2008-08-13  Thiago Macieira  <tjmaciei@trolltech.com>

        Reviewed by Simon.

        Fix encoding of [ and ] in the host part of the URL

        Until QUrl is fixed (making QUrl's tolerant parser more tolerant), we have to
        add this workaround to the QUrl <> WebCore::KURL conversion operator so that it
        doesn't encode [ and ] when they are found in the host part. That is, the
        following URL:
        http://[::1]/
        is valid and should not be reencoded to:
        http://%5b::1%5d/

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35705 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-12  Sam Weinig  <sam@webkit.org>

        Reviewed by Anders Carlsson.

        * platform/mac/Skipped: Add fast/dom/cssTarget-crash.html to the skipped list.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35704 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add peephole optimisation to 'op_not... jfalse...' (eg. if(!...) )

Reviewed by Geoff Garen

This is a very slight win in sunspider, and a fairly substantial win
in hot code that does if(!...), etc.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35703 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Not reviewed.

Revert the last SVGLineElement commit. It's not needed, the first run on buildbot showed them as failures
the next build doesn't complain anymore (this was before adjusting the offsets.) No idea what's going on,
but if it's fixed I'm happy.

While I'm at it, regenerate struct-image-02-b.svg (invisible differences) to have a clean baseline.
All SVG pixel test results are accurate on Leopard now.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35702 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Not reviewed. The last SVGLineElement* fix attempt, fixed 4/8 failing tests.
Trying two variations in two tests to find out the right offset to make the other four tests work on Tiger.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35701 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Eric.
Add new dynamic-update layout tests covering SVGRectElement.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35700 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Not reviewed. Trying to modify mouseevent target location, to work around known CG/Tiger hit test problem.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35699 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Sam.
Forgot to include these in my last commit.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35698 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

2008-08-12  Mihnea Ovidenie  <mihnea@adobe.com>

        Fix for https://bugs.webkit.org/show_bug.cgi?id=19891
        Broken HTML object elements cause de-reference of pointer to freed memory.
        If we fail to load an image for an object tag and we no longer believe the object tag points at
        an image, then clear m_imageLoader in the HTMLObjectElement so that we attempt to render the
        fall back content.

        Reviewed by Dave Hyatt and Alexey Proskuryakov.

        Test: http/tests/misc/object-image-error-with-onload.html

        * html/HTMLObjectElement.cpp:
        (WebCore::HTMLObjectElement::renderFallbackContent):
        * page/Frame.cpp:
        (WebCore::Frame::Frame):

LayoutTests:

2008-08-12  Mihnea Ovidenie  <mihnea@adobe.com>

        Test for https://bugs.webkit.org/show_bug.cgi?id=19891
        Broken HTML object elements cause de-reference of pointer to freed memory.

        Reviewed by Dave Hyatt and Alexey Proskuryakov.

        * http/tests/misc/object-image-error-with-onload-expected.txt: Added.
        * http/tests/misc/object-image-error-with-onload.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35697 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Eric.

Fix typo in most files: "mesage" -> "message".
Remove unused code in SVGTestCase.js

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35696 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Eric.

Add new dynamic-update layout tests covering SVGPatternElement.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35695 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Dave.

Fixes: https://bugs.webkit.org/show_bug.cgi?id=19798
Masks are translated, and the mask images are swapped on the y-axis.

Turned out that http://trac.webkit.org/changeset/31830/trunk/WebCore/svg/graphics/cg/SVGResourceMaskerCg.mm
is guilty. GraphicsContext::clipToImageBuffer() does some extra transformations that SVGResourcesMaskerCg does not want.

Long term goal is to remove the SVGResource*/SVGPaintServer* classes anyway, so it's okay to duplicate
the "clip to image buffer" functionality, in the CG specific SVGResourceMaskerCg class - as it was before.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35694 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Sam.

Speculative fix for hit testing problems in dynamic-updates/SVGLineElement* on Tiger.
Supposed to fix the hanging tests on the buildbot.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35693 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-12  Sam Weinig  <sam@webkit.org>

        * fast/dom/Window/window-properties-expected.txt: Update results.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35692 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScriptCore:

        - JavaScriptCore part of <rdar://problem/6121636>
          Make fast*alloc() abort() on failure and add "try" variants that
          return NULL on failure.

        Reviewed by Darin Adler.

        * JavaScriptCore.exp: Exported tryFastCalloc().
        * VM/RegisterFile.h:
        (KJS::RegisterFile::RegisterFile): Removed an ASSERT().
        * kjs/JSArray.cpp:
        (KJS::JSArray::putSlowCase): Changed to use tryFastRealloc().
        (KJS::JSArray::increaseVectorLength): Ditto.
        * kjs/ustring.cpp:
        (KJS::allocChars): Changed to use tryFastMalloc().
        (KJS::reallocChars): Changed to use tryFastRealloc().
        * wtf/FastMalloc.cpp:
        (WTF::fastZeroedMalloc): Removed null checking of fastMalloc()'s result
        and removed extra call to InvokeNewHook().
        (WTF::tryFastZeroedMalloc): Added. Uses tryFastMalloc().
        (WTF::tryFastMalloc): Renamed fastMalloc() to this.
        (WTF::fastMalloc): Added. This version abort()s if allocation fails.
        (WTF::tryFastCalloc): Renamed fastCalloc() to this.
        (WTF::fastCalloc): Added. This version abort()s if allocation fails.
        (WTF::tryFastRealloc): Renamed fastRealloc() to this.
        (WTF::fastRealloc): Added. This version abort()s if allocation fails.
        (WTF::do_malloc): Made this a function template. When the abortOnFailure
        template parameter is set, the function abort()s on failure to allocate.
        Otherwise, it sets errno to ENOMEM and returns zero.
        (WTF::TCMallocStats::fastMalloc): Defined to abort() on failure.
        (WTF::TCMallocStats::tryFastMalloc): Added. Does not abort() on
        failure.
        (WTF::TCMallocStats::fastCalloc): Defined to abort() on failure.
        (WTF::TCMallocStats::tryFastCalloc): Added. Does not abort() on
        failure.
        (WTF::TCMallocStats::fastRealloc): Defined to abort() on failure.
        (WTF::TCMallocStats::tryFastRealloc): Added. Does not abort() on
        failure.
        * wtf/FastMalloc.h: Declared the "try" variants.

WebCore:

        - WebCore part of <rdar://problem/6121636>
          Make fast*alloc() abort() on failure and add "try" variants that
          return NULL on failure.

        Reviewed by Darin Adler.

        * platform/Arena.cpp:
        (WebCore::ArenaAllocate): Removed null checking of fastMalloc()'s
        result.
        * platform/graphics/cg/ImageBufferCG.cpp:
        (WebCore::ImageBuffer::create): Changed to use tryFastCalloc().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35691 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin Adler.

        - fix https://bugs.webkit.org/show_bug.cgi?id=19348
          <rdar://problem/5978447> REGRESSION (r34193): Setting the size of a frame with javascript document.body.row no longer works

        Test: fast/frames/frameset-style-recalc.html

        * html/HTMLFrameSetElement.cpp:
        (WebCore::HTMLFrameSetElement::recalcStyle): Changed to call the base
        class implementation after marking for layout.

LayoutTests:

        Reviewed by Darin Adler.

        - test for https://bugs.webkit.org/show_bug.cgi?id=19348
          <rdar://problem/5978447> REGRESSION (r34193): Setting the size of a frame with javascript document.body.row no longer works

        * fast/frames/frameset-style-recalc.html: Added.
        * platform/mac/fast/frames/frameset-style-recalc-expected.checksum: Added.
        * platform/mac/fast/frames/frameset-style-recalc-expected.png: Added.
        * platform/mac/fast/frames/frameset-style-recalc-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35690 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Remove unneeded header imports from some Web Inspector files.

        Reviewed by Adam Roben.

        * WebCoreSupport/WebInspectorClient.mm:
        * WebInspector/WebInspector.mm:
        * WebInspector/WebNodeHighlightView.m:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35689 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Remove the call to WKNSWindowMakeBottomCornersSquare on the
        Web Inspector's window. This isn't needed anymore since the
        window style masks used always have square bottom corners.

        Reviewed by Adam Roben.

        * WebCoreSupport/WebInspectorClient.mm:
        (-[WebInspectorWindowController window]): Remove the call
        to WKNSWindowMakeBottomCornersSquare.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35688 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Make attaching and detaching the Web Inspector instantaneous.
        This also preserves the current view, other state, and keeps
        the script debugger attached.

        https://bugs.webkit.org/show_bug.cgi?id=19301

        Reviewed by Adam Roben.

        * WebCoreSupport/WebInspectorClient.mm:
        (-[WebInspectorWindowController close]): Only call setWindowVisible(false)
        when not moving windows.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35687 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Export WTFLog in Debug builds

        Reviewed by John Sullivan.

        * WebKit.vcproj/WebKit_debug.def:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35686 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by John Sullivan.

        - move shouldUpdateWhenOffscreen from Settings to FrameView and rename it shouldUpdateWhileHidden

        * WebCore.base.exp:
        * page/FrameView.cpp:
        (WebCore::FrameViewPrivate::FrameViewPrivate):
        (WebCore::FrameView::shouldUpdateWhileHidden):
        (WebCore::FrameView::setShouldUpdateWhileHidden):
        * page/FrameView.h:
        * page/Settings.cpp:
        * page/Settings.h:

WebKit/mac:

        Reviewed by John Sullivan.

        - replace -[WebPreferences updatesWhenOffscreen] with -[WebView shouldUpdateWhileHidden]

        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::transitionToCommittedForNewPage):
        * WebView/WebFrame.mm:
        (-[WebFrame _updateBackgroundAndUpdatesWhileHidden]):
        * WebView/WebFrameInternal.h:
        * WebView/WebHTMLView.mm:
        (-[WebHTMLView windowWillOrderOnScreen:]):
        * WebView/WebPreferenceKeysPrivate.h:
        * WebView/WebPreferences.mm:
        (+[WebPreferences initialize]):
        * WebView/WebPreferencesPrivate.h:
        * WebView/WebView.mm:
        (-[WebViewPrivate init]):
        (-[WebView _preferencesChangedNotification:]):
        (-[WebView setBackgroundColor:]):
        (-[WebView setDrawsBackground:]):
        (-[WebView shouldUpdateWhileHidden]):
        (-[WebView setShouldUpdateWhileHidden:]):
        * WebView/WebViewPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35685 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Windows build fix

        * bindings/js/JSSVGPODTypeWrapper.h: Align
        JSSVGDynamicPODTypeWrapper's and JSSVGStaticPODTypeWrapperWithParent's
        members on 16-byte boundaries to avoid an alignment warning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35684 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Anders.

Add new dynamic-update layout tests covering SVGCursorElement.

Scale svg/webarchive/resources/green-checker.png to a value < 128x128,
as this is the maximum hardcoded cursor size (fixes manual testing of svg-cursor-subresources.png).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35683 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Oliver.

Add new dynamice-update layout tests covering SVGMaskElement.
Move pixel tests results of SVGMarkerElement* to the right positions, while I'm at it.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35682 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-12  Urs Wolfer  <uwolfer@kde.org>

        Reviewed by Simon.

        https://bugs.webkit.org/show_bug.cgi?id=20357

        Fix crash in QWebPage in case contextMenuEvent has been overwritten
        because context menu has the view as parent and thus is deleted too early.

        Add testcase for this crash.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35681 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Oliver.

Add new dynamice-update layout tests covering SVGMarkerElement.
Fix bug: SVGMarkerElement's SVG DOM function calls don't update rendering.
Fix orientAngle/orientType confusion: "auto" orient should always return "0" as angle.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35680 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Oliver.

Add new dynamic-update layout tests covering SVGImageElement.
Fix bug: SVGImageElement doesn't react on 'preserveAspectRatio' changes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35679 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Changed Option/Alt-Up or Down in CSS editing when the value is
        near zero to jump to the next integer.

        Reviewed by Tim Hatcher.

        https://bugs.webkit.org/show_bug.cgi?id=20326

        * page/inspector/StylesSidebarPane.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35678 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Changed the line highlight transition for an easier animation.

        Reviewed by Tim Hatcher.

        * page/inspector/SourceFrame.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35677 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Added support for some Firebug Command Line APIs.

        Reviewed by Tim Hatcher.

        https://bugs.webkit.org/show_bug.cgi?id=19867
        https://bugs.webkit.org/show_bug.cgi?id=19868
        https://bugs.webkit.org/show_bug.cgi?id=19869
        https://bugs.webkit.org/show_bug.cgi?id=19875
        https://bugs.webkit.org/show_bug.cgi?id=19876
        https://bugs.webkit.org/show_bug.cgi?id=19880

        * page/inspector/Console.js:
        (WebInspector.Console.prototype._evalInInspectedWindow):
        Create an object on the inspected window that holds the console
        command line API functions. This object is used in a with statement
        around the typed expression.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35676 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rubber stamped by Oliver.

Add a bunch of new dynamic-updates tests covering SVGLineElement & SVGRadialGradientElement.
No functional changes needed, all work out of the box.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35675 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Antti.

Fixes: http://bugs.webkit.org/show_bug.cgi?id=17736
JS wrapper objects around SVG POD types, that contain other SVG POD types with writable properties
failed to update. Modification of the values were completly ignored (ie. transform.matrix.a = 50, didn't take any effect)

Added tests: svg/custom/svg-modify-currentTranslate.html
             svg/custom/tearoffs-with-tearoffs.html
             svg/custom/immutable-properties.html

Fixed tests: svg/dynamic-updates/SVGLinearGradientElement-svgdom-gradientTransform-prop.html

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35674 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-11  Brady Eidson  <beidson@apple.com>

        Reviewed by John and Anders

        Fix for <rdar://problem/6141797>

        When WebArchives were entirely a WebKit concept, there was a guarantee that a WebResource
        would never have nil data.

        When they were pushed down into WebCore, that guarantee was lost, subtly changing a few
        semantics with some WebKit applications.

        The guarantee was a good one and should be restored.

        Note that ApplicationCacheResource doesn't need any updates to follow this rule as it already
        creates an empty data object in the case of null data for its own purposes.

        * loader/SubstituteResource.h:
        (WebCore::SubstituteResource::SubstituteResource): ASSERT that the data is not null.  This
          well help any future subclassers not make this mistake.

        * loader/archive/ArchiveResource.cpp:
        (WebCore::ArchiveResource::create): Return 0 if the data is null.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35673 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-11  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dave Hyatt

        https://bugs.webkit.org/show_bug.cgi?id=20328
        Fix a problem when an 'all' transition transition with more than
        one property changing is interrupted, and did some AnimationController
        cleanup.

        Test: transitions/interrupted-all-transition.html

        * page/AnimationController.cpp:
        (WebCore::ImplicitAnimation::ImplicitAnimation):
        (WebCore::AnimationControllerPrivate::blendProperties):
        (WebCore::CompositeAnimation::updateTransitions):
        (WebCore::CompositeAnimation::cleanupFinishedAnimations):
        (WebCore::CompositeAnimation::setTransitionStartTime):
        (WebCore::CompositeAnimation::overrideImplicitAnimations):
        (WebCore::CompositeAnimation::resumeOverriddenImplicitAnimations):
        (WebCore::ImplicitAnimation::animate):
        (WebCore::ImplicitAnimation::onAnimationEnd):
        (WebCore::ImplicitAnimation::sendTransitionEvent):
        (WebCore::ImplicitAnimation::affectsProperty):
        (WebCore::KeyframeAnimation::endAnimation):
        (WebCore::KeyframeAnimation::onAnimationEnd):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35672 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-11  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Tim.

        - Because console messages have group levels now, newly created messages
        that do not specify the level lose their message since the number of
        arguments is wrong.

        * page/inspector/Console.js:
        * page/inspector/Resource.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35671 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-11  Alp Toker  <alp@nuanti.com>

        Build fix. Add new files from r35666 (WebKitAnimationEvent). Also take
        the opportunity to sort the sources lists.

        * GNUmakefile.am:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35670 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Speed up the the JavaScript syntax highlighter by generating
        the finders only once per script instead of per line.

        https://bugs.webkit.org/show_bug.cgi?id=20346

        Reviewed by Adam Roben.

        * page/inspector/SourceFrame.js:
        (WebInspector.SourceFrame.prototype._syntaxHighlightJavascriptLine):
        Removed, factored into syntaxHighlightJavascript as an inline function.
        (WebInspector.SourceFrame.prototype.syntaxHighlightJavascript):
        Pulled in the _syntaxHighlightJavascriptLine so it will create a closure.
        Generate the finders before iterating the lines.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35669 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Windows build fix

        * WebCore.vcproj/WebCore.vcproj: Added JSWebKitAnimationEvent.cpp and
        JSWebKitTransitionEvent.cpp to the project. Let VS reformat the file,
        too.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35668 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin Adler.

        - fix <rdar://problem/6131096> Reproducible crash in CounterNode::isReset under guard malloc

        Test: fast/css/counters/invalidate-cached-counter-node.html

        * rendering/RenderContainer.cpp:
        (WebCore::RenderContainer::invalidateCounters): Added. Invalidates all
        RenderCounters in :before and :after content.
        * rendering/RenderContainer.h:
        * rendering/RenderCounter.cpp:
        (WebCore::RenderCounter::isCounter): Renamed isRenderCounter() to this
        to match the RenderObject method.
        (WebCore::RenderCounter::invalidate): Added. Resets the cached
        CounterNode and invalidates the object's layout and preferred widths.
        (WebCore::destroyCounterNodeChildren): Added a call to
        invalidateCounters().
        * rendering/RenderCounter.h:
        * rendering/RenderObject.h:
        (WebCore::RenderObject::invalidateCounters):

LayoutTests:

        Reviewed by Darin Adler.

        - test for <rdar://problem/6131096> Reproducible crash in CounterNode::isReset under guard malloc

        * fast/css/counters/invalidate-cached-counter-node.html: Added.
        * platform/mac/fast/css/counters/invalidate-cached-counter-node-expected.checksum: Added.
        * platform/mac/fast/css/counters/invalidate-cached-counter-node-expected.png: Added.
        * platform/mac/fast/css/counters/invalidate-cached-counter-node-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35667 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Implement CSS Animation and Transition Events
https://bugs.webkit.org/show_bug.cgi?id=20337

Implement the events defined in the CSS Animations
and Transitions specifications so code can react
to animations and transitions.

Reviewed by Tim Hatcher and Dave Hyatt.

* DerivedSources.make:
* GNUmakefile.am:
* WebCore.pro:
* WebCore.vcproj/WebCore.vcproj:
* WebCore.xcodeproj/project.pbxproj:
* WebCoreSources.bkl:
    Build configs for new files

* bindings/js/JSDOMWindowBase.cpp:
* bindings/js/JSDOMWindowBase.h:
* bindings/js/JSEventCustom.cpp:
* dom/Document.h:
* dom/Event.cpp:
* dom/Event.h:
* dom/EventTarget.cpp:
* dom/EventTargetNode.cpp:
* dom/EventTargetNode.h:
* html/HTMLElement.cpp:
* page/AnimationController.cpp:
    do all the new event stuff

* html/HTMLAttributeNames.in:
    the onwebkitanimation* and onwebkittransitionend attrs

* dom/WebKitAnimationEvent.cpp: Added.
* dom/WebKitAnimationEvent.h: Added.
* dom/WebKitAnimationEvent.idl: Added.
* dom/WebKitTransitionEvent.cpp: Added.
* dom/WebKitTransitionEvent.h: Added.
* dom/WebKitTransitionEvent.idl: Added.
    New files for the events

* manual-tests/transition-events.html: Added.
    New testfile

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35666 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Move WTF::notFound into its own header so that it can be used independently of Vector

JavaScriptCore:

        Move WTF::notFound into its own header so that it can be used
        independently of Vector

        Rubberstamped by Darin Adler.

        * JavaScriptCore.vcproj/WTF/WTF.vcproj:
        * JavaScriptCore.xcodeproj/project.pbxproj:
        Added NotFound.h to the project.
        * wtf/NotFound.h: Added. Moved the notFound constant here...
        * wtf/Vector.h: ...from here.

WebCore:

        Add a ForwardingHeader for wtf/NotFound.h

        Rubberstamped by Darin Adler.

        * ForwardingHeaders/wtf/NotFound.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35665 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Fixes a bug where error bubbles in JavaScript resources would
        be clobbered by the syntax highlighter.

        https://bugs.webkit.org/show_bug.cgi?id=20345

        Reviewed by Adam Roben.

        * manual-tests/inspector/resources/script-error.js: Added.
        * manual-tests/inspector/styled-error-bubbles-in-scripts.html: Added.
        * page/inspector/SourceFrame.js:
        (WebInspector.SourceFrame.prototype._addMessageToSource):
        Check the nodeType and not the nodeName, this is less fragile.
        (WebInspector.SourceFrame.prototype._syntaxHighlightJavascriptLine):
        Check if the lastChild is an error bubble, if so remove it before
        getting the line's textContent. Add the error bubble back at the end.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35664 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Mark Rowe.

        <rdar://problem/6130393> REGRESSION: PhotoBooth hangs after launching under TOT Webkit

        * API/JSContextRef.cpp: (JSGlobalContextRelease): Corrected a comment.

        * kjs/collector.cpp: (KJS::Heap::~Heap): Ensure that JSGlobalData is not deleted while
        sweeping the heap.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35663 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Roll over three large ChangeLogs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35662 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Eric Seidel.

        - fix https://bugs.webkit.org/show_bug.cgi?id=20339
          REGRESSION (r35531-r35615): Acid3 crashes on Windows in CachedFont::getSVGFontById

        * dom/make_names.pl: Removed "using namespace WebCore::*Names"
        from *ElementFactory.cpp files.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35661 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix crash in Acid3 in Windows and Gtk ports, introduced in r35590.

Reviewed by Cameron Zwarich.

* bindings/js/ScriptControllerGtk.cpp: Fix the logic.
* bindings/js/ScriptControllerWin.cpp: Ditto.
* bindings/js/ScriptControllerQt.cpp: Switch to isPluginView for consistency with Gtk and Windows.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35660 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Reviewed by Darin Adler.

        - fix <rdar://problem/6074587> REGRESSION (34722): Unable to view AT&T wireless bill - window title contains markup, window is empty

        Test: fast/tokenizer/ampersand-in-special-tag.html

        * html/HTMLTokenizer.cpp:
        (WebCore::HTMLTokenizer::parseSpecial): Changed to only advance
        lastDecodedEntityPosition if an entity was found and decoded.

LayoutTests:

        Reviewed by Darin Adler.

        - test for <rdar://problem/6074587> REGRESSION (34722): Unable to view AT&T wireless bill - window title contains markup, window is empty

        * fast/tokenizer/ampersand-in-special-tag-expected.txt: Added.
        * fast/tokenizer/ampersand-in-special-tag.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35659 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-10  Jan Michael Alonzo  <jmalonzo@webkit.org>

        Reviewed (and updated) by Alp Toker.

        https://bugs.webkit.org/show_bug.cgi?id=16620
        [GTK] Autotools make dist and make check support

        Get make dist working.

        Note that not all possible configurations have been tested yet.

        * GNUmakefile.am:

2008-08-10  Alp Toker  <alp@nuanti.com>

        Remove leftover qmake/GTK+ build files.

        * DumpRenderTree/gtk/DumpRenderTree.pro: Removed.
        * GtkLauncher/GtkLauncher.pro: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35658 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Sam Weinig.

        Added same heap debug checks to more code paths.

        * kjs/JSActivation.cpp:
        (KJS::JSActivation::put):
        (KJS::JSActivation::putWithAttributes):
        * kjs/JSGlobalObject.cpp:
        (KJS::JSGlobalObject::putWithAttributes):
        * kjs/JSObject.h:
        (KJS::JSObject::putDirect):
        * kjs/JSVariableObject.h:
        (KJS::JSVariableObject::symbolTablePut):
        (KJS::JSVariableObject::symbolTablePutWithAttributes):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35657 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-09  Marco Barisione  <marco.barisione@collabora.co.uk>

        Reviewed by Alp Toker.

        http://bugs.webkit.org/show_bug.cgi?id=19045
        [gtk] ./configure doesn't check for x toolkit availability

        Some old versions of Xt do not provide xt.pc, in this case use
        AC_CHECK_LIB to check if the library is installed and then fall back
        to using just -lXt.

        * configure.ac:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35656 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-09  Alp Toker  <alp@nuanti.com>

        GTK+ build fixes.

        Back out r20338 libxslt check (conditional check already exists and
        this broke the build).

        Make the flex version check a warning rather than fatal to fix the
        build bot.

        * configure.ac:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35655 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Jan Michael Alonzo  <jmalonzo@webkit.org>

        Reviewed by Mark Rowe.

        http://bugs.webkit.org/show_bug.cgi?id=20338
        [Gtk] Autotools should check for libxslt

        Autotools check for libxslt

        * configure.ac:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35654 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-09  Cameron Zwarich  <cwzwarich@uwaterloo.ca>

        Reviewed by Maciej.

        Fix some style issues in the sampling tool.

        * VM/SamplingTool.cpp:
        (KJS::sleepForMicroseconds):
        (KJS::SamplingTool::dump):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35653 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-09  Cameron Zwarich  <cwzwarich@uwaterloo.ca>

        Reviewed by Oliver.

        Revision 35651, despite being a rather trivial change, introduced a
        large regression on the regexp-dna SunSpider test. This regression
        stemmed from an increase in the size of CodeBlock::dump(). There is
        no reason for this method (and several related methods) to be compiled
        in non-debug builds with the sampling tool disabled. This patch
        conditionally compiles them, reversing the regression on SunSpider.

        * JavaScriptCore.exp:
        * VM/CodeBlock.cpp:
        * VM/CodeBlock.h:
        * VM/Machine.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35652 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Cameron Zwarich  <cwzwarich@uwaterloo.ca>

        Reviewed by Oliver.

        Bug 20330: JSCore crash loading any filehurricane media page
        <https://bugs.webkit.org/show_bug.cgi?id=20330>

        Fix a typo in the constant loading patch. Also, add a case for
        op_unexpected_load to CodeBlock::dump().

        JavaScriptCore:

        * VM/CodeBlock.cpp:
        (KJS::CodeBlock::dump):
        * VM/CodeGenerator.cpp:
        (KJS::CodeGenerator::addUnexpectedConstant):

        LayoutTests:

        * fast/js/resources/unexpected-constant-crash.js: Added.
        * fast/js/unexpected-constant-crash-expected.txt: Added.
        * fast/js/unexpected-constant-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35651 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by John Sullivan.

        - fix <rdar://problem/6130216> Exception "windowRegionBeingDrawn != nil" in NSView when caching image of a subframe

        This change reintroduces <https://bugs.webkit.org/show_bug.cgi?id=5195> on Leopard.

        * WebView/WebHTMLView.mm:
        (-[WebHTMLView _recursive:displayRectIgnoringOpacity:inContext:topView:]): Removed an
        unsuccessful workaround for <rdar://problem/5668489>, because invoking
        layout may change the view hierarchy during the drawing operation,
        which is not supported on Leopard.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35650 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Beth Dakin  <bdakin@apple.com>

        Reviewed by Darin.

        This patch makes the findString and markAllMatchesForText functions
        work with disconnected frames that are contained within
        overflow:hidden blocks.

        * editing/Editor.cpp:
        (WebCore::Editor::rangeVisibility):
        (WebCore::Editor::firstVisibleRange):
        (WebCore::Editor::lastVisibleRange):
        * editing/Editor.h:
        (WebCore::Editor::):
        * page/Frame.cpp:
        (WebCore::Frame::findString):
        (WebCore::Frame::markAllMatchesForText):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35649 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Maxime Britto  <britto@apple.com>

        Reviewed by Adele.

        Test: fast/events/scroll-to-anchor-in-overflow-hidden.html
        https://bugs.webkit.org/show_bug.cgi?id=20270
        Jump to an anchor wasn't working when the overflow:hidden CSS attribute was set on the page.

        * WebCore.base.exp:
        * dom/Element.cpp:
        (WebCore::Element::scrollIntoView):
        (WebCore::Element::scrollIntoViewIfNeeded):
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::gotoAnchor):
        * page/EventHandler.cpp:
        (WebCore::EventHandler::handleMousePressEvent):
        (WebCore::EventHandler::handleMouseDraggedEvent):
        * page/Frame.cpp:
        (WebCore::Frame::revealSelection):
        (WebCore::Frame::revealCaret):
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::scrollRectToVisible): Takes a new parameter "scrollToAnchor" since it's a special case of scroll. It passes this new bool to canBeProgrammaticallyScrolled.
        (WebCore::RenderLayer::autoscroll):
        * rendering/RenderLayer.h:
        * rendering/RenderListBox.h:
        (WebCore::RenderListBox::canBeProgramaticallyScrolled):
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::canBeProgramaticallyScrolled): If the call is coming from a scrollToAnchor we don't need scrollBars to accept.
        * rendering/RenderObject.h:
        * rendering/RenderTextControl.h:
        (WebCore::RenderTextControl::canBeProgramaticallyScrolled):

2008-08-08  Maxime Britto  <britto@apple.com>

        Reviewed by Adele.

        * WebView/WebFrame.mm:
        (-[WebFrame _scrollDOMRangeToVisible:]):

2008-08-08  Maxime Britto  <britto@apple.com>

        Reviewed by Adele.
        https://bugs.webkit.org/show_bug.cgi?id=20270
        This test is automatic and can also be run manually.

        * fast/events/scroll-to-anchor-in-overflow-hidden-expected.txt: Added.
        * fast/events/scroll-to-anchor-in-overflow-hidden.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35648 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Maxime Britto  <britto@apple.com>

        Reviewed by Adele.

        <rdar://problem/6137022>
        <https://bugs.webkit.org/show_bug.cgi?id=20331>
        REGRESSION(r35177-r35203): Scrolling slowly over a scrollable field scrolls the page instead of the field (20331)

        * page/EventHandler.cpp:
        (WebCore::scrollAndAcceptEvent): fixed the wrong cast for the wheel delta which was truncated to 0 on slow wheel scrolls.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35647 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Execute CSS Animations using new AnimationController
https://bugs.webkit.org/show_bug.cgi?id=20119

I still need to file some follow-on bugs that EricS noticed.

Reviewed by Dave Hyatt

Some new tests:
* manual-tests/animate-duration.html: Added.
* manual-tests/animate-left.html: Added.
* manual-tests/animate-none.html: Added.

* css/CSSStyleSelector.cpp:
* dom/Document.cpp:
    - don't redo style when page is place into cache
* page/AnimationController.cpp:
* page/AnimationController.h:
    - execute the animations
* rendering/RenderObject.cpp:
* rendering/style/RenderStyle.h:
    - make border radii const (for blending)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35646 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Tim.

        <rdar://problem/6136326> JSProfiler: Start/stop profiling button in the
        Web Inspector does not match the menu.

        * page/InspectorController.cpp:
        (WebCore::InspectorController::toggleRecordButton): Toggle the record
        button in the webInspector.
        (WebCore::InspectorController::startUserInitiatedProfiling):
        (WebCore::InspectorController::stopUserInitiatedProfiling):
        * page/InspectorController.h:
        * page/inspector/ProfilesPanel.js: Move the UI part of the record button
        into its funtion so that it can be called from the Inspector controller.
        * page/inspector/inspector.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35645 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Matt Lilek  <webkit@mattlilek.com>

        Not reviewed, build fix.

        * JavaScriptCore.exp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35644 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Brady Eidson  <beidson@apple.com>

        Reviewed by Sam Weinig

        Fix for <rdar://problem/5865504>

        This bug was actually *largely* fixed by http://trac.webkit.org/changeset/35538.
        But with that same patch, it became possible for a WebResource to fail to initialize.
        Therefore we were trying to add nil to an NSCFArray for certain situations, which is bad.
        Lets fix that, shall we?

        * WebView/WebArchive.mm:
        (-[WebArchive subresources]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35643 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Tim.

        <rdar://problem/6131187> Profiles don't show up immediately when run
        with the UI.

        * page/InspectorController.cpp:
        (WebCore::startProfiling): Added so that the inspector's JS code could
        call through to the same function used by the Develop menu.  This way
        both can call didFinishAllExecution() which puts the profile in the
        Web Inspector.
        (WebCore::stopProfiling): Ditto
        (WebCore::InspectorController::windowScriptObjectAvailable):
        (WebCore::InspectorController::stopUserInitiatedProfiling): Now calls
        didFinishAllExecution() to put the profile in the Web Inspector.  This
        should be safe since the UIs shouldn't work if JS is still running.
        * page/inspector/ProfilesPanel.js: Use the new functions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35642 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Darin Adler.

        - fix <rdar://problem/5961260> GDI text: transparent fill color text does not render if it it has a stroke or shadow

        * platform/graphics/win/FontCGWin.cpp:
        (WebCore::Font::drawGlyphs):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35641 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Improve performance of arithmetic operators

Reviewed by Cameron Zwarich

Added a fast (non-virtual) mechanism to determine if a non-immediate JSValue*
is a JSNumberCell.  We then use this to allow improved specialisation in many
arithmetic operators.  SunSpider reports a 2.5% progression overall, with greater
than 10% progressions on a number of arithmetic heavy tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35639 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-08  Adam Treat  <treat@kde.org>

        Reviewed by Simon.

        Return the canonical file path in Qt's listDirectory as WebCore::PluginDatabase
        uses this method to retrieve a list of possible plugins.  Many linux systems
        have nested symbolic links scattered throughout the plugin directories which
        resulted in the PluginDatabase loading the same plugin N times depending on
        the number of symbolic links.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35638 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix bug numbers.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35636 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Change highlight time from 300ms to 500ms and fade effect from 1s to 2s

        https://bugs.webkit.org/show_bug.cgi?id=17589

        Reviewed by Timothy Hatcher.

        * page/inspector/SourceFrame.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35633 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    The search is only performed if the field contains more than 3
        letters or if Enter is pressed.

        Search is now performed through a keyup event.

        Added a WebInspector.lastQuery to remember the last query and not
        perform it again (eg. the user hits a modifier key).

        https://bugs.webkit.org/show_bug.cgi?id=18548

        Reviewed by Timothy Hatcher.

        * page/inspector/inspector.html: Removed the onsearch and incremental
        attributes
        * page/inspector/inspector.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35632 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Compile derived sources through an all-in-one cpp file.
        Fixes Windows release build on non-x64 systems by reducing the size of WebCore.lib from 1.5GB to 1GB.

        Reviewed by Mark Rowe.

        * DerivedSources.cpp: Added.
        * WebCore.vcproj/WebCore.vcproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35629 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-07  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Adam.

        <rdar://problem/5712917> Confusing console error message clear
        behavior (16791)
        https://bugs.webkit.org/show_bug.cgi?id=16791
        <rdar://problem/5831899> The Web Inspector does not correctly clear the
        console messages

        * page/InspectorController.cpp: Clear the message so they don't show up
        in the inspector after it has cleared them.
        (WebCore::clearMessages):
        (WebCore::InspectorController::clearConsoleMessages):
        (WebCore::InspectorController::windowScriptObjectAvailable):
        * page/InspectorController.h:
        * page/inspector/Console.js: Tell the InspectorController to clear the
        messages it's holding onto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35628 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-07  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Sam Weinig

        Rename Marquee to RenderMarquee and move to its own file.
        https://bugs.webkit.org/show_bug.cgi?id=20319

        * GNUmakefile.am:
        * WebCore.pro:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:
        * html/HTMLMarqueeElement.cpp:
        * rendering/RenderBlock.cpp:
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::styleChanged):
        * rendering/RenderLayer.h:
        * rendering/RenderMarquee.cpp: Added.
        (WebCore::RenderMarquee::RenderMarquee):
        (WebCore::RenderMarquee::marqueeSpeed):
        (WebCore::RenderMarquee::direction):
        (WebCore::RenderMarquee::isHorizontal):
        (WebCore::RenderMarquee::computePosition):
        (WebCore::RenderMarquee::start):
        (WebCore::RenderMarquee::suspend):
        (WebCore::RenderMarquee::stop):
        (WebCore::RenderMarquee::updateMarqueePosition):
        (WebCore::RenderMarquee::updateMarqueeStyle):
        (WebCore::RenderMarquee::timerFired):
        * rendering/RenderMarquee.h: Added.
        (WebCore::RenderMarquee::speed):
        (WebCore::RenderMarquee::reverseDirection):
        (WebCore::RenderMarquee::setEnd):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35627 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by Dave Hyatt.

        - fix <rdar://problem/6118816> Initial empty page or bookmarks view window has white strip where vertical scrollbar would be (which vanishes on resize)

        * page/Frame.cpp:
        (WebCore::Frame::setZoomFactor): If the view has never had a layout with
        the old zoom factor, then there is no need to explicitly call layout()
        here.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35626 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-07  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Geoff.

        <rdar://problem/6115981> Start the profiler in Heavy view

        * page/inspector/ProfileView.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35625 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Delete backing stores before 5 second WM_TIMER fires if more than 2 are already queued.

        WM_TIMER-based timers can be starved if enough other events are keeping
        our message loop busy. This change prevents many backing store deletes from stacking
        up in this case.

        Reviewed by Geoff Garen.

        * WebView.cpp:
        (WebView::deleteBackingStore):
        (WebView::deleteBackingStoreSoon):
        (WebView::cancelDeleteBackingStoreSoon):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35624 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-07  Alp Toker  <alp@nuanti.com>

        Reviewed by Eric Seidel.

        https://bugs.webkit.org/show_bug.cgi?id=20313
        Add null check in String::fromUTF8()

        Make String::fromUTF8() consistent with other constructors by
        returning a null String when the input is null instead of crashing.

        * platform/text/String.cpp:
        (WebCore::String::fromUTF8):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35623 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-07  Kalle Vahlman  <zuh@iki.fi>

        Reviewed by David Kilzer.

        Check for correct flex version to avoid faulty builds

        A correct Webkit build requires flex 2.5.33 but autotools build doesn't check for it
        https://bugs.webkit.org/show_bug.cgi?id=20253

        acinclude.m4: Import AX_COMPARE_VERSION macro
        configure.ac: Check for flex version

        * acinclude.m4:
        * configure.ac:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35622 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-07  Simon Hausmann  <hausmann@webkit.org>

        Rubber-stamped by Lars.

        Added API tests for QWebPage/QWebFrame based on QTestLib.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35621 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Kevin Ollivier.

Create a wxWebFrame API to match other ports and to prepare for frames support.
Also fixes a frame leak in wx port on trunk.

https://bugs.webkit.org/show_bug.cgi?id=19041

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35620 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35619 268f45cc-cd09-0410-ab3c-d52691b4dbfc

wx build fix after Frame -> ScriptController moves. Also, remove a character which was in an unknown encoding and replace it with an ANSI representation (which the author used for their Bugzilla account too).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35618 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-06  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Tim.

        - Make XHR console message clearer.

        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::didFinishLoading):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35617 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-06  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Tim.

        - Created a new function to create DOM links from text without
        converting back and forth from HTML and messing with escaping special
        characters.

        * page/inspector/Console.js:
        * page/inspector/inspector.js:
        * page/inspector/utilities.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35616 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by hyatt.

        Fix a large animation leak found on the buildbot
        (m_animations and m_transitions were never deleted)
        Drag RenderStyle (kicking and screaming) into the 21st century
        of memory management with a little application of OwnPtr.

        * rendering/style/RenderStyle.cpp:
        (WebCore::FillLayer::cullEmptyLayers):
        (WebCore::RenderStyle::diff):
        (WebCore::RenderStyle::contentDataEquivalent):
        (WebCore::RenderStyle::setContent):
        (WebCore::BindingURI::BindingURI):
        (WebCore::RenderStyle::setBoxShadow):
        (WebCore::ShadowData::ShadowData):
        (WebCore::RenderStyle::counterDirectives):
        (WebCore::RenderStyle::accessCounterDirectives):
        (WebCore::RenderStyle::adjustAnimations):
        (WebCore::RenderStyle::adjustTransitions):
        (WebCore::RenderStyle::accessAnimations):
        (WebCore::RenderStyle::accessTransitions):
        * rendering/style/RenderStyle.h:
        (WebCore::RenderStyle::clearAnimations):
        (WebCore::RenderStyle::clearTransitions):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35615 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-06  Brady Eidson  <beidson@apple.com>

        Reviewed by MitzPettel

        <rdar://problem/6078543> - Many LocalStorageThreads may be started

        The original decision to key LocalStorage sets to PageGroups was a poor one.

        It seems a much better solution is to key them to the actual path of on-disk storage,
        which also solves all of the avenues that could lead to a boundless number of threads.

        In this patch, I am not attempting to completely sever the attachment of LocalStorage
        to PageGroups as that would require a more substantial re-working of the lifetime
        between pages, groups, and local storage threads.

        * page/PageGroup.cpp:
        (WebCore::PageGroup::addPage):

        * storage/LocalStorage.cpp:
        (WebCore::localStorageMap):  Map paths to LocalStorage objects here.
        (WebCore::LocalStorage::localStorage): Return the shared LocalStorage object for this path,
          or create a new one if it doesn't yet exist.
        (WebCore::LocalStorage::LocalStorage):
        (WebCore::LocalStorage::~LocalStorage):  Add to remove the path->LocalStorage mapping
        * storage/LocalStorage.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35614 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-06  Beth Dakin  <bdakin@apple.com>

        Reviewed by Adele.

        Fix for <rdar://problem/6130724>

        Disconnected frames should not be allowed to resize the window.

        * page/DOMWindow.cpp:
        (WebCore::DOMWindow::resizeBy):
        (WebCore::DOMWindow::resizeTo):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35613 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Added bug URLs

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35612 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Bug 20038: REGRESSION (r35151): Can't post comments on flickr.com

        <https://bugs.webkit.org/show_bug.cgi?id=20038>
        <rdar://problem/6092270>

        Reviewed by Eric Seidel.

        Test: fast/forms/submit-to-url-fragment.html

        The problem was that isFormSubmission was not being set to true in
        FrameLoader::loadWithDocumentLoader() when we were actually
        submitting a form, causing the page to scroll instead of the form to
        be submitted.

        The isFormSubmission variable wasn't set to true because a FormState
        object was not being created in
        FrameLoader::loadFrameRequestWithFormAndValues().

        The FormState object was not being created because
        HTMLFormElement::submit(Event*, bool activateSubmitButton) would
        only set FrameLoader::m_formAboutToBeSubmitted to the current form
        if there was a "text field" element in the form (type = text,
        password, search or isindex).

        Thus when FrameLoader::submitForm(const FrameLoadRequest&, Event*)
        called FrameLoader::loadFrameRequestWithFormAndValues(), a null
        HTMLFormElement pointer would be sent and cause the above failures.

        * html/HTMLFormElement.cpp:
        (WebCore::HTMLFormElement::submit): Call new
        FrameLoader::setFormAboutToBeSubmitted() method outside the for
        loop so we always set FrameLoader::m_formAboutToBeSubmitted exactly
        once for any form submission.  The FrameLoader::recordFormValue()
        method is only called with the name/value pair of each text field.

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::setFormAboutToBeSubmitted): Added method that
        only sets m_formAboutToBeSubmitted.
        (WebCore::FrameLoader::recordFormValue): Removed
        PassRefPtr<HTMLFormElement> argument since this method only sets
        name/value pairs on m_formValuesAboutToBeSubmitted now.
        (WebCore::FrameLoader::loadFrameRequestWithFormAndValues): Create a
        FormState object as long as submitForm is not null so that other
        FrameLoader methods know when a form is being submitted.
        * loader/FrameLoader.h:

LayoutTests:

        Bug 20038: REGRESSION (r35151): Can't post comments on flickr.com

        <https://bugs.webkit.org/show_bug.cgi?id=20038>
        <rdar://problem/6092270>

        Reviewed by Eric Seidel.

        * fast/forms/submit-to-url-fragment-expected.txt: Added.
        * fast/forms/submit-to-url-fragment.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35611 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    BUILD FIX: Add HTMLPlugInImageElement.cpp to all other build systems

        * GNUmakefile.am:
        * WebCore.pro:
        * WebCoreSources.bkl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35610 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    - Windows build fix

        * WebCore.vcproj/WebCore.vcproj: Added PlugInImageElement.*

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35609 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Reviewed by darin.

        Fix leaks seen on build-bot by fixing memory management of AccessibilityUIElement.

        I've made AccessibilityUIElement a stack object for now.  It could be
        made RefCounted, but I figured that might be overkill for the moment.
        Essentially it's just a RefPtr itself. :)

        I also fixed a few typos, such as AccessibilityUIElement() instead of ~AccessibilityUIElement()
        which was causing additional leaks. :)

        I added toAXElement to remove a bunch of copy/paste code.

        * DumpRenderTree/AccessibilityController.h:
        * DumpRenderTree/AccessibilityUIElement.cpp:
        (toAXElement):
        (allAttributesCallback):
        (attributesOfLinkedUIElementsCallback):
        (attributesOfChildrenCallback):
        (lineForIndexCallback):
        (boundsForRangeCallback):
        (childAtIndexCallback):
        (getRoleCallback):
        (getTitleCallback):
        (getDescriptionCallback):
        (getWidthCallback):
        (getHeightCallback):
        (getIntValueCallback):
        (getMinValueCallback):
        (getMaxValueCallback):
        (getInsertionPointLineNumberCallback):
        (finalize):
        * DumpRenderTree/AccessibilityUIElement.h:
        (AccessibilityUIElement::platformUIElement):
        * DumpRenderTree/mac/AccessibilityControllerMac.mm:
        (AccessibilityController::focusedElement):
        * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
        (AccessibilityUIElement::AccessibilityUIElement):
        (AccessibilityUIElement::getLinkedUIElements):
        (AccessibilityUIElement::getChildren):
        (AccessibilityUIElement::getChildAtIndex):
        (AccessibilityUIElement::attributesOfLinkedUIElements):
        (AccessibilityUIElement::attributesOfChildren):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35608 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-06  Kevin McCullough  <kmccullough@apple.com>

        Reviewed by Tim and Darin.

        <rdar://problem/6126160> URLs in the console should be clickable.
        - Implemented and now use linkifyString() to parse console messages for
        URLs.

        * page/inspector/Console.js:
        * page/inspector/inspector.js:
        * page/inspector/utilities.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35607 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebCore:

        Images using QT plugin do not display correctly

        <https://bugs.webkit.org/show_bug.cgi?id=9214>
        <rdar://problem/5683413>

        Reviewed by Darin.

        Tests: fast/images/embed-image.html
               fast/images/object-image.html

        Use RenderImage instead of RenderPartObject for a renderer when
        the <embed> tag references an image.  Extract common code from
        HTMLEmbedElement and HTMLObjectElement into HTMLPlugInImageElement.

        * WebCore.xcodeproj/project.pbxproj: Added new source files.

        * html/HTMLEmbedElement.cpp: Cleaned up #include statements.
        (WebCore::HTMLEmbedElement::HTMLEmbedElement): Extend
        HTMLPlugInImageElement which extends HTMLPlugInElement.
        (WebCore::HTMLEmbedElement::parseMappedAttribute): Split codeAttr
        and srcAttr cases.  Create or clear HTMLImageLoader as needed for
        srcAttr.
        (WebCore::HTMLEmbedElement::rendererIsNeeded): Use
        HTMLPlugInElement::rendererIsNeeded() for images.
        (WebCore::HTMLEmbedElement::createRenderer): Create a
        RenderImage instead of a RenderPartObject for images.
        (WebCore::HTMLEmbedElement::attach): Create HTMLImageLoader,
        load image and populate RenderImage for images.
        (WebCore::HTMLEmbedElement::updateWidget): Don't call
        RenderPartObject::updateWidget() for images.
        (WebCore::HTMLEmbedElement::imageSourceAttributeName): Added.
        * html/HTMLEmbedElement.h: Removed unneeded namespace and class
        declarations.  Extend HTMLPlugInImageElement which extends
        HTMLPlugInElement.
        (WebCore::HTMLEmbedElement::imageSourceAttributeName): Added.
        (WebCore::HTMLEmbedElement::serviceType): Moved to
        HTMLPlugInImageElement.
        (WebCore::HTMLEmbedElement::url): Ditto.

        * html/HTMLObjectElement.cpp: Cleaned up #include statements.
        (WebCore::HTMLObjectElement::attach): Removed local variable.
        (WebCore::HTMLObjectElement::isImageType): Moved to
        HTMLPlugInImageElement.
        * html/HTMLObjectElement.h: Extend HTMLPlugInImageElement which
        extends HTMLPlugInElement.
        (WebCore::HTMLObjectElement::isImageType): Moved to
        HTMLPlugInImageElement.
        (WebCore::HTMLObjectElement::serviceType): Ditto.
        (WebCore::HTMLObjectElement::url): Ditto.

        * html/HTMLPlugInImageElement.cpp: Added.
        (WebCore::HTMLPlugInImageElement::HTMLPlugInImageElement): Added.
        (WebCore::HTMLPlugInImageElement::~HTMLPlugInImageElement): Added.
        (WebCore::HTMLPlugInImageElement::isImageType): Moved from
        HTMLObjectElement.
        * html/HTMLPlugInImageElement.h: Added.
        (WebCore::HTMLPlugInImageElement::isImageType): Added from
        HTMLObjectElement.
        (WebCore::HTMLPlugInImageElement::serviceType): Ditto.
        (WebCore::HTMLPlugInImageElement::url): Ditto.

        * rendering/HitTestResult.cpp:
        (WebCore::HitTestResult::absoluteImageURL): Return urlString
        for embed elements with images.

LayoutTests:

        Images using QT plugin do not display correctly

        <https://bugs.webkit.org/show_bug.cgi?id=9214>
        <rdar://problem/5683413>

        Reviewed by Darin.

        Test images rendered by both embed and object elements.

        * fast/images/embed-image.html: Added.
        * fast/images/object-image.html: Added.
        * platform/mac/fast/images/embed-image-expected.checksum: Added.
        * platform/mac/fast/images/embed-image-expected.png: Added.
        * platform/mac/fast/images/embed-image-expected.txt: Added.
        * platform/mac/fast/images/object-image-expected.checksum: Added.
        * platform/mac/fast/images/object-image-expected.png: Added.
        * platform/mac/fast/images/object-image-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35606 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix an accessibility layout test that was failing on different platforms

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35605 268f45cc-cd09-0410-ab3c-d52691b4dbfc

    Try to fix the Windows build bots

        * API/JSBase.cpp: Touch this to force JSC to rebuild and re-copy the
        WTF headers.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35604 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-06  Benjamin C Meyer  <ben@meyerhome.net>

        Reviewed by Simon.

        During the drag operation only accept the event if the action is not ignore action.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35603 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Revert change 35595.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35602 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-06  Marco Barisione  <marco.barisione@collabora.co.uk>

        Reviewed by Eric Seidel.

        http://bugs.webkit.org/show_bug.cgi?id=20295
        RegularExpression::match should not crash when a null string is passed

        RegularExpression::match used to pass null subject strings to
        jsRegExpExecute causing an ASSERT failure and then a crash.
        There is no need to fix also RegularExpression:search and
        RegularExpression::searchRev as they just call
        RegularExpression::match.

        * platform/text/RegularExpression.cpp:
        (WebCore::RegularExpression::match): Return -1 if the string is null.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35601 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2008-08-06  Marco Barisione  <marco.barisione@collabora.co.uk>

        Reviewed by Eric Seidel.

        http://bugs.webkit.org/show_bug.cgi?id=20276
        [GTK] Makefile rules for the enum GTypes files depend on Makefile
        instead of GNUmakefile

        The patch committed in revision 34649 added the code needed to
        generate GTypes for enums but the makefile rules for
        stamp-webkitenumtypes.h and webkitenumtypes.cpp were depending on
        Makefile (used by other ports) instead of GNUmakefile.

        * GNUmakefile.am: Depend on GNUmakefile instead of Makefile.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@35600 268f45cc-cd09-0410-ab3c-d52691b4dbfc