WebKit-https.git
3 years ago[Threaded Compositor] Modernize and simplify threaded compositor code
carlosgc@webkit.org [Tue, 14 Jun 2016 07:52:25 +0000 (07:52 +0000)]
[Threaded Compositor] Modernize and simplify threaded compositor code
https://bugs.webkit.org/show_bug.cgi?id=158615

Reviewed by Žan Doberšek.

* Shared/CoordinatedGraphics/threadedcompositor/CompositingRunLoop.cpp:
(WebKit::CompositingRunLoop::performTask): Use NoncopyableFunction.
(WebKit::CompositingRunLoop::performTaskSync): Ditto.
(WebKit::CompositingRunLoop::startUpdateTimer): Just renamed to start instead of set.
(WebKit::CompositingRunLoop::run): Expose run/stop methods instead of the internal RunLoop object.
(WebKit::CompositingRunLoop::stop): Also stop the update timer instead of relying on the caller to do it.
(WebKit::CompositingRunLoop::setUpdateTimer): Deleted.
(WebKit::CompositingRunLoop::stopUpdateTimer): Deleted.
* Shared/CoordinatedGraphics/threadedcompositor/CompositingRunLoop.h:
* Shared/CoordinatedGraphics/threadedcompositor/ThreadedCompositor.cpp:
(WebKit::ThreadedCompositor::setNativeSurfaceHandleForCompositing): Protects this directly in lambda capture.
(WebKit::ThreadedCompositor::setDeviceScaleFactor): Ditto.
(WebKit::ThreadedCompositor::didChangeViewportSize): Ditto.
(WebKit::ThreadedCompositor::didChangeViewportAttribute): Ditto.
(WebKit::ThreadedCompositor::didChangeContentsSize): Ditto.
(WebKit::ThreadedCompositor::scrollTo): Ditto.
(WebKit::ThreadedCompositor::scrollBy): Ditto.
(WebKit::ThreadedCompositor::updateViewport): Use startUpdateTimer().
(WebKit::ThreadedCompositor::scheduleDisplayImmediately): Ditto.
(WebKit::ThreadedCompositor::didChangeVisibleRect): Improve lambda captures.
(WebKit::ThreadedCompositor::renderLayerTree): Use m_viewportController directly.
(WebKit::ThreadedCompositor::createCompositingThread): Use createThread() version that receives a function.
(WebKit::ThreadedCompositor::runCompositingThread): Use run method and don't stop the update timer when the run
loop finishes.
(WebKit::ThreadedCompositor::terminateCompositingThread): Use stop method.
(WebKit::ThreadedCompositor::ThreadedCompositor): Deleted.
(WebKit::ThreadedCompositor::compositingThreadEntry): Deleted.
* Shared/CoordinatedGraphics/threadedcompositor/ThreadedCompositor.h:
(WebKit::ThreadedCompositor::viewportController): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202038 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years ago[Threaded Compositor] Flickering and rendering artifacts when resizing the web view
carlosgc@webkit.org [Tue, 14 Jun 2016 07:48:54 +0000 (07:48 +0000)]
[Threaded Compositor] Flickering and rendering artifacts when resizing the web view
https://bugs.webkit.org/show_bug.cgi?id=154070

Reviewed by Žan Doberšek.

Resizing the web view is expected to be a sync operation, the UI process creates a new backing store state ID,
sends UpdateBackingStoreState message with the flag RespondImmediately to the web process and waits up to 500ms
for the reply (DidUpdateBackingStoreState message). When using the threaded compositor, we schedule a task in
the compositing thread to update the viewport size, and return immediately, so that we reply to the UI process
before the compositing thread has actually updated its size. There's a moment in which sizes are out of sync
causing the flickering and rendering artifacts, the UI process continues rendering at the new size, while the
web process is still rendering at the previous size. We can prevent this from happening just by making the
resize task synchronous in the threaded compositor.

* Shared/CoordinatedGraphics/threadedcompositor/CompositingRunLoop.cpp:
(WebKit::CompositingRunLoop::performTaskSync): Add sync version of performTask().
* Shared/CoordinatedGraphics/threadedcompositor/CompositingRunLoop.h:
* Shared/CoordinatedGraphics/threadedcompositor/ThreadedCompositor.cpp:
(WebKit::ThreadedCompositor::didChangeViewportSize): Use performTaskSync().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202037 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUnreviewed. Fix Soup downloads after r201943.
carlosgc@webkit.org [Tue, 14 Jun 2016 07:43:24 +0000 (07:43 +0000)]
Unreviewed. Fix Soup downloads after r201943.

This is a follow up of r201943. The DownloadClient used in DownloadSoup was not updated to the new API of the
ResourceHandleClient because it was not using override on the virtual methods, so it was unnoticed. That broke
the downloads soup implementation, because didReceiveResponse is no longer used in the DownloadClient. This
patch updates the DownloadClient to the new ResourceHandleClient API adding also override to all the virtual
methods to prevent this from happening in the future.

* NetworkProcess/Downloads/soup/DownloadSoup.cpp:
(WebKit::Download::start):
(WebKit::Download::startWithHandle):
(WebKit::DownloadClient::DownloadClient):
(WebKit::DownloadClient::downloadFailed):
(WebKit::DownloadClient::didReceiveResponse):
(WebKit::DownloadClient::didReceiveData):
(WebKit::DownloadClient::didFinishLoading):
(WebKit::DownloadClient::didFail):
(WebKit::DownloadClient::wasBlocked): Deleted.
(WebKit::DownloadClient::cannotShowURL): Deleted.
(WebKit::DownloadClient::cancel):
(WebKit::DownloadClient::handleResponse):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202036 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoCrash inside firstPositionInNode in checkLoadCompleteForThisFrame
rniwa@webkit.org [Tue, 14 Jun 2016 07:35:53 +0000 (07:35 +0000)]
Crash inside firstPositionInNode in checkLoadCompleteForThisFrame
https://bugs.webkit.org/show_bug.cgi?id=158724

Reviewed by Alex Christensen.

Added null checks for document and document element since they could be nullptr here.

* loader/FrameLoader.cpp:
(WebCore::FrameLoader::checkLoadCompleteForThisFrame):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202035 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoModernize DumpRenderTreeMac.h
ddkilzer@apple.com [Tue, 14 Jun 2016 07:10:20 +0000 (07:10 +0000)]
Modernize DumpRenderTreeMac.h
<https://webkit.org/b/158721>

Reviewed by Andy Estes.

* DumpRenderTree/mac/DumpRenderTreeMac.h:
- Update copyright.
- Update license.
- Use #pragma once.
- Use OBJC_CLASS macro.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202034 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agosetUpStaticFunctionSlot does not handle Builtin|Accessor properties
barraclough@apple.com [Tue, 14 Jun 2016 06:57:15 +0000 (06:57 +0000)]
setUpStaticFunctionSlot does not handle Builtin|Accessor properties
https://bugs.webkit.org/show_bug.cgi?id=158637

Reviewed by Geoff Garen.

setUpStaticFunctionSlot contains a duplicate copy of the body of the function reifyStaticProperty
- however it is missing handling for Accessor type under Builtin functions.
Fix the bug by de-duplicating - setUpStaticFunctionSlot should just call reifyStaticProperty.

* runtime/Lookup.cpp:
(JSC::setUpStaticFunctionSlot):
    - should just call reifyStaticProperty.
* runtime/Lookup.h:
(JSC::lookupPut):
(JSC::reifyStaticProperty):
    - changed reifyStaticProperty to take PropertyName.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202033 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoRemove hasStaticPropertyTable (part 3: JSLocation::putDelegate)
barraclough@apple.com [Tue, 14 Jun 2016 06:36:30 +0000 (06:36 +0000)]
Remove hasStaticPropertyTable (part 3: JSLocation::putDelegate)
https://bugs.webkit.org/show_bug.cgi?id=158431

Unreviewed build fix.

* bindings/js/JSLocationCustom.cpp:
(WebCore::JSLocation::putDelegate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202032 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoRemove hasStaticPropertyTable (part 4: JSHTMLDocument & JSStorage)
barraclough@apple.com [Tue, 14 Jun 2016 06:18:32 +0000 (06:18 +0000)]
Remove hasStaticPropertyTable (part 4: JSHTMLDocument & JSStorage)
https://bugs.webkit.org/show_bug.cgi?id=158431

Reviewed by Chris Dumez.

All uses of hasStaticPropertyTable flag generated by bindings are wrong.

JSHTMLDocument & JSStorage contain a number of static_asserts claiming that
various methods do not support static properties. These asserts were likely
correct at the time they were added, as JSObject::getOwnPropertySlot and
JSObject::deleteProperty did not support getting / deleting static value.
This is no longer the case, and these asserts are now incorrect.

* bindings/js/JSHTMLDocumentCustom.cpp:
(WebCore::JSHTMLDocument::getOwnPropertySlot):
* bindings/js/JSStorageCustom.cpp:
(WebCore::JSStorage::deleteProperty):
(WebCore::JSStorage::deletePropertyByIndex):
(WebCore::JSStorage::putDelegate):
    - remove incorrect static_asserts.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202031 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoRemove hasStaticPropertyTable (part 3: JSLocation::putDelegate)
barraclough@apple.com [Tue, 14 Jun 2016 06:17:05 +0000 (06:17 +0000)]
Remove hasStaticPropertyTable (part 3: JSLocation::putDelegate)
https://bugs.webkit.org/show_bug.cgi?id=158431

Reviewed by Geoff Garen.

All uses of hasStaticPropertyTable flag generated by bindings are wrong.

JSLocation::putDelegate checks the static property table redundantly.

In the case of same origin access, if the property is not in the static
table the method will call JSObject::put and return true (indicating the
delegate handled the put). If the property is in the static table, the
method will return false (indicating the the delegate did not handle the
access) - in which case the calling function will call JSObject::put.
Checking for the property in the static table is redundant - same origin
access does not require any special handling, and should just always
return false & let the caller handle the put.

In the case of cross origin access, if the property is not in the static
table we return true (indicating the access was handled, and silently
blocking it). If it is a static property, we check the name, and if the
name is not 'href' we also return true, silently blocking. In the case
that the name is 'href' we'll return false, indicating to the caller
that the access was not handled by the delegate, resulting in it taking
place. The additional check of the static table is redundant, since we
only have special behaviour in the case of 'href'. (Moreover it is
unnecesszarily fragile, since if we made a change such that 'href' was no
longer implemented as a static property with would fail.)

- for same origin, always return false.
- for cross origin, return false for 'href', otherwise return true.

* bindings/js/JSLocationCustom.cpp:
(WebCore::JSLocation::putDelegate):
    - restructure & remove static table check.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202030 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoRemove hasStaticPropertyTable (part 2: JSPluginElement)
barraclough@apple.com [Tue, 14 Jun 2016 06:09:53 +0000 (06:09 +0000)]
Remove hasStaticPropertyTable (part 2: JSPluginElement)
https://bugs.webkit.org/show_bug.cgi?id=158431

Reviewed by Chris Dumez.

All uses of hasStaticPropertyTable flag generated by bindings are wrong.

The check in pluginElementCustomGetOwnPropertySlot was somewhat dubious in the
first place (for types with static properties it would give precedence to both
static and also property storage properties; for types without static properties
it would check neither - an odd asymetry in the case of values in the storage
array, and was depending on an implementation detail that could change).

This is all now redundant anyway. None of these types have static properties.
All properties are now corretcly on the prototype (which is handled appropriately
below). This is just dead code.

* bindings/js/JSPluginElementFunctions.h:
(WebCore::pluginElementCustomGetOwnPropertySlot):
    - remove dead code.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202029 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoRemove hasStaticPropertyTable (part 1: DOM bindings)
barraclough@apple.com [Tue, 14 Jun 2016 06:08:56 +0000 (06:08 +0000)]
Remove hasStaticPropertyTable (part 1: DOM bindings)
https://bugs.webkit.org/show_bug.cgi?id=158431

Reviewed by Chris Dumez.

All uses of hasStaticPropertyTable flag generated by bindings are wrong.

* bindings/js/JSDOMBinding.h:
(WebCore::getStaticValueSlotEntryWithoutCaching): Deleted.
(WebCore::getStaticValueSlotEntryWithoutCaching<JSDOMObject>): Deleted.
    - this method is not used anywhere.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202028 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoJSBoundSlotBaseFunction no longer binds slot base
barraclough@apple.com [Tue, 14 Jun 2016 06:04:32 +0000 (06:04 +0000)]
JSBoundSlotBaseFunction no longer binds slot base
https://bugs.webkit.org/show_bug.cgi?id=157978

Reviewed by Geoff Garen.

This class is basically currently named after a bug. We should never have
been binding function to slot bases - this was not ever correct behavior.
This was fixed earlier in the year, but there is still some cruft including
the class name to clean up.

    - renamed JSBoundSlotBaseFunction -> JSCustomGetterSetterFunction
    - removed m_boundSlotBase - don't retain the original slot base
      (we were not really using it anyway).
    - ASSERT customGetterSetter->getter/setter are non-null, rather than checking.
    - Store the PropertyName such that we can pass this to the getter
      (we're currently reperforming the String->Identifier conversion every time).
    - Removed JSFunction::lookUpOrCreateNativeExecutable - this is just overhead,
      and not used consistently.

* CMakeLists.txt:
* JavaScriptCore.xcodeproj/project.pbxproj:
* runtime/JSBoundSlotBaseFunction.cpp: Removed.
* runtime/JSBoundSlotBaseFunction.h: Removed.
    - JSBoundSlotBaseFunction -> JSCustomGetterSetterFunction
* runtime/JSCustomGetterSetterFunction.cpp: Copied from Source/JavaScriptCore/runtime/JSBoundSlotBaseFunction.cpp.
(JSC::JSCustomGetterSetterFunction::customGetterSetterFunctionCall):
    - made a static function on JSCustomGetterSetterFunction such that accessor
      to member properties could be made private. Call variant of callCustomSetter
      that does not require slotBase, ASSERT getter/setter present, pass stored
      PropertyName to getter.
(JSC::JSCustomGetterSetterFunction::JSCustomGetterSetterFunction):
    - renamed, store propertyName.
(JSC::JSCustomGetterSetterFunction::create):
    - use same function name to Executable as is being passed to Function::finishCreation.
(JSC::JSCustomGetterSetterFunction::visitChildren):
(JSC::JSCustomGetterSetterFunction::finishCreation):
    - removed m_boundSlotBase.
* runtime/JSCustomGetterSetterFunction.h: Copied from Source/JavaScriptCore/runtime/JSBoundSlotBaseFunction.h.
(JSC::JSCustomGetterSetterFunction::customGetterSetter):
(JSC::JSCustomGetterSetterFunction::isSetter):
    - made private.
(JSC::JSCustomGetterSetterFunction::propertyName):
    - new accessor.
(JSC::JSBoundSlotBaseFunction::boundSlotBase): Deleted.
    - removed.
* runtime/JSFunction.cpp:
(JSC::JSFunction::create):
(JSC::JSFunction::lookUpOrCreateNativeExecutable): Deleted.
    - removed lookUpOrCreateNativeExecutable. This inconsistently used wrapper was providing no value, only bloat.
* runtime/JSFunction.h:
* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init):
(JSC::JSGlobalObject::visitChildren):
    - renamed JSBoundSlotBaseFunction -> JSCustomGetterSetterFunction, etc.
* runtime/JSGlobalObject.h:
(JSC::JSGlobalObject::customGetterSetterFunctionStructure):
(JSC::JSGlobalObject::boundSlotBaseFunctionStructure): Deleted.
    - renamed JSBoundSlotBaseFunction -> JSCustomGetterSetterFunction, etc.
* runtime/JSNativeStdFunction.cpp:
(JSC::JSNativeStdFunction::create):
    - removed lookUpOrCreateNativeExecutable.
* runtime/JSObject.cpp:
(JSC::getCustomGetterSetterFunctionForGetterSetter):
(JSC::JSObject::getOwnPropertyDescriptor):
(JSC::getBoundSlotBaseFunctionForGetterSetter): Deleted.
    - renamed JSBoundSlotBaseFunction -> JSCustomGetterSetterFunction, etc.
* runtime/VM.h:
    - renamed JSBoundSlotBaseFunction -> JSCustomGetterSetterFunction, etc.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202027 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWebRTC: Imlement MediaEndpointPeerConnection::replaceTrack()
adam.bergkvist@ericsson.com [Tue, 14 Jun 2016 06:04:29 +0000 (06:04 +0000)]
WebRTC: Imlement MediaEndpointPeerConnection::replaceTrack()
https://bugs.webkit.org/show_bug.cgi?id=158688

Reviewed by Eric Carlson.

Source/WebCore:

Implement MediaEndpointPeerConnection::replaceTrack() that is the MediaEndpoint implementation
of RTCRtpSender.replaceTrack() [1].

[1] https://w3c.github.io/webrtc-pc/archives/20160513/webrtc.html#dom-rtcrtpsender-replacetrack

Updated fast/mediastream/RTCRtpSender-replaceTrack.html

* Modules/mediastream/MediaEndpointPeerConnection.cpp:
(WebCore::MediaEndpointPeerConnection::replaceTrack):
(WebCore::MediaEndpointPeerConnection::replaceTrackTask):
Implemented.
* Modules/mediastream/MediaEndpointPeerConnection.h:
* Modules/mediastream/PeerConnectionBackend.h:
* Modules/mediastream/RTCPeerConnection.cpp:
(WebCore::RTCPeerConnection::replaceTrack):
* Modules/mediastream/RTCPeerConnection.h:
Move the MediaStreamTrack instance of sending a reference to it. This change is the main
reason many files are touched by this change.
* Modules/mediastream/RTCRtpSender.h:
* Modules/mediastream/RTCRtpSender.idl:
* platform/mediastream/MediaEndpoint.h:
Use mid instead of mdescIndex to identify the media description in the backend.
* platform/mock/MockMediaEndpoint.cpp:
(WebCore::MockMediaEndpoint::replaceSendSource):
* platform/mock/MockMediaEndpoint.h:

LayoutTests:

Update existing test for RTCRtpSender.replaceTrack.

* fast/mediastream/RTCRtpSender-replaceTrack-expected.txt:
* fast/mediastream/RTCRtpSender-replaceTrack.html:
Add test cases where a "not yet negotiated track" is directly replaced. Also check that the
old track id is used in later offers.
* fast/mediastream/resources/promise-utils.js:
(promiseShouldResolve):
Added utility method to test promise expressions that are expected to resolve.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202026 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWeb Inspector: Show Exception Stack in UncaughtExceptionReporter view
commit-queue@webkit.org [Tue, 14 Jun 2016 04:33:06 +0000 (04:33 +0000)]
Web Inspector: Show Exception Stack in UncaughtExceptionReporter view
https://bugs.webkit.org/show_bug.cgi?id=158657
<rdar://problem/26754441>

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-06-13
Reviewed by Darin Adler.

* UserInterface/Debug/UncaughtExceptionReporter.css:
(.uncaught-exception-sheet li):
Make newlines significant.

* UserInterface/Debug/UncaughtExceptionReporter.js:
(unblockEventHandlers):
(handleError):
(handleLinkClick):
(formattedEntry):
Format the exception and a possible stack with a bit of sanitizing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202025 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years ago[Mac] Web Content service with a restricted entitlement may load arbitrary dylibs
mitz@apple.com [Tue, 14 Jun 2016 04:03:02 +0000 (04:03 +0000)]
[Mac] Web Content service with a restricted entitlement may load arbitrary dylibs
https://bugs.webkit.org/show_bug.cgi?id=156668
<rdar://problem/26714558>

Reviewed by Anders Carlsson.

* Configurations/WebContentService.xcconfig: Enable library validation when the Web Content
  service is given the XPC domain extension entitlement when targeting macOS Sierra or later.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202024 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agowindow.onerror should pass the ErrorEvent's 'error' property as the 5th argument...
commit-queue@webkit.org [Tue, 14 Jun 2016 03:28:10 +0000 (03:28 +0000)]
window.onerror should pass the ErrorEvent's 'error' property as the 5th argument to the event handler
https://bugs.webkit.org/show_bug.cgi?id=55092
<rdar://problem/25731279>

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-06-13
Reviewed by Dean Jackson.

LayoutTests/imported/w3c:

* web-platform-tests/html/dom/interfaces-expected.txt:

Source/WebCore:

This includes the actual Error in window.error / ErrorEvent:
https://html.spec.whatwg.org/multipage/webappapis.html#the-errorevent-interface

This is useful for scripts to be able to get an error stack
from uncaught exceptions, by checking the error itself.

Tests: fast/events/window-onerror17.html
       http/tests/security/cross-origin-script-error-event-redirected.html
       http/tests/security/cross-origin-script-error-event.html
       http/tests/security/script-crossorigin-error-event-information.html
       http/tests/security/script-no-crossorigin-error-event-should-be-sanitized.html
       userscripts/window-onerror-for-isolated-world-3.html

* CMakeLists.txt:
* WebCore.xcodeproj/project.pbxproj:
* bindings/js/JSBindingsAllInOne.cpp:
Add new custom error event file.

* bindings/js/JSDOMBinding.cpp:
(WebCore::reportException):
Include the JSC::Exception when reporting exceptions, so the error value is available.

* bindings/js/JSErrorEventCustom.cpp:
(WebCore::JSErrorEvent::error):
Sanitized access to the ErrorEvent's error property to prevent leaking objects
across isolated world boundaries. This is like CustomEvent's data property.

* bindings/js/JSErrorHandler.cpp:
(WebCore::JSErrorHandler::handleEvent):
* bindings/js/JSErrorHandler.h:
Include the error object as the 4th argument to the window.onerror event handler.

* dom/ScriptExecutionContext.cpp:
(WebCore::ScriptExecutionContext::sanitizeScriptError):
(WebCore::ScriptExecutionContext::reportException):
(WebCore::ScriptExecutionContext::dispatchErrorEvent):
* dom/ScriptExecutionContext.h:
Include the error object in the ErrorEvent constructed when dispatching error events.

* dom/ErrorEvent.cpp:
(WebCore::ErrorEvent::ErrorEvent):
(WebCore::ErrorEvent::sanitizedErrorValue):
(WebCore::ErrorEvent::trySerializeError):
* dom/ErrorEvent.h:
* dom/ErrorEvent.idl:
Include an any "error" property on the ErrorEvent, and allow it in initialization.

* bindings/js/WorkerScriptController.cpp:
(WebCore::WorkerScriptController::evaluate):
* workers/WorkerMessagingProxy.cpp:
(WebCore::WorkerMessagingProxy::postExceptionToWorkerObject):
Within the Worker world, the error is included in the event.
When re-dispatching the error on the world object in the world that spawned the
Worker the event does not include an error object. This matches other browsers
right now, but could be improved to have the same cross world serialization
as isolated worlds have with the error data.

* dom/CustomEvent.h:
Remove unimplemented stale method.

LayoutTests:

* platform/wk2/TestExpectations:
Skip new userscript test on WebKit2 like others.

* fast/events/constructors/error-event-constructor-expected.txt:
* fast/events/constructors/error-event-constructor.html:
Test for the new any "error" initialization property.

* fast/events/event-leak-objects-expected.txt:
* fast/events/event-leak-objects.html:
* fast/events/event-properties-gc-expected.txt:
* fast/events/event-properties-gc.html:
Like CustomEvent.details, ErrorEvent.error should not leak
objects across isolated worlds.

* fast/events/window-onerror-exception-in-attr-expected.txt:
* fast/events/window-onerror-exception-in-attr.html:
* fast/events/window-onerror-syntax-error-in-attr-expected.txt:
* fast/events/window-onerror-syntax-error-in-attr.html:
* fast/events/window-onerror1-expected.txt:
* fast/events/window-onerror1.html:
* fast/events/window-onerror10-expected.txt:
* fast/events/window-onerror10.html:
* fast/events/window-onerror11-expected.txt:
* fast/events/window-onerror11.html:
* fast/events/window-onerror12-expected.txt:
* fast/events/window-onerror12.html:
* fast/events/window-onerror13.html:
* fast/events/window-onerror14.html:
* fast/events/window-onerror16.html:
* fast/events/window-onerror17-expected.txt: Added.
* fast/events/window-onerror17.html: Copied from LayoutTests/fast/events/window-onerror12.html.
* fast/events/window-onerror2-expected.txt:
* fast/events/window-onerror2.html:
* fast/events/window-onerror3-expected.txt:
* fast/events/window-onerror3.html:
* fast/events/window-onerror4-expected.txt:
* fast/events/window-onerror4.html:
* fast/events/window-onerror5-expected.txt:
* fast/events/window-onerror5.html:
* fast/events/window-onerror6-expected.txt:
* fast/events/window-onerror6.html:
* fast/events/window-onerror7-expected.txt:
* fast/events/window-onerror7.html:
* fast/events/window-onerror8-expected.txt:
* fast/events/window-onerror8.html:
* fast/events/window-onerror9-expected.txt:
* fast/events/window-onerror9.html:
Include output for the column number and error object where possible.

* fast/files/resources/setup-for-read-common.js:
* fast/files/workers/worker-apply-blob-url-to-xhr.html:
* fast/workers/resources/worker-error-in-handling-script-error.js:
* fast/workers/resources/worker-script-error-bubbled.js:
* fast/workers/resources/worker-script-error-handled.js:
* fast/workers/resources/worker-script-error.js:
* fast/workers/worker-script-error-expected.txt:
Test "onerror" within Workers, and the "error" event on
the Worker from the spawning context.

* http/tests/security/cross-origin-script-error-event-expected.txt: Added.
* http/tests/security/cross-origin-script-error-event-redirected-expected.txt: Added.
* http/tests/security/cross-origin-script-error-event-redirected.html: Added.
* http/tests/security/cross-origin-script-error-event.html: Added.
* http/tests/security/cross-origin-script-window-onerror-expected.txt:
* http/tests/security/cross-origin-script-window-onerror-redirected-expected.txt:
* http/tests/security/cross-origin-script-window-onerror-redirected.html:
* http/tests/security/cross-origin-script-window-onerror.html:
* http/tests/security/script-crossorigin-error-event-information-expected.txt: Added.
* http/tests/security/script-crossorigin-error-event-information.html: Added.
* http/tests/security/script-crossorigin-onerror-information-expected.txt:
* http/tests/security/script-crossorigin-onerror-information.html:
* http/tests/security/script-no-crossorigin-error-event-should-be-sanitized-expected.txt: Added.
* http/tests/security/script-no-crossorigin-error-event-should-be-sanitized.html: Copied from LayoutTests/http/tests/security/script-no-crossorigin-onerror-should-be-sanitized.html.
* http/tests/security/script-no-crossorigin-onerror-should-be-sanitized-expected.txt:
* http/tests/security/script-no-crossorigin-onerror-should-be-sanitized.html:
* http/tests/security/window-onerror-exception-in-iframe-expected.txt:
* http/tests/security/window-onerror-exception-in-iframe.html:
* http/tests/workers/worker-importScriptsOnError-expected.txt:
* js/dom/exception-line-number-expected.txt:
* js/dom/script-tests/exception-line-number.js:
(window.onerror):

* userscripts/window-onerror-for-isolated-world-1-expected.txt:
* userscripts/window-onerror-for-isolated-world-1.html:
* userscripts/window-onerror-for-isolated-world-2-expected.txt:
* userscripts/window-onerror-for-isolated-world-2.html:
* userscripts/window-onerror-for-isolated-world-3-expected.txt: Added.
* userscripts/window-onerror-for-isolated-world-3.html: Added.
Add an explicit test for error objects crossing isolated world boundaries.
In the Isolated World case, serializable values are cloned, but unserializable
values cause the error object to return to null to avoid leaking across worlds.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202023 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoSVG elements don't blend correctly into HTML
dino@apple.com [Tue, 14 Jun 2016 03:05:21 +0000 (03:05 +0000)]
SVG elements don't blend correctly into HTML
https://bugs.webkit.org/show_bug.cgi?id=158718
<rdar://problem/26782004>

Reviewed by Antoine Quint.

Source/WebCore:

We were not creating any transparency layers for the root SVG nodes.
This is ok if the SVG is the root document, because it is the backdrop.
However, if it is inline SVG, it needs to apply the operation in
order to composite into the document.

Test: svg/css/mix-blend-mode-with-inline-svg.html

* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::beginTransparencyLayers):

LayoutTests:

* svg/css/mix-blend-mode-with-inline-svg-expected.html: Added.
* svg/css/mix-blend-mode-with-inline-svg.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202022 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoThe sampling profiler should further protect itself against certain forms of sampling...
sbarati@apple.com [Tue, 14 Jun 2016 02:29:26 +0000 (02:29 +0000)]
The sampling profiler should further protect itself against certain forms of sampling bias that arise due to the sampling interval being in sync with some other system process
https://bugs.webkit.org/show_bug.cgi?id=158678

Reviewed by Benjamin Poulain.

I first became aware of this problem when I read this paper:
http://plv.colorado.edu/papers/mytkowicz-pldi10.pdf

To provide background for this change, I'll quote a paragraph
from section 6.2:
"One statically sound method for collecting random samples is to collect a
sample at every t + r milliseconds, where t is the desired sampling interval
and r is a random number between −t and t. One might think that sampling every
t seconds is enough (i.e., drop the r component) but it is not: specifically,
if a profiler samples every t seconds, the sampling rate would be synchronized
with any program or system activity that occurs at regular time intervals [17].
For example, if the thread scheduler switches between threads every 10ms and our
sampling interval was also 10ms, then we may always take samples immediately after
a thread switch. Because performance is often different immediately after a thread
switch than at other points (e.g., due to cache and TLB warm-up effects) we would
get biased data. The random component, r, guards against such situations."

* runtime/SamplingProfiler.cpp:
(JSC::SamplingProfiler::timerLoop):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202021 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoFalse-positive over-release of WebView in destroyWebViewAndOffscreenWindow() in DumpR...
ddkilzer@apple.com [Tue, 14 Jun 2016 02:27:08 +0000 (02:27 +0000)]
False-positive over-release of WebView in destroyWebViewAndOffscreenWindow() in DumpRenderTree.mm
<https://webkit.org/b/158716>

Reviewed by Andy Estes.

* DumpRenderTree/mac/DumpRenderTree.mm:
(destroyWebViewAndOffscreenWindow): Change this function to take
a WebView, and assert that it's the same as [mainFrame webView].
(dumpRenderTree): Pass 'webView' into
destroyWebViewAndOffscreenWindow().
* DumpRenderTree/mac/DumpRenderTreeMac.h:
(createWebViewAndOffscreenWindow): Add NS_RETURNS_RETAINED since
this function returns a +1 WebView.  Declare this method only for
Objective-C[++] source.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202020 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agostorage/indexeddb/modern/leaks-1.html leaks the database connection handle.
beidson@apple.com [Tue, 14 Jun 2016 00:34:11 +0000 (00:34 +0000)]
storage/indexeddb/modern/leaks-1.html leaks the database connection handle.
https://bugs.webkit.org/show_bug.cgi?id=158643

Reviewed by Alex Christensen.

Source/WebCore:

Tested by changes to existing test.

* Modules/indexeddb/IDBDatabase.cpp:
(WebCore::IDBDatabase::hasPendingActivity):

* dom/EventTarget.h:
(WebCore::EventTarget::eventTargetData):
(WebCore::EventTarget::hasEventListeners):

LayoutTests:

* storage/indexeddb/modern/leak-1-expected.txt:
* storage/indexeddb/modern/resources/leak-1.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202019 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoREGRESSION(r201956): Failure to initialize new internal settings produced random...
enrica@apple.com [Tue, 14 Jun 2016 00:17:44 +0000 (00:17 +0000)]
REGRESSION(r201956): Failure to initialize new internal settings produced random test failures in release.
https://bugs.webkit.org/show_bug.cgi?id=158713
rdar://26769957

Reviewed by Simon Fraser.

Failed to initialize the new member variable in both Settings and InternalSettings classes.

* page/Settings.cpp:
(WebCore::Settings::Settings):
* testing/InternalSettings.cpp:
(WebCore::InternalSettings::Backup::Backup):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202018 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoDrop HipChat hack introduced in r197548
cdumez@apple.com [Tue, 14 Jun 2016 00:11:16 +0000 (00:11 +0000)]
Drop HipChat hack introduced in r197548
https://bugs.webkit.org/show_bug.cgi?id=158711

Reviewed by Geoffrey Garen.

Drop HipChat hack introduced in r197548. This hack is no longer needed
as the bug was fixed in HipChat since then:
https://support.atlassian.com/servicedesk/customer/portal/32/HCP-7532

I have confirmed locally that the latest version (4.0.12.665) is able
to connect without the hack.

* bindings/js/JSLocationCustom.cpp:
(WebCore::JSLocation::putDelegate): Deleted.
* platform/RuntimeApplicationChecks.h:
* platform/RuntimeApplicationChecks.mm:
(WebCore::MacApplication::isHipChat): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202017 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoMinor style updates for figcaptions and post forewords.
jond@apple.com [Mon, 13 Jun 2016 23:52:09 +0000 (23:52 +0000)]
Minor style updates for figcaptions and post forewords.

* wp-content/themes/webkit/style.css:
(article .foreword):
(figure.aligncenter figcaption):
(figure.aligncenter figcaption::before):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202016 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoDFG Validation fails when performing a concatenation with only a single entry
oliver@apple.com [Mon, 13 Jun 2016 23:17:00 +0000 (23:17 +0000)]
DFG Validation fails when performing a concatenation with only a single entry
https://bugs.webkit.org/show_bug.cgi?id=158699

Reviewed by Saam Barati.

Fairly simple short circuiting of a single replacement template string
without any padding to be planted as a simple to string rather than
op_strcat.

* bytecompiler/NodesCodegen.cpp:
(JSC::TemplateLiteralNode::emitBytecode):
* tests/stress/template-literal.js:
(testSingleNode):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202015 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAX: CrashTracer: com.apple.WebKit.WebContent at WebCore::AccessibilityRenderObject...
cfleizach@apple.com [Mon, 13 Jun 2016 23:05:20 +0000 (23:05 +0000)]
AX: CrashTracer: com.apple.WebKit.WebContent at WebCore::AccessibilityRenderObject::remoteSVGRootElement const + 227
https://bugs.webkit.org/show_bug.cgi?id=158685

Reviewed by David Kilzer.

Crash reports show a null access at a line that tries to dereference a pointer.
I still don't have a way to layout test this, as it seems tied to tear down of the main document.

* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::remoteSVGRootElement):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202014 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUse two video layer solution only on mac.
commit-queue@webkit.org [Mon, 13 Jun 2016 22:57:31 +0000 (22:57 +0000)]
Use two video layer solution only on mac.
https://bugs.webkit.org/show_bug.cgi?id=158705
rdar://problem/26776360

Patch by Jeremy Jones <jeremyj@apple.com> on 2016-06-13
Reviewed by Jer Noble.

Two video layer solution is only useful on the mac to prevent flicker, so don't do it elsewhere.

* platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::MediaPlayerPrivateAVFoundationObjC::createAVPlayerLayer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202013 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAdds support for foreword custom meta and display styles on blog posts.
jond@apple.com [Mon, 13 Jun 2016 22:53:57 +0000 (22:53 +0000)]
Adds support for foreword custom meta and display styles on blog posts.
https://bugs.webkit.org/show_bug.cgi?id=158707.

Reviewed by Timothy Hatcher.

* wp-content/themes/webkit/functions.php:
* wp-content/themes/webkit/style.css:
(article .foreword):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202012 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWeb Inspector: Filter Records not applying to new records
mattbaker@apple.com [Mon, 13 Jun 2016 22:44:26 +0000 (22:44 +0000)]
Web Inspector: Filter Records not applying to new records
https://bugs.webkit.org/show_bug.cgi?id=158213
<rdar://problem/26543912>

Reviewed by Timothy Hatcher.

This patch makes the following improvements to timeline grid filtering:

    - Records are filtered as they are added to the grid.
    - Timeline view no longer triggers a filter refresh while the
      current time changes while recording.
    - Filters are refreshed whenever the current timeline view changes.

* UserInterface/Views/DataGrid.js:
(WebInspector.DataGrid.prototype.hasFilters):
Helper function to check for custom and text filters.

(WebInspector.DataGrid.prototype._applyFiltersToNodeAndDispatchEvent):
Encapsulate node state change and event dispatch, which was previously
done in multiple places.

(WebInspector.DataGrid.prototype.insertChild):
Filter incoming node if needed.

(WebInspector.DataGrid.prototype._updateFilter.createIteratorForNodesToBeFiltered):
(WebInspector.DataGrid.prototype._updateFilter):
(WebInspector.DataGrid.prototype.yieldableTaskWillProcessItem):
Use new convenience functions.

(WebInspector.DataGrid.prototype.hasCustomFilters): Deleted.
Renamed `hasFilters`.

* UserInterface/Views/TimelineRecordingContentView.js:
(WebInspector.TimelineRecordingContentView.prototype._currentContentViewDidChange):
Refresh grid filters on view change.

* UserInterface/Views/TimelineView.js:
(WebInspector.TimelineView.prototype.set startTime):
(WebInspector.TimelineView.prototype.set endTime):
Schedule filter change notification when selection bounds changes.

(WebInspector.TimelineView.prototype._timesDidChange):
(WebInspector.TimelineView.prototype._scheduleFilterDidChange):
(WebInspector.TimelineView):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202011 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWeb Inspector: Call Trees view should have a 'Top Functions'-like mode
sbarati@apple.com [Mon, 13 Jun 2016 22:43:20 +0000 (22:43 +0000)]
Web Inspector: Call Trees view should have a 'Top Functions'-like mode
https://bugs.webkit.org/show_bug.cgi?id=158555
<rdar://problem/26712544>

Reviewed by Timothy Hatcher.

This patch adds a Top Functions view that is similar to Instruments'
Top Functions view. I really wanted to add this view because I've
been doing a lot of performance debugging and I've exclusively
used the Top Functions view and I want the Inspector to also have
this view. I like to think of it as a more sophisticated version of the bottom-up view.

Top Functions works by treating every frame as a root in the tree.
Top functions view then presents a list of "roots". This is the same
as all other views, which also present a list of roots, but in this case,
every frame is a root. Bottom Up is great for nailing in specific performance
problems in exactly one frame. But Bottom Up doesn't give you good context about where
a specific frame is in the call tree and how frames are related by having
a caller or some path of shared callers. For example, consider this call tree:
         (program)
         /        \
        /          \
   (many nodes...)
      /
     /
    (parent)
    /   \
   /     \
 (leaf1)  (leaf2)

Suppose that 'leaf1' is super hot, and 'leaf2' is moderately hot.
If we look at this through Bottom Up view, we will see 'leaf1'
is super hot, but it will take more scrolling to see that 'leaf2'
is moderately hot. Lets say that 'parent' is also moderately hot,
but that the majority of its time isn't self time. With Bottom Up view,
there is no good way to see that 'leaf1' and 'leaf2' are both nodes under 'parent'.
With Top Down, you can find this information, but it requires a ton of drilling down into
the tree (i.e, you must expand past the 'many nodes...' I drew above). It's inconvenient to
use Top Down here for indentation alone. Bottom up will tell you that 'leaf1' is super hot,
and that 'leaf2' and 'parent' are moderately hot, but it doesn't show how they're related
in the original tree. It's important to see that 'parent's total time is very high
because it itself is moderately hot, and it has a child node that is super hot, and
another child that's moderately 'hot'. For the sake of this example, let's pretend
that 85% of the program's time is spent inside 'parent'. Seeing this information through
'Top Functions' is easy because this information filters to the top of the list. Specifically,
when using 'Top Functions' sorted by Total Time. Because every node is a root, there will be
a top-level entry for every frame in the program. Specifically, there will be a top-level node
for 'parent' in my above example. Because I've sorted this view by Total Time, I will see '(program)'
first. That's because 100% of execution time is under the '(program)' frame. Then, I might see
a few other nodes that also run the entire time because '(program)' calls them, and they eventually
call into other things that never leave the stack. These will also have time ranges near 100%.
But, only a few nodes after that, I'll see 'parent' in the list because it accounts for 85% of
execution time. Immediately, I will see that it has some self time, and that it has two child
nodes that have self time. This is really helpful.

Let's consider another example where it's not easy in Top Down to get the full picture of 'parent':
           (program)
            /  |  \
         (... many nodes...)
          /           \
    (many nodes...)   (many nodes...)
         /             \
       parent         parent
         |              |
        leaf1          leaf2

If we viewed this program in Top Down, we don't get a full picture of 'parent'
because it has its time distributed in two different subsections of the tree.
Specifically, lets say it has 70% of time in the leaf1 path, and 30% of the
time in the leaf2 path. We want a way to see these things together. It's impossible
to do this in Top Down or Bottom Up. But, in Top Functions view, we get the view that
we want to see because we treat 'parent' as a root of the tree. Because we do this,
we will create the following sub tree in the Top Functions view:
        parent
       /      \
     leaf1   leaf2
This happens naturally because when 'parent' is a root, we add all its children
to its subtree.

Constructing this tree is really easy. What we do is take any arbitrary stack
trace of length n, and treat is as n separate stack traces. Specifically, we
perform the following operation for any stack trace S.

S = [A, B, C, D]
(A is the entry frame, and D is the top of the stack).
We will transform this into a list of stack traces S' like so:
S' = [[A, B, C, D], [B, C, D], [C, D], [D]]

If we then run the normal top down tree algorithm on this set of stack
traces, all nodes get treated as roots, and voila, we get the Top Functions view.

* Localizations/en.lproj/localizedStrings.js:
* UserInterface/Controllers/TimelineManager.js:
* UserInterface/Main.html:
* UserInterface/Models/CallingContextTree.js:
* UserInterface/Models/TimelineRecording.js:
* UserInterface/Views/ScriptProfileTimelineView.js:
* UserInterface/Views/TextToggleButtonNavigationItem.css: Added.
* UserInterface/Views/TextToggleButtonNavigationItem.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202010 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWeb Inspector: Add ability to show/hide DataGird columns
mattbaker@apple.com [Mon, 13 Jun 2016 22:41:54 +0000 (22:41 +0000)]
Web Inspector: Add ability to show/hide DataGird columns
https://bugs.webkit.org/show_bug.cgi?id=158676
<rdar://problem/26761573>

Reviewed by Timothy Hatcher.

Make it possible to show/hide grid columns using the grid header
context menu. This patch enables the new behavior for most of the
timeline grids.

* UserInterface/Views/DataGrid.js:
(WebInspector.DataGrid):
(WebInspector.DataGrid.prototype.get identifier):
(WebInspector.DataGrid.prototype.set identifier):
An identifier for the grid instance, for managing per-grid settings.
Setting the id causes settings to be created, and their values to be
applied to the grid.

(WebInspector.DataGrid.prototype.get columnChooserEnabled):
(WebInspector.DataGrid.prototype.set columnChooserEnabled):
Enable showing/hiding columns via the grid header.

(WebInspector.DataGrid.prototype.insertColumn):
(WebInspector.DataGrid.prototype.showColumn):
Set column visibility and hidden column setting, then perform layout.

(WebInspector.DataGrid.prototype._collapseColumnGroupWithCell):
(WebInspector.DataGrid.prototype._contextMenuInHeader):
Create column chooser menu items if necessary.

(WebInspector.DataGrid.prototype._showColumn): Deleted.
(WebInspector.DataGrid.prototype._hideColumn): Deleted.
Replaced by `showColumn`.

* UserInterface/Views/LayoutTimelineView.js:
(WebInspector.LayoutTimelineView):
Always show "type" and "name" columns.

* UserInterface/Views/NetworkTimelineView.js:
(WebInspector.NetworkTimelineView):
* UserInterface/Views/RenderingFrameTimelineView.js:
(WebInspector.RenderingFrameTimelineView):
* UserInterface/Views/ScriptDetailsTimelineView.js:
(WebInspector.ScriptDetailsTimelineView):
Always show "name" column.

* UserInterface/Views/TimelineDataGrid.js:
(WebInspector.TimelineDataGrid):
Enable column chooser.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202009 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoBuild fix. Strip out "use strict" everywhere so that the perf dashboard works on...
rniwa@webkit.org [Mon, 13 Jun 2016 22:25:52 +0000 (22:25 +0000)]
Build fix. Strip out "use strict" everywhere so that the perf dashboard works on the shipping Safari.

* tools/bundle-v3-scripts.py:
(main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202008 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoFTL::Output methods should be out-of-line whenever possible
fpizlo@apple.com [Mon, 13 Jun 2016 22:17:50 +0000 (22:17 +0000)]
FTL::Output methods should be out-of-line whenever possible
https://bugs.webkit.org/show_bug.cgi?id=158704

Reviewed by Benjamin Poulain.

These methods turn into a non-trivial amount of code because of the template-based B3 API.
Inlining them didn't achieve any performance advantages for the FTL, but it did make the
code larger. This outlines most methods in FTL::Output. It makes FTL::LowerDFGToB3 smaller
and it doesn't change performance.

* ftl/FTLOutput.cpp:
(JSC::FTL::Output::appendTo):
(JSC::FTL::Output::framePointer):
(JSC::FTL::Output::lockedStackSlot):
(JSC::FTL::Output::constBool):
(JSC::FTL::Output::constInt32):
(JSC::FTL::Output::constInt64):
(JSC::FTL::Output::constDouble):
(JSC::FTL::Output::phi):
(JSC::FTL::Output::add):
(JSC::FTL::Output::sub):
(JSC::FTL::Output::mul):
(JSC::FTL::Output::div):
(JSC::FTL::Output::chillDiv):
(JSC::FTL::Output::mod):
(JSC::FTL::Output::chillMod):
(JSC::FTL::Output::neg):
(JSC::FTL::Output::doubleAdd):
(JSC::FTL::Output::doubleSub):
(JSC::FTL::Output::doubleMul):
(JSC::FTL::Output::doubleDiv):
(JSC::FTL::Output::doubleMod):
(JSC::FTL::Output::bitAnd):
(JSC::FTL::Output::bitOr):
(JSC::FTL::Output::bitXor):
(JSC::FTL::Output::shl):
(JSC::FTL::Output::aShr):
(JSC::FTL::Output::lShr):
(JSC::FTL::Output::bitNot):
(JSC::FTL::Output::logicalNot):
(JSC::FTL::Output::ctlz32):
(JSC::FTL::Output::doubleAbs):
(JSC::FTL::Output::doubleCeil):
(JSC::FTL::Output::doubleFloor):
(JSC::FTL::Output::doubleTrunc):
(JSC::FTL::Output::doubleSin):
(JSC::FTL::Output::doubleCos):
(JSC::FTL::Output::doublePow):
(JSC::FTL::Output::doublePowi):
(JSC::FTL::Output::doubleSqrt):
(JSC::FTL::Output::doubleLog):
(JSC::FTL::Output::hasSensibleDoubleToInt):
(JSC::FTL::Output::doubleToUInt):
(JSC::FTL::Output::signExt32To64):
(JSC::FTL::Output::zeroExt):
(JSC::FTL::Output::intToDouble):
(JSC::FTL::Output::unsignedToDouble):
(JSC::FTL::Output::castToInt32):
(JSC::FTL::Output::doubleToFloat):
(JSC::FTL::Output::floatToDouble):
(JSC::FTL::Output::load):
(JSC::FTL::Output::load8SignExt32):
(JSC::FTL::Output::baseIndex):
(JSC::FTL::Output::equal):
(JSC::FTL::Output::notEqual):
(JSC::FTL::Output::above):
(JSC::FTL::Output::aboveOrEqual):
(JSC::FTL::Output::below):
(JSC::FTL::Output::belowOrEqual):
(JSC::FTL::Output::greaterThan):
(JSC::FTL::Output::greaterThanOrEqual):
(JSC::FTL::Output::lessThan):
(JSC::FTL::Output::lessThanOrEqual):
(JSC::FTL::Output::doubleEqual):
(JSC::FTL::Output::doubleEqualOrUnordered):
(JSC::FTL::Output::doubleNotEqualOrUnordered):
(JSC::FTL::Output::doubleLessThan):
(JSC::FTL::Output::doubleLessThanOrEqual):
(JSC::FTL::Output::doubleGreaterThan):
(JSC::FTL::Output::doubleGreaterThanOrEqual):
(JSC::FTL::Output::doubleNotEqualAndOrdered):
(JSC::FTL::Output::doubleLessThanOrUnordered):
(JSC::FTL::Output::doubleLessThanOrEqualOrUnordered):
(JSC::FTL::Output::doubleGreaterThanOrUnordered):
(JSC::FTL::Output::doubleGreaterThanOrEqualOrUnordered):
(JSC::FTL::Output::isZero32):
(JSC::FTL::Output::notZero32):
(JSC::FTL::Output::isZero64):
(JSC::FTL::Output::notZero64):
(JSC::FTL::Output::select):
(JSC::FTL::Output::jump):
(JSC::FTL::Output::branch):
(JSC::FTL::Output::check):
(JSC::FTL::Output::ret):
(JSC::FTL::Output::unreachable):
(JSC::FTL::Output::speculate):
(JSC::FTL::Output::speculateAdd):
(JSC::FTL::Output::speculateSub):
(JSC::FTL::Output::speculateMul):
(JSC::FTL::Output::patchpoint):
(JSC::FTL::Output::trap):
(JSC::FTL::Output::anchor):
(JSC::FTL::Output::bitCast):
(JSC::FTL::Output::fround):
* ftl/FTLOutput.h:
(JSC::FTL::Output::setOrigin):
(JSC::FTL::Output::origin):
(JSC::FTL::Output::constIntPtr):
(JSC::FTL::Output::doubleNeg):
(JSC::FTL::Output::zeroExtPtr):
(JSC::FTL::Output::load32NonNegative):
(JSC::FTL::Output::isNull):
(JSC::FTL::Output::notNull):
(JSC::FTL::Output::testIsZeroPtr):
(JSC::FTL::Output::testNonZeroPtr):
(JSC::FTL::Output::call):
(JSC::FTL::Output::operation):
(JSC::FTL::Output::branch):
(JSC::FTL::Output::switchInstruction):
(JSC::FTL::Output::addIncomingToPhi):
(JSC::FTL::Output::framePointer): Deleted.
(JSC::FTL::Output::constBool): Deleted.
(JSC::FTL::Output::constInt32): Deleted.
(JSC::FTL::Output::constInt64): Deleted.
(JSC::FTL::Output::constDouble): Deleted.
(JSC::FTL::Output::phi): Deleted.
(JSC::FTL::Output::add): Deleted.
(JSC::FTL::Output::sub): Deleted.
(JSC::FTL::Output::mul): Deleted.
(JSC::FTL::Output::div): Deleted.
(JSC::FTL::Output::chillDiv): Deleted.
(JSC::FTL::Output::mod): Deleted.
(JSC::FTL::Output::chillMod): Deleted.
(JSC::FTL::Output::doubleAdd): Deleted.
(JSC::FTL::Output::doubleSub): Deleted.
(JSC::FTL::Output::doubleMul): Deleted.
(JSC::FTL::Output::doubleDiv): Deleted.
(JSC::FTL::Output::doubleMod): Deleted.
(JSC::FTL::Output::bitAnd): Deleted.
(JSC::FTL::Output::bitOr): Deleted.
(JSC::FTL::Output::bitXor): Deleted.
(JSC::FTL::Output::shl): Deleted.
(JSC::FTL::Output::aShr): Deleted.
(JSC::FTL::Output::lShr): Deleted.
(JSC::FTL::Output::ctlz32): Deleted.
(JSC::FTL::Output::addWithOverflow32): Deleted.
(JSC::FTL::Output::subWithOverflow32): Deleted.
(JSC::FTL::Output::mulWithOverflow32): Deleted.
(JSC::FTL::Output::addWithOverflow64): Deleted.
(JSC::FTL::Output::subWithOverflow64): Deleted.
(JSC::FTL::Output::mulWithOverflow64): Deleted.
(JSC::FTL::Output::doubleAbs): Deleted.
(JSC::FTL::Output::doubleCeil): Deleted.
(JSC::FTL::Output::doubleFloor): Deleted.
(JSC::FTL::Output::doubleSin): Deleted.
(JSC::FTL::Output::doubleCos): Deleted.
(JSC::FTL::Output::doublePow): Deleted.
(JSC::FTL::Output::doubleSqrt): Deleted.
(JSC::FTL::Output::doubleLog): Deleted.
(JSC::FTL::Output::signExt32To64): Deleted.
(JSC::FTL::Output::zeroExt): Deleted.
(JSC::FTL::Output::intToDouble): Deleted.
(JSC::FTL::Output::castToInt32): Deleted.
(JSC::FTL::Output::doubleToFloat): Deleted.
(JSC::FTL::Output::floatToDouble): Deleted.
(JSC::FTL::Output::equal): Deleted.
(JSC::FTL::Output::notEqual): Deleted.
(JSC::FTL::Output::above): Deleted.
(JSC::FTL::Output::aboveOrEqual): Deleted.
(JSC::FTL::Output::below): Deleted.
(JSC::FTL::Output::belowOrEqual): Deleted.
(JSC::FTL::Output::greaterThan): Deleted.
(JSC::FTL::Output::greaterThanOrEqual): Deleted.
(JSC::FTL::Output::lessThan): Deleted.
(JSC::FTL::Output::lessThanOrEqual): Deleted.
(JSC::FTL::Output::doubleEqual): Deleted.
(JSC::FTL::Output::doubleEqualOrUnordered): Deleted.
(JSC::FTL::Output::doubleNotEqualOrUnordered): Deleted.
(JSC::FTL::Output::doubleLessThan): Deleted.
(JSC::FTL::Output::doubleLessThanOrEqual): Deleted.
(JSC::FTL::Output::doubleGreaterThan): Deleted.
(JSC::FTL::Output::doubleGreaterThanOrEqual): Deleted.
(JSC::FTL::Output::doubleNotEqualAndOrdered): Deleted.
(JSC::FTL::Output::doubleLessThanOrUnordered): Deleted.
(JSC::FTL::Output::doubleLessThanOrEqualOrUnordered): Deleted.
(JSC::FTL::Output::doubleGreaterThanOrUnordered): Deleted.
(JSC::FTL::Output::doubleGreaterThanOrEqualOrUnordered): Deleted.
(JSC::FTL::Output::isZero32): Deleted.
(JSC::FTL::Output::notZero32): Deleted.
(JSC::FTL::Output::isZero64): Deleted.
(JSC::FTL::Output::notZero64): Deleted.
(JSC::FTL::Output::select): Deleted.
(JSC::FTL::Output::extractValue): Deleted.
(JSC::FTL::Output::jump): Deleted.
(JSC::FTL::Output::ret): Deleted.
(JSC::FTL::Output::unreachable): Deleted.
(JSC::FTL::Output::speculate): Deleted.
(JSC::FTL::Output::speculateAdd): Deleted.
(JSC::FTL::Output::speculateSub): Deleted.
(JSC::FTL::Output::speculateMul): Deleted.
(JSC::FTL::Output::patchpoint): Deleted.
(JSC::FTL::Output::trap): Deleted.
(JSC::FTL::Output::anchor): Deleted.
(JSC::FTL::Output::bitCast): Deleted.
(JSC::FTL::Output::fround): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202007 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoDecrease PiP flicker by not removing window prematurely.
commit-queue@webkit.org [Mon, 13 Jun 2016 22:00:49 +0000 (22:00 +0000)]
Decrease PiP flicker by not removing window prematurely.
https://bugs.webkit.org/show_bug.cgi?id=158436
<rdar://problem/19052639>

Patch by Jeremy Jones <jeremyj@apple.com> on 2016-06-13
Reviewed by Darin Adler.

UIWindow shouldn't be removed until cleanupFullscreen, so the video layer has a chance
to be reparented in the DOM first.

* platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
(WebVideoFullscreenInterfaceAVKit::didStopPictureInPicture):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202006 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAdd WebSocketProvider stub
achristensen@apple.com [Mon, 13 Jun 2016 21:58:53 +0000 (21:58 +0000)]
Add WebSocketProvider stub
https://bugs.webkit.org/show_bug.cgi?id=158702

Reviewed by Brady Eidson.

Source/WebCore:

No new tests.  No change in behavior.

* WebCore.xcodeproj/project.pbxproj:
* dom/DocumentMarkerController.cpp:
* dom/ScriptedAnimationController.cpp:
* html/HTMLMediaElement.cpp:
* html/MediaDocument.cpp:
* html/shadow/MediaControlElements.cpp:
* html/shadow/MediaControls.cpp:
* html/shadow/MediaControls.h:
* html/shadow/MediaControlsApple.cpp:
* inspector/InspectorInstrumentation.cpp:
* inspector/InspectorInstrumentation.h:
* inspector/InspectorOverlay.cpp:
(WebCore::InspectorOverlay::overlayPage):
* loader/EmptyClients.h:
* loader/FrameLoader.cpp:
* loader/FrameLoader.h:
* loader/appcache/ApplicationCacheHost.cpp:
* loader/cache/CachedResource.cpp:
* page/FrameView.cpp:
* page/Page.cpp:
(WebCore::Page::Page):
* page/Page.h:
(WebCore::Page::applicationCacheStorage):
(WebCore::Page::databaseProvider):
(WebCore::Page::socketProvider):
(WebCore::Page::storageNamespaceProvider):
* page/PageConfiguration.cpp:
(WebCore::PageConfiguration::PageConfiguration):
* page/PageConfiguration.h:
* page/ResourceUsageOverlay.cpp:
* page/SocketProvider.h: Added.
(WebCore::SocketProvider::~SocketProvider):
* page/cocoa/ResourceUsageOverlayCocoa.mm:
* rendering/RenderElement.cpp:
* rendering/RenderLayerBacking.cpp:
* style/StyleResolveForDocument.cpp:
* style/StyleTreeResolver.cpp:
* svg/graphics/SVGImage.cpp:
(WebCore::SVGImage::dataChanged):
* testing/MockPageOverlayClient.cpp:

Source/WebKit:

* WebKit.xcodeproj/project.pbxproj:

Source/WebKit/mac:

* Misc/WebSocketProvider.h: Added.
* WebView/WebView.mm:
(-[WebView _commonInitializationWithFrameName:groupName:]):
(-[WebView initSimpleHTMLDocumentWithStyle:frame:preferences:groupName:]):

Source/WebKit/win:

* WebSocketProvider.h: Added.
* WebView.cpp:
(WebView::initWithFrame):

Source/WebKit2:

* WebKit2.xcodeproj/project.pbxproj:
* WebProcess/Network/WebSocketProvider.h: Added.
* WebProcess/WebPage/WebPage.cpp:
(WebKit::m_shouldDispatchFakeMouseMoveEvents):

Source/WTF:

* wtf/UniqueRef.h:
(WTF::UniqueRef::operator->):
(WTF::UniqueRef::operator T&):
(WTF::UniqueRef::operator const T&):
Added operator T& to avoid unneeded .get() calls.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202005 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUnreviewed, Cloop build fix.
keith_miller@apple.com [Mon, 13 Jun 2016 21:55:42 +0000 (21:55 +0000)]
Unreviewed, Cloop build fix.

* bytecode/BytecodeList.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202004 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAdd new builtin opcode tailCallForwardArguments
keith_miller@apple.com [Mon, 13 Jun 2016 21:05:36 +0000 (21:05 +0000)]
Add new builtin opcode tailCallForwardArguments
https://bugs.webkit.org/show_bug.cgi?id=158666

Reviewed by Filip Pizlo.

We should support the ability to have a builtin forward its
arguments to a helper without allocating an arguments object. This
patch adds a new bytecode intrinsic @tailCallForwardArguments that
takes two values. The first is the target of the call and the
second is the new this value. This opcode will tail call to the
passed function without triggering an allocation of an arguments
object for the caller function.

In the LLInt and Baseline this function acts the same way a normal
tail call does.  The bytecode will allocate a new stack frame
copying all the arguments of the caller function into the new
frame, along with the new this. Then when the actual call happens
the new frame is copied over the caller frame. While this is not
necessary, it allows the target function to have more arguments
than the caller function via arity fixup.

Once we get to the DFG we reuse existing DFG Nodes for forwarding
arguments, although there were some minor changes. This patch
swaps the meaning of the second and third children for each DFG
varargs node, exchanging the argmuments and this child,
respectively. It also makes the arguments child for each varargs
node, as well as the ForwardVarargs node optional. If the optional
child is missing, then forwarding node assumes that the arguments
for the node's inlineCallFrame should be used instead. Finally,
when inlining the target of an inlined
op_tail_call_forward_arguments we make sure the arguments of the
forwarding function are marked as non-unboxable since this would
normally be done by the caller's create arguments object node,
which does not exist in this case.

* bytecode/BytecodeIntrinsicRegistry.h:
* bytecode/BytecodeList.json:
* bytecode/BytecodeUseDef.h:
(JSC::computeUsesForBytecodeOffset):
(JSC::computeDefsForBytecodeOffset):
* bytecode/CallLinkInfo.h:
(JSC::CallLinkInfo::callTypeFor):
* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpBytecode):
(JSC::CodeBlock::finishCreation):
* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::emitCallForwardArgumentsInTailPosition):
(JSC::BytecodeGenerator::emitCallVarargs):
* bytecompiler/BytecodeGenerator.h:
* bytecompiler/NodesCodegen.cpp:
(JSC::BytecodeIntrinsicNode::emit_intrinsic_tailCallForwardArguments):
(JSC::BytecodeIntrinsicNode::emit_intrinsic_tryGetById):
* dfg/DFGArgumentsEliminationPhase.cpp:
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::getPredictionWithoutOSRExit):
(JSC::DFG::ByteCodeParser::handleCall):
(JSC::DFG::ByteCodeParser::handleVarargsCall):
(JSC::DFG::ByteCodeParser::handleInlining):
(JSC::DFG::ByteCodeParser::parseBlock):
* dfg/DFGCapabilities.cpp:
(JSC::DFG::capabilityLevel):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
* dfg/DFGNode.h:
(JSC::DFG::Node::hasArgumentsChild):
(JSC::DFG::Node::argumentsChild):
* dfg/DFGPreciseLocalClobberize.h:
(JSC::DFG::PreciseLocalClobberizeAdaptor::readTop):
* dfg/DFGPredictionPropagationPhase.cpp:
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileForwardVarargs):
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::emitCall):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::emitCall):
* dfg/DFGVarargsForwardingPhase.cpp:
* ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::LowerDFGToB3::compileCallOrConstructVarargs):
(JSC::FTL::DFG::LowerDFGToB3::compileForwardVarargs):
* interpreter/Interpreter.cpp:
(JSC::sizeFrameForForwardArguments):
(JSC::setupForwardArgumentsFrame):
(JSC::setupForwardArgumentsFrameAndSetThis):
* interpreter/Interpreter.h:
* jit/JIT.cpp:
(JSC::JIT::privateCompileMainPass):
(JSC::JIT::privateCompileSlowCases):
* jit/JIT.h:
* jit/JITCall.cpp:
(JSC::JIT::compileSetupVarargsFrame):
(JSC::JIT::compileOpCall):
(JSC::JIT::compileOpCallSlowCase):
(JSC::JIT::emit_op_tail_call_forward_arguments):
(JSC::JIT::emitSlow_op_tail_call_forward_arguments):
* jit/JITCall32_64.cpp:
(JSC::JIT::emitSlow_op_tail_call_forward_arguments):
(JSC::JIT::emit_op_tail_call_forward_arguments):
(JSC::JIT::compileSetupVarargsFrame):
(JSC::JIT::compileOpCall):
* jit/JITOperations.cpp:
* jit/JITOperations.h:
* llint/LLIntSlowPaths.cpp:
(JSC::LLInt::LLINT_SLOW_PATH_DECL):
(JSC::LLInt::varargsSetup):
* llint/LLIntSlowPaths.h:
* llint/LowLevelInterpreter.asm:
* tests/stress/tailCallForwardArguments.js: Added.
(putFuncToPrivateName.createBuiltin):
(putFuncToPrivateName):
(createTailCallForwardingFuncWith):
(baz):
(baz2):
(baz3):
(let.bodyText):
(baz4):
(baz5):
(arrayEq):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202003 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoMake HashMap and HashSet work with Refs
commit-queue@webkit.org [Mon, 13 Jun 2016 20:06:22 +0000 (20:06 +0000)]
Make HashMap and HashSet work with Refs
https://bugs.webkit.org/show_bug.cgi?id=158638

Patch by Sam Weinig <sam@webkit.org> on 2016-06-13
Reviewed by Darin Adler.

Source/WTF:

* wtf/GetPtr.h:
Move HashTableDeletedValueType and HashTableEmptyValueType here, as they are now shared
by more than one smart pointer. This file should probably be renamed to something else
at some point to indicate that it contains helpers for pointer and ref related functionality.

* wtf/HashFunctions.h:
Add a DefaultHash for Refs. Customize the PtrHash to indicate that it is not safe to compare
to empty or deleted Refs.

* wtf/HashMap.h:
(WTF::HashMapTranslator::equal):
(WTF::HashMapTranslator::translate):
(WTF::HashMapEnsureTranslator::equal):
(WTF::HashMapEnsureTranslator::translate):
* wtf/HashSet.h:
(WTF::IdentityExtractor::extract):
(WTF::HashSetTranslator::hash):
(WTF::HashSetTranslator::equal):
(WTF::HashSetTranslator::translate):
* wtf/HashTable.h:
(WTF::IdentityHashTranslator::hash):
(WTF::IdentityHashTranslator::equal):
(WTF::IdentityHashTranslator::translate):
Use the new assignToEmpty trait function to allow uninitialized Ref's to be safely assigned to.

* wtf/HashTraits.h:
(WTF::HashTraits<Ref<P>>::emptyValue):
(WTF::HashTraits<Ref<P>>::isEmptyValue):
(WTF::HashTraits<Ref<P>>::assignToEmpty):
(WTF::HashTraits<Ref<P>>::peek):
(WTF::HashTraits<Ref<P>>::customDeleteBucket):
Add custom HashTraits for Ref. Also, introduce a new trait function, assignToEmpty, for use
in translation assignments. This is necessary since the default assignment operator for Ref
will not allow assignment to the empty Ref, which we need to do here.

* wtf/Ref.h:
(WTF::Ref::operator=):
(WTF::Ref::Ref):
(WTF::Ref::isHashTableDeletedValue):
(WTF::Ref::hashTableDeletedValue):
(WTF::Ref::isHashTableEmptyValue):
(WTF::Ref::hashTableEmptyValue):
Add explicit constructors/predicates for making deleted/empty Refs.

(WTF::Ref::assignToHashTableEmptyValue):
Add a special function that allows assignment to an empty Ref, which the
assignment operator does not.

(WTF::IsSmartPtr):
Add an IsSmartPtr override to indicate that Ref is a smart pointer.

* wtf/RefPtr.h:
Move HashTableDeletedValueType to GetPtr.h.

Tools:

* TestWebKitAPI/Tests/WTF/HashMap.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WTF/HashSet.cpp:
(TestWebKitAPI::TEST):
Add tests for using Refs in HashMaps (both as key and value) and HashSets.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202002 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoInvalid token error when trying to create an A/B analysis for a range
rniwa@webkit.org [Mon, 13 Jun 2016 19:47:38 +0000 (19:47 +0000)]
Invalid token error when trying to create an A/B analysis for a range
https://bugs.webkit.org/show_bug.cgi?id=158679

Reviewed by Chris Dumez.

The problem in this particular case was due to another website overriding cookies for our subdomain.
Make PrivilegedAPI robust against its token becoming invalid in general to fix the bug since the cookie
is only available under /privileged-api/ and the v3 UI can't access it for security reasons.

This patch factors out PrivilegedAPI out of remote.js so that it can be tested separately in server tests
as well as unit tests even though RemoteAPI itself is implemented differently in each case.

* init-database.sql: Added a forgotten default value "false" to run_marked_outlier.
* public/v3/index.html:
* public/v3/privileged-api.js: Added. Extracted out of public/v3/remote.js.
(PrivilegedAPI.sendRequest): Fixed the bug. When the initial request fails with "InvalidToken" error,
re-generate the token and re-issue the request.
(PrivilegedAPI.requestCSRFToken):
* public/v3/remote.js:
(RemoteAPI.postJSON): Added to match tools/js/remote.js.
(RemoteAPI.postJSONWithStatus): Ditto.
(PrivilegedAPI): Moved to privileged-api.js.
* server-tests/api-measurement-set-tests.js: Removed the unused require for crypto.
* server-tests/privileged-api-upate-run-status.js: Added tests for /privileged-api/update-run-status.
* server-tests/resources/test-server.js:
(TestServer.prototype.inject): Clear the cookies as well as tokens in PrivilegedAPI.
* tools/js/remote.js:
(RemoteAPI): Added the support for PrivilegedAPI by making cookie set by the server persist.
(RemoteAPI.prototype.clearCookies): Added for tests.
(RemoteAPI.prototype.postJSON): Make sure sendHttpRequest always sends a valid JSON.
(RemoteAPI.prototype.postJSONWithStatus): Added since this API is used PrivilegedAPI.
(RemoteAPI.prototype.sendHttpRequest): Retain the cookie set by the server and send it back in each request.
* tools/js/v3-models.js:
* unit-tests/privileged-api-tests.js: Added unit tests for PrivilegedAPI.
* unit-tests/resources/mock-remote-api.js:
(MockRemoteAPI.postJSON): Added for unit testing.
(MockRemoteAPI.postJSONWithStatus): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202001 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years ago/admin/tests is very slow
rniwa@webkit.org [Mon, 13 Jun 2016 19:40:01 +0000 (19:40 +0000)]
/admin/tests is very slow
https://bugs.webkit.org/show_bug.cgi?id=158682

Reviewed by Chris Dumez.

The slowness came from TestNameResolver::__construct, which was fetching the entire table of test_configurations,
which at this point contains more than 32,000 rows. Don't fetch the entire table in the constructor. Instead,
fetch a subset of rows as needed in configurations_for_metric_and_platform. Even though this results in many SQL
queries being issued, that's a lot more efficient in practice because we only fetch a few dozen rows in practice.

Also removed a whole bunch of features from /admin/tests to simplify the page. In particular, the ability to update
the list of triggerables has been removed now that sync-buildbot.js automatically updates that for us. This removed
the last use of test_exists_on_platform, which was also dependent on fetching test_configurations upfront.

* public/admin/tests.php:
* public/include/test-name-resolver.php:
(TestNameResolver::__construct): Don't fetch the entire table of test_configurations.
(TestNameResolver::configurations_for_metric_and_platform): Just issue a SQL query for the specified platform and metric.
(TestNameResolver::test_exists_on_platform): Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@202000 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWebRTC: Add test that verifies that RTCPeerConnection.addTrack can reuse an existing...
adam.bergkvist@ericsson.com [Mon, 13 Jun 2016 19:39:01 +0000 (19:39 +0000)]
WebRTC: Add test that verifies that RTCPeerConnection.addTrack can reuse an existing RTCRtpSender
https://bugs.webkit.org/show_bug.cgi?id=158621

Reviewed by Eric Carlson.

Add test to verify that RTCPeerConnection.addTrack reuses an existing RTCRtpSender
under the right circumstances.

* fast/mediastream/RTCPeerConnection-addTrack-reuse-sender-expected.txt: Added.
* fast/mediastream/RTCPeerConnection-addTrack-reuse-sender.html: Added.
* platform/mac/TestExpectations:
The mac port is not building with WEB_RTC yet.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201999 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWebRTC: Add media setup test with complete offer/answer dialog
adam.bergkvist@ericsson.com [Mon, 13 Jun 2016 19:38:37 +0000 (19:38 +0000)]
WebRTC: Add media setup test with complete offer/answer dialog
https://bugs.webkit.org/show_bug.cgi?id=158620

Reviewed by Eric Carlson.

Test a complete SDP offer/answer dialog between two RTCPeerConnection instances in the same
page. Media is set up in both direction with a single dialog. Check the local- and
remoteDescription attributes and the RTCPeerConnection states.

* fast/mediastream/RTCPeerConnection-media-setup-single-dialog-expected.txt: Added.
* fast/mediastream/RTCPeerConnection-media-setup-single-dialog.html: Added.
* platform/mac/TestExpectations:
The mac port is not building with WEB_RTC yet.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201998 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoCrashes in WebCore::IDBServer::UniqueIDBDatabase::executeNextDatabaseTask.
beidson@apple.com [Mon, 13 Jun 2016 18:56:43 +0000 (18:56 +0000)]
Crashes in WebCore::IDBServer::UniqueIDBDatabase::executeNextDatabaseTask.
<rdar://problem/26768449> and https://bugs.webkit.org/show_bug.cgi?id=158696

Reviewed by David Kilzer.

No new tests (Covered by all existing tests in Gmalloc/ASAN configs).

* Modules/indexeddb/server/UniqueIDBDatabase.cpp:
(WebCore::IDBServer::UniqueIDBDatabase::executeNextDatabaseTask):
(WebCore::IDBServer::UniqueIDBDatabase::executeNextDatabaseTaskReply):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201997 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoFix round-down goof in Air.js's ShuffleCustom.forEachArg
fpizlo@apple.com [Mon, 13 Jun 2016 18:08:36 +0000 (18:08 +0000)]
Fix round-down goof in Air.js's ShuffleCustom.forEachArg
https://bugs.webkit.org/show_bug.cgi?id=158674

Reviewed by Michael Saboff.

x / 3 * 3 is not how you round down to multiples of 3 in JavaScript. You need to do
Math.floor(x / 3) * 3 instead.

This is a benign change, because having extra arguments to a Shuffle is not something we
actually take advantage of yet. But I think it's best to match the original C++ code's
intent.

* Air.js/custom.js:
(const.ShuffleCustom.forEachArg):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201996 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoModern IDB: IDBOpenDBRequest objects leak.
beidson@apple.com [Mon, 13 Jun 2016 17:32:50 +0000 (17:32 +0000)]
Modern IDB: IDBOpenDBRequest objects leak.
https://bugs.webkit.org/show_bug.cgi?id=158694

Reviewed by Alex Christensen.

No new tests (Currently have no testing strategy for guaranteeing lifetime of WebCore DOM objects)

* Modules/indexeddb/client/IDBConnectionProxy.cpp:
(WebCore::IDBClient::IDBConnectionProxy::completeOpenDBRequest): At this point we never need the
    request again, so remove it from the map.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201995 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoMake sure HTTPHeaderMap gets a move constructor / assignment operator
cdumez@apple.com [Mon, 13 Jun 2016 17:28:29 +0000 (17:28 +0000)]
Make sure HTTPHeaderMap gets a move constructor / assignment operator
https://bugs.webkit.org/show_bug.cgi?id=158695
<rdar://problem/26729511>

Reviewed by Alex Christensen.

Make sure HTTPHeaderMap gets a move constructor / assignment operator.
It was not getting an implicit one because of its user-declared
destructor. This patch drops the user-declared destructor so that
HTTPHeaderMap now gets an implicit move constructor / assignment
operator.

Not having a move constructor / assignment operator is an issue because
we rely on HTTPHeaderMap::isolatedCopy() / WTFMove() since r201623 to
pass HTTPHeaderMap across thread.

* platform/network/HTTPHeaderMap.cpp:
(WebCore::HTTPHeaderMap::~HTTPHeaderMap): Deleted.
* platform/network/HTTPHeaderMap.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201994 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUnreviewed, follow up patch for r201964
utatane.tea@gmail.com [Mon, 13 Jun 2016 16:52:37 +0000 (16:52 +0000)]
Unreviewed, follow up patch for r201964
https://bugs.webkit.org/show_bug.cgi?id=158619

Fix typo in the comment.

* runtime/MathCommon.h:
(JSC::toInt32):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201993 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoRemove useless parameter from GenerateParametersCheck signature
nael.ouedraogo@crf.canon.fr [Mon, 13 Jun 2016 16:46:04 +0000 (16:46 +0000)]
Remove useless parameter from GenerateParametersCheck signature
https://bugs.webkit.org/show_bug.cgi?id=158692

Reviewed by Chris Dumez.

Remove one parameter which is passed to GenerateParametersCheck
but never used in the caller code.

* bindings/scripts/CodeGeneratorJS.pm:
(GenerateImplementation):
(GenerateParametersCheck):
(GenerateConstructorDefinition):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201992 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoStack overflow at RefPtr::release on Windows port since r201782
commit-queue@webkit.org [Mon, 13 Jun 2016 16:27:50 +0000 (16:27 +0000)]
Stack overflow at RefPtr::release on Windows port since r201782
https://bugs.webkit.org/show_bug.cgi?id=158687

Patch by Fujii Hironori <Hironori.Fujii@sony.com> on 2016-06-13
Reviewed by Chris Dumez.

RefPtr::release calls RefPtr::RefPtr, and RefPtr::RefPtr calls
RefPtr::release.

RefPtr::RefPtr does not need to call RefPtr::release.

* wtf/RefPtr.h:
(WTF::RefPtr::RefPtr): Do not call RefPtr::release.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201991 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAdd a mechanism for collecting LLINT stats.
mark.lam@apple.com [Mon, 13 Jun 2016 15:53:42 +0000 (15:53 +0000)]
Add a mechanism for collecting LLINT stats.
https://bugs.webkit.org/show_bug.cgi?id=158668

Reviewed by Filip Pizlo.

Source/JavaScriptCore:

This patch will add a mechanism for collecting the stats on LLINT opcode
execution counts.  The changes made to enable this are:

1. Refactored how Options availability work so that we can add a new category:
   Configurable (in addition to the pre-existing Normal and Restricted
   availability).
       Normal options - always available.
       Restricted options - only available on debug builds.
       Configurable options - depends on #define flag options.

   This change is necessary so that:
   a. we won't have to rebuild the world when we want to enable that #define flag
      to make that Configurable option available.
   b. when the #define flag is disabled, the option will be invisible to the user.

   With this, we add our first configurable option, JSC_reportLLIntStats, which
   is dependent on the ENABLE_LLINT_STATS flag.  See next.

2. Added the ENABLE_LLINT_STATS flag in LLIntCommon.h.  To enable LLINT stats
   collection, we'll need to set this flag to a non-zero value, and rebuilding
   the project.  By design, this will only require a minimal set of files to
   be rebuilt.

   ENABLE_LLINT_STATS is 0 (i.e. disabled) by default.

3. Added a slow path callback to the LLINT's traceExecution() macro, to call
   _llint_count_opcode(), which in turns counts the opcode.  This callback will
   only be built into the LLINT if ENABLE_LLINT_STATS is non-zero.

4. Added s_opcodeStatsArray to LLInt::Data.  This is where the stats are
   recorded and stored.

5. Added calls to LLInt::Data::dumpStats() in jsc.cpp and DumpRenderTree.mm
   to dump the LLINT stats if enabled.  If enabled, the LLINT stats will be
   sorted and dumped (via dataLog) before the programs terminate.

* interpreter/Interpreter.h:
* jsc.cpp:
(main):
* llint/LLIntCommon.h:
* llint/LLIntData.cpp:
(JSC::LLInt::initialize):
(JSC::LLInt::Data::dumpStats):
* llint/LLIntData.h:
(JSC::LLInt::Data::opcodeStats):
* llint/LLIntOfflineAsmConfig.h:
* llint/LLIntSlowPaths.cpp:
(JSC::LLInt::llint_crash):
(JSC::LLInt::LLINT_SLOW_PATH_DECL):
* llint/LLIntSlowPaths.h:
* llint/LowLevelInterpreter.asm:
* runtime/Options.cpp:
(JSC::parse):
(JSC::Options::isAvailable):
(JSC::overrideOptionWithHeuristic):
(JSC::scaleJITPolicy):
(JSC::Options::initialize):
(JSC::Options::setOptionWithoutAlias):
(JSC::Options::dumpAllOptions):
(JSC::Options::dumpOption):
* runtime/Options.h:
(JSC::Option::Option):
(JSC::Option::operator!=):
(JSC::Option::id):

Tools:

* DumpRenderTree/mac/DumpRenderTree.mm:
(DumpRenderTreeMain):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201990 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years ago[GTK] Enabling Shadow DOM by default
commit-queue@webkit.org [Mon, 13 Jun 2016 14:10:18 +0000 (14:10 +0000)]
[GTK] Enabling Shadow DOM by default
https://bugs.webkit.org/show_bug.cgi?id=158686

Patch by Romain Bellessort <romain.bellessort@crf.canon.fr> on 2016-06-13
Reviewed by Carlos Garcia Campos.

.:

Added support for enabling Shadow DOM by default for GTK.

* Source/cmake/OptionsGTK.cmake:

Tools:

Added support for enabling Shadow DOM by default for GTK.
Changed GTK test expectations for fast/shadow-dom tests (most tests
pass).

* Scripts/webkitperl/FeatureList.pm:

LayoutTests:

Changed GTK test expectations for fast/shadow-dom tests (most tests
pass).

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201989 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoImprove code generator for functions with variadic parameters
nael.ouedraogo@crf.canon.fr [Mon, 13 Jun 2016 13:40:26 +0000 (13:40 +0000)]
Improve code generator for functions with variadic parameters
https://bugs.webkit.org/show_bug.cgi?id=158529

Reviewed by Darin Adler.

JS bindings code of functions with variadic parameters is improved.

Functions with variadic parameters are skipped for ObjC and GObject code generators.

* bindings/scripts/CodeGeneratorGObject.pm:
(SkipFunction): Skip functions with variadic parameters.
* bindings/scripts/CodeGeneratorJS.pm:
(GenerateParametersCheck):
* bindings/scripts/CodeGeneratorObjC.pm:
(SkipFunction): Skip functions with variadic parameters.
* bindings/scripts/test/GObject/WebKitDOMTestObj.cpp:
(webkit_dom_test_obj_any): Deleted.
(webkit_dom_test_obj_attach_shadow_root): Deleted.
(webkit_dom_test_obj_get_read_only_long_attr): Deleted.
(webkit_dom_test_obj_get_read_only_string_attr): Deleted.
* bindings/scripts/test/GObject/WebKitDOMTestObj.h:
* bindings/scripts/test/JS/JSTestObj.cpp:
(WebCore::jsTestObjPrototypeFunctionOverloadedMethod12):
(WebCore::jsTestObjPrototypeFunctionVariadicStringMethod):
(WebCore::jsTestObjPrototypeFunctionVariadicDoubleMethod):
(WebCore::jsTestObjPrototypeFunctionVariadicNodeMethod):
* bindings/scripts/test/ObjC/DOMTestObj.h:
* bindings/scripts/test/ObjC/DOMTestObj.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201988 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUnreviewed test gardening.
peavo@outlook.com [Mon, 13 Jun 2016 09:06:43 +0000 (09:06 +0000)]
Unreviewed test gardening.

Unskip passing fast/workers tests.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201987 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agorun-safari/run-webkit-app fail to quit iOS simulator after Xcode installation
aakash_jain@apple.com [Mon, 13 Jun 2016 03:57:23 +0000 (03:57 +0000)]
run-safari/run-webkit-app fail to quit iOS simulator after Xcode installation
https://bugs.webkit.org/show_bug.cgi?id=158651
rdar://problem/26499824

Reviewed by Daniel Bates.

This change is to make our tools more robust. Sometimes run-safari/run-webkit-app fails
to start the tests as one of the first thing these scripts do is to relaunch iOS Simulators,
and in process call quitIOSSimulator. In case the Simulator is not registered in LaunchServices
for some reason, quitIOSSimulator fails and the script stops. This change will make the scripts
still continue when they fail to quit iOS Simuator, and modify the way Simulator is opened,
so that Simulator is implicitly registered with LaunchServices while opening it.

* Scripts/webkitdirs.pm:
(relaunchIOSSimulator):
Launch iOS Simulator using complete path. Partially reverting http://trac.webkit.org/changeset/184202.
(quitIOSSimulator):
Do not die if quitting ios Simulator fails, so that relaunchIOSSimulator can still
attempt to launch the simulator.
(waitUntilIOSSimulatorDeviceIsInState):
Added a FIXME.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201986 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoCleanup RenderBlock::removePositionedObjects
zalan@apple.com [Sun, 12 Jun 2016 20:57:11 +0000 (20:57 +0000)]
Cleanup RenderBlock::removePositionedObjects
https://bugs.webkit.org/show_bug.cgi?id=158670

Reviewed by Simon Fraser.

No change in functionality.

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::insertPositionedObject):
(WebCore::RenderBlock::removePositionedObject):
(WebCore::RenderBlock::removePositionedObjects):
* rendering/RenderBlock.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201985 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoRemove positioned descendants when RenderBlock is no longer a containing block.
zalan@apple.com [Sun, 12 Jun 2016 20:45:44 +0000 (20:45 +0000)]
Remove positioned descendants when RenderBlock is no longer a containing block.
https://bugs.webkit.org/show_bug.cgi?id=158655
<rdar://problem/26510032>

Reviewed by Simon Fraser.

Normally the RenderView is the containing block for fixed positioned renderers.
However when a renderer acquires some transform related properties, it becomes the containing
block for all the fixed positioned renderers in its descendant tree.
When the last transform related property is removed, the renderer is no longer a containing block
and we need to remove all these positioned renderers from the descendant tracker map (gPositionedDescendantsMap).
They will be inserted back into the tracker map during the next layout (either under the RenderView or
under the next transformed renderer in the ancestor chain).

Source/WebCore:

Test: fast/block/fixed-position-reparent-when-transition-is-removed.html

* rendering/RenderBlock.cpp:
(WebCore::RenderBlock::removePositionedObjectsIfNeeded):

LayoutTests:

* fast/block/fixed-position-reparent-when-transition-is-removed-expected.txt: Added.
* fast/block/fixed-position-reparent-when-transition-is-removed.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201984 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoCrash in com.apple.WebKit.WebContent at std::__1::__call_once_proxy<std::__1::tuple...
ddkilzer@apple.com [Sun, 12 Jun 2016 13:22:53 +0000 (13:22 +0000)]
Crash in com.apple.WebKit.WebContent at std::__1::__call_once_proxy<std::__1::tuple<CrashReporterSupportLibrary()::$_0&&> >
<https://webkit.org/b/158660>
<rdar://problem/25652686>

Reviewed by Darin Adler.

* bmalloc/Logging.cpp: Switch to use
BSOFT_LINK_PRIVATE_FRAMEWORK() to link
CrashReporterSupport.framework.
* bmalloc/darwin/BSoftLinking.h:
(BSOFT_LINK_PRIVATE_FRAMEWORK): Rename from BSOFT_LINK_FRAMEWORK.
Switch to use /System/Library/PrivateFrameworks/.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201983 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAddressing post-review comments after r201978.
mmaxfield@apple.com [Sun, 12 Jun 2016 01:14:58 +0000 (01:14 +0000)]
Addressing post-review comments after r201978.
https://bugs.webkit.org/show_bug.cgi?id=158649
<rdar://problem/13258122>

Unreviewed.

Source/WebCore:

* platform/graphics/FontCache.cpp:
(WebCore::FontCache::alternateFamilyName):
* platform/graphics/cocoa/FontCacheCoreText.cpp:
(WebCore::FontCache::platformAlternateFamilyName):

Source/WTF:

* wtf/text/StringCommon.h:
(WTF::equal):
(WTF::naiveEqualWithoutPerformingUnicodeNormalization): Deleted.

Tools:

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201982 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUnreviewed leak fix after r201863.
ap@apple.com [Sat, 11 Jun 2016 23:44:50 +0000 (23:44 +0000)]
Unreviewed leak fix after r201863.

* DumpRenderTree/mac/DumpRenderTree.mm:
(dumpFramesAsText):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201981 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoTighten code to build set of tag names
darin@apple.com [Sat, 11 Jun 2016 23:35:23 +0000 (23:35 +0000)]
Tighten code to build set of tag names
https://bugs.webkit.org/show_bug.cgi?id=158662

Reviewed by Alexey Proskuryakov.

* dom/Element.cpp:
(WebCore::canAttachAuthorShadowRoot): Use an array of pointers that the loader
can initialize as part of loading the library, rather than an array that needs
to be initialized with code at runtime.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201980 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years ago[Win] [EFL] Build fix after r201978.
mmaxfield@apple.com [Sat, 11 Jun 2016 22:55:40 +0000 (22:55 +0000)]
[Win] [EFL] Build fix after r201978.
https://bugs.webkit.org/show_bug.cgi?id=158649
<rdar://problem/13258122>

Unreviewed

* platform/graphics/freetype/FontCacheFreeType.cpp:
(WebCore::FontCache::platformAlternateFamilyName):
* platform/graphics/win/FontCacheWin.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201979 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years ago[Cocoa] Map commonly used Chinese Windows font names to names present on Cocoa operat...
mmaxfield@apple.com [Sat, 11 Jun 2016 21:48:15 +0000 (21:48 +0000)]
[Cocoa] Map commonly used Chinese Windows font names to names present on Cocoa operating systems
https://bugs.webkit.org/show_bug.cgi?id=158649
<rdar://problem/13258122>

Reviewed by Darin Adler.

Source/WebCore:

There are many Chinese websites which hardcode Windows font names.
We should map these to fonts which best match them on Cocoa operating
systems. We can do this by using our existing fallback font name
infrastructure.

Tests: fast/text/chinese-font-name-aliases-2.html
       fast/text/chinese-font-name-aliases.html

* platform/graphics/FontCache.cpp:
(WebCore::FontCache::alternateFamilyName):
(WebCore::alternateFamilyName): Deleted.
* platform/graphics/FontCache.h:
* platform/graphics/cocoa/FontCacheCoreText.cpp:
(WebCore::FontCache::platformAlternateFamilyName):
* platform/graphics/freetype/FontCacheFreeType.cpp:
(WebCore::FontCache::platformAlternateFamilyName):
* platform/graphics/win/FontCacheWin.cpp:
(WebCore::FontCache::platformAlternateFamilyName):

Source/WTF:

* wtf/text/StringCommon.h:
(WTF::naiveEqualWithoutPerformingUnicodeNormalization): Added.

Tools:

* TestWebKitAPI/Tests/WTF/WTFString.h:
(TestWebKitAPI::TEST):

LayoutTests:

Because this tests platform-specific font names, the test must
be platform-specific. Because it's a ref test, make two tests which
are mutually-exclusive on Cocoa platforms, and expected to fail
everywhere else.

* fast/text/chinese-font-name-aliases-2-expected.html: Added.
* fast/text/chinese-font-name-aliases-2.html: Added.
* fast/text/chinese-font-name-aliases-expected.html: Added.
* fast/text/chinese-font-name-aliases.html: Added.
* platform/efl/TestExpectations:
* platform/gtk/TestExpectations:
* platform/ios-simulator/TestExpectations:
* platform/mac/TestExpectations:
* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201978 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUnreviewed, rolling out r201967, r201968, and r201972.
commit-queue@webkit.org [Sat, 11 Jun 2016 20:05:55 +0000 (20:05 +0000)]
Unreviewed, rolling out r201967, r201968, and r201972.
https://bugs.webkit.org/show_bug.cgi?id=158665

Caused flaky failures on IndexedDB tests (Requested by ap on
#webkit).

Reverted changesets:

"Vary:Cookie validation doesn't work in private browsing"
https://bugs.webkit.org/show_bug.cgi?id=158616
http://trac.webkit.org/changeset/201967

"Build fix."
http://trac.webkit.org/changeset/201968

"WinCairo build fix attempt."
http://trac.webkit.org/changeset/201972

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201977 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoMinimize the amount of memcpy done for allocating Error stacks.
mark.lam@apple.com [Sat, 11 Jun 2016 19:58:07 +0000 (19:58 +0000)]
Minimize the amount of memcpy done for allocating Error stacks.
https://bugs.webkit.org/show_bug.cgi?id=158664

Reviewed by Darin Adler.

Currently, Vector<StackFrame> are being copied around multiple times in the
process of creating Error stacks.

This patch avoids this unnecessary copying by:
1. Sizing the StackFrame vector correctly to begin with, and skipping
   undesirable top frames before filling in the vector.
2. Using perfect forwarding or passing by reference to pass the vector data around
   instead of copying the vectors.
3. Changing the Exception object to take a Vector<StackFrame> instead of a
   RefCountedArray<StackFrame>.

This patch has passed the JSC and layout tests.  Benchmarks show that perf is
neutral.

* API/tests/testapi.mm:
(testObjectiveCAPI):
* inspector/ScriptCallStackFactory.cpp:
(Inspector::createScriptCallStackFromException):
* interpreter/Interpreter.cpp:
(JSC::GetStackTraceFunctor::GetStackTraceFunctor):
(JSC::GetStackTraceFunctor::operator()):
(JSC::Interpreter::getStackTrace):
(JSC::Interpreter::stackTraceAsString):
(JSC::findExceptionHandler):
* interpreter/Interpreter.h:
* runtime/Error.cpp:
(JSC::addErrorInfoAndGetBytecodeOffset):
* runtime/Exception.cpp:
(JSC::Exception::finishCreation):
* runtime/Exception.h:
(JSC::Exception::valueOffset):
(JSC::Exception::value):
(JSC::Exception::stack):
(JSC::Exception::didNotifyInspectorOfThrow):
(JSC::Exception::setDidNotifyInspectorOfThrow):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201976 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUnreviewed, skip Air.js. It got unskipped when I renamed it from JSAir.
fpizlo@apple.com [Sat, 11 Jun 2016 19:48:48 +0000 (19:48 +0000)]
Unreviewed, skip Air.js. It got unskipped when I renamed it from JSAir.

* Skipped:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201975 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoTests that overflows the stack should not be run with the sampling profiler.
mark.lam@apple.com [Sat, 11 Jun 2016 18:53:58 +0000 (18:53 +0000)]
Tests that overflows the stack should not be run with the sampling profiler.
https://bugs.webkit.org/show_bug.cgi?id=158663

Reviewed by Saam Barati.

The sampling profiler will be sampling the whole stack, and the amount of memory
churn will make this tests time out, especially with debug builds.  Hence,
let's not run the test with the sampling profiler configuration.

* tests/stress/mutual-tail-call-no-stack-overflow.js:
(shouldThrow):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201974 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoFixed compilation of LocaleICU with ENABLE(DATE_AND_TIME_INPUT_TYPES)
annulen@yandex.ru [Sat, 11 Jun 2016 18:07:05 +0000 (18:07 +0000)]
Fixed compilation of LocaleICU with ENABLE(DATE_AND_TIME_INPUT_TYPES)
https://bugs.webkit.org/show_bug.cgi?id=158659

Reviewed by Darin Adler.

No new tests needed.

* platform/text/LocaleICU.cpp:
(WebCore::getFormatForSkeleton):
(WebCore::LocaleICU::monthFormat):
(WebCore::LocaleICU::shortMonthFormat):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201973 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWinCairo build fix attempt.
antti@apple.com [Sat, 11 Jun 2016 17:58:51 +0000 (17:58 +0000)]
WinCairo build fix attempt.

* platform/network/NetworkStorageSession.cpp:
* platform/network/NetworkStorageSession.h:
* platform/network/NetworkStorageSessionStub.cpp:
(WebCore::NetworkStorageSession::NetworkStorageSession):
(WebCore::NetworkStorageSession::context):
(WebCore::NetworkStorageSession::createPrivateBrowsingSession):
(WebCore::NetworkStorageSession::switchToNewTestingSession):
(WebCore::NetworkStorageSession::~NetworkStorageSession): Deleted.
(WebCore::defaultSession): Deleted.
(WebCore::NetworkStorageSession::defaultStorageSession): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201972 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoDeleting a CSSOM style rule invalidates any previously-added FontFaces
mmaxfield@apple.com [Sat, 11 Jun 2016 17:52:35 +0000 (17:52 +0000)]
Deleting a CSSOM style rule invalidates any previously-added FontFaces
https://bugs.webkit.org/show_bug.cgi?id=158450

Reviewed by Darin Adler.

Source/WebCore:

This patch has two pieces: updating the CSSOM when the FontFace changes, and
updating the FontFace when the CSSOM changes.

1: Updating the CSSOM when the FontFace changes: CSSFontFaces already have a RefPtr
to their StyleRuleFontFace which represents their CSS-connection. When changing a
property of the CSSFontFace, we simply reach into the StyleRule and update it to
match. Our existing infrastructure of invalidation due to the attribute changes
makes sure that all the necessary updates occur.

2. Updating the FontFace when the CSSOM changes: If the CSSOM changes in a trivial
way (for example, a new @font-face is appended to the end of the last <style>
element), we can handle it directly. However, when something more invasive occurs,
we end up clearing the entire CSSFontSelector, and then adding all the style rules
from scratch. This involves three steps:
    a) CSSFontSelector::buildStarted() is run, which means "we're about to start
       building up all the @font-face rules from scratch." We take this opportunity
       to purge as many fonts as possible. This is valuable because, for example,
       this function gets run when the page gets put into the page cache, so we
       want to destroy as much as possible. Not everything can be purged, however -
       only CSS-connected fonts which have never been inspected by script are
       purgeable. We don't allow fonts inspected by script to be purged because
       purging might result in a font appearing from JavaScript to transition from
       a success -> failure state, which we don't allow.
    b) Upon style recalc (possibly asynchronously) CSSFontSelector::addFontFaceRule()
       is called for each @font-face rule. We actually detect that we're in the
       middle of a style rebuild, and defer this step.
    c) When we're done adding all the font face rules, we call
       CSSFontSelector::buildCompleted(). This is where we compare the newly built-
       up list of font faces with what existed previously (as remembered in
       CSSFontSelector::buildStarted()) in order to detect font faces which were
       deleted from the document. Fonts which were newly added to the document
       are handled naturally.
       Fonts which have a property modified on them are created as if they were new.
       However, instead of simply adding the CSSFontFace, we search for the existing
       CSSFontFace (by CSS connection pointer) and tell the existing FontFace to
       adopt this new CSSFontFace. This means that the JavaScript object will just
       pick up any newly-written values in the CSSOM. It also means that the
       "status" attribute of the JavaScript object is reset, but this is expected
       and allowed by the spec. (For example, if you change the "src" attribute of
       an @font-face block via the CSSOM, all bets are off when you inspect the
       FontFace JS object representing that block.)

Test: fast/text/font-face-set-cssom.html

* css/CSSFontFace.cpp:
(WebCore::CSSFontFace::CSSFontFace):
(WebCore::CSSFontFace::setFamilies):
(WebCore::CSSFontFace::setStyle):
(WebCore::CSSFontFace::setWeight):
(WebCore::CSSFontFace::setUnicodeRange):
(WebCore::CSSFontFace::setVariantLigatures):
(WebCore::CSSFontFace::setVariantPosition):
(WebCore::CSSFontFace::setVariantCaps):
(WebCore::CSSFontFace::setVariantNumeric):
(WebCore::CSSFontFace::setVariantAlternates):
(WebCore::CSSFontFace::setVariantEastAsian):
(WebCore::CSSFontFace::setFeatureSettings):
(WebCore::CSSFontFace::initializeWrapper):
(WebCore::CSSFontFace::wrapper):
(WebCore::CSSFontFace::setWrapper):
(WebCore::CSSFontFace::purgeable):
(WebCore::CSSFontFace::updateStyleIfNeeded):
* css/CSSFontFace.h:
* css/CSSFontFaceSet.cpp:
(WebCore::CSSFontFaceSet::remove):
(WebCore::CSSFontFaceSet::containsCSSConnection):
(WebCore::CSSFontFaceSet::purge):
* css/CSSFontFaceSet.h:
* css/CSSFontSelector.cpp:
(WebCore::CSSFontSelector::buildStarted):
(WebCore::CSSFontSelector::buildCompleted):
(WebCore::CSSFontSelector::addFontFaceRule):
* css/CSSFontSelector.h:
* css/FontFace.cpp:
(WebCore::FontFace::family):
(WebCore::FontFace::style):
(WebCore::FontFace::weight):
(WebCore::FontFace::unicodeRange):
(WebCore::FontFace::variant):
(WebCore::FontFace::featureSettings):
(WebCore::FontFace::adopt):
* css/FontFace.h:

LayoutTests:

* fast/text/font-face-set-cssom-expected.txt: Added.
* fast/text/font-face-set-cssom.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201971 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWorkerNavigator is missing some attributes
cdumez@apple.com [Sat, 11 Jun 2016 17:44:20 +0000 (17:44 +0000)]
WorkerNavigator is missing some attributes
https://bugs.webkit.org/show_bug.cgi?id=158593
<rdar://problem/26731334>

Reviewed by Darin Adler.

Source/WebCore:

Add attributes that are missing on WorkerNavigator:
- appCodeName
- hardwareConcurrency
- language
- product
- productSub
- vendor
- vendorSub

Firefox and Chrome already expose those attributes.

Relevant specification:
https://html.spec.whatwg.org/multipage/workers.html#the-workernavigator-object

This patch also refactors the IDL to match the specification more
closely and promote sharing between Navigator and WorkerNavigator.

No new tests, updated existing test.

* CMakeLists.txt:
* DerivedSources.make:
Add new supplemental IDL files.

* page/Navigator.cpp:
* page/Navigator.h:
Moved language() / hardwareConcurrency() from Navigator to NavigatorBase
so that it can be used by NavigatorWorker as well.

* page/NavigatorBase.h:
* page/NavigatorBase.cpp:
(WebCore::NavigatorBase::language):
The implementation still calls defaultLanguage() but I updated it to be
thread safe on all platforms.

(WebCore::NavigatorBase::hardwareConcurrency):
Use std::call_once() for thread safety.

* page/Navigator.idl:
* page/NavigatorConcurrentHardware.idl: Copied from Source/WebCore/page/WorkerNavigator.idl.
* page/NavigatorID.idl: Copied from Source/WebCore/page/WorkerNavigator.idl.
* page/NavigatorLanguage.idl: Copied from Source/WebCore/page/WorkerNavigator.idl.
* page/NavigatorOnLine.idl: Copied from Source/WebCore/page/WorkerNavigator.idl.
* page/WorkerNavigator.idl:
Move several attributes to their own supplemental interfaces to match
the specification and promote sharing with WorkerNavigator.

* platform/Language.cpp:
(WebCore::userPreferredLanguages):
* platform/Language.h:
Made thread-safe on all platforms.

Source/WTF:

* wtf/PlatformUserPreferredLanguages.h:
* wtf/PlatformUserPreferredLanguagesWin.cpp:
(WTF::platformLanguage):
(WTF::platformUserPreferredLanguages):

Make platformUserPreferredLanguages() thread safe on Windows. The
Mac and Unix implementations are already thread-safe.

LayoutTests:

Improved test coverage for WorkerNavigator.

* fast/workers/resources/worker-navigator.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201970 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoImplement logging for RELEASE_BASSERT_WITH_MESSAGE() in BAssert.h
ddkilzer@apple.com [Sat, 11 Jun 2016 11:46:19 +0000 (11:46 +0000)]
Implement logging for RELEASE_BASSERT_WITH_MESSAGE() in BAssert.h
<http://webkit.org/b/155992>

Reviewed by Geoff Garen.

* bmalloc/BAssert.h:
(BLOG_ERROR): Add method to always log error messages.
(RELEASE_BASSERT_WITH_MESSAGE): Use BLOG_ERROR() to implement
logging in Debug builds.
* bmalloc/BPlatform.h:
(BPLATFORM_MAC): Add.
(BUSE): Add BUSE() macro.
(BATTRIBUTE_PRINTF): Add.
(BUSE_OS_LOG): Add.
* bmalloc/Logging.cpp:
(bmalloc::reportAssertionFailureWithMessage): Add.  Logs to
stderr.
* bmalloc/Logging.h:
(bmalloc::reportAssertionFailureWithMessage): Add declaration.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201969 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoBuild fix.
antti@apple.com [Sat, 11 Jun 2016 11:20:23 +0000 (11:20 +0000)]
Build fix.

* platform/network/cf/NetworkStorageSessionCFNet.cpp:
(WebCore::NetworkStorageSession::switchToNewTestingSession):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201968 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoVary:Cookie validation doesn't work in private browsing
antti@apple.com [Sat, 11 Jun 2016 11:09:54 +0000 (11:09 +0000)]
Vary:Cookie validation doesn't work in private browsing
https://bugs.webkit.org/show_bug.cgi?id=158616
Source/WebCore:

rdar://problem/26755067

Reviewed by Darin Adler.

This wasn't implemented because there was no way to get NetworkStorageSession from
a SessionID on WebCore side.

The patch adds a simple WebCore level weak map that allows getting NetworkStorageSessions
from SessionID. This seemed like the cleanest way to do this without a big refactoring
around the currently WebKit2 level SessionTracker.

* CMakeLists.txt:
* WebCore.xcodeproj/project.pbxproj:
* platform/network/CacheValidation.cpp:
(WebCore::headerValueForVary):

    Get NetworkStorageSession from SessionID for cookies

(WebCore::verifyVaryingRequestHeaders):
* platform/network/NetworkStorageSession.cpp: Added.

    Add platform independent .cpp for NetworkStorageSession.
    Implement a weak map for SessionID -> NetworkStorageSession.

(WebCore::sessionsMap):
(WebCore::NetworkStorageSession::NetworkStorageSession):
(WebCore::NetworkStorageSession::~NetworkStorageSession):
(WebCore::NetworkStorageSession::forSessionID):

    Get NetworkStorageSession for sessionID.

* platform/network/NetworkStorageSession.h:
(WebCore::NetworkStorageSession::sessionID):
(WebCore::NetworkStorageSession::credentialStorage):
* platform/network/cf/NetworkStorageSessionCFNet.cpp:
(WebCore::NetworkStorageSession::NetworkStorageSession):

    Call to common constructor.

(WebCore::defaultNetworkStorageSession):
* platform/network/soup/NetworkStorageSessionSoup.cpp:
(WebCore::NetworkStorageSession::NetworkStorageSession):

    Call to common constructor.

(WebCore::defaultSession):
(WebCore::NetworkStorageSession::~NetworkStorageSession): Deleted.

LayoutTests:

Reviewed by Darin Adler.

* http/tests/cache/disk-cache/disk-cache-vary-cookie-expected.txt:
* http/tests/cache/disk-cache/disk-cache-vary-cookie.html:

Exapand the existing test to cover memory cache and private browsing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201967 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUnreviewed, attempt to fix r201964 failure on Apple ports
utatane.tea@gmail.com [Sat, 11 Jun 2016 06:01:48 +0000 (06:01 +0000)]
Unreviewed, attempt to fix r201964 failure on Apple ports
https://bugs.webkit.org/show_bug.cgi?id=158619

Reviewed by Mark Lam.

Add Private attributes to MathCommon.h.

* JavaScriptCore.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201966 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWeb Inspector: <template> content should not be hidden as Shadow Content
commit-queue@webkit.org [Sat, 11 Jun 2016 05:32:44 +0000 (05:32 +0000)]
Web Inspector: <template> content should not be hidden as Shadow Content
https://bugs.webkit.org/show_bug.cgi?id=158654

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-06-10
Reviewed by Ryosuke Niwa.

* Localizations/en.lproj/localizedStrings.js:
New "Template Content" string.

* UserInterface/Models/DOMNode.js:
(WebInspector.DOMNode):
Don't treat template content as shadow content.

* UserInterface/Views/DOMTreeElement.js:
(WebInspector.DOMTreeElement.prototype._nodeTitleInfo):
Custom string for template content fragment.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201965 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years ago[JSC] Inline JSC::toInt32 to improve kraken
utatane.tea@gmail.com [Sat, 11 Jun 2016 05:31:48 +0000 (05:31 +0000)]
[JSC] Inline JSC::toInt32 to improve kraken
https://bugs.webkit.org/show_bug.cgi?id=158619

Reviewed by Mark Lam.

Several kraken benchmarks show that JSC::toInt32 is frequently called.
For example, stanford-crypto-pbkdf2 reports that the hottest runtime function is JSC::toInt32.

The data is below (taken by Linux perf tools).
5.50%  jsc      libJavaScriptCore.so.1.0.0  [.] _ZN3JSC7toInt32Ed
3.96%  jsc      libJavaScriptCore.so.1.0.0  [.] _ZN3JSC20arrayProtoFuncConcatEPNS_9ExecStateE
2.48%  jsc      libJavaScriptCore.so.1.0.0  [.] _ZN3JSC19arrayProtoFuncSliceEPNS_9ExecStateE
1.69%  jsc      libJavaScriptCore.so.1.0.0  [.] _ZNK3JSC9Structure27holesMustForwardToPrototypeERNS_2VME

This is because of CommonSlowPaths' bit operations's JSValue::toInt32.
Due to the slow path, in `value | 0`, `value` may be a double number value. In that case, JSC::toInt32 is called.

While JSC::toIn32 is hot, the function itself is very small. It's worth inlining.

This change offers the following kraken improvements.

                                                 baseline                  patched
Kraken:
   audio-beat-detection                       47.492+-1.701             46.657+-1.232           might be 1.0179x faster
   stanford-crypto-aes                        43.669+-0.210      ^      42.862+-0.115         ^ definitely 1.0188x faster
   stanford-crypto-ccm                        45.213+-1.424             44.490+-1.290           might be 1.0162x faster
   stanford-crypto-pbkdf2                    107.665+-0.581      ^     106.229+-0.807         ^ definitely 1.0135x faster

This patch only focused on the call to toInt32 from the runtime functions.
So JSC::toInt32 calls from the baseline / DFG remain.
We ensure that JIT code uses operationToInt32 instead of JSC::toInt32 since JSC::toInt32 is now marked as ALWAYS_INLINE.
Linux perf profiler also finds that this `operationToInt32` is frequently called in the above benchmarks.
It may be good to introduce asm emit for that instead of calling JSC::toInt32 operation in the separated patch.

* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileValueToInt32):
(JSC::DFG::SpeculativeJIT::compilePutByValForIntTypedArray):
* ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::LowerDFGToB3::doubleToInt32):
(JSC::FTL::DFG::LowerDFGToB3::sensibleDoubleToInt32):
* runtime/JSCJSValue.cpp:
(JSC::toInt32): Deleted.
* runtime/JSCJSValueInlines.h:
* runtime/MathCommon.cpp:
(JSC::operationToInt32):
* runtime/MathCommon.h:
(JSC::toInt32):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201964 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUse the video element's video box when getting the inline video rect in WebVideoFulls...
adachan@apple.com [Sat, 11 Jun 2016 04:37:06 +0000 (04:37 +0000)]
Use the video element's video box when getting the inline video rect in WebVideoFullscreenManager
https://bugs.webkit.org/show_bug.cgi?id=158351
<rdar://problem/26567938>

Reviewed by Darin Adler.

Source/WebCore:

* WebCore.xcodeproj/project.pbxproj:
Change the visibility of RenderVideo.h and RenderMedia.h since we'll be importing RenderVideo.h from WebKit2.
* rendering/RenderVideo.h:

Source/WebKit2:

* WebProcess/cocoa/WebVideoFullscreenManager.mm:
(WebKit::inlineVideoFrame):
(WebKit::WebVideoFullscreenManager::enterVideoFullscreenForVideoElement):
(WebKit::WebVideoFullscreenManager::exitVideoFullscreenForVideoElement):
(WebKit::WebVideoFullscreenManager::fullscreenMayReturnToInline):
(WebKit::WebVideoFullscreenManager::setVideoLayerFrameFenced):
(WebKit::clientRectForElement): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201963 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoFix CMake build.
achristensen@apple.com [Sat, 11 Jun 2016 04:24:27 +0000 (04:24 +0000)]
Fix CMake build.

* PlatformMac.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201962 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoMake it easy to package Air.js
fpizlo@apple.com [Sat, 11 Jun 2016 01:52:58 +0000 (01:52 +0000)]
Make it easy to package Air.js
https://bugs.webkit.org/show_bug.cgi?id=158652

Reviewed by Benjamin Poulain.

If you want to give Air.js to someone, you can now do:

cd PerformanceTests/Air.js
./make_dist.sh

This will create Air.js.tar.gz and a directory called Air.js. If you want to send someone
the benchmark, just send either of those things and tell them to look at index.html.

You may have to edit make_dist.sh for wherever you have Markdown.pl. You can get that from
https://daringfireball.net/projects/markdown/

* Air.js/make_dist.sh: Added.
* Air.js/test.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201961 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoUnreviewed, make the link to test.html into an actual link.
fpizlo@apple.com [Sat, 11 Jun 2016 01:32:47 +0000 (01:32 +0000)]
Unreviewed, make the link to test.html into an actual link.

* Air.js/README.md:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201960 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoFix some text in Air.js/README.md
fpizlo@apple.com [Sat, 11 Jun 2016 01:20:21 +0000 (01:20 +0000)]
Fix some text in Air.js/README.md
https://bugs.webkit.org/show_bug.cgi?id=158650

Reviewed by Benjamin Poulain.

I read the text again and found bugs:

- We never actually say how to run the benchmark. This change adds a blurb about how to run
  it.

- We both say that allocateStack is responsible for the bulk of the running time and that
  we haven't measured where the bulk of the time is spent. This changes the text to say that
  it was a goal to make allocateStack be the hottest part of the benchmark, but that we did
  not measure this.

* Air.js/README.md:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201959 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAdd support for passive event listeners on touch events
benjamin@webkit.org [Sat, 11 Jun 2016 01:18:37 +0000 (01:18 +0000)]
Add support for passive event listeners on touch events
https://bugs.webkit.org/show_bug.cgi?id=158601

Patch by Benjamin Poulain <bpoulain@apple.com> on 2016-06-10
Reviewed by Simon Fraser.

Source/WebCore:

This patch wires "passive" state of EventTarget to the delivery of touch
events in WebKit2.

Instead of having a NonFastScrollableRegion, we have a pair of regions
in EventTrackingRegions.
The "asynchronousDispatchRegion" tracks the area for which all event
listeners are passive. For those, events should be dispatched asynchronously.
The "synchronousDispatchRegion" tracks the area for which there is at
least one active event listener. Events have to be dispatched synchronously
for correctness.

Tests: fast/events/touch/ios/tap-with-active-listener-on-elements.html
       fast/events/touch/ios/tap-with-active-listener-on-window.html
       fast/events/touch/ios/tap-with-passive-listener-on-elements.html
       fast/events/touch/ios/tap-with-passive-listener-on-window.html

* WebCore.xcodeproj/project.pbxproj:
* dom/Document.cpp:
(WebCore::Document::wheelEventHandlersChanged):
(WebCore::Document::Document): Deleted.
* dom/Document.h:

* dom/EventListenerMap.cpp:
(WebCore::EventListenerMap::containsActive):
If a Target has multiple listener for an event type, we want to know
if any of them is active.

* dom/EventListenerMap.h:
* dom/EventTarget.cpp:
(WebCore::EventTarget::hasActiveEventListeners):
(WebCore::EventTarget::hasActiveTouchEventListeners):
* dom/EventTarget.h:

* page/DebugPageOverlays.cpp:
(WebCore::NonFastScrollableRegionOverlay::updateRegion):
I did not change the debug overlays.
The NonFastScrollable area is the region for which events needs
synchronous dispatch. Everything else should scroll without delay.

* page/FrameView.cpp:
(WebCore::FrameView::scrollableAreaSetChanged):
* page/Page.cpp:
(WebCore::Page::nonFastScrollableRects):
* page/scrolling/AsyncScrollingCoordinator.cpp:
(WebCore::AsyncScrollingCoordinator::setEventTrackingRegionsDirty):
(WebCore::AsyncScrollingCoordinator::willCommitTree):
(WebCore::AsyncScrollingCoordinator::updateEventTrackingRegions):
(WebCore::AsyncScrollingCoordinator::frameViewLayoutUpdated):
(WebCore::AsyncScrollingCoordinator::frameViewEventTrackingRegionsChanged):
(WebCore::AsyncScrollingCoordinator::scrollingStateTreeAsText):
(WebCore::AsyncScrollingCoordinator::setNonFastScrollableRegionDirty): Deleted.
(WebCore::AsyncScrollingCoordinator::updateNonFastScrollableRegion): Deleted.
(WebCore::AsyncScrollingCoordinator::frameViewNonFastScrollableRegionChanged): Deleted.
* page/scrolling/AsyncScrollingCoordinator.h:
(WebCore::AsyncScrollingCoordinator::eventTrackingRegionsDirty):
(WebCore::AsyncScrollingCoordinator::nonFastScrollableRegionDirty): Deleted.

* page/scrolling/ScrollingCoordinator.cpp:
(WebCore::ScrollingCoordinator::absoluteEventTrackingRegionsForFrame):
(WebCore::ScrollingCoordinator::absoluteEventTrackingRegions):
(WebCore::ScrollingCoordinator::absoluteNonFastScrollableRegionForFrame): Deleted.
(WebCore::ScrollingCoordinator::absoluteNonFastScrollableRegion): Deleted.
I intentionally left the Wheel event with synchronous dispatch.
This use case will need its own set of tests.

* page/scrolling/ScrollingCoordinator.h:
(WebCore::ScrollingCoordinator::frameViewEventTrackingRegionsChanged):
(WebCore::ScrollingCoordinator::frameViewNonFastScrollableRegionChanged): Deleted.
* page/scrolling/ScrollingStateFrameScrollingNode.cpp:
(WebCore::ScrollingStateFrameScrollingNode::ScrollingStateFrameScrollingNode):
(WebCore::ScrollingStateFrameScrollingNode::setEventTrackingRegions):
(WebCore::ScrollingStateFrameScrollingNode::dumpProperties):
(WebCore::ScrollingStateFrameScrollingNode::setNonFastScrollableRegion): Deleted.
* page/scrolling/ScrollingStateFrameScrollingNode.h:
* page/scrolling/ScrollingTree.cpp:
(WebCore::ScrollingTree::shouldHandleWheelEventSynchronously):
(WebCore::ScrollingTree::commitNewTreeState):
(WebCore::ScrollingTree::eventTrackingTypeForPoint):
(WebCore::ScrollingTree::isPointInNonFastScrollableRegion): Deleted.
* page/scrolling/ScrollingTree.h:
* page/scrolling/mac/ScrollingCoordinatorMac.mm:
(WebCore::ScrollingCoordinatorMac::scheduleTreeStateCommit):
* platform/EventTrackingRegions.h: Added.
(WebCore::EventTrackingRegions::isEmpty):
(WebCore::EventTrackingRegions::trackingTypeForPoint):
(WebCore::operator==):

Source/WebKit2:

* Shared/Scrolling/RemoteScrollingCoordinatorTransaction.cpp:
(ArgumentCoder<ScrollingStateFrameScrollingNode>::encode):
(ArgumentCoder<ScrollingStateFrameScrollingNode>::decode):
(WebKit::dump):
* Shared/WebCoreArgumentCoders.cpp:
(IPC::ArgumentCoder<EventTrackingRegions>::encode):
(IPC::ArgumentCoder<EventTrackingRegions>::decode):
* Shared/WebCoreArgumentCoders.h:
* UIProcess/Scrolling/RemoteScrollingCoordinatorProxy.cpp:
(WebKit::RemoteScrollingCoordinatorProxy::eventTrackingTypeForPoint):
(WebKit::RemoteScrollingCoordinatorProxy::isPointInNonFastScrollableRegion): Deleted.
* UIProcess/Scrolling/RemoteScrollingCoordinatorProxy.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::touchEventTrackingType):
The type is the most restrictive type of all the touch points.
If any touch point needs synchronous tracking, everything has to be synchronous.

(WebKit::WebPageProxy::handleTouchEventSynchronously):
This is a very unfortunate design.

The upper layer dealing with UIKit knows about gesture and when something
can be dispatched asynchrnously because a gesture has started.
This layer knows about tracking and can use that extra inforamtion
to dispatch an event asynchrnously anyway.

In the future, we should refactor this to let WebPageProxy
always decide of the dispatch type. I am just concerned changing
the upper layer while we are stabilizing WebKit.

(WebKit::WebPageProxy::handleTouchEventAsynchronously):
(WebKit::WebPageProxy::handleTouchEvent):
(WebKit::WebPageProxy::resetState):
(WebKit::WebPageProxy::WebPageProxy): Deleted.
(WebKit::WebPageProxy::shouldStartTrackingTouchEvents): Deleted.
* UIProcess/WebPageProxy.h:

LayoutTests:

* fast/events/touch/ios/tap-with-active-listener-on-elements-expected.txt: Added.
* fast/events/touch/ios/tap-with-active-listener-on-elements.html: Added.
* fast/events/touch/ios/tap-with-active-listener-on-window-expected.txt: Added.
* fast/events/touch/ios/tap-with-active-listener-on-window.html: Added.
* fast/events/touch/ios/tap-with-passive-listener-inside-active-listener-expected.txt: Added.
* fast/events/touch/ios/tap-with-passive-listener-on-elements-expected.txt: Added.
* fast/events/touch/ios/tap-with-passive-listener-on-elements.html: Added.
* fast/events/touch/ios/tap-with-passive-listener-on-window-expected.txt: Added.
* fast/events/touch/ios/tap-with-passive-listener-on-window.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201958 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAir.js should have some documentation
fpizlo@apple.com [Sat, 11 Jun 2016 00:48:43 +0000 (00:48 +0000)]
Air.js should have some documentation
https://bugs.webkit.org/show_bug.cgi?id=158648

Reviewed by Keith Miller.

I want to be able to point people at a document if they want to know more about this
benchmark.

* Air.js/README.md: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201957 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoREGRESSION(r198177): Cannot paste an image when the pasteboard format is mime type.
enrica@apple.com [Sat, 11 Jun 2016 00:21:16 +0000 (00:21 +0000)]
REGRESSION(r198177): Cannot paste an image when the pasteboard format is mime type.
https://bugs.webkit.org/show_bug.cgi?id=158590
rdar://problem/25471371

Reviewed by Darin Adler.

Source/WebCore:

When creating a fragment from an image resource, the resource needs to
be added to the document loader before setting the src attribute to the
image element, otherwise loading is triggered and the loading fails.
In r198177 the order of the operations was changed causing the bug.
This patch adds support to test the scenario where the image in the pasteboard
is available only as mime type (not WebArchive or RTFD), a situation that occurs
more frequently on iOS.

Test: editing/pasteboard/image-in-iframe.html

* editing/ios/EditorIOS.mm:
(WebCore::Editor::createFragmentForImageResourceAndAddResource):
* editing/mac/EditorMac.mm:
(WebCore::Editor::WebContentReader::readWebArchive):
(WebCore::Editor::WebContentReader::readRTFD):
(WebCore::Editor::WebContentReader::readRTF):
(WebCore::Editor::createFragmentForImageResourceAndAddResource):
* page/Settings.cpp:
(WebCore::Settings::setImagesEnabled):
(WebCore::Settings::setPreferMimeTypeForImages):
(WebCore::Settings::setForcePendingWebGLPolicy):
* page/Settings.h:
(WebCore::Settings::areImagesEnabled):
(WebCore::Settings::preferMimeTypeForImages):
(WebCore::Settings::arePluginsEnabled):
* testing/InternalSettings.cpp:
(WebCore::InternalSettings::Backup::restoreTo):
(WebCore::InternalSettings::setLangAttributeAwareFormControlUIEnabled):
(WebCore::InternalSettings::setPreferMimeTypeForImages):
(WebCore::InternalSettings::setImagesEnabled):
* testing/InternalSettings.h:
* testing/InternalSettings.idl:

LayoutTests:

* editing/pasteboard/image-in-iframe-expected.txt: Added.
* editing/pasteboard/image-in-iframe.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201956 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoFix CMake build.
achristensen@apple.com [Sat, 11 Jun 2016 00:20:05 +0000 (00:20 +0000)]
Fix CMake build.

* TestWebKitAPI/PlatformMac.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201955 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoFix WinCairo build after r201943
achristensen@apple.com [Sat, 11 Jun 2016 00:15:11 +0000 (00:15 +0000)]
Fix WinCairo build after r201943

* platform/network/curl/MultipartHandle.cpp:
(WebCore::MultipartHandle::didReceiveResponse):
* platform/network/curl/ResourceHandleManager.cpp:
(WebCore::handleLocalReceiveResponse):
(WebCore::headerCallback):
(WebCore::ResourceHandleManager::dispatchSynchronousJob):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201954 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoPerformanceTests:
fpizlo@apple.com [Fri, 10 Jun 2016 23:54:33 +0000 (23:54 +0000)]
PerformanceTests:
Rename JSAir to Air.js.

Rubber stamped by Mark Lam.

* Air.js: Copied from PerformanceTests/JSAir.
* Air.js/airjs-tests.yaml: Copied from PerformanceTests/JSAir/jsair-tests.yaml.
* Air.js/benchmark.js: Copied from PerformanceTests/JSAir/benchmark.js.
(Benchmark):
* Air.js/jsair-tests.yaml: Removed.
* Air.js/payload-airjs-ACLj8C.js: Copied from PerformanceTests/JSAir/payload-jsair-ACLj8C.js.
* Air.js/payload-jsair-ACLj8C.js: Removed.
* Air.js/stress-test.js: Copied from PerformanceTests/JSAir/stress-test.js.
* Air.js/test.html: Copied from PerformanceTests/JSAir/test.html.
* Air.js/test.js: Copied from PerformanceTests/JSAir/test.js.
* JSAir: Removed.
* JSAir/all.js: Removed.
* JSAir/allocate_stack.js: Removed.
* JSAir/arg.js: Removed.
* JSAir/basic_block.js: Removed.
* JSAir/benchmark.js: Removed.
* JSAir/code.js: Removed.
* JSAir/custom.js: Removed.
* JSAir/frequented_block.js: Removed.
* JSAir/insertion_set.js: Removed.
* JSAir/inst.js: Removed.
* JSAir/jsair-tests.yaml: Removed.
* JSAir/liveness.js: Removed.
* JSAir/opcode.js: Removed.
* JSAir/payload-gbemu-executeIteration.js: Removed.
* JSAir/payload-imaging-gaussian-blur-gaussianBlur.js: Removed.
* JSAir/payload-jsair-ACLj8C.js: Removed.
* JSAir/payload-typescript-scanIdentifier.js: Removed.
* JSAir/reg.js: Removed.
* JSAir/stack_slot.js: Removed.
* JSAir/stress-test.js: Removed.
* JSAir/strip-hash.rb: Removed.
* JSAir/symbols.js: Removed.
* JSAir/test.html: Removed.
* JSAir/test.js: Removed.
* JSAir/tmp.js: Removed.
* JSAir/tmp_base.js: Removed.
* JSAir/util.js: Removed.

Tools:
Rename JSAir to Air.js

Rubber stamped by Mark Lam.

* Scripts/run-javascriptcore-tests:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201953 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoMarking imported/blink/storage/indexeddb/blob-delete-objectstore-db.html as flaky...
ryanhaddad@apple.com [Fri, 10 Jun 2016 23:29:10 +0000 (23:29 +0000)]
Marking imported/blink/storage/indexeddb/blob-delete-objectstore-db.html as flaky on Yosemite Release WK2
https://bugs.webkit.org/show_bug.cgi?id=158639

Unreviewed test gardening.

* platform/mac-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201952 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWeb Inspector: heap snapshot comparison view shows nothing if later snapshot is chose...
mattbaker@apple.com [Fri, 10 Jun 2016 23:05:52 +0000 (23:05 +0000)]
Web Inspector: heap snapshot comparison view shows nothing if later snapshot is chosen first
https://bugs.webkit.org/show_bug.cgi?id=158584
<rdar://problem/26727849>

Reviewed by Timothy Hatcher.

ALways compare the the later snapshot to the earlier snapshot,
regardless of selection order.

* UserInterface/Views/HeapAllocationsTimelineView.js:
(WebInspector.HeapAllocationsTimelineView):
(WebInspector.HeapAllocationsTimelineView.prototype._dataGridNodeSelected):
Swap shapshots if the first has a greater identifier than the second.
Drive-by cleanup: null out `_baselineDataGridNode` since it isn't used
after the comparison completes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201949 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoJSC Stress Test failing: jsair-tests.yaml/test.js.ftl-eager-no-cjit
fpizlo@apple.com [Fri, 10 Jun 2016 22:55:01 +0000 (22:55 +0000)]
JSC Stress Test failing: jsair-tests.yaml/test.js.ftl-eager-no-cjit
https://bugs.webkit.org/show_bug.cgi?id=158571

Reviewed by Keith Miller.

PerformanceTests:

Introduce a different harness for run-jsc-stress-tests, which runs for a minimum of 10
iterations and then tries to do any number of "bonus" iterations until it's been running for
two seconds. Since this is the sort of test that isn't really meant to stress anything in
particular, I think it's OK if it is time-limited in this way. The worst case is that some
of its failures will be flaky, but I think that they would have been flaky anyway given the
complexity of the test.

* JSAir/benchmark.js:
(Benchmark):
(Benchmark.prototype.runIteration):
(benchmark): Deleted.
* JSAir/jsair-tests.yaml:
* JSAir/stress-test.js: Added.
(preciseTime):
* JSAir/test.html:
* JSAir/test.js:

Tools:

Unskip the JSAir test.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201947 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agohandleDataURL is only used by curl
achristensen@apple.com [Fri, 10 Jun 2016 22:26:34 +0000 (22:26 +0000)]
handleDataURL is only used by curl
https://bugs.webkit.org/show_bug.cgi?id=158636

Reviewed by Tim Horton.

* CMakeLists.txt:
* platform/network/DataURL.cpp: Removed.
* platform/network/DataURL.h: Removed.
* platform/network/curl/ResourceHandleManager.cpp:
(WebCore::ResourceHandleManager::startScheduledJobs):
(WebCore::handleDataURL):
(WebCore::ResourceHandleManager::dispatchSynchronousJob):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201946 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAdd requestExitFullScreen to WKFullScreenWindowController
bdakin@apple.com [Fri, 10 Jun 2016 22:18:04 +0000 (22:18 +0000)]
Add requestExitFullScreen to WKFullScreenWindowController
https://bugs.webkit.org/show_bug.cgi?id=158633
-and corresponding-
rdar://problem/26564036

Reviewed by Tim Horton.

* UIProcess/mac/WKFullScreenWindowController.h:
* UIProcess/mac/WKFullScreenWindowController.mm:
(-[WKFullScreenWindowController requestExitFullScreen]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201945 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoRe-disable the UserMedia tests which are timing out.
weinig@apple.com [Fri, 10 Jun 2016 22:15:24 +0000 (22:15 +0000)]
Re-disable the UserMedia tests which are timing out.

They used to be accidentally disabled, due to not including
the FeatureDefines correctly, but not we explicitly disable it.

* TestWebKitAPI/Tests/WebKit2/UserMedia.cpp:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201944 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoReduce ResourceResponse copying
achristensen@apple.com [Fri, 10 Jun 2016 22:02:55 +0000 (22:02 +0000)]
Reduce ResourceResponse copying
https://bugs.webkit.org/show_bug.cgi?id=158232

Reviewed by Darin Adler.

Source/WebCore:

No new tests.  No change in behavior except removing an unnecessary copy on cocoa platforms.

* loader/ResourceLoader.cpp:
(WebCore::ResourceLoader::didSendData):
(WebCore::ResourceLoader::didReceiveResponse):
* loader/ResourceLoader.h:
* loader/appcache/ApplicationCacheGroup.cpp:
(WebCore::ApplicationCacheGroup::createResourceHandle):
(WebCore::ApplicationCacheGroup::didReceiveResponse):
* loader/appcache/ApplicationCacheGroup.h:
* platform/graphics/PlatformMediaResourceLoader.h:
(WebCore::PlatformMediaResourceClient::~PlatformMediaResourceClient):
(WebCore::PlatformMediaResourceClient::responseReceived):
(WebCore::PlatformMediaResourceClient::redirectReceived):
(WebCore::PlatformMediaResourceClient::shouldCacheResponse):
(WebCore::PlatformMediaResourceClient::dataSent):
* platform/graphics/avfoundation/objc/WebCoreAVFResourceLoader.h:
* platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:
(ResourceHandleStreamingClient::willSendRequest):
(ResourceHandleStreamingClient::didReceiveResponse):
* platform/network/BlobResourceHandle.cpp:
(WebCore::BlobResourceHandle::notifyResponseOnSuccess):
(WebCore::BlobResourceHandle::notifyResponseOnError):
(WebCore::BlobResourceHandle::notifyReceiveData):
* platform/network/DataURL.cpp:
(WebCore::handleDataURL):
* platform/network/PingHandle.h:
(WebCore::PingHandle::PingHandle):
* platform/network/ResourceHandleClient.cpp:
(WebCore::ResourceHandleClient::willSendRequestAsync):
(WebCore::ResourceHandleClient::didReceiveResponseAsync):
* platform/network/ResourceHandleClient.h:
(WebCore::ResourceHandleClient::didSendData):
(WebCore::ResourceHandleClient::didReceiveResponse):
(WebCore::ResourceHandleClient::didReceiveData):
* platform/network/ResourceResponseBase.cpp:
(WebCore::ResourceResponseBase::ResourceResponseBase):
(WebCore::ResourceResponseBase::includeCertificateInfo):
(WebCore::ResourceResponseBase::suggestedFilename):
(WebCore::ResourceResponseBase::certificateInfo): Deleted.
* platform/network/ResourceResponseBase.h:
(WebCore::ResourceResponseBase::certificateInfo):
(WebCore::ResourceResponseBase::encode):
(WebCore::ResourceResponseBase::decode):
(WebCore::ResourceResponseBase::containsCertificateInfo): Deleted.
* platform/network/SynchronousLoaderClient.cpp:
(WebCore::SynchronousLoaderClient::canAuthenticateAgainstProtectionSpace):
(WebCore::SynchronousLoaderClient::didReceiveResponse):
(WebCore::SynchronousLoaderClient::didReceiveData):
* platform/network/SynchronousLoaderClient.h:
* platform/network/cf/ResourceHandleCFURLConnectionDelegateWithOperationQueue.cpp:
(WebCore::ResourceHandleCFURLConnectionDelegateWithOperationQueue::didReceiveResponse):
* platform/network/cf/SynchronousResourceHandleCFURLConnectionDelegate.cpp:
(WebCore::SynchronousResourceHandleCFURLConnectionDelegate::didReceiveResponse):
(WebCore::SynchronousResourceHandleCFURLConnectionDelegate::didReceiveData):
* platform/network/mac/WebCoreResourceHandleAsDelegate.mm:
(-[WebCoreResourceHandleAsDelegate connection:didReceiveResponse:]):
* platform/network/mac/WebCoreResourceHandleAsOperationQueueDelegate.mm:
(-[WebCoreResourceHandleAsOperationQueueDelegate connection:didReceiveResponse:]):
* platform/network/soup/ResourceHandleSoup.cpp:
(WebCore::nextMultipartResponsePartCallback):
(WebCore::sendRequestCallback):

Source/WebKit2:

* NetworkProcess/Downloads/PendingDownload.h:
* NetworkProcess/NetworkDataTask.h:
* NetworkProcess/NetworkLoad.cpp:
(WebKit::NetworkLoad::continueDidReceiveResponse):
(WebKit::NetworkLoad::sharedDidReceiveResponse):
(WebKit::NetworkLoad::sharedWillSendRedirectedRequest):
(WebKit::NetworkLoad::didReceiveChallenge):
(WebKit::NetworkLoad::didReceiveResponseNetworkSession):
(WebKit::NetworkLoad::didReceiveResponseAsync):
* NetworkProcess/NetworkLoad.h:
* NetworkProcess/NetworkLoadClient.h:
* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::retrieveCacheEntry):
(WebKit::NetworkResourceLoader::abort):
* NetworkProcess/NetworkResourceLoader.h:
* NetworkProcess/PingLoad.h:
* NetworkProcess/cache/NetworkCacheCoders.h:
(WebKit::NetworkCache::Coder<Optional<T>>::encode):
(WebKit::NetworkCache::Coder<Optional<T>>::decode):
* NetworkProcess/cache/NetworkCacheSpeculativeLoad.cpp:
(WebKit::NetworkCache::SpeculativeLoad::willSendRedirectedRequest):
* NetworkProcess/cache/NetworkCacheSpeculativeLoad.h:
* NetworkProcess/cocoa/NetworkDataTaskCocoa.mm:
(WebKit::NetworkDataTask::didCompleteWithError):
(WebKit::NetworkDataTask::didReceiveResponse):
* NetworkProcess/cocoa/NetworkSessionCocoa.mm:
(-[WKNetworkSessionDelegate URLSession:dataTask:didReceiveResponse:completionHandler:]):
* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::dispatchDidCommitLoad):
* WebProcess/WebPage/WebFrame.cpp:
(WebKit::WebFrame::url):
(WebKit::WebFrame::certificateInfo):
(WebKit::WebFrame::innerText):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201943 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAdd SPI to disable spellchecking on auto-fillable text fields
rniwa@webkit.org [Fri, 10 Jun 2016 22:00:23 +0000 (22:00 +0000)]
Add SPI to disable spellchecking on auto-fillable text fields
https://bugs.webkit.org/show_bug.cgi?id=158611

Reviewed by Anders Carlsson.

Source/WebCore:

Added a boolean flag m_isSpellCheckingEnabled to HTMLInputElement. This flag defaults to true, and can be set
to false by WebKit2 C API.

* editing/Editor.cpp:
(WebCore::Editor::isSpellCheckingEnabledFor): Fixed a bug that we were calling isSpellCheckingEnabled on
the div inside an input element's shadow tree instead of the input element itself.
* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::HTMLInputElement): Initialize m_spellcheckEnabled to true (it's a bit field).
(WebCore::HTMLInputElement::isSpellCheckingEnabled): Added. Return false if m_spellcheckEnabled is false.
* html/HTMLInputElement.h:
(WebCore::HTMLInputElement::setSpellcheckEnabled): Added.

Source/WebKit2:

Added WKBundleNodeHandleSetHTMLInputElementSpellcheckEnabled to disable spellchecking on a text field.

This is used by WebKit2 client which desires to disable spellchecking and notably autocorrection on
login forms, etc... where such feature would interfere with user's actions.

* WebProcess/InjectedBundle/API/c/WKBundleNodeHandle.cpp:
(WKBundleNodeHandleSetHTMLInputElementSpellcheckEnabled): Added.
* WebProcess/InjectedBundle/API/c/WKBundleNodeHandlePrivate.h:
* WebProcess/InjectedBundle/DOM/InjectedBundleNodeHandle.cpp:
(WebKit::InjectedBundleNodeHandle::setHTMLInputElementSpellcheckEnabled): Added.
* WebProcess/InjectedBundle/DOM/InjectedBundleNodeHandle.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201942 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoIntroduce WTF::UniqueRef
achristensen@apple.com [Fri, 10 Jun 2016 21:26:00 +0000 (21:26 +0000)]
Introduce WTF::UniqueRef
https://bugs.webkit.org/show_bug.cgi?id=158596

Reviewed by Brady Eidson.

Source/WebCore:

No new tests.  No change in behavior.

* inspector/InspectorOverlay.cpp:
(WebCore::InspectorOverlay::overlayPage):
* loader/EmptyClients.cpp:
(WebCore::fillWithEmptyClients):
* page/Page.cpp:
(WebCore::Page::Page):
* page/Page.h:
(WebCore::Page::canStartMedia):
(WebCore::Page::editorClient):
(WebCore::Page::plugInClient):
(WebCore::Page::mainFrame):
(WebCore::Page::groupPtr): Deleted.
* page/PageConfiguration.cpp:
(WebCore::PageConfiguration::PageConfiguration):
* page/PageConfiguration.h:
* svg/graphics/SVGImage.cpp:
(WebCore::SVGImage::dataChanged):

Source/WebKit/mac:

* WebView/WebView.mm:
(-[WebView _commonInitializationWithFrameName:groupName:]):
(-[WebView initSimpleHTMLDocumentWithStyle:frame:preferences:groupName:]):

Source/WebKit/win:

* WebView.cpp:
(WebView::initWithFrame):

Source/WebKit2:

* WebProcess/WebPage/WebPage.cpp:
(WebKit::m_shouldDispatchFakeMouseMoveEvents):

Source/WTF:

WTF::UniqueRef is like a std::unique_ptr that is guaranteed to be non-null.
std::make_unique returns a non-null value that is put into a std::unique_ptr, a type
that could contain null values.  To be able to pass such values around and store them
without wondering if they are null, we now have WTF::UniqueRef which cannot be null.

* WTF.xcodeproj/project.pbxproj:
* wtf/CMakeLists.txt:
* wtf/UniqueRef.h: Added.
(WTF::makeUniqueRef):
(WTF::UniqueRef::UniqueRef):
(WTF::UniqueRef::get):
(WTF::UniqueRef::operator&):
(WTF::UniqueRef::operator->):

Tools:

* TestWebKitAPI/CMakeLists.txt:
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WTF/UniqueRef.cpp: Added.
(TestWebKitAPI::B::B):
(TestWebKitAPI::C::C):
(TestWebKitAPI::function):
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201941 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoAdd an IDB leak test.
beidson@apple.com [Fri, 10 Jun 2016 21:01:55 +0000 (21:01 +0000)]
Add an IDB leak test.
https://bugs.webkit.org/show_bug.cgi?id=158632

Reviewed by Alex Christensen.

* storage/indexeddb/modern/leak-1-expected.txt: Added.
* storage/indexeddb/modern/leak-1.html: Added.
* storage/indexeddb/modern/resources/leak-1.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201940 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWeb Inspector: Cleanup InspectorIndexedDBAgent a bit
commit-queue@webkit.org [Fri, 10 Jun 2016 20:59:38 +0000 (20:59 +0000)]
Web Inspector: Cleanup InspectorIndexedDBAgent a bit
https://bugs.webkit.org/show_bug.cgi?id=158598

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2016-06-10
Reviewed by Darin Adler.

* inspector/InspectorIndexedDBAgent.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201939 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoRefactor TestWebKitAPI to allow just testing WTF
weinig@apple.com [Fri, 10 Jun 2016 20:38:35 +0000 (20:38 +0000)]
Refactor TestWebKitAPI to allow just testing WTF
https://bugs.webkit.org/show_bug.cgi?id=158625

Reviewed by Tim Horton.

Extract all the WTF tests into a new target to allow a faster build / test / fix
cycle when working on WTF bugs and features.

By calling `run-api-tests --wtf-only`, you don't need to have a build of JavaScriptCore/WebCore/etc
to test WTF only changes.

* Scripts/build-api-tests:
* Scripts/run-api-tests:
(buildTestTool):
(testToolPaths):
Add new options to build-api-tests and run-api-tests to only build/run the WTF test runner.

* TestWebKitAPI/CMakeLists.txt:
* TestWebKitAPI/Configurations/TestWTF.xcconfig: Added.
* TestWebKitAPI/Configurations/TestWTFLibrary.xcconfig: Added.
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
Add two new targets, TestWTFLibrary which contains all the WTF tests, and TestWTF,
the test runner that runs the WTF tests.

* TestWebKitAPI/Tests/WTF/RunLoop.cpp:
Use Utilities.h rather than PlatformUtilities.h.

* TestWebKitAPI/Utilities.h: Added.
* TestWebKitAPI/PlatformUtilities.h:
* TestWebKitAPI/cocoa/PlatformUtilitiesCocoa.mm:
* TestWebKitAPI/cocoa/UtilitiesCocoa.mm: Added.
Move the sleep and run functions into a new Utilities.h/cpp file. This file contains utilities
that don't depend on anything in the WebKit project.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201938 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoWeb Inspector: rendering frames timeline "60fps" horizontal guide is mispositioned...
mattbaker@apple.com [Fri, 10 Jun 2016 19:59:20 +0000 (19:59 +0000)]
Web Inspector: rendering frames timeline "60fps" horizontal guide is mispositioned after switching timeline modes
https://bugs.webkit.org/show_bug.cgi?id=158587
<rdar://problem/26728174>

Reviewed by Timothy Hatcher.

When the view mode changes we iterate over all overview graphs, updating
their visibility as needed. Graphs being shown perform layout, even though
the overview's state is in transition: its offset height may still change.

A simple fix is to use TimelineOverviewGraph.height instead of the graph
element's offset height when performing layout.

* UserInterface/Views/RenderingFrameTimelineOverviewGraph.js:
(WebInspector.RenderingFrameTimelineOverviewGraph.prototype.layout):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201937 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 years agoThe backend should be happy to compile Unreachable even if AI didn't prove it to...
fpizlo@apple.com [Fri, 10 Jun 2016 19:56:18 +0000 (19:56 +0000)]
The backend should be happy to compile Unreachable even if AI didn't prove it to be unreachable
https://bugs.webkit.org/show_bug.cgi?id=158631

Reviewed by Keith Miller.

We've been slowly making the DFG Unreachable opcode behave like a grown-up. When we first
added it, it was a hack for Throw, and we could always rely on AI proving that Unreachable
was not reachable. But then we started using Unreachable as a proper Unreachable opcode,
like Oops in B3 for example, which has a more nuanced meaning: you use it whenever you
emit code that *you* know will not return, and you need some way of terminating the basic
block. The DFG is not a proof-carrying compiler, and it never will be. So, when you have
proved that something is not reachable, you should be able to use Unreachable even if
there is no guarantee that the compiler will later be able to replicate your proof. This
means that the backend may find itself compiling Unreachable because AI did not prove that
it was unreachable.

Prior to this change, we would crash compiling Unreachable because we would rely on AI
preventing us from reaching Unreachable in the backend. But that's silly! We don't want
users of Unreachable to have to also convince AI that their Unreachable is really
Unreachable.

This fixes crashes on real websites. I couldn't work out how to turn them into a reduced
test.

* assembler/AbortReason.h:
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::emitInvalidationPoint):
(JSC::DFG::SpeculativeJIT::unreachable):
(JSC::DFG::SpeculativeJIT::terminateSpeculativeExecution):
* dfg/DFGSpeculativeJIT.h:
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::LowerDFGToB3::compileNode):
(JSC::FTL::DFG::LowerDFGToB3::compilePutDynamicVar):
(JSC::FTL::DFG::LowerDFGToB3::compileUnreachable):
(JSC::FTL::DFG::LowerDFGToB3::compareEqObjectOrOtherToObject):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@201936 268f45cc-cd09-0410-ab3c-d52691b4dbfc