempt to fix the Windows bot · 9ae791c1
      benjamin@webkit.org authored 
      Unreviewed.

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreExports.def:
* JavaScriptCore.vcxproj/JavaScriptCoreExportGenerator/JavaScriptCoreExports.def.in:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147828 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      9ae791c1
    • benjamin@webkit.org's avatar
      Attempt to fix the Windows bot · d483d10a
      benjamin@webkit.org authored 
      Unreviewed.

Patch by Benjamin Poulain <bpoulain@apple.com> on 2013-04-05

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreExports.def:
* JavaScriptCore.vcxproj/JavaScriptCoreExportGenerator/JavaScriptCoreExports.def.in:
r147825 removed the symbol for nullptr_t. Add it back.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147827 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d483d10a
    • rniwa@webkit.org's avatar
      Re-enable testing on the commit queue · be641bf5
      rniwa@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=114078

Reviewed by Darin Adler.

Re-enabled the feature and the associated tests.

* Scripts/webkitpy/tool/bot/commitqueuetask.py:
(CommitQueueTask.run):
* Scripts/webkitpy/tool/bot/commitqueuetask_unittest.py:
(CommitQueueTaskTest._run_through_task):
* Scripts/webkitpy/tool/commands/queues_unittest.py:
(CommitQueueTest._mock_test_result):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147826 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      be641bf5
    • roger_fong@apple.com's avatar
      Build fix. · bfd5b21c
      roger_fong@apple.com authored 
      * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCoreExports.def:
* JavaScriptCore.vcxproj/JavaScriptCoreExportGenerator/JavaScriptCoreExports.def.in:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147825 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      bfd5b21c
    • cfleizach@apple.com's avatar
      AX: VoiceOver can't press on items · c0bb1500
      cfleizach@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=114071

Reviewed by Tim Horton.

Source/WebCore:

The AXScrollToVisibleAction should not be used on earlier Mac versions because
it confuses VoiceOver.

Updated existing tests.

* accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(-[WebAccessibilityObjectWrapper accessibilityActionNames]):

LayoutTests:

* platform/mac-future/accessibility: Added.
* platform/mac-future/accessibility/press-action-is-first-expected.txt: Added.
* platform/mac/accessibility/press-action-is-first-expected.txt:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147824 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c0bb1500
    • commit-queue@webkit.org's avatar
      [WK2] --no-tiled-backing-store build fails because FloatPoint is not declared · 478b414b
      commit-queue@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=113618

Patch by Ed Bartosh <bartosh@gmail.com> on 2013-04-05
Reviewed by Simon Fraser.

* WebProcess/Plugins/Plugin.h: Forward declared FloatPoint

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147823 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      478b414b
    • andersca@apple.com's avatar
      Stop building Objective-C bindings for the SVG DOM · 8b914b1d
      andersca@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=114076

Reviewed by Geoffrey Garen.

The DOM SVG Objective-C bindings have never been exposed as API, and not building them cuts WebCore build by 10%.

Source/WebCore:

* Configurations/FeatureDefines.xcconfig:
* Configurations/WebCore.xcconfig:
* WebCore.exp.in:
* WebCore.xcodeproj/project.pbxproj:
* bindings/objc/DOM.mm:
(WebCore::createElementClassMap):
(kitClass):
* bindings/objc/DOMCSS.mm:
(kitClass):
* bindings/objc/DOMEvents.mm:
* bindings/objc/DOMSVG.h: Removed.
* bindings/objc/DOMSVGException.h: Removed.
* bindings/objc/DOMSVGPathSegInternal.mm: Removed.
* bindings/objc/ExceptionHandlers.mm:
* html/HTMLEmbedElement.idl:
* html/HTMLFrameElement.idl:
* html/HTMLIFrameElement.idl:
* html/HTMLObjectElement.idl:

Source/WebKit/mac:

* Configurations/FeatureDefines.xcconfig:
* MigrateHeaders.make:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147822 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      8b914b1d
    • oliver@apple.com's avatar
      Build fix. · 1caa93af
      oliver@apple.com authored 
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147820 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      1caa93af
    • benjamin@webkit.org's avatar
      Change css1/basic/comments-expected.html to be a ref-test · fc3d5a36
      benjamin@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=113996

Patch by Benjamin Poulain <bpoulain@apple.com> on 2013-04-05
Reviewed by Jer Noble.

The test css1/basic/comments.html verifies the handling
of comments inside a CSS declaration.

The test is changed to a ref-test by simply defining the style
directly on each element on the reference.

* css1/basic/comments-expected.html: Added.
* platform/efl/css1/basic/comments-expected.png: Removed.
* platform/efl/css1/basic/comments-expected.txt: Removed.
* platform/gtk/css1/basic/comments-expected.txt: Removed.
* platform/mac/css1/basic/comments-expected.png: Removed.
* platform/mac/css1/basic/comments-expected.txt: Removed.
* platform/qt/css1/basic/comments-expected.png: Removed.
* platform/qt/css1/basic/comments-expected.txt: Removed.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147819 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      fc3d5a36
    • oliver@apple.com's avatar
      Unify the many and varied stack trace mechanisms, and make the result sane. · 3980d396
      oliver@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=114072

Reviewed by Filip Pizlo.

Source/JavaScriptCore:

Makes JSC::StackFrame record the bytecode offset and other necessary data
rather than requiring us to perform eager evaluation of the line number, etc.
Then remove most of the users of retrieveLastCaller, as most of them were
using it to create a stack trace in a fairly incomplete and inefficient way.

StackFrame now also has a couple of helpers to get the line and column info.

* API/JSContextRef.cpp:
(JSContextCreateBacktrace):
* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::emitDebugHook):
* interpreter/Interpreter.cpp:
(JSC):
(JSC::Interpreter::dumpRegisters):
(JSC::Interpreter::unwindCallFrame):
(JSC::getBytecodeOffsetForCallFrame):
(JSC::getCallerInfo):
(JSC::StackFrame::line):
(JSC::StackFrame::column):
(JSC::StackFrame::expressionInfo):
(JSC::StackFrame::toString):
(JSC::Interpreter::getStackTrace):
(JSC::Interpreter::addStackTraceIfNecessary):
(JSC::Interpreter::retrieveCallerFromVMCode):
* interpreter/Interpreter.h:
(StackFrame):
(Interpreter):
* runtime/Error.cpp:
(JSC::throwError):
* runtime/JSGlobalData.h:
(JSC):
(JSGlobalData):
* runtime/JSGlobalObject.cpp:
(JSC::DynamicGlobalObjectScope::DynamicGlobalObjectScope):

Source/WebCore:

Now that we've fleshed out the StackFrames from Interpreter::getStackTrace
WebCore can just ask us for a stack trace rather than implementing its own
stack walking.

* bindings/js/ScriptCallStackFactory.cpp:
(WebCore::createScriptCallStack):
* inspector/ScriptCallFrame.cpp:
(WebCore::ScriptCallFrame::isEqual):
* inspector/ScriptCallFrame.h:
(ScriptCallFrame):
(WebCore::ScriptCallFrame::columnNumber):

Tools:

The commandline jsc executable no longer requires arguments, so
I've made run-jsc work without them.

* Scripts/run-jsc:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147818 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      3980d396
    • mhahnenberg@apple.com's avatar
      tryCacheGetByID sets StructureStubInfo accessType to an incorrect value · d0be74de
      mhahnenberg@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=114068

Reviewed by Geoffrey Garen.

In the case where we have a non-Value cacheable property, we set the StructureStubInfo accessType to 
get_by_id_self, but then we don't patch self and instead patch in a get_by_id_self_fail. This leads to 
incorrect profiling data so when the DFG compiles the function, it uses a GetByOffset rather than a GetById, 
which leads to loading a GetterSetter directly out of an object.

Source/JavaScriptCore: 

* jit/JITStubs.cpp:
(JSC::tryCacheGetByID):
(JSC::DEFINE_STUB_FUNCTION):

LayoutTests: 

* fast/js/jit-set-profiling-access-type-only-for-get-by-id-self-expected.txt: Added.
* fast/js/jit-set-profiling-access-type-only-for-get-by-id-self.html: Added.
* fast/js/script-tests/jit-set-profiling-access-type-only-for-get-by-id-self.js: Added.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147816 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d0be74de
    • kov@webkit.org's avatar
      Translation of WebKitGTK+ for Malayalam (ml) · 94a61e69
      kov@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=113280

Patch by Ani Peter <peter.ani@gmail.com> on 2013-04-05
Reviewed by Gustavo Noronha Silva.

* ml.po: Added.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147814 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      94a61e69
    • kov@webkit.org's avatar
      webkit translations for gujarati · dbf4ee36
      kov@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=113190

Patch by Sweta Kothari <swkothar@redhat.com> on 2013-04-05
Reviewed by Gustavo Noronha Silva.

* gu.po:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147812 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      dbf4ee36
    • kov@webkit.org's avatar
      [l10n] [hi] Updated WebKit Translation in Hindi [hi] language · 03c636ea
      kov@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=113180

Patch by Rajesh Ranjan <rajeshkajha@yahoo.com> on 2013-04-05
Reviewed by Gustavo Noronha Silva.

* hi.po:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147811 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      03c636ea
    • commit-queue@webkit.org's avatar
      [BlackBerry] FatFingers Don't check the shadow tree for all elements. · ed89a9c6
      commit-queue@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=114066

Patch by Tiancheng Jiang <tijiang@rim.com> on 2013-04-05
Reviewed by Rob Buis.

BlackBerry PR 263078
Internally reviewed by Genevieve Mak.

No need to check shadow tree since sliders use touch events.

* WebKitSupport/FatFingers.cpp:
(BlackBerry::WebKit::FatFingers::isElementClickable):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147807 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      ed89a9c6
    • bdakin@apple.com's avatar
      Setting a header or footer should cause a relayout · fc6af6b0
      bdakin@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=114070

Reviewed by Sam Weinig.

Mark the RenderView as needing layout in order to adjust everything for the new 
header/footer.

* page/FrameView.cpp:
(WebCore::FrameView::setHeaderHeight):
(WebCore::FrameView::setFooterHeight):



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147806 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      fc6af6b0
    • cfleizach@apple.com's avatar
      AX: Make SVG Group containers accessible elements · b2b0b4d9
      cfleizach@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=113939

Reviewed by Simon Fraser.

Source/WebCore: 

SVG group containers are currently not exposed in the AX tree. 
When an SVG element has a <title> child, that is not being used for the AX label.
When hit-testing, from Accessibility, SVG groups are never returned.

All these issues are fixed for Mac and iOS.
There are two tests because right now iOS skips all the top-level accessibility tests
due to fundamental architectural differences.

Tests: accessibility/svg-group-element-with-title.html
       platform/iphone-simulator/accessibility/svg-group-element-with-title.html

* accessibility/AccessibilityNodeObject.cpp:
(WebCore::AccessibilityNodeObject::alternativeText):
(WebCore::AccessibilityNodeObject::accessibilityDescription):
* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::accessibilityHitTest):
(WebCore::AccessibilityRenderObject::determineAccessibilityRole):
* accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
(-[WebAccessibilityObjectWrapper isSVGGroupElement]):
(-[WebAccessibilityObjectWrapper determineIsAccessibilityElement]):
(-[WebAccessibilityObjectWrapper accessibilityClickPoint]):
* rendering/HitTestRequest.h:
    Add a new option, AccessibilityHitTest, that will allow us to override normal behavior.
* rendering/svg/RenderSVGContainer.cpp:
(WebCore::RenderSVGContainer::nodeAtFloatPoint):
   Allow a SVG group to be the node at a point, for accessibility hit tests.

Tools: 

Implement axController.elementAtPoint for iOS.
Implement a correct version of clickPoint for iOS.

* DumpRenderTree/ios/AccessibilityControllerIOS.mm:
(AccessibilityController::elementAtPoint):
* DumpRenderTree/ios/AccessibilityUIElementIOS.mm:
(AccessibilityUIElement::clickPointX):
(AccessibilityUIElement::clickPointY):

LayoutTests: 

* accessibility/svg-group-element-with-title-expected.txt: Added.
* accessibility/svg-group-element-with-title.html: Added.
* platform/iphone-simulator/accessibility/svg-group-element-with-title-expected.txt: Added.
* platform/iphone-simulator/accessibility/svg-group-element-with-title.html: Added.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147802 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      b2b0b4d9
    • cfleizach@apple.com's avatar
      WebSpeech: event.elapsedTime is always 1364484608 · 16399b98
      cfleizach@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=114045

Reviewed by Tim Horton.

Source/WebCore:

We were calculating elapsed time incorrectly.

Test: platform/mac/fast/speechsynthesis/speech-synthesis-elapsed-time.html

* Modules/speech/SpeechSynthesis.cpp:
(WebCore::SpeechSynthesis::fireEvent):

LayoutTests:

* platform/mac/fast/speechsynthesis/speech-synthesis-elapsed-time-expected.txt: Added.
* platform/mac/fast/speechsynthesis/speech-synthesis-elapsed-time.html: Added.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147801 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      16399b98
    • cfleizach@apple.com's avatar
      WebSpeech: crash at WebCore::SpeechSynthesis::speak · e73c07ab
      cfleizach@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=113937

Reviewed by Tim Horton.

Source/WebCore:

Protect against the case when invalid data is passed in.

Test: platform/mac/fast/speechsynthesis/speech-synthesis-crash-on-bad-utterance.html

* Modules/speech/SpeechSynthesis.cpp:
(WebCore::SpeechSynthesis::speak):
* Modules/speech/SpeechSynthesisUtterance.cpp:
(WebCore::SpeechSynthesisUtterance::setVoice):

LayoutTests:

* platform/mac/fast/speechsynthesis/speech-synthesis-crash-on-bad-utterance-expected.txt: Added.
* platform/mac/fast/speechsynthesis/speech-synthesis-crash-on-bad-utterance.html: Added.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147800 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      e73c07ab
    • commit-queue@webkit.org's avatar
      [GTK] Change from PLATFORM(GTK) to USE(GLIB) in WTF/CurrentTime.cpp · ee2dcfba
      commit-queue@webkit.org authored 
      https://bugs.webkit.org/show_bug.cgi?id=114061

Patch by Jesus Sanchez-Palencia <jesus.palencia@openbossa.org> on 2013-04-05
Reviewed by Gustavo Noronha Silva.

Now other ports using GLib can use the correct implementation
and not the fallback one without adding any other PLATFORM guards
here.

* wtf/CurrentTime.cpp:
(WTF):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147799 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      ee2dcfba
    • fpizlo@apple.com's avatar
      If CallFrame::trueCallFrame() knows that it's about to read garbage instead of... · e62f04d9
      fpizlo@apple.com authored 
      If CallFrame::trueCallFrame() knows that it's about to read garbage instead of a valid CodeOrigin/InlineCallFrame, then it should give up and return 0 and all callers should be robust against this
https://bugs.webkit.org/show_bug.cgi?id=114062

Reviewed by Oliver Hunt.

* bytecode/CodeBlock.h:
(JSC::CodeBlock::canGetCodeOrigin):
(CodeBlock):
* interpreter/CallFrame.cpp:
(JSC::CallFrame::trueCallFrame):
* interpreter/Interpreter.cpp:
(JSC::Interpreter::getStackTrace):



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147798 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      e62f04d9
    • antti@apple.com's avatar
      Throttle compositing layer flushes during page loading · a6775c18
      antti@apple.com authored 
      https://bugs.webkit.org/show_bug.cgi?id=113786

Reviewed by Simon Fraser.
        
Page content can change rapidly during page loading triggering excessive layer flushes and repainting. We should avoid this unnecessary work.
        
This patch reduces layer flushes (and painting) during loading by 50-70% on many popular pages.

* loader/FrameLoader.cpp:
(WebCore::FrameLoader::loadProgressingStatusChanged):        
* loader/FrameLoader.h:
* loader/ProgressTracker.cpp:
(WebCore::ProgressTracker::ProgressTracker):
(WebCore::ProgressTracker::reset):
(WebCore::ProgressTracker::progressStarted):
(WebCore::ProgressTracker::finalProgressComplete):
(WebCore::ProgressTracker::isLoadProgressing):
(WebCore::ProgressTracker::progressHeartbeatTimerFired):
* loader/ProgressTracker.h:
        
    Track if the document load is progressing. This is done with a heartbeat timer that checks every 100ms if we have received more than 1k of data.
    If four heartbeats pass without progress then we consider the load stalled.

* page/FrameView.cpp:
(WebCore::FrameView::resetDeferredRepaintDelay):
        
    Disable throttling temporary on user interaction so the page stays as responsive as possible even during loading.

(WebCore::FrameView::updateLayerFlushThrottling):
        
    Enable throttling when the load is progressing, disable otherwise.

* page/FrameView.h:
* platform/graphics/GraphicsLayer.h:
(WebCore::GraphicsLayer::canThrottleLayerFlush):
* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::platformCALayerDidCreateTiles):
(WebCore::GraphicsLayerCA::canThrottleLayerFlush):
        
    Don't throttle if new tiles have been added by the tile controller. They may have stale content and need to be flushed immediately.

(WebCore::GraphicsLayerCA::noteLayerPropertyChanged):
        
    Set the new TilesAdded change flag.

* platform/graphics/ca/GraphicsLayerCA.h:
* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::notifyFlushRequired):
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::RenderLayerCompositor):
(WebCore::RenderLayerCompositor::notifyFlushRequired):
(WebCore::RenderLayerCompositor::scheduleLayerFlushNow):
        
    Factor the actual flush scheduling to private function.

(WebCore::RenderLayerCompositor::scheduleLayerFlush):
        
    Mark the compositor for flush and return without flushing if the flushes are currently being throttled.

(WebCore::RenderLayerCompositor::flushPendingLayerChanges):
        
    After a flush, start the throtting timer (currently 0.5s) coalescing the subsequent flushes.

(WebCore::RenderLayerCompositor::didChangeVisibleRect):
        
    Do immediately flush if needed.

(WebCore::RenderLayerCompositor::setLayerFlushThrottlingEnabled):
        
    Flush immediately if disabled.

(WebCore::RenderLayerCompositor::disableLayerFlushThrottlingTemporarilyForInteraction):
(WebCore::RenderLayerCompositor::isThrottlingLayerFlushes):
(WebCore::RenderLayerCompositor::startLayerFlushTimerIfNeeded):
(WebCore::RenderLayerCompositor::layerFlushTimerFired):
        
    Flush when the timer fires timer.

* rendering/RenderLayerCompositor.h:



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@147797 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      a6775c18