Timeline
Apr 23, 2014:
- 11:37 PM Changeset in webkit [167750] by
-
- 10 edits in trunk
[MediaStream] Implement MediaStream active attribute
https://bugs.webkit.org/show_bug.cgi?id=131973
Patch by Praveen R Jadhav <praveen.j@samsung.com> on 2014-04-23
Reviewed by Eric Carlson.
Source/WebCore:
MediaStream .active attribute are introduced which will replace
.ended attribute. This patch implements the newly introduced attributes.
MediaStream-add-remove-tracks.html is updated to handle this scenario.
- Modules/mediastream/MediaStream.cpp:
(WebCore::MediaStream::active): Added.
(WebCore::MediaStream::setActive): Added.
(WebCore::MediaStream::addTrack): Propagates 'onactive' event when required.
(WebCore::MediaStream::removeTrack): Propagates 'oninactive' event when required.
(WebCore::MediaStream::trackDidEnd): Propagates 'oninactive' event when required.
(WebCore::MediaStream::streamDidEnd):
(WebCore::MediaStream::setStreamIsActive): Added.
- Modules/mediastream/MediaStream.h:
- Modules/mediastream/MediaStream.idl:
- dom/EventNames.h:
- platform/mediastream/MediaStreamPrivate.cpp:
(WebCore::MediaStreamPrivate::MediaStreamPrivate): Initialize .active attribute
(WebCore::MediaStreamPrivate::setEnded):
(WebCore::MediaStreamPrivate::setActive): Added.
- platform/mediastream/MediaStreamPrivate.h:
(WebCore::MediaStreamPrivate::active): Added.
LayoutTests:
MediaStream .onended attribute will be replaced with .active attribute.
Patch updates the test case to verify the .active attribute.
- fast/mediastream/MediaStream-add-remove-tracks-expected.txt:
- fast/mediastream/MediaStream-add-remove-tracks.html:
- 10:16 PM Changeset in webkit [167749] by
-
- 5 edits in trunk/Source/WebKit2
[iOS WK2] Missing content inside long overflow-scrolling:touch
https://bugs.webkit.org/show_bug.cgi?id=132105
Reviewed by Tim Horton.
PlatformCALayerRemote needs to track bounds origin as well as size,
since iOS uses bounds origin for overflow scrolling.
- Shared/mac/RemoteLayerTreePropertyApplier.mm:
(WebKit::applyPropertiesToLayer):
- Shared/mac/RemoteLayerTreeTransaction.h:
- Shared/mac/RemoteLayerTreeTransaction.mm:
(WebKit::RemoteLayerTreeTransaction::LayerProperties::LayerProperties):
(WebKit::RemoteLayerTreeTransaction::LayerProperties::encode):
(WebKit::RemoteLayerTreeTransaction::LayerProperties::decode):
(WebKit::dumpChangedLayers):
- WebProcess/WebPage/mac/PlatformCALayerRemote.cpp:
(WebKit::PlatformCALayerRemote::updateBackingStore):
(WebKit::PlatformCALayerRemote::bounds):
(WebKit::PlatformCALayerRemote::setBounds):
- 9:49 PM Changeset in webkit [167748] by
-
- 5 edits in trunk/Source
Versioning.
- 9:47 PM Changeset in webkit [167747] by
-
- 1 copy in tags/Safari-538.31
New Tag.
- 9:46 PM Changeset in webkit [167746] by
-
- 12 edits in trunk/Source/WebCore
[Cocoa] fix CF leaks found by code inspection
https://bugs.webkit.org/show_bug.cgi?id=132106
Reviewed by Andreas Kling.
- page/CaptionUserPreferencesMediaAF.cpp:
(WebCore::trackDisplayName): Added a missing adoptCF.
- platform/Language.cpp:
(WebCore::displayNameForLanguageLocale): Added a missing adoptCF.
- platform/graphics/FontPlatformData.cpp:
(WebCore::FontPlatformData::openTypeTable): Added a missing adoptCF.
- platform/graphics/avfoundation/cf/InbandTextTrackPrivateAVCF.cpp:
(WebCore::InbandTextTrackPrivateAVCF::label): Added two missing adoptCF.
- platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
(WebCore::AVFWrapper::createImageForTimeInRect): Added two missing adoptCF.
- platform/graphics/cg/PDFDocumentImage.cpp:
(WebCore::PDFDocumentImage::createPDFDocument): Added missing adoptCF.
- platform/graphics/cocoa/FontPlatformDataCocoa.mm:
(WebCore::cascadeToLastResortFontDescriptor): Added two missing adoptCF.
- platform/graphics/mac/FontMac.mm:
(WebCore::Font::primaryFontDataIsSystemFont): Added missing adoptCF.
- platform/graphics/mac/SimpleFontDataMac.mm:
(WebCore::hasCustomTracking): Added missing adoptCF.
- platform/image-decoders/ImageDecoder.h:
(WebCore::ImageDecoder::qcmsOutputDeviceProfile): Added CFRelease.
- plugins/mac/PluginPackageMac.cpp:
(WebCore::readPListFile): Added two missing adoptCF.
- 9:42 PM Changeset in webkit [167745] by
-
- 2 edits in trunk/Source/WebKit/mac
Remove Apple Dictionary workaround in WebFrameLoaderClient.
<https://webkit.org/b/132095>
Kill a hack to work around a WK1 client bug that got fixed years ago.
See <rdar://problem/6471058> for backstory.
Reviewed by Darin Adler.
- WebCoreSupport/WebFrameLoaderClient.mm:
(WebFrameLoaderClient::dispatchDidLoadResourceFromMemoryCache):
(WebFrameLoaderClient::dispatchWillSendRequest):
(applyAppleDictionaryApplicationQuirkNonInlinePart): Deleted.
(applyAppleDictionaryApplicationQuirk): Deleted.
- 9:24 PM Changeset in webkit [167744] by
-
- 2 edits in trunk/Source/WebCore
REGRESSION (r157328): popover to check into flight ba.com dismisses instantly when focusing form
https://bugs.webkit.org/show_bug.cgi?id=131949
Address the review comment.
- dom/EventDispatcher.cpp:
(WebCore::EventRelatedNodeResolver::findHostOfTreeScopeInTargetTreeScope):
- 9:11 PM Changeset in webkit [167743] by
-
- 4 edits in trunk/Source/WebCore
CachedResourceLoader hoards URLs indefinitely for no good reason.
<https://webkit.org/b/132102>
<rdar://problem/16708265>
Since we don't care about CachedResourceLoader::m_validatedURL's after
the Document has finished dispatching its initial load event, clear the
set at that point, and don't add any new URLs to it.
Reviewed by Anders Carlsson.
- dom/Document.cpp:
(WebCore::Document::dispatchWindowLoadEvent):
- loader/cache/CachedResourceLoader.cpp:
(WebCore::CachedResourceLoader::requestResource):
(WebCore::CachedResourceLoader::documentDidFinishLoadEvent):
- loader/cache/CachedResourceLoader.h:
- 8:56 PM Changeset in webkit [167742] by
-
- 10 edits in trunk/Source/WebKit2
Remove unused WKPageSetMemoryCacheClientCallsEnabled API.
<https://webkit.org/b/132098>
This API was added to WK2 back in 2011 but never actually used.
Reviewed by Anders Carlsson.
- Shared/WebPageCreationParameters.cpp:
(WebKit::WebPageCreationParameters::encode):
(WebKit::WebPageCreationParameters::decode):
- Shared/WebPageCreationParameters.h:
- UIProcess/API/C/WKPage.cpp:
(WKPageSetMemoryCacheClientCallsEnabled): Deleted.
- UIProcess/API/C/WKPage.h:
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::WebPageProxy):
(WebKit::WebPageProxy::creationParameters):
(WebKit::WebPageProxy::setMemoryCacheClientCallsEnabled): Deleted.
- UIProcess/WebPageProxy.h:
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::WebPage):
(WebKit::WebPage::setMemoryCacheMessagesEnabled): Deleted.
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/WebPage.messages.in:
- 8:43 PM Changeset in webkit [167741] by
-
- 3 edits in trunk/Source/WebCore
Canvas cache of clean URLs can grow without bounds.
<https://webkit.org/b/132091>
<rdar://problem/16695665>
Remove a silly "optimization" that kept a cache of clean URLs
that can be drawn into a canvas without tainting it, all to avoid
the "expensive" checks to determine whether it would taint.
Reviewed by Benjamin Poulain.
- html/canvas/CanvasRenderingContext.cpp:
(WebCore::CanvasRenderingContext::wouldTaintOrigin):
- html/canvas/CanvasRenderingContext.h:
- 8:31 PM Changeset in webkit [167740] by
-
- 3 edits in trunk/Source/WebKit2
[iOS WK2] Make -webkit-overflow-scrolling:touch work
https://bugs.webkit.org/show_bug.cgi?id=132097
Reviewed by Tim Horton.
Scrolling of UIScrollViews in content (for -webkit-overflow-scrolling:touch)
depends on them getting hit-tested correctly. UIKit hit testing assumes
that subviews are enclosed by their ancestors, but this is not true of
web content. In addition, we had a root layer that was zero sized.
Fix by overriding hitTest:withEvent: on our content WKViews to hit
test subviews even if they are not enclosed.
- UIProcess/ios/RemoteLayerTreeHostIOS.mm:
(-[UIView _recursiveFindDescendantViewAtPoint:withEvent:]):
(-[UIView _findDescendantViewAtPoint:withEvent:]):
(-[WKCompositingView hitTest:withEvent:]):
(-[WKTransformView hitTest:withEvent:]):
(-[WKRemoteView hitTest:withEvent:]):
(WebKit::RemoteLayerTreeHost::createLayer):
- WebProcess/WebPage/mac/RemoteLayerTreeDrawingArea.mm:
(WebKit::RemoteLayerTreeDrawingArea::updatePreferences):
Update debug borders on the root layer.
(WebKit::RemoteLayerTreeDrawingArea::mainFrameContentSizeChanged):
Size the root layer to the contents size. This isn't strictly necessary
given the hit testing overrides, but seems sensible to do anyway.
- 8:31 PM Changeset in webkit [167739] by
-
- 2 edits in trunk/Source/WebKit2
Nonopaque layers double-paint with UI-side compositing
https://bugs.webkit.org/show_bug.cgi?id=132101
Reviewed by Tim Horton.
The non-IOSurface code path was failing to clear the backing store
before painting.
- Shared/mac/RemoteLayerBackingStore.mm:
(WebKit::RemoteLayerBackingStore::display):
(WebKit::RemoteLayerBackingStore::drawInContext):
- 7:37 PM Changeset in webkit [167738] by
-
- 4 edits in trunk/Source/WebKit2
[iOS][WK2] Image and text documents sometime gets the webpage viewport configuration
https://bugs.webkit.org/show_bug.cgi?id=132099
Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-04-23
Reviewed by Tim Horton.
When didReceiveMobileDocType(), if the doctype was not XHTML mobile, we were setting
the viewport configuration to webpageParameters(). This is obviously not correct for
Image and Text documents.
This patch moves the code deciding the default configuration out of WebPage::didCommitLoad()
and use it from WebPage::didCommitLoad() and WebPage::didReceiveMobileDocType().
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::didCommitLoad):
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::didReceiveMobileDocType):
(WebKit::WebPage::resetViewportDefaultConfiguration):
- 7:25 PM Changeset in webkit [167737] by
-
- 5 edits in trunk/Source/WebKit/efl
[EFL][WK1] Add *const* keyword to _get() APIs
https://bugs.webkit.org/show_bug.cgi?id=132094
Patch by Hyowon Kim <hw1008.kim@samsung.com> on 2014-04-23
Reviewed by Gyuyoung Kim.
Some _get() EFL APIs don't have *const* keyword though the _get() APIs don't modify passed argument.
We need to add *const* keyword to those APIs.
- ewk/ewk_history.cpp:
(ewk_history_limit_get):
- ewk/ewk_history.h:
- ewk/ewk_js.cpp:
(ewk_js_object_type_get):
- ewk/ewk_js.h:
- 7:23 PM Changeset in webkit [167736] by
-
- 2 edits in trunk/Tools
Try to fix iOS build.
- DumpRenderTree/mac/DumpRenderTree.mm:
- 7:13 PM Changeset in webkit [167735] by
-
- 12 edits in trunk/Source
[iOS][WK2] Fix a few mistakes affecting the initial layout and the initial unobscured rect
https://bugs.webkit.org/show_bug.cgi?id=132093
Source/WebCore:
Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-04-23
Reviewed by Tim Horton.
Change the minimum layout size to float point values to account for size defined on retina displays.
The minimum layout size supports half-pixels, the value is rounded later when computing the layout size
in document coordinates.
- WebCore.exp.in:
- page/ViewportConfiguration.cpp:
(WebCore::ViewportConfiguration::ViewportConfiguration):
Setting the initial content size is incorrect. The layout size computation already take into account
empty size for the first layout.
Setting the content size upfront make the first computation incorrect when the viewport arguments specify
the initial scale.
(WebCore::ViewportConfiguration::setMinimumLayoutSize):
- page/ViewportConfiguration.h:
(WebCore::ViewportConfiguration::minimumLayoutSize):
Source/WebKit2:
<rdar://problem/16703237>
Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-04-23
Reviewed by Tim Horton.
The user of WKWebView can setup a size on device pixels, which can add half a point to the minimum
layout size. By rounding this up before applying the page scale, we ended up with rounding errors
on the layout size and the transitory unobscured content rect.
This patch fixes changes the minimal layout size to float point values to reduce the rounding problems.
- UIProcess/API/Cocoa/WKWebView.mm:
(setViewportConfigurationMinimumLayoutSize):
(-[WKWebView _setMinimumLayoutSizeOverrideForMinimalUI:]):
(-[WKWebView _beginAnimatedResizeWithUpdates:]):
- UIProcess/WebPageProxy.h:
- UIProcess/ios/WebPageProxyIOS.mm:
(WebKit::WebPageProxy::dynamicViewportSizeUpdate):
(WebKit::WebPageProxy::setViewportConfigurationMinimumLayoutSize):
(WebKit::WebPageProxy::setMinimumLayoutSizeForMinimalUI):
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::didCommitLoad):
Defaulting the content size to the minimum layout size only works if the initial-scale is 1.
ViewportConfiguration knows exactly what to do before the first layout.
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/WebPage.messages.in:
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::setViewportConfigurationMinimumLayoutSize):
(WebKit::WebPage::setMinimumLayoutSizeForMinimalUI):
(WebKit::WebPage::dynamicViewportSizeUpdate):
A few fixes here:
-setZoomedOutPageScaleFactor() was incorrectly using the initial scale. 99% of the time, initial scale
and minimum scale are equal, but the page can specify something different with the viewport meta tag.
-Use floating point for manipulating the minimum layout sizes, then round the value.
-minimumLayoutSizeInDocumentCoordinate was scaled the wrong way around.
(WebKit::WebPage::viewportConfigurationChanged):
- 6:23 PM Changeset in webkit [167734] by
-
- 10 edits in trunk/Source/WebCore
[Mac, iOS] Stop buffering media when on an inactive tab.
https://bugs.webkit.org/show_bug.cgi?id=132077
Reviewed by Eric Carlson.
- html/HTMLMediaElement.cpp: Rename 'm_isDisplaySleepDisablingSuspended'
to 'm_elementIsHidden'.
(WebCore::HTMLMediaElement::HTMLMediaElement):
(WebCore::HTMLMediaElement::visibilityStatusChanged): Notify the
media session that the element is (or is not) hidden.
(WebCore::HTMLMediaElement::setShouldBufferData): Added.
- html/HTMLMediaElement.h:
- platform/audio/MediaSession.cpp:
(WebCore::MediaSession::clientWillBeginPlayback): Tell media to
buffer if not hidden or playing.
(WebCore::MediaSession::clientWillPausePlayback): Ditto.
(WebCore::MediaSession::visibilityChanged): Added. Client API, just relays call
to updateClientDataBuffering.
(WebCore::MediaSession::updateClientDataBuffering): Tell client it should only
buffer data if it's currently playing, or not hidden.
- platform/audio/MediaSession.h:
- platform/graphics/MediaPlayer.cpp:
(WebCore::MediaPlayer::setShouldBufferData): Added: Just relays to
MediaPlayerPrivate object.
- platform/graphics/MediaPlayer.h:
- platform/graphics/MediaPlayerPrivate.h:
(WebCore::MediaPlayerPrivateInterface::setShouldBufferData):
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.h:
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::MediaPlayerPrivateAVFoundationObjC):
(WebCore::MediaPlayerPrivateAVFoundationObjC::setShouldBufferData): Added.
Detach the player item from the player if we don't want to continue
buffering or other background tasks.
- 5:43 PM Changeset in webkit [167733] by
-
- 3 edits in trunk/Source/JavaScriptCore
The GC should only resume compiler threads that it suspended in the same GC pass.
<https://webkit.org/b/132088>
Reviewed by Mark Hahnenberg.
Previously, this scenario can occur:
- Thread 1 starts a GC and tries to suspend DFG worklist threads. However, no worklists were created yet at the that time.
- Thread 2 starts to compile some functions and creates a DFG worklist, and acquires the worklist thread's lock.
- Thread 1's GC completes and tries to resume suspended DFG worklist thread. This time, it sees the worklist created by Thread 2 and ends up unlocking the worklist thread's lock that is supposedly held by Thread 2.
Thereafter, chaos ensues.
The fix is to cache the worklists that were actually suspended by each GC pass,
and only resume those when the GC is done.
This issue was discovered by enabling COLLECT_ON_EVERY_ALLOCATION and running
the fast/workers layout tests.
- heap/Heap.cpp:
(JSC::Heap::visitCompilerWorklists):
(JSC::Heap::deleteAllCompiledCode):
(JSC::Heap::suspendCompilerThreads):
(JSC::Heap::resumeCompilerThreads):
- heap/Heap.h:
- 5:32 PM Changeset in webkit [167732] by
-
- 22 edits1 delete in trunk
Eliminate internals.setMockScrollbarsEnabled()
https://bugs.webkit.org/show_bug.cgi?id=132085
Reviewed by Tim Horton.
Source/WebCore:
This was essentially unused, and also didnt work.
- testing/InternalSettings.cpp:
(WebCore::InternalSettings::Backup::restoreTo):
(WebCore::InternalSettings::setMockScrollbarsEnabled): Deleted.
- testing/InternalSettings.h:
- testing/InternalSettings.idl:
Tools:
- DumpRenderTree/mac/DumpRenderTree.mm: (resetWebPreferencesToConsistentValues):
No need to reset the preference, there is no way it can change.
LayoutTests:
Remove calls, all of which were resetting it to true (which is the default, and
the only value that makes sense).
- compositing/geometry/fixed-position.html:
- compositing/geometry/horizontal-scroll-composited.html:
- compositing/geometry/vertical-scroll-composited.html:
- compositing/iframes/resources/scrollgrandchild-inner.html:
- compositing/overflow/fixed-position-ancestor-clip.html:
- compositing/resources/mock_scrollbars.js: Removed.
- css3/filters/blur-filter-page-scroll-parents.html:
- css3/filters/blur-filter-page-scroll-self.html:
- css3/filters/blur-filter-page-scroll.html:
- fast/dom/window-scroll-scaling.html:
- fast/events/touch/gesture/gesture-scrollbar.html:
- fast/sub-pixel/sub-pixel-iframe-copy-on-scroll.html:
- fast/sub-pixel/transformed-iframe-copy-on-scroll.html:
- media/video-controls-rendering.html:
- media/video-zoom.html:
- scrollingcoordinator/resources/non-fast-scrollable-region-testing.js:
- 5:08 PM Changeset in webkit [167731] by
-
- 2 edits in trunk/Tools
Unreviewed. Moving myself from Contributors to Committers list.
- 4:17 PM Changeset in webkit [167730] by
-
- 10 edits1 copy in trunk/Source
Don't migrate the WKView.h header from WebCore to WebKit
https://bugs.webkit.org/show_bug.cgi?id=132086
Reviewed by Dan Bernstein.
Source/WebCore:
- WebCore.xcodeproj/project.pbxproj:
Add WAKViewInternal.h.
- platform/WAKViewInternal.h: Added.
- platform/ios/wak/WAKClipView.m:
Import WAKViewInternal.h instead of WAKViewPrivate.h.
- platform/ios/wak/WAKScrollView.mm:
Import WAKViewInternal.h instead of WAKViewPrivate.h.
- platform/ios/wak/WAKView.h:
Move ivars to a class extension in WAKViewInternal.h and remove WKView.h import.
- platform/ios/wak/WAKView.mm:
Import WAKViewInternal.h instead of WAKViewPrivate.h.
- platform/ios/wak/WAKViewPrivate.h:
Import WKViewPrivate.h.
Source/WebKit/mac:
- MigrateHeaders.make:
Remove WKView.h from the list of headers to migrate.
- WebView/WebView.mm:
Import WKView.h.
- 3:35 PM Changeset in webkit [167729] by
-
- 2 edits1 add in trunk/Source/JavaScriptCore
Arguments::copyBackingStore needs to update m_registers in tandem with m_registerArray
https://bugs.webkit.org/show_bug.cgi?id=132079
Reviewed by Michael Saboff.
Since we're moving the register backing store, we don't want to leave a dangling pointer into a random CopiedBlock.
Also added a test that previously triggered this bug.
- runtime/Arguments.cpp:
(JSC::Arguments::copyBackingStore): D'oh!
- tests/stress/arguments-copy-register-array-backing-store.js: Added.
(foo):
(bar):
- 3:22 PM Changeset in webkit [167728] by
-
- 2 edits in trunk/Tools
Mock scrollbars don't always work in DumpRenderTree
https://bugs.webkit.org/show_bug.cgi?id=132082
Reviewed by Alexey Proskuryakov.
- DumpRenderTree/mac/DumpRenderTree.mm:
(prepareConsistentTestingEnvironment):
Set the mock scrollbar setting before creating the WebView, as this setting
cannot be dynamically changed, so any scrollbars that are created during
WebView initialization and are not replaced before snapshotting will end
up being platform scrollbars instead of mock ones.
WebKitTestRunner already sets this preference before creating WKViews,
so we're just matching its behavior.
- 3:19 PM Changeset in webkit [167727] by
-
- 3 edits in trunk/Source/WebCore
[New Multicolumn] fast/multicol/fixed-column-percent-logical-height-orthogonal-writing-mode.html fails
https://bugs.webkit.org/show_bug.cgi?id=132078
Reviewed by Anders Carlsson.
- rendering/RenderBox.cpp:
(WebCore::RenderBox::skipContainingBlockForPercentHeightCalculation):
Add a parameter to test for orthogonal writing modes. If we're perpendicular,
then we should not skip the flow thread, since we resolve relative to the column width,
and that is always set.
(WebCore::RenderBox::computePercentageLogicalHeight):
Patched to pass in whether or not the box and the ancestor block are perpendicular.
- rendering/RenderBox.h:
Patched the signature of skipContainingBlockForPercentHeightCalculation
- 2:55 PM Changeset in webkit [167726] by
-
- 5 edits in branches/safari-537.76-branch/Source
Versioning.
- 2:48 PM Changeset in webkit [167725] by
-
- 1 copy in branches/safari-537.76-branch
New branch.
- 2:33 PM Changeset in webkit [167724] by
-
- 4 edits in trunk/Source/WebCore
[OS X] Make checking if a font is the system font more robust
https://bugs.webkit.org/show_bug.cgi?id=132030
Reviewed by Dean Jackson.
Instead of inspecting a font's name to determine if it is a system font,
on OS X we can ask the system directly.
This patch also moves a platform-specific check into platform-specific
code, so that other platforms don't check for OS X-specific behavior.
Covered by existing tests.
- platform/graphics/Font.cpp:
(WebCore::Font::hasValidAverageCharWidth):
- platform/graphics/Font.h:
- platform/graphics/mac/FontMac.mm:
(WebCore::Font::primaryFontDataIsSystemFont):
- 2:27 PM Changeset in webkit [167723] by
-
- 3 edits in trunk/Source/WebCore
[New Multicolumn] Assertion failure in huge-column-count.html
https://bugs.webkit.org/show_bug.cgi?id=132071
Reviewed by Dean Jackson.
- rendering/RenderBlock.cpp:
(WebCore::RenderBlock::regionAtBlockOffset):
Back out this change, since it wasn't general enough.
- rendering/RenderFlowThread.cpp:
(WebCore::RenderFlowThread::getRegionRangeForBox):
The real issue was that this loop needed to consider the actual box
rather than starting from the parent. This was a non-issue for normal
regions (which cannot have nested flow threads), but for columns, you
have to consider the fact that the box could itself be a flow thread.
- 2:22 PM Changeset in webkit [167722] by
-
- 2 edits in trunk/Source/WebCore
[iOS] Memory pressure notification should fire on main thread.
<https://webkit.org/b/132074>
Rejig the memory pressure notification to fire on the main queue
directly instead of rerouting it manually.
Reviewed by Mark Rowe.
- platform/cocoa/MemoryPressureHandlerCocoa.mm:
(WebCore::MemoryPressureHandler::install):
- 1:54 PM Changeset in webkit [167721] by
-
- 3 edits in trunk/Source/WebCore
Unreviewed, rolling out r167720.
https://bugs.webkit.org/show_bug.cgi?id=132075
broke eight newmulticol tests (Requested by thorton on
#webkit).
Reverted changeset:
"[New Multicolumn] Assertion failure in huge-column-
count.html"
https://bugs.webkit.org/show_bug.cgi?id=132071
http://trac.webkit.org/changeset/167720
- 1:01 PM Changeset in webkit [167720] by
-
- 3 edits in trunk/Source/WebCore
[New Multicolumn] Assertion failure in huge-column-count.html
https://bugs.webkit.org/show_bug.cgi?id=132071
Reviewed by Dean Jackson.
- rendering/RenderBlock.cpp:
(WebCore::RenderBlock::regionAtBlockOffset):
Remove the code that returned 0 here, since we're going to patch a lower-level
function to catch all cases.
- rendering/RenderFlowThread.cpp:
(WebCore::RenderFlowThread::getRegionRangeForBox):
Don't allow in-flow RenderFlowThreads to ever have a region range. The sets
are what should have ranges... the flow thread needs to just be ignored.
- 12:38 PM Changeset in webkit [167719] by
-
- 2 edits in trunk/Source/WebKit2
Support encoding/decoding NSUInteger arguments in WKRemoteObjectCoder.
https://bugs.webkit.org/show_bug.cgi?id=132040
Add support for encoding/decoding NSUInteger arguments.
Patch by Yongjun Zhang <yongjun_zhang@apple.com> on 2014-04-23
Reviewed by Sam Weinig.
- Shared/API/Cocoa/WKRemoteObjectCoder.mm:
(encodeInvocation):
(decodeInvocationArguments):
- 11:55 AM Changeset in webkit [167718] by
-
- 3 edits2 adds in trunk
[New Multicolumn] Crasher when clearing out a flow thread in multicolumn layout.
https://bugs.webkit.org/show_bug.cgi?id=132069
Reviewed by Dean Jackson.
Source/WebCore:
This is imported from a patch Morten did for Blink, but I had to change it a fair
bit. deleteLines() is used to handle simple line box layout instead of just calling
deleteLineBoxTree.
I also had to disable the layout state to stop asserts on repaint when the children
get moved. Not sure why Blink didn't hit this, but it's simple enough to add a
LayoutStateDisabler to stop the assert.
Added fast/multicol/inline-children-crash.html
- rendering/RenderMultiColumnFlowThread.cpp:
(WebCore::RenderMultiColumnFlowThread::evacuateAndDestroy):
LayoutTests:
- fast/multicol/inline-children-crash-expected.txt: Added.
- fast/multicol/inline-children-crash.html: Added.
- 11:47 AM Changeset in webkit [167717] by
-
- 4 edits1 add in trunk/Source/WebCore
[iOS WebKit2] IOSurfacePool should force CA to actually garbage collect surfaces.
<https://webkit.org/b/132065>
<rdar://problem/16110687>
Add a platformGarbageCollectNow() helper function to IOSurfacePool that
triggers a sweep of the IOSurfaces. Call this from collectionTimerFired()
and discardAllSurfaces().
This lets us drop all otherwise-unused 420f surfaces on memory pressure.
Reviewed by Tim Horton.
- WebCore.xcodeproj/project.pbxproj:
- platform/graphics/cg/IOSurfacePool.cpp:
(WebCore::IOSurfacePool::collectionTimerFired):
(WebCore::IOSurfacePool::discardAllSurfaces):
- platform/graphics/cg/IOSurfacePool.h:
- platform/graphics/cocoa/IOSurfacePoolCocoa.mm: Added.
(WebCore::IOSurfacePool::platformGarbageCollectNow):
- 11:39 AM Changeset in webkit [167716] by
-
- 3 edits2 adds in trunk
REGRESSION (Safari 6 - ToT): Incorrectly assumes that RenderStyle data can be shared
https://bugs.webkit.org/show_bug.cgi?id=113058
Patch by Morten Stenshorne <mstensho@opera.com> on 2014-04-23
Reviewed by David Hyatt.
Source/WebCore:
Before sharing CSS properties with an element in the cache, we need to
check that the new element is suitable for this, just like we check
elements before inserting them into the cache.
Test: fast/css/identical-logical-height-decl.html
- css/StyleResolver.cpp:
(WebCore::StyleResolver::applyMatchedProperties):
LayoutTests:
- fast/css/identical-logical-height-decl-expected.html: Added.
- fast/css/identical-logical-height-decl.html: Added.
- 11:36 AM Changeset in webkit [167715] by
-
- 4 edits in trunk/Source/WebCore
Unreviewed, rolling out r167713.
https://bugs.webkit.org/show_bug.cgi?id=132070
broke hundreds of tests (Requested by thorton on #webkit).
Reverted changeset:
"[OS X] Make checking if a font is the system font more
robust"
https://bugs.webkit.org/show_bug.cgi?id=132030
http://trac.webkit.org/changeset/167713
- 11:05 AM Changeset in webkit [167714] by
-
- 6 edits in trunk/Source/WebCore
[New Multicolumn] Nested columns not working at all.
https://bugs.webkit.org/show_bug.cgi?id=131805
Reviewed by Dean Jackson.
Add support for nested pagination contexts, allowing for an arbitrary level
of nesting of multicolumn layouts. There were a number of things that had to
be patched in order for this to work.
- rendering/RenderBlock.cpp:
(WebCore::RenderBlock::regionAtBlockOffset):
Make sure RenderMultiColumnFlowThreads just return null for regions at any
block offset. Individual region sets will be created as you cross ancestor
regions eventually, so this is just getting in the way.
- rendering/RenderLayer.cpp:
(WebCore::RenderLayer::enclosingPaginationLayerInSubtree):
Add a new helper method for obtaining an enclosingPaginationLayer when
constrained by some root. This function ensures you don't accidentally
cross your subtree root when looking for enclosing pagination layers.
(WebCore::RenderLayer::collectFragments):
Patch collectFragments to know how to recur to collect ancestor fragments
in order to apply nested splitting as you cross pagination boundaries.
(WebCore::RenderLayer::updatePaintingInfoForFragments):
(WebCore::RenderLayer::calculateClipRects):
- rendering/RenderLayer.h:
(WebCore::LayerFragment::LayerFragment):
(WebCore::LayerFragment::setRects):
(WebCore::LayerFragment::moveBy):
(WebCore::LayerFragment::intersect):
Improve the LayerFragment so that it caches transformed bounding boxes as
well. This is needed to fix intersectsDamageRect so that it doesn't grab
the wrong bounding box when checking inline layers that are paginated.
- rendering/RenderMultiColumnFlowThread.cpp:
(WebCore::RenderMultiColumnFlowThread::flowThreadDescendantInserted):
Ignore inserted flow threads inside an ancestor flow thread, since we only
care about what the sets do.
- rendering/RenderObject.cpp:
(WebCore::RenderObject::insertedIntoTree):
Make sure that nested flow thread layers return themselves when a child
is inserted directly under them.
- 10:49 AM Changeset in webkit [167713] by
-
- 4 edits in trunk/Source/WebCore
[OS X] Make checking if a font is the system font more robust
https://bugs.webkit.org/show_bug.cgi?id=132030
Reviewed by Dean Jackson.
Instead of inspecting a font's name to determine if it is a system font,
on OS X we can ask the system directly.
This patch also moves a platform-specific check into platform-specific
code, so that other platforms don't check for OS X-specific behavior.
Covered by existing tests.
- platform/graphics/Font.cpp:
(WebCore::Font::hasValidAverageCharWidth):
- platform/graphics/Font.h:
- platform/graphics/mac/FontMac.mm:
(WebCore::Font::primaryFontDataIsSystemFont):
- 9:44 AM Changeset in webkit [167712] by
-
- 5 edits in trunk/Source/WebCore
[GTK] Focus management API was moved from HTMLDocument to Document in r166668
https://bugs.webkit.org/show_bug.cgi?id=132060
Patch by Philippe Normand <pnormand@igalia.com> on 2014-04-23
Reviewed by Carlos Garcia Campos.
Deprecate the removed methods in WebKitDOMHTMLDocument to keep API compatibility.
- bindings/gobject/WebKitDOMDeprecated.cpp:
(webkit_dom_html_document_get_active_element):
(webkit_dom_html_document_has_focus):
- bindings/gobject/WebKitDOMDeprecated.h:
- bindings/gobject/WebKitDOMDeprecated.symbols:
- bindings/gobject/webkitdom.symbols:
- 9:19 AM Changeset in webkit [167711] by
-
- 2 edits in trunk/Source/WebCore
[iOS WebKit1] MemoryPressureHandler::respondToMemoryPressure called on wrong thread.
<https://webkit.org/b/132041>
<rdar://problem/16687238>
Always dispatch the memory pressure relief code on the main queue.
This is really only needed for iOS/WK1, but we might as well share the code.
Reviewed by Michael Saboff.
- platform/cocoa/MemoryPressureHandlerCocoa.mm:
(WebCore::MemoryPressureHandler::install):
- 8:38 AM Changeset in webkit [167710] by
-
- 2 edits in trunk/LayoutTests
[GTK] Unreviewed GTK gardening
https://bugs.webkit.org/show_bug.cgi?id=132058
Patch by Lorenzo Tilve <ltilve@igalia.com> on 2014-04-23
- platform/gtk/TestExpectations:
- 7:39 AM Changeset in webkit [167709] by
-
- 2 edits in trunk/Source/JavaScriptCore
[Mac] REGRESSION (r164823): Building JavaScriptCore creates files under /tmp/JavaScriptCore.dst
<https://webkit.org/b/132053>
Reviewed by Dan Bernstein.
- JavaScriptCore.xcodeproj/project.pbxproj: Don't try to create a symlink at /usr/local/bin/jsc inside
the DSTROOT unless we're building to the deployment location. Also remove the unnecessary -x argument
from /bin/sh since that generates unnecessary output.
- 6:43 AM Changeset in webkit [167708] by
-
- 6 edits in trunk
[GTK] ShadowRoot API was removed in r164131
https://bugs.webkit.org/show_bug.cgi?id=132059
Source/WebCore:
Patch by Philippe Normand <pnormand@igalia.com> on 2014-04-23
Reviewed by Gustavo Noronha Silva.
Add ShadowRoot removed API to WebKitDOMDeprecated to keep API/ABI compatibility.
- bindings/gobject/WebKitDOMDeprecated.cpp:
(webkit_dom_shadow_root_class_init):
(webkit_dom_shadow_root_init):
(webkit_dom_shadow_root_element_from_point):
(webkit_dom_shadow_root_get_active_element):
(webkit_dom_shadow_root_get_apply_author_styles):
(webkit_dom_shadow_root_get_element_by_id):
(webkit_dom_shadow_root_get_elements_by_class_name):
(webkit_dom_shadow_root_get_elements_by_tag_name):
(webkit_dom_shadow_root_get_elements_by_tag_name_ns):
(webkit_dom_shadow_root_get_inner_html):
(webkit_dom_shadow_root_get_reset_style_inheritance):
(webkit_dom_shadow_root_get_selection):
(webkit_dom_shadow_root_set_apply_author_styles):
(webkit_dom_shadow_root_set_inner_html):
(webkit_dom_shadow_root_set_reset_style_inheritance):
- bindings/gobject/WebKitDOMDeprecated.h:
- bindings/gobject/WebKitDOMDeprecated.symbols:
Tools:
Patch by Carlos Garcia Campos <cgarcia@igalia.com> on 2014-04-23
Reviewed by Gustavo Noronha Silva.
Add WebKitDOMShadowRoot to the list of deleted classes to generate
the API docs.
- gtk/webkitdom.py:
(WebKitDOMDocGenerator):
- 6:36 AM Changeset in webkit [167707] by
-
- 7 edits4 adds in trunk
[New Multicolumn] fast/multicol/overflow-content.html displays red
https://bugs.webkit.org/show_bug.cgi?id=131809
Patch by Morten Stenshorne <mstensho@opera.com> on 2014-04-23
Reviewed by David Hyatt.
Source/WebCore:
Insert a break at end of content on our own in the multicol code, to make sure
that overflow is accounted for, and also to make sure that we account for all
content in non-final sets (i.e. those preceding spanners).
In other words, this will additionally fix balancing issues in sets preceding
a spanner. Added a test for that.
Tests: fast/multicol/break-in-columns-before-spanner.html
fast/multicol/newmulticol/compare-with-old-impl/overflow-content.html
- rendering/RenderFlowThread.cpp:
(WebCore::RenderFlowThread::regionInRange): Deleted.
- rendering/RenderFlowThread.h:
- rendering/RenderMultiColumnSet.cpp:
(WebCore::RenderMultiColumnSet::distributeImplicitBreaks):
- rendering/RenderNamedFlowThread.cpp:
(WebCore::RenderNamedFlowThread::applyBreakAfterContent):
- rendering/RenderNamedFlowThread.h:
LayoutTests:
- fast/multicol/break-in-columns-before-spanner-expected.html: Added.
- fast/multicol/break-in-columns-before-spanner.html: Added.
- fast/multicol/newmulticol/compare-with-old-impl/overflow-content-expected.html: Added.
- fast/multicol/newmulticol/compare-with-old-impl/overflow-content.html: Added.
- 6:32 AM Changeset in webkit [167706] by
-
- 6 edits4 adds3 deletes in trunk
Overflow propagation broken in BTT and RTL writing-modes
https://bugs.webkit.org/show_bug.cgi?id=113781
Patch by Morten Stenshorne <mstensho@opera.com> on 2014-04-23
Reviewed by David Hyatt.
Source/WebCore:
Overflow rectangles are not quite physical, not quite logical. This
means that we cannot use clientBoxRect() directly to represent a
rectangle that expresses exactly no overflow. This rectangle is the
padding box (relative to the border box) in vertical-lr and
horizontal-tb, but the block-direction borders need to be flipped in
vertical-rl and horizontal-bt.
Tests: fast/css/overflow-btt-border-after.html
fast/css/overflow-rtl-border-after.html
- WebCore.exp.in:
- rendering/RenderBlock.cpp:
(WebCore::RenderBlock::computeOverflow):
- rendering/RenderBox.cpp:
(WebCore::RenderBox::addLayoutOverflow):
(WebCore::RenderBox::addVisualOverflow):
(WebCore::RenderBox::layoutOverflowRectForPropagation):
- rendering/RenderBox.h:
(WebCore::RenderBox::layoutOverflowRect):
LayoutTests:
- fast/css/overflow-btt-border-after-expected.txt: Added.
- fast/css/overflow-btt-border-after.html: Added.
- fast/css/overflow-rtl-border-after-expected.txt: Added.
- fast/css/overflow-rtl-border-after.html: Added.
- fast/multicol/vertical-rl/rules-with-border-before-expected.png: Removed.
- platform/gtk/fast/multicol/vertical-rl/rules-with-border-before-expected.png: Removed.
- platform/mac/fast/multicol/vertical-rl/rules-with-border-before-expected.png: Removed.
Removed incorrect screenshots that now differ from the actual (correct) rendering.
- 5:57 AM WebKitGTK/KeepingTheTreeGreen edited by
- (diff)
- 5:30 AM Changeset in webkit [167705] by
-
- 3 edits in trunk/Source/WebCore
[CSS Regions] Improve the debugging infrastructure
https://bugs.webkit.org/show_bug.cgi?id=132042
Reviewed by Mihnea Ovidenie.
This patch improves the debugging code for CSS Regions.
Tests: No function change. No new tests.
- rendering/RenderObject.cpp:
(WebCore::RenderObject::showRegionsInformation): Print brackets around the region range when dumping
the render tree to the console.
- rendering/RootInlineBox.cpp:
(WebCore::RootInlineBox::containingRegion): Convert the ASSERT to an ASSERT_WITH_SECURITY_IMPLICATION.
- 4:19 AM Changeset in webkit [167704] by
-
- 3 edits2 copies1 move4 deletes in trunk/LayoutTests
Unreviewed EFL gardening.
- platform/efl-wk1/css3/flexbox/flexbox-baseline-expected.png: Removed.
- platform/efl-wk1/css3/flexbox/flexbox-baseline-expected.txt: Removed.
- platform/efl-wk1/editing/input/reveal-caret-of-multiline-contenteditable-expected.txt: Removed.
- platform/efl-wk1/fast/block/basic/020-expected.txt: Removed.
- platform/efl-wk2/TestExpectations:
- platform/efl/TestExpectations:
- platform/efl/css3/flexbox/flexbox-baseline-expected.png: Renamed from LayoutTests/platform/efl-wk2/css3/flexbox/flexbox-baseline-expected.png.
- platform/efl/css3/flexbox/flexbox-baseline-expected.txt: Renamed from LayoutTests/platform/efl-wk2/css3/flexbox/flexbox-baseline-expected.txt.
- platform/efl/editing/input/reveal-caret-of-multiline-contenteditable-expected.txt: Renamed from LayoutTests/platform/efl-wk2/editing/input/reveal-caret-of-multiline-contenteditable-expected.txt.
- 4:11 AM Changeset in webkit [167703] by
-
- 2 edits1 add in trunk/LayoutTests
[EFL] Add expectations of new added tests.
https://bugs.webkit.org/show_bug.cgi?id=132052
Unreviewed EFL gardening.
Patch by Krzysztof Wolanski <k.wolanski@samsung.com> on 2014-04-23
- platform/efl-wk2/TestExpectations:
- platform/efl/fast/repaint/spanner-with-margin-expected.txt: Added. Rebaseline after r167439
- 2:39 AM Changeset in webkit [167702] by
-
- 2 edits in trunk/Tools
[GTK] run-webkit-tests script is failing to invoke run-launcher
https://bugs.webkit.org/show_bug.cgi?id=132051
Patch by Lorenzo Tilve <ltilve@igalia.com> on 2014-04-23
Reviewed by Philippe Normand.
After the WK1 support was dropped, if run-webkit-tests is executed
without --no-show-results parameter, the launcher is being called
using a -2 parameter that is it causing the run-launcher to break.
- Scripts/webkitpy/port/gtk.py:
(GtkPort.show_results_html_file):
- 12:34 AM Changeset in webkit [167701] by
-
- 4 edits2 adds7 deletes in trunk/LayoutTests
Refactoring editing/selection/13804.html
https://bugs.webkit.org/show_bug.cgi?id=131986
Reviewed by Darin Adler.
This test does not need pixel baselines (MAC, EFL, GTK) to just
check whether spelling marker appears on 'adf'.
According to bug 125688, the test starts using asynchronous
spellchecking.
Since the test mostly verifies spelling behaviour, it was moved
to editing/spelling directory.
Its name was changed to spelling-backward.html.
- editing/selection/13804.html: Removed.
- editing/spelling/spelling-backward-expected.txt: Added, renamed from editing/selection/13804.html.
- editing/spelling/spelling-backward.html: Added.
- platform/efl/editing/selection/13804-expected.png: Removed.
- platform/efl/editing/selection/13804-expected.txt: Removed.
- platform/gtk/editing/selection/13804-expected.png: Removed.
- platform/gtk/editing/selection/13804-expected.txt: Removed.
- platform/mac/editing/selection/13804-expected.png: Removed.
- platform/mac/editing/selection/13804-expected.txt: Removed.
Get rid of platform specific expectation.
- platform/gtk/TestExpectations:
- platform/mac-wk2/TestExpectations:
- platform/win/TestExpectations:
Mark spelling-backward.html as failure due to missing asynchronous spellchecking.
Apr 22, 2014:
- 11:11 PM Changeset in webkit [167700] by
-
- 6 edits in trunk/Source
Cursor doesn't change back to pointer when leaving the Safari window
https://bugs.webkit.org/show_bug.cgi?id=132038
Reviewed by Alexey Proskuryakov.
Source/WebCore:
r147739 incorrectly added an early exit in EventHandler::selectCursor when hit test result didn't have
any node associated with it. Since we will hit this code when the cursor is outside of the WebView,
we still need to take the CURSOR_AUTO path as did the code before r147739.
No new test is added since this behavior can't be tested in DRT or WTR.
- page/EventHandler.cpp:
(WebCore::EventHandler::selectCursor):
Source/WebKit/mac:
Since the cursor type is now updated asynchronously after r147739,
[window windowNumber] != [NSWindow windowNumberAtPoint:[NSEvent mouseLocation] belowWindowWithWindowNumber:0]
evalutes to false depending on how fast cursor is moving.
Instead, check whether the NSWindow of the WebView is the key window or not since
key window appears to control the cursor style in Cocoa as far as I've tested:
https://developer.apple.com/library/mac/documentation/Cocoa/Conceptual/WinPanel/Concepts/ChangingMainKeyWindow.html
- WebCoreSupport/WebChromeClient.mm:
(WebChromeClient::setCursor):
Source/WebKit2:
Since the cursor type is now updated asynchronously after r147739,
[window windowNumber] != [NSWindow windowNumberAtPoint:[NSEvent mouseLocation] belowWindowWithWindowNumber:0]
evalutes to false depending on how fast cursor is moving.
Instead, check whether the NSWindow of the WebView is the key window or not since
key window appears to control the cursor style in Cocoa as far as I've tested:
https://developer.apple.com/library/mac/documentation/Cocoa/Conceptual/WinPanel/Concepts/ChangingMainKeyWindow.html
- UIProcess/mac/PageClientImpl.mm:
(WebKit::PageClientImpl::setCursor):
- 10:59 PM Changeset in webkit [167699] by
-
- 3 edits in trunk/Source/WebKit2
Add SPI to expose provisional URL from Frame.
https://bugs.webkit.org/show_bug.cgi?id=132014
Add methods to return the provisional URL from WKWebProcessPlugInFrame.
Patch by Yongjun Zhang <yongjun_zhang@apple.com> on 2014-04-22
Reviewed by Dan Bernstein.
- WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInFrame.mm:
(-[WKWebProcessPlugInFrame _provisionalURL]):
- WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInFramePrivate.h:
- 10:53 PM Changeset in webkit [167698] by
-
- 14 edits in trunk/Source/WebKit2
[iOS][WK2] Split iOS touch event dispatch for the regular touch event dispatch
https://bugs.webkit.org/show_bug.cgi?id=132033
Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-04-22
Reviewed by Simon Fraser.
The generic touch event dispatching does not work well with iOS gesture recognizers.
Some events are required to be synchronous, while other needs can be asynchronous and
can even use event coalescing. We only need UI callback for the synchronous events.
Previously, that was implemented by changing the property ShouldSendEventsSynchronously
on WebPageProxy. That is causing issues when synchronous events are sent right after asynchronous
events because the responses in WebPageProxy::didReceiveEvent() easily get out of order.
This patch completely split the UIGestureRecognizerDispatching from the normal asynchronous event dispatching.
Synchronous events are sent directly to the WebProcess and report the result to the PageClient right away.
Asynchronous events are sent to the EventDispatcher where they are coalesced until the main thread is free
to process events.
There can be a race where a synchronous event comes before the asynchronous events are processed. In that case,
the synchronous event handler starts by getting the events out of the EventDispatcher and dispatch them first.
- Shared/NativeWebTouchEvent.h:
(WebKit::NativeWebTouchEvent::uniqueId): Deleted.
- Shared/ios/NativeWebTouchEventIOS.mm:
(WebKit::NativeWebTouchEvent::NativeWebTouchEvent):
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::handleSynchronousTouchEvent):
(WebKit::WebPageProxy::sendAsynchronousTouchEvent):
(WebKit::WebPageProxy::didReceiveEvent):
(WebKit::WebPageProxy::resetStateAfterProcessExited):
- UIProcess/WebPageProxy.h:
- UIProcess/ios/WKContentViewInteraction.h:
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _webTouchEventsRecognized:]):
(-[WKContentView _webTouchEvent:WebKit::preventsNativeGestures:]):
- WebProcess/WebPage/EventDispatcher.cpp:
(WebKit::EventDispatcher::EventDispatcher):
(WebKit::EventDispatcher::clearQueuedTouchEventsForPage):
(WebKit::EventDispatcher::getQueuedTouchEventsForPage):
(WebKit::EventDispatcher::touchEvent):
(WebKit::EventDispatcher::dispatchTouchEvents):
- WebProcess/WebPage/EventDispatcher.h:
- WebProcess/WebPage/EventDispatcher.messages.in:
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::dispatchTouchEvent):
(WebKit::WebPage::touchEventSync):
(WebKit::WebPage::touchEvent):
(WebKit::WebPage::touchEventSyncForTesting):
(WebKit::WebPage::didCommitLoad):
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/WebPage.messages.in:
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::dispatchAsynchronousTouchEventsQueue):
- 10:28 PM Changeset in webkit [167697] by
-
- 2 edits in trunk/Source/WebKit/efl
[EFL][WK1] There are many failures in fast/forms since r163033
https://bugs.webkit.org/show_bug.cgi?id=132039
Reviewed by Gyuyoung Kim.
- ewk/ewk_view.cpp: Added break statements which is missing while remove COMPILE_ASSERT_MATCHING_ENUM
(ewk_view_navigation_policy_decision):
- 10:03 PM Changeset in webkit [167696] by
-
- 3 edits in trunk/LayoutTests
Unreviewed EFL gardening. media tests has been broken since r166395 on both EFL wk1 and wk2.
- platform/efl-wk2/TestExpectations:
- platform/efl/TestExpectations:
- 8:13 PM Changeset in webkit [167695] by
-
- 2 edits in trunk/LayoutTests
Skip fast/events/ghostly-mousemoves-in-subframe.html on Mac/WK1.
- platform/mac-wk1/TestExpectations:
- 7:44 PM Changeset in webkit [167694] by
-
- 5 edits in trunk
Do not paint border image when the border rect is empty.
https://bugs.webkit.org/show_bug.cgi?id=131988
Reviewed by Darin Adler.
http://trac.webkit.org/changeset/167351 introduced an early return when border
rect is empty. This patch ensures that border image is not painted either in that case.
Source/WebCore:
Modified padding-margin-negative-border.html to cover border-image case.
- rendering/RenderBoxModelObject.cpp:
(WebCore::RenderBoxModelObject::paintBorder):
LayoutTests:
- fast/css/padding-margin-negative-border-expected.html: Borders in -expected.html does not get painted either, just
ensure that they are explicitly different colors.
- fast/css/padding-margin-negative-border.html:
- 6:37 PM Changeset in webkit [167693] by
-
- 2 edits in trunk/Source/WebCore
ASSERTION FAILED: scrollerImp == scrollbarPainterForScrollbar(_scrollbar) on two API tests
https://bugs.webkit.org/show_bug.cgi?id=132034
<rdar://problem/16624332>
Reviewed by Simon Fraser.
- platform/mac/ScrollAnimatorMac.mm:
(-[WebScrollbarPainterDelegate shouldUseLayerPerPartForScrollerImp:]):
scrollerImpWithStyle:controlSize:horizontal:replacingScrollerImp: moves the delegate
from the old scrollerImp to the new one, and also happens to call shouldUseLayerPerPartForScrollerImp.
Since scrollerImpWithStyle: has not returned yet, the scrollbarPainterForScrollbar still returns the old scrollerImp,
so this assertion fires. It's safe to remove this because supportsUpdateOnSecondaryThread doesn't make use of the imp.
- 5:36 PM Changeset in webkit [167692] by
-
- 2 edits in trunk/Source/JavaScriptCore
DFG::Worklist should acquire the m_lock before iterating DFG plans.
<https://webkit.org/b/132032>
Reviewed by Filip Pizlo.
Currently, there's a rightToRun mechanism that ensures that no compilation
threads are running when the GC is iterating through the DFG worklists.
However, this does not prevent a Worker thread from doing a DFG compilation
and modifying the plans in the worklists thereby invalidating the plan
iterator that the GC is using. This patch fixes the issue by acquiring
the worklist m_lock before iterating the worklist plans.
This issue was uncovered by running the fast/workers layout tests with
COLLECT_ON_EVERY_ALLOCATION enabled.
- dfg/DFGWorklist.cpp:
(JSC::DFG::Worklist::isActiveForVM):
(JSC::DFG::Worklist::visitChildren):
- 4:54 PM Changeset in webkit [167691] by
-
- 2 edits in trunk/Tools
DumpRenderTree generates black snapshots on some platforms
https://bugs.webkit.org/show_bug.cgi?id=132029
<rdar://problem/16679363>
Reviewed by Simon Fraser.
- DumpRenderTree/mac/PixelDumpSupportMac.mm:
(createBitmapContextFromWebView):
DRT checks with the WebView to determine whether it can use the window's
backing store directly or (because of compositing) needs to take a WindowServer
snapshot. However, on some platforms and in some situations, the window can
be layer backed without the WebView necessarily currently being composited.
We should check with the window in addition to the WebView before going
ahead and using its backing store.
- 4:51 PM Changeset in webkit [167690] by
-
- 2 edits in trunk/Source/WebKit2
AX: iOS hit-testing does not work when page is zoomed
https://bugs.webkit.org/show_bug.cgi?id=132013
Reviewed by Simon Fraser.
Implement accessibilityHitTest so that the point can be converted correctly.
- WebProcess/WebPage/WKAccessibilityWebPageObjectIOS.mm:
(-[WKAccessibilityWebPageObject accessibilityHitTest:]):
- 4:25 PM Changeset in webkit [167689] by
-
- 3 edits3 adds in trunk
REGRESSION (r157328): popover to check into flight ba.com dismisses instantly when focusing form
https://bugs.webkit.org/show_bug.cgi?id=131949
Reviewed by Darin Adler.
Source/WebCore:
The regression was caused by two bugs:
- The event didn't stop propagating itself even when it should. If the related target is same as the event origin, the event propagation should stop when the event reaches the root of the related target's tree scope. Otherwise, it should stop when it reaches the related target.
- Mouse event's related target exposed nodes inside a user-agent shadow DOM when the related target appeared inside the origin.
Fixed the bugs by re-introducing path shrinkage algorithm removed in r157328 into EventPath::setRelatedTarget
and adding an algorithm to determine the least common ancestor of the related target and the current target
in moveToParentOrShadowHost. The latter algorithm doesn't match the shadow DOM specification:
http://www.w3.org/TR/2013/WD-shadow-dom-20130514/
but it's good enough in terms of the Web exposed behavior as we don't support author defined insertion points.
Test: fast/events/shadow-event-path.html
- dom/EventDispatcher.cpp:
(WebCore::EventRelatedNodeResolver::moveToParentOrShadowHost):
(WebCore::EventRelatedNodeResolver::findHostOfTreeScopeInTargetTreeScope): Added.
(WebCore::EventDispatcher::dispatchEvent):
(WebCore::EventPath::setRelatedTarget):
LayoutTests:
Add a test that dumps the event target and the related target of every mouse event
when dispatched inside an input element inside a details element.
This catches the regression as well as other bugs I encountered while fixing the bug.
We need a WK2 specific results because WK1 mac results contain an extra fake mouse move event.
- fast/events/shadow-event-path-expected.txt: Added.
- fast/events/shadow-event-path.html: Added.
- platform/mac-wk2/fast/events/shadow-event-expected.txt: Added.
- 4:05 PM Changeset in webkit [167688] by
-
- 3 edits2 deletes in trunk
Rollout r156635 since the old behavior was intentional.
Source/WebCore:
- page/EventHandler.cpp:
(WebCore::EventHandler::selectCursor):
LayoutTests:
- editing/caret/selection-with-caret-type-progress-expected.txt: Removed.
- editing/caret/selection-with-caret-type-progress.html: Removed.
- 3:44 PM Changeset in webkit [167687] by
-
- 3 edits in trunk/Tools
DumpRenderTree generates white snapshots on some platforms
https://bugs.webkit.org/show_bug.cgi?id=132026
<rdar://problem/16679363>
Reviewed by Simon Fraser.
- DumpRenderTree/mac/Configurations/DumpRenderTree.xcconfig:
Link DRT against QuartzCore.
- DumpRenderTree/mac/PixelDumpSupportMac.mm:
(createBitmapContextFromWebView):
Force CoreAnimation to synchronously update layers in a more explicit
way that doesn't depend on AppKit behavior. This is very similar to what
WebKitTestRunner does (indirectly) during its force repaint.
- 3:05 PM Changeset in webkit [167686] by
-
- 17 edits in trunk/Source
Unreviewed, rolling out r167674.
https://bugs.webkit.org/show_bug.cgi?id=132025
Going a different way with this (Requested by bradee-oh on
#webkit).
Reverted changeset:
"Change Image Controls replacement to use selection and paste"
https://bugs.webkit.org/show_bug.cgi?id=131992
http://trac.webkit.org/changeset/167674
- 2:58 PM Changeset in webkit [167685] by
-
- 12 edits in trunk
[Win] Support Python 2.7 in Cygwin
https://bugs.webkit.org/show_bug.cgi?id=132023
Reviewed by Michael Saboff.
Source/JavaScriptCore:
- DerivedSources.make: Use a conditional variable to define
the path to Python/Perl.
Source/WebCore:
- DerivedSources.make: Use proper path to Cygwin on
all platforms.
Source/WebInspectorUI:
- WebInspectorUI.vcxproj/build-webinspectorui.sh: use
proper path to Python under Cygwin.
Source/WebKit:
- WebKit.vcxproj/WebKit/WebKitVersion.cmd: Use correct
path to Python under Cygwin.
Source/WTF:
- WTF.vcxproj/WTFGenerated.make: Use proper path to Python.
Tools:
- Scripts/webkitdirs.pm:
(checkInstalledTools): Remove check for Python 2.7, since
it now works properly.
- 2:55 PM Changeset in webkit [167684] by
-
- 3 edits4 adds in trunk
REGRESSION (r151839): Subframe keeps getting mousemove events with the same coordinates after hiding a hovered element.
<https://webkit.org/b/131974>
<rdar://problem/15907469>
Source/WebCore:
When the currently hovered element disappears as a result of style recalc,
we send a fake mousemove event to the page, to see if anything newly added
should become hovered.
The faking mechanism lives in EventHandler and simply synthesizes a new
mousemove event using the last seen mouse location.
The problem here is that we were sending this fake mousemove event to the
subframe where the hovered element lived. Since subframes aren't kept up
to date on recent mouse locations, this could cause some strange behavior
where a subframe would dispatch mousemove events with stale coordinates.
The solution is to always dispatch fake mousemove events from the main
frame's event handler. This is how real event delivery happens, and hit
testing will then find the appropriate subframe, if any.
Reviewed by Benjamin Poulain.
Test: fast/events/ghostly-mousemoves-in-subframe.html
- dom/Document.cpp:
(WebCore::Document::recalcStyle):
LayoutTests:
Add a test that triggers the weirdness where removing the renderer from
a hovered element in a subframe would leave the subframe's EventHandler
in a state where it could dispatch fake mousemove events with stale
coordinates in response to style recalc.
Note that the final 500ms delay is because fake mousemove events are
sent on 250ms delay timers so we need to give it some time to catch up.
Reviewed by Benjamin Poulain.
- fast/events/ghostly-mousemoves-in-subframe-expected.txt: Added.
- fast/events/ghostly-mousemoves-in-subframe.html: Added.
- platform/mac-wk2/fast/events/resources/ghostly-mousemoves-in-subframe-the-actual-subframe.html: Added.
- 2:07 PM Changeset in webkit [167683] by
-
- 2 edits in trunk/Source/WebCore
[OS X] Glyph spacing for system fonts may be incorrect
https://bugs.webkit.org/show_bug.cgi?id=131967
Unreviewed iOS build fix after r167679.
- platform/graphics/mac/SimpleFontDataMac.mm:
- 1:51 PM Changeset in webkit [167682] by
-
- 2 edits in trunk/Source/WebKit2
Remove unused local variable
https://bugs.webkit.org/show_bug.cgi?id=132020
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2014-04-22
Reviewed by Alexey Proskuryakov.
- UIProcess/ios/forms/WKFormSelectPopover.mm:
(-[WKSelectPopover initWithView:hasGroups:]):
- 1:49 PM Changeset in webkit [167681] by
-
- 2 edits in trunk/Source/WebCore
[OS X] Glyph spacing for system fonts may be incorrect
https://bugs.webkit.org/show_bug.cgi?id=131967
Unreviewed build fix after r167679.
- platform/graphics/mac/SimpleFontDataMac.mm:
(WebCore::hasCustomTracking):
- 1:40 PM Changeset in webkit [167680] by
-
- 2 edits in trunk/LayoutTests
REGRESSION: JSRegress's js/slow-stress/new-spread.html fails sometimes
https://bugs.webkit.org/show_bug.cgi?id=132022
<rdar://problem/16690671>
Mark the test as flaky pass/fail/timeout, because that's what it is.
- 1:38 PM Changeset in webkit [167679] by
-
- 5 edits in trunk
[OS X] Glyph spacing for system fonts may be incorrect
https://bugs.webkit.org/show_bug.cgi?id=131967
Reviewed by Simon Fraser.
Source/WebCore:
Covered by existing tests.
- platform/graphics/mac/SimpleFontDataMac.mm:
(WebCore::SimpleFontData::platformWidthForGlyph): Update to use CTFontGetAdvancesForGlyphs() for system fonts
(WebCore::hasCustomTracking):
LayoutTests:
Updating test to be more robust.
- fast/forms/search/intrinsic-search-width-with-decoration-border-padding-expected.txt:
- fast/forms/search/intrinsic-search-width-with-decoration-border-padding.html:
- 1:31 PM Changeset in webkit [167678] by
-
- 2 edits in trunk/Source/WebCore
[New Multicolumn] widows/orphans cause assertion failures.
https://bugs.webkit.org/show_bug.cgi?id=131233
Reviewed by Dean Jackson.
- rendering/RenderMultiColumnSet.cpp:
(WebCore::RenderMultiColumnSet::recordSpaceShortage):
Fix an obvious bug where the space shortage is recorded twice (and the
if statement that kept it from being negative is accidentally ignored).
- 1:05 PM Changeset in webkit [167677] by
-
- 2 edits in trunk/Source/WebCore
[New Multicolumn] Make sure columnTranslationForOffset has the same column-span-aware
translation that fragment collection does.
https://bugs.webkit.org/show_bug.cgi?id=131738
Reviewed by Anders Carlsson.
- rendering/RenderMultiColumnSet.cpp:
(WebCore::RenderMultiColumnSet::collectLayerFragments):
Remove a FIXME that is no longer needed, since the translation offset of the
column set did get patched.
(WebCore::RenderMultiColumnSet::columnTranslationForOffset):
Add the exact same fix to columnTranslationForOffset that was applied to collectLayerFragments.
- 12:58 PM Changeset in webkit [167676] by
-
- 2 edits in trunk/Source/WebCore
WebCore::HTMLMediaElement::ensureMediaControlsInjectedScript() needs to acquire the JSLock before calling into JS.
<https://webkit.org/b/132021>
Reviewed by Mark Hahnenberg.
Covered by existing layout test.
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::parseAttribute):
- 12:48 PM Changeset in webkit [167675] by
-
- 4 edits in trunk
REGRESSION (r167652): Broke fast/regions/cssom/region-range-for-box-crash.html in debug mode
https://bugs.webkit.org/show_bug.cgi?id=131982
Reviewed by David Hyatt.
Source/WebCore:
The problem was creating the Range from the arguments received at RenderView::setSelection(). Specifically
in this test endPos is 1 when the element has not children which creates an invalid Range.
- rendering/RenderView.cpp:
(WebCore::RenderView::splitSelectionBetweenSubtrees): Pass 0 as startPos and endPos to Range::create() as we are
not interested in the positions at this point.
LayoutTests:
- TestExpectations: Unskip fast/regions/cssom/region-range-for-box-crash.html.
- 12:22 PM Changeset in webkit [167674] by
-
- 17 edits in trunk/Source
Change Image Controls replacement to use selection and paste
<rdar://problem/16302722> and https://bugs.webkit.org/show_bug.cgi?id=131992
Reviewed by Tim Horton.
Source/WebCore:
- WebCore.exp.in: Remove deleted symbol.
- html/shadow/mac/ImageControlsButtonElementMac.cpp:
(WebCore::ImageControlsButtonElementMac::defaultEventHandler):
- page/ContextMenuController.cpp:
(WebCore::ContextMenuController::showImageControlsMenu): Select the image element
before showing the menu.
(WebCore::ContextMenuController::replaceControlledImage): Deleted.
- page/ContextMenuController.h:
Source/WebKit/mac:
- Misc/WebSharingServicePickerController.mm:
(-[WebSharingServicePickerController sharingService:didShareItems:]): Write the
replacement data to a special pasteboard, then replace the current selection from
that pasteboard.
Source/WebKit2:
- UIProcess/mac/WebContextMenuProxyMac.mm:
(-[WKSharingServicePickerDelegate sharingService:didShareItems:]): Write the
replacement data to a special pasteboard, then replace the current selection from
that pasteboard.
Remove all other code used to support the old replacement mechanism:
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::replaceControlledImage): Deleted.
- UIProcess/WebPageProxy.h:
- UIProcess/mac/WebContextMenuProxyMac.h:
(WebKit::WebContextMenuProxyMac::page):
(WebKit::WebContextMenuProxyMac::replaceControlledImage): Deleted.
- WebProcess/WebPage/WebContextMenu.cpp:
(WebKit::WebContextMenu::replaceControlledImage): Deleted.
- WebProcess/WebPage/WebContextMenu.h:
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::replaceControlledImage): Deleted.
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/WebPage.messages.in:
- 12:00 PM Changeset in webkit [167673] by
-
- 2 edits in trunk/Source/WebCore
REGRESSION: Hitting asserts in new flow thread selection code.
https://bugs.webkit.org/show_bug.cgi?id=132019
Reviewed by Anders Carlsson.
Null checking flowThreadController is not enough. The list of threads
can be null even when the flowThreadController is not. That's why you
use hasRenderNamedFlowThreads. I patched the two bad places in the code
to use hasRenderNamedFlowThreads to stop the crashes.
- rendering/RenderView.cpp:
(WebCore::RenderView::selectionBounds):
(WebCore::RenderView::repaintSelection):
- 11:36 AM Changeset in webkit [167672] by
-
- 3 edits2 adds in trunk
Check (rather than assume) element is a RenderTableSection before using it
https://bugs.webkit.org/show_bug.cgi?id=121858
Reviewed by David Kilzer.
Source/WebCore:
Test: fast/table/table-insert-object-before-td-crash.html
- rendering/RenderTable.cpp:
(WebCore::RenderTable::addChild): Check that lastBox is a table section before
using the node as if it was one.
LayoutTests:
- fast/table/table-insert-object-before-td-crash-expected.txt: Added.
- fast/table/table-insert-object-before-td-crash.html: Added.
- 11:36 AM Changeset in webkit [167671] by
-
- 6 edits in trunk
Switch the LLVMForJSC target to using the LLVM in /usr/local rather than /usr/local/LLVMForJavaScriptCore on iOS
https://bugs.webkit.org/show_bug.cgi?id=130867
<rdar://problem/16432456>
Reviewed by Mark Hahnenberg.
Source/JavaScriptCore:
- Configurations/Base.xcconfig:
- Configurations/LLVMForJSC.xcconfig:
Tools:
- Scripts/build-webkit:
- Scripts/copy-webkitlibraries-to-product-directory:
- 11:22 AM Changeset in webkit [167670] by
-
- 7 edits2 adds in trunk/Source
Begin implementation of video using Media Foundation.
https://bugs.webkit.org/show_bug.cgi?id=131830
Reviewed by Brent Fulgham.
Source/WebCore:
- WebCore.vcxproj/WebCore.vcxproj:
- WebCore.vcxproj/WebCore.vcxproj.filters:
Added MediaPlayerPrivateMediaFoundation files.
- platform/graphics/MediaPlayer.cpp:
Include MediaPlayerPrivateMediaFoundation inside new USE(MEDIA_FOUNDATION) flag.
- platform/graphics/win/MediaPlayerPrivateMediaFoundation.cpp: Added.
- platform/graphics/win/MediaPlayerPrivateMediaFoundation.h: Added.
Source/WebKit/win:
- FullscreenVideoController.cpp:
- WebView.cpp:
Added new USE(MEDIA_FOUNDATION) flag to prevent using the unsupported fullscreen api.
- 11:12 AM Changeset in webkit [167669] by
-
- 2 edits in trunk/Tools
[GTK] Fails to build documentation when output is redirected
https://bugs.webkit.org/show_bug.cgi?id=131178
Reviewed by Martin Robinson.
- gtk/generate-gtkdoc: set the output for stdout and stderr to UTF-8
explicitly to avoid encoding issues when they are redirected.
- 11:08 AM Changeset in webkit [167668] by
-
- 2 edits in trunk/Source/JavaScriptCore
[Win] Unreviewed build fix after my r167666.
- JavaScriptCore.vcxproj/LLInt/LLIntOffsetsExtractor/LLIntOffsetsExtractorCommon.props:
Added ../../../ again to include headers in Source/JavaScriptCore.
- 10:38 AM Changeset in webkit [167667] by
-
- 2 edits in trunk/Tools
[GTK] YCM choose a newer compile_commands.json in between Release or Debug
https://bugs.webkit.org/show_bug.cgi?id=131911
Reviewed by Martin Robinson.
common.get_build_path returns release path even though Debug configuration
is newer than Release. So YouCompleteMe is used to refer old build setup inadvertently.
- gtk/ycm_extra_conf.py:
(get_build_path): Compare modified time of Release and Debug. And return a recent modified path.
(FlagsForFile):
- 10:30 AM Changeset in webkit [167666] by
-
- 8 edits1 delete in trunk
Removed old stdbool and inttypes headers.
https://bugs.webkit.org/show_bug.cgi?id=131966
Reviewed by Brent Fulgham.
.:
- Source/cmake/OptionsWindows.cmake:
- Source/cmake/WebKitPackaging.cmake:
Removed references to os-win32 directory.
Source/JavaScriptCore:
- JavaScriptCore.vcxproj/LLInt/LLIntOffsetsExtractor/LLIntOffsetsExtractorCommon.props:
- JavaScriptCore.vcxproj/testRegExp/testRegExpCommon.props:
Removed references to os-win32 directory.
- os-win32: Removed.
- os-win32/inttypes.h: Removed.
- os-win32/stdbool.h: Removed.
WebKitLibraries:
- win/tools/vsprops/WinCairo.props:
Removed references to os-win32 directory.
- 10:27 AM April 2014 Meeting edited by
- (diff)
- 10:27 AM 2014 Meeting Subpixel Layout created by
- 10:24 AM 2014 Meeting Pointers and Iterators created by
- 10:18 AM April 2014 Meeting edited by
- (diff)
- 10:04 AM Changeset in webkit [167665] by
-
- 2 edits in trunk/Source/WebKit2
[EFL][WK2] Remove duplicated code to set size to the DrawingArea.
https://bugs.webkit.org/show_bug.cgi?id=131972
Reviewed by Andreas Kling.
DrawingArea's size is set in the WebView.cpp when EwkView's size is
changed, so we don't have to set again in the PageViewportController.
- UIProcess/CoordinatedGraphics/PageViewportController.cpp:
(WebKit::PageViewportController::didChangeViewportSize):
- 9:56 AM Changeset in webkit [167664] by
-
- 3 edits in trunk/Source/WebCore
Unreviewed, rolling out r167658.
https://bugs.webkit.org/show_bug.cgi?id=131993
Caused many crashes on media tests (Requested by ap on
#webkit).
Reverted changeset:
"[Mac] don't ask for AVAssetTrack properties before they are
available"
https://bugs.webkit.org/show_bug.cgi?id=131902
http://trac.webkit.org/changeset/167658
- 9:08 AM Changeset in webkit [167663] by
-
- 3 edits2 adds in trunk
[New Multicolumn] fast/multicol/multicol-with-child-renderLayer-for-input.html puts the textfield in the wrong place
https://bugs.webkit.org/show_bug.cgi?id=131808
Reviewed by Brady Eidson.
Source/WebCore:
Make sure FrameView is patched to pass in the update pagination flag when subtree layout
is done on textfields. It had code for the old multicolumn system but did not have a check
for the new multicolumn system.
Added fast/multicol/newmulticol/compare-with-old-impl/multicol-with-child-renderLayer-for-input.html
- page/FrameView.cpp:
(WebCore::updateLayerPositionFlags):
LayoutTests:
- fast/multicol/newmulticol/compare-with-old-impl/multicol-with-child-renderLayer-for-input-expected.html: Added.
- fast/multicol/newmulticol/compare-with-old-impl/multicol-with-child-renderLayer-for-input.html: Added.
- 8:51 AM Changeset in webkit [167662] by
-
- 2 edits in trunk/Source/WebKit2
WKThumbnailView doesn't quite work with topContentInset
https://bugs.webkit.org/show_bug.cgi?id=131976
<rdar://problem/16642127>
Reviewed by Darin Adler.
- UIProcess/API/Cocoa/_WKThumbnailView.mm:
(-[_WKThumbnailView initWithFrame:fromWKView:]):
_WKThumbnailView always needs a layer.
We don't want to use snapshots by default, this was debugging accidentally landed in the last change.
(-[_WKThumbnailView _requestSnapshotIfNeeded]):
Don't request a new snapshot if the scale is the same as the existing one.
Reduce the height of the snapshot by the topContentInset.
Having the origin at (0, 0) is still correct, as the inset content is in negative Y space,
and (0, 0) is the true origin of the content.
(-[_WKThumbnailView setScale:]):
When not snapshotting, we've decided that WKThumbnailView should apply the scale to the WKView layer tree,
instead of depending on clients to do so.
(-[_WKThumbnailView setUsesSnapshot:]):
We *do* need to apply thumbnail scale in the web process even when snapshotting,
because we use the live layer tree temporarily while waiting for the asynchronous
snapshot to return. This was a misstep in the last thumbnail view patch.
- 8:39 AM Changeset in webkit [167661] by
-
- 4 edits in trunk/Source/WebKit/efl
[EFL][WK1] There are many warnings in layout test since r167428
https://bugs.webkit.org/show_bug.cgi?id=131985
Reviewed by Gyuyoung Kim.
Since r167428, implementation of flush method was removed, but ewk_view_frame_main_cleared
still checked and generated many warnings.
Now, ewk_view_frame_main_cleared is not required because flush is already
meaningless and others are covered by ewk_frame_view_create_for_view.
So, this patch removed ewk_view_frame_main_cleared and related code which is already dead.
- WebCoreSupport/FrameLoaderClientEfl.cpp:
(WebCore::FrameLoaderClientEfl::transitionToCommittedForNewPage):
- ewk/ewk_view.cpp:
(ewk_view_frame_main_cleared): Deleted.
(ewk_view_did_first_visually_nonempty_layout): Deleted.
(ewk_view_dispatch_did_finish_loading): Deleted.
- ewk/ewk_view.h: Removed flush smart method.
- 8:19 AM Changeset in webkit [167660] by
-
- 2 edits in trunk/Source/WebKit2
[WebKit2] Cleanup the build from unused parameter in WebProcess Module
https://bugs.webkit.org/show_bug.cgi?id=131970
Patch by Shivakumar JM <shiva.jm@samsung.com> on 2014-04-22
Reviewed by Darin Adler.
Fix a build warning by removing the parameters in function paintLayerContents() to fix build warning.
- WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.cpp:
(WebKit::CoordinatedLayerTreeHost::paintLayerContents):
- 7:17 AM Changeset in webkit [167659] by
-
- 4 edits2 adds in trunk
border-style: double rendered as solid when combined with border-radius
https://bugs.webkit.org/show_bug.cgi?id=131927
Reviewed by Simon Fraser.
BorderEdge::getDoubleBorderStripeWidths() should return the calculated inner, outer border widths.
Source/WebCore:
Test: fast/borders/hidpi-double-border-with-border-radius-always-produce-solid-line.html
- rendering/BorderEdge.cpp:
(WebCore::BorderEdge::getDoubleBorderStripeWidths):
- rendering/BorderEdge.h:
LayoutTests:
- fast/borders/hidpi-double-border-with-border-radius-always-produce-solid-line-expected.html: Added.
- fast/borders/hidpi-double-border-with-border-radius-always-produce-solid-line.html: Added.
- 7:06 AM Changeset in webkit [167658] by
-
- 3 edits in trunk/Source/WebCore
[Mac] don't ask for AVAssetTrack properties before they are available
https://bugs.webkit.org/show_bug.cgi?id=131902
<rdar://problem/16505076>
Reviewed by Jer Noble.
No new tests, the behavior this changes can not be tested with a layout test.
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.h:
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::MediaPlayerPrivateAVFoundationObjC): Initialize
m_cachedTotalBytes.
(WebCore::MediaPlayerPrivateAVFoundationObjC::beginLoadingMetadata): Don't report that
metadata has been loaded until the track properties we need have been loaded too.
(WebCore::MediaPlayerPrivateAVFoundationObjC::totalBytes): Cache totalBytes instead
of recalculating it every time.
(WebCore::MediaPlayerPrivateAVFoundationObjC::tracksDidChange): Invalidate cached
total bytes.
(WebCore::assetTrackMetadataKeyNames): Array of AVAssetTrack properties we use.
- 6:39 AM Changeset in webkit [167657] by
-
- 2 edits in releases/WebKitGTK/webkit-2.4/Source/WebKit2
Merge r167656 - [GTK] Crash after getting web view context property with g_object_get
https://bugs.webkit.org/show_bug.cgi?id=131983
Reviewed by Philippe Normand.
The problem is that the getter is using g_value_take_object() and
the default context is destroyed when the caller releases its ref.
- UIProcess/API/gtk/WebKitWebView.cpp:
(webkitWebViewGetProperty): Use g_value_set_object() instead of
g_value_take_object().
- 6:36 AM Changeset in webkit [167656] by
-
- 2 edits in trunk/Source/WebKit2
[GTK] Crash after getting web view context property with g_object_get
https://bugs.webkit.org/show_bug.cgi?id=131983
Reviewed by Philippe Normand.
The problem is that the getter is using g_value_take_object() and
the default context is destroyed when the caller releases its ref.
- UIProcess/API/gtk/WebKitWebView.cpp:
(webkitWebViewGetProperty): Use g_value_set_object() instead of
g_value_take_object().
- 5:00 AM Changeset in webkit [167655] by
-
- 2 edits in trunk/LayoutTests
Unreviewed gardening.
- TestExpectations: fast/regions/cssom/region-range-for-box-crash.html is crashing on debug mode after r167652.
- 4:03 AM Changeset in webkit [167654] by
-
- 2 edits in trunk/Source/WebCore
Buildfix after r167650 broke non-Cocoa builds
https://bugs.webkit.org/show_bug.cgi?id=131980
Reviewed by Mario Sanchez Prada.
- page/EventHandler.cpp:
(WebCore::EventHandler::handleWheelEvent):
- 3:51 AM Changeset in webkit [167653] by
-
- 19 edits in trunk/Source/WebKit2
[EFL][WK2] Remove unnecessary includes in files of Source/WebKit2/UIProcess/API/efl
https://bugs.webkit.org/show_bug.cgi?id=131979
Reviewed by Gyuyoung Kim.
Removed unnecessary includes of header files.
In addition, tried to use forward declarations.
- UIProcess/API/efl/EwkView.cpp:
- UIProcess/API/efl/EwkView.h:
- UIProcess/API/efl/EwkViewCallbacks.h:
- UIProcess/API/efl/GestureRecognizer.h:
- UIProcess/API/efl/ewk_context_menu_item_private.h:
- UIProcess/API/efl/ewk_context_menu_private.h:
- UIProcess/API/efl/ewk_download_job.cpp:
- UIProcess/API/efl/ewk_download_job_private.h:
- UIProcess/API/efl/ewk_form_submission_request.cpp:
- UIProcess/API/efl/ewk_form_submission_request_private.h:
- UIProcess/API/efl/ewk_navigation_data.cpp:
- UIProcess/API/efl/ewk_navigation_data_private.h:
- UIProcess/API/efl/ewk_page_group_private.h:
- UIProcess/API/efl/ewk_popup_menu_private.h:
- UIProcess/API/efl/ewk_view.cpp:
- UIProcess/API/efl/ewk_window_features_private.h:
- UIProcess/efl/PageUIClientEfl.cpp:
- UIProcess/efl/ViewClientEfl.cpp:
- 3:22 AM Changeset in webkit [167652] by
-
- 15 edits4 adds in trunk
[CSS Regions] Selection highlight doesn't match DOM selection
https://bugs.webkit.org/show_bug.cgi?id=131511
Reviewed by David Hyatt.
Source/WebCore:
This patch implements a subtrees approach to fix selection interaction in CSS Regions.
The idea is to divide the Render tree in subtrees. One subtree will be the RenderView (which is always present),
then for each RenderNamedFlowThread (which are children of RenderView) we will have an extra subtree.
For each subtree it determines the start and end positions according to the DOM tree. So, every subtree will
always have a start and end position, and the selection algorithm will be able to reach the end from the start
without problems (this allows us to remove the changes introduced in r155058).
Then selection related methods in RenderView are adapted to ensure that they are called for each subtree. Making
the selection highlight to match DOM selection as expected.
Test: fast/regions/selection/selecting-text-including-region.html
- CMakeLists.txt: Add new class SelectionSubtreeRoot to build file.
- WebCore.vcxproj/WebCore.vcxproj: Ditto.
- WebCore.vcxproj/WebCore.vcxproj.filters: Ditto.
- WebCore.xcodeproj/project.pbxproj: Ditto.
- rendering/RenderNamedFlowThread.h: Inherit from SelectionSubtreeRoot.
- rendering/RenderObject.cpp:
(WebCore::RenderObject::selectionRoot): New method to return the RenderNamedFlowThread if any or the RenderView
otherwise.
(WebCore::RenderObject::selectionStartEnd): If it's inside a RenderNamedFlowThread use specific method instead
of the RenderView's one.
- rendering/RenderObject.h: Add new method signature.
- rendering/RenderView.cpp:
(WebCore::RenderView::selectionBounds): Adapt method to subtrees approach.
(WebCore::RenderView::subtreeSelectionBounds): Modify previous method in order to use SelectionSubtreeRoot
objects.
(WebCore::RenderView::repaintSelection): Adapt method to subtrees approach.
(WebCore::RenderView::repaintSubtreeSelection): Modify previous method in order to use SelectionSubtreeRoot
objects.
(getNextOrPrevRenderObjectBasedOnDirection): Deleted method from r155058 as it is not needed anymore. Each
subtree will have start and end positions, so the selection algorithm will always find the end.
(WebCore::RenderView::setSelection): Adapt method to subtrees approach.
(WebCore::RenderView::splitSelectionBetweenSubtrees): Determine start and end positions for each subtree and
call setSubtreeSelection().
(WebCore::RenderView::setSubtreeSelection): Modify previous method in order to use SelectionSubtreeRoot objects.
- rendering/RenderView.h: Inherit from SelectionSubtreeRoot. Add signature for new helper methods receiving
SelectionSubtreeRoot objects.
- rendering/SelectionSubtreeRoot.cpp: Added.
(WebCore::SelectionSubtreeRoot::SelectionSubtreeRoot):
(WebCore::SelectionSubtreeRoot::selectionStartEndPositions):
- rendering/SelectionSubtreeRoot.h: Added. New class to identify root elements for the subtrees. With some basic
attributes and methods to store the start and end positions and offsets.
(WebCore::SelectionSubtreeRoot::selectionStart):
(WebCore::SelectionSubtreeRoot::selectionStartPos):
(WebCore::SelectionSubtreeRoot::selectionEnd):
(WebCore::SelectionSubtreeRoot::selectionEndPos):
(WebCore::SelectionSubtreeRoot::setSelectionStart):
(WebCore::SelectionSubtreeRoot::setSelectionStartPos):
(WebCore::SelectionSubtreeRoot::setSelectionEnd):
(WebCore::SelectionSubtreeRoot::setSelectionEndPos):
LayoutTests:
Add new test to check highlight of render objects moved under RenderNamedFlowThread in the render tree.
Updated some tests to avoid issues with selection gap filling in the new implementation.
- fast/regions/selection/selecting-text-including-region-expected.html: Added.
- fast/regions/selection/selecting-text-including-region.html: Added.
- fast/regions/selection/selecting-text-through-different-region-flows-2-expected.html: Updated.
- fast/regions/selection/selecting-text-through-different-region-flows-2.html: Updated.
- fast/regions/selection/selection-direction-expected.html: Updated.
- fast/regions/selection/selection-direction.html: Updated.
- 1:29 AM Changeset in webkit [167651] by
-
- 2 edits in trunk
Fix a typo in WebKitFeatures.cmake -- INITALVALUE -> INITIALVALUE.
Rubber-stamped by Carlos Garcia Campos.
- Source/cmake/WebKitFeatures.cmake:
