Timeline
Jul 27, 2015:
- 11:01 PM Changeset in webkit [187483] by
-
- 8 edits3 adds in trunk/Source/JavaScriptCore
[ES6] Implement Reflect.enumerate
https://bugs.webkit.org/show_bug.cgi?id=147347
Reviewed by Sam Weinig.
This patch implements Reflect.enumerate.
It returns the iterator that iterates the enumerable keys of the given object.
It follows the for-in's enumeration order.
To implement it, we write down the same logic to the for-in's enumeration code in C++.
- CMakeLists.txt:
- JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
- JavaScriptCore.vcxproj/JavaScriptCore.vcxproj.filters:
- JavaScriptCore.xcodeproj/project.pbxproj:
- runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
(JSC::JSGlobalObject::visitChildren):
- runtime/JSGlobalObject.h:
(JSC::JSGlobalObject::propertyNameIteratorStructure):
- runtime/JSPropertyNameIterator.cpp: Added.
(JSC::JSPropertyNameIterator::JSPropertyNameIterator):
(JSC::JSPropertyNameIterator::clone):
(JSC::JSPropertyNameIterator::create):
(JSC::JSPropertyNameIterator::finishCreation):
(JSC::JSPropertyNameIterator::visitChildren):
(JSC::JSPropertyNameIterator::next):
(JSC::propertyNameIteratorFuncNext):
- runtime/JSPropertyNameIterator.h: Added.
(JSC::JSPropertyNameIterator::createStructure):
- runtime/ReflectObject.cpp:
(JSC::reflectObjectEnumerate):
- tests/stress/reflect-enumerate.js: Added.
(shouldBe):
(shouldThrow):
- 9:52 PM Changeset in webkit [187482] by
-
- 2 edits in branches/safari-601.1-branch/Source/WebKit2
Roll out r187376. rdar://problem/22000647
- 9:35 PM Changeset in webkit [187481] by
-
- 2 edits in trunk/Source/WebKit2
[iOS] REGRESSION (r187376): Form editing buttons are missing from the Shortcut Bar
https://bugs.webkit.org/show_bug.cgi?id=147351
Reviewed by Adele Peterson.
We were relying on -[WKContentView inputAccessoryView] getting called to initialize the form
accessory view, even when that method would always return nil, as it does on iPad. After
r187376, that method changed to initializing the view only when returning it, which caused
this bug.
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _displayFormNodeInputView]): Call the new -_ensureFormAccessoryView from
here, when we are going to need the view.
(-[WKContentView _ensureFormAccessoryView]): Moved initialization code from
-inputAccessoryView here, and removed outdated +instancesRespondToSelector: check.
(-[WKContentView inputAccessoryView]): Now calls through -formAccessoryView.
(-[WKContentView formAccessoryView]): Added call to -_ensureFormAccessoryView here.
- 9:29 PM Changeset in webkit [187480] by
-
- 2 edits in trunk/Source/WebInspectorUI
Web Inspector: Record button in Timelines view is clipped on non-retina. Hovering over the dot draws it correctly.
https://bugs.webkit.org/show_bug.cgi?id=147346
Reviewed by Joseph Pecoraro.
Since the hovered button isn't clipped (-webkit-filter is applied in that case), add a filter that
doesn't alter the button's appearance to the non-hovered selector.
- UserInterface/Views/TimelineSidebarPanel.css:
(.sidebar > .panel.navigation.timeline > .status-bar > .record-glyph):
Added brightness 100% filter to non-hovered selector.
- 9:23 PM Changeset in webkit [187479] by
-
- 2 edits1 add in trunk/Source/JavaScriptCore
[ES6] Implement Reflect.preventExtensions
https://bugs.webkit.org/show_bug.cgi?id=147331
Reviewed by Sam Weinig.
Implement Reflect.preventExtensions.
This is different from Object.preventExensions.
- When preventExtensions is called onto the non-object, it raises the TypeError.
- Reflect.preventExtensions does not raise the TypeError when the preventExtensions operation is failed.
For the (2) case, since there is no Proxy implementation currently, Reflect.preventExtensions always succeed.
- runtime/ReflectObject.cpp:
(JSC::reflectObjectPreventExtensions):
- tests/stress/reflect-prevent-extensions.js: Added.
(shouldBe):
(shouldThrow):
- 8:26 PM Changeset in webkit [187478] by
-
- 2 edits in trunk/Source/WebKit2
WebsiteDataStore should clean up its storage in the network process when destroyed.
<https://webkit.org/b/147349>
<rdar://problem/21838764>
Reviewed by Darin Adler.
Have ~WebsiteDataStore() send a DestroyPrivateBrowsingSession message to all networking processes
for ephemeral sessions. This plugs a NetworkStorageSession leak that could retain a large
CFNetwork object graph.
This complements r187115 which did the same for network process storage owned by
API::Session objects.
- UIProcess/WebsiteData/WebsiteDataStore.cpp:
(WebKit::WebsiteDataStore::~WebsiteDataStore):
- 7:19 PM Changeset in webkit [187477] by
-
- 5 edits in branches/safari-600.1.4.17-branch/Source
Versioning.
- 7:18 PM Changeset in webkit [187476] by
-
- 5 edits in branches/safari-600.8-branch/Source
Versioning.
- 7:17 PM Changeset in webkit [187475] by
-
- 5 edits in branches/safari-601.1-branch/Source
Versioning.
- 6:02 PM Changeset in webkit [187474] by
-
- 1 copy in tags/Safari-601.1.44
New tag.
- 6:02 PM Changeset in webkit [187473] by
-
- 1 copy in tags/Safari-600.8.7
New tag.
- 6:02 PM Changeset in webkit [187472] by
-
- 1 copy in tags/Safari-600.1.4.17.5
New tag.
- 5:43 PM Changeset in webkit [187471] by
-
- 2 edits in trunk/Source/WebKit2
First in-window viewStateChange synchronously blocks despite not previously being in-window
https://bugs.webkit.org/show_bug.cgi?id=147344
<rdar://problem/22021772>
Reviewed by Simon Fraser.
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::updateViewState):
(WebKit::WebPageProxy::dispatchViewStateChange):
The whole point of m_viewWasEverInWindow was so that we would not
synchronously wait when a view was added to a window for the first time,
only all subsequent times.
However, since m_viewWasEverInWindow was being set *before* being
checked in dispatchViewStateChange, we were always blocking. This is
a huge waste of main-thread time, because there's no reason to wait
for the first paint if you've never seen the view before (and shouldn't
expect it to have content).
Instead, set the flag after dispatching a view state change, so it becomes
"have we ever sent a view state with IsInWindow set" instead.
- 5:32 PM Changeset in webkit [187470] by
-
- 11 edits in trunk
Use Ninja on Windows.
https://bugs.webkit.org/show_bug.cgi?id=147228
Reviewed by Martin Robinson.
.:
- Source/cmake/OptionsWindows.cmake:
Only use /MP when using generated Visual Studio solution files to build.
It makes compiling parallel in MSVC, but Ninja doesn't like it.
Source/JavaScriptCore:
- CMakeLists.txt:
Set the working directory when generating LowLevelInterpreterWin.asm to put LowLevelInterpreterWin.asm.sym in the right place.
Source/WebKit:
- PlatformWin.cmake:
Added /NODEFAULTLIB:LIBCMT to link WebKit.dll successfully with Ninja.
Source/WebKit/win:
- WebView.h:
Add more #if USE(CA) to protect the CACFLayer code.
Tools:
- WinLauncher/CMakeLists.txt:
WinLauncherLib must make WinLauncher.dll, not WinLauncherLib.dll.
- Scripts/build-webkit:
Use Ninja to build WinCairo by default, but fall back to Visual Studio if Ninja is not on the system.
- 5:20 PM Changeset in webkit [187469] by
-
- 2 edits in trunk/Source/WebCore
Adding Web Components to feature status page.
https://bugs.webkit.org/show_bug.cgi?id=147270
Reviewed by Timothy Hatcher.
- features.json:
- 5:13 PM Changeset in webkit [187468] by
-
- 13 edits in trunk/Source/WebInspectorUI
Web Inspector: Clicking a frame in the Rendering Frames timeline should select the corresponding grid row
https://bugs.webkit.org/show_bug.cgi?id=147013
Reviewed by Timothy Hatcher.
Added ability to select (or highlight) a particular frame by clicking in the overview graph. A tick mark
appears above the selected frame inside the ruler, and the frame element's background color changes to a light
blue vertical highlight. Frame selection is synced across the timeline view's data grid, overview graph, and
navigation bar path components.
A few behaviors:
- Clicking a frame outside the current ruler selection selects the frame.
- Clicking the selected frame will deselect it.
- Resizing the ruler selection to exclude the selected frame will deselect it.
- UserInterface/Views/DataGrid.js:
(WebInspector.DataGridNode.prototype.select):
No need to generate two SelectedNodeChanged events when selecting a node causes
the currently selected node to be deselected.
- UserInterface/Views/LayoutTimelineOverviewGraph.js:
(WebInspector.LayoutTimelineOverviewGraph):
- UserInterface/Views/NetworkTimelineOverviewGraph.js:
(WebInspector.NetworkTimelineOverviewGraph):
- UserInterface/Views/ScriptTimelineOverviewGraph.js:
(WebInspector.ScriptTimelineOverviewGraph): Set TimelineOverview during construction.
Set TimelineOverview during construction.
- UserInterface/Views/RenderingFrameTimelineOverviewGraph.css:
(.timeline-overview-graph.rendering-frame > .frame-marker):
(body.window-inactive .timeline-overview-graph.rendering-frame > .frame-marker):
Added styles for selected frame marker.
- UserInterface/Views/RenderingFrameTimelineOverviewGraph.js:
(WebInspector.RenderingFrameTimelineOverviewGraph):
Set TimelineOverview during construction.
(WebInspector.RenderingFrameTimelineOverviewGraph.prototype._mouseClicked):
Added frame marker element and frame selection via mouse click.
- UserInterface/Views/RenderingFrameTimelineView.js:
(WebInspector.RenderingFrameTimelineView.prototype.treeElementDeselected):
Deselect grid node when frame is deselected from the overview graph.
(WebInspector.RenderingFrameTimelineView.prototype.filterDidChange): Deleted.
Removed unnecessary override.
- UserInterface/Views/TimelineOverview.js:
Added selectRecord method and RecordSelected event.
(WebInspector.TimelineOverview):
(WebInspector.TimelineOverview.prototype._timelineRulerMouseDown):
(WebInspector.TimelineOverview.prototype._timelineRulerMouseClicked):
Since timeline ruler specifies pointer-events: all, click events must be forwarded to the
underlying overview graph. Click events following changes to the ruler's selection are ignored.
(WebInspector.TimelineOverview.prototype._timeRangeSelectionChanged):
(WebInspector.TimelineOverview.prototype._recordSelected):
Forward RecordSelected events from child graphs to clients of the overview.
- UserInterface/Views/TimelineOverviewGraph.js:
Added RecordSelected event.
(WebInspector.TimelineOverviewGraph):
(WebInspector.TimelineOverviewGraph.prototype.get visible):
(WebInspector.TimelineOverviewGraph.prototype.get selectedRecord):
(WebInspector.TimelineOverviewGraph.prototype.set selectedRecord):
Added property for selecting a record in the graph. Derived graph types can implement record selection
and choose how to style the selected record.
(WebInspector.TimelineOverviewGraph.prototype.updateLayout):
(WebInspector.TimelineOverviewGraph.prototype.needsLayout):
Update frame marker during an animation frame.
(WebInspector.TimelineOverviewGraph.prototype.dispatchSelectedRecordChangedEvent):
Dispatch wrapper to simplify things.
(WebInspector.TimelineOverviewGraph.prototype.updateSelectedRecord):
(WebInspector.TimelineOverviewGraph.prototype._needsSelectedRecordLayout.update):
(WebInspector.TimelineOverviewGraph.prototype._needsSelectedRecordLayout):
Inform derived overview graph to optionally style the selected record.
(WebInspector.TimelineOverviewGraph.prototype.set timelineOverview): Deleted.
No longer needed. TimelineOverview set during construction.
- UserInterface/Views/TimelineRecordFrame.css:
(.timeline-record-frame):
(.timeline-record-frame.tall):
Limit the "fade out" effect to frames that are >= 95% of the graph height, otherwise the
gradient mask is noticeable when selecting a shorter frame.
(.timeline-record-frame.selected):
(body.window-inactive .timeline-record-frame.selected):
(.timeline-record-frame.selected > .frame):
Added highlight (and dimmed inactive highlight) for selected frame element.
- UserInterface/Views/TimelineRecordFrame.js:
(WebInspector.TimelineRecordFrame.prototype.get selected):
(WebInspector.TimelineRecordFrame.prototype.set selected):
Added property for setting selected style.
- UserInterface/Views/TimelineRecordingContentView.js:
(WebInspector.TimelineRecordingContentView):
(WebInspector.TimelineRecordingContentView.prototype._contentViewSelectionPathComponentDidChange):
Select record in overview in response to changing navigation path.
(WebInspector.TimelineRecordingContentView.prototype._recordSelected.get console):
(WebInspector.TimelineRecordingContentView.prototype._recordSelected):
Update selected tree element in response to changing record selection in the overview graph.
- 5:12 PM Changeset in webkit [187467] by
-
- 4 edits in trunk/Source/WebCore
ShrinkToFit() all the DFA's vectors when we no longer need to add new nodes to it
https://bugs.webkit.org/show_bug.cgi?id=147341
Patch by Benjamin Poulain <bpoulain@apple.com> on 2015-07-27
Reviewed by Geoffrey Garen.
Add a finalizer to the DFA to reduce the total memory needed
to store it.
On real test lists, this cuts the peak memory usage by about 1 megabyte.
The NFA already had a finalizer, no need to modify that one.
- contentextensions/DFA.cpp:
(WebCore::ContentExtensions::DFA::finalize):
- contentextensions/DFA.h:
- contentextensions/NFAToDFA.cpp:
(WebCore::ContentExtensions::NFAToDFA::convert):
- 5:06 PM Changeset in webkit [187466] by
-
- 4 edits in trunk/Source/WebCore
Crash in WebCore::DocumentLoader::willSendRequest() with ContentFilter and AppCache.
<rdar://problem/21960398> and https://bugs.webkit.org/show_bug.cgi?id=147339
Reviewed by Alexey Proskuryakov.
No new tests (Not yet proven to be possible to test this).
- loader/DocumentLoader.cpp:
(WebCore::DocumentLoader::willSendRequest): Grab the identifier from the CachedResource directly, not from the null ResourceLoader.
(WebCore::DocumentLoader::continueAfterNavigationPolicy): Null check the ResourceLoader, as it can definitely be gone by this point.
- loader/cache/CachedResource.cpp:
(WebCore::CachedResource::clearLoader): Save off the identifier for later use.
- loader/cache/CachedResource.h:
(WebCore::CachedResource::identifierForLoadWithoutResourceLoader): Expose the identifier that the ResourceLoader had when it went away.
- 4:53 PM Changeset in webkit [187465] by
-
- 2 edits in trunk/Source/WebInspectorUI
Web Inspector: Revert change to DOMTreeElement.prototype.remove
https://bugs.webkit.org/show_bug.cgi?id=147336
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2015-07-27
Reviewed by Timothy Hatcher.
- UserInterface/Views/DOMTreeElement.js:
(WebInspector.DOMTreeElement.prototype.remove.removeNodeCallback):
This change was accidental, and just resulted in the function calling
itself again. It didn't infinite loops because another path would have
removed the DOMTreeElement, but lets correct this code.
- 4:51 PM Changeset in webkit [187464] by
-
- 5 edits3 adds in trunk
SparseValueMap check is skipped when the butterfly's vectorLength is larger than the access-requested index
https://bugs.webkit.org/show_bug.cgi?id=147265
Reviewed by Geoffrey Garen.
Source/JavaScriptCore:
JSObject's vector holds the indexed values and we leverage it to represent stored values and holes.
By checking that the given index is in-bound of the vector's length, we can look up the property fast.
And for the sparse array, we have also the separated SparseValueMap to hold the pairs.
And we need to take care that the length of the vector should not overlap the indices stored in the SparseValueMap.
The vector only holds the pure JS values to avoid additional checking for accessors when looking up the value
from the vector. To achieve this, we also store the accessors (and attributed properties) to SparseValueMap
even the index is less than MIN_SPARSE_ARRAY_INDEX.
As a result, if the length of the vector overlaps the indices of the accessors stored in the SparseValueMap,
we accidentally skip the phase looking up from the SparseValueMap. Instead, we just load from the vector and
if the loaded value is an array hole, we decide the given object does not have the value for the given index.
This patch fixes the problem.
When defining the attributed value that index is smaller than the length of the vector, we throw away the vector
and change the object to DictionaryIndexingMode. Since we can assume that indexed accessors rarely exist in
practice, we expect this does not hurt the performance while keeping the fast property access system without
checking the sparse map.
- runtime/JSObject.cpp:
(JSC::JSObject::putDirectIndexBeyondVectorLength):
- tests/stress/sparse-map-non-overlapping.js: Added.
(shouldBe):
(testing):
(object.get 1000):
- tests/stress/sparse-map-non-skip-getter-overriding.js: Added.
(shouldBe):
(obj.get 1):
(testing):
- tests/stress/sparse-map-non-skip.js: Added.
(shouldBe):
(testing):
(testing2):
(.get for):
LayoutTests:
The previous test "basic-computed-property-name.js" has the issue[1] and
the issue is fixed by this patch.
[1]: https://bugs.webkit.org/show_bug.cgi?id=142943
- js/basic-computed-property-name-expected.txt:
- js/script-tests/basic-computed-property-name.js:
- 4:38 PM Changeset in webkit [187463] by
-
- 3 edits2 adds in trunk
Crash happens when calling removeEventListener for an SVG element which has an instance inside a <defs> element of shadow tree
https://bugs.webkit.org/show_bug.cgi?id=147290
Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2015-07-27
Reviewed by Simon Fraser.
Source/WebCore:
When the shadow tree is built for a <use> element, all the SVG elements
are allowed to be cloned in the shadow tree but later some of the elements
are disallowed and removed. Make sure, when disallowing an element in the
shadow tree, to reset the correspondingElement relationship between all
the disallowed descendant SVG elements and all their original elements.
Test: svg/custom/remove-event-listener-shadow-disallowed-element.svg
- svg/SVGUseElement.cpp:
(WebCore::removeDisallowedElementsFromSubtree):
LayoutTests:
Make sure we do not crash when when calling removeEventListener() for an
element which is cloned under a disallowed parent inside the shadow tree
of another <use> element.
- svg/custom/remove-event-listener-shadow-disallowed-element-expected.txt: Added.
- svg/custom/remove-event-listener-shadow-disallowed-element.svg: Added.
- 4:27 PM Changeset in webkit [187462] by
-
- 2 edits in trunk/Source/WebKit2
[iOS] Long press or link click can sometimes trigger during/after a preview
https://bugs.webkit.org/show_bug.cgi?id=147338
<rdar://problem/22020770>
Reviewed by Enrica Casucci.
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _longPressRecognized:]):
Bail if we're previewing.
(-[WKContentView _dataForPreviewItemController:atPosition:type:]):
(-[WKContentView _presentedViewControllerForPreviewItemController:]):
Always avoid allowing clicks and start interaction when previewing a link.
- 4:25 PM Changeset in webkit [187461] by
-
- 5 edits in trunk/Source/WebCore
Rename AnimationBase::isTransformFunctionListValid()
https://bugs.webkit.org/show_bug.cgi?id=147334
Reviewed by Sam Weinig.
isTransformFunctionListValid() was a confusing name for this state. It's true
when the transform function lists for an animation or transition are matched,
so rename it to transformFunctionListsMatch().
No behavior change.
- page/animation/AnimationBase.h:
(WebCore::AnimationBase::transformFunctionListsMatch):
(WebCore::AnimationBase::isTransformFunctionListValid): Deleted.
- page/animation/CSSPropertyAnimation.cpp:
(WebCore::blendFunc):
- page/animation/ImplicitAnimation.cpp:
(WebCore::ImplicitAnimation::computeExtentOfTransformAnimation):
(WebCore::ImplicitAnimation::validateTransformFunctionList):
- page/animation/KeyframeAnimation.cpp:
(WebCore::KeyframeAnimation::computeExtentOfTransformAnimation):
(WebCore::KeyframeAnimation::validateTransformFunctionList):
- 4:04 PM Changeset in webkit [187460] by
-
- 13 edits in trunk/Source/JavaScriptCore
Reduce execution time for "let" and "const" tests
https://bugs.webkit.org/show_bug.cgi?id=147291
Reviewed by Geoffrey Garen.
We don't need to loop so many times for things that will not make it
into the DFG. Also, we can loop a lot less for almost all the tests
because they're mostly testing the bytecode generator.
- tests/stress/const-and-with-statement.js:
- tests/stress/const-exception-handling.js:
- tests/stress/const-loop-semantics.js:
- tests/stress/const-not-strict-mode.js:
- tests/stress/const-semantics.js:
- tests/stress/const-tdz.js:
- tests/stress/lexical-let-and-with-statement.js:
- tests/stress/lexical-let-exception-handling.js:
(assert):
- tests/stress/lexical-let-loop-semantics.js:
(assert):
(shouldThrowTDZ):
(.):
- tests/stress/lexical-let-not-strict-mode.js:
- tests/stress/lexical-let-semantics.js:
(.):
- tests/stress/lexical-let-tdz.js:
(shouldThrowTDZ):
(.):
- 4:01 PM Changeset in webkit [187459] by
-
- 2 edits in trunk/Source/WebKit2
[iOS] Occasional crashes in WebPage::elementDidBlur()'s async block.
<https://webkit.org/b/147281>
<rdar://problem/21701858>
Reviewed by Anders Carlsson.
Protect the WebPage object until the dispatch_async block has finished.
I have no way to repro this, but there's evidence of some crashing here.
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::elementDidBlur):
- 3:39 PM Changeset in webkit [187458] by
-
- 14 edits1 add in trunk
Progress towards building AppleWin with CMake
https://bugs.webkit.org/show_bug.cgi?id=147325
Reviewed by Martin Robinson.
.:
- Source/cmake/OptionsWindows.cmake:
Link with 64-bit libraries if building 64-bit binaries.
Don't run regular expressions on empty strings when using Ninja.
Removed DebugSuffix Visual Studio environment variables.
Source/WebCore:
- PlatformAppleWin.cmake: Added.
- PlatformWin.cmake:
- PlatformWinCairo.cmake:
Moved a lot of common directories from PlatformWinCairo to PlatformWin.
Source/WebKit:
- PlatformWin.cmake:
Removed references to non-existing files.
Source/WebKit/win:
- WebError.cpp:
- WebKitDLL.cpp:
- WebPreferences.cpp:
- WebURLResponse.cpp:
Fix some include quirks.
Source/WTF:
- wtf/PlatformWin.cmake:
Link with CoreFoundation.lib.
- 3:30 PM Changeset in webkit [187457] by
-
- 11 edits in branches/safari-600.8-branch
Roll out r180020. rdar://problem/21716372
- 3:25 PM Changeset in webkit [187456] by
-
- 7 edits1 copy2 deletes in branches/safari-600.8-branch
Roll out r182829. rdar://problem/21716372
- 3:18 PM Changeset in webkit [187455] by
-
- 4 edits in trunk/Source/WebKit2
[Seccomp] Set appropriate filters when trapping syscalls by default
https://bugs.webkit.org/show_bug.cgi?id=142983
If we trap syscalls by default, we must not set separate filters to trap
anything here, since it will fail causing us to crash.
But also, there are some things we must allow unconditionally even when
trapping by default. sigreturn, obviously. Also, let's whitelist brk
here instead of in platform-specific code.
Reviewed by Žan Doberšek.
- Shared/linux/SeccompFilters/SeccompBroker.cpp:
(WebKit::SeccompBroker::launchProcess): Don't trap sigprocmask or sigaction unless allow is
the default action. Also, allow sigreturn and brk is allow is not the default.
- Shared/linux/SeccompFilters/SeccompFilters.cpp: Added
(WebKit::SeccompFilters::defaultAction):
- Shared/linux/SeccompFilters/SeccompFilters.h: Added defaultAction
- 3:10 PM Changeset in webkit [187454] by
-
- 11 edits in branches/safari-600.1.4.17-branch
Roll out r180020. rdar://problem/21716511
- 3:07 PM Changeset in webkit [187453] by
-
- 7 edits1 copy2 deletes in branches/safari-600.1.4.17-branch
Roll out r182829. rdar://problem/21716511
- 2:52 PM Changeset in webkit [187452] by
-
- 7 edits5 adds in trunk/LayoutTests
[Win] Test gardening after r187380.
https://bugs.webkit.org/show_bug.cgi?id=121902
Unreviewed.
- platform/win/TestExpectations:
- platform/win/fast/css/text-overflow-ellipsis-text-align-center-expected.png:
- platform/win/fast/css/text-overflow-ellipsis-text-align-center-expected.txt:
- platform/win/fast/css/text-overflow-ellipsis-text-align-left-expected.png:
- platform/win/fast/css/text-overflow-ellipsis-text-align-left-expected.txt:
- platform/win/fast/css/text-overflow-ellipsis-text-align-right-expected.png:
- platform/win/fast/css/text-overflow-ellipsis-text-align-right-expected.txt:
- platform/win/fast/css/vertical-text-overflow-ellipsis-text-align-center-expected.png:
- platform/win/fast/css/vertical-text-overflow-ellipsis-text-align-center-expected.txt:
- platform/win/fast/css/vertical-text-overflow-ellipsis-text-align-right-expected.png:
- platform/win/fast/css/vertical-text-overflow-ellipsis-text-align-right-expected.txt:
- 2:25 PM Changeset in webkit [187451] by
-
- 2 edits in trunk/Source/WebCore
Fix the build after r187383
- platform/spi/cocoa/QuartzCoreSPI.h:
- 2:20 PM Changeset in webkit [187450] by
-
- 3 edits in trunk/Source/WebInspectorUI
Web Inspector: support live editing of rule selectors
https://bugs.webkit.org/show_bug.cgi?id=139153
Reviewed by Timothy Hatcher.
- UserInterface/Controllers/DOMTreeManager.js:
(WebInspector.DOMTreeManager.prototype.highlightSelector):
Moved from CSSStyleDeclarationSection.
- UserInterface/Views/CSSStyleDeclarationSection.js:
(WebInspector.CSSStyleDeclarationSection):
(WebInspector.CSSStyleDeclarationSection.prototype.get _currentSelectorText):
Returns the current selector text, either from the style ownerRule or the selector element text.
(WebInspector.CSSStyleDeclarationSection.prototype._highlightNodesWithSelector):
Now highlights all nodes matching the current selector instead of the ownerRule's selector.
(WebInspector.CSSStyleDeclarationSection.prototype._hideDOMNodeHighlight):
(WebInspector.CSSStyleDeclarationSection.prototype._handleMouseOver):
(WebInspector.CSSStyleDeclarationSection.prototype._handleMouseOut):
(WebInspector.CSSStyleDeclarationSection.prototype._handleKeyDown):
If the character is not a tab, highlight all nodes matching the current selector text.
(WebInspector.CSSStyleDeclarationSection.prototype._handleKeyUp):
(WebInspector.CSSStyleDeclarationSection.prototype._hideHighlightOnNodesWithSelector): Deleted.
- 2:08 PM Changeset in webkit [187449] by
-
- 2 edits1 add in trunk/Tools
Test for bug #147164 "[iOS] Keyboard bold/italic/underline keys don't
highlight after being tapped to style a selection"
https://bugs.webkit.org/show_bug.cgi?id=147278
Reviewed by Ryosuke Niwa.
- TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
- TestWebKitAPI/Tests/mac/WordBoundaryTypingAttributes.mm: Added.
(-[WordBoundaryTypingAttributesController webView:didFinishLoadForFrame:]):
(TestWebKitAPI::TEST):
Load a page with the string "Hello, world!", then select "world" and
underline it. Test that the typing attributes for the current selection
include the underline attribute.
- 1:56 PM Changeset in webkit [187448] by
-
- 2 edits in trunk/Source/WebCore
WKWebsiteDataStore remove methods don't properly delete cookies
https://bugs.webkit.org/show_bug.cgi?id=147333
rdar://problem/21948230
Reviewed by Tim Horton.
If there are multiple cookies for a single domain, make sure to delete all of them
and not just the first one we find.
Fix this by keeping a mapping from domain to a list of cookies.
- platform/network/mac/CookieJarMac.mm:
(WebCore::deleteCookiesForHostnames):
- 12:55 PM Changeset in webkit [187447] by
-
- 2 edits in branches/safari-601.1-branch/Source/WebKit2
Merge r187412. rdar://problem/22003112
- 12:55 PM Changeset in webkit [187446] by
-
- 2 edits in branches/safari-601.1-branch/Source/WebKit2
Merge r187376. rdar://problem/22000647
- 12:43 PM Changeset in webkit [187445] by
-
- 2 edits in branches/safari-601.1-branch/LayoutTests
More test gardening after r187346 ('>>' descendant combinator).
- platform/mac/TestExpectations:
- 12:40 PM Changeset in webkit [187444] by
-
- 7 edits2 copies in branches/safari-601.1-branch
Merge r187443.
2015-07-27 Alexey Proskuryakov <ap@apple.com>
Test result gardening for El Capitan.
- platform/mac-yosemite/fast/text/international/bold-bengali-expected.png: Copied from LayoutTests/platform/mac/fast/text/international/bold-bengali-expected.png.
- platform/mac-yosemite/fast/text/international/bold-bengali-expected.txt: Copied from LayoutTests/platform/mac/fast/text/international/bold-bengali-expected.txt.
- platform/mac/fast/text/font-weights-zh-expected.txt:
- platform/mac/fast/text/international/bold-bengali-expected.png:
- platform/mac/fast/text/international/bold-bengali-expected.txt:
- platform/mac/fast/text/international/danda-space-expected.png:
- platform/mac/fast/text/international/danda-space-expected.txt:
- 12:37 PM Changeset in webkit [187443] by
-
- 6 edits2 copies in trunk/LayoutTests
Test result gardening for El Capitan.
- platform/mac-yosemite/fast/text/international/bold-bengali-expected.png: Copied from LayoutTests/platform/mac/fast/text/international/bold-bengali-expected.png.
- platform/mac-yosemite/fast/text/international/bold-bengali-expected.txt: Copied from LayoutTests/platform/mac/fast/text/international/bold-bengali-expected.txt.
- platform/mac/fast/text/font-weights-zh-expected.txt:
- platform/mac/fast/text/international/bold-bengali-expected.png:
- platform/mac/fast/text/international/bold-bengali-expected.txt:
- platform/mac/fast/text/international/danda-space-expected.png:
- platform/mac/fast/text/international/danda-space-expected.txt:
- 12:27 PM Changeset in webkit [187442] by
-
- 3 edits in branches/safari-601.1-branch
2015-07-27 Alexey Proskuryakov <ap@apple.com>
More test expectations for rdar://problem/22014672
- platform/mac/TestExpectations:
2015-07-27 Alexey Proskuryakov <ap@apple.com>
Add test expectations for rdar://problem/22014672
- platform/mac/TestExpectations:
- 12:24 PM Changeset in webkit [187441] by
-
- 2 edits in trunk/LayoutTests
More test expectations for rdar://problem/22014672
- platform/mac/TestExpectations:
- 12:10 PM Changeset in webkit [187440] by
-
- 5 edits in trunk/Source/JavaScriptCore
Rename PropertyNameMode::Both to PropertyNameMode::StringsAndSymbols
https://bugs.webkit.org/show_bug.cgi?id=147311
Reviewed by Sam Weinig.
To make the meaning clear in the user side (PropertyNameArray array(exec, PropertyNameMode::StringsAndSymbols)),
this patch renames PropertyNameMode::Both to PropertyNameMode::StringsAndSymbols.
- bytecode/ObjectAllocationProfile.h:
(JSC::ObjectAllocationProfile::possibleDefaultPropertyCount):
- runtime/EnumerationMode.h:
- runtime/ObjectConstructor.cpp:
(JSC::ownEnumerablePropertyKeys):
(JSC::defineProperties):
(JSC::objectConstructorSeal):
(JSC::objectConstructorFreeze):
(JSC::objectConstructorIsSealed):
(JSC::objectConstructorIsFrozen):
(JSC::ownPropertyKeys):
- runtime/ReflectObject.cpp:
(JSC::reflectObjectOwnKeys):
- 12:10 PM Changeset in webkit [187439] by
-
- 2 edits in trunk/LayoutTests
Add test expectations for rdar://problem/22014672
- platform/mac/TestExpectations:
- 11:54 AM Changeset in webkit [187438] by
-
- 2 edits in trunk/Source/WebCore
Enhance Animation logging slightly
https://bugs.webkit.org/show_bug.cgi?id=147324
Reviewed by Anders Carlsson.
Include the value of "param" in animation logging, because sometimes it's negative,
and log state changes in a few more places.
- page/animation/AnimationBase.cpp:
(WebCore::AnimationBase::updateStateMachine):
- 11:52 AM Changeset in webkit [187437] by
-
- 2 edits in trunk/Source/JavaScriptCore
Added a comment explaining that all "addVar()"s should happen before
emitting bytecode for a function's default parameter expressions
Rubber Stamped by Mark Lam.
- bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::BytecodeGenerator):
- 11:48 AM Changeset in webkit [187436] by
-
- 6 edits in trunk/Source
[Curl] Crash in CurlDownload::didReceiveHeader when downloading file.
https://bugs.webkit.org/show_bug.cgi?id=146832
Reviewed by Darin Adler.
Source/WebCore:
Only call ResourceResponse::setMimeType from the main thread.
Also, CurlDownload should be reference counted to make sure it
still lives when a function call is invoked on the main thread
from the download thread.
- platform/network/curl/CurlDownload.cpp:
(WebCore::CurlDownloadManager::downloadThread):
(WebCore::CurlDownload::CurlDownload):
(WebCore::CurlDownload::start):
(WebCore::CurlDownload::didReceiveHeader):
(WebCore::CurlDownload::didReceiveData):
- platform/network/curl/CurlDownload.h:
(WebCore::CurlDownloadListener::didFail):
(WebCore::CurlDownload::setListener):
Source/WebKit/win:
CurlDownload should be reference counted to make sure it still
lives when a function call is invoked on the main thread from
the download thread.
- WebDownload.h:
- WebDownloadCurl.cpp:
(WebDownload::init):
(WebDownload::start):
(WebDownload::cancel):
(WebDownload::deletesFileUponFailure):
(WebDownload::setDeletesFileUponFailure):
(WebDownload::setDestination):
(WebDownload::didReceiveResponse):
- 11:41 AM Changeset in webkit [187435] by
-
- 6 edits in trunk/Source
Source/WebCore:
Remove duplicate vectors inside of UserMediaRequest
https://bugs.webkit.org/show_bug.cgi?id=147321
<rdar://problem/22011290>
Patch by Matthew Daiter <mdaiter@apple.com> on 2015-07-27
Reviewed by Eric Carlson.
- Modules/mediastream/UserMediaRequest.h:
(WebCore::UserMediaRequest::deviceUIDsVideo): Deleted.
(WebCore::UserMediaRequest::deviceUIDsAudio): Deleted.
Source/WebKit2:
Renamed duplicate vectors inside UserMediaPermissionRequestProxy
https://bugs.webkit.org/show_bug.cgi?id=147321
<rdar://problem/22011290>
Patch by Matthew Daiter <mdaiter@apple.com> on 2015-07-27
Reviewed by Eric Carlson.
- UIProcess/UserMediaPermissionRequestProxy.cpp:
(WebKit::UserMediaPermissionRequestProxy::UserMediaPermissionRequestProxy):
- UIProcess/UserMediaPermissionRequestProxy.h:
(WebKit::UserMediaPermissionRequestProxy::videoDeviceUIDs):
(WebKit::UserMediaPermissionRequestProxy::audioDeviceUIDs):
(WebKit::UserMediaPermissionRequestProxy::deviceUIDsVideo): Deleted.
(WebKit::UserMediaPermissionRequestProxy::deviceUIDsAudio): Deleted.
- WebProcess/MediaStream/UserMediaPermissionRequestManager.cpp:
(WebKit::UserMediaPermissionRequestManager::startRequest):
- 1:19 AM Changeset in webkit [187434] by
-
- 2 edits in trunk/Source/WebInspectorUI
Web Inspector: " = $0" in Elements panel can cause a jerk by pushing nodes below it
https://bugs.webkit.org/show_bug.cgi?id=147318
Reviewed by Timothy Hatcher.
- UserInterface/Views/DOMTreeOutline.css:
(.dom-tree-outline li.selected > span::after):
- 1:14 AM Changeset in webkit [187433] by
-
- 5 edits in trunk/LayoutTests
[Streams API] Strategy tests need sync with the reference
https://bugs.webkit.org/show_bug.cgi?id=147267
Reviewed by Sam Weinig.
The reference accepted a two new tests for the constructor and another for the size method for both strategies
that we need to bring to WebKit to ensure compliance. Expectations updated accordingly.
- streams/reference-implementation/byte-length-queuing-strategy-expected.txt:
- streams/reference-implementation/byte-length-queuing-strategy.html:
- streams/reference-implementation/count-queuing-strategy-expected.txt:
- streams/reference-implementation/count-queuing-strategy.html:
- 1:07 AM Changeset in webkit [187432] by
-
- 8 edits in trunk/Source
[GTK] Pass a GstInstallPluginsContext to gst_install_plugins_async
https://bugs.webkit.org/show_bug.cgi?id=147103
Reviewed by Philippe Normand.
Source/WebCore:
- platform/graphics/gstreamer/GUniquePtrGStreamer.h: Allow to use
GUniquePtr with GstInstallPluginsContext.
Source/WebKit2:
This allows PackageKit to properly position the window and make it
transient to the web view window.
- UIProcess/API/gtk/PageClientImpl.cpp:
(WebKit::PageClientImpl::setCursor): Disambiguate Cursor now that
we include gtkx.h.
(WebKit::PageClientImpl::createGstInstallPluginsContext): Create a
new GstInstallPluginsContext and set the web view window XID when
running on X11.
- UIProcess/API/gtk/PageClientImpl.h:
- UIProcess/PageClient.h:
- UIProcess/efl/WebViewEfl.h:
- UIProcess/gstreamer/WebPageProxyGStreamer.cpp:
(WebKit::WebPageProxy::requestInstallMissingMediaPlugins): Use
PageClient::createGstInstallPluginsContext() to create a new
GstInstallPluginsContext and pass it to gst_install_plugins_async().
- 12:10 AM Changeset in webkit [187431] by
-
- 4 edits in branches/safari-601.1-branch/LayoutTests
Merged r187395. rdar://problem/21964842
- 12:10 AM Changeset in webkit [187430] by
-
- 3 edits2 copies in branches/safari-601.1-branch
Merged r187352. rdar://problem/21964842
- 12:09 AM Changeset in webkit [187429] by
-
- 2 edits in branches/safari-601.1-branch/Source/WebCore
Merged r187393. rdar://problem/21905839
- 12:08 AM Changeset in webkit [187428] by
-
- 4 edits in branches/safari-601.1-branch/Source/WebCore
Merged r187392. rdar://problem/21905839
- 12:07 AM Changeset in webkit [187427] by
-
- 3 edits in branches/safari-601.1-branch/Source/WebCore
Merged r187391. rdar://problem/21905839
- 12:06 AM Changeset in webkit [187426] by
-
- 2 edits in branches/safari-601.1-branch/Source/WebCore
Merged r187386. rdar://problem/21992826
- 12:05 AM Changeset in webkit [187425] by
-
- 18 edits2 copies in branches/safari-601.1-branch
Merged r187371. rdar://problem/19918933
- 12:04 AM Changeset in webkit [187424] by
-
- 6 edits in branches/safari-601.1-branch/Source/WebCore
Merged r187367. rdar://problem/21394410
- 12:03 AM Changeset in webkit [187423] by
-
- 3 edits in branches/safari-601.1-branch/Source/WebCore
Merged r187366. rdar://problem/21429111
- 12:02 AM Changeset in webkit [187422] by
-
- 2 edits in branches/safari-601.1-branch/Source/WebKit2
Merged r187364. rdar://problem/21423353
- 12:01 AM Changeset in webkit [187421] by
-
- 2 edits in branches/safari-601.1-branch/Source/WebCore
Merged r187358. rdar://problem/20689512
Jul 26, 2015:
- 11:59 PM Changeset in webkit [187420] by
-
- 2 edits in branches/safari-601.1-branch/Source/WebKit2
Merged r187345. rdar://problem/21409047
- 11:59 PM Changeset in webkit [187419] by
-
- 3 edits in branches/safari-601.1-branch/Source/WebKit2
Merged r187178.
- 11:58 PM Changeset in webkit [187418] by
-
- 3 edits in branches/safari-601.1-branch/Source/bmalloc
Merged r187270. rdar://problem/21971820
- 11:57 PM Changeset in webkit [187417] by
-
- 4 edits1 copy in branches/safari-601.1-branch
Merged r187244. rdar://problem/21966931
- 11:34 PM Changeset in webkit [187416] by
-
- 3 edits in trunk/Source/WebKit2
[ThreadedCompositor] Unreviewed build fix after r186059
https://bugs.webkit.org/show_bug.cgi?id=147315
- Shared/CoordinatedGraphics/threadedcompositor/ThreadedCompositor.cpp:
(WebKit::ThreadedCompositor::glContext):
Explicitly cast u_int64 to GLNativeWindowType.
- WebProcess/WebPage/CoordinatedGraphics/ThreadedCoordinatedLayerTreeHost.h:
(WebKit::ThreadedCoordinatedLayerTreeHost::create):
Update declaration to match with modified definition.
- 10:54 PM Changeset in webkit [187415] by
-
- 5 edits in branches/safari-600.1.4.17-branch/Source
Versioning.
- 10:53 PM Changeset in webkit [187414] by
-
- 5 edits in branches/safari-600.8-branch/Source
Versioning.
- 10:41 PM Changeset in webkit [187413] by
-
- 5 edits in branches/safari-601.1-branch/Source
Versioning.
- 10:08 PM Changeset in webkit [187412] by
-
- 2 edits in trunk/Source/WebKit2
[WK2][iOS] WebContent process main thread should have fixed priority
https://bugs.webkit.org/show_bug.cgi?id=147313
<rdar://problem/22003112>
Reviewed by Darin Adler.
WebContent process main thread should have fixed priority on iOS 9.
Priority decay regresses PLT and fixing the main thread's priority
gives up a ~3% progression on warm PLT (tested on iPhone 5s).
- WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::platformInitializeWebProcess):
- 9:52 PM Changeset in webkit [187411] by
-
- 2 edits in trunk/Source/JavaScriptCore
Add missing builtin files to the JavaScriptCore Xcode project
https://bugs.webkit.org/show_bug.cgi?id=147312
Patch by Sam Weinig <sam@webkit.org> on 2015-07-26
Reviewed by Darin Adler.
- JavaScriptCore.xcodeproj/project.pbxproj:
Add missing files.
- 2:17 PM Changeset in webkit [187410] by
-
- 2 edits1 add in trunk/Source/JavaScriptCore
[ES6] Implement Reflect.isExtensible
https://bugs.webkit.org/show_bug.cgi?id=147308
Reviewed by Sam Weinig.
This patch implements Reflect.isExtensible.
It is similar to Object.isExtensible.
The difference is that it raises an error if the first argument is not an object.
- runtime/ReflectObject.cpp:
(JSC::reflectObjectIsExtensible):
- tests/stress/reflect-is-extensible.js: Added.
(shouldBe):
(shouldThrow):
- 2:10 PM Changeset in webkit [187409] by
-
- 2 edits in trunk/Source/JavaScriptCore
Unreviewed, fix the debug build due to touching the non-declared variable in ASSERT
https://bugs.webkit.org/show_bug.cgi?id=147307
- runtime/ObjectConstructor.cpp:
(JSC::ownPropertyKeys):
- 2:02 PM Changeset in webkit [187408] by
-
- 4 edits1 add in trunk/Source/JavaScriptCore
[ES6] Implement Reflect.ownKeys
https://bugs.webkit.org/show_bug.cgi?id=147307
Reviewed by Sam Weinig.
This patch implements Reflect.ownKeys.
In this patch, we refactor the existing code to list up own keys in the object.
Such code is used by Object.getOwnPropertyNames, Object.getOwnPropertyKeys, Object.keys and @ownEnumerableKeys.
We factor out the listing up own keys as ownPropertyKeys function and also use it in Reflect.ownKeys.
- runtime/ObjectConstructor.cpp:
(JSC::objectConstructorGetOwnPropertyNames):
(JSC::objectConstructorGetOwnPropertySymbols):
(JSC::objectConstructorKeys):
(JSC::ownEnumerablePropertyKeys):
(JSC::ownPropertyKeys):
- runtime/ObjectConstructor.h:
- runtime/ReflectObject.cpp:
(JSC::reflectObjectOwnKeys):
- tests/stress/reflect-own-keys.js: Added.
(shouldBe):
(shouldThrow):
(shouldBeArray):
- 1:52 PM Changeset in webkit [187407] by
-
- 4 edits1 add in trunk/Source/JavaScriptCore
[ES6] Implement Reflect.apply
https://bugs.webkit.org/show_bug.cgi?id=147306
Reviewed by Sam Weinig.
Implement Reflect.apply.
The large part of this can be implemented by the @apply builtin annotation.
The only thing which is different from the Funciton.prototype.apply is the third parameter,
"argumentsList" is needed to be an object.
- builtins/ReflectObject.js:
(apply):
(deleteProperty):
- runtime/ReflectObject.cpp:
- tests/stress/reflect-apply.js: Added.
(shouldBe):
(shouldThrow):
(get shouldThrow):
(.get shouldThrow):
(get var.array.get length):
(get var.array.get 0):
(.get var):
- tests/stress/reflect-delete-property.js:
- 12:14 PM Changeset in webkit [187406] by
-
- 3 edits in trunk/Source/WebKit2
Suppressed an Xcode 7 warning about including a nib in an iOS build product.
- Configurations/BaseLegacyProcess.xcconfig: Exclude xib files on iOS.
- Configurations/BaseXPCService.xcconfig: Ditto.
