Timeline



Mar 19, 2015:

11:01 PM Changeset in webkit [181784] by aestes@apple.com
  • 1 edit in trunk/LayoutTests/ChangeLog

Fixed a ChangeLog misspelling pointed out by Andreas.

10:58 PM Changeset in webkit [181783] by aestes@apple.com
  • 8 edits
    2 adds in trunk/LayoutTests

[Content Filtering] Give contentfiltering tests a JavaScript harness
​https://bugs.webkit.org/show_bug.cgi?id=142899

Reviewed by Andreas Kling.

Create a JavaScript function to run the contentfiltering/ tests. This consolodates shared logic and will make
it easier to write new tests.

  • TestExpectations: Skipped a test that times out with the new harness (but it also failed before).
  • contentfiltering/allow-after-add-data.html: Called testContentFiltering().
  • contentfiltering/allow-after-finished-adding-data.html: Ditto.
  • contentfiltering/allow-after-response.html: Ditto.
  • contentfiltering/block-after-add-data.html: Ditto.
  • contentfiltering/block-after-finished-adding-data.html: Ditto.
  • contentfiltering/block-after-response.html: Ditto.
  • contentfiltering/resources/contentfiltering.js: Added.
10:45 PM Changeset in webkit [181782] by Chris Dumez
  • 8 edits in trunk/Source

[WK2] We should allow stale content when restoring a killed / crashed WebProcess
​https://bugs.webkit.org/show_bug.cgi?id=142893
<rdar://problem/19633737>
<rdar://problem/19869257>

Reviewed by Antti Koivisto.

Source/WebCore:

Always use ReturnCacheDataElseLoad cache policy for history navigations.
Previously, we would use the default protocol cache policy for history
navigations if m_stateMachine.committedFirstRealDocumentLoad() returned
false so that we do not display stale content when restoring the session
state of Safari statup (rdar://problem/8131355). This is now handled
before reaching this method by no longer using "IndexedBackForward"
FrameLoadType when restoring session tabs, and using "Standard" load
type instead.

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::loadDifferentDocumentItem):

Source/WebKit2:

Allow stale content when restoring a killed / crashed WebProcess
(similarly to a history navigation). This is actually what was intended
as we call WebPage::goToBackForwardItem() in this case, which uses
IndexedBackForward load type. However, this value was then ignored in
FrameLoader::loadDifferentDocumentItem().

This patch gets rid of the special casing in
FrameLoader::loadDifferentDocumentItem() and makes it clear at call
sites when calling goToBackForwardItem() if we should allow stale
content or not. If goToBackForwardItem() is called with the
allowStaleContent flag off, we'll keep using the IndexedBackForward
frame load type. However, if the flag is on, we'll use the Standard
frame load type and do a fresh load.

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::reattachToWebProcessForReload):
Set allowStaleContent flag to true as we want to allow stale content
when reloading a WebProcess after it crashed or was killed (behavior
change).

(WebKit::WebPageProxy::reattachToWebProcessWithItem):
Add allowStaleContent parameter so that call sites can decide which
policy to use.

(WebKit::WebPageProxy::goForward):
(WebKit::WebPageProxy::goBack):
Set allowStaleContent flag to true as this is our policy for history
navigations.

(WebKit::WebPageProxy::goToBackForwardItem):
Add allowStaleContent parameter so that the call site can let us know
what to do. This is useful as this is called not only for indexed
back / forward navigations (allowing stale content) but also for
restoring session state (forbidding state content:
<rdar://problem/8131355>).

(WebKit::WebPageProxy::restoreFromSessionState):
Call goToBackForwardItem() with allowStaleContent flag set to false
as users expect fresh contents in this case. This method is called
when:

  • Tabs are restored on Safari / MobileSafari startup
  • Tabs are restored via "History > Reopen All windows from last session"
  • UIProcess/WebPageProxy.h:
  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::goToBackForwardItem):
Add allowStaleContent parameter and use "Standard" frame load type
instead of "IndexedBackForward" if it is false. This is needed as
"IndexedBackForward" frame load type allows using stale content.

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

Add allowStaleContent flag to GoToBackForwardItem message.

10:22 PM Changeset in webkit [181781] by aestes@apple.com
  • 6 edits
    1 copy
    2 moves in trunk/Source/WebCore

[Content Filtering] Move ContentFilter to from platform/ to loader/
​https://bugs.webkit.org/show_bug.cgi?id=142897

Reviewed by Dan Bernstein.

ContentFilterCollection never contained any platform code, so it doesn't make sense for it to be in platform.
Move it to loader/, since ContentFilters are owned by DocumentLoaders, and just call it ContentFilter. Extract
the abstract part of ContentFilter into a new base class called PlatformContentFilter that each of the
individual filters inherit from.

  • WebCore.xcodeproj/project.pbxproj:
  • loader/ContentFilter.cpp: Renamed from Source/WebCore/platform/ContentFilter.cpp.

(WebCore::ContentFilter::types): Moved from ContentFilterCollection.
(WebCore::ContentFilter::createIfNeeded): Ditto.
(WebCore::ContentFilter::ContentFilter): Ditto.
(WebCore::ContentFilter::addData): Ditto.
(WebCore::ContentFilter::finishedAddingData): Ditto.
(WebCore::ContentFilter::needsMoreData): Ditto.
(WebCore::ContentFilter::didBlockData): Ditto.
(WebCore::ContentFilter::getReplacementData): Ditto.
(WebCore::ContentFilter::unblockHandler): Ditto.

  • loader/ContentFilter.h: Copied from Source/WebCore/platform/ContentFilter.h.

(WebCore::ContentFilter::type): Replaced a use of ContentFilterCollection.

  • platform/PlatformContentFilter.h: Renamed from Source/WebCore/platform/ContentFilter.h.
  • platform/cocoa/NetworkExtensionContentFilter.h: Inherited from PlatformContentFilter.
  • platform/cocoa/ParentalControlsContentFilter.h: Ditto.
  • testing/MockContentFilter.h: Ditto.
10:06 PM Changeset in webkit [181780] by Brent Fulgham
  • 2 edits in trunk/LayoutTests

[Win] Try to get the Debug bots green.

  • platform/win/TestExpectations:
10:02 PM Changeset in webkit [181779] by Brent Fulgham
  • 2 edits in trunk/LayoutTests

[Win] Skip an <attachment> test that times out on Windows.

Also mark two flakes I noticed on the test bots.

  • platform/win/TestExpectations:
9:47 PM Changeset in webkit [181778] by dino@apple.com
  • 12 edits
    4 adds in trunk

CSS Animations with triggers should map scroll position to duration
​https://bugs.webkit.org/show_bug.cgi?id=142870
<rdar://problem/20227244>

Reviewed by Simon Fraser.

Source/WebCore:

Expose a prototype implementation of what will eventually be called
animation-timebase, mapping the scroll location to the duration of an
animation. This only applies if the animation has a start and end
trigger.

Test: animations/trigger-container-scroll-boundaries.html
Test: animations/trigger-container-scroll-empty.html

  • css/CSSComputedStyleDeclaration.cpp:

(WebCore::createAnimationTriggerValue): Change to downcast.

  • css/CSSToStyleMap.cpp:

(WebCore::CSSToStyleMap::mapAnimationTrigger): Fix a bug where I wasn't
telling the ScrollAnimationTrigger object whether or not it had
an end value. This wasn't visible before because we were not using
the end value.

  • page/animation/AnimationBase.cpp:

(WebCore::AnimationBase::fireAnimationEventsIfNeeded): Add a comment and
change to using a downcast.
(WebCore::AnimationBase::timeToNextService): Change to downcast.
(WebCore::AnimationBase::freezeAtTime): Whitespace cleanup.
(WebCore::AnimationBase::getElapsedTime): Calculate the elapsedTime based
on the scroll position relative to the start and end trigger spots.

  • page/animation/AnimationController.cpp:

(WebCore::AnimationControllerPrivate::updateAnimations): Since this
can potentially call into beginAnimationUpdateTime, it should have
an update blocker. This fixes the assert we were seeing in tests.
(WebCore::AnimationControllerPrivate::scrollWasUpdated): Keep track of the scroll
position so we don't need to ask for it all the time.

  • page/animation/AnimationControllerPrivate.h:

(WebCore::AnimationControllerPrivate::scrollPosition): New accessor.

  • platform/animation/AnimationTrigger.h: Add downcast macros.

(WebCore::ScrollAnimationTrigger::create): Calculate hasEndValue from the passed
parameters rather than a separate value.
(WebCore::ScrollAnimationTrigger::hasEndValue):
(WebCore::ScrollAnimationTrigger::ScrollAnimationTrigger):
(WebCore::ScrollAnimationTrigger::setHasEndValue): Deleted.

  • platform/graphics/ca/GraphicsLayerCA.cpp:

(WebCore::GraphicsLayerCA::animationCanBeAccelerated): If we have an
animation of this sort, we can't use a CAAnimation.

LayoutTests:

Add a new test that exercises an animation trigger over a scroll
region, and a test of the boundary condition. Also, unskip a test that
was failing due to a bug fixed in this patch.

  • animations/trigger-container-scroll-boundaries-expected.txt: Added.
  • animations/trigger-container-scroll-boundaries.html: Added.
  • animations/trigger-container-scroll-empty-expected.txt: Added.
  • animations/trigger-container-scroll-empty.html: Added.
  • platform/efl/TestExpectations:
  • platform/gtk/TestExpectations:
  • platform/mac/TestExpectations:
  • platform/win/TestExpectations:
8:51 PM Changeset in webkit [181777] by eric.carlson@apple.com
  • 9 edits
    4 moves in trunk/Source

[Mac] Move MediaPlaybackTargetPicker
​https://bugs.webkit.org/show_bug.cgi?id=142889

Source/WebCore:

Move WebMediaPlaybackTargetPickerProxy from WK2 and rename as MediaPlaybackTargetPicker so it
can be used by both WK2 and WK1.

Reviewed by Tim Horton.

  • WebCore.xcodeproj/project.pbxproj:
  • platform/graphics/MediaPlaybackTargetPicker.cpp: Copied from Source/WebKit2/UIProcess/WebMediaPlaybackTargetPickerProxy.cpp.
  • platform/graphics/MediaPlaybackTargetPicker.h: Copied from Source/WebKit2/UIProcess/WebMediaPlaybackTargetPickerProxy.h.
  • platform/graphics/avfoundation/objc/MediaPlaybackTargetPickerMac.h: Copied from Source/WebKit2/UIProcess/mac/WebMediaPlaybackTargetPickerProxyMac.h.
  • platform/graphics/avfoundation/objc/MediaPlaybackTargetPickerMac.mm: Copied from Source/WebKit2/UIProcess/mac/WebMediaPlaybackTargetPickerProxyMac.mm.

Source/WebKit2:

Move WebMediaPlaybackTargetPickerProxy to WebCore so it can be used by
both WK2 and WK1.

Reviewed by Tim Horton.

  • UIProcess/PageClient.h:
  • UIProcess/WebMediaPlaybackTargetPickerProxy.cpp: Removed.
  • UIProcess/WebMediaPlaybackTargetPickerProxy.h: Removed.
  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::devicePickerProxy):

  • UIProcess/WebPageProxy.h:
  • UIProcess/mac/PageClientImpl.h:
  • UIProcess/mac/PageClientImpl.mm:

(WebKit::PageClientImpl::createPlaybackTargetPicker):

  • UIProcess/mac/WebMediaPlaybackTargetPickerProxyMac.h: Removed.
  • UIProcess/mac/WebMediaPlaybackTargetPickerProxyMac.mm: Removed.
  • WebKit2.xcodeproj/project.pbxproj:
6:52 PM Changeset in webkit [181776] by jberlin@webkit.org
  • 2 edits in trunk/Source/WebKit2

32-bit build fix after r181770.

  • UIProcess/mac/WKFullScreenWindowController.mm:

(-[WKFullScreenWindowController initWithWindow:webView:]):

6:37 PM Changeset in webkit [181775] by andersca@apple.com
  • 2 edits in trunk/Source/WebKit2

Add support for deleting individual cache entries
​https://bugs.webkit.org/show_bug.cgi?id=142886

Reviewed by Antti Koivisto.

  • NetworkProcess/NetworkProcess.cpp:

(WebKit::fetchDiskCacheEntries):
Remove an unnecessary call to cfURLCacheOrigins().

(WebKit::clearDiskCacheEntries):
Helper function for clearing disk cache entries. This handles both the old and new cache implementations.

(WebKit::NetworkProcess::deleteWebsiteDataForOrigins):
Call clearDiskCacheEntries if needed.

5:53 PM Changeset in webkit [181774] by rniwa@webkit.org
  • 2 edits in trunk/Tools

Unreviewed build fix. Don't try to report stats for all benchmarks.
Just report the geometric mean of preferred means.

  • Scripts/run-jsc-benchmarks:
5:47 PM Changeset in webkit [181773] by enrica@apple.com
  • 3 edits
    2 adds in trunk

Source/WebCore:
REGRESSION (r109593): Clicking after last inline element could cause a crash.
​https://bugs.webkit.org/show_bug.cgi?id=142880
rdar://problem/17222294

Reviewed by Ryosuke Niwa.

Test: editing/selection/click-after-last-inline-crash.html

  • rendering/RootInlineBox.cpp:

(WebCore::RootInlineBox::closestLeafChildForLogicalLeftPosition):

LayoutTests:
Web Inspector: Adopt ES6 Class Syntax for all Model Objects
​https://bugs.webkit.org/show_bug.cgi?id=142858

Patch by Joseph Pecoraro <Joseph Pecoraro> on 2015-03-19
Reviewed by Timothy Hatcher.

  • inspector/model/parse-script-syntax-tree.html:

This test was calling a constructor without "new". Class
syntax enforces "new" and threw an exception.

5:44 PM Changeset in webkit [181772] by enrica@apple.com
  • 2 edits in trunk/Source/WebCore

One more build fix after r181760.

  • page/EventHandler.cpp:

(WebCore::EventHandler::handleDrag):

5:38 PM Changeset in webkit [181771] by jberlin@webkit.org
  • 2 edits in trunk/Source/WebCore

Build fix after r181760.

  • rendering/RenderAttachment.h:
5:20 PM Changeset in webkit [181770] by jer.noble@apple.com
  • 6 edits in trunk/Source

[WK2][Mac] Fullscreen animations with mismatched aspect ratios are "squished".
​https://bugs.webkit.org/show_bug.cgi?id=142132

Reviewed by Tim Horton.

Source/WebCore:

Add a new utility method to calculate a rect with a specific aspect ratio wrapping a smaller one.

  • platform/graphics/GeometryUtilities.cpp:

(WebCore::smallestRectWithAspectRatioAroundRect):

  • platform/graphics/GeometryUtilities.h:

Source/WebKit2:

Use CALayer animations for the transition into and out of fullscreen.

The fullscreen transition consists of three separate animations:

  1. An opacity animation for the black backdrop behind the fullscreen content
  2. A scale/translation animation from fullscreen element's initial screen to its final one.
  3. A clip animation from the fullscreen element's initial aspect ratio to its final one.

The opacity animation will apply to the fullscreen window's content view's layer's
background color. To separately animate the transform and mask of the web view's content, a
layer-backed subview is added to the content view, and the scale/translation & clip
animations are applied to its transform property and mask layer respectively.

Utility methods to create each animation have been added, and each includes a parameter for
the direction of the animation, so that the same methods can be used for entering and
exiting fullscreen transitions.

The user-visible changes to this new transition are when the aspect ratios of the initial
and final screen locations are different: previously the transition would use a scale
transform to "squish" the fullscreen content into the initial aspect ratio. The new
transition instead "clips" the fullscreen content to the initial aspect ratio. For common
operations such a <video> element with a different aspect ratio than the screen, this makes
the black letterbox "grow" during the transition, and makes the transition overall much
smoother.

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

(-[WKFullScreenWindowController initWithWindow:webView:]): Create and initialze the clipping view.
(-[WKFullScreenWindowController applicationDidChangeScreenParameters:]): _backgroundWindow was removed.
(-[WKFullScreenWindowController enterFullScreen:]): Add the webView to the _clipView, not the contentView.
(-[WKFullScreenWindowController beganEnterFullScreenWithInitialFrame:finalFrame:]): _backgroundWindow,

_fadeAnimation, and _scaleAnimation are all removed.

(-[WKFullScreenWindowController finishedEnterFullScreenAnimation:]): Ditto.
(-[WKFullScreenWindowController finishedExitFullScreenAnimation:]): Ditto. Hide the contentView's

layer. Pause visibility updates.

(-[WKFullScreenWindowController completeFinishExitFullScreenAnimationAfterRepaint]): Resume visibility updates.
(-[WKFullScreenWindowController close]): _fadeAnimation and _scaleAnimation are removed.
(-[WKFullScreenWindowController customWindowsToEnterFullScreenForWindow:]): Return only the fullscreen

window.

(-[WKFullScreenWindowController customWindowsToExitFullScreenForWindow:]): Ditto.
(zoomAnimation): Added.
(maskAnimation): Added.
(fadeAnimation): Added.
(-[WKFullScreenWindowController _startEnterFullScreenAnimationWithDuration:]): Use the utility

methods above to set up the animation.

(-[WKFullScreenWindowController _startExitFullScreenAnimationWithDuration:]): Ditto.
(createBackgroundFullscreenWindow): Deleted.
(windowFrameFromApparentFrames): Deleted.

5:18 PM Changeset in webkit [181769] by Joseph Pecoraro
  • 182 edits in trunk

Web Inspector: Adopt ES6 Class Syntax for all Model Objects
​https://bugs.webkit.org/show_bug.cgi?id=142858

Reviewed by Timothy Hatcher.

Source/WebInspectorUI:

  • Convert WebInspector.Object to a class
  • Convert all UserInterface/Models objects to classes
  • Convert constructor functions to constructor methods
  • Convert "constructor.method" to class static methods where possible
  • Convert all methods to method syntax, eliminate commas between methods
  • Convert all superclass calls in classes to use "super"
  • Workaround <​https://webkit.org/b/142862> and add empty constructors
  • Added "deprecated" prefix to addConstructorFunctions, since it is not needed with classes
  • Added many missing calls to super in constructors
  • Added FIXME to WebInspector.Object subclasses not yet moved to classes.
  • Cleaned up SourceMap global properties, moved to constructor instead of prototype
  • Cleaned up Timeline factory constructor to static "create" factory method
  • Fixed any style issues noticed in the mass edit
  • Fixed strict mode issues now that classes enforce strict mode
    • RunLoopTimelineRecord.js was missing a var for a local variable
    • "const" is not allowed, converted to "var"
    • "arguments.callee" is not allowed in strict mode
  • UserInterface//*.js:

Many files modified mostly mechanically.

LayoutTests:

  • inspector/model/parse-script-syntax-tree.html:

This test was calling a constructor without "new". Class
syntax enforces "new" and threw an exception.

4:52 PM Changeset in webkit [181768] by jonowells@apple.com
  • 3 edits in trunk/Source/WebInspectorUI

Web Inspector: Debugger sidebar header should say "Scripts" instead of "Breakpoints", appear only on pause
​https://bugs.webkit.org/show_bug.cgi?id=142847

Reviewed by Timothy Hatcher.

Add a paused class on the debugger sidebar panel. Use that class to adjust the sidebar styles so that, when
the debugger is not paused: there is no header for the scripts section, there is no border beneath the scripts
section, and the scripts section cannot be collapsed (which could be done while the debugger is paused). The
header has changed to say "Scripts" instead of "Breakpoints" when it is showing.

  • UserInterface/Views/DebuggerSidebarPanel.css:

(.sidebar > .panel.navigation.debugger .details-section.scripts):
(.sidebar > .panel.navigation.debugger .details-section.scripts .header):
(.sidebar > .panel.navigation.debugger .details-section.scripts.collapsed > .content):
(.sidebar > .panel.navigation.debugger.paused .details-section.scripts):
(.sidebar > .panel.navigation.debugger.paused .details-section.scripts .header):
(.sidebar > .panel.navigation.debugger.paused .details-section.scripts.collapsed > .content):
Adjust styles to hide header and border.

  • UserInterface/Views/DebuggerSidebarPanel.js:

(WebInspector.DebuggerSidebarPanel):
(WebInspector.DebuggerSidebarPanel.prototype._debuggerDidPause):
(WebInspector.DebuggerSidebarPanel.prototype._debuggerDidResume):
Add a paused class to the debugger sidebar when it is paused.

4:33 PM Changeset in webkit [181767] by matthew_hanson@apple.com
  • 2 edits in tags/Safari-601.1.23.3/Source/WebCore

Merge r181662. rdar://problem/20144648

4:26 PM Changeset in webkit [181766] by Chris Dumez
  • 5 edits in trunk/Source/WebCore

Get rid of "CachePolicyCache" cache policy
​https://bugs.webkit.org/show_bug.cgi?id=142881

Reviewed by Antti Koivisto.

Get rid of "CachePolicyCache" cache policy as it is unused.

  • loader/cache/CachePolicy.h:
  • loader/cache/CachedResource.cpp:

(WebCore::CachedResource::mustRevalidateDueToCacheHeaders):

  • page/DiagnosticLoggingKeys.cpp:

(WebCore::DiagnosticLoggingKeys::mustRevalidateIsExpiredKey): Deleted.

  • page/DiagnosticLoggingKeys.h:
4:01 PM Changeset in webkit [181765] by akling@apple.com
  • 3 edits in trunk/Source/JavaScriptCore

JSCallee unnecessarily overrides a bunch of things in the method table.
<​https://webkit.org/b/142855>

Reviewed by Geoffrey Garen.

Remove JSCallee method table overrides that simply call to base class.
This makes JSFunction property slot lookups slightly more efficient since
they can take the fast path when passing over JSCallee in the base class chain.

  • runtime/JSCallee.cpp:

(JSC::JSCallee::getOwnPropertySlot): Deleted.
(JSC::JSCallee::getOwnNonIndexPropertyNames): Deleted.
(JSC::JSCallee::put): Deleted.
(JSC::JSCallee::deleteProperty): Deleted.
(JSC::JSCallee::defineOwnProperty): Deleted.

  • runtime/JSCallee.h:
3:46 PM Changeset in webkit [181764] by rniwa@webkit.org
  • 2 edits in trunk/Tools

run-jsc-benchmarks should only report preferred mean
​https://bugs.webkit.org/show_bug.cgi?id=142874

Reviewed by Geoffrey Garen.

Just collect statistics for the preferred mean.

  • Scripts/run-jsc-benchmarks:
3:39 PM Changeset in webkit [181763] by enrica@apple.com
  • 2 edits in trunk/Source/WebCore

[Content Extensions] Test regular expression parse failures.
​https://bugs.webkit.org/show_bug.cgi?id=142872

Patch by Alex Christensen <​achristensen@webkit.org> on 2015-03-19
Reviewed by Benjamin Poulain.

  • contentextensions/URLFilterParser.cpp:

(WebCore::ContentExtensions::GraphBuilder::finalize):
(WebCore::ContentExtensions::GraphBuilder::quantifyAtom):
(WebCore::ContentExtensions::URLFilterParser::addPattern):
(WebCore::ContentExtensions::URLFilterParser::statusString):

  • contentextensions/URLFilterParser.h:

Changed failures that can never happen to assertions.

3:37 PM Changeset in webkit [181762] by achristensen@apple.com
  • 5 edits in trunk

[Content Extensions] Test regular expression parse failures.
​https://bugs.webkit.org/show_bug.cgi?id=142872

Reviewed by Benjamin Poulain.

Source/WebCore:

  • contentextensions/URLFilterParser.cpp:

(WebCore::ContentExtensions::GraphBuilder::finalize):
(WebCore::ContentExtensions::GraphBuilder::quantifyAtom):
(WebCore::ContentExtensions::URLFilterParser::addPattern):
(WebCore::ContentExtensions::URLFilterParser::statusString):

  • contentextensions/URLFilterParser.h:

Changed failures that can never happen to assertions.

Tools:

  • TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:

(TestWebKitAPI::testPatternStatus):
(TestWebKitAPI::TEST_F):
Added tests for each parsing status.

3:36 PM Changeset in webkit [181761] by rniwa@webkit.org
  • 5 edits in trunk/Websites/perf.webkit.org

Unreviewed build fixes.

  • public/include/manifest.php:

(Manifest::generate): These should be {} instead of [] when they're empty.

  • public/v2/data.js:

(Measurement.prototype.formattedRevisions): Don't assume previousRevisions[repositoryId] exits.

  • public/v2/manifest.js:

(App.Metric.fullName): Fixed the typo.

  • tests/admin-regenerate-manifest.js: Fixed the test.
2:59 PM Changeset in webkit [181760] by enrica@apple.com
  • 32 edits
    2 adds in trunk

<attachment> should put URLs on the pasteboard so that Finder can accept drops.
​https://bugs.webkit.org/show_bug.cgi?id=142801
rdar://problem/19982527

Reviewed by Tim Horton.

Source/WebCore:

Test: editing/pasteboard/drag-and-drop-attachment-contenteditable.html

This patch adds all the necessary support to write attachment elements into
the pasteboard, including the promised type.

  • WebCore.xcodeproj/project.pbxproj:
  • page/DragActions.h:
  • page/DragClient.h:

(WebCore::DragClient::declareAndWriteAttachment):

  • page/DragController.cpp:

(WebCore::DragController::draggableElement):
(WebCore::DragController::startDrag):

  • page/DragController.h:

(WebCore::DragController::draggingAttachmentURL):

  • page/EventHandler.cpp:

(WebCore::EventHandler::dragHysteresisExceeded):
(WebCore::EventHandler::handleDrag):

  • page/mac/DragControllerMac.mm:

(WebCore::DragController::declareAndWriteAttachment):

  • page/win/DragControllerWin.cpp:

(WebCore::DragController::declareAndWriteAttachment):

  • platform/URL.cpp:

(WebCore::URL::fileURLWithFileSystemPath):

  • platform/URL.h:
  • rendering/HitTestResult.cpp:

(WebCore::HitTestResult::absoluteAttachmentURL):

  • rendering/HitTestResult.h:

Source/WebKit/mac:

This patch adds all the necessary support to write attachment elements into
the pasteboard, including the promised type in WK1.

  • Misc/WebNSPasteboardExtras.mm:

(-[NSPasteboard _web_declareAndWriteDragImageForElement:URL:title:archive:source:]):

  • WebCoreSupport/WebDragClient.h:
  • WebCoreSupport/WebDragClient.mm:

(WebDragClient::declareAndWriteAttachment):

  • WebView/WebHTMLView.mm:

(-[WebHTMLView namesOfPromisedFilesDroppedAtDestination:]):

  • WebView/WebUIDelegate.h:

Source/WebKit2:

This patch adds all the necessary support to write attachment elements into
the pasteboard, including the promised type in WK2.

  • UIProcess/API/mac/WKView.mm:

(-[WKView _setPromisedDataForImage:withFileName:withExtension:withTitle:withURL:withVisibleURL:withArchive:forPasteboard:]):
(-[WKView _setPromisedDataForAttachment:withExtension:withTitle:withURL:withVisibleURL:forPasteboard:]):
(-[WKView namesOfPromisedFilesDroppedAtDestination:]):
(-[WKView _setPromisedData:withFileName:withExtension:withTitle:withURL:withVisibleURL:withArchive:forPasteboard:]): Deleted.

  • UIProcess/API/mac/WKViewInternal.h:
  • UIProcess/PageClient.h:
  • UIProcess/WebPageProxy.h:
  • UIProcess/WebPageProxy.messages.in:
  • UIProcess/mac/PageClientImpl.h:
  • UIProcess/mac/PageClientImpl.mm:

(WebKit::PageClientImpl::setPromisedDataForImage):
(WebKit::PageClientImpl::setPromisedDataForAttachment):
(WebKit::PageClientImpl::setPromisedData): Deleted.

  • UIProcess/mac/WebPageProxyMac.mm:

(WebKit::WebPageProxy::setPromisedDataForImage):
(WebKit::WebPageProxy::setPromisedDataForAttachment):
(WebKit::WebPageProxy::setPromisedData): Deleted.

  • WebProcess/WebCoreSupport/WebDragClient.h:
  • WebProcess/WebCoreSupport/mac/WebDragClientMac.mm:

(WebKit::WebDragClient::declareAndWriteAttachment):
(WebKit::WebDragClient::declareAndWriteDragImage):

LayoutTests:

  • editing/pasteboard/drag-and-drop-attachment-contenteditable-expected.txt: Added.
  • editing/pasteboard/drag-and-drop-attachment-contenteditable.html: Added.
  • platform/mac-wk2/TestExpectations: Skipping new test on WK2 since it uses eventSender.
2:44 PM Changeset in webkit [181759] by jonowells@apple.com
  • 2 edits in trunk/Source/WebInspectorUI

Web Inspector: FilterBar for debugger sidebar hides breakpoints for displayed resources
​https://bugs.webkit.org/show_bug.cgi?id=142777

Reviewed by Brian Burg.

Typing into the text input on the filter bar for the debugger sidebar no longer incorrectly hides the
breakpoints for scripts that match the text input.

  • UserInterface/Views/BreakpointTreeElement.js:

(WebInspector.BreakpointTreeElement.prototype.get filterableData): Added.

2:19 PM Changeset in webkit [181758] by akling@apple.com
  • 10 edits
    2 deletes in trunk/Source

DFGAllocator should use bmalloc's aligned allocator.
<​https://webkit.org/b/142871>

Reviewed by Geoffrey Garen.

Source/JavaScriptCore:

Switch DFGAllocator to using bmalloc through fastAlignedMalloc().

  • dfg/DFGAllocator.h:

(JSC::DFG::Allocator<T>::allocateSlow):
(JSC::DFG::Allocator<T>::freeRegionsStartingAt):

  • heap/CopiedSpace.h:
  • heap/MarkedBlock.h:
  • heap/MarkedSpace.h:

Source/WTF:

Remove PageAllocationAligned since it has no more clients.

  • WTF.vcxproj/WTF.vcxproj:
  • WTF.vcxproj/WTF.vcxproj.filters:
  • WTF.xcodeproj/project.pbxproj:
  • wtf/CMakeLists.txt:
  • wtf/PageAllocationAligned.cpp: Removed.
  • wtf/PageAllocationAligned.h: Removed.
2:17 PM Changeset in webkit [181757] by Chris Dumez
  • 2 edits in trunk/Source/WebCore

Unreviewed, fix linking error after r181753.

Drop WEBCORE_EXPORT for Settings::minimumDOMTimerInterval() as it is
now inlined.

  • page/Settings.h:

(WebCore::Settings::minimumDOMTimerInterval):

2:16 PM Changeset in webkit [181756] by andersca@apple.com
  • 3 edits in trunk/Source/WebKit2

Cache::remove should take a Key instead of an entry
​https://bugs.webkit.org/show_bug.cgi?id=142878

Reviewed by Antti Koivisto.

  • NetworkProcess/cache/NetworkCache.cpp:

(WebKit::NetworkCache::Cache::remove):

  • NetworkProcess/cache/NetworkCache.h:
1:50 PM Changeset in webkit [181755] by achristensen@apple.com
  • 2 edits in trunk/Tools

Fix API tests after r181754.

  • TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:

(TestWebKitAPI::testRequest):

12:36 PM Changeset in webkit [181754] by beidson@apple.com
  • 11 edits in trunk

Part of content extensions should apply css selectors
​https://bugs.webkit.org/show_bug.cgi?id=142604

Reviewed by Alex Christensen.

Source/WebCore:

Add a new action for applying the cached global stylesheet.

  • contentextensions/CompiledContentExtension.cpp:

(WebCore::ContentExtensions::CompiledContentExtension::globalDisplayNoneSelectors):

  • contentextensions/ContentExtensionActions.h:
  • contentextensions/ContentExtensionCompiler.cpp:

(WebCore::ContentExtensions::serializeActions):

  • contentextensions/ContentExtensionParser.cpp:

(WebCore::ContentExtensions::loadAction):

  • contentextensions/ContentExtensionRule.cpp:

(WebCore::ContentExtensions::Action::deserialize):

  • contentextensions/ContentExtensionRule.h:

(WebCore::ContentExtensions::Action::Action):
(WebCore::ContentExtensions::Action::operator==):
(WebCore::ContentExtensions::Action::stringArgument):
(WebCore::ContentExtensions::Action::cssSelector): Deleted.

  • contentextensions/ContentExtensionsBackend.cpp:

(WebCore::ContentExtensions::ContentExtensionsBackend::actionsForResourceLoad):

  • loader/cache/CachedResourceLoader.cpp:

(WebCore::CachedResourceLoader::requestResource):

Tools:

  • TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:

(WebCore::ContentExtensions::operator<<):

12:22 PM Changeset in webkit [181753] by Chris Dumez
  • 15 edits in trunk/Source

Clean up DOMTimer related settings
​https://bugs.webkit.org/show_bug.cgi?id=142837

Reviewed by Darin Adler.

Source/WebCore:

Clean up DOMTimer related settings:

  • Make default minimum timer interval 4ms instead of 10ms as all ports were setting it to 4ms and the specifications says 4ms.
  • Get rid of setters for default DOMTimer interval values as they are not set anymore.
  • Move default DOMTimer interval values from Settings to DOMTimer as those are global (not per-page) and not really settings.
  • Stop using abbreviations in the names (e.g. min -> minimum).
  • Move DOMTimer settings members from Page to Settings. There is no reason for those to be stored on the Page. The getters were private so call-sites had to go via Settings anyway. Also, call-sites already had to do a null check on the page to get the settings. It seems unfortunate for the settings to do another null-check on the page for retrieving the setting value from the page.
  • Inline the DOMTimer settings getters as these are trivial.
  • dom/Document.cpp:

(WebCore::Document::minimumTimerInterval):
(WebCore::Document::timerAlignmentInterval):

  • dom/ScriptExecutionContext.cpp:

(WebCore::ScriptExecutionContext::minimumTimerInterval):
(WebCore::ScriptExecutionContext::timerAlignmentInterval):

  • page/DOMTimer.h:
  • page/Page.cpp:

(WebCore::Page::Page):
(WebCore::Page::setTimerThrottlingEnabled):
(WebCore::Page::setMinimumTimerInterval): Deleted.
(WebCore::Page::minimumTimerInterval): Deleted.

  • page/Page.h:

(WebCore::Page::timerAlignmentInterval): Deleted.

  • page/Settings.cpp:

(WebCore::Settings::Settings):
(WebCore::Settings::setMinimumDOMTimerInterval):
(WebCore::Settings::setDOMTimerAlignmentInterval):
(WebCore::Settings::setHiddenPageDOMTimerAlignmentInterval): Deleted.
(WebCore::Settings::hiddenPageDOMTimerAlignmentInterval): Deleted.
(WebCore::Settings::setDefaultMinDOMTimerInterval): Deleted.
(WebCore::Settings::defaultMinDOMTimerInterval): Deleted.
(WebCore::Settings::setMinDOMTimerInterval): Deleted.
(WebCore::Settings::minDOMTimerInterval): Deleted.
(WebCore::Settings::setDefaultDOMTimerAlignmentInterval): Deleted.
(WebCore::Settings::defaultDOMTimerAlignmentInterval): Deleted.
(WebCore::Settings::domTimerAlignmentInterval): Deleted.

  • page/Settings.h:

(WebCore::Settings::minimumDOMTimerInterval):
(WebCore::Settings::domTimerAlignmentInterval):

  • testing/InternalSettings.cpp:

(WebCore::InternalSettings::Backup::Backup):
(WebCore::InternalSettings::Backup::restoreTo):
(WebCore::InternalSettings::setMinimumTimerInterval):

Source/WebKit/mac:

Stop setting the DOMTimers' default minimum interval to 4ms as this
is now the default.

  • WebView/WebView.mm:

(-[WebView _commonInitializationWithFrameName:groupName:]):

Source/WebKit/win:

Stop setting the DOMTimers' default minimum interval to 4ms as this
is now the default.

  • WebView.cpp:

(WebView::initWithFrame):
(WebView::defaultMinimumTimerInterval):

Source/WebKit2:

Stop setting the DOMTimers' default minimum interval to 4ms as this
is now the default.

  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::WebPage):

11:13 AM Changeset in webkit [181752] by Lucas Forschler
  • 5 edits in branches/safari-600.1.17-branch/Source

Versioning.

11:11 AM Changeset in webkit [181751] by Lucas Forschler
  • 1 copy in tags/Safari-600.1.17.10

New Tag.

10:52 AM Changeset in webkit [181750] by msaboff@apple.com
  • 7 edits in branches/safari-600.1.17-branch/Source

Merged r181628. <rdar://problem/19804738>

2015-03-16 Michael Saboff <​msaboff@apple.com>

Windows X86-64 should use the fixed executable allocator
​https://bugs.webkit.org/show_bug.cgi?id=142749

Reviewed by Filip Pizlo.

Source/JavaScriptCore:

Added jit/ExecutableAllocatorFixedVMPool.cpp to Windows build.

  • JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
  • JavaScriptCore.vcxproj/JavaScriptCore.vcxproj.filters:
  • jit/ExecutableAllocatorFixedVMPool.cpp: Don't include unistd.h on Windows.

Source/WTF:

Set ENABLE_EXECUTABLE_ALLOCATOR_FIXED for Windows.
Needed to export MetaAllocator::currentStatistics() for use in JavaScriptCore.

  • wtf/MetaAllocator.h:
  • wtf/Platform.h:
10:14 AM Changeset in webkit [181749] by dino@apple.com
  • 2 edits in trunk/LayoutTests

​http://webkit.org/b/142790

Temporarily skip animations/trigger-container-scroll-simple.html
while it is crashing.

  • platform/mac/TestExpectations:
9:42 AM Changeset in webkit [181748] by Brent Fulgham
  • 2 edits in trunk/Tools

[Win] Activate tests on EWS
​https://bugs.webkit.org/show_bug.cgi?id=142850

Reviewed by Alexey Proskuryakov.

  • Scripts/webkitpy/common/config/ews.json: Add me as

the EWS bot watcher, and activate tests.

9:40 AM Changeset in webkit [181747] by Brent Fulgham
  • 2 edits in trunk/LayoutTests

[Win] Test gardening.

  • platform/win/TestExpectations:
6:25 AM Changeset in webkit [181746] by commit-queue@webkit.org
  • 2 edits in trunk/LayoutTests

[GTK] Skip Legacy WebNotifications test
​https://bugs.webkit.org/show_bug.cgi?id=141962

The revision r177073 skips all legacy tests for notifications except from
http/tests/notifications/legacy/notification-request-permission-then-navigate.html
which was failing. We skip this test now too.

Patch by Marcos Chavarría Teijeiro <​chavarria1991@gmail.com> on 2015-03-19
Reviewed by Carlos Garcia Campos.

  • platform/gtk/TestExpectations:
3:40 AM Changeset in webkit [181745] by Carlos Garcia Campos
  • 2 edits in releases/WebKitGTK/webkit-2.8/Source/WebCore

Merge r181744 - [GTK] Scrollbars look bad with GTK+ 3.16
​https://bugs.webkit.org/show_bug.cgi?id=140800

Reviewed by Sergio Villar Senin.

Take margin into account when rendering scrollbars. This fixes the
huge scrollbars rendered with GTK+ 3.16. We don't need to check
the GTK+ version because in previous versions the marging were 0,
so the same code just works.

  • platform/gtk/ScrollbarThemeGtk.cpp:

(WebCore::adjustRectAccordingToMargin):
(WebCore::ScrollbarThemeGtk::paintTrackBackground):
(WebCore::ScrollbarThemeGtk::paintThumb):

3:24 AM Changeset in webkit [181744] by Carlos Garcia Campos
  • 2 edits in trunk/Source/WebCore

[GTK] Scrollbars look bad with GTK+ 3.16
​https://bugs.webkit.org/show_bug.cgi?id=140800

Reviewed by Sergio Villar Senin.

Take margin into account when rendering scrollbars. This fixes the
huge scrollbars rendered with GTK+ 3.16. We don't need to check
the GTK+ version because in previous versions the marging were 0,
so the same code just works.

  • platform/gtk/ScrollbarThemeGtk.cpp:

(WebCore::adjustRectAccordingToMargin):
(WebCore::ScrollbarThemeGtk::paintTrackBackground):
(WebCore::ScrollbarThemeGtk::paintThumb):

2:01 AM Changeset in webkit [181743] by bshafiei@apple.com
  • 4 edits in tags/Safari-601.1.23.3/Source

Merged r181679. rdar://problem/20032670

1:58 AM Changeset in webkit [181742] by bshafiei@apple.com
  • 2 edits in tags/Safari-601.1.23.3/Source/WebKit/mac

Merged r181674. rdar://problem/20032670

1:57 AM Changeset in webkit [181741] by bshafiei@apple.com
  • 6 edits in tags/Safari-601.1.23.3/Source

Merged r181668. rdar://problem/20032670

1:55 AM Changeset in webkit [181740] by bshafiei@apple.com
  • 3 edits in tags/Safari-601.1.23.3/Source/WebCore

Merged r181703. rdar://problem/19446938

1:53 AM Changeset in webkit [181739] by bshafiei@apple.com
  • 2 edits in tags/Safari-601.1.23.3/Source/WebCore

Merged r181666. rdar://problem/19445803

1:51 AM Changeset in webkit [181738] by bshafiei@apple.com
  • 2 edits in tags/Safari-601.1.23.3/Source/WebCore

Merged r181616. rdar://problem/20184456

1:49 AM Changeset in webkit [181737] by bshafiei@apple.com
  • 7 edits
    4 copies in tags/Safari-601.1.23.3

Merged r181608. rdar://problem/20184456

1:27 AM Changeset in webkit [181736] by youenn.fablet@crf.canon.fr
  • 8 edits in trunk

[Streams API] Update ReadableStream API according new version of the specification
​https://bugs.webkit.org/show_bug.cgi?id=142822

Reviewed by Benjamin Poulain.

Source/WebCore:

This patch updates the IDL of ReadableStream according the new version of the spec, which splits functionality between ReadableStream and ReadableStreamReader.
In particular, this patch removes read(), ready, closed and state from ReadableStream and it adds the getReader method.

Covered by updated readablestream-constructor.html test.

  • Modules/streams/ReadableStream.cpp:

(WebCore::ReadableStream::ReadableStream):
(WebCore::ReadableStream::state): Deleted.
(WebCore::ReadableStream::closed): Deleted.
(WebCore::ReadableStream::ready): Deleted.

  • Modules/streams/ReadableStream.h:

(WebCore::ReadableStream::internalState): Added to make mac build system happy, to be used by ReadableStreamReader.

  • Modules/streams/ReadableStream.idl:
  • bindings/js/JSReadableStreamCustom.cpp:

(WebCore::JSReadableStream::cancel):
(WebCore::JSReadableStream::getReader):
(WebCore::JSReadableStream::read): Deleted.
(WebCore::getOrCreatePromiseDeferredFromObject): Deleted.
(WebCore::readyPromiseSlotName): Deleted.
(WebCore::JSReadableStream::ready): Deleted.
(WebCore::closedPromiseSlotName): Deleted.
(WebCore::JSReadableStream::closed): Deleted.

LayoutTests:

Removing tests checking ready and closed.
Removing assertions checking read(), ready, closed and state.
Adding assertions to test getReader() and parameters of remaining methods.

  • streams/readablestream-constructor-expected.txt:
  • streams/readablestream-constructor.html:
1:19 AM Changeset in webkit [181735] by Csaba Osztrogonác
  • 4 edits in trunk

[GTK] Fix inspector userinterface related incremental build issue
​https://bugs.webkit.org/show_bug.cgi?id=142849

Reviewed by Carlos Garcia Campos.

.:

  • Source/cmake/OptionsGTK.cmake: Revert r181733.

Tools:

  • Scripts/webkitdirs.pm:

(shouldRemoveCMakeCache):

12:54 AM Changeset in webkit [181734] by Chris Dumez
  • 4 edits in trunk

The network cache should ignore "cache-control: no-cache, must-revalidate" on history navigation
​https://bugs.webkit.org/show_bug.cgi?id=142831
<rdar://problem/20209957>

Reviewed by Antti Koivisto.

Source/WebKit2:

The network cache should ignore "no-cache" and "must-revalidate" as
Cache-Control policy for history navigations to avoid unnecessary
revalidation in this case. Our memory cache already behaves this way.

On history navigation, our disk cache implementation currently
revalidates unconditionally resources if "cache-control: no-cache" is
used. It also revalidates expired resources if "cache-control:
must-revalidate" is used. This is sub-optimal.

RFC 7234 does not require us to do so [1]:
"""

User agents often have history mechanisms, such as "Back" buttons and
history lists, that can be used to redisplay a representation
retrieved earlier in a session.

The freshness model (Section 4.2) does not necessarily apply to
history mechanisms. That is, a history mechanism can display a
previous representation even if it has expired.

"""

Chrome and IE9+ [2] already ignore "cache-control: must-revalidate" on history navigation.

[1] ​https://tools.ietf.org/html/rfc7234#page-32
[2] ​http://blogs.msdn.com/b/ie/archive/2010/07/14/caching-improvements-in-internet-explorer-9.aspx

Test: http/tests/cache/disk-cache-validation-back-navigation-policy.html

  • NetworkProcess/cache/NetworkCache.cpp:

(WebKit::NetworkCache::decodeStorageEntry):

LayoutTests:

Rebaseline disk-cache-validation-back-navigation-policy.html test now
that we ignore "cache-control: no-cache, must-revalidate" on history
navigation. This is a progression.

  • http/tests/cache/disk-cache-validation-back-navigation-policy-expected.txt:

Mar 18, 2015:

11:30 PM Changeset in webkit [181733] by Csaba Osztrogonác
  • 2 edits in trunk

Unreviewed, kick the GTK bots to regenerate makefile.

  • Source/cmake/OptionsGTK.cmake:
11:22 PM Changeset in webkit [181732] by Brent Fulgham
  • 2 edits in trunk/LayoutTests

[Win] Windows gardening for debug bots.

  • platform/win/TestExpectations:
11:14 PM Changeset in webkit [181731] by bshafiei@apple.com
  • 5 edits in tags/Safari-601.1.23.3/Source

Versioning.

11:10 PM Changeset in webkit [181730] by bshafiei@apple.com
  • 1 copy in tags/Safari-601.1.23.3

New tag.

11:09 PM Changeset in webkit [181729] by Manuel Rego Casasnovas
  • 2 edits in trunk/Source/WebCore

Unreviewed. GTK build fix after r181720.

  • rendering/RenderImage.cpp:

(WebCore::RenderImage::styleDidChange):

10:21 PM Changeset in webkit [181728] by Antti Koivisto
  • 10 edits
    2 adds in trunk

Test disk cache behavior when using back navigation cache policy
​https://bugs.webkit.org/show_bug.cgi?id=142848

Reviewed by Chris Dumez.

Source/WebCore:

Add a way to override the normal cache policy so we can test reload behavior of cache without actually reloading.

Test: http/tests/cache/disk-cache-validation-back-navigation-policy.html

  • loader/FrameLoader.cpp:

(WebCore::FrameLoader::addExtraFieldsToRequest):

  • loader/FrameLoader.h:

(WebCore::FrameLoader::setOverrideCachePolicyForTesting):

  • testing/Internals.cpp:

(WebCore::Internals::setOverrideCachePolicy):

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

LayoutTests:

  • http/tests/cache/disk-cache-validation-back-navigation-policy-expected.txt: Added.
  • http/tests/cache/disk-cache-validation-back-navigation-policy.html: Added.

Test permutations of cache headers on reload including Cache-control: must-revalidate.

  • platform/mac-wk1/TestExpectations:
  • platform/win/TestExpectations:
10:16 PM Changeset in webkit [181727] by Yusuke Suzuki
  • 2 edits in trunk/Source/WebCore

TextFragment#start() is always >= 0 since its type is unsigned
​https://bugs.webkit.org/show_bug.cgi?id=142860

Reviewed by Andreas Kling.

Since TextFragment::m_start is unsigned, the result of
TextFragment::start() is always >= 0 and assertion is not effective.
This patch removes this assertion to suppress warnings.

  • rendering/SimpleLineLayoutTextFragmentIterator.cpp:

(WebCore::SimpleLineLayout::TextFragmentIterator::revertToFragment):

9:57 PM Changeset in webkit [181726] by commit-queue@webkit.org
  • 13 edits in trunk

[ContentExtensions] Prepare for compiling stylesheets of selectors to be used on every page.
​https://bugs.webkit.org/show_bug.cgi?id=142799

Patch by Alex Christensen <​achristensen@webkit.org> on 2015-03-18
Reviewed by Brady Eidson.

Source/WebCore:

  • WebCore.xcodeproj/project.pbxproj:

Make private headers to use with API tests.

  • contentextensions/CompiledContentExtension.cpp:

(WebCore::ContentExtensions::CompiledContentExtension::globalDisplayNoneSelectors):

  • contentextensions/CompiledContentExtension.h:

Added method to get only the selectors from the root of the DFA, which apply to all URLs.

  • contentextensions/ContentExtensionCompiler.cpp:

(WebCore::ContentExtensions::compileRuleList):
Added checking if the trigger will match everything.
These actions can be put directly on the root of the DFA without adding extra epsilon transitions to the NFA.

  • contentextensions/DFA.h:

(WebCore::ContentExtensions::DFA::nodeAt):

  • contentextensions/DFABytecodeInterpreter.cpp:

(WebCore::ContentExtensions::DFABytecodeInterpreter::actionsFromDFARoot):
(WebCore::ContentExtensions::DFABytecodeInterpreter::interpret):

  • contentextensions/DFABytecodeInterpreter.h:
  • contentextensions/NFA.h:
  • contentextensions/URLFilterParser.cpp:

(WebCore::ContentExtensions::Term::quantifier):
Sink terms to a vector then add nodes to NFA when finalizing after checking for regexes that match everything.
(WebCore::ContentExtensions::GraphBuilder::GraphBuilder):
(WebCore::ContentExtensions::GraphBuilder::finalize):
(WebCore::ContentExtensions::GraphBuilder::parseStatus):
(WebCore::ContentExtensions::GraphBuilder::atomPatternCharacter):
(WebCore::ContentExtensions::GraphBuilder::atomBuiltInCharacterClass):
(WebCore::ContentExtensions::GraphBuilder::quantifyAtom):
(WebCore::ContentExtensions::GraphBuilder::atomBackReference):
(WebCore::ContentExtensions::GraphBuilder::assertionBOL):
(WebCore::ContentExtensions::GraphBuilder::assertionWordBoundary):
(WebCore::ContentExtensions::GraphBuilder::atomCharacterClassAtom):
(WebCore::ContentExtensions::GraphBuilder::atomCharacterClassRange):
(WebCore::ContentExtensions::GraphBuilder::atomCharacterClassBuiltIn):
(WebCore::ContentExtensions::GraphBuilder::atomParentheticalAssertionBegin):
(WebCore::ContentExtensions::GraphBuilder::disjunction):
(WebCore::ContentExtensions::GraphBuilder::hasError):
(WebCore::ContentExtensions::GraphBuilder::fail):
(WebCore::ContentExtensions::GraphBuilder::sinkFloatingTermIfNecessary):
(WebCore::ContentExtensions::URLFilterParser::addPattern):
(WebCore::ContentExtensions::URLFilterParser::statusString):
(WebCore::ContentExtensions::GraphBuilder::errorMessage): Deleted.

  • contentextensions/URLFilterParser.h:

Use an enum instead of strings for the status to avoid checking strings when we have a regex that matches everything.

Tools:

  • TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:

(TestWebKitAPI::testPattern):
(TestWebKitAPI::TEST_F):
Start testing regex failures.

9:05 PM Changeset in webkit [181725] by Yusuke Suzuki
  • 2 edits in trunk/Source/WebCore

Fix build failure due to FALLTHROUGH in unreachable code
​https://bugs.webkit.org/show_bug.cgi?id=142703

Reviewed by Benjamin Poulain.

FALLTHROUGH should not exist in unreachable code marked by
ASSERT_NOT_REACHABLE.

When !ENABLE(CSS_SELECTOR_JIT) and Debug mode is enabled,
both ASSERT_NOT_REACHED and FALLTHROUGH are instantiated
and it cause build failure.

Here, since CompiledSingle/CompiledSingleWithRootFilter are
not set when !ENABLE(CSS_SELECTOR_JIT). So dropping FALLTHROUGH
and leave ASSERT_NOT_REACHED.
And since ASSERT_NOT_ReACHED is stripped when Release build,
I've inserted FALLTHROUGH() if ASSERT_DISABLED.
This fix is the same to r162906.

  • dom/SelectorQuery.cpp:

(WebCore::SelectorDataList::execute):

8:27 PM Changeset in webkit [181724] by Joseph Pecoraro
  • 5 edits in trunk

ES6 Classes: Extends should accept an expression without parenthesis
​https://bugs.webkit.org/show_bug.cgi?id=142840

Reviewed by Ryosuke Niwa.

Source/JavaScriptCore:

  • parser/Parser.cpp:

(JSC::Parser<LexerType>::parseClass):
"extends" allows a LeftHandExpression (new expression / call expression,
which includes a member expression), not a primary expression. Our
parseMemberExpression does all of these.

LayoutTests:

  • js/class-syntax-extends-expected.txt:
  • js/script-tests/class-syntax-extends.js:

Extend the test to test cases of extends with different types of expressions.

8:26 PM Changeset in webkit [181723] by Joseph Pecoraro
  • 7 edits
    1 delete in trunk/Source/WebInspectorUI

Web Inspector: Replace last use of ObjectPropertiesSection with ObjectTreeView
​https://bugs.webkit.org/show_bug.cgi?id=142834

Reviewed by Timothy Hatcher.

  • UserInterface/Views/DOMNodeDetailsSidebarPanel.js:

(WebInspector.DOMNodeDetailsSidebarPanel.prototype._refreshProperties):
Use properties only ObjectTreeView instead of ObjectPropertiesSection.
This doesn't fix the functionality issues, but fixes the appearance.

  • UserInterface/Views/DetailsSection.css:

(.details-section > .content > .group:first-child > .row.simple:first-child > *): Deleted.
(body.mac-platform.legacy .details-section > .content > .group:first-child > .row.simple:first-child > *): Deleted.
(.details-section > .content > .group > .row.properties:not(.empty)): Deleted.
(body.mac-platform.legacy .details-section > .content > .group > .row.properties:not(.empty)): Deleted.
Re-add back the small padding at the top of sections. This actually
broke padding in a few sections (Event Listeners) at the expense
of eliminating a few pixels of whitespace at the top of other sections.
We should focus on addressing the extra whitespace separately.

  • UserInterface/Main.html:
  • UserInterface/Views/ObjectPropertiesSection.js: Removed.
  • UserInterface/Views/TypePropertiesSection.js:
  • WebInspectorUI.vcxproj/WebInspectorUI.vcxproj:
  • WebInspectorUI.vcxproj/WebInspectorUI.vcxproj.filters:

Remove new unused class.

8:26 PM Changeset in webkit [181722] by Joseph Pecoraro
  • 7 edits in trunk/Source

Web Inspector: Debugger Popovers and Probes should use FormattedValue/ObjectTreeView instead of Custom/ObjectPropertiesSection
​https://bugs.webkit.org/show_bug.cgi?id=142830

Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

  • inspector/agents/InspectorDebuggerAgent.cpp:

(Inspector::InspectorDebuggerAgent::breakpointActionProbe):
Give Probe Samples object previews.

Source/WebInspectorUI:

  • UserInterface/Views/ProbeSetDataGrid.css:

(.details-section.probe-set .data-grid .object-tree > :matches(.title, .object-preview)::before):
Another line-height fix for object tree disclosure triangles.

  • UserInterface/Views/ProbeSetDataGridNode.js:

(WebInspector.ProbeSetDataGridNode.prototype.createCellContent):
Create an ObjectTree / FormattedValue for the RemoteObject.

  • UserInterface/Views/SourceCodeTextEditor.css:

(.popover .debugger-popover-content > .title):
(.popover .debugger-popover-content > .body):
(.popover .debugger-popover-content.function > .body):
Be more specific and don't accidentally style ".title" within the body.

  • UserInterface/Views/SourceCodeTextEditor.js:

(WebInspector.SourceCodeTextEditor.prototype._showPopoverForObject):
Show a properties only ObjectTree instead of an ObjectPropertiesSection.

(WebInspector.SourceCodeTextEditor.prototype._showPopoverForString): Deleted.
(WebInspector.SourceCodeTextEditor.prototype._showPopoverForRegExp): Deleted.
(WebInspector.SourceCodeTextEditor.prototype._showPopoverForNumber): Deleted.
(WebInspector.SourceCodeTextEditor.prototype._showPopoverForBoolean): Deleted.
(WebInspector.SourceCodeTextEditor.prototype._showPopoverForNull): Deleted.
(WebInspector.SourceCodeTextEditor.prototype._showPopoverForUndefined): Deleted.
(WebInspector.SourceCodeTextEditor.prototype._showPopoverWithFormattedValue):
Reduce most of these to a single popover for formatted values.

7:16 PM Changeset in webkit [181721] by Simon Fraser
  • 2 edits in trunk/Source/WebKit2

Don't paint PDFs on every scroll
​https://bugs.webkit.org/show_bug.cgi?id=142857
rdar://problem/20130207

Reviewed by Tim Horton.

Don't push a scale factor onto the PDF layer controller if it hasn't
changed, since doing so causes a repaint.

  • WebProcess/Plugins/PDF/PDFPlugin.mm:

(WebKit::PDFPlugin::updatePageAndDeviceScaleFactors):

7:15 PM Changeset in webkit [181720] by commit-queue@webkit.org
  • 5 edits
    2 adds in trunk

Switching between two SVG images with no intrinsic sizes causes them to get the default SVG size instead of the container size.
​https://bugs.webkit.org/show_bug.cgi?id=142805.

Patch by Said Abou-Hallawa <​sabouhallawa@apple.com> on 2015-03-18
Reviewed by Darin Adler.
Source/WebCore:

The bug happens due to wrong logic in RenderImage::imageDimensionsChanged().
This function decides to setNeedsLayout() if the intrinsic size of the image
changes. If the size does not change, it only repaints the image rectangle.
When switching the src of the an image between two SVG images and both of
them have no intrinsic size, we do not updateInnerContentRect() and this
means an SVGImageForContainer is not going to be created for this image.
When the image is drawn, it is drawn directly from the SVGImage. And this
means the drawing has to be scaled by container_size / SVG_default_intrinsic_size

After figuring out that I need to updateInnerContentRect() to fix this bug,
I found out Blink has already changed this code to do the same thing. But
they also did more clean-up in this function. Here is the link
​https://codereview.chromium.org/114323004. I think their change seems correct
although they did not say what exactly they were trying to fix.

The plan for repaintOrMarkForLayout(), which is the new name of this function,
is the following:

-- setNeedLayout() if the intrinsic size changes and it affects the size

of the image.

-- updateInnerContentRect() if the intrinsic size did not change but the

image has exiting layout.

-- repaint the image rectangle if layout is not needed.

This change also removes the call to computeLogicalWidthInRegion(), which is
almost running a layout for the image. This call figures out whether the image
needs to setNeedsLayout(). This call is unnecessary; the image needs to run a
layout if the intrinsic size has changed and it affects the size of the image.

Test: svg/as-image/svg-no-intrinsic-size-switching.html

  • rendering/RenderImage.cpp:

(WebCore::RenderImage::styleDidChange): Change the function call.
(WebCore::RenderImage::imageChanged): Rename local variable and change the
function call.

(WebCore::RenderImage::updateIntrinsicSizeIfNeeded): Simplify this function.
Call setIntrinsicSize() with the new size unless the image is in error state.

(WebCore::RenderImage::repaintOrMarkForLayout): This a better name for this
function since it is called even if the intrinsic size was not changed.
(WebCore::RenderImage::imageDimensionsChanged): Deleted.

  • rendering/RenderImage.h: Rename imageDimensionsChanged() and change the

updateIntrinsicSizeIfNeeded() to return void.

  • rendering/svg/RenderSVGForeignObject.cpp:

(WebCore::RenderSVGForeignObject::paint): Code cleanup. This function can
only handle the paint phases PaintPhaseForeground and PaintPhaseSelection.
Use this information to simplify the logic and order of painting there.

LayoutTests:

  • svg/as-image/svg-no-intrinsic-size-switching-expected.html: Added.
  • svg/as-image/svg-no-intrinsic-size-switching.html: Added.

Ensure that switching the source of an <img> element between two SVG images,
which have no intrinsic sizes, gets the image the size of the container and
not the default SVG intrinsic size which is 300x150 pixels.

6:28 PM Changeset in webkit [181719] by bshafiei@apple.com
  • 2 edits in tags/Safari-601.1.23.2/Source/WebCore

Merged r181713. rdar://problem/20178504

6:25 PM Changeset in webkit [181718] by bshafiei@apple.com
  • 5 edits in tags/Safari-601.1.23.2/Source

Versioning.

6:10 PM Changeset in webkit [181717] by diorahman@rockybars.com
  • 4 edits in trunk

Source/WTF:
Fix StringView misplaced implementations after r181525 and r181558
​https://bugs.webkit.org/show_bug.cgi?id=142772

Reviewed by Darin Adler.

Fix misplaced implementations in StringView::startsWithIgnoringASCIICase and
StringView::endsWith.

Add some tests in [1] to cover the StringView::startsWith,
StringView::startsWithIgnoringASCIICase, StringView::endsWith and
StringView::endsWithIgnoringASCIICase implementations.

[1] Tools/TestWebKitAPI/Tests/WTF/StringView.cpp.

  • wtf/text/StringView.cpp:

(WTF::StringView::startsWithIgnoringASCIICase):
(WTF::StringView::endsWith):
(WTF::StringView::endsWithIgnoringASCIICase):

Tools:
Fix StringView typos after r181525 and r181558
​https://bugs.webkit.org/show_bug.cgi?id=142772

Reviewed by Darin Adler.

Add some tests in [1] to cover the StringView::startsWith,
StringView::startsWithIgnoringASCIICase, StringView::endsWith and
StringView::endsWithIgnoringASCIICase implementations.

  • TestWebKitAPI/Tests/WTF/StringView.cpp:

(TestWebKitAPI::TEST):

5:49 PM Changeset in webkit [181716] by bshafiei@apple.com
  • 1 copy in tags/Safari-601.1.23.2

New tag.

5:14 PM Changeset in webkit [181715] by dburkart@apple.com
  • 2 edits in trunk/Tools

Remove extraneous import to fix the build.

Unreviewed.

5:01 PM Changeset in webkit [181714] by dburkart@apple.com
  • 2 edits in trunk/Tools

ASAN_OPTIONS=allocator_may_return_null=1 needs to be set
​https://bugs.webkit.org/show_bug.cgi?id=142547

Reviewed by Alexey Proskuryakov.

4:35 PM Changeset in webkit [181713] by jeremyj-wk@apple.com
  • 2 edits in trunk/Source/WebCore

Fix typo in playerViewControllerWillCancelOptimizedFullscree.
​https://bugs.webkit.org/show_bug.cgi?id=142745

Reviewed by Darin Adler.

  • platform/ios/WebVideoFullscreenInterfaceAVKit.mm:

(-[WebAVPlayerController playerViewControllerWillCancelOptimizedFullscreen:]):
(-[WebAVPlayerController playerViewControllerWillCancelOptimizedFullscree:]): Deleted.

4:29 PM Changeset in webkit [181712] by ap@apple.com
  • 2 edits in trunk/LayoutTests

webaudio/convolution-mono-mono.html fails on some machines
rdar://problem/19996807

Rubber-stampted by Jer Noble.

  • webaudio/resources/convolution-testing.js: (checkTriangularPulse): Slightly increase

the tolerance. It's still extremely low.

3:39 PM Changeset in webkit [181711] by andersca@apple.com
  • 2 edits in trunk/Source/WebKit2

Use && instead of & when deciding if the network cache should be cleared
​https://bugs.webkit.org/show_bug.cgi?id=142845

Reviewed by Antti Koivisto.

  • NetworkProcess/NetworkProcess.cpp:

(WebKit::NetworkProcess::deleteWebsiteData):

3:17 PM Changeset in webkit [181710] by Simon Fraser
  • 5 edits
    2 adds in trunk

Avoid repaints when changing transform on an element with multiple background images
​https://bugs.webkit.org/show_bug.cgi?id=142841

Reviewed by Zalan Bujtas.

Source/WebCore:

Replace the cheap test for changed images in RenderElement::updateFillImages()
with an exhaustive test that walks the entire list of background images,
since any ensuing repaint is way more expensive than a slightly more expensive check here.

Test: fast/repaint/multiple-backgrounds-style-change.html

  • rendering/RenderElement.cpp:

(WebCore::RenderElement::updateFillImages):

  • rendering/style/FillLayer.cpp:

(WebCore::layerImagesIdentical): See if both images are the same (either none
or both mask images, and same image pointer).
(WebCore::FillLayer::imagesIdentical): Walk the two FillLayer lists, checking the images
on each one. Returns false if we reach the end of one list before the other, or the images
are different.

  • rendering/style/FillLayer.h: New static function; static because

it compares two FillLayer lists, and I think that makes more sense than
a member function.

LayoutTests:

Test that changes transform on a composited element with 2 background images,
and tests for no repaints.

  • fast/repaint/multiple-backgrounds-style-change-expected.txt: Added.
  • fast/repaint/multiple-backgrounds-style-change.html: Added.
2:06 PM Changeset in webkit [181709] by andersca@apple.com
  • 2 edits in trunk/Source/WebCore

Pass cookies by reference in CookieHash functions
​https://bugs.webkit.org/show_bug.cgi?id=142839

Reviewed by Sam Weinig.

  • platform/Cookie.h:

(WebCore::CookieHash::hash):
(WebCore::CookieHash::equal):

1:42 PM Changeset in webkit [181708] by ap@apple.com
  • 2 edits in trunk/Tools

Tweak how AppleSystemFontOSSubversion default is added
​https://bugs.webkit.org/show_bug.cgi?id=142835

Reviewed by Dan Bernstein.

  • DumpRenderTree/mac/DumpRenderTree.mm:

(setDefaultsToConsistentValuesForTesting):

12:54 PM Changeset in webkit [181707] by andersca@apple.com
  • 11 edits
    2 copies in trunk/Source/WebKit2

Add WKContextGetWebsiteDataStore API
​https://bugs.webkit.org/show_bug.cgi?id=142832
rdar://problem/16544715

Reviewed by Sam Weinig.

Also add WKWebsiteDataStoreRef which is toll-free bridged to _WKWebsiteDataStore.

  • Shared/API/c/WKBase.h:

Add new declaration.

  • UIProcess/API/APIWebsiteDataStore.cpp:

(API::WebsiteDataStore::create):

  • UIProcess/API/APIWebsiteDataStore.h:

Add a new function that takes a configuration.

  • UIProcess/API/C/WKAPICast.h:

Add cast.

  • UIProcess/API/C/WKContext.cpp:

(WKContextGetWebsiteDataStore):
Return the context's website data store.

  • UIProcess/API/C/WKContext.h:

Add new declaration.

  • UIProcess/API/C/WKWebsiteDataStoreRef.cpp: Added.
  • UIProcess/API/C/WKWebsiteDataStoreRef.h: Added.

Add new files.

  • UIProcess/WebKeyValueStorageManager.cpp:

(WebKit::WebKeyValueStorageManager::getKeyValueStorageOrigins):
(WebKit::WebKeyValueStorageManager::getStorageDetailsByOrigin):
(WebKit::WebKeyValueStorageManager::deleteEntriesForOrigin):
(WebKit::WebKeyValueStorageManager::deleteAllEntries):
Update now that WebProcessPool holds on to an API::WebsiteDataStore.

  • UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::WebProcessPool):
(WebKit::WebProcessPool::createWebPage):

  • UIProcess/WebProcessPool.h:

WebProcessPool should hold on to an API::WebsiteDataStore instead of a WebKit::WebsiteDataStore.

  • WebKit2.xcodeproj/project.pbxproj:

Add new files.

12:07 PM Changeset in webkit [181706] by timothy_horton@apple.com
  • 2 edits in trunk/Source/WebCore

Temporarily fix the !ENABLE(CSS_SELECTOR_JIT) and assertions-enabled build

  • dom/SelectorQuery.cpp:

(WebCore::SelectorDataList::execute):

11:40 AM Changeset in webkit [181705] by timothy_horton@apple.com
  • 2 edits in trunk/Source/WebKit2

Cannot invoke action menus anymore
​https://bugs.webkit.org/show_bug.cgi?id=142797
<rdar://problem/20032670>

  • UIProcess/API/mac/WKView.mm:

Need one more respondsToSelector check.

11:30 AM Changeset in webkit [181704] by Joseph Pecoraro
  • 8 edits
    1 delete in trunk/Source/WebInspectorUI

Web Inspector: Scopes sidebar should use new ObjectTreeView and not ObjectPropertiesSection
​https://bugs.webkit.org/show_bug.cgi?id=142808

Reviewed by Timothy Hatcher.

  • UserInterface/Models/PropertyPath.js:

(WebInspector.PropertyPath):
(WebInspector.PropertyPath.emptyPropertyPathForScope):
Allow a special property empty path for "Scopes". This way for a
"<scopeObject>.property" we can show just the tooltip "property".

  • UserInterface/Views/ObjectTreeView.css:

(.object-tree.properties-only > :matches(.title, .object-preview)):
(.object-tree.properties-only .object-tree-outline):
(.object-tree.properties-only .object-tree-property .property-name):
Tweak styles for only properties view, which won't have a top-level
preview and doesn't fade out enumerable properties.

  • UserInterface/Views/ObjectTreeView.js:

(WebInspector.ObjectTreeView.prototype.get treeOutline):
Access the TreeOutline.

(WebInspector.ObjectTreeView.prototype.showOnlyProperties):
Properties only view modifies the display slightly.

(WebInspector.ObjectTreeView.prototype.appendExtraPropertyDescriptor):
(WebInspector.ObjectTreeView.prototype._updateProperties):
Allow the client to add its own property descriptors to display
as a property in this ObjectTreeView.

  • UserInterface/Views/ScopeChainDetailsSidebarPanel.js:

(WebInspector.ScopeChainDetailsSidebarPanel.prototype.refresh):
Switch to using an ObjectTreeView.

(WebInspector.ScopeChainDetailsSidebarPanel.prototype._propertyPathIdentifierForTreeElement):
(WebInspector.ScopeChainDetailsSidebarPanel.prototype._objectTreeAddHandler):
(WebInspector.ScopeChainDetailsSidebarPanel.prototype._objectTreeExpandHandler):
(WebInspector.ScopeChainDetailsSidebarPanel.prototype._objectTreeCollapseHandler):
Keep track of what properties were expanded so we can auto-expand
them again when the sidebar refreshes.

  • UserInterface/Main.html:
  • UserInterface/Views/ScopeVariableTreeElement.js: Removed.
  • WebInspectorUI.vcxproj/WebInspectorUI.vcxproj:
  • WebInspectorUI.vcxproj/WebInspectorUI.vcxproj.filters:

Remove the now unused ScopeVariableTreeElement.js.

11:12 AM Changeset in webkit [181703] by mmaxfield@apple.com
  • 3 edits in trunk/Source/WebCore

Call CTFontSetRenderingParameters before rendering text
​https://bugs.webkit.org/show_bug.cgi?id=142816

Reviewed by Darin Adler.

No new tests.

  • platform/graphics/cocoa/FontCascadeCocoa.mm:

(WebCore::showGlyphsWithAdvances):

  • platform/spi/cocoa/CoreTextSPI.h:
10:50 AM Changeset in webkit [181702] by Chris Dumez
  • 5 edits in trunk/Source

[WK2] Log the number of network cache requests that we have never seen before
​https://bugs.webkit.org/show_bug.cgi?id=142828
<rdar://problem/19632130>

Reviewed by Antti Koivisto.

Source/WebCore:

Add diagnostic logging key for network cache efficacy logging.

  • page/DiagnosticLoggingKeys.cpp:

(WebCore::DiagnosticLoggingKeys::neverSeenBeforeKey):
(WebCore::DiagnosticLoggingKeys::requestKey):

  • page/DiagnosticLoggingKeys.h:

Source/WebKit2:

Log the number of network cache requests that we have never seen before
using diagnostic logging.

  • NetworkProcess/cache/NetworkCacheStatisticsCocoa.mm:

(WebKit::NetworkCache::Statistics::recordNotUsingCacheForRequest):
(WebKit::NetworkCache::Statistics::recordRetrievalFailure):

10:40 AM Changeset in webkit [181701] by commit-queue@webkit.org
  • 2 edits in trunk/LayoutTests

Unreviewed GTK Gardening 18th March
​https://bugs.webkit.org/show_bug.cgi?id=142824

Unreviewed.

Patch by Marcos Chavarría Teijeiro <​chavarria1991@gmail.com> on 2015-03-18

  • platform/gtk/TestExpectations:
10:40 AM Changeset in webkit [181700] by Antti Koivisto
  • 6 edits in trunk/Source/WebKit2

Prune least valuable cache entries first
​https://bugs.webkit.org/show_bug.cgi?id=142810
rdar://problem/19632130

Reviewed by Darin Adler.

When pruning the cache estimate relative value of each entry using formula

age = current time - creation time
accessAge = last access time - creation time
value = accessAge / age

That is, we value entries that have been accessed recently and survived in the cache longest.

The most valuable entries don't get deleted at all while the deletion probablity ramps up for
lower value entries.

  • NetworkProcess/cache/NetworkCacheFileSystemPosix.h:

(WebKit::NetworkCache::fileTimes):

Get the creation time and access time for a file.

(WebKit::NetworkCache::updateFileAccessTimeIfNeeded):

Update access time manually if the file system doesn't do it automatically.

  • NetworkProcess/cache/NetworkCacheIOChannel.h:

(WebKit::NetworkCache::IOChannel::path):
(WebKit::NetworkCache::IOChannel::type):

  • NetworkProcess/cache/NetworkCacheIOChannelCocoa.mm:

(WebKit::NetworkCache::IOChannel::IOChannel):
(WebKit::NetworkCache::IOChannel::open):

Remember the file path and move most of the work to constructor.

  • NetworkProcess/cache/NetworkCacheStorage.cpp:

(WebKit::NetworkCache::Storage::Storage):
(WebKit::NetworkCache::Storage::remove):
(WebKit::NetworkCache::Storage::updateFileAccessTime):
(WebKit::NetworkCache::Storage::dispatchReadOperation):
(WebKit::NetworkCache::deletionProbability):

Compute the probability based on entry age and access time.

(WebKit::NetworkCache::Storage::shrinkIfNeeded):

  • NetworkProcess/cache/NetworkCacheStorage.h:

(WebKit::NetworkCache::Storage::serialBackgroundIOQueue):
(WebKit::NetworkCache::Storage::deleteQueue): Deleted.

More generic name.

10:27 AM Changeset in webkit [181699] by Yusuke Suzuki
  • 4 edits
    4 adds in trunk

Use filterRootId in SelectorQuery even if CSS JIT is not enabled
​https://bugs.webkit.org/show_bug.cgi?id=142703

Reviewed by Benjamin Poulain.

Source/WebCore:

filterRootId pruning path is not inherently related to CSS JIT.
This patch enables filterRootId even in the environment
where CSS JIT is not enabled.

  • dom/SelectorQuery.cpp:

(WebCore::filterRootById):
(WebCore::SelectorDataList::executeSingleSelectorData):
(WebCore::SelectorDataList::execute):

  • dom/SelectorQuery.h:

LayoutTests:

This is covered by the following layout-tests.
CompilableSingleWithRootFilter and CompilableSingle are covered by them.

  • fast/selectors/filter-root-node-with-selector-contains-adjacents.html
  • fast/selectors/querySelector-id-filtering.html
  • fast/selectors/querySelector-scope-filtered-root.html

And add tests to cover the following cases.

  • Failures on selectorForIdLookup()'s rootNode.inDocument().
  • Failures on selectorForIdLookup()'s rootNode.document().inQuirksMode().
  • fast/selectors/querySelector-with-id-for-detached-element-expected.txt: Added.
  • fast/selectors/querySelector-with-id-for-detached-element.html: Added.
  • fast/selectors/querySelector-with-id-in-quirks-mode-expected.txt: Added.
  • fast/selectors/querySelector-with-id-in-quirks-mode.html: Added.
9:51 AM Changeset in webkit [181698] by Alan Bujtas
  • 3 edits in trunk/Source/WebCore

Simple line layout: Use Vector<>::const_iterator instead of custom FlowContents::Iterator.
​https://bugs.webkit.org/show_bug.cgi?id=142809

Reviewed by Antti Koivisto.

FlowContents::Iterator simply iterates on a vector<>. No need to custom implement it.

No change in functionality.

  • rendering/SimpleLineLayoutFlowContents.h:

(WebCore::SimpleLineLayout::FlowContents::begin):
(WebCore::SimpleLineLayout::FlowContents::end):
(WebCore::SimpleLineLayout::FlowContents::Iterator::Iterator): Deleted.
(WebCore::SimpleLineLayout::FlowContents::Iterator::operator++): Deleted.
(WebCore::SimpleLineLayout::FlowContents::Iterator::operator--): Deleted.
(WebCore::SimpleLineLayout::FlowContents::Iterator::operator==): Deleted.
(WebCore::SimpleLineLayout::FlowContents::Iterator::operator!=): Deleted.
(WebCore::SimpleLineLayout::FlowContents::Iterator::operator*): Deleted.
(WebCore::SimpleLineLayout::FlowContents::Iterator::operator->): Deleted.

  • rendering/SimpleLineLayoutTextFragmentIterator.cpp:

(WebCore::SimpleLineLayout::TextFragmentIterator::skipToNextPosition):

9:47 AM Changeset in webkit [181697] by Alan Bujtas
  • 4 edits in trunk/Source/WebCore

Simple line layout: Change FlowContents::segmentForPosition() to segmentForRun().
​https://bugs.webkit.org/show_bug.cgi?id=142785

Reviewed by Antti Koivisto.

This is in transition to support <br>. A particular position could point to multiple
segments when <br> is directly followed by text.

No change in functionality.

  • rendering/SimpleLineLayoutFlowContents.cpp:

(WebCore::SimpleLineLayout::FlowContents::segmentIndexForRunSlow):
(WebCore::SimpleLineLayout::FlowContents::segmentIndexForPositionSlow): Deleted.

  • rendering/SimpleLineLayoutFlowContents.h:

(WebCore::SimpleLineLayout::FlowContents::segmentForRun):
(WebCore::SimpleLineLayout::FlowContents::segmentIndexForPosition): Deleted.
(WebCore::SimpleLineLayout::FlowContents::segmentForPosition): Deleted.

  • rendering/SimpleLineLayoutResolver.cpp:

(WebCore::SimpleLineLayout::RunResolver::Run::text):

9:27 AM Changeset in webkit [181696] by mitz@apple.com
  • 2 edits in trunk/Tools

prepare-ChangeLog doesn't understand C string literals split across multiple lines with \
​https://bugs.webkit.org/show_bug.cgi?id=142815

Reviewed by Darin Adler.

  • Scripts/prepare-ChangeLog:

(get_function_line_ranges_for_cpp): If the line ends with a backslash instead of a matching
quotation mark, use new variable $quotation_mark to remember what we are looking for, and
keep consuming the quoted text until the matching quotation mark is reached. Emit the
warning only if a line ends without a backslash before the matching quotation mark was found.

9:22 AM Changeset in webkit [181695] by Simon Fraser
  • 2 edits in trunk/Source/WebCore

Skip trying to paint overlay scrollbars when there are none or they are clipped out
​https://bugs.webkit.org/show_bug.cgi?id=142811
rdar://problem/20200725

Reviewed by Darin Adler.

In some content with lots of layers and overflow:scroll, we could spend 20% of
the time under paintOverflowControlsForFragments() setting up an (empty) clip,
and then trying to draw scrollbars that we don't have.

Avoid calling paintOverflowControlsForFragments() if there are no scrollbars,
and don't both setting up an empty clip just to paint nothing.

  • rendering/RenderLayer.cpp:

(WebCore::RenderLayer::paintLayerContents):
(WebCore::RenderLayer::paintOverflowControlsForFragments):

9:15 AM Changeset in webkit [181694] by peavo@outlook.com
  • 2 edits in trunk/Source/WebCore

[WinCairo] Unreviewed build fix after r181665.

  • platform/graphics/win/MediaPlayerPrivateMediaFoundation.cpp:

(WebCore::MediaPlayerPrivateMediaFoundation::setSize):

8:56 AM Changeset in webkit [181693] by commit-queue@webkit.org
  • 5 edits in trunk

Remove unused "preprocessor" parameter to sub-CodeGenerators
​https://bugs.webkit.org/show_bug.cgi?id=142793

Patch by Joseph Pecoraro <Joseph Pecoraro> on 2015-03-18
Reviewed by Darin Adler.

Source/WebCore:

  • bindings/scripts/CodeGenerator.pm:

(ProcessDocument):
(FileNamePrefix):

  • bindings/scripts/CodeGeneratorObjC.pm:

Tools:

  • WebKitTestRunner/InjectedBundle/Bindings/CodeGeneratorTestRunner.pm:

(new):

7:48 AM Changeset in webkit [181692] by Alan Bujtas
  • 7 edits
    2 adds in trunk

Simple line layout: Split fragments on renderer boundary on the fly.
​https://bugs.webkit.org/show_bug.cgi?id=142579

Reviewed by Antti Koivisto.

Fragment splitting at renderers' boundary at the end of the line is no longer needed.
This patch ensures that TextFragmentIterator::nextTextFragment() does not concatenate
fragments across renderer boundary.

Source/WebCore:

Test: fast/text/simple-line-with-multiple-renderers.html

  • rendering/SimpleLineLayout.cpp:

(WebCore::SimpleLineLayout::revertRuns):
(WebCore::SimpleLineLayout::LineState::isEmpty):
(WebCore::SimpleLineLayout::LineState::appendFragmentAndCreateRunIfNeeded): Check if fragments need a new run and
whether neighboring fragments need collapsing.
(WebCore::SimpleLineLayout::LineState::revertUncommitted): Revert fragments over multiple renderers if they form
a fragment continuation.
(WebCore::SimpleLineLayout::LineState::removeTrailingWhitespace):
(WebCore::SimpleLineLayout::forceFragmentToLine): Ensure that if a fragment has continuation across multiple renderers,
they all get added to the current line.
(WebCore::SimpleLineLayout::createLineRuns):
(WebCore::SimpleLineLayout::LineState::appendFragment): Deleted.
(WebCore::SimpleLineLayout::FragmentForwardIterator::FragmentForwardIterator): Deleted.
(WebCore::SimpleLineLayout::FragmentForwardIterator::operator++): Deleted.
(WebCore::SimpleLineLayout::FragmentForwardIterator::operator!=): Deleted.
(WebCore::SimpleLineLayout::FragmentForwardIterator::operator*): Deleted.
(WebCore::SimpleLineLayout::begin): Deleted.
(WebCore::SimpleLineLayout::end): Deleted.
(WebCore::SimpleLineLayout::preWrap): Deleted.
(WebCore::SimpleLineLayout::removeTrailingWhitespace): Deleted.
(WebCore::SimpleLineLayout::updateLineConstrains): Deleted.
(WebCore::SimpleLineLayout::splitRunsAtRendererBoundary): Deleted.
(WebCore::SimpleLineLayout::createTextRuns): Deleted.
(WebCore::SimpleLineLayout::create): Deleted.
(WebCore::SimpleLineLayout::Layout::create): Deleted.

  • rendering/SimpleLineLayoutFlowContents.h:

(WebCore::SimpleLineLayout::FlowContents::Iterator::operator+):
(WebCore::SimpleLineLayout::FlowContents::Iterator::operator*):
(WebCore::SimpleLineLayout::FlowContents::Iterator::operator->):

  • rendering/SimpleLineLayoutFunctions.cpp:

(WebCore::SimpleLineLayout::paintFlow): Unrelated fix on non-collapsed tab position.

  • rendering/SimpleLineLayoutTextFragmentIterator.cpp:

(WebCore::SimpleLineLayout::TextFragmentIterator::TextFragmentIterator):
(WebCore::SimpleLineLayout::TextFragmentIterator::nextTextFragment):
(WebCore::SimpleLineLayout::TextFragmentIterator::revertToFragment):
(WebCore::SimpleLineLayout::TextFragmentIterator::textWidth):
(WebCore::SimpleLineLayout::TextFragmentIterator::skipToNextPosition):

  • rendering/SimpleLineLayoutTextFragmentIterator.h:

(WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::TextFragment):
(WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::overlapsToNextRenderer):
(WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::isCollapsible):
(WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::operator==):
(WebCore::SimpleLineLayout::TextFragmentIterator::isLineBreak):
(WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::split): Deleted.

LayoutTests:

  • fast/text/simple-line-with-multiple-renderers-expected.html: Added.
  • fast/text/simple-line-with-multiple-renderers.html: Added.
6:47 AM Changeset in webkit [181691] by Manuel Rego Casasnovas
  • 11 edits
    4 adds in trunk

Flex and grid items should be painted as inline-blocks
​https://bugs.webkit.org/show_bug.cgi?id=142266

Reviewed by Darin Adler.

Source/WebCore:

Based on Blink r157004 by <​cbiesinger@chromium.org>.
​https://src.chromium.org/viewvc/blink?revision=157004&view=revision

Both flexbox and grid specs define that the painting order of flex/grid
items is the same as inline blocks. See
​http://dev.w3.org/csswg/css-flexbox/#painting and
​http://dev.w3.org/csswg/css-grid/#z-order.

Extracted inline blocks painting code from InlineElementBox and moved to
a helper method that will be reused for flexboxes and grids.

Tests: css3/flexbox/flex-item-text-background-not-interleaved.html

fast/css-grid-layout/grid-item-text-background-not-interleaved.html

  • rendering/InlineElementBox.cpp:

(WebCore::InlineElementBox::paint): Move code to
RenderElement::paintAsInlineBlock().

  • rendering/RenderBlock.cpp:

(WebCore::RenderBlock::paintChild): Add new argument to paint children
as inline blocks.

  • rendering/RenderBlock.h: Define PaintType enmu and modify paintChild()

signature to add the new argument.

  • rendering/RenderElement.cpp:

(WebCore::paintPhase): Paint element in a phase.
(WebCore::RenderElement::paintAsInlineBlock): Code extracted from
InlineElementBox::paint().

  • rendering/RenderElement.h: Add new method signature.
  • rendering/RenderFlexibleBox.cpp:

(WebCore::RenderFlexibleBox::paintChildren): Call
RenderBlock::paintChild() with the new argument.

  • rendering/RenderGrid.cpp:

(WebCore::RenderGrid::paintChildren): Ditto.

LayoutTests:

  • css3/flexbox/flex-item-text-background-not-interleaved-expected.html: Added.
  • css3/flexbox/flex-item-text-background-not-interleaved.html: Added.
  • fast/css-grid-layout/float-not-protruding-into-next-grid-item-expected.html:

Add some vertical space to avoid issues with backgrounds.

  • fast/css-grid-layout/float-not-protruding-into-next-grid-item.html:

Ditto.

  • fast/css-grid-layout/grid-item-text-background-not-interleaved-expected.html: Added.
  • fast/css-grid-layout/grid-item-text-background-not-interleaved.html: Added.
6:21 AM WebKitGTK/Gardening/Calendar edited by chavarria1991@gmail.com
(diff)
3:47 AM WebKitGTK/Gardening/Calendar edited by chavarria1991@gmail.com
(diff)
2:32 AM Changeset in webkit [181690] by Csaba Osztrogonác
  • 3 edits
    1 add
    1 delete in trunk/Tools

[EFL] Bump gstreamer version to 1.4.4
​https://bugs.webkit.org/show_bug.cgi?id=142782

Reviewed by Philippe Normand.

  • efl/jhbuild.modules:
  • efl/patches/gst-libav.patch:
  • efl/patches/gst-plugins-bad-remove-gnustep-support.patch: Added.
  • efl/patches/gst-prevent-neon-check-in-configure-from-passing-under-aarch64.patch: Removed.
2:31 AM March 2015 Meeting edited by jfernandez@igalia.com
(diff)
1:51 AM March 2015 Meeting edited by adam.bergkvist@ericsson.com
(diff)
1:48 AM webrtc_in_webkit_2015.pdf attached to March 2015 Meeting by adam.bergkvist@ericsson.com
Slides from the WebRTC session at the 2015 WebKit contributors meeting
12:27 AM Changeset in webkit [181689] by Chris Dumez
  • 7 edits in trunk/Source

[WK2] Log total number of network cache queries using diagnostic logging
​https://bugs.webkit.org/show_bug.cgi?id=142803
<rdar://problem/19632130>

Reviewed by Antti Koivisto.

Source/WebCore:

Add diagnostic logging key needed for network efficacy logging.

  • page/DiagnosticLoggingKeys.cpp:

(WebCore::DiagnosticLoggingKeys::retrievalRequestKey):

  • page/DiagnosticLoggingKeys.h:

Source/WebKit2:

Log total number of network cache queries using diagnostic logging.

  • NetworkProcess/cache/NetworkCache.cpp:

(WebKit::NetworkCache::Cache::retrieve):

  • NetworkProcess/cache/NetworkCacheStatistics.h:
  • NetworkProcess/cache/NetworkCacheStatisticsCocoa.mm:

(WebKit::NetworkCache::Statistics::recordCacheRetrievalRequest):

12:26 AM Changeset in webkit [181688] by Chris Dumez
  • 5 edits in trunk/Source/WebKit2

[WK2] We should not even try the network cache for non-HTTP protocol requests
​https://bugs.webkit.org/show_bug.cgi?id=142802
<rdar://problem/19632130>

Reviewed by Antti Koivisto.

We should not even try the network cache for non-HTTP protocol requests
and the network cache cannot handle those.

  • NetworkProcess/NetworkResourceLoader.cpp:

(WebKit::NetworkResourceLoader::start):

  • NetworkProcess/cache/NetworkCache.cpp:

(WebKit::NetworkCache::canRetrieve):
(WebKit::NetworkCache::Cache::retrieve):

  • NetworkProcess/cache/NetworkCache.h:
  • NetworkProcess/cache/NetworkCacheStatisticsCocoa.mm:

(WebKit::NetworkCache::retrieveDecisionToDiagnosticKey):

12:19 AM Changeset in webkit [181687] by timothy_horton@apple.com
  • 3 edits in trunk/Source/WebCore

Fix the build after r181660

  • page/EventHandler.cpp:

(WebCore::EventHandler::setImmediateActionStage):

  • page/EventHandler.h:

(WebCore::EventHandler::setImmediateActionStage):
Un-inline setImmediateActionStage, because it is exported to WebCore, and
exporting inline thing causes random weak symbol errors on some platforms.

12:05 AM Changeset in webkit [181686] by Conrad Shultz
  • 6 edits in trunk/Source

Ignore some deprecation warnings
​https://bugs.webkit.org/show_bug.cgi?id=142813

Reviewed by Mark Rowe.

Fixing the deprecations is tracked by: <rdar://problem/20201450>

Source/WebCore:

  • editing/cocoa/HTMLConverter.mm:

(HTMLConverter::_processElement):

Source/WebKit/mac:

  • WebCoreSupport/PopupMenuMac.mm:

(PopupMenuMac::populate):

Source/WebKit2:

  • UIProcess/mac/WebPopupMenuProxyMac.mm:

(WebKit::WebPopupMenuProxyMac::populate):

Note: See TracTimeline for information about the timeline view.