GitLab

https://bugs.webkit.org/show_bug.cgi?id=108253

Reviewed by Mark Hahnenberg.

When accessing methodTable() we now always make sure that our
structure _could_ be valid.  Added a separate method to get a
classes methodTable during destruction as it's not possible to
validate the structure at that point.  This separation might
also make it possible to improve the performance of methodTable
access more generally in future.

* heap/MarkedBlock.cpp:
(JSC::MarkedBlock::callDestructor):
* runtime/JSCell.h:
(JSCell):
* runtime/JSCellInlines.h:
(JSC::JSCell::methodTableForDestruction):
(JSC):
(JSC::JSCell::methodTable):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141190 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108261

Reviewed by Oliver Hunt.
        
Backends shouldn't override each other's methods. That's not cool.

* offlineasm/mips.rb:



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141189 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108232

Unreviewed, gardening.

After https://codereview.chromium.org/12045002/, these
tests are expected to fail.  Disabling and creating bugs for follow-up.

* platform/chromium/TestExpectations:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141188 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reviewed by Simon Fraser.

Only do the work if the tiled scrolling indicator/map is enabled.

* platform/graphics/ca/mac/TileCache.mm:
(WebCore::TileCache::setScrollingModeIndication):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141187 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108228

Reviewed by Ryosuke Niwa.

Source/WebCore:

Comboboxes behave much like textfields, and so they need
to respond like text controls.

Test: platform/mac/accessibility/combox-box-value.html

* accessibility/AccessibilityObject.cpp:
(WebCore::AccessibilityObject::isTextControl):
(WebCore):
* accessibility/AccessibilityObject.h:
(AccessibilityObject):

LayoutTests:

* platform/mac/accessibility/combox-box-value-expected.txt: Added.
* platform/mac/accessibility/combox-box-value.html: Added.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141186 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108251

Reviewed by Mark Hahnenberg.
        
Revert http://trac.webkit.org/changeset/141178 and rename 'dump' to 'clDump'.
        
Also made trivial build fixes for !ENABLE(JIT).

* offlineasm/cloop.rb:
* runtime/Executable.h:
(ExecutableBase):
(JSC::ExecutableBase::intrinsicFor):
* runtime/JSGlobalData.h:



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141185 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108218

Reviewed by Adam Barth.

Requested by darin in https://bugs.webkit.org/show_bug.cgi?id=107845
This also allows enabling -Wreturn-type-c-linkage again, but I'd like to
wait for the next clang roll (which tweaks this warning) before undoing
r140800 (which removed that warning).

No behavior change.

* bindings/v8/npruntime.cpp:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141184 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108258

Reviewed by Beth Dakin.

* Platform/CoreIPC/Connection.h:
(CoreIPC::Connection::send):
(CoreIPC::Connection::sendSync):
(CoreIPC::Connection::waitForAndDispatchImmediately):
* Platform/CoreIPC/MessageSender.h:
(CoreIPC::MessageSender::send):
* Scripts/webkit2/messages.py:
(surround_in_condition):
(message_to_struct_declaration):
(generate_messages_header):
* Scripts/webkit2/messages_unittest.py:
* Scripts/webkit2/model.py:
(Message.__init__):
* Shared/ChildProcessProxy.h:
(WebKit::ChildProcessProxy::send):
* Shared/WebConnection.cpp:
(WebKit::WebConnection::postMessage):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141183 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108240

Patch by Alan Cutter <alancutter@chromium.org> on 2013-01-29
Reviewed by Dirk Pranke.

Added the ability to have MockWeb serve mock web data.

* Scripts/webkitpy/common/host_mock.py:
(MockHost.__init__):
* Scripts/webkitpy/common/net/web_mock.py:
(MockWeb.__init__):
(MockWeb.get_binary):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141182 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108249

* platform/win/TestExpectations:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141181 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108241

Patch by Kiran Muppala <cmuppala@apple.com> on 2013-01-29
Reviewed by Anders Carlsson.

* UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::addExistingWebPage): Remove erroneous
semicolon following the if condition.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141180 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108245

Reviewed by Filip Pizlo.

../JavaScriptCore: 

* GNUmakefile.list.am:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj.filters:
* JavaScriptCore.xcodeproj/project.pbxproj:
* dfg/DFGRepatch.cpp:
(JSC::DFG::emitPutReplaceStub):
(JSC::DFG::emitPutTransitionStub):
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::writeBarrier):
* dfg/DFGSpeculativeJIT.h:
(SpeculativeJIT):
* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* heap/CardSet.h: Removed.
* heap/Heap.cpp:
(JSC::Heap::markRoots):
(JSC::Heap::collect):
* heap/Heap.h:
(Heap):
(JSC::Heap::shouldCollect):
(JSC::Heap::isWriteBarrierEnabled):
(JSC):
(JSC::Heap::writeBarrier):
* heap/MarkedBlock.h:
(MarkedBlock):
(JSC):
* heap/MarkedSpace.cpp:
(JSC):
* jit/JITPropertyAccess.cpp:
(JSC::JIT::emitWriteBarrier):

../WebCore: 

* ForwardingHeaders/heap/CardSet.h: Removed.
* WebCore.vcxproj/WebCore.vcxproj:
* WebCore.vcxproj/WebCore.vcxproj.filters:


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141179 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108247

Reviewed by Oliver Hunt.
        
Makes offlineasm dumping easier to read and less likely to cause assertion failures.
Also fixes the strange situation where cloop.rb and ast.rb both defined dump methods,
but cloop.rb was winning.

* offlineasm/cloop.rb:



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141178 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108244

Reviewed by Beth Dakin.

* Platform/CoreIPC/MessageID.h:
(CoreIPC::MessageID::MessageID):
* Platform/CoreIPC/mac/ConnectionMac.cpp:
(CoreIPC):
(CoreIPC::Connection::sendOutgoingMessage):
(CoreIPC::createMessageDecoder):
(CoreIPC::Connection::receiveSourceEventHandler):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141177 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=107839

Reviewed by Oliver Hunt.

JSContext has a JSWrapperMap, which has an NSMutableDictionary m_classMap, which has values that
are JSObjCClassInfo objects, which have strong references to two JSValue *'s, m_prototype and
m_constructor, which in turn have strong references to the JSContext, creating a reference cycle.
We should make m_prototype and m_constructor Weak<JSObject>. This gets rid of the strong reference
to the JSContext and also prevents clients from accidentally creating reference cycles by assigning
to the prototype of the constructor. If Weak<JSObject> fields are ever garbage collected, we will
reallocate them.

* API/JSContext.mm:
(-[JSContext wrapperMap]):
* API/JSContextInternal.h:
* API/JSWrapperMap.mm:
(-[JSObjCClassInfo initWithContext:forClass:superClassInfo:]):
(-[JSObjCClassInfo dealloc]):
(-[JSObjCClassInfo allocateConstructorAndPrototypeWithSuperClassInfo:]):
(-[JSObjCClassInfo allocateConstructorAndPrototype]):
(-[JSObjCClassInfo wrapperForObject:]):
(-[JSObjCClassInfo constructor]):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141176 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108209

Reviewed by Dimitri Glazkov.

Source/WebCore:

Many of the exception cases for ShadowRoot are actually impossible and
should be asserts instead. We can also move the one case of exception logic,
for elements that don't allow author shadows into Element::createShadowRoot
instead of having it all over the ShadowRoot and ElementShadow classes. This
is the first step in centralizing all ShadowRoot creation inside ElementShadow.

No new tests, covered by existing tests.

* WebCore.exp.in:
* dom/Element.cpp:
(WebCore::Element::createShadowRoot): Be explicit about what kind of ShadowRoot you're creating.
(WebCore::Element::ensureUserAgentShadowRoot): No more exceptions.
* dom/ElementShadow.cpp:
(WebCore::ElementShadow::addShadowRoot): This never actually throws exceptions, remove ExceptionCode.
* dom/ElementShadow.h:
(ElementShadow):
* dom/ShadowRoot.cpp:
(WebCore::determineUsageType): Merge with Element::createShadowRoot.
(WebCore):
(WebCore::ShadowRoot::ShadowRoot): Moved Histogram logic here since it's actually about the constructor anyway.
(WebCore::ShadowRoot::create): Removed overload that made the code less obvious.
* dom/ShadowRoot.h:
(ShadowRoot):
* html/HTMLKeygenElement.cpp:
(WebCore::HTMLKeygenElement::HTMLKeygenElement):
* html/shadow/TextFieldDecorationElement.cpp:
(WebCore::getDecorationRootAndDecoratedRoot):
* testing/Internals.cpp:
(WebCore::Internals::ensureShadowRoot):
(WebCore::Internals::createShadowRoot):

Source/WebKit/win:

* WebKit.vcproj/WebKitExports.def.in: Swap ShadowRoot::create export with Element::createShadowRoot.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141175 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108234

Reviewed by Beth Dakin.

Add explicit message receiver name equality checks instead of using MessageID::is.

* NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::didReceiveMessage):
(WebKit::NetworkConnectionToWebProcess::didReceiveSyncMessage):
* Platform/CoreIPC/MessageID.h:
* PluginProcess/WebProcessConnection.cpp:
(WebKit::WebProcessConnection::didReceiveMessage):
(WebKit::WebProcessConnection::didReceiveSyncMessage):
* Shared/mac/SecItemShim.cpp:
(WebKit::SecItemShim::didReceiveMessageOnConnectionWorkQueue):
* UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::didReceiveMessage):
* UIProcess/WebContext.cpp:
(WebKit::WebContext::didReceiveMessage):
(WebKit::WebContext::didReceiveSyncMessage):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::didReceiveMessage):
(WebKit::WebPageProxy::didReceiveSyncMessage):
* UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::didReceiveMessage):
(WebKit::WebProcessProxy::didReceiveSyncMessage):
(WebKit::WebProcessProxy::didReceiveMessageOnConnectionWorkQueue):
* UIProcess/mac/SecItemShimProxy.cpp:
(WebKit::SecItemShimProxy::didReceiveMessageOnConnectionWorkQueue):
* WebProcess/Network/NetworkProcessConnection.cpp:
(WebKit::NetworkProcessConnection::didReceiveMessage):
(WebKit::NetworkProcessConnection::didReceiveSyncMessage):
* WebProcess/Plugins/PluginProcessConnection.cpp:
(WebKit::PluginProcessConnection::didReceiveSyncMessage):
* WebProcess/WebPage/EventDispatcher.cpp:
(WebKit::EventDispatcher::didReceiveMessageOnConnectionWorkQueue):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::didReceiveMessage):
* WebProcess/WebProcess.cpp:
(WebKit::WebProcess::didReceiveMessage):
(WebKit::WebProcess::didReceiveMessageOnConnectionWorkQueue):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141174 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=103558
<rdar://problem/12763112>

Reviewed by Benjamin Poulain.

Source/WebKit/mac:

Clients may want to behave differently when their WebView/WKView has been swapped out by a placeholder
view when in full screen mode.  Add a simple accessor for the existing placeholder view to
WebFullScreenController and WebView.

* WebView/WebFullScreenController.h:
* WebView/WebFullScreenController.mm:
(-[WebFullScreenController webViewPlaceholder]): Added simple accessor.
* WebView/WebView.mm:
(-[WebView fullScreenPlaceholderView]): Added simple accessor.

Source/WebKit2:

Clients may want to behave differently when their WebView/WKView has been swapped out by a placeholder
view when in full screen mode.  Add a simple accessor for the existing placeholder view to
WKFullScreenWindowController and WKView.

* UIProcess/API/mac/WKView.mm:
(-[WKView fullScreenPlaceholderView]): Added simple accessor.
* UIProcess/API/mac/WKViewPrivate.h:
* UIProcess/mac/WKFullScreenWindowController.h:
* UIProcess/mac/WKFullScreenWindowController.mm:
(-[WKFullScreenWindowController webViewPlaceholder]): Added simple accessor.

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141173 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108210

Reviewed by Ryosuke Niwa.

* Scripts/run-api-tests:
Add the option and use it to set the configuration product directory. Since it is supposed
to point to the built products, do not build the products if root is specified (this mimics
the behavior of run-javascriptcore-tests).

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141172 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[WK2] Call LayerTreeHost::deviceOrPageScaleFactorChanged() when a device or page scale factor is changed.
https://bugs.webkit.org/show_bug.cgi?id=107802

Patch by Huang Dongsung <luxtella@company100.net> on 2013-01-29
Reviewed by Simon Fraser.

Currently, LayerTreeHostMac and *GTK call deviceOrPageScaleFactorChanged()
of the non compositing GraphicsLayer when a device scale factor is changed.

There are two problems.
1. We don't notify LayerTreeHost when a page scale factor is changed.
2. When using TiledCoreAnimationDrawingAreaProxy, LayerTreeHostMac does
not receive the device scale factor changed callback.

So this patch changes three points.
1. Rename from deviceScaleFactorDidChange() to deviceOrPageScaleFactorChanged()
in LayerTreeHost.
2. WebPage::setDeviceScaleFactor() calls LayerTreeHost::deviceScaleFactorDidChange()
because of dealing with TiledCoreAnimationDrawingAreaProxy.
3. WebPage::pageScaleFactor() calls LayerTreeHost::deviceScaleFactorDidChange()
to call deviceOrPageScaleFactorChanged() of the non compositing GraphicsLayer.

Unfortunately, I couldn't think of a way to test this in an automated fashion.

* WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.h:
(WebKit::CoordinatedLayerTreeHost::deviceOrPageScaleFactorChanged):
* WebProcess/WebPage/DrawingAreaImpl.cpp:
(WebKit::DrawingAreaImpl::updateBackingStoreState):
    Don't call LayerTreeHost::deviceScaleFactorDidChange() because this
    method calls WebPage::setDeviceScaleFactor() and then
    LayerTreeHost::deviceScaleFactorDidChange() is called.
* WebProcess/WebPage/LayerTreeHost.h:
(LayerTreeHost):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::scalePage):
(WebKit::WebPage::setDeviceScaleFactor):
* WebProcess/WebPage/gtk/LayerTreeHostGtk.cpp:
(WebKit::LayerTreeHostGtk::deviceOrPageScaleFactorChanged):
* WebProcess/WebPage/gtk/LayerTreeHostGtk.h:
(LayerTreeHostGtk):
* WebProcess/WebPage/mac/LayerTreeHostMac.h:
(LayerTreeHostMac):
* WebProcess/WebPage/mac/LayerTreeHostMac.mm:
(WebKit::LayerTreeHostMac::deviceOrPageScaleFactorChanged):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141171 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108235

Reviewed by Beth Dakin.

Just check for the message receiver name and message name explicitly instead.

* Platform/CoreIPC/MessageID.h:
(CoreIPC::MessageID::is):
* Shared/CoreIPCSupport/WebContextMessageKinds.h:
(WebContextLegacyMessage::messageReceiverName):
(WebContextLegacyMessage):
(WebContextLegacyMessage::postMessageMessageName):
(WebContextLegacyMessage::postSynchronousMessageMessageName):
* UIProcess/WebContext.cpp:
(WebKit::WebContext::didReceiveMessage):
(WebKit::WebContext::didReceiveSyncMessage):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141170 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108097

Reviewed by Geoffrey Garen.

Source/JavaScriptCore:

LiteralParser was accepting a bogus 'var a.b = c' statement

* runtime/LiteralParser.cpp:
(JSC::::tryJSONPParse):

LayoutTests:

Add parser test for invalid var a.b syntax

* fast/js/parser-syntax-check-expected.txt:
* fast/js/script-tests/parser-syntax-check.js:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141168 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108054

Patch by Kiran Muppala <cmuppala@apple.com> on 2013-01-29
Reviewed by Anders Carlsson.

Provide a preference to enable process suppression in WebProcesses with
no visible pages even if the application is not completely occluded.
This provides more opportunities for process suppression to take effect.

Replace the messaging of application occlusion status from UI process to
ChildProcesses with messages that indicate the current required process
suppression state.  WebProcessProxy should determine if the process is
eligible for process suppression based on both application occlusion
and page visibility.  When either of these parameters changes,
the proxy should message the child process of the updated process
suppression state.

* NetworkProcess/NetworkProcess.messages.in: Rename
SetApplicationIsOccluded to SetProcessSuppressionEnabled.
* PluginProcess/PluginProcess.messages.in: Ditto.
* Shared/ChildProcess.h:
(WebKit::ChildProcess::processSuppressionEnabled): Rename
applicationIsOccluded.
* Shared/WebPreferencesStore.h: Add a new preference
pageVisibilityBasedProcessSuppressionEnabled, to enabled/disable
process suppression of WebProcesses when all pages are hidden.
* Shared/mac/ChildProcessMac.mm:
(WebKit::ChildProcess::setProcessSuppressionEnabled): Rename
setApplicationIsOccluded.
(WebKit::ChildProcess::platformInitialize): Replace call to
setApplicationIsOccluded with setProcessSuppressionEnabled.
* SharedWorkerProcess/SharedWorkerProcess.messages.in: Rename
SetApplicationIsOccluded to SetProcessSuppressionEnabled.
* UIProcess/API/C/WKPreferences.cpp:
(WKPreferencesSetPageVisibilityBasedProcessSuppressionEnabled):
(WKPreferencesGetPageVisibilityBasedProcessSuppressionEnabled):
* UIProcess/API/C/WKPreferencesPrivate.h:
* UIProcess/Network/NetworkProcessProxy.cpp:
(WebKit::NetworkProcessProxy::didFinishLaunching): Use setter
to update process suppression state instead of messaging directly.
Use WebContext::canEnableProcessSuppressionForNetworkProcess() helper
method to determine if process suppression should be enabled.
* UIProcess/Network/NetworkProcessProxy.h: Rename
setApplicationIsOccluded to setProcessSuppresionEnabled.
* UIProcess/Network/mac/NetworkProcessProxyMac.mm:
(WebKit::NetworkProcessProxy::setProcessSuppressionEnabled): Ditto.
* UIProcess/Plugins/PluginProcessManager.h: Ditto.
* UIProcess/Plugins/PluginProcessProxy.cpp:
(WebKit::PluginProcessProxy::didFinishLaunching): Use setter
to update process suppression state instead of messaging directly.
Use WebContext::canEnableProcessSuppressionForGlobalChildProcesses()
helper method to determine if process suppression should be enabled.
* UIProcess/Plugins/PluginProcessProxy.h: Rename
setApplicationIsOccluded to setProcessSuppresionEnabled.
* UIProcess/Plugins/mac/PluginProcessManagerMac.mm:
(WebKit::PluginProcessManager::setProcessSuppressionEnabled): Ditto.
* UIProcess/Plugins/mac/PluginProcessProxyMac.mm:
(WebKit::PluginProcessProxy::setProcessSuppressionEnabled): Ditto.
* UIProcess/SharedWorkers/SharedWorkerProcessManager.h: Ditto.
* UIProcess/SharedWorkers/SharedWorkerProcessProxy.cpp:
(WebKit::SharedWorkerProcessProxy::didFinishLaunching): Use setter
to update process suppression state instead of messaging directly.
Use WebContext::canEnableProcessSuppressionForGlobalChildProcesses()
helper method to determine if process suppression should be enabled.
* UIProcess/SharedWorkers/SharedWorkerProcessProxy.h: Rename
setApplicationIsOccluded to setProcessSuppresionEnabled.
* UIProcess/SharedWorkers/mac/SharedWorkerProcessManagerMac.mm:
(WebKit::SharedWorkerProcessManager::setProcessSuppressionEnabled):
Ditto.
* UIProcess/SharedWorkers/mac/SharedWorkerProcessProxyMac.mm:
(WebKit::SharedWorkerProcessProxy::setProcessSuppressionEnabled): Ditto.
* UIProcess/WebContext.h: Replace applicationIsOccluded() getter with
helper methods to determine if a child process can have process
suppression enabled.
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::viewStateDidChange): Notify WebProcessProxy of
page visibility change.
(WebKit::WebPageProxy::preferencesDidChange): Notify WebProcessProxy of
change in preferences.
* UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::WebProcessProxy): Initialize member variable
tracking process suppression state to false.
(WebKit::WebProcessProxy::createWebPage): Update set of pages which can
be process suppressed and the resulting process suppression state for
the child process.
(WebKit::WebProcessProxy::addExistingWebPage): Ditto.
(WebKit::WebProcessProxy::removeWebPage): Ditto.
(WebKit::WebProcessProxy::pageVisibilityChanged): Ditto.
(WebKit::WebProcessProxy::pagePreferencesChanged): Ditto.
(WebKit::WebProcessProxy::didFinishLaunching): Call
updateProcessSuppressionState().
* UIProcess/WebProcessProxy.h:
* UIProcess/mac/WebContextMac.mm:
(WebKit::updateProcessSuppressionStateOfGlobalChildProcesses): Use new
helper method canEnableProcessSuppressionForGlobalChildProcesses() to
determine if process suppression should be enabled.
(WebKit::applicationOcclusionStateChanged): Update renamed methods
and variables.
(WebKit::enableOcclusionNotifications): Move OmitProcessSuppression
user default check into canEnableProcessSuppression methods.
(WebKit::omitProcessSuppression):
(WebKit::WebContext::updateProcessSuppressionStateOfChildProcesses):
Use new helper method canEnableProcessSuppressionForNetworkProcess() to
determine if process suppression should be enabled for NetworkProcess.
Let WebProcess update process suppression state based on application
occlusion state and page visibility.
(WebKit::WebContext::canEnableProcessSuppressionForNetworkProcess):
(WebKit::WebContext::canEnableProcessSuppressionForWebProcess):
(WebKit::WebContext::canEnableProcessSuppressionForGlobalChildProcesses):
(WebKit::WebContext::processSuppressionEnabledChanged): Reevaluate if
process suppression is enabled for all contexts and update process
suppression state of global child processes.
* UIProcess/mac/WebProcessProxyMac.mm:
(WebKit::WebProcessProxy::pageIsProcessSuppressible):
(WebKit::WebProcessProxy::allPagesAreProcessSuppressible):
(WebKit::WebProcessProxy::updateProcessSuppressionState):
* WebProcess/WebProcess.messages.in:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141167 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=107784.

Reviewed by Geoffrey Garen.

DatabaseContext will no longer be a Supplement of ScriptExecutionContext.
Instead we will maintain a mutex guarded contextMap in the DatabaseManager
which maps ScriptExecutionContexts to DatabaseContexts.

Also cleaned up the shutdown mechanism of the DatabaseContext,
DatabaseThread, and Databases when their owner ScriptExecutionContext
destructs.

No new tests.

* Modules/webdatabase/AbstractDatabase.cpp:
(WebCore::AbstractDatabase::AbstractDatabase):
* Modules/webdatabase/AbstractDatabase.h:
(WebCore::AbstractDatabase::databaseContext):
(AbstractDatabase):
* Modules/webdatabase/Database.cpp:
(WebCore::Database::Database):
* Modules/webdatabase/Database.h:
(WebCore):
(Database):
* Modules/webdatabase/DatabaseContext.cpp:
(WebCore):
(WebCore::DatabaseContext::DatabaseContext):
(WebCore::DatabaseContext::~DatabaseContext):
(WebCore::DatabaseContext::contextDestroyed):
(WebCore::DatabaseContext::stop):
(WebCore::DatabaseContext::databaseThread):
(WebCore::DatabaseContext::stopDatabases):
* Modules/webdatabase/DatabaseContext.h:
(DatabaseContext):
(WebCore::DatabaseContext::scriptExecutionContext):
(WebCore::DatabaseContext::hasOpenDatabases):
(WebCore::DatabaseContext::stopDatabases):
* Modules/webdatabase/DatabaseManager.cpp:
(WebCore::DatabaseManager::manager):
(WebCore::DatabaseManager::DatabaseManager):
(WebCore::DatabaseManager::getExistingDatabaseContext):
(WebCore):
(WebCore::DatabaseManager::getDatabaseContext):
(WebCore::DatabaseManager::registerDatabaseContext):
(WebCore::DatabaseManager::unregisterDatabaseContext):
(WebCore::DatabaseManager::notifyDatabaseContextConstructed):
(WebCore::DatabaseManager::notifyDatabaseContextDestructed):
(WebCore::DatabaseManager::openDatabase):
(WebCore::DatabaseManager::openDatabaseSync):
(WebCore::DatabaseManager::hasOpenDatabases):
(WebCore::DatabaseManager::stopDatabases):
(WebCore::DatabaseManager::interruptAllDatabasesForContext):
* Modules/webdatabase/DatabaseManager.h:
(WebCore):
(DatabaseManager):
(WebCore::DatabaseManager::notifyDatabaseContextConstructed):
(WebCore::DatabaseManager::notifyDatabaseContextDestructed):
* Modules/webdatabase/DatabaseSync.cpp:
(WebCore::DatabaseSync::DatabaseSync):
* Modules/webdatabase/DatabaseSync.h:
(WebCore):
(DatabaseSync):
* Modules/webdatabase/DatabaseThread.cpp:
(WebCore::DatabaseThread::~DatabaseThread):
(WebCore::DatabaseThread::requestTermination):
* Modules/webdatabase/DatabaseTracker.cpp:
(WebCore::DatabaseTracker::canEstablishDatabase):
* dom/ActiveDOMObject.cpp:
(WebCore::ActiveDOMObject::~ActiveDOMObject):



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141166 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108227

Reviewed by Beth Dakin.

Stop using the flags in MessageID and store the flags directly in the message instead.
This is another step towards eliminating MessageID.

* Platform/CoreIPC/Connection.cpp:
(CoreIPC::Connection::SyncMessageState::processIncomingMessage):
(CoreIPC::Connection::createSyncMessageEncoder):
(CoreIPC::Connection::sendMessage):
(CoreIPC::Connection::sendSyncMessage):
(CoreIPC::Connection::sendSyncMessageFromSecondaryThread):
(CoreIPC::Connection::dispatchSyncMessage):
(CoreIPC::Connection::dispatchMessage):
* Platform/CoreIPC/Connection.h:
* Platform/CoreIPC/MessageDecoder.cpp:
(CoreIPC::MessageDecoder::MessageDecoder):
(CoreIPC::MessageDecoder::isSyncMessage):
(CoreIPC):
(CoreIPC::MessageDecoder::shouldDispatchMessageWhenWaitingForSyncReply):
* Platform/CoreIPC/MessageDecoder.h:
(MessageDecoder):
* Platform/CoreIPC/MessageEncoder.cpp:
(CoreIPC):
(CoreIPC::MessageEncoder::MessageEncoder):
(CoreIPC::MessageEncoder::~MessageEncoder):
(CoreIPC::MessageEncoder::setIsSyncMessage):
(CoreIPC::MessageEncoder::setShouldDispatchMessageWhenWaitingForSyncReply):
* Platform/CoreIPC/MessageEncoder.h:
(MessageEncoder):
* Platform/CoreIPC/MessageFlags.h: Copied from Source/WebKit2/Platform/CoreIPC/MessageEncoder.h.
(CoreIPC):
* Platform/CoreIPC/MessageID.h:
* WebKit2.xcodeproj/project.pbxproj:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141165 268f45cc-cd09-0410-ab3c-d52691b4dbfc

They are all timing out due to some configuration problem on the bots.

* platform/win/TestExpectations:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141164 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[CSS Grid Layout] Make resolveContentBasedTrackSizingFunctionsForItems reuse distributeSpaceToTracks
https://bugs.webkit.org/show_bug.cgi?id=108110

Reviewed by Tony Chang.

This change makes us match more closely the specification by reusing distributeSpaceToTracks inside
resolveContentBasedTrackSizingFunctionsForItems. This also removes some existing code duplication.

Refactoring covered by existing tests.

* rendering/RenderGrid.cpp:
(WebCore::RenderGrid::computedUsedBreadthOfGridTracks):
Updated after distributeSpaceToTracks new arguments. Copying the tracks to a Vector<GridTrack*> is
now done here.

(WebCore::RenderGrid::resolveContentBasedTrackSizingFunctionsForItems):
Removed code duplication and switched to using distributeSpaceToTracks.

(WebCore::RenderGrid::distributeSpaceToTracks):
Refactored distributeSpaceToTracks to implement the distribution of any extra space above max breadth
as it was required to pass the tests (required to properly handling min-content > max). Also changed
the arguments of the function to better match the intent of the function.

* rendering/RenderGrid.h: Updated distributeSpaceToTracks's arguments.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141163 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108207

Reviewed by Dimitri Glazkov.

There's no reason to expose removeAllShadowRoots since the only legitimate
place to call it is right as we're destroying the ElementShadow.

No new tests, just refactoring.

* dom/Element.cpp:
(WebCore::Element::~Element): Remove call to removeAllShadowRoots()
* dom/ElementShadow.h:
(WebCore::ElementShadow::~ElementShadow): Call removeAllShadowRoots().
(ElementShadow): Make removeAllShadowRoots() private.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141162 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=107833

Reviewed by Simon Fraser.

Source/WebCore:

Regression caused by http://trac.webkit.org/changeset/135629
The regression was due to faulty occlusion logic that was assuming
that drawing the background color of a render box background layer
could be skipped when the same layer also has an opaque image attached.
In the case where the background color is drawn for the purpose of
rendering a box shadow, the shadow is typically not
completely occluded by the background image because of the shadow
blur and/or offset.  This patch fixes the problem by not culling a
background draw if it is used to draw a box shadow.

Test: fast/backgrounds/gradient-background-shadow.html

* rendering/RenderBoxModelObject.cpp:
(WebCore::RenderBoxModelObject::paintFillLayerExtended):
Changing occlusion culling test to never cull background color
draw if it is used to draw a box shadow. This is because box shadows
can draw outside the border fill region.

LayoutTests:

New ref test verifies that box shadow is drawn when
background is an opaque image. Test uses an blue gradient
as background image. Reference uses blue background color.

* fast/backgrounds/gradient-background-shadow-expected.html: Added.
* fast/backgrounds/gradient-background-shadow.html: Added.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141160 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108211

Reviewed by Simon Fraser.

No new tests, no change in functionality.

* rendering/RenderLayer.h:
(WebCore::RenderLayer::enclosingStackingContainer):
  This function is similar to RenderLayer::stackingContainer, but may return the
  layer itself if it's a stacking container.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141159 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=97592

Patch by Laszlo Gombos <l.gombos@samsung.com> on 2013-01-29
Reviewed by Kenneth Rohde Christiansen.

Capture the minimum version information for the tools that are required
to build WebKit for all CMake based build systems.

* CMakeLists.txt:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141157 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108094

Reviewed by Ryosuke Niwa.

TextBreakIteratorWinCE does not contain any Windows CE specific code.
Rename it to TextBreakIteratorWchar to match the name in wtf/unicode.

* platform/text/wchar/TextBreakIteratorWchar.cpp: Renamed from Source/WebCore/platform/text/wince/TextBreakIteratorWinCE.cpp.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141156 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108212

Reviewed by Mark Hahnenberg.

Add a ContiguousData type that we use to represent contiguous property
storage.  In release builds it is simply a pointer to the correct type,
but in debug builds it also carries the data length and performs bounds
checks.  This means we don't have to add as many manual bounds assertions
when performing operations over contiguous data.

* dfg/DFGOperations.cpp:
* runtime/ArrayStorage.h:
(ArrayStorage):
(JSC::ArrayStorage::vector):
* runtime/Butterfly.h:
(JSC::ContiguousData::ContiguousData):
(ContiguousData):
(JSC::ContiguousData::operator[]):
(JSC::ContiguousData::data):
(JSC::ContiguousData::length):
(JSC):
(JSC::Butterfly::contiguousInt32):
(Butterfly):
(JSC::Butterfly::contiguousDouble):
(JSC::Butterfly::contiguous):
* runtime/JSArray.cpp:
(JSC::JSArray::sortNumericVector):
(ContiguousTypeAccessor):
(JSC::ContiguousTypeAccessor::getAsValue):
(JSC::ContiguousTypeAccessor::setWithValue):
(JSC::ContiguousTypeAccessor::replaceDataReference):
(JSC):
(JSC::JSArray::sortCompactedVector):
(JSC::JSArray::sort):
(JSC::JSArray::fillArgList):
(JSC::JSArray::copyToArguments):
* runtime/JSArray.h:
(JSArray):
* runtime/JSObject.cpp:
(JSC::JSObject::copyButterfly):
(JSC::JSObject::visitButterfly):
(JSC::JSObject::createInitialInt32):
(JSC::JSObject::createInitialDouble):
(JSC::JSObject::createInitialContiguous):
(JSC::JSObject::convertUndecidedToInt32):
(JSC::JSObject::convertUndecidedToDouble):
(JSC::JSObject::convertUndecidedToContiguous):
(JSC::JSObject::convertInt32ToDouble):
(JSC::JSObject::convertInt32ToContiguous):
(JSC::JSObject::genericConvertDoubleToContiguous):
(JSC::JSObject::convertDoubleToContiguous):
(JSC::JSObject::rageConvertDoubleToContiguous):
(JSC::JSObject::ensureInt32Slow):
(JSC::JSObject::ensureDoubleSlow):
(JSC::JSObject::ensureContiguousSlow):
(JSC::JSObject::rageEnsureContiguousSlow):
(JSC::JSObject::ensureLengthSlow):
* runtime/JSObject.h:
(JSC::JSObject::ensureInt32):
(JSC::JSObject::ensureDouble):
(JSC::JSObject::ensureContiguous):
(JSC::JSObject::rageEnsureContiguous):
(JSObject):
(JSC::JSObject::indexingData):
(JSC::JSObject::currentIndexingData):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141154 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=107599

Reviewed by Adam Barth.

When applyDeviceScaleFactorInCompositor, targetScale should exclude device scale factor.
When applyPageScaleFactorInCompositor, caret size and content sizes are in css pixels and they should be in the viewport of the new scale.

* src/WebViewImpl.cpp:
(WebKit::WebViewImpl::scrollFocusedNodeIntoRect):
(WebKit):
(WebKit::WebViewImpl::computeScaleAndScrollForFocusedNode): Extracted from scrollFocusedNodeIntoRect() to ease testing.
* src/WebViewImpl.h:
(WebViewImpl):
* tests/WebFrameTest.cpp: Updated test DivScrollEditableTest
* tests/data/get_scale_for_zoom_into_editable_test.html: Moved the logic of onload script (which seems not to work) into WebFrameTest.cpp.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141153 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108122

Reviewed by Antonio Gomes.

Adds fullscreen mode toggle (F11) option to MiniBrowser using native API.

* MiniBrowser/efl/main.c:
(on_key_down): Add 'F11' key handler.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141152 268f45cc-cd09-0410-ab3c-d52691b4dbfc

MediaStream API: A MediaStreamComponent should be able to return the MediaStreamDescriptor it belongs to
https://bugs.webkit.org/show_bug.cgi?id=108173

Reviewed by Adam Barth.

Source/Platform:

* chromium/public/WebMediaStreamComponent.h:
(WebKit):
(WebMediaStreamComponent):

Source/WebCore:

To be able to return the MediaStreamDescriptor a MediaStreamComponent belongs to the "ownership"
of the MediaStreamDescriptor needed to move from a MediaStreamTrack to the MediaStreamComponent.
This is also better from an architectonic view as well.

Patch covered by existing tests.

* Modules/mediastream/MediaStream.cpp:
(WebCore::MediaStream::MediaStream):
(WebCore::MediaStream::addTrack):
(WebCore::MediaStream::addRemoteTrack):
* Modules/mediastream/MediaStreamTrack.cpp:
(WebCore::MediaStreamTrack::create):
(WebCore::MediaStreamTrack::MediaStreamTrack):
(WebCore::MediaStreamTrack::setEnabled):
* Modules/mediastream/MediaStreamTrack.h:
(MediaStreamTrack):
* Modules/mediastream/RTCStatsRequestImpl.cpp:
(WebCore::RTCStatsRequestImpl::RTCStatsRequestImpl):
* platform/chromium/support/WebMediaStreamComponent.cpp:
(WebKit::WebMediaStreamComponent::stream):
(WebKit):
* platform/mediastream/MediaStreamComponent.h:
(WebCore):
(WebCore::MediaStreamComponent::create):
(MediaStreamComponent):
(WebCore::MediaStreamComponent::stream):
(WebCore::MediaStreamComponent::setStream):
(WebCore::MediaStreamComponent::MediaStreamComponent):
* platform/mediastream/MediaStreamDescriptor.h:
(WebCore::MediaStreamDescriptor::MediaStreamDescriptor):


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141151 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=108208

Reviewed by Beth Dakin.

This is another step towards getting rid of MessageID.

* Platform/CoreIPC/ArgumentDecoder.cpp:
(CoreIPC::ArgumentDecoder::decodeUInt8):
(CoreIPC):
* Platform/CoreIPC/ArgumentDecoder.h:
(ArgumentDecoder):
(CoreIPC::ArgumentDecoder::decode):
(CoreIPC):
* Platform/CoreIPC/ArgumentEncoder.cpp:
(CoreIPC::ArgumentEncoder::encode):
(CoreIPC):
* Platform/CoreIPC/ArgumentEncoder.h:
(ArgumentEncoder):
* Platform/CoreIPC/Connection.cpp:
(CoreIPC::Connection::sendMessage):
* Platform/CoreIPC/MessageDecoder.cpp:
(CoreIPC::MessageDecoder::MessageDecoder):
* Platform/CoreIPC/MessageDecoder.h:
(CoreIPC::MessageDecoder::messageSendFlags):
(MessageDecoder):
* Platform/CoreIPC/MessageEncoder.cpp:
(CoreIPC):
(CoreIPC::MessageEncoder::MessageEncoder):
(CoreIPC::MessageEncoder::~MessageEncoder):
(CoreIPC::MessageEncoder::setMessageSendFlags):
* Platform/CoreIPC/MessageEncoder.h:
(MessageEncoder):
* UIProcess/Authentication/AuthenticationChallengeProxy.h:
(CoreIPC):

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141150 268f45cc-cd09-0410-ab3c-d52691b4dbfc

* JavaScriptCore.vcxproj/JavaScriptCoreExports.def: Update symbols
to match JavaScriptCore.vcproj version.



git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141149 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Current error reporting method used by check-layout.js should not affect subsequent sub-tests using checking data-offset-y.
https://bugs.webkit.org/show_bug.cgi?id=105407

Patch by Pravin D <pravind.2k4@gmail.com> on 2013-01-29
Reviewed by Tony Chang.

When a testcase is processed by check-layout.js, the result is written just after the node being processed. This causes
offsetTop of subsequent sub-testcases to altered. If however if we process the nodes in the reverse order in which they
appear in the DOM tree, the result of node being processed will not affect the offsetTop of susequent nodes.

* resources/check-layout.js:

git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141147 268f45cc-cd09-0410-ab3c-d52691b4dbfc

https://bugs.webkit.org/show_bug.cgi?id=103920

Reviewed by Dimitri Glazkov.

Source/WebCore:

Update IDBCallbacks::onSuccess and IDBCallbacks::onUpgradeNeeded to
pass through a metadata parameter. While there, remove the unused
IDBTransactionBackendInterface parameter to onUpgradeNeeded.

As this is another step in the IDB refactor, I've simplified future cleanup
work by making the WebKit API code still use the old API. This
will make it possible to outright remove code on the chromium side rather
than another three-step checkin.

No new tests, as this is more refactoring.

* Modules/indexeddb/IDBCallbacks.h:
(WebCore::IDBCallbacks::onUpgradeNeeded): new method signature.
(WebCore::IDBCallbacks::onSuccess): new method signature.
* Modules/indexeddb/IDBDatabaseBackendImpl.cpp:
(WebCore::IDBDatabaseBackendImpl::VersionChangeOperation::perform):
(WebCore::IDBDatabaseBackendImpl::processPendingCalls):
(WebCore::IDBDatabaseBackendImpl::openConnection):
* Modules/indexeddb/IDBOpenDBRequest.cpp:
(WebCore::IDBOpenDBRequest::onUpgradeNeeded): use passed-in metadata.
(WebCore::IDBOpenDBRequest::onSuccess): use passed-in metadata.
* Modules/indexeddb/IDBOpenDBRequest.h:
(IDBOpenDBRequest):

Source/WebKit/chromium:

Support the new IDBCallbacks::onSuccess and IDBCallbacks::onUpgradeNeeded
while maintaining chromium compatibility by shimming in the old API
in the WebKit side. Future code will clean this up so that it is just a
pass-through as it was before.

* public/WebIDBCallbacks.h:
(WebKit):
(WebKit::WebIDBCallbacks::onSuccess): new method signature.
(WebKit::WebIDBCallbacks::onUpgradeNeeded): new method signature.
* src/IDBCallbacksProxy.cpp:
(WebKit::IDBCallbacksProxy::onSuccess): call on new method signature proxies through old API.
(WebKit):
(WebKit::IDBCallbacksProxy::onUpgradeNeeded): call on new method signature proxies through old API.
* src/IDBCallbacksProxy.h:
(IDBCallbacksProxy):
* src/WebIDBCallbacksImpl.cpp:
(WebKit::WebIDBCallbacksImpl::onSuccess): call on old WebKit proxy signature calls new API.
(WebKit):
(WebKit::WebIDBCallbacksImpl::onUpgradeNeeded): call on old WebKit proxy signature calls new API.
* src/WebIDBCallbacksImpl.h:
(WebIDBCallbacksImpl):
* tests/IDBAbortOnCorruptTest.cpp: new method signature.
(WebCore::MockIDBCallbacks::onSuccess):
* tests/IDBDatabaseBackendTest.cpp: new method signature.


git-svn-id: http://svn.webkit.org/repository/webkit/trunk@141142 268f45cc-cd09-0410-ab3c-d52691b4dbfc