Timeline
Jul 8, 2018:
- 10:57 PM Changeset in webkit [233636] by
-
- 4 edits in trunk/Source/WebCore
Optimize packing of RenderSVGViewportContainer
https://bugs.webkit.org/show_bug.cgi?id=187458
Reviewed by Zalan Bujtas.
Shrink RenderSVGViewportContainer from 296 bytes to 280 bytes, saving about 6KB on nytimes.com.
- rendering/svg/RenderSVGContainer.cpp:
(WebCore::RenderSVGContainer::RenderSVGContainer):
- rendering/svg/RenderSVGContainer.h:
- rendering/svg/RenderSVGViewportContainer.h:
- 10:35 PM Changeset in webkit [233635] by
-
- 2 edits in trunk/Source/WebCore
Shrink RenderTableSection
https://bugs.webkit.org/show_bug.cgi?id=187457
Reviewed by Zalan Bujtas.
Shrink RenderTableSection from 344 to 336 bytes, saving about 3.6KB on nytimes.com.
- rendering/RenderTableSection.h:
- 10:09 PM Changeset in webkit [233634] by
-
- 9 edits in trunk/Source/WebCore
Optimize packing of RootInlineBox
https://bugs.webkit.org/show_bug.cgi?id=187430
Reviewed by Zalan Bujtas.
In InlineBox, move the m_logicalWidth float up next to the m_expansion float with m_topLeft next; this
avoids padding of 4 bytes after this float.
In InlineFlowBox, move the bitfields before the pointers so they can snug up into the
4 bytes after m_expansion in the base class.
The comment about m_lineBreakPos's padding in RootInlineBox is wrong; just move it to the end
to avoid padding before the m_lineBreakObj pointer.
Make m_logicalWidth private and have derived classes use the accessor.
Make EllipsisBox 4 bytes smaller too.
- rendering/EllipsisBox.cpp:
(WebCore::EllipsisBox::EllipsisBox):
(WebCore::EllipsisBox::paintMarkupBox):
(WebCore::EllipsisBox::nodeAtPoint):
- rendering/EllipsisBox.h:
- rendering/InlineBox.h:
(WebCore::InlineBox::InlineBox):
- rendering/InlineFlowBox.cpp:
- rendering/InlineFlowBox.h:
(WebCore::InlineFlowBox::InlineFlowBox):
(WebCore::InlineFlowBox::frameRectIncludingLineHeight const):
- rendering/InlineTextBox.cpp:
(WebCore::InlineTextBox::localSelectionRect const):
(WebCore::InlineTextBox::paint):
(WebCore::InlineTextBox::paintPlatformDocumentMarker):
(WebCore::InlineTextBox::paintMarkedTextBackground):
(WebCore::InlineTextBox::paintCompositionUnderline const):
- rendering/RootInlineBox.cpp:
(WebCore::RootInlineBox::RootInlineBox):
- rendering/RootInlineBox.h:
- 9:37 PM Changeset in webkit [233633] by
-
- 3 edits in trunk/Source/JavaScriptCore
Unreviewed, build fix for debug builds after r233630
https://bugs.webkit.org/show_bug.cgi?id=187441
- jit/JIT.cpp:
(JSC::JIT::frameRegisterCountFor):
- llint/LLIntEntrypoint.cpp:
(JSC::LLInt::frameRegisterCountFor):
- 9:10 PM Changeset in webkit [233632] by
-
- 7 edits in trunk
[Web Animations] A number of tests report an incorrect computed offset
https://bugs.webkit.org/show_bug.cgi?id=187410
<rdar://problem/41905790>
Patch by Antoine Quint <Antoine Quint> on 2018-07-08
Reviewed by Dean Jackson.
LayoutTests/imported/w3c:
Mark 16 new WPT progressions.
- web-platform-tests/web-animations/interfaces/Animatable/animate-expected.txt:
- web-platform-tests/web-animations/interfaces/KeyframeEffect/constructor-expected.txt:
- web-platform-tests/web-animations/interfaces/KeyframeEffect/processing-a-keyframes-argument-001-expected.txt:
- web-platform-tests/web-animations/interfaces/KeyframeEffect/setKeyframes-expected.txt:
Source/WebCore:
While we would correctly avoid computing missing offsets when processing the first keyframe following the last
keyframes with a specified offset, we were forgetting to update the index of the last keyframe with a specified
offset which meant we would accidentally override a specified offset with an automically-computed one.
- animation/KeyframeEffectReadOnly.cpp:
(WebCore::computeMissingKeyframeOffsets):
- 4:18 PM Changeset in webkit [233631] by
-
- 2 edits in trunk/Tools
run-benchmark should run ARES-6 1.0.1
https://bugs.webkit.org/show_bug.cgi?id=187452
Reviewed by Sam Weinig.
We forgot to update this when we released 1.0.1!
- Scripts/webkitpy/benchmark_runner/data/plans/ares6.plan:
- 10:59 AM Changeset in webkit [233630] by
-
- 20 edits in trunk/Source/JavaScriptCore
[JSC] Optimize layout of CodeBlock to reduce padding
https://bugs.webkit.org/show_bug.cgi?id=187441
Reviewed by Mark Lam.
Arrange the order of members to reduce the size of CodeBlock from 552 to 544.
We also make SourceCodeRepresentation 1 byte since CodeBlock has a vector of this,
Vector<SourceCodeRepresentation> m_constantsSourceCodeRepresentation.
We also move m_numCalleeLocals and m_numVars from
publictoprivatein CodeBlock.
- bytecode/BytecodeDumper.cpp:
(JSC::BytecodeDumper<Block>::dumpBlock):
- bytecode/BytecodeUseDef.h:
(JSC::computeDefsForBytecodeOffset):
- bytecode/CodeBlock.cpp:
(JSC::CodeBlock::CodeBlock):
- bytecode/CodeBlock.h:
(JSC::CodeBlock::numVars const):
- bytecode/UnlinkedCodeBlock.h:
(JSC::UnlinkedCodeBlock::numVars const):
- dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::ByteCodeParser):
(JSC::DFG::ByteCodeParser::flushForTerminalImpl):
(JSC::DFG::ByteCodeParser::handleRecursiveTailCall):
(JSC::DFG::ByteCodeParser::inlineCall):
(JSC::DFG::ByteCodeParser::handleGetById):
(JSC::DFG::ByteCodeParser::handlePutById):
(JSC::DFG::ByteCodeParser::parseBlock):
- dfg/DFGGraph.h:
(JSC::DFG::Graph::forAllLocalsLiveInBytecode):
- dfg/DFGOSREntrypointCreationPhase.cpp:
(JSC::DFG::OSREntrypointCreationPhase::run):
- dfg/DFGVariableEventStream.cpp:
(JSC::DFG::VariableEventStream::reconstruct const):
- ftl/FTLOSREntry.cpp:
(JSC::FTL::prepareOSREntry):
- ftl/FTLState.cpp:
(JSC::FTL::State::State):
- interpreter/Interpreter.cpp:
(JSC::Interpreter::dumpRegisters):
- jit/JIT.cpp:
(JSC::JIT::frameRegisterCountFor):
- jit/JITOpcodes.cpp:
(JSC::JIT::emit_op_enter):
- jit/JITOpcodes32_64.cpp:
(JSC::JIT::emit_op_enter):
- jit/JITOperations.cpp:
- llint/LLIntEntrypoint.cpp:
(JSC::LLInt::frameRegisterCountFor):
- llint/LLIntSlowPaths.cpp:
(JSC::LLInt::traceFunctionPrologue):
(JSC::LLInt::LLINT_SLOW_PATH_DECL):
- runtime/JSCJSValue.h:
- 10:39 AM Changeset in webkit [233629] by
-
- 7 edits in trunk/Source
[JSC] Optimize padding of UnlinkedCodeBlock to shrink
https://bugs.webkit.org/show_bug.cgi?id=187448
Reviewed by Saam Barati.
Source/JavaScriptCore:
We optimize the size of CodeType and TriState. And we arrange the layout of UnlinkedCodeBlock.
These optimizations reduce the size of UnlinkedCodeBlock from 304 to 288.
- bytecode/CodeType.h:
- bytecode/UnlinkedCodeBlock.cpp:
(JSC::UnlinkedCodeBlock::UnlinkedCodeBlock):
- bytecode/UnlinkedCodeBlock.h:
(JSC::UnlinkedCodeBlock::codeType const):
(JSC::UnlinkedCodeBlock::didOptimize const):
(JSC::UnlinkedCodeBlock::setDidOptimize):
- bytecode/VirtualRegister.h:
Source/WTF:
- wtf/TriState.h:
- 9:58 AM Changeset in webkit [233628] by
-
- 3 edits5 adds in trunk
DOMMatrix.invertSelf() returns garbage values for a non-invertible matrix
<https://webkit.org/b/187446>
<rdar://problem/41853187>
Reviewed by Daniel Bates.
Source/WebCore:
Test: http/wpt/css/geometry/DOMMatrix-invertSelf.html
- css/DOMMatrix.cpp:
(WebCore::DOMMatrix::invertSelf): Add missing
else, and
restructure to use positive logic.
LayoutTests:
- http/wpt/css/geometry/DOMMatrix-invertSelf-expected.txt: Add.
- http/wpt/css/geometry/DOMMatrix-invertSelf.html: Add.
- http/wpt/css/geometry/support/dommatrix-test-util.js: Copy from LayoutTests/imported/w3c/web-platform-tests/css/geometry/support/dommatrix-test-util.js.
- 9:41 AM Changeset in webkit [233627] by
-
- 4 edits in trunk/Source/JavaScriptCore
[JSC] Optimize padding of InferredTypeTable by using cellLock
https://bugs.webkit.org/show_bug.cgi?id=187447
Reviewed by Mark Lam.
Use cellLock() in InferredTypeTable to guard changes of internal structures.
This is the same usage to SparseArrayValueMap. By using cellLock(), we can
reduce the size of InferredTypeTable from 40 to 32.
- runtime/InferredTypeTable.cpp:
(JSC::InferredTypeTable::visitChildren):
(JSC::InferredTypeTable::get):
(JSC::InferredTypeTable::willStoreValue):
(JSC::InferredTypeTable::makeTop):
- runtime/InferredTypeTable.h:
Using enum class and using. And remove
isEmpty()since it is not used.
- runtime/Structure.h:
- 5:29 AM Changeset in webkit [233626] by
-
- 3 edits in trunk/Source/JavaScriptCore
[JSC] Optimize layout of SourceProvider to reduce padding
https://bugs.webkit.org/show_bug.cgi?id=187440
Reviewed by Mark Lam.
Arrange members of SourceProvider to reduce the size from 80 to 72.
- parser/SourceProvider.cpp:
(JSC::SourceProvider::SourceProvider):
- parser/SourceProvider.h:
- 2:54 AM Changeset in webkit [233625] by
-
- 2 edits in trunk/Source/JavaScriptCore
PropertyTable::skipDeletedEntries() should guard against iterating past the table end.
https://bugs.webkit.org/show_bug.cgi?id=187444
<rdar://problem/41282849>
Reviewed by Saam Barati.
PropertyTable supports C++ iteration by offering begin() and end() methods, and
an iterator class. The begin() methods and the iterator operator++() method uses
PropertyTable::skipDeletedEntries() to skip over deleted entries in the table.
However, PropertyTable::skipDeletedEntries() does not prevent the iteration
pointer from being incremented past the end of the table. As a result, we can
iterate past the end of the table. Note that the C++ iteration protocol tests
for the iterator not being equal to the end() value. It does not do a <= test.
If the iterator ever shoots past end, the loop will effectively not terminate.
This issue can manifest if and only if the last entry in the table is a deleted
one, and the key field of the PropertyMapEntry shaped space at the end of the
table (the one beyond the last) contains a 1 (i.e. PROPERTY_MAP_DELETED_ENTRY_KEY)
value.
No test because manifesting this issue requires uncontrollable happenstance where
memory just beyond the end of the table looks like a deleted entry.
- runtime/PropertyMapHashTable.h:
(JSC::PropertyTable::begin):
(JSC::PropertyTable::end):
(JSC::PropertyTable::begin const):
(JSC::PropertyTable::end const):
(JSC::PropertyTable::skipDeletedEntries):
Jul 7, 2018:
- 3:29 PM Changeset in webkit [233624] by
-
- 2 edits in trunk/Tools
WPT importer should create dummy HTML files for *.window.js script files
https://bugs.webkit.org/show_bug.cgi?id=187426
Reviewed by Ryosuke Niwa.
Some WPT tests use *.window.js tests, which should be accessed as *.window.html.
Do the same strategy as done for worker.js/any.js files.
- Scripts/webkitpy/w3c/test_importer.py:
(TestImporter.write_html_files_for_templated_js_tests):
- 2:05 PM Changeset in webkit [233623] by
-
- 13 edits1 add in trunk
Introduce a layout milestone to track when the document contains a large number of rendered characters
https://bugs.webkit.org/show_bug.cgi?id=187412
<rdar://problem/41744338>
Reviewed by Ryosuke Niwa.
Source/WebCore:
Implements a new layout milestone:
DidRenderSignificantAmountOfText. This is similar to the existing
DidFirstVisuallyNonEmptyLayoutmilestone, but with a few important additional constraints:
- The minimum threshold of rendered characters is much larger (3000 instead of 200).
- The mean length of a text node must be large (at least 50).
This is a (computationally) lightweight heuristic intended to estimate the likelihood that a page is capable of
being presented in Reader mode. In many article-like pages, the average length of a text node is significantly
longer than other types of pages; thus, on pages where the average length of a text node is very large *and*
there is a large quantity of text, we can use this as a cue to opportunistically detect and enter Safari reader
mode on watchOS.
Test: RenderingProgressTests.DidRenderSignificantAmountOfText
- page/FrameView.cpp:
(WebCore::FrameView::reset):
Clear out new state members that keep track of whether the significant rendered text milestone has been met.
(WebCore::elementOverflowRectIsLargerThanThreshold):
Factor out logic to grab the overflow height of an element into a separate helper function.
(WebCore::FrameView::qualifiesAsVisuallyNonEmpty const):
(WebCore::FrameView::updateSignificantRenderedTextMilestoneIfNeeded):
Determines whether to fire this new milestone after the next layout pass is finished. The logic here is similar
to the logic in qualifiesAsVisuallyNonEmpty, requiring that there are not still pending stylesheets in the head
and that the document height is at least a minimum threshold. However, unlike the first visually non-empty
layout, this milestone is never guaranteed to fire.
(WebCore::FrameView::fireLayoutRelatedMilestonesIfNeeded):
- page/FrameView.h:
(WebCore::FrameView::incrementVisuallyNonEmptyCharacterCount):
- page/LayoutMilestones.h:
Source/WebKit:
Add plumbing for the new significant rendered text layout milestone in the modern WebKit client layer. See
WebCore/ChangeLog for more details.
- Shared/API/Cocoa/_WKRenderingProgressEvents.h:
- Shared/API/Cocoa/_WKRenderingProgressEventsInternal.h:
(renderingProgressEvents):
- Shared/API/c/WKPageLoadTypes.h:
- Shared/API/c/WKSharedAPICast.h:
(WebKit::toWKLayoutMilestones):
(WebKit::toLayoutMilestones):
- UIProcess/API/Cocoa/WKWebView.mm:
(layoutMilestones):
Tools:
Add a new API test to check that we observe the new rendering milestone in an article-like document.
- TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
- TestWebKitAPI/Tests/WebKitCocoa/significant-text-milestone.html: Added.
- TestWebKitAPI/Tests/ios/RenderingProgressTests.mm:
(TEST):
- 1:46 PM Changeset in webkit [233622] by
-
- 2 edits in trunk/Source/JavaScriptCore
[JSC] Optimize layout of SymbolTable to reduce padding
https://bugs.webkit.org/show_bug.cgi?id=187437
Reviewed by Mark Lam.
Arrange the layout of SymbolTable to reduce the size from 88 to 72.
- runtime/SymbolTable.h:
- 1:29 PM Changeset in webkit [233621] by
-
- 5 edits in trunk/Source/JavaScriptCore
[JSC] Optimize layout of RegExp to reduce padding
https://bugs.webkit.org/show_bug.cgi?id=187438
Reviewed by Mark Lam.
Reduce the size of RegExp from 168 to 144.
- runtime/RegExp.cpp:
(JSC::RegExp::RegExp):
- runtime/RegExp.h:
- runtime/RegExpKey.h:
- yarr/YarrErrorCode.h:
- 1:29 PM Changeset in webkit [233620] by
-
- 2 edits in trunk/Source/JavaScriptCore
[JSC] Optimize layout of ValueProfile to reduce padding
https://bugs.webkit.org/show_bug.cgi?id=187439
Reviewed by Mark Lam.
Reduce the size of ValueProfile from 40 to 32 by reordering members.
- bytecode/ValueProfile.h:
(JSC::ValueProfileBase::ValueProfileBase):
- 11:40 AM Changeset in webkit [233619] by
-
- 7 edits in branches/safari-606.1.24-branch/Source
Versioning.
- 11:39 AM Changeset in webkit [233618] by
-
- 1 copy in tags/Safari-606.1.24.4
Tag Safari-606.1.24.4.
- 11:38 AM Changeset in webkit [233617] by
-
- 2 edits in branches/safari-606.1.24-branch/Source/WebKit
Cherry-pick r233616. rdar://problem/41931915
REGRESSION (r233561): MobileMail crashes when replying to an email
https://bugs.webkit.org/show_bug.cgi?id=187436
rdar://problem/41931915
Reviewed by Alan Bujtas.
r233561 added a RELEASE_ASSERT(isMainThread()) in this function. This is fine most of the time,
until Mail spawns a WebThread for mail compose, at which point the meaning of isMainThread changes
and the RELEASE_ASSERT fires.
For now, just remove the RELEASE_ASSERT.
- UIProcess/API/Cocoa/WKWebView.mm: (-[WKWebView _scheduleVisibleContentRectUpdateAfterScrollInView:]):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233616 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 10:50 AM Changeset in webkit [233616] by
-
- 2 edits in trunk/Source/WebKit
REGRESSION (r233561): MobileMail crashes when replying to an email
https://bugs.webkit.org/show_bug.cgi?id=187436
rdar://problem/41931915
Reviewed by Alan Bujtas.
r233561 added a RELEASE_ASSERT(isMainThread()) in this function. This is fine most of the time,
until Mail spawns a WebThread for mail compose, at which point the meaning of isMainThread changes
and the RELEASE_ASSERT fires.
For now, just remove the RELEASE_ASSERT.
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _scheduleVisibleContentRectUpdateAfterScrollInView:]):
- 9:20 AM Changeset in webkit [233615] by
-
- 2 edits in trunk/Source/WTF
iOS port should define HAVE_RUNLOOP_TIMER
<https://webkit.org/b/187370>
Reviewed by Simon Fraser.
- wtf/Platform.h:
(HAVE_DTRACE): Remove unused macro since r200568.
(HAVE_RUNLOOP_TIMER): Define for PLATFORM(COCOA), not just
PLATFORM(MAC). Alphabetize macros.
- 8:52 AM Changeset in webkit [233614] by
-
- 5 edits1 copy3 adds in trunk/Tools
dump-class-layout reports wrong padding in many cases
https://bugs.webkit.org/show_bug.cgi?id=185801
Reviewed by Dan Bates.
Enhance dump-class-layout, fixing bugs and adding unit tests. This also includes
the patch from bug 187141.
Move the code that uses lldb to fetch the class layout to its own module,
lldb_dump_class_layout.py, so it can be unit tested. Change this code
to build up a data structure for the class. This is necessary since correct
padding computation is easier with second traversal over the data structure.
Try to deal with the empty base class optimization, which is necessary to report
correct padding in classes using std::unique_ptr, and correctly handle virtual base classes.
The ClassLayoutBase class knows how to test for equality and generate a string
or string list representation of itself. The derived ClassLayout class knows
how to build up the class layout using the lldb Python bindings.
Finally we wrap the lldb debugger instance in LLDBDebuggerInstance so that we don't
have to make a new one for each unit test. The tests have to run serially.
Add to lldbWebKitTester a .cpp file that contains a set of C++ classes to unit-test
dump-class-layout.
- Scripts/dump-class-layout:
(main):
- lldb/dump_class_layout_unittest.py:
(destroy_cached_debug_session):
(TestDumpClassLayout):
(TestDumpClassLayout.setUpClass):
(TestDumpClassLayout.setUp):
(TestDumpClassLayout.serial_test_BasicClassLayout):
(serial_test_PaddingBetweenClassMembers):
(serial_test_BoolPaddingClass):
(serial_test_ClassWithEmptyClassMembers):
(serial_test_SimpleVirtualClass):
(serial_test_SimpleVirtualClassWithNonVirtualBase):
(serial_test_InterleavedVirtualNonVirtual):
(serial_test_ClassWithTwoVirtualBaseClasses):
(serial_test_ClassWithClassMembers):
(serial_test_ClassWithBitfields):
(serial_test_ClassWithUniquePtrs):
(serial_test_ClassWithOptionals):
(TestDumpClassLayout.test_BasicClassLayout): Deleted.
- lldb/lldbWebKitTester/DumpClassLayoutTesting.cpp:
(EmptyClass::doStuff):
(VirtualBaseClass::~VirtualBaseClass):
(VirtualBaseClass2::~VirtualBaseClass2):
(SimpleVirtualClass::~SimpleVirtualClass):
(SimpleVirtualClass::doStuff):
(SimpleVirtualClassWithNonVirtualBase::~SimpleVirtualClassWithNonVirtualBase):
(SimpleVirtualClassWithNonVirtualBase::doStuff):
(avoidClassDeadStripping):
- lldb/lldb_dump_class_layout.py:
(ansi_colors):
(ClassLayoutBase):
(ClassLayoutBase.init):
(ClassLayoutBase.ne):
(ClassLayoutBase.eq):
(ClassLayoutBase._to_string_recursive):
(ClassLayoutBase.as_string_list):
(ClassLayoutBase.as_string):
(ClassLayoutBase.dump):
(ClassLayoutExpected):
(ClassLayoutExpected.init):
(ClassLayout):
(ClassLayout.init):
(ClassLayout._has_polymorphic_base_class):
(ClassLayout._parse):
(ClassLayout._probably_has_empty_base_class_optimization):
(ClassLayout._compute_padding_recursive):
(ClassLayout._compute_padding):
(LLDBDebuggerInstance):
(LLDBDebuggerInstance.init):
(LLDBDebuggerInstance.del):
(LLDBDebuggerInstance._get_first_file_architecture):
(LLDBDebuggerInstance.layout_for_classname):
(ClassLayoutDumper): Deleted.
(ClassLayoutDumper.init): Deleted.
(ClassLayoutDumper._get_first_file_architecture): Deleted.
(ClassLayoutDumper.verify_type): Deleted.
(ClassLayoutDumper.verify_type_recursive): Deleted.
(ClassLayoutDumper._class_layout_as_string): Deleted.
(ClassLayoutDumper.dump_to_string): Deleted.
(ClassLayoutDumper.dump): Deleted.
- 12:10 AM Changeset in webkit [233613] by
-
- 2 edits in trunk/Source/WTF
[linux] ASSERT: Using an alternative signal stack is not supported. Consider disabling the concurrent GC.
https://bugs.webkit.org/show_bug.cgi?id=187297
Reviewed by Mark Lam.
This patch relaxes the JSC's limitation: accepting an alternative signal stack mechanism.
- wtf/ThreadingPthreads.cpp:
(WTF::getApproximateStackPointer):
Fix approximate stack pointer function to make it valid.
(WTF::Thread::signalHandlerSuspendResume):
Use StackBounds::contains to check whether the given stack pointer is in range of StackBounds.
If it is out of range, it seems that this stack pointer is pointing an alternative signal stack.
(WTF::Thread::suspend):
Repeatedly retry suspension by using Thread::yield().
(WTF::isOnAlternativeSignalStack): Deleted.
Jul 6, 2018:
- 8:43 PM Changeset in webkit [233612] by
-
- 13 edits in trunk/Source
Semantic colors don't update when accessibility Increase Contrast mode is enabled.
https://bugs.webkit.org/show_bug.cgi?id=187425
rdar://problem/39948240
Reviewed by Tim Horton.
Source/WebCore:
Added a listener for the accessibility change notification to invalidate our color caches.
Removed calls to RenderTheme::singleton().platformColorsDidChange() in Page, since that
ended up calling Page::updateStyleForAllPagesAfterGlobalChangeInEnvironment multiple times.
Instead, changed the functions to use the new instance version instead.
- page/Page.cpp:
(WebCore::Page::updateStyleAfterChangeInEnvironment): Added. Gives Page a direct way to do this work
per instance instead of on all pages (since appearance can be difference per view).
(WebCore::Page::updateStyleForAllPagesAfterGlobalChangeInEnvironment): Call updateStyleAfterChangeInEnvironment.
(WebCore::Page::setUseSystemAppearance): Call updateStyleAfterChangeInEnvironment.
(WebCore::Page::setUseDarkAppearance): Added. Call updateStyleAfterChangeInEnvironment.
- page/Page.h:
(WebCore::Page::setUseDarkAppearance): Moved to the implementation file.
- platform/mac/LocalDefaultSystemAppearance.mm:
(WebCore::LocalDefaultSystemAppearance::LocalDefaultSystemAppearance): Removed recursive check since it was interfering
with the setting of m_usingDarkAppearance and causing the wrong color cache to be used.
(WebCore::LocalDefaultSystemAppearance::~LocalDefaultSystemAppearance): Ditto.
- rendering/RenderThemeMac.mm:
(-[WebCoreRenderThemeNotificationObserver init]): Listen for NSWorkspaceAccessibilityDisplayOptionsDidChangeNotification.
(-[WebCoreRenderThemeNotificationObserver systemColorsDidChange:]): Removed assert since multiple notifications are used now.
Source/WebKit:
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]): Moved call to setUseDarkAppearance to WebViewImpl.
(-[WKWebView _setUseSystemAppearance:]): No need to call setUseDarkAppearance here anymore.
(-[WKWebView viewDidChangeEffectiveAppearance]): Added. This is the proper NSView method to use.
(-[WKWebView _effectiveAppearanceIsDark]): Deleted.
(-[WKWebView effectiveAppearanceDidChange]): Deleted. This method is a deprecated name.
- UIProcess/API/mac/WKView.mm:
(-[WKView viewDidChangeEffectiveAppearance]): Added. This is the proper NSView method to use.
(-[WKView _setUseSystemAppearance:]): No need to call setUseDarkAppearance here anymore.
(-[WKView _effectiveAppearanceIsDark]): Deleted.
(-[WKView effectiveAppearanceDidChange]): Deleted. This method is a deprecated name.
- UIProcess/Cocoa/WebViewImpl.h:
- UIProcess/Cocoa/WebViewImpl.mm:
(WebKit::WebViewImpl::WebViewImpl): Call setUseDarkAppearance before page config is sent in initializeWebPage.
(WebKit::WebViewImpl::effectiveAppearanceDidChange): Added.
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::setUseDarkAppearance): WebCore::Page::setUseDarkAppearance now handles the style changes.
The RenderTheme color caches also don't need cleared with platformColorsDidChange(), since we cache light
and dark colors seperatly in RenderThemeMac.
Source/WebKitLegacy/mac:
- WebView/WebView.mm:
(-[WebView _commonInitializationWithFrameName:groupName:]): Removed call to
_updateDefaultAppearance, which called setUseDarkAppearance. There is already
a call in this method to setUseDarkAppearance.
(-[WebView _setUseSystemAppearance:]): Just call the page, not setUseDarkAppearance.
(-[WebView viewDidChangeEffectiveAppearance]): Added. This is the proper NSView method to use.
(-[WebView _updateDefaultAppearance]): Deleted.
(-[WebView effectiveAppearanceDidChange]): Deleted. This method is a deprecated name.
- 7:56 PM Changeset in webkit [233611] by
-
- 16 edits in trunk/Source/WebCore
Make ReasonForSuspension enum class
https://bugs.webkit.org/show_bug.cgi?id=187405
Reviewed by Zalan Bujtas.
Made ReasonForSuspension an enum class instead of enum and moved out of ActiveDOMObject.
- Modules/geolocation/Geolocation.cpp:
(WebCore::Geolocation::suspend):
- Modules/websockets/WebSocket.cpp:
(WebCore::WebSocket::suspend):
- dom/ActiveDOMObject.h:
- dom/Document.cpp:
(WebCore::Document::didBecomeCurrentDocumentInFrame):
(WebCore::Document::suspendActiveDOMObjects):
(WebCore::Document::resumeActiveDOMObjects):
(WebCore::Document::suspend):
(WebCore::Document::resume):
(WebCore::Document::suspendScheduledTasks):
(WebCore::Document::resumeScheduledTasks):
- dom/Document.h:
- dom/ScriptExecutionContext.cpp:
(WebCore::ScriptExecutionContext::suspendActiveDOMObjects):
(WebCore::ScriptExecutionContext::resumeActiveDOMObjects):
- dom/ScriptExecutionContext.h:
(WebCore::ScriptExecutionContext::reasonForSuspendingActiveDOMObjects const):
- history/CachedFrame.cpp:
(WebCore::CachedFrameBase::restore):
(WebCore::CachedFrame::CachedFrame):
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::suspend):
- html/HTMLSourceElement.cpp:
(WebCore::HTMLSourceElement::suspend):
- inspector/PageScriptDebugServer.cpp:
(WebCore::PageScriptDebugServer::setJavaScriptPaused):
- loader/FrameLoader.cpp:
(WebCore::FrameLoader::commitProvisionalLoad):
- page/Frame.cpp:
(WebCore::Frame::suspendActiveDOMObjectsAndAnimations):
(WebCore::Frame::resumeActiveDOMObjectsAndAnimations):
- page/PageGroupLoadDeferrer.cpp:
(WebCore::PageGroupLoadDeferrer::PageGroupLoadDeferrer):
(WebCore::PageGroupLoadDeferrer::~PageGroupLoadDeferrer):
- xml/XMLHttpRequest.cpp:
(WebCore::XMLHttpRequest::suspend):
- 7:24 PM Changeset in webkit [233610] by
-
- 2 edits in trunk/Tools
[WinCairo] WebKit MiniBrowser crashes when attempting to navigate to certain URLs
https://bugs.webkit.org/show_bug.cgi?id=187167
Reviewed by Alex Christensen.
A null byte was written past the end of the buffer causing the crash.
Some of the heap allocated buffers were also not getting deleted.
- MiniBrowser/win/WebKitBrowserWindow.cpp:
- 5:48 PM Changeset in webkit [233609] by
-
- 3 edits in trunk/Source/ThirdParty/libwebrtc
libWebRTC GetThreadCpuTimeNanos() leaks mach_ports
https://bugs.webkit.org/show_bug.cgi?id=187403
<rdar://problem/41741599>
Reviewed by Simon Fraser.
- Source/webrtc/rtc_base/cpu_time.cc: Call mach_port_deallocate to
to ensure mach_port is deleted.
- libwebrtc.xcodeproj/project.pbxproj: Stop compiling this file since
this is not used except by libwebrtc tests.
- 5:44 PM Changeset in webkit [233608] by
-
- 2 edits in trunk/Source/WebInspectorUI
Web Inspector: Dark Mode: resource search field has white text on white background
https://bugs.webkit.org/show_bug.cgi?id=187423
Reviewed by Timothy Hatcher.
Match the colors of sidebar filter fields.
- UserInterface/Views/DarkMode.css:
(@media (prefers-dark-interface)):
(.find-banner > input[type="search"]):
- 5:34 PM Changeset in webkit [233607] by
-
- 4 edits2 adds in trunk
AX: add a subrole for meter elements on macOS
https://bugs.webkit.org/show_bug.cgi?id=187409
<rdar://problem/41905702>
Reviewed by Chris Fleizach.
Source/WebCore:
Test: accessibility/mac/meter-subrole.html
- accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(-[WebAccessibilityObjectWrapper subrole]):
LayoutTests:
- accessibility/mac/meter-subrole-expected.txt: Added.
- accessibility/mac/meter-subrole.html: Added.
- platform/mac/accessibility/roles-exposed-expected.txt:
- 5:32 PM Changeset in webkit [233606] by
-
- 7 edits in branches/safari-606.1.24-branch/Source
Versioning.
- 5:15 PM Changeset in webkit [233605] by
-
- 4 edits in trunk/Source/WebCore
For <video> element, use video fullscreen instead of element fullscreen.
https://bugs.webkit.org/show_bug.cgi?id=186093
rdar://problem/40144751
Patch by Jeremy Jones <jeremyj@apple.com> on 2018-07-06
Reviewed by Jer Noble.
Some pages use webkitRequestFullscreen directly on the video element, without any custom controls.
This will prevent native fullscreeen presentation mode conrols from being used.
This change makes webkitRequestFullscreen on a video element call webKitSetPresentationMode.
- dom/Element.h:
- html/HTMLVideoElement.cpp:
(WebCore::HTMLVideoElement::webkitRequestFullscreen):
- html/HTMLVideoElement.h:
- 4:47 PM Changeset in webkit [233604] by
-
- 10 edits2 deletes in trunk
WebRTC MediaStreamTrack Enable / Disable causes video delay / lag
https://bugs.webkit.org/show_bug.cgi?id=186889
<rdar://problem/41370285>
Reviewed by Eric Carlson.
Source/WebCore:
Libwebrtc expects a continuous flow of calls for audio data since the API
does not provide any possiblity to give timestamps.
We were optimizing previously when a source is muted so that we would not transmit audio data.
This breaks synchronization between audio and video frames (which are timestamped).
This patch reverts the optimization and instead makes sure to send zeros for silenced audio tracks.
This requires MediaStreamTrackPrivate to send audio data even if disabled,
so that RealtimeOutgoingAudioSource will continue sending zeros at the correct pace.
This also requires WebAudioSourceProviderAVFObjC to exit early if its track is disabled.
Covered by existing tests.
Manual testing shows that synchronization is kept.
- platform/mediastream/MediaStreamTrackPrivate.cpp:
(WebCore::MediaStreamTrackPrivate::audioSamplesAvailable):
- platform/mediastream/RealtimeOutgoingAudioSource.cpp:
(WebCore::RealtimeOutgoingAudioSource::RealtimeOutgoingAudioSource):
(WebCore::RealtimeOutgoingAudioSource::initializeConverter):
(WebCore::RealtimeOutgoingAudioSource::stop):
(WebCore::RealtimeOutgoingAudioSource::sourceMutedChanged):
(WebCore::RealtimeOutgoingAudioSource::sourceEnabledChanged):
(WebCore::RealtimeOutgoingAudioSource::handleMutedIfNeeded): Deleted.
- platform/mediastream/RealtimeOutgoingAudioSource.h:
(WebCore::RealtimeOutgoingAudioSource::pullAudioData):
(WebCore::RealtimeOutgoingAudioSource::isSilenced const):
(WebCore::RealtimeOutgoingAudioSource::sendSilence): Deleted.
- platform/mediastream/mac/RealtimeOutgoingAudioSourceCocoa.cpp:
(WebCore::RealtimeOutgoingAudioSourceCocoa::pullAudioData):
(WebCore::RealtimeOutgoingAudioSourceCocoa::handleMutedIfNeeded): Deleted.
(WebCore::RealtimeOutgoingAudioSourceCocoa::sendSilence): Deleted.
- platform/mediastream/mac/RealtimeOutgoingAudioSourceCocoa.h:
- platform/mediastream/mac/WebAudioSourceProviderAVFObjC.mm:
(WebCore::WebAudioSourceProviderAVFObjC::audioSamplesAvailable):
LayoutTests:
Test is no longer valid since we are now sending 0 bytes for audio tracks.
- webrtc/audio-muted-stats2-expected.txt: Removed.
- webrtc/audio-muted-stats2.html: Removed.
- 4:40 PM Changeset in webkit [233603] by
-
- 2 edits in trunk/LayoutTests
Skip imported/w3c/web-platform-tests/infrastructure/assumptions/html-elements.html on debug.
https://bugs.webkit.org/show_bug.cgi?id=187093
Unreviewed test gardening.
- 4:30 PM Changeset in webkit [233602] by
-
- 2 edits in trunk/Source/WebCore
Make RealtimeOutgoingVideoSource use DestructionThread::Main
https://bugs.webkit.org/show_bug.cgi?id=187402
Reviewed by Chris Dumez.
No change of behavior.
- platform/mediastream/RealtimeOutgoingVideoSource.h:
- 4:26 PM Changeset in webkit [233601] by
-
- 3 edits in trunk/Source/WebKit
Add release assertions to try and catch a possible HashMap corruption in NetworkConnectionToWebProcess
https://bugs.webkit.org/show_bug.cgi?id=187417
Reviewed by Ryosuke Niwa.
Add assertions to make sure we:
- Always use NetworkConnectionToWebProcess::m_networkResourceLoaders from the main thread
- Never use a 0 identifier as key for NetworkConnectionToWebProcess::m_networkResourceLoaders
We see crashes (rdar://problem/39265927) that only seem to make sense if this HashMap was
somehow getting corrupted. Let's try and catch the most common reasons for HashMap corruption.
- NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::didCleanupResourceLoader):
(WebKit::NetworkConnectionToWebProcess::didReceiveMessage):
(WebKit::NetworkConnectionToWebProcess::scheduleResourceLoad):
(WebKit::NetworkConnectionToWebProcess::performSynchronousLoad):
(WebKit::NetworkConnectionToWebProcess::removeLoadIdentifier):
(WebKit::NetworkConnectionToWebProcess::setDefersLoading):
(WebKit::NetworkConnectionToWebProcess::convertMainResourceLoadToDownload):
- WebProcess/Network/WebResourceLoader.cpp:
(WebKit::WebResourceLoader::messageSenderDestinationID):
- 4:11 PM Changeset in webkit [233600] by
-
- 2 edits in trunk/Source/WebKit
WebDriver: hang in Selenium test alerts_tests.py::testShouldHandleAlertOnPageBeforeUnload
https://bugs.webkit.org/show_bug.cgi?id=187418
<rdar://problem/41909520>
Reviewed by Timothy Hatcher.
Per the W3C specification, “User prompts that are spawned from beforeunload event handlers,
are dismissed implicitly upon navigation or close window, regardless of the defined user prompt handler.”
This behavior differs from legacy Selenium behavior where (as in the test) it's expected that
a test can accept or dismiss a beforeunload alert manually.
Prior to this patch, beforeunload alerts hang because Safari uses a nested modal run loop, which does
not process incoming Automation messages, so there was no way for the test to manually dismiss
the alert.
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::runBeforeUnloadConfirmPanel):
Don't show beforeunload alerts for pages being controlled by automation.
- 3:54 PM Changeset in webkit [233599] by
-
- 2 edits in trunk/LayoutTests
This test is flaky. This unreviewed patch is my attempt to make it less so.
- storage/indexeddb/modern/opendatabase-after-storage-crash.html:
- 3:23 PM Changeset in webkit [233598] by
-
- 2 edits in trunk/Source/WebKit
ASSERTION FAILED: contextConnection under WebCore::SWServer::terminateWorkerInternal
https://bugs.webkit.org/show_bug.cgi?id=187348
<rdar://problem/41853270>
Reviewed by Youenn Fablet.
Make sure we mark corresponding SWServerWorkers as terminated when we terminate a
Service Worker context connection once it is no longer needed.
- StorageProcess/StorageProcess.cpp:
(WebKit::StorageProcess::swContextConnectionMayNoLongerBeNeeded):
- 3:22 PM Changeset in webkit [233597] by
-
- 2 edits in trunk/LayoutTests
Layout Test imported/w3c/web-platform-tests/IndexedDB/interfaces.any.worker.html is flaky.
https://bugs.webkit.org/show_bug.cgi?id=187372
Unreviewed test gardening.
Added [ Slow ] to the existing [ Pass Timeout ] expectation.
- platform/mac-wk1/TestExpectations:
- 3:19 PM Changeset in webkit [233596] by
-
- 3 edits in trunk/Source/WebCore
Make HTMLMediaElement::remove*Track take a Ref<>&&
https://bugs.webkit.org/show_bug.cgi?id=187407
Reviewed by Zalan Bujtas.
Make these functions take Ref<>&& since they can delete track objects.
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::removeAudioTrack):
(WebCore::HTMLMediaElement::removeTextTrack):
(WebCore::HTMLMediaElement::removeVideoTrack):
(WebCore::HTMLMediaElement::forgetResourceSpecificTracks):
- html/HTMLMediaElement.h:
- 2:50 PM Changeset in webkit [233595] by
-
- 2 edits in trunk/LayoutTests
Rebaseline accessibility/ios-simulator/strong-password-field.html after r233578.
Unreviewed test gardening.
- accessibility/ios-simulator/strong-password-field-expected.txt:
- 2:28 PM Changeset in webkit [233594] by
-
- 2 edits in trunk/Tools
[WPE][GTK] Build php zlib plugin and enable sockets in flatpak
https://bugs.webkit.org/show_bug.cgi?id=187401
Patch by Thibault Saunier <tsaunier@igalia.com> on 2018-07-06
Reviewed by Philippe Normand.
- flatpak/org.webkit.WebKit.yaml:
- 2:01 PM Changeset in webkit [233593] by
-
- 1 copy in tags/Safari-606.1.24.3
Tag Safari-606.1.24.3.
- 1:38 PM Changeset in webkit [233592] by
-
- 2 edits in branches/safari-606.1.24-branch/Source/WebKit
Cherry-pick r233561. rdar://problem/41878797
Address two possible causes of missing tiles in iOS Safari, and add logging to gather more data about other possible causes
https://bugs.webkit.org/show_bug.cgi?id=187376
rdar://problem/40941118
Reviewed by Tim Horton.
We have continual reports of users experiencing missing tiles in MobileSafari, where loading a page
shows the tiles at the top, but we don't render new tiles as the user scrolls down. This is consistent
with failing to dispatch visible content rect updates via -[WKWebView _updateVisibleContentRects].
This patch addresses two possible (but unlikely) causes. First, it resets _currentlyAdjustingScrollViewInsetsForKeyboard
after a web process crash. Second, it catches exceptions thrown by [webView _updateVisibleContentRects]
and resets _hasScheduledVisibleRectUpdate.
This patch also adds release logging that fires if over 1s has elapsed between scheduling
a visible content rect update and trying to re-schedule, and logging for all reasons that
-_updateVisibleContentRects returns early.
- UIProcess/API/Cocoa/WKWebView.mm: (-[WKWebView _initializeWithConfiguration:]): (-[WKWebView _processDidExit]): (-[WKWebView _addUpdateVisibleContentRectPreCommitHandler]): (-[WKWebView _scheduleVisibleContentRectUpdateAfterScrollInView:]): (-[WKWebView _updateVisibleContentRects]):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233561 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 1:37 PM Changeset in webkit [233591] by
-
- 2 edits in trunk/LayoutTests
Layout Test imported/w3c/web-platform-tests/WebCryptoAPI/generateKey/successes_RSA-PSS.https.any.worker.html is flaky timeout.
https://bugs.webkit.org/show_bug.cgi?id=187408
Unreviewed test gardening.
- platform/mac/TestExpectations:
- 1:37 PM Changeset in webkit [233590] by
-
- 7 edits in branches/safari-606.1.24-branch/Source
Versioning.
- 1:13 PM Changeset in webkit [233589] by
-
- 2 edits in trunk/Source/WebKit
[iOS] WKPDFView should show the Data Detectors sheet when long-pressing a URL that Data Detectors can handle
https://bugs.webkit.org/show_bug.cgi?id=187396
<rdar://problem/41786980>
Reviewed by Dan Bernstein.
If a URL can be presented by Data Detectors, show the Data Detectors sheet instead
of the link sheet. Also implement the optional WKActionSheetAssistantDelegate
method that asks the UI delegate for its data detection context.
- UIProcess/ios/WKPDFView.mm:
(-[WKPDFView _showActionSheetForURL:atLocation:withAnnotationRect:]):
(-[WKPDFView dataDetectionContextForActionSheetAssistant:]):
- 12:39 PM Changeset in webkit [233588] by
-
- 4 edits in trunk
[Web Animations] Make WPT test at interfaces/KeyframeEffect/processing-a-keyframes-argument-002.html pass reliably
https://bugs.webkit.org/show_bug.cgi?id=186502
<rdar://problem/41000229>l
Reviewed by Dean Jackson.
LayoutTests/imported/w3c:
Mark WPT progression.
- web-platform-tests/web-animations/interfaces/KeyframeEffect/processing-a-keyframes-argument-002-expected.txt:
Source/WebCore:
Ensure we only attempt to convert defined objects or we risk to throw errors in valid situations.
- animation/KeyframeEffectReadOnly.cpp:
(WebCore::processKeyframeLikeObject):
- 12:33 PM Changeset in webkit [233587] by
-
- 3 edits in trunk/Source/WebKit
WebResourceLoadStatisticsStore::callHasStorageAccessForFrameHandler() should take in a CompletionHandler
https://bugs.webkit.org/show_bug.cgi?id=187392
Reviewed by Youenn Fablet.
- UIProcess/WebResourceLoadStatisticsStore.cpp:
(WebKit::WebResourceLoadStatisticsStore::callHasStorageAccessForFrameHandler):
- UIProcess/WebResourceLoadStatisticsStore.h:
- 12:24 PM Changeset in webkit [233586] by
-
- 4 edits in trunk/Tools
[ews-build] EWS should run javascriptcore tests
https://bugs.webkit.org/show_bug.cgi?id=187365
Reviewed by David Kilzer.
- BuildSlaveSupport/ews-build/steps.py: Added step to run JavaScriptCoreTests.
- BuildSlaveSupport/ews-build/steps_unittest.py: Added unit-tests.
- BuildSlaveSupport/ews-build/factories.py:
- 11:43 AM Changeset in webkit [233585] by
-
- 8 edits in trunk
[Web Animations] Make WPT test at interfaces/Animation/finish.html pass reliably
https://bugs.webkit.org/show_bug.cgi?id=186496
<rdar://problem/41000179>
Reviewed by Dean Jackson.
LayoutTests/imported/w3c:
Mark a WPT progression.
- web-platform-tests/web-animations/interfaces/Animation/finish-expected.txt:
Source/WebCore:
We used to only resolve animations that had a target element, but animations need not have a target and their
current time should still advance so that their finished promise may resolve. We now maintain a list of animations
without targets and we iterate through them as well as animations with targets in DocumentTimeline::updateAnimations().
- animation/AnimationTimeline.cpp:
(WebCore::AnimationTimeline::addAnimation):
(WebCore::AnimationTimeline::removeAnimation):
(WebCore::AnimationTimeline::animationWasAddedToElement):
(WebCore::AnimationTimeline::animationWasRemovedFromElement):
- animation/AnimationTimeline.h:
(WebCore::AnimationTimeline:: const):
- animation/DocumentTimeline.cpp:
(WebCore::DocumentTimeline::updateAnimations):
- animation/WebAnimation.cpp:
(WebCore::WebAnimation::resolve):
- animation/WebAnimation.h:
- 11:37 AM Changeset in webkit [233584] by
-
- 6 edits in trunk
[Web Animations] Make WPT test at interfaces/Animation/finished.html pass reliably
https://bugs.webkit.org/show_bug.cgi?id=186497
<rdar://problem/41000193>
Reviewed by Dean Jackson.
LayoutTests/imported/w3c:
Mark WPT progression.
- web-platform-tests/web-animations/interfaces/Animation/finished-expected.txt:
Source/WebCore:
We need to ensure the finished state is updated as a result of any timing property changing.
- animation/AnimationEffectReadOnly.cpp:
(WebCore::AnimationEffectReadOnly::timingDidChange):
- animation/WebAnimation.cpp:
(WebCore::WebAnimation::effectTimingPropertiesDidChange):
- animation/WebAnimation.h:
- 11:31 AM Changeset in webkit [233583] by
-
- 9 edits4 adds in trunk
[Web Animations] Using a Web Animation leaks the Document
https://bugs.webkit.org/show_bug.cgi?id=187088
<rdar://problem/41392046>
Reviewed by Darin Adler.
Source/WebCore:
Test: webanimations/leak-document-with-web-animation.html
We need to ensure that any remaining animation is cleared when the DocumentTimeline is detached from its Document.
We rename WebAnimation::prepareAnimationForRemoval() to WebAnimation::remove() since it really actively disassociates
the animation from its timeline.
An earlier version of this patch (r233349) was rolled out due to crashes caught in the ASan configuration. The following
changes were made to make it safe:
- We protect the DocumentAnimationScheduler instance in displayRefreshFired() against code that might run in a
requestAnimationFrame() callback that would trigger the object to be deleted.
- We protect the WebAnimation instance in remove() against setEffectInternal() or setTimelineInternal() potentially
causing the object to be deleted. Similar protections were addede to setEffect() and setTimeline().
- We changed ~DocumentTimeline() to a default implementation to ensure it calls ~DisplayRefreshMonitorClient()
to avoid callbacks after the object has been marked for deletion.
- animation/AnimationTimeline.cpp:
(WebCore::AnimationTimeline::removeAnimationsForElement): We no longer need the call to removeAnimation()
since the new WebAnimation::remove() method will also set the timeline to null which will eventually call
removeAnimation() on the disassociated timeline.
- animation/DeclarativeAnimation.cpp:
(WebCore::DeclarativeAnimation::remove):
(WebCore::DeclarativeAnimation::prepareAnimationForRemoval): Deleted.
- animation/DeclarativeAnimation.h:
- animation/DocumentAnimationScheduler.cpp:
(WebCore::DocumentAnimationScheduler::displayRefreshFired):
- animation/DocumentTimeline.cpp:
(WebCore::DocumentTimeline::detachFromDocument): Call remove() on all known animations.
(WebCore::DocumentTimeline::~DocumentTimeline): Deleted.
- animation/WebAnimation.cpp:
(WebCore::WebAnimation::remove): Set the timeline to null to fully disassociate this animation from its timeline.
(WebCore::WebAnimation::setEffect):
(WebCore::WebAnimation::setEffectInternal):
(WebCore::WebAnimation::setTimeline): Factor the internal timeline-association code out of this JS API method so
that we can call this code without any JS-facing implications.
(WebCore::WebAnimation::setTimelineInternal):
(WebCore::WebAnimation::prepareAnimationForRemoval): Deleted.
- animation/WebAnimation.h:
LayoutTests:
- webanimations/leak-document-with-web-animation-expected.txt: Added.
- webanimations/leak-document-with-web-animation.html: Added.
- webanimations/resources/web-animation-leak-iframe.html: Added.
- 11:25 AM Changeset in webkit [233582] by
-
- 2 edits in trunk/Source/WebKit
[iOS] Assertion failure in WebProcessProxy::allProcesses() (isMainThread())
https://bugs.webkit.org/show_bug.cgi?id=187394
Reviewed by Dan Bernstein.
Use RunLoop::isMain() instead of isMainThread() in the assertion to fix issues for
applications using both WebKit and WebKitLegacy.
- UIProcess/WebProcessProxy.cpp:
- 10:44 AM Changeset in webkit [233581] by
-
- 2 edits in trunk/LayoutTests
Layout Test editing/selection/navigation-clears-editor-state.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=187389
Reviewed by Mark Lam.
- editing/selection/navigation-clears-editor-state.html:
Make the test's internal time out mechanism longer since
it's being hit on the bots.
- 10:29 AM Changeset in webkit [233580] by
-
- 3 edits in trunk/LayoutTests
AX: setValue on text controls should send out key events
https://bugs.webkit.org/show_bug.cgi?id=186690
<rdar://problem/41169985>
Reviewed by Chris Fleizach.
Editor's clearText function will also dispatch the InputEvent. In the test,
we need to make sure each test case finishes handling all the input events
before moving to the next test case.
- accessibility/mac/set-value-editable-dispatch-events-expected.txt:
- accessibility/mac/set-value-editable-dispatch-events.html:
- 10:29 AM Changeset in webkit [233579] by
-
- 2 edits in trunk/LayoutTests
Layout Test imported/w3c/web-platform-tests/2dcontext/imagebitmap/createImageBitmap-origin.sub.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=187393
Unreviewed test gardening.
Patch by Truitt Savell <Truitt Savell> on 2018-07-06
- 9:29 AM Changeset in webkit [233578] by
-
- 39 edits10 deletes in trunk
Remove Strong Confirmation Password button
https://bugs.webkit.org/show_bug.cgi?id=187306
<rdar://problem/41795185>
Reviewed by Sam Weinig.
Remove support for the Strong Confirmation Password button because we never made use of it.
Source/WebCore:
- English.lproj/Localizable.strings:
- accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
(-[WebAccessibilityObjectWrapper _accessibilityIsStrongPasswordField]):
- accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(-[WebAccessibilityObjectWrapper accessibilityAttributeValue:]):
- html/HTMLInputElement.h:
(WebCore::HTMLInputElement::hasAutoFillStrongPasswordButton const):
- html/HTMLTextFormControlElement.h:
- html/TextFieldInputType.cpp:
(WebCore::autoFillButtonTypeToAccessibilityLabel):
(WebCore::autoFillButtonTypeToAutoFillButtonText):
(WebCore::autoFillButtonTypeToAutoFillButtonPseudoClassName):
(WebCore::isAutoFillButtonTypeChanged):
- platform/LocalizedStrings.cpp:
(WebCore::AXAutoFillStrongConfirmationPasswordLabel): Deleted.
- platform/LocalizedStrings.h:
- rendering/RenderTextControlSingleLine.cpp:
(WebCore::RenderTextControlSingleLine::layout):
- testing/Internals.cpp:
(WebCore::toAutoFillButtonType):
(WebCore::toInternalsAutoFillButtonType):
- testing/Internals.h:
- testing/Internals.idl:
Source/WebKit:
- WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInNodeHandle.mm:
(toAutoFillButtonType):
(toWKAutoFillButtonType):
- WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInNodeHandlePrivate.h:
- WebProcess/InjectedBundle/API/c/WKBundleNodeHandle.cpp:
(toAutoFillButtonType):
(toWKAutoFillButtonType):
- WebProcess/InjectedBundle/API/c/WKBundleNodeHandlePrivate.h:
LayoutTests:
- accessibility/ios-simulator/strong-password-field-expected.txt:
- accessibility/ios-simulator/strong-password-field.html:
- fast/forms/auto-fill-button/auto-fill-strong-password-button-when-maxlength-changes-expected.html:
- fast/forms/auto-fill-button/auto-fill-strong-password-button-when-maxlength-changes.html:
- fast/forms/auto-fill-button/auto-fill-strong-password-button-when-minlength-changes-expected.html:
- fast/forms/auto-fill-button/auto-fill-strong-password-button-when-minlength-changes.html:
- fast/forms/auto-fill-button/hide-auto-fill-strong-password-button-when-form-is-reset-expected.html:
- fast/forms/auto-fill-button/hide-auto-fill-strong-password-button-when-form-is-reset.html:
- fast/forms/auto-fill-button/hide-auto-fill-strong-password-button-when-value-changes-expected.html:
- fast/forms/auto-fill-button/hide-auto-fill-strong-password-button-when-value-changes.html:
- fast/forms/auto-fill-button/hide-strong-password-when-field-type-changes-expected.html:
- fast/forms/auto-fill-button/hide-strong-password-when-field-type-changes.html:
- fast/forms/auto-fill-button/input-autofilled-disabled-strong-password-auto-fill-button-expected.html: Removed.
- fast/forms/auto-fill-button/input-autofilled-disabled-strong-password-auto-fill-button.html: Removed.
- fast/forms/auto-fill-button/input-autofilled-readonly-strong-password-auto-fill-button-expected.html: Removed.
- fast/forms/auto-fill-button/input-autofilled-readonly-strong-password-auto-fill-button.html: Removed.
- fast/forms/auto-fill-button/input-disabled-strong-password-and-strong-confirmation-password-auto-fill-buttons-expected.html:
- fast/forms/auto-fill-button/input-disabled-strong-password-and-strong-confirmation-password-auto-fill-buttons.html:
- fast/forms/auto-fill-button/input-readonly-strong-password-and-strong-confirmation-password-auto-fill-buttons-expected.html:
- fast/forms/auto-fill-button/input-readonly-strong-password-and-strong-confirmation-password-auto-fill-buttons.html:
- fast/forms/auto-fill-button/input-strong-confirmation-password-auto-fill-button.html: Removed.
- fast/forms/auto-fill-button/last-auto-fill-button-type-expected.txt:
- fast/forms/auto-fill-button/last-auto-fill-button-type.html:
- platform/gtk/TestExpectations:
- platform/gtk/fast/forms/auto-fill-button/input-strong-confirmation-password-auto-fill-button-expected.txt: Removed.
- platform/ios/fast/forms/auto-fill-button/input-strong-confirmation-password-auto-fill-button-expected.txt: Removed.
- platform/mac/fast/forms/auto-fill-button/input-strong-confirmation-password-auto-fill-button-expected.txt: Removed.
- platform/win/TestExpectations:
- platform/win/fast/forms/auto-fill-button/input-strong-confirmation-password-auto-fill-button-expected.txt: Removed.
- platform/wincairo/fast/forms/auto-fill-button/input-strong-confirmation-password-auto-fill-button-expected.txt: Removed.
- 4:49 AM Changeset in webkit [233577] by
-
- 3 edits in trunk/Source/WTF
WTF's internal std::optional implementation should abort() on bad optional access
https://bugs.webkit.org/show_bug.cgi?id=186536
Patch by Frederic Wang <fwang@igalia.com> on 2018-07-06
Reviewed by Michael Catanzaro.
Currently, some ports built with recent compilers will cause the program to abort when one
tries to access the value of an unset std:optional (i.e. std::nullopt) as specified by C++17.
However, most ports still use WTF's internal std::optional implementation, which does not
verify illegal access. Hence it's not possible for developers working on these ports to
detect issues like bugs #186189, #186535, #186752, #186753, #187139, #187145, #187243 or
#187382.
WTF's version of std::optional was introduced in bug #164199 but it was not possible to
verify the availability of the value inside constexpr member functions because the assert
might involve asm declarations. This commit introduces a new ASSERT_UNDER_CONSTEXPR_CONTEXT
macro (a simplified version of ASSERT that can be used in constexpr context) and uses it in
WTF's implementation of std::optional.
- wtf/Assertions.h: Define ASSERT_UNDER_CONSTEXPR_CONTEXT as a version of ASSERT that can be
used in constexpr context (in particular avoids asm declarations).
- wtf/Optional.h:
(std::optional::operator ->): Add an assert to ensure the optional value is available.
(std::optional::operator *): Ditto.
(std::optional::value const): Ditto.
(std::optional::value): Ditto.
(std::optional<T::value const): Ditto.
- 3:51 AM Changeset in webkit [233576] by
-
- 2 edits in trunk/Tools
[GTK][WPE] REGRESSION(r233111): Test /webkit/WebKitUserContentManager/script-message-received fails
https://bugs.webkit.org/show_bug.cgi?id=187383
Reviewed by Žan Doberšek.
This is actually a problem in the test itself. The problem is that we run js, connect to message-received and
quite the main loop when the message is received, which happens before the run js operation has been
completed. The test finishes early and since r233111 an IPC message is still pending (I guess the run js
callback) and the web page is still protected. Just adding a run loop operation before finishing the test would
fix it. We should wait for the js run to complete, instead of the message.
- TestWebKitAPI/Tests/WebKitGLib/TestWebKitUserContentManager.cpp:
(UserScriptMessageTest::scriptMessageReceived):
(UserScriptMessageTest::waitUntilMessageReceived):
(UserScriptMessageTest::runJavaScriptFinished):
(UserScriptMessageTest::postMessageAndWaitUntilReceived):
- 3:38 AM Changeset in webkit [233575] by
-
- 2 edits in trunk/Source/WebCore
[GTK][WPE] REGRESSION(r233239): proxy settings not applied after r233239
https://bugs.webkit.org/show_bug.cgi?id=187381
Patch by Carlos Garcia Campos <cgarcia@igalia.com> on 2018-07-06
Reviewed by Yusuke Suzuki.
This is because the static variable for proxy settings was replaced by a NeverDestroyed, but the function is not
returning a reference, but a copy. This is causing several unit tests to fail.
- platform/network/soup/SoupNetworkSession.cpp:
(WebCore::proxySettings):
- 3:10 AM Changeset in webkit [233574] by
-
- 2 edits in trunk/Source/WebCore
[Crash] Illegal use of uninitialized std::optional value in WebCore::AnimationBase::updateStateMachine
https://bugs.webkit.org/show_bug.cgi?id=187382
Patch by Frederic Wang <fwang@igalia.com> on 2018-07-06
Reviewed by Carlos Garcia Campos.
WebCore::AnimationBase::updateStateMachine has two potential places where the use of an
uninitialized std:optional value is possible and one of them is hit when using Google drive.
Since that old animation code is going to be removed soon, we just quickly patch this issue
via value_or() so that we can restore the ASSERT added in bug 186536.
No new tests, code is going to be removed soon.
- page/animation/AnimationBase.cpp:
(WebCore::AnimationBase::updateStateMachine): Use value_or(0) to avoid potential crashes.
- 12:59 AM Changeset in webkit [233573] by
-
- 1 copy in tags/Safari-606.1.24.2
Tag Safari-606.1.24.2.
- 12:48 AM Changeset in webkit [233572] by
-
- 4 edits in trunk/Source/WebCore
Rename HTMLTreeBuilder::didCreateCustomOrCallbackElement
https://bugs.webkit.org/show_bug.cgi?id=187317
Patch by Frederic Wang <fwang@igalia.com> on 2018-07-06
Reviewed by Ryosuke Niwa.
This function is called with a new custom element or the fallback HTMLUnknownElement element
so we rename it didCreateCustomOrFallbackElement.
No new tests, behavior unchanged.
- html/parser/HTMLDocumentParser.cpp:
(WebCore::HTMLDocumentParser::runScriptsForPausedTreeBuilder):
- html/parser/HTMLTreeBuilder.cpp:
(WebCore::HTMLTreeBuilder::didCreateCustomOrFallbackElement):
(WebCore::HTMLTreeBuilder::didCreateCustomOrCallbackElement): Deleted.
- html/parser/HTMLTreeBuilder.h:
- 12:37 AM Changeset in webkit [233571] by
-
- 5 edits in trunk/Source/WebCore
REGRESSION(r233496): macOS Sierra hits debug assertions in TrackListBase::suspend
https://bugs.webkit.org/show_bug.cgi?id=187378
<rdar://problem/41878885>
Reviewed by Eric Carlson.
These assertions were wrong since inline debugger, alert, etc... can suspend active DOM objects without
calling canSuspendForDocumentSuspension. Fixed the bug by removing the assertion and instead suspending
and resuming the generic event queue.
- Modules/mediasource/MediaSource.cpp:
(WebCore::MediaSource::suspend):
(WebCore::MediaSource::resume):
- Modules/mediasource/SourceBuffer.cpp:
(WebCore::SourceBuffer::suspend):
(WebCore::SourceBuffer::resume):
- Modules/mediasource/SourceBufferList.cpp:
(WebCore::SourceBufferList::suspend):
(WebCore::SourceBufferList::resume):
- html/track/TrackListBase.cpp:
(WebCore::TrackListBase::suspend):
(WebCore::TrackListBase::resume):
Jul 5, 2018:
- 11:43 PM Changeset in webkit [233570] by
-
- 2 edits in branches/safari-606.1.24-branch/Source/WebCore
Cherry-pick r233557. rdar://problem/41870910
REGRESSION(r233496): Crash in WebCore::VideoTrack::clearClient()
https://bugs.webkit.org/show_bug.cgi?id=187377
Reviewed by Simon Fraser.
Clear m_client of an audio track or a video track before removing it from the list
since TrackListBase::m_inbandTracks may hold the last ref to the track.
- html/HTMLMediaElement.cpp: (WebCore::HTMLMediaElement::removeAudioTrack): (WebCore::HTMLMediaElement::removeVideoTrack):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233557 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 11:43 PM Changeset in webkit [233569] by
-
- 3 edits in branches/safari-606.1.24-branch/Source/WebKit
Cherry-pick r233546. rdar://problem/41867306
Upstream hover gesture implementation
https://bugs.webkit.org/show_bug.cgi?id=187366
Reviewed by Wenson Hsieh.
- UIProcess/ios/WKContentViewInteraction.h:
- UIProcess/ios/WKContentViewInteraction.mm: (-[WKContentView setupInteraction]): (-[WKContentView cleanupInteraction]): (-[WKContentView _removeDefaultGestureRecognizers]): (-[WKContentView _addDefaultGestureRecognizers]): (-[WKContentView _hoverGestureRecognizerChanged:]):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233546 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 11:40 PM Changeset in webkit [233568] by
-
- 7 edits in branches/safari-606.1.24-branch/Source
Versioning.
- 11:31 PM Changeset in webkit [233567] by
-
- 1 copy in branches/safari-606.1.24-branch
New branch.
- 11:28 PM Changeset in webkit [233566] by
-
- 1 edit1 delete in trunk/LayoutTests
Unreviewed GTK gardening. Remove platform expectations for imported/w3c/web-platform-tests/web-animations/interfaces/Animatable/animate-no-browsing-context.html
Since r233512 we are getting the same results as other ports.
- platform/gtk/imported/w3c/web-platform-tests/web-animations/interfaces/Animatable/animate-no-browsing-context-expected.txt: Removed.
- 11:12 PM Changeset in webkit [233565] by
-
- 7 edits in tags/Safari-606.1.24/Source
Versioning.
- 10:57 PM Changeset in webkit [233564] by
-
- 3 edits in trunk/Websites/perf.webkit.org
MeasurementSet should merge last four segments into two if values are identical.
https://bugs.webkit.org/show_bug.cgi?id=187311
Reviewed by Ryosuke Niwa.
If last four segments has the same value, MeasurmentSet should merge them.
- public/v3/models/measurement-set.js: Added the logic to conditionally merge last segments.
(MeasurementSet.prototype.fetchSegmentation):
- unit-tests/measurement-set-tests.js: Added a unit test for this change.
- 10:50 PM Changeset in webkit [233563] by
-
- 14 edits2 adds in trunk
REGRESSION (r230843): Flash doesn't work; Produces blue box on page
https://bugs.webkit.org/show_bug.cgi?id=187346
<rdar://problem/41773974>
Reviewed by Ryosuke Niwa.
Source/WebCore:
Introduce clearPagesPluginData used by WebProcess when a plugin policy changes.
Add a new internals API to get the plugin count as navigator.plugin is filtering plugins.
Test: http/tests/plugins/plugin-allow-then-reload.html
- plugins/PluginData.h:
- plugins/PluginInfoProvider.cpp:
(WebCore::PluginInfoProvider::clearPagesPluginData):
- plugins/PluginInfoProvider.h:
- testing/Internals.cpp:
(WebCore::Internals::pluginCount):
- testing/Internals.h:
- testing/Internals.idl:
Source/WebKit:
Reset page plugins when the load client policies are updated.
This will ensure that pages will rebuild their plugin lists based on the new policies.
- WebProcess/Plugins/WebPluginInfoProvider.cpp:
(WebKit::WebPluginInfoProvider::setPluginLoadClientPolicy):
(WebKit::WebPluginInfoProvider::clearPluginClientPolicies):
Tools:
- WebKitTestRunner/TestController.cpp:
(WTR::TestController::resetStateToConsistentValues):
(WTR::TestController::setBlockAllPlugins):
When not blocked, set the load policy to allow once.
- WebKitTestRunner/TestController.h:
LayoutTests:
Skip new test on WK1 since it does not have the ability to block plugins.
- http/tests/plugins/plugin-allow-then-reload-expected.txt: Added.
- http/tests/plugins/plugin-allow-then-reload.html: Added.
- platform/mac-wk1/TestExpectations:
- 10:48 PM Changeset in webkit [233562] by
-
- 19 edits2 adds in trunk
IndexedDB operations in a Page fail after a StorageProcess crash.
<rdar://problem/41626526> and https://bugs.webkit.org/show_bug.cgi?id=187123
Reviewed by Alex Christensen.
Source/WebCore:
Test: storage/indexeddb/modern/opendatabase-after-storage-crash.html
When the connection to a StorageProcess goes away, explicitly tell all of the WebPages
in the WebProcess about it.
This puts Documents/Workers in an error mode where requests fail instead of timeout.
It also clears the Page's connection so *new* Documents and Workers will get a fresh
new connection that works.
- Modules/indexeddb/client/IDBConnectionToServer.cpp:
(WebCore::IDBClient::IDBConnectionToServer::callResultFunctionLater):
(WebCore::IDBClient::IDBConnectionToServer::deleteDatabase):
(WebCore::IDBClient::IDBConnectionToServer::openDatabase):
(WebCore::IDBClient::IDBConnectionToServer::createObjectStore):
(WebCore::IDBClient::IDBConnectionToServer::deleteObjectStore):
(WebCore::IDBClient::IDBConnectionToServer::renameObjectStore):
(WebCore::IDBClient::IDBConnectionToServer::clearObjectStore):
(WebCore::IDBClient::IDBConnectionToServer::createIndex):
(WebCore::IDBClient::IDBConnectionToServer::deleteIndex):
(WebCore::IDBClient::IDBConnectionToServer::renameIndex):
(WebCore::IDBClient::IDBConnectionToServer::putOrAdd):
(WebCore::IDBClient::IDBConnectionToServer::getRecord):
(WebCore::IDBClient::IDBConnectionToServer::getAllRecords):
(WebCore::IDBClient::IDBConnectionToServer::getCount):
(WebCore::IDBClient::IDBConnectionToServer::deleteRecord):
(WebCore::IDBClient::IDBConnectionToServer::openCursor):
(WebCore::IDBClient::IDBConnectionToServer::iterateCursor):
(WebCore::IDBClient::IDBConnectionToServer::establishTransaction):
(WebCore::IDBClient::IDBConnectionToServer::commitTransaction):
(WebCore::IDBClient::IDBConnectionToServer::didFinishHandlingVersionChangeTransaction):
(WebCore::IDBClient::IDBConnectionToServer::abortTransaction):
(WebCore::IDBClient::IDBConnectionToServer::didFireVersionChangeEvent):
(WebCore::IDBClient::IDBConnectionToServer::confirmDidCloseFromServer):
(WebCore::IDBClient::IDBConnectionToServer::connectionToServerLost):
(WebCore::IDBClient::IDBConnectionToServer::openDBRequestCancelled):
(WebCore::IDBClient::IDBConnectionToServer::databaseConnectionPendingClose):
(WebCore::IDBClient::IDBConnectionToServer::databaseConnectionClosed):
(WebCore::IDBClient::IDBConnectionToServer::abortOpenAndUpgradeNeeded):
(WebCore::IDBClient::IDBConnectionToServer::getAllDatabaseNames):
- Modules/indexeddb/client/IDBConnectionToServer.h:
- Modules/indexeddb/shared/IDBError.h:
(WebCore::IDBError::serverConnectionLostError):
- page/Page.cpp:
(WebCore::Page::setSessionID):
(WebCore::Page::idbConnection):
(WebCore::Page::optionalIDBConnection):
(WebCore::Page::clearIDBConnection):
- page/Page.h:
Source/WebKit:
When the connection to a StorageProcess goes away, explicitly tell all of the WebPages
in the WebProcess about it.
This puts Documents/Workers in an error mode where requests fail instead of timeout.
It also clears the Page's connection so *new* Documents and Workers will get a fresh
new connection that works.
- UIProcess/API/C/WKContext.cpp:
(WKContextTerminateStorageProcess):
- UIProcess/API/Cocoa/WKProcessPool.mm:
(-[WKProcessPool _terminateStorageProcess]):
- UIProcess/Storage/StorageProcessProxy.cpp:
(WebKit::StorageProcessProxy::terminateForTesting):
- UIProcess/Storage/StorageProcessProxy.h:
- UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::terminateStorageProcessForTesting):
(WebKit::WebProcessPool::terminateStorageProcess): Deleted.
- UIProcess/WebProcessPool.h:
- WebProcess/Storage/WebToStorageProcessConnection.cpp:
(WebKit::WebToStorageProcessConnection::didClose):
- WebProcess/Storage/WebToStorageProcessConnection.h:
(WebKit::WebToStorageProcessConnection::existingIDBConnectionToServerForIdentifier):
- WebProcess/WebProcess.cpp:
(WebKit::WebProcess::webToStorageProcessConnectionClosed):
LayoutTests:
- storage/indexeddb/modern/opendatabase-after-storage-crash.html: Added.
- 9:56 PM Changeset in webkit [233561] by
-
- 2 edits in trunk/Source/WebKit
Address two possible causes of missing tiles in iOS Safari, and add logging to gather more data about other possible causes
https://bugs.webkit.org/show_bug.cgi?id=187376
rdar://problem/40941118
Reviewed by Tim Horton.
We have continual reports of users experiencing missing tiles in MobileSafari, where loading a page
shows the tiles at the top, but we don't render new tiles as the user scrolls down. This is consistent
with failing to dispatch visible content rect updates via -[WKWebView _updateVisibleContentRects].
This patch addresses two possible (but unlikely) causes. First, it resets _currentlyAdjustingScrollViewInsetsForKeyboard
after a web process crash. Second, it catches exceptions thrown by [webView _updateVisibleContentRects]
and resets _hasScheduledVisibleRectUpdate.
This patch also adds release logging that fires if over 1s has elapsed between scheduling
a visible content rect update and trying to re-schedule, and logging for all reasons that
-_updateVisibleContentRects returns early.
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]):
(-[WKWebView _processDidExit]):
(-[WKWebView _addUpdateVisibleContentRectPreCommitHandler]):
(-[WKWebView _scheduleVisibleContentRectUpdateAfterScrollInView:]):
(-[WKWebView _updateVisibleContentRects]):
- 8:20 PM Changeset in webkit [233560] by
-
- 8 edits2 adds in trunk
Youtube video pages crash after a couple of minutes
https://bugs.webkit.org/show_bug.cgi?id=187316
Reviewed by Antti Koivisto.
Source/WebCore:
The crash was caused by HTMLMediaElement::stopWithoutDestroyingMediaPlayer invoking updatePlaybackControlsManager,
which traverses all media players across different documents including the one in the main frame while its iframe
is getting removed (to update the Touch Bar's media control).
Fixed the bug by making this code async in both stopWithoutDestroyingMediaPlayer and ~HTMLMediaElement. To do this,
this patch moves the timer to update the playback controls manager from HTMLMediaElement to Page since scheduling
a timer owned by HTMLMediaElement in its destructor wouldn't work as the timer would get destructed immediately.
Also replaced the call to clientWillPausePlayback by a call to stopSession in stopWithoutDestroyingMediaPlayer
since the former also updates the layout synchronously via updateNowPlayingInfo; the latter function schedules
a timer via scheduleUpdateNowPlayingInfo instead.
Test: media/remove-video-best-media-element-in-main-frame-crash.html
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::~HTMLMediaElement): Call scheduleUpdatePlaybackControlsManager now that timer has been
moved to Page.
(WebCore::HTMLMediaElement::bestMediaElementForShowingPlaybackControlsManager): Made this return a RefPtr instead of
a raw pointer while we're at it.
(WebCore::HTMLMediaElement::clearMediaPlayer): Call scheduleUpdatePlaybackControlsManager.
(WebCore::HTMLMediaElement::stopWithoutDestroyingMediaPlayer): Ditto. Also invoke stopSession instead of
clientWillPausePlayback on MediaSession since clientWillPausePlayback will synchronously try to update the layout.
(WebCore::HTMLMediaElement::contextDestroyed):
(WebCore::HTMLMediaElement::stop):
(WebCore::HTMLMediaElement::schedulePlaybackControlsManagerUpdate): Renamed from scheduleUpdatePlaybackControlsManager.
(WebCore::HTMLMediaElement::updatePlaybackControlsManager): Moved to Page::playbackControlsManagerUpdateTimerFired.
- html/HTMLMediaElement.h:
- page/Page.cpp:
(WebCore::Page::Page):
(WebCore::Page::schedulePlaybackControlsManagerUpdate): Added.
(WebCore::Page::playbackControlsManagerUpdateTimerFired): Moved from HTMLMediaElement::updatePlaybackControlsManager.
- page/Page.h:
- testing/Internals.cpp:
(WebCore::Internals::bestMediaElementForShowingPlaybackControlsManager):
- testing/Internals.h:
LayoutTests:
Added a regression test to remove an iframe with a video while there is a main content
which is eligible to be shown in the Touch Bar.
- media/remove-video-best-media-element-in-main-frame-crash-expected.txt: Added.
- media/remove-video-best-media-element-in-main-frame-crash.html: Added.
- 7:44 PM Changeset in webkit [233559] by
-
- 1 copy in tags/Safari-606.1.24.1
Tag Safari-606.1.24.1.
- 7:39 PM Changeset in webkit [233558] by
-
- 7 edits in tags/Safari-606.1.24/Source
Versioning.
- 7:32 PM Changeset in webkit [233557] by
-
- 2 edits in trunk/Source/WebCore
REGRESSION(r233496): Crash in WebCore::VideoTrack::clearClient()
https://bugs.webkit.org/show_bug.cgi?id=187377
Reviewed by Simon Fraser.
Clear m_client of an audio track or a video track before removing it from the list
since TrackListBase::m_inbandTracks may hold the last ref to the track.
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::removeAudioTrack):
(WebCore::HTMLMediaElement::removeVideoTrack):
- 6:56 PM Changeset in webkit [233556] by
-
- 4 edits in trunk
REGRESSION(r233495) [cairo] drawGlyphsShadow should use the fast path for zero blur-radius
https://bugs.webkit.org/show_bug.cgi?id=187336
Reviewed by Carlos Garcia Campos.
Source/WebCore:
drawGlyphsShadow has a fast path to draw a shadow without a shadow
layer. Since r233495, it is not used for zero blur-radius shadow.
In Bug 187244, I changed ShadowState::isRequired not to check
blur-radius is zero. This is not a right fix.
This change reverts r233495.
No new tests (Covered by existing tests).
- platform/graphics/cairo/CairoOperations.cpp:
(WebCore::Cairo::ShadowState::isRequired const): Return false if
'blur' is zero.
(WebCore::Cairo::fillRectWithRoundedHole): Use
ShadowState::isVisible instead of ShadowState::isRequired to check
a shadow should be drawn.
LayoutTests:
- platform/gtk/TestExpectations: Unmarked fast/text/multiple-text-shadow-overflow-layout-rect.html.
- 6:35 PM Changeset in webkit [233555] by
-
- 3 edits in trunk/LayoutTests
[mac-wk1] Layout Test editing/selection/navigation-clears-editor-state.html is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=187309
Reviewed by Mark Lam.
Try to make this test less flaky by making percents threshold of freed
objects smaller.
- editing/selection/navigation-clears-editor-state-expected.txt:
- editing/selection/navigation-clears-editor-state.html:
- 5:43 PM Changeset in webkit [233554] by
-
- 4 edits in trunk/LayoutTests
Layout Test http/tests/workers/service/basic-messageport.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=182630
Reviewed by Alex Christensen.
Test is flaky due to a race between serviceWorkerClient.postMessage and MessagePort.postMessage.
Both are going through UIProcess but MessagePort events are grabbed asynchronously.
Fixed the test to not expect order between these two message exchange channels.
- http/tests/workers/service/basic-messageport-expected.txt:
- http/tests/workers/service/resources/basic-messageport.js:
(channel.port1.onmessage):
(then):
- platform/mac-wk2/TestExpectations:
- 5:40 PM Changeset in webkit [233553] by
-
- 2 edits in trunk/Source/WebKit
iPad: Scrolling with hardware keyboard while SELECT popover is visible scrolls the page, detaches popover
https://bugs.webkit.org/show_bug.cgi?id=187363
<rdar://problem/41106306>
Patch by Olivia Barnett <obarnett@apple.com> on 2018-07-05
Reviewed by Tim Horton.
Added extra check to prevent keyboard scrolling when the select box is open.
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _scrollOffsetForEvent:]):
- 5:38 PM Changeset in webkit [233552] by
-
- 44 edits in trunk/Source
Rename and flip the meaning of defaultAppearance to be useDarkAppearance.
https://bugs.webkit.org/show_bug.cgi?id=187369
rdar://problem/41870420
Reviewed by Tim Horton.
Source/WebCore:
- Modules/mediasession/WebMediaSessionManager.cpp:
(WebCore::WebMediaSessionManager::showPlaybackTargetPicker):
- css/MediaQueryEvaluator.cpp:
(WebCore::prefersDarkInterfaceEvaluate):
- css/StyleColor.h:
- dom/Document.cpp:
(WebCore::Document::useDarkAppearance const):
(WebCore::Document::styleColorOptions const):
(WebCore::Document::useDefaultAppearance const): Deleted.
- dom/Document.h:
- inspector/InspectorOverlay.cpp:
(WebCore::InspectorOverlay::paint):
- page/Page.cpp:
(WebCore::Page::useDarkAppearance const):
(WebCore::Page::defaultAppearance const): Deleted.
- page/Page.h:
(WebCore::Page::setUseDarkAppearance):
(WebCore::Page::setDefaultAppearance): Deleted.
- platform/Theme.h:
- platform/graphics/MediaPlaybackTargetPicker.h:
- platform/graphics/avfoundation/objc/MediaPlaybackTargetPickerMac.h:
- platform/graphics/avfoundation/objc/MediaPlaybackTargetPickerMac.mm:
(WebCore::MediaPlaybackTargetPickerMac::showPlaybackTargetPicker):
- platform/mac/DragImageMac.mm:
(WebCore::createDragImageForLink):
- platform/mac/LocalDefaultSystemAppearance.h:
- platform/mac/LocalDefaultSystemAppearance.mm:
(WebCore::LocalDefaultSystemAppearance::LocalDefaultSystemAppearance):
- platform/mac/ThemeMac.h:
- platform/mac/ThemeMac.mm:
(WebCore::ThemeMac::paint):
- platform/mock/MediaPlaybackTargetPickerMock.cpp:
(WebCore::MediaPlaybackTargetPickerMock::showPlaybackTargetPicker):
- platform/mock/MediaPlaybackTargetPickerMock.h:
- rendering/InlineTextBox.cpp:
(WebCore::InlineTextBox::resolveStyleForMarkedText):
- rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::paintContents):
- rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::paintContents):
- rendering/RenderTheme.cpp:
(WebCore::RenderTheme::paint):
- rendering/RenderThemeMac.mm:
(WebCore::RenderThemeMac::documentViewFor const):
(WebCore::RenderThemeMac::platformActiveSelectionBackgroundColor const):
(WebCore::RenderThemeMac::platformInactiveSelectionBackgroundColor const):
(WebCore::RenderThemeMac::supportsSelectionForegroundColors const):
(WebCore::RenderThemeMac::platformActiveSelectionForegroundColor const):
(WebCore::RenderThemeMac::platformInactiveSelectionForegroundColor const):
(WebCore::RenderThemeMac::platformActiveListBoxSelectionBackgroundColor const):
(WebCore::RenderThemeMac::platformInactiveListBoxSelectionBackgroundColor const):
(WebCore::RenderThemeMac::platformActiveListBoxSelectionForegroundColor const):
(WebCore::RenderThemeMac::platformInactiveListBoxSelectionForegroundColor const):
(WebCore::RenderThemeMac::platformActiveTextSearchHighlightColor const):
(WebCore::RenderThemeMac::colorCache const):
(WebCore::RenderThemeMac::systemColor const):
(WebCore::RenderThemeMac::paintCellAndSetFocusedElementNeedsRepaintIfNecessary):
(WebCore::RenderThemeMac::paintSliderThumb):
- svg/graphics/SVGImage.cpp:
(WebCore::SVGImage::draw):
Source/WebKit:
- Shared/WebPageCreationParameters.cpp:
(WebKit::WebPageCreationParameters::encode const):
(WebKit::WebPageCreationParameters::decode):
- Shared/WebPageCreationParameters.h:
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _effectiveAppearanceIsDark]):
(-[WKWebView _initializeWithConfiguration:]):
(-[WKWebView _setUseSystemAppearance:]):
(-[WKWebView effectiveAppearanceDidChange]):
(-[WKWebView _defaultAppearance]): Deleted.
- UIProcess/API/mac/WKView.mm:
(-[WKView _effectiveAppearanceIsDark]):
(-[WKView effectiveAppearanceDidChange]):
(-[WKView _setUseSystemAppearance:]):
(-[WKView _defaultAppearance]): Deleted.
(-[WKView _setDefaultAppearance:]): Deleted.
- UIProcess/Cocoa/WebViewImpl.h:
- UIProcess/Cocoa/WebViewImpl.mm:
(WebKit::WebViewImpl::effectiveAppearanceIsDark):
(WebKit::WebViewImpl::setUseDarkAppearance):
(WebKit::WebViewImpl::useDefaultAppearance): Deleted.
(WebKit::WebViewImpl::setDefaultAppearance): Deleted.
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::creationParameters):
(WebKit::WebPageProxy::setUseDarkAppearance):
(WebKit::WebPageProxy::showPlaybackTargetPicker):
(WebKit::WebPageProxy::setDefaultAppearance): Deleted.
- UIProcess/WebPageProxy.h:
(WebKit::WebPageProxy::useDarkAppearance const):
(WebKit::WebPageProxy::defaultAppearance const): Deleted.
- UIProcess/mac/WKPrintingView.mm:
(-[WKPrintingView drawRect:]):
- WebProcess/InjectedBundle/DOM/InjectedBundleRangeHandle.cpp:
(WebKit::InjectedBundleRangeHandle::renderedImage):
- WebProcess/Plugins/PDF/PDFPlugin.mm:
(WebKit::PDFPlugin::paintControlForLayerInContext):
- WebProcess/WebPage/WebPage.cpp:
(WebKit::m_credentialsMessenger):
(WebKit::WebPage::drawRect):
(WebKit::WebPage::setUseDarkAppearance):
(WebKit::WebPage::setDefaultAppearance): Deleted.
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/WebPage.messages.in:
Source/WebKitLegacy/mac:
- WebView/WebMediaPlaybackTargetPicker.mm:
(WebMediaPlaybackTargetPicker::showPlaybackTargetPicker):
- WebView/WebView.mm:
(-[WebView _commonInitializationWithFrameName:groupName:]):
(-[WebView _effectiveAppearanceIsDark]):
(-[WebView _updateDefaultAppearance]):
(-[WebView _defaultAppearance]): Deleted.
- 5:25 PM Changeset in webkit [233551] by
-
- 2 edits in trunk/LayoutTests
LayoutTest imported/w3c/web-platform-tests/IndexedDB/interfaces.any.worker.html is flaky.
https://bugs.webkit.org/show_bug.cgi?id=187372
Unreviewed test gardening.
- platform/mac-wk1/TestExpectations:
- 5:19 PM Changeset in webkit [233550] by
-
- 3 edits in trunk/Source/bmalloc
Build fix (r233547): Disable IsoHeaps when Gigacage is off
<https://webkit.org/b/187160>
- bmalloc/PerThread.cpp: Add #if !HAVE_PTHREAD_MACHDEP_H/#endif
around variables only used when that macro is 0. Include what
you use: Cache.h and Heap.h.
- bmalloc/PerThread.h: Include <memory> for std::once_flag.
- 4:24 PM Changeset in webkit [233549] by
-
- 8 edits2 deletes in trunk
Unreviewed, rolling out r233539.
https://bugs.webkit.org/show_bug.cgi?id=187368
Changes made for youtube crashes has caused 15 api failures
(Requested by Truitt on #webkit).
Reverted changeset:
"Youtube video pages crash after a couple of minutes"
https://bugs.webkit.org/show_bug.cgi?id=187316
https://trac.webkit.org/changeset/233539
- 4:17 PM Changeset in webkit [233548] by
-
- 2 edits in trunk/Source/WebKit
REGRESSION: ASSERT under WebAutomationSessionProxy::computeElementLayout when elementInViewClientCenterPoint returns nullopt
https://bugs.webkit.org/show_bug.cgi?id=187367
<rdar://problem/41861346>
Reviewed by Timothy Hatcher.
- WebProcess/Automation/WebAutomationSessionProxy.cpp:
(WebKit::WebAutomationSessionProxy::computeElementLayout):
There's no reason to unwrap this optional, as the IPC argument type is std::optional<IntPoint>.
- 4:12 PM Changeset in webkit [233547] by
-
- 5 edits1 add in trunk/Source/bmalloc
Disable IsoHeaps when Gigacage is off
https://bugs.webkit.org/show_bug.cgi?id=187160
Reviewed by Saam Barati.
Updated determineMallocFallbackState to base enabling of Iso Heaps on Gigacage
being enabled. We do this because if Gigacage is disabled, it may be due to lack
of address space.
To work around a compiler issue uncovered by the change above, I added explicit
instantiation of PerThread's static variables. Defined the same explicit
instantiated static variables with export scope in the new file PerThread.cpp
to eliminate separate variables allocations in each linked framework / library.
- CMakeLists.txt:
- bmalloc.xcodeproj/project.pbxproj:
- bmalloc/IsoTLS.cpp:
(bmalloc::IsoTLS::determineMallocFallbackState):
- bmalloc/PerThread.cpp: Added.
- bmalloc/PerThread.h:
- 3:28 PM Changeset in webkit [233546] by
-
- 3 edits in trunk/Source/WebKit
Upstream hover gesture implementation
https://bugs.webkit.org/show_bug.cgi?id=187366
Reviewed by Wenson Hsieh.
- UIProcess/ios/WKContentViewInteraction.h:
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView setupInteraction]):
(-[WKContentView cleanupInteraction]):
(-[WKContentView _removeDefaultGestureRecognizers]):
(-[WKContentView _addDefaultGestureRecognizers]):
(-[WKContentView _hoverGestureRecognizerChanged:]):
- 3:20 PM Changeset in webkit [233545] by
-
- 3 edits in trunk/Source/WebCore
Update Element API to use qualifiedName instead of name/localName where applicable
https://bugs.webkit.org/show_bug.cgi?id=187347
Reviewed by Darin Adler.
Update Element API to use qualifiedName instead of name/localName where applicable. Naming
is now consistent with the IDL and the specification:
Our implementation properly deals with qualified names already, even though the parameters
were named localName.
- dom/Element.cpp:
(WebCore::Element::getAttribute const):
(WebCore::Element::toggleAttribute):
(WebCore::Element::setAttribute):
(WebCore::Element::removeAttribute):
(WebCore::Element::getAttributeNode):
(WebCore::Element::hasAttribute const):
- dom/Element.h:
- 3:20 PM Changeset in webkit [233544] by
-
- 5 edits in trunk/Source/WebKit
Fix some -Wdocumentation warnings in WebKit
https://bugs.webkit.org/show_bug.cgi?id=187318
Reviewed by Dan Bernstein.
Fix most of the -Wdocumentation warnings in WebKit. The remaining
ones are slightly mysterious, so we can't turn the warning on yet.
- Configurations/BaseTarget.xcconfig:
Define U_HIDE_DEPRECATED_API. It's defined in all other WebKit projects,
and the doc comments in the deprecated ICU API have some errors.
- UIProcess/API/Cocoa/WKUIDelegate.h:
- UIProcess/API/Cocoa/WKWebView.h:
- UIProcess/API/Cocoa/WKWebsiteDataStore.h:
Adjust comments or parameter names to be consistent with each other.
- 2:51 PM Changeset in webkit [233543] by
-
- 7 edits in trunk/Source
Versioning.
- 2:46 PM Changeset in webkit [233542] by
-
- 3 edits in trunk/Source/WTF
Unreviewed, rolling out r233417 and r233418.
https://bugs.webkit.org/show_bug.cgi?id=187364
Introduced an occasional crash on Google Drive (Requested by
rniwa on #webkit).
Reverted changesets:
"WTF's internal std::optional implementation should abort() on
bad optional access"
https://bugs.webkit.org/show_bug.cgi?id=186536
https://trac.webkit.org/changeset/233417
"WTF's internal std::optional implementation should abort() on
bad optional access"
https://bugs.webkit.org/show_bug.cgi?id=186536
https://trac.webkit.org/changeset/233418
- 2:43 PM Changeset in webkit [233541] by
-
- 4 edits in tags/Safari-606.1.24
Revert r233417. rdar://problem/41863878
- 2:35 PM Changeset in webkit [233540] by
-
- 2 edits in trunk/Source/JavaScriptCore
ProgramExecutable may be collected as we checkSyntax on it
https://bugs.webkit.org/show_bug.cgi?id=187359
<rdar://problem/41832135>
Reviewed by Mark Lam.
The bug was we were passing in a reference to the SourceCode field on ProgramExecutable as
the ProgramExecutable itself may be collected. The fix here is to make a copy
of the field instead of passing in a reference inside of ParserError::toErrorObject.
No new tests here as this was already caught by our iOS JSC testers.
- parser/ParserError.h:
(JSC::ParserError::toErrorObject):
- 2:02 PM Changeset in webkit [233539] by
-
- 9 edits2 adds in trunk
Youtube video pages crash after a couple of minutes
https://bugs.webkit.org/show_bug.cgi?id=187316
Reviewed by Antti Koivisto.
Source/WebCore:
The crash was caused by HTMLMediaElement::stopWithoutDestroyingMediaPlayer invoking updatePlaybackControlsManager,
which traverses all media players across different documents including the one in the main frame while its iframe
is getting removed (to update the Touch Bar's media control).
Fixed the bug by making this code async in both stopWithoutDestroyingMediaPlayer and ~HTMLMediaElement. To do this,
this patch moves the timer to update the playback controls manager from HTMLMediaElement to Page since scheduling
a timer owned by HTMLMediaElement in its destructor wouldn't work as the timer would get destructed immediately.
Also replaced the call to clientWillPausePlayback by a call to stopSession in stopWithoutDestroyingMediaPlayer
since the former also updates the layout synchronously via updateNowPlayingInfo; the latter function schedules
a timer via scheduleUpdateNowPlayingInfo instead.
Test: media/remove-video-best-media-element-in-main-frame-crash.html
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::~HTMLMediaElement): Call scheduleUpdatePlaybackControlsManager now that timer has been
moved to Page.
(WebCore::HTMLMediaElement::bestMediaElementForShowingPlaybackControlsManager): Made this return a RefPtr instead of
a raw pointer while we're at it.
(WebCore::HTMLMediaElement::clearMediaPlayer): Call scheduleUpdatePlaybackControlsManager.
(WebCore::HTMLMediaElement::stopWithoutDestroyingMediaPlayer): Ditto. Also invoke stopSession instead of
clientWillPausePlayback on MediaSession since clientWillPausePlayback will synchronously try to update the layout.
(WebCore::HTMLMediaElement::contextDestroyed):
(WebCore::HTMLMediaElement::stop):
(WebCore::HTMLMediaElement::scheduleUpdatePlaybackControlsManager):
(WebCore::HTMLMediaElement::updatePlaybackControlsManager): Moved to Page::schedulePlaybackControlsManagerUpdate.
- html/HTMLMediaElement.h:
- page/Page.cpp:
(WebCore::Page::schedulePlaybackControlsManagerUpdate): Added.
- page/Page.h:
- testing/Internals.cpp:
(WebCore::Internals::bestMediaElementForShowingPlaybackControlsManager):
- testing/Internals.h:
Source/WebKitLegacy/mac:
- WebView/WebView.mm:
(-[WebView _preferencesChanged:]):
LayoutTests:
Added a regression test to remove an iframe with a video while there is a main content
which is eligible to be shown in the Touch Bar.
- media/remove-video-best-media-element-in-main-frame-crash-expected.txt: Added.
- media/remove-video-best-media-element-in-main-frame-crash.html: Added.
- 1:41 PM Changeset in webkit [233538] by
-
- 2 edits in trunk/Source/WebKit
Regression(r232886): WebsiteDataStore objects may get destroyed on a background thread
https://bugs.webkit.org/show_bug.cgi?id=187356
<rdar://problem/41854555>
Reviewed by Geoffrey Garen.
As of r232886, CallbackAggregators in WebsiteDataStore hold a Ref<> to their WebsiteDataStore. This
is an issue because CallbackAggregator objects can get destroyed on a background thread and may be
the last ones holding a ref to the data store. When this happens, the WebsiteDataStore would get
destroyed on a background store and potentially cause crashes. Note that even if the callback
aggregator would not be the last one to hold a ref to the store, it still would not be safe to deref
the store on the background thread since WebsiteDataStore is not ThreadSafeRefCounted.
To address the issue, this patch updates the CallbackAggregators' destructor to deref their data
store member on the main thread. Note that we could also have WebsiteDataStore subclass
ThreadSafeRefCounted<T, DestructionThread::Main> but the data store technically does not need to
be ThreadSafeRefCounted at the moment.
- UIProcess/WebsiteData/WebsiteDataStore.cpp:
(WebKit::WebsiteDataStore::WebsiteDataStore):
(WebKit::WebsiteDataStore::~WebsiteDataStore):
(WebKit::WebsiteDataStore::fetchDataAndApply):
(WebKit::WebsiteDataStore::removeData):
- 1:28 PM Changeset in webkit [233537] by
-
- 4 edits2 adds in trunk
Do not assume that hypen's width can be computed using the simplified text measure codepath.
https://bugs.webkit.org/show_bug.cgi?id=187352
<rdar://problem/40821283>
Reviewed by Simon Fraser.
Source/WebCore:
Just because the text content is qualified for simplified text measure, it does not necessarily mean the hyphen is simple enough as well.
Test: fast/text/hyphen-is-complex-crash.html
- rendering/SimpleLineLayoutTextFragmentIterator.cpp:
(WebCore::SimpleLineLayout::TextFragmentIterator::Style::Style):
(WebCore::SimpleLineLayout::TextFragmentIterator::TextFragmentIterator):
- rendering/SimpleLineLayoutTextFragmentIterator.h:
LayoutTests:
- fast/text/hyphen-is-complex-crash-expected.txt: Added.
- fast/text/hyphen-is-complex-crash.html: Added.
- 12:45 PM Changeset in webkit [233536] by
-
- 2 edits in trunk/Source/WebKit
[macOS] REGRESSION: Development WebContent service has restricted entitlements, rendering it useless for development
https://bugs.webkit.org/show_bug.cgi?id=187355
Reviewed by Anders Carlsson.
- Configurations/WebContentService.Development.xcconfig: Set WK_USE_RESTRICTED_ENTITLEMENTS to NO for the Development service.
- 12:42 PM Changeset in webkit [233535] by
-
- 2 edits in trunk/Source/WebCore
Video sometimes flickers when playing to AppleTV
https://bugs.webkit.org/show_bug.cgi?id=187193
<rdar://problem/40153865>
Unreviewed, fix a broken API test.
- platform/audio/ios/AudioSessionIOS.mm:
(WebCore::AudioSession::setCategory): None should set the category to AVAudioSessionCategoryAmbient,
because it signals that WebKit is not playing audio so we want allow other apps to play.
- 12:39 PM Changeset in webkit [233534] by
-
- 2 edits in trunk/Source/WebCore
Use ENABLE(WEBPROCESS_WINDOWSERVER_BLOCKING) in ScrollAnimatorMac.mm
https://bugs.webkit.org/show_bug.cgi?id=187334
Reviewed by Alex Christensen.
Use ENABLE(WEBPROCESS_WINDOWSERVER_BLOCKING) instead of MAC_OS_X_VERSION_MIN_REQUIRED when deciding
to use NSTimer or NSAnimation for scrollbar animation.
No new tests. No change in behavior.
- platform/mac/ScrollAnimatorMac.mm:
(-[WebScrollbarPartAnimation initWithScrollbar:featureToAnimate:animateFrom:animateTo:duration:]):
(-[WebScrollbarPartAnimation startAnimation]):
(-[WebScrollbarPartAnimation setCurrentProgress:setCurrentProgress:]):
- 12:28 PM Changeset in webkit [233533] by
-
- 2 edits in trunk/LayoutTests
[WK1] editing/spelling/markers.html times out on High Sierra after r233439
https://bugs.webkit.org/show_bug.cgi?id=187351
Reviewed by Tim Horton.
This is currently timing out on High Sierra, because the
verifyUnexpectedMarkersportion of the
test fails to actually make use of the override spell checking results. Fix this by passing the
results when calling the function.
- editing/spelling/markers.html:
- 12:11 PM Changeset in webkit [233532] by
-
- 5 edits in trunk/Source/WebCore
Don't invert text color for selections in light mode.
https://bugs.webkit.org/show_bug.cgi?id=187349
rdar://problem/41297946
Reviewed by Tim Horton.
Fix color caching in RenderTheme so we don't cache a dark mode color
for a light appearance, or vise versa. Use the new color caching
in RenderThemeMac, and clear the color caches in purgeCaches.
Allow supportsSelectionForegroundColors to conditionalize on StyleColor::Options.
Return true only in dark mode.
- rendering/RenderTheme.cpp:
(WebCore::RenderTheme::activeSelectionBackgroundColor const):
(WebCore::RenderTheme::inactiveSelectionBackgroundColor const):
(WebCore::RenderTheme::activeSelectionForegroundColor const):
(WebCore::RenderTheme::inactiveSelectionForegroundColor const):
(WebCore::RenderTheme::activeListBoxSelectionBackgroundColor const):
(WebCore::RenderTheme::inactiveListBoxSelectionBackgroundColor const):
(WebCore::RenderTheme::activeListBoxSelectionForegroundColor const):
(WebCore::RenderTheme::inactiveListBoxSelectionForegroundColor const):
(WebCore::RenderTheme::purgeCaches):
(WebCore::RenderTheme::platformColorsDidChange):
(WebCore::RenderTheme::activeTextSearchHighlightColor const):
(WebCore::RenderTheme::inactiveTextSearchHighlightColor const):
- rendering/RenderTheme.h:
(WebCore::RenderTheme::supportsSelectionForegroundColors const):
(WebCore::RenderTheme::supportsListBoxSelectionForegroundColors const):
(WebCore::RenderTheme::colorCache const):
(WebCore::RenderTheme::purgeCaches): Deleted.
- rendering/RenderThemeMac.h:
- rendering/RenderThemeMac.mm:
(WebCore::RenderThemeMac::purgeCaches):
(WebCore::RenderThemeMac::supportsSelectionForegroundColors const):
(WebCore::RenderThemeMac::platformActiveSelectionForegroundColor const):
(WebCore::RenderThemeMac::platformInactiveSelectionForegroundColor const):
(WebCore::RenderThemeMac::platformColorsDidChange):
(WebCore::RenderThemeMac::colorCache const):
(WebCore::RenderThemeMac::systemColor const):
- 11:36 AM Changeset in webkit [233531] by
-
- 3 edits in trunk/Tools
[ews-build] Bindings-tests-EWS should output result in json
https://bugs.webkit.org/show_bug.cgi?id=187330
Reviewed by David Kilzer.
- BuildSlaveSupport/ews-build/steps.py:
(RunBindingsTests): Passed --json-output parameter to run-bindings-tests.
- BuildSlaveSupport/ews-build/steps_unittest.py: Updated unit-tests.
- 11:18 AM Changeset in webkit [233530] by
-
- 2 edits in trunk/Source/WebCore
SimpleLineLayout::FlowContents wastes 54KB of Vector capacity on nytimes.com
https://bugs.webkit.org/show_bug.cgi?id=186709
<rdar://problem/41173793>
Reviewed by Simon Fraser.
The size of the m_segments vector in SimpleLineLayoutFlowContents is alway pre-computed and don't change after the initial append.
Not testable.
- rendering/SimpleLineLayoutFlowContents.h:
- 10:52 AM Changeset in webkit [233529] by
-
- 1 copy in tags/Safari-606.1.24
Tag Safari-606.1.24.
- 10:49 AM Changeset in webkit [233528] by
-
- 2 edits in trunk/Tools
[ews-build] WebKitPy-Tests-EWS should output result in json
https://bugs.webkit.org/show_bug.cgi?id=187331
Follow-up fix to r233527. Extracting jsonFileName as a variable and re-using it.
- 10:25 AM Changeset in webkit [233527] by
-
- 3 edits in trunk/Tools
[ews-build] WebKitPy-Tests-EWS should output result in json
https://bugs.webkit.org/show_bug.cgi?id=187331
Reviewed by David Kilzer.
- BuildSlaveSupport/ews-build/steps.py:
(RunWebKitPyTests): Passed --json-output parameter to test-webkitpy.
- BuildSlaveSupport/ews-build/steps_unittest.py: Updated unit-tests.
- 10:22 AM Changeset in webkit [233526] by
-
- 7 edits in tags/Safari-605.3.8.1/Source
Versioning.
- 10:20 AM Changeset in webkit [233525] by
-
- 3 edits2 adds in trunk
AX: setValue on text controls should send out key events
https://bugs.webkit.org/show_bug.cgi?id=186690
<rdar://problem/41169985>
Reviewed by Chris Fleizach.
Source/WebCore:
Use Editor's insertText function when the text control element is being
focused, so that the InputEvent will be dispatched properly.
Test: accessibility/mac/set-value-editable-dispatch-events.html
- accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::setValue):
LayoutTests:
- accessibility/mac/set-value-editable-dispatch-events-expected.txt: Added.
- accessibility/mac/set-value-editable-dispatch-events.html: Added.
- 10:09 AM Changeset in webkit [233524] by
-
- 1 copy in tags/Safari-605.3.8.1
New tag.
- 10:07 AM Changeset in webkit [233523] by
-
- 2 edits in trunk/LayoutTests
Layout Test svg/animations/animateMotion-additive-2c.svg and svg/animations/animateMotion-additive-2d.svg is flaky
https://bugs.webkit.org/show_bug.cgi?id=187345
Unreviewed test gardening.
Patch by Truitt Savell <Truitt Savell> on 2018-07-05
- platform/mac-wk2/TestExpectations:
- 9:59 AM Changeset in webkit [233522] by
-
- 7 edits in branches/safari-605-branch/Source
Versioning.
- 9:05 AM Changeset in webkit [233521] by
-
- 2 edits in trunk/Source/WebCore
ASSERT(m_dictionaryStack.size() == 1) assertion hit in ~KeyedDecoderCF() when decoding fails
https://bugs.webkit.org/show_bug.cgi?id=187152
Reviewed by Brent Fulgham.
Whenever a call to KeyedDecoderCF::beginArrayElement() succeeds (returns true), it
appends an item to m_dictionaryStack which is expected to get popped later on via
a matching call to KeyedDecoderCF::endArrayElement(). The ASSERT(m_dictionaryStack.size() == 1)
assertion in the destructor protects against a begin / end mismatch.
However, there was a bug in KeyedDecoder::decodeObjects(), which would cause us to return early
after a call to beginArrayElement() when the lambda function call returns false, causing us
not to call the matching endArrayElement() in this case. This patch addresses the issue by
calling endArrayElement() before returning early.
This bug was uncovered by r233309 which slightly changed the ITP encoding / decoding format.
Because empty HashCountedSets are now expected to be encoded as empty array, decoding of old
data fails and we hit the buggy code path.
- platform/KeyedCoding.h:
(WebCore::KeyedDecoder::decodeObjects):
- 8:46 AM Changeset in webkit [233520] by
-
- 16 edits2 adds in trunk/Source/WebCore
Move CSSParserContext to a file of its own
https://bugs.webkit.org/show_bug.cgi?id=187342
Reviewed by Zalan Bujtas.
Move it out from CSSParserMode.h and CSSParser.cpp.
- Sources.txt:
- WebCore.xcodeproj/project.pbxproj:
- contentextensions/ContentExtensionParser.cpp:
- css/PropertySetCSSStyleDeclaration.h:
- css/StyleProperties.h:
- css/StyleSheetContents.h:
- css/parser/CSSDeferredParser.h:
- css/parser/CSSParser.cpp:
(WebCore::strictCSSParserContext): Deleted.
(WebCore::CSSParserContext::CSSParserContext): Deleted.
(WebCore::operator==): Deleted.
- css/parser/CSSParser.h:
- css/parser/CSSParserContext.cpp: Added.
(WebCore::strictCSSParserContext):
(WebCore::CSSParserContext::CSSParserContext):
(WebCore::operator==):
Also add the missing textAutosizingEnabled bit.
- css/parser/CSSParserContext.h: Copied from Source/WebCore/css/parser/CSSParserMode.h.
(WebCore::CSSParserContextHash::hash):
Also add the missing bits to the hash function.
(WebCore::isQuirksModeBehavior): Deleted.
(WebCore::isUASheetBehavior): Deleted.
(WebCore::isUnitLessValueParsingEnabledForMode): Deleted.
(WebCore::isCSSViewportParsingEnabledForMode): Deleted.
(WebCore::strictToCSSParserMode): Deleted.
(WebCore::isStrictParserMode): Deleted.
- css/parser/CSSParserIdioms.h:
- css/parser/CSSParserImpl.h:
- css/parser/CSSParserMode.h:
(WebCore::CSSParserContext::completeURL const): Deleted.
(WebCore::operator!=): Deleted.
(WebCore::CSSParserContextHash::hash): Deleted.
(WebCore::CSSParserContextHash::equal): Deleted.
(WTF::HashTraits<WebCore::CSSParserContext>::constructDeletedValue): Deleted.
(WTF::HashTraits<WebCore::CSSParserContext>::isDeletedValue): Deleted.
(WTF::HashTraits<WebCore::CSSParserContext>::emptyValue): Deleted.
- css/parser/CSSPropertyParserHelpers.h:
- css/parser/CSSSelectorParser.cpp:
- css/parser/CSSVariableParser.cpp:
- 7:56 AM Changeset in webkit [233519] by
-
- 6 edits in trunk
[Web Animations] The ready promise should initially be resolved
https://bugs.webkit.org/show_bug.cgi?id=187341
<rdar://problem/41844711>
Reviewed by Dean Jackson.
LayoutTests/imported/w3c:
Mark WPT progressions.
- web-platform-tests/web-animations/interfaces/Animation/finish-expected.txt:
- web-platform-tests/web-animations/timing-model/animations/current-time-expected.txt:
- web-platform-tests/web-animations/timing-model/animations/set-the-animation-start-time-expected.txt:
Source/WebCore:
The Web Animations spec mandates that the ready promise is initially resolved.
- animation/WebAnimation.cpp:
(WebCore::WebAnimation::WebAnimation):
- 7:18 AM Changeset in webkit [233518] by
-
- 4 edits1 delete in trunk/Source/WebKit
[GTK] Remove soup/DownloadSoupErrors.h
https://bugs.webkit.org/show_bug.cgi?id=187339
This header is not used anymore.
Patch by Rob Buis <rbuis@igalia.com> on 2018-07-05
Reviewed by Frédéric Wang.
- NetworkProcess/Downloads/soup/DownloadSoupErrors.h: Removed.
- NetworkProcess/soup/NetworkDataTaskSoup.cpp:
- PlatformGTK.cmake:
- 6:43 AM Changeset in webkit [233517] by
-
- 2 edits in trunk/Tools
[Flatpak] Add gst-libav to dependencies
https://bugs.webkit.org/show_bug.cgi?id=187340
Reviewed by Žan Doberšek.
- flatpak/org.webkit.WebKit.yaml: Use tarballs instead of git
clones because the FFMPeg git server doesn't support shallow
clones.
- 5:41 AM Changeset in webkit [233516] by
-
- 4 edits in trunk/Source/WebDriver
[GTK][WPE] Add --host option to WebDriver process
https://bugs.webkit.org/show_bug.cgi?id=187288
Reviewed by Žan Doberšek.
We currently allow to pass a port number and the server always listens for connections on localhost. We should
allow to optionally pass a host IP address to be able to use WebDriver remotely.
- HTTPServer.h:
- WebDriverService.cpp:
(WebDriver::printUsageStatement): Update usage to include --host option.
(WebDriver::WebDriverService::run): Parse --host option and pass it to HTTPServer::listen().
- soup/HTTPServerSoup.cpp:
(WebDriver::soupServerListen): Helper to call soup_server_listen_local(), soup_server_listen_all() or
soup_server_listen() depending on the given host.
(WebDriver::HTTPServer::listen): Use soupServerListen() helper.
- 4:42 AM Changeset in webkit [233515] by
-
- 3 edits in trunk/Source/WebCore
[WebVR] Fix VRDisplayEvent implementation
https://bugs.webkit.org/show_bug.cgi?id=187337
Reviewed by Žan Doberšek.
VRDisplayEvent implementation had two issues. First there were no attributes storing the
VRDisplay the event refers to. Secondly the VRDisplayEventReason is not mandatory so it
should be handled via an optional.
- Modules/webvr/VRDisplayEvent.cpp:
(WebCore::VRDisplayEvent::VRDisplayEvent):
(WebCore::VRDisplayEvent::display const): Deleted.
(WebCore::VRDisplayEvent::reason const): Deleted.
- Modules/webvr/VRDisplayEvent.h:
- 1:07 AM Changeset in webkit [233514] by
-
- 2 edits in trunk/Tools
[WPE][GTK]: Call gtk platform 'gtk' by default as this is what PortFactory expects
https://bugs.webkit.org/show_bug.cgi?id=187322
Patch by Thibault Saunier <tsaunier@igalia.com> on 2018-07-05
Reviewed by Žan Doberšek.
- flatpak/flatpakutils.py:
(WebkitFlatpak.init):
- 12:40 AM Changeset in webkit [233513] by
-
- 2 edits in trunk/Tools
Unreviewed, rolling out r233511.
Proper fix provided in bug #187322
Reverted changeset:
"Unreviewed. Fixing a call in flatpakutils.py to
PortFactory.get(),"
https://trac.webkit.org/changeset/233511
Jul 4, 2018:
- 11:54 PM Changeset in webkit [233512] by
-
- 11 edits2 adds in trunk
REGRESSION(r233325): [GTK] Broke 40 animations tests
https://bugs.webkit.org/show_bug.cgi?id=187264
Reviewed by Žan Doberšek.
Source/WebCore:
This is because DocumentAnimationScheduler uses DisplayRefreshMonitor, that is only used by GTK port when
accelerated compositing is enabled. Add a GTK display refresh monitor implementation to be used as the default.
- PlatformGTK.cmake: Add new file to compilation.
- platform/graphics/DisplayRefreshMonitor.cpp:
(WebCore::DisplayRefreshMonitor::createDefaultDisplayRefreshMonitor): Create DisplayRefreshMonitorGtk in GTK.
- platform/graphics/gtk/DisplayRefreshMonitorGtk.cpp: Added.
(WebCore::DisplayRefreshMonitorGtk::DisplayRefreshMonitorGtk):
(WebCore::DisplayRefreshMonitorGtk::~DisplayRefreshMonitorGtk):
(WebCore::onFrameClockUpdate):
(WebCore::DisplayRefreshMonitorGtk::requestRefreshCallback):
(WebCore::DisplayRefreshMonitorGtk::displayLinkFired):
- platform/graphics/gtk/DisplayRefreshMonitorGtk.h: Added.
(WebCore::DisplayRefreshMonitorGtk::create):
Source/WebKit:
Use a different PlatformDisplayID for compositing in the GTK port. This way, when entering AC a window screen
changed is emitted and the GTK default display refresh monitor is destroyed, ensuring the one created by the
threaded compositor is used.
- Shared/CoordinatedGraphics/threadedcompositor/ThreadedCompositor.cpp:
(WebKit::ThreadedCompositor::create): Remove unused WebPage parameter and add PlatformDisplayID.
(WebKit::ThreadedCompositor::ThreadedCompositor): Create the display refresh monitor with the given displayID.
- Shared/CoordinatedGraphics/threadedcompositor/ThreadedCompositor.h:
- Shared/CoordinatedGraphics/threadedcompositor/ThreadedDisplayRefreshMonitor.cpp:
(WebKit::ThreadedDisplayRefreshMonitor::ThreadedDisplayRefreshMonitor): Pass the given displayID to the parent constructor.
- Shared/CoordinatedGraphics/threadedcompositor/ThreadedDisplayRefreshMonitor.h:
(WebKit::ThreadedDisplayRefreshMonitor::create): Add PlatformDisplayID parameter.
- WebProcess/WebPage/CoordinatedGraphics/ThreadedCoordinatedLayerTreeHost.cpp:
(WebKit::ThreadedCoordinatedLayerTreeHost::ThreadedCoordinatedLayerTreeHost): Use compositingDisplayID when
creating the threaded compositor and notify about the window screen change.
(WebKit::ThreadedCoordinatedLayerTreeHost::setIsDiscardable): Use primaryDisplayID when leaving AC and
compositingDisplayID when re-entering.
LayoutTests:
Remove expectations for tests that are passing now.
- platform/gtk/TestExpectations:
- 10:48 PM Changeset in webkit [233511] by
-
- 2 edits in trunk/Tools
Unreviewed. Fixing a call in flatpakutils.py to PortFactory.get(),
enforcing the passed platform argument to always be lowercase, as is
expected from the PortFactory code.
- flatpak/flatpakutils.py:
(WebkitFlatpak.clean_args):
- 8:50 PM Changeset in webkit [233510] by
-
- 7 edits2 adds in trunk
Reparse user stylesheets when _useSystemAppearance changes
https://bugs.webkit.org/show_bug.cgi?id=187312
<rdar://problem/38565834>
Reviewed by Tim Horton.
Source/WebCore:
This setting may affect user stylesheet parsing. Reparse if it changes.
Test: fast/media/use-system-appearance-user-stylesheet-parsing.html
- page/Page.cpp:
(WebCore::Page::setUseSystemAppearance):
- page/Page.h:
(WebCore::Page::setUseSystemAppearance): Deleted.
- testing/Internals.cpp:
(WebCore::Internals::setUseSystemAppearance):
- testing/Internals.h:
- testing/Internals.idl:
LayoutTests:
- fast/media/use-system-appearance-user-stylesheet-parsing-expected.txt: Added.
- fast/media/use-system-appearance-user-stylesheet-parsing.html: Added.
- 1:12 PM Changeset in webkit [233509] by
-
- 2 edits in trunk/Source/WebCore
RealtimeIncomingVideoSources: Call stop() directly in the destructor
https://bugs.webkit.org/show_bug.cgi?id=187303
Patch by Thibault Saunier <tsaunier@igalia.com> on 2018-07-04
Reviewed by Youenn Fablet.
Avoiding to call stopProducingData when we are not actually producing data.
This fixes tests that will run once the WPE/GTK PeerConnection implementation is merged.
- platform/mediastream/RealtimeIncomingVideoSource.h:
(WebCore::RealtimeIncomingVideoSource::~RealtimeIncomingVideoSource):
- 1:08 PM Changeset in webkit [233508] by
-
- 2 edits in trunk/Source/WebCore
[GTK][WPE] Stop pushing the StreamCollection as event in GStreamerMediaStreamSource
https://bugs.webkit.org/show_bug.cgi?id=187304
Patch by Thibault Saunier <tsaunier@igalia.com> on 2018-07-04
Reviewed by Philippe Normand.
We let parsebin build the StreamCollection for us and pass extra metadata through the
TAG_EVENTpushing the collection ourself is not bringing anything but adds extra work in the
MediaPlayerGStreamer duplicating the STREAM_COLLECTION message for no good reason.
MediaStream layout tests already test that code path.
- platform/mediastream/gstreamer/GStreamerMediaStreamSource.cpp:
(WebCore::webkitMediaStreamSrcPadProbeCb):
- 11:45 AM Changeset in webkit [233507] by
-
- 3 edits in trunk/LayoutTests
Unreviewed GTK test gardening.
https://bugs.webkit.org/show_bug.cgi?id=187328
- platform/gtk/TestExpectations:
- platform/wpe/TestExpectations:
- 8:47 AM Changeset in webkit [233506] by
-
- 2 edits in trunk/Source/ThirdParty/libwebrtc
[libwebrtc] Allow IP mismatch for local connections on localhost
https://bugs.webkit.org/show_bug.cgi?id=187302
Patch by Thibault Saunier <tsaunier@igalia.com> on 2018-07-04
Reviewed by Youenn Fablet.
The rest of the code allows it, but there was an unecessary assert
- Source/webrtc/p2p/base/tcpport.cc:
- 7:17 AM Changeset in webkit [233505] by
-
- 9 copies1 add in releases/Apple/Safari Technology Preview 60
Added a tag for Safari Technology Preview release 60.
- 5:54 AM Changeset in webkit [233504] by
-
- 2 edits in trunk/Source/WTF
FunctionTraits: Make cCallArity() constant on 32-bits.
https://bugs.webkit.org/show_bug.cgi?id=187292
Reviewed by Yusuke Suzuki.
On X86, in Source/JavaScriptCore/jit/CCallHelpers.h we have a
static_assert that uses cCallArity(), so it needs to be constant to
avoid a compilation error. This is achieved by changing an ASSERT into
a static_assert.
- wtf/FunctionTraits.h:
(WTF::slotsForCCallArgument):
- 2:49 AM Changeset in webkit [233503] by
-
- 2 edits in trunk/Source/WebCore
[WebIDL] Remove custom bindings for HTMLDocument
https://bugs.webkit.org/show_bug.cgi?id=173444
Unreviewed, add missing bugzilla link for FIXME comments.
Patch by Frederic Wang <fwang@igalia.com> on 2018-07-04
- dom/Document.cpp:
(WebCore::Document::closeForBindings):
(WebCore::Document::write):
(WebCore::Document::writeln):
- 2:43 AM Changeset in webkit [233502] by
-
- 2 edits in trunk/Source/WebKit
A WKWebView in a UIViewController that is presented modally and uses a file picker will be incorrectly dismissed by the system
https://bugs.webkit.org/show_bug.cgi?id=185257
<rdar://problem/40819252>
Patch by Olivia Barnett <obarnett@apple.com> on 2018-07-04
Reviewed by Tim Horton.
Call to dismiss the presented modal was being called on the parent and not the child, which caused the file picker to be incorrectly dismissed along with the modal. A test for this bug-fix could not be completed without additional tools as TestWebKitAPI is not a UI application.
- UIProcess/ios/forms/WKFileUploadPanel.mm:
(-[WKFileUploadPanel _dismissDisplayAnimated:]):
- 2:06 AM Changeset in webkit [233501] by
-
- 67 edits in trunk
Introduce PLATFORM(IOSMAC)
https://bugs.webkit.org/show_bug.cgi?id=187315
Reviewed by Dan Bernstein.
Source/bmalloc:
- Configurations/Base.xcconfig:
Source/JavaScriptCore:
- Configurations/Base.xcconfig:
- Configurations/FeatureDefines.xcconfig:
Source/ThirdParty/ANGLE:
- Configurations/Base.xcconfig:
Source/WebCore:
- Configurations/Base.xcconfig:
- Configurations/FeatureDefines.xcconfig:
- Configurations/WebCore.xcconfig:
- Modules/geolocation/ios/GeolocationPositionIOS.mm:
(WebCore::GeolocationPosition::GeolocationPosition):
- page/cocoa/MemoryReleaseCocoa.mm:
(WebCore::platformReleaseMemory):
- platform/audio/ios/AudioSessionIOS.mm:
(WebCore::AudioSession::setCategory):
(WebCore::AudioSession::routingContextUID const):
- platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::createAVPlayer):
(WebCore::MediaPlayerPrivateAVFoundationObjC::setShouldDisableSleep):
- platform/ios/PlatformPasteboardIOS.mm:
(WebCore::registerItemToPasteboard):
- platform/ios/WebItemProviderPasteboard.mm:
(-[WebItemProviderRegistrationInfoList itemProvider]):
Source/WebCore/PAL:
- Configurations/Base.xcconfig:
- Configurations/FeatureDefines.xcconfig:
- pal/cf/CoreMediaSoftLink.cpp:
- pal/cf/CoreMediaSoftLink.h:
- pal/spi/mac/AVFoundationSPI.h:
Source/WebInspectorUI:
- Configurations/Base.xcconfig:
Source/WebKit:
- Configurations/Base.xcconfig:
- Configurations/FeatureDefines.xcconfig:
- Configurations/NetworkService.xcconfig:
- Configurations/WebContentService.xcconfig:
- Configurations/WebKit.xcconfig:
- NetworkProcess/cocoa/NetworkProcessCocoa.mm:
(WebKit::NetworkProcess::sourceApplicationAuditData const):
- NetworkProcess/ios/NetworkProcessIOS.mm:
- NetworkProcess/mac/NetworkProcessMac.mm:
(WebKit::NetworkProcess::initializeProcessName):
(WebKit::overrideSystemProxies):
- Platform/mac/LayerHostingContext.mm:
(WebKit::LayerHostingContext::createForExternalHostingProcess):
- Platform/spi/ios/UIKitSPI.h:
- Shared/ios/ChildProcessIOS.mm:
(WebKit::ChildProcess::initializeSandbox):
- Shared/ios/InteractionInformationAtPosition.h:
- Shared/ios/InteractionInformationAtPosition.mm:
(WebKit::InteractionInformationAtPosition::encode const):
(WebKit::InteractionInformationAtPosition::decode):
- Shared/mac/ChildProcessMac.mm:
(WebKit::ChildProcess::setApplicationIsDaemon):
(WebKit::enableSandboxStyleFileQuarantine):
- Shared/mac/CodeSigning.mm:
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _webViewPrintFormatter]):
- UIProcess/API/Cocoa/WKWebViewInternal.h:
- UIProcess/RemoteLayerTree/RemoteLayerTreeHost.mm:
(WebKit::RemoteLayerTreeHost::updateLayerTree):
- UIProcess/RemoteLayerTree/ios/RemoteLayerTreeHostIOS.mm:
(-[WKRemoteView initWithFrame:contextID:]):
- UIProcess/_WKWebViewPrintFormatter.mm:
- UIProcess/_WKWebViewPrintFormatterInternal.h:
- UIProcess/ios/WKContentView.mm:
(-[WKContentView _didExitStableState]):
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView gestureRecognizer:canBePreventedByGestureRecognizer:]):
(-[WKContentView gestureRecognizer:shouldRecognizeSimultaneouslyWithGestureRecognizer:]):
(-[WKContentView canPerformActionForWebView:withSender:]):
(-[WKContentView _defineForWebView:]):
(-[WKContentView setSelectedTextRange:]):
(-[WKContentView closestPositionToPoint:]):
(-[WKContentView _updateChangedSelection:]):
- UIProcess/ios/WKLegacyPDFView.mm:
- UIProcess/ios/WKPDFView.mm:
- WebKit.xcodeproj/project.pbxproj:
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::selectWithGesture):
(WebKit::WebPage::getPositionInformation):
- WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::initializeSandbox):
- config.h:
Source/WebKitLegacy/ios:
- Misc/WebGeolocationCoreLocationProvider.mm:
(-[WebGeolocationCoreLocationProvider requestGeolocationAuthorization]):
Source/WebKitLegacy/mac:
- Configurations/Base.xcconfig:
- Configurations/FeatureDefines.xcconfig:
Source/WTF:
- Configurations/Base.xcconfig:
- wtf/FeatureDefines.h:
- wtf/Platform.h:
Tools:
- TestWebKitAPI/Configurations/Base.xcconfig:
- TestWebKitAPI/Configurations/FeatureDefines.xcconfig:
- 12:58 AM Changeset in webkit [233500] by
-
- 7 edits in trunk
[Web Animations] Make WPT test at timing-model/timelines/timelines.html pass reliably
https://bugs.webkit.org/show_bug.cgi?id=186508
<rdar://problem/41000260>
Reviewed by Dean Jackson.
LayoutTests/imported/w3c:
Record WPT test progressions.
- web-platform-tests/web-animations/interfaces/Animation/finished-expected.txt:
- web-platform-tests/web-animations/timing-model/timelines/timelines-expected.txt:
Source/WebCore:
The remaining failing assertion in this test assumes that the ready promise for an animation is fired
prior to requestAnimationFrame callbacks being fired in the current event loop such that registering
the requestAnimationFrame() call within the ready promise has its callback fired with the same timeline
time. To ensure that is true, we perform a microtask checkpoint as soon as we're done running pending
tasks, which will resolve the ready promise for any play-pending animations.
- animation/DocumentTimeline.cpp:
(WebCore::DocumentTimeline::updateAnimations):
LayoutTests:
We're now passing this test reliably.
- 12:43 AM Changeset in webkit [233499] by
-
- 2 edits in trunk/Source/WebKit
50 failing WebDriver tests after r233417
https://bugs.webkit.org/show_bug.cgi?id=187263
Patch by Carlos Garcia Campos <cgarcia@igalia.com> on 2018-07-04
Reviewed by Frédéric Wang.
A std::optional value is used while being std::nullopt. This is because the wrong input source is used when
transitioning keyboard input states.
- UIProcess/Automation/SimulatedInputDispatcher.cpp:
(WebKit::SimulatedInputDispatcher::transitionInputSourceToState):
