2009-04-09  Jian Li  <jianli@chromium.org>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=25084
        Upstream changes to V8 bindings for XHR so that it can work under either DOMWindow or WorkerContext.

        * bindings/v8/WorkerContextExecutionProxy.cpp:
        (WebCore::WorkerContextExecutionProxy::dispose):
        (WebCore::WorkerContextExecutionProxy::retrieve):
        (WebCore::WorkerContextExecutionProxy::initContextIfNeeded):
        (WebCore::WorkerContextExecutionProxy::findOrCreateEventListenerHelper):
        (WebCore::WorkerContextExecutionProxy::FindOrCreateEventListener):
        (WebCore::WorkerContextExecutionProxy::findOrCreateObjectEventListener):
        * bindings/v8/WorkerContextExecutionProxy.h:
        * bindings/v8/custom/V8XMLHttpRequestConstructor.cpp:
        (WebCore::CALLBACK_FUNC_DECL):
        * bindings/v8/custom/V8XMLHttpRequestCustom.cpp:
        (WebCore::getEventListener):
        (WebCore::ACCESSOR_SETTER):
        (WebCore::CALLBACK_FUNC_DECL):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42373 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-04-09  Darin Adler  <darin@apple.com>

        Reviewed by Anders Carlsson and Sam Weinig.

        Part of <rdar://problem/5438063> Saving history containing 100,000 entries causes pauses of 2s while browsing

        Longer term solution is to change the design so Safari doesn't read and write all of history.
        This patch is step one: Do the serializing, which is done on the main thread, much faster.

        * WebCore.base.exp: Added new entry points.
        * WebCore.xcodeproj/project.pbxproj: Added new source files.

        * history/cf: Added.

        * history/cf/HistoryPropertyList.cpp: Added.
        * history/cf/HistoryPropertyList.h: Added. Code to write history files. In the future we'll also
        have code for reading here too.

        * platform/cf/BinaryPropertyList.cpp: Added.
        * platform/cf/BinaryPropertyList.h: Added. Code to write binary property list files.

WebKit/mac:

2009-04-09  Darin Adler  <darin@apple.com>

        Reviewed by Anders Carlsson and Sam Weinig.

        Part of <rdar://problem/5438063> Saving history containing 100,000 entries causes pauses of 2s while browsing

        Longer term solution is to change the design so Safari doesn't read and write all of history.
        This patch is step one: Do the serializing, which is done on the main thread, much faster.

        * History/WebHistory.mm:
        (-[WebHistoryPrivate data]): Added. Returns the NSData object containing serialized history.
        For creating new SPI so you can get the data in memory instead of on disk. Uses WebHistoryWriter.
        (-[WebHistoryPrivate saveToURL:error:]): Changed to call [self data
        (-[WebHistory _data]): Added.
        (WebHistoryWriter::WebHistoryWriter): Added.
        (WebHistoryWriter::writeHistoryItems): Added.

        * History/WebHistoryPrivate.h: Added a new _data method.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42372 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Antti Koivisto

        https://bugs.webkit.org/show_bug.cgi?id=25122

        If deferred repainting is active, we need to ensure that there are no repaints
        pending at the end of layoutIfNeededRecursive, which is called just before
        painting via -viewWillDraw.

        * page/FrameView.cpp:
        (WebCore::FrameView::layoutIfNeededRecursive):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42371 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Removed duplicate header include.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42370 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Mike Thole  <mthole@apple.com>

        Rubber-stamped by Mark Rowe.

        Fix 64-bit build

        * Panels/WebAuthenticationPanel.h:
            Declare the 'separateRealmLabel' IBOutlet as an NSTextField instead of 'id'.
            This fixes a duplicate method warning (setAutoresizingMask: on NSView and CALayer)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42369 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  John Grabowski  <jrg@chromium.org>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=25117
        Fix Mac Chromium form controls.

        * rendering/RenderThemeChromiumMac.mm:
        When rendering native controls in Mac Chromium, use a clean NSGraphicsContext.
        Properly save and restore native graphics context.
        (WebCore::RenderThemeChromiumMac::paintCheckbox):
        (WebCore::RenderThemeChromiumMac::paintRadio):
        (WebCore::RenderThemeChromiumMac::paintMenuList):
        (WebCore::RenderThemeChromiumMac::paintSearchFieldCancelButton):
        (WebCore::RenderThemeChromiumMac::paintSearchFieldResultsDecoration):
        (WebCore::RenderThemeChromiumMac::paintSearchFieldResultsButton):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42368 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Drew Wilson  <atwilson@google.com>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=25030
        Refactor MessagePort code to enable Chrome to use a MessagePortProxy object
        instead of a real MessagePort for the entangledPort, since it could
        theoretically live in a separate process.

        The goal of this refactoring is to remove all assumptions that the object
        representing the entangled port is actually of type MessagePort. All
        operations that were previously directly modifying private MessagePort member
        variables in the entangled port (i.e. entangle(), postMessage()) have been changed
        to invoke functions in the  MessagePortProxy interface instead which will allow
        the Chrome bindings to proxy these calls to another process.

        No new layout tests, as this is just a refactoring with no bug fixes
        or new functionality.

        * GNUmakefile.am:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * bindings/js/JSMessagePortCustom.cpp:
        (WebCore::JSMessagePort::mark):
        * dom/MessagePort.cpp:
        (WebCore::MessagePort::clone):
        (WebCore::MessagePort::postMessage):
        (WebCore::MessagePort::deliverMessage):
        (WebCore::MessagePort::startConversation):
        (WebCore::MessagePort::close):
        (WebCore::MessagePort::entangle):
        (WebCore::MessagePort::unentangle):
        * dom/MessagePort.h:
        (WebCore::MessagePort::entangledPort):
        * dom/MessagePortProxy.h: Added.
        (WebCore::MessagePortProxy::~MessagePortProxy):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42367 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Ben Murdoch  <benm@google.com>

        Reviewed by Alexey Proskuryakov.

        The Android platform requires threads to be registered with the VM.
        This patch implements this behaviour inside ThreadingPthreads.cpp.

        * wtf/ThreadingPthreads.cpp: Add a level above threadEntryPoint that takes care of (un)registering threads with the VM.
        (WTF::runThreadWithRegistration): register the thread and run entryPoint. Unregister the thread afterwards.
        (WTF::createThreadInternal): call runThreadWithRegistration instead of entryPoint directly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42366 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-04-08  Sam Weinig  <sam@webkit.org>

        Reviewed by Geoffrey "Big Boy" Garen.

        Fix for <rdar://problem/5745677> Possible to stop load during an unload event
        Also fixes https://bugs.webkit.org/show_bug.cgi?id=20605

        Tests: fast/events/stop-load-in-unload-handler-using-document-write.html
               fast/events/stop-load-in-unload-handler-using-window-stop.html

        Don't allow calling methods that would stop the new load inside the unload
        event.

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::FrameLoader):
        (WebCore::FrameLoader::stopLoading):
        (WebCore::FrameLoader::stopAllLoaders):
        * loader/FrameLoader.h:

LayoutTests:

2009-04-08  Sam Weinig  <sam@webkit.org>

        Reviewed by Geoffrey "Big Boy" Garen.

        Tests for <rdar://problem/5745677> Possible to stop load during an unload event

        * fast/events/resources/pass.html: Added.
        * fast/events/resources/subframe-stop-load-in-unload-handler-using-document-write.html: Added.
        * fast/events/resources/subframe-stop-load-in-unload-handler-using-window-stop.html: Added.
        * fast/events/stop-load-in-unload-handler-using-document-write-expected.txt: Added.
        * fast/events/stop-load-in-unload-handler-using-document-write.html: Added.
        * fast/events/stop-load-in-unload-handler-using-window-stop-expected.txt: Added.
        * fast/events/stop-load-in-unload-handler-using-window-stop.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42365 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Mike Thole  <mthole@apple.com>

        Reviewed by Adam Roben.

        * WebKit.xcodeproj/project.pbxproj:
            Set the Localizable.strings encoding to UTF-16.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42364 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Mike Thole  <mthole@apple.com>

        Reviewed by Ada Chan.

        <rdar://problem/5697111> Basic authentication dialog spoofing vulnerability

        * Panels/WebAuthenticationPanel.h:
            Added IBOutlet for separateRealmLabel
        * Panels/WebAuthenticationPanel.m:
        (-[WebAuthenticationPanel setUpForChallenge:]):
            In the realm case, decide if it is a simple or complex realm name.  A realm name
            is considered complex if it has any whitespace or newline characters.  Present
            alternative text and layout for the complex case, where the realm name isn't inline
            with the rest of the sheet's text.

        * Panels/English.lproj/WebAuthenticationPanel.nib/designable.nib:
        * Panels/English.lproj/WebAuthenticationPanel.nib/keyedobjects.nib:
            Updated the nib with a new 'separateRealmLabel' outlet.
            Updated the File's Owner to correctly be WebAuthenticationPanel.
            Fixed springs on the sheet's icon to keep it from moving during a resize.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42363 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Update results for createDocument-namespace-err.html

        It looks like the test was changed without updating the results before
        it was committed.

        Rubber-stamped by Adele Peterson.

        * fast/dom/DOMImplementation/createDocument-namespace-err-expected.txt:
        Updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42362 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reinstating <rdar://problem/6718589> Option to turn off SVG DOM Objective-C bindings

        Rolled r42345 back in.  The build failure was caused by an
        internal script which had not been updated the same way that
        build-webkit was updated.

JavaScriptCore:

        * Configurations/JavaScriptCore.xcconfig:

WebCore:

        * Configurations/WebCore.xcconfig:
        * DerivedSources.make:
        * bindings/objc/DOM.mm:
        (WebCore::createElementClassMap):
        (+[DOMNode _wrapNode:WebCore::]):
        * bindings/objc/DOMCSS.mm:
        (+[DOMCSSValue _wrapCSSValue:WebCore::]):
        * bindings/objc/DOMEvents.mm:
        (+[DOMEvent _wrapEvent:WebCore::]):
        * bindings/objc/DOMInternal.h:
        * bindings/objc/ExceptionHandlers.mm:
        (WebCore::raiseDOMException):
        * html/HTMLEmbedElement.idl:
        * html/HTMLFrameElement.idl:
        * html/HTMLIFrameElement.idl:
        * html/HTMLObjectElement.idl:

WebKit/mac:

        * Configurations/WebKit.xcconfig:
        * DOM/WebDOMOperations.mm:
        * MigrateHeaders.make:

WebKitTools:

        * Scripts/build-webkit:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42361 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Darin Adler

        https://bugs.webkit.org/show_bug.cgi?id=25109

        Eliminate CompositeAnimationPrivate, which was a needless level
        of abstraction. Remove the methods like addToStyleAvailableWaitList()
        that simply turn around and call the animation controller.

        Rename animationControllerPriv() to animationController().
        Rename willNeedService() to timeToNextService().

        * page/animation/AnimationBase.cpp:
        (WebCore::AnimationBase::~AnimationBase):
        (WebCore::AnimationBase::updateStateMachine):
        (WebCore::AnimationBase::timeToNextService):
        (WebCore::AnimationBase::beginAnimationUpdateTime):
        * page/animation/AnimationBase.h:
        * page/animation/AnimationController.cpp:
        (WebCore::AnimationControllerPrivate::updateAnimationTimer):
        * page/animation/CompositeAnimation.cpp:
        (WebCore::CompositeAnimation::~CompositeAnimation):
        (WebCore::CompositeAnimation::clearRenderer):
        (WebCore::CompositeAnimation::updateTransitions):
        (WebCore::CompositeAnimation::updateKeyframeAnimations):
        (WebCore::CompositeAnimation::animate):
        (WebCore::CompositeAnimation::getAnimatedStyle):
        (WebCore::CompositeAnimation::setAnimating):
        (WebCore::CompositeAnimation::timeToNextService):
        (WebCore::CompositeAnimation::getAnimationForProperty):
        (WebCore::CompositeAnimation::cleanupFinishedAnimations):
        (WebCore::CompositeAnimation::suspendAnimations):
        (WebCore::CompositeAnimation::resumeAnimations):
        (WebCore::CompositeAnimation::overrideImplicitAnimations):
        (WebCore::CompositeAnimation::resumeOverriddenImplicitAnimations):
        (WebCore::CompositeAnimation::isAnimatingProperty):
        (WebCore::CompositeAnimation::pauseAnimationAtTime):
        (WebCore::CompositeAnimation::pauseTransitionAtTime):
        (WebCore::CompositeAnimation::numberOfActiveAnimations):
        * page/animation/CompositeAnimation.h:
        (WebCore::CompositeAnimation::animationController):
        (WebCore::CompositeAnimation::isSuspended):
        (WebCore::CompositeAnimation::hasAnimations):
        (WebCore::CompositeAnimation::CompositeAnimation):
        * page/animation/ImplicitAnimation.cpp:
        (WebCore::ImplicitAnimation::sendTransitionEvent):
        (WebCore::ImplicitAnimation::timeToNextService):
        * page/animation/ImplicitAnimation.h:
        * page/animation/KeyframeAnimation.cpp:
        (WebCore::KeyframeAnimation::sendAnimationEvent):
        (WebCore::KeyframeAnimation::timeToNextService):
        * page/animation/KeyframeAnimation.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42360 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  David Levin  <levin@chromium.org>

        Reviewed by David Kilzer.

        https://bugs.webkit.org/show_bug.cgi?id=25101

        Launch DumpRenderTree and the image diff tool in a way that will work from perl threads.
        This is a workaround due to a perl limitation. When perl does open* calls on a thread, it ignores
        the environment of the current thread and simply uses the environment of the main thread instead.

        * Scripts/execAppWithEnv: Added.
        * Scripts/run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42359 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reverting <rdar://problem/6718589> Option to turn off SVG DOM Objective-C bindings.
        It broke Mac build, and I don't know how to fix it.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42358 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Xan Lopez  <xlopez@igalia.com>

        Unreviewed build fix.

        RenderView::docWidth() was made private by r42334. That commit
        changes its usage for overflowWidth(), so do that here too.

        * page/PrintContext.cpp:
        (WebCore::PrintContext::computePageRects):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42357 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Xan Lopez  <xlopez@igalia.com>

        Unreviewed build fix.

        Checking for __GLIBCXX__ being bigger than some date is not enough
        to get std::tr1, C++0x has to be in use too. Add another check for
        __GXX_EXPERIMENTAL_CXX0X__.

        * wtf/TypeTraits.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42356 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=21546
        [GTK] ATK accessibility enhancements

        Based on a patch by Alp Toker.

        Implement AtkObject::ref_state_set. Still quite a few states not
        implemented.

        * page/gtk/AccessibilityObjectWrapperAtk.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42355 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=21546
        [GTK] ATK accessibility enhancements

        Based on a patch by Alp Toker.

        Move fallback object creation to its own function, as it will be
        used in several places.

        * page/gtk/AccessibilityObjectWrapperAtk.cpp:
        (fallbackObject):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42354 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=21546
        [GTK] ATK accessibility enhancements

        Based on a patch by Alp Toker.

        Move AccessibilityObject methods to their file.

        * page/gtk/AccessibilityObjectAtk.cpp:
        (WebCore::AccessibilityObject::wrapper):
        (WebCore::AccessibilityObject::setWrapper):
        * page/gtk/AccessibilityObjectWrapperAtk.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42353 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=21546
        [GTK] ATK accessibility enhancements

        Rework accessibility type generation code, based on Mozilla a11y
        implementation.

        Have one base a11y type, and generate derived types that implement
        only the necessary interfaces at runtime, based on the specific
        object we are wrapping. This allows to save a lot of code while
        covering all possible cases.

        * page/gtk/AccessibilityObjectWrapperAtk.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42352 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Oliver Hunt.

        Fix document.implementation.createDocument(null, "a:b") not to crash!
        https://bugs.webkit.org/show_bug.cgi?id=25096

        Test: fast/dom/DOMImplementation/createDocument-namespace-err.html

        * dom/ContainerNode.cpp:
        (WebCore::ContainerNode::addChild): add ASSERT() to give a better crash next time
        * dom/DOMImplementation.cpp:
        (WebCore::DOMImplementation::createDocument): check the exception code before using the result

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42351 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Make <select>s in application chrome documents match the Vista system dialog look

        Reviewed by Dave Hyatt.

        * rendering/RenderThemeWin.cpp:
        (WebCore::documentIsInApplicationChromeMode): Added this simple helper
        function.
        (WebCore::RenderThemeWin::getThemeData): In application chrome mode,
        the drop down button never uses any look but the "normal" look.
        (WebCore::RenderThemeWin::paintMenuList): In application chrome mode,
        use the "readonly" style to match system dialogs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42350 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix <rdar://6592515> Crash closing tab with Top Sites or Cover Flow showing after clicking "Go Back" in phishing alert

        RenderButton's m_timer is not stopped when its Document enters the
        page cache. When the WebView was torn down, the timer would fire after
        the Frame's Page had been destroyed, leading to a crash. This patch
        just makes the timer do nothing when it fires. Making the timer not
        fire when in the page cache is covered by bug 25110.

        Reviewed by Dave Hyatt.

        No test possible.

        * rendering/RenderButton.cpp:
        (WebCore::RenderButton::timerFired): Do nothing when we're in the page
        cache.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42349 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-09  David Hyatt  <hyatt@apple.com>

        Reviewed by Adam Roben

        Fix bugs where we would mistakenly turn off blitting for iframes if an external paint
        caused the checks to be triggered too early in the loading process of the frame.

        * rendering/RenderBoxModelObject.cpp:
        (WebCore::RenderBoxModelObject::paintFillLayerExtended):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42348 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix assertion failure in function.apply

Reviewed by Adam Roben.

The result of excess arguments to function.apply is irrelevant
so we don't need to provide a result register.  We were providing
temporary result register but not ref'ing it resulting in an
assertion failure.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42347 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Cameron Zwarich

        https://bugs.webkit.org/show_bug.cgi?id=25106

        Fix signature of virtual function on AnimationBase so that it matches
        those in the derived classes, by removing unused parameter.

        * page/animation/AnimationBase.h:
        (WebCore::AnimationBase::endAnimation):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42346 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    <rdar://problem/6718589> Option to turn off SVG DOM Objective-C bindings

        Reviewed by Darin Adler and Maciej Stachowiak.

        Introduce the ENABLE_SVG_DOM_OBJC_BINDINGS feature define so
        that SVG DOM Objective-C bindings may be optionally disabled.

JavaScriptCore:

        * Configurations/JavaScriptCore.xcconfig: Added
        ENABLE_SVG_DOM_OBJC_BINDINGS variable and use it in
        FEATURE_DEFINES.

WebCore:

        * Configurations/WebCore.xcconfig: Added
        ENABLE_SVG_DOM_OBJC_BINDINGS variable and use it in
        FEATURE_DEFINES and EXCLUDED_SOURCE_FILE_NAMES.  When SVG DOM
        Objective-C bindings are disabled, two generated header files no
        longer exist and none of the generated DOMSVG*.mm files should
        be compiled.
        * DerivedSources.make: All of the WebCore.SVG*.exp files contain
        SVG DOM Objective-C bindings classes, so exclude them unless
        ENABLE_SVG_DOM_OBJC_BINDINGS is set.
        * bindings/objc/DOM.mm: Switched from using ENABLE(SVG) to using
        ENABLE(SVG_DOM_OBJC_BINDINGS).
        (WebCore::createElementClassMap):
        (+[DOMNode _wrapNode:WebCore::]):
        * bindings/objc/DOMCSS.mm: Ditto.
        (+[DOMCSSValue _wrapCSSValue:WebCore::]):
        * bindings/objc/DOMEvents.mm: Ditto.
        (+[DOMEvent _wrapEvent:WebCore::]):
        * bindings/objc/DOMInternal.h: Ditto.
        * bindings/objc/ExceptionHandlers.mm: Added use of
        ENABLE(SVG_DOM_OBJC_BINDINGS).
        (WebCore::raiseDOMException):
        * html/HTMLEmbedElement.idl: Excluded -getSVGDocument method
        in generated Objective-C DOM classes if SVG DOM Objective-C
        bindings are disabled.
        * html/HTMLFrameElement.idl: Ditto.
        * html/HTMLIFrameElement.idl: Ditto.
        * html/HTMLObjectElement.idl: Ditto.

WebKit/mac:

        * Configurations/WebKit.xcconfig: Added
        ENABLE_SVG_DOM_OBJC_BINDINGS variable and use it in
        FEATURE_DEFINES.
        * DOM/WebDOMOperations.mm: Removed unused header.
        * MigrateHeaders.make: Switched from using ENABLE_SVG to using
        ENABLE_SVG_DOM_OBJC_BINDINGS.

WebKitTools:

        * Scripts/build-webkit: Added --[no-]svg-dom-objc-bindings
        command-line argument.  Set special ENABLE_SVG_DOM_OBJC_BINDINGS
        Xcode variable required by WebCore.xcconfig.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42345 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Paul Pedriana <ppedriana@ea.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=20422
        Allow custom memory allocation control.

        * wtf/FastAllocBase.h:
        New added file. Implements allocation base class.
        * wtf/TypeTraits.h:
        Augments existing type traits support as needed by FastAllocBase.
        * wtf/FastMalloc.h:
        Changed to support FastMalloc match validation.
        * wtf/FastMalloc.cpp:
        Changed to support FastMalloc match validation.
        * wtf/Platform.h:
        Added ENABLE_FAST_MALLOC_MATCH_VALIDATION; defaults to 0.
        * GNUmakefile.am:
        Updated to include added FastAllocBase.h.
        * JavaScriptCore.xcodeproj/project.pbxproj:
        Updated to include added FastAllocBase.h.
        * JavaScriptCore.vcproj/WTF/WTF.vcproj:
        Updated to include added FastAllocBase.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42344 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Skip fast/js/toString-stack-overflow.html on Windows

        This test sometimes crashes in Debug builds.

        Rubber-stamped by Geoff Garen.

        * platform/win/Skipped: Skip toString-stack-overflow.html.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42343 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Sam Weinig.

        - limit the workaround for <rdar://problem/5781372> to Mac OS X versions
          that need it

        * platform/mac/WebFontCache.mm:
        (fixUpWeight): Changed this function to be a no-op post-Leopard.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42342 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Dave Hyatt.

        - fix <rdar://problem/6771547> Appending to a RenderText causes all its
          lines to be dirtied

        Test: fast/repaint/text-append-dirty-lines.html

        * rendering/RenderText.cpp:
        (WebCore::RenderText::setTextWithOffset): If no text boxes intersected
        with or came after the new text, dirty the last root box, since the new
        text may begin there. Having at least one dirty line prevents the
        incorrect dirtying of all lines later during layoutInlineChildren().

LayoutTests:

        Reviewed by Dave Hyatt.

        - test for <rdar://problem/6771547> Appending to a RenderText causes all its
          lines to be dirtied

        * fast/repaint/text-append-dirty-lines.html: Added.
        * platform/mac/fast/repaint/text-append-dirty-lines-expected.checksum: Added.
        * platform/mac/fast/repaint/text-append-dirty-lines-expected.png: Added.
        * platform/mac/fast/repaint/text-append-dirty-lines-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42341 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Darin Fisher  <darin@chromium.org>

        Reviewed by Eric Seidel.

        https://bugs.webkit.org/show_bug.cgi?id=25100

        Fixes a Chromium-specific crash in WebCore::ImageBuffer::context that
        occurs when rendering semi-transparent RTL text.

        Test: fast/text/complex-text-opacity.html

        * platform/graphics/chromium/FontChromiumWin.cpp: Handle RTL when computing text bounds.
        (WebCore::TransparencyAwareFontPainter::TransparencyAwareUniscribePainter::estimateTextBounds):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42340 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Sam Weinig  <sam@webkit.org>

        Add Windows specific results.

        * platform/win/http/tests/security: Added.
        * platform/win/http/tests/security/XFrameOptions: Added.
        * platform/win/http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-expected.txt: Added.
        * platform/win/http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-in-body-expected.txt: Added.
        * platform/win/http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-allow-expected.txt: Added.
        * platform/win/http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-deny-expected.txt: Added.
        * platform/win/http/tests/security/XFrameOptions/x-frame-options-parent-same-origin-allow-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42339 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Sam Weinig  <sam@webkit.org>

        Add Tiger specific results.

        * platform/mac-tiger/http/tests/security: Added.
        * platform/mac-tiger/http/tests/security/XFrameOptions: Added.
        * platform/mac-tiger/http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-expected.txt: Added.
        * platform/mac-tiger/http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-in-body-expected.txt: Added.
        * platform/mac-tiger/http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-allow-expected.txt: Added.
        * platform/mac-tiger/http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-deny-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42338 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Improve function.apply performance

Reviewed by Geoff Garen.

Jump through a few hoops to improve performance of function.apply in the general case.

In the case of zero or one arguments, or if there are only two arguments and the
second is an array literal we treat function.apply as function.call.

Otherwise we use the new opcodes op_load_varargs and op_call_varargs to do the .apply call
without re-entering the virtual machine.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42337 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  David Hyatt  <hyatt@apple.com>

        Reviewed by Beth Dakin

        Fix a porting error in my previous patch.  A != was supposed to be an ==.

        * platform/ScrollView.cpp:
        (WebCore::ScrollView::updateScrollbars):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42336 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Skip fast/cookies/local-file-can-set-cookies.html on Windows

        Reviewed by John Sullivan.

        * platform/win/Skipped: Skip local-file-can-set-cookies.html, which is
        failing due to a bug in one of our dependencies.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42335 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-04-08  David Hyatt  <hyatt@apple.com>

        Reviewed by Adam Roben and Darin Adler

        Fix for https://bugs.webkit.org/show_bug.cgi?id=12440, repaints inconsistent for
        fixed positioned elements.

        Rewrite the updateScrollers method to be more correct in its results.

        Test: fast/block/positioning/fixed-positioning-scrollbar-bug.html

        * dom/Document.cpp:
        (WebCore::Document::recalcStyle):
        (WebCore::Document::implicitClose):
        * page/FrameView.cpp:
        (WebCore::FrameView::createScrollbar):
        (WebCore::FrameView::layout):
        (WebCore::FrameView::adjustPageHeight):
        * page/FrameView.h:
        * page/win/FrameWin.cpp:
        (WebCore::computePageRectsForFrame):
        * platform/ScrollView.cpp:
        (WebCore::ScrollView::ScrollView):
        (WebCore::ScrollView::updateScrollbars):
        * platform/ScrollView.h:
        * rendering/RenderView.cpp:
        (WebCore::RenderView::layout):
        (WebCore::RenderView::docHeight):
        (WebCore::RenderView::docWidth):
        * rendering/RenderView.h:

WebKit/mac:

2009-04-08  David Hyatt  <hyatt@apple.com>

        Reviewed by Adam Roben and Darin Adler

        Fix for https://bugs.webkit.org/show_bug.cgi?id=12440, fixed positioned elements end up in
        inconsistent positions.  Rewrite updateScrollers to improve the correctness.

        * WebView/WebDynamicScrollBarsView.h:
        * WebView/WebDynamicScrollBarsView.m:
        (-[WebDynamicScrollBarsView updateScrollers]):

LayoutTests:

2009-04-08  David Hyatt  <hyatt@apple.com>

        Reviewed by Adam Roben and Darin Adler

        Updated results and new tests for https://bugs.webkit.org/show_bug.cgi?id=12440.

        * fast/block/positioning/fixed-positioning-scrollbar-bug.html: Added.
        * platform/mac/fast/block/positioning/fixed-positioning-scrollbar-bug-expected.checksum: Added.
        * platform/mac/fast/block/positioning/fixed-positioning-scrollbar-bug-expected.png: Added.
        * platform/mac/fast/block/positioning/fixed-positioning-scrollbar-bug-expected.txt: Added.
        * platform/mac/fast/flexbox/016-expected.checksum:
        * platform/mac/fast/flexbox/016-expected.png:
        * platform/mac/fast/flexbox/016-expected.txt:
        * platform/mac/fast/flexbox/flex-hang-expected.checksum:
        * platform/mac/fast/flexbox/flex-hang-expected.png:
        * platform/mac/fast/flexbox/flex-hang-expected.txt:
        * platform/mac/fast/lists/001-expected.checksum:
        * platform/mac/fast/lists/001-expected.png:
        * platform/mac/fast/lists/001-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42334 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-04-08  Sam Weinig  <sam@webkit.org>

        Reviewed by Anders Carlsson.

        Fix for <rdar://problem/6226200> Implement Microsoft's X-FRAME-OPTIONS anti-framing defense

        Tests: http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-in-body.html
               http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-allow.html
               http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-deny.html
               http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag.html
               http/tests/security/XFrameOptions/x-frame-options-deny.html
               http/tests/security/XFrameOptions/x-frame-options-parent-same-origin-allow.html
               http/tests/security/XFrameOptions/x-frame-options-parent-same-origin-deny.html

        * dom/Document.cpp:
        (WebCore::Document::processHttpEquiv): Stop the current load and redirect to about:blank
        if an X-FRAME-OPTIONS <meta> tag http-equiq dictates we should.
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::shouldInterruptLoadForXFrameOptions): Add logic to parse
        the X-FRAME-OPTIONS parameter.
        * loader/FrameLoader.h:
        * loader/MainResourceLoader.cpp:
        (WebCore::MainResourceLoader::didReceiveResponse): Stop the current load if framed and
        a X-FRAME-OPTIONS header and its parameter dictate that we should.

LayoutTests:

2009-04-08  Sam Weinig  <sam@webkit.org>

        Reviewed by Anders Carlsson.

        Tests for <rdar://problem/6226200> Implement Microsoft's X-FRAME-OPTIONS anti-framing defense

        * http/tests/security/XFrameOptions: Added.
        * http/tests/security/XFrameOptions/resources: Added.
        * http/tests/security/XFrameOptions/resources/x-frame-options-deny-meta-tag-subframe-in-body.html: Added.
        * http/tests/security/XFrameOptions/resources/x-frame-options-deny-meta-tag-subframe-parent-same-origin-allow.html: Added.
        * http/tests/security/XFrameOptions/resources/x-frame-options-deny-meta-tag-subframe-parent-same-origin-deny.html: Added.
        * http/tests/security/XFrameOptions/resources/x-frame-options-deny-meta-tag-subframe.html: Added.
        * http/tests/security/XFrameOptions/resources/x-frame-options-deny.cgi: Added.
        * http/tests/security/XFrameOptions/resources/x-frame-options-parent-same-origin-allow.cgi: Added.
        * http/tests/security/XFrameOptions/resources/x-frame-options-parent-same-origin-deny.cgi: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny-expected.txt: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-expected.txt: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-in-body-expected.txt: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-in-body.html: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-allow-expected.txt: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-allow.html: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-deny-expected.txt: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag-parent-same-origin-deny.html: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny-meta-tag.html: Added.
        * http/tests/security/XFrameOptions/x-frame-options-deny.html: Added.
        * http/tests/security/XFrameOptions/x-frame-options-parent-same-origin-allow-expected.txt: Added.
        * http/tests/security/XFrameOptions/x-frame-options-parent-same-origin-allow.html: Added.
        * http/tests/security/XFrameOptions/x-frame-options-parent-same-origin-deny-expected.txt: Added.
        * http/tests/security/XFrameOptions/x-frame-options-parent-same-origin-deny.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42333 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix security/block-test-no-port.html on Windows

        Reviewed by Beth Dakin.

        * WebMutableURLRequest.cpp:
        (WebMutableURLRequest::mainDocumentURL): Return the ResourceRequest's
        mainDocumentURL, not its URL.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42332 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix http/tests/xmlhttprequest/xmlhttprequest-unsafe-redirect.html on Windows

        Reviewed by Sam Weinig.

        * platform/network/cf/ResourceHandleCFNet.cpp:
        (WebCore::ResourceHandle::loadResourceSynchronously): Mimic the
        ResourceHandleMac code by propagating the CFError's error code for
        kCFErrorDomainCFNetwork errors to the ResourceResponse.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42331 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Adam Barth  <abarth@webkit.org>

        Reviewed by Darin Adler.

        More tests for our handling of MIME types and style sheets.

        * http/tests/mime/resources/style-with-charset.php: Added.
        * http/tests/mime/resources/style-with-empty-content-type.cgi: Renamed from LayoutTests/http/tests/mime/resources/style.cgi.
        * http/tests/mime/resources/style-with-text-plain.php: Added.
        * http/tests/mime/standard-mode-does-not-load-stylesheet-with-text-plain-and-css-extension-expected.txt: Added.
        * http/tests/mime/standard-mode-does-not-load-stylesheet-with-text-plain-and-css-extension.html: Added.
        * http/tests/mime/standard-mode-does-not-load-stylesheet-with-text-plain-expected.txt: Added.
        * http/tests/mime/standard-mode-does-not-load-stylesheet-with-text-plain.html: Added.
        * http/tests/mime/standard-mode-loads-stylesheet-with-charset-and-css-extension-expected.txt: Copied from LayoutTests/http/tests/mime/standard-mode-loads-stylesheet-without-content-type-header-expected.txt.
        * http/tests/mime/standard-mode-loads-stylesheet-with-charset-and-css-extension.html: Added.
        * http/tests/mime/standard-mode-loads-stylesheet-with-charset-expected.txt: Copied from LayoutTests/http/tests/mime/standard-mode-loads-stylesheet-without-content-type-header-expected.txt.
        * http/tests/mime/standard-mode-loads-stylesheet-with-charset.html: Added.
        * http/tests/mime/standard-mode-loads-stylesheet-with-empty-content-type-expected.txt: Renamed from LayoutTests/http/tests/mime/standard-mode-loads-stylesheet-without-content-type-header-expected.txt.
        * http/tests/mime/standard-mode-loads-stylesheet-with-empty-content-type.html: Added.
        * http/tests/mime/standard-mode-loads-stylesheet-without-content-type-header.html: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42330 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Anders Carlsson  <andersca@apple.com>

        Try to fix the PPC build.

        * plugins/npfunctions.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42329 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Darin Adler  <darin@apple.com>

        Reviewed by Sam Weinig.

        <rdar://problem/5908616> tokenizer resizer should handle edge cases consistently

        * html/HTMLTokenizer.cpp:
        (WebCore::HTMLTokenizer::enlargeBuffer): Handle edge cases the same way as a failed fastMalloc.
        (WebCore::HTMLTokenizer::enlargeScriptBuffer): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42328 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  David Hyatt  <hyatt@apple.com>

        Reviewed by Darin Adler

        Make sure that cloneNode works with SVG attributes that require synchronization.  This fixes
        issues with <use> element shadow tree cloning where the clones can end up not obtaining the
        correct attribute values.

        A subsequent patch of mine will expose this problem and cause it to be covered by existing
        layout tests, so no test is required at this time for this problem.

        * dom/Element.cpp:
        (WebCore::Element::cloneElementWithoutChildren):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42327 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Check in Windows-specific results for an SVG test that was failing due to rounding issues

        Rubber-stamped by Dave Hyatt.

        * platform/win/svg/W3C-SVG-1.1/coords-trans-01-b-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42326 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Timothy Hatcher.

        https://bugs.webkit.org/show_bug.cgi?id=25093
        REGRESSION: some scripts are showing in the Inspector with only "true"
        as the source.

        * inspector/InspectorController.cpp:
        (WebCore::InspectorController::didParseSource): Fixed conversion issue.
        (WebCore::InspectorController::failedToParseSource): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42325 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reverting r42321 because it broke some layout tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42324 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Add a failing test to the Windows Skipped file

        * platform/win/Skipped: Added media/video-size-intrinsic-scale.html.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42323 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-04-08  Adam Barth  <abarth@webkit.org>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=24904

        Verify the MIME type of stylesheets (in standards mode) by looking
        at the Content-Type header directly.  This bypasses any content
        sniffing that might be confusing the issue.

        Test: http/tests/mime/standard-mode-loads-stylesheet-without-content-type-header.html

        * loader/CachedCSSStyleSheet.cpp:
        (WebCore::CachedCSSStyleSheet::canUseSheet):

LayoutTests:

2009-04-08  Adam Barth  <abarth@webkit.org>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=24904

        Test that we apply style sheets in standards mode even if
        those style sheets have an empty Content-Type header.

        * http/tests/mime/resources/style.cgi: Added.
        * http/tests/mime/standard-mode-loads-stylesheet-without-content-type-header-expected.txt: Added.
        * http/tests/mime/standard-mode-loads-stylesheet-without-content-type-header.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42322 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Anantanarayanan Iyengar  <ananta@chromium.org>

        Reviewed by Darin Fisher.

        Added support for returning plugins which support the wildcard mime type (*).
        This is on the same lines as Firefox, which uses a wildcard plugin (npnul32.dll) for
        plugin installation. The changes to support wildcard mime type matching are in
        PluginDatabase.cpp which is used by Webkit's plugin implementation, and
        PluginData.cpp/.h which is used by Chromium.

        https://bugs.webkit.org/show_bug.cgi?id=24988

        Test: plugins/netscape-plugin-wildcard-plugin.html

        * dom/DOMImplementation.cpp:
        (WebCore::DOMImplementation::createDocument):
        * plugins/PluginData.cpp:
        (WebCore::PluginData::supportsMimeType):
        (WebCore::PluginData::supportsMimeTypeIgnoringWildCards):
        * plugins/PluginData.h:
        * plugins/PluginDatabase.cpp:
        (WebCore::PluginDatabase::pluginForMIMEType):
        * plugins/win/PluginPackageWin.cpp:
        (WebCore::PluginPackage::isPluginBlacklisted):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42321 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Eric Carlson  <eric.carlson@apple.com>

        Reviewed by Simon Fraser.

        https://bugs.webkit.org/show_bug.cgi?id=25041
        <rdar://problem/6697754>
        Bug 25041: HTMLMediaElement: OSX autoplay broken by r41907

        OSX media player should set readyState to HAVE_ENOUGH_DATA when the movie's load state
        reaches QTMovieLoadStatePlaythroughOK, not when the entire file has been downloaded,
        so autoplay can begin when it should.

        * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
        (WebCore::MediaPlayerPrivate::updateStates): Set readState to
        HAVE_ENOUGH_DATA when QTMovie load state reaches QTMovieLoadStatePlaythroughOK.
        (WebCore::MediaPlayerPrivate::paint): Add braces missed in r42203.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42320 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by David Kilzer. Based on a patch by Grace Kloba.

        https://bugs.webkit.org/show_bug.cgi?id=24656
        cacheControlContainsNoCache() in ResourceResponseBase.h is wrong

        * platform/network/ResourceResponseBase.cpp:
        (WebCore::ResourceResponseBase::parseCacheControlDirectives): Removed misused code for
        parsing directive values, fixing parsing of directives that we care about.

        * platform/network/ResourceResponseBase.h:
        (WebCore::ResourceResponseBase::cacheControlContainsNoCache): Fixed a copy/paste mistake,
        m_cacheControlContainsMustRevalidate was returned instead of m_cacheControlContainsNoCache.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42319 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-08  Alejandro Piñeiro  <apinheiro@igalia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=21546
        [GTK] ATK accessibility enhancements

        Use getOrCreate instead of just get, as we need to actually create
        the wrapper object the first time this function is called.

        * webkit/webkitwebview.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42318 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Make text fields match the system look on Vista

        Reviewed by Dave Hyatt.

        * rendering/RenderThemeWin.cpp:
        (WebCore::RenderThemeWin::getThemeData): Use the new-to-Vista
        EP_EDITBORDER_NOSCROLL part for text fields.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42317 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Print the number of files being compiled when using pdevenv

        * Scripts/parallelcl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42316 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=25073
        JavaScriptCore tests don't run if time zone is not PST

        * API/tests/testapi.c:
        (timeZoneIsPST): Added a function that checks whether the time zone is PST, using the same
        method as functions in DateMath.cpp do for formatting the result.
        (main): Skip date string format test if the time zone is not PST.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42315 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Maciej Stachowiak.

        https://bugs.webkit.org/show_bug.cgi?id=24957
        REGRESSION: Google calendar widget no longer shows upcoming events

        Test: http/tests/xmlhttprequest/authorization-header.html

        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequestStaticData::XMLHttpRequestStaticData): Removed Authorization from
        the list of forbidden headers.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42314 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Ignore passwords.json to prevent it from accidentally being checked in.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42313 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Land the current build.webkit.org configuration for safekeeping.

* BuildSlaveSupport/build.webkit.org-config/README: Added.
* BuildSlaveSupport/build.webkit.org-config/Makefile:
* BuildSlaveSupport/build.webkit.org-config/buildbot.tac:
* BuildSlaveSupport/build.webkit.org-config/config.json: Added.
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
* BuildSlaveSupport/build.webkit.org-config/public_html/buildbot.css: Renamed from BuildSlaveSupport/build.webkit.org-config/buildbot.css.
* BuildSlaveSupport/build.webkit.org-config/public_html/index.html: Added.
* BuildSlaveSupport/build.webkit.org-config/public_html/robots.txt: Added.
* BuildSlaveSupport/build.webkit.org-config/webkit/__init__.py: Removed.
* BuildSlaveSupport/build.webkit.org-config/webkit/auth.py: Removed.
* BuildSlaveSupport/build.webkit.org-config/webkit/basesteps.py: Removed.
* BuildSlaveSupport/build.webkit.org-config/webkit/builders.py: Removed.
* BuildSlaveSupport/build.webkit.org-config/webkit/factories.py: Removed.
* BuildSlaveSupport/build.webkit.org-config/webkit/schedulers.py: Removed.
* BuildSlaveSupport/build.webkit.org-config/webkit/status.py: Removed.
* BuildSlaveSupport/build.webkit.org-config/webkit/steps.py: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42312 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Ignore the mysterious exception that Cygwin's Python is throwing when attempting
to remove the layout-test-results directory.

* BuildSlaveSupport/test-result-archive:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42311 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Kevin Ollivier.

Remove no longer needed debug warning and Windows-specific codepaths
(as we now use the Safari/Win port's timer there)

https://bugs.webkit.org/show_bug.cgi?id=25070

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42310 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Change the working directory before removing the test result directory
in the hopes of making the Windows build slaves happier.

* BuildSlaveSupport/test-result-archive:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42309 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Make the links to tests in the uploaded results.html files on build.webkit.org point to the test files.

Reviewed by Adam Roben.

* Scripts/VCSUtils.pm: Add a function that will return the location of a file relative to the root of the working copy.
* Scripts/run-webkit-tests: Convert the path in to a remote URL if the path is below the layout tests directory.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42308 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Adam Roben.

        - make this test insensitive to platform differences in Thai word
          breaking dictionaries

        * editing/selection/thai-word-at-document-end.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42307 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fix many <video> regression tests on Windows

        Reviewed by Dave Hyatt.

        * platform/graphics/win/QTMovieWin.cpp:
        (QTMovieWin::getNaturalSize): Call GetMovieNaturalBoundsRect, which
        returns the movie's natural bounds (duh), rather than GetMovieBox,
        which returns the movie's current displayed size/position.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42306 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove the layout-test-results directory after we have archived it so that
log files and results are reset between runs of the tests.

* BuildSlaveSupport/test-result-archive:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42305 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Check in files that were forgotten in r42293

        * platform/win/http/tests/xmlhttprequest/web-apps/012-expected.txt:
        * platform/win/http/tests/xmlhttprequest/web-apps/013-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42304 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Brady Eidson  <beidson@apple.com>

        Update layout test results for new exceptions.

        * fast/dom/Window/window-properties-expected.txt:
        * platform/mac-leopard/fast/dom/Window/window-properties-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42303 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-04-07  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin Adler and Sam Weinig

        <rdar://problem/5968249> "Private Browsing" doesn't affect HTML 5 LocalStorage

        The expected behavior for persistent storage when private browsing mode is undefined by the spec.
        For now we're going with a "read-only when in private browsing" policy until we can get feedback
        and/or get the behavior specified in the spec.

        Note that I purposefully made the change to restrict SessionStorage to read-only, also, with the
        understanding that the spec allows for SessionStorage to persist across relaunches, even though
        our implementation currently doesn't do this.

        * dom/DOMCoreException.idl: Add some new ExceptionCodes defined in HTML5, one of which is needed
          for LocalStorage changes (QUOTA_EXCEEDED_ERR)
        * dom/ExceptionCode.cpp:
        * dom/ExceptionCode.h:

        * storage/StorageArea.cpp:
        (WebCore::StorageArea::internalSetItem): If private browsing is enabled, throw the QUOTA_EXCEEDED_ERR
          exception as the spec allows.
        (WebCore::StorageArea::internalRemoveItem): If private browsing is enabled, silently fail to remove
          the item.
        (WebCore::StorageArea::internalClear): If private browsing is enabled, silently fail to clear the area.

LayoutTests:

2009-04-07  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin Adler

        Tests for <rdar://problem/5968249> "Private Browsing" doesn't affect HTML 5 LocalStorage

        * storage/domstorage/localstorage/private-browsing-affects-storage-expected.txt: Added.
        * storage/domstorage/localstorage/private-browsing-affects-storage.html: Added.
        * storage/domstorage/sessionstorage/private-browsing-affects-storage-expected.txt: Added.
        * storage/domstorage/sessionstorage/private-browsing-affects-storage.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42302 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Dean Jackson  <dino@apple.com>

        Reviewed by Simon Fraser.

        <rdar://problem/6148255>

        Standalone media documents can not always handle everything
        using a <video> element. In these cases switch to an <embed>
        element that uses the QuickTime plugin, in the hope that it
        will have more luck. Typical media types that trigger this
        are streamed media and QTVR.

        Equivalent Windows fix to come once reviewed. Note also that
        this area of code needs a cleanup:
        https://bugs.webkit.org/show_bug.cgi?id=25085

        * html/HTMLMediaElement.cpp:
        (WebCore::HTMLMediaElement::mediaPlayerSawUnsupportedTracks):
        * html/HTMLMediaElement.h:
        * loader/MediaDocument.cpp:
        (WebCore::MediaDocument::mediaElementSawUnsupportedTracks):
        (WebCore::MediaDocument::replaceVideoWithEmbed):
        * loader/MediaDocument.h:
            - removes the <video> element and puts in an <embed>
        * platform/graphics/MediaPlayer.h:
        (WebCore::MediaPlayerClient::mediaPlayerSawUnsupportedTracks):
            - new internal method for flagging a problem
        * platform/graphics/mac/MediaPlayerPrivateQTKit.h:
        * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
        (WebCore::MediaPlayerPrivate::MediaPlayerPrivate):
        (WebCore::MediaPlayerPrivate::updateStates):
            - looks for the types of content we don't handle
        (WebCore::MediaPlayerPrivate::rateChanged):
        (WebCore::MediaPlayerPrivate::sizeChanged):
        (WebCore::MediaPlayerPrivate::timeChanged):
        (WebCore::MediaPlayerPrivate::didEnd):
        (WebCore::MediaPlayerPrivate::disableUnsupportedTracks):
        (WebCore::MediaPlayerPrivate::sawUnsupportedTracks):
            - inform the MediaClient that we have a potential problem

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42301 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Anders Carlsson  <andersca@apple.com>

        Fix Tiger build for real this time.

        * Plugins/WebNetscapePluginView.h:
        * Plugins/WebNetscapePluginView.mm:
        (-[WebNetscapePluginView getVariable:forURL:value:length:]):
        (-[WebNetscapePluginView setVariable:forURL:value:length:]):
        (-[WebNetscapePluginView getAuthenticationInfoWithProtocol:host:port:scheme:realm:username:usernameLength:password:passwordLength:]):
        * Plugins/npapi.mm:
        (NPN_GetValueForURL):
        (NPN_SetValueForURL):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42300 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Helper script for uploading layout test results for display on build.webkit.org.

* BuildSlaveSupport/test-result-archive: Copied from WebKitTools/BuildSlaveSupport/built-product-archive.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42299 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        - fix <rdar://problem/6767428> Resurrect
          +[WebFontCache fontWithFamily:traits:size:] for clients that call it
          directly

        WebFontCache is a class in the WebCore framework and is not WebKit API,
        but apparently (perhaps mistakenly) some clients call this private
        interface.

        This change resurrects the interface and gives it reasonable behavior
        (however, note that WebCore will not call the old interface, so hacks that
        override it will have no effect on WebKit clients).

        * platform/mac/WebFontCache.h:
        * platform/mac/WebFontCache.mm:
        (+[WebFontCache fontWithFamily:traits:size:]): Added. Call through to
        +fontWithFamily:traits:weight:size:.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42298 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-04-07  David Hyatt  <hyatt@apple.com>

        Reviewed by Adam Roben

        Fix ScrollView's algorithm that determines whether or not to show scrollbars.  There was a bug
        with the method mainly because RenderView will size itself to the viewport's size, making it
        impossible to really know the document's "true" size.

        In order to solve this problem ScrollView now has a minimumContentsSize method.  This method
        returns the document size excluding the RenderView from the calculation.  This width/height
        is now cached on RenderView in m_docWidth/m_docHeight member variables.

        * WebCore.base.exp:
        * page/FrameView.cpp:
        (WebCore::FrameView::adjustPageHeight):
        (WebCore::FrameView::minimumContentsSize):
        * page/FrameView.h:
        * platform/ScrollView.cpp:
        (WebCore::ScrollView::updateScrollbars):
        * platform/ScrollView.h:
        (WebCore::ScrollView::minimumContentsSize):
        * rendering/RenderView.cpp:
        (WebCore::RenderView::layout):
        (WebCore::RenderView::calcDocHeight):
        (WebCore::RenderView::calcDocWidth):
        * rendering/RenderView.h:
        (WebCore::RenderView::docWidth):
        (WebCore::RenderView::docHeight):

WebKit/mac:

2009-04-07  David Hyatt  <hyatt@apple.com>

        Reviewed by Adam Roben

        Mac portion of fix to make DumpRenderTree always produce accurate scrollbar results.  Change
        updateScrollers to call minimumContentsSize when a WebHTMLView is inside the WebDynamicScrollbarsView.

        * WebView/WebDynamicScrollBarsView.m:
        (-[WebDynamicScrollBarsView updateScrollers]):

LayoutTests:

2009-04-07  David Hyatt  <hyatt@apple.com>

        Reviewed by Adam Roben

        Land updated layout test results now that DRT is doing the right thing
        with scrollbar results.

        * platform/mac/css2.1/t0803-c5502-mrgn-r-02-c-expected.checksum:
        * platform/mac/css2.1/t0803-c5502-mrgn-r-02-c-expected.png:
        * platform/mac/css2.1/t0803-c5502-mrgn-r-02-c-expected.txt:
        * platform/mac/css2.1/t0803-c5505-mrgn-02-c-expected.checksum:
        * platform/mac/css2.1/t0803-c5505-mrgn-02-c-expected.png:
        * platform/mac/css2.1/t0803-c5505-mrgn-02-c-expected.txt:
        * platform/mac/fast/block/float/013-expected.checksum:
        * platform/mac/fast/block/float/013-expected.png:
        * platform/mac/fast/block/float/013-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42297 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Dimitri Glazkov  <dglazkov@chromium.org>

        Build fix, not reviewed.

        * WebCore.xcodeproj/project.pbxproj: Made ScriptObject.h private.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42296 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Timothy Hatcher.

        https://bugs.webkit.org/show_bug.cgi?id=25063
        Refactor InspectorController to use ScriptObject/State.

        * GNUmakefile.am: Added ScriptState.cpp.
        * WebCore.pro: Ditto.
        * WebCore.vcproj/WebCore.vcproj: Ditto.
        * WebCore.xcodeproj/project.pbxproj: Ditto.
        * WebCoreSources.bkl: Ditto.
        * bindings/js/JSInspectorControllerCustom.cpp:
        (WebCore::JSInspectorController::addSourceToFrame): Fixed argument index typo.
        * bindings/js/ScriptFunctionCall.cpp:
        (WebCore::ScriptFunctionCall::call): Added reportExceptions parameter.
        (WebCore::ScriptFunctionCall::construct): Ditto.
        * bindings/js/ScriptFunctionCall.h: Ditto.
        * bindings/js/ScriptObject.cpp:
        (WebCore::handleException): Added exception-reporting heloper.
        (WebCore::ScriptObject::set): Changed to use handleException.
        (WebCore::ScriptGlobalObject::set): Ditto,
        * bindings/js/ScriptObject.cpp: Added ScriptGlobalObject.
        (WebCore::ScriptGlobalObject::set): Added.
        (WebCore::ScriptGlobalObject::getObject): Added.
        * bindings/js/ScriptObject.h: Added ScriptGlobalObject decls.
        (WebCore::ScriptGlobalObject::ScriptGlobalObject): Added.
        * bindings/js/ScriptState.cpp: Added.
        * bindings/js/ScriptState.h: Added scriptStateFromPage decl.
        * inspector/ConsoleMessage.cpp:
        (WebCore::ConsoleMessage::addToConsole): Changed to not report exceptions
            to avoid re-entrancy.
        * inspector/InspectorController.cpp: Refactored to use ScriptState/Object.
        * inspector/InspectorController.h: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42295 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Adam Langley  <agl@google.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=25083

        Skia was previously calculating the leading value incorrectly, leading
        us to add fHeight to get the height of a line of text. Now that Skia
        is calculating fLeading correctly, we can remove fHeight.

        This doesn't affect any layout tests.

        * platform/graphics/chromium/SimpleFontDataLinux.cpp:
        (WebCore::SimpleFontData::platformInit):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42294 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2009-04-07  Brian Weinstein  <bweinstein@gmail.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=24851

        This fix removes the status code from the status message in a ResourceResponse, which allows a few more tests in http to pass.

        * platform/network/cf/ResourceResponseCFNet.cpp:
        (WebCore::ResourceResponse::platformLazyInit):

LayoutTests:

2009-04-07  Brian Weinstein  <bweinstein@gmail.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=24851

        Tests in http/tests/xmlhttprequest/web-apps were taking results from
        platform/mac instead of LayoutTests, so added the values from LayoutTests
        into platform/win, and removed a few tests from the skip list that are fixed
        by the fix to platform/network/cf/ResourceResponseCFNet.cpp

        * platform/win/Skipped:
        * platform/win/http: Added.
        * platform/win/http/tests: Added.
        * platform/win/http/tests/xmlhttprequest: Added.
        * platform/win/http/tests/xmlhttprequest/web-apps: Added.
        * platform/win/http/tests/xmlhttprequest/web-apps/012-expected.txt: Copied from LayoutTests/http/tests/xmlhttprequest/web-apps/012-expected.txt.
        * platform/win/http/tests/xmlhttprequest/web-apps/013-expected.txt: Copied from LayoutTests/http/tests/xmlhttprequest/web-apps/013-expected.txt.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42293 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-11  Brian Weinstein  <bweinstein@gmail.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=24908

        Allows timeout to be set in run-webkit-tests as a command line argument.

        * Scripts/run-webkit-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42292 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dan Bernstein

        https://bugs.webkit.org/show_bug.cgi?id=25082

        Clean up repaint logic when RenderLayers become and stop being composited.

        * rendering/RenderLayer.h:
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::rendererContentChanged):
        (WebCore::RenderLayer::repaintIncludingNonCompositingDescendants):
        New compositing-only method that repaints a layer and all its non-composited descendants.
        Takes a repaintContainer for performance; all the repaints necessarily share the same
        repaintContainer.

        * rendering/RenderLayerCompositor.h:
        * rendering/RenderLayerCompositor.cpp:
        (WebCore::RenderLayerCompositor::updateLayerCompositingState):
        Rather than use calculateCompositedBounds() to compute the repaint rect (which gave
        the wrong answer when the composited layer tree was in flux), use the new
        repaintOnCompositingChange() method.

        (WebCore::RenderLayerCompositor::repaintOnCompositingChange):
        Call repaintIncludingNonCompositingDescendants(), and if necessary make sure
        that the view/layer drawing synchronization happens.

        (WebCore::RenderLayerCompositor::computeCompositingRequirements):
        Do a repaintOnCompositingChange() when a layer is going to be composited. This is
        a good place because we haven't started to change the compositing tree.

        (WebCore::RenderLayerCompositor::rebuildCompositingLayerTree):
        After we've finished updating all the descendant layers, we can do a repaint for
        layers that ceased compositing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42291 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Adam Langley  <agl@google.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=25081

        Skia is changing from reporting the XRange to reporting the XMin and
        XMax.  Obviously these are equivalent, but other users of the metrics
        need the two numbers to be separate.

        This doesn't affect any layout tests.

        * platform/graphics/chromium/SimpleFontDataLinux.cpp:
        (WebCore::SimpleFontData::platformInit):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42290 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Make the look of <select>s match the system look on Vista

        Reviewed by Dave Hyatt.

        * rendering/RenderThemeWin.cpp: Added new constants.
        (WebCore::RenderThemeWin::getThemeData): On Vista, use the
        CP_DROPDOWNBUTTONRIGHT part for drop down buttons.
        (WebCore::RenderThemeWin::paintMenuList): On Vista, use the CP_BORDER
        part to draw the border of <select>s.
        (WebCore::RenderThemeWin::paintMenuListButton): On Vista, outset the
        drop down button to make its border coincide with the <select>'s
        border.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42289 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Move isRunningOnVistaOrLater to a shared location

        Rubber-stamped in advance by Dave Hyatt.

        * WebCore.vcproj/WebCore.vcproj: Added SystemInfo.{cpp,h}, and let VS
        have its way with the rest of the file.
        * platform/win/ScrollbarThemeWin.cpp: Moved isRunningOnVistaOrLater
        from here...
        * platform/win/SystemInfo.cpp: Added.
        (WebCore::isRunningOnVistaOrLater): ...to here.
        * platform/win/SystemInfo.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42288 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Paul Godavari  <paul@chromium.org>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=24692
        Enable PopupMenuChromium to handle HTML select popups using native
        cocoa controls on the Mac, all other platforms are unchanged.

        We also split out the storage container for the popup menu items
        so that ChromeClientChromium can access them for forwarding to
        the embedding host (Chromium.app or test_shell).

        * page/chromium/ChromeClientChromium.h:
        * platform/chromium/PopupMenuChromium.cpp:
        (WebCore::PopupListBox::items):
        (WebCore::PopupContainer::PopupContainer):
        (WebCore::PopupContainer::~PopupContainer):
        (WebCore::PopupContainer::showPopup):
        (WebCore::PopupContainer::showExternal):
        (WebCore::PopupContainer::menuItemHeight):
        (WebCore::popupData):
        (WebCore::PopupListBox::pointToRowIndex):
        (WebCore::PopupListBox::getRowBounds):
        (WebCore::PopupListBox::isSelectableItem):
        (WebCore::PopupListBox::updateFromElement):
        (WebCore::PopupListBox::layout):
        (WebCore::PopupListBox::clear):
        (WebCore::PopupMenu::show):
        * platform/chromium/PopupMenuChromium.h:
        (WebCore::PopupItem::):
        (WebCore::PopupItem::PopupItem):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42287 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Geoffrey Garen  <ggaren@apple.com>

        Reviewed by Oliver Hunt.

        Added some double-checking to event handler marking. Now, when the
        wrapper that's responsible for marking an event handler's JavaScript
        function gets collected, it clears the event handler's JavaScript function
        so we won't dereference it after it's collected.

        In theory, we would never dereference a JavaScript event handler after
        its node's wrapper was collected anyway, but it's nice to be safe.

        * bindings/js/JSDOMApplicationCacheCustom.cpp:
        (WebCore::JSDOMApplicationCache::mark):
        * bindings/js/JSEventListener.cpp:
        (WebCore::JSEventListener::JSEventListener):
        (WebCore::JSEventListener::jsFunction):
        (WebCore::JSEventListener::invalidateJSFunction):
        (WebCore::JSEventListener::markJSFunction):
        * bindings/js/JSEventListener.h:
        * bindings/js/JSMessagePortCustom.cpp:
        (WebCore::JSMessagePort::mark):
        * bindings/js/JSNodeCustom.cpp:
        (WebCore::markEventListeners):
        * bindings/js/JSWorkerContextCustom.cpp:
        (WebCore::JSWorkerContext::mark):
        * bindings/js/JSWorkerCustom.cpp:
        (WebCore::JSWorker::mark):
        * bindings/js/JSXMLHttpRequestCustom.cpp:
        (WebCore::JSXMLHttpRequest::mark):
        * bindings/js/JSXMLHttpRequestUploadCustom.cpp:
        (WebCore::JSXMLHttpRequestUpload::mark):
        * bindings/scripts/CodeGeneratorJS.pm:
        * dom/EventListener.h:
        (WebCore::EventListener::invalidateJSFunction):
        (WebCore::EventListener::markJSFunction):
        (WebCore::markIfNotNull):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42286 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Gustavo Noronha Silva  <gustavo.noronha@collabora.co.uk>

        Reviewed by Anders Carlsson.

        Trying to fix Tiger build.

        * Plugins/WebNetscapePluginView.h:
        * Plugins/WebNetscapePluginView.mm:
        (-[WebNetscapePluginView getAuthenticationInfoWithProtocol:host:port:scheme:realm:username:usernameLength:password:passwordLength:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42285 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKit:

2009-04-07  Anders Carlsson  <andersca@apple.com>

        Try to fix the Leopard build once more.

        * WebKit.xcodeproj/project.pbxproj:

WebKit/mac:

2009-04-07  Anders Carlsson  <andersca@apple.com>

        Try to fix the Leopard build once more.

        * Plugins/WebBaseNetscapePluginView.h:
        * Plugins/WebBaseNetscapePluginView.mm:
        * Plugins/WebNetscapePluginView.mm:
        (-[WebNetscapePluginView getVariable:forURL:value:length:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42284 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  David Levin  <levin@chromium.org>

        Build fix.

        r44280 undid the change in r42268 that made CookieJar.h private.
        This redoes the change.

        * WebCore.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42283 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  David Levin  <levin@chromium.org>

        Reviewed by Sam Weinig and Geoff Garen.

        https://bugs.webkit.org/show_bug.cgi?id=25039
        UString refactoring to support UChar* sharing.

        No change in sunspider perf.

        * runtime/SmallStrings.cpp:
        (JSC::SmallStringsStorage::SmallStringsStorage):
        * runtime/UString.cpp:
        (JSC::initializeStaticBaseString):
        (JSC::initializeUString):
        (JSC::UString::BaseString::isShared):
        Encapsulate the meaning behind the refcount == 1 checks because
        this needs to do slightly more when sharing is added.
        (JSC::concatenate):
        (JSC::UString::append):
        (JSC::UString::operator=):
        * runtime/UString.h:
        Make m_baseString part of a union to get rid of casts, but make it protected because
        it is tricky to use it correctly since it is only valid when the Rep is not a BaseString.
        The void* will be filled in when sharing is added.

        Add constructors due to the making members protected and it make ensuring proper
        initialization work better (like in SmallStringsStorage).
        (JSC::UString::Rep::create):
        (JSC::UString::Rep::Rep):
        (JSC::UString::Rep::):
        (JSC::UString::BaseString::BaseString):
        (JSC::UString::Rep::setBaseString):
        (JSC::UString::Rep::baseString):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42282 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Jian Li  <jianli@chromium.org>

        Reviewed by Darin Fisher.

        Upstream changes to WorkerContextExecutionProxy for enabling V8 preemption.
        https://bugs.webkit.org/show_bug.cgi?id=25034

        * bindings/v8/WorkerContextExecutionProxy.cpp:
        (WebCore::WorkerContextExecutionProxy::evaluate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42281 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin Adler.

        - fix <rdar://problem/6764359> Thai word selection misbehaves for the
          last sentence in the document

        Test: editing/selection/thai-word-at-document-end.html

        * editing/visible_units.cpp:
            - Added a BoundarySearchContextAvailability enum used to indicate
            whether the caller to a boundary search function may be able to provide
            additional text in the search direction and call again.
            - Added a named type for the boundary search function signature. Changed
            the signature to take a context availability parameter and a boolean
            out parameter indicating whether more context is needed to perform
            the search.
        (WebCore::previousBoundary): If the beginning of the range was reached
            but the last search wanted more context, perform the search again, this
            time indicating that there is no earlier text.
        (WebCore::nextBoundary): Similarly, in the other direction.
        (WebCore::startWordBoundary): Check whether more context may be available
            and ask for more context if needed.
        (WebCore::endWordBoundary): Ditto.
        (WebCore::previousWordPositionBoundary): Ditto.
        (WebCore::nextWordPositionBoundary): Ditto.
        (WebCore::startSentenceBoundary):  Updated signature.
        (WebCore::endSentenceBoundary): Ditto.
        (WebCore::previousSentencePositionBoundary): Ditto.
        (WebCore::nextSentencePositionBoundary): Ditto.

LayoutTests:

        Reviewed by Darin Adler.

        - test for <rdar://problem/6764359> Thai word selection misbehaves for the
          last sentence in the document

        * editing/selection/thai-word-at-document-end-expected.txt: Added.
        * editing/selection/thai-word-at-document-end.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42280 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Geoffrey Garen  <ggaren@apple.com>

        Rubber stamped by Sam Weinig.

        A little renaming:

        function => jsFunction
        listener => jsFunction

        * bindings/js/JSEventListener.cpp:
        (WebCore::JSAbstractEventListener::handleEvent):
        (WebCore::JSEventListener::JSEventListener):
        (WebCore::JSEventListener::~JSEventListener):
        (WebCore::JSEventListener::jsFunction):
        (WebCore::JSEventListener::mark):
        (WebCore::JSProtectedEventListener::JSProtectedEventListener):
        (WebCore::JSProtectedEventListener::~JSProtectedEventListener):
        (WebCore::JSProtectedEventListener::jsFunction):
        * bindings/js/JSEventListener.h:
        (WebCore::JSProtectedEventListener::create):
        * bindings/js/JSLazyEventListener.cpp:
        (WebCore::JSLazyEventListener::jsFunction):
        (WebCore::JSLazyEventListener::parseCode):
        * dom/EventListener.h:
        (WebCore::EventListener::jsFunction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42279 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Anders Carlsson  <andersca@apple.com>

        ...and try to fix the Leopard build.

        * Plugins/npapi.mm:
        (NPN_GetAuthenticationInfo):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42278 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Brady Eidson  <beidson@apple.com>

        Reviewed by Darin Adler

        While working on <rdar://problem/5968249>, noticed some glaring problems with LocalStorage.

        * page/DOMWindow.cpp:
        (WebCore::DOMWindow::localStorage): Return the cached m_localStorage object if it exists to
          avoid creating multiple representations for the same underlying StorageArea.
        * page/DOMWindow.h:
        (WebCore::DOMWindow::optionalLocalStorage): Return m_localStorage, not m_sessionStorage.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42277 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Darin Adler  <darin@apple.com>

        Roll out incorrect build fix.

        * WebCore.NPAPI.exp: Reverted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42276 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Darin Adler  <darin@apple.com>

        Another build fix.

        * WebCore.NPAPI.exp: Added new functions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42275 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2009-04-07  Anders Carlsson  <andersca@apple.com>

        Another build fix.

        * plugins/npfunctions.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@42274 268f45cc-cd09-0410-ab3c-d52691b4dbfc