Timeline

Sep 8, 2020:

10:14 PM Changeset in webkit [266775] by Lauro Moura

• 2 edits in trunk/Tools

kill-old-processes: Remove stale Apache2 shm segments
​https://bugs.webkit.org/show_bug.cgi?id=216068

Reviewed by Alexey Proskuryakov.

Based on original patch by Diego Pino Garcia.

When Apache2 terminates abruptely, it may not free up allocated shared
memory segments. When later Apache2 is started again, if its pid
matches the pid of orphan shm segments it fails believing there's
already an Apache2 instance running. To avoid this situation, after
killing Apache2 we check for potential orphaned shm segments and
remove them.

If any error happens during the cleanup, log the failure and continue
the script, addressing the issue that caused the initial revert in
r266536.

• BuildSlaveSupport/kill-old-processes:

(removeOrphanShmSegments):
(orphanedShmSegmentsByUser):
(main):

9:37 PM Changeset in webkit [266774] by Fujii Hironori

• 3 edits in trunk/Tools

[WinCairo] TestWTF.WTF.UniqueRef is crashing in Debug builds
​https://bugs.webkit.org/show_bug.cgi?id=216303

Reviewed by Ryosuke Niwa.

The class A of UniqueRef.cpp and CompactUniquePtrTuple.cpp are
conflicting. CRT heap threw an exception due to allocator
mismatching while deleting class A object.

• TestWebKitAPI/Tests/WTF/CompactUniquePtrTuple.cpp: Added a new namespace CompactUniquePtrTupleTest.
• TestWebKitAPI/Tests/WTF/UniqueRef.cpp: Added a new namespace UniqueRefTest.

(TestWebKitAPI::function): Deleted an unused function.

9:32 PM Changeset in webkit [266773] by Ryan Haddad

• 2 edits in trunk/LayoutTests

Unreviewed test gardening after r266761.

• platform/mac/TestExpectations: This should have been Mojave, not HighSierra.

9:27 PM Changeset in webkit [266772] by timothy_horton@apple.com

• 3 edits in trunk/LayoutTests

[ iOS wk2 ] fast/attachment/attachment-thumbnail-preserves-aspect-ratio.html is a constant failure
​https://bugs.webkit.org/show_bug.cgi?id=216300
<rdar://problem/68542670>

Reviewed by Wenson Hsieh.

• fast/attachment/attachment-thumbnail-preserves-aspect-ratio.html:

shouldBecomeDifferent only works if both of the strings it evals are evaluable in the global scope.

• platform/ios-wk2/TestExpectations:

Remove the failure expectation.

9:12 PM Changeset in webkit [266771] by Chris Dumez

• 9 edits in trunk/Source/WebKit

Regression(r260614) Power usage has increased due to extra thread hopping
​https://bugs.webkit.org/show_bug.cgi?id=216296
<rdar://problem/67719299>

Reviewed by Simon Fraser.

Power usage has increased after r260614 due to extra thread hopping. To recover,
we now process the DisplayWasRefreshed IPC to a background queue if there is
scrolling going on and responsiveness is thus critical. In the common case, where
the user is not scrolling, we keep processing the IPC on the main thread, like
before r260614. This avoids extra thread hopping and saves power.

• UIProcess/WebPageProxy.cpp:

(WebKit::ScrollingObserver::willSendWheelEvent):
(WebKit::ScrollingObserver::ScrollingObserver):
(WebKit::ScrollingObserver::singleton):
(WebKit::WebPageProxy::sendWheelEvent):

• UIProcess/mac/DisplayLink.cpp:

(WebKit::DisplayLink::displayLinkCallback):

• UIProcess/mac/DisplayLink.h:

(WebKit::DisplayLink::setShouldSendIPCOnBackgroundQueue):

• WebProcess/WebPage/EventDispatcher.cpp:

(WebKit::EventDispatcher::notifyScrollingTreesDisplayWasRefreshed):
(WebKit::EventDispatcher::displayWasRefreshed):

• WebProcess/WebPage/EventDispatcher.h:
• WebProcess/WebProcess.cpp:

(WebKit::WebProcess::displayWasRefreshed):

• WebProcess/WebProcess.h:
• WebProcess/WebProcess.messages.in:

8:48 PM Changeset in webkit [266770] by ysuzuki@apple.com

• 7 edits
  1 add in trunk

[JSC] returnEarlyFromInfiniteLoopsForFuzzing should return object
​https://bugs.webkit.org/show_bug.cgi?id=216289
<rdar://problem/68496533>

Reviewed by Saam Barati.

JSTests:

• stress/construct-return-early-from-infinite-loop-for-fuzzer.js: Added.

(foo):

Source/JavaScriptCore:

When returning early with returnEarlyFromInfiniteLoopsForFuzzing, we are returning with undefined.
But this is wrong when the callee is constructor since constructor is strongly assumed that it returns an object.
We should return some object from returnEarlyFromInfiniteLoopsForFuzzing. In this patch, we return global object
associated to this callee instead of undefined

• bytecode/CodeBlock.cpp:

(JSC::CodeBlock::finishCreation):
(JSC::CodeBlock::~CodeBlock):

• dfg/DFGSpeculativeJIT64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

• ftl/FTLLowerDFGToB3.cpp:

(JSC::FTL::DFG::LowerDFGToB3::compileLoopHint):

• jit/JITOpcodes.cpp:

(JSC::JIT::emit_op_loop_hint):

• llint/LowLevelInterpreter64.asm:

8:32 PM Changeset in webkit [266769] by rniwa@webkit.org

• 9 edits in trunk/Source/WebCore

Having an iframe as a descendent node shouldn't require ElementRareData
​https://bugs.webkit.org/show_bug.cgi?id=216264

Reviewed by Darin Adler.

Store the number of connected frames in the descendent nodes directly in Node using CompactUniquePtrTuple
in the same space as where we store the NodeRareData pointer. This avoids creating ElementRareData on every
ancestor element and shadow host / document of an iframe.

Also moved TabIndexState there to simply NodeFlags, and created CustomElementState to replace the existing
flags in NodeFlags to match the latest terminology used in the specification:
​https://dom.spec.whatwg.org/#concept-element-custom-element-state

CustomElementState has four states: "uncustomized" (default; builtin elements), "undefined" (i.e. element has
a valid custom element but it hasn't been defined or upgraded yet), "custom" (a valid custom element instance),
and "failed" (upgrading has resulted in an error). Before this patch, "uncustomized" meant that neither
IsCustomElement nor IsEditingTextOrUndefinedCustomElementFlag is set, "undefined" had IsCustomElement
and IsEditingTextOrUndefinedCustomElementFlag set, and "custom" had IsCustomElement set but
IsEditingTextOrUndefinedCustomElementFlag unset whereas "failed" had the opposite.

No new tests since there should be no observable behavioral change.

• cssjit/SelectorCompiler.cpp:

(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsNthChild): Mask out the bits stored in
the pointer in the 64-bit architecture. In the 32-bit architecture, the pointer is the first component
without any extra bits stored in it.

• dom/Element.cpp:

(WebCore::Node::setTabIndexState): Moved from Node.h. Updated to use the newly introduced bitfields.
(WebCore::Node::setCustomElementState): Added.
(WebCore::Element::setIsDefinedCustomElement): Now updates CustomElementState in the newly added bitfields.
(WebCore::Element::setIsFailedCustomElementWithoutClearingReactionQueue): Ditto.
(WebCore::Element::setIsCustomElementUpgradeCandidate): Ditto.

• dom/ElementRareData.cpp:
• dom/ElementRareData.h:

(WebCore::ElementRareData): Moved m_unusualTabIndex and m_childIndex to NodeRareData for better packing.

• dom/Node.cpp:

(WebCore::Node::materializeRareData):
(WebCore::Node::clearRareData):
(WebCore::Node::connectedSubframeCount const): Moved to the header file.
(WebCore::Node::incrementConnectedSubframeCount): Now updates the newly added bitfields.
(WebCore::Node::decrementConnectedSubframeCount): Ditto.

• dom/Node.h:

(WebCore::Node::isUndefinedCustomElement const): Now uses CustomElementState in the newly added bitfields.
(WebCore::Node::isCustomElementUpgradeCandidate const): Ditto.
(WebCore::Node::isDefinedCustomElement const): Ditto.
(WebCore::Node::isFailedCustomElement const): Ditto.
(WebCore::Node::isEditingText const):
(WebCore::Node::connectedSubframeCount const): Moved here from cpp file.
(WebCore::Node::rareDataMemoryOffset):
(WebCore::Node::rareDataPointerMask): Added.
(WebCore::Node::CustomElementState): Added.
(WebCore::Node::RareDataBitFields): Added.
(WebCore::Node::rareDataBitfields const): Added.
(WebCore::Node::setRareDataBitfields): Added.
(WebCore::Node::tabIndexState const): Updated to use rareDataBitfields.
(WebCore::Node::setTabIndexState): Moved to Element.cpp.
(WebCore::Node::customElementState const): Added.
(WebCore::Node::hasRareData const):
(WebCore::Node::rareData const):

• dom/NodeRareData.cpp:
• dom/NodeRareData.h:

(WebCore::NodeRareData::UseType): Removed ConnectedFrameCount.
(WebCore::NodeRareData::NodeRareData):
(WebCore::NodeRareData::useTypes const):
(WebCore::NodeRareData::connectedSubframeCount const): Deleted.
(WebCore::NodeRareData::incrementConnectedSubframeCount): Deleted.
(WebCore::NodeRareData::decrementConnectedSubframeCount): Deleted.
(WebCore::NodeRareData): Moved m_unusualTabIndex and m_childIndex here for better packing in 64-bit architecture.

8:14 PM Changeset in webkit [266768] by timothy_horton@apple.com

• 2 edits in trunk/Source/WebCore

WebCore UnifiedSource81 builds for upwards of 10 minutes
​https://bugs.webkit.org/show_bug.cgi?id=216297

Reviewed by Simon Fraser.

• bindings/js/WebCoreBuiltinNames.h:

(WebCore::WebCoreBuiltinNames::WebCoreBuiltinNames):
Work around an LLVM regression by disabling debugging in the
WebCoreBuiltinNames constructor. This is tracked by <rdar://68246686>.

This drops the WebCore release build time on my iMac Pro from 14 minutes to 6.

6:35 PM Changeset in webkit [266767] by Hector Lopez

• 2 edits in trunk/LayoutTests

[ iOS wk2 ] fast/attachment/attachment-thumbnail-preserves-aspect-ratio.html is a constant failure
​https://bugs.webkit.org/show_bug.cgi?id=216300

Unreviewed test gardening.

• platform/ios-wk2/TestExpectations:

6:00 PM Changeset in webkit [266766] by Megan Gardner

• 2 edits in trunk/Source/WebKit

Removed unneeded respondsToSelector checks on UIWKTextInteractionAssistant in WKContentViewInteraction.
​https://bugs.webkit.org/show_bug.cgi?id=216204

Reviewed by Wenson Hsieh.

Over the years, we have had many selection and text interaction assistants, that may not have all
responded to the selectors we expect them to. This has left many respondsToSelector checks around that
should be removed, as they are unnecessary and an unneeded performance hit.

No behavior change, covered by existing tests.

• UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView _willStartScrollingOrZooming]):
(-[WKContentView _didEndScrollingOrZooming]):

5:52 PM Changeset in webkit [266765] by commit-queue@webkit.org

• 2 edits in trunk/LayoutTests

[ macOS iOS ] svg/animations/smil-leak-element-instances-noBaseValRef.svg is a flaky failure
​https://bugs.webkit.org/show_bug.cgi?id=215353

Patch by Said Abou-Hallawa <​sabouhallawa@apple.com> on 2020-09-08
Reviewed by Geoffrey Garen.

This patch reverts r265780 and goes in the opposite direction: allocating
lots and lots of objects, then proving that the number of live objects
did not grow a lot.

GCController can be conservative sometimes and leaves some objects here
and there. So we should not be checking that *all* the objects were freed
to verify there was no leak in allocating the SVG properties objects.

Instead we can check for an end-to-end test the memory did not grow a lot
despite the huge number of allocations.

• svg/animations/smil-leak-element-instances-noBaseValRef.svg:

5:41 PM Changeset in webkit [266764] by Hector Lopez

• 3 edits in trunk/LayoutTests

[ macOS iOS ] imported/w3c/web-platform-tests/IndexedDB/blob-valid-after-deletion.any.html is a flaky failure
​https://bugs.webkit.org/show_bug.cgi?id=216298

Unreviewed test gardening.

• platform/ios-wk2/TestExpectations:
• platform/mac/TestExpectations:

5:23 PM Changeset in webkit [266763] by Chris Dumez

• 4 edits
  2 adds in trunk

Drop non-standard name / units attributes from AudioParam
​https://bugs.webkit.org/show_bug.cgi?id=216290

Reviewed by Geoffrey Garen.

Source/WebCore:

Drop non-standard name / units attributes from AudioParam:

• ​https://www.w3.org/TR/webaudio/#AudioParam

Blink does not have them either.

Test: webaudio/audioparam-obsolete-api.html

• Modules/webaudio/AudioParam.h:
• Modules/webaudio/AudioParam.idl:

LayoutTests:

Add layout test coverage.

• webaudio/audioparam-obsolete-api-expected.txt: Added.
• webaudio/audioparam-obsolete-api.html: Added.

5:16 PM Changeset in webkit [266762] by Chris Dumez

• 3 edits in trunk/Source/WebCore

Drop unnecessary AudioParam::m_smoothingConstant data member
​https://bugs.webkit.org/show_bug.cgi?id=216291

Reviewed by Geoffrey Garen.

Drop unnecessary AudioParam::m_smoothingConstant data member. This constant is never changed.

No new tests, no web-facing behavior change.

• Modules/webaudio/AudioParam.cpp:

(WebCore::AudioParam::AudioParam):
(WebCore::AudioParam::smooth):

• Modules/webaudio/AudioParam.h:

4:47 PM Changeset in webkit [266761] by Ryan Haddad

• 2 edits in trunk/LayoutTests

Update imported/w3c/web-platform-tests/css/css-text/line-break/line-break-{strict,normal}-015.xht
​https://bugs.webkit.org/show_bug.cgi?id=209250

Unreviewed test gardening.

• platform/mac/TestExpectations: Limit image failure expectation to High Sierra & Catalina.

4:41 PM Changeset in webkit [266760] by Ryan Haddad

• 2 edits in trunk/LayoutTests

[ Mac ] fast/history/page-cache-webdatabase-pending-transaction.html is timing out
​https://bugs.webkit.org/show_bug.cgi?id=207010

Unreviewed test gardening.

• platform/mac/TestExpectations: The test isn't consistently timing out,

so mark it as flaky instead.

4:39 PM Changeset in webkit [266759] by Hector Lopez

• 2 edits in trunk/LayoutTests

[ macOS wk2 Release ] imported/w3c/web-platform-tests/css/css-flexbox/quirks-auto-block-size-with-percentage-item.html is a flaky failure
​https://bugs.webkit.org/show_bug.cgi?id=216292

Unreviewed test gardening.

• platform/mac-wk2/TestExpectations:

4:32 PM Changeset in webkit [266758] by Ryan Haddad

• 4 edits in trunk/LayoutTests

http/tests/history/back-with-fragment-change.php fails
​https://bugs.webkit.org/show_bug.cgi?id=68278

Unreviewed test gardening.

• platform/mac-wk1/TestExpectations: Mark test as consistently failing for macOS WK1.
• platform/mac-wk2/TestExpectations: Mark test as flaky for macOS WK2.
• platform/mac/TestExpectations: Remove general Mac expectation.

4:03 PM Changeset in webkit [266757] by Lauro Moura

• 2 edits in trunk/Source/WebKit

NetworkProcess: Check for invalid handle before writing salt file
​https://bugs.webkit.org/show_bug.cgi?id=216285

Reviewed by Sihui Liu.

Covered by existing tests.

Fixes crashes in GTK/WPE after r266742.

• NetworkProcess/cache/NetworkCacheData.cpp:

(WebKit::NetworkCache::readOrMakeSalt):

3:58 PM Changeset in webkit [266756] by beidson@apple.com

• 2 edits in trunk/Source/WebKit

Crashtracer inside PDFPlugin::createScrollbar.
<rdar://problem/68474062> and ​https://bugs.webkit.org/show_bug.cgi?id=216286

Reviewed by Tim Horton.

To quote Tim from r264945:
No new tests; timing is such that I can't reproduce without inserting
intentional delays into the main thread hops, which is further than
I'm willing to go for a test.

This is a speculative fix due to the aforementioned reproducibility issue.

• WebProcess/Plugins/PDF/PDFPlugin.mm:

(WebKit::PDFPlugin::createScrollbar): Plugin hasn't been destroyed, and there's still a PluginView,

but the PluginView isn't in a frame. So null check that.

3:09 PM Changeset in webkit [266755] by Alan Coon

• 1 copy in tags/Safari-610.1.28.10.1

Tag Safari-610.1.28.10.1.

2:51 PM Changeset in webkit [266754] by Chris Dumez

• 19 edits in trunk/LayoutTests/imported/w3c

Unreviewed, rebaseline flaky webaudio WPT tests.

• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/k-rate-biquad-connection-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/k-rate-biquad-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/k-rate-delay-connections-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/k-rate-delay-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/k-rate-gain-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/k-rate-stereo-panner-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-biquadfilternode-interface/no-dezippering-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-constantsourcenode-interface/constant-source-output-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-convolvernode-interface/realtime-conv-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-delaynode-interface/no-dezippering-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-gainnode-interface/no-dezippering-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-iirfilternode-interface/ctor-iirfilter-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-oscillatornode-interface/detune-limiting-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-pannernode-interface/automation-changes-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-pannernode-interface/panner-distance-clamping-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-pannernode-interface/panner-equalpower-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-scriptprocessornode-interface/simple-input-output-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-stereopanner-interface/no-dezippering-expected.txt:

2:46 PM Changeset in webkit [266753] by weinig@apple.com

• 40 edits
  1 move
  9 adds
  2 deletes in trunk/Source/WebCore

[WebIDL] Add support for partial interface mixins
​https://bugs.webkit.org/show_bug.cgi?id=216255

Reviewed by Darin Adler.

Add support for WebIDL 'partial interface mixins' and adopt them for a few
supplements of the DocumentOrShadowRoot mixin.

• CMakeLists.txt:
• DerivedSources.make:
• WebCore.xcodeproj/project.pbxproj:
• DerivedSources-input.xcfilelist:
• DerivedSources-output.xcfilelist:
• dom/DocumentOrShadowRoot.idl:
• css/DocumentOrShadowRootStyleSheets.idl: Added.
• dom/DocumentOrShadowRootPointerLock.idl: Added.
• Modules/pictureinpicture/DocumentOrShadowRootPictureInPicture.idl: Added.
• animation/DocumentOrShadowRootAnimations.idl: Added.

Add new split-out IDLs.

• bindings/scripts/CodeGenerator.pm:

(GenerateEmptyHeaderAndCpp):
Move empty file generation here (from generate-bindings.pl), allowing
for a simplified compilation model, where all IDLs given to generate-bindings.pl
now instantiate the CodeGenerator and it is the CodeGenerators job to figure
out if the file needs anything generated. This allows the dependencies file
to get out of the business of indicating that an IDL doesn't need generation,
and stick to just listing dependencies. This is important now that dependencies
like interface mixins can now have depencies of their own, partial interface mixins.

(ProcessInterfaces):
(ProcessDictionaries):
Move logic to avoid generation of mixins and partials here.

(ProcessInterfaceSupplementalDependencies):
Support recursively processing supplemental dependencies for partial interface
mixins. Ensure that we don't evaluate exposure when doing this, as that should
only happen on the final interface.

(shouldPropertyBeExposed):
Move all logic for exposure here, rather than splitting it up between here
and ProcessInterfaceSupplementalDependencies/ProcessDictionarySupplementalDependencies

• bindings/scripts/generate-bindings.pl:

(generateBindings):
(generateEmptyHeaderAndCpp): Deleted.
Remove support for unused additionalIdlFiles option and simplify things
by having CodeGenerator work out whether it should generate an empty set of
files based on the files themselves rather than inferring it from the dependency
information.

• bindings/scripts/preprocess-idls.pl:

(getPartialNamesFromIDL):
Add support for properly parsing partial interface mixins and stop disallowing
dependencies, like a mixin interface, from having its own dependences.

• bindings/scripts/test/JS/JSTestConditionalIncludes.cpp:
• bindings/scripts/test/JS/JSTestConditionalIncludes.h:
• bindings/scripts/test/JS/JSTestIncludes.cpp: Removed.
• bindings/scripts/test/JS/JSTestIncludes.h: Removed.
• bindings/scripts/test/JS/JSTestInterface.cpp:
• bindings/scripts/test/JS/JSTestInterface.h:
• bindings/scripts/test/JS/JSTestMixinInterface.cpp: Added.
• bindings/scripts/test/JS/JSTestMixinInterface.h: Added.
• bindings/scripts/test/JS/JSTestPartialMixinInterface.cpp: Added.
• bindings/scripts/test/JS/JSTestPartialMixinInterface.h: Added.
• bindings/scripts/test/SupplementalDependencies.dep:
• bindings/scripts/test/TestConditionalIncludes.idl:
• bindings/scripts/test/TestIncludes.idl: Removed.
• bindings/scripts/test/TestInterface.idl:
• bindings/scripts/test/TestMixinInterface.idl: Added.
• bindings/scripts/test/TestObj.idl:
• bindings/scripts/test/TestPartialMixinInterface.idl: Added.

Rename TestIncludes to TestMixinInterface (since that is what it is testing)
and add TestPartialMixinInterface to test partial interface mixins.

2:40 PM Changeset in webkit [266752] by Peng Liu

• 2 edits in trunk/Source/WebCore

Web process crashes at WebCore::HTMLMediaElement::prepareForVideoFullscreenStandby
​https://bugs.webkit.org/show_bug.cgi?id=216283

Reviewed by Alex Christensen.

• html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::prepareForVideoFullscreenStandby):

2:33 PM Changeset in webkit [266751] by Jonathan Bedard

• 3 edits in trunk/Tools

[webkitcorepy] Register another package in the WebKit repository (Follow-up fix)
​https://bugs.webkit.org/show_bug.cgi?id=216013
<rdar://problem/68097461>

Reviewed by Dewei Zhu.

• Scripts/libraries/webkitcorepy/webkitcorepy/init.py: Bump version.
• Scripts/libraries/webkitcorepy/webkitcorepy/autoinstall.py:

(AutoInstall.register): Use pypiname instead of the package name.

2:10 PM Changeset in webkit [266750] by commit-queue@webkit.org

• 2 edits in trunk/Source/WebKit

Null check m_layerHostingContext in TiledCoreAnimationDrawingArea::sendDidFirstLayerFlushIfNeeded
​https://bugs.webkit.org/show_bug.cgi?id=216282
<rdar://problem/64933878>

Patch by Alex Christensen <​achristensen@webkit.org> on 2020-09-08
Reviewed by Tim Horton.

There's no reason to dereference null here. Let's not crash.

• WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:

(WebKit::TiledCoreAnimationDrawingArea::sendDidFirstLayerFlushIfNeeded):

2:10 PM Changeset in webkit [266749] by achristensen@apple.com

• 2 edits in trunk/Source/WebCore

REGRESSION(r266681) [GTK] constexpr build failure in LTS/Debian with GCC 8
​https://bugs.webkit.org/show_bug.cgi?id=216280

Patch by Lauro Moura <Lauro Moura> on 2020-09-08
Reviewed by Darin Adler and Alex Christensen.

• platform/text/TextCodecSingleByte.cpp:

(WebCore::tableForDecoding):

2:08 PM Changeset in webkit [266748] by achristensen@apple.com

• 8 edits in trunk

new URL("#") should throw an error
​https://bugs.webkit.org/show_bug.cgi?id=216115

Reviewed by Yusuke Suzuki and Darin Adler.

Source/WebCore:

This aligns the DOM URL object with the specification and Firefox.
Covered by adding to fast/dom/DOMURL/url-constructor.html.

• html/DOMURL.cpp:

(WebCore::DOMURL::create):

• html/DOMURL.h:

LayoutTests:

• fast/dom/DOMURL/url-constructor-expected.txt:
• fast/dom/DOMURL/url-constructor.html:
• inspector/unit-tests/url-utilities.html:
• inspector/unit-tests/url-utilities-expected.txt:

1:32 PM Changeset in webkit [266747] by sbarati@apple.com

• 2 edits in trunk/Source/JavaScriptCore

re-enable TCSM on all OSs
​https://bugs.webkit.org/show_bug.cgi?id=216281

Reviewed by Tadeu Zagallo.

• runtime/Options.cpp:

(JSC::defaultTCSMValue):

1:24 PM Changeset in webkit [266746] by Chris Dumez

• 17 edits in trunk

Fix precision issues in AudioParamTimeline when event times are very close
​https://bugs.webkit.org/show_bug.cgi?id=216279

Reviewed by Darin Adler.

LayoutTests/imported/w3c:

• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/audioparam-close-expected.txt:

Rebaseline test that is now passing.

• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/audioparam-exponentialRampToValueAtTime-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/audioparam-linearRampToValueAtTime-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/k-rate-panner-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-pannernode-interface/panner-automation-position-expected.txt:

Rebaseline tests whose output has changed very slightly.

Source/WebCore:

Fix precision issues in AudioParamTimeline when event times are very close. To achieve this,
AudioParamTimeline::valuesForTimeRange() now uses frames for the range instead of time.
frames have the benefit of being more exact.

This change is based on audio_param_timeline.cc from Blink:

• ​https://github.com/chromium/chromium/blob/master/third_party/blink/renderer/modules/webaudio/audio_param_timeline.cc

No new tests, rebaselined existing test.

• Modules/webaudio/AudioParam.cpp:

(WebCore::AudioParam::calculateTimelineValues):

• Modules/webaudio/AudioParamTimeline.cpp:

(WebCore::AudioParamTimeline::valueForContextTime):
(WebCore::AudioParamTimeline::valuesForTimeRange):
(WebCore::AudioParamTimeline::valuesForTimeRangeImpl):
(WebCore::AudioParamTimeline::isEventCurrent const):

• Modules/webaudio/AudioParamTimeline.h:

LayoutTests:

Rebaseline tests whose output has changed very slightly.

• webaudio/audioparam-cancel-and-hold-expected.txt:
• webaudio/oscillator-custom-expected.wav:
• webaudio/oscillator-sawtooth-expected.wav:
• webaudio/oscillator-square-expected.wav:
• webaudio/oscillator-triangle-expected.wav:

1:24 PM Changeset in webkit [266745] by youenn@apple.com

• 12 edits in trunk/Source

Tighten checks when creating an audio buffer list
​https://bugs.webkit.org/show_bug.cgi?id=216237
<rdar://problem/68271376>

Reviewed by Geoffrey Garen.

Source/WebCore:

Add a routine to check there is no multiplication integer overflow.

• platform/audio/cocoa/WebAudioBufferList.cpp:

(WebCore::computeBufferSize):
(WebCore::WebAudioBufferList::isSupportedDescription):
(WebCore::WebAudioBufferList::setSampleCount):

• platform/audio/cocoa/WebAudioBufferList.h:

Source/WebKit:

Add message checks to verify that no message integer overflows happen when processing audio buffer list messages.

• GPUProcess/GPUConnectionToWebProcess.cpp:

(WebKit::GPUConnectionToWebProcess::audioTrackRendererManager):

• GPUProcess/webrtc/RemoteAudioMediaStreamTrackRenderer.cpp:

(WebKit::RemoteAudioMediaStreamTrackRenderer::RemoteAudioMediaStreamTrackRenderer):
(WebKit::RemoteAudioMediaStreamTrackRenderer::audioSamplesStorageChanged):
(WebKit::RemoteAudioMediaStreamTrackRenderer::audioSamplesAvailable):

• GPUProcess/webrtc/RemoteAudioMediaStreamTrackRenderer.h:
• GPUProcess/webrtc/RemoteAudioMediaStreamTrackRendererManager.cpp:

(WebKit::RemoteAudioMediaStreamTrackRendererManager::RemoteAudioMediaStreamTrackRendererManager):
(WebKit::RemoteAudioMediaStreamTrackRendererManager::createRenderer):

• GPUProcess/webrtc/RemoteAudioMediaStreamTrackRendererManager.h:
• GPUProcess/webrtc/RemoteMediaRecorder.cpp:

(WebKit::RemoteMediaRecorder::audioSamplesStorageChanged):
(WebKit::RemoteMediaRecorder::audioSamplesAvailable):

• GPUProcess/webrtc/RemoteMediaRecorder.h:
• WebProcess/cocoa/RemoteCaptureSampleManager.cpp:

(WebKit::RemoteCaptureSampleManager::RemoteAudio::audioSamplesAvailable):

12:49 PM Changeset in webkit [266744] by Fujii Hironori

• 2 edits in trunk/Tools

[Win] MiniBrowser URL bar can't insert a longer text than the window width
​https://bugs.webkit.org/show_bug.cgi?id=216252

Reviewed by Per Arne Vollan.

• MiniBrowser/win/MainWindow.cpp:

(MainWindow::createToolbar): Set ES_AUTOHSCROLL and for the URL
bar edit control. Removed ES_AUTOVSCROLL for it since it doesn't
scroll vertically.

12:47 PM Changeset in webkit [266743] by timothy_horton@apple.com

• 20 edits
  3 adds in trunk

iOS: <attachment>'s QuickLook thumbnails can appear squished
​https://bugs.webkit.org/show_bug.cgi?id=216209
<rdar://problem/67817706>

Reviewed by Wenson Hsieh.

Source/WebCore:

Test: fast/attachment/attachment-thumbnail-preserves-aspect-ratio.html

• html/HTMLAttachmentElement.idl:
• testing/Internals.cpp:

(WebCore::Internals::attachmentThumbnailInfo):

• testing/Internals.h:
• testing/Internals.idl:

Expose the attachment thumbnail size via Internals.

• rendering/RenderThemeIOS.mm:

(WebCore::RenderAttachmentInfo::RenderAttachmentInfo):
Allow the thumbnail aspect ratio to vary, instead of assuming it is always square.

Source/WebKit:

• UIProcess/Cocoa/WebPageProxyCocoa.mm:

(WebKit::convertPlatformImageToBitmap):
Propagate an image of the same aspect ratio that QuickLook provided,
instead of squishing it to square.

• UIProcess/QuickLookThumbnailLoader.mm:

(-[WKQLThumbnailLoadOperation start]):
Only request full thumbnails; we do not want the icon form, since <attachment>
already has one without QuickLook's help; if we can't get a full thumbnail,
we'll just leave it alone.

Tools:

• TestRunnerShared/UIScriptContext/Bindings/UIScriptController.idl:
• TestRunnerShared/UIScriptContext/UIScriptController.h:

(WTR::UIScriptController::insertAttachmentForFilePath):

• WebKitTestRunner/TestController.cpp:

(WTR::TestController::currentTestURL const):

• WebKitTestRunner/TestController.h:
• WebKitTestRunner/cocoa/UIScriptControllerCocoa.h:
• WebKitTestRunner/cocoa/UIScriptControllerCocoa.mm:

(WTR::UIScriptControllerCocoa::insertAttachmentForFilePath):
Make it possible to insert an attachment wrapping a file on disk
via UIScriptController.

LayoutTests:

• fast/attachment/attachment-thumbnail-preserves-aspect-ratio-expected.txt: Added.
• fast/attachment/attachment-thumbnail-preserves-aspect-ratio.html: Added.
• fast/attachment/resources/400x200-circle.png: Added.
• platform/ios/fast/attachment/attachment-thumbnail-preserves-aspect-ratio-expected.txt: Added.
• resources/ui-helper.js:

(window.UIHelper.insertAttachmentForFilePath):
Add a test that dumps the thumbnail size for a 400x200 attachment.
We only run it on iOS, because on macOS, QuickLook always returns
an image of the size we ask for (400x400), padded with whitespace,
so the problem does not reproduce and the test doesn't work right there.
On iOS, the result used to be 400x400 and now is 400x200.

I tried and failed to make a more useful test (a ref test, actually
testing the presentation) because it's quite hard to match the
native <attachment> painting.

12:37 PM Changeset in webkit [266742] by sihui_liu@apple.com

• 8 edits in trunk/Source

Remove storage WorkQueue in NetworkProcess
​https://bugs.webkit.org/show_bug.cgi?id=216199

Reviewed by Youenn Fablet.

Source/WebCore:

Move origin fetching code from NetworkProcess to IDBServer.

• Modules/indexeddb/server/IDBServer.cpp:

(WebCore::IDBServer::IDBServer::collectOriginsForVersion const):
(WebCore::IDBServer::IDBServer::getOrigins const):

• Modules/indexeddb/server/IDBServer.h:

Source/WebKit:

The only meaningful usage of this WorkQueue is iterating IndexedDB directories and collecting origins, while
there can be some IDB thread accessing the same directories. To remove the race, Let's just drop this WorkQueue
and fetch origins via WebIDBServer.

• NetworkProcess/IndexedDB/WebIDBServer.cpp:

(WebKit::WebIDBServer::getOrigins):

• NetworkProcess/IndexedDB/WebIDBServer.h:
• NetworkProcess/NetworkProcess.cpp:

(WebKit::NetworkProcess::addSessionStorageQuotaManager):
(WebKit::NetworkProcess::fetchWebsiteData):
(WebKit::NetworkProcess::deleteAndRestrictWebsiteDataForRegistrableDomains):
(WebKit::NetworkProcess::registrableDomainsWithWebsiteData):
(WebKit::NetworkProcess::addIndexedDatabaseSession):
(WebKit::NetworkProcess::addServiceWorkerSession):
(WebKit::NetworkProcess::ensurePathExists): Deleted.
(WebKit::NetworkProcess::postStorageTask): Deleted.
(WebKit::NetworkProcess::performNextStorageTask): Deleted.
(WebKit::NetworkProcess::collectIndexedDatabaseOriginsForVersion): Deleted.
(WebKit::NetworkProcess::indexedDatabaseOrigins): Deleted.

• NetworkProcess/NetworkProcess.h:

12:29 PM Changeset in webkit [266741] by aakash_jain@apple.com

• 3 edits in trunk/Tools

EWS should set cq- flag when a patch fails to build or introduces layout-test failures
​https://bugs.webkit.org/show_bug.cgi?id=214194

Reviewed by Jonathan Bedard.

• BuildSlaveSupport/ews-build/steps.py:

(AnalyzeCompileWebKitResults.start): Set cq- on patch when patch fails to build.
(AnalyzeLayoutTestsResults.report_failure): Set cq- on patch when patch introduces layout-test failures.

• BuildSlaveSupport/ews-build/steps_unittest.py: Updated unit-tests.

12:25 PM Changeset in webkit [266740] by commit-queue@webkit.org

• 5 edits
  8 adds in trunk

CoreImage Implementation of CSS Filters invert(), opacity(), brightness(), contrast()
​https://bugs.webkit.org/show_bug.cgi?id=215956

Patch by Frank Yang <​guowei_yang@apple.com> on 2020-09-08
Reviewed by Darin Adler.

Source/WebCore:

Adding CoreImage support for 4 more CSS filters, invert, opacity, brightness, contrast,
which all uses FEComponentTransfer and the linear transfer function. Partially implemented
FEComponentTransfer to support linear function type so that the 4 CSS filters work properly

Tests: css3/filters/effect-brightness-square.html

css3/filters/effect-contrast-square.html
css3/filters/effect-invert-square.html
css3/filters/effect-opacity-square.html

• platform/graphics/coreimage/FilterEffectRendererCoreImage.h:
• platform/graphics/coreimage/FilterEffectRendererCoreImage.mm:

(WebCore::FilterEffectRendererCoreImage::isNullOrLinearComponentTransferFunction):

currently CoreImage can only render FEComponentTransfer filters if the transfer funtion
is only linear. A FEComponentTransfer filter using only linear function would have it's
red, green, blue, alpha functions types either be linear or null.

(WebCore::FilterEffectRendererCoreImage::supportsCoreImageRendering):

currently CoreImage can only render FEComponentTransfer filters if the transfer funtion
is only linear.

(WebCore::FilterEffectRendererCoreImage::connectCIFilters):

now returns image for FEComponentTransfer

(WebCore::FilterEffectRendererCoreImage::imageForFEComponentTransfer):

function that uses CIFilters to render FEComponentTransfer.

• platform/graphics/filters/FEComponentTransfer.h:

(isType): added SPECIALIZE_TYPE_TRAITS so that we can downcast to FEComponentTransfer

LayoutTests:

• css3/filters/effect-brightness-square-expected.html: Added.
• css3/filters/effect-brightness-square.html: Added.
• css3/filters/effect-contrast-square-expected.html: Added.
• css3/filters/effect-contrast-square.html: Added.
• css3/filters/effect-invert-square-expected.html: Added.
• css3/filters/effect-invert-square.html: Added.
• css3/filters/effect-opacity-square-expected.html: Added.
• css3/filters/effect-opacity-square.html: Added.

12:22 PM Changeset in webkit [266739] by Aditya Keerthi

• 4 edits
  2 adds in trunk

[macOS] Date/time inputs should preserve focus on value change
​https://bugs.webkit.org/show_bug.cgi?id=216272

Reviewed by Wenson Hsieh.

Source/WebCore:

When the value of a date/time input is changed programmatically, the
inner elements of the input are rebuilt. Removing an inner element
that is focused results in the document's activeElement being nulled
out, and the date/time input loses focus.

To fix this issue, check to see whether the input has a focused field
prior to rebuilding. If there is a focused field, make an attempt to
focus the new field with the matching pseudo identifier. This ensures
that the focused field is preserved. If no matching identifier is found,
focus the first field to ensure the date/time input retains focus. The
second case can occur when the new value has different fields than the
old value, for example the presence/lack of a millisecond field in a
time input.

Test: fast/forms/date/date-preserve-focus-value-change.html

• html/shadow/DateTimeEditElement.cpp:

(WebCore::DateTimeEditElement::focusedFieldElement const):
(WebCore::DateTimeEditElement::layout):

• html/shadow/DateTimeEditElement.h:

LayoutTests:

• fast/forms/date/date-preserve-focus-value-change-expected.txt: Added.
• fast/forms/date/date-preserve-focus-value-change.html: Added.

12:11 PM Changeset in webkit [266738] by youenn@apple.com

• 2 edits in trunk/LayoutTests

Fix Internals::supportsVCPEncoder on BigSur
​https://bugs.webkit.org/show_bug.cgi?id=216174
<rdar://problem/66492801>

Reviewed by Geoffrey Garen.

• platform/mac/webrtc/captureCanvas-webrtc-software-encoder.html:

Increase timeout.

12:03 PM Changeset in webkit [266737] by youenn@apple.com

• 2 edits in trunk/LayoutTests

Do not use mouse events in http/wpt/html/dom/elements/images/hover-image-change.html
​https://bugs.webkit.org/show_bug.cgi?id=216265

Reviewed by Antti Koivisto.

• http/wpt/html/dom/elements/images/hover-image-change.html:

Mouse events do not work in iOS. Instead, use style change through changing class of an element.

11:55 AM Changeset in webkit [266736] by Megan Gardner

• 4 edits in trunk/Source

Switch from deprecated secondarySelectedControlColor to unemphasizedSelectedContentBackgroundColor.
​https://bugs.webkit.org/show_bug.cgi?id=216207

Reviewed by Darin Adler.

Source/WebCore:

No behavior change.

• rendering/RenderThemeMac.mm:

(WebCore::RenderThemeMac::platformInactiveSelectionBackgroundColor const):
(WebCore::RenderThemeMac::platformInactiveListBoxSelectionBackgroundColor const):
(WebCore::RenderThemeMac::systemColor const):

Source/WebKitLegacy/mac:

• Plugins/WebPluginController.mm:

(-[WebPluginController webPlugInContainerSelectionColor]):

11:49 AM Changeset in webkit [266735] by commit-queue@webkit.org

• 4 edits in trunk/Source/WebInspectorUI

Web Inspector: Incorrect decimal separator in Network web inspector
​https://bugs.webkit.org/show_bug.cgi?id=216177

Patch by Patrick Angle <Patrick Angle> on 2020-09-08
Reviewed by Brian Burg.

Removed separate bytes and suffix pieces for resources sizes, allowing the suffix to be localized and the number
to be correctly formatted for the locale.

• UserInterface/Base/Utilities.js: Number.bytesToString(...) now accepts a byte threshold to support

ResourceSizesContentView where we want less jitter between B and KB as you up/down through resources.

• UserInterface/Views/ResourceSizesContentView.css:

(.resource-sizes > .content > section.network .bytes): Adopted text-align from .bytes-group.
(.resource-sizes > .content > section.network .bytes-group): Deleted.
(.resource-sizes > .content > section.network .suffix): Deleted.

• UserInterface/Views/ResourceSizesContentView.js: No longer using a separate suffix element for sizes.

(WI.ResourceSizesContentView.prototype.initialLayout.createSizeComponents):
(WI.ResourceSizesContentView.prototype.initialLayout):
(WI.ResourceSizesContentView.prototype._formattedSizeComponent): Replaces _sizeComponents.
(WI.ResourceSizesContentView.prototype._refreshResourceSizeSection):
(WI.ResourceSizesContentView.prototype._sizeComponents): Deleted.

11:33 AM Changeset in webkit [266734] by Alan Coon

• 8 edits in branches/safari-610.1.28.10-branch/Source

Versioning.

WebKit-7610.1.28.10.1

11:29 AM Changeset in webkit [266733] by Alan Coon

• 1 copy in branches/safari-610.1.28.10-branch

New branch.

11:07 AM Changeset in webkit [266732] by Alan Coon

• 1 copy in tags/Safari-610.1.28.1.8

Tag Safari-610.1.28.1.8.

11:03 AM Changeset in webkit [266731] by Alan Coon

• 8 edits in branches/safari-610.1.28.1-branch/Source

Versioning.

WebKit-7610.1.28.1.8

10:56 AM Changeset in webkit [266730] by aakash_jain@apple.com

• 6 edits in trunk/Tools

Use https urls instead of http in webkitpy
​https://bugs.webkit.org/show_bug.cgi?id=216276

Reviewed by Jonathan Bedard.

• Scripts/webkitpy/common/checkout/scm/scm_mock.py:
• Scripts/webkitpy/common/config/urls.py:
• Scripts/webkitpy/common/config/urls_unittest.py: Changed few http urls to https, leaving a mix of http and https for testing.
• Scripts/webkitpy/common/config/watchlist:
• Scripts/webkitpy/style/checkers/cpp.py:

10:51 AM Changeset in webkit [266729] by commit-queue@webkit.org

• 4 edits in trunk/Source/WebCore

Derive big5, jis0208 and jis0212 tables from ICU
​https://bugs.webkit.org/show_bug.cgi?id=216253

Patch by Alex Christensen <​achristensen@webkit.org> on 2020-09-08
Reviewed by Darin Adler.

This substantially reduces the size of WebCore's binary data without changing behavior.
The EUC-KR table has so many differences from ICU it isn't worth the extraction.
Verified by many assertions and web platform tests.

• platform/text/EncodingTables.cpp:

(WebCore::jis0208):
(WebCore::jis0212):
(WebCore::big5):
(WebCore::eucKR):
(WebCore::checkEncodingTableInvariants):

• platform/text/EncodingTables.h:

(WebCore::stableSortByFirst):
(WebCore::findLastInSortedPairs): Deleted.

• platform/text/TextCodecCJK.cpp:

(WebCore::jis0208EncodeIndex):
(WebCore::codePointJIS0208):
(WebCore::codePointJIS0212):
(WebCore::eucJPEncode):
(WebCore::iso2022JPEncode):
(WebCore::shiftJISEncode):
(WebCore::eucKREncodingIndex):
(WebCore::TextCodecCJK::eucKRDecode):
(WebCore::big5EncodeIndex):
(WebCore::big5Encode):
(WebCore::TextCodecCJK::big5Decode):
(WebCore::jis0208DecodeIndex): Deleted.
(WebCore::big5DecodeIndex): Deleted.

10:38 AM Changeset in webkit [266728] by Peng Liu

• 21 edits in trunk/Source

Clean up functions and state variables related to the picture-in-picture implementation
​https://bugs.webkit.org/show_bug.cgi?id=215972

Reviewed by Jer Noble.

Source/WebCore:

This patch cleans up variables and functions related to changing video presentation mode
in HTMLVideoElement and HTMLMediaElement. It cleans up some event-related code as well.

This patch also simplifies the process of entering the picture-in-picture mode.
The XPC message VideoFullscreenManagerProxy::DidEnterFullscreen from the UI process
will provide the picture-in-picture window size if the video is entering picture-in-picture,
so that we can get rid of the intermediate state (waiting for the initial picture-in-picture
window frame).

With this patch, all presentation mode change requests need to go through the HTMLVideoElement.
Normally, video presentation mode change requests come from the web process where a page
calls the fullscreen/picture-in-picture API or a user interacts with the modern media controls.
However, through gestures, the UI process may request to change the presentation mode
(e.g., from fullscreen to picture-in-picture). The web process side will be like a dry run
because the tasks in the UI process are mostly completed before the web process receives
the request. The motivation to do that is to guarantee correct events are fired.
All scenarios sharing the same code path also help us avoid regressions because we don't
have a good way to create regression tests for the scenarios where presentation mode change
requests come from the UI process.

Covered by existing tests.

• html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::dispatchEvent):
(WebCore::HTMLMediaElement::enterFullscreen):
(WebCore::HTMLMediaElement::exitFullscreen):
(WebCore::HTMLMediaElement::setVideoFullscreenStandby):
(WebCore::HTMLMediaElement::willBecomeFullscreenElement):
Simplify the implementation.
(WebCore::HTMLMediaElement::willStopBeingFullscreenElement):
(WebCore::HTMLMediaElement::isVideoLayerInline):
(WebCore::HTMLMediaElement::setVideoFullscreenFrame):
(WebCore::HTMLMediaElement::setFullscreenMode):
Delete function HTMLVideoElement::fullscreenModeChanged() and move its code
to this function.
(WebCore::HTMLMediaElement::fullscreenModeChanged): Deleted.

• html/HTMLMediaElement.h:

(WebCore::HTMLMediaElement::fullscreenMode const):
Replace fullscreenModeChanged() with a private function setFullscreenMode().
Change "void setVideoFullscreenFrame(FloatRect)" to "void setVideoFullscreenFrame(const FloatRect&)".

• html/HTMLVideoElement.cpp:

(WebCore::HTMLVideoElement::webkitDisplayingFullscreen):
(WebCore::HTMLVideoElement::toPresentationMode):
(WebCore::HTMLVideoElement::webkitSetPresentationMode):
(WebCore::HTMLVideoElement::setPresentationMode):
(WebCore::HTMLVideoElement::didEnterFullscreenOrPictureInPicture):
(WebCore::HTMLVideoElement::didExitFullscreenOrPictureInPicture):
(WebCore::HTMLVideoElement::didBecomeFullscreenElement):
(WebCore::HTMLVideoElement::setVideoFullscreenFrame):
(WebCore::toPresentationMode): Deleted.
(WebCore::HTMLVideoElement::setFullscreenMode): Deleted.
(WebCore::HTMLVideoElement::fullscreenModeChanged): Deleted.
(WebCore::HTMLVideoElement::didStopBeingFullscreenElement): Deleted.

• html/HTMLVideoElement.h:

Only HTMLVIdeoElement::setPresentationMode() can be used by other modules (PlaybackSessionManager
or VideoFullscreenManager) to change the video presentation mode.
This patch defines different callbacks regarding the fullscreen mode change transition
for FullscreenManager and VideoFullscreenManager.

• didBecomeFullscreenElement for FullscreenManager.
• didEnterFullscreenOrPictureInPicture and didExitFullscreenOrPictureInPicture for

VideoFullscreenManager.
This patch removes m_isWaitingForPictureInPictureWindowFrame and renames m_isChangingPresentationMode
to m_isChangingVideoFullscreenMode.

• platform/cocoa/PlaybackSessionModelMediaElement.mm:

(WebCore::PlaybackSessionModelMediaElement::togglePictureInPicture):

• platform/cocoa/VideoFullscreenChangeObserver.h:

• platform/cocoa/VideoFullscreenModelVideoElement.mm:

(WebCore::VideoFullscreenModelVideoElement::requestFullscreenMode):
(WebCore::VideoFullscreenModelVideoElement::fullscreenModeChanged):
(WebCore::VideoFullscreenModelVideoElement::setHasVideo):
(WebCore::VideoFullscreenModelVideoElement::setVideoDimensions):
(WebCore::VideoFullscreenModelVideoElement::willEnterPictureInPicture):
(WebCore::VideoFullscreenModelVideoElement::didEnterPictureInPicture):
(WebCore::VideoFullscreenModelVideoElement::failedToEnterPictureInPicture):
(WebCore::VideoFullscreenModelVideoElement::willExitPictureInPicture):
(WebCore::VideoFullscreenModelVideoElement::didExitPictureInPicture):
Iterate a copy of the HashSet of VideoFullscreenModelClient pointers because a callback may
remove the client itself during the iteration.

• platform/ios/VideoFullscreenInterfaceAVKit.h:
• platform/ios/VideoFullscreenInterfaceAVKit.mm:

(-[WebAVPlayerLayer calculateTargetVideoFrame]):
(-[WebAVPlayerLayer layoutSublayers]):
(WebAVPlayerLayerView_startRoutingVideoToPictureInPicturePlayerLayerView):
(VideoFullscreenInterfaceAVKit::cleanupFullscreen):
(VideoFullscreenInterfaceAVKit::didStartPictureInPicture):
(VideoFullscreenInterfaceAVKit::failedToStartPictureInPicture):
(VideoFullscreenInterfaceAVKit::doSetup):
(VideoFullscreenInterfaceAVKit::doEnterFullscreen):
(VideoFullscreenInterfaceAVKit::willEnterStandbyFromPictureInPicture): Deleted.
(VideoFullscreenInterfaceAVKit::setWillEnterStandbyFromPictureInPicture): Deleted.
Add a property isPictureInPicture to WebAVPlayerLayer, which will set to YES if
a layer is created for picture-in-picture.
This property is used in -[WebAVPlayerLayer layoutSublayers] to skip
the unnecessary calculation of "targetVideoFrame" because the PiP window will
always have the same aspect ratio as the video content, so we can set
"targetVideoFrame" to be the bounds of the layer. Moreover, the calculation may
lead to rounding errors for some numbers and the errors may trigger unnecessary
picture-in-picture window resize events.

• platform/ios/WebVideoFullscreenControllerAVKit.mm:

This file has been moved to WebCore so we don't need to include WebCore headers
in a way like "#include <WebCore/XXX.h>".

• platform/mac/VideoFullscreenInterfaceMac.mm:

(-[WebVideoFullscreenInterfaceMacObjC enterPIP]):
(-[WebVideoFullscreenInterfaceMacObjC boundsDidChangeForVideoViewContainer:]):
(WebCore::VideoFullscreenInterfaceMac::enterFullscreen):
Add PIPState::EnteringPIP. Notify the observer that the video has entered
picture-in-picture after the PiP window size is finalized.

Source/WebKit:

This patch adds an optional FloatSize parameter to the VideoFullscreenManager::DidEnterFullscreen message.
Also, it enables the picture-in-picture and fullscreen events when a video element enters picture-in-picture
from the fullscreen mode.

• UIProcess/Cocoa/VideoFullscreenManagerProxy.h:
• UIProcess/Cocoa/VideoFullscreenManagerProxy.mm:

(WebKit::VideoFullscreenModelContext::didEnterFullscreen):
(WebKit::VideoFullscreenModelContext::prepareToExitFullscreen):
(WebKit::VideoFullscreenModelContext::willEnterPictureInPicture):
(WebKit::VideoFullscreenModelContext::didEnterPictureInPicture):
(WebKit::VideoFullscreenModelContext::failedToEnterPictureInPicture):
(WebKit::VideoFullscreenModelContext::willExitPictureInPicture):
(WebKit::VideoFullscreenModelContext::didExitPictureInPicture):
(WebKit::VideoFullscreenManagerProxy::ensureClientForContext):
(WebKit::VideoFullscreenManagerProxy::setupFullscreenWithID):
(WebKit::VideoFullscreenManagerProxy::enterFullscreen):
(WebKit::VideoFullscreenManagerProxy::didEnterFullscreen):
Use ensureClientForContext() instead of addClientForContext() in setupFullscreenWithID()
to avoid leaking VideoFullscreenInterfaceAVKit instances.
Iterate a copy of the HashSet of VideoFullscreenModelClient pointers because a callback may
remove the client itself during the iteration.

• UIProcess/ios/fullscreen/WKFullScreenWindowControllerIOS.mm:

(-[WKFullScreenWindowController prepareToExitPictureInPicture]):

• UIProcess/mac/WKFullScreenWindowController.h:
• UIProcess/mac/WKFullScreenWindowController.mm:

(-[WKFullScreenWindowController didExitPictureInPicture]):
Fix leaking of VideoFullscreenInterfaceMac.

• WebProcess/cocoa/VideoFullscreenManager.h:
• WebProcess/cocoa/VideoFullscreenManager.messages.in:
• WebProcess/cocoa/VideoFullscreenManager.mm:

(WebKit::VideoFullscreenManager::fullscreenModeChanged):
We need to change the fullscreen mode in the VideoFullscreenInterfaceContext instance as well.
(WebKit::VideoFullscreenManager::didEnterFullscreen):
(WebKit::VideoFullscreenManager::didCleanupFullscreen):
Call didExitFullscreenOrPictureInPicture() instead of didStopBeingFullscreenElement(), which
is for the FullscreenManager.

9:46 AM Changeset in webkit [266727] by Simon Fraser

• 1 edit
  3 adds in trunk/LayoutTests

Add a standalone version of the test runtime tree map viewer, that can accept stats.json from a local test run
​https://bugs.webkit.org/show_bug.cgi?id=189396

Reviewed by Darin Adler.

Add the viewer and a sample stats.json file. Devs can drag a stats.json file into test-duration-treemap.html to
see the map (or specify it as a url parameter).

A future patch will hook this up to results.html.

• fast/harness/stats.json: Added.
• fast/harness/test-duration-treemap-expected.txt: Added.
• fast/harness/test-duration-treemap.html: Added.

8:25 AM Changeset in webkit [266726] by commit-queue@webkit.org

• 2 edits in trunk/Source/WebKit

[GTK] AcceleratedBackingStoreWayland might erroneously try to use SHM with older wpebackend-fdo
​https://bugs.webkit.org/show_bug.cgi?id=216254

Patch by Mike Gorse <​mgorse@suse.com> on 2020-09-08
Reviewed by Adrian Perez de Castro.

• UIProcess/gtk/AcceleratedBackingStoreWayland.cpp:

(WebKit::tryInitializeSHM): Return false for wpebackend-fdo < 1.7.0.

8:16 AM Changeset in webkit [266725] by Chris Lord

• 2 edits in trunk/LayoutTests

[GStreamer] Some convolvernode tests are flaky crashing
​https://bugs.webkit.org/show_bug.cgi?id=216118

Unreviewed test gardening.

Some tests are causing random, subsequently run tests to crash. Skip
these tests on affected platforms.

• platform/glib/TestExpectations:

7:53 AM Changeset in webkit [266724] by commit-queue@webkit.org

• 8 edits
  4 adds in trunk/Tools

[Flatpak SDK] Follow-up fixes for 0.3 branch
​https://bugs.webkit.org/show_bug.cgi?id=216239

Patch by Philippe Normand <​pnormand@igalia.com> on 2020-09-08
Reviewed by Žan Doberšek.

• buildstream/elements/filtered.bst: Enable sccache on aarch64 too.
• buildstream/elements/flatpak/sdk.bst: Enable the rust extension, can be useful eg, when

using gst-build to have the Rust plugins enabled.

• buildstream/elements/freedesktop-sdk.bst: Switch to tracking branch instead of tags.
• buildstream/elements/sdk-build-depends/rapidjson.bst: Bump to current git master and add a

patch fixing aarch64 build.

• buildstream/elements/sdk-platform.bst: Add procps, needed at least for test-webkitpy
• buildstream/elements/sdk/gst-plugin-dav1d.bst: Added.
• buildstream/elements/sdk/gst-plugins-bad.bst: Fix build for aarch64.
• buildstream/elements/sdk/sccache.bst: Disable dist server, we need the client only. The

server builds only on x86_64 anyway.

• buildstream/files/gst-plugin-dav1d/Cargo.lock: Added.
• buildstream/patches/rapidjson-0001-Fix-build-warnings-emitted-by-GCC-10-on-Aarch64.patch: Added.

7:24 AM Changeset in webkit [266723] by Karl Rackler

• 2 edits in trunk/LayoutTests

Removed 6 forms and 3 tables test expectations for Big Sur
rdar://68372418

Unreviewed test gardening.

• platform/mac/TestExpectations:

6:26 AM Changeset in webkit [266722] by Alan Bujtas

• 3 edits in trunk/Source/WebCore

[LFC][IFC] Remove redundant FormattingContext::Quirks::lineDescentNeedsCollapsing
​https://bugs.webkit.org/show_bug.cgi?id=216259

Reviewed by Antti Koivisto.

This function is not needed anymore. InlineBoxes start collapsed by default and they get stretched when they gain content
(and the root inline box has no imaginary strut in non-standard mode).

• layout/inlineformatting/InlineFormattingContext.h:
• layout/inlineformatting/InlineFormattingContextQuirks.cpp:

(WebCore::Layout::InlineFormattingContext::Quirks::lineDescentNeedsCollapsing const): Deleted.

6:18 AM Changeset in webkit [266721] by commit-queue@webkit.org

• 6 edits in trunk/Source/WebCore

[GStreamer] Follow-up of "Convert custom GObject subclasses to WEBKIT_DEFINE_TYPE"
​https://bugs.webkit.org/show_bug.cgi?id=204673
<rdar://problem/68455348>

Patch by Víctor Manuel Jáquez Leal <​vjaquez@igalia.com> on 2020-09-08
Reviewed by Carlos Garcia Campos.

This a continuation of 204673 since a post-merge review came up: it
was also needed to remove g_type_class_add_private() calls.

No new tests needed.

• platform/audio/gstreamer/WebKitWebAudioSourceGStreamer.cpp:

(_WebKitWebAudioSrcPrivate::_WebKitWebAudioSrcPrivate):
(webkit_web_audio_src_class_init):

• platform/graphics/gstreamer/VideoSinkGStreamer.cpp:

(webkit_video_sink_class_init):

• platform/graphics/gstreamer/eme/WebKitClearKeyDecryptorGStreamer.cpp:

(webkit_media_clear_key_decrypt_class_init):

• platform/graphics/gstreamer/eme/WebKitCommonEncryptionDecryptorGStreamer.cpp:

(webkit_media_common_encryption_decrypt_class_init):

• platform/graphics/gstreamer/eme/WebKitThunderDecryptorGStreamer.cpp:

(webkit_media_thunder_decrypt_class_init):

4:59 AM Changeset in webkit [266720] by commit-queue@webkit.org

• 8 edits in trunk

Fix crash in image-loading-lazy-slow.html WPT test
​https://bugs.webkit.org/show_bug.cgi?id=215998

Patch by Rob Buis <​rbuis@igalia.com> on 2020-09-08
Reviewed by Youenn Fablet.

LayoutTests/imported/w3c:

Adjust image-loading-lazy-slow.html to enable lazy image loading.

• web-platform-tests/html/semantics/embedded-content/the-img-element/image-loading-lazy-slow.html:

Source/WebCore:

When setting images to an empty src, make sure
to reset any lazy image load information.

• html/LazyLoadImageObserver.h:
• loader/ImageLoader.cpp:

(WebCore::ImageLoader::updateFromElement):
(WebCore::ImageLoader::elementDidMoveToNewDocument):
(WebCore::ImageLoader::resetLazyImageLoading):

• loader/ImageLoader.h:

(WebCore::ImageLoader::isDeferred const):

LayoutTests:

Mark image-loading-lazy-slow.html as ImageOnlyFailure.

• TestExpectations:

3:49 AM Changeset in webkit [266719] by Diego Pino Garcia

• 2 edits in trunk/Tools

[GTK] Unreviewed gardening. Mark several API-GTK tests as failures after r266697.

• TestWebKitAPI/glib/TestExpectations.json:

1:52 AM Changeset in webkit [266718] by Pablo Saavedra

• 11 edits
  2 adds in trunk/Source/WebKit

[GLIB] RemoteInspectorServer is not started if WebKitWebContext is not created already created
​https://bugs.webkit.org/show_bug.cgi?id=216120

Reviewed by Carlos Garcia Campos.

Added a WebKit/UIProcess/API/glib/WebKitInitialize what implements
its own webkitInitialize() to ensure the inspector server is
initialized as early as possible, before other api calls that would
depend on the inspector server being running.

The RemoteInspectorServer initialization is not longer
asociated to the initialization of the WebProcessPoolGLib.

The webkitInitialize() extends the shared InitializeWebKit2();
and also includes the initialization of the RemoteInspectorServer
so the initialization of the remote inspector server can be now
triggered by the initialization of many other API objects too:
WebKitInputMethodContext, WebKitSettings, WebKitUserContentManager,
WebKitWebContext, WebKitWebsiteDataManager,
WebKitUserContentFilterStore, WebKitWebViewBase.

• SourcesGTK.txt:
• SourcesWPE.txt:
• UIProcess/API/glib/WebKitInitialize.cpp: Added.

(WebKit::initializeRemoteInspectorServer):
(WebKit::webkitInitialize):

• UIProcess/API/glib/WebKitInitialize.h: Added.
• UIProcess/API/glib/WebKitInputMethodContext.cpp:

(webkit_input_method_context_class_init):

• UIProcess/API/glib/WebKitSettings.cpp:

(webkit_settings_class_init):

• UIProcess/API/glib/WebKitUserContentFilterStore.cpp:

(webkit_user_content_filter_store_class_init):

• UIProcess/API/glib/WebKitUserContentManager.cpp:
• UIProcess/API/glib/WebKitWebContext.cpp:

(webkit_web_context_class_init):

• UIProcess/API/glib/WebKitWebsiteDataManager.cpp:

(webkit_website_data_manager_class_init):

• UIProcess/API/gtk/WebKitWebViewBase.cpp:

(webkit_web_view_base_class_init):

• UIProcess/glib/WebProcessPoolGLib.cpp:

(WebKit::WebProcessPool::platformInitialize):
(WebKit::initializeRemoteInspectorServer): Deleted.

1:48 AM Changeset in webkit [266717] by Antti Koivisto

• 6 edits in trunk/Source/WebCore

Comparing styles with large but identical custom property maps is slow
​https://bugs.webkit.org/show_bug.cgi?id=216241
<rdar://problem/67946605>

Reviewed by Darin Adler.

In a full style rebuild we lose sharing of equal StyleCustomPropertyDatas between the old and the new style.
Custom properties are usually defined in document element and this lack of sharing inherits to all elements.
Without sharing equality comparisons end up requiring deep comparisons, which can be slow if there are thousands of properties.

Fix by deduplicating identical property maps between the old the new style. All the descendants that inherit the properties will
now be cheap to compare.

The patch also contains some other basic custom property optimizations.

Going to full screen and back on youtube, this patch reduces time in the longest style update from ~460ms to ~90ms.

• css/CSSCustomPropertyValue.cpp:

(WebCore::CSSCustomPropertyValue::equals const):

Check for pointer equality first.

• rendering/style/RenderStyle.cpp:

(WebCore::RenderStyle::deduplicateInheritedCustomProperties):

Deduplicate if the properties are equal but not shared.

(WebCore::RenderStyle::setInheritedCustomPropertyValue):
(WebCore::RenderStyle::setNonInheritedCustomPropertyValue):

Check if the value is already in the map before triggering copy-on-write.

• rendering/style/RenderStyle.h:

(WebCore::RenderStyle::setInheritedCustomPropertyValue): Deleted.
(WebCore::RenderStyle::setNonInheritedCustomPropertyValue): Deleted.

• style/StyleBuilder.cpp:

(WebCore::Style::Builder::applyCustomProperty):

Don't create unnecesary copies of CSSCustomPropertyValue.

• style/StyleTreeResolver.cpp:

(WebCore::Style::TreeResolver::createAnimatedElementUpdate):

Try to deduplicate before comparing the styles.

12:35 AM Changeset in webkit [266716] by svillar@igalia.com

• 6 edits in trunk

[css-flex] Don't skip flexboxes with auto height for percentage computations in quirks mode
​https://bugs.webkit.org/show_bug.cgi?id=216249

Reviewed by Darin Adler.

LayoutTests/imported/w3c:

Fixed the (FAIL) expectations of two WPT tests that are now passing.

• web-platform-tests/css/css-flexbox/percentage-size-quirks-expected.txt: Replaced FAIL with PASS.
• web-platform-tests/css/css-flexbox/quirks-auto-block-size-with-percentage-item-expected.txt: Ditto.

Source/WebCore:

We shouldn't skip flexboxes with height:auto when going up the ancestor chain as we do with other types as grid.

Apart from that, we force flexible box to follow strict mode even when document is in quirks mode as flexbox
was invented way after quirky browsers. This matches Chromium, Edge (pre-Chromium) and Firefox behaviour.

• rendering/RenderBox.cpp:

(WebCore::RenderBox::skipContainingBlockForPercentHeightCalculation const): Don't skip flexboxes.

• rendering/RenderBoxModelObject.cpp:

(WebCore::RenderBoxModelObject::hasAutoHeightOrContainingBlockWithAutoHeight const): Ignore quirks mode for flexboxes.

12:21 AM Changeset in webkit [266715] by ysuzuki@apple.com

• 6 edits
  4 adds in trunk

[JSC] Special property caching should check Structure's cacheability
​https://bugs.webkit.org/show_bug.cgi?id=216222

Reviewed by Saam Barati.

JSTests:

• stress/module-namespace-object-caching.js: Added.

(shouldBe):
(async fn):

• stress/not-cache-over-uncacheable-dictionary.js: Added.

(shouldBe):

• stress/resources/to-string-module.js: Added.

(export.toString):

• stress/resources/value-of-module.js: Added.

(export.valueOf):

Source/JavaScriptCore:

While StructureRareData::cacheSpecialPropertySlow caches properties, the way it takes is incomplete.
It is not checking Structure's cacheability. We were caching miss condition even if structure is !propertyAccessesAreCacheableForAbsence.
We should perform the same check done in IC case. Strictly speaking, we can cache value for uncacheable-dictionary because we are setting
property change watchpoint (which will fire). But it sounds not so profitable if this structure is uncacheable.

• runtime/JSObject.cpp:

(JSC::JSObject::convertToUncacheableDictionary):

• runtime/JSObject.h:
• runtime/StructureRareData.cpp:

(JSC::StructureRareData::cacheSpecialPropertySlow):

• tools/JSDollarVM.cpp:

(JSC::functionToUncacheableDictionary):
(JSC::JSDollarVM::finishCreation):

Sep 7, 2020:

9:30 PM Changeset in webkit [266714] by rniwa@webkit.org

• 11 edits in trunk/Source/WebCore

Move all remaining flags from ElementRareData to Node to reduce the frequency
​https://bugs.webkit.org/show_bug.cgi?id=216208

Reviewed by Darin Adler.

This patch moves all remaining bit flags from ElementRareData to Node::m_nodeFlags. It also stores
tabIndex of 0 and -1 in Node::m_nodeFlags to avoid creating ElementRareData for these common values.

No new tests since there should be no behavioral change.

• dom/Element.cpp:

(WebCore::Element::clearTabIndexExplicitlyIfNeeded): Deleted. Merged into setTabIndexExplicitly.
(WebCore::Element::setTabIndexExplicitly): Now takes Optional<int>. Stores 0 and -1 as node flags.
(WebCore::Element::tabIndexSetExplicitly const): Added the special casing for 0 and -1.
(WebCore::Element::removedFromAncestor):
(WebCore::Element::containsFullScreenElement const): Deleted.
(WebCore::Element::setContainsFullScreenElement):
(WebCore::Element::hasPendingResources const): Deleted.
(WebCore::Element::setHasPendingResources): Deleted.
(WebCore::Element::clearHasPendingResources): Deleted.
(WebCore::Element::hasCSSAnimation const): Deleted.
(WebCore::Element::setHasCSSAnimation): Deleted.
(WebCore::Element::clearHasCSSAnimation): Deleted.
(WebCore::Element::createElementIdentifier):

• dom/Element.h:

(WebCore::Element::hasPendingResources const): Moved here from Element.cpp.
(WebCore::Element::setHasPendingResources): Ditto.
(WebCore::Element::clearHasPendingResources): Ditto.
(WebCore::Element::hasCSSAnimation const): Ditto.
(WebCore::Element::setHasCSSAnimation): Ditto.
(WebCore::Element::clearHasCSSAnimation): Ditto.
(WebCore::Element::containsFullScreenElement const): Ditto.

• dom/ElementRareData.cpp:
• dom/ElementRareData.h:

(WebCore::ElementRareData::tabIndex const): Deleted.
(WebCore::ElementRareData::setTabIndexExplicitly): Deleted.
(WebCore::ElementRareData::tabIndexSetExplicitly const): Deleted.
(WebCore::ElementRareData::clearTabIndexExplicitly): Deleted.
(WebCore::ElementRareData::containsFullScreenElement): Deleted.
(WebCore::ElementRareData::setContainsFullScreenElement): Deleted.
(WebCore::ElementRareData::hasPendingResources const): Deleted.
(WebCore::ElementRareData::setHasPendingResources): Deleted.
(WebCore::ElementRareData::hasCSSAnimation const): Deleted.
(WebCore::ElementRareData::setHasCSSAnimation): Deleted.
(WebCore::ElementRareData::hasElementIdentifier const): Deleted.
(WebCore::ElementRareData::setHasElementIdentifier): Deleted.
(WebCore::ElementRareData::useTypes const): Added the missing use types for style maps, part list,
and part names.
(WebCore::ElementRareData::ElementRareData):
(WebCore::ElementRareData::unusualTabIndex const): Renamed from tabIndex.
(WebCore::ElementRareData::setUnusualTabIndex): Ditto from setTabIndexExplicitly.

• dom/Node.cpp:

(WebCore::stringForRareDataUseType): Added the missing strings.

• dom/Node.h:

(WebCore::Node::NodeFlags): Moved bit flags from ElementRareData and reserved the two highest bits
to store TabIndexState.
(WebCore::Node::s_tabIndexStateBitOffset): Added.
(WebCore::Node::s_tabIndexStateBitMask): Added.
(WebCore::Node::TabIndexState): Added.
(WebCore::Node::tabIndexState const): Added.
(WebCore::Node::setTabIndexState): Added.

• dom/NodeRareData.h:

(WebCore::NodeRareData::UseType): Added the missing use types: style maps, part list, and part names.

• html/HTMLElement.cpp:

(WebCore::HTMLElement::parseAttribute):

• mathml/MathMLElement.cpp:

(WebCore::MathMLElement::parseAttribute):

• svg/SVGElement.cpp:

(WebCore::SVGElement::parseAttribute):

8:50 PM Changeset in webkit [266713] by rniwa@webkit.org

• 5 edits in trunk

Make CompactUniquePtrTuple actually work with subclassing and custom deleter
​https://bugs.webkit.org/show_bug.cgi?id=216225

Reviewed by Darin Adler.

Source/WTF:

Fixed bugs in CompactUniquePtrTuple which prevented subclassing and custom deleter to work.

• wtf/CompactPointerTuple.h:

(WTF::CompactPointerTuple::CompactPointerTuple): Added move constructor with implicit cast
of a convertible pointer type.

• wtf/CompactUniquePtrTuple.h:

(WTF::makeCompactUniquePtr): Added the missing deleter from the return type.
(WTF::CompactUniquePtrTuple::CompactUniquePtrTuple): Allow Deleter to be different if it's
the default deleter in the move constructor so that CompactUniquePtrTuple<U, Type> could be
moved to CompactUniquePtrTuple<T, Type> if U is convertible to T without having to specify
the same deleter (std::default_delete<U> is not same as std::default_delete<T> but allow it).
(WTF::CompactUniquePtrTuple::operator=): Ditto.
(WTF::CompactUniquePtrTuple::setPointer): Ditto from std::unique_ptr.
(WTF::CompactUniquePtrTuple): Friend declare all other specializations of CompactUniquePtrTuple
so that the above fixes work.

Tools:

• TestWebKitAPI/Tests/WTF/CompactUniquePtrTuple.cpp:

(TestWebKitAPI::A::~A): Make this virtual.
(TestWebKitAPI::B): Added.
(TestWebKitAPI::B::B): Added.
(TestWebKitAPI::B::~B): Added.
(WTF_CompactUniquePtrTuple.Subclassing): Added. Tests subclassing.
(TestWebKitAPI::ADeleter): Added.
(TestWebKitAPI::ADeleter::operator() const):
(WTF_CompactUniquePtrTuple.Deleter): Added. Tests a custom deleter.

7:46 PM Debugging With Visual Studio edited by Fujii Hironori

(diff)

6:35 PM Changeset in webkit [266712] by Chris Dumez

• 12 edits
  5 adds
  2 deletes in trunk

AudioParam.cancelAndHoldAtTime() is missing
​https://bugs.webkit.org/show_bug.cgi?id=215947
<rdar://problem/68362061>

Reviewed by Darin Adler.

LayoutTests/imported/w3c:

Rebaseline WPT tests that are now passing.

• web-platform-tests/webaudio/idlharness.https.window-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/audioparam-setValueCurveAtTime-expected.txt:
• web-platform-tests/webaudio/the-audio-api/the-audioparam-interface/retrospective-setValueCurveAtTime-expected.txt:

Source/WebCore:

Add implementation for AudioParam.cancelAndHoldAtTime():

• ​https://www.w3.org/TR/webaudio/#dom-audioparam-cancelandholdattime

This patch is based on the following Blink commit by Raymond Toy:

• ​https://codereview.chromium.org/1533103002

Tests: webaudio/audioparam-cancel-and-hold.html

webaudio/cancel-values-crash-913217.html

• Modules/webaudio/AudioParam.cpp:
• Modules/webaudio/AudioParam.h:
• Modules/webaudio/AudioParam.idl:
• Modules/webaudio/AudioParamTimeline.cpp:

(WebCore::AudioParamTimeline::setValueAtTime):
(WebCore::AudioParamTimeline::linearRampToValueAtTime):
(WebCore::AudioParamTimeline::exponentialRampToValueAtTime):
(WebCore::AudioParamTimeline::setTargetAtTime):
(WebCore::AudioParamTimeline::setValueCurveAtTime):
(WebCore::AudioParamTimeline::insertEvent):
(WebCore::AudioParamTimeline::cancelAndHoldAtTime):
(WebCore::AudioParamTimeline::removeCancelledEvents):
(WebCore::AudioParamTimeline::valueForContextTime):
(WebCore::AudioParamTimeline::valuesForTimeRangeImpl):
(WebCore::AudioParamTimeline::linearRampAtTime):
(WebCore::AudioParamTimeline::exponentialRampAtTime):
(WebCore::AudioParamTimeline::valueCurveAtTime):
(WebCore::AudioParamTimeline::handleCancelValues):

• Modules/webaudio/AudioParamTimeline.h:

(WebCore::AudioParamTimeline::ParamEvent::createSetValueEvent):
(WebCore::AudioParamTimeline::ParamEvent::createLinearRampEvent):
(WebCore::AudioParamTimeline::ParamEvent::createExponentialRampEvent):
(WebCore::AudioParamTimeline::ParamEvent::createSetTargetEvent):
(WebCore::AudioParamTimeline::ParamEvent::createSetValueCurveEvent):
(WebCore::AudioParamTimeline::ParamEvent::createCancelValuesEvent):
(WebCore::AudioParamTimeline::ParamEvent::ParamEvent):
(WebCore::AudioParamTimeline::ParamEvent::type const):
(WebCore::AudioParamTimeline::ParamEvent::savedEvent):
(WebCore::AudioParamTimeline::ParamEvent::setCancelledValue):
(WebCore::AudioParamTimeline::ParamEvent::hasDefaultCancelledValue const):
(WebCore::AudioParamTimeline::ParamEvent::curvePointsPerSecond const):
(WebCore::AudioParamTimeline::ParamEvent::curveEndValue const):

LayoutTests:

• webaudio/audioparam-cancel-and-hold-expected.txt: Added.
• webaudio/audioparam-cancel-and-hold.html: Added.
• webaudio/cancel-values-crash-913217-expected.txt: Added.
• webaudio/cancel-values-crash-913217.html: Added.
• webaudio/resources/audio-param.js: Added.

Import layout test coverage from Blink.

• webaudio/audioparam-setValueCurveAtTime-expected.txt: Removed.
• webaudio/audioparam-setValueCurveAtTime.html: Removed.

Remove outdated test. This test is now part of web-platform-tests at
webaudio/the-audio-api/the-audioparam-interface/audioparam-setValueCurveAtTime.html
and is now passing.

3:35 PM Changeset in webkit [266711] by Karl Rackler

• 1 edit
  9 deletes in trunk/LayoutTests

Remove duplicate test expectaions from platform/mac-bigsur
rdar://68372418

Unreviewed test gardening.

• platform/mac-bigsur/fast/css/text-overflow-input-expected.txt: Removed.
• platform/mac-bigsur/fast/forms/button-sizes-expected.txt: Removed.
• platform/mac-bigsur/fast/forms/button-table-styles-expected.txt: Removed.
• platform/mac-bigsur/fast/forms/input-button-sizes-expected.txt: Removed.
• platform/mac-bigsur/fast/forms/listbox-bidi-align-expected.txt: Removed.
• platform/mac-bigsur/fast/text/international/pop-up-button-text-alignment-and-direction-expected.txt: Removed.
• platform/mac-bigsur/tables/mozilla/bugs/bug2479-3-expected.txt: Removed.
• platform/mac-bigsur/tables/mozilla/bugs/bug33855-expected.txt: Removed.
• platform/mac-bigsur/tables/mozilla_expected_failures/bugs/bug2479-5-expected.txt: Removed.

1:55 PM Changeset in webkit [266710] by commit-queue@webkit.org

• 6 edits in trunk/Source

Unreviewed, reverting r266645.
​https://bugs.webkit.org/show_bug.cgi?id=216251

Caused MotionMark regression

Reverted changeset:

"Move lazy DisplayLink tear down logic from the WebProcess to
the UIProcess"
​https://bugs.webkit.org/show_bug.cgi?id=216195
​https://trac.webkit.org/changeset/266645

12:36 PM Changeset in webkit [266709] by weinig@apple.com

• 12 edits
  12 adds in trunk

[WebIDL] Fix issues found by preprocess-idls.pl parser validation and enabled parser validation by default for the tests
​https://bugs.webkit.org/show_bug.cgi?id=216246

Reviewed by Darin Adler.

Source/WebCore:

Enabling preprocess-idls.pl's validation parser for run-bindings-tests revealed a few issues
that this change fixes:

• Includes statements (as well as any other non-extended attribute related statement) in a comment were being treated as being part of the document. Fixed this by moving the existing c/c++ style comment removal from getInterfaceExtendedAttributesFromIDL to processIDL so it is run for all IDLs. Since getInterfaceExtendedAttributesFromIDL was called for most files anyway, this doesn't measurably change the performance.
• Some test IDLs did not have their primary interface name matching the name of the IDL file. This resulted in a number of tests not actually producing any output. Updated the tests themselves to match name.

To ensure we don't regress here, the validation parser is now always enabled when running
the tests, and preprocess-idls.pl has been updated to be a little less noisy when running
with the validation parser enabled (unless you specify --verbose) so that the test output
is not disturbed.

• bindings/scripts/preprocess-idls.pl:

(processIDL):
(getPartialNamesFromIDL):
(getIncludedInterfacesFromIDL):
(isCallbackInterfaceFromIDL):
(isMixinInterfaceFromIDL):
(containsIterableInterfaceFromIDL):
(containsInterfaceOrCallbackInterfaceFromIDL):
(containsInterfaceWithConstantsFromIDL):
(getInterfaceExtendedAttributesFromIDL):
(getUndefinedBaseDictionariesFromIDL):

• bindings/scripts/test/JS/JSTestDomainSecurity.cpp: Added.
• bindings/scripts/test/JS/JSTestDomainSecurity.h: Added.
• bindings/scripts/test/JS/JSTestMapLike.cpp: Added.
• bindings/scripts/test/JS/JSTestMapLike.h: Added.
• bindings/scripts/test/JS/JSTestReadOnlyMapLike.cpp: Added.
• bindings/scripts/test/JS/JSTestReadOnlyMapLike.h: Added.
• bindings/scripts/test/JS/JSTestReadOnlySetLike.cpp: Added.
• bindings/scripts/test/JS/JSTestReadOnlySetLike.h: Added.
• bindings/scripts/test/JS/JSTestReportExtraMemoryCost.cpp: Added.
• bindings/scripts/test/JS/JSTestReportExtraMemoryCost.h: Added.
• bindings/scripts/test/JS/JSTestSetLike.cpp: Added.
• bindings/scripts/test/JS/JSTestSetLike.h: Added.
• bindings/scripts/test/SupplementalDependencies.dep:
• bindings/scripts/test/TestDomainSecurity.idl:
• bindings/scripts/test/TestMapLike.idl:
• bindings/scripts/test/TestObj.idl:
• bindings/scripts/test/TestReadOnlyMapLike.idl:
• bindings/scripts/test/TestReadOnlySetLike.idl:
• bindings/scripts/test/TestReportExtraMemoryCost.idl:
• bindings/scripts/test/TestSetLike.idl:

Tools:

• Scripts/webkitpy/bindings/main.py:

(BindingsTests.generate_supplemental_dependency):
Enable the parser validation for preprocess-idls.pl to catch accidental differences between
the fast regex based scanner in preprocess-idls.pl and the real IDL parser.

12:04 PM Changeset in webkit [266708] by youenn@apple.com

• 2 edits in trunk/Source/WebCore

Fix ​https://trac.webkit.org/changeset/266508 as per post-commit comments
​https://bugs.webkit.org/show_bug.cgi?id=216242

Reviewed by Alex Christensen.

No change of behavior.

• platform/mediastream/libwebrtc/LibWebRTCProvider.cpp:

(WebCore::toRTCRtpCapabilities):

11:54 AM Changeset in webkit [266707] by Alan Bujtas

• 3 edits in trunk/LayoutTests

Unreviewed rebaselining.

The test output (render tree dump) is sensitive to float precision.

r266509 slightly changed the way we compute the run's horizontal position
and its new position is 0.000061px more to the right. The integral ceil that we apply
to the run's width computation in the dump-render-tree-as-text makes the final value go from 39px to 40px.
(Note that final width value has dependency on the horizontal positon.)

• platform/ios-wk2/TestExpectations:
• platform/ios/css2.1/t0905-c5525-fltwidth-00-c-g-expected.txt:

11:42 AM Changeset in webkit [266706] by weinig@apple.com

• 17 edits
  1 copy
  3 adds in trunk/Source/WebCore

[WebIDL] Support extended attributes on includes statements to allow for conditionalized inclusion
​https://bugs.webkit.org/show_bug.cgi?id=216235

Reviewed by Darin Adler.

In some circumstancs, such as with the AnimationFrameProvider interface mixin, it is
useful to only include a mixin based on some conditional. For AnimationFrameProvider,
we currently only want to include it in DedicatedWorkerGlobalScope if ENABLE(OFFSCREEN_CANVAS)
is set at compile time, and the runtime feature flag offscreenCanvasEnabled() is turned
on. To support that, this change adds support for specifying extended attributes on the
includes statement itself, and having those extended attributes applied to each member
of the mixin. For AnimationFrameProvider in DedicatedWorkerGlobalScope, that looks like:

[Conditional=OFFSCREEN_CANVAS, EnabledAtRuntime=OffscreenCanvas] DedicatedWorkerGlobalScope includes AnimationFrameProvider;

• CMakeLists.txt:
• DerivedSources-input.xcfilelist:
• DerivedSources-output.xcfilelist:
• DerivedSources.make:
• WebCore.xcodeproj/project.pbxproj:

Add new files, update xcfilelists.

• animation/AnimationFrameProvider.idl: Added.
• page/DOMWindow.idl:
• workers/DedicatedWorkerGlobalScope.idl:

Split AnimationFrameProvider.idl out into its own mixin (as specified) and included it in
both DOMWindow and DedicatedWorkerGlobalScope (conditionally).

• bindings/scripts/CodeGenerator.pm:

(ProcessDocument):
(ProcessDictionaryAndEnumerationImplementedAsOverrides):
(ProcessInterfaces):
(ProcessCallbackFunctions):
(ProcessDictionaries):
(ProcessEnumerations):
Split processing of each different type of top level type into its own function,
and defer processing of supplemental dependencies until the primary type has been
selected.

(MergeExtendedAttributesFromSupplemental):
(IsValidSupplementalInterface):
(IsValidSupplementalDictionary):
(ProcessInterfaceSupplementalDependencies):
(ProcessDictionarySupplementalDependencies):
Split ProcessSupplementalDependencies into ProcessInterfaceSupplementalDependencies
and ProcessDictionarySupplementalDependencies, which now get passed the primary
interface or dictionary that needs supplementing. ProcessInterfaceSupplementalDependencies
now also supports merging extended attributes from includes statements and checking
that all includes are listed in the dependency set.

• bindings/scripts/IDLAttributes.json:

Allow setting use of Conditional, EnabledAtRuntime and EnabledBySetting for
includes statements.

• bindings/scripts/IDLParser.pm:

(parseIncludesStatement):
Fix ordering of struct members so that extendedAttributes is always last (no real
good reason other than most of the structs already did this and the lack of consistency
bothered me). Add extendedAttributes to IDLIncludesStatement and set it when
parsing.

• bindings/scripts/preprocess-idls.pl:

(getIncludedInterfacesFromIDL):
Fix bug in regex (found easily by the validation parser!) that incorrectly required the interface
name in the includes statement to be the first word on a line.

• bindings/scripts/test/BindingTestGlobalConstructors.idl:
• bindings/scripts/test/JS/JSTestConditionalIncludes.cpp: Added.
• bindings/scripts/test/JS/JSTestConditionalIncludes.h: Added.
• bindings/scripts/test/JS/JSTestGlobalObject.cpp:
• bindings/scripts/test/JS/JSTestInterface.cpp:
• bindings/scripts/test/SupplementalDependencies.dep:
• bindings/scripts/test/TestConditionalIncludes.idl: Added.
• bindings/scripts/test/TestIncludes.idl:

Add tests for extended attributes on includes statements.

10:55 AM Changeset in webkit [266705] by commit-queue@webkit.org

• 4 edits in trunk/Source/WebCore

Add memory cache with/without validation to logResourceResponseSource
​https://bugs.webkit.org/show_bug.cgi?id=216244
<rdar://problem/64184239>

Patch by Alex Christensen <​achristensen@webkit.org> on 2020-09-07
Reviewed by Darin Adler.

• loader/ResourceLoader.cpp:

(WebCore::logResourceResponseSource):

• page/DiagnosticLoggingKeys.cpp:

(WebCore::DiagnosticLoggingKeys::memoryCacheKey):
(WebCore::DiagnosticLoggingKeys::memoryCacheAfterValidationKey):

• page/DiagnosticLoggingKeys.h:

10:43 AM Changeset in webkit [266704] by Darin Adler

• 2 edits in trunk/LayoutTests

• platform/mac-wk2/TestExpectations: Revert change I accidentally included in my last patch.

10:11 AM Changeset in webkit [266703] by Darin Adler

• 14 edits in trunk

Make TextCodecCJK and TextCodecSingleByte thread-safe and refactor a bit to share code
​https://bugs.webkit.org/show_bug.cgi?id=216229

Reviewed by Sam Weinig.

LayoutTests/imported/w3c:

• web-platform-tests/encoding/eof-utf-8-one.html:
• web-platform-tests/encoding/eof-utf-8-three.html:
• web-platform-tests/encoding/eof-utf-8-two.html:

Updated these files with fresh copies. Somehow, the trailing invalid
UTF-8 sequences from these files must have gotten lost when we
re-synced web-platform-tests/encoding from upstream. If we do that
again, then these tests will start failing again.

Source/WebCore:

The text encoding machinery is usable on multiple threads, but our new TextCodec
classes have some global data structures that need to be guarded to keep that intact.

• platform/text/EncodingTables.cpp:

(WebCore::checkEncodingTableInvariants): One-time check for invariants that clients
of these tables depend on.

• platform/text/EncodingTables.h: Added checkEncodingTableInvariants. Also Added

function templates for encoding tables: Added findFirstInSortedPairs,
findLastInSortedPairs and findInSortedPairs for searching a sorted array of pairs
used as a map. Added sortByFirst to aid in the creation of such a sorted array.
And added isSortedByFirst and sortedFirstsAreUnique so we can assert those invariants.
One of the good features of the findInSortedPairs functions is that they handle integer
values that don't fit in an integral key type, returning WTF::nullopt in that case.
That lets us pass code point values when looking in tables that use code units as their
key, without separately checking if they are in range, which otherwise requires
converting to UChar and checking for equality or calling U_IS_BMP.

• platform/text/TextCodecCJK.cpp:

(WebCore::TextCodecCJK::TextCodecCJK): Call checkEncodingTableInvariants.
(WebCore::TextCodecCJK::encode const): Ditto.
(WebCore::jis0208DecodeIndex): Use std::call_once for thread safety.
Use sortByFirst and sortedFirstsAreUnique.
(WebCore::codePointJIS0208): Use findFirstInSortedPairs.
(WebCore::codePointJIS0212): Ditto.
(WebCore::TextCodecCJK::eucJPDecode): Cast to char instad of LChar when adding
an ASCII character to a StringBuilder.
(WebCore::eucJPEncode): Use the name codePoint instead of c to match other
surrounding code. Use findLastInSortedPairs.
(WebCore::iso2022JPEncode): Removed some unneeded casts to uint8_t when appending
bytes to a Vector. Added a static_assert to check the size of iso2022JPKatakana.
Use findLastInSortedPairs.
(WebCore::shiftJISEncode): Use findInSortedPairs.
(WebCore::eucKREncodingIndex): Use std::call_once for thread safety.
Use sortByFirst and sortedFirstsAreUnique.
(WebCore::eucKREncode): Removed some unneeded casts to uint8_t when appending
bytes to a Vector. Use findFirstInSortedPairs.
(WebCore::TextCodecCJK::eucKRDecode): Use findFirstInSortedPairs.
(WebCore::big5Encode): Use findInSortedPairs. Also renamed c to codePoint.
(WebCore::big5DecodeIndex): Use std::call_once for thread safety.
Use sortByFirst and sortedFirstsAreUnique.
(WebCore::TextCodecCJK::big5Decode): Use findFirstInSortedPairs. Cast to char
instad of LChar when adding an ASCII character to a StringBuilder.

• platform/text/TextCodecSingleByte.cpp:

(WebCore::tableForEncoding): Return an IteratorRange instead of a pair of
pointer and size. This works with std::begin/end. Also make table with actual
encoded bytes, by adding 0x80 here, rather than doing that when using the table.
(WebCore::encode): Use std::call_once for thread safety.
Use sortByFirst and sortedFirstsAreUnique. The code before was not sorting at
all, which means it probably didn't work in any cases where the code units
happen to not be in ascending order. We should add some test cases.
(WebCore::decode): Use findFirstInSortedPairs. Also use StringView::codePoints
because these are likely to be 8-bit strings and we don't need to temporarily
upconvert them to 16-bit just to encode them. Should probably later measure if
the use of StringView::upconvertedCharacters plus CodePointIterator<UChar>
instead of StringView::codePoints is better for performance in the CJK encoding
functions. This approach means more branching inside the loop, but the other
version involves memory allocation and a second loop when the characters are
all 8-bit.

Tools:

• TestWebKitAPI/Tests/WTF/StringView.cpp:

(TestWebKitAPI::TEST): Removed an unused variable. At one point, with
code that I wrote and now have rolled out, this was causing a build failure.

LayoutTests:

• TestExpectations: Removed expectation that skips the three eof-utf-8

tests. They were failing because the tests were imported into the source
tree improperly and are passing now that is fixed.

10:02 AM Changeset in webkit [266702] by commit-queue@webkit.org

• 2 edits in trunk/Source/WebKit

Build failure; cannot find seccomp.h
​https://bugs.webkit.org/show_bug.cgi?id=216226

Patch by Mike Gorse <​mgorse@suse.com> on 2020-09-07
Reviewed by Darin Adler.

No new tests (build fix only).

• PlatformGTK.cmake: Add LIBSECCOMP_INCLUDE_DIRS to

WebKit_SYSTEM_INCLUDE_DIRECTORIES.

9:11 AM Changeset in webkit [266701] by youenn@apple.com

• 5 edits in trunk

Fix Internals::supportsVCPEncoder on BigSur
​https://bugs.webkit.org/show_bug.cgi?id=216174

Reviewed by Eric Carlson.

Source/WebCore:

Fixed by unflaked test.

• testing/Internals.cpp:

(WebCore::Internals::supportsVCPEncoder):
All MacOS supported versions now supports VCP.

LayoutTests:

• platform/mac-wk2/TestExpectations:
• platform/mac/webrtc/captureCanvas-webrtc-software-encoder-expected.txt:
• platform/mac/webrtc/captureCanvas-webrtc-software-encoder.html:

Slight modification to make the test run faster and output more comprehensive error message if error happens.

9:11 AM Changeset in webkit [266700] by youenn@apple.com

• 2 edits in trunk/Source/WebCore

Update IDLAttributes.json to forbid Exposed on dictionary
​https://bugs.webkit.org/show_bug.cgi?id=216243

Reviewed by Alex Christensen.

• bindings/scripts/IDLAttributes.json:

8:51 AM Changeset in webkit [266699] by youenn@apple.com

• 7 edits
  12 adds in trunk

Safari takes too long to fetch images from memory cache
​https://bugs.webkit.org/show_bug.cgi?id=216048
<rdar://problem/68260952>

Reviewed by Antti Koivisto.

LayoutTests/imported/w3c:

• web-platform-tests/fetch/stale-while-revalidate/stale-image.html:
• web-platform-tests/html/dom/elements/images/bypass-cache-revalidation-expected.txt: Added.
• web-platform-tests/html/dom/elements/images/bypass-cache-revalidation.html: Added.
• web-platform-tests/html/dom/elements/images/image.py: Added.

(main):

Source/WebCore:

In case of image resource, we skip revalidation if the stored image is from the same document.
This is inline with ​https://html.spec.whatwg.org/#updating-the-image-data:list-of-available-images
which defines a list of available images for each Document.
In case fetch mode is different, we do not skip revalidation as we might otherwise bypass security checks.

Tests: http/wpt/html/dom/elements/images/bypass-cache-redirection-revalidation.html

http/wpt/html/dom/elements/images/hover-image-change.html
imported/w3c/web-platform-tests/html/dom/elements/images/bypass-cache-revalidation.html

• loader/cache/CachedImage.cpp:

(WebCore::CachedImage::load):
(WebCore::CachedImage::canSkipRevalidation const):

• loader/cache/CachedImage.h:
• loader/cache/CachedResourceLoader.cpp:

(WebCore::CachedResourceLoader::determineRevalidationPolicy const):

LayoutTests:

• http/wpt/html/dom/elements/images/bypass-cache-redirection-revalidation-expected.txt: Added.
• http/wpt/html/dom/elements/images/bypass-cache-redirection-revalidation.html: Added.
• http/wpt/html/dom/elements/images/hover-image-change-expected.html: Added.
• http/wpt/html/dom/elements/images/hover-image-change.html: Added.
• http/wpt/html/dom/elements/images/redirect.py: Added.

5:37 AM Changeset in webkit [266698] by youenn@apple.com

• 27 edits
  8 adds in trunk

Add missing members to RTCIceCandidate
​https://bugs.webkit.org/show_bug.cgi?id=216075

Reviewed by Eric Carlson.

LayoutTests/imported/w3c:

• web-platform-tests/webrtc/RTCIceCandidate-constructor-expected.txt:
• web-platform-tests/webrtc/RTCIceCandidate-constructor.html:
• web-platform-tests/webrtc/idlharness.https.window-expected.txt:

Source/ThirdParty/libwebrtc:

• Configurations/libwebrtc.iOS.exp:
• Configurations/libwebrtc.iOSsim.exp:
• Configurations/libwebrtc.mac.exp:

Source/WebCore:

Sync IDL with latest spec.
Add missing enums and make use of libwebrtc parse routine to get fields from candidate SDP.
Covered by updated tests.

• CMakeLists.txt:
• DerivedSources-input.xcfilelist:
• DerivedSources-output.xcfilelist:
• DerivedSources.make:
• Modules/mediastream/RTCIceCandidate.cpp:

(WebCore::RTCIceCandidate::RTCIceCandidate):
(WebCore::RTCIceCandidate::create):

• Modules/mediastream/RTCIceCandidate.h:
• Modules/mediastream/RTCIceCandidate.idl:
• Modules/mediastream/RTCIceCandidateInit.h:
• Modules/mediastream/RTCIceCandidateInit.idl:
• Modules/mediastream/RTCIceCandidateType.h: Added.
• Modules/mediastream/RTCIceCandidateType.idl: Added.
• Modules/mediastream/RTCIceComponent.h: Added.
• Modules/mediastream/RTCIceComponent.idl: Added.
• Modules/mediastream/RTCIceProtocol.h: Added.
• Modules/mediastream/RTCIceProtocol.idl: Added.
• Modules/mediastream/RTCIceTcpCandidateType.h: Added.
• Modules/mediastream/RTCIceTcpCandidateType.idl: Added.
• Modules/mediastream/RTCStatsReport.h:
• Modules/mediastream/RTCStatsReport.idl:
• Modules/mediastream/libwebrtc/LibWebRTCStatsCollector.cpp:

(WebCore::iceCandidateState):
(WebCore::fillRTCIceCandidateStats):

• Modules/mediastream/libwebrtc/LibWebRTCUtils.cpp:

(WebCore::toRTCIceComponent):
(WebCore::toRTCIceProtocol):
(WebCore::toRTCIceTcpCandidateType):
(WebCore::toRTCIceCandidateType):
(WebCore::parseIceCandidateSDP):

• Sources.txt:
• WebCore.xcodeproj/project.pbxproj:
• testing/MockLibWebRTCPeerConnection.cpp:

(WebCore::MockLibWebRTCPeerConnectionForIceCandidates::sendCandidates):

LayoutTests:

• fast/mediastream/RTCPeerConnection-icecandidate-event-expected.txt:

3:09 AM Changeset in webkit [266697] by commit-queue@webkit.org

• 6 edits in trunk/Source/WebCore

[GStreamer] Convert custom GObject subclasses to WEBKIT_DEFINE_TYPE
​https://bugs.webkit.org/show_bug.cgi?id=204673

Patch by Víctor Manuel Jáquez Leal <​vjaquez@igalia.com> on 2020-09-07
Reviewed by Philippe Normand.

Instead of using G_DEFINE_TYPE macro use rather WEBKIT_DEFINE_TYPE
since it handles private structures.

No new tests since no functional changes.

• platform/audio/gstreamer/WebKitWebAudioSourceGStreamer.cpp:

Renamed the private structure to match with class name so it can
be autogeneterated by the macro.
(_WebKitWebAudioSrcPrivate::_WebKitWebAudioSrcPrivate): add
constructure.
(_WebKitWebAudioSrcPrivate::~_WebKitWebAudioSrcPrivate): add
destructor.
(webkit_web_audio_src_class_init): remove duplicated
initialization.
(webKitWebAudioSrcConstructed): add constructed vmethod to
initalize some private members.
(webKitWebAudioSrcSetProperty): renamed private structure.
(webKitWebAudioSrcGetProperty): ditto.
(webKitWebAudioSrcAllocateBuffersAndRenderAudio): ditto.
(webKitWebAudioSrcLoop): ditto.

• platform/graphics/gstreamer/VideoSinkGStreamer.cpp:

(webkitVideoSinkConstructed): add constructed vmethod to initalize
some private members.
(webkit_video_sink_class_init): remove duplicated initialization.

• platform/graphics/gstreamer/eme/WebKitClearKeyDecryptorGStreamer.cpp:

(webkit_media_clear_key_decrypt_class_init): remove duplicated
initialization.

• platform/graphics/gstreamer/eme/WebKitCommonEncryptionDecryptorGStreamer.cpp:

(webkit_media_common_encryption_decrypt_class_init): remove
duplicated initialization.
(constructed): add constructed vmethod to initalize some private
members.

• platform/graphics/gstreamer/eme/WebKitThunderDecryptorGStreamer.cpp:

(webkit_media_thunder_decrypt_class_init): remove duplicated
initialization.

2:54 AM Changeset in webkit [266696] by svillar@igalia.com

• 7 edits in trunk

[css-flex] Allow indefinite size flex items to be definite wrt resolving percentages inside them
​https://bugs.webkit.org/show_bug.cgi?id=212264

Reviewed by Manuel Rego Casasnovas.

LayoutTests/imported/w3c:

• web-platform-tests/css/css-flexbox/percentage-heights-003-expected.txt: There were 3 subtests

marked as failures. All of them work fine now.

Source/WebCore:

Implement ​https://github.com/w3c/csswg-drafts/commit/5b5db39d21f3658ae2f4d7992daaf822aca178d8 which modified
the way percentages were resolved in flexible items with indefinite sizes. From now on we can pretend that
they're really definite.

This allows us to mark 3 tests which were testing percentages in flex items as correct.

Based on Blink's crrev.com/1247184 by <​cbiesinger@chromium.org>

This is a reland of r263399 which got reverted due to bug 214655. This same patch was previously relanded in
r262124 which got reverted due to the bug fixed in r263389.

• rendering/RenderFlexibleBox.cpp:

(WebCore::RenderFlexibleBox::mainSizeForPercentageResolution): Do only check flex container main size
definiteness when computing the main size for percentage resolution, no need to check flex basis at all.

LayoutTests:

• css3/flexbox/definite-main-size-expected.txt: Updated expectation.
• css3/flexbox/definite-main-size.html: Updated comment.

2:20 AM Changeset in webkit [266695] by svillar@igalia.com

• 7 edits
  2 deletes in trunk

[css-flexbox] min-height: auto not applied to nested flexboxes.
​https://bugs.webkit.org/show_bug.cgi?id=210089

Reviewed by Daniel Bates.

Source/WebCore:

Nested flexboxes with column direction were not computing correctly min-size:auto because
we were explicitly preventing them from doing so in the code. Implemented the required bits to
make it work correctly and thus removed the retriction. The idea is to set an indefinite override
containing block size so that percentages would be resolved to auto as spec'ed. The code which
decides whether to apply min-size:auto was refactored in the shouldApplyMinSizeAutoForChild() method.

In order not to cause regressions some other two additional changes were also implemented. First we
had to adjust childHasIntrinsicMainAxisSize() so that it also takes into account the cases where
shouldApplyMinSizeAutoForChild() is true and return true. Secondly we had to add an additional case
to mainAxisLengthIsDefinite() so that it returns false for column flows where the flexBasis is intrinsic.

Inspired by Blink's crrev.com/c/1641510, crrev.com/c/1269995 & crrev.com/c/1786297 by <​cbiesinger@chromium.org>

• rendering/RenderFlexibleBox.cpp:

(WebCore::RenderFlexibleBox::shouldApplyMinSizeAutoForChild const): Refactored from
RenderFlexibleBox::adjustChildSizeForMinAndMax.
(WebCore::RenderFlexibleBox::mainAxisLengthIsDefinite const): Additional case for column flows.
(WebCore::RenderFlexibleBox::layoutFlexItems): Reset m_hasDefiniteHeight to Unknown after calling
constructFlexItem() because the latter might set now an override containing block height which basically
potentially makes any cached size value incorrect.
(WebCore::RenderFlexibleBox::adjustChildSizeForMinAndMax): Directly call shouldApplyMinSizeAutoForChild().
(WebCore::RenderFlexibleBox::constructFlexItem): Set an indefinite override containing block size for
children with percentage sizes so that they're resolved as auto.
(WebCore::RenderFlexibleBox::childHasIntrinsicMainAxisSize const): Return true for those cases where the
main axis length is indefinite and also when shouldApplyMinSizeAutoForChild().

• rendering/RenderFlexibleBox.h:

LayoutTests:

Apart from enabling some tests we're removing a test which is now invalid as it was added
under the condition that we were not matching the specs wrt percentage height computation in
column flexboxes.

• TestExpectations: Removed two test cases that are passing now.
• fast/flexbox/nested-column-intrinsic-min-disabled-expected.html: Removed.
• fast/flexbox/nested-column-intrinsic-min-disabled.html: Removed.
• platform/ios/fast/forms/auto-fill-button/input-strong-password-auto-fill-button-expected.txt: Updated.
• platform/mac/fast/forms/auto-fill-button/input-strong-password-auto-fill-button-expected.txt: Ditto.

1:28 AM Changeset in webkit [266694] by jh718.park@samsung.com

• 2 edits in trunk/Source/JavaScriptCore

Unreviewed. Remove the build warning below since r266567.
warning: parameter ‘hint’ set but not used [-Wunused-but-set-parameter]

• runtime/JSObject.cpp:

(JSC::callToPrimitiveFunction):