Make sure LibWebRTCMediaEndpoint is always destroyed on the main thread
https://bugs.webkit.org/show_bug.cgi?id=187702

Reviewed by Youenn Fablet.

Make sure LibWebRTCMediaEndpoint is always constructed and destructed on the main thread since
it has a Timer data member and it would not be safe otherwise. LibWebRTCMediaEndpoint is
ThreadSafeRefCounted and frequently passed to other threads.

* Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:
(WebCore::LibWebRTCMediaEndpoint::LibWebRTCMediaEndpoint):
* Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233857 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add --target-path option to dump-class-layout
https://bugs.webkit.org/show_bug.cgi?id=187687

Reviewed by Simon Fraser.

We add an escape hatch to dump-class-layout for specifying target path directly.
This `--target-path` allows us to use dump-class-layout in the other ports
like JSCOnly.

We can dump class layout if we build the target with clang by using the following command.

Tools/Scripts/dump-class-layout \
    --architecture=x86_64 \
    --target-path=path/to/libJavaScriptCore.so \
    JavaScriptCore \
    ScopeNode

* Scripts/dump-class-layout:
(main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233856 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[JSC] Generator and AsyncGeneratorMethod's prototype is incorrect
https://bugs.webkit.org/show_bug.cgi?id=187585

Reviewed by Darin Adler.

JSTests:

* stress/default-proto-for-async-generator.js: Added.
(shouldBe):
(async.asyncGenerator):
* stress/default-proto-for-generator.js: Added.
(shouldBe):
(generator):
* stress/prototype-for-async-generator.js: Added.
(shouldBe):
(async.asyncGenerator):
(A.prototype.async.asyncGenerator):
(A):
* test262/expectations.yaml:

Source/JavaScriptCore:

This patch fixes Generator and AsyncGenerator's prototype issues.

1. Generator's default prototype is incorrect when `generator.prototype = null` is performed.
We fix this by changing JSFunction::prototypeForConstruction.

2. AsyncGeneratorMethod is not handled. We change the name isAsyncGeneratorFunctionParseMode
to isAsyncGeneratorWrapperParseMode since it is aligned to Generator's code. And use it well
to fix `prototype` issues for AsyncGeneratorMethod.

* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::emitPutAsyncGeneratorFields):
(JSC::BytecodeGenerator::emitNewFunction):
* bytecompiler/NodesCodegen.cpp:
(JSC::FunctionNode::emitBytecode):
* parser/ASTBuilder.h:
(JSC::ASTBuilder::createFunctionMetadata):
* parser/Parser.cpp:
(JSC::getAsynFunctionBodyParseMode):
(JSC::Parser<LexerType>::parseInner):
(JSC::Parser<LexerType>::parseAsyncGeneratorFunctionSourceElements):
* parser/ParserModes.h:
(JSC::isAsyncGeneratorParseMode):
(JSC::isAsyncGeneratorWrapperParseMode):
(JSC::isAsyncGeneratorFunctionParseMode): Deleted.
* runtime/FunctionExecutable.h:
* runtime/JSFunction.cpp:
(JSC::JSFunction::prototypeForConstruction):
(JSC::JSFunction::getOwnPropertySlot):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233855 268f45cc-cd09-0410-ab3c-d52691b4dbfc

jsc shell's noFTL utility test function should be more robust.
https://bugs.webkit.org/show_bug.cgi?id=187704
<rdar://problem/42231988>

Reviewed by Michael Saboff and Keith Miller.

* jsc.cpp:
(functionNoFTL):
- only setNeverFTLOptimize() if the function is actually a JS function.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233854 268f45cc-cd09-0410-ab3c-d52691b4dbfc

IndexedDB: closeAndDeleteDatabasesForOrigins should remove all databases for those origins
https://bugs.webkit.org/show_bug.cgi?id=187631
<rdar://problem/42164227>

Reviewed by Brady Eidson.

Source/WebCore:

When asked to delete database for an origin, we deleted the databases whose mainFrameOrigin
is that origin. Given that the origin may create IndexedDB from subframes, we should delete
databases whose openingOrigin is that origin too.

Covered by modified API test: WebKit.WebsiteDataStoreCustomPaths.

* Modules/indexeddb/server/IDBServer.cpp:
(WebCore::IDBServer::IDBServer::performCloseAndDeleteDatabasesForOrigins):

Source/WebKit:

We need to return all origins, both openingOrigin and mainFrameOrigin, of IndexedDB so users
could be better aware of which origins are using databases and decide what they want to
remove.

* StorageProcess/StorageProcess.cpp:
(WebKit::StorageProcess::indexedDatabaseOrigins):
* StorageProcess/StorageProcess.h:

Tools:

* TestWebKitAPI/Tests/WebKitCocoa/WebsiteDataStoreCustomPaths.mm:
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233853 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233852 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Shrink some font-related classes and enums
https://bugs.webkit.org/show_bug.cgi?id=187686

Reviewed by Myles Maxfield.

Use enum class for enums in TextFlags.h and make them one byte big.

Re-order members of Font to shrink it from 360 to 328 bytes.

* css/CSSPrimitiveValueMappings.h:
(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
(WebCore::CSSPrimitiveValue::operator FontSmoothingMode const):
(WebCore::CSSPrimitiveValue::operator FontSmallCaps const):
(WebCore::CSSPrimitiveValue::operator TextRenderingMode const):
* platform/graphics/Font.cpp:
(WebCore::Font::Font):
(WebCore::Font::verticalRightOrientationFont const):
* platform/graphics/Font.h:
* platform/graphics/FontCascade.cpp:
(WebCore::offsetToMiddleOfGlyph):
* platform/graphics/FontCascade.h:
(WebCore::FontCascade::advancedTextRenderingMode const):
* platform/graphics/FontCascadeFonts.cpp:
(WebCore::FontCascadeFonts::glyphDataForSystemFallback):
(WebCore::FontCascadeFonts::glyphDataForVariant):
(WebCore::glyphPageFromFontRanges):
* platform/graphics/FontDescription.cpp:
(WebCore::FontCascadeDescription::FontCascadeDescription):
* platform/graphics/FontDescription.h:
(WebCore::FontDescription::setTextRenderingMode):
(WebCore::FontDescription::setOrientation):
(WebCore::FontDescription::setWidthVariant):
(WebCore::FontCascadeDescription::setFontSmoothing):
(WebCore::FontCascadeDescription::initialSmallCaps):
(WebCore::FontCascadeDescription::initialFontSmoothing):
(WebCore::FontCascadeDescription::initialTextRenderingMode):
* platform/graphics/FontPlatformData.h:
(WebCore::FontPlatformData::isForTextCombine const):
* platform/graphics/cocoa/FontCacheCoreText.cpp:
(WebCore::preparePlatformFont):
* platform/graphics/cocoa/FontCascadeCocoa.mm:
(WebCore::showLetterpressedGlyphsWithAdvances):
(WebCore::showGlyphsWithAdvances):
(WebCore::FontCascade::drawGlyphs):
(WebCore::FontCascade::fontForCombiningCharacterSequence const):
* platform/graphics/cocoa/FontCocoa.mm:
(WebCore::Font::platformInit):
(WebCore::Font::platformBoundsForGlyph const):
(WebCore::Font::platformWidthForGlyph const):
* platform/graphics/cocoa/FontPlatformDataCocoa.mm:
(WebCore::FontPlatformData::hash const):
(WebCore::mapFontWidthVariantToCTFeatureSelector):
(WebCore::FontPlatformData::ctFont const):
(WebCore::FontPlatformData::description const):
* platform/graphics/freetype/FontPlatformDataFreeType.cpp:
(WebCore::FontPlatformData::buildScaledFont):
* platform/graphics/freetype/SimpleFontDataFreeType.cpp:
(WebCore::Font::platformInit):
(WebCore::Font::platformWidthForGlyph const):
* platform/graphics/harfbuzz/ComplexTextControllerHarfBuzz.cpp:
(WebCore::fontFeatures):
(WebCore::ComplexTextController::collectComplexTextRunsForCharacters):
* platform/graphics/mac/SimpleFontDataCoreText.cpp:
(WebCore::Font::getCFStringAttributes const):
* platform/graphics/win/FontCGWin.cpp:
(WebCore::FontCascade::drawGlyphs):
* platform/graphics/win/FontCascadeDirect2D.cpp:
(WebCore::FontCascade::drawGlyphs):
* platform/graphics/win/GlyphPageTreeNodeDirect2D.cpp:
(WebCore::GlyphPage::fill):
* platform/graphics/win/SimpleFontDataDirect2D.cpp:
(WebCore::Font::platformInit):
(WebCore::Font::platformBoundsForGlyph const):
(WebCore::Font::platformWidthForGlyph const):
* platform/text/TextFlags.h:
* rendering/RenderCombineText.cpp:
(WebCore::RenderCombineText::combineTextIfNeeded):
* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::calculateClipRects const):
* rendering/TextPainter.cpp:
(WebCore::TextPainter::paintTextWithShadows):
* rendering/TextPainter.h:
* rendering/style/RenderStyle.cpp:
(WebCore::RenderStyle::fontAndGlyphOrientation):
* rendering/svg/RenderSVGInlineText.cpp:
(WebCore::RenderSVGInlineText::computeNewScaledFontForStyle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233851 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[ iOS ] Layout Test fast/forms/submit-change-fragment.html is a flaky Timeout
https://bugs.webkit.org/show_bug.cgi?id=187699

Unreviewed test gardening.

Patch by Truitt Savell <tsavell@apple.com> on 2018-07-16

* platform/ios-simulator-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233850 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[ EWS ] http/tests/security/contentSecurityPolicy/userAgentShadowDOM/allow-audio.html is Crashing on Win-EWS
https://bugs.webkit.org/show_bug.cgi?id=187700

Unreviewed test gardening.

Patch by Truitt Savell <tsavell@apple.com> on 2018-07-16

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: Dark Mode: Console filter field buttons should be darker
https://bugs.webkit.org/show_bug.cgi?id=187626
<rdar://problem/42142744>

Reviewed by Brian Burg.

* UserInterface/Views/FindBanner.css:
(@media (prefers-dark-interface)):
(.find-banner > button.segmented):
(.find-banner > button.segmented > .glyph):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233848 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: Dark Mode: selected item background color is too light
https://bugs.webkit.org/show_bug.cgi?id=187691
<rdar://problem/42225308>

Reviewed by Brian Burg.

* UserInterface/Views/DarkMode.css:
(@media (prefers-dark-interface)):
(:root):
(.tree-outline.dom li.elements-drag-over .selection-area):
(.tree-outline.dom:focus li.selected .selection-area):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233847 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[WebVR] Add support for connect/disconnect and mount/unmount device events
https://bugs.webkit.org/show_bug.cgi?id=187343

Reviewed by Žan Doberšek.

WebVR specs define a series of events as part of the Window Interface Extension. We're
adding support for the connect/disconnect and mount/unmount events both at the module level
and the platform level using OpenVR.

In order to do that we need to keep lists of VRPlatformDisplays at platform level and
VRDisplays at bindings level. We then update those lists accordingly to detect potential
additions/removals, and emit the corresponding signals. A new client interface
VRPlatformDisplayClient was also defined so that VRPlatformDisplay implementations could
notify their clients (typically a VRDisplay).

Last but not least, NavigatorWebVR was updated so it supplements Navigator instead of
supplementing Page.

* Modules/webvr/NavigatorWebVR.cpp: Supplement Navigator not Page.
(WebCore::NavigatorWebVR::getVRDisplays): Keep a list of VRDisplays and update them
conveniently, also emitting the required events under certain conditions (like device
disconnection).
(WebCore::NavigatorWebVR::supplementName): New method.
(WebCore::NavigatorWebVR::from): Ditto.
* Modules/webvr/NavigatorWebVR.h: Supplement Navigator not Page.
* Modules/webvr/VRDisplay.cpp:
(WebCore::VRDisplay::create): Moved suspendIfNeeded() to constructor.
(WebCore::VRDisplay::VRDisplay): Set itself as VRPlatformDisplay client.
(WebCore::VRDisplay::~VRDisplay): Unset as  VRPlatformDisplay client.
(WebCore::VRDisplay::VRPlatformDisplayConnected): Dispatch event on DOM window.
(WebCore::VRDisplay::VRPlatformDisplayDisconnected): Ditto.
(WebCore::VRDisplay::VRPlatformDisplayMounted): Ditto.
(WebCore::VRDisplay::VRPlatformDisplayUnmounted): Ditto.
* Modules/webvr/VRDisplay.h: Extend from VRPlatformDisplayClient.
(WebCore::VRDisplay::document):
* Modules/webvr/VRDisplayEvent.cpp: Updated Copyright.
* Modules/webvr/VRDisplayEvent.h: Ditto.
* Sources.txt: Added the two new files.
* WebCore.xcodeproj/project.pbxproj: Ditto.
* platform/vr/VRManager.cpp:
(WebCore::VRManager::getVRDisplays): Keep a list of VRPlatformDisplays and update them conveniently,
also emitting the required events under certain conditions (like device disconnection).
* platform/vr/VRManager.h:
* platform/vr/VRPlatformDisplay.cpp: New file with common implementations for VRPlatformDisplays.
(WebCore::VRPlatformDisplay::setClient):
(WebCore::VRPlatformDisplay::notifyVRPlatformDisplayEvent):
* platform/vr/VRPlatformDisplay.h: Added a generic method to notify about different
events. Added the client pointer.
* platform/vr/VRPlatformDisplayClient.h: New file. VRPlatformDisplay implementations will
call the client methods in the event of some circumstances happening.
(WebCore::VRPlatformDisplayClient::VRPlatformDisplayConnected):
(WebCore::VRPlatformDisplayClient::VRPlatformDisplayDisconnected):
(WebCore::VRPlatformDisplayClient::VRPlatformDisplayMounted):
(WebCore::VRPlatformDisplayClient::VRPlatformDisplayUnmounted):
* platform/vr/openvr/VRPlatformDisplayOpenVR.cpp:
(WebCore::VRPlatformDisplayOpenVR::updateDisplayInfo): Poll the device for new events to
detect connection/disconnections or device activations/deactivations (HMD
mounted/unmounted).
* platform/vr/openvr/VRPlatformDisplayOpenVR.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233846 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Nicosia] Add Nicosia::PlatformLayer, Nicosia::CompositionLayer classes
https://bugs.webkit.org/show_bug.cgi?id=187693

Reviewed by Carlos Garcia Campos.

Add the Nicosia::PlatformLayer class. This will be the base platform
layer class from which different derivatives will be created, addressing
different use cases. The generic PlatformLayer type alias will point to
this class in the future.

First class deriving from Nicosia::PlatformLayer is
Nicosia::CompositionLayer, purpose of which will be to mirror the state
that's stored in the platform-specific GraphicsLayer derivative. It will
also allow making thread-safe updates to that state.

CoordinatedGraphicsLayer implementation now spawns a CompositionLayer
object and tracks state changes in a separate
CompositionLayer::LayerState::Delta object. During flushing, the changed
state is applied to the layer's pending state before the delta is nulled
out. The updated state isn't used anywhere yet, but future changes will
implement committing this state into the rendering pipeline.

There's bits of state not yet being managed by CompositionLayer, e.g.
debug visuals, filters and animations. These will be addressed later.

The m_solidColor member variable is added to CoordinatedGraphicsLayer in
order to properly store the solid color value. Normally this would be
contained by the parent GraphicsLayer class, but no such member variable
exists there.

* platform/TextureMapper.cmake:
* platform/graphics/nicosia/NicosiaPlatformLayer.cpp: Added.
(Nicosia::PlatformLayer::PlatformLayer):
(Nicosia::CompositionLayer::CompositionLayer):
* platform/graphics/nicosia/NicosiaPlatformLayer.h: Added.
(Nicosia::PlatformLayer::isCompositionLayer const):
(Nicosia::PlatformLayer::id const):
(Nicosia::CompositionLayer::create):
(Nicosia::CompositionLayer::LayerState::Flags::Flags):
(Nicosia::CompositionLayer::updateState):
* platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.cpp:
(WebCore::CoordinatedGraphicsLayer::CoordinatedGraphicsLayer):
(WebCore::CoordinatedGraphicsLayer::setPosition):
(WebCore::CoordinatedGraphicsLayer::setAnchorPoint):
(WebCore::CoordinatedGraphicsLayer::setSize):
(WebCore::CoordinatedGraphicsLayer::setTransform):
(WebCore::CoordinatedGraphicsLayer::setChildrenTransform):
(WebCore::CoordinatedGraphicsLayer::setPreserves3D):
(WebCore::CoordinatedGraphicsLayer::setMasksToBounds):
(WebCore::CoordinatedGraphicsLayer::setDrawsContent):
(WebCore::CoordinatedGraphicsLayer::setContentsVisible):
(WebCore::CoordinatedGraphicsLayer::setContentsOpaque):
(WebCore::CoordinatedGraphicsLayer::setBackfaceVisibility):
(WebCore::CoordinatedGraphicsLayer::setOpacity):
(WebCore::CoordinatedGraphicsLayer::setContentsRect):
(WebCore::CoordinatedGraphicsLayer::setContentsTileSize):
(WebCore::CoordinatedGraphicsLayer::setContentsTilePhase):
(WebCore::CoordinatedGraphicsLayer::setContentsToSolidColor):
(WebCore::CoordinatedGraphicsLayer::setMaskLayer):
(WebCore::CoordinatedGraphicsLayer::setReplicatedByLayer):
(WebCore::CoordinatedGraphicsLayer::syncChildren):
(WebCore::CoordinatedGraphicsLayer::flushCompositingStateForThisLayerOnly):
* platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233845 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GLIB] Add API to evaluate code using a given object to store global symbols
https://bugs.webkit.org/show_bug.cgi?id=187639

Reviewed by Michael Catanzaro.

Source/JavaScriptCore:

Add jsc_context_evaluate_in_object(). It returns a new object as an out parameter. Global symbols in the
evaluated script are added as properties to the new object instead of to the context global object. This is
similar to JS::Evaluate in spider monkey when a scopeChain parameter is passed, but JSC doesn't support using a
scope for assignments, so we have to create a new context and get its global object. This patch also updates
jsc_context_evaluate_with_source_uri() to receive the starting line number for consistency with the new
jsc_context_evaluate_in_object().

* API/glib/JSCContext.cpp:
(jsc_context_evaluate): Pass 0 as line number to jsc_context_evaluate_with_source_uri().
(evaluateScriptInContext): Helper function to evaluate a script in a JSGlobalContextRef.
(jsc_context_evaluate_with_source_uri): Use evaluateScriptInContext().
(jsc_context_evaluate_in_object): Create a new context and set the main context global object as extension
scope of it. Evaluate the script in the new context and get its global object to be returned as parameter.
* API/glib/JSCContext.h:
* API/glib/docs/jsc-glib-4.0-sections.txt:

Tools:

Add a new test case.

* TestWebKitAPI/Tests/JavaScriptCore/glib/TestJSC.cpp:
(testJSCEvaluateInObject):
(testJSCExceptions):
(main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233844 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[webkitpy] run-web-platform-tests should allow specifying custom WPT metadata directories
https://bugs.webkit.org/show_bug.cgi?id=187353

Reviewed by Youenn Fablet.

When using run-web-platform-tests, allow specifying custom WPT metadata
directory. This will avoid generating such metadata-filled directory
from the port-specific JSON expectations file, and instead will search
for the metadata files and the include manifest inside the specified
directory. The MANIFEST.json file will also be generated under there.

Rationale for this change is prototyping using custom metadata
directories and avoiding generating all this content from a JSON file.
Using this by default in the future would avoid sporradic changes in how
WPT handles metadata .ini files and releases us from the obligation of
adjusting the generator for every such change. This would also allow
managing this metadata content in a separate repository, avoiding
polluting the main webkit.org repository with thousands of .ini files.

If this turns out to be the preferable way of managing the metadata
content, the JSON files under WebPlatformTests/ and the related code in
wpt_runner.py would be removed.

* Scripts/webkitpy/w3c/wpt_runner.py:
(parse_args):
(WPTRunner._wpt_run_paths):
(WPTRunner.run):
* Scripts/webkitpy/w3c/wpt_runner_unittest.py:
(WPTRunnerTest.test_run_with_specified_options):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233843 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[SOUP] http/tests/misc/bubble-drag-events.html crashes
https://bugs.webkit.org/show_bug.cgi?id=182352

Reviewed by Youenn Fablet.

PingLoad is not refcounted and deletes itself when the load finishes. The problem is that the network data
task can also finish the load, causing the PingLoad to be deleted early. We tried to fix it in r233032 in the
network data task soup implementation, but it caused regressions.

* NetworkProcess/PingLoad.cpp:
(WebKit::PingLoad::didReceiveChallenge): Create a weak ref for the ping load before calling the operation completion
handler and return early after the completion handler if it was deleted.
(WebKit::PingLoad::didReceiveResponseNetworkSession): Ditto.
* NetworkProcess/PingLoad.h:
* NetworkProcess/soup/NetworkDataTaskSoup.cpp:
(WebKit::NetworkDataTaskSoup::continueAuthenticate): Revert r233032.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233842 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[iOS apps on macOS] Playing embedded Twitter videos in the News app crashes the web process
https://bugs.webkit.org/show_bug.cgi?id=187690
<rdar://problem/41869703>

Reviewed by Tim Horton.

Work around unexpected behavior when soft-linking AVFoundation. After using `dlopen_preflight` to check for the
existence of a library prior to loading the library using `dlopen`, `dlsym` subsequently returns null for some
symbols that would otherwise be available. This causes us to RELEASE_ASSERT later down the road when we try to
load AVAudioSessionModeDefault in AudioSessionIOS.mm.

To fix this for now, simply check for the library directly instead of using the more lightweight preflight
check. See clone: <rdar://problem/42224780> for more detail.

* platform/graphics/avfoundation/objc/AVFoundationMIMETypeCache.mm:
(WebCore::AVFoundationMIMETypeCache::isAvailable const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233841 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Shrink StyleFillData, StyleStrokeData and StyleMiscData
https://bugs.webkit.org/show_bug.cgi?id=187681

Reviewed by Anders Carlsson.

Shrink these data structures by making more enum classes one byte in size, and
re-ordering. StyleFillData goes from 56 to 48, StyleStrokeData from 80 to 72,
StyleMiscData from 40 to 32 bytes.

* rendering/style/SVGRenderStyleDefs.cpp:
(WebCore::StyleFillData::StyleFillData):
(WebCore::StyleFillData::operator== const):
(WebCore::StyleStrokeData::StyleStrokeData):
(WebCore::StyleStrokeData::operator== const):
(WebCore::StyleStopData::operator== const):
(WebCore::StyleMiscData::StyleMiscData):
* rendering/style/SVGRenderStyleDefs.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233840 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Shrink CachedResource and subclasses
https://bugs.webkit.org/show_bug.cgi?id=187546

Reviewed by Daniel Bates.

Shrink CachedResource down from 1384 to 1336 bytes, CachedImage from 1480 to
1424 bytes, and CachedFont a little.

This saves about 23KB on cnn.com.

* loader/ResourceLoaderOptions.h:
* loader/cache/CachedFont.h:
* loader/cache/CachedImage.h:
* loader/cache/CachedResource.cpp:
(WebCore::CachedResource::CachedResource):
* loader/cache/CachedResource.h:
* platform/network/CacheValidation.h:
(WebCore::RedirectChainCacheStatus::RedirectChainCacheStatus):
* platform/network/NetworkLoadMetrics.h:
* platform/network/ParsedContentRange.h:
* platform/network/ResourceRequestBase.h:
* platform/network/ResourceResponseBase.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233839 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Shrink some style-related classes and enums
https://bugs.webkit.org/show_bug.cgi?id=187680

Reviewed by Antti Koivisto.

Make all the enum classes in RenderStyleConstants be one byte big (all have less
than 256 values).

Shrink DocumentRuleSet from 384 to 368 bytes by re-ordering, which also helps shrink
StyleResolver from 1024 to 952 bytes.

Shrink BorderValue by re-ordering (now that the layout of Color has changed) which
shrinks BorderData from 168 to 136 bytes.

Convert a couple of other enums to enum class so that they can have explicit size.

* css/DocumentRuleSets.h:
* css/MediaQueryMatcher.cpp:
(WebCore::MediaQueryMatcher::documentElementUserAgentStyle const):
* css/StyleMedia.cpp:
(WebCore::StyleMedia::matchMedium const):
* css/StyleResolver.cpp:
(WebCore::StyleResolver::StyleResolver):
(WebCore::StyleResolver::State::State):
(WebCore::StyleResolver::styleForElement):
(WebCore::StyleResolver::cascadedPropertiesForRollback):
(WebCore::StyleResolver::applyProperty):
(WebCore::cascadeLevelForIndex):
* css/StyleResolver.h:
* rendering/style/BorderValue.h:
* rendering/style/RenderStyle.cpp:
* rendering/style/RenderStyleConstants.h:
* style/StyleTreeResolver.cpp:
(WebCore::Style::TreeResolver::styleForElement):
* svg/SVGElementRareData.h:
(WebCore::SVGElementRareData::overrideComputedStyle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233838 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Ensure WebKit stack is ad-hoc signed
https://bugs.webkit.org/show_bug.cgi?id=187667

Patch by Kocsen Chung <kocsen_chung@apple.com> on 2018-07-14
Reviewed by Alexey Proskuryakov.

* Configurations/Base.xcconfig:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233837 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Replace TR2_OPTIONAL_ASSERTED_EXPRESSION macro in <wtf/Optional.h>
<https://webkit.org/b/187672>

Reviewed by Frédéric Wang.

* wtf/Optional.h:
(std::optional::operator -> const):
(std::optional::operator * const):
(std::optional<T::operator-> const):
(std::optional<T::operator* const):
- Replace TR2_OPTIONAL_ASSERTED_EXPRESSION macro with
  ASSERT_UNDER_CONSTEXPR_CONTEXT macro and return statement.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233836 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[css-masking] Fully support -webkit-clip-path on SVG elements
https://bugs.webkit.org/show_bug.cgi?id=185829

Reviewed by Simon Fraser.

Source/WebCore:

-webkit-clip-path contributes to SVG elements with boxes, shapes and now with
element references to <clipPath> elements as well. Make sure that all types
contribute to hit-testing of the SVG element as well as they should.

Tests: svg/clip-path/webkit-clip-path-after-expected.svg
       svg/clip-path/webkit-clip-path-after.svg
       svg/clip-path/webkit-clip-path-before-expected.svg
       svg/clip-path/webkit-clip-path-before.svg
       svg/dynamic-updates/SVGClipPath-prefixed-influences-hitTesting.html
       svg/dynamic-updates/SVGClipPath-prefixed-path-influences-hitTesting.html
       svg/dynamic-updates/SVGClipPathElement-prefixed-css-transform-influences-hitTesting.html
       svg/dynamic-updates/SVGClipPathElement-prefixed-transform-influences-hitTesting.html

* rendering/svg/SVGRenderSupport.cpp: Share code as much as possible.
(WebCore::clipPathReferenceBox):
(WebCore::isPointInCSSClippingArea): Take -webkit-clip-path into account.
(WebCore::SVGRenderSupport::clipContextToCSSClippingArea):
(WebCore::SVGRenderSupport::pointInClippingArea):
* rendering/svg/SVGRenderSupport.h:
* rendering/svg/SVGRenderingContext.cpp: Clip to -webkit-clip-path boxes, shapes and references.
(WebCore::SVGRenderingContext::prepareToRenderSVGContent):
* rendering/svg/SVGResources.cpp: Add -webkit-clip-path references to cached resources. Mimic SVG clip-path.
(WebCore::SVGResources::buildCachedResources):

LayoutTests:

Test -webkit-clip-path element references on SVG elements. Make sure, -webkit-clip-path
contributes to hit testing for element references and basic shapes.

* svg/clip-path/webkit-clip-path-after-expected.svg: Added.
* svg/clip-path/webkit-clip-path-after.svg: Added.
* svg/clip-path/webkit-clip-path-before-expected.svg: Added.
* svg/clip-path/webkit-clip-path-before.svg: Added.
* svg/dynamic-updates/SVGClipPath-prefixed-influences-hitTesting-expected.txt: Added.
* svg/dynamic-updates/SVGClipPath-prefixed-influences-hitTesting.html: Added.
* svg/dynamic-updates/SVGClipPath-prefixed-path-influences-hitTesting-expected.txt: Added.
* svg/dynamic-updates/SVGClipPath-prefixed-path-influences-hitTesting.html: Added.
* svg/dynamic-updates/SVGClipPathElement-prefixed-css-transform-influences-hitTesting-expected.txt: Added.
* svg/dynamic-updates/SVGClipPathElement-prefixed-css-transform-influences-hitTesting.html: Added.
* svg/dynamic-updates/SVGClipPathElement-prefixed-transform-influences-hitTesting-expected.txt: Added.
* svg/dynamic-updates/SVGClipPathElement-prefixed-transform-influences-hitTesting.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233835 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[WinCairo] run-api-tests is timing out for almost all test cases
https://bugs.webkit.org/show_bug.cgi?id=187547

Reviewed by Fujii Hironori.

* Scripts/webkitpy/port/server_process.py:
(ServerProcess._wait_for_data_and_update_buffers_using_win32_apis):
Still attempt to read from stdout/stderr even if the process has exited, as long as it's before the deadline.

* Scripts/webkitpy/port/server_process_unittest.py:
(TestServerProcess.test_read_after_process_exits):
Add test case.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233834 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Avoid fetching visitedDependentColor() so many times in editing code
https://bugs.webkit.org/show_bug.cgi?id=187676

Reviewed by Zalan Bujtas.

editingAttributedStringFromRange called style.visitedDependentColor() twice for each property,
and fontAttributesForSelectionStart() called it two or three times. Use a local Color variable
to avoid so many calls. Also replace a call to alpha() with isVisible() which makes the usage more clear.

No behavior change.

* editing/cocoa/EditorCocoa.mm:
(WebCore::Editor::fontAttributesForSelectionStart const):
* editing/cocoa/HTMLConverter.mm:
(WebCore::editingAttributedStringFromRange):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233833 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add TestExpectations and baselines for iOS 12
https://bugs.webkit.org/show_bug.cgi?id=187628

Reviewed by Alexey Proskuryakov.

Tools:

* Scripts/webkitpy/port/ios.py:
(IOSPort): Bump current version to 12.

LayoutTests:

* platform/ios-12/TestExpectations: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233832 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add TestExpectations and baselines for Mojave.
https://bugs.webkit.org/show_bug.cgi?id=187620

Reviewed by Alexey Proskuryakov.

Tools:

* Scripts/webkitpy/port/mac.py:
(MacPort): Bump current version to 10.14.

LayoutTests:

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233831 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Flesh out WebSocket cookie tests to cover cookie policy for third-party resources
https://bugs.webkit.org/show_bug.cgi?id=187541
<rdar://problem/42048729>

Reviewed by Alex Christensen.

* http/tests/cookies/resources/cookie-utilities.js:
    Added a function for setting a cookie in a WebSocket handshake.
* http/tests/websocket/tests/hybi/cookie_wsh.py:
(web_socket_do_extra_handshake):
    Now sets the root path for new cookies so that they can be seen by
    for example cookies/resources/echo-cookies.php.
* http/tests/websocket/tests/hybi/websocket-allowed-setting-cookie-as-third-party-expected.txt: Added.
* http/tests/websocket/tests/hybi/websocket-allowed-setting-cookie-as-third-party.html: Added.
* http/tests/websocket/tests/hybi/websocket-blocked-from-setting-cookie-as-third-party-expected.txt: Added.
* http/tests/websocket/tests/hybi/websocket-blocked-from-setting-cookie-as-third-party.html: Added.
* http/tests/websocket/tests/hybi/websocket-cookie-overwrite-behavior-expected.txt:
* http/tests/websocket/tests/hybi/websocket-cookie-overwrite-behavior.html:
    Now tests under the condition where localhost as third-party is
    allowed to set a new cookie as third party. It also makes sure to use
    cookies with the path set to the root so that all cookies are visible.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233830 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Support connecting a MediaStreamAudioDestinationNode to RTCPeerConnection
https://bugs.webkit.org/show_bug.cgi?id=187627
<rdar://problem/35334400>

Reviewed by Jer Noble.

Source/WebCore:

When MediaStreamAudioSource is called to read new audio samples,
convert these samples to a WebAudioBufferList and call RealtimeMediaSource::audioSamplesAvailable.
This makes its observers to get the audio data.

Test: webrtc/peer-connection-createMediaStreamDestination.html

* Modules/mediastream/MediaStream.cpp:
(WebCore::MediaStream::create): Minor refactoring.
* Modules/webaudio/MediaStreamAudioDestinationNode.cpp:
(WebCore::createMediaStream):
(WebCore::MediaStreamAudioDestinationNode::MediaStreamAudioDestinationNode):
(WebCore::MediaStreamAudioDestinationNode::process):
* Modules/webaudio/MediaStreamAudioDestinationNode.h:
* Modules/webaudio/MediaStreamAudioSource.cpp:
(WebCore::MediaStreamAudioSource::MediaStreamAudioSource):
(WebCore::MediaStreamAudioSource::consumeAudio):
* Modules/webaudio/MediaStreamAudioSource.h:
* Modules/webaudio/MediaStreamAudioSourceCocoa.cpp: Added.
(WebCore::streamDescription):
(WebCore::MediaStreamAudioSource::consumeAudio):
* WebCore.xcodeproj/project.pbxproj:
* platform/audio/AudioDestinationConsumer.h: Removed.
* platform/mediastream/MediaStreamPrivate.cpp:
(WebCore::MediaStreamPrivate::create):
* platform/mediastream/MediaStreamPrivate.h:

LayoutTests:

* webrtc/peer-connection-createMediaStreamDestination-expected.txt: Added.
* webrtc/peer-connection-createMediaStreamDestination.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233829 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add _drawsBackground to WKWebViewConfiguration.
https://bugs.webkit.org/show_bug.cgi?id=187665
rdar://problem/42182268

Reviewed by Tim Horton.

Source/WebKit:

* UIProcess/API/APIPageConfiguration.cpp:
(API::PageConfiguration::copy const): Copy m_drawsBackground, and some missing values.
* UIProcess/API/APIPageConfiguration.h:
(API::PageConfiguration::drawsBackground const): Added.
(API::PageConfiguration::setDrawsBackground): Added.
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]): Transfer _drawsBackground to page config.
* UIProcess/API/Cocoa/WKWebViewConfiguration.mm:
(-[WKWebViewConfiguration init]): Set _drawsBackground to YES.
(-[WKWebViewConfiguration encodeWithCoder:]): Encode _drawsBackground.
(-[WKWebViewConfiguration initWithCoder:]): Decode _drawsBackground.
(-[WKWebViewConfiguration copyWithZone:]): Copy _drawsBackground.
(-[WKWebViewConfiguration _drawsBackground]): Added.
(-[WKWebViewConfiguration _setDrawsBackground:]): Added.
* UIProcess/API/Cocoa/WKWebViewConfigurationPrivate.h:
* UIProcess/WebPageProxy.cpp: Set m_drawsBackground based on configuration.

Tools:

* TestWebKitAPI/Tests/WebKitCocoa/Configuration.mm:
(TestWebKitAPI.WebKit.ConfigurationDrawsBackground): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233828 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[WinCairo] Enable WebGL when Accelerated Compositing is disabled
https://bugs.webkit.org/show_bug.cgi?id=187664

Reviewed by Fujii Hironori.

AC was disabled for WinCairo in r233725 but it can still run WebGL without AC.

* html/HTMLCanvasElement.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233827 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: REGRESSION (r195723): Improve support for resources with '\r' and '\r\n' line endings
https://bugs.webkit.org/show_bug.cgi?id=186453
<rdar://problem/39689180>

Reviewed by Joseph Pecoraro.

Now that the frontend no longer uses offsets from the original source
file to calculate positions within CodeMirror, it is possible to support
resources with '\r' and '\r\n' line endings in the editor.

* UserInterface/Views/CodeMirrorEditor.js:
(WI.CodeMirrorEditor.create):
(WI.CodeMirrorEditor):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233824 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Crash under ApplicationCacheGroup::didFailLoadingEntry()
https://bugs.webkit.org/show_bug.cgi?id=187661
<rdar://problem/42179755>

Reviewed by Youenn Fablet.

If ApplicationCacheResourceLoader::create() fails synchronously with
ApplicationCacheResourceLoader::Error::CannotCreateResource error, then
m_entryLoader will be null when didFailLoadingEntry() is called. However,
didFailLoadingEntry() fails to null check m_entryLoader before using it.

* loader/appcache/ApplicationCacheGroup.cpp:
(WebCore::ApplicationCacheGroup::didFailLoadingEntry):
(WebCore::ApplicationCacheGroup::startLoadingEntry):
* loader/appcache/ApplicationCacheGroup.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233823 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKit sometimes holds WiFi/BT assertions while the Networking process is suspended
https://bugs.webkit.org/show_bug.cgi?id=187662
<rdar://problem/41999448>

Reviewed by Chris Dumez.

Inform WiFiAssertions when the Networking process is first going into the background,
so it has a chance of dropping its assertions even in cases where the system
suspends the process before we receive prepareToSuspend.

* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::processDidTransitionToForeground):
(WebKit::NetworkProcess::processDidTransitionToBackground):
* NetworkProcess/NetworkProcess.h:
* NetworkProcess/NetworkProcess.messages.in:
* NetworkProcess/curl/NetworkProcessCurl.cpp:
(WebKit::NetworkProcess::platformProcessDidTransitionToForeground):
(WebKit::NetworkProcess::platformProcessDidTransitionToBackground):
* NetworkProcess/soup/NetworkProcessSoup.cpp:
(WebKit::NetworkProcess::platformProcessDidTransitionToForeground):
(WebKit::NetworkProcess::platformProcessDidTransitionToBackground):
* UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::sendProcessDidTransitionToForeground):
(WebKit::NetworkProcessProxy::sendProcessDidTransitionToBackground):
* UIProcess/Network/NetworkProcessProxy.h:
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::updateProcessAssertions):
Plumb the foreground/background transition to NetworkProcess.

* NetworkProcess/cocoa/NetworkProcessCocoa.mm:
(WebKit::NetworkProcess::platformPrepareToSuspend):
(WebKit::NetworkProcess::platformProcessDidTransitionToBackground):
(WebKit::NetworkProcess::platformProcessDidTransitionToForeground):
Make use of SuspensionReason to explain to WiFiAssertions the
difference between prepareToSuspend and didTransitionToBackground,
so that it can adjust the timing of dropping the assertion.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233822 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add release assertion to check thread in TimerBase::setNextFireTime
https://bugs.webkit.org/show_bug.cgi?id=187666

Reviewed by Ryosuke Niwa.

This should give us insight into what is causing <rdar://problem/33352721>

* platform/Timer.cpp:
(WebCore::TimerBase::setNextFireTime):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233821 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: Basic blocks highlighting should use line/column locations instead of offsets
https://bugs.webkit.org/show_bug.cgi?id=187613
<rdar://problem/42131808>

Reviewed by Joseph Pecoraro.

* UserInterface/Controllers/BasicBlockAnnotator.js:
Basic blocks sent from the backend include offsets into the original
file, rather than line/column locations. In order to translate to positions
within CodeMirror, we need to calculate the original line and column
for each block.

(WI.BasicBlockAnnotator.prototype.insertAnnotations):
(WI.BasicBlockAnnotator.prototype._calculateBasicBlockPositions.offsetToPosition):
(WI.BasicBlockAnnotator.prototype._calculateBasicBlockPositions):
(WI.BasicBlockAnnotator.prototype._annotateBasicBlockExecutionRanges.):
(WI.BasicBlockAnnotator.prototype._annotateBasicBlockExecutionRanges):
(WI.BasicBlockAnnotator.prototype._highlightTextForBasicBlock):

* UserInterface/Models/SourceCodePosition.js:
(WI.SourceCodePosition.prototype.offsetColumn):

* UserInterface/Views/TextEditor.js:
(WI.TextEditor.prototype.getTextInRange):
(WI.TextEditor.prototype.addStyleToTextRange):
Better encapsulation for CodeMirror positions.

* UserInterface/Workers/Formatter/FormatterUtilities.js:
(get if):
Update String.prototype.lineEndings to support additional line separators.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233820 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[ MacOS WK1 ] Layout Tests in media/media-fragments/ are flaky
https://bugs.webkit.org/show_bug.cgi?id=187557

Unreviewed test gardening.

Patch by Truitt Savell <tsavell@apple.com> on 2018-07-13

* platform/mac-wk1/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233819 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (r231676): watchOS WebKit usually doesn't load in the background
https://bugs.webkit.org/show_bug.cgi?id=187663
<rdar://problem/42181185>

Reviewed by Chris Dumez.

* UIProcess/ios/WebPageProxyIOS.mm:
(WebKit::WebPageProxy::applicationDidEnterBackground):
"Screen lock" is very aggressive on watchOS; we want to do our usual
30 seconds of loading in the background when you drop your wrist,
so disable this power optimization on that platform.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233818 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: Execution highlighting in the frontend should be line/column-based
https://bugs.webkit.org/show_bug.cgi?id=187532
<rdar://problem/42035580>

Reviewed by Joseph Pecoraro.

Source code offsets from Esprima should not be used to calculate ranges
in CodeMirror for expression highlighting.

This also fixes a long standing bug when adjusting for the starting
position of an inline script. Previously the start offset from the script
TextRange was used for this purpose, but the value is often incorrect (see
https://bugs.webkit.org/show_bug.cgi?id=187532#c5). By using the starting
line/column instead, we avoid the problem.

* UserInterface/Models/ScriptSyntaxTree.js:
(WI.ScriptSyntaxTree.prototype.containersOfPosition):
(WI.ScriptSyntaxTree.prototype.containersOfOffset): Deleted.

* UserInterface/Models/SourceCodePosition.js:
(WI.SourceCodePosition.prototype.offsetColumn):

* UserInterface/Views/SourceCodeTextEditor.js:
(WI.SourceCodeTextEditor.prototype.textEditorExecutionHighlightRange.toInlineScriptPosition):
(WI.SourceCodeTextEditor.prototype.textEditorExecutionHighlightRange.fromInlineScriptPosition):
(WI.SourceCodeTextEditor.prototype.textEditorExecutionHighlightRange):
(WI.SourceCodeTextEditor.prototype.textEditorExecutionHighlightRange.convertRangeOffsetsToSourceCodeOffsets): Deleted.

* UserInterface/Views/TextEditor.js:
(WI.TextEditor.prototype._updateExecutionRangeHighlight):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233817 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[iOS] [WK1] Crash inside IOSurfacePool::platformGarbageCollectNow() in WebThread
https://bugs.webkit.org/show_bug.cgi?id=187635
<rdar://problem/34297065>

Reviewed by Simon Fraser.

r167717 added code to trigger a CA commit in the web process via platformGarbageCollectNow() in order to free IOSurface-related memory.
However, that code is also running in the web thread in apps using WebKit1, causing unwanted UIView layout on the web thread.

Fix by not triggering this CA commit if it's called on the web thread.

* platform/graphics/cocoa/IOSurfacePoolCocoa.mm:
(WebCore::IOSurfacePool::platformGarbageCollectNow):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233816 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebResourceLoader may try to send a IPC with a destination ID that is 0
https://bugs.webkit.org/show_bug.cgi?id=187654
<rdar://problem/39265927>

Reviewed by Brady Eidson.

WebResourceLoader may try to send a IPC with a destination ID that is 0 according to release
assertion hits we see in the wild. This can lead to HashMap corruption on recipient side when
trying to lookup a key that is 0.

As per the crash traces, we see it starts from WebLoaderStrategy::internallyFailedLoadTimerFired()
which is likely due to the Network process crashing. WebLoaderStrategy::networkProcessCrashed()
calls scheduleInternallyFailedLoad() on each ResourceLoader and clears the WebResourceLoaders in
m_webResourceLoaders. When a ResourceLoader is cancelled (marked as failed), we clear its identifier
and we normally null out the WebLoaderStrategy's coreLoader. However, if we cannot find the
WebResourceLoader in m_webResourceLoaders (because that map was already cleared) then the
WebResourceLoader's coreLoader may not get nulled out, even if the coreLoader's identifier has
been reset. We have 2 lambdas in WebResourceLoader which keep the WebResourceLoader alive and
try to send IPC and are merely doing a null check on the coreLoader before trying to send the
IPC.

To address the issue, we now clear the WebResourceLoader's coreLoader in
WebLoaderStrategy::networkProcessCrashed(). For robustness, we also check that a coreLoader's
identifier is not 0 before trying to send IPC.

* WebProcess/Network/WebLoaderStrategy.cpp:
(WebKit::WebLoaderStrategy::networkProcessCrashed):
* WebProcess/Network/WebResourceLoader.cpp:
(WebKit::WebResourceLoader::willSendRequest):
(WebKit::WebResourceLoader::didReceiveResponse):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233815 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Dark Mode: document markers are difficult to see
https://bugs.webkit.org/show_bug.cgi?id=187632
<rdar://problem/41099719>

Reviewed by Simon Fraser.

We update the way we draw the document markers for macOS and use more constrasting colors in dark mode.
Paving the way for future improvements, we move the drawLineForDocumentMarker() method from GraphicsContext
to RenderTheme and implement a first version in RenderThemeMac. The circles used for the underline are now
drawn directly with Core Graphics and we no longer use an image resource. To allow both GraphicsContext
and RenderTheme to have different versions of the drawLineForDocumentMarker() method, the DocumentMarkerLineStyle
enum is now an "enum class".

No new test due to webkit.org/b/105616, webkit.org/b/187655 was raised to track the creation of new tests
when it becomes possible again.

* platform/graphics/GraphicsContext.h:
* platform/graphics/GraphicsContextImpl.h:
* platform/graphics/cairo/CairoOperations.cpp:
(WebCore::Cairo::drawLineForDocumentMarker):
* platform/graphics/cairo/CairoOperations.h:
* platform/graphics/cairo/GraphicsContextImplCairo.cpp:
(WebCore::GraphicsContextImplCairo::drawLineForDocumentMarker):
* platform/graphics/cairo/GraphicsContextImplCairo.h:
* platform/graphics/cocoa/GraphicsContextCocoa.mm:
(WebCore::GraphicsContext::drawLineForDocumentMarker):
* platform/graphics/displaylists/DisplayListItems.h:
(WebCore::DisplayList::DrawLineForDocumentMarker::create):
(WebCore::DisplayList::DrawLineForDocumentMarker::DrawLineForDocumentMarker):
* platform/graphics/displaylists/DisplayListRecorder.cpp:
(WebCore::DisplayList::Recorder::drawLineForDocumentMarker):
* platform/graphics/displaylists/DisplayListRecorder.h:
* platform/graphics/nicosia/cairo/NicosiaCairoOperationRecorder.cpp:
(Nicosia::CairoOperationRecorder::drawLineForDocumentMarker):
* platform/graphics/nicosia/cairo/NicosiaCairoOperationRecorder.h:
* platform/graphics/win/GraphicsContextCGWin.cpp:
(WebCore::GraphicsContext::drawLineForDocumentMarker):
* rendering/InlineTextBox.cpp:
(WebCore::InlineTextBox::paintPlatformDocumentMarker): Call drawLineForDocumentMarker() on the RenderTheme on
macOS and on GraphicsContext in all other cases.
* rendering/RenderTheme.cpp:
(WebCore::RenderTheme::drawLineForDocumentMarker):
* rendering/RenderTheme.h:
* rendering/RenderThemeMac.h:
* rendering/RenderThemeMac.mm:
(WebCore::colorForStyle): Provide different colors for light and dark modes.
(WebCore::RenderThemeMac::drawLineForDocumentMarker): A new macOS-specific version of drawLineForDocumentMarker()
where we paint circles using Core Graphics directly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233814 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[ WK2 ] Layout Test editing/selection/update-selection-by-style-change.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=187649

Unreviewed test gardening.

Patch by Truitt Savell <tsavell@apple.com> on 2018-07-13

* platform/mac-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233813 268f45cc-cd09-0410-ab3c-d52691b4dbfc

libwebrtc.dylib Objective-C classes conflict with third-party frameworks
<https://webkit.org/b/187653>

Reviewed by Alex Christensen.

* Source/webrtc/sdk/WebKit/WebKitUtilities.mm:
- Manually add an attribute to change the class name.

* Source/webrtc/sdk/objc/Framework/Classes/Common/RTCUIApplicationStatusObserver.h:
* Source/webrtc/sdk/objc/Framework/Classes/Metal/RTCMTLI420Renderer.h:
* Source/webrtc/sdk/objc/Framework/Classes/Metal/RTCMTLNV12Renderer.h:
* Source/webrtc/sdk/objc/Framework/Classes/Metal/RTCMTLRenderer.h:
* Source/webrtc/sdk/objc/Framework/Classes/PeerConnection/RTCDtmfSender+Private.h:
* Source/webrtc/sdk/objc/Framework/Classes/PeerConnection/RTCVideoRendererAdapter.h:
* Source/webrtc/sdk/objc/Framework/Classes/PeerConnection/RTCWrappedNativeVideoDecoder.h:
* Source/webrtc/sdk/objc/Framework/Classes/PeerConnection/RTCWrappedNativeVideoEncoder.h:
* Source/webrtc/sdk/objc/Framework/Classes/UI/RTCEAGLVideoView.m:
* Source/webrtc/sdk/objc/Framework/Classes/Video/RTCAVFoundationVideoCapturerInternal.h:
* Source/webrtc/sdk/objc/Framework/Classes/Video/RTCDefaultShader.h:
* Source/webrtc/sdk/objc/Framework/Classes/Video/RTCI420TextureCache.h:
* Source/webrtc/sdk/objc/Framework/Classes/Video/RTCNV12TextureCache.h:
* Source/webrtc/sdk/objc/Framework/Classes/Video/objc_frame_buffer.h:
* Source/webrtc/sdk/objc/Framework/Classes/VideoToolbox/objc_video_decoder_factory.h:
* Source/webrtc/sdk/objc/Framework/Classes/VideoToolbox/objc_video_encoder_factory.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAVFoundationVideoSource.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAudioSession.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAudioSessionConfiguration.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAudioSource.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCAudioTrack.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCCameraPreviewView.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCCameraVideoCapturer.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCConfiguration.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCDataChannel.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCDataChannelConfiguration.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCDispatcher.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCDtmfSender.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCEAGLVideoView.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCFileLogger.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCFileVideoCapturer.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCIceCandidate.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCIceServer.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCIntervalRange.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCLegacyStatsReport.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCMTLNSVideoView.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCMTLVideoView.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCMediaConstraints.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCMediaSource.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCMediaStream.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCMediaStreamTrack.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCMetricsSampleInfo.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCNSGLVideoView.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCPeerConnection.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCPeerConnectionFactory.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCPeerConnectionFactoryOptions.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCRtpCodecParameters.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCRtpEncodingParameters.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCRtpParameters.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCRtpReceiver.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCRtpSender.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCSessionDescription.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoCapturer.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoCodec.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoCodecFactory.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoCodecH264.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoDecoderVP8.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoDecoderVP9.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoEncoderVP8.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoEncoderVP9.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoFrame.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoFrameBuffer.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoRenderer.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoSource.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoTrack.h:
* Source/webrtc/sdk/objc/Framework/Headers/WebRTC/RTCVideoViewShading.h:
- Apply two shell scripts (see bug) to add an attribute to
  change the name of all classes and protocols.

* WebKit/0012-Add-WK-prefix-to-Objective-C-classes-and-protocols.patch: Add.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233812 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Allow BOCOM and ABC plug-ins to run unsandboxed
https://bugs.webkit.org/show_bug.cgi?id=187652
rdar://problem/42149182

Patch by Zach Li <zachli@apple.com> on 2018-07-13
Reviewed by Youenn Fablet.

* UIProcess/Plugins/mac/PluginInfoStoreMac.mm:
(WebKit::PluginInfoStore::shouldAllowPluginToRunUnsandboxed):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233811 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (r233155): Remove last references to click_annotate.cc and rtpcat.cc

* libwebrtc.xcodeproj/project.pbxproj: Let Xcode have its way
with the project file by removing orphaned entries.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233810 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (r222476): Add missing semi-colons to EXPORTED_SYMBOLS_FILE variables

* Configurations/libwebrtc.xcconfig:
(EXPORTED_SYMBOLS_FILE): Add missing semi-colons.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233809 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add more threading release assertions
https://bugs.webkit.org/show_bug.cgi?id=187647

Reviewed by Alex Christensen.

Add more threading release assertions to help debug <rdar://problem/39265927>.

* NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::NetworkConnectionToWebProcess):
(WebKit::NetworkConnectionToWebProcess::~NetworkConnectionToWebProcess):
* UIProcess/WebProcessProxy.cpp:
(WebKit::m_isInPrewarmedPool):
(WebKit::WebProcessProxy::~WebProcessProxy):
(WebKit::WebProcessProxy::shutDown):
(WebKit::WebProcessProxy::didFinishLaunching):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233808 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[WinCairo] Add windows storage process connection implementation
https://bugs.webkit.org/show_bug.cgi?id=187531

Reviewed by Fujii Hironori.

* NetworkProcess/NetworkProcess.cpp:
* Platform/IPC/Attachment.h:
* StorageProcess/StorageProcess.cpp:
* UIProcess/Storage/StorageProcessProxy.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233807 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[32bit JSC tests]  stress/cow-convert-double-to-contiguous.js and stress/cow-convert-int32-to-contiguous.js are failing
https://bugs.webkit.org/show_bug.cgi?id=187561

Reviewed by Darin Adler.

This patch fixes the issue that CoW array handling is not introduced in 32bit put_by_val code.
We clean up 32bit put_by_val code.

1. We remove inline out-of-bounds recording code since it is done in C operation code. This change
aligns 32bit implementation to 64bit implementation.

2. We add CoW array checking, which is done in 64bit implementation.

* jit/JITPropertyAccess.cpp:
(JSC::JIT::emit_op_put_by_val):
* jit/JITPropertyAccess32_64.cpp:
(JSC::JIT::emit_op_put_by_val):
(JSC::JIT::emitSlow_op_put_by_val):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233806 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GStreamer] Use smart pointers for GstByteReader
https://bugs.webkit.org/show_bug.cgi?id=187638

Reviewed by Xabier Rodriguez-Calvar.

* platform/graphics/gstreamer/GUniquePtrGStreamer.h: Add
specialisation for GstByteReader.
* platform/graphics/gstreamer/eme/WebKitClearKeyDecryptorGStreamer.cpp:
Use the new smart pointer class to avoid needing to remember where
to call gst_byte_reader_free.
(webKitMediaClearKeyDecryptorDecrypt):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233805 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[macOS] REGRESSION (r233536): Development WebContent service got opted back into Library Validation
https://bugs.webkit.org/show_bug.cgi?id=187640

Reviewed by Daniel Bates.

* Scripts/process-webcontent-entitlements.sh: Add the
  com.apple.security.cs.disable-library-validation to the Development service regardless of
  whether restricted entitlements are to be used, because that entitlement is not restricted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233804 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Curl] Move HTTP Setup logic from CurlRequest to CurlHandle for reuse.
https://bugs.webkit.org/show_bug.cgi?id=187427

Reviewed by Fujii Hironori.

CurlContext will be used by Secure WebSocket client, but HTTP setup code is
in CurlRequest, which is only for regular HTTP/HTTPS transaction. This patch
allows wss client to setup CurlHandle for HTTPS communication, such as TLS,
proxy or authentication.

No new tests because there's no behavior change.

* platform/network/curl/CurlContext.cpp:
(WebCore::CurlHandle::CurlHandle):
(WebCore::CurlHandle::enableSSLForHost):
(WebCore::CurlHandle::willSetupSslCtx):
(WebCore::CurlHandle::willSetupSslCtxCallback):
(WebCore::CurlHandle::sslErrors const):
(WebCore::CurlHandle::setUrl):
(WebCore::CurlHandle::enableHttp):
(WebCore::CurlHandle::enableConnectionOnly):
(WebCore::CurlHandle::certificateInfo const):
(WebCore::CurlHandle::enableStdErrIfUsed):
(WebCore::CurlHandle::initialize): Deleted.
* platform/network/curl/CurlContext.h:
(WebCore::CurlHandle::url const):
* platform/network/curl/CurlRequest.cpp:
(WebCore::CurlRequest::setupTransfer):
(WebCore::CurlRequest::didReceiveHeader):
(WebCore::CurlRequest::didCompleteTransfer):
(WebCore::CurlRequest::finalizeTransfer):
(WebCore::CurlRequest::willSetupSslCtx): Deleted.
(WebCore::CurlRequest::willSetupSslCtxCallback): Deleted.
* platform/network/curl/CurlRequest.h:
* platform/network/curl/CurlSSLVerifier.cpp:
(WebCore::CurlSSLVerifier::CurlSSLVerifier):
(WebCore::CurlSSLVerifier::verify):
* platform/network/curl/CurlSSLVerifier.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233803 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[ MacOS Debug ] Layout Test inspector/view/asynchronous-layout.html is a Flaky Timeout
https://bugs.webkit.org/show_bug.cgi?id=187622

Unreviewed test gardening.

Patch by Truitt Savell <tsavell@apple.com> on 2018-07-13

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233802 268f45cc-cd09-0410-ab3c-d52691b4dbfc

NetworkConnectionToWebProcess::m_networkResourceLoaders should use Ref<> for its values
https://bugs.webkit.org/show_bug.cgi?id=187629

Reviewed by Youenn Fablet.

NetworkConnectionToWebProcess::m_networkResourceLoaders should use Ref<> for its values
since they cannot be null.

* NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::didClose):
(WebKit::NetworkConnectionToWebProcess::scheduleResourceLoad):
(WebKit::NetworkConnectionToWebProcess::performSynchronousLoad):
* NetworkProcess/NetworkConnectionToWebProcess.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233801 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: SourceCodePosition.js missing from Test.html
https://bugs.webkit.org/show_bug.cgi?id=187644

Reviewed by Brian Burg.

* UserInterface/Test.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233800 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add a FrameLoaderClient willInjectUserScriptForFrame callback
https://bugs.webkit.org/show_bug.cgi?id=187565
<rdar://problem/42095701>

Unreviewed WKTR fix after r233782.

* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233799 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reduce size of WebCore::URL
https://bugs.webkit.org/show_bug.cgi?id=186820
<rdar://problem/42087508>

Unreviewed, re-generates Service worker registrations database used by API tests after r233789.

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKitCocoa/SimpleServiceWorkerRegistrations-2.sqlite3: Removed.
* TestWebKitAPI/Tests/WebKitCocoa/SimpleServiceWorkerRegistrations-2.sqlite3-shm: Removed.
* TestWebKitAPI/Tests/WebKitCocoa/SimpleServiceWorkerRegistrations-3.sqlite3: Renamed from Tools/TestWebKitAPI/Tests/WebKitCocoa/SimpleServiceWorkerRegistrations-2.sqlite3-wal.
* TestWebKitAPI/Tests/WebKitCocoa/WebsiteDataStoreCustomPaths.mm:
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233798 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GStreamer][MSE] Add GstFlowCombiner to handle non-linked inactive branches
https://bugs.webkit.org/show_bug.cgi?id=187636

Reviewed by Carlos Garcia Campos.

When we have more than one source buffer, only one will be
rendered and the inactive branch will report linking errors that
we have to deal with.

* platform/graphics/gstreamer/GUniquePtrGStreamer.h: Added GstFlowCombiner.
* platform/graphics/gstreamer/mse/WebKitMediaSourceGStreamer.cpp:
(webkitMediaSrcChain): Combine the flow in the flow combiner.
(webkit_media_src_init): Initialize the flow combiner.
(webKitMediaSrcLinkStreamToSrcPad): Add the proxypad to the
combiner and set the chain function.
* platform/graphics/gstreamer/mse/WebKitMediaSourceGStreamerPrivate.h:
Declare the flow combiner.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233797 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GStreamer] Add GstBufferMapped abstraction
https://bugs.webkit.org/show_bug.cgi?id=187600

Reviewed by Xabier Rodriguez-Calvar.

There is a similar abstraction called `mapGstBuffer` and friends,
which have a slightly different use-case: wanting a buffer that is
mapped for a longer lifetime without have to keep track of the map
infos separately. They could be subsumed by this abstraction, but
everytime they need to write to the buffer, they'd have to remap
the memory blocks.

This abstraction is more for one-short reads and writes saving the user
from remembering to unmap the buffer and having to manage to
auxiliary GstMapInfo structures.

* platform/graphics/gstreamer/GStreamerCommon.h:
(WebCore::GstMappedBuffer::GstMappedBuffer):
(WebCore::GstMappedBuffer::~GstMappedBuffer):
(WebCore::GstMappedBuffer::data):
(WebCore::GstMappedBuffer::size const):
(WebCore::GstMappedBuffer::operator bool const):
* platform/graphics/gstreamer/InbandTextTrackPrivateGStreamer.cpp:
(WebCore::InbandTextTrackPrivateGStreamer::notifyTrackOfSample):
* platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
(WebCore::MediaPlayerPrivateGStreamerBase::handleSyncMessage):
(WebCore::MediaPlayerPrivateGStreamerBase::initializationDataEncountered):
* platform/graphics/gstreamer/eme/WebKitClearKeyDecryptorGStreamer.cpp:
(webKitMediaClearKeyDecryptorSetupCipher):
(webKitMediaClearKeyDecryptorDecrypt):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233796 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Signal.cpp's activeThreads() should only create its ThreadGroup once.
https://bugs.webkit.org/show_bug.cgi?id=187634
<rdar://problem/40662311>

Reviewed by Yusuke Suzuki.

registerThreadForMachExceptionHandling() is relying on the activeThreads()
ThreadGroup being a singleton.

* wtf/threads/Signals.cpp:
(WTF::activeThreads):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233795 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Make it easier to hit the significant rendered text layout milestone on pages with main article elements
https://bugs.webkit.org/show_bug.cgi?id=187578
<rdar://problem/42104637>

Reviewed by Ryosuke Niwa.

Source/WebCore:

Our current heuristics for triggering the significant rendered text layout milestone are very conservative, with
the intention of avoiding false positives. In practice, we can relax some of these constraints when we've
detected the presence of a main article element on the page. (e.g. in New York Times articles). See per-method
changes below for more detail.

Test: RenderingProgressTests.DidRenderSignificantAmountOfText

* dom/Document.cpp:
(WebCore::Document::registerArticleElement):
(WebCore::Document::unregisterArticleElement):
(WebCore::Document::updateMainArticleElementAfterLayout):

As a post layout task, update the main article element by looping through the articles (up to a maximum limit of
10) in search of an article element that is several times larger than the second largest article element.

* dom/Document.h:

Store a set containing the article elements in the document, as well as the current main article on the page.

(WebCore::Document::hasMainArticleElement const):
* html/Element.cpp:
(WebCore::Element::insertedIntoAncestor):
(WebCore::Element::removedFromAncestor):

Keep track of the article elements that exist in the document whenever elements with the article tag are added
to or removed from the document.

* page/FrameView.cpp:

Add new minimum thresholds for firing the significant rendered text milestone when there exists a main article.

(WebCore::FrameView::performPostLayoutTasks):
(WebCore::FrameView::updateSignificantRenderedTextMilestoneIfNeeded):

Take the main article element into consideration when determining whether to fire the significant text
layout milestone.

Tools:

Tweak an existing layout test to additionally check that the significant text layout milestone is fired on a
page with an article element that is clearly the "main" content on the page (i.e. much taller than all other
articles).

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKitCocoa/significant-text-milestone-article.html: Copied from Tools/TestWebKitAPI/Tests/WebKitCocoa/significant-text-milestone.html.
* TestWebKitAPI/Tests/WebKitCocoa/significant-text-milestone.html:
* TestWebKitAPI/Tests/ios/RenderingProgressTests.mm:
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233794 268f45cc-cd09-0410-ab3c-d52691b4dbfc

JavaScript URL gives incorrect result when frame is navigated
https://bugs.webkit.org/show_bug.cgi?id=187203
<rdar://problem/41438443>

Reviewed by David Kilzer.

* loader/SubframeLoader.cpp:
(WebCore::SubframeLoader::requestFrame):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233793 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: Type token positioning should use line/column locations from Esprima instead of offsets
https://bugs.webkit.org/show_bug.cgi?id=187612
<rdar://problem/42131910>

Reviewed by Joseph Pecoraro.

* UserInterface/Controllers/TypeTokenAnnotator.js:
(WI.TypeTokenAnnotator.prototype.insertAnnotations):
(WI.TypeTokenAnnotator.prototype._insertTypeToken):
(WI.TypeTokenAnnotator.prototype._insertToken):
Use line/column locations, instead of offsets, from the AST when calculating
token positions for CodeMirror. Once in CodeMirror's string space, we
can safely convert to/from offsets.

* UserInterface/Models/ScriptSyntaxTree.js:
Retrieve line/column locations for AST nodes, in addition to offsets.
Offsets into the original file are still needed for getting type information
from the profiler in the backend.

(WI.ScriptSyntaxTree):
(WI.ScriptSyntaxTree.prototype.filterByRange):
Filter by positions, which can be safely used from CodeMirror, instead of offsets.

(WI.ScriptSyntaxTree.prototype._createInternalSyntaxTree):
(WI.ScriptSyntaxTree.prototype.filterByRange.filterForNodesInRange): Deleted.

* UserInterface/Models/SourceCodePosition.js:
Add convenience methods for comparing line/column positions, and for
converting to the format expected by CodeMirror. SourceCodePosition could
be made to interoperate with CodeMirror by exposing properties `line`
and `ch`, but making the conversion explicit improves code readability.

(WI.SourceCodePosition.prototype.equals):
(WI.SourceCodePosition.prototype.isBefore):
(WI.SourceCodePosition.prototype.isAfter):
(WI.SourceCodePosition.prototype.isWithin):
(WI.SourceCodePosition.prototype.toCodeMirror):
(WI.SourceCodePosition):

* UserInterface/Views/TextEditor.js:
(WI.TextEditor.prototype.visibleRangePositions):
(WI.TextEditor.prototype.originalPositionToCurrentPosition):
(WI.TextEditor.prototype.currentOffsetToCurrentPosition):
(WI.TextEditor.prototype.currentPositionToCurrentOffset):
(WI.TextEditor.prototype.setInlineWidget):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233792 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Change my status to be a WebKit reviewer.

Unreviewed status update.

* Scripts/webkitpy/common/config/contributors.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233791 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed contributors.json update

Just run "validate-committer-lists -c" to canonicalize the style.

* Scripts/webkitpy/common/config/contributors.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233790 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reduce size of WebCore::URL
https://bugs.webkit.org/show_bug.cgi?id=186820

Reviewed by Yusuke Suzuki and Youenn Fablet.

Source/WebCore:

We were using 32 bits for the length of the port, which is always between 0 and 5 inclusive
because port numbers are missing or between 0 and 65535.  Let's just use 3 bits here.
We were using 32 bits for the length of the scheme, which is usually 3-5 characters and can be
longer for some custom schemes, but I've never seen one more than 20 characters.  If we assume
schemes are always less than 64MB, we can save 8 bytes per URL!

No change in behavior, just less memory use!

To restore the IPC encoding to how it was before r221165, I just encode the string and reparse it.

* platform/URL.cpp:
(WebCore::URL::invalidate):
(WebCore::URL::lastPathComponent const):
(WebCore::URL::port const):
(WebCore::URL::protocolHostAndPort const):
(WebCore::URL::path const):
(WebCore::URL::removePort):
(WebCore::URL::setPort):
(WebCore::URL::setHostAndPort):
(WebCore::URL::setPath):
* platform/URL.h:
(WebCore::URL::encode const):
(WebCore::URL::decode):
(WebCore::URL::hasPath const):
(WebCore::URL::pathStart const):
* platform/URLParser.cpp:
(WebCore::URLParser::copyBaseWindowsDriveLetter):
(WebCore::URLParser::urlLengthUntilPart):
(WebCore::URLParser::copyURLPartsUntil):
(WebCore::URLParser::shouldPopPath):
(WebCore::URLParser::popPath):
(WebCore::URLParser::parse):
(WebCore::URLParser::parsePort):
(WebCore::URLParser::parseHostAndPort):
(WebCore::URLParser::allValuesEqual):
(WebCore::URLParser::internalValuesConsistent):
* workers/service/server/RegistrationDatabase.cpp:
Increment the service worker registration schema version because of the URL encoding change.

Source/WebKit:

* NetworkProcess/cache/NetworkCacheStorage.h:
Increment cache version because of URL encoding change.

Tools:

* TestWebKitAPI/Tests/WebCore/URLParser.cpp:
(TestWebKitAPI::TEST_F):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233789 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Cocoa] Make sure NetworkProcess::createNetworkConnectionToWebProcess() returns a valid mach port
https://bugs.webkit.org/show_bug.cgi?id=187625

Reviewed by Youenn Fablet.

Make sure NetworkProcess::createNetworkConnectionToWebProcess() returns a valid mach port to help
debug <rdar://problem/41995022>.

* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::createNetworkConnectionToWebProcess):
* UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::didCreateNetworkConnectionToWebProcess):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233788 268f45cc-cd09-0410-ab3c-d52691b4dbfc

PingLoad does not need to ref the NetworkConnectionToWebProcess
https://bugs.webkit.org/show_bug.cgi?id=187624

Reviewed by Youenn Fablet.

PingLoad does not need to ref the NetworkConnectionToWebProcess and keep it alive longer than it needs to be.
Instead, ref the IPC::Connection which is lighter weight and gets invalidated when the NetworkConnectionToWebProcess
is destroyed.

* NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::loadPing):
(WebKit::NetworkConnectionToWebProcess::didFinishPingLoad): Deleted.
* NetworkProcess/NetworkConnectionToWebProcess.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233787 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Make process-swap-on-navigation an experimental feature.
<rdar://problem/41995053> and https://bugs.webkit.org/show_bug.cgi?id=187558

Reviewed by Chris Dumez.

Source/WebKit:

* Shared/WebPreferences.yaml:

* UIProcess/API/C/WKPreferences.cpp:
(WKPreferencesSetProcessSwapOnNavigationEnabled):
(WKPreferencesGetProcessSwapOnNavigationEnabled):
* UIProcess/API/C/WKPreferencesRef.h:

* UIProcess/WebPreferences.cpp:
(WebKit::WebPreferences::updateBoolValueForExperimentalFeatureKey):

* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::createWebPage): If the new web page has PSON enabled via WebPreferences,
  absorb that setting into this process pool's configuration.

Tools:

Make sure WKTR doesn't enable ProcessSwapOnNavigation by default simply because it's an experimental feature.

* WebKitTestRunner/TestController.cpp:
(WTR::TestController::createWebViewWithOptions):
(WTR::TestController::resetPreferencesToConsistentValues):

* WebKitTestRunner/TestOptions.h:
(WTR::TestOptions::shouldEnableProcessSwapOnNavigation const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233786 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Simplify code in NetworkConnectionToWebProcess::didReceiveMessage()
https://bugs.webkit.org/show_bug.cgi?id=187621

Reviewed by Youenn Fablet.

Simplify code in NetworkConnectionToWebProcess::didReceiveMessage() by not using HashMap iterators and
calling get() directly.

* NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::didReceiveMessage):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233785 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Make sure WebProcess::ensureNetworkProcessConnection() is always called on the main thread
https://bugs.webkit.org/show_bug.cgi?id=187607

Reviewed by Alex Christensen.

Add release assertion to make sure that ensureNetworkProcessConnection() is always called on the main
thread. Calling it on a background thread would not be safe. It would not be safe because:
1. We check if we have a network process connection and then create one if we don't without any locking.
2. It is not safe to construct or use a NetworkProcessConnection object from a non-main thread

* WebProcess/WebProcess.cpp:
(WebKit::WebProcess::ensureNetworkProcessConnection):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233784 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Assert that the IPC::Connection is valid in Connection::dispatchMessage(Decoder&)
https://bugs.webkit.org/show_bug.cgi?id=187617

Reviewed by Youenn Fablet.

Assert that the IPC::Connection is valid in Connection::dispatchMessage(Decoder&) as
m_client would be stale otherwise.

* Platform/IPC/Connection.cpp:
(IPC::Connection::dispatchMessage):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233783 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add a FrameLoaderClient willInjectUserScriptForFrame callback
https://bugs.webkit.org/show_bug.cgi?id=187565

Reviewed by Alex Christensen.

Source/WebCore:

Test: http/tests/contentextensions/injected-script-callback.html.

* loader/FrameLoaderClient.h:
* page/Frame.cpp:
(WebCore::Frame::injectUserScriptImmediately):
Calling the new callback whenever being about to inject a new script.

Source/WebKit:

Introduce a new WKBundlePageLoaderClient callback that is called everytime a user script is injected.
Implement WebFrameLoaderClient::willInjectUserScript by calling this new callback.

* WebProcess/InjectedBundle/API/APIInjectedBundlePageLoaderClient.h:
(API::InjectedBundle::PageLoaderClient::willInjectUserScriptForFrame):
* WebProcess/InjectedBundle/API/c/WKBundlePageLoaderClient.h:
* WebProcess/InjectedBundle/API/mac/WKWebProcessPlugInBrowserContextController.mm:
(setUpPageLoaderClient):
* WebProcess/InjectedBundle/InjectedBundlePageLoaderClient.cpp:
(WebKit::InjectedBundlePageLoaderClient::globalObjectIsAvailableForFrame):
(WebKit::InjectedBundlePageLoaderClient::willInjectUserScriptForFrame):
* WebProcess/InjectedBundle/InjectedBundlePageLoaderClient.h:
* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::willInjectUserScript):
* WebProcess/WebCoreSupport/WebFrameLoaderClient.h:

Tools:

Added new test runner API to check for willInjectUserScriptForFrame callback.

* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
* WebKitTestRunner/InjectedBundle/InjectedBundle.h:
(WTR::InjectedBundle::resetUserScriptInjectedCount):
(WTR::InjectedBundle::increaseUserScriptInjectedCount):
(WTR::InjectedBundle::userScriptInjectedCount const):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
(WTR::InjectedBundlePage::InjectedBundlePage):
(WTR::InjectedBundlePage::resetAfterTest):
(WTR::InjectedBundlePage::willInjectUserScriptForFrame):
* WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::userScriptInjectedCount const):
(WTR::TestRunner::injectUserScript):
* WebKitTestRunner/InjectedBundle/TestRunner.h:
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::injectUserScript):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
* WebKitTestRunner/cocoa/TestControllerCocoa.mm:
(WTR::TestController::injectUserScript):

LayoutTests:

* http/tests/contentextensions/injected-script-callback-expected.txt: Added.
* http/tests/contentextensions/injected-script-callback.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233782 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[iOS] When bringing MobileSafari to the foreground, images are drawn asynchronously after removing a snapshot that included them
https://bugs.webkit.org/show_bug.cgi?id=187374
<rdar://problem/41249545>

Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2018-07-12
Reviewed by Tim Horton.

UIProcess should block committing all the layer tree transactions for
immediate update until it receives a one whose activityStateChangeID is
greater than or equal to the one it sends with SetActivityState message.

* Scripts/webkit/messages.py:
Fix the messages code generator to include DrawingAreaInfo.h for the WebKit
typedef ActivityStateChangeID.

* Shared/DrawingAreaInfo.h:
Define ActivityStateChangeID to be uint64_t. Add a definition for the case
when the UIProcess won't be blocked for a reply back from the WebProcess.

* Shared/RemoteLayerTree/RemoteLayerTreeTransaction.h:
(WebKit::RemoteLayerTreeTransaction::activityStateChangeID const):
(WebKit::RemoteLayerTreeTransaction::setActivityStateChangeID):
Add a member for activityStateChangeID in the LayerTreeTransaction.

* Shared/RemoteLayerTree/RemoteLayerTreeTransaction.mm:
(WebKit::RemoteLayerTreeTransaction::encode const):
(WebKit::RemoteLayerTreeTransaction::decode):
Make LayerTreeTransaction know how to encode and decode activityStateChangeID.

* UIProcess/DrawingAreaProxy.h:
(WebKit::DrawingAreaProxy::waitForDidUpdateActivityState):
* UIProcess/RemoteLayerTree/RemoteLayerTreeDrawingAreaProxy.h:
* UIProcess/RemoteLayerTree/RemoteLayerTreeDrawingAreaProxy.mm:
(WebKit::RemoteLayerTreeDrawingAreaProxy::commitLayerTree):
(WebKit::RemoteLayerTreeDrawingAreaProxy::waitForDidUpdateActivityState):
Make the DrawingArea in the UIProcess wait the layer tree with a certain
activityStateChangeID.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::dispatchActivityStateChange):
(WebKit::WebPageProxy::waitForDidUpdateActivityState):
If m_activityStateChangeWantsSynchronousReply is true, generate a new
activityStateChangeID and send it in the SetActivityState message.

* UIProcess/WebPageProxy.h:
(WebKit::WebPageProxy::takeNextActivityStateChangeID):
A simple way to generate a new activityStateChangeID.

* UIProcess/mac/TiledCoreAnimationDrawingAreaProxy.h:
* UIProcess/mac/TiledCoreAnimationDrawingAreaProxy.mm:
(WebKit::TiledCoreAnimationDrawingAreaProxy::waitForDidUpdateActivityState):
Fix the UIProcess functions' prototype for Mac.

* WebProcess/WebPage/AcceleratedDrawingArea.cpp:
(WebKit::AcceleratedDrawingArea::activityStateDidChange):
* WebProcess/WebPage/AcceleratedDrawingArea.h:
Fix the UIProcess functions' prototype for GTK.

* WebProcess/WebPage/DrawingArea.h:
(WebKit::DrawingArea::activityStateDidChange):
* WebProcess/WebPage/RemoteLayerTree/RemoteLayerTreeDrawingArea.h:
Replace wantsDidUpdateActivityState with activityStateChangeID. Treat
activityStateChangeID == ActivityStateChangeAsynchronous as if
wantsDidUpdateActivityState == false.

* WebProcess/WebPage/RemoteLayerTree/RemoteLayerTreeDrawingArea.mm:
(WebKit::RemoteLayerTreeDrawingArea::flushLayers):
Make the WebPrcess pass the activityStateChangeID which it receives from
the SetActivityState message to the LayerTreeTransaction.

(WebKit::RemoteLayerTreeDrawingArea::activityStateDidChange):
When receiving the SetActivityState, treat activityStateChangeID !=
ActivityStateChangeAsynchronous as if wantsDidUpdateActivityState == true.

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::reinitializeWebPage):
(WebKit::WebPage::setActivityState):
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:
Replace the boolean wantsDidUpdateActivityState with activityStateChangeID
in the SetActivityState message.

* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.h:
* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::TiledCoreAnimationDrawingArea):
(WebKit::TiledCoreAnimationDrawingArea::activityStateDidChange):
(WebKit::TiledCoreAnimationDrawingArea::didUpdateActivityStateTimerFired):
Fix the WebProcess functions' prototype for Mac.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233781 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (r232356): After zooming a page in and out, it's possible to temporarily have missing tiles (google.com, apple.com)
https://bugs.webkit.org/show_bug.cgi?id=187553
<rdar://problem/41863803>

Reviewed by Simon Fraser.

We ignore zoom scale from UI process if it doesn't match the last remote layer tree transaction. However nothing
guarantees that we receive the real scale again until user interacts with the page. This means the web process
scale and the UI process scale can be wildly out of sync.

This problem becomes more likely to be visible when
- we are under memory pressure so we don't update tiles during zooming and rely on the one final update afterwards
- the page generates lots of layer tree transactions

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::willCommitLayerTree):

Remember the last transaction where we actually changed the page scale.

* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::scaleFromUIProcess const):

Only reject the UI process scale if there has been a scale changing transaction meanwhile.
The transaction id test was added in r218149 and the problem it fixed remains fixed. This change also matches better
what it was purpoted to implement.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233780 268f45cc-cd09-0410-ab3c-d52691b4dbfc

UTF-16 XHTML files need svn:mime-type to be handled correctly by Windows SVN
https://bugs.webkit.org/show_bug.cgi?id=187571

Reviewed by Alex Christensen.

`svn propset`s cannot be committed with `webkit-patch land`.
This commit actually does what the previous one claimed to.

* dom/xhtml/level3/core/documentgetinputencoding03.xhtml:
* dom/xhtml/level3/core/entitygetinputencoding02.xhtml:
* dom/xhtml/level3/core/entitygetxmlversion02.xhtml:
* dom/xhtml/level3/core/external_foo.ent:
Added property svn:mime-type and removed property svn:eol-style (to align with documentgetxmlencoding03.xhtml).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233779 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Keep Selections within Shadow DOM boundaries
https://bugs.webkit.org/show_bug.cgi?id=187556
<rdar://problem/41664567>

Reviewed by Ryosuke Niwa.

Source/WebCore:

Expose needed functionality to WebKit to help with determing editing and shadow dom boundries.

Only exposing functionality to WebKit.

* dom/TreeScope.h:
* editing/Editing.h:
* editing/VisibleSelection.cpp:
(WebCore::VisibleSelection::adjustPositionForEnd const):
(WebCore::VisibleSelection::adjustPositionForStart const):
(WebCore::adjustPositionForEnd): Deleted.
(WebCore::adjustPositionForStart): Deleted.
* editing/VisibleSelection.h:

Source/WebKit:

Update rangeForPosition to take into account both editing and shadow DOM boundries.

* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::rangeForPoint):
(WebKit::WebPage::updateSelectionWithTouches):
(WebKit::rangeForPosition): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233778 268f45cc-cd09-0410-ab3c-d52691b4dbfc

IndexedDB: database file of subframe cannot be removed
https://bugs.webkit.org/show_bug.cgi?id=187564

Reviewed by Youenn Fablet.

Source/WebCore:

For IndexedDB, if openingOrigin is different from mainFrameOrigin, there will be another
layer in the database file path.
IDBServer should delete database files recursively to make sure all files are removed.

Covered by modified test: WebKit.WebsiteDataStoreCustomPaths.

* Modules/indexeddb/server/IDBServer.cpp:
(WebCore::IDBServer::removeAllDatabasesForOriginPath):

Tools:

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKitCocoa/WebsiteDataStoreCustomPaths.mm:
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233777 268f45cc-cd09-0410-ab3c-d52691b4dbfc

UTF-16 XHTML files need svn:mime-type to be handled correctly by Windows SVN
https://bugs.webkit.org/show_bug.cgi?id=187571

Reviewed by Alex Christensen.

* dom/xhtml/level3/core/documentgetinputencoding03.xhtml:
* dom/xhtml/level3/core/entitygetinputencoding02.xhtml:
* dom/xhtml/level3/core/entitygetxmlversion02.xhtml:
* dom/xhtml/level3/core/external_foo.ent:
Added property svn:mime-type and removed property svn:eol-style (to align with documentgetxmlencoding03.xhtml).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233776 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Make WebKit.LinkColorWithSystemAppearance work across system versions.
https://bugs.webkit.org/show_bug.cgi?id=187573

Reviewed by Wenson Hsieh.

* TestWebKitAPI/Tests/WebKitCocoa/SystemColors.mm:
(TestWebKitAPI.WebKit.LinkColorWithSystemAppearance): Dynamically generate the expected string.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233775 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Revert: iOS port should define HAVE_RUNLOOP_TIMER
<https://webkit.org/b/187370>

* wtf/Platform.h: This is only needed on macOS when using
-[WebView scheduleInRunLoop:forMode:] and
-[WebView unscheduleFromRunLoop:forMode:] to make WK1 WebViews
load synchronously using -loadHTMLString: or -loadData:.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233774 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Disable IsoHeaps when Gigacage is off
https://bugs.webkit.org/show_bug.cgi?id=187160

Reviewed by Saam Barati.

Relanding change sets 233547 and 233550 with the added fix that Gigacage is also
enabled for DumpRenderTree.

Updated determineMallocFallbackState to base enabling of Iso Heaps on Gigacage
being enabled.  We do this because if Gigacage is disabled, it may be due to lack
of address space.

To work around a compiler issue uncovered by the change above, I added explicit
instantiation of PerThread's static variables.  Defined the same explicit
instantiated static variables with export scope in the new file PerThread.cpp
to eliminate separate variables allocations in each linked framework / library.

* CMakeLists.txt:
* bmalloc.xcodeproj/project.pbxproj:
* bmalloc/IsoTLS.cpp:
(bmalloc::IsoTLS::determineMallocFallbackState):
* bmalloc/PerThread.cpp: Added.
* bmalloc/PerThread.h:
* bmalloc/ProcessCheck.mm:
(bmalloc::gigacageEnabledForProcess):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233773 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Need to handle CodeBlock::replacement() being null.
https://bugs.webkit.org/show_bug.cgi?id=187569
<rdar://problem/41468692>

Reviewed by Saam Barati.

CodeBlock::replacement() may return a nullptr.  Some of our code already checks
for this while others do not.  We should add null checks in all the places that
need it.

* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::hasOptimizedReplacement):
(JSC::CodeBlock::jettison):
(JSC::CodeBlock::numberOfDFGCompiles):
(JSC::CodeBlock::setOptimizationThresholdBasedOnCompilationResult):
* dfg/DFGOperations.cpp:
* dfg/DFGToFTLDeferredCompilationCallback.cpp:
(JSC::DFG::ToFTLDeferredCompilationCallback::compilationDidBecomeReadyAsynchronously):
(JSC::DFG::ToFTLDeferredCompilationCallback::compilationDidComplete):
* jit/JITOperations.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233772 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (r230163): Videos cannot be seen full screen in Complete Anatomy app
https://bugs.webkit.org/show_bug.cgi?id=187527
<rdar://problem/40511527>

Reviewed by Ryosuke Niwa.

Do not enable element fullscreen mode unless apps specifically opt-in. The Fullscreen API is
an experimental feature on iOS, but not on Mac, but we can't simply not return the ExperimentalFeature
object from the list of experimental features, as this keeps Safari from being able to apply a
NSUserDefault value for that feature. Instead, add a property to API::ExperimentalFeature and
_WKExperimentalFeature called "hidden", which signals to clients whether to display the feature
in their UI.

* Scripts/GeneratePreferences.rb:
* Scripts/PreferencesTemplates/WebPreferencesExperimentalFeatures.cpp.erb:
* Shared/WebPreferences.yaml:
* Shared/WebPreferencesDefaultValues.h:
* UIProcess/API/APIExperimentalFeature.cpp:
(API::ExperimentalFeature::create):
(API::ExperimentalFeature::ExperimentalFeature):
* UIProcess/API/APIExperimentalFeature.h:
* UIProcess/API/Cocoa/_WKExperimentalFeature.h:
* UIProcess/API/Cocoa/_WKExperimentalFeature.mm:
(-[_WKExperimentalFeature isHidden]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233771 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Mac] Run Unlock Keychain more reliably
https://bugs.webkit.org/show_bug.cgi?id=187604

Reviewed by Dan Bernstein.

Move unlocking to a separate target, and make it a dependency for all targets that
produce binaries.

This is more than strictly required to fix the build, but this way, we won't need
to remember about keychain when project structure changes. Another consideration that
wasn't initially apparent to me is that is should be possible to build any target,
not just "All", so projects aren't necessarily built in the same order even now.

* WebKit.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233770 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add compile guard for enabling NSRunLoop in the WebContent process.
https://bugs.webkit.org/show_bug.cgi?id=187563

Reviewed by Chris Dumez.

Source/WebCore:

No new tests, no change in behavior.

* platform/mac/EventLoopMac.mm:
(WebCore::EventLoop::cycle):

Source/WebKit:

Also, replace __MAC_OS_X_VERSION_MIN_REQUIRED >= 101400 with ENABLE(WEBPROCESS_WINDOWSERVER_BLOCKING) where appropriate.

* Platform/mac/LayerHostingContext.mm:
(WebKit::LayerHostingContext::createForExternalHostingProcess):
* Shared/ChildProcess.h:
* Shared/mac/ChildProcessMac.mm:
* Shared/mac/HangDetectionDisablerMac.mm:
(WebKit::setClientsMayIgnoreEvents):
* UIProcess/Cocoa/WebProcessPoolCocoa.mm:
(WebKit::WebProcessPool::registerNotificationObservers):
(WebKit::WebProcessPool::unregisterNotificationObservers):
* UIProcess/Launcher/mac/ProcessLauncherMac.mm:
(WebKit::shouldLeakBoost):
* UIProcess/WebProcessPool.h:
* WebProcess/Plugins/PDF/PDFLayerControllerSPI.h:
* WebProcess/Plugins/PDF/PDFPlugin.mm:
(WebKit::PDFPlugin::PDFPlugin):
* WebProcess/WebProcess.h:
* WebProcess/WebProcess.messages.in:
* WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::platformInitializeWebProcess):
(WebKit::WebProcess::stopRunLoop):

Source/WTF:

* wtf/FeatureDefines.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233769 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed GTK+ and WPE gardening after r233765.

* platform/gtk/TestExpectations:
* platform/wpe/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233768 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Newly added float should trigger full layout on the block.
https://bugs.webkit.org/show_bug.cgi?id=187251
<rdar://problem/41726137>

Reviewed by David Kilzer.

Source/WebCore:

RenderBlockFlow::determineStartPosition() is one of the places where we decide the extent of the line layout for the current block.
In here we try to figure out the first line in the block that requires layout. In certain cases when floats are present,
(due to their intrusive behavior) we just trigger a full layout on the entire block.
One of the special cases is when a new float is added to the block. determineStartPosition() checks for such floats (floats inserted
after the "last known float") and marks the block for full layout. However it missed the case when other, unrelated mutations happened
in addition to this newly inserted float. This patch fixes this case by checking if the floats after the "last know float" actually need layout.

Test: fast/inline/new-float-needs-layout-when-line-is-dirty.html

* rendering/RenderBlockLineLayout.cpp:
(WebCore::RenderBlockFlow::determineStartPosition):

LayoutTests:

* fast/inline/new-float-needs-layout-when-line-is-dirty-expected.txt: Added.
* fast/inline/new-float-needs-layout-when-line-is-dirty.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233767 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GStreamer] Add pads to the GstFlowCombiner in MediaStreamSrc
https://bugs.webkit.org/show_bug.cgi?id=187552

Patch by Thibault Saunier <tsaunier@igalia.com> on 2018-07-12
Reviewed by Alejandro G. Castro.

This was overlooked and it is the way the API is supposed to be used.

* platform/mediastream/gstreamer/GStreamerMediaStreamSource.cpp:
(WebCore::webkitMediaStreamSrcAddPad):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233766 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[JSC] Thread VM& to JSCell::methodTable(VM&)
https://bugs.webkit.org/show_bug.cgi?id=187548

Reviewed by Saam Barati.

Source/JavaScriptCore:

This patch threads VM& to methodTable(VM&) and remove methodTable().
We add VM& parameter to estimatedSize() to thread VM& in estimatedSize implementations.

* API/APICast.h:
(toJS):
* API/JSCallbackObject.h:
* API/JSCallbackObjectFunctions.h:
(JSC::JSCallbackObject<Parent>::className):
* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::estimatedSize):
* bytecode/CodeBlock.h:
* bytecode/UnlinkedCodeBlock.cpp:
(JSC::UnlinkedCodeBlock::estimatedSize):
* bytecode/UnlinkedCodeBlock.h:
* debugger/DebuggerScope.cpp:
(JSC::DebuggerScope::className):
* debugger/DebuggerScope.h:
* heap/Heap.cpp:
(JSC::GatherHeapSnapshotData::GatherHeapSnapshotData):
(JSC::GatherHeapSnapshotData::operator() const):
(JSC::Heap::gatherExtraHeapSnapshotData):
* heap/HeapSnapshotBuilder.cpp:
(JSC::HeapSnapshotBuilder::json):
* runtime/ArrayPrototype.cpp:
(JSC::arrayProtoFuncToString):
* runtime/ClassInfo.h:
* runtime/DirectArguments.cpp:
(JSC::DirectArguments::estimatedSize):
* runtime/DirectArguments.h:
* runtime/HashMapImpl.cpp:
(JSC::HashMapImpl<HashMapBucket>::estimatedSize):
* runtime/HashMapImpl.h:
* runtime/JSArrayBuffer.cpp:
(JSC::JSArrayBuffer::estimatedSize):
* runtime/JSArrayBuffer.h:
* runtime/JSBigInt.cpp:
(JSC::JSBigInt::estimatedSize):
* runtime/JSBigInt.h:
* runtime/JSCell.cpp:
(JSC::JSCell::dump const):
(JSC::JSCell::estimatedSizeInBytes const):
(JSC::JSCell::estimatedSize):
(JSC::JSCell::className):
* runtime/JSCell.h:
* runtime/JSCellInlines.h:
* runtime/JSGenericTypedArrayView.h:
* runtime/JSGenericTypedArrayViewInlines.h:
(JSC::JSGenericTypedArrayView<Adaptor>::estimatedSize):
* runtime/JSObject.cpp:
(JSC::JSObject::estimatedSize):
(JSC::JSObject::className):
(JSC::JSObject::toStringName):
(JSC::JSObject::calculatedClassName):
* runtime/JSObject.h:
* runtime/JSProxy.cpp:
(JSC::JSProxy::className):
* runtime/JSProxy.h:
* runtime/JSString.cpp:
(JSC::JSString::estimatedSize):
* runtime/JSString.h:
* runtime/RegExp.cpp:
(JSC::RegExp::estimatedSize):
* runtime/RegExp.h:
* runtime/WeakMapImpl.cpp:
(JSC::WeakMapImpl<WeakMapBucket>::estimatedSize):
* runtime/WeakMapImpl.h:

Source/WebCore:

* bindings/js/JSDOMConstructorBase.h:
(WebCore::JSDOMConstructorBase::className):
* bindings/js/JSPluginElementFunctions.cpp:
(WebCore::pluginElementCustomGetCallData):
* bindings/scripts/CodeGeneratorJS.pm:
(GenerateHeader):
(GenerateImplementation):
* bindings/scripts/test/JS/JSInterfaceName.cpp:
(WebCore::JSInterfaceName::estimatedSize):
* bindings/scripts/test/JS/JSInterfaceName.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233765 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, add myself as a WebKit committer.

* Scripts/webkitpy/common/config/contributors.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233764 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[WPE] Pass the backend library name as command line parameter to the web process
https://bugs.webkit.org/show_bug.cgi?id=186841

Reviewed by Žan Doberšek.

Source/WebKit:

* UIProcess/Launcher/glib/ProcessLauncherGLib.cpp:
(WebKit::ProcessLauncher::launchProcess): Add --backend-library parameter when launching the web process.
* WebProcess/wpe/WebProcessMainWPE.cpp: Call wpe_loader_init() with the library passed as --backend-library parameter.

Tools:

* MiniBrowser/wpe/main.cpp:
(main): Stop using WPE_BACKEND_LIBRARY env var.
* Scripts/run-wpe-tests:
(WPETestRunner.__init__): Ditto.
* Scripts/webkitpy/port/wpe.py:
(WPEPort.setup_environ_for_server): Ditto.
* wpe/backends/ViewBackend.cpp:
(WPEToolingBackends::ViewBackend::ViewBackend): Call wpe_loader_init() to ensure WPEBackend-fdo is used.
* wpe/jhbuild.modules: Upgrade WPEBackend to version 0.2.0.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233763 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, rolling out r233714.
https://bugs.webkit.org/show_bug.cgi?id=187579

it made tests time out (Requested by pizlo on #webkit).

Reverted changeset:

"Change the reoptimization backoff base to 1.3 from 2"
https://bugs.webkit.org/show_bug.cgi?id=187540
https://trac.webkit.org/changeset/233714

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233762 268f45cc-cd09-0410-ab3c-d52691b4dbfc

MediaDevices should derive from EventTarget in its IDL
https://bugs.webkit.org/show_bug.cgi?id=187575

Reviewed by Chris Dumez.

Source/WebCore:

Test: fast/mediastream/MediaDevices-addEventListener.html

* Modules/mediastream/MediaDevices.idl:

LayoutTests:

* fast/mediastream/MediaDevices-addEventListener-expected.txt: Added.
* fast/mediastream/MediaDevices-addEventListener.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233761 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GLIB] Add API to allow creating variadic functions
https://bugs.webkit.org/show_bug.cgi?id=187517

Reviewed by Michael Catanzaro.

Source/JavaScriptCore:

Add a _variadic alternate method for jsc_class_add_constructor, jsc_class_add_method and
jsc_value_new_function. In that case the callback always receives a GPtrArray of JSCValue.

* API/glib/JSCCallbackFunction.cpp:
(JSC::JSCCallbackFunction::create): Make the parameters optional.
(JSC::JSCCallbackFunction::JSCCallbackFunction): Ditto.
(JSC::JSCCallbackFunction::call): Handle the case of parameters being nullopt by creating a GPtrArray of
JSCValue for the arguments.
(JSC::JSCCallbackFunction::construct): Ditto.
* API/glib/JSCCallbackFunction.h:
* API/glib/JSCClass.cpp:
(jscClassCreateConstructor): Make the parameters optional.
(jsc_class_add_constructor_variadic): Pass nullopt as parameters to jscClassCreateConstructor.
(jscClassAddMethod): Make the parameters optional.
(jsc_class_add_method_variadic): Pass nullopt as parameters to jscClassAddMethod.
* API/glib/JSCClass.h:
* API/glib/JSCValue.cpp:
(jsc_value_object_define_property_accessor): Update now that parameters are optional.
(jscValueFunctionCreate): Make the parameters optional.
(jsc_value_new_function_variadic): Pass nullopt as parameters to jscValueFunctionCreate.
* API/glib/JSCValue.h:
* API/glib/docs/jsc-glib-4.0-sections.txt:

Tools:

Add test cases.

* TestWebKitAPI/Tests/JavaScriptCore/glib/TestJSC.cpp:
(testJSCFunction):
(fooCreateWithFooV):
(multiplyFooV):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233760 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GLIB] Add jsc_context_get_global_object() to GLib API
https://bugs.webkit.org/show_bug.cgi?id=187515

Reviewed by Michael Catanzaro.

Source/JavaScriptCore:

This wasn't exposed because we have convenient methods in JSCContext to get and set properties on the global
object. However, getting the global object could be useful in some cases, for example to give it a well known
name like 'window' in browsers and GJS.

* API/glib/JSCContext.cpp:
(jsc_context_get_global_object):
* API/glib/JSCContext.h:
* API/glib/docs/jsc-glib-4.0-sections.txt:

Tools:

Add test cases.

* TestWebKitAPI/Tests/JavaScriptCore/glib/TestJSC.cpp:
(testJSCGlobalObject):
(main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233759 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GLIB] Handle G_TYPE_STRV in glib API
https://bugs.webkit.org/show_bug.cgi?id=187512

Reviewed by Michael Catanzaro.

Source/JavaScriptCore:

Add jsc_value_new_array_from_strv() and handle G_TYPE_STRV types in function parameters.

* API/glib/JSCContext.cpp:
(jscContextGValueToJSValue):
(jscContextJSValueToGValue):
* API/glib/JSCValue.cpp:
(jsc_value_new_array_from_strv):
* API/glib/JSCValue.h:
* API/glib/docs/jsc-glib-4.0-sections.txt:

Tools:

Add test cases.

* TestWebKitAPI/Tests/JavaScriptCore/glib/TestJSC.cpp:
(testJSCTypes):
(joinFunction):
(testJSCFunction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233758 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[ews-build] EWS should unapply patch and re-run jsc tests when patch fails jsc tests
https://bugs.webkit.org/show_bug.cgi?id=187549

Reviewed by Alexey Proskuryakov.

* BuildSlaveSupport/ews-build/factories.py:
(JSCTestsFactory.__init__):
* BuildSlaveSupport/ews-build/steps.py:
(CompileJSCOnlyToT): Build ToT without patch.
(CompileJSCOnlyToT.doStepIf): Run this step only if patch failed to build.
(RunJavaScriptCoreTests.evaluateCommand): Set patchFailedJSCTests property appropriately.
(ReRunJavaScriptCoreTests): Re-runs JSC tests.
(ReRunJavaScriptCoreTests.doStepIf): Run this step only if patch failed JSC tests.
(ReRunJavaScriptCoreTests.evaluateCommand): Set patchFailedJSCTests property appropriately.
(RunJavaScriptCoreTestsToT): Run JSC tests on ToT without patch.
(RunJavaScriptCoreTestsToT.doStepIf): Run this step only if patch failed JSC tests.
* BuildSlaveSupport/ews-build/steps_unittest.py: Added unit-tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233757 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Update iOS fullscreen alert text
https://bugs.webkit.org/show_bug.cgi?id=187576
rdar://problem/42052284

Reviewed by Ryosuke Niwa.

Source/WebCore:

* English.lproj/Localizable.strings:

Source/WebKit:

* UIProcess/ios/fullscreen/WKFullScreenViewController.mm:
(-[WKFullScreenViewController _showPhishingAlert]): Update text and string keys.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@233756 268f45cc-cd09-0410-ab3c-d52691b4dbfc