Timeline
Jun 6, 2015:
- 9:42 PM Changeset in webkit [185299] by
-
- 2 edits in trunk/Source/WebKit2
Don’t enable manual sandboxing in the iOS Simulator. This should fix all the tests timing out.
- Configurations/DebugRelease.xcconfig:
- 7:08 PM Changeset in webkit [185298] by
-
- 13 edits in trunk/Source/WebKit2
[iOS] Can’t debug with Development XPC services due to automatic sandboxing
https://bugs.webkit.org/show_bug.cgi?id=145731
Reviewed by Darin Adler.
In iOS, the XPC services are automatically sandboxed based on their entitlements. The
sandbox doesn’t allow them to re-exec, which is necessary for linking against engineering
builds of the WebKit frameworks placed in a non-system location.
The solution is to opt Engineering builds into manual sandboxing.
- Configurations/BaseTarget.xcconfig:
- Add ENABLE_MANUAL_SANDBOXING to GCC_PREPROCESSOR_DEFINITIONS when WK_MANUAL_SANDBOXING_ENABLED is YES.
- Set WK_MANUAL_SANDBOXING_ENABLED to YES for OS X, because we always use manual sandboxing on that platform.
- Configurations/DebugRelease.xcconfig: Removed unused definition of DEBUG_OR_RELEASE. Set WK_MANUAL_SANDBOXING_ENABLED to YES for all engineering builds.
- Configurations/BaseXPCService.xcconfig: Define CODE_SIGN_ENTITLEMENTS for iOS to be empty if the service is not installed; to use the base entitlements file from Configurations if manual sandboxing is not enabled; and to use a derived entitlements file, which doesn’t specify a sandbox profile, if manual sandboxing is enabled.
- Configurations/DatabaseService.Development.xcconfig: Define WK_XPC_SERVICE_IOS_ENTITLEMENTS_BASE and let BaseXPCService.xcconfig define CODE_SIGN_ENTITLEMENTS for iOS.
- Configurations/DatabaseService.xcconfig: Ditto.
- Configurations/NetworkService.Development.xcconfig: Ditto.
- Configurations/NetworkService.xcconfig: Ditto.
- Configurations/WebContentService.Development.xcconfig: Ditto.
- Configurations/WebContentService.xcconfig: Ditto.
- WebKit2.xcodeproj/project.pbxproj:
- In the Derived Sources target, added a Derive Entitlements for Manual Sandboxing script build phase. When building for iOS with manual sandboxing enabled, it creates copies of the iOS entitlements files with a -no-sandbox suffix and the seatbelt-profiles key removed.
- In the WebKit target, added a Copy iOS Sandbox Profiles for Manual Sandboxing script build phase. When building for iOS with manual sandboxing enabled, it copies the iOS sandbox profiles into the framework, so that the services can apply them to themselves at runtime.
- Shared/ios/ChildProcessIOS.mm:
(WebKit::ChildProcess::initializeSandbox): Use ENABLE(MANUAL_SANDBOXING) and don’t define it
here.
- WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::initializeSandbox): Use ENABLE(MANUAL_SANDBOXING) instead of defining
and using ENABLE_MANUAL_WEBPROCESS_SANDBOXING here.
- 7:07 PM Changeset in webkit [185297] by
-
- 2 edits in trunk/Source/WebKit/mac
Restored binary compatibility with shipping Safari.
Suggested by Anders Carlsson.
- Misc/WebKitNSStringExtras.mm:
(-[NSString _webkit_fixedCarbonPOSIXPath]): Added an implementation that returns self, so
that Safari doesn’t crash when trying to call this method.
- 5:09 PM Changeset in webkit [185296] by
-
- 9 edits1 add in trunk/LayoutTests
Clean up tests for blocking mixed content
https://bugs.webkit.org/show_bug.cgi?id=145718
Reviewed by Darin Adler.
Stop using timeouts, which were making the tests flaky.
- http/tests/xmlhttprequest/access-control-response-with-body.html: This wasn't even
a mixed content test, but the file was reused in one. I now made a separate helper
document, so this is just a regular test now, not another test's resource.
- http/tests/security/mixedContent/redirect-http-to-https-iframe-in-main-frame.html:
- http/tests/security/mixedContent/redirect-https-to-http-iframe-in-main-frame.html:
- http/tests/security/mixedContent/resources/frame-with-insecure-frame.html:
- http/tests/security/mixedContent/resources/frame-with-redirect-http-to-https-frame.html:
- http/tests/security/mixedContent/resources/frame-with-redirect-https-to-http-frame.html:
Refactored to not depend on timeouts.
- http/tests/security/mixedContent/insecure-xhr-in-main-frame-expected.txt:
- http/tests/security/mixedContent/insecure-xhr-in-main-frame.html:
- http/tests/security/mixedContent/resources/insecure-xhr-in-main-frame-window.html: Added.
This test still needs a timeout, because of a WebKit bug. I just made it much longer.
- 3:55 PM Changeset in webkit [185295] by
-
- 2 edits in trunk/Source/WebKit/mac
Fixed the build.
- Plugins/WebBasePluginPackage.mm:
- 2:22 PM Changeset in webkit [185294] by
-
- 10 edits in trunk/Source
Remove support from getting plug-in information from Carbon resources
https://bugs.webkit.org/show_bug.cgi?id=145730
Reviewed by Dan Bernstein.
Source/WebCore:
Remove a now unused function.
- platform/mac/WebCoreNSStringExtras.h:
- platform/mac/WebCoreNSStringExtras.mm:
(stringEncodingForResource): Deleted.
Source/WebKit/mac:
No current plug-in versions use this feature, and both Chrome and Mozilla for
Mac stopped looking at Carbon resources years ago.
- Misc/WebKitNSStringExtras.h:
- Misc/WebKitNSStringExtras.mm:
(+[NSString _web_encodingForResource:]): Deleted.
- Plugins/WebBasePluginPackage.mm:
- Plugins/WebNetscapePluginPackage.mm:
(-[WebNetscapePluginPackage _initWithPath:]):
(-[WebNetscapePluginPackage stringForStringListID:andIndex:]): Deleted.
(-[WebNetscapePluginPackage getPluginInfoFromResources]): Deleted.
Source/WebKit2:
No current plug-in versions use this feature, and both Chrome and Mozilla for
Mac stopped looking at Carbon resources years ago.
- Shared/Plugins/Netscape/mac/NetscapePluginModuleMac.mm:
(WebKit::NetscapePluginModule::getPluginInfo):
(WebKit::ResourceMap::ResourceMap): Deleted.
(WebKit::ResourceMap::~ResourceMap): Deleted.
(WebKit::ResourceMap::isValid): Deleted.
(WebKit::getStringListResource): Deleted.
(WebKit::getPluginInfoFromCarbonResources): Deleted.
- 1:01 PM Changeset in webkit [185293] by
-
- 6 edits in trunk/Source/WebKit/mac
Remove now unused category methods
https://bugs.webkit.org/show_bug.cgi?id=145729
Reviewed by Dan Bernstein.
- Configurations/WebKitLegacy.xcconfig:
- Misc/WebKitNSStringExtras.h:
- Misc/WebKitNSStringExtras.mm:
(-[NSString _webkit_fixedCarbonPOSIXPath]): Deleted.
- Misc/WebNSFileManagerExtras.h:
- Misc/WebNSFileManagerExtras.mm:
(-[NSFileManager _webkit_startupVolumeName]): Deleted.
- 12:50 PM Changeset in webkit [185292] by
-
- 2 edits in trunk/Source/WebKit/mac
Fix build.
- WebKit.mac.exp:
- 12:45 PM Changeset in webkit [185291] by
-
- 3 edits in trunk/Source/WebKit/mac
Stop using _webkit_fixedCarbonPOSIXPath in plug-in code
https://bugs.webkit.org/show_bug.cgi?id=145728
Reviewed by Dan Bernstein.
This code was put in place to support old plug-ins that would pass Carbon paths to NPAPI functions.
WebKit2 never supported this and we've never had any problems so let's just get rid of it.
- Plugins/Hosted/NetscapePluginInstanceProxy.mm:
(WebKit::NetscapePluginInstanceProxy::loadURL):
- Plugins/WebNetscapePluginView.mm:
(-[WebNetscapePluginView _postURL:target:len:buf:file:notifyData:sendNotification:allowHeaders:]):
- 12:37 PM Changeset in webkit [185290] by
-
- 3 edits2 deletes in trunk/Source/WebKit
Remove SPI that's no longer used
https://bugs.webkit.org/show_bug.cgi?id=145727
Reviewed by Dan Bernstein.
Source/WebKit:
- WebKit.xcodeproj/project.pbxproj:
Source/WebKit/mac:
- Plugins/WebPluginsPrivate.h: Removed.
- Plugins/WebPluginsPrivate.m: Removed.
- 12:22 PM Changeset in webkit [185289] by
-
- 2 edits in trunk/Source/WebKit/mac
Only allow plug-ins to create custom MIME property lists on older versions of OS X
https://bugs.webkit.org/show_bug.cgi?id=145726
Reviewed by Dan Bernstein.
- Plugins/WebBasePluginPackage.mm:
(-[WebBasePluginPackage getPluginInfoFromPLists]):
- 1:43 AM Changeset in webkit [185288] by
-
- 2 edits in trunk/Source/WebCore
Unreviewed, rolling out r185284.
https://bugs.webkit.org/show_bug.cgi?id=145723
it needs some discussion before landing (Requested by dino on
#webkit).
Reverted changeset:
"[iOS] Don't force compositing layers for no-op 3D transforms
on low-memory devices."
https://bugs.webkit.org/show_bug.cgi?id=145719
http://trac.webkit.org/changeset/185284
- 1:21 AM Changeset in webkit [185287] by
-
- 3 edits in trunk/Source/WebCore
Typing is slow in Gmail on iPads
https://bugs.webkit.org/show_bug.cgi?id=145686
Reviewed by Enrica Casucci.
The bug was caused by nextCandidate and nextVisuallyDistinctCandidate traversing through each character
in a text node without a renderer. Skip any node that doesn't have a renderer in both of those functions
and corresponding previous* functions.
It's fine to skip unrendered nodes in PositionIterator because only other clients of PositionIterator
are Position::upstream and Position::downstream and they don't care about un-rendered nodes either.
- dom/PositionIterator.cpp:
(WebCore::PositionIterator::increment):
(WebCore::PositionIterator::decrement):
- editing/htmlediting.cpp:
(WebCore::nextVisuallyDistinctCandidate):
(WebCore::previousVisuallyDistinctCandidate):
- 12:55 AM Changeset in webkit [185286] by
-
- 6 edits in trunk
Returned Exception* values need to be initialized to nullptr when no exceptions are thrown.
https://bugs.webkit.org/show_bug.cgi?id=145720
Reviewed by Dan Bernstein.
Source/JavaScriptCore:
- debugger/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::evaluate):
Source/WebCore:
No new tests because this issue is already covered by the existing inspector-protocol tests.
- bindings/js/WorkerScriptController.cpp:
(WebCore::WorkerScriptController::evaluate):
LayoutTests:
- platform/win/TestExpectations:
- Undoing skipped tests.
Jun 5, 2015:
- 9:40 PM Changeset in webkit [185285] by
-
- 2 edits in trunk/Tools
From r185273 on WTF::RefCounter exposes the actual refcount instead of a boolean value.
Reviewed by Dan Bernstein.
- TestWebKitAPI/Tests/WTF/RefCounter.cpp:
(TestWebKitAPI::TEST):
- 8:11 PM Changeset in webkit [185284] by
-
- 2 edits in trunk/Source/WebCore
[iOS] Don't force compositing layers for no-op 3D transforms on low-memory devices.
<https://webkit.org/b/145719>
<rdar://problem/19973042>
Reviewed by Geoffrey Garen.
We put elements with 3D identity transforms onto compositing layers because
we anticipate that they will somehow animate in the future.
This can get extremely expensive, especially on low-memory devices.
This patch makes WebKit stop handing out compositing layers for this kinda thing:
-webkit-transform: translate3d(0,0,0)
..on devices with 512MB or less. This dramatically improves stability on some
web pages.
- rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::requiresCompositingForTransform):
- 6:23 PM Changeset in webkit [185283] by
-
- 2 edits in trunk/LayoutTests
Gardening: skip some inspector-protocol tests while the issue is being investigated.
https://bugs.webkit.org/show_bug.cgi?id=145720
Not reviewed.
- platform/win/TestExpectations:
- 6:23 PM Changeset in webkit [185282] by
-
- 3 edits in trunk/Source/WebInspectorUI
Web Inspector: Fixing code style and adding more limitations for bug 141262
https://bugs.webkit.org/show_bug.cgi?id=145668
Patch by Devin Rousso <Devin Rousso> on 2015-06-05
Reviewed by Timothy Hatcher.
- UserInterface/Models/CSSCompletions.js:
(WebInspector.CSSCompletions.prototype.isValidPropertyName): Loops through the full property list and returns true only if a property exactly matches the given property name.
(WebInspector.CSSCompletions): Added isValidPropertyName function.
- UserInterface/Views/CSSStyleDeclarationTextEditor.js:
(WebInspector.CSSStyleDeclarationTextEditor.prototype._createTextMarkerForPropertyIfNeeded): Added logic to limit the invalid class marker to only the property value if the property name is an actual property and to prevent invalid style from being applied incorrectly.
- 6:19 PM Changeset in webkit [185281] by
-
- 2 edits in trunk/Tools
[EFL] API test has been broken since r184357
https://bugs.webkit.org/show_bug.cgi?id=145635
Reviewed by Darin Adler.
r184357 suppress cmake build warning using TARGET_FILE property. However
many EFL API test cases have been broken after applying it.
- TestWebKitAPI/CMakeLists.txt:
- 6:18 PM Changeset in webkit [185280] by
-
- 2 edits in trunk/LayoutTests
[Win] Skip fast/text/decorations-vertical-underline-expected.html
Unreviewed gardening.
- platform/win/TestExpectations:
- 6:16 PM Changeset in webkit [185279] by
-
- 4 edits in trunk/Source/WebInspectorUI
Web Inspector: Highlighting a CSS section does not deselect previously highlighted CSS sections
https://bugs.webkit.org/show_bug.cgi?id=145399
Patch by Devin Rousso <Devin Rousso> on 2015-06-05
Reviewed by Timothy Hatcher.
- UserInterface/Views/CSSStyleDeclarationSection.js:
(WebInspector.CSSStyleDeclarationSection): Added delegate variable to constructor.
(WebInspector.CSSStyleDeclarationSection.prototype.clearSelection):
(WebInspector.CSSStyleDeclarationSection.prototype.cssStyleDeclarationTextEditorFocused):
- UserInterface/Views/CSSStyleDeclarationTextEditor.js:
(WebInspector.CSSStyleDeclarationTextEditor): Added event listener to codeMirror on "focus".
(WebInspector.CSSStyleDeclarationTextEditor.prototype.clearSelection): Sets the cursor of the codeMirror to character 0 on line 0.
(WebInspector.CSSStyleDeclarationTextEditor.prototype._editorFocused): Function that calls the delegate function "editorFocused" when the codeMirror is focused
- UserInterface/Views/RulesStyleDetailsPanel.js:
(WebInspector.RulesStyleDetailsPanel.prototype.refresh.appendStyleSection):
(WebInspector.RulesStyleDetailsPanel.prototype.cssStyleDeclarationSectionEditorFocused): Loops through all sections except for the ignoredSection (parameter) and clears the selected text in each of them.
- 6:06 PM Changeset in webkit [185278] by
-
- 2 edits in trunk/Source/WebCore
Web Inspector: "Other" time in the framerate table is often negative
https://bugs.webkit.org/show_bug.cgi?id=145712
Reviewed by Timothy Hatcher.
The Inspector frontend flattens the timeline event tree it receives from the backend, which can contain nested
paint records. The nested records represent a single paint event, but were being interpreted as two separate
events for purposes of calculating total layout time. This caused the calculated "other" time to be less than
it should be (and in some cases negative).
- inspector/InspectorTimelineAgent.cpp:
(WebCore::InspectorTimelineAgent::addRecordToTimeline):
Paint records are no longer nested, we simply drop the child paint event.
- 5:33 PM Changeset in webkit [185277] by
-
- 4 edits1 add in trunk/Source/JavaScriptCore
Subclasses of JSNonFinalObject with gc'able children need to implement visitChildren().
https://bugs.webkit.org/show_bug.cgi?id=145709
Reviewed by Geoffrey Garen.
- jsc.cpp:
(functionSetElementRoot):
- The Element class has a member of type Root which extends JSDestructibleObject. It should be stored in a WriteBarrier, and visited by visitChildren().
- runtime/ClonedArguments.cpp:
(JSC::ClonedArguments::materializeSpecialsIfNecessary):
(JSC::ClonedArguments::visitChildren):
- runtime/ClonedArguments.h:
- Add missing visitChildren().
- tests/stress/cloned-arguments-should-visit-callee-during-gc.js: Added.
(makeTransientFunction.transientFunc):
(makeTransientFunction):
- 5:04 PM Changeset in webkit [185276] by
-
- 1 edit2 adds in trunk/LayoutTests
[Win] Update expected results.
Unreviewed gardening.
- platform/win/fast/forms/select-element-focus-ring-expected.txt: Added.
- platform/win/fast/text/emoji-expected.txt: Added.
- 4:50 PM Changeset in webkit [185275] by
-
- 2 edits in trunk/Source/WebCore
Unreviewed, speculative Windows build fix after r185273.
- platform/VNodeTracker.h:
- 4:49 PM Changeset in webkit [185274] by
-
- 6 edits in trunk/Source/WebKit2
Only allow plug-ins to create custom MIME property lists on older versions of OS X
https://bugs.webkit.org/show_bug.cgi?id=145716
Reviewed by Dan Bernstein.
- PluginProcess/EntryPoint/mac/LegacyProcess/PluginProcessMain.mm:
- Shared/Plugins/Netscape/NetscapePluginModule.h:
- Shared/Plugins/Netscape/mac/NetscapePluginModuleMac.mm:
(WebKit::getMIMETypesFromPluginBundle):
- UIProcess/Plugins/PluginProcessProxy.h:
- UIProcess/Plugins/mac/PluginProcessProxyMac.mm:
- 4:20 PM Changeset in webkit [185273] by
-
- 10 edits3 adds in trunk/Source
[WK2][iOS] Limit the number of vnodes used by the WebContent processes
https://bugs.webkit.org/show_bug.cgi?id=145672
<rdar://problem/21126637>
Reviewed by Antti Koivisto.
Source/WebCore:
Limit the number of vnodes used by the WebContent processes to reduce
the chance of getting killed due to the system running out of vnodes.
We sometimes see the WebContent process use up to 50% of the system's
vnode limit on some tests on iOS, which seems excessive. Most vnodes
are due to CachedResources which are mmap'd from the WebKit disk cache
and kept alive due to caches such as the Memory Cache / PageCache.
This patch adds tracking for the number of SharedBuffer objects that are
backed by a CFDataRef as this should track most of the vnodes used by
the process. The WebContent process registers a vnode pressure handler
upon initialization (likely it already does for memory pressure
handling). This handler gets called when the process uses over 15% of
of system's vnode limit and the "critical" flag is set if it uses over
20% of the system limit. The vnode pressure handler currently calls the
memory pressure handler as freeing our caches (especially PageCache,
MemoryCache) and doing a JS GC frees up vnodes.
On iPhone 6 Plus, the system limit is at 3000, which would lead to the
following limits being used:
soft: 450 / hard: 600
However, on older hardware, the system limit can be as low as 2000,
which would lead to the following limits being used:
soft: 300 / hard: 400
vnode tracking is currently only enabled on iOS because system vnode
limits are usually much higher on Mac (e.g. 473695 on Mac Pro, 9984
on a 2011 MacBook Air) and we normally don't get killed.
- WebCore.xcodeproj/project.pbxproj:
Add new files to the XCode project.
- platform/SharedBuffer.h:
Add a VNodeTracker::Token member next to the CFDataRef member that we
initialize if the SharedBuffer is constructed from a CFDataRef.
- platform/VNodeTracker.cpp: Added.
(WebCore::VNodeTracker::singleton):
(WebCore::VNodeTracker::VNodeTracker):
(WebCore::VNodeTracker::checkPressureState):
(WebCore::VNodeTracker::pressureWarningTimerFired):
(WebCore::VNodeTracker::nextPressureWarningInterval):
(WebCore::VNodeTracker::platformInitialize):
- platform/VNodeTracker.h: Added.
(WebCore::VNodeTracker::setPressureHandler):
Add a new VNodeTracker singleton to keep track the number of vnodes
used by the process since the system does not issue warnings before
reaching its vnode limit. Call sites should request Tokens for
every vnode they use for tracking purposes. Each process can then
set a vnode pressure handler function that will be called when the
process is over the designated limits.
- platform/cf/SharedBufferCF.cpp:
(WebCore::SharedBuffer::SharedBuffer):
Take a VNodeToken from the VNodeTracker when constructing the
SharedBuffer from a CFDataRef as these usually use mmap.
- platform/cocoa/VNodeTrackerCocoa.cpp: Added.
(WebCore::VNodeTracker::platformInitialize):
Get the system's vnode limit and uses 15% of that value as soft
limit for the process and 20% of that value as hard limit.
Source/WebKit2:
Have the WebContent process register a vnode pressure handler on iOS,
which calls the memory pressure handler.
On non-critical pressure, it will prune dead resources from the memory
cache, which should free up some vnodes. On critical pressure, the
handler will clear the PageCache and do a JS GC, which should free even
more vnodes.
- WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::platformInitializeWebProcess):
Source/WTF:
- wtf/RefCounter.h:
(WTF::RefCounter::value):
Expose the actual refcount instead of a boolean value.
- 4:16 PM Changeset in webkit [185272] by
-
- 3 edits in trunk/Source/WebCore
Lingering checkbox animations when rapidly switching between views on WK1
https://bugs.webkit.org/show_bug.cgi?id=145715
<rdar://problem/17381121>
Reviewed by Simon Fraser.
On WK1, the NSView we use for NSButtonCells is parented in the ScrollView.
Our checkbox and radiobutton animations cause AppKit to create a CALayer which
is parented into that view. As we render the animations, we see the CALayer getting
drawn also.
The fix was to use an unparented layer on WK1 when drawing these controls, the same
way WK2 does. This technique may apply to all controls, but restricting it
to just the buttons at the moment is less risky.
This only happens during the swap between a checked and non-checked state,
and we have no way to test it :(
- platform/mac/ThemeMac.h: New optional parameter to ensuredView.
- platform/mac/ThemeMac.mm:
(WebCore::paintToggleButton): If we're on Yosemite or highter, use an unparented view.
(WebCore::ThemeMac::ensuredView): Return either a parented or unparented view.
- 3:59 PM Changeset in webkit [185271] by
-
- 4 edits in trunk/Source/WebInspectorUI
REGRESSION (r185213): TypeError: undefined is not an object (evaluating 'WebInspector.clearLogOnReload.value')
https://bugs.webkit.org/show_bug.cgi?id=145695
Patch by Joseph Pecoraro <Joseph Pecoraro> on 2015-06-05
Reviewed by Mark Lam.
Move and rename the clear log on reload setting property from
WebInspector (Main.js) to LogManager (shared with tests).
- UserInterface/Base/Main.js:
(WebInspector.loaded): Deleted.
- UserInterface/Controllers/LogManager.js:
(WebInspector.LogManager):
(WebInspector.LogManager.prototype._delayedMessagesCleared):
- UserInterface/Views/LogContentView.js:
(WebInspector.LogContentView.prototype._sessionStarted):
(WebInspector.LogContentView.prototype._handleContextMenuEvent):
(WebInspector.LogContentView.prototype._toggleClearLogOnReloadSetting):
- 3:33 PM Changeset in webkit [185270] by
-
- 4 edits in trunk/Source/WebCore
MediaSessions should keep track of their participating media elements.
https://bugs.webkit.org/show_bug.cgi?id=145703
Patch by Matt Rajca <mrajca@apple.com> on 2015-06-05
Reviewed by Eric Carlson.
- Modules/mediasession/MediaSession.cpp:
(WebCore::MediaSession::addMediaElement): Add the given media element to the media session.
(WebCore::MediaSession::removeMediaElement): Remove the given media element from the media session.
(WebCore::MediaSession::activeParticipatingElements): Dynamically build a vector of actively-playing media
elements.
- Modules/mediasession/MediaSession.h:
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::~HTMLMediaElement): Ensure the element is removed from its media session.
(WebCore::HTMLMediaElement::setSession): Update sessions' references to their media elements.
- 3:10 PM Changeset in webkit [185269] by
-
- 13 edits15 deletes in trunk
Regression(r176212): Broke app switching on iCloud.com
https://bugs.webkit.org/show_bug.cgi?id=145708
<rdar://problem/21235277>
Reviewed by Simon Fraser.
Source/WebCore:
Roll out r176212 and follow-up fixes for now, to fix iCloud.com.
We can reconsider later how to do this in a safer way.
- bindings/js/JSCSSStyleDeclarationCustom.cpp:
(WebCore::JSCSSStyleDeclaration::putDelegate): Deleted.
(WebCore::JSCSSStyleDeclaration::getOwnPropertyNames): Deleted.
- dom/Element.cpp:
- dom/Element.h:
- dom/ElementRareData.cpp:
- dom/ElementRareData.h:
(WebCore::ElementRareData::ElementRareData):
(WebCore::ElementRareData::~ElementRareData): Deleted.
- dom/Node.cpp:
(WebCore::Node::materializeRareData):
- html/HTMLCanvasElement.cpp:
(WebCore::HTMLCanvasElement::notifyObserversCanvasChanged): Deleted.
- page/DOMTimer.cpp:
(WebCore::DOMTimerFireState::scriptMadeNonUserObservableChanges): Deleted.
(WebCore::DOMTimerFireState::scriptMadeUserObservableChanges): Deleted.
(WebCore::NestedTimersMap::instanceForContext): Deleted.
(WebCore::DOMTimer::install): Deleted.
(WebCore::DOMTimer::fired): Deleted.
(WebCore::DOMTimer::alignedFireTime): Deleted.
(WebCore::DOMTimer::activeDOMObjectName): Deleted.
- page/DOMTimer.h:
- page/FrameView.cpp:
(WebCore::FrameView::reset): Deleted.
(WebCore::FrameView::viewportContentsChanged): Deleted.
(WebCore::FrameView::autoSizeIfEnabled): Deleted.
- page/FrameView.h:
LayoutTests:
Remove layout tests covering DOM Timer throttling.
- fast/canvas/canvas-outside-viewport-timer-throttling-expected.txt: Removed.
- fast/canvas/canvas-outside-viewport-timer-throttling.html: Removed.
- fast/dom/nested-timer-display-none-element-throttling-expected.txt: Removed.
- fast/dom/nested-timer-display-none-element-throttling.html: Removed.
- fast/dom/repeating-timer-display-none-element-throttling-expected.txt: Removed.
- fast/dom/repeating-timer-display-none-element-throttling.html: Removed.
- fast/dom/repeating-timer-element-overflow-hidden-throttling-expected.txt: Removed.
- fast/dom/repeating-timer-element-overflow-hidden-throttling.html: Removed.
- fast/dom/resources/timer-throttling-iframe.html: Removed.
- fast/dom/timer-throttle-on-scrolling-iframe-away-expected.txt: Removed.
- fast/dom/timer-throttle-on-scrolling-iframe-away.html: Removed.
- fast/dom/timer-unthrottle-on-layout-expected.txt: Removed.
- fast/dom/timer-unthrottle-on-layout.html: Removed.
- fast/dom/timer-unthrottle-on-scroll-expected.txt: Removed.
- fast/dom/timer-unthrottle-on-scroll.html: Removed.
- 3:03 PM Changeset in webkit [185268] by
-
- 2 edits in trunk/Source/JavaScriptCore
DropAllLocks RELEASE_ASSERT on iOS
https://bugs.webkit.org/show_bug.cgi?id=139654
Reviewed by Mark Lam.
- runtime/JSLock.cpp:
(JSC::JSLock::dropAllLocks): Removed a comment because it duplicated
the code beneath it. Removed a FIXME because we can't ASSERT that
we're holding the lock. WebKit1 on iOS drops the lock before calling to
delegates, not knowing whether it holds the lock or not.
(JSC::JSLock::DropAllLocks::DropAllLocks): Only ASSERT that we are not
GC'ing if we hold the lock. If we do not hold the lock, it is perfectly
valid for some other thread, which does hold the lock, to be GC'ing.
What is not valid is to drop the lock in the middle of GC, since GC
must be atomic.
- 3:00 PM Changeset in webkit [185267] by
-
- 2 edits in trunk/Source/JavaScriptCore
speculateRealNumber() should early exit if you're already a real number, not if you're already a real double.
Rubber stamped by Mark Lam.
This was causing: https://build.webkit.org/results/Apple%20Yosemite%20Debug%20WK1%20(Tests)/r185261%20(5180)/webaudio/note-grain-on-timing-crash-log.txt
- dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::speculateRealNumber):
- 2:46 PM Changeset in webkit [185266] by
-
- 6 edits1 copy1 move1 add in trunk
[iOS] Emoji overlap preceeding lines
https://bugs.webkit.org/show_bug.cgi?id=145678
<rdar://problem/10684914>
Reviewed by Enrica Casucci.
Source/WebCore:
In WebKit, we first lay out lines, and then we lay out text inside those lines. Text is vertically centered in
its containing line. If the author has not specified a particular line-height, the height of a line comes from
the font metrics for the primary font of the line.
In iOS, we were specifically hardcoding the line height metric of Apple Color Emoji to be 0. This means that,
if Apple Color Emoji is the primary font for a line, and an explicit line-height was not specified, the lines
are laid out with 0 height. Then, when we vertically center the text inside the line, the top half of the text
overlaps the preceeding line.
I'm not sure exactly why were were hardcoding this value to 0, as it is surely wrong. I'm going to ask Enrica
to review this patch; hopefully she knows the answer.
Test: fast/text/emoji.html
- platform/graphics/cocoa/FontCocoa.mm:
(WebCore::Font::platformInit):
LayoutTests:
- fast/text/emoji.html:
- platform/mac/TestExpectations: Removed unnecessary line.
- platform/mac-mavericks/fast/text/emoji-expected.html:
- platform/mac/fast/text/emoji-expected.html:
- platform/ios-simulator/fast/text/emoji-expected.html:
- TestExpectations: Removed unnecessary line.
- 2:42 PM Changeset in webkit [185265] by
-
- 2 edits in trunk/Source/WebKit2
Fix refacto, don't put IndexedDB databases in the cache directory
https://bugs.webkit.org/show_bug.cgi?id=145714
Reviewed by Antti Koivisto.
- UIProcess/API/Cocoa/APIWebsiteDataStoreCocoa.mm:
(API::WebsiteDataStore::defaultIndexedDBDatabaseDirectory):
- 2:37 PM Changeset in webkit [185264] by
-
- 5 edits in trunk/Source
[Web Timing] Use new SPI to enable data collection.
https://bugs.webkit.org/show_bug.cgi?id=145650
rdar://problem/21203358
Reviewed by Chris Dumez.
Source/WebCore:
- platform/network/cf/ResourceHandleCFNet.cpp:
(WebCore::ResourceHandle::createCFURLConnection):
Gather web timing data on iOS.
- platform/network/mac/ResourceHandleMac.mm:
(WebCore::ResourceHandle::createNSURLConnection):
(WebCore::ResourceHandle::setCollectsTimingData):
Gather web timing data on Mac.
Source/WTF:
- wtf/Platform.h:
Added HAVE_TIMINGDATAOPTIONS for platform and version dependent feature.
- 2:36 PM Changeset in webkit [185263] by
-
- 2 edits in trunk/Source/WebInspectorUI
Web Inspector: [REGRESSION] Dashboard's resource count does not reset when main resource navigates
https://bugs.webkit.org/show_bug.cgi?id=144553
Add missing _mainResourceDidChange, which was removed after switching to the tabs UI.
Reviewed by Timothy Hatcher.
- UserInterface/Models/DefaultDashboard.js:
(WebInspector.DefaultDashboard):
_waitingForFirstMainResourceToStartTrackingSize is not used anywhere, remove it.
(WebInspector.DefaultDashboard.prototype._mainResourceDidChange): Added.
(WebInspector.DefaultDashboard.prototype._resourceSizeDidChange): Removed.
Unused.
- 2:31 PM Changeset in webkit [185262] by
-
- 13 edits in trunk
Make the network cache directory part of WKContextConfigurationRef
https://bugs.webkit.org/show_bug.cgi?id=145711
Reviewed by Antti Koivisto.
Source/WebKit2:
Also, make the UI process be in control of where we place the new network cache.
When using the Modern WebKit API, put the cache in ~/Library/Caches/<Bundle ID>/WebKit/NetworkCache,
and put the cache in ~/Library/Caches/<Bundle ID>/WebKitCache otherwise (this is just for Safari).
- NetworkProcess/cache/NetworkCacheStorage.cpp:
(WebKit::NetworkCache::Storage::open):
- UIProcess/API/APIProcessPoolConfiguration.cpp:
(API::ProcessPoolConfiguration::createWithLegacyOptions):
(API::ProcessPoolConfiguration::ProcessPoolConfiguration):
(API::ProcessPoolConfiguration::copy):
- UIProcess/API/APIProcessPoolConfiguration.h:
- UIProcess/API/C/WKContext.cpp:
(WKContextSetDiskCacheDirectory): Deleted.
- UIProcess/API/C/WKContextConfigurationRef.cpp:
(WKContextConfigurationCopyDiskCacheDirectory):
(WKContextConfigurationSetDiskCacheDirectory):
- UIProcess/API/C/WKContextConfigurationRef.h:
- UIProcess/API/C/WKContextPrivate.h:
- UIProcess/Cocoa/WebProcessPoolCocoa.mm:
(WebKit::WebProcessPool::platformInitializeNetworkProcess):
(WebKit::WebProcessPool::legacyPlatformDefaultNetworkCacheDirectory):
(WebKit::WebProcessPool::isNetworkCacheEnabled):
(WebKit::WebProcessPool::platformDefaultDiskCacheDirectory): Deleted.
- UIProcess/WebProcessPool.cpp:
(WebKit::legacyWebsiteDataStoreConfiguration):
(WebKit::WebProcessPool::WebProcessPool):
(WebKit::WebProcessPool::ensureNetworkProcess):
(WebKit::WebProcessPool::diskCacheDirectory): Deleted.
- UIProcess/WebProcessPool.h:
Tools:
- WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
- 12:27 PM Changeset in webkit [185261] by
-
- 11 edits in trunk
Make the application cache directory part of WKContextConfigurationRef
https://bugs.webkit.org/show_bug.cgi?id=145704
Reviewed by Geoffrey Garen.
Source/WebKit2:
- UIProcess/API/APIProcessPoolConfiguration.cpp:
(API::ProcessPoolConfiguration::createWithLegacyOptions):
(API::ProcessPoolConfiguration::ProcessPoolConfiguration):
(API::ProcessPoolConfiguration::copy):
- UIProcess/API/APIProcessPoolConfiguration.h:
- UIProcess/API/C/WKContext.cpp:
(WKContextSetApplicationCacheDirectory): Deleted.
- UIProcess/API/C/WKContextConfigurationRef.cpp:
(WKContextConfigurationCopyApplicationCacheDirectory):
(WKContextConfigurationSetApplicationCacheDirectory):
- UIProcess/API/C/WKContextConfigurationRef.h:
- UIProcess/API/C/WKContextPrivate.h:
- UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::WebProcessPool):
(WebKit::WebProcessPool::createNewWebProcess):
(WebKit::WebProcessPool::applicationCacheDirectory): Deleted.
- UIProcess/WebProcessPool.h:
Tools:
- WebKitTestRunner/TestController.cpp:
(WTR::TestController::initialize):
- 12:08 PM Changeset in webkit [185260] by
-
- 5 edits in trunk/Source
[Streams API] ReadableStream should store callbacks as a Deque
https://bugs.webkit.org/show_bug.cgi?id=145641
Reviewed by Darin Adler.
Source/WebCore:
Covered by existing tests.
- Modules/streams/ReadableStream.cpp:
(WebCore::ReadableStream::resolveReadCallback): Use of Deque.takeFirst
- Modules/streams/ReadableStream.h: Using Deque in lieu of Vector for storing callbacks.
Source/WTF:
- wtf/Deque.h:
(WTF::Deque::append): Adding not templated append method, forwarding to templated append method.
- 11:52 AM Changeset in webkit [185259] by
-
- 101 edits44 adds in trunk
finally blocks should not set the exception stack trace when re-throwing the exception.
https://bugs.webkit.org/show_bug.cgi?id=145525
Reviewed by Geoffrey Garen.
Source/JavaScriptCore:
How exceptions presently work:
=============================
- op_throw can throw any JSValue.
- the VM tries to capture the stack at the throw point and propagate that as needed.
- finally blocks are implemented using op_catch to catch the thrown value, and throws it again using op_throw.
What's wrong with how it presently works:
========================================
- finally's makes for bad exception throw line numbers in the Inspector console.
The op_throw in finally will throw the value anew i.e. it captures a stack from the re-throw point.
As a result, the Inspector sees the finally block as the throw point. The original stack is lost.
- finally's breaks the Inspector's "Breaks on Uncaught Exception"
This is because finally blocks are indistinguishable from catch blocks. As a result, a try-finally,
which should break in the Inspector on the throw, does not because the Inspector thought the
exception was "caught".
- finally's yields confusing break points when the Inspector "Breaks on All Exceptions"
- In a try-finally scenario, the Inspector breaks 2 times: 1 at the throw, 1 at the finally.
- In a for-of loop (which has synthesized finallys), the Inspector will do another break. Similarly for other cases of JS code which synthesize finallys.
- At VM re-entry boundaries (e.g. js throws & returns to native code, which returns to js), the Inspector will do another break if there's an uncaught exception.
How this patch fixes the issues:
===============================
- We introduce an Exception object that wraps the thrown value and the exception stack.
When throwing an exception, the VM will check if the thrown value is an Exception
object or not. If it is not an Exception object, then we must be throwing a new
exception. The VM will create an Exception object to wrap the thrown value and
capture the current stack for it.
If the thrown value is already an Exception object, then the requested throw operation
must be a re-throw. The VM will not capture a new stack for it.
- op_catch will now populate 2 locals: 1 for the Exception, 1 for the thrown JSValue.
The VM is aware of the Exception object and uses it for rethrows in finally blocks.
JS source code is never aware of the Exception object.
JS code is aware of the thrown value. If it throws the caught thrown value, that
constitutes a new throw, and a new Exception object will be created for it.
- The VM no longer tracks the thrown JSValue and the exception stack. It will only track a m_exception field which is an Exception*.
- The BytecodeGenerator has already been updated in a prior patch to distinguish between Catch, Finally, and SynthesizedFinally blocks. The interpreter runtime will now report to the debugger whether we have a Catch handler, not just any handlers.
The debugger will use this detail to determine whether to break or not. "Break on
uncaught exceptions" will only break if no Catch handler was found.
This solves the issue of the debugger breaking at finally blocks, and for-of statements.
- The Exception object will also have a flag to indicate whether the debugger has been notified of the Exception being thrown. Once the Interpreter notifies the debugger of the Exception object, it will mark this flag and not repeat the notify the debugger again of the same Exception.
This solves the issue of the debugger breaking at VM re-entry points due to uncaught
exceptions.
- The life-cycle of the captured exception stack trace will now follow the life-cycle of the Exception object.
Other changes:
- Change all clients of the VM::exception() to expect an Exception* instead of JSValue.
- Fixed a few bugs where thrown exceptions are not cleared before exiting the VM.
- Also renamed some variables and classes to better describe what they are.
- API/JSBase.cpp:
(JSEvaluateScript):
(JSCheckScriptSyntax):
- API/JSObjectRef.cpp:
(handleExceptionIfNeeded):
- The functions below all do the same exception check. Added this helper to simplify the code.
(JSClassCreate):
(JSObjectMakeFunction):
(JSObjectMakeArray):
(JSObjectMakeDate):
(JSObjectMakeError):
(JSObjectMakeRegExp):
(JSObjectGetProperty):
(JSObjectSetProperty):
(JSObjectGetPropertyAtIndex):
(JSObjectSetPropertyAtIndex):
(JSObjectDeleteProperty):
(JSObjectCallAsFunction):
(JSObjectCallAsConstructor):
- API/JSScriptRef.cpp:
- API/JSValue.mm:
(JSContainerConvertor::take):
(reportExceptionToInspector):
- API/JSValueRef.cpp:
(handleExceptionIfNeeded):
- The functions below all do the same exception check. Added this helper to simplify the code.
(evernoteHackNeeded):
(JSValueIsEqual):
(JSValueIsInstanceOfConstructor):
(JSValueCreateJSONString):
(JSValueToNumber):
(JSValueToStringCopy):
(JSValueToObject):
- CMakeLists.txt:
- JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
- JavaScriptCore.vcxproj/JavaScriptCore.vcxproj.filters:
- JavaScriptCore.xcodeproj/project.pbxproj:
- Added new files Exception.h and Exception.cpp.
- bindings/ScriptFunctionCall.cpp:
(Deprecated::ScriptFunctionCall::call):
- bindings/ScriptFunctionCall.h:
- bytecode/BytecodeList.json:
- op_catch now had 2 operands: the exception register, and the thrown value register.
- bytecode/BytecodeUseDef.h:
(JSC::computeDefsForBytecodeOffset):
- bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpBytecode):
(JSC::CodeBlock::handlerForBytecodeOffset):
- bytecode/CodeBlock.h:
- handlerForBytecodeOffset() now can look for just Catch handlers only.
- bytecode/HandlerInfo.h:
- Cleaned up some white space I accidentally added in a previous patch.
- bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::pushTry):
(JSC::BytecodeGenerator::popTryAndEmitCatch):
(JSC::BytecodeGenerator::emitThrowReferenceError):
(JSC::BytecodeGenerator::emitEnumeration):
- bytecompiler/BytecodeGenerator.h:
(JSC::BytecodeGenerator::emitThrow):
- bytecompiler/NodesCodegen.cpp:
(JSC::TryNode::emitBytecode):
- Adding support for op_catch's 2 operands.
- debugger/Debugger.cpp:
(JSC::Debugger::hasBreakpoint):
(JSC::Debugger::pauseIfNeeded):
(JSC::Debugger::exception):
- debugger/Debugger.h:
- debugger/DebuggerCallFrame.cpp:
(JSC::DebuggerCallFrame::thisValue):
(JSC::DebuggerCallFrame::evaluate):
- debugger/DebuggerCallFrame.h:
(JSC::DebuggerCallFrame::isValid):
- inspector/InjectedScriptManager.cpp:
(Inspector::InjectedScriptManager::createInjectedScript):
- inspector/InspectorEnvironment.h:
- inspector/JSGlobalObjectInspectorController.cpp:
(Inspector::JSGlobalObjectInspectorController::appendAPIBacktrace):
(Inspector::JSGlobalObjectInspectorController::reportAPIException):
- inspector/JSGlobalObjectInspectorController.h:
- inspector/JSGlobalObjectScriptDebugServer.h:
- inspector/JSJavaScriptCallFrame.cpp:
(Inspector::JSJavaScriptCallFrame::evaluate):
- inspector/JavaScriptCallFrame.h:
(Inspector::JavaScriptCallFrame::vmEntryGlobalObject):
(Inspector::JavaScriptCallFrame::thisValue):
(Inspector::JavaScriptCallFrame::evaluate):
- inspector/ScriptCallStackFactory.cpp:
(Inspector::extractSourceInformationFromException):
(Inspector::createScriptCallStackFromException):
- inspector/ScriptCallStackFactory.h:
- inspector/ScriptDebugServer.cpp:
(Inspector::ScriptDebugServer::evaluateBreakpointAction):
(Inspector::ScriptDebugServer::handleBreakpointHit):
(Inspector::ScriptDebugServer::handleExceptionInBreakpointCondition):
- inspector/ScriptDebugServer.h:
- interpreter/CallFrame.h:
(JSC::ExecState::clearException):
(JSC::ExecState::exception):
(JSC::ExecState::hadException):
(JSC::ExecState::atomicStringTable):
(JSC::ExecState::propertyNames):
(JSC::ExecState::clearSupplementaryExceptionInfo): Deleted.
- interpreter/Interpreter.cpp:
(JSC::unwindCallFrame):
(JSC::Interpreter::stackTraceAsString):
(JSC::GetCatchHandlerFunctor::GetCatchHandlerFunctor):
(JSC::GetCatchHandlerFunctor::operator()):
(JSC::Interpreter::unwind):
- Added a check for didNotifyInspectorOfThrow() here to prevent duplicate reports of the same Exception to the debugger.
(JSC::GetExceptionHandlerFunctor::GetExceptionHandlerFunctor): Deleted.
(JSC::GetExceptionHandlerFunctor::operator()): Deleted.
- Renamed GetExceptionHandlerFunctor to GetCatchHandlerFunctor since the debugger is only interested in knowing whether we have Catch handlers.
- interpreter/Interpreter.h:
(JSC::SuspendExceptionScope::SuspendExceptionScope):
(JSC::SuspendExceptionScope::~SuspendExceptionScope):
(JSC::Interpreter::sampler):
(JSC::ClearExceptionScope::ClearExceptionScope): Deleted.
(JSC::ClearExceptionScope::~ClearExceptionScope): Deleted.
- Renamed ClearExceptionScope to SuspendExceptionScope because "clear" implies that we're purging the exception. Instead, we're merely suspending any handling of that exception for a period defined by the scope.
- jit/AssemblyHelpers.cpp:
(JSC::AssemblyHelpers::emitExceptionCheck):
- jit/JITExceptions.cpp:
(JSC::genericUnwind):
- Removed the exception argument. It is always the value in VM::exception() anyway. genericUnwind() can just get it from the VM, and save everyone some work.
- jit/JITExceptions.h:
- jit/JITOpcodes.cpp:
(JSC::JIT::emit_op_catch):
- jit/JITOpcodes32_64.cpp:
(JSC::JIT::privateCompileCTINativeCall):
(JSC::JIT::emit_op_catch):
- Add support for the new op_catch operands.
- jit/JITOperations.cpp:
- jit/ThunkGenerators.cpp:
(JSC::nativeForGenerator):
- jsc.cpp:
(functionRun):
(functionLoad):
(runWithScripts):
(runInteractive):
- llint/LLIntOffsetsExtractor.cpp:
- llint/LLIntSlowPaths.cpp:
(JSC::LLInt::LLINT_SLOW_PATH_DECL):
- llint/LowLevelInterpreter32_64.asm:
- llint/LowLevelInterpreter64.asm:
- Add support for the new op_catch operands. Also update the code to handle VM::m_exception being an Exception pointer, not a JSValue.
- parser/NodeConstructors.h:
(JSC::TryNode::TryNode):
- parser/Nodes.h:
- runtime/CallData.cpp:
(JSC::call):
- runtime/CallData.h:
- runtime/Completion.cpp:
(JSC::evaluate):
- runtime/Completion.h:
(JSC::evaluate):
- Change evaluate() to take a reference to the returned exception value instead of a pointer. In all but 2 or 3 cases, we want the returned exception anyway. Might as well simplify the code by requiring the reference.
- runtime/Error.h:
(JSC::throwVMError):
(JSC::throwVMTypeError):
- runtime/Exception.cpp: Added.
(JSC::Exception::create):
(JSC::Exception::destroy):
(JSC::Exception::createStructure):
(JSC::Exception::visitChildren):
(JSC::Exception::Exception):
(JSC::Exception::~Exception):
- runtime/Exception.h: Added.
(JSC::Exception::valueOffset):
(JSC::Exception::cast):
(JSC::Exception::value):
(JSC::Exception::stack):
(JSC::Exception::didNotifyInspectorOfThrow):
(JSC::Exception::setDidNotifyInspectorOfThrow):
- runtime/ExceptionHelpers.cpp:
(JSC::createTerminatedExecutionException):
(JSC::isTerminatedExecutionException):
(JSC::createStackOverflowError):
- runtime/ExceptionHelpers.h:
- runtime/GetterSetter.cpp:
(JSC::callGetter):
- runtime/IteratorOperations.cpp:
(JSC::iteratorClose):
- runtime/JSObject.cpp:
- runtime/JSPromiseConstructor.cpp:
(JSC::constructPromise):
- runtime/JSPromiseDeferred.cpp:
(JSC::updateDeferredFromPotentialThenable):
(JSC::abruptRejection):
- runtime/JSPromiseReaction.cpp:
(JSC::ExecutePromiseReactionMicrotask::run):
- runtime/VM.cpp:
(JSC::VM::VM):
(JSC::VM::releaseExecutableMemory):
(JSC::VM::throwException):
(JSC::VM::setStackPointerAtVMEntry):
(JSC::VM::getExceptionInfo): Deleted.
(JSC::VM::setExceptionInfo): Deleted.
(JSC::VM::clearException): Deleted.
(JSC::clearExceptionStack): Deleted.
- runtime/VM.h:
(JSC::VM::targetMachinePCForThrowOffset):
(JSC::VM::clearException):
(JSC::VM::setException):
(JSC::VM::exception):
(JSC::VM::addressOfException):
(JSC::VM::exceptionStack): Deleted.
- runtime/VMEntryScope.cpp:
(JSC::VMEntryScope::VMEntryScope):
(JSC::VMEntryScope::setEntryScopeDidPopListener):
Source/WebCore:
Update to use the new JSC::Exception object.
Test: inspector/debugger/break-on-exceptions.html
- ForwardingHeaders/runtime/Exception.h: Added.
- bindings/js/JSCallbackData.cpp:
(WebCore::JSCallbackData::invokeCallback):
- bindings/js/JSCustomXPathNSResolver.cpp:
(WebCore::JSCustomXPathNSResolver::lookupNamespaceURI):
- bindings/js/JSDOMBinding.cpp:
(WebCore::jsArray):
(WebCore::reportException):
(WebCore::reportCurrentException):
- bindings/js/JSDOMBinding.h:
- bindings/js/JSErrorHandler.cpp:
(WebCore::JSErrorHandler::handleEvent):
- bindings/js/JSEventListener.cpp:
(WebCore::JSEventListener::handleEvent):
- bindings/js/JSMainThreadExecState.cpp:
(WebCore::JSMainThreadExecState::didLeaveScriptContext):
(WebCore::functionCallHandlerFromAnyThread):
(WebCore::evaluateHandlerFromAnyThread):
- bindings/js/JSMainThreadExecState.h:
(WebCore::JSMainThreadExecState::currentState):
(WebCore::JSMainThreadExecState::call):
(WebCore::JSMainThreadExecState::evaluate):
(WebCore::JSMainThreadExecState::runTask):
- bindings/js/JSMediaDevicesCustom.cpp:
(WebCore::JSMediaDevices::getUserMedia):
- Fixed a bug where the exception was not cleared before entering the VM to call JS code.
- bindings/js/JSMutationCallback.cpp:
(WebCore::JSMutationCallback::call):
- bindings/js/ReadableJSStream.cpp:
(WebCore::getPropertyFromObject):
(WebCore::callFunction):
(WebCore::ReadableJSStream::Source::start):
- bindings/js/ScheduledAction.cpp:
(WebCore::ScheduledAction::executeFunctionInContext):
- bindings/js/ScriptController.cpp:
(WebCore::ScriptController::evaluateInWorld):
- bindings/js/SerializedScriptValue.cpp:
(WebCore::SerializedScriptValue::create):
(WebCore::SerializedScriptValue::deserialize):
- bindings/js/WorkerScriptController.cpp:
(WebCore::WorkerScriptController::evaluate):
(WebCore::WorkerScriptController::setException):
(WebCore::WorkerScriptController::scheduleExecutionTermination):
- bindings/js/WorkerScriptController.h:
(WebCore::WorkerScriptController::workerGlobalScopeWrapper):
- bindings/js/WorkerScriptDebugServer.cpp:
(WebCore::WorkerScriptDebugServer::runEventLoopWhilePaused):
(WebCore::WorkerScriptDebugServer::reportException):
- bindings/js/WorkerScriptDebugServer.h:
- bindings/objc/WebScriptObject.mm:
(WebCore::createJSWrapper):
(WebCore::addExceptionToConsole):
(-[WebScriptObject callWebScriptMethod:withArguments:]):
(-[WebScriptObject evaluateWebScript:]):
- Changed to call a version of JSMainThreadExecState::evaluate() that provides a stub returnedException because evaluateWebScript: doesn't need the exception.
- inspector/PageScriptDebugServer.cpp:
(WebCore::PageScriptDebugServer::isContentScript):
(WebCore::PageScriptDebugServer::reportException):
- inspector/PageScriptDebugServer.h:
- workers/WorkerGlobalScope.cpp:
(WebCore::WorkerGlobalScope::importScripts):
Source/WebKit/mac:
- WebView/WebView.mm:
(+[WebView _reportException:inContext:]):
(WebKitInitializeApplicationCachePathIfNecessary):
- Changed to use the new Exception object.
Source/WebKit/win:
- WebView.cpp:
(WebView::reportException):
- Changed to use the new Exception object.
Source/WebKit2:
- WebProcess/InjectedBundle/InjectedBundle.cpp:
(WebKit::InjectedBundle::reportException):
- Changed to use the new Exception object.
LayoutTests:
- TestExpectations:
- Skip the new tests until webkit.org/b/145090 is fixed.
- fast/dom/regress-131530-expected.txt:
- Rebased results because we now have a proper line number.
- http/tests/inspector/inspector-test.js:
(InspectorTestProxy.clearResults):
(InspectorTestProxy.reportUncaughtException):
- Add the feature to sanitize the url reported by reportUncaughtException() since we can have tests that do expect uncaught exceptions, and we need the test results to be invariant. Sanitization of the url, in this case means, stripping off the preceding path.
- inspector/debugger/break-on-exception-expected.txt: Added.
- inspector/debugger/break-on-exception.html: Added.
- inspector/debugger/break-on-exception-catch-expected.txt: Added.
- inspector/debugger/break-on-exception-catch.html: Added.
- inspector/debugger/break-on-exception-finally-expected.txt: Added.
- inspector/debugger/break-on-exception-finally.html: Added.
- inspector/debugger/break-on-exception-native-expected.txt: Added.
- inspector/debugger/break-on-exception-native.html: Added.
- inspector/debugger/break-on-exception-throw-in-promise-expected.txt: Added.
- inspector/debugger/break-on-exception-throw-in-promise.html: Added.
- inspector/debugger/break-on-exception-throw-in-promise-with-catch-expected.txt: Added.
- inspector/debugger/break-on-exception-throw-in-promise-with-catch.html: Added.
- inspector/debugger/break-on-exception-throw-in-promise-then-expected.txt: Added.
- inspector/debugger/break-on-exception-throw-in-promise-then.html: Added.
- inspector/debugger/break-on-exception-throw-in-promise-then-with-catch-expected.txt: Added.
- inspector/debugger/break-on-exception-throw-in-promise-then-with-catch.html: Added.
- inspector/debugger/break-on-exception-throw-in-promise-rethrow-in-catch-expected.txt: Added.
- inspector/debugger/break-on-exception-throw-in-promise-rethrow-in-catch.html: Added.
- inspector/debugger/break-on-exception-window-onerror-expected.txt: Added.
- inspector/debugger/break-on-exception-window-onerror.html: Added.
- inspector/debugger/break-on-uncaught-exception-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception.html: Added.
- inspector/debugger/break-on-uncaught-exception-catch-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception-catch.html: Added.
- inspector/debugger/break-on-uncaught-exception-finally-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception-finally.html: Added.
- inspector/debugger/break-on-uncaught-exception-native-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception-native.html: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise.html: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise-with-catch-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise-with-catch.html: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise-then-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise-then.html: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise-then-with-catch-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise-then-with-catch.html: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise-rethrow-in-catch-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception-throw-in-promise-rethrow-in-catch.html: Added.
- inspector/debugger/break-on-uncaught-exception-window-onerror-expected.txt: Added.
- inspector/debugger/break-on-uncaught-exception-window-onerror.html: Added.
- inspector/debugger/resources/break-on-exception-tests.js: Added.
(doThrow):
(testCatch):
(testFinally):
(testThrowingThruNativeCode):
(testThrowingInPromise):
(testThrowingInPromiseWithCatch):
(testThrowingInPromiseThen):
(testThrowingInPromiseThenWithCatch):
(testThrowingInPromiseWithRethrowInCatch):
- 11:39 AM Changeset in webkit [185258] by
-
- 5 edits in trunk
Layout tests fullscreen/video-controls-drag.html and media/video-fullscreeen-only-controls.html
fail on WK1 and WK2
https://bugs.webkit.org/show_bug.cgi?id=145675
Reviewed by Dean Jackson.
Source/WebCore:
No new tests, changes covered by existing tests.
- Modules/mediacontrols/mediaControlsApple.js:
(Controller.prototype.updateControls): Use shouldHaveControls() to test if controls should
be visible.
(Controller.prototype.handleWrapperMouseMove): Don't return early when in fullscreen.
(Controller.prototype.updateFullscreenButtons): Drive-by fix: always show the (exit) fullscreen
button when in fullscreen.
LayoutTests:
- TestExpectations: Unskip tests.
- fullscreen/video-controls-drag.html: Minor cleanup, force layout before testing control location.
- 11:35 AM Changeset in webkit [185257] by
-
- 6 edits in trunk/LayoutTests
[iOS] Rebaseline expected results.
Unreviewed gardening.
- platform/ios-simulator/compositing/overflow/overflow-positioning-expected.txt:
- platform/ios-simulator/css3/selectors3/xhtml/css3-modsel-15c-expected.txt:
- platform/ios-simulator/css3/selectors3/xml/css3-modsel-15c-expected.txt:
- platform/ios-simulator/ios/scrolling/overflow-div-scrolling-expected.txt:
- platform/ios-simulator/ios/touch/touch-handler-on-window-expected.txt:
- 11:26 AM Changeset in webkit [185256] by
-
- 3 edits2 adds in trunk
Underlines too close in vertical Chinese text.
https://bugs.webkit.org/show_bug.cgi?id=145651
<rdar://problem/11105920>
Reviewed by Simon Fraser.
Source/WebCore:
Added fast/text/decorations-vertical-underline.html
- style/InlineTextBoxStyle.cpp:
(WebCore::computeUnderlineOffset):
Make sure the to map text-underline-position: auto to under when a line has an ideographic baseline.
LayoutTests:
- fast/text/decorations-vertical-underline-expected.html: Added.
- fast/text/decorations-vertical-underline.html: Added.
- 11:11 AM Changeset in webkit [185255] by
-
- 4 edits in trunk/Source/WebInspectorUI
Web Inspector: Switch to doughnut chart and remove idle time in the Rendering Frames timeline
https://bugs.webkit.org/show_bug.cgi?id=145671
Reviewed by Timothy Hatcher.
- Localizations/en.lproj/localizedStrings.js:
"Idle" string no longer used.
- UserInterface/Views/TimelineRecordFrame.js:
Reduced maximum frame width by 2px.
- UserInterface/Views/TimelineSidebarPanel.js:
(WebInspector.TimelineSidebarPanel):
(WebInspector.TimelineSidebarPanel.prototype._refreshFrameSelectionChart):
Added inner radius to chart and removed "Idle" value.
- 11:09 AM Changeset in webkit [185254] by
-
- 2 edits in trunk/LayoutTests
Layout test http/tests/misc/acid3.html times out on Yosemite WK2 Debug.
https://bugs.webkit.org/show_bug.cgi?id=145702
- platform/mac-wk2/TestExpectations: Marking as flaky for now.
- 10:41 AM Changeset in webkit [185253] by
-
- 6 edits2 adds in trunk/LayoutTests
[iOS] Rebaseline expected results for svg font tests
Reviewed by Daniel Bates.
- platform/ios-simulator/svg/custom/svg-fonts-no-latin-glyph-expected.txt: Added.
- platform/ios-simulator/svg/custom/svg-fonts-segmented-expected.txt:
- platform/ios-simulator/svg/custom/svg-fonts-with-no-element-reference-expected.txt: Added.
- platform/ios-simulator/svg/text/text-overflow-ellipsis-svgfont-expected.txt:
- platform/ios-simulator/svg/text/text-overflow-ellipsis-svgfont-kerning-ligatures-expected.txt:
- svg/custom/svg-fonts-no-latin-glyph-expected.txt:
- svg/custom/svg-fonts-no-latin-glyph.html:
- 10:19 AM Changeset in webkit [185252] by
-
- 3 edits in trunk/Source/WebCore
[GTK] [Wayland] The MiniBrowser crashes inside Weston.
https://bugs.webkit.org/show_bug.cgi?id=145645
Reviewed by Žan Doberšek.
No new tests, no behavior changes.
The issue was caused because we were not properly initializing the
Wayland display, and we were returning a nullptr in PlatformDisplayWayland::create(),
therefore a crash was happening on WebKit2/UIProcess/API/gtk/WebKitWebViewBase.cpp:1136
<http://trac.webkit.org/browser/trunk/Source/WebKit2/UIProcess/API/gtk/WebKitWebViewBase.cpp?rev=185245#L1136>
- platform/graphics/wayland/PlatformDisplayWayland.cpp:
(WebCore::PlatformDisplayWayland::create): When connecting to the Wayland
server socket we should use the default value (NULL). This allows the value
to be overwritten at run-time by the environment variables WAYLAND_DISPLAY or
WAYLAND_SOCKET that the Wayland compositor should define. Also add a couple
of warnings to get information about what is going wrong when we fail to
initialize the display instead of just crashing without giving more information.
- platform/graphics/wayland/PlatformDisplayWayland.h:
(WebCore::PlatformDisplayWayland::isInitialized): Not longer check for m_webkitgtk
to signal that the display is ready because the UIProcess embedded Wayland
subcompositer is still not implemented. Added a FIXME note about it.
- 10:19 AM Changeset in webkit [185251] by
-
- 9 edits in trunk/Source/WebKit2
Disable the CFNetwork cache in the web process
https://bugs.webkit.org/show_bug.cgi?id=145700
rdar://problem/20243343
Reviewed by Geoffrey Garen.
- NetworkProcess/cache/NetworkCacheStorage.cpp:
(WebKit::NetworkCache::Storage::open):
- Shared/WebProcessCreationParameters.cpp:
(WebKit::WebProcessCreationParameters::WebProcessCreationParameters): Deleted.
(WebKit::WebProcessCreationParameters::encode): Deleted.
(WebKit::WebProcessCreationParameters::decode): Deleted.
- Shared/WebProcessCreationParameters.h:
- UIProcess/Cocoa/WebProcessPoolCocoa.mm:
(WebKit::WebProcessPool::platformInitializeWebProcess): Deleted.
- UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::createNewWebProcess): Deleted.
- WebProcess/WebProcess.cpp:
(WebKit::WebProcess::WebProcess): Deleted.
(WebKit::WebProcess::initializeWebProcess): Deleted.
- WebProcess/WebProcess.h:
- WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::platformClearResourceCaches):
(WebKit::WebProcess::platformInitializeWebProcess):
(WebKit::WebProcess::platformSetCacheModel): Deleted.
(WebKit::WebProcess::platformTerminate): Deleted.
- 10:17 AM Changeset in webkit [185250] by
-
- 2 edits in trunk/LayoutTests
[iOS] Mask layer now is part of the layer tree output.
Unreviewed gardening.
- platform/ios-simulator/compositing/masks/compositing-clip-path-change-no-repaint-expected.txt:
- 10:08 AM Changeset in webkit [185249] by
-
- 3 edits3 adds in trunk
[Mac] Popup focus is cut off
https://bugs.webkit.org/show_bug.cgi?id=145688
<rdar://problem/18313375>
Reviewed by Brent Fulgham.
Source/WebCore:
Yosemite increased the width of focus rings, causing
some ugly clipping on Yosemite. The solution was to
not set a clip (which was code introduced for Leopard).
Test: fast/forms/select-element-focus-ring.html
- rendering/RenderThemeMac.mm:
(WebCore::RenderThemeMac::paintMenuList): Don't clip on Yosemite.
LayoutTests:
- fast/forms/select-element-focus-ring.html: Added.
- platform/mac/fast/forms/select-element-focus-ring-expected.png: Added.
- platform/mac/fast/forms/select-element-focus-ring-expected.txt: Added.
- 9:50 AM Changeset in webkit [185248] by
-
- 4 edits in trunk/Source/WebKit2
Consolidate default website data directory computations
https://bugs.webkit.org/show_bug.cgi?id=145698
Reviewed by Andreas Kling.
- UIProcess/API/APIProcessPoolConfiguration.cpp:
(API::ProcessPoolConfiguration::ProcessPoolConfiguration):
- UIProcess/API/APIWebsiteDataStore.h:
- UIProcess/API/Cocoa/APIWebsiteDataStoreCocoa.mm:
(API::WebsiteDataStore::defaultApplicationCacheDirectory):
(API::WebsiteDataStore::defaultNetworkCacheDirectory):
(API::WebsiteDataStore::defaultIndexedDBDatabaseDirectory):
(API::WebsiteDataStore::defaultLocalStorageDirectory):
(API::WebsiteDataStore::defaultMediaKeysStorageDirectory):
(API::WebsiteDataStore::defaultWebSQLDatabaseDirectory):
(API::WebsiteDataStore::defaultDataStoreConfiguration):
(API::applicationCacheDirectory): Deleted.
- 9:19 AM Changeset in webkit [185247] by
-
- 1 edit5 adds in trunk/LayoutTests
Add WK1 versions of tests from Bugs 145637 and 145642
https://bugs.webkit.org/show_bug.cgi?id=145649
Reviewed by Zalan Bujtas.
- platform/mac/fast/scrolling/iframe_in_iframe-expected.txt: Added.
- platform/mac/fast/scrolling/iframe_in_iframe.html: Added.
- platform/mac/fast/scrolling/resources/inner_content.html: Added.
- platform/mac/fast/scrolling/resources/testContent.html: Added.
- platform/mac/fast/scrolling/resources/testImage.png: Added.
- 8:19 AM Changeset in webkit [185246] by
-
- 5 edits in trunk
[CSS Grid Layout] Support dots sequences in grid-template-areas
https://bugs.webkit.org/show_bug.cgi?id=145693
Reviewed by Sergio Villar Senin.
Source/WebCore:
The CSS WG agreed to consider a sequence of "." characters a single null
cell token for grid-template-areas. That way authors can match the
number of dots to the width of the column.
For more information see:
https://lists.w3.org/Archives/Public/www-style/2015May/0175.html
Added new test cases to check the new behavior.
- css/CSSParser.cpp:
(WebCore::containsOnlyDots): New helper function to check that the grid
area name is composed by only dots.
(WebCore::CSSParser::parseGridTemplateAreasRow): Use the new helper
function to look for unnamed areas.
LayoutTests:
Added new test cases to check the new behavior.
- fast/css-grid-layout/grid-template-areas-get-set-expected.txt:
- fast/css-grid-layout/grid-template-areas-get-set.html:
- 4:06 AM Changeset in webkit [185245] by
-
- 3 edits in trunk/Tools
check-webkit-style should recommend using nullptr instead of recommending using 0 for the null pointer in C++ code.
https://bugs.webkit.org/show_bug.cgi?id=145680
Reviewed by Brent Fulgham.
- Scripts/webkitpy/style/checkers/cpp.py:
(check_for_null):
- Scripts/webkitpy/style/checkers/cpp_unittest.py:
(WebKitStyleTest.test_null_false_zero):
- 2:11 AM Changeset in webkit [185244] by
-
- 3 edits in trunk/Tools
Make the web server more robust to timing issues..
Reviewed by Ryosuke Niwa.
Cleaned up exception handling.
Test server is actually serving pages before returning.
Clean up some style issues.
- Scripts/webkitpy/benchmark_runner/http_server_driver/http_server/twisted_http_server.py:
- Scripts/webkitpy/benchmark_runner/http_server_driver/simple_http_server_driver.py:
(SimpleHTTPServerDriver.init):
(SimpleHTTPServerDriver.serve): check output, wait for pages to serve
(SimpleHTTPServerDriver.baseUrl):
(SimpleHTTPServerDriver.fetchResult):
(SimpleHTTPServerDriver.killServer):
(SimpleHTTPServerDriver.getReturnCode):
- 2:11 AM Changeset in webkit [185243] by
-
- 6 edits in trunk/Tools
Add errors to run-benchmark exception handling.
Reviewed by Ryosuke Niwa.
- Scripts/run-benchmark:
(main):
- Scripts/webkitpy/benchmark_runner/benchmark_builder/generic_benchmark_builder.py:
(GenericBenchmarkBuilder.prepare):
(GenericBenchmarkBuilder._runCreateScript):
(GenericBenchmarkBuilder._applyPatch):
(GenericBenchmarkBuilder.clean):
- Scripts/webkitpy/benchmark_runner/benchmark_runner.py:
(BenchmarkRunner.init):
(BenchmarkRunner.execute):
(BenchmarkRunner):
(BenchmarkRunner.cleanup):
- Scripts/webkitpy/benchmark_runner/generic_factory.py:
(GenericFactory.iterateGetItem):
(GenericFactory.create):
- Scripts/webkitpy/benchmark_runner/utils.py:
(loadModule):
(loadJSONFromFile):
(forceRemove):
- 2:11 AM Changeset in webkit [185242] by
-
- 3 edits in trunk/Tools
Streamline run-benchmark browser
Reviewed by Ryosuke Niwa.
Add Caffeinate so computer won't sleep while running.
Remove calls to collect output since we were just dumping to stdout anyway.
Add xpc dyld_path.
Default to system safari.
- Scripts/webkitpy/benchmark_runner/browser_driver/osx_browser_driver.py:
(OSXBrowserDriver.launchProcess):
(OSXBrowserDriver.terminateProcesses):
(OSXBrowserDriver):
(OSXBrowserDriver.launchCaffeinateForProcess):
- Scripts/webkitpy/benchmark_runner/browser_driver/osx_safari_driver.py:
(OSXSafariDriver.launchUrl):
(OSXSafariDriver.closeBrowsers):
