<rdar://problem/9028929> REGRESSION (r75897): Scaling applied twice to an iframe with a transformed ancestor

Reviewed by Maciej Stachowiak.

Source/WebCore:

Test: fast/frames/iframe-scale-applied-twice.html

* page/FrameView.cpp:
(WebCore::FrameView::create): Set the initial bounds of the view to match the
frame size.
* platform/ScrollView.cpp:
(WebCore::ScrollView::visibleContentRect): Based on bounds, not frame size.
(WebCore::ScrollView::updateScrollbars): Ditto.
(WebCore::ScrollView::setFrameRect): Moved code that really handles bounds size
change to setBoundsSize().
(WebCore::ScrollView::setBoundsSize): Added.
(WebCore::ScrollView::setInitialBoundsSize): Added. Sets the bounds size but does
not update anything.
(WebCore::ScrollView::frameRectsChanged): Based on bounds, not frame size.
(WebCore::ScrollView::scrollbarCornerPresent): Ditto.
* platform/ScrollView.h:
(WebCore::ScrollView::boundsSize): Added this getter.
* platform/Widget.h:
(WebCore::Widget::resize): Set the bounds size to the frame size.
* platform/mac/ScrollbarThemeMac.mm:
(WebCore::ScrollbarThemeMac::paint): Fixed an error in the indirect drawing code
path where the buffer rect was resized to capture only the damaged part, but was
still drawn in the original location.

LayoutTests:

* fast/frames/iframe-scale-applied-twice-expected.txt: Added.
* fast/frames/iframe-scale-applied-twice.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79167 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-17  Ojan Vafai  <ojan@chromium.org>

        Reviewed by Adam Barth.

        use shift+j/k to focus next/previous line
        https://bugs.webkit.org/show_bug.cgi?id=54723

        Tested on a large Hyatt change to confirm it performed fine.
        (https://bugs.webkit.org/attachment.cgi?id=82533&action=review)

        * code-review.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79166 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-17  Ojan Vafai  <ojan@chromium.org>

        Reviewed by Adam Barth.

        fix escape handling in code review tool
        https://bugs.webkit.org/show_bug.cgi?id=54722

        This broke with http://trac.webkit.org/changeset/78948.
        Apparently jquery events get a wrapped object that
        does not expose keyIdentifier.

        * code-review.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79165 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-17  Ojan Vafai  <ojan@chromium.org>

        Reviewed by Adam Barth.

        create comments when hitting enter with a focused diff
        https://bugs.webkit.org/show_bug.cgi?id=54719

        * code-review.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79164 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-17  Ojan Vafai  <ojan@chromium.org>

        Reviewed by Adam Barth.

        use actual browser focus in the code review tool
        https://bugs.webkit.org/show_bug.cgi?id=54726

        This makes keyboard handling play better with focusable element
        (i.e. links/textareas/buttons). Also, in theory, this makes
        the review tool more amenable to screen readers (I think).

        * PrettyPatch/PrettyPatch.rb:
        * code-review.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79163 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-20  Ami Fischman  <fischman@chromium.org>

        Reviewed by Kent Tamura.

        Remove flakiness from media/video-poster.html by using event callbacks
        instead of 100ms timeouts and hoping that states have transitioned in
        time (!).
        Removed seemingly-outdated BUG_DRT annotation from test_expectation.

        https://bugs.webkit.org/show_bug.cgi?id=54685
        http://crbug.com/60845

        * media/video-poster.html:
        * platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79161 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-21 Benjamin Kalman  <kalman@chromium.org>

        Reviewed by Ryosuke Niwa.

        Make editing/selection/extend-selection tests more human readable and hand-runnable
        https://bugs.webkit.org/show_bug.cgi?id=54712

        Three changes to achieve this:
          - Column-align the "right:", "left:", "forward:", and "backward:" labels in the output.
          - Add buttons to switch between LTR and RTL (the test will switch everything to RTL but LTR is useful).
          - Fix typo causing failing tests to have a js error (rather than layout error).

        Expected results need to change due to column alignment.

        * editing/selection/extend-selection-character-expected.txt:
        * editing/selection/extend-selection-word-expected.txt:
        * editing/selection/resources/extend-selection.js:
        (logMismatchingPositions):
        (runSelectionTestsWithGranularity):
        (getTestNodeContainer.createButton.button.onmousedown):
        (getTestNodeContainer.createButton):
        (getTestNodeContainer.createButtonToSwitchDirection):
        (getTestNodeContainer):
        (createNode):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79160 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Eric Seidel.

        Tighten up access permissions by using libxslt API
        https://bugs.webkit.org/show_bug.cgi?id=52688
        <rdar://problem/8909191>

        * xml/XSLTProcessorLibxslt.cpp: (WebCore::XSLTProcessor::transformToString): We are only
        interested in a string result, so let libxslt know about that.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79159 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-20  Sheriff Bot  <webkit.review.bot@gmail.com>

        Unreviewed, rolling out r79104.
        http://trac.webkit.org/changeset/79104
        https://bugs.webkit.org/show_bug.cgi?id=54835

        Might have caused flaky canaries (Requested by tonyg-cr on
        #webkit).

        * html/parser/HTMLDocumentParser.cpp:
        (WebCore::HTMLDocumentParser::canTakeNextToken):
        (WebCore::HTMLDocumentParser::pumpTokenizer):
        * html/parser/HTMLParserScheduler.cpp:
        (WebCore::isLayoutTimerActive):
        * html/parser/HTMLParserScheduler.h:
        (WebCore::HTMLParserScheduler::checkForYieldBeforeToken):
        * page/FrameView.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79158 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-20  Anders Carlsson  <andersca@apple.com>

        Reviewed by Maciej Stachowiak.

        Crash when a plug-in requests a javascript: url that destroys the plug-in
        https://bugs.webkit.org/show_bug.cgi?id=54837
        <rdar://problem/9005475>

        Add new plug-in test.

        * DumpRenderTree/DumpRenderTree.gypi:
        * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
        * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
        (PluginTest::NPN_GetURL):
        * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
        * DumpRenderTree/TestNetscapePlugIn/Tests/GetURLWithJavaScriptURLDestroyingPlugin.cpp: Added.
        (GetURLWithJavaScriptURLDestroyingPlugin::GetURLWithJavaScriptURLDestroyingPlugin):
        (GetURLWithJavaScriptURLDestroyingPlugin::NPP_New):
        * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
        * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
2011-02-20  Anders Carlsson  <andersca@apple.com>

        Reviewed by Maciej Stachowiak.

        Crash when a plug-in requests a javascript: url that destroys the plug-in
        https://bugs.webkit.org/show_bug.cgi?id=54837
        <rdar://problem/9005475>

        Add test.

        * plugins/get-url-with-javascript-destroying-plugin-expected.txt: Added.
        * plugins/get-url-with-javascript-destroying-plugin.html: Added.
2011-02-20  Anders Carlsson  <andersca@apple.com>

        Reviewed by Maciej Stachowiak.

        Crash when a plug-in requests a javascript: url that destroys the plug-in
        https://bugs.webkit.org/show_bug.cgi?id=54837
        <rdar://problem/9005475>

        * WebProcess/Plugins/PluginProxy.cpp:
        (WebKit::PluginProxy::destroy):
        Null out m_pluginController.

        * WebProcess/Plugins/PluginView.cpp:
        (WebKit::PluginView::performJavaScriptURLRequest):
        Don't access the frame through m_pluginElement since it will be nulled out
        when the plug-in is destroyed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79157 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-20  Anders Carlsson  <andersca@apple.com>

        Reviewed by Sam Weinig.

        Fix another crash when a plug-in is destroyed when evaluating JavaScript
        https://bugs.webkit.org/show_bug.cgi?id=54834
        <rdar://problem/9013384>

        This contains three separate fixes:

        - In unprotectPluginFromDestruction we can't just destroy the plug-in once
          the count reaches zero, because the plug-in might still have code executing
          on the stack. To fix this we use a zero-delay timer to defer destruction
          of the plug-in.

        - Trying to get the NPObject for the window using NPN_GetValue would return
          NPERR_NO_ERROR, even if the returned NPObject was null. Fix this to return
          NPERR_GENERIC_ERROR instead.

        - Protect the plug-in from destruction anytime an NPAPI call that sends a
          synchronous IPC message is made.

        * PluginProcess/PluginControllerProxy.cpp:
        (WebKit::PluginControllerProxy::PluginControllerProxy):
        (WebKit::PluginControllerProxy::destroy):
        (WebKit::PluginControllerProxy::unprotectPluginFromDestruction):
        * PluginProcess/PluginControllerProxy.h:
        * WebProcess/Plugins/Netscape/NetscapeBrowserFuncs.cpp:
        (WebKit::PluginDestructionProtector::PluginDestructionProtector):
        (WebKit::NPN_GetValue):
        (WebKit::NPN_Evaluate):
        (WebKit::NPN_GetProperty):
        (WebKit::NPN_SetProperty):
        (WebKit::NPN_RemoveProperty):
        (WebKit::NPN_HasProperty):
        (WebKit::NPN_HasMethod):
        (WebKit::NPN_Enumerate):
        (WebKit::NPN_Construct):
        (WebKit::NPN_GetValueForURL):
        (WebKit::NPN_SetValueForURL):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79156 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-20  Dirk Schulze  <krit@webkit.org>

        Reviewed by Nikolas Zimmermann.

        SVG animation - analyze attribute type for animation
        https://bugs.webkit.org/show_bug.cgi?id=54542

        Check dynamic update of attributeName. The test fails at the moment, because of some wrong
        internal mapping. Tested update manually with gdb. Working on a fix to get this test pass
        as well.

        * svg/animations/animate-dynamic-update-attributeName-expected.txt: Added.
        * svg/animations/animate-dynamic-update-attributeName.html: Added.
        * svg/animations/script-tests/animate-dynamic-update-attributeName.js: Added.
        (sample1):
        (sample2):
        (sample3):
        (executeTest):
2011-02-20  Dirk Schulze  <krit@webkit.org>

        Reviewed by Nikolas Zimmermann.

        SVG animation - analyze attribute type for animation
        https://bugs.webkit.org/show_bug.cgi?id=54542

        Analyze animated attribute type to determine the kind of animation. Removed enum PropertyType and replace it
        with AnimatedAttributeType instead. More cleanup of the animation code.
        Added missing fillAttributeToPropertyTypeMap() for SVGGElement.

        Use QualifiedName instead of AtomicStrings for attribute names. This makes it easier to handle animated attributes
        from a different namespace like xlink:href. The rest of the SVG code is using QualifiedNames as well.

        No change of functionality on the animation code. The changes for the <g>-element affect animateTransform. Sadly
        animateTransform is not supported by our SVG animation test script at the moment. Can still be tested manually
        with the W3C test suite in trunk.

        The change on SVGUseElement gets covered by animate-elem-39-t.svg

        Test: svg/animations/animate-dynamic-update-attributeName.html

        * svg/SVGAnimateElement.cpp:
        (WebCore::SVGAnimateElement::SVGAnimateElement):
        (WebCore::SVGAnimateElement::hasValidAttributeType):
        (WebCore::SVGAnimateElement::determineAnimatedAttributeType):
        (WebCore::SVGAnimateElement::calculateAnimatedValue):
        (WebCore::inheritsFromProperty):
        (WebCore::SVGAnimateElement::calculateFromAndToValues):
        (WebCore::SVGAnimateElement::calculateFromAndByValues):
        (WebCore::SVGAnimateElement::resetToBaseValue):
        (WebCore::SVGAnimateElement::applyResultsToTarget):
        (WebCore::SVGAnimateElement::calculateDistance):
        * svg/SVGAnimateElement.h:
        * svg/SVGAnimateMotionElement.cpp:
        (WebCore::SVGAnimateMotionElement::hasValidAttributeType):
        (WebCore::parsePoint):
        (WebCore::SVGAnimateMotionElement::resetToBaseValue):
        (WebCore::SVGAnimateMotionElement::calculateAnimatedValue):
        (WebCore::SVGAnimateMotionElement::calculateDistance):
        * svg/SVGAnimateMotionElement.h:
        * svg/SVGAnimateTransformElement.cpp:
        (WebCore::SVGAnimateTransformElement::hasValidAttributeType):
        (WebCore::SVGAnimateTransformElement::determineAnimatedAttributeType):
        (WebCore::SVGAnimateTransformElement::resetToBaseValue):
        (WebCore::SVGAnimateTransformElement::calculateAnimatedValue):
        (WebCore::SVGAnimateTransformElement::calculateFromAndByValues):
        (WebCore::SVGAnimateTransformElement::applyResultsToTarget):
        (WebCore::SVGAnimateTransformElement::calculateDistance):
        * svg/SVGAnimateTransformElement.h:
        * svg/SVGAnimationElement.cpp:
        (WebCore::parseKeyTimes):
        (WebCore::parseKeySplines):
        (WebCore::SVGAnimationElement::isTargetAttributeCSSProperty):
        (WebCore::SVGAnimationElement::setTargetAttributeAnimatedValue):
        (WebCore::SVGAnimationElement::calculateKeyTimesForCalcModePaced):
        (WebCore::solveEpsilon):
        (WebCore::SVGAnimationElement::calculatePercentFromKeyPoints):
        (WebCore::SVGAnimationElement::currentValuesFromKeyPoints):
        (WebCore::SVGAnimationElement::currentValuesForValuesAnimation):
        (WebCore::SVGAnimationElement::startedActiveInterval):
        * svg/SVGAnimationElement.h:
        * svg/SVGFilterElement.cpp:
        (WebCore::SVGFilterElement::fillAttributeToPropertyTypeMap):
        * svg/SVGGElement.cpp:
        (WebCore::SVGGElement::attributeToPropertyTypeMap):
        (WebCore::SVGGElement::fillAttributeToPropertyTypeMap): Added to fill the animated attribute map for this element.
        * svg/SVGGElement.h:
        * svg/SVGUseElement.cpp:
        (WebCore::SVGUseElement::fillAttributeToPropertyTypeMap): Changed attribute type to AnimatedString for xlink:href.
        * svg/animation/SMILTimeContainer.cpp:
        (WebCore::SMILTimeContainer::baseValueFor):
        (WebCore::SMILTimeContainer::updateAnimations):
        * svg/animation/SMILTimeContainer.h:
        * svg/animation/SVGSMILElement.cpp:
        (WebCore::SVGSMILElement::SVGSMILElement):
        (WebCore::constructQualifiedName): Create a QualifiedName from a attribute name.
        (WebCore::SVGSMILElement::insertedIntoDocument): Update local varaible of attribute name.
        (WebCore::SVGSMILElement::removedFromDocument): Ditto.
        (WebCore::SVGSMILElement::attributeChanged): Ditto.
        * svg/animation/SVGSMILElement.h:
        (WebCore::SVGSMILElement::attributeName):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79155 268f45cc-cd09-0410-ab3c-d52691b4dbfc

BUILD FIX: Add missing UNUSED_PARAM() when !ENABLE(FULLSCREEN_API)

Not reviewed.

* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::requiresCompositingForFullScreen):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79154 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-20  Carlos Garcia Campos  <cgarcia@igalia.com>

        Reviewed by Martin Robinson.

        [GTK] Implement PlatformKeyboardEvent::getCurrentModifierState()
        https://bugs.webkit.org/show_bug.cgi?id=52880

        * platform/gtk/KeyEventGtk.cpp:
        (WebCore::PlatformKeyboardEvent::getCurrentModifierState):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79151 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-20  Carlos Garcia Campos  <cgarcia@igalia.com>

        Reviewed by Martin Robinson.

        [GTK] Bump GTK+3 dependency to 3.0
        https://bugs.webkit.org/show_bug.cgi?id=54394

        * configure.ac:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79150 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Set the svn:mime-type property of some PNG files to image/png

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79149 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Andreas Kling  <kling@webkit.org>

        Reviewed by Oliver Hunt.

        DragController should use Color::serialized() when passing colors to CSSStyleDeclaration
        https://bugs.webkit.org/show_bug.cgi?id=54545

        * page/DragController.cpp:
        (WebCore::DragController::concludeEditDrag):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79145 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Laszlo Gombos  <laszlo.1.gombos@nokia.com>

        Unreviewed build fix for Symbian.

        [Symbian] Revert the removal of linking
        against hal after r79126.

        Dependency on the hal library can not be removed
        as it is still used (e.g. in MarkStackSymbian.cpp).

        * JavaScriptCore.pri:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79144 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Lukasz Slachciak  <l.slachciak@samsung.com>

        Reviewed by Andreas Kling.

        [EFL] Eliminated error when .webkit directory doesn't exist.
        https://bugs.webkit.org/show_bug.cgi?id=54813

        * ewk/ewk_main.cpp:
        (_ewk_init_body):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79143 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Interpreter build fix.

* runtime/ArrayConstructor.cpp:
* runtime/BooleanPrototype.cpp:
* runtime/DateConstructor.cpp:
* runtime/ErrorPrototype.cpp:
* runtime/FunctionPrototype.cpp:
* runtime/Lookup.cpp:
* runtime/NumberPrototype.cpp:
* runtime/ObjectConstructor.cpp:
* runtime/ObjectPrototype.cpp:
* runtime/RegExpPrototype.cpp:
* runtime/StringConstructor.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79142 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add phase in addition to momentumPhase to platform wheel events
Part of <rdar://problem/8945362>

Reviewed by Anders Carlsson.

Rename existing phase to momentumPhase.

../WebCore:

* platform/PlatformWheelEvent.h:
(WebCore::PlatformWheelEvent::PlatformWheelEvent):
(WebCore::PlatformWheelEvent::momentumPhase):
* platform/mac/ScrollAnimatorMac.mm:
(WebCore::ScrollAnimatorMac::handleWheelEvent):
(WebCore::ScrollAnimatorMac::smoothScrollWithEvent):
* platform/mac/WheelEventMac.mm:
(WebCore::momentumPhaseForEvent):
(WebCore::phaseForEvent):
(WebCore::PlatformWheelEvent::PlatformWheelEvent):

../WebKit2:

* Shared/WebEvent.h:
(WebKit::WebWheelEvent::momentumPhase):
* Shared/WebEventConversion.cpp:
(WebKit::WebKit2PlatformWheelEvent::WebKit2PlatformWheelEvent):
* Shared/WebWheelEvent.cpp:
(WebKit::WebWheelEvent::WebWheelEvent):
(WebKit::WebWheelEvent::encode):
(WebKit::WebWheelEvent::decode):
* Shared/mac/WebEventFactory.mm:
(WebKit::phaseForEvent):
(WebKit::momentumPhaseForEvent):
(WebKit::WebEventFactory::createWebWheelEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79141 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Qt build fix.

* bridge/runtime_method.h:
(JSC::RuntimeMethod::createStructure):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79140 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Qt build fix.

* bridge/qt/qt_instance.cpp:
* bridge/qt/qt_pixmapruntime.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79139 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Build fix!!

* JavaScriptCore.exp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79138 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Windows build fix!!

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79137 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Qt build fix.

* bridge/qt/qt_instance.cpp:
(JSC::Bindings::QtInstance::getMethod):
* bridge/qt/qt_pixmapruntime.cpp:
(JSC::Bindings::QtPixmapInstance::getMethod):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79136 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Windows build fix!

* JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79135 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Build fix!

* JavaScriptCore.exp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79134 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Anders Carlsson  <andersca@apple.com>

        Reviewed by Dan Bernstein.

        Crash when trying to take a snapshot of an uninitialized plug-in
        https://bugs.webkit.org/show_bug.cgi?id=54812

        * WebProcess/Plugins/PluginView.cpp:
        (WebKit::PluginView::notifyWidget):
        Don't try to create a snapshot of the plug-in if it hasn't been initialized.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79133 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Bug 54786 - Devirtualize JSCell::classInfo()

Reviewed by Sam Weinig.

Instead of making a virtual function call, add a pointer to the ClassInfo
onto Structure.

This removes a virtual function call, and paves the way towards removing all
the createStructure methods, and StructureFlags/AnonymousSlotCount properties
(these should be able to move onto ClassInfo).

Calls to Structure::create must now pass a pointer to the ClassInfo for the
structure. All objects now have a ClassInfo pointer, non-object cell types
still do not.

Changes are most mechanical, involving three steps:
    * Remove virtual classInfo() methods.
    * Add &s_info parameter to calls to Structure::create.
    * Rename ClassInfo static members on classes from 'info' to 's_info',
      for consistency.

Source/JavaScriptCore:

* API/JSCallbackConstructor.cpp:
* API/JSCallbackConstructor.h:
* API/JSCallbackFunction.cpp:
* API/JSCallbackFunction.h:
* API/JSCallbackObject.cpp:
* API/JSCallbackObject.h:
* API/JSCallbackObjectFunctions.h:
* API/JSObjectRef.cpp:
* API/JSValueRef.cpp:
* API/JSWeakObjectMapRefPrivate.cpp:
* JavaScriptCore.exp:
* debugger/Debugger.cpp:
* debugger/DebuggerActivation.h:
* debugger/DebuggerCallFrame.cpp:
* interpreter/Interpreter.cpp:
* jit/JITCall32_64.cpp:
* jit/JITOpcodes.cpp:
* jit/JITStubs.cpp:
* profiler/Profiler.cpp:
* runtime/Arguments.cpp:
* runtime/Arguments.h:
* runtime/ArrayConstructor.cpp:
* runtime/ArrayPrototype.cpp:
* runtime/ArrayPrototype.h:
* runtime/BooleanObject.cpp:
* runtime/BooleanObject.h:
* runtime/BooleanPrototype.cpp:
* runtime/DateConstructor.cpp:
* runtime/DateInstance.cpp:
* runtime/DateInstance.h:
* runtime/DatePrototype.cpp:
* runtime/DatePrototype.h:
* runtime/ErrorInstance.cpp:
* runtime/ErrorInstance.h:
* runtime/ErrorPrototype.cpp:
* runtime/FunctionPrototype.cpp:
* runtime/FunctionPrototype.h:
* runtime/GetterSetter.h:
* runtime/GlobalEvalFunction.h:
* runtime/InternalFunction.cpp:
* runtime/InternalFunction.h:
* runtime/JSAPIValueWrapper.h:
* runtime/JSActivation.cpp:
* runtime/JSActivation.h:
* runtime/JSArray.cpp:
* runtime/JSArray.h:
* runtime/JSByteArray.cpp:
* runtime/JSByteArray.h:
* runtime/JSCell.cpp:
* runtime/JSCell.h:
* runtime/JSFunction.cpp:
* runtime/JSFunction.h:
* runtime/JSGlobalData.cpp:
* runtime/JSGlobalObject.cpp:
* runtime/JSGlobalObject.h:
* runtime/JSNotAnObject.h:
* runtime/JSONObject.cpp:
* runtime/JSONObject.h:
* runtime/JSObject.cpp:
* runtime/JSObject.h:
* runtime/JSObjectWithGlobalObject.h:
* runtime/JSPropertyNameIterator.h:
* runtime/JSStaticScopeObject.h:
* runtime/JSString.h:
* runtime/JSVariableObject.h:
* runtime/JSWrapperObject.h:
* runtime/JSZombie.cpp:
* runtime/JSZombie.h:
* runtime/Lookup.cpp:
* runtime/MathObject.cpp:
* runtime/MathObject.h:
* runtime/NativeErrorConstructor.cpp:
* runtime/NativeErrorConstructor.h:
* runtime/NumberConstructor.cpp:
* runtime/NumberConstructor.h:
* runtime/NumberObject.cpp:
* runtime/NumberObject.h:
* runtime/NumberPrototype.cpp:
* runtime/ObjectConstructor.cpp:
* runtime/ObjectPrototype.cpp:
* runtime/RegExpConstructor.cpp:
* runtime/RegExpConstructor.h:
* runtime/RegExpObject.cpp:
* runtime/RegExpObject.h:
* runtime/RegExpPrototype.cpp:
* runtime/ScopeChain.cpp:
* runtime/StringConstructor.cpp:
* runtime/StringObject.cpp:
* runtime/StringObject.h:
* runtime/StringObjectThatMasqueradesAsUndefined.h:
* runtime/StringPrototype.cpp:
* runtime/StringPrototype.h:
* runtime/Structure.cpp:
* runtime/Structure.h:

Source/JavaScriptGlue:

* JSUtils.cpp:
(KJSValueToJSObject):
(KJSValueToCFTypeInternal):
* UserObjectImp.cpp:
* UserObjectImp.h:
(UserObjectImp::createStructure):

Source/WebCore:

* WebCore.exp.in:
* bindings/js/JSAudioConstructor.cpp:
* bindings/js/JSAudioConstructor.h:
* bindings/js/JSDOMBinding.cpp:
* bindings/js/JSDOMBinding.h:
* bindings/js/JSDOMGlobalObject.cpp:
* bindings/js/JSDOMGlobalObject.h:
* bindings/js/JSDOMWindowBase.cpp:
* bindings/js/JSDOMWindowBase.h:
* bindings/js/JSDOMWindowShell.cpp:
* bindings/js/JSDOMWindowShell.h:
* bindings/js/JSGeolocationCustom.cpp:
* bindings/js/JSImageConstructor.cpp:
* bindings/js/JSImageConstructor.h:
* bindings/js/JSImageDataCustom.cpp:
* bindings/js/JSOptionConstructor.cpp:
* bindings/js/JSOptionConstructor.h:
* bindings/js/JSWorkerContextBase.cpp:
* bindings/js/JSWorkerContextBase.h:
* bindings/js/SerializedScriptValue.cpp:
* bindings/scripts/CodeGeneratorJS.pm:
* bridge/c/CRuntimeObject.cpp:
* bridge/c/CRuntimeObject.h:
* bridge/c/c_instance.cpp:
* bridge/jni/jsc/JNIUtilityPrivate.cpp:
* bridge/jni/jsc/JavaInstanceJSC.cpp:
* bridge/jni/jsc/JavaRuntimeObject.cpp:
* bridge/jni/jsc/JavaRuntimeObject.h:
* bridge/jsc/BridgeJSC.cpp:
* bridge/objc/ObjCRuntimeObject.h:
* bridge/objc/ObjCRuntimeObject.mm:
* bridge/objc/objc_instance.mm:
* bridge/objc/objc_runtime.h:
* bridge/objc/objc_runtime.mm:
* bridge/runtime_array.cpp:
* bridge/runtime_array.h:
* bridge/runtime_method.cpp:
* bridge/runtime_method.h:
* bridge/runtime_object.cpp:
* bridge/runtime_object.h:

Source/WebKit/mac:

* Plugins/Hosted/ProxyInstance.mm:
(WebKit::ProxyRuntimeMethod::ProxyRuntimeMethod):
(WebKit::ProxyRuntimeMethod::createStructure):
* Plugins/Hosted/ProxyRuntimeObject.h:
(WebKit::ProxyRuntimeObject::createStructure):
* Plugins/Hosted/ProxyRuntimeObject.mm:
(WebKit::ProxyRuntimeObject::ProxyRuntimeObject):
* WebView/WebView.mm:
(aeDescFromJSValue):

Source/WebKit2:

* WebProcess/Plugins/Netscape/JSNPMethod.cpp:
(WebKit::JSNPMethod::JSNPMethod):
* WebProcess/Plugins/Netscape/JSNPMethod.h:
(WebKit::JSNPMethod::createStructure):
* WebProcess/Plugins/Netscape/JSNPObject.cpp:
(WebKit::JSNPObject::JSNPObject):
* WebProcess/Plugins/Netscape/JSNPObject.h:
(WebKit::JSNPObject::createStructure):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79132 268f45cc-cd09-0410-ab3c-d52691b4dbfc

<webkit.org/b/54808> Change jsc target to build directly into JavaScriptCore.framework/Resources/jsc

Reviewed by Dan Bernstein.

* Configurations/Base.xcconfig: Added
JAVASCRIPTCORE_FRAMEWORKS_DIR variable.
* Configurations/JavaScriptCore.xcconfig: Used
JAVASCRIPTCORE_FRAMEWORKS_DIR to define INSTALL_PATH.
* JavaScriptCore.xcodeproj/project.pbxproj: Set the INSTALL_PATH
for Production configuration of jsc target.
(Copy Into Framework): Removed old build phase.
(Fix Framework Reference): Renamed build phase to "Copy Into
Framework".  Added "set -x" call to make the script print the
commands it is running.  Added code to exit early for Production
builds since this was never intended for them.  Added code to
copy jsc into the JavaScriptCore.framework/Resources directory.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79131 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Ryuan Choi  <ryuan.choi@samsung.com>

        Reviewed by Martin Robinson.

        [GTK] Remove mandatory Enchant dependency
        https://bugs.webkit.org/show_bug.cgi?id=51587

        Add enable-spellcheck option which decides whether to use enchant.
        If disabled, WebKitGTK+ will not use enchant library.

        * configure.ac:
2011-02-19  Ryuan Choi  <ryuan.choi@samsung.com>

        Reviewed by Martin Robinson.

        [GTK] Remove mandatory Enchant dependency
        https://bugs.webkit.org/show_bug.cgi?id=51587

        Extract TextCheckerClientEnchant From EditorClientGtk and remove
        enchant dependency from other files.

        * GNUmakefile.am:
        * WebCoreSupport/EditorClientGtk.cpp:
        (WebKit::EditorClient::EditorClient):
        * WebCoreSupport/EditorClientGtk.h:
        (WebKit::EditorClient::textChecker):
        * WebCoreSupport/TextCheckerClientEnchant.cpp: Added.
        (WebKit::TextCheckerClientEnchant::TextCheckerClientEnchant):
        (WebKit::TextCheckerClientEnchant::~TextCheckerClientEnchant):
        (WebKit::TextCheckerClientEnchant::ignoreWordInSpellDocument):
        (WebKit::TextCheckerClientEnchant::learnWord):
        (WebKit::TextCheckerClientEnchant::checkSpellingOfString):
        (WebKit::TextCheckerClientEnchant::getAutoCorrectSuggestionForMisspelledWord):
        (WebKit::TextCheckerClientEnchant::checkGrammarOfString):
        (WebKit::TextCheckerClientEnchant::getGuessesForWord):
        (WebKit::getAvailableDictionariesCallback):
        (WebKit::TextCheckerClientEnchant::updateSpellCheckingLanguage):
        (WebKit::TextCheckerClientEnchant::freeSpellCheckingLanguage):
        * WebCoreSupport/TextCheckerClientEnchant.h: Added.
        (WebKit::TextCheckerClientEnchant::requestCheckingOfString):
        * webkit/webkitwebsettings.cpp:
        (webkit_web_settings_finalize):
        (webkit_web_settings_set_property):
        * webkit/webkitwebview.cpp:
        (webkit_web_view_settings_notify):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79130 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Zan Dobersek  <zandobersek@gmail.com>

        Reviewed by Martin Robinson.

        [cairo][canvas] Drawing from/into float rectangles with width or height in range 0 to 1 fails
        https://bugs.webkit.org/show_bug.cgi?id=54491

        Enable another passing test after fixing float values rounding.

        * platform/gtk/Skipped:
2011-02-19  Zan Dobersek  <zandobersek@gmail.com>

        Reviewed by Martin Robinson.

        [cairo][canvas] Drawing from/into float rectangles with width or height in range 0 to 1 fails
        https://bugs.webkit.org/show_bug.cgi?id=54491

        When width or height in float rectangle are in range (0, 0.5) or (-0.5, 0)
        and would round to 0, alter the behaviour to ensure that width or height are
        at least 1 pixel in size in these cases.

        * platform/graphics/cairo/GraphicsContextCairo.cpp:
        (WebCore::GraphicsContext::roundToDevicePixels):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79129 268f45cc-cd09-0410-ab3c-d52691b4dbfc

LLVM Compiler build fix.

* platform/graphics/ShadowBlur.cpp:
(WebCore::ShadowBlur::blurLayerImage): Eliminated a file static that generated
a global initializer since the compiler does not evaluate sqrtf at compile time.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79128 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rename performLookupAtCurrentMouseLocation to performDictionaryLookupAtCurrentMouseLocation

Reviewed by Anders Carlsson.

* UIProcess/API/mac/WKView.mm:
(-[WKView performDictionaryLookupAtCurrentMouseLocation]):
* UIProcess/API/mac/WKViewPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79127 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Siddharth Mathur  <siddharth.mathur@nokia.com>

        Reviewed by Laszlo Gombos.

        [Symbian] OSAllocator implementation for Symbian OS.
        Manages both data and code region requests. V8 and Sunspider tested
        OK with interpreter. Not tested with JSC JIT yet as it has unrelated
        failures. Also no thread safety yet.
        https://bugs.webkit.org/show_bug.cgi?id=51128

        * JavaScriptCore.pri: removed HAL linkage
        * wtf/Bitmap.h:
        (WTF::::findRunOfZeros): find run of zeros in a bitmap. quick n dirty
        * wtf/OSAllocator.h:
        (WTF::OSAllocator::decommitAndRelease): decommit explicitly
        * wtf/OSAllocatorSymbian.cpp: Impl. of OSAllocator interface
        (WTF::allocateCodeChunk): utility for code chunks
        (WTF::deallocateCodeChunk): utility for code chunks
        (WTF::dataAllocatorInstance): getter for data allocator instance
        (WTF::OSAllocator::reserveUncommitted):
        (WTF::OSAllocator::releaseDecommitted):
        (WTF::OSAllocator::commit):
        (WTF::OSAllocator::decommit):
        (WTF::OSAllocator::reserveAndCommit):
        (WTF::PageAllocatorSymbian::PageAllocatorSymbian): maps requests
        to one underlying Symbian chunk
        (WTF::PageAllocatorSymbian::~PageAllocatorSymbian):
        (WTF::PageAllocatorSymbian::reserve):
        (WTF::PageAllocatorSymbian::release):
        (WTF::PageAllocatorSymbian::commit):
        (WTF::PageAllocatorSymbian::decommit):
        (WTF::PageAllocatorSymbian::contains):
        * wtf/PageAllocatorSymbian.h: Added.
        (WTF::SymbianChunk::SymbianChunk): wrapper around RChunk
        (WTF::SymbianChunk::~SymbianChunk):
        (WTF::SymbianChunk::contains):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79126 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Alejandro G. Castro  <alex@igalia.com>

        Unreviewed, fixed typo in the GTK+ skipped file.

        * platform/gtk/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79125 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Yong Li  <yoli@rim.com>

        Reviewed by Eric Seidel.

        https://bugs.webkit.org/show_bug.cgi?id=54687
        When being built with armcc, "int" bit fields are treated as
        unsigned integers, which will fail the comparisons like "m_offset == -1".
        Using "signed" fixes the problem.

        * assembler/ARMAssembler.h:
        * assembler/ARMv7Assembler.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79124 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Brian Ryner  <bryner@chromium.org>

        Reviewed by Adam Barth.

        Replace the #include of DocumentLoader.h in Document.h with a
        forward declaration, and add a missing #include now that this
        transitive include is gone.
        https://bugs.webkit.org/show_bug.cgi?id=50489

        No new tests required.

        * bindings/ScriptControllerBase.cpp:
        * dom/Document.h:
2011-02-19  Brian Ryner  <bryner@chromium.org>

        Reviewed by Adam Barth.

        Add missing #includes of DocumentLoader.h, needed since the transitive
        include through Document.h is removed.
        https://bugs.webkit.org/show_bug.cgi?id=50489

        * src/WebPageSerializerImpl.cpp:
        * src/WebSearchableFormData.cpp:
2011-02-19  Brian Ryner  <bryner@chromium.org>

        Reviewed by Adam Barth.

        Add missing #include of DocumentLoader.h, needed since the transitive
        include through Document.h is removed.
        https://bugs.webkit.org/show_bug.cgi?id=50489

        * WebView/WebFrame.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79123 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19 Patrick Gansterer <paroga@webkit.org>

        Reviewed by Antonio Gomes.

        [EFL] Remove WebCore::currentTime()
        https://bugs.webkit.org/show_bug.cgi?id=53886

        There is no declaration for it and we use WTF::currentTime() everywhere.

        * platform/efl/SystemTimeEfl.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79122 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19 Patrick Gansterer <paroga@webkit.org>

        Reviewed by Andreas Kling.

        Move KeygenWinCE from wince into win directory
        https://bugs.webkit.org/show_bug.cgi?id=54804

        Move this file into the win directory, since it can be used on all windows platforms.

        * CMakeListsWinCE.txt:
        * platform/win/SSLKeyGeneratorWin.cpp: Renamed from platform/wince/KeygenWinCE.cpp.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79121 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Patrick Gansterer  <paroga@webkit.org>

        Reviewed by Andreas Kling.

        Make WinCE a core builder
        https://bugs.webkit.org/show_bug.cgi?id=54801

        * Scripts/webkitpy/common/net/buildbot/buildbot.py:
        * Scripts/webkitpy/common/net/buildbot/buildbot_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79120 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Patrick Gansterer  <paroga@webkit.org>

        Reviewed by Andreas Kling.

        Make FileChooserWin.cpp more portable and use it on WinCE
        https://bugs.webkit.org/show_bug.cgi?id=54803

        * CMakeListsWinCE.txt:
        * platform/win/FileChooserWin.cpp:
        (WebCore::FileChooser::basenameForWidth):
        * platform/wince/FileChooserWinCE.cpp: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79119 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Benjamin Poulain  <ikipou@gmail.com>

        Reviewed by Andreas Kling.

        [Qt] Enable the skipped tests combining workers and database
        https://bugs.webkit.org/show_bug.cgi?id=54802

        Enable the tests that do not seem to be flaky anymore.

        * platform/qt/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79118 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  James Simonsen  <simonjam@chromium.org>

        Reviewed by Adam Barth.

        Make ScriptElement match the HTML5 spec
        https://bugs.webkit.org/show_bug.cgi?id=54676

        This implements the "prepare a script" section of the HTML5 spec in ScriptElement::prepareScript().
        http://www.whatwg.org/specs/web-apps/current-work/multipage/scripting-1.html#prepare-a-script

        There are a couple of things missing from the spec that would be new functionality. These will be added later.
        - Support for async=false
        - Empty src attribute should dispatch an error.

        There are a couple of slight behavioral changes to match the spec.
        - An XHTML script that is loaded then copied will not fire load on the copy.
        - If onbeforeload moves the script to a different document, then it will not try to execute again because wasAlreadyStarted is true.

        The parsers were updated to use the new API, but not rewritten to look like the spec. That will be done separately.

        * fast/dom/HTMLScriptElement/move-in-beforeload.html: Original author says test was only meant to check for crashes. Beforeload is not specified by HTML5. Test was modified to assume moved script wouldn't execute because wasAlreadyStarted is set.
        * fast/dom/script-clone-rerun-src.xhtml: According to HTML5 spec, load should only fire after a script executes. Test was modified to match that. The old broken behavior was that load fired on the cloned element, even though it didn't execute.
2011-02-19  James Simonsen  <simonjam@chromium.org>

        Reviewed by Adam Barth.

        Make ScriptElement match the HTML5 spec
        https://bugs.webkit.org/show_bug.cgi?id=54676

        This implements the "prepare a script" section of the HTML5 spec in ScriptElement::prepareScript().
        http://www.whatwg.org/specs/web-apps/current-work/multipage/scripting-1.html#prepare-a-script

        There are a couple of things missing from the spec that would be new functionality. These will be added later.
        - Support for async=false
        - Empty src attribute should dispatch an error.

        There are a couple of slight behavioral changes to match the spec.
        - An XHTML script that is loaded then copied will not fire load on the copy.
        - If onbeforeload moves the script to a different document, then it will not try to execute again because wasAlreadyStarted is true.

        The parsers were updated to use the new API, but not rewritten to look like the spec. That will be done separately.

        Test: All existing tests.

        * dom/ScriptElement.cpp: Rewritten to match HTML5 spec.
        (WebCore::ScriptElement::ScriptElement):
        (WebCore::ScriptElement::insertedIntoDocument): Logic moved to prepareScript.
        (WebCore::ScriptElement::childrenChanged): Logic moved to prepareScript.
        (WebCore::ScriptElement::handleSourceAttribute): Logic moved to prepareScript.
        (WebCore::isLegacySupportedJavaScriptLanguage): Added to support old script types in layout tests.
        (WebCore::ScriptElement::isScriptTypeSupported): Derived from old shouldExecuteAsJavaScript().
        (WebCore::ScriptElement::prepareScript): START HERE. Main change. Should read exactly like HTML5's "prepare a script." Legacy type support needed for layout tests using XML parser.
        (WebCore::ScriptElement::requestScript): Most logic moved to prepareScript. Check security settings here.
        (WebCore::ScriptElement::executeScript): Combined evaluateScript() and executeScript() from old code. Logic moved to prepareScript.
        (WebCore::ScriptElement::stopLoadRequest): Ignore parser executed scripts.
        (WebCore::ScriptElement::execute): Renamed executeScript.
        (WebCore::ScriptElement::notifyFinished): We should only listen for non-parser executed scripts.
        (WebCore::ScriptElement::ignoresLoadRequest): New variable names.
        (WebCore::ScriptElement::childrenAreCommentsOrEmptyText): Added for HTML5 compliance.
        (WebCore::ScriptElement::scriptCharset): Use HTML5 variables.
        * dom/ScriptElement.h:
        (WebCore::ScriptElement::willBeParserExecuted): Added.
        (WebCore::ScriptElement::readyToBeParserExecuted): Added.
        (WebCore::ScriptElement::willExecuteWhenDocumentFinishedParsing): Added.
        (WebCore::ScriptElement::cachedScript): prepareScript() is the only place that should load scripts. This accessor lets the parsers listen for when loads finish.
        (WebCore::ScriptElement::isParserInserted): Added.
        * dom/XMLDocumentParserLibxml2.cpp:
        (WebCore::XMLDocumentParser::endElementNs): Should behave the same. Offloads much of its work to prepareScript().
        * dom/XMLDocumentParserQt.cpp:
        (WebCore::XMLDocumentParser::parseEndElement): Identical to libxml2 changes.
        * html/HTMLScriptElement.cpp:
        (WebCore::HTMLScriptElement::insertedIntoDocument): No longer needs url.
        (WebCore::HTMLScriptElement::hasSourceAttribute): Added.
        * html/HTMLScriptElement.h:
        * html/parser/HTMLScriptRunner.cpp:
        (WebCore::HTMLScriptRunner::requestPendingScript): Requesting scripts offloaded to ScriptElement.
        (WebCore::HTMLScriptRunner::runScript): Should behave the same. Offloads much of its work to prepareScript().
        * svg/SVGScriptElement.cpp:
        (WebCore::SVGScriptElement::svgAttributeChanged): New ScriptElement function names.
        (WebCore::SVGScriptElement::insertedIntoDocument): No longer needs url.
        (WebCore::SVGScriptElement::finishParsingChildren): ScriptElement::finishParsingChildren is gone.
        (WebCore::SVGScriptElement::hasSourceAttribute): Added.
        (WebCore::SVGScriptElement::dispatchLoadEvent): New ScriptElement function names.
        * svg/SVGScriptElement.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79114 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Lukasz Slachciak  <l.slachciak@samsung.com>

        Unreviewed build fix.

        SoupSession build break fix
        https://bugs.webkit.org/show_bug.cgi?id=54796

        * ewk/ewk_settings.cpp:
        (ewk_settings_proxy_uri_set):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79113 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Marc-Antoine Ruel  <maruel@chromium.org>

        Reviewed by James Robinson.

        Split webcore_rendering off webcore_remaining to reduce its size for WPO builds
        https://bugs.webkit.org/show_bug.cgi?id=54789

        * WebCore.gyp/WebCore.gyp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79112 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Updated test expectations (more fallout from r78846).

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79111 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Bill Budge  <bbudge@chromium.org>

        Reviewed by David Levin.

        ThreadableLoaderClient needs willSendRequest method
        https://bugs.webkit.org/show_bug.cgi?id=54688

        No new tests. Exposes no new functionality

        * WebCore.gypi:
        * loader/DocumentThreadableLoader.cpp:
        (WebCore::DocumentThreadableLoader::willSendRequest):
        (WebCore::DocumentThreadableLoader::didReceiveData):
        (WebCore::DocumentThreadableLoader::didReceiveCachedMetadata):
        * loader/DocumentThreadableLoaderClient.h: Added.
        (WebCore::DocumentThreadableLoaderClient::isDocumentThreadableLoaderClient):
        (WebCore::DocumentThreadableLoaderClient::willSendRequest):
        * loader/ThreadableLoaderClient.h:
        (WebCore::ThreadableLoaderClient::isDocumentThreadableLoaderClient):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79110 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Lukasz Slachciak  <l.slachciak@samsung.com>

        Reviewed by Andreas Kling.

        [EFL] Use standard WebCore::homeDirectoryPath method instead of reimplementing it
        https://webkit.org/b/54781

        * ewk/ewk_main.cpp:
        (_ewk_init_body):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79108 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Tests that going forward within the same document does not stop loading.

        * http/tests/navigation/forward-to-fragment-fires-onload.html: Added.
        * http/tests/navigation/forward-to-fragment-fires-onload-expected.txt: Added.
        * http/tests/navigation/resources/forward-to-fragment-fires-onload-2.html: Added.
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Avoid stopping all loaders in goToItem for same document navigations
        or pseudo-back-forward URLs.  Make HistoryController::goToItem private
        to force callers to go through Page::goToItem.  Also add a callback to
        FrameLoaderClient to let clients decide whether to stop loading first.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * loader/EmptyClients.h:
        * loader/FrameLoader.h:
        * loader/FrameLoaderClient.h:
        * loader/HistoryController.cpp:
        * loader/HistoryController.h:
        * page/Page.cpp:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Call Page::goToItem instead of HistoryController::goToItem, so that
        FrameLoader::stopAllLoaders is called first.  Also adds a callback in
        FrameLoaderClient for checking for pseudo-back-forward URLs.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * src/FrameLoaderClientImpl.cpp:
        (WebKit::FrameLoaderClientImpl::shouldStopLoadingForHistoryItem): Added.
        * src/FrameLoaderClientImpl.h:
        * src/WebFrameImpl.cpp:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Add a FrameLoaderClient callback for whether to stop loading before goToItem.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * WebCoreSupport/FrameLoaderClientEfl.cpp:
        (WebCore::FrameLoaderClientEfl::shouldStopLoadingForHistoryItem): Added.
        * WebCoreSupport/FrameLoaderClientEfl.h:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Add a FrameLoaderClient callback for whether to stop loading before goToItem.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * WebCoreSupport/FrameLoaderClientGtk.cpp:
        (WebKit::FrameLoaderClientGtk::shouldStopLoadingForHistoryItem): Added.
        * WebCoreSupport/FrameLoaderClientGtk.h:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Add a FrameLoaderClient callback for whether to stop loading before goToItem.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * WebCoreSupport/FrameLoaderClientHaiku.cpp:
        (WebCore::FrameLoaderClientHaiku::shouldStopLoadingForHistoryItem): Added.
        * WebCoreSupport/FrameLoaderClientHaiku.h:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Add a FrameLoaderClient callback for whether to stop loading before goToItem.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::shouldStopLoadingForHistoryItem): Added.
        * WebCoreSupport/WebFrameLoaderClient.h:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Add a FrameLoaderClient callback for whether to stop loading before goToItem.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * WebCoreSupport/FrameLoaderClientQt.cpp:
        (WebCore::FrameLoaderClientQt::shouldStopLoadingForHistoryItem): Added.
        * WebCoreSupport/FrameLoaderClientQt.h:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Add a FrameLoaderClient callback for whether to stop loading before goToItem.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * WebCoreSupport/WebFrameLoaderClient.cpp:
        (WebFrameLoaderClient::shouldStopLoadingForHistoryItem): Added.
        * WebCoreSupport/WebFrameLoaderClient.h:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Add a FrameLoaderClient callback for whether to stop loading before goToItem.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * WebCoreSupport/FrameLoaderClientWinCE.cpp:
        (WebKit::FrameLoaderClientWinCE::shouldStopLoadingForHistoryItem): Added.
        * WebCoreSupport/FrameLoaderClientWinCE.h:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Add a FrameLoaderClient callback for whether to stop loading before goToItem.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * WebKitSupport/FrameLoaderClientWx.cpp:
        (WebCore::FrameLoaderClientWx::shouldStopLoadingForHistoryItem): Added.
        * WebKitSupport/FrameLoaderClientWx.h:
2011-02-19  Charlie Reis  <creis@chromium.org>

        Reviewed by Mihai Parparita.

        Ensure loading has stopped in HistoryController::goToItem
        https://bugs.webkit.org/show_bug.cgi?id=54517

        Add a FrameLoaderClient callback for whether to stop loading before goToItem.

        Test: http/tests/navigation/forward-to-fragment-fires-onload.html

        * WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
        (WebFrameLoaderClient::shouldStopLoadingForHistoryItem): Added.
        * WebProcess/WebCoreSupport/WebFrameLoaderClient.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79107 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Adam Barth  <abarth@webkit.org>

        Reviewed by Daniel Bates.

        Fix xssAuditor/iframe-injection.html
        https://bugs.webkit.org/show_bug.cgi?id=54591

        Update expected results to show that we pass.

        * http/tests/security/xssAuditor/iframe-injection-expected.txt:
2011-02-19  Adam Barth  <abarth@webkit.org>

        Reviewed by Daniel Bates.

        Fix xssAuditor/iframe-injection.html
        https://bugs.webkit.org/show_bug.cgi?id=54591

        We should block the iframe src attribute.  Although this technically
        can't be used to run script, it's a pretty easy vector for stealing
        passwords.

        * html/parser/XSSFilter.cpp:
        (WebCore::XSSFilter::filterTokenInitial):
        (WebCore::XSSFilter::filterIframeToken):
        * html/parser/XSSFilter.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79106 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-19  Lukasz Slachciak  <l.slachciak@samsung.com>

        Reviewed by Andreas Kling.

        Fixed problem with launching EWebLauncher with "run-launcher --efl". Incorrect path referring to non-existing .libs directory ws set.
        https://webkit.org/b/54778

        * Scripts/webkitdirs.pm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79105 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Tony Gentilcore  <tonyg@chromium.org>

        Reviewed by Eric Seidel.

        Let the parser yield for layout before running scripts
        https://bugs.webkit.org/show_bug.cgi?id=54355

        Prior to this patch, the parser would yield to perform a layout/paint before running a
        script only if the script or a stylesheet blocking the script is not loaded yet. Since we
        don't preload scan into the body while parsing the head, typically we'll block on a script
        early in the body that causes us to yield to do the first paint within a reasonable time.

        However, I'm planning to change the PreloadScanner to scan into the body from the head.
        That significantly improves overall load time, but would hurt first paint time because
        fewer scripts would be blocked during parsing and thus wouldn't yield.

        This change causes us to yield before running scripts if we haven't painted yet (regardless
        of whether or not the script is loaded). In addition to allowing the above mentioned
        PreloadScanner change to be implemented without regressing first paint time, this also
        improves first paint time by itself.

        I tested Alexa's top 45 websites using Web Page Replay to control the content and simulate
        bandwidth. This patch improved average first paint time by 1% over an unlimited connection,
        6% over a 1Mbps connection and 11% over a 5Mbps connection. There was no statistically
        signifcant change in page load time.

        Within the pages tested, 33 had no statistically significant change in time to first paint,
        12 improved, and none regressed. Of the improved, some of the standouts from the 1Mbps set
        are: 20% on youtube, 37% on wiki, 27% on ebay, 13% on cnn, 16% on espn, 74% on sohu.

        * html/parser/HTMLDocumentParser.cpp:
        (WebCore::HTMLDocumentParser::canTakeNextToken): This is the new yield point.
        (WebCore::HTMLDocumentParser::pumpTokenizer): Remove ASSERT that we are not paused. isPaused
        means that we are waiting for a script. Bug 54574 changed pumpTokenizer() so that it does
        the right thing whether we are just before a token or waiting for a script. Now that we may
        yield before a token or before a script, this may be called while paused.
        * html/parser/HTMLParserScheduler.cpp:
        (WebCore::isLayoutTimerActive): Added a FIXME because r52919 changed minimumLayoutDelay()
        to return m_extraLayoutDelay instead of 0 as a minimum. So checking !minimumLayoutDelay()
        no longer works. The fix is to change it to check minimumLayoutDelay() ==
        m_extraLayoutDelay. But this is all the more reason to move this method onto Document. I'll
        do this in a follow up.
        (WebCore::HTMLParserScheduler::checkForYieldBeforeScript): Added.
        * page/FrameView.h:
        (WebCore::FrameView::hasEverPainted): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79104 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Dawit Alemayehu  <adawit@kde.org>

        Reviewed by Andreas Kling.

        [Qt]  Button Element is rendered w/ text off-center.
        https://bugs.webkit.org/show_bug.cgi?id=53373

        Test: LayoutTests/fast/forms/button-white-space.html

        * platform/qt/RenderThemeQt.cpp:
        (WebCore::RenderThemeQt::adjustButtonStyle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79103 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  David Levin  <levin@chromium.org>

        Reviewed by Eric Seidel.

        check-webkit-style: Misses brace style error when the line with the { has a } in it.
        https://bugs.webkit.org/show_bug.cgi?id=54769

        * Scripts/webkitpy/style/checkers/cpp.py: Fixed the check for the close brace
          to only look after the last open brace, so that the open brace in this line
          "} else {" will still be able to trigger the error.
        * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added related unit tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79102 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed test expectations fix.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79101 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed test expectations fix.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79099 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Fabrizio Machado  <fabrizio.machado@nokia.com>

        Reviewed by Eric Seidel.

        Remove reduntant checks.
        https://bugs.webkit.org/show_bug.cgi?id=54764

        * WebCoreSupport/DumpRenderTreeSupportQt.cpp:
        (DumpRenderTreeSupportQt::elementDoesAutoCompleteForElementWithId):
        * WebCoreSupport/NotificationPresenterClientQt.cpp:
        (WebCore::NotificationPresenterClientQt::toPage):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79098 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed. Mark object-clip-rects-assertion.html as timing out.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79097 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Zan Dobersek  <zandobersek@gmail.com>

        Reviewed by Martin Robinson.

        [Gtk] Re-enable meter tag support
        https://bugs.webkit.org/show_bug.cgi?id=54762

        Enable meter tag support in the configure.ac. This feature was
        previously enabled by default in the makefile, but should be
        enabled here after changes in r78981.

        * configure.ac:
2011-02-18  Zan Dobersek  <zandobersek@gmail.com>

        Reviewed by Martin Robinson.

        [Gtk] Re-enable meter tag support
        https://bugs.webkit.org/show_bug.cgi?id=54762

        Meter tag is supported on the Gtk port, enable it by default.

        * Scripts/build-webkit:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79096 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  David Levin  <levin@chromium.org>

        Reviewed by Eric Seidel.

        check-webkit-style falsely complains about WebKitGTK+ public headers
        https://bugs.webkit.org/show_bug.cgi?id=54650

        * Scripts/webkitpy/style/checker.py: Add some exceptions for the gtk files.
        * Scripts/webkitpy/style/checkers/cpp.py: Restrict the checks for WEBKIT_API
          to the chromium directory (and improved the checks slightly).
        * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added corresponding tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79095 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed chromium expectations change.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79094 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Jonathan Backer  <backer@chromium.org>

        Reviewed by Eric Seidel.

        [chromium] Fix leak of texture IDs in compositor.
        https://bugs.webkit.org/show_bug.cgi?id=54750

        No new tests. It is extremely unlikely that this leak would have
        any user visible impact because only a few bytes of space are wasted
        (we're leaking texture IDs, not actual textures) and the space of
        texture IDs is large (32 bits).

       * platform/graphics/chromium/TextureManager.cpp:
        (WebCore::TextureManager::requestTexture):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79093 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Dominic Mazzoni  <dmazzoni@google.com>

        Reviewed by Dimitri Glazkov.

        Marks any actions triggered by performDefaultAction as a user
        gesture, so that a user with a screen reader isn't prevented from
        performing operations that must be triggered by a user gesture.
        https://bugs.webkit.org/show_bug.cgi?id=54759

        * src/WebAccessibilityObject.cpp:
        (WebKit::WebAccessibilityObject::performDefaultAction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79092 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed expectations change for chromium.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79090 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed expectations change because my rebaselines didn't seem to work.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79088 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Geoffrey Garen  <ggaren@apple.com>

        Reviewed by Sam Weinig.

        Made MarkedSpace block iteration size-class agnostic
        https://bugs.webkit.org/show_bug.cgi?id=54792

        SunSpider reports no change.

        * runtime/MarkedSpace.cpp:
        (JSC::MarkedSpace::clearMarks):
        (JSC::MarkedSpace::sweep):
        (JSC::MarkedSpace::objectCount):
        (JSC::MarkedSpace::size):
        (JSC::MarkedSpace::capacity):
        * runtime/MarkedSpace.h:
        (JSC::MarkedSpace::forEach): Iterate blocks in hashing order instead of
        size class list order. This is a much simpler convention in a world
        of many different size classes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79087 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Yi Shen  <yi.4.shen@nokia.com>

        Reviewed by Tor Arne Vestbø.

        Always display the media controls when requiresFullscreenForVideoPlayback() is true
        https://bugs.webkit.org/show_bug.cgi?id=54308

        For video element, it should have controls when
        Chrome::requiresFullscreenForVideoPlayback() is true.

        * html/HTMLMediaElement.cpp:
        (WebCore::HTMLMediaElement::controls):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79085 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Adrienne Walker  <enne@google.com>

        Reviewed by Kenneth Russell.

        Adjust expectations for chromium GPU tests.
        https://bugs.webkit.org/show_bug.cgi?id=54409
        https://bugs.webkit.org/show_bug.cgi?id=54509

        Correct baselines for the 54409 images will be generated after this lands.

        * platform/chromium/test_expectations.txt:
2011-02-18  Adrienne Walker  <enne@google.com>

        Reviewed by Kenneth Russell.

        [chromium] Use nearest-neighbor filtering for root layer.
        https://bugs.webkit.org/show_bug.cgi?id=54409
        https://bugs.webkit.org/show_bug.cgi?id=54509

        This setting creates more consistent images for LayoutTests and
        prevents small floating point errors in texture coordinates from
        creating off-by-one pixel color differences.

        * platform/graphics/chromium/LayerTilerChromium.cpp:
        (WebCore::LayerTilerChromium::update):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79084 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Kristian Amlie  <kristian.amlie@nokia.com>

        Reviewed by Andreas Kling.

        Added friend exception to Qt platform, which also compiles Windows.

        QtWebKit fails to compile on Windows XP with msvc-2008
        https://bugs.webkit.org/show_bug.cgi?id=54746

        * bytecode/CodeBlock.h:
        * runtime/RegExpObject.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79083 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Gyuyoung Kim  <gyuyoung.kim@samsung.com>

        Reviewed by Kent Tamura.

        [EFL] Fix coding style errors in RenderThemeEfl.h
        https://bugs.webkit.org/show_bug.cgi?id=54693

        Fix style errors in RenderThemeEfl.h

        * platform/efl/RenderThemeEfl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79080 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Geoffrey Garen  <ggaren@apple.com>

        (Rolled back in r79022 with crash fixed.)

        Reviewed by Sam Weinig.

        Use hashing instead of linear search in the conservative pointer test
        https://bugs.webkit.org/show_bug.cgi?id=54767

        SunSpider reports no change.

        * runtime/MarkedSpace.cpp:
        (JSC::MarkedSpace::destroy): No need to explicitly clear the blocks array,
        since freeBlock removes items for us.

        (JSC::MarkedSpace::freeBlock): Fixed a typo that always removed the last
        block from the block set instead of the block being freed. Changed to
        remove a block from our data structures before deallocating it, since
        this is slightly cleaner.

        * runtime/MarkedSpace.h:
        (JSC::MarkedSpace::contains): Variable-sized objects will use more,
        smaller blocks, so it's important for the contains check not to be O(n)
        in the number of blocks.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79079 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>

        Reviewed by Andreas Kling.

        [Qt] The localized vendor name for Qt SIS packages should be "Nokia"
        https://bugs.webkit.org/show_bug.cgi?id=54742

        This change was applied in the Qt repository (qt/src/3rdparty/webkit),
        so we should do the same here in QtWebKit.

        Patch by Eckhart Koppen <eckhart.koppen@nokia.com>
        a8a84f1667966acfa093c4be0b7d4b0900ddd3d9:

        The previously used name "Nokia, Qt" was not usable for Nokia
        Content Signing, which only allows "Nokia" as the visible vendor
        name. The unique vendor ID remains as "Nokia, Qt"

        * WebCore.pro:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79077 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Gyuyoung Kim  <gyuyoung.kim@samsung.com>

        Reviewed by Kent Tamura.

        [EFL] Fix coding style errors in ewk_frame.h
        https://bugs.webkit.org/show_bug.cgi?id=54718

        Fix style errors in ewk_frame.h.

        * ewk/ewk_frame.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79075 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Alexis Menard  <alexis.menard@openbossa.org>

        Reviewed by Andreas Kling.

        [Qt] REGRESSION(r67516) : on www.gmail.com a strange rendering issue appears on the
        menu bar due to flash.
        https://bugs.webkit.org/show_bug.cgi?id=54741

        Only show plugins with a valid size. We then don't involve X11 if there is
        nothing to see anyway.

        * plugins/qt/PluginViewQt.cpp:
        (WebCore::PluginView::updatePluginWidget):
        (WebCore::PluginView::platformStart):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79074 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Kristian Amlie  <kristian.amlie@nokia.com>

        Reviewed by Kenneth Rohde Christiansen.

        Avoided UiTools dependency if the module is not present.

        [Qt] WebKit patches required to work with a modularized version of Qt
        https://bugs.webkit.org/show_bug.cgi?id=53916

        * Source/WebKit.pri:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79072 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Gyuyoung Kim  <gyuyoung.kim@samsung.com>

        Reviewed by Antonio Gomes.

        [EFL] Remove proxy feature of soup when proxy is null
        https://bugs.webkit.org/show_bug.cgi?id=54621

        When proxy value is NULL, remove proxy feature in soup.

        * ewk/ewk_settings.cpp:
        (ewk_settings_proxy_uri_set):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79070 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Ryuan Choi  <ryuan.choi@samsung.com>

        Reviewed by Kent Tamura.

        [EFL] Remove GDK dependency.
        https://bugs.webkit.org/show_bug.cgi?id=53978

        Remove checking GDK library.

        * Source/cmake/OptionsEfl.cmake:
2011-02-18  Ryuan Choi  <ryuan.choi@samsung.com>

        Reviewed by Kent Tamura.

        [EFL] Remove GDK dependency.
        https://bugs.webkit.org/show_bug.cgi?id=53978

        Remove GLIB_SUPPORT macro in GDK related code.
        Regardless of GLIB_SUPPORT, getDefaultFontOptions() will works same way.

        * CMakeListsEfl.txt:
        * platform/graphics/freetype/FontPlatformDataFreeType.cpp:
        (WebCore::getDefaultFontOptions):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79069 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Gyuyoung Kim  <gyuyoung.kim@samsung.com>

        Reviewed by Kent Tamura.

        [EFL] Fix coding style errors in ewk_view.h
        https://bugs.webkit.org/show_bug.cgi?id=54624

        Fix style errors in ewk_view.h.

        * ewk/ewk_view.cpp:
        (ewk_view_pre_render_relative_radius):
        * ewk/ewk_view.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79067 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Peter Kasting  <pkasting@google.com>

        Reviewed by James Robinson.

        https://bugs.webkit.org/show_bug.cgi?id=54605
        Remove some time limits in a test, let test harnesses kill the test if
        it runs long.  Also changes a busy loop from 1 ms to 5 ms repeats since
        this should be much less CPU at little visible cost.

        * http/tests/security/resources/cross-frame-access.js:
        (canAccessFrame.test):
        (canAccessFrame):
        (cannotAccessFrame.test):
        (cannotAccessFrame):
        (closeWindowAndNotifyDone.doneHandler):
        (closeWindowAndNotifyDone):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79066 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Real fix for <rdar://problem/9025723> CrashTracer: [USER]
1 crash in WebProcess at com.apple.WebKit2:
WebKit::WebPage::didChangeScrollOffsetForMainFrame + 31

Reviewed by Adele Peterson.

* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::didChangeScrollOffset):
Add null check for the FrameView. This function can be called when the
FrameView is being torn down during a transition to a new FrameView.

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::didChangeScrollOffsetForMainFrame):
Revert last attempt.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79064 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Noel Gordon  <noel.gordon@gmail.com>

        Reviewed by James Robinson.

        [Chromium] Add elliptical gradient support to GradientSkia
        https://bugs.webkit.org/show_bug.cgi?id=51841

        Covered by existing tests, these need new rebaselines once this patch
        lands for chrome linux, win32

        fast/gradients/css3-color-stop-units.html
        fast/gradients/css3-color-stops.html
        fast/gradients/css3-linear-angle-gradients.html
        fast/gradients/css3-radial-gradients.html
        fast/gradients/css3-radial-gradients2.html
        fast/gradients/css3-radial-gradients3.html
        fast/gradients/css3-repeating-radial-gradients.html

        * platform/graphics/skia/GradientSkia.cpp:
        (WebCore::Gradient::platformGradient):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79063 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Dirk Pranke  <dpranke@chromium.org>

        Reviewed by James Robinson.

        new-run-webkit-tests: enable multiple processes by default on
        mac. This change removes the artificial restrictions we placed
        on the # of child processes we used with the old threading
        model, and switches to the new message-based model and multiple
        processes, where available. If multiple processes are not
        available (Leopard / Python 2.5), then we use the 'inline' model
        instead of the 'old-threads' model on the 'mac' port or one
        process on the 'chromium-mac' port. We need additional testing
        to see if the new 'threads' model works reliably and is worth
        supporting, or if we should just live with things being slightly
        slow.

        https://bugs.webkit.org/show_bug.cgi?id=54596

        * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
        * Scripts/webkitpy/layout_tests/port/chromium_mac_unittest.py:
        * Scripts/webkitpy/layout_tests/port/mac.py:
        * Scripts/webkitpy/layout_tests/port/mac_unittest.py:
        * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79062 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  chris reiss  <christopher.reiss@nokia.com>

        Reviewed by Andreas Kling.

        REGRESSION: Date.parse("Tue Nov 23 20:40:05 2010 GMT") returns NaN
        https://bugs.webkit.org/show_bug.cgi?id=49989

        * fast/js/date-parse-test-expected.txt:
        * fast/js/script-tests/date-parse-test.js:
2011-02-18  chris reiss  <christopher.reiss@nokia.com>

        Reviewed by Andreas Kling.

        REGRESSION: Date.parse("Tue Nov 23 20:40:05 2010 GMT") returns NaN
        https://bugs.webkit.org/show_bug.cgi?id=49989

        updated test fast/js/script-tests/date-parse-test.js

        * wtf/DateMath.cpp:
        (WTF::parseDateFromNullTerminatedCharacters):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79060 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Hang trying to load nytimes.com with Flash installed
<rdar://problem/9018113>

Reviewed by Sam Weinig.

* Platform/CoreIPC/Connection.cpp:
(CoreIPC::Connection::sendSyncMessage):
Call beginWaitForSyncReply() before sending the message. Otherwise, in some cases we could
have gotten back a synchronous request before calling beginWaitForSyncReply(), and then we wouldn't
process it correctly. (Which would lead to the hang).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79059 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Kevin Decker.

        <rdar://problem/9021296> Some plug-ins are the wrong size after zooming

        Use the same frame/bounds adjustment that we use for NSView-based
        plugins in WebKit2, which allows plug-ins and their snapshots to
        render with the correct size after scaling.

        * WebProcess/Plugins/PluginView.cpp:
        (WebKit::PluginView::setBoundsSize):
        (WebKit::PluginView::viewGeometryDidChange):
        (WebKit::PluginView::clipRectInWindowCoordinates):
        * WebProcess/Plugins/PluginView.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79058 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix for <rdar://problem/9025723> CrashTracer: [USER]
1 crash in WebProcess at com.apple.WebKit2:
WebKit::WebPage::didChangeScrollOffsetForMainFrame + 31

Rubber-stamped by Adele Peterson.

Simple null-check.
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::didChangeScrollOffsetForMainFrame):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79057 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  James Robinson  <jamesr@chromium.org>

        Fix typo in boundary test in ASSERT() - test is for an inclusive range, not exclusive.

        * platform/graphics/chromium/ContentLayerChromium.cpp:
        (WebCore::ContentLayerChromium::updateTextureIfNeeded):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79056 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Andrew Wilson  <atwilson@chromium.org>

        Unreviewed, rolling out r79047.
        http://trac.webkit.org/changeset/79047
        https://bugs.webkit.org/show_bug.cgi?id=54596

        Broke canary bots - please watch the canaries the next time
        you land this

        * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
        * Scripts/webkitpy/layout_tests/port/mac.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79055 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Viatcheslav Ostapenko  <ostapenko.viatcheslav@nokia.com>

        Reviewed by Kenneth Rohde Christiansen.

        Tiled backing store area is too big.
        Error in area calculcation causes size of backing store
        up to 6 times bigger than viewport with default multipliers.
        https://bugs.webkit.org/show_bug.cgi?id=54587

        * platform/graphics/TiledBackingStore.cpp:
        (WebCore::TiledBackingStore::createTiles):
2011-02-18  Viatcheslav Ostapenko  <ostapenko.viatcheslav@nokia.com>

        Reviewed by Kenneth Rohde Christiansen.

        Tiled backing store area is too big.
        Error in area calculcation causes size of backing store
        up to 8 times bigger than viewport with default multipliers.
        https://bugs.webkit.org/show_bug.cgi?id=54587

        * UIProcess/TiledDrawingAreaProxy.cpp:
        (WebKit::TiledDrawingAreaProxy::calculateKeepRect):
        (WebKit::TiledDrawingAreaProxy::calculateCoverRect):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79054 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix for <rdar://problem/9018729> Horizontal scroller doesn't
appear when loading a page with a Horizontal scrollbar from
the back/forward cache.

Reviewed by Sam Weinig.

This patch adds a new  bool member variable to FrameView to
keep track of whether we are loading a page from the back/
forward cache. If we are, don't suppress scrollbars on
first layout.
* history/CachedFrame.cpp:
(WebCore::CachedFrameBase::restore):
* page/FrameView.cpp:
(WebCore::FrameView::FrameView):
(WebCore::FrameView::reset):
(WebCore::FrameView::layout):
* page/FrameView.h:
(WebCore::FrameView::setIsRestoringFromBackForward):
(WebCore::FrameView::isRestoringFromBackForward):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79053 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Patrick Gansterer  <paroga@webkit.org>

        Unreviewed WinCE build fix for r78846.

        * platform/graphics/wince/FontWinCE.cpp:
        (WebCore::TextRunComponent::TextRunComponent):
        * platform/graphics/wince/GraphicsContextWinCE.cpp:
        (WebCore::GraphicsContext::drawLineForText):
        (WebCore::GraphicsContext::drawLineForTextChecking):
        (WebCore::GraphicsContext::drawText):
        * platform/wince/FileChooserWinCE.cpp:
        (WebCore::FileChooser::basenameForWidth):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79052 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Patrick Gansterer  <paroga@webkit.org>

        Unreviewed build fix after r78634 and r78786.

        * CMakeLists.txt:
2011-02-18  Patrick Gansterer  <paroga@webkit.org>

        Unreviewed build fix after r78634 and r78786.

        * CMakeListsEfl.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79051 268f45cc-cd09-0410-ab3c-d52691b4dbfc

<rdar://problem/9026169>
https://bugs.webkit.org/show_bug.cgi?id=54780
pendingAPIRequestURL isn't updated for WKPageGoToBackForwardListItem

Reviewed by Maciej Stachowiak.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::goToBackForwardItem):
Update the pending API URL.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79050 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Emil A Eklund  <eae@chromium.org>

        Reviewed by Darin Adler.

        Crash in EventHandler::sendContextMenuEventForKey
        https://bugs.webkit.org/show_bug.cgi?id=54495

        Add test for triggering the context menu key event (VK_APPS) on a hidden
        element.

        * fast/events/menu-keydown-on-hidden-element-expected.txt: Added.
        * fast/events/menu-keydown-on-hidden-element.html: Added.
2011-02-18  Emil A Eklund  <eae@chromium.org>

        Reviewed by Darin Adler.

        Crash in EventHandler::sendContextMenuEventForKey
        https://bugs.webkit.org/show_bug.cgi?id=54495

        Test: fast/events/menu-keydown-on-hidden-element.html

        * page/EventHandler.cpp:
        (WebCore::EventHandler::sendContextMenuEventForKey): Add null check.
2011-02-18  Emil A Eklund  <eae@chromium.org>

        Reviewed by Darin Adler.

        Crash in EventHandler::sendContextMenuEventForKey
        https://bugs.webkit.org/show_bug.cgi?id=54495

        Add support for the context menu key (VK_APPS) to EventSender for
        platforms that support the key.

        * DumpRenderTree/chromium/EventSender.cpp:
        (EventSender::keyDown):
        * DumpRenderTree/gtk/EventSender.cpp:
        (keyDownCallback):
        * DumpRenderTree/qt/EventSenderQt.cpp:
        (EventSender::keyDown):
        * DumpRenderTree/win/EventSender.cpp:
        (keyDownCallback):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79049 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Accelerated content fades in when using a layer backed WKView
<rdar://problem/9021586>

Reviewed by Sam Weinig.

* UIProcess/API/mac/WKView.mm:
(-[WKView _enterAcceleratedCompositingMode:]):
Make a new nested CATransaction and disable animations when adding the layer
hosting subview. This avoids an implicit fade animation that would otherwise occur.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79048 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Dirk Pranke  <dpranke@chromium.org>

        Reviewed by Ojan Vafai.

        new-run-webkit-tests: enable multiple processes by default on mac
        https://bugs.webkit.org/show_bug.cgi?id=54596

        * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
        * Scripts/webkitpy/layout_tests/port/mac.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79047 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Dirk Pranke  <dpranke@chromium.org>

        Reviewed by Adam Roben.

        Skip pretty patch unit tests if ruby isn't installed.

        https://bugs.webkit.org/show_bug.cgi?id=54699

        * Scripts/webkitpy/common/prettypatch_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79046 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-02-18  Dirk Pranke  <dpranke@chromium.org>

        Reviewed by Adam Roben.

        Skip mac.test_skipped_file_paths() when running on win32; the
        test has hardcoded '/' directory separators and since it's a
        port-specific test and we have coverage on other ports it's not
        worth it to make the test more generic.

        https://bugs.webkit.org/show_bug.cgi?id=54698

        * Scripts/webkitpy/layout_tests/port/mac_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@79045 268f45cc-cd09-0410-ab3c-d52691b4dbfc