2010-12-01 Geoffrey Garen Build fix: Added a forwarding header. * ForwardingHeaders/wtf/OSAllocator.h: Added. 2010-12-01 Jia Pu Reviewed by Darin Adler. Support multiple correction candidates panel for misspelled word on Mac OS X. https://bugs.webkit.org/show_bug.cgi?id=50137 Test: platform/mac/editing/spelling/delete-autocorrected-word-1.html Manual test: WebCore/manual-tests/autocorrection/dismiss-multiple-guesses.html WebCore/manual-tests/autocorrection/select-from-multiple-guesses.html This patch is for supporting multiple correction suggestion panel on Mac OS X. The behavior and implementation is similar to that of reversion candiate panel. * WebCore.exp.in: Added and removed symbols. * editing/CorrectionPanelInfo.h: Added new correction panel type, PanelTypeSpellingSuggestions. * editing/Editor.cpp: (WebCore::markerTypesForAutocorrection): Convenient function for getting marker types to be added to autocorrected word. (WebCore::boundingBoxForRange): Convenient function for calculating the bounding box of a range. (WebCore::markerTypesForReplacement): Convenient function for getting marker types to be added to word that has been replaced by various correction panel. (WebCore::Editor::respondToChangedSelection): Added code to trigger timer for multiple suggestion panel. (WebCore::Editor::guessesForMisspelledSelection): Adopted new signature of EditorClient::getGuessesForWord(). (WebCore::Editor::markMisspellingsAfterTypingToWord): Adopted new signature of applyCorrectionPanelInfo(). (WebCore::Editor::markAllMisspellingsAndBadGrammarInRanges): Ditto. (WebCore::Editor::correctionPanelTimerFired): Added code to handle new multiple suggestion panel. (WebCore::Editor::handleCorrectionPanelResult): Replaced handleRejectedCorrection(), since with new AppKit correction panel API, not only rejection, but also confirmation can be handled in callback. (WebCore::Editor::stopCorrectionPanelTimer): Moved the code that clears correction panel info from dismissCorrectionPanel() to this function to fix a bug that would occur when deleting autocorrected word. (WebCore::Editor::dismissCorrectionPanel): Ditto. (WebCore::Editor::applyCorrectionPanelInfo): Added code to handle new multiple suggestion panel. * editing/Editor.h: Replaced handleRejectedCorrection() with handleCorrectionPanelResult(). Changed the signature of applyCorrectionPanelInfo() to hanlde new correction panel type. * editing/TextCheckingHelper.cpp: (WebCore::TextCheckingHelper::guessesForMisspelledOrUngrammaticalRange): Adopted new EditorClient::getGuessesForWord(). * editing/TextCheckingHelper.h: Made paragraphRange() public. * loader/EmptyClients.h: (WebCore::EmptyEditorClient::showCorrectionPanel): Adopted new signature defined in base class. (WebCore::EmptyEditorClient::getGuessesForWord): Ditto. * manual-tests/autocorrection/dismiss-multiple-guesses.html: Added. * manual-tests/autocorrection/select-from-multiple-guesses.html: Added. * page/EditorClient.h: Changed showCorrectionPanel() to allow passing multiple correction candidates. Changed getGuessesForWord() to allow passing in contextual text for better correction suggestions (for those spellcheckers that utilizes such contextual information). 2010-12-01 David Hyatt Reviewed by James Robinson. https://bugs.webkit.org/show_bug.cgi?id=50360 Make sure to include the origin adjustment code that I added on Mac to keep truncation-rtl.html passing. * platform/ScrollView.cpp: (WebCore::ScrollView::updateScrollbars): 2010-12-01 David Hyatt Reviewed by Dan Bernstein. https://bugs.webkit.org/show_bug.cgi?id=46645 Generalize overflow section scrollOriginX code to be a point and to work in both horizontal and vertical modes. Not testable yet, since the rightmost/leftmost/topmost/lowestPosition functions are returning horrendously wrong values in the vertical text case for overflow:auto objects. * rendering/RenderLayer.cpp: (WebCore::RenderLayer::RenderLayer): (WebCore::RenderLayer::scrollToOffset): (WebCore::RenderLayer::computeScrollDimensions): (WebCore::RenderLayer::updateOverflowStatus): * rendering/RenderLayer.h: (WebCore::RenderLayer::scrolledContentOffset): (WebCore::RenderLayer::scrollXOffset): (WebCore::RenderLayer::scrollYOffset): (WebCore::RenderLayer::scrollToXOffset): (WebCore::RenderLayer::scrollToYOffset): 2010-12-01 Steve Falkenburg Reviewed by Adam Roben. WinCairo build should not use link-time code generation (LTCG) https://bugs.webkit.org/show_bug.cgi?id=50353 * WebCore.vcproj/QTMovieWin.vcproj: * WebCore.vcproj/WebCore.vcproj: 2010-12-01 David Hyatt Reviewed by Darin Adler. https://bugs.webkit.org/show_bug.cgi?id=46645 Part 1 of making scrolling work with vertical text. This patch generalizes the scrollOriginX concept recently added for horizontal RTL documents and applies it to writing modes as well. Now the scrollOrigin is a point, since you can start off locked to the bottom or locked to the right. This patch also fixes numerous bugs with the scrollOrigin code using native Mac NSScrollViews and makes them behave the same as the cross-platform code (allowing for cross-platform results to be landed). * page/FrameView.cpp: (WebCore::FrameView::adjustViewSize): * page/mac/WebCoreFrameView.h: * platform/ScrollView.cpp: (WebCore::ScrollView::ScrollView): (WebCore::ScrollView::maximumScrollPosition): (WebCore::ScrollView::minimumScrollPosition): (WebCore::ScrollView::valueChanged): (WebCore::ScrollView::updateScrollbars): (WebCore::ScrollView::wheelEvent): * platform/ScrollView.h: (WebCore::ScrollView::scrollOrigin): * platform/mac/ScrollViewMac.mm: (WebCore::ScrollView::platformSetScrollPosition): (WebCore::ScrollView::platformSetScrollOrigin): * rendering/RenderBox.cpp: (WebCore::RenderBox::paintRootBoxDecorations): * rendering/RenderView.cpp: (WebCore::RenderView::layout): (WebCore::RenderView::docTop): (WebCore::RenderView::docHeight): (WebCore::RenderView::docLeft): (WebCore::RenderView::docWidth): * rendering/RenderView.h: 2010-11-30 Steve Block Reviewed by Jeremy Orlow. Upstream recent changes to WebCore/platform/network/android https://bugs.webkit.org/show_bug.cgi?id=50224 Android implementation changes only, no new tests. * platform/network/NetworkingContext.h: * platform/network/android/CookieJarAndroid.cpp: (WebCore::setCookies): (WebCore::cookies): (WebCore::cookieRequestHeaderFieldValue): (WebCore::cookiesEnabled): * platform/network/android/ProxyServerAndroid.cpp: Added. (WebCore::proxyServersForURL): * platform/network/android/ResourceHandleAndroid.cpp: (WebCore::ResourceHandle::start): (WebCore::ResourceHandle::loadResourceSynchronously): * platform/network/android/ResourceRequestAndroid.cpp: (WebCore::initializeMaximumHTTPConnectionCountPerHost): 2010-12-01 Steve Falkenburg Reviewed by Adam Roben. vcproj changes can't be applied cleanly by the Windows EWS bot https://bugs.webkit.org/show_bug.cgi?id=50328 * WebCore.vcproj/QTMovieWin.vcproj: Modified property svn:eol-style. * WebCore.vcproj/QTMovieWinCommon.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCore.sln: Modified property svn:eol-style. * WebCore.vcproj/WebCore.submit.sln: Modified property svn:eol-style. * WebCore.vcproj/WebCore.vcproj: Modified property svn:eol-style. * WebCore.vcproj/WebCoreCFNetwork.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCoreCG.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCoreCURL.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCoreCairo.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCoreCommon.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCoreGenerated.vcproj: Modified property svn:eol-style. * WebCore.vcproj/WebCoreGeneratedCairo.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCoreGeneratedCommon.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCoreMediaQT.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCorePthreads.vsprops: Added property svn:eol-style. * WebCore.vcproj/WebCoreQuartzCore.vsprops: Added property svn:eol-style. 2010-11-30 Vangelis Kokkevis Reviewed by Darin Fisher. Adds accelerated compositing trigger flags to Settings to allow ports to set them via command line flags, etc. https://bugs.webkit.org/show_bug.cgi?id=50301 * page/Settings.cpp: (WebCore::Settings::Settings): (WebCore::Settings::setAcceleratedCompositingFor3DTransformsEnabled): (WebCore::Settings::setAcceleratedCompositingForVideoEnabled): (WebCore::Settings::setAcceleratedCompositingForPluginsEnabled): (WebCore::Settings::setAcceleratedCompositingForCanvasEnabled): (WebCore::Settings::setAcceleratedCompositingForAnimationEnabled): * page/Settings.h: (WebCore::Settings::acceleratedCompositingFor3DTransformsEnabled): (WebCore::Settings::acceleratedCompositingForVideoEnabled): (WebCore::Settings::acceleratedCompositingForPluginsEnabled): (WebCore::Settings::acceleratedCompositingForCanvasEnabled): (WebCore::Settings::acceleratedCompositingForAnimationEnabled): 2010-12-01 Darin Adler Fixed warning seen on buildbot. * platform/graphics/chromium/ContentLayerChromium.cpp: (WebCore::ContentLayerChromium::SharedValues::SharedValues): Added #if. 2010-12-01 Ryosuke Niwa Reviewed by Darin Adler, Tony Chang, and unofficially by Enrica Casucci. chrome.dll!WebCore::RangeBoundaryPoint::toPosition ReadAV@NULL (cf0d0f28bc56f2591cc74f71b46036ea) https://bugs.webkit.org/show_bug.cgi?id=47808 The crash was caused by moveParagraphWithClones' not copying the entire paragraph when called by doApplyForSingleParagraph, which was induced by a bug in cloneParagraphUnderNewElement which ended the cloning process early when the start node and the end node didn't share the same parent node. Fixed the bug in cloneParagraphUnderNewElement by calling traverseNextSibling(outerNode) instead of nextSibling(). Also added an early exit to InsertListCommand::doApply when lastSelectionRange is null. Tests: editing/execCommand/switch-multiple-list-items-crash.html editing/execCommand/switch-multiple-list-items.html * editing/CompositeEditCommand.cpp: (WebCore::CompositeEditCommand::cloneParagraphUnderNewElement): * editing/InsertListCommand.cpp: (WebCore::InsertListCommand::doApply): 2010-12-01 Kevin Ollivier [wx] Build fixes after recent trunk changes. * platform/graphics/wx/ImageBufferWx.cpp: (WebCore::ImageBuffer::ImageBuffer): * platform/wx/LocalizedStringsWx.cpp: (WebCore::platformDefaultLanguage): 2010-12-01 Sheriff Bot Unreviewed, rolling out r73047. http://trac.webkit.org/changeset/73047 https://bugs.webkit.org/show_bug.cgi?id=50339 missing bug number (Requested by rniwa on #webkit). * editing/CompositeEditCommand.cpp: (WebCore::CompositeEditCommand::cloneParagraphUnderNewElement): * editing/InsertListCommand.cpp: (WebCore::InsertListCommand::doApply): 2010-12-01 Ryosuke Niwa Reviewed by Darin Adler, Tony Chang, and unofficially by Enrica Casucci. The crash was caused by moveParagraphWithClones' not copying the entire paragraph when called by doApplyForSingleParagraph, which was induced by a bug in cloneParagraphUnderNewElement which ended the cloning process early when the start node and the end node didn't share the same parent node. Fixed the bug in cloneParagraphUnderNewElement by calling traverseNextSibling(outerNode) instead of nextSibling(). Also added an early exit to InsertListCommand::doApply when lastSelectionRange is null. Tests: editing/execCommand/switch-multiple-list-items-crash.html editing/execCommand/switch-multiple-list-items.html * editing/CompositeEditCommand.cpp: (WebCore::CompositeEditCommand::cloneParagraphUnderNewElement): * editing/InsertListCommand.cpp: (WebCore::InsertListCommand::doApply): 2010-12-01 Grace Kloba Reviewed by Kenneth Russell. [chromium] ContentLayerChromium shader should honor the platform Sk_x_SHIFT value instead of assuming BGRA color https://bugs.webkit.org/show_bug.cgi?id=50189 * platform/graphics/chromium/ContentLayerChromium.cpp: (WebCore::ContentLayerChromium::SharedValues::SharedValues):Use SK_B32_SHIFT to decide whether to use RGBA or BGRA shader for Skia port. 2010-12-01 Nico Weber Reviewed by Dimitri Glazkov. Fix clang warnings caused by -Wlogical-op-parentheses https://bugs.webkit.org/show_bug.cgi?id=50324 * platform/graphics/mac/ComplexTextController.cpp: (WebCore::ComplexTextController::adjustGlyphsAndAdvances): 2010-12-01 Martin Robinson Build fix for GTK+. * platform/gtk/PopupMenuGtk.cpp: (WebCore::PopupMenuGtk::typeAheadFind): Uncomment mistakenly commented line. 2010-12-01 Apelete Seketeli and Martin Robinson Reviewed by Xan Lopez. [Gtk] Open menulists should support typeahead find https://bugs.webkit.org/show_bug.cgi?id=27443 Added typeahead find support for open GTK+ menulists. Test: platform/gtk/fast/forms/menulist-typeahead-find.html * platform/gtk/PopupMenuGtk.cpp: (WebCore::PopupMenuGtk::PopupMenuGtk): Initialize new members. (WebCore::PopupMenuGtk::show): Track currently selected menu item via a signal handler. (WebCore::PopupMenuGtk::typeAheadFind): Added. (WebCore::PopupMenuGtk::menuUnmapped): Reset typeahead find state when menu is unmapped. (WebCore::PopupMenuGtk::resetTypeAheadFindState): Added. (WebCore::PopupMenuGtk::selectItemCallback): Added. (WebCore::PopupMenuGtk::keyPressEventCallback): Added. * platform/gtk/PopupMenuGtk.h: Added new members to track typeahead find state. 2010-12-01 Nikolas Zimmermann Reviewed by Xan Lopez. Split DECLARE_ANIMATED_PROPERTY* in DECLARE/DEFINE parts for SVGAnimatedNumber https://bugs.webkit.org/show_bug.cgi?id=50323 Continuing the work on bug 42025: Introduce DECLARE_ANIMATED_NUMBER / DEFINE_ANIMATED_NUMBER. The animated property declaration lives in the header, the definition in the cpp file, to avoid the dependency on SVGNames.h in all headers. No functional changes, thus no new tests. * svg/SVGAnimatedNumber.h: * svg/SVGComponentTransferFunctionElement.cpp: * svg/SVGComponentTransferFunctionElement.h: * svg/SVGFECompositeElement.cpp: * svg/SVGFECompositeElement.h: * svg/SVGFEConvolveMatrixElement.cpp: * svg/SVGFEConvolveMatrixElement.h: * svg/SVGFEDiffuseLightingElement.cpp: * svg/SVGFEDiffuseLightingElement.h: * svg/SVGFEDisplacementMapElement.cpp: * svg/SVGFEDisplacementMapElement.h: * svg/SVGFEGaussianBlurElement.cpp: * svg/SVGFEGaussianBlurElement.h: * svg/SVGFELightElement.cpp: * svg/SVGFELightElement.h: * svg/SVGFEMorphologyElement.cpp: * svg/SVGFEMorphologyElement.h: * svg/SVGFEOffsetElement.cpp: * svg/SVGFEOffsetElement.h: * svg/SVGFESpecularLightingElement.cpp: * svg/SVGFESpecularLightingElement.h: * svg/SVGFETurbulenceElement.cpp: * svg/SVGFETurbulenceElement.h: * svg/SVGPathElement.cpp: * svg/SVGPathElement.h: * svg/SVGStopElement.cpp: * svg/SVGStopElement.h: (WebCore::SVGStopElement::isGradientStop): 2010-12-01 Martin Robinson Reviewed by Xan Lopez. Sliders are not transformed after r50188 https://bugs.webkit.org/show_bug.cgi?id=50317 Need to call updateLayerTransform during slider layout, as sliders may be transformed. No new tests. This patch should make fast/forms/slider-transformed.html pass on the GTK+ bots. * rendering/RenderSlider.cpp: (WebCore::RenderSlider::layout): Call updateLayerTransform here. 2010-11-30 Pavel Feldman Reviewed by Yury Semikhatsky. Web Inspector: do not make helpScreen.css lazily loaded. https://bugs.webkit.org/show_bug.cgi?id=50304 * inspector/front-end/HelpScreen.js: (WebInspector.HelpScreen): (WebInspector.HelpScreen.prototype._onBlur): * inspector/front-end/inspector.html: 2010-12-01 Nikolas Zimmermann Reviewed by Xan Lopez. Split DECLARE_ANIMATED_PROPERTY* in DECLARE/DEFINE parts for SVGLengthList/SVGNumberList/SVGTransformList https://bugs.webkit.org/show_bug.cgi?id=50322 Continuing the work on bug 42025: Introduce DECLARE_ANIMATED_(LENGTH|NUMBER|TRANSFORM)_LIST / DEFINE_ANIMATED_(LENGTH|NUMBER|TRANSFORM). The animated property declaration lives in the header, the definition in the cpp file, to avoid the dependency on SVGNames.h in all headers. No functional changes, thus no new tests. * svg/SVGAnimatedLengthList.h: * svg/SVGAnimatedNumberList.h: * svg/SVGAnimatedTransformList.h: * svg/SVGComponentTransferFunctionElement.cpp: * svg/SVGComponentTransferFunctionElement.h: * svg/SVGFEColorMatrixElement.cpp: * svg/SVGFEColorMatrixElement.h: * svg/SVGFEConvolveMatrixElement.cpp: * svg/SVGFEConvolveMatrixElement.h: * svg/SVGGradientElement.cpp: * svg/SVGGradientElement.h: (WebCore::SVGGradientElement::needsPendingResourceHandling): * svg/SVGPatternElement.cpp: * svg/SVGPatternElement.h: * svg/SVGPolygonElement.cpp: * svg/SVGPolylineElement.cpp: * svg/SVGStyledTransformableElement.cpp: * svg/SVGStyledTransformableElement.h: * svg/SVGTextElement.cpp: (WebCore::SVGTextElement::SVGTextElement): * svg/SVGTextElement.h: (WebCore::SVGTextElement::localCoordinateSpaceTransform): * svg/SVGTextPositioningElement.cpp: * svg/SVGTextPositioningElement.h: * svg/properties/SVGAnimatedPropertyMacros.h: 2010-12-01 Alexander Pavlov Reviewed by Yury Semikhatsky. Web Inspector: name-only property appears in style declaration but not in the Styles pane https://bugs.webkit.org/show_bug.cgi?id=49663 Check that at least one property has been parsed from the user input. * inspector/InspectorStyleSheet.cpp: (WebCore::InspectorStyle::setPropertyText): 2010-12-01 Andrey Kosyakov Reviewed by Pavel Feldman. Web Inspector: disable cookies tab in network resource view by default, enable for chromium https://bugs.webkit.org/show_bug.cgi?id=50249 * inspector/front-end/NetworkItemView.js: (WebInspector.NetworkItemView): (WebInspector.NetworkItemView.prototype.resize): * inspector/front-end/Settings.js: 2010-12-01 Nikolas Zimmermann Reviewed by Xan Lopez. Split DECLARE_ANIMATED_PROPERTY* in DECLARE/DEFINE parts for SVGLength https://bugs.webkit.org/show_bug.cgi?id=50316 Introduce DECLARE_ANIMATED_LENGTH / DEFINE_ANIMATED_LENGTH, splitting up the original DECLARE_ANIMATED_PROPERTY_NEW macro in a part that goes in the .cpp file and another that remains in the header. This avoids having to include SVGNames.h in all SVG*Element files. No functional changes, thus no new tests. * svg/SVGAnimatedLength.h: * svg/SVGCircleElement.cpp: * svg/SVGCircleElement.h: (WebCore::SVGCircleElement::isValid): * svg/SVGCursorElement.cpp: * svg/SVGCursorElement.h: (WebCore::SVGCursorElement::isValid): * svg/SVGEllipseElement.cpp: * svg/SVGEllipseElement.h: (WebCore::SVGEllipseElement::isValid): * svg/SVGFilterElement.cpp: * svg/SVGFilterElement.h: * svg/SVGFilterPrimitiveStandardAttributes.cpp: * svg/SVGFilterPrimitiveStandardAttributes.h: * svg/SVGForeignObjectElement.cpp: (WebCore::SVGForeignObjectElement::SVGForeignObjectElement): * svg/SVGForeignObjectElement.h: (WebCore::SVGForeignObjectElement::isValid): * svg/SVGImageElement.cpp: * svg/SVGImageElement.h: (WebCore::SVGImageElement::isValid): * svg/SVGLineElement.cpp: * svg/SVGLineElement.h: (WebCore::SVGLineElement::isValid): (WebCore::SVGLineElement::supportsMarkers): * svg/SVGLinearGradientElement.cpp: * svg/SVGLinearGradientElement.h: * svg/SVGMarkerElement.cpp: (WebCore::SVGMarkerElement::SVGMarkerElement): * svg/SVGMarkerElement.h: * svg/SVGMaskElement.cpp: * svg/SVGMaskElement.h: (WebCore::SVGMaskElement::isValid): (WebCore::SVGMaskElement::needsPendingResourceHandling): * svg/SVGPatternElement.cpp: * svg/SVGPatternElement.h: (WebCore::SVGPatternElement::isValid): (WebCore::SVGPatternElement::needsPendingResourceHandling): * svg/SVGRadialGradientElement.cpp: * svg/SVGRadialGradientElement.h: * svg/SVGRectElement.cpp: * svg/SVGRectElement.h: (WebCore::SVGRectElement::isValid): * svg/SVGSVGElement.cpp: (WebCore::SVGSVGElement::SVGSVGElement): * svg/SVGSVGElement.h: (WebCore::SVGSVGElement::isValid): (WebCore::SVGSVGElement::setContainerSize): (WebCore::SVGSVGElement::containerSize): (WebCore::SVGSVGElement::hasSetContainerSize): (WebCore::SVGSVGElement::currentTranslate): (WebCore::SVGSVGElement::timeContainer): (WebCore::SVGSVGElement::isSVG): (WebCore::SVGSVGElement::rendererIsNeeded): * svg/SVGTextContentElement.cpp: * svg/SVGTextContentElement.h: (WebCore::SVGTextContentElement::isValid): (WebCore::SVGTextContentElement::isTextContent): * svg/SVGTextPathElement.cpp: * svg/SVGTextPathElement.h: * svg/SVGTextPositioningElement.h: * svg/SVGUseElement.cpp: * svg/SVGUseElement.h: (WebCore::SVGUseElement::isValid): (WebCore::SVGUseElement::setUpdatesBlocked): (WebCore::SVGUseElement::isPendingResource): * svg/properties/SVGAnimatedPropertyMacros.h: 2010-12-01 Anton D'Auria Reviewed by Kevin Decker. Deny access to local and session storage in private browsing mode. https://bugs.webkit.org/show_bug.cgi?id=49329 * storage/Storage.cpp: (WebCore::Storage::length): (WebCore::Storage::key): (WebCore::Storage::getItem): (WebCore::Storage::contains): 2010-12-01 Philippe Normand Reviewed by Martin Robinson. Volume control not correctly initialized https://bugs.webkit.org/show_bug.cgi?id=36299 Replaced the mute/volume Timers with g_timeouts which are (for now, at least) more reliable than Timers for one-shot-fire-now actions. Test: media/video-volume.html * platform/graphics/MediaPlayer.cpp: (WebCore::MediaPlayer::~MediaPlayer): Reset the raw pointers to 0 when destructing the player. * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp: (WebCore::mediaPlayerPrivateVolumeChangeTimeoutCallback): (WebCore::mediaPlayerPrivateMuteChangeTimeoutCallback): (WebCore::MediaPlayerPrivateGStreamer::~MediaPlayerPrivateGStreamer): (WebCore::MediaPlayerPrivateGStreamer::notifyPlayerOfVolumeChange): (WebCore::MediaPlayerPrivateGStreamer::volumeChanged): (WebCore::MediaPlayerPrivateGStreamer::notifyPlayerOfMute): (WebCore::MediaPlayerPrivateGStreamer::muteChanged): (WebCore::MediaPlayerPrivateGStreamer::createGSTPlayBin): Set playbin2 volume/mute base on MediaPlayer related values. * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.h: 2010-12-01 Patrick Gansterer Reviewed by Andreas Kling. Cleanup main CMakeLists.txt after r72664, r72672 and r72667 https://bugs.webkit.org/show_bug.cgi?id=50232 * CMakeLists.txt: Removed obsolete dependencies. 2010-12-01 Nikolas Zimmermann Reviewed by Dirk Schulze. Cleanup SVGPropertyTraits https://bugs.webkit.org/show_bug.cgi?id=50229 Move all template specializations for non-POD types (eg. SVGLength, SVGLengthList, etc.) from SVGPropertyTraits into the right SVG* classes, to avoid having to include a dozen SVG*.h files in SVGPropertyTraits.h, which is included by all DOM/JS/V8 binding files and SVGElement.h. No new functionality, thus no new tests. * GNUmakefile.am: * WebCore.gypi: * WebCore.pro: * WebCore.xcodeproj/project.pbxproj: * bindings/scripts/CodeGenerator.pm: * bindings/scripts/CodeGeneratorObjC.pm: * bindings/scripts/CodeGeneratorV8.pm: * svg/SVGAngle.h: * svg/SVGLength.h: * svg/SVGLengthList.h: * svg/SVGMarkerElement.h: * svg/SVGNumberList.h: * svg/SVGPatternElement.h: * svg/SVGPointList.h: * svg/SVGPreserveAspectRatio.h: * svg/SVGRect.h: Added. * svg/SVGSVGElement.h: * svg/SVGStringList.h: * svg/SVGSymbolElement.h: * svg/SVGTransformList.h: * svg/SVGViewElement.h: * svg/SVGViewSpec.h: * svg/properties/SVGListProperty.h: * svg/properties/SVGPropertyTraits.h: * svg/properties/SVGTransformListPropertyTearOff.h: 2010-11-30 Kent Tamura Reviewed by Darin Adler. Refactor HTMLInputElement: Move a part of HTMLInputElement:: defaultEventHandler() to InputTypes https://bugs.webkit.org/show_bug.cgi?id=50097 Move all of keyboard-related event handling to InputType. No new tests because this should not change any behavior. * html/BaseButtonInputType.cpp: (WebCore::BaseButtonInputType::handleKeydownEvent): (WebCore::BaseButtonInputType::handleKeypressEvent): (WebCore::BaseButtonInputType::handleKeyupEvent): * html/BaseButtonInputType.h: * html/BaseCheckableInputType.cpp: (WebCore::BaseCheckableInputType::handleKeydownEvent): (WebCore::BaseCheckableInputType::handleKeypressEvent): * html/BaseCheckableInputType.h: * html/CheckboxInputType.cpp: (WebCore::CheckboxInputType::handleKeyupEvent): * html/CheckboxInputType.h: * html/HTMLInputElement.cpp: (WebCore::HTMLInputElement::defaultEventHandler): We remove !implicitSubmission check for callBaseClassEarly because implicitSubmission can not be true for keydown event and keypress event at this point. * html/InputType.cpp: (WebCore::InputType::shouldSubmitImplicitly): Returns true for a keypress event with "\r". Note that we don't overload this function for BUTTON, FILE, IMAGE, RESET, and SUBMIT because a keypress event with "\r" for these types returns from HTMLInputElement::defaultEventHandler() before implicit submission checking. (WebCore::InputType::handleKeypressEvent): (WebCore::InputType::handleKeyupEvent): * html/InputType.h: * html/RadioInputType.cpp: (WebCore::RadioInputType::handleKeydownEvent): (WebCore::RadioInputType::handleKeyupEvent): * html/RadioInputType.h: * html/TextFieldInputType.cpp: (WebCore::TextFieldInputType::shouldSubmitImplicitly): * html/TextFieldInputType.h: >>>>>>> .r73054 2010-11-30 Patrick Gansterer Reviewed by Andreas Kling. Move CSS generator logic into main CMakeLists.txt file https://bugs.webkit.org/show_bug.cgi?id=50230 * CMakeLists.txt: 2010-11-30 Ojan Vafai Revert r72876. It caused a ~30% perf regression in chromium's bloat-http test https://bugs.webkit.org/show_bug.cgi?id=50288 2010-11-29 Adam Barth Reviewed by Darin Adler. Introduce the notion of a "display-isolated" URL scheme for use by Chrome-internal URLs https://bugs.webkit.org/show_bug.cgi?id=50182 A display-isolated URL can only be displayed (e.g., put in an iframe, hyperlinked to) by documents from that scheme. In a sense, this is a generalization of some of the protections we give file URLs, but instead of lumping them all together into one "local" bucket, this patch creates a separate bucket for each scheme. For a while, I tried using a separate bucket for each origin. That would have played nicely with what Blob URLs are trying to do, but some "chrome" URL pages rely on being able to display other chrome URL pages, even in different origins. For example, the New Tab Page shows thumbnails from the "thumbnail" host. This patch also removes a bunch of unused code. I've also propagated the "deprecated" status of deprecatedCanDisplay to deprecatedShouldTreatURLAsLocal because that method has no other callers and is really asking for uppercase/lowercase bugs. I dream of someday removing these functions. page/SecurityOrigin.cpp: (WebCore::SecurityOrigin::canDisplay): (WebCore::SecurityOrigin::deprecatedCanDisplay): platform/SchemeRegistry.cpp: (WebCore::displayIsolatedURLSchemes): (WebCore::SchemeRegistry::registerURLSchemeAsLocal): (WebCore::SchemeRegistry::deprecatedShouldTreatURLAsLocal): (WebCore::SchemeRegistry::shouldTreatURLSchemeAsLocal): (WebCore::SchemeRegistry::registerURLSchemeAsDisplayIsolated): (WebCore::SchemeRegistry::shouldTreatURLSchemeAsDisplayIsolated): platform/SchemeRegistry.h: 2010-11-29 Adam Barth Reviewed by Darin Adler. Introduce the notion of a "display-isolated" URL scheme for use by Chrome-internal URLs https://bugs.webkit.org/show_bug.cgi?id=50182 This patch adds a Chromium API for registering schemes as display-isolated. In a subsequent patch, I'll change the "chrome" scheme in Chrome to be display isolated instead of local. That will prevent file URLs from linking to chrome URLs. public/WebSecurityPolicy.h: src/WebSecurityPolicy.cpp: (WebKit::WebSecurityPolicy::registerURLSchemeAsDisplayIsolated): * page/SecurityOrigin.cpp: (WebCore::SecurityOrigin::canDisplay): (WebCore::SecurityOrigin::deprecatedCanDisplay): * platform/SchemeRegistry.cpp: (WebCore::SchemeRegistry::registerURLSchemeAsLocal): (WebCore::SchemeRegistry::removeURLSchemeRegisteredAsLocal): (WebCore::SchemeRegistry::localURLSchemes): (WebCore::SchemeRegistry::shouldTreatURLAsLocal): (WebCore::SchemeRegistry::shouldTreatURLSchemeAsLocal): * platform/SchemeRegistry.h: 2010-11-30 Chris Fleizach Reviewed by Beth Dakin. WebKit duplicates AXValue and AXDescription on abbr https://bugs.webkit.org/show_bug.cgi?id=50260 stringValue() and accessibilityDescription() were returning the same value for static text with an aria-label. Static text elements should not have an accessibility description, they should only have a stringValue. Test: platform/mac/accessibility/static-text-with-aria-label.html * accessibility/AccessibilityRenderObject.cpp: (WebCore::AccessibilityRenderObject::ariaAccessibilityDescription): (WebCore::AccessibilityRenderObject::accessibilityDescription): (WebCore::AccessibilityRenderObject::text): * accessibility/AccessibilityRenderObject.h: 2010-11-30 Ryosuke Niwa Reviewed by Darin Adler. [HTML5] Rename member variables of HTMLScriptElement https://bugs.webkit.org/show_bug.cgi?id=49705 Renamed m_wasCreatedByParser to m_wasInsertedByParser, m_requested to m_isExternalScript, m_isEvaluated to m_wasAlreadyStarted, and m_firedLoad to m_haveFiredLoad. Also made ScriptElement's haveFiredLoadEvent and removed HTMLScriptElement's haveFiredLoadEvent. This effectively adds the same function to SVGScriptElement, which isn't currently used anywhere. No new tests are added since this is a refactoring. * dom/ScriptElement.cpp: (WebCore::ScriptElement::ScriptElement): (WebCore::ScriptElement::insertedIntoDocument): (WebCore::ScriptElement::childrenChanged): (WebCore::ScriptElement::finishParsingChildren): (WebCore::ScriptElement::requestScript): (WebCore::ScriptElement::evaluateScript): (WebCore::ScriptElement::executeScript): (WebCore::ScriptElement::ignoresLoadRequest): * dom/ScriptElement.h: (WebCore::ScriptElement::haveFiredLoadEvent): (WebCore::ScriptElement::setHaveFiredLoadEvent): (WebCore::ScriptElement::wasInsertedByParser): (WebCore::ScriptElement::wasAlreadyStarted): * html/HTMLScriptElement.cpp: (WebCore::HTMLScriptElement::HTMLScriptElement): (WebCore::HTMLScriptElement::create): (WebCore::HTMLScriptElement::cloneElementWithoutAttributesAndChildren): * html/HTMLScriptElement.h: * svg/SVGScriptElement.cpp: (WebCore::SVGScriptElement::SVGScriptElement): (WebCore::SVGScriptElement::create): (WebCore::SVGScriptElement::svgAttributeChanged): (WebCore::SVGScriptElement::insertedIntoDocument): (WebCore::SVGScriptElement::dispatchLoadEvent): (WebCore::SVGScriptElement::cloneElementWithoutAttributesAndChildren): * svg/SVGScriptElement.h: 2010-11-30 Darin Adler Rolled back most of the FileReaderLoader changes to fix 32-bit builds. * fileapi/FileReaderLoader.cpp: Changed types back to unsigned from unsigned long long * fileapi/FileReaderLoader.h: Ditto. 2010-11-30 Dan Bernstein Reviewed by Darin Adler. Text search should treat all matches as word-start matches when the target begins with a separator character https://bugs.webkit.org/show_bug.cgi?id=50302 * editing/TextIterator.cpp: (WebCore::SearchBuffer::SearchBuffer): Disable the AtWordStarts option if it was specified and the target string begins with one of the "separator" characters. 2010-11-30 Pavel Feldman Reviewed by Dave Hyatt. Web Inspector: [CRASH] on "Revert to this revision" of style with import. https://bugs.webkit.org/show_bug.cgi?id=50256 Rolled back one line from http://trac.webkit.org/changeset/36904 as agreed with David. * css/CSSImportRule.cpp: (WebCore::CSSImportRule::insertedIntoParent): 2010-11-30 Martin Robinson Reviewed by Xan Lopez. [GTK] fillRect shadow has incorrect behavior when using ctx.scale(x,y) https://bugs.webkit.org/show_bug.cgi?id=50283 Adjust the mask rectangle for Cairo shadows, so that the origin is not affected by the scale of the transformation matrix. * platform/graphics/ContextShadow.h: Make this Qt-only member global. * platform/graphics/cairo/ContextShadowCairo.cpp: (WebCore::ContextShadow::beginShadowLayer): Save the original layer area. (WebCore::ContextShadow::endShadowLayer): If the target context has a scale transform, make sure not to scale the blur distance when blitting the shadow. 2010-11-30 Darin Adler Reviewed by Alexey Proskuryakov. Make ProgressEvent 64-bit instead of 32-bit https://bugs.webkit.org/show_bug.cgi?id=50289 * dom/ProgressEvent.cpp: (WebCore::ProgressEvent::ProgressEvent): Changed arguments to unsigned long long. (WebCore::ProgressEvent::initProgressEvent): Ditto. * dom/ProgressEvent.h: Fixed formatting. Changed types to unsigned long long. Made isProgressEvent override private. * dom/ProgressEvent.idl: Tweaked formatting and changed types to unsigned long long. * fileapi/FileReader.cpp: Removed unneeded includes. Used static on const so we'd get an error if the same constant was in a header. (WebCore::FileReader::didReceiveData): Removed unneeded WTF prefix. (WebCore::FileReader::fireEvent): Got rid of local variables that would narrow bytes loaded and total bytes to 32-bit. * fileapi/FileReaderLoader.cpp: (WebCore::FileReaderLoader::didReceiveResponse): Restructured code a bit and added comments. Also added handling for when ArrayBuffer::create returns 0. (WebCore::FileReaderLoader::didReceiveData): Fixed handling of case where more data arrives than is expected so it won't overrun the buffer. Also broke an assertion into two and improved a comment. * fileapi/FileReaderLoader.h: Changed bytesLoaded and totalBytes to unsigned long long. * fileapi/FileWriter.cpp: (WebCore::FileWriter::fireEvent): Removed narrowing casts in call to ProgressEvent::create. * html/canvas/ArrayBuffer.cpp: (WebCore::ArrayBuffer::tryAllocate): Added a comment because the code here seems wrong. * xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::didSendData): Removed narrowing casts in call to XMLHttpRequestProgressEvent::create. (WebCore::XMLHttpRequest::didReceiveData): Removed narrowing casts in call to dispatchProgressEvent. * xml/XMLHttpRequestProgressEvent.h: Fixed formatting. Changed types to unsigned long long. Made isXMLHttpRequestProgressEvent override private. * xml/XMLHttpRequestProgressEvent.idl: Tweaked formatting and changed types to unsigned long long. * xml/XMLHttpRequestProgressEventThrottle.cpp: (WebCore::XMLHttpRequestProgressEventThrottle::dispatchProgressEvent): Changed types to unsigned long long. * xml/XMLHttpRequestProgressEventThrottle.h: Changed types to unsigned long long. 2010-11-30 Yael Aharon Reviewed by Darin Adler. Progress and meter elements should take a form in their constructor like any other form control. https://bugs.webkit.org/show_bug.cgi?id=50195 Tests: fast/dom/HTMLMeterElement/meter-element-form.html fast/dom/HTMLProgressElement/progress-element-form.html * html/HTMLMeterElement.cpp: (WebCore::HTMLMeterElement::HTMLMeterElement): (WebCore::HTMLMeterElement::create): * html/HTMLMeterElement.h: * html/HTMLProgressElement.h: * html/HTMLTagNames.in: 2010-11-30 Tony Chang Reviewed by Adam Barth. [chromium] fix get-int-identifier-special-values.html using TestNetscapePlugIn https://bugs.webkit.org/show_bug.cgi?id=49036 No new tests: Covered by plugins/npruntime/get-int-identifier-special-values.html In http://codereview.chromium.org/11574, the forked chromium test plugin changed how it testGetIntIdentifier to work around a change caused by a silverlight specific plugin fix: http://codereview.chromium.org/11569/diff/1/2 It's not clear to me that this change is needed anymore since silverlight has had multiple releases, testing without this patch works on a test windows build, and safari win seems to run silverlight without this either. Since there's no test for this behavior, I'm rolling it out so we can match the upstream test. * bindings/v8/V8NPUtils.cpp: (WebCore::convertV8ObjectToNPVariant): 2010-11-30 Chris Guillory Reviewed by Chris Fleizach. Computing style on a stale node while sending pending accessibility notification. https://bugs.webkit.org/show_bug.cgi?id=50162 Retain node pointer members of AccessibilityImageMapLink. Test: accessibility/image-map-title-causes-crash.html * accessibility/AccessibilityImageMapLink.cpp: (WebCore::AccessibilityImageMapLink::parentObject): (WebCore::AccessibilityImageMapLink::anchorElement): (WebCore::AccessibilityImageMapLink::url): (WebCore::AccessibilityImageMapLink::elementRect): * accessibility/AccessibilityImageMapLink.h: (WebCore::AccessibilityImageMapLink::areaElement): (WebCore::AccessibilityImageMapLink::mapElement): (WebCore::AccessibilityImageMapLink::node): * dom/Document.cpp: (WebCore::Document::clearAXObjectCache): 2010-11-30 James Robinson Reviewed by Tony Chang. Build fix - add parens in foo || bar && baz to make gcc happy https://bugs.webkit.org/show_bug.cgi?id=50276 gcc 4.4 warns that statements of the form "foo || bar && baz". It wants parens around the "bar && baz" part to make it clear which way the logical operators bind. * editing/TextIterator.cpp: (WebCore::SearchBuffer::search): 2010-11-30 Daniel Bates Reviewed by David Hyatt. Focused should use CSS properties of instead of associated https://bugs.webkit.org/show_bug.cgi?id=49888 Fixes an issue where the CSS properties of an were not used when the was focused. Currently, when focusing an (say by pressing option + tab in Mac Safari) we use the CSS style information for the associated with the . Instead, we should use the CSS style information for the focused . In particular, this will make the us honor the outline- color property of a focused when drawing its focus ring. Tests: fast/images/imagemap-focus-ring-outline-color-explicitly-inherited-from-map.html fast/images/imagemap-focus-ring-outline-color-not-inherited-from-map.html fast/images/imagemap-focus-ring-outline-color.html * rendering/RenderImage.cpp: (WebCore::RenderImage::paintFocusRing): Use the style information of the focused when drawing the focus ring for it. 2010-11-30 Tony Chang Reviewed by James Robinson. [chromium] enable WebCore compiler warnings on Linux take 3 https://bugs.webkit.org/show_bug.cgi?id=50258 * WebCore.gyp/WebCore.gyp: 2010-11-30 Mario Sanchez Prada Reviewed by Chris Fleizach. [GTK] Implement ROLE_COMBO_BOX https://bugs.webkit.org/show_bug.cgi?id=25678 Implement the remaining bits for combo boxes. This patch finishes the implementation, from the point of view of the Atk library, of those objects exposed to ATK as combo boxes, and their related elements (menus and menu items). It therefore implements the proper interfaces for each type of object related to combo boxes (AtkSelection for the combo box, AtkText for every menu item and AtkAction for the combo box, the menu and the menu items), takes care of emitting the proper signals when focus or a given a selection changes and adds a new unit test to check all this new stuff. Make possible to ask an AccessibleMenuListOption for a sensible string representation, so far only available through the private and MSAA related method nameForMSAA. Just moved the implementation of that method to an overriden version of stringValue(), which is platform independent, and called that from nameForMSAA(). * accessibility/AccessibilityMenuListOption.h: * accessibility/AccessibilityMenuListOption.cpp: (WebCore::AccessibilityMenuListOption::nameForMSAA): Just call to stringValue(), which holds from now on that used to be here. (WebCore::AccessibilityMenuListOption::stringValue): New, override of AccessibilityObject::stringValue() to return a proper string. Emit the missing signals when a selection is made. * accessibility/gtk/AXObjectCacheAtk.cpp: (WebCore::AXObjectCache::postPlatformNotification): Emit the usual 'focus' signals when a selection is made over the combo box. * accessibility/gtk/AccessibilityObjectWrapperAtk.cpp: (setAtkStateSetFromCoreObject): Ensure the EXPANDABLE and EXPANDED Atk states are added to the state set when needed. (listObjectForSelection): New, returns the proper list object (the one holding the list of available options as its children) for an specific AtkObject implementing AtkSelection. This is needed because sometimes the selectable options are not directly children of the AtkSelection object (i.e. a combo box has a 'menu' object as its only child of it, holding the list of options as children). (optionFromList): Use listObjectForSelection() to get the actual object holding the list of children as the available options. (optionFromSelection): Add support for combo boxes. (webkit_accessible_selection_add_selection): Ditto. (webkit_accessible_selection_clear_selection): Ditto. (webkit_accessible_selection_get_selection_count): Ditto. (webkit_accessible_selection_is_child_selected): Ditto. (webkit_accessible_selection_remove_selection): Ditto. (webkit_accessible_text_get_text): Makes sure stringValue() is considered to get the result substring when it was already considered when checking the maximum text length for the object. (getInterfaceMaskFromObject): Make sure the AtkSelection interface is implemented for the combo boxes, that the AtkText is implemented for the menu items and that the AtkAction interface is now implemented for every object (WebCore will decide what to do). Avoid a segfault crash when using this from unit tests. * platform/gtk/PopupMenuGtk.cpp: (WebCore::PopupMenuGtk::show): Make sure we got a valid GdkWindow before calling gdk_window_get_origin() over it. 2010-11-30 Andreas Kling Reviewed by Tor Arne Vestbø. [Qt] GraphicsContext: Avoid unnecessarily copying the painter's QTransform Stash the transform in a const reference (QPainter::transform() returns one) instead of making a copy in the shadow drawing parts of fillPath, strokePath and fillRect. * platform/graphics/qt/GraphicsContextQt.cpp: (WebCore::GraphicsContext::fillPath): (WebCore::GraphicsContext::strokePath): (WebCore::GraphicsContext::fillRect): 2010-11-30 Vangelis Kokkevis Reviewed by Simon Fraser. Provide more fine grained control to ports over when to turn on accelerated compositing. https://bugs.webkit.org/show_bug.cgi?id=49998 The RenderLayerCompositor now checks with the chrome client whether 3d transforms, video, plugins, canvas and animation require switching over to the composited path. Layers that overlap composited content will still turn into composited layers regardless of the setting. * css/MediaQueryEvaluator.cpp: (WebCore::transform_3dMediaFeatureEval): * page/ChromeClient.h: (WebCore::ChromeClient::allowedCompositingTriggers): * rendering/RenderLayer.cpp: (WebCore::RenderLayer::canRender3DTransforms): (WebCore::RenderLayer::updateTransform): (WebCore::RenderLayer::currentTransform): * rendering/RenderLayer.h: * rendering/RenderLayerBacking.cpp: (WebCore::RenderLayerBacking::updateLayerTransform): * rendering/RenderLayerCompositor.cpp: (WebCore::RenderLayerCompositor::RenderLayerCompositor): (WebCore::RenderLayerCompositor::cacheAcceleratedCompositingFlags): (WebCore::RenderLayerCompositor::canRender3DTransforms): (WebCore::RenderLayerCompositor::requiresCompositingLayer): (WebCore::RenderLayerCompositor::requiresCompositingForTransform): (WebCore::RenderLayerCompositor::requiresCompositingForVideo): (WebCore::RenderLayerCompositor::requiresCompositingForCanvas): (WebCore::RenderLayerCompositor::requiresCompositingForPlugin): (WebCore::RenderLayerCompositor::requiresCompositingForAnimation): * rendering/RenderLayerCompositor.h: 2010-11-30 Xan Lopez Reviewed by Martin Robinson. [GTK] Plug leak in SoupCache https://bugs.webkit.org/show_bug.cgi?id=50142 Free the cache file contents when we don't need them anymore. * platform/network/soup/cache/webkit/soup-cache.c: (webkit_soup_cache_load): 2010-11-30 Andras Becsi Reviewed by Csaba Osztrogonác. [Qt][V8] Make QtWebKit with V8 build on Linux. https://bugs.webkit.org/show_bug.cgi?id=50244 No new tests needed. * WebCore.pri: add missing idl file * WebCore.pro: enable Linux build * bindings/scripts/CodeGeneratorV8.pm: add missing headers * bindings/v8/ScriptControllerQt.cpp: fix instantiation (WebCore::ScriptController::qtScriptEngine): * bindings/v8/V8GCController.cpp: fix guard (WebCore::V8GCController::checkMemoryUsage): * inspector/CodeGeneratorInspector.pm: add missing header * loader/ResourceLoadScheduler.cpp: ditto * loader/cache/MemoryCache.cpp: ditto * loader/loader.cpp: ditto 2010-11-30 David Hyatt Reviewed by Simon Fraser. https://bugs.webkit.org/show_bug.cgi?id=50183 Code got commented out by accident during render tree refactoring last year. Put the code back in, since it hurts performance to have that line commented out. In order to put this code back in, I had to fix some bugs with preferred width computations and overflow: scroll. Make the pref widths computation create the scrollbar if it doesn't exist yet. Make sure the scrollbar size is included prior to checking for defined width/min-width/max-width values. Otherwise the scrollbar inflates the width beyond the fixed size that was specified. For table cells, there is an additional wrinkle. Make sure not to include the scrollbar size as part of the minimum intrinsic width so that we don't incorrectly grow. * rendering/RenderBlock.cpp: (WebCore::RenderBlock::layoutPositionedObjects): (WebCore::RenderBlock::computePreferredLogicalWidths): * rendering/RenderFlexibleBox.cpp: (WebCore::RenderFlexibleBox::computePreferredLogicalWidths): 2010-11-30 Vitaly Repeshko Unreviewed. Attempt to fix windows debug chromium build. * html/TimeInputType.cpp: Added include providing isfinite. 2010-11-30 Darin Adler Reviewed by Jon Honeycutt. DOM binding code generator scripts import unused File::stat Perl module https://bugs.webkit.org/show_bug.cgi?id=50250 * bindings/scripts/CodeGeneratorCPP.pm: Removed "using File::stat". * bindings/scripts/CodeGeneratorJS.pm: Ditto. * bindings/scripts/CodeGeneratorObjC.pm: Ditto. * bindings/scripts/CodeGeneratorV8.pm: Ditto. 2010-11-30 Pavel Feldman Reviewed by Yury Semikhatsky. Web Inspector: implement "Revert to this revision" in resources panel. https://bugs.webkit.org/show_bug.cgi?id=50107 * inspector/Inspector.idl: * inspector/InspectorCSSAgent.cpp: (WebCore::InspectorCSSAgent::setStyleSheetText2): * inspector/InspectorCSSAgent.h: * inspector/InspectorStyleSheet.cpp: (WebCore::InspectorStyleSheet::reparseStyleSheet): * inspector/front-end/CSSStyleModel.js: (WebInspector.CSSStyleModel.prototype._styleSheetChanged.callback): (WebInspector.CSSStyleModel.prototype._styleSheetChanged): (WebInspector.CSSStyleModel.prototype._onRevert): * inspector/front-end/Resource.js: (WebInspector.Resource.prototype.get contentTimestamp): (WebInspector.Resource.prototype.setInitialContent): (WebInspector.Resource.prototype.isLocallyModified): (WebInspector.Resource.prototype.setContent.else.finished): (WebInspector.Resource.prototype.setContent): (WebInspector.Resource.prototype.revertToThis.callback): (WebInspector.Resource.prototype.revertToThis): (WebInspector.Resource.prototype.get baseRevision): (WebInspector.Resource.prototype._innerRequestContent): * inspector/front-end/ResourcesPanel.js: (WebInspector.ResourcesPanel.prototype._showResourceView): (WebInspector.FrameResourceTreeElement.prototype._contentChanged): (WebInspector.ResourceRevisionTreeElement): (WebInspector.ResourceRevisionTreeElement.prototype.onattach): (WebInspector.ResourceRevisionTreeElement.prototype._ondragstart): (WebInspector.ResourceRevisionTreeElement.prototype._handleContextMenuEvent): * inspector/front-end/ScriptsPanel.js: (WebInspector.ScriptsPanel.prototype.editScriptSource.mycallback): (WebInspector.ScriptsPanel.prototype.editScriptSource): * inspector/front-end/SourceView.js: (WebInspector.SourceView.prototype._editLine): (WebInspector.SourceView.prototype._editLineComplete): (WebInspector.SourceView.prototype._revertEditLine): 2010-11-30 Adam Roben Clean up WebCore.vcproj's bindings filter I added files that we compile (via an all-in-one file) but weren't in the .vcproj, removed files that no longer exist, and added the scripts we use. Fixes WebCore.vcproj's bindings filter is out-of-date Reviewed by Darin Adler. * WebCore.vcproj/WebCore.vcproj: 2010-11-30 Adam Roben Move JSIDB* files to the DerivedSources filter This correctly reflects their location on disk. (They were formerly listed under bindings/js.) * WebCore.vcproj/WebCore.vcproj: 2010-11-30 Xiaomei Ji Reviewed by David Hyatt. Remove clampNegativeToZero in adjustScrollPositionWithinRange. https://bugs.webkit.org/show_bug.cgi?id=50203 * platform/ScrollView.cpp: (WebCore::ScrollView::adjustScrollPositionWithinRange): 2010-11-30 Adam Roben Remove duplicate files from WebCore.vcproj Rubber-stamped by Steve Falkenburg. * WebCore.vcproj/WebCore.vcproj: 2010-11-30 Adam Roben Let VS do its thang with WebCore.vcproj I just removed and readded AnimationBase.cpp. Rubber-stamped by Steve Falkenburg. * WebCore.vcproj/WebCore.vcproj: 2010-11-30 Mario Sanchez Prada Reviewed by Chris Fleizach. ASSERT failing for combo boxes when selection changes https://bugs.webkit.org/show_bug.cgi?id=50062 Don't call listBoxOnChange for elements using menu lists. Test: fast/forms/select-option-accesskey-crash.html * dom/SelectElement.cpp: (WebCore::SelectElement::accessKeySetSelectedIndex): Extra check to decide whether to call menuListOnChange() or listBoxOnChange() instead of always calling listBoxOnChange(). 2010-11-30 Andreas Kling Reviewed by Dirk Schulze. GraphicsContext: Remove "current path" and have strokePath, fillPath and clipPath take a Path argument. https://bugs.webkit.org/show_bug.cgi?id=48516 No new tests, this is refactoring only. fillPath() -> fillPath(const Path&) strokePath() -> strokePath(const Path&) clipPath(WindRule) -> clipPath(const Path&, WindRule) drawPath() was only used by CG, so made it PLATFORM(CG) specific. The Qt, CG and Cairo implementations are best-effort. For the rest, I've put beginPath/addPath and FIXMEs in the three affected methods. * WebCore.exp.in: * html/canvas/CanvasRenderingContext2D.cpp: (WebCore::CanvasRenderingContext2D::fill): (WebCore::CanvasRenderingContext2D::stroke): * inspector/InspectorController.cpp: (WebCore::drawOutlinedQuad): * mathml/RenderMathMLRoot.cpp: (WebCore::RenderMathMLRoot::paint): * mathml/RenderMathMLSquareRoot.cpp: (WebCore::RenderMathMLSquareRoot::paint): * platform/graphics/GraphicsContext.h: * platform/graphics/GraphicsContextPrivate.h: * platform/graphics/cairo/GraphicsContextCairo.cpp: (WebCore::GraphicsContext::fillPath): (WebCore::GraphicsContext::strokePath): (WebCore::GraphicsContext::clipPath): * platform/graphics/cairo/GraphicsContextPlatformPrivateCairo.h: * platform/graphics/cg/GraphicsContextCG.cpp: (WebCore::GraphicsContext::drawEllipse): (WebCore::addConvexPolygonToPath): (WebCore::GraphicsContext::drawConvexPolygon): (WebCore::GraphicsContext::clipConvexPolygon): (WebCore::GraphicsContext::drawPath): (WebCore::GraphicsContext::fillPath): (WebCore::GraphicsContext::strokePath): (WebCore::GraphicsContext::fillRoundedRect): (WebCore::GraphicsContext::clipPath): * platform/graphics/haiku/GraphicsContextHaiku.cpp: (WebCore::GraphicsContext::strokePath): (WebCore::GraphicsContext::fillPath): * platform/graphics/openvg/GraphicsContextOpenVG.cpp: (WebCore::GraphicsContext::fillPath): (WebCore::GraphicsContext::strokePath): (WebCore::GraphicsContext::clipPath): * platform/graphics/qt/GraphicsContextQt.cpp: (WebCore::GraphicsContext::restorePlatformState): (WebCore::GraphicsContext::fillPath): (WebCore::GraphicsContext::strokePath): (WebCore::GraphicsContext::clipPath): (WebCore::GraphicsContext::strokeRect): (WebCore::GraphicsContext::translate): (WebCore::GraphicsContext::rotate): (WebCore::GraphicsContext::scale): (WebCore::GraphicsContext::concatCTM): * platform/graphics/skia/GraphicsContextSkia.cpp: (WebCore::GraphicsContext::clipPath): (WebCore::GraphicsContext::fillPath): (WebCore::GraphicsContext::strokePath): * platform/graphics/wince/GraphicsContextWinCE.cpp: (WebCore::GraphicsContext::fillPath): (WebCore::GraphicsContext::strokePath): (WebCore::GraphicsContext::clipPath): * platform/graphics/wx/GraphicsContextWx.cpp: (WebCore::GraphicsContext::clipPath): (WebCore::GraphicsContext::fillPath): (WebCore::GraphicsContext::strokePath): * rendering/RenderBoxModelObject.cpp: (WebCore::RenderBoxModelObject::paintBorder): (WebCore::RenderBoxModelObject::paintBoxShadow): * rendering/RenderEmbeddedObject.cpp: (WebCore::RenderEmbeddedObject::paintReplaced): * rendering/RenderObject.cpp: (WebCore::RenderObject::drawBoxSideFromPath): * rendering/RenderSVGResource.h: (WebCore::RenderSVGResource::postApplyResource): * rendering/RenderSVGResourceClipper.cpp: (WebCore::RenderSVGResourceClipper::pathOnlyClipping): * rendering/RenderSVGResourceFilter.cpp: (WebCore::RenderSVGResourceFilter::postApplyResource): * rendering/RenderSVGResourceFilter.h: * rendering/RenderSVGResourceGradient.cpp: (WebCore::RenderSVGResourceGradient::applyResource): * rendering/RenderSVGResourceGradient.h: * rendering/RenderSVGResourcePattern.cpp: (WebCore::RenderSVGResourcePattern::postApplyResource): * rendering/RenderSVGResourcePattern.h: * rendering/RenderSVGResourceSolidColor.cpp: (WebCore::RenderSVGResourceSolidColor::postApplyResource): * rendering/RenderSVGResourceSolidColor.h: * rendering/SVGRenderSupport.cpp: (WebCore::SVGRenderSupport::finishRenderSVGContent): * rendering/svg/RenderSVGPath.cpp: (WebCore::RenderSVGPath::fillAndStrokePath): * rendering/svg/SVGInlineTextBox.cpp: (WebCore::SVGInlineTextBox::releasePaintingResource): (WebCore::SVGInlineTextBox::restoreGraphicsContextAfterTextPainting): (WebCore::SVGInlineTextBox::paintDecorationWithStyle): * rendering/svg/SVGInlineTextBox.h: * svg/SVGFont.cpp: (WebCore::Font::drawTextUsingSVGFont): 2010-11-30 Alexey Proskuryakov Reviewed by Darin Adler and Geoff Garen. https://bugs.webkit.org/show_bug.cgi?id=44152 CSSOM should match DOM in discarding wrapper-less parents. We have existing behaviors where parent objects in detached subtrees are not preserved: - if a root of a node tree doesn't have a wrapper, it's immediately destroyed, making its children parentNode attribute null; - relationship between a