WebKit-https.git
4 years agoDFG LICM needs to go all-in on the idea that some loops can't be LICMed
fpizlo@apple.com [Sun, 3 Jul 2016 17:46:39 +0000 (17:46 +0000)]
DFG LICM needs to go all-in on the idea that some loops can't be LICMed
https://bugs.webkit.org/show_bug.cgi?id=159388

Reviewed by Mark Lam.

Some time ago I acknowledged that LICM required loops to meet certain requirements that
may get broken by the time we do LICM, like that the terminal of the pre-header is ExitOK.
It used to be that we just ignored that requirement and would hoist anyway, but since
r189126 we've stopped hoisting out of loops that don't have ExitOK.  We also added tests
for the case that the pre-header doesn't exist or is invalid.

It turns out that this patch didn't go far enough: even though it made LICM avoid loops
that had an invalid pre-header, the part that updated the AI state in nested loops still
assumed that these loops had valid pre-headers.  We would crash in null dereference in
that loop if a nested loop had an invalid pre-header.

The fix is simple: don't update the AI state of nested loops that don't have pre-headers,
since we won't try to hoist out of those loops anyway.

* dfg/DFGLICMPhase.cpp:
(JSC::DFG::LICMPhase::attemptHoist):
* tests/stress/licm-no-pre-header-nested.js: Added. This would always crash before this fix.
(foo):
* tests/stress/licm-pre-header-cannot-exit-nested.js: Added. This was a failed attempt at a test, but I figure it's good to have weird code anyway.
(foo):
(valueOf):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202790 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Cocoa] Adopt modern API availability macros
mitz@apple.com [Sun, 3 Jul 2016 17:08:20 +0000 (17:08 +0000)]
[Cocoa] Adopt modern API availability macros
https://bugs.webkit.org/show_bug.cgi?id=159378

Reviewed by Sam Weinig.

iOS 10 and macOS Sierra introduce new macros for declaring API availability, defined in
CFAvailability.h, including API_AVAILABLE, API_DEPRECATED, and
API_DEPRECATED_WITH_REPLACEMENT. Rather than having positional parameters for macOS and iOS
versions, the new macros take a variable number of function-like parameters, where the
function specifies the platform to which it applies.

To have WebKit2’s postprocessed headers use the new macros, we make the following changes:

- Replace instances of WK_AVAILABLE(_mac, _ios) with WK_API_AVAILABLE(macosx(_mac), ios(_ios)),
  eliding a platform if its current availability is NA.
- Change instances of WK_CLASS_AVAILABLE(_mac, _ios) to
  WK_CLASS_AVAILABLE(macosx(_mac), ios(_ios)), eliding a platform if its current
  availability is NA.
- Replace instances of WK_DEPRECATED(_macIntro, _macDep, _iosIntro, _iosDep, ...) with
  WK_API_DEPRECATED_WITH_REPLACEMENT(_replacement, macosx(_macIntro, _macDep), ios(_iosIntro, _iosDep))
  when there is a one-for-one replacement, and with
  WK_API_DEPRECATED(_message, macosx(_macIntro, _macDep), ios(_iosIntro, _iosDep)) when the
  replacement isn’t one-for-one. Again, if the API doesn’t apply to a platform, then it’s
  elided.
- Replace instances of WK_CLASS_DEPRECATED with WK_CLASS_DEPRECATED_WITH_REPLACEMENT,
  following the same rules.
- Replace instances of the WK_ENUM_* macros with corresponding WK_API_* macros, now that
  all supported compilers support availability attributes on enum values.

* Configurations/WebKit.xcconfig: Disable header postprocessing when building for iOS 9.x,
  which doesn’t have the new macros.

* mac/postprocess-framework-headers.sh: Handle the new WK_* macros: WK_API_AVAILABLE becomes
  API_AVAILABLE; WK_API_DEPRECATED(_WITH_REPLACEMENT) become API_DEPRECATED(_WITH_REPLACEMENT);
  WK_CLASS_AVAILABLE and WK_CLASS_DEPRECATED(_WITH_REPLACEMENT) change to the corresponding
  API_ macros, prefixed with WK_EXTERN.
* Shared/API/Cocoa/WKFoundation.h: When header postprocessing is disabled, provide
  definitions of the new WK_* macros we use.

* UIProcess/API/Cocoa/WKWebViewConfiguration.h: In addition to the above, changed one
  instance of (iOS) 10_0 to WK_IOS_TBA.

Made the above changes in the following headers:

* Shared/API/Cocoa/WKBrowsingContextHandle.h:
* Shared/API/Cocoa/_WKFrameHandle.h:
* Shared/API/Cocoa/_WKHitTestResult.h:
* Shared/API/Cocoa/_WKRemoteObjectInterface.h:
* Shared/API/Cocoa/_WKRemoteObjectRegistry.h:
* Shared/API/Cocoa/_WKRenderingProgressEvents.h:
* Shared/API/Cocoa/_WKSameDocumentNavigationType.h:
* UIProcess/API/Cocoa/WKBackForwardList.h:
* UIProcess/API/Cocoa/WKBackForwardListItem.h:
* UIProcess/API/Cocoa/WKBrowsingContextController.h:
* UIProcess/API/Cocoa/WKBrowsingContextGroup.h:
* UIProcess/API/Cocoa/WKConnection.h:
* UIProcess/API/Cocoa/WKError.h:
* UIProcess/API/Cocoa/WKErrorPrivate.h:
* UIProcess/API/Cocoa/WKFrameInfo.h:
* UIProcess/API/Cocoa/WKFrameInfoPrivate.h:
* UIProcess/API/Cocoa/WKMenuItemIdentifiersPrivate.h:
* UIProcess/API/Cocoa/WKNavigation.h:
* UIProcess/API/Cocoa/WKNavigationAction.h:
* UIProcess/API/Cocoa/WKNavigationActionPrivate.h:
* UIProcess/API/Cocoa/WKNavigationData.h:
* UIProcess/API/Cocoa/WKNavigationDelegate.h:
* UIProcess/API/Cocoa/WKNavigationDelegatePrivate.h:
* UIProcess/API/Cocoa/WKNavigationResponse.h:
* UIProcess/API/Cocoa/WKOpenPanelParameters.h:
* UIProcess/API/Cocoa/WKPreferences.h:
* UIProcess/API/Cocoa/WKPreferencesPrivate.h:
* UIProcess/API/Cocoa/WKPreviewActionItem.h:
* UIProcess/API/Cocoa/WKPreviewActionItemIdentifiers.h:
* UIProcess/API/Cocoa/WKPreviewElementInfo.h:
* UIProcess/API/Cocoa/WKProcessGroup.h:
* UIProcess/API/Cocoa/WKProcessPool.h:
* UIProcess/API/Cocoa/WKProcessPoolPrivate.h:
* UIProcess/API/Cocoa/WKScriptMessage.h:
* UIProcess/API/Cocoa/WKSecurityOrigin.h:
* UIProcess/API/Cocoa/WKTypeRefWrapper.h:
* UIProcess/API/Cocoa/WKUIDelegate.h:
* UIProcess/API/Cocoa/WKUIDelegatePrivate.h:
* UIProcess/API/Cocoa/WKUserContentController.h:
* UIProcess/API/Cocoa/WKUserContentControllerPrivate.h:
* UIProcess/API/Cocoa/WKUserScript.h:
* UIProcess/API/Cocoa/WKUserScriptPrivate.h:
* UIProcess/API/Cocoa/WKViewPrivate.h:
* UIProcess/API/Cocoa/WKWebView.h:
* UIProcess/API/Cocoa/WKWebViewConfigurationPrivate.h:
* UIProcess/API/Cocoa/WKWebViewPrivate.h:
* UIProcess/API/Cocoa/WKWebsiteDataRecord.h:
* UIProcess/API/Cocoa/WKWebsiteDataRecordPrivate.h:
* UIProcess/API/Cocoa/WKWebsiteDataStore.h:
* UIProcess/API/Cocoa/WKWebsiteDataStorePrivate.h:
* UIProcess/API/Cocoa/WKWindowFeatures.h:
* UIProcess/API/Cocoa/_WKActivatedElementInfo.h:
* UIProcess/API/Cocoa/_WKAutomationSession.h:
* UIProcess/API/Cocoa/_WKContextMenuElementInfo.h:
* UIProcess/API/Cocoa/_WKDiagnosticLoggingDelegate.h:
* UIProcess/API/Cocoa/_WKDownload.h:
* UIProcess/API/Cocoa/_WKElementAction.h:
* UIProcess/API/Cocoa/_WKExperimentalFeature.h:
* UIProcess/API/Cocoa/_WKFindOptions.h:
* UIProcess/API/Cocoa/_WKFormInputSession.h:
* UIProcess/API/Cocoa/_WKInputDelegate.h:
* UIProcess/API/Cocoa/_WKLayoutMode.h:
* UIProcess/API/Cocoa/_WKOverlayScrollbarStyle.h:
* UIProcess/API/Cocoa/_WKProcessPoolConfiguration.h:
* UIProcess/API/Cocoa/_WKSessionState.h:
* UIProcess/API/Cocoa/_WKThumbnailView.h:
* UIProcess/API/Cocoa/_WKUserContentExtensionStore.h:
* UIProcess/API/Cocoa/_WKUserContentFilter.h:
* UIProcess/API/Cocoa/_WKUserContentWorld.h:
* UIProcess/API/Cocoa/_WKUserStyleSheet.h:
* UIProcess/API/Cocoa/_WKVisitedLinkProvider.h:
* UIProcess/API/Cocoa/_WKVisitedLinkStore.h:
* UIProcess/API/Cocoa/_WKWebsiteDataRecord.h:
* UIProcess/API/Cocoa/_WKWebsiteDataSize.h:
* UIProcess/API/Cocoa/_WKWebsiteDataStore.h:
* WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInFormDelegatePrivate.h:
* WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInFrame.h:
* WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInHitTestResult.h:
* WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInNodeHandle.h:
* WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInPageGroup.h:
* WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInScriptWorld.h:
* WebProcess/InjectedBundle/API/mac/WKDOMDocument.h:
* WebProcess/InjectedBundle/API/mac/WKDOMElement.h:
* WebProcess/InjectedBundle/API/mac/WKDOMNode.h:
* WebProcess/InjectedBundle/API/mac/WKDOMRange.h:
* WebProcess/InjectedBundle/API/mac/WKDOMText.h:
* WebProcess/InjectedBundle/API/mac/WKDOMTextIterator.h:
* WebProcess/InjectedBundle/API/mac/WKWebProcessPlugIn.h:
* WebProcess/InjectedBundle/API/mac/WKWebProcessPlugInBrowserContextController.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202789 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[win] Update TestExpectations after the MathML refactoring
fred.wang@free.fr [Sun, 3 Jul 2016 16:27:09 +0000 (16:27 +0000)]
[win] Update TestExpectations after the MathML refactoring
https://bugs.webkit.org/show_bug.cgi?id=159386

Reviewed by Alex Christensen.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202788 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUpdate windows test expectations for some MathML tests.
fred.wang@free.fr [Sun, 3 Jul 2016 16:18:03 +0000 (16:18 +0000)]
Update windows test expectations for some MathML tests.

Unreviewed test gardening.

* platform/win/TestExpectations:
* platform/win/mathml/opentype/horizontal-expected.txt:
* platform/win/mathml/opentype/horizontal-munderover-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202787 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agofast/css/image-set-unprefixed.html fails on iOS Simulator
ddkilzer@apple.com [Sun, 3 Jul 2016 11:59:58 +0000 (11:59 +0000)]
fast/css/image-set-unprefixed.html fails on iOS Simulator

Follow-up fix for:

"image-src" support is missing. We only support "-webkit-image-src"
https://bugs.webkit.org/show_bug.cgi?id=159373
<rdar://problem/27140443>

* platform/ios-simulator/fast/css/image-set-unprefixed-expected.txt: Added.
Cursor tests are expected to fail on iOS Simulator.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202786 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: CallFrame console.assert seen in tests (sourceCodeLocation instanceof...
commit-queue@webkit.org [Sun, 3 Jul 2016 03:36:10 +0000 (03:36 +0000)]
Web Inspector: CallFrame console.assert seen in tests (sourceCodeLocation instanceof SourceCodeLocation)
https://bugs.webkit.org/show_bug.cgi?id=159368

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-07-02
Reviewed by Timothy Hatcher.

* UserInterface/Test.html:
SourceCodeLocation was already included earlier. This redefines it and
thereby breaks some things.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202785 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoinspector/debugger/command-line-api-exception.html flakily times out on mac
commit-queue@webkit.org [Sun, 3 Jul 2016 03:33:34 +0000 (03:33 +0000)]
inspector/debugger/command-line-api-exception.html flakily times out on mac
https://bugs.webkit.org/show_bug.cgi?id=152029
<rdar://problem/23813812>

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-07-02
Reviewed by Timothy Hatcher.

Source/WebInspectorUI:

* UserInterface/Base/Object.js:
(WebInspector.Object.addEventListener):
Return the listener, this makes it cleaner to write tests.

LayoutTests:

* inspector/debugger/command-line-api-exception-expected.txt:
* inspector/debugger/command-line-api-exception-nested-catch-expected.txt:
* inspector/debugger/command-line-api-exception-nested-catch.html:
* inspector/debugger/command-line-api-exception.html:
* platform/mac/TestExpectations:
Rewrite these tests to use the new test harness.
This also is more careful about Pause/Resume timing differences
between Release and Debug builds.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202784 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Fix inspector/debugger/pause-reason.html
commit-queue@webkit.org [Sun, 3 Jul 2016 03:32:18 +0000 (03:32 +0000)]
Web Inspector: Fix inspector/debugger/pause-reason.html
https://bugs.webkit.org/show_bug.cgi?id=159325

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-07-02
Reviewed by Timothy Hatcher.

* platform/mac/TestExpectations:
* inspector/debugger/pause-reason.html:
This test was pausing on inspector code (evaluateInPage) and
auto-resuming. Instead change the test to pause on the page
doing the setTimeout, so we get the expected pause on next
statement evaluation.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202783 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBuild fix.
mitz@apple.com [Sat, 2 Jul 2016 17:52:09 +0000 (17:52 +0000)]
Build fix.

* wtf/Assertions.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202782 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, roll back unintentional commit in r202778.
fpizlo@apple.com [Sat, 2 Jul 2016 17:47:52 +0000 (17:47 +0000)]
Unreviewed, roll back unintentional commit in r202778.

* benchmarks/LockFairnessTest.cpp:
(main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202781 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMake it straightforward to start the Web Platform Tests HTTP server using run-webkit...
commit-queue@webkit.org [Sat, 2 Jul 2016 17:45:57 +0000 (17:45 +0000)]
Make it straightforward to start the Web Platform Tests HTTP server using run-webkit-httpd
https://bugs.webkit.org/show_bug.cgi?id=152486

Patch by Youenn Fablet <youennf@gmail.com> on 2016-07-02
Reviewed by Daniel Bates.

Add launch of web-platform-tests server by run-webkit-httpd.
Add two options to disable starting httpd and web-platform-tests servers.

* Scripts/run-webkit-httpd:
(parse_args):
(main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202780 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSynchronous preflight checker should set loading options to not use credentials
commit-queue@webkit.org [Sat, 2 Jul 2016 17:44:42 +0000 (17:44 +0000)]
Synchronous preflight checker should set loading options to not use credentials
https://bugs.webkit.org/show_bug.cgi?id=159351

Patch by Youenn Fablet <youenn@apple.com> on 2016-07-02
Reviewed by Alex Christensen.

Like for asynchronous preflighting, synchronous preflighting loading options should disqble any credentials.

No change of behavior as preflight request is expressly set to not use credentials in
createAccessControlPreflightRequest.

* loader/CrossOriginPreflightChecker.cpp:
(WebCore::CrossOriginPreflightChecker::doPreflight):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202779 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoScopes that are not under TDZ should still push their variables onto the TDZ stack...
fpizlo@apple.com [Sat, 2 Jul 2016 17:43:56 +0000 (17:43 +0000)]
Scopes that are not under TDZ should still push their variables onto the TDZ stack so that lifting TDZ doesn't bypass that scope
https://bugs.webkit.org/show_bug.cgi?id=159332
rdar://problem/27018958

Reviewed by Saam Barati.

This fixes an instacrash in this code:

    try{}catch(e){}print(e);let e;

We lift TDZ for "e" in "catch (e){}", but since that scope doesn't push anything onto the
TDZ stack, we lift TDZ from "let e".

The problem is that we weren't tracking the set of variables that do not have TDZ. We need
to track them to "block" the traversal that lifts TDZ. This change fixes this issue by
using a map that tracks all known variables, and tells you if they are under TDZ or not.

* bytecode/CodeBlock.h:
(JSC::CodeBlock::numParameters):
* bytecode/CodeOrigin.h:
* bytecompiler/BytecodeGenerator.cpp:
(JSC::Label::setLocation):
(JSC::Variable::dump):
(JSC::BytecodeGenerator::generate):
(JSC::BytecodeGenerator::BytecodeGenerator):
(JSC::BytecodeGenerator::pushLexicalScopeInternal):
(JSC::BytecodeGenerator::popLexicalScope):
(JSC::BytecodeGenerator::popLexicalScopeInternal):
(JSC::BytecodeGenerator::prepareLexicalScopeForNextForLoopIteration):
(JSC::BytecodeGenerator::variable):
(JSC::BytecodeGenerator::needsTDZCheck):
(JSC::BytecodeGenerator::liftTDZCheckIfPossible):
(JSC::BytecodeGenerator::pushTDZVariables):
(JSC::BytecodeGenerator::getVariablesUnderTDZ):
(JSC::BytecodeGenerator::endGenerator):
(WTF::printInternal):
* bytecompiler/BytecodeGenerator.h:
(JSC::Variable::isConst):
(JSC::Variable::setIsReadOnly):
* interpreter/CallFrame.h:
(JSC::ExecState::topOfFrame):
* tests/stress/lift-tdz-bypass-catch.js: Added.
(foo):
(catch):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202778 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUpdate EFL expectations for some MathML tests.
fred.wang@free.fr [Sat, 2 Jul 2016 11:29:42 +0000 (11:29 +0000)]
Update EFL expectations for some MathML tests.

Unreviewed test gardening.

* platform/efl/TestExpectations: Mark mathml/presentation/radical-bar-visibility.html as failing.
* platform/efl/mathml/opentype/horizontal-munderover-expected.txt: Update expectation using latest buildbot results.
* platform/efl/mathml/presentation/menclose-notation-values-expected.txt: Ditto.
* platform/efl/mathml/presentation/mspace-children-expected.txt: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202777 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUpdate windows expectations for some MathML tests.
fred.wang@free.fr [Sat, 2 Jul 2016 08:24:51 +0000 (08:24 +0000)]
Update windows expectations for some MathML tests.

Unreviewed test gardening.

* platform/win/TestExpectations:
* platform/win/mathml/opentype/large-operators-expected.txt: Added.
* platform/win/mathml/opentype/opentype-stretchy-expected.txt: Added.
* platform/win/mathml/opentype/opentype-stretchy-horizontal-expected.txt: Removed.
* platform/win/mathml/opentype/vertical-expected.txt: Added.
* platform/win/mathml/presentation/menclose-notation-default-longdiv-expected.txt:
* platform/win/mathml/presentation/mo-stretch-expected.txt: Added.
* platform/win/mathml/presentation/roots-expected.txt: Added.
* platform/win/mathml/radical-fallback-expected.png: Added.
* platform/win/mathml/radical-fallback-expected.txt: Added.
* platform/efl/mathml/radical-fallback-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202776 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUpdate EFL expectations for some MathML tests.
fred.wang@free.fr [Sat, 2 Jul 2016 07:54:19 +0000 (07:54 +0000)]
Update EFL expectations for some MathML tests.

Unreviewed test gardening.

* platform/efl/accessibility/math-mo-fences-and-separators-expected.txt: Added.
* platform/efl/mathml/opentype/horizontal-expected.txt:
* platform/efl/mathml/opentype/horizontal-munderover-expected.txt:
* platform/efl/mathml/opentype/large-operators-expected.png: Removed.
* platform/efl/mathml/opentype/large-operators-expected.txt:
* platform/efl/mathml/opentype/opentype-stretchy-expected.png: Removed.
* platform/efl/mathml/opentype/opentype-stretchy-expected.txt:
* platform/efl/mathml/opentype/vertical-expected.png: Removed.
* platform/efl/mathml/opentype/vertical-expected.txt:
* platform/efl/mathml/presentation/menclose-notation-default-longdiv-expected.png: Removed.
* platform/efl/mathml/presentation/menclose-notation-default-longdiv-expected.txt:
* platform/efl/mathml/presentation/menclose-notation-values-expected.txt:
* platform/efl/mathml/presentation/mo-stretch-expected.png: Removed.
* platform/efl/mathml/presentation/mo-stretch-expected.txt:
* platform/efl/mathml/presentation/mspace-children-expected.txt:
* platform/efl/mathml/presentation/roots-expected.png: Removed.
* platform/efl/mathml/presentation/roots-expected.txt:
* platform/efl/mathml/radical-fallback-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202775 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUpdate iOS expectations for some MathML tests.
fred.wang@free.fr [Sat, 2 Jul 2016 07:11:34 +0000 (07:11 +0000)]
Update iOS expectations for some MathML tests.

Unreviewed test gardening.

* platform/ios-simulator-wk1/TestExpectations:
* platform/ios-simulator-wk2/TestExpectations:
* platform/ios-simulator/TestExpectations:
* platform/ios-simulator/mathml/mathml-in-dashboard-actual.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202774 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[EFL] Many layout tests are timeout after r202532.
hs85.jeong@samsung.com [Sat, 2 Jul 2016 03:03:25 +0000 (03:03 +0000)]
[EFL] Many layout tests are timeout after r202532.
https://bugs.webkit.org/show_bug.cgi?id=159337

Reviewed by Gyuyoung Kim.

WebPageProxy::forceRepaint() wait for the next commit from the web process
after r202532, but WebPageProxy::callAfterNextPresentationUpdate() is not
implemented in CoordinatedDrawingAreaProxy.

* UIProcess/CoordinatedGraphics/CoordinatedDrawingAreaProxy.cpp:
(WebKit::CoordinatedDrawingAreaProxy::contentsRect):
(WebKit::CoordinatedDrawingAreaProxy::dispatchAfterEnsuringDrawing):
(WebKit::CoordinatedDrawingAreaProxy::sizeDidChange):
* UIProcess/CoordinatedGraphics/CoordinatedDrawingAreaProxy.h:
(WebKit::CoordinatedDrawingAreaProxy::page):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202773 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION(r189668): Notification tests are flakey
mmaxfield@apple.com [Sat, 2 Jul 2016 02:53:11 +0000 (02:53 +0000)]
REGRESSION(r189668): Notification tests are flakey
https://bugs.webkit.org/show_bug.cgi?id=159375
<rdar://problem/22760990>

Reviewed by Alexey Proskuryakov.

Source/WebKit2:

Implement WKNotificationManagerGetLocalID(). For more information, see the entry in Tools/ChangeLog.

* UIProcess/API/C/WKNotificationManager.cpp:
(WKNotificationManagerGetLocalID):
* UIProcess/Notifications/WebNotificationManagerProxy.cpp:
(WebKit::WebNotificationManagerProxy::notificationLocalIDForTesting):
* UIProcess/Notifications/WebNotificationManagerProxy.h:
* WebKit2.xcodeproj/project.pbxproj:

Tools:

Notifications are objects which must exist in both the UI Process and the Web Process. Each process
identifies a notification object by a unique ID. When the Web Process sends a message regarding a
notification to the UI Process, the UI Process's WebNotificationManagerProxy holds a map from
(Page ID, Web Process notification ID) -> UI Process notification ID. This works as intended.

Our tests, however, include an additional method, simulateWebNotificationClick(), which is implemented
by WebKitTestRunner in the Web Process via the Injected Bundle. This method involves sending a message
to the UI process, to handle the simulated click. However, that RPC didn't perform the same local ->
global notification ID mapping, causing the wrong notification to be investigated.

The solution is for WebNotificationProvider, implemented in WebKitTestRunner in the UI Process, to
manually perform this same mapping. Luckily, this object already receives callbacks every time a
notification is created or destroyed. However, because this object is implemented outside WebKit,
it isn't privy to the internal Web Process ID used inside WebNotificationmanagerProxy. Therefore,
this patch adds a private testing function which returns this internal ID. Once given this intenal ID,
WebNotificationProvider can properly map between the different IDs.

* WebKitTestRunner/TestController.cpp:
(WTR::TestController::runTestingServerLoop):
(WTR::TestController::simulateWebNotificationClick):
* WebKitTestRunner/WebNotificationProvider.cpp:
(WTR::WebNotificationProvider::showWebNotification):
(WTR::removeGlobalIDFromIDMap):
(WTR::WebNotificationProvider::closeWebNotification):
(WTR::WebNotificationProvider::removeNotificationManager):
(WTR::WebNotificationProvider::simulateWebNotificationClick):
(WTR::WebNotificationProvider::reset):
* WebKitTestRunner/WebNotificationProvider.h:

LayoutTests:

* platform/mac/TestExpectations:
* platform/mac-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202772 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r202766.
commit-queue@webkit.org [Sat, 2 Jul 2016 02:32:39 +0000 (02:32 +0000)]
Unreviewed, rolling out r202766.
https://bugs.webkit.org/show_bug.cgi?id=159382

The new test asserts every time (Requested by ap on #webkit).

Reverted changeset:

"Web Inspector: Sending XHR with UTF8 encoded data shows
garbled data in Resource sidebar"
https://bugs.webkit.org/show_bug.cgi?id=159358
http://trac.webkit.org/changeset/202766

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202771 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[JSC] RegExp.compile is not returning the regexp when it succeed
commit-queue@webkit.org [Sat, 2 Jul 2016 02:20:39 +0000 (02:20 +0000)]
[JSC] RegExp.compile is not returning the regexp when it succeed
https://bugs.webkit.org/show_bug.cgi?id=159381

Patch by Benjamin Poulain <bpoulain@apple.com> on 2016-07-01
Reviewed by Mark Lam.

Source/JavaScriptCore:

Spec:
-https://tc39.github.io/ecma262/#sec-regexp.prototype.compile
-https://tc39.github.io/ecma262/#sec-regexpinitialize

* runtime/RegExpPrototype.cpp:
(JSC::regExpProtoFuncCompile):

LayoutTests:

* js/regexp-compile-expected.txt:
* js/script-tests/regexp-compile.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202770 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoprepareForDestruction() always needs to be called before destroying the Document...
zalan@apple.com [Sat, 2 Jul 2016 01:15:44 +0000 (01:15 +0000)]
prepareForDestruction() always needs to be called before destroying the Document object.
https://bugs.webkit.org/show_bug.cgi?id=159372
rdar://problem/26788150

Reviewed by Antti Koivisto.

We should never start destroying the Document object without calling prepareForDestruction() first.
It ensures that render tree gets nuked before we start tearing down the node tree.

Source/WebCore:

Test: fast/history/page-cache-destroy-document.html

* dom/Document.cpp:
(WebCore::Document::removedLastRef):

LayoutTests:

* fast/history/page-cache-destroy-document-expected.txt: Added.
* fast/history/page-cache-destroy-document.html: Added.
* fast/history/resources/page-cache-destroy-helper.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202769 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agofix "ASSERTION FAILED: currentOffset() >= currentLineStartOffset()"
sbarati@apple.com [Sat, 2 Jul 2016 00:59:38 +0000 (00:59 +0000)]
fix "ASSERTION FAILED: currentOffset() >= currentLineStartOffset()"
https://bugs.webkit.org/show_bug.cgi?id=158572
<rdar://problem/26884092>

Reviewed by Mark Lam.

Source/JavaScriptCore:

There is a bug in our lexer when we notice the pattern:
```<return|continue|break|...etc> // some comment here```
Our code will say that the token for the comment is a semicolon.
This is the correct semantics, however, it would give the semicolon
a start offset of the comment, but it will give its line start offset
the newline after the comment.  This breaks the invariant in the lexer/parser
that the offset for the current line starting point must be less than or equal to
than the start offset of any token on that line. This invariant was broken because
the line start offset was greater than the token start offset. To maintain this
invariant, we claim that the semicolon token is located where the comment starts,
and that its line start offset is the line start offset for the line with the
comment on it.  There are other solutions that maintain this invariant, but this
solution provides the best error messages.

* parser/Lexer.cpp:
(JSC::Lexer<T>::lex):
* parser/Parser.h:
(JSC::Parser::internalSaveLexerState):
* tests/stress/obscure-error-message-dont-crash.js: Added.
(try.eval.or.catch):

LayoutTests:

* js/parser-syntax-check-expected.txt:
* js/script-tests/parser-syntax-check.js:
(invalid.or.break.catch):
(catch): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202768 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMarking fast/history/page-cache-webdatabase-pending-transaction.html as flaky on Mac
ryanhaddad@apple.com [Sat, 2 Jul 2016 00:56:08 +0000 (00:56 +0000)]
Marking fast/history/page-cache-webdatabase-pending-transaction.html as flaky on Mac
https://bugs.webkit.org/show_bug.cgi?id=159379

Unreviewed test gardening.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202767 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Sending XHR with UTF8 encoded data shows garbled data in Resource...
commit-queue@webkit.org [Sat, 2 Jul 2016 00:55:44 +0000 (00:55 +0000)]
Web Inspector: Sending XHR with UTF8 encoded data shows garbled data in Resource sidebar
https://bugs.webkit.org/show_bug.cgi?id=159358

Patch by Johan K. Jensen <jj@johanjensen.dk> on 2016-07-01
Reviewed by Joseph Pecoraro.

Source/WebCore:

Test: http/tests/inspector/network/xhr-request-data-encoded-correctly.html

* inspector/InspectorNetworkAgent.cpp:
(WebCore::buildObjectForResourceRequest):

LayoutTests:

Based on Chromium patch:
<https://chromium.googlesource.com/chromium/src/+/bcbb663864624ab38b36731eb2edc839a90f9e65%5E%21/#F2>

* http/tests/inspector/network/xhr-request-data-encoded-correctly-expected.txt:
* http/tests/inspector/network/xhr-request-data-encoded-correctly.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202766 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago"image-src" support is missing. We only support "-webkit-image-src"
dino@apple.com [Sat, 2 Jul 2016 00:22:15 +0000 (00:22 +0000)]
"image-src" support is missing. We only support "-webkit-image-src"
https://bugs.webkit.org/show_bug.cgi?id=159373
<rdar://problem/27140443>

Patch by Brent Fulgham and Dean Jackson.
Reviewed by Dean Jackson and Brent Fulgham.

Source/WebCore:

Support unprefixed image-set.

Test: fast/css/image-set-unprefixed.html

* css/CSSImageSetValue.cpp:
(WebCore::CSSImageSetValue::customCSSText):
* css/CSSParser.cpp:
(WebCore::isImageSetFunctionValue): New helper function
that checks prefixed and unprefixed form.
(WebCore::CSSParser::parseValue): Use the helper.
(WebCore::CSSParser::parseContent):
(WebCore::CSSParser::parseFillImage):
(WebCore::CSSParser::parseBorderImage):

LayoutTests:

Convert some existing tests to use the unprefixed form,
and add a test that checks we are compatible with
the prefixed form.

* cssom/cssvalue-comparison-expected.txt:
* cssom/cssvalue-comparison.html:
* fast/css/cursor-parsing-image-set-expected.txt:
* fast/css/cursor-parsing-image-set.html:
* fast/css/image-set-setting-expected.txt:
* fast/css/image-set-unprefixed-expected.txt: Added.
* fast/css/image-set-unprefixed.html: Added.
* fast/css/script-tests/image-set-setting.js:
(testImageSetRule):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202765 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoLand test expectations for rdar://problem/27141291.
ryanhaddad@apple.com [Fri, 1 Jul 2016 23:58:30 +0000 (23:58 +0000)]
Land test expectations for rdar://problem/27141291.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202764 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION(r201405): Fullscreen video no longer enters low-power mode
jer.noble@apple.com [Fri, 1 Jul 2016 23:41:45 +0000 (23:41 +0000)]
REGRESSION(r201405): Fullscreen video no longer enters low-power mode
https://bugs.webkit.org/show_bug.cgi?id=159220
<rdar://problem/26701056>

Reviewed by Beth Dakin.

In r201405, we removed the call to setTopContentInset(0) to avoid a flash
during repainting where the WebProcess still thought it had an inset and
the UIProcess did not, but the >0 inset breaks low power video playback
in fullscreen. So, instead, fix the repaint issue by setting a CALayer
fence which makes sure both the UIProcess and WebProcess paint simultaneously
after resizing and changing the top content inset.

This requires a new message from WebPageProxy -> WebPage: setTopContentInsetFenced().

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::setTopContentInset):
* UIProcess/mac/WKFullScreenWindowController.h:
* UIProcess/mac/WKFullScreenWindowController.mm:
(-[WKFullScreenWindowController enterFullScreen:]):
(-[WKFullScreenWindowController finishedEnterFullScreenAnimation:]):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::setTopContentInsetFenced):
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202763 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoPossible null Range dereference under AXObjectCache::visiblePositionFromCharacterOffset()
cdumez@apple.com [Fri, 1 Jul 2016 23:25:04 +0000 (23:25 +0000)]
Possible null Range dereference under AXObjectCache::visiblePositionFromCharacterOffset()
https://bugs.webkit.org/show_bug.cgi?id=159330
<rdar://problem/27123752>

Reviewed by Benjamin Poulain.

rangeForUnorderedCharacterOffsets() can return a null Range but we failed
to do a null check before dereferencing it.

* accessibility/AXObjectCache.cpp:
(WebCore::AXObjectCache::visiblePositionFromCharacterOffset):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202762 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRegression(r199087): window.focus() / window.close() can no longer be called by a...
cdumez@apple.com [Fri, 1 Jul 2016 23:22:23 +0000 (23:22 +0000)]
Regression(r199087): window.focus() / window.close() can no longer be called by a Window's opener
https://bugs.webkit.org/show_bug.cgi?id=159364
<rdar://problem/27117169>

Reviewed by Gavin Barraclough.

Source/WebCore:

window.focus() / window.close() could no longer be called by a Window's opener
after r199087, which would break focusing of open iWork documents on icloud.com.

Before r199087, we would construct a new function in the caller's context every
time window.focus and window.close was accessed. r199087 fixed the issue so that
we always call the same function. However, those functions are using
[CallWith=Document] and they are were no longer passed the *caller*'s document
as a result. This broke focus / close permission checking as the code needed the
caller's document to do the check.

This patch introduces [CallWith=CallerDocument] and [CallWith=CallerWindow] so
that the implementation can now pass the caller's Document / Window to the
implementation. The bindings rely on JSDOMWindow's callerDOMWindow() to get the
caller DOMWindow / document. This new functionality is now used for window.close
and window.focus to unbreak their permission checking.

Test: fast/dom/Window/child-window-focus.html

* bindings/scripts/CodeGeneratorJS.pm:
(GenerateCallWith):
* bindings/scripts/IDLAttributes.txt:
* page/DOMWindow.cpp:
(WebCore::DOMWindow::focus):
* page/DOMWindow.h:
* page/DOMWindow.idl:

LayoutTests:

Add layout test coverage.

* fast/dom/Window/child-window-focus-expected.txt: Added.
* fast/dom/Window/child-window-focus.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202761 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoTest gardening for Sierra WK1
ryanhaddad@apple.com [Fri, 1 Jul 2016 23:15:48 +0000 (23:15 +0000)]
Test gardening for Sierra WK1

Unreviewed test gardening.

* platform/mac-wk1/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202760 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemove failure expectation for mathml/presentation/bug97990.html on ios-simulator
fred.wang@free.fr [Fri, 1 Jul 2016 21:50:13 +0000 (21:50 +0000)]
Remove failure expectation for mathml/presentation/bug97990.html on ios-simulator

Unreviewed test gardening.

Patch by Frederic Wang <fwang@igalia.com> on 2016-07-01

* platform/ios-simulator-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202759 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Mac] Get rid of support for old SDKs
mitz@apple.com [Fri, 1 Jul 2016 21:43:04 +0000 (21:43 +0000)]
[Mac] Get rid of support for old SDKs
https://bugs.webkit.org/show_bug.cgi?id=159369

Reviewed by Anders Carlsson.

* Shared/API/Cocoa/WKFoundation.h: Removed the WK_ARRAY, WK_DICTIONARY, and WK_SET macros,
  which were only needed for pre-10.11 SDKs.

* Shared/mac/WebEventFactory.mm:
(WebKit::mouseButtonForEvent): Removed checking for pre-10.10.3 SDKs.
(WebKit::globalPointForEvent): Ditto.
(WebKit::pointForEvent): Ditto.
(WebKit::WebEventFactory::createWebMouseEvent): Ditto.

* UIProcess/API/Cocoa/WKBackForwardList.h: Replaced WK_ARRAY(...) with NSArray<...>.

* UIProcess/API/Cocoa/WKPreferences.mm: Ditto.
(-[WKPreferences _setFixedPitchFontFamily:]):
(+[WKPreferences _experimentalFeatures]):
* UIProcess/API/Cocoa/WKPreferencesPrivate.h: Ditto.

* UIProcess/API/Cocoa/WKUIDelegate.h: Ditto.
* UIProcess/API/Cocoa/WKUIDelegatePrivate.h: Ditto.

* UIProcess/API/Cocoa/WKUserContentController.h: Ditto.
* UIProcess/API/Cocoa/WKUserContentControllerPrivate.h: Ditto.

* UIProcess/API/Cocoa/WKUserScriptPrivate.h: Ditto.

* UIProcess/API/Cocoa/WKWebsiteDataRecord.h: Replaced WK_SET(...) with NSSet<...>.
* UIProcess/API/Cocoa/WKWebsiteDataStore.h: Ditto.
* UIProcess/API/Cocoa/WKWebsiteDataStore.mm: Ditto.
(-[WKWebsiteDataStore fetchDataRecordsOfTypes:completionHandler:]):
(-[WKWebsiteDataStore _fetchDataRecordsOfTypes:withOptions:completionHandler:]):

* UIProcess/API/Cocoa/WKWebsiteDataStorePrivate.h: Ditto.

* UIProcess/API/Cocoa/_WKProcessPoolConfiguration.h: Replaced WK_ARRAY(...) with NSArray<...>.

* UIProcess/API/Cocoa/_WKUserStyleSheet.h: Ditto.
* UIProcess/API/Cocoa/_WKUserStyleSheet.mm: Ditto.
(-[_WKUserStyleSheet initWithSource:forMainFrameOnly:]):
(-[_WKUserStyleSheet initWithSource:forMainFrameOnly:legacyWhitelist:legacyBlacklist:userContentWorld:]):
(-[_WKUserStyleSheet initWithSource:forMainFrameOnly:legacyWhitelist:legacyBlacklist:baseURL:userContentWorld:]):

* UIProcess/API/Cocoa/_WKWebsiteDataSize.h: Replaced WK_SET(...) with NSSet<...>.

* UIProcess/Cocoa/WebViewImpl.mm:
(WebKit::WebViewImpl::pressureChangeWithEvent): Removed checking for pre-10.10.3 SDKs.

* WebProcess/InjectedBundle/API/Cocoa/WKWebProcessPlugInFrame.h: Replaced WK_ARRAY(...) with
  NSArray<...>.

* mac/postprocess-framework-headers.sh: Removed sed rules for WK_ARRAY, WK_DICTIONARY and
  WK_SET.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202758 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] Possible null Range dereference under computeAutocorrectionContext()
cdumez@apple.com [Fri, 1 Jul 2016 20:59:08 +0000 (20:59 +0000)]
[iOS] Possible null Range dereference under computeAutocorrectionContext()
https://bugs.webkit.org/show_bug.cgi?id=159328
<rdar://problem/26766720>

Reviewed by Benjamin Poulain.

Source/WebCore:

* editing/Editor.cpp:
(WebCore::Editor::compositionRange):
* editing/Editor.h:
Update to return a RefPtr instead of a PassRefPtr and use nullptr
instead of 0 in the implementation.

Source/WebKit2:

The code in computeAutocorrectionContext() was checking Editor::hasComposition()
before dereferencing Editor::compositionRange(). However, compositionRange()
can also return null in other cases (e.g. compositionStart == compositionEnd).

Drop the check for hasComposition() and do a null check on the value returned
by compositionRange() instead.

* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::computeAutocorrectionContext):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202757 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdded a colorful new image and fixed the Open Graph attribute.
jond@apple.com [Fri, 1 Jul 2016 20:58:12 +0000 (20:58 +0000)]
Added a colorful new image and fixed the Open Graph attribute.
https://bugs.webkit.org/show_bug.cgi?id=159359

Reviewed by Timothy Hatcher.

The Open Graph image tag should use the property attribute not
the itemprop attribute.

* wp-content/plugins/social-meta.php:
* wp-content/themes/webkit/images/ogimage.png:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202756 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago__defineGetter__/__defineSetter__ should throw exceptions
commit-queue@webkit.org [Fri, 1 Jul 2016 20:55:55 +0000 (20:55 +0000)]
__defineGetter__/__defineSetter__ should throw exceptions
https://bugs.webkit.org/show_bug.cgi?id=142934

Patch by Benjamin Poulain <bpoulain@apple.com> on 2016-07-01
Reviewed by Mark Lam.

Source/JavaScriptCore:

* runtime/ObjectPrototype.cpp:
(JSC::objectProtoFuncDefineGetter):
(JSC::objectProtoFuncDefineSetter):

LayoutTests:

* js/dom/dom-properties-are-configurable-expected.txt: Added.
* js/dom/dom-properties-are-configurable.html: Added.
* js/object-literal-duplicate-properties-expected.txt:
* js/property-getters-and-setters-expected.txt:
* js/script-tests/object-literal-duplicate-properties.js:
* js/script-tests/property-getters-and-setters.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202755 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoPrevent crash when attempting to copy an image
bfulgham@apple.com [Fri, 1 Jul 2016 20:52:30 +0000 (20:52 +0000)]
Prevent crash when attempting to copy an image
https://bugs.webkit.org/show_bug.cgi?id=159288
<rdar://problem/23507828>

Reviewed by Brady Eidson.

WebKit could crash if the HTMLImageElement had a null cachedImage member.
A little investigation showed that the URL we retrieve from the cachedImage
is never used, so rather than adding a null check, let's just pass a
default argument. The actual URL gets read from elsewhere in the pasteboard
processing code.

* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::performActionOnElement): Don't bother passing the
unused URL argument to the pasteboard method.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202754 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSource/JavaScriptCore:
jond@apple.com [Fri, 1 Jul 2016 20:51:02 +0000 (20:51 +0000)]
Source/JavaScriptCore:
Moved Web Animations and Resource Timing feature entries to WebCore.
https://bugs.webkit.org/show_bug.cgi?id=159356

Reviewed by Timothy Hatcher.

* features.json:

Source/WebCore:
Updated Picture element and WOFF 2 status
https://bugs.webkit.org/show_bug.cgi?id=159356

Reviewed by Timothy Hatcher.

Status updates and clean-up to move Web Animations and Resource Timing entries from JSC to WebCore.

* features.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202753 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[JSC] Date.toGMTString should be the Date.toUTCString function
commit-queue@webkit.org [Fri, 1 Jul 2016 20:49:31 +0000 (20:49 +0000)]
[JSC] Date.toGMTString should be the Date.toUTCString function
https://bugs.webkit.org/show_bug.cgi?id=159318

Patch by Benjamin Poulain <bpoulain@apple.com> on 2016-07-01
Reviewed by Mark Lam.

Source/JavaScriptCore:

See https://tc39.github.io/ecma262/#sec-date.prototype.togmtstring

* runtime/DatePrototype.cpp:
(JSC::DatePrototype::finishCreation):
(JSC::dateProtoFuncToGMTString): Deleted.

LayoutTests:

* js/date-prototype-expected.txt: Added.
* js/date-prototype.html: Added.
* js/script-tests/date-prototype.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202752 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd early return when processing content extensions if there aren't any.
akling@apple.com [Fri, 1 Jul 2016 20:46:14 +0000 (20:46 +0000)]
Add early return when processing content extensions if there aren't any.
<https://webkit.org/b/159363>

Reviewed by Antti Koivisto.

Short-circuit outta there if there aren't any extensions to query.

* contentextensions/ContentExtensionsBackend.cpp:
(WebCore::ContentExtensions::ContentExtensionsBackend::processContentExtensionRulesForLoad):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202751 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRebaseline some MathML tests.
fred.wang@free.fr [Fri, 1 Jul 2016 20:44:40 +0000 (20:44 +0000)]
Rebaseline some MathML tests.

Unreviewed test gardening.

Patch by Frederic Wang <fwang@igalia.com> on 2016-07-01

* platform/gtk/mathml/opentype/munderover-window-resize-expected.txt: Removed.
* platform/ios-simulator/mathml/presentation/roots-expected.png: Added.
* platform/mac/mathml/opentype/munderover-window-resize-expected.png: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202750 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoHTMLMediaElement::resume() may cause JavaScript execution
eric.carlson@apple.com [Fri, 1 Jul 2016 20:31:18 +0000 (20:31 +0000)]
HTMLMediaElement::resume() may cause JavaScript execution
https://bugs.webkit.org/show_bug.cgi?id=159327
<rdar://problem/27131641>

Reviewed by Jer Noble.

HTMLMediaElement::updatePlayState can cause an element to begin playing and enter fullscreen,
which can result in a call to the media controls and JavaScript execution. Javascript is not
allowed allowed to run when a page resumes, so make the call to updatePlayState asynchronous.

No new tests, I wasn't able to create a test that triggers the crash.

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::scheduleDelayedAction): Support UpdatePlayState.
(WebCore::HTMLMediaElement::pendingActionTimerFired): Ditto.
(WebCore::HTMLMediaElement::setReadyState): UpdateMediaState -> UpdateState.
(WebCore::HTMLMediaElement::playInternal): Don't call updateMediaController, it is called
  by updatePlayState.
(WebCore::HTMLMediaElement::setMuted): UpdateMediaState -> UpdateState.
(WebCore::HTMLMediaElement::mediaPlayerTimeChanged): Ditto.
(WebCore::HTMLMediaElement::mediaEngineWasUpdated): Update media state asynchronously.
(WebCore::HTMLMediaElement::updatePlayState): Add parameter to allow update to happen
  asynchronously.
(WebCore::HTMLMediaElement::setPlaying): UpdateMediaState -> UpdateState.
(WebCore::HTMLMediaElement::setPausedInternal): Update media state asynchronously.
(WebCore::HTMLMediaElement::mediaPlayerCurrentPlaybackTargetIsWirelessChanged):
  UpdateMediaState -> UpdateState.
(WebCore::HTMLMediaElement::removeEventListener): Ditto.
(WebCore::HTMLMediaElement::enqueuePlaybackTargetAvailabilityChangedEvent): Ditto.
(WebCore::HTMLMediaElement::updateMediaState): UpdateMediaState -> UpdateState
* html/HTMLMediaElement.h:
* html/HTMLMediaElementEnums.h: Add UpdatePlayState.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202749 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUpdate OS X and iOS expectations for some MathML pixel tests.
fred.wang@free.fr [Fri, 1 Jul 2016 19:41:30 +0000 (19:41 +0000)]
Update OS X and iOS expectations for some MathML pixel tests.
https://bugs.webkit.org/show_bug.cgi?id=159339

Unreviewed test gardening.

Patch by Frederic Wang <fwang@igalia.com> on 2016-07-01

* mathml/opentype/fraction-line.html: Add a comment about font requirements.
* mathml/opentype/horizontal-munderover.html: Ditto.
* mathml/opentype/horizontal.html: Ditto.
* mathml/opentype/large-operators-italic-correction.html: Ditto.
* mathml/opentype/large-operators.html: Ditto.
* mathml/opentype/vertical.html: Ditto.
* mathml/presentation/attributes-mathvariant.html: Ditto.
* mathml/presentation/fractions-linethickness.html: Ditto.
* mathml/presentation/tokenElements-mathvariant.html: Ditto.
* mathml/radical-fallback.html: Ditto.
* platform/ios-simulator-wk1/TestExpectations: Update expectations.
* platform/ios-simulator-wk2/TestExpectations: Ditto.
* platform/ios-simulator-wk2/mathml/opentype/large-operators-expected.png: Removed.
* platform/ios-simulator-wk2/mathml/opentype/vertical-expected.png: Removed.
* platform/ios-simulator-wk2/mathml/presentation/mo-stretch-expected.png: Removed.
* platform/ios-simulator/TestExpectations:
* platform/ios-simulator/mathml/opentype/horizontal-expected.txt: Removed.
* platform/ios-simulator/mathml/opentype/horizontal-munderover-expected.txt: Removed.
* platform/ios-simulator/mathml/opentype/large-operators-expected.txt: Removed.
* platform/ios-simulator/mathml/opentype/opentype-stretchy-expected.png: Added.
* platform/ios-simulator/mathml/opentype/opentype-stretchy-horizontal-expected.png: Added.
* platform/ios-simulator/mathml/opentype/opentype-stretchy-horizontal-expected.txt:
* platform/ios-simulator/mathml/opentype/vertical-expected.txt: Removed.
* platform/ios-simulator/mathml/presentation/menclose-notation-default-longdiv-expected.png: Added.
* platform/ios-simulator/mathml/presentation/menclose-notation-values-expected.png: Added.
* platform/ios-simulator/mathml/presentation/mo-stretch-expected.png: Added.
* platform/ios-simulator/mathml/presentation/mo-stretch-expected.txt:
* platform/ios-simulator/mathml/presentation/mspace-children-expected.png: Added.
* platform/ios-simulator/mathml/radical-fallback-expected.png:
* platform/mac/TestExpectations: Update expectations.
* platform/mac/mathml/opentype/horizontal-expected.txt: Removed.
* platform/mac/mathml/opentype/horizontal-munderover-expected.txt: Removed.
* platform/mac/mathml/opentype/large-operators-expected.txt: Removed.
* platform/mac/mathml/opentype/opentype-stretchy-expected.png: Added.
* platform/mac/mathml/opentype/vertical-expected.txt: Removed.
* platform/mac/mathml/presentation/menclose-notation-values-expected.png:
* platform/mac/mathml/presentation/menclose-notation-values-expected.txt:
* platform/mac/mathml/presentation/mo-stretch-expected.png:
* platform/mac/mathml/presentation/mo-stretch-expected.txt:
* platform/mac/mathml/presentation/mspace-children-expected.png: Added.
* platform/mac/mathml/presentation/roots-expected.png:
* platform/mac/mathml/presentation/roots-expected.txt:
* platform/mac/mathml/radical-fallback-expected.png:
* platform/mac/mathml/radical-fallback-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202748 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoBlob content type not preserved when retrieving blobs from IndexedDB.
beidson@apple.com [Fri, 1 Jul 2016 19:10:35 +0000 (19:10 +0000)]
Blob content type not preserved when retrieving blobs from IndexedDB.
<rdar://problem/27057357> and https://bugs.webkit.org/show_bug.cgi?id=159360

Reviewed by Alex Christensen.

Source/WebCore:

Test: storage/indexeddb/modern/blob-svg-image.html

* fileapi/Blob.cpp:
(WebCore::Blob::Blob):

* fileapi/ThreadableBlobRegistry.cpp:
(WebCore::postToMainThread):
(WebCore::ThreadableBlobRegistry::registerBlobURLOptionallyFileBacked): Pass along the content type
  to the blob registry so that if the file-backed blob takes over, it has the content type.
(WebCore::threadableQueue): Deleted.
* fileapi/ThreadableBlobRegistry.h:

* platform/network/BlobRegistry.h:

* platform/network/BlobRegistryImpl.cpp:
(WebCore::BlobRegistryImpl::registerBlobURL):
(WebCore::BlobRegistryImpl::registerBlobURLOptionallyFileBacked):
* platform/network/BlobRegistryImpl.h:

Source/WebKit2:

* NetworkProcess/FileAPI/NetworkBlobRegistry.cpp:
(WebKit::NetworkBlobRegistry::registerBlobURLOptionallyFileBacked):
* NetworkProcess/FileAPI/NetworkBlobRegistry.h:

* NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::registerBlobURLOptionallyFileBacked):
* NetworkProcess/NetworkConnectionToWebProcess.h:
* NetworkProcess/NetworkConnectionToWebProcess.messages.in:

* WebProcess/FileAPI/BlobRegistryProxy.cpp:
(WebKit::BlobRegistryProxy::registerBlobURLOptionallyFileBacked):
* WebProcess/FileAPI/BlobRegistryProxy.h:

LayoutTests:

* storage/indexeddb/modern/blob-svg-image-expected.txt: Added.
* storage/indexeddb/modern/blob-svg-image.html: Added.
* storage/indexeddb/modern/resources/blob-svg-image.js: Added.
* storage/indexeddb/modern/resources/blob-svg-image1.html: Added.
* storage/indexeddb/modern/resources/blob-svg-image2.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202747 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Cocoa] Get rid of WK_NULLABLE_SPECIFIER now that all supported compilers understand...
mitz@apple.com [Fri, 1 Jul 2016 19:07:20 +0000 (19:07 +0000)]
[Cocoa] Get rid of WK_NULLABLE_SPECIFIER now that all supported compilers understand _Nullable
https://bugs.webkit.org/show_bug.cgi?id=159361

Reviewed by Anders Carlsson.

* Shared/API/Cocoa/WKFoundation.h:
* UIProcess/API/Cocoa/WKNavigationDelegate.h:
* UIProcess/API/Cocoa/WKUIDelegate.h:
* UIProcess/API/Cocoa/WKWebView.h:
* mac/postprocess-framework-headers.sh:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202746 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoASSERTION FAILED: info.bmBitsPixel == 32
pvollan@apple.com [Fri, 1 Jul 2016 18:08:30 +0000 (18:08 +0000)]
ASSERTION FAILED: info.bmBitsPixel == 32
https://bugs.webkit.org/show_bug.cgi?id=17737

Reviewed by Brent Fulgham.

The ::SelectObject call will fail if m_backingStoreBitmap is already selected into a device context.
This happens when this method is called indirectly from WebView::updateBackingStore during normal
painting. There is no point continuing, since we would just be scrolling a 1x1 bitmap which is
selected into the device context by default. We can just scroll by repainting the scroll rectangle.

* WebView.cpp:
(WebView::scrollBackingStore): Invalidate the scroll rectangle if the ::SelectObject call fails.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202744 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSimplify LayoutTestRelay
ap@apple.com [Fri, 1 Jul 2016 17:08:38 +0000 (17:08 +0000)]
Simplify LayoutTestRelay
https://bugs.webkit.org/show_bug.cgi?id=159353

Reviewed by Daniel Bates.

1. Use UDID to pass the device from webkitpy to LayoutTestRelay, searching for it
by name and runtime makes no sense.
2. Removed code to create devices, webkitpy always does this in advance.
3. Don't create copies of testing apps - we can install the same one into
each simulator, now that we don't run multiple tester processes in a single simulator.
4. Pass information on how to find I/O pipes via environment, not via Info.plist.
5. Removed code that kept testing apps alive in the background.
6. Slightly updated coding style in functions that I touched.

* LayoutTestRelay/LayoutTestRelay/CoreSimulatorSPI.h:
Updated for what we acually use at this point.

* LayoutTestRelay/LayoutTestRelay/LTRelayController.h:
* LayoutTestRelay/LayoutTestRelay/LTRelayController.m:
(-[LTRelayController initWithDevice:productDir:appPath:deviceUDID:dumpToolArguments:]):
(-[LTRelayController ipcIdentifier]):
(-[LTRelayController processName]):
(-[LTRelayController didReceiveStdoutData:]):
(-[LTRelayController didCrashWithMessage:]):
(-[LTRelayController installApp]):
(-[LTRelayController _environmentVariables]):
(-[LTRelayController launchApp]):
(-[LTRelayController start]):
(-[LTRelayController initWithDevice:productDir:appPath:identifierSuffix:dumpToolArguments:]): Deleted.
(-[LTRelayController uniqueAppPath]): Deleted.
(-[LTRelayController uniqueAppURL]): Deleted.
(-[LTRelayController uniqueAppIdentifier]): Deleted.
(-[LTRelayController createUniqueApp]): Deleted.
* LayoutTestRelay/LayoutTestRelay/main.m:
(usage):
(getRequiredStringArgument):
(main):
(getTestingSimDevice): Deleted.
* Scripts/webkitpy/port/driver.py:
(IOSSimulatorDriver.cmd_line):
* WebKitTestRunner/ios/TestControllerIOS.mm:
(WTR::TestController::platformInitialize):
* WebKitTestRunner/ios/mainIOS.mm:
(-[WebKitTestRunnerApp applicationDidEnterBackground:]):
* DumpRenderTree/mac/DumpRenderTree.mm:
(dumpRenderTree):
(-[DumpRenderTree applicationDidEnterBackground:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202743 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Make enabling protocol logging in inspector tests easy to remember
commit-queue@webkit.org [Fri, 1 Jul 2016 17:06:57 +0000 (17:06 +0000)]
Web Inspector: Make enabling protocol logging in inspector tests easy to remember
https://bugs.webkit.org/show_bug.cgi?id=159323

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-07-01
Reviewed by Brian Burg.

Now you just need to type InspectorTest.debug() or
ProtocolTest.debug() and it will include stderr logging.

* UserInterface/Test/FrontendTestHarness.js:
(FrontendTestHarness.prototype.debug):
* UserInterface/Test/ProtocolTestHarness.js:
(ProtocolTestHarness.prototype.debug):
(ProtocolTestHarness):
* UserInterface/Test/TestHarness.js:
(TestHarness.prototype.debug):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202742 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMake ResourceLoaderOptions derive from FetchOptions
commit-queue@webkit.org [Fri, 1 Jul 2016 17:03:27 +0000 (17:03 +0000)]
Make ResourceLoaderOptions derive from FetchOptions
https://bugs.webkit.org/show_bug.cgi?id=159345

Patch by Youenn Fablet <youenn@apple.com> on 2016-07-01
Reviewed by Alex Christensen.

No change of behavior.

* Modules/fetch/FetchLoader.cpp:
(WebCore::FetchLoader::start):
* loader/CrossOriginPreflightChecker.cpp:
(WebCore::CrossOriginPreflightChecker::startPreflight):
* loader/ResourceLoaderOptions.h:
(WebCore::ResourceLoaderOptions::fetchOptions): Deleted.
(WebCore::ResourceLoaderOptions::setFetchOptions): Deleted.
* loader/SubresourceLoader.cpp:
(WebCore::SubresourceLoader::willSendRequestInternal):
* loader/ThreadableLoader.h: Removing securityOrigin field (left over from https://bugs.webkit.org/show_bug.cgi?id=159221)

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202741 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Cleanup inspector/console/console-message.html
commit-queue@webkit.org [Fri, 1 Jul 2016 17:00:48 +0000 (17:00 +0000)]
Web Inspector: Cleanup inspector/console/console-message.html
https://bugs.webkit.org/show_bug.cgi?id=159324

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-07-01
Reviewed by Brian Burg.

* inspector/console/console-message.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202740 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[cmake] Build TestNetscapePlugin only if NPAPI is enabled.
annulen@yandex.ru [Fri, 1 Jul 2016 16:43:02 +0000 (16:43 +0000)]
[cmake] Build TestNetscapePlugin only if NPAPI is enabled.
https://bugs.webkit.org/show_bug.cgi?id=159344

Reviewed by Alex Christensen.

* DumpRenderTree/CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202739 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win] Animations tests are crashing in debug mode.
pvollan@apple.com [Fri, 1 Jul 2016 16:39:53 +0000 (16:39 +0000)]
[Win] Animations tests are crashing in debug mode.
https://bugs.webkit.org/show_bug.cgi?id=159335

Reviewed by Alex Christensen.

A MSVC runtime check fails because an uninitialized variable is being used.

* css/StyleResolver.cpp:
(WebCore::StyleResolver::keyframeStylesForAnimation):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202738 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUpdate JSC_functionOverrides to handle the new SourceCode strings that have params.
mark.lam@apple.com [Fri, 1 Jul 2016 16:38:11 +0000 (16:38 +0000)]
Update JSC_functionOverrides to handle the new SourceCode strings that have params.
https://bugs.webkit.org/show_bug.cgi?id=159321

Reviewed by Geoffrey Garen.

And add tests so that this won't fail silently and bit rot anymore.

* API/tests/FunctionOverridesTest.cpp: Added.
(testFunctionOverrides):
* API/tests/FunctionOverridesTest.h: Added.
* API/tests/testapi-function-overrides.js: Added.
* API/tests/testapi.c:
(main):
* JavaScriptCore.xcodeproj/project.pbxproj:
* bytecode/UnlinkedFunctionExecutable.cpp:
(JSC::UnlinkedFunctionExecutable::link):
* shell/PlatformWin.cmake:
* tools/FunctionOverrides.cpp:
(JSC::FunctionOverrides::FunctionOverrides):
(JSC::FunctionOverrides::reinstallOverrides):
(JSC::initializeOverrideInfo):
(JSC::FunctionOverrides::initializeOverrideFor):
* tools/FunctionOverrides.h:
(JSC::FunctionOverrides::clear):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202737 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDeadlock inside -[WebCoreNSURLSession dealloc]
jer.noble@apple.com [Fri, 1 Jul 2016 15:40:39 +0000 (15:40 +0000)]
Deadlock inside -[WebCoreNSURLSession dealloc]
https://bugs.webkit.org/show_bug.cgi?id=159331
<rdar://problem/27122716>

Reviewed by Alex Christensen.

A Function<> object can wrap any callable type, including a C++ lambda.

dispatchFunctionsFromMainThread() holds a lock while iterating over the functions in
functionQueue(), and during ths iteration, the previous callable object is destroyed by
assigning the result of functionQueue().takeFirst(). Because lambdas (and other callables,
like functors) can own objects, destroying this callable can have side effects, and if one
of those side effects is to call callOnMainThread(), this can deadlock.

Move this side-effect-having call outside the locked block by clearing the function object
immediately after calling it.

* wtf/MainThread.cpp:
(WTF::dispatchFunctionsFromMainThread):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202736 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[cmake] Make LEGACY_WEB_AUDIO depend on WEB_AUDIO.
annulen@yandex.ru [Fri, 1 Jul 2016 15:29:09 +0000 (15:29 +0000)]
[cmake] Make LEGACY_WEB_AUDIO depend on WEB_AUDIO.
https://bugs.webkit.org/show_bug.cgi?id=159338

Reviewed by Michael Catanzaro.

Enabling LEGACY_WEB_AUDIO makes no sense when WEB_AUDIO is
disabled.

* Source/cmake/WebKitFeatures.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202735 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoES6: Implement HasRestrictedGlobalProperty when checking for global lexical tier...
commit-queue@webkit.org [Fri, 1 Jul 2016 15:25:35 +0000 (15:25 +0000)]
ES6: Implement HasRestrictedGlobalProperty when checking for global lexical tier conflicts
https://bugs.webkit.org/show_bug.cgi?id=148763

Patch by Caio Lima <ticaiolima@gmail.com> on 2016-07-01
Reviewed by Saam Barati

I've implemented the ES6 spec 8.1.1.4.14
(http://www.ecma-international.org/ecma-262/6.0/index.html#sec-hasrestrictedglobalproperty)
that defines when a global property can be shadowed.

Added some test cases into global-lexical-redeclare-variable.js

* runtime/Executable.cpp:
(JSC::ProgramExecutable::initializeGlobalProperties):
* tests/stress/global-lexical-redeclare-variable.js:
(catch):
* tests/stress/multiple-files-tests/global-lexical-redeclare-variable/eighth.js: Added.
* tests/stress/multiple-files-tests/global-lexical-redeclare-variable/nineth.js: Added.
* tests/stress/multiple-files-tests/global-lexical-redeclare-variable/seventh.js: Added.
* tests/stress/multiple-files-tests/global-lexical-redeclare-variable/sixth.js:
* tests/stress/multiple-files-tests/global-lexical-redeclare-variable/tenth.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202734 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRebaseline opentype-stretchy-horizontal.html for EFL.
fred.wang@free.fr [Fri, 1 Jul 2016 12:46:58 +0000 (12:46 +0000)]
Rebaseline opentype-stretchy-horizontal.html for EFL.
https://bugs.webkit.org/show_bug.cgi?id=135561

Unreviewed test gardening.

Patch by Frederic Wang <fwang@igalia.com> on 2016-07-01

* platform/efl/TestExpectations:
* platform/efl/mathml/opentype/opentype-stretchy-horizontal-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202733 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRegenerate EFL expectation for foreign-element-in-token.html
fred.wang@free.fr [Fri, 1 Jul 2016 12:36:20 +0000 (12:36 +0000)]
Regenerate EFL expectation for foreign-element-in-token.html
https://bugs.webkit.org/show_bug.cgi?id=133972

Unreviewed test gardening.

Patch by Frederic Wang <fwang@igalia.com> on 2016-07-01

* platform/efl/TestExpectations:
* platform/efl/mathml/presentation/foreign-element-in-token-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202732 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix warnings about non existent paths in TestExpectations
fred.wang@free.fr [Fri, 1 Jul 2016 12:07:31 +0000 (12:07 +0000)]
Fix warnings about non existent paths in TestExpectations

Unreviewed test gardening.

Patch by Frederic Wang <fwang@igalia.com> on 2016-07-01

* platform/efl/TestExpectations:
* platform/gtk/TestExpectations:
* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202731 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFix error in r202729
fred.wang@free.fr [Fri, 1 Jul 2016 10:30:28 +0000 (10:30 +0000)]
Fix error in r202729

Unreviewed test gardening.

Patch by Frederic Wang <fwang@igalia.com> on 2016-07-01

* platform/ios-simulator/TestExpectations: Revert unintentional change of horizontal-munderover.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202730 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoReorganize and improve comments for MathML test expectations on ios and mac.
fred.wang@free.fr [Fri, 1 Jul 2016 09:52:22 +0000 (09:52 +0000)]
Reorganize and improve comments for MathML test expectations on ios and mac.

Unreviewed test gardening.

Patch by Frederic Wang <fwang@igalia.com> on 2016-07-01

* platform/ios-simulator-wk2/TestExpectations:
* platform/ios-simulator/TestExpectations:
* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202729 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoAdd a runtime flag for DOM iterators
commit-queue@webkit.org [Fri, 1 Jul 2016 07:47:14 +0000 (07:47 +0000)]
Add a runtime flag for DOM iterators
https://bugs.webkit.org/show_bug.cgi?id=159300

Patch by Youenn Fablet <youennf@gmail.com> on 2016-07-01
Reviewed by Alex Christensen.

Source/JavaScriptCore:

* runtime/CommonIdentifiers.h:

Source/WebCore:

* Modules/fetch/FetchHeaders.idl: Making iterator runtime-enabled.
* bindings/generic/RuntimeEnabledFeatures.h:
(WebCore::RuntimeEnabledFeatures::setDOMIteratorEnabled):
(WebCore::RuntimeEnabledFeatures::domIteratorEnabled):
* bindings/scripts/CodeGeneratorJS.pm:
(ToMethodName): Fixing dOM -> dom casing issue.
(GenerateImplementation): Using addIterableProperties new method.
(addIterableProperties): Activating property addition according runtime flag if iterator is rnutime flagged.
* bindings/scripts/IDLParser.pm:
(parseOptionalIterableInterface): Adding extendedAttributes to iterable.
* bindings/scripts/test/JS/JSTestNode.cpp:
(WebCore::JSTestNodePrototype::finishCreation):
* bindings/scripts/test/JS/JSTestObj.cpp:
(WebCore::JSTestObjPrototype::finishCreation):
* bindings/scripts/test/TestNode.idl: Making iterator runtime-enabled.
* bindings/scripts/test/TestObj.idl: Ditto.
* css/FontFaceSet.idl: Ditto.
* dom/NodeList.idl: Ditto.

Source/WebKit/mac:

* WebView/WebPreferenceKeysPrivate.h:
* WebView/WebPreferences.mm:
(-[WebPreferences DOMIteratorEnabled]):
(-[WebPreferences setDOMIteratorEnabled:]):
* WebView/WebPreferencesPrivate.h:
* WebView/WebView.mm:
(-[WebView _preferencesChanged:]):

Source/WebKit/win:

* WebPreferenceKeysPrivate.h:
* WebPreferences.cpp:
(WebPreferences::domIteratorEnabled):
(WebPreferences::setDOMIteratorEnabled):
(WebView::notifyPreferencesChanged):
* WebPreferencesreferences.h:
* WebView.cpp:
(WebView::notifyPreferencesChanged):

Source/WebKit2:

* Shared/WebPreferencesDefinitions.h:
* WebProcess/InjectedBundle/InjectedBundle.cpp:
(WebKit::InjectedBundle::overrideBoolPreferenceForTestRunner):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::updatePreferences):

Tools:

* DumpRenderTree/mac/DumpRenderTree.mm:
(resetWebPreferencesToConsistentValues):
* DumpRenderTree/win/DumpRenderTree.cpp:
(resetWebPreferencesToConsistentValues):
* Scripts/webkitperl/FeatureList.pm:
* WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::beginTesting):
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::setDOMIteratorEnabled):
* WebKitTestRunner/InjectedBundle/TestRunner.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202728 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoEliminate trailing whitespace in MathML code
fred.wang@free.fr [Fri, 1 Jul 2016 07:29:53 +0000 (07:29 +0000)]
Eliminate trailing whitespace in MathML code
https://bugs.webkit.org/show_bug.cgi?id=159091

Patch by Frederic Wang <fwang.igalia.com> on 2016-07-01
Reviewed by Alex Christensen.

No new tests, behavior is unchanged.

* rendering/mathml/RenderMathMLBlock.cpp:
(WebCore::RenderMathMLBlock::baselinePosition):
(WebCore::RenderMathMLBlock::paint):
(WebCore::parseMathMLNamedSpace):
* rendering/mathml/RenderMathMLBlock.h:
* rendering/mathml/RenderMathMLFenced.cpp:
(WebCore::RenderMathMLFenced::updateFromElement):
(WebCore::RenderMathMLFenced::addChild):
* rendering/mathml/RenderMathMLFenced.h:
* rendering/mathml/RenderMathMLFraction.cpp:
(WebCore::RenderMathMLFraction::styleDidChange):
(WebCore::RenderMathMLFraction::paint):
* rendering/mathml/RenderMathMLFraction.h:
* rendering/mathml/RenderMathMLMath.h:
* rendering/mathml/RenderMathMLMenclose.h:
* rendering/mathml/RenderMathMLOperator.cpp:
* rendering/mathml/RenderMathMLOperator.h:
* rendering/mathml/RenderMathMLRoot.cpp:
(WebCore::RenderMathMLRoot::paint):
* rendering/mathml/RenderMathMLScripts.cpp:
* rendering/mathml/RenderMathMLSpace.cpp:
* rendering/mathml/RenderMathMLSpace.h:
* rendering/mathml/RenderMathMLToken.h:
* rendering/mathml/RenderMathMLUnderOver.cpp:
* rendering/mathml/RenderMathMLUnderOver.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202727 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSmall cleanup: Remove unused functions RenderObject::isRenderMathML*Wrapper
fred.wang@free.fr [Fri, 1 Jul 2016 07:06:42 +0000 (07:06 +0000)]
Small cleanup: Remove unused functions RenderObject::isRenderMathML*Wrapper
https://bugs.webkit.org/show_bug.cgi?id=159333

Patch by Frederic Wang <fwang@igalia.com> on 2016-07-01
Reviewed by Alex Christensen.

After the refactoring of RenderMathMLRoot and RenderMathMLScripts, the anonymous flexbox
wrappers used in the old layout implementation have been removed. We thus remove the
corresponding isRender* function from RenderObject.

No new tests, behavior is unchanged.

* rendering/RenderObject.h:
(WebCore::RenderObject::isRenderMathMLRootWrapper): Deleted.
(WebCore::RenderObject::isRenderMathMLScriptsWrapper): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202726 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Mac] Get rid of the old timey rubber-banding linen pattern.
akling@apple.com [Fri, 1 Jul 2016 06:58:29 +0000 (06:58 +0000)]
[Mac] Get rid of the old timey rubber-banding linen pattern.
<https://webkit.org/b/159329>

Reviewed by Benjamin Poulain.

Source/WebCore:

Remove the "ScrollingOverhang" custom GraphicsLayer appearance since that was only used to
install the old timey linen pattern behind the web content.

We now always just set the overhang area's background color to the document background color.

This fixes an issue where we could end up loading the linen pattern and keeping it in memory
despite never actually showing it on screen.

* platform/ScrollbarTheme.h:
(WebCore::ScrollbarTheme::setUpOverhangAreasLayerContents): Deleted.
* platform/graphics/GraphicsLayer.cpp:
* platform/graphics/GraphicsLayer.h:
* platform/graphics/ca/cocoa/PlatformCALayerCocoa.mm:
(PlatformCALayerCocoa::updateCustomAppearance):
* platform/mac/ScrollbarThemeMac.h:
* platform/mac/ScrollbarThemeMac.mm:
(WebCore::linenBackgroundColor): Deleted.
(WebCore::ScrollbarThemeMac::setUpOverhangAreaBackground): Deleted.
(WebCore::ScrollbarThemeMac::removeOverhangAreaBackground): Deleted.
(WebCore::ScrollbarThemeMac::setUpOverhangAreasLayerContents): Deleted.
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::updateOverflowControlsLayers):
(WebCore::RenderLayerCompositor::setRootExtendedBackgroundColor):

Source/WebKit2:

* Shared/mac/RemoteLayerTreePropertyApplier.mm:
(WebKit::updateCustomAppearance):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202725 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSource/WebKit2:
commit-queue@webkit.org [Fri, 1 Jul 2016 05:23:02 +0000 (05:23 +0000)]
Source/WebKit2:
Add SPI to allow restoring session state without navigation in WKPage
https://bugs.webkit.org/show_bug.cgi?id=159326

Patch by Tina Liu <iting_liu@apple.com> on 2016-06-30
Reviewed by Brady Eidson.

* UIProcess/API/C/WKPage.cpp:
(restoreFromSessionState):
Added a helper function that takes a bool parameter to specify whether
to navigate or not when restoring from the session state.
(WKPageRestoreFromSessionState):
Calling restoreFromSessionState, passing navigate = true.
(WKPageRestoreFromSessionStateWithoutNavigation):
Ditto but passing navigate = false.
* UIProcess/API/C/WKPagePrivate.h:

Tools:
Add an API test for WKPageRestoreFromSessionStateWithoutNavigation.
https://bugs.webkit.org/show_bug.cgi?id=159326

Patch by Tina Liu <iting_liu@apple.com> on 2016-06-30
Reviewed by Brady Eidson.

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit2/RestoreSessionStateWithoutNavigation.cpp: Added.
(TestWebKitAPI::didFinishLoadForFrame):
(TestWebKitAPI::didChangeBackForwardListForPage):
(TestWebKitAPI::setPageLoaderClient):
Set the page loader client and register for didFinishLoadForFrame and
didChangeBackForwardList callbacks.
(TestWebKitAPI::createSessionStateData):
Load a webpage ("simple.html") and return the session state for this page.
(TestWebKitAPI::TEST):
Restore the page session state with that of "simple.html" without navigation.
Verify that the committed URL is NULL since there's no navigation involved.
Verify that the current item in the back forward list, which should be what
we restored from the session state, has the expected URL.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202724 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] WebContent processes do not exit promptly
cdumez@apple.com [Fri, 1 Jul 2016 01:29:47 +0000 (01:29 +0000)]
[iOS] WebContent processes do not exit promptly
https://bugs.webkit.org/show_bug.cgi?id=159301
<rdar://problem/26965488>

Reviewed by Anders Carlsson.

WebContent processes do not exit promptly, they hang around for 10 seconds
until the watchdog forcefully calls exit().

This patch addresses the issue by asking XPC to exit when clean. It also
fixes 2 XPC transactions that were leaking so that XPC can become clean.

* DatabaseProcess/EntryPoint/mac/XPCService/DatabaseServiceEntryPoint.mm:
(DatabaseServiceInitializer):
* NetworkProcess/EntryPoint/mac/XPCService/NetworkServiceEntryPoint.mm:
(NetworkServiceInitializer):
* PluginProcess/EntryPoint/mac/XPCService/PluginServiceEntryPoint.mm:
(PluginServiceInitializer):
* WebProcess/EntryPoint/mac/XPCService/WebContentServiceEntryPoint.mm:
(WebContentServiceInitializer):
Add extra priorityBoostMessage parameter which is then passed along to
ChildProcess via initialization parameters. ChildProcess is now in
charge of retaining the message for as long as it needs the priority
boost. In particular, ChildProcess now takes care of releasing the
boost message before existing to avoid leaking an XPC transaction.

* Shared/ChildProcess.cpp:
(WebKit::ChildProcess::initialize):
Retain priorityBoostMessage as a data member.

(WebKit::ChildProcess::stopRunLoop):
(WebKit::ChildProcess::platformStopRunLoop):
* Shared/ios/ChildProcessIOS.mm:
(WebKit::ChildProcess::platformStopRunLoop):
On iOS, call XPCServiceExit() to exit instead of RunLoop::main().stop()
which did not work.

* Shared/ChildProcess.h:
* Shared/EntryPointUtilities/mac/XPCService/XPCServiceEntryPoint.h:
(WebKit::XPCServiceInitializer):
Set priorityBoostMessage on ChildProcessInitializationParameters.

* Shared/EntryPointUtilities/mac/XPCService/XPCServiceEntryPoint.mm:
(WebKit::XPCServiceExit):
Add XPCServiceExit() function which destroys the priority boost
message, calls xpc_transaction_end() to balance the
xpc_transaction_begin() in XPCServiceInitializer() and then call
xpc_transaction_exit_clean() to ask XPC to exit when clean.

* Shared/EntryPointUtilities/mac/XPCService/XPCServiceMain.mm:
(WebKit::XPCServiceEventHandler):
Leaking the "pre-bootstrap" event for priority boosting would cause us to
leak an XPC transaction, which would prevent XPC from becoming clean and
exiting. Instead, we now pass it along to the initialization function.
We then pass it to ChildProcess which manages the lifetime of this message
instead of leaking it.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202723 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed build fix.
mattbaker@apple.com [Fri, 1 Jul 2016 00:13:49 +0000 (00:13 +0000)]
Unreviewed build fix.

* Localizations/en.lproj/localizedStrings.js:
Add "Area" localized string.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202722 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoSkipping flaky inspector/debugger/break-on-* tests
ryanhaddad@apple.com [Fri, 1 Jul 2016 00:07:48 +0000 (00:07 +0000)]
Skipping flaky inspector/debugger/break-on-* tests
https://bugs.webkit.org/show_bug.cgi?id=159272

Unreviewed test gardening.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202720 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoCreate a generic "linked-on-or-after" check for new CSP Rules
jiewen_tan@apple.com [Fri, 1 Jul 2016 00:03:58 +0000 (00:03 +0000)]
Create a generic "linked-on-or-after" check for new CSP Rules
https://bugs.webkit.org/show_bug.cgi?id=159322
<rdar://problem/27117220>

Reviewed by Brent Fulgham.

Source/WebCore:

Create a generic "linked-on-or-after" check for new CSP Rules and cleanup
quirks for Ecobee, Quora and XtraMat.

* platform/RuntimeApplicationChecks.h:
* platform/RuntimeApplicationChecks.mm:
(WebCore::IOSApplication::isEcobee): Deleted.
(WebCore::IOSApplication::isQuora): Deleted.
(WebCore::IOSApplication::isXtraMath): Deleted.

Source/WebKit/mac:

* WebView/WebView.mm:
(shouldAllowContentSecurityPolicySourceStarToMatchAnyProtocol):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202719 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMoving flaky expectation for inspector/codemirror/prettyprinting-css-rules.html from...
ryanhaddad@apple.com [Thu, 30 Jun 2016 23:48:50 +0000 (23:48 +0000)]
Moving flaky expectation for inspector/codemirror/prettyprinting-css-rules.html from mac-wk1 to mac.
https://bugs.webkit.org/show_bug.cgi?id=153460

Unreviewed test gardening.

* platform/mac-wk1/TestExpectations:
* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202718 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Wrong function name next to scope
commit-queue@webkit.org [Thu, 30 Jun 2016 23:24:58 +0000 (23:24 +0000)]
Web Inspector: Wrong function name next to scope
https://bugs.webkit.org/show_bug.cgi?id=158210
<rdar://problem/26543093>

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-06-30
Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

* CMakeLists.txt:
* JavaScriptCore.xcodeproj/project.pbxproj:
Add DebuggerLocation. A helper for describing a unique location.

* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::setConstantRegisters):
When compiled with debug info, add a SymbolTable rare data pointer
back to the CodeBlock. This will be used later to get JSScope debug
info if Web Inspector pauses.

* runtime/SymbolTable.h:
* runtime/SymbolTable.cpp:
(JSC::SymbolTable::cloneScopePart):
(JSC::SymbolTable::prepareForTypeProfiling):
(JSC::SymbolTable::uniqueIDForVariable):
(JSC::SymbolTable::uniqueIDForOffset):
(JSC::SymbolTable::globalTypeSetForOffset):
(JSC::SymbolTable::globalTypeSetForVariable):
Rename rareData and include a CodeBlock pointer.

(JSC::SymbolTable::rareDataCodeBlock):
(JSC::SymbolTable::setRareDataCodeBlock):
Setter and getter for the rare data. It should only be set once.

(JSC::SymbolTable::visitChildren):
Visit the rare data code block if we have one.

* runtime/JSSymbolTableObject.h:
* runtime/JSSymbolTableObject.cpp:
(JSC::JSSymbolTableObject::deleteProperty):
(JSC::JSSymbolTableObject::getOwnNonIndexPropertyNames):
Give JSSymbolTable its own class info. JSWithScope was unexpectedly
inheriting from JSSymbolTable since it did not have its own and
was using JSScope's class info. Also do a bit of cleanup.

* debugger/DebuggerLocation.cpp: Added.
(JSC::DebuggerLocation::DebuggerLocation):
* debugger/DebuggerLocation.h: Added.
(JSC::DebuggerLocation::DebuggerLocation):
Construction from a ScriptExecutable.

* runtime/JSScope.cpp:
(JSC::JSScope::symbolTable):
* runtime/JSScope.h:
* debugger/DebuggerScope.h:
* debugger/DebuggerScope.cpp:
(JSC::DebuggerScope::name):
(JSC::DebuggerScope::location):
Name and location for a scope. This uses:
JSScope -> SymbolTable -> CodeBlock -> Executable

* inspector/protocol/Debugger.json:
* inspector/InjectedScriptSource.js:
(InjectedScript.CallFrameProxy.prototype._wrapScopeChain):
(InjectedScript.CallFrameProxy._createScopeJson):
* inspector/JSJavaScriptCallFrame.cpp:
(Inspector::valueForScopeType):
(Inspector::valueForScopeLocation):
(Inspector::JSJavaScriptCallFrame::scopeDescriptions):
(Inspector::JSJavaScriptCallFrame::scopeType): Deleted.
* inspector/JSJavaScriptCallFrame.h:
* inspector/JSJavaScriptCallFramePrototype.cpp:
(Inspector::JSJavaScriptCallFramePrototype::finishCreation):
(Inspector::jsJavaScriptCallFramePrototypeFunctionScopeDescriptions):
(Inspector::jsJavaScriptCallFramePrototypeFunctionScopeType): Deleted.
Simplify this code to build the objects we will send across the protocol
to descript a Scope.

Source/WebInspectorUI:

* UserInterface/Controllers/DebuggerManager.js:
(WebInspector.DebuggerManager.prototype._scopeChainNodeFromPayload):
Include new payload data in the construction call.
All the new data is optional, so we gracefully handle
legacy backends.

* UserInterface/Models/ScopeChainNode.js:
(WebInspector.ScopeChainNode):
(WebInspector.ScopeChainNode.prototype.get type):
(WebInspector.ScopeChainNode.prototype.get objects):
(WebInspector.ScopeChainNode.prototype.get name):
(WebInspector.ScopeChainNode.prototype.get location):
(WebInspector.ScopeChainNode.prototype.get hash):
Hash is a rough (name:sourceId:line:column) string for quick comparisons.

(WebInspector.ScopeChainNode.prototype.makeLocalScope):
Make this an action you take on a scope, to avoid having to
do it at construction time, or making it a generic setting.

* UserInterface/Views/ScopeChainDetailsSidebarPanel.js:
(WebInspector.ScopeChainDetailsSidebarPanel.prototype._generateCallFramesSection):
This was wrong before. Move the work to CallFrame
and change it to be correct.

* UserInterface/CallFrame.js:
(WebInspector.CallFrame.prototype.mergedScopeChain):

This transforms the scope chain for a call frame from:

     scope1  scope2  scope3  scope4  scope5  scope6  scope7
      foo     foo     foo     bar     bar      -       -
     Block  Closure Closure Closure Closure   GLE     GBL

To:
     scope1  scope2&3   scope4&5  scope6  scope7
      foo      foo*       bar*      -       -
     Block    Local     Closure    GLE     GBL

Doing a few things:

    - Merge the first two Closure scopes sharing a location.
      These are the "var" and "let" Closure scopes in a function,
      and it is better to present these together in the UI.

    - Mark the first Closure scope within a function (*). When
      this is displayed in the UI, we can provide the name of
      the function: "Closure Scope (name)", and we even have
      location information that we can use to display a goto
      arrow if needed.

    - Make the first Closure scope the Local scope if it
      matches the Call Frame's function name. This lets us
      display the section as "Local Variables".

LayoutTests:

* inspector/debugger/paused-scopes-expected.txt: Added.
* inspector/debugger/paused-scopes.html: Added.
* inspector/debugger/resources/paused-scopes.js: Added.
Test dumps the call frames and scope chains for each call frame
when pausing at different locations in a program. Outputting
the hashes we can see even identically named functions have
different hashes because their location is different.

* inspector/model/scope-chain-node.html:
Make the test more stable on debug builds which run slower.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202717 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWebContent crash due to RELEASE_ASSERT(!m_inLoadPendingImages) in StyleResolver:...
antti@apple.com [Thu, 30 Jun 2016 23:12:03 +0000 (23:12 +0000)]
WebContent crash due to RELEASE_ASSERT(!m_inLoadPendingImages) in StyleResolver::~StyleResolver()
https://bugs.webkit.org/show_bug.cgi?id=159307
<rdar://problem/26184868>

Reviewed by Andreas Kling.

Pseudo elements are resolved in RenderTreeUpdater (instead of Style::TreeResolver). Their resolution may trigger
resource loads which can cause synchronous layout (when failing synchronously) and lead to destruction of the
the style resolver in post layout task.

No known reliable way to test this.

* style/RenderTreeUpdater.cpp:
(WebCore::RenderTreeUpdater::commit):

    Use PostResolutionCallbackDisabler in RenderTreeUpdater similarly to Style::TreeResolver. This prevents
    post layout tasks from running synchronously and closes this particular crash path.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202716 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agomissing exception checks in arrayProtoFuncReverse
sbarati@apple.com [Thu, 30 Jun 2016 22:35:30 +0000 (22:35 +0000)]
missing exception checks in arrayProtoFuncReverse
https://bugs.webkit.org/show_bug.cgi?id=159319
<rdar://problem/27083696>

Reviewed by Filip Pizlo.

* runtime/ArrayPrototype.cpp:
(JSC::arrayProtoFuncToString):
(JSC::arrayProtoFuncReverse):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202714 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: add pixel area column to layout timeline view
mattbaker@apple.com [Thu, 30 Jun 2016 22:29:32 +0000 (22:29 +0000)]
Web Inspector: add pixel area column to layout timeline view
https://bugs.webkit.org/show_bug.cgi?id=148677
<rdar://problem/22518762>

Reviewed by Timothy Hatcher.

The view can simply add the column. LayoutTimelineDataGridNode
already includes area data and cell formatting.

* UserInterface/Views/LayoutTimelineView.js:
(WebInspector.LayoutTimelineView):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202713 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoDrawing an SVG image into a <canvas> that is not in the DOM draws the wrong region
commit-queue@webkit.org [Thu, 30 Jun 2016 22:18:12 +0000 (22:18 +0000)]
Drawing an SVG image into a <canvas> that is not in the DOM draws the wrong region
https://bugs.webkit.org/show_bug.cgi?id=159276

Patch by Antoine Quint <graouts@apple.com> on 2016-06-30
Reviewed by Dean Jackson.

Source/WebCore:

In the event where the <img> element that we are passing to CanvasRenderingContext2D.drawImage()
points to an SVG resource, we ensure that the container for the SVG image is sized to match the
HTML element. The necessity for setting this container size, explained in webkit.org/b/148845,
is that we must ensure a cached image does not have an outdated container size.

Tests: svg/as-image/img-with-svg-resource-in-dom-and-drawImage.html
       svg/as-image/img-with-svg-resource-in-dom-no-size-and-drawImage.html
       svg/as-image/img-with-svg-resource-not-in-dom-and-drawImage.html
       svg/as-image/img-with-svg-resource-not-in-dom-no-size-and-drawImage.html

* html/canvas/CanvasRenderingContext2D.cpp:
(WebCore::CanvasRenderingContext2D::drawImage):

LayoutTests:

Adding a series of new tests to check we correctly respect mismatching source and
destination rectangles with SVG images as sources, both with the source <img> element
being present and absent from the DOM, and explicit sizes being set or not set.

* svg/as-image/img-with-svg-resource-in-dom-and-drawImage-expected.html: Added.
* svg/as-image/img-with-svg-resource-in-dom-and-drawImage.html: Added.
* svg/as-image/img-with-svg-resource-in-dom-no-size-and-drawImage-expected.html: Added.
* svg/as-image/img-with-svg-resource-in-dom-no-size-and-drawImage.html: Added.
* svg/as-image/img-with-svg-resource-not-in-dom-and-drawImage-expected.html: Added.
* svg/as-image/img-with-svg-resource-not-in-dom-and-drawImage.html: Added.
* svg/as-image/img-with-svg-resource-not-in-dom-no-size-and-drawImage-expected.html: Added.
* svg/as-image/img-with-svg-resource-not-in-dom-no-size-and-drawImage.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202712 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: the last normal tab cannot be closed with "Close Tab" context menu...
bburg@apple.com [Thu, 30 Jun 2016 22:14:20 +0000 (22:14 +0000)]
Web Inspector: the last normal tab cannot be closed with "Close Tab" context menu item
https://bugs.webkit.org/show_bug.cgi?id=159317
<rdar://problem/27042491>

Reviewed by Timothy Hatcher.

* UserInterface/Views/TabBarItem.js:
(WebInspector.TabBarItem.prototype._handleContextMenuEvent):
(WebInspector.TabBarItem): fix the expression that controls whether the item is disabled.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202711 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoget_by_id_with_this does not trigger a to_this in caller.
sbarati@apple.com [Thu, 30 Jun 2016 22:06:44 +0000 (22:06 +0000)]
get_by_id_with_this does not trigger a to_this in caller.
https://bugs.webkit.org/show_bug.cgi?id=159226

Reviewed by Keith Miller.

This is a bug if the caller is in sloppy mode and the callee is in strict
mode. This can't happen with ES6 classes because they're all in strict mode,
but it can happen with method syntax on an object literal. The caller must
to_this on |this| when it knows that it performs super property accesses.

* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::BytecodeGenerator):
* tests/stress/super-property-access-object-literal-to-this-2.js: Added.
(assert):
(test):
(let.o1.get foo):
(let.o2.a):
(let.o2.aa):
* tests/stress/super-property-access-object-literal-to-this.js: Added.
(assert):
(test):
(let.o1.get foo):
(let.o2.a):
(let.o2.aa):
(let.o2.b):
(let.o2.bb):
* tests/stress/super-property-access-to-this.js: Added.
(assert):
(test):
(Base.prototype.get foo):
(Base):
(Child.prototype.a):
(Child.prototype.b):
(Child):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202710 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoTab pointed out a typo that made the overdamped case fail.
dino@apple.com [Thu, 30 Jun 2016 21:50:40 +0000 (21:50 +0000)]
Tab pointed out a typo that made the overdamped case fail.

* demos/spring/spring.js:
(SpringSolver.prototype.solve):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202709 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agogetUserMedia() exposed, but not functional
eric.carlson@apple.com [Thu, 30 Jun 2016 21:20:27 +0000 (21:20 +0000)]
getUserMedia() exposed, but not functional
https://bugs.webkit.org/show_bug.cgi?id=158393
<rdar://problem/26642259>

Reviewed by Dean Jackson.

Set default value of the Media Stream runtime flag to false on Mac OS X and iOS until the
browser support is in place.

* bindings/generic/RuntimeEnabledFeatures.cpp:
(WebCore::RuntimeEnabledFeatures::RuntimeEnabledFeatures): Disable media stream by default
on Mac OS X and iOS.
* bindings/generic/RuntimeEnabledFeatures.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202704 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWeb Inspector: Add "spring" to transition-timing-function value autocompletion
nvasilyev@apple.com [Thu, 30 Jun 2016 21:19:04 +0000 (21:19 +0000)]
Web Inspector: Add "spring" to transition-timing-function value autocompletion
https://bugs.webkit.org/show_bug.cgi?id=159313
<rdar://problem/27114269>

Reviewed by Timothy Hatcher.

* UserInterface/Models/CSSKeywordCompletions.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202702 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoREGRESSION(r202329): WebInspectorProxy's WKPagePolicyClient callbacks are not being...
bburg@apple.com [Thu, 30 Jun 2016 21:17:05 +0000 (21:17 +0000)]
REGRESSION(r202329): WebInspectorProxy's WKPagePolicyClient callbacks are not being called
https://bugs.webkit.org/show_bug.cgi?id=159308
<rdar://problem/27111764>

Reviewed by Tim Horton.

Adopt WKPageNavigationClient and ditch WKPolicyClient and WKPageLoaderClient.
After r202329, it's not possible to set a WKPolicyClient for a WKPage/WebPageProxy that backs a WKWebView.

* UIProcess/WebInspectorProxy.cpp:
(WebKit::isMainOrTestInspectorPage): Take a WebCore::ResourceRequest.
(WebKit::processDidCrash): Deleted.
(WebKit::webProcessDidCrash): Rename to match new client callback name.
(WebKit::decidePolicyForNavigationAction): Rework this to take new argument types.
(WebKit::WebInspectorProxy::eagerlyCreateInspectorPage):
Install a WKPageNavigationClient.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202700 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r202676.
ryanhaddad@apple.com [Thu, 30 Jun 2016 20:44:42 +0000 (20:44 +0000)]
Unreviewed, rolling out r202676.
https://bugs.webkit.org/show_bug.cgi?id=159314

This change caused storage/websql tests to crash on Mac and
iOS WK1 (Requested by ryanhaddad on #webkit).

Reverted changeset:

"Purge PassRefPtr in Modules/webdatabase"
https://bugs.webkit.org/show_bug.cgi?id=159255
http://trac.webkit.org/changeset/202676

Patch by Commit Queue <commit-queue@webkit.org> on 2016-06-30

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202698 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoFixed a typo in the link to the other version.
dino@apple.com [Thu, 30 Jun 2016 20:29:26 +0000 (20:29 +0000)]
Fixed a typo in the link to the other version.

* blog-files/color-gamut/index.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202697 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] Process suspension is prevented for 30 seconds after closing a tab
cdumez@apple.com [Thu, 30 Jun 2016 20:17:30 +0000 (20:17 +0000)]
[iOS] Process suspension is prevented for 30 seconds after closing a tab
https://bugs.webkit.org/show_bug.cgi?id=159280
<rdar://problem/27014867>

Reviewed by Gavin Barraclough.

When closing a tab, ConnectionTerminationWatchdog takes a background
assertion on behalf of the WebContent process and only releases it
after 30 seconds, allowing for the WebContent process to exit cleanly
without worrying about getting suspended too early.

However, the child process normally exits much sooner than this and
we end up holding a process assertion for no reason for a full 30
seconds anyway. This patch addresses the issue by registering an
invalidation handler with the BKSProcessAssertion and releasing our
our assertion in such case. The invalidation handler gets called
as soon as the child process exits.

* Platform/spi/ios/AssertionServicesSPI.h:
* UIProcess/ProcessAssertion.cpp:
(WebKit::ProcessAssertion::ProcessAssertion):
* UIProcess/ProcessAssertion.h:
(WebKit::ProcessAssertion::ProcessAssertion):
(WebKit::ProcessAssertion::validity):
* UIProcess/ios/ProcessAssertionIOS.mm:
(WebKit::ProcessAssertion::ProcessAssertion):
(WebKit::ProcessAssertion::~ProcessAssertion):
(WebKit::ProcessAndUIAssertion::updateRunInBackgroundCount):
(WebKit::ProcessAndUIAssertion::ProcessAndUIAssertion):
(WebKit::ProcessAssertion::setState): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202696 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoKeep track of when a WKWebView is blank before the initial non-empty layout
timothy_horton@apple.com [Thu, 30 Jun 2016 19:15:15 +0000 (19:15 +0000)]
Keep track of when a WKWebView is blank before the initial non-empty layout
https://bugs.webkit.org/show_bug.cgi?id=159217
<rdar://problem/26071766>

Reviewed by Beth Dakin.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]):
(-[WKWebView _setIsBlankBeforeFirstNonEmptyLayout:]):
(-[WKWebView _didFirstVisuallyNonEmptyLayoutForMainFrame]):
* UIProcess/API/Cocoa/WKWebViewInternal.h:
* UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::didFirstVisuallyNonEmptyLayoutForMainFrame):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202695 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[iOS] Media controls are too cramped with small video
commit-queue@webkit.org [Thu, 30 Jun 2016 19:10:36 +0000 (19:10 +0000)]
[iOS] Media controls are too cramped with small video
https://bugs.webkit.org/show_bug.cgi?id=158815
<rdar://problem/26824238>

Patch by Antoine Quint <graouts@apple.com> on 2016-06-30
Reviewed by Eric Carlson.

Source/WebCore:

In updateLayoutForDisplayedWidth(), we try to ensure a minimum width is guaranteed
for the progress indicator. However, we were not accounting for the width used by
the current and remaining time labels on either side of it, so we would incorrectly
conclude that we were guaranteeing the minimum time and yield incorrect layouts since
we were trying to fit more buttons than we had room for.

In order to correctly compute the available width for the progress indicator, we now
have clones of the current and remaining time labels, hidden from video and VoiceOver,
that we update along with the originals. The same styles apply to both clones and
originals, so we may measure the clones to determine the space used by the time labels.
The reason we need to use clones is that if the time labels had previously been hidden
from view, precisely because there was not enough space to display them along with the
progress indicator, then trying to obtain metrics from them would yield 0 since they had
"display: none" styles applied. In order to avoid extra layouts and possible flashing, we
use the clones so that we never have to toggle the "display" property of the originals
just to obtain their measurements.

As a result of this change, we adjust the constant used to set the minimum required
width available to display the progress indicator after all other essential controls
and labels have been measured. That constant used to account for the width of the
time labels, and this is no longer correct.

Test: media/video-controls-drop-and-restore-timeline.html

* Modules/mediacontrols/mediaControlsApple.css:
(::-webkit-media-controls-time-remaining-display.clone):
* Modules/mediacontrols/mediaControlsApple.js:
(Controller):
(Controller.prototype.createTimeClones):
(Controller.prototype.removeTimeClass):
(Controller.prototype.addTimeClass):
(Controller.prototype.updateDuration):
(Controller.prototype.updateLayoutForDisplayedWidth):
(Controller.prototype.updateTime):
(Controller.prototype.updateControlsWhileScrubbing):
* Modules/mediacontrols/mediaControlsiOS.css:
(::-webkit-media-controls-time-remaining-display.clone):
* Modules/mediacontrols/mediaControlsiOS.js:

LayoutTests:

Adjust the output of a couple of tests to account for the time label clones, ensure the video
is wide enough to always have its timeline visible for tests that rely on the timeline being
visible to drag and seek, and finally add a new test.

* media/controls-drag-timebar.html:
* media/media-controls-drag-timeline-set-controls-property.html:
* media/video-controls-drop-and-restore-timeline-expected.txt: Added.
* media/video-controls-drop-and-restore-timeline.html: Added.
* platform/mac-yosemite/http/tests/media/hls/video-controls-live-stream-expected.txt:
* platform/mac/http/tests/media/hls/video-controls-live-stream-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202694 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoWe need to to_this when an inner arrow function uses 'this'
sbarati@apple.com [Thu, 30 Jun 2016 18:46:23 +0000 (18:46 +0000)]
We need to to_this when an inner arrow function uses 'this'
https://bugs.webkit.org/show_bug.cgi?id=159290
<rdar://problem/27058322>

Reviewed by Geoffrey Garen.

We put the |this| value into the closure object when there
is an arrow function that uses |this|. However, an arrow function
using |this| wasn't causing the creator of the closure that
holds |this| to to_this its value before putting it in the
closure. That's a huge bug because it means some arrow functions
can capture the raw |this| value, which might be a JSLexicalEnvironment.
This patch fixes this by adding an easy to check to see if any
inner arrow functions use |this|, and if any do, it will to_this
the |this| value.

* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::BytecodeGenerator):
* tests/stress/to-this-before-arrow-function-closes-over-this-that-starts-as-lexical-environment.js: Added.
(assert):
(obj):
(foo.capture):
(foo.wrapper.let.x.):
(foo2.capture):
(foo2.wrapper.let.x.):
(foo2.wrapper.bar):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202693 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoRemoving duplicated line from Mac TestExpectations file.
ryanhaddad@apple.com [Thu, 30 Jun 2016 18:19:41 +0000 (18:19 +0000)]
Removing duplicated line from Mac TestExpectations file.

Unreviewed test gardening.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202692 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoCleanup ProcessAssertion RunInBackground management
barraclough@apple.com [Thu, 30 Jun 2016 18:15:47 +0000 (18:15 +0000)]
Cleanup ProcessAssertion RunInBackground management
https://bugs.webkit.org/show_bug.cgi?id=159284

Reviewed by Chris Dumez.

Factor out code that handles taking a background assertion on the UI app into
ProcessAndUIAssertion::updateRunInBackgroundCount helper.

This patch just cleanup to enable a futher fix.

* UIProcess/ProcessAssertion.h:
(WebKit::ProcessAssertion::state):
    - Added ProcessAndUIAssertion::updateRunInBackgroundCount
    - Added ProcessAndUIAssertion::m_isHoldingBackgroundAssertionOnApp
* UIProcess/ios/ProcessAssertionIOS.mm:
(WebKit::ProcessAndUIAssertion::updateRunInBackgroundCount):
    - helper to manage calls to incrementNeedsToRunInBackgroundCount/decrementNeedsToRunInBackgroundCount
(WebKit::ProcessAndUIAssertion::ProcessAndUIAssertion):
    - call updateRunInBackgroundCount
(WebKit::ProcessAndUIAssertion::~ProcessAndUIAssertion):
    changed to use m_isHoldingBackgroundAssertionOnApp
(WebKit::ProcessAndUIAssertion::setState):
    - call updateRunInBackgroundCount

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202690 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoGenerators violate bytecode liveness validation
fpizlo@apple.com [Thu, 30 Jun 2016 18:13:26 +0000 (18:13 +0000)]
Generators violate bytecode liveness validation
https://bugs.webkit.org/show_bug.cgi?id=159279

Reviewed by Yusuke Suzuki.
PerformanceTests:

Add Basic to our test harness.

Also made some cosmetic changes to the benchmark harness.

* ES6SampleBench/Basic/basic-tests.yaml: Added.
* ES6SampleBench/Basic/stress-test.js: Added.
(preciseTime):
* ES6SampleBench/driver.js:
(Driver):
(Driver.prototype.start):
(Driver.prototype.reportError):
* ES6SampleBench/glue.js:
* ES6SampleBench/index.html:

Source/JavaScriptCore:

Fix a liveness bug found by Basic. The problem is that resume's intended liveness rule is:
"live-in is just the token argument", but the liveness analysis thought that the rule was
"live-in is live-out minus defs plus live-at-catch". Clearly these two rules are quite
different. The way this sort of worked before is that we would define the defs of resume
as being equal to our prediction of what the live-outs would be. We did this in the hope
that we would subtract all live-outs. But, this misses the live-at-catch part. So, this
change adds another hack to neutralize live-at-catch.

This would make a lot more sense if we wrote a new liveness analysis that was just for
generator conversion. It could reuse BytecodeUseDef but otherwise it would be a new thing.
It would be easy to write crazy rules for save/resume in such an analysis, especially if
that analysis rewrote the bytecode. We could then just have an op_yield that is a no-op.
We would just record the live-outs of op_yield and use that for rewriting the code in terms
of a switch statement.

* bytecode/BytecodeLivenessAnalysis.cpp:
(JSC::stepOverInstruction):
(JSC::BytecodeLivenessAnalysis::dumpResults):
* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpBytecode):

Tools:

Add Basic to our test harness.

* Scripts/run-javascriptcore-tests:
(runJSCStressTests):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202689 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, fix the macOS Sierra Release configuration after r202642.
bburg@apple.com [Thu, 30 Jun 2016 18:06:12 +0000 (18:06 +0000)]
Unreviewed, fix the macOS Sierra Release configuration after r202642.

* platform/audio/mac/MediaSessionManagerMac.mm:
(WebCore::MediaSessionManagerMac::updateNowPlayingInfo):
Add missing UNUSED_PARAM for when logging is not enabled.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202688 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMarking fast/shrink-wrap/rect-shrink-wrap.html as flaky on ios-simulator
ryanhaddad@apple.com [Thu, 30 Jun 2016 17:49:27 +0000 (17:49 +0000)]
Marking fast/shrink-wrap/rect-shrink-wrap.html as flaky on ios-simulator
https://bugs.webkit.org/show_bug.cgi?id=159275

Unreviewed test gardening.

* platform/ios-simulator-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202687 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[Win][Debug] Assertion fails in TestWTF.
pvollan@apple.com [Thu, 30 Jun 2016 17:48:50 +0000 (17:48 +0000)]
[Win][Debug] Assertion fails in TestWTF.
https://bugs.webkit.org/show_bug.cgi?id=159299

Reviewed by Alex Christensen.

The assert in the & operator in the UniqueRef class fails, because the instance has been moved.
The Windows version of the UNUSED_PARAM macro is getting the address of the parameter, causing
the assertion failure. Since the variable is actually used, there is no need to use the
UNUSED_PARAM macro.

* TestWebKitAPI/Tests/WTF/UniqueRef.cpp:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202686 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoMail crashes on launch at WebCore::PaymentCoordinator::~PaymentCoordinator() + 20
andersca@apple.com [Thu, 30 Jun 2016 17:47:20 +0000 (17:47 +0000)]
Mail crashes on launch at WebCore::PaymentCoordinator::~PaymentCoordinator() + 20
https://bugs.webkit.org/show_bug.cgi?id=159306
rdar://problem/27092107

Reviewed by Tim Horton.

* WebView/WebView.mm:
(-[WebView initSimpleHTMLDocumentWithStyle:frame:preferences:groupName:]):
Create a WebPaymentCoordinatorClient. Remove an unneeded #import.

(-[WebView _commonInitializationWithFrameName:groupName:]):
Remove an unneeded #import.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202685 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r202659.
ryanhaddad@apple.com [Thu, 30 Jun 2016 17:31:23 +0000 (17:31 +0000)]
Unreviewed, rolling out r202659.
https://bugs.webkit.org/show_bug.cgi?id=159305

The test for this change times out on mac-wk2 debug and caused
an existing test to crash. (Requested by ryanhaddad on
#webkit).

Reverted changeset:

"Web Inspector: Wrong function name next to scope"
https://bugs.webkit.org/show_bug.cgi?id=158210
http://trac.webkit.org/changeset/202659

Patch by Commit Queue <commit-queue@webkit.org> on 2016-06-30

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202684 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[JSC] Date.setYear() misses timeClip()
commit-queue@webkit.org [Thu, 30 Jun 2016 17:16:39 +0000 (17:16 +0000)]
[JSC] Date.setYear() misses timeClip()
https://bugs.webkit.org/show_bug.cgi?id=159289

Patch by Benjamin Poulain <bpoulain@apple.com> on 2016-06-30
Reviewed by Geoffrey Garen.

Source/JavaScriptCore:

* runtime/DatePrototype.cpp:
(JSC::dateProtoFuncSetYear):

LayoutTests:

* js/date-timeClip-large-values-expected.txt:
* js/script-tests/date-timeClip-large-values.js:
The failures are caused by bugs related to GregorianDateTime.
The last test case is the one covering this patch.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202683 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoUnreviewed, rolling out r202679.
commit-queue@webkit.org [Thu, 30 Jun 2016 16:19:32 +0000 (16:19 +0000)]
Unreviewed, rolling out r202679.
https://bugs.webkit.org/show_bug.cgi?id=159302

Still causing timeouts on media/controls-drag-timebar.html
(Requested by ap on #webkit).

Reverted changeset:

"[iOS] Media controls are too cramped with small video"
https://bugs.webkit.org/show_bug.cgi?id=158815
http://trac.webkit.org/changeset/202679

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202682 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years agoLayoutTests intermittently failing to run due to issues starting Web Platform Test...
commit-queue@webkit.org [Thu, 30 Jun 2016 16:14:44 +0000 (16:14 +0000)]
LayoutTests intermittently failing to run due to issues starting Web Platform Test server
https://bugs.webkit.org/show_bug.cgi?id=159166
<rdar://problem/27051759>

Patch by Youenn Fablet <youennf@gmail.com> on 2016-06-30
Reviewed by Alexey Proskuryakov.

LayoutTests/imported/w3c:

Pregenerating certificates.

* resources/_wpt_certs/.gitignore: Added.
* resources/_wpt_certs/localhost.key: Added.
* resources/_wpt_certs/localhost.pem: Added.
* resources/config.json:

Tools:

Setting the certificate directory to the copied config.json.

* Scripts/webkitpy/layout_tests/servers/web_platform_test_server.py:
(WebPlatformTestServer._copy_webkit_test_files):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202681 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 years ago[JSC] Implement isFinite / isNaN in JS and make DFG ToNumber accept non number values
utatane.tea@gmail.com [Thu, 30 Jun 2016 15:26:47 +0000 (15:26 +0000)]
[JSC] Implement isFinite / isNaN in JS and make DFG ToNumber accept non number values
https://bugs.webkit.org/show_bug.cgi?id=154022

Reviewed by Filip Pizlo.

Source/JavaScriptCore:

We aim at optimizing @toInteger operation.
While it still has an unoptimized part[1], this patch should be a first step.

We introduce the @toNumber builtin intrinsic operation.
This converts the given value to the JS number by emitting op_to_number bytecode.
Previously @toInteger called C++ @Number constructor for that purpose.

And in DFG, op_to_number is converted to DFG ToNumber node.
During DFG, we attempt to convert this to edge filtering and Identity, but if we fail,
we just fall back to calling the C++ function.

To utilize ToNumber in user-land side, we add a path attempting to convert Number constructor calls
to ToNumber DFG nodes. This conversion is useful because `Number(value)` is used to convert a value to a number in JS.

Before this patch, we emit simple edge filtering (NumberUse) instead of emitting DFG node like ToNumber for op_to_number.
But emitting ToNumber is useful, because in the case of `Number(value)`, considering `value` may not be a number is reasonable.

By leveraging @toNumber operation, we rewrite Number.{isFinite, isNaN}, global.{isFinite, isNaN} and @toInteger.

ToNumber DFG node has a value profiling. This profiling is leveraged to determine the result number type of the ToNumber operation.
This value profiling is provided from either NumberConstructor's call operation or op_to_number.

The results (with the added performance tests) show that, while existing cases are performance neutral, the newly added cases gain the performance benefit.
And ASMBench/n-body.c also shows stable ~2% progression.

[1]: https://bugs.webkit.org/show_bug.cgi?id=153738

* CMakeLists.txt:
* DerivedSources.make:
* JavaScriptCore.xcodeproj/project.pbxproj:
* builtins/BuiltinNames.h:
* builtins/GlobalObject.js:
(globalPrivate.isFinite):
(globalPrivate.isNaN):
(globalPrivate.toInteger): Deleted.
(globalPrivate.toLength): Deleted.
(globalPrivate.isDictionary): Deleted.
(globalPrivate.speciesGetter): Deleted.
(globalPrivate.speciesConstructor): Deleted.
* builtins/GlobalOperations.js: Copied from Source/JavaScriptCore/builtins/GlobalObject.js.
(globalPrivate.toInteger):
(globalPrivate.toLength):
(globalPrivate.isDictionary):
(globalPrivate.speciesGetter):
(globalPrivate.speciesConstructor):
* builtins/NumberConstructor.js: Added.
(isFinite):
(isNaN):
* bytecode/BytecodeIntrinsicRegistry.cpp:
(JSC::BytecodeIntrinsicRegistry::BytecodeIntrinsicRegistry):
* bytecode/BytecodeIntrinsicRegistry.h:
* bytecode/BytecodeList.json:
* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpBytecode):
(JSC::CodeBlock::finishCreation):
* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::emitUnaryOp):
(JSC::BytecodeGenerator::emitUnaryOpProfiled):
* bytecompiler/BytecodeGenerator.h:
(JSC::BytecodeGenerator::emitToNumber):
* bytecompiler/NodesCodegen.cpp:
(JSC::BytecodeIntrinsicNode::emit_intrinsic_toNumber):
(JSC::UnaryPlusNode::emitBytecode):
* dfg/DFGAbstractInterpreterInlines.h:
(JSC::DFG::AbstractInterpreter<AbstractStateType>::executeEffects):
* dfg/DFGBackwardsPropagationPhase.cpp:
(JSC::DFG::BackwardsPropagationPhase::propagate):
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::attemptToInlineCall):
(JSC::DFG::ByteCodeParser::handleConstantInternalFunction):
(JSC::DFG::ByteCodeParser::parseBlock):
We use `getPrediction()` to retrieve the heap prediction from the to_number bytecode.
According to the benchmark results, choosing `getPredictionWithoutOSRExit()` causes performance regression (1.5%) in kraken stanford-crypto-aes.

* dfg/DFGClobberize.h:
(JSC::DFG::clobberize):
* dfg/DFGConstantFoldingPhase.cpp:
(JSC::DFG::ConstantFoldingPhase::foldConstants):
* dfg/DFGDoesGC.cpp:
(JSC::DFG::doesGC):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
(JSC::DFG::FixupPhase::fixupToNumber):
* dfg/DFGNode.h:
(JSC::DFG::Node::hasHeapPrediction):
* dfg/DFGNodeType.h:
* dfg/DFGOperations.cpp:
* dfg/DFGOperations.h:
* dfg/DFGPredictionPropagationPhase.cpp:
Always on the heap prediction.

* dfg/DFGSafeToExecute.h:
(JSC::DFG::safeToExecute):
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
As of 64bit version, we carefully manage the register reuse. The largest difference between 32bit and 64bit is
`branchIfNotNumber()` requires the temporary register. We should not use the result registers for that since
it may be reuse the argument registers and it can break the argument registers before using them to call the operation.
Currently, we allocate the additional temporary register for that scratch register.

* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
Reuse the argument register for the result if possible. And manually decrement the use count in the middle of the node.
This is similar technique used in ToPrimitive. Typically, the child of ToNumber is only used by this ToNumber node since
we would like to perform the type conversion onto this child node here. So this careful register reuse effectively removes
the spills to call the operation. The example of the actually emitted code is the following.

76:<!2:loc11>     ToNumber(Untyped:@68, JS|MustGen|UseAsOther, DoubleimpurenanTopEmpty, R:World, W:Heap, Exits, ClobbersExit, bc#48)  predicting DoubleimpurenanTopEmpty
    0x7f986d5fe693: test %rax, %r14
    0x7f986d5fe696: jz 0x7f986d5fe6a1
    0x7f986d5fe69c: jmp 0x7f986d5fe6d1
    0x7f986d5fe6a1: mov %rax, %rsi
    0x7f986d5fe6a4: mov %rbp, %rdi
    0x7f986d5fe6a7: mov $0x2, 0x24(%rbp)
    0x7f986d5fe6ae: mov $0x7f98711ea5f0, %r11
    0x7f986d5fe6b8: call *%r11
    0x7f986d5fe6bb: mov $0x7f982d3f72d0, %r11
    0x7f986d5fe6c5: mov (%r11), %r11
    0x7f986d5fe6c8: test %r11, %r11
    0x7f986d5fe6cb: jnz 0x7f986d5fe88c

It effectively removes the unnecessary spill to call the operation!

* ftl/FTLCapabilities.cpp:
(JSC::FTL::canCompile):
* ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::LowerDFGToB3::compileNode):
(JSC::FTL::DFG::LowerDFGToB3::compileToNumber):
(JSC::FTL::DFG::LowerDFGToB3::compileCompareStrictEq):
* jit/AssemblyHelpers.h:
(JSC::AssemblyHelpers::branchIfNumber):
(JSC::AssemblyHelpers::branchIfNotNumber):
* jit/JITOpcodes.cpp:
(JSC::JIT::emit_op_to_number):
* jit/JITOpcodes32_64.cpp:
(JSC::JIT::emit_op_to_number):
* llint/LowLevelInterpreter32_64.asm:
* llint/LowLevelInterpreter64.asm:
* parser/Nodes.h:
(JSC::UnaryOpNode::opcodeID):
* runtime/CommonSlowPaths.cpp:
(JSC::SLOW_PATH_DECL):
* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
* runtime/JSGlobalObjectFunctions.cpp:
(JSC::globalFuncIsNaN): Deleted.
(JSC::globalFuncIsFinite): Deleted.
* runtime/JSGlobalObjectFunctions.h:
* runtime/MathCommon.h:
(JSC::maxSafeInteger):
(JSC::minSafeInteger):
* runtime/NumberConstructor.cpp:
(JSC::NumberConstructor::finishCreation):
(JSC::numberConstructorFuncIsFinite): Deleted.
(JSC::numberConstructorFuncIsNaN): Deleted.
* runtime/NumberConstructor.h:
* tests/stress/Number-isNaN-basics.js: Added.
(numberIsNaNOnInteger):
(testNumberIsNaNOnIntegers):
(verifyNumberIsNaNOnIntegerWithOtherTypes):
(numberIsNaNOnDouble):
(testNumberIsNaNOnDoubles):
(verifyNumberIsNaNOnDoublesWithOtherTypes):
(numberIsNaNNoArguments):
(numberIsNaNTooManyArguments):
(testNumberIsNaNOnConstants):
(numberIsNaNStructTransition):
(Number.isNaN):
* tests/stress/global-is-finite.js: Added.
(shouldBe):
* tests/stress/global-is-nan.js: Added.
(shouldBe):
* tests/stress/global-isNaN-basics.js: Added.
(isNaNOnInteger):
(testIsNaNOnIntegers):
(verifyIsNaNOnIntegerWithOtherTypes):
(isNaNOnDouble):
(testIsNaNOnDoubles):
(verifyIsNaNOnDoublesWithOtherTypes):
(verifyIsNaNOnCoercedTypes):
(isNaNNoArguments):
(isNaNTooManyArguments):
(testIsNaNOnConstants):
(isNaNTypeCoercionSideEffects):
(i.value.isNaNTypeCoercionSideEffects.valueOf):
(isNaNStructTransition):
(isNaN):
* tests/stress/number-is-finite.js: Added.
(shouldBe):
(test2):
(test3):
* tests/stress/number-is-nan.js: Added.
(shouldBe):
(test2):
(test3):
* tests/stress/to-number-basics.js: Added.
(shouldBe):
* tests/stress/to-number-convert-identity-without-execution.js: Added.
(shouldBe):
(object.valueOf):
(valueOf):
* tests/stress/to-number-int52.js: Added.
(shouldBe):
(object.valueOf):
* tests/stress/to-number-intrinsic-convert-to-identity-without-execution.js: Added.
(shouldBe):
(object.valueOf):
(valueOf):
* tests/stress/to-number-intrinsic-int52.js: Added.
(shouldBe):
(object.valueOf):
* tests/stress/to-number-intrinsic-object-without-execution.js: Added.
(shouldBe):
(object.valueOf):
* tests/stress/to-number-intrinsic-value-profiling.js: Added.
(shouldBe):
(object.valueOf):
* tests/stress/to-number-object-without-execution.js: Added.
(shouldBe):
(object.valueOf):
* tests/stress/to-number-object.js: Added.
(shouldBe):
(test12):
(object1.valueOf):
(test2):
(test22):
(object2.valueOf):
(test3):
(test32):
(object3.valueOf):
* tests/stress/to-number-value-profiling.js: Added.
(shouldBe):
(object.valueOf):

LayoutTests:

* js/regress/Number-isNaN-expected.txt: Added.
* js/regress/Number-isNaN.html: Added.
* js/regress/global-isNaN-expected.txt: Added.
* js/regress/global-isNaN.html: Added.
* js/regress/script-tests/Number-isNaN.js: Added.
* js/regress/script-tests/global-isNaN.js: Added.
* js/regress/script-tests/many-foreach-calls.js:
(i.4.forEach):
(i.array.forEach): Deleted.
* js/regress/script-tests/to-number-constructor-number-string-number-string.js: Added.
(test):
* js/regress/script-tests/to-number-constructor-only-number.js: Added.
(test):
* js/regress/script-tests/to-number-constructor-only-string.js: Added.
(test):
* js/regress/script-tests/to-number-constructor-string-number-string-number.js: Added.
(test):
* js/regress/script-tests/to-number-number-string-number-string.js: Added.
(test):
* js/regress/script-tests/to-number-only-number.js: Added.
(test):
* js/regress/script-tests/to-number-only-string.js: Added.
(test):
* js/regress/script-tests/to-number-string-number-string-number.js: Added.
(test):
* js/regress/to-number-constructor-number-string-number-string-expected.txt: Added.
* js/regress/to-number-constructor-number-string-number-string.html: Added.
* js/regress/to-number-constructor-only-number-expected.txt: Added.
* js/regress/to-number-constructor-only-number.html: Added.
* js/regress/to-number-constructor-only-string-expected.txt: Added.
* js/regress/to-number-constructor-only-string.html: Added.
* js/regress/to-number-constructor-string-number-string-number-expected.txt: Added.
* js/regress/to-number-constructor-string-number-string-number.html: Added.
* js/regress/to-number-number-string-number-string-expected.txt: Added.
* js/regress/to-number-number-string-number-string.html: Added.
* js/regress/to-number-only-number-expected.txt: Added.
* js/regress/to-number-only-number.html: Added.
* js/regress/to-number-only-string-expected.txt: Added.
* js/regress/to-number-only-string.html: Added.
* js/regress/to-number-string-number-string-number-expected.txt: Added.
* js/regress/to-number-string-number-string-number.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202680 268f45cc-cd09-0410-ab3c-d52691b4dbfc