Timeline

Nov 21, 2019:

10:46 PM Changeset in webkit [252765] by Ross Kirsling

• 2 edits in trunk/JSTests

Unreviewed test262 gardening following r252680.

• test262/expectations.yaml:

Mark eight test cases as passing.

10:01 PM Changeset in webkit [252764] by mitz@apple.com

• 9 copies
  1 add in releases/Apple/Safari Technology Preview/Safari Technology Preview 96

Added a tag for Safari Technology Preview release 96.

9:41 PM Changeset in webkit [252763] by sbarati@apple.com

• 14 edits
  9 adds in trunk

GetByStatus should not say it took the slow path for multiple identifiers and should have a way to indicate if the StructureStubInfo it saw took the slow path
​https://bugs.webkit.org/show_bug.cgi?id=204435

Reviewed by Tadeu Zagallo.

JSTests:

• microbenchmarks/get-by-val-polymorphic-ic-1.js: Added.

(assert):
(test1.foo):
(test1):

• microbenchmarks/get-by-val-polymorphic-ic-2.js: Added.

(assert):
(test2.foo):
(test2):

• microbenchmarks/get-by-val-polymorphic-ic-3.js: Added.

(assert):
(test3.foo):
(test3.args):
(test3):

• microbenchmarks/get-by-val-polymorphic-ic-4.js: Added.

(assert):
(test4.foo):
(test4.capture):
(test4.args):
(test4):

• microbenchmarks/get-by-val-polymorphic-ic-5.js: Added.

(assert):
(test5.foo):
(test5.capture):
(test5.args):
(test5):

• microbenchmarks/get-by-val-polymorphic-ic-6.js: Added.

(assert):
(test6.foo):
(test6.capture):
(test6.args):
(test6):

• microbenchmarks/get-by-val-polymorphic-ic-7.js: Added.

(assert):
(test7.foo):
(test7):

• microbenchmarks/get-by-val-polymorphic-ic-8.js: Added.

(assert):
(test7.foo):
(test7):

• microbenchmarks/get-by-val-polymorphic-ic-9.js: Added.

(assert):
(test7.foo):
(test7):

Source/JavaScriptCore:

I discovered some issues with get by val ICs when re-running the microbenchmarks
I wrote. I noticed that we were faster when not running with the DFG. The reason
for this is that we were only emitting a get by val IC in the DFG/FTL when we
observe the GetByStatus says it didn't "go to the slow path". The logic in GetByStatus
for building up a variant list was wrong for ICs with multiple identifiers. We have
a consistency check when building up the list to ensure that no two variants have
structure sets which overlap, because we wouldn't know which one to choose. However,
we were accidentally saying two GetByIdVariants overlap when they had different identifiers.
This patch fixes that bug by also doing an identifier comparison check. Two GetByIdVariants
with different identifiers do not overlap.

We also used to say a GetByStatus "goes to the slow path" if any of the cases were an
array-like load. I wrote that code thinking that ArrayProfile would just handle it.
However, sometimes we have a get by val IC that both has string properties and int32 properties.
In these kinds of scenarios, an IC is super profitable. This patch now distinguishes
between a GetByStatus saying "we're a slow path" and if we actually observed the StructureStubInfo
go to the slow path. In the DFG/FTL, we only forgo emitting a get by val IC when observing a
prior StructureStubInfo that went to the slow path.

I also realized are call to StructureStubInfo::considerCaching was wrong for get by val ICs.
We were only considering the Structure in isolation, not the { Structure, Identifier }
pair. For get by val, we need to consider the pair together, since {s1, "a"}
and {s1, "b"} will be two different access cases.

This patch demonstrates that on these microbenchmarks, get by val ICs can
be between 50-200% faster.

• bytecode/GetByIdVariant.cpp:

(JSC::GetByIdVariant::dumpInContext const):

• bytecode/GetByIdVariant.h:

(JSC::GetByIdVariant::overlaps):

• bytecode/GetByStatus.cpp:

(JSC::GetByStatus::GetByStatus):
(JSC::GetByStatus::computeForStubInfoWithoutExitSiteFeedback):
(JSC::GetByStatus::makesCalls const):
(JSC::GetByStatus::slowVersion const):
(JSC::GetByStatus::merge):
(JSC::GetByStatus::dump const):

• bytecode/GetByStatus.h:

(JSC::GetByStatus::GetByStatus):
(JSC::GetByStatus::takesSlowPath const):
(JSC::GetByStatus::observedStructureStubInfoSlowPath const):

• bytecode/ICStatusUtils.h:

(JSC::appendICStatusVariant):

• bytecode/InByIdVariant.h:

(JSC::InByIdVariant::overlaps):

• bytecode/InstanceOfVariant.h:

(JSC::InstanceOfVariant::overlaps):

• bytecode/PutByIdVariant.h:

(JSC::PutByIdVariant::overlaps):

• bytecode/StructureStubInfo.cpp:

(JSC::StructureStubInfo::visitWeakReferences):

• bytecode/StructureStubInfo.h:

(JSC::StructureStubInfo::considerCaching):

• dfg/DFGByteCodeParser.cpp:

(JSC::DFG::ByteCodeParser::parseBlock):

• jit/JITOperations.cpp:

9:17 PM Changeset in webkit [252762] by jer.noble@apple.com

• 26 edits
  4 adds in trunk

Experimental support for HDR media query
​https://bugs.webkit.org/show_bug.cgi?id=204422

Reviewed by Eric Carlson.

Source/WebCore:

Test: fast/media/mq-highdynamicrange-live-update.html

Add support for the dynamic-range: media query.

• SourcesCocoa.txt:
• WebCore.xcodeproj/project.pbxproj:
• css/CSSValueKeywords.in:
• css/MediaFeatureNames.h:
• css/MediaQueryEvaluator.cpp:

(WebCore::dynamicRangeEvaluate):

• css/MediaQueryExpression.cpp:

(WebCore::featureWithValidIdent):

• page/Chrome.cpp:

(WebCore::Chrome::windowScreenDidChange):

• page/Settings.yaml:
• page/SettingsBase.h:
• platform/PlatformScreen.h:
• platform/ScreenProperties.h:

(WebCore::ScreenData::encode const):
(WebCore::ScreenData::decode):

• platform/ios/PlatformScreenIOS.mm:

(WebCore::screenSupportsHighDynamicRange):

• platform/mac/PlatformScreenMac.mm:

(WebCore::collectScreenProperties):
(WebCore::screenSupportsHighDynamicRange):

• testing/InternalSettings.cpp:

(WebCore::InternalSettings::forcedSupportsHighDynamicRangeValue const):
(WebCore::InternalSettings::setForcedSupportsHighDynamicRangeValue):

• testing/InternalSettings.h:
• testing/InternalSettings.idl:

Source/WebCore/PAL:

• PAL.xcodeproj/project.pbxproj:
• pal/PlatformMac.cmake:
• pal/cocoa/MediaToolboxSoftLink.cpp: Renamed from Source/WebCore/platform/cocoa/MediaToolboxSoftLink.h.
• pal/cocoa/MediaToolboxSoftLink.h: Renamed from Source/WebCore/platform/cocoa/MediaToolboxSoftLink.cpp.

Source/WebKit:

Notify the WebContent process of updated screen properties when the screen's
color space changes.

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

(-[WKWindowVisibilityObserver startObserving:]):
(-[WKWindowVisibilityObserver stopObserving:]):
(-[WKWindowVisibilityObserver _screenDidChangeColorSpace:]):
(WebKit::WebViewImpl::screenDidChangeColorSpace):

• WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::screenPropertiesDidChange):

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

(WebKit::WebProcess::setScreenProperties):

LayoutTests:

• fast/media/mq-highdynamicrange-live-update-expected.txt: Added.
• fast/media/mq-highdynamicrange-live-update.html: Added.

9:09 PM Changeset in webkit [252761] by rniwa@webkit.org

• 7 edits in trunk

Scrolling to fragment shouldn't happen as a part of updating style
​https://bugs.webkit.org/show_bug.cgi?id=203982

Reviewed by Simon Fraser.

Source/WebCore:

Don't scroll to the current URL's fragment at the end of resolveStyle. Instead, schedule a task
to scroll to the current URL's fragment when all pending stylesheets have finished loading.

This patch also moves the code which sets a Document's m_gotoAnchorNeededAfterStylesheetsLoadflag
from FrameView to FrameLoader as FrameView shouldn't be relying on the states of pending stylesheets.

• dom/Document.cpp:

(WebCore::Document::resolveStyle): Removed the code to scroll to the current URL's fragment.
(WebCore::Document::didRemoveAllPendingStylesheet): Added a code to schedule a task to scoll.

• loader/FrameLoader.cpp:

(WebCore::FrameLoader::scrollToFragmentWithParentBoundary): Moved the code to trigger the code
in didRemoveAllPendingStylesheet from FrameView.

• page/FrameView.cpp:

(WebCore::FrameView::scrollToFragment):
(WebCore::FrameView::scrollToFragmentInternal): Renamed from scrollToAnchor since this function sets
the current anchor. Also removed the code which defers the scrolling based on pending stylesheets'
states since such a code doesn't belong in FrameView.

• page/FrameView.h:

LayoutTests:

Made an existing test more robust.

• fast/parser/adoption-agency-unload-iframe-4.html: Made the iframe's data URL not cachable.

7:56 PM Changeset in webkit [252760] by mmaxfield@apple.com

• 38 edits
  2 deletes in trunk

Remove font-variant @font-face descriptor
​https://bugs.webkit.org/show_bug.cgi?id=203179

Reviewed by Simon Fraser.

Source/WebCore:

As per ​https://github.com/w3c/csswg-drafts/issues/2531

Deleted relevant tests.

• css/CSSFontFace.cpp:

(WebCore::CSSFontFace::font):
(WebCore::CSSFontFace::setVariantLigatures): Deleted.
(WebCore::CSSFontFace::setVariantPosition): Deleted.
(WebCore::CSSFontFace::setVariantCaps): Deleted.
(WebCore::CSSFontFace::setVariantNumeric): Deleted.
(WebCore::CSSFontFace::setVariantAlternates): Deleted.
(WebCore::CSSFontFace::setVariantEastAsian): Deleted.

• css/CSSFontFace.h:
• css/CSSFontFaceSource.cpp:

(WebCore::CSSFontFaceSource::load):
(WebCore::CSSFontFaceSource::font):

• css/CSSFontFaceSource.h:
• css/CSSFontSelector.cpp:

(WebCore::CSSFontSelector::addFontFaceRule):

• css/FontFace.cpp:

(WebCore::FontFace::create):
(WebCore::FontFace::setVariant): Deleted.
(WebCore::FontFace::variant const): Deleted.

• css/FontFace.h:
• css/FontFace.idl:
• loader/cache/CachedFont.cpp:

(WebCore::CachedFont::createFont):
(WebCore::CachedFont::platformDataFromCustomData):

• loader/cache/CachedFont.h:
• loader/cache/CachedSVGFont.cpp:

(WebCore::CachedSVGFont::createFont):
(WebCore::CachedSVGFont::platformDataFromCustomData):

• loader/cache/CachedSVGFont.h:
• platform/graphics/FontCache.cpp:

(WebCore::FontPlatformDataCacheKey::FontPlatformDataCacheKey):
(WebCore::FontPlatformDataCacheKey::operator== const):
(WebCore::FontPlatformDataCacheKeyHash::hash):
(WebCore::FontCache::getCachedFontPlatformData):
(WebCore::FontCache::fontForFamily):

• platform/graphics/FontCache.h:

(WebCore::FontCache::fontForFamily):
(WebCore::FontCache::getCachedFontPlatformData):
(WebCore::FontCache::createFontPlatformDataForTesting):

• platform/graphics/cocoa/FontCacheCoreText.cpp:

(WebCore::preparePlatformFont):
(WebCore::fontWithFamily):
(WebCore::FontCache::createFontPlatformData):
(WebCore::FontCache::systemFallbackForCharacters):

• platform/graphics/cocoa/FontCacheCoreText.h:
• platform/graphics/cocoa/FontFamilySpecificationCoreText.cpp:

(WebCore::FontFamilySpecificationCoreText::fontRanges const):

• platform/graphics/mac/FontCustomPlatformData.cpp:

(WebCore::FontCustomPlatformData::fontPlatformData):

• platform/graphics/mac/FontCustomPlatformData.h:

LayoutTests:

Delete tests for the removed feature.

• css3/font-variant-font-face-all-expected.html: Deleted.
• css3/font-variant-font-face-all.html: Deleted.
• css3/font-variant-font-face-override-expected.html:
• css3/font-variant-font-face-override.html:
• fast/text/font-face-empty-string-expected.txt:
• fast/text/font-face-empty-string.html:
• fast/text/font-face-javascript-expected.txt:
• fast/text/font-face-javascript.html:

7:01 PM Changeset in webkit [252759] by BJ Burg

• 2 edits in trunk/Source/WebKit

Unreviewed, sort WebKit's project file after changes in r252702 and r252637.

• WebKit.xcodeproj/project.pbxproj:

6:16 PM Changeset in webkit [252758] by mark.lam@apple.com

• 4 edits
  1 add
  1 delete in trunk

Fix broken String.prototype.replace() and replaceAll().
​https://bugs.webkit.org/show_bug.cgi?id=204479
<rdar://problem/57354854>

Reviewed by Ross Kirsling and Yusuke Suzuki.

JSTests:

• ChakraCore.yaml:
• ChakraCore/test/Strings/replace.baseline-jsc: Removed.
• We no longer need this because we've fixed the spec compliance issue in replace().
• stress/string-replaceAll-2.js: Added.

Source/JavaScriptCore:

String.prototype.replace() regressed due to r252683: <​https://trac.webkit.org/r252683>
for webkit.org/b/202471. The patch failed to handle InternalFunctions.

This patch also fixed a spec compliance bug for String.prototype.replace() i.e.
the replaceValue needs to be evaluated before we check if there's a match in the
source string.
Ref: 21.1.3.16-6 at ​https://www.ecma-international.org/ecma-262/10.0/#sec-string.prototype.replace

For String.prototype.replaceAll(), make sure it "behaves just like
String.prototype.replace if searchValue is a global regular expression".
Ref: ​https://github.com/tc39/proposal-string-replaceall

r252683 also made replaceUsingStringSearch() work the same way as
replaceUsingRegExpSearch(). I think this is the wrong trade off to make.
replaceUsingRegExpSearch() expects each search leg to do a RegExp search, which
is inherently expensive. We shouldn't make string searches slower just because
the RegExp search does it that way.

However, at ​https://bugs.webkit.org/show_bug.cgi?id=202471#c22, Ross pointed out
that JetStream 2 results appeared to be neutral. I think we should double check
with a micro-benchmark as well. I'll leave this for a later patch. For now, the
goal of this patch is simply to achieve correctness.
Ref: ​https://bugs.webkit.org/show_bug.cgi?id=204481

• runtime/StringPrototype.cpp:

(JSC::replaceUsingRegExpSearch):
(JSC::replaceUsingStringSearch):

5:32 PM Changeset in webkit [252757] by sihui_liu@apple.com

• 2 edits in trunk/Source/WebKit

Crash in com.apple.WebKit.WebContent at WebKit: WebKit::StorageAreaMap::loadValuesIfNeeded
​https://bugs.webkit.org/show_bug.cgi?id=204459
<rdar://problem/57383446>

Reviewed by Geoffrey Garen.

If m_storageMapID can be null if connect() fails.

• WebProcess/WebStorage/StorageAreaMap.cpp:

(WebKit::StorageAreaMap::setItem):
(WebKit::StorageAreaMap::removeItem):
(WebKit::StorageAreaMap::clear):
(WebKit::StorageAreaMap::loadValuesIfNeeded):

4:54 PM Changeset in webkit [252756] by pvollan@apple.com

• 6 edits in trunk/Source/JavaScriptCore

Fix Win64 compile errors
​https://bugs.webkit.org/show_bug.cgi?id=204471

Reviewed by Brent Fulgham.

Fix warnings being treated as errors.

• bytecode/AccessCase.cpp:

(JSC::AccessCase::requiresIdentifierNameMatch const):
(JSC::AccessCase::requiresInt32PropertyCheck const):
(JSC::AccessCase::needsScratchFPR const):
(JSC::AccessCase::doesCalls const):
(JSC::AccessCase::canReplace const):

• dfg/DFGOSRExit.cpp:

(JSC::DFG::OSRExit::executeOSRExit):

• jit/JITOperations.cpp:

(JSC::profiledAdd):

• jit/Repatch.cpp:

(JSC::appropriateOptimizingGetByFunction):
(JSC::appropriateGetByFunction):

• tools/JSDollarVM.cpp:

(JSC::functionCallWithStackSize):

4:32 PM Changeset in webkit [252755] by commit-queue@webkit.org

• 7 edits in trunk/Source

Ignore the requests to enter/exit video fullscreen/picture-in-picture if the requesting mode is the same as the current mode
​https://bugs.webkit.org/show_bug.cgi?id=204461

Patch by Peng Liu <Peng Liu> on 2019-11-21
Reviewed by Jer Noble.

Source/WebCore:

Change the order of enum VideoPresentationMode and add some log messages regarding presentation mode change.
Those changes make the debugging easier.

Covered by existing test cases.

• html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::enterFullscreen):
(WebCore::HTMLMediaElement::didBecomeFullscreenElement):
(WebCore::HTMLMediaElement::waitForPreparedForInlineThen):
(WebCore::HTMLMediaElement::setVideoFullscreenLayer):

• html/HTMLVideoElement.cpp:

(WebCore::HTMLVideoElement::webkitSetPresentationMode):
(WebCore::HTMLVideoElement::setFullscreenMode):
(WebCore::HTMLVideoElement::fullscreenModeChanged):

• html/HTMLVideoElement.h:
• html/HTMLVideoElement.idl:

Source/WebKit:

In enterVideoFullscreenForVideoElement() and exitVideoFullscreenForVideoElement(),
we can ignore the request if the targeting mode is the same as the current one.

• WebProcess/cocoa/VideoFullscreenManager.mm:

(WebKit::VideoFullscreenManager::enterVideoFullscreenForVideoElement):
(WebKit::VideoFullscreenManager::exitVideoFullscreenForVideoElement):

3:35 PM Changeset in webkit [252754] by ysuzuki@apple.com

• 10 edits
  1 add in trunk

Unreviewed, rolling in again, regression is not caused by it
​https://bugs.webkit.org/show_bug.cgi?id=202471

JSTests:

• stress/string-replaceall.js: Added.

(shouldBe):
(shouldThrowTypeError):
(shouldBe.string_appeared_here.replaceAll.Symbol.match):

Source/JavaScriptCore:

• builtins/BuiltinNames.h:
• builtins/StringPrototype.js:

(replaceAll):

• runtime/StringPrototype.cpp:

(JSC::StringPrototype::finishCreation):
(JSC::jsSpliceSubstringsWithSeparators):
(JSC::replaceUsingStringSearch):
(JSC::replace):
(JSC::stringProtoFuncReplaceUsingStringSearch):
(JSC::stringProtoFuncReplaceAllUsingStringSearch):

Source/WTF:

• wtf/text/StringCommon.h:

(WTF::findCommon):

LayoutTests:

• js/Object-getOwnPropertyNames-expected.txt:
• js/script-tests/Object-getOwnPropertyNames.js:

3:18 PM Changeset in webkit [252753] by commit-queue@webkit.org

• 10 edits
  1 delete in trunk

Unreviewed, rolling out r252683 and r252721.
​https://bugs.webkit.org/show_bug.cgi?id=204475

13% regression in JetStream2/prepack-wtb (Requested by
yusukesuzuki on #webkit).

Reverted changesets:

"Implement String.prototype.replaceAll"
​https://bugs.webkit.org/show_bug.cgi?id=202471
​https://trac.webkit.org/changeset/252683

"Unreviewed, address Darin's feedback on r252683."
​https://trac.webkit.org/changeset/252721

2:42 PM Changeset in webkit [252752] by Alan Coon

• 1 copy in tags/Safari-608.5.1

Tag Safari-608.5.1.

1:47 PM Changeset in webkit [252751] by Justin Fan

• 2 edits in trunk/Source/WebCore

[WebGL] Fix framebufferTexture2D with DEPTH_STENCIL_ATTACHMENT on ANGLE
​https://bugs.webkit.org/show_bug.cgi?id=204433

Reviewed by Dean Jackson.

ANGLE expects requests for DEPTH_STENCIL_ATTACHMENT to not be split into two calls, as previous WebGL 1 backend required.

No new tests. Prevents conformance/extensions/webgl-depth-texture.html regression when run with the ANGLE backend.

• html/canvas/WebGLRenderingContextBase.cpp:

(WebCore::WebGLRenderingContextBase::framebufferTexture2D):

1:42 PM Changeset in webkit [252750] by Alan Bujtas

• 3 edits in trunk/Source/WebCore

[LFC][IFC] Cache trailing trimmable width.
​https://bugs.webkit.org/show_bug.cgi?id=204466
<rdar://problem/57403888>

Reviewed by Antti Koivisto.

Line::trailingTrimmableWidth() is called every time we pass a run to the line breaker because their accumulated widths are part of the "available width context".

• layout/inlineformatting/InlineLine.cpp:

(WebCore::Layout::Line::removeTrailingTrimmableContent):
(WebCore::Layout::Line::appendTextContent):
(WebCore::Layout::Line::appendNonReplacedInlineBox):
(WebCore::Layout::Line::TrimmableContent::append):
(WebCore::Layout::Line::trailingTrimmableWidth const): Deleted.

• layout/inlineformatting/InlineLine.h:

(WebCore::Layout::Line::trailingTrimmableWidth const):
(WebCore::Layout::Line::TrimmableContent::width const):
(WebCore::Layout::Line::TrimmableContent::runs):
(WebCore::Layout::Line::TrimmableContent::isEmpty const):
(WebCore::Layout::Line::TrimmableContent::clear):

1:30 PM Changeset in webkit [252749] by Simon Fraser

• 25 edits
  1 move in trunk/LayoutTests

Flaky Test: compositing/backing/backing-store-attachment-with-rotation.html
​https://bugs.webkit.org/show_bug.cgi?id=204394

Reviewed by Myles C. Maxfield.

Move compositing-overlap-utils.js to compositing/resources and use its layerTreeWithoutTransforms()
function to not dump transforms for the flakey test.

Adjust other tests which use this JS file.

• compositing/backing/backing-store-attachment-with-rotation-expected.txt:
• compositing/backing/backing-store-attachment-with-rotation.html:
• compositing/layer-creation/animation-overlap-with-children.html:
• compositing/layer-creation/mismatched-rotated-transform-animation-overlap.html:
• compositing/layer-creation/mismatched-rotated-transform-transition-overlap.html:
• compositing/layer-creation/mismatched-transform-transition-overlap.html:
• compositing/layer-creation/multiple-keyframes-animation-overlap.html:
• compositing/layer-creation/scale-rotation-animation-overlap.html:
• compositing/layer-creation/scale-rotation-transition-overlap.html:
• compositing/layer-creation/translate-animation-overlap.html:
• compositing/layer-creation/translate-scale-animation-overlap.html:
• compositing/layer-creation/translate-scale-transition-overlap.html:
• compositing/layer-creation/translate-transition-overlap.html:
• compositing/resources/compositing-test-utils.js: Renamed from LayoutTests/compositing/layer-creation/resources/compositing-overlap-utils.js.

(makeDots):
(layerTreeWithoutTransforms):
(dumpLayersWithoutTransforms):

• legacy-animation-engine/compositing/layer-creation/animation-overlap-with-children.html:
• legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-animation-overlap.html:
• legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-transition-overlap.html:
• legacy-animation-engine/compositing/layer-creation/mismatched-transform-transition-overlap.html:
• legacy-animation-engine/compositing/layer-creation/multiple-keyframes-animation-overlap.html:
• legacy-animation-engine/compositing/layer-creation/scale-rotation-animation-overlap.html:
• legacy-animation-engine/compositing/layer-creation/scale-rotation-transition-overlap.html:
• legacy-animation-engine/compositing/layer-creation/translate-animation-overlap.html:
• legacy-animation-engine/compositing/layer-creation/translate-scale-animation-overlap.html:
• legacy-animation-engine/compositing/layer-creation/translate-scale-transition-overlap.html:
• legacy-animation-engine/compositing/layer-creation/translate-transition-overlap.html:

1:26 PM Changeset in webkit [252748] by Chris Fleizach

• 3 edits in trunk/Source/WebCore

AX: support more attributes for AXIsolatedTreeNode
​https://bugs.webkit.org/show_bug.cgi?id=204241
<rdar://problem/57237606>

Reviewed by Zalan Bujtas.

Add more attributes to the accessibility isolated tree model.

• accessibility/isolatedtree/AXIsolatedTreeNode.cpp:

(WebCore::AXIsolatedObject::initializeAttributeData):

• accessibility/isolatedtree/AXIsolatedTreeNode.h:

1:11 PM Changeset in webkit [252747] by Nikita Vasilyev

• 8 edits in trunk

Web Inspector: Outline sRGB-safe areas on P3 color picker
​https://bugs.webkit.org/show_bug.cgi?id=203533
<rdar://problem/56688057>

Reviewed by Brian Burg.

Source/WebInspectorUI:

Visualize the edge of sRGB gamut as a white line.

• Localizations/en.lproj/localizedStrings.js:

• UserInterface/Models/Color.js:

(WI.Color.displayP3toSRGB.multiplyMatrixByVector):
(WI.Color.displayP3toSRGB):
(WI.Color._toLinearLight):
(WI.Color._gammaCorrect):
Use equations from CSS Color Module Level 4.

• UserInterface/Views/ColorSquare.css:

(.color-square):
(.color-square > .crosshair):
Place the crosshair above the sRGB edge line.

(.color-square .svg-root):
(.color-square .srgb-edge):
(.color-square .srgb-label):
(.color-square .srgb-label:hover):
(.color-square .srgb-label:hover + .svg-root > .srgb-edge):
(@media (-webkit-device-pixel-ratio: 1)):
Make the optical weight of the line on non-retina screens the same as on retina screens.

• UserInterface/Views/ColorSquare.js:

(WI.ColorSquare):
(WI.ColorSquare.prototype.set tintedColor):
(WI.ColorSquare.prototype._updateBaseColor):
(WI.ColorSquare.prototype._drawSRGBOutline):

LayoutTests:

Test WI.Color.displayP3toSRGB.

LayoutTests:

• inspector/model/color-expected.txt:
• inspector/model/color.html:

12:27 PM Changeset in webkit [252746] by dbates@webkit.org

• 2 edits in trunk/Source/WebKitLegacy/mac

Remove unneeded code that annotated DOMHTMLElement as conforming to UITextInputTraits protocol
​https://bugs.webkit.org/show_bug.cgi?id=204464

Reviewed by Wenson Hsieh.

DOMHTMLElement never really conformed to the UITextInputTraits protocol. It was annotated as doing so
in order to register with the runtime as part of advertising that it implemented two UITextInputTraits_Private
messages in r222487: -acceptsAutofilledLoginCredentials and -representingPageURL. These messages were
subsequently removed in r222991 (over years ago). So, we do not need to continue advertising UITextInputTraits
conformance.

• DOM/DOMHTMLInputElement.h:

12:06 PM Changeset in webkit [252745] by Devin Rousso

• 7 edits in trunk/Source/WebInspectorUI

Web Inspector: call frames from blackboxed scripts should be visually distinct
​https://bugs.webkit.org/show_bug.cgi?id=204424

Reviewed by Timothy Hatcher.

Seeing "Deferred pause from blackboxed script" in the Pause Reason does indicate that
something in the Call Stack was blackboxed, but it doesn't clarify what was blackboxed.

Given that the debugger effectively ignores blackboxed scripts, we should distinguish call
frames that were ignored so that it's clear how the current pause location was reached.

• UserInterface/Views/CallFrameTreeElement.js:

(WI.CallFrameTreeElement):
(WI.CallFrameTreeElement.prototype.onattach):

• UserInterface/Views/CallFrameTreeElement.css:

(.tree-outline .item.call-frame.blackboxed:not(.selected)): Added.
Make blackboxed call frames 50% opaque when not selected. Add additional explanatory text to
the end of the tooltip when hovering blackboxed call frame tree elements.

• UserInterface/Views/ContextMenuUtilities.js:

(WI.appendContextMenuItemsForSourceCode):

• UserInterface/Views/SourceCodeTreeElement.js:

(WI.SourceCodeTreeElement.prototype._updateToggleBlackboxImageElementState):
Update tooltip/contextmenu text to include "blacboxing" so functionality is clearer.

• UserInterface/Models/SourceCodeLocation.js:

(WI.SourceCodeLocation.prototype.populateLiveDisplayLocationTooltip):
Allow a suffix to be added to the end of the live location.

• Localizations/en.lproj/localizedStrings.js:

11:53 AM Changeset in webkit [252744] by Kate Cheney

• 2 edits in trunk/LayoutTests

REGRESSION (r234440): [Mac WK2 ] Layout Test http/tests/storageAccess/request-and-grant-access-cross-origin-sandboxed-iframe-from-prevalent-domain-with-user-interaction-but-access-from-wrong-frame.html is flaky timeout (194164)
<​https://bugs.webkit.org/show_bug.cgi?id=194164>
<rdar://problem/48011649>

Reviewed by John Wilander.

Test no longer flaky.

• platform/mac-wk2/TestExpectations:

11:50 AM Changeset in webkit [252743] by Wenson Hsieh

• 5 edits in trunk

[iOS] Web process crashes when attempting to request document context in a textarea
​https://bugs.webkit.org/show_bug.cgi?id=204455
<rdar://problem/57383975>

Reviewed by Tim Horton.

Source/WebCore:

Make WebCore::rangesOverlap robust in the case where the ranges' container nodes are in different tree scopes.
Currently, compareBoundaryPoints returns an exception in this scenario, which rangesOverlap does not expect.
As a result, we crash when releasing the return value after boundary point comparison. To fix this, just upgrade
the ownerDocument check in rangesOverlap to check for tree scopes instead, via the
areNodesConnectedInSameTreeScope helper function.

Test: WebKit.DocumentEditingContextSpatialRequestInTextField

• dom/Range.cpp:

(WebCore::rangesOverlap):

Tools:

Add a test to verify that we don't crash (and are still able to retrieve character rects).

• TestWebKitAPI/Tests/WebKitCocoa/DocumentEditingContext.mm:

(-[UIWKDocumentContext markedTextRects]):
(-[UIWKDocumentContext textRects]):

• TestWebKitAPI/ios/UIKitSPI.h:

11:31 AM Changeset in webkit [252742] by Antti Koivisto

• 5 edits in trunk/Source/WebCore

Tighten ScrollingTreeNode tree mutations
​https://bugs.webkit.org/show_bug.cgi?id=204451

Reviewed by Simon Fraser.

• Use plain Vector instead of std::unique_ptr<Vector> for children
• Use Vector<Ref> instead of Vector<RefPtr>
• Remove non-const ScrollingTreeNode::children() accessor
• Add relase assert that all mutations happen during ScrollingTree::commitTreeState

• page/scrolling/ScrollingTree.cpp:

(WebCore::ScrollingTree::commitTreeState):
(WebCore::ScrollingTree::updateTreeFromStateNode):
(WebCore::ScrollingTree::applyLayerPositionsRecursive):
(WebCore::ScrollingTree::notifyRelatedNodesRecursive):

• page/scrolling/ScrollingTree.h:

(WebCore::ScrollingTree::inCommitTreeState const):

• page/scrolling/ScrollingTreeNode.cpp:

(WebCore::ScrollingTreeNode::appendChild):
(WebCore::ScrollingTreeNode::removeChild):
(WebCore::ScrollingTreeNode::removeAllChildren):

Encapsulate mutations to these functions.

(WebCore::ScrollingTreeNode::dump const):
(WebCore::ScrollingTreeNode::scrollingNodeForPoint const):

• page/scrolling/ScrollingTreeNode.h:

(WebCore::ScrollingTreeNode::children const):
(WebCore::ScrollingTreeNode::children): Deleted.

10:53 AM Changeset in webkit [252741] by dino@apple.com

• 5 edits in trunk/Source/WebCore

Implement accelerated video-to-texture upload path for ANGLE backend for WebGL on desktop
​https://bugs.webkit.org/show_bug.cgi?id=200904

Support GPU-accelerated video uploads to WebGL textures with ANGLE.

Uses the IOSurface path, as the CVOpenGLTextureCache seems to have
been disabled at the OS level, even on top-of-tree WebKit without
using ANGLE.

Uses the EGL_ANGLE_iosurface_client_buffer extension to import
IOSurfaces; reuses all of the existing shader and OpenGL code.

Necessary other fixes to GraphicsContext3D's ANGLE backend: supporting
context sharing, and querying the selected EGLConfig.

Covered by WebGL conformance tests. Verified speedup with
"Animating textures in WebGL" MDN article.

Patch by Kenneth Russell <​kbr@chromium.org> on 2019-11-21
Reviewed by Dean Jackson.

• platform/graphics/GraphicsContext3D.h:

(WebCore::GraphicsContext3D::platformDisplay const):
(WebCore::GraphicsContext3D::platformConfig const):

• platform/graphics/cocoa/GraphicsContext3DCocoa.mm:

(WebCore::GraphicsContext3D::GraphicsContext3D):

• platform/graphics/cv/VideoTextureCopierCV.cpp:

(WebCore::VideoTextureCopierCV::initializeContextObjects):
(WebCore::VideoTextureCopierCV::initializeUVContextObjects):
(WebCore::VideoTextureCopierCV::attachIOSurfaceToTexture):
(WebCore::VideoTextureCopierCV::detachIOSurfaceFromTexture):
(WebCore::VideoTextureCopierCV::copyImageToPlatformTexture):
(WebCore::VideoTextureCopierCV::copyVideoTextureToPlatformTexture):

• platform/graphics/cv/VideoTextureCopierCV.h:

9:52 AM Changeset in webkit [252740] by Matt Lewis

• 2 edits in trunk/Tools

Fix triggers for catalina debug and webgl testers.
​https://bugs.webkit.org/show_bug.cgi?id=204456

Reviewed by Jonathan Bedard.

• BuildSlaveSupport/build.webkit.org-config/config.json:

9:39 AM Changeset in webkit [252739] by Jonathan Bedard

• 2 edits in trunk/Tools

build.webkit.org: Forward results database credentials to jsc test runs
​https://bugs.webkit.org/show_bug.cgi?id=204453

Reviewed by Aakash Jain.

• BuildSlaveSupport/build.webkit.org-config/steps.py:

(RunJavaScriptCoreTests.init): Do not log environment with credentials in it.
(RunJavaScriptCoreTests.start): Add results database API key to environment.
(RunLLINTCLoopTests.init): Do not log environment with credentials in it.
(RunLLINTCLoopTests.start): Add results database API key to environment.
(Run32bitJSCTests.init): Do not log environment with credentials in it.
(Run32bitJSCTests.start): Add results database API key to environment.

9:08 AM Changeset in webkit [252738] by Jonathan Bedard

• 3 edits in trunk/Tools

run-webkit-tests: Clean-up simulator processes after tearing down simulators
​https://bugs.webkit.org/show_bug.cgi?id=204416

Reviewed by Simon Fraser.

• Scripts/webkitpy/common/system/executive_mock.py:

(MockExecutive.running_pids): Make function Python 3 compatible.

• Scripts/webkitpy/xcode/simulated_device.py:

(SimulatedDeviceManager.tear_down): Kill any CoreSimulator processes.

9:07 AM Changeset in webkit [252737] by youenn@apple.com

• 5 edits in trunk

Regression (r252660): Layout Test platform/ios/mediastream/audio-muted-in-background-tab.html is failing
​https://bugs.webkit.org/show_bug.cgi?id=204408
<rdar://problem/57364353>

Reviewed by Eric Carlson.

Source/WebCore:

Covered by no longer failing test.

• platform/mediastream/mac/CoreAudioCaptureSource.h:
• platform/mock/MockRealtimeMediaSourceCenter.cpp:

Now that we are using CoreAudioCaptureSource, we need to use its factory to interrupt/uninterrupt the active source.

LayoutTests:

• platform/ios/TestExpectations:

8:50 AM Changeset in webkit [252736] by Antti Koivisto

• 13 edits in trunk/Source/WebCore

MediaQueryEvaluator shouldn't know about style resolver
​https://bugs.webkit.org/show_bug.cgi?id=204449

Reviewed by Zalan Bujtas.

Gather dynamic dependencies from MediaQueryEvaluator into a struct rather than writing them directly to a resolver.

• css/MediaQueryEvaluator.cpp:

(WebCore::MediaQueryEvaluator::evaluate const):

Delete a duplicate version of evaluate() that returned partial dependencies.

• css/MediaQueryEvaluator.h:

(WebCore::MediaQueryDynamicResults::append):

• css/parser/SizesAttributeParser.cpp:

(WebCore::SizesAttributeParser::mediaConditionMatches):

This doesn't need resolver, the picture element implementation takes care of the media query dependencies.

• html/HTMLImageElement.cpp:

(WebCore::HTMLImageElement::bestFitSourceFromPictureElement):

• html/HTMLPictureElement.cpp:

(WebCore::HTMLPictureElement::viewportChangeAffectedPicture const):
(WebCore::HTMLPictureElement::appearanceChangeAffectedPicture const):

Store MediaQueryDynamicResults struct instead of individual vectors.

• style/RuleSet.cpp:

(WebCore::Style::RuleSet::addChildRules):
(WebCore::Style::RuleSet::addRulesFromSheet):

• style/RuleSet.h:
• style/StyleResolver.cpp:

(WebCore::Style::Resolver::appendAuthorStyleSheets):
(WebCore::Style::Resolver::addMediaQueryDynamicResults):
(WebCore::Style::Resolver::hasMediaQueriesAffectedByViewportChange const):
(WebCore::Style::Resolver::hasMediaQueriesAffectedByAccessibilitySettingsChange const):
(WebCore::Style::Resolver::hasMediaQueriesAffectedByAppearanceChange const):
(WebCore::Style::Resolver::addViewportDependentMediaQueryResult): Deleted.
(WebCore::Style::Resolver::addAccessibilitySettingsDependentMediaQueryResult): Deleted.
(WebCore::Style::Resolver::addAppearanceDependentMediaQueryResult): Deleted.

Store MediaQueryDynamicResults struct instead of individual vectors.

• style/StyleResolver.h:

(WebCore::Style::Resolver::hasViewportDependentMediaQueries const):
(WebCore::Style::Resolver::hasAccessibilitySettingsDependentMediaQueries const):
(WebCore::Style::Resolver::hasAppearanceDependentMediaQueries const):

• style/StyleScopeRuleSets.cpp:

(WebCore::Style::ScopeRuleSets::updateUserAgentMediaQueryStyleIfNeeded const):
(WebCore::Style::ScopeRuleSets::initializeUserStyle):
(WebCore::Style::ScopeRuleSets::collectRulesFromUserStyleSheets):
(WebCore::Style::ScopeRuleSets::appendAuthorStyleSheets):

• style/StyleScopeRuleSets.h:

8:32 AM Changeset in webkit [252735] by Alan Bujtas

• 3 edits in trunk/Source/WebCore

[LFC] Add support for LFC layout only.
​https://bugs.webkit.org/show_bug.cgi?id=204441
<rdar://problem/57382929>

Reviewed by Antti Koivisto.

Clean up the render tree state when RenderView::layout is not called.

• page/FrameViewLayoutContext.cpp:

(WebCore::FrameViewLayoutContext::layoutUsingFormattingContext):

• rendering/updating/RenderTreeUpdater.cpp:

(WebCore::RenderTreeUpdater::updateRendererStyle):

8:21 AM Changeset in webkit [252734] by Pablo Saavedra

• 2 edits in trunk/Tools

Fix Tools/Scripts/run-javascriptcore-tests after r252490
​https://bugs.webkit.org/show_bug.cgi?id=204452

Rubber-stamped by Philippe Normand

Regression introduced in:

results.webkit.org: Report JSC tests to the results database
​https://bugs.webkit.org/show_bug.cgi?id=204091
<rdar://problem/49778900>

WebKitGTK and WPE build environment is defined by the dependencies
policy document [1] as a Debian Stable. For Debian the uname binary
is installed in the /bin/ directory and this directory is included
in the $PATH for any regular or system user. This fix removes the
path for the uname command in the run-javascriptcore-tests script.

[1] ​https://trac.webkit.org/wiki/WebKitGTK/DependenciesPolicy

• Scripts/run-javascriptcore-tests:

(configurationForUpload):

7:11 AM Changeset in webkit [252733] by pvollan@apple.com

• 2 edits in trunk/LayoutTests

Update test expectations for flaky tests

Unreviewed test gardening.

• platform/win/TestExpectations:

6:12 AM Changeset in webkit [252732] by Andres Gonzalez

• 5 edits in trunk/Source/WebCore

Support boundingBoxRect and elementRect in AXIsolatedObject.
​https://bugs.webkit.org/show_bug.cgi?id=204431

Reviewed by Chris Fleizach.

• Implementation of boundingBoxRect and elementRect in AXIsolatedObject.
• Templatized rectAttributeValue generalize retrieval of cached rects for different types of rects.

• accessibility/AccessibilityObject.h:
• accessibility/AccessibilityObjectInterface.h:

(WebCore::AXCoreObject::pixelSnappedBoundingBoxRect const): Moved from AccessibilityObject.
(WebCore::AXCoreObject::size const): Moved from AccessibilityObject.

• accessibility/isolatedtree/AXIsolatedTreeNode.cpp:

(WebCore::AXIsolatedObject::initializeAttributeData): Caching of boundingBox and element rects.
(WebCore::AXIsolatedObject::rectAttributeValue const): Templatized version.

• accessibility/isolatedtree/AXIsolatedTreeNode.h:

4:59 AM Changeset in webkit [252731] by commit-queue@webkit.org

• 2 edits in trunk/Source/WTF

add ASSERT_NOT_REACHED_WITH_MESSAGE and RELEASE_ASSERT_NOT_REACHED_WITH_MESSAGE
​https://bugs.webkit.org/show_bug.cgi?id=204445

Patch by Tuomas Karkkainen <​tuomas.webkit@apple.com> on 2019-11-21
Reviewed by Antti Koivisto.

Add assertions that combine ASSERT_NOT_REACHED and ASSERT_WITH_MESSAGE.

• wtf/Assertions.h:

4:14 AM Changeset in webkit [252730] by Devin Rousso

• 5 edits in trunk

Web Inspector: removing the blackbox for a specific script doesn't actually remove the blackbox
​https://bugs.webkit.org/show_bug.cgi?id=204428

Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

Previously, when updating the blackbox state of each existing script, we would only tell the
Debugger about when scripts should be blackboxed, not when they shouldn't. This means that
when a given script is un-blackboxed, the Debugger would never get told about it and would
therefore still defer pauses as if it was blackboxed.

The solution to this is simple; update the blackboxed state of every script, not just those
that should be blackboxed, and tell the Debugger about each.

• inspector/agents/InspectorDebuggerAgent.cpp:

(Inspector::InspectorDebuggerAgent::setShouldBlackboxURL):

LayoutTests:

• inspector/debugger/setShouldBlackboxURL.html:
• inspector/debugger/setShouldBlackboxURL-expected.txt:

2:56 AM WebKitGTK/2.26.x edited by Philippe Normand

(diff)

2:56 AM WebKitGTK/2.26.x edited by Philippe Normand

(diff)

Nov 20, 2019:

11:25 PM Changeset in webkit [252729] by Conrad Shultz

• 2 edits in trunk/Source/WebCore

Crash at com.apple.WebCore: -[WebRevealHighlight revealContext:drawRectsForItem:]
​https://bugs.webkit.org/show_bug.cgi?id=204434

Reviewed by Megan Gardner.

_attributedString was not retained, leading to a potential use-after-dealloc. Fix
this by wrapping it in a RetainPtr. While we're here, remove explicit ivars and
property synthesis and copy the string during assignment.

Reveal functionality is currently not testable.

• editing/cocoa/DictionaryLookup.mm:

(SOFT_LINK_CLASS_OPTIONAL):
(-[WebRevealHighlight initWithHighlightRect:useDefaultHighlight:attributedString:]):

10:02 PM Changeset in webkit [252728] by ysuzuki@apple.com

• 3 edits in trunk/Source/JavaScriptCore

[JSC] Extend MacroAssemblerARM64::load/store for datasize = 16
​https://bugs.webkit.org/show_bug.cgi?id=204442
<rdar://problem/57366761>

Reviewed by Mark Lam.

Our void load16(const void* address, RegisterID dest) and void store16(RegisterID src, const void* address) are not aware of
the condition that passed register can be memoryTempRegister, while MacroAssemblerARM64::{load,store} handles it correctly, e.g.
load invalidates cachedMemoryTempRegister if destination register is memoryTempRegister. As a result, when we are emitting
or16(TrustedImm32 imm, AbsoluteAddress address) with address where the address's value does not fit in imm, the generated code
is reusing memoryTempRegister incorrectly.

0xedf8d4fb4: mov x17, #0x7af0
0xedf8d4fb8: movk x17, #0xd5a, lsl #16
0xedf8d4fbc: movk x17, #0x1, lsl #32 Construct imm register on x17.
0xedf8d4fc0: ldrh w17, [x17] Load half word from x17 to w17 (we should invalidate x17 memoryTempRegister here).
0xedf8d4fc4: mov w16, #0x1b
0xedf8d4fc8: orr w16, w17, w16
0xedf8d4fcc: strh w16, [x17] x17 memoryTempRegister is reused while its content is invalid.

The problem is that load and store functions are not supporting datasize = 16 case. This patch extends MacroAssemblerARM64::{load,store}
to support 16 so that or16 implementation looks is similar to or32 etc.

• assembler/MacroAssemblerARM64.h:

(JSC::MacroAssemblerARM64::load16):
(JSC::MacroAssemblerARM64::store16):
(JSC::MacroAssemblerARM64::load):
(JSC::MacroAssemblerARM64::store):

• assembler/testmasm.cpp:

(JSC::testOrImmMem):

9:12 PM Changeset in webkit [252727] by Jonathan Bedard

• 2 edits in trunk/Tools

run-webkit-tests: Make usability test on booting simulators more robust
​https://bugs.webkit.org/show_bug.cgi?id=204423
<rdar://problem/57364735>

Reviewed by Alexey Proskuryakov.

• Scripts/webkitpy/xcode/simulated_device.py:

(SimulatedDevice.is_usable): Use regular expressions to find home-screen service.

8:34 PM Changeset in webkit [252726] by Fujii Hironori

• 2 edits in trunk/Source/WebCore

[MSVC] error C2039: 'weakPtrFactory': is not a member of 'WebCore::DocumentStorageAccess::requestStorageAccess::<lambda_3f2cfd7704f93d8fe19d5b5f064f8add>'
​https://bugs.webkit.org/show_bug.cgi?id=204437

Unreviewed build fix.

MSVC has a bug of a outer lambda's captured 'this' can't be used as
a lambda capture initializer of a inner lambda.

• dom/DocumentStorageAccess.cpp:

(WebCore::DocumentStorageAccess::requestStorageAccess): Replaced '*this' in the lambda capture initializer with '*weakThis'.

6:00 PM Changeset in webkit [252725] by sbarati@apple.com

• 3 edits in trunk/Source/JavaScriptCore

Baseline JIT should fill in StructureStubInfo's propertyIsInt32 and the slow path should update the array profile more frequently
​https://bugs.webkit.org/show_bug.cgi?id=204432

Reviewed by Tadeu Zagallo.

When I added inline caching for get by val, I removed code which updated the
ArrayProfile with some frequency. This patch adds code that does that back,
which recovers some of the JetStream2 regressions we are seeing.

• jit/JITOperations.cpp:
• jit/JITPropertyAccess.cpp:

(JSC::JIT::emit_op_get_by_val):

5:48 PM Changeset in webkit [252724] by Simon Fraser

• 31 edits in trunk

getComputedStyle returns "auto" for zIndex property even after it has been set, on non-positioned elements
​https://bugs.webkit.org/show_bug.cgi?id=15562

Reviewed by Antti Koivisto.

LayoutTests/imported/w3c:

More passing tests.

• web-platform-tests/css/css-position/inheritance-expected.txt:
• web-platform-tests/css/css-values/calc-integer-expected.txt:
• web-platform-tests/css/css-values/calc-z-index-fractions-001-expected.txt:
• web-platform-tests/css/css-values/minmax-integer-computed-expected.txt:

Source/WebCore:

The computed value of z-index should be the specified value, but Style::Adjuster::adjust() set z-index
to auto on non-positioned elements, so we reported the incorrect computed value.

Fix by storing the specified value in RenderStyle as "zIndex", and the used value as "effectiveZIndex", and
converting all rendering code to use the "effective" variants. getComputedStyle reads "zIndex".

Style::Adjuster::adjust() copies zIndex to effectiveZIndex for positioned elements.

Tests: fast/css/getComputedStyle/getComputedStyle-zIndex-auto.html

• animation/KeyframeEffect.cpp:

(WebCore::KeyframeEffect::apply):

• css/CSSComputedStyleDeclaration.cpp:

(WebCore::ComputedStyleExtractor::valueForPropertyInStyle):

• css/CSSProperties.json:
• html/HTMLInputElement.cpp:

(WebCore::HTMLInputElement::createInnerTextStyle):

• page/animation/CSSPropertyAnimation.cpp:

(WebCore::CSSPropertyAnimationWrapperMap::CSSPropertyAnimationWrapperMap):

• page/animation/CompositeAnimation.cpp:

(WebCore::CompositeAnimation::animate):

• rendering/RenderBox.cpp:

(WebCore::isCandidateForOpaquenessTest):

• rendering/RenderBox.h:
• rendering/RenderElement.cpp:

(WebCore::RenderElement::styleWillChange):

• rendering/RenderFullScreen.cpp:

(WebCore::createFullScreenStyle):

• rendering/RenderLayer.cpp:

(WebCore::canCreateStackingContext):
(WebCore::RenderLayer::shouldBeCSSStackingContext const):
(WebCore::RenderLayer::calculateClipRects const):

• rendering/RenderLayer.h:
• rendering/RenderLayerCompositor.cpp:

(WebCore::RenderLayerCompositor::logLayerInfo):

• rendering/RenderLayerModelObject.cpp:

(WebCore::RenderLayerModelObject::styleWillChange):

• rendering/RenderTreeAsText.cpp:

(WebCore::RenderTreeAsText::writeRenderObject):

• rendering/style/RenderStyle.cpp:

(WebCore::RenderStyle::changeRequiresLayerRepaint const):

• rendering/style/RenderStyle.h:

(WebCore::RenderStyle::specifiedZIndex const):
(WebCore::RenderStyle::hasAutoSpecifiedZIndex const):
(WebCore::RenderStyle::setSpecifiedZIndex):
(WebCore::RenderStyle::setHasAutoSpecifiedZIndex):
(WebCore::RenderStyle::usedZIndex const):
(WebCore::RenderStyle::hasAutoUsedZIndex const):
(WebCore::RenderStyle::setUsedZIndex):
(WebCore::RenderStyle::setHasAutoUsedZIndex):
(WebCore::RenderStyle::hasAutoZIndex const): Deleted.
(WebCore::RenderStyle::setHasAutoZIndex): Deleted.
(WebCore::RenderStyle::zIndex const): Deleted.
(WebCore::RenderStyle::setZIndex): Deleted.

• rendering/style/StyleBoxData.cpp:

(WebCore::StyleBoxData::StyleBoxData):
(WebCore::StyleBoxData::operator== const):

• rendering/style/StyleBoxData.h:

(WebCore::StyleBoxData::specifiedZIndex const):
(WebCore::StyleBoxData::hasAutoSpecifiedZIndex const):
(WebCore::StyleBoxData::usedZIndex const):
(WebCore::StyleBoxData::hasAutoUsedZIndex const):
(WebCore::StyleBoxData::zIndex const): Deleted.
(WebCore::StyleBoxData::hasAutoZIndex const): Deleted.

• style/StyleAdjuster.cpp:

(WebCore::Style::Adjuster::adjust const):

LayoutTests:

• fast/css-grid-layout/grid-item-z-index-support-expected.txt:
• fast/css-grid-layout/grid-item-z-index-support.html:
• fast/css/getComputedStyle/getComputedStyle-zIndex-auto-expected.txt:
• fast/css/getComputedStyle/getComputedStyle-zIndex-auto.html:

5:47 PM Changeset in webkit [252723] by rniwa@webkit.org

• 34 edits in trunk/Source/WebCore

MicrotaskQueue should be accessed via EventLoop
​https://bugs.webkit.org/show_bug.cgi?id=204397

Reviewed by Antti Koivisto.

This patch refactors the existing code so that a microtask is always queued via EventLoopTaskGroup.
It preserves all other (broken) semantics and behavior like all origins sharing a single microtask queue.

The singleton MicrotaskQueue for the main thread has been moved from MicrotaskQueue::mainThreadQueue
to WindowEventLoop, and an instance of MicrotaskQueue for each worker has been moved from WorkerGlobalScope
to WorkerEventLoop.

• animation/DocumentTimeline.cpp:

(WebCore::DocumentTimeline::internalUpdateAnimationsAndSendEvents):

• animation/WebAnimation.cpp:

(WebCore::WebAnimation::updateFinishedState):

• bindings/js/JSDOMGlobalObjectTask.cpp:

(WebCore::JSGlobalObjectTask::JSGlobalObjectTask):

• bindings/js/JSDOMWindowBase.cpp:

(WebCore::JSDOMWindowBase::queueMicrotaskToEventLoop): Renamed from queueTaskToEventLoop.

• bindings/js/JSDOMWindowBase.h:
• bindings/js/JSExecState.cpp:

(WebCore::JSExecState::didLeaveScriptContext):

• bindings/js/JSRemoteDOMWindowBase.cpp:
• bindings/js/JSWorkerGlobalScopeBase.cpp:

(WebCore::JSWorkerGlobalScopeBase::queueMicrotaskToEventLoop): Renamed from queueTaskToEventLoop.

• bindings/js/JSWorkerGlobalScopeBase.h:
• bindings/js/JSWorkletGlobalScopeBase.cpp:
• bindings/js/JSWorkletGlobalScopeBase.h:
• dom/CustomElementReactionQueue.cpp:

(WebCore::BackupElementQueueMicrotask): Deleted.
(WebCore::CustomElementReactionQueue::enqueueElementOnAppropriateElementQueue):
(WebCore::CustomElementReactionQueue::ensureBackupQueue):

• dom/CustomElementReactionQueue.h:
• dom/Document.cpp:

(WebCore::Document::finishedParsing):

• dom/DocumentStorageAccess.cpp:

(WebCore::DocumentStorageAccess::requestStorageAccess):

• dom/EventLoop.cpp:

(WebCore::EventLoop::queueMicrotask): Added.
(WebCore::EventLoop::performMicrotaskCheckpoint): Added.
(WebCore::EventLoopTaskGroup::queueMicrotaskCallback): Added.
(WebCore::EventLoopTaskGroup::queueMicrotask): Added.
(WebCore::EventLoopTaskGroup::performMicrotaskCheckpoint): Added.

• dom/EventLoop.h:

(WebCore::EventLoopTaskGroup::microtaskQueue):

• dom/Microtasks.cpp:

(WebCore::MicrotaskQueue::mainThreadQueue): Deleted.
(WebCore::MicrotaskQueue::contextQueue): Deleted.

• dom/Microtasks.h:
• dom/MutationObserver.cpp:

(WebCore::MutationObserverMicrotask): Deleted.
(WebCore::MutationObserver::queueMutationObserverCompoundMicrotask): Made this a member function
so that it can call notifyMutationObservers in its lambda.
(WebCore::MutationObserver::enqueueMutationRecord):
(WebCore::MutationObserver::enqueueSlotChangeEvent):
(WebCore::MutationObserver::setHasTransientRegistration):

• dom/MutationObserver.h:
• dom/MutationObserverRegistration.cpp:

(WebCore::MutationObserverRegistration::observedSubtreeNodeWillDetach):

• dom/WindowEventLoop.cpp:

(WebCore::WindowEventLoop::microtaskQueue):

• dom/WindowEventLoop.h:
• html/parser/HTMLDocumentParser.cpp:

(WebCore::HTMLDocumentParser::runScriptsForPausedTreeBuilder):

• html/parser/HTMLScriptRunner.cpp:

(WebCore::HTMLScriptRunner::executePendingScriptAndDispatchEvent):
(WebCore::HTMLScriptRunner::runScript):

• inspector/agents/InspectorCanvasAgent.cpp:

(WebCore::InspectorCanvasAgent::recordCanvasAction):

• testing/Internals.cpp:

(WebCore::Internals::queueMicroTask):

• workers/WorkerEventLoop.cpp:

(WebCore::WorkerEventLoop::~WorkerEventLoop):
(WebCore::WorkerEventLoop::microtaskQueue):
(WebCore::WorkerEventLoop::clearMicrotaskQueue):

• workers/WorkerEventLoop.h:
• workers/WorkerGlobalScope.cpp:

(WebCore::WorkerGlobalScope::WorkerGlobalScope):
(WebCore::WorkerGlobalScope::prepareForTermination):

• workers/WorkerGlobalScope.h:

(WebCore::WorkerGlobalScope::microtaskQueue const): Deleted.

5:25 PM Changeset in webkit [252722] by Jonathan Bedard

• 2 edits in trunk/Tools

run-webkit-tests: Do not create global SystemHost objects
​https://bugs.webkit.org/show_bug.cgi?id=204426

Reviewed by Aakash Jain.

If created, SystemHost objects should be created on-demand, not globally
shared between all instances of a function.

• Scripts/webkitpy/xcode/simulated_device.py:

(SimulatedDeviceManager.populate_available_devices):
(SimulatedDeviceManager.available_devices):
(SimulatedDeviceManager.device_by_filter):
(SimulatedDeviceManager._create_or_find_device_for_request):
(SimulatedDeviceManager._boot_device):
(SimulatedDeviceManager.device_count_for_type):
(SimulatedDeviceManager.initialize_devices):
(SimulatedDeviceManager.max_supported_simulators):
(SimulatedDeviceManager.swap):
(SimulatedDeviceManager.tear_down):

5:16 PM Changeset in webkit [252721] by Ross Kirsling

• 6 edits in trunk

Unreviewed, address Darin's feedback on r252683.

JSTests:

• stress/string-replaceall.js:

Source/JavaScriptCore:

• runtime/StringPrototype.cpp:

(JSC::replaceUsingStringSearch):
(JSC::replace):
(JSC::stringProtoFuncReplaceUsingStringSearch):
(JSC::stringProtoFuncReplaceAllUsingStringSearch):

Source/WTF:

• wtf/text/StringCommon.h:

(WTF::findCommon):

5:09 PM Changeset in webkit [252720] by Kate Cheney

• 3 edits in trunk/LayoutTests

[ Jazz ] http/tests/resourceLoadStatistics/cookie-deletion.html is timing out
​https://bugs.webkit.org/show_bug.cgi?id=203813
<rdar://problem/54316765>

Reviewed by John Wilander.

Added random dummy value to the end of the cookie-redirect query to ensure
the third party url isn't failing to set the cookie and redirect back
to the correct file because the url is already in the cache from previous runs.

• http/tests/resourceLoadStatistics/cookie-deletion-expected.txt:
• http/tests/resourceLoadStatistics/cookie-deletion.html:

5:08 PM Changeset in webkit [252719] by Wenson Hsieh

• 8 edits in trunk/Tools

Correct argument types in UIScriptController::activateDataListSuggestion after r252062
​https://bugs.webkit.org/show_bug.cgi?id=203116

Reviewed by Tim Horton.

Use unsigned to represent unsigned long arguments declared in the IDL.

• DumpRenderTree/mac/UIScriptControllerMac.h:
• DumpRenderTree/mac/UIScriptControllerMac.mm:

(WTR::UIScriptControllerMac::activateDataListSuggestion):

• TestRunnerShared/UIScriptContext/UIScriptController.h:

(WTR::UIScriptController::activateDataListSuggestion):

• WebKitTestRunner/ios/UIScriptControllerIOS.h:
• WebKitTestRunner/ios/UIScriptControllerIOS.mm:

(WTR::UIScriptControllerIOS::activateDataListSuggestion):

• WebKitTestRunner/mac/UIScriptControllerMac.h:
• WebKitTestRunner/mac/UIScriptControllerMac.mm:

(WTR::UIScriptControllerMac::activateDataListSuggestion):

4:28 PM Changeset in webkit [252718] by Wenson Hsieh

• 2 edits in trunk/Source/WebCore

Unreviewed, remove an unnecessary null check after r252561
​https://bugs.webkit.org/show_bug.cgi?id=204287

Address post-review feedback.

• Modules/async-clipboard/Clipboard.cpp:

(WebCore::Clipboard::writeText):

3:59 PM Changeset in webkit [252717] by ChangSeok Oh

• 33 edits
  1 copy
  1 add in trunk

[GTK] Add ANGLE backend to GTK port
​https://bugs.webkit.org/show_bug.cgi?id=199060

Reviewed by Žan Doberšek.

.:

This change brings ANGLE support for WebGL to GTK port. USE_ANGLE_WEBGL
is newly defined in OptionGTK.cmake, and turns on the USE_ANGLE_EGL flag.

• Source/cmake/OptionsGTK.cmake: USE_ANGLE_WEBGL added.

Source/ThirdParty/ANGLE:

• CMakeLists.txt: When USE_ANGLE_WEBGL is enabled, necessary glesv2 entry points headers are installed in the derived source directory. And then adjust-angle-include-path.sh is applied to change include paths in the entry points header files.
• GLESv2.cmake: libglesv2_entry_points_headers is newly defined, which is a gathering of entry point headers of libGLESv2.
• PlatformGTK.cmake: In this file, we append gl and glx related angle code as a build target. Since we want gl context and gles context simultaneously, we remove gl prototype calls from the generated libGLESv2 library. Instead, we invoke egl and gl calls of ANGLE via their non-prototype function names.
• adjust-angle-include-paths.sh: sed is slightly different in unix and linux systems. In particular, its inplace option needs to be changed for better compatibility on linux.
• include/CMakeLists.txt: ANGLE_WEBGL_HEADERS is newly defined. It gathers header files of ANGLE for WebGL support.

Source/WebCore:

This change aims to bring ANGLE support for WebGL to GTK port. The port
wants to have a gl context for texture mapper and a gles/egl context for webgl
simultaneously, we adopt the readPixel-copyToTexture approach for now because
sharing textures between the two different contextes is not a feasible direction.
Also, to avoid conflicts between gl and gles calls, we use non-prototype functions
of ANGLE for WebGL context (i.e., GraphicsContext3D). Although many combinations
of graphic configurations exist, only default configuration (Nicosia layers
running on ANGLE + GLX is supported. Other combinations like ANGLE + GLES or WPE
will be covered by following patches.

No new tests since no functionality changed.

• Headers.cmake:
• PlatformGTK.cmake:
• SourcesGTK.txt:
• platform/TextureMapper.cmake:
• platform/graphics/ANGLEWebKitBridge.h:
• platform/graphics/GLContext.cpp:
• platform/graphics/GLContext.h:
• platform/graphics/GraphicsContext3D.h:
• platform/graphics/OpenGLShims.cpp:

(WebCore::initializeOpenGLShims):

• platform/graphics/OpenGLShims.h:
• platform/graphics/PlatformDisplay.cpp:
• platform/graphics/angle/GraphicsContext3DANGLE.cpp:

(WebCore::GraphicsContext3D::reshapeFBOs):
(WebCore::GraphicsContext3D::prepareTexture):

• platform/graphics/nicosia/texmap/NicosiaGC3DANGLELayer.cpp: Added.

(Nicosia::GC3DANGLELayer::ANGLEContext::errorString):
(Nicosia::GC3DANGLELayer::ANGLEContext::lastErrorString):
(Nicosia::GC3DANGLELayer::ANGLEContext::createContext):
(Nicosia::GC3DANGLELayer::ANGLEContext::ANGLEContext):
(Nicosia::GC3DANGLELayer::ANGLEContext::~ANGLEContext):
(Nicosia::GC3DANGLELayer::ANGLEContext::makeContextCurrent):
(Nicosia::GC3DANGLELayer::ANGLEContext::platformContext):
(Nicosia::GC3DANGLELayer::GC3DANGLELayer):
(Nicosia::GC3DANGLELayer::~GC3DANGLELayer):
(Nicosia::GC3DANGLELayer::makeContextCurrent):
(Nicosia::GC3DANGLELayer::platformContext):

• platform/graphics/nicosia/texmap/NicosiaGC3DANGLELayer.h: Copied from Source/WebCore/platform/graphics/nicosia/texmap/NicosiaGC3DLayer.h.
• platform/graphics/nicosia/texmap/NicosiaGC3DLayer.cpp:

(Nicosia::GC3DLayer::GC3DLayer):
(Nicosia::GC3DLayer::swapBuffersIfNeeded):

• platform/graphics/nicosia/texmap/NicosiaGC3DLayer.h:
• platform/graphics/texmap/BitmapTextureGL.cpp:

(WebCore::BitmapTextureGL::setPendingContents):
(WebCore::BitmapTextureGL::updatePendingContents):

• platform/graphics/texmap/BitmapTextureGL.h:
• platform/graphics/texmap/GraphicsContext3DTextureMapper.cpp:

(WebCore::GraphicsContext3D::create):
(WebCore::GraphicsContext3D::GraphicsContext3D):
(WebCore::GraphicsContext3D::~GraphicsContext3D):

• platform/graphics/texmap/TextureMapperPlatformLayerBuffer.cpp:

(WebCore::TextureMapperPlatformLayerBuffer::paintToTextureMapper):

• platform/graphics/texmap/TextureMapperPlatformLayerProxy.cpp:

(WebCore::TextureMapperPlatformLayerProxy::pushNextBuffer):

Source/WebKit:

• UIProcess/API/glib/WebKitProtocolHandler.cpp: Extention3DANGLE is used instead where ANGLE for WebGL is enabled.
• UIProcess/gtk/AcceleratedBackingStoreWayland.cpp: Append gl extensions for ANGLE.

(WebKit::AcceleratedBackingStoreWayland::checkRequirements):

Source/WTF:

The GTK port uses TEXTURE_MAPPER that has a gl context for accelerated rendering,
and ANGLE has an egl context for WebGL. We want to make both live together
so an exception is made where TEXTURE_MAPPER is enabled.

• wtf/Platform.h:

3:52 PM Changeset in webkit [252716] by Alan Bujtas

• 15 edits in trunk

Flexbox sizing logic triggers full repaint on the flex items.
​https://bugs.webkit.org/show_bug.cgi?id=204380
<rdar://problem/57236404>

Reviewed by Simon Fraser.

Source/WebCore:

RenderFlexibleBox::applyStretchAlignmentToChild explicitly sets the child renderer's height to 0 before issuing layout on it.
This confuses the child's repaint logic and could trigger unnecessary repaints on complete subtrees.

Many ::layout functions plant a LayoutRepainter stack object to track and report paint invalidations.
It works as long as the renderer's geometry change happens within the scope of this LayoutRepainter.
When the parent (RenderFlexibleBox) mutates the renderer's geometry, the LayoutRepainter object sees
this already mutated state as the initial state and will happily issue repaints even when the final
geometry remains the same.

This patch addresses the redundant repaint by pushing the height reset from the parent down to the child, inside the LayoutRepainter scope.

• rendering/RenderBlock.h: Restrict it to RenderBlock level for now. It might need to go all the way up to RenderBox.

(WebCore::RenderBlock::shouldResetChildLogicalHeightBeforeLayout const):

• rendering/RenderBlockFlow.cpp:

(WebCore::RenderBlockFlow::layoutBlock):

• rendering/RenderBlockFlow.h:
• rendering/RenderBox.cpp:

(WebCore::RenderBox::resetLogicalHeightBeforeLayoutIfNeeded):

• rendering/RenderBox.h:

(WebCore::RenderBox::shouldResetLogicalHeightBeforeLayout const):

• rendering/RenderDeprecatedFlexibleBox.cpp:

(WebCore::RenderDeprecatedFlexibleBox::layoutBlock):

• rendering/RenderFlexibleBox.cpp:

(WebCore::RenderFlexibleBox::layoutBlock):
(WebCore::RenderFlexibleBox::applyStretchAlignmentToChild):

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

(WebCore::RenderGrid::layoutBlock):

• rendering/RenderGrid.h:
• rendering/RenderTable.cpp:

(WebCore::RenderTable::layout):

• rendering/RenderTable.h:

LayoutTests:

• fast/repaint/align-items-change-expected.txt: progression.

3:31 PM Changeset in webkit [252715] by Justin Fan

• 2 edits in trunk/LayoutTests

webgl/1.0.3/conformance/context/context-release-upon-reload.html is flaky
​https://bugs.webkit.org/show_bug.cgi?id=203984

Unreviewed test gardening. Skip the 2.0.0 version of this test as well.

• platform/mac-highsierra-wk1/TestExpectations:

3:21 PM Changeset in webkit [252714] by Fujii Hironori

• 3 edits in trunk/Tools

run-api-tests reports AssertionError os.pathsep not in value in _append_value_colon_separated in Cygwin Python
​https://bugs.webkit.org/show_bug.cgi?id=204400

Reviewed by Jonathan Bedard.

r249500 changed Port.environment_for_api_tests to use
Port._append_value_colon_separated to append a build path to some
enviroment variables. _append_value_colon_separated has a
assertion to ensure the given value doesn't iclude ':'. However,
Port._build_path contains ':' on Cygwin Python.

Those enviroment variables are only for macOS and iOS. Move the code to darwin.py.

• Scripts/webkitpy/port/base.py:

(Port.environment_for_api_tests):

• Scripts/webkitpy/port/darwin.py:

(DarwinPort.environment_for_api_tests):

3:12 PM Changeset in webkit [252713] by Caio Lima

• 18 edits in trunk

[JSC] OSR exit to LLInt is broken on MIPS
​https://bugs.webkit.org/show_bug.cgi?id=203737

Reviewed by Yusuke Suzuki.

JSTests:

Unskipping broken tests due to OSR to LLInt bug.

• microbenchmarks/call-spread-call.js:
• microbenchmarks/throw.js:
• stress/allocation-sinking-hints-are-valid-ssa-2.js:
• stress/allocation-sinking-hints-are-valid-ssa.js:
• stress/arith-profile-for-negate-can-see-non-number-due-to-dfg-osr-exit-profiling.js:
• stress/arrowfunction-lexical-bind-supercall-4.js:
• stress/arrowfunction-tdz-3.js:
• stress/function-constructor-semantics.js:
• stress/global-import-function-should-return-a-promise-when-clearing-exceptions.js:
• stress/stress-cleared-calllinkinfo.js:
• stress/typedarray-configure-index.js:
• stress/v8-deltablue-strict.js:

PerformanceTests/SunSpider:

• tests/v8-v6/v8-deltablue.js:

Source/JavaScriptCore:

This patch is adjusting the OSR to LLInt mechanism to MIPS. When we
use PIC on this architecture, we need to properly configure $gp
at some places to be able to access global variables. This is required
on LLInt to access Global Offset Table (got). According to MIPS ABI,
the $gp can be recalculated during function prologue using caller
register $t9. We also emit such instructions (we can see this as
OFFLINE_ASM_CPLOAD macro) immediately after a non-local label on
LLInt. With the introduction of OSR to LLInt mechanism, we now have
return location labels that are reached from ret LLInt instructions.
Such return locations are used to properly return to LLInt execution
whenever an OSR exits from inlined call on DFG or FTL to LLInt. When
OSR is materializing LLInt stack frames for inlined functions (or
accessors), it sets return address to its return location label.
This means that for such labels, we need to adjust $gp
using $ra instead of $t9, given that LLInt ret operation uses
jr $ra to jump the execution to there.
To implement this, we changed mipsAddPICCode to emit code
using the correct register required to recalculate $gp.
We also changed callTargetFunction to use the stubs as return
location points, since the declaration of global labels will emmit
OFFLINE_ASM_CPLOAD($ra) and we don't want to execute it during
normal LLInt execution.

• llint/LowLevelInterpreter.asm:
• offlineasm/mips.rb:

3:01 PM Changeset in webkit [252712] by Chris Dumez

• 3 edits in trunk/Source/WebKit

[iOS] Make sure WebContent process does not get suspended while it is holding a process assertion for the UIProcess
​https://bugs.webkit.org/show_bug.cgi?id=204418

Reviewed by Jer Noble.

Make sure WebContent process does not get suspended while it is holding a process assertion for the UIProcess. We
see this happening in sysdiagnoses, and it means the system ends up killing the WebContent process because it leaked
a process assertion.

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

(WebKit::WebProcess::processTaskStateDidChange):
(WebKit::WebProcess::releaseProcessWasResumedAssertions):

2:45 PM Changeset in webkit [252711] by Jonathan Bedard

• 3 edits
  1 add in trunk

[WebGL] Add --webgl-test-suite flag for run-webkit-tests
​https://bugs.webkit.org/show_bug.cgi?id=204327

Rubber-stamped by Aakash Jain.

Tools:

• Scripts/webkitpy/layout_tests/run_webkit_tests.py:

(parse_args): Add the --webgl macro command, which is the equivalent of:
'Run-webkit-tests --additional-expectations=LayoutTests/webgl/TestExpectations webgl'

LayoutTests:

• webgl/TestExpectations: Added.

2:29 PM Changeset in webkit [252710] by mmaxfield@apple.com

• 4 edits
  2 adds in trunk

[Cocoa] Add ui-sans-serif as a sibling to ui-serif, ui-monospace, and ui-rounded
​https://bugs.webkit.org/show_bug.cgi?id=204188

Reviewed by Simon Fraser.

Source/WebCore:

As per ​https://github.com/w3c/csswg-drafts/commit/cd0059a32a74c5a48630ce7c77292642dac70fd9#diff-dd2d54c50c872e1e303daebd4794d642

Test: fast/text/design-system-ui-sans-serif.html

• platform/graphics/cocoa/FontCacheCoreText.cpp:

(WebCore::fontWithFamilySpecialCase):

• platform/graphics/cocoa/FontDescriptionCocoa.cpp:

(WebCore::matchSystemFontUse):

LayoutTests:

• fast/text/design-system-ui-sans-serif-expected.html: Added.
• fast/text/design-system-ui-sans-serif.html: Added.

2:25 PM Changeset in webkit [252709] by Alan Coon

• 1 copy in tags/Safari-608.4.9.1.4

Tag Safari-608.4.9.1.4.

2:19 PM Changeset in webkit [252708] by Alan Coon

• 7 edits in branches/safari-608.4.9.1-branch/Source

Versioning.

2:18 PM Changeset in webkit [252707] by ChangSeok Oh

• 2 edits in trunk/Tools

Reactivate my committer status.

Unreviewed.

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

2:13 PM Changeset in webkit [252706] by Truitt Savell

• 1 edit
  3 adds in trunk/LayoutTests

Add platform specific expectations for High Sierra and Mojave for svg/custom/glyph-selection-arabic-forms.svg
​https://bugs.webkit.org/show_bug.cgi?id=204249

Unreviewed test gardening.

• platform/mac-highsierra/svg/custom/glyph-selection-arabic-forms-expected.txt: Added.
• platform/mac-mojave/svg/custom/glyph-selection-arabic-forms-expected.txt: Added.

2:12 PM Changeset in webkit [252705] by Matt Lewis

• 2 edits in trunk/Tools

Add trigger for WebGL bot to config.

RS config fix by Jonathan Bedard.

• BuildSlaveSupport/build.webkit.org-config/config.json:

1:23 PM Changeset in webkit [252704] by yurys@chromium.org

• 8 edits in trunk/Source/WebInspectorUI

Web Inspector: REGRESSION(r250618): main resource view is empty when pausing on inline 'debugger' statement
​https://bugs.webkit.org/show_bug.cgi?id=204086

Reviewed by Devin Rousso.

SourceCodeRevision.currentRevision doesn't create new revisions under the hood anymore.
This allows to avoid undesirable side effects when e.g. text editor tries to read current
content of the SourceCode which results in a new revision (with empty content).
New method editableRevision is introduced for applying changes to the resource while
keeping original content revision intact.

• UserInterface/Controllers/CSSManager.js:

(WI.CSSManager.prototype._resourceContentDidChange.applyStyleSheetChanges.styleSheetFound):
(WI.CSSManager.prototype._resourceContentDidChange.applyStyleSheetChanges):
(WI.CSSManager.prototype._resourceContentDidChange):
(WI.CSSManager.prototype._updateResourceContent.fetchedStyleSheetContent):

• UserInterface/Models/SourceCode.js:

(WI.SourceCode.prototype.get currentRevision):
(WI.SourceCode.prototype.get editableRevision):
(WI.SourceCode.prototype.get content):
(WI.SourceCode.prototype.revisionContentDidChange):

• UserInterface/Views/FontResourceContentView.js:

(WI.FontResourceContentView.prototype.dropZoneHandleDrop):

• UserInterface/Views/ImageResourceContentView.js:

(WI.ImageResourceContentView.prototype.dropZoneHandleDrop):

• UserInterface/Views/ResourceContentView.js:

(WI.ResourceContentView.prototype._contentAvailable):
(WI.ResourceContentView.prototype._handleImportLocalResourceOverride):

• UserInterface/Views/ScriptContentView.js:

(WI.ScriptContentView.prototype._handleTextEditorContentDidChange):

• UserInterface/Views/TextResourceContentView.js:

(WI.TextResourceContentView.prototype._textEditorContentDidChange):

1:04 PM Changeset in webkit [252703] by Matt Lewis

• 3 edits in trunk/Tools

Bring up WebGL Test Queue.
​https://bugs.webkit.org/show_bug.cgi?id=204374

Reviewed by Jonathan Bedard.

• BuildSlaveSupport/build.webkit.org-config/config.json:
• BuildSlaveSupport/build.webkit.org-config/steps_unittest.py:

12:35 PM Changeset in webkit [252702] by BJ Burg

• 14 edits
  1 add in trunk/Source

[Cocoa] Add _WKRemoteWebInspectorViewController SPI to set diagnostic logging delegate
​https://bugs.webkit.org/show_bug.cgi?id=204371

Reviewed by Devin Rousso.

Source/WebCore:

Expose whether a diagnostic logging delegate has been configured for the frontend's WKWebView.
In a later patch, WI.DiagnosticController will start and stop recording as availability changes.

• inspector/InspectorFrontendClient.h:

(WebCore::InspectorFrontendClient::diagnosticLoggingAvailable):

• inspector/InspectorFrontendHost.cpp:

(WebCore::InspectorFrontendHost::diagnosticLoggingAvailable):

• inspector/InspectorFrontendHost.h:

Source/WebKit:

• UIProcess/WebInspectorProxy.h:
• UIProcess/RemoteWebInspectorProxy.h:
• UIProcess/RemoteWebInspectorProxy.cpp:

(WebKit::RemoteWebInspectorProxy::setDiagnosticLoggingAvailable):
Hook up the plumbing to forward this state change to the frontend.

• UIProcess/API/Cocoa/_WKRemoteWebInspectorViewControllerPrivate.h: Added.
• UIProcess/API/Cocoa/_WKRemoteWebInspectorViewController.mm:

(-[_WKRemoteWebInspectorViewController _setDiagnosticLoggingDelegate:]):
Add new SPI to set the diagnostic logging delegate. This is preferable to directly
setting the delegate on the inspector WKWebView (via .webView._diagnosticLoggingDelegate),
because the controller can notify the frontend that diagnostic logging is now available.

• WebKit.xcodeproj/project.pbxproj: Add new header file and make it private.

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

(WebKit::WebInspectorUI::setDiagnosticLoggingAvailable):
Cache this value and use it for InspectorFrontendClient::diagnosticLoggingAvailable().

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

(WebKit::RemoteWebInspectorUI::setDiagnosticLoggingAvailable):
Duplicate what is done for the local case.

12:33 PM Changeset in webkit [252701] by Simon Fraser

• 7 edits in trunk

Turn off antialiasing when rendering with Ahem, for testing
​https://bugs.webkit.org/show_bug.cgi?id=204197
Source/WebCore:

<rdar://problem/57190712>

Reviewed by Myles C. Maxfield.

If the font family name matches "Ahem", turn off antialiasing when rendering text,
because many CSS tests assume that Ahem glyphs can be compared with positioned elements
in ref tests.

Since Ahem is a font only used for testing, I don't think this behavior needs to be
controlled by a setting.

• platform/graphics/Font.cpp:

(WebCore::Font::Font):

• platform/graphics/Font.h:

(WebCore::Font::allowsAntialiasing const):

• platform/graphics/cocoa/FontCascadeCocoa.mm:

(WebCore::FontCascade::drawGlyphs):

• platform/graphics/cocoa/FontCocoa.mm:

(WebCore::isAhemFont):
(WebCore::Font::platformInit):

LayoutTests:

Reviewed by Myles C. Maxfield.

• TestExpectations:

12:14 PM Changeset in webkit [252700] by Jonathan Bedard

• 2 edits in trunk/Tools

[EWS] iOS layout-tests fails with 30+ failures on re-run and on clean-tree
​https://bugs.webkit.org/show_bug.cgi?id=204409

Reviewed by Aakash Jain.

• BuildSlaveSupport/ews-build/steps.py:

(RunWebKitPerlTests.evaluateCommand): Add killOldProcesses before running test suite.
(RunJavaScriptCoreTests.evaluateCommand): Ditto.
(ReRunJavaScriptCoreTests.evaluateCommand): Ditto.
(RunWebKitTests.evaluateCommand): Ditto.
(ReRunWebKitTests.evaluateCommand): Ditto.
(RunAPITests.evaluateCommand): Ditto.
(ReRunAPITests.evaluateCommand): Ditto.

12:06 PM Changeset in webkit [252699] by rmorisset@apple.com

• 2 edits in trunk/Source/JavaScriptCore

Fix load<16> on ARM64
​https://bugs.webkit.org/show_bug.cgi?id=204326

Reviewed by Mark Lam.

On ARM64 I used load<16> in ​https://bugs.webkit.org/show_bug.cgi?id=202832.
Unfortunately it turns out to call ldr<16>, and ldr<n> asserts that n is either 32 or 64.
This fix simply calls ldrh/strh directly.

• assembler/MacroAssemblerARM64.h:

(JSC::MacroAssemblerARM64::load16):
(JSC::MacroAssemblerARM64::store16):

11:52 AM Changeset in webkit [252698] by pvollan@apple.com

• 2 edits in trunk/LayoutTests

Layout Test fast/text/font-antialiasing-save-restore.html is failing
​https://bugs.webkit.org/show_bug.cgi?id=204413

Unreviewed test gardening.

• platform/win/TestExpectations:

11:28 AM Changeset in webkit [252697] by Alan Coon

• 2 edits in tags/Safari-609.1.10.2/Source/WebKit

Cherry-pick r252333. rdar://problem/57079557

[iOS] Unable to view .pages files
​https://bugs.webkit.org/show_bug.cgi?id=204076
<rdar://problem/57079557>

Reviewed by Brent Fulgham.

This is caused by a syscall being blocked by the sandbox. The syscall was previously unused in the
WebContent process.

• Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb:

git-svn-id: ​https://svn.webkit.org/repository/webkit/trunk@252333 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10:57 AM Changeset in webkit [252696] by Alan Bujtas

• 5 edits in trunk/Source/WebCore

[LFC][IFC] Cache InlinteTextItem width when possible
​https://bugs.webkit.org/show_bug.cgi?id=204412
<rdar://problem/57359366>

Reviewed by Antti Koivisto.

This helps when subsequent inline layouts use the same set of InlineTextItems and they would
keep re-measuring them on each layout frame (e.g. view resize).

• layout/inlineformatting/InlineLineLayout.cpp:

(WebCore::Layout::inlineItemWidth):

• layout/inlineformatting/InlineTextItem.cpp:

(WebCore::Layout::InlineTextItem::createAndAppendTextItems):
(WebCore::Layout::InlineTextItem::createWhitespaceItem):
(WebCore::Layout::InlineTextItem::createNonWhitespaceItem):
(WebCore::Layout::InlineTextItem::createSegmentBreakItem):
(WebCore::Layout::InlineTextItem::InlineTextItem):
(WebCore::Layout::InlineTextItem::left const):
(WebCore::Layout::InlineTextItem::right const):

• layout/inlineformatting/InlineTextItem.h:

(WebCore::Layout::InlineTextItem::width const):

• layout/inlineformatting/text/TextUtil.h:

(WebCore::Layout::TextUtil::width):

10:53 AM Changeset in webkit [252695] by Truitt Savell

• 2 edits in trunk/LayoutTests

Regression (r252660): Layout Test platform/ios/mediastream/audio-muted-in-background-tab.html is failing
​https://bugs.webkit.org/show_bug.cgi?id=204408

Unreviewed test gardening.

• platform/ios/TestExpectations:

10:50 AM Changeset in webkit [252694] by mark.lam@apple.com

• 4 edits in trunk

Flaky JSC test: stress/stack-overflow-in-yarr-byteCompile.js.dfg-eager.
​https://bugs.webkit.org/show_bug.cgi?id=204405

Reviewed by Alexey Proskuryakov.

JSTests:

The test was timing out because it's slow. We don't really need to run this test
on all configurations. So, we'll change it to on run with the default
configuration. Also declare the test as "@ slow!".

• stress/stack-overflow-in-yarr-byteCompile.js:

Source/JavaScriptCore:

$vm.allWithStackSize() manipulates the stack in ways that will freak out ASan.
So, add SUPPRESS_ASAN to functionCallWithStackSize() to tell ASan to ignore it.

• tools/JSDollarVM.cpp:

10:39 AM Changeset in webkit [252693] by Truitt Savell

• 2 edits in trunk/LayoutTests

Rebase svg/custom/glyph-selection-arabic-forms.svg
​https://bugs.webkit.org/show_bug.cgi?id=204249

Unreviewed test gardening.

• svg/custom/glyph-selection-arabic-forms-expected.txt:

10:32 AM Changeset in webkit [252692] by jer.noble@apple.com

• 7 edits
  2 adds in trunk

HTML5 audio .ended event not fired when app in background or phone screen is off
​https://bugs.webkit.org/show_bug.cgi?id=173332
<rdar://problem/32757402>

Reviewed by Eric Carlson.

Source/WebCore:

Test: media/audio-background-playback-playlist.html

In addition to the necessary WebKit part of this patch, there are behaviors which prevent
websites from enqueuing new playback while in the background. Namely, the platform will
prevent background playback from any application which is not currently the "Now Playing"
application, so in order to allow pages to switch sources, we must ensure we do not give
up "Now Playing" status. To do so, we will change the implementation of canProduceAudio()
to include any media element which previously could produce audio but currently has no
source.

Also, MediaElementSession::canShowControlsManager() will be modified to only check for
a RequireUserGestureToControlControlsManager restriction if the purpose passed in is
is "ControlsManager" and not "NowPlaying".

• html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::canProduceAudio const):

• html/MediaElementSession.cpp:

(WebCore::MediaElementSession::canShowControlsManager const):

Source/WebKit:

When a WebPage goes from audible to inaudible, allow a short grace period before removing
the activity token, to give the page a chance to (e.g.) move to the next item in a playlist
before the process is suspended when in the background.

• UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::close):
(WebKit::WebPageProxy::updateThrottleState):
(WebKit::WebPageProxy::clearAudibleActivity):

• UIProcess/WebPageProxy.h:

LayoutTests:

• media/audio-background-playback-playlist-expected.txt: Added.
• media/audio-background-playback-playlist.html: Added.

10:25 AM Changeset in webkit [252691] by Jonathan Bedard

• 2 edits in trunk/Tools

kill-old-processes: Log currently running processes
​https://bugs.webkit.org/show_bug.cgi?id=204415

Unreviewed infrastructure management.

• BuildSlaveSupport/kill-old-processes:

(main): Log 'ps aux' for Mac and Linux.

9:48 AM Changeset in webkit [252690] by Caio Lima

• 6 edits in trunk/Source/JavaScriptCore

Regression (r252680): JSCOnly build broken: no matching function for call to JSC::DFG::SpeculativeJIT::jsValueResult
​https://bugs.webkit.org/show_bug.cgi?id=204404

Reviewed by Saam Barati.

Adjusting build after changes from r252684 and r252680.

• bytecode/AccessCase.cpp:

(JSC::AccessCase::generateWithGuard):

• dfg/DFGSpeculativeJIT.cpp:

(JSC::DFG::SpeculativeJIT::compileIncOrDec):

• dfg/DFGSpeculativeJIT32_64.cpp:

(JSC::DFG::SpeculativeJIT::compile):

• jit/JITOpcodes32_64.cpp:

(JSC::JIT::emit_op_to_numeric):

• jit/JITPropertyAccess32_64.cpp:

(JSC::JIT::emit_op_get_by_val):
(JSC::JIT::emitSlow_op_get_by_val):

9:12 AM Changeset in webkit [252689] by Simon Fraser

• 3 edits
  2 adds in trunk

REGRESSION (r252161): box-shadow with inset and rounded borders is clipped
​https://bugs.webkit.org/show_bug.cgi?id=204348

Reviewed by Zalan Bujtas.

Source/WebCore:

r252161 introduced a bug for inset shadows where we passed the wrong rect to areaCastingShadowInHole().
Clean the code up a little, and make the xOffset stuff more similar to the outset shadow path.
We could still do a better job of pixel snapping here.

Test: fast/box-shadow/inset-box-shadow.html

• rendering/RenderBoxModelObject.cpp:

(WebCore::RenderBoxModelObject::paintBoxShadow):

LayoutTests:

• fast/box-shadow/inset-box-shadow-expected.html: Added.
• fast/box-shadow/inset-box-shadow.html: Added.

8:53 AM Changeset in webkit [252688] by pvollan@apple.com

• 2 edits in trunk/LayoutTests

Layout Test fast/text/font-antialiasing-save-restore.html is failing
​https://bugs.webkit.org/show_bug.cgi?id=204413

Unreviewed test gardening.

• platform/win/TestExpectations:

8:04 AM Changeset in webkit [252687] by Fujii Hironori

• 5 edits
  1 delete in trunk/Source/WTF

[Win] Implement WTF::ThreadSpecific in WTF::Thread
​https://bugs.webkit.org/show_bug.cgi?id=204341

Reviewed by Brent Fulgham and Yusuke Suzuki.

Thread::destructTLS had a tricky code to defer destroying
WTF::Thread in TLS in order to ensure WTF::Thread is destructed
after other ThreadSpecific are destructed, which is a part of
cause of nasty hanging issue in the process termination (Bug 204192).

This change implements WTF::ThreadSpecific in WTF::Thread by
adding a new class Thread::SpecificStorage to manage TLS. Simplify
Thread::destructTLS. Remove threadMapMutex in ThreadingWin.cpp

• wtf/PlatformWin.cmake:
• wtf/ThreadSpecific.h:

(WTF::canBeGCThread>::ThreadSpecific):
(WTF::canBeGCThread>::get):
(WTF::canBeGCThread>::setInTLS):
(WTF::canBeGCThread>::destroy):
(WTF::canBeGCThread>::~ThreadSpecific): Deleted.

• wtf/Threading.h:

(WTF::Thread::specificStorage):
(WTF::Thread::current):

• wtf/win/ThreadSpecificWin.cpp: Removed.
• wtf/win/ThreadingWin.cpp:

(WTF::Thread::initializeTLSKey):
(WTF::Thread::initializeTLS):
(WTF::Thread::destructTLS):
(WTF::Thread::SpecificStorage::allocateKey):
(WTF::Thread::SpecificStorage::get):
(WTF::Thread::SpecificStorage::set):
(WTF::Thread::SpecificStorage::destroySlots):
(): Deleted.
(WTF::Thread::currentDying): Deleted.
(WTF::Thread::get): Deleted.

7:18 AM Changeset in webkit [252686] by Alan Bujtas

• 9 edits
  1 copy in trunk/Source/WebCore

[LFC][IFC] TextUtil::width should use FontCascade::widthForSimpleText when applicable
​https://bugs.webkit.org/show_bug.cgi?id=204393
<rdar://problem/57347144>

Reviewed by Antti Koivisto.

This is one of reasons why simple line layout performs better than complex line layout.

• WebCore.xcodeproj/project.pbxproj:
• layout/blockformatting/BlockFormattingContext.cpp:

(WebCore::Layout::BlockFormattingContext::computeHeightAndMargin):

• layout/inlineformatting/InlineLine.cpp:

(WebCore::Layout::Line::Run::expand):
(WebCore::Layout::Line::appendTextContent):

• layout/inlineformatting/InlineTextItem.cpp:

(WebCore::Layout::moveToNextNonWhitespacePosition):
(WebCore::Layout::InlineTextItem::createAndAppendTextItems):

• layout/inlineformatting/text/TextUtil.cpp:

(WebCore::Layout::TextUtil::width):
(WebCore::Layout::TextUtil::fixedPitchWidth):

• layout/inlineformatting/text/TextUtil.h:
• layout/layouttree/LayoutBox.cpp:

(WebCore::Layout::Box::Box):
(WebCore::Layout::Box::setTextContext):
(WebCore::Layout::Box::hasTextContent const):
(WebCore::Layout::Box::textContext const):
(WebCore::Layout::Box::setTextContent): Deleted.
(WebCore::Layout::Box::textContent const): Deleted.

• layout/layouttree/LayoutBox.h:
• layout/layouttree/LayoutTreeBuilder.cpp:

(WebCore::Layout::canUseSimplifiedTextMeasuring):
(WebCore::Layout::TreeBuilder::createLayoutBox):
(WebCore::Layout::outputLayoutBox):