Timeline

Apr 13, 2019:

11:46 PM Changeset in webkit [244244] by zandobersek@gmail.com

• 4 edits in trunk/Source/bmalloc

[bmalloc][Linux] Add support for memory status calculation
​https://bugs.webkit.org/show_bug.cgi?id=195938

Reviewed by Carlos Garcia Campos.

Memory status and under-memory-pressure capabilities in bmalloc can be
implemented on Linux by reading and parsing the statm file under the
proc filesystem.

We retrieve the resident set size from the statm file and multiply it
with the page size. This gives an upper-bound estimate of the memory
that's being consumed by the process.

The statm-based estimate seems preferable to other alternatives. One
such alternative would be reading and parsing more-detailed smaps file,
also exposed under the proc filesystem. This is at the moment being done
in WTF's MemoryFootprint implementation for Linux systems, but on Linux
ports this operation is being throttled to only execute once per second
because of the big computing expense required to read and parse out the
data. A future MemoryFootprint implementation could simply retrieve the
memory footprint value from bmalloc.

Another alternative is the Linux taskstats interface. This one would
require utilizing a netlink socket to retrieve the necessary statistics,
but it requires the process to have elevated privileges, which is a
blocker.

• bmalloc/AvailableMemory.cpp:

(bmalloc::LinuxMemory::singleton):
(bmalloc::LinuxMemory::footprint const):
(bmalloc::computeAvailableMemory):
(bmalloc::memoryStatus):

• bmalloc/AvailableMemory.h:

(bmalloc::isUnderMemoryPressure):

• bmalloc/bmalloc.h:

9:22 PM Changeset in webkit [244243] by Chris Dumez

• 3 edits in trunk/Source/WebKit

[ Mac Debug ] TestWebKitAPI.ProcessSwap.ReuseSuspendedProcessForRegularNavigationRetainBundlePage is a flaky crash
​https://bugs.webkit.org/show_bug.cgi?id=196548
<rdar://problem/49567254>

Reviewed by Darin Adler.

Update ProvisionalPageProxy methods to more consistently ignore unexpected IPC from the process. Previously,
some of the methods were doing this, but some other like didFailProvisionalLoadForFrame() weren't and this
was leading to this flaky crash. The issue is that if we do the load in an existing process that was recently
doing, there may be leftover IPC for the same pageID and this IPC gets received by the ProvisionalPageProxy
even though it is from a previous navigation. For this reason, the ProvisionalPageProxy should ignore all
incoming IPC that is not for its associated navigation.

• UIProcess/ProvisionalPageProxy.cpp:

(WebKit::ProvisionalPageProxy::didPerformClientRedirect):
(WebKit::ProvisionalPageProxy::didStartProvisionalLoadForFrame):
(WebKit::ProvisionalPageProxy::didFailProvisionalLoadForFrame):
(WebKit::ProvisionalPageProxy::didCommitLoadForFrame):
(WebKit::ProvisionalPageProxy::didNavigateWithNavigationData):
(WebKit::ProvisionalPageProxy::didChangeProvisionalURLForFrame):
(WebKit::ProvisionalPageProxy::decidePolicyForNavigationActionAsync):
(WebKit::ProvisionalPageProxy::decidePolicyForResponse):
(WebKit::ProvisionalPageProxy::didPerformServerRedirect):
(WebKit::ProvisionalPageProxy::didReceiveServerRedirectForProvisionalLoadForFrame):
(WebKit::ProvisionalPageProxy::decidePolicyForNavigationActionSync):

12:44 PM Changeset in webkit [244242] by Wenson Hsieh

• 2 edits in trunk/Source/WebKit

Unreviewed, try to fix the internal build after r244239

Force the bots to regenerate WKWebpagePreferences.h by touching the file. Adds a Foundation/Foundation.h
import that was missing anyways.

• UIProcess/API/Cocoa/WKWebpagePreferences.h:

1:54 AM Changeset in webkit [244241] by Tadeu Zagallo

• 6 edits
  1 add in trunk

CodeCache should check that the UnlinkedCodeBlock was successfully created before caching it
​https://bugs.webkit.org/show_bug.cgi?id=196880

Reviewed by Yusuke Suzuki.

JSTests:

• stress/bytecode-cache-syntax-error.js: Added.

(catch):

Source/JavaScriptCore:

CodeCache should not tell the SourceProvider to cache the bytecode if it failed
to create the UnlinkedCodeBlock.

• runtime/CodeCache.cpp:

(JSC::CodeCache::getUnlinkedGlobalCodeBlock):

Tools:

Add a new function for bytecode cache tests that does not forceDiskCache
for the second run: runBytecodeCacheNoAssetion. This is necessary for the
test added in this patch, since the code is invalid and therefore won't be
cached. It should also be useful for tests that evaluate dynamically
generated code.

• Scripts/jsc-stress-test-helpers/bytecode-cache-test-helper.sh:
• Scripts/run-jsc-stress-tests:

1:00 AM Changeset in webkit [244240] by graouts@webkit.org

• 6 edits in trunk/Source/WebCore

Provide a quirk to disable Pointer Events
​https://bugs.webkit.org/show_bug.cgi?id=196877
<rdar://problem/49863470>

Reviewed by Dean Jackson.

Add a quirk to disable Pointer Events. We also opt a website that has compatibility issues with Pointer Events into this new quirk.

• dom/PointerEvent.idl:
• page/Quirks.cpp:

(WebCore::Quirks::shouldDisablePointerEventsQuirk const):

• page/Quirks.h:
• page/scrolling/ScrollingCoordinator.cpp:

(WebCore::ScrollingCoordinator::absoluteEventTrackingRegionsForFrame const):

• style/StyleTreeResolver.cpp:

(WebCore::Style::TreeResolver::resolveElement):

Apr 12, 2019:

8:54 PM Changeset in webkit [244239] by Wenson Hsieh

• 13 edits in trunk/Source

Enable modern compatibility mode by default in WKWebView on some devices
​https://bugs.webkit.org/show_bug.cgi?id=196883
<rdar://problem/49864527>

Reviewed by Tim Horton.

Source/WebCore:

Add a new helper function to determine whether an app is pre-installed on iOS, for the purposes of ensuring
compatibility with existing Apple apps that are not affected by linked-on-or-after. This involves all apps with
a bundle ID that begins with "com.apple.".

• platform/RuntimeApplicationChecks.h:
• platform/cocoa/RuntimeApplicationChecksCocoa.mm:

(WebCore::setApplicationBundleIdentifier):
(WebCore::applicationBundleStartsWith):
(WebCore::IOSApplication::isAppleApplication):

Source/WebKit:

Make some minor adjustments to new API.

• Shared/WebCompatibilityMode.h:

Rename WebCompatibilityMode::Default to WebCompatibilityMode::Recommended.

• Shared/WebPreferences.yaml:
• Shared/WebPreferencesDefaultValues.h:

Now that the role of the UseModernCompatibilityModeByDefault debug preference is limited to bypassing linked-on-
or-after and app bundle compatibility hacks, we no longer make this default to true in iOSMac.

• UIProcess/API/APIWebsitePolicies.h:
• UIProcess/API/Cocoa/WKNavigationDelegate.h:

Rename the withPreferences: label to just preferences:.

• UIProcess/API/Cocoa/WKWebViewConfiguration.h:
• UIProcess/Cocoa/NavigationState.mm:

(WebKit::NavigationState::setNavigationDelegate):
(WebKit::NavigationState::NavigationClient::decidePolicyForNavigationAction):

• UIProcess/Cocoa/VersionChecks.h:
• UIProcess/ios/WebPageProxyIOS.mm:

8:04 PM Changeset in webkit [244238] by sbarati@apple.com

• 7 edits
  1 add in trunk

r244079 logically broke shouldSpeculateInt52
​https://bugs.webkit.org/show_bug.cgi?id=196884

Reviewed by Yusuke Suzuki.

JSTests:

• microbenchmarks/int52-rand-function.js: Added.

(Math.random):

Source/JavaScriptCore:

In r244079, I changed shouldSpeculateInt52 to only return true
when the prediction is isAnyInt52Speculation(). However, it was
wrong to not to include SpecInt32 in this for two reasons:

1. We diligently write code that first checks if we should speculate Int32.

For example:
if (shouldSpeculateInt32()) ...
else if (shouldSpeculateInt52()) ...

It would be wrong not to fall back to Int52 if we're dealing with the union of
Int32 and Int52.

It would be a performance mistake to not include Int32 here because
data flow can easily tell us that we have variables that are the union
of Int32 and Int52 values. It's better to speculate Int52 than Double
in that situation.

2. We also write code where we ask if the inputs can be Int52, e.g, if

we know via profiling that an Add overflows, we may not emit an Int32 add.
However, we only emit such an add if both inputs can be Int52, and Int32
can trivially become Int52.

This patch recovers the 0.5-1% regression r244079 caused on JetStream 2.

• bytecode/SpeculatedType.h:

(JSC::isInt32SpeculationForArithmetic):
(JSC::isInt32OrBooleanSpeculationForArithmetic):
(JSC::isInt32OrInt52Speculation):

• dfg/DFGFixupPhase.cpp:

(JSC::DFG::FixupPhase::observeUseKindOnNode):

• dfg/DFGNode.h:

(JSC::DFG::Node::shouldSpeculateInt52):

• dfg/DFGPredictionPropagationPhase.cpp:
• dfg/DFGVariableAccessData.cpp:

(JSC::DFG::VariableAccessData::couldRepresentInt52Impl):

5:05 PM Changeset in webkit [244237] by sbarati@apple.com

• 2 edits in trunk/Source/JavaScriptCore

Unreviewed. Build fix after r244233.

• assembler/CPU.cpp:

4:45 PM Changeset in webkit [244236] by BJ Burg

• 2 edits in trunk/Source/WebKit

WebDriver: fix typo in EnterFullscreen.js in error-handling code
​https://bugs.webkit.org/show_bug.cgi?id=196882
<rdar://problem/49867122>

Reviewed by Devin Rousso.

• UIProcess/Automation/atoms/EnterFullscreen.js:

(enterFullscreen):

4:40 PM Changeset in webkit [244235] by Justin Fan

• 7 edits in trunk/Source/WebCore

[Web GPU] Prevent narrowing conversions during Metal function calls on 32-bit platforms
​https://bugs.webkit.org/show_bug.cgi?id=196793

Reviewed by Darin Adler.

On 32-bit platforms, NSUInteger is 32-bit, which limits certain Web GPU parameters.
Ensure that valid parameters are properly converted to NSUInteger for Metal calls, regardless of platform.

• platform/graphics/gpu/GPUBuffer.h:

(WebCore::GPUBuffer::byteLength const):

• platform/graphics/gpu/cocoa/GPUBindGroupMetal.mm:

(WebCore::tryGetResourceAsBufferBinding):
(WebCore::setBufferOnEncoder):

• platform/graphics/gpu/cocoa/GPUBufferMetal.mm:

(WebCore::GPUBuffer::validateBufferUsage):
(WebCore::GPUBuffer::tryCreate):
(WebCore::GPUBuffer::GPUBuffer):
(WebCore::GPUBuffer::setSubData):

• platform/graphics/gpu/cocoa/GPUCommandBufferMetal.mm:

(WebCore::GPUCommandBuffer::copyBufferToBuffer):
(WebCore::GPUCommandBuffer::copyBufferToTexture):
(WebCore::GPUCommandBuffer::copyTextureToBuffer):

• platform/graphics/gpu/cocoa/GPURenderPassEncoderMetal.mm:

(WebCore::GPURenderPassEncoder::drawIndexed):

• platform/graphics/gpu/cocoa/GPURenderPipelineMetal.mm:

(WebCore::trySetInputStateForPipelineDescriptor):

4:32 PM Changeset in webkit [244234] by Ross Kirsling

• 4 edits in trunk/Source

Unreviewed fix for non-unified build.

Source/WebCore:

• dom/ScriptedAnimationController.h:

Add missing include from r244182.

Source/WebKit:

• Shared/PrintInfo.cpp:

Add missing include from r244202.

4:26 PM Changeset in webkit [244233] by sbarati@apple.com

• 10 edits
  1 add in trunk/Source

Sometimes we need to user fewer CPUs in our threading calculations
​https://bugs.webkit.org/show_bug.cgi?id=196794
<rdar://problem/49389497>

Reviewed by Yusuke Suzuki.

Source/JavaScriptCore:

• JavaScriptCore.xcodeproj/project.pbxproj:
• Sources.txt:
• assembler/CPU.cpp: Added.

(JSC::isKernTCSMAvailable):
(JSC::enableKernTCSM):
(JSC::kernTCSMAwareNumberOfProcessorCores):

• assembler/CPU.h:

(JSC::isKernTCSMAvailable):
(JSC::enableKernTCSM):
(JSC::kernTCSMAwareNumberOfProcessorCores):

• heap/MachineStackMarker.h:

(JSC::MachineThreads::addCurrentThread):

• runtime/JSLock.cpp:

(JSC::JSLock::didAcquireLock):

• runtime/Options.cpp:

(JSC::computeNumberOfWorkerThreads):
(JSC::computePriorityDeltaOfWorkerThreads):

• wasm/WasmWorklist.cpp:

(JSC::Wasm::Worklist::Worklist):

Source/WebKit:

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

3:44 PM Changeset in webkit [244232] by Alan Coon

• 1 copy in tags/Safari-607.2.4

Tag Safari-607.2.4.

3:31 PM Changeset in webkit [244231] by Ross Kirsling

• 2 edits in trunk/LayoutTests

[WinCairo][WKL] Unreviewed test gardening.

• platform/wincairo-wk1/TestExpectations:

r244182 causes all animations tests to time out.

1:46 PM Changeset in webkit [244230] by Devin Rousso

• 2 edits in trunk/Source/WebKit

WebDriver: evaluating javascript shouldn't fail if a dialog is shown
​https://bugs.webkit.org/show_bug.cgi?id=196847
<rdar://problem/49609396>

Reviewed by Brian Burg.

• UIProcess/Automation/WebAutomationSession.cpp:

(WebKit::WebAutomationSession::willShowJavaScriptDialog):

1:37 PM Changeset in webkit [244229] by rniwa@webkit.org

• 2 edits in trunk/Source/WebCore

WebThread should run at a higher priority than user initiated
​https://bugs.webkit.org/show_bug.cgi?id=196849
<rdar://problem/46851062>

Reviewed by Geoffrey Garen.

Use QOS_CLASS_USER_INTERACTIVE on WebThread with -10 relative priority so that WebThread
won't wait for other threads with priority 30-37 but does not content with the main thread.

Also removed the call to pthread_attr_setschedparam which disables QoS.

This improves the blocked time in StartWebThread from 2~3ms to 250μs while cold launching
iBooks to an opened book.

• platform/ios/wak/WebCoreThread.mm:

(StartWebThread): Replaced 200 * 4096 by 800 * KB for a better readability.

11:52 AM Changeset in webkit [244228] by Simon Fraser

• 28 edits in trunk

[iOS WK2] Make -webkit-overflow-scrolling be a no-op
​https://bugs.webkit.org/show_bug.cgi?id=196803
rdar://problem/49078202

Reviewed by Antti Koivisto.

Source/WebKit:

Flip LegacyOverflowScrollingTouchEnabled to false.

• Shared/WebPreferences.yaml:

LayoutTests:

Rebase some tests, if I thought that non-stacking context overflow was interesting.
For others, make overflow be stacking context by adding z-index.
Add <!-- webkit-test-runner [ internal:AsyncOverflowScrollingEnabled=true ] --> to those that
need it.

• compositing/overflow/scrolling-content-clip-to-viewport.html:
• compositing/rtl/rtl-scrolling-with-transformed-descendants.html:
• fast/scrolling/ios/change-scrollability-on-content-resize-nested.html:
• fast/scrolling/ios/overflow-scroll-inherited-expected.txt:
• fast/scrolling/ios/overflow-scrolling-ancestor-clip-size.html:
• fast/scrolling/ios/overflow-scrolling-ancestor-clip.html:
• fast/scrolling/ios/reconcile-layer-position-recursive-expected.txt:
• fast/scrolling/ios/reconcile-layer-position-recursive.html:
• fast/scrolling/ios/scrolling-content-clip-to-viewport.html:
• fast/scrolling/ios/subpixel-overflow-scrolling-with-ancestor.html:
• fast/scrolling/ios/touch-scroll-pointer-events-none.html:
• fast/scrolling/ios/touch-scroll-visibility-hidden.html:
• platform/ios/compositing/overflow/scrolling-content-clip-to-viewport-expected.txt:
• platform/ios/fast/scrolling/ios/scrolling-content-clip-to-viewport-expected.txt:
• scrollingcoordinator/scrolling-tree/coordinated-frame-gain-scrolling-ancestor.html:
• scrollingcoordinator/scrolling-tree/coordinated-frame-in-fixed.html:
• scrollingcoordinator/scrolling-tree/coordinated-frame-lose-scrolling-ancestor.html:
• scrollingcoordinator/scrolling-tree/coordinated-frame.html:
• scrollingcoordinator/scrolling-tree/resources/doc-with-sticky.html:

11:34 AM Changeset in webkit [244227] by Ryan Haddad

• 2 edits in trunk/LayoutTests

Unreviewed test gardening for Windows.

• platform/win/TestExpectations:

11:26 AM Changeset in webkit [244226] by rniwa@webkit.org

• 2 edits in trunk/Source/WebCore

Add CSS Shadow Parts as a feature under consideration
​https://bugs.webkit.org/show_bug.cgi?id=196835

Reviewed by Antti Koivisto.

This feature is under consideration.

• features.json:

11:18 AM Changeset in webkit [244225] by Ross Kirsling

• 43 edits in trunk/Source

WebKit should build successfully even with -DENABLE_UNIFIED_BUILDS=OFF
​https://bugs.webkit.org/show_bug.cgi?id=196845

Reviewed by Ryosuke Niwa.

Source/WebCore:

• html/canvas/CanvasRenderingContext2DBase.cpp:

(WebCore::CanvasRenderingContext2DBase::FontProxy::initialize):
(WebCore::CanvasRenderingContext2DBase::FontProxy::fontMetrics const):
(WebCore::CanvasRenderingContext2DBase::FontProxy::fontDescription const):
(WebCore::CanvasRenderingContext2DBase::FontProxy::width const):
(WebCore::CanvasRenderingContext2DBase::FontProxy::drawBidiText const):
(WebCore::CanvasRenderingContext2DBase::beginCompositeLayer):
(WebCore::CanvasRenderingContext2DBase::endCompositeLayer):
Remove inline specifier to address linking errors (regardless of CMake platform).
Doing this in a .cpp file interferes with symbol creation.

• Modules/mediastream/MediaStreamTrack.cpp:
• Modules/webvr/VREyeParameters.cpp:
• Modules/webvr/VRFrameData.cpp:
• Modules/webvr/VRPose.cpp:
• accessibility/AccessibilityList.cpp:
• accessibility/isolatedtree/AXIsolatedTree.cpp:
• accessibility/isolatedtree/AXIsolatedTreeNode.cpp:
• bindings/js/JSDOMConvertWebGL.cpp:
• bindings/js/JSHistoryCustom.cpp:
• bindings/js/JSIDBCursorWithValueCustom.cpp:
• bindings/js/JSPerformanceObserverCustom.cpp:
• bindings/js/WindowProxy.cpp:
• platform/ColorData.gperf:
• platform/mediastream/RealtimeMediaSourceSettings.cpp:
• platform/network/DNSResolveQueue.cpp:
• workers/service/ServiceWorkerClientQueryOptions.h:
• workers/service/ServiceWorkerContainer.cpp:

Add missing includes to address compiler errors on GTK.

Source/WebKit:

• NetworkProcess/Classifier/ResourceLoadStatisticsDatabaseStore.h:
• NetworkProcess/NetworkCORSPreflightChecker.cpp:
• NetworkProcess/NetworkDataTask.cpp:
• NetworkProcess/NetworkHTTPSUpgradeChecker.cpp:
• NetworkProcess/NetworkHTTPSUpgradeChecker.h: Include wtf/Forward.h for String and CompletionHandler fwd decls.
• NetworkProcess/NetworkProcess.cpp:
• NetworkProcess/NetworkResourceLoadMap.cpp:
• NetworkProcess/NetworkResourceLoadMap.h:
• NetworkProcess/NetworkResourceLoader.cpp:
• NetworkProcess/PingLoad.h:
• Shared/WebCompiledContentRuleListData.cpp:
• Shared/gtk/WebEventFactory.cpp:
• UIProcess/API/C/WKWebsiteDataStoreRef.cpp:

(WKWebsiteDataStoreStatisticsClearInMemoryAndPersistentStore):
(WKWebsiteDataStoreStatisticsClearInMemoryAndPersistentStoreModifiedSinceHours):
(WKWebsiteDataStoreStatisticsResetToConsistentState):

• UIProcess/Downloads/DownloadProxyMap.cpp:
• UIProcess/InspectorTargetProxy.cpp:
• UIProcess/PageClient.h:
• UIProcess/ProcessAssertion.cpp:
• UIProcess/ProvisionalPageProxy.h:
• UIProcess/WebPageInspectorTargetAgent.h:
• UIProcess/geoclue/GeoclueGeolocationProvider.cpp:
• WebProcess/Cache/WebCacheStorageConnection.cpp:

(WebKit::WebCacheStorageConnection::updateQuotaBasedOnSpaceUsage):

• WebProcess/InjectedBundle/API/c/WKBundle.cpp:

(WKBundleSetDatabaseQuota):

• WebProcess/Storage/WebServiceWorkerFetchTaskClient.h:

Add missing includes / forward declarations to address compiler errors on GTK / WinCairo.

11:14 AM Changeset in webkit [244224] by Alan Bujtas

• 2 edits in trunk/Source/WebCore

REGRESSION (r244098): [ Mac WK1 ] Layout Test fast/dynamic/paused-event-dispatch.html is Timing out
​https://bugs.webkit.org/show_bug.cgi?id=196789
<rdar://problem/49855255>

Reviewed by Tim Horton.

Disable auto-sizing mode at the start of each test.

• testing/Internals.cpp:

(WebCore::Internals::resetToConsistentState):

11:04 AM Changeset in webkit [244223] by eric.carlson@apple.com

• 13 edits in trunk

Update AudioSession route sharing policy
​https://bugs.webkit.org/show_bug.cgi?id=196776
<rdar://problem/46501611>

Reviewed by Jer Noble.

Source/WebCore:

No new tests, updated an API test.

• platform/audio/AudioSession.cpp:

(WebCore::convertEnumerationToString):

• platform/audio/AudioSession.h:

(WTF::LogArgument<WebCore::RouteSharingPolicy>::toString):
(WTF::LogArgument<WebCore::AudioSession::CategoryType>::toString):

• platform/audio/cocoa/MediaSessionManagerCocoa.mm:

(MediaSessionManagerCocoa::updateSessionState):

• platform/audio/ios/AudioSessionIOS.mm:

(WebCore::AudioSession::setCategory):
(WebCore::AudioSession::routeSharingPolicy const):

• platform/audio/mac/AudioSessionMac.cpp:

(WebCore::AudioSession::setCategory):

Source/WebKit:

• UIProcess/ios/forms/WKAirPlayRoutePicker.mm:

(-[WKAirPlayRoutePicker showFromView:routeSharingPolicy:routingContextUID:hasVideo:]):

Source/WTF:

• wtf/Platform.h: Define HAVE_ROUTE_SHARING_POLICY_LONG_FORM_VIDEO.

Tools:

• TestWebKitAPI/Tests/WebKitLegacy/ios/AudioSessionCategoryIOS.mm:

(TestWebKitAPI::routeSharingPolicyLongFormVideo):
(TestWebKitAPI::routeSharingPolicyLongFormAudio):
(TestWebKitAPI::TEST):

11:02 AM Changeset in webkit [244222] by rmorisset@apple.com

• 3 edits in trunk/Source/JavaScriptCore

Use padding at end of ArrayBuffer
​https://bugs.webkit.org/show_bug.cgi?id=196823

Reviewed by Filip Pizlo.

• runtime/ArrayBuffer.h:

10:59 AM Changeset in webkit [244221] by Wenson Hsieh

• 2 edits in trunk/Source/WebKit

WebsitePolicies doesn't copy its media source policy in WebsitePolicies::copy
​https://bugs.webkit.org/show_bug.cgi?id=196862

Reviewed by Darin Adler.

Add a missing bit of code to copy m_mediaSourcePolicy over when making a copy of WebsitePolicies. This doesn't
have any observable effect, since m_mediaSourcePolicy isn't currently exposed as SPI on WKWebpagePreferences.

• UIProcess/API/APIWebsitePolicies.cpp:

(API::WebsitePolicies::copy const):

10:30 AM Changeset in webkit [244220] by Wenson Hsieh

• 20 edits in trunk

[iOS] Software keyboard is shown too frequently on some websites
​https://bugs.webkit.org/show_bug.cgi?id=195856
<rdar://problem/49191395>

Reviewed by Darin Adler.

Source/WebCore/PAL:

Declare new GraphicsServices SPI.

• pal/spi/ios/GraphicsServicesSPI.h:

Source/WebKit:

On some websites, hidden editable elements are very frequently focused upon user interaction. Currently, this
causes the software keyboard to pop in and out unexpectedly; luckily, these same sites also apply
inputmode="none" to the hidden editable element, which ought to ensure that the software keyboard doesn't appear
when the element is focused.

However, since we disabled support for inputmode="none" in r240497, the software keyboard is no longer
suppressed, and becomes a big nuissance. r240497 removed support for this feature because, when using a hardware
keyboard, pressing the globe key no longer showed UI for switching languages. However, support for inputmode
none makes a much larger impact when a software keyboard is used (since the entire software keyboard animates in
and out), whereas a hardware keyboard only displays an input accessory view. For this reason, we can mitigate
this bug without reintroducing <rdar://problem/47406553> by re-enabling inputmode="none", but only when a
hardware keyboard is not attached.

• UIProcess/API/Cocoa/WKWebView.mm:

(hardwareKeyboardAvailabilityChangedCallback):

• UIProcess/ios/WKContentViewInteraction.h:
• UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView shouldShowAutomaticKeyboardUI]):

Don't show the keyboard if inputmode is none and a hardware keyboard is not attached.

(-[WKContentView _hardwareKeyboardAvailabilityChanged]):

Reload input views if the inputmode is none to ensure that if a hardware keyboard is attached while editing an
element with inputmode="none", we'll show the input accessory view once again.

Tools:

Add support for attaching or detaching the hardware keyboard on iOS in layout tests.

• DumpRenderTree/ios/UIScriptControllerIOS.mm:

(WTR::UIScriptController::setHardwareKeyboardAttached):

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

(WTR::UIScriptController::setHardwareKeyboardAttached):

• TestRunnerShared/UIScriptContext/UIScriptController.h:
• WebKitTestRunner/Configurations/WebKitTestRunnerApp.xcconfig:

Additionally link against GraphicsServices in WebKitTestRunner.

• WebKitTestRunner/ios/UIScriptControllerIOS.mm:

(WTR::TestController::platformResetStateToConsistentValues):
(WTR::UIScriptController::setHardwareKeyboardAttached):

WebKitLibraries:

Add a symbol for GSEventSetHardwareKeyboardAttached.

• WebKitPrivateFrameworkStubs/iOS/12/GraphicsServices.framework/GraphicsServices.tbd:

LayoutTests:

Fix a failing layout test, which (among other reasons) is currently failing because support for inputmode="none"
is disabled.

• fast/forms/ios/inputmode-none-expected.txt:
• fast/forms/ios/inputmode-none.html:
• resources/ui-helper.js:

Add a UIHelper method for attaching or detaching the hardware keyboard.

(window.UIHelper.setHardwareKeyboardAttached):
(window.UIHelper):

10:12 AM Changeset in webkit [244219] by Ryan Haddad

• 2 edits in trunk/LayoutTests

[macOS WK1] ASSERTION FAILED: formData in WebCore::ResourceRequest::doUpdateResourceHTTPBody()
​https://bugs.webkit.org/show_bug.cgi?id=196864

Unreviewed test gardening.

• platform/mac-wk1/TestExpectations: Skip test since it consistently crashes, update bug number.

9:52 AM Changeset in webkit [244218] by graouts@webkit.org

• 13 edits
  1 copy in trunk/Source

Opt some websites into the simulated mouse events dispatch quirk when in modern compatibility mode
​https://bugs.webkit.org/show_bug.cgi?id=196830
<rdar://problem/49124313>

Reviewed by Wenson Hsieh.

Source/WebCore:

We add a new policy to determine whether simulated mouse events dispatch are allowed and use it to determine whether the
simulated mouse events dispatch quirk can be used for a given website. We then check the domain name for the current page's
document to see if it matches some known websites that require this quirk.

We needed to add some calls into Quirks::shouldDispatchSimulateMouseEvents() where we used to only consult the RuntimeEnabledFeature
flag to ensure we correctly created touch regions for simulated mouse events.

• dom/EventNames.h:

(WebCore::EventNames::isTouchRelatedEventType const):

• dom/Node.cpp:

(WebCore::Node::moveNodeToNewDocument):
(WebCore::tryAddEventListener):
(WebCore::tryRemoveEventListener):
(WebCore::Node::defaultEventHandler):

• loader/DocumentLoader.h:

(WebCore::DocumentLoader::simulatedMouseEventsDispatchPolicy const):
(WebCore::DocumentLoader::setSimulatedMouseEventsDispatchPolicy):

• page/DOMWindow.cpp:

(WebCore::DOMWindow::addEventListener):
(WebCore::DOMWindow::removeEventListener):

• page/Quirks.cpp:

(WebCore::Quirks::shouldDispatchSimulatedMouseEvents const):
(WebCore::Quirks::shouldDispatchSimulateMouseEvents const): Deleted.

• page/Quirks.h:

Source/WebKit:

We add a new policy to determine whether simulated mouse events dispatch are allowed.

• Shared/WebsitePoliciesData.cpp:

(WebKit::WebsitePoliciesData::encode const):
(WebKit::WebsitePoliciesData::decode):
(WebKit::WebsitePoliciesData::applyToDocumentLoader):

• Shared/WebsitePoliciesData.h:
• Shared/WebsiteSimulatedMouseEventsDispatchPolicy.h: Added.
• UIProcess/API/APIWebsitePolicies.cpp:

(API::WebsitePolicies::copy const):
(API::WebsitePolicies::data):

• UIProcess/API/APIWebsitePolicies.h:
• WebKit.xcodeproj/project.pbxproj:

9:11 AM Changeset in webkit [244217] by Chris Dumez

• 4 edits in trunk/Source/WebKit

[iOS Sim Debug] ASSERTION FAILED: m_downloads.isEmpty() Layout Test http/tests/websocket/tests/hybi/network-process-crash-error.html is a flaky crash
​https://bugs.webkit.org/show_bug.cgi?id=196781
<rdar://problem/49789381>

Reviewed by Darin Adler.

When the network process gets terminated by the client, the NetworkProcessProxy object (and thus its DownloadProxyMap member) get
destroyed right away, before we get a call to didClose(IPC::Connection&). As a result, if there are ongoing downloads at the time
of the termination, we will hit the assertion above. To address the issue, update the NetworkProcessProxy destructor to invalidate
its DownloadProxyMap member, similator to what it does in didClose(IPC::Connection&).

• UIProcess/Downloads/DownloadProxyMap.cpp:

(WebKit::DownloadProxyMap::invalidate):
(WebKit::DownloadProxyMap::processDidClose): Deleted.

• UIProcess/Downloads/DownloadProxyMap.h:
• UIProcess/Network/NetworkProcessProxy.cpp:

(WebKit::NetworkProcessProxy::~NetworkProcessProxy):
(WebKit::NetworkProcessProxy::didClose):

8:54 AM Changeset in webkit [244216] by commit-queue@webkit.org

• 1 edit
  26 adds in trunk/LayoutTests/imported/w3c

Import WPT preload tests
​https://bugs.webkit.org/show_bug.cgi?id=196439

Patch by Rob Buis <​rbuis@igalia.com> on 2019-04-12
Reviewed by Youenn Fablet.

• web-platform-tests/preload/avoid-delaying-onload-link-preload-expected.txt: Added.
• web-platform-tests/preload/avoid-delaying-onload-link-preload.html: Added.
• web-platform-tests/preload/delaying-onload-link-preload-after-discovery-expected.txt: Added.
• web-platform-tests/preload/delaying-onload-link-preload-after-discovery.html: Added.
• web-platform-tests/preload/dynamic-adding-preload-expected.txt: Added.
• web-platform-tests/preload/dynamic-adding-preload.html: Added.
• web-platform-tests/preload/preload-csp.sub-expected.txt: Added.
• web-platform-tests/preload/preload-csp.sub.html: Added.
• web-platform-tests/preload/preload-default-csp.sub-expected.txt: Added.
• web-platform-tests/preload/preload-default-csp.sub.html: Added.
• web-platform-tests/preload/resources/dummy.css: Added.
• web-platform-tests/preload/resources/dummy.css.sub.headers: Added.
• web-platform-tests/preload/resources/dummy.js: Added.
• web-platform-tests/preload/resources/dummy.js.sub.headers: Added.
• web-platform-tests/preload/resources/dummy.xml: Added.
• web-platform-tests/preload/resources/dummy.xml.sub.headers: Added.
• web-platform-tests/preload/resources/foo.vtt: Added.
• web-platform-tests/preload/resources/preload_helper.js: Added.

(verifyPreloadAndRTSupport):
(getAbsoluteURL):
(verifyNumberOfResourceTimingEntries):
(verifyLoadedAndNoDoubleDownload):

• web-platform-tests/preload/resources/sound_5.oga: Added.
• web-platform-tests/preload/resources/square.png: Added.
• web-platform-tests/preload/resources/white.mp4: Added.
• web-platform-tests/preload/single-download-late-used-preload-expected.txt: Added.
• web-platform-tests/preload/single-download-late-used-preload.html: Added.
• web-platform-tests/preload/single-download-preload-expected.txt: Added.
• web-platform-tests/preload/single-download-preload.html: Added.

7:54 AM Changeset in webkit [244215] by Simon Fraser

• 4 edits
  12 adds in trunk

[iOS WK2] Wrong scrolling behavior for nested absolute position elements inside overflow scroll
​https://bugs.webkit.org/show_bug.cgi?id=196146

Reviewed by Antti Koivisto.

Source/WebCore:

computeCoordinatedPositioningForLayer() failed to handle nested positions elements
inside overflow scroll, because it only walked up to the first containing block of
a nested position:absolute. We need to walk all the way up the ancestor layer chain,
looking at containing block, scroller and composited ancestor relationships.

Make this code easier to understand by writing it in terms of "is foo scrolled by bar", rather than
trying to collapse all the logic into a single ancestor walk, which was really hard. This is a few
more ancestor traversals, but we now only run this code if there's composited scrolling
in the ancestor chain.

Tests: scrollingcoordinator/scrolling-tree/nested-absolute-in-absolute-overflow.html

scrollingcoordinator/scrolling-tree/nested-absolute-in-overflow.html
scrollingcoordinator/scrolling-tree/nested-absolute-in-relative-in-overflow.html
scrollingcoordinator/scrolling-tree/nested-absolute-in-sc-overflow.html

• rendering/RenderLayerCompositor.cpp:

(WebCore::enclosingCompositedScrollingLayer):
(WebCore::isScrolledByOverflowScrollLayer):
(WebCore::isNonScrolledLayerInsideScrolledCompositedAncestor):
(WebCore::RenderLayerCompositor::layerContainingBlockCrossesCoordinatedScrollingBoundary):
(WebCore::collectStationaryLayerRelatedOverflowNodes):
(WebCore::RenderLayerCompositor::computeCoordinatedPositioningForLayer const):
(WebCore::collectRelatedCoordinatedScrollingNodes):
(WebCore::layerParentedAcrossCoordinatedScrollingBoundary): Deleted.

LayoutTests:

Dump the scrolling tree for various configurations of positioned, overflow and stacking context
elements.

• fast/scrolling/ios/overflow-scroll-overlap-6-expected.txt: Progressed results.
• platform/ios-wk2/scrollingcoordinator/scrolling-tree/nested-absolute-in-absolute-overflow-expected.txt: Added.
• platform/ios-wk2/scrollingcoordinator/scrolling-tree/nested-absolute-in-overflow-expected.txt: Added.
• platform/ios-wk2/scrollingcoordinator/scrolling-tree/nested-absolute-in-relative-in-overflow-expected.txt: Added.
• platform/ios-wk2/scrollingcoordinator/scrolling-tree/nested-absolute-in-sc-overflow-expected.txt: Added.
• scrollingcoordinator/scrolling-tree/nested-absolute-in-absolute-overflow-expected.txt: Added.
• scrollingcoordinator/scrolling-tree/nested-absolute-in-absolute-overflow.html: Added.
• scrollingcoordinator/scrolling-tree/nested-absolute-in-overflow-expected.txt: Added.
• scrollingcoordinator/scrolling-tree/nested-absolute-in-overflow.html: Added.
• scrollingcoordinator/scrolling-tree/nested-absolute-in-relative-in-overflow-expected.txt: Added.
• scrollingcoordinator/scrolling-tree/nested-absolute-in-relative-in-overflow.html: Added.
• scrollingcoordinator/scrolling-tree/nested-absolute-in-sc-overflow-expected.txt: Added.
• scrollingcoordinator/scrolling-tree/nested-absolute-in-sc-overflow.html: Added.

6:29 AM Changeset in webkit [244214] by commit-queue@webkit.org

• 3 edits in trunk/Tools

[GTK][WPE] Fix pacman install-dependencies packages
​https://bugs.webkit.org/show_bug.cgi?id=196817

Patch by Ludovico de Nittis <​ludovico.denittis@collabora.com> on 2019-04-12
Reviewed by Žan Doberšek.

python3-setuptools is called python-setuptools
ruby-highline is only available on AUR
libv4l-devel is incorporated in v4l-utils

• gtk/install-dependencies:
• wpe/install-dependencies:

5:04 AM Changeset in webkit [244213] by Manuel Rego Casasnovas

• 17 edits
  2 moves
  13 adds in trunk

[css-flex][css-grid] Fix synthesized baseline
​https://bugs.webkit.org/show_bug.cgi?id=196312

Reviewed by Javier Fernandez.

LayoutTests/imported/w3c:

Imported some tests from WPT css-align test suite that are fixed with this patch.

• resources/import-expectations.json:
• web-platform-tests/css/css-align/baseline-rules/synthesized-baseline-flexbox-001-expected.txt: Added.
• web-platform-tests/css/css-align/baseline-rules/synthesized-baseline-flexbox-001.html: Added.
• web-platform-tests/css/css-align/baseline-rules/synthesized-baseline-grid-001-expected.txt: Added.
• web-platform-tests/css/css-align/baseline-rules/synthesized-baseline-grid-001.html: Added.
• web-platform-tests/css/css-align/baseline-rules/synthesized-baseline-inline-block-001-expected.txt: Added.
• web-platform-tests/css/css-align/baseline-rules/synthesized-baseline-inline-block-001.html: Added.
• web-platform-tests/css/css-align/baseline-rules/w3c-import.log: Added.

Source/WebCore:

When a flex or grid container has no baseline,
its baseline should be synthesized from the border edges.
The same happens for flex and grid items.

Right now we were using the content box in some cases
and even using the margin box in a particular scenario.
The patch fixes this.

At the same time this is also fixing the baseline for
inline flex/grid containers to make it interoperable with Firefox.
Inline blocks have a special behavior per legacy reasons,
which applies to inline flex/grid containers when they have no items;
otherwise the items should be used to compute its baseline.
See more at: ​https://github.com/w3c/csswg-drafts/issues/3416

Note that we need to keep current behavior for buttons,
as the flexbox spec doesn't apply to them.

Tests: css3/flexbox/flexbox-baseline-margins.html

fast/css-grid-layout/grid-baseline-margins-1.html
fast/css-grid-layout/grid-baseline-margins-2.html
imported/w3c/web-platform-tests/css/css-align/baseline-rules/synthesized-baseline-flexbox-001.html
imported/w3c/web-platform-tests/css/css-align/baseline-rules/synthesized-baseline-grid-001.html
imported/w3c/web-platform-tests/css/css-align/baseline-rules/synthesized-baseline-inline-block-001.html

• rendering/RenderButton.cpp:

(WebCore::synthesizedBaselineFromContentBox):
(WebCore::RenderButton::baselinePosition const):

• rendering/RenderButton.h:
• rendering/RenderFlexibleBox.cpp:

(WebCore::synthesizedBaselineFromBorderBox):
(WebCore::RenderFlexibleBox::baselinePosition const):
(WebCore::RenderFlexibleBox::firstLineBaseline const):
(WebCore::RenderFlexibleBox::inlineBlockBaseline const):

• rendering/RenderGrid.cpp:

(WebCore::RenderGrid::baselinePosition const):
(WebCore::RenderGrid::inlineBlockBaseline const):

LayoutTests:

Some of the tests were not checking the right behavior changed them to test the expected one.
We need new rebaselines for some tests.

• TestExpectations:
• css3/flexbox/flexbox-baseline-margins.html:
• fast/css-grid-layout/grid-baseline-expected.html:
• fast/css-grid-layout/grid-baseline-margins-1-expected.html: Renamed from LayoutTests/fast/css-grid-layout/grid-baseline-margins-expected.html.
• fast/css-grid-layout/grid-baseline-margins-1.html: Renamed from LayoutTests/fast/css-grid-layout/grid-baseline-margins.html.
• fast/css-grid-layout/grid-baseline-margins-2-expected.html: Added.
• fast/css-grid-layout/grid-baseline-margins-2.html: Added.
• fast/css-grid-layout/grid-baseline.html: This test is modified and split in two parts as it doesn't fit in the viewport.
• platform/gtk/css3/flexbox/flexbox-baseline-margins-expected.png:
• platform/gtk/css3/flexbox/flexbox-baseline-margins-expected.txt:
• platform/ios/css3/flexbox/flexbox-baseline-margins-expected.png: Added.
• platform/ios/css3/flexbox/flexbox-baseline-margins-expected.txt:
• platform/mac/css3/flexbox/flexbox-baseline-margins-expected.png: Added.
• platform/mac/css3/flexbox/flexbox-baseline-margins-expected.txt:
• platform/win/css3/flexbox/flexbox-baseline-margins-expected.png: Added.
• platform/win/css3/flexbox/flexbox-baseline-margins-expected.txt:

1:46 AM Changeset in webkit [244212] by Carlos Garcia Campos

• 11 edits in trunk

[GTK] REGRESSION(r243860): Many tests failing
​https://bugs.webkit.org/show_bug.cgi?id=196791

Reviewed by Joanmarie Diggs.

Source/WebKit:

Calling updateAccessibilityTree() on document loaded was causing a re-layout because of the backing store update
that confused all those tests. We shouldn't need to update the accessibility tree on document load, it should
happen automatically when root object is attached/detached. This patch emits children-changed::add when the root
object wrapper is attached and children-changed::remove when the root object is detached. That way ATs are
notified of the changes in the accessibility tree.

• WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:

(WebKit::WebFrameLoaderClient::dispatchDidFinishDocumentLoad): Remove call to WebPage::updateAccessibilityTree().

• WebProcess/WebPage/WebPage.h: Remove updateAccessibilityTree().
• WebProcess/WebPage/atk/WebKitWebPageAccessibilityObject.cpp:

(coreRootObjectWrapperDetachedCallback): Emit children-changed::remove.
(rootWebAreaWrapper): Helper to get the root WebArea wrapper.
(accessibilityRootObjectWrapper): Set the parent here when root object is created and emit children-changed::add.
(webkitWebPageAccessibilityObjectRefChild): Dot no set the parent here, it's now set when the root object is created.

• WebProcess/WebPage/atk/WebKitWebPageAccessibilityObject.h: Remove webkitWebPageAccessibilityObjectRefresh().
• WebProcess/WebPage/gtk/WebPageGtk.cpp:

Tools:

Rework the accessibility unit test to use DBus for the communication with the server. This way we can load
multiple documents and check that accessibility hierarchy is updated after a navigation.

• TestWebKitAPI/Tests/WebKitGtk/AccessibilityTestServer.cpp:

(loadChangedCallback):

• TestWebKitAPI/Tests/WebKitGtk/TestWebKitAccessibility.cpp:

(AccessibilityTest::AccessibilityTest):
(AccessibilityTest::~AccessibilityTest):
(AccessibilityTest::loadHTMLAndWaitUntilFinished):
(AccessibilityTest::findTestServerApplication):
(AccessibilityTest::findDocumentWeb):
(AccessibilityTest::findRootObject):
(AccessibilityTest::waitUntilChildrenRemoved):
(AccessibilityTest::ensureProxy):
(testAtspiBasicHierarchy):
(beforeAll):
(afterAll):

LayoutTests:

Remove expectations for tests that pass now.

• platform/gtk/TestExpectations: