Merge 92908 - Readback composited webgl results for printing

Source/WebCore/ChangeLog

Index: Source/WebCore/ChangeLog
===================================================================
--- Source/WebCore/ChangeLog	(revision 93076)
+++ Source/WebCore/ChangeLog	(working copy)
@@ -1,20 +1,439 @@
-2011-08-12  Denis Lagno  <dilmah@chromium.org>
+2011-08-11  John Bauman  <jbauman@chromium.org>
 
-        Need a way to selectively use hixie-76 for websocket connections depending on destination and/or origin
-        https://bugs.webkit.org/show_bug.cgi?id=65835
+        Readback composited webgl results for printing
+        https://bugs.webkit.org/show_bug.cgi?id=65658
 
-        Reviewed by Alexey Proskuryakov and Yuta Kitamura.
+        Reviewed by James Robinson.
 
-        No new tests. (because it is temporary solution for transient period between hixie-76 and hybi-NN).
+        The real composited results may be locked inside the compositor
+        context's version of a texture because the drawing buffer was
+        automatically cleared, so read from there to get the actual presented
+        version to draw.
 
-        * page/Settings.cpp:
-        (WebCore::Settings::addHixie76WebSocketProtocolException):
-        (WebCore::Settings::useHixie76WebSocketProtocol):
-        * page/Settings.h:
-        (WebCore::Settings::setUseHixie76WebSocketProtocol):
-        * websockets/WebSocketChannel.cpp:
-        (WebCore::WebSocketChannel::WebSocketChannel):
+        * html/canvas/WebGLRenderingContext.cpp:
+        (WebCore::WebGLRenderingContext::paintRenderingResultsToCanvas):
+        * platform/graphics/GraphicsContext3D.h:
+        * platform/graphics/chromium/Extensions3DChromium.h:
+        * platform/graphics/chromium/WebGLLayerChromium.cpp:
+        (WebCore::WebGLLayerChromium::paintRenderedResultsToCanvas):
+        * platform/graphics/chromium/WebGLLayerChromium.h:
+        * platform/graphics/opengl/GraphicsContext3DOpenGL.cpp:
+        (WebCore::GraphicsContext3D::paintCompositedResultsToCanvas):
 
+2011-08-11  Andrew Wason  <rectalogic@rectalogic.com>
+
+        REGRESSION: Qt JavaScript bridge signal connection fails
+        https://bugs.webkit.org/show_bug.cgi?id=66097
+
+        Reviewed by Gavin Barraclough.
+
+        Tests in WebKit/qt/tests/qwebframe
+
+        Call toThisObject() on exec->lexicalGlobalObject() so the correct
+        object is passed to the JS signal handler.
+
+        * bridge/qt/qt_runtime.cpp:
+        (JSC::Bindings::QtRuntimeConnectionMethod::call):
+
+2011-08-11  Nico Weber  <thakis@chromium.org>
+
+        Remove incorrect comment about m_wheelEventHandlerCount
+        https://bugs.webkit.org/show_bug.cgi?id=66117
+
+        Reviewed by Simon Fraser.
+
+        From what I can tell, this number is changed only in
+        Frame::notifyChromeClientWheelEventHandlerCountChanged(),
+        and there it just reads the number modified in
+        Document::didAdd/RemoveWheelEventHandler(). So this is just the number
+        of event handlers (and not related to horizontal scrollbars), and the
+        variable name expresses that already.
+
+        * dom/Document.h:
+
+2011-08-11  James Robinson  <jamesr@chromium.org>
+
+        [chromium] Defer managed texture creation and destruction until updateCompositorResources
+        https://bugs.webkit.org/show_bug.cgi?id=64772
+
+        Reviewed by Kenneth Russell.
+
+        This adds support for deferring GraphicsContext3D calls for texture creation and destruction for managed
+        textures and defers these calls until updateCompositorResources() for contents textures.  The primary benefit of
+        this change is that it decouples managing the texture budget for the next frame (which has to happen before
+        painting layer contents) from the actual GL calls that create/destroy textures.  That way, in the threaded
+        compositing world we can continue to use textures from the previous frame while software painting contents for
+        the next frame into software buffers or SkPictures.
+
+        Also renames LayerTexture to ManagedTexture to better reflect what it is.  Not all ManagedTextures we create are
+        necessarily associated with layers, but they are all managed by a TextureManager.
+
+        Covered by compositing/ layout tests.
+
+        * WebCore.gypi:
+        * platform/graphics/chromium/ContentLayerChromium.h:
+        * platform/graphics/chromium/ImageLayerChromium.cpp:
+        (WebCore::ImageLayerTextureUpdater::updateTextureRect):
+        * platform/graphics/chromium/LayerRendererChromium.cpp:
+        (WebCore::LayerRendererChromium::releaseTextures):
+        (WebCore::LayerRendererChromium::drawLayers):
+        (WebCore::LayerRendererChromium::updateLayers):
+        (WebCore::LayerRendererChromium::getOffscreenLayerTexture):
+        (WebCore::LayerRendererChromium::useRenderSurface):
+        (WebCore::LayerRendererChromium::initializeSharedObjects):
+        * platform/graphics/chromium/LayerRendererChromium.h:
+        * platform/graphics/chromium/LayerTextureUpdater.h:
+        * platform/graphics/chromium/LayerTextureUpdaterCanvas.cpp:
+        (WebCore::LayerTextureUpdaterBitmap::updateTextureRect):
+        (WebCore::LayerTextureUpdaterSkPicture::updateTextureRect):
+        * platform/graphics/chromium/LayerTextureUpdaterCanvas.h:
+        * platform/graphics/chromium/LayerTilerChromium.cpp:
+        (WebCore::LayerTilerChromium::getSingleTexture):
+        (WebCore::LayerTilerChromium::createTile):
+        (WebCore::LayerTilerChromium::updateRect):
+        (WebCore::LayerTilerChromium::drawTiles):
+        * platform/graphics/chromium/LayerTilerChromium.h:
+        (WebCore::LayerTilerChromium::Tile::Tile):
+        (WebCore::LayerTilerChromium::Tile::texture):
+        * platform/graphics/chromium/ManagedTexture.cpp: Renamed from Source/WebCore/platform/graphics/chromium/LayerTexture.cpp.
+        (WebCore::ManagedTexture::ManagedTexture):
+        (WebCore::ManagedTexture::~ManagedTexture):
+        (WebCore::ManagedTexture::isValid):
+        (WebCore::ManagedTexture::reserve):
+        (WebCore::ManagedTexture::unreserve):
+        (WebCore::ManagedTexture::bindTexture):
+        (WebCore::ManagedTexture::framebufferTexture2D):
+        * platform/graphics/chromium/ManagedTexture.h: Renamed from Source/WebCore/platform/graphics/chromium/LayerTexture.h.
+        (WebCore::ManagedTexture::create):
+        (WebCore::ManagedTexture::format):
+        (WebCore::ManagedTexture::isReserved):
+        * platform/graphics/chromium/RenderSurfaceChromium.h:
+        * platform/graphics/chromium/TextureManager.cpp:
+        (WebCore::TextureManager::TextureManager):
+        (WebCore::TextureManager::deleteEvictedTextures):
+        (WebCore::TextureManager::removeTexture):
+        (WebCore::TextureManager::allocateTexture):
+        (WebCore::TextureManager::requestTexture):
+        * platform/graphics/chromium/TextureManager.h:
+        (WebCore::TextureManager::create):
+        (WebCore::TextureManager::setAssociatedContextDebugOnly):
+        (WebCore::TextureManager::associatedContextDebugOnly):
+        * platform/graphics/chromium/TiledLayerChromium.h:
+        * platform/graphics/chromium/cc/CCHeadsUpDisplay.cpp:
+        (WebCore::CCHeadsUpDisplay::draw):
+        * platform/graphics/chromium/cc/CCHeadsUpDisplay.h:
+        * platform/graphics/chromium/cc/CCRenderSurface.cpp:
+        (WebCore::CCRenderSurface::prepareContentsTexture):
+        (WebCore::CCRenderSurface::drawSurface):
+        * platform/graphics/chromium/cc/CCRenderSurface.h:
+        (WebCore::CCRenderSurface::contentsTexture):
+        * platform/graphics/chromium/cc/CCTiledLayerImpl.cpp:
+        (WebCore::CCTiledLayerImpl::bindContentsTexture):
+
+2011-08-11  David Kilzer  <ddkilzer@apple.com>
+
+        <http://webkit.org/b/66113> [CFNetwork] willCacheResponse() leaks CFCachedURLResponseRef if delegate changes CacheStoragePolicy
+
+        Reviewed by Joseph Pecoraro.
+
+        This first appeared in ToT WebKit r23462 (Windows merge).
+
+        * platform/network/cf/ResourceHandleCFNet.cpp:
+        (WebCore::willCacheResponse): Only retain cachedResponse if we
+        are returning the same object passed into the method, otherwise
+        we end up double-retaining the new object created.
+
+2011-08-05  Nat Duca  <nduca@chromium.org>
+
+        [chromium] Make WebViewImpl point at CCLayerTreeHost and related separation
+        https://bugs.webkit.org/show_bug.cgi?id=65791
+
+        With this patch, LayerRendererChromium becomes increasingly responsible
+        for rendering, while render scheduling and tree hosting moves into the
+        CCLayerTreeHost.
+
+        Reviewed by James Robinson.
+
+        * platform/graphics/chromium/ContentLayerChromium.cpp:
+        (WebCore::ContentLayerChromium::createTextureUpdaterIfNeeded):
+        * platform/graphics/chromium/LayerChromium.h:
+        * platform/graphics/chromium/LayerRendererChromium.cpp:
+        (WebCore::LayerRendererChromium::create):
+        (WebCore::LayerRendererChromium::LayerRendererChromium):
+        (WebCore::LayerRendererChromium::initialize):
+        (WebCore::LayerRendererChromium::releaseTextures):
+        (WebCore::LayerRendererChromium::updateRootLayerContents):
+        (WebCore::LayerRendererChromium::drawRootLayer):
+        (WebCore::LayerRendererChromium::invalidateRootLayerRect):
+        (WebCore::LayerRendererChromium::rootLayerChanged):
+        (WebCore::LayerRendererChromium::viewportChanged):
+        (WebCore::LayerRendererChromium::updateLayers):
+        (WebCore::LayerRendererChromium::drawLayers):
+        (WebCore::LayerRendererChromium::drawLayersInternal):
+        (WebCore::LayerRendererChromium::getFramebufferPixels):
+        (WebCore::LayerRendererChromium::getOffscreenLayerTexture):
+        (WebCore::LayerRendererChromium::copyOffscreenTextureToDisplay):
+        (WebCore::LayerRendererChromium::useRenderSurface):
+        (WebCore::LayerRendererChromium::setScissorToRect):
+        (WebCore::LayerRendererChromium::layerTreeAsText):
+        (WebCore::LayerRendererChromium::dumpRenderSurfaces):
+        * platform/graphics/chromium/LayerRendererChromium.h:
+        (WebCore::LayerRendererChromium::settings):
+        (WebCore::LayerRendererChromium::owner):
+        (WebCore::LayerRendererChromium::rootLayer):
+        (WebCore::LayerRendererChromium::skiaContext):
+        * platform/graphics/chromium/WebGLLayerChromium.cpp:
+        (WebCore::WebGLLayerChromium::setTextureUpdated):
+        * platform/graphics/chromium/cc/CCHeadsUpDisplay.cpp:
+        (WebCore::CCHeadsUpDisplay::draw):
+        (WebCore::CCHeadsUpDisplay::enabled):
+        * platform/graphics/chromium/cc/CCHeadsUpDisplay.h:
+        * platform/graphics/chromium/cc/CCLayerTreeHost.cpp:
+        (WebCore::CCLayerTreeHost::create):
+        (WebCore::CCLayerTreeHost::CCLayerTreeHost):
+        (WebCore::CCLayerTreeHost::initialize):
+        (WebCore::CCLayerTreeHost::animateAndLayout):
+        (WebCore::CCLayerTreeHost::createLayerTreeHostCommitter):
+        (WebCore::CCLayerTreeHost::createLayerTreeHostImpl):
+        (WebCore::CCLayerTreeHost::context):
+        (WebCore::CCLayerTreeHost::compositeAndReadback):
+        (WebCore::CCLayerTreeHost::createRootLayerPainter):
+        (WebCore::CCLayerTreeHost::finishAllRendering):
+        (WebCore::CCLayerTreeHost::invalidateRootLayerRect):
+        (WebCore::CCLayerTreeHost::setNeedsCommitAndRedraw):
+        (WebCore::CCLayerTreeHost::setNeedsRedraw):
+        (WebCore::CCLayerTreeHost::setRootLayer):
+        (WebCore::CCLayerTreeHost::setViewport):
+        (WebCore::CCLayerTreeHost::setVisible):
+        (WebCore::CCLayerTreeHost::doComposite):
+        (WebCore::CCLayerTreeHost::composite):
+        (WebCore::CCLayerTreeHost::reallocateRenderer):
+        * platform/graphics/chromium/cc/CCLayerTreeHost.h:
+        (WebCore::CCLayerTreeHost::animating):
+        (WebCore::CCLayerTreeHost::setAnimating):
+        (WebCore::CCLayerTreeHost::rootLayer):
+        (WebCore::CCLayerTreeHost::settings):
+        (WebCore::CCLayerTreeHost::viewportContentRect):
+        (WebCore::CCLayerTreeHost::viewportScrollPosition):
+        (WebCore::CCLayerTreeHost::viewportVisibleRect):
+        * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
+        (WebCore::CCLayerTreeHostImpl::create):
+        (WebCore::CCLayerTreeHostImpl::CCLayerTreeHostImpl):
+        (WebCore::CCLayerTreeHostImpl::drawLayers):
+        (WebCore::CCLayerTreeHostImpl::drawLayersOnMainThread):
+        * platform/graphics/chromium/cc/CCLayerTreeHostImpl.h:
+        * platform/graphics/chromium/cc/CCLayerTreeHostImplProxy.cpp:
+        (WebCore::CCLayerTreeHostImplProxy::initImplOnCCThread):
+
+2011-08-11  Jeff Miller  <jeffm@apple.com>
+
+        WebCore::AVFWrapper fails to remove observer for kCACFContextNeedsFlushNotification
+        https://bugs.webkit.org/show_bug.cgi?id=66116
+
+        We need to pass 0 as the object parameter to CFNotificationCenterRemoveObserver() when removing
+        kCACFContextNeedsFlushNotification to match what we do when we registered for the same
+        notification with CFNotificationCenterAddObserver().
+
+        Reviewed by John Sullivan.
+
+        No new tests, uses existing media tests.
+
+        * platform/graphics/avfoundation/MediaPlayerPrivateAVFoundationCF.cpp:
+        (WebCore::AVFWrapper::disconnectAndDeleteAVFWrapper): Pass 0 as the object parameter to CFNotificationCenterRemoveObserver().
+        (WebCore::AVFWrapper::notificationCallback): Remove obsolete FIXME comment.
+
+2011-08-11  Ryosuke Niwa  <rniwa@webkit.org>
+
+        Share code between isStyleSpanOrSpanWithOnlyStyleAttribute, isUnstyledStyleSpan,
+        isSpanWithoutAttributesOrUnstyleStyleSpan and replaceWithSpanOrRemoveIfWithoutAttributes
+        https://bugs.webkit.org/show_bug.cgi?id=66091
+
+        Reviewed by Tony Chang.
+
+        Extracted common code as hasNoAttributeOrOnlyStyleAttribute. The only behavioral difference is that
+        replaceWithSpanOrRemoveIfWithoutAttributes will now remove elements with class="Apple-style-span",
+        for which I'm adding a test.
+
+        Test: editing/style/remove-styled-element-with-style-span.html
+
+        * editing/ApplyStyleCommand.cpp:
+        (WebCore::hasNoAttributeOrOnlyStyleAttribute):
+        (WebCore::isStyleSpanOrSpanWithOnlyStyleAttribute):
+        (WebCore::isUnstyledStyleSpan):
+        (WebCore::isSpanWithoutAttributesOrUnstyleStyleSpan):
+        (WebCore::ApplyStyleCommand::replaceWithSpanOrRemoveIfWithoutAttributes):
+
+2011-08-11  Tom Zakrajsek  <tomz@codeaurora.org>
+
+        Add HTMLUnknownElement interface as defined in
+        http://www.w3.org/TR/html5/elements.html#elements-in-the-dom.
+        https://bugs.webkit.org/show_bug.cgi?id=41841
+
+        Reviewed by Adam Barth.
+
+        Test: fast/html/unknown-tag.html
+
+        * CMakeLists.txt:
+        * CodeGenerators.pri:
+        * DerivedSources.cpp:
+        * DerivedSources.make:
+        * GNUmakefile.list.am:
+        * WebCore.gypi:
+        * WebCore.vcproj/WebCore.vcproj:
+        * WebCore.xcodeproj/project.pbxproj:
+        * bindings/scripts/CodeGeneratorV8.pm:
+        (IsDOMNodeType):
+        * dom/make_names.pl:
+        (defaultParametersHash):
+        (buildConstructorMap):
+        (printJSElementIncludes):
+        (printElementIncludes):
+        (printWrapperFunctions):
+        (printWrapperFactoryCppFile):
+        * html/HTMLTagNames.in:
+        * html/HTMLUnknownElement.h: Added.
+        (WebCore::HTMLUnknownElement::create):
+        (WebCore::HTMLUnknownElement::HTMLUnknownElement):
+        * html/HTMLUnknownElement.idl: Added.
+        * mathml/mathtags.in:
+        * page/DOMWindow.idl:
+        * svg/svgtags.in:
+
+2011-08-11  Levi Weintraub  <leviw@chromium.org>
+
+        Remove dead code: borderInnerRect
+        https://bugs.webkit.org/show_bug.cgi?id=66087
+
+        Reviewed by Adam Barth.
+
+        Removing an old and currently unused function: RenderObject::borderInnerRect.
+
+        No new tests since I'm just pruning dead code.
+
+        * rendering/RenderObject.cpp:
+        * rendering/RenderObject.h:
+
+2011-08-11  Simon Fraser  <simon.fraser@apple.com>
+
+        Avoid adding the visible wash layer twice
+        https://bugs.webkit.org/show_bug.cgi?id=66098
+
+        Reviewed by Chris Marrin.
+
+        Fix two issues; only create a m_visibleTileWashLayer if we don't have one
+        already, and avoid adding multiple m_visibleTileWashLayers to the sublayers
+        array when flipping between tiled and non-tiled.
+
+        * platform/graphics/ca/GraphicsLayerCA.cpp:
+        (WebCore::GraphicsLayerCA::swapFromOrToTiledLayer):
+
+2011-08-11  Benjamin Poulain  <benjamin@webkit.org>
+
+        The PageSerializer should always use absolute URL
+        https://bugs.webkit.org/show_bug.cgi?id=66006
+
+        Reviewed by Adam Barth.
+
+        Page serializer resolves every URL for storing the resources.
+        The MarkupAccumulator used there should use the same rules in order to
+        create a valid file.
+
+        No new tests because the test infrastructure for this
+        does not exist with layout tests and the Chromium unit test for
+        the serializer are disabled.
+
+        * page/PageSerializer.cpp:
+        (WebCore::SerializerMarkupAccumulator::SerializerMarkupAccumulator):
+
+2011-08-01  Adrienne Walker  <enne@google.com>
+
+        Enable fast path scrolling for composited iframes
+        https://bugs.webkit.org/show_bug.cgi?id=65056
+
+        Reviewed by James Robinson.
+
+        iframes with composited layers for content should be able to do fast
+        path scrolling because they draw into their own backing. To enable
+        this, useSlowRepaints is modified to return true only for the cases
+        that apply to composited layers. The fast path scrolling function is
+        modified to send the invalidation to the backing itself rather than to
+        the window.
+
+        Renamed a number of "slow repaint" functions to be the more accurate
+        "cannot blit to window" as it will not always force slow path
+        scrolling.
+
+        * page/FrameView.cpp:
+        (WebCore::FrameView::reset):
+        (WebCore::FrameView::useSlowRepaints):
+        (WebCore::FrameView::useSlowRepaintsIfNotOverlapped):
+        (WebCore::FrameView::contentsInCompositedLayer):
+        (WebCore::FrameView::setCannotBlitToWindow):
+        (WebCore::FrameView::scrollContentsFastPath):
+        (WebCore::FrameView::scrollContentsSlowPath):
+        * page/FrameView.h:
+        * rendering/RenderLayer.cpp:
+        (WebCore::RenderLayer::cannotBlitToWindow):
+        * rendering/RenderLayer.h:
+        * rendering/RenderView.cpp:
+        (WebCore::RenderView::paintBoxDecorations):
+
+2011-08-10  Adrienne Walker  <enne@google.com>
+
+        Support creating compositing layers for scrollable frames and iframes
+        https://bugs.webkit.org/show_bug.cgi?id=55257
+
+        Reviewed by James Robinson.
+
+        When force compositing mode is turned on, iframes with overflow will
+        become composited (in order to enable faster scrolling).
+
+        This patch was originally by Daniel Sievers <sievers@chromium.org>.
+
+        Tests: platform/chromium/compositing/force-compositing-mode/no-overflow-iframe-layer.html
+               platform/chromium/compositing/force-compositing-mode/overflow-hidden-iframe-layer.html
+               platform/chromium/compositing/force-compositing-mode/overflow-iframe-enter-compositing.html
+               platform/chromium/compositing/force-compositing-mode/overflow-iframe-layer.html
+               platform/chromium/compositing/force-compositing-mode/overflow-iframe-leave-compositing.html
+
+        * rendering/RenderLayerCompositor.cpp:
+        (WebCore::RenderLayerCompositor::cacheAcceleratedCompositingFlags):
+        (WebCore::RenderLayerCompositor::requiresCompositingForScrollableFrame):
+        * rendering/RenderLayerCompositor.h:
+
+2011-08-11  Yong Li  <yoli@rim.com>
+
+        https://bugs.webkit.org/show_bug.cgi?id=65944
+        Replace QuotesData::operator==() with QuotesData::equal(const QuotesData*
+        , const QuotesData*) because QuotesData::operator==() were using non-standard
+        C++ code that is not supported by all compilers.
+
+        Reviewed by Alexey Proskuryakov.
+
+        No new tests because some existing test cases can show the issue,
+        for example, fast/css/content/content-quotes-01.html.
+
+        * rendering/RenderQuote.cpp:
+        (WebCore::RenderQuote::styleDidChange):
+        * rendering/style/QuotesData.cpp:
+        (WebCore::QuotesData::equal):
+        * rendering/style/QuotesData.h:
+        * rendering/style/RenderStyle.cpp:
+        (WebCore::RenderStyle::setQuotes):
+        * rendering/style/StyleRareInheritedData.cpp:
+        (WebCore::StyleRareInheritedData::operator==):
+
+2011-08-11  David Hyatt  <hyatt@apple.com>
+
+        https://bugs.webkit.org/show_bug.cgi?id=66075
+
+        Turn on regions and exclusions by default. Fix one stray ifdef that I missed.
+
+        Reviewed by James Robinson.
+
+        * css/CSSComputedStyleDeclaration.cpp:
+        (WebCore::contentToCSSValue):
+
 2011-08-11  David Reveman  <reveman@chromium.org>
 
         [Chromium] Temporarily disable layer anti-aliasing on ChromeOS.