GitLab

* src/FrameLoaderClientImpl.cpp:
(WebKit::FrameLoaderClientImpl::objectContentType):



git-svn-id: svn://svn.chromium.org/blink/trunk@81917 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        * fast/images/embed-image-plugins-disabled.html: Added.
        * platform/mac/fast/images/embed-image-expected.checksum:
        * platform/mac/fast/images/embed-image-expected.png:
        * platform/mac/fast/images/embed-image-expected.txt:
        * platform/mac/fast/images/embed-image-plugins-disabled-expected.checksum: Added.
        * platform/mac/fast/images/embed-image-plugins-disabled-expected.png: Added.
        * platform/mac/fast/images/embed-image-plugins-disabled-expected.txt: Added.
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * src/FrameLoaderClientImpl.cpp:
        (WebKit::FrameLoaderClientImpl::objectContentType):
        * src/FrameLoaderClientImpl.h:
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * WebCoreSupport/FrameLoaderClientQt.cpp:
        (WebCore::FrameLoaderClientQt::objectContentType):
        * WebCoreSupport/FrameLoaderClientQt.h:
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * WebCoreSupport/FrameLoaderClientGtk.cpp:
        (WebKit::FrameLoaderClient::objectContentType):
        * WebCoreSupport/FrameLoaderClientGtk.h:
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * WebFrame.cpp:
        (WebFrame::objectContentType):
        * WebFrame.h:
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * WebKitSupport/FrameLoaderClientWx.cpp:
        (WebCore::FrameLoaderClientWx::objectContentType):
        * WebKitSupport/FrameLoaderClientWx.h:
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * WebCoreSupport/WebFrameLoaderClient.h:
        * WebCoreSupport/WebFrameLoaderClient.mm:
        (WebFrameLoaderClient::objectContentType):
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * WebCoreSupport/FrameLoaderClientHaiku.cpp:
        (WebCore::FrameLoaderClientHaiku::objectContentType):
        * WebCoreSupport/FrameLoaderClientHaiku.h:
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * WebCoreSupport/FrameLoaderClientWinCE.cpp:
        (WebKit::FrameLoaderClientWinCE::objectContentType):
        * WebCoreSupport/FrameLoaderClientWinCE.h:
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * WebCoreSupport/FrameLoaderClientEfl.cpp:
        (WebCore::FrameLoaderClientEfl::objectContentType):
        * WebCoreSupport/FrameLoaderClientEfl.h:
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        AC_QuickTime.js uses the common <object>/<embed> paradigm to embed the
        QuickTime plug-in in web pages. r70748 removed our mapping of classids
        to MIME types, which causes WebKit to fall back from the object to the
        embed tag when QuickTime is embedded by this script. The script emits
        the following embed tag to embed a QuickTime movie with a poster frame:

        <embed src="poster-frame.png" target="quicktimeplayer" href="movie.mov">

        The expectation is that a QuickTime plug-in is instantiated to display the
        poster frame, since QuickTime registers for many common image MIME
        types. This is how Gecko behaves for embed. However, WebKit prefers to
        use its native image rendering for image embeds, in which case no movie
        is played when the poster frame is clicked.

        Fix this by changing embed to check for a plug-in that can handle the
        image type before rendering the image natively. This matches Gecko.

        Test: fast/images/embed-image-plugins-disabled.html

        * html/HTMLObjectElement.cpp:
        (WebCore::HTMLObjectElement::parametersForPlugin):
        * html/HTMLPlugInImageElement.cpp:
        (WebCore::HTMLPlugInImageElement::HTMLPlugInImageElement):
        (WebCore::HTMLPlugInImageElement::isImageType):
        (WebCore::HTMLPlugInImageElement::wouldLoadAsNetscapePlugin):
        * html/HTMLPlugInImageElement.h:
        (WebCore::HTMLPlugInImageElement::preferPluginsForImages):
        * loader/EmptyClients.h:
        (WebCore::EmptyFrameLoaderClient::objectContentType):
        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::defaultObjectContentType):
        * loader/FrameLoader.h:
        * loader/FrameLoaderClient.h:
        * loader/SubframeLoader.cpp:
        (WebCore::SubframeLoader::resourceWillUsePlugin):
        (WebCore::SubframeLoader::requestPlugin):
        (WebCore::SubframeLoader::requestObject):
        (WebCore::SubframeLoader::shouldUsePlugin):
        * loader/SubframeLoader.h:
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Darin Adler.

        REGRESSION (r70748): latest nightly builds kills AC_QuickTime.js
        https://bugs.webkit.org/show_bug.cgi?id=49016

        Update objectContentType() implementation to handle the
        shouldPreferPlugInsForImages flag.

        * WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
        (WebKit::WebFrameLoaderClient::objectContentType):
        * WebProcess/WebCoreSupport/WebFrameLoaderClient.h:


git-svn-id: svn://svn.chromium.org/blink/trunk@81916 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by James Robinson.

        Move lifetime management of grContext from a global to being
        per-SharedGraphicsContext3D, which correctly is 1:1 with the
        underlying opengl context.
        https://bugs.webkit.org/show_bug.cgi?id=54330

        No new tests. Existing <canvas> tests exercise this:

        * platform/graphics/chromium/DrawingBufferChromium.cpp:
        (WebCore::DrawingBuffer::DrawingBuffer):
        (WebCore::DrawingBuffer::~DrawingBuffer):
        (WebCore::DrawingBuffer::publishToPlatformLayer):
        (WebCore::DrawingBuffer::setGrContext):
        * platform/graphics/gpu/DrawingBuffer.h:
        * platform/graphics/gpu/SharedGraphicsContext3D.cpp:
        (WebCore::SharedGraphicsContext3D::SharedGraphicsContext3D):
        (WebCore::SharedGraphicsContext3D::~SharedGraphicsContext3D):
        (WebCore::SharedGraphicsContext3D::getGrContext):
        * platform/graphics/gpu/SharedGraphicsContext3D.h:
        * platform/graphics/skia/PlatformContextSkia.cpp:
        (WebCore::PlatformContextSkia::~PlatformContextSkia):
        (WebCore::PlatformContextSkia::setSharedGraphicsContext3D):

git-svn-id: svn://svn.chromium.org/blink/trunk@81915 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Add a stub implementation of the TestInvocation dumping logic.
Update project files to reflect new file.

* WebKitTestRunner/cairo: Added.
* WebKitTestRunner/cairo/TestInvocationCairo.cpp: Added.
(WTR::TestInvocation::dumpPixelsAndCompareWithExpected):
* WebKitTestRunner/win/WebKitTestRunner.vcproj:



git-svn-id: svn://svn.chromium.org/blink/trunk@81914 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Dmitry Titov.

        Add an overloaded WebFrame::createAssociatedURLLoader method that takes WebURLLoaderOptions, so clients can specify CORS options to allow cross-origin requests.
        https://bugs.webkit.org/show_bug.cgi?id=56885

        No new tests. Exposes no new functionality.

        * public/WebFrame.h:
        * src/WebFrameImpl.cpp:
        (WebKit::WebFrameImpl::createAssociatedURLLoader):
        * src/WebFrameImpl.h:

git-svn-id: svn://svn.chromium.org/blink/trunk@81913 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Fixes <http://webkit.org/b/57060> WebKit2.vcproj should compile LayerTreeHostCA

Reviewed by Anders Carlsson.

* WebProcess/WebPage/ca/win/LayerTreeHostCAWin.cpp: Added.
(WebKit::LayerTreeHostCA::platformInitialize):
(WebKit::LayerTreeHostCA::scheduleLayerFlush):
(WebKit::LayerTreeHostCA::platformInvalidate):
(WebKit::LayerTreeHostCA::platformSizeDidChange):
(WebKit::LayerTreeHostCA::platformForceRepaint):
(WebKit::LayerTreeHostCA::platformDidPerformScheduledLayerFlush):
Stubbed out.

* win/WebKit2.vcproj: Added new files. Let VS reorder existing files.

* win/WebKit2Apple.vsprops: Added WebProcess\WebPage\ca to the include path.

git-svn-id: svn://svn.chromium.org/blink/trunk@81912 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Remove reference to LayerTreeHostMac.h from WebFullScreenManagerMac.mm.

* WebProcess/FullScreen/mac/WebFullScreenManagerMac.mm:

git-svn-id: svn://svn.chromium.org/blink/trunk@81911 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Windows build fix take 2: Add new symobl.
        
        (I should have used the EWS bots for this!)

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:



git-svn-id: svn://svn.chromium.org/blink/trunk@81910 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Darin Adler.

        Untangle dependency between event ancestor chain computation and InspectorDOMAgent.
        https://bugs.webkit.org/show_bug.cgi?id=57050

        Inspector's list of event listeners does not need to invoke Node::getEventListeners,
        because it simply wants to collect all ancestors and never uses EventContext bits.

        No functional change, covered by existing tests.

        * dom/Node.cpp:
        (WebCore::getEventAncestors): Converted into a static function.
        (WebCore::Node::dispatchGenericEvent): Changed to pass node to getEventAncestors.
        * dom/Node.h: Removed decl, moved EventDispatchBehavior enum inside.
        * inspector/InspectorDOMAgent.cpp:
        (WebCore::InspectorDOMAgent::getEventListenersForNode): Replaced the call
            to getEventAncestors with a simple ancestor traversal loop.

git-svn-id: svn://svn.chromium.org/blink/trunk@81909 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Tony Chang.

        new-run-webkit-tests: clean up worker model defaults. This
        changes the default behavior to 'threads' instead of
        'old-threads', and clarifies that chromium-win-* and
        chromium-mac-leopard are exceptions.

        https://bugs.webkit.org/show_bug.cgi?id=56971

        * Scripts/webkitpy/layout_tests/port/base.py:
        * Scripts/webkitpy/layout_tests/port/chromium_linux.py:
        * Scripts/webkitpy/layout_tests/port/chromium_mac.py:
        * Scripts/webkitpy/layout_tests/port/chromium_win.py:
        * Scripts/webkitpy/layout_tests/port/test.py:
        * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:

git-svn-id: svn://svn.chromium.org/blink/trunk@81908 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Windows build fix take 1: Removed old symobl.

        * JavaScriptCore.vcproj/JavaScriptCore/JavaScriptCore.def:



git-svn-id: svn://svn.chromium.org/blink/trunk@81907 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by David Levin.

        [fileapi] Make FileError and FileException accessible from WorkerContext
        https://bugs.webkit.org/show_bug.cgi?id=57041

        Fixed some test expectations related to FileError/FileExcetion and
        exposed while testing this change under Chromium.

        * http/tests/filesystem/workers/resolve-url-sync-expected.txt:
2011-03-24  Adam Klein  <adamk@chromium.org>

        Reviewed by David Levin.

        [fileapi] Make FileError and FileException accessible from WorkerContext
        https://bugs.webkit.org/show_bug.cgi?id=57041

        * workers/WorkerContext.idl:

git-svn-id: svn://svn.chromium.org/blink/trunk@81906 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Anders Carlsson.

        Remove legacy version of findPlainText.
        https://bugs.webkit.org/show_bug.cgi?id=57056

        * src/WebFrameImpl.cpp:
        (WebKit::WebFrameImpl::scopeStringMatches):
        Replace use of legacy findPlainText with a version that
        takes an options parameter.
2011-03-24  Sam Weinig  <sam@webkit.org>

        Reviewed by Anders Carlsson.

        Remove legacy version of findPlainText.
        https://bugs.webkit.org/show_bug.cgi?id=57056

        * editing/TextIterator.cpp:
        * editing/TextIterator.h:
        Remove legacy overload of findPlainText that doesn't
        take an options parameter.


git-svn-id: svn://svn.chromium.org/blink/trunk@81905 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Oliver Hunt.

        Ensure that all compilation takes place within a dynamic global object scope
        https://bugs.webkit.org/show_bug.cgi?id=57054
        <rdar://problem/9083011>        

        Otherwise, entry to the global object scope might throw away the code
        we just compiled, causing a crash.

        * JavaScriptCore.exp: Updated for signature change.

        * debugger/Debugger.cpp:
        (JSC::evaluateInGlobalCallFrame):
        * debugger/DebuggerCallFrame.cpp:
        (JSC::DebuggerCallFrame::evaluate): Removed explicit compilation calls
        here because (a) they took place outside a dynamic global object scope
        and (b) they were redundant.

        * interpreter/CachedCall.h:
        (JSC::CachedCall::CachedCall): Updated for signature change.

        * interpreter/Interpreter.cpp:
        (JSC::Interpreter::execute):
        (JSC::Interpreter::executeCall):
        (JSC::Interpreter::executeConstruct): Declare our dynamic global object
        scope earlier, to ensure that compilation takes place within it.

        * runtime/Completion.cpp:
        (JSC::evaluate): Removed explicit compilation calls here because (a)
        they took place outside a dynamic global object scope and (b) they were
        redundant.

        * runtime/Executable.h:
        (JSC::EvalExecutable::compile):
        (JSC::ProgramExecutable::compile):
        (JSC::FunctionExecutable::compileForCall):
        (JSC::FunctionExecutable::compileForConstruct): Added an ASSERT to
        verify our new invariant that all compilation takes place within a
        dynamic global object scope.

        * runtime/JSGlobalObject.cpp:
        (JSC::DynamicGlobalObjectScope::DynamicGlobalObjectScope):
        * runtime/JSGlobalObject.h: Changed the signature of DynamicGlobalObjectScope
        to require a JSGlobalData instead of an ExecState* since it is often
        easier to provide the former, and the latter was not necessary.
2011-03-24  Geoffrey Garen  <ggaren@apple.com>

        Reviewed by Oliver Hunt.

        Ensure that all compilation takes place within a dynamic global object scope
        https://bugs.webkit.org/show_bug.cgi?id=57054

        * WebView/WebScriptDebugDelegate.mm:
        (-[WebScriptCallFrame evaluateWebScript:]): Updated for signature change.
2011-03-24  Geoffrey Garen  <ggaren@apple.com>

        Reviewed by Oliver Hunt.

        Ensure that all compilation takes place within a dynamic global object scope
        https://bugs.webkit.org/show_bug.cgi?id=57054

        * bindings/js/JSErrorHandler.cpp:
        (WebCore::JSErrorHandler::handleEvent):
        * bindings/js/JSEventListener.cpp:
        (WebCore::JSEventListener::handleEvent): Updated for signature change.


git-svn-id: svn://svn.chromium.org/blink/trunk@81904 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Fixes <http://webkit.org/b/57046> LayerTreeHostMac's code should be shareable with Windows

Reviewed by Anders Carlsson.

* WebKit2.xcodeproj/project.pbxproj:

* WebProcess/WebPage/ca/LayerTreeHostCA.cpp: Renamed from Source/WebKit2/WebProcess/WebPage/ca/LayerTreeHostCA.mm.
(WebKit::LayerTreeHostCA::LayerTreeHostCA): Moved some Mac-specific code to
LayerTreeHostCAMac.mm.
(WebKit::LayerTreeHostCA::~LayerTreeHostCA): Wrapped some Mac-specific code in
PLATFORM(MAC).

(WebKit::LayerTreeHostCA::invalidate):
(WebKit::LayerTreeHostCA::sizeDidChange):
(WebKit::LayerTreeHostCA::forceRepaint):
Moved some Mac-specific code to LayerTreeHostCAMac.mm.

(WebKit::LayerTreeHostCA::performScheduledLayerFlush): Renamed from
flushPendingLayerChangesRunLoopObserverCallback. Moved some code from here...
(WebKit::LayerTreeHostCA::didPerformScheduledLayerFlush): ...to here. Moved some
Mac-specific code to LayerTreeHostCAMac.mm.

* WebProcess/WebPage/ca/LayerTreeHostCA.h: Added new functions, wrapped some Mac-specific
declarations in PLATFORM(MAC).

* WebProcess/WebPage/ca/mac/LayerTreeHostCAMac.mm: Added.
(WebKit::LayerTreeHostCA::platformInitialize):
(WebKit::LayerTreeHostCA::scheduleLayerFlush):
(WebKit::LayerTreeHostCA::platformInvalidate):
(WebKit::LayerTreeHostCA::platformSizeDidChange):
(WebKit::LayerTreeHostCA::platformForceRepaint):
(WebKit::LayerTreeHostCA::flushPendingLayerChangesRunLoopObserverCallback):
(WebKit::LayerTreeHostCA::platformDidPerformScheduledLayerFlush):
Code was extracted from LayerTreeHostCA.cpp.

git-svn-id: svn://svn.chromium.org/blink/trunk@81903 bbb929c8-8fbe-4397-9dbb-9b2b20218538

This is the first step toward sharing code with Windows.

Fixes <http://webkit.org/b/57051>.

Reviewed by Anders Carlsson.

* WebKit2.xcodeproj/project.pbxproj:
* WebProcess/WebPage/LayerTreeHost.cpp:
(WebKit::LayerTreeHost::create):
Updated for rename.

* WebProcess/WebPage/ca/LayerTreeHostCA.h: Renamed from Source/WebKit2/WebProcess/WebPage/mac/LayerTreeHostMac.h.
* WebProcess/WebPage/ca/LayerTreeHostCA.mm: Renamed from Source/WebKit2/WebProcess/WebPage/mac/LayerTreeHostMac.mm.

git-svn-id: svn://svn.chromium.org/blink/trunk@81902 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Maciej Stachowiak.

        WebKit2: Cancelling full screen early leaves full screen window up.
        https://bugs.webkit.org/show_bug.cgi?id=56589

        No new tests, as WebKitTestRunner does not currently support the new Full Screen API.

        Notify the UIProcess when it needs to tear down its layer hosting view, turn off
        the background layer when not in accelerated rendering mode and don't swap out
        the web view unnecessarily

        * UIProcess/mac/WKFullScreenWindowController.mm:
        (-[WKFullScreenWindowController beganExitFullScreenAnimation]): Check before swapping _webView.
        (-[WKFullScreenWindowController enterAcceleratedCompositingMode:]): Add the _layerHostingView
            to the full screen window's animationView, not its contentsView.
        (-[WKFullScreenWindowController exitAcceleratedCompositingMode]): Hide the background layer.
        * WebProcess/FullScreen/mac/WebFullScreenManagerMac.mm:
        (WebKit::WebFullScreenManagerMac::setRootFullScreenLayer): If given a null rootLayer, tell
            the client to exit accelerated mode.

git-svn-id: svn://svn.chromium.org/blink/trunk@81901 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Geoffrey Garen.

        REGRESSION (r79987-r80210): Crash in JSWeakObjectMapClear
        https://bugs.webkit.org/show_bug.cgi?id=55671

        This is no longer necessary, and it seems that with the new weakmap
        model it's simply unsafe, so this reduces it to a no-op.

        * API/JSWeakObjectMapRefPrivate.cpp:

git-svn-id: svn://svn.chromium.org/blink/trunk@81900 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Kenneth Russell.

        preserveDrawingBuffer=true is ignored
        https://bugs.webkit.org/show_bug.cgi?id=56987

        Add code to the V8 and JSC bindings to support grabbing the value of
        preserveDrawingBuffer from the input context attributes. Also, in
        WebGLRenderingContext use the WebGLContextAttributes that were input
        directly, not those from the GraphicsContext3D which could have been
        changed.

        No new tests, as this can't be tested with DRT. However, this works
        when tested manually.

        * bindings/js/JSHTMLCanvasElementCustom.cpp:
        (WebCore::JSHTMLCanvasElement::getContext):
        * bindings/v8/custom/V8HTMLCanvasElementCustom.cpp:
        (WebCore::V8HTMLCanvasElement::getContextCallback):
        * html/canvas/WebGLRenderingContext.cpp:
        (WebCore::WebGLRenderingContext::clearIfComposited):
        (WebCore::WebGLRenderingContext::paintRenderingResultsToCanvas):

git-svn-id: svn://svn.chromium.org/blink/trunk@81899 bbb929c8-8fbe-4397-9dbb-9b2b20218538

https://bugs.webkit.org/show_bug.cgi?id=56975
<rdar://problem/8915066>
        
Reviewed by Alexey Proskuryakov.

Source/WebCore: 

Adding support in WebCore to implement readSelectionFromPasteboard
to support Mac OS X services from WebKit2.

* WebCore.exp.in:
* editing/Editor.h:
* editing/mac/EditorMac.mm:
(WebCore::Editor::readSelectionFromPasteboard): Added entry point
to call the paste functions with the specified pasteboard.

Source/WebKit2: 

Adding missing entry point to support Mac OS X services in WebKit2.

* UIProcess/API/mac/WKView.mm:
(-[WKView readSelectionFromPasteboard:]): Added.
* UIProcess/WebPageProxy.h:
* UIProcess/mac/WebPageProxyMac.mm:
(WebKit::WebPageProxy::readSelectionFromPasteboard): Added.
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in: Added synchronous
message.
* WebProcess/WebPage/mac/WebPageMac.mm:
(WebKit::WebPage::readSelectionFromPasteboard): Added.



git-svn-id: svn://svn.chromium.org/blink/trunk@81898 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Pavel Feldman.

        Web Inspector: Inspector does not show correct transfer size for synchronous requests
        https://bugs.webkit.org/show_bug.cgi?id=56951

        Fixed transfer size for synchronous load.

        * http/tests/inspector/network/network-size-sync-expected.txt: Added.
        * http/tests/inspector/network/network-size-sync.html: Added.
        * platform/chromium/test_expectations.txt:
        * platform/gtk/Skipped:
        * platform/mac/Skipped:
        * platform/qt/Skipped:
        * platform/win/Skipped:
2011-03-24  Vsevolod Vlasov  <vsevik@chromium.org>

        Reviewed by Pavel Feldman.

        Web Inspector: Inspector does not show correct transfer size for synchronous requests
        https://bugs.webkit.org/show_bug.cgi?id=56951

        Fixed transfer size for synchronous load.

        Test: http/tests/inspector/network/network-size-sync.html

        * loader/FrameLoader.cpp:
        (WebCore::FrameLoader::willLoadMediaElementURL):
        (WebCore::FrameLoader::commitProvisionalLoad):
        (WebCore::FrameLoader::loadResourceSynchronously):
        (WebCore::FrameLoader::loadedResourceFromMemoryCache):
        * loader/ResourceLoadNotifier.cpp:
        (WebCore::ResourceLoadNotifier::sendRemainingDelegateMessages):
        * loader/ResourceLoadNotifier.h:

git-svn-id: svn://svn.chromium.org/blink/trunk@81897 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Darin Adler.

        Dismissed reversion suggestion is incorrectly learned.
        https://bugs.webkit.org/show_bug.cgi?id=57039

        CorrectionPanel should always use [NSSpellChecker dismissCorrectionBubbleForView].
        [NSSpellChecker cancelCorrectionBubbleForView] is reserved for situation where correction panel
        is dismissed explicitly by ESC key or clicking the dimiss button. Misusing these causes
        incorrect automatic learning.

        * WebCoreSupport/CorrectionPanel.mm:
        (CorrectionPanel::dismissInternal):
2011-03-24  Jia Pu  <jpu@apple.com>

        Reviewed by Darin Adler.

        Dismissed reversion suggestion is incorrectly learned.
        https://bugs.webkit.org/show_bug.cgi?id=57039

        CorrectionPanel should always use [NSSpellChecker dismissCorrectionBubbleForView].
        [NSSpellChecker cancelCorrectionBubbleForView] is reserved for situation where correction panel
        is dismissed explicitly by ESC key or clicking the dimiss button. Misusing these causes
        incorrect automatic learning.

        * UIProcess/mac/CorrectionPanel.mm:
        (WebKit::CorrectionPanel::dismissInternal):

git-svn-id: svn://svn.chromium.org/blink/trunk@81896 bbb929c8-8fbe-4397-9dbb-9b2b20218538

https://bugs.webkit.org/show_bug.cgi?id=49830
<rdar://problem/8705951>

Reviewed by Adam Roben.

Copy code from WebKit1 to WebKit2 to handle initializing fake scrollbars so 
IBM machines with a trackpad send us WM_VSCROLL and WM_HSCROLL messages.

Listen for the WM_VSCROLL and WM_HSCROLL messages, and turn the values into
ScrollDirection and ScrollGranularity, and send a scroll command to the
WebProcess.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::scrollBy): Send a message to the WebProcess.
* UIProcess/WebPageProxy.h:
* UIProcess/win/WebView.cpp:
(WebKit::WebView::wndProc): Add WM_VSCROLL and WM_HSCROLL message handling.
(WebKit::WebView::initialize): Call shouldInitializeTrackPointHack.
(WebKit::WebView::onHorizontalScroll): Turn wParam into a ScrollDirection and ScrollGranularity. 
(WebKit::WebView::onVerticalScroll): Ditto.
(WebKit::WebView::shouldInitializeTrackPointHack): Check the registry for keys that indicate
    the machine has a IBM Trackpoint driver.
* UIProcess/win/WebView.h:
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::scrollBy): Call scroll method.
(WebKit::WebPage::scroll): Moved from WebPageMac and WebPageWin.
(WebKit::WebPage::logicalScroll): Ditto.
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in: Add a new scrollBy message. 
* WebProcess/WebPage/mac/WebPageMac.mm: Remove scroll and logicalScroll, they are now in WebPage.cpp.
* WebProcess/WebPage/win/WebPageWin.cpp: Ditto.



git-svn-id: svn://svn.chromium.org/blink/trunk@81895 bbb929c8-8fbe-4397-9dbb-9b2b20218538

* platform/qt/editing/pasteboard/5065605-expected.txt:
* platform/qt/editing/pasteboard/display-block-on-spans-expected.txt:
* platform/qt/editing/pasteboard/paste-text-011-expected.txt:
* platform/qt/editing/pasteboard/paste-text-at-tabspan-002-expected.txt:



git-svn-id: svn://svn.chromium.org/blink/trunk@81893 bbb929c8-8fbe-4397-9dbb-9b2b20218538

REGRESSION (r81782): http/tests/inspector/extensions-resources-redirect.html sometimes crashes WebKit2's 
web process while handling a WebIconDatabaseProxy::ReceivedIconLoadDecision message

Reviewed by Adam Roben.

The callback objects had some bogus ASSERTs and missed a relevant null check. If a callback was waiting on
a message back from the UIProcess, but was invalidated from within the WebProcess in the meantime, it's
perfectly valid to attempt to performCallback() after the callback function pointer has been cleared.

* loader/icon/IconDatabaseBase.h:
(WebCore::EnumCallback::performCallback):
(WebCore::EnumCallback::invalidate):
(WebCore::EnumCallback::EnumCallback):
(WebCore::ObjectCallback::performCallback):
(WebCore::ObjectCallback::invalidate):
(WebCore::ObjectCallback::ObjectCallback):

LayoutTests: Resolving https://bugs.webkit.org/show_bug.cgi?id=57030, unskip the busted test.

Reviewed by Adam Roben.

* platform/mac-wk2/Skipped:



git-svn-id: svn://svn.chromium.org/blink/trunk@81892 bbb929c8-8fbe-4397-9dbb-9b2b20218538

<rdar://problem/9078569>
https://bugs.webkit.org/show_bug.cgi?id=56995

Reviewed by Darin Adler.

../WebCore: 

* WebCore.exp.in:
Add some editing related exports needed by WebKit2.

../WebKit2: 

* Shared/DictionaryPopupInfo.cpp:
(WebKit::DictionaryPopupInfo::encode):
(WebKit::DictionaryPopupInfo::decode):
* Shared/DictionaryPopupInfo.h:
Add options dictionary.

* UIProcess/API/mac/PageClientImpl.mm:
(WebKit::PageClientImpl::didPerformDictionaryLookup):
Add path that can pass options through.

* WebProcess/WebCoreSupport/mac/WebContextMenuClientMac.mm:
(WebKit::WebContextMenuClient::lookUpInDictionary):
Use the new performDictionaryLookupForSelection which can extract context
if supported.

* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/mac/WebPageMac.mm:
(WebKit::characterRangeAtPositionForPoint):
(WebKit::characterRangeAtPoint):
Split functionality out of characterRangeAtPoint and into 
characterRangeAtPositionForPoint to avoid doing duplicate work 
if you already have the position.

(WebKit::isPositionInRange):
(WebKit::shouldUseSelection):
Add predicate to determine if we should use the selection instead
of expanding to find the word we are over.

(WebKit::WebPage::performDictionaryLookupAtLocation):
If available, use the surrounding paragraph as context to get better extraction
and to get lexicographical information about the word. Also, clean up and use 
editing APIs to make the code more concise and understandable.

(WebKit::WebPage::performDictionaryLookupForSelection):
Use similar logic as in performDictionaryLookupAtLocation to extract additional
details from a selection for use in the dictionary popup.

(WebKit::WebPage::performDictionaryLookupForRange):
Pass options to WebProcess if available.



git-svn-id: svn://svn.chromium.org/blink/trunk@81890 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Eric Seidel.

        REGRESSION (r80231): Bad cast in HTMLTreeBuilder with closed </form> tags
        https://bugs.webkit.org/show_bug.cgi?id=56836

        During fragment parsing, HTMLConstructionSite holds a reference to the
        fragment context's closest form ancestor. If a misnested form end tag is
        then encountered as the first node of the fragment, we will check to see
        if a corresponding form start tag is in scope even though no such tag
        exists. This led to isScope() walking the HTMLElementStack all the way
        to the root DocumentFragment and attempting to cast it to Element*.

        Fix this by ensuring that the inScope() family of functions operate in
        terms of ContainerNodes to account for the fragment case.

        Test: fast/parser/fragment-closest-form-ancestor.html

        * html/parser/HTMLElementStack.cpp:
        (WebCore::HTMLNames::isRootNode):
        (WebCore::HTMLNames::isScopeMarker):
        (WebCore::HTMLNames::isTableScopeMarker):
        (WebCore::HTMLNames::isTableBodyScopeMarker):
        (WebCore::HTMLNames::isTableRowScopeMarker):
        (WebCore::HTMLElementStack::hasOnlyHTMLElementsInScope):
        (WebCore::HTMLElementStack::inScope):
2011-03-24  Andy Estes  <aestes@apple.com>

        Reviewed by Eric Seidel.

        REGRESSION (r80231): Bad cast in HTMLTreeBuilder with closed </form> tags
        https://bugs.webkit.org/show_bug.cgi?id=56836

        * fast/parser/fragment-closest-form-ancestor-expected.txt: Added.
        * fast/parser/fragment-closest-form-ancestor.html: Added.


git-svn-id: svn://svn.chromium.org/blink/trunk@81889 bbb929c8-8fbe-4397-9dbb-9b2b20218538

<rdar://problem/8690506>
https://bugs.webkit.org/show_bug.cgi?id=56874
        
Reviewed by Darin Adler.

When we calculate the style to apply at the insertion point we compare the initial
style at the insertion point against the style calculated at the span we wrap the
copied markup fragment with. We could end up with a series of unnecessary spans
to remove the initial style that simply grow our markup.
The consists in moving the insertion point outside any inline element that could
affect the fragment being inserted when we are not pasting and matching the style.

Test: editing/pasteboard/paste-text-with-style.html

* editing/ReplaceSelectionCommand.cpp:
(WebCore::isInlineNodeWithStyle): Added.
(WebCore::ReplaceSelectionCommand::doApply): Added logic to change the insertion
point according to the new rules.

LayoutTests: Repeated copy and paste-in-place operation results in increasingly verbose HTML.
<rdar://problem/8690506>
https://bugs.webkit.org/show_bug.cgi?id=56874

Reviewed by Darin Adler.

* editing/pasteboard/paste-text-with-style-expected.txt: Added.
* editing/pasteboard/paste-text-with-style.html: Added.
The following are new results for existing tests that now produce
a different markup.
* platform/mac/editing/pasteboard/5065605-expected.txt:
* platform/mac/editing/pasteboard/display-block-on-spans-expected.txt:
* platform/mac/editing/pasteboard/paste-text-011-expected.txt:
* platform/mac/editing/pasteboard/paste-text-at-tabspan-001-expected.txt:
* platform/mac/editing/pasteboard/paste-text-at-tabspan-002-expected.txt:



git-svn-id: svn://svn.chromium.org/blink/trunk@81887 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] When we render WebGL offscreen, color conversion cost a lot of CPU cycles
        https://bugs.webkit.org/show_bug.cgi?id=40884

        The software fallback is now only needed for corner cases like a manual rendering
        of the page to QImage.

        Keeping the image with the last pixel values is no longer needed. Removing it reduce the
        performance for real-time rendering on software surface, but this case should no longer be
        supported.

        The conversion from OpenGL color space and coordinates is done manually for performance. This
        also fix the bug of the inverted X axis due to the transformation.

        The tests and benchmarks are done through Qt API tests.

        * platform/graphics/qt/GraphicsContext3DQt.cpp:
        (WebCore::swapBgrToRgb):
        (WebCore::GraphicsContext3DInternal::paint):
        (WebCore::GraphicsContext3D::reshape):
2011-03-24  Benjamin Poulain  <benjamin.poulain@nokia.com>

        Reviewed by Kenneth Rohde Christiansen.

        [Qt] When we render WebGL offscreen, color conversion cost a lot of CPU cycles
        https://bugs.webkit.org/show_bug.cgi?id=40884

        Add tests and benchmarks for the software fallback of WebGL.

        * tests/benchmarks/webgl/10000_triangles.html: Added.
        * tests/benchmarks/webgl/tst_webgl.cpp: Added.
        (GraphicsView::GraphicsView):
        (GraphicsView::resizeEvent):
        (tst_WebGlPerformance::init):
        (tst_WebGlPerformance::cleanup):
        (tst_WebGlPerformance::benchSoftwareFallbackRgb16):
        (tst_WebGlPerformance::benchSoftwareFallbackRgb32):
        (tst_WebGlPerformance::benchSoftwareFallbackArgb32):
        (tst_WebGlPerformance::benchSoftwareFallbackArgb32Premultiplied):
        (tst_WebGlPerformance::benchmarkFrameRenderingOnImage):
        * tests/benchmarks/webgl/tst_webgl.qrc: Added.
        * tests/benchmarks/webgl/webgl.pro: Added.
        * tests/qgraphicswebview/qgraphicswebview.pro:
        * tests/qgraphicswebview/resources/pointing_right.html: Added.
        * tests/qgraphicswebview/resources/pointing_up.html: Added.
        * tests/qgraphicswebview/tst_qgraphicswebview.cpp:
        (compareImagesFuzzyPixelCount):
        (GraphicsView::GraphicsView):
        (tst_QGraphicsWebView::webglSoftwareFallbackVerticalOrientation):
        (tst_QGraphicsWebView::webglSoftwareFallbackHorizontalOrientation):
        (tst_QGraphicsWebView::compareCanvasToImage):
        * tests/qgraphicswebview/tst_qgraphicswebview.qrc:
        * tests/tests.pro:

git-svn-id: svn://svn.chromium.org/blink/trunk@81886 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by James Robinson.

        [chromium] Remove bool that forces compositor HUD to always be enabled
        https://bugs.webkit.org/show_bug.cgi?id=57034

        * platform/graphics/chromium/cc/CCHeadsUpDisplay.h:
        (WebCore::CCHeadsUpDisplay::enabled):

git-svn-id: svn://svn.chromium.org/blink/trunk@81884 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Reviewed by Joseph Pecoraro.

* Plugins/WebPluginContainerPrivate.h: Changed value of
ENABLE_PLUGIN_PROXY_FOR_VIDEO from 1 to 0.

git-svn-id: svn://svn.chromium.org/blink/trunk@81883 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Darin Adler.

        https://bugs.webkit.org/show_bug.cgi?id=20302
        Correct implementation of signbit on Solaris

        * wtf/MathExtras.h:
        (signbit):

git-svn-id: svn://svn.chromium.org/blink/trunk@81882 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Kenneth Russell.

        [chromium] Add traceEvents to compositor
        https://bugs.webkit.org/show_bug.cgi?id=56965

        * WebCore.gypi:
        * platform/chromium/TraceEvent.h: Added.
        (WebCore::internal::ScopeTracer::ScopeTracer):
        (WebCore::internal::ScopeTracer::~ScopeTracer):
        * platform/graphics/chromium/LayerRendererChromium.cpp:
        (WebCore::LayerRendererChromium::updateRootLayerContents):
        (WebCore::LayerRendererChromium::updateRootLayerScrollbars):
        (WebCore::LayerRendererChromium::updateLayers):
        (WebCore::LayerRendererChromium::drawLayers):
        (WebCore::LayerRendererChromium::finish):
        (WebCore::LayerRendererChromium::present):
        * platform/graphics/chromium/LayerTilerChromium.cpp:
        (WebCore::LayerTilerChromium::update):
2011-03-24  Nat Duca  <nduca@chromium.org>

        Reviewed by Kenneth Russell.

        [chromium] Add traceEvents to compositor
        https://bugs.webkit.org/show_bug.cgi?id=56965

        * src/WebViewImpl.cpp:
        (WebKit::WebViewImpl::composite):

git-svn-id: svn://svn.chromium.org/blink/trunk@81881 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Darin Adler.

        WebKit2: Tabbing from the last focused field to a non-webpage element leaves the selection in a weird state
        <rdar://problem/8553962>

        * WebProcess/WebPage/WebPage.cpp:
        (WebKit::WebPage::setFocused): When the page loses focus, clear out any selection in the frame

git-svn-id: svn://svn.chromium.org/blink/trunk@81880 bbb929c8-8fbe-4397-9dbb-9b2b20218538

Rubber-stamped by Sam Weinig.

* WebProcess/WebPage/mac/WebPageMac.mm:
(WebKit::WebPage::performDefaultBehaviorForKeyEvent): It should be a logical or,
    not a bitwise or.



git-svn-id: svn://svn.chromium.org/blink/trunk@81879 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Yury Semikhatsky.

        Web Inspector: render XHRs matching JSON regex as JSON.
        https://bugs.webkit.org/show_bug.cgi?id=57035

        * English.lproj/localizedStrings.js:
        * WebCore.gypi:
        * WebCore.vcproj/WebCore.vcproj:
        * inspector/front-end/NetworkItemView.js:
        (WebInspector.NetworkItemView):
        * inspector/front-end/RemoteObject.js:
        (WebInspector.LocalJSONObject.prototype.get description.switch.case):
        (WebInspector.LocalJSONObject.prototype.get description):
        (WebInspector.LocalJSONObject.prototype._concatenate):
        (WebInspector.LocalJSONObject.prototype.getProperties):
        (WebInspector.LocalJSONObject.prototype._children):
        * inspector/front-end/ResourceJSONView.js: Added.
        (WebInspector.ResourceJSONView):
        (WebInspector.ResourceJSONView.parseJSON.WebInspector.ResourceJSONView.prototype.hasContent):
        (WebInspector.ResourceJSONView.parseJSON.WebInspector.ResourceJSONView.prototype.show):
        (WebInspector.ResourceJSONView.parseJSON.WebInspector.ResourceJSONView.prototype._initialize):
        * inspector/front-end/WebKit.qrc:
        * inspector/front-end/inspector.css:
        (.resource-view.json):
        * inspector/front-end/inspector.html:

git-svn-id: svn://svn.chromium.org/blink/trunk@81878 bbb929c8-8fbe-4397-9dbb-9b2b20218538

* WebProcess/InjectedBundle/win/InjectedBundleWin.cpp:
(WebKit::InjectedBundle::setHostAllowsAnyHTTPSCertificate): USE(CF_NETWORK) should be USE(CFNETWORK)



git-svn-id: svn://svn.chromium.org/blink/trunk@81877 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Benjamin Poulain.

        Avoided ASCII-cast warnings for WebKit.

        Normally they won't be enabled anyway, but if you build webkit from
        within the Qt mother repository it will pick up Qt's default build
        settings, which do enable it. We need to disable them because
        warnings are treated as errors and there are way too many of them in
        the WebKit code.

        [Qt] Avoid ASCII-cast warnings for WebKit.
        https://bugs.webkit.org/show_bug.cgi?id=57016

        * QtWebKit.pro:

git-svn-id: svn://svn.chromium.org/blink/trunk@81876 bbb929c8-8fbe-4397-9dbb-9b2b20218538

        Reviewed by Dimitri Glazkov.

        Add accessibilityObject accessor to WebDocument (needed for
        building an accessibility tree that includes iframes).
        https://bugs.webkit.org/show_bug.cgi?id=56984

        * public/WebDocument.h:
        * src/WebDocument.cpp:
        (WebKit::WebDocument::accessibilityObject):

git-svn-id: svn://svn.chromium.org/blink/trunk@81875 bbb929c8-8fbe-4397-9dbb-9b2b20218538

The failures are tracked by <http://webkit.org/b/57027> and <http://webkit.org/b/57030>.

* platform/mac-wk2/Skipped: Added http/tests/inspector/extensions-resources-redirect.html.
* platform/win-wk2/Skipped: Added inspector/debugger/source-frame.html.

git-svn-id: svn://svn.chromium.org/blink/trunk@81874 bbb929c8-8fbe-4397-9dbb-9b2b20218538