1. 16 Jan, 2009 8 commits
    • hyatt@apple.com's avatar
      2009-01-16 David Hyatt <hyatt@apple.com> · aa59fcd9
      hyatt@apple.com authored
              Add a new flag (similar to first-line and first-letter) to Document to track whether any before/after rules
              are being used.  If not, don't waste time probing for those pseudo element styles.
              Reviewed by Sam Weinig
              * css/CSSGrammar.y:
              * dom/Document.cpp:
              * dom/Document.h:
              * rendering/RenderBlock.cpp:
              * rendering/RenderContainer.cpp:
              * rendering/RenderInline.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39992 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • hyatt@apple.com's avatar
      2009-01-16 David Hyatt <hyatt@apple.com> · 97dafec2
      hyatt@apple.com authored
              Fix badness with the cursor CSS property.  For all links and form controls, resetting to a value of auto
              was resulting in the creation of a cursor list.  clearCursorList was creating a new list rather than just
              nulling out the pointer.  Since addCursor already lazily recreates the list, in the case of simple cursor types,
              the clear method was wastefully creating a completely unnecessary list for all of these elements.
              Also optimized the setting of simple cursor values to not alter the RenderStyle's cursor setting if it is already
              set to the same value.  Since this is the common case, it prevents a fault and copy of StyleInheritedData for
              all links and form controls.
              This is a PLT speedup.
              Reviewed by Oliver Hunt
              * css/CSSStyleSelector.cpp:
              * rendering/style/RenderStyle.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39990 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • mitz@apple.com's avatar
      Reviewed by Kevin Decker and Adam Roben. · 073deee0
      mitz@apple.com authored
              - fix a regression that resulted in a crash when plug-ins were
              * rendering/RenderPartObject.cpp:
              (WebCore::RenderPartObject::updateWidget): Added a null check.
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39988 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • hyatt@apple.com's avatar
      2009-01-16 David Hyatt <hyatt@apple.com> · cfa6d61d
      hyatt@apple.com authored
              Split continuation() into virtual and non-virtual versions so that the non-virtual version can
              be used whereever possible.  Eventually the virtual method can be eliminated completely.
              Reviewed by Sam Weinig and Oliver Hunt
              * html/HTMLAnchorElement.cpp:
              * page/AccessibilityRenderObject.cpp:
              * rendering/InlineFlowBox.cpp:
              * rendering/RenderBlock.cpp:
              * rendering/RenderContainer.cpp:
              * rendering/RenderFlow.h:
              * rendering/RenderInline.cpp:
              * rendering/RenderObject.cpp:
              * rendering/RenderObject.h:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39987 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • hyatt@apple.com's avatar
      2009-01-16 David Hyatt <hyatt@apple.com> · 92162d4b
      hyatt@apple.com authored
              Eliminate a bunch of save/restore pairs from commonly used GraphicsContext methods.  The most common reason
              for needing save/restore in the line-drawing methods was antialiasing, so this has now been made part of
              the cross-platform graphics state.  By being able to query for the current antialiasing setting, antialiasing
              can be temporarily changed and then restored back to the real value without needing to do a save/restore.
              The line drawing code also set the fill color to the stroke color and relied on save/restore to prevent
              a permanent mutation of the fill color in the CGContext.   The new code just restores the fill color back
              to its original value without doing the much more expensive save/restore.
              Reviewed by Oliver Hunt
              * platform/graphics/GraphicsContext.cpp:
              * platform/graphics/GraphicsContext.h:
              * platform/graphics/GraphicsContextPrivate.h:
              * platform/graphics/cairo/GraphicsContextCairo.cpp:
              * platform/graphics/cg/GraphicsContextCG.cpp:
              * platform/graphics/qt/GraphicsContextQt.cpp:
              * platform/graphics/skia/GraphicsContextSkia.cpp:
              * platform/graphics/wx/GraphicsContextWx.cpp:
              * rendering/RenderPath.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39981 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • hyatt@apple.com's avatar
      2009-01-16 David Hyatt <hyatt@apple.com> · 4e2bbb7d
      hyatt@apple.com authored
              Optimize RenderText::styleDidChange.  Don't call the base class, which does extra meaningless work.
              Instead just do the one check we need to do (for a diff of layout) and mark dirty as appropriate.
              Clean up the base class method also to just bail early if no m_parent is set instead of repeatedly checking
              for the existence of m_parent in each if clause.
              Reviewed by Darin Adler
              * rendering/RenderObject.cpp:
              * rendering/RenderText.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39979 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • cmarrin@apple.com's avatar
      Fixed https://bugs.webkit.org/show_bug.cgi?id=23088. · 815d5bf3
      cmarrin@apple.com authored
              This was happening because I was calling setChanged() from inside updateRendering()
              which causes an infinite loop. I fixed this by deferring the setChanged to the next
              run loop iteration. That made it not infinite loop, but it still retriggers the
              transition forever. The problem is that there is both an 'all' and specific transition
              on 'opacity'. This tickled a bug in AnimationController which causes the opacity 
              transition to get constantly cancelled and then retriggered. The problem is that
              the specific opacity transition has a duration of 0. I got rid of the logic to 
              flush out 0 duration transitions and it is no longer constantly triggered. The
              logic to flush them was just an optimization, and you really need to keep them
              around to make the logic to override earlier animations by later ones work. And there is
              very little overhead in this case anyway, so the optimization was not that useful.
              I made a LayoutTest from the original testcase which tests both the infinite
              loop and constantly triggering animation cases.
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39974 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • ariya@webkit.org's avatar
      2009-01-16 Ariya Hidayat <ariya.hidayat@trolltech.com> · becca4b4
      ariya@webkit.org authored
              Reviewed by Adam Roben.
              When plugins are disabled, plugin data should not be available for
              the page. This ensures that navigator.plugins would not see the
              additional MIME types supported by the plugins.
              * page/Page.cpp:
              * loader/FrameLoader.cpp:
              * rendering/RenderPartObject.cpp:
      2009-01-16  Ariya Hidayat  <ariya.hidayat@trolltech.com>
              Reviewed by Adam Roben.
              [Qt] Prevent crash if plugin data is NULL.
              * WebCoreSupport/FrameLoaderClientQt.cpp: NULL check.
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39969 268f45cc-cd09-0410-ab3c-d52691b4dbfc
  2. 15 Jan, 2009 1 commit
    • hyatt@apple.com's avatar
      2009-01-15 David Hyatt <hyatt@apple.com> · ef5fbc78
      hyatt@apple.com authored
              Eliminate an extra virtual function call that occurs for every single RenderText by moving the SVG-specific
              code that was placed into RenderText::styleDidChange into RenderSVGInlineText::styleDidChange.
              Reviewed by Oliver Hunt
              * rendering/RenderSVGInlineText.cpp:
              * rendering/RenderSVGInlineText.h:
              * rendering/RenderText.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39953 268f45cc-cd09-0410-ab3c-d52691b4dbfc
  3. 14 Jan, 2009 1 commit
    • adele@apple.com's avatar
      WebCore: · f32f267c
      adele@apple.com authored
      2009-01-14  Adele Peterson  <adele@apple.com>
              Reviewed by Darin Adler.
              Fix for https://bugs.webkit.org/show_bug.cgi?id=23335
              <rdar://problem/6247650> Update <input type="search"> for RenderThemeWin
              No new tests added.  In the future, we need to update DumpRenderTree to be able to run with different RenderThemes and
              we'll need a separate set of results for different themes.
              * css/themeWin.css: Renable search fields.  Added new rules for sub-element placement.
              * rendering/RenderTextControlSingleLine.cpp:
              (WebCore::RenderTextControlSingleLine::textBlockWidth): Consider margin set on the sub-elements when computing the
              desired width for the text block.
              * rendering/RenderThemeWin.cpp:
              (WebCore::RenderThemeWin::supportsFocus): Treat SearchFieldPart the same as TextFieldPart.
              (WebCore::RenderThemeWin::determineState): ditto.
              (WebCore::RenderThemeWin::getClassicThemeData): ditto.
              (WebCore::RenderThemeWin::getThemeData): ditto.
              (WebCore::RenderThemeWin::paintSearchFieldCancelButton): Draw a scaled, vertically centered, version of the new artwork.
              (WebCore::RenderThemeWin::paintSearchFieldResultsDecoration): ditto.
              (WebCore::RenderThemeWin::paintSearchFieldResultsButton): ditto.
              (WebCore::RenderThemeWin::adjustSearchFieldCancelButtonStyle): Set the scaled width and height.
              (WebCore::RenderThemeWin::adjustSearchFieldResultsDecorationStyle): ditto.
              (WebCore::RenderThemeWin::adjustSearchFieldResultsButtonStyle): ditto.
      2009-01-14  Adele Peterson  <adele@apple.com>
              Reviewed by Darin Adler.
              Fix for https://bugs.webkit.org/show_bug.cgi?id=23335
              <rdar://problem/6247650> Update <input type="search"> for RenderThemeWin
              Added artwork.
              * WebKit.vcproj/WebKit.rc:
              * WebKit.vcproj/WebKit.vcproj:
              * WebKit.vcproj/resource.h:
              * WebKit.vcproj/searchCancel.png: Added.
              * WebKit.vcproj/searchCancelPressed.png: Added.
              * WebKit.vcproj/searchMagnifier.png: Added.
              * WebKit.vcproj/searchMagnifierResults.png: Added.
              * WebKitDLL.cpp:(loadResourceIntoBuffer):
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39924 268f45cc-cd09-0410-ab3c-d52691b4dbfc
  4. 13 Jan, 2009 2 commits
    • darin@apple.com's avatar
      2009-01-13 Darin Adler <darin@apple.com> · 5f680147
      darin@apple.com authored
              Reviewed by Sam Weinig.
              Bug 23102: turn on unused parameter warnings in WebCore
              Last step: Turn on the warning and fix the last few instances.
              * Configurations/Base.xcconfig: Removed -Wno-unused-parameter from
              WARNING_CFLAGS_BASE, so the unused parameter warning will fire.
              * bindings/js/JSDOMWindowBase.cpp:
              (jsDOMWindowBaseMessageChannel): Separate the version for use when
              CHANNEL_MESSAGING is turned off, so we can omit the argument names.
              It would be better to not even have this property in that case,
              rather than having the function return undefined, but I don't know
              how to do that with the bindings script.
              * bindings/scripts/CodeGeneratorJS.pm: Leave out argument names in
              functions where they're not used (the SVG context, and the property
              name in the index getter). Use UNUSED_PARAM where arguments are
              sometimes not used.
              * bridge/runtime_object.cpp:
              (JSC::RuntimeObjectImp::put): Use UNUSED_PARAM for non-QT.
              * css/MediaQueryEvaluator.cpp:
              (WebCore::animationMediaFeatureEval): Omit argument names for style and frame.
              (WebCore::transitionMediaFeatureEval): Ditto.
              (WebCore::transform_2dMediaFeatureEval): Ditto.
              (WebCore::transform_3dMediaFeatureEval): Ditto.
              * dom/make_names.pl: Leave out the createdByParser property name in
              cases where it's not used.
              * inspector/JavaScriptDebugServer.cpp:
              (WebCore::JavaScriptDebugServer::setJavaScriptPaused): Separate the
              version for the Mac platform so we can omit the argument names.
              * loader/EmptyClients.h:
              (WebCore::EmptyChromeClient::contentsSizeChanged): Removed argument name.
              * loader/FrameLoaderClient.h:
              (WebCore::FrameLoaderClient::shouldUsePluginDocument): Commented out
              argument name.
              * loader/icon/IconFetcher.cpp:
              (WebCore::IconFetcher::didReceiveResponse): Use ASSERT_UNUSED.
              (WebCore::IconFetcher::didReceiveData): Ditto.
              (WebCore::IconFetcher::didFinishLoading): Ditto.
              (WebCore::IconFetcher::didFail): Ditto.
              * platform/KURL.cpp:
              (WebCore::assertProtocolIsGood): Separate out the NDEBUG version so
              we can omit the argument names.
              * platform/Timer.cpp:
              (WebCore::TimerHeapIterator::checkConsistency): Use ASSERT_UNUSED.
              * platform/graphics/GlyphBuffer.h:
              (WebCore::GlyphBuffer::offsetAt): Use UNUSED_PARAM in the non-WIN case.
              (WebCore::GlyphBuffer::add): Use UNUSED_PARAM in the non-CAIRO case.
              * platform/graphics/cg/ImageSourceCG.cpp:
              (WebCore::ImageSource::clear): Removed argument name.
              * rendering/SVGRenderSupport.cpp:
              (WebCore::prepareToRenderSVGContent): Use UNUSED_PARAM when SVG_FILTERS
              is not enabled.
              (WebCore::finishRenderSVGContent): Ditto.
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39880 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • vestbo@webkit.org's avatar
      2009-01-13 Tor Arne Vestbø <tavestbo@trolltech.com> · d203bbc2
      vestbo@webkit.org authored
              Reviewed by Simon Hausmann.
              Change how themes adjust mediaControls.css to match html4/quicks.css
              Instead of providing the full style sheet, the themes provide extra
              overrides to the default style defined in UserAgentStyleSheetsData.
              Also, merge WebKitResources.qrc and WebCoreResources.qrc to speed up
              build time for the Qt port.
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39854 268f45cc-cd09-0410-ab3c-d52691b4dbfc
  5. 12 Jan, 2009 4 commits
    • simon.fraser@apple.com's avatar
      2009-01-12 Simon Fraser <simon.fraser@apple.com> · 2fdf2727
      simon.fraser@apple.com authored
              Reviewed by Dan Bernstein
              In RenderBox::repaintLayerRectsForImage(), the repaint rect does not have
              to be computed in absolute coordintes. Instead, we compute a repaintRect
              relative to the RenderObject, which can repaint itself.
              * rendering/RenderBox.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39847 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • hyatt@apple.com's avatar
      2009-01-12 David Hyatt <hyatt@apple.com> · b6c79dd4
      hyatt@apple.com authored
              Optimize the retrieval of line height.  If the document is not using any first-line rules, just immediately
              default to the normal case.  Tiny speedup on PLT.
              Reviewed by Oliver Hunt
              * rendering/RenderFlow.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39837 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • simon.fraser@apple.com's avatar
      2009-01-12 Simon Fraser <simon.fraser@apple.com> · 6528b50e
      simon.fraser@apple.com authored
              Reviewed by Dan Bernstein
              Merge layoutDelta into LayoutState. This requires that we
              push even when layoutState is disabled, so that layoutDelta
              can still be stored.
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39834 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • darin@apple.com's avatar
      2009-01-11 Darin Adler <darin@apple.com> · faced26e
      darin@apple.com authored
              Reviewed by Oliver Hunt.
              Bug 23102: turn on unused parameter warnings in WebCore
              Fourth step: Remove arguments entirely in the many cases where they are not
              needed at all. This step is a little less mechanical than the earlier ones.
              * dom/DOMImplementation.cpp:
              (WebCore::DOMImplementation::isXMLMIMEType): Pass TextCaseSensitive when creating
              the regular expression, since we already specify a-zA-Z in our expression.
              * dom/Document.cpp:
              (WebCore::Document::databaseThread): Don't pass in the document when creating
              a DatabaseThread.
              * dom/Element.cpp:
              (WebCore::Element::createRareData): Don't pass the element when creating an
              (WebCore::Element::detach): Don't pass the element when calling resetComputedStyle.
              (WebCore::Element::recalcStyle): Ditto.
              * dom/ElementRareData.h: Remove the unused Element* from the constructor and
              the resetComputedStyle function.
              * dom/EventTargetNode.cpp:
              (WebCore::EventTargetNode::dispatchEvent): Don't pass ec to dispatchGenericEvent.
              (WebCore::EventTargetNode::dispatchGenericEvent): Remove the unused ec argument.
              (WebCore::EventTargetNode::dispatchWindowEvent): Don't pass ec to dispatchGenericEvent.
              * dom/EventTargetNode.h: Remove the ExceptionCode& argument form disaptchGenericEvent.
              * editing/ApplyStyleCommand.cpp: Remove the empty removeBlockStyle function.
              (WebCore::ApplyStyleCommand::pushDownTextDecorationStyleAroundNode): Remove the unused
              start and end arguments.
              (WebCore::ApplyStyleCommand::pushDownTextDecorationStyleAtBoundaries): Don't pass start
              and end to pushDownTextDecorationStyleAroundNode.
              * editing/ApplyStyleCommand.h: Remove the empty removeBlockStyle function. Removed the
              unused start and end arguments from pushDownTextDecorationStyleAroundNode.
              * editing/htmlediting.cpp: Removed include of unused RegularExpression.h header.
              * html/HTMLTokenizer.cpp:
              (WebCore::Token::addAttribute): Removed unused Document* argument.
              (WebCore::HTMLTokenizer::parseTag): Don't pass document to addAttribute.
              * html/HTMLTokenizer.h: Removed unused Document* argument to to addAttribute.
              * page/DOMWindow.cpp:
              (WebCore::DOMWindow::localStorage): Don't pass the frame to storageArea.
              * page/DragController.cpp:
              (WebCore::DragController::dragIsMove): Removed unused DragData argument.
              (WebCore::DragController::performDrag): Call concludeEditDrag rather than
              (WebCore::DragController::tryDocumentDrag): Don't pass dragData to dragIsMove.
              (WebCore::DragController::concludeEditDrag): Renamed to make it clear this
              should only be called for edit drags. Removed the action mask argument. Also
              changed dragIsMove call sites to not pass dragData.
              * page/DragController.h: Updated for above changes and tweaked comment format.
              * page/Frame.cpp:
              (WebCore::createRegExpForLabels): Pass TextCaseSensitive and TextCaseInsensitive
              when creating regular expressions for clarity. Call match instead of search,
              since both do the same thing except for with negative start offsets, and none
              used here are negative.
              (WebCore::Frame::matchLabelsAgainstElement): Ditto.
              * page/Page.h: Moved TextCaseSensitivity from here to StringImpl.h in platform.
              * page/mac/AccessibilityObjectWrapper.mm:
              (AXAttributeStringSetHeadingLevel): Removed unused object argument.
              (AXAttributedStringAppendText): Ditto.
              (-[AccessibilityObjectWrapper doAXAttributedStringForTextMarkerRange:]): Ditto.
              * page/mac/FrameMac.mm:
              (WebCore::regExpForLabels): Pass TextCaseSensitive and TextCaseInsensitive
              when creating regular expressions for clarity. Call match instead of search,
              since both do the same thing except for with negative start offsets, and none
              used here are negative.
              (WebCore::Frame::matchLabelsAgainstElement): Ditto.
              * platform/graphics/SegmentedFontData.cpp:
              (WebCore::SegmentedFontData::containsCharacter): Added.
              (WebCore::SegmentedFontData::containsCharacters): Updated to check all characters
              instead of just the first one, using containsCharacter.
              * platform/graphics/SegmentedFontData.h: Made more members private.
              Added containsCharacter function member.
              * platform/mac/DragImageMac.mm:
              (WebCore::dragImageSize): Changed to use RetainPtr<NSImage> directly instead of
              DragImageRef for clarity, since this is platform-specific code.
              (WebCore::deleteDragImage): Ditto. Also remove argument name to get rid of unused
              variable warning.
              (WebCore::scaleDragImage): Ditto.
              (WebCore::dissolveDragImageToFraction): Ditto.
              (WebCore::createDragImageFromImage): Ditto.
              (WebCore::createDragImageIconForCachedImage): Ditto.
              * platform/mac/WebFontCache.mm:
              (acceptableChoice): Removed the unused weight-related arguments.
              (betterChoice): Ditto.
              (+[WebFontCache internalFontWithFamily:traits:weight:size:]): Ditto.
              * platform/posix/FileSystemPOSIX.cpp: Removed the implementation of listDirectory
              because we'd rather get a link error rather than a runtime error if we use it
              for anything, and right now it's not used.
              * platform/text/RegularExpression.cpp: Removed unused private data,
              including pattern, lastMatchString, lastMatchOffsets, lastMatchCount,
              and lastMatchPos. Made the RegularExpression::Private::regex be private
              and use an accessor function to get at it.
              (WebCore::RegularExpression::RegularExpression): Removed the const char* version
              of the constructor because it's not any more efficient. Changed the boolean
              constructor argument to use TextCaseSensitivity instead to make it less likely
              we'll use it wrong.
              (WebCore::RegularExpression::operator=): Since the only member is a single RefPtr,
              just use plain old assignment on it. The only reason we do this instead of using
              the compiler generated assignment operator is that we want the Private structure
              to be internal to the .cpp file.
              (WebCore::RegularExpression::match): Added a null check so we won't crash if
              we fail to compile the regular expression. Removed the code to handle multiple
              match strings, since that was unused.
              * platform/text/RegularExpression.h: Changed the case sensitivity argument to use
              TextCaseSensitivity and not have a default value. Removed the default constructor,
              and the pattern, match, and pos functions.
              * platform/text/StringImpl.h: Moved the TextCaseSensitivity enum here from
              Page.h, because we should eventually use it in all the places where we have
              an argument to control case sensitivity, starting with the functions in this
              * platform/text/mac/ShapeArabic.c:
              (shapeUnicode): Removed the unused options and pErrorCode arguments. Even though
              this is code borrowed from ICU, this seems like a simple safe change to make.
              (shapeArabic): Don't pass options or pErrorCode in to shapeUnicode.
              * rendering/RenderPath.cpp:
              (WebCore::RenderPath::RenderPath): Removed the unused style argument.
              Also removed the type check. The argument already has the correct type,
              so it's not helpful to upcast it and then check its type again, unless
              we have some reason to believe the type is not reliable. And we don't.
              * rendering/RenderPath.h: Removed the RenderStyle* argument to the
              constructor. Also removed the explicit declaration of the destructor.
              It's not helpful or needed.
              * rendering/RenderSVGInlineText.cpp:
              (WebCore::RenderSVGInlineText::localCaretRect): Updated the comment here to
              make it clear that just returning an empty rect is not a good implementation.
              Removed the argument names to get rid of the warning.
              * rendering/RenderTreeAsText.cpp:
              (WebCore::printBorderStyle): Removed the unused RenderObject argument.
              (WebCore::operator<<): Updated for the change to printBorderStyle function.
              * storage/DatabaseThread.cpp:
              (WebCore::DatabaseThread::DatabaseThread): Removed the unused document argument.
              * storage/DatabaseThread.h: Ditto.
              * storage/LocalStorage.cpp:
              (WebCore::LocalStorage::storageArea): Removed the unused Frame* argument.
              * storage/LocalStorage.h: Ditto.
              * svg/SVGElement.cpp:
              (WebCore::SVGElement::sendSVGLoadEventIfPossible): Don't pass an exception
              code location to dispatchGenericEvent.
              * svg/SVGElementInstance.cpp:
              (WebCore::SVGElementInstance::dispatchEvent): Ditto.
              * svg/SVGStyledTransformableElement.cpp: Removed include of unused RegularExpression.h.
              (WebCore::SVGStyledTransformableElement::createRenderer): Don't pass style in to
              the RenderPath constructor.
              * svg/SVGTransformable.cpp: Removed include of unused RegularExpression.h.
              * xml/XSLTProcessor.cpp:
              (WebCore::createFragmentFromSource): Removed unused sourceNode argument.
              (WebCore::XSLTProcessor::transformToFragment): Don't pass sourceNode to createFragmentFromSource.
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39818 268f45cc-cd09-0410-ab3c-d52691b4dbfc
  6. 11 Jan, 2009 7 commits
  7. 10 Jan, 2009 1 commit
  8. 09 Jan, 2009 1 commit
    • adele@apple.com's avatar
      2009-01-09 Adele Peterson <adele@apple.com> · 524122c2
      adele@apple.com authored
              Reviewed by Darin Adler.
              Fix infinite recursion in clientPaddingLeft and clientPaddingRight.  This changes the code back to
              how it was before RenderTextControlSingleLine was split out from RenderTextControl.
              No test added since this code is only exercised when trying to place a search field's recent searches popup menu.
              * rendering/RenderTextControlSingleLine.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39772 268f45cc-cd09-0410-ab3c-d52691b4dbfc
  9. 08 Jan, 2009 1 commit
    • eric.carlson@apple.com's avatar
      2009-01-08 Eric Carlson <eric.carlson@apple.com> · cae10900
      eric.carlson@apple.com authored
              Reviewed by Adele Peterson.
              Simplify Mac interfaces for drawing media controller elements
              * WebCore.base.exp: update for changed WebKitSystemInterface media controller functions
              * platform/mac/WebCoreSystemInterface.h: Ditto
              * platform/mac/WebCoreSystemInterface.mm: Ditto
              * rendering/MediaControlElements.h:
              (WebCore::): Add MediaControlElements enum
              * rendering/RenderThemeMac.mm:
              (WebCore::RenderThemeMac::paintMediaFullscreenButton):  update for changed WebKitSystemInterface
              media controller functions
              (WebCore::RenderThemeMac::paintMediaMuteButton): Ditto
              (WebCore::RenderThemeMac::paintMediaPlayButton): Ditto
              (WebCore::RenderThemeMac::paintMediaSeekBackButton): Ditto
              (WebCore::RenderThemeMac::paintMediaSeekForwardButton): Ditto
              (WebCore::RenderThemeMac::paintMediaSliderTrack): Ditto
              (WebCore::RenderThemeMac::paintMediaSliderThumb): Ditto
      2009-01-08  Eric Carlson  <eric.carlson@apple.com>
              Reviewed by Adele Peterson.
              Simplify Mac interfaces for drawing media controller elements
              * WebCoreSupport/WebSystemInterface.m:
              (InitWebCoreSystemInterface): Update for changes to media controller functions
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39714 268f45cc-cd09-0410-ab3c-d52691b4dbfc
  10. 07 Jan, 2009 1 commit
  11. 05 Jan, 2009 4 commits
    • simon.fraser@apple.com's avatar
      2009-01-05 Simon Fraser <simon.fraser@apple.com> · 6b8fefe6
      simon.fraser@apple.com authored
              Reviewed by Dave Hyatt
              Add an assertion that clip rects are being used when painting with the same
              rootLayer that they were computed with.
              Fix two issues detected by the assertion:
              RenderLayer::updateClipRects() should not unconditionally update the clip rects
              on its parent, but stop when reaching rootLayer (just like calculateClipRects()).
              We need to pass the temporaryClipRects flag down through reflection painting
              to handle the case of nested reflections.
              Also use temporary clip rects in RenderTreeAsText, since that code does not
              reset the painting root for transformed layers, so cached clip rects will not
              match those used for painting.
              * rendering/RenderLayer.cpp:
              * rendering/RenderLayer.h:
              * rendering/RenderTreeAsText.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39640 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • dino@apple.com's avatar
      2009-01-05 Dean Jackson <dino@apple.com> · 0cf2dc9d
      dino@apple.com authored
              Reviewed by Darin Adler.
              Rename AffineTransform to TransformationMatrix in
              preparation for future enhancements (non-affine matrices)
              * GNUmakefile.am:
              * WebCore.pro:
              * WebCore.scons:
              * WebCore.vcproj/WebCore.vcproj:
              * WebCore.xcodeproj/project.pbxproj:
              * WebCoreSources.bkl:
              * platform/graphics/transforms/AffineTransform.cpp: Removed.
              * platform/graphics/transforms/AffineTransform.h: Removed.
              * platform/graphics/transforms/TransformationMatrix.cpp: Copied from WebCore/platform/graphics/transforms/AffineTransform.cpp.
              * platform/graphics/transforms/TransformationMatrix.h: Copied from WebCore/platform/graphics/transforms/AffineTransform.h.
              * lots of other files with s/AffineTransform/TransformationMatrix/g
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39635 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • dino@apple.com's avatar
      2008-01-05 Dean Jackson <dino@apple.com> · 563bd1f6
      dino@apple.com authored
              Reviewed by David Hyatt.
              Implement 'pointer-events' for HTML content. This involved
              adding a new value 'auto' which behaves as 'visiblePainted'
              in SVG content.
              Moved the property out of the SVG CSS code and into
              the general CSS (both parsing and RenderStyle).
              Changes to the hit testing functionality of the Render tree,
              specifically the nodeAtPoint methods. Where they used to
              test for visibility, they now use a helper function defined
              on base classes (RenderObject and InlineBox) that checks both
              visibility and pointer-events.
              Tests: fast/events/pointer-events-2.html
              * css/CSSComputedStyleDeclaration.cpp:
              * css/CSSParser.cpp:
              * css/CSSPrimitiveValueMappings.h:
              (WebCore::CSSPrimitiveValue::operator EPointerEvents):
              * css/CSSStyleSelector.cpp:
              * css/SVGCSSComputedStyleDeclaration.cpp:
              * css/SVGCSSStyleSelector.cpp:
              * manual-tests/pointer-events.html: Added.
              * rendering/EllipsisBox.cpp:
              * rendering/InlineBox.h:
              * rendering/InlineFlowBox.cpp:
              * rendering/InlineTextBox.cpp:
              * rendering/PointerEventsHitRules.cpp:
              * rendering/PointerEventsHitRules.h:
              * rendering/RenderBlock.cpp:
              * rendering/RenderBox.cpp:
              * rendering/RenderObject.h:
              * rendering/RenderPath.cpp:
              * rendering/RenderSVGImage.cpp:
              * rendering/RenderSVGText.cpp:
              * rendering/RootInlineBox.cpp:
              * rendering/TextControlInnerElements.cpp:
              * rendering/style/RenderStyle.h:
              * rendering/style/RenderStyleConstants.h:
              * rendering/style/SVGRenderStyle.h:
              * rendering/style/SVGRenderStyleDefs.h:
              * css/CSSValueKeywords.in:
              * css/SVGCSSValueKeywords.in:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39634 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • darin@apple.com's avatar
      2009-01-05 Darin Adler <darin@apple.com> · 98a7ac69
      darin@apple.com authored
              Reviewed by Alexey Proskuryakov.
              Bug 23102: turn on unused parameter warnings in WebCore
              First step: Fix the simple cases where we can just remove an argument name.
              Also made a few more things protected and private.
              * many files
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39601 268f45cc-cd09-0410-ab3c-d52691b4dbfc
  12. 04 Jan, 2009 2 commits
  13. 03 Jan, 2009 1 commit
  14. 02 Jan, 2009 3 commits
    • eric.carlson@apple.com's avatar
      2009-01-02 Eric Carlson <eric.carlson@apple.com> · de645cee
      eric.carlson@apple.com authored
              Reviewed by Darin Adler
              Fix for https://bugs.webkit.org/show_bug.cgi?id=23066 & <rdar://problem/6028417> 
              Full Page Zoom: a <video> element that doesn't include width/height attribute does not scale
              * rendering/RenderVideo.cpp:
              (WebCore::RenderVideo::calcReplacedWidth): Multiply aspect ratio width by style()->effectiveZoom()
              (WebCore::RenderVideo::calcReplacedHeight): Multiply aspect ratio height by style()->effectiveZoom()
      2009-01-02  Eric Carlson  <eric.carlson@apple.com>
              Reviewed by Darin Adler
              https://bugs.webkit.org/show_bug.cgi?id=23066 & <rdar://problem/6028417> 
              Full Page Zoom: a <video> element that doesn't include width/height attribute does not scale
              Add test for <video> zoom with and without width/height attributes
              * media/video-zoom.html: Added.
              * platform/mac/media/video-zoom-expected.txt: Added.
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39562 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • ddkilzer@apple.com's avatar
      Bug 23081: These files are no longer part of the KDE libraries · 6242b775
      ddkilzer@apple.com authored
              Reviewed by Darin Adler.
              Removed "This file is part of the KDE libraries" comment from
              source files.  Added or updated Apple copyrights as well.
              * parser/Lexer.h:
              * wtf/HashCountedSet.h:
              * wtf/RetainPtr.h:
              * wtf/VectorTraits.h:
              No tests since there are only changes to comments.
              * bridge/testbindings.cpp:
              * html/HTMLEntityNames.gperf:
              * loader/Cache.h:
              * loader/CachedCSSStyleSheet.h:
              * loader/CachedResourceClient.h:
              * loader/CachedResourceClientWalker.cpp:
              * loader/CachedResourceClientWalker.h:
              * loader/CachedScript.cpp:
              * loader/CachedScript.h:
              * loader/CachedXBLDocument.cpp:
              * loader/CachedXBLDocument.h:
              * loader/CachedXSLStyleSheet.cpp:
              * loader/CachedXSLStyleSheet.h:
              * rendering/GapRects.h:
              * xml/XMLSerializer.cpp:
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39556 268f45cc-cd09-0410-ab3c-d52691b4dbfc
    • simon.fraser@apple.com's avatar
      2009-01-02 Simon Fraser <simon.fraser@apple.com> · f7d140d5
      simon.fraser@apple.com authored
              Reviewed by Darin Adler
              Repaints inside of an SVGImage were thrown on the floor,
              which broke incremental painting due to loading, or SVG animation.
              Fix this by assigning the SVGImage a ChromeClient subclass that passes
              along repaints via a new method on ImageObserver, which also takes
              a rect parameter for the changed rect, allowing incremental repaints.
              Fix RenderImage::imageChanged to take advantage of this changedRect to
              only repaint the changed parts of the image.
              This also enables incremental painting for canvas-as-image, so
              add tests for that too.
              Tests: fast/backgrounds/animated-svg-as-background.html
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@39555 268f45cc-cd09-0410-ab3c-d52691b4dbfc
  15. 30 Dec, 2008 1 commit
  16. 28 Dec, 2008 2 commits