Timeline

Dec 3, 2021:

10:12 PM Changeset in webkit [286526] by Said Abou-Hallawa

• 13 edits in trunk

[GPU Process] (REGRESSION r285597): Fix SVGFilter clamping calculation
​https://bugs.webkit.org/show_bug.cgi?id=233843

Reviewed by Simon Fraser.

Source/WebCore:

In this patch:

1. lastEffect() is made virtual function of Filter and it's already implemented as non-virtual function in both CSSFilter and SVGFilter.

2. The clamping calculations for SVGFilter and CSSFilter are merged in one function called Filter::clampFilterRegionIfNeeded(). This merge makes the new function is the only caller to FilterEffect::maxEffectRect().

3. Remove FilterEffect::maxEffectRect() . It can be calculated by clipping the FilterEffect::filterPrimitiveSubregion() to filterRegion() and then scaling the result by filterScale().

4. In RenderSVGResourceFilter::applyResource() we need to pass the non- scaled drawingRegion to ImageBuffer::sizeNeedsClamping() because this function starts by scaling the argument 'size' by the argument 'scale'. So we were doubling the scaling by passing 'absoluteDrawingRegion'

5. SVGRenderingContext::createImageBuffer() needs to use ImageBuffer:: sizeNeedsClamping() instead of using ImageBuffer::clampedSize(). The former function clamps the size to the MaxClampedArea. But the later function shrinks the size to { MaxClampedLength, MaxClampedLength }.

When calling createImageBuffer() from RenderSVGResourceFilter

applyResource() we do not expect any further clamping. But if we pass a size with a side greater than MaxClampedLength to clampedSize(), it will be clamped even if the area is less than MaxClampedArea. And for filters we clamp to the area.

• platform/graphics/filters/Filter.cpp:

(WebCore::Filter::clampFilterRegionIfNeeded):

• platform/graphics/filters/Filter.h:
• platform/graphics/filters/FilterEffect.cpp:

(WebCore::FilterEffect::determineFilterPrimitiveSubregion):
(WebCore::FilterEffect::apply):

• platform/graphics/filters/FilterEffect.h:

(WebCore::FilterEffect::maxEffectRect const): Deleted.
(WebCore::FilterEffect::setMaxEffectRect): Deleted.

• rendering/CSSFilter.cpp:

(WebCore::CSSFilter::lastEffect const):
(WebCore::CSSFilter::lastEffect): Deleted.
(WebCore::CSSFilter::determineFilterPrimitiveSubregion): Deleted.

• rendering/CSSFilter.h:
• rendering/RenderLayerFilters.cpp:

(WebCore::RenderLayerFilters::beginFilterEffect):

• rendering/svg/RenderSVGResourceFilter.cpp:

(WebCore::RenderSVGResourceFilter::applyResource):

• rendering/svg/SVGRenderingContext.cpp:

(WebCore::SVGRenderingContext::createImageBuffer):

• svg/graphics/filters/SVGFilter.h:

LayoutTests:

Unskip two layout tests which were skipped in r285597.

• TestExpectations:

7:30 PM Changeset in webkit [286525] by Lauro Moura

• 3 edits in trunk/LayoutTests

[GLIB] Gardening some consistent timeouts

Unreviewed test gardening.

• platform/glib/TestExpectations:
• platform/wpe/TestExpectations:

7:08 PM Changeset in webkit [286524] by Russell Epstein

• 1 copy in tags/Safari-612.4.2

Tag Safari-612.4.2.

7:05 PM Changeset in webkit [286523] by Russell Epstein

• 8 edits in branches/safari-612-branch/Source

Versioning.

WebKit-7612.4.2

5:35 PM Changeset in webkit [286522] by Wenson Hsieh

• 4 edits in trunk

[iOS] Web content process sometimes crashes under WebPage::positionInformation()
​https://bugs.webkit.org/show_bug.cgi?id=233841
rdar://85917212

Reviewed by Geoffrey Garen.

Source/WebKit:

Add a missing null check in the case where the hit-tested node is null in WebPage::positionInformation. This
can happen in a number of ways (one of which is exercised by the new API test).

Test: ImageAnalysisTests.DoNotCrashWhenHitTestingOutsideOfWebView

• WebProcess/WebPage/ios/WebPageIOS.mm:

(WebKit::WebPage::positionInformation):

Tools:

Add a test that exercises hit-testing for Live Text, but induces this crash by hit-testing outside of the bounds
of the web view (such that the hit-tested node ends up as null).

• TestWebKitAPI/Tests/WebKitCocoa/ImageAnalysisTests.mm:

(TestWebKitAPI::swizzledLocationInView):
(TestWebKitAPI::TEST):

5:14 PM Changeset in webkit [286521] by Russell Epstein

• 1 copy in tags/Safari-612.4.1

Tag Safari-612.4.1.

4:09 PM Changeset in webkit [286520] by Chris Dumez

• 2 edits in trunk/Source/WebCore

Add BroadcastChannel, COOP, COEP and Web Locks to features.json
​https://bugs.webkit.org/show_bug.cgi?id=233839

Reviewed by Alex Christensen.

• features.json:

4:02 PM Changeset in webkit [286519] by wilander@apple.com

• 24 edits in trunk

PCM: Unlinkable tokens for triggering event to prevent fraud
​https://bugs.webkit.org/show_bug.cgi?id=233673
<rdar://79426347>

Reviewed by Alex Christensen.

Source/WebCore:

This patch allows triggering events to have an optional query parameter like this:
​https://destination.example/.well-known/private-click-measurement/trigger-attribution/11?attributionDestinationNonce=ABCDEFabcdef0123456789

The presence of an attributionDestinationNonce triggers PrivateClickMeasurementManager to
ask the server to sign an unlinkable token for the triggering event and then stores the
corresponding secret token plus the signature for later inclusion in the attribution report.

The WebCore changes introduce these new structs and data fields in
WebCore::PrivateClickMeasurement:

• EphemeralDestinationNonce
• UnlinkableToken (base class for the specific ones for source and destination)
• SourceUnlinkableToken
• DestinationUnlinkableToken
• SecretToken (base class for the specific ones for source and destination)
• SourceSecretToken
• DestinationSecretToken

WebCore::PrivateClickMeasurement::AttributionTriggerData now has these optional fields:

• ephemeralDestinationNonce
• destinationUnlinkableToken
• destinationSecretToken
• destinationSite

The remaining changes are setters/getters and creator functions for those data fields.

New API/unit test PrivateClickMeasurement.DestinationClickFraudPrevention.

• html/HTMLAnchorElement.cpp:

(WebCore::HTMLAnchorElement::parsePrivateClickMeasurement const):

• loader/PrivateClickMeasurement.cpp:

(WebCore::PrivateClickMeasurement::SecretToken::isolatedCopy const):
(WebCore::PrivateClickMeasurement::SourceSecretToken::isolatedCopy const):
(WebCore::PrivateClickMeasurement::DestinationSecretToken::isolatedCopy const):
(WebCore::PrivateClickMeasurement::EphemeralNonce::isolatedCopy const):
(WebCore::PrivateClickMeasurement::UnlinkableToken::isolatedCopy const):
(WebCore::PrivateClickMeasurement::SourceUnlinkableToken::isolatedCopy const):
(WebCore::PrivateClickMeasurement::DestinationUnlinkableToken::isolatedCopy const):
(WebCore::PrivateClickMeasurement::parseAttributionRequestQuery):
(WebCore::PrivateClickMeasurement::parseAttributionRequest):
(WebCore::PrivateClickMeasurement::attributionReportJSON const):
(WebCore::PrivateClickMeasurement::EphemeralNonce::isValid const):
(WebCore::PrivateClickMeasurement::setEphemeralSourceNonce):
(WebCore::PrivateClickMeasurement::tokenPublicKeyURL):
(WebCore::PrivateClickMeasurement::tokenPublicKeyURL const):
(WebCore::PrivateClickMeasurement::tokenSignatureURL):
(WebCore::PrivateClickMeasurement::tokenSignatureURL const):
(WebCore::PrivateClickMeasurement::AttributionTriggerData::tokenSignatureJSON const):
(WebCore::PrivateClickMeasurement::SecretToken::isValid const):
(WebCore::PrivateClickMeasurement::setDestinationSecretToken):
(WebCore::PrivateClickMeasurement::EphemeralSourceNonce::isolatedCopy const): Deleted.
(WebCore::PrivateClickMeasurement::EphemeralSourceNonce::isValid const): Deleted.
(WebCore::PrivateClickMeasurement::SourceSecretToken::isValid const): Deleted.

• loader/PrivateClickMeasurement.h:

(WebCore::PrivateClickMeasurement::AttributionTriggerData::AttributionTriggerData):
(WebCore::PrivateClickMeasurement::AttributionTriggerData::setDestinationUnlinkableTokenValue):
(WebCore::PrivateClickMeasurement::AttributionTriggerData::setDestinationSecretToken):
(WebCore::PrivateClickMeasurement::AttributionTriggerData::tokenPublicKeyURL const):
(WebCore::PrivateClickMeasurement::AttributionTriggerData::tokenSignatureURL const):
(WebCore::PrivateClickMeasurement::ephemeralSourceNonce const):
(WebCore::PrivateClickMeasurement::sourceUnlinkableToken):
(WebCore::PrivateClickMeasurement::decode):
(WebCore::PrivateClickMeasurement::EphemeralNonce::encode const):
(WebCore::PrivateClickMeasurement::EphemeralNonce::decode):
(WebCore::PrivateClickMeasurement::AttributionTriggerData::encode const):
(WebCore::PrivateClickMeasurement::AttributionTriggerData::decode):
(WebCore::PrivateClickMeasurement::EphemeralSourceNonce::encode const): Deleted.
(WebCore::PrivateClickMeasurement::EphemeralSourceNonce::decode): Deleted.

• loader/cocoa/PrivateClickMeasurementCocoa.mm:

(WebCore::PrivateClickMeasurement::calculateAndUpdateSourceUnlinkableToken):
(WebCore::PrivateClickMeasurement::calculateAndUpdateDestinationUnlinkableToken):
(WebCore::PrivateClickMeasurement::calculateAndUpdateUnlinkableToken):
(WebCore::PrivateClickMeasurement::calculateAndUpdateSourceSecretToken):
(WebCore::PrivateClickMeasurement::calculateAndUpdateDestinationSecretToken):
(WebCore::PrivateClickMeasurement::calculateAndUpdateSecretToken):

Source/WebKit:

This patch allows triggering events to have an optional query parameter like this:
​https://destination.example/.well-known/private-click-measurement/trigger-attribution/11?attributionDestinationNonce=ABCDEFabcdef0123456789

The presence of an attributionDestinationNonce triggers PrivateClickMeasurementManager to
ask the server to sign an unlinkable token for the triggering event and then stores the
corresponding secret token plus the signature for later inclusion in the attribution report.

New API/unit test PrivateClickMeasurement.DestinationClickFraudPrevention.

• NetworkProcess/DatabaseUtilities.cpp:

(WebKit::DatabaseUtilities::buildPrivateClickMeasurementFromDatabase const):

Now handles destination tokens.

• NetworkProcess/NetworkResourceLoader.cpp:

(WebKit::NetworkResourceLoader::willSendRedirectedRequest):

Now sets destinationSite in WebCore::PrivateClickMeasurement::AttributionTriggerData.

• NetworkProcess/PrivateClickMeasurement/PrivateClickMeasurementDatabase.cpp:

(WebKit::PCM::Database::Database):

Now calls addDestinationTokenColumnsIfNecessary().

(WebKit::PCM::Database::insertPrivateClickMeasurement):

Now does insertion of optional destination tokens.

(WebKit::PCM::Database::addDestinationTokenColumnsIfNecessary):

Database migration.

(WebKit::PCM::Database::addBundleIDColumnIfNecessary): Deleted.

Temporary migration function for non-customer builds. No longer needed.

• NetworkProcess/PrivateClickMeasurement/PrivateClickMeasurementDatabase.h:
• NetworkProcess/PrivateClickMeasurement/PrivateClickMeasurementManager.cpp:

(WebKit::PrivateClickMeasurementManager::storeUnattributed):

Just name changes.

(WebKit::PrivateClickMeasurementManager::getTokenPublicKey):

Not changed afaik. Seems like the diff is confused.

(WebKit::PrivateClickMeasurementManager::configureForTokenSigning):

New function to extract shared code.

(WebKit::PrivateClickMeasurementManager::getSignatureBase64URLFromTokenSignatureResponse):

New function to extract shared code.

(WebKit::PrivateClickMeasurementManager::getSignedUnlinkableTokenForSource):

Renamed to add "ForSource" and all shared code extracted.

(WebKit::PrivateClickMeasurementManager::getSignedUnlinkableTokenForDestination):

Different than for the source in that it results in an attribution call.

(PrivateClickMeasurementManager::handleAttribution):

Now does signing of destination token if the optional nonce is in the triggering event.

(PrivateClickMeasurementManager::fireConversionRequest):
(WebKit::PrivateClickMeasurementManager::setPCMFraudPreventionValuesForTesting):

Added empty initializer strings for the destination test token values not in use yet.

(WebKit::PrivateClickMeasurementManager::getSignedUnlinkableToken): Deleted.

Renamed PrivateClickMeasurementManager::getSignedUnlinkableTokenForSource().

• NetworkProcess/PrivateClickMeasurement/PrivateClickMeasurementManager.h:

Tools:

New test PrivateClickMeasurement.DestinationClickFraudPrevention.
The other changes are a report version bump 2->3 and a name change
FraudPrevention–>SourceClickFraudPrevention.

• TestWebKitAPI/Tests/WebCore/PrivateClickMeasurement.cpp:

(TestWebKitAPI::TEST):

• TestWebKitAPI/Tests/WebCore/cocoa/PrivateClickMeasurementCocoa.mm:

(TestWebKitAPI::TEST):

• TestWebKitAPI/Tests/WebKitCocoa/EventAttribution.mm:

(TestWebKitAPI::runBasicPCMTest):
(TestWebKitAPI::triggerAttributionWithSubresourceRedirect):

Extracted shared code.

(TestWebKitAPI::runDestinationWebView):

Extracted shared code.

(TestWebKitAPI::signUnlinkableTokenAndSendSecretToken):

Extracted shared code.

(TestWebKitAPI::TEST):

New test PrivateClickMeasurement.DestinationClickFraudPrevention.
Extracted shared code from PrivateClickMeasurement.SourceClickFraudPrevention.

LayoutTests:

These changes are a report version bump 2->3.

• http/tests/privateClickMeasurement/attribution-conversion-through-fetch-keepalive-expected.txt:
• http/tests/privateClickMeasurement/expired-attribution-report-gets-sent-on-session-start-expected.txt:
• http/tests/privateClickMeasurement/multiple-app-bundle-ids-expected.txt:
• http/tests/privateClickMeasurement/send-attribution-conversion-request-expected.txt:
• http/tests/privateClickMeasurement/store-private-click-measurement-with-source-nonce-expected.txt:
• http/tests/privateClickMeasurement/triggering-event-with-attribution-source-through-fetch-keepalive-expected.txt:
• platform/ios/http/tests/privateClickMeasurement/attribution-conversion-through-fetch-keepalive-expected.txt:

3:42 PM Changeset in webkit [286518] by Alan Bujtas

• 4 edits in trunk/Source/WebCore

[LFC][IFC] Move the (bidi)display boxes horizontally by the inline box margin, border and padding end as needed
​https://bugs.webkit.org/show_bug.cgi?id=233744

Reviewed by Antti Koivisto.

Keep track of the inline box range in the context of display boxes so that we can adjust the content horizontally
with the margin/border/padding end of the "last" inline box.
We also set the width of the (fragmented inline box type of) display boxes here based on their content right position.

• layout/formattingContexts/inline/InlineDisplayContentBuilder.cpp:

(WebCore::Layout::InlineDisplayContentBuilder::processBidiContent):

• layout/formattingContexts/inline/display/InlineDisplayBox.h:

(WebCore::InlineDisplay::Box::setLogicalRight):

• layout/layouttree/LayoutBoxGeometry.h:

(WebCore::Layout::BoxGeometry::borderAndPaddingEnd const):

3:21 PM Changeset in webkit [286517] by commit-queue@webkit.org

• 3 edits in trunk/Source/WebKit

Fix playstation build after r286484
​https://bugs.webkit.org/show_bug.cgi?id=233823

unreviewed build fix.

Patch by ujwal koneru <​ujwal.koneru@sony.com> on 2021-12-03

• UIProcess/playstation/WebProcessPoolPlayStation.cpp:

(WebKit::WebProcessPool::platformInitializeWebProcess):

• WebProcess/playstation/WebProcessPlayStation.cpp:

(WebKit::WebProcess::platformInitializeWebProcess):

2:48 PM Changeset in webkit [286516] by fpizlo@apple.com

• 5 edits
  1 add in trunk/Source/bmalloc

[libpas] Bitfit allocator has a wrong assertion when a page's max_free is enough for the size of an allocation, not enough for that allocation's size class, and the object of that size is not aligned to the currently requested alignment
​https://bugs.webkit.org/show_bug.cgi?id=233831

Reviewed by Yusuke Suzuki.

What a combination of conditions:

• We just failed bitfit allocation in a page, which gives us some max_free (aka largest_available), and the allocation had nontrivial alignment.
• The max_free is smaller than the size class.
• The max_free is larger than the requested size.
• The max_free object is not aligned to the requested alignment.

The code handles this fine, but has a wrong assertion about it.

This change fixes the assertion and adds a test that deterministically reproduced the issue.

• libpas/libpas.xcodeproj/project.pbxproj:
• libpas/src/libpas/pas_bitfit_allocator.c:

(pas_bitfit_allocator_finish_failing):

• libpas/src/libpas/pas_bitfit_allocator_inlines.h:

(pas_bitfit_allocator_try_allocate):

• libpas/src/test/BitfitTests.cpp: Added.

(std::getBitfitSizeClasses):
(std::assertSizeClasses):
(std::testAllocateAlignedSmallerThanSizeClassAndSmallerThanLargestAvailable):
(addBitfitTests):

• libpas/src/test/TestHarness.cpp:

(main):

1:58 PM Changeset in webkit [286515] by Russell Epstein

• 1 copy in tags/Safari-613.1.10

Tag Safari-613.1.10.

1:46 PM Changeset in webkit [286514] by Chris Dumez

• 2 edits in trunk/Source/WTF

[WK2] Turn on Web Locks API support
​https://bugs.webkit.org/show_bug.cgi?id=233816

Reviewed by Darin Adler.

• Scripts/Preferences/WebPreferencesExperimental.yaml:

1:14 PM Changeset in webkit [286513] by timothy_horton@apple.com

• 2 edits in trunk/Source/WebKit

Fix the build

• NetworkProcess/storage/NetworkStorageManager.cpp:

1:10 PM Changeset in webkit [286512] by timothy_horton@apple.com

• 4 edits in trunk/Source/WebKit

Add more logging for MomentumEventDispatcher
​https://bugs.webkit.org/show_bug.cgi?id=233811

Reviewed by Simon Fraser.

Add a temporary event log to MomentumEventDispatcher, to debug delta/offset/curve issues.

• WebProcess/WebPage/EventDispatcher.h:

(WebKit::EventDispatcher::queue):

• WebProcess/WebPage/MomentumEventDispatcher.cpp:

(WebKit::MomentumEventDispatcher::handleWheelEvent):
Accumulate event deltas. Also accumulate event deltas for the
fingers-down phase in the "generated" offset. Store the phase and
momentum phase smooshed into a single field.

(WebKit::MomentumEventDispatcher::dispatchSyntheticMomentumEvent):
Accumulate generated deltas.

(WebKit::MomentumEventDispatcher::didEndMomentumPhase):
Attempt to dump the log 1 second after each momentum phase. We'll skip
it if another scroll has started since.

(WebKit::MomentumEventDispatcher::setScrollingAccelerationCurve):
(WebKit::MomentumEventDispatcher::startDisplayLink):
(WebKit::MomentumEventDispatcher::stopDisplayLink):
(WebKit::MomentumEventDispatcher::consumeDeltaForCurrentTime):
(WebKit::MomentumEventDispatcher::buildOffsetTableWithInitialDelta):
(WebKit::MomentumEventDispatcher::computeNextDelta):
Adopt more MOMENTUM_EVENT_DISPATCHER_TEMPORARY_LOGGING. Error logging
we leave outside of this, because that we'll keep around.

(WebKit::MomentumEventDispatcher::pushLogEntry):
(WebKit::MomentumEventDispatcher::flushLog):
Dump the event log in an easy-to-copy-into-a-CSV format.

• WebProcess/WebPage/MomentumEventDispatcher.h:

12:37 PM Changeset in webkit [286511] by mmaxfield@apple.com

• 2 edits in trunk/Source/WebCore

Fix internal Apple builds

Unreviewed.

Turns out -weak_framework works if the framework isn't present at runtime, but
still requires the framework to be present at build time.

No new tests because there is no behavior change.

• Configurations/WebCore.xcconfig:

12:29 PM Changeset in webkit [286510] by Russell Epstein

• 11 edits in branches/safari-612-branch

Cherry-pick r286505. rdar://problem/85918531

Follow-up to r286479 to add API test and address issues found by the test
​https://bugs.webkit.org/show_bug.cgi?id=233798

Reviewed by Darin Adler.

Source/WebKit:

Add functionality needed for API testing and fix issues found by the API test.

• UIProcess/API/Cocoa/WKWebViewPrivateForTesting.h:
• UIProcess/API/Cocoa/WKWebViewTesting.mm: (-[WKWebView _isLayerTreeFrozenForTesting:]):
• UIProcess/WebPageProxy.cpp: (WebKit::WebPageProxy::destroyProvisionalPage): (WebKit::WebPageProxy::isLayerTreeFrozen):
• UIProcess/WebPageProxy.h:
• WebProcess/WebPage/WebPage.cpp: (WebKit::WebPage::isLayerTreeFrozen):
• WebProcess/WebPage/WebPage.h:
• WebProcess/WebPage/WebPage.messages.in: Add new SPI to check if the layer tree is frozen in the WebProcess so that I could write an API test for this.

• UIProcess/ProvisionalPageProxy.cpp: (WebKit::ProvisionalPageProxy::ProvisionalPageProxy): Make sure m_provisionalLoadURL gets initialized when the ProvisionalPageProxy gets constructed *after* the provisional load has already started (which is the case when the process swap is triggered by COOP).

(WebKit::ProvisionalPageProxy::didFailProvisionalLoadForFrame):
If the provisional load load fails in the provisional process, and the ProvisionalPageProxy
was constructed on resource response (COOP case), then no longer forward the
didFailProvisionalLoadForFrame() to the WebPageProxy. Instead, we destroy the
ProvisionalPageProxy. This is to avoid duplicate calls to didFailProvisionalLoadForFrame().
In this case, there is still a provisional load ongoing in the committed process and the
ProvisionalPageProxy destructor will take care of stopping that provisional load (due to
r286479), which will cause the committed process to send its own
didFailProvisionalLoadForFrame IPC.

Tools:

Add API test coverage.

• TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@286505 268f45cc-cd09-0410-ab3c-d52691b4dbfc

12:29 PM Changeset in webkit [286509] by Russell Epstein

• 4 edits in branches/safari-612-branch/Source/WebKit

Cherry-pick r286479. rdar://problem/85918531

Regression(r283179) Google Drive freezes after downloading a folder
​https://bugs.webkit.org/show_bug.cgi?id=233783
<rdar://85918531>

Reviewed by Darin Adler.

When process-swapping on a navigation response due to COOP, we create a new ProvisionalPageProxy
to trigger a provisional load in a new WebProcess. In the common case, the ProvisionalPageProxy
gets committed, we process-swap and everything works fine. However, if the client decides to
convert the navigation into a download (like in the Google Drive case), then the
ProvisionalPageProxy gets destroyed without being committed. The issue is that the committed
process still thinks at this point that it is in the middle of a navigation and its layer
tree is thus frozen with reason=PageTransition. This is what makes Google Drive look frozen
after the download. This is not an issue with PSON because the process-swap happens in
decidePolicyForNavigationAction() and we tell the previous process to ignore the navigation
when we process-swap.

To address the issue, we now tell the committed process to cancel its navigation if the
ProvisionalPageProxy ends up getting destroyed without being committed. This lets the
committed process know there is no point in waiting for this navigation to happen and allows
it to unfreeze its layer tree.

• UIProcess/ProvisionalPageProxy.cpp: (WebKit::ProvisionalPageProxy::ProvisionalPageProxy): (WebKit::ProvisionalPageProxy::~ProvisionalPageProxy):
• UIProcess/ProvisionalPageProxy.h: (WebKit::ProvisionalPageProxy::isProcessSwappingOnNavigationResponse const): (WebKit::ProvisionalPageProxy::shouldClosePreviousPageAfterCommit const): Deleted.
• UIProcess/WebPageProxy.cpp: (WebKit::WebPageProxy::commitProvisionalPage): (WebKit::WebPageProxy::continueNavigationInNewProcess):

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@286479 268f45cc-cd09-0410-ab3c-d52691b4dbfc

12:23 PM Changeset in webkit [286508] by Robert Jenner

• 6 edits in trunk/Source/WebKit

Unreviewed, reverting r286487.

Broke multiple builds.

Reverted changeset:

"Rename DerivedSources/WebKit2 to DerivedSources/WebKit"
​https://bugs.webkit.org/show_bug.cgi?id=233803
​https://commits.webkit.org/r286487

12:17 PM Changeset in webkit [286507] by sihui_liu@apple.com

• 29 edits
  1 add in trunk

Fetch and remove file system data via WKWebsiteDataStore
​https://bugs.webkit.org/show_bug.cgi?id=233567

Reviewed by Youenn Fablet.

Source/WebCore:

• Modules/filesystemaccess/FileSystemStorageConnection.h:

Source/WebKit:

Introduce a new WebsiteDataType value FileSystem for FileSystemAccess data. Network process now can fetch and
delete this type of data when fetching and deleteing website data (if FileSystem type is included in target
types).

To track origins that have FileSystem data, this patch introduces a new file named origin in the origin's
directory. This file will be created when OriginStorageManager is created.

To delete existing FileSystem data, network process finds origins that are requested to be deleted and have data
on disk, closes active access handles, and deletes the files. The origin file mentioned above will be deleted if
there is no other file left in the same directory, and empty directories will be deleted.

New API tests: FileSystemAccess.FetchAndRemoveData

FileSystemAccess.RemoveDataByModificationTime
FileSystemAccess.FetchDataForThirdParty

• CMakeLists.txt:
• DerivedSources-input.xcfilelist:
• DerivedSources-output.xcfilelist:
• DerivedSources.make:
• NetworkProcess/Classifier/WebResourceLoadStatisticsStore.cpp:

(WebKit::WebResourceLoadStatisticsStore::monitoredDataTypes):

• NetworkProcess/NetworkProcess.cpp:

(WebKit::NetworkProcess::fetchWebsiteData):
(WebKit::NetworkProcess::deleteWebsiteData):
(WebKit::NetworkProcess::deleteWebsiteDataForOrigins):
(WebKit::NetworkProcess::deleteAndRestrictWebsiteDataForRegistrableDomains):
(WebKit::NetworkProcess::registrableDomainsWithWebsiteData):

• NetworkProcess/storage/FileSystemStorageHandle.h:

(WebKit::FileSystemStorageHandle::activeSyncAccessHandle const):

• NetworkProcess/storage/FileSystemStorageManager.cpp:

(WebKit::FileSystemStorageManager::~FileSystemStorageManager):
(WebKit::FileSystemStorageManager::close):

• NetworkProcess/storage/FileSystemStorageManager.h:
• NetworkProcess/storage/NetworkStorageManager.cpp:

(WebKit::readOriginFromFile):
(WebKit::writeOriginToFileIfNecessary):
(WebKit::deleteOriginFileIfNecessary):
(WebKit::originDirectoryPath):
(WebKit::originFilePath):
(WebKit::NetworkStorageManager::localOriginStorageManager):
(WebKit::NetworkStorageManager::removeOriginStorageManagerIfPossible):
(WebKit::toWebsiteDataType):
(WebKit::NetworkStorageManager::forEachOriginDirectory):
(WebKit::NetworkStorageManager::fetchDataFromDisk):
(WebKit::NetworkStorageManager::fetchData):
(WebKit::NetworkStorageManager::deleteDataOnDisk):
(WebKit::NetworkStorageManager::deleteData):
(WebKit::NetworkStorageManager::deleteDataModifiedSince):
(WebKit::NetworkStorageManager::deleteDataForRegistrableDomains):
(WebKit::originPath): Deleted.

• NetworkProcess/storage/NetworkStorageManager.h:
• NetworkProcess/storage/OriginStorageManager.cpp:

(WebKit::OriginStorageManager::StorageBucket::toStorageIdentifier):
(WebKit::OriginStorageManager::StorageBucket::typeStoragePath const):
(WebKit::OriginStorageManager::StorageBucket::fileSystemStorageManager):
(WebKit::OriginStorageManager::StorageBucket::isActive):
(WebKit::OriginStorageManager::StorageBucket::deleteData):
(WebKit::OriginStorageManager::StorageBucket::deleteFileSystemStorageData):
(WebKit::OriginStorageManager::OriginStorageManager):
(WebKit::OriginStorageManager::isActive):
(WebKit::OriginStorageManager::deleteData):

• NetworkProcess/storage/OriginStorageManager.h:
• Shared/WebsiteData/WebsiteData.cpp:

(WebKit::WebsiteData::ownerProcess):

• Shared/WebsiteData/WebsiteDataType.h:
• UIProcess/API/Cocoa/WKWebsiteDataRecord.mm:

(dataTypesToString):

• UIProcess/API/Cocoa/WKWebsiteDataRecordInternal.h:

(WebKit::toWebsiteDataType):
(WebKit::toWKWebsiteDataTypes):

• UIProcess/API/Cocoa/WKWebsiteDataRecordPrivate.h:
• UIProcess/API/Cocoa/WKWebsiteDataStore.mm:

(+[WKWebsiteDataStore _allWebsiteDataTypesIncludingPrivate]):

• WebKit.xcodeproj/project.pbxproj:
• WebProcess/Network/NetworkProcessConnection.cpp:

(WebKit::NetworkProcessConnection::didReceiveMessage):

• WebProcess/WebCoreSupport/WebFileSystemStorageConnection.messages.in: Added.

Source/WTF:

• wtf/FileSystem.cpp:

(WTF::FileSystemImpl::readEntireFile): Read whole file content into a Vector.
(WTF::FileSystemImpl::deleteAllFilesModifiedSince): Recursively delete files and folders modified after
specified time in a directory.

• wtf/FileSystem.h:

Tools:

• TestWebKitAPI/Tests/WebKitCocoa/FileSystemAccess.mm:

11:52 AM Changeset in webkit [286506] by Russell Epstein

• 8 edits in branches/safari-613.1.8-branch/Source

Versioning.

WebKit-7613.1.8.4

11:39 AM Changeset in webkit [286505] by Chris Dumez

• 11 edits in trunk

Follow-up to r286479 to add API test and address issues found by the test
​https://bugs.webkit.org/show_bug.cgi?id=233798

Reviewed by Darin Adler.

Source/WebKit:

Add functionality needed for API testing and fix issues found by the API test.

• UIProcess/API/Cocoa/WKWebViewPrivateForTesting.h:
• UIProcess/API/Cocoa/WKWebViewTesting.mm:

(-[WKWebView _isLayerTreeFrozenForTesting:]):

• UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::destroyProvisionalPage):
(WebKit::WebPageProxy::isLayerTreeFrozen):

• UIProcess/WebPageProxy.h:
• WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::isLayerTreeFrozen):

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

Add new SPI to check if the layer tree is frozen in the WebProcess so that I could
write an API test for this.

• UIProcess/ProvisionalPageProxy.cpp:

(WebKit::ProvisionalPageProxy::ProvisionalPageProxy):
Make sure m_provisionalLoadURL gets initialized when the ProvisionalPageProxy gets
constructed *after* the provisional load has already started (which is the case
when the process swap is triggered by COOP).

(WebKit::ProvisionalPageProxy::didFailProvisionalLoadForFrame):
If the provisional load load fails in the provisional process, and the ProvisionalPageProxy
was constructed on resource response (COOP case), then no longer forward the
didFailProvisionalLoadForFrame() to the WebPageProxy. Instead, we destroy the
ProvisionalPageProxy. This is to avoid duplicate calls to didFailProvisionalLoadForFrame().
In this case, there is still a provisional load ongoing in the committed process and the
ProvisionalPageProxy destructor will take care of stopping that provisional load (due to
r286479), which will cause the committed process to send its own
didFailProvisionalLoadForFrame IPC.

Tools:

Add API test coverage.

• TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

11:21 AM Changeset in webkit [286504] by timothy_horton@apple.com

• 2 edits in trunk/Source/WTF

Enable Momentum Event Generator by default
​https://bugs.webkit.org/show_bug.cgi?id=233807
<rdar://problem/85262515>

Reviewed by Simon Fraser.

• Scripts/Preferences/WebPreferencesInternal.yaml:

Turn it on for modern WebKit. Adjust the description slightly.

11:06 AM Changeset in webkit [286503] by Lauro Moura

• 2 edits in trunk/LayoutTests

[GLIB] WPT css-transform transform-interpolation-rotate-slerp.html is flaky img failing after updated in r286002

Unreviewed test gardening.

• platform/glib/TestExpectations:

10:30 AM Changeset in webkit [286502] by keith_miller@apple.com

• 11 edits
  1 delete in trunk/Source/JavaScriptCore

Remove StructureIDBlob
​https://bugs.webkit.org/show_bug.cgi?id=233723

Reviewed by Yusuke Suzuki.

StructureIDBlob isn't very useful now that StructureIDs are just the
bottom bits of the pointer on 64 bit platforms. In a follow up patch
I'll change the layout of JSCell and Structure so that TypeInfo creation
can be a single load platforms that allow (and don't penalize) misaligned loads.

• CMakeLists.txt:
• JavaScriptCore.xcodeproj/project.pbxproj:
• ftl/FTLAbstractHeapRepository.h:
• ftl/FTLLowerDFGToB3.cpp:

(JSC::FTL::DFG::LowerDFGToB3::compileCompareStrictEq):

• jit/AssemblyHelpers.h:

(JSC::AssemblyHelpers::emitStoreStructureWithTypeInfo):

• jit/JITPropertyAccess.cpp:

(JSC::JIT::emit_op_put_to_scope):

• runtime/Structure.cpp:

(JSC::Structure::Structure):
(JSC::Structure::addNewPropertyTransition):
(JSC::Structure::removeNewPropertyTransition):
(JSC::Structure::attributeChangeTransition):
(JSC::Structure::nonPropertyTransitionSlow):
(JSC::Structure::setBrandTransition):

• runtime/Structure.h:

(JSC::Structure::id const):
(JSC::Structure::objectInitializationBlob const):
(JSC::Structure::idBlob const):
(JSC::Structure::isProxy const):
(JSC::Structure::typeInfo const):
(JSC::Structure::indexingType const):
(JSC::Structure::indexingMode const):
(JSC::Structure::fencedIndexingMode):
(JSC::Structure::indexingModeIncludingHistory const):
(JSC::Structure::indexingModeIncludingHistoryOffset):
(JSC::Structure::structureIDOffset): Deleted.

• runtime/StructureIDBlob.h: Removed.
• runtime/StructureInlines.h:

(JSC::Structure::hasIndexingHeader const):

• tools/VMInspectorInlines.h:

(JSC::VMInspector::verifyCellSize):

10:29 AM Changeset in webkit [286501] by Wenson Hsieh

• 2 edits in trunk/Source/WebKit

[ Monterey wk2 ] http/tests/media/video-webm-stall.html (layout-test) is a constant text failure
​https://bugs.webkit.org/show_bug.cgi?id=233717
rdar://85933254

Reviewed by Peng Liu.

After r274117, we call PlatformMediaSessionManager::setWebMFormatReaderEnabled in GPUProcess with
m_webMParserEnabled instead of the value of m_webMFormatReaderEnabled. Since we (arbitrarily) update WebM
parser enablement before WebM format reader enablement, this means that the first call to update preferences
leaves the WebM format reader disabled on PlatformMediaSessionManager even though it should be enabled. Prior to
the more recent change in r285477, this would be corrected on subsequent calls into
GPUProcessProxy::updatePreferences(), which previously sent GPUProcess::SetWebMFormatReaderEnabled()
messages even if the value of the preference hadn't changed.

r285477 removed those redundant IPC messages when the enablement value hasn't changed, so we can no longer
rely on PlatformMediaSessionManager being eventually updated to true due to m_webMParserEnabled being set to
true. Simply fix this by using m_webMFormatReaderEnabled instead of m_webMParserEnabled.

• GPUProcess/GPUProcess.cpp:

(WebKit::GPUProcess::setWebMFormatReaderEnabled):

10:15 AM Changeset in webkit [286500] by Alan Bujtas

• 4 edits in trunk/Source/WebCore

[LFC][IFC] Set the first/last box flag on the (bidi fragmented) inline box type of display boxes
​https://bugs.webkit.org/show_bug.cgi?id=233743

Reviewed by Antti Koivisto.

Keep track of the last constructed inline box type of display box position so that when we
happen to create another fragment for this same inline box, we can change the 'isLastBox' box to false.

• layout/formattingContexts/inline/InlineDisplayContentBuilder.cpp:

(WebCore::Layout::InlineDisplayContentBuilder::insertInlineBoxDisplayBoxForBidiBoundary):
(WebCore::Layout::InlineDisplayContentBuilder::processBidiContent):

• layout/formattingContexts/inline/InlineDisplayContentBuilder.h:

10:00 AM Changeset in webkit [286499] by Chris Dumez

• 5 edits in trunk/Source/WebCore

Introduce WorkerGlobalScope::type() function
​https://bugs.webkit.org/show_bug.cgi?id=233793

Reviewed by Alex Christensen.

Introduce WorkerGlobalScope::type() function to replace isDedicatedWorkerGlobalScope() / isServiceWorkerGlobalScope()
/ isSharedWorkerGlobalScope(). This is extracted from Alex's large patch at Bug 230382.

• workers/shared/SharedWorkerGlobalScope.h:

(isType):

9:45 AM Changeset in webkit [286498] by BJ Burg

• 3 edits in trunk/Source/WebKit

Web Inspector: Web Inspector^{2 crashes after closing if Web Inspector}1 closed first
​https://bugs.webkit.org/show_bug.cgi?id=233293
<rdar://problem/85526508>

Reviewed by Devin Rousso.

Cache the inspected page's identifier. During frontend teardown, use the cached indentifier
to remove the message receiver that was added to receive messages from the inspected page.

Other operations using m_inspectedPage should be guarded in case that the inspected
page already been closed and destroyed.

• UIProcess/Inspector/WebInspectorUIProxy.cpp:

(WebKit::WebInspectorUIProxy::createFrontendPage):
(WebKit::WebInspectorUIProxy::closeFrontendPageAndWindow):

• UIProcess/Inspector/WebInspectorUIProxy.h:

9:36 AM Changeset in webkit [286497] by BJ Burg

• 3 edits in trunk/Source/WebKit

[Cocoa] Web Inspector: fix completion handler type signature for _WKInspectorExtension methods
​https://bugs.webkit.org/show_bug.cgi?id=233792
<rdar://problem/85995314>

Reviewed by Devin Rousso.

• UIProcess/API/Cocoa/_WKInspectorExtension.h:
• UIProcess/API/Cocoa/_WKInspectorExtension.mm:

(-[_WKInspectorExtension evaluateScript:frameURL:contextSecurityOrigin:useContentScriptContext:completionHandler:]):
(-[_WKInspectorExtension evaluateScript:inTabWithIdentifier:completionHandler:]):
Replace the completionHandler's NSDictionary argument type with id.

A dictionary is not necessarily returned here because the implementation deserializes a SerializedScriptValue.
This could be a string, number, array, dictionary, or other serializable values.

While the frontend does return a dictionary payload, this is unpacked on the WebProcess-side
before being sent back out via the _WKInspectorExtension APi.

9:35 AM Changeset in webkit [286496] by Chris Dumez

• 6 edits in trunk/Source/WebCore

Improve dictionary-based DOMPromiseDeferred::settle
​https://bugs.webkit.org/show_bug.cgi?id=233751

Patch by Youenn Fablet <​youenn@apple.com> on 2021-12-03
Reviewed by Chris Dumez.

Make settle a templated function so that it can adapt to any ExceptionOr based type.
Update call sites that use settle for non ExceptionOr types as they should either use reject or resolve directly.
Covered by existing tests.

• Modules/applepay/ApplePaySetup.cpp:

(WebCore::ApplePaySetup::getSetupFeatures):
(WebCore::ApplePaySetup::begin):
(WebCore::ApplePaySetup::stop):

• Modules/filesystemaccess/FileSystemDirectoryHandle.cpp:

(WebCore::FileSystemDirectoryHandle::getFileHandle):
(WebCore::FileSystemDirectoryHandle::getDirectoryHandle):
(WebCore::FileSystemDirectoryHandle::resolve):

• Modules/filesystemaccess/FileSystemFileHandle.cpp:

(WebCore::FileSystemFileHandle::createSyncAccessHandle):

• bindings/js/JSDOMPromiseDeferred.h:

(WebCore::DOMPromiseDeferred::settle):
(WebCore::DOMPromiseDeferred<IDLDictionary<T>>::resolve): Deleted.
(WebCore::DOMPromiseDeferred<IDLDictionary<T>>::settle): Deleted.
(WebCore::DOMPromiseDeferred<IDLInterface<T>>::resolve): Deleted.
(WebCore::DOMPromiseDeferred<IDLInterface<T>>::settle): Deleted.

• workers/service/ServiceWorkerContainer.cpp:

(WebCore::ServiceWorkerContainer::getNavigationPreloadState):

9:19 AM Changeset in webkit [286495] by Antti Koivisto

• 2 edits in trunk/Source/WTF

[:has() pseudo-class] Enable by default
​https://bugs.webkit.org/show_bug.cgi?id=233809

Reviewed by Simon Fraser.

Flip the switch.

• Scripts/Preferences/WebPreferencesExperimental.yaml:

9:10 AM Changeset in webkit [286494] by Antti Koivisto

• 3 edits in trunk/Source/WebCore

[:has() pseudo-class] Improve result caching
​https://bugs.webkit.org/show_bug.cgi?id=233806

Reviewed by Simon Fraser.

Improve caching to to avoid O(n^{2) cases during invalidation and style resolution.}

• css/SelectorChecker.cpp:

(WebCore::SelectorChecker::matchHasPseudoClass const):

Cache matches as well as failures.
Cache failures that don't apply to subtrees.
Check cached matches and failures for all :has() match element types.

• style/SelectorMatchingState.h:

9:08 AM Changeset in webkit [286493] by fpizlo@apple.com

• 76 edits
  7 adds in trunk/Source/bmalloc

[libpas] Update to 96f9f4c28dc119695311c7c6bd81ed1f3f4e260c (allow more specialization of partial versus exclusive allocation)
​https://bugs.webkit.org/show_bug.cgi?id=233782

Reviewed by Yusuke Suzuki.

Libpas's main optimization for lots of isoheaps is partial views of shared pages: a single page may get
chopped up so that different heaps get different ranges of the page. This works out, in a not-too-gross
way, by reusing the segregated heap page management logic.

But previously, pages used exclusively versus pages that were shared looked largely the same; the only
way to tell them apart is that they had a different owner. Notably, those pages would get allocated out
of the same megapages, so a megapage lookup could not tell you if you were exclusive or shared.

This change introduces the concept of pas_segregated_page_role, which is this distinction between
exclusive and shared, and spreads it across all page management logic. So:

-> Page allocation knows the page's role, so it can use different megapages for exclusive or shared.

This enables exclusive and shared to have different deallocation fast paths.

-> Page_kind is now different depending on whether the page is exclusive or shared.

-> Many of the entrypoints into segregated_page that are required to know page_config are now also

required to know page_role. This makes sense, because the combination of megapages and page_kinds
knowing about roles means that anyone who ends up calling segregated_page API will now typically
know the role.

-> It's still possible to dynamically deduce the role using the page's owner, but hardly any code does

that.

This is leveraged in this patch to introduce some tiny space-saving optimizations to isoheaps:

-> Partial deallocation no longer uses the deallocation log.

-> Partial deallocation no longer uses the empty_word_eligibility_optimization.

This change is a stepping stone for introducing more different kinds of roles in the future.

• CMakeLists.txt:
• bmalloc.xcodeproj/project.pbxproj:
• libpas/libpas.xcodeproj/project.pbxproj:
• libpas/src/libpas/bmalloc_heap_config.h:
• libpas/src/libpas/hotbit_heap_config.h:
• libpas/src/libpas/iso_heap_config.h:
• libpas/src/libpas/iso_test_heap_config.h:
• libpas/src/libpas/jit_heap_config.c:

(jit_heap_config_dump_shared_page_directory_arg):

• libpas/src/libpas/jit_heap_config.h:
• libpas/src/libpas/minalign32_heap_config.h:
• libpas/src/libpas/pagesize64k_heap_config.h:
• libpas/src/libpas/pas_basic_heap_page_caches.h:
• libpas/src/libpas/pas_bitfit_allocator.c:

(pas_bitfit_allocator_commit_view):
(pas_bitfit_allocator_finish_failing):

• libpas/src/libpas/pas_bitfit_allocator.h:
• libpas/src/libpas/pas_bitfit_allocator_inlines.h:

(pas_bitfit_allocator_try_allocate):
(pas_bitfit_allocator_reset): Deleted.
(pas_bitfit_allocator_assert_reset): Deleted.

• libpas/src/libpas/pas_bitfit_heap.c:

(pas_bitfit_heap_construct_and_insert_size_class):
(pas_bitfit_heap_ensure_size_class): Deleted.

• libpas/src/libpas/pas_bitfit_heap.h:
• libpas/src/libpas/pas_bitfit_page.c:

(pas_bitfit_page_construct):

• libpas/src/libpas/pas_bitfit_page_config.h:
• libpas/src/libpas/pas_bitfit_page_config_inlines.h:
• libpas/src/libpas/pas_bitfit_page_config_kind.c:

(pas_bitfit_page_config_kind_get_string):

• libpas/src/libpas/pas_bitfit_page_config_kind.h:

(pas_bitfit_page_config_kind_get_string): Deleted.

• libpas/src/libpas/pas_bitfit_page_inlines.h:

(pas_bitfit_page_allocate):

• libpas/src/libpas/pas_bitfit_size_class.c:

(pas_bitfit_size_class_create): Deleted.

• libpas/src/libpas/pas_bitfit_size_class.h:
• libpas/src/libpas/pas_create_basic_heap_page_caches_with_reserved_memory.c:

(pas_create_basic_heap_page_caches_with_reserved_memory):

• libpas/src/libpas/pas_deallocate.c:

(deallocate_segregated):
(pas_try_deallocate_slow_no_cache):

• libpas/src/libpas/pas_deallocate.h:

(pas_deallocate_known_segregated):
(pas_try_deallocate_not_small_exclusive_segregated):
(pas_try_deallocate_impl):
(pas_deallocate_known_segregated_impl): Deleted.
(pas_try_deallocate_not_small): Deleted.

• libpas/src/libpas/pas_enumerate_segregated_heaps.c:

(pas_enumerate_segregated_heaps):

• libpas/src/libpas/pas_fast_megapage_kind.h:

(pas_fast_megapage_kind_get_string):

• libpas/src/libpas/pas_fast_megapage_table.c:

(pas_fast_megapage_table_set_by_index):

• libpas/src/libpas/pas_fast_megapage_table.h:

(pas_fast_megapage_table_get_by_index):

• libpas/src/libpas/pas_get_allocation_size.h:

(pas_get_allocation_size):

• libpas/src/libpas/pas_get_heap.h:

(pas_get_heap):
(pas_get_heap_known_segregated): Deleted.

• libpas/src/libpas/pas_get_object_kind.h:

(pas_get_object_kind):

• libpas/src/libpas/pas_get_page_base.h:

(pas_get_page_base):

• libpas/src/libpas/pas_get_page_base_and_kind_for_small_other_in_fast_megapage.h: Added.

(pas_get_page_base_and_kind_for_small_other_in_fast_megapage):

• libpas/src/libpas/pas_has_object.h:

(pas_has_object):

• libpas/src/libpas/pas_heap_config.h:
• libpas/src/libpas/pas_heap_config_inlines.h:
• libpas/src/libpas/pas_heap_config_kind.c:

(pas_heap_config_kind_get_string):

• libpas/src/libpas/pas_heap_config_kind.h:

(pas_heap_config_kind_get_string): Deleted.

• libpas/src/libpas/pas_heap_config_utils.h:
• libpas/src/libpas/pas_heap_config_utils_inlines.h:
• libpas/src/libpas/pas_internal_config.h:
• libpas/src/libpas/pas_local_allocator.c:

(pas_local_allocator_construct):
(pas_local_allocator_refill_with_bitfit): Deleted.
(pas_local_allocator_finish_refill_with_bitfit): Deleted.

• libpas/src/libpas/pas_local_allocator_inlines.h:

(pas_local_allocator_scan_bits_to_set_up_free_bits):
(pas_local_allocator_set_up_free_bits):
(pas_local_allocator_set_up_primordial_bump):
(pas_local_allocator_start_allocating_in_primordial_partial_view):
(pas_local_allocator_bless_primordial_partial_view_before_stopping):
(pas_local_allocator_try_allocate_in_primordial_partial_view):
(pas_local_allocator_refill_with_known_config):
(pas_local_allocator_return_memory_to_page_set_bit_callback):
(pas_local_allocator_return_memory_to_page_for_role):
(pas_local_allocator_return_memory_to_page):
(pas_local_allocator_try_allocate_small_segregated_slow_impl):
(pas_local_allocator_try_allocate_out_of_line_cases):
(pas_local_allocator_try_allocate_slow_impl):

• libpas/src/libpas/pas_object_kind.h:

(pas_object_kind_for_page_kind):

• libpas/src/libpas/pas_page_base_and_kind.h: Added.

(pas_page_base_and_kind_create):
(pas_page_base_and_kind_create_empty):

• libpas/src/libpas/pas_page_base_config.c: Added.

(pas_page_base_config_get_kind_string):

• libpas/src/libpas/pas_page_base_config.h:

(pas_page_base_config_is_segregated):
(pas_page_base_config_is_bitfit):
(pas_page_base_config_get_config_kind): Deleted.

• libpas/src/libpas/pas_page_kind.h:

(pas_page_kind_get_string):
(pas_page_kind_get_config_kind):
(pas_page_kind_get_segregated_variant):
(pas_page_kind_get_segregated_role):
(pas_page_kind_for_segregated_variant_and_role):
(pas_page_kind_for_bitfit_variant):

• libpas/src/libpas/pas_segregated_deallocation_logging_mode.h: Added.

(pas_segregated_deallocation_logging_mode_does_logging):
(pas_segregated_deallocation_logging_mode_is_size_aware):
(pas_segregated_deallocation_logging_mode_is_checked):

• libpas/src/libpas/pas_segregated_heap.c:

(max_object_size_for_page_config):
(pas_segregated_heap_ensure_size_directory_for_size):

• libpas/src/libpas/pas_segregated_page.c:

(pas_segregated_page_construct):
(pas_segregated_page_and_config_for_address_and_heap_config):

• libpas/src/libpas/pas_segregated_page_config.h:

(pas_segregated_page_config_enable_empty_word_eligibility_optimization_for_role):
(pas_segregated_page_config_logging_mode_for_role):

• libpas/src/libpas/pas_segregated_page_config_inlines.h:
• libpas/src/libpas/pas_segregated_page_config_kind.c:

(pas_segregated_page_config_kind_get_string):

• libpas/src/libpas/pas_segregated_page_config_kind.def:
• libpas/src/libpas/pas_segregated_page_config_kind.h:

(pas_segregated_page_config_kind_get_string): Deleted.

• libpas/src/libpas/pas_segregated_page_config_kind_and_role.c: Added.

(pas_segregated_page_config_kind_and_role_get_string):

• libpas/src/libpas/pas_segregated_page_config_kind_and_role.h: Added.

(pas_segregated_page_config_kind_and_role_create):

• libpas/src/libpas/pas_segregated_page_config_utils.h:

(pas_segregated_page_config_verify_dealloc): Deleted.

• libpas/src/libpas/pas_segregated_page_inlines.h:

(pas_segregated_page_deallocate_with_page):
(pas_segregated_page_deallocate):
(pas_segregated_page_get_directory_for_address_in_page):
(pas_segregated_page_get_directory_for_address_and_page_config):
(pas_segregated_page_get_object_size_for_address_in_page):
(pas_segregated_page_get_object_size_for_address_and_page_config):
(pas_segregated_page_log_or_deallocate):

• libpas/src/libpas/pas_segregated_page_role.h: Added.

(pas_segregated_page_role_get_string):

• libpas/src/libpas/pas_segregated_shared_view.c:

(pas_segregated_shared_view_commit_page):

• libpas/src/libpas/pas_segregated_size_directory.c:

(pas_segregated_size_directory_create):

• libpas/src/libpas/pas_segregated_size_directory.h:

(pas_segregated_size_directory_is_bitfit):
(pas_segregated_size_directory_get_bitfit_size_class):

• libpas/src/libpas/pas_segregated_view.c:

(should_be_eligible):

• libpas/src/libpas/pas_segregated_view.h:

(pas_segregated_view_get_page_role_for_owner):
(pas_segregated_view_get_page_role_for_allocator):

• libpas/src/libpas/pas_segregated_view_allocator_inlines.h:

(pas_segregated_view_will_start_allocating):

• libpas/src/libpas/pas_segregated_view_kind.h:

(pas_segregated_view_kind_get_role_for_owner):
(pas_segregated_view_kind_get_role_for_allocator):

• libpas/src/libpas/pas_thread_local_cache.c:

(process_deallocation_log_with_config):
(flush_deallocation_log):
(pas_thread_local_cache_append_deallocation_slow):

• libpas/src/libpas/pas_thread_local_cache.h:

(pas_thread_local_cache_encode_object):
(pas_thread_local_cache_append_deallocation):
(pas_thread_local_cache_append_deallocation_with_size):

• libpas/src/libpas/pas_try_reallocate.h:

(pas_try_reallocate_table_segregated_case):
(pas_try_reallocate):

• libpas/src/libpas/pas_try_shrink.h:

(pas_try_shrink):

• libpas/src/libpas/pas_utility_heap.c:

(pas_utility_heap_deallocate):

• libpas/src/libpas/pas_utility_heap_config.c:

(pas_utility_heap_allocate_page):
(pas_utility_heap_shared_page_directory_selector):
(pas_utility_heap_config_for_each_shared_page_directory):
(pas_utility_heap_config_dump_shared_page_directory_arg):

• libpas/src/libpas/pas_utility_heap_config.h:

(pas_utility_heap_shared_page_directory_selector): Deleted.

• libpas/src/libpas/pas_utils.h:

(pas_assertion_failed_noreturn_silencer):

• libpas/src/libpas/thingy_heap_config.h:
• libpas/src/test/ThingyAndUtilityHeapAllocationTests.cpp:

(std::addMargeBitfitTests):

8:50 AM Changeset in webkit [286492] by achristensen@apple.com

• 15 edits in trunk

Add room for more bytecode in WKContentRuleList file format
​https://bugs.webkit.org/show_bug.cgi?id=233780

Reviewed by Tim Hatcher.

Source/WebCore:

For rdar://72203352 we will need more bytecode. This adds room for it without putting anything there yet.
As long as we are updating the version number and forcing a recompile, I'm making the string serialization
more like the other action serializations to be able to simplify the interpreter later. I'm also shifting
around a few bits in the bytecode to be more organized and reasonable.

• contentextensions/CompiledContentExtension.h:
• contentextensions/ContentExtensionStringSerialization.cpp:

(WebCore::ContentExtensions::deserializeString):
(WebCore::ContentExtensions::serializeString):
(WebCore::ContentExtensions::stringSerializedLength):

• contentextensions/DFABytecode.h:

(WebCore::ContentExtensions::smallestPossibleJumpSize): Deleted.

• contentextensions/DFABytecodeCompiler.cpp:

(WebCore::ContentExtensions::smallestPossibleJumpSize):

Source/WebKit:

• Shared/WebCompiledContentRuleList.cpp:

(WebKit::WebCompiledContentRuleList::frameURLFiltersBytecode const):

• Shared/WebCompiledContentRuleList.h:
• Shared/WebCompiledContentRuleListData.cpp:

(WebKit::WebCompiledContentRuleListData::encode const):
(WebKit::WebCompiledContentRuleListData::decode):

• Shared/WebCompiledContentRuleListData.h:

(WebKit::WebCompiledContentRuleListData::WebCompiledContentRuleListData):

• UIProcess/API/APIContentRuleListStore.cpp:

(API::ContentRuleListMetaData::fileSize const):
(API::encodeContentRuleListMetaData):
(API::decodeContentRuleListMetaData):
(API::compiledToFile):
(API::createExtension):
(API::getContentRuleListSourceFromMappedFile):

• UIProcess/API/APIContentRuleListStore.h:

Tools:

• TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:

8:35 AM Changeset in webkit [286491] by eric.carlson@apple.com

• 20 edits
  2 adds in trunk

DisplayCaptureSource doesn't initialize base class correctly
​https://bugs.webkit.org/show_bug.cgi?id=233722

Reviewed by Youenn Fablet.

Source/WebCore:

Test: fast/mediastream/get-display-media-settings.html

• platform/mediastream/RealtimeMediaSourceCenter.cpp:

(WebCore::RealtimeMediaSourceCenter::createMediaStream): Pass hashSalt through.
(WebCore::RealtimeMediaSourceCenter::getDisplayMediaDevices): Ditto.
(WebCore::RealtimeMediaSourceCenter::validateRequestConstraintsAfterEnumeration): Ditto.

• platform/mediastream/RealtimeMediaSourceCenter.h:

• platform/mediastream/RealtimeMediaSourceFactory.h:
• platform/mediastream/cocoa/DisplayCaptureSourceCocoa.cpp:

(WebCore::DisplayCaptureSourceCocoa::create): Ditto.
(WebCore::DisplayCaptureSourceCocoa::DisplayCaptureSourceCocoa): Ditto.

• platform/mediastream/cocoa/DisplayCaptureSourceCocoa.h:

• platform/mediastream/gstreamer/GStreamerCaptureDeviceManager.h:
• platform/mediastream/gstreamer/GStreamerDisplayCaptureDeviceManager.cpp:

(WebCore::GStreamerDisplayCaptureDeviceManager::createDisplayCaptureSource): Ditto.

• platform/mediastream/gstreamer/GStreamerVideoCaptureSource.cpp:
• platform/mediastream/mac/RealtimeMediaSourceCenterMac.cpp:
• platform/mock/MockRealtimeMediaSourceCenter.cpp:
• platform/mock/MockRealtimeMediaSourceCenter.cpp:
• platform/mock/MockRealtimeVideoSource.cpp:

(WebCore::MockRealtimeVideoSource::settings):

Source/WebKit:

• UIProcess/Cocoa/UserMediaCaptureManagerProxy.cpp:

(WebKit::UserMediaCaptureManagerProxy::createMediaSourceForCaptureDeviceWithConstraints):
Pass the hash salt to the factory.

• WebProcess/cocoa/RemoteRealtimeDisplaySource.cpp:

(WebKit::RemoteRealtimeDisplaySource::create): Remove 'name' parameter because
CaptureDevice already has the name.
(WebKit::RemoteRealtimeDisplaySource::RemoteRealtimeDisplaySource): Pass name to
base class from device.label().
(WebKit::RemoteRealtimeDisplaySource::createRemoteMediaSource): Don't call setName
with settings, it was already set in the constructor.

• WebProcess/cocoa/RemoteRealtimeDisplaySource.h:

• WebProcess/cocoa/UserMediaCaptureManager.cpp:

(WebKit::UserMediaCaptureManager::DisplayFactory::createDisplayCaptureSource): Pass
hash salt to RemoteRealtimeDisplaySource::create.

• WebProcess/cocoa/UserMediaCaptureManager.h:

LayoutTests:

• fast/mediastream/get-display-media-settings-expected.txt: Added.
• fast/mediastream/get-display-media-settings.html: Added.

7:30 AM Changeset in webkit [286490] by svillar@igalia.com

• 9 edits
  2 adds in trunk/LayoutTests

[css-flexbox] WPT import. December 2021 edition
​https://bugs.webkit.org/show_bug.cgi?id=233810

Reviewed by Rob Buis.

LayoutTests/imported/w3c:

• resources/resource-files.json:
• web-platform-tests/css/css-flexbox/negative-available-size-crash-expected.txt: Added.
• web-platform-tests/css/css-flexbox/negative-available-size-crash.html: Added.
• web-platform-tests/css/css-flexbox/scrollbars-auto-expected.html:
• web-platform-tests/css/css-flexbox/scrollbars-expected.html:
• web-platform-tests/css/css-flexbox/table-as-item-inflexible-in-column-2.html:
• web-platform-tests/css/css-flexbox/table-as-item-inflexible-in-row-2.html:
• web-platform-tests/css/css-flexbox/w3c-import.log:

LayoutTests:

• TestExpectations: Unskipped a test that works fine after the changes in the import.

6:12 AM Changeset in webkit [286489] by Alan Bujtas

• 3 edits in trunk/Source/WebCore

[LFC][IFC] Move the (bidi)display boxes horizontally by the inline box margin, border and padding start as needed
​https://bugs.webkit.org/show_bug.cgi?id=233742

Reviewed by Antti Koivisto.

Re-visit the newly constructed display boxes and move them horizontally using the inline box margin, border and padding start values.
Note that in this patch we don't yet handle the end side of the inline boxes.

• layout/formattingContexts/inline/InlineDisplayContentBuilder.cpp:

(WebCore::Layout::InlineDisplayContentBuilder::processBidiContent):

6:01 AM Changeset in webkit [286488] by youenn@apple.com

• 16 edits in trunk

Persist NavigationPreloadState in service worker registration database
​https://bugs.webkit.org/show_bug.cgi?id=233698
Source/WebCore:

Reviewed by Chris Dumez.

Whenever changing a service worker registration navigation state, trigger storing of the changes in the database.
To do so, we add this state to ServiceWorkerContextData.
When creating a SWServerRegistration, we set default values to false and "true"
as per ​https://w3c.github.io/ServiceWorker/#service-worker-registration-navigation-preload-header-value.

Covered by updated test.

• workers/service/ServiceWorkerContextData.cpp:
• workers/service/ServiceWorkerContextData.h:
• workers/service/server/RegistrationDatabase.cpp:
• workers/service/server/RegistrationStore.h:
• workers/service/server/SWServer.cpp:
• workers/service/server/SWServer.h:
• workers/service/server/SWServerJobQueue.cpp:
• workers/service/server/SWServerRegistration.cpp:
• workers/service/server/SWServerRegistration.h:
• workers/service/server/SWServerWorker.cpp:

Tools:

<rdar://problem/85963120>

Reviewed by Chris Dumez.

Increase counter from 7 to 8.

• TestWebKitAPI/Tests/WebKitCocoa/ServiceWorkerBasic.mm:

LayoutTests:

Reviewed by Chris Dumez.

• http/wpt/service-workers/service-worker-networkprocess-crash-expected.txt:
• http/wpt/service-workers/service-worker-networkprocess-crash.html:

1:31 AM Changeset in webkit [286487] by mmaxfield@apple.com

• 6 edits in trunk/Source/WebKit

Rename DerivedSources/WebKit2 to DerivedSources/WebKit
​https://bugs.webkit.org/show_bug.cgi?id=233803

Reviewed by Tim Horton.

Mechanical rename.

No new tests because there is no behavior change.

• Configurations/BaseTarget.xcconfig:
• Scripts/generate-derived-sources.sh:
• Scripts/generate-unified-sources.sh:
• UnifiedSources-output.xcfilelist:
• WebKit.xcodeproj/project.pbxproj:

12:50 AM Changeset in webkit [286486] by commit-queue@webkit.org

• 2 edits in trunk/LayoutTests

[GStreamer] test http/tests/security/webaudio-render-remote-audio-allowed-crossorigin.html fails
​https://bugs.webkit.org/show_bug.cgi?id=233737

Patch by Arcady Goldmints-Orlov <Arcady Goldmints-Orlov> on 2021-12-03
Reviewed by Philippe Normand.

Give some more time to the frequency analyzer to detect activity.

• http/tests/security/webaudio-render-remote-audio-allowed-crossorigin.html:

12:25 AM Changeset in webkit [286485] by aboya@igalia.com

• 3 edits
  2 adds in trunk

[MSE] Fix erase range to prevent accidental deletion in files with changing durations
​https://bugs.webkit.org/show_bug.cgi?id=233528

Reviewed by Xabier Rodriguez-Calvar.

Source/WebCore:

In didReceiveSample(), eraseBeginTime was being set to
highestPresentationTimestamp minus a tolerance. The tolerance is not
needed since highestPresentationTimestamp is loaded from exact frame
timestamps, and can cause accidental frame erasure in situations where
there are frames with frames smaller than the tolerance, which is the
case for certain MP4 files.

Test: media/media-source/media-source-append-tiny-durations.html

• platform/graphics/SourceBufferPrivate.cpp:

(WebCore::SourceBufferPrivate::didReceiveSample):

LayoutTests:

• media/media-source/media-source-append-tiny-durations-expected.txt: Added.
• media/media-source/media-source-append-tiny-durations.html: Added.

12:06 AM Changeset in webkit [286484] by commit-queue@webkit.org

• 29 edits
  2 copies in trunk/Source

Network process does not seem to initialize logging
​https://bugs.webkit.org/show_bug.cgi?id=233360
<rdar://problem/85765854>

Patch by Kimmo Kinnunen <​kkinnunen@apple.com> on 2021-12-03
Reviewed by Fujii Hironori.

Source/WebCore:

Export WebCore::logLevelString(), it's the canonical
method to obtain WebCore log configuration string.
It's already used in WebKit, but not on Cocoa platform
where the export matters.

• platform/LogInitialization.h:

Source/WebKit:

Move GPUProcess and WebProcess process creation arguments
related to logging strings to new AuxiliaryProcessCreationParameters.
Use this in all auxiliary processes GPUProcess, WebProcess,
AuthnProcess and NetworkProcess.

Use the canonical WTF::logLevelString(), WebCore::logLevelString(),
WebKit::logLevelString() to initialize Cocoa WebKit2 logging settings
for process creation arguments.

Fixes bugs where

Tools/Scripts/run-minibrowser --debug -WebKit2Logging IPCMessages

would not print IPCMessages logging for NetworkProcess.

• GPUProcess/GPUProcess.cpp:

(WebKit::GPUProcess::initializeGPUProcess):

• GPUProcess/GPUProcessCreationParameters.cpp:

(WebKit::GPUProcessCreationParameters::encode const):
(WebKit::GPUProcessCreationParameters::decode):

• GPUProcess/GPUProcessCreationParameters.h:
• NetworkProcess/NetworkProcess.cpp:

(WebKit::NetworkProcess::initializeNetworkProcess):

• NetworkProcess/NetworkProcessCreationParameters.cpp:

(WebKit::NetworkProcessCreationParameters::encode const):
(WebKit::NetworkProcessCreationParameters::decode):

• NetworkProcess/NetworkProcessCreationParameters.h:
• Shared/AuxiliaryProcess.cpp:

(WebKit::AuxiliaryProcess::applyProcessCreationParameters):

• Shared/AuxiliaryProcess.h:
• Shared/AuxiliaryProcessCreationParameters.cpp: Copied from Source/WebKit/WebAuthnProcess/WebAuthnProcessCreationParameters.cpp.

(WebKit::AuxiliaryProcessCreationParameters::encode const):
(WebKit::AuxiliaryProcessCreationParameters::decode):

• Shared/AuxiliaryProcessCreationParameters.h: Copied from Source/WebKit/WebAuthnProcess/WebAuthnProcessCreationParameters.h.
• Shared/WebProcessCreationParameters.cpp:

(WebKit::WebProcessCreationParameters::encode const):
(WebKit::WebProcessCreationParameters::decode):

• Shared/WebProcessCreationParameters.h:
• Sources.txt:
• UIProcess/AuxiliaryProcessProxy.cpp:

(WebKit::AuxiliaryProcessProxy::auxiliaryProcessParameters):

• UIProcess/AuxiliaryProcessProxy.h:
• UIProcess/Cocoa/GPUProcessProxyCocoa.mm:

(WebKit::GPUProcessProxy::platformInitializeGPUProcessParameters):

• UIProcess/Cocoa/WebProcessPoolCocoa.mm:

(WebKit::WebProcessPool::platformInitializeWebProcess):

• UIProcess/GPU/GPUProcessProxy.cpp:

(WebKit::GPUProcessProxy::GPUProcessProxy):
(WebKit::GPUProcessProxy::platformInitializeGPUProcessParameters):

• UIProcess/GPU/GPUProcessProxy.h:
• UIProcess/Network/NetworkProcessProxy.cpp:

(WebKit::NetworkProcessProxy::sendCreationParametersToNewProcess):

• UIProcess/WebAuthentication/WebAuthnProcessProxy.cpp:

(WebKit::WebAuthnProcessProxy::singleton):

• UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::initializeNewWebProcess):

• WebAuthnProcess/WebAuthnProcess.cpp:

(WebKit::WebAuthnProcess::initializeWebAuthnProcess):

• WebAuthnProcess/WebAuthnProcessCreationParameters.cpp:

(WebKit::WebAuthnProcessCreationParameters::encode const):
(WebKit::WebAuthnProcessCreationParameters::decode):

• WebAuthnProcess/WebAuthnProcessCreationParameters.h:
• WebKit.xcodeproj/project.pbxproj:
• WebProcess/cocoa/WebProcessCocoa.mm:

(WebKit::WebProcess::platformInitializeWebProcess):

• WebProcess/playstation/WebProcessPlayStation.cpp:

(WebKit::WebProcess::platformInitializeWebProcess):

Dec 2, 2021:

11:25 PM Changeset in webkit [286483] by timothy_horton@apple.com

• 6 edits in trunk/Source/WebKit

MomentumEventDispatcher curve sometimes doesn't match the system curve
​https://bugs.webkit.org/show_bug.cgi?id=233801
<rdar://problem/85870287>

Reviewed by Simon Fraser.

Three small changes to get us closer to the system curve:

1) Fetch the momentum event dispatch interval from the system, and use
it to scale the momentum start event's delta (which we use as our
initial velocity) into the "ideal" curve's frame rate (which is always
60fps regardless).

2) Back-date the animation start time to the fingers-down phase end event,
so that the momentum start phase has a delta. This seems to match what
the system does.

3) Switch to MonotonicTime for the animation time, since it doesn't need
to be in the same timebase as events (which are oddly in WallTime), and
certainly should be monotonic.

• Shared/ScrollingAccelerationCurve.cpp:

(WebKit::ScrollingAccelerationCurve::ScrollingAccelerationCurve):
(WebKit::ScrollingAccelerationCurve::interpolate):
(WebKit::ScrollingAccelerationCurve::encode const):
(WebKit::ScrollingAccelerationCurve::decode):
(WebKit::operator<<):

• Shared/ScrollingAccelerationCurve.h:

(WebKit::ScrollingAccelerationCurve::frameRate):
(WebKit::ScrollingAccelerationCurve::operator== const):

• Shared/mac/ScrollingAccelerationCurveMac.mm:

(WebKit::fromIOHIDCurve):
(WebKit::fromIOHIDCurveArrayWithAcceleration):
(WebKit::fromIOHIDDevice):
Fetch and propagate the dispatch frame rate. This isn't *really* a
ScrollingAccelerationCurve property, but neither is resolution;
this is just currently the only place we look up HID properties
and push them to the Web Content process; some re-architecture is
warranted here in the future.

• WebProcess/WebPage/MomentumEventDispatcher.cpp:

(WebKit::MomentumEventDispatcher::handleWheelEvent):
Keep track of the last fingers-down phase ended event timestamp.

(WebKit::MomentumEventDispatcher::didStartMomentumPhase):
Backdate the start time so that the first event has the appropriate delta.
Divide out the dispatch frame rate so that the initial velocity is
as if the curve were running at 60fps (since we *will* run it at 60fps and interpolate).

(WebKit::MomentumEventDispatcher::didEndMomentumPhase):
(WebKit::MomentumEventDispatcher::consumeDeltaForTime):
(WebKit::MomentumEventDispatcher::displayWasRefreshed):
Factor the consume-a-delta-from-the-ideal-curve code out from displayWasRefreshed
so we can use it in the start phase too.

(WebKit::MomentumEventDispatcher::offsetAtTime):
(WebKit::MomentumEventDispatcher::computeNextDelta):
Rename idealCurveFrameRate->idealCurveFrameInterval for accuracy.

• WebProcess/WebPage/MomentumEventDispatcher.h:

10:39 PM Changeset in webkit [286482] by Chris Dumez

• 13 edits in trunk

html/semantics/forms/constraints/form-validation-validity-valid.html WPT test is failing
​https://bugs.webkit.org/show_bug.cgi?id=233778

Reviewed by Darin Adler.

LayoutTests/imported/w3c:

Rebaseline WPT tests now that more checks are passing.

• web-platform-tests/html/semantics/forms/constraints/form-validation-validity-valid-expected.txt:

Source/WebCore:

html/semantics/forms/constraints/form-validation-validity-valid.html WPT test is failing in WebKit
but passing in Blink & Gecko. The issue was caused by incorrect willValidate() checks inside
HTMLInputElement::isValid() & HTMLSelectElement::valueMissing().

No new tests, rebaselined existing tests.

• html/HTMLInputElement.cpp:

(WebCore::HTMLInputElement::isValid const):

• html/HTMLSelectElement.cpp:

(WebCore::HTMLSelectElement::valueMissing const):

LayoutTests:

Rebaseline WPT tests now that more checks are passing.

• platform/mac-wk2/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-validity-valid-expected.txt:
• platform/mac-wk2/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-validity-valueMissing-expected.txt:

10:05 PM Changeset in webkit [286481] by Simon Fraser

• 6 edits in trunk/Source

A Safari tab can rarely get stuck in a state where rendering updates stop happening
​https://bugs.webkit.org/show_bug.cgi?id=233784
rdar://85445072

Reviewed by Chris Dumez.

Sometimes a Safari tab can get into a state where rendering updates cease to happen,
which manifests as partially broken scrolling, blank tiles revealed when scrolling,
and somewhat broken page updates. I was able to sometimes reproduce this by clicking
on links in eBay emails from Mail on a system with two displays.

From the one time I reproduce with logging, the output indicated that DisplayRefreshMonitor::displayLinkFired()
would early return because isPreviousFrameDone() was false. The only way for that to occur,
barring memory corruption, is if DisplayRefreshMonitorMac::dispatchDisplayDidRefresh() returned early,
which it does if the callback comes twice in a single event loop; this may explain the rarity.

So fix DisplayRefreshMonitorMac::dispatchDisplayDidRefresh() call setIsPreviousFrameDone(true)
so the next callback can make progress

Also add some locking annotations and fix one missing lock, and some release logging.

Source/WebCore:

• platform/graphics/DisplayRefreshMonitor.cpp:

(WebCore::DisplayRefreshMonitor::stop):
(WebCore::DisplayRefreshMonitor::firedAndReachedMaxUnscheduledFireCount):
(WebCore::DisplayRefreshMonitor::displayLinkFired):

• platform/graphics/DisplayRefreshMonitor.h:

(WebCore::DisplayRefreshMonitor::WTF_REQUIRES_LOCK):
(WebCore::DisplayRefreshMonitor::WTF_GUARDED_BY_LOCK):
(WebCore::DisplayRefreshMonitor::setMaxUnscheduledFireCount): Deleted.
(WebCore::DisplayRefreshMonitor::isScheduled const): Deleted.
(WebCore::DisplayRefreshMonitor::setIsScheduled): Deleted.
(WebCore::DisplayRefreshMonitor::isPreviousFrameDone const): Deleted.
(WebCore::DisplayRefreshMonitor::setIsPreviousFrameDone): Deleted.

Source/WebKit:

• WebProcess/WebPage/mac/DisplayRefreshMonitorMac.cpp:

(WebKit::DisplayRefreshMonitorMac::dispatchDisplayDidRefresh):

8:49 PM Changeset in webkit [286480] by commit-queue@webkit.org

• 2 edits
  2 adds in trunk/LayoutTests

[GLIB] Update test expectations and baselines. Unreviewed test gardening.
​https://bugs.webkit.org/show_bug.cgi?id=233799

Add a platform specific baseline for imported/w3c/web-platform-tests/service-workers/service-worker/navigation-preload/broken-chunked-encoding.https.html
as the tests pass on GTK and WPE.

Patch by Arcady Goldmints-Orlov <Arcady Goldmints-Orlov> on 2021-12-02

• platform/glib/TestExpectations:
• platform/glib/imported/w3c/web-platform-tests/service-workers/service-worker/navigation-preload/broken-chunked-encoding.https-expected.txt: Added.

8:37 PM Changeset in webkit [286479] by Chris Dumez

• 4 edits in trunk/Source/WebKit

Regression(r283179) Google Drive freezes after downloading a folder
​https://bugs.webkit.org/show_bug.cgi?id=233783
<rdar://85918531>

Reviewed by Darin Adler.

When process-swapping on a navigation response due to COOP, we create a new ProvisionalPageProxy
to trigger a provisional load in a new WebProcess. In the common case, the ProvisionalPageProxy
gets committed, we process-swap and everything works fine. However, if the client decides to
convert the navigation into a download (like in the Google Drive case), then the
ProvisionalPageProxy gets destroyed without being committed. The issue is that the committed
process still thinks at this point that it is in the middle of a navigation and its layer
tree is thus frozen with reason=PageTransition. This is what makes Google Drive look frozen
after the download. This is not an issue with PSON because the process-swap happens in
decidePolicyForNavigationAction() and we tell the previous process to ignore the navigation
when we process-swap.

To address the issue, we now tell the committed process to cancel its navigation if the
ProvisionalPageProxy ends up getting destroyed without being committed. This lets the
committed process know there is no point in waiting for this navigation to happen and allows
it to unfreeze its layer tree.

• UIProcess/ProvisionalPageProxy.cpp:

(WebKit::ProvisionalPageProxy::ProvisionalPageProxy):
(WebKit::ProvisionalPageProxy::~ProvisionalPageProxy):

• UIProcess/ProvisionalPageProxy.h:

(WebKit::ProvisionalPageProxy::isProcessSwappingOnNavigationResponse const):
(WebKit::ProvisionalPageProxy::shouldClosePreviousPageAfterCommit const): Deleted.

• UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::commitProvisionalPage):
(WebKit::WebPageProxy::continueNavigationInNewProcess):

8:24 PM Changeset in webkit [286478] by ysuzuki@apple.com

• 3 edits
  1 add in trunk

[JSC] shell's $.globalObjectFor is not safe for non object cells
​https://bugs.webkit.org/show_bug.cgi?id=233794

Reviewed by Saam Barati.

JSTests:

• stress/dollar-global-object.js: Added.

Source/JavaScriptCore:

Only Object cells can have Structures having JSGlobalObject.

• jsc.cpp:

(JSC_DEFINE_HOST_FUNCTION):

8:17 PM Changeset in webkit [286477] by Tyler Wilcock

• 6 edits in trunk

AX: In AccessibilityRenderObject::documentLinks, use existing image-map document links instead of always creating new ones
​https://bugs.webkit.org/show_bug.cgi?id=233767

Reviewed by Chris Fleizach.

The current implementation of AccessibilityRenderObject::documentLinks always
creates new image-map links rather than using the existing ones created via
AccessibilityRenderObject::addImageMapChildren. This is problematic for two reasons:

1. It is wasteful in terms of memory usage as we will endlessly accumulate AccessibilityImageMapLink objects in the cache each time AccessibilityRenderObject::documentLinks is called.

2. It breaks <area> document links in isolated tree mode, since the objects created this way aren't a child of any other object, and thus don't have any representation in the isolated tree. Concretely, this means AXIsolatedTree::nodeForID never returns anything for these objects.

In this patch, we try to find the existing image-map link children
in AccessibilityRenderObject::documentLinks before falling back to creating new ones.

Source/WebCore:

Fixes test accessibility/mac/document-links.html in isolated tree mode.

• accessibility/AccessibilityRenderObject.cpp:

(WebCore::AccessibilityRenderObject::documentLinks):
Try to find the existing image-map link children before falling back to creating new ones.

LayoutTests:

• accessibility/mac/document-links-expected.txt:

Update expectations based on new variable names.

• accessibility/mac/document-links.html:

Modernize test and improve variable naming.

• resources/accessibility-helper.js:

(debugEscaped): Added.

7:51 PM Changeset in webkit [286476] by Jean-Yves Avenard

• 2 edits in trunk/Tools

Compilation error on Tools/TestWebKitAPI/Tests/WebKitCocoa/WebPushDaemon.mm
​https://bugs.webkit.org/show_bug.cgi?id=233797
rdar://problem/86000569

Unreviewed API test build fix.

• TestWebKitAPI/Tests/WebKitCocoa/WebPushDaemon.mm:

7:01 PM Changeset in webkit [286475] by Alan Bujtas

• 3 edits in trunk/Source/WebCore

[LFC][IFC] Create display boxes for fragmented inline boxes on bidi boundaries
​https://bugs.webkit.org/show_bug.cgi?id=233736

Reviewed by Antti Koivisto.

In this patch we insert inline box type of display boxes in-between bidi display boxes as needed.
We don't yet compute visual geometry and/or offset the rest of the content with the injected margins, borders and padding.

• layout/formattingContexts/inline/InlineDisplayContentBuilder.cpp:

(WebCore::Layout::InlineDisplayContentBuilder::insertInlineBoxDisplayBoxForBidiBoundary):
(WebCore::Layout::InlineDisplayContentBuilder::processBidiContent):
(WebCore::Layout::InlineDisplayContentBuilder::appendInlineBoxDisplayBoxForBidiBoundary): Deleted.

• layout/formattingContexts/inline/InlineDisplayContentBuilder.h:

6:54 PM Changeset in webkit [286474] by Chris Dumez

• 3 edits in trunk/Source/WebCore

Introduce a MainThreadBridge class inside WebLockManager
​https://bugs.webkit.org/show_bug.cgi?id=233789

Reviewed by Darin Adler.

Extract all the logic to hop to and back from the main thread out of the WebLockManager and into a MainThreadBridge
internal class. This is similar to what was done for BroadcastChannel already. It provides better structuring and
slightly better performance by calling isolatedCopy() on the ClientOrigin only once instead of on every call.

No new tests, no web-facing behavior change.

• Modules/web-locks/WebLockManager.cpp:

(WebCore::WebLockManager::MainThreadBridge::create):
(WebCore::WebLockManager::MainThreadBridge::MainThreadBridge):
(WebCore::WebLockManager::MainThreadBridge::requestLock):
(WebCore::WebLockManager::MainThreadBridge::releaseLock):
(WebCore::WebLockManager::MainThreadBridge::abortLockRequest):
(WebCore::WebLockManager::MainThreadBridge::query):
(WebCore::WebLockManager::MainThreadBridge::clientIsGoingAway):
(WebCore::WebLockManager::MainThreadBridge::ensureOnMainThread):
(WebCore::WebLockManager::WebLockManager):
(WebCore::WebLockManager::request):
(WebCore::WebLockManager::didCompleteLockRequest):
(WebCore::WebLockManager::query):
(WebCore::WebLockManager::signalToAbortTheRequest):
(WebCore::WebLockManager::clientIsGoingAway):
(WebCore::WebLockManager::requestLockOnMainThread): Deleted.
(WebCore::WebLockManager::releaseLockOnMainThread): Deleted.
(WebCore::WebLockManager::abortLockRequestOnMainThread): Deleted.
(WebCore::WebLockManager::queryOnMainThread): Deleted.
(WebCore::WebLockManager::ensureOnMainThread): Deleted.

• Modules/web-locks/WebLockManager.h:

6:35 PM Changeset in webkit [286473] by Russell Epstein

• 1 copy in tags/Safari-612.3.6.1.7

Tag Safari-612.3.6.1.7.

6:14 PM Changeset in webkit [286472] by Robert Jenner

• 2 edits in trunk/LayoutTests

[ Monterey ] http/tests/workers/service/serviceworker-websocket.https.html (layout-test) is constant text failure
​https://bugs.webkit.org/show_bug.cgi?id=233665

Unreviewed test gardening.

• platform/mac-wk2/TestExpectations:

6:06 PM Changeset in webkit [286471] by sbarati@apple.com

• 2 edits in trunk/Source/JavaScriptCore

Lower structureHeapAddressSize on more memory limited ARM64 devices
​https://bugs.webkit.org/show_bug.cgi?id=233786
<rdar://85940431>

Reviewed by Yusuke Suzuki.

Some processes using JSC are failing the mmap for the 1GB region.
Let's lower the region size to 512MB on lower memory iOS devices.

• runtime/JSCConfig.h:

5:42 PM Changeset in webkit [286470] by commit-queue@webkit.org

• 2 edits in trunk/Source/WebKit

Make enum values consisten in ApplicationManifest::Icon::Purpose
and _WKApplicationManifestIconPurpose
​https://bugs.webkit.org/show_bug.cgi?id=233764
rdar://85976270

Patch by Rachel Ginsberg <​rginsberg@apple.com> on 2021-12-02
Reviewed by Brent Fulgham.

The ApplicationManifest tests already check for 1, 2, and 4 respectively.

• UIProcess/API/Cocoa/_WKApplicationManifest.h:

Set the _WKApplicationManifestIconPurpose enum values to
1, 2, 4 to match the enum values set in ApplicationManifest::Icon::Purpose.
Those values are used in ApplicationManifest because an
OptionSet is used.

5:14 PM Changeset in webkit [286469] by pvollan@apple.com

• 3 edits in trunk/Source/WebKit

[WP][iOS] Add telemetry to determine which system calls are only used on process launch
​https://bugs.webkit.org/show_bug.cgi?id=233649
<rdar://problem/85875930>

Reviewed by Brent Fulgham.

Add telemetry to determine which system calls are only used on WebContent process launch on iOS.

• Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb.in:
• WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::create):

5:03 PM Changeset in webkit [286468] by Jonathan Bedard

• 2 edits in trunk/Tools

[resultsdbpy] Python 3.10 compatibility
​https://bugs.webkit.org/show_bug.cgi?id=233771
<rdar://problem/85978234>

Reviewed by Dewei Zhu.

• Tools/Scripts/libraries/resultsdbpy/resultsdbpy/model/configuration_context.py:

Canonical link: ​https://commits.webkit.org/244808@main

4:56 PM Changeset in webkit [286467] by Jonathan Bedard

• 4 edits in trunk/Tools

[git-webkit] Handle ambiguous branch names in land
​https://bugs.webkit.org/show_bug.cgi?id=233763
<rdar://problem/85973434>

Reviewed by Dewei Zhu.

• Scripts/libraries/webkitscmpy/setup.py: Bump version.
• Scripts/libraries/webkitscmpy/webkitscmpy/init.py: Ditto.
• Scripts/libraries/webkitscmpy/webkitscmpy/local/git.py:

(Git.rebase): Be explicit about the name of the target branch when modifying timestamps.

Canonical link: ​https://commits.webkit.org/244807@main

4:37 PM Changeset in webkit [286466] by Said Abou-Hallawa

• 12 edits
  1 add in trunk/Source/WebCore

[GPU Process] Move the FilterEffect boundaries to a new class FilterEffectGeometry
​https://bugs.webkit.org/show_bug.cgi?id=233618
rdar://85881969

Reviewed by Cameron McCormack.

The plan is to have Filter own its geometry and the geometry of the
FilterEffects. So move the boundaries and their flags from FilterEffect
to a new class called FilterEffectGeometry.

SVGFilterBuilder will calculate the FilterEffectGeometry for each FilterEffect
and store it in map { key: FilterEffect, value: FilterEffectGeometry }.
Eventually this map is moved to SVGFilter which will be used in the
virtual function effectGeometry(). determineFilterPrimitiveSubregion()
will call this virtual function to calculate the primitive subregion.

• Headers.cmake:
• WebCore.xcodeproj/project.pbxproj:
• platform/graphics/filters/Filter.h:

(WebCore::Filter::effectGeometry const):

• platform/graphics/filters/FilterEffect.cpp:

(WebCore::FilterEffect::determineFilterPrimitiveSubregion):
(WebCore::FilterEffect::apply):

• platform/graphics/filters/FilterEffect.h:

(WebCore::FilterEffect::hasX const): Deleted.
(WebCore::FilterEffect::setHasX): Deleted.
(WebCore::FilterEffect::hasY const): Deleted.
(WebCore::FilterEffect::setHasY): Deleted.
(WebCore::FilterEffect::hasWidth const): Deleted.
(WebCore::FilterEffect::setHasWidth): Deleted.
(WebCore::FilterEffect::hasHeight const): Deleted.
(WebCore::FilterEffect::setHasHeight): Deleted.
(WebCore::FilterEffect::effectBoundaries const): Deleted.
(WebCore::FilterEffect::setEffectBoundaries): Deleted.

• platform/graphics/filters/FilterEffectGeometry.h: Added.

(WebCore::FilterEffectGeometry::FilterEffectGeometry):
(WebCore::FilterEffectGeometry::x const):
(WebCore::FilterEffectGeometry::y const):
(WebCore::FilterEffectGeometry::width const):
(WebCore::FilterEffectGeometry::height const):

• svg/SVGFilterPrimitiveStandardAttributes.cpp:

(WebCore::SVGFilterPrimitiveStandardAttributes::setStandardAttributes const): Deleted.

• svg/SVGFilterPrimitiveStandardAttributes.h:
• svg/graphics/filters/SVGFilter.cpp:

(WebCore::SVGFilter::create):
(WebCore::SVGFilter::effectGeometry const):

• svg/graphics/filters/SVGFilter.h:
• svg/graphics/filters/SVGFilterBuilder.cpp:

(WebCore::boundarySetFlagsForElement):
(WebCore::SVGFilterBuilder::buildFilterEffects):

• svg/graphics/filters/SVGFilterBuilder.h:

(WebCore::SVGFilterBuilder::takeEffectGeometryMap):

4:35 PM Changeset in webkit [286465] by pvollan@apple.com

• 6 edits in trunk/Source

[WP] Strengthen sandbox when AppCache is disabled
​https://bugs.webkit.org/show_bug.cgi?id=233746
<rdar://problem/85953893>

Reviewed by Brent Fulgham.

Source/WebKit:

When AppCache is disabled, we can remove access to some resources in the WebContent process' sandbox.
This is implemented using a new sandbox state variable representing the AppCache state.

• Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb.in:
• WebProcess/WebPage/WebPage.cpp:
• WebProcess/com.apple.WebProcess.sb.in:

Source/WTF:

Add HAVE macro for sandbox state flags support.

• wtf/PlatformHave.h:

4:32 PM Changeset in webkit [286464] by Chris Dumez

• 6 edits
  6 moves
  1 add in trunk/Source/WebCore

Move Shared Worker classes to WebCore/workers/shared
​https://bugs.webkit.org/show_bug.cgi?id=233769

Reviewed by Alex Christensen.

Move Shared Worker classes to WebCore/workers/shared, for better structuring and consistency
with WebCore/workers/service.

• CMakeLists.txt:
• DerivedSources.make:
• WebCore.xcodeproj/project.pbxproj:
• workers/shared/SharedWorker.cpp: Renamed from Source/WebCore/workers/SharedWorker.cpp.
• workers/shared/SharedWorker.h: Renamed from Source/WebCore/workers/SharedWorker.h.
• workers/shared/SharedWorker.idl: Renamed from Source/WebCore/workers/SharedWorker.idl.
• workers/shared/SharedWorkerGlobalScope.cpp: Renamed from Source/WebCore/workers/SharedWorkerGlobalScope.cpp.
• workers/shared/SharedWorkerGlobalScope.h: Renamed from Source/WebCore/workers/SharedWorkerGlobalScope.h.
• workers/shared/SharedWorkerGlobalScope.idl: Renamed from Source/WebCore/workers/SharedWorkerGlobalScope.idl.

4:14 PM Changeset in webkit [286463] by Alan Coon

• 1 edit in branches/safari-612-branch/Source/WebCore/page/csp/ContentSecurityPolicy.cpp

Unreviewed build fix. rdar://problem/85943167

candidate function not viable: no known conversion from 'const WTF::String' to 'const WTF::URL' for 3rd argument

4:12 PM Changeset in webkit [286462] by sbarati@apple.com

• 3 edits
  1 add in trunk

Fix OOM crash in JSValue::toWTFStringForConsole
​https://bugs.webkit.org/show_bug.cgi?id=233775
<rdar://85259423>

Reviewed by Mark Lam.

JSTests:

• stress/pretty-print-oom.js: Added.

Source/JavaScriptCore:

• runtime/JSCJSValue.cpp:

(JSC::JSValue::toWTFStringForConsole const):

3:49 PM Changeset in webkit [286461] by Megan Gardner

• 33 edits
  10 adds in trunk

Re-add support of image control menus.
​https://bugs.webkit.org/show_bug.cgi?id=233305

Reviewed by Tim Horton.

Source/WebCore:

Test: fast/images/image-controls-basic.html

Image control menus are needed for mail transition to modern WebKit, so re-add them
after the removal in ​https://trac.webkit.org/changeset/272117/webkit

• DerivedSources-input.xcfilelist:
• DerivedSources.make:
• Sources.txt:
• WebCore.xcodeproj/project.pbxproj:
• css/CSSPrimitiveValueMappings.h:

(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):

• css/CSSProperties.json:
• css/CSSValueKeywords.in:
• css/parser/CSSParserIdioms.cpp:

(WebCore::isValueAllowedInMode):

• dom/mac/ImageControlsMac.cpp: Added.

(WebCore::ImageControlsMac::imageControlsElementIdentifier):
(WebCore::ImageControlsMac::imageControlsButtonIdentifier):
(WebCore::ImageControlsMac::hasControls):
(WebCore::ImageControlsMac::imageControlHost):
(WebCore::ImageControlsMac::isImageControlsButtonElement):
(WebCore::ImageControlsMac::containerRect):
(WebCore::ImageControlsMac::createImageControls):

• dom/mac/ImageControlsMac.h: Added.
• html/HTMLAttributeNames.in:
• html/HTMLImageElement.cpp:

(WebCore::HTMLImageElement::HTMLImageElement):
(WebCore::HTMLImageElement::parseAttribute):
(WebCore::HTMLImageElement::didAttachRenderers):
(WebCore::HTMLImageElement::updateImageControls):
(WebCore::HTMLImageElement::tryCreateImageControls):
(WebCore::HTMLImageElement::destroyImageControls):
(WebCore::HTMLImageElement::hasImageControls const):
(WebCore::HTMLImageElement::childShouldCreateRenderer const):

• html/HTMLImageElement.h:

(WebCore::HTMLImageElement::hasShadowControls const):

• html/shadow/mac/ImageControlsMac.css: Added.

(#if defined(ENABLE_SERVICE_CONTROLS) && ENABLE_SERVICE_CONTROLS):
(button#image-controls-button):
(div#image-controls:hover button#image-controls-button):

• html/shadow/mac/ImageControlsRootElementMac.cpp: Added.

(WebCore::xWebkitImageControlsElementIdentifier):
(WebCore::RenderImageControls::RenderImageControls):
(WebCore::RenderImageControls::updateLogicalWidth):
(WebCore::RenderImageControls::computeLogicalHeight const):
(WebCore::hasControl):
(WebCore::imageControlHost):
(WebCore::isImageControls):
(WebCore::ImageControlsRootElement::tryCreate):
(WebCore::ImageControlsRootElementMac::ImageControlsRootElementMac):
(WebCore::ImageControlsRootElementMac::createElementRenderer):

• html/shadow/mac/ImageControlsRootElementMac.h: Added.
• page/ContextMenuController.cpp:

(WebCore::imageFromImageElementNode):
(WebCore::ContextMenuController::maybeCreateContextMenu):
(WebCore::ContextMenuController::showImageControlsMenu):

• page/ContextMenuController.h:
• page/EventHandler.cpp:

(WebCore::EventHandler::textRecognitionCandidateElement const):

• platform/ThemeTypes.cpp:

(WebCore::operator<<):

• platform/ThemeTypes.h:
• rendering/RenderImage.cpp:

(WebCore::RenderImage::RenderImage):

• rendering/RenderTheme.cpp:

(WebCore::RenderTheme::adjustStyle):
(WebCore::RenderTheme::paint):
(WebCore::RenderTheme::paintBorderOnly):
(WebCore::RenderTheme::paintDecorations):
(WebCore::RenderTheme::adjustImageControlsButtonStyle const):

• rendering/RenderTheme.h:

(WebCore::RenderTheme::imageControlsButtonSize const):
(WebCore::RenderTheme::imageControlsButtonPositionOffset const):
(WebCore::RenderTheme::paintImageControlsButton):

• rendering/RenderThemeMac.h:
• rendering/RenderThemeMac.mm:

(WebCore::RenderThemeMac::documentViewFor const):
(WebCore::RenderThemeMac::adjustImageControlsButtonStyle const):
(WebCore::RenderThemeMac::servicesRolloverButtonCell const):
(WebCore::RenderThemeMac::paintImageControlsButton):
(WebCore::RenderThemeMac::imageControlsButtonSize const):
(WebCore::RenderThemeMac::imageControlsButtonPositionOffset const):

• style/UserAgentStyle.cpp:
• style/UserAgentStyle.h:

Source/WebCore/PAL:

Move SPI to SPI specific file.

• PAL.xcodeproj/project.pbxproj:
• pal/spi/mac/NSServicesRolloverButtonCellSPI.h: Added.

Source/WebKit:

Image control menus are needed for mail transition to modern WebKit, so re-add them
after the removal in ​https://trac.webkit.org/changeset/272117/webkit

• Platform/spi/mac/AppKitSPI.h:
• UIProcess/API/C/WKPreferences.cpp:

(WKPreferencesSetImageControlsEnabled):
(WKPreferencesGetImageControlsEnabled):

• UIProcess/API/Cocoa/WKWebView.mm:

(-[WKWebView _setupPageConfiguration:]):

• UIProcess/API/Cocoa/WKWebViewConfiguration.mm:

(-[WKWebViewConfiguration init]):
(-[WKWebViewConfiguration copyWithZone:]):
(-[WKWebViewConfiguration _imageControlsEnabled]):
(-[WKWebViewConfiguration _setImageControlsEnabled:]):

• UIProcess/API/Cocoa/WKWebViewConfigurationPrivate.h:

Source/WTF:

• Scripts/Preferences/WebPreferences.yaml:

LayoutTests:

Update tests to account for image controls.

• fast/images/image-controls-basic.html: Added.
• fast/images/image-controls-basic-expected.txt: Added.
• platform/mac/fast/images/image-controls-basic-expected.txt: Added.

3:48 PM Changeset in webkit [286460] by pvollan@apple.com

• 7 edits in trunk/Source/WebKit

[iOS] Add required system call to WebKit sandboxes
​https://bugs.webkit.org/show_bug.cgi?id=233730
<rdar://85266065>

Reviewed by Brent Fulgham.

• GPUProcess/mac/com.apple.WebKit.GPUProcess.sb.in:
• NetworkProcess/mac/com.apple.WebKit.NetworkProcess.sb.in:
• Resources/SandboxProfiles/ios/com.apple.WebKit.GPU.sb:
• Resources/SandboxProfiles/ios/com.apple.WebKit.Networking.sb:
• Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb.in:
• WebProcess/com.apple.WebProcess.sb.in:

3:37 PM Changeset in webkit [286459] by mmaxfield@apple.com

• 32 edits in trunk/Source

[WebGPU] Hook up navigator.gpu.requestAdapter()
​https://bugs.webkit.org/show_bug.cgi?id=233619

Source/WebCore:

Reviewed by Dean Jackson.

Hook up Navigator.GPU to WebKitLegacy/WebKit (via ChromeClient) so WebKit can provide a
RemoteGPUProxy "backing" object.

I have a test for this! But unfortunately I can't commit it yet because HAVE(WEBGPU_IMPLEMENTATION)
has to be off until Apple configures its internal build system. As soon as that's done, I'll add
the test for this!!! (The test passes when run locally.)

• Configurations/WebCore.xcconfig: Weak link with WebGPU.framework. This will

succeed even if WebGPU.framework doesn't exist.

• Modules/WebGPU/GPU.cpp:

(WebCore::GPU::setBacking): This is for WorkerNavigator. To create a GPU object, we have to
go through ChromeClient, which means we have to use the main thread. This means that creating
the GPU object has to be asynchronous on workers. We handle this by having
WorkerNavigator::gpu() return a shell object with no "backing," and asynchronously kick off a
task to construct a backing object and connect it to the shell object. If content calls
requestAdapter() on the shell, that function is already async, so we just fold in the wait for
the backing into the already-existing wait we have to do as a matter of course. This implements
that.
(WebCore::GPU::requestAdapter):

• Modules/WebGPU/GPU.h:

(WebCore::GPU::create):
(WebCore::GPU::backing): Deleted.
(WebCore::GPU::backing const): Deleted.
(WebCore::GPU::GPU): Deleted.

• Modules/WebGPU/GPUDevice.cpp:

(WebCore::GPUDevice::popErrorScope):

• page/Chrome.cpp:

(WebCore::Chrome::createGPUForWebGPU const): Ask WebKit / WebKitLegacy what kind of backing we
should be using.

• page/Chrome.h:
• page/ChromeClient.h:

(WebCore::ChromeClient::createGPUForWebGPU const):

• page/Navigator.cpp:

(WebCore::Navigator::gpu):

• page/Navigator.h:
• page/NavigatorBase.h:

(WebCore::NavigatorBase::gpu): Deleted.

• page/WorkerNavigator.cpp:

(WebCore::WorkerNavigator::gpu):

• page/WorkerNavigator.h:

Source/WebGPU:

Just hook up enough for Instance::requestAdapter() to not fail.

Reviewed by Dean Jackson.

• WebGPU/Adapter.mm:

(WebGPU::Adapter::getLimits):

• WebGPU/Instance.mm:

(WebGPU::Instance::requestAdapter):
(wgpuCreateInstance):
(wgpuInstanceRequestAdapter):

Source/WebKit:

Reviewed by Dean Jackson.

This hooks up a new message, GPUConnectionToWebProcess::createRemoteGPU(), which starts
the whole WebGPU GPU process infrastructure.

• Configurations/BaseTarget.xcconfig: We have to be able to #include files from WebGPU.
• GPUProcess/GPUConnectionToWebProcess.cpp:

(WebKit::GPUConnectionToWebProcess::createRemoteGPU): GPUConnectionToWebProcess owns
the RemoteGPUs.

• GPUProcess/GPUConnectionToWebProcess.h:
• GPUProcess/GPUConnectionToWebProcess.messages.in:
• Sources.txt:
• WebProcess/GPU/graphics/WebGPU/RemoteGPUProxy.cpp:

(WebKit::RemoteGPUProxy::RemoteGPUProxy): Creating a RemoteGPUProxy needs to send a
message to the GPU process to start the GPU process infrastructure. We block while we're
waiting for the reply from the GPU process. This was copied from RemoteGraphicsContextGL.
(WebKit::RemoteGPUProxy::gpuProcessConnectionDidClose):

• WebProcess/GPU/graphics/WebGPU/RemoteGPUProxy.h:
• WebProcess/WebCoreSupport/WebChromeClient.cpp: Hook up RemoteGPUProxy to WebCore.

(WebKit::WebChromeClient::createGPUForWebGPU const):

• WebProcess/WebCoreSupport/WebChromeClient.h:

Source/WebKitLegacy/mac:

Reviewed by Dean Jackson.

• Configurations/WebKitLegacy.xcconfig: We have to be able to #include files from WebGPU.
• WebCoreSupport/WebChromeClient.h:
• WebCoreSupport/WebChromeClient.mm:

(WebChromeClient::createGPUForWebGPU const): Hook up GPUImpl to WebCore.

Source/WTF:

Reviewed by Dean Jackson.

• wtf/PlatformHave.h: Create a HAVE() macro, by default off. As we bring

more infrastructure online, this will eventually flip to on.

3:26 PM Changeset in webkit [286458] by Devin Rousso

• 14 edits
  1 add in trunk

[css-values-4] Support *vi (inline) and *vb (block) viewport units
​https://bugs.webkit.org/show_bug.cgi?id=232895
<rdar://problem/85179134>

Reviewed by Simon Fraser.

Source/WebCore:

*vi is equal to 1% of the size of the UA-default/large/small/dynamic viewport size in the direction of the root element’s inline axis.
*vb is equal to 1% of the size of the initial containing block UA-default/small/large/dynamic viewport size in the direction of the root element’s block axis.
As an example, on about:blank:

• vi would be equal to vw
• vb would be equal to vh
• svi would be equal to svw
• svb would be equal to svh
• lvi would be equal to lvw
• lvb would be equal to lvh
• dvi would be equal to dvw
• dvb would be equal to dvh

Spec: <​https://drafts.csswg.org/css-values-4/#viewport-variants>

Tests: CSSViewportUnits.AllSame

CSSViewportUnits.EmptyUnobscuredSizeOverrides
CSSViewportUnits.SameUnobscuredSizeOverrides
CSSViewportUnits.DifferentUnobscuredSizeOverrides
CSSViewportUnits.SVGDocument

• css/CSSUnits.h:
• css/CSSUnits.cpp:

(WebCore::operator<<):

• css/CSSPrimitiveValue.h:

(WebCore::CSSPrimitiveValue::isViewportPercentageLength):

• css/CSSPrimitiveValue.cpp:

(WebCore::isValidCSSUnitTypeForDoubleConversion):
(WebCore::isStringType):
(WebCore::CSSPrimitiveValue::cleanup):
(WebCore::lengthOfViewportPhysicalAxisForLogicalAxis): Added.
(WebCore::CSSPrimitiveValue::computeUnzoomedNonCalcLengthDouble):
(WebCore::CSSPrimitiveValue::computeNonCalcLengthDouble):
(WebCore::CSSPrimitiveValue::unitTypeString):
(WebCore::CSSPrimitiveValue::formatNumberForCustomCSSText const):
(WebCore::CSSPrimitiveValue::equals const):

• css/parser/CSSParserToken.cpp:

(WebCore::cssPrimitiveValueUnitFromTrie):

• css/parser/CSSPropertyParserHelpers.cpp:

(WebCore::CSSPropertyParserHelpers::consumeLengthRawWithKnownTokenTypeDimension):

• css/calc/CSSCalcCategoryMapping.cpp:

(WebCore::calcUnitCategory):
(WebCore::calculationCategoryForCombination):
(WebCore::hasDoubleValue):

Source/WebInspectorUI:

• UserInterface/Models/CSSCompletions.js:

Tools:

• TestWebKitAPI/Tests/WebKitCocoa/CSSViewportUnits.html:
• TestWebKitAPI/Tests/WebKitCocoa/CSSViewportUnits.svg: Added.
• TestWebKitAPI/Tests/WebKitCocoa/CSSViewportUnits.mm:

(TEST.CSSViewportUnits.AllSame):
(TEST.CSSViewportUnits.EmptyUnobscuredSizeOverrides):
(TEST.CSSViewportUnits.SameUnobscuredSizeOverrides):
(TEST.CSSViewportUnits.DifferentUnobscuredSizeOverrides):
(TEST.CSSViewportUnits.SVGDocument): Added.

• TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:

3:07 PM Changeset in webkit [286457] by Russell Epstein

• 8 edits in branches/safari-612-branch/Source

Unreviewed versioning fix.

rdar://85990909

3:05 PM Changeset in webkit [286456] by Chris Dumez

• 2 edits in trunk/Source/WebCore

Unreviewed follow-up to r286453 to address comments from Darin.

• dom/MessageChannel.h:

(WebCore::MessageChannel::port1 const):
(WebCore::MessageChannel::port2 const):

2:59 PM Changeset in webkit [286455] by Chris Dumez

• 38 edits
  2 copies
  5 adds in trunk

[WK2] Make Web Lock API work across multiple WebProcesses
​https://bugs.webkit.org/show_bug.cgi?id=233719

Reviewed by Alex Christensen.

Source/WebCore:

Make Web Lock API work across multiple WebProcesses. To achieve this, WebLockRegistry is now
an interface with 2 concrete subclasses:

• LocalWebLockRegistry: In charge of the actual storage of the lock registrations. This is basically a rename of the class that used to be named WebLockRegistry. In the WebKitLegacy case, we use a global LocalWebLockRegistry for regular browsing views and another global LocalWebLockRegistry for private browsing views. In the modern WebKit case, the LocalWebLockRegistry is held in the UIProcess by the WebsiteDataStore. It is held by the WebsiteDataStore because we don't want different sessions to share locks. It lives in the UIProcess because it needs to be shared by several WebProcesses. The NetworkProcess may have been suitable too but it would have meant dealing with crash recovery.
• RemoteWebLockRegistry: It has the same interface has LocalWebLockRegistry but merely proxies calls to the UIProcess via IPC.

In the UIProcess, each WebProcessProxy has a WebLockRegistryProxy instance which listens for
IPC from the RemoteWebLockRegistry in the WebProcess. Upon receiving IPC, it validates the
parameters (since the WebProcesses are not trusted) and calls the corresponding functions on
the LocalWebLockRegistry held by the WebProcessProxy's WebsiteDataStore. The WebLockRegistryProxy
is also in charge of notifying the RemoteWebLockRegistry in the WebProcess whenever a lock
request completes and whenever a held lock gets stolen.

• Modules/web-locks/WebLockManager.cpp:

(WebCore::WebLockManager::request):
(WebCore::WebLockManager::didCompleteLockRequest):
(WebCore::WebLockManager::requestLockOnMainThread):
(WebCore::WebLockManager::releaseLockOnMainThread):
(WebCore::WebLockManager::abortLockRequestOnMainThread):
(WebCore::WebLockManager::queryOnMainThread):
(WebCore::WebLockManager::clientIsGoingAway):
(WebCore::WebLockManager::ensureOnMainThread):

• Modules/web-locks/WebLockManager.h:
• Modules/web-locks/WebLockManagerSnapshot.h:

(WebCore::WebLockManagerSnapshot::Info::encode const):
(WebCore::WebLockManagerSnapshot::encode const):
(WebCore::WebLockManagerSnapshot::Info::decode):
(WebCore::WebLockManagerSnapshot::decode):

• Modules/web-locks/WebLockMode.h:
• Modules/web-locks/WebLockRegistry.cpp:

(WebCore::LocalWebLockRegistry::ensureRegistryForOrigin):
(WebCore::LocalWebLockRegistry::existingRegistryForOrigin const):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::create):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::PerOriginRegistry):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::~PerOriginRegistry):
(WebCore::LocalWebLockRegistry::requestLock):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::requestLock):
(WebCore::LocalWebLockRegistry::releaseLock):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::releaseLock):
(WebCore::LocalWebLockRegistry::abortLockRequest):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::abortLockRequest):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::isGrantable const):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::processLockRequestQueue):
(WebCore::LocalWebLockRegistry::snapshot):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::snapshot):
(WebCore::LocalWebLockRegistry::clientIsGoingAway):
(WebCore::LocalWebLockRegistry::PerOriginRegistry::clientsAreGoingAway):
(WebCore::LocalWebLockRegistry::clientsAreGoingAway):
(): Deleted.
(WebCore::WebLockRegistry::registryForOrigin): Deleted.
(WebCore::WebLockRegistry::WebLockRegistry): Deleted.
(WebCore::WebLockRegistry::~WebLockRegistry): Deleted.
(WebCore::WebLockRegistry::requestLock): Deleted.
(WebCore::WebLockRegistry::releaseLock): Deleted.
(WebCore::WebLockRegistry::abortLockRequest): Deleted.
(WebCore::WebLockRegistry::isGrantable const): Deleted.
(WebCore::WebLockRegistry::processLockRequestQueue): Deleted.
(WebCore::WebLockRegistry::snapshot): Deleted.
(WebCore::WebLockRegistry::clientIsGoingAway): Deleted.

• Modules/web-locks/WebLockRegistry.h:

(WebCore::WebLockRegistry::~WebLockRegistry):

• WebCore.xcodeproj/project.pbxproj:
• loader/EmptyClients.cpp:

(WebCore::pageConfigurationWithEmptyClients):

• page/Page.cpp:

(WebCore::Page::Page):

• page/Page.h:

(WebCore::Page::webLockRegistry):

• page/PageConfiguration.cpp:

(WebCore::PageConfiguration::PageConfiguration):

• page/PageConfiguration.h:

Source/WebKit:

See WebCore changelog.

• CMakeLists.txt:
• DerivedSources-input.xcfilelist:
• DerivedSources-output.xcfilelist:
• DerivedSources.make:
• Scripts/webkit/messages.py:

(types_that_cannot_be_forward_declared):

• Sources.txt:
• UIProcess/WebLockRegistryProxy.cpp: Added.

(WebKit::WebLockRegistryProxy::WebLockRegistryProxy):
(WebKit::WebLockRegistryProxy::~WebLockRegistryProxy):
(WebKit::WebLockRegistryProxy::requestLock):
(WebKit::WebLockRegistryProxy::releaseLock):
(WebKit::WebLockRegistryProxy::abortLockRequest):
(WebKit::WebLockRegistryProxy::snapshot):
(WebKit::WebLockRegistryProxy::clientIsGoingAway):
(WebKit::WebLockRegistryProxy::processDidExit):

• UIProcess/WebLockRegistryProxy.h: Added.
• UIProcess/WebLockRegistryProxy.messages.in: Added.
• UIProcess/WebProcessProxy.cpp:

(WebKit::m_webLockRegistry):
(WebKit::WebProcessProxy::shutDown):

• UIProcess/WebProcessProxy.h:
• UIProcess/WebsiteData/WebsiteDataStore.cpp:

(WebKit::WebsiteDataStore::WebsiteDataStore):

• UIProcess/WebsiteData/WebsiteDataStore.h:

(WebKit::WebsiteDataStore::webLockRegistry):

• WebKit.xcodeproj/project.pbxproj:
• WebProcess/Storage/WebSWContextManagerConnection.cpp:

(WebKit::WebSWContextManagerConnection::installServiceWorker):

• WebProcess/WebCoreSupport/RemoteWebLockRegistry.cpp: Added.

(WebKit::RemoteWebLockRegistry::LocksSnapshot::isEmpty const):
(WebKit::RemoteWebLockRegistry::RemoteWebLockRegistry):
(WebKit::RemoteWebLockRegistry::~RemoteWebLockRegistry):
(WebKit::RemoteWebLockRegistry::requestLock):
(WebKit::RemoteWebLockRegistry::releaseLock):
(WebKit::RemoteWebLockRegistry::abortLockRequest):
(WebKit::RemoteWebLockRegistry::snapshot):
(WebKit::RemoteWebLockRegistry::clientIsGoingAway):
(WebKit::RemoteWebLockRegistry::didCompleteLockRequest):
(WebKit::RemoteWebLockRegistry::didStealLock):

• WebProcess/WebCoreSupport/RemoteWebLockRegistry.h: Added.
• WebProcess/WebCoreSupport/RemoteWebLockRegistry.messages.in: Added.
• WebProcess/WebPage/WebPage.cpp:

(WebKit::m_appHighlightsVisible):

• WebProcess/WebProcess.cpp:

(WebKit::WebProcess::WebProcess):

• WebProcess/WebProcess.h:

(WebKit::WebProcess::webLockRegistry):

Source/WebKitLegacy/mac:

See WebCore changelog.

• WebView/WebView.mm:

(getOrCreateWebLockRegistry):
(-[WebView _commonInitializationWithFrameName:groupName:]):
(-[WebView initSimpleHTMLDocumentWithStyle:frame:preferences:groupName:]):

Source/WebKitLegacy/win:

See WebCore Changelog.

• WebView.cpp:

(getOrCreateWebLockRegistry):
(WebView::initWithFrame):

Tools:

Add API test coverage.

• TestWebKitAPI/SourcesCocoa.txt:
• TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
• TestWebKitAPI/Tests/WebKitCocoa/WebLocks.mm: Added.

(TestWebKitAPI::TEST):

2:27 PM Changeset in webkit [286454] by Devin Rousso

• 2 edits in trunk/Source/WebCore

[Apple Pay] recurring line items appear way in the future
​https://bugs.webkit.org/show_bug.cgi?id=233779
<rdar://problem/85976407>

Reviewed by Wenson Hsieh.

JavaScript Date uses a milliseconds-based value, whereas native NSDate is seconds-based.

• Modules/applepay/cocoa/PaymentSummaryItemsCocoa.mm:

(WebCore::toDate):

2:22 PM Changeset in webkit [286453] by Chris Dumez

• 5 edits in trunk/Source/WebCore

MessageChannel::port1() / port2() should return references instead of pointers
​https://bugs.webkit.org/show_bug.cgi?id=233772

Reviewed by Alex Christensen.

This was extracted from Alex Christensen's large patch at Bug 230382.

• Modules/webaudio/AudioWorkletNode.cpp:

(WebCore::AudioWorkletNode::create):

• bindings/js/JSMessageChannelCustom.cpp:

(WebCore::JSMessageChannel::visitAdditionalChildren):

• dom/MessageChannel.cpp:

(WebCore::generateMessagePorts):
(WebCore::MessageChannel::MessageChannel):

• dom/MessageChannel.h:

(WebCore::MessageChannel::port1 const):
(WebCore::MessageChannel::port2 const):

2:21 PM Changeset in webkit [286452] by Devin Rousso

• 13 edits in trunk

[Payment Request] Validate payment method data on construction
​https://bugs.webkit.org/show_bug.cgi?id=233292
<rdar://problem/85736007>

Reviewed by Andy Estes.

Source/WebCore:

This will allow developers to replace any ApplePaySession.supportsVersion check(s) by
attempting to create a PaymentRequest object with data specific to Apple Pay. This is
actually an improvement, because it'll also allow developers to catch most errors in that
data earlier, as previously any errors would only be thrown when show() is called.

Spec: <​https://github.com/w3c/payment-request/pull/976>

Test: http/tests/ssl/applepay/PaymentRequest.https.html

• Modules/paymentrequest/PaymentRequest.cpp:

(WebCore::PaymentRequest::show):

• Modules/paymentrequest/PaymentHandler.h:

Adjust PaymentHandler::convertData to take a Document so that additional validation
can be performed on the provided payment method(s) (e.g. check the Apple Pay version).

• Modules/applepay/PaymentRequestValidator.h:
• Modules/applepay/PaymentRequestValidator.mm:

(WebCore::PaymentRequestValidator::validate):
Allow callers to choose what fields to validate. This is needed because when the
ApplePaySessionPaymentRequest created/converted when validating the payment method data in
the PaymentRequest constructor isn't able to get fields from the PaymentRequest as it
hasn't been created yet.

• Modules/applepay/ApplePaySession.cpp:

(WebCore::convertAndValidate):

• Modules/applepay/paymentrequest/ApplePayPaymentHandler.h:
• Modules/applepay/paymentrequest/ApplePayPaymentHandler.cpp:

(WebCore::convertAndValidateApplePayRequest):
(WebCore::ApplePayPaymentHandler::convertData):
(WebCore::ApplePayPaymentHandler::show):

• Modules/applepay-ams-ui/ApplePayAMSUIPaymentHandler.h:
• Modules/applepay-ams-ui/ApplePayAMSUIPaymentHandler.cpp:

(WebCore::convertAndValidateApplePayAMSUIRequest):
(WebCore::ApplePayAMSUIPaymentHandler::convertData):

LayoutTests:

• http/tests/ssl/applepay/PaymentRequest.https.html:
• http/tests/ssl/applepay/PaymentRequest.https-expected.txt:

1:51 PM Changeset in webkit [286451] by Fujii Hironori

• 9 edits in trunk/LayoutTests

[WinCairo] Unreviewed test gardening

• platform/wincairo-wk1/TestExpectations:
• platform/wincairo/TestExpectations:
• platform/wincairo/fast/box-shadow/basic-shadows-expected.txt:
• platform/wincairo/fast/dom/52776-expected.txt:
• platform/wincairo/fast/text/atsui-rtl-override-selection-expected.txt:
• platform/wincairo/fast/text/international/bidi-LDB-2-formatting-characters-expected.txt:
• platform/wincairo/fast/text/international/bidi-ignored-for-first-child-inline-expected.txt:
• platform/wincairo/fast/text/international/bidi-neutral-run-expected.txt:

1:33 PM Changeset in webkit [286450] by Wenson Hsieh

• 4 edits in trunk

[ Monterey ] fast/images/text-recognition/mac/image-overlay-text-disables-app-highlight-menu-items.html (layout-test) is a constant text failure
​https://bugs.webkit.org/show_bug.cgi?id=233658
rdar://problem/85886424

Reviewed by Tim Horton.

Source/WTF:

Define ENABLE_APP_HIGHLIGHTS for iOS 15+ and macOS Monterey+. This ensures that
internals.appHighlightContextMenuItemTitles is defined in layout tests on Monterey.

• wtf/PlatformEnableCocoa.h:

Tools:

Add a forward declaration for open source builds.

• TestWebKitAPI/ios/UIKitSPI.h:

1:26 PM Changeset in webkit [286449] by commit-queue@webkit.org

• 3 edits in trunk/Source/WebCore

Verify borderRect is Renderable
​https://bugs.webkit.org/show_bug.cgi?id=232650

Patch by Brandon Stewart <Brandon> on 2021-12-02
Reviewed by Alan Bujtas.

Verify that borderRect is renderable.
Align isRenderable() implementations in RoundedRect and FloatRoundedRect.

• platform/graphics/RoundedRect.cpp:

(WebCore::RoundedRect::isRenderable const):

• rendering/RenderBoxModelObject.cpp:

(WebCore::RenderBoxModelObject::paintBoxShadow):

1:25 PM Changeset in webkit [286448] by Alan Bujtas

• 2 edits in trunk/Source/WebCore

[LFC][IFC] Redundant line run type checking in InlineDisplayContentBuilder::processNonBidiContent
​https://bugs.webkit.org/show_bug.cgi?id=233597

Reviewed by Antti Koivisto.

• layout/formattingContexts/inline/InlineDisplayContentBuilder.cpp:

(WebCore::Layout::InlineDisplayContentBuilder::processNonBidiContent):

1:13 PM Changeset in webkit [286447] by commit-queue@webkit.org

• 35 edits in trunk/Source/WebCore

Remove the multipart parameter to FormAssociatedElement::appendFormData, InputType::appendFormData and HTMLFormElement::constructEntryList
​https://bugs.webkit.org/show_bug.cgi?id=233774

Patch by Andreu Botella <​andreu@andreubotella.com> on 2021-12-02
Reviewed by Chris Dumez.

These parameters were only needed because of an optimization where file inputs were being
added to the DOMFormData as string values containing their filenames if the entry list was
being constructed to be used in a urlencoded or text/plain form submission. This
optimization was removed in r286427 and those parameters can now be dropped.

No behavior change.

• html/BaseButtonInputType.cpp:

(WebCore::BaseButtonInputType::appendFormData const):

• html/BaseButtonInputType.h:
• html/BaseCheckableInputType.cpp:

(WebCore::BaseCheckableInputType::appendFormData const):

• html/BaseCheckableInputType.h:
• html/DOMFormData.cpp:

(WebCore::DOMFormData::create):

• html/FileInputType.cpp:

(WebCore::FileInputType::appendFormData const):

• html/FileInputType.h:
• html/FormAssociatedElement.h:

(WebCore::FormAssociatedElement::appendFormData):

• html/HTMLButtonElement.cpp:

(WebCore::HTMLButtonElement::appendFormData):

• html/HTMLButtonElement.h:
• html/HTMLFormControlElement.h:
• html/HTMLFormElement.cpp:

(WebCore::HTMLFormElement::constructEntryList):

• html/HTMLFormElement.h:
• html/HTMLInputElement.cpp:

(WebCore::HTMLInputElement::appendFormData):

• html/HTMLInputElement.h:
• html/HTMLKeygenElement.cpp:

(WebCore::HTMLKeygenElement::appendFormData):

• html/HTMLKeygenElement.h:
• html/HTMLObjectElement.cpp:

(WebCore::HTMLObjectElement::appendFormData):

• html/HTMLObjectElement.h:
• html/HTMLSelectElement.cpp:

(WebCore::HTMLSelectElement::appendFormData):

• html/HTMLSelectElement.h:
• html/HTMLTextAreaElement.cpp:

(WebCore::HTMLTextAreaElement::appendFormData):

• html/HTMLTextAreaElement.h:
• html/HiddenInputType.cpp:

(WebCore::HiddenInputType::appendFormData const):

• html/HiddenInputType.h:
• html/ImageInputType.cpp:

(WebCore::ImageInputType::appendFormData const):

• html/ImageInputType.h:
• html/InputType.cpp:

(WebCore::InputType::appendFormData const):

• html/InputType.h:
• html/SubmitInputType.cpp:

(WebCore::SubmitInputType::appendFormData const):

• html/SubmitInputType.h:
• html/TextFieldInputType.cpp:

(WebCore::TextFieldInputType::appendFormData const):

• html/TextFieldInputType.h:
• loader/FormSubmission.cpp:

(WebCore::FormSubmission::create):

1:04 PM Changeset in webkit [286446] by Alan Bujtas

• 2 edits in trunk/Source/WebCore

[LFC][IFC] Introduce createDisplayBoxesInVisualOrderForContentRuns
​https://bugs.webkit.org/show_bug.cgi?id=233733

Reviewed by Antti Koivisto.

Revert some of the inline box related bidi changes and start with
a very simple "let's create the display boxes for all _content_ runs in visual order".
This is very similar to InlineDisplayContentBuilder::processNonBidiContent() where we just
loop through the content and construct display boxes.
This is also in preparation for advancing with the inline box case.

• layout/formattingContexts/inline/InlineDisplayContentBuilder.cpp:

(WebCore::Layout::InlineDisplayContentBuilder::processNonBidiContent):
(WebCore::Layout::InlineDisplayContentBuilder::processBidiContent):

12:22 PM Changeset in webkit [286445] by Alan Bujtas

• 3 edits in trunk/Source/WebCore

[LFC][IFC] Move inline box geometry (BoxGeometry) update logic to a dedicated function
​https://bugs.webkit.org/show_bug.cgi?id=233615

Reviewed by Antti Koivisto.

This is also in preparation for computing the BoxGeometry for bidi inline boxes.

• layout/formattingContexts/inline/InlineDisplayContentBuilder.cpp:

(WebCore::Layout::InlineDisplayContentBuilder::setInlineBoxGeometry):
(WebCore::Layout::InlineDisplayContentBuilder::appendInlineBoxDisplayBox):
(WebCore::Layout::InlineDisplayContentBuilder::appendSpanningInlineBoxDisplayBox):

• layout/formattingContexts/inline/InlineDisplayContentBuilder.h:

12:09 PM Changeset in webkit [286444] by Chris Dumez

• 4 edits in trunk

html/semantics/forms/constraints/input-number-validity-dynamic-value-no-change.html WPT test is failing
​https://bugs.webkit.org/show_bug.cgi?id=233765

Reviewed by Darin Adler.

LayoutTests/imported/w3c:

Rebaseline WPT test that is now passing.

• web-platform-tests/html/semantics/forms/constraints/input-number-validity-dynamic-value-no-change-expected.txt:

Source/WebCore:

Per ​https://html.spec.whatwg.org/#concept-input-min-zero, step base needs to be computed based on
the 'min' attribute, then fallback to the 'value' attribute and finally fallback to the default
step base. WebKit was missing the "fallback to the 'value' attribute logic.

No new tests, rebaselined existing test.

• html/NumberInputType.cpp:

(WebCore::NumberInputType::createStepRange const):

11:54 AM Changeset in webkit [286443] by Russell Epstein

• 3 edits in branches/safari-612-branch

Revert r286394. rdar://problem/83070565

11:54 AM Changeset in webkit [286442] by Russell Epstein

• 3 edits
  2 deletes in branches/safari-612-branch

Revert r286395. rdar://problem/83070565

11:49 AM Changeset in webkit [286441] by Ryan Haddad

• 3 edits
  2 adds in trunk

Fix crash in GraphicsContextCG::endTransparencyLayer
​https://bugs.webkit.org/show_bug.cgi?id=230230

Patch by Tim Nguyen <​ntim@apple.com> on 2021-12-02
Reviewed by Myles C. Maxfield.

Source/WebCore:

The crash was due to unbalanced calls to begin and end transparency layers.

A branch handling ancestors of transparent layers that are transform root needed to be
aware of the top layer. Opacity on ancestors don't affect top layer elements so calling
beginTransparencyLayers on parent() is incorrect.

Also fix transparentPaintingAncestor() to be top layer aware to avoid flickering layers
while scrolling.

Test: fast/layers/top-layer-ancestor-opacity-and-transform-crash.html

• rendering/RenderLayer.cpp:

(WebCore::RenderLayer::transparentPaintingAncestor):
(WebCore::RenderLayer::paintLayerWithEffects):

LayoutTests:

• fast/layers/top-layer-ancestor-opacity-and-transform-crash-expected.txt: Added.
• fast/layers/top-layer-ancestor-opacity-and-transform-crash.html: Added.

11:39 AM Changeset in webkit [286440] by beidson@apple.com

• 2 edits in trunk/Tools

REGRESSION: r286075 [ Big Sur x86] 2 TestWebKitAPI.WebPushD (API-Tests) are constant failures
​https://bugs.webkit.org/show_bug.cgi?id=233726

Unreviewed gardening.

• Disabling these tests on all Mac platforms until we can figure out the code signing/etc situation
• Fix the truly broken "basic" test everywhere by adapting to messaging changes

• TestWebKitAPI/Tests/WebKitCocoa/WebPushDaemon.mm:

(TestWebKitAPI::TEST):

11:35 AM Changeset in webkit [286439] by Robert Jenner

• 3 edits
  2 deletes in trunk

Unreviewed, reverting r285570.

Broke StarD build

Reverted changeset:

"Fix crash in GraphicsContextCG::endTransparencyLayer"
​https://bugs.webkit.org/show_bug.cgi?id=230230
​https://commits.webkit.org/r285570

11:19 AM Changeset in webkit [286438] by commit-queue@webkit.org

• 6 edits in trunk/Source

Unreviewed, reverting r286437.
​https://bugs.webkit.org/show_bug.cgi?id=233776

Introduced test failure

Reverted changeset:

"[WP] Strengthen sandbox when AppCache is disabled"
​https://bugs.webkit.org/show_bug.cgi?id=233746
​https://commits.webkit.org/r286437

10:44 AM Changeset in webkit [286437] by pvollan@apple.com

• 6 edits in trunk/Source

[WP] Strengthen sandbox when AppCache is disabled
​https://bugs.webkit.org/show_bug.cgi?id=233746
<rdar://problem/85953893>

Reviewed by Brent Fulgham.

Source/WebKit:

When AppCache is disabled, we can remove access to some resources in the WebContent process' sandbox.
This is implemented using a new sandbox state variable representing the AppCache state.

• Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb.in:
• WebProcess/WebPage/WebPage.cpp:
• WebProcess/com.apple.WebProcess.sb.in:

Source/WTF:

Add HAVE macro for sandbox state flags support.

• wtf/PlatformHave.h:

9:57 AM Changeset in webkit [286436] by Russell Epstein

• 2 edits in branches/safari-613.1.10-branch/Source/WebKit

Cherry-pick r286381. rdar://problem/85832755

[WP] Sandbox telemetry is missing for some system calls
​https://bugs.webkit.org/show_bug.cgi?id=233594
<rdar://problem/85832755>

Reviewed by Brent Fulgham.

Sandbox telemetry is missing for some system calls, since telemetry rules are automatically overridden in some cases.
This patch is addressing this by disabling system call inference.

• WebProcess/com.apple.WebProcess.sb.in:

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@286381 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9:57 AM Changeset in webkit [286435] by Russell Epstein

• 2 edits in branches/safari-613.1.10-branch/Source/WebKit

Cherry-pick r286266. rdar://problem/85832755

[WP] Sandbox telemetry is missing for some system calls
​https://bugs.webkit.org/show_bug.cgi?id=233594
<rdar://problem/85832755>

Reviewed by Brent Fulgham.

Sandbox telemetry is missing for some system calls, since telemetry rules are automatically overridden in some cases.
This patch is addressing this by disabling system call inference.

• Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb.in:

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@286266 268f45cc-cd09-0410-ab3c-d52691b4dbfc

9:48 AM Changeset in webkit [286434] by Aditya Keerthi

• 2 edits in trunk/Source/WTF

[macCatalyst] Enable support for date/time inputs
​https://bugs.webkit.org/show_bug.cgi?id=233710
rdar://80108114

Reviewed by Tim Horton.

Date/time inputs were unsupported due to the lack of support for custom
views inside UIContextMenuInteractions on macCatalyst.

However, after date/time pickers transitioned to use
_UIDatePickerOverlayPresentation in iOS 15, there is nothing preventing
them from functioning on macCatalyst.

• Scripts/Preferences/WebPreferencesInternal.yaml:

9:36 AM Changeset in webkit [286433] by Antti Koivisto

• 6 edits
  8 adds in trunk

[:has() pseudo-class] Invalidation in non-subject position
​https://bugs.webkit.org/show_bug.cgi?id=233758

Reviewed by Simon Fraser.

LayoutTests/imported/w3c:

• web-platform-tests/css/selectors/invalidation/has-in-adjacent-position-expected.txt: Added.
• web-platform-tests/css/selectors/invalidation/has-in-adjacent-position.html: Added.
• web-platform-tests/css/selectors/invalidation/has-in-ancestor-position-expected.txt: Added.
• web-platform-tests/css/selectors/invalidation/has-in-ancestor-position.html: Added.
• web-platform-tests/css/selectors/invalidation/has-in-parent-position-expected.txt: Added.
• web-platform-tests/css/selectors/invalidation/has-in-parent-position.html: Added.
• web-platform-tests/css/selectors/invalidation/has-in-sibling-position-expected.txt: Added.
• web-platform-tests/css/selectors/invalidation/has-in-sibling-position.html: Added.

Source/WebCore:

Invalidation for selectors like '.ancestor:has(.foo) #target'.

Tests: imported/w3c/web-platform-tests/css/selectors/invalidation/has-in-adjacent-position.html

imported/w3c/web-platform-tests/css/selectors/invalidation/has-in-ancestor-position.html
imported/w3c/web-platform-tests/css/selectors/invalidation/has-in-parent-position.html
imported/w3c/web-platform-tests/css/selectors/invalidation/has-in-sibling-position.html

• style/ChildChangeInvalidation.cpp:

(WebCore::Style::ChildChangeInvalidation::invalidateForChangedElement):
(WebCore::Style::needsTraversal):
(WebCore::Style::needsDescendantTraversal):

• style/RuleFeature.cpp:

(WebCore::Style::isSiblingOrSubject):
(WebCore::Style::isHasPseudoClassMatchElement):
(WebCore::Style::computeHasPseudoClassMatchElement):
(WebCore::Style::computeSubSelectorMatchElement):

Use new MatchElement::HasNonSubject as a catch-all for cases where :has() is in a non-subject position.
In the future this can be optimized better by computing both the :has match element and the overall
match element separately.

• style/RuleFeature.h:
• style/StyleInvalidator.cpp:

(WebCore::Style::Invalidator::invalidateStyleWithMatchElement):

Worst-case invalidation.

8:44 AM Changeset in webkit [286432] by youenn@apple.com

• 3 edits in trunk/Source/WebKit

Add some logging to NetworkRTCProvider
​https://bugs.webkit.org/show_bug.cgi?id=233759

Reviewed by Eric Carlson.

No change of behavior.

• NetworkProcess/webrtc/NetworkRTCProvider.cpp:

(WebKit::NetworkRTCProvider::addSocket):

• NetworkProcess/webrtc/NetworkRTCUDPSocketCocoa.mm:

(WebKit::NetworkRTCUDPSocketCocoaConnections::NetworkRTCUDPSocketCocoaConnections):

8:33 AM Changeset in webkit [286431] by ap@apple.com

• 2 edits in trunk/Source/WTF

macOS Monterey build fix after r286420 (suggested by Brady).

• wtf/PlatformHave.h:

8:28 AM Changeset in webkit [286430] by pvollan@apple.com

• 2 edits in trunk/Source/WebKit

REGRESSION(r285722): Introduced performance regression
​https://bugs.webkit.org/show_bug.cgi?id=233718

Unreviewed partial revert of r285722.

• Resources/SandboxProfiles/ios/com.apple.WebKit.GPU.sb:

7:46 AM Changeset in webkit [286429] by commit-queue@webkit.org

• 6 edits in trunk/Source/bmalloc

[libpas] Define and use PAS_RISCV
​https://bugs.webkit.org/show_bug.cgi?id=233753

Patch by Zan Dobersek <​zdobersek@igalia.com> on 2021-12-02
Reviewed by Yusuke Suzuki.

Define PAS_RISCV as appropriate and use it accordingly to avoid custom
x86-64 assembly, much like it's avoided for PAS_ARM. This gets libpas
compiling for RISC-V targets.

• libpas/src/libpas/pas_config.h:
• libpas/src/libpas/pas_config_prefix.h:
• libpas/src/libpas/pas_segregated_page_inlines.h:

(pas_segregated_page_deallocate_with_page):

• libpas/src/libpas/pas_utils.h:
• libpas/src/libpas/pas_utils_prefix.h:

(pas_fence):

7:41 AM Changeset in webkit [286428] by commit-queue@webkit.org

• 2 edits in trunk/Source/JavaScriptCore

[RISCV64] Fix effective address loading for LabelReferences with offsets
​https://bugs.webkit.org/show_bug.cgi?id=233754

Patch by Zan Dobersek <​zdobersek@igalia.com> on 2021-12-02
Reviewed by Yusuke Suzuki.

r286345 (and subsequent change in r286372) introduced a load from a
label address with an additional offset. To properly handle this,
RISCV64 offlineasm generates the desired load-effective-address
instruction but now also generates an additional add instruction when
an offset value is present for that lea.

• offlineasm/riscv64.rb:

7:34 AM Changeset in webkit [286427] by commit-queue@webkit.org

• 8 edits in trunk

File inputs in non-multipart form submissions show up as string values in the formdata event
​https://bugs.webkit.org/show_bug.cgi?id=233725

Patch by Andreu Botella <​andreu@andreubotella.com> on 2021-12-02
Reviewed by Chris Dumez.

LayoutTests/imported/w3c:

Added tests in web-platform-tests/html/semantics/forms/form-submission-0/constructing-form-data-set.html
to check that file inputs show up as File entries in the formdata event.

• web-platform-tests/html/semantics/forms/form-submission-0/constructing-form-data-set-expected.txt:
• web-platform-tests/html/semantics/forms/form-submission-0/constructing-form-data-set.html:
• web-platform-tests/html/semantics/forms/form-submission-0/text-plain.window-expected.txt:
• web-platform-tests/html/semantics/forms/form-submission-0/urlencoded2.window-expected.txt:

Source/WebCore:

When a form's entry list was constructed form a non-multipart/form-data form submission,
file inputs were added to the DOMFormData object as string values consisting of the file's
filename. Before r280310, this optimization was unobservable, since the DOMFormData object
was never exposed to the user halfway through a form submission.

r280310 changed this by adding the formdata event, which fires at the end of the "construct
the entry list" algorithm and can be used to both inspect and modify the DOMFormData object.
This means that it is now observable that file inputs in urlencoded and text/plain form
submissions are encoded as their filenames, and that any new File entries added to the
DOMFormData object the event listeners will be skipped when building the form payload.

This change fixes that by always adding file inputs to DOMFormData as File entries, and
using a File's filename as the string value in FormData::appendNonMultipartKeyValuePairItems

Tests: imported/w3c/web-platform-tests/html/semantics/forms/form-submission-0/constructing-form-data-set.html

imported/w3c/web-platform-tests/html/semantics/forms/form-submission-0/text-plain.window.html
imported/w3c/web-platform-tests/html/semantics/forms/form-submission-0/urlencoded2.window.html

• html/FileInputType.cpp:

(WebCore::FileInputType::appendFormData const): Changed to always add File entries to the
DOMFormData object, rather than adding the filenames as string values when multipart is
false.

• platform/network/FormData.cpp:

(WebCore::FormData::appendNonMultiPartKeyValuePairItems): Changed to encode File entry
values as their filenames, rather than skipping them.

7:25 AM Changeset in webkit [286426] by Tyler Wilcock

• 5 edits in trunk

AX Isolated Tree: Document links should be stored in web areas, not the root
​https://bugs.webkit.org/show_bug.cgi?id=233728

Reviewed by Chris Fleizach.

Source/WebCore:

In WebAccessibilityObjectWrapperMac.mm, we expose the document links attribute (AXLinkUIElements)
on web areas. However, in AXIsolatedObject::initializeAttributeData we currently only store
AXPropertyName::DocumentLinks for the root element, which is not a web area.

In this patch, we store document links for web area elements instead
of the root element.

This patch fixes accessibility/visible-elements.html, which tests document links.

​https://github.com/WebKit/WebKit/blob/109c8234854a0fa28955d81994bb98daa1fa14fe/Source/WebCore/accessibility/mac/WebAccessibilityObjectWrapperMac.mm#L2235#L2237

• accessibility/isolatedtree/AXIsolatedObject.cpp:

(WebCore::AXIsolatedObject::initializeAttributeData):

LayoutTests:

• accessibility/visible-elements.html:
• accessibility/visible-elements-expected.txt:

Make this test async so it passes in isolated tree mode.
This test must be async because it includes JS that dynamically
changes page content.

7:21 AM Changeset in webkit [286425] by commit-queue@webkit.org

• 6 edits in trunk

Add a fast path for empty string to setInnerHTML()
​https://bugs.webkit.org/show_bug.cgi?id=233647

Patch by Alexey Shvayka <​ashvayka@apple.com> on 2021-12-02
Reviewed by Geoff Garen.

LayoutTests/imported/w3c:

• web-platform-tests/html/semantics/scripting-1/the-template-element/innerhtml-on-templates/innerhtml-expected.txt:
• web-platform-tests/html/semantics/scripting-1/the-template-element/innerhtml-on-templates/innerhtml.html:

Source/WebCore:

This patch adds a fast path for element.innerHTML = "", which is a common idiom for
removing all children from an element, while ensuring <template> contents is modified
rather than its children [1], and that mutation records are enqueued.

Although there are quite a few insertion modes [1], parsing empty string creates
additional elements only inside <html> container (please see "Anything else" clauses).

Bypassing parser instantiation results in 3.2x progression of the attached microbenchmark.
Both Gecko and Blink has this optimization, which was reported to progress Speedometer2
score (especially Vanilla / jQuery / Preact subtests).

Our textContent and innerText setters are already fast enough for empty strings.

[1] ​https://w3c.github.io/DOM-Parsing/#dom-innerhtml-innerhtml (setting, step 3)
[2] ​https://html.spec.whatwg.org/multipage/parsing.html#parsing-main-inhtml

No behavior change.

• dom/Element.cpp:

(WebCore::Element::setInnerHTML):

• dom/ShadowRoot.cpp:

(WebCore::ShadowRoot::setInnerHTML):

6:26 AM Changeset in webkit [286424] by commit-queue@webkit.org

• 13 edits in trunk/Source/JavaScriptCore

[JSC] Generated code size reductions for baseline JIT (all architectures)
​https://bugs.webkit.org/show_bug.cgi?id=233474

Patch by Geza Lore <Geza Lore> on 2021-12-02
Reviewed by Yusuke Suzuki.

This patch introduces a few improvements that reduce the generated
code size.

Target independent improvements to the Baseline JIT:

1. Some bytecodes that are very frequent (e.g.: get_by_id, call) share

the same instructions at the tail end of the fast and slow paths.
Instead of duplicating these in the slow path, then branch to the next
sequential bytecode on the fast path, make the slow path branch to and
reuse these common instructions, which then naturally fall through to
the next sequential bytecode.

2. Minor tweaks in a few places to remove redundant reloading of

immediates and remove redundant moves.

3. Remove a small number of redundant unconditional branches from some

DataIC fast paths.

ARMv7/Thumb-2 specific improvements:

4. Add assembler support for LDRD and STRD (load/store a pair of

32-bit GPRs) and use them throughout via loadValue/storeValue. This
yields denser code as it often eliminates repeated temporary register
setups (especially for a BaseIndex access), and also due to point 4
below. This is also potentially a performance improvement on
micro-architectures with a 64-bit LSU data-path.

5. Instructions using only r0-r7 as operands can often use a short,

16-bit encoding in Thumb-2, so prefer to use low order registers
as temporaries wherever possible.

The net effect of this patch is that the emitted baseline code during
a run of JetStream2 is ~6.6% smaller on x86_64, ~5.1% smaller on
ARM64, and ~24% smaller on ARMv7/Thumb-2. On ARMv7/Thumb-2, DFG code
is also ~5.3% smaller, while on other architectures the DFG code is
unaffected.

On ARMv7/Thumb-2, this patch also yields an ~2% improvement in
JetStream2 scores on my test machine.

• assembler/ARMv7Assembler.h:

(JSC::ARMv7Assembler::ldrd):
(JSC::ARMv7Assembler::strd):
(JSC::ARMv7Assembler::ARMInstructionFormatter::twoWordOp12Reg4Reg4Reg4Imm8):

• assembler/MacroAssembler.h:

(JSC::MacroAssembler::addPtr):

• assembler/MacroAssemblerARMv7.h:

(JSC::MacroAssemblerARMv7::bestTempRegister):
(JSC::MacroAssemblerARMv7::scratchRegister):
(JSC::MacroAssemblerARMv7::add32):
(JSC::MacroAssemblerARMv7::sub32):
(JSC::MacroAssemblerARMv7::loadPair32):
(JSC::MacroAssemblerARMv7::store32):
(JSC::MacroAssemblerARMv7::storePair32):
(JSC::MacroAssemblerARMv7::compare32AndSetFlags):
(JSC::MacroAssemblerARMv7::test32):
(JSC::MacroAssemblerARMv7::branch32):
(JSC::MacroAssemblerARMv7::farJump):
(JSC::MacroAssemblerARMv7::call):
(JSC::MacroAssemblerARMv7::compare32):

• assembler/MacroAssemblerMIPS.h:

(JSC::MacroAssemblerMIPS::loadPair32):
(JSC::MacroAssemblerMIPS::storePair32):

• jit/AssemblyHelpers.h:

(JSC::AssemblyHelpers::storeValue):
(JSC::AssemblyHelpers::loadValue):

• jit/JIT.cpp:

(JSC::JIT::privateCompileSlowCases):

• jit/JIT.h:
• jit/JITCall.cpp:

(JSC::JIT::compileCallEval):
(JSC::JIT::compileCallEvalSlowCase):
(JSC::JIT::compileOpCall):
(JSC::JIT::compileOpCallSlowCase):
(JSC::JIT::emitSlow_op_iterator_open):
(JSC::JIT::emitSlow_op_iterator_next):

• jit/JITInlineCacheGenerator.cpp:

(JSC::generateGetByIdInlineAccess):
(JSC::JITPutByIdGenerator::generateBaselineDataICFastPath):

• jit/JITInlineCacheGenerator.h:
• jit/JITInlines.h:

(JSC::JIT::setFastPathResumePoint):
(JSC::JIT::fastPathResumePoint const):

• jit/JITOpcodes.cpp:

(JSC::JIT::emit_op_enter):

• jit/JITPropertyAccess.cpp:

(JSC::JIT::emit_op_get_by_val):
(JSC::JIT::generateGetByValSlowCase):
(JSC::JIT::emit_op_get_private_name):
(JSC::JIT::emitSlow_op_get_private_name):
(JSC::JIT::emit_op_try_get_by_id):
(JSC::JIT::emitSlow_op_try_get_by_id):
(JSC::JIT::emit_op_get_by_id_direct):
(JSC::JIT::emitSlow_op_get_by_id_direct):
(JSC::JIT::emit_op_get_by_id):
(JSC::JIT::emitSlow_op_get_by_id):
(JSC::JIT::emit_op_get_by_id_with_this):
(JSC::JIT::emitSlow_op_get_by_id_with_this):
(JSC::JIT::emit_op_in_by_id):
(JSC::JIT::emitSlow_op_in_by_id):
(JSC::JIT::emit_op_in_by_val):
(JSC::JIT::emitSlow_op_in_by_val):
(JSC::JIT::emitHasPrivate):
(JSC::JIT::emitHasPrivateSlow):
(JSC::JIT::emitSlow_op_has_private_name):
(JSC::JIT::emitSlow_op_has_private_brand):
(JSC::JIT::emit_op_enumerator_get_by_val):
(JSC::JIT::emitWriteBarrier):

6:15 AM Changeset in webkit [286423] by commit-queue@webkit.org

• 19 edits
  2 copies in trunk

GraphicsContextGLANGLE should not have Cocoa Mac specific display reconfiguration code
​https://bugs.webkit.org/show_bug.cgi?id=233695

Patch by Kimmo Kinnunen <​kkinnunen@apple.com> on 2021-12-02
Reviewed by Antti Koivisto.

Source/WebCore:

Move code to observe Mac display reconfiguration
away from GraphicsContextGLANGLE to GraphicsContextGLCocoa.
Apply the display reconfiguration listener only for
WebProcessGraphicsContextGLCocoa, it is not needed for the
GPU process contexts.

Move the listener notification code away from GraphicsContextGLOpenGLManager
into its own class. This way GraphicsContextGLOpenGLManager does not need
to know Mac specific interfaces. Later on, GraphicsContextGLOpenGLManager can
be removed altogether.

Changes behavior so that simulateEventForTesting invokes the simulated context
change only for the context being tested. This simplifies the implementation.

No new tests, refactor.

• PlatformMac.cmake:
• SourcesCocoa.txt:
• WebCore.xcodeproj/project.pbxproj:
• platform/graphics/GraphicsContextGLAttributes.h:

(WebCore::GraphicsContextGLAttributes::effectivePowerPreference const):

• platform/graphics/angle/GraphicsContextGLANGLE.cpp:

(WebCore::GraphicsContextGLANGLE::simulateEventForTesting):

• platform/graphics/angle/GraphicsContextGLANGLE.h:
• platform/graphics/cocoa/GraphicsContextGLCocoa.h:
• platform/graphics/cocoa/GraphicsContextGLCocoa.mm:

(WebCore::initializeEGLDisplay):
(WebCore::GraphicsContextGLCocoa::create):
(WebCore::GraphicsContextGLCocoa::GraphicsContextGLCocoa):
(WebCore::GraphicsContextGLCocoa::~GraphicsContextGLCocoa):
(WebCore::GraphicsContextGLANGLE::GraphicsContextGLANGLE):
(WebCore::GraphicsContextGLCocoa::setContextVisibility):
(WebCore::GraphicsContextGLCocoa::updateContextOnDisplayReconfiguration):
(WebCore::GraphicsContextGLANGLE::setContextVisibility): Deleted.
(WebCore::GraphicsContextGLANGLE::displayWasReconfigured): Deleted.

• platform/graphics/cocoa/WebProcessGraphicsContextGLCocoa.mm:
• platform/graphics/mac/DisplayConfigurationMonitor.cpp: Copied from Source/WebCore/platform/graphics/opengl/GraphicsContextGLOpenGLManager.h.

(WebCore::DisplayConfigurationMonitor::singleton):
(WebCore::DisplayConfigurationMonitor::addClient):
(WebCore::DisplayConfigurationMonitor::removeClient):
(WebCore::DisplayConfigurationMonitor::dispatchDisplayWasReconfigured):

• platform/graphics/mac/DisplayConfigurationMonitor.h: Copied from Source/WebCore/platform/graphics/opengl/GraphicsContextGLOpenGLManager.h.
• platform/graphics/opengl/GraphicsContextGLOpenGLManager.cpp:

(WebCore::GraphicsContextGLOpenGLManager::displayWasReconfigured): Deleted.

• platform/graphics/opengl/GraphicsContextGLOpenGLManager.h:

Source/WebKit:

Previous GraphicsContextGLANGLE::displayWasReconfigured()
was renamed to updateContextOnDisplayReconfiguration().
The displayWasReconfigured() was taken into use in the notification interface
DisplayConfigurationMonitor::Client::displayWasReconfigured().

• GPUProcess/graphics/RemoteGraphicsContextGL.cpp:

(WebKit::RemoteGraphicsContextGL::displayWasReconfigured):

• WebProcess/cocoa/WebProcessCocoa.mm:

(WebKit::WebProcess::displayConfigurationChanged):

LayoutTests:

Change the test to account for the behavior change
where the simulated context change event is delivered
only to the context being tested, not to all contexts.

• fast/canvas/webgl/webglcontextchangedevent.html:

6:11 AM Changeset in webkit [286422] by commit-queue@webkit.org

• 2 edits in trunk/Source/WebCore

REGRESSION(r286209) [GTK][WPE] crashes in offscreen canvas tests that try to create a GL context
​https://bugs.webkit.org/show_bug.cgi?id=233712

Patch by Kimmo Kinnunen <​kkinnunen@apple.com> on 2021-12-02
Reviewed by Antti Koivisto.

Fix a crash for ports that use offscreen canvas WebGL.
Currently we do not have code to navigate from offscreen canvas
CanvasBase to ChromeClient. For this case it doesn't matter, as
GPUP does not support offscreen canvas in meaningful way
anyway.

No new tests, tested by existing test.

• html/canvas/WebGLRenderingContextBase.cpp:

(WebCore::WebGLRenderingContextBase::create):

4:51 AM Changeset in webkit [286421] by svillar@igalia.com

• 3 edits in trunk/Source/WebCore

Reset height definiteness when constructing flex items
​https://bugs.webkit.org/show_bug.cgi?id=233458

Reviewed by Alan Bujtas.

Flexbox renderer uses an attribute called m_hasDefiniteHeight in order to properly
consider as definite some sizes that are normally considered as indefinite. For example,
a percentage of an intrinsic size is normally indefinite unless the containing block has
an overriding size in which case can be considered as definite.

The value of m_hasDefiniteHeight is properly reset every time we construct a flex item or
just after finishing the layout. However we should do it as well just before computing
the flex basis if the flex item is also a flex container. In that case it might happen that
we try to compute the flex basis value with a value for m_hasDefiniteHeight which can
be potentially outdated.

• rendering/RenderFlexibleBox.cpp:

(WebCore::RenderFlexibleBox::layoutBlock): Use resetHasDefiniteHeight().
(WebCore::RenderFlexibleBox::layoutFlexItems): Ditto.
(WebCore::RenderFlexibleBox::constructFlexItem): Call resetHasDefiniteHeight() if the
flex item is also a flex container.

• rendering/RenderFlexibleBox.h:

(WebCore::RenderFlexibleBox::resetHasDefiniteHeight): Added.

1:02 AM Changeset in webkit [286420] by beidson@apple.com

• 4 edits in trunk/Source

AVDelegatingPlaybackCoordinator error compiling on macOS 11
​https://bugs.webkit.org/show_bug.cgi?id=232454

Reviewed by Tim Horton.

Source/WebCore/PAL:

• pal/cocoa/AVFoundationSoftLink.h: Don't softlink on platforms that don't have it (No other source changes were necessary)

Source/WTF:

• wtf/PlatformHave.h:

12:57 AM Changeset in webkit [286419] by youenn@apple.com

• 58 edits
  5 copies
  13 adds in trunk

Add support for NavigationPreloadManager
​https://bugs.webkit.org/show_bug.cgi?id=233523

Reviewed by Chris Dumez.

LayoutTests/imported/w3c:

• web-platform-tests/service-workers/idlharness.https.any-expected.txt:
• web-platform-tests/service-workers/idlharness.https.any.worker-expected.txt:
• web-platform-tests/service-workers/service-worker/navigation-preload/broken-chunked-encoding.https-expected.txt: Added.
• web-platform-tests/service-workers/service-worker/navigation-preload/chunked-encoding.https-expected.txt: Added.
• web-platform-tests/service-workers/service-worker/navigation-preload/empty-preload-response-body.https-expected.txt: Added.
• web-platform-tests/service-workers/service-worker/navigation-preload/get-state.https-expected.txt: Added.
• web-platform-tests/service-workers/service-worker/navigation-preload/navigationPreload.https-expected.txt: Added.
• web-platform-tests/service-workers/service-worker/navigation-preload/redirect.https-expected.txt: Added.
• web-platform-tests/service-workers/service-worker/navigation-preload/request-headers.https-expected.txt: Added.
• web-platform-tests/service-workers/service-worker/navigation-preload/resource-timing.https-expected.txt: Added.
• web-platform-tests/service-workers/service-worker/navigation-preload/samesite-cookies.https-expected.txt: Added.

Source/WebCore:

Add service worker navigation preload IDL and implementation.
This includes in particular state management (read and write) on API side and SWServer side.
This state is used in WK2 network code path to enable or not preloads.
We are not yet persisting preload registration state on disk, this will be done as a follow-up.

The FetchEvent preloadResponse returns a promise which has either undefined or a response as resolved value.
In case there is a preloader for that FetchEvent, we start the load when preloadResponse is retrieved by JS.

To implement the retrieval of the preload response, we use FetchResponse::fetch.
To do so, we add an initiator parameter as this preload initiator is the navigation task and not a fetch call.

Covered by enabled tests.

Tests: http/wpt/service-workers/service-worker-iframe-preload.https.html.

• CMakeLists.txt:
• DerivedSources-input.xcfilelist:
• DerivedSources-output.xcfilelist:
• DerivedSources.make:
• Modules/cache/DOMCache.cpp:
• Modules/fetch/FetchLoader.cpp:
• Modules/fetch/FetchLoader.h:
• Modules/fetch/FetchResponse.cpp:
• Modules/fetch/FetchResponse.h:
• Modules/fetch/WindowOrWorkerGlobalScopeFetch.cpp:
• Sources.txt:
• WebCore.xcodeproj/project.pbxproj:
• bindings/js/DOMPromiseProxy.h:
• bindings/js/WebCoreBuiltinNames.h:
• loader/cache/CachedResource.cpp:
• loader/cache/CachedResourceRequestInitiators.cpp:

(WebCore::CachedResourceRequestInitiators::CachedResourceRequestInitiators):

• loader/cache/CachedResourceRequestInitiators.h:
• platform/network/HTTPHeaderNames.in:
• workers/service/FetchEvent.cpp:
• workers/service/FetchEvent.h:
• workers/service/FetchEvent.idl:
• workers/service/NavigationPreloadManager.cpp: Added.
• workers/service/NavigationPreloadManager.h: Added.
• workers/service/NavigationPreloadManager.idl: Added.
• workers/service/NavigationPreloadState.h: Added.
• workers/service/NavigationPreloadState.idl: Added.
• workers/service/SWClientConnection.h:
• workers/service/ServiceWorkerContainer.cpp:
• workers/service/ServiceWorkerContainer.h:
• workers/service/ServiceWorkerRegistration.cpp:
• workers/service/ServiceWorkerRegistration.h:
• workers/service/ServiceWorkerRegistration.idl:
• workers/service/WorkerSWClientConnection.cpp:
• workers/service/WorkerSWClientConnection.h:
• workers/service/context/ServiceWorkerFetch.cpp:
• workers/service/context/ServiceWorkerFetch.h:
• workers/service/context/ServiceWorkerThread.cpp:
• workers/service/context/ServiceWorkerThread.h:
• workers/service/context/ServiceWorkerThreadProxy.cpp:
• workers/service/context/ServiceWorkerThreadProxy.h:
• workers/service/server/SWServerRegistration.cpp:
• workers/service/server/SWServerRegistration.h:

Source/WebKit:

Make use of ServiceWorkerNavigationPreloader when preload flag is true.
We only start as soon as created the preload in case of main frame navigation when preload flag is true.
Update ServiceWorkerNavigationPreloader to differentiate explicit preloads (enabled by NavigationPreloadManager) from implict preloads.
Implict preloads can be used when fetch event is not handled, which is not guaranteed for explicit preloads, given the Service-Worker-Navigation-Preload header.
Add IPC to read and set preload state (enabled and header value).

• NetworkProcess/ServiceWorker/ServiceWorkerFetchTask.cpp:
• NetworkProcess/ServiceWorker/ServiceWorkerNavigationPreloader.cpp:
• NetworkProcess/ServiceWorker/ServiceWorkerNavigationPreloader.h:
• NetworkProcess/ServiceWorker/WebSWServerConnection.cpp:
• NetworkProcess/ServiceWorker/WebSWServerConnection.h:
• NetworkProcess/ServiceWorker/WebSWServerConnection.messages.in:
• WebProcess/Storage/WebSWClientConnection.cpp:
• WebProcess/Storage/WebSWClientConnection.h:
• WebProcess/Storage/WebSWContextManagerConnection.cpp:
• WebProcess/Storage/WebSWContextManagerConnection.h:
• WebProcess/Storage/WebSWContextManagerConnection.messages.in:

Source/WTF:

• Scripts/Preferences/WebPreferencesExperimental.yaml:

LayoutTests:

• TestExpectations:

12:37 AM Changeset in webkit [286418] by Carlos Garcia Campos

• 2 edits in trunk/Tools

Unreviewed. [GTK] Use nullptr instead of 0 in JSStringCreateWithCharacters()

• WebKitTestRunner/InjectedBundle/atspi/AccessibilityUIElementAtspi.cpp:

(WTR::AccessibilityUIElement::attributesOfLinkedUIElements):
(WTR::AccessibilityUIElement::attributesOfDocumentLinks):
(WTR::AccessibilityUIElement::stringDescriptionOfAttributeValue):
(WTR::AccessibilityUIElement::stringAttributeValue):
(WTR::AccessibilityUIElement::parameterizedAttributeNames):
(WTR::AccessibilityUIElement::role):
(WTR::AccessibilityUIElement::subrole):
(WTR::AccessibilityUIElement::computedRoleString):
(WTR::AccessibilityUIElement::stringValue):
(WTR::AccessibilityUIElement::language):
(WTR::AccessibilityUIElement::helpText const):
(WTR::AccessibilityUIElement::speakAs):
(WTR::AccessibilityUIElement::ariaDropEffects const):
(WTR::AccessibilityUIElement::rangeForLine):
(WTR::AccessibilityUIElement::rangeForPosition):
(WTR::AccessibilityUIElement::boundsForRange):
(WTR::AccessibilityUIElement::stringForRange):
(WTR::AccessibilityUIElement::attributedStringForRange):
(WTR::AccessibilityUIElement::attributesOfColumnHeaders):
(WTR::AccessibilityUIElement::attributesOfRowHeaders):
(WTR::AccessibilityUIElement::attributesOfColumns):
(WTR::AccessibilityUIElement::attributesOfRows):
(WTR::AccessibilityUIElement::attributesOfVisibleCells):
(WTR::AccessibilityUIElement::attributesOfHeader):
(WTR::AccessibilityUIElement::rowIndexRange):
(WTR::AccessibilityUIElement::columnIndexRange):
(WTR::AccessibilityUIElement::selectedTextRange):
(WTR::AccessibilityUIElement::accessibilityValue const):
(WTR::AccessibilityUIElement::documentEncoding):
(WTR::AccessibilityUIElement::documentURI):
(WTR::AccessibilityUIElement::url):
(WTR::AccessibilityUIElement::stringForTextMarkerRange):
(WTR::AccessibilityUIElement::rectsForTextMarkerRange):
(WTR::AccessibilityUIElement::characterAtOffset):

12:18 AM Changeset in webkit [286417] by Martin Robinson

• 5 edits
  8 adds in trunk

sticky th or td in table does not stops at specified top
​https://bugs.webkit.org/show_bug.cgi?id=232431
<rdar://problem/85011818>

Reviewed by Simon Fraser.

LayoutTests/imported/w3c:

Add new WPT tests that test the behavior of sticky <td> elements in the middle of
tables.

• web-platform-tests/css/css-position/sticky/position-sticky-table-td-bottom-expected.html: Added.
• web-platform-tests/css/css-position/sticky/position-sticky-table-td-bottom.html: Added.
• web-platform-tests/css/css-position/sticky/position-sticky-table-td-left-expected.html: Added.
• web-platform-tests/css/css-position/sticky/position-sticky-table-td-left.html: Added.
• web-platform-tests/css/css-position/sticky/position-sticky-table-td-right-expected.html: Added.
• web-platform-tests/css/css-position/sticky/position-sticky-table-td-right.html: Added.
• web-platform-tests/css/css-position/sticky/position-sticky-table-td-top-expected.html: Added.
• web-platform-tests/css/css-position/sticky/position-sticky-table-td-top.html: Added.

Source/WebCore:

Tests: imported/w3c/web-platform-tests/css/css-position/sticky/position-sticky-table-td-bottom.html

imported/w3c/web-platform-tests/css/css-position/sticky/position-sticky-table-td-left.html
imported/w3c/web-platform-tests/css/css-position/sticky/position-sticky-table-td-right.html
imported/w3c/web-platform-tests/css/css-position/sticky/position-sticky-table-td-top.html

• rendering/RenderBox.h: Make frameRectForStickyPositioning not final.
• rendering/RenderTableCell.cpp:

(WebCore::RenderTableCell::frameRectForStickyPositioning const): Added this implementation
that takes into account that RenderTableCell is positioned relatively to the RenderTableSection.

• rendering/RenderTableCell.h: Added method declaration.

12:16 AM Changeset in webkit [286416] by Carlos Garcia Campos

• 10 edits
  2 adds in trunk

[GTK][a11y] Add implementation of table and table cell interfaces when building with ATSPI
​https://bugs.webkit.org/show_bug.cgi?id=233694

Reviewed by Adrian Perez de Castro.

Source/WebCore:

• SourcesGTK.txt:
• accessibility/AccessibilityTableColumn.cpp:

(WebCore::AccessibilityTableColumn::computeAccessibilityIsIgnored const): Do not expose column for ATSPI.

• accessibility/AccessibilityTableHeaderContainer.cpp:

(WebCore::AccessibilityTableHeaderContainer::computeAccessibilityIsIgnored const): Do not expose header container either.

• accessibility/atspi/AccessibilityObjectAtspi.cpp:

(WebCore::AccessibilityObjectAtspi::interfacesForObject):
(WebCore::AccessibilityObjectAtspi::path):
(WebCore::AccessibilityObjectAtspi::wrapperVector const):
(WebCore::AccessibilityObjectAtspi::children const):
(WebCore::AccessibilityObjectAtspi::buildInterfaces const):

• accessibility/atspi/AccessibilityObjectAtspi.h:
• accessibility/atspi/AccessibilityObjectTableAtspi.cpp: Added.

(WebCore::AccessibilityObjectAtspi::rowCount const):
(WebCore::AccessibilityObjectAtspi::columnCount const):
(WebCore::AccessibilityObjectAtspi::cell const):
(WebCore::AccessibilityObjectAtspi::tableCaption const):
(WebCore::AccessibilityObjectAtspi::cellIndex const):
(WebCore::AccessibilityObjectAtspi::rowAtIndex const):
(WebCore::AccessibilityObjectAtspi::columnAtIndex const):
(WebCore::AccessibilityObjectAtspi::rowHeader const):
(WebCore::AccessibilityObjectAtspi::columnHeader const):
(WebCore::AccessibilityObjectAtspi::rowDescription const):
(WebCore::AccessibilityObjectAtspi::columnDescription const):
(WebCore::AccessibilityObjectAtspi::rowExtent const):
(WebCore::AccessibilityObjectAtspi::columnExtent const):
(WebCore::AccessibilityObjectAtspi::cells const):
(WebCore::AccessibilityObjectAtspi::rows const):
(WebCore::AccessibilityObjectAtspi::rowHeaders const):
(WebCore::AccessibilityObjectAtspi::columnHeaders const):

• accessibility/atspi/AccessibilityObjectTableCellAtspi.cpp: Added.

(WebCore::AccessibilityObjectAtspi::cellRowHeaders const):
(WebCore::AccessibilityObjectAtspi::cellColumnHeaders const):
(WebCore::AccessibilityObjectAtspi::rowSpan const):
(WebCore::AccessibilityObjectAtspi::columnSpan const):
(WebCore::AccessibilityObjectAtspi::cellPosition const):

• accessibility/atspi/xml/TableCell.xml:

Tools:

Add unit tests and WTR implementation for tables.

• TestWebKitAPI/Tests/WebKitGtk/TestWebKitAccessibility.cpp:

(testTableBasic):
(beforeAll):

• WebKitTestRunner/InjectedBundle/atspi/AccessibilityUIElementAtspi.cpp:

(WTR::AccessibilityUIElement::rowAtIndex):
(WTR::elementsVector):
(WTR::attributesOfElements):
(WTR::AccessibilityUIElement::attributesOfChildren):
(WTR::AccessibilityUIElement::numberAttributeValue):
(WTR::makeJSArray):
(WTR::AccessibilityUIElement::rowHeaders const):
(WTR::AccessibilityUIElement::columnHeaders const):
(WTR::AccessibilityUIElement::attributesOfColumnHeaders):
(WTR::AccessibilityUIElement::attributesOfRowHeaders):
(WTR::AccessibilityUIElement::attributesOfRows):
(WTR::AccessibilityUIElement::attributesOfVisibleCells):
(WTR::AccessibilityUIElement::rowCount):
(WTR::AccessibilityUIElement::columnCount):
(WTR::AccessibilityUIElement::rowIndexRange):
(WTR::AccessibilityUIElement::columnIndexRange):
(WTR::AccessibilityUIElement::cellForColumnAndRow):

12:06 AM Changeset in webkit [286415] by Manuel Rego Casasnovas

• 4 edits
  2 adds in trunk

[selectors] :focus-visible should stop matching after blur
​https://bugs.webkit.org/show_bug.cgi?id=233688

Reviewed by Antti Koivisto.

LayoutTests/imported/w3c:

Add new test to check that :focus-visible stop matching after blur.

• web-platform-tests/css/selectors/focus-visible-023-expected.txt: Added.
• web-platform-tests/css/selectors/focus-visible-023.html: Added.

Source/WebCore:

Move PseudoClassChangeInvalidation for :focus-visible together with the one for :focus,
as we need to keep both flags in sync when computing the invalidations
(it's not possible to have an element that matches :focus-visible but not :focus,
and that was happening before this change).

Test: imported/w3c/web-platform-tests/css/selectors/focus-visible-023.html

• dom/Element.cpp:

(WebCore::Element::setFocus): Add PseudoClassChangeInvalidation for :focus-visible change,
and put it in the same scope as the flags changes.
(WebCore::Element::setHasFocusVisible): Remove PseudoClassChangeInvalidation from this method.

• page/EventHandler.cpp:

(WebCore::EventHandler::internalKeyEvent): Use PseudoClassChangeInvalidation for :focus-visible change.