Timeline
Dec 26, 2021:
- 11:46 PM Changeset in webkit [287458] by
-
- 4 edits in trunk/Source/JavaScriptCore
[JSC] Use SlowPathReturnType instead of EncodedMatchResult
https://bugs.webkit.org/show_bug.cgi?id=234686
Reviewed by Filip Pizlo.
This patch replaces YarrJIT's EncodedMatchResult with SlowPathReturnType so that
CCallHelpers can handle it correctly.
- runtime/MatchResult.h:
(JSC::MatchResult::MatchResult):
(JSC::MatchResult::failed):
- runtime/SlowPathReturnType.h:
(JSC::decodeResult):
- yarr/YarrJIT.h:
- 11:38 PM Changeset in webkit [287457] by
-
- 2 edits in trunk/Source/WebCore
[GTK][a11y] Test accessibility/aria-inherits-presentational.html is failing with ATSPI
https://bugs.webkit.org/show_bug.cgi?id=234685
Reviewed by Adrian Perez de Castro.
We are ignoring list items inheriting presentational role, but we need to expose its contents.
- accessibility/atspi/AccessibilityObjectAtspi.cpp:
(WebCore::AccessibilityObjectAtspi::effectiveRole const): Use section as effective role of list items inheriting
presentational role.
(WebCore::AccessibilityObject::accessibilityPlatformIncludesObject const): Include list items inheriting
presentational role.
- 11:37 PM Changeset in webkit [287456] by
-
- 8 edits in trunk
[GTK][a11y] Stop setting manages-descendants on web process root object with ATSPI
https://bugs.webkit.org/show_bug.cgi?id=234663
Reviewed by Adrian Perez de Castro.
Source/WebCore:
Emit parent-changed and children-changed signals instead, like all other objects, to properly update the ATSPI
cache.
- accessibility/atspi/AccessibilityAtspi.cpp:
(WebCore::AccessibilityAtspi::parentChanged): Emit the signal for the root object.
(WebCore::AccessibilityAtspi::childrenChanged): Ditto.
- accessibility/atspi/AccessibilityAtspi.h:
- accessibility/atspi/AccessibilityObjectAtspi.cpp:
(WebCore::AccessibilityObjectAtspi::cacheDestroyed): Notify the root object too.
(WebCore::AccessibilityObjectAtspi::elementDestroyed): Ditto.
(WebCore::AccessibilityObjectAtspi::setParent): Ditto.
(WebCore::AccessibilityObjectAtspi::serialize const): Use AccessibilityRootAtspi::applicationReference().
- accessibility/atspi/AccessibilityRootAtspi.cpp:
(WebCore::AccessibilityRootAtspi::embedded): Call AccessibilityAtspi::parentChanged().
(WebCore::AccessibilityRootAtspi::parentReference const): Added.
(WebCore::AccessibilityRootAtspi::childAdded): Call AccessibilityAtspi::childrenChanged().
(WebCore::AccessibilityRootAtspi::childRemoved): Ditto.
(WebCore::AccessibilityRootAtspi::serialize const): Do not set manages-descendants state.
- accessibility/atspi/AccessibilityRootAtspi.h:
Tools:
Update the unit test to remove the ATSPI ifdef.
- TestWebKitAPI/Tests/WebKitGtk/TestWebKitAccessibility.cpp:
(testAccessibleBasicHierarchy):
- 11:36 PM Changeset in webkit [287455] by
-
- 3 edits in trunk/Source/WebCore
[GTK][a11y] Test /webkit/WebKitAccessibility/selection/listbox is flaky with ATSPI
https://bugs.webkit.org/show_bug.cgi?id=234665
Reviewed by Adrian Perez de Castro.
The problem is that selection-changed can be called multiple times for list boxes. When the notification is
deferred by the AXObject cache, it's ensured to be called only one for every element in the following run loop
iteration, but when not deferred the signal is emitted for every notification.
- accessibility/atspi/AccessibilityObjectAtspi.h:
- accessibility/atspi/AccessibilityObjectSelectionAtspi.cpp:
(WebCore::AccessibilityObjectAtspi::selectionChanged): Ensure we don't emit the selection-changed signal more
than once in the same run loop iteration.
- 11:34 PM Changeset in webkit [287454] by
-
- 2 edits in trunk/Tools
[GTK][a11y] WTR: handle xml-roles attributes to return the right role for landmark with ATSPI
https://bugs.webkit.org/show_bug.cgi?id=234638
Reviewed by Adrian Perez de Castro.
- WebKitTestRunner/InjectedBundle/atspi/AccessibilityUIElementAtspi.cpp:
(WTR::xmlRoleValueString):
(WTR::AccessibilityUIElement::role):
- 10:43 PM Changeset in webkit [287453] by
-
- 11 edits in trunk
[LFC][IFC] Do not trim the hanging trailing content
https://bugs.webkit.org/show_bug.cgi?id=234679
Reviewed by Antti Koivisto.
Source/WebCore:
This is the final step of making the hanging content handling inline with the spec.
(https://www.w3.org/TR/css-text-3/#hang)
Since we don't (fake)trim such content when closing the line, the content width matches
the actual content on the line which also helps when it is aligned to support RTL ordering.
- layout/formattingContexts/inline/InlineLine.cpp:
(WebCore::Layout::Line::Run::removeTrailingWhitespace):
(WebCore::Layout::Line::visuallyCollapseHangingOverflowingGlyphs): Deleted.
(WebCore::Layout::Line::Run::visuallyCollapseTrailingWhitespace): Deleted.
- layout/formattingContexts/inline/InlineLine.h:
- layout/formattingContexts/inline/InlineLineBoxBuilder.cpp:
(WebCore::Layout::horizontalAlignmentOffset):
- layout/formattingContexts/inline/InlineLineBuilder.cpp:
(WebCore::Layout::LineBuilder::close):
LayoutTests:
- platform/mac/fast/text/whitespace/pre-wrap-overflow-selection-expected.txt: Progression. Matches FF and Chrome selection behavior.
- 4:46 PM Changeset in webkit [287452] by
-
- 12 edits1 delete in trunk
Replace WebKit/CocoaColor.h with WebCore/ColorCocoa.h, not confusing at all
https://bugs.webkit.org/show_bug.cgi?id=234693
Reviewed by Simon Fraser.
Source/WebKit:
We don't need both. The WebCore one can work for both.
- Platform/cocoa/CocoaColor.h: Removed.
- Shared/Cocoa/ArgumentCodersCocoa.mm:
(IPC::typeFromObject):
(IPC::encodeColorInternal):
(IPC::encodeObject):
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView themeColor]):
(-[WKWebView underPageBackgroundColor]):
(-[WKWebView setUnderPageBackgroundColor:]):
(-[WKWebView _themeColor]):
(-[WKWebView _pageExtendedBackgroundColor]):
(-[WKWebView _sampledPageTopColor]):
- UIProcess/API/Cocoa/_WKApplicationManifest.mm:
(-[_WKApplicationManifest initWithCoder:]):
(-[_WKApplicationManifest themeColor]):
- UIProcess/Cocoa/WKSafeBrowsingWarning.h:
- UIProcess/Cocoa/WKSafeBrowsingWarning.mm:
(colorForItem):
(-[WKSafeBrowsingBox setSafeBrowsingBackgroundColor:]):
(-[WKSafeBrowsingWarning showDetailsClicked]):
(-[WKSafeBrowsingTextView initWithAttributedString:forWarning:]):
- WebKit.xcodeproj/project.pbxproj:
- WebProcess/Plugins/PDF/PDFPlugin.mm:
(WebKit::PDFPlugin::PDFPlugin):
Tools:
Use WebCore/ColorCocoa.h instead.
- TestWebKitAPI/Tests/WebKitCocoa/CopyHTML.mm:
- TestWebKitAPI/Tests/WebKitCocoa/WKWebViewThemeColor.mm:
- TestWebKitAPI/Tests/WebKitCocoa/WKWebViewUnderPageBackgroundColor.mm:
- 4:03 PM Changeset in webkit [287451] by
-
- 8 edits2 adds in trunk
Make -webkit-text-combine an inherited property
https://bugs.webkit.org/show_bug.cgi?id=234694
Reviewed by Simon Fraser.
Source/WebCore:
-webkit-text-combine is said to be inherited in CSSProperties.json, but it is not actually the case.
It should be inherited per-spec (for text-combine-upright).
https://drafts.csswg.org/css-writing-modes/#text-combine-upright
Test: fast/text/text-combine-inherited.html
- rendering/style/RenderStyle.cpp:
(WebCore::rareNonInheritedDataChangeRequiresLayout):
(WebCore::rareInheritedDataChangeRequiresLayout):
- rendering/style/RenderStyle.h:
(WebCore::RenderStyle::textCombine const):
(WebCore::RenderStyle::setTextCombine):
- rendering/style/StyleRareInheritedData.cpp:
(WebCore::StyleRareInheritedData::StyleRareInheritedData):
(WebCore::StyleRareInheritedData::operator== const):
- rendering/style/StyleRareInheritedData.h:
- rendering/style/StyleRareNonInheritedData.cpp:
(WebCore::StyleRareNonInheritedData::StyleRareNonInheritedData):
(WebCore::StyleRareNonInheritedData::operator== const):
- rendering/style/StyleRareNonInheritedData.h:
LayoutTests:
There are also WPT that already exist for the standard version (text-combine-upright):
- css/css-writing-modes/text-combine-upright-inherit-all-001.html
- css/css-writing-modes/text-combine-upright-inherit-all-002.html
- fast/text/text-combine-inherited-expected.html: Added.
- fast/text/text-combine-inherited.html: Added.
- 3:11 PM Changeset in webkit [287450] by
-
- 2 edits in trunk/Source/WebKit
RemoteInspectorProtocolHandler::inspect() expects ConnectionID as uint32_t
https://bugs.webkit.org/show_bug.cgi?id=234657
Reviewed by Don Olmstead.
connectionID a parameter of ScriptMessageClient::didPostMessage() is casteed as <int>.
Since ConnectionID is define as uint32_t, it should be casted as <uint32_t>.
- UIProcess/Inspector/socket/RemoteInspectorProtocolHandler.cpp: Fixed cast for connectionID.
- 1:53 PM Changeset in webkit [287449] by
-
- 2 edits in trunk/Source/WebCore
null ptr deref in WebCore::indexForVisiblePosition()
https://bugs.webkit.org/show_bug.cgi?id=234255
Patch by Frédéric Wang <fwang@igalia.com> on 2021-12-26
Reviewed by Darin Adler.
- editing/InsertListCommand.cpp:
(WebCore::InsertListCommand::doApply): Ensure endOfSelection did not become orphan before
calling indexForVisiblePosition(). This may happen because the VisiblePosition constructor
is called in the while loop, triggering re-layout and post-layout tasks.
- 1:40 PM Changeset in webkit [287448] by
-
- 4 edits in trunk/Source/WebCore
Remove -webkit-{border-fit/margin-collapse} leftovers in StyleRareNonInheritedData
https://bugs.webkit.org/show_bug.cgi?id=234692
Reviewed by Simon Fraser.
r285615 removed -webkit-border-fit.
r287429 removed -webkit-margin-collapse properties.
- rendering/style/RenderStyle.cpp:
(WebCore::RenderStyle::hashForTextAutosizing const):
(WebCore::RenderStyle::equalForTextAutosizing const):
(WebCore::rareNonInheritedDataChangeRequiresLayout):
- rendering/style/StyleRareNonInheritedData.cpp:
(WebCore::StyleRareNonInheritedData::StyleRareNonInheritedData):
(WebCore::StyleRareNonInheritedData::operator== const):
- rendering/style/StyleRareNonInheritedData.h:
- 11:24 AM Changeset in webkit [287447] by
-
- 8 edits in trunk/Source/WebCore
[IFC][Integration] Update text renderer's needsVisualReordering bit
https://bugs.webkit.org/show_bug.cgi?id=234688
Reviewed by Antti Koivisto.
This is similar to legacy line layout where the RenderText's needsVisualReordering is
updated as the (bidi) text box is being placed on the line.
Here we update this bit right after the line layout, when we finished constructing the display boxes.
- layout/formattingContexts/inline/InlineItemsBuilder.cpp:
(WebCore::Layout::InlineItemsBuilder::handleTextContent):
- layout/integration/LayoutIntegrationBoxTree.cpp:
(WebCore::LayoutIntegration::BoxTree::buildTree):
- layout/integration/LayoutIntegrationInlineContentBuilder.cpp:
(WebCore::LayoutIntegration::InlineContentBuilder::InlineContentBuilder):
(WebCore::LayoutIntegration::InlineContentBuilder::build const):
- layout/integration/LayoutIntegrationInlineContentBuilder.h:
- layout/layouttree/LayoutInlineTextBox.cpp:
(WebCore::Layout::InlineTextBox::InlineTextBox):
(WebCore::Layout::m_canUseSimplifiedContentMeasuring):
(WebCore::Layout::m_containsBidiText): Deleted.
- layout/layouttree/LayoutInlineTextBox.h:
(WebCore::Layout::InlineTextBox::canUseSimplifiedContentMeasuring const):
(WebCore::Layout::InlineTextBox::containsBidiText const): Deleted. No need to cache this value on the layout box
InlineTextItems more or less have the same lifecycle as their associated layout boxes.
- layout/layouttree/LayoutTreeBuilder.cpp:
(WebCore::Layout::TreeBuilder::createTextBox):
- 10:47 AM Changeset in webkit [287446] by
-
- 2 edits in trunk/Source/WebCore
Update writing-mode property values in CSSProperties.json
https://bugs.webkit.org/show_bug.cgi?id=234689
Reviewed by Simon Fraser.
horizontal-tb/vertical-lr/vertical-rl have been supported for a while now.
horizontal-bt has too, but it is non-standard.
- css/CSSProperties.json:
- 10:45 AM Changeset in webkit [287445] by
-
- 8 edits in trunk
[:has() pseudo-class] Support :disabled and :enabled pseudo-class invalidation
https://bugs.webkit.org/show_bug.cgi?id=234636
Reviewed by Simon Fraser.
LayoutTests/imported/w3c:
- web-platform-tests/css/selectors/invalidation/has-pseudo-class.html:
Source/WebCore:
Use Style::PseudoClassChangeInvalidation to support invalidation with :has(:disabled).
- html/HTMLFormControlElement.cpp:
(WebCore::HTMLFormControlElement::setAncestorDisabled):
(WebCore::HTMLFormControlElement::parseAttribute):
(WebCore::HTMLFormControlElement::disabledStateChanged):
- html/HTMLOptGroupElement.cpp:
(WebCore::HTMLOptGroupElement::isDisabledFormControl const):
Use a member bit instead of checking the attribute directly. This allows invalidation to be scoped over the state change.
(WebCore::HTMLOptGroupElement::parseAttribute):
Optgroup can flip the disabled status of the associated option elements too so handle that specifically.
- html/HTMLOptGroupElement.h:
- html/HTMLOptionElement.cpp:
(WebCore::HTMLOptionElement::parseAttribute):
- 9:21 AM Changeset in webkit [287444] by
-
- 8 edits in trunk/Source/WebCore
makeTextLogicalOrderCacheIfNeeded is only interested in whether the content needs visual reordering
https://bugs.webkit.org/show_bug.cgi?id=234687
Reviewed by Antti Koivisto.
This renaming patch is in preparation for fixing the caching codepath for LFC bidi.
- layout/formattingContexts/inline/text/TextUtil.cpp:
(WebCore::Layout::TextUtil::containsStrongDirectionalityText):
(WebCore::Layout::TextUtil::containsBidiText): Deleted.
- layout/formattingContexts/inline/text/TextUtil.h:
- layout/integration/InlineIteratorLogicalOrderTraversal.cpp:
(WebCore::InlineIterator::makeTextLogicalOrderCacheIfNeeded):
- layout/integration/LayoutIntegrationBoxTree.cpp:
(WebCore::LayoutIntegration::BoxTree::buildTree):
- layout/layouttree/LayoutTreeBuilder.cpp:
(WebCore::Layout::TreeBuilder::createTextBox):
- rendering/RenderText.cpp:
(WebCore::RenderText::RenderText):
(WebCore::RenderText::positionLineBox):
- rendering/RenderText.h:
(WebCore::RenderText::needsVisualReordering const):
(WebCore::RenderText::setNeedsVisualReordering):
(WebCore::RenderText::containsBidiText const): Deleted.
(WebCore::RenderText::setContainsBidiText): Deleted.
- 7:02 AM Changeset in webkit [287443] by
-
- 4 edits in trunk/Source/WebCore
[LFC][IFC] ContinuousContent should tell InlineContentBreaker if the candidate content is hanging
https://bugs.webkit.org/show_bug.cgi?id=234671
Reviewed by Antti Koivisto.
ContinuousContent now keeps track of whether the candidate run is a hanging content.
It enables InlineContentBreaker to just simply look at the flag to decide if the overflowing
(hanging) content should stay on the line or not.
- layout/formattingContexts/inline/InlineContentBreaker.cpp:
(WebCore::Layout::InlineContentBreaker::processOverflowingContent const):
(WebCore::Layout::InlineContentBreaker::ContinuousContent::append):
(WebCore::Layout::InlineContentBreaker::ContinuousContent::reset):
- layout/formattingContexts/inline/InlineContentBreaker.h:
(WebCore::Layout::InlineContentBreaker::ContinuousContent::isHangingContent const):
- layout/formattingContexts/inline/InlineLineBuilder.cpp:
(WebCore::Layout::LineCandidate::InlineContent::appendInlineItem):
Dec 25, 2021:
- 10:16 PM Changeset in webkit [287442] by
-
- 4 edits in trunk/Source/WebCore
[LFC][IFC] Turn InlineContentBreaker::ContinuousContent's leading/trailing member variables to std::optional<InlineLayoutUnit>
https://bugs.webkit.org/show_bug.cgi?id=234676
Reviewed by Antti Koivisto.
While an empty (logical width = 0) ContinuousContent takes up no space, it should not be mistaken for collapsed content
(i.e when m_logicalWidth(0) == m_trailingCollapsibleWidth(0), hasCollapsibleContent() should still return false)
- layout/formattingContexts/inline/InlineContentBreaker.cpp:
(WebCore::Layout::InlineContentBreaker::processOverflowingContent const):
(WebCore::Layout::InlineContentBreaker::tryBreakingOverflowingRun const):
(WebCore::Layout::InlineContentBreaker::tryBreakingPreviousNonOverflowingRuns const):
(WebCore::Layout::InlineContentBreaker::ContinuousContent::appendToRunList): Add a dedicated function to deal with append and logicalWidth.
(WebCore::Layout::InlineContentBreaker::ContinuousContent::resetTrailingWhitespace): Turn trailing whitespace into leading if applicable.
(WebCore::Layout::InlineContentBreaker::ContinuousContent::append): Add dedicated functions for each inline item types.
- layout/formattingContexts/inline/InlineContentBreaker.h:
(WebCore::Layout::InlineContentBreaker::ContinuousContent::leadingCollapsibleWidth const):
(WebCore::Layout::InlineContentBreaker::ContinuousContent::trailingCollapsibleWidth const):
(WebCore::Layout::InlineContentBreaker::ContinuousContent::hasCollapsibleContent const):
(WebCore::Layout::InlineContentBreaker::ContinuousContent::isFullyCollapsible const):
- layout/formattingContexts/inline/InlineLineBuilder.cpp:
(WebCore::Layout::LineCandidate::InlineContent::appendInlineItem):
- 10:10 PM Changeset in webkit [287441] by
-
- 2 edits in trunk/LayoutTests
[iOS] Rebaselining after r287440.
Unreviewed.
- platform/ios/fast/borders/rtl-border-01-expected.txt: Progression. Content does not overflow the line anymore.
- 12:55 PM Changeset in webkit [287440] by
-
- 3 edits in trunk/Source/WebCore
[LFC][IFC] Remove redundant InlineContentBreaker::shouldKeepEndOfLineWhitespace
https://bugs.webkit.org/show_bug.cgi?id=234668
Reviewed by Antti Koivisto.
This is in preparation for moving pre-wrap checking out of InlineContentBreaker.
- layout/formattingContexts/inline/InlineContentBreaker.cpp:
(WebCore::Layout::InlineContentBreaker::processOverflowingContent const):
(WebCore::Layout::InlineContentBreaker::shouldKeepEndOfLineWhitespace const): Deleted.
- layout/formattingContexts/inline/InlineContentBreaker.h:
- 11:18 AM Changeset in webkit [287439] by
-
- 2 edits in trunk
Add my github username to contributors.json
https://bugs.webkit.org/show_bug.cgi?id=231666
Patch by Sergey Rubanov <Sergey Rubanov> on 2021-12-25
Reviewed by Alexey Proskuryakov.
- metadata/contributors.json:
- 10:34 AM Changeset in webkit [287438] by
-
- 2 edits in trunk/Source/WebCore
[LFC][IFC] Fully collapsible whitespace content should always stay on the current line
https://bugs.webkit.org/show_bug.cgi?id=234667
Reviewed by Antti Koivisto.
Fully collapsible trailing whitespace content (normal/nowrap/preline) gets removed at
the trimming phase so it always "fits" the current line.
This is also in preparation for decoupling removeable and hanging trailing content.
- layout/formattingContexts/inline/InlineContentBreaker.cpp:
(WebCore::Layout::InlineContentBreaker::processOverflowingContent const):
- 7:36 AM Changeset in webkit [287437] by
-
- 5 edits in trunk/Source/WebCore
[LFC][IFC] InlineContentBreaker should know about the hanging content width
https://bugs.webkit.org/show_bug.cgi?id=234662
Reviewed by Antti Koivisto.
Pass in the trailing hanging content width to InlineContentBreaker as this specific type
of content (e.g. pre-wrap whitespace) should be ignored when checking for content fit.
(Note that we already ignore it through the shouldKeepEndOfLineWhitespace() call in
InlineContentBreaker::processOverflowingContent, but this patch is in preparation for making hanging content logic
more inline with the spec. -and also removing shouldKeepEndOfLineWhitespace needs some more changes.)
- layout/formattingContexts/inline/InlineContentBreaker.cpp:
(WebCore::Layout::InlineContentBreaker::processOverflowingContent const):
- layout/formattingContexts/inline/InlineContentBreaker.h:
- layout/formattingContexts/inline/InlineLine.h:
(WebCore::Layout::Line::hangingTrailingContentWidth const):
(WebCore::Layout::Line::hangingWhitespaceWidth const): Deleted.
- layout/formattingContexts/inline/InlineLineBuilder.cpp:
(WebCore::Layout::LineBuilder::layoutInlineContent):
(WebCore::Layout::LineBuilder::handleInlineContent):
- 2:13 AM Changeset in webkit [287436] by
-
- 3 edits in trunk/Source/WebKit
Unreviewed, reverting r285088.
https://bugs.webkit.org/show_bug.cgi?id=234680
From A/B test, we probably need to revert both, 285087 and
285088
Reverted changeset:
"Web process shouldn't crash if
ImageBuffer::ensureBackendCreated() fails"
https://bugs.webkit.org/show_bug.cgi?id=232520
https://commits.webkit.org/r285088
