2009-03-17 Darin Adler <darin@apple.com>

[WebKit-https.git] / WebCore / ChangeLog

2009-03-17  Darin Adler  <darin@apple.com>

        Earlier version reviewed by Adele Peterson.

        Bug 24304: REGRESSION (r39864): Hitting the space bar to select an <input type=radio>
        or push an <input type=button> or <button> causes the page to scroll down.

        Would be best to add a regression test for Windows eventually; tested that this has
        no effect on the Mac OS X platform.

        * html/HTMLInputElement.cpp:
        (WebCore::HTMLInputElement::defaultEventHandler): Added FIXMEs and tweaked formatting.
        Use the code that calls the base class's defaultEventHandler early only in the cases
        where it's needed: keydown and keypress events in text fields. In other cases, do the
        more typical thing and call the default handler only at the end of the function.
        This function already had code to make sure the keypress event for space never gets
        through, but it was running too late since the scrolling code was moved into the
        base class default event handler.

2009-03-17  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Darin Adler

        https://bugs.webkit.org/show_bug.cgi?id=24396

        * config.h:
        Add WTF_USE_ACCELERATED_COMPOSITING, defined to 0 for now, and add some
        comments to make the #ifdefs more readable.

        * css/CSSComputedStyleDeclaration.cpp:
        (WebCore::computedTransform):
        Add a comment to mention that we don't flatten the matrix.
        
        * css/MediaQueryEvaluator.cpp:
        (WebCore::transform_3dMediaFeatureEval):
        Have the 'transform-3d' media query evaluate to 'true' if 3d-rendering
        is supported.
        
        * platform/graphics/mac/GraphicsLayerCA.mm:
        (WebCore::GraphicsLayerCA::animateTransform):
        No need for the #ifdef here. If we don't support 3d, we will have already flattened
        the matrix.
        
        * platform/graphics/transforms/TransformationMatrix.cpp:
        (WebCore::TransformationMatrix::makeAffine):
        * platform/graphics/transforms/TransformationMatrix.h:
        New method to convert the matrix to an affine matrix by throwing a way the non-affine
        parts.
        
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::updateTransform):
        (WebCore::RenderLayer::currentTransform):
        * rendering/RenderLayerBacking.cpp:
        (WebCore::RenderLayerBacking::updateLayerTransform):
        If 3d rendering is not supported, convert the matrix to an affine matrix
        which can be rendered, and used for hit testing.
        
        * rendering/RenderLayerCompositor.cpp:
        Change the name of the exported symbol that webkitdirs.pm uses to know if
        3d rendering is supported. There is no other 3d-rendering-specific symbol we can sniff.
        
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::transformFromContainer):
        Only take perspective into account if 3d rendering is supported.

        * rendering/RenderObject.h:
        (WebCore::makeMatrixRenderable):
        Utility method that flattens a matrix if 3d rendering is not supported.

2009-03-17  Kevin Ollivier  <kevino@theolliviers.com>

        wx build fix. Fix typo after mouse wheel changes.

        * platform/wx/MouseWheelEventWx.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):

2009-03-17  Darin Adler  <darin@apple.com>

        * inspector/ConsoleMessage.cpp:
        (WebCore::ConsoleMessage::isEqual): Fix build, remove stray parenthesis.

2009-03-17  Darin Adler  <darin@apple.com>

        Fix crash seen right away when running run-webkit-tests.

        * inspector/ConsoleMessage.cpp:
        (WebCore::ConsoleMessage::isEqual): Restore assertion to its behavior pre-refactoring.
        Also tweaked formatting a bit.

2009-03-17  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        HTMLSelectElement::add() doesn't look at exception code returned from insertBefore(), so 
        it doesn't need to zero it out before calling.

        * html/HTMLSelectElement.cpp: (WebCore::HTMLSelectElement::add): Removed "ec = 0" line.

2009-03-17  Dan Bernstein  <mitz@apple.com>

        Reviewed by Adam Roben.

        - WebCore part of adding a mechanism for controlling the caching of
          responses through WebFrameLoaderClient

        Mac already has such a mechanism, and this adds one for CFNetwork ports.

        * WebCore.vcproj/WebCore.vcproj: Added EmptyClients.h
        and ResourceLoaderCFNet.cpp.

        * loader/EmptyClients.h:
        (WebCore::EmptyFrameLoaderClient::shouldCacheResponse): Added an
        implementation that always returns true.

        * loader/FrameLoaderClient.h: Declared shouldCacheResponse().

        * loader/ResourceLoader.h: Ditto.

        * loader/cf/ResourceLoaderCFNet.cpp: Added.
        (WebCore::ResourceLoader::shouldCacheResponse): Added. Calls through to
        FrameLoaderClient::shouldCacheResponse().

        * platform/network/ResourceHandleClient.h:
        (WebCore::ResourceHandleClient::shouldCacheResponse): Added an
        implementation that always returns true.

        * platform/network/cf/ResourceHandleCFNet.cpp:
        (WebCore::willCacheResponse): Added a call to
        ResourceHandleClient::shouldCacheResponse(). If the client returns
        false, return 0, which will prevent CFNetwork from caching the response.

2009-03-17  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=13287
        Cannot change SELECT to a dynamically created option

        Tests: fast/forms/add-and-remove-option.html
               fast/forms/add-remove-option-modification-event.html
               fast/forms/add-selected-option.html
               fast/forms/select-cache-desynchronization.html

        * dom/ContainerNode.cpp:
        (WebCore::dispatchChildInsertionEvents): Increment DOM tree version. This will happen when
        dispatching DOMSubtreeModified again, but the version should be incremented for event
        listeners to have an up to date view of the DOM.
        (WebCore::dispatchChildRemovalEvents): Ditto.

        * html/HTMLOptionElement.cpp: (WebCore::HTMLOptionElement::insertedIntoTree):
        Make sure that the select element knows about its new selected option.

        * html/HTMLOptionElement.h: Use insertedIntoTree() instead of insertedIntoDocument(),
        because DOM also needs to be updated for forms that are not in document yet. Similar
        problems exist for node removing, but removedFromTree() is called at a wrong time, so
        those problems cannot be fixed without deeper refactoring.

        * html/HTMLSelectElement.cpp:
        (WebCore::HTMLSelectElement::setRecalcListItems): Reset m_activeSelectionAnchorIndex - it
        doesn't make sense to keep the anchor after programmatically changing the selection, and
        keeping it was causing a failure in fast/forms/listbox-selection.html.

        * html/HTMLSelectElement.h: Removed overrides for ContainerNode methods that only called
        base class versions.

2009-03-17  Steve Falkenburg  <sfalken@apple.com>

        <rdar://problem/6690324> Accessing FTP sites reads unallocated memory, can result in garbled entries or crashes
        
        Reviewed by Darin Adler.

        * loader/FTPDirectoryDocument.cpp:
        (WebCore::FTPDirectoryTokenizer::parseAndAppendOneLine): Assign CString to a local while we hold pointers into it.

2009-03-16  David Hyatt  <hyatt@apple.com>

        <rdar://problem/6648411> REGRESSION: Layout of page is wrong at http://www.popcap.com/

        Make sure that the initial shouldPaint check that looks at enclosingLayers properly skips over
        layers that don't paint themselves.  This is done by adding a new enclosingSelfPaintingLayer method
        so that RenderObjects can walk up the enclosing layer chain and skip any layers that don't paint
        themselves.

        Reviewed by Darin Adler.

        Added fast/block/float/overlapping-floats-with-overflow-hidden.html

        * WebCore.base.exp:
        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::addOverhangingFloats):
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::enclosingSelfPaintingLayer):
        * rendering/RenderObject.h:

2009-03-17  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24592
        [GTK] Crash in FcPatternHash

        Style fixes.

        * platform/graphics/gtk/FontPlatformDataGtk.cpp:
        (WebCore::FontPlatformData::operator=):
        * platform/graphics/gtk/FontPlatformDataPango.cpp:

2009-03-17  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24592
        [GTK] Crash in FcPatternHash

        Sanitize memory management in pango fonts.

        Release memory allocated by FontPlatformDataPango in its own
        destructor instead of doing it from other classes, and add copy
        constructor and '=' operator to be able to track referenced
        objects properly.

        * platform/graphics/gtk/FontPlatformDataPango.cpp:
        (WebCore::FontPlatformData::~FontPlatformData):
        (WebCore::FontPlatformData::operator=):
        (WebCore::FontPlatformData::FontPlatformData):
        * platform/graphics/gtk/SimpleFontDataPango.cpp:

2009-03-17  Darin Adler  <darin@apple.com>

        Reviewed by Alexey Proskuryakov.

        Bug 24624: Crash in imageLoadEventTimerFired after adoptNode used on <img>,
        seen with inspector, which uses adoptNode
        https://bugs.webkit.org/show_bug.cgi?id=24624
        rdar://problem/6422850

        Test: fast/dom/HTMLImageElement/image-load-cross-document.html

        * dom/Document.cpp:
        (WebCore::Document::Document): Removed m_imageLoadEventTimer.
        (WebCore::Document::detach): Removed m_imageLoadEventDispatchSoonList and
        m_imageLoadEventDispatchingList.
        (WebCore::Document::implicitClose): Called ImageLoader::dispatchPendingLoadEvents
        instead of dispatchImageLoadEventsNow.

        * dom/Document.h: Removed ImageLoader, dispatchImageLoadEventSoon,
        dispatchImageLoadEventsNow, removeImage, m_imageLoadEventDispatchSoonList,
        m_imageLoadEventDispatchingList, m_imageLoadEventTimer, and imageLoadEventTimerFired.

        * loader/ImageLoader.cpp:
        (WebCore::loadEventSender): Added. Returns the single global ImageLoadEventSender
        object used privately as the target of the load event timer.
        (WebCore::ImageLoader::~ImageLoader): Call ImageLoadEventSender::cancelLoadEvent
        rather than Document::removeImage.
        (WebCore::ImageLoader::setImage): Use m_element directly, not element().
        (WebCore::ImageLoader::updateFromElement): Ditto. Also name the local variable
        document instead of doc.
        (WebCore::ImageLoader::notifyFinished): Call ImageLoadEventSender::dispatchLoadEventSoon
        rather than Document::dispatchImageLoadEventSoon.
        (WebCore::ImageLoader::dispatchPendingLoadEvent): Added. Handles the common logic
        about when load events can be dispatched so that dispatchLoadEvent only has to
        have the specific part for each derived class. This includes a check that the
        document is attached, which used to be handled by having documents empty out the
        image load event vectors in the detach function.
        (WebCore::ImageLoader::dispatchPendingLoadEvents): Added. Calls the appropriate
        function on the ImageLoadEventSender, which avoids the need to have that class be
        public in the ImageLoader header.
        (WebCore::ImageLoadEventSender::ImageLoadEventSender): Added. Has the code that
        was previously in the Document constructor.
        (WebCore::ImageLoadEventSender::dispatchLoadEventSoon): Added. Has the code that
        was previously in Document::dispatchImageLoadEventSoon.
        (WebCore::ImageLoadEventSender::cancelLoadEvent): Added. Has the code that was
        previously in Document::removeImage.
        (WebCore::ImageLoadEventSender::dispatchPendingLoadEvents): Added. Has the code
        that was previously in Document::dispatchImageLoadEventsNow.
        (WebCore::ImageLoadEventSender::timerFired): Added. Calls dispatchPendingLoadEvents.

        * loader/ImageLoader.h: Improved comments. Made the virtual functions private
        or protected rather than public. Added static dispatchPendingLoadEvents function
        for use by Document and private dispatchPendingLoadEvent function for use by
        ImageLoadEventSender. Made setLoadingImage private and eliminated
        setHaveFiredLoadEvent since that can all be done inside the class without any
        member functions.

        * html/HTMLImageLoader.cpp:
        (WebCore::HTMLImageLoader::dispatchLoadEvent): Removed logic to check whether a
        load event already fired and whether image() is 0. These are now both base class
        responsibilities.
        * svg/SVGImageLoader.cpp:
        (WebCore::SVGImageLoader::dispatchLoadEvent): Ditto.
        * wml/WMLImageLoader.cpp:
        (WebCore::WMLImageLoader::dispatchLoadEvent): Ditto.

2009-03-17  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Timothy Hatcher.

        https://bugs.webkit.org/show_bug.cgi?id=24623
        Refactor ConsoleMessage to use ScriptFuncitonCall and eliminate JSC
        dependencies.

        * bindings/js/ScriptFunctionCall.cpp:
        (WebCore::ScriptFunctionCall::appendArgument): Added uint and ScriptString-taking methods.
        * bindings/js/ScriptFunctionCall.h:
        * bindings/js/ScriptObjectQuarantine.cpp:
        (WebCore::quarantineValue): Added generic ScriptValue quarantine helper.
        * bindings/js/ScriptObjectQuarantine.h:
        * bindings/js/ScriptValue.cpp:
        (WebCore::ScriptValue::isEqual): Added.
        * bindings/js/ScriptValue.h:
        * inspector/ConsoleMessage.cpp:
        (WebCore::ConsoleMessage::ConsoleMessage):
        (WebCore::ConsoleMessage::addToConsole): Added.
        (WebCore::ConsoleMessage::isEqual): Changed to use ScriptValue::isEqual.
        * inspector/ConsoleMessage.h:
        (WebCore::ConsoleMessage::incrementCount): Added.
        * inspector/InspectorController.cpp:
        (WebCore::InspectorController::addConsoleMessage): Changed to use ConsoleMessage::addToConsole.
        (WebCore::InspectorController::populateScriptObjects): Ditto.
        * inspector/InspectorController.h:

2009-03-17  Kevin Ollivier  <kevino@theolliviers.com>

        Reviewed by Mark Rowe.

        Get BUILDING_ON_* defines from Platform.h.
        
        https://bugs.webkit.org/show_bug.cgi?id=24630

        * WebCorePrefix.h:

2009-03-16  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24592
        [GTK] Crash in FcPatternHash

        Sanitize memory management in gtk fonts.

        Release memory allocated by FontPlatformDataGtk in its own
        destructor instead of doing it from other classes, and add copy
        constructor and '=' operator to be able to track referenced
        objects properly.

        * platform/graphics/gtk/FontPlatformData.h:
        * platform/graphics/gtk/FontPlatformDataGtk.cpp:
        (WebCore::FontPlatformData::operator=):
        (WebCore::FontPlatformData::FontPlatformData):
        (WebCore::FontPlatformData::~FontPlatformData):
        * platform/graphics/gtk/SimpleFontDataGtk.cpp:
        (WebCore::SimpleFontData::platformDestroy):

2009-03-17  Ariya Hidayat  <ariya.hidayat@nokia.com>

        Build fix for Qt < 4.5.

        As reported by Yael Aharon  <yael.aharon@nokia.com>

        * platform/graphics/qt/GraphicsContextQt.cpp:
        (WebCore::GraphicsContext::drawLine):

2009-03-17  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Sam Weinig.

        https://bugs.webkit.org/show_bug.cgi?id=24614
        Access control checks are different in cached and uncached cases

        Test: http/tests/xmlhttprequest/access-control-basic-non-simple-deny-cached.html

        * loader/CrossOriginAccessControl.cpp:
        (WebCore::isOnAccessControlSimpleRequestMethodWhitelist): Factored out simple method
        check for use in both cached and uncached cases. In cached case, an old definition that
        omitted HEAD was still used.
        (WebCore::isOnAccessControlSimpleRequestHeaderWhitelist): Check that content type has an
        allowed value. This is needed in all call sites. Also changed to compare MIME type, not
        content type.
        (WebCore::isSimpleCrossOriginAccessRequest): Use the above methods.

        * loader/CrossOriginAccessControl.h: Expose isOnAccessControlSimpleRequestMethodWhitelist.

        * loader/CrossOriginPreflightResultCache.cpp:
        (WebCore::CrossOriginPreflightResultCacheItem::allowsCrossOriginMethod):
        (WebCore::CrossOriginPreflightResultCacheItem::allowsCrossOriginHeaders):
        Use the new checks for simple method and header.

2009-03-16  Gustavo Noronha Silva  <gns@gnome.org> and Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com>

        Reviewed by Mark Rowe.

        https://bugs.webkit.org/show_bug.cgi?id=24638
        [GTK] HTML5 media tags do not work

        Add a repaint-requested signal to the video sink, and use it to
        call MediaPlayerPrivate::repaint, so that the video actually
        plays.

        * platform/graphics/gtk/MediaPlayerPrivateGStreamer.cpp:
        (WebCore::mediaPlayerPrivateRepaintCallback):
        (WebCore::MediaPlayerPrivate::createGSTPlayBin):
        * platform/graphics/gtk/VideoSinkGStreamer.cpp:
        (webkit_video_sink_idle_func):
        (webkit_video_sink_render):
        (webkit_video_sink_class_init):

2009-03-16  Gustavo Noronha Silva  <gns@gnome.org> and Thadeu Lima de Souza Cascardo <cascardo@holoscopio.com>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24638
        [GTK] HTML5 media tags do not work

        Work-around the fact that gst_element_query_duration returns true even
        though it is unable to figure out the duration when in stream (push)
        mode.

        * platform/graphics/gtk/MediaPlayerPrivateGStreamer.cpp:
        (WebCore::MediaPlayerPrivate::duration):

2009-03-16  Darin Adler  <darin@apple.com>

        Reviewed by Kevin Decker.

        <rdar://problem/6642742> Top Sites malfunction when switching text zoom mode

        * page/Frame.cpp:
        (WebCore::Frame::setNeedsReapplyStyles): Don’t do anything if the frame is
        currently showing a non-HTML view.

2009-03-16  Darin Adler  <darin@apple.com>

        Reviewed by Adele Peterson.

        Bug 24629: moving forward or backward a paragraph fails at edge of document
        https://bugs.webkit.org/show_bug.cgi?id=24629
        rdar://problem/6544413

        Test: editing/selection/move-paragraph-document-edges.html

        * editing/visible_units.cpp:
        (WebCore::previousParagraphPosition): Use the last result from
        previousLinePosition rather than going all the way back to what was originally
        passed in when we hit exception cases like null or not moving. This correctly
        inherits the behavior of previousLinePosition when we are in a paragraph at the
        edge of a document.
        (WebCore::nextParagraphPosition): Ditto.

2009-03-16  Darin Adler  <darin@apple.com>

        Reviewed by Adele Peterson.

        Bug 24619: RenderObject::selectionStartEnd does not need to be a virtual function
        https://bugs.webkit.org/show_bug.cgi?id=24619

        * rendering/RenderObject.h: Remove virtual keyword from selectionStartEnd declaration.
        * rendering/RenderView.h: Ditto.

2009-03-16  Peter Kasting  <pkasting@google.com>

        Reviewed by David Hyatt.

        https://bugs.webkit.org/show_bug.cgi?id=24368
        DOM scroll events should be based off the actual number of wheel
        ticks, not off the number of lines scrolled.  This matches IE.

        * dom/Node.cpp:
        (WebCore::Node::dispatchWheelEvent):
        * dom/WheelEvent.cpp:
        (WebCore::WheelEvent::WheelEvent):
        * dom/WheelEvent.h:
        (WebCore::WheelEvent::create):
        * platform/PlatformWheelEvent.h:
        (WebCore::PlatformWheelEvent::wheelTicksX):
        (WebCore::PlatformWheelEvent::wheelTicksY):
        * platform/gtk/WheelEventGtk.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/mac/WheelEventMac.mm:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/qt/WheelEventQt.cpp:
        * platform/win/WheelEventWin.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/wx/MouseWheelEventWx.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):

2009-03-16  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Eric Seidel
        
        Clean up a few issues in the Animation code:

        * page/animation/AnimationBase.cpp:
        (WebCore::AnimationBase::updateStateMachine):
        Whitespace
        
        (WebCore::AnimationBase::willNeedService):
        Don't round to float, use std::max

        * page/animation/AnimationController.cpp:
        (WebCore::AnimationControllerPrivate::startTimeResponse):
        Fix erroneously copied line to null out m_lastResponseWaiter.

2009-03-12  David Hyatt  <hyatt@apple.com>

        Reviewed by Eric Seidel

        https://bugs.webkit.org/show_bug.cgi?id=13632

        Overflow scrolling needs to account for the bottom/right padding on the object itself as well
        as for bottom/right margins on children.

        Existing tests cover this.

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::lowestPosition):
        (WebCore::RenderBlock::rightmostPosition):

2009-03-16  Sam Weinig  <sam@webkit.org>

        Reviewed by Anders Carlsson.

        Fix for <rdar://problem/6320555>
        Add an upper limit for setting HTMLSelectElement.length.

        Test: fast/forms/select-max-length.html

        * html/HTMLSelectElement.cpp:
        (WebCore::HTMLSelectElement::setOption):
        (WebCore::HTMLSelectElement::setLength):

2009-03-16  Eric Carlson  <eric.carlson@apple.com>

        Reviewed by Simon Fraser.

        <rdar://problem/6686721> Media document crash in 64-bit WebKit

        * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
        (WebCore::MediaPlayerPrivate::createQTMovieView): QTMovieContentViewClass is only used when
        rendering inline with old versions of QuickTime, so don't look for it when we are in a 
        media document.

2009-03-16  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Timothy Hatcher.

        https://bugs.webkit.org/show_bug.cgi?id=24590
        Refactor InspectorDOMStorageResource to use ScriptFunctionCall.

        * bindings/js/ScriptFunctionCall.cpp:
        (WebCore::ScriptFunctionCall::appendArgument): Added method for bool argument.
        * bindings/js/ScriptFunctionCall.h: Ditto, also cleaned up.
        * bindings/js/ScriptObjectQuarantine.cpp:
        (WebCore::getQuarantinedScriptObject): Added Storage helper.
        * bindings/js/ScriptObjectQuarantine.h: Ditto.
        * inspector/InspectorController.cpp:
        (WebCore::InspectorController::populateScriptObjects): Changed to use bind method.
        (WebCore::InspectorController::resetScriptObjects): Changed to use unbind method.
        (WebCore::InspectorController::didUseDOMStorage): Changed to use isSameHostAndType and bind methods.
        * inspector/InspectorController.h: Removed add/remove methods for DOM storage.
        * inspector/InspectorDOMStorageResource.cpp:
        (WebCore::InspectorDOMStorageResource::InspectorDOMStorageResource):
        (WebCore::InspectorDOMStorageResource::isSameHostAndType): Added.
        (WebCore::InspectorDOMStorageResource::bind): Added.
        (WebCore::InspectorDOMStorageResource::unbind): Added.
        * inspector/InspectorDOMStorageResource.h:

2009-03-16   Mike Belshe  <mike@belse.com>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24580
        Fix query() to match KURL behavior, this time with the code that
        compiles.

        * platform/KURLGoogle.cpp:
        (WebCore::KURL::query): Fix copy/paste mistake.

2009-03-16  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=21752
        REGRESSION: referencing XHR constructor for a not yet loaded frame permanently breaks it

        Test: fast/dom/Window/window-early-properties-xhr.html

        For some transitions, the Window object is not replaced, but Document is. When this happened,
        window.document property was updated, but references to Document kept in cached constructors
        were not.

        * bindings/js/JSAudioConstructor.cpp:
        (WebCore::JSAudioConstructor::JSAudioConstructor):
        (WebCore::JSAudioConstructor::document):
        (WebCore::JSAudioConstructor::mark):
        * bindings/js/JSAudioConstructor.h:
        * bindings/js/JSImageConstructor.cpp:
        (WebCore::JSImageConstructor::JSImageConstructor):
        (WebCore::JSImageConstructor::document):
        (WebCore::JSImageConstructor::mark):
        * bindings/js/JSImageConstructor.h:
        * bindings/js/JSMessageChannelConstructor.cpp:
        (WebCore::JSMessageChannelConstructor::JSMessageChannelConstructor):
        (WebCore::JSMessageChannelConstructor::scriptExecutionContext):
        (WebCore::JSMessageChannelConstructor::mark):
        * bindings/js/JSMessageChannelConstructor.h:
        * bindings/js/JSOptionConstructor.cpp:
        (WebCore::JSOptionConstructor::JSOptionConstructor):
        (WebCore::JSOptionConstructor::document):
        (WebCore::JSOptionConstructor::mark):
        * bindings/js/JSOptionConstructor.h:
        * bindings/js/JSXMLHttpRequestConstructor.cpp:
        (WebCore::JSXMLHttpRequestConstructor::JSXMLHttpRequestConstructor):
        (WebCore::JSXMLHttpRequestConstructor::scriptExecutionContext):
        (WebCore::JSXMLHttpRequestConstructor::mark):
        * bindings/js/JSXMLHttpRequestConstructor.h:
        Changed cached constructors to keep a reference to Window, not Document.

2009-03-15  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=24549
        Impose a limit on Access-Control-Max-Age value

        * loader/CrossOriginPreflightResultCache.cpp:
        (WebCore::CrossOriginPreflightResultCacheItem::parse):

2009-03-15  Greg Bolsinga  <bolsinga@apple.com>

        Reviewed by David Kilzer.

        <rdar://problem/6668875> Normalize Geolocation results

        * platform/mac/GeolocationServiceMac.mm:
        (-[WebCoreCoreLocationObserver locationManager:didUpdateToLocation:fromLocation:]):

2009-03-15  Greg Bolsinga  <bolsinga@apple.com>

        Reviewed by David Kilzer.

        Update ::toString format as suggested by Darin Adler.

        * page/Geoposition.cpp:
        (WebCore::Geoposition::toString):

2009-03-15  David Kilzer  <ddkilzer@apple.com>

        <rdar://problem/6668238> WebCore is registering text encodings needlessly from KURL constructor.

        Reviewed by Darin Adler.

        Yet another case where we would trigger extended encoding loading needlessly.

        * platform/text/TextEncoding.cpp:
        (WebCore::TextEncoding::encodingForFormSubmission):

2009-03-15  Simon Fraser  <simon.fraser@apple.com>

        Build fix: no review.

        * rendering/style/ContentData.h:

2009-03-15  David Kilzer  <ddkilzer@apple.com>

        Bug 24542: Improve ContentData encapsulation

        <https://bugs.webkit.org/show_bug.cgi?id=24542>

        Reviewed by Simon Fraser.

        No tests since there is no change in behavior.

        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::createObject): Used getter methods
        instead of data members on ContentData class.  Used isImage()
        convenience method.
        * rendering/RenderObjectChildList.cpp:
        (WebCore::RenderObjectChildList::updateBeforeAfterContent): Ditto.

        * rendering/style/ContentData.cpp:
        (WebCore::ContentData::clear): Extracted code into
        deleteContent() method.
        (WebCore::ContentData::dataEquivalent): Added.  Extracted code
        from StyleRareNonInheritedData::contentDataEquivalent().
        (WebCore::ContentData::deleteContent): Added.  Used by setter
        methods.
        * rendering/style/ContentData.h: Made m_type, m_content and
        m_next private.
        (WebCore::ContentData::isCounter): Added.
        (WebCore::ContentData::isImage): Added.
        (WebCore::ContentData::isNone): Added.
        (WebCore::ContentData::isText): Added.
        (WebCore::ContentData::type): Added.
        (WebCore::ContentData::dataEquivalent): Added.
        (WebCore::ContentData::image): Added.
        (WebCore::ContentData::setImage): Added.
        (WebCore::ContentData::text): Added.
        (WebCore::ContentData::setText): Added.
        (WebCore::ContentData::counter): Added.
        (WebCore::ContentData::setCounter): Added.
        (WebCore::ContentData::next): Added.
        (WebCore::ContentData::setNext): Added.

        * rendering/style/CounterContent.h:
        (WebCore::operator!=): Removed.
        (WebCore::operator==): Renamed operator!=() and reversed its
        logic after extracting code from
        StyleRareNonInheritedData::contentDataEquivalent() to create
        ContentData::dataEquivalent().

        * rendering/style/RenderStyle.cpp:
        (WebCore::RenderStyle::setContent): Used new getters and setters
        on ContentData class.  Changed first argument from a StringImpl*
        to a PassRefPtr<StrimgImpl>.  Used isText() convenience method.
        * rendering/style/RenderStyle.h:
        (WebCore::RenderStyle::setContent): Updated declaration.
        * rendering/style/StyleRareNonInheritedData.cpp:
        (WebCore::StyleRareNonInheritedData::contentDataEquivalent):
        Extracted most logic in while() loop into
        ContentData::dataEquivalent().

2009-03-15  Gustavo Noronha Silva  <gns@gnome.org>

        Reviewed by Anders Carlsson.

        https://bugs.webkit.org/show_bug.cgi?id=24602
        [Gtk] Searching in thepiratebay.org doesn't work with more than 1 word

        Reintroduce the URI into the soup message after having set it in
        the KURL, on redirects, to make sure it is properly encoded. This
        fixes bad request problems when servers give bad URIs on their
        response's Location header.

        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::restartedCallback):

2009-03-15  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=19737
        No cursor and paste not enabled right clicking text field/area

        Test: fast/events/right-click-focus.html

        * page/EventHandler.cpp:
        (WebCore::EventHandler::handleMousePressEventSingleClick):
        (WebCore::EventHandler::handleMousePressEvent):
        Take normal code path for right clicks (we were taking it for Ctrl-clicks anyway).

2009-03-15  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Adele Peterson.

        https://bugs.webkit.org/show_bug.cgi?id=23949
        HTMLSelectElement is in inconsistent state when handling mutation events

        Test: fast/forms/mutation-event-recalc.html

        * dom/ContainerNode.cpp:
        (WebCore::ContainerNode::insertBefore):
        (WebCore::ContainerNode::appendChild):
        Call childrenChanged() before dispatching modification events, not after.

        * html/HTMLOptGroupElement.cpp:
        (WebCore::HTMLOptGroupElement::insertBefore):
        (WebCore::HTMLOptGroupElement::replaceChild):
        (WebCore::HTMLOptGroupElement::removeChild):
        (WebCore::HTMLOptGroupElement::appendChild):
        (WebCore::HTMLOptGroupElement::removeChildren):
        * html/HTMLSelectElement.cpp:
        (WebCore::HTMLSelectElement::add):
        (WebCore::HTMLSelectElement::remove):
        (WebCore::HTMLSelectElement::insertBefore):
        (WebCore::HTMLSelectElement::replaceChild):
        (WebCore::HTMLSelectElement::removeChild):
        (WebCore::HTMLSelectElement::appendChild):
        (WebCore::HTMLSelectElement::removeChildren):
        Remove calls to recalcSelectOptions(). It is too late to recalc now, after mutation events
        were already dispatched.

2009-03-14  Greg Bolsinga  <bolsinga@apple.com>

        <rdar://problem/6683465>

        Geolocation has to be able to handle NULL Frames.

        Reviewed by Mark Rowe
    
2009-03-14  Greg Bolsinga  <bolsinga@apple.com>

        Reviewed by David Kilzer.

        Geoposition::toString was missing the Coordinates after the update.

        * page/Geoposition.cpp:
        (WebCore::Geoposition::toString):

2009-03-14  Jan Michael Alonzo  <jmalonzo@webkit.org>

        Reviewed by Mark Rowe.

        [Gtk] Build fix - pass a Coordinate to Geoposition::create
        https://bugs.webkit.org/show_bug.cgi?id=24603

        Gtk build fix per r41650
        https://bugs.webkit.org/show_bug.cgi?id=24506
        Geolocation in Safari differs from the spec, no Coordinates attribute on Position

        Pass a Coordinate to Geoposition::create.

        * platform/gtk/GeolocationServiceGtk.cpp:
        (WebCore::GeolocationServiceGtk::updatePosition):

2009-03-14  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=24545
        Make cross-site XHR simple request definition match current spec draft

        Tests: http/tests/xmlhttprequest/access-control-basic-post-fail-non-simple-content-type.html
               http/tests/xmlhttprequest/access-control-basic-whitelist-request-headers.html

        * loader/CrossOriginAccessControl.cpp:
        (WebCore::isOnAccessControlSimpleRequestHeaderWhitelist): Added Content-Language.
        (WebCore::isSimpleCrossOriginAccessRequest): Added HEAD. Restricted content types to those
        that could be sent via form submission.

2009-03-14  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=15172
        object fallback - empty string for first argument of setAttributeNS does not work like null

        https://bugs.webkit.org/show_bug.cgi?id=24548
        createElementNS("", name) should create an element in null namespace

        Tests: fast/dom/createElementNS-empty-namespace.html
               fast/dom/setAttributeNS-empty-namespace.html

        * dom/QualifiedName.h: (WebCore::QualifiedName::QualifiedNameImpl::QualifiedNameImpl):
        Convert empty namespaces to null.

2009-03-13  Mark Rowe  <mrowe@apple.com>

        Rubber-stamped by Dan Bernstein.

        Take advantage of the ability of recent versions of Xcode to easily switch the active
        architecture.

        * Configurations/DebugRelease.xcconfig:

2009-03-13  John Abd-El-Malek  <jam@google.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=24593
        Added requestorID so we can track the request to its WebView without using frame (which was removed).

        * platform/network/chromium/ResourceRequest.h:
        (WebCore::ResourceRequest::ResourceRequest):
        (WebCore::ResourceRequest::requestorID):
        (WebCore::ResourceRequest::setRequestorID):
        (WebCore::ResourceRequest::setTargetType):
        (WebCore::ResourceRequest::policyURL):
        (WebCore::ResourceRequest::setPolicyURL):
        (WebCore::ResourceRequest::requestorPid):
        (WebCore::ResourceRequest::setRequestorPid):

2009-03-13  Mike Belshe  <mike@belshe.com>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24561
        Add custom V8 bindings for HTMLElementCanvas, Location.

        * bindings/v8/custom/V8HTMLCanvasElementCustom.cpp: Added.
        * bindings/v8/custom/V8LocationCustom.cpp: Added.

2009-03-13  Stephen White  <senorblanco@chromium.org>

        Reviewed by Eric Seidel.

        https://bugs.webkit.org/show_bug.cgi?id=24584

        Fix transparent text rendering on Chromium.  FontChromiumWin
        was calling beginTransparencyLayer()/endTransparencyLayer(), with
        a TransparencyWin inside to do GDI ClearType rendering over an
        opaque background.  TransparencyWin does its special sauce
        in the destructor, but it was being called too late to be used
        correctly in the layer.  Put the special sauce into a new function,
        composite(), and call that explicitly instead.

        * platform/graphics/chromium/FontChromiumWin.cpp:
        (WebCore::TransparencyAwareFontPainter::TransparencyAwareFontPainter::~TransparencyAwareFontPainter):
        * platform/graphics/chromium/TransparencyWin.cpp:
        (WebCore::TransparencyWin::~TransparencyWin):
        (WebCore::TransparencyWin::composite):
        (WebCore::TransparencyWin::init):
        * platform/graphics/chromium/TransparencyWin.h:
        * rendering/RenderThemeChromiumWin.cpp:
        (WebCore::):

2009-03-13  Mike Belshe  <mike@belshe.com>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24562
        Add custom implementation for getCSSCanvasContext to V8 bindings.

        * bindings/v8/custom/V8DocumentCustom.cpp: Added new method.

2009-03-13  Mike Belshe  <mike@belse.com>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24580
        Fix query() to match KURL behavior.

        * platform/KURLGoogle.cpp:
        (WebCore::KURL::query): remove extra logic around question mark.

2009-03-13  Chris Fleizach  <cfleizach@apple.com>

        Reviewed by Beth Dakin.

        Bug 24474: AX: in multi-body tables, asking for a cell at a specific coordinate can return nil
        https://bugs.webkit.org/show_bug.cgi?id=24474

        Test: platform/mac-snowleopard/accessibility/table-multi-bodies.html

        * page/AccessibilityTable.cpp:
        (WebCore::AccessibilityTable::cellForColumnAndRow):

2009-03-13  Jian Li  <jianli@chromium.org>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24589
        Upstream changes to V8 event listeners (Chromium r11133) in order to
        fix worker functionality break in Chromium.

        * bindings/v8/V8AbstractEventListener.cpp:
        (WebCore::V8AbstractEventListener::invokeEventHandler):
        (WebCore::V8AbstractEventListener::handleEvent):
        * bindings/v8/V8AbstractEventListener.h:
        * bindings/v8/V8WorkerContextEventListener.cpp:
        (WebCore::V8WorkerContextEventListener::handleEvent):

2009-03-13  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Timothy Hatcher.

        https://bugs.webkit.org/show_bug.cgi?id=24524
        Introduce ScriptObject and ScriptFunctionCall abstractions.

        * GNUmakefile.am: Added ScriptObject and ScriptFunctionCall to project.
        * WebCore.pro: Ditto.
        * WebCore.vcproj/WebCore.vcproj: Ditto.
        * WebCore.xcodeproj/project.pbxproj: Ditto.
        * WebCoreSources.bkl: Ditto.
        * bindings/js/ScriptFunctionCall.cpp: Added.
        * bindings/js/ScriptFunctionCall.h: Added.
        * bindings/js/ScriptObject.cpp: Added.
        * bindings/js/ScriptObject.h: Added.
        * bindings/js/ScriptObjectQuarantine.cpp: Added.
        * bindings/js/ScriptObjectQuarantine.h: Added.
        * inspector/InspectorController.cpp:
        (WebCore::InspectorController::populateScriptObjects): Changed to use InspectorDatabaseResource::bind.
        (WebCore::InspectorController::resetScriptObjects): Changed to use InspectorDatabaseResource::unbind.
        (WebCore::InspectorController::didOpenDatabase): Changed to use InspectorDatabaseResource::unbind.
        * inspector/InspectorController.h: Removed addScriptDatabaseResource and
        removeScriptDatabaseResource declarations.
        * inspector/InspectorDatabaseResource.cpp:
        (WebCore::InspectorDatabaseResource::InspectorDatabaseResource):
        (WebCore::InspectorDatabaseResource::bind): Added.
        (WebCore::InspectorDatabaseResource::unbind): Added.
        * inspector/InspectorDatabaseResource.h: Added bind and unbind declarations.

2009-03-13  Peter Kasting  <pkasting@google.com>

        Reviewed by Eric Seidel.

        https://bugs.webkit.org/show_bug.cgi?id=24467
        Make Skia drawRect() and stroke behavior match CG.

        This makes drawRect() ignore the stroke width (like CG does), and
        adds a warning comment about that to the appropriate header.

        It also eliminates some hacky code in Skia's stroke preparation,
        which tried to adjust odd-width strokes to fall on pixel boundaries.
        Not only did this not match CG, it wouldn't necessarily work right,
        because there could be other transforms (e.g. full-page zoom) that
        would affect the stroke before it reached the device pixel level.

        * platform/graphics/GraphicsContext.h:
        * platform/graphics/skia/GraphicsContextSkia.cpp:
        (WebCore::GraphicsContext::drawRect):
        (WebCore::GraphicsContext::fillRect):
        * platform/graphics/skia/PlatformContextSkia.cpp:
        (PlatformContextSkia::drawRect):
        (PlatformContextSkia::setupPaintForStroking):

2009-03-13  Jian Li  <jianli@chromium.org>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24583
        Need to change scope of constructor and destructor of V8ObjectEventListener from private
        to protected so that they could be used in its derived class V8WorkerContextEventListener.

        * bindings/v8/V8ObjectEventListener.h:

2009-03-13  David Levin  <levin@chromium.org>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24579
        Many V8*Event* files are misplaced.

        This code is infrastructure for supporting the bindings and thus should
        be in the bindings/v8 directory.

        * bindings/v8/V8AbstractEventListener.cpp: Renamed from WebCore/bindings/v8/custom/V8AbstractEventListener.cpp.
        * bindings/v8/V8AbstractEventListener.h: Renamed from WebCore/bindings/v8/custom/V8AbstractEventListener.h.
        * bindings/v8/V8LazyEventListener.cpp: Renamed from WebCore/bindings/v8/custom/V8LazyEventListener.cpp.
        * bindings/v8/V8LazyEventListener.h: Renamed from WebCore/bindings/v8/custom/V8LazyEventListener.h.
        * bindings/v8/V8ObjectEventListener.cpp: Renamed from WebCore/bindings/v8/custom/V8ObjectEventListener.cpp.
        * bindings/v8/V8ObjectEventListener.h: Renamed from WebCore/bindings/v8/custom/V8ObjectEventListener.h.
        * bindings/v8/V8WorkerContextEventListener.cpp: Renamed from WebCore/bindings/v8/custom/V8WorkerContextEventListener.cpp.
        * bindings/v8/V8WorkerContextEventListener.h: Renamed from WebCore/bindings/v8/custom/V8WorkerContextEventListener.h.

2009-03-13  Adele Peterson  <adele@apple.com>

        Reviewed by Justin Garcia.

        Fix for <rdar://problem/5089327> Color of quoted content is wrong when pasted inside other quoted content

        Test: editing/pasteboard/5089327.html

        Handle spans being pasted within a quoted region in the same way we handle "paste as quotation" content.

        * editing/ReplaceSelectionCommand.cpp:
        (WebCore::handleStyleSpansBeforeInsertion):
        (WebCore::ReplaceSelectionCommand::handleStyleSpans):

2009-03-13  Greg Bolsinga  <bolsinga@apple.com>

        Reviewed by Simon Fraser.

        Update Geolocation perimission dialogs to be asynchronous.
        https://bugs.webkit.org/show_bug.cgi?id=24505

        Geolocation now requests permission from the Chrome asynchronously.
        The Chrome is passed the Geolocation object, and the Chrome sets the permission
        on the Geolocation. Geolocation also tracks if the Chrome should clear its cache
        of SecurityOrigins with geolocation permission. This is so that the GeolocationService
        can inform the Chrome of its request, and the Chrome is also free to implement its
        own policy.
        
        * WebCore.base.exp:
        * WebCore.xcodeproj/project.pbxproj:
        * page/Chrome.cpp:
        (WebCore::Chrome::requestGeolocationPermissionForFrame):
        * page/Chrome.h:
        * page/ChromeClient.h:
        (WebCore::ChromeClient::requestGeolocationPermissionForFrame):
        * page/Geolocation.cpp:
        (WebCore::Geolocation::Geolocation):
        (WebCore::Geolocation::getCurrentPosition):
        (WebCore::Geolocation::watchPosition):
        (WebCore::Geolocation::setIsAllowed):
        (WebCore::Geolocation::displayChallengeIfNecessary):
        (WebCore::Geolocation::geolocationServicePositionChanged):
        * page/Geolocation.h:
        (WebCore::Geolocation::isAllowed):
        (WebCore::Geolocation::setShouldClearCache):
        (WebCore::Geolocation::shouldClearCache):
        (WebCore::Geolocation::):

2009-03-13  Anders Carlsson  <andersca@apple.com>

        Reviewed by Dan Bernstein.

        <rdar://problem/6610666> Revise the Cocoa event model text API
        
        Add a NPCocoaEventTextInput event type. Remove the text input variables.
        
        * bridge/npapi.h:
        (_NPCocoaEvent::):

2009-03-13  Kevin Decker  <kdecker@apple.com>

        Reviewed by Anders.

        <rdar://problem/6630340> REGRESSION (39114-39115): Unity Web Player no longer works if Flip4Mac is also installed
        
        The code assumed if we have a plug-in that supports "application/x-oleobject" we should always prefer the object tag
        over of an embed tag. That assumption can cause the Mac platform to load the wrong plug-in, as Flip4Mac claims supports
        for x-oleobject.
        
        * rendering/RenderPartObject.cpp:
        (WebCore::shouldUseEmbedDescendant): Made the Mac platform always return true here. 

2009-03-13  Jian Li  <jianli@chromium.org>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24559
        Need to port JSC fix (r41565) to V8 in order to fix layout test onload-single-line-comment.html.

        * bindings/v8/custom/V8LazyEventListener.cpp:
        (WebCore::V8LazyEventListener::getListenerFunction):
        (WebCore::V8LazyEventListener::getWrappedListenerFunction):

2009-03-13  Jian Li  <jianli@chromium.org>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24557
        This is to support running multiple workers in a single worker process in chromium.

        * bindings/v8/custom/V8WorkerContextEventListener.cpp:
        (WebCore::V8WorkerContextEventListener::handleEvent):
        Add locker to v8 event listener of worker context.

2009-03-13  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Sam Weinig.

        https://bugs.webkit.org/show_bug.cgi?id=24575
        XHR response is incorrect after a network error.

        Tests: http/tests/xmlhttprequest/cross-site-denied-response-sync-2.html
               http/tests/xmlhttprequest/cross-site-denied-response-sync.html
               http/tests/xmlhttprequest/cross-site-denied-response.html

        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::networkError): Call internalAbort() to fully reset the request.
        (WebCore::XMLHttpRequest::didFailRedirectCheck): No need to call internalAbort() here. Note
        that since internalAbort() can drop GC protection, it is not safe to use the object after
        this call.
        (WebCore::XMLHttpRequest::didReceiveData): Check that the request wasn't aborted, and
        return early if it was. This can happen during sync requests, as the loader does not know
        that it was aborted, and just synthesizes all callbacks.

2009-03-13  Adam Bergkvist <adam.bergkvist@ericsson.com>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=3547
        XMLHttpRequest.statusText returns always "OK"

        Covered by existing tests (which now pass on all platforms but Mac).

        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::statusText): Return ResourceResponse status text. It is now up
        to each platform to correctly set the status text or set it to "OK" to retain current
        behavior.

2009-03-13  Adam Bergkvist <adam.bergkvist@ericsson.com>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=24349
        [QT] HTTP status text is never set

        Set HTTP status text to the reason phrase attribute of QNetworkReply.

        * platform/network/qt/QNetworkReplyHandler.cpp:
        (WebCore::QNetworkReplyHandler::sendResponseIfNeeded):

2009-03-12  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Mark Rowe

        <rdar://problem/6622300>: Reproducible crash on
        <http://www.editgrid.com/explore/tnc/dave/FusionChart%3A_Candlestick>
        
        Prevent CSSStyleSheet::checkLoaded() writing to freed memory when it gets
        deleted from under itself. The sheetLoaded() notification can allow scripts
        to run via HTMLTokenizer::executeScriptsWaitingForStylesheets(),
        which can cause the last ref to the CSSStyleSheet to be released.

        * css/CSSStyleSheet.cpp:
        (WebCore::CSSStyleSheet::checkLoaded):

2009-03-12  Kevin Ollivier  <kevino@theolliviers.com>

        wx build fix. Allow make-generated-sources.sh to pass args down to DerivedSources.make

        * make-generated-sources.sh:

2009-03-12  Adam Treat  <adam.treat@torchmobile.com>

        Reviewed by Oliver Hunt.

        https://bugs.webkit.org/show_bug.cgi?id=24498
        Fix the Qt port to use the same algorithm for drawing dashed and dotted
        borders as the other ports.  This makes the Qt port pixel-for-pixel perfect
        compared to border drawing with Apple's canonical mac port and much closer
        to konqueror and firefox behavior.

        * platform/graphics/qt/GraphicsContextQt.cpp:
        (WebCore::GraphicsContext::drawLine):

2009-02-26  Eric Seidel  <eric@webkit.org>

        Reviewed by Justin Garcia.

        Clean up DOMSelection to use some helper functions
        making the code smaller, and less error-prone.
        https://bugs.webkit.org/show_bug.cgi?id=19221

        I tried moving DOMSelection off of rangeCompliantEquivalent
        but failed.  VisibleSelection holds positions like (table, 1) to mean
        "after the table".

        * page/DOMSelection.cpp:
        (WebCore::DOMSelection::visibleSelection):
        (WebCore::anchorPosition):
        (WebCore::focusPosition):
        (WebCore::basePosition):
        (WebCore::extentPosition):
        (WebCore::DOMSelection::anchorNode):
        (WebCore::DOMSelection::anchorOffset):
        (WebCore::DOMSelection::focusNode):
        (WebCore::DOMSelection::focusOffset):
        (WebCore::DOMSelection::baseNode):
        (WebCore::DOMSelection::baseOffset):
        (WebCore::DOMSelection::extentNode):
        (WebCore::DOMSelection::extentOffset):
        * page/DOMSelection.h:

2009-03-12  Dmitry Titov  <dimich@chromium.org>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24563
        Change ResourceRequest to be a struct rather then a class to match other declarations.

        * platform/network/chromium/ResourceRequest.h:

2009-03-12  David Levin  <levin@chromium.org>

        Reviewed by  Darin Fisher.

        Bug 24530: width100percent-searchfield.html should be fixed for chromium.
        <https://bugs.webkit.org/show_bug.cgi?id=24530>

        Compensate for r39924 in chromium, which broke LayoutTests/fast/replaced/width100percent-searchfield.html

        * css/themeChromiumWin.css: Added.
        Overrides the changes to start supporting input[type="search"] on Windows (since
        Chromium doesn't do anything special for this).

        * rendering/RenderThemeChromiumWin.cpp:
        (WebCore::supportsFocus):
        (WebCore::RenderThemeChromiumWin::extraDefaultStyleSheet):
        (WebCore::RenderThemeChromiumWin::determineState):
        (WebCore::RenderThemeChromiumWin::getThemeData):

2009-03-12  Greg Bolsinga  <bolsinga@apple.com>

        Reviewed by Antti Koivisto.

        https://bugs.webkit.org/show_bug.cgi?id=24506
        Geolocation in Safari differs from the spec, no Coordinates attribute on Position
        
        Get up to date.

        * DerivedSources.make:
        * GNUmakefile.am:
        * WebCore.pro:
        * WebCore.scons:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:
        * bindings/js/JSGeolocationCustom.cpp:
        (WebCore::createPositionOptions):
        * page/Coordinates.cpp: Added.
        (WebCore::Coordinates::toString):
        * page/Coordinates.h: Added.
        (WebCore::Coordinates::create):
        (WebCore::Coordinates::latitude):
        (WebCore::Coordinates::longitude):
        (WebCore::Coordinates::altitude):
        (WebCore::Coordinates::accuracy):
        (WebCore::Coordinates::altitudeAccuracy):
        (WebCore::Coordinates::heading):
        (WebCore::Coordinates::speed):
        (WebCore::Coordinates::Coordinates):
        * page/Coordinates.idl: Added.
        * page/Geoposition.cpp:
        (WebCore::Geoposition::toString):
        * page/Geoposition.h:
        (WebCore::Geoposition::create):
        (WebCore::Geoposition::coords):
        (WebCore::Geoposition::Geoposition):
        * page/Geoposition.idl:
        * page/PositionOptions.h:
        (WebCore::PositionOptions::create):
        (WebCore::PositionOptions::maximumAge):
        (WebCore::PositionOptions::setMaximumAge):
        (WebCore::PositionOptions::PositionOptions):
        * platform/mac/GeolocationServiceMac.mm:
        (-[WebCoreCoreLocationObserver locationManager:didUpdateToLocation:fromLocation:]):

2009-03-12  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=24496
        Fix console logging of non-string values by coercing the argument to
        a string, for V8 bindings.

        Test: fast/js/console-non-string-values.html

        * bindings/v8/ScriptValue.cpp: Added toString method.
        (WebCore::ScriptValue::toString):
        * bindings/v8/ScriptValue.h: Added PlatformString, ScriptState includes
            and toString declaration.

2009-03-12  Darin Fisher  <darin@chromium.org>

        Reviewed by Dimitri Glazkov.

        Eliminate ChromiumBridge::uiResourceProtocol
        https://bugs.webkit.org/show_bug.cgi?id=24558

        * platform/chromium/ChromiumBridge.h:

2009-03-12  Greg Bolsinga  <bolsinga@apple.com>

        Reviewed by Simon Fraser.

        Missing breaks in switch statement
        https://bugs.webkit.org/show_bug.cgi?id=24556

        * platform/mac/GeolocationServiceMac.mm:
        (-[WebCoreCoreLocationObserver locationManager:didFailWithError:]):

2009-03-12  Darin Fisher  <darin@chromium.org>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24554
        Remove some unused functions from ChromiumBridge.

        * platform/chromium/ChromiumBridge.h:

2009-03-02  Eric Seidel  <eric@webkit.org>

        Reviewed by Darin Fisher and Justin Garcia.

        Safari crashes during drag and drop in Google presentations
        due to mutation event handlers removing DOM content during insertNode
        https://bugs.webkit.org/show_bug.cgi?id=22634

        Added a bunch of "null" checks to make sure nodes are still
        in the document before we operate on them.  This is an
        inelegant solution, but it's the best we have for now.
        
        Test: editing/selection/crash-on-drag-with-mutation-events.html

        * editing/CompositeEditCommand.cpp:
        (WebCore::CompositeEditCommand::insertNodeAt):
        * editing/ReplaceSelectionCommand.cpp:
        (WebCore::ReplaceSelectionCommand::doApply):

2009-03-12  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Adam Treat.

        https://bugs.webkit.org/show_bug.cgi?id=24525
        REGRESSION: Inspector window doesn't close when inspected page is
        destroyed. This is a revert of r41158, which became unnecessary when
        InspectorController became ref-counted in r41462.

        * inspector/InspectorController.cpp:
        (WebCore::InspectorController::inspectedPageDestroyed): Reset m_inspectedPage
            after calling close().
        (WebCore::InspectorController::stopUserInitiatedProfiling): Remove
            m_inspectedPage check guard around profile logic.

2009-03-12  Peter Kasting  <pkasting@google.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=24502
        Make horizontal scrolling on Windows always go the correct direction.

        * platform/PlatformWheelEvent.h:
        * platform/win/WheelEventWin.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):

2009-03-12  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Simon Fraser.

        https://bugs.webkit.org/show_bug.cgi?id=24496
        Fix console logging of non-string values by coercing the argument to
        a string.

        Test: fast/js/console-non-string-values.html

        * bindings/js/ScriptValue.cpp: Removed PlatformString include.
        * bindings/js/ScriptValue.h: Added toString method.
        (WebCore::ScriptValue::toString):
        * page/Console.cpp:
        (WebCore::getFirstArgumentAsString): Changed firstArgumentAsString method to use
            ScriptValue::toString.
        (WebCore::Console::addMessage): added extra ScriptState argument to callsite.
        (WebCore::Console::count): Ditto.

2009-03-12  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Adele Peterson.
        
        https://bugs.webkit.org/show_bug.cgi?id=24508
        
        Fix updating of text field with placeholder text when value is set
        by ensuring that updatePlaceholderVisibility() is called after
        the value has been updated, not before.

        Test: fast/forms/placeholder-set-value.html

        * html/HTMLInputElement.cpp:
        (WebCore::HTMLInputElement::setValue):

2009-03-12  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dave Hyatt
        
        https://bugs.webkit.org/show_bug.cgi?id=24552
        
        Hit testing involving normal flow, self-painting layers (e.g. reflections)
        was broken because it was not null-testing the result of hitTestLayer(),
        so bailing early with a nil hit layer.

        Test: fast/layers/normal-flow-hit-test.html

        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::hitTestLayer):

2009-03-12  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Cameron Zwarich.

        Removed zero-sized files, left over from improperly applied patch.

        * platform/graphics/chromium/ThemeHelperChromiumWin.cpp: Removed.
        * platform/graphics/chromium/ThemeHelperChromiumWin.h: Removed.

2009-03-12  Adam Treat  <adam.treat@torchmobile.com>

        Reviewed by George Staikos.

        Do not ignore alpha color in optimization.  Hard to believe, but there
        are websites that fill the entire page with a tiled image consisting of
        nothing but a 100% transparent 1x1 image. <cough>orbitz.com</cough>

        * platform/graphics/qt/ImageQt.cpp:
        (WebCore::BitmapImage::checkForSolidColor):

2009-03-12  Julien Chaffraix  <jchaffraix@webkit.org>

        Reviewed by Darin Adler.

        Bug 24110: cloneNode should call cloneElement and not the reverse

        - Splitted the code from cloneNode into cloneElementWithChildren and cloneElementWithChildren.
          Now cloneNode calls one of the 2 previous methods.

        - Renamed cloneElement to cloneElementWithoutChildren as it was the previous behaviour.

        - Moved cloneNode to the Element private section so that WebCore callers cannot use it.

        - Removed Element::cloneNode usage through WebCore.

        * dom/Element.cpp:
        (WebCore::Element::cloneNode): Moved to Element's private section and it
        now calls the two next methods.
        (WebCore::Element::cloneElementWithChildren): Added.
        (WebCore::Element::cloneElementWithoutChildren): Renamed from cloneElement
        to avoid ambiguity.
        * dom/Element.h:

        * editing/ApplyStyleCommand.cpp:
        (WebCore::ApplyStyleCommand::addInlineStyleIfNeeded): Changed call to cloneElement
        to call to cloneElementWithoutChildren.
        * editing/BreakBlockquoteCommand.cpp:
        (WebCore::BreakBlockquoteCommand::doApply): Ditto.
        * editing/IndentOutdentCommand.cpp:
        (WebCore::IndentOutdentCommand::indentRegion): Ditto.
        * editing/InsertParagraphSeparatorCommand.cpp:
        (WebCore::InsertParagraphSeparatorCommand::doApply): Ditto.
        * editing/ModifySelectionListLevel.cpp:
        (WebCore::IncreaseSelectionListLevelCommand::doApply): Ditto.
        * editing/SplitElementCommand.cpp:
        (WebCore::SplitElementCommand::doApply): Ditto.
        * editing/markup.cpp:
        (WebCore::createFragmentFromText): Ditto.
        * svg/SVGUseElement.cpp:
        (WebCore::SVGUseElement::buildShadowTree): Ditto.
        (WebCore::SVGUseElement::expandUseElementsInShadowTree): Ditto.

2009-03-12  Dirk Schulze  <krit@webkit.org>

        Reviewed by Oliver Hunt.

        Fixed a crash on the 1x1 pixel check for background image in gtk.
        We need to load the image first and need to be sure that the image
        is not null before checking the type. 

        * platform/graphics/cairo/ImageCairo.cpp:
        (WebCore::BitmapImage::checkForSolidColor):

2009-03-12  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=24544
        Rename setUrl to setURL

        Use proper case for setUrl method, rename it to setURL.

        * platform/network/ResourceResponseBase.cpp:
        (WebCore::ResourceResponseBase::adopt):
        * platform/network/ResourceResponseBase.h:
        * platform/network/curl/ResourceHandleManager.cpp:
        (WebCore::writeCallback):
        (WebCore::headerCallback):
        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::fillResponseFromMessage):
        (WebCore::queryInfoCallback):

2009-03-12  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=24519
        [GTK] Use two argument KURL ctor in ResourceHandleSoup

        Use two argument KURL ctor, the single argument ctor expects its
        input to be the output of a previous KURL::parse call, which is
        not the case here.

        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::fillResponseFromMessage):
        (WebCore::queryInfoCallback):

2009-03-11  Jon Honeycutt  <jhoneycutt@apple.com>

        Fix for <rdar://6418681>
        https://bugs.webkit.org/show_bug.cgi?id=22644

        Reviewed by Steve Falkenburg.

        * plugins/win/PluginPackageWin.cpp:
        (WebCore::PluginPackage::isPluginBlacklisted): Add the Citrix ICA
        Client plug-in to the blacklist; it requires a Mozilla-based browser.

2009-03-12  Oliver Hunt  <oliver@apple.com>

        Reviewed by NOBODY (build fix).

        build fix.

        * html/CanvasRenderingContext2D.h:

2009-03-11  Oliver Hunt  <oliver@apple.com>

        Reviewed by Adele Peterson.

        <rdar://problem/6673852> Widget "Simple World Clock" does not display with Safari 4

        A number of dashboard widgets rely on canvas failing silently when given invalid
        colors for gradient stops, but both the spec and firefox throw exceptions here.
        So we work around this by creating a dashboard quirk that will only suppress the
        exception in dashboard compatibility mode.

        Test: fast/canvas/canvas-gradient-addStop-error.html

        * html/CanvasGradient.cpp:
        (WebCore::CanvasGradient::CanvasGradient):
        (WebCore::CanvasGradient::addColorStop):
        * html/CanvasGradient.h:
        (WebCore::CanvasGradient::setDashboardCompatibilityMode):
        * html/CanvasRenderingContext2D.cpp:
        (WebCore::CanvasRenderingContext2D::prepareGradientForDashboard):
        (WebCore::CanvasRenderingContext2D::createLinearGradient):
        (WebCore::CanvasRenderingContext2D::createRadialGradient):
        * html/CanvasRenderingContext2D.h:

2009-03-11  David Kilzer  <ddkilzer@apple.com>

        Clarify comments regarding order of FEATURE_DEFINES

        Rubber-stamped by Mark Rowe.

        * Configurations/WebCore.xcconfig: Added warning about the
        consequences when FEATURE_DEFINES are not kept in sync.

2009-03-11  Gustavo Noronha Silva  <gns@gnome.org>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24515
        [GTK] soup backend should use GMappedFile instead of mmap directly

        Rewrite mmap code used to upload files to use GMappedFile, so that
        it is platform-independent.

        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::ResourceHandle::startHttp):

2009-03-11  Dan Bernstein  <mitz@apple.com>

        Reviewed by Darin Adler.

        - WebCore part of fixing <rdar://problem/3919124> Thai text selection
          in Safari is incorrect

        Test: platform/mac/editing/selection/word-thai.html

        Tested on Mac, but should work on all ICU platforms (or more precisely
        on all platforms that implement
        WTF::Unicode::hasLineBreakingPropertyComplexContext() correctly).

        * editing/TextIterator.cpp:
        (WebCore::BackwardsCharacterIterator::BackwardsCharacterIterator):
        Added.
        (WebCore::BackwardsCharacterIterator::range): Added.
        (WebCore::BackwardsCharacterIterator::advance): Added.

        * editing/TextIterator.h: Added BackwardsCharacterIterator with the
        minimal functionality required for this patch.
        (WebCore::BackwardsCharacterIterator::atEnd): Added.

        * editing/visible_units.cpp:
        (WebCore::firstNonComplexContextLineBreak): Added this helper function
        that returns the index of the first character in the string whose
        Unicode line breaking property value is not SA ("Complex Context"), or
        the length of the string if there is no such character.
        (WebCore::lastNonComplexContextLineBreak): Added this helper function
        that returns the index of the last character in the string whose
        Unicode line breaking property value is not SA ("Complex Context"), or
        -1 if there is no such character.
        (WebCore::previousBoundary): Changed the signature of the search
        function to include an offset parameter. Renamed the 'exception' local
        variable to 'ec' and changed its type to ExceptionCode. Extend the
        string forwards until the first character with Unicode line breaking
        property value other than SA. This gives the boundary search function
        enough context in the forward direction. Changed to use a
        BackwardsCharacterIterator for translating the backwards offset into
        a position.
        (WebCore::nextBoundary): Changed the signature of the search
        function to include an offset parameter. Extend the string backwards
        until the first character with Unicode line breaking property value
        other than SA. This gives the boundary search function enough context in
        the backwards direction. Restricted the workaround for <rdar://5192593>
        only to the line break case, because it was causing an extra character
        to be selected following a word that ended with a combining mark.
        (WebCore::startWordBoundary): Added an offset parameter. Check that
        going backwards from the given offset, there is a character with Unicode
        line breaking property value other than SA, and otherwise return 0 to
        request more context.
        (WebCore::endWordBoundary): Similar, but in reverse.
        (WebCore::previousWordPositionBoundary): Similar.
        (WebCore::nextWordPositionBoundary): Similar.
        (WebCore::startSentenceBoundary): Updated for the new search function
        signature.
        (WebCore::endSentenceBoundary): Ditto.
        (WebCore::previousSentencePositionBoundary): Ditto.
        (WebCore::nextSentencePositionBoundary): Ditto.

2009-03-11  Darin Adler  <darin@apple.com>

        Reviewed by Mark Rowe.

        Don't use "::" in thread names because it is ugly!

        * loader/icon/IconDatabase.cpp:
        (WebCore::IconDatabase::open): Use "WebCore: " instead of "WebCore::".
        * platform/network/cf/ResourceHandleCFNet.cpp:
        (WebCore::ResourceHandle::loaderRunLoop): Call this "WebCore: CFNetwork Loader" to make
        it clear this is a WebCore-created thread.
        * storage/DatabaseThread.cpp:
        (WebCore::DatabaseThread::start): Use "WebCore: " instead of "WebCore::".
        * storage/LocalStorageThread.cpp:
        (WebCore::LocalStorageThread::start): Use "WebCore: " instead of "WebCore::".
        * workers/WorkerThread.cpp:
        (WebCore::WorkerThread::start): Use "WebCore: " instead of "WebCore::".

2009-03-11  Mark Rowe  <mrowe@apple.com>

        Rubber-stamped by Darin Adler.

        <rdar://problem/6653286> WebCore's implementation of screenIsMonochrome is needlessly complex

        * platform/mac/PlatformScreenMac.mm:
        (WebCore::screenIsMonochrome): Return false.

2009-03-11  Beth Dakin  <bdakin@apple.com>

        Reviewed by Dan Bernstein.

        Fix for https://bugs.webkit.org/show_bug.cgi?id=14749 percentage 
        top value on position:relative child not calculated correctly if 
        parent has percentage height
        - and corresponding -
        <rdar://problem/6172925>

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

2009-03-11  Jian Li  <jianli@chromium.org>

        Reviewed by Dimitri Glazkov.

        Implement ScriptValue::hasNoValue for V8.
        https://bugs.webkit.org/show_bug.cgi?id=24507

        * bindings/v8/ScriptValue.h:
        (WebCore::ScriptValue::hasNoValue):

2009-03-11  David Levin  <levin@chromium.org>

        Reviewed by Dimitri Glazkov.

        Bug 24459: Add v8 bindings for event.
        <https://bugs.webkit.org/show_bug.cgi?id=24459>

        Adding custom event bindings.

        * bindings/v8/custom/V8AbstractEventListener.cpp: Added.
        * bindings/v8/custom/V8AbstractEventListener.h: Added.
        * bindings/v8/custom/V8AttrCustom.cpp:
        Fixed messed up copyright header.
        * bindings/v8/custom/V8CustomEventListener.cpp: Added.
        * bindings/v8/custom/V8CustomEventListener.h:
        Replaced forwarding header with the real implementation.
        * bindings/v8/custom/V8ElementCustom.cpp:
        * bindings/v8/custom/V8LazyEventListener.cpp: Added.
        * bindings/v8/custom/V8LazyEventListener.h: Added.
        * bindings/v8/custom/V8ObjectEventListener.cpp: Added.
        * bindings/v8/custom/V8ObjectEventListener.h: Added.
        * bindings/v8/custom/V8WorkerContextEventListener.cpp: Added.
        * bindings/v8/custom/V8WorkerContextEventListener.h: Added.

2009-03-11  Darin Fisher  <darin@chromium.org>

        Reviewed by Dimitri Glazkov.

        https://bugs.webkit.org/show_bug.cgi?id=24521
        Add some missing includes for the non-JSC build.

        * platform/text/PlatformString.h: added OwnPtr.h
        * xml/XMLHttpRequest.cpp: added SecurityOrigin.h

2009-03-11  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Darin Adler

        https://bugs.webkit.org/show_bug.cgi?id=24531
        
        Make sure that AnimationControllerPrivate::getAnimatedStyleForRenderer() never
        returns a null style if it has a valid renderer, which could happen if a
        CompositeAnimation existed, but wasn't running any animations or transitions.

        * page/animation/AnimationController.cpp:
        (WebCore::AnimationControllerPrivate::getAnimatedStyleForRenderer):

2009-03-11  Simon Fraser  <simon.fraser@apple.com>

        Fix Scons build. Patch from Alexander Guy.

        * WebCore.scons:

2009-03-11  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dan Bernstein
       
        https://bugs.webkit.org/show_bug.cgi?id=23093
        
        When the cancel button in the search field is shown or hidden,
        we need to explicitly setStyle() on the cancel button's renderer,
        to ensure that repaint happens. Changing the style without telling
        the renderer won't work.

        Test: fast/repaint/search-field-cancel.html

        * rendering/RenderTextControlSingleLine.cpp:
        (WebCore::RenderTextControlSingleLine::subtreeHasChanged):
        (WebCore::RenderTextControlSingleLine::updateFromElement):
        (WebCore::RenderTextControlSingleLine::createCancelButtonStyle):
        (WebCore::RenderTextControlSingleLine::updateCancelButtonVisibility):
        (WebCore::RenderTextControlSingleLine::visibilityForCancelButton):
        * rendering/RenderTextControlSingleLine.h:

2009-03-11  Hironori Bono  <hbono@chromium.org>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=15790
        Ligatures aren't accounted for when manipulating VisiblePositions

        Changed cursorMovementIterator rules to prevent a cursor from moving in the middle of an
        Indic ligature which uses a virama sign.
        
        Test: editing/deleting/skip-virama-001.html

        * platform/text/TextBreakIteratorICU.cpp:
        (WebCore::cursorMovementIterator):

2009-03-11  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by George Staikos.

        https://bugs.webkit.org/show_bug.cgi?id=24513
        REGRESSION: Cursor movement doesn't work on Tiger (ICU 3.2)

        Fixes existing tests on Tiger.

        * platform/text/TextBreakIteratorICU.cpp: (WebCore::cursorMovementIterator):
        Use characterBreakIterator on Tiger (like we used to do).

2009-03-11  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Oliver Hunt.

        <rdar://problem/4072827> Downloaded non-ASCII file name becomes garbled

        * platform/network/mac/ResourceRequestMac.mm:
        (WebCore::ResourceRequest::doUpdatePlatformRequest): Correct selector name
        in instancesRespondToSelector.

2009-03-11  Jan Michael Alonzo  <jmalonzo@webkit.org>

        Reviewed by Holger Freyther.

        [GTK]DumpRenderTree doesn't compile for non-X11 GTK ports anymore
        https://bugs.webkit.org/show_bug.cgi?id=2260

        Add missing stdio.h include to get the build going

        * plugins/gtk/PluginPackageGtk.cpp:

2009-03-10  Justin Garcia  <justin.garcia@apple.com>

        Reviewed by Ken Kocienda.

        * dom/Range.cpp:
        (WebCore::Range::addLineBoxRects): If the range ended at [div, 0], addLineBoxRects
        would include rects for the divs contents.  Set the correct RenderObject to stop at.

2009-03-10  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Sam Weinig.

        <rdar://problem/6123770> Restrict access to document.cookie when making a cross-site XHR

        * xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::responseXML): Removed an incorrect
        comment about cookie support. Firefox doesn't expose cookies on responseXML at all, and
        there are security concerns with exposing them for cross-origin requests, so it's not clear
        if we want to change anything here.

2009-03-10  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Darin Adler

        https://bugs.webkit.org/show_bug.cgi?id=24503
        
        Fix hit testing of absolutely positioned single line text controls by
        ensuring that we set result.innerNode() correctly. If the hit node is
        a descendant of the inner text element or if it is the <input> itself,
        then we say we hit the innerTextElement.
        
        Rename hitInnerTextBlock() to hitInnerTextElement() to match the
        'innerTextElement' terminology used elsewhere.
        
        Assert that if renderer()->hitTest() returns false, no-one set
        result.innerNode().

        Test: fast/forms/search-abs-pos-cancel-button.html

        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::hitTestContents):
        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::hitInnerTextElement):
        * rendering/RenderTextControl.h:
        * rendering/RenderTextControlMultiLine.cpp:
        (WebCore::RenderTextControlMultiLine::nodeAtPoint):
        * rendering/RenderTextControlSingleLine.cpp:
        (WebCore::RenderTextControlSingleLine::nodeAtPoint):

2009-03-10  Antti Koivisto  <antti@apple.com>

        Reviewed by Mark Rowe.

        <rdar://problem/6668238> WebCore is registering text encodings needlessly from KURL constructor.
        
        Another case where we would trigger extended encoding loading needlessly.

        * platform/text/TextEncoding.cpp:
        (WebCore::TextEncoding::isNonByteBasedEncoding):

2009-03-10  Beth Dakin  <bdakin@apple.com>

        Reviewed by Darin Adler.

        Fix for <rdar://problem/6624769> REGRESSION (Safari 4 PB): No 
        scroll bar appears for long line of text with NOWRAP set

        This is a regression from http://trac.webkit.org/changeset/32226
        I talked with Dan about the original change, and we decided that 
        the best fix was to remove his small potential-optimization that 
        only created a separate line box for whitespace under certain 
        circumstances. This new code will always create a separate line 
        box.

        * rendering/bidi.cpp:
        (WebCore::RenderBlock::layoutInlineChildren):

2009-03-10  Sam Weinig  <sam@webkit.org>

        Reviewed by Geoffrey Garen.

        Fix for <rdar://problem/6166844>
        https://bugs.webkit.org/show_bug.cgi?id=24495

        Use same rule for loading java applets as we do for images.

        * html/HTMLAppletElement.cpp:
        (WebCore::HTMLAppletElement::createRenderer):
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::loadItem):

2009-03-10  Justin Garcia  <justin.garcia@apple.com>

        Reviewed by David Kilzer.

        Get rid of revealCaret and use revealSelection instead.

        * WebCore.base.exp:
        * editing/SelectionController.cpp:
        (WebCore::SelectionController::setSelection):
        * page/Frame.cpp:
        (WebCore::Frame::revealSelection):
        * page/Frame.h:

2009-03-10  Cameron Zwarich  <cwzwarich@uwaterloo.ca>

        Reviewed by Geoff Garen.

        Bug 23736: WebKit Crashes on http://g-conquest.fr/~server2
        <https://bugs.webkit.org/show_bug.cgi?id=23736>
        <rdar://problem/6560278>

        Before replacing the document, FrameLoader::executeIfJavaScriptURL()
        should call stopAllLoaders(), just like continueLoadAfterNavigationPolicy().

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::executeIfJavaScriptURL):

2009-03-10  Darin Adler  <darin@apple.com>

        Reviewed by Alexey Proskuryakov.

        Bug 24494: crash when deleting at end of document and merging paragraphs
        https://bugs.webkit.org/show_bug.cgi?id=24494
        rdar://problem/6571537

        Test: editing/deleting/merge-at-end-of-document.html

        * dom/Range.cpp:
        (WebCore::Range::compareBoundaryPoints): Split out assertion. It's better not to
        use && in assertions since we'd like to know which condition is failing.

        * editing/ApplyStyleCommand.cpp:
        (WebCore::ApplyStyleCommand::applyInlineStyleToRange): Added a null check before
        calling compareBoundaryPoints, since a 0 for the node is ambiguous and so the
        function doesn't know which value to return.

2009-03-10  Dirk Schulze  <krit@webkit.org>

        Reviewed by Eric Seidel. Math checked by Oliver Hunt.

        Implement the Cairo version of the checkForSolidColor() method. This halfes the
        time to draw on 1x1 px background images.
        I added two new calls to Color for the pixel manipulation on cairo_surface's.
        They are neede to premultiply/unpremultiply the colors of the surface.

        [CAIRO] Introduce single-pixel image optimizations
        https://bugs.webkit.org/show_bug.cgi?id=17284

        * platform/graphics/Color.cpp:
        (WebCore::colorFromPremultipliedARGB):
        (WebCore::premultipliedARGBFromColor):
        * platform/graphics/Color.h:
        * platform/graphics/cairo/ImageBufferCairo.cpp:
        (WebCore::ImageBuffer::getImageData):
        (WebCore::ImageBuffer::putImageData):
        * platform/graphics/cairo/ImageCairo.cpp:
        (WebCore::BitmapImage::checkForSolidColor):

2009-03-06  Darin Fisher  <darin@chromium.org>

        Reviewed by Sam Weinig.

        https://bugs.webkit.org/show_bug.cgi?id=24437

        Add support for registering noAccess URL schemes:
        1-  Add FrameLoader::registerURLSchemeAsNoAccess, and have SecurityOrigin check
        that list upon construction (similar to how isLocal is implemented).
        2-  Make InspectorController call grantUniversalAccess on its Document's
        SecurityOrigin at the time when windowScriptObjectAvailable is called.

        This enables content such as the inspector to be loaded from a custom (non-file)
        URL, which is how Chromium loads the inspector.  It also allows other URL schemes
        to be treated like data: URLs, which Chromium utilizes for its various HTML-based
        UI panels.

        * inspector/InspectorController.cpp:
        (WebCore::InspectorController::windowScriptObjectAvailable):
        * loader/FrameLoader.cpp:
        (WebCore::localSchemes):
        (WebCore::noAccessSchemes):
        (WebCore::FrameLoader::loadItem):
        * loader/FrameLoader.h:
        * page/SecurityOrigin.cpp:
        (WebCore::SecurityOrigin::SecurityOrigin):
        (WebCore::SecurityOrigin::isLocal):

2009-03-10  Xan Lopez  <xlopez@igalia.com>

        Build fix, no review

        * page/mac/AccessibilityObjectWrapper.mm:
        (textMarkerForVisiblePosition):
        (visiblePositionForTextMarker):

2009-03-10  Xan Lopez  <xlopez@igalia.com>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=24491
        Rename posOffset to m_offset, get rid of accessor

        As suggested in dom/Position.h, this patch gets rid of the
        'offset()' accessor and renames posOffset to m_offset. I've used
        m_offset instead of offset to follow the style guide lines, since
        Position is still a class and not a structure. If the long term
        plan is still to make it a structure it would be pretty easy to
        just s/m_offset/offset/ globally when that is achieved.

        * dom/Position.cpp:
        (WebCore::Position::previous):
        (WebCore::Position::next):
        (WebCore::Position::atStart):
        (WebCore::Position::atEnd):
        (WebCore::Position::renderedOffset):
        (WebCore::Position::isCandidate):
        (WebCore::Position::inRenderedText):
        (WebCore::Position::isRenderedCharacter):
        (WebCore::Position::rendersInDifferentPosition):
        (WebCore::Position::leadingWhitespacePosition):
        (WebCore::Position::debugPosition):
        (WebCore::Position::formatForDebugger):
        * dom/Position.h:
        (WebCore::Position::Position):
        * dom/PositionIterator.h:
        (WebCore::PositionIterator::PositionIterator):
        * dom/Range.cpp:
        * dom/RangeBoundaryPoint.h:
        (WebCore::RangeBoundaryPoint::position):
        (WebCore::RangeBoundaryPoint::set):
        (WebCore::RangeBoundaryPoint::setOffset):
        (WebCore::RangeBoundaryPoint::setToChild):
        (WebCore::RangeBoundaryPoint::setToStart):
        (WebCore::RangeBoundaryPoint::setToEnd):
        (WebCore::RangeBoundaryPoint::childBeforeWillBeRemoved):
        * editing/ApplyStyleCommand.cpp:
        (WebCore::ApplyStyleCommand::applyRelativeFontStyleChange):
        (WebCore::ApplyStyleCommand::applyInlineStyleToRange):
        (WebCore::ApplyStyleCommand::removeInlineStyle):
        (WebCore::ApplyStyleCommand::nodeFullySelected):
        (WebCore::ApplyStyleCommand::nodeFullyUnselected):
        (WebCore::ApplyStyleCommand::splitTextAtStartIfNeeded):
        (WebCore::ApplyStyleCommand::splitTextAtEndIfNeeded):
        (WebCore::ApplyStyleCommand::splitTextElementAtStartIfNeeded):
        (WebCore::ApplyStyleCommand::splitTextElementAtEndIfNeeded):
        (WebCore::ApplyStyleCommand::mergeStartWithPreviousIfIdentical):
        (WebCore::ApplyStyleCommand::mergeEndWithNextIfIdentical):
        (WebCore::ApplyStyleCommand::joinChildTextNodes):
        * editing/BreakBlockquoteCommand.cpp:
        (WebCore::BreakBlockquoteCommand::doApply):
        * editing/CompositeEditCommand.cpp:
        (WebCore::CompositeEditCommand::insertNodeAt):
        (WebCore::CompositeEditCommand::positionOutsideTabSpan):
        (WebCore::CompositeEditCommand::rebalanceWhitespaceAt):
        (WebCore::CompositeEditCommand::prepareWhitespaceAtPositionForSplit):
        (WebCore::CompositeEditCommand::deleteInsignificantText):
        (WebCore::CompositeEditCommand::removePlaceholderAt):
        (WebCore::CompositeEditCommand::moveParagraphs):
        (WebCore::CompositeEditCommand::breakOutOfEmptyMailBlockquotedParagraph):
        * editing/DeleteSelectionCommand.cpp:
        (WebCore::updatePositionForNodeRemoval):
        (WebCore::updatePositionForTextRemoval):
        (WebCore::DeleteSelectionCommand::handleGeneralDelete):
        (WebCore::DeleteSelectionCommand::fixupWhitespace):
        * editing/Editor.cpp:
        (WebCore::Editor::setComposition):
        (WebCore::Editor::advanceToNextMisspelling):
        (WebCore::Editor::getCompositionSelection):
        * editing/InsertLineBreakCommand.cpp:
        (WebCore::InsertLineBreakCommand::doApply):
        * editing/InsertParagraphSeparatorCommand.cpp:
        (WebCore::InsertParagraphSeparatorCommand::doApply):
        * editing/InsertTextCommand.cpp:
        (WebCore::InsertTextCommand::performTrivialReplace):
        (WebCore::InsertTextCommand::input):
        (WebCore::InsertTextCommand::insertTab):
        * editing/MoveSelectionCommand.cpp:
        (WebCore::MoveSelectionCommand::doApply):
        * editing/ReplaceSelectionCommand.cpp:
        (WebCore::ReplaceSelectionCommand::doApply):
        * editing/SelectionController.cpp:
        (WebCore::SelectionController::debugRenderer):
        * editing/TextIterator.cpp:
        (WebCore::TextIterator::rangeFromLocationAndLength):
        * editing/TypingCommand.cpp:
        (WebCore::TypingCommand::deleteKeyPressed):
        (WebCore::TypingCommand::forwardDeleteKeyPressed):
        * editing/VisiblePosition.cpp:
        (WebCore::VisiblePosition::characterAfter):
        (WebCore::VisiblePosition::debugPosition):
        (WebCore::makeRange):
        (WebCore::setStart):
        (WebCore::setEnd):
        * editing/VisibleSelection.cpp:
        (WebCore::VisibleSelection::toNormalizedRange):
        (WebCore::makeSearchRange):
        (WebCore::VisibleSelection::debugPosition):
        (WebCore::VisibleSelection::showTreeForThis):
        * editing/htmlediting.cpp:
        (WebCore::comparePositions):
        (WebCore::rangeCompliantEquivalent):
        (WebCore::isFirstPositionAfterTable):
        (WebCore::isLastPositionBeforeTable):
        * editing/visible_units.cpp:
        (WebCore::previousBoundary):
        (WebCore::nextBoundary):
        (WebCore::startPositionForLine):
        (WebCore::startOfLine):
        (WebCore::endPositionForLine):
        (WebCore::nextLinePosition):
        (WebCore::startOfParagraph):
        (WebCore::endOfParagraph):
        * page/AccessibilityObject.cpp:
        (WebCore::updateAXLineStartForVisiblePosition):
        * page/AccessibilityRenderObject.cpp:
        (WebCore::AccessibilityRenderObject::indexForVisiblePosition):
        * page/DOMSelection.cpp:
        (WebCore::DOMSelection::anchorOffset):
        (WebCore::DOMSelection::baseOffset):
        (WebCore::DOMSelection::focusOffset):
        (WebCore::DOMSelection::extentOffset):
        * page/EventHandler.cpp:
        (WebCore::EventHandler::handleMousePressEventSingleClick):
        * page/Frame.cpp:
        (WebCore::Frame::selectionLayoutChanged):
        * rendering/RenderTextControl.cpp:
        (WebCore::RenderTextControl::indexForVisiblePosition):
        * rendering/RenderTreeAsText.cpp:
        (WebCore::writeSelection):

2009-03-10  Darin Adler  <darin@apple.com>

        Reviewed by Dan Bernstein.

        Bug 23564: REGRESSION (r39230-39286): crash loading page that changes <input> display type and then calls innerHTML
        https://bugs.webkit.org/show_bug.cgi?id=23564
        rdar://problem/6537238

        Test: fast/dom/HTMLElement/innerHTML-selection-crash.html

        * editing/markup.cpp: (WebCore::createMarkup): Added updateLayoutIgnorePendingStylesheets
        call to the one of the two overloads of this function that wasn't calling it. This fixes
        this crash and other possible crashes inside innerHTML.

2009-03-10  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=24285
        Text resource loading checks for BOM twice

        This removes TextDecoder class, since its only purpose was to check for BOM, which is
        already done in TextResourceDecoder. Callers that use TextEncoding::decode() won't get
        BOM checked, but I didn't find any cases where it would significantly change behavior.

        * GNUmakefile.am:
        * WebCore.pro:
        * WebCore.scons:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:
        Remove TextDecoder.{h,cpp}.

        * platform/text/TextDecoder.cpp: Removed.
        * platform/text/TextDecoder.h: Removed.

        * loader/TextResourceDecoder.h: Instead of a TextDecoder object, keep TextEncoding and
        TextCodec separately.

        * loader/TextResourceDecoder.cpp:
        (WebCore::TextResourceDecoder::TextResourceDecoder): Updated for m_encoding being a member.
        (WebCore::TextResourceDecoder::setEncoding): Ditto.
        (WebCore::TextResourceDecoder::checkForBOM): Removed a FIXME saying that a BOM could override
        even a user-chosen encoding - this is how it already worked due to TextDecoder checking for
        BOM again. Made this function return the detected BOM length.
        (WebCore::TextResourceDecoder::decode): Skip the BOM if it was found at the beginning of
        a text resource.
        (WebCore::TextResourceDecoder::flush): Reset m_checkedForBOM, so that re-decoding the same
        resource again (as frequently done by CachedResource subclasses) will skip the BOM correctly.

        * platform/text/TextEncoding.cpp: (WebCore::TextEncoding::decode):
        Use TextCodec directly without a TextDecoder wrapper. This means that this method no longer
        checks for BOM, which was a counter-intuitive feature.

        * loader/CachedScript.cpp:
        (WebCore::CachedScript::CachedScript):
        (WebCore::CachedScript::setEncoding):
        (WebCore::CachedScript::encoding):
        (WebCore::CachedScript::script):
        * loader/CachedScript.h:
        * loader/appcache/ManifestParser.cpp:
        (WebCore::parseManifest):
        Use TextResourceDecoder, as TextEncoding::decode() no longer checks for BOM.
        A side effect of this is that these resources will now be subject to encoding auto-detection.

        * loader/CachedFont.cpp: (WebCore::CachedFont::ensureSVGFontData):
        * page/Page.cpp: (WebCore::Page::userStyleSheet):
        Be sure to flush TextResourceDecoder, pushing any remaining bytes out, and making the decoder
        re-usable (for repeated decoding of the same resource).

        * platform/text/TextEncodingRegistry.h: Updated comments for newTextCodec().

2009-03-10  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=24488
        An assertion failure when updating an offline application cache after emptying caches

        Emptying appcache cannot be performed automatically, so no test.

        * loader/appcache/ApplicationCacheGroup.cpp:
        (WebCore::ApplicationCacheGroup::checkIfLoadIsComplete): Ensure that the cache is in disk
        storage, even if the application wasn't updated server side.

2009-03-10  Oliver Hunt  <oliver@apple.com>

        Reviewed by Alexey Proskuryakov.

        <rdar://problem/6425807> Implement WorkerUtils.importScripts()
        <https://bugs.webkit.org/show_bug.cgi?id=22721>

        Implement importScripts, currently uses a series of synchronous loads
        to fetch the scripts, but this is simpler than a synchronous load of
        multiple loads in parallel.  In future we'll want to switch to parallel
        loading, but this will do for now.
        
        Test: http/tests/workers/worker-importScripts.html

        * GNUmakefile.am:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * bindings/js/JSWorkerContextCustom.cpp:
        (WebCore::JSWorkerContext::importScripts):
        * bindings/js/ScriptValue.h:
        (WebCore::ScriptValue::hasNoValue):
        * bindings/js/WorkerScriptController.cpp:
        (WebCore::WorkerScriptController::evaluate):
        (WebCore::WorkerScriptController::setException):
        * bindings/js/WorkerScriptController.h:
        * dom/Document.cpp:
        (WebCore::Document::scriptImported):
        * dom/Document.h:
        * dom/ScriptExecutionContext.h:
        * inspector/InspectorController.cpp:
        (WebCore::InspectorController::scriptImported):
        * inspector/InspectorController.h:
        * inspector/InspectorResource.cpp:
        (WebCore::InspectorResource::setScriptProperties):
        * inspector/InspectorResource.h:
        * workers/WorkerContext.cpp:
        (WebCore::WorkerContext::scriptImported):
        (WebCore::WorkerContext::importScripts):
        * workers/WorkerContext.h:
        * workers/WorkerContext.idl:
        * workers/WorkerImportScriptsClient.cpp: Added.
        (WebCore::WorkerImportScriptsClient::didReceiveResponse):
        (WebCore::WorkerImportScriptsClient::didReceiveData):
        (WebCore::WorkerImportScriptsClient::didFinishLoading):
        (WebCore::WorkerImportScriptsClient::didFail):
        (WebCore::WorkerImportScriptsClient::didFailRedirectCheck):
        (WebCore::WorkerImportScriptsClient::didReceiveAuthenticationCancellation):
        * workers/WorkerImportScriptsClient.h: Added.
        (WebCore::WorkerImportScriptsClient::WorkerImportScriptsClient):
        (WebCore::WorkerImportScriptsClient::script):
        (WebCore::WorkerImportScriptsClient::failed):

2009-03-10  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=24462
        Move cross-origin access control code out of XMLHttpRequest

        Step 1: move preflight result cache and access control helper functions.

        No change in behavior, so no test.

        * GNUmakefile.am:
        * WebCore.pro:
        * WebCore.scons:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:
        * loader/CrossOriginAccessControl.cpp: Added.
        (WebCore::isOnAccessControlSimpleRequestHeaderWhitelist):
        (WebCore::isSimpleCrossOriginAccessRequest):
        (WebCore::createAllowedCrossOriginResponseHeadersSet):
        (WebCore::isOnAccessControlResponseHeaderWhitelist):
        (WebCore::passesAccessControlCheck):
        * loader/CrossOriginAccessControl.h: Added.
        * loader/CrossOriginPreflightResultCache.cpp: Added.
        (WebCore::parseAccessControlMaxAge):
        (WebCore::addToAccessControlAllowList):
        (WebCore::parseAccessControlAllowList):
        (WebCore::CrossOriginPreflightResultCacheItem::parse):
        (WebCore::CrossOriginPreflightResultCacheItem::allowsCrossOriginMethod):
        (WebCore::CrossOriginPreflightResultCacheItem::allowsCrossOriginHeaders):
        (WebCore::CrossOriginPreflightResultCacheItem::allowsRequest):
        (WebCore::CrossOriginPreflightResultCache::shared):
        (WebCore::CrossOriginPreflightResultCache::appendEntry):
        (WebCore::CrossOriginPreflightResultCache::canSkipPreflight):
        * loader/CrossOriginPreflightResultCache.h: Added.
        (WebCore::CrossOriginPreflightResultCacheItem::CrossOriginPreflightResultCacheItem):
        (WebCore::CrossOriginPreflightResultCache::CrossOriginPreflightResultCache):
        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequestStaticData::XMLHttpRequestStaticData):
        (WebCore::XMLHttpRequest::createRequest):
        (WebCore::XMLHttpRequest::makeCrossOriginAccessRequest):
        (WebCore::XMLHttpRequest::makeSimpleCrossOriginAccessRequest):
        (WebCore::XMLHttpRequest::makeCrossOriginAccessRequestWithPreflight):
        (WebCore::XMLHttpRequest::didReceiveResponse):
        (WebCore::XMLHttpRequest::didReceiveResponsePreflight):
        * xml/XMLHttpRequest.h:

2009-03-10  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=19797
        Bring the list of forbidden headers in sync with XMLHttpRequest spec draft

        Added new checks to http/tests/xmlhttprequest/set-dangerous-headers.html.

        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequestStaticData::XMLHttpRequestStaticData): Added all headers from
        XMLHttpRequest 2 draft.

2009-03-10  Dan Bernstein  <mitz@apple.com>

        Reviewed by Adele Peterson.

        - fix https://bugs.webkit.org/show_bug.cgi?id=23975
          <rdar://problem/6593610> REGRESSION: Scrollbar arrow not repainted

        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::invalidateScrollbarRect): Corrected the vertical
        scrollbar rect calculation.

2009-03-09  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dan Bernstein, Oliver Hunt

        https://bugs.webkit.org/show_bug.cgi?id=24475
        
        When repainting replaced elements, need to union the selection
        and overflow rects because either may extend outside the other.
        
        Test: fast/repaint/transform-replaced-shadows.html

        * rendering/RenderReplaced.cpp:
        (WebCore::RenderReplaced::clippedOverflowRectForRepaint):

2009-03-09  Stephanie Lewis  <slewis@apple.com>

        Fix build.

        * editing/Editor.cpp:
        (WebCore::Editor::advanceToNextMisspelling):
        (WebCore::markAllMisspellingsAndBadGrammarInRanges):

2009-03-09  Kim Christensen  <kimworking@gmail.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=24276
        Dropdown / select boxes do not consider the windows task bar in Chromium.

        * platform/chromium/PopupMenuChromium.cpp:
        (WebCore::PopupContainer::showPopup):

2009-03-09  Dan Bernstein  <mitz@apple.com>

        Reviewed by Adele Peterson and Dave Hyatt.

        - fix <rdar://problem/6654344> REGRESSION (r41041): Zillow.com has
          misaligned edit fields

        Test: fast/forms/legend-display-none.html

        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::adjustRenderStyle): Change the display
        property of legend elements to 'block' unless it is 'none'.
        * css/html4.css: Do not force the display property of legend elements to
        'block' using CSS, because that prevents style rules from setting it to
        'none'.

2009-03-09  Alpha Lam  <hclam@chromium.org>

        Reviewed by Dave Hyatt.

        https://bugs.webkit.org/show_bug.cgi?id=23024

        Fix an incorrect criteria in the if statement that leaves references to
        deleted floating objects in RenderBlock.

        Test: fast/block/float/crash-on-absolute-positioning.html

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::markAllDescendantsWithFloatsForLayout):

2009-03-09  Mike Belshe  <mike@belshe.com>

        Reviewed by Eric Seidel.

        Unify JSC/V8 idl file.
        https://bugs.webkit.org/show_bug.cgi?id=24424

        * page/Location.idl:

2009-03-09  Mike Belshe  <mike@belshe.com>

        Reviewed by Eric Seidel.

        Unify JSC/V8 for History.idl.

        * page/History.idl:

2009-03-09  Mike Belshe  <mike@belshe.com>

        Reviewed by Eric Seidel.

        Unify JSC/V8 implementations of DOMWindow.idl.

        * page/DOMWindow.idl:

2009-03-09  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dave Hyatt

        https://bugs.webkit.org/show_bug.cgi?id=24163
        
        Implement mapping of FloatQuads through 3D transforms.
        
        Removed the redundant localToContainerQuad() methods, which can now 
        shared code with the old mapLocalToAbsolutePoint(), which was
        renamed to mapLocalToContainer(). This can now convert a point,
        and optionally a FloatQuad, which are carried along in the TransformState.
        
        Optimized TransformState to reduce to simple FloatPoint.move()
        if there are no transforms, and to heap-allocate a transform only if
        necessary to accumulate transforms (when using preserve-3d).
        
        Tested by 3d point mapping tests, and the inspector highlight (which now shows
        the correct quads for 3d-transformed elements).

        * platform/graphics/transforms/TransformationMatrix.cpp:
        (WebCore::TransformationMatrix::projectQuad):
        (WebCore::TransformationMatrix::translateRight):
        (WebCore::TransformationMatrix::translateRight3d):
        * platform/graphics/transforms/TransformationMatrix.h:
        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::mapLocalToContainer):
        (WebCore::RenderBox::mapAbsoluteToLocalPoint):
        * rendering/RenderBox.h:
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::localToAbsolute):
        (WebCore::RenderObject::absoluteToLocal):
        (WebCore::RenderObject::mapLocalToContainer):
        (WebCore::RenderObject::localToContainerQuad):
        * rendering/RenderObject.h:
        * rendering/RenderTableCell.cpp:
        (WebCore::RenderTableCell::mapLocalToContainer):
        * rendering/RenderTableCell.h:
        * rendering/RenderView.cpp:
        (WebCore::RenderView::mapLocalToContainer):
        * rendering/RenderView.h:
        * rendering/TransformState.cpp:
        (WebCore::TransformState::move):
        (WebCore::TransformState::applyTransform):
        (WebCore::TransformState::flatten):
        (WebCore::TransformState::mappedPoint):
        (WebCore::TransformState::mappedQuad):
        (WebCore::TransformState::flattenWithTransform):
        (WebCore::HitTestingTransformState::mappedQuad):
        * rendering/TransformState.h:
        (WebCore::TransformState::TransformState):
        (WebCore::TransformState::move):
        (WebCore::TransformState::lastPlanarPoint):
        (WebCore::TransformState::lastPlanarQuad):
        * rendering/style/RenderStyle.h:
        (WebCore::InheritedFlags::preserves3D):

2009-03-09  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dave Hyatt

        https://bugs.webkit.org/show_bug.cgi?id=24433
        
        During hit testing with transforms, if a layer has backface-visibility: hidden,
        we need to check for it whether or not the layer has a transform, because it
        might be an ancestor that is transformed.

        Test: transforms/3d/hit-testing/backface-no-transform-hit-test.html

        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::hitTestLayer):

2009-03-09  Laszlo Gombos  <laszlo.1.gombos@nokia.com>

        Reviewed by George Staikos.

        https://bugs.webkit.org/show_bug.cgi?id=24353
        Allow to overrule default build options for Qt build.

        * WebCore.pro: Allow to overrule ENABLE_VIDEO and ENABLE_NETSCAPE_PLUGIN_API

2009-03-09  Adam Treat  <adam.treat@torchmobile.com>

        Reviewed by George Staikos.

        https://bugs.webkit.org/show_bug.cgi?id=24463
        WebCore::qstring is detaching and copying twice for every single
        WebCore::TextRun that is processed and drawn.  This elevates this method
        to one of the top-ten most expensive methods in all of QtWebKit according
        to profiling.  This changes the method so that QString only detaches
        when absolutely necessary.

        * platform/graphics/qt/FontQt.cpp:
        (WebCore::qstring):
        (WebCore::fixSpacing):
        (WebCore::Font::drawComplexText):
        (WebCore::Font::floatWidthForComplexText):
        (WebCore::Font::offsetForPositionForComplexText):
        (WebCore::Font::selectionRectForComplexText):

2009-03-08  Mark Rowe  <mrowe@apple.com>

        Reviewed by Oliver Hunt.

        Split ScrollAlignment and ScrollBehavior out of RenderLayer.h so that
        Frame.h no longer needs to include it.  This cuts the size of the symbols
        for a debug build by around 3%.

        * dom/Element.cpp:
        (WebCore::Element::scrollIntoView):
        (WebCore::Element::scrollIntoViewIfNeeded):
        * editing/Editor.cpp:
        (WebCore::Editor::insertTextWithoutSendingTextEvent):
        (WebCore::Editor::revealSelectionAfterEditingOperation):
        * editing/SelectionController.cpp:
        (WebCore::SelectionController::setSelection):
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::gotoAnchor):
        * page/ContextMenuController.cpp:
        (WebCore::ContextMenuController::contextMenuItemSelected):
        * page/Frame.cpp:
        (WebCore::Frame::revealSelection):
        (WebCore::Frame::revealCaret):
        * page/Frame.h:
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::getRectToExpose):
        (WebCore::RenderLayer::autoscroll):
        * rendering/RenderLayer.h:
        * rendering/ScrollBehavior.cpp: Added.
        (WebCore::):
        * rendering/ScrollBehavior.h: Added.
        (WebCore::):
        (WebCore::ScrollAlignment::getVisibleBehavior):
        (WebCore::ScrollAlignment::getPartialBehavior):
        (WebCore::ScrollAlignment::getHiddenBehavior):

        Add the new files to the build systems:
        * GNUmakefile.am:
        * WebCore.base.exp:
        * WebCore.pro:
        * WebCore.scons:
        * WebCore.vcproj/WebCore.vcproj:
        * WebCore.xcodeproj/project.pbxproj:
        * WebCoreSources.bkl:

        Add includes that were previously pulled in by way of RenderLayer.h from Frame.h:
        * dom/Clipboard.cpp:
        * editing/EditorCommand.cpp:
        * editing/InsertLineBreakCommand.cpp:
        * editing/TypingCommand.cpp:
        * html/HTMLFormControlElement.cpp:
        * loader/archive/cf/LegacyWebArchive.cpp:
        * loader/icon/IconLoader.cpp:
        * page/animation/AnimationController.cpp:
        * rendering/RenderSlider.cpp:
        * rendering/RenderTextControl.cpp:


2009-03-08  Mark Rowe  <mrowe@apple.com>

        Reviewed Dan Bernstein.

        Push the include of Frame.h out of RenderView.h and down to those files
        that need it.  This cuts the size of symbols for RenderFoo object files
        by around 15% in a debug build.

        * bindings/objc/DOM.mm:
        * rendering/RenderFileUploadControl.cpp:
        * rendering/RenderInline.cpp:
        * rendering/RenderObject.cpp:
        * rendering/RenderReplaced.cpp:
        * rendering/RenderText.cpp:
        * rendering/RenderView.cpp:
        (WebCore::RenderView::zoomFactor):
        * rendering/RenderView.h:

2009-03-08  Mark Rowe  <mrowe@apple.com>

        Rubber-stamped by Dan Bernstein.

        Remove some unnecessary or redundant includes from files related to the JS bindings.
        This cuts the size of the symbols for a debug build by around 3%.

        * bindings/js/JSCustomPositionCallback.cpp:
        * bindings/js/JSCustomPositionErrorCallback.cpp:
        * bindings/js/JSCustomSQLStatementCallback.cpp:
        * bindings/js/JSCustomSQLStatementErrorCallback.cpp:
        * bindings/js/JSCustomSQLTransactionCallback.cpp:
        * bindings/js/JSCustomSQLTransactionErrorCallback.cpp:
        * bindings/js/JSCustomVoidCallback.cpp:
        * bindings/js/JSCustomXPathNSResolver.cpp:
        * bindings/js/JSDocumentCustom.cpp:
        * bindings/js/JSPluginElementFunctions.cpp:
        * bindings/js/JSSVGPODTypeWrapper.h:

2009-03-08  Cameron Zwarich  <cwzwarich@uwaterloo.ca>

        Reviewed by Oliver Hunt.

        Remove the unused methods previousHistoryItem(), setPreviousHistoryItem()
        and provisionalHistoryItem() from FrameLoader.

        * WebCore.base.exp:
        * WebCore.order:
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::frameLoadCompleted):
        (WebCore::FrameLoader::loadItem):
        * loader/FrameLoader.h:

2009-03-08  Gustavo Noronha Silva  <gns@gnome.org>

        Reviewed by Holger Freyther.

        Implement proper logging for the GTK+ port, inspired by the one in
        Qt. We use the WEBKIT_DEBUG environment variable which was already
        being used to decide whether to setup a logger for soup or not.

        * platform/gtk/LoggingGtk.cpp:
        (WebCore::getChannelFromName):
        (WebCore::InitializeLoggingChannelsIfNecessary):
        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::ensureSessionIsInitialized):

2009-03-07  Dan Bernstein  <mitz@apple.com>

        Reviewed by Mark Rowe.

        - WebCore part of removing build-time and run-time support for legacy
          versions of CFNetwork and Core Graphics

        * platform/graphics/win/FontCGWin.cpp:
        (WebCore::Font::drawGlyphs):
        * platform/graphics/win/FontCustomPlatformData.cpp:
        (WebCore::FontCustomPlatformData::fontPlatformData):
        * platform/graphics/win/FontPlatformDataCGWin.cpp:
        (WebCore::FontPlatformData::platformDataInit):
        * platform/network/cf/FormDataStreamCFNet.cpp:
        (WebCore::setHTTPBody):
        (WebCore::httpBodyFromRequest):
        * platform/network/cf/ResourceHandleCFNet.cpp:
        (WebCore::didSendBodyData):
        (WebCore::shouldUseCredentialStorageCallback):
        (WebCore::ResourceHandle::start):
        * platform/network/win/CookieJarCFNetWin.cpp:
        (WebCore::filterCookies):

2009-03-07  Adele Peterson  <adele@apple.com>

        Reviewed by Dan Bernstein.

        Fix for https://bugs.webkit.org/show_bug.cgi?id=24444
        No multiple JS property for <input type="file" multiple />

        * html/HTMLInputElement.cpp:
        (WebCore::HTMLInputElement::multiple):
        (WebCore::HTMLInputElement::setMultiple):
        * html/HTMLInputElement.h:
        * html/HTMLInputElement.idl:

2009-03-07  Xan Lopez  <xan@gnome.org>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24358
        [GTK] Scrollbars not clipped correctly

        ScrollView scrollbars in subframes are *not* native, so take that
        into account again.

        * platform/gtk/ScrollbarGtk.cpp:
        (ScrollbarGtk::getLocationInParentWindow):
        (ScrollbarGtk::frameRectsChanged):
        (ScrollbarGtk::paint):
        * platform/gtk/ScrollbarGtk.h:

2009-03-06  Peter Kasting  <pkasting@google.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=24405
        Horizontal scrolling on Windows was reversed from Mac (and intuition).

        * platform/gtk/WheelEventGtk.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/win/WheelEventWin.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):

2009-03-06  Peter Kasting  <pkasting@google.com>

        Reviewed by Sam Weinig.

        Convert some C-style casts to static_cast<>()s.

        * platform/gtk/WheelEventGtk.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/mac/WheelEventMac.mm:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/win/WheelEventWin.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/wx/MouseWheelEventWx.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):

2009-03-06  Adele Peterson  <adele@apple.com>

        Reviewed by Oliver Hunt.

        Fix for <rdar://problem/6607524> REGRESSION (Safari 3-4): I can't tab back to the URL field in an empty window (key loop is broken)

        I haven't been able to make a test for this since the problem is not reproducible within an empty iframe.

        * page/EventHandler.cpp: (WebCore::eventTargetNodeForDocument): We used to ensure that every html document had a body element.
          That is no longer true, so we should return the document element for a truly empty document.

2009-03-06  Jay Campan  <jcampan@google.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=24306

        Adding a flag to ResourceRequestBase to indicate whether or not upload
        progress notifications are needed for a resource. This is useful to
        avoid sending these notifications when there are no consumers
        (especially in the Chromium case where IPC is involved). 

        * platform/network/ResourceRequestBase.h:
        (WebCore::ResourceRequestBase::reportUploadProgress):
        (WebCore::ResourceRequestBase::setReportUploadProgress):
        (WebCore::ResourceRequestBase::ResourceRequestBase):
        * xml/XMLHttpRequest.cpp:
        (WebCore::XMLHttpRequest::loadRequestAsynchronously):

2009-03-06  Dmitry Titov  <dimich@chromium.org>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=24150
        Add virtual ScriptExecutionContext::encoding()

        Test: http/tests/workers/text-encoding.html

        * dom/Document.cpp:
        (WebCore::Document::encoding):
        * dom/Document.h:
        (WebCore::Document::inputEncoding):
        (WebCore::Document::charset):
        (WebCore::Document::characterSet):
        Add new virtual method and route DOM synonym functions on Document through it.

        * dom/ScriptExecutionContext.h:
        * workers/Worker.cpp:
        (WebCore::Worker::Worker):
        (WebCore::Worker::notifyFinished):
        * workers/WorkerContext.cpp:
        (WebCore::WorkerContext::WorkerContext):
        (WebCore::WorkerContext::encoding):
        (WebCore::WorkerContext::completeURL): Added comment on why this is different from Document::completeURL
        * workers/WorkerContext.h:
        (WebCore::WorkerContext::create):
        * workers/WorkerContextProxy.h:
        * workers/WorkerMessagingProxy.cpp:
        (WebCore::WorkerMessagingProxy::startWorkerContext):
        * workers/WorkerMessagingProxy.h:
        * workers/WorkerThread.cpp:
        (WebCore::WorkerThreadStartupData::create):
        (WebCore::WorkerThreadStartupData::WorkerThreadStartupData):
        (WebCore::WorkerThread::create):
        (WebCore::WorkerThread::WorkerThread):
        (WebCore::WorkerThread::workerThread):
        * workers/WorkerThread.h:
        All of the above route the 'encoding' parameter of parent context to the new
        instance of WorkerContext - from Worker::notifyFinished() via WorkerMessagingProxy
        through WorkerThread through WorkerThreadStartupData and into constructor of WorkerContext.

2009-03-06  Peter Kasting  <pkasting@google.com>

        Build bustage fix.

        * dom/Node.cpp:
        (WebCore::Node::dispatchWheelEvent):

2009-03-06  Eric Carlson  <eric.carlson@apple.com>

        Reviewed by Antti Koivisto.

        * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
        (WebCore::MediaPlayerPrivate::createQTMovieView):  Delay callback while setting up movieview.

2009-03-06  Douglas R. Davidson  <ddavidso@apple.com>

        Reviewed by Justin Garcia.
        
        https://bugs.webkit.org/show_bug.cgi?id=24108

        Update spelling and grammar checking to use the new combined text 
        checking (with automatic language identification) on Snow Leopard. 
        Tested manually in Mail and Safari; automated tests to come later. 
        
        * editing/Editor.cpp:
        (WebCore::findFirstMisspellingOrBadGrammarInRange):
        (WebCore::Editor::advanceToNextMisspelling):
        (WebCore::guessesForMisspelledOrUngrammaticalRange):
        (WebCore::Editor::guessesForMisspelledOrUngrammaticalSelection):
        (WebCore::Editor::markMisspellingsAfterTypingToPosition):
        (WebCore::markAllMisspellingsAndBadGrammarInRanges):
        (WebCore::Editor::markMisspellingsAndBadGrammar):
        * editing/Editor.h:
        * loader/EmptyClients.h:
        (WebCore::EmptyEditorClient::checkSpellingAndGrammarOfParagraph):
        * page/EditorClient.h:
        * page/Frame.cpp:
        (WebCore::Frame::respondToChangedSelection):
        * platform/ContextMenu.cpp:
        (WebCore::ContextMenu::populate):

2009-03-06  Peter Kasting  <pkasting@google.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=24407
        Windows scroll amount was too small, and wheel scroll distance
        conversion code was overly complex.

        * page/EventHandler.cpp:
        (WebCore::scrollAndAcceptEvent):
        * platform/PlatformWheelEvent.h:
        (WebCore::):
        * platform/ScrollView.cpp:
        (WebCore::ScrollView::wheelEvent):
        * platform/Scrollbar.h:
        * platform/gtk/WheelEventGtk.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/mac/WheelEventMac.mm:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/qt/WheelEventQt.cpp:
        * platform/win/WheelEventWin.cpp:
        (WebCore::horizontalScrollChars):
        (WebCore::verticalScrollLines):
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):
        * platform/wx/MouseWheelEventWx.cpp:
        (WebCore::PlatformWheelEvent::PlatformWheelEvent):

2009-03-06  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Adam Roben.

        Generate valid bindings with HTML5 database support disabled. 

        * inspector/InspectorController.idl: Added feature guard around 
        databaseTableNames definition.

2009-03-04  Xan Lopez  <xan@gnome.org>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24358
        [GTK] Scrollbars not clipped correctly

        Do not take into account the case of being a ScrollView scrollbar,
        since those are native in our case.

        * platform/gtk/ScrollbarGtk.cpp:
        (ScrollbarGtk::frameRectsChanged):

2009-03-04  Xan Lopez  <xan@gnome.org>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24358
        [GTK] Scrollbars not clipped correctly

        Move Widget::paint to ScrollbarGtk::paint, since it's scrollbar
        specific and it's our only Widget anyway.

        * platform/gtk/ScrollbarGtk.cpp:
        (ScrollbarGtk::paint):
        * platform/gtk/ScrollbarGtk.h:
        * platform/gtk/WidgetGtk.cpp:

2009-03-04  Xan Lopez  <xan@gnome.org>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24358
        [GTK] Scrollbars not clipped correctly

        Use correct clip rectangle and apply coordinate translation needed
        for non-ScrollView scrollbars.

        We were ignoring the clip rectangle passed as parameter, which is
        wrong in the case of non coalesced expose events. This, in turn,
        uncovers the fact that we were not applying coordinate translation
        to our position.

        * platform/gtk/WidgetGtk.cpp:
        (WebCore::Widget::paint):

2009-03-06  Eric Carlson  <eric.carlson@apple.com>

        Build fix, no review

        * platform/graphics/chromium/MediaPlayerPrivateChromium.h: fix setSize declaration

2009-03-06  Eric Carlson  <eric.carlson@apple.com>

        Reviewed by Simon Fraser.

        https://bugs.webkit.org/show_bug.cgi?id=22790
        Bug 22790: [Transforms] MediaPlayer::setRect() makes no sense with transforms
        Replace media engine setRect with setSize since they don't use about the 
        position anyway.

        * platform/graphics/MediaPlayer.cpp:
        (WebCore::NullMediaPlayerPrivate::setSize): Changed from setRect.
        (WebCore::MediaPlayer::setSize): Ditto.
        * platform/graphics/MediaPlayer.h:
        (WebCore::MediaPlayer::size): Changed from rect().

        * platform/graphics/MediaPlayerPrivate.h: Changed setRect to setSize.

        * platform/graphics/chromium/MediaPlayerPrivateChromium.h: Ditto.

        * platform/graphics/gtk/MediaPlayerPrivateGStreamer.cpp: 
        (WebCore::MediaPlayerPrivate::MediaPlayerPrivate): Initialize m_size instead of m_rect
        (WebCore::MediaPlayerPrivate::setSize): Changed from setRect
        (WebCore::MediaPlayerPrivate::paint): update comment
        * platform/graphics/gtk/MediaPlayerPrivateGStreamer.h: m_rect -> m_size.

        * platform/graphics/mac/MediaPlayerPrivateQTKit.h:  m_rect
        * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
        (WebCore::MediaPlayerPrivate::MediaPlayerPrivate): Initialize m_rect.
        (WebCore::MediaPlayerPrivate::createQTMovieView): setRect-> setSize.
        (WebCore::MediaPlayerPrivate::setSize): Changed from setRect
        (WebCore::MediaPlayerPrivate::paint): Call view:setFrame: when in a media document so 
        the movie is drawn in the correct location.

        * platform/graphics/qt/MediaPlayerPrivatePhonon.cpp:
        (WebCore::MediaPlayerPrivate::setSize):  Changed from setRect
        * platform/graphics/qt/MediaPlayerPrivatePhonon.h:

        * platform/graphics/win/MediaPlayerPrivateQuickTimeWin.cpp:
        (WebCore::MediaPlayerPrivate::setSize):  Changed from setRect
        * platform/graphics/win/MediaPlayerPrivateQuickTimeWin.h:

        * rendering/RenderVideo.cpp:
        (WebCore::RenderVideo::updatePlayer): Call setSize instead of setRect.

2009-03-06  Darin Adler  <darin@apple.com>

        Reviewed by Darin Fisher.

        Bug 24422: REGRESSION: null-URL crash in FrameLoader setting location.hash on new window
        https://bugs.webkit.org/show_bug.cgi?id=24422
        rdar://problem/6402208

        Test: fast/dom/location-new-window-no-crash.html

        The issue here is empty (or null) URLs. I picked the "schedule navigation" bottleneck
        to add some checks for empty URLs. We could also put the empty URL checks at some
        other bottleneck level and add more assertions over time. I tried adding a few more
        assertions to functions like loadURL and hit them while running the regression tests,
        so it's probably going to be a bit tricky to clean this up throughout the loader.

        * loader/FrameLoader.cpp:
        (WebCore::ScheduledRedirection::ScheduledRedirection): Explicitly marked this struct
        immutable by making all its members const. Added assertions about the arguments,
        including that the URL is not empty. Initialized one uninitialized member in one of
        the constructors.
        (WebCore::FrameLoader::scheduleHTTPRedirection): Added an early exit to make this
        a no-op if passed an empty URL.
        (WebCore::FrameLoader::scheduleLocationChange): Ditto.
        (WebCore::FrameLoader::scheduleRefresh): Ditto.

2009-03-06  Gustavo Noronha Silva  <gns@gnome.org>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24423
        Use new soup_message_body_set_accumulate API in soup backend

        Disable accumulating chunks for request_body on file uploads,
        using the new soup API.

        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::ResourceHandle::startHttp):

2009-03-06  Gustavo Noronha Silva  <gns@gnome.org>

        Reviewed by Holger Freyther.

        Replace use of deprecated SOUP_MESSAGE_OVERWRITE_CHUNKS flag with
        the new soup_message_body_set_accumulate API in soup.

        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::gotHeadersCallback):
        (WebCore::ResourceHandle::startHttp):

2009-03-06  Gustavo Noronha Silva  <gns@gnome.org>

        Reviewed by Holger Freyther.

        https://bugs.webkit.org/show_bug.cgi?id=24051
        Soup backend needs content sniffing capabilities

        Perform content sniffing when using soup, so that we have a chance
        of figuring out the Content-Type of the file if it's not sent by
        the server.

        * platform/network/ResourceHandleInternal.h:
        (WebCore::ResourceHandleInternal::ResourceHandleInternal):
        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::gotHeadersCallback):
        (WebCore::gotChunkCallback):

2009-03-06  Hironori Bono  <hbono@chromium.org>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=24342
        Cannot insert a Thai character after a Thai prepend character when using ICU 4.0

        This change creates a new break iterator "cursorMovementIterator" for
        moving cursors and use it when moving an input cursor.
        In "TextBreakIteratorICU.cpp", this break iterator uses custom ruleset
        based on the one of ICU 3.8.
        On the other hand, in "TextBreakIteratorQt.cpp", this break iterator
        just calls the characterBreakIterator() function.

        Test: editing/inserting/insert-thai-characters-001.html

        * platform/text/TextBreakIterator.h: Added a new function cursorMovementIterator().
        * platform/text/TextBreakIteratorICU.cpp: Implemented the cursorMovementIterator() function for ICU.
        (WebCore::setUpIteratorWithRules): Ditto.
        (WebCore::cursorMovementIterator): Ditto.
        * platform/text/qt/TextBreakIteratorQt.cpp: Implemented the cursorMovementIterator() function for Qt.
        (WebCore::cursorMovementIterator): Ditto.
        * rendering/RenderText.cpp: Call the cursorMovementIterator() function when moving an input cursor.
        (WebCore::RenderText::previousOffset): Ditto.
        (WebCore::RenderText::nextOffset): Ditto.

2009-03-05  Alexey Proskuryakov  <ap@webkit.org>

        Reviewed by Oliver Hunt.

        <rdar://problem/6621701> Safari 4 Beta Breaks XMLHttpRequest Response Text With Special
        Characters (a compatibility issue with widgets).

        Test: http/tests/xmlhttprequest/broken-xml-encoding.html

        Revert part of an Acid 3 fix - now we are no longer strict when decoding XMLHttpRequest XML
        responses.

        * loader/TextResourceDecoder.cpp:
        (WebCore::TextResourceDecoder::TextResourceDecoder):
        (WebCore::TextResourceDecoder::decode):
        (WebCore::TextResourceDecoder::flush):
        * loader/TextResourceDecoder.h:
        (WebCore::TextResourceDecoder::useLenientXMLDecoding):
        Don't stop on XML decoding errors if useLenientXMLDecoding() was called.

        * xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::didReceiveData): Don't stop on XML
        decoding errors. This behavior is now limited to other kinds of XML content.

2009-03-05  Simone Fiorentino  <simone.fiorentino@consulenti.fastweb.it>

        Bug 24382: request to add SH4 platform

        <https://bugs.webkit.org/show_bug.cgi?id=24382>

        Reviewed by David Kilzer.

        * platform/text/AtomicString.cpp:
        (WebCore::equal): Aligned memory access on SH4 platform.

2009-03-05  Jeremy Moskovich  <jeremy@chromium.org>

        Reviewed by Simon Fraser.

        Fix for https://bugs.webkit.org/show_bug.cgi?id=24215

        Gears expects an object tag with display:none to instantiate the plugin,
        so we add a workaround to make this work and fix Gears on WebKit trunk.

        * html/HTMLObjectElement.cpp:
        (WebCore::HTMLObjectElement::rendererIsNeeded):

2009-03-05  Avi Drissman  <avi@chromium.org>

        Reviewed by Darin Fisher.

        Need to have Chromium Mac match Safari Mac's accesskey handling
        https://bugs.webkit.org/show_bug.cgi?id=24404

        * page/chromium/EventHandlerChromium.cpp:
        (WebCore::EventHandler::accessKeyModifiers): Share access key modifiers with Mac Safari when building for the Mac.

2009-03-05  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dave Hyatt

        https://bugs.webkit.org/show_bug.cgi?id=24412
        
        Fix crash when hit-testing elements with -webkit-transform-style: preserve-3d
        but no transform. We need to make localTransformState if we see preserve-3d.
        Also need to call update3DTransformedDescendantStatus() before we test
        m_has3DTransformedDescendant.

        Test: transforms/3d/hit-testing/hit-preserves-3d.html

        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::hitTestLayer):

2009-03-05  Eric Seidel  <eric@webkit.org>

        Reviewed by David Hyatt.

        Changes to RenderLayer destruction to hopefully help catch an elusive crasher
        https://bugs.webkit.org/show_bug.cgi?id=24409
        
        Added a new RenderBoxModelObject::destroyLayer() call which is
        now the only way which RenderLayers should ever be destroyed.
        This ensures that the pointer to the layer is cleared in the
        RenderObject after destruction, allowing us to ASSERT in the
        RenderBoxModelObject destructor.

        * rendering/RenderBox.cpp:
        (WebCore::RenderBox::calcAbsoluteHorizontalReplaced):
        * rendering/RenderBoxModelObject.cpp:
        (WebCore::RenderBoxModelObject::~RenderBoxModelObject):
        (WebCore::RenderBoxModelObject::destroyLayer):
        (WebCore::RenderBoxModelObject::destroy):
        (WebCore::RenderBoxModelObject::styleDidChange):
        * rendering/RenderBoxModelObject.h:
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::stackingContext):
        (WebCore::RenderLayer::destroy):
        (WebCore::RenderLayer::removeOnlyThisLayer):
        * rendering/RenderLayer.h:
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::destroy):
        * rendering/RenderWidget.cpp:
        (WebCore::RenderWidget::destroy):

2009-03-05  Eric Seidel  <eric@webkit.org>

        Reviewed by David Hyatt.

        Remove old, unused IE 5.5 scrollbar-* CSS properties.
        Sort the unimplemented getComputedStyle properties so it's
        easier to see which ones actually need implementation.

        * css/CSSComputedStyleDeclaration.cpp:
        (WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):
        * css/CSSParser.cpp:
        (WebCore::CSSParser::parseValue):
        * css/CSSPropertyNames.in:
        * css/CSSStyleSelector.cpp:
        (WebCore::CSSStyleSelector::applyProperty):

2009-03-05  Justin Garcia  <justin.garcia@apple.com>

        Reviewed by Darin Adler.
        
        WebViewDidChangeSelectionNotifications weren't being sent for commands that change the selection's position
        within the document without changing its position in the DOM.  For example, pressing return in (caret marked by ^):
        <div contentEditable="true"><div>^Hello</div></div>
        Undo was being enabled, shouldDeleteDOMRange called, etc. when doing no-op deletes (a delete in an empty document 
        for example). 

        Changes to layout tests demonstrate fix.
        
        * editing/EditCommand.cpp:
        (WebCore::EditCommand::apply): Don't call applyEditing for a TypingCommand.  The TypingCommand knows whether or
        not it did work that needs to be applied.
        * editing/Editor.cpp:
        (WebCore::Editor::appliedEditing): Moved code (but did not alter) to changeSelectionAfterCommand.
        (WebCore::Editor::unappliedEditing): Ditto.
        (WebCore::Editor::reappliedEditing): Ditto.
        (WebCore::Editor::changeSelectionAfterCommand): Moved code from *appliedEditing into here.  Also call out to
        EditorClient::respondToChangedSelection() for commands that changed the selection's position in the document
        even if they did not change it's position in the DOM.  Any TypingCommand that gets this far changed it's position
        in the document.
        * editing/Editor.h:
        * editing/TypingCommand.cpp:
        (WebCore::TypingCommand::TypingCommand): Removed unused m_appliedEditing.
        (WebCore::TypingCommand::typingAddedToOpenCommand): Always apply editing.  We won't get this far if we don't need to.
        (WebCore::TypingCommand::deleteKeyPressed): Don't do any of the things that only make sense for Range selections, like
        adding to the killring and responding to a change in selections if the delete was a no-op.
        (WebCore::TypingCommand::forwardDeleteKeyPressed): Ditto.
        * editing/TypingCommand.h:

2009-03-05  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Adam Roben.

        Eliminate a ref-counting leak in InspectorController.

        * inspector/InspectorController.h: Made constructor private, added static create method.
        * page/Page.cpp:
        (WebCore::Page::Page): Change initializer to use static create method.

2009-03-05  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Timothy Hatcher.

        https://bugs.webkit.org/show_bug.cgi?id=24355
        Add InspectorController.idl and convert InspectorController.cpp to use
        JSC bindings.

        * DerivedSources.make: Added InspectorController.idl
        * WebCore.xcodeproj/project.pbxproj: Added idl, bindings files.
        * bindings/js/JSInspectorControllerCustom.cpp: Added.
        * bindings/scripts/CodeGeneratorJS.pm: Added Array to no-header types.
        * inspector/InspectorController.cpp: Removed hand-rolled bindings.
        * inspector/InspectorController.h: Added platform and addSourceToFrame methods.
        * inspector/InspectorController.idl: Added.
        * page/Page.h: Changed member to RefPtr since InspectorController is now ref-counted.

2009-03-05  Eric Carlson  <eric.carlson@apple.com>

        Reviewed by Simon Fraser.

        https://bugs.webkit.org/show_bug.cgi?id=24400
        Bug 24400: Remove "start", "end", "loopStart", "loopEnd", "currentLoop", and "playCount" 
        media element attributes

        Test: media/video-loop.html

        * html/HTMLMediaElement.cpp:
        (WebCore::HTMLMediaElement::HTMLMediaElement):
        (WebCore::HTMLMediaElement::loadInternal):
        (WebCore::HTMLMediaElement::setNetworkState):
        (WebCore::HTMLMediaElement::seek):
        (WebCore::HTMLMediaElement::playInternal):
        (WebCore::HTMLMediaElement::loop):
        (WebCore::HTMLMediaElement::setLoop):
        (WebCore::HTMLMediaElement::mediaPlayerTimeChanged):
        (WebCore::HTMLMediaElement::endedPlayback):
        (WebCore::HTMLMediaElement::updatePlayState):
        * html/HTMLMediaElement.h:
        * html/HTMLMediaElement.idl:

2009-03-05  Steve Falkenburg  <sfalken@apple.com>

        <rdar://problem/6651112> Safari asks about re-posting a form even when page is cached
        
        Reviewed by Ada Chan.

        * platform/network/cf/ResourceHandleCFNet.cpp:
        (WebCore::ResourceHandle::willLoadFromCache): Ported from Mac version.

2009-03-05  David Hyatt  <hyatt@apple.com>

        Reviewed by Eric Seidel

        https://bugs.webkit.org/show_bug.cgi?id=24248

        Make sure painting of overflow controls checks that visibility:visible is set on the block before
        painting.  Pixel tests caught this regression.

        Make sure resizer painting pushes a clip of the corner rect.  It was relying on the clip layers happened
        to do to their bounds (which had nothing to do with overflow).

        * rendering/RenderBlock.cpp:
        (WebCore::RenderBlock::paint):
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::paintResizer):

2009-03-05  Yong Li  <yong.li@torchmobile.com>

        Reviewed by Simon Fraser.

        https://bugs.webkit.org/show_bug.cgi?id=24386
        A faster implementation of extractMIMETypeFromMediaType.

        * platform/network/HTTPParsers.cpp:
        (WebCore::extractMIMETypeFromMediaType):

2009-03-05  Yong Li  <yong.li@torchmobile.com>

        Reviewed by Antti Koivisto.

        https://bugs.webkit.org/show_bug.cgi?id=24392
        Do not get the current time for unless we're doing PRELOAD_DEBUG as this
        can be unnecessarily expensive.

        * html/PreloadScanner.cpp:
        (WebCore::PreloadScanner::write):

2009-03-04  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Dave Hyatt

        https://bugs.webkit.org/show_bug.cgi?id=24327
        
        When mapping points and hit testing through transforms, work
        correctly when acclerated animations of transforms are running.
        
        Tested by LayoutTests/animations/animation-hit-test-transform.html,
        which only failed when ACCELERATED_COMPOSITING was turned on.

        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::currentTransform):
        * rendering/RenderLayer.h:
        * rendering/RenderObject.cpp:
        (WebCore::RenderObject::transformFromContainer):

2009-03-05  Mike Belshe  <mike@belshe.com>

        Reviewed by Darin Fisher.

        https://bugs.webkit.org/show_bug.cgi?id=24391
        Frame.cpp uses JSC specific includes

        * page/Frame.cpp:

2009-03-05  Gustavo Noronha Silva  <gns@gnome.org>

        Reviewed by Alexey Proskuryakov.

        https://bugs.webkit.org/show_bug.cgi?id=24389
        WebKitGTK+ crashes when cancelling plugin loads

        Remove bogus calls to the client's didFinishLoading method from
        our ResourceHandle::cancel implementation. Calling
        didFinishLoading here is mostly inoffensive for most loads, but
        causes crashes when plugin loads are cancelled.

        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::ResourceHandle::cancel):

2009-03-05  Steve Falkenburg  <sfalken@apple.com>

        Windows build fix.
        
        * WebCore.vcproj/QTMovieWin.vcproj:

2009-03-05  Adam Treat  <adam.treat@torchmobile.com>

        Build fix for when ENABLE_NETSCAPE_PLUGIN_API = 0.  The method
        ScriptController::jsObjectForPluginElement(HTMLPlugInElement*); is not
        protected by an #if and uses HTMLPlugInElement so it must be included.

        * bindings/js/ScriptController.cpp:

2009-03-05  Dimitri Glazkov  <dglazkov@chromium.org>

        GTK Build fix.

        * inspector/ConsoleMessage.h: Use proper header.

2009-03-05  Dimitri Glazkov  <dglazkov@chromium.org>

        Reviewed by Timothy Hatcher.

        https://bugs.webkit.org/show_bug.cgi?id=24376
        Split InspectorController.cpp file into separate classes.

        * GNUmakefile.am: Modified to include new files.
        * WebCore.pro: Ditto.
        * WebCore.scons: Ditto.
        * WebCore.vcproj/WebCore.vcproj: Ditto.
        * WebCore.xcodeproj/project.pbxproj: Ditto.
        * WebCoreSources.bkl: Ditto.
        * inspector/ConsoleMessage.cpp: Added.
        * inspector/ConsoleMessage.h: Added.
        * inspector/InspectorController.cpp:
        * inspector/InspectorDOMStorageResource.cpp: Added.
        * inspector/InspectorDOMStorageResource.h: Added.
        * inspector/InspectorDatabaseResource.cpp: Added.
        * inspector/InspectorDatabaseResource.h: Added.
        * inspector/InspectorResource.cpp: Added.
        * inspector/InspectorResource.h: Added.

2009-03-05  Eric Carlson  <eric.carlson@apple.com>

        Reviewed by Darin Adler

        https://bugs.webkit.org/show_bug.cgi?id=24387
        Remove media element bufferingRate attribute. No test necessary as there
        were none for this attribute.

        * html/HTMLMediaElement.cpp:
        (WebCore::HTMLMediaElement::HTMLMediaElement): Don't initialize m_bufferingRate.
        (WebCore::HTMLMediaElement::loadInternal): Ditto.
        (WebCore::HTMLMediaElement::setNetworkState): Ditto.
        (WebCore::HTMLMediaElement::progressEventTimerFired): Don't calculate m_bufferingRate.
        * html/HTMLMediaElement.h: Remove m_bufferingRate.
        * html/HTMLMediaElement.idl: Ditto.

2009-03-05  Xan Lopez  <xan@gnome.org>

        Reviewed by Mark Rowe.

        https://bugs.webkit.org/show_bug.cgi?id=24377
        [GTK] Comply better with coding style guidelines in ResourceHandleSoup

        Use C++ booleans, 0 instead of NULL, prefix boolean variables with
        'is', do not use 'else if' if the previous if had a return, delete
        trailing whitespace, etc.

        * platform/network/ResourceHandleInternal.h:
        (WebCore::ResourceHandleInternal::ResourceHandleInternal):
        * platform/network/soup/ResourceHandleSoup.cpp:
        (WebCore::WebCoreSynchronousLoader::WebCoreSynchronousLoader):
        (WebCore::fillResponseFromMessage):
        (WebCore::restartedCallback):
        (WebCore::finishedCallback):
        (WebCore::parseDataUrl):
        (WebCore::ensureSessionIsInitialized):
        (WebCore::ResourceHandle::startHttp):
        (WebCore::reportUnknownProtocolError):
        (WebCore::ResourceHandle::start):
        (WebCore::cleanupGioOperation):
        (WebCore::closeCallback):
        (WebCore::readCallback):
        (WebCore::openCallback):
        (WebCore::queryInfoCallback):
        (WebCore::ResourceHandle::startGio):

2009-03-04  Simon Fraser  <simon.fraser@apple.com>

        Reviewed by Eric Seidel

        https://bugs.webkit.org/show_bug.cgi?id=24328
        
        If an element has backface-visibility: hidden, hit testing should not
        hit the back sides of elements. Test for that by inverting the accumulated
        transform and looking at the z vector.

        Tested by transforms/3d/hit-testing/backface-hit-test.html
        
        * rendering/RenderLayer.cpp:
        (WebCore::RenderLayer::hitTestLayer):

2009-03-04  Adam Langley  <agl@google.com>

        Reviewed by Darin Fisher.

        r41362 mistakenly added functions "RenderThemeWin::" into
        RenderThemeChromiumGtk.cpp. When correcting this, I somehow changed
        them to void return types. This patch switches them back to ints.

        https://bugs.webkit.org/show_bug.cgi?id=24360

        * rendering/RenderThemeChromiumLinux.cpp:
        (WebCore::RenderThemeChromiumLinux::buttonInternalPaddingLeft):
        (WebCore::RenderThemeChromiumLinux::buttonInternalPaddingRight):
        (WebCore::RenderThemeChromiumLinux::buttonInternalPaddingTop):
        (WebCore::RenderThemeChromiumLinux::buttonInternalPaddingBottom):

2009-03-04  Adam Langley  <agl@google.com>

        Reviewed by Darin Fisher.

        Chromium Linux: change some metrics to better match Windows after
        r41416. Aesthetically this may be worse, but web-compat is king.

        Also, the previous code had an off by one error when drawing
        scrollbars which caused the scrollbar to overflow it's bounds by 1px
        at the right and bottom edges.

        https://bugs.webkit.org/show_bug.cgi?id=24369

        * rendering/RenderThemeChromiumLinux.cpp:
        (WebCore::):
        (WebCore::RenderThemeChromiumLinux::systemFont):
        * platform/chromium/ScrollbarThemeChromiumLinux.cpp:
        (WebCore::ScrollbarThemeChromium::paintTrackPiece):

2009-03-04  Eric Seidel  <eric@webkit.org>

        Reviewed by Adam Roben.

        Add an ASSERT to better demonstrate the cause of the crash in
        https://bugs.webkit.org/show_bug.cgi?id=23736

        * loader/DocLoader.cpp:
        (WebCore::DocLoader::~DocLoader):

2009-03-04  David Levin  <levin@chromium.org>

        Reviewed by Alexey Proskuryakov.

        Bug 24337: Assert when doing sync XHR in a worker for a cacheable response.
        <https://bugs.webkit.org/show_bug.cgi?id=24337>

        Tests: http/tests/xmlhttprequest/access-control-basic-denied-preflight-cache.html
               http/tests/xmlhttprequest/workers/xmlhttprequest-file-not-found.html

        * loader/mac/ResourceLoaderMac.mm:
        (WebCore::ResourceLoader::willCacheResponse):
        An identifier is only asssigned if resource load callbacks are done.  So don't send
        willCacheResponse which uses identifier if resource load callbacks aren't being sent.

2009-03-04  David Levin  <levin@chromium.org>

        Reviewed by Alexey Proskuryakov.

        Bug 24330: Sync xhr in workers should send an abort exception when the worker is terminated.
        <https://bugs.webkit.org/show_bug.cgi?id=24330>

        Test: http/tests/xmlhttprequest/workers/abort-exception-assert.html

        * dom/ExceptionCode.cpp:
        (WebCore::xmlHttpRequestExceptionNames):
        Added missing ABORT_ERR whose absence caused an assert.

        * loader/WorkerThreadableLoader.cpp:
        (WebCore::WorkerThreadableLoader::loadResourceSynchronously):
        (WebCore::WorkerThreadableLoader::MainThreadBridge::cancel):
        Add more logic to handle the termination case for sync xhr.

2009-03-04  Eric Carlson  <eric.carlson@apple.com>

        Reviewed by Antti Koivisto.

        https://bugs.webkit.org/show_bug.cgi?id=24364
        Add support for HTMLMediaElement canPlayType method. Make
        MediaPlayer::supportsType take a ContentType instead of a
        separate mime type and codecs parameter.

        Test: media/video-can-play-type.html

        * dom/DOMImplementation.cpp:
        (WebCore::DOMImplementation::createDocument):
        * html/HTMLMediaElement.cpp:
        (WebCore::HTMLMediaElement::canPlayType): Add canPlayType method.
        (WebCore::HTMLMediaElement::loadInternal): selectMediaURL returns a ContentType.
        (WebCore::HTMLMediaElement::selectMediaURL): Return a ContentType instead of raw MIME type.
        (WebCore::HTMLMediaElement::initialURL): selectMediaURL returns a ContentType.
        * html/HTMLMediaElement.h: Add prototype.
        * html/HTMLMediaElement.idl: Ditto.
        * platform/graphics/MediaPlayer.cpp:
        (WebCore::MediaPlayer::load):&nb