WTF should have an Xcode project
https://bugs.webkit.org/show_bug.cgi?id=71752

Patch by Mark Rowe <mrowe@apple.com> on 2011-11-19
Reviewed by Adam Barth.

This adds an Xcode project that includes only Stub.cpp and Stub.h.
They’re built in to a library at the appropriate path for each
configuration (WebKitBuild/{Debug,Release}/libWTF.a and
/usr/local/lib/libWTF.a) and headers are installed in to the
appropriate location (WebKitBuild/{Debug,Release}/usr/local/include/wtf
and /usr/local/include/wtf). I tested building WTF in this project and
everything appears to build except for DateMath.cpp (due to bug 71747).
I have not yet done any work on making JavaScriptCore and other
projects use the built products of this new project.

* Configurations: Added.
* Configurations/Base.xcconfig: Copied from Source/JavaScriptCore/Configurations/Base.xcconfig.
* Configurations/CompilerVersion.xcconfig: Copied from Source/JavaScriptCore/Configurations/CompilerVersion.xcconfig.
* Configurations/DebugRelease.xcconfig: Copied from Source/JavaScriptCore/Configurations/DebugRelease.xcconfig.
* Configurations/WTF.xcconfig: Copied from Source/WebKit2/Configurations/Shim.xcconfig.
* WTF.xcodeproj: Added.
* WTF.xcodeproj/project.pbxproj: Added.
* config.h: Copied from Source/JavaScriptCore/config.h.
* icu: Added.
* icu/LICENSE: Copied from Source/JavaScriptCore/icu/LICENSE.
* icu/README: Copied from Source/JavaScriptCore/icu/README.
* icu/unicode: Added.
* icu/unicode/parseerr.h: Copied from Source/JavaScriptCore/icu/unicode/parseerr.h.
* icu/unicode/platform.h: Copied from Source/JavaScriptCore/icu/unicode/platform.h.
* icu/unicode/putil.h: Copied from Source/JavaScriptCore/icu/unicode/putil.h.
* icu/unicode/uchar.h: Copied from Source/JavaScriptCore/icu/unicode/uchar.h.
* icu/unicode/ucnv.h: Copied from Source/JavaScriptCore/icu/unicode/ucnv.h.
* icu/unicode/ucnv_err.h: Copied from Source/JavaScriptCore/icu/unicode/ucnv_err.h.
* icu/unicode/ucol.h: Copied from Source/JavaScriptCore/icu/unicode/ucol.h.
* icu/unicode/uconfig.h: Copied from Source/JavaScriptCore/icu/unicode/uconfig.h.
* icu/unicode/uenum.h: Copied from Source/JavaScriptCore/icu/unicode/uenum.h.
* icu/unicode/uiter.h: Copied from Source/JavaScriptCore/icu/unicode/uiter.h.
* icu/unicode/uloc.h: Copied from Source/JavaScriptCore/icu/unicode/uloc.h.
* icu/unicode/umachine.h: Copied from Source/JavaScriptCore/icu/unicode/umachine.h.
* icu/unicode/unorm.h: Copied from Source/JavaScriptCore/icu/unicode/unorm.h.
* icu/unicode/urename.h: Copied from Source/JavaScriptCore/icu/unicode/urename.h.
* icu/unicode/uscript.h: Copied from Source/JavaScriptCore/icu/unicode/uscript.h.
* icu/unicode/uset.h: Copied from Source/JavaScriptCore/icu/unicode/uset.h.
* icu/unicode/ustring.h: Copied from Source/JavaScriptCore/icu/unicode/ustring.h.
* icu/unicode/utf.h: Copied from Source/JavaScriptCore/icu/unicode/utf.h.
* icu/unicode/utf16.h: Copied from Source/JavaScriptCore/icu/unicode/utf16.h.
* icu/unicode/utf8.h: Copied from Source/JavaScriptCore/icu/unicode/utf8.h.
* icu/unicode/utf_old.h: Copied from Source/JavaScriptCore/icu/unicode/utf_old.h.
* icu/unicode/utypes.h: Copied from Source/JavaScriptCore/icu/unicode/utypes.h.
* icu/unicode/uversion.h: Copied from Source/JavaScriptCore/icu/unicode/uversion.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100850 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Chromium] [WebSocket] export WebSocketChannel interface for plugins
https://bugs.webkit.org/show_bug.cgi?id=72016

Source/WebCore:

Patch by Takashi Toyoshima <toyoshim@chromium.org> on 2011-11-18
Reviewed by Darin Fisher.

Add a interface to send raw binary data.
This interface is used by WebWebSocketChannel implementation
in WebKit API.

No new tests because just export a interface.

* websockets/WebSocketChannel.cpp:
(WebCore::WebSocketChannel::send):
* websockets/WebSocketChannel.h:

Source/WebKit/chromium:

Add WebSocket and WebSocketClient to WebKit API.

Currently, only WebSocketStreamHandle class is exported to WebKit API.
WebSocketStreamHandle implements bi-directional communication ports.
If plugins handles the WebSocket protocol, WebSocketChannel and
WebSocketChannelClient must be exported as WebSocket because they are
the classes which implement the WebSocket protocol stack.

Normally, we use WebWebSocketChannel and WebWebSocketChannelClient as
their class names. But here we use just WebSocket and WebSocketClient
because of avoiding WebWeb prefix and class name confliction on
header include.

Patch by Takashi Toyoshima <toyoshim@chromium.org> on 2011-11-18
Reviewed by Darin Fisher.

* WebKit.gyp:
* public/WebSocket.h: Added.
(WebKit::WebSocket::~WebSocket):
* public/WebSocketClient.h: Added.
(WebKit::WebSocketClient::~WebSocketClient):
* src/WebRuntimeFeatures.cpp: Insert websockets directory to include path in order to avoid filename confliction.
(WebKit::WebRuntimeFeatures::enableSockets): Use WebCore namespace explicitly.
* src/WebSocket.cpp: Added.
(WebKit::WebSocket::create):
* src/WebSocketImpl.cpp: Added.
(WebKit::WebSocketImpl::WebSocketImpl):
(WebKit::WebSocketImpl::~WebSocketImpl):
(WebKit::WebSocketImpl::connect):
(WebKit::WebSocketImpl::subprotocol):
(WebKit::WebSocketImpl::sendText):
(WebKit::WebSocketImpl::sendBinary):
(WebKit::WebSocketImpl::bufferedAmount):
(WebKit::WebSocketImpl::close):
(WebKit::WebSocketImpl::fail):
(WebKit::WebSocketImpl::disconnect):
(WebKit::WebSocketImpl::didConnect):
(WebKit::WebSocketImpl::didReceiveMessage):
(WebKit::WebSocketImpl::didReceiveBinaryData):
(WebKit::WebSocketImpl::didReceiveMessageError):
(WebKit::WebSocketImpl::didStartClosingHandshake):
(WebKit::WebSocketImpl::didClose):
* src/WebSocketImpl.h: Added.
(WebKit::WebSocketImpl::isNull):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Pixel results missing for fast/regions/render-region-custom-style-mark.html
https://bugs.webkit.org/show_bug.cgi?id=72775

Reviewed by Adam Barth.

* platform/mac-snowleopard/fast/regions/render-region-custom-style-mark-expected.png: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100848 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=72591
Remove document.width / document.height

Patch by Vineet Chaudhary <vineet.chaudhary@motorola.com> on 2011-11-18
Reviewed by Darin Adler.

Source/WebCore:

Removed document.width/document.height from JS bindings,
but keeping the same for ObjC bindings. Also it should use
document.body.clientWidth and document.body.clientHeight instead.

* html/HTMLDocument.idl:

LayoutTests:

As support for document.width / document.height has been
removed it should use document.body.clientWidth and document.body.clientHeight
respectively.

* fast/dom/document-width-height-force-layout-expected.txt:
* fast/dom/document-width-height-force-layout.html:
Modified test case to use document.body.clientWidth and document.body.clientHeight.
* fullscreen/video-specified-size-expected.txt:
* fullscreen/video-specified-size.html:
Modified test case to use document.body.clientWidth.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100847 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the GTK+ build.

* page/Navigator.idl: Properly disable the webkitGamepads API if gamepad
isn't enabled at compile time.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100846 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Unreviewed, add one failing test to the Skipped list.

* platform/qt/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100845 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add CMake build infrastructure for the BlackBerry port
https://bugs.webkit.org/show_bug.cgi?id=72768

.:

Reviewed by Adam Barth and Antonio Gomes.

* Source/CMakeLists.txt: Add the BlackBerry port and support for building on QNX.
* Source/cmake/OptionsBlackBerry.cmake: Added.
* Source/cmake/OptionsCommon.cmake: Add QNX QCC-variant of CODE_GENERATOR_PREPROCESSOR.
  We may be able to fold this into the non-MSVC CODE_GENERATOR_PREPROCESSOR.

Source/JavaScriptCore:

Reviewed by Antonio Gomes.

* PlatformBlackBerry.cmake: Added.
* shell/PlatformBlackBerry.cmake: Added.
* wtf/PlatformBlackBerry.cmake: Added.

Source/WebCore:

Reviewed by Antonio Gomes.

* CMakeLists.txt: At this time the BlackBerry port doesn't support generating
  DOM bindings from the SVG IDLs. See WebKit bug #72773.
* PlatformBlackBerry.cmake: Added.

Source/WebKit:

Reviewed by Antonio Gomes.

* blackberry: Added.
* blackberry/CMakeListsBlackBerry.txt: Added.

Tools:

Reviewed by Antonio Gomes.

Add optional --blackberry command line argument to build-webkit to build the
BlackBerry port.

* Scripts/build-webkit:
* Scripts/webkitdirs.pm:
(determineBaseProductDir):
(argumentsForConfiguration):
(builtDylibPathForName):
(determineIsQt):
(isBlackBerry): Added.
(determineIsBlackBerry): Added.
(blackberryTargetArchitecture): Added.
(isAppleWebKit):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100844 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add another isfinite assertion failure for fast/borders/
https://bugs.webkit.org/show_bug.cgi?id=72053

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100843 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, rolling out r100826.
http://trac.webkit.org/changeset/100826
https://bugs.webkit.org/show_bug.cgi?id=72786

Broke Chromium Mac build (Requested by aklein on #webkit).

Patch by Sheriff Bot <webkit.review.bot@gmail.com> on 2011-11-18

Source/WebCore:

* page/ChromeClient.h:
* page/FrameView.cpp:
* page/FrameView.h:
* platform/ScrollView.cpp:
(WebCore::ScrollView::wheelEvent):
* platform/ScrollView.h:
* platform/ScrollableArea.h:
(WebCore::ScrollableArea::scrollbarStyleChanged):
* platform/mac/ScrollAnimatorMac.mm:
(WebCore::ScrollAnimatorMac::updateScrollerStyle):

Source/WebKit2:

* UIProcess/API/mac/PageClientImpl.h:
* UIProcess/API/mac/PageClientImpl.mm:
* UIProcess/PageClient.h:
* UIProcess/WebPageProxy.cpp:
* UIProcess/WebPageProxy.h:
* UIProcess/WebPageProxy.messages.in:
* WebProcess/Plugins/PDF/BuiltInPDFView.cpp:
(WebKit::BuiltInPDFView::scrollbarStyleChanged):
* WebProcess/Plugins/PDF/BuiltInPDFView.h:
* WebProcess/WebCoreSupport/WebChromeClient.cpp:
* WebProcess/WebCoreSupport/WebChromeClient.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100842 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline after r100819.

Unreviewed gardening.

* platform/chromium-mac-leopard/fast/dom/rtl-scroll-to-leftmost-and-resize-expected.png: Added.
* platform/chromium-mac-snowleopard/fast/dom/rtl-scroll-to-leftmost-and-resize-expected.png: Renamed from LayoutTests/platform/chromium-mac/fast/dom/rtl-scroll-to-leftmost-and-resize-expected.png.
* platform/chromium-mac-snowleopard/fast/dom/rtl-scroll-to-leftmost-and-resize-expected.txt: Renamed from LayoutTests/platform/chromium-mac/fast/dom/rtl-scroll-to-leftmost-and-resize-expected.txt.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100841 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix reviewed by lines in ChangeLogs

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100840 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[chromium] composited layers are blurry with a zoom-in page scale factor
https://bugs.webkit.org/show_bug.cgi?id=71225

Source/WebCore:

Patch by Alpha Lam <hclam@chromium.org> on 2011-11-18
Reviewed by James.

Pass contents scale factor to the compositor such that it can:
1. Adjust contentBounds() of the compositoer layers with content scale.
2. Apply the content scale in the painter for texture update in better resolution.
3. Apply the content scale to the dirty rect in CanvasLayerTextureUpdater.

This change fixed blurry problem for all tiled layer types and is not limited to
position:fixed layers.

Tests: compositing/geometry/fixed-position-composited-page-scale-down.html
       compositing/geometry/fixed-position-composited-page-scale.html
       compositing/geometry/fixed-position-iframe-composited-page-scale-down.html
       compositing/geometry/fixed-position-iframe-composited-page-scale.html
       compositing/geometry/fixed-position-transform-composited-page-scale-down.html
       compositing/geometry/fixed-position-transform-composited-page-scale.html

* platform/graphics/chromium/GraphicsLayerChromium.cpp:
(WebCore::GraphicsLayerChromium::setTransform):
(WebCore::GraphicsLayerChromium::updateLayerPreserves3D):
(WebCore::GraphicsLayerChromium::updateContentsScale):
(WebCore::GraphicsLayerChromium::contentsScale):
(WebCore::GraphicsLayerChromium::deviceOrPageScaleFactorChanged):
* platform/graphics/chromium/GraphicsLayerChromium.h:
* platform/graphics/chromium/ImageLayerChromium.cpp:
(WebCore::ImageLayerTextureUpdater::prepareToUpdate):
(WebCore::ImageLayerChromium::needsContentsScale):
* platform/graphics/chromium/ImageLayerChromium.h:
* platform/graphics/chromium/LayerChromium.cpp:
(WebCore::LayerChromium::LayerChromium):
(WebCore::LayerChromium::setContentsScale):
* platform/graphics/chromium/LayerChromium.h:
(WebCore::LayerChromium::needsContentsScale):
(WebCore::LayerChromium::contentsScale):
* platform/graphics/chromium/LayerTextureUpdater.h:
* platform/graphics/chromium/LayerTextureUpdaterCanvas.cpp:
(WebCore::LayerTextureUpdaterCanvas::paintContents):
(WebCore::LayerTextureUpdaterBitmap::prepareToUpdate):
(WebCore::LayerTextureUpdaterSkPicture::prepareToUpdate):
* platform/graphics/chromium/LayerTextureUpdaterCanvas.h:
* platform/graphics/chromium/TiledLayerChromium.cpp:
(WebCore::TiledLayerChromium::needsContentsScale):
(WebCore::TiledLayerChromium::contentBounds):
(WebCore::TiledLayerChromium::prepareToUpdate):
* platform/graphics/chromium/TiledLayerChromium.h:

Source/WebKit/chromium:

Patch by Alpha Lam <hclam@chromium.org> on 2011-11-18
Reviewed by James.

Implements GraphicsLayerClient to return values for pageScaleFactor and deviceScaleFactor. They are used to determine contents scale factor compositer layers.

* src/PageOverlay.cpp:
(WebKit::OverlayGraphicsLayerClientImpl::deviceScaleFactor):
(WebKit::OverlayGraphicsLayerClientImpl::pageScaleFactor):

LayoutTests:

Added test cases that has position:fixed and page scale.

Patch by Alpha Lam <hclam@chromium.org> on 2011-11-18
Reviewed by James.

* compositing/geometry/fixed-position-composited-page-scale-down-expected.txt: Added.
* compositing/geometry/fixed-position-composited-page-scale-down.html: Added.
* compositing/geometry/fixed-position-composited-page-scale-expected.txt: Added.
* compositing/geometry/fixed-position-composited-page-scale.html: Added.
* compositing/geometry/fixed-position-iframe-composited-page-scale-down-expected.txt: Added.
* compositing/geometry/fixed-position-iframe-composited-page-scale-down.html: Added.
* compositing/geometry/fixed-position-iframe-composited-page-scale-expected.txt: Added.
* compositing/geometry/fixed-position-iframe-composited-page-scale.html: Added.
* compositing/geometry/fixed-position-transform-composited-page-scale-down-expected.txt: Added.
* compositing/geometry/fixed-position-transform-composited-page-scale-down.html: Added.
* compositing/geometry/fixed-position-transform-composited-page-scale-expected.txt: Added.
* compositing/geometry/fixed-position-transform-composited-page-scale.html: Added.
* platform/chromium-linux/compositing/geometry/fixed-position-composited-page-scale-down-expected.png: Added.
* platform/chromium-linux/compositing/geometry/fixed-position-composited-page-scale-expected.png: Added.
* platform/chromium-linux/compositing/geometry/fixed-position-iframe-composited-page-scale-down-expected.png: Added.
* platform/chromium-linux/compositing/geometry/fixed-position-iframe-composited-page-scale-expected.png: Added.
* platform/chromium-linux/compositing/geometry/fixed-position-transform-composited-page-scale-down-expected.png: Added.
* platform/chromium-linux/compositing/geometry/fixed-position-transform-composited-page-scale-expected.png: Added.
* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100839 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Move manual-tests out of WebCore.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100838 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, rolling out r100835.
http://trac.webkit.org/changeset/100835
https://bugs.webkit.org/show_bug.cgi?id=72778

Broke some page scale tests

* DumpRenderTree/chromium/EventSender.cpp:
(EventSender::scalePageBy):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100837 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Prepare to move manual-tests out of WebCore
https://bugs.webkit.org/show_bug.cgi?id=72782

Reviewed by Eric Seidel.

.:

Update these files to refer to the new location of ManualTests.

* .gitattributes:
* Source/cmake/WebKitPackaging.cmake:

Tools:

Update this file to refer to the new location of ManualTests.

* Scripts/webkitpy/common/config/build.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100836 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[chromium] Set min/max page scale factors when calling eventSender.scalePageBy to avoid clamping
https://bugs.webkit.org/show_bug.cgi?id=72778

Reviewed by Kenneth Russell.

Set the min/max page scale clamps when a test calls eventSender.scalePageBy() so that the scale isn't clamped.

* DumpRenderTree/chromium/EventSender.cpp:
(EventSender::scalePageBy):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100835 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[WK2][Qt] Move Accelerated Composite animations to UIProcess
https://bugs.webkit.org/show_bug.cgi?id=72753

Source/WebCore:

Add helper method to synchronize animations in TextureMapper.

Reviewed by Noam Rosenthal.

* platform/graphics/texmap/TextureMapperNode.cpp:
(WebCore::TextureMapperNode::syncAnimationsRecursively):
* platform/graphics/texmap/TextureMapperNode.h:

Source/WebKit2:

Currently, all the accelerated compositing animations are applied in WebProcess and
for each frame generated, WebProcess needs to sync with UIProcess.

This patch moves the accelerated compositing animations to UIProcess reducing the
amount of synchronization messages between WebProcess and UIProcess, this is done
sending animations information to UIProcess.

Reviewed by Noam Rosenthal.

* Shared/WebCoreArgumentCoders.cpp:
(CoreIPC::::encode):
(CoreIPC::::decode):

Encode/Decode RefPtr<Animation> instead of Animation class. This facilitates
the data serialization.

* Shared/WebCoreArgumentCoders.h:
* Shared/WebLayerTreeInfo.cpp:
(WebKit::WebLayerInfo::encode):
(WebKit::WebLayerInfo::decode):
(WebKit::WebLayerAnimation::encode):
(WebKit::WebLayerAnimation::decode):

Create WebLayerAnimation struct, it is used to send animation information
to UIProcess.

* Shared/WebLayerTreeInfo.h:
(WebKit::WebLayerAnimation::WebLayerAnimation):
* UIProcess/LayerTreeHostProxy.h:
* UIProcess/qt/LayerTreeHostProxyQt.cpp:
(WebKit::LayerTreeHostProxy::LayerTreeHostProxy):
(WebKit::LayerTreeHostProxy::paintToCurrentGLContext):
(WebKit::LayerTreeHostProxy::updateSceneGraphUpdateTimerFired):
(WebKit::LayerTreeHostProxy::syncLayerParameters):
* WebProcess/WebCoreSupport/WebGraphicsLayer.cpp:
(WebCore::WebGraphicsLayer::addAnimation):
(WebCore::WebGraphicsLayer::pauseAnimation):
(WebCore::WebGraphicsLayer::removeAnimation):
(WebCore::WebGraphicsLayer::syncCompositingStateForThisLayerOnly):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100834 268f45cc-cd09-0410-ab3c-d52691b4dbfc

IDL changes for gamepad support
https://bugs.webkit.org/show_bug.cgi?id=71753

Patch by Scott Graham <scottmg@chromium.org> on 2011-11-18
Reviewed by Adam Barth.

Source/WebCore:

IDL changes and associated plumbing to expose list of gamepad objects
on navigator object (per current spec). Full patch is
https://bugs.webkit.org/show_bug.cgi?id=69451. Only basic existence
test until more plumbing in future patches.

Test: gamepad/gamepad-api.html

* WebCore.gypi:
* bindings/generic/RuntimeEnabledFeatures.h:
(WebCore::RuntimeEnabledFeatures::setWebkitGamepadsEnabled):
(WebCore::RuntimeEnabledFeatures::webkitGamepadsEnabled):
* page/Gamepad.cpp: Added.
(WebCore::Gamepad::Gamepad):
(WebCore::Gamepad::axes):
(WebCore::Gamepad::buttons):
(WebCore::Gamepad::~Gamepad):
* page/Gamepad.h: Added.
* page/Gamepad.idl: Added.
* page/GamepadList.cpp: Added.
(WebCore::GamepadList::~GamepadList):
(WebCore::GamepadList::set):
(WebCore::GamepadList::length):
(WebCore::GamepadList::item):
* page/GamepadList.h: Added.
(WebCore::GamepadList::create):
(WebCore::GamepadList::GamepadList):
* page/GamepadList.idl: Added.
* page/Navigator.cpp:
(WebCore::Navigator::webkitGamepads):
* page/Navigator.h:
* page/Navigator.idl:

Source/WebKit/chromium:

Update to use vendor-prefixed enable.

* src/WebRuntimeFeatures.cpp:
(WebKit::WebRuntimeFeatures::enableGamepad):
(WebKit::WebRuntimeFeatures::isGamepadEnabled):

Tools:

Runtime enable gamepad in chromium test shell.

* DumpRenderTree/chromium/TestShell.cpp:
(TestShell::TestShell):

LayoutTests:

Add basic api existence test.

* gamepad/gamepad-api-expected.txt: Added.
* gamepad/gamepad-api.html: Added.
* gamepad/gamepad-test.js: Added.
(logConsole):
(testExpected):
(reportExpected):
(waitForEventAndEnd):
(waitForEvent._eventCallback):
(waitForEvent):
(waitForEventAndTest._eventCallback):
(waitForEventAndTest):
(waitForEventTestAndEnd):
(endTest):
(logResult):
(consoleWrite):
* platform/chromium/fast/dom/navigator-detached-no-crash-expected.txt:
* platform/efl/Skipped:
* platform/gtk/Skipped:
* platform/mac/Skipped:
* platform/qt/Skipped:
* platform/win/Skipped:
* platform/wincairo/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100833 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, rolling out r100693.
http://trac.webkit.org/changeset/100693
https://bugs.webkit.org/show_bug.cgi?id=72779

This patch caused a rendering regression (see bug 72770)
(Requested by philip__ on #webkit).

Patch by Sheriff Bot <webkit.review.bot@gmail.com> on 2011-11-18

Source/WebCore:

* rendering/RenderObject.cpp:
(WebCore::RenderObject::setStyle):
* rendering/style/RenderStyle.cpp:
(WebCore::RenderStyle::diff):

LayoutTests:

* fast/layers/layer-absolute-parent-opacity-expected.png: Removed.
* fast/layers/layer-absolute-parent-opacity-expected.txt: Removed.
* fast/layers/layer-absolute-parent-opacity.html: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100832 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reflection on composited element doesn't update if the element changes to show no content
https://bugs.webkit.org/show_bug.cgi?id=72774

Source/WebCore:

Reviewed by Chris Marrin.

When a style change results in a reflected element losing its backing store,
we need to clear the backing stores on the layer clones as well.

Test: compositing/reflections/become-simple-composited-reflection.html

* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::updateLayerDrawsContent):

LayoutTests:

Reviewed by Chris Marrin.

Test that removes the background-color of a reflected, composited element.

* compositing/reflections/become-simple-composited-reflection-expected.png: Added.
* compositing/reflections/become-simple-composited-reflection-expected.txt: Added.
* compositing/reflections/become-simple-composited-reflection.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100831 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Convert the encoding of JavaScriptGlue/Change from MacRoman to UTF-8 as it should be.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100830 268f45cc-cd09-0410-ab3c-d52691b4dbfc

DFG JIT fails speculation on InstanceOf if the base is not an object
https://bugs.webkit.org/show_bug.cgi?id=72709

Reviewed by Geoff Garen.

InstanceOf already leverages the fact that we only allow the default
hasInstance implementation. So, if the base is predicted to possibly
be not an object and the CFA has not yet proven otherwise, InstanceOf
will abstain from speculating cell and instead return false if the
base is not a cell.

This appears to be a 1% speed-up on V8 on the V8 harness. 3-4% or so
speed-up in earley-boyer. Neutral according to bencher on SunSpider,
V8, and Kraken. In 32-bit, it's a 0.5% win on SunSpider and a 1.9%
win on V8 even on my harness, due to a 12.5% win on earley-boyer.

I also took this opportunity to make the code for InstanceOf common
between the two JITs. This was partially successful, in that the
"common code" has a bunch of #if's, but overall it seems like a code
size reduction.

* dfg/DFGAbstractState.cpp:
(JSC::DFG::AbstractState::execute):
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileInstanceOfForObject):
(JSC::DFG::SpeculativeJIT::compileInstanceOf):
* dfg/DFGSpeculativeJIT.h:
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100829 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[chromium] Use correct backing store for ChromiumDataObject in pasteboard writes.
https://bugs.webkit.org/show_bug.cgi?id=72767

Source/WebCore:

Reviewed by Tony Chang.

Covered by existing tests.

* editing/chromium/EditorChromium.cpp:
(WebCore::Editor::newGeneralClipboard):
* page/chromium/EventHandlerChromium.cpp:
(WebCore::EventHandler::createDraggingClipboard):
* platform/chromium/ChromiumDataObject.cpp:
(WebCore::ChromiumDataObject::types):
(WebCore::ChromiumDataObject::getData):
(WebCore::ChromiumDataObject::containsFilenames):
(WebCore::ChromiumDataObject::ChromiumDataObject):
* platform/chromium/ChromiumDataObject.h:
(WebCore::ChromiumDataObject::createFromPasteboard):
(WebCore::ChromiumDataObject::create):
(WebCore::ChromiumDataObject::storageMode):
* platform/chromium/ClipboardChromium.cpp:
(WebCore::ClipboardChromium::hasData):

Source/WebKit/chromium:

Reviewed by Tony Chang.

* src/WebDragData.cpp:
(WebKit::WebDragData::initialize):
(WebKit::WebDragData::assign):

LayoutTests:

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100828 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Crash with ranges across a detached, reparented node tree
https://bugs.webkit.org/show_bug.cgi?id=72757

Reviewed by Adam Barth.

Source/WebCore:

Test: fast/dom/move-detached-child-in-range.html

* dom/RangeBoundaryPoint.h:
(WebCore::RangeBoundaryPoint::childBefore): protect the raw child node from getting pulled from under us.

LayoutTests:

* fast/dom/move-detached-child-in-range-expected.txt: Added.
* fast/dom/move-detached-child-in-range.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100827 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=72551
When the recommended scrollbar style changes, WKView's tracking options should
adjust accordingly
-and corresponding-
<rdar://problem/10409328>

Reviewed by Darin Adler.

Source/WebCore:

This new ChromeClient function is called when the recommended scrollbar style
changes. This way, WebKit can respond to the change by adjusting its mouse
tracking.
* page/ChromeClient.h:
(WebCore::ChromeClient::recommendedScrollbarStyleDidChange):

Existing ScrollableArea function scrollbarStyleChanged() now takes an int
indicating the new scrollbar style and a bool indicating whether it is necessary
to force an update. It used to be the case that this function was ONLY used to
force an update (and only called when an updated was needed), but now that it must
also call into the ChromeClient, it is necessary to include a bool tracking
whether we need to force an update. New implementation on FrameView is responsible
for calling ChromeClient, and then that calls into the pre-existing ScrollView
function for the forceUpdate part.
* page/FrameView.cpp:
(WebCore::FrameView::scrollbarStyleChanged):
* page/FrameView.h:
* platform/ScrollView.cpp:
(WebCore::ScrollView:: scrollbarStyleChanged):
* platform/ScrollView.h:
* platform/ScrollableArea.h:
(WebCore::ScrollableArea::scrollbarStyleChanged):
* platform/mac/ScrollAnimatorMac.mm:
(WebCore::ScrollAnimatorMac::updateScrollerStyle):

Source/WebKit2:

These new functions take care of passing along the
recommendedScrollbarStyleDidChange() message that originates in the ChromeClient.
* UIProcess/API/mac/PageClientImpl.h:
* UIProcess/PageClient.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::recommendedScrollbarStyleDidChange):
* UIProcess/WebPageProxy.h:
* UIProcess/WebPageProxy.messages.in:
* WebProcess/WebCoreSupport/WebChromeClient.cpp:
(WebKit::WebChromeClient::recommendedScrollbarStyleDidChange):
* WebProcess/WebCoreSupport/WebChromeClient.h:

This is where we actually respond to the recommendedScrollbarStyleDidChange
message. We remove the existing tracking area and create a new tracking area with
the appropriate tracking options.
* UIProcess/API/mac/PageClientImpl.mm:
(WebKit::PageClientImpl::recommendedScrollbarStyleDidChange):

BuiltInPDFView inherits from WebCore::ScrollableArea, so scrollbarStyleChanged()
must now take two parameters like the one in ScrollableArea.
* WebProcess/Plugins/PDF/BuiltInPDFView.cpp:
(WebKit::BuiltInPDFView::scrollbarStyleChanged):
* WebProcess/Plugins/PDF/BuiltInPDFView.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100826 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Use canplay instead of canplaythrough in http/tests/media to prevent timeout
https://bugs.webkit.org/show_bug.cgi?id=72758

Allow these tests to timeout until the bug is fixed.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100825 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fixes several more void functions in RenderObject that return values.
https://bugs.webkit.org/show_bug.cgi?id=72750

Reviewed by Adam Barth.

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::updateBeforeAfterContent):
(WebCore::RenderBlock::addChildToContinuation):
(WebCore::RenderBlock::addChildToAnonymousColumnBlocks):
(WebCore::RenderBlock::addChild):
(WebCore::RenderBlock::addChildIgnoringContinuation):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100824 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[chromium] Add some useful text to existing debug dump
https://bugs.webkit.org/show_bug.cgi?id=72576

Patch by Shawn Singh <shawnsingh@chromium.org> on 2011-11-18
Reviewed by James Robinson.

No new tests needed.

* platform/graphics/chromium/cc/CCLayerImpl.cpp:
(WebCore::CCLayerImpl::dumpLayerProperties):
* platform/graphics/chromium/cc/CCRenderSurface.cpp:
(WebCore::CCRenderSurface::dumpSurface):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100823 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Forgot to completely de-virtualize isDynamicScope
https://bugs.webkit.org/show_bug.cgi?id=72763

Reviewed by Darin Adler.

* runtime/JSActivation.h: Removed virtual keyword.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100822 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2011-10-24  Darin Fisher  <darin@chromium.org>

        [chromium] Improve fullscreen API
        https://bugs.webkit.org/show_bug.cgi?id=70477

        Reviewed by Adam Barth.

        Move fullscreen entry points from WebView to WebWidget since they are
        just about widget geometry.

        Delay webkitfullscreenchange event until we know that we have entered
        fullscreen mode.  This is accomplished by expanding the WebKit APIs to
        differentiate will{Enter,Exit} from did{Enter,Exit} state changes.

        WebViewImpl is forced to keep a reference to the requested fullscreen
        element while we are requesting to enter fullscreen mode.  Once in
        fullscreen mode, we switch to keeping only a reference to the Frame
        containing the fullscreen element, which we drop upon exiting
        fullscreen mode.

        * public/WebView.h:
        * public/WebViewClient.h:
        * public/WebWidget.h:
        (WebKit::WebWidget::willEnterFullScreen):
        (WebKit::WebWidget::didEnterFullScreen):
        (WebKit::WebWidget::willExitFullScreen):
        (WebKit::WebWidget::didExitFullScreen):
        * public/WebWidgetClient.h:
        (WebKit::WebWidgetClient::enterFullScreen):
        (WebKit::WebWidgetClient::exitFullScreen):
        * src/ChromeClientImpl.cpp:
        (WebKit::ChromeClientImpl::supportsFullscreenForNode):
        (WebKit::ChromeClientImpl::enterFullscreenForNode):
        (WebKit::ChromeClientImpl::exitFullscreenForNode):
        (WebKit::ChromeClientImpl::supportsFullScreenForElement):
        (WebKit::ChromeClientImpl::enterFullScreenForElement):
        (WebKit::ChromeClientImpl::exitFullScreenForElement):
        * src/WebViewImpl.cpp:
        (WebKit::WebViewImpl::willEnterFullScreen):
        (WebKit::WebViewImpl::didEnterFullScreen):
        (WebKit::WebViewImpl::willExitFullScreen):
        (WebKit::WebViewImpl::didExitFullScreen):
        (WebKit::WebViewImpl::enterFullScreenForElement):
        (WebKit::WebViewImpl::exitFullScreenForElement):
        * src/WebViewImpl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100821 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Crash in JSC::DFG::OSRExitCompiler::compileExit(JSC::DFG::OSRExit const&, JSC::DFG::SpeculationRecovery*)
https://bugs.webkit.org/show_bug.cgi?id=72292

Reviewed by Darin Adler.

Fix this for 32_64.

* dfg/DFGOSRExitCompiler32_64.cpp:
(JSC::DFG::OSRExitCompiler::compileExit):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100820 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION: rtl horizontal scrollbar / resize bug - Body shifts on resize when scrolled all the way to the left
https://bugs.webkit.org/show_bug.cgi?id=70395

Reviewed by Tony Chang.

Source/WebCore:

This patch fixes the problem in Mac and Chromium Mac.

The existing test fast/dom/rtl-scroll-to-leftmost-and-resize.html seems does not really work in Mac DRT
(the browswer window is not resized). It works in Chromium-Mac. And Chromium-Mac's code
is forked from Mac.

* platform/chromium/ScrollAnimatorChromiumMac.mm:
(WebCore::ScrollAnimatorChromiumMac::immediateScrollToPoint):
* platform/mac/ScrollAnimatorMac.mm:
(WebCore::ScrollAnimatorMac::immediateScrollToPoint):

LayoutTests:

The test still marks as FAIL in test_expectations because cg-mac might need rebase.

* platform/chromium-mac/fast/dom/rtl-scroll-to-leftmost-and-resize-expected.png: Added.
* platform/chromium-mac/fast/dom/rtl-scroll-to-leftmost-and-resize-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100819 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (r99924): broke 2 pasteboard tests on GTK
https://bugs.webkit.org/show_bug.cgi?id=72131

Reviewed by Tony Chang.

Source/WebCore:

* platform/gtk/ClipboardGtk.cpp:
(WebCore::ClipboardGtk::clearData): Call clearAll now.
(WebCore::ClipboardGtk::clearAllData): Call clearAllExceptFilenames now.
* platform/gtk/DataObjectGtk.cpp:
(WebCore::DataObjectGtk::clearAllExceptFilenames): Renamed from clear.
(WebCore::DataObjectGtk::clearAll): Added this method which also clear filenames.
* platform/gtk/DataObjectGtk.h:
* platform/gtk/PasteboardGtk.cpp: Call clear before setting new clipboard data.
(WebCore::Pasteboard::writeSelection): Ditto.
(WebCore::Pasteboard::writePlainText): Ditto.
(WebCore::Pasteboard::writeURL): Ditto.
(WebCore::Pasteboard::writeImage): Ditto.
* platform/gtk/PasteboardHelper.cpp:
(WebCore::clearClipboardContentsCallback): Use the clearAll method now.

Source/WebKit/gtk:

* WebCoreSupport/EditorClientGtk.cpp:
(WebKit::setSelectionPrimaryClipboardIfNeeded): Use the new clearAll method.

LayoutTests:

* platform/gtk/Skipped: Unskip passing tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100817 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt][Wk2] Add an API test for scroll request from javascript
https://bugs.webkit.org/show_bug.cgi?id=72407

New API test added for QQuickWebView for javascript scroll request
Patch by Dinu Jacob <dinu.jacob@nokia.com> on 2011-11-18
Reviewed by Simon Hausmann.

* UIProcess/API/qt/tests/html/scroll.html: Added.
* UIProcess/API/qt/tests/qquickwebview/tst_qquickwebview.cpp:
(tst_QQuickWebView::scrollRequest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100816 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Layout Test accessibility/anonymous-render-block-in-continuation-causes-crash.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=72761

Mark test as PASS TIMEOUT as it occasionally times out on the bots.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100815 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[chromium] The exclusions tests are all passing. Remove the
test_expectations.txt entry for them.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100812 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[chromium] Test fast/events/dropzone-002.html is failing on Leopard
https://bugs.webkit.org/show_bug.cgi?id=61625

Test is no longer failing since the end of October.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100811 268f45cc-cd09-0410-ab3c-d52691b4dbfc

De-virtualize ExecutableBase::intrinsic
https://bugs.webkit.org/show_bug.cgi?id=72548

Reviewed by Oliver Hunt.

* runtime/Executable.cpp:
(JSC::ExecutableBase::intrinsic): Dynamic cast to NativeExecutable. If successful, call intrinsic, otherwise return default value.
* runtime/Executable.h:
* runtime/JSCell.h:
(JSC::jsDynamicCast): Add jsDynamicCast that duplicates the functionality of dynamic_cast in C++ but uses ClassInfo
rather than requiring C++ RTTI.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100810 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline chromium expectations after r100800.
Note that previous rebaseline was actually meant for r100786.

Unreviewed gardening.

* platform/chromium-win/animations/additive-transform-animations-expected.png: Added.
* platform/chromium-win/animations/additive-transform-animations-expected.txt: Added.
* platform/chromium/animations/additive-transform-animations-expected.png: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100808 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline chromium expectations after r100800.

Unreviewed gardening.

* platform/chromium-linux/css3/images/cross-fade-overflow-position-expected.png: Added.
* platform/chromium-mac-snowleopard/css3/images/cross-fade-overflow-position-expected.png: Added.
* platform/chromium-win/css3/images/cross-fade-overflow-position-expected.png: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100807 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Layout Test fast/writing-mode/japanese-rl-text-with-broken-font.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=66900

Re-adding SNOWLEOPARD flakiness on CG bots that seems to have been
erroneously removed.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100806 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Access key should work on all elements.
https://bugs.webkit.org/show_bug.cgi?id=71854

Patch by Vineet Chaudhary <vineet.chaudhary@motorola.com> on 2011-11-18
Reviewed by Ryosuke Niwa.

Source/WebCore:

All HTML elements can have the accesskey content attribute set.
Specification http://dev.w3.org/html5/spec/Overview.html#the-accesskey-attribute
Adding "accessKey" attribute to HTMLElement.idl file as [Reflect].

Test: fast/forms/access-key-for-all-elements.html

* bindings/objc/PublicDOMInterfaces.h: Moved properties form subclass to base class.
* html/BaseButtonInputType.cpp:
(WebCore::BaseButtonInputType::accessKeyAction): Renamed variable sendToAnyElement to sendMouseEvents.
* html/BaseButtonInputType.h: Ditto
* html/BaseCheckableInputType.cpp:
(WebCore::BaseCheckableInputType::accessKeyAction): Ditto
* html/BaseCheckableInputType.h: Ditto
* html/HTMLAnchorElement.cpp:
(WebCore::HTMLAnchorElement::accessKeyAction): Ditto
* html/HTMLAnchorElement.h: Ditto
* html/HTMLAnchorElement.idl: Removed redundant IDL attribute entries.
* html/HTMLAreaElement.idl: Removed redundant IDL attribute entries.
* html/HTMLButtonElement.cpp:
(WebCore::HTMLButtonElement::accessKeyAction): Renamed variable sendToAnyElement to sendMouseEvents.
* html/HTMLButtonElement.h: Ditto
* html/HTMLButtonElement.idl: Removed redundant IDL attribute entries.
* html/HTMLElement.cpp:
(WebCore::HTMLElement::accessKeyAction): Renamed variable sendToAnyElement to sendMouseEvents.
If the element does not have a defined activation behavior, fire a click event at the element.
* html/HTMLElement.h: Ditto
* html/HTMLElement.idl: Added accessKey IDL attribute.
* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::accessKeyAction): Renamed variable sendToAnyElement to sendMouseEvents.
* html/HTMLInputElement.h: Ditto
* html/HTMLInputElement.idl: Removed redundant IDL attribute entries.
* html/HTMLLabelElement.cpp:
(WebCore::HTMLLabelElement::accessKeyAction): Renamed variable sendToAnyElement to sendMouseEvents.
* html/HTMLLabelElement.h: Ditto
* html/HTMLLabelElement.idl: Removed redundant IDL attribute entries.
* html/HTMLLegendElement.cpp:
(WebCore::HTMLLegendElement::accessKeyAction): Renamed variable sendToAnyElement to sendMouseEvents.
* html/HTMLLegendElement.h: Ditto
* html/HTMLLegendElement.idl: Removed redundant IDL attribute entries.
* html/HTMLOptGroupElement.h: Renamed variable sendToAnyElement to sendMouseEvents.
* html/HTMLSelectElement.cpp:
(WebCore::HTMLSelectElement::accessKeyAction): Ditto
* html/HTMLSelectElement.h: Ditto
* html/HTMLTextAreaElement.h: Ditto
* html/HTMLTextAreaElement.idl: Removed redundant IDL attribute entries.
* html/HiddenInputType.h: Renamed variable sendToAnyElement to sendMouseEvents.
* html/InputType.h: Ditto
* html/RangeInputType.cpp:
(WebCore::RangeInputType::accessKeyAction): Ditto
* html/RangeInputType.h: Ditto

LayoutTests:

Added test case to check whether all elements can also have accessKey attribute.

* fast/forms/access-key-expected.txt:
* fast/forms/access-key-for-all-elements-expected.txt: Added.
* fast/forms/access-key-for-all-elements.html: Added.
* fast/forms/access-key.html: Added more test coverage for select and textarea elements.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100805 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Layout Test accessibility/loading-iframe-sends-notification.html is slow on Win dbg
https://bugs.webkit.org/show_bug.cgi?id=72747

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100804 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove expectation for fast/js/delete-syntax.html, which seems to have
been fixed by http://code.google.com/p/v8/source/detail?r=9993.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100803 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove IMAGE suppression for css3/images/cross-fade-overflow-position.html
to let me use garden-o-matic for rebaselining.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100802 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Infinite recursion in WebCore::RenderInline::computeRectForRepaint
https://bugs.webkit.org/show_bug.cgi?id=52144

Reviewed by Tony Chang.

The bug got fixed, landing the test. As I could not test the old test case, it is
landed as-is (apart from making it dumpAsText which was deemed safe).

* editing/execCommand/infinite-recursion-computeRectForRepaint-expected.txt: Added.
* editing/execCommand/infinite-recursion-computeRectForRepaint.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100801 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Source/WebCore: Appearance of compound transform animations under apps linked on SnowLeopard is incorrect
https://bugs.webkit.org/show_bug.cgi?id=72641
and
<rdar://problem/10314267>

Reviewed by Dean Jackson.

GraphicsLayerCA contains a "linked on or after" check to account for a bug in
Core Animation on SnowLeopard and earlier, which is that CA would apply the list
of animations in reverse order.

Our previous fix was incorrect, because it only adjusted the 'additive' property
of the animation list based on ordering, rather than flipping the entire list.
This change reverses the list of animations before giving them to CA, which fixes
the bug.

Test: animations/additive-transform-animations.html

* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::appendToUncommittedAnimations):
(WebCore::GraphicsLayerCA::createTransformAnimationsFromKeyframes):
* platform/graphics/ca/GraphicsLayerCA.h:

LayoutTests: Appearance of compound transform animations under apps linked on SnowLeopard is incorrect
https://bugs.webkit.org/show_bug.cgi?id=72641

Reviewed by Dean Jackson.

Pixel test for a hardware-animated tranform, where order of operations matters.

* animations/additive-transform-animations-expected.png: Added.
* animations/additive-transform-animations.html: Added.
* platform/mac/animations/additive-transform-animations-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100800 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Ignore an ANGLE leak that is not WebKit's fault

* Scripts/old-run-webkit-tests:
(countAndPrintLeaks):
* Scripts/webkitpy/layout_tests/port/leakdetector.py:
(LeakDetector._callstacks_to_exclude_from_leaks):
Added ScanFromString to the call stacks to exclude on all OS versions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100799 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Suppress IMAGE failures for css3/images/cross-fade-overflow-position.html
Needs rebaseline after r100786.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100794 268f45cc-cd09-0410-ab3c-d52691b4dbfc

bool WebChromeClient::shouldInterruptJavaScript casts pointer to bool and always returns true
https://bugs.webkit.org/show_bug.cgi?id=72568

Reviewed by David Kilzer.

WebChromeClient::shouldInterruptJavaScript called CallUIDelegate, which used an id
as the return value, which was then cast to a bool and always returned true. This
patch uses CallUIDelegateReturningBoolean instead of CallUIDelegate, with the default
return value of NO.

* WebCoreSupport/WebChromeClient.mm:
(WebChromeClient::shouldInterruptJavaScript):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100792 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Ignore another Lion-specific leak that is not WebKit's fault

* Scripts/old-run-webkit-tests:
(countAndPrintLeaks):
* Scripts/webkitpy/layout_tests/port/leakdetector.py:
(LeakDetector._callstacks_to_exclude_from_leaks):
Added SecTransformExecute to the call stacks to exclude.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100790 268f45cc-cd09-0410-ab3c-d52691b4dbfc

No'am Rosenthal  <noam.rosenthal@nokia.com>

WebKit2: Enable serializing of data types needed for cross-process accelerated compositing
https://bugs.webkit.org/show_bug.cgi?id=61694

Add an ArgumentCoder for KeyframeValueList, and modify the TimingFunction ArgumentCoder to
allow encoding const TimingFunctions and not just RefPtr<TimingFunction>.

Reviewed by Simon Hausmann.

* Scripts/webkit2/messages.py:
* Shared/WebCoreArgumentCoders.cpp:
(CoreIPC::::encode):
(CoreIPC::::decode):
* Shared/WebCoreArgumentCoders.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100789 268f45cc-cd09-0410-ab3c-d52691b4dbfc

ASSERTION FAILED: fontCache()->generation() == m_generation (running new-run-webkit-tests)
https://bugs.webkit.org/show_bug.cgi?id=59552

Added CRASH to expectations for fast/encoding/parser-tests-50.html.
It's only crashing on Snow Leopard, but I didn't want to split the
existing line in two.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100788 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[CMake] Remove duplicate dtoa files from CMakeLists.txt
https://bugs.webkit.org/show_bug.cgi?id=72711

Reviewed by Brent Fulgham.

* wtf/CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100787 268f45cc-cd09-0410-ab3c-d52691b4dbfc

-webkit-cross-fade rendered incorrectly in overflow divs
https://bugs.webkit.org/show_bug.cgi?id=72693
<rdar://problem/10468564>

Reviewed by Simon Fraser.

Respect the desired source rectangle when rendering the cross-fade.

Test: css3/images/cross-fade-overflow-position.html

* platform/graphics/CrossfadeGeneratedImage.cpp:
(WebCore::CrossfadeGeneratedImage::drawCrossfade):
(WebCore::CrossfadeGeneratedImage::draw):
(WebCore::CrossfadeGeneratedImage::drawPattern):
* platform/graphics/CrossfadeGeneratedImage.h:

Add a test to make sure that -webkit-cross-fade renders correctly inside overflowed areas.

* css3/images/cross-fade-overflow-position.html: Added.
* css3/images/resources/half-circles.svg: Added.
* platform/mac/css3/images/cross-fade-overflow-position-expected.png: Added.
* platform/mac/css3/images/cross-fade-overflow-position-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100786 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[CMake] Add possibility to overwrite WEBKIT_PORT_DIR in platform specific files
https://bugs.webkit.org/show_bug.cgi?id=72710

Reviewed by Brent Fulgham.

.:

* Source/CMakeLists.txt:

Source/WebKit:

* CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100785 268f45cc-cd09-0410-ab3c-d52691b4dbfc

ASSERTION FAILED: fontCache()->generation() == m_generation (running new-run-webkit-tests)
https://bugs.webkit.org/show_bug.cgi?id=59552

Add CRASH expectation for
fast/frames/sandboxed-iframe-parsing-space-characters.html on Mac 10.6
dbg.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100784 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] Accessibility API tests not loading Gtk's accessibility implementation
https://bugs.webkit.org/show_bug.cgi?id=72732

Reviewed by Philippe Normand.

Ensure that the GAIL module is being loaded by avoiding using
gtk_test_init() directly, but a modified version of it.

* tests/testatk.c:
(testWebkitAtkParentForRootObject): Use a normal GtkWidget (a
GtkWindow) as the parent container for the webView, instead of
using a dummy webView for that, which was hackish.
(testWebkitAtkSetParentForObject): Use normal GtkWidgets instead
of dummy webViews (a GtkWindow and a GtkButton).
(initializeTestingFramework): New function, identical to
gtk_test_init(), but loading GAIL as the only GTK module.
(main): Use initializeTestingFramework().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100783 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed build fix; r100686 broke the EFL build when Geolocation
support is enabled.

Patch by Raphael Kubo da Costa <kubo@profusion.mobi> on 2011-11-18

* platform/efl/GeolocationServiceEfl.cpp: Only declare
s_factoryFunction if CLIENT_BASED_GEOLOCATION is off, otherwise it is
also declared in GeolocationService.cpp.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100782 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[TextureMapper] computePerspectiveTransformIfNeeded is called twice in TextureMapperNode::syncCompositingState
https://bugs.webkit.org/show_bug.cgi?id=72727

TextureMapperNode::computeAllTransforms already has a call to TextureMapperNode::computePerspectiveTransformIfNeeded,
so it does not need to be called again after TextureMapperNode::computeAllTransforms.

Reviewed by Noam Rosenthal.

* platform/graphics/texmap/TextureMapperNode.cpp:
(WebCore::TextureMapperNode::syncCompositingState):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100781 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Chromium] fast/filesystem/file-writer-abort-depth.html crashes occasionally on Win dbg
https://bugs.webkit.org/show_bug.cgi?id=72631

Add CRASH expectation for fast/filesystem/file-writer-abort.html on Linux.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100780 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Guard for the viewportInteractionEngine as it doesn't always exist.

Reviewed by Antonio Gomes.

The engine is non-existent when the view is configured to use desktop
behavior.

* UIProcess/qt/QtWebPageProxy.cpp:
(QtWebPageProxy::handleWheelEvent):
(QtWebPageProxy::didFindZoomableArea):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100779 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add test for strange scrollbars with non-integer scaled SVG image
https://bugs.webkit.org/show_bug.cgi?id=43018

Patch by Philip Rogers <pdr@google.com> on 2011-11-18
Reviewed by Nikolas Zimmermann.

* svg/as-image/resources/green-non-integer-size-rect.svg: Added.
* svg/as-image/svg-non-integer-scaled-image-expected.png: Added.
* svg/as-image/svg-non-integer-scaled-image-expected.txt: Added.
* svg/as-image/svg-non-integer-scaled-image.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100778 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Another attempt at rebaselining fast/box-shadow for Chromium Mac 10.5.
It seems the rebaseline tool made a mess of
http://trac.webkit.org/changeset/100742.

Unreviewed gardening.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100777 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Inspector tests flakily crashing on Linux dbg canary
https://bugs.webkit.org/show_bug.cgi?id=72651

Add CRASH and TIMEOUT expectations for severla more tests.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100776 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Chromium] fast/canvas/canvas-alphaImageData-behavior.html has TEXT failure on Mac 10.5 CG
https://bugs.webkit.org/show_bug.cgi?id=72728

Add a TEXT expectation for canvas/philip/tests/2d.imageData.put.unchanged.html.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100775 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Chromium] fast/canvas/canvas-alphaImageData-behavior.html has TEXT failure on Mac 10.5 CG
https://bugs.webkit.org/show_bug.cgi?id=72728

Add a TEXT expectation.

Unreviewed gardening.

* platform/chromium/test_expectations.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100774 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, skip another GTK crasher.

* platform/gtk/Skipped: Skip fast/borders/inline-mask-overlay-image-outset.html

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100773 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline remaining fast/ tests for Chromium Mac 10.5 CG after crrev.com/110543

Unreviewed gardening.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100772 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] Send DidCreateDestination to DownloadProxy when destination has been created
https://bugs.webkit.org/show_bug.cgi?id=72724

Reviewed by Martin Robinson.

* WebProcess/Downloads/soup/DownloadSoup.cpp:
(WebKit::DownloadClient::didReceiveResponse): Call
didCreateDestination() when destination URI has been successfully
created.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100771 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] Cancel current download operation before download object is deleted
https://bugs.webkit.org/show_bug.cgi?id=72721

Reviewed by Martin Robinson.

It fixes a crash in the web process when a download finishes early
due to a failure.

* WebProcess/Downloads/soup/DownloadSoup.cpp:
(WebKit::Download::cancel): Delete current ResourceHandle.
(WebKit::Download::platformInvalidate): If there's a
ResourceHandle active, cancel the download operation and delete
the handle.
(WebKit::Download::platformDidFinish): Delete current ResourceHandle.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100770 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] Downloads not started by WebView only fail for transport errors
https://bugs.webkit.org/show_bug.cgi?id=72718

Reviewed by Martin Robinson.

Source/WebKit/gtk:

Abort the download with an error if the response contains an
error code. It adds a new test case to the download unit tests.

* tests/testdownload.c:
(notifyDownloadStatusCallback):
(serverCallback):
(test_webkit_download_not_found):
(main):
* webkit/webkitdownload.cpp:
(DownloadClient::didReceiveResponse): Check whether http status of
the response is an error code and emit error signal in such case.

Source/WebKit2:

* WebProcess/Downloads/soup/DownloadSoup.cpp:
(WebKit::DownloadClient::didReceiveResponse): Check if the http
status of the response is an error code and finish the download
with an error in such case.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100769 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline remaining svg/ tests for Chromium Mac 10.5 CG after crrev.com/110543

Unreviewed gardening.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100768 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Support customizing JS alert/confirm/prompt dialogs using QML
https://bugs.webkit.org/show_bug.cgi?id=72319

Reviewed by Simon Hausmann.

Adds alertDialog, confirmDialog and promptDialog properties to QQuickWebView's
privateObject. These are QML components that are created by the webview when the
corresponding function is called in JS.

The dialogs are created in a context that contains a model object, similar to
ListView delegates. The "message" and "defaultValue" parameters are available
in the model object, as well as slots expected to be called by the dialog.

This commit removes the old code for supporting QWidget builtin dialogs. The code
wasn't working properly (closing any dialog was closing the app after QWindow
refactoring in Qt5).

* Target.pri:
* UIProcess/API/qt/qquickwebview.cpp:
(QQuickWebViewPrivate::runJavaScriptAlert):
(QQuickWebViewPrivate::runJavaScriptConfirm):
(QQuickWebViewPrivate::runJavaScriptPrompt):
(QQuickWebViewExperimental::alertDialog):
(QQuickWebViewExperimental::setAlertDialog):
(QQuickWebViewExperimental::confirmDialog):
(QQuickWebViewExperimental::setConfirmDialog):
(QQuickWebViewExperimental::promptDialog):
(QQuickWebViewExperimental::setPromptDialog):
* UIProcess/API/qt/qquickwebview_p.h:
* UIProcess/API/qt/qquickwebview_p_p.h:
* UIProcess/API/qt/tests/qmltests/WebView/tst_javaScriptDialogs.qml: Added.
* UIProcess/API/qt/tests/qmltests/common/alert.html: Added.
* UIProcess/API/qt/tests/qmltests/common/confirm.html: Added.
* UIProcess/API/qt/tests/qmltests/common/prompt.html: Added.
* UIProcess/API/qt/tests/qmltests/qmltests.pro:
* UIProcess/qt/QtDialogRunner.cpp: Added.
(QtDialogRunner::QtDialogRunner):
(QtDialogRunner::~QtDialogRunner):
(DialogContextObject::DialogContextObject):
(DialogContextObject::message):
(DialogContextObject::defaultValue):
(DialogContextObject::dismiss):
(DialogContextObject::accept):
(DialogContextObject::reject):
(QtDialogRunner::initForAlert):
(QtDialogRunner::initForConfirm):
(QtDialogRunner::initForPrompt):
(QtDialogRunner::createDialog):
* UIProcess/qt/QtDialogRunner.h: Added.
(QtDialogRunner::wasAccepted):
(QtDialogRunner::result):
(QtDialogRunner::onAccepted):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100767 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline fast/writing-mode for Chromium Mac 10.5 CG after crrev.com/110543

Unreviewed gardening.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100766 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix compilation warning in ComplexTextControllerCoreText.mm
https://bugs.webkit.org/show_bug.cgi?id=72702

Reviewed by Andreas Kling.

No functionality changed, so no new tests.

* platform/graphics/mac/ComplexTextControllerCoreText.mm:
(WebCore::ComplexTextController::collectComplexTextRunsForCharactersCoreText):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100765 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Support wheel event together with resizesToContents

Reviewed by Simon Hausmann.

* UIProcess/API/qt/qquickwebpage.cpp:
* UIProcess/qt/QtViewportInteractionEngine.cpp:
(WebKit::QtViewportInteractionEngine::wheelEvent):
* UIProcess/qt/QtViewportInteractionEngine.h:
* UIProcess/qt/QtWebPageProxy.cpp:
(QtWebPageProxy::handleWheelEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100764 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Enable WebSocket hybi tests
https://bugs.webkit.org/show_bug.cgi?id=72687

Reviewed by Simon Hausmann.

Source/WebKit/qt:

* WebCoreSupport/DumpRenderTreeSupportQt.cpp:
(DumpRenderTreeSupportQt::defaultHixie76WebSocketProtocolEnabled):
(DumpRenderTreeSupportQt::setHixie76WebSocketProtocolEnabled):
Call WebCore::Settings::setUseHixie76WebSocketProtocol() to switch WebSocket protocols.
* WebCoreSupport/DumpRenderTreeSupportQt.h:

Tools:

Let LayoutTestControllerQt be able to switch WebSocket protocols via overridePreference().

* DumpRenderTree/qt/DumpRenderTreeQt.cpp:
(WebCore::WebPage::resetSettings):
* DumpRenderTree/qt/LayoutTestControllerQt.cpp:
(LayoutTestController::overridePreference):

LayoutTests:

* platform/qt/Skipped:
Unskip WebSocket hybi tests. Skip some of them which are known to fail probably due to
a bug in SocketStreamHandleQt implementation.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100763 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt][WK2] Layer violation: WebPopupMenuProxyQtDesktop.cpp uses files from WebKit/qt https://bugs.webkit.org/show_bug.cgi?id=72696

Reviewed by Kenneth Rohde Christiansen.

Implement the WK2 desktop popup by simply subclassing QComboBox and doing the
same thing as QtWebComboBox from WK1 (a few lines of code).

* UIProcess/qt/WebPopupMenuProxyQtDesktop.cpp:
(WebKit::WebPopupMenuProxyQtDesktop::WebPopupMenuProxyQtDesktop):
(WebKit::WebPopupMenuProxyQtDesktop::~WebPopupMenuProxyQtDesktop):
(WebKit::WebPopupMenuProxyQtDesktop::showPopupMenu):
(WebKit::WebPopupMenuProxyQtDesktop::hidePopupMenu):
(WebKit::WebPopupMenuProxyQtDesktop::eventFilter):
(WebKit::WebPopupMenuProxyQtDesktop::setSelectedIndex):
(WebKit::WebPopupMenuProxyQtDesktop::populate):
* UIProcess/qt/WebPopupMenuProxyQtDesktop.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100762 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline fast/repaint for Chromium Mac 10.5 CG after crrev.com/110543

Unreviewed gardening.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100761 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt][WK2] setViewportInteractionEngine must also set the InteractionEngine for the tap recognizer.
https://bugs.webkit.org/show_bug.cgi?id=72717

Reviewed by Kenneth Rohde Christiansen.

* UIProcess/qt/QtTapGestureRecognizer.cpp:
(WebKit::QtTapGestureRecognizer::setViewportInteractionEngine):
* UIProcess/qt/QtTapGestureRecognizer.h:
* UIProcess/qt/QtWebPageProxy.cpp:
(QtWebPageProxy::setViewportInteractionEngine):
* UIProcess/qt/QtWebPageProxy.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100760 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Not reviewed: restore front-end compilability via updating externs and JS generator.

* inspector/front-end/ElementsPanel.js:
(WebInspector.ElementsPanel.prototype.jumpToNextSearchResult):
(WebInspector.ElementsPanel.prototype.jumpToPreviousSearchResult):
* inspector/front-end/externs.js:
(WebInspector.showPanelForAnchorNavigation):
* inspector/generate-protocol-externs:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100759 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt][WK2] Fix popups for select elements.

Reviewed by Tor Arne Vestbø.

* UIProcess/qt/QtWebPageProxy.cpp:
(QtWebPageProxy::createPopupMenuProxy): Fall back to the "desktop" implementation
for the moment. Better than nothing :)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100758 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Unreviewed debug build fix

* UIProcess/qt/QtWebPageProxy.cpp:
(QtWebPageProxy::QtWebPageProxy): Don't assert on a non-existant
variable.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100757 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: dispatch messages from the front-end to the backend asynchronously.
https://bugs.webkit.org/show_bug.cgi?id=72621

Source/WebCore:

We should align the way we dispatch messages from the front-end to backend across the environments:
  - WebKit has it synchronoulsly
  - Chromium has it asynchronously
  - Remote debugging has it asynchronously
Making it asynchronous made a number of flaky Qt tests pass.

Tests uncovered console agent problem that was also fixed.

Reviewed by Yury Semikhatsky.

* inspector/InspectorConsoleAgent.cpp:
(WebCore::InspectorConsoleAgent::clearFrontend):
* inspector/InspectorFrontendClient.h:
* inspector/InspectorFrontendClientLocal.cpp:
(WebCore::InspectorBackendDispatchTask::InspectorBackendDispatchTask):
(WebCore::InspectorBackendDispatchTask::dispatch):
(WebCore::InspectorBackendDispatchTask::reset):
(WebCore::InspectorBackendDispatchTask::onTimer):
(WebCore::InspectorFrontendClientLocal::InspectorFrontendClientLocal):
(WebCore::InspectorFrontendClientLocal::windowObjectCleared):
(WebCore::InspectorFrontendClientLocal::sendMessageToBackend):
* inspector/InspectorFrontendClientLocal.h:

Source/WebKit/chromium:

Reviewed by Yury Semikhatsky.

* src/InspectorFrontendClientImpl.cpp:
(WebKit::InspectorFrontendClientImpl::performDisconnectFromBackend):
* src/InspectorFrontendClientImpl.h:

LayoutTests:

Now that dispatching is asynchronous, timeline test needed a fix.

Reviewed by Yury Semikhatsky.

* inspector/timeline/timeline-script-tag-1.html:
* platform/qt/Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100756 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK][WK2] Invalid TestNetscapePlugin path when using new-run-webkit-tests
https://bugs.webkit.org/show_bug.cgi?id=72707

Reviewed by Philippe Normand.

Tools:

Make new-run-webkit-tests use the right path to locate the test
plugin used by WebKitTestRunner.

* Scripts/webkitpy/layout_tests/port/gtk.py:

LayoutTests:

Unskipped a test that is passing now.

* platform/gtk-wk2/Skipped: unskipped plugins/get-url-notify-with-url-that-fails-to-load.html

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100755 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Unreviewed build fix: Export QQuickWebPage/Preferences private API
needed for the qml plugin.

* Source/qtwebkit-export.map:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100754 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline svg/custom for Chromium Mac 10.5 CG after crrev.com/110543

Unreviewed gardening.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100753 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline svg/clip-path for Chromium Mac 10.5 CG after crrev.com/110543

Unreviewed gardening.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100752 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[chromium] Pass screen refresh rate into compositor.
https://bugs.webkit.org/show_bug.cgi?id=71040

Patch by Iain Merrick <husky@google.com> on 2011-11-18
Reviewed by Tony Gentilcore.

Source/WebCore:

Covered by CCLayerTreeHostTest.

* platform/PlatformScreen.h:
* platform/chromium/PlatformScreenChromium.cpp:
(WebCore::screenRefreshRate):
* platform/chromium/PlatformSupport.h:
* platform/graphics/chromium/cc/CCLayerTreeHost.h:
(WebCore::CCSettings::CCSettings):
* platform/graphics/chromium/cc/CCThreadProxy.cpp:
(WebCore::CCThreadProxy::initializeImplOnImplThread):

Source/WebKit/chromium:

* public/WebScreenInfo.h:
(WebKit::WebScreenInfo::WebScreenInfo):
* src/PlatformSupport.cpp:
(WebCore::PlatformSupport::screenRefreshRate):
* src/WebViewImpl.cpp:
(WebKit::WebViewImpl::setIsAcceleratedCompositingActive):
* tests/CCLayerTreeHostTest.cpp:
(WTF::CCLayerTreeHostTest::runTest):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100751 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[CSSShaders] Implement the computed style for mesh parameters of the custom() filter
https://bugs.webkit.org/show_bug.cgi?id=72478

Patch by Alexandru Chiculita <achicu@adobe.com> on 2011-11-18
Reviewed by Dean Jackson.

* css3/filters/custom-filter-property-computed-style-expected.txt:
* css3/filters/custom-filter-property-parsing-invalid-expected.txt:
* css3/filters/script-tests/custom-filter-property-computed-style.js:
* css3/filters/script-tests/custom-filter-property-parsing-invalid.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100750 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[CSSShaders] Implement the computed style for mesh parameters of the custom() filter
https://bugs.webkit.org/show_bug.cgi?id=72478

Reviewed by Dean Jackson.

Source/WebCore:

Added parsing and computed style for the mesh rows, columns, mesh box type
(filter-box, border-box, content-box and padding-box) and the detached mode.

Also fixed a case where "custom(none, 10, 10 filter-box)" was incorrectly
treated as "custom(none, 10)".

* css/CSSComputedStyleDeclaration.cpp:
(WebCore::CSSComputedStyleDeclaration::valueForFilter):
* css/CSSParser.cpp:
(WebCore::CSSParser::parseCustomFilter): Fixed a case where invalid syntax was parsed as correct syntax.
* css/CSSPrimitiveValueMappings.h:
(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
(WebCore::CSSPrimitiveValue::operator CustomFilterOperation::MeshBoxType):
* css/CSSStyleSelector.cpp:
(WebCore::CSSStyleSelector::createCustomFilterOperation):
* css/CSSValueList.h:
(WebCore::CSSValueListIterator::isPrimitiveValue):
* platform/graphics/filters/CustomFilterOperation.h:

LayoutTests:

* css3/filters/custom-filter-property-computed-style-expected.txt:
* css3/filters/custom-filter-property-parsing-invalid-expected.txt:
* css3/filters/script-tests/custom-filter-property-computed-style.js:
* css3/filters/script-tests/custom-filter-property-parsing-invalid.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100749 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Unreviewed build fix: Export QQuickWebViewExperimental private API
that was added since I made the export patch.

* Source/qtwebkit-export.map:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100748 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] Accessibility API tests failing because of using non-WebKit GtkWidgets
https://bugs.webkit.org/show_bug.cgi?id=72708

Reviewed by Xan Lopez.

Do not use non-WebKit GtkWidget's in unit tests, to avoid problems
when GTK's accessibility implementation is not being loaded.

* tests/testatk.c:
(testWebkitAtkParentForRootObject): Use a dummy WebKitWebView
widget as the parent container needed for this unit test.
(testWebkitAtkSetParentForObject): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100747 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] WTR and InjectedBundle should not link statically against JSC/WTF https://bugs.webkit.org/show_bug.cgi?id=72697

Reviewed by Kenneth Rohde Christiansen.

.:

* Source/api.pri: On Linux, use a GNU LD version script to manage the
symbol exports, which includes the public WK1/WK2 API as well as a bunch
of private WTF symbols needed for WTR and the Injected Bundle.
* Source/qtwebkit-export.map: Added.

Tools:

* Tools.pro: Make WTR linux only for the moment.
* WebKitTestRunner/InjectedBundle/Target.pri: Remove static jsc/wtf linkage.
* WebKitTestRunner/Target.pri: Ditto.
* qmake/mkspecs/features/default_post.prf: Don't build the main webkit sources
with ELF visibility on Linux. Instead we're temporarily using an LD version script.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100746 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Oops, fix ChangeLog text

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100745 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Removal of QtViewInterface part 3

Reviewed by Kenneth Rohde Christiansen.

Move the remaining forwarded calls from ClientImpl that went through
ViewInterface and then PageProxy or WebView to go straight to
QtWebPageProxy. QtViewInterface is empty now and therefore removed.

* Target.pri:
* UIProcess/API/qt/qquickwebpage_p.h:
* UIProcess/API/qt/qquickwebpage_p_p.h:
* UIProcess/API/qt/qquickwebview.cpp:
(QQuickWebViewPrivate::initialize):
(QQuickWebViewPrivate::chooseFiles):
* UIProcess/API/qt/qquickwebview_p.h:
* UIProcess/API/qt/qquickwebview_p_p.h:
* UIProcess/qt/ClientImpl.cpp:
(qt_wk_runJavaScriptAlert):
(qt_wk_runJavaScriptConfirm):
(qt_wk_runJavaScriptPrompt):
(qt_wk_setStatusText):
(qt_wk_runOpenPanel):
(qt_wk_mouseDidMoveOverElement):
(setupPageUiClient):
* UIProcess/qt/QtViewInterface.cpp: Removed.
* UIProcess/qt/QtViewInterface.h: Removed.
* UIProcess/qt/QtWebPageProxy.cpp:
(QtWebPageProxy::QtWebPageProxy):
(QtWebPageProxy::didChangeStatusText):
(QtWebPageProxy::didMouseMoveOverElement):
(QtWebPageProxy::showContextMenu):
(QtWebPageProxy::hideContextMenu):
(QtWebPageProxy::runJavaScriptAlert):
(QtWebPageProxy::runJavaScriptConfirm):
(QtWebPageProxy::runJavaScriptPrompt):
(QtWebPageProxy::chooseFiles):
(QtWebPageProxy::createContextMenuProxy):
(QtWebPageProxy::didChangeLoadProgress):
* UIProcess/qt/QtWebPageProxy.h:
(QtWebPageProxy::contextMenuItemSelected):
* UIProcess/qt/WebContextMenuProxyQt.cpp:
(WebKit::WebContextMenuProxyQt::WebContextMenuProxyQt):
(WebKit::WebContextMenuProxyQt::create):
(WebKit::WebContextMenuProxyQt::hideContextMenu):
* UIProcess/qt/WebContextMenuProxyQt.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100744 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt] Change default size of the MiniBrowser

Reviewed by Simon Hausmann.

As we lay out desktop pages using the width of 980, change our default
size to 980x735 so that we are not scaling down by default.

* MiniBrowser/qt/MiniBrowserApplication.h:
(WindowOptions::WindowOptions):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100743 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rebaseline fast/box-shadow for Chromium Mac 10.5 after crrev.com/110543

Unreviewed gardening.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100742 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Qt][WK2] UI process segfaults, when visiblity properties are set while the WebProcess is dead.
https://bugs.webkit.org/show_bug.cgi?id=72699

The drawing area may be null, if the web process has crashed, but not yet recovered.
Always null check against drawing area.

Patch by Zalan Bujtas <zbujtas@gmail.com> on 2011-11-18
Reviewed by Simon Hausmann.

* UIProcess/qt/QtWebPageProxy.cpp:
(QtWebPageProxy::setDrawingAreaSize):
(QtWebPageProxy::setVisibleContentRectAndScale):
(QtWebPageProxy::setVisibleContentRectTrajectoryVector):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@100741 268f45cc-cd09-0410-ab3c-d52691b4dbfc