upgrade webkit 20251007 #113
Draft
+89,729
−86,708
Conversation
This file contains hidden or 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
https://bugs.webkit.org/show_bug.cgi?id=300011 rdar://161796662 Reviewed by Richard Robinson. Comments on WebKit#51245 suggest that this is a better way to structure this modulemap. Canonical link: https://commits.webkit.org/300948@main
https://bugs.webkit.org/show_bug.cgi?id=300069 Reviewed by Timothy Hatcher. Instead of decoding data: URLs manually, use the WebCore::DataURLDecoder to decode data: urls, and return the resulting URL. * Source/WebKit/UIProcess/Extensions/Cocoa/WebExtensionCocoa.mm: (WebKit::WebExtension::resourceDataForPath): Canonical link: https://commits.webkit.org/300949@main
https://bugs.webkit.org/show_bug.cgi?id=300073 rdar://161868858 Reviewed by Timothy Hatcher. In 299859@main I introduced a lifetime model of _WKJSHandle that allowed them to be put into arrays and dictionaries with later messages from JS able to remove the same object representation from the containers. This worked well, except it introduced a race condition if you are rapidly creating and destroying _WKJSHandles: If a _WKJSHandle is destroyed near the same time when the web process is sending a message to the UI process with the same handle, the web process would receive a message saying the handle was destroyed near the same time as when the UI process receives a message with the _WKJSHandle, which would then fail when used because the JSC::Strong had already been removed from the map, so lookup with the identifier would fail. This caused assertions when running TestWebKitAPI.SiteIsolation.HitTesting: ASSERTION FAILED: objectMap().contains(strong.get()) .../Source/WebCore/page/WebKitJSHandle.cpp(71) : static void WebCore::WebKitJSHandle::jsHandleDestroyed(JSHandleIdentifier) I fix this issue by removing the unneeded === equality in JS and pointer equality in ObjC while maintaining the isEqual equality in ObjC, which is what really mattered. I make a ref counting system that allows multiple handles to point to the same object and be destroyed whenever they are destroyed. I also rename the JS entry point from jsHandle back to createJSHandle because we are creating new handles each time again. Tests: Tools/TestWebKitAPI/Tests/WebKitCocoa/JSHandle.mm Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebViewSnapshot.mm Tools/TestWebKitAPI/Tests/WebKitCocoa/WebTransport.mm * Source/WebCore/page/WebKitJSHandle.cpp: (WebCore::WebKitJSHandle::create): (WebCore::WebKitJSHandle::jsHandleDestroyed): (WebCore::WebKitJSHandle::WebKitJSHandle): (WebCore::objectMap): Deleted. (WebCore::WebKitJSHandle::getOrCreate): Deleted. * Source/WebCore/page/WebKitJSHandle.h: * Source/WebCore/page/WebKitNamespace.cpp: (WebCore::WebKitNamespace::createJSHandle): (WebCore::WebKitNamespace::jsHandle): Deleted. * Source/WebCore/page/WebKitNamespace.h: * Source/WebCore/page/WebKitNamespace.idl: * Source/WebKit/Shared/JSHandleInfo.cpp: (WebKit::JSHandleInfo::operator== const): * Source/WebKit/Shared/JSHandleInfo.h: * Source/WebKit/Shared/JavaScriptEvaluationResult.cpp: (WebKit::JavaScriptEvaluationResult::APIInserter::toAPI): * Source/WebKit/Shared/JavaScriptEvaluationResult.mm: (WebKit::JavaScriptEvaluationResult::ObjCInserter::toID): * Source/WebKit/UIProcess/API/APIJSHandle.cpp: (API::JSHandle::create): (API::JSHandle::JSHandle): (API::JSHandle::~JSHandle): (API::handleMap): Deleted. (API::JSHandle::getOrCreate): Deleted. * Source/WebKit/UIProcess/API/APIJSHandle.h: * Source/WebKit/UIProcess/API/Cocoa/WKWebView.mm: (-[WKWebView _hitTestAtPoint:inFrameCoordinateSpace:completionHandler:]): * Source/WebKit/UIProcess/API/Cocoa/_WKContentWorldConfiguration.h: * Source/WebKit/UIProcess/API/Cocoa/_WKJSHandle.h: * Source/WebKit/UIProcess/API/Cocoa/_WKJSHandle.mm: (-[_WKJSHandle isEqual:]): (-[_WKJSHandle hash]): (-[_WKJSHandle copyWithZone:]): * Source/WebKit/WebProcess/WebPage/WebPage.cpp: (WebKit::WebPage::hitTestAtPoint): * Tools/TestWebKitAPI/Tests/WebKitCocoa/JSHandle.mm: (TestWebKitAPI::TEST(JSHandle, Basic)): (TestWebKitAPI::TEST(JSHandle, Equality)): (TestWebKitAPI::TEST(JSHandle, WebpagePreferences)): * Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebViewSnapshot.mm: (TestWebKitAPI::TEST(WKWebView, SnapshotNodeByJSHandle)): * Tools/TestWebKitAPI/Tests/WebKitCocoa/WebTransport.mm: (TestWebKitAPI::TEST(WebTransport, CreateStreamsBeforeReady)): * Tools/WebKitTestRunner/TestController.cpp: Canonical link: https://commits.webkit.org/300950@main
https://bugs.webkit.org/show_bug.cgi?id=300114 rdar://161902154 Unreviewed. The test was timing out sometimes. The test doesn't need to check that stream closing works, it just needs to test that the connection succeeds. Simplify the test to make it stop timing out. * Tools/TestWebKitAPI/Tests/WebKitCocoa/WebTransport.mm: (TestWebKitAPI::TEST(WebTransport, ServerCertificateHashes)): Canonical link: https://commits.webkit.org/300951@main
https://bugs.webkit.org/show_bug.cgi?id=300106 rdar://161895900 Reviewed by Yusuke Suzuki. When zero initializing locals, the full 16-byte stack slot needs to be zeroed since the local could be a v128. * JSTests/wasm/stress/simd-get-set.js: (Test.that.uninitialized.v128.local.is.fully.zeroed.all.16.bytes.func.async test): (Test.global.and.global.func.async test.setI32x4): Deleted. (Test.global.and.global.func.async test.getI32x4): Deleted. (Test.global.and.global.func.async test): Deleted. * Source/JavaScriptCore/llint/InPlaceInterpreter64.asm: Canonical link: https://commits.webkit.org/300952@main
https://bugs.webkit.org/show_bug.cgi?id=298896 rdar://160640689 Reviewed by Jonathan Bedard. From buildbot v4.1.0 onwards: getBuilder and getBuilders now return a BuilderModel instead of a dictionary. Reference: https://docs.buildbot.net/current/relnotes/index.html This change should be compatible with both older and newer buildbot versions. * Tools/CISupport/ews-build/events.py: (Events.buildStarted): (Events.buildFinished): Canonical link: https://commits.webkit.org/300953@main
https://bugs.webkit.org/show_bug.cgi?id=300101 Reviewed by Ryosuke Niwa. Files that were not unskipped have false positives that are currently being addressed in the checker. * Source/WebKit/Platform/mac/MenuUtilities.h: * Source/WebKit/Platform/mac/MenuUtilities.mm: (WebKit::menuItemTitleForTelephoneNumberGroup): * Source/WebKit/SaferCPPExpectations/UnretainedCallArgsCheckerExpectations: * Source/WebKit/UIProcess/API/APIAttachment.h: * Source/WebKit/UIProcess/API/Cocoa/APIAttachmentCocoa.mm: (API::Attachment::associatedElementNSData const): * Source/WebKit/UIProcess/mac/WebContextMenuProxyMac.mm: (WebKit::WebContextMenuProxyMac::setupServicesMenu): (WebKit::menuItemIdentifier): (WebKit::createMenuActionItem): (WebKit::WebContextMenuProxyMac::getContextMenuItem): (WebKit::WebContextMenuProxyMac::window const): * Source/WebKit/UIProcess/mac/WebPopupMenuProxyMac.mm: (WebKit::WebPopupMenuProxyMac::showPopupMenu): Canonical link: https://commits.webkit.org/300954@main
https://bugs.webkit.org/show_bug.cgi?id=300021 Reviewed by Justin Michaud. This makes the intent clearer and means the 32-bit version can stay closer to the 64-bit code. Canonical link: https://commits.webkit.org/300955@main
https://bugs.webkit.org/show_bug.cgi?id=300042 Reviewed by Antoine Quint and Antti Koivisto. Started adding property specific types for AcceleratedEffect values, initially focusing on removing the remaining uses of WebCore::Length and WebCore::LengthPoint in AcceleratedEffectValues. Added layering violation comments to indicate further work needed to resolve the layering. * Source/WebCore/CMakeLists.txt: * Source/WebCore/Headers.cmake: * Source/WebCore/Sources.txt: * Source/WebCore/WebCore.xcodeproj/project.pbxproj: * Source/WebCore/platform/LengthFunctions.cpp: * Source/WebCore/platform/LengthFunctions.h: * Source/WebCore/platform/animation/AcceleratedEffect.cpp: * Source/WebCore/platform/animation/AcceleratedEffectValues.cpp: * Source/WebCore/platform/animation/AcceleratedEffectValues.h: * Source/WebCore/platform/animation/AnimationUtilities.h: * Source/WebCore/platform/animation/values/AcceleratedEffectOffsetAnchor.cpp: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectOffsetAnchor.h: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectOffsetDistance.cpp: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectOffsetDistance.h: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectOffsetPosition.cpp: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectOffsetPosition.h: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectOffsetRotate.cpp: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectOffsetRotate.h: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectOpacity.cpp: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectOpacity.h: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectTransformBox.h: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectTransformOrigin.cpp: Added. * Source/WebCore/platform/animation/values/AcceleratedEffectTransformOrigin.h: Added. * Source/WebCore/rendering/MotionPath.cpp: * Source/WebCore/rendering/MotionPath.h: * Source/WebCore/rendering/style/RenderStyle.cpp: * Source/WebCore/rendering/style/StyleMiscNonInheritedData.h: * Source/WebCore/rendering/style/StyleRareNonInheritedData.h: * Source/WebCore/style/StyleExtractorConverter.h: * Source/WebCore/style/StyleExtractorSerializer.h: * Source/WebCore/style/values/color/StyleOpacity.h: * Source/WebCore/style/values/motion/StyleOffsetAnchor.cpp: * Source/WebCore/style/values/motion/StyleOffsetAnchor.h: * Source/WebCore/style/values/motion/StyleOffsetDistance.h: * Source/WebCore/style/values/motion/StyleOffsetPosition.cpp: * Source/WebCore/style/values/motion/StyleOffsetPosition.h: * Source/WebCore/style/values/motion/StyleOffsetRotate.cpp: * Source/WebCore/style/values/motion/StyleOffsetRotate.h: * Source/WebCore/style/values/primitives/StylePosition.cpp: * Source/WebCore/style/values/primitives/StylePosition.h: * Source/WebKit/Shared/WebCoreArgumentCoders.serialization.in: * Source/WebKit/UIProcess/RemoteLayerTree/RemoteAcceleratedEffectStack.mm: Canonical link: https://commits.webkit.org/300956@main
…ed symbol "WebCore::Node::document() const" rdar://161863117 https://bugs.webkit.org/show_bug.cgi?id=300121 Unreviewed build fix; add missing include. * Source/WebKit/WebProcess/Plugins/PDF/UnifiedPDF/UnifiedPDFPlugin.mm: Canonical link: https://commits.webkit.org/300957@main
https://bugs.webkit.org/show_bug.cgi?id=299702 Reviewed by Darin Adler. Converts the 'font-size-adjust' property to use strong style types. * Source/WebCore/Headers.cmake: * Source/WebCore/Sources.txt: * Source/WebCore/WebCore.xcodeproj/project.pbxproj: * Source/WebCore/css/CSSProperties.json: * Source/WebCore/platform/graphics/FontCascadeDescription.h: * Source/WebCore/platform/graphics/FontDescription.cpp: * Source/WebCore/rendering/style/RenderStyle.cpp: * Source/WebCore/rendering/style/RenderStyle.h: * Source/WebCore/rendering/style/RenderStyleInlines.h: * Source/WebCore/style/StyleBuilderConverter.h: * Source/WebCore/style/StyleBuilderCustom.h: * Source/WebCore/style/StyleBuilderState.h: * Source/WebCore/style/StyleBuilderStateInlines.h: * Source/WebCore/style/StyleExtractorConverter.h: * Source/WebCore/style/StyleExtractorSerializer.h: * Source/WebCore/style/StyleFontSizeFunctions.cpp: * Source/WebCore/style/StyleFontSizeFunctions.h: * Source/WebCore/style/StyleInterpolationFunctions.h: * Source/WebCore/style/StyleInterpolationWrappers.h: * Source/WebCore/style/StyleResolveForFont.cpp: * Source/WebCore/style/StyleResolveForFont.h: * Source/WebCore/style/values/fonts/StyleFontSizeAdjust.cpp: Added. * Source/WebCore/style/values/fonts/StyleFontSizeAdjust.h: Added. Canonical link: https://commits.webkit.org/300958@main
WebKit-7623.1.11 Canonical link: https://commits.webkit.org/300959@main
https://bugs.webkit.org/show_bug.cgi?id=299488 rdar://161273712 Reviewed by Daniel Liu. As announced on September 9th, the SoCs used in the next generation of iPhones will include support for ARM'S Memory Tagging Extension functionality. As part of Apple's MIE (Memory Integrity Enforcement) feature, libpas should thus implement support for MTE and related memory-safety functionality to ensure that WebKit is up to par with the new memory safety standards set by the rest of the system. In particular, this patch ensures that when possible we allocate memory with backing MTE tag pages and tag allocations made within them prior to returning allocation memory to the caller. Not all memory can be tagged this way: in particular, objects >= 32K and objects which may be referenced via compact pointers cannot be MTE tagged. There are other exceptions as well, depending on process/object-type/platform. It also implements a variety of hardening strategies to further strengthen the feature and prevent certain well-known kinds of attacks. Trying this again, since the previous attempt caused build breakages in certain configurations. Canonical link: https://commits.webkit.org/300960@main
https://bugs.webkit.org/show_bug.cgi?id=300058 rdar://problem/161856764 Reviewed by Alan Baradlay. This patch adds the overall logic that we will use to support aligning grid items after we have determined their sizes. Note that since alignment is done with respect to the grid area, which acts as the containing block for the grid item, the border box position that is returned will be in the coordinate space of the grid area. With this patch, we only support start alignment for the flex items, so there is realistically no movement that is needed for the grid item's margin box, so the final border box position will just be the same as the value of the top margin. In order to support other alignment values, we will likely need the size of the grid area along with the border box dimensions of the grid item. * Source/WebCore/layout/formattingContexts/grid/GridFormattingContext.cpp: (WebCore::Layout::GridFormattingContext::constructPlacedGridItems const): In order to perform alignment, we will need the align-self and justify-self values for the grid item. If the value for this property is not auto, then that is the value that will be used; otherwise, we refer to the value of align-items/justify-items on the grid. * Source/WebCore/layout/formattingContexts/grid/GridLayout.cpp: (WebCore::Layout::GridLayout::layout): (WebCore::Layout::GridLayout::performInlineAxisSelfAlignment): (WebCore::Layout::GridLayout::performBlockAxisSelfAlignment): The basic idea between these two functions is the same: first, we need to find out what the margin box position will be to satisfy the needs of the alignment value, and from there, we can just add in the start margin to get the final position of the item's border box. Since the most basic version of start alignment (without considering any sort of safety) does not move the item from the start of the grid area, we can just return a value of 0 to indicate that it is at the start of the grid area. Canonical link: https://commits.webkit.org/300961@main
https://bugs.webkit.org/show_bug.cgi?id=300054 rdar://161853921 Reviewed by Yusuke Suzuki. Add support for GC struct and arrays with v128 field and element type. Testing: run JSTests/wasm/gc/simd.js with --useWasmIPIntSIMD=true * Source/JavaScriptCore/llint/InPlaceInterpreter64.asm: * Source/JavaScriptCore/wasm/WasmIPIntSlowPaths.cpp: (JSC::IPInt::WASM_IPINT_EXTERN_CPP_DECL): * Source/JavaScriptCore/wasm/WasmIPIntSlowPaths.h: * Source/JavaScriptCore/wasm/WasmOperationsInlines.h: (JSC::Wasm::arrayNew): (JSC::Wasm::tryCopyElementsInReverse): (JSC::Wasm::arrayNewFixed): (JSC::Wasm::arrayGet): (JSC::Wasm::arraySet): (JSC::Wasm::structNew): (JSC::Wasm::structGet): (JSC::Wasm::structSet): * Source/JavaScriptCore/wasm/js/JSWebAssemblyArray.h: * Source/JavaScriptCore/wasm/js/JSWebAssemblyArrayInlines.h: (JSC::JSWebAssemblyArray::getVector): * Source/JavaScriptCore/wasm/js/JSWebAssemblyStruct.cpp: (JSC::JSWebAssemblyStruct::get const): (JSC::JSWebAssemblyStruct::getVector const): * Source/JavaScriptCore/wasm/js/JSWebAssemblyStruct.h: Canonical link: https://commits.webkit.org/300962@main
…axedSIMD=1 https://bugs.webkit.org/show_bug.cgi?id=300112 rdar://161901342 Reviewed by Daniel Liu. Let's not block enabling --useWasmIPIntSIMD by default on relaxed SIMD support in IPInt. However, we won't be able to remove the path that tiers up immediately to BBQ until the relaxed SIMD support is added to IPInt (or we stop testing --useWasmRelaxedSIMD=1 until the relaxed SIMD support is completed). * JSTests/wasm/stress/simd-const-relaxed-f32-madd.js: * JSTests/wasm/stress/simd-const-relaxed-f32-trunc.js: * JSTests/wasm/stress/simd-const-relaxed-f64-madd.js: * JSTests/wasm/stress/simd-const-relaxed-f64-trunc.js: * JSTests/wasm/stress/simd-const-relaxed-lane-select.js: * JSTests/wasm/stress/simd-const-relaxed-swizzle.js: * Source/JavaScriptCore/runtime/Options.cpp: (JSC::Options::assertOptionsAreCoherent): Canonical link: https://commits.webkit.org/300963@main
…ocess cache https://bugs.webkit.org/show_bug.cgi?id=299916 Reviewed by Ben Nham. Made shared process for site isolation work with the Web process cache. WebProcessCache is extended to maintain a separate map of shared processes per top-level domain in addition to the existing cache of "main" processes. The size and the capacity of WebProcessCache now takes into account both types of the process caches. In addition to being segregated by the top-level domain, when a given domain receives an user interaction from the user, the existing shared process cache which had loaded the same domain will no longer be eligible for the cache reuse. This PR also fixes the bug that a wrong BrowsingContextGroup is used during a back/forward navigation. Store the relevant BrowsingContextGroup in WebBackForwardListItem and use this BrowsingContextGroup when back/forward navigating to the item. In addition, this PR also fixes the bug in WebPageProxy::receivedNavigationActionPolicyDecision that we were creating a new BrowsingContextGroup even when the result of redirect chain ended up navigating back to the same site. We now avoid creating a new BrowsingContextGroup when main frame's site did not change per navigation. Test: TestWebKitAPI.SiteIsolation.SharedProcessBasicWebProcessCache TestWebKitAPI.SiteIsolation.SharedProcessWebProcessCacheSharedProcessForSiteWithUserInteraction Test: Tools/TestWebKitAPI/Tests/WebKitCocoa/SiteIsolation.mm * Source/WebKit/Shared/WebBackForwardListItem.cpp: (WebKit::WebBackForwardListItem::create): Optionally store BrowsingContextGroup. (WebKit::WebBackForwardListItem::WebBackForwardListItem): * Source/WebKit/Shared/WebBackForwardListItem.h: (WebKit::WebBackForwardListItem::browsingContextGroup const): * Source/WebKit/UIProcess/BrowsingContextGroup.cpp: (WebKit::BrowsingContextGroup::sharedProcessForSite): (WebKit::BrowsingContextGroup::ensureProcessForSite): * Source/WebKit/UIProcess/BrowsingContextGroup.h: * Source/WebKit/UIProcess/FrameProcess.cpp: (WebKit::FrameProcess::FrameProcess): * Source/WebKit/UIProcess/FrameProcess.h: (WebKit::FrameProcess::sharedProcessMainFrameSite const): (WebKit::FrameProcess::create): * Source/WebKit/UIProcess/WebBackForwardList.cpp: (WebKit::WebBackForwardList::backForwardAddItemShared): * Source/WebKit/UIProcess/WebFrameProxy.cpp: (WebKit::WebFrameProxy::prepareForProvisionalLoadInProcess): * Source/WebKit/UIProcess/WebPageProxy.cpp: (WebKit::WebPageProxy::launchProcess): (WebKit::WebPageProxy::initializeWebPage): (WebKit::WebPageProxy::receivedNavigationActionPolicyDecision): Fixed a subtle bug that this code was using a wrong BrowsingContextGroup to find the shared process when the navigation is supposed to swap the BrowsingContextGroup. (WebKit::WebPageProxy::continueNavigationInNewProcess): (WebKit::WebPageProxy::triggerBrowsingContextGroupSwitchForNavigation): * Source/WebKit/UIProcess/WebPageProxy.h: * Source/WebKit/UIProcess/WebProcessCache.cpp: Improved the logging code in this file to use SENSITIVE_LOG_STRING of PRIVATE_LOG_STRING since it logs the site name. (WebKit::WebProcessCache::canCacheProcess const): (WebKit::WebProcessCache::addProcess): (WebKit::WebProcessCache::evictAtRandomIfNeeded): Extracted from addProcess. We try to evict a shared process before trying to evict the "main" process since the main process needs to launch sooner. (WebKit::WebProcessCache::takeProcess): (WebKit::WebProcessCache::takeSharedProcess): Added. * Source/WebKit/UIProcess/WebProcessCache.h: * Source/WebKit/UIProcess/WebProcessPool.cpp: (WebKit::WebProcessPool::processForSite): Added the logic to reuse a cached shared process. (WebKit::WebProcessPool::createWebPage): (WebKit::WebProcessPool::processForNavigation): (WebKit::WebProcessPool::prepareProcessForNavigation): (WebKit::WebProcessPool::processForNavigationInternal): * Source/WebKit/UIProcess/WebProcessPool.h: * Source/WebKit/UIProcess/WebProcessProxy.cpp: (WebKit::WebProcessProxy::addSharedProcessDomain): Added. (WebKit::WebProcessProxy::didStartUsingProcessForSiteIsolation): * Source/WebKit/UIProcess/WebProcessProxy.h: * Tools/TestWebKitAPI/Tests/WebKitCocoa/SiteIsolation.mm: (TestWebKitAPI::siteIsolatedViewWithSharedProcess): (TestWebKitAPI::(SiteIsolation, SharedProcessBasicWebProcessCache)): Added (TestWebKitAPI::(SiteIsolation, SharedProcessWithResourceLoadStatistics)): (TestWebKitAPI::(SiteIsolation, SharedProcessAfterClick)): (TestWebKitAPI::(SiteIsolation, SharedProcessAfterKeyDown)): (TestWebKitAPI::(SiteIsolation, SharedProcessAfterUserInteractionInSharedProcesss)): (TestWebKitAPI::(SiteIsolation, SharedProcessWebProcessCacheSharedProcessForSiteWithUserInteraction)): Added. Canonical link: https://commits.webkit.org/300964@main
https://bugs.webkit.org/show_bug.cgi?id=300124 rdar://160150633 Reviewed by Matthew Finkel. Before this PR, if the network failed or the server terminated the connection group, JS would receive no notification that that happened. Now it does. Tests: Tools/TestWebKitAPI/NetworkConnection.h Tools/TestWebKitAPI/NetworkConnection.mm Tools/TestWebKitAPI/Tests/WebKitCocoa/WebTransport.mm * Source/WebCore/Modules/webtransport/WebTransport.cpp: (WebCore::WebTransport::didFail): (WebCore::WebTransport::networkProcessCrashed): Deleted. * Source/WebCore/Modules/webtransport/WebTransport.h: * Source/WebCore/Modules/webtransport/WebTransportSessionClient.h: * Source/WebCore/Modules/webtransport/WorkerWebTransportSession.cpp: (WebCore::WorkerWebTransportSession::didFail): (WebCore::WorkerWebTransportSession::networkProcessCrashed): Deleted. * Source/WebCore/Modules/webtransport/WorkerWebTransportSession.h: * Source/WebKit/NetworkProcess/webtransport/cocoa/NetworkTransportSessionCocoa.mm: (WebKit::NetworkTransportSession::initialize): * Source/WebKit/WebProcess/Network/WebTransportSession.cpp: (WebKit::WebTransportSession::didFail): (WebKit::WebTransportSession::networkProcessCrashed): Deleted. * Source/WebKit/WebProcess/Network/WebTransportSession.h: * Source/WebKit/WebProcess/Network/WebTransportSession.messages.in: * Source/WebKit/WebProcess/WebProcess.cpp: (WebKit::WebProcess::networkProcessConnectionClosed): * Tools/TestWebKitAPI/NetworkConnection.h: * Tools/TestWebKitAPI/NetworkConnection.mm: (TestWebKitAPI::ConnectionGroup::cancel): * Tools/TestWebKitAPI/Tests/WebKitCocoa/WebTransport.mm: (TestWebKitAPI::TEST(WebTransport, ServerConnectionTermination)): Canonical link: https://commits.webkit.org/300965@main
https://bugs.webkit.org/show_bug.cgi?id=299915 <rdar://161690817> Reviewed by Abrar Rahman Protyasha. Use localized string for image controls label. Also detect system RTL language usage and adjust layout as appropriate. * Source/WebCore/html/shadow/SpatialImageControls.cpp: (WebCore::SpatialImageControls::ensureSpatialControls): * Source/WebCore/html/shadow/spatialImageControls.css: (div#label): (span#spatial-glyph): (span#pano-glyph): Mirror layout iff system is in RTL language. * Source/WebCore/en.lproj/Localizable.strings: * Source/WebCore/platform/LocalizedStrings.cpp: (WebCore::imageControlsLabelSpatial): (WebCore::imageControlsLabelPanorama): * Source/WebCore/platform/LocalizedStrings.h: Make label text localizable strings. * LayoutTests/TestExpectations: * LayoutTests/fast/images/spatial-image-controls-expected.html: * LayoutTests/fast/images/spatial-image-controls-rtl-expected.html: * LayoutTests/fast/images/spatial-image-controls-rtl.html: * LayoutTests/platform/visionos/TestExpectations: Add test coverage for RTL support. Canonical link: https://commits.webkit.org/300966@main
https://bugs.webkit.org/show_bug.cgi?id=300133 rdar://161539144 Reviewed by Yusuke Suzuki and Dan Hecht. The global.get Wasm opcode in constant expression position currently doesn't support v128. Testing covered by additions to JSTests/wasm/gc/simd.js. * JSTests/wasm/gc/simd.js: (testSIMDGlobal): * Source/JavaScriptCore/wasm/WasmConstExprGenerator.cpp: (JSC::Wasm::ConstExprGenerator::getGlobal): Canonical link: https://commits.webkit.org/300967@main
https://bugs.webkit.org/show_bug.cgi?id=300055 rdar://161854148 Reviewed by Abrar Rahman Protyasha. Avoids unsafe downcast of key{up/down/press} events to KeyboardEvent, which was crashing on iOS simulated keyboard events. Crashes were occurring in these two API tests: - KeyboardInputTests.InsertDictationAlternativesSimulatingKeyboardInput - KeyboardInputTests.InsertTextSimulatingKeyboardInput (these tests are not currently being exercised by automation because the feature is currently behind a feature flag) * Source/WebCore/page/LocalDOMWindow.cpp: (WebCore::LocalDOMWindow::computeInteractionID): (WebCore::LocalDOMWindow::finalizeEventTimingEntry): Canonical link: https://commits.webkit.org/300968@main
https://bugs.webkit.org/show_bug.cgi?id=300036 Reviewed by Abrar Rahman Protyasha. Update UITextInput protocol in SPI headers (this protocol is actually API, but WKWebView still uses a pre-API internal protocol). This needs to be visible from Swift, so add it to the bridging header. Unlike our frameworks, the test bundle is not itself a module, so we do not need to modularize UIKitSPIForTesting.h and friends to be able to use it from swift code. Along the way, fix a couple of public build issues, noted below. Test: Tools/TestWebKitAPI/Tests/WebKit Swift/TestWebKitAPIBundle-Bridging-Header.h Test: Tools/TestWebKitAPI/Tests/WebKit Swift/TestWebKitAPIBundle-Bridging-Header.h * Configurations/CommonBase.xcconfig: Change the definition of SDKDB_HAS_148943382 to be internal only. When we generate partial SDKDBs from the public SDK, they do not exhibit the same bug. This change reverts an earlier attempt to work around the bug in 300877@main. * Source/WebCore/Configurations/AllowedSPI-legacy.toml: Revert an earlier attempt to fix the above problem (300877@main). * Tools/TestRunnerShared/spi/UIKitSPIForTesting.h: * Tools/TestWebKitAPI/Tests/WebKit Swift/TestWebKitAPIBundle-Bridging-Header.h: * WebKitLibraries/SDKs/iphoneos26.0-additions.sdk/System/Library/PrivateFrameworks/UIKitCore.framework/UIKitCore.tbd: Removed. The public build gets UIKit symbols from UIKit.tbd directly, not by binding to UIKitCore. This sidesteps an issue where WebKitSwift binds to different UIKit symbols in debug and release configurations, which confuses extract-tbds-from-internal-sdk. * WebKitLibraries/SDKs/iphonesimulator26.0-additions.sdk/System/Library/PrivateFrameworks/UIKitCore.framework/UIKitCore.tbd: Removed. Canonical link: https://commits.webkit.org/300969@main
https://bugs.webkit.org/show_bug.cgi?id=300063 rdar://161639791 Reviewed by Matthew Finkel. Fix CoreIPC fuzz blocker by checking that the NSURLRequest from the ResourceRequest in NetworkSessionCocoa is not nil before use. Test: ipc/networksessioncocoa-empty-resource-request.html Test: ipc/networksessioncocoa-empty-resource-request.html * LayoutTests/ipc/networksessioncocoa-empty-resource-request-expected.txt: Added. * LayoutTests/ipc/networksessioncocoa-empty-resource-request.html: Added. * Source/WebKit/NetworkProcess/cocoa/NetworkSessionCocoa.mm: (WebKit::NetworkSessionCocoa::createWebSocketTask): Canonical link: https://commits.webkit.org/300970@main
https://bugs.webkit.org/show_bug.cgi?id=291839 rdar://149683032 Reviewed by Chris Dumez. WebContent logs during launch are not emitted when logd is blocked in the sandbox because the log streaming channel to the UI process is not set up until WebProcess::platformInitializeWebProcess is called. This patch addresses this by sending logs emitted in this time interval over the XPC connection to the UI process and emit the logs there. Test: Tools/TestWebKitAPI/Tests/WebKit/LogForwarding.mm * Source/JavaScriptCore/API/JSValue.mm: (createStructHandlerMap): * Source/JavaScriptCore/API/JSWrapperMap.mm: (parsePropertyAttributes): * Source/JavaScriptCore/API/ObjcRuntimeExtras.h: (adoptSystem): Deleted. * Source/WTF/wtf/SystemFree.h: (WTF::adoptSystem): * Source/WebKit/Shared/Cocoa/LaunchLogHook.h: Added. * Source/WebKit/Shared/Cocoa/LaunchLogHook.mm: Added. (WebKit::LaunchLogHook::singleton): (WebKit::LaunchLogHook::initialize): (WebKit::LaunchLogHook::disable): * Source/WebKit/Shared/Cocoa/LaunchLogMessages.h: Added. * Source/WebKit/Shared/EntryPointUtilities/Cocoa/XPCService/XPCServiceMain.mm: (WebKit::initializeLogd): (WebKit::XPCServiceEventHandler): * Source/WebKit/UIProcess/API/Cocoa/WKWebViewPrivateForTesting.h: * Source/WebKit/UIProcess/API/Cocoa/WKWebViewTesting.mm: (-[WKWebView _launchLogsCountForTesting]): * Source/WebKit/UIProcess/Cocoa/WebProcessProxyCocoa.mm: (WebKit::WebProcessProxy::xpcEventHandler const): (WebKit::WebProcessProxy::XPCEventHandler::handleXPCEvent const): (WebKit::WebProcessProxy::XPCEventHandler::XPCEventHandler): * Source/WebKit/UIProcess/Launcher/cocoa/ProcessLauncherCocoa.mm: * Source/WebKit/UIProcess/WebProcessProxy.h: * Source/WebKit/WebKit.xcodeproj/project.pbxproj: * Source/WebKit/WebProcess/cocoa/WebProcessCocoa.mm: (WebKit::registerLogClient): * Tools/TestWebKitAPI/Tests/WebKit/LogForwarding.mm: (TEST(WebKit, LaunchLogs)): Canonical link: https://commits.webkit.org/300971@main
https://bugs.webkit.org/show_bug.cgi?id=299961 rdar://problem/161742137 Reviewed by Brian Weinstein and Timothy Hatcher. Add support for browser.runtime.getVersion(). This was proposed in the WECG: w3c/webextensions#878 Test: Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebExtensionAPIRuntime.mm * Source/WebKit/WebProcess/Extensions/API/Cocoa/WebExtensionAPIRuntimeCocoa.mm: (WebKit::WebExtensionAPIRuntime::getVersion): * Source/WebKit/WebProcess/Extensions/API/WebExtensionAPIRuntime.h: * Source/WebKit/WebProcess/Extensions/Interfaces/WebExtensionAPIRuntime.idl: * Tools/TestWebKitAPI/Tests/WebKitCocoa/WKWebExtensionAPIRuntime.mm: (TestWebKitAPI::TEST(WKWebExtensionAPIRuntime, GetVersion)): Canonical link: https://commits.webkit.org/300972@main
…dentifiers.end() https://bugs.webkit.org/show_bug.cgi?id=299996 rdar://161779919 Reviewed by Eric Carlson. Calling AudioVideoRendererAVFObjC::addTrack could trigger a call to notifyRequiresFlushToResume, before returning: as such the track identifier wouldn't have been inserted in the hash map yet causing the assertion. We ensure that the callback `notifyRequiresFlushToResume` isn't called when we just added the video track, as no video data would have been enqueued yet: so there's nothing to flush yet. Re-enabling tests. * LayoutTests/platform/ios/TestExpectations: * LayoutTests/platform/mac/TestExpectations: * Source/WebCore/platform/graphics/avfoundation/AudioVideoRendererAVFObjC.h: * Source/WebCore/platform/graphics/avfoundation/AudioVideoRendererAVFObjC.mm: (WebCore::AudioVideoRendererAVFObjC::stageVideoRenderer): Canonical link: https://commits.webkit.org/300973@main
https://bugs.webkit.org/show_bug.cgi?id=299838 rdar://161612789 Reviewed by Jer Noble. Enable preference for testing, covered by existing webm tests. * Source/WTF/Scripts/Preferences/UnifiedWebPreferences.yaml: * Source/WebCore/platform/graphics/MediaPlayer.cpp: (WebCore::WTF_REQUIRES_LOCK): Swap the order of registration of MediaPlayerFactory. MSE and MP4 are more commonly used as webm. It makes more sense to try those first. * Source/WebCore/platform/graphics/cocoa/MediaPlayerPrivateWebM.mm: (WebCore::MediaPlayerPrivateWebM::cancelLoad): We incorrectly set the networkState to idle when the load was cancelled. When running in the GPUP, it didn't make a difference as the MediaPlayerPrivateRemote stopped accepting new events from the GPUP once unloaded. But when running in the content process, it would cause the element to incorrectly fire the "suspend" event and causing tests to fail. Canonical link: https://commits.webkit.org/300974@main
https://bugs.webkit.org/show_bug.cgi?id=299580 Reviewed by Sihui Liu. Introduce a new DOM event, webkitshadowrootattached, which gets fired on an element whenever a non-user-agent shadow root attached to the element. We dispatch this event whenever DOM wrapper world is supposed to have access to all shadow roots regardless of its mode. Tests: Tools/TestWebKitAPI/Tests/WebKitCocoa/FormValidation.mm Tools/TestWebKitAPI/Tests/WebKitCocoa/UserContentController.mm * LayoutTests/fast/dom/gc-dom-tree-lifetime-shadow-tree.html: Fixed the test to wait with a 0s timer so that we perform the liveness check after webkitshadowrootattached has been dispatched. * Source/WebCore/dom/Element.cpp: (WebCore::Element::addShadowRoot): (WebCore::Element::enqueueShadowRootAttachedEvent): (WebCore::Element::dispatchShadowRootAttachedEvent): * Source/WebCore/dom/Element.h: (WebCore::Element::didDispatchShadowRootAttachedEvent): * Source/WebCore/dom/Event.h: (WebCore::Event::isShadowRootAttachedEvent): (WebCore::Event::setIsShadowRootAttachedEvent): * Source/WebCore/dom/EventNames.json: * Source/WebCore/dom/EventTarget.cpp: (WebCore::EventTarget::innerInvokeEventListeners): * Source/WebCore/dom/MutationObserver.cpp: (WebCore::MutationObserver::enqueueShadowRootAttachedEvent): (WebCore::MutationObserver::notifyMutationObservers): * Source/WebCore/dom/MutationObserver.h: * Source/WebCore/dom/Node.h: * Source/WebCore/dom/WindowEventLoop.h: * Tools/TestWebKitAPI/Tests/WebKitCocoa/FormValidation.mm: (TestWebKitAPI::TEST(WebKit, DidAssociateFormControls)): Deleted. * Tools/TestWebKitAPI/Tests/WebKitCocoa/UserContentController.mm: (TEST(WKUserContentController, DidAssociateFormControls)): (TEST(WKUserContentController, ShadowRootAttachedEvent)): Canonical link: https://commits.webkit.org/300975@main
https://bugs.webkit.org/show_bug.cgi?id=300134 <rdar://161909905> Reviewed by Aditya Keerthi and Abrar Rahman Protyasha. Guard relevant recent changes with PANORAMA_IMAGE_CONTROLS feature flag. * Source/WebCore/html/shadow/SpatialImageControls.cpp: (WebCore::SpatialImageControls::shouldHaveSpatialControls): (WebCore::SpatialImageControls::ensureSpatialControls): * Source/WebCore/html/shadow/spatialImageControls.css: Add new compile-time feature flag. * LayoutTests/platform/visionos/TestExpectations: Skip layout test since necessary changes are now compiled out. Canonical link: https://commits.webkit.org/300976@main
https://bugs.webkit.org/show_bug.cgi?id=299813 rdar://161584294 Unreviewed build fix * Source/WebGPU/Configurations/WebGPU.xcconfig: Canonical link: https://commits.webkit.org/300977@main
|
/build-preview |
|
🔄 Preview build started Release tag: |
|
✅ Preview build completed Release: autobuild-preview-pr-113-68ce1d24 |
|
/build-preview |
|
🔄 Preview build started Release tag: |
|
✅ Preview build completed Release: autobuild-preview-pr-113-58e9ebc3 |
|
/build-preview |
|
🔄 Preview build started Release tag: |
|
✅ Preview build completed Release: autobuild-preview-pr-113-341e207c |
|
/build-preview |
|
🔄 Preview build started Release tag: |
|
✅ Preview build completed Release: autobuild-preview-pr-113-5434fc4c |
|
/build-preview |
|
🔄 Preview build started Release tag: |
|
✅ Preview build completed Release: autobuild-preview-pr-113-22206137 |
|
/build-preview |
|
🔄 Preview build started Release tag: |
|
✅ Preview build completed Release: autobuild-preview-pr-113-c3634e29 |
|
/build-preview |
|
🔄 Preview build started Release tag: |
|
✅ Preview build completed Release: autobuild-preview-pr-113-a8c96c80 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
74 participants
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.
No description provided.