Merge upstream 2023 12 18 #20
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…dio-session-mode.html is a flaky crash https://bugs.webkit.org/show_bug.cgi?id=265954 rdar://119265443 Unreviewed test gardening. Updating test expectation * LayoutTests/platform/mac-wk2/TestExpectations: Canonical link: https://commits.webkit.org/272003@main
https://bugs.webkit.org/show_bug.cgi?id=266311 rdar://119582604 Reviewed by Justin Michaud. Let's not fallback to slow C++ JSBoundFunction code, which invokes function through interpreter again. We can just materialize JIT code from JSBoundFunction thunk and continue running from JS world instead of jumping to C++ and jumping back to JS. * Source/JavaScriptCore/jit/JITOperations.cpp: (JSC::materializeTargetCode): (JSC::JSC_DEFINE_JIT_OPERATION): * Source/JavaScriptCore/jit/JITOperations.h: * Source/JavaScriptCore/jit/ThunkGenerators.cpp: (JSC::boundFunctionCallGenerator): * Source/JavaScriptCore/runtime/VM.cpp: (JSC::thunkGeneratorForIntrinsic): Canonical link: https://commits.webkit.org/272004@main
https://bugs.webkit.org/show_bug.cgi?id=266355 rdar://119621360 Reviewed by Chris Dumez. Store the String to be parsed in a member variable instead of only storing pointers to the start and end of the String to avoid crashes like the one caused by caused by 270868@main. * Source/WebCore/html/track/VTTScanner.cpp: (WebCore::VTTScanner::VTTScanner): * Source/WebCore/html/track/VTTScanner.h: Canonical link: https://commits.webkit.org/272005@main
* Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebViewCloseAllMediaPresentations.mm: Canonical link: https://commits.webkit.org/272006@main
https://bugs.webkit.org/show_bug.cgi?id=266374 rdar://119632464 Broke the TestWebKitAPI build Unreviewed build fix. * Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebViewCloseAllMediaPresentations.mm: Canonical link: https://commits.webkit.org/272007@main
https://bugs.webkit.org/show_bug.cgi?id=266376 Introduced perf regression Reverted changeset: "Unreviewed, reverting 271900@main." https://bugs.webkit.org/show_bug.cgi?id=266340 https://commits.webkit.org/271984@main Canonical link: https://commits.webkit.org/272008@main
https://bugs.webkit.org/show_bug.cgi?id=265990 Reviewed by Žan Doberšek. Add an unique ID to cairo surfaces created in ImageBackingStoreCairo. This unique ID allows us to differentiate when the cairo surface that's backing an image has changed, so we are sure that we're not using an incorrect value cached inside some ImageBackingStore. * Source/WebCore/platform/graphics/cairo/CairoUtilities.cpp: (WebCore::attachSurfaceUniqueID): (WebCore::getSurfaceUniqueID): * Source/WebCore/platform/graphics/cairo/CairoUtilities.h: * Source/WebCore/platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.cpp: (WebCore::CoordinatedGraphicsLayer::flushCompositingStateForThisLayerOnly): * Source/WebCore/platform/image-decoders/cairo/ImageBackingStoreCairo.cpp: (WebCore::ImageBackingStore::image const): Canonical link: https://commits.webkit.org/272009@main
…ain/contain-size-inline-block-003.html only fail on GTK. https://bugs.webkit.org/show_bug.cgi?id=266345 rdar://problem/119617786 Unreviewed, test gardening. * LayoutTests/TestExpectations: * LayoutTests/platform/gtk/TestExpectations: Canonical link: https://commits.webkit.org/272010@main
https://bugs.webkit.org/show_bug.cgi?id=266357 Reviewed by Antti Koivisto. Let's hold on to ruby base min/max width and make sure the associated annotation box's min/max values are taken into account. * Source/WebCore/rendering/RenderBlockFlow.cpp: (WebCore::RenderBlockFlow::computeInlinePreferredLogicalWidths const): Canonical link: https://commits.webkit.org/272011@main
https://bugs.webkit.org/show_bug.cgi?id=266294 rdar://119563251 Reviewed by Justin Michaud. FOR_EACH_BYTECODE_HELPER_ID was mistakenly lumped in with FOR_EACH_LLINT_NATIVE_HELPER, which does not populate their entries in g_opcodeMapWide16 and g_opcodeMapWide32. This patch fixes that. This was causing CLoop runs to crash on nullptr derefs on these non-initialized entries. * Source/JavaScriptCore/llint/LLIntOpcode.h: * Source/JavaScriptCore/llint/LowLevelInterpreter.cpp: (JSC::CLoop::execute): Canonical link: https://commits.webkit.org/272012@main
…splay-sampling-profiler-output https://bugs.webkit.org/show_bug.cgi?id=266327 rdar://119597908 Reviewed by Justin Michaud. Add --top-bytecode-count and --top-function-count options to display-sampling-profiler-output to see more information easily. * Tools/Scripts/display-sampling-profiler-output: Canonical link: https://commits.webkit.org/272013@main
…ting AVSampleBufferAttachContentKey https://bugs.webkit.org/show_bug.cgi?id=266375 rdar://119345487 Reviewed by Jer Noble. On iOS a key may be usable for decryption even when its status isn't KeyStatus::Usable. Accounted for this by reusing the logic in CDMInstanceFairPlayStreamingAVFObjC::isAnyKeyUsable to determine if a key is usable. Also asserted that there is no usable key whenever CDMInstanceSessionFairPlayStreamingAVFObjC::contentKeyForSample returns nil. * Source/WebCore/platform/graphics/avfoundation/objc/CDMInstanceFairPlayStreamingAVFObjC.h: * Source/WebCore/platform/graphics/avfoundation/objc/CDMInstanceFairPlayStreamingAVFObjC.mm: (WebCore::isPotentiallyUsableKeyStatus): (WebCore::CDMInstanceFairPlayStreamingAVFObjC::isAnyKeyUsable const): (WebCore::CDMInstanceSessionFairPlayStreamingAVFObjC::isAnyKeyUsable const): (WebCore::CDMInstanceSessionFairPlayStreamingAVFObjC::contentKeyForSample): Canonical link: https://commits.webkit.org/272014@main
https://bugs.webkit.org/show_bug.cgi?id=266370 Reviewed by Ryosuke Niwa. Stop using CheckedPtr / CheckedRef in TreeScopeOrderedMap.h. Use WeakPtr / WeakRef instead to generate more actionable crashes. This tested as performance neutral on Speedometer 2 & 3. * Source/WebCore/dom/TreeScope.cpp: (WebCore::TreeScope::getAllElementsById const): (WebCore::TreeScope::labelElementsForId): * Source/WebCore/dom/TreeScope.h: * Source/WebCore/dom/TreeScopeOrderedMap.cpp: (WebCore::TreeScopeOrderedMap::add): (WebCore::TreeScopeOrderedMap::remove): (WebCore::TreeScopeOrderedMap::get const): (WebCore::TreeScopeOrderedMap::getAll const): (WebCore::TreeScopeOrderedMap::getElementsByLabelForAttribute const): (WebCore::TreeScopeOrderedMap::getAllElementsById const): * Source/WebCore/dom/TreeScopeOrderedMap.h: Canonical link: https://commits.webkit.org/272015@main
https://bugs.webkit.org/show_bug.cgi?id=265721 Reviewed by Justin Michaud. This patch fixes two related issues. The main issue is that there is a cache for type definition unrollings that didn't hold the values strongly, and since these values are not held by the Wasm instance they could get de-allocated. This could cause correctness issues or crashes. That bug was in practice hard to trigger (though it did trigger on some large compiled-to-wasm examples), because it was hidden by a second bug. The placeholders in recursive types (used to delay type expansion to cut off infinite expansion of recursive types) were also not held strongly by the TypeInformation store, so they could also get de-allocated when an instance gets collected. This bug makes memoization less effective but didn't affect correctness. This patch makes the placeholders held by the TypeInformation store itself, similar to how special type signatures are held by the store. They are also only allocated via a helper function now, which ensures the store holds the typedef. It also fixes the cache by using RefPtr for the value type. * JSTests/wasm/gc/typedef.js: * Source/JavaScriptCore/wasm/WasmParser.h: (JSC::Wasm::Parser<SuccessType>::parseValueType): * Source/JavaScriptCore/wasm/WasmSectionParser.cpp: (JSC::Wasm::SectionParser::parseSubtype): * Source/JavaScriptCore/wasm/WasmTypeDefinition.cpp: (JSC::Wasm::TypeInformation::getPlaceholderProjection): * Source/JavaScriptCore/wasm/WasmTypeDefinition.h: Canonical link: https://commits.webkit.org/272016@main
https://bugs.webkit.org/show_bug.cgi?id=266056 Reviewed by Justin Michaud. Use subtyping instead of type equality for annotated select arguments. * JSTests/wasm/gc/bug266056.js: Added. * Source/JavaScriptCore/wasm/WasmFunctionParser.h: (JSC::Wasm::FunctionParser<Context>::parseExpression): Canonical link: https://commits.webkit.org/272017@main
https://bugs.webkit.org/show_bug.cgi?id=266302 rdar://119568570 Reviewed by Tim Horton. Adopt `WKSE*`-prefixed objects, which are aliased to the new classes, protocols, or types and constants exposed from the ServiceExtensions framework (if available), or otherwise aliased to the `UIAsync*` SPI names. No change in behavior; see below for more details. * Source/WebCore/WebCore.xcodeproj/project.pbxproj: * Source/WebCore/platform/ios/WebEvent.mm: (webEventType): (isChangingKeyModifiers): (-[WebEvent initWithKeyEvent:]): (-[WebEvent originalKeyEvent]): (-[WebEvent initWithUIKeyEvent:]): Deleted. (-[WebEvent originalUIKeyEvent]): Deleted. * Source/WebCore/platform/ios/WebEventPrivate.h: * Source/WebCore/platform/ios/WebSEDefinitions.h: Copied from Source/WebKit/UIProcess/API/ios/WKSEDefinitions.h. * Source/WebKit/Platform/spi/ios/UIKitSPI.h: * Source/WebKit/Shared/DocumentEditingContext.h: * Source/WebKit/Shared/DocumentEditingContext.mm: (WebKit::setOptionalEditingContextProperties): (WebKit::DocumentEditingContext::toLegacyPlatformContext): (WebKit::DocumentEditingContext::toPlatformContext): * Source/WebKit/UIProcess/API/ios/WKSEDefinitions.h: * Source/WebKit/UIProcess/API/ios/WKWebViewPrivateForTestingIOS.h: * Source/WebKit/UIProcess/WebPageProxy.h: * Source/WebKit/UIProcess/ios/WKContentViewInteraction.h: * Source/WebKit/UIProcess/ios/WKContentViewInteraction.mm: (-[WKContentView _updateRuntimeProtocolConformanceIfNeeded]): (-[WKContentView textInteractionGesture:shouldBeginAtPoint:]): (toGestureType): (toWKSEGestureType): (toSelectionTouch): (toWKSESelectionTouch): (toWKSESelectionFlags): (toSelectionFlags): (selectionChangedWithGesture): (selectionChangedWithTouch): (-[WKContentView changeSelectionWithGestureAt:withGesture:withState:]): (-[WKContentView changeSelectionWithGestureAt:withGesture:withState:withFlags:]): (-[WKContentView changeSelectionWithTouchAt:withSelectionTouch:baseIsStart:withFlags:]): (-[WKContentView changeSelectionWithTouchesFrom:to:withGesture:withState:]): (logTextInteraction): (-[WKContentView insertTextSuggestion:]): (shiftKeyState): (-[WKContentView _deferKeyEventToInputMethodEditing:]): (-[WKContentView _interpretKeyEvent:isCharEvent:]): (-[WKContentView dataListTextSuggestions]): (-[WKContentView setDataListTextSuggestions:]): (-[WKContentView updateTextSuggestionsForInputDelegate]): (-[WKContentView _provideSuggestionsToInputDelegate:]): (-[WKContentView _provideUITextSuggestionsToInputDelegate:]): (toWebRequest): (-[WKContentView requestDocumentContext:completionHandler:]): Make the completion handler here take an `NSObject`, representing either a `UIWKDocumentRequest` or the replacement class from ServiceExtensions. This allows us to maintain both binary and source compatibility in all the following cases: - Against iOS 17, which doesn't have async text input support at all, or against versions of iOS 17 that have `UIAsyncTextInput` but not the replacement APIs in ServiceExtensions. We fall back to `UIWKDocumentContext` here. - Building/running against versions of iOS 17 that have both `UIAsyncTextInput` and the requisite APIs in ServiceExtensions, but have `UIKit/async_text_input` disabled by default — here, we use the new ServiceExtensions API objects and flags (which, for the most part, maintain binary compatibility), with the exception of `WKSETextDocumentRequest`. In the completion handler, if async text input is disabled, we return a `UIWKDocumentRequest` to maintain binary compatibility. - Lastly, building/running against versions of iOS 17 that have both `UIAsyncTextInput` and the requisite APIs in ServiceExtensions, and enable `UIKit/async_text_input` by default. Here, we return the new `WKSETextDocumentRequest`. (-[WKContentView selectPositionAtPoint:withContextRequest:completionHandler:]): (-[WKContentView asyncSystemInputDelegate]): (-[WKContentView setAsyncSystemInputDelegate:]): (-[WKContentView handleAsyncKeyEvent:withCompletionHandler:]): (-[WKContentView replaceText:withText:options:withCompletionHandler:]): (-[WKContentView requestTextContextForAutocorrectionWithCompletionHandler:]): Refactor this code to use `WebKit::DocumentEditingContext` to create the document context, to avoid having to duplicate messy logic for initializing either of the two `WKSETextDocumentRequest` types, depending on the SDK version. Note: we don't need to consult the async text input feature flag here and return a legacy `UIWKDocumentRequest`, since this codepath is only ever invoked in the async text input codepath to begin with. (-[WKContentView adjustSelection:completionHandler:]): (-[WKContentView shiftKeyStateChangedFrom:to:]): (-[WKContentView selectionWillChange:]): (-[WKContentView selectionDidChange:]): (toUIWKGestureType): Deleted. (toUIWKSelectionTouch): Deleted. (toUIWKSelectionFlags): Deleted. * Source/WebKit/UIProcess/ios/WKExtendedTextInputTraits.h: * Source/WebKit/UIProcess/ios/WKTextInteractionWrapper.h: * Source/WebKit/UIProcess/ios/WKTextInteractionWrapper.mm: (-[WKTextInteractionWrapper initWithView:]): (-[WKTextInteractionWrapper selectionChangedWithGestureAt:withGesture:withState:withFlags:]): (-[WKTextInteractionWrapper selectionChangedWithTouchAt:withSelectionTouch:withFlags:]): (SOFT_LINK_CLASS_OPTIONAL): Deleted. * Source/WebKit/UIProcess/ios/WebDataListSuggestionsDropdownIOS.h: * Source/WebKit/UIProcess/ios/WebDataListSuggestionsDropdownIOS.mm: (+[WKDataListTextSuggestion textSuggestionWithInputText:]): * Source/WebKit/UIProcess/ios/WebPageProxyIOS.mm: (WebKit::WebPageProxy::requestDocumentEditingContext): Drive-by fixes: make both the completion handler and method itself take rvalue references, instead of passing by value. * Source/WebKit/WebProcess/WebPage/WebPage.h: * Source/WebKit/WebProcess/WebPage/ios/WebPageIOS.mm: (WebKit::WebPage::requestDocumentEditingContext): Canonical link: https://commits.webkit.org/272018@main
https://bugs.webkit.org/show_bug.cgi?id=266388 rdar://119634195 Reviewed by Pascoe. In 270267@main I introduced a null dereference that introduced a possible crash. Covered by the test in the radar. * Source/WebCore/loader/FrameLoader.cpp: (WebCore::FrameLoader::updateFirstPartyForCookies): Canonical link: https://commits.webkit.org/272019@main
https://bugs.webkit.org/show_bug.cgi?id=266281 rdar://119555352 Reviewed by Aditya Keerthi. Make RenderThemeIOS build on ControlStyle::State rather than ControlStates::States. The ControlStates architecture is dated, does not account for shadow hosts, and should probably be disappeared entirely. This is an incremental step towards that goal. Then provide the switch's thumb painting code has access to the system outline color. And ensure that by default we don't draw an outline for a switch on iOS. Web developers can override this by using a non-auto outline-style value, but that will not draw the outline around the thumb. * Source/WebCore/rendering/RenderTheme.cpp: (WebCore::RenderTheme::extractControlStyleStatesForRendererInternal const): (WebCore::adjustRendererForAppearance): (WebCore::RenderTheme::extractControlStyleStatesForRenderer const): (WebCore::RenderTheme::extractControlStyleForRenderer const): * Source/WebCore/rendering/RenderTheme.h: * Source/WebCore/rendering/RenderThemeIOS.h: * Source/WebCore/rendering/RenderThemeIOS.mm: (WebCore::renderThemePaintSwitchThumb): (WebCore::renderThemePaintSwitchTrack): (WebCore::RenderThemeIOS::adjustSwitchStyle const): (WebCore::RenderThemeIOS::paintSwitchThumb): (WebCore::RenderThemeIOS::paintSwitchTrack): (WebCore::RenderThemeIOS::checkboxRadioBorderColor): (WebCore::RenderThemeIOS::checkboxRadioBackgroundColor): (WebCore::RenderThemeIOS::checkboxRadioBackgroundGradient): (WebCore::RenderThemeIOS::checkboxRadioIndicatorColor): (WebCore::RenderThemeIOS::paintCheckboxRadioInnerShadow): (WebCore::RenderThemeIOS::paintCheckbox): (WebCore::RenderThemeIOS::paintRadio): Canonical link: https://commits.webkit.org/272020@main
https://bugs.webkit.org/show_bug.cgi?id=266278 Reviewed by Carlos Garcia Campos. Changeset 271330@main makes use of function 'zwp_linux_dmabuf_v1_get_surface_feeback', which is only available since wayland-protocols v1.24. Also, update 'wayland' to 1.18 since it's the version available in the oldest system supported at this time, Ubuntu 20.04. Update JHBuild modules accordingly. * Source/cmake/OptionsGTK.cmake: * Source/cmake/OptionsWPE.cmake: * Tools/gtk/jhbuild.modules: * Tools/jhbuild/jhbuild-minimal.modules: * Tools/wpe/jhbuild.modules: Canonical link: https://commits.webkit.org/272021@main
… and `height` are animated https://bugs.webkit.org/show_bug.cgi?id=211986 rdar://63309680 Reviewed by Simon Fraser. When animating the `transform` or `translate` property, we resolve percentage values for translate operations to fixed values. However, if the element also has an animation of `width` or `height`, the resolved translation values will no longer be correct while the animation is in flight. In that case, we need to detect such a case and turn off acceleration support. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/transform-percent-with-width-and-height-expected.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/transform-percent-with-width-and-height-ref.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/transform-percent-with-width-and-height-separate-expected.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/transform-percent-with-width-and-height-separate-ref.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/transform-percent-with-width-and-height-separate.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/transform-percent-with-width-and-height.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/translate-percent-with-width-and-height-expected.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/translate-percent-with-width-and-height-ref.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/translate-percent-with-width-and-height-separate-expected.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/translate-percent-with-width-and-height-separate-ref.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/translate-percent-with-width-and-height-separate.html: Added. * LayoutTests/imported/w3c/web-platform-tests/css/css-transforms/animation/translate-percent-with-width-and-height.html: Added. * Source/WebCore/animation/KeyframeEffect.cpp: (WebCore::KeyframeEffect::setBlendingKeyframes): (WebCore::KeyframeEffect::canBeAccelerated const): (WebCore::KeyframeEffect::applyPendingAcceleratedActions): (WebCore::KeyframeEffect::computeHasSizeDependentTransform): * Source/WebCore/animation/KeyframeEffect.h: Canonical link: https://commits.webkit.org/272022@main
https://bugs.webkit.org/show_bug.cgi?id=266350 rdar://119217913 Reviewed by Brent Fulgham. Call _RegisterApplication on Catalyst while holding a XPC sandbox extension. Calling this function without the sandbox extension will crash. * Source/WebCore/PAL/pal/spi/cocoa/LaunchServicesSPI.h * Source/WebKit/UIProcess/Cocoa/WebProcessPoolCocoa.mm * Source/WebKit/WebProcess/WebProcess.h * Source/WebKit/WebProcess/cocoa/WebProcessCocoa.mm * Source/WebKit/WebProcess/cocoa/WebProcessCocoa.mm: (WebKit::WebProcess::platformInitializeWebProcess): Canonical link: https://commits.webkit.org/272023@main
https://bugs.webkit.org/show_bug.cgi?id=250454 rdar://104377727 Reviewed by Chris Dumez. In case a service worker is located in the same process as a controlled document, they both share the same memory cache. Before the patch, in case of preloads triggered byt he speculative parser and the preloads are not cacheable, the following would happen: - Trigger a preload, which gets registered in the memory cache. - Let the service worker handle the preload, which triggers a fetch of the same resource. - This fetch will remove the preload resource from the memory cache and do the actual network load. - The real load of the HTML resource kicks in and tries to load the resource from the memory cache. - Given the resource is not a preload and is not reusable according the HTTP headers, a new load is triggered. This makes it so that two loads are happening for the same resource (one for the preload and one for the actual load). To resolve the issue, we do not add the service worker fetch resource in the memory cache in the following case: - There is an existing cached resource, which is a preload, and not a service worker request. - The new request is a service worker request. This ensures that we keep the preload cached resource in the memory cache which allows reuse when the actual load happens. * LayoutTests/http/wpt/service-workers/resources/serviceworker-and-preloads-iframe.html: Added. * LayoutTests/http/wpt/service-workers/resources/serviceworker-preload-img.png: Added. * LayoutTests/http/wpt/service-workers/resources/serviceworker-preload-img.png.headers: Added. * LayoutTests/http/wpt/service-workers/resources/serviceworker-preload-script.js: Added. * LayoutTests/http/wpt/service-workers/resources/serviceworker-preload-script.js.headers: Added. * LayoutTests/http/wpt/service-workers/serviceworker-and-preloads-expected.txt: Added. * LayoutTests/http/wpt/service-workers/serviceworker-and-preloads-worker.js: Added. * LayoutTests/http/wpt/service-workers/serviceworker-and-preloads.html: Added. * Source/WebCore/loader/cache/CachedResourceLoader.cpp: (WebCore::CachedResourceLoader::updateCachedResourceWithCurrentRequest): (WebCore::computeShouldAddToMemoryCache): (WebCore::CachedResourceLoader::requestResource): (WebCore::CachedResourceLoader::loadResource): * Source/WebCore/loader/cache/CachedResourceLoader.h: Canonical link: https://commits.webkit.org/272024@main
https://bugs.webkit.org/show_bug.cgi?id=266385 rdar://119615258 Reviewed by Jer Noble. Changes in 271514@main exposed an underlying existing issue. The MediaSource's duration was initialised to NaN but the GPU process one was initialised to 0. Following 271514@main the inital value would be identical. If the media source duration was set prior any sourcebuffer being created, we wouldn't set the SourceBuffer initial duration, and the step `If the media segment contains data beyond the current duration, then run the duration change algorithm with new duration set to the maximum of the current duration and the [[group end timestamp]].` would always be skipped. We set the initial duration to the sourceBuffer when created. Add tests. * LayoutTests/media/media-source/media-source-append-buffer-durationchange-expected.txt: Added. * LayoutTests/media/media-source/media-source-append-buffer-durationchange.html: Added. * Source/WebCore/platform/graphics/avfoundation/objc/MediaSourcePrivateAVFObjC.mm: (WebCore::MediaSourcePrivateAVFObjC::addSourceBuffer): * Source/WebCore/platform/graphics/gstreamer/mse/MediaSourcePrivateGStreamer.cpp: (WebCore::MediaSourcePrivateGStreamer::addSourceBuffer): * Source/WebCore/platform/mock/mediasource/MockMediaSourcePrivate.cpp: (WebCore::MockMediaSourcePrivate::addSourceBuffer): Canonical link: https://commits.webkit.org/272025@main
https://bugs.webkit.org/show_bug.cgi?id=266394 Reviewed by Carlos Garcia Campos. Fixes linking error. Symbol 'wl_proxy_marshal_flags' is only defined since wayland 1.20. * Source/cmake/OptionsGTK.cmake: * Source/cmake/OptionsWPE.cmake: * Tools/gtk/jhbuild.modules: * Tools/jhbuild/jhbuild-minimal.modules: * Tools/wpe/jhbuild.modules: Canonical link: https://commits.webkit.org/272026@main
https://bugs.webkit.org/show_bug.cgi?id=266330 rdar://119599179 Unreviewed test expectations gardening. A lot of tests under `http/tests/paymentrequest/` were previously skipped on both macOS and iOS because they tested for features yet to be available in public versions of PassKit. This is no longer the case, so we can safely unskip these tests. The only kink comes from having to single out some tests to skip on macOS Monterey. These tests happen to be testing features available from macOS Ventura-onwards only. Some more tests still do not require the DumpJSConsoleLogInStdErr classification, either because they are not producing JS console logs or because their expectations have been rebaselined to reflect the console log. We also unskip these tests. * LayoutTests/platform/ios-wk2/TestExpectations: * LayoutTests/platform/mac-monterey-wk2/TestExpectations: * LayoutTests/platform/mac-ventura-wk2/TestExpectations: * LayoutTests/platform/mac-wk2/TestExpectations: Canonical link: https://commits.webkit.org/272027@main
…r::equal(KeyTraits::emptyValue(), key) https://bugs.webkit.org/show_bug.cgi?id=266333 Reviewed by Xabier Rodriguez-Calvar. HashMap doesn't like to store key that evaluate to false, so a TrackID with value 0 would trigger an ASSERT. So switch back to unordered_map for tracks storage, and avoid std::views for access, which is not available in clang-14. * LayoutTests/platform/glib/TestExpectations: * Source/WebCore/platform/graphics/gstreamer/mse/MediaSourcePrivateGStreamer.cpp: (WebCore::MediaSourcePrivateGStreamer::startPlaybackIfHasAllTracks): * Source/WebCore/platform/graphics/gstreamer/mse/SourceBufferPrivateGStreamer.cpp: (WebCore::SourceBufferPrivateGStreamer::removedFromMediaSource): (WebCore::SourceBufferPrivateGStreamer::flush): (WebCore::SourceBufferPrivateGStreamer::enqueueSample): (WebCore::SourceBufferPrivateGStreamer::isReadyForMoreSamples): (WebCore::SourceBufferPrivateGStreamer::notifyClientWhenReadyForMoreSamples): (WebCore::SourceBufferPrivateGStreamer::allSamplesInTrackEnqueued): (WebCore::SourceBufferPrivateGStreamer::precheckInitializationSegment): * Source/WebCore/platform/graphics/gstreamer/mse/SourceBufferPrivateGStreamer.h: Canonical link: https://commits.webkit.org/272028@main
https://bugs.webkit.org/show_bug.cgi?id=265791 rdar://119133276 Reviewed by Eric Carlson. Migrate code from deprecated webrtc::AsyncResolverFactory to webrtc::AsyncDnsResolverFactory. This in particular simplifies how we do memory management, since the peer connection factory is now owned by the peer connection. We keep using a map of WeakPtr to allow to handle IPC responses from network process. We work around a header conflict between SDK and libwebrtc by isolating absl::InvokeCall in LibWebRTCDnsResolverFactory.cpp. * Source/WebCore/platform/mediastream/libwebrtc/LibWebRTCProvider.cpp: (WebCore::LibWebRTCProvider::createPeerConnection): * Source/WebCore/platform/mediastream/libwebrtc/LibWebRTCProvider.h: * Source/WebKit/Sources.txt: * Source/WebKit/WebKit.xcodeproj/project.pbxproj: * Source/WebKit/WebProcess/Network/webrtc/LibWebRTCDnsResolverFactory.cpp: Added. (WebKit::LibWebRTCDnsResolverFactory::CreateAndResolve): (WebKit::LibWebRTCDnsResolverFactory::Create): (WebKit::LibWebRTCDnsResolverFactory::Resolver::Start): * Source/WebKit/WebProcess/Network/webrtc/LibWebRTCDnsResolverFactory.h: Added. * Source/WebKit/WebProcess/Network/webrtc/LibWebRTCProvider.cpp: (WebKit::LibWebRTCProvider::createPeerConnection): (WebKit::RTCSocketFactory::CreateAsyncDnsResolver): (WebKit::RTCSocketFactory::CreateAsyncResolver): Deleted. * Source/WebKit/WebProcess/Network/webrtc/LibWebRTCResolver.cpp: (WebKit::LibWebRTCResolver::~LibWebRTCResolver): (WebKit::LibWebRTCResolver::start): (WebKit::LibWebRTCResolver::result const): (WebKit::LibWebRTCResolver::setResolvedAddress): (WebKit::LibWebRTCResolver::setError): (WebKit::LibWebRTCResolver::Start): Deleted. (WebKit::LibWebRTCResolver::Destroy): Deleted. (WebKit::LibWebRTCResolver::doDestroy): Deleted. * Source/WebKit/WebProcess/Network/webrtc/LibWebRTCResolver.h: * Source/WebKit/WebProcess/Network/webrtc/LibWebRTCSocketFactory.cpp: (WebKit::LibWebRTCSocketFactory::createAsyncDnsResolver): (WebKit::LibWebRTCSocketFactory::createAsyncResolver): Deleted. * Source/WebKit/WebProcess/Network/webrtc/LibWebRTCSocketFactory.h: (WebKit::LibWebRTCSocketFactory::resolver): (WebKit::LibWebRTCSocketFactory::removeResolver): (WebKit::LibWebRTCSocketFactory::takeResolver): Deleted. * Source/WebKit/WebProcess/Network/webrtc/WebRTCResolver.cpp: (WebKit::WebRTCResolver::setResolvedAddress): (WebKit::WebRTCResolver::resolvedAddressError): Canonical link: https://commits.webkit.org/272029@main
…change CPU frequency governor for RPis https://bugs.webkit.org/show_bug.cgi?id=266353 Reviewed by Philippe Normand. This updates the Yocto layers used for the RPi build with the following major changes: - meta-webkit: include libbacktrace and libportal into the base build. libbacktrace is required by the default build of WPE now and libportal is required by Cog since bug 266351 - meta-raspberrypi: stop setting "powersave" as the default CPU frequency governor and use "schedutil" instead. This will likely mean a big change in the performance results of all tests. See agherzan/meta-raspberrypi#1243 for context. - poky, meta-openembedded: just update to the last version of the Mickledore branch - backport a change landed in meta-openembedded master to the Mickledore branch that updates libbacktrace to allow building it on riscv64. * Tools/yocto/meta-openembedded_backport-libbacktrace-patch-to-mickledore-branch.patch: Added. * Tools/yocto/riscv/manifest.xml: * Tools/yocto/rpi/manifest.xml: * Tools/yocto/targets.conf: Canonical link: https://commits.webkit.org/272030@main
…nlineFormattingUtils::inlineItemWidth https://bugs.webkit.org/show_bug.cgi?id=266387 Reviewed by Antti Koivisto. Do not try to measure out-of-flow content while rebuilding the line due to "revert to last breaking position" type of content. * LayoutTests/fast/inline/inline-content-with-out-of-flow-assert-expected.txt: Added. * LayoutTests/fast/inline/inline-content-with-out-of-flow-assert.html: Added. * Source/WebCore/layout/formattingContexts/inline/InlineLineBuilder.cpp: (WebCore::Layout::LineBuilder::rebuildLineWithInlineContent): Canonical link: https://commits.webkit.org/272031@main
…teBindGroupLayout.html are failing https://bugs.webkit.org/show_bug.cgi?id=266318 <radar://119586689> Reviewed by Tadeu Zagallo. Add validation such tat createBindGroupLayout.html passes and add the passing expectation. Add passing expectations for validation/createBindGroup.html and add some validation so that the CTS test passes. * LayoutTests/http/tests/webgpu/webgpu/api/validation/createBindGroup-expected.txt: * LayoutTests/http/tests/webgpu/webgpu/api/validation/createBindGroupLayout-expected.txt: * Source/WebCore/Modules/WebGPU/GPUStorageTextureAccess.h: (WebCore::convertToBacking): * Source/WebCore/Modules/WebGPU/GPUStorageTextureAccess.idl: * Source/WebCore/Modules/WebGPU/Implementation/WebGPUConvertToBackingContext.cpp: (WebCore::WebGPU::ConvertToBackingContext::convertToBacking): * Source/WebCore/Modules/WebGPU/Implementation/WebGPUDeviceImpl.cpp: (WebCore::WebGPU::DeviceImpl::createSampler): (WebCore::WebGPU::DeviceImpl::createBindGroup): * Source/WebCore/Modules/WebGPU/InternalAPI/WebGPU.serialization.in: * Source/WebCore/Modules/WebGPU/InternalAPI/WebGPUStorageTextureAccess.h: * Source/WebGPU/WGSL/GlobalVariableRewriter.cpp: (WGSL::convertAccess): (WGSL::bindingMemberForGlobal): * Source/WebGPU/WGSL/WGSL.h: * Source/WebGPU/WebGPU/BindGroup.mm: (WebGPU::hasProperUsageFlags): (WebGPU::hasBinding): (WebGPU::formatType): (WebGPU::formatIsFloat): (WebGPU::formatIsDepth): (WebGPU::formatIsSignedInt): (WebGPU::formatIsUnsignedInt): (WebGPU::validateTextureSampleType): (WebGPU::validateTextureViewDimension): (WebGPU::validateStorageTextureViewFormat): (WebGPU::validateSamplerType): (WebGPU::Device::createBindGroup): * Source/WebGPU/WebGPU/BindGroupLayout.h: (WebGPU::BindGroupLayout::create): (WebGPU::BindGroupLayout::createInvalid): * Source/WebGPU/WebGPU/BindGroupLayout.mm: (WebGPU::createArgumentDescriptor): (WebGPU::Device::createBindGroupLayout): (WebGPU::BindGroupLayout::BindGroupLayout): (WebGPU::BindGroupLayout::uniformBuffersPerStage const): (WebGPU::BindGroupLayout::storageBuffersPerStage const): (WebGPU::BindGroupLayout::samplersPerStage const): (WebGPU::BindGroupLayout::texturesPerStage const): (WebGPU::BindGroupLayout::storageTexturesPerStage const): (WebGPU::BindGroupLayout::device const): (WebGPU::BindGroupLayout::isAutoGenerated const): * Source/WebGPU/WebGPU/Buffer.h: * Source/WebGPU/WebGPU/Buffer.mm: (WebGPU::Buffer::isDestroyed const): * Source/WebGPU/WebGPU/ComputePipeline.mm: (WebGPU::createComputePipelineState): (WebGPU::returnInvalidComputePipeline): (WebGPU::Device::createComputePipeline): (WebGPU::Device::createComputePipelineAsync): * Source/WebGPU/WebGPU/Device.h: * Source/WebGPU/WebGPU/PipelineLayout.mm: (WebGPU::Device::createPipelineLayout): * Source/WebGPU/WebGPU/RenderPipeline.mm: (WebGPU::convertToMTLCompare): (WebGPU::convertAccess): (WebGPU::convertFormat): (WebGPU::Device::addPipelineLayouts): (WebGPU::returnInvalidRenderPipeline): (WebGPU::Device::createRenderPipeline): (WebGPU::Device::createRenderPipelineAsync): * Source/WebGPU/WebGPU/Sampler.mm: (WebGPU::compareFunction): * Source/WebGPU/WebGPU/ShaderModule.mm: (WebGPU::wgslAccess): (WebGPU::wgslFormat): (WebGPU::convertBindingLayout): * Source/WebGPU/WebGPU/Texture.h: * Source/WebGPU/WebGPU/Texture.mm: (WebGPU::Texture::hasStorageBindingCapability): (WebGPU::Device::errorValidatingTextureCreation): (WebGPU::Texture::isDestroyed const): (WebGPU::hasStorageBindingCapability): Deleted. * Source/WebGPU/WebGPU/TextureView.h: * Source/WebGPU/WebGPU/TextureView.mm: (WebGPU::TextureView::usage const): (WebGPU::TextureView::sampleCount const): (WebGPU::TextureView::format const): (WebGPU::TextureView::mipLevelCount const): (WebGPU::TextureView::isDestroyed const): * Source/WebGPU/WebGPU/WebGPU.h: Canonical link: https://commits.webkit.org/272032@main
…LocalFrame lifetime https://bugs.webkit.org/show_bug.cgi?id=266105 rdar://116202445 Reviewed by Pascoe. Before this PR we were calling addRootFrame in an ad-hoc manner, and we weren't calling remoteRootFrame at all except confusingly in the LocalFrame constructor, which was quite incorrect. This caused memory leaks and crashes when site isolation was enabled. To fix this, I call addRootFrame in the LocalFrame constructor and removeRootFrame in the LocalFrame destructor. During the destructor call, we can't reliably check the parent frame because it may have also been destroyed, so I cache the value of whether the LocalFrame is a root frame using a const bool member function. I use the ChromeClient to tell the WebPage's DrawingArea about the creation and destruction of root frames. In order to successfully draw after this root frame lifetime management change, I needed the WebPage to already have a DrawingArea when the main frame is constructed, which required moving it up in the list of constructed members of the WebPage. When I did this, though, the calls to WebPage::corePage in the DrawingArea constructors returned null because the DrawingArea was being constructed before the WebCore::Page. To fix this, I just moved the settings update code from the DrawingArea constructors to the WebPage constructor after creating the WebCore::Page. Additionally, a few members needed to be initialized before the DrawingArea constructor might use them, so I moved them around, and the call to TiledCoreAnimationDrawingArea::setViewExposedRect needs the main frame to have been initialized, so I moved that to after the main frame was initialized. I need the Frame to be destroyed when it transitions between LocalFrame and RemoteFrame, and FocusController was keeping it alive sometimes, which caused crashes. I changed FocusController to have a WeakPtr instead of a RefPtr keeping the Frame alive. When the Frame is destroyed after having called Frame::takeWindowProxyFrom on it, it does some teardown on its WindowProxy. To make that teardown not happen on a WindowProxy that is still being used by another Frame, I added a call to frame.resetWindowProxy() but without the detachFromFrame() call. This is needed to keep the WindowProxy pointing to the right Frame and DOMWindow after the transition when the Frame lifetime is managed better. * Source/WebCore/loader/EmptyClients.h: * Source/WebCore/page/ChromeClient.h: * Source/WebCore/page/FrameTree.h: (WebCore::FrameTree::thisFrame const): * Source/WebCore/page/LocalFrame.cpp: (WebCore::isRootFrame): (WebCore::LocalFrame::LocalFrame): (WebCore::LocalFrame::~LocalFrame): (WebCore::LocalFrame::isRootFrame const): * Source/WebCore/page/LocalFrame.h: * Source/WebCore/page/Page.cpp: (WebCore::createMainFrame): (WebCore::Page::addRootFrame): (WebCore::Page::removeRootFrame): * Source/WebKit/WebProcess/WebCoreSupport/WebChromeClient.cpp: (WebKit::WebChromeClient::rootFrameAdded): (WebKit::WebChromeClient::rootFrameRemoved): * Source/WebKit/WebProcess/WebCoreSupport/WebChromeClient.h: * Source/WebKit/WebProcess/WebPage/CoordinatedGraphics/DrawingAreaCoordinatedGraphics.cpp: (WebKit::DrawingAreaCoordinatedGraphics::DrawingAreaCoordinatedGraphics): * Source/WebKit/WebProcess/WebPage/DrawingArea.h: (WebKit::DrawingArea::removeRootFrame): (WebKit::DrawingArea::shouldForceCompositingMode const): * Source/WebKit/WebProcess/WebPage/RemoteLayerTree/RemoteLayerTreeDrawingArea.h: * Source/WebKit/WebProcess/WebPage/RemoteLayerTree/RemoteLayerTreeDrawingArea.mm: (WebKit::RemoteLayerTreeDrawingArea::RemoteLayerTreeDrawingArea): (WebKit::RemoteLayerTreeDrawingArea::removeRootFrame): * Source/WebKit/WebProcess/WebPage/WebFrame.cpp: (WebKit::WebFrame::transitionToLocal): * Source/WebKit/WebProcess/WebPage/WebPage.cpp: (WebKit::WebPage::WebPage): (WebKit::m_historyItemClient): (WebKit::addRootFramesToNewDrawingArea): (WebKit::WebPage::reinitializeWebPage): * Source/WebKit/WebProcess/WebPage/WebPage.h: * Source/WebKit/WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm: (WebKit::TiledCoreAnimationDrawingArea::TiledCoreAnimationDrawingArea): * Source/WebKitLegacy/mac/WebCoreSupport/WebChromeClient.h: Canonical link: https://commits.webkit.org/272224@main
https://bugs.webkit.org/show_bug.cgi?id=244054 Reviewed by Jonathan Bedard. * Tools/CISupport/ews-app/ews/common/github.py: (GitHubEWS): Canonical link: https://commits.webkit.org/272225@main
https://bugs.webkit.org/show_bug.cgi?id=260039 Reviewed by Justin Michaud. Use subtyping for module linking in general. Also add and re-enable more tests about linking. Note: we don't add subtyping for exception tags for now, because the exception proposal is in flux. * JSTests/wasm/gc-spec-tests/type-subtyping.wast.js: * JSTests/wasm/gc/linking.js: Added. (testLinking): * Source/JavaScriptCore/wasm/js/WebAssemblyModuleRecord.cpp: (JSC::WebAssemblyModuleRecord::initializeImports): Canonical link: https://commits.webkit.org/272226@main
… set of out-of-flow boxes https://bugs.webkit.org/show_bug.cgi?id=266580 Reviewed by Antti Koivisto. Replace this visual tree dump with slightly more flexible, inline content. * LayoutTests/TestExpectations: * LayoutTests/fast/ruby/ruby-expansion-cjk-2-expected.html: * LayoutTests/fast/ruby/ruby-expansion-cjk-2.html: * LayoutTests/fast/ruby/ruby-expansion-cjk-3-expected.html: * LayoutTests/fast/ruby/ruby-expansion-cjk-3.html: * LayoutTests/fast/ruby/ruby-expansion-cjk-5-expected.html: * LayoutTests/fast/ruby/ruby-expansion-cjk-5.html: * LayoutTests/fast/ruby/ruby-expansion-cjk-expected.html: * LayoutTests/fast/ruby/ruby-expansion-cjk.html: * LayoutTests/fast/ruby/ruby-justification-expected.html: * LayoutTests/fast/ruby/ruby-justification.html: Canonical link: https://commits.webkit.org/272227@main
https://bugs.webkit.org/show_bug.cgi?id=266583 Reviewed by Don Olmstead. Build fix -- revert 263486@main and 263783@main for PlayStation. * Source/WebCore/platform/network/curl/CurlContext.cpp: (WebCore::CurlHandle::enableHttp): * Source/WebCore/platform/network/curl/CurlRequest.cpp: (WebCore::CurlRequest::didReceiveHeader): (WebCore::CurlRequest::didReceiveData): * Source/cmake/OptionsPlayStation.cmake: Canonical link: https://commits.webkit.org/272228@main
…er double tapping trackpad causes page to scroll down https://bugs.webkit.org/show_bug.cgi?id=266529 rdar://116419609 Reviewed by Tim Horton. The smart magnify gesture (two-finger double tap on the trackpad) results in some combination of scrolling and zooming. `RemoteLayerTreeDrawingAreaProxyMac::commitTransientZoom()` uses a CAAnimation to animate to the final state, but the completion handler of that animation put the transform directly onto the `layerForPageScale`. That's OK if the gesture results in a scale change; that transform will get replaced by one that comes from the web process. However, if a gesture just resulted in a scroll with no scale change, we never get a new transform from the web process, so we leave that transform on `layerForPageScale`. This transform has a translation baked in, so this causes a permanent scroll offset. So we have to stop shoving the transform onto this layer. This revealed an issue where there was a jump when the animation was removed. Fixing this just requires two things: 1. Delay removing the transient zoom animations until we know the web process has committed the new scale and scroll position (via a `callAfterNextPresentationUpdate()`). 2. Override any scroll position changes that come from the web process before this by adding a CAAnimation onto the scrolled contents layer which overrides the `position` property. If we don't do this, scroll position changes cause a flash of offset content. This requires tracking m_pageScrollingLayerID, but that information was already in the transaction for banner hookup. * LayoutTests/view-gestures/smart-magnify/double-tap-zoom-scroll-above-top-expected.html: Added. * LayoutTests/view-gestures/smart-magnify/double-tap-zoom-scroll-above-top.html: Added. * Source/WebKit/UIProcess/RemoteLayerTree/mac/RemoteLayerTreeDrawingAreaProxyMac.h: * Source/WebKit/UIProcess/RemoteLayerTree/mac/RemoteLayerTreeDrawingAreaProxyMac.mm: (WebKit::RemoteLayerTreeDrawingAreaProxyMac::didCommitLayerTree): (WebKit::fillFowardsAnimationWithKeyPath): (WebKit::transientZoomTransformOverrideAnimation): (WebKit::RemoteLayerTreeDrawingAreaProxyMac::applyTransientZoomToLayer): (WebKit::RemoteLayerTreeDrawingAreaProxyMac::removeTransientZoomFromLayer): (WebKit::RemoteLayerTreeDrawingAreaProxyMac::commitTransientZoom): (WebKit::RemoteLayerTreeDrawingAreaProxyMac::sendCommitTransientZoom): Canonical link: https://commits.webkit.org/272229@main
https://bugs.webkit.org/show_bug.cgi?id=266562 rdar://119793659 Reviewed by Pascoe. Add a test to make sure the feature `verifyWindowOpenUserGestureFromUIProcess` works correctly when a mouse event is used to open a window. The import in WKWebExtensionAPIDeclarativeNetRequest.mm is needed for unified sources. * Tools/TestWebKitAPI/SourcesCocoa.txt: * Tools/TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: * Tools/TestWebKitAPI/Tests/WebKitCocoa/VerifyUserGestureFromUIProcess.mm: Added. (TestWebKitAPI::TEST): * Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebExtensionAPIDeclarativeNetRequest.mm: Canonical link: https://commits.webkit.org/272230@main
https://bugs.webkit.org/show_bug.cgi?id=264737 rdar://117805319 Reviewed by Brent Fulgham. Until now, when we are purging inactive font data, we would just clear the glyph page cache if we had to purge system fallback font. This means that we consider glyph page cache would only point to fonts from system fonts fallback. When we are handling unicode's in the Private-User-Area (PUA) block, we shouldn't fallback to system fonts searching for a font that can render it, per spec: https://www.w3.org/TR/css-fonts-4/#char-handling-issues Instead, we render the glyph 0 with the last resort font. However, this font is just added to the custom font cache, and its font pointer in the Glyph Page cache is not cleared during memory pressure. We should add this font to the system font fallback set, to make sure that the associated font pointer is removed from the glyph page cache during memory pressure. * LayoutTests/fonts/font-cache-memory-pressure-crash.html: Added. * Source/WebCore/platform/graphics/FontCascadeFonts.cpp: (WebCore::FontCascadeFonts::glyphDataForVariant): * LayoutTests/fonts/font-cache-memory-pressure-crash-expected.txt: Added. Originally-landed-as: a595ddd8348d. rdar://119598929 Canonical link: https://commits.webkit.org/272231@main
https://bugs.webkit.org/show_bug.cgi?id=266585 Reviewed by Tim Nguyen. Link to spec was broken; fixed it. * Source/WebCore/rendering/RenderGrid.cpp: (WebCore::RenderGrid::placeSpecifiedMajorAxisItemsOnGrid): Canonical link: https://commits.webkit.org/272232@main
https://bugs.webkit.org/show_bug.cgi?id=266578 rdar://119816733 Reviewed by Aditya Keerthi. Prepare our API tests for async text input enablement on iOS, many of which currently call into legacy SPI methods that crash with the feature flag enabled. To avoid these release assertions, we refactor all tests which explicitly call into these codepaths, such that they'll invoke their `UIAsyncTextInput` counterparts when `UIKit/async_text_input` is enabled. See below for more details. * Source/WebKit/UIProcess/ios/WKContentViewInteraction.mm: (-[WKContentView moveInDirection:byGranularity:]): (-[WKContentView extendInDirection:byGranularity:]): Also drive-by fix an actual bug that was caught by `SelectionTests.ModifyByParagraphBoundary`. Instead of just moving backward or forward by paragraph granularity in the `UIAsyncTextInput` method `-extendInDirection:byGranularity:`, we need to additionally move backwards or forwards by character before doing so in order to maintain parity with `-_moveToEndOfParagraph:withHistory:` and `_moveToStartOfParagraph:withHistory:`. * Source/WebKit/UIProcess/ios/WKSEDefinitions.h: Renamed from Source/WebKit/UIProcess/API/ios/WKSEDefinitions.h. Also move `WKSEDefinitions.h` to `UIProcess/ios` instead of `UIProcess/API/ios`, since the latter just holds iOS-specific files related to WebKit API objects (in this case, only `WKWebView`). Putting this in `UIProcess/ios` seems more appropriate, and also allows us to include it directly in WebKitTestRunner. * Source/WebKit/WebKit.xcodeproj/project.pbxproj: * Tools/TestRunnerShared/spi/UIKitSPIForTesting.h: * Tools/TestWebKitAPI/Tests/WebKitCocoa/DocumentEditingContext.mm: (-[TestWKWebView synchronouslyAdjustSelectionWithDelta:]): Deleted. * Tools/TestWebKitAPI/Tests/WebKitCocoa/ImageAnalysisTests.mm: (TestWebKitAPI::simulateEditContextMenuAppearance): * Tools/TestWebKitAPI/Tests/WebKitCocoa/WKContentViewEditingActions.mm: * Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebViewEditActions.mm: * Tools/TestWebKitAPI/Tests/WebKitCocoa/iOSMouseSupport.mm: (TestWebKitAPI::simulateEditContextMenuAppearance): * Tools/TestWebKitAPI/Tests/ios/AutocorrectionTestsIOS.mm: (-[TestWKWebView autocorrectionRectsForString:]): Deleted. (-[TestWKWebView synchronouslyRequestAutocorrectionContext]): Deleted. * Tools/TestWebKitAPI/Tests/ios/EnterKeyHintTests.mm: (-[TestWKWebView test:enterKeyHint:returnKeyType:]): * Tools/TestWebKitAPI/Tests/ios/KeyboardInputTestsIOS.mm: * Tools/TestWebKitAPI/Tests/ios/SelectionModifyByParagraphBoundary.mm: * Tools/TestWebKitAPI/Tests/ios/TextServicesTests.mm: * Tools/TestWebKitAPI/Tests/ios/WKScrollViewTests.mm: * Tools/TestWebKitAPI/cocoa/TestWKWebView.h: * Tools/TestWebKitAPI/cocoa/TestWKWebView.mm: Add a bunch of new category helpers on `WKWebView`, which replace many of the direct calls into `WKContentView` in our API tests when exercising various `UIWKInteractionViewProtocol` SPIs. These helper methods will call directly into these legacy SPIs when async text input is disabled, but will call into their new `UIAsyncTextInput` counterparts when async text input is enabled. (-[WKWebView defineSelection]): (-[WKWebView shareSelection]): (-[WKWebView hasAsyncTextInput]): (-[WKWebView selectionClipRect]): (-[WKWebView moveSelectionToStartOfParagraph]): (-[WKWebView extendSelectionToStartOfParagraph]): (-[WKWebView moveSelectionToEndOfParagraph]): (-[WKWebView extendSelectionToEndOfParagraph]): (-[WKWebView asyncTextInput]): (-[WKWebView extendedTextInputTraits]): (-[WKWebView autocorrectionRectsForString:]): (-[WKWebView autocorrectionContext]): (-[WKWebView effectiveTextInputTraits]): (-[WKWebView replaceText:withText:shouldUnderline:completion:]): (wrap): (unwrap): (-[WKWebView synchronouslyAdjustSelectionWithDelta:]): (-[WKWebView selectTextForContextMenuWithLocationInView:completion:]): (-[WKWebView handleKeyEvent:completion:]): * Tools/WebKitTestRunner/Configurations/Base.xcconfig: * Tools/WebKitTestRunner/ios/UIScriptControllerIOS.h: * Tools/WebKitTestRunner/ios/UIScriptControllerIOS.mm: (WTR::UIScriptControllerIOS::asyncTextInput const): Canonical link: https://commits.webkit.org/272233@main
https://bugs.webkit.org/show_bug.cgi?id=265432 <radar://118866310> Reviewed by Dan Glastonbury. https://webkit.org/demos/webgpu/ was pointing to old samples which do not function as they were built against outdated versions of the specification. Update index.html to direct to the publically maintained GitHub samples which function in ToT builds of WebKit. * Websites/webkit.org/demos/webgpu/index.html: * Websites/webkit.org/demos/webgpu/resources/animometer.png: Added. * Websites/webkit.org/demos/webgpu/resources/babylon-helmet.png: Removed. * Websites/webkit.org/demos/webgpu/resources/babylon-sphere.png: Removed. * Websites/webkit.org/demos/webgpu/resources/bitonic-sort.png: Added. * Websites/webkit.org/demos/webgpu/resources/compute-blur.png: * Websites/webkit.org/demos/webgpu/resources/compute-boids.png: * Websites/webkit.org/demos/webgpu/resources/hello-cube.png: * Websites/webkit.org/demos/webgpu/resources/hello-triangle.png: * Websites/webkit.org/demos/webgpu/resources/motionmark.jpeg: Removed. * Websites/webkit.org/demos/webgpu/resources/photon-mapping.png: Added. * Websites/webkit.org/demos/webgpu/resources/textured-cube.png: * Websites/webkit.org/demos/webgpu/resources/video-uploading.png: Added. * Websites/webkit.org/demos/webgpu/resources/webkit-logo.png: Removed. Canonical link: https://commits.webkit.org/272234@main
* LayoutTests/platform/wincairo/TestExpectations: Canonical link: https://commits.webkit.org/272235@main
…KWebExtensionAPITabs.SendMessageWithPromiseReply timeout https://webkit.org/b/266518 rdar://problem/119740501 Reviewed by Brian Weinstein. Adopt EagerCallbackAggregator for the reply handler. This was added on the UIProcess side for the message sending in 271636@main. It can also be used here, with the help of an ObjC wrapper object that can keep the aggregator in scope for the required blocks. We can't use BlockPtr for these since the JSValue ObjC translation requires blocks to have a signature to translate them into JavaScript functions, which is only for compiled blocks. Without this, the aggregator falls out of scope and is released. Added additional tests to make sure async replies work in addition to sync replies. Also adopted optional chaining for some tests since an exception is thrown when response is null. * Source/WebKit/WebProcess/Extensions/API/Cocoa/WebExtensionAPIRuntimeCocoa.mm: (-[_WKReplyCallbackAggregator initWithAggregator:]): Added. (-[_WKReplyCallbackAggregator aggregator]): Added. (WebKit::WebExtensionContextProxy::internalDispatchRuntimeMessageEvent): * Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebExtensionAPIRuntime.mm: (TestWebKitAPI::TEST): * Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebExtensionAPITabs.mm: (TestWebKitAPI::TEST): Canonical link: https://commits.webkit.org/272236@main
https://webkit.org/b/261848 rdar://problem/115809590 Unreviewed test gardening. Enable the WKWebExtensionAPITabs.RemovedEvent test, since it no longer fails after hundreds of iterations. * Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebExtensionAPITabs.mm: (TestWebKitAPI::TEST): Canonical link: https://commits.webkit.org/272237@main
… tests in inspector/timeline are flaky crash with ASSERTION FAILED in void WebCore::InspectorTimelineAgent::didCompleteCurrentRecord(TimelineRecordType) https://bugs.webkit.org/show_bug.cgi?id=266591 rdar://119828743 Unreviewed test gardening. Adding test expectation * LayoutTests/platform/mac-wk2/TestExpectations: Canonical link: https://commits.webkit.org/272238@main
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This brings qtwebkit up to date with upstream webkit as of 2023-12-18.