1. 14 Mar, 2010 2 commits
  2. 13 Mar, 2010 24 commits
  3. 12 Mar, 2010 14 commits
    • barraclough@apple.com's avatar
      * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: · 5ac12f3f
      barraclough@apple.com authored
      Reviewed by NOBODY (build fix).
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55948 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      5ac12f3f
    • enrica@apple.com's avatar
      Fixed broken build on Windows. · d6d7e351
      enrica@apple.com authored
      Added contditional compilation for accelerated compositing.
      
      * WebView.cpp:
      (WebView::deleteBackingStore):
      (WebView::addToDirtyRegion):
      (WebView::updateBackingStore):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55947 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d6d7e351
    • bdakin@apple.com's avatar
      Fix for https://bugs.webkit.org/show_bug.cgi?id=34942 Fullscreen · e7bfed76
      bdakin@apple.com authored
      API naming is inconsistent
      -and corresponding-
      <rdar://problem/7729165>
      
      Reviewed by Simon Fraser.
      
      WebCore: 
      
      This patch changes all occurrences of "fullScreen" to the more 
      popular "fullscreen." webkitEnterFullScreen and 
      webkitExitFullScreen have been maintained for now for backwards 
      compatibility. 
      
      * html/HTMLVideoElement.cpp:
      (WebCore::HTMLVideoElement::webkitEnterFullscreen):
      (WebCore::HTMLVideoElement::webkitExitFullscreen):
      * html/HTMLVideoElement.h:
      (WebCore::HTMLVideoElement::webkitEnterFullScreen):
      (WebCore::HTMLVideoElement::webkitExitFullScreen):
      * html/HTMLVideoElement.idl:
      * platform/graphics/mac/MediaPlayerProxy.h:
      
      WebKit/mac: 
      
      This patch changes all occurrences of "fullScreen" to the more 
      popular "fullscreen."
      
      * Plugins/Hosted/NetscapePluginHostProxy.h:
      (WebKit::NetscapePluginHostProxy::isFullscreenWindowShowing):
      * Plugins/Hosted/NetscapePluginHostProxy.mm:
      (WebKit::NetscapePluginHostProxy::NetscapePluginHostProxy):
      (WebKit::NetscapePluginHostProxy::didEnterFullscreen):
      (WebKit::NetscapePluginHostProxy::didExitFullscreen):
      (WebKit::NetscapePluginHostProxy::setFullscreenWindowIsShowing):
      (WKPCSetFullscreenWindowIsShowing):
      * Plugins/Hosted/WebKitPluginClient.defs:
      * Plugins/WebNetscapePluginView.mm:
      (-[WebNetscapePluginView _workaroundSilverlightFullscreenBug:]):
      (-[WebNetscapePluginView _createPlugin]):
      (-[WebNetscapePluginView _destroyPlugin]):
      * WebView/WebVideoFullscreenHUDWindowController.mm:
      (-[WebVideoFullscreenHUDWindowController windowDidLoad]):
      
      WebKit/win: 
      
      This patch changes all occurrences of "fullScreen" to the more 
      popular "fullscreen."
      
      * FullscreenVideoController.cpp:
      (FullscreenVideoController::onMouseDown):
      (FullscreenVideoController::onMouseMove):
      (FullscreenVideoController::onMouseUp):
      * FullscreenVideoController.h:
      (FullscreenVideoController::fullscreenToHUDCoordinates):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55946 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      e7bfed76
    • jamesr@google.com's avatar
      2010-03-12 James Robinson <jamesr@chromium.org> · b095f1c0
      jamesr@google.com authored
              Reviewed by Adam Barth.
      
              Add virtual destructor to DOMWrapperWorld
              https://bugs.webkit.org/show_bug.cgi?id=36077
      
              DOMWrapperWorld is RefCounted<DOMWrapperWorld>.  IsolatedWorld
              inherits from DOMWrapperWorld and has member variables
              with destructors, so DOMWrapperWorld needs to have a declared virtual
              d'tor to ensure that its derived class's destructors are called.
      
              No new tests, no change in behavior
      
              * bindings/v8/DOMWrapperWorld.h:
              (WebCore::DOMWrapperWorld::~DOMWrapperWorld):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55945 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      b095f1c0
    • barraclough@apple.com's avatar
      * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def: · 69ca0715
      barraclough@apple.com authored
      Reviewed by NOBODY (build fix).
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55944 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      69ca0715
    • barraclough@apple.com's avatar
      Bug 36075 - Clean up screwyness re static string impls & Identifiers. · c24293a1
      barraclough@apple.com authored
      Reviewed by Oliver Hunt.
      
      JavaScriptCore: 
      
      * API/JSClassRef.cpp:
      (OpaqueJSClass::~OpaqueJSClass): Classname may be null/empty, and these are an identifer.  This is okay, since the null/empty strings are shared across all threads.
      * JavaScriptCore.exp:
      * runtime/Identifier.cpp:
      (JSC::Identifier::add): No need to explicitly hash null reps, this is done in the ststic UStringImpl constructor.
      (JSC::Identifier::addSlowCase): UStringImpl::empty() handled & checkCurrentIdentifierTable now called in the header.
      (JSC::Identifier::checkCurrentIdentifierTable): Replaces checkSameIdentifierTable (this no longer checked the rep since the identifierTable pointer was removed from UString::Rep long ago).
      * runtime/Identifier.h:
      (JSC::Identifier::add): Replace call to checkSameIdentifierTable with call to checkCurrentIdentifierTable at head of function.
      * runtime/UStringImpl.cpp:
      (JSC::UStringImpl::~UStringImpl): Remove call to checkConsistency - this function no longer checks anything interesting.
      * runtime/UStringImpl.h:
      (JSC::UStringOrRopeImpl::UStringOrRopeImpl): Set s_refCountFlagIsIdentifier in static constructor.
      (JSC::UStringImpl::UStringImpl): remove calls to checkConsistency (see above), add new ASSERT to substring constructor.
      (JSC::UStringImpl::setHash): ASSERT not static (static strings set the hash in their constructor, should not reach this code path).
      (JSC::UStringImpl::create): Add missing ASSERT.
      (JSC::UStringImpl::setIsIdentifier): ASSERT !isStatic() (static strings hash set in constructor).
      
      WebCore: 
      
      * platform/text/StringImpl.cpp:
      (WebCore::StringImpl::~StringImpl): Add ASSERT
      (WebCore::StringImpl::sharedBuffer): Add ASSERT
      * platform/text/StringImpl.h:
      (WebCore::StringImpl::setHash): Add ASSERT
      (WebCore::StringImpl::isStatic): added.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55943 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c24293a1
    • bdakin@apple.com's avatar
      Fix for https://bugs.webkit.org/show_bug.cgi?id=33739 Fullscreen · c167ac58
      bdakin@apple.com authored
      video HUD stays on top when switching to another window (e.g. via 
      Alt-Tab)
      -and corresponding-
      <rdar://problem/7547574>
      
      Reviewed by Adam Roben.
      
      The HUD was always on top because it had the WS_EX_TOPMOST style. 
      So I removed the style and made m_videoWindow the owner of 
      m_hudWindow. This keeps m_hudWindow on top only when m_videoWindow 
      is the focused window.
      
      * FullscreenVideoController.cpp:
      (FullscreenVideoController::exitFullscreen): ASSERT that movie()->exitFullscreen() also destroyed the hud.
      (FullscreenVideoController::createHUDWindow):
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55942 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      c167ac58
    • enrica@apple.com's avatar
      Content of 3D tests appears at the bottom right corner sometimes. · 401b2ab7
      enrica@apple.com authored
      <rdar://problem/7556244>
      <https://bugs.webkit.org/show_bug.cgi?id=36027>
      
      Reviewed by Simon Fraser.
      
      WebCore: 
      
      There were two problems to solve here:
      - the incorrect anchoring of the rootChildLayer that was causing the composited
        content to be positioned incorrectly
      - the failure to paint the non composited content into the backing store when
        animating composited content.
                
      The first problem has been solved by leaving the original anchor point for the
      rootChildLayer and splitting the tasks of clipping and scrolling using two separate layers.
      The second problem has been solved leveraging the knowledge that WebView has of the dirty region
      of the backing store to pass this information to the layer renderer. This allows the renderer to force
      a paint into the backing store before moving to the compositing.
              
      Tests: compositing/geometry/horizontal-scroll-composited.html
             compositing/geometry/vertical-scroll-composited.html
      
      * manual-tests/win/horizontal-scroll-composited.html: Removed. This is now a layout test.
      * manual-tests/win/milliondollar.html: Added.
      * platform/graphics/win/WKCACFLayerRenderer.cpp:
      (WebCore::WKCACFLayerRenderer::WKCACFLayerRenderer): Added initialization of dirty flag.
      (WebCore::WKCACFLayerRenderer::setScrollFrame): 
      (WebCore::WKCACFLayerRenderer::updateScrollFrame): Updated to resize and position the clip and scroll layers.
      (WebCore::WKCACFLayerRenderer::setRootChildLayer):
      (WebCore::WKCACFLayerRenderer::createRenderer): Added new layer hierarchy.
      (WebCore::WKCACFLayerRenderer::destroyRenderer): Remove clip layer on destroy.
      (WebCore::WKCACFLayerRenderer::resize):
      (WebCore::WKCACFLayerRenderer::paint): Forcing paint massage to trigger paint into the backing store.
      * platform/graphics/win/WKCACFLayerRenderer.h:
      (WebCore::WKCACFLayerRenderer::setBackingStoreDirty): Added.
      
      WebKit/win: 
      
      See detailed comments in WebCore/ChangeLog.
      
      * WebView.cpp:
      (WebView::deleteBackingStore): Reset the dirty flag when deleting the backing store.
      (WebView::addToDirtyRegion): Set the dirty flag when adding dirty rectangles to the
      backing store dirty region.
      (WebView::updateBackingStore): Reset the dirty flag after painting into the backing store.
      (WebView::setAcceleratedCompositing): Removed unnecessary call to updateRootLayerContents.
      (WebView::updateRootLayerContents): Changed the way we pass parameters to setScrollFrame.
      We are passing width and height of the view content together with the offset for the scrolling.
      It was confusing to pass it all as a rectangle, when it is not a rectangle.
      
      LayoutTests: 
      
      * compositing/geometry/horizontal-scroll-composited.html: Added.
      * compositing/geometry/vertical-scroll-composited.html: Added.
      * compositing/resources/apple.jpg: Added.
      * platform/mac/compositing/geometry/horizontal-scroll-composited-expected.checksum: Added.
      * platform/mac/compositing/geometry/horizontal-scroll-composited-expected.png: Added.
      * platform/mac/compositing/geometry/horizontal-scroll-composited-expected.txt: Added.
      * platform/mac/compositing/geometry/vertical-scroll-composited-expected.checksum: Added.
      * platform/mac/compositing/geometry/vertical-scroll-composited-expected.png: Added.
      * platform/mac/compositing/geometry/vertical-scroll-composited-expected.txt: Added.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55941 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      401b2ab7
    • dpranke@chromium.org's avatar
      2010-03-12 Dirk Pranke <dpranke@chromium.org> · 0d7d7721
      dpranke@chromium.org authored
              Reviewed by Dimitri Glazkov.
      
              Fix typo in websocket_server (path_from_base instead of
              path_from_chromium_base).
      
              https://bugs.webkit.org/show_bug.cgi?id=36074
      
              * Scripts/webkitpy/layout_tests/port/websocket_server.py:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55940 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      0d7d7721
    • robert@webkit.org's avatar
      2010-03-12 Robert Hogan <robert@webkit.org> · 3fb2faa4
      robert@webkit.org authored
              Not reviewed, build fix.
      
              Revert http://trac.webkit.org/projects/webkit/changeset/55374 which broke
              the !ENABLE(DATABASE) build on all platforms when attempting to fix the
              --minimal build on Qt.
      
              Support for SQLite now seems to be non-negotiable for the Qt build but making
              it mandatory requires review, so re-break --minimal Qt build for now.
      
              Qt build issue now tracked at https://bugs.webkit.org/show_bug.cgi?id=36073
      
              * page/GeolocationPositionCache.cpp:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55939 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      3fb2faa4
    • dpranke@chromium.org's avatar
      2010-03-12 Dirk Pranke <dpranke@chromium.org> · d5415dfd
      dpranke@chromium.org authored
              Reviewed by Adam Barth.
      
              new-run-webkit-tests --new-baseline doesn't work at all.
      
              It attempts to call a method that isn't defined. To fix it, I
              removed the unnecessary and unnecessarily confusing 'platform'
              argument to the test_type constructor and use the Port object that
              is passed in instead, since we are only ever generating a baseline
              from the port that is currently executing.
      
              https://bugs.webkit.org/show_bug.cgi?id=36046
      
              * Scripts/webkitpy/layout_tests/port/mac.py:
              * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
              * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
              * Scripts/webkitpy/layout_tests/test_types/test_type_base.py:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55938 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d5415dfd
    • dpranke@chromium.org's avatar
      2010-03-12 Dirk Pranke <dpranke@chromium.org> · f00c0851
      dpranke@chromium.org authored
              Reviewed by Adam Barth.
      
              Fix new-run-webkit-tests --run-singly
      
              This script option is currently broken - the script attempts to
              dereference methods and variables that don't exist, which causes
              the Chromium Linux valgrind bot to be quite unhappy. This has been
              broken since r54449 when I renamed Port.start_test_driver to
              Port.start_driver.
      
              https://bugs.webkit.org/show_bug.cgi?id=36042
      
              * Scripts/webkitpy/layout_tests/layout_package/test_shell_thread.py:
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55937 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      f00c0851
    • krit@webkit.org's avatar
      2010-03-12 Dirk Schulze <krit@webkit.org> · d1efe04a
      krit@webkit.org authored
              Unreviewed LayoutTest update.
      
              Result was not updated with r55930.
      
              * platform/mac/svg/custom/dominant-baseline-hanging-expected.txt:
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55936 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      d1efe04a
    • mitz@apple.com's avatar
      <rdar://problem/7709115> REGRESSION: toolbar is missing at http://www.glom.org/ · 83571990
      mitz@apple.com authored
      https://bugs.webkit.org/show_bug.cgi?id=35507
      
      Reviewed by Alexey Proskuryakov.
      
      * css/CSSImportRule.cpp:
      (WebCore::CSSImportRule::setCSSStyleSheet): Extend the change made in
      <http://trac.webkit.org/changeset/48818> to detect the two variants of
      KHTMLFixes.css in @import rules as well as in <link> elements.
      
      
      
      git-svn-id: http://svn.webkit.org/repository/webkit/trunk@55935 268f45cc-cd09-0410-ab3c-d52691b4dbfc
      83571990