1. 07 Feb, 2012 40 commits
    • tony@chromium.org's avatar
      merge DashboardSupportCSSPropertyNames.in into CSSPropertyNames.in · a769aebd
      tony@chromium.org authored
      https://bugs.webkit.org/show_bug.cgi?id=78036
      
      Reviewed by Darin Adler.
      
      Source/JavaScriptCore:
      
      * Configurations/FeatureDefines.xcconfig: Add ENABLE_DASHBOARD_SUPPORT to FEATURE_DEFINES.
      
      Source/WebCore:
      
      In r89362, we started running the preprocessor through CSSPropertyNames.in.
      Now we can move DashboardSupportCSSPropertyNames.in into CSSPropertyNames.in
      and wrap it in an #if.
      
      No new tests, build refactoring.
      
      * Configurations/FeatureDefines.xcconfig: Add ENABLE_DASHBOARD_SUPPORT to FEATURE_DEFINES.
      * DerivedSources.make: Remove DashboardSupportCSSPropertyNames.in.
      * DerivedSources.pri: Remove DashboardSupportCSSPropertyNames.in.
      * WebCore.xcodeproj/project.pbxproj: Remove DashboardSupportCSSPropertyNames.in.
      * css/CSSPropertyNames.in: Wrap -webkit-dashboard-region in an #if.
      * css/DashboardSupportCSSPropertyNames.in: Removed.
      
      Source/WebKit/mac:
      
      * Configurations/FeatureDefines.xcconfig: Add ENABLE_DASHBOARD_SUPPORT to FEATURE_DEFINES.
      
      Source/WebKit2:
      
      * Configurations/FeatureDefines.xcconfig: Add ENABLE_DASHBOARD_SUPPORT to FEATURE_DEFINES.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107026 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      a769aebd
    • commit-queue@webkit.org's avatar
      Enable IPP for FFTFrame · afa83730
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=75522
      
      Patch by Xingnan Wang <xingnan.wang@intel.com> on 2012-02-07
      Reviewed by Tony Chang.
      
      Add the FFTFrame implementation using Intel IPP's DFT algorithm.
      
      * WebCore.gyp/WebCore.gyp:
      * WebCore.gypi:
      * platform/audio/FFTFrame.h:
      (FFTFrame):
      * platform/audio/FFTFrameStub.cpp:
      * platform/audio/ipp/FFTFrameIPP.cpp: Added.
      (WebCore):
      (WebCore::FFTFrame::FFTFrame):
      (WebCore::FFTFrame::initialize):
      (WebCore::FFTFrame::cleanup):
      (WebCore::FFTFrame::~FFTFrame):
      (WebCore::FFTFrame::multiply):
      (WebCore::FFTFrame::doFFT):
      (WebCore::FFTFrame::doInverseFFT):
      (WebCore::FFTFrame::realData):
      (WebCore::FFTFrame::imagData):
      (WebCore::FFTFrame::getUpToDateComplexData):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107025 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      afa83730
    • enne@google.com's avatar
      Properly detect top level frames when propogating compositing · 3cafe50a
      enne@google.com authored
      https://bugs.webkit.org/show_bug.cgi?id=78033
      
      Reviewed by James Robinson.
      
      There's no need to enumerate all tag names when searching for a
      top-level frame. If a render view's document has a frame, then that
      frame is not the top-level one.
      
      * rendering/RenderLayerCompositor.cpp:
      (WebCore::RenderLayerCompositor::shouldPropagateCompositingToEnclosingFrame):
      * rendering/RenderLayerCompositor.h:
      (RenderLayerCompositor):
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107024 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      3cafe50a
    • commit-queue@webkit.org's avatar
      Resolve crash in FrameLoader::checkTimerFired. · 667ee8ba
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=77907
      
      Patch by Chris Palmer <palmer@google.com> on 2012-02-07
      Reviewed by Eric Seidel.
      
      Source/WebCore:
      
      Test is LayoutTests/http/tests/appcache/deferred-events-delete-while-raising-timer.html.
      
      * loader/FrameLoader.cpp:
      (WebCore::FrameLoader::checkTimerFired):
      
      LayoutTests:
      
      * http/tests/appcache/deferred-events-delete-while-raising-timer-expected.txt: Added.
      * http/tests/appcache/deferred-events-delete-while-raising-timer.html: Added.
      * http/tests/appcache/resources/deferred-events-delete-while-raising-timer-1.html: Added.
      * http/tests/appcache/resources/deferred-events-delete-while-raising-timer-2.html: Added.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107023 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      667ee8ba
    • commit-queue@webkit.org's avatar
      [Chromium] add option for 2d canvas defered rendering to DumpRenderTree · 51f0d5fe
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=78039
      
      Patch by Justin Novosad <junov@chromium.org> on 2012-02-07
      Reviewed by Stephen White.
      
      Adding the --enable-deferred-2d-canvas switch to DumpRenderTree
      
      * DumpRenderTree/chromium/DumpRenderTree.cpp:
      (main):
      * DumpRenderTree/chromium/TestShell.cpp:
      (TestShell::TestShell):
      (TestShell::resetWebSettings):
      (TestShell::runFileTest):
      * DumpRenderTree/chromium/TestShell.h:
      (TestShell::setDeferred2dCanvasEnabled):
      (TestShell):
      * DumpRenderTree/chromium/WebPreferences.cpp:
      (WebPreferences::reset):
      (WebPreferences::applyTo):
      * DumpRenderTree/chromium/WebPreferences.h:
      (WebPreferences):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107022 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      51f0d5fe
    • commit-queue@webkit.org's avatar
      [BlackBerry] NetworkJob should stop redirecting when the request is cleared by client · dc51972d
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=78029
      
      Patch by Yong Li <yoli@rim.com> on 2012-02-07
      Reviewed by Rob Buis.
      
      When a redirect is rejected by security origin check, the ResourceRequest
      will be cleared (see DocumentThreadableLoader::redirectReceived()). In this
      case, we should stop handling the request.
      
      No new tests because existing tests (like http:/tests/xmlhttprequest/redirect-cross
      -origin-tripmine.html) can cover this.
      
      * platform/network/blackberry/NetworkJob.cpp:
      (WebCore::NetworkJob::startNewJobWithRequest):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107021 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      dc51972d
    • gyuyoung.kim@samsung.com's avatar
      [CMAKE] Use *bin* and *lib* directories for executable and libraries. · 42a03531
      gyuyoung.kim@samsung.com authored
      https://bugs.webkit.org/show_bug.cgi?id=77928
      
      Reviewed by Daniel Bates.
      
      CMake has used *Programs* directory for executable. In addition, shared libraries are being
      built in source directory. It is better to set common places in order to maintain executable
      and libraries. *bin* is for executable and *lib* is for library.
      
      .:
      
      * CMakeLists.txt: Set library output path.
      * Source/cmake/WebKitFS.cmake: Remove *Programs* creation.
      
      Source/JavaScriptCore:
      
      * shell/CMakeLists.txt: Change *Programs* with *bin*.
      
      Source/WebKit2:
      
      * CMakeLists.txt:
      
      Tools:
      
      * DumpRenderTree/efl/CMakeLists.txt:
      * EWebLauncher/CMakeLists.txt:
      * Scripts/run-launcher:
      * Scripts/webkitdirs.pm:
      (jscProductDir):
      * Scripts/webkitpy/layout_tests/port/efl.py:
      (EflPort._path_to_driver):
      (EflPort._path_to_image_diff):
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107020 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      42a03531
    • commit-queue@webkit.org's avatar
      Remove extraneous MathML code before bug 52444 fix · 7190c446
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=78034
      
      Source/WebCore:
      
      Patch by David Barton <dbarton@mathscribe.com> on 2012-02-07
      Reviewed by Eric Seidel.
      
      Per Darin Adler, I am breaking up the patch fixing bug 52444 into smaller pieces.
      This patch removes a couple unused functions, some extra blank lines, unused #include
      directives, etc., and adds a very few WebKit-standard changes to these files.
      
      No new tests.
      
      * rendering/mathml/RenderMathMLBlock.cpp:
      (WebCore):
      * rendering/mathml/RenderMathMLBlock.h:
      (WebCore::RenderMathMLBlock::getBoxModelObjectHeight):
          - changed to a static member function since 'this' is unused;
            removed redundant non-const version
      (WebCore::RenderMathMLBlock::getBoxModelObjectWidth):
          - changed to a static member function since 'this' is unused;
            removed redundant non-const version
      (WebCore):
      * rendering/mathml/RenderMathMLFraction.cpp:
      * rendering/mathml/RenderMathMLMath.cpp:
      * rendering/mathml/RenderMathMLMath.h:
      * rendering/mathml/RenderMathMLOperator.h:
      (WebCore):
      * rendering/mathml/RenderMathMLRoot.cpp:
      * rendering/mathml/RenderMathMLRow.cpp:
      * rendering/mathml/RenderMathMLRow.h:
      * rendering/mathml/RenderMathMLSquareRoot.cpp:
      * rendering/mathml/RenderMathMLSquareRoot.h:
      * rendering/mathml/RenderMathMLSubSup.cpp:
      * rendering/mathml/RenderMathMLSubSup.h:
      * rendering/mathml/RenderMathMLUnderOver.cpp:
      * rendering/mathml/RenderMathMLUnderOver.h:
      
      LayoutTests:
      
      Patch by David Barton <dbarton@mathscribe.com> on 2012-02-07
      Reviewed by Eric Seidel.
      
      The rendering of the scrollbar in row-expected.png was flaky, presumably due to a test
      environment timing issue, so I removed it.
      
      * mathml/presentation/row.xhtml:
      * platform/mac/mathml/presentation/row-expected.png:
      * platform/mac/mathml/presentation/row-expected.txt:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107019 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      7190c446
    • commit-queue@webkit.org's avatar
      Override expected-failure... · 20bf2a9a
      commit-queue@webkit.org authored
      Override expected-failure mac-leopard/media/media-blocked-by-beforeload-expected.txt with copy of expected-success expectation, for chromium-mac-leopard.
      https://bugs.webkit.org/show_bug.cgi?id=78030
      
      Unreviewed.
      
      Patch by Ami Fischman <fischman@chromium.org> on 2012-02-07
      
      * platform/chromium-mac-leopard/media/media-blocked-by-beforeload-expected.txt: Added.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107018 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      20bf2a9a
    • timothy@apple.com's avatar
      Initilize the WebPageProxy intrinsicDeviceScaleFactor on creation of the WKView. · bdf0adba
      timothy@apple.com authored
      If WKView was created and loaded a page with canvas elements before beinging moved
      to a window, the canvas elements would be locked in at a pixel ratio of 1. With this
      change the WKView will use the main screen's pixel ratio until it moves to a window.
      
      Reviewed by Anders Carlsson.
      
      * UIProcess/API/mac/WKView.mm:
      (-[WKView initWithFrame:contextRef:pageGroupRef:]): Call setIntrinsicDeviceScaleFactor.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107017 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      bdf0adba
    • commit-queue@webkit.org's avatar
      Get rid of Source/autotools/webkit.m4 · cf419a46
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=77833
      
      Patch by Priit Laes <plaes@plaes.org> on 2012-02-07
      Reviewed by Martin Robinson.
      
      Move checks from webkit.m4 to configure.ac and modernize macros:
        . Check for visibility flags after CXX has been defined
        . Use PKG_PROG_PKG_CONFIG for pkg-config lookup
        . Drop unused LIBFFTW_REQUIRED_VERSION
        . Update ICU libs for mingw (bug 77837).
        . Require Pango 1.21.0 (as per unicode backend)
        . Bump minimum glib version due to g_thread_init changes
      
      * Source/autotools/webkit.m4: Removed.
      * configure.ac:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107016 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      cf419a46
    • commit-queue@webkit.org's avatar
      [Chromium] REGRESSION(r101854): Causing large amounts of unnecessary repainting. · 74b0108a
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=78020
      
      Patch by David Reveman <reveman@chromium.org> on 2012-02-07
      Reviewed by James Robinson.
      
      Source/WebCore:
      
      Revert r101854.
      
      This patch is tested by the following unit test:
      - TextureManagerTest.requestTextureExceedingPreferredLimit
      
      * platform/graphics/chromium/ManagedTexture.cpp:
      (WebCore::ManagedTexture::reserve):
      * platform/graphics/chromium/TextureManager.cpp:
      (WebCore::TextureManager::requestTexture):
      * platform/graphics/chromium/TextureManager.h:
      (TextureManager):
      
      Source/WebKit/chromium:
      
      Update TextureManagerTest.requestTextureExceedingPreferredLimit test.
      
      * tests/TextureManagerTest.cpp:
      (WTF::TextureManagerTest::requestTexture):
      (WTF::TEST_F):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107014 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      74b0108a
    • andersca@apple.com's avatar
      Use the non-fast-scrollable region to detect when we can't do fast scrolling · 858c9c88
      andersca@apple.com authored
      https://bugs.webkit.org/show_bug.cgi?id=78056
      <rdar://problem/10247932>
      
      Reviewed by Sam Weinig.
      
      * page/scrolling/ScrollingCoordinator.cpp:
      (WebCore::ScrollingCoordinator::frameViewLayoutUpdated):
      Actually set the non-fast scrollable region on the scrolling tree state.
      
      * page/scrolling/ScrollingTree.cpp:
      (WebCore::ScrollingTree::tryToHandleWheelEvent):
      Check if the wheel event's position is inside the non-fast-scrollable region
      and return false if it is.
      
      (WebCore::ScrollingTree::updateMainFrameScrollPosition):
      Store the cached main frame scroll position so we can use it in tryToHandleWheelEvent.
      
      * platform/graphics/Region.cpp:
      * platform/graphics/Region.h:
      Add a simple contains(const IntPoint&) member function.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107013 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      858c9c88
    • mitz@apple.com's avatar
      <rdar://problem/10475450> Synthetic bold is illegible under some scaling transforms · 83ff1e35
      mitz@apple.com authored
      https://bugs.webkit.org/show_bug.cgi?id=78044
      
      Reviewed by Beth Dakin.
      
      Source/WebCore: 
      
      Tests: fast/text/synthetic-bold-transformed-expected.html
             fast/text/synthetic-bold-transformed.html
      
      * platform/graphics/mac/FontMac.mm:
      (WebCore::Font::drawGlyphs): Changed to interpret syntheticBoldOffset as a length in device pixels.
      
      LayoutTests: 
      
      * fast/text/synthetic-bold-transformed-expected.html: Added.
      * fast/text/synthetic-bold-transformed.html: Added.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107012 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      83ff1e35
    • commit-queue@webkit.org's avatar
      [EFL] Add missing initialization of editable variable. · 41b8dda2
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=77948
      
      Patch by Krzysztof Czech <k.czech@samsung.com> on 2012-02-07
      Reviewed by Eric Seidel.
      
      The value of editable in Ewk_Frame_Smart_Data struct is not set.
      
      * ewk/ewk_frame.cpp:
      (ewk_frame_editable_set):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107010 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      41b8dda2
    • leviw@chromium.org's avatar
      Update LayoutUnit usage in ColumnInfo and RenderFrameSet · fd3a383b
      leviw@chromium.org authored
      https://bugs.webkit.org/show_bug.cgi?id=77914
      
      Reviewed by Eric Seidel.
      
      Updating ColumnInfo and RenderFrameSet to use LayoutUnits
      instead of directly referencing integers for locations and
      distances.
      
      No new tests. No changed behavior.
      
      * rendering/ColumnInfo.h:
      (WebCore::ColumnInfo::forcedBreakOffset):
      (WebCore::ColumnInfo::maximumDistanceBetweenForcedBreaks):
      (ColumnInfo):
      * rendering/RenderFrameSet.cpp:
      (WebCore::RenderFrameSet::paintColumnBorder):
      (WebCore::RenderFrameSet::paintRowBorder):
      * rendering/RenderFrameSet.h:
      (RenderFrameSet):
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107009 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      fd3a383b
    • adamk@chromium.org's avatar
      Add JSC support for delivering mutations when the outermost script context exits · 82abc774
      adamk@chromium.org authored
      https://bugs.webkit.org/show_bug.cgi?id=70289
      
      Reviewed by Eric Seidel.
      
      Source/WebCore:
      
      The meat of this change is in JSMainThreadExecState, where a counter
      is incremented every time WebCore calls into JSC and decremented every
      time it returns. When the counter reaches zero, any pending mutations
      are delivered (this mirrors very similar code in V8Proxy and V8RecursionScope).
      
      The rest of the changes are of two sorts: compilation/logic fixes for
      JSC code when ENABLE(MUTATION_OBSERVERS) is true, and additional
      usages of JSMainThreadExecState so as to trigger the above
      increment/decrements at the appropriate times.
      
      * bindings/js/JSCustomXPathNSResolver.cpp:
      (WebCore::JSCustomXPathNSResolver::lookupNamespaceURI):
      Use JSMainThreadExecState instead of JSC::call.
      * bindings/js/JSDictionary.cpp:
      (WebCore::JSDictionary::convertValue): Add support
      for tryGetProperty with a HashMap<AtomicString>.
      * bindings/js/JSDictionary.h:
      * bindings/js/JSErrorHandler.cpp:
      (WebCore::JSErrorHandler::handleEvent):
      Use JSMainThreadExecState instead of JSC::call.
      * bindings/js/JSHTMLDocumentCustom.cpp:
      (WebCore::JSHTMLDocument::open):
      Use JSMainThreadExecState instead of JSC::call.
      * bindings/js/JSMainThreadExecState.cpp:
      (WebCore::JSMainThreadExecState::didLeaveScriptContext):
      * bindings/js/JSMainThreadExecState.h:
      (WebCore::JSMainThreadExecState::JSMainThreadExecState):
      Increment a static recursion level counter.
      (WebCore::JSMainThreadExecState::~JSMainThreadExecState):
      Decrement a static recursion level counter and, if we are
      at zero (the outermost script invocation), deliver any
      outstanding mutation records.
      * bindings/js/JSNodeFilterCondition.cpp:
      (WebCore::JSNodeFilterCondition::acceptNode):
      Use JSMainThreadExecState instead of JSC::call.
      * bindings/js/JSWebKitMutationObserverCustom.cpp:
      (WebCore::JSWebKitMutationObserver::observe):
      Fix JSDictionary logic, add support for attributeFilter.
      
      Tools:
      
      * DumpRenderTree/mac/EventSendingController.mm: Add support for
      eventSender.scheduleAsynchronousKeyDown.
      (+[EventSendingController isSelectorExcludedFromWebScript:]):
      (+[EventSendingController webScriptNameForSelector:]):
      (-[EventSendingController keyDownWrapper:withModifiers:withLocation:]):
      (-[EventSendingController scheduleAsynchronousKeyDown:withModifiers:withLocation:]):
      * Scripts/build-webkit: Properly alphabetize --mutation-observers in the --help output.
      
      LayoutTests:
      
      With the various fixes in this change, 8/10 tests in fast/mutation
      pass under WebKit/Mac. Of the failing tests, only one is due to
      a deficiency in the Mac port's code (end-of-task-delivery.html);
      the other is due to lack of support for a feature (FILE_SYSTEM)
      exercised by the test (non-event-delivery.html).
      
      * fast/mutation/non-event-delivery.html: Made it fail fast if FileSystem support isn't available.
      * fast/mutation/observe-attributes.html: Fixed calls to removeEventListener.
      * fast/mutation/observe-characterdata.html: ditto.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107008 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      82abc774
    • andersca@apple.com's avatar
      Fix build. · e1df2ad7
      andersca@apple.com authored
      * platform/ScrollableArea.h:
      (WebCore::ScrollableArea::scrollableAreaBoundingBox):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107006 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      e1df2ad7
    • leviw@chromium.org's avatar
      [SVG] Use element disappears after scripted change · c0c54ad1
      leviw@chromium.org authored
      https://bugs.webkit.org/show_bug.cgi?id=74392
      
      Reviewed by Eric Seidel.
      
      Source/WebCore:
      
      Solution uncovered by Nikolas Zimmermann. Removing an early return that caused
      SVGUseElements to not update the shadow root's style, and therefor not render
      correctly.
      
      Test: svg/custom/use-disappears-after-style-update.svg
      
      * svg/SVGUseElement.cpp:
      (WebCore::SVGUseElement::didRecalcStyle):
      
      LayoutTests:
      
      * platform/mac-snowleopard/svg/custom/use-disappears-after-style-update-expected.png: Added.
      * platform/mac-snowleopard/svg/custom/use-disappears-after-style-update-expected.txt: Added.
      * svg/custom/use-disappears-after-style-update.svg: Added.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107005 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c0c54ad1
    • haraken@chromium.org's avatar
      Rename [v8OnProto] IDL attribute to [V8OnProto] IDL attribute · 76f68c9e
      haraken@chromium.org authored
      https://bugs.webkit.org/show_bug.cgi?id=77973
      
      Reviewed by Adam Barth.
      
      This patch renames [v8OnProto] to [V8OnProto], since V8 specific IDL
      attributes should be prefixed by "V8".
      
      No tests. No change in behavior.
      
      * bindings/scripts/CodeGeneratorV8.pm:
      (GenerateNormalAttrGetter):
      (GenerateNormalAttrSetter):
      (GenerateSingleBatchedAttribute):
      (GenerateImplementation):
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107004 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      76f68c9e
    • jer.noble@apple.com's avatar
      media/audio-data-url.html test broken on Lion · 0474d5cf
      jer.noble@apple.com authored
      https://bugs.webkit.org/show_bug.cgi?id=69779
      
      Reviewed by Eric Carlson.
      
      Do not use "OpenForPlayback" attribute on data:// urls, as CoreMedia/QuickTime X cannot
      handle those URLs.
      
      * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
      (WebCore::MediaPlayerPrivateQTKit::createQTMovie):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107002 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      0474d5cf
    • andersca@apple.com's avatar
      Scrolling tree should keep track of region we can't do fast scrolling for · 6c3da667
      andersca@apple.com authored
      https://bugs.webkit.org/show_bug.cgi?id=78050
      
      Reviewed by Dan Bernstein.
      
      We currently won't do fast scrolling for subframes and other types of scrollable areas.
      Because of this, we'll have the scrolling tree keep a region of the page for which we can't
      do fast scrolling. This region will be updated after layout.
      
      * page/FrameView.cpp:
      (WebCore::FrameView::scrollableAreaBoundingBox):
      Return the bounding box.
      
      * page/scrolling/ScrollingCoordinator.cpp:
      (WebCore::ScrollingCoordinator::frameViewLayoutUpdated):
      Go through all the scrollable areas in this frame view and compute the region which we can't do
      fast scrolling for.
      
      * page/scrolling/ScrollingTree.cpp:
      (WebCore::ScrollingTree::commitNewTreeState):
      Update the non-fast-scrollable region.
      
      * page/scrolling/ScrollingTreeState.cpp:
      (WebCore::ScrollingTreeState::setNonFastScrollableRegion):
      Set the non-fast-scrollable region if it's changed.
      
      * platform/ScrollableArea.h:
      Add scrollableAreaBoundingBox member function.
      
      * rendering/RenderLayer.cpp:
      (WebCore::RenderLayer::scrollableAreaBoundingBox):
      Return the bounding box.
      
      * rendering/RenderListBox.cpp:
      (WebCore::RenderListBox::scrollableAreaBoundingBox):
      Return the bounding box.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107001 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      6c3da667
    • leviw@chromium.org's avatar
      unicode-bidi:plaintext is supposed to be effective on display:inline elements too · e7812f3d
      leviw@chromium.org authored
      https://bugs.webkit.org/show_bug.cgi?id=73310
      
      Reviewed by Eric Seidel.
      
      Source/WebCore:
      
      Adding support for unicode-bidi: plaintext as a property on inlines. These are treated
      like unicode-bidi: isolate with the addition of their directionality being determined
      by the UBA.
      
      Tests: fast/text/international/inline-plaintext-is-isolated-expected.html
             fast/text/international/inline-plaintext-is-isolated.html
             fast/text/international/inline-plaintext-relayout-with-leading-neutrals-expected.html
             fast/text/international/inline-plaintext-relayout-with-leading-neutrals.html
             fast/text/international/inline-plaintext-with-generated-content-expected.html
             fast/text/international/inline-plaintext-with-generated-content.html
      
      * platform/text/UnicodeBidi.h:
      (WebCore::isIsolated): Added this convenience function as Plaintext and Isolate Unicode-Bidi values
      are both treated as isolated content.
      * rendering/InlineIterator.h:
      (WebCore::notifyObserverEnteredObject): Inline now supports Unicode-Bidi Plaintext.
      (WebCore::notifyObserverWillExitObject): Ditto.
      (WebCore::bidiFirstSkippingEmptyInlines): Changed to support being called without a resolver.
      (WebCore::isIsolatedInline): Inline now supports Unicode-Bidi: Plaintext.
      * rendering/RenderBlockLineLayout.cpp:
      (WebCore::determineDirectionality): Generalized for inlines.
      (WebCore::constructBidiRuns): Added support for Unicode-Bidi: Plaintext as an isolated inline.
      (WebCore::RenderBlock::layoutRunsAndFloatsInRange): Fixed comment.
      (WebCore::RenderBlock::determineStartPosition): Fixed comment and switched to updated
      bidiFirstSkippingEmptyInlines.
      
      LayoutTests:
      
      Ref tests for unicode-bidi: plaintext on inlines.
      
      * fast/text/international/inline-plaintext-is-isolated-expected.html: Added.
      * fast/text/international/inline-plaintext-is-isolated.html: Added.
      * fast/text/international/inline-plaintext-relayout-with-leading-neutrals-expected.html: Added.
      * fast/text/international/inline-plaintext-relayout-with-leading-neutrals.html: Added.
      * fast/text/international/inline-plaintext-with-generated-content-expected.html: Added.
      * fast/text/international/inline-plaintext-with-generated-content.html: Added.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@107000 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      e7812f3d
    • barraclough@apple.com's avatar
      Crash on http://www.rickshawbags.com/ · cce352d7
      barraclough@apple.com authored
      https://bugs.webkit.org/show_bug.cgi?id=78045
      
      Reviewed by Darin Adler.
      
      Problem URL is: http://www.rickshawbags.com/customize/custom-bag#!thl=rickshaw/bag()
              
      This is a bug introduced by https://bugs.webkit.org/show_bug.cgi?id=71933,
      isVariableObject() checks were excluding StaticScopeObjects, this patch
      inadvertently changed them to be included.
      
      * runtime/JSType.h:
          - sort JSType enum such that StaticScopeObjectType comes before VariableObjectType,
            and thus is excluded from isVariableObject() checks.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106999 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      cce352d7
    • haraken@chromium.org's avatar
      [Refactoring] Use the [IsWorkerContext] IDL in CodeGeneratorV8.pm · a1ad8b75
      haraken@chromium.org authored
      https://bugs.webkit.org/show_bug.cgi?id=77957
      
      Reviewed by Adam Barth.
      
      This patch replaces IsSubType("WorkerContext") and something equivalent that
      with the [IsWorkerContext] IDL.
      
      No tests. No change in behavior.
      
      * bindings/scripts/CodeGeneratorV8.pm:
      (GetInternalFields):
      (GenerateConstructorGetter):
      (GenerateImplementation):
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106998 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      a1ad8b75
    • commit-queue@webkit.org's avatar
      [Chromium] Crash when using per-tile painting on Windows. · be4980b5
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=75715
      
      Patch by David Reveman <reveman@chromium.org> on 2012-02-07
      Reviewed by James Robinson.
      
      PlatformCanvas constructor on win32 expects forth argument to be a
      shared section handle. Passing a pointer to a system memory causes
      it to crash. Fix this by not using the PlatformCanvas API for
      SkCanvas construction in per-tile texture uploader.
      
      Tested with manual tests.
      
      * platform/graphics/chromium/BitmapSkPictureCanvasLayerTextureUpdater.cpp:
      (WebCore::BitmapSkPictureCanvasLayerTextureUpdater::Texture::prepareRect):
      (WebCore::BitmapSkPictureCanvasLayerTextureUpdater::Texture::updateRect):
      * platform/graphics/chromium/BitmapSkPictureCanvasLayerTextureUpdater.h:
      (Texture):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106997 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      be4980b5
    • jer.noble@apple.com's avatar
      Unreviewed build fix; make OSStatus the explicit return type for CMTimebase functions. · bd3b4ab0
      jer.noble@apple.com authored
      * platform/mac/PlatformClockCM.mm:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106996 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      bd3b4ab0
    • jchaffraix@webkit.org's avatar
      Unreviewed gardening after r106982. · 0492744b
      jchaffraix@webkit.org authored
      * platform/chromium/test_expectations.txt:
      We don't have at least one of the layoutTestController
      functions used in this test so we are receiving an
      exception and timing out.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106995 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      0492744b
    • commit-queue@webkit.org's avatar
      [chromium] Unit test for bug #77996 · 25c8ef6e
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=78019
      
      Patch by Dana Jansens <danakj@chromium.org> on 2012-02-07
      Reviewed by James Robinson.
      
      * tests/CCLayerTreeHostTest.cpp:
      (CCLayerTreeHostTestManySurfaces):
      (WTF::CCLayerTreeHostTestManySurfaces::CCLayerTreeHostTestManySurfaces):
      (WTF::CCLayerTreeHostTestManySurfaces::beginTest):
      (WTF::CCLayerTreeHostTestManySurfaces::afterTest):
      (WTF):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106994 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      25c8ef6e
    • commit-queue@webkit.org's avatar
      [BlackBerry] Upstream DumpRenderTreeSupport · cf709dde
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=77991
      
      Patch by Rob Buis <rbuis@rim.com> on 2012-02-07
      Reviewed by Antonio Gomes.
      
      No functionalities changed, no new tests.
      
      * blackberry/WebKitSupport/DumpRenderTreeSupport.cpp: Added.
      (toGeolocationClientMock):
      (DumpRenderTreeSupport::DumpRenderTreeSupport):
      (DumpRenderTreeSupport::~DumpRenderTreeSupport):
      (DumpRenderTreeSupport::javaScriptObjectsCount):
      (DumpRenderTreeSupport::garbageCollectorCollect):
      (DumpRenderTreeSupport::garbageCollectorCollectOnAlternateThread):
      (DumpRenderTreeSupport::setLinksIncludedInFocusChain):
      (DumpRenderTreeSupport::linksIncludedInFocusChain):
      (DumpRenderTreeSupport::dumpConfigurationForViewport):
      (DumpRenderTreeSupport::numberOfPendingGeolocationPermissionRequests):
      (DumpRenderTreeSupport::resetGeolocationMock):
      (DumpRenderTreeSupport::setMockGeolocationError):
      (DumpRenderTreeSupport::setMockGeolocationPermission):
      (DumpRenderTreeSupport::setMockGeolocationPosition):
      (DumpRenderTreeSupport::scalePageBy):
      (DumpRenderTreeSupport::computedStyleIncludingVisitedInfo):
      * blackberry/WebKitSupport/DumpRenderTreeSupport.h: Added.
      (WebKit):
      (WebCore):
      (WTF):
      (DumpRenderTreeSupport):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106993 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      cf709dde
    • jamesr@google.com's avatar
      [chromium] Allow retaining texture across frames in composited video playback... · 19b38d50
      jamesr@google.com authored
      [chromium] Allow retaining texture across frames in composited video playback and correctly handle lost context
      https://bugs.webkit.org/show_bug.cgi?id=77923
      
      Reviewed by Kenneth Russell.
      
      Thanks to r106840, we can improve the video playback mode a bit. Instead of creating a new texture on every
      frame, this attempts to reuse the texture from the previous frame unless the context is lost. Also improves
      error checking in case the TextureManager cannot successfully reserve memory for the texture.
      
      Tested manually by killing the GPU process with an html5 video playing and verifying that the video playback
      continues and that we don't create a new set of textures for each plane on each frame.
      
      * platform/graphics/chromium/cc/CCVideoLayerImpl.cpp:
      (WebCore::CCVideoLayerImpl::draw):
      (WebCore::CCVideoLayerImpl::reserveTextures):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106992 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      19b38d50
    • commit-queue@webkit.org's avatar
      IndexedDB createIndex should fail if options arg is invalid · d9337708
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=58467
      
      Patch by Eugene Girard <girard@chromium.org> on 2012-02-07
      Reviewed by Tony Chang.
      
      This ticket was fixed as part of https://bugs.webkit.org/show_bug.cgi?id=58471
      Adding unit tests to verify correct operation.
      
      * storage/indexeddb/index-basics-expected.txt:
      * storage/indexeddb/index-basics-workers-expected.txt:
      * storage/indexeddb/resources/index-basics.js:
      (deleteExisting):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106991 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d9337708
    • jchaffraix@webkit.org's avatar
      Unreviewed gardening. · 4bda3526
      jchaffraix@webkit.org authored
      * platform/chromium/test_expectations.txt:
      Mark those 2 tests as crashing in Debug as they seem
      to be hitting an ASSERT.
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106990 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      4bda3526
    • ossy@webkit.org's avatar
      [Qt] Unreviewed gardening, paint the bots green. · f5439173
      ossy@webkit.org authored
      * platform/qt/Skipped: Skip failing/crashing tests, because we like regressions in trunk.
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106989 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      f5439173
    • mdelaney@apple.com's avatar
      https://bugs.webkit.org/show_bug.cgi?id=77912 · b4f13703
      mdelaney@apple.com authored
      Removing old CG shadow code.
              
      A CG-specific shadow offset hack was added in http://trac.webkit.org/changeset/34317
      for this particular setShadow method. However, this shadow offset adjustment for CG
      has since moved down into platform specific code. Thus, this offset adjustment here
      is now redundant.
      
      The CG-only shadow setting code block in this setShadow method is now redundant.
      Since it sets the shadow values to the CGContext directly - and not to the State object -
      it will be overwritten later by any subsequent calls to setting shadow values such as
      blur, offset, or shadow color.
      
      Reviewed by Simon Fraser.
      
      No new tests. Current canvas tests cover this path.
      
      * html/canvas/CanvasRenderingContext2D.cpp:
      (WebCore::CanvasRenderingContext2D::setShadow):
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106988 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      b4f13703
    • commit-queue@webkit.org's avatar
      Process Scroll-gesture events from the compositor. · 634e0de9
      commit-queue@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=77477
      
      Patch by Sadrul Habib Chowdhury <sadrul@chromium.org> on 2012-02-07
      Reviewed by James Robinson.
      
      * src/WebCompositorInputHandlerImpl.cpp:
      (WebKit::WebCompositorInputHandlerImpl::WebCompositorInputHandlerImpl):
      (WebKit::WebCompositorInputHandlerImpl::handleInputEvent):
      * src/WebCompositorInputHandlerImpl.h:
      (WebCompositorInputHandlerImpl):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106987 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      634e0de9
    • jamesr@google.com's avatar
      [chromium] Gracefully handle compositor initialization failure in single-threaded proxy · 4e1451dd
      jamesr@google.com authored
      https://bugs.webkit.org/show_bug.cgi?id=78013
      
      Reviewed by Kenneth Russell.
      
      If compositor initialization fails it's not safe to proceed through the rest of the frame process. This adds
      some early outs.
      
      Tested manually by forcing the first makeContextCurrent() call fail.
      
      * platform/graphics/chromium/cc/CCLayerTreeHost.cpp:
      (WebCore::CCLayerTreeHost::updateLayers):
      * platform/graphics/chromium/cc/CCLayerTreeHost.h:
      (CCLayerTreeHost):
      * platform/graphics/chromium/cc/CCSingleThreadProxy.cpp:
      (WebCore::CCSingleThreadProxy::compositeAndReadback):
      (WebCore::CCSingleThreadProxy::compositeImmediately):
      (WebCore::CCSingleThreadProxy::commitIfNeeded):
      * platform/graphics/chromium/cc/CCSingleThreadProxy.h:
      (CCSingleThreadProxy):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106986 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      4e1451dd
    • ap@apple.com's avatar
      Keep a reference in didReceiveInvalidMessage. · b12895fc
      ap@apple.com authored
              https://bugs.webkit.org/show_bug.cgi?id=78024
      
              Reviewed by Anders Carlsson.
      
              * UIProcess/WebConnectionToWebProcess.cpp:
              (WebKit::WebConnectionToWebProcess::didReceiveInvalidMessage):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106985 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      b12895fc
    • crogers@google.com's avatar
      Add Chris Rogers to reviewers section · 44007a1f
      crogers@google.com authored
      https://bugs.webkit.org/show_bug.cgi?id=78016
      
      Reviewed by Adam Barth.
      
      * Scripts/webkitpy/common/config/committers.py:
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106984 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      44007a1f
    • zimmermann@webkit.org's avatar
      layoutTestController.display() is flaky for SVG tests · 3a5ef64c
      zimmermann@webkit.org authored
      https://bugs.webkit.org/show_bug.cgi?id=78021
      
      Reviewed by Dan Bernstein.
      
      Fix race condition in repaint rect tracking. When running a test using the repaint.js
      harness right after a test starting with "svg/W3C-SVG-1.1" the repaint test will fail,
      the gray overlay isn't contained in the pixel dump anymore. This is because of a specific
      hack that forces 480x360 instead of 800x600 pixel test dumps for any test starting with
      svg/W3C-SVG-1.1. The resizing of the web view was done when dumping the render tree, after
      the repaint test already run, thus invalidating the previously tracked repaint rects.
      
      * DumpRenderTree/mac/DumpRenderTree.mm:
      (dump):
      (runTest):
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@106983 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      3a5ef64c