Commits · WebKit/WebKit

https://bugs.webkit.org/show_bug.cgi?id=251259 Reviewed by Kimmo Kinnunen. * LayoutTests/platform/ios-simulator/webgl/webgl-draft-extensions-flag-default-expected.txt: * LayoutTests/platform/ios-simulator/webgl/webgl-draft-extensions-flag-off-expected.txt: * LayoutTests/platform/ios-simulator/webgl/webgl-draft-extensions-flag-on-expected.txt: * LayoutTests/webgl/resources/webgl-draft-extensions-flag.js: * LayoutTests/webgl/webgl-draft-extensions-flag-default-expected.txt: * LayoutTests/webgl/webgl-draft-extensions-flag-off-expected.txt: * LayoutTests/webgl/webgl-draft-extensions-flag-on-expected.txt: * Source/WebCore/html/canvas/WebGL2RenderingContext.cpp: (WebCore::WebGL2RenderingContext::getExtension): (WebCore::WebGL2RenderingContext::getSupportedExtensions): Canonical link: https://commits.webkit.org/259499@main

https://bugs.webkit.org/show_bug.cgi?id=251077 Reviewed by Chris Dumez. Replaced the use of raw pointers to RenderSVGResourceContainer by WeakPtr. * Source/WebCore/rendering/svg/LegacyRenderSVGRoot.cpp: (WebCore::LegacyRenderSVGRoot::layout): (WebCore::LegacyRenderSVGRoot::addResourceForClientInvalidation): * Source/WebCore/rendering/svg/LegacyRenderSVGRoot.h: * Source/WebCore/rendering/svg/RenderSVGRoot.cpp: (WebCore::RenderSVGRoot::layoutChildren): * Source/WebCore/rendering/svg/RenderSVGRoot.h: * Source/WebCore/rendering/svg/SVGResources.cpp: (WebCore::SVGResources::removeClientFromCache const): (WebCore::SVGResources::resourceDestroyed): (WebCore::SVGResources::buildSetOfResources): (WebCore::SVGResources::dump): * Source/WebCore/rendering/svg/SVGResources.h: (WebCore::SVGResources::clipper const): (WebCore::SVGResources::markerStart const): (WebCore::SVGResources::markerMid const): (WebCore::SVGResources::markerEnd const): (WebCore::SVGResources::masker const): (WebCore::SVGResources::filter const): (WebCore::SVGResources::fill const): (WebCore::SVGResources::stroke const): (WebCore::SVGResources::linkedResource const): (): Deleted. * Source/WebCore/rendering/svg/SVGResourcesCache.cpp: (WebCore::SVGResourcesCache::addResourcesFromRenderer): (WebCore::SVGResourcesCache::removeResourcesFromRenderer): * Source/WebCore/rendering/svg/SVGResourcesCycleSolver.cpp: (WebCore::SVGResourcesCycleSolver::resourceContainsCycles const): (WebCore::SVGResourcesCycleSolver::resolveCycles): * Source/WebCore/rendering/svg/SVGResourcesCycleSolver.h: Canonical link: https://commits.webkit.org/259498@main

Fix negative shadow repaint issue https://bugs.webkit.org/show_bug.cgi?id=251176 Reviewed by Simon Fraser. Merge - https://src.chromium.org/viewvc/blink?view=revision&revision=164709 When a box with negative v-shadow changes height, the repaint rect is inadequate resulting some part of shadow is not repainted. Fix the issue by considering negative shadow-top/shadow-left. * Source/WebCore/rendering/RenderElement.cpp: (RenderElement::repaintAfterLayout): Add consideration for negative shadows * LayoutTests/fast/box-shadow/negative-shadow-box-shrink.html: Add Test Case * LayoutTests/fast/box-shadow/negative-shadow-box-shrink-expected.txt: Add Test Case Expectation * LayoutTests/fast/box-shadow/negative-shadow-box-expand.html: Add Test Case * LayoutTests/fast/box-shadow/negative-shadow-box-shrink-expand.txt: Add Test Case Expectation Canonical link: https://commits.webkit.org/259497@main

…n to SecurityOriginData https://bugs.webkit.org/show_bug.cgi?id=251048 rdar://104578586 Unreviewed, revert 259432@main as it regressed ProcessSwap.GetUserMediaCaptureState API test. * Source/WTF/wtf/Markable.h: (WTF::add): Deleted. * Source/WebCore/page/SecurityOrigin.cpp: (WebCore::SecurityOrigin::SecurityOrigin): (WebCore::SecurityOrigin::isSameOriginDomain const): (WebCore::SecurityOrigin::isSameOriginAs const): (WebCore::SecurityOrigin::create): (WebCore::SecurityOrigin::equal const): * Source/WebCore/page/SecurityOrigin.h: (WebCore::SecurityOrigin::isOpaque const): (WebCore::add): * Source/WebCore/page/SecurityOriginData.cpp: (WebCore::SecurityOriginData::securityOrigin const): (WebCore::SecurityOriginData::isolatedCopy const): (WebCore::SecurityOriginData::isolatedCopy): (WebCore::operator==): * Source/WebCore/page/SecurityOriginData.h: (WebCore::SecurityOriginData::SecurityOriginData): (WebCore::SecurityOriginData::isOpaque const): (WebCore::add): (): Deleted. (WebCore::SecurityOriginData::createOpaque): Deleted. (WebCore::SecurityOriginData::encode const): Deleted. (WebCore::SecurityOriginData::decode): Deleted. * Source/WebKit/Shared/WebCoreArgumentCoders.serialization.in: Canonical link: https://commits.webkit.org/259496@main

…nfigurationCallBack() https://bugs.webkit.org/show_bug.cgi?id=251233 rdar://104719183 Reviewed by Tim Horton. WebProcessPool had both screenPropertiesStateChanged() and code in displayReconfigurationCallBack() that did similar things, sending new screen properties to the WebProcess and, on macOS, the GPU process, but the code was scattered around. Clean it up by moving the code in displayReconfigurationCallBack() into a WebProcessPool member function; this callback tells you about a specific display and what changed, but in it we gather properties for all displays, but keep that behavior for now. screenPropertiesStateChanged() is called via accessibility callbacks that indicate that screen-related settings like "invert colors" changed. To show its relationship to WebProcessPool::displayPropertiesChanged() put the two functions next to each other, but it was tricky to share code because this callback doesn't get a displayID. Add a DisplayLink::displayPropertiesChanged() stub that will be implemented in a future patch. * Source/WebKit/UIProcess/Cocoa/WebProcessPoolCocoa.mm: (WebKit::WebProcessPool::registerNotificationObservers): (WebKit::WebProcessPool::screenPropertiesChanged): (WebKit::WebProcessPool::displayPropertiesChanged): (WebKit::displayReconfigurationCallBack): * Source/WebKit/UIProcess/WebPageProxy.cpp: (WebKit::WebPageProxy::accessibilitySettingsDidChange): * Source/WebKit/UIProcess/WebProcessPool.cpp: (WebKit::WebProcessPool::screenPropertiesStateChanged): Deleted. * Source/WebKit/UIProcess/WebProcessPool.h: * Source/WebKit/UIProcess/mac/DisplayLink.cpp: (WebKit::DisplayLink::displayPropertiesChanged): * Source/WebKit/UIProcess/mac/DisplayLink.h: * Source/WebKit/UIProcess/mac/WebViewImpl.mm: (WebKit::WebViewImpl::screenDidChangeColorSpace): Canonical link: https://commits.webkit.org/259495@main

…s/dir-pseudo-on-input-element.html https://bugs.webkit.org/show_bug.cgi?id=251275 rdar://104748241 Unreviewed test gardening. This test was re-named as part of the re-sync in 258889@main, but the mac-wk1 baseline wasn't updated to match. * LayoutTests/platform/mac-wk1/imported/w3c/web-platform-tests/css/selectors/dir-pseudo-on-input-element-expected.txt: Renamed from LayoutTests/platform/mac-wk1/imported/w3c/web-platform-tests/css/css-pseudo/dir-pseudo-on-input-element-expected.txt. Canonical link: https://commits.webkit.org/259494@main

https://bugs.webkit.org/show_bug.cgi?id=251222 rdar://104710968 Reviewed by Jonathan Bedard. To skip obsolete builds and PRs with skip-ews more frequently. * Tools/CISupport/ews-build/factories.py: (Factory.__init__): * Tools/CISupport/ews-build/factories_unittest.py: (TestExpectedBuildSteps): Canonical link: https://commits.webkit.org/259493@main

…c/web-platform-tests/fs/FileSystemFileHandle-create-sync-access-handle.https.tentative.window.html is a consistent failure webkit.org/b/251274 Unreviewed test gardening. * LayoutTests/platform/ios/TestExpectations: Canonical link: https://commits.webkit.org/259492@main

https://bugs.webkit.org/show_bug.cgi?id=251244 <rdar://problem/104725544> Reviewed by Yusuke Suzuki. With FPAC, the aut instruction will validate its own result. Hence, the manual validation that we used to do after the aut is now redundant. * Source/JavaScriptCore/assembler/CPU.cpp: (JSC::isARM64E_FPAC): * Source/JavaScriptCore/assembler/CPU.h: (JSC::isARM64E_FPAC): (JSC::isARM64_LSE): * Source/JavaScriptCore/assembler/MacroAssemblerARM64E.h: (JSC::MacroAssemblerARM64E::validateUntaggedPtr): (JSC::MacroAssemblerARM64E::untagArrayPtr): (JSC::MacroAssemblerARM64E::untagArrayPtrLength64): * Source/JavaScriptCore/runtime/InitializeThreading.cpp: (JSC::initialize): * Source/JavaScriptCore/runtime/JSCConfig.h: Canonical link: https://commits.webkit.org/259491@main

…sOrPointer. https://bugs.webkit.org/show_bug.cgi?id=251251 <rdar://problem/104731974> Reviewed by Yusuke Suzuki. The rest of FixedBitVector's code relies on this relationship. So, let's make it explicit. Also, fixed a bug in BitMap's API tests: testBitmapConcurrentTestAndSet() and testBitmapConcurrentTestAndClear() were exercising testAndSet() and testAndClear() respectively instead of their concurrent versions. This is now fixed. This issue was originally found by Yijia Huang in https://bugs.webkit.org/show_bug.cgi?id=250847. * Source/WTF/wtf/FixedBitVector.h: * Tools/TestWebKitAPI/Tests/WTF/Bitmap.cpp: (TestWebKitAPI::testBitmapConcurrentTestAndSet): (TestWebKitAPI::testBitmapConcurrentTestAndClear): Canonical link: https://commits.webkit.org/259490@main

https://bugs.webkit.org/show_bug.cgi?id=251252 rdar://problem/104733721 Reviewed by Chris Dumez. Use cced25519_make_pub to generate the public key. * LayoutTests/imported/w3c/web-platform-tests/WebCryptoAPI/import_export/okp_importKey.https.any-expected.txt: * LayoutTests/imported/w3c/web-platform-tests/WebCryptoAPI/import_export/okp_importKey.https.any.worker-expected.txt: * LayoutTests/imported/w3c/web-platform-tests/WebCryptoAPI/wrapKey_unwrapKey/wrapKey_unwrapKey.https.any-expected.txt: * LayoutTests/imported/w3c/web-platform-tests/WebCryptoAPI/wrapKey_unwrapKey/wrapKey_unwrapKey.https.any.worker-expected.txt: * Source/WebCore/crypto/mac/CryptoKeyOKPCocoa.cpp: (WebCore::CryptoKeyOKP::generateJwkX const): Canonical link: https://commits.webkit.org/259489@main

…specializations https://bugs.webkit.org/show_bug.cgi?id=250777 Reviewed by Kimmo Kinnunen. Instead of finding better or worse fixed-sized-type matches for size values in different ArgumentCoder specializations, the types of those size values should be used. This means simply encoding values of size_t or unsigned types. The target type of the size value should be noted in the code, just to provide the necessary context of what is being encoded and have it match the decoding side. This means the size type is explicitly used either in any temporary variable where the value is stored, or an additional static_cast is used to explicitly specify the type when the value is passed to the stream insertion operator. This still leaves open the possibility of using mismatched types, i.e. the size value type not matching what the encoding enforces. Some sort of validation could be implemented in the encoder class to make sure no undesired narrowing conversions occur. * Source/WebKit/Platform/IPC/ArgumentCoders.cpp: (IPC::ArgumentCoder<CString>::encode): (IPC::ArgumentCoder<CString>::decode): (IPC::ArgumentCoder<String>::encode): (IPC::decodeStringText): (IPC::ArgumentCoder<String>::decode): (IPC::ArgumentCoder<StringView>::encode): * Source/WebKit/Platform/IPC/ArgumentCoders.h: Canonical link: https://commits.webkit.org/259488@main

…only test https://bugs.webkit.org/show_bug.cgi?id=251253 rdar://104698580 Unreviewed test gardening. Skipping playform-specific directory. * LayoutTests/platform/ios/TestExpectations: Canonical link: https://commits.webkit.org/259487@main

https://bugs.webkit.org/show_bug.cgi?id=251250 Reviewed by Yusuke Suzuki. Noticed incorrect spelling while browsing the code today. Update errorMesasgeForTransfer to errorMessageForTransfer. * Source/JavaScriptCore/runtime/ArrayBuffer.cpp: (JSC::errorMessageForTransfer): (JSC::errorMesasgeForTransfer): Deleted. * Source/JavaScriptCore/runtime/ArrayBuffer.h: * Source/WebCore/bindings/js/SerializedScriptValue.cpp: (WebCore::SerializedScriptValue::create): Canonical link: https://commits.webkit.org/259486@main

https://bugs.webkit.org/show_bug.cgi?id=251227 rdar://104203486 Reviewed by Wenson Hsieh. This seems to have regressed with 255459@main, which updated the logic to restore portrait orientation between tests when we added support for the screen orientation API. There are 2 ways to restore the orientation: 1. Call [UIScene requestGeometryUpdateWithPreferences:] 2. Lock the orientation to portrait then unlock In 255459@main, I had replaced method 1 with method 2 because method 1 didn't work properly with the screen orientation lock tests. However, it broke the tests in fast/events/iOS/rotation which rely on `[UIScene requestGeometryUpdateWithPreferences:]` to simulate screen rotation during the tests. After several hours of trying to find a common mechanism to restore orientation that would work for both screen rotation and orientation lock tests, I still wasn't able to make it work. For this reason, in my this, I am using method 2 only for tests that lock the screen orientation and keep using the old method 1 for all other tests. As a result, this restores pre-255459@main behavior and fixes existing tests. It also doesn't regress the screen orientation lock tests which were added in 255459@main. * Tools/WebKitTestRunner/TestController.h: * Tools/WebKitTestRunner/ios/TestControllerIOS.mm: (-[WindowDidRotateObserver initWithCallback:]): (-[WindowDidRotateObserver dealloc]): (-[WindowDidRotateObserver _windowDidRotate]): (WTR::TestController::restorePortraitOrientationIfNeeded): (WTR::TestController::lockScreenOrientation): Canonical link: https://commits.webkit.org/259485@main

https://bugs.webkit.org/show_bug.cgi?id=251045 rdar://103361530 Reviewed by Chris Fleizach. Covered by existing tests. The crash happens in ITM because AXObjectCache::remove updates the isolated tree by calling AXIsolatedTree::removeNode, that calls parentObjectUnignored(), which results in a call to textUnderElement which cannot be called during a layout. The solution in this patch is to make the removal of the object in question asynchronously. Some code cleanup by splitting the handling of an element's renderer creation from AXObjectCache::get(Node*). * Source/WebCore/accessibility/AXObjectCache.cpp: (WebCore::AXObjectCache::get): (WebCore::AXObjectCache::onRendererCreated): (WebCore::AXObjectCache::handleLiveRegionCreated): (WebCore::AXObjectCache::performDeferredCacheUpdate): (WebCore::AXObjectCache::updateCacheAfterNodeIsAttached): Deleted. * Source/WebCore/accessibility/AXObjectCache.h: (WebCore::AXObjectCache::onRendererCreated): (WebCore::AXObjectCache::updateCacheAfterNodeIsAttached): Deleted. * Source/WebCore/rendering/updating/RenderTreeUpdater.cpp: (WebCore::RenderTreeUpdater::createRenderer): Canonical link: https://commits.webkit.org/259484@main

* Source/WTF/wtf/glib/GWeakPtr.h: (WTF::GWeakPtr::operator=): Add missing std:: Canonical link: https://commits.webkit.org/259483@main

…t_add|remove_weak_pointer https://bugs.webkit.org/show_bug.cgi?id=251258 Reviewed by Michael Catanzaro and Adrian Perez de Castro. It's very common to forget to remove a weak references, GWeakPtr ensures weak references are always removed and makes easier to create weak GObjects. * Source/WTF/wtf/PlatformGTK.cmake: * Source/WTF/wtf/PlatformWPE.cmake: * Source/WTF/wtf/glib/GWeakPtr.h: Added. (WTF::GWeakPtr::GWeakPtr): (WTF::GWeakPtr::~GWeakPtr): (WTF::GWeakPtr::get const): (WTF::GWeakPtr::reset): * Source/WebKit/UIProcess/API/glib/WebKitDownload.cpp: (webkitDownloadCreate): (webkit_download_get_web_view): (_WebKitDownloadPrivate::~_WebKitDownloadPrivate): Deleted. * Source/WebKit/UIProcess/API/glib/WebKitFaviconDatabase.cpp: (webkitFaviconDatabaseGetLoadDecisionForIcon): * Source/WebKit/UIProcess/API/glib/WebKitIconLoadingClient.cpp: * Source/WebKit/UIProcess/API/glib/WebKitUIClient.cpp: (UIClient::~UIClient): Deleted. * Tools/TestWebKitAPI/PlatformGTK.cmake: * Tools/TestWebKitAPI/PlatformWPE.cmake: * Tools/TestWebKitAPI/Tests/WTF/glib/GWeakPtr.cpp: Added. (TestWebKitAPI::TEST): Canonical link: https://commits.webkit.org/259482@main

…ped because of skip-ews label (follow-up) https://bugs.webkit.org/show_bug.cgi?id=251221 <rdar://problem/104708261> Unreviewed follow-up fix. * Tools/CISupport/ews-app/ews/common/github.py: (GitHubEWS.github_status_for_queue): Canonical link: https://commits.webkit.org/259481@main

…ould never return NULL for base cache and data directories https://bugs.webkit.org/show_bug.cgi?id=251255 Reviewed by Adrian Perez de Castro. When NULL is passed to network session constructor the default directory is provided to website data manager to ensure the getters don't return NULL. * Source/WebKit/UIProcess/API/glib/WebKitNetworkSession.cpp: (webkitNetworkSessionSetProperty): (webkitNetworkSessionConstructed): * Source/WebKit/UIProcess/API/glib/WebKitWebsiteDataManager.cpp: (webkitWebsiteDataManagerSetProperty): (webkitWebsiteDataManagerConstructed): (webkitWebsiteDataManagerGetDataStore): (webkitWebsiteDataManagerCreate): (webkit_website_data_manager_get_base_data_directory): (webkit_website_data_manager_get_base_cache_directory): (webkitWebsiteDataManagerGetFaviconDatabasePath): * Source/WebKit/UIProcess/API/glib/WebKitWebsiteDataManagerPrivate.h: * Source/WebKit/UIProcess/WebsiteData/WebsiteDataStore.h: * Source/WebKit/UIProcess/glib/WebsiteDataStoreGLib.cpp: (WebKit::WebsiteDataStore::defaultBaseCacheDirectory): (WebKit::WebsiteDataStore::defaultBaseDataDirectory): (WebKit::WebsiteDataStore::cacheDirectoryFileSystemRepresentation): (WebKit::WebsiteDataStore::websiteDataDirectoryFileSystemRepresentation): * Tools/TestWebKitAPI/Tests/WebKitGLib/TestWebsiteData.cpp: (testWebsiteDataConfiguration): Canonical link: https://commits.webkit.org/259480@main

…building easily for a set of curated embedded targets with Yocto https://bugs.webkit.org/show_bug.cgi?id=249604 Reviewed by Philippe Normand. This adds a new script named 'cross-toolchain-helper' that automates the process of building a cross-toolchain and an image with Yocto. The script also is able to automatically deploy the Yocto workdir and set the target build configs from a set of configuration files. The main configuration file for this script is at the path 'Tools/yocto/targets.conf'. Inside this file several targets can be defined. Each target defines a set of Yocto layers and local.conf configurations. It can be also defined the default WebKit parameters for each target. The following WebKit tool scripts are also patched to be aware of the cross-target builds: - Tools/Scripts/update-webkitwpe-libs - Tools/Scripts/update-webkitgtk-libs - Tools/Scripts/build-webkit - Tools/CISupport/built-product-archive To enable the cross-target builds the flag '--cross-target=target_name' should be passed. Alternatively the environment variable 'WEBKIT_CROSS_TARGET' can be used instead. The list of targets defined at 'Tools/yocto/targets.conf' is curated and is expected to build and work without issues. In the near future a post-commit bot will be checking this. The tool is designed to automatically wipe and recreate the cross-toolchain and images when it detects the configuration for any target has changed. However, this behaviour can be overriden by optional parameters. For more info (including examples on how to use this) check the file: Tools/yocto/README.md * Tools/CISupport/built-product-archive: * Tools/Scripts/build-webkit: * Tools/Scripts/cross-toolchain-helper: Added. * Tools/Scripts/update-webkitgtk-libs: * Tools/Scripts/update-webkitwpe-libs: * Tools/Scripts/webkitdirs.pm: (determineCrossTarget): (determineConfigurationProductDir): (determinePassedConfiguration): (runInCrossTargetEnvironment): (inCrossTargetEnvironment): (getCrossTargetName): (shouldBuildForCrossTarget): (wrapperPrefixIfNeeded): (shouldUseFlatpak): (cmakeCachePath): (cmakeFilesPath): (shouldRemoveCMakeCache): (cmakeGeneratedBuildfile): (generateBuildSystemFromCMakeProject): (productDirForCMake): (buildCMakeGeneratedProject): (cleanCMakeGeneratedProject): * Tools/yocto/README.md: Added. * Tools/yocto/rpi/bblayers.conf: Added. * Tools/yocto/rpi/fix-nativesdk-image-defs.patch: Added. * Tools/yocto/rpi/local-rpi3-32bits-mesa.conf: Added. * Tools/yocto/rpi/local-rpi3-32bits-userland.conf: Added. * Tools/yocto/rpi/local-rpi3-64bits-mesa.conf: Added. * Tools/yocto/rpi/local-rpi4-32bits-mesa.conf: Added. * Tools/yocto/rpi/local-rpi4-64bits-mesa.conf: Added. * Tools/yocto/rpi/manifest.xml: Added. * Tools/yocto/targets.conf: Added. Canonical link: https://commits.webkit.org/259479@main

…nnected https://bugs.webkit.org/show_bug.cgi?id=231787 rdar://problem/84529041 Reviewed by Eric Carlson. We receive a remote pause command when BT is disconnected. We also get remote commands from keyboard and the current heuristic does not work well with video conference websites that have multiple media elements playing at the same time. We introduce a new heuristic in that case, where instead of pausing/playing media elements, we mute/unmute capture and audio rendering. This allow users for instance to restart capture/audio using Safari UI. We update WebPlaybackControlsManager setPlaying to always send an IPC message since calling playing may unmute WebProcess. Covered by API test. * Source/WebCore/dom/Document.cpp: (WebCore::Document::updateIsPlayingMedia): * Source/WebCore/dom/Document.h: (WebCore::Document::activeMediaElementsWithMediaStreamCount const): * Source/WebCore/html/MediaElementSession.cpp: (WebCore::isDocumentPlayingSeveralMediaStreams): (WebCore::processRemoteControlCommandIfPlayingMediaStreams): (WebCore::MediaElementSession::didReceiveRemoteControlCommand): (WebCore::MediaElementSession::nowPlayingInfo const): * Source/WebCore/platform/mac/WebPlaybackControlsManager.mm: (-[WebPlaybackControlsManager setPlaying:]): * Tools/TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * Tools/TestWebKitAPI/Tests/WebKit/GetUserMedia.mm: * Tools/TestWebKitAPI/Tests/WebKitCocoa/webrtc-remote.html: Added. Canonical link: https://commits.webkit.org/259478@main

https://bugs.webkit.org/show_bug.cgi?id=251205 <rdar://problem/104691861> Reviewed by Myles C. Maxfield. Add a new pass that removes global variables and creates structs compatible with Metal argument buffers. For now, the pass only generates the structs based on the globals available in the source program, but we should use the pipeline layout in the future. Fow now what it does is: 1. Find all the globals in the source and remove them 2. Creates a struct per bind group to receive the globals 3. Traverse the call graph and identify which functions use which globals 4. Insert parameters in the functions that need to receive the globals 5. Rewrite the global accesses with struct accesses to load the globals from the structs received via the parameters. In order for this patch to work correct, GlobalVariableRewriter needs to be updated to use a ReferenceType, which is being introduce in PR #9174. This PR also includes a fair amount of FIXMEs, but it was getting pretty large so I paused here and will upload follow up patches shortly. * Source/WebGPU/WGSL/AST/ASTFunctionDecl.h: * Source/WebGPU/WGSL/AST/ASTIdentifierExpression.h: * Source/WebGPU/WGSL/AST/ASTStructureDecl.h: * Source/WebGPU/WGSL/GlobalVariableRewriter.cpp: Added. (WGSL::RewriteGlobalVariables::RewriteGlobalVariables): (WGSL::RewriteGlobalVariables::run): (WGSL::RewriteGlobalVariables::visit): (WGSL::replace): (WGSL::RewriteGlobalVariables::collectGlobals): (WGSL::RewriteGlobalVariables::visitEntryPoint): (WGSL::RewriteGlobalVariables::requiredGroups): (WGSL::RewriteGlobalVariables::insertStructs): (WGSL::RewriteGlobalVariables::insertParameters): (WGSL::RewriteGlobalVariables::def): (WGSL::RewriteGlobalVariables::read): (WGSL::RewriteGlobalVariables::argumentBufferParameterName): (WGSL::RewriteGlobalVariables::argumentBufferStructName): (WGSL::rewriteGlobalVariables): * Source/WebGPU/WGSL/GlobalVariableRewriter.h: Added. * Source/WebGPU/WGSL/Metal/MetalFunctionWriter.cpp: (WGSL::Metal::FunctionDefinitionWriter::visit): (WGSL::Metal::FunctionDefinitionWriter::visitArgumentBufferParameter): * Source/WebGPU/WGSL/WGSL.cpp: (WGSL::prepare): * Source/WebGPU/WebGPU.xcodeproj/project.pbxproj: Canonical link: https://commits.webkit.org/259477@main

https://bugs.webkit.org/show_bug.cgi?id=251030 Reviewed by Kimmo Kinnunen. The 'WantsConnection' IPC message attribute imposed using the wants-connection variants of IPC::handleMessage(), passing the Connection reference as the first argument to the method handling a given IPC message. The attribute can be removed, along with the relevant IPC::handleMessage() variants. Instead, the initial Connection reference parameter on the handler method can be indicated on the MethodSignatureValidation template class that's already used for validation purposes. Existing variants do a constexpr check and adjust the method invocation to pass the Connection reference as the first argument to the method when necessary. * Source/WebKit/NetworkProcess/NetworkBroadcastChannelRegistry.messages.in: * Source/WebKit/NetworkProcess/storage/NetworkStorageManager.messages.in: * Source/WebKit/Platform/IPC/HandleMessage.h: (IPC::handleMessage): (IPC::handleMessageSynchronous): (IPC::handleMessageAsync): (IPC::handleMessageWantsConnection): Deleted. (IPC::handleMessageSynchronousWantsConnection): Deleted. (IPC::handleMessageAsyncWantsConnection): Deleted. * Source/WebKit/Scripts/webkit/messages.py: (async_message_statement): (sync_message_statement): * Source/WebKit/Scripts/webkit/model.py: * Source/WebKit/Scripts/webkit/tests/TestWithLegacyReceiver.messages.in: * Source/WebKit/Scripts/webkit/tests/TestWithLegacyReceiverMessageReceiver.cpp: (WebKit::TestWithLegacyReceiver::didReceiveSyncTestWithLegacyReceiverMessage): * Source/WebKit/Scripts/webkit/tests/TestWithSuperclass.messages.in: * Source/WebKit/Scripts/webkit/tests/TestWithSuperclassMessageReceiver.cpp: (WebKit::TestWithSuperclass::didReceiveMessage): * Source/WebKit/Scripts/webkit/tests/TestWithoutAttributes.messages.in: * Source/WebKit/Scripts/webkit/tests/TestWithoutAttributesMessageReceiver.cpp: (WebKit::TestWithoutAttributes::didReceiveSyncMessage): * Source/WebKit/Shared/IPCTester.messages.in: * Source/WebKit/Shared/Notifications/NotificationManagerMessageHandler.messages.in: * Source/WebKit/UIProcess/WebGeolocationManagerProxy.messages.in: * Source/WebKit/UIProcess/WebPageProxy.messages.in: * Source/WebKit/UIProcess/WebPasteboardProxy.messages.in: * Source/WebKit/UIProcess/WebProcessPool.messages.in: Canonical link: https://commits.webkit.org/259476@main

https://bugs.webkit.org/show_bug.cgi?id=251253 rdar://104698580 Unreviewed. Update Test Expectations to skip tests except in mac-wk2. * LayoutTests/TestExpectations: * LayoutTests/platform/mac-wk2/TestExpectations: Canonical link: https://commits.webkit.org/259475@main

https://bugs.webkit.org/show_bug.cgi?id=251202 <rdar://problem/104689979> Reviewed by Myles C. Maxfield. Remove the existing `TypeReference` node, which was a workaround to have a node that points to another type node, but it's now no longer necessary. Replace it with a `ReferenceType` node. The similar name is a bit confusing, but this actually refers to a reference to type, i.e. `T&` in source. The node still isn't used in this patch, but another patch will follow this one. * Source/WebGPU/WGSL/AST/ASTForward.h: * Source/WebGPU/WGSL/AST/ASTNode.h: * Source/WebGPU/WGSL/AST/ASTStringDumper.cpp: (WGSL::AST::StringDumper::visit): * Source/WebGPU/WGSL/AST/ASTStringDumper.h: * Source/WebGPU/WGSL/AST/ASTTypeDecl.h: (isType): * Source/WebGPU/WGSL/AST/ASTVisitor.cpp: (WGSL::AST::Visitor::visit): * Source/WebGPU/WGSL/AST/ASTVisitor.h: * Source/WebGPU/WGSL/EntryPointRewriter.cpp: (WGSL::EntryPointRewriter::getResolvedType): (WGSL::EntryPointRewriter::constructInputStruct): * Source/WebGPU/WGSL/Metal/MetalFunctionWriter.cpp: (WGSL::Metal::FunctionDefinitionWriter::visit): Canonical link: https://commits.webkit.org/259474@main

https://bugs.webkit.org/show_bug.cgi?id=251145 Reviewed by Xabier Rodriguez-Calvar. * Source/WebCore/Modules/mediastream/gstreamer/GStreamerMediaEndpoint.cpp: (WebCore::GStreamerMediaEndpoint::addRemoteStream): The msid property can be null, so needs checking. (WebCore::GStreamerMediaEndpoint::requestAuxiliarySender): Remove auxiliary sender caching. This code is called once per transport stream, ownership of the sender is passed to the caller. Canonical link: https://commits.webkit.org/259473@main

https://bugs.webkit.org/show_bug.cgi?id=251204 <rdar://problem/104691207> Reviewed by Myles C. Maxfield. The `Identifier`s generated by the parser keep a reference and offset to the input source, but no references are kept to make sure the source stays around, which resulted in it being deallocated while the AST was still in use. I added a test that crashes if Guard Malloc is enabled. * Source/WebGPU/WGSL/AST/ASTShaderModule.h: * Source/WebGPU/WGSL/Parser.cpp: (WGSL::parse): (WGSL::Parser<Lexer>::parseShader): * Source/WebGPU/WGSL/Parser.h: * Source/WebGPU/WGSL/ParserPrivate.h: * Tools/TestWebKitAPI/Tests/WGSL/ParserTests.cpp: (TestWGSLAPI::TEST): Canonical link: https://commits.webkit.org/259472@main

https://bugs.webkit.org/show_bug.cgi?id=251254 Unreviewed, follow-up from 259429@main listing all the required patches for the FDO SDK junction. * Tools/buildstream/elements/freedesktop-sdk.bst: Canonical link: https://commits.webkit.org/259471@main

https://bugs.webkit.org/show_bug.cgi?id=251249 <rdar://problem/104730730> Reviewed by Yusuke Suzuki. Most Interpreter execute methods already have unique names e.g. executeProgram, executeCall, etc. The only 2 that are still not disambiguated this way are for eval and cached calls. Rename these to executeEval and executeCachedCall so that they are more easily grep'able. * Source/JavaScriptCore/debugger/DebuggerCallFrame.cpp: (JSC::DebuggerCallFrame::evaluateWithScopeExtension): * Source/JavaScriptCore/interpreter/CachedCall.h: (JSC::CachedCall::call): * Source/JavaScriptCore/interpreter/Interpreter.cpp: (JSC::eval): (JSC::Interpreter::executeEval): (JSC::Interpreter::execute): Deleted. * Source/JavaScriptCore/interpreter/Interpreter.h: * Source/JavaScriptCore/interpreter/InterpreterInlines.h: (JSC::Interpreter::executeCachedCall): (JSC::Interpreter::execute): Deleted. * Source/JavaScriptCore/runtime/JSGlobalObjectFunctions.cpp: (JSC::JSC_DEFINE_HOST_FUNCTION): * Source/JavaScriptCore/runtime/ShadowRealmPrototype.cpp: (JSC::JSC_DEFINE_HOST_FUNCTION): Canonical link: https://commits.webkit.org/259470@main

https://bugs.webkit.org/show_bug.cgi?id=251236 rdar://problem/104721662 Reviewed by Matt Woodrow. Move WebGLStateTracker files to the same location as the rest of WebGL. * Source/WebCore/Headers.cmake: * Source/WebCore/Sources.txt: * Source/WebCore/WebCore.xcodeproj/project.pbxproj: * Source/WebCore/html/canvas/WebGLStateTracker.cpp: Renamed from Source/WebCore/platform/WebGLStateTracker.cpp. (WebCore::WebGLStateTracker::WebGLStateTracker): (WebCore::m_stateChangeHandler): (WebCore::WebGLStateTracker::token): (WebCore::WebGLStateTracker::updateWebGLState): * Source/WebCore/html/canvas/WebGLStateTracker.h: Renamed from Source/WebCore/platform/WebGLStateTracker.h. Canonical link: https://commits.webkit.org/259469@main

…ws you to omit redundant information https://bugs.webkit.org/show_bug.cgi?id=251239 rdar://104722324 Reviewed by Ryosuke Niwa. We probably don't want people using -> in function signatures all over the place where it isn't necessary. * Websites/webkit.org/code-style.md: Canonical link: https://commits.webkit.org/259468@main

https://bugs.webkit.org/show_bug.cgi?id=251235 rdar://104720292 Reviewed by Simon Fraser. Since layout tests use mock scrollbars, we also need to hide native scrollbars for this case. To do so we check DeprecatedGlobalSettings::mockScrollbarsEnabled in Scrollbar::nativeScrollbarVisibility. We also refactor some of the code that handles turning off the native scrollbars to use the getters on ScrollingTreeScrollingNode. * Source/WebCore/page/scrolling/ScrollingTreeScrollingNodeDelegate.h: (WebCore::ScrollingTreeScrollingNodeDelegate::getScrollbarLayersForStateNode): Deleted. * Source/WebCore/page/scrolling/mac/ScrollingTreeScrollingNodeDelegateMac.h: * Source/WebCore/page/scrolling/mac/ScrollingTreeScrollingNodeDelegateMac.mm: (WebCore::ScrollingTreeScrollingNodeDelegateMac::getScrollbarLayersForStateNode): Deleted. * Source/WebCore/platform/Scrollbar.cpp: (WebCore::Scrollbar::nativeScrollbarVisibility): * Source/WebKit/UIProcess/RemoteLayerTree/mac/ScrollingTreeFrameScrollingNodeRemoteMac.cpp: (WebKit::ScrollingTreeFrameScrollingNodeRemoteMac::commitStateBeforeChildren): * Source/WebKit/UIProcess/RemoteLayerTree/mac/ScrollingTreeOverflowScrollingNodeRemoteMac.cpp: (WebKit::ScrollingTreeOverflowScrollingNodeRemoteMac::commitStateBeforeChildren): Canonical link: https://commits.webkit.org/259467@main

https://bugs.webkit.org/show_bug.cgi?id=251094 Reviewed by Youenn Fablet. The Clear-Site-Data HTTP header should obey origin partition. If shouldn't be possible for an iframe of origin A under top origin B to be able to clear site data from top origin A (and vice-versa). Our storages are partitioned and the request to clear site data should respect that. * LayoutTests/http/wpt/clear-site-data/partitioning-expected.txt: Added. * LayoutTests/http/wpt/clear-site-data/partitioning.html: Added. * LayoutTests/http/wpt/clear-site-data/resources/partitioning-popup.html: Added. * LayoutTests/imported/w3c/web-platform-tests/clear-site-data/support/test_utils.sub.js: * Source/WebCore/loader/cache/MemoryCache.cpp: (WebCore::MemoryCache::removeResourcesWithOrigin): * Source/WebCore/loader/cache/MemoryCache.h: * Source/WebCore/platform/network/NetworkStorageSession.cpp: (WebCore::NetworkStorageSession::deleteCookies): * Source/WebCore/platform/network/NetworkStorageSession.h: * Source/WebCore/platform/network/cocoa/NetworkStorageSessionCocoa.mm: (WebCore::NetworkStorageSession::deleteCookiesMatching): (WebCore::NetworkStorageSession::deleteCookies): (WebCore::NetworkStorageSession::deleteCookiesForHostnames): * Source/WebCore/workers/service/ServiceWorkerRegistrationKey.cpp: (WebCore::ServiceWorkerRegistrationKey::matchesOriginWithCachePartition const): * Source/WebCore/workers/service/ServiceWorkerRegistrationKey.h: * Source/WebCore/workers/service/server/SWServer.cpp: (WebCore::SWServer::clear): (WebCore::SWServer::clearInternal): * Source/WebCore/workers/service/server/SWServer.h: * Source/WebKit/NetworkProcess/NetworkConnectionToWebProcess.cpp: (WebKit::NetworkConnectionToWebProcess::deleteWebsiteDataForOrigin): (WebKit::NetworkConnectionToWebProcess::deleteWebsiteDataForOrigins): Deleted. * Source/WebKit/NetworkProcess/NetworkConnectionToWebProcess.h: * Source/WebKit/NetworkProcess/NetworkProcess.cpp: (WebKit::NetworkProcess::deleteWebsiteDataForOrigin): * Source/WebKit/NetworkProcess/NetworkProcess.h: * Source/WebKit/NetworkProcess/NetworkResourceLoader.cpp: (WebKit::NetworkResourceLoader::processClearSiteDataHeader): * Source/WebKit/NetworkProcess/storage/NetworkStorageManager.cpp: (WebKit::NetworkStorageManager::deleteData): * Source/WebKit/NetworkProcess/storage/NetworkStorageManager.h: * Source/WebKit/WebProcess/Network/NetworkProcessConnection.cpp: (WebKit::NetworkProcessConnection::deleteWebsiteDataForOrigin): (WebKit::NetworkProcessConnection::deleteWebsiteDataForOrigins): Deleted. * Source/WebKit/WebProcess/Network/NetworkProcessConnection.h: * Source/WebKit/WebProcess/Network/NetworkProcessConnection.messages.in: * Source/WebKit/WebProcess/WebProcess.cpp: (WebKit::WebProcess::deleteWebsiteDataForOrigin): * Source/WebKit/WebProcess/WebProcess.h: Canonical link: https://commits.webkit.org/259466@main

…rd library proposal https://bugs.webkit.org/show_bug.cgi?id=251228 rdar://104717280 Reviewed by Yusuke Suzuki. http://wg21.link/P2527 and an upcoming r2 have shifted the design slightly to use std::integral_constant and remove inline. This PR just uses the latest changes to gain some "implementation experience" and verify it can be used in practice. * Source/WTF/wtf/StdLibExtras.h: Canonical link: https://commits.webkit.org/259465@main

Commits on Jan 27, 2023