Timeline
Jul 11, 2014:
- 10:08 PM Changeset in webkit [171020] by
-
- 4 edits in trunk/Source/WebCore
Partition the CSS rules based on the most specific filter of the rightmost fragment
https://bugs.webkit.org/show_bug.cgi?id=134828
Reviewed by Andreas Kling.
Previously, RuleSet was partitioning each rule based on the rightmost filter.
While fast, this had the side effect of putting many selectors with ID match in the class
bucket (because the selectors are generally written starting with the ID).
This patch replace the code of findBestRuleSetAndAdd() by a simple loop going over all
the simple selectors in the rightmost fragment to find the best bucket.
- css/ElementRuleCollector.cpp:
(WebCore::ElementRuleCollector::ruleMatches):
- css/RuleSet.cpp:
(WebCore::isSelectorMatchingHTMLBasedOnRuleHash):
I unified ruleData.hasRightmostSelectorMatchingHTMLBasedOnRuleHash() and hasMultipartSelector().
(WebCore::RuleData::RuleData):
(WebCore::rulesCountForName):
(WebCore::RuleSet::addRule):
I removed the recursive part of findBestRuleSetAndAdd() (which was wrong anyway). The function
was useless so I just moved the algorithm to addRule() directly.
We first loop over all the CSSSelectors related by SubSelector, this correspond to the rightmost fragment.
If a filter with high specificity is found, we add the rule immediately and end there.
If a filter that is not very specific is found, we keep a pointer to the selector to use it later.
(WebCore::RuleSet::findBestRuleSetAndAdd): Deleted.
- css/RuleSet.h:
(WebCore::RuleData::hasMultipartSelector): Deleted.
- 8:54 PM Changeset in webkit [171019] by
-
- 2 edits in trunk/Source/WebCore
[WinCairo] Unreviewed build fix after r170937.
- WebCore.vcxproj/WebCoreGeneratedWinCairo.make:
Update and apply changes applied to WebCoreGenerated.make in r170937.
- 6:49 PM Changeset in webkit [171018] by
-
- 2 edits in trunk/Source/WebCore
Web Inspector: Crash when using a stale InspectableNode Node
https://bugs.webkit.org/show_bug.cgi?id=134849
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2014-07-11
Reviewed by Timothy Hatcher.
- inspector/PageConsoleAgent.cpp:
- 5:53 PM Changeset in webkit [171017] by
-
- 5 edits in trunk/Source/WebCore
[MSE] Separate MediaSource logging into its own log channel
https://bugs.webkit.org/show_bug.cgi?id=134809
Reviewed by Eric Carlson.
Separate out MediaSource logging (which can be quite verbose) into its own log channel,
distinct from the Media log channel. Add some per-sample logging and further sequester that
logging into its own MediaSourceSamples log channel.
Add the MediaSource and MediaSourceSamples channels:
- platform/Logging.h:
Move LOG(Media) -> LOG(MediaSource):
- Modules/mediasource/MediaSource.cpp:
(WebCore::MediaSource::MediaSource):
(WebCore::MediaSource::~MediaSource):
(WebCore::MediaSource::setReadyState):
(WebCore::MediaSource::addSourceBuffer):
(WebCore::MediaSource::removeSourceBuffer):
(WebCore::MediaSource::isTypeSupported):
- Modules/mediasource/SourceBuffer.cpp:
(WebCore::SourceBuffer::sourceBufferPrivateSeekToTime):
(WebCore::SourceBuffer::sourceBufferPrivateAppendComplete):
(WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample):
(WebCore::SourceBuffer::sourceBufferPrivateDidBecomeReadyForMoreSamples):
(WebCore::SourceBuffer::provideMediaData):
(WebCore::SourceBuffer::monitorBufferingRate):
(WebCore::SourceBuffer::textTrackAddCues):
(WebCore::SourceBuffer::hasFutureTime):
- platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:
(WebCore::SourceBufferPrivateAVFObjC::didParseStreamDataAsAsset):
(WebCore::SourceBufferPrivateAVFObjC::didFailToParseStreamDataWithError):
(WebCore::SourceBufferPrivateAVFObjC::processCodedFrame):
(WebCore::SourceBufferPrivateAVFObjC::didProvideContentKeyRequestInitializationDataForTrackID):
(WebCore::SourceBufferPrivateAVFObjC::append):
(WebCore::SourceBufferPrivateAVFObjC::layerDidReceiveError):
(WebCore::SourceBufferPrivateAVFObjC::rendererDidReceiveError):
(WebCore::SourceBufferPrivateAVFObjC::flushAndEnqueueNonDisplayingSamples):
(WebCore::SourceBufferPrivateAVFObjC::enqueueSample):
- 4:54 PM Changeset in webkit [171016] by
-
- 2 edits in trunk/Source/WebCore
Phone numbers that span two lines are not detected.
<rdar://problem/17601146> and https://bugs.webkit.org/show_bug.cgi?id=134808
Reviewed by Tim Horton.
- editing/Editor.cpp:
(WebCore::Editor::scanSelectionForTelephoneNumbers): After scanning a range from the TextIterator,
create an "edge range" window around the end of the TextIterator range, and scan it.
Also make sure to not accumulate duplicate ranges that might have showed up in both the
TextIterator range and the edge window range.
- 4:24 PM Changeset in webkit [171015] by
-
- 8 edits in trunk/Source
Implement textStylingAtPosition in WK2.
https://bugs.webkit.org/show_bug.cgi?id=134843
<rdar://problem/17614981>
Source/WebCore:
Reviewed by Benjamin Poulain.
Adding some exports and making styleForSelectionStart public.
- WebCore.exp.in:
- editing/Editor.h:
Source/WebKit2:
Reviewed by Benjamin Poulain.
Adding information about typing attributes to EditorState so
that we can implement textStylingAtPosition.
- Shared/EditorState.cpp:
(WebKit::EditorState::encode):
(WebKit::EditorState::decode):
- Shared/EditorState.h:
(WebKit::EditorState::EditorState):
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView textStylingAtPosition:inDirection:]):
(-[WKContentView canPerformAction:withSender:]):
(-[WKContentView toggleBoldface:]):
(-[WKContentView toggleItalics:]):
(-[WKContentView toggleUnderline:]):
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::editorState):
- 3:36 PM Changeset in webkit [171014] by
-
- 9 edits2 adds in trunk
REGRESSION (r168868): eBay 'see all' links fail due to different JS bindings conversion behavior.
https://bugs.webkit.org/show_bug.cgi?id=134841
<rdar://problem/17577563>
Reviewed by Simon Fraser.
Revert back to type int for Element.scroll* APIs in order to preserve exception handling
behavior.
Source/WebCore:
Test: cssom/non-subpixel-scroll-top-left-values.html
- dom/Element.cpp:
(WebCore::Element::scrollLeft):
(WebCore::Element::scrollTop):
(WebCore::Element::setScrollLeft):
(WebCore::Element::setScrollTop):
(WebCore::Element::scrollWidth):
(WebCore::Element::scrollHeight):
- dom/Element.h:
- dom/Element.idl:
- html/HTMLBodyElement.cpp:
(WebCore::adjustForZoom):
(WebCore::HTMLBodyElement::scrollLeft):
(WebCore::HTMLBodyElement::setScrollLeft):
(WebCore::HTMLBodyElement::scrollTop):
(WebCore::HTMLBodyElement::setScrollTop):
(WebCore::HTMLBodyElement::scrollHeight):
(WebCore::HTMLBodyElement::scrollWidth):
- html/HTMLBodyElement.h:
LayoutTests:
- cssom/non-subpixel-scroll-top-left-values-expected.txt: Added.
- cssom/non-subpixel-scroll-top-left-values.html: Added.
- 2:54 PM Changeset in webkit [171013] by
-
- 9 edits in trunk/Source
Web Inspector: Debugger Pause button does not work
https://bugs.webkit.org/show_bug.cgi?id=134785
Reviewed by Timothy Hatcher.
Source/JavaScriptCore:
- CMakeLists.txt:
- DerivedSources.make:
Minification strips the sourceURL command. Add it back with minification.
Source/WebCore:
- CMakeLists.txt:
- DerivedSources.make:
- inspector/CommandLineAPIModuleSource.js:
Minification strips the sourceURL command. Add it back with minification.
Source/WebInspectorUI:
Previously we were automatically resuming if we had no recognized
call frames in the call stack we get when paused. However, we
were accidentally skipping over anonymous scripts instead of
just Web Inspector originated call frames.
Now that we allow anonymous scripts we can step into anonymous
scripts (e.g. evals, new Function(...)) as expected.
Also, remove the condition of automatically resuming if the
call stack was empty. The UI handles this with "No Call Frames".
- UserInterface/Controllers/DebuggerManager.js:
(WebInspector.DebuggerManager.prototype.debuggerDidPause):
- 2:10 PM Changeset in webkit [171012] by
-
- 2 edits in trunk/Source/WebCore
Use the bare minimum tile coverage rect when under memory pressure.
<https://webkit.org/b/134837>
When the browser is under critical memory pressure, don't generate any
more tiles than are needed to cover the exposed viewport rect.
Reviewed by Pratik Solanki.
- page/FrameView.cpp:
(WebCore::FrameView::computeCoverageRect):
- 2:04 PM Changeset in webkit [171011] by
-
- 2 edits in trunk/Source/WebInspectorUI
Web Inspector: CSS Autocompletion: "background" only has colors, should support others
https://bugs.webkit.org/show_bug.cgi?id=134839
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2014-07-11
Reviewed by Timothy Hatcher.
Update and consolidate the background related CSS properties
and include sub-property completions in the shorthand property.
Also take the opportunity to add or update other properties
(-webkit-alt, -webkit-text-decoration-skip, -webkit-aspect-ratio)
and remove unnecessary newlines / commas in the list.
- UserInterface/Models/CSSKeywordCompletions.js:
- 1:26 PM Changeset in webkit [171010] by
-
- 13 edits2 adds in trunk
[CSS Grid Layout] Implement justify-self css property
https://bugs.webkit.org/show_bug.cgi?id=134419
Reviewed by Dean Jackson.
This change adds the justify-self property from CSS 3 Box Alignment
and implements the parsing.
From Blink r164685 by <jchaffraix@chromium.org>
Source/WebCore:
Test: fast/css/parse-justify-self.html
- css/CSSComputedStyleDeclaration.cpp:
(WebCore::ComputedStyleExtractor::propertyValue):
- css/CSSParser.cpp:
(WebCore::isValidKeywordPropertyAndValue):
(WebCore::CSSParser::parseValue):
(WebCore::isItemPositionKeyword):
(WebCore::CSSParser::parseJustifySelf):
- css/CSSParser.h:
- css/CSSPrimitiveValueMappings.h:
(WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
(WebCore::CSSPrimitiveValue::operator EJustifySelf):
(WebCore::CSSPrimitiveValue::operator EJustifySelfOverflowAlignment):
- css/CSSPropertyNames.in:
- css/CSSValueKeywords.in:
- css/DeprecatedStyleBuilder.cpp:
(WebCore::DeprecatedStyleBuilder::DeprecatedStyleBuilder):
- css/StyleResolver.cpp:
(WebCore::StyleResolver::applyProperty):
- rendering/style/RenderStyle.h:
- rendering/style/RenderStyleConstants.h:
- rendering/style/StyleRareNonInheritedData.cpp:
(WebCore::StyleRareNonInheritedData::StyleRareNonInheritedData):
(WebCore::StyleRareNonInheritedData::operator==):
- rendering/style/StyleRareNonInheritedData.h:
LayoutTests:
- fast/css/parse-justify-self-expected.txt: Added.
- fast/css/parse-justify-self.html: Added.
- 12:47 PM Changeset in webkit [171009] by
-
- 6 edits in tags/Safari-538.46
Rollout r170675. <rdar://problem/17643689>.
- 12:00 PM Changeset in webkit [171008] by
-
- 2 edits in trunk/Source/WebCore
[Feature Queries] Feature Query CSS Grammar Productions Should Return a Value
https://bugs.webkit.org/show_bug.cgi?id=134810
Reviewed by Antti Koivisto.
The parsing return values are used to determine if a feature query is valid. Two
grammar productions have been updated to correctly return a value.
Tests exist for this feature but will not be run until the feature is turned on.
See https://bugs.webkit.org/show_bug.cgi?id=134404.
- css/CSSGrammar.y.in: Adding return values for two valid feature query productions.
- 11:56 AM Changeset in webkit [171007] by
-
- 2 edits in tags/Safari-538.46/Source/WebKit2
Merged r171003. <rdar://problem/17642184>
- 11:54 AM Changeset in webkit [171006] by
-
- 10 edits in tags/Safari-538.46
Merged r171001. <rdar://problem/17435776>
- 11:53 AM Changeset in webkit [171005] by
-
- 7 edits in trunk/Source
[Win] Enable DFG JIT.
https://bugs.webkit.org/show_bug.cgi?id=123615
Patch by peavo@outlook.com <peavo@outlook.com> on 2014-07-11
Reviewed by Mark Lam.
Source/JavaScriptCore:
When the return type of a JIT generated function call is larger than 64-bit (e.g. SlowPathReturnType),
the normal call() implementation cannot be used on 64-bit Windows, because the 64-bit Windows ABI is different in this case.
Also, when generating calls with double arguments, we need to make sure the arguments are put in the correct registers,
since the register allocation differs on 64-bit Windows.
- assembler/MacroAssemblerX86_64.h:
(JSC::MacroAssemblerX86_64::callWithSlowPathReturnType): Added method to handle function calls where the return value type size is larger than 64-bit.
- jit/CCallHelpers.h:
(JSC::CCallHelpers::setupArgumentsWithExecState): Move arguments to correct registers when there are floating point arguments.
(JSC::CCallHelpers::setupArgumentsWithExecStateForCallWithSlowPathReturnType): Added method.
- jit/JIT.h:
(JSC::JIT::appendCallWithSlowPathReturnType): Added method.
- jit/JITInlines.h:
(JSC::JIT::appendCallWithExceptionCheckAndSlowPathReturnType): Added method.
(JSC::JIT::callOperation): Call new method.
Source/WTF:
- wtf/Platform.h:
- 10:32 AM Changeset in webkit [171004] by
-
- 11 edits in trunk
Use a separate backdrop element to allow cues to have highlight and background color
https://bugs.webkit.org/show_bug.cgi?id=134821
<rdar://problem/15999721>
Reviewed by Eric Carlson.
Source/WebCore:
Add a new <div> element wrapping the existing cue <span>. This allows
us to have a highlight on the cue (in the <span> background), as well
as an overall background color.
- Modules/mediacontrols/mediaControlsApple.css:
(video::-webkit-media-text-track-display-backdrop): New markup for
the backdrop element of the caption.
- html/track/VTTCue.cpp:
(WebCore::VTTCue::cueBackdropShadowPseudoId): Added to
allow user customization of the cue backdrop.
(WebCore::VTTCue::initialize): Rename the old "m_cueBackgroundBox" to
"m_cueHighlightBox" and add a new "m_cueBackdropBox" member.
(WebCore::VTTCue::updateDisplayTree): Update for m_cueHighlightBox.
(WebCore::VTTCue::getDisplayTree): Make m_cueHighlightBox a child
of the new m_cueBackdropBox element, and add m_cueBackdropBox to
the display tree.
- html/track/VTTCue.h:
(WebCore::VTTCue::element):
- page/CaptionUserPreferencesMediaAF.cpp:
(WebCore::CaptionUserPreferencesMediaAF::setInterestedInCaptionPreferenceChanges):
Fix for missing caption style updates. Even if we are already
listening for caption changes, we still want to update the new
instance's style sheet to match.
(WebCore::CaptionUserPreferencesMediaAF::captionsStyleSheetOverride):
- rendering/RenderVTTCue.cpp:
(WebCore::RenderVTTCue::initializeLayoutParameters): Take the new
<div> into consideration when looking for the Cue text element.
LayoutTests:
Updated tests for new formatting logic.
- platform/mac/media/track/track-cue-rendering-horizontal-expected.png:
- platform/mac/media/track/track-cue-rendering-horizontal-expected.txt:
- 10:15 AM Changeset in webkit [171003] by
-
- 2 edits in trunk/Source/WebKit2
Tighten WebContent sandbox
https://bugs.webkit.org/show_bug.cgi?id=134834
Reviewed by Sam Weinig.
Define a much tighter sandbox profile for the WebContent process
- Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb:
- 9:29 AM Changeset in webkit [171002] by
-
- 2 edits in trunk/Source/WebKit2
REGRESSION (r170163?): Web content shifts revealing space equivalent to the find bar when clicking a link while a phrase is targeted via Cmd+F
https://bugs.webkit.org/show_bug.cgi?id=134833
<rdar://problem/17580021>
Reviewed by Zalan Bujtas.
Some versions of OS X Safari can't handle the new unfreeze timing. Revert back to DidFirstLayout on Mac.
- WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::dispatchDidLayout):
- 8:27 AM Changeset in webkit [171001] by
-
- 10 edits in trunk
Subpixel layout: return integral results for offset*, client*, scroll* by default.
https://bugs.webkit.org/show_bug.cgi?id=134651
Reviewed by Simon Fraser.
Revert to returning integral values for Element.offset* client* scroll* by default.
Fractional values break number of sites(tight design) and JS frameworks(fail to handle fractional values).
Since snapped dimension depends on both the original point and the width/height of the box,
we need to call RenderBoxModelObject::pixelSnapped*() helpers, instead of round().
Covered by existing tests
Source/WebCore:
- dom/Element.cpp:
(WebCore::subpixelMetricsEnabled):
(WebCore::convertToNonSubpixelValueIfNeeded):
(WebCore::Element::offsetLeft):
(WebCore::Element::offsetTop):
(WebCore::Element::offsetWidth):
(WebCore::Element::offsetHeight):
(WebCore::Element::clientLeft):
(WebCore::Element::clientTop):
(WebCore::Element::clientWidth):
(WebCore::Element::clientHeight):
- page/Settings.in:
Source/WebKit/mac:
- WebView/WebPreferences.mm:
(+[WebPreferences initialize]):
Source/WebKit2:
- Shared/WebPreferencesDefinitions.h:
- UIProcess/API/C/WKPreferencesRefPrivate.h:
LayoutTests:
- cssom/subpixel-offsetleft-top-width-height-values-expected.txt:
- 7:19 AM Changeset in webkit [171000] by
-
- 4 edits2 adds in trunk
Subpixel rendering: icloud.com password arrow has clipped circle at some window sizes.
https://bugs.webkit.org/show_bug.cgi?id=134824
<rdar://problem/17612729>
Reviewed by Simon Fraser.
Transparent layer clipping needs device pixel offset adjusting.
Source/WebCore:
Test: fast/layers/wrong-clipping-semi-transparent-compositing-layer-on-subpixel-position.html
- rendering/RenderLayer.cpp:
(WebCore::RenderLayer::beginTransparencyLayers):
(WebCore::RenderLayer::paintLayer):
(WebCore::RenderLayer::paintLayerContents):
(WebCore::RenderLayer::paintBackgroundForFragments):
(WebCore::RenderLayer::paintForegroundForFragments):
- rendering/RenderLayer.h:
LayoutTests:
- fast/layers/wrong-clipping-semi-transparent-compositing-layer-on-subpixel-position-expected.html: Added.
- fast/layers/wrong-clipping-semi-transparent-compositing-layer-on-subpixel-position.html: Added.
- 5:26 AM Changeset in webkit [170999] by
-
- 2 edits in trunk/Source/WTF
Unreviewed, rolling out r170995.
https://bugs.webkit.org/show_bug.cgi?id=134831
Causing odd crashes in debug builds (Requested by zdobersek on
#webkit).
Reverted changeset:
"[WTF] Add the move constructor, move assignment operator for
HashTable"
https://bugs.webkit.org/show_bug.cgi?id=130772
http://trac.webkit.org/changeset/170995
- 5:15 AM Changeset in webkit [170998] by
-
- 1 edit1 delete in trunk/Source/WebCore
Unreviewed. Remove unused file.
- html/shadow/MediaControlsGtk.h: Removed.
- 5:12 AM Changeset in webkit [170997] by
-
- 2 edits in trunk
[GTK] Enable VIDEO_TRACK by default
https://bugs.webkit.org/show_bug.cgi?id=134801
Reviewed by Philippe Normand.
- Source/cmake/OptionsGTK.cmake:
- 3:42 AM Changeset in webkit [170996] by
-
- 33 edits in trunk
[CSS Grid Layout] Update grid-auto-flow to the new syntax
https://bugs.webkit.org/show_bug.cgi?id=134057
Reviewed by Sergio Villar Senin.
Source/WebCore:
In last versions of the spec grid-auto-flow syntax has changed. New
syntax is:
[ row | column ] && dense? | stack && [ row | column ]?
Implemented parsing for new syntax and added/modified test cases in
current layout tests.
For the moment, the implementation keeps working in the same way, but
using "stack" value for grid-auto-flow property instead of "none". This
should be fixed in a follow-up patch once "stack" is properly
implemented.
Also "dense" needs to be reviewed. Right now auto-placement algorithm is
always "dense" and never "sparse".
No new tests, update current tests to add new cases.
- css/CSSComputedStyleDeclaration.cpp:
(WebCore::ComputedStyleExtractor::propertyValue): Adapt to new syntax.
- css/CSSParser.cpp:
(WebCore::isValidKeywordPropertyAndValue): Removed grid-auto-flow as it
is not a keyword anymore.
(WebCore::isKeywordPropertyID): Ditto.
(WebCore::CSSParser::parseValue): Add specific method for parsing
grid-auto-flow.
(WebCore::CSSParser::parseGridShorthand): Adapt parsing for
grid-auto-flow property inside the shorthand.
(WebCore::isValidGridAutoFlowId): Helper method to check if the CSSValue
id is a valid keyword for grid-auto-flow property.
(WebCore::CSSParser::parseGridAutoFlow): Implement new parsing for
grid-auto-flow.
- css/CSSParser.h: Method header for grid-auto-flow parsing.
- css/CSSPrimitiveValueMappings.h:
(WebCore::CSSPrimitiveValue::operator GridAutoFlow): Deleted.
(WebCore::CSSPrimitiveValue::CSSPrimitiveValue): Deleted.
- css/CSSValueKeywords.in: Add new keywords required by grid-auto-flow:
"dense" and "stack".
- css/DeprecatedStyleBuilder.cpp:
(WebCore::DeprecatedStyleBuilder::DeprecatedStyleBuilder): Remove
default handler as grid-auto-flow is not a keyword now.
- css/StyleResolver.cpp:
(WebCore::StyleResolver::applyProperty): Implement particular handler
for new grid-auto-flow syntax.
- rendering/RenderGrid.cpp:
(WebCore::RenderGrid::placeItemsOnGrid): Keep old "none" behavior as
"stack" behavior for the moment.
(WebCore::RenderGrid::autoPlacementMajorAxisDirection): Use the new
helper methods in RenderStyle.
(WebCore::RenderGrid::autoPlacementMinorAxisDirection): Ditto.
- rendering/style/GridResolvedPosition.cpp:
(WebCore::GridResolvedPosition::resolveGridPositionsFromStyle): Again
keep old "none" behavior for "stack".
- rendering/style/RenderStyle.h: Add new helper methods to know the
direction and algorithm of grid-auto-flow property.
- rendering/style/RenderStyleConstants.h: Redefine GridAutoFlow enum
using flags for algorithm and direction.
- rendering/style/StyleGridData.h: Change m_gridAutoFlow type to
unsigned.
LayoutTests:
Adapted current tests to the new syntax. Keeping old "none" behavior
using "stack".
Add new use cases to cover the new syntax to grid-auto-flow-get-set.html
and grid-shorthand-get-set.html.
- fast/css-grid-layout/grid-auto-flow-get-set-expected.txt:
- fast/css-grid-layout/grid-auto-flow-get-set.html:
- fast/css-grid-layout/grid-auto-flow-resolution.html:
- fast/css-grid-layout/grid-auto-flow-update-expected.txt:
- fast/css-grid-layout/grid-auto-flow-update.html:
- fast/css-grid-layout/grid-item-named-grid-area-resolution.html:
- fast/css-grid-layout/grid-item-z-index-change-repaint.html:
- fast/css-grid-layout/grid-item-z-index-stacking-context.html:
- fast/css-grid-layout/grid-shorthand-get-set-expected.txt:
- fast/css-grid-layout/grid-shorthand-get-set.html:
- fast/css-grid-layout/named-grid-lines-with-named-grid-areas-dynamic-get-set.html:
- fast/css-grid-layout/named-grid-lines-with-named-grid-areas-resolution.html:
- fast/css-grid-layout/resources/grid.css:
(.gridAutoFlowStack):
(.gridAutoFlowNone): Deleted.
- fast/css/getComputedStyle/computed-style-expected.txt:
- fast/css/getComputedStyle/computed-style-without-renderer-expected.txt:
- ietestcenter/css3/grid/grid-column-001.htm:
- ietestcenter/css3/grid/grid-column-002.htm:
- ietestcenter/css3/grid/grid-column-003.htm:
- svg/css/getComputedStyle-basic-expected.txt:
- 1:16 AM Changeset in webkit [170995] by
-
- 2 edits in trunk/Source/WTF
[WTF] Add the move constructor, move assignment operator for HashTable
https://bugs.webkit.org/show_bug.cgi?id=130772
Reviewed by Darin Adler.
HashTable has both copy constructor and copy assignment operator, meaning that the move constructor
and move assignment operator are implicitly deleted. This patch defines both to avoid unnecessary
copies when moves can be performed.
- wtf/HashTable.h:
(WTF::KeyTraits>::HashTable):
Jul 10, 2014:
- 11:17 PM Changeset in webkit [170994] by
-
- 10 edits in tags/Safari-538.46/Source
Merge patch from <rdar://problem/17631537>.
- 11:01 PM Changeset in webkit [170993] by
-
- 5 edits in trunk/Source
Versioning.
- 10:57 PM Changeset in webkit [170992] by
-
- 1 copy in tags/Safari-538.46
New Tag.
- 10:06 PM Changeset in webkit [170991] by
-
- 3 edits in trunk/Source/WebInspectorUI
Web Inspector: Debugger Continue button is incorrectly sized
https://bugs.webkit.org/show_bug.cgi?id=134786
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2014-07-10
Reviewed by Timothy Hatcher.
Before this patch everything in the DebuggerDashboardView was flexing.
Normally, if the contents were not too wide, this would just behave
as expected and be centered. If the content was too wide, then each
element would flex uniformly and overlap uniformly. That was not expected.
This patch changes makes everything but the link have a non-flexible
size. If the link (at the end) flexs it will have a nice ellipsis
text truncation, and the rest of the content will be left justified.
This also fixes a few other minor issues seen in the view.
- UserInterface/Views/DebuggerDashboardView.css:
(.dashboard.debugger .navigation-bar):
Reduce the complexity, give the nav bar (the resume button) some simple padding.
(.dashboard.debugger > div):
Make everything non-flex by default.
(.dashboard.debugger > .location):
The location will have an inner element flex, so make it flex.
(.dashboard.debugger > .location > *):
Make everything in .location non-flex by default.
(.dashboard.debugger > .location .go-to-link):
Make the link flex and truncate cleanly.
(.dashboard.debugger > .location .function-name):
Fix vertical alignment when the toolbar has small icons.
- UserInterface/Views/DebuggerDashboardView.js:
(WebInspector.DebuggerDashboardView.prototype._rebuildLocation):
Fix class name spelling.
- 9:26 PM Changeset in webkit [170990] by
-
- 2 edits in trunk/Source/WebKit2
Unreviewed EFL build fix after r170970.
- UIProcess/efl/WebContextEfl.cpp:
(WebKit::WebContext::platformDefaultOpenGLCacheDirectory):
- 8:50 PM Changeset in webkit [170989] by
-
- 4 edits in trunk/Source/WebKit2
REGRESSION(r170970) REGRESSION(r170974): [GTK] Build broken.
https://bugs.webkit.org/show_bug.cgi?id=134825
Unreviewed GTK build fix.
Patch by Carlos Alberto Lopez Perez <clopez@igalia.com> on 2014-07-10
- Shared/SessionState.h: Put ifdefs for ViewSnapshot on Mac port.
- Shared/WebBackForwardListItem.h: Idem.
- UIProcess/gtk/WebContextGtk.cpp:
(WebKit::WebContext::platformDefaultOpenGLCacheDirectory): Implement skeleton.
- 7:29 PM Changeset in webkit [170988] by
-
- 2 edits in trunk/Source/WebKit2
[iOS][WK2] It should be safe to call WKContentViewInteraction's cleanupInteraction multiple times
https://bugs.webkit.org/show_bug.cgi?id=134820
Patch by Benjamin Poulain <bpoulain@apple.com> on 2014-07-10
Reviewed by Andreas Kling.
If a view is destroyed just after a crash, "cleanupInteraction" is called twice: once on crash,
once on dealloc.
The code handling _interactionViewsContainerView is using KVO to monitor transform changes. It is not safe
to remove the observer if we are not already observing on that view.
To solve the problem, this patch makes the cleanup actually remove the view so that setup and cleanup
are completely symmetrical. If cleanup is called twice, the second time would not enter the branch because
the view is already nil.
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView setupInteraction]):
(-[WKContentView cleanupInteraction]):
- 5:49 PM Changeset in webkit [170987] by
-
- 3 edits2 adds in trunk/Source/WebKit2
[iOS WK2] Move WKInspectorHighlightView to its own file
https://bugs.webkit.org/show_bug.cgi?id=134819
Reviewed by Joseph Pecoraro.
WKInspectorHighlightView brought a lot of path/quad-related code into WKContentView.mm,
so move it into its own file.
- UIProcess/WKInspectorHighlightView.h: Added.
- UIProcess/WKInspectorHighlightView.mm: Added.
(-[WKInspectorHighlightView dealloc]):
(-[WKInspectorHighlightView _removeAllLayers]):
(-[WKInspectorHighlightView _createLayers:]):
(findIntersectionOnLineBetweenPoints):
(quadIntersection):
(layerPathWithHole):
(layerPath):
(-[WKInspectorHighlightView _layoutForNodeHighlight:]):
(-[WKInspectorHighlightView _layoutForRectsHighlight:]):
(-[WKInspectorHighlightView update:]):
- UIProcess/ios/WKContentView.mm:
(-[WKInspectorHighlightView initWithFrame:]): Deleted.
(-[WKInspectorHighlightView dealloc]): Deleted.
(-[WKInspectorHighlightView _removeAllLayers]): Deleted.
(-[WKInspectorHighlightView _createLayers:]): Deleted.
(findIntersectionOnLineBetweenPoints): Deleted.
(quadIntersection): Deleted.
(layerPathWithHole): Deleted.
(layerPath): Deleted.
(-[WKInspectorHighlightView _layoutForNodeHighlight:]): Deleted.
(-[WKInspectorHighlightView _layoutForRectsHighlight:]): Deleted.
(-[WKInspectorHighlightView update:]): Deleted.
- WebKit2.xcodeproj/project.pbxproj:
- 5:12 PM Changeset in webkit [170986] by
-
- 5 edits3 adds in trunk
[iOS] Pasting rich content does not perform a two-step paste
https://bugs.webkit.org/show_bug.cgi?id=134795
Reviewed by Benjamin Poulain.
Source/WebCore:
UIKit requires writing something (anything!) to a particular pasteboard
during copy in order to flag the content as rich. When a paste occurs,
UIKit will read this flag and perform a two-step paste.
Test: platform/ios-sim/editing/pasteboard/two-step-paste.html
- platform/ios/PlatformPasteboardIOS.mm:
(WebCore::PlatformPasteboard::write):
LayoutTests:
Copy and paste some text, and make sure that the two step paste occurred.
The mechanism for detecting this is whether or not styles have changed appropriately.
- platform/ios-sim/editing/pasteboard/two-step-paste-expected.txt: Added.
- platform/ios-sim/editing/pasteboard/two-step-paste.html: Added.
Tools:
Two-step paste requires setting InputViewObeysDOMFocus on the UIWebBrowserView.
- DumpRenderTree/mac/DumpRenderTree.mm:
(createWebViewAndOffscreenWindow):
- 5:04 PM Changeset in webkit [170985] by
-
- 1 edit in trunk/Source/WebCore/ChangeLog
Update the ChangeLog per Pratik's review comments
- 5:02 PM Changeset in webkit [170984] by
-
- 5 edits in trunk/Source
[iOS] Some QuickLook documents are not displayed as previews
https://bugs.webkit.org/show_bug.cgi?id=134788
Reviewed by Pratik Solanki.
ResourceRequest::isMainResourceRequest() is not reliable (e.g. ResourceRequests reconstituted from an
NSURLRequest have no way to set this bit properly) and should probably be removed. Instead of using
isMainResourceRequest() to check if a QuickLookHandle should be created when the Network Process is in use,
just check if the current ResourceLoader is the same as the DocumentLoader's ResourceLoader.
No new tests. QuickLook is not testable from WebKit.
- platform/network/ios/QuickLook.mm:
(WebCore::QuickLookHandle::create):
- 4:57 PM Changeset in webkit [170983] by
-
- 2 edits in tags/Safari-538.45.1/Source/WebKit2
Merged r170979. <rdar://problem/17632687>
- 4:46 PM Changeset in webkit [170982] by
-
- 2 edits in trunk/Source/WebKit2
REGRESSION (r170935): WKWebView is always transparent until the first layer tree commit
https://bugs.webkit.org/show_bug.cgi?id=134818
<rdar://problem/17632468>
Reviewed by Anders Carlsson.
- UIProcess/API/Cocoa/WKWebView.mm:
(scrollViewBackgroundColor):
r170935 made it so that we would initialize the scroll view background color to an
invalid color until the first layer tree commit. We should go with white instead.
- 4:26 PM Changeset in webkit [170981] by
-
- 12 edits in trunk/Source/WebKit2
Add a mechanism to notify the UIProcess when an editing command is done executing.
https://bugs.webkit.org/show_bug.cgi?id=134807
Reviewed by Tim Horton.
Some editing commands have an effect on some parts of the system that
run inside the UIProcess. A good example are the cursor movement commands
that require an update of the autocorrection/autosuggestion machinery.
This patch adds a way to reliably know when the command has been executed
in the WebProcess. A previous attempt at solving this problem was added in
r170858 and was partially reverted in r170948.
The change also removes the selectionWillChange notification added in r170858.
- UIProcess/PageClient.h:
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::editorStateChanged):
- UIProcess/WebPageProxy.h:
- UIProcess/ios/PageClientImplIOS.h:
- UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::selectionWillChange): Deleted.
- UIProcess/ios/WKContentViewInteraction.h:
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView executeEditCommandWithCallback:]):
(-[WKContentView _moveUp:withHistory:]):
(-[WKContentView _moveDown:withHistory:]):
(-[WKContentView _moveLeft:withHistory:]):
(-[WKContentView _moveRight:withHistory:]):
(-[WKContentView _moveToStartOfWord:withHistory:]):
(-[WKContentView _moveToStartOfParagraph:withHistory:]):
(-[WKContentView _moveToStartOfLine:withHistory:]):
(-[WKContentView _moveToStartOfDocument:withHistory:]):
(-[WKContentView _moveToEndOfWord:withHistory:]):
(-[WKContentView _moveToEndOfParagraph:withHistory:]):
(-[WKContentView _moveToEndOfLine:withHistory:]):
(-[WKContentView _moveToEndOfDocument:withHistory:]):
(-[WKContentView _selectionWillChange]): Deleted.
- UIProcess/ios/WebPageProxyIOS.mm:
(WebKit::WebPageProxy::executeEditCommand):
(WebKit::WebPageProxy::notifySelectionWillChange): Deleted.
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/WebPage.messages.in:
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::executeEditCommandWithCallback):
- 4:20 PM Changeset in webkit [170980] by
-
- 4 edits in trunk/Source
[Mac] NSWindow warning: adding an unknown subview opening detached Inspector
https://bugs.webkit.org/show_bug.cgi?id=134813
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2014-07-10
Reviewed by Timothy Hatcher.
Source/WebKit/mac:
- WebCoreSupport/WebInspectorClient.mm:
(-[WebInspectorWindowController window]):
Use a selector that will avoid the warning message.
Source/WebKit2:
- UIProcess/mac/WebInspectorProxyMac.mm:
(WebKit::WebInspectorProxy::createInspectorWindow):
Use a selector that will avoid the warning message.
- 4:17 PM Changeset in webkit [170979] by
-
- 2 edits in trunk/Source/WebKit2
Remove use of container relative restrictions in the network process sandbox
https://bugs.webkit.org/show_bug.cgi?id=134816
Reviewed by Anders Carlsson.
As i'm tidying up the various sandboxes and that's meaning we
need to reduce some file restrictions in the network process.
- Resources/SandboxProfiles/ios/com.apple.WebKit.Networking.sb:
- 3:26 PM Changeset in webkit [170978] by
-
- 2 edits in trunk/Source/WebCore
[iOS WebKit2] Don't uninstall the memory pressure listener after first signal.
<https://webkit.org/b/134814>
The call to uninstall() should not be made on iOS since iOS doesn't use the
hold-off mechanism in MemoryPressureHandler.
Reviewed by Benjamin Poulain.
- platform/cocoa/MemoryPressureHandlerCocoa.mm:
(WebCore::MemoryPressureHandler::respondToMemoryPressure):
- 2:35 PM Changeset in webkit [170977] by
-
- 2 edits in trunk/Source/WebKit2
Unreviewed iOS build fix after r170974. Define id if building a non ObjC file.
- UIProcess/mac/ViewSnapshotStore.h:
- 2:28 PM Changeset in webkit [170976] by
-
- 5 edits in trunk/Source/WebCore
[MSE] Overlapping appended ranges must cause the decoder to flush and re-enqueue.
https://bugs.webkit.org/show_bug.cgi?id=134805
Reviewed by Eric Carlson.
When appending a range of media data which overlaps with an existing range, SourceBuffer
must cause the decoder to flush and re-enqueue samples. Those samples removed by appending
the overlapping range may have already been enqueued for display, and unless the decode
queue is flushed, corruption or decode errors may occur as the new samples are enqueued for
display.
Add a boolean flag onto TrackBuffer to indicate whether the decoder needs to be flushed and
re-enqueued the next time an append operation completes. Set this flag whenever samples are
removed due to overlapping or an explicit call to removeCodedFrames(). Move the contents of
sourceBufferPrivateSeekToTime() (which previously did flushing and re-enqueueing) into a new
function, reenqueueMediaForTime(), which can be called from sourceBufferPrivateAppendComplete().
Drive-by fix: findSyncSampleAfterDecodeIterator() would return the passed in iterator if that
sample is a sync-sample. Fix this to correctly return the next sync sample.
Drive-by fix: Use a SampleMap, rather than a DecodeOrderSampleMap, to track erased samples
so that the erasedSamples are correctly accounted for in both presentation and decode orders.
- Modules/mediasource/SampleMap.cpp:
(WebCore::SampleMap::empty): Add convenience method.
(WebCore::DecodeOrderSampleMap::findSyncSampleAfterDecodeIterator): Drive-by fix.
- Modules/mediasource/SampleMap.h:
(WebCore::SampleMap::decodeOrder): Added const accessor.
(WebCore::SampleMap::presentationOrder): Ditto.
(WebCore::SampleMap::addRange): Added.
- Modules/mediasource/SourceBuffer.cpp:
(WebCore::SourceBuffer::TrackBuffer::TrackBuffer): Add needsReenqueeing flag.
(WebCore::SourceBuffer::sourceBufferPrivateSeekToTime): Move contents into reenqueueMediaForTime().
(WebCore::SourceBuffer::sourceBufferPrivateAppendComplete): Call reenqueMediaForTime() if necessary.
(WebCore::SourceBuffer::removeCodedFrames): Set needsReenqueing.
(WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample): Ditto.
(WebCore::SourceBuffer::provideMediaData): Drive-by fix.
(WebCore::SourceBuffer::reenqueueMediaForTime): Moved from sourceBufferPrivateSeekToTime().
- Modules/mediasource/SourceBuffer.h:
- 1:48 PM Changeset in webkit [170975] by
-
- 2 edits in trunk/Source/WebCore
ASSERT in SharedBuffer::maybeAppendDataArray() on MobileSafari launch
https://bugs.webkit.org/show_bug.cgi?id=134812
<rdar://problem/17628434>
Reviewed by Joseph Pecoraro.
Fix bug in my fix in r170930. Initialize the badly named m_shouldUsePurgeableMemory field to
false. This field indicates when it is okay to use purgeable memory and is set to true once
the resource is finished loading. By setting it to true in the constructor we were creating
purgeable memory while the resource was still being loaded and this triggered the assert.
No new tests. Should be covered by existing tests.
- platform/cf/SharedBufferCF.cpp:
(WebCore::SharedBuffer::SharedBuffer):
- 1:33 PM Changeset in webkit [170974] by
-
- 20 edits in trunk/Source/WebKit2
Store ViewSnapshots directly on the WebBackForwardListItem
https://bugs.webkit.org/show_bug.cgi?id=134667
<rdar://problem/17082639>
Reviewed by Dan Bernstein.
Make ViewSnapshot a refcounted class. Store it directly on the back-forward item
instead of in a side map referenced by UUID. Switch to a very simple LRU eviction model for now.
This fixes a ton of snapshot management bugs; for example, we would start throwing out snapshots
in the page that was actively being interacted with *first* when evicting snapshots, instead of
preferring older snapshots. Additionally, we would not throw away snapshots when back forward items
became unreachable.
There is definitely room for improvement of the eviction mechanism, but this is closer to a time-tested implementation.
- Shared/SessionState.h:
Keep a ViewSnapshot instead of a UUID on the BackForwardListItemState.
- Shared/WebBackForwardListItem.h:
Fix some indented namespace contents.
(WebKit::WebBackForwardListItem::snapshot):
(WebKit::WebBackForwardListItem::setSnapshot):
(WebKit::WebBackForwardListItem::setSnapshotUUID): Deleted.
(WebKit::WebBackForwardListItem::snapshotUUID): Deleted.
Switch the snapshot getter/setter to operate on ViewSnapshots instead of UUIDs.
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _takeViewSnapshot]):
- UIProcess/API/Cocoa/WKWebViewInternal.h:
- UIProcess/API/mac/WKView.mm:
(-[WKView _takeViewSnapshot]):
- UIProcess/API/mac/WKViewInternal.h:
- UIProcess/PageClient.h:
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::takeViewSnapshot):
- UIProcess/WebPageProxy.h:
- UIProcess/ios/PageClientImplIOS.h:
- UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::takeViewSnapshot):
- UIProcess/mac/PageClientImpl.h:
- UIProcess/mac/PageClientImpl.mm:
(WebKit::PageClientImpl::takeViewSnapshot):
Adopt ViewSnapshot::create, return a PassRefPtr, and class-ify ViewSnapshot.
- UIProcess/ios/ViewGestureControllerIOS.mm:
(WebKit::ViewGestureController::beginSwipeGesture):
(WebKit::ViewGestureController::endSwipeGesture):
- UIProcess/mac/ViewGestureController.h:
- UIProcess/mac/ViewGestureControllerMac.mm:
(WebKit::ViewGestureController::shouldUseSnapshotForSize):
(WebKit::ViewGestureController::beginSwipeGesture):
(WebKit::ViewGestureController::endSwipeGesture):
Grab the ViewSnapshot directly from the WebBackForwardListItem, and adopt the new functions.
- UIProcess/ios/WebMemoryPressureHandlerIOS.mm:
(WebKit::WebMemoryPressureHandler::WebMemoryPressureHandler):
Rename discardSnapshots to discardSnapshotImages, because we're really only discarding
the images; the render tree size/background color "snapshot" remains and is useful.
- UIProcess/mac/ViewSnapshotStore.h:
(WebKit::ViewSnapshot::setRenderTreeSize):
(WebKit::ViewSnapshot::renderTreeSize):
(WebKit::ViewSnapshot::setBackgroundColor):
(WebKit::ViewSnapshot::backgroundColor):
(WebKit::ViewSnapshot::setDeviceScaleFactor):
(WebKit::ViewSnapshot::deviceScaleFactor):
(WebKit::ViewSnapshot::imageSizeInBytes):
(WebKit::ViewSnapshot::surface):
(WebKit::ViewSnapshot::size):
(WebKit::ViewSnapshot::creationTime):
Make ViewSnapshot a refcounted class.
Add create functions which take an image (or slot ID), and relevant sizes.
It is expected that a ViewSnapshot is created with an image, and it is only possible
to remove that image, never to replace it. A new ViewSnapshot is required in that case.
Add setters for things that ViewSnapshotStore sets on the snapshot after the PageClient
retrieves it from the view. Add getters for things that the ViewGestureControllers need.
Remove removeSnapshotImage, getSnapshot, and the snapshot map.
- UIProcess/mac/ViewSnapshotStore.mm:
(WebKit::ViewSnapshotStore::~ViewSnapshotStore):
(WebKit::ViewSnapshotStore::didAddImageToSnapshot):
(WebKit::ViewSnapshotStore::willRemoveImageFromSnapshot):
Manage m_snapshotCacheSize and m_snapshotsWithImages via didAddImageToSnapshot and willRemoveImageFromSnapshot.
willRemoveImageFromSnapshot will -always- be called before the ViewSnapshot is destroyed.
(WebKit::ViewSnapshotStore::pruneSnapshots):
Switch to a simple LRU eviction model. As previously mentioned, it's possible to do better, but
this is much less broken than the previous implementation.
(WebKit::ViewSnapshotStore::recordSnapshot):
(WebKit::ViewSnapshotStore::discardSnapshotImages):
(WebKit::ViewSnapshot::create):
(WebKit::ViewSnapshot::ViewSnapshot):
(WebKit::ViewSnapshot::~ViewSnapshot):
(WebKit::ViewSnapshot::hasImage):
(WebKit::ViewSnapshot::clearImage):
(WebKit::ViewSnapshot::asLayerContents):
If a surface is Empty when it comes back from being volatile, throw away the surface
and notify the Store to remove it from m_snapshotCacheSize (via clearImage()).
(WebKit::ViewSnapshotStore::removeSnapshotImage): Deleted.
(WebKit::ViewSnapshotStore::getSnapshot): Deleted.
(WebKit::ViewSnapshotStore::discardSnapshots): Deleted.
- 1:24 PM Changeset in webkit [170973] by
-
- 4 edits in trunk/Source/WebCore
[iOS WebKit2] Some memory pressure relief tweaks.
<https://webkit.org/b/134811>
Split memory pressure relief into critical and non-critical sections.
Non-critical relief is for clearing out things that are really not
essential, e.g unused font data, text measurement caches, etc.
On iOS, only flip the "WebKit is under memory pressure" flag when we
are under *critical* memroy pressure, rather than doing it early on
and gimping ourselves because other processes are too big.
Also added logging for when we transition in/out of system pressure.
Reviewed by Geoffrey Garen.
- platform/MemoryPressureHandler.cpp:
(WebCore::MemoryPressureHandler::releaseNoncriticalMemory):
(WebCore::MemoryPressureHandler::releaseCriticalMemory):
(WebCore::MemoryPressureHandler::releaseMemory):
- platform/MemoryPressureHandler.h:
(WebCore::MemoryPressureHandler::ReliefLogger::loggingEnabled):
- platform/cocoa/MemoryPressureHandlerCocoa.mm:
(WebCore::MemoryPressureHandler::install):
- 1:04 PM Changeset in webkit [170972] by
-
- 16 edits in trunk/Source
Need Setting/WKPreference that allows clients to prevent scrollbars from drawing
on a secondary thread
https://bugs.webkit.org/show_bug.cgi?id=134778
-and corresponding-
<rdar://problem/17595333>
Reviewed by Tim Horton.
Source/WebCore:
This is a requirement for some types of performance tests. The patch adds a new
virtual function to ScrollableArea that forces subclasses to indicate the value of
the Setting. This is required because Scrollbar and ScrollableArea can’t get to
Settings on their own.
- page/FrameView.cpp:
(WebCore::FrameView::forceUpdateScrollbarsOnMainThreadForPerformanceTesting):
- page/FrameView.h:
- page/Settings.in:
- platform/ScrollableArea.h:
- platform/Scrollbar.cpp:
(WebCore::Scrollbar::supportsUpdateOnSecondaryThread):
- platform/win/PopupMenuWin.h:
- rendering/RenderLayer.cpp:
(WebCore::RenderLayer::forceUpdateScrollbarsOnMainThreadForPerformanceTesting):
- rendering/RenderLayer.h:
- rendering/RenderListBox.cpp:
(WebCore::RenderListBox::forceUpdateScrollbarsOnMainThreadForPerformanceTesting):
- rendering/RenderListBox.h:
Source/WebKit2:
This is a requirement for some types of performance tests.
New pref.
- Shared/WebPreferencesDefinitions.h:
PDFPlugin has to implement this new ScrollableArea virtual function to indicate
the Setting’s value.
- WebProcess/Plugins/PDF/PDFPlugin.h:
- WebProcess/Plugins/PDF/PDFPlugin.mm:
(WebKit::PDFPlugin::forceUpdateScrollbarsOnMainThreadForPerformanceTesting):
New pref.
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::updatePreferences):
- 12:31 PM Changeset in webkit [170971] by
-
- 2 edits in trunk/Source/WebKit2
[iOS] Frequent assertion failures when swiping back
Reviewed by Dan Bernstein.
- UIProcess/mac/RemoteLayerTreeDrawingAreaProxy.mm:
(WebKit::RemoteLayerTreeDrawingAreaProxy::dispatchAfterEnsuringDrawing):
Don't create an unused VoidCallback. It will assert when destroyed without being called.
- 11:33 AM Changeset in webkit [170970] by
-
- 7 edits in trunk/Source/WebKit2
Pass sandbox extension for GL cache over to webprocess
https://bugs.webkit.org/show_bug.cgi?id=134806
Reviewed by Anders Carlsson.
Add additional WebProcess parameters to pass an extension
that allows access to the opengl cache directory in the
host application's container.
- Shared/WebProcessCreationParameters.cpp:
(WebKit::WebProcessCreationParameters::encode):
(WebKit::WebProcessCreationParameters::decode):
- Shared/WebProcessCreationParameters.h:
- UIProcess/WebContext.cpp:
(WebKit::WebContext::createNewWebProcess):
(WebKit::WebContext::openGLCacheDirectory):
- UIProcess/WebContext.h:
- UIProcess/mac/WebContextMac.mm:
(WebKit::WebContext::platformDefaultOpenGLCacheDirectory):
- WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::platformInitializeWebProcess):
- 11:24 AM Changeset in webkit [170969] by
-
- 1 edit in trunk/Source/WebKit2/ChangeLog
Fix my mis-merged ChangeLog from r170966
- 10:52 AM Changeset in webkit [170968] by
-
- 4 edits in trunk
Enable same features on all Windows ports.
https://bugs.webkit.org/show_bug.cgi?id=134715
Patch by Alex Christensen <achristensen@webkit.org> on 2014-07-10
Reviewed by Brent Fulgham.
Source/WTF:
- wtf/Platform.h:
Use REQUEST_ANIMATION_FRAME_TIMER on WinCairo.
WebKitLibraries:
- win/tools/vsprops/FeatureDefinesCairo.props:
Made identical to FeatureDefines.props.
- 10:07 AM Changeset in webkit [170967] by
-
- 2 edits in trunk/Source/WebKit2
iOS build fix.
- WebProcess/WebPage/ServicesOverlayController.h:
- 9:19 AM Changeset in webkit [170966] by
-
- 6 edits in trunk/Source
Phone number highlights should always be visible if the mouse hovers over.
<rdar://problem/17527476> and https://bugs.webkit.org/show_bug.cgi?id=134784
Reviewed by Tim Horton.
Source/WebCore:
- WebCore.exp.in:
Source/WebKit2:
This is a fairly extensive rewrite of ServicesOverlayController.
It allows one selection highlight for the entire selection, and as many telephone number highlights as there are numbers.
If a telephone number highlight is hovered over, it wins and is painted.
If no telephone number highlight is hovered but the selection highlight is, then it is painted.
The purposes of each method are self evident by their name, and the concepts are mostly the same as they used to be.
The exception is establishHoveredTelephoneHighlight which gets a more detailed explanation below.
- Platform/Logging.h: Add a Services logging channel.
- WebProcess/WebPage/ServicesOverlayController.h:
(WebKit::TelephoneNumberData::TelephoneNumberData):
- WebProcess/WebPage/mac/ServicesOverlayController.mm:
(WebKit::ServicesOverlayController::ServicesOverlayController):
(WebKit::ServicesOverlayController::selectionRectsDidChange):
(WebKit::ServicesOverlayController::selectedTelephoneNumberRangesChanged):
(WebKit::ServicesOverlayController::clearHighlightState):
(WebKit::ServicesOverlayController::drawRect):
(WebKit::ServicesOverlayController::drawSelectionHighlight):
(WebKit::ServicesOverlayController::maybeDrawTelephoneNumberHighlight):
(WebKit::ServicesOverlayController::drawHighlight):
(WebKit::ServicesOverlayController::clearSelectionHighlight):
(WebKit::ServicesOverlayController::clearHoveredTelephoneNumberHighlight):
(WebKit::ServicesOverlayController::establishHoveredTelephoneHighlight): Starts walking the telephone number ranges and
creating a highlight for each one that doesn’t already have a highlight. If that highlight is also being hovered by
the mouse, then it is set as the hovered telephone number highlight and the method stops creating new highlights.
(WebKit::ServicesOverlayController::maybeCreateSelectionHighlight):
(WebKit::ServicesOverlayController::mouseEvent):
(WebKit::ServicesOverlayController::handleClick):
(WebKit::ServicesOverlayController::drawTelephoneNumberHighlight): Deleted.
(WebKit::ServicesOverlayController::drawCurrentHighlight): Deleted.
- 9:13 AM Changeset in webkit [170965] by
-
- 5 edits in trunk
[GTK][CMake] Add a 'distcheck' target
https://bugs.webkit.org/show_bug.cgi?id=130675
.:
Reviewed by Gustavo Noronha Silva.
- Source/PlatformGTK.cmake: Add distcheck target.
Tools:
Patch by Carlos Garcia Campos <cgarcia@igalia.com> and Martin Robinson <mrobinson@igalia.com> on 2014-07-10
Reviewed by Gustavo Noronha Silva.
Add --check option to make-dist script that builds and installs the tarball.
- gtk/make-dist.py:
(Distcheck.init):
(Distcheck.extract_tarball): Extratc the tarball.
(Distcheck.configure): Run cmake.
(Distcheck.configure.create_dir): Helper function to create a directory.
(Distcheck.build): Run make.
(Distcheck.install): Run make install.
(Distcheck.clean): Clean up the extracted tarball.
(Distcheck.check): Run all previous methods in order.
- gtk/manifest.txt: Add missing file.
- 9:11 AM Changeset in webkit [170964] by
-
- 4 edits1 add in trunk
[GTK] Use the same default options for production builds that previous stable releases
https://bugs.webkit.org/show_bug.cgi?id=134589
Reviewed by Martin Robinson.
.:
Change the default value of some features to match our stable releases.
Add FindCairoGL to find cairo-gl libraries and make accelearetd 2D
canvas depend on whether cairo-gl is found.
- Source/cmake/FindCairoGL.cmake: Added.
- Source/cmake/OptionsGTK.cmake:
Tools:
Make sure all features enabled in production builds are also
enabled in developer mode.
- Scripts/webkitperl/FeatureList.pm: Enable ACCELERATED_2D_CANVAS
and CSS_FILTERS for the GTK port.
- 8:39 AM Changeset in webkit [170963] by
-
- 4 edits in trunk/Source
Assertions or crashes under _takeViewSnapshot when restoring windows
https://bugs.webkit.org/show_bug.cgi?id=134792
Reviewed by Simon Fraser.
- UIProcess/API/mac/WKView.mm:
(-[WKView _takeViewSnapshot]):
Taking a window-server snapshot of a non-visible window tends to not succeed.
- platform/graphics/cocoa/IOSurface.mm:
(IOSurface::createFromImage):
- 5:23 AM Changeset in webkit [170962] by
-
- 1 edit4 adds in trunk/LayoutTests
[EFL] Unreviewed EFL gardening.
Rebaseline tests after r170304 and r170619.
- platform/efl/fast/multicol/pagination/RightToLeft-rl-dynamic-expected.png: Added after r170304.
- platform/efl/fast/multicol/pagination/RightToLeft-rl-dynamic-expected.txt: Added after r170304.
- platform/efl/fast/text/indic-expected.png: Added after r170619.
- platform/efl/fast/text/indic-expected.txt: Added after r170619.
- 5:14 AM Changeset in webkit [170961] by
-
- 1 edit1 add in trunk/LayoutTests
[EFL] Unreviewed EFL gardening.
Rebaseline tests after r170207.
- platform/efl/fast/text/khmer-lao-font-expected.txt: Added.
- 3:09 AM Changeset in webkit [170960] by
-
- 3 edits3 adds in trunk
[XHR] overrideMimeType() should be able to change encoding in HEADERS RECEIVED state
https://bugs.webkit.org/show_bug.cgi?id=128968
Patch by Youenn Fablet <youenn.fablet@crf.canon.fr> on 2014-07-10
Reviewed by Alexey Proskuryakov.
Source/WebCore:
Moved response encoding computation from didReceiveResponse to didReceiveData, just before the decoder is instantiated.
This allows overrideMimeType to be changed within readystatechange event callback and have an impact on selected encoding.
Test: http/tests/xmlhttprequest/overridemimetype-headers-received-state-force-shiftjis.html
- xml/XMLHttpRequest.cpp:
(WebCore::XMLHttpRequest::didReceiveResponse): Removed m_responseEncoding computation from response and mimeTypeOverride.
(WebCore::XMLHttpRequest::didReceiveData): Added m_responseEncoding computation from response and mimeTypeOverride.
LayoutTests:
Added test equivalent to http://w3c-test.org/web-platform-tests/master/XMLHttpRequest/overridemimetype-headers-received-state-force-shiftjis.htm
The test sets overrideMimeType when xhr state is changed to HEADERS RECEIVED.
- http/tests/xmlhttprequest/overridemimetype-headers-received-state-force-shiftjis-expected.txt: Added.
- http/tests/xmlhttprequest/overridemimetype-headers-received-state-force-shiftjis.html: Added.
- http/tests/xmlhttprequest/resources/get-content.php: Added.
