WebKit-https.git
5 years ago[Win] Turn off SSE Instructions for 32-bit Builds
bfulgham@apple.com [Fri, 19 Sep 2014 20:19:57 +0000 (20:19 +0000)]
[Win] Turn off SSE Instructions for 32-bit Builds
https://bugs.webkit.org/show_bug.cgi?id=136959
<rdar://problem/18375370>

Reviewed by Dean Jackson.

* win/tools/vsprops/common.props: Conditionally exclude the enhanced (SSE) instruction
set when building for 32-bit Windows. Otherwise, continue using SSE.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173768 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION(r173464): It broke the !ENABLE(PICTURE_SIZES) build
evab.u-szeged@partner.samsung.com [Fri, 19 Sep 2014 18:48:38 +0000 (18:48 +0000)]
REGRESSION(r173464): It broke the !ENABLE(PICTURE_SIZES) build
https://bugs.webkit.org/show_bug.cgi?id=136942

Reviewed by Simon Fraser.

* html/HTMLImageElement.cpp:
(WebCore::HTMLImageElement::parseAttribute):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173767 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoProvide a default argument for the most commonly used HitTestRequest variant
simon.fraser@apple.com [Fri, 19 Sep 2014 18:43:14 +0000 (18:43 +0000)]
Provide a default argument for the most commonly used HitTestRequest variant
https://bugs.webkit.org/show_bug.cgi?id=136653

Reviewed by Darin Adler.

Make the default HitTestRequest flags be ReadOnly | Active | DisallowShadowContent
and change call sites using the default flags to use a HitTestRequest temporary.

Source/WebCore:

* dom/TreeScope.cpp:
(WebCore::nodeFromPoint):
* editing/FrameSelection.cpp:
(WebCore::FrameSelection::contains):
* page/DragController.cpp:
(WebCore::elementUnderMouse):
* page/EventHandler.cpp:
(WebCore::EventHandler::handleMouseDraggedEvent):
(WebCore::EventHandler::handleMousePressEvent):
* rendering/HitTestRequest.h:
(WebCore::HitTestRequest::HitTestRequest):

Source/WebKit2:

* WebProcess/WebPage/ViewGestureGeometryCollector.cpp:
(WebKit::ViewGestureGeometryCollector::collectGeometryForSmartMagnificationGesture):
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::dynamicViewportSizeUpdate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173766 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAllow DOM methods to return references instead of pointers
cdumez@apple.com [Fri, 19 Sep 2014 18:33:16 +0000 (18:33 +0000)]
Allow DOM methods to return references instead of pointers
https://bugs.webkit.org/show_bug.cgi?id=136931

Source/WebCore:

Reviewed by Sam Weinig.

It is common practice in WebKit to have methods return a reference
instead of a pointer if the pointer can never be null. However, this
unfortunately did not work for DOM methods (functions called by JS
bindings). This prevented further refactoring.

This patch brings support for having DOM methods to return references
instead of pointers when the pointer cannot be null. The generated
bindings were calling WTF::getPtr() on the pointer type returned by
the implementation already (in case it was a smart pointer type).
This patch leverages this by having WTF::getPtr() convert reference
arguments into raw pointers.

This patch also updates a few DOM methods on Document and Element
classes to return a reference instead of a pointer, to test the change.
There are likely more DOM methods that can be updated though.

No new tests, no behavior change.

* accessibility/AccessibilityObject.cpp:
(WebCore::AccessibilityObject::classList):
* bindings/js/JSDOMBinding.h:
(WTF::getPtr): Deleted.
* dom/Document.cpp:
(WebCore::Document::implementation):
(WebCore::Document::webkitGetNamedFlows):
(WebCore::Document::namedFlows):
(WebCore::Document::setXMLVersion):
(WebCore::Document::setXMLStandalone):
(WebCore::Document::securityPolicy):
(WebCore::Document::styleSheets):
* dom/Document.h:
(WebCore::Document::timing):
* dom/Element.cpp:
(WebCore::Element::classList):
(WebCore::Element::dataset):
* dom/Element.h:
* html/shadow/MediaControlElements.cpp:
(WebCore::MediaControlPanelElement::setPosition):
(WebCore::MediaControlPanelElement::resetPosition):
(WebCore::MediaControlClosedCaptionsTrackListElement::updateDisplay):
* html/track/VTTRegion.cpp:
(WebCore::VTTRegion::displayLastTextTrackCueBox):
(WebCore::VTTRegion::willRemoveTextTrackCueBox):
* inspector/InspectorCSSAgent.cpp:
(WebCore::InspectorCSSAgent::getAllStyleSheets):
(WebCore::InspectorCSSAgent::getNamedFlowCollection):
* page/PerformanceTiming.cpp:
(WebCore::PerformanceTiming::documentTiming):
* rendering/FlowThreadController.cpp:
(WebCore::FlowThreadController::ensureRenderFlowThreadWithName):

Source/WTF:

Add support for having WTF::getPtr() transform reference arguments
into raw pointers so that DOM methods can now return references when
appropriate and so that the generated bindings code can handle this
via WTF::getPtr().

This patch had to alter the way getPtr() was overloaded for smart
pointer types so that we don't call &p on smart pointers but p.get().
This was needed because the new WTF::getPtr(T&) was being called for
RefPtr<T> arguments instead of the getPtr(const RefPtr<T>&) overload.
This was addressed using traits and template specialization to
distinguish WTF smart pointers from other types.

Reviewed by Sam Weinig.

* wtf/GetPtr.h:
(WTF::getPtr):
* wtf/OwnPtr.h:
(WTF::getPtr): Deleted.
* wtf/PassOwnPtr.h:
(WTF::getPtr): Deleted.
* wtf/PassRefPtr.h:
(WTF::getPtr): Deleted.
* wtf/Ref.h:
* wtf/RefPtr.h:
(WTF::getPtr): Deleted.
* wtf/gobject/GRefPtr.h:
(WTF::getPtr): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173765 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoVideos with controls enabled never receive 'dragstart' events.
jer.noble@apple.com [Fri, 19 Sep 2014 18:15:05 +0000 (18:15 +0000)]
Videos with controls enabled never receive 'dragstart' events.
https://bugs.webkit.org/show_bug.cgi?id=136837

Reviewed by Sam Weinig.

* media/video-controls-drag-expected.txt: Added.
* media/video-controls-drag.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173764 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAlways assume internal SDK when building configuration Production
dbates@webkit.org [Fri, 19 Sep 2014 18:10:32 +0000 (18:10 +0000)]
Always assume internal SDK when building configuration Production
https://bugs.webkit.org/show_bug.cgi?id=136925
<rdar://problem/18362399>

Reviewed by Dan Bernstein.

Source/bmalloc:

* Configurations/Base.xcconfig:

Source/JavaScriptCore:

As a side effect of this change we will always enable ENABLE_TOUCH_EVENTS, ENABLE_IOS_{GESTURE, TOUCH}_EVENTS,
and ENABLE_XSLT when either building configuration Production or building with the Internal SDK.

* Configurations/Base.xcconfig:

Source/ThirdParty/ANGLE:

* Configurations/Base.xcconfig:

Source/WebCore:

As a side effect of this change we will always enable ENABLE_TOUCH_EVENTS, ENABLE_IOS_{GESTURE, TOUCH}_EVENTS,
and ENABLE_XSLT when either building configuration Production or building with the Internal SDK.

* Configurations/Base.xcconfig:

Source/WebInspectorUI:

* Configurations/DebugRelease.xcconfig: Also move #include of HaveInternalSDK.xcconfig from DebugRelease.xcconfig
to Base.xcconfig so as to be consistent with the format of other Base.xcconfig/DebugRelease.xcconfig files
(e.g. Source/WebCore/Configurations/Base.xcconfig and Source/WebCore/Configurations/DebugRelease.xcconfig).

Source/WebKit/mac:

As a side effect of this change we will always enable ENABLE_TOUCH_EVENTS, ENABLE_IOS_{GESTURE, TOUCH}_EVENTS,
and ENABLE_XSLT when either building configuration Production or building with the Internal SDK.

* Configurations/Base.xcconfig:

Source/WebKit2:

As a side effect of this change we will always enable ENABLE_TOUCH_EVENTS, ENABLE_IOS_{GESTURE, TOUCH}_EVENTS,
and ENABLE_XSLT when either building configuration Production or building with the Internal SDK.

* Configurations/Base.xcconfig:

Source/WTF:

* Configurations/Base.xcconfig:

Tools:

* DumpRenderTree/mac/Configurations/DebugRelease.xcconfig:
* MiniBrowser/Configurations/DebugRelease.xcconfig:
* TestWebKitAPI/Configurations/DebugRelease.xcconfig:
* WebKitTestRunner/Configurations/DebugRelease.xcconfig:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173763 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoFix the Mac Mavericks build. Only iOS cares about fencing.
simon.fraser@apple.com [Fri, 19 Sep 2014 18:04:33 +0000 (18:04 +0000)]
Fix the Mac Mavericks build. Only iOS cares about fencing.

* Platform/mac/LayerHostingContext.h:
* Platform/mac/LayerHostingContext.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173762 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoSimple ES6 feature:String prototype additions
commit-queue@webkit.org [Fri, 19 Sep 2014 17:47:50 +0000 (17:47 +0000)]
Simple ES6 feature:String prototype additions
https://bugs.webkit.org/show_bug.cgi?id=131704

Patch by Diego Pino Garcia <dpino@igalia.com> on 2014-09-19
Reviewed by Darin Adler.

Source/JavaScriptCore:

* runtime/StringPrototype.cpp:
(JSC::StringPrototype::finishCreation):
(JSC::stringProtoFuncStartsWith): Added.
(JSC::stringProtoFuncEndsWith): Added.
(JSC::stringProtoFuncContains): Added.

Source/WTF:

* wtf/text/StringImpl.cpp:
(WTF::StringImpl::find):
(WTF::equalInner): Added.
(WTF::StringImpl::startsWith): Add implementation that supports
'startOffset' parameter.
(WTF::StringImpl::endsWith): Add implementation that supports
'endOffset' parameter.
* wtf/text/StringImpl.h:
* wtf/text/WTFString.h:
(WTF::String::contains): Modify current implementation to allow
setting a startOffset, 0 by default.
(WTF::String::startsWith):
(WTF::String::endsWith):

LayoutTests:

Test ES6 functions: string.startsWith(), string.endsWith() and
string.contains().

* js/Object-getOwnPropertyNames-expected.txt:
* js/script-tests/Object-getOwnPropertyNames.js:
* js/script-tests/string-contains.js: Added.
(stringToSearchIn.toString):
(startOffset.valueOf):
(matchString.toString):
(endOffset.valueOf):
* js/string-contains-expected.txt: Added.
* js/string-contains.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173761 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWTR and DRT didReceiveAuthenticationChallengeInFrame should print messages consistently
commit-queue@webkit.org [Fri, 19 Sep 2014 17:38:13 +0000 (17:38 +0000)]
WTR and DRT didReceiveAuthenticationChallengeInFrame should print messages consistently
https://bugs.webkit.org/show_bug.cgi?id=136847

Tools:

Added protection space host and port logging to WTR didReceiveAuthenticationChallengeInFrame.

Patch by Youenn Fablet <youenn.fablet@crf.canon.fr> on 2014-09-19
Reviewed by Alexey Proskuryakov.

* WebKitTestRunner/TestController.cpp:
(WTR::TestController::didReceiveAuthenticationChallengeInFrame):

LayoutTests:

Updated WK2 expected.txt files according new logging of didReceiveAuthenticationChallengeInFrame.
Factorized efl/gtk platform specific expected files as WK2 expected files.

Patch by Youenn Fablet <youenn.fablet@crf.canon.fr> on 2014-09-19
Reviewed by Alexey Proskuryakov.

* platform/gtk/http/tests/loading/basic-auth-resend-wrong-credentials-expected.txt: Removed.
* platform/gtk/http/tests/loading/basic-credentials-sent-automatically-expected.txt: Removed.
* platform/gtk/http/tests/media/video-auth-expected.txt: Removed.
* platform/gtk/http/tests/misc/401-alternative-content-expected.txt: Removed.
* platform/gtk/http/tests/misc/authentication-redirect-1/authentication-sent-to-redirect-cross-origin-expected.txt: Removed.
* platform/gtk/http/tests/misc/authentication-redirect-2/authentication-sent-to-redirect-same-origin-expected.txt: Removed.
* platform/gtk/http/tests/misc/authentication-redirect-4/authentication-sent-to-redirect-same-origin-url-expected.txt: Removed.
* platform/gtk/http/tests/security/401-logout/401-logout-expected.txt: Removed.
* platform/gtk/http/tests/xmlhttprequest/failed-auth-expected.txt: Removed.
* platform/gtk/http/tests/xmlhttprequest/remember-bad-password-expected.txt: Removed.
* platform/wk2/http/tests/loading/authentication-after-redirect-stores-wrong-credentials/authentication-after-redirect-stores-wrong-credentials-expected.txt:
* platform/wk2/http/tests/loading/basic-auth-resend-wrong-credentials-expected.txt:
* platform/wk2/http/tests/loading/basic-credentials-sent-automatically-expected.txt:
* platform/wk2/http/tests/media/video-auth-expected.txt: Renamed from LayoutTests/platform/efl/http/tests/media/video-auth-expected.txt.
* platform/wk2/http/tests/misc/401-alternative-content-expected.txt:
* platform/wk2/http/tests/misc/authentication-redirect-1/authentication-sent-to-redirect-cross-origin-expected.txt:
* platform/wk2/http/tests/misc/authentication-redirect-2/authentication-sent-to-redirect-same-origin-expected.txt:
* platform/wk2/http/tests/misc/authentication-redirect-4/authentication-sent-to-redirect-same-origin-url-expected.txt:
* platform/wk2/http/tests/security/401-logout/401-logout-expected.txt:
* platform/wk2/http/tests/xmlhttprequest/failed-auth-expected.txt:
* platform/wk2/http/tests/xmlhttprequest/remember-bad-password-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173760 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Missing inspector files in the tarball
carlosgc@webkit.org [Fri, 19 Sep 2014 15:58:13 +0000 (15:58 +0000)]
[GTK] Missing inspector files in the tarball
https://bugs.webkit.org/show_bug.cgi?id=136952

Reviewed by Gustavo Noronha Silva.

Source/WebKit2:

* PlatformGTK.cmake: Use APPEND instead of GLOB to add individual
and generated files to the list of inspector files.

Tools:

* gtk/manifest.txt: Add
Source/WebInspectorUI/Localizations/en.lproj/localizedStrings.js
to the tarball.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173755 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoOptimize MarkupAccumulator::appendText()
cdumez@apple.com [Fri, 19 Sep 2014 15:48:36 +0000 (15:48 +0000)]
Optimize MarkupAccumulator::appendText()
https://bugs.webkit.org/show_bug.cgi?id=136935

Reviewed by Benjamin Poulain.

MarkupAccumulator::appendText() was calling the generic appendNodeValue()
method to append the text. appendNodeValue() supports any kind of Node and
thus calls the Node::nodeValue() virtual function.

Since appendText() is the only caller to appendNodeValue(), this patch
inlines the code in appendText() and makes it specific to Text Node, and
thus more efficient. The code now calls CharacterData::data() instead of
the Node::nodeValue() virtual function. This patch also updates
CharacterData::data() to return a const reference to avoid copying the
return value.

No new tests, no behavior change.

* dom/CharacterData.h:
(WebCore::CharacterData::data):
* editing/MarkupAccumulator.cpp:
(WebCore::MarkupAccumulator::appendText):
(WebCore::MarkupAccumulator::appendNodeValue): Deleted.
* editing/MarkupAccumulator.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173754 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWebkitgtk 2.5.3: tarball contains APPLE LICENSE file - stating not distributable
carlosgc@webkit.org [Fri, 19 Sep 2014 15:40:04 +0000 (15:40 +0000)]
Webkitgtk 2.5.3: tarball contains APPLE LICENSE file - stating not distributable
https://bugs.webkit.org/show_bug.cgi?id=136632

Reviewed by Gustavo Noronha Silva.

Source/WebInspectorUI:

Add free icons for the GTK port. These icons are from the GNOME
project, some of them adapted by Andres Gomez for the inspector.
This is only a subset of the icons, since we still don't have a
free replacement for all of them, but at least it makes the
inspector kind of usable.

* UserInterface/Images/gtk/AUTHORS: Added.
* UserInterface/Images/gtk/BackArrow.svg: Added.
* UserInterface/Images/gtk/BreakpointActionAdd.svg: Added.
* UserInterface/Images/gtk/BreakpointActionRemove.svg: Added.
* UserInterface/Images/gtk/BreakpointButton.svg: Added.
* UserInterface/Images/gtk/BreakpointInactiveButton.svg: Added.
* UserInterface/Images/gtk/Breakpoints.svg: Added.
* UserInterface/Images/gtk/COPYING: Added.
* UserInterface/Images/gtk/Close.svg: Added.
* UserInterface/Images/gtk/CloseWhite.svg: Added.
* UserInterface/Images/gtk/ContentFlow.svg: Added.
* UserInterface/Images/gtk/Crosshair.svg: Added.
* UserInterface/Images/gtk/DOMCharacterData.svg: Added.
* UserInterface/Images/gtk/DOMComment.svg: Added.
* UserInterface/Images/gtk/DOMDocument.svg: Added.
* UserInterface/Images/gtk/DOMDocumentType.svg: Added.
* UserInterface/Images/gtk/DOMElement.svg: Added.
* UserInterface/Images/gtk/DOMNode.svg: Added.
* UserInterface/Images/gtk/DOMTextNode.svg: Added.
* UserInterface/Images/gtk/DOMTree.svg: Added.
* UserInterface/Images/gtk/DisclosureTriangleSmallClosed.svg: Added.
* UserInterface/Images/gtk/DisclosureTriangleSmallOpen.svg: Added.
* UserInterface/Images/gtk/DisclosureTriangleTinyClosed.svg: Added.
* UserInterface/Images/gtk/DisclosureTriangleTinyOpen.svg: Added.
* UserInterface/Images/gtk/DockBottom.svg: Added.
* UserInterface/Images/gtk/DockRight.svg: Added.
* UserInterface/Images/gtk/DocumentCSS.png: Added.
* UserInterface/Images/gtk/DocumentCSS@2x.png: Added.
* UserInterface/Images/gtk/DocumentFont.png: Added.
* UserInterface/Images/gtk/DocumentFont@2x.png: Added.
* UserInterface/Images/gtk/DocumentGeneric.png: Added.
* UserInterface/Images/gtk/DocumentGeneric@2x.png: Added.
* UserInterface/Images/gtk/DocumentImage.png: Added.
* UserInterface/Images/gtk/DocumentImage@2x.png: Added.
* UserInterface/Images/gtk/DocumentJS.png: Added.
* UserInterface/Images/gtk/DocumentJS@2x.png: Added.
* UserInterface/Images/gtk/DocumentMarkup.png: Added.
* UserInterface/Images/gtk/DocumentMarkup@2x.png: Added.
* UserInterface/Images/gtk/DownloadArrow.svg: Added.
* UserInterface/Images/gtk/Error.svg: Added.
* UserInterface/Images/gtk/Errors.svg: Added.
* UserInterface/Images/gtk/ErrorsEnabled.svg: Added.
* UserInterface/Images/gtk/EventListener.svg: Added.
* UserInterface/Images/gtk/Exception.svg: Added.
* UserInterface/Images/gtk/FilterFieldGlyph.svg: Added.
* UserInterface/Images/gtk/FolderGeneric.png: Added.
* UserInterface/Images/gtk/FolderGeneric@2x.png: Added.
* UserInterface/Images/gtk/ForwardArrow.svg: Added.
* UserInterface/Images/gtk/Function.svg: Added.
* UserInterface/Images/gtk/GoToArrow.svg: Added.
* UserInterface/Images/gtk/Resources.svg: Added.
* UserInterface/Images/gtk/SortIndicatorDownArrow.svg: Added.
* UserInterface/Images/gtk/SortIndicatorUpArrow.svg: Added.
* UserInterface/Images/gtk/UpDownArrows.svg: Added.

Source/WebKit2:

* PlatformGTK.cmake: Use only the icons from the gtk directory since they are free.

Tools:

* gtk/generate-inspector-gresource-manifest.py: Use an alias for
the gtk icons.
* gtk/manifest.txt: Only include the free icons in the tarball.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173753 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] generate-inspector-gresource-manifest.py should receive the list of files from...
carlosgc@webkit.org [Fri, 19 Sep 2014 09:21:16 +0000 (09:21 +0000)]
[GTK] generate-inspector-gresource-manifest.py should receive the list of files from the makefile
https://bugs.webkit.org/show_bug.cgi?id=136943

Reviewed by Philippe Normand.

Source/WebKit2:

* PlatformGTK.cmake: Fix the inspector file list and pass it to
the generate-inspector-gresource-manifest.py script.

Tools:

Update the script to receive the list of files from the command
line instead of scanning the given directories.

* gtk/generate-inspector-gresource-manifest.py:
(get_filenames): Convert filenames to use relative paths.
(find_all_files_in_directory): Deleted.
(find_all_files_in_directory.select_file): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173751 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[EFL][WK2] Minibrowser : Support Zoom on Control key + scroll
commit-queue@webkit.org [Fri, 19 Sep 2014 06:58:54 +0000 (06:58 +0000)]
[EFL][WK2] Minibrowser : Support Zoom on Control key + scroll
https://bugs.webkit.org/show_bug.cgi?id=136739

Patch by Tanay C <tanay.c@samsung.com> on 2014-09-18
Reviewed by Gyuyoung Kim.

* MiniBrowser/efl/main.c: Add support for zoom on control key + mouse scroll
(on_mouse_wheel): Modified.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173750 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Dot not allow to create delete-on-destroy GMainLoopSources
carlosgc@webkit.org [Fri, 19 Sep 2014 06:51:24 +0000 (06:51 +0000)]
[GTK] Dot not allow to create delete-on-destroy GMainLoopSources
https://bugs.webkit.org/show_bug.cgi?id=136923

Reviewed by Gustavo Noronha Silva.

Source/WebCore:

* platform/gtk/GtkDragAndDropHelper.cpp:
(WebCore::GtkDragAndDropHelper::handleDragLeave): Use GMainLoopSource::scheduleAndDeleteOnDestroy().

Source/WebKit2:

* Platform/gtk/WorkQueueGtk.cpp:
(WorkQueue::dispatch): Use GMainLoopSource::scheduleAndDeleteOnDestroy().
(WorkQueue::dispatchAfter): Use GMainLoopSource::scheduleAfterDelayAndDeleteOnDestroy().

Source/WTF:

We have several asserts to ensure that delete-on-destroy sources
are not misused, like not scheduling socket sources on a
delete-on-destroy GMainLoopSource or not allowing to cancel them
before they have been dispatched. It's better to ensure all those
things at compile time, using static methods to schedule sources
creating a delete-on-destroy GMainLoopSource that is not returned
to the user.

* wtf/gobject/GMainLoopSource.cpp:
(WTF::GMainLoopSource::create): Private static method to create a
delete-on-destroy GMainLoopSource.
(WTF::GMainLoopSource::cancelWithoutLocking): Return early in case
of delete-on-destroy source, since they can't be cancelled.
(WTF::GMainLoopSource::schedule): Remove assertion to ensure
socket sources are not scheduled on a delete-on-destroy GMainLoopSource.
(WTF::GMainLoopSource::scheduleAndDeleteOnDestroy):
(WTF::GMainLoopSource::scheduleAfterDelayAndDeleteOnDestroy):
(WTF::GMainLoopSource::createAndDeleteOnDestroy): Deleted.
* wtf/gobject/GMainLoopSource.h:
* wtf/gtk/MainThreadGtk.cpp:
(WTF::scheduleDispatchFunctionsOnMainThread): Use GMainLoopSource::scheduleAndDeleteOnDestroy()
* wtf/gtk/RunLoopGtk.cpp:
(WTF::RunLoop::wakeUp): Ditto.

Tools:

* TestWebKitAPI/Tests/WTF/gobject/GMainLoopSource.cpp:
(TestWebKitAPI::TEST): Use the new API that doesn't allow to use
the source.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173749 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[EFL][CoordinatedGraphics] Remove setBackgroundColor message
ryuan.choi@gmail.com [Fri, 19 Sep 2014 04:41:49 +0000 (04:41 +0000)]
[EFL][CoordinatedGraphics] Remove setBackgroundColor message
https://bugs.webkit.org/show_bug.cgi?id=136824

Reviewed by Gyuyoung Kim.

CoordinatedGraphics used setBackgroundColor() to draw the background of contents
although tiles are not ready.

This patch removes setBackgroundColor() and uses pageExtendedBackgroundColor()
which provides similar functionality.

* UIProcess/API/efl/EwkView.cpp:
Enables or disables WKPageSetBackgroundExtendsBeyondPage to use pageExtendedBackgroundColor()
(EwkView::EwkView):
(EwkView::setBackgroundColor):
* UIProcess/CoordinatedGraphics/CoordinatedGraphicsScene.cpp:
(WebCore::CoordinatedGraphicsScene::CoordinatedGraphicsScene):
(WebCore::CoordinatedGraphicsScene::paintToCurrentGLContext): Passed pageExtendedBackgroundColor.
(WebCore::CoordinatedGraphicsScene::paintToGraphicsContext): Ditto.
(WebCore::CoordinatedGraphicsScene::setBackgroundColor): Deleted.
* UIProcess/CoordinatedGraphics/CoordinatedGraphicsScene.h:
(WebCore::CoordinatedGraphicsScene::setDrawsBackground): Deleted.
* UIProcess/CoordinatedGraphics/CoordinatedLayerTreeHostProxy.cpp:
(WebKit::CoordinatedLayerTreeHostProxy::setBackgroundColor): Deleted.
* UIProcess/CoordinatedGraphics/CoordinatedLayerTreeHostProxy.h:
* UIProcess/CoordinatedGraphics/CoordinatedLayerTreeHostProxy.messages.in:
* UIProcess/CoordinatedGraphics/WebView.cpp:
(WebKit::WebView::paintToCurrentGLContext):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::pageExtendedBackgroundColorDidChange):
Refactored to use const reference for the Color parameter.
* UIProcess/WebPageProxy.h:
* UIProcess/efl/WebViewEfl.cpp:
(WebKit::WebViewEfl::paintToCairoSurface):
* WebProcess/WebCoreSupport/WebChromeClient.cpp:
(WebKit::WebChromeClient::pageExtendedBackgroundColorDidChange):
Added PLATFORM(EFL) to pass PageExtendedBackgroundColorDidChange message.
* WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.cpp:
(WebKit::CoordinatedLayerTreeHost::setBackgroundColor): Deleted.
* WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.h:
* WebProcess/WebPage/LayerTreeHost.h:
(WebKit::LayerTreeHost::setBackgroundColor): Deleted.
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::mainFrameDidLayout):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173748 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Windows] Media tests all fail because they cannot find localized strings.
roger_fong@apple.com [Fri, 19 Sep 2014 01:47:20 +0000 (01:47 +0000)]
[Windows] Media tests all fail because they cannot find localized strings.

* rendering/RenderThemeSafari.cpp:
(WebCore::RenderThemeSafari::mediaControlsScript):
Make sure to use localized strings files in RenderThemeSafari.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173746 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemove generator/yield terminology in RenderInline's GeneratorContext
simon.fraser@apple.com [Fri, 19 Sep 2014 01:01:39 +0000 (01:01 +0000)]
Remove generator/yield terminology in RenderInline's GeneratorContext
https://bugs.webkit.org/show_bug.cgi?id=119278

Reviewed by Sam Weinig.

Get rid of the icky Generator/yield syntax used for collecting rects in RenderInline.
Just replace the operator() with an explicit addRect().

Also some drive-by nullptr cleanup, and removing default arguments from
addChildIgnoringContinuation() calls.

* rendering/RenderInline.cpp:
(WebCore::RenderInline::willBeDestroyed):
(WebCore::RenderInline::styleDidChange):
(WebCore::RenderInline::splitInlines):
(WebCore::RenderInline::addChildToContinuation):
(WebCore::RenderInline::generateLineBoxRects):
(WebCore::RenderInline::generateCulledLineBoxRects):
* rendering/RenderInline.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173743 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoImprove fullscreen video rotation animation.
commit-queue@webkit.org [Thu, 18 Sep 2014 23:40:46 +0000 (23:40 +0000)]
Improve fullscreen video rotation animation.
https://bugs.webkit.org/show_bug.cgi?id=136870

Patch by Jeremy Jones <jeremyj@apple.com> on 2014-09-18
Reviewed by Simon Fraser.

Source/WebCore:

Instead of setting the frame on the video layer, set position and bounds separately. This allows the position to be synchronized with the rest of the animation.
When using a fencePort to synchronize animations, if the fence times out, pivoting around the center provides a better fallback.

* platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
(-[WebAVVideoLayer setVideoSublayer:]): added
(-[WebAVVideoLayer videoSublayer]): added
(-[WebAVVideoLayer setBounds:]):  set position and bounds insted of frame.
(WebVideoFullscreenInterfaceAVKit::setupFullscreen): use setVideoSublayer
* platform/ios/WebVideoFullscreenModelVideoElement.mm:
(WebVideoFullscreenModelVideoElement::setVideoFullscreenLayer): set bounds and anchorPoint instead of frame
(WebVideoFullscreenModelVideoElement::setVideoLayerFrame): set bounds instead of frame.

Source/WebKit2:

Remove WKCAContext from LayerHostingContext and use a fencePort to synchronize animation between processes.

* Platform/mac/LayerHostingContext.h: add setFencePort. Remove WKCAContext.
* Platform/mac/LayerHostingContext.mm:
(WebKit::LayerHostingContext::createForPort): Remove WKCAContext.
(WebKit::LayerHostingContext::createForExternalHostingProcess): ditto
(WebKit::LayerHostingContext::setRootLayer): ditto
(WebKit::LayerHostingContext::rootLayer): ditto
(WebKit::LayerHostingContext::contextID): ditto
(WebKit::LayerHostingContext::invalidate): ditto
(WebKit::LayerHostingContext::setColorSpace): ditto
(WebKit::LayerHostingContext::colorSpace): ditto
(WebKit::LayerHostingContext::setFencePort): added
* UIProcess/ios/WebVideoFullscreenManagerProxy.mm:
(WebKit::WebVideoFullscreenManagerProxy::setVideoLayerFrame):
* WebProcess/ios/WebVideoFullscreenManager.h:
* WebProcess/ios/WebVideoFullscreenManager.messages.in: add fence port parameter
* WebProcess/ios/WebVideoFullscreenManager.mm: ditto
(WebKit::WebVideoFullscreenManager::didSetupFullscreen): ditto
(WebKit::WebVideoFullscreenManager::setVideoLayerFrameFenced): ditto

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173741 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoExpand the srcset demo to have 3x and 4x images.
dino@apple.com [Thu, 18 Sep 2014 22:58:51 +0000 (22:58 +0000)]
Expand the srcset demo to have 3x and 4x images.

Unreviewed.

* demos/srcset/image-1x.png:
* demos/srcset/image-2x.png:
* demos/srcset/image-3x.png: Added.
* demos/srcset/image-4x.png: Added.
* demos/srcset/index.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173740 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoText laid out with the SVG -> OTF font converter does not have the same metrics as...
mmaxfield@apple.com [Thu, 18 Sep 2014 22:52:21 +0000 (22:52 +0000)]
Text laid out with the SVG -> OTF font converter does not have the same metrics as with the SVG font code path
https://bugs.webkit.org/show_bug.cgi?id=136907

Reviewed by Darin Adler.

There are three things that are causing text laid out with the SVG -> OTF font converter to not have metrics that
measure our existing SVG codepath. They are:

1. Creating a font with a 0 ascent or descent makes OS X think that something is wrong with the font, and take
a different codepath when trying to parse ascent and descent information. This patch checks for this condition
and sets the ascent/descent to 1 FUnit instead (which is generally much smaller than a pixel).
2. Our SVG font codepath hardcodes a line gap of 1/10th of the font size for every font. This patch makes the
font converter obey this.
3. The converter was not allowing for default glyph advances as per the SVG font specification. This patch
does so.

No new tests yet, but they will come soon! I promise!

* svg/SVGToOTFFontConversion.cpp:
(WebCore::SVGToOTFFontConverter::appendHHEATable):
(WebCore::SVGToOTFFontConverter::SVGToOTFFontConverter):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173739 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoREGRESSION (r168921): SVG elements may be unnecessarily rebuilt
dbates@webkit.org [Thu, 18 Sep 2014 22:49:24 +0000 (22:49 +0000)]
REGRESSION (r168921): SVG elements may be unnecessarily rebuilt
https://bugs.webkit.org/show_bug.cgi?id=136373

Reviewed by David Kilzer.

Source/WebCore:

Fixes an issue where a SVG element may be unnecessarily rebuilt. In particular,
it is unnecessary to rebuild a SVG element that will be removed from the document.

Test: svg/custom/remove-subtree-including-use-with-child-textpath-that-references-earlier-path.html

* svg/SVGDocumentExtensions.cpp:
(WebCore::SVGDocumentExtensions::removeAllElementReferencesForTarget): Modified to remove a SVG element
from the list of elements to be rebuilt.

LayoutTests:

Add a test to ensure that we don't rebuild a SVG element that will be removed from the document.

* svg/custom/remove-subtree-including-use-with-child-textpath-that-references-earlier-path-expected.txt: Added.
* svg/custom/remove-subtree-including-use-with-child-textpath-that-references-earlier-path.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173738 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAvoid divide by zero when we get an empty rect while snapshotting a node
simon.fraser@apple.com [Thu, 18 Sep 2014 22:31:43 +0000 (22:31 +0000)]
Avoid divide by zero when we get an empty rect while snapshotting a node
https://bugs.webkit.org/show_bug.cgi?id=136932
rdar://problem/18383473

Reviewed by Dean Jackson.

renderer()->paintingRootRect() can return an empty rect for some inline configurations,
for example an <a> containing a floated <span>

When this occurs, avoid a divide by zero and return a null snapshot image.

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::snapshotNode):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173737 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed rollout r173731. Broke multiple builds.
joepeck@webkit.org [Thu, 18 Sep 2014 22:06:26 +0000 (22:06 +0000)]
Unreviewed rollout r173731. Broke multiple builds.

Source/JavaScriptCore:

* inspector/JSGlobalObjectInspectorController.cpp:
(Inspector::JSGlobalObjectInspectorController::JSGlobalObjectInspectorController):
(Inspector::JSGlobalObjectInspectorController::connectFrontend):
* inspector/JSGlobalObjectInspectorController.h:
* inspector/remote/RemoteInspector.h:
* inspector/remote/RemoteInspector.mm:
(Inspector::RemoteInspector::RemoteInspector):
(Inspector::RemoteInspector::setupFailed):
(Inspector::RemoteInspector::start):
(Inspector::RemoteInspector::stopInternal):
(Inspector::RemoteInspector::setupXPCConnectionIfNeeded):
(Inspector::RemoteInspector::xpcConnectionReceivedMessage):
(Inspector::RemoteInspector::xpcConnectionFailed):
(Inspector::RemoteInspector::receivedSetupMessage):
(Inspector::globalAutomaticInspectionState): Deleted.
(Inspector::RemoteInspector::updateDebuggableAutomaticInspectCandidate): Deleted.
(Inspector::RemoteInspector::sendAutomaticInspectionCandidateMessage): Deleted.
(Inspector::RemoteInspector::setupSucceeded): Deleted.
(Inspector::RemoteInspector::waitingForAutomaticInspection): Deleted.
(Inspector::RemoteInspector::receivedAutomaticInspectionConfigurationMessage): Deleted.
(Inspector::RemoteInspector::receivedAutomaticInspectionRejectMessage): Deleted.
* inspector/remote/RemoteInspectorConstants.h:
* inspector/remote/RemoteInspectorDebuggable.cpp:
(Inspector::RemoteInspectorDebuggable::setRemoteDebuggingAllowed):
(Inspector::RemoteInspectorDebuggable::pauseWaitingForAutomaticInspection): Deleted.
* inspector/remote/RemoteInspectorDebuggable.h:
* inspector/remote/RemoteInspectorDebuggableConnection.h:
* inspector/remote/RemoteInspectorDebuggableConnection.mm:
(Inspector::RemoteInspectorDebuggableConnection::setup):
* runtime/JSGlobalObjectDebuggable.cpp:
(JSC::JSGlobalObjectDebuggable::connect):
(JSC::JSGlobalObjectDebuggable::pauseWaitingForAutomaticInspection): Deleted.
* runtime/JSGlobalObjectDebuggable.h:

Source/WebCore:

* WebCore.exp.in:
* inspector/InspectorController.cpp:
(WebCore::InspectorController::connectFrontend):
(WebCore::InspectorController::show):
* inspector/InspectorController.h:
* page/PageDebuggable.cpp:
(WebCore::PageDebuggable::connect):
* page/PageDebuggable.h:
* testing/Internals.cpp:
(WebCore::Internals::openDummyInspectorFrontend):

Source/WTF:

* wtf/Assertions.cpp:
* wtf/Assertions.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173736 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoASSERT in RenderBox::instrinsicScrollbarLogicalWidth opening the inspector
bdakin@apple.com [Thu, 18 Sep 2014 21:37:39 +0000 (21:37 +0000)]
ASSERT in RenderBox::instrinsicScrollbarLogicalWidth opening the inspector
https://bugs.webkit.org/show_bug.cgi?id=136929

Reviewed by Simon Fraser.

This is a regression from http://trac.webkit.org/changeset/173668
This code is wrong after that change because it assumes that overflow:scroll will
always have a scrollbar. That is no longer the case. We need to check to make sure
it’s an overflow:scroll that does NOT have overflow:auto behavior.

* rendering/RenderBox.cpp:
(WebCore::RenderBox::instrinsicScrollbarLogicalWidth):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173734 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Should be able to attach a debugger to a JSContext before anything...
commit-queue@webkit.org [Thu, 18 Sep 2014 19:47:55 +0000 (19:47 +0000)]
Web Inspector: Should be able to attach a debugger to a JSContext before anything is executed
https://bugs.webkit.org/show_bug.cgi?id=136893

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-09-18
Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

Adds new remote inspector protocol handling for automatic inspection.
Debuggers can signal they have enabled automatic inspection, and
when debuggables are created the current application will pause to
see if the debugger will inspect or decline to inspect the debuggable.

* inspector/remote/RemoteInspectorConstants.h:
* inspector/remote/RemoteInspector.h:
* inspector/remote/RemoteInspector.mm:
(Inspector::globalAutomaticInspectionState):
(Inspector::RemoteInspector::RemoteInspector):
(Inspector::RemoteInspector::start):
When first starting, check the global "is there an auto-inspect" debugger state.
This is necessary so that the current application knows if it should pause or
not when a debuggable is created, even without having connected to webinspectord yet.

(Inspector::RemoteInspector::updateDebuggableAutomaticInspectCandidate):
When a debuggable has enabled remote inspection, take this path to propose
it as an automatic inspection candidate if there is an auto-inspect debugger.

(Inspector::RemoteInspector::sendAutomaticInspectionCandidateMessage):
Send the automatic inspection candidate message.

(Inspector::RemoteInspector::receivedSetupMessage):
(Inspector::RemoteInspector::setupFailed):
(Inspector::RemoteInspector::setupSucceeded):
After attempting to open an inspector, unpause if it was for the
automatic inspection candidate.

(Inspector::RemoteInspector::waitingForAutomaticInspection):
When running a nested runloop, check if we should remain paused.

(Inspector::RemoteInspector::setupXPCConnectionIfNeeded):
If by the time we connect to webinspectord we have a candidate, then
immediately send the candidate message.

(Inspector::RemoteInspector::stopInternal):
(Inspector::RemoteInspector::xpcConnectionFailed):
In error cases, clear our state.

(Inspector::RemoteInspector::xpcConnectionReceivedMessage):
(Inspector::RemoteInspector::receivedAutomaticInspectionConfigurationMessage):
(Inspector::RemoteInspector::receivedAutomaticInspectionRejectMessage):
Update state when receiving new messages.

* inspector/remote/RemoteInspectorDebuggable.h:
* inspector/remote/RemoteInspectorDebuggable.cpp:
(Inspector::RemoteInspectorDebuggable::setRemoteDebuggingAllowed):
Special case when a debuggable is newly allowed to be debuggable.

(Inspector::RemoteInspectorDebuggable::pauseWaitingForAutomaticInspection):
Run a nested run loop while this is an automatic inspection candidate.

* inspector/JSGlobalObjectInspectorController.h:
* inspector/JSGlobalObjectInspectorController.cpp:
(Inspector::JSGlobalObjectInspectorController::JSGlobalObjectInspectorController):
(Inspector::JSGlobalObjectInspectorController::connectFrontend):
When the inspector starts via automatic inspection automatically pause.
We plan on removing this condition by having the frontend signal to the
backend when it is completely initialized.

* inspector/remote/RemoteInspectorDebuggableConnection.h:
* inspector/remote/RemoteInspectorDebuggableConnection.mm:
(Inspector::RemoteInspectorDebuggableConnection::setup):
Pass on the flag of whether or not this was automatic inspection.

* runtime/JSGlobalObjectDebuggable.h:
* runtime/JSGlobalObjectDebuggable.cpp:
(JSC::JSGlobalObjectDebuggable::connect):
(JSC::JSGlobalObjectDebuggable::pauseWaitingForAutomaticInspection):
When pausing in a JSGlobalObject we need to release the API lock.

Source/WebCore:

Automatic inspection is currently disabled for web pages.
This just updates the interfaces that changed.

* WebCore.exp.in:
* inspector/InspectorController.cpp:
(WebCore::InspectorController::connectFrontend):
(WebCore::InspectorController::show):
* inspector/InspectorController.h:
* page/PageDebuggable.cpp:
(WebCore::PageDebuggable::connect):
* page/PageDebuggable.h:
* testing/Internals.cpp:
(WebCore::Internals::openDummyInspectorFrontend):

Source/WTF:

Currently automatic inspection only happens in processes that have a
debugger attached. That condition may change in the future, but this
function can stand on its own in WTF. It may be useful in the future
to perhaps continue though ASSERTs if you have a debugger attached.

* wtf/Assertions.cpp:
* wtf/Assertions.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173731 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoCSS JIT: The backtracking register can be ignored from the minimumRegisterRequirements
benjamin@webkit.org [Thu, 18 Sep 2014 19:36:36 +0000 (19:36 +0000)]
CSS JIT: The backtracking register can be ignored from the minimumRegisterRequirements
https://bugs.webkit.org/show_bug.cgi?id=136906
rdar://problem/18368294

Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-09-18
Reviewed by Darin Adler.

* fast/selectors/nth-child-with-backtracking-expected.txt:
* fast/selectors/nth-child-with-backtracking.html:
The use case is well covered by existing tests. This is updating the backtracking test
with a simpler selector similar to the one in the bug report.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173730 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoDon’t skip all the media tests on Windows.
roger_fong@apple.com [Thu, 18 Sep 2014 19:33:30 +0000 (19:33 +0000)]
Don’t skip all the media tests on Windows.
* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173729 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed. Windows rebaselining to account for subpixel layout Part 3.
roger_fong@apple.com [Thu, 18 Sep 2014 19:12:18 +0000 (19:12 +0000)]
Unreviewed. Windows rebaselining to account for subpixel layout Part 3.
* platform/win/css1/box_properties/acid_test-expected.txt: Added.
* platform/win/fast/block/margin-collapse: Added.
* platform/win/fast/block/margin-collapse/103-expected.txt: Added.
* platform/win/fast/css/empty-pseudo-class-expected.txt:
* platform/win/fast/css/first-child-pseudo-class-expected.txt:
* platform/win/fast/css/last-child-pseudo-class-expected.txt:
* platform/win/fast/css/only-child-pseudo-class-expected.txt:
* platform/win/fast/forms/basic-inputs-expected.txt: Added.
* platform/win/fast/forms/file/file-input-direction-expected.txt:
* platform/win/fast/forms/formmove-expected.txt: Added.
* platform/win/fast/forms/formmove2-expected.txt: Added.
* platform/win/fast/repaint/reflection-redraw-expected.txt:
* platform/win/fast/replaced/width100percent-button-expected.txt:
* platform/win/fast/ruby/bopomofo-expected.txt: Added.
* platform/win/fast/ruby/bopomofo-letter-spacing-expected.txt: Added.
* platform/win/fast/ruby/bopomofo-rl-expected.txt: Added.
* platform/win/fast/table/multiple-captions-display-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173728 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Win] Unreviewed build fix.
bfulgham@apple.com [Thu, 18 Sep 2014 18:10:13 +0000 (18:10 +0000)]
[Win] Unreviewed build fix.

* gtest/msvc/gtest-md.vcxproj: Make sure we use static linking for gtest on
normal debug build to match the rest of WebKit.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173727 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoIncorrect pending API request URL when loading an HTML string
carlosgc@webkit.org [Thu, 18 Sep 2014 17:36:45 +0000 (17:36 +0000)]
Incorrect pending API request URL when loading an HTML string
https://bugs.webkit.org/show_bug.cgi?id=136913

Reviewed by Alexey Proskuryakov.

Use about:blank as pending API request URL when the given base URL
is empty, otherwise use the given base URL.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::loadHTMLString):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173726 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoFix countFailures of RunLLINTCLoopTests and Run32bitJSCTests (II)
ossy@webkit.org [Thu, 18 Sep 2014 17:15:01 +0000 (17:15 +0000)]
Fix countFailures of RunLLINTCLoopTests and Run32bitJSCTests (II)
https://bugs.webkit.org/show_bug.cgi?id=136856

Reviewed by Darin Adler.

* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(RunLLINTCLoopTests.countFailures):
(Run32bitJSCTests.countFailures):
* BuildSlaveSupport/build.webkit.org-config/mastercfg_unittest.py:
(RunLLINTCLoopTestsTest.test_failures):
(RunLLINTCLoopTestsTest.test_failure):
(RunLLINTCLoopTestsTest.test_no_failure):
(Run32bitJSCTestsTest.test_failures):
(Run32bitJSCTestsTest.test_failure):
(Run32bitJSCTestsTest.test_no_failure):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173725 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUse fastHasAttribute() / fastGetAttribute() when possible
cdumez@apple.com [Thu, 18 Sep 2014 16:52:10 +0000 (16:52 +0000)]
Use fastHasAttribute() / fastGetAttribute() when possible
https://bugs.webkit.org/show_bug.cgi?id=136900

Reviewed by Benjamin Poulain.

Use fastHasAttribute() / fastGetAttribute() when possible, that is when
the attribute is not SVG-animated or the |style| attribute, to avoid
synchronizing attributes unnecessarily.

Also avoid calling hasAttribute(xxxAttr) then getAttribute(xxxAttr) as
it causes 2 linear searches. It is best to call getAttribute(xxxAttr)
directly and check if it returns the nullAtom.

No new tests, no behavior change. Element::fastAttributeLookupAllowed()
is there as a safety set for debug builds.

* accessibility/AccessibilityListBoxOption.cpp:
(WebCore::AccessibilityListBoxOption::isEnabled):
* accessibility/AccessibilityNodeObject.cpp:
(WebCore::AccessibilityNodeObject::alternativeTextForWebArea):
(WebCore::AccessibilityNodeObject::hierarchicalLevel):
* accessibility/AccessibilityObject.cpp:
(WebCore::AccessibilityObject::contentEditableAttributeIsEnabled):
* dom/Document.cpp:
(WebCore::Document::hasManifest):
* dom/DocumentOrderedMap.cpp:
(WebCore::keyMatchesLabelForAttribute):
* dom/DocumentStyleSheetCollection.cpp:
(WebCore::DocumentStyleSheetCollection::collectActiveStyleSheets):
* dom/Element.cpp:
(WebCore::Element::imageSourceURL):
(WebCore::Element::pseudo):
(WebCore::Element::spellcheckAttributeState):
* editing/ApplyStyleCommand.cpp:
(WebCore::ApplyStyleCommand::removeEmbeddingUpToEnclosingBlock):
* editing/EditingStyle.cpp:
(WebCore::HTMLAttributeEquivalent::attributeValueAsCSSValue):
(WebCore::HTMLFontSizeEquivalent::attributeValueAsCSSValue):
* editing/SplitElementCommand.cpp:
(WebCore::SplitElementCommand::doUnapply):
* html/HTMLAnchorElement.cpp:
(WebCore::HTMLAnchorElement::draggable):
(WebCore::HTMLAnchorElement::sendPings):
* html/HTMLAppletElement.cpp:
(WebCore::HTMLAppletElement::updateWidget):
* html/HTMLBodyElement.cpp:
(WebCore::HTMLBodyElement::addSubresourceAttributeURLs):
* html/HTMLButtonElement.cpp:
(WebCore::HTMLButtonElement::value):
* html/HTMLDocument.cpp:
(WebCore::HTMLDocument::dir):
* html/HTMLEmbedElement.cpp:
(WebCore::HTMLEmbedElement::imageSourceURL):
(WebCore::HTMLEmbedElement::addSubresourceAttributeURLs):
* html/HTMLFieldSetElement.cpp:
(WebCore::HTMLFieldSetElement::~HTMLFieldSetElement):
(WebCore::updateFromControlElementsAncestorDisabledStateUnder):
(WebCore::HTMLFieldSetElement::disabledAttributeChanged):
(WebCore::HTMLFieldSetElement::disabledStateChanged):
(WebCore::HTMLFieldSetElement::childrenChanged):
(WebCore::HTMLFieldSetElement::didMoveToNewDocument):
* html/HTMLFormControlElement.cpp:
(WebCore::HTMLFormControlElement::computeIsDisabledByFieldsetAncestor):
* html/HTMLFormElement.cpp:
(WebCore::HTMLFormElement::action):
* html/HTMLFrameElement.cpp:
(WebCore::HTMLFrameElement::noResize):
* html/HTMLFrameElementBase.cpp:
(WebCore::HTMLFrameElementBase::location):
* html/HTMLHRElement.cpp:
(WebCore::HTMLHRElement::collectStyleForPresentationAttribute):
* html/HTMLHtmlElement.cpp:
(WebCore::HTMLHtmlElement::insertedByParser):
* html/HTMLImageElement.cpp:
(WebCore::HTMLImageElement::altText):
(WebCore::HTMLImageElement::alt):
(WebCore::HTMLImageElement::draggable):
(WebCore::HTMLImageElement::src):
(WebCore::HTMLImageElement::addSubresourceAttributeURLs):
* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::finishParsingChildren):
(WebCore::HTMLInputElement::altText):
(WebCore::HTMLInputElement::reset):
* html/HTMLLabelElement.cpp:
(WebCore::HTMLLabelElement::control):
* html/HTMLLinkElement.cpp:
(WebCore::HTMLLinkElement::process):
(WebCore::HTMLLinkElement::rel):
* html/HTMLMetaElement.cpp:
(WebCore::HTMLMetaElement::content):
(WebCore::HTMLMetaElement::httpEquiv):
* html/HTMLObjectElement.cpp:
(WebCore::HTMLObjectElement::imageSourceURL):
(WebCore::HTMLObjectElement::addSubresourceAttributeURLs):
* html/HTMLOptGroupElement.cpp:
(WebCore::HTMLOptGroupElement::groupLabelText):
* html/HTMLProgressElement.cpp:
(WebCore::HTMLProgressElement::max):
* html/HTMLScriptElement.cpp:
(WebCore::HTMLScriptElement::sourceAttributeValue):
(WebCore::HTMLScriptElement::charsetAttributeValue):
(WebCore::HTMLScriptElement::languageAttributeValue):
(WebCore::HTMLScriptElement::forAttributeValue):
(WebCore::HTMLScriptElement::eventAttributeValue):
* html/HTMLTableCellElement.cpp:
(WebCore::HTMLTableCellElement::addSubresourceAttributeURLs):
* html/HTMLTableElement.cpp:
(WebCore::HTMLTableElement::rules):
(WebCore::HTMLTableElement::summary):
(WebCore::HTMLTableElement::addSubresourceAttributeURLs):
* html/HTMLTableSectionElement.cpp:
(WebCore::HTMLTableSectionElement::align):
(WebCore::HTMLTableSectionElement::ch):
(WebCore::HTMLTableSectionElement::vAlign):
* html/HTMLTextAreaElement.cpp:
(WebCore::HTMLTextAreaElement::maxLength):
* html/SearchInputType.cpp:
(WebCore::SearchInputType::searchEventsShouldBeDispatched):
* loader/FormSubmission.cpp:
(WebCore::FormSubmission::create):
* page/Frame.cpp:
(WebCore::Frame::matchLabelsAgainstElement):
* page/PageSerializer.cpp:
(WebCore::PageSerializer::serializeFrame):
* platform/win/PasteboardWin.cpp:
(WebCore::Pasteboard::writeImageToDataObject):
* rendering/HitTestResult.cpp:
(WebCore::HitTestResult::altDisplayString):
* rendering/RenderSearchField.cpp:
(WebCore::RenderSearchField::autosaveName):
* rendering/RenderTableCell.cpp:
(WebCore::RenderTableCell::computePreferredLogicalWidths):
* rendering/svg/RenderSVGResourceClipper.cpp:
(WebCore::RenderSVGResourceClipper::drawContentIntoMaskImage):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173724 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoHave LiveNodeLists / HTMLCollections's elementMatches() take a reference
cdumez@apple.com [Thu, 18 Sep 2014 16:32:41 +0000 (16:32 +0000)]
Have LiveNodeLists / HTMLCollections's elementMatches() take a reference
https://bugs.webkit.org/show_bug.cgi?id=136902

Reviewed by Darin Adler.

Have LiveNodeLists / HTMLCollections's elementMatches() take a reference
in argument instead of a pointer as the Element can never be null.

No new tests, no behavior change.

* dom/ClassNodeList.h:
(WebCore::ClassNodeList::elementMatches):
* dom/DocumentOrderedMap.cpp:
(WebCore::keyMatchesWindowNamedItem):
(WebCore::keyMatchesDocumentNamedItem):
* dom/Element.cpp:
(WebCore::Element::updateNameForDocument):
(WebCore::Element::updateIdForDocument):
* dom/LiveNodeList.cpp:
(WebCore::LiveNodeList::namedItem):
* dom/LiveNodeList.h:
(WebCore::CachedLiveNodeList<NodeListType>::collectionBegin):
(WebCore::CachedLiveNodeList<NodeListType>::collectionLast):
(WebCore::CachedLiveNodeList<NodeListType>::collectionTraverseForward):
(WebCore::CachedLiveNodeList<NodeListType>::collectionTraverseBackward):
* dom/NameNodeList.h:
(WebCore::NameNodeList::elementMatches):
* dom/TagNodeList.h:
(WebCore::TagNodeList::elementMatches):
(WebCore::HTMLTagNodeList::elementMatches):
* html/HTMLCollection.cpp:
(WebCore::isMatchingElement):
* html/HTMLNameCollection.cpp:
(WebCore::WindowNameCollection::elementMatchesIfNameAttributeMatch):
(WebCore::WindowNameCollection::elementMatches):
(WebCore::DocumentNameCollection::elementMatchesIfIdAttributeMatch):
(WebCore::DocumentNameCollection::elementMatchesIfNameAttributeMatch):
(WebCore::DocumentNameCollection::elementMatches):
* html/HTMLNameCollection.h:
* html/LabelsNodeList.cpp:
(WebCore::LabelsNodeList::elementMatches):
* html/LabelsNodeList.h:
* html/RadioNodeList.cpp:
(WebCore::RadioNodeList::checkElementMatchesRadioNodeListFilter):
(WebCore::RadioNodeList::elementMatches):
* html/RadioNodeList.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173723 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoOccasional unreproducible crashes in MessageReceiverMap::dispatchMessage
timothy_horton@apple.com [Thu, 18 Sep 2014 16:26:27 +0000 (16:26 +0000)]
Occasional unreproducible crashes in MessageReceiverMap::dispatchMessage
https://bugs.webkit.org/show_bug.cgi?id=136909
<rdar://problem/17758325>

Reviewed by Anders Carlsson.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView dealloc]):
We're seeing some messages dispatched on the WebContext that end up
attempting to call dispatchMessage on a freed MessageReceiver.
The WKRemoteObjectRegistry message receiver is added to the WebContext
message receiver map, but never removed, despite the WebContext easily
outliving the WKWebView that owns the remote object registry.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173722 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoFix "Tools/Scripts/build-webkit --efl --no-inspector" build
evab.u-szeged@partner.samsung.com [Thu, 18 Sep 2014 16:22:52 +0000 (16:22 +0000)]
Fix "Tools/Scripts/build-webkit --efl --no-inspector" build
https://bugs.webkit.org/show_bug.cgi?id=136912

Reviewed by Darin Adler.

* runtime/TypeSet.cpp:
(JSC::TypeSet::leastCommonAncestor):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173721 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoGMainLoopSource is exposed to race conditions
zandobersek@gmail.com [Thu, 18 Sep 2014 13:06:59 +0000 (13:06 +0000)]
GMainLoopSource is exposed to race conditions
https://bugs.webkit.org/show_bug.cgi?id=135800

Reviewed by Carlos Garcia Campos.

Source/WTF:

GMainLoopSource objects can be dispatching tasks on one thread
while having a new task scheduled on a different thread. This
can for instance occur in WebKitVideoSink, where the timeout
callback can be called on main thread while at the same time
it is being rescheduled on a different thread (created through
GStreamer).

The initial solution is to use GMutex to prevent parallel data
access from different threads. In the future I plan to look at
the possibility of creating thread-specific GMainLoopSource
objects that wouldn't require the use of GMutex.

GSource, GCancellable and std::function<> objects are now packed
into an internal Context structure. Using the C++11 move semantics
it's simple to, at the time of dispatch, move the current context
out of the GMainLoopSource object in case the dispatch causes a
rescheduling on that same object.

Also added in the Context struct is a new GCancellable. The pointer
of that object is shared with the GMainLoopSource before the Context
is moved out for the callback dispatch. This makes it safe to cancel
or even delete the GMainLoopSource during the dispatch and prevents
use-after-delete on GMainLoopSource once the dispatch is done in
the GMainLoopSource::*Callback() methods.

All the schedule*() methods and the cancelWithoutLocking() method
callers now lock the GMutex to ensure no one else is accessing the
data at that moment. Similar goes for the dispatch methods, but those
do the dispatch and possible destruction duties with the mutex unlocked.
The dispatch can cause rescheduling on the same GMainLoopSource object,
which must not be done with a locked mutex.

* wtf/gobject/GMainLoopSource.cpp:
(WTF::GMainLoopSource::GMainLoopSource):
(WTF::GMainLoopSource::~GMainLoopSource):
(WTF::GMainLoopSource::cancel):
(WTF::GMainLoopSource::cancelWithoutLocking):
(WTF::GMainLoopSource::scheduleIdleSource):
(WTF::GMainLoopSource::schedule):
(WTF::GMainLoopSource::scheduleTimeoutSource):
(WTF::GMainLoopSource::scheduleAfterDelay):
(WTF::GMainLoopSource::voidCallback):
(WTF::GMainLoopSource::boolCallback):
(WTF::GMainLoopSource::socketCallback):
(WTF::GMainLoopSource::socketSourceCallback):
(WTF::GMainLoopSource::Context::destroySource):
(WTF::GMainLoopSource::reset): Deleted.
(WTF::GMainLoopSource::destroy): Deleted.
* wtf/gobject/GMainLoopSource.h:

Tools:

Add unit tests for GMainLoopSource.

The tests check correct behavior of GMainLoopSource in various conditions --
from the most simple rescheduling to rescheduling during dispatch, cancelling
or destroying the GMainLoopSource during dispatch, proper destroy callback
dispatching etc.

Scheduling both void (one-time) and bool (repeatable) callbacks is tested.
State of the GMainLoopSource object (either ready, sheduled or active) is
thoroughly tested throughout the lifetime of that object.

Still missing are tests for socket callbacks, which are a bit trickier because
they rely on a GSocket object. The delete-on-destroy GMainLoopSource objects
are also not tested thoroughly, simply because it is at the moment impossible
to test that the objects are actually destroyed when the corresponding source
is finally deleted.

* TestWebKitAPI/PlatformGTK.cmake:
* TestWebKitAPI/Tests/WTF/gobject/GMainLoopSource.cpp: Added.
(TestWebKitAPI::GMainLoopSourceTest::GMainLoopSourceTest):
(TestWebKitAPI::GMainLoopSourceTest::~GMainLoopSourceTest):
(TestWebKitAPI::GMainLoopSourceTest::runLoop):
(TestWebKitAPI::GMainLoopSourceTest::delayedFinish):
(TestWebKitAPI::GMainLoopSourceTest::finish):
(TestWebKitAPI::GMainLoopSourceTest::source):
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173720 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemove unused methods from MediaPlayerClient
carlosgc@webkit.org [Thu, 18 Sep 2014 11:16:23 +0000 (11:16 +0000)]
Remove unused methods from MediaPlayerClient
https://bugs.webkit.org/show_bug.cgi?id=136884

Reviewed by Eric Carlson.

Remove unused methods from MediaPlayerClient that introduce
dependencies on WebCore types.

* html/HTMLMediaElement.cpp:
* html/HTMLMediaElement.h:
* platform/graphics/MediaPlayer.h:
(WebCore::MediaPlayerClient::mediaPlayerOwningDocument): Deleted.
(WebCore::MediaPlayerClient::mediaPlayerHostWindow): Deleted.
(WebCore::MediaPlayerClient::mediaPlayerWindowClipRect): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173719 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[EFL][WK2] Use the correct enum for control key in Ewk_Event_Modifiers in ewk_navigat...
commit-queue@webkit.org [Thu, 18 Sep 2014 06:23:03 +0000 (06:23 +0000)]
[EFL][WK2] Use the correct enum for control key in Ewk_Event_Modifiers in ewk_navigation_policy_decision.cpp
https://bugs.webkit.org/show_bug.cgi?id=136740

Patch by Rohit Kumar <kumar.rohit@samsung.com> on 2014-09-17
Reviewed by Gyuyoung Kim.

* UIProcess/API/efl/ewk_navigation_policy_decision.cpp:
(toEwkEventModifiers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173717 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Add Wayland support to GLContextEGL
zandobersek@gmail.com [Thu, 18 Sep 2014 06:21:55 +0000 (06:21 +0000)]
[GTK] Add Wayland support to GLContextEGL
https://bugs.webkit.org/show_bug.cgi?id=136829

Reviewed by Martin Robinson.

* platform/graphics/egl/GLContextEGL.cpp:
(WebCore::sharedEGLDisplay): Special-case the shared EGL display initialization
for the GTK port when running under Wayland. In that case the pointer to the
native wl_display object should be passed to eglGetDisplay(). If not running
under Wayland we should fall back to using the shared X11 display (if support
for that display system is enabled) or the EGL_DEFAULT_DISPLAY.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173716 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Remove unnecessary prefixes in WebInspectorUI CSS
commit-queue@webkit.org [Thu, 18 Sep 2014 02:32:21 +0000 (02:32 +0000)]
Web Inspector: Remove unnecessary prefixes in WebInspectorUI CSS
https://bugs.webkit.org/show_bug.cgi?id=136905

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-09-17
Reviewed by Timothy Hatcher.

Removed vendor prefixes from justify-content and a few lingering
occurances with transition, box-shadow, and display.

* UserInterface/Views/BreakpointActionView.css:
* UserInterface/Views/BreakpointTreeElement.css:
* UserInterface/Views/ButtonNavigationItem.css:
* UserInterface/Views/CSSStyleDetailsSidebarPanel.css:
* UserInterface/Views/ControlToolbarItem.css:
* UserInterface/Views/DebuggerDashboardView.css:
* UserInterface/Views/DefaultDashboardView.css:
* UserInterface/Views/FontResourceContentView.css:
* UserInterface/Views/ImageResourceContentView.css:
* UserInterface/Views/LayerTreeSidebarPanel.css:
* UserInterface/Views/Main.css:
* UserInterface/Views/NavigationBar.css:
* UserInterface/Views/NavigationSidebarPanel.css:
* UserInterface/Views/Popover.css:
* UserInterface/Views/ReplayDashboardView.css:
* UserInterface/Views/Toolbar.css:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173715 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Windows] Unreviewed rebaselining of some failing JS tests.
roger_fong@apple.com [Thu, 18 Sep 2014 01:26:03 +0000 (01:26 +0000)]
[Windows] Unreviewed rebaselining of some failing JS tests.
* platform/win/js/dom/deep-recursion-test-expected.txt: Added.
* platform/win/js/dom/global-constructors-attributes-shared-worker-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173712 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed, Update my email in contributors.json
ryuan.choi@gmail.com [Thu, 18 Sep 2014 01:23:09 +0000 (01:23 +0000)]
Unreviewed, Update my email in contributors.json

* Scripts/webkitpy/common/config/contributors.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173711 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed, rolling out r173695.
dbates@webkit.org [Thu, 18 Sep 2014 01:15:02 +0000 (01:15 +0000)]
Unreviewed, rolling out r173695.

Broke building third-party Legacy WebKit apps; will
investigate offline

Reverted changeset:

"[iOS] Make WebCore build with public iOS SDK"
https://bugs.webkit.org/show_bug.cgi?id=136487
http://trac.webkit.org/changeset/173695

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173710 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed. Use a power-of-2 image size instead so that WebGL tests won’t complain.
roger_fong@apple.com [Thu, 18 Sep 2014 00:59:56 +0000 (00:59 +0000)]
Unreviewed. Use a power-of-2 image size instead so that WebGL tests won’t complain.
* fast/canvas/image-potential-subsample.html:
* fast/canvas/resources/image-6400x6400.jpg: Copied from fast/canvas/resources/image-8000x8000.jpg.
* fast/canvas/resources/image-8000x8000.jpg: Removed.
* fast/canvas/webgl/resources/tex-image-and-sub-image-2d-with-potentially-subsampled-image.js:
(.init):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173709 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed. Windows rebaselining to account for subpixel layout Part 2.
roger_fong@apple.com [Thu, 18 Sep 2014 00:26:42 +0000 (00:26 +0000)]
Unreviewed. Windows rebaselining to account for subpixel layout Part 2.

* platform/win/editing/apply-style-iframe-crash-expected.txt: Added.
* platform/win/editing/selection/move-left-right-expected.txt:
* platform/win/fast/table/dynamic-caption-add-remove-before-child-expected.txt:
* platform/win/tables/mozilla/bugs/bug4527-expected.txt: Added.
* platform/win/tables/mozilla/marvin/body_col-expected.txt:
* platform/win/tables/mozilla/marvin/x_th_valign_baseline-expected.txt:
* platform/win/tables/mozilla/other/body_col-expected.txt:
* platform/win/tables/mozilla_expected_failures/core: Added.
* platform/win/tables/mozilla_expected_failures/core/captions1-expected.txt: Added.
* platform/win/tables/mozilla_expected_failures/core/captions2-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173708 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWebInpector can use unprefixed flexbox
dino@apple.com [Wed, 17 Sep 2014 22:40:21 +0000 (22:40 +0000)]
WebInpector can use unprefixed flexbox
https://bugs.webkit.org/show_bug.cgi?id=136899

Reviewed by Joseph Pecoraro.

Now the inspector UI itself can use a non-prefixed flexbox. This was
a basic search and replace for the following properties:
- align-content
- align-items
- align-self
- flex-basis
- flex-direction
- flex-wrap
- flex-grow
- flex-shrink
- flex
- flex-flow
- justify

* UserInterface/Views/ButtonToolbarItem.css:
* UserInterface/Views/CSSStyleDetailsSidebarPanel.css:
* UserInterface/Views/ContentBrowser.css:
* UserInterface/Views/ControlToolbarItem.css:
* UserInterface/Views/DebuggerDashboardView.css:
* UserInterface/Views/DefaultDashboardView.css:
* UserInterface/Views/FilterBar.css:
* UserInterface/Views/FlexibleSpaceNavigationItem.css:
* UserInterface/Views/FontResourceContentView.css:
* UserInterface/Views/HierarchicalPathComponent.css:
* UserInterface/Views/ImageResourceContentView.css:
* UserInterface/Views/LayerTreeSidebarPanel.css:
* UserInterface/Views/LogContentView.css:
* UserInterface/Views/Main.css:
* UserInterface/Views/NavigationBar.css:
* UserInterface/Views/NavigationSidebarPanel.css:
* UserInterface/Views/Popover.css:
* UserInterface/Views/QuickConsole.css:
* UserInterface/Views/ReplayDashboardView.css:
* UserInterface/Views/ResourceSidebarPanel.css:
* UserInterface/Views/Toolbar.css:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173707 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoChange CallFrame to use Callee instead of JSScope to implement vm()
msaboff@apple.com [Wed, 17 Sep 2014 22:34:14 +0000 (22:34 +0000)]
Change CallFrame to use Callee instead of JSScope to implement vm()
https://bugs.webkit.org/show_bug.cgi?id=136894

Reviewed by Geoffrey Garen.

Added JSCell::vm() method that can be used on any JSObject.  Changed CallFrame::vm() to
use JSCell::vm with the Callee.  Made similar changes in the LLInt.
In support of this, changed JSGlobalObject::init() to take a VM& parameter, as there is
a chicken/egg problem with trying to use the Callee in the global exec before the Callee
has been create.  Besides, the vm is readily available in finishCreation(), the caller of
init().

* llint/LowLevelInterpreter32_64.asm:
* llint/LowLevelInterpreter64.asm:
Changed the calculation of CallFrame::VM to use the Callee instead of JSScope.

* runtime/JSCell.h:
* runtime/JSCellInlines.h:
(JSC::JSCell::vm): New method for getting VM from the pointer.
(JSC::ExecState::vm): Moved this method from JSScope.h to here since this file
contains the implementation of JSCell::vm(), this file is included by all users
of CallFrame::vm, and lastly putting it in CallFrameInlines.h required changing
many other .h files and possible the WebCore generator generate-bindings.pl.

* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
* runtime/JSGlobalObject.h:
(JSC::JSGlobalObject::finishCreation):
Changed init() to take a VM parameter.

* runtime/JSScope.h:
(JSC::ExecState::vm): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173706 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAdd unprefixed flexbox display values and property names/values
dino@apple.com [Wed, 17 Sep 2014 22:06:03 +0000 (22:06 +0000)]
Add unprefixed flexbox display values and property names/values
https://bugs.webkit.org/show_bug.cgi?id=136880
<rdar://problem/18361137>

Reviewed by Benjamin Poulain with some extra comments by Joseph Pecoraro.

* UserInterface/Models/CSSKeywordCompletions.js: Switch to all the unprefixed
flex properties.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173705 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Windows] Unreviewed gardening. Content Security Policy 1.1 (ENABLE_CSP_NEXT) is...
roger_fong@apple.com [Wed, 17 Sep 2014 21:37:30 +0000 (21:37 +0000)]
[Windows] Unreviewed gardening. Content Security Policy 1.1 (ENABLE_CSP_NEXT) is not enabled

* platform/win/TestExpectations: Tests sometimes crash or timeout in addition to just failing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173704 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoSpeculative fix for a fast\dom\html-collections-named-getter failing only in Debug...
roger_fong@apple.com [Wed, 17 Sep 2014 21:24:40 +0000 (21:24 +0000)]
Speculative fix for a fast\dom\html-collections-named-getter failing only in Debug builds.

* dom/Node.cpp:
(WebCore::Document::invalidateNodeListAndCollectionCaches):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173703 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoFix an issue that caused video playback to not always use high resolution streams
simon.fraser@apple.com [Wed, 17 Sep 2014 21:18:04 +0000 (21:18 +0000)]
Fix an issue that caused video playback to not always use high resolution streams
https://bugs.webkit.org/show_bug.cgi?id=136879
rdar://problem/18316542

Reviewed by Eric Carlson.

The underlying video framework expects to be able to use -[CALayer convertRect:toLayer:nil] on
the video layer to obtain information about the display resolution of the video,
including the device scale factor. This doesn't work for layers in the web process,
which are remotely hosted into the UI process.

For now, use a workaround which involves putting a scale transform equivalent to
the device scale factor on the web process layer, and the inverse scale on the
hosting layer in the UI process.

We have to do this in two places, for in-page video, and for fullscreen video.

* Shared/mac/RemoteLayerTreeTransaction.h: Add hostingDeviceScaleFactor to the LayerCreation
properties.
* Shared/mac/RemoteLayerTreeTransaction.mm:
(WebKit::RemoteLayerTreeTransaction::LayerCreationProperties::LayerCreationProperties):
(WebKit::RemoteLayerTreeTransaction::LayerCreationProperties::encode):
(WebKit::RemoteLayerTreeTransaction::LayerCreationProperties::decode):
* UIProcess/ios/RemoteLayerTreeHostIOS.mm:
(-[WKRemoteView initWithFrame:contextID:hostingDeviceScaleFactor:]): Set the inverse scale
transform on the layer.
(WebKit::RemoteLayerTreeHost::createLayer):
(-[WKRemoteView initWithFrame:contextID:]): Deleted.
* UIProcess/ios/WebVideoFullscreenManagerProxy.h:
* UIProcess/ios/WebVideoFullscreenManagerProxy.messages.in:
* UIProcess/ios/WebVideoFullscreenManagerProxy.mm:
(WebKit::WebVideoFullscreenManagerProxy::setupFullscreenWithID):
* WebProcess/WebPage/mac/PlatformCALayerRemoteCustom.mm:
(WebKit::PlatformCALayerRemoteCustom::PlatformCALayerRemoteCustom):
* WebProcess/WebPage/mac/RemoteLayerTreeContext.h:
(WebKit::RemoteLayerTreeContext::deviceScaleFactor):
* WebProcess/WebPage/mac/RemoteLayerTreeContext.mm:
(WebKit::RemoteLayerTreeContext::layerWasCreated):
* WebProcess/ios/WebVideoFullscreenManager.mm:
(WebKit::WebVideoFullscreenManager::enterFullscreenForNode):
(WebKit::WebVideoFullscreenManager::didSetupFullscreen):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173702 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAdd parsing for :nth-child(An+B of selector)
benjamin@webkit.org [Wed, 17 Sep 2014 20:03:14 +0000 (20:03 +0000)]
Add parsing for :nth-child(An+B of selector)
https://bugs.webkit.org/show_bug.cgi?id=136845

Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-09-17
Reviewed by Antti Koivisto.

Source/WebCore:

Add support for parsing :nth-child(of). The implementation of selector matching
will be in a follow up, there are enough edge cases here already.

Spec here: http://dev.w3.org/csswg/selectors4/#the-nth-child-pseudo

Test: fast/css/parsing-css-nth-child-of.html

* css/CSSParser.cpp:
(WebCore::CSSParser::detectFunctionTypeToken):
(WebCore::CSSParser::realLex):
The parser generate two new kinds of tokens:
-NTHCHILDFUNCTION, matching the "nth-child" identifier.
-NTHCHILDSELECTORSEPARATOR, matching the "of" keyword separating An+B from the selector list.

NTHCHILDFUNCTION is used to extend the new syntax only for :nth-child() and not all the "nth" functions
(e.g. :nth-of-type()).

NTHCHILDSELECTORSEPARATOR exists for two reasons:
-We must clear the "parsingMode" before parsing selectorList. If we failed to do that, the complex selectors
 would be parsed in NthChildMode, which has all kind of bad side effects.
-The second reason is differentiacting "of" for all the other identifiers. Arguably, this could have been done
 in the grammar, but it is clearer this way since we already need the branches for the parsingMode.

* css/CSSGrammar.y.in:
Those are pretty much duplicates of the existing FUNCTION rules but supporting the new syntax.

* css/CSSParserValues.cpp:
(WebCore::selectorListMatchesPseudoElement):
(WebCore::CSSParserSelector::matchesPseudoElement):
Matching pseudo elements does not make much sense for those selectorList (e.g. :nth-child(2n of ::before)).
Add helper function to fail parsing of those cases.

* css/CSSParserValues.h:
* css/CSSSelector.cpp:
(WebCore::CSSSelector::selectorText):
* css/CSSSelector.h:
(WebCore::CSSSelector::selectorList):
Updated serialization code for CSS OM.

LayoutTests:

* fast/css/parsing-css-nth-child-of-expected.txt: Added.
* fast/css/parsing-css-nth-child-of.html: Added.
Two new tests to cover the new syntax.

* fast/css/css-selector-text-expected.txt:
* fast/css/css-selector-text.html:
* fast/css/css-set-selector-text-expected.txt:
* fast/css/css-set-selector-text.html:
The usual tests for CSS OM.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173698 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoCSS value in whitespace-separated list attribute selector (~=) mishandles tab/newline...
benjamin@webkit.org [Wed, 17 Sep 2014 20:01:09 +0000 (20:01 +0000)]
CSS value in whitespace-separated list attribute selector (~=) mishandles tab/newline/etc.
https://bugs.webkit.org/show_bug.cgi?id=136807

Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-09-17
Reviewed by Antti Koivisto.

Source/WebCore:

The spec defines selector lists as token separated by white space.

The definition of white space for CSS Selectors is:
"White space in Selectors consists of the code points SPACE (U+0020),
TAB (U+0009), LINE FEED (U+000A), CARRIAGE RETURN (U+000D), and FORM FEED (U+000C)
can occur in whitespace. Other space-like code points, such as EM SPACE (U+2003)
and IDEOGRAPHIC SPACE (U+3000), are never part of white space."

The old code was only matching SPACE (U+0020) which is buggy. This patch
adds support for the other caracters.

Tests: fast/selectors/attr-list-01.html
       fast/selectors/attribute-list-with-whitespace-in-selector.html
       fast/selectors/attribute-list-with-whitespace-in-value.html

* css/SelectorChecker.cpp:
(WebCore::attributeValueMatches):
* cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::constructFragments):
(WebCore::SelectorCompiler::attributeValueSpaceSeparetedListContains):
Add the missing checks.

* html/track/VTTRegion.cpp:
(WebCore::VTTRegion::setRegionSettings):
(WebCore::VTTRegion::parseSettingValue):
* html/track/WebVTTParser.cpp:
(WebCore::WebVTTParser::hasRequiredFileIdentifier):
(WebCore::WebVTTParser::collectTimingsAndSettings):
* html/track/WebVTTParser.h:
While looking for affected code, I discovered WebVTT redefines isHTMLSpace().
I cleaned that code.

Source/WTF:

* wtf/text/AtomicString.h:
(WTF::AtomicString::find):

LayoutTests:

* fast/selectors/attr-list-01-expected.html: Added.
* fast/selectors/attr-list-01.html: Added.
This test is from a similar fix from Opera for the blink fork.
Review URL: https://chromiumcodereview.appspot.com/14980012
No other part of that patch is included.

* fast/selectors/attribute-list-with-whitespace-in-selector-expected.txt: Added.
* fast/selectors/attribute-list-with-whitespace-in-selector.html: Added.
Test whitespace characters in the selector itself.

* fast/selectors/attribute-list-with-whitespace-in-value-expected.txt: Added.
* fast/selectors/attribute-list-with-whitespace-in-value.html: Added.
Test whitespace characters in the attribute value.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173697 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAways have a PageThrottler (sometimes have a UserActivity::Impl)
barraclough@apple.com [Wed, 17 Sep 2014 18:27:42 +0000 (18:27 +0000)]
Aways have a PageThrottler (sometimes have a UserActivity::Impl)
https://bugs.webkit.org/show_bug.cgi?id=136892

Reviewed by Geoffrey Garen.

Some instances of Page have a PageThrottler to track activity on the Page, and some do not.
(Specifically, those created from WebPage in WK2 do, those related to WK1, SVG & inspector do not).

We do this for three reasons:
    * We do not need to take AppNap assertions on WK1.
    * Some Pages do not track their visibility correctly, and would always claim to be visible.
    * Page VisibilityState is not set until after construction, so if we instantiate

Creating the entire PageThrottler lazily has the drawback that we have to check for its existence at
numerous points throughout the code, and we'll miss activity that occurs between the Page being created
& the PageThrottler added to it (really a theoretical problem right now, since they're currently always
created pretty much back to back).

Instead, Page should always have a PageThrottler, & instead make the UserActivity::Impl on the
PageThrottler be added later.

Source/WebCore:

* WebCore.exp.in:
    - createPageThrottler -> enablePageThrottler
* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::parseAttribute):
    - Page::pageThrottler now returns a reference rather than a pointer
* loader/FrameLoader.cpp:
(WebCore::FrameLoader::started):
    - Page::pageThrottler now returns a reference rather than a pointer
* page/Page.cpp:
(WebCore::Page::Page):
    - instantiate PageThrottler in constructor
(WebCore::Page::enablePageThrottler):
    - renamed from createPageThrottler, instead of creating the PageThrottler this method now
      instruct PageThrottler to create a UserActivity::Impl
(WebCore::Page::setViewState):
    - m_pageThrottler is not a pointer
(WebCore::Page::createPageThrottler): Deleted.
    - renamed to enablePageThrottler
* page/Page.h:
(WebCore::Page::pageThrottler):
    - now returns a reference
* page/PageThrottler.cpp:
(WebCore::PageThrottler::PageThrottler):
    - no longer instantiate m_activity
(WebCore::PageThrottler::createUserActivity):
    - lazily instantiate m_activity, call updateUserActivity
(WebCore::PageThrottler::updateUserActivity):
    - m_activity may not yet be instantiated; added early return
* page/PageThrottler.h:
    - m_activity becomes a std::unique_ptr, added createUserActivity

Source/WebKit2:

* WebProcess/Plugins/Netscape/NPRuntimeObjectMap.cpp:
(WebKit::NPRuntimeObjectMap::evaluate):
    - createPageThrottler -> enablePageThrottler
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::WebPage):
(WebKit::WebPage::mouseEvent):
(WebKit::WebPage::wheelEvent):
(WebKit::WebPage::keyEvent):
    - Page::pageThrottler now returns a reference rather than a pointer

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173696 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[iOS] Make WebCore build with public iOS SDK
dbates@webkit.org [Wed, 17 Sep 2014 18:18:02 +0000 (18:18 +0000)]
[iOS] Make WebCore build with public iOS SDK
https://bugs.webkit.org/show_bug.cgi?id=136487

Reviewed by David Kilzer.

Source/WebCore:

Add headers that wrap Mac and iOS SPIs and update WebCore source files to include these
headers instead of the corresponding actual SPI header(s).

* WebCore.vcxproj/WebCore.vcxproj:
* WebCore.vcxproj/WebCore.vcxproj.filters:
* WebCore.vcxproj/WebCoreCG.props:
* WebCore.vcxproj/copyForwardingHeaders.cmd:
* WebCore.xcodeproj/project.pbxproj:
* bindings/objc/DOMInternal.mm:
* bridge/objc/objc_instance.mm:
* platform/audio/ios/MediaSessionManagerIOS.mm:
* platform/cocoa/DisplaySleepDisablerCocoa.cpp:
* platform/cocoa/MemoryPressureHandlerCocoa.mm:
* platform/graphics/SimpleFontData.h:
* platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
* platform/graphics/ca/mac/PlatformCALayerMac.mm:
* platform/graphics/cg/BitmapImageCG.cpp:
* platform/graphics/cg/ColorCG.cpp:
* platform/graphics/cg/GraphicsContextCG.cpp:
* platform/graphics/cg/ImageSourceCG.cpp:
* platform/graphics/cg/PDFDocumentImage.cpp:
* platform/graphics/cocoa/FontPlatformDataCocoa.mm:
* platform/graphics/ios/FontCacheIOS.mm:
* platform/graphics/ios/FontServicesIOS.mm:
* platform/graphics/ios/SimpleFontDataIOS.mm:
* platform/graphics/mac/FontMac.mm:
* platform/graphics/mac/GlyphPageTreeNodeMac.cpp:
* platform/graphics/mac/MediaPlayerProxy.h:
* platform/graphics/mac/SimpleFontDataMac.mm:
* platform/ios/LegacyTileCache.mm:
* platform/ios/LegacyTileGrid.mm:
* platform/ios/LegacyTileGridTile.mm:
* platform/ios/PlatformEventFactoryIOS.mm:
* platform/ios/PlatformScreenIOS.mm:
* platform/ios/WebEvent.mm:
* platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
* platform/ios/wak/WAKAppKitStubs.h:
* platform/ios/wak/WAKView.h:
* platform/ios/wak/WKGraphics.h:
* platform/ios/wak/WKGraphics.mm:
* platform/ios/wak/WebCoreThread.mm:
* platform/network/ios/QuickLook.mm:
* platform/spi/ca/CALayerSPI.h: Added.
* platform/spi/ca/CATiledLayerSPI.h: Added.
* platform/spi/cf/CFLocaleSPI.h: Modified to use EXTERN_C.
* platform/spi/cf/CFNetworkConnectionCacheSPI.h: Ditto. Also changed header visibility from private to project.
* platform/spi/cf/CFURLRequestSPI.h: Ditto. Also modified to unconditionally include header <CoreFoundation/CFBase.h>.
* platform/spi/cg/CGColorTransformSPI.h: Added.
* platform/spi/cg/CGContextSPI.h: Added.
* platform/spi/cg/CGFloatSPI.h: Added.
* platform/spi/cg/CGFontGlyphSupportSPI.h: Added.
* platform/spi/cg/CGFontInfoSPI.h: Added.
* platform/spi/cg/CGFontRenderingSPI.h: Added.
* platform/spi/cg/CGFontUnicodeSupportSPI.h: Added.
* platform/spi/cg/CGImageSPI.h: Added.
* platform/spi/cg/CGSRegionSPI.h: Added.
* platform/spi/cocoa/DispatchSPI.h: Added.
* platform/spi/cocoa/IOPMLibSPI.h: Added.
* platform/spi/cocoa/MachVMSPI.h: Added.
* platform/spi/cocoa/NSFileManagerSPI.h: Added.
* platform/spi/cocoa/NSGeometrySPI.h: Added.
* platform/spi/cocoa/NSPointerFunctionsSPI.h: Added.
* platform/spi/cocoa/dyldSPI.h: Added.
* platform/spi/ios/AVPlayerControllerSPI.h: Added.
* platform/spi/ios/AVPlayerViewControllerSPI.h: Added.
* platform/spi/ios/AVValueTimingSPI.h: Added.
* platform/spi/ios/AVVideoLayerSPI.h: Added.
* platform/spi/ios/CTFontDescriptorSPI.h: Added.
* platform/spi/ios/CTFontSPI.h: Added.
* platform/spi/ios/CUICatalogSPI.h: Added.
* platform/spi/ios/CUIStyleEffectConfigurationSPI.h: Added.
* platform/spi/ios/MPAVRoutingControllerSPI.h: Added.
* platform/spi/ios/MobileGestaltSPI.h: Added.
* platform/spi/ios/QLPreviewConverterSPI.h: Added.
* platform/spi/ios/QuickLookSPI.h: Added.
* rendering/RenderThemeIOS.mm:

Source/WebKit/ios:

Substitute <WebCore/CGFloatSPI.h> for <CoreGraphics/CGFloat.h>.

* WebCoreSupport/WebFixedPositionContent.mm:

Source/WebKit/mac:

Update WebKit1 source files to include the wrapper SPI headers instead of the corresponding
actual SPI header(s).

* DOM/WebDOMOperationsPrivate.h:
* MigrateHeaders.make: Modified to only modify the include path of a "<WebCore/"-prefixed header
file if its filename doesn't end in "SPI.h" as an alternative to having MigrateHeaders.make copy
such files into the WebKit framework as private headers.
* WebCoreSupport/WebFrameLoaderClient.mm:
* WebInspector/WebNodeHighlight.mm:
* WebView/WebIndicateLayer.mm:
* WebView/WebRenderLayer.h:
* WebView/WebView.mm:

Source/WebKit2:

Update WebKit2 source files to include the wrapper SPI headers instead of the corresponding
actual SPI header(s).

* Platform/mac/SharedMemoryMac.cpp:
* Shared/ios/NativeWebTouchEventIOS.mm:
* Shared/mac/RemoteLayerBackingStore.mm:
* Shared/mac/RemoteLayerTreePropertyApplier.mm:
* UIProcess/API/Cocoa/WKWebView.mm:
* UIProcess/ios/WKPDFPageNumberIndicator.mm:
* UIProcess/ios/WKScrollView.mm:
* UIProcess/ios/WebMemoryPressureHandlerIOS.mm:
* UIProcess/ios/forms/WKAirPlayRoutePicker.mm:
* WebProcess/WebPage/WebPage.cpp:

Source/WTF:

Define convenience macro EXTERN_C that can be used to specify the C linkage convention
for a declaration. For example, the declaration "EXTERN_C const int x;" will expand to:

extern const int x;

and

extern "C" const int x;

when used in a C and C++ file, respectively.

* wtf/Compiler.h:

Tools:

Substitute <WebCore/MachVMSPI.h> for <mach/mach_vm.h>.

* DumpRenderTree/mac/CheckedMalloc.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173695 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoDOM timer throttling for hidden plugins
barraclough@apple.com [Wed, 17 Sep 2014 17:00:24 +0000 (17:00 +0000)]
DOM timer throttling for hidden plugins
https://bugs.webkit.org/show_bug.cgi?id=136197

Patch by Gavin Barraclough <barraclough@apple.com> on 2014-09-17
Reviewed by Geoff Garen & Andreas Kling.

For non-visible web pages we slow timers (since they can't be driving animations, etc).
We should do the same for plugins.

Source/WebCore:

There are a few steps to this:
    - JSPluginElementFunctions notifies DOMTimer when script interacts with a plugin.
    - DOMTimerFireState keeps a record of events that occur while a timer is firing.
    - DOMTimer::fired detects timers that interact with invisible/inaudible plugins, and flags itself for throtting.
    - DOMTimer::intervalClampedToMinimum sets timer intervals appropriately.

* bindings/js/JSPluginElementFunctions.cpp:
(WebCore::pluginScriptObject):
    - notify DOMTimer when script interacts with a plugin.
* html/HTMLPlugInElement.cpp:
(WebCore::HTMLPlugInElement::isDetectable):
* html/HTMLPlugInElement.h:
    - added isDetectable, to check for visible / audible plugins.
* page/DOMTimer.cpp:
(WebCore::DOMTimerFireState::DOMTimerFireState):
(WebCore::DOMTimerFireState::~DOMTimerFireState):
    - track current DOMTimerFireState.
(WebCore::DOMTimer::DOMTimer):
    - initialize m_throttleState.
(WebCore::DOMTimer::scriptDidInteractWithPlugin):
    - if interaction with a plugin occurs while a timer is firing, record on the DOMTimerFireState.
(WebCore::DOMTimer::fired):
    - set DOMTimerFireState, and update m_throttleState accordingly.
(WebCore::DOMTimer::updateTimerIntervalIfNecessary):
    - remove redundant check of maxTimerNestingLevel (covered by intervalClampedToMinimum).
(WebCore::DOMTimer::intervalClampedToMinimum):
    - Also take m_throttleState into account when throttling.
* page/DOMTimer.h:
    - added scriptDidInteractWithPlugin, m_throttleState.
* platform/audio/AudioHardwareListener.h:
    - enum -> enum class.
* plugins/PluginViewBase.h:
(WebCore::PluginViewBase::audioHardwareActivity):
    - expose audioHardwareActivity on PluginViewBase (previously available in subclass).

Source/WebKit2:

* PluginProcess/WebProcessConnection.cpp:
* WebProcess/Plugins/Netscape/NPRuntimeObjectMap.cpp:
    - remove now redundant includes.
* WebProcess/Plugins/PluginView.h:
    - audioHardwareActivity is now virtual, override.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173694 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoExclude page visibility from PageThrottler's hysteresis
barraclough@apple.com [Wed, 17 Sep 2014 16:32:10 +0000 (16:32 +0000)]
Exclude page visibility from PageThrottler's hysteresis
https://bugs.webkit.org/show_bug.cgi?id=136866

Reviewed by Geoff Garen.

Including visibility in the hysteresis mechanism has the effect of prolonging the visually
idle timeout, and causing the page hide event to be run at foreground priority. Neither of
these are particularly desirable. Instead separate visibility from the rest of the page
activities we track (and apply hysteresis to), and feed this directly into determination
of the UserActivity state.

* page/PageThrottler.cpp:
(WebCore::PageThrottler::PageThrottler):
    - when the PageThrottler is instantiated call updateUserActivity to set initial UserActivity.
(WebCore::PageThrottler::incrementActivityCount):
    - simplified - when m_activityCount becomes non-zero, start m_hysteresis.
(WebCore::PageThrottler::decrementActivityCount):
    - simplified - when m_activityCount becomes zero, stop m_hysteresis.
(WebCore::PageThrottler::updateUserActivity):
    - end the UserActivity (allow AppNap) if visually idle and no page activity is taking place.
(WebCore::PageThrottler::setViewState):
    - when the visually idle state changed call updateUserActivity to update the UserActivity.
(WebCore::PageThrottler::started):
(WebCore::PageThrottler::stopped):
    - when the hysteresis state changed call updateUserActivity to update the UserActivity.
(WebCore::PageThrottler::updateHysteresis): Deleted.
    - removed: simplified the hysteresis trigger, we now incorporate visually idle state in updateUserActivity.
* page/PageThrottler.h:
    - removed updateHysteresis, added updateUserActivity.
* platform/HysteresisActivity.h:
(WebCore::HysteresisActivity::state):
    - determine the curent state of the HysteresisActivity - started, waiting, or stopped.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173693 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Fix runtime critical warnings when writing to the clipboard after r173687
carlosgc@webkit.org [Wed, 17 Sep 2014 16:12:50 +0000 (16:12 +0000)]
[GTK] Fix runtime critical warnings when writing to the clipboard after r173687
https://bugs.webkit.org/show_bug.cgi?id=136891

Reviewed by Philippe Normand.

* platform/gtk/PasteboardHelper.cpp:
(WebCore::PasteboardHelper::writeClipboardContents): Check passed
callback is not null before trying to ref/unref it since
g_closure_ref/unref are not null safe.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173692 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] WebKitWebView::load-failed-with-tls-errors should receive the failing URI inste...
carlosgc@webkit.org [Wed, 17 Sep 2014 16:11:12 +0000 (16:11 +0000)]
[GTK] WebKitWebView::load-failed-with-tls-errors should receive the failing URI instead of a host
https://bugs.webkit.org/show_bug.cgi?id=136889

Reviewed by Gustavo Noronha Silva.

Source/WebKit2:

We were passing a host for two different reasons, first because
it's more convenient to add an exception with
webkit_web_context_allow_tls_certificate_for_host(), but also
because we were assuming the active URI is the failing URI in case
of failure. This assumption is correct because our current code
does that, but I'm not sure we are doing it on purpose. That
behaviour is not documented anywhere and it's not what WebKit2 does
internaly.

* UIProcess/API/gtk/WebKitWebView.cpp:
(webkit_web_view_class_init):
(webkitWebViewLoadFailedWithTLSErrors):
* UIProcess/API/gtk/WebKitWebView.h:

Tools:

* TestWebKitAPI/Tests/WebKit2Gtk/TestSSL.cpp:
(testLoadFailedWithTLSErrors): Check that LoadFailedWithTLSErrors
event was added to the events vector.
* TestWebKitAPI/gtk/WebKit2Gtk/LoadTrackingTest.cpp:
(loadFailedCallback): Do not assume the web view URI is the
failing URI when the load fails before the committed state.
(loadFailedWithTLSErrorsCallback): Handle the case of load failure
because of TLS errors can call LoadTrackingTest::loadFailedWithTLSErrors.
(LoadTrackingTest::LoadTrackingTest): Connect to WebKitWebView::load-failed-with-tls-errors.
(LoadTrackingTest::loadFailedWithTLSErrors): Add
LoadFailedWithTLSErrors event to the events vector.
* TestWebKitAPI/gtk/WebKit2Gtk/LoadTrackingTest.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173691 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Add a helper function for display system deduction
zandobersek@gmail.com [Wed, 17 Sep 2014 12:35:08 +0000 (12:35 +0000)]
[GTK] Add a helper function for display system deduction
https://bugs.webkit.org/show_bug.cgi?id=136849

Reviewed by Martin Robinson.

getDisplaySystemType() is a helper function that checks the type of the
default GdkDisplay object and initializes a static variable to the
corresponding value. It then returns the value of that variable.

When building with the GTK+2 API (still the case with the GTK+2 plugin
process) we default to X11.

The function is declared and implemented in GtkUtilities files.

* platform/gtk/GtkUtilities.cpp:
(WebCore::widgetIsOnscreenToplevelWindow):
(WebCore::getDisplaySystemType):
* platform/gtk/GtkUtilities.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173690 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[EFL][GTK] Remove WebKit1 related codes
commit-queue@webkit.org [Wed, 17 Sep 2014 09:29:31 +0000 (09:29 +0000)]
[EFL][GTK] Remove WebKit1 related codes
https://bugs.webkit.org/show_bug.cgi?id=136853

Patch by Renato Nagy <rnagy@inf.u-szeged.hu> on 2014-09-17
Reviewed by Csaba Osztrogonác.

Gtk and EFL WebKit1 ports no longer exist. Webkit1 related codes were deleted
from scripts.

.:

* Source/PlatformGTK.cmake:

Tools:

* BuildSlaveSupport/build.webkit.org-config/config.json:
* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(Compile32bitJSC):
(BuildAndTestWebKit2Factory):
(CompileWebKit1Only): Deleted.
(CompileWebKit2Only): Deleted.
(BuildAndTestWebKit1OnlyFactory): Deleted.
(BuildAndTestWebKit2OnlyFactory): Deleted.
* Scripts/build-webkit:
* Scripts/webkitperl/webkitdirs_unittest/checkForArgumentAndRemoveFromArrayRefGettingValue.pl:
* Scripts/webkitpy/common/config/ports.py:
(GtkWK2Port.build_webkit_command):
(EflWK2Port.build_webkit_command):
* Scripts/webkitpy/common/config/ports_unittest.py:
(DeprecatedPortTest.test_gtk_wk2_port):
* Scripts/webkitpy/port/efl.py:
(EflPort.build_webkit_command):
* Scripts/webkitpy/port/gtk.py:
(GtkPort._built_libraries_path):
(GtkPort._search_paths):
(GtkPort.test_expectations_file_position):
(GtkPort.build_webkit_command):
(GtkPort.run_webkit_tests_command):
(GtkPort.warn_if_bug_missing_in_test_expectations): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173689 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agofilterRootById accidentally clears inAdjacentChain flag
utatane.tea@gmail.com [Wed, 17 Sep 2014 08:54:50 +0000 (08:54 +0000)]
filterRootById accidentally clears inAdjacentChain flag
https://bugs.webkit.org/show_bug.cgi?id=136851

Reviewed by Benjamin Poulain.

Source/WebCore:

Test: fast/selectors/filter-root-node-with-selector-contains-adjacents.html

* dom/SelectorQuery.cpp:
(WebCore::filterRootById):

LayoutTests:

When SubSelector comes, `inAdjacentChain` flag is cleared to false.
So for example, `document.querySelector("span#id + ok")` doesn't work correctly.
In this patch, we handles SubSelector relation correctly; don't change `inAdjacentChain` flag
to filter the right search root node.

* fast/selectors/filter-root-node-with-selector-contains-adjacents-expected.txt: Added.
* fast/selectors/filter-root-node-with-selector-contains-adjacents.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173688 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Fix layering violations in PasteboardGtk
carlosgc@webkit.org [Wed, 17 Sep 2014 07:05:45 +0000 (07:05 +0000)]
[GTK] Fix layering violations in PasteboardGtk
https://bugs.webkit.org/show_bug.cgi?id=136802

Reviewed by Darin Adler.

Source/WebCore:

Refactor the Pasteboard code moving the WebCore parts to the Editor.

* PlatformGTK.cmake: Add new file to compilation.
* editing/Editor.cpp:
(WebCore::Editor::performCutOrCopy):
(WebCore::Editor::copyImage):
* editing/Editor.h:
* editing/gtk/EditorGtk.cpp: Added.
(WebCore::createFragmentFromPasteBoardData):
(WebCore::Editor::pasteWithPasteboard):
(WebCore::getImageAndURLForElement):
(WebCore::Editor::writeImageToPasteboard):
(WebCore::Editor::writeSelectionToPasteboard):
* page/DragController.cpp:
(WebCore::DragController::startDrag):
* page/gtk/DragControllerGtk.cpp:
(WebCore::DragController::declareAndWriteDragImage):
* platform/Pasteboard.h:
* platform/gtk/DataObjectGtk.cpp: Remove range member, the caller
should call setText() + setMarkup() intead.
(WebCore::DataObjectGtk::setText):
(WebCore::DataObjectGtk::setMarkup):
(WebCore::DataObjectGtk::clearText):
(WebCore::DataObjectGtk::clearMarkup):
(WebCore::DataObjectGtk::clearAllExceptFilenames):
* platform/gtk/DataObjectGtk.h:
(WebCore::DataObjectGtk::hasText):
(WebCore::DataObjectGtk::hasMarkup):
(WebCore::DataObjectGtk::clearImage):
(WebCore::DataObjectGtk::text):
(WebCore::DataObjectGtk::markup):
(WebCore::DataObjectGtk::setRange): Deleted.
* platform/gtk/GtkDragAndDropHelper.h:
* platform/gtk/PasteboardGtk.cpp:
(WebCore::PasteboardImage::PasteboardImage):
(WebCore::PasteboardImage::~PasteboardImage):
(WebCore::Pasteboard::write):
(WebCore::Pasteboard::writeSelection): Deleted.
(WebCore::getURLForImageElement): Deleted.
(WebCore::Pasteboard::writeImage): Deleted.
(WebCore::Pasteboard::documentFragment): Deleted.
* platform/gtk/PasteboardHelper.cpp:
(WebCore::displayFromFrame): Deleted.
(WebCore::PasteboardHelper::getPrimarySelectionClipboard): Deleted.
* platform/gtk/PasteboardHelper.h:

Source/WebKit2:

* WebProcess/WebCoreSupport/gtk/WebEditorClientGtk.cpp:
(WebKit::WebEditorClient::updateGlobalSelection): Use new API to
update the global selection.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173687 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoDragData should not depend on Clipboard, DocumentFragment, and Document
carlosgc@webkit.org [Wed, 17 Sep 2014 06:27:46 +0000 (06:27 +0000)]
DragData should not depend on Clipboard, DocumentFragment, and Document
https://bugs.webkit.org/show_bug.cgi?id=21358

Reviewed by Darin Adler.

Source/WebCore:

Refactor DragData to not depend on Frame, DocumentFragment,
Document and Range. DragData::asFragment() has been moved to
DragController and implemented in the platform specific file.
DragData methods using Frame has been reworked in Mac to not use
Frame, since all other ports were ignoreing the frame parameter.

* page/DragController.cpp:
(WebCore::DragController::documentFragmentFromDragData): Make it
a static member to be able to use createFragmentFromDragData().
(WebCore::DragController::performDragOperation): Update to
DragData API changes.
(WebCore::DragController::dispatchTextInputEventFor): Ditto.
(WebCore::DragController::concludeEditDrag): Ditto.
(WebCore::documentFragmentFromDragData): Deleted.
* page/DragController.h:
* page/efl/DragControllerEfl.cpp:
(WebCore::DragController::createFragmentFromDragData): Move
DragData::asFragment() implementation here.
* page/gtk/DragControllerGtk.cpp:
(WebCore::DragController::dragOperation):
(WebCore::DragController::createFragmentFromDragData): Ditto.
* page/mac/DragControllerMac.mm:
(WebCore::DragController::dragOperation):
(WebCore::DragController::createFragmentFromDragData): Ditto.
* page/win/DragControllerWin.cpp:
(WebCore::DragController::createFragmentFromDragData): Ditto.
* platform/DragData.h:
* platform/efl/DragDataEfl.cpp:
(WebCore::DragData::asPlainText): Remove Frame parameter.
(WebCore::DragData::containsURL): Ditto.
(WebCore::DragData::asURL): Ditto.
(WebCore::DragData::asFragment): Deleted.
* platform/gtk/DragDataGtk.cpp:
(WebCore::DragData::asPlainText): Remove Frame parameter.
(WebCore::DragData::containsCompatibleContent):
(WebCore::DragData::containsURL): Ditto.
(WebCore::DragData::asURL): Ditto.
(WebCore::DragData::asFragment): Deleted.
* platform/mac/DragDataMac.mm:
(WebCore::DragData::asPlainText): Same implementation as the
Editor.
(WebCore::DragData::containsURL): Remove Frame parameter.
(WebCore::DragData::asURL): Use URLByCanonicalizingURL() instead
of using the Editor client.
(WebCore::DragData::asFragment): Deleted.
* platform/win/DragDataWin.cpp:
(WebCore::DragData::containsURL): Remove Frame parameter.
(WebCore::DragData::asURL): Ditto.
(WebCore::DragData::asPlainText): Ditto.
(WebCore::DragData::containsCompatibleContent):
(WebCore::DragData::asFragment): Deleted.

Source/WebKit2:

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::performDragControllerAction): Update to API
change in DragData::asURL().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173686 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed, disable native inlining because it causes build failures.
fpizlo@apple.com [Wed, 17 Sep 2014 04:32:32 +0000 (04:32 +0000)]
Unreviewed, disable native inlining because it causes build failures.

Source/JavaScriptCore:

* JavaScriptCore.xcodeproj/project.pbxproj:

Source/WTF:

* wtf/Platform.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173685 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRename Node::childNode(index) to traverseToChildAt(index) for clarity
cdumez@apple.com [Wed, 17 Sep 2014 02:00:40 +0000 (02:00 +0000)]
Rename Node::childNode(index) to traverseToChildAt(index) for clarity
https://bugs.webkit.org/show_bug.cgi?id=136825

Reviewed by Benjamin Poulain.

Rename Node::childNode(index) to traverseToChildAt(index) to make it
clearer that the method is actually traversing the children and thus
potentially expensive.

This patch also avoids calling traverseToChildAt() in a couple of
easily avoidable cases.

No new tests, no behavior change.

* WebCore.exp.in:
* WebCore.order:
Remove symbol for ContainerNode::childNode() as it was renamed. It does
not seem we need to expose ContainerNode::traverseToChildAt().

* accessibility/AccessibilityObject.cpp:
(WebCore::AccessibilityObject::stringForVisiblePositionRange):
(WebCore::AccessibilityObject::lengthForVisiblePositionRange):
* accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
(-[WebAccessibilityObjectWrapper arrayOfTextForTextMarkers:attributed:]):
* accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(-[WebAccessibilityObjectWrapper doAXAttributedStringForTextMarkerRange:]):
Mechanical renaming.

* dom/ContainerNode.cpp:
(WebCore::ContainerNode::traverseToChildAt):
(WebCore::ContainerNode::childNode): Deleted.
* dom/ContainerNode.h:
(WebCore::Node::traverseToChildAt):
(WebCore::Node::childNode): Deleted.
* dom/Node.h:
- Rename Node / ContainerNode's childNode(index) to
  traverseToChildAt(index) to make it clear that it is actually
  traversing the children and thus potentially expensive.
- Clean up the implementation of ContainerNode::traverseToChildAt()
  to avoid the use of an extra |i| variable.

* dom/Position.cpp:
(WebCore::Position::computeNodeBeforePosition):
Avoid calling traverseToChildAt(-1). Relying on the unsigned argument
wrapping and the method returning null in this case is a bit obscure
and causes unnecessary traversal of all children.

(WebCore::Position::computeNodeAfterPosition):
(WebCore::Position::previous):
(WebCore::Position::next):
Mechanical renaming and update variable names to stop using
abbreviations as per coding style.

* dom/PositionIterator.h:
(WebCore::PositionIterator::PositionIterator):
* dom/Range.cpp:
(WebCore::Range::insertNode):
(WebCore::Range::checkNodeWOffset):
(WebCore::Range::firstNode):
(WebCore::Range::pastLastNode):
* dom/RangeBoundaryPoint.h:
(WebCore::RangeBoundaryPoint::set):
* editing/BreakBlockquoteCommand.cpp:
(WebCore::BreakBlockquoteCommand::doApply):
* editing/CompositeEditCommand.cpp:
(WebCore::CompositeEditCommand::removeChildrenInRange):
* editing/DeleteSelectionCommand.cpp:
(WebCore::DeleteSelectionCommand::handleGeneralDelete):
* editing/InsertParagraphSeparatorCommand.cpp:
(WebCore::InsertParagraphSeparatorCommand::doApply):
* editing/TextIterator.cpp:
(WebCore::nextInPreOrderCrossingShadowBoundaries):
(WebCore::TextIterator::node):
(WebCore::SimplifiedBackwardsTextIterator::SimplifiedBackwardsTextIterator):
* editing/VisibleUnits.cpp:
(WebCore::nextLinePosition):
* editing/cocoa/HTMLConverter.mm:
(WebCore::editingAttributedStringFromRange):
Mechanical renaming.

* inspector/DOMPatchSupport.cpp:
(WebCore::DOMPatchSupport::innerPatchChildren):
Iterate over children while incrementing the |i| variable to avoid
calling traverseToChildAt(index) repeatedly and thus traversing the
children from the beginning every time.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173684 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Reduce a bit of churn setting initial remote inspection state
commit-queue@webkit.org [Wed, 17 Sep 2014 00:49:20 +0000 (00:49 +0000)]
Web Inspector: Reduce a bit of churn setting initial remote inspection state
https://bugs.webkit.org/show_bug.cgi?id=136875

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-09-16
Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

* API/JSContextRef.cpp:
(JSGlobalContextCreateInGroup):
Set the defaultl remote debuggable state at the API boundary.

* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
Do not set remote debuggable state here. Let clients set it.

Source/WebCore:

* bindings/js/JSDOMGlobalObject.cpp:
(WebCore::JSDOMGlobalObject::finishCreation):
We no longer need to toggle this state.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173683 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoShould have an editing behavior specific for IOS.
enrica@apple.com [Wed, 17 Sep 2014 00:48:01 +0000 (00:48 +0000)]
Should have an editing behavior specific for IOS.
https://bugs.webkit.org/show_bug.cgi?id=136876

Reviewed by Sam Weinig.

This patch introduces a new editing behavior type to be able
to perform editing tasks that are specific to iOS.
All the existing EditingBehavior methods return the same boolean
value for for Mac and iOS. A new one has been introduced to support
a specific behavior in the DeleteSelectionCommand.

* editing/DeleteSelectionCommand.cpp:
(WebCore::DeleteSelectionCommand::doApply):
* editing/EditingBehavior.h:
(WebCore::EditingBehavior::shouldConsiderSelectionAsDirectional):
(WebCore::EditingBehavior::shouldCenterAlignWhenSelectionIsRevealed):
(WebCore::EditingBehavior::shouldToggleStyleBasedOnStartOfSelection):
(WebCore::EditingBehavior::shouldAlwaysGrowSelectionWhenExtendingToBoundary):
(WebCore::EditingBehavior::shouldSelectOnContextualMenuClick):
(WebCore::EditingBehavior::shouldExtendSelectionByWordOrLineAcrossCaret):
(WebCore::EditingBehavior::shouldRebalanceWhiteSpacesInSecureField):
* editing/EditingBehaviorTypes.h:
* editing/FrameSelection.cpp:
(WebCore::FrameSelection::positionForPlatform):
* page/Settings.cpp:
(WebCore::editingBehaviorTypeForPlatform):
* testing/InternalSettings.cpp:
(WebCore::InternalSettings::setEditingBehavior):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173682 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoPromise: Drop Promise.cast
utatane.tea@gmail.com [Wed, 17 Sep 2014 00:32:18 +0000 (00:32 +0000)]
Promise: Drop Promise.cast
https://bugs.webkit.org/show_bug.cgi?id=136222

Reviewed by Sam Weinig.

Promise.cast is dropped and Promise.resolve is replaced with old Promise.cast.

Source/JavaScriptCore:

* runtime/CommonIdentifiers.h:
* runtime/JSPromiseConstructor.cpp:
(JSC::JSPromiseConstructorFuncResolve):
(JSC::JSPromiseConstructorFuncRace):
(JSC::JSPromiseConstructorFuncAll):
(JSC::JSPromiseConstructorFuncCast): Deleted.

LayoutTests:

* js/dom/Promise-static-cast-expected.txt: Removed.
* js/dom/Promise-static-cast.html: Removed.
* js/dom/Promise-static-resolve-expected.txt:
* js/dom/Promise-static-resolve.html:
* js/dom/Promise-types-expected.txt:
* js/dom/Promise-types.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173681 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Add Versioned Inspector protocol JSON files for iOS 8.0
commit-queue@webkit.org [Tue, 16 Sep 2014 23:47:20 +0000 (23:47 +0000)]
Web Inspector: Add Versioned Inspector protocol JSON files for iOS 8.0
https://bugs.webkit.org/show_bug.cgi?id=136873

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-09-16
Reviewed by Timothy Hatcher.

* Scripts/update-LegacyInspectorBackendCommands.rb:
* UserInterface/Protocol/Legacy/8.0/InspectorJSBackendCommands.js: Added.
* UserInterface/Protocol/Legacy/8.0/InspectorWebBackendCommands.js: Added.
* Versions/InspectorJS-iOS-8.0.json: Added.
* Versions/InspectorWeb-iOS-8.0.json: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173680 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed, rolling out r173670.
enrica@apple.com [Tue, 16 Sep 2014 23:32:56 +0000 (23:32 +0000)]
Unreviewed, rolling out r173670.
https://bugs.webkit.org/show_bug.cgi?id=136871

Creates layering violation (Requested by enrica on #webkit).

Reverted changeset:

"Move HTMLConverter from editing/cocoa to platform/cocoa."
https://bugs.webkit.org/show_bug.cgi?id=136474
http://trac.webkit.org/changeset/173670

Patch by Commit Queue <commit-queue@webkit.org> on 2014-09-16

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173679 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Mac] MediaPlayerPrivateAVFoundationObjC::paintWithVideoOutput doesn't work with...
bfulgham@apple.com [Tue, 16 Sep 2014 23:27:22 +0000 (23:27 +0000)]
[Mac] MediaPlayerPrivateAVFoundationObjC::paintWithVideoOutput doesn't work with rotated movies
https://bugs.webkit.org/show_bug.cgi?id=136872

Patch by Eric Carlson <eric.carlson@apple.com> on 2014-09-16
Reviewed by Tim Horton.

* platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::paintWithImageGenerator): Add logging.
(WebCore::MediaPlayerPrivateAVFoundationObjC::paintWithVideoOutput): Apply the video track
    preferred transform.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173678 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agobmalloc: moved line caches from the deallocator to the allocator
ggaren@apple.com [Tue, 16 Sep 2014 22:37:15 +0000 (22:37 +0000)]
bmalloc: moved line caches from the deallocator to the allocator
https://bugs.webkit.org/show_bug.cgi?id=136868

Reviewed by Gavin Barraclough.

I did this mostly as a simplification, to make it easier to change the
allocation strategy.

No throughput change on MallocBench. Saves about 50kB.

Since the deallocator needs to lock the heap when freeing lines anyway,
there isn't much benefit to giving the deallocator a local cache of
deallocated lines.

We still give the allocator a local cache of lines because that does
reduce the frequency at which it needs to lock the heap in order to
acquire more lines.

* bmalloc/Allocator.cpp:
(bmalloc::Allocator::scavenge):
(bmalloc::Allocator::allocateSmallLine):
(bmalloc::Allocator::allocateMediumLine):
(bmalloc::Allocator::allocateMedium):
(bmalloc::Allocator::allocateSlowCase):
* bmalloc/Allocator.h:
* bmalloc/Deallocator.cpp:
(bmalloc::Deallocator::Deallocator):
(bmalloc::Deallocator::scavenge):
(bmalloc::Deallocator::processObjectLog):
(bmalloc::Deallocator::deallocateSmallLine): Deleted.
(bmalloc::Deallocator::allocateSmallLine): Deleted.
(bmalloc::Deallocator::deallocateMediumLine): Deleted.
(bmalloc::Deallocator::allocateMediumLine): Deleted.
* bmalloc/Deallocator.h:

* bmalloc/Sizes.h:
* bmalloc/VMAllocate.h: Took the opportunity to make the line cache size
exactly one page in size. That's about what we were shooting for anyway,
and it may make it easier to switch to per-page allocation in future.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173675 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoMake spelling tests more reliable by using a word that is misspelled everywhere
mmaxfield@apple.com [Tue, 16 Sep 2014 22:36:41 +0000 (22:36 +0000)]
Make spelling tests more reliable by using a word that is misspelled everywhere
https://bugs.webkit.org/show_bug.cgi?id=136865

Reviewed by Simon Fraser.

Some platforms recognize "wellcome" as spelled correctly.

* editing/spelling/context-menu-suggestions.html:
* editing/spelling/spelling-with-punctuation-selection-expected.txt:
* editing/spelling/spelling-with-punctuation-selection.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173674 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[iOS] Add iOS SDK name and version suffix to WebKitSystemInterface debug/release...
dbates@webkit.org [Tue, 16 Sep 2014 22:29:44 +0000 (22:29 +0000)]
[iOS] Add iOS SDK name and version suffix to WebKitSystemInterface debug/release product name
https://bugs.webkit.org/show_bug.cgi?id=136862

Reviewed by David Kilzer.

Source/WebKit/mac:

* Configurations/DebugRelease.xcconfig:

Source/WebKit2:

* Configurations/DebugRelease.xcconfig:

Tools:

* DumpRenderTree/mac/Configurations/DebugRelease.xcconfig:
* WebKitTestRunner/Configurations/DebugRelease.xcconfig:
* WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj: Remove hardcoded iOS 8.0
IPHONEOS_DEPLOYMENT_TARGET to ensure we link against the latest version of WebKitSystemInterface.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173673 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoLocal OSR availability calculation should be reusable
fpizlo@apple.com [Tue, 16 Sep 2014 22:18:23 +0000 (22:18 +0000)]
Local OSR availability calculation should be reusable
https://bugs.webkit.org/show_bug.cgi?id=136860

Reviewed by Oliver Hunt.

Previously, the FTL lowering repeated some of the logic of the OSR availability analysis
phase. Humorously, it actually did this logic a bit differently; for example the phase
would claim that a SetLocal makes both the flush and the node available while the FTL
only claimed that the flush was available. This different was benign, but still: yuck!

Also, previously if you wanted to use availability information then you'd have to repeat
some of the logic that both the phase itself and the FTL lowering already had.
Presumably, you could get epic style points for finding other benign ways in which to
make your copy of the logic different from the other two!

This reduces the amount of style points one could conceivably get in the future when
hacking JSC, by creating a single reusable thingy for computing local OSR availability.

* dfg/DFGOSRAvailabilityAnalysisPhase.cpp:
(JSC::DFG::OSRAvailabilityAnalysisPhase::run):
(JSC::DFG::LocalOSRAvailabilityCalculator::LocalOSRAvailabilityCalculator):
(JSC::DFG::LocalOSRAvailabilityCalculator::~LocalOSRAvailabilityCalculator):
(JSC::DFG::LocalOSRAvailabilityCalculator::beginBlock):
(JSC::DFG::LocalOSRAvailabilityCalculator::executeNode):
* dfg/DFGOSRAvailabilityAnalysisPhase.h:
* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::LowerDFGToLLVM):
(JSC::FTL::LowerDFGToLLVM::compileBlock):
(JSC::FTL::LowerDFGToLLVM::compileNode):
(JSC::FTL::LowerDFGToLLVM::compileSetLocal):
(JSC::FTL::LowerDFGToLLVM::compileInvalidationPoint):
(JSC::FTL::LowerDFGToLLVM::appendOSRExit):
(JSC::FTL::LowerDFGToLLVM::buildExitArguments):
(JSC::FTL::LowerDFGToLLVM::availability):
(JSC::FTL::LowerDFGToLLVM::compileMovHint): Deleted.
(JSC::FTL::LowerDFGToLLVM::compileZombieHint): Deleted.
(JSC::FTL::LowerDFGToLLVM::initializeOSRExitStateForBlock): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173672 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoMac queue slaves should reboot every so often
rniwa@webkit.org [Tue, 16 Sep 2014 22:14:42 +0000 (22:14 +0000)]
Mac queue slaves should reboot every so often
https://bugs.webkit.org/show_bug.cgi?id=136791

Reviewed by Alexey Proskuryakov.

Reboot Mac queue slaves every night between 1 A.M. and 6 A.M.

* EWSTools/start-queue-mac.sh:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173671 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoMove HTMLConverter from editing/cocoa to platform/cocoa.
enrica@apple.com [Tue, 16 Sep 2014 21:00:53 +0000 (21:00 +0000)]
Move HTMLConverter from editing/cocoa to platform/cocoa.
https://bugs.webkit.org/show_bug.cgi?id=136474

Reviewed by Benjamin Poulain.

This is a platform specific class and it belongs to the platform folder.

* WebCore.xcodeproj/project.pbxproj:
* editing/cocoa: Removed.
* editing/cocoa/HTMLConverter.h: Removed.
* editing/cocoa/HTMLConverter.mm: Removed.
* platform/cocoa/HTMLConverter.h: Copied from editing/cocoa/HTMLConverter.h.
* platform/cocoa/HTMLConverter.mm: Copied from editing/cocoa/HTMLConverter.mm.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173670 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemove PLATFORM(IOS) from WebCore/editing (Part 3).
enrica@apple.com [Tue, 16 Sep 2014 20:38:46 +0000 (20:38 +0000)]
Remove PLATFORM(IOS) from WebCore/editing (Part 3).
https://bugs.webkit.org/show_bug.cgi?id=136474

Reviewed by Benjamin Poulain.

This is the updated version of the patch that was landed in r173340.
This patch removes the use of PLATFORM(IOS) from TextAffinity.h
and removes the assumption that EAffinity values match NSSelectionAffinity
values. It also removes the includes in TextAffinity.h, creating the need to
include explicitly the files when necessary.

Source/WebCore:

* editing/TextAffinity.h:
(kit): Deleted.
(core): Deleted.
* editing/cocoa/HTMLConverter.mm:
* page/mac/WebCoreFrameView.h:

Source/WebKit/mac:

* WebCoreSupport/WebEditorClient.h:
(kit):
(core):

Source/WebKit2:

* WebProcess/InjectedBundle/API/mac/WKDOMInternals.mm:
* WebProcess/WebPage/WKAccessibilityWebPageObjectIOS.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173669 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agooverflow:scroll should not leave space for a scroll corner with overlay scrollbars
bdakin@apple.com [Tue, 16 Sep 2014 20:32:29 +0000 (20:32 +0000)]
overflow:scroll should not leave space for a scroll corner with overlay scrollbars
https://bugs.webkit.org/show_bug.cgi?id=136861

Reviewed by Sam Weinig.

overflow:scroll should behave like overflow:auto when the scrollbar will render as
an overlay scrollbar.

Re-name hasAutoVerticalScrollbar()/Horizontal to
hasVerticalScrollbarWithAutoBehavior()/Horizontal, and return true for
overflow:scroll scrollbars that will render as overlay scrollbars.
* rendering/RenderBox.cpp:
(WebCore::RenderBox::hasVerticalScrollbarWithAutoBehavior):
(WebCore::RenderBox::hasHorizontalScrollbarWithAutoBehavior):
* rendering/RenderBox.h:
(WebCore::RenderBox::scrollsOverflowX):
(WebCore::RenderBox::scrollsOverflowY):
(WebCore::RenderBox::hasAutoVerticalScrollbar): Deleted.
(WebCore::RenderBox::hasAutoHorizontalScrollbar): Deleted.

Re-name overflowRequiresScrollbar() to styleRequiresScrollbar() and also re-name
overflowDefinesAutomaticScrollbar() to styleDefinesAutomaticScrollbar(), and make
these functions take into account the fact that overflow:scroll should act like
overflow:auto when the scrollbar will render as an overlay scrollbar.
* rendering/RenderLayer.cpp:
(WebCore::styleRequiresScrollbar):
(WebCore::styleDefinesAutomaticScrollbar):
(WebCore::RenderLayer::updateScrollbarsAfterLayout):
(WebCore::RenderLayer::calculateClipRects):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173668 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoWeb Inspector: Fix ESLint no-extra-bind warnings
commit-queue@webkit.org [Tue, 16 Sep 2014 19:54:12 +0000 (19:54 +0000)]
Web Inspector: Fix ESLint no-extra-bind warnings
https://bugs.webkit.org/show_bug.cgi?id=136659

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-09-16
Reviewed by Timothy Hatcher.

* .eslintrc:
* UserInterface/Controllers/LayerTreeManager.js:
(WebInspector.LayerTreeManager.prototype.layersForNode):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173667 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Curl] Sometimes incomplete or empty content can be loaded from cache.
commit-queue@webkit.org [Tue, 16 Sep 2014 19:22:39 +0000 (19:22 +0000)]
[Curl] Sometimes incomplete or empty content can be loaded from cache.
https://bugs.webkit.org/show_bug.cgi?id=136855

Patch by peavo@outlook.com <peavo@outlook.com> on 2014-09-16
Reviewed by Alex Christensen.

Sometimes, when two requests with the same url are started at the same time,
there is a possibility of loading incomplete or empty content from the cache.
This happens because the method CurlCacheEntry::isLoading() is returning the wrong status
in the time period between the headers are received, and the content data is received.
This can be fixed by using a flag for the load status, instead of checking whether
the content file is open.

* platform/network/curl/CurlCacheEntry.cpp:
(WebCore::CurlCacheEntry::CurlCacheEntry): Initialize loading flag.
(WebCore::CurlCacheEntry::isLoading): Return loading flag.
(WebCore::CurlCacheEntry::didFail): Call new method to set loading flag.
(WebCore::CurlCacheEntry::didFinishLoading): Ditto.
(WebCore::CurlCacheEntry::setIsLoading): Added new method to set loading flag.
* platform/network/curl/CurlCacheEntry.h: Added loading flag and new method to set it.
* platform/network/curl/CurlCacheManager.cpp:
(WebCore::CurlCacheManager::didReceiveResponse): Call new method to set loading flag.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173666 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRename Node::nodeIndex() to computeNodeIndex() for clarity
cdumez@apple.com [Tue, 16 Sep 2014 18:28:57 +0000 (18:28 +0000)]
Rename Node::nodeIndex() to computeNodeIndex() for clarity
https://bugs.webkit.org/show_bug.cgi?id=136826

Reviewed by Benjamin Poulain.

Rename Node::nodeIndex() to computeNodeIndex() to make it clear that
index is being computed and that calling this method is potentially
expensive.

No new tests, no behavior change.

* WebCore.exp.in:
* WebCore.order:
Rename exported symbol for Node::nodeIndex() as well.

* dom/Document.cpp:
(WebCore::Document::caretRangeFromPoint):
* dom/Node.cpp:
(WebCore::Node::computeNodeIndex):
(WebCore::Node::nodeIndex): Deleted.
* dom/Node.h:
* dom/NodeWithIndex.h:
(WebCore::NodeWithIndex::index):
Clean up implementation to reduce the scope of the Node variable and
to follow coding style (variable naming, spacing, star placement).

* dom/Position.cpp:
(WebCore::Position::computeOffsetInContainerNode):
(WebCore::Position::previous):
(WebCore::Position::next):
(WebCore::Position::upstream):
(WebCore::Position::downstream):
* dom/Position.h:
(WebCore::positionInParentBeforeNode):
(WebCore::positionInParentAfterNode):
* dom/Range.cpp:
(WebCore::Range::compareNode):
(WebCore::Range::intersectsNode):
(WebCore::Range::processContents):
(WebCore::Range::insertNode):
(WebCore::Range::setStartAfter):
(WebCore::Range::setEndBefore):
(WebCore::Range::setEndAfter):
(WebCore::Range::setStartBefore):
* dom/RangeBoundaryPoint.h:
(WebCore::RangeBoundaryPoint::ensureOffsetIsValid):
* editing/ApplyStyleCommand.cpp:
(WebCore::ApplyStyleCommand::mergeStartWithPreviousIfIdentical):
(WebCore::ApplyStyleCommand::mergeEndWithNextIfIdentical):
* editing/DeleteSelectionCommand.cpp:
(WebCore::DeleteSelectionCommand::handleGeneralDelete):
* editing/Editor.cpp:
(WebCore::Editor::avoidIntersectionWithDeleteButtonController):
* editing/FrameSelection.cpp:
(WebCore::FrameSelection::selectFrameElementInParentIfFullySelected):
* editing/ReplaceSelectionCommand.cpp:
(WebCore::ReplaceSelectionCommand::doApply):
* editing/TextIterator.cpp:
(WebCore::TextIterator::range):
(WebCore::SimplifiedBackwardsTextIterator::handleReplacedElement):
(WebCore::SimplifiedBackwardsTextIterator::handleNonTextNode):
* editing/htmlediting.cpp:
(WebCore::updatePositionForNodeRemoval):
* html/HTMLTextFormControlElement.cpp:
(WebCore::setContainerAndOffsetForRange):
* page/DOMSelection.cpp:
(WebCore::DOMSelection::getRangeAt):
(WebCore::DOMSelection::containsNode):
(WebCore::DOMSelection::shadowAdjustedOffset):
* rendering/RenderBlock.cpp:
(WebCore::positionForPointRespectingEditingBoundaries):
* rendering/RenderTreeAsText.cpp:
(WebCore::nodePosition):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173665 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoVideos with controls enabled never receive 'dragstart' events.
jer.noble@apple.com [Tue, 16 Sep 2014 18:09:21 +0000 (18:09 +0000)]
Videos with controls enabled never receive 'dragstart' events.
https://bugs.webkit.org/show_bug.cgi?id=136837

Reviewed by Eric Carlson.

Fixes a crash regression introduced by r173631. Node::contains() has an interesting property of returning false
if `this` is NULL. Rather than depending on this behavior, explicitly NULL-check state.source and bail early.

* page/DragController.cpp:
(WebCore::DragController::startDrag):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173664 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[Win] Unreviewed. Skip failing WebGL conformance test.
roger_fong@apple.com [Tue, 16 Sep 2014 18:08:13 +0000 (18:08 +0000)]
[Win] Unreviewed. Skip failing WebGL conformance test.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173663 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoStandardized the return type of a delegate method.
mitz@apple.com [Tue, 16 Sep 2014 18:00:22 +0000 (18:00 +0000)]
Standardized the return type of a delegate method.

Rubber-stamped by Anders Carlsson.

* UIProcess/API/Cocoa/WKUIDelegatePrivate.h: Changed a return type from bool to BOOL.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173662 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[as] Updated Assamese translations of WebKitGtk+
gns@gnome.org [Tue, 16 Sep 2014 17:27:02 +0000 (17:27 +0000)]
[as] Updated Assamese translations of WebKitGtk+
https://bugs.webkit.org/show_bug.cgi?id=136480

Patch by Nilamdyuti Goswami <ngoswami@redhat.com> on 2014-09-16
Reviewed by Gustavo Noronha.

* as.po: updated.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173661 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoURTBF after r173659.
ossy@webkit.org [Tue, 16 Sep 2014 16:18:59 +0000 (16:18 +0000)]
URTBF after r173659.

After r173659 convertToIntLength() is used only inside ENABLE(DASHBOARD_SUPPORT), so it
became unused function and caused build failure if ENABLE(DASHBOARD_SUPPORT) is false.

Source/WebCore/css/StyleResolver.cpp:1471:15: error: 'WebCore::Length WebCore::convertToIntLength(const WebCore::CSSPrimitiveValue*, const WebCore::CSSToLengthConversionData&)'
defined but not used [-Werror=unused-function] cc1plus: all warnings being treated as errors

* css/StyleResolver.cpp: Missing ENABLE(DASHBOARD_SUPPORT) guard added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173660 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoSource/WebCore:
darin@apple.com [Tue, 16 Sep 2014 15:58:12 +0000 (15:58 +0000)]
Source/WebCore:
Scroll snap properties don't handle inherit and initial propertly.
https://bugs.webkit.org/show_bug.cgi?id=136643

Reviewed by Beth Dakin.

Restructured scroll snap code both to handle inherit and initial, and to reduce the
total code size a bit, fixing a few minor bugs along the way. Updated the regression
tests to cover the new parts.

* css/CSSComputedStyleDeclaration.cpp:
(WebCore::addValuesForNamedGridLinesAtIndex): Use a modern for loop.
(WebCore::scrollSnapDestination): Changed to take a LengthSize. Also streamlined.
(WebCore::scrollSnapPoints): Changed to to take a ScrollSnapPoints. Also streamlined and
added code to handle the "uses elements" case; serialize that as an identifier, not the
string "elements", fixing a minor bug.
(WebCore::scrollSnapCoordinates): Changed to take a Vector<LengthSize>. Also streamlined
and added code to handle the empty vector case; serialize that as an identifier, not the
string "none", fixing a minor bug.
(WebCore::ComputedStyleExtractor::getFontSizeCSSValuePreferringKeyword): Use nullptr.
(WebCore::counterToCSSValue): Use nullptr.
(WebCore::ComputedStyleExtractor::styledNode): Use nullptr.
(WebCore::ComputedStyleExtractor::propertyValue): Use nullptr. Moved CSS_SCROLL_SNAP
to a more appropriate section of the switch statement. Made it just call through to
the functions above in a more straightforward way.
(WebCore::CSSComputedStyleDeclaration::getPropertyValue): Reversed logic to use early
return and to call emptyString.
(WebCore::CSSComputedStyleDeclaration::item): Use emptyString.
(WebCore::ComputedStyleExtractor::getCSSPropertyValuesForSidesShorthand): Use nullptr.
(WebCore::CSSComputedStyleDeclaration::parentRule): Use nullptr.
(WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue): Use nullptr.
(WebCore::CSSComputedStyleDeclaration::getPropertyPriority): Use emptyString.
(WebCore::CSSComputedStyleDeclaration::getPropertyShorthand): Use emptyString.

* css/DeprecatedStyleBuilder.cpp:
(WebCore::ApplyPropertyMarqueeIncrement::applyValue): Use the switch statement for the
invalid value as well as for the valid values, rather than a separate if statement.
Call convertToLength directly rather than through a StyleResolver static member function.

* css/PropertySetCSSStyleDeclaration.cpp:
(WebCore::PropertySetCSSStyleDeclaration::getPropertyValue): Fixed a formatting mistake.

* css/StyleResolver.cpp:
(WebCore::convertToIntLength): Made this function local to this file instead of a
static member function.
(WebCore::convertToFloatLength): Ditto.
(WebCore::StyleResolver::parseSnapCoordinate): Added. Used to keep snap point parsing
code small.
(WebCore::StyleResolver::parseSnapCoordinatePair): Ditto.
(WebCore::StyleResolver::parseSnapPoints): Added. Parses an entire snap points property.
(WebCore::StyleResolver::applyProperty): Removed various bits of redundant code that
were not using the primitiveValue value already in a local variable. Added the
HANDLE_INHERIT_AND_INITIAL to all the scroll snap properties and updated since the
RenderStyle structure now more closely matches the CSS properties.

* css/StyleResolver.h: Updated for above changes, and removed a redundant "private:".

* page/scrolling/AxisScrollSnapOffsets.cpp:
(WebCore::appendChildSnapOffsets): Use auto& instead of SnapCoordinate since that type
is now gone. Updated since coordinates are now LengthSize.
(WebCore::updateFromStyle): Changed to take a reference instead of a pointer and
also to fetch the data from RenderStyle in a new way that more closely matches the
CSS properties.
(WebCore::updateSnapOffsetsForScrollableArea): Ditto. Also moved some memory allocation
inside if statements so we don't wastefully allocate and destroy.

* rendering/style/RenderStyle.cpp: Added an include of StyleScrollSnapPoints.h now
that the header doesn't include them. Moved the conditional includes out of the main
include paragraph (even though I don't think conditional includes are usually a good
idea).
(WebCore::RenderStyle::initialScrollSnapPointsX): Added.
(WebCore::RenderStyle::initialScrollSnapPointsY): Added.
(WebCore::RenderStyle::initialScrollSnapDestination): Added.
(WebCore::RenderStyle::initialScrollSnapCoordinates): Added.
(WebCore::RenderStyle::scrollSnapPointsX): Added.
(WebCore::RenderStyle::scrollSnapPointsY): Added.
(WebCore::RenderStyle::scrollSnapDestination): Added.
(WebCore::RenderStyle::scrollSnapCoordinates): Added.
(WebCore::RenderStyle::setScrollSnapPointsX): Added.
(WebCore::RenderStyle::setScrollSnapPointsY): Added.
(WebCore::RenderStyle::setScrollSnapDestination): Added.
(WebCore::RenderStyle::setScrollSnapCoordinates): Added.

* rendering/style/RenderStyle.h: Replaced the include of StyleScrollSnapPoints.h with
a forward declaration of ScrollSnapPoints. Removed getters and setters for the old
style properties and replaced them with new ones that match the CSS properties more
closely. Tweaked formatting a bit. Added initial value functions for the the scroll snap
properties.

* rendering/style/StyleRareNonInheritedData.cpp: Added include of StyleScrollSnapPoints.h
now that RenderStyle.h doesn't include it.
(WebCore::StyleRareNonInheritedData::StyleRareNonInheritedData): Updated since
m_scrollSnapType is now unsigned instead of ScrollSnapType.

* rendering/style/StyleRareNonInheritedData.h: Fixed type of m_scrollSnapType to be
unsigned instead of ScrollSnapType. This should have been causing problems on Windows,
but I think we got lucky and had no detectable problems.

* rendering/style/StyleScrollSnapPoints.cpp:
(WebCore::ScrollSnapPoints::ScrollSnapPoints): Added.
(WebCore::operator==): Added overloads for both ScrollSnapPoints and also for
StyleScrollSnapPoints; using a free function for both since both have public data.
(WebCore::defaultScrollSnapDestination): Added, replaces a static member function.
(WebCore::StyleScrollSnapPoints::StyleScrollSnapPoints): Updated both the default
constructor and the copy constructor for the new design.

* rendering/style/StyleScrollSnapPoints.h: Added a ScrollSnapPoints struct for the
type of the snap-points-x and snap-points-y CSS properties. Changed the data members
of StyleScrollSnapPoints to match the various CSS properties rather than breaking
out all the data into separate data members.
(WebCore::StyleScrollSnapPoints::defaultRepeatOffset): Deleted. Replaced by the
default constructor for ScrollSnapPoints.
(WebCore::StyleScrollSnapPoints::defaultDestinationOffset): Deleted. Replaced by
a function inside the .cpp file.

LayoutTests:
Scroll snap properties do not handle inherit and initial properly.
https://bugs.webkit.org/show_bug.cgi?id=136643

Reviewed by Beth Dakin.

* css3/scroll-snap/scroll-snap-property-computed-style-expected.txt: Added
expected passing results from new tests.
* css3/scroll-snap/scroll-snap-property-computed-style.js: Added new test
cases that cover inherit and initial. Before, these would crash or fail.

* css3/scroll-snap/scroll-snap-property-parsing-expected.txt: Added expected
passing results from new tests, and also updated to expect pass rather than
fail for the tests that had incorrect expected results.

* css3/scroll-snap/scroll-snap-property-parsing.js: Added new test cases that
cover inherit and initial. Also updated expected results for certain pixel
repeat cases that were incorrect.

* platform/mac/TestExpectations: Fixed an unrelated syntax error in this file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173659 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Typo in webkit_security_manager_register_uri_scheme_as_empty_document document...
commit-queue@webkit.org [Tue, 16 Sep 2014 12:36:41 +0000 (12:36 +0000)]
[GTK] Typo in  webkit_security_manager_register_uri_scheme_as_empty_document documentation
https://bugs.webkit.org/show_bug.cgi?id=136852

Patch by Tomas Popela <tpopela@redhat.com> on 2014-09-16
Reviewed by Carlos Garcia Campos.

* UIProcess/API/gtk/WebKitSecurityManager.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173658 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUnreviewed gardening, skip one more tests on Windows.
ossy@webkit.org [Tue, 16 Sep 2014 09:29:45 +0000 (09:29 +0000)]
Unreviewed gardening, skip one more tests on Windows.

* Scripts/run-javascriptcore-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173657 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoJSC test gardening
ossy@webkit.org [Tue, 16 Sep 2014 08:58:06 +0000 (08:58 +0000)]
JSC test gardening
https://bugs.webkit.org/show_bug.cgi?id=136823

Reviewed by Geoffrey Garen.

Source/JavaScriptCore:

* tests/mozilla/mozilla-tests.yaml: Unskip passing tests.

Tools:

* Scripts/run-javascriptcore-tests: Skip tests on only the affected platforms.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173656 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoFix FindICU.cmake
commit-queue@webkit.org [Tue, 16 Sep 2014 08:57:12 +0000 (08:57 +0000)]
Fix FindICU.cmake
https://bugs.webkit.org/show_bug.cgi?id=136820

Patch by Tibor Meszaros <tmeszaros.u-szeged@partner.samsung.com> on 2014-09-16
Reviewed by Csaba Osztrogonác.

* Source/cmake/FindICU.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173655 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GStreamer] http/tests/media/video-auth.html is failing
commit-queue@webkit.org [Tue, 16 Sep 2014 08:46:10 +0000 (08:46 +0000)]
[GStreamer] http/tests/media/video-auth.html is failing
https://bugs.webkit.org/show_bug.cgi?id=126619

Patch by Youenn Fablet <youenn.fablet@crf.canon.fr> on 2014-09-16
Reviewed by Philippe Normand.

Source/WebCore:

Test: http/tests/media/video-auth-cors.html

Use stored credentials when retrieving videos following image loading behavior
(i.e. stored credentials are always used except when in cors anonymous mode),
except that requesting credentials from user remains disabled for cross-origin requests.

Patch is covered by http/tests/media/video-auth.html.

* platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:
(CachedResourceStreamingClient::CachedResourceStreamingClient): Enable using stored credentials except when using cors anonymous mode.

LayoutTests:

* platform/efl/TestExpectations: Removed failure expectation for http/tests/media/video-auth.html.
* platform/efl/http/tests/media/video-auth-expected.txt: Added.
* platform/gtk/TestExpectations: Removed failure expectation for http/tests/media/video-auth.html.
* platform/gtk/http/tests/media/video-auth-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173654 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRemove shouldAbortEarly cruft from master.cfg
evab.u-szeged@partner.samsung.com [Tue, 16 Sep 2014 08:16:36 +0000 (08:16 +0000)]
Remove shouldAbortEarly cruft from master.cfg
https://bugs.webkit.org/show_bug.cgi?id=136779

Reviewed by Csaba Osztrogonác.

* BuildSlaveSupport/build.webkit.org-config/master.cfg:
(ConfigureBuild.start):
(RunWebKitTests.start):
(RunWebKit2Tests.start): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173653 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoAdd version of drawFocusIfNeeded that take a Path2D parameter.
commit-queue@webkit.org [Tue, 16 Sep 2014 08:13:06 +0000 (08:13 +0000)]
Add version of drawFocusIfNeeded that take a Path2D parameter.
https://bugs.webkit.org/show_bug.cgi?id=136846

Patch by Jinho Bang <jinho.bang@samsung.com> on 2014-09-16
Reviewed by Dirk Schulze.

Source/WebCore:

Tests: fast/canvas/draw-focus-if-needed-with-path.html

* html/canvas/CanvasRenderingContext2D.cpp:
(WebCore::CanvasRenderingContext2D::drawFocusIfNeeded):
(WebCore::CanvasRenderingContext2D::drawFocusIfNeededInternal):
* html/canvas/CanvasRenderingContext2D.h:
* html/canvas/CanvasRenderingContext2D.idl:

LayoutTests:

* fast/canvas/draw-focus-if-needed-expected.txt: Added.
* fast/canvas/draw-focus-if-needed-with-path-expected.txt: Added.
* fast/canvas/draw-focus-if-needed-with-path.html: Added.
* fast/canvas/draw-focus-if-needed.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173652 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years ago[GTK] Add WaylandDisplay
zandobersek@gmail.com [Tue, 16 Sep 2014 06:53:30 +0000 (06:53 +0000)]
[GTK] Add WaylandDisplay
https://bugs.webkit.org/show_bug.cgi?id=136216

Reviewed by Martin Robinson.

Add the WaylandDisplay class. This class is used in the LayerTreeHostGtk,
in the WebProcess, and controls the connection to the nested Wayland
compositor in the UIProcess. Only one instance of the class is used in the
WebProcess, accessible via the static WaylandDisplay::instance() method.

The WaylandDisplay constructor performs the EGL initialization while also
querying the Wayland register, resulting in initializing client-side objects
that can interact with the wl_compositor and wl_webkitgtk interfaces
implemented in the nested Wayland compositor. The single class instance is
only deemed properly initialized if it picked up the Wayland interface objects
and has properly acquired the EGL display and config.

WaylandDisplay::createSurface() is called by the LayerTreeHostGtk during
its own initialization. The method creates a new wl_surface object via the
wl_compositor interface and a new wl_egl_window object that's based on that
surface. For the wl_egl_window object, we fall back to a width or height of 1
in case the passed-in value for either is 0. This avoids problems in Mesa
where widths or heights of 0 are not supported.

We associate the created surface with the passed-in widget ID, as provided
by LayerTreeHostGtk, via the wl_webkitgtk interface. This enables proper
mapping of Wayland surfaces and the GtkWidgets in the UIProcess and makes
it possible for the nested Wayland compositor to correctly determine which
GtkWidget has to be redrawn after some surface has been committed.

WaylandDisplay::createSharingGLContext() creates a new GLContextEGL object
that's to be used as a sharing context. The method creates a new surface
via the wl_compositor interface and uses it to create a dummy native EGL
window that's 1x1px in size. The GLContextEGL object is then created
through the static GLContextEGL::createWindowContext() method.

* PlatformGTK.cmake:
* platform/graphics/wayland/WaylandDisplay.cpp: Added.
(WebCore::WaylandDisplay::globalCallback):
(WebCore::WaylandDisplay::globalRemoveCallback):
(WebCore::WaylandDisplay::instance):
(WebCore::WaylandDisplay::WaylandDisplay):
(WebCore::WaylandDisplay::createSurface):
(WebCore::WaylandDisplay::createSharingGLContext):
* platform/graphics/wayland/WaylandDisplay.h: Added.
(WebCore::WaylandDisplay::nativeDisplay):
(WebCore::WaylandDisplay::eglDisplay):
* platform/graphics/wayland/WaylandSurface.cpp:
(WebCore::WaylandSurface::~WaylandSurface): Assert that the WaylandDisplay
instance is present before going on to destroy the surface resources.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173651 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoRename LiveNodeLists / HTMLCollections's nodeMatches() to elementMatches()
cdumez@apple.com [Tue, 16 Sep 2014 05:36:37 +0000 (05:36 +0000)]
Rename LiveNodeLists / HTMLCollections's nodeMatches() to elementMatches()
https://bugs.webkit.org/show_bug.cgi?id=136844

Reviewed by Sam Weinig.

Rename LiveNodeLists / HTMLCollections's nodeMatches() to elementMatches()
for clarity as the argument is an Element.

No new tests, no behavior change.

* dom/ClassNodeList.h:
(WebCore::ClassNodeList::elementMatches):
(WebCore::ClassNodeList::nodeMatches): Deleted.
* dom/DocumentOrderedMap.cpp:
(WebCore::keyMatchesWindowNamedItem):
(WebCore::keyMatchesDocumentNamedItem):
* dom/Element.cpp:
(WebCore::Element::updateNameForDocument):
(WebCore::Element::updateIdForDocument):
* dom/LiveNodeList.cpp:
(WebCore::LiveNodeList::namedItem):
* dom/LiveNodeList.h:
(WebCore::CachedLiveNodeList<NodeListType>::collectionBegin):
(WebCore::CachedLiveNodeList<NodeListType>::collectionLast):
(WebCore::CachedLiveNodeList<NodeListType>::collectionTraverseForward):
(WebCore::CachedLiveNodeList<NodeListType>::collectionTraverseBackward):
* dom/NameNodeList.h:
(WebCore::NameNodeList::elementMatches):
(WebCore::NameNodeList::nodeMatches): Deleted.
* dom/TagNodeList.h:
(WebCore::TagNodeList::elementMatches):
(WebCore::HTMLTagNodeList::elementMatches):
(WebCore::TagNodeList::nodeMatches): Deleted.
(WebCore::HTMLTagNodeList::nodeMatches): Deleted.
* html/HTMLCollection.cpp:
(WebCore::isMatchingElement):
* html/HTMLNameCollection.cpp:
(WebCore::WindowNameCollection::elementMatchesIfNameAttributeMatch):
(WebCore::WindowNameCollection::elementMatches):
(WebCore::DocumentNameCollection::elementMatchesIfIdAttributeMatch):
(WebCore::DocumentNameCollection::elementMatchesIfNameAttributeMatch):
(WebCore::DocumentNameCollection::elementMatches):
(WebCore::WindowNameCollection::nodeMatchesIfNameAttributeMatch): Deleted.
(WebCore::WindowNameCollection::nodeMatches): Deleted.
(WebCore::DocumentNameCollection::nodeMatchesIfIdAttributeMatch): Deleted.
(WebCore::DocumentNameCollection::nodeMatchesIfNameAttributeMatch): Deleted.
(WebCore::DocumentNameCollection::nodeMatches): Deleted.
* html/HTMLNameCollection.h:
* html/LabelsNodeList.cpp:
(WebCore::LabelsNodeList::elementMatches):
(WebCore::LabelsNodeList::nodeMatches): Deleted.
* html/LabelsNodeList.h:
* html/RadioNodeList.cpp:
(WebCore::RadioNodeList::elementMatches):
(WebCore::RadioNodeList::nodeMatches): Deleted.
* html/RadioNodeList.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173649 268f45cc-cd09-0410-ab3c-d52691b4dbfc

5 years agoUse an AtomicString as key for caching ClassNodeList objects
cdumez@apple.com [Tue, 16 Sep 2014 04:09:58 +0000 (04:09 +0000)]
Use an AtomicString as key for caching ClassNodeList objects
https://bugs.webkit.org/show_bug.cgi?id=136830

Reviewed by Benjamin Poulain.

Use an AtomicString as key for caching ClassNodeList objects instead of
a String. ClassNodeList is the only type using a String instead of an
AtomicString as key in the cache HashTable. This brings some
complexity.

I believe this was done to avoid unnecessarily atomizing the String,
for performance reasons. However, at the moment, the String gets
atomized anyway when constructing the ClassNodeList object. This is
because the ClassNodeList::m_classNames member is of SpaceSplitString
type and the SpaceSplitString constructor takes an AtomicString in
argument.

Using an AtomicString to cache ClassNodeLists simplifies the code quite
a bit and decreases the size of NodeListsNodeData as well.

Test: fast/dom/getElementsByClassName/conflict-tag-name.html

* WebCore.order:
Remove symbol corresponding to addCacheWithName() as it was removed.

* dom/ClassNodeList.cpp:
(WebCore::ClassNodeList::~ClassNodeList):
Update the constructor to take an AtomicString in argument instead of
a String, for clarity. The String gets atomized when initializing
m_classNames anyway.

(WebCore::ClassNodeList::ClassNodeList):
Call removeCacheWithAtomicName() instead of removeCacheWithName() now
that m_originalClassNames is an AtomicString.

* dom/ClassNodeList.h:
Use AtomicString instead of String type for classNames, in both the
constructor argument and the m_originalClassNames data member.

* dom/ContainerNode.cpp:
(WebCore::ContainerNode::getElementsByClassName):
Call addCacheWithAtomicName() instead of addCacheWithName() now that
addCacheWithName() has been removed.

* dom/Node.cpp:
(WebCore::NodeListsNodeData::invalidateCaches):
Stop invalidating m_nameCaches as this HashMap no longer exists.

* dom/NodeRareData.h:
(WebCore::NodeListsNodeData::NodeListCacheMapEntryHash::hash):
(WebCore::NodeListsNodeData::NodeListCacheMapEntryHash::equal):
(WebCore::NodeListsNodeData::isEmpty):
(WebCore::NodeListsNodeData::adoptDocument):
(WebCore::NodeListsNodeData::namedNodeListKey):
(WebCore::NodeListsNodeData::deleteThisAndUpdateNodeRareDataIfAboutToRemoveLastList):
(WebCore::NodeListsNodeData::addCacheWithName): Deleted.
(WebCore::NodeListsNodeData::removeCacheWithName): Deleted.
- Drop addCacheWithName() / removeCacheWithName() now that no NodeList
  uses a String as HashMap key.
- Drop m_nameCaches now that ClassNodeLists are cached in
  m_atomicNameCaches instead.
- Remove StringType template parameter and hardcode AtomicString
  instead.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@173648 268f45cc-cd09-0410-ab3c-d52691b4dbfc