Timeline

Nov 14, 2018:

10:01 PM Changeset in webkit [238219] by Keith Rollin

• 9 edits
  10 adds in trunk

Move scripts for Derived and Unified Sources to external files
​https://bugs.webkit.org/show_bug.cgi?id=191670
<rdar://problem/46082278>

Reviewed by Keith Miller.

Move the scripts in the Generate Derived Sources and Generate Unified
Sources Run Script phases from the Xcode projects to external shell
script files. Then invoke those scripts from the Run Script phases.
This refactoring is being performed to support later work that will
invoke these scripts in other contexts.

The scripts were maintained as-is when making the move. I did a little
reformatting and added 'set -e' to the top of each file, but that's
it.

Source/JavaScriptCore:

• JavaScriptCore.xcodeproj/project.pbxproj:
• Scripts/generate-derived-sources.sh: Added.
• Scripts/generate-unified-sources.sh: Added.

Source/WebCore:

No new tests -- no changed functionality.

• Scripts/generate-derived-sources.sh: Added.
• Scripts/generate-unified-sources.sh: Added.
• WebCore.xcodeproj/project.pbxproj:

Source/WebKit:

• Scripts/generate-derived-sources.sh: Added.
• Scripts/generate-unified-sources.sh: Added.
• WebKit.xcodeproj/project.pbxproj:

Tools:

• DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
• DumpRenderTree/Scripts/generate-derived-sources.sh: Added.
• WebKitTestRunner/Scripts/generate-derived-sources.sh: Added.
• WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:

7:35 PM Changeset in webkit [238218] by Chris Dumez

• 6 edits in trunk

WebKit.DecidePolicyForNavigationActionForHyperlinkThatRedirects API fails when PSON is enabled
​https://bugs.webkit.org/show_bug.cgi?id=191640

Reviewed by Alex Christensen.

Source/WebKit:

The issue was that when process-swapping on a redirect, the parameters provided by the new
WebContent process for navigationActionData / originatingFrameInfoData / frameSecurityOrigin
are not correct because it does not have sufficient information (it does not know a swap
happened).

To address the issue, we now store the navigationActionData / originatingFrameInfoData /
frameSecurityOrigin on the Navigation object whenever decidePolicyForNavigationAction is
called. If the decidePolicyForNavigationAction is for a redirect, use the ones we stored
on the Navigation object instead of the ones provided by the WebContent process. In case
of redirect, those do not change anyway.

This fixes all WebKit.DecidePolicyForNavigationAction* tests with PSON enabled.

• UIProcess/API/APINavigation.h:

(API::Navigation::currentRequestIsRedirect const):
(API::Navigation::wasUserInitiated const):
(API::Navigation::shouldForceDownload const):
(API::Navigation::treatAsSameOriginNavigation const):
(API::Navigation::hasOpenedFrames const):
(API::Navigation::openedViaWindowOpenWithOpener const):
(API::Navigation::opener const):
(API::Navigation::requesterOrigin const):
(API::Navigation::lockHistory const):
(API::Navigation::lockBackForwardList const):
(API::Navigation::clientRedirectSourceForHistory const):
(API::Navigation::setLastNavigationAction):
(API::Navigation::lastNavigationAction const):
(API::Navigation::setOriginatingFrameInfo):
(API::Navigation::originatingFrameInfo const):
(API::Navigation::setDestinationFrameSecurityOrigin):
(API::Navigation::destinationFrameSecurityOrigin const):

• UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::decidePolicyForNavigationActionAsync):
(WebKit::WebPageProxy::decidePolicyForNavigationAction):
(WebKit::WebPageProxy::decidePolicyForNavigationActionSync):

• UIProcess/WebPageProxy.h:

Tools:

Add API test coverage.

• TestWebKitAPI/Tests/WebKitCocoa/DecidePolicyForNavigationAction.mm:

(runDecidePolicyForNavigationActionForHyperlinkThatRedirects):
(TEST):

7:33 PM Changeset in webkit [238217] by Keith Rollin

• 4 edits in trunk/Source

Fix #end vs. #endif typo.
​https://bugs.webkit.org/show_bug.cgi?id=191668
<rdar://problem/46081704>

Reviewed by Alexey Proskuryakov.

Source/WebCore/SourcesCocoa.txt had a #end that should have been a
#endif. Fix this, an add a check to generate-unified-source-bundles.rb
to detect similar typos.

Source/WebCore:

No new tests -- no changed functionality.

• SourcesCocoa.txt:

Source/WTF:

• Scripts/generate-unified-source-bundles.rb:

7:32 PM Changeset in webkit [238216] by Chris Dumez

• 2 edits in trunk/Tools

REGRESSION (r238115): [iOS] TestWebKitAPI.ProcessSwap.NavigateToInvalidURL is failing
​https://bugs.webkit.org/show_bug.cgi?id=191660

Reviewed by Ryosuke Niwa.

The test tries to navigate to an invalid URL and makes sure that we do not process swap
for such load. The test was inherently racing because it was doing:
<body onload="setTimeout(() => alert('DONE'), 0); location.href = '​http://A=a%B=b'">

Its load event handler tries to navigate to '​http://A=a%B=b' and sets a 0 timer to
log an alert, which causes the API test to keep going and check that we did not swap
process. The issue is that a 0 timer may not be sufficient to trigger the load, as
the policy decision is asynchronous and relies on the UIProcess responding to it.

To address the issue, the test now waits for the next policy decision, spins the
run loop and then makes sure we did not swap process. It no longer relies on a message
from the page.

• TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

(-[PSONNavigationDelegate webView:decidePolicyForNavigationAction:decisionHandler:]):

7:30 PM Changeset in webkit [238215] by Chris Dumez

• 4 edits in trunk

WebKit.WebsiteDataStoreCustomPaths API test is failing when enabling process prewarming
​https://bugs.webkit.org/show_bug.cgi?id=191638

Reviewed by Alex Christensen.

Source/WebKit:

WebProcessPool::prewarmProcess() should not create the default WebSite Data Store if it
does not exist yet. This is bad for memory consumption and it is what was causing this
API test to fail.

WebProcessPool::prewarmProcess() now tries to use the following data stores in this
order of preference:

1. WebProcessPool::m_websiteDataStore (aka this process pool's primary data store)
2. The data store of the last WebProcessProxy that was created
3. The default data store if it exists

If no suitable data store is found, we cancel the process prewarming and log a console
message instead.

• UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::prewarmProcess):

Tools:

Add API test coverage.

• TestWebKitAPI/Tests/WebKitCocoa/WebsiteDataStoreCustomPaths.mm:

(runWebsiteDataStoreCustomPaths):
(TEST):

6:40 PM Changeset in webkit [238214] by Keith Rollin

• 2 edits in trunk/Source/WebCore

Remove VideoFullscreenLayerManager.mm from WebCore/SourcesCocoa.txt
​https://bugs.webkit.org/show_bug.cgi?id=191667
<rdar://problem/46081286>

Reviewed by Eric Carlson.

VideoFullscreenLayerManager.mm no longer exists.

No new tests -- No changed functionality.

• SourcesCocoa.txt:

6:27 PM Changeset in webkit [238213] by timothy@apple.com

• 4 edits in trunk/Source

Enabled dark mode CSS support by default.
​https://bugs.webkit.org/show_bug.cgi?id=191609
rdar://problem/46046861

Reviewed by Megan Gardner.

Source/WebCore:

• page/RuntimeEnabledFeatures.h: Set m_isDarkModeCSSEnabled to true.

Source/WebKit:

• Shared/WebPreferences.yaml: Set DarkModeCSSEnabled's defaultValue to true.

5:48 PM Changeset in webkit [238212] by timothy@apple.com

• 24 edits
  2 adds in trunk

Default the view background color and text color to different values when in dark mode.
​https://bugs.webkit.org/show_bug.cgi?id=191607
rdar://problem/46045854

Reviewed by Dean Jackson.

Source/WebCore:

Test: css-dark-mode/default-colors.html

• css/html.css:

(html): Set color: text on macOS.

• dom/Document.cpp:

(WebCore::Document::processSupportedColorSchemes): Call recalculateBaseBackgroundColor().

• editing/EditingStyle.cpp:

(WebCore::caretColorFromStyle): Added.
(WebCore::EditingStyle::prepareToApplyAt): Use equalIgnoringSemanticColor. Check for
caret-color directly since removeEquivalentProperties fails with semantic colors.
(WebCore::extractPropertiesNotIn): Use equalIgnoringSemanticColor. Check for caret-color
directly since removeEquivalentProperties fails with semantic colors.

• page/Frame.cpp:

(WebCore::Frame::createView): Drop backgroundColor.

• page/Frame.h:
• page/FrameView.cpp:

(WebCore::FrameView::recalculateBaseBackgroundColor): Added.
(WebCore::FrameView::updateBackgroundRecursively): Drop backgroundColor argument.
Calculate the backgroundColor based on the transparent argument only.

• page/FrameView.h:
• platform/graphics/Color.h:

(WebCore::equalIgnoringSemanticColor): Added for EditingStyle.

• rendering/RenderBox.cpp:

(WebCore::RenderBox::styleDidChange): Call recalculateBaseBackgroundColor().

• rendering/RenderLayerBacking.cpp:

(WebCore::RenderLayerBacking::shouldDumpPropertyForLayer const): Use isWhiteColor()
since it ignores the semantic color flag.

• testing/Internals.cpp:

(WebCore::Internals::setViewIsTransparent): Drop backgroundColor.
(WebCore::Internals::viewBaseBackgroundColor): Added.

• testing/Internals.h:
• testing/Internals.idl: Added viewBaseBackgroundColor.

Source/WebKit:

• UIProcess/Cocoa/WebViewImpl.mm:

(WebKit::WebViewImpl::backgroundColor const): Use controlBackgroundColor.
(WebKit::WebViewImpl::updateLayer): Simplified and use backgroundColor() getter.

• WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:

(WebKit::WebFrameLoaderClient::transitionToCommittedForNewPage): Drop backgroundColor.

• WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::setDrawsBackground): Drop backgroundColor.

Source/WebKitLegacy/win:

• WebFrame.cpp:

(WebFrame::updateBackground): Drop backgroundColor.

• WebCoreSupport/WebFrameLoaderClient.cpp:

(WebFrameLoaderClient::transitionToCommittedForNewPage): Drop backgroundColor.

LayoutTests:

• css-dark-mode/default-colors-expected.txt: Added.
• css-dark-mode/default-colors.html: Added.
• inspector/css/shadow-scoped-style-expected.txt: Rebaseline.
• platform/mac/printing/width-overflow-expected.txt: Rebaseline.

5:38 PM Changeset in webkit [238211] by Alan Coon

• 7 edits in trunk/Source

Versioning.

5:27 PM Changeset in webkit [238210] by Alan Coon

• 7 edits in trunk/Source

Revert "Versioning."

This reverts commit 657b4918e31469e5d37af91eca7b4a07f0335d9c.

5:25 PM Changeset in webkit [238209] by mitz@apple.com

• 9 copies
  1 add in releases/Apple/Safari Technology Preview 70

Added a tag for Safari Technology Preview release 70.

4:33 PM Changeset in webkit [238208] by Justin Fan

• 12 edits in trunk/Source/WebCore

[WebGPU] Code quality concerns raised for 191291: [WebGPU] Experimental prototype for WebGPURenderPipeline and WebGPUSwapChain
​https://bugs.webkit.org/show_bug.cgi?id=191383

Reviewed by Dean Jackson.

Covered by existing WebGPU tests introduced in original patch.

• Modules/webgpu/GPUDevice.h:
• Modules/webgpu/GPUPipelineStageDescriptor.h:
• Modules/webgpu/GPURenderPipelineDescriptor.h: Now a base struct with a guaranteed vertex stage member.

(): Refactored into enum class.
(WebCore::GPURenderPipelineDescriptor::GPURenderPipelineDescriptor): Removed in favor of init-list construction.
(WebCore::GPURenderPipelineDescriptor::primitiveTopology): Now a proper enum class member.

• Modules/webgpu/GPUShaderModule.h:
• Modules/webgpu/WebGPUDevice.cpp:

(WebCore::WebGPUDevice::createRenderPipeline const):

• Modules/webgpu/WebGPUShaderModule.h:

(WebCore::WebGPUShaderModule::module const):

• Modules/webgpu/WebGPUShaderStage.h: Replaced enum with constants to better reflect IDL.
• Modules/webgpu/cocoa/GPURenderPipeline.h:
• Modules/webgpu/cocoa/GPURenderPipelineMetal.mm:

(WebCore::setFunctionsForPipelineDescriptor):
(WebCore::GPURenderPipeline::create):

• Modules/webgpu/cocoa/GPUSwapChain.h:
• WebCore.xcodeproj/project.pbxproj: Removed GPUPipelineDescriptorBase.

4:32 PM Changeset in webkit [238207] by Ryan Haddad

• 2 edits in trunk/LayoutTests

Layout test fast/layers/no-clipping-overflow-hidden-added-after-transform.html is a flaky failure on EWS bots
​https://bugs.webkit.org/show_bug.cgi?id=191658

Unreviewed test gardening.

• platform/mac-wk2/TestExpectations: Mark this test as flaky on Sierra Release WK2 to reduce the number of retries on mac-wk2-ews.

3:56 PM Changeset in webkit [238206] by Joseph Pecoraro

• 34 edits in trunk/Source

Web Inspector: Pass Inspector::FrontendChannel as a reference connect/disconnect methods
​https://bugs.webkit.org/show_bug.cgi?id=191612

Reviewed by Matt Baker.

Source/JavaScriptCore:

• inspector/InspectorFrontendRouter.cpp:

(Inspector::FrontendRouter::connectFrontend):
(Inspector::FrontendRouter::disconnectFrontend):

• inspector/InspectorFrontendRouter.h:
• inspector/JSGlobalObjectInspectorController.cpp:

(Inspector::JSGlobalObjectInspectorController::connectFrontend):
(Inspector::JSGlobalObjectInspectorController::disconnectFrontend):

• inspector/JSGlobalObjectInspectorController.h:
• inspector/remote/RemoteControllableTarget.h:
• inspector/remote/cocoa/RemoteConnectionToTargetCocoa.mm:

(Inspector::RemoteConnectionToTarget::setup):
(Inspector::RemoteConnectionToTarget::close):

• inspector/remote/glib/RemoteConnectionToTargetGlib.cpp:

(Inspector::RemoteConnectionToTarget::setup):
(Inspector::RemoteConnectionToTarget::close):

• runtime/JSGlobalObjectDebuggable.cpp:

(JSC::JSGlobalObjectDebuggable::connect):
(JSC::JSGlobalObjectDebuggable::disconnect):

• runtime/JSGlobalObjectDebuggable.h:

Source/WebCore:

• inspector/InspectorController.cpp:

(WebCore::InspectorController::connectFrontend):
(WebCore::InspectorController::disconnectFrontend):
(WebCore::InspectorController::show):

• inspector/InspectorController.h:
• inspector/WorkerInspectorController.cpp:

(WebCore::WorkerInspectorController::connectFrontend):
(WebCore::WorkerInspectorController::disconnectFrontend):

• page/PageDebuggable.cpp:

(WebCore::PageDebuggable::connect):
(WebCore::PageDebuggable::disconnect):

• page/PageDebuggable.h:
• testing/Internals.cpp:

(WebCore::InspectorStubFrontend::InspectorStubFrontend):
(WebCore::InspectorStubFrontend::closeWindow):

• workers/service/context/ServiceWorkerDebuggable.cpp:

(WebCore::ServiceWorkerDebuggable::connect):
(WebCore::ServiceWorkerDebuggable::disconnect):

• workers/service/context/ServiceWorkerDebuggable.h:
• workers/service/context/ServiceWorkerInspectorProxy.cpp:

(WebCore::ServiceWorkerInspectorProxy::connectToWorker):
(WebCore::ServiceWorkerInspectorProxy::disconnectFromWorker):

• workers/service/context/ServiceWorkerInspectorProxy.h:

Source/WebKit:

• UIProcess/Automation/WebAutomationSession.cpp:

(WebKit::WebAutomationSession::connect):
(WebKit::WebAutomationSession::disconnect):
(WebKit::WebAutomationSession::terminate):

• UIProcess/Automation/WebAutomationSession.h:
• UIProcess/WebPageDebuggable.cpp:

(WebKit::WebPageDebuggable::connect):
(WebKit::WebPageDebuggable::disconnect):

• UIProcess/WebPageDebuggable.h:
• UIProcess/WebPageInspectorController.cpp:

(WebKit::WebPageInspectorController::connectFrontend):
(WebKit::WebPageInspectorController::disconnectFrontend):

• UIProcess/WebPageInspectorController.h:
• WebProcess/WebPage/WebInspector.cpp:

(WebKit::WebInspector::close):

• WebProcess/WebPage/WebPageInspectorTarget.cpp:

(WebKit::WebPageInspectorTarget::connect):
(WebKit::WebPageInspectorTarget::disconnect):

Source/WebKitLegacy/mac:

• WebCoreSupport/WebInspectorClient.mm:

(-[WebInspectorWindowController destroyInspectorView]):

Source/WebKitLegacy/win:

• WebCoreSupport/WebInspectorClient.cpp:

(WebInspectorFrontendClient::destroyInspectorView):

3:40 PM Changeset in webkit [238205] by Alan Coon

• 7 edits in trunk/Source

Versioning.

3:31 PM Changeset in webkit [238204] by Alan Coon

• 1 copy in tags/Safari-607.1.14

Tag Safari-607.1.14.

2:54 PM Changeset in webkit [238203] by Matt Baker

• 3 edits in trunk/Source/WebInspectorUI

Web Inspector: Table should recalculate scrollable height when resized
​https://bugs.webkit.org/show_bug.cgi?id=191328
<rdar://problem/45854412>

Reviewed by Devin Rousso.

• UserInterface/Views/NetworkTableContentView.js:

(WI.NetworkTableContentView.prototype._hideDetailView):
Force table columns to layout after hiding the detail view.

• UserInterface/Views/Table.js:

(WI.Table):
(WI.Table.prototype.showColumn):
Call _resizeColumnsAndFiller instead of resize. The latter also cleared
_cachedWidth, which isn't unnecessary since the width hasn't changed.

(WI.Table.prototype.layout):
previously _resizeColumnsAndFiller always occurred after _updateVisibleRows,
it was just a matter of whether the cached width and height were cleared first.
That now happens in sizeDidChange.

(WI.Table.prototype.sizeDidChange):
(WI.Table.prototype._updateVisibleRows):
(WI.Table.prototype.resize): Deleted.
Internally this cleared the cached width and height, which moved to sizeDidChange.
Externally this was used by NetworkTableContentView to force columns to
layout after hiding the details view. Calling updateLayout on the table
accomplishes the same thing using the standard View API.

(WI.Table.prototype._resizeColumnsAndFiller):
Only update the cached width, since the height will already have been
updated. The number of rows doesn't change during this method, so that
check was removed.

2:39 PM Changeset in webkit [238202] by Devin Rousso

• 2 edits in trunk/Source/WebInspectorUI

Web Inspector: REGRESSION(r236540): Debugger: breakpoints are sorted alphabetically instead of numerically
​https://bugs.webkit.org/show_bug.cgi?id=191560

Reviewed by Joseph Pecoraro.

When sorting breakpoints, if the WI.TreeElements being compared both contain instances of
WI.Breakpoint, sort by WI.SourceCodeLocation instead of title.

We only do this for WI.Breakpoint because they are the only debuggable item that has a
WI.SourceCodeLocation (e.g. WI.EventBreakpoint is just an event name string).

• UserInterface/Views/DebuggerSidebarPanel.js:

(WI.DebuggerSidebarPanel.prototype._addTreeElement):
(WI.DebuggerSidebarPanel.prototype._compareBreakpointTreeElements): Added.
(WI.DebuggerSidebarPanel.prototype._addIssue):
(WI.DebuggerSidebarPanel.prototype._addTreeElement.comparator): Deleted.
(WI.DebuggerSidebarPanel.prototype._compareTreeElements): Deleted.

2:28 PM Changeset in webkit [238201] by Nikita Vasilyev

• 4 edits in trunk/Source/WebInspectorUI

Web Inspector: Styles: shift-clicking on a property should extend selection
​https://bugs.webkit.org/show_bug.cgi?id=191575
<rdar://problem/46012855>

Reviewed by Devin Rousso.

When there's at least one property is selected, Shift-clicking another property should extend the selection.
Pressing Up or Down keys while holding Shift also extends the selection.

Remove unnecessary _startedSelection. Use hasSelectedProperties and _isMousePressed instead.

• UserInterface/Views/SpreadsheetCSSStyleDeclarationEditor.js:

_isMousePressed was set to true by mistake, causing numerous assertion fails.

(WI.SpreadsheetCSSStyleDeclarationEditor.prototype.set focusIndex):
(WI.SpreadsheetCSSStyleDeclarationEditor.prototype.hasSelectedProperties): Renamed from _hasSelectedProperties.
(WI.SpreadsheetCSSStyleDeclarationEditor.prototype.extendSelectedProperties): Added.
(WI.SpreadsheetCSSStyleDeclarationEditor.prototype.spreadsheetStylePropertyMouseLeave): Removed.
(WI.SpreadsheetCSSStyleDeclarationEditor.prototype.spreadsheetStylePropertyCopy):

(WI.SpreadsheetCSSStyleDeclarationEditor.prototype._handleKeyDown):
Implement Shift-ArrowUp/ArrowDown keys.

• UserInterface/Views/SpreadsheetCSSStyleDeclarationSection.js:

(WI.SpreadsheetCSSStyleDeclarationSection):
(WI.SpreadsheetCSSStyleDeclarationSection.prototype.spreadsheetCSSStyleDeclarationEditorPropertyBlur):
(WI.SpreadsheetCSSStyleDeclarationSection.prototype.spreadsheetCSSStyleDeclarationEditorPropertyMouseLeave): Removed.
This was only used to set _startedSelection.

(WI.SpreadsheetCSSStyleDeclarationSection.prototype._handleMouseDown):
(WI.SpreadsheetCSSStyleDeclarationSection.prototype._handleWindowClick):
(WI.SpreadsheetCSSStyleDeclarationSection.prototype._handleClick):
Remove _startedSelection and use hasSelectedProperties method instead.

• UserInterface/Views/SpreadsheetStyleProperty.js:

(WI.SpreadsheetStyleProperty):
Remove spreadsheetStylePropertyMouseLeave since it is no longer used.

2:24 PM Changeset in webkit [238200] by timothy@apple.com

• 6 edits in trunk

Update prefers-color-scheme media query matching based on GitHub issue #3278.
​https://bugs.webkit.org/show_bug.cgi?id=191654
rdar://problem/46074307

Reviewed by Simon Fraser.

Source/WebCore:

Test: css-dark-mode/prefers-color-scheme.html

• css/MediaQueryEvaluator.cpp:

(WebCore::prefersColorSchemeEvaluate): Return true when there is no value. Return false
for no-preference since there is no macOS option for no user preference.

• css/MediaQueryExpression.cpp:

(WebCore::isFeatureValidWithoutValue): Added prefers-color-scheme.

LayoutTests:

• css-dark-mode/prefers-color-scheme-expected.txt: Rebaseline.
• css-dark-mode/prefers-color-scheme.html: Updates to better match the spec.

2:03 PM Changeset in webkit [238199] by Devin Rousso

• 9 edits in trunk

Web Inspector: Canvas: send a call stack with each action instead of an array of call frames
​https://bugs.webkit.org/show_bug.cgi?id=191628

Reviewed by Dean Jackson.

Source/WebCore:

Updated existing test: inspector/model/recording.html

• inspector/InspectorCanvas.h:
• inspector/InspectorCanvas.cpp:

(WebCore::InspectorCanvas::indexForData):
(WebCore::InspectorCanvas::buildInitialState):
(WebCore::InspectorCanvas::buildAction):
Drive-by: prevent de-duplicated objects from being destroyed while recording.

Source/WebInspectorUI:

• UserInterface/Models/Recording.js:

(WI.Recording.prototype.async swizzle):

• UserInterface/Models/RecordingAction.js:

(WI.RecordingAction.fromPayload):
(WI.RecordingAction.prototype.async swizzle):

LayoutTests:

• inspector/model/recording.html:

2:01 PM Changeset in webkit [238198] by Devin Rousso

• 16 edits in trunk/Source/WebInspectorUI

Web Inspector: add drag+drop for importing Audits and Recordings
​https://bugs.webkit.org/show_bug.cgi?id=191566

Reviewed by Joseph Pecoraro.

Reorganize the file loading logic for the Audit and Canvas tabs/managers. Add "drop" event
handlers to the tabs that add and show the corresponding object for the imported data.

• UserInterface/Views/AuditTabContentView.js:

(WI.AuditTabContentView):
(WI.AuditTabContentView.prototype._handleDragOver): Added.
(WI.AuditTabContentView.prototype._handleDrop): Added.

• UserInterface/Views/CanvasTabContentView.js:

(WI.CanvasTabContentView):
(WI.CanvasTabContentView.prototype._addCanvas):
(WI.CanvasTabContentView.prototype._recordingImportedOrStopped):
(WI.CanvasTabContentView.prototype._handleDragOver): Added.
(WI.CanvasTabContentView.prototype._handleDrop): Added.

• UserInterface/Controllers/AuditManager.js:

(WI.AuditManager.prototype.async processJSON): Added.
(WI.AuditManager.prototype.export):
(WI.AuditManager.prototype.import): Deleted.

• UserInterface/Controllers/CanvasManager.js:

(WI.CanvasManager.prototype.processJSON): Added.
(WI.CanvasManager.prototype.importRecording): Deleted.

• UserInterface/Views/AuditNavigationSidebarPanel.js:

(WI.AuditNavigationSidebarPanel):
(WI.AuditNavigationSidebarPanel.prototype.showDefaultContentView):
(WI.AuditNavigationSidebarPanel.prototype._handleImportButtonNavigationItemClicked):

• UserInterface/Views/CanvasSidebarPanel.js:

(WI.CanvasSidebarPanel):
(WI.CanvasSidebarPanel.prototype._handleImportButtonNavigationItemClicked): Added.
(WI.CanvasSidebarPanel.prototype._recordingChanged):
(WI.CanvasSidebarPanel.prototype._updateRecordNavigationItem):

• UserInterface/Views/CanvasOverviewContentView.js:

(WI.CanvasOverviewContentView):
(WI.CanvasOverviewContentView.prototype._handleImportButtonNavigationItemClicked): Added.

• UserInterface/Base/FileUtilities.js:

(WI.FileUtilities.save): Added.
(WI.FileUtilities.importText): Added.
(WI.FileUtilities.importJSON): Added.
(WI.FileUtilities.async readText): Added.
(WI.FileUtilities.async readJSON): Added.
(WI.saveDataToFile): Deleted.
(WI.loadDataFromFile): Deleted.

• UserInterface/Base/Main.js:

(WI._contextMenuRequested):
(WI._save):
(WI._saveAs):

• UserInterface/Views/ContextMenuUtilities.js:

(WI.appendContextMenuItemsForSourceCode):

• UserInterface/Views/LogContentView.js:

(WI.LogContentView.prototype._handleContextMenuEvent):

• UserInterface/Views/NetworkTableContentView.js:

(WI.NetworkTableContentView.prototype._exportHAR):

• UserInterface/Views/RecordingContentView.js:

(WI.RecordingContentView.prototype._exportRecording):

• UserInterface/Views/SpreadsheetCSSStyleDeclarationSection.js:

(WI.SpreadsheetCSSStyleDeclarationSection.prototype._save):
Move file related functions into a static class container.

• Localizations/en.lproj/localizedStrings.js:

1:59 PM Changeset in webkit [238197] by Joseph Pecoraro

• 5 edits in trunk/Source/WebInspectorUI

Web Inspector: Include target identifier in protocol logging
​https://bugs.webkit.org/show_bug.cgi?id=191478

Reviewed by Devin Rousso.

• UserInterface/Protocol/Connection.js:

(InspectorBackend.Connection.prototype._dispatchResponse):
(InspectorBackend.Connection.prototype._dispatchEvent):
(InspectorBackend.Connection.prototype._sendMessageToBackend):
Call tracer log methods with the connection.

• UserInterface/Debug/CapturingProtocolTracer.js:
• UserInterface/Protocol/ProtocolTracer.js:
• UserInterface/Protocol/LoggingProtocolTracer.js:

(WI.LoggingProtocolTracer):
(WI.LoggingProtocolTracer.prototype.logFrontendException):
(WI.LoggingProtocolTracer.prototype.logFrontendRequest):
(WI.LoggingProtocolTracer.prototype.logWillHandleResponse):
(WI.LoggingProtocolTracer.prototype.logDidHandleResponse):
(WI.LoggingProtocolTracer.prototype.logDidHandleEvent):
Include the connection in log methods.

(WI.LoggingProtocolTracer.prototype.logProtocolError):
Removed, this is never called.

(WI.LoggingProtocolTracer.prototype._processEntry):
Include the target identifier in entry messages.

1:54 PM Changeset in webkit [238196] by stephan.szabo@sony.com

• 3 edits in trunk/Source/WebCore

[Win] Compile Service Worker support
​https://bugs.webkit.org/show_bug.cgi?id=191409

Reviewed by Youenn Fablet.

Fix compilation errors when ENABLE(SERVICE_WORKER)
on Windows with clang-cl. Clang on dllexport
platforms does not support specifying the
dllexport on both a class and members of the class
and unistd.h isn't provided but also appeared to
not be used.

No new tests, should be covered by existing tests.

• workers/service/ServiceWorkerProvider.h:
• workers/service/context/SWContextManager.cpp:

1:41 PM Changeset in webkit [238195] by Joseph Pecoraro

• 3 edits in trunk/Source/WebKit

Unreviewed attempted wincairo build fix after 238192.

• WebProcess/WebPage/WebPageInspectorTargetController.h:
• WebProcess/WebPage/WebPageInspectorTargetFrontendChannel.cpp:

1:37 PM Changeset in webkit [238194] by Ryan Haddad

• 3 edits in trunk/LayoutTests

Unreviewed test gardening, move a skip expectation to a more specific file.

• TestExpectations:
• platform/mac-wk2/TestExpectations:

1:03 PM Changeset in webkit [238193] by Joseph Pecoraro

• 8 edits in trunk/Source/WebKit

Web Inspector: Follow-up: Keep Disabling Process Swap for Local Web Inspector
​https://bugs.webkit.org/show_bug.cgi?id=191494
<rdar://problem/45469854>

Reviewed by Devin Rousso.

Continue to have the Local Inspector disable PSON while open
until support is added.

• UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::resetState):

• UIProcess/WebPageProxy.h:

(WebKit::WebPageProxy::hasLocalInspectorFrontend const):
(WebKit::WebPageProxy::setHasLocalInspectorFrontend):

• UIProcess/WebPageProxy.messages.in:
• UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::processForNavigationInternal):

• WebProcess/WebPage/WebInspector.cpp:

(WebKit::WebInspector::openFrontendConnection):
(WebKit::WebInspector::closeFrontendConnection):

• WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::setHasLocalInspectorFrontend):

• WebProcess/WebPage/WebPage.h:

1:03 PM Changeset in webkit [238192] by Joseph Pecoraro

• 54 edits
  16 copies
  1 move
  6 adds in trunk

Web Inspector: Keep Web Inspector window alive across process swaps (PSON) (Remote Inspector)
​https://bugs.webkit.org/show_bug.cgi?id=191494
<rdar://problem/45469854>

Reviewed by Devin Rousso.

Source/JavaScriptCore:

• CMakeLists.txt:
• DerivedSources.make:
• JavaScriptCore.xcodeproj/project.pbxproj:
• Sources.txt:

New domain and resources.

• inspector/protocol/Target.json: Added.

New protocol domain, modeled after Worker.json, to allow for
multiplexing between different targets.

• inspector/InspectorTarget.h:

Each target will instantiate an InspectorTarget and must
provide an identifier, type, and means of connecting/disconnecting
to a frontend channel.

• inspector/agents/InspectorTargetAgent.cpp: Added.

(Inspector::InspectorTargetAgent::InspectorTargetAgent):
(Inspector::InspectorTargetAgent::didCreateFrontendAndBackend):
(Inspector::InspectorTargetAgent::willDestroyFrontendAndBackend):
(Inspector::InspectorTargetAgent::exists):
(Inspector::InspectorTargetAgent::initialized):
(Inspector::InspectorTargetAgent::sendMessageToTarget):
(Inspector::InspectorTargetAgent::sendMessageFromTargetToFrontend):
(Inspector::targetTypeToProtocolType):
(Inspector::buildTargetInfoObject):
(Inspector::InspectorTargetAgent::targetCreated):
(Inspector::InspectorTargetAgent::targetTerminated):
(Inspector::InspectorTargetAgent::connectToTargets):
(Inspector::InspectorTargetAgent::disconnectFromTargets):

• inspector/agents/InspectorTargetAgent.h: Added.

TargetAgent holds a list of targets, and connects/disconnects to each
of the targets when a frontend connects/disconnects.

• inspector/scripts/codegen/generator.py:

Better enum casing of ServiceWorker.

Source/WebCore:

• inspector/InspectorClient.h:

(WebCore::InspectorClient::allowRemoteInspectionToPageDirectly const):
Provide a hook so that a client may wish to allow direct remote inspection of the Page.
This is used by WebKitLegacy only.

• page/Page.cpp:

(Page::Page):
Only enable the PageDebuggable if the client wishes remote inspection of the Page directly.
This is used by WebKitLegacy only.

• inspector/InspectorController.cpp:

(WebCore::InspectorController::connectFrontend):

• inspector/InspectorController.h:
• page/PageDebuggable.cpp:

(WebCore::PageDebuggable::connect):
(WebCore::PageDebuggable::disconnect):

• page/PageDebuggable.h:

When a frontend connects, always enable the developer extras for the Page.
This is pretty much only for the remote path, which allows inspection if developer
extras were not already enabled (iOS). This simplifies the logic, and toggling
developer extras after it was already enabled is not really important.

Source/WebInspectorUI:

This starts introducing multi-target support into the Web Inspector frontend.
Specifically a backend connection that is persistent, but has the ability to
connect to and transition between Page targets received through that backend
connection.

This patch introduces the concept of a "Backend Target" which is the single
connection that the frontend contains to a backend. The old way of connecting
directly to a target is still supported. In that case the frontend constructs
a DirectBackendTarget for the debuggable type the frontend was spawns for.
However, if the frontend opens and has a TargetAgent, then it is likely
connected to a multi-target supporting backend and instead constructs a
MultiplexingBackendTarget, and will receive further information about
sub-targets to connect to. The only sub-target at the moment is a Page
sub-target.

As part of bringing up multi-target support this adds a few measures to
handle situations where the frontend is playing fast and loose with agents.
When the frontend does FooAgent.method it intends that to be performed
on the "main" target being debugged. Likewise when the frontend loops
over targets it expects them to be the debuggable targets. This patch
profiles a new implementation of WI.mainTarget and WI.targets to
match the assumptions being made by the frontend.

• In a direct-target world, there is a single target which should be used for global agents and in WI.targets.

• In a multi-target world, the page target is the one that should be used for global agents and WI.targets is the list of sub-targets (excluding the MultiplexingBackendTarget).

In a multi-target world, there are now commonly two Targets. The
MultiplexingBackendTarget and a PageTarget sub-target. In the future
this may include more targets, such as ServiceWorkers, DedicatedWorkers,
and perhaps even frames.

In a multi-target world, the frontend is immediately told about targets
as soon as it opens (via Target.targetCreated). In order to support
this, frontend initialization happens without a main target being available.
So this makes a few small changes to frontend initialization to perform
a bit of work once we know what the main target is.

During a page transition the frontend is told to destroy existing targets
and is soon after told about any new page targets (via Target.targetDestroyed
and Target.targetCreated). The frontend special cases this page transition.
It expects only one Page target to be alive at any time, accessible through
WI.pageTarget. When a page transition happens the WI.pageTarget changes, and
the frontend performs a bit of work to prepare the UI to handle the transition:
<Manager>.transitionPageTarget / WI.Notification.TransitionPageTarget.
For the most part the UI behaves fine as long once there are main frame
change and main resource change events, but those other events allow
the frontend to respond to the specific page transition cases.

• UserInterface/Base/Main.js:

(WI.loaded):
(WI.initializeBackendTarget):
(WI.initializePageTarget):
(WI.transitionPageTarget):
(WI.terminatePageTarget):
(WI.resetMainExecutionContext):
(WI.redirectGlobalAgentsToConnection):
(WI.contentLoaded):
New global functions for target initialization
and page transitioning.

• UserInterface/Test/Test.js:

(WI.loaded):
(WI.initializeBackendTarget):
(WI.initializePageTarget):
(WI.transitionPageTarget):
(WI.terminatePageTarget):
(WI.resetMainExecutionContext):
(WI.redirectGlobalAgentsToConnection):
New global functions for target initialization.
Tests continue to be a direct connection to the Page.

• UserInterface/Protocol/TargetObserver.js:

(WI.TargetObserver.prototype.targetCreated):
(WI.TargetObserver.prototype.targetDestroyed):
(WI.TargetObserver.prototype.dispatchMessageFromTarget):
New observer goes to the manager.

• UserInterface/Controllers/TargetManager.js:

(WI.TargetManager):
(WI.TargetManager.prototype.get targets):
(WI.TargetManager.prototype.get allTargets):
(WI.TargetManager.prototype.targetForIdentifier):
(WI.TargetManager.prototype.targetCreated):
(WI.TargetManager.prototype.targetDestroyed):
(WI.TargetManager.prototype.dispatchMessageFromTarget):
(WI.TargetManager.prototype.createMultiplexingBackendTarget):
(WI.TargetManager.prototype.createDirectBackendTarget):
(WI.TargetManager.prototype._addTarget):
(WI.TargetManager.prototype._removeTarget):
(WI.TargetManager.prototype._createTarget):
(WI.TargetManager.prototype._checkAndHandlePageTargetTransition):
(WI.TargetManager.prototype._checkAndHandlePageTargetTermination):
(WI.TargetManager.prototype.addTarget): Deleted.
(WI.TargetManager.prototype.removeTarget): Deleted.
(WI.TargetManager.prototype.initializeMainTarget): Deleted.
TargetManager is where we handle creating and destroying targets
and their connections. In order to simplify things a bit we make
WI.targets, which goes through get targets() an array instead
of a Set. And this includes only the sub-targets.

• UserInterface/Controllers/WorkerManager.js:

(WI.WorkerManager.prototype.workerCreated):
(WI.WorkerManager.prototype.workerTerminated):
Workers are still special-target-like things that multiplex through
WorkerAgent instead of TargetAgent. We'd like to promote these to
be full targets in the future.

• UserInterface/Protocol/DirectBackendTarget.js: Renamed from Source/WebInspectorUI/UserInterface/Protocol/MainTarget.js.

(WI.DirectBackendTarget):
(WI.DirectBackendTarget.connectionInfoForDebuggable):
(WI.DirectBackendTarget.prototype.get mainResource):
(WI.DirectBackendTarget.prototype.set mainResource):
This is the only "MainTarget" class. It is the backend target for a direct connection.

• UserInterface/Protocol/MultiplexingBackendTarget.js:

(WI.MultiplexingBackendTarget):
(WI.MultiplexingBackendTarget.prototype.initialize):
(WI.MultiplexingBackendTarget.prototype.get name):
(WI.MultiplexingBackendTarget.prototype.get executionContext):
(WI.MultiplexingBackendTarget.prototype.get mainResource):
This is the new backend target for a multi-target connection.
We don't expect it to be treated like other targets, so we don't
expect anyone to ask it for resources/executionContext/name info.

• UserInterface/Controllers/RuntimeManager.js:

(WI.TargetManager.prototype.evaluateInInspectedWindow):
This can be triggered by watch expressions before any target, and
therefore execution context, is available. Just return null, when
an execution context is available those clients will try again.

• UserInterface/Debug/Bootstrap.js:

Provide an WI.isEngineeringBuild boolean that can be used for various
debugging features.

• UserInterface/Main.html:
• UserInterface/Test.html:

New resources.

• UserInterface/Protocol/InspectorBackend.js:

(InspectorBackendClass.prototype.dispatch):
(InspectorBackendClass.prototype.runAfterPendingDispatches):
(InspectorBackend.Agent):
InspectorBackend.mainConnection was renamed InspectorBackend.backendConnection.

• UserInterface/Protocol/Connection.js:

(InspectorBackend.Connection):
(InspectorBackend.Connection.prototype._dispatchResponse):
(InspectorBackend.Connection.prototype._sendCommandToBackendWithCallback):
(InspectorBackend.Connection.prototype._sendCommandToBackendExpectingPromise):
(InspectorBackend.BackendConnection):
(InspectorBackend.BackendConnection.prototype.sendMessageToBackend):
(InspectorBackend.WorkerConnection):
(InspectorBackend.TargetConnection):
(InspectorBackend.TargetConnection.sendMessageToBackend):
Use a global sequence id to make filtering a bit easier in protocol tracing.
TargetConnection is identical to WorkerConnection except it uses TargetAgent
instead of WorkerAgent to perform multiplexing.

• UserInterface/Protocol/JavaScriptContextTarget.js:

(WI.JavaScriptContextTarget):

• UserInterface/Protocol/PageTarget.js:

(WI.PageTarget):
(WI.PageTarget.prototype.get displayName):
Specialized target types.

• UserInterface/Views/DebuggerSidebarPanel.js:

(WI.DebuggerSidebarPanel):
(WI.DebuggerSidebarPanel.prototype._targetAdded):
(WI.DebuggerSidebarPanel.prototype._targetRemoved):
(WI.DebuggerSidebarPanel.prototype._updateCallStackTreeOutline):

• UserInterface/Views/SourceCodeTextEditor.js:

(WI.SourceCodeTextEditor.prototype._targetAdded):
(WI.SourceCodeTextEditor.prototype._targetRemoved):
(WI.SourceCodeTextEditor.prototype._callFramesDidChange):
(WI.SourceCodeTextEditor.prototype._updateThreadIndicatorWidget):
(WI.SourceCodeTextEditor.prototype._reinsertAllThreadIndicators):

• UserInterface/Views/QuickConsole.js:

(WI.QuickConsole.prototype.initializeMainExecutionContextPathComponent):
(WI.QuickConsole.prototype._targetAdded):
(WI.QuickConsole.prototype._targetRemoved):
We make target added get called with all targets, including the
MultiplexingBackendTarget and PageTargets, both of which would
not have happened before. Before it was only WorkerTargets. Make
these sites a little more robust for the type of target they expect
to be able to handle.

• UserInterface/Base/Object.js:
• UserInterface/Controllers/DOMManager.js:

(WI.DOMManager):
(WI.DOMManager.prototype.transitionPageTarget):
(WI.DOMManager.prototype.requestDocument):
(WI.DOMManager.prototype._setDocument):

• UserInterface/Controllers/NetworkManager.js:

(WI.NetworkManager):
(WI.NetworkManager.prototype.transitionPageTarget):
(WI.NetworkManager.prototype.executionContextCreated):
(WI.NetworkManager.prototype._processMainFrameResourceTreePayload):

• UserInterface/Models/DefaultDashboard.js:

(WI.DefaultDashboard):
(WI.DefaultDashboard.prototype._mainResourceDidChange):
(WI.DefaultDashboard.prototype._transitionPageTarget):

• UserInterface/Views/NetworkTableContentView.js:

(WI.NetworkTableContentView):
(WI.NetworkTableContentView.prototype._mainResourceDidChange):
(WI.NetworkTableContentView.prototype._transitionPageTarget):
Special case handling when performing a page transition.

• UserInterface/Views/SettingsTabContentView.js:
• UserInterface/Debug/UncaughtExceptionReporter.js:

Document reloads are not supported right now.

Source/WebKit:

To support process swapping a slim Web Inspector backend lives in the UIProcess.
The Web Inspector frontend connects to it and is told about sub-targets, namely
pages, that it can further connect to. When performing a process swap the backend
tells the frontend to destroy existing targets and create new targets.

In the UIProcess the WebPageProxy has a WebPageInspectorController, with a single
TargetAgent holding InspectorTargetProxies to targets it knows about. Inspector
protocol messages go through this inspector controller and are routed to the
WebPage and its WebCore::Page's InspectorController. The WebPageProxy decides
when to close and expose new page targets during process swap, or basically
any time it reconnects to a WebProcess. So this patch also makes Web Inspector
stay alive and reconnect to a page when the inspected page crashes!

In the WebContentProcess the WebPage has a WebPageInspectorTarget. It also
has a WebPageInspectorTargetController in anticipation of further sub-targets
within the page (workers, frames) but none exist at the moment. The WebPage
relies on the WebPageProxy to know when to expose this target as a debuggable.

• Sources.txt:
• WebKit.xcodeproj/project.pbxproj:

New files.

• Shared/WebPageCreationParameters.cpp:

(WebKit::WebPageCreationParameters::encode const):
(WebKit::WebPageCreationParameters::decode):

• Shared/WebPageCreationParameters.h:

Remote inspector state can now stay in the UIProcess and does not need to
be passed down to the WebContentProcess.

• UIProcess/WebPageDebuggable.cpp: Copied from Source/WebCore/page/PageDebuggable.cpp.

(WebKit::WebPageDebuggable::WebPageDebuggable):
(WebKit::WebPageDebuggable::name const):
(WebKit::WebPageDebuggable::url const):
(WebKit::WebPageDebuggable::hasLocalDebugger const):
(WebKit::WebPageDebuggable::connect):
(WebKit::WebPageDebuggable::disconnect):
(WebKit::WebPageDebuggable::dispatchMessageFromRemote):
(WebKit::WebPageDebuggable::setIndicating):
(WebKit::WebPageDebuggable::setNameOverride):

• UIProcess/WebPageDebuggable.h: Copied from Source/WebCore/page/PageDebuggable.h.

Remote debuggable entry point into the UIProcess for a page.
This is pretty much identical to the PageDebuggable in WebCore.

• Scripts/webkit/messages.py:
• UIProcess/WebPageProxy.messages.in:
• UIProcess/WebPageProxy.cpp:

(WebKit::m_resetRecentCrashCountTimer):
(WebKit::WebPageProxy::finishAttachingToWebProcess):
(WebKit::WebPageProxy::close):
(WebKit::WebPageProxy::createInspectorTarget):
(WebKit::WebPageProxy::destroyInspectorTarget):
(WebKit::WebPageProxy::sendMessageToInspectorFrontend):
(WebKit::WebPageProxy::setIndicating):
(WebKit::WebPageProxy::allowsRemoteInspection const):
(WebKit::WebPageProxy::setAllowsRemoteInspection):
(WebKit::WebPageProxy::remoteInspectionNameOverride const):
(WebKit::WebPageProxy::setRemoteInspectionNameOverride):
(WebKit::WebPageProxy::remoteInspectorInformationDidChange):
(WebKit::WebPageProxy::clearInspectorTargets):
(WebKit::WebPageProxy::createInspectorTargets):
(WebKit::WebPageProxy::didCommitLoadForFrame):
(WebKit::WebPageProxy::didReceiveTitleForFrame):
(WebKit::WebPageProxy::creationParameters):

• UIProcess/WebPageProxy.h:

(WebKit::WebPageProxy::inspectorController):
(WebKit::WebPageProxy::allowsRemoteInspection const): Deleted.
(WebKit::WebPageProxy::remoteInspectionNameOverride const): Deleted.
Own more inspector state in the UIProcess including a debuggable and inspector controller.

• UIProcess/WebPageInspectorController.h: Added.
• UIProcess/WebPageInspectorController.cpp: Added.

(WebKit::WebPageInspectorController::WebPageInspectorController):
(WebKit::WebPageInspectorController::pageClosed):
(WebKit::WebPageInspectorController::hasLocalFrontend const):
(WebKit::WebPageInspectorController::hasRemoteFrontend const):
(WebKit::WebPageInspectorController::connectFrontend):
(WebKit::WebPageInspectorController::disconnectFrontend):
(WebKit::WebPageInspectorController::disconnectAllFrontends):
(WebKit::WebPageInspectorController::dispatchMessageFromFrontend):
(WebKit::WebPageInspectorController::setIndicating):
(WebKit::WebPageInspectorController::clearTargets):
(WebKit::WebPageInspectorController::createInspectorTarget):
(WebKit::WebPageInspectorController::destroyInspectorTarget):
(WebKit::WebPageInspectorController::sendMessageToInspectorFrontend):
InspectorController with a single TargetAgent in the UIProcess.

• UIProcess/WebPageInspectorTargetAgent.h:
• UIProcess/WebPageInspectorTargetAgent.cpp:

(WebKit::WebPageInspectorTargetAgent::WebPageInspectorTargetAgent):
(WebKit::WebPageInspectorTargetAgent::frontendChannel):
Target agent implementation.

• UIProcess/InspectorTargetProxy.cpp:

(WebKit::InspectorTargetProxy::create):
(WebKit::InspectorTargetProxy::InspectorTargetProxy):
(WebKit::InspectorTargetProxy::connect):
(WebKit::InspectorTargetProxy::disconnect):
(WebKit::InspectorTargetProxy::sendMessageToTargetBackend):

• UIProcess/InspectorTargetProxy.h:

UIProcess proxy for an InspectorTarget in the WebContentProcess.

• UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::processForNavigationInternal):

• WebProcess/WebPage/WebPage.cpp:

(WebKit::m_shouldAttachDrawingAreaOnPageTransition):
(WebKit::WebPage::connectInspector):
(WebKit::WebPage::disconnectInspector):
(WebKit::WebPage::sendMessageToTargetBackend):
(WebKit::WebPage::setIndicating):
(WebKit::WebPage::setAllowsRemoteInspection): Deleted.
(WebKit::WebPage::setRemoteInspectionNameOverride): Deleted.

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

• WebProcess/WebPage/WebPageInspectorTarget.h:
• WebProcess/WebPage/WebPageInspectorTarget.cpp:

(WebKit::WebPageInspectorTarget::WebPageInspectorTarget):
(WebKit::WebPageInspectorTarget::identifier const):
(WebKit::WebPageInspectorTarget::connect):
(WebKit::WebPageInspectorTarget::disconnect):
(WebKit::WebPageInspectorTarget::sendMessageToTargetBackend):
InspectorTarget for this WebPage.

• WebProcess/WebPage/WebPageInspectorTargetController.cpp: Added.

(WebKit::WebPageInspectorTargetController::WebPageInspectorTargetController):
(WebKit::WebPageInspectorTargetController::~WebPageInspectorTargetController):
(WebKit::WebPageInspectorTargetController::addTarget):
(WebKit::WebPageInspectorTargetController::removeTarget):
(WebKit::WebPageInspectorTargetController::connectInspector):
(WebKit::WebPageInspectorTargetController::disconnectInspector):
(WebKit::WebPageInspectorTargetController::sendMessageToTargetBackend):
(WebKit::WebPageInspectorTargetController::sendMessageToTargetFrontend):

• WebProcess/WebPage/WebPageInspectorTargetController.h:
• WebProcess/WebPage/WebPageInspectorTargetFrontendChannel.h:
• WebProcess/WebPage/WebPageInspectorTargetFrontendChannel.cpp:

(WebKit::WebPageInspectorTargetFrontendChannel::create):
(WebKit::WebPageInspectorTargetFrontendChannel::WebPageInspectorTargetFrontendChannel):
(WebKit::WebPageInspectorTargetFrontendChannel::sendMessageToFrontend):
Preparation for more target managment in the WebContentProcess.

Source/WebKitLegacy/mac:

• WebCoreSupport/WebInspectorClient.h:

WebKitLegacy will still have remote inspection of the Page directly.

LayoutTests:

• inspector/unit-tests/target-manager.html:

WI.targets has switched to being an array instead of a set.

12:00 PM Changeset in webkit [238191] by Truitt Savell

• 2 edits in trunk/LayoutTests

[ Sierra Release WK2 ] Layout Test fast/workers/worker-cloneport.html is flaky.
​https://bugs.webkit.org/show_bug.cgi?id=191644.

Unreviewed test Gardening.

• platform/mac-wk2/TestExpectations:

11:47 AM Changeset in webkit [238190] by achristensen@apple.com

• 23 edits
  2 moves in trunk

Add SPI to show a safe browsing warning
​https://bugs.webkit.org/show_bug.cgi?id=191604

Reviewed by Andy Estes.

Source/WebKit:

Safari has a case where it shows safe browsing warnings for a reason other than being told to by the SafeBrowsing framework.
Because safe browsing is moving to WebKit, we will have Safari call this WebKit SPI in that case.
The warning looks a little different, so I had to change the structure of a SafeBrowsingResult and the location of the string processing,
but the existing behavior is the same as verified by existing unit tests. I added a unit test to verify behavior of the new SPI.

• SourcesCocoa.txt:
• UIProcess/API/Cocoa/WKWebView.mm:

(-[WKWebView _showSafeBrowsingWarningWithTitle:warning:details:completionHandler:]):
(-[WKWebView _showSafeBrowsingWarning:completionHandler:]):
(+[WKWebView _confirmMalwareSentinel]):
(+[WKWebView _visitUnsafeWebsiteSentinel]):

• UIProcess/API/Cocoa/WKWebViewInternal.h:
• UIProcess/API/Cocoa/WKWebViewPrivate.h:
• UIProcess/Cocoa/SafeBrowsingResultCocoa.mm: Removed.
• UIProcess/Cocoa/SafeBrowsingWarningCocoa.mm: Copied from Source/WebKit/UIProcess/Cocoa/SafeBrowsingResultCocoa.mm.

(WebKit::malwareDetailsBase):
(WebKit::learnMoreURL):
(WebKit::reportAnErrorBase):
(WebKit::localizedProvider):
(WebKit::SafeBrowsingWarning::visitUnsafeWebsiteSentinel):
(WebKit::SafeBrowsingWarning::confirmMalwareSentinel):
(WebKit::replace):
(WebKit::addLinkAndReplace):
(WebKit::reportAnErrorURL):
(WebKit::malwareDetailsURL):
(WebKit::safeBrowsingTitleText):
(WebKit::safeBrowsingWarningText):
(WebKit::safeBrowsingDetailsText):
(WebKit::SafeBrowsingWarning::SafeBrowsingWarning):
(WebKit::learnMore): Deleted.
(WebKit::SafeBrowsingResult::SafeBrowsingResult): Deleted.

• UIProcess/Cocoa/WKSafeBrowsingWarning.h:
• UIProcess/Cocoa/WKSafeBrowsingWarning.mm:

(-[WKSafeBrowsingWarning initWithFrame:safeBrowsingWarning:completionHandler:]):
(-[WKSafeBrowsingWarning addContent]):
(-[WKSafeBrowsingWarning showDetailsClicked]):
(-[WKSafeBrowsingWarning clickedOnLink:]):
(-[WKSafeBrowsingTextView initWithAttributedString:forWarning:]):
(confirmMalwareSentinel): Deleted.
(visitUnsafeWebsiteSentinel): Deleted.
(replace): Deleted.
(addLinkAndReplace): Deleted.
(reportAnErrorURL): Deleted.
(malwareDetailsURL): Deleted.
(titleText): Deleted.
(warningText): Deleted.
(detailsText): Deleted.
(-[WKSafeBrowsingWarning initWithFrame:safeBrowsingResult:completionHandler:]): Deleted.

• UIProcess/Cocoa/WebPageProxyCocoa.mm:

(WebKit::WebPageProxy::beginSafeBrowsingCheck):

• UIProcess/Cocoa/WebViewImpl.h:
• UIProcess/Cocoa/WebViewImpl.mm:

(WebKit::WebViewImpl::showSafeBrowsingWarning):

• UIProcess/PageClient.h:

(WebKit::PageClient::showSafeBrowsingWarning):

• UIProcess/SafeBrowsingResult.h: Removed.
• UIProcess/SafeBrowsingWarning.h: Copied from Source/WebKit/UIProcess/SafeBrowsingResult.h.

(WebKit::SafeBrowsingWarning::create):
(WebKit::SafeBrowsingWarning::title const):
(WebKit::SafeBrowsingWarning::warning const):
(WebKit::SafeBrowsingWarning::details const):
(WebKit::SafeBrowsingResult::create): Deleted.
(WebKit::SafeBrowsingResult::isPhishing const): Deleted.
(WebKit::SafeBrowsingResult::isMalware const): Deleted.
(WebKit::SafeBrowsingResult::isUnwantedSoftware const): Deleted.
(WebKit::SafeBrowsingResult::url const): Deleted.
(WebKit::SafeBrowsingResult::provider const): Deleted.
(WebKit::SafeBrowsingResult::localizedProviderName const): Deleted.
(WebKit::SafeBrowsingResult::malwareDetailsURLBase const): Deleted.
(WebKit::SafeBrowsingResult::reportAnErrorURLBase const): Deleted.
(WebKit::SafeBrowsingResult::learnMoreURL const): Deleted.
(WebKit::SafeBrowsingResult::needsSafeBrowsingWarning const): Deleted.
(): Deleted.

• UIProcess/WebFramePolicyListenerProxy.cpp:

(WebKit::WebFramePolicyListenerProxy::didReceiveSafeBrowsingResults):
(WebKit::WebFramePolicyListenerProxy::use):

• UIProcess/WebFramePolicyListenerProxy.h:
• UIProcess/WebFrameProxy.cpp:

(WebKit::WebFrameProxy::WebFrameProxy):
(WebKit::WebFrameProxy::setUpPolicyListenerProxy):
(WebKit::WebFrameProxy::didHandleContentFilterUnblockNavigation):

• UIProcess/WebFrameProxy.h:

(WebKit::WebFrameProxy::create):
(WebKit::WebFrameProxy::page const):

• UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::swapToWebProcess):
(WebKit::WebPageProxy::didCreateMainFrame):
(WebKit::WebPageProxy::didCreateSubframe):
(WebKit::WebPageProxy::decidePolicyForNavigationAction):
(WebKit::WebPageProxy::decidePolicyForNewWindowAction):
(WebKit::WebPageProxy::decidePolicyForResponse):

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

(WebKit::PageClientImpl::showSafeBrowsingWarning):

• UIProcess/mac/PageClientImplMac.h:
• UIProcess/mac/PageClientImplMac.mm:

(WebKit::PageClientImpl::showSafeBrowsingWarning):

• UIProcess/mac/WebDataListSuggestionsDropdownMac.mm:
• WebKit.xcodeproj/project.pbxproj:

Tools:

• TestWebKitAPI/Tests/WebKitCocoa/SafeBrowsing.mm:

(TEST):

11:46 AM Changeset in webkit [238189] by Antti Koivisto

• 2 edits in trunk/LayoutTests

REGRESSION(r238178): fast/forms/access-key-mutated.html and fast/forms/access-key-case-insensitive.html are timing out
​https://bugs.webkit.org/show_bug.cgi?id=191642

• TestExpectations: Additionally skip accessibility/mac/selection-notification-focus-change.html

11:27 AM Changeset in webkit [238188] by Wenson Hsieh

• 10 edits in trunk

[iOS] A few drag and drop tests are crashing after r238146
​https://bugs.webkit.org/show_bug.cgi?id=191617

Reviewed by Dean Jackson.

Source/WebKit:

The notion of temporarily suppressing the selection assistant was introduced during iOS drag and drop
development as a way of allowing both the selection view and dropped content snapshot to fade in simultaneously
during a drop in an editable element. r238146 piggy-backs on this mechanism by changing selection suppression
state when an element is focused, when the selection changes, and when an element is blurred, depending on
whether the currently focused element is transparent.

However, in the case where the selection assistant is suppressed due to a running drop animation, if focus moves
to an element that is not fully transparent, we end up prematurely unsuppressing the text selection assistant.
This subsequently causes selection UI to immediately show up after a drop instead of animating in alongside a
snapshot of the inserted document fragment, if the drop moved focus to an editable element.

A number of drag and drop tests on iOS exercised this codepath by dragging content into editable fields and/or
moving content between editable elements in a web view. These tests began to crash due to selection views and
the accompanying callout bar appearing earlier than usual, which triggers an unrelated UIKit assertion in
<​https://webkit.org/b/190401>.

This patch fixes the failing tests by refactoring our selection assistant suppression code. Instead of
maintaining a single BOOL flag indicating whether the selection is suppressed, we use an OptionSet of
SuppressSelectionAssistantReasons, which (at the moment) only include (1) a running drop animation, and (2)
focusing a transparent element. The text selection assistant is considered suppressed when either of the reasons
apply. This allows us to correctly handle a drop animation that occurs simultaneously as an element is focused
without unsuppressing the selection assistant early, and also allows us to handle selection assistant
suppression in more nuanced ways, depending on the suppression reason.

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

(-[WKContentView cleanupInteraction]):
(-[WKContentView _displayFormNodeInputView]):

Only prevent zooming to the focused element during drop if we're suppressing the selection assistant due to
focusing a transparent element. In the case of a drop, we still want to allow scrolling and zooming.

(-[WKContentView canShowNonEmptySelectionView]):
(-[WKContentView hasSelectablePositionAtPoint:]):
(-[WKContentView pointIsNearMarkedText:]):
(-[WKContentView textInteractionGesture:shouldBeginAtPoint:]):
(-[WKContentView _startAssistingKeyboard]):
(-[WKContentView _startAssistingNode:userIsInteracting:blurPreviousNode:changingActivityState:userObject:]):
(-[WKContentView _stopAssistingNode]):
(-[WKContentView _updateChangedSelection:]):
(-[WKContentView _shouldSuppressSelectionCommands]):
(-[WKContentView _beginSuppressingSelectionAssistantForReason:]):
(-[WKContentView _stopSuppressingSelectionAssistantForReason:]):

Add helper methods for adding or removing selection assistant suppression reasons. When the last selection
assistant suppression reason is removed, we activate the selection assistant, and conversely, when the first
suppression reason is added, we deactivate the selection assistant.

(-[WKContentView _didConcludeEditDataInteraction:]):
(-[WKContentView _didPerformDragOperation:]):
(-[WKContentView dropInteraction:performDrop:]):
(-[WKContentView suppressAssistantSelectionView]): Deleted.
(-[WKContentView setSuppressAssistantSelectionView:]): Deleted.

Tools:

Augment these crashing tests to verify that selection commands are suppressed during drop over editable elements
via more robust means. These tests currently hit an assertion when revealing the callout bar too early, because
TestWebKitAPI is not a UI application (see <​https://webkit.org/b/190401>).

Instead of relying on this other bug, directly ask the text input whether it is suppressing selection commands
during a drop, and remember the answer via DragAndDropSimulator.

• TestWebKitAPI/Tests/ios/DragAndDropTestsIOS.mm:

(TestWebKitAPI::TEST):

Add to the existing tests that started failing after r238146.

• TestWebKitAPI/cocoa/DragAndDropSimulator.h:
• TestWebKitAPI/cocoa/TestWKWebView.h:
• TestWebKitAPI/cocoa/TestWKWebView.mm:

(-[TestWKWebView textInputContentView]):

• TestWebKitAPI/ios/DragAndDropSimulatorIOS.mm:

(-[DragAndDropSimulator _resetSimulatedState]):
(-[DragAndDropSimulator _webView:dataInteractionOperationWasHandled:forSession:itemProviders:]):

• TestWebKitAPI/ios/UIKitSPI.h:

11:18 AM Changeset in webkit [238187] by Antti Koivisto

• 2 edits in trunk/LayoutTests

REGRESSION(r238178): fast/forms/access-key-mutated.html and fast/forms/access-key-case-insensitive.html are timing out
​https://bugs.webkit.org/show_bug.cgi?id=191642

• TestExpectations: Skipping for now.

11:00 AM Changeset in webkit [238186] by Wenson Hsieh

• 12 edits in trunk

Source/WebKit:
[Cocoa] [WebKit2] Hook up some more editing actions (-_pasteAndMatchStyle:, -makeTextWritingDirectionNatural:)
​https://bugs.webkit.org/show_bug.cgi?id=191605
<rdar://problem/45813991>

Reviewed by Dean Jackson.

Hook up a couple of editing actions on iOS and macOS. The -pasteAndMatchStyle: selector is privately declared
in UIKit, but does not exist on macOS; as such, add -_pasteAndMatchStyle: as SPI on WKWebView, and
additionally implement the unprefixed version, -pasteAndMatchStyle:, on iOS.

Since -makeTextWritingDirectionNatural: is already declared on both iOS and macOS (SPI on UIResponder and API
on macOS, respectively) we simply implement the unprefixed variant on both platforms.

• Platform/spi/ios/UIKitSPI.h:
• UIProcess/API/Cocoa/WKWebView.mm:
• UIProcess/API/Cocoa/WKWebViewPrivate.h:
• UIProcess/ios/WKContentViewInteraction.h:
• UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView makeTextWritingDirectionNaturalForWebView:]):
(-[WKContentView canPerformActionForWebView:withSender:]):

Tools:
[Cocoa] [WebKit2] Hook up some more editing actions (-pasteAndMatchStyle:, -makeTextWritingDirectionNatural:)
​https://bugs.webkit.org/show_bug.cgi?id=191605
<rdar://problem/45813991>

Reviewed by Dean Jackson.

Adds two new API tests to exercise -_pasteAndMatchStyle: and -makeTextWritingDirectionNatural:.

• TestWebKitAPI/Tests/WebKitCocoa/WKAttachmentTests.mm:

(-[TestWKWebView _synchronouslyExecuteEditCommand:argument:]): Deleted.

• TestWebKitAPI/Tests/WebKitCocoa/WKWebViewEditActions.mm:

(TestWebKitAPI::TEST):

• TestWebKitAPI/cocoa/TestWKWebView.h:
• TestWebKitAPI/cocoa/TestWKWebView.mm:

Pull -_synchronouslyExecuteEditCommand:argument: out of WKAttachmentTests and into shared code, in a new
category on WKWebView declared in TestWKWebView. Also, fix an unrelated typo in a nearby category.

(-[WKWebView _synchronouslyExecuteEditCommand:argument:]):

• TestWebKitAPI/ios/UIKitSPI.h:

10:49 AM Changeset in webkit [238185] by yusukesuzuki@slowstart.org

• 9 edits in trunk/Source/JavaScriptCore

Unreviewed, rolling in CodeCache in r237254
​https://bugs.webkit.org/show_bug.cgi?id=190340

Land the CodeCache part without adding an additional hash value.

• bytecode/UnlinkedFunctionExecutable.cpp:

(JSC::UnlinkedFunctionExecutable::fromGlobalCode):

• bytecode/UnlinkedFunctionExecutable.h:
• parser/SourceCodeKey.h:

(JSC::SourceCodeKey::SourceCodeKey):
(JSC::SourceCodeKey::operator== const):

• runtime/CodeCache.cpp:

(JSC::CodeCache::getUnlinkedGlobalCodeBlock):
(JSC::CodeCache::getUnlinkedGlobalFunctionExecutable):

• runtime/CodeCache.h:
• runtime/FunctionConstructor.cpp:

(JSC::constructFunctionSkippingEvalEnabledCheck):

• runtime/FunctionExecutable.cpp:

(JSC::FunctionExecutable::fromGlobalCode):

• runtime/FunctionExecutable.h:

10:43 AM Changeset in webkit [238184] by pvollan@apple.com

• 22 edits
  2 adds in trunk

REGRESSION (WEBPROCESS_WINDOWSERVER_BLOCKING): requestAnimationFrame Stops Completing
​https://bugs.webkit.org/show_bug.cgi?id=190884

Reviewed by Dean Jackson.

Source/WebCore:

Only notify display refresh monitors with matching display ID.

Test: fast/animation/request-animation-frame-in-two-pages.html

• platform/graphics/DisplayRefreshMonitorManager.cpp:

(WebCore::DisplayRefreshMonitorManager::displayWasUpdated):

• platform/graphics/DisplayRefreshMonitorManager.h:

Source/WebKit:

Since the Web page currently owns the display link on the UI process side, the display link will be
deleted in the UI process when a Web page is closed. This is incorrect, since a display link can be
reused between Web pages in the WebContent process. This patch moves the ownership of the UI process
display links to the Web process object to address this issue. Also, the display ID of the display
being updated is included in the message to the WebContent process.

• UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::close):
(WebKit::WebPageProxy::resetStateAfterProcessExited):

• UIProcess/WebPageProxy.h:
• UIProcess/WebPageProxy.messages.in:
• UIProcess/WebProcessProxy.h:
• UIProcess/WebProcessProxy.messages.in:
• UIProcess/mac/DisplayLink.cpp:

(WebKit::DisplayLink::DisplayLink):
(WebKit::DisplayLink::displayLinkCallback):

• UIProcess/mac/DisplayLink.h:

(WebKit::DisplayLink::displayID const):

• UIProcess/mac/WebPageProxyMac.mm:

(WebKit::WebPageProxy::startDisplayLink): Deleted.
(WebKit::WebPageProxy::stopDisplayLink): Deleted.

• UIProcess/mac/WebProcessProxyMac.mm:

(WebKit::WebProcessProxy::startDisplayLink):
(WebKit::WebProcessProxy::stopDisplayLink):

• WebProcess/WebPage/DrawingArea.h:
• WebProcess/WebPage/DrawingArea.messages.in:
• WebProcess/WebPage/mac/DrawingAreaMac.cpp:

(WebKit::DisplayRefreshMonitorMac::create):
(WebKit::DisplayRefreshMonitorMac::DisplayRefreshMonitorMac):
(WebKit::DisplayRefreshMonitorMac::~DisplayRefreshMonitorMac):
(WebKit::DisplayRefreshMonitorMac::requestRefreshCallback):
(WebKit::DrawingArea::createDisplayRefreshMonitor):
(WebKit::DrawingArea::displayWasRefreshed): Deleted.

• WebProcess/WebProcess.h:
• WebProcess/WebProcess.messages.in:
• WebProcess/cocoa/WebProcessCocoa.mm:

(WebKit::WebProcess::displayWasRefreshed):

LayoutTests:

• TestExpectations:
• fast/animation/request-animation-frame-in-two-pages-expected.txt: Added.
• fast/animation/request-animation-frame-in-two-pages.html: Added.
• platform/mac-wk2/TestExpectations:

10:13 AM Changeset in webkit [238183] by Chris Dumez

• 4 edits in trunk/Source/WebKit

Client should be able to disable PSON even if the experimental feature is on by default
​https://bugs.webkit.org/show_bug.cgi?id=191634

Reviewed by Dean Jackson.

Instead of overriding the APIProcessPoolConfiguration's m_processSwapsOnNavigation flag to true
when the experimental is enabled, use a separate flag to store this information. If the client
explicitly sets the APIProcessPoolConfiguration's m_processSwapsOnNavigation flag then we obey
the client's request, otherwise, we fall back to using the state from experimental features.

This allows:

• API tests to explicitely disable PSON / Process prewarming even if those are on by default in experimental features.
• If the client does not set those flags on the APIProcessPoolConfiguration (Safari for e.g.), then the experimental feature flag still fully controls the feature.

• UIProcess/API/APIProcessPoolConfiguration.cpp:

(API::ProcessPoolConfiguration::copy):

• UIProcess/API/APIProcessPoolConfiguration.h:
• UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::createWebPage):

9:53 AM Changeset in webkit [238182] by youenn@apple.com

• 12 edits in trunk

Convert libwebrtc error types to DOM exceptions
​https://bugs.webkit.org/show_bug.cgi?id=191590

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

• web-platform-tests/webrtc/RTCConfiguration-rtcpMuxPolicy-expected.txt:
• web-platform-tests/webrtc/RTCPeerConnection-setLocalDescription-answer-expected.txt:
• web-platform-tests/webrtc/RTCRtpTransceiver.https-expected.txt:

Source/ThirdParty/libwebrtc:

• Configurations/libwebrtc.iOS.exp:
• Configurations/libwebrtc.iOSsim.exp:
• Configurations/libwebrtc.mac.exp:

Source/WebCore:

Make use of overloaded callback method that provides an error type.
This type is then used to create a DOM exception with the correct type.
Covered by existing tests.

• Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:

(WebCore::LibWebRTCMediaEndpoint::doSetRemoteDescription):
(WebCore::LibWebRTCMediaEndpoint::createSessionDescriptionFailed):
(WebCore::LibWebRTCMediaEndpoint::setLocalSessionDescriptionFailed):
(WebCore::LibWebRTCMediaEndpoint::setRemoteSessionDescriptionFailed):

• Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.h:
• Modules/mediastream/libwebrtc/LibWebRTCObservers.h:

(WebCore::toExceptionCode):

9:50 AM Changeset in webkit [238181] by youenn@apple.com

• 4 edits in trunk/Source/WebCore

Allow to remove MediaStreamPrivate observers when iterating over observers
​https://bugs.webkit.org/show_bug.cgi?id=187256

Reviewed by Eric Carlson.

Migrate the observer list from a Vector to a HashSet.
This is more robust to multiple observing and keeping of order of observers is not required.
Copy the set of observers to a vector before iterating over it.
This allows to remove an observer while iterating, which is now used in UserMediaRequest.

Covered by existing tests.

• Modules/mediastream/UserMediaRequest.cpp:

(WebCore::UserMediaRequest::mediaStreamIsReady):

• platform/mediastream/MediaStreamPrivate.cpp:

(WebCore::MediaStreamPrivate::addObserver):
(WebCore::MediaStreamPrivate::removeObserver):
(WebCore::MediaStreamPrivate::forEachObserver const):
(WebCore::MediaStreamPrivate::updateActiveState):
(WebCore::MediaStreamPrivate::addTrack):
(WebCore::MediaStreamPrivate::removeTrack):
(WebCore::MediaStreamPrivate::characteristicsChanged):

• platform/mediastream/MediaStreamPrivate.h:

9:35 AM Changeset in webkit [238180] by youenn@apple.com

• 9 edits in trunk

Calling removeTrack on different RTCPeerConnection should throw InvalidAccessError
​https://bugs.webkit.org/show_bug.cgi?id=191603

Reviewed by Chris Dumez.

LayoutTests/imported/w3c:

• web-platform-tests/webrtc/RTCPeerConnection-removeTrack.https-expected.txt:

Source/WebCore:

Make sure to check that the sender peer connection backend is matching.
Covered by rebased WPT test.

• Modules/mediastream/RTCPeerConnection.cpp:

(WebCore::RTCPeerConnection::removeTrack):

• Modules/mediastream/RTCRtpSender.cpp:

(WebCore::RTCRtpSender::isCreatedBy const):

• Modules/mediastream/RTCRtpSender.h:

LayoutTests:

Removed obsolete test.

• fast/mediastream/RTCPeerConnection-add-removeTrack-expected.txt:
• fast/mediastream/RTCPeerConnection-add-removeTrack.html:

8:58 AM Changeset in webkit [238179] by Chris Dumez

• 9 edits in trunk

WebKit.ApplicationManifestBasic API test is failing when enabling PSON
​https://bugs.webkit.org/show_bug.cgi?id=191602

Reviewed by Alex Christensen.

Source/WebKit:

Add support for process swapping for a [WKWebView loadHTML:] load by storing
the necessary data on the API::Navigation and doing a loadData() instead of
a loadRequest() after process swapping when this data is present on the
navigation.

• UIProcess/API/APINavigation.cpp:

(API::Navigation::Navigation):

• UIProcess/API/APINavigation.h:

(API::Navigation::create):
(API::Navigation::substituteData const):

• UIProcess/WebNavigationState.cpp:

(WebKit::WebNavigationState::createLoadDataNavigation):

• UIProcess/WebNavigationState.h:
• UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::loadData):
(WebKit::WebPageProxy::continueNavigationInNewProcess):

Tools:

Add API test coverage.

• TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

8:53 AM Changeset in webkit [238178] by Antti Koivisto

• 3 edits in trunk/Source/WebKit

Align Mac WK2 layer flush throttling with iOS
​https://bugs.webkit.org/show_bug.cgi?id=191632

Reviewed by Zalan Bujtas.

Currently Mac WK2 uses WebCore side throttling implementation in RenderLayerCompositor. This code has
throttling timer per-frame while the actual decision making and layer flushes itself are per page. These
timers generate way more flushes than expected in presence of multiple frames. There are also bugs in how
flushing state is updated when frames are created dynamically.

On iOS WK2 throttling is implemented on WebKit side and controlled by a per-page timer. Recent fixes also
make this implementation visually fast. We should align the Mac implementation and eventually unify them.

This patch implements throttling in TiledCoreAnimationDrawingArea mirroring the iOS RemoteLayerTreeDrawingArea
implementation. There are some adjustments for platform differences (local vs remote layers) and we continue
using runloop observer for the actual flushes. Timings are as in the existing Mac code.

The patch appears to be a significant performance progression.

• WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.h:
• WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:

(WebKit::TiledCoreAnimationDrawingArea::TiledCoreAnimationDrawingArea):
(WebKit::TiledCoreAnimationDrawingArea::setLayerTreeStateIsFrozen):

Schedule an immediate flush when layers are unfrozen.

(WebKit::TiledCoreAnimationDrawingArea::scheduleCompositingLayerFlush):
(WebKit::TiledCoreAnimationDrawingArea::scheduleCompositingLayerFlushImmediately):

Track pending flushes with m_hasPendingFlush bit.
Delay flush if the flush throttling timer is active. Start it if we throttling but it is not active yet.

(WebKit::TiledCoreAnimationDrawingArea::flushLayers):

Clear m_hasPendingFlush if the flush succeeded.
Restart the throttling timer if we are still throttling.
Manage runloop observer invalidation here instead of the caller (and stop returning value).

(WebKit::TiledCoreAnimationDrawingArea::layerFlushRunLoopCallback):
(WebKit::TiledCoreAnimationDrawingArea::adjustLayerFlushThrottling):

Returning 'true' here disables WebCore side throttling code.

Start or stop the throttling timer on state changes.

(WebKit::TiledCoreAnimationDrawingArea::layerFlushThrottlingIsActive const):

This is used to control style and layout timers on WebCore side. Return false on frozen
state since unfreezing depends on style and layout.

(WebKit::TiledCoreAnimationDrawingArea::startLayerFlushThrottlingTimer):
(WebKit::TiledCoreAnimationDrawingArea::layerFlushThrottlingTimerFired):

8:24 AM Changeset in webkit [238177] by Jonathan Bedard

• 6 edits
  1 add in trunk/Tools

webkitpy: Refactor port code for devices
​https://bugs.webkit.org/show_bug.cgi?id=191568
<rdar://problem/46005821>

Reviewed by Alex Christensen.

watchOS and iOS device ports share many characteristics in testing, mostly due
to the fact that both are devices.

• Scripts/webkitpy/port/device_port.py: Added.

(DevicePort): Add base class that all simulated/physical device ports share.
(DevicePort.init):
(DevicePort.driver_cmd_line_for_logging): Moved from iOS port.
(DevicePort._generate_all_test_configurations): Ditto.
(DevicePort.child_processes): Ditto.
(DevicePort.driver_name): Ditto.
(DevicePort.target_host): Generalize worker access to attached devices and
simulated devices based on the device manager.
(DevicePort.devices): Return a list of all devices associated with this port.
(DevicePort._create_devices):
(DevicePort._build_driver_flags): Moved from iOSSimulatorPort/iOSDevicePort.
(DevicePort._install): Moved from iOS port.
(DevicePort.setup_test_run): Ditto.
(DevicePort.clean_up_test_run): Ditto.
(DevicePort.did_spawn_worker): Ditto.
(DevicePort.setup_environ_for_server): Ditto.

• Scripts/webkitpy/port/ios.py:

(IOSPort):
(IOSPort._device_for_worker_number_map): Replaced by DevicePort.devices().
(IOSPort.driver_cmd_line_for_logging): Moved to DevicePort.
(IOSPort.driver_name): Ditto.
(IOSPort._generate_all_test_configurations): Ditto.
(IOSPort.child_processes): Ditto.
(IOSPort._testing_device): Ditto.
(IOSPort.target_host): Ditto.
(IOSPort._create_devices): Ditto.
(IOSPort.setup_test_run): Ditto.
(IOSPort.clean_up_test_run): Ditto.
(IOSPort.did_spawn_worker): Ditto.

• Scripts/webkitpy/port/ios_device.py:

(IOSDevicePort): IOSDevicePort uses a device manager from apple_additions.
(IOSDevicePort._look_for_all_crash_logs_in_log_dir): Use DevicePort.devices().
(IOSDevicePort._get_crash_log): Ditto.
(IOSDevicePort.ios_version): Ditto.
(IOSDevicePort._create_devices): Ditto.
(IOSDevicePort._device_for_worker_number_map): Replaced by DevicePort.devices().
(IOSSimulatorPort._build_driver_flags): Moved to DevicePort.

• Scripts/webkitpy/port/ios_device_unittest.py:

(IOSDeviceTest.test_get_crash_log):

• Scripts/webkitpy/port/ios_simulator.py:

(IOSSimulatorPort): IOSSimulatorPort uses the SimulatedDeviceManager.
(IOSSimulatorPort._create_devices): Clean up logging statement.
(IOSSimulatorPort.setup_environ_for_server): Some of environment set-up is shared
across all devices.
(IOSSimulatorPort._device_for_worker_number_map): Replaced by DevicePort.devices().
(IOSSimulatorPort._build_driver_flags): Moved to DevicePort.

• Scripts/webkitpy/port/ios_simulator_unittest.py:

(IOSSimulatorTest.test_32bit):
(IOSSimulatorTest.test_64bit):

8:14 AM Changeset in webkit [238176] by Fujii Hironori

• 4 edits
  1 move
  1 delete in trunk/Source/WebCore

[curl] Unify CookieJarCurlDatabase and the abstract class CookieJarCurl
​https://bugs.webkit.org/show_bug.cgi?id=191620

Reviewed by Alex Christensen.

Remove a abstract class CookieJarCurl which is not needed anymore.
And, rename CookieJarCurlDatabase to CookieJarCurl.

No new tests because there's no behaviour change in WebCore.

• platform/Curl.cmake: Replaced CookieJarCurlDatabase.cpp with CookieJarCurl.cpp.
• platform/network/curl/CookieJarCurl.cpp: Renamed from Source/WebCore/platform/network/curl/CookieJarCurlDatabase.cpp.
• platform/network/curl/CookieJarCurl.h: Merged CookieJarCurl.h and CookieJarCurlDatabase.h.
• platform/network/curl/CookieJarCurlDatabase.h: Removed.
• platform/network/curl/NetworkStorageSessionCurl.cpp:

(WebCore::NetworkStorageSession::NetworkStorageSession): Replaced CookieJarCurlDatabase with CookieJarCurl.

8:13 AM Changeset in webkit [238175] by chris.reid@sony.com

• 2 edits in trunk/Source/WebCore

[WPE] Remove glib usage in PlatformKeyboardEventWPE.cpp
​https://bugs.webkit.org/show_bug.cgi?id=191606

Reviewed by Michael Catanzaro.

No behavior change.

Use StringBuilder::append(UChar32) as a generic way to convert a uint32_t code point to WTFString.

• platform/wpe/PlatformKeyboardEventWPE.cpp:

(WebCore::PlatformKeyboardEvent::keyValueForWPEKeyCode):
(WebCore::PlatformKeyboardEvent::singleCharacterString):

7:50 AM Changeset in webkit [238174] by Chris Dumez

• 4 edits in trunk

WebKit.WKHTTPCookieStoreWithoutProcessPool API test is failing with process prewarming is enabled
​https://bugs.webkit.org/show_bug.cgi?id=191624

Reviewed by Alex Christensen.

Source/WebKit:

Stop setting the WebProcessPool's primary data store (m_websiteDataStore) to the default one in
WebProcessPool::prewarmProcess(). We did not really need to, we can pass the default data store
to the new WebPageProxy without having to set m_websiteDataStore. m_websiteDataStore only gets
set upon constructor if thr default data store already exists or later on when creating a WebPage
that uses the default data store.

In the case of the API test, the following was happening:

1. Create an ephemeral data store EDS
2. Create a WebView V1 using datastore EDS
3. Do a load in V1
4. Process prewarming would kick in and wrongly associated V1's process pool PP1 with the default data store
5. Create/Get the default datastore and set a few cookies on it
6. Create a WebView V2 using default datastore and a fresh new process pool PP2
7. Do a load in V2 and expect the cookies to be there

In HTTPCookieStore::setCookie() that is called at step 5, we call:
m_owningDataStore->processPoolForCookieStorageOperations()

In this case, m_owningDataStore is the default datastore and this call would previously return null because
there is no WebProcessPool yet associated with the default datastore. However, with the process prewarming
bug at step 4, the process pool PP1 would be returned since it was wrongly associated with the default
data store. As a result, we would call setCookie() on PP1's WebCookieManagerProxy and this would fail
because PP1's network process knows nothing about this session ID (it was only ever used with an ephemeral
session).

• UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::prewarmProcess):

Tools:

Add API test coverage.

• TestWebKitAPI/Tests/WebKitCocoa/WKHTTPCookieStore.mm:

(runWKHTTPCookieStoreWithoutProcessPool):
(TEST):

7:31 AM Changeset in webkit [238173] by Alan Bujtas

• 5 edits in trunk/Source/WebCore

[LFC][IFC] Construct dedicated runs when the inline element requires it (part 2)
​https://bugs.webkit.org/show_bug.cgi?id=191623

Reviewed by Antti Koivisto.

This patch expands the breaking behaviour to support separate start/end breaks.

<span>parent </span><span style="padding: 10px;">start<span> middle </span>end</span><span> parent</span>

input to line breaking -> <parent start middle end parent>
output of line breaking (considering infinite constraint) -> <parent start middle end parent>
due to padding, final runs -> <parent><start middle end><parent>

"parent" -> n/a
"start" -> BreakAtStart
" middle " -> n/a
"end" -> BreakAtEnd
"parent" -> n/a

Another example:
<span>parent </span><span style="padding-right: 10px;">start<span> middle </span>end</span><span> parent</span>

line breaking -> <parent start middle end parent>
due to padding-right, final runs -> <parent start middle end><parent>

"parent" -> n/a
"start" -> n/a
" middle " -> n/a
"end" -> BreakAtEnd
"parent" -> n/a

• layout/inlineformatting/InlineFormattingContext.cpp:

(WebCore::Layout::InlineFormattingContext::splitInlineRunIfNeeded const):
(WebCore::Layout::InlineFormattingContext::collectInlineContent const): Move to a recursive algorithm (which is fine, inline contents don't tend to be too deep)
(WebCore::Layout::InlineFormattingContext::contentRequiresSeparateRun const): Deleted.

• layout/inlineformatting/InlineFormattingContext.h:
• layout/inlineformatting/InlineFormattingState.cpp:

(WebCore::Layout::InlineFormattingState::detachingRules const):

• layout/inlineformatting/InlineFormattingState.h:

(WebCore::Layout::InlineFormattingState::lastInlineItem const):
(WebCore::Layout::InlineFormattingState::addDetachingRule):

7:21 AM Changeset in webkit [238172] by youenn@apple.com

• 6 edits in trunk

Add support for RTCRtpCodecParameters.sdpFmtpLine
​https://bugs.webkit.org/show_bug.cgi?id=191591

Reviewed by Eric Carlson.

LayoutTests/imported/w3c:

• web-platform-tests/webrtc/protocol/video-codecs.https-expected.txt:

Source/WebCore:

Covered by rebased test.

• Modules/mediastream/RTCRtpCodecParameters.h:
• Modules/mediastream/RTCRtpCodecParameters.idl:
• Modules/mediastream/libwebrtc/LibWebRTCUtils.cpp:

(WebCore::toRTCCodecParameters):

6:57 AM Changeset in webkit [238171] by youenn@apple.com

• 15 edits in trunk

Add support for transport and peerConnection stats
​https://bugs.webkit.org/show_bug.cgi?id=191592

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

• web-platform-tests/webrtc/RTCPeerConnection-getStats.https-expected.txt:
• web-platform-tests/webrtc/RTCRtpReceiver-getStats.https-expected.txt:
• web-platform-tests/webrtc/RTCRtpSender-getStats.https-expected.txt:
• web-platform-tests/webrtc/getstats-expected.txt:

Source/ThirdParty/libwebrtc:

• Configurations/libwebrtc.iOS.exp:
• Configurations/libwebrtc.iOSsim.exp:
• Configurations/libwebrtc.mac.exp:

Source/WebCore:

Covered by rebased tests.

• Modules/mediastream/RTCStatsReport.h:

(WebCore::RTCStatsReport::TransportStats::TransportStats):
(WebCore::RTCStatsReport::PeerConnectionStats::PeerConnectionStats):

• Modules/mediastream/RTCStatsReport.idl:
• Modules/mediastream/libwebrtc/LibWebRTCStatsCollector.cpp:

(WebCore::fillRTCTransportStats):
(WebCore::fillRTCPeerConnectionStats):
(WebCore::LibWebRTCStatsCollector::OnStatsDelivered):

LayoutTests:

• webrtc/video-stats.html:

5:13 AM Changeset in webkit [238170] by ajuma@chromium.org

• 3 edits
  2 adds in trunk

Transform of composited layer not updated when layer also needs repaint
​https://bugs.webkit.org/show_bug.cgi?id=191598

Reviewed by Simon Fraser.

Source/WebCore:

Trigger a compositing geometry update whenever a RenderLayer's transform changes,
even when other parts of its style have changed in a way that produces a
StyleDifference greater than RecompositeLayer.

Test: compositing/geometry/transform-and-repaint-updates-geometry.html

• rendering/RenderLayerCompositor.cpp:

(WebCore::RenderLayerCompositor::layerStyleChanged):

LayoutTests:

• compositing/geometry/transform-and-repaint-updates-geometry-expected.html: Added.
• compositing/geometry/transform-and-repaint-updates-geometry.html: Added.

4:01 AM Changeset in webkit [238169] by Devin Rousso

• 3 edits in trunk/Source/WebInspectorUI

Web Inspector: Canvas: don't start with auto-capture enabled
​https://bugs.webkit.org/show_bug.cgi?id=191553

Reviewed by Matt Baker.

• UserInterface/Views/CanvasOverviewContentView.js:

(WI.CanvasOverviewContentView.prototype._setRecordingAutoCaptureFrameCount):
(WI.CanvasOverviewContentView.prototype._updateRecordingAutoCaptureInputElementSize):
(WI.CanvasOverviewContentView.prototype._handleRecordingAutoCaptureInput):

• UserInterface/Views/CanvasOverviewContentView.css:

(.navigation-bar > .item.canvas-recording-auto-capture > label > input):

1:26 AM Changeset in webkit [238168] by Devin Rousso

• 4 edits in trunk/Source/WebInspectorUI

Web Inspector: Network Graphs are missing minimum sizes, might have no graph at all
​https://bugs.webkit.org/show_bug.cgi?id=191208

Reviewed by Joseph Pecoraro.

WI.NetworkTableContentView used to rely on the WI.timelineManager.persistentNetworkTimeline
for events when a new resource is added. It also listened for WI.Frame.Event.MainResourceDidChange
on it's own, which was also listened to by WI.timelineManager.persistentNetworkTimeline.
Due to the order in which these listeners are added, the new main resource would get added
to the WI.timelineManager.persistentNetworkTimeline (and the WI.NetworkTableContentView
shortly after), and right after that the WI.NetworkTableContentView would reset the graph
in it's own listener for WI.Frame.Event.MainResourceDidChange.

This patch removes WI.timelineManager.persistentNetworkTimeline and instead makes it so
that WI.NetworkTableContentView listens for resource added and main frame changed events
on its own (similar to other views that follow this pattern), ensuring that there are no
event races.

Also relaxes the "requirement" that prevented 0 width blocks from being drawn, allowing
requests served from memory/disk to appear in the graph (their time is effectively 0).

• UserInterface/Views/NetworkTableContentView.js:

(WI.NetworkTableContentView):
(WI.NetworkTableContentView.prototype.closed):
(WI.NetworkTableContentView.prototype._populateWaterfallGraph):
(WI.NetworkTableContentView.prototype._populateWaterfallGraph.appendBlock):
(WI.NetworkTableContentView.prototype._updateWaterfallTimeRange): Added.
(WI.NetworkTableContentView.prototype._resourceLoadingDidFinish):
(WI.NetworkTableContentView.prototype._resourceLoadingDidFail):
(WI.NetworkTableContentView.prototype._handleResourceAdded): Added.
(WI.NetworkTableContentView.prototype._insertResourceAndReloadTable):
(WI.NetworkTableContentView.prototype._handleNodeDidFireEvent):
(WI.NetworkTableContentView.prototype._handleNodeLowPowerChanged):

• UserInterface/Views/NetworkTableContentView.css:

(.network-table .waterfall .block:matches(.mouse-tracking, .filler) + .block:not(.mouse-tracking, .filler)): Added.
(.network-table .waterfall .block.filler + .block): Deleted.

• UserInterface/Controllers/TimelineManager.js:

(WI.TimelineManager):
(WI.TimelineManager.prototype._mainResourceDidChange):
(WI.TimelineManager.prototype._resourceWasAdded):
(WI.TimelineManager.prototype.get persistentNetworkTimeline): Deleted.

Nov 13, 2018:

11:36 PM Changeset in webkit [238167] by Dewei Zhu

• 7 edits in trunk/Websites/perf.webkit.org

Add cache for CommitLog objects to avoid refetching same commit.
​https://bugs.webkit.org/show_bug.cgi?id=191621

Reviewed by Ryosuke Niwa.

Added a cache for fully fetched commit log objects to avoid refetching.

• public/v3/models/commit-log.js:

(CommitLog): Added assertion for id.
Removed unused 'remoteId' as it has been removed since r198479.
(CommitLog.async.fetchBetweenRevisions): Turned it into async function.
(CommitLog.async.fetchForSingleRevision): Added the logic to check cache before fetching.
(CommitLog._constructFromRawData): Added logic to add entries to cache.

• public/v3/models/commit-set.js: Fixed measurement set not passing commit id while constructing a

commit log object.

• public/v3/models/repository.js: Added the ability to track fetched commit for certain repository.

(Repository.commitForRevision): Returns a commit for given revision.
(Repository.setCommitForRevision): Sets commit for a given revision.

• unit-tests/commit-log-tests.js: Added unit tests for this change.

Fixed existing tests.

• unit-tests/commit-set-range-bisector-tests.js: Fixed unit tests.
• unit-tests/commit-set-tests.js: Fixed unit tests.

10:54 PM Changeset in webkit [238166] by jiewen_tan@apple.com

• 31 edits
  5 copies
  1 move
  24 adds in trunk

[WebAuthN] Support CTAP HID authenticators on macOS
​https://bugs.webkit.org/show_bug.cgi?id=188623
<rdar://problem/43353777>

Reviewed by Brent Fulgham and Chris Dumez.

Source/WebCore:

This patch removes AuthenticatorCoordinatorClient::~AuthenticatorCoordinatorClient to ignore
any incompleted CompletionHandlers as calling them in destructors could cause unexpected cyclic
dependency. Also, it adds a hack to temporarily deal with nullable userhandle.

Tests: http/wpt/webauthn/ctap-hid-failure.https.html

http/wpt/webauthn/ctap-hid-success.https.html
http/wpt/webauthn/public-key-credential-create-failure-hid-silent.https.html
http/wpt/webauthn/public-key-credential-create-failure-hid.https.html
http/wpt/webauthn/public-key-credential-create-success-hid.https.html
http/wpt/webauthn/public-key-credential-get-failure-hid-silent.https.html
http/wpt/webauthn/public-key-credential-get-failure-hid.https.html
http/wpt/webauthn/public-key-credential-get-success-hid.https.html

• Modules/webauthn/AuthenticatorCoordinatorClient.cpp:

(WebCore::AuthenticatorCoordinatorClient::~AuthenticatorCoordinatorClient): Deleted.

• Modules/webauthn/AuthenticatorCoordinatorClient.h:
• Modules/webauthn/PublicKeyCredentialCreationOptions.h:
• Modules/webauthn/fido/DeviceResponseConverter.cpp:

(fido::readCTAPGetAssertionResponse):

• Modules/webauthn/fido/FidoConstants.h:

Source/WebKit:

This patch introduces a primitive support of CTAP HID authenticators for WebAuthN in macOS.
It involves low level HID device management&communication, high level CTAP HID authenticator
management&communication, and mock testing. The above three aspects will be covered in details:
1) Low level HID device management&communication: HidService&HidConnection
It relies on IOHIDManager to discover appropriate hid devices by passing a matching dictionary:
{ PrimaryUsagePage: 0xf1d0, PrimaryUsage: 0x01}. For communication, it utilizes HID reports.
To send a report, it calls IOHIDDeviceSetReport since the async version is not implemented.
To recieve a report, it calls IOHIDDeviceRegisterInputReportCallback to asynchronously wait
for incoming reports.
Here is the corresponding reference:
​https://developer.apple.com/library/archive/documentation/DeviceDrivers/Conceptual/HID/new_api_10_5/tn2187.html#//apple_ref/doc/uid/TP40000970-CH214-SW2
2) High level CTAP HID authenticator management&communication: HidService&CtapHidDriver
Whenever an appropriate hid device is discovered by IOHIDManager, an AuthenticatorGetInfo command
is sent to the device to determine properties of the authenticator, says, which version of protocol
it supports, i.e. CTAP or U2F. So far, we only support CTAP authenticators. Once the authenticator
is determined to support CTAP, we then instantiate CtapHidAuthenticator which will then take care
of even higher level WebAuthN requests&responses.
Binaries are constructed and packaged according to the CTAP HID porotocol. CtapHidDriver takes care
of concurrency and channels, i.e. allocating channel and establishing the actual request/response
transaction. At the meantime, CtapHidDriver::Worker is then responsible for each single transaction.
Here is the corresponding reference:
​https://fidoalliance.org/specs/fido-v2.0-ps-20170927/fido-client-to-authenticator-protocol-v2.0-ps-20170927.html#usb.
3) Mock Testing: MockHidService & MockHidConnection
A CTAP hid authenticator is simulated within MockHidConnection with options of specifying specific
error scenarios and of course could take care of successful cases. Four stages are presented in the
simulated authenticator to reflect: a) allocating channel for AuthenticatorGetInfo, b) sending
AuthenticatorGetInfo, c) allocating channel for actual request, and d) sending the actual request.

Besides implementing the above, it also does a few other things:
1) Make AuthenticatorManager::clearState asynchronous to avoid cyclic dependency:
Authenticator::returnResponse => AuthenticatorManager::respondReceived => AuthenticatorManager::clearState
=> Authenticator::~Authenticator.
2) Reorganize unified build sources to make it clear that which files are .mm and which are .cpp.
3) Import LocalAuthentication.framework in LocalAuthenticationSoftLink instead of being scattered.

• Sources.txt:
• SourcesCocoa.txt:
• Shared/CoordinatedGraphics/threadedcompositor/ThreadedCompositor.h:
• UIProcess/API/C/WKWebsiteDataStoreRef.cpp:

(WKWebsiteDataStoreSetWebAuthenticationMockConfiguration):

• UIProcess/WebAuthentication/AuthenticatorManager.cpp:

(WebKit::AuthenticatorManagerInternal::collectTransports):
(WebKit::AuthenticatorManager::clearStateAsync):
(WebKit::AuthenticatorManager::respondReceived):
(WebKit::AuthenticatorManager::initTimeOutTimer):

• UIProcess/WebAuthentication/AuthenticatorManager.h:
• UIProcess/WebAuthentication/AuthenticatorTransportService.cpp:

(WebKit::AuthenticatorTransportService::create):
(WebKit::AuthenticatorTransportService::createMock):
(WebKit::AuthenticatorTransportService::startDiscovery):
(WebKit::AuthenticatorTransportService::startDiscovery const): Deleted.

• UIProcess/WebAuthentication/AuthenticatorTransportService.h:
• UIProcess/WebAuthentication/Cocoa/HidConnection.h: Copied from Source/WebKit/UIProcess/WebAuthentication/AuthenticatorTransportService.h.
• UIProcess/WebAuthentication/Cocoa/HidConnection.mm: Added.

(WebKit::reportReceived):
(WebKit::HidConnection::HidConnection):
(WebKit::HidConnection::~HidConnection):
(WebKit::HidConnection::initialize):
(WebKit::HidConnection::terminate):
(WebKit::HidConnection::send):
(WebKit::HidConnection::registerDataReceivedCallback):
(WebKit::HidConnection::unregisterDataReceivedCallback):
(WebKit::HidConnection::receiveReport):
(WebKit::HidConnection::consumeReports):
(WebKit::HidConnection::registerDataReceivedCallbackInternal):

• UIProcess/WebAuthentication/Cocoa/HidService.h: Copied from Source/WebKit/UIProcess/WebAuthentication/Cocoa/LocalService.h.
• UIProcess/WebAuthentication/Cocoa/HidService.mm: Added.

(WebKit::deviceAddedCallback):
(WebKit::deviceRemovedCallback):
(WebKit::HidService::HidService):
(WebKit::HidService::~HidService):
(WebKit::HidService::startDiscoveryInternal):
(WebKit::HidService::platformStartDiscovery):
(WebKit::HidService::createHidConnection const):
(WebKit::HidService::deviceAdded):
(WebKit::HidService::continueAddDeviceAfterGetInfo):

• UIProcess/WebAuthentication/Cocoa/LocalAuthenticationSoftLink.h:
• UIProcess/WebAuthentication/Cocoa/LocalConnection.mm:
• UIProcess/WebAuthentication/Cocoa/LocalService.h:
• UIProcess/WebAuthentication/Cocoa/LocalService.mm:

(WebKit::LocalService::startDiscoveryInternal):
(WebKit::LocalService::startDiscoveryInternal const): Deleted.

• UIProcess/WebAuthentication/Mock/MockAuthenticatorManager.cpp:

(WebKit::MockAuthenticatorManager::respondReceivedInternal):

• UIProcess/WebAuthentication/Mock/MockHidConnection.cpp: Added.

(WebKit::MockHidConnection::MockHidConnection):
(WebKit::MockHidConnection::initialize):
(WebKit::MockHidConnection::terminate):
(WebKit::MockHidConnection::send):
(WebKit::MockHidConnection::registerDataReceivedCallbackInternal):
(WebKit::MockHidConnection::assembleRequest):
(WebKit::MockHidConnection::parseRequest):
(WebKit::MockHidConnection::feedReports):
(WebKit::MockHidConnection::stagesMatch const):
(WebKit::MockHidConnection::shouldContinueFeedReports):
(WebKit::MockHidConnection::continueFeedReports):

• UIProcess/WebAuthentication/Mock/MockHidConnection.h: Copied from Source/WebKit/UIProcess/WebAuthentication/Mock/MockLocalConnection.h.
• UIProcess/WebAuthentication/Mock/MockHidService.cpp: Copied from Source/WebKit/UIProcess/WebAuthentication/Mock/MockLocalService.cpp.

(WebKit::MockHidService::MockHidService):
(WebKit::MockHidService::platformStartDiscovery):
(WebKit::MockHidService::createHidConnection const):

• UIProcess/WebAuthentication/Mock/MockHidService.h: Copied from Source/WebKit/UIProcess/WebAuthentication/Mock/MockLocalConnection.h.
• UIProcess/WebAuthentication/Mock/MockLocalConnection.h:
• UIProcess/WebAuthentication/Mock/MockLocalConnection.mm:
• UIProcess/WebAuthentication/Mock/MockLocalService.mm: Renamed from Source/WebKit/UIProcess/WebAuthentication/Mock/MockLocalService.cpp.

(WebKit::MockLocalService::MockLocalService):
(WebKit::MockLocalService::platformStartDiscovery const):
(WebKit::MockLocalService::createLocalConnection const):

• UIProcess/WebAuthentication/Mock/MockWebAuthenticationConfiguration.h:
• UIProcess/WebAuthentication/fido/CtapHidAuthenticator.cpp: Added.

(WebKit::CtapHidAuthenticator::CtapHidAuthenticator):
(WebKit::CtapHidAuthenticator::makeCredential):
(WebKit::CtapHidAuthenticator::continueMakeCredentialAfterResponseReceived const):
(WebKit::CtapHidAuthenticator::getAssertion):
(WebKit::CtapHidAuthenticator::continueGetAssertionAfterResponseReceived const):

• UIProcess/WebAuthentication/fido/CtapHidAuthenticator.h: Copied from Source/WebKit/UIProcess/WebAuthentication/Cocoa/LocalService.h.
• UIProcess/WebAuthentication/fido/CtapHidDriver.cpp: Added.

(WebKit::CtapHidDriver::Worker::Worker):
(WebKit::CtapHidDriver::Worker::~Worker):
(WebKit::CtapHidDriver::Worker::transact):
(WebKit::CtapHidDriver::Worker::write):
(WebKit::CtapHidDriver::Worker::read):
(WebKit::CtapHidDriver::Worker::returnMessage):
(WebKit::CtapHidDriver::CtapHidDriver):
(WebKit::CtapHidDriver::transact):
(WebKit::CtapHidDriver::continueAfterChannelAllocated):
(WebKit::CtapHidDriver::continueAfterResponseReceived):
(WebKit::CtapHidDriver::returnResponse):

• UIProcess/WebAuthentication/fido/CtapHidDriver.h: Added.
• UIProcess/mac/WebDataListSuggestionsDropdownMac.mm:
• WebProcess/WebPage/CoordinatedGraphics/ThreadedCoordinatedLayerTreeHost.h:
• WebKit.xcodeproj/project.pbxproj:

Tools:

This patch adds support for the mock testing and entitlements to allow minibrowser to talk
to hid devices.

• MiniBrowser/MiniBrowser.entitlements:
• WebKitTestRunner/InjectedBundle/TestRunner.cpp:

(WTR::TestRunner::setWebAuthenticationMockConfiguration):

LayoutTests:

• http/wpt/webauthn/ctap-hid-failure.https-expected.txt: Added.
• http/wpt/webauthn/ctap-hid-failure.https.html: Added.
• http/wpt/webauthn/ctap-hid-success.https-expected.txt: Added.
• http/wpt/webauthn/ctap-hid-success.https.html: Added.
• http/wpt/webauthn/public-key-credential-create-failure-hid-silent.https-expected.txt: Added.
• http/wpt/webauthn/public-key-credential-create-failure-hid-silent.https.html: Added.
• http/wpt/webauthn/public-key-credential-create-failure-hid.https-expected.txt: Added.
• http/wpt/webauthn/public-key-credential-create-failure-hid.https.html: Added.
• http/wpt/webauthn/public-key-credential-create-success-hid.https-expected.txt: Added.
• http/wpt/webauthn/public-key-credential-create-success-hid.https.html: Added.
• http/wpt/webauthn/public-key-credential-get-failure-hid-silent.https-expected.txt: Added.
• http/wpt/webauthn/public-key-credential-get-failure-hid-silent.https.html: Added.
• http/wpt/webauthn/public-key-credential-get-failure-hid.https-expected.txt: Added.
• http/wpt/webauthn/public-key-credential-get-failure-hid.https.html: Added.
• http/wpt/webauthn/public-key-credential-get-success-hid.https-expected.txt: Added.
• http/wpt/webauthn/public-key-credential-get-success-hid.https.html: Added.
• http/wpt/webauthn/resources/util.js:
• platform/ios-wk2/TestExpectations:

10:14 PM Changeset in webkit [238165] by rniwa@webkit.org

• 3 edits in trunk/Tools

WebKit.GeolocationTransitionToLowAccuracy API crashes when enabling PSON
​https://bugs.webkit.org/show_bug.cgi?id=191616

Reviewed by Chris Dumez.

The crash was caused by WKView in autorelease pool invoking stopUpdatingCallback after
GeolocationTransitionToLowAccuracyStateTracker in the stack had been destroyed,
resulting in the use-after-free.

Made the tests more robust by clearing geolocation provider before exiting each test
since we can't really prevent WKView from entering an autorelease pool.

Also made WebKit.GeolocationTransitionToLowAccuracy wait for the success callback
instead of simply the end of the navigation so that the test would continue to work
even if a web content process was created for the second web view (lowAccuracyWebView)

• TestWebKitAPI/Tests/WebKit/Geolocation.cpp:

(TestWebKitAPI::setupGeolocationProvider): Moved "*" to match the WebKit coding style guideline.
(TestWebKitAPI::clearGeolocationProvider): Added.
(TestWebKitAPI::runJavaScriptAlert): Added.
(TestWebKitAPI::didFinishNavigation): Deleted.

• TestWebKitAPI/Tests/WebKit/geolocationWatchPosition.html:

9:31 PM Changeset in webkit [238164] by Dewei Zhu

• 4 edits in trunk/Websites/perf.webkit.org

commit time returned by '/api/measurement-set' should match the one returned by '/api/commits'.
​https://bugs.webkit.org/show_bug.cgi?id=191457

Reviewed by Dean Jackson and Ryosuke Niwa.

Commit time returned by '/api/measurement-set' sometimes is calculated by 'epoch from ..'.
This function will return a floating number with 5 or 6 decimal digits due to double precision limitations.
However, some commits may be reported with 6 decimal decimal.
So the commit time for those commits will be rounded to 5 decimal digits.
In order to avoid front end assertion failure in CommitLog, Database::to_js_time need to
match this behavior.

• public/include/db.php: Change the behavior to match that of postgres.

Added logic to avoid losing precision in php.

• server-tests/api-measurement-set-tests.js: Added unit tests for this bug.

(queryPlatformAndMetric): Fix a bug that arguments are not used at all.

8:58 PM Changeset in webkit [238163] by sbarati@apple.com

• 2 edits in trunk/Source/JavaScriptCore

ProxyObject should check for VMInquiry and return early before throwing a stack overflow exception
​https://bugs.webkit.org/show_bug.cgi?id=191601

Reviewed by Mark Lam.

This doesn't fix any bugs today, but it may reduce future bugs. It was
always weird that ProxyObject::getOwnPropertySlot with VMInquiry might
throw a stack overflow error instead of just returning false like it
normally does when VMInquiry is passed in.

• runtime/ProxyObject.cpp:

(JSC::ProxyObject::getOwnPropertySlotCommon):

8:57 PM Changeset in webkit [238162] by sbarati@apple.com

• 13 edits
  1 add in trunk

TypeProfileLog::processLogEntries should stash away any pending exceptions and re-apply them to the VM
​https://bugs.webkit.org/show_bug.cgi?id=191600

Reviewed by Mark Lam.

JSTests:

• stress/type-profiler-log-should-defer-pending-exceptions.js: Added.

(foo):
(test):
(bar):

Source/JavaScriptCore:

processLogEntries will call into calculatedClassName, which will clear
any exceptions it encounters (it assumes that they're stack overflow exceptions).
However, this code may be called when an exception is already pending on the
VM (e.g, when we throw an exception in the DFG, we compile an OSR exit
offramp, which may compile a baseline codeblock, which will process
the type profiler log). To get around this, processLogEntires should stash
away and re-apply any pending exceptions.

• dfg/DFGDriver.cpp:

(JSC::DFG::compileImpl):

• dfg/DFGOperations.cpp:
• inspector/agents/InspectorRuntimeAgent.cpp:

(Inspector::InspectorRuntimeAgent::getRuntimeTypesForVariablesAtOffsets):

• jit/JIT.cpp:

(JSC::JIT::doMainThreadPreparationBeforeCompile):

• jit/JITOperations.cpp:
• runtime/CommonSlowPaths.cpp:

(JSC::SLOW_PATH_DECL):

• runtime/TypeProfilerLog.cpp:

(JSC::TypeProfilerLog::processLogEntries):

• runtime/TypeProfilerLog.h:
• runtime/VM.cpp:

(JSC::VM::dumpTypeProfilerData):

• runtime/VM.h:

(JSC::VM::DeferExceptionScope::DeferExceptionScope):

• tools/JSDollarVM.cpp:

(JSC::functionFindTypeForExpression):
(JSC::functionReturnTypeFor):

8:52 PM Changeset in webkit [238161] by Ross Kirsling

• 2 edits in trunk/Source/WebKit

Unreviewed correction to previous build fix to avoid any internal/downstream repercussions.

• WebProcess/Network/webrtc/LibWebRTCProvider.h:

8:22 PM Changeset in webkit [238160] by Ross Kirsling

• 2 edits in trunk/Source/WebKit

Unreviewed GTK/WPE build fix for r238159.

• WebProcess/Network/webrtc/LibWebRTCProvider.h:

7:57 PM Changeset in webkit [238159] by Ross Kirsling

• 6 edits
  2 moves
  1 delete in trunk/Source/WebCore

[WebRTC] Provide default implementation of LibWebRTCProvider
​https://bugs.webkit.org/show_bug.cgi?id=191611

Reviewed by Michael Catanzaro.

Refactor LibWebRTCProvider such that platform-specific implementations need not worry about specifying defaults.

• PlatformWin.cmake:
• platform/GStreamer.cmake:
• platform/SourcesGLib.txt:
• platform/mediastream/libwebrtc/LibWebRTCProvider.cpp:
• platform/mediastream/libwebrtc/LibWebRTCProviderCocoa.cpp:
• platform/mediastream/libwebrtc/LibWebRTCProviderGStreamer.cpp: Renamed from Source/WebCore/platform/mediastream/libwebrtc/LibWebRTCProviderGlib.cpp.
• platform/mediastream/libwebrtc/LibWebRTCProviderGStreamer.h: Renamed from Source/WebCore/platform/mediastream/libwebrtc/LibWebRTCProviderGlib.h.
• platform/mediastream/libwebrtc/LibWebRTCProviderWin.cpp: Removed.

7:48 PM Changeset in webkit [238158] by Chris Dumez

• 2 edits in trunk/Tools

WKProcessPool.InitialWarmedProcessUsed API is failing with PSON enabled
​https://bugs.webkit.org/show_bug.cgi?id=191618

Reviewed by Ryosuke Niwa.

Update API test to explicitly disable automatic process prewarning since it is testing
manual prewarming.

• TestWebKitAPI/Tests/WebKitCocoa/ProcessPreWarming.mm:

(TEST):

7:42 PM Changeset in webkit [238157] by Chris Dumez

• 2 edits in trunk/Tools

Several Service Worker API tests are failing when enabling PSON
​https://bugs.webkit.org/show_bug.cgi?id=191619

Reviewed by Youenn Fablet.

Update tests to use _webProcessCountIgnoringPrewarmed instead of _webProcessCount so that they are
not impacted by process prewarming.

• TestWebKitAPI/Tests/WebKitCocoa/ServiceWorkerBasic.mm:

5:47 PM Changeset in webkit [238156] by Ryan Haddad

• 9 edits
  1 delete in trunk

Unreviewed, rolling out r238132.

The test added with this change is timing out on Debug JSC
bots.

Reverted changeset:

"[BigInt] JSBigInt::createWithLength should throw when length
is greater than JSBigInt::maxLength"
​https://bugs.webkit.org/show_bug.cgi?id=190836
​https://trac.webkit.org/changeset/238132

5:15 PM Changeset in webkit [238155] by timothy@apple.com

• 17 edits
  2 adds in trunk

Use a light scrollbar for transparent web views in dark mode.
​https://bugs.webkit.org/show_bug.cgi?id=191559
rdar://problem/46000489

Reviewed by Dean Jackson.

Source/WebCore:

Test: css-dark-mode/supported-color-schemes-scrollbar.html

• css/CSSProperties.json: Marked supported-color-schemes as a custom Value.
• css/StyleBuilderCustom.h:

(WebCore::StyleBuilderCustom::applyValueSupportedColorSchemes):

• editing/cocoa/WebContentReaderCocoa.mm: Use FrameView's useDarkAppearance().

(WebCore::createFragment):

• inspector/InspectorOverlay.cpp:

(WebCore::InspectorOverlay::paint): Use FrameView's useDarkAppearance().

• page/FrameView.cpp:

(WebCore::FrameView::recalculateScrollbarOverlayStyle): Use a light scrollbar for
transparent web views in dark mode.
(WebCore::FrameView::rendererForSupportedColorSchemes const): Added.
Return the body for document element renderer.
(WebCore::FrameView::useDarkAppearance const): Use rendererForSupportedColorSchemes.
(WebCore::FrameView::styleColorOptions const): Added. Ditto.

• page/FrameView.h:
• rendering/style/RenderStyle.cpp:

(WebCore::rareInheritedDataChangeRequiresRepaint): Drive-by fix. Added supportedColorSchemes.

• rendering/style/RenderStyle.h:

(WebCore::RenderStyle::setHasExplicitlySetSupportedColorSchemes): Added.
(WebCore::RenderStyle::hasExplicitlySetSupportedColorSchemes const): Added.
(WebCore::RenderStyle::NonInheritedFlags::operator== const): Added supportedColorSchemes.

• svg/graphics/SVGImage.cpp:

(WebCore::SVGImage::draw): Use FrameView's useDarkAppearance().

• testing/Internals.cpp:

(WebCore::Internals::setViewIsTransparent): Added.
(WebCore::Internals::scrollbarOverlayStyle const): Added.

• testing/Internals.h:
• testing/Internals.idl: Added setViewIsTransparent and scrollbarOverlayStyle.

Source/WebKit:

• WebProcess/InjectedBundle/DOM/InjectedBundleRangeHandle.cpp:

(WebKit::InjectedBundleRangeHandle::renderedImage): Use FrameView's useDarkAppaearance().

• WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::drawRect): Ditto.

LayoutTests:

• css-dark-mode/supported-color-schemes-scrollbar-expected.txt: Added.
• css-dark-mode/supported-color-schemes-scrollbar.html: Added.

3:42 PM Changeset in webkit [238154] by Kocsen Chung

• 1 copy in tags/Safari-606.3.4.1.4

Tag Safari-606.3.4.1.4.

3:30 PM Changeset in webkit [238153] by Kocsen Chung

• 4 edits in branches/safari-606.3.4.1-branch/Source

Revert r237971. rdar://problem/46043804

3:28 PM Changeset in webkit [238152] by Ross Kirsling

• 3 edits in trunk/Source/WebCore

[AppleWin] Unreviewed build fix after r238108.

• platform/graphics/ca/win/PlatformCALayerWin.cpp:

(printLayer):
(PlatformCALayerWin::embeddedViewID const):

• platform/graphics/ca/win/PlatformCALayerWin.h:

3:23 PM Changeset in webkit [238151] by Kocsen Chung

• 7 edits in branches/safari-606.3.4.1-branch/Source

Versioning.

3:22 PM Changeset in webkit [238150] by youenn@apple.com

• 10 edits in trunk

RTCPeerConnection.getTransceivers is not always exposing all transceivers
​https://bugs.webkit.org/show_bug.cgi?id=191589

Reviewed by Eric Carlson.

LayoutTests/imported/w3c:

• web-platform-tests/webrtc/RTCRtpTransceiver.https-expected.txt:

Source/WebCore:

Implement the collect transceiver algorithm using libwebrtc backend.
Call this algorithm everytime transceivers are retrieved from JS.
For Plan B, make this a no-op as this is not supported.
Introduce senders/receivers/transceivers getters where we just look at already created transceivers.

Covered by existing and rebased tests.

• Modules/mediastream/PeerConnectionBackend.h:

(WebCore::PeerConnectionBackend::collectTransceivers):

• Modules/mediastream/RTCPeerConnection.cpp:

(WebCore::RTCPeerConnection::getSenders const):
(WebCore::RTCPeerConnection::getReceivers const):
(WebCore::RTCPeerConnection::getTransceivers const):

• Modules/mediastream/RTCPeerConnection.h:
• Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:

(WebCore::LibWebRTCMediaEndpoint::collectTransceivers):

• Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.h:
• Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.cpp:

(WebCore::LibWebRTCPeerConnectionBackend::addTrack):
(WebCore::LibWebRTCPeerConnectionBackend::existingTransceiver):
(WebCore::LibWebRTCPeerConnectionBackend::collectTransceivers):
(WebCore::LibWebRTCPeerConnectionBackend::applyRotationForOutgoingVideoSources):
(WebCore::LibWebRTCPeerConnectionBackend::shouldOfferAllowToReceive const):

• Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.h:

2:46 PM Changeset in webkit [238149] by Alan Coon

• 4 edits in branches/safari-606-branch/Source

Apply patch. rdar://problem/45996792

2:46 PM Changeset in webkit [238148] by Alan Coon

• 8 edits
  2 adds in branches/safari-606-branch

Cherry-pick r237837. rdar://problem/45996796

[MediaStream] An audio track should be muted when capture is interrupted by the OS.
Source/WebCore:

​https://bugs.webkit.org/show_bug.cgi?id= 191283

<rdar://problem/45773103>

Patch by Eric Carlson <​eric.carlson@apple.com> on 2018-11-05
Reviewed by Jon Lee.

Test: fast/mediastream/media-stream-track-interrupted.html

• platform/mediastream/RealtimeMediaSource.cpp: (WebCore::RealtimeMediaSource::setInterruptedForTesting):
• platform/mediastream/RealtimeMediaSource.h:
• platform/mediastream/mac/CoreAudioCaptureSource.cpp: (WebCore::CoreAudioCaptureSource::beginInterruption): (WebCore::CoreAudioCaptureSource::endInterruption):
• testing/Internals.cpp: (WebCore::Internals::setMediaStreamSourceInterrupted):
• testing/Internals.h:
• testing/Internals.idl:

LayoutTests:

​https://bugs.webkit.org/show_bug.cgi?id=191283

<rdar://problem/45773103>

Patch by Eric Carlson <​eric.carlson@apple.com> on 2018-11-05
Reviewed by Jon Lee.

• fast/mediastream/media-stream-track-interrupted-expected.txt: Added.
• fast/mediastream/media-stream-track-interrupted.html: Added.

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@237837 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2:35 PM Changeset in webkit [238147] by Ross Kirsling

• 2 edits in trunk/LayoutTests

[WinCairo] Unreviewed layout test gardening.

• platform/wincairo/TestExpectations:

2:30 PM Changeset in webkit [238146] by Wenson Hsieh

• 19 edits
  10 adds in trunk

[iOS] Do not show selection UI for editable elements with opacity near zero
​https://bugs.webkit.org/show_bug.cgi?id=191442
<rdar://problem/45958625>

Reviewed by Simon Fraser.

Source/WebCore:

Tests: editing/selection/ios/do-not-zoom-to-focused-hidden-contenteditable.html

editing/selection/ios/hide-selection-after-hiding-contenteditable.html
editing/selection/ios/hide-selection-in-contenteditable-nested-transparency.html
editing/selection/ios/hide-selection-in-hidden-contenteditable-frame.html
editing/selection/ios/hide-selection-in-hidden-contenteditable.html

• rendering/RenderObject.cpp:

(WebCore::RenderObject::isTransparentRespectingParentFrames const):

Add a helper function to determine whether a RenderObject is contained within a transparent layer, taking parent
frames into account. A layer is considered transparent if its opacity is less than a small threshold (i.e. 0.01).
Opacity on ancestor elements is applied multiplicatively.

• rendering/RenderObject.h:

Source/WebKit:

Add support for suppressing native selection UI (for instance, selection highlight views, selection handles, and
selection-related gestures) when the selection is inside a transparent editable element. This helps maintain
compatibility with text editors that work by capturing key events and input events hidden contenteditable
elements, and reflect these changes in different document or different part of the document.

Since selection UI is rendered in the UI process on iOS using element geometry propagated from the web process,
selection rendering is entirely decoupled from the process of painting in the web process. This means that if
the editable root has an opacity of 0, we would correctly hide the caret and selection on macOS, but draw over
the transparent element on iOS. When these hidden editable elements are focused, this often results in unwanted
behaviors, such as double caret painting, native and custom selection UI from the page being drawn on top of one
another, and the ability to change selection via tap and loupe gestures within hidden text.

To fix this, we compute whether the focused element is transparent when an element is focused, or when the
selection changes, and send this information over to the UI process via AssistedNodeInformation and
EditorState. In the UI process, we then respect this information by suppressing the selection assistant if the
focused element is transparent; this disables showing and laying out selection views, as well as gestures
associated with selection overlays. However, this still allows for contextual autocorrection and spell checking.

• Shared/AssistedNodeInformation.cpp:

(WebKit::AssistedNodeInformation::encode const):
(WebKit::AssistedNodeInformation::decode):

• Shared/AssistedNodeInformation.h:
• Shared/EditorState.cpp:

(WebKit::EditorState::PostLayoutData::encode const):
(WebKit::EditorState::PostLayoutData::decode):

• Shared/EditorState.h:

Add elementIsTransparent flags, and also add boilerplate IPC code.

• UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView _displayFormNodeInputView]):

Prevent zooming to the focused element if the focused element is hidden.

(-[WKContentView hasSelectablePositionAtPoint:]):
(-[WKContentView pointIsNearMarkedText:]):
(-[WKContentView textInteractionGesture:shouldBeginAtPoint:]):

Don't allow these text interaction gestures to begin while suppressing the selection assistant.

(-[WKContentView _startAssistingNode:userIsInteracting:blurPreviousNode:changingActivityState:userObject:]):

When an element is focused, begin suppressing the selection assistant if the element is fully transparent.

(-[WKContentView _stopAssistingNode]):

When the focused element is blurred, reset state by ending selection assistant suppression (additionally
reactivating the selection assistant if needed). This ensures that selection in non-editable text isn't broken
after focusing a hidden editable element.

(-[WKContentView _updateChangedSelection:]):

If needed, suppress or un-suppress the selection assistant when the selection changes. On certain rich text
editors, a combination of custom selection UI and native selection UI is used. For instance, on Microsoft Office
365, caret selections are rendered using the native caret view, but as soon as the selection becomes ranged, the
editable root becomes fully transparent, and Office's selection UI takes over.

(-[WKContentView _shouldSuppressSelectionCommands]):

Override this UIKit SPI hook to suppress selection commands (e.g. the callout bar) when suppressing the
selection assistant.

• WebProcess/WebPage/ios/WebPageIOS.mm:

(WebKit::WebPage::platformEditorState const):
(WebKit::WebPage::getAssistedNodeInformation):

Compute and set elementIsTransparent using the assisted node.

Tools:

Add a couple of new testing helpers to UIScriptController.

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

(WTR::UIScriptController::textSelectionRangeRects const):
(WTR::UIScriptController::selectionCaretViewRect const):
(WTR::UIScriptController::selectionRangeViewRects const):

• TestRunnerShared/UIScriptContext/UIScriptController.h:
• WebKitTestRunner/ios/UIScriptControllerIOS.mm:

(WTR::UIScriptController::textSelectionRangeRects const):

Rename selectionRangeViewRects to textSelectionRangeRects. This allows us to draw a distinction between
textSelectionRangeRects/textSelectionCaretRect, which retrieve information about selection rects known
to the text interaction assistant, and selectionCaretViewRect/selectionRangeViewRects, which retrieve the
actual frames of the selection views used to draw overlaid selection UI. This difference is important in the
new layout tests added in this patch, which only suppress caret rendering (i.e. selection views remain hidden).

Also, drive-by fix a leaked NSMutableArray.

(WTR::UIScriptController::selectionStartGrabberViewRect const):
(WTR::UIScriptController::selectionEndGrabberViewRect const):
(WTR::UIScriptController::selectionCaretViewRect const):
(WTR::UIScriptController::selectionRangeViewRects const):

Testing helpers to grab the frames of caret and selection views, in WKContentView's coordinate space. These
rects are also clamped to WKContentView bounds.

LayoutTests:

Add 5 new layout tests. See below for more details.

• editing/selection/character-granularity-rect.html:

Adjust for a renamed UIScriptController function.

• editing/selection/ios/do-not-zoom-to-focused-hidden-contenteditable-expected.txt: Added.
• editing/selection/ios/do-not-zoom-to-focused-hidden-contenteditable.html: Added.

Add a test to verify that we don't zoom to fit the focused element, if the focused element is completely
transparent.

• editing/selection/ios/hide-selection-after-hiding-contenteditable-expected.txt: Added.
• editing/selection/ios/hide-selection-after-hiding-contenteditable.html: Added.

Add a test to verify that selection UI is hidden after making an editable root transparent, and shown again when
the editable root becomes opaque.

• editing/selection/ios/hide-selection-in-contenteditable-nested-transparency-expected.txt: Added.
• editing/selection/ios/hide-selection-in-contenteditable-nested-transparency.html: Added.

Add a test to verify that transparency applied on an editable root via nested transparent containers causes
selection UI to be suppressed.

• editing/selection/ios/hide-selection-in-hidden-contenteditable-expected.txt: Added.
• editing/selection/ios/hide-selection-in-hidden-contenteditable-frame-expected.txt: Added.
• editing/selection/ios/hide-selection-in-hidden-contenteditable-frame.html: Added.

Add a test to verify that selection UI is suppressed when an editable element inside a subframe is focused. This
test checks that the caret, selection rects and selection handle views are not shown, and additionally verifies
that the selection in a hidden contenteditable area cannot be changed via tap gesture.

• editing/selection/ios/hide-selection-in-hidden-contenteditable.html: Added.

Same test as above, but in a regular editable element in the main document instead of a subframe.

• resources/ui-helper.js:

(window.UIHelper.getUISelectionRects.return.new.Promise.):
(window.UIHelper.getUISelectionRects.return.new.Promise):
(window.UIHelper.getUISelectionRects):
(window.UIHelper.getUICaretViewRect.return.new.Promise.):
(window.UIHelper.getUICaretViewRect.return.new.Promise):
(window.UIHelper.getUICaretViewRect):

Add new UIHelper wrapper methods. See Tools/ChangeLog for more detail.

2:21 PM Changeset in webkit [238145] by eric.carlson@apple.com

• 5 edits in trunk/Source/WebCore

[MediaStream] Observer AVCaptureDevice "suspended" property
​https://bugs.webkit.org/show_bug.cgi?id=191587
<rdar://problem/46030598>

Reviewed by Youenn Fablet.

No new tests, AVCapture can only be tested manually.

• platform/mediastream/mac/AVCaptureDeviceManager.h:
• platform/mediastream/mac/AVCaptureDeviceManager.mm:

(WebCore::AVCaptureDeviceManager::captureDevicesInternal): Don't notify of devices "changes"
the first time the device list is scanned.
(WebCore::deviceIsAvailable): Don't check for "isInUseByAnotherApplication", it doesn't
change device availability.
(WebCore::AVCaptureDeviceManager::beginObservingDevices): New, observe "suspended" on all
devices and add them to the cached list.
(WebCore::AVCaptureDeviceManager::stopObservingDevices): New, opposite of above.
(WebCore::AVCaptureDeviceManager::refreshCaptureDevices): Watch for changes in the list of
devices.
(WebCore::AVCaptureDeviceManager::~AVCaptureDeviceManager): Stop observing all cached devices.
(WebCore::AVCaptureDeviceManager::registerForDeviceNotifications):
(-[WebCoreAVCaptureDeviceManagerObserver disconnect]):
(-[WebCoreAVCaptureDeviceManagerObserver deviceConnectedDidChange:]):
(-[WebCoreAVCaptureDeviceManagerObserver observeValueForKeyPath:ofObject:change:context:]):
(WebCore::AVCaptureDeviceManager::refreshAVCaptureDevicesOfType): Deleted.
(WebCore::AVCaptureDeviceManager::deviceConnected): Deleted.
(WebCore::AVCaptureDeviceManager::deviceDisconnected): Deleted.
(-[WebCoreAVCaptureDeviceManagerObserver deviceDisconnected:]): Deleted.
(-[WebCoreAVCaptureDeviceManagerObserver deviceConnected:]): Deleted.

• platform/mediastream/mac/AVVideoCaptureSource.h:
• platform/mediastream/mac/AVVideoCaptureSource.mm:

(WebCore::AVVideoCaptureSource::~AVVideoCaptureSource): Stop observing "running" (not "rate")
and "suspended".
(WebCore::AVVideoCaptureSource::setupSession): Observe "running" (not "rate"), and "suspended".
(WebCore::AVVideoCaptureSource::captureDeviceSuspendedDidChange):
(-[WebCoreAVVideoCaptureSourceObserver observeValueForKeyPath:ofObject:change:context:]):

2:06 PM Changeset in webkit [238144] by Devin Rousso

• 2 edits in trunk/Source/WebCore

Web Inspector: REGRESSION(r238122): fetching the CertificateInfo triggers an ASSERT in workers
​https://bugs.webkit.org/show_bug.cgi?id=191597

Reviewed by Joseph Pecoraro.

When WebInspector is open, the CertificateInfo for every ResourceResponse is now fetched,
meaning that we may try to fetch in situations previously unexpected.

• platform/network/cocoa/ResourceResponseCocoa.mm:

(WebCore::ResourceResponse::platformCertificateInfo const):

1:33 PM Changeset in webkit [238143] by mark.lam@apple.com

• 8 edits
  1 add in trunk

Add OOM detection to StringPrototype's substituteBackreferences().
​https://bugs.webkit.org/show_bug.cgi?id=191563
<rdar://problem/45720428>

Reviewed by Saam Barati.

JSTests:

• stress/regress-191563.js: Added.

Source/JavaScriptCore:

• dfg/DFGStrengthReductionPhase.cpp:

(JSC::DFG::StrengthReductionPhase::handleNode):

• runtime/StringPrototype.cpp:

(JSC::substituteBackreferencesSlow):
(JSC::substituteBackreferencesInline):
(JSC::substituteBackreferences):
(JSC::replaceUsingRegExpSearch):
(JSC::replaceUsingStringSearch):

• runtime/StringPrototype.h:

Source/WTF:

Enhanced StringBuilder::toString() to skip the shrinkToFit(), reifyString(), and
the hasOverflowed() check if m_string is not null. When m_string is not null,
the StringBuilder either only has a single String in m_string (with m_buffer being
null), or reifyString() has already been called (resulting in a non-null m_string
with a possibly non-null m_buffer).

We can skip the overflow check because:

1. if the StringBuilder only has a single String, then there cannot be an overflow.
2. if reifyString() has already been called, then the hasOverflowed() checked has already been done because every code path that calls reifyString() first does the hasOverflowed() check.

We can skip shrinkToFit() because it only applies to m_buffer.

1. if the StringBuilder only has a single String, then there's no m_buffer to shrink.
2. if reifyString() has already been called, then we either came down
   1. the toString() path with a null m_string previously, where we would have already called shrinkToFit() before reifyString(), or
   2. the toStringPreserveCapacity() path where we don't want to shrinkToFit().

We can skip reifyString() because:

1. if the StringBuilder only has a single String, then the string is already reified.
2. if reifyString() has been already called, then the string is already reified.

Note that if m_string is the null string and m_buffer is null, reifyString() will
replace it with the empty string. For this reason, we cannot solely check for
!m_buffer because we need to reify the null string into the empty string.

Note also that if m_string is null and m_buffer is non-null, reifyString() will
create a String and set m_string to it. However, m_buffer remains non-null.
For this reason, we cannot assert !m_buffer alone when m_string is non-null.
We add a m_isReified flag (only when assertions are enabled) to track the reified
case where both m_buffer and m_string are non-null.

• wtf/text/StringBuilder.cpp:

(WTF::StringBuilder::reifyString const):

• wtf/text/StringBuilder.h:

(WTF::StringBuilder::toString):

1:04 PM Changeset in webkit [238142] by Ryan Haddad

• 3 edits in trunk/Source/WebKit

Unreviewed, rolling out r238137.

Introduced API test failures on macOS.

Reverted changeset:

"Enable process swap on cross-site navigation by default on
macOS"
​https://bugs.webkit.org/show_bug.cgi?id=191572
​https://trac.webkit.org/changeset/238137

12:48 PM Changeset in webkit [238141] by mark.lam@apple.com

• 3 edits
  1 add in trunk

LLIntSlowPath's llint_loop_osr and llint_replace should set the topCallFrame.
​https://bugs.webkit.org/show_bug.cgi?id=191579
<rdar://problem/45942472>

Reviewed by Saam Barati.

JSTests:

• stress/regress-191579.js: Added.

Source/JavaScriptCore:

Both of these functions do a lot of work. It would be good for the topCallFrame
to be correct should we need to throw an exception.

For example, we've observed the following crash trace:

• frame #0: WTFCrash() at Assertions.cpp:253 frame #1: ... frame #2: JSC::StructureIDTable::get(this=0x00006040000162f0, structureID=1874583248) at StructureIDTable.h:129 frame #3: JSC::VM::getStructure(this=0x0000604000016210, id=4022066896) at VM.h:705 frame #4: JSC::JSCell::structure(this=0x00007ffeefbbde30, vm=0x0000604000016210) const at JSCellInlines.h:125 frame #5: JSC::JSCell::classInfo(this=0x00007ffeefbbde30, vm=0x0000604000016210) const at JSCellInlines.h:335 frame #6: JSC::JSCell::inherits(this=0x00007ffeefbbde30, vm=0x0000604000016210, info=0x0000000105eaf020) const at JSCellInlines.h:302 frame #7: JSC::JSObject* JSC::jsCast<JSC::JSObject*, JSC::JSCell>(from=0x00007ffeefbbde30) at JSCast.h:36 frame #8: JSC::asObject(cell=0x00007ffeefbbde30) at JSObject.h:1299 frame #9: JSC::asObject(value=JSValue @ 0x00007ffeefbba380) at JSObject.h:1304 frame #10: JSC::Register::object(this=0x00007ffeefbbdd58) const at JSObject.h:1514 frame #11: JSC::ExecState::jsCallee(this=0x00007ffeefbbdd40) const at CallFrame.h:107 frame #12: JSC::ExecState::isStackOverflowFrame(this=0x00007ffeefbbdd40) const at CallFrameInlines.h:36 frame #13: JSC::StackVisitor::StackVisitor(this=0x00007ffeefbba860, startFrame=0x00007ffeefbbdd40, vm=0x0000631000000800) at StackVisitor.cpp:52 frame #14: JSC::StackVisitor::StackVisitor(this=0x00007ffeefbba860, startFrame=0x00007ffeefbbdd40, vm=0x0000631000000800) at StackVisitor.cpp:41 frame #15: void JSC::StackVisitor::visit<(JSC::StackVisitor::EmptyEntryFrameAction)0, JSC::Interpreter::getStackTrace(JSC::JSCell*, WTF::Vector<JSC::StackFrame, 0ul, WTF::CrashOnOverflow, 16ul>&, unsigned long, unsigned long)::$_3>(startFrame=0x00007ffeefbbdd40, vm=0x0000631000000800, functor=0x00007ffeefbbaa60)::$_3 const&) at StackVisitor.h:147 frame #16: JSC::Interpreter::getStackTrace(this=0x0000602000005db0, owner=0x000062d00020cbe0, results=0x00006020000249d0, framesToSkip=0, maxStackSize=1) at Interpreter.cpp:437 frame #17: JSC::getStackTrace(exec=0x000062d00002c048, vm=0x0000631000000800, obj=0x000062d00020cbe0, useCurrentFrame=true) at Error.cpp:170 frame #18: JSC::ErrorInstance::finishCreation(this=0x000062d00020cbe0, exec=0x000062d00002c048, vm=0x0000631000000800, message=0x00007ffeefbbb800, useCurrentFrame=true) at ErrorInstance.cpp:119 frame #19: JSC::ErrorInstance::create(exec=0x000062d00002c048, vm=0x0000631000000800, structure=0x000062d0000f5730, message=0x00007ffeefbbb800, appender=0x0000000000000000, type=TypeNothing, useCurrentFrame=true)(WTF::String const&, WTF::String const&, JSC::RuntimeType, JSC::ErrorInstance::SourceTextWhereErrorOccurred), JSC::RuntimeType, bool) at ErrorInstance.h:49 frame #20: JSC::createRangeError(exec=0x000062d00002c048, globalObject=0x000062d00002c000, message=0x00007ffeefbbb800, appender=0x0000000000000000)(WTF::String const&, WTF::String const&, JSC::RuntimeType, JSC::ErrorInstance::SourceTextWhereErrorOccurred)) at Error.cpp:68 frame #21: JSC::createRangeError(exec=0x000062d00002c048, globalObject=0x000062d00002c000, message=0x00007ffeefbbb800) at Error.cpp:316 frame #22: JSC::createStackOverflowError(exec=0x000062d00002c048, globalObject=0x000062d00002c000) at ExceptionHelpers.cpp:77 frame #23: JSC::createStackOverflowError(exec=0x000062d00002c048) at ExceptionHelpers.cpp:72 frame #24: JSC::throwStackOverflowError(exec=0x000062d00002c048, scope=0x00007ffeefbbbaa0) at ExceptionHelpers.cpp:335 frame #25: JSC::ProxyObject::getOwnPropertySlotCommon(this=0x000062d000200e40, exec=0x000062d00002c048, propertyName=PropertyName @ 0x00007ffeefbbba80, slot=0x00007ffeefbbc720) at ProxyObject.cpp:372 frame #26: JSC::ProxyObject::getOwnPropertySlot(object=0x000062d000200e40, exec=0x000062d00002c048, propertyName=PropertyName @ 0x00007ffeefbbbd40, slot=0x00007ffeefbbc720) at ProxyObject.cpp:395 frame #27: JSC::JSObject::getNonIndexPropertySlot(this=0x000062d000200e40, exec=0x000062d00002c048, propertyName=PropertyName @ 0x00007ffeefbbbea0, slot=0x00007ffeefbbc720) at JSObjectInlines.h:150 frame #28: bool JSC::JSObject::getPropertySlot<false>(this=0x000062d000200e40, exec=0x000062d00002c048, propertyName=PropertyName @ 0x00007ffeefbbc320, slot=0x00007ffeefbbc720) at JSObject.h:1424 frame #29: JSC::JSObject::calculatedClassName(object=0x000062d000200e40) at JSObject.cpp:535 frame #30: JSC::Structure::toStructureShape(this=0x000062d000007410, value=JSValue @ 0x00007ffeefbbcae0, sawPolyProtoStructure=0x00007ffeefbbcf60) at Structure.cpp:1142 frame #31: JSC::TypeProfilerLog::processLogEntries(this=0x000060400000a950, reason=0x00007ffeefbbd5c0) at TypeProfilerLog.cpp:89 frame #32: JSC::JIT::doMainThreadPreparationBeforeCompile(this=0x0000619000034da0) at JIT.cpp:951 frame #33: JSC::JITWorklist::Plan::Plan(this=0x0000619000034d80, codeBlock=0x000062d0001d88c0, loopOSREntryBytecodeOffset=0) at JITWorklist.cpp:43 frame #34: JSC::JITWorklist::Plan::Plan(this=0x0000619000034d80, codeBlock=0x000062d0001d88c0, loopOSREntryBytecodeOffset=0) at JITWorklist.cpp:42 frame #35: JSC::JITWorklist::compileLater(this=0x0000616000001b80, codeBlock=0x000062d0001d88c0, loopOSREntryBytecodeOffset=0) at JITWorklist.cpp:256 frame #36: JSC::LLInt::jitCompileAndSetHeuristics(codeBlock=0x000062d0001d88c0, exec=0x00007ffeefbbde30, loopOSREntryBytecodeOffset=0) at LLIntSlowPaths.cpp:391 frame #37: llint_replace(exec=0x00007ffeefbbde30, pc=0x00006040000161ba) at LLIntSlowPaths.cpp:516 frame #38: llint_entry at LowLevelInterpreter64.asm:98 frame #39: vmEntryToJavaScript at LowLevelInterpreter64.asm:296 ...

This crash occurred because StackVisitor was seeing an invalid topCallFrame while
trying to capture the Error stack while throwing a StackOverflowError below
llint_replace. While in this specific example, it is questionable whether we
should be executing JS code below TypeProfilerLog::processLogEntries(), it is
correct to have set the topCallFrame in llint_replace. We do this by calling
LLINT_BEGIN_NO_SET_PC() at the top of llint_replace.

We also do the same for llint_osr.

Note: both of these LLInt slow path functions are called with a fully initialized
CallFrame. Hence, there's no issue with setting topCallFrame to their CallFrames
for these functions.

• llint/LLIntSlowPaths.cpp:

(JSC::LLInt::LLINT_SLOW_PATH_DECL):

12:44 PM Changeset in webkit [238140] by Matt Baker

• 5 edits in trunk

Web Inspector: Table should support select all (Cmd-A)
​https://bugs.webkit.org/show_bug.cgi?id=190299
<rdar://problem/45029170>

Reviewed by Devin Rousso.

Source/WebInspectorUI:

• UserInterface/Views/Table.js:

(WI.Table.prototype.selectAll):
(WI.Table.prototype._handleKeyDown):

LayoutTests:

• inspector/table/table-selection-expected.txt:
• inspector/table/table-selection.html:

Add tests that selectAll works when multiple selection is enabled,
and does nothing when multiple selection is disabled.

11:41 AM Changeset in webkit [238139] by Matt Baker

• 2 edits in trunk/Source/WebInspectorUI

Web Inspector: Table with no selection should select the first/last row on down/up arrow key
​https://bugs.webkit.org/show_bug.cgi?id=190100
<rdar://problem/44879243>

Reviewed by Devin Rousso.

Pressing the up or down arrow key when nothing is selected should select
the last or first row, respectively. After selecting the row make sure
it is visible by calling the new Table method revealRow.

• UserInterface/Views/Table.js:

(WI.Table.prototype.revealRow):
(WI.Table.prototype._handleKeyDown):
(WI.Table.prototype._selectRowsFromArrowKey):

10:44 AM Changeset in webkit [238138] by timothy@apple.com

• 4 edits
  2 adds in trunk

Treat supported-color-schemes as the second highest priority property.
​https://bugs.webkit.org/show_bug.cgi?id=191556
rdar://problem/46000076

Reviewed by Dean Jackson.

Source/WebCore:

Test: css-dark-mode/supported-color-schemes-priority.html

• css/CSSProperties.json: Make supported-color-schemes high-priority and add a comment.
• css/StyleResolver.cpp:

(WebCore::StyleResolver::applyMatchedProperties): Manually handle supported-color-schemes
after -webkit-ruby-position, before other properties, so it can affect resolved colors.

LayoutTests:

• css-dark-mode/supported-color-schemes-priority-expected.txt: Added.
• css-dark-mode/supported-color-schemes-priority.html: Added.

10:43 AM Changeset in webkit [238137] by rniwa@webkit.org

• 3 edits in trunk/Source/WebKit

Enable process swap on cross-site navigation by default on macOS
​https://bugs.webkit.org/show_bug.cgi?id=191572

Reviewed by Chris Dumez.

Enabled the feature by default on macOS.

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

10:43 AM Changeset in webkit [238136] by youenn@apple.com

• 58 edits
  1 copy
  1 move
  5 adds
  29 deletes in trunk/LayoutTests

Refresh WPT webrtc tests to ToT
​https://bugs.webkit.org/show_bug.cgi?id=191564

Reviewed by Eric Carlson.

LayoutTests/imported/w3c:

Also remove QUIC specific tests that are not related to webrtc-pc.

• resources/import-expectations.json:
• web-platform-tests/webrtc: refreshed.

LayoutTests:

Refresh webrtc tests

• tests-options.json:

10:40 AM Changeset in webkit [238135] by Nikita Vasilyev

• 3 edits in trunk/Source/WebInspectorUI

Web Inspector: Styles: Command-A should select all properties
​https://bugs.webkit.org/show_bug.cgi?id=191435
<rdar://problem/45921373>

Reviewed by Devin Rousso.

When focused on a style property, Command-A on Mac (Control-A on other platforms)
should select all properties of the style rule.

• UserInterface/Base/Utilities.js:
• UserInterface/Views/SpreadsheetCSSStyleDeclarationEditor.js:

(WI.SpreadsheetCSSStyleDeclarationEditor.prototype.selectProperties):
(WI.SpreadsheetCSSStyleDeclarationEditor.prototype._handleKeyDown):

9:35 AM Changeset in webkit [238134] by dbates@webkit.org

• 15 edits
  1 delete in trunk

Consolidate WebKit UIKitSPI.h and UIKitTestSPI.h
​https://bugs.webkit.org/show_bug.cgi?id=173341
<rdar://problem/32752890>

Reviewed by Simon Fraser.

Source/WebKit:

• Platform/spi/ios/UIKitSPI.h:

Tools:

There is little value to making a distinction between forward declarations
of UIKit SPI needed in the WebKit layer and just for testing. Moreover,
the contents of these two headers can conflict with each other following
r218275 as DumpRenderTree includes both headers. Instead we should remove
UIKitTestSPI.h and have WebKitTestRunner include the WebKit variant, UIKitSPI.h.

• DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
• DumpRenderTree/mac/DumpRenderTree.mm:
• TestRunnerShared/spi/UIKitTestSPI.h: Removed.
• WebKitTestRunner/Configurations/Base.xcconfig:
• WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
• WebKitTestRunner/cocoa/TestRunnerWKWebView.mm:
• WebKitTestRunner/ios/GeneratedTouchesDebugWindow.mm:
• WebKitTestRunner/ios/HIDEventGenerator.h:
• WebKitTestRunner/ios/HIDEventGenerator.mm:
• WebKitTestRunner/ios/PlatformWebViewIOS.mm:
• WebKitTestRunner/ios/TestControllerIOS.mm:
• WebKitTestRunner/ios/UIScriptControllerIOS.mm:
• WebKitTestRunner/ios/mainIOS.mm:

9:16 AM Changeset in webkit [238133] by achristensen@apple.com

• 2 edits in trunk/Source/WebKit

Fix internal builds after r238115
​https://bugs.webkit.org/show_bug.cgi?id=191441

• UIProcess/Cocoa/WKSafeBrowsingWarning.mm:

(-[WKSafeBrowsingTextView intrinsicContentSize]):
Some iOS-like operating systems don't have safe browsing. Don't use symbols that don't exist on those systems.

7:46 AM Changeset in webkit [238132] by Caio Lima

• 9 edits
  1 add in trunk

[BigInt] JSBigInt::createWithLength should throw when length is greater than JSBigInt::maxLength
​https://bugs.webkit.org/show_bug.cgi?id=190836

Reviewed by Saam Barati.

JSTests:

• stress/big-int-out-of-memory-tests.js: Added.

Source/JavaScriptCore:

In this patch we are creating a new method called JSBigInt::createWithLengthUnchecked
where we allocate a BigInt trusting the length received as argument.
With this additional method, we now check if length passed to
JSBigInt::createWithLength is not greater than JSBigInt::maxLength.
When the length is greater than maxLength, we then throw OOM
exception.
This required change the interface of some JSBigInt operations to
receive ExecState* instead of VM&. We changed only operations that
can throw because of OOM.
We beleive that this approach of throwing instead of finishing the
execution abruptly is better because JS programs can catch such
exception and handle this issue properly.

• dfg/DFGOperations.cpp:
• jit/JITOperations.cpp:
• runtime/CommonSlowPaths.cpp:

(JSC::SLOW_PATH_DECL):

• runtime/JSBigInt.cpp:

(JSC::JSBigInt::createZero):
(JSC::JSBigInt::tryCreateWithLength):
(JSC::JSBigInt::createWithLengthUnchecked):
(JSC::JSBigInt::createFrom):
(JSC::JSBigInt::multiply):
(JSC::JSBigInt::divide):
(JSC::JSBigInt::copy):
(JSC::JSBigInt::unaryMinus):
(JSC::JSBigInt::remainder):
(JSC::JSBigInt::add):
(JSC::JSBigInt::sub):
(JSC::JSBigInt::bitwiseAnd):
(JSC::JSBigInt::bitwiseOr):
(JSC::JSBigInt::bitwiseXor):
(JSC::JSBigInt::absoluteAdd):
(JSC::JSBigInt::absoluteSub):
(JSC::JSBigInt::absoluteDivWithDigitDivisor):
(JSC::JSBigInt::absoluteDivWithBigIntDivisor):
(JSC::JSBigInt::absoluteLeftShiftAlwaysCopy):
(JSC::JSBigInt::absoluteBitwiseOp):
(JSC::JSBigInt::absoluteAddOne):
(JSC::JSBigInt::absoluteSubOne):
(JSC::JSBigInt::toStringGeneric):
(JSC::JSBigInt::rightTrim):
(JSC::JSBigInt::allocateFor):
(JSC::JSBigInt::createWithLength): Deleted.

• runtime/JSBigInt.h:
• runtime/Operations.cpp:

(JSC::jsAddSlowCase):

• runtime/Operations.h:

(JSC::jsSub):
(JSC::jsMul):

6:45 AM Changeset in webkit [238131] by cturner@igalia.com

• 7 edits in trunk/Source/WebCore

[EME][GStreamer] Make CDMInstance's available in decryptors, and factor out some EME utility classes.
​https://bugs.webkit.org/show_bug.cgi?id=191316

Reviewed by Xabier Rodriguez-Calvar.

Another preparation in patch getting ready to move the decryption
logic behind the CDMInstance and out of the GStreamer decryptors
themselves. The first step taken here is to arrange for the
instances to always be available in the decryptors when they need
to decrypt content.

In doing so, there were a number of hairy bits of code that could
use some abstraction, so the opportunity was taken to do that as
well.

Covered by tests in media/encrypted-media and
imported/w3c/web-platform-tests/encrypted-media.

• platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:

(WebCore::MediaPlayerPrivateGStreamer::handleMessage): Remove
drm-key-needed since it was not being used anywhere.

• platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:

(WebCore::MediaPlayerPrivateGStreamerBase::handleSyncMessage):
Factor out the parsing of decryption system information from
GStreamer, since it was not clear what that code was doing unless
you squinted pretty hard. Also remove the duplicated
initialization-data-encountered posting.
(WebCore::MediaPlayerPrivateGStreamerBase::initializationDataEncountered):
Refactored to make it a more general method and usable in more
situations. It now has an optional to stop it from eliding init
datas for a different key system. This is required the first time
we post them, since if a CDM instance has already been set, and if
the stream init datas are for different systems, we ended up never
posting an encrypted event.
(WebCore::MediaPlayerPrivateGStreamerBase::attemptToDecryptWithLocalInstance):
Actually send a CDMInstance now when in regular playback mode.
(WebCore::MediaPlayerPrivateGStreamerBase::dispatchDecryptionKey):
Remove m_needToSendCredentials, it was not being used.
(WebCore::MediaPlayerPrivateGStreamerBase::handleProtectionEvent):
Refactored to use the new initializationDataEncountered.
(WebCore::MediaPlayerPrivateGStreamerBase::reportWaitingForKey):
Log the waiting state, since it was currently not clear what that
logging message was even telling you!
(WebCore::extractEventsAndSystemsFromMessage): Deleted.

• platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.h:
• platform/graphics/gstreamer/eme/GStreamerEMEUtilities.h:

(WebCore::InitData::InitData): New class that encapsulates both
single instantiation and streaming instantiation.
(WebCore::InitData::append): Used for the streaming mode, when you
are concatenating init datas together.
(WebCore::InitData::payload const):
(WebCore::InitData::systemId const):
(WebCore::InitData::payloadContainerType const):
(WebCore::InitData::isFromDifferentContainer):
(WebCore::ProtectionSystemEvents::ProtectionSystemEvents):
(WebCore::ProtectionSystemEvents::events const):
(WebCore::ProtectionSystemEvents::availableSystems const):

• platform/graphics/gstreamer/eme/WebKitCommonEncryptionDecryptorGStreamer.cpp:

(webkitMediaCommonEncryptionDecryptTransformInPlace): If you post
waiting-for-key after requesting a CDM instance, it will flap back
to not waiting for a key almost immediately, didn't make sense
positing after requesting an instance. Also post key-received when
we receive the key.
(webkitMediaCommonEncryptionDecryptSinkEventHandler): It has now
been arranged that a CDMInstance will always be present in an OOB
message, so parse it out here.

• platform/graphics/gstreamer/mse/MediaPlayerPrivateGStreamerMSE.cpp:

(WebCore::MediaPlayerPrivateGStreamerMSE::attemptToDecryptWithInstance):
As above, make sure when posting the OOB that a CDMInstance is present.

6:35 AM Changeset in webkit [238130] by cturner@igalia.com

• 4 edits in trunk/Source/WebCore

Various compiler warnings/errors fixes.
​https://bugs.webkit.org/show_bug.cgi?id=191583

Reviewed by Frédéric Wang.

• Modules/indexeddb/server/MemoryIDBBackingStore.cpp:

(WebCore::IDBServer::MemoryIDBBackingStore::clearObjectStore):
ASSERT is only compiled in DEBUG mode, so guarding with
!LOG_DISABLED is wrong.

• rendering/RenderLayerCompositor.cpp:

(WebCore::RenderLayerCompositor::updateCompositingLayers):
showPaintOrderTree is only compiled in ENABLE(TREE_DEBUGGING)
mode, so guarding with !LOG_DISABLED was wrong.
(WebCore::RenderLayerCompositor::computeCompositingRequirements):
Ditto, this time with member .depth.
(WebCore::RenderLayerCompositor::traverseUnchangedSubtree): Ditto.

• rendering/svg/SVGRenderSupport.cpp:

(WebCore::SVGRenderSupport::styleChanged): Add another unused
parameter.

5:07 AM Changeset in webkit [238129] by ajuma@chromium.org

• 2 edits in trunk/Source/WebKit

Turn Intersection Observer on by default
​https://bugs.webkit.org/show_bug.cgi?id=191569

Reviewed by Simon Fraser.

• Shared/WebPreferences.yaml:

4:28 AM WebKitGTK/2.22.x edited by cadubentzen@gmail.com

(diff)

4:27 AM WebKitGTK/2.22.x edited by cadubentzen@gmail.com

(diff)

2:22 AM Changeset in webkit [238128] by graouts@webkit.org

• 10 edits
  2 adds in trunk

[Web Animations] Don't schedule animation frames or update style while an accelerated animation is running
​https://bugs.webkit.org/show_bug.cgi?id=191542
<rdar://problem/45356027>

Reviewed by Simon Fraser.

Source/WebCore:

Test: animations/no-style-recalc-during-accelerated-animation.html

In order to be more power-efficient, we stop scheduling calls to updateAnimationsAndSendEvents() when running only accelerated
animations. To do that, we prevent scheduling further animation resolution if we're in the process of updating animations, and
when we are done, call the new DocumentTimeline::scheduleNextTick() method that will check whether we have only accelerated
animations running, and in that case check which of those animations needs an update the soonest and starts a timer scheduled
for that time when we'll schedule animation resolution.

By default, animations compute the time until their natural completion but in the case of CSS Animations, we want to make sure
we also update animations in-flight to dispatch "animationiteration" events.

• animation/AnimationEffect.h: Make the simpleIterationProgress() public so it can be called by WebAnimation::timeToNextTick().
• animation/DocumentTimeline.cpp:

(WebCore::DocumentTimeline::DocumentTimeline): Create the m_tickScheduleTimer and set it up to call scheduleAnimationResolutionIfNeeded().
(WebCore::DocumentTimeline::suspendAnimations): If we don't already have a cached current time, cache the current time.
(WebCore::DocumentTimeline::resumeAnimations): Reset the cached current time to ensure we'll get a fresh one when updating animations next.
(WebCore::DocumentTimeline::liveCurrentTime const): Factor the code to compute the current time out of currentTime() so that we can
cache the current time in suspendAnimations() without also automatically clearing the current time.
(WebCore::DocumentTimeline::currentTime): Use liveCurrentTime() and cacheCurrentTime() since much of the code from this function has been
factored out into those. Additionally, we were failing to clear the current time if called inside an animation frame, which we now do correctly
by virtue of using cacheCurrentTime(). This fixes some flakiness.
(WebCore::DocumentTimeline::cacheCurrentTime): Factor the code to cache the current time out of currentTime().
(WebCore::DocumentTimeline::maybeClearCachedCurrentTime): No need to clear the current time if we get suspended.
(WebCore::DocumentTimeline::scheduleAnimationResolutionIfNeeded): Prevent scheduling an animation update if we're in the middle of one already,
scheduleNextTick() will be called after animations are updated to see if we should schedule an animation update instead.
(WebCore::DocumentTimeline::unscheduleAnimationResolution): Cancel the m_tickScheduleTimer if we need to unschedule animation resolution.
(WebCore::DocumentTimeline::animationResolutionTimerFired): Factor the call to applyPendingAcceleratedAnimations() out of updateAnimationsAndSendEvents()
and call scheduleNextTick().
(WebCore::DocumentTimeline::updateAnimationsAndSendEvents): Set the new m_isUpdatingAnimations member variable to true while this function is running.
(WebCore::DocumentTimeline::scheduleNextTick): Schedule an animation update immediately if we have any relevant animation that is not accelerated.
Otherwise, iterate through all animations to figure out the earliest moment at which we need to update animations.
(WebCore::DocumentTimeline::updateListOfElementsWithRunningAcceleratedAnimationsForElement): Use the new WebAnimation::isRunningAccelerated() function.

• animation/DocumentTimeline.h:
• animation/WebAnimation.cpp:

(WebCore::WebAnimation::isRunningAccelerated const): Since we end up checking if an animation is running with an accelerated effect, we introduce a new
function to get that information directly through the WebAnimation object without bothering about its effect.
(WebCore::WebAnimation::resolve): We should only call updateFinishedState() here since timingDidChange() would also notify the timeline about a potential
change in relevance, which is not necessary and which would schedule an animation frame even for animations that are accelerated.
(WebCore::WebAnimation::timeToNextTick const): Compute the time until our animation completion or, in the case of CSS animations, the next iteration.

• animation/WebAnimation.h:

LayoutTests:

Add a test that checks that we make only minimal style updates and still dispatch events while an accelerated animation is running.

• animations/no-style-recalc-during-accelerated-animation-expected.txt: Added.
• animations/no-style-recalc-during-accelerated-animation.html: Added.
• fast/layers/no-clipping-overflow-hidden-added-after-transform-expected.html:
• fast/layers/no-clipping-overflow-hidden-added-after-transform.html: Change the colors to avoid a tiny ImageOnlyFailure.
• platform/win/TestExpectations: Mark some regressions tracked by webkit.org/b/191584.

2:05 AM Changeset in webkit [238127] by commit-queue@webkit.org

• 2 edits in trunk/Tools

Remove WebKitTestRunnerLib's build warnings.
​https://bugs.webkit.org/show_bug.cgi?id=191580

Patch by Takashi Komori <​Takashi.Komori@sony.com> on 2018-11-13
Reviewed by Fujii Hironori.

Include cmakeconfig.h in precompiled header.

• WebKitTestRunner/WebKitTestRunnerPrefix.h:

12:58 AM Changeset in webkit [238126] by magomez@igalia.com

• 3 edits in trunk/Source/WebCore

[GTK][WPE] Incorrect tile coverage when resizing a layer out of the visible area
​https://bugs.webkit.org/show_bug.cgi?id=191545

Reviewed by Žan Doberšek.

Keep track of layer size changes even if they happen when the layer is not in the visible
area, so we can update edge tiles when the layer gets visible.

• platform/graphics/texmap/coordinated/TiledBackingStore.cpp:

(WebCore::TiledBackingStore::createTiles):

• platform/graphics/texmap/coordinated/TiledBackingStore.h: