Purge PassRefPtr create() factory functions in html
[WebKit-https.git] / Source / WebCore / ChangeLog
index ea22507..2325709 100644 (file)
@@ -1,3 +1,917 @@
+2015-05-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
+
+        Purge PassRefPtr create() factory functions in html 
+        https://bugs.webkit.org/show_bug.cgi?id=144522
+
+        Reviewed by Darin Adler.
+
+        Return Ref instead of PassRefPtr in create() factory functions in html, because the factory
+        can't return null. This patch is a first step to purge in html. Following patch is going to
+        be uploaded.
+
+        No new tests, no behavior changes.
+
+        * html/FileInputType.cpp:
+        (WebCore::UploadButtonElement::create):
+        (WebCore::UploadButtonElement::createForMultiple):
+        * html/MediaController.cpp:
+        (MediaController::create):
+        * html/MediaController.h:
+        * html/canvas/WebGLBuffer.cpp:
+        (WebCore::WebGLBuffer::create):
+        * html/canvas/WebGLBuffer.h:
+        * html/canvas/WebGLContextAttributes.cpp:
+        (WebCore::WebGLContextAttributes::create):
+        * html/canvas/WebGLContextAttributes.h:
+        * html/canvas/WebGLContextGroup.cpp:
+        (WebCore::WebGLContextGroup::create):
+        * html/canvas/WebGLContextGroup.h:
+        * html/canvas/WebGLFramebuffer.cpp:
+        (WebCore::WebGLFramebuffer::create):
+        * html/canvas/WebGLFramebuffer.h:
+        * html/canvas/WebGLProgram.cpp:
+        (WebCore::WebGLProgram::create):
+        * html/canvas/WebGLProgram.h:
+        * html/canvas/WebGLQuery.cpp:
+        (WebCore::WebGLQuery::create):
+        * html/canvas/WebGLQuery.h:
+        * html/canvas/WebGLRenderbuffer.cpp:
+        (WebCore::WebGLRenderbuffer::create):
+        * html/canvas/WebGLRenderbuffer.h:
+        * html/canvas/WebGLSampler.cpp:
+        (WebCore::WebGLSampler::create):
+        * html/canvas/WebGLSampler.h:
+        * html/canvas/WebGLShader.cpp:
+        (WebCore::WebGLShader::create):
+        * html/canvas/WebGLShader.h:
+        * html/canvas/WebGLShaderPrecisionFormat.cpp:
+        (WebCore::WebGLShaderPrecisionFormat::create):
+        * html/canvas/WebGLShaderPrecisionFormat.h:
+        * html/canvas/WebGLSync.cpp:
+        (WebCore::WebGLSync::create):
+        * html/canvas/WebGLSync.h:
+        * html/canvas/WebGLTexture.cpp:
+        (WebCore::WebGLTexture::create):
+        * html/canvas/WebGLTexture.h:
+        * html/canvas/WebGLTransformFeedback.cpp:
+        (WebCore::WebGLTransformFeedback::create):
+        * html/canvas/WebGLTransformFeedback.h:
+        * html/canvas/WebGLUniformLocation.cpp:
+        (WebCore::WebGLUniformLocation::create):
+        * html/canvas/WebGLUniformLocation.h:
+        * html/canvas/WebGLVertexArrayObject.cpp:
+        (WebCore::WebGLVertexArrayObject::create):
+        * html/canvas/WebGLVertexArrayObject.h:
+        * html/canvas/WebGLVertexArrayObjectOES.cpp:
+        (WebCore::WebGLVertexArrayObjectOES::create):
+        * html/canvas/WebGLVertexArrayObjectOES.h:
+
+2015-05-02  Andreas Kling  <akling@apple.com>
+
+        Style recalc always causes detach from shared StyleRareNonInheritedData.
+        <https://webkit.org/b/144530>
+
+        Reviewed by Darin Adler.
+
+        StyleResolver::adjustStyleForMaskImages() was always calling RenderStyle::ensureMaskLayers()
+        when we were in style recalc. This caused the RenderStyle to COW off from the previously
+        shared StyleRareNonInheritedData even if nothing actually changes.
+
+        This patch adds an early return to adjustStyleForMaskImages() when the previous style had
+        no mask images. Further refinements could be made, but this covers the most common case where
+        we don't have any mask images at all.
+
+        * css/StyleResolver.cpp:
+        (WebCore::StyleResolver::adjustStyleForMaskImages):
+
+2015-05-02  Alexey Proskuryakov  <ap@apple.com>
+
+        fast/frames/flattening/iframe-flattening-resize-event-count.html times out on Yosemite WK2
+        https://bugs.webkit.org/show_bug.cgi?id=144155
+
+        Unreviewed, will roll out soon.
+
+        Added some temporary logging to better see if alignment is indeed the culprit.
+
+        * page/Settings.cpp: (WebCore::Settings::setDOMTimerAlignmentInterval):
+
+2015-05-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
+
+        Use Ref instead of PassRefPtr in WebCore/svg 
+        https://bugs.webkit.org/show_bug.cgi?id=144479
+
+        Reviewed by Darin Adler.
+
+        Remove remained PassRefPtr in svg.
+
+        No new tests, no behavior changes.
+
+        * svg/SVGFontFaceSrcElement.cpp:
+        (WebCore::SVGFontFaceSrcElement::srcValue):
+        * svg/SVGFontFaceSrcElement.h:
+        * svg/properties/SVGPropertyInfo.h:
+
+2015-05-02  Simon Fraser  <simon.fraser@apple.com>
+
+        Small cleanup in BitmapImage
+        https://bugs.webkit.org/show_bug.cgi?id=144515
+
+        Reviewed by Darin Adler.
+
+        Use an enum for the internalAdvanceAnimation() argument, with a default.
+        
+        Wrap the clearing of the timer in a clearTimer() function, called from
+        stopAnimation(). This fixes the jarring reading of BitmapImage::internalAdvanceAnimation(),
+        which used to start by calling stopAnimation().
+
+        * platform/graphics/BitmapImage.cpp:
+        (WebCore::BitmapImage::clearTimer):
+        (WebCore::BitmapImage::startAnimation):
+        (WebCore::BitmapImage::stopAnimation):
+        (WebCore::BitmapImage::advanceAnimation):
+        (WebCore::BitmapImage::internalAdvanceAnimation):
+        * platform/graphics/BitmapImage.h:
+        * platform/graphics/Image.h:
+
+2015-05-01  Simon Fraser  <simon.fraser@apple.com>
+
+        Avoid compositing updates after style recalcs which have no compositing implications
+        https://bugs.webkit.org/show_bug.cgi?id=144502
+
+        Reviewed by Darin Adler.
+        
+        After r183461, we have reliable information about whether a style change with zero
+        diff can be reliably ignored. Use that information to track whether a given
+        recalcStyle() does anything which should force a compositing update.
+        
+        This eliminates up to 40% of the post-recalcStyle compositing updates on some pages.
+        
+        Add Internals API to test.
+
+        Test: compositing/updates/no-style-change-updates.html
+
+        * dom/Document.cpp:
+        (WebCore::Document::recalcStyle): Tell the FrameView we're going to recalc style.
+        * page/FrameView.cpp:
+        (WebCore::FrameView::willRecalcStyle): Pass it on to the compositor.
+        (WebCore::FrameView::updateCompositingLayersAfterStyleChange): Move the code
+        that was here into RenderLayerCompositor::didRecalcStyleWithNoPendingLayout().
+        * page/FrameView.h:
+        * rendering/RenderLayerCompositor.cpp:
+        (WebCore::RenderLayerCompositor::willRecalcStyle): Reset the m_layerNeedsCompositingUpdate flag.
+        (WebCore::RenderLayerCompositor::didRecalcStyleWithNoPendingLayout): Bail on the update if
+        no layers changed.
+        (WebCore::RenderLayerCompositor::updateCompositingLayers): Logging. Increment m_compositingUpdateCount,
+        which is used for testing.
+        (WebCore::RenderLayerCompositor::layerStyleChanged): Set the m_layerNeedsCompositingUpdate flag.
+        (WebCore::RenderLayerCompositor::startTrackingCompositingUpdates): Reset the counter.
+        (WebCore::RenderLayerCompositor::compositingUpdateCount):
+        * rendering/RenderLayerCompositor.h:
+        * testing/Internals.cpp:
+        (WebCore::Internals::startTrackingCompositingUpdates):
+        (WebCore::Internals::compositingUpdateCount):
+        * testing/Internals.h:
+        * testing/Internals.idl:
+
+2015-05-01  Andreas Kling  <akling@apple.com>
+
+        Reproducible crash removing name attribute from <img> node
+        <https://webkit.org/b/144371>
+        <rdar://problem/17198583>
+
+        Reviewed by Darin Adler.
+
+        The problem here was with HTMLImageElement::getNameAttribute(), which relies
+        on Element::hasName() to avoid slow attribute lookups when the attribute
+        is already known not to be present. Unfortunately hasName() uses an ElementData
+        flag that wasn't getting updated until after the call to parseAttribute().
+
+        This patch fixes the issue by moving the code that updates the hasName() flag
+        before the parseAttribute() virtual dispatch.
+
+        Test: fast/dom/HTMLImageElement/remove-name-id-attribute-from-image.html
+
+        * dom/Element.cpp:
+        (WebCore::Element::attributeChanged):
+
+2015-05-01  Eric Carlson  <eric.carlson@apple.com>
+
+        Postpone caption style sheet creation
+        https://bugs.webkit.org/show_bug.cgi?id=144499
+
+        Reviewed by Simon Fraser.
+        
+        Generating and inserting the caption user style sheet is expensive so don't do it until
+        we see a text track, and only do it for the first video element in a PageGroup.
+
+        Test: media/track/track-user-stylesheet.html
+
+        * html/HTMLMediaElement.cpp:
+        (WebCore::HTMLMediaElement::registerWithDocument): Only register for caption preferences
+        changes if we have done so before.
+        (WebCore::HTMLMediaElement::unregisterWithDocument): Only unregister for caption preferences
+        changes if we registered for them.
+        (WebCore::HTMLMediaElement::addTextTrack): Register for caption preference changes.
+        * html/HTMLMediaElement.h:
+
+        * page/CaptionUserPreferencesMediaAF.cpp:
+        (WebCore::CaptionUserPreferencesMediaAF::setInterestedInCaptionPreferenceChanges): Only
+        generate the style sheet when called for the first time.
+
+2015-05-01  Dean Jackson  <dino@apple.com>
+
+        [iOS] Scrubber display is broken if the buffered range is empty
+        https://bugs.webkit.org/show_bug.cgi?id=144511
+
+        Reviewed by Eric Carlson.
+
+        Make sure that if video.buffered returns an empty list, we
+        still draw the scrubber background correctly.
+
+        * Modules/mediacontrols/mediaControlsiOS.js:
+        (ControllerIOS.prototype.drawTimelineBackground): Ensure that buffered
+        is always at least as big as the current time.
+
+2015-05-01  Roger Fong  <roger_fong@apple.com>
+
+        Media control volume slider should be vertical to avoid flickering issues.
+        https://bugs.webkit.org/show_bug.cgi?id=144496.
+        <rdar://problem/20451328>
+
+        Reviewed by Darin Adler.
+
+        * Modules/mediacontrols/mediaControlsApple.css:
+        Resize and reposition the volume box so that it does not have to rotated via a transform.
+        (audio::-webkit-media-controls-panel):
+        (audio::-webkit-media-controls-panel .volume-box): Also set overflow: hidden so we can't interact
+             with the volume slider while the volume box is hidden.
+        (audio::-webkit-media-controls-panel .volume-box:active):
+        Rotate and position the volume slider via transforms. Make sure it's not transformed in the fullscreen controls.
+        (video::-webkit-media-controls-volume-slider):
+        (video:-webkit-full-screen::-webkit-media-controls-volume-slider):
+
+2015-05-01  Brent Fulgham  <bfulgham@apple.com>
+
+        Expand test infrastructure to support scrolling tests (Part 4): Scroll Snap Support
+        https://bugs.webkit.org/show_bug.cgi?id=144482
+
+        Reviewed by Simon Fraser.
+
+        New tests will be landed in a second patch.
+
+        The new WheelEventTestTriggers do not work properly with scroll snap points, because some test deferral notifications
+        need to be triggered for state changes on the scrolling thread. This required the following changes:
+        1. A flag indicating that we want to know about WheelEvent state changes needed to be propagated to the scrolling thread,
+           so that tests could be deferred until Scrolling thread rubberband or scroll-snap animations had completed.
+        2. The ScrollingNodeID needed to be used for registering and clearing deferrals.
+        3. The scrolling thread needed a way to dispatch messages to the main thread indicating that we should (or should not)
+           defer tests due to scrolling actions.
+
+        Note that a future patch will extend this support to include the RemoteScrollingTree.
+
+        * page/WheelEventTestTrigger.cpp:
+        (WebCore::WheelEventTestTrigger::deferTestsForReason): Added some logging features.
+        (WebCore::WheelEventTestTrigger::removeTestDeferralForReason): Ditto.
+        (WebCore::dumpState): Helper function for logging test deferral state.
+        (WebCore::WheelEventTestTrigger::triggerTestTimerFired): Added some logging features.
+        * page/scrolling/AsyncScrollingCoordinator.cpp:
+        (WebCore::AsyncScrollingCoordinator::frameViewLayoutUpdated): Make sure that the scrolling thread is
+        told that it needs to send scrolling state back to the main thread. Only do this if we are in testing mode.
+        (WebCore::AsyncScrollingCoordinator::updateScrollPositionAfterAsyncScroll): The scrollingNodeID should be
+        used as the identifier for this operation, not the page address.
+        (WebCore::AsyncScrollingCoordinator::deferTestsForReason): Added. receives messages from the scrolling thread
+        and notifies the testing infrastructure.
+        (WebCore::AsyncScrollingCoordinator::removeTestDeferralForReason): Ditto.
+        * page/scrolling/AsyncScrollingCoordinator.h:
+        * page/scrolling/ScrollingStateScrollingNode.cpp:
+        (WebCore::ScrollingStateScrollingNode::ScrollingStateScrollingNode): When cloning the ScrollingStateScrollingNode,
+        include the testing mode state.
+        (WebCore::ScrollingStateScrollingNode::setExpectsWheelEventTestTrigger): Added.
+        * page/scrolling/ScrollingStateScrollingNode.h:
+        (WebCore::ScrollingStateScrollingNode::expectsWheelEventTestTrigger): Added.
+        * page/scrolling/ScrollingTree.h:
+        (WebCore::ScrollingTree::deferTestsForReason): Added stub.
+        (WebCore::ScrollingTree::removeTestDeferralForReason): Ditto.
+        * page/scrolling/ThreadedScrollingTree.cpp:
+        (WebCore::ThreadedScrollingTree::deferTestsForReason): Added. Dispatches messages from the scrolling thread to the
+        main thread.
+        (WebCore::ThreadedScrollingTree::removeTestDeferralForReason): Ditto.
+        * page/scrolling/ThreadedScrollingTree.h:
+        * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.h:
+        * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
+        (WebCore::ScrollingTreeFrameScrollingNodeMac::ScrollingTreeFrameScrollingNodeMac): Initialize testing state flag.
+        (WebCore::ScrollingTreeFrameScrollingNodeMac::updateBeforeChildren): If the 'ExpectsWheelEventTestTrigger' flag is
+        set, update local state.
+        (WebCore::ScrollingTreeFrameScrollingNodeMac::handleWheelEvent): If we are testing, and the wheel event should be handled
+        asynchronously, tell the testing infrastructure we need to wait for the thread state to sync.
+        (WebCore::ScrollingTreeFrameScrollingNodeMac::deferTestsForReason): Added. Notifies test system we need to defer tests
+        until we notify them. Also used by the ScrollController during animations.
+        (WebCore::ScrollingTreeFrameScrollingNodeMac::removeTestDeferralForReason): Ditto.
+        * platform/Logging.h:
+        * platform/ScrollAnimator.cpp:
+        (WebCore::ScrollAnimator::deferTestsForReason): Added. Used by ScrollController during animation. This updates the
+        testing infrastructure directly, since it is running in the same process and main thread.
+        (WebCore::ScrollAnimator::removeTestDeferralForReason): Ditto.
+        * platform/ScrollAnimator.h:
+        * platform/cocoa/ScrollController.h:
+        (WebCore::ScrollControllerClient::deferTestsForReason): Change client API.
+        (WebCore::ScrollControllerClient::removeTestDeferralForReason): Ditto.
+        (WebCore::ScrollControllerClient::testTrigger): Deleted.
+        * platform/cocoa/ScrollController.mm:
+        (WebCore::ScrollController::startSnapRubberbandTimer): Use new client API.
+        (WebCore::ScrollController::stopSnapRubberbandTimer): Ditto.
+        (WebCore::ScrollController::startScrollSnapTimer): Ditto.
+        (WebCore::ScrollController::stopScrollSnapTimer): Ditto.
+
+2015-05-01  Joseph Pecoraro  <pecoraro@apple.com>
+
+        Inline some small methods instead of them being EXPORT'd and out of line
+        https://bugs.webkit.org/show_bug.cgi?id=144498
+
+        Reviewed by Darin Adler.
+
+        * page/Page.cpp:
+        (WebCore::Page::testTrigger): Deleted.
+        (WebCore::Page::clearTrigger): Deleted.
+        (WebCore::Page::expectsWheelEventTriggers): Deleted.
+        * page/Page.h:
+        (WebCore::Page::testTrigger):
+        (WebCore::Page::clearTrigger):
+        (WebCore::Page::expectsWheelEventTriggers):
+
+2015-05-01  Dan Bernstein  <mitz@apple.com>
+
+        WebCore part of <rdar://problem/8636045> Back/forward navigation to an error page in Safari breaks the back-forward list
+        https://bugs.webkit.org/show_bug.cgi?id=144501
+
+        Reviewed by Darin Adler.
+
+        Test: TestWebKitAPI/Tests/WebKit2Cocoa/LoadAlternateHTMLString.mm
+
+        Normally, loading substitute data (such as an error page) creates a new back-forward list
+        item. FrameLoader has a mechanism that detects when a substitute data load occurs during
+        handling of a provisional load error and prevents the creation of a new back-forwards list
+        item in that case if the unreachable URL is the same as the failing provisional URL. This
+        mechanism was broken in WebKit2, where handling the provisional load error is asynchronous.
+
+        The fix is to capture some state (namely, the failing provisional URL) when dispatching the
+        load error and allow it to be restored when loading the substitute data.
+
+        * loader/FrameLoader.cpp:
+        (WebCore::FrameLoader::FrameLoader): Removed initialization of
+        m_delegateIsHandlingProvisionalLoadError.
+        (WebCore::FrameLoader::shouldReloadToHandleUnreachableURL): Instead of checking
+        m_delegateIsHandlingProvisionalLoadError and if true using the provisional document loader’s
+        URL, check m_provisionalLoadErrorBeingHandledURL.
+        (WebCore::FrameLoader::checkLoadCompleteForThisFrame): Instead of checking and setting
+        m_delegateIsHandlingProvisionalLoadError, use m_provisionalLoadErrorBeingHandledURL.
+        * loader/FrameLoader.h:
+        (WebCore::FrameLoader::provisionalLoadErrorBeingHandledURL): Added this getter. The client
+        can call this from its override of dispatchDidFailProvisionalLoad and store the result.
+        (WebCore::FrameLoader::setProvisionalLoadErrorBeingHandledURL): Added this setter. The
+        client can call this prior to loading substitute data if it’s done as part of handling a
+        previously-dispatched didFailProvisionalLoad.
+
+2015-05-01  Martin Robinson  <mrobinson@igalia.com>
+
+        USE(...) macro should expect unprefixed variables
+        https://bugs.webkit.org/show_bug.cgi?id=144454
+
+        Reviewed by Daniel Bates.
+
+        * DerivedSources.make: Replace all occurrences WTF_USE with USE.
+        * PlatformEfl.cmake: Ditto.
+        * PlatformGTK.cmake: Ditto.
+        * bindings/objc/PublicDOMInterfaces.h: Ditto.
+        * config.h: Ditto.
+        * crypto/mac/SerializedCryptoKeyWrapMac.mm: Ditto.
+        * page/AlternativeTextClient.h: Ditto.
+        * platform/efl/FileSystemEfl.cpp: Ditto.
+        (WebCore::fileSystemRepresentation): Ditto.
+        * platform/graphics/cg/ImageBufferDataCG.h: Ditto.
+        * platform/graphics/cg/PDFDocumentImage.h: Ditto.
+        * platform/ios/wak/WAKAppKitStubs.h: Ditto.
+        * rendering/RenderLayerCompositor.cpp: Ditto.
+
+2015-05-01  Myles C. Maxfield  <mmaxfield@apple.com>
+
+        [OS X] Text dilation parameters are not restored after being set
+        https://bugs.webkit.org/show_bug.cgi?id=144507
+        <rdar://problem/19446938>
+
+        Reviewed by Simon Fraser and Ned Holbrook.
+
+        Create a RAII class who is in charge of saving, applying, and restoring dilation parameters. Use this class
+        around our text drawing routines.
+
+        Note that this migrates our use of CTFontSetRenderingParameters() to CTFontSetRenderingStyle().
+
+        Test: fast/text/dilation-save-restore.html
+
+        * platform/graphics/cocoa/FontCascadeCocoa.mm:
+        (WebCore::RenderingStyleSaver::RenderingStyleSaver):
+        (WebCore::RenderingStyleSaver::~RenderingStyleSaver):
+        (WebCore::showGlyphsWithAdvances):
+        * platform/spi/cg/CoreGraphicsSPI.h:
+        * platform/spi/cocoa/CoreTextSPI.h:
+
+2015-05-01  Commit Queue  <commit-queue@webkit.org>
+
+        Unreviewed, rolling out r183687.
+        https://bugs.webkit.org/show_bug.cgi?id=144505
+
+        Appears to have caused assertions (Requested by ap on
+        #webkit).
+
+        Reverted changeset:
+
+        "Media control volume slider is opaque."
+        https://bugs.webkit.org/show_bug.cgi?id=144470
+        http://trac.webkit.org/changeset/183687
+
+2015-05-01  Jeremy Jones  <jeremyj@apple.com>
+
+        Stop using deprecated AVKit SPI for enter and exit fullscreen.
+        https://bugs.webkit.org/show_bug.cgi?id=144493
+
+        Reviewed by Jer Noble.
+
+        Replace 
+            -enterFullScreenWithCompletionHandler:
+            -exitFullScreenWithCompletionHandler:
+        with
+            -enterFullScreenAnimated:completionHandler:
+            -exitFullScreenAnimated:completionHandler:
+        
+        * platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
+        (WebVideoFullscreenInterfaceAVKit::enterFullscreenStandard):
+        (WebVideoFullscreenInterfaceAVKit::exitFullscreenInternal):
+        * platform/spi/cocoa/AVKitSPI.h:
+
+2015-05-01  Roger Fong  <roger_fong@apple.com>
+
+        Media control volume slider is opaque.
+        https://bugs.webkit.org/show_bug.cgi?id=144470.
+        <rdar://problem/20770350>
+
+        Rubberstamped by Simon Fraser.
+
+        * Modules/mediacontrols/mediaControlsApple.css:
+        (audio::-webkit-media-controls-panel): 
+        Don't create a stacking context just for the controls.
+        Doing so causes the mix-blend-mode on the volume box to fail.
+
+2015-05-01  Joanmarie Diggs  <jdiggs@igalia.com>
+
+        [GTK] REGRESSION(183368): It made editing tests assert
+        https://bugs.webkit.org/show_bug.cgi?id=144447
+
+        Reviewed by Chris Fleizach.
+
+        Don't notify platforms of editing actions of type EditActionUnspecified.
+        If we don't know what the action is, platform accessibility APIs certainly
+        won't.
+
+        Add AXTextEditTypeAttributesChange and AXTextAttributesChanged as catch-alls
+        for formatting changes. ATK lacks finely-grained signals through which to
+        specify which text attributes have changed. If other platforms have such
+        signals, new types can always be added.
+
+        Don't post an accessible replacement notification for AXTextEditTypeCut.
+        The notification is handled by DeleteSelectionCommand. It is possible for
+        white space to be inserted in preparation for a cut which was triggering
+        a not-reached assert, but insertion of whitespace in preparation for a
+        cut does not seem notification-worthy.
+
+        Treat AXTextEditTypeDictation the same as other insertion types (insert,
+        paste, typing) with respect to replacement notifications.
+
+        No new tests. The hundreds of editing tests that had been crashing seem sufficient.
+
+        * accessibility/AXObjectCache.cpp:
+        (WebCore::AXObjectCache::showIntent):
+        (WebCore::AXObjectCache::textChangeForEditType):
+        * accessibility/AXObjectCache.h:
+        * accessibility/AXTextStateChangeIntent.h:
+        * accessibility/atk/AXObjectCacheAtk.cpp:
+        (WebCore::AXObjectCache::nodeTextChangePlatformNotification):
+        * editing/AppendNodeCommand.cpp:
+        (WebCore::AppendNodeCommand::doApply):
+        (WebCore::AppendNodeCommand::doUnapply):
+        * editing/DeleteFromTextNodeCommand.cpp:
+        (WebCore::DeleteFromTextNodeCommand::doApply):
+        (WebCore::DeleteFromTextNodeCommand::doUnapply):
+        * editing/EditCommand.cpp:
+        (WebCore::EditCommand::applyEditType):
+        (WebCore::EditCommand::unapplyEditType):
+        * editing/InsertIntoTextNodeCommand.cpp:
+        (WebCore::InsertIntoTextNodeCommand::doApply):
+        (WebCore::InsertIntoTextNodeCommand::doUnapply):
+        * editing/InsertNodeBeforeCommand.cpp:
+        (WebCore::InsertNodeBeforeCommand::doApply):
+        (WebCore::InsertNodeBeforeCommand::doUnapply):
+        * editing/ReplaceInsertIntoTextNodeCommand.cpp:
+        (WebCore::ReplaceInsertIntoTextNodeCommand::notifyAccessibilityForTextChange):
+
+2015-05-01  Brady Eidson  <beidson@apple.com>
+
+        Add API to disable meta refreshes.
+        <rdar://problem/20333198> and https://bugs.webkit.org/show_bug.cgi?id=144269
+
+        Reviewed by Alexey Proskuryakov.
+
+        Test: loader/meta-refresh-disabled.html
+
+        * dom/Document.cpp:
+        (WebCore::Document::processHttpEquiv): Bail early if Settings have meta refreshes disabled.
+        * page/Settings.in:
+
+2015-05-01  Martin Robinson  <mrobinson@igalia.com>
+
+        [Freetype] Properly support synthetic oblique in vertical text
+        https://bugs.webkit.org/show_bug.cgi?id=144492
+
+        Reviewed by Sergio Villar Senin.
+
+        No new tests. Covered by existing tests.
+
+        * platform/graphics/freetype/FontPlatformDataFreeType.cpp:
+        (WebCore::FontPlatformData::initializeWithFontFace): Skew vertical when using synthetic
+        oblique for vertical text.
+
+2015-05-01  Joanmarie Diggs  <jdiggs@igalia.com>
+
+        AX: [ATK] REGRESSION: accessibility/canvas-fallback-content.html now crashes
+        https://bugs.webkit.org/show_bug.cgi?id=144481
+
+        Reviewed by Chris Fleizach.
+
+        Canvas fallback content is allowed to be focusable if the canvas is displayed
+        and visible. Update the style when creating an accessible object for a node in
+        a canvas subtree and before Element::isFocusable() gets called.
+
+        No new tests. The existing, crashing test no longer crashes.
+
+        * accessibility/AXObjectCache.cpp:
+        (WebCore::AXObjectCache::getOrCreate):
+
+2015-05-01  Alex Christensen  <achristensen@webkit.org>
+
+        [Content Extensions] Add CombinedURLFilters debugging code.
+        https://bugs.webkit.org/show_bug.cgi?id=144491
+
+        Reviewed by Daniel Bates.
+
+        No change in behavior.
+
+        * contentextensions/CombinedURLFilters.cpp:
+        (WebCore::ContentExtensions::recursiveMemoryUsed):
+        (WebCore::ContentExtensions::CombinedURLFilters::memoryUsed):
+        (WebCore::ContentExtensions::prefixTreeVertexToString):
+        (WebCore::ContentExtensions::recursivePrint):
+        (WebCore::ContentExtensions::CombinedURLFilters::print):
+        (WebCore::ContentExtensions::CombinedURLFilters::addPattern):
+        (WebCore::ContentExtensions::generateNFAForSubtree):
+        (WebCore::ContentExtensions::CombinedURLFilters::processNFAs):
+        * contentextensions/CombinedURLFilters.h:
+        * contentextensions/NFA.cpp:
+        (WebCore::ContentExtensions::NFA::memoryUsed):
+        * contentextensions/NFA.h:
+        * contentextensions/Term.h:
+        (WebCore::ContentExtensions::quantifierToString):
+        (WebCore::ContentExtensions::Term::toString):
+
+2015-05-01  Eric Carlson  <eric.carlson@apple.com>
+
+        Fix text track language selection logic
+        https://bugs.webkit.org/show_bug.cgi?id=144467
+
+        Reviewed by Brent Fulgham.
+
+        No new tests, media/track/track-language-preference.html was updated.
+
+        * html/HTMLMediaElement.cpp:
+        (WebCore::HTMLMediaElement::configureTextTrackGroup): Correct a minor style typo.
+        (WebCore::HTMLMediaElement::configureTextTrackDisplay): Update logging.
+
+        * page/CaptionUserPreferences.cpp:
+        (WebCore::CaptionUserPreferences::textTrackSelectionScore): Minor cleanup.
+        (WebCore::CaptionUserPreferences::textTrackLanguageSelectionScore): Give exact matches a
+        higher score.
+
+        * page/CaptionUserPreferencesMediaAF.cpp:
+        (WebCore::CaptionUserPreferencesMediaAF::textTrackSelectionScore): Update for 
+        indexOfBestMatchingLanguageInList change.
+
+        * platform/Language.cpp:
+        (WebCore::indexOfBestMatchingLanguageInList): Add parameter for exact match. Convert the
+        passed language to lower case as we do with the preferred languages.
+        * platform/Language.h:
+
+2015-05-01  Martin Robinson  <mrobinson@igalia.com>
+
+        [Freetype] Add support for the font-synthesis property
+        https://bugs.webkit.org/show_bug.cgi?id=144472
+
+        Reviewed by Sergio Villar Senin.
+
+        No new tests. This causes fast/css3-text/font-synthesis.html to pass
+        for WebKitGTK+.
+
+        * platform/graphics/freetype/FontPlatformDataFreeType.cpp:
+        (WebCore::FontPlatformData::FontPlatformData): Only use synthetic bold when the font description allows it.
+        (WebCore::FontPlatformData::initializeWithFontFace): Ditto for synthetic oblique.
+
+2015-05-01  Martin Robinson  <mrobinson@igalia.com>
+
+        [GTK] New CORS tests from r183280 fail on WebKitGTK+
+        https://bugs.webkit.org/show_bug.cgi?id=144469
+
+        Reviewed by Sergio Villar Senin.
+
+        No new tests. This causes failing tests to pass.
+
+        * platform/network/soup/ResourceHandleSoup.cpp:
+        (WebCore::doRedirect): Clear the origin header on cross-origin redirects.
+
+2015-04-30  Alex Christensen  <achristensen@webkit.org>
+
+        Compile fix when using content extensions debugging code.
+
+        * contentextensions/ContentExtensionCompiler.cpp:
+        (WebCore::ContentExtensions::compileRuleList):
+        * contentextensions/ContentExtensionsDebugging.h:
+        * contentextensions/DFA.cpp:
+        (WebCore::ContentExtensions::printTransitions):
+        (WebCore::ContentExtensions::DFA::debugPrintDot):
+        * contentextensions/DFANode.h:
+        * contentextensions/NFA.cpp:
+        (WebCore::ContentExtensions::NFA::memoryUsed):
+        (WebCore::ContentExtensions::NFA::debugPrintDot):
+
+2015-04-30  Dan Bernstein  <mitz@apple.com>
+
+        Fixed the build for <rdar://problem/20758514>
+
+        * platform/graphics/avfoundation/AudioSourceProviderAVFObjC.mm:
+        (WebCore::AudioSourceProviderAVFObjC::destroyMix): Don’t pass nil to
+        -[AVMutableAudioMix setInputParameters:].
+
+2015-04-30  Daniel Bates  <dabates@apple.com>
+
+        Clean up: Remove unnecessary runtime computation of string length
+        https://bugs.webkit.org/show_bug.cgi?id=144483
+
+        Reviewed by Joseph Pecoraro.
+
+        Following <http://trac.webkit.org/changeset/183649>, WebCore::fullyQualifiedInfoTableName()
+        computes strlen() of the first string literal as part of concatenating two string literals.
+        It is sufficient to use sizeof() - 1 instead of strlen() to compute the length of the first
+        string literal because the size of the string literal is known at compile time.
+
+        * Modules/webdatabase/DatabaseBackendBase.cpp:
+        (WebCore::fullyQualifiedInfoTableName):
+
+2015-04-30  Dean Jackson  <dino@apple.com>
+
+        Expose -apple-system as a font family
+        https://bugs.webkit.org/show_bug.cgi?id=144484
+        <rdar://problem/20767330>
+
+        Reviewed by Tim Horton.
+
+        Accept "-apple-system" for the font-family property, and
+        rename "-apple-system-font-monospaced-numbers" to
+        "-apple-system-monospaced-numbers".
+
+        Also change the media controls to use the new name.
+
+        Covered by existing tests and this new one:
+            fast/text/system-font-legacy-name.html
+
+        * Modules/mediacontrols/mediaControlsApple.css:
+        (audio::-webkit-media-controls-time-remaining-display):
+        * Modules/mediacontrols/mediaControlsiOS.css:
+        (::-webkit-media-controls):
+        (audio::-webkit-media-controls-time-remaining-display):
+        (audio::-webkit-media-controls-status-display):
+        * platform/graphics/ios/FontCacheIOS.mm: Add support for the new name.
+        (WebCore::createCTFontWithFamilyNameAndWeight):
+        * platform/graphics/mac/FontCacheMac.mm: Ditto.
+        (WebCore::fontWithFamily):
+
+2015-04-30  Javier Fernandez  <jfernandez@igalia.com>
+
+        [CSS Grid Layout] overflow-position keyword for align and justify properties.
+        https://bugs.webkit.org/show_bug.cgi?id=144235
+
+        Reviewed by Sergio Villar Senin.
+
+        When the alignment subject is larger than the alignment container,
+        it will overflow. Some alignment modes, if honored in this
+        situation, may cause data loss; an overflow alignment mode can be
+        explicitly specified to avoid this.
+
+        This patch implements overflow-keyword handling for Grid Layout on
+        align-self and justify-self properties.
+
+        Test: fast/css-grid-layout/grid-align-justify-overflow.html
+
+        * rendering/RenderGrid.cpp:
+        (WebCore::computeOverflowAlignmentOffset):
+        (WebCore::RenderGrid::rowPositionForChild):
+        (WebCore::RenderGrid::columnPositionForChild):
+        (WebCore::RenderGrid::rowAxisPositionForChild): Deleted.
+        * rendering/style/RenderStyle.cpp:
+        (WebCore::resolveAlignmentData):
+        (WebCore::resolveJustificationData):
+        (WebCore::RenderStyle::resolveAlignment):
+        (WebCore::RenderStyle::resolveAlignmentOverflow):
+        (WebCore::RenderStyle::resolveJustification):
+        (WebCore::RenderStyle::resolveJustificationOverflow):
+        * rendering/style/RenderStyle.h:
+
+2015-04-30  Jon Honeycutt  <jhoneycutt@apple.com>
+
+        Rebaseline bindings tests results after r183648.
+
+        Unreviewed.
+
+        * bindings/scripts/test/JS/JSTestActiveDOMObject.cpp:
+        (WebCore::jsTestActiveDOMObjectPrototypeFunctionPostMessage):
+        * bindings/scripts/test/JS/JSTestCustomNamedGetter.cpp:
+        (WebCore::jsTestCustomNamedGetterPrototypeFunctionAnotherFunction):
+        * bindings/scripts/test/JS/JSTestInterface.cpp:
+        (WebCore::JSTestInterfaceConstructor::constructJSTestInterface):
+        (WebCore::setJSTestInterfaceConstructorImplementsStaticAttr):
+        (WebCore::setJSTestInterfaceImplementsStr2):
+        (WebCore::setJSTestInterfaceConstructorSupplementalStaticAttr):
+        (WebCore::setJSTestInterfaceSupplementalStr2):
+        (WebCore::jsTestInterfacePrototypeFunctionImplementsMethod2):
+        (WebCore::jsTestInterfacePrototypeFunctionSupplementalMethod2):
+        * bindings/scripts/test/JS/JSTestNamedConstructor.cpp:
+        (WebCore::JSTestNamedConstructorNamedConstructor::constructJSTestNamedConstructor):
+        * bindings/scripts/test/JS/JSTestNondeterministic.cpp:
+        (WebCore::setJSTestNondeterministicNondeterministicWriteableAttr):
+        (WebCore::setJSTestNondeterministicNondeterministicExceptionAttr):
+        (WebCore::setJSTestNondeterministicNondeterministicGetterExceptionAttr):
+        (WebCore::setJSTestNondeterministicNondeterministicSetterExceptionAttr):
+        * bindings/scripts/test/JS/JSTestObj.cpp:
+        (WebCore::setJSTestObjConstructorStaticStringAttr):
+        (WebCore::setJSTestObjStringAttr):
+        (WebCore::setJSTestObjReflectedStringAttr):
+        (WebCore::setJSTestObjReflectedURLAttr):
+        (WebCore::setJSTestObjReflectedCustomURLAttr):
+        (WebCore::setJSTestObjStringAttrWithGetterException):
+        (WebCore::setJSTestObjStringAttrWithSetterException):
+        (WebCore::jsTestObjPrototypeFunctionVoidMethodWithArgs):
+        (WebCore::jsTestObjPrototypeFunctionByteMethodWithArgs):
+        (WebCore::jsTestObjPrototypeFunctionOctetMethodWithArgs):
+        (WebCore::jsTestObjPrototypeFunctionLongMethodWithArgs):
+        (WebCore::jsTestObjPrototypeFunctionObjMethodWithArgs):
+        (WebCore::jsTestObjPrototypeFunctionMethodThatRequiresAllArgsAndThrows):
+        (WebCore::jsTestObjPrototypeFunctionMethodWithOptionalString):
+        (WebCore::jsTestObjPrototypeFunctionMethodWithOptionalStringIsUndefined):
+        (WebCore::jsTestObjPrototypeFunctionMethodWithOptionalStringIsNullString):
+        (WebCore::jsTestObjPrototypeFunctionOverloadedMethod1):
+        (WebCore::jsTestObjPrototypeFunctionOverloadedMethod3):
+        (WebCore::jsTestObjPrototypeFunctionOverloadedMethod11):
+        (WebCore::jsTestObjConstructorFunctionOverloadedMethod12):
+        (WebCore::jsTestObjPrototypeFunctionStrictFunction):
+        (WebCore::jsTestObjPrototypeFunctionVariadicStringMethod):
+        * bindings/scripts/test/JS/JSTestOverloadedConstructors.cpp:
+        (WebCore::JSTestOverloadedConstructorsConstructor::constructJSTestOverloadedConstructors4):
+        * bindings/scripts/test/JS/JSTestTypedefs.cpp:
+        (WebCore::JSTestTypedefsConstructor::constructJSTestTypedefs):
+        (WebCore::setJSTestTypedefsStringAttrWithGetterException):
+        (WebCore::setJSTestTypedefsStringAttrWithSetterException):
+        (WebCore::jsTestTypedefsPrototypeFunctionSetShadow):
+
+2015-04-30  Simon Fraser  <simon.fraser@apple.com>
+
+        Fixed elements end up in the middle of the view with pageScale < 1
+        https://bugs.webkit.org/show_bug.cgi?id=144428
+        rdar://problem/20404982
+
+        Reviewed by Tim Horton.
+
+        When pageScale is < 1, we used fixed layout mode, and FrameView::fixedElementsLayoutRelativeToFrame()
+        returns true. However, the scrolling thread was calling the static scrollOffsetForFixedPosition()
+        hardcoding 'false' for this parameter.
+        
+        Fix by sending the value of fixedElementsLayoutRelativeToFrame over to the scrolling thread,
+        so we can use it when doing scrolling-thread fixed position stuff.
+
+        Not testable.
+
+        * page/scrolling/AsyncScrollingCoordinator.cpp:
+        (WebCore::AsyncScrollingCoordinator::frameViewLayoutUpdated):
+        * page/scrolling/ScrollingStateFrameScrollingNode.cpp:
+        (WebCore::ScrollingStateFrameScrollingNode::ScrollingStateFrameScrollingNode):
+        (WebCore::ScrollingStateFrameScrollingNode::setFixedElementsLayoutRelativeToFrame):
+        * page/scrolling/ScrollingStateFrameScrollingNode.h:
+        * page/scrolling/ScrollingTreeFrameScrollingNode.cpp:
+        (WebCore::ScrollingTreeFrameScrollingNode::ScrollingTreeFrameScrollingNode):
+        (WebCore::ScrollingTreeFrameScrollingNode::updateBeforeChildren):
+        * page/scrolling/ScrollingTreeFrameScrollingNode.h:
+        (WebCore::ScrollingTreeFrameScrollingNode::fixedElementsLayoutRelativeToFrame):
+        (WebCore::ScrollingTreeFrameScrollingNode::shouldUpdateScrollLayerPositionSynchronously): Deleted.
+        * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
+        (WebCore::ScrollingTreeFrameScrollingNodeMac::setScrollLayerPosition):
+
+2015-04-30  Beth Dakin  <bdakin@apple.com>
+
+        Remove invalid assertion from MouseEvent::create()
+        https://bugs.webkit.org/show_bug.cgi?id=144477
+
+        Reviewed by Tim Horton.
+
+        mouseforcechanged events and mouseforcewillbegin trigger this assertion. The 
+        assertion does not seem valuable, so rather than changing it, just remove it.
+        * dom/MouseEvent.cpp:
+        (WebCore::MouseEvent::create):
+
+2015-04-30  Dean Jackson  <dino@apple.com>
+
+        -apple-system-font-monospaced-numbers doesn't work on iOS
+        https://bugs.webkit.org/show_bug.cgi?id=144478
+        <rdar://problem/20544940>
+
+        Reviewed by Brent Fulgham.
+
+        Make sure to start from the system font descriptor when
+        asking for a monospaced numeric alternate.
+
+        Test: platform/ios-simulator/fast/text/system-monospaced-numbers.html
+
+        * platform/graphics/ios/FontCacheIOS.mm:
+        (WebCore::createCTFontWithFamilyNameAndWeight):
+
+2015-04-30  Oliver Hunt  <oliver@apple.com>
+
+        DOM bindings should not be using a reference type to point to a temporary object
+        https://bugs.webkit.org/show_bug.cgi?id=144474
+
+        Reviewed by Beth Dakin.
+
+        The DOM bindings will currently try and use a local reference to point
+        to a temporary object. This currently works as a by product of the compiler's
+        stack layout. This patch removes the dependency on undefined behaviour
+        by ensuring that we use a value rather than reference type.
+
+        * bindings/scripts/CodeGeneratorJS.pm:
+        (GenerateParametersCheck):
+        (GetNativeTypeForCallbacks):
+
+2015-04-30  Brady Eidson  <beidson@apple.com>
+
+        Build fix after r183646 for less enlightened platforms.
+
+        Unreviewed.
+
+        * Modules/webdatabase/DatabaseBackendBase.cpp:
+        (WebCore::fullyQualifiedInfoTableName): Windows doesn’t have stpcpy :(
+
+2015-04-30  Brady Eidson  <beidson@apple.com>
+
+        Javascript using WebSQL can create their own WebKit info table.
+        <rdar://problem/20688792> and https://bugs.webkit.org/show_bug.cgi?id=144466
+
+        Reviewed by Alex Christensen.
+
+        Test: storage/websql/alter-to-info-table.html
+
+        * Modules/webdatabase/DatabaseBackendBase.cpp:
+        (WebCore::DatabaseBackendBase::databaseInfoTableName): Return the info table name.
+        (WebCore::fullyQualifiedInfoTableName): Append "main." to the info table name.
+        (WebCore::DatabaseBackendBase::DatabaseBackendBase): Use the fully qualified name.
+        (WebCore::DatabaseBackendBase::performOpenAndVerify): Ditto.
+        (WebCore::DatabaseBackendBase::getVersionFromDatabase): Ditto.
+        (WebCore::DatabaseBackendBase::setVersionInDatabase): Ditto.
+
+2015-04-30  Beth Dakin  <bdakin@apple.com>
+
+        Should choose UIScrollView indicatorStyle based on the document background color
+        https://bugs.webkit.org/show_bug.cgi?id=144473
+        -and corresponding-
+        rdar://problem/19897699
+
+        Reviewed by Simon Fraser.
+
+        Export this function.
+        * platform/graphics/Color.h:
+
+2015-04-30  Brent Fulgham  <bfulgham@apple.com>
+
+        [Win] Allow WebKit to build without ANGLE support
+        https://bugs.webkit.org/show_bug.cgi?id=144459
+        <rdar://problem/20707307>
+
+        Reviewed by Dean Jackson.
+
+        * platform/graphics/GLContext.cpp: Use the ENABLE(GRAPHICS_CONTEXT_3D)
+        macro to avoid compiling this code if not using WEBGL/3D contexts.
+
 2015-04-29  David Hyatt  <hyatt@apple.com>
 
         Avoid containingBlock() calls when no writing mode flipping is needed.