WebKit-https.git
7 years ago[Win] Unreviewed build fix.
bfulgham@apple.com [Sun, 17 Aug 2014 15:35:44 +0000 (15:35 +0000)]
[Win] Unreviewed build fix.

Upload static VS2013 version of WebKitSystemInterface.

* win/lib32/WebKitSystemInterface.lib:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172695 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoRemove the obsolete WebKitEnableCoalescedUpdatesPreferenceKey
mjs@apple.com [Sat, 16 Aug 2014 22:52:58 +0000 (22:52 +0000)]
Remove the obsolete WebKitEnableCoalescedUpdatesPreferenceKey
https://bugs.webkit.org/show_bug.cgi?id=136024

Reviewed by Filip Pizlo.

* WebView/WebFrameView.mm:
(-[WebFrameView initWithFrame:]): Remove code that respects
this key. It was SPI and is no longer used; also it was
only relevant for versions of OS X that WebKit no longer supports.
* WebView/WebPreferenceKeysPrivate.h: Remove declaration.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172694 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoREGRESSION (r164173): Safari AutoFill button in popover, default buttons on iCloud...
dino@apple.com [Sat, 16 Aug 2014 20:32:53 +0000 (20:32 +0000)]
REGRESSION (r164173): Safari AutoFill button in popover, default buttons on iCloud control panel and Safari Safe Browsing pages don't have default appearance
https://bugs.webkit.org/show_bug.cgi?id=136011

Reviewed by Sam Weinig.

When we render some controls, we need to make sure the NSWindow subclass we
use as the host of the AppKit views identifies itself as the key window (if
the control state is correct). We already have a flag tracking this for
key appearance - just make sure we do the same for isKeyWindow.

* platform/mac/ThemeMac.mm:
(-[WebCoreThemeWindow isKeyWindow]): New override to mirror the static
variable for key appearance.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172693 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoUpdate contributors.json for my new email
fpizlo@apple.com [Sat, 16 Aug 2014 19:26:07 +0000 (19:26 +0000)]
Update contributors.json for my new email
https://bugs.webkit.org/show_bug.cgi?id=136021

Patch by Mark Hahnenberg <mhahnenb@gmail.com> on 2014-08-16
Reviewed by Filip Pizlo.

* Scripts/webkitpy/common/config/contributors.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172692 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoRename DiskCacheMonitor to NetworkDiskCacheMonitor
psolanki@apple.com [Sat, 16 Aug 2014 19:01:37 +0000 (19:01 +0000)]
Rename DiskCacheMonitor to NetworkDiskCacheMonitor
https://bugs.webkit.org/show_bug.cgi?id=135897

Reviewed by Andreas Kling.

In preparation for moving DiskCacheMonitor code to WebCore in bug 135896, rename the WebKit2
class to NetworkDiskCacheMonitor.

* NetworkProcess/mac/NetworkDiskCacheMonitor.h: Renamed from Source/WebKit2/NetworkProcess/mac/DiskCacheMonitor.h.
(WebKit::NetworkDiskCacheMonitor::resourceRequest):
* NetworkProcess/mac/NetworkDiskCacheMonitor.mm: Renamed from Source/WebKit2/NetworkProcess/mac/DiskCacheMonitor.mm.
(WebKit::NetworkDiskCacheMonitor::monitorFileBackingStoreCreation):
(WebKit::NetworkDiskCacheMonitor::NetworkDiskCacheMonitor):
(WebKit::NetworkDiskCacheMonitor::messageSenderConnection):
(WebKit::NetworkDiskCacheMonitor::messageSenderDestinationID):
* NetworkProcess/mac/NetworkResourceLoaderMac.mm:
(WebKit::NetworkResourceLoader::willCacheResponseAsync):
* WebKit2.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172691 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[GTK] build fails with error: cannot allocate an object of abstract type 'WebKit...
commit-queue@webkit.org [Sat, 16 Aug 2014 14:26:17 +0000 (14:26 +0000)]
[GTK] build fails with error: cannot allocate an object of abstract type 'WebKit::PageClientImpl'
https://bugs.webkit.org/show_bug.cgi?id=136017

Patch by Byungseon Shin <sun.shin@lge.com> on 2014-08-16
Reviewed by Gyuyoung Kim.

* UIProcess/API/gtk/PageClientImpl.cpp:
(WebKit::PageClientImpl::didFirstVisuallyNonEmptyLayoutForMainFrame):
(WebKit::PageClientImpl::didFinishLoadForMainFrame):
* UIProcess/API/gtk/PageClientImpl.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172690 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Improve performance of selection range changes viewing Scripts timeline
commit-queue@webkit.org [Sat, 16 Aug 2014 09:25:27 +0000 (09:25 +0000)]
Web Inspector: Improve performance of selection range changes viewing Scripts timeline
https://bugs.webkit.org/show_bug.cgi?id=136015

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-08-16
Reviewed by Timothy Hatcher.

* UserInterface/Base/Utilities.js:
(clamp):
Helper function to clamp a value between a min and max.

* UserInterface/Models/ProfileNode.js:
(WebInspector.ProfileNode.prototype.get startTime):
(WebInspector.ProfileNode.prototype.get endTime):
These can be quickly calculated, so avoid full calculation to grab the values.

* UserInterface/Views/ScriptTimelineView.js:
(WebInspector.ScriptTimelineView.prototype.updateLayout):
Update ranges with a smart helper, instead of blindly doing it and needing refresh.

* UserInterface/Views/ProfileNodeDataGridNode.js:
(WebInspector.ProfileNodeDataGridNode.prototype.get rangeEndTime):
(WebInspector.ProfileNodeDataGridNode.prototype.get data):
(WebInspector.ProfileNodeDataGridNode.prototype.updateRangeTimes):
(WebInspector.ProfileNodeDataGridNode.prototype.set rangeStartTime): Deleted.
(WebInspector.ProfileNodeDataGridNode.prototype.set rangeEndTime): Deleted.
* UserInterface/Views/ScriptTimelineDataGridNode.js:
(WebInspector.ScriptTimelineDataGridNode.prototype.updateRangeTimes):
(WebInspector.ScriptTimelineDataGridNode.prototype.set rangeStartTime): Deleted.
(WebInspector.ScriptTimelineDataGridNode.prototype.set rangeEndTime): Deleted.
When updating the range selection, compare to the last range selection
on this DataGridNode. If the visible portion of this node's time range
changes, we need a refresh. Otherwise, we don't need a refresh.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172689 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Update Scripts/update-LegacyInspectorBackendCommands.rb for the new...
commit-queue@webkit.org [Sat, 16 Aug 2014 05:49:19 +0000 (05:49 +0000)]
Web Inspector: Update Scripts/update-LegacyInspectorBackendCommands.rb for the new generator
https://bugs.webkit.org/show_bug.cgi?id=136006

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-08-15
Reviewed by Timothy Hatcher.

* Scripts/update-LegacyInspectorBackendCommands.rb:
Use the new generator and pass the appropriate flags.

* UserInterface/Protocol/Legacy/6.0/InspectorWebBackendCommands.js:
* UserInterface/Protocol/Legacy/7.0/InspectorWebBackendCommands.js:
Newly generated files using the new generator.

* Versions/Inspector-iOS-6.0.json:
Remove unsupported features.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172688 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: rewrite CodeGeneratorInspector to be modular and testable
burg@cs.washington.edu [Sat, 16 Aug 2014 04:38:00 +0000 (04:38 +0000)]
Web Inspector: rewrite CodeGeneratorInspector to be modular and testable
https://bugs.webkit.org/show_bug.cgi?id=131596

Unreviewed gardening to rebaseline inspector generator tests after addressing review comments.

* inspector/scripts/tests/expected/commands-with-async-attribute.json-result:
* inspector/scripts/tests/expected/commands-with-optional-call-return-parameters.json-result:
* inspector/scripts/tests/expected/domains-with-varying-command-sizes.json-result:
* inspector/scripts/tests/expected/events-with-optional-parameters.json-result:
* inspector/scripts/tests/expected/same-type-id-different-domain.json-result:
* inspector/scripts/tests/expected/shadowed-optional-type-setters.json-result:
* inspector/scripts/tests/expected/type-declaration-aliased-primitive-type.json-result:
* inspector/scripts/tests/expected/type-declaration-array-type.json-result:
* inspector/scripts/tests/expected/type-declaration-enum-type.json-result:
* inspector/scripts/tests/expected/type-declaration-object-type.json-result:
* inspector/scripts/tests/expected/type-requiring-runtime-casts.json-result:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172687 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoUnreviewed build fix for some GTK bots after r172655.
burg@cs.washington.edu [Sat, 16 Aug 2014 04:36:42 +0000 (04:36 +0000)]
Unreviewed build fix for some GTK bots after r172655.

Some bots use Python 2.6, which lacks the 'flags' named parameter for re.sub.

* inspector/scripts/codegen/generator.py:
(Generator.stylized_name_for_enum_value): Do things the old-school way.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172686 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[EFL] Add ewk_intro.h to describe main page of doxygen
ryuan.choi@samsung.com [Sat, 16 Aug 2014 04:35:59 +0000 (04:35 +0000)]
[EFL] Add ewk_intro.h to describe main page of doxygen
https://bugs.webkit.org/show_bug.cgi?id=136013

Reviewed by Gyuyoung Kim.

* PlatformEfl.cmake:
* UIProcess/API/efl/EWebKit2.h.in:
* UIProcess/API/efl/ewk_intro.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172685 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Remove forced layouts during recording caused by striped background...
commit-queue@webkit.org [Sat, 16 Aug 2014 04:05:50 +0000 (04:05 +0000)]
Web Inspector: Remove forced layouts during recording caused by striped background element resizing
https://bugs.webkit.org/show_bug.cgi?id=136009

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-08-15
Reviewed by Timothy Hatcher.

We used to manually resize and a custom striped background behind
the timeline tree outline content area. This can happen frequently
during timeline recordings causing forced layouts. We can eliminate
the hack and instead put the stripe background on the tree outline.

* UserInterface/Views/TimelineSidebarPanel.js:
(WebInspector.TimelineSidebarPanel.prototype.updateCustomContentOverflow): Deleted.
* UserInterface/Views/TimelineSidebarPanel.css:
(body.mac-platform.legacy .sidebar > .panel.navigation.timeline > .content .item > .status > .status-button):
(.sidebar > .panel.navigation.timeline > .content > .stripe-background): Deleted.
(.sidebar > .panel.navigation.timeline.timeline-content-view-showing > .content > .stripe-background): Deleted.
Eliminate the striped background element.

* UserInterface/Views/TimelineView.css:
(.panel.navigation.timeline > .content > .navigation-sidebar-panel-content-tree-outline):
(.panel.navigation.timeline.timeline-content-view-showing > .content > .navigation-sidebar-panel-content-tree-outline):
Give the timeline content tree outlines the striped background.

* UserInterface/Views/NavigationSidebarPanel.js:
(WebInspector.NavigationSidebarPanel.prototype._updateContentOverflowShadowVisibility):
(WebInspector.NavigationSidebarPanel.prototype.updateCustomContentOverflow): Deleted.
Remove the now unused updateCustomContentOverflow method.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172682 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoDo not use FloatRect::infiniteRect() to flag full repaints.
zalan@apple.com [Sat, 16 Aug 2014 04:03:42 +0000 (04:03 +0000)]
Do not use FloatRect::infiniteRect() to flag full repaints.
https://bugs.webkit.org/show_bug.cgi?id=135900

Reviewed by Simon Fraser.

Converting FloatRect::infiniteRect() to IntRect leads to value overflow
and we end up with invalid repaint rectangle. Use a boolean flag to indicate
full repaint request.

Covered by existing tests.

Source/WebCore:

* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::GraphicsLayerCA):
(WebCore::GraphicsLayerCA::setNeedsDisplay):
(WebCore::GraphicsLayerCA::setNeedsDisplayInRect):
(WebCore::GraphicsLayerCA::repaintLayerDirtyRects):
* platform/graphics/ca/GraphicsLayerCA.h:
* platform/graphics/ca/PlatformCALayer.h:
* platform/graphics/ca/TileGrid.cpp:
(WebCore::TileGrid::setTileNeedsDisplayInRect):
* platform/graphics/ca/mac/PlatformCALayerMac.h:
* platform/graphics/ca/mac/PlatformCALayerMac.mm:
(PlatformCALayerMac::setNeedsDisplay):
(PlatformCALayerMac::setNeedsDisplayInRect):

Source/WebKit2:

* WebProcess/WebPage/mac/PlatformCALayerRemote.cpp:
(WebKit::PlatformCALayerRemote::setNeedsDisplayInRect):
(WebKit::PlatformCALayerRemote::setNeedsDisplay):
* WebProcess/WebPage/mac/PlatformCALayerRemote.h:
* WebProcess/WebPage/mac/PlatformCALayerRemoteCustom.h:
* WebProcess/WebPage/mac/PlatformCALayerRemoteCustom.mm:
(WebKit::PlatformCALayerRemoteCustom::setNeedsDisplayInRect):
(WebKit::PlatformCALayerRemoteCustom::setNeedsDisplay):
* WebProcess/WebPage/mac/PlatformCALayerRemoteTiledBacking.cpp:
(WebKit::PlatformCALayerRemoteTiledBacking::setNeedsDisplayInRect):
(WebKit::PlatformCALayerRemoteTiledBacking::setNeedsDisplay):
* WebProcess/WebPage/mac/PlatformCALayerRemoteTiledBacking.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172681 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoREGRESSION (r172660): WebKit2.TerminateTwice asserts
ap@apple.com [Sat, 16 Aug 2014 02:57:58 +0000 (02:57 +0000)]
REGRESSION (r172660): WebKit2.TerminateTwice asserts
https://bugs.webkit.org/show_bug.cgi?id=136012

Source/WebKit2:

* UIProcess/WebPageProxy.cpp: (WebKit::WebPageProxy::terminateProcess):
Added bug number to a fixme.

Tools:

* TestWebKitAPI/Tests/WebKit2/TerminateTwice.cpp: Skipped the test in debug mode.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172680 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoUnify the modes style resolution modes SharingRules and StyleInvalidation
benjamin@webkit.org [Sat, 16 Aug 2014 02:49:30 +0000 (02:49 +0000)]
Unify the modes style resolution modes SharingRules and StyleInvalidation
https://bugs.webkit.org/show_bug.cgi?id=135969

Reviewed by Andreas Kling.

Unifying the Collector and SelectorCheckers was done in previous patches,
this is just removing the enums.

* css/ElementRuleCollector.cpp:
(WebCore::ElementRuleCollector::collectMatchingRules):
(WebCore::ElementRuleCollector::hasAnyMatchingRules):
* css/SelectorChecker.cpp:
(WebCore::SelectorChecker::match):
* css/SelectorChecker.h:
* css/StyleInvalidationAnalysis.cpp:
(WebCore::invalidateStyleRecursively):
* cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateMarkPseudoStyleForPseudoElement):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172679 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoUnreviewed build fix on the EFL port since r172635
ryuan.choi@samsung.com [Sat, 16 Aug 2014 02:32:54 +0000 (02:32 +0000)]
Unreviewed build fix on the EFL port since r172635

* UIProcess/WebPageProxy.cpp: Guard removeNavigationGestureSnapshot with PLATFORM(MAC) macro.
* UIProcess/WebPageProxy.h: Ditto.
* UIProcess/efl/WebViewEfl.h: Added didFirstVisuallyNonEmptyLayoutForMainFrame and didFinishLoadForMainFrame as dummy.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172678 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoUnreviewed build fix attempt on the CMake based pots since r172665
ryuan.choi@samsung.com [Sat, 16 Aug 2014 02:09:37 +0000 (02:09 +0000)]
Unreviewed build fix attempt on the CMake based pots since r172665

* TestWebKitAPI/CMakeLists.txt: Added JavaScriptCore/interpreter to the include list.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172672 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoChange callToJavaScript and callToNativeFunction so their callFrames match the native...
msaboff@apple.com [Sat, 16 Aug 2014 01:45:40 +0000 (01:45 +0000)]
Change callToJavaScript and callToNativeFunction so their callFrames match the native calling conventions
https://bugs.webkit.org/show_bug.cgi?id=131578

Reviewed by Geoffrey Garen.

Renamed callToJavaScript and callToNativeFunction to vmEntryToJavaScript and vmEntryToNative,
respectively.  Eliminated the sentinel frame and replaced it with the structure VMEntryRecord
that appears in the "locals" area of a VM entry stack frame.  Changed the order that
vmEntryToJavaScript and vmEntryToNative creates their stack frames to be native calling
convention compliant.  That is to save prior frame pointer, save callee save registers, then
allocate and populate the VMEntryRecord, and finally allocate a CallFrame for the JS function
that vmEntryToJavaScript will invoke.  The top most vm entry frame pointer is saved in
VM::topVMEntryFrame.  The vmEntry functions save prior contents of VM::topVMEntryFrame
along with the VM and VM::topCallFrame in the VMEntryRecord it places on the stack.  Starting
at VM::topCallFrame, the stack can be walked using these VMEntryRecords.

Arbitrary stack unwinding is now handled either iteratively by loading VM::topVMEntryFrame
into a local variable and using CallFrame::callerFrame(VMEntryFrame*&) or by using StackVisitor.
Given that the stack is effectively a singly linked list, general stack unwinding needs to use
one of these two methods.

* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj.filters:
* JavaScriptCore.xcodeproj/project.pbxproj:
Addition of VMEntryRecord.h

* bytecode/BytecodeList.json:
Renaming of llint helper opcodes due to renaming callToJavaScript and callToNativeFunction.

* debugger/Debugger.cpp:
(JSC::Debugger::stepOutOfFunction):
(JSC::Debugger::returnEvent):
(JSC::Debugger::didExecuteProgram):
* jsc.cpp:
(functionDumpCallFrame):
* jit/JITOperations.cpp:
Changed unwinding to use CallFrame::callerFrame(VMEntryFrame*&).

* bytecode/CodeBlock.cpp:
(JSC::RecursionCheckFunctor::RecursionCheckFunctor):
(JSC::RecursionCheckFunctor::operator()):
(JSC::RecursionCheckFunctor::didRecurse):
(JSC::CodeBlock::noticeIncomingCall):
* debugger/DebuggerCallFrame.cpp:
(JSC::FindCallerMidStackFunctor::FindCallerMidStackFunctor):
(JSC::FindCallerMidStackFunctor::operator()):
(JSC::FindCallerMidStackFunctor::getCallerFrame):
(JSC::DebuggerCallFrame::callerFrame):
* interpreter/VMInspector.cpp:
(JSC::CountFramesFunctor::CountFramesFunctor):
(JSC::CountFramesFunctor::operator()):
(JSC::CountFramesFunctor::count):
(JSC::VMInspector::countFrames):
* runtime/VM.cpp:
(JSC::VM::VM):
(JSC::FindFirstCallerFrameWithCodeblockFunctor::FindFirstCallerFrameWithCodeblockFunctor):
(JSC::FindFirstCallerFrameWithCodeblockFunctor::operator()):
(JSC::FindFirstCallerFrameWithCodeblockFunctor::foundCallFrame):
(JSC::FindFirstCallerFrameWithCodeblockFunctor::index):
(JSC::VM::throwException):
Changed unwinding to use StackVisitor including added functor classes.

* interpreter/CallFrame.cpp:
(JSC::CallFrame::callerFrame):
Added new flavor of callerFrame() that can iteratively unwind the stack.

* interpreter/CallFrame.h:
(JSC::ExecState::callerFrame): Changed callerFrame() to use private common helper.
(JSC::ExecState::callerFrameOrVMEntryFrame): Deleted.
(JSC::ExecState::isVMEntrySentinel): Deleted.
(JSC::ExecState::vmEntrySentinelCallerFrame): Deleted.
(JSC::ExecState::initializeVMEntrySentinelFrame): Deleted.
(JSC::ExecState::callerFrameSkippingVMEntrySentinel): Deleted.
(JSC::ExecState::vmEntrySentinelCodeBlock): Deleted.

* interpreter/CallFrame.h:
(JSC::ExecState::init):
(JSC::ExecState::topOfFrame):
(JSC::ExecState::currentVPC):
(JSC::ExecState::setCurrentVPC):
Eliminated unneded checking of sentinel frame.

* interpreter/Interpreter.cpp:
(JSC::unwindCallFrame):
(JSC::Interpreter::getStackTrace): Updated for unwidning changes.
(JSC::Interpreter::unwind): Eliminated unneeded sentinel frame check.

* interpreter/Interpreter.cpp:
(JSC::Interpreter::executeCall):
(JSC::Interpreter::executeConstruct):
* jit/JITStubs.h:
* llint/LLIntThunks.cpp:
(JSC::callToJavaScript): Deleted.
(JSC::callToNativetion): Deleted.
(JSC::vmEntryToJavaScript):
(JSC::vmEntryToNative):
* llint/LLIntThunks.h:
Updated for vmEntryToJavaScript and vmEntryToNative name changes.

* interpreter/Interpreter.h:
(JSC::TopCallFrameSetter::TopCallFrameSetter):
(JSC::TopCallFrameSetter::~TopCallFrameSetter):
Eliminated unneeded sentinel frame check.

* interpreter/Interpreter.h:
(JSC::NativeCallFrameTracer::NativeCallFrameTracer):
Removed sentinel specific constructor.

* interpreter/StackVisitor.cpp:
(JSC::StackVisitor::StackVisitor):
(JSC::StackVisitor::readFrame):
(JSC::StackVisitor::readNonInlinedFrame):
(JSC::StackVisitor::readInlinedFrame):
(JSC::StackVisitor::Frame::print):
* interpreter/StackVisitor.h:
(JSC::StackVisitor::Frame::callerIsVMEntry):
Changes for unwinding using CallFrame::callerFrame(VMEntryFrame*&).  Also added field that
indicates when about to step over a VM entry frame.

* interpreter/VMEntryRecord.h: Added.
(JSC::VMEntryRecord::prevTopCallFrame):
(JSC::VMEntryRecord::prevTopVMEntryFrame):
New struct to record prior state of VM's notion of VM entry and top call frames.

* jit/JITCode.cpp:
(JSC::JITCode::execute):
Use new vmEntryToJavaScript and vmEntryToNative name.

* llint/LLIntOffsetsExtractor.cpp: Added include for VMEntryRecord.h.

* llint/LowLevelInterpreter.asm:
* llint/LowLevelInterpreter32_64.asm:
* llint/LowLevelInterpreter64.asm:
Offline assembly implementation of creating stack frame with VMEntryRecord and well as restoring
relevent VM fields when exiting the VM.  Added a helper that returns a VMEntryRecord given
a pointer to the VM entry frame.

* llint/LLIntThunks.cpp:
(JSC::vmEntryRecord):
* llint/LowLevelInterpreter.cpp:
(JSC::CLoop::execute):
C Loop changes to mirror the assembly changes.

* runtime/VM.h:
Added topVMEntryFrame field.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172665 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoCSS JIT: compile :-webkit-full-page-media
benjamin@webkit.org [Sat, 16 Aug 2014 01:37:29 +0000 (01:37 +0000)]
CSS JIT: compile :-webkit-full-page-media
https://bugs.webkit.org/show_bug.cgi?id=135968

Reviewed by Alexey Proskuryakov.

Add the pseudo class :-webkit-full-page-media with an unoptimized function call.

It would be possible to inline it, but this selector is so uncommon I don't think
it is worth having any additional complexity for it.

* css/SelectorChecker.cpp:
(WebCore::SelectorChecker::checkOne):
* css/SelectorCheckerTestFunctions.h:
(WebCore::isMediaDocument):
* cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::addPseudoClassType):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172663 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Inspect Element doesn't always highlight the correct element when...
jonowells@apple.com [Sat, 16 Aug 2014 00:56:44 +0000 (00:56 +0000)]
Web Inspector: Inspect Element doesn't always highlight the correct element when inspector is closed before inspecting
https://bugs.webkit.org/show_bug.cgi?id=135995

Reviewed by Joseph Pecoraro.

Change so if an element is inspected via the Inspect Element context menu, selected
node restoration is disallowed until the main resource in the main frame changes.
This way an attempt by a sub-frame will not change the highlighted element in the
Main Frame's DOM tree unexpectedly.

* UserInterface/Controllers/DOMTreeManager.js:
(WebInspector.DOMTreeManager):
(WebInspector.DOMTreeManager.prototype._mainResourceDidChange):
(WebInspector.DOMTreeManager.prototype.inspectElement):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172662 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[Win] Link to static VS2013 runtime library.
bfulgham@apple.com [Sat, 16 Aug 2014 00:42:24 +0000 (00:42 +0000)]
[Win] Link to static VS2013 runtime library.
https://bugs.webkit.org/show_bug.cgi?id=136007
<rdar://problem/18037766>

Reviewed by Dean Jackson.

* win/tools/vsprops/common.props: Update link command.
* win/tools/vsprops/debugsuffix.props: Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172661 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoImprove page to process relationship tracking
ap@apple.com [Fri, 15 Aug 2014 23:56:57 +0000 (23:56 +0000)]
Improve page to process relationship tracking
https://bugs.webkit.org/show_bug.cgi?id=135996
<rdar://problem/16991213>

Reviewed by Sam Weinig.

* UIProcess/VisitedLinkProvider.cpp:
(WebKit::VisitedLinkProvider::removeAll):
(WebKit::VisitedLinkProvider::pendingVisitedLinksTimerFired):
(WebKit::VisitedLinkProvider::sendTable):
Added assertions for m_processes only having valid entries.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::reattachToWebProcess): When attaching to a new process,
tell the old process that the page is not longer associated with it, avoiding
a potential stale pointer.
If re-attached to an existing process, make sure that we perform all the same
registrations as after having launched a new process. This substantially improves
the behavior when the number of open tabs is over process limit.
(WebKit::WebPageProxy::reattachToWebProcessWithItem): Added ASSERT(!isValid())
to avoid confusion. All other calls to reattachToWebProcess() have this as a
runtime check, but reattachToWebProcessWithItem() is only called for valid pages.
(WebKit::WebPageProxy::terminateProcess): Added an assertion with a FIXME for
something that will need to be fixed another day.

* UIProcess/WebPageProxy.h: Removed an unimplemented function.

* UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::addExistingWebPage): Added assertions for page map sanity.
(WebKit::WebProcessProxy::removeWebPage): Added a check for page state being Terminated
already. This avoids an assertion failure that happened under the new call to
removeWebPage() in reattachToWebProcess(), as we are now calling it for terminated
processes that are not in WebContext::m_processes any more.
(WebKit::WebProcessProxy::didFinishLaunching): Added an assertion that page agrees
about using this process.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172660 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoFix gliding in AxisScrollSnapAnimator for Mac
commit-queue@webkit.org [Fri, 15 Aug 2014 23:35:21 +0000 (23:35 +0000)]
Fix gliding in AxisScrollSnapAnimator for Mac
https://bugs.webkit.org/show_bug.cgi?id=135971

Patch by Wenson Hsieh <wenson_hsieh@apple.com> on 2014-08-15
Reviewed by Dean Jackson.

Previously, momentum events that are handled immediately after finishing a glide animation cause a second, extremely
slow glide animation to trigger. To fix this, I added a new state DestinationReached. During this state, additional
momentum events that are handled after the glide animation completes will not fill up the momentum window and trigger
a second glide event.

* platform/mac/AxisScrollSnapAnimator.h:
* platform/mac/AxisScrollSnapAnimator.mm:
(WebCore::AxisScrollSnapAnimator::AxisScrollSnapAnimator):
(WebCore::AxisScrollSnapAnimator::handleWheelEvent):
(WebCore::AxisScrollSnapAnimator::scrollSnapAnimationUpdate):
(WebCore::AxisScrollSnapAnimator::endScrollSnapAnimation):
(WebCore::AxisScrollSnapAnimator::computeGlideDelta):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172659 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[iOS] Teach postprocess-headers.sh to remove code guarded by ENABLE_IOS_GESTURE_EVENTS
dbates@webkit.org [Fri, 15 Aug 2014 23:19:44 +0000 (23:19 +0000)]
[iOS] Teach postprocess-headers.sh to remove code guarded by ENABLE_IOS_GESTURE_EVENTS
and ENABLE_TOUCH_EVENTS if applicable
https://bugs.webkit.org/show_bug.cgi?id=135993

Reviewed by Andy Estes.

Add support infrastructure to remove code from public and private API headers that is
guarded by macro define ENABLE_{IOS_GESTURE, TOUCH}_EVENTS when these features are disabled.

We aren't using this functionality yet, but we will soon.

* postprocess-headers.sh:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172658 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[MSE] Implement a maximum buffer size for SourceBuffer
eric.carlson@apple.com [Fri, 15 Aug 2014 23:09:19 +0000 (23:09 +0000)]
[MSE] Implement a maximum buffer size for SourceBuffer
https://bugs.webkit.org/show_bug.cgi?id=135614

Reviewed by Jer Noble.

Implement the MSE coded frame eviction algorithm: when new buffers are appended attempt
to keep the amount of data buffered below a maximum size (which is determined by the media
session manager) by freeing coded frames as follows:
    1 - Free frames in 30 second chunks from 0 to 30 seconds before the current time.
    2 - If there are time ranges after the range with the current time, free frames in
        30 second chunks from duration back to the beginning of the time range after
        current time.

For now we DO NOT throw a QUOTA_EXCEEDED_ERR when we are not able to free up enough
space to remain below the prescribed quota, because some big name, widely deployed,
code bases ignore the error and continue appending data as though the failed append
succeeded, leading to a corrupted bitstream and failure to play.

* Modules/mediasource/SampleMap.cpp:
(WebCore::SampleMap::addSample): Drive-by performance optimization: sample->presentationTime()
    is used more than once, stash it in a local variable.
(WebCore::SampleMap::removeSample): Ditto.

* Modules/mediasource/SourceBuffer.cpp:
(WebCore::logRanges): Debug-only function to log TimeRanges.
(WebCore::SourceBuffer::SourceBuffer):
(WebCore::SourceBuffer::remove): Optimize logging. Buffer full and coded frame eviction logic
    is in SourceBuffer instead of SourceBufferPrivate.
(WebCore::SourceBuffer::appendBufferInternal): Ditto.
(WebCore::SourceBuffer::sourceBufferPrivateAppendComplete): Set m_bufferFull when more data
    has been buffered than allowed.
(WebCore::removeSamplesFromTrackBuffer): Add logging.
(WebCore::SourceBuffer::removeCodedFrames): Improve logging. Avoid debug build assert when
    current time is after last enqueued presentation time.
(WebCore::SourceBuffer::evictCodedFrames): The coded frame eviction algorithm.
(WebCore::SourceBuffer::maximumBufferSize): Return the maximum amount of data that can
    be buffered.
(WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample): Improve logging. Don't attempt
    to create a PlatformTimeRanges with inverted start and end to avoid an assert in debug.
(WebCore::SourceBuffer::hasAudio): New, return true if there are any audio tracks.
(WebCore::SourceBuffer::hasVideo): New, return true if there are any video tracks.
(WebCore::SourceBuffer::sourceBufferPrivateHasAudio): Call hasAudio.
(WebCore::SourceBuffer::sourceBufferPrivateHasVideo): Call hasVideo.
(WebCore::SourceBuffer::hasCurrentTime): Return true if currentTime is greater than duration.
(WebCore::SourceBuffer::hasFutureTime): Ditto.
(WebCore::SourceBuffer::extraMemoryCost): Return the amount of data buffered: the size of
    the input buffer plus the size of all track samples.
(WebCore::SourceBuffer::reportExtraMemoryCost): Move buffered size calculation to extraMemoryCost.
(WebCore::SourceBuffer::document): Document accessor.
* Modules/mediasource/SourceBuffer.h: Drive-by size optimization by moving all bool member
    variables to the end of class.

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::refreshCachedTime): Drive-by removal of overly chatty logging.
(WebCore::HTMLMediaElement::maximumSourceBufferSize): New, maximum source buffer size.
* html/HTMLMediaElement.h:

* html/HTMLMediaSession.cpp:
(WebCore::HTMLMediaSession::maximumMediaSourceBufferSize): New, get maximum source buffer
    size from settings, return full amount for an SourceBuffer with video and audio, return 5%
    of the maximum for an audio-only SourceBuffer.
* html/HTMLMediaSession.h:

* page/Settings.in: Add maximumSourceBufferSize. Default value is enough for approximately
    five minutes of 1080p video and stereo audio.

* platform/graphics/PlatformTimeRanges.cpp:
(WebCore::PlatformTimeRanges::totalDuration): Drive-by optimization.
(WebCore::PlatformTimeRanges::dump): New, allow a PlatformTimeRanges to be printed.
* platform/graphics/PlatformTimeRanges.h:

* platform/graphics/SourceBufferPrivate.h: Delete evictCodedFrames and isFull, that logic
    is not in SourceBuffer.
* platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.h:
* platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:
(WebCore::SourceBufferPrivateAVFObjC::processCodedFrame): Drive-by logging fix.
(WebCore::SourceBufferPrivateAVFObjC::evictCodedFrames): Deleted.
(WebCore::SourceBufferPrivateAVFObjC::isFull): Deleted.

* platform/mock/mediasource/MockSourceBufferPrivate.cpp:
(WebCore::MockSourceBufferPrivate::evictCodedFrames): Deleted.
(WebCore::MockSourceBufferPrivate::isFull): Deleted.
* platform/mock/mediasource/MockSourceBufferPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172657 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoREGRESSION: Parts of the route/route options windows are invisible at maps.google.com
zalan@apple.com [Fri, 15 Aug 2014 22:43:50 +0000 (22:43 +0000)]
REGRESSION: Parts of the route/route options windows are invisible at maps.google.com
https://bugs.webkit.org/show_bug.cgi?id=135977
<rdar://problem/17961698>

Reviewed by Andreas Kling.

When the style change requires both layout and repaint, the style diff is set to
'layout is needed' (and masks the repaint change) as normally layout is followed by a repaint.
However, in compositing context, layout may not be followed by repaint, so when the style diff
indicates layout, we need to figure out whether it implies repaint too.
This logic should eventually be moved from RenderElement to RenderLayerModelObject/RenderLayer.
-tracked here: webkit.org/b/135990

Source/WebCore:

Test: compositing/cliprect-and-position-change-on-compositing-layer.html

* rendering/RenderElement.cpp:
(WebCore::RenderElement::setNeedsPositionedMovementLayout):
* rendering/style/RenderStyle.cpp:
(WebCore::RenderStyle::changeRequiresLayerRepaint):
(WebCore::RenderStyle::diffRequiresLayerRepaint):
(WebCore::RenderStyle::diffRequiresRepaint): Deleted.
* rendering/style/RenderStyle.h:
* rendering/style/RenderStyleConstants.h:

LayoutTests:

* compositing/cliprect-and-position-change-on-compositing-layer-expected.html: Added.
* compositing/cliprect-and-position-change-on-compositing-layer.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172656 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: rewrite CodeGeneratorInspector to be modular and testable
burg@cs.washington.edu [Fri, 15 Aug 2014 22:40:33 +0000 (22:40 +0000)]
Web Inspector: rewrite CodeGeneratorInspector to be modular and testable
https://bugs.webkit.org/show_bug.cgi?id=131596

Reviewed by Joseph Pecoraro.

Source/JavaScriptCore:

Replace CodeGeneratorInspector.py with generate-inspector-protocol-bindings.py.
The new generator decouples parsing and typechecking a model of the protocol from
code generation. Each generated file is created by a different subclass of Generator.
Helper methods to compute various type signatures are shared among generators.

This patch introduces a test harness and a test suite that covers all functionality.

Aside from hooking up the new inspector bindings generator to the build system,
there are a few comingled changes that would be painful to split from the main
patch:

Convert protocol enumeration types from struct-namespaced enums to C++ scoped enums.

Move all runtimeCast(), assertValueHasExpectedType(), and RuntimeCastHelper methods to static
methods of BindingTraits specializations.

Together, these changes reduce duplication and make it possible to forward-declare
all protocol enum and object types, reducing weird ordering dependencies between domains.

* CMakeLists.txt:
* DerivedSources.make:
* JavaScriptCore.vcxproj/copy-files.cmd:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj.filters: Add inspector scripts to solution filters.
* JavaScriptCore.xcodeproj/project.pbxproj:
* inspector/ConsoleMessage.cpp: Convert to scoped enums.
(Inspector::messageSourceValue):
(Inspector::messageTypeValue):
(Inspector::messageLevelValue):
* inspector/InjectedScript.cpp: Convert to scoped enums and BindingTraits.
(Inspector::InjectedScript::getFunctionDetails):
(Inspector::InjectedScript::getProperties):
(Inspector::InjectedScript::getInternalProperties):
(Inspector::InjectedScript::wrapCallFrames):
(Inspector::InjectedScript::wrapObject):
(Inspector::InjectedScript::wrapTable):
* inspector/InjectedScriptBase.cpp: Convert InspectorValue::Type to a scoped enum.
(Inspector::InjectedScriptBase::makeEvalCall):
* inspector/InjectedScriptManager.cpp:
(Inspector::InjectedScriptManager::injectedScriptForObjectId):
* inspector/InspectorTypeBuilder.h:
(Inspector::TypeBuilder::Array::create):
(Inspector::TypeBuilder::StructItemTraits::pushRefPtr):
(Inspector::TypeBuilder::ArrayItemHelper<String>::Traits::pushRaw):
(Inspector::TypeBuilder::ArrayItemHelper<int>::Traits::pushRaw):
(Inspector::TypeBuilder::ArrayItemHelper<double>::Traits::pushRaw):
(Inspector::TypeBuilder::ArrayItemHelper<bool>::Traits::pushRaw):
(Inspector::TypeBuilder::ArrayItemHelper<InspectorValue>::Traits::pushRefPtr):
(Inspector::TypeBuilder::ArrayItemHelper<InspectorObject>::Traits::pushRefPtr):
(Inspector::TypeBuilder::ArrayItemHelper<InspectorArray>::Traits::pushRefPtr):
(Inspector::TypeBuilder::PrimitiveBindingTraits::assertValueHasExpectedType):
(Inspector::TypeBuilder::BindingTraits<TypeBuilder::Array<T>>::runtimeCast):
(Inspector::TypeBuilder::BindingTraits<TypeBuilder::Array<T>>::assertValueHasExpectedType):
(Inspector::TypeBuilder::BindingTraits<InspectorValue>::assertValueHasExpectedType):
(Inspector::TypeBuilder::BindingTraits<int>::assertValueHasExpectedType):
(Inspector::TypeBuilder::ExactlyInt::ExactlyInt): Deleted. It was not used.
(Inspector::TypeBuilder::ExactlyInt::operator int): Deleted.
(Inspector::TypeBuilder::ExactlyInt::cast_to_int): Deleted.
(Inspector::TypeBuilder::ExactlyInt::cast_to_int<int>): Deleted.
(Inspector::TypeBuilder::int>): Deleted.
(Inspector::TypeBuilder::RuntimeCastHelper::assertType): Deleted.
(Inspector::TypeBuilder::RuntimeCastHelper::assertAny): Deleted.
(Inspector::TypeBuilder::RuntimeCastHelper::assertInt): Deleted.
(Inspector::TypeBuilder::Array::runtimeCast): Deleted.
(Inspector::TypeBuilder::Array::assertCorrectValue): Deleted.
(Inspector::TypeBuilder::StructItemTraits::assertCorrectValue): Deleted.
(Inspector::TypeBuilder::ArrayItemHelper<String>::Traits::assertCorrectValue): Deleted.
(Inspector::TypeBuilder::ArrayItemHelper<int>::Traits::assertCorrectValue): Deleted.
(Inspector::TypeBuilder::ArrayItemHelper<double>::Traits::assertCorrectValue): Deleted.
(Inspector::TypeBuilder::ArrayItemHelper<bool>::Traits::assertCorrectValue): Deleted.
(Inspector::TypeBuilder::ArrayItemHelper<InspectorValue>::Traits::assertCorrectValue): Deleted.
(Inspector::TypeBuilder::ArrayItemHelper<InspectorObject>::Traits::assertCorrectValue): Deleted.
(Inspector::TypeBuilder::ArrayItemHelper<InspectorArray>::Traits::assertCorrectValue): Deleted.
(Inspector::TypeBuilder::ArrayItemHelper<TypeBuilder::Array<T>>::Traits::assertCorrectValue): Deleted.

* inspector/InspectorValues.cpp: Convert InspectorValue::Type to a scoped enum.
(Inspector::InspectorValue::writeJSON):
(Inspector::InspectorBasicValue::asBoolean):
(Inspector::InspectorBasicValue::asNumber):
(Inspector::InspectorBasicValue::writeJSON):
(Inspector::InspectorString::writeJSON):
(Inspector::InspectorObjectBase::InspectorObjectBase):
(Inspector::InspectorObjectBase::setArray): Take InspectorArrayBase.
(Inspector::InspectorObjectBase::setObject): Take InspectorObjectBase.
(Inspector::InspectorArrayBase::InspectorArrayBase):
* inspector/InspectorValues.h:

* inspector/agents/InspectorDebuggerAgent.cpp: Convert to scoped enums.
(Inspector::InspectorDebuggerAgent::schedulePauseOnNextStatement):
(Inspector::InspectorDebuggerAgent::breakProgram):
* inspector/agents/InspectorDebuggerAgent.h:
* inspector/agents/InspectorRuntimeAgent.cpp:
(Inspector::InspectorRuntimeAgent::parse):
* inspector/agents/InspectorRuntimeAgent.h:

* inspector/scripts/CodeGeneratorInspector.py: Removed.
* inspector/scripts/codegen/__init__.py: Added.
* inspector/scripts/codegen/generate_backend_commands.py: Added.
(BackendCommandsGenerator):
(BackendCommandsGenerator.__init__):
(BackendCommandsGenerator.model):
(BackendCommandsGenerator.output_filename):
(BackendCommandsGenerator.generate_license):
(BackendCommandsGenerator.generate_output):
(BackendCommandsGenerator.generate_domain):
(BackendCommandsGenerator.generate_domain.is_anonymous_enum_member):
(BackendCommandsGenerator.generate_domain.generate_parameter_object):
* inspector/scripts/codegen/generate_backend_dispatcher_header.py: Added.
(BackendDispatcherHeaderGenerator):
(BackendDispatcherHeaderGenerator.__init__):
(BackendDispatcherHeaderGenerator.model):
(BackendDispatcherHeaderGenerator.output_filename):
(BackendDispatcherHeaderGenerator.generate_license):
(BackendDispatcherHeaderGenerator.generate_output):
(BackendDispatcherHeaderGenerator.generate_output.for):
(BackendDispatcherHeaderGenerator._generate_handler_declarations_for_domain):
(BackendDispatcherHeaderGenerator._generate_anonymous_enum_for_parameter):
(BackendDispatcherHeaderGenerator._generate_handler_declaration_for_command):
(BackendDispatcherHeaderGenerator._generate_async_handler_declaration_for_command):
(BackendDispatcherHeaderGenerator._generate_dispatcher_declarations_for_domain):
(BackendDispatcherHeaderGenerator._generate_dispatcher_declaration_for_command):
* inspector/scripts/codegen/generate_backend_dispatcher_implementation.py: Added.
(BackendDispatcherImplementationGenerator):
(BackendDispatcherImplementationGenerator.__init__):
(BackendDispatcherImplementationGenerator.model):
(BackendDispatcherImplementationGenerator.output_filename):
(BackendDispatcherImplementationGenerator.generate_license):
(BackendDispatcherImplementationGenerator.generate_output):
(BackendDispatcherImplementationGenerator._generate_handler_class_destructor_for_domain):
(BackendDispatcherImplementationGenerator._generate_dispatcher_implementations_for_domain):
(BackendDispatcherImplementationGenerator._generate_small_dispatcher_switch_implementation_for_domain):
(BackendDispatcherImplementationGenerator._generate_large_dispatcher_switch_implementation_for_domain):
(BackendDispatcherImplementationGenerator._generate_async_dispatcher_class_for_domain):
(BackendDispatcherImplementationGenerator._generate_dispatcher_implementation_for_command):
* inspector/scripts/codegen/generate_frontend_dispatcher_header.py: Added.
(FrontendDispatcherHeaderGenerator):
(FrontendDispatcherHeaderGenerator.__init__):
(FrontendDispatcherHeaderGenerator.model):
(FrontendDispatcherHeaderGenerator.output_filename):
(FrontendDispatcherHeaderGenerator.generate_license):
(FrontendDispatcherHeaderGenerator.generate_output):
(FrontendDispatcherHeaderGenerator._generate_anonymous_enum_for_parameter):
(FrontendDispatcherHeaderGenerator._generate_dispatcher_declarations_for_domain):
(FrontendDispatcherHeaderGenerator._generate_dispatcher_declaration_for_event):
* inspector/scripts/codegen/generate_frontend_dispatcher_implementation.py: Added.
(FrontendDispatcherImplementationGenerator):
(FrontendDispatcherImplementationGenerator.__init__):
(FrontendDispatcherImplementationGenerator.model):
(FrontendDispatcherImplementationGenerator.output_filename):
(FrontendDispatcherImplementationGenerator.generate_license):
(FrontendDispatcherImplementationGenerator.generate_output):
(FrontendDispatcherImplementationGenerator._generate_dispatcher_implementations_for_domain):
(FrontendDispatcherImplementationGenerator._generate_dispatcher_implementation_for_event):
* inspector/scripts/codegen/generate_type_builder_header.py: Added.
(TypeBuilderHeaderGenerator):
(TypeBuilderHeaderGenerator.__init__):
(TypeBuilderHeaderGenerator.model):
(TypeBuilderHeaderGenerator.output_filename):
(TypeBuilderHeaderGenerator.generate_license):
(TypeBuilderHeaderGenerator.generate_output):
(TypeBuilderHeaderGenerator._generate_forward_declarations):
(_generate_typedefs):
(_generate_typedefs_for_domain):
(_generate_builders_for_domain):
(_generate_class_for_object_declaration):
(_generate_struct_for_enum_declaration):
(_generate_struct_for_anonymous_enum_member):
(_generate_struct_for_anonymous_enum_member.apply_indentation):
(_generate_struct_for_enum_type):
(_generate_builder_state_enum):
(_generate_builder_setter_for_member):
(_generate_unchecked_setter_for_member):
(_generate_forward_declarations_for_binding_traits):
* inspector/scripts/codegen/generate_type_builder_implementation.py: Added.
(TypeBuilderImplementationGenerator):
(TypeBuilderImplementationGenerator.__init__):
(TypeBuilderImplementationGenerator.model):
(TypeBuilderImplementationGenerator.output_filename):
(TypeBuilderImplementationGenerator.generate_license):
(TypeBuilderImplementationGenerator.generate_output):
(TypeBuilderImplementationGenerator._generate_enum_mapping):
(TypeBuilderImplementationGenerator._generate_open_field_names):
(TypeBuilderImplementationGenerator._generate_builders_for_domain):
(TypeBuilderImplementationGenerator._generate_runtime_cast_for_object_declaration):
(TypeBuilderImplementationGenerator._generate_assertion_for_object_declaration):
(TypeBuilderImplementationGenerator._generate_assertion_for_enum):
* inspector/scripts/codegen/generator.py: Added.
(ucfirst):
(Generator):
(Generator.__init__):
(Generator.model):
(Generator.generate_license):
(Generator.domains_to_generate):
(Generator.generate_output):
(Generator.output_filename):
(Generator.encoding_for_enum_value):
(Generator.assigned_enum_values):
(Generator.type_needs_runtime_casts):
(Generator.type_has_open_fields):
(Generator.type_needs_shape_assertions):
(Generator.calculate_types_requiring_shape_assertions):
(Generator.calculate_types_requiring_shape_assertions.gather_transitively_referenced_types):
(Generator._traverse_and_assign_enum_values):
(Generator._assign_encoding_for_enum_value):
(Generator.wrap_with_guard_for_domain):
(Generator.stylized_name_for_enum_value):
(Generator.stylized_name_for_enum_value.replaceCallback):
(Generator.keyed_get_method_for_type):
(Generator.keyed_set_method_for_type):
(Generator.type_builder_string_for_type):
(Generator.type_builder_string_for_type_member):
(Generator.type_string_for_unchecked_formal_in_parameter):
(Generator.type_string_for_checked_formal_event_parameter):
(Generator.type_string_for_type_member):
(Generator.type_string_for_type_with_name):
(Generator.type_string_for_formal_out_parameter):
(Generator.type_string_for_formal_async_parameter):
(Generator.type_string_for_stack_in_parameter):
(Generator.type_string_for_stack_out_parameter):
(Generator.assertion_method_for_type_member):
(Generator.assertion_method_for_type_member.assertion_method_for_type):
(Generator.cpp_name_for_primitive_type):
(Generator.js_name_for_parameter_type):
(Generator.should_use_wrapper_for_return_type):
(Generator.should_pass_by_copy_for_return_type):
* inspector/scripts/codegen/generator_templates.py: Added.
(GeneratorTemplates):
(void):
(HashMap):
(Builder):
(Inspector):
* inspector/scripts/codegen/models.py: Added.
(ucfirst):
(ParseException):
(TypecheckException):
(Framework):
(Framework.__init__):
(Framework.setting):
(Framework.fromString):
(Frameworks):
(TypeReference):
(TypeReference.__init__):
(TypeReference.referenced_name):
(Type):
(Type.__init__):
(Type.__eq__):
(Type.__hash__):
(Type.raw_name):
(Type.is_enum):
(Type.type_domain):
(Type.qualified_name):
(Type.resolve_type_references):
(PrimitiveType):
(PrimitiveType.__init__):
(PrimitiveType.__repr__):
(PrimitiveType.type_domain):
(PrimitiveType.qualified_name):
(AliasedType):
(AliasedType.__init__):
(AliasedType.__repr__):
(AliasedType.is_enum):
(AliasedType.type_domain):
(AliasedType.qualified_name):
(AliasedType.resolve_type_references):
(EnumType):
(EnumType.__init__):
(EnumType.__repr__):
(EnumType.is_enum):
(EnumType.type_domain):
(EnumType.enum_values):
(EnumType.qualified_name):
(EnumType.resolve_type_references):
(ArrayType):
(ArrayType.__init__):
(ArrayType.__repr__):
(ArrayType.type_domain):
(ArrayType.qualified_name):
(ArrayType.resolve_type_references):
(ObjectType):
(ObjectType.__init__):
(ObjectType.__repr__):
(ObjectType.type_domain):
(ObjectType.qualified_name):
(check_for_required_properties):
(Protocol):
(Protocol.__init__):
(Protocol.parse_specification):
(Protocol.parse_domain):
(Protocol.parse_type_declaration):
(Protocol.parse_type_member):
(Protocol.parse_command):
(Protocol.parse_event):
(Protocol.parse_call_or_return_parameter):
(Protocol.resolve_types):
(Protocol.lookup_type_for_declaration):
(Protocol.lookup_type_reference):
(Domain):
(Domain.__init__):
(Domain.resolve_type_references):
(Domains):
(TypeDeclaration):
(TypeDeclaration.__init__):
(TypeDeclaration.resolve_type_references):
(TypeMember):
(TypeMember.__init__):
(TypeMember.resolve_type_references):
(Parameter):
(Parameter.__init__):
(Parameter.resolve_type_references):
(Command):
(Command.__init__):
(Command.resolve_type_references):
(Event):
(Event.__init__):
(Event.resolve_type_references):
* inspector/scripts/generate-inspector-protocol-bindings.py: Added.
(IncrementalFileWriter):
(IncrementalFileWriter.__init__):
(IncrementalFileWriter.write):
(IncrementalFileWriter.close):
(generate_from_specification):
(generate_from_specification.load_specification):
* inspector/scripts/tests/commands-with-async-attribute.json: Added.
* inspector/scripts/tests/commands-with-optional-call-return-parameters.json: Added.
* inspector/scripts/tests/domains-with-varying-command-sizes.json: Added.
* inspector/scripts/tests/events-with-optional-parameters.json: Added.
* inspector/scripts/tests/expected/commands-with-async-attribute.json-result: Added.
* inspector/scripts/tests/expected/commands-with-optional-call-return-parameters.json-result: Added.
* inspector/scripts/tests/expected/domains-with-varying-command-sizes.json-result: Added.
* inspector/scripts/tests/expected/events-with-optional-parameters.json-result: Added.
* inspector/scripts/tests/fail-on-duplicate-type-declarations.json-error: Added.
* inspector/scripts/tests/fail-on-enum-with-no-values.json-error: Added.
* inspector/scripts/tests/fail-on-type-declaration-using-type-reference.json-error: Added.
* inspector/scripts/tests/fail-on-type-with-lowercase-name.json-error: Added.
* inspector/scripts/tests/fail-on-unknown-type-reference-in-type-declaration.json-error: Added.
* inspector/scripts/tests/fail-on-unknown-type-reference-in-type-member.json-error: Added.
* inspector/scripts/tests/expected/same-type-id-different-domain.json-result: Added.
* inspector/scripts/tests/expected/type-declaration-aliased-primitive-type.json-result: Added.
* inspector/scripts/tests/expected/type-declaration-array-type.json-result: Added.
* inspector/scripts/tests/expected/type-declaration-enum-type.json-result: Added.
* inspector/scripts/tests/expected/type-declaration-object-type.json-result: Added.
* inspector/scripts/tests/expected/type-requiring-runtime-casts.json-result: Added.
* inspector/scripts/tests/fail-on-duplicate-type-declarations.json: Added.
* inspector/scripts/tests/fail-on-enum-with-no-values.json: Added.
* inspector/scripts/tests/fail-on-type-declaration-using-type-reference.json: Added.
* inspector/scripts/tests/fail-on-type-with-lowercase-name.json: Added.
* inspector/scripts/tests/fail-on-unknown-type-reference-in-type-declaration.json: Added.
* inspector/scripts/tests/fail-on-unknown-type-reference-in-type-member.json: Added.
* inspector/scripts/tests/same-type-id-different-domain.json: Added.
* inspector/scripts/tests/type-declaration-aliased-primitive-type.json: Added.
* inspector/scripts/tests/type-declaration-array-type.json: Added.
* inspector/scripts/tests/type-declaration-enum-type.json: Added.
* inspector/scripts/tests/type-declaration-object-type.json: Added.
* inspector/scripts/tests/type-requiring-runtime-casts.json: Added.

Source/WebCore:

Aside from hooking up the new inspector bindings generator to the build system,
there are a few comingled changes that would be painful to split from the main
patch. Together, these changes make it possible to forward-declare all protocol
enumerations as scoped enums, reducing weird ordering dependencies between domains.

All runtimeCast() methods are now static methods of BindingTraits specializations,
rather than as static methods of the type builder classes.

All protocol enumeration types are scoped enums rather than struct-namespaced enums.

Remove some implicit protocol type and enumeration definitions because they trigger
compiler bugs or introduce unnecessary complexity in the bindings generator.

No new tests, no behavior changed.

* CMakeLists.txt:
* DerivedSources.make: Update script dependencies and the bindings generator script name.

* inspector/CommandLineAPIHost.cpp:
(WebCore::CommandLineAPIHost::inspectImpl): Convert to BindingTraits.

* inspector/InspectorCSSAgent.cpp:
(WebCore::InspectorCSSAgent::detectOrigin):
(WebCore::InspectorCSSAgent::buildArrayForRegions):
* inspector/InspectorCSSAgent.h:
* inspector/InspectorDOMAgent.cpp:
(WebCore::InspectorDOMAgent::buildObjectForAccessibilityProperties):
* inspector/InspectorPageAgent.cpp:
(WebCore::InspectorPageAgent::resourceTypeJson):
(WebCore::InspectorPageAgent::cachedResourceTypeJson):
(WebCore::InspectorPageAgent::getScriptExecutionStatus): Convert to scoped enums.

(WebCore::InspectorPageAgent::buildObjectForFrameTree): Extract FrameResourceTree::Resources
into its own explicitly-defined object type. The new generator does not support implicitly
defining object types as elements of arrays.
* inspector/InspectorPageAgent.h:
* inspector/InspectorReplayAgent.cpp: Disambiguate between internal and protocol types.
(WebCore::buildInspectorObjectForSessionState):
(WebCore::buildInspectorObjectForSegmentState):
(WebCore::InspectorReplayAgent::sessionState):
(WebCore::InspectorReplayAgent::didCreateFrontendAndBackend):
(WebCore::InspectorReplayAgent::frameNavigated):
(WebCore::InspectorReplayAgent::frameDetached):
(WebCore::InspectorReplayAgent::willDispatchEvent):
(WebCore::InspectorReplayAgent::startCapturing):
(WebCore::InspectorReplayAgent::stopCapturing):
(WebCore::InspectorReplayAgent::replayToPosition):
(WebCore::InspectorReplayAgent::replayToCompletion):
(WebCore::InspectorReplayAgent::pausePlayback):
(WebCore::InspectorReplayAgent::cancelPlayback):
(WebCore::InspectorReplayAgent::switchSession):
(WebCore::InspectorReplayAgent::insertSessionSegment):
(WebCore::InspectorReplayAgent::removeSessionSegment):
(WebCore::InspectorReplayAgent::currentReplayState):
(WebCore::InspectorReplayAgent::getAvailableSessions):
(WebCore::InspectorReplayAgent::getSessionData):
(WebCore::InspectorReplayAgent::getSegmentData):
* inspector/InspectorReplayAgent.h:
* inspector/InspectorResourceAgent.cpp:
(WebCore::InspectorResourceAgent::willSendRequest):
* inspector/InspectorStyleSheet.cpp: Convert to scoped enums. Extract CSSProperty::Status
because scoped enums defined in a class scope trigger compiler bugs in GCC 4.8.
(WebCore::buildMediaObject):
(WebCore::InspectorStyle::styleWithProperties): Work around compiler disagreements.
(WebCore::InspectorStyleSheet::create):
(WebCore::InspectorStyleSheet::InspectorStyleSheet):
(WebCore::InspectorStyleSheetForInlineStyle::create):
(WebCore::InspectorStyleSheetForInlineStyle::InspectorStyleSheetForInlineStyle):
* inspector/InspectorStyleSheet.h:

* inspector/InspectorTimelineAgent.cpp: Convert to scoped enums and BindingTraits.
(WebCore::toProtocol):
(WebCore::InspectorTimelineAgent::addRecordToTimeline):
(WebCore::InspectorTimelineAgent::sendEvent):

* inspector/protocol/CSS.json: Extract CSSProperty::Status.
* inspector/protocol/Page.json: Extract FrameResourceTree::Resources.

Source/WebInspectorUI:

Also apply type extraction refactorings to old protocol versions for consistency.

* Versions/Inspector-iOS-6.0.json:
* Versions/Inspector-iOS-7.0.json:

Tools:

* Scripts/run-inspector-generator-tests: Added.
(main):
* Scripts/webkitpy/inspector/__init__.py: Added.
* Scripts/webkitpy/inspector/main.py: Added.
(InspectorGeneratorTests):
(InspectorGeneratorTests.__init__):
(InspectorGeneratorTests.generate_from_json):
(InspectorGeneratorTests.write_error_file):
(InspectorGeneratorTests.detect_changes):
(InspectorGeneratorTests.run_tests):
(InspectorGeneratorTests.main):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172655 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoFix plugin visibility check.
barraclough@apple.com [Fri, 15 Aug 2014 21:22:13 +0000 (21:22 +0000)]
Fix plugin visibility check.
https://bugs.webkit.org/show_bug.cgi?id=135991

D'oh, more unreviewed typo fix. :-(

* PluginProcess/PluginControllerProxy.cpp:
(WebKit::PluginControllerProxy::visibilityDidChange):
(WebKit::PluginControllerProxy::windowVisibilityChanged):
(WebKit::PluginControllerProxy::updateVisibilityActivity):
(WebKit::PluginControllerProxy::updateVisiblityActivity): Deleted.
* PluginProcess/PluginControllerProxy.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172653 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[EME][Mac] Adopt AVStreamSession.
jer.noble@apple.com [Fri, 15 Aug 2014 21:14:14 +0000 (21:14 +0000)]
[EME][Mac] Adopt AVStreamSession.
https://bugs.webkit.org/show_bug.cgi?id=135983

Reviewed by Eric Carlson.

Add AVStreamDataParsers to a new AVStreamSession instance, and expire that session when asked to release
the session's keys.

* platform/graphics/avfoundation/objc/CDMSessionMediaSourceAVFObjC.h:
* platform/graphics/avfoundation/objc/CDMSessionMediaSourceAVFObjC.mm:
(WebCore::CDMSessionMediaSourceAVFObjC::~CDMSessionMediaSourceAVFObjC):
(WebCore::CDMSessionMediaSourceAVFObjC::releaseKeys):
(WebCore::CDMSessionMediaSourceAVFObjC::update):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172651 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Probe details sidebar data grids do not show data
timothy@apple.com [Fri, 15 Aug 2014 21:12:06 +0000 (21:12 +0000)]
Web Inspector: Probe details sidebar data grids do not show data
https://bugs.webkit.org/show_bug.cgi?id=135992

Reviewed by Joseph Pecoraro.

* UserInterface/Views/ProbeSetDataGrid.js:
(WebInspector.ProbeSetDataGrid): Make the DataGrid be inline to show the data correctly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172650 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoImplement snapping behavior for iOS
commit-queue@webkit.org [Fri, 15 Aug 2014 21:09:13 +0000 (21:09 +0000)]
Implement snapping behavior for iOS
https://bugs.webkit.org/show_bug.cgi?id=135769

Patch by Wenson Hsieh <wenson_hsieh@apple.com> on 2014-08-15
Reviewed by Brent Fulgham.

Source/WebCore:

Added support for snap points on iOS. This entails sending snap points from the web process to the UI process
by encoding and decoding snap points and updating ScrollingTreeOverflowScrollingNodeIOS and WKWebView to support
animation to snap points in the cases of mainframe and overflow scrolling, respectively.

There are no new tests yet -- we'll need to find a way to test this!

* WebCore.exp.in:
* page/scrolling/AsyncScrollingCoordinator.cpp:
(WebCore::setStateScrollingNodeSnapOffsetsAsFloat): Helper functions used to convert LayoutUnits to floats for snap offsets.
(WebCore::AsyncScrollingCoordinator::frameViewLayoutUpdated): Update snap offsets for mainframe scrolling, as well as the state node.
(WebCore::AsyncScrollingCoordinator::updateOverflowScrollingNode): Added snap offsets.
* page/scrolling/AxisScrollSnapOffsets.h:
(WebCore::closestSnapOffset): Computes the closest snap offset given a sorted vector of comparable types (e.g. float, LayoutUnit).
* page/scrolling/ScrollingCoordinator.h:
* page/scrolling/ScrollingStateScrollingNode.cpp:
(WebCore::ScrollingStateScrollingNode::ScrollingStateScrollingNode):
(WebCore::ScrollingStateScrollingNode::setHorizontalSnapOffsets):
(WebCore::ScrollingStateScrollingNode::setVerticalSnapOffsets):
* page/scrolling/ScrollingStateScrollingNode.h:
(WebCore::ScrollingStateScrollingNode::horizontalSnapOffsets):
(WebCore::ScrollingStateScrollingNode::verticalSnapOffsets):
* page/scrolling/ScrollingTreeScrollingNode.cpp:
(WebCore::ScrollingTreeScrollingNode::updateBeforeChildren):
* page/scrolling/ScrollingTreeScrollingNode.h:
(WebCore::ScrollingTreeScrollingNode::horizontalSnapOffsets):
(WebCore::ScrollingTreeScrollingNode::verticalSnapOffsets):
* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::updateScrollInfoAfterLayout): Update snap offsets for overflow scrolling.
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::updateScrollCoordinatedLayer): Updates the state node's snap offsets for overflow scrolling.

Source/WebKit2:

Added support for snap points on iOS.

* Shared/Scrolling/RemoteScrollingCoordinatorTransaction.cpp:
(ArgumentCoder<ScrollingStateScrollingNode>::encode): Added snap offset encoding.
(ArgumentCoder<ScrollingStateScrollingNode>::decode): Added snap offset decoding.
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView scrollViewWillEndDragging:withVelocity:targetContentOffset:]): For mainframe scrolling, retargets the scroll destination to the appropriate snap point.
* UIProcess/Scrolling/RemoteScrollingCoordinatorProxy.h: Added methods to extract relevant information for WKWebView from snap offsets in mainframe scrolling.
* UIProcess/Scrolling/ios/ScrollingTreeOverflowScrollingNodeIOS.mm:
(-[WKOverflowScrollViewDelegate scrollViewWillEndDragging:withVelocity:targetContentOffset:]): For overflow scrolling, retargets the scroll destination to the appropriate snap point.
(WebKit::ScrollingTreeOverflowScrollingNodeIOS::updateAfterChildren): Sets the deceleration factor to FAST if snap points are active.
* UIProcess/ios/RemoteScrollingCoordinatorProxyIOS.mm:
(WebKit::RemoteScrollingCoordinatorProxy::adjustTargetContentOffsetForSnapping): Used by WKWebView for mainframe scrolling to retarget the scroll destination, if necessary.
(WebKit::RemoteScrollingCoordinatorProxy::shouldSetScrollViewDecelerationRateFast): Used by WKWebView for mainframe scrolling to determine whether to set deceleration rate to the fast value.
(WebKit::RemoteScrollingCoordinatorProxy::shouldSnapForMainFrameScrolling): Helper method.
(WebKit::RemoteScrollingCoordinatorProxy::closestSnapOffsetForMainFrameScrolling): Helper method.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172649 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoMade native inlining errors not segfault.
mmirman@apple.com [Fri, 15 Aug 2014 21:04:31 +0000 (21:04 +0000)]
Made native inlining errors not segfault.
https://bugs.webkit.org/show_bug.cgi?id=135988

Reviewed by Geoffrey Garen.

* ftl/FTLAbbreviations.h:
(JSC::FTL::disposeMessage): Added.
* ftl/FTLLowerDFGToLLVM.cpp:
(JSC::FTL::LowerDFGToLLVM::compilePutById):
abstracted out Options::verboseCompilation as was the case in the rest of the file.
(JSC::FTL::LowerDFGToLLVM::compileNativeCallOrConstruct):
(JSC::FTL::LowerDFGToLLVM::getModuleByPathForSymbol):
added output error messages for llvm module loading.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172648 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoTurn r/rx/ry to presentation attributes
krit@webkit.org [Fri, 15 Aug 2014 20:48:35 +0000 (20:48 +0000)]
Turn r/rx/ry to presentation attributes
https://bugs.webkit.org/show_bug.cgi?id=135978

Source/WebCore:

Unreviewed typo.

Patch by Dirk Schulze <krit@webkit.org> on 2014-08-15

* rendering/svg/RenderSVGEllipse.cpp:
(WebCore::RenderSVGEllipse::calculateRadiiAndCenter):

LayoutTests:

Patch by Dirk Schulze <krit@webkit.org> on 2014-08-15
Reviewed by Dean Jackson.

Rebaseline test.

* svg/css/parse-length-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172647 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoFix plugin visibility check.
barraclough@apple.com [Fri, 15 Aug 2014 20:37:15 +0000 (20:37 +0000)]
Fix plugin visibility check.
https://bugs.webkit.org/show_bug.cgi?id=135991

Unreviewed typo fix.

* PluginProcess/PluginControllerProxy.cpp:
(WebKit::PluginControllerProxy::PluginControllerProxy):
(WebKit::PluginControllerProxy::visibilityDidChange):
(WebKit::PluginControllerProxy::windowVisibilityChanged):
(WebKit::PluginControllerProxy::updateVisiblityActivity):
* PluginProcess/PluginControllerProxy.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172646 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoFix plugin visibility check.
barraclough@apple.com [Fri, 15 Aug 2014 20:32:14 +0000 (20:32 +0000)]
Fix plugin visibility check.
https://bugs.webkit.org/show_bug.cgi?id=135991

Reviewed by Andreas Kling.

We should only take a UserActivity if the plugin is in a visible window,
and also is itself visible.

* PluginProcess/PluginControllerProxy.cpp:
(WebKit::PluginControllerProxy::PluginControllerProxy):
    - initialize member state.
(WebKit::PluginControllerProxy::visibilityDidChange):
(WebKit::PluginControllerProxy::windowVisibilityChanged):
    - update member state & call updateVisiblityActivity.
(WebKit::PluginControllerProxy::updateVisiblityActivity):
    - enable the UserActivity only if both visibilities are true.
* PluginProcess/PluginControllerProxy.h:
    - added m_isVisible, m_isWindowVisible, updateVisibilityActivity.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172645 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[Services with UI] Selections are incorrect when selecting three lines.
enrica@apple.com [Fri, 15 Aug 2014 19:53:19 +0000 (19:53 +0000)]
[Services with UI] Selections are incorrect when selecting three lines.
https://bugs.webkit.org/show_bug.cgi?id=135989
<rdar://problem/18032571>

Reviewed by Tim Horton.

The stitching algorithm did not handle correctly the case of selections
over three lines if the middle line is composed of only one rectangle.

* WebProcess/WebPage/mac/ServicesOverlayController.mm:
(WebKit::stitchRects):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172643 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoTurn r/rx/ry to presentation attributes
krit@webkit.org [Fri, 15 Aug 2014 19:42:38 +0000 (19:42 +0000)]
Turn r/rx/ry to presentation attributes
https://bugs.webkit.org/show_bug.cgi?id=135978

Patch by Dirk Schulze <krit@webkit.org> on 2014-08-15
Reviewed by Dean Jackson.

This follows the patch for width and height presentation attributes and
Source/WebCore:

turns rx, ry and r to presentation attributes as well:

​http://trac.webkit.org/changeset/171341

Extended existing tests.

* css/CSSComputedStyleDeclaration.cpp:
(WebCore::ComputedStyleExtractor::propertyValue):
* css/CSSParser.cpp:
(WebCore::isSimpleLengthPropertyID):
* css/DeprecatedStyleBuilder.cpp:
(WebCore::DeprecatedStyleBuilder::DeprecatedStyleBuilder):
* css/SVGCSSParser.cpp:
(WebCore::CSSParser::parseSVGValue):
* css/SVGCSSPropertyNames.in:
* page/animation/CSSPropertyAnimation.cpp:
(WebCore::CSSPropertyAnimationWrapperMap::CSSPropertyAnimationWrapperMap):
* rendering/style/RenderStyle.h:
* rendering/style/SVGRenderStyle.h:
(WebCore::SVGRenderStyle::setR):
(WebCore::SVGRenderStyle::setRx):
(WebCore::SVGRenderStyle::setRy):
(WebCore::SVGRenderStyle::r):
(WebCore::SVGRenderStyle::rx):
(WebCore::SVGRenderStyle::ry):
* rendering/style/SVGRenderStyleDefs.cpp:
(WebCore::StyleLayoutData::StyleLayoutData):
(WebCore::StyleLayoutData::operator==):
* rendering/style/SVGRenderStyleDefs.h:
* rendering/svg/RenderSVGEllipse.cpp:
(WebCore::RenderSVGEllipse::calculateRadiiAndCenter):
* rendering/svg/SVGPathData.cpp:
(WebCore::updatePathFromCircleElement):
(WebCore::updatePathFromEllipseElement):
(WebCore::updatePathFromRectElement):
(WebCore::updatePathFromLineElement): Deleted.
* svg/SVGCircleElement.cpp:
(WebCore::SVGCircleElement::svgAttributeChanged):
* svg/SVGElement.cpp:
(WebCore::populateAttributeNameToCSSPropertyIDMap):
(WebCore::populateCSSPropertyWithSVGDOMNameToAnimatedPropertyTypeMap):
* svg/SVGEllipseElement.cpp:
(WebCore::SVGEllipseElement::svgAttributeChanged):
* svg/SVGRectElement.cpp:
(WebCore::SVGRectElement::svgAttributeChanged):

LayoutTests:

turns r, rx and ry to presentation attributes as well:

​http://trac.webkit.org/changeset/171341

* svg/css/parse-length.html:
* transitions/svg-layout-transition-expected.txt:
* transitions/svg-layout-transition.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172642 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoTurn cx/cy to presentation attributes
krit@webkit.org [Fri, 15 Aug 2014 19:38:10 +0000 (19:38 +0000)]
Turn cx/cy to presentation attributes
https://bugs.webkit.org/show_bug.cgi?id=135975

Reviewed by Dean Jackson.

This follows the patch for width and height presentation attributes and
Source/WebCore:

turns x and y to presentation attributes as well:

​http://trac.webkit.org/changeset/171341

Extended existing tests.

* css/CSSComputedStyleDeclaration.cpp:
(WebCore::ComputedStyleExtractor::propertyValue):
* css/CSSParser.cpp:
(WebCore::isSimpleLengthPropertyID):
* css/DeprecatedStyleBuilder.cpp:
(WebCore::DeprecatedStyleBuilder::DeprecatedStyleBuilder):
* css/SVGCSSParser.cpp:
(WebCore::CSSParser::parseSVGValue):
* css/SVGCSSPropertyNames.in:
* page/animation/CSSPropertyAnimation.cpp:
(WebCore::CSSPropertyAnimationWrapperMap::CSSPropertyAnimationWrapperMap):
* rendering/style/RenderStyle.h:
* rendering/style/SVGRenderStyle.h:
(WebCore::SVGRenderStyle::setCx):
(WebCore::SVGRenderStyle::setCy):
(WebCore::SVGRenderStyle::cx):
(WebCore::SVGRenderStyle::cy):
* rendering/style/SVGRenderStyleDefs.cpp:
(WebCore::StyleLayoutData::StyleLayoutData):
(WebCore::StyleLayoutData::operator==):
* rendering/style/SVGRenderStyleDefs.h:
* rendering/svg/RenderSVGEllipse.cpp:
(WebCore::RenderSVGEllipse::calculateRadiiAndCenter):
* rendering/svg/SVGPathData.cpp:
(WebCore::updatePathFromCircleElement):
(WebCore::updatePathFromEllipseElement):
* svg/SVGCircleElement.cpp:
(WebCore::SVGCircleElement::svgAttributeChanged):
(WebCore::SVGCircleElement::selfHasRelativeLengths): Deleted.
* svg/SVGCircleElement.h:
* svg/SVGElement.cpp:
(WebCore::populateAttributeNameToCSSPropertyIDMap):
(WebCore::populateCSSPropertyWithSVGDOMNameToAnimatedPropertyTypeMap):
* svg/SVGEllipseElement.cpp:
(WebCore::SVGEllipseElement::svgAttributeChanged):
(WebCore::SVGEllipseElement::selfHasRelativeLengths): Deleted.
* svg/SVGEllipseElement.h:

LayoutTests:

turns cx and cy to presentation attributes as well:

​http://trac.webkit.org/changeset/171341

* svg/css/parse-length-expected.txt:
* svg/css/parse-length.html:
* transitions/svg-layout-transition-expected.txt:
* transitions/svg-layout-transition.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172641 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Legacy Mac toolbar appearance looks wrong
timothy@apple.com [Fri, 15 Aug 2014 19:11:55 +0000 (19:11 +0000)]
Web Inspector: Legacy Mac toolbar appearance looks wrong
https://bugs.webkit.org/show_bug.cgi?id=135962

Reviewed by Joseph Pecoraro.

* UserInterface/Views/Toolbar.css:
(body.mac-platform.legacy .toolbar): Don't draw a gradient background when not docked.
Move border-bottom-color from body.mac-platform.legacy.docked .toolbar to here so it
applies all the time as it should have.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172640 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoService overlays stay fixed when <iframe> scrolls
timothy_horton@apple.com [Fri, 15 Aug 2014 19:10:04 +0000 (19:10 +0000)]
Service overlays stay fixed when <iframe> scrolls
https://bugs.webkit.org/show_bug.cgi?id=135959
<rdar://problem/17957716>

Reviewed by Enrica Casucci.

* WebProcess/WebPage/mac/ServicesOverlayController.mm:
(WebKit::ServicesOverlayController::highlightsAreEquivalent):
Tiny correctness fix; highlights can only be equivalent if their
types are equivalent as well.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172639 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agorun-webkit-tests should skip 'support' directories
commit-queue@webkit.org [Fri, 15 Aug 2014 19:03:59 +0000 (19:03 +0000)]
run-webkit-tests should skip 'support' directories
https://bugs.webkit.org/show_bug.cgi?id=135910

Patch by Rebecca Hauck <rhauck@adobe.com> on 2014-08-15
Reviewed by Bem Jones-Bey.

This patch adds the 'support' directory to the list
of dirs to skip when looking for runnable tests.
the W3C uses this name for support files, so tests
imported from the W3C will not work without this change
or without many edits.

* Scripts/webkitpy/port/base.py:
(Port._real_tests):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172638 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agode-PLATFORM(COCOA) a couple of methods in plugin code
barraclough@apple.com [Fri, 15 Aug 2014 18:56:52 +0000 (18:56 +0000)]
de-PLATFORM(COCOA) a couple of methods in plugin code
https://bugs.webkit.org/show_bug.cgi?id=135987

Reviewed by Sam Weinig.

Pass through the visibility & focus on all platforms, to simplify & avoid
any weird inconsistencies between platforms.

* PluginProcess/PluginControllerProxy.cpp:
(WebKit::PluginControllerProxy::windowFocusChanged):
(WebKit::PluginControllerProxy::windowVisibilityChanged):
* PluginProcess/PluginControllerProxy.h:
* PluginProcess/PluginControllerProxy.messages.in:
* PluginProcess/mac/PluginControllerProxyMac.mm:
(WebKit::PluginControllerProxy::windowFocusChanged): Deleted.
(WebKit::PluginControllerProxy::windowVisibilityChanged): Deleted.
* WebProcess/Plugins/Netscape/NetscapePlugin.cpp:
(WebKit::NetscapePlugin::windowFocusChanged):
(WebKit::NetscapePlugin::windowVisibilityChanged):
* WebProcess/Plugins/Netscape/NetscapePlugin.h:
* WebProcess/Plugins/Plugin.h:
* WebProcess/Plugins/PluginProxy.cpp:
(WebKit::PluginProxy::windowAndViewFramesChanged):
* WebProcess/Plugins/PluginProxy.h:
* WebProcess/Plugins/PluginView.cpp:
(WebKit::PluginView::viewStateDidChange):
(WebKit::PluginView::platformViewStateDidChange): Deleted.
* WebProcess/Plugins/PluginView.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172637 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoService overlays stay fixed when <iframe> scrolls
timothy_horton@apple.com [Fri, 15 Aug 2014 18:47:51 +0000 (18:47 +0000)]
Service overlays stay fixed when <iframe> scrolls
https://bugs.webkit.org/show_bug.cgi?id=135959
<rdar://problem/17957716>

Reviewed by Enrica Casucci.

* WebProcess/WebPage/PageOverlay.cpp:
(WebKit::PageOverlay::didScrollFrame):
* WebProcess/WebPage/PageOverlay.h:
(WebKit::PageOverlay::Client::didScrollFrame):
* WebProcess/WebPage/PageOverlayController.cpp:
(WebKit::PageOverlayController::didScrollFrame):
Push didScrollFrame down to the overlays.

* WebProcess/WebPage/ServicesOverlayController.h:
* WebProcess/WebPage/mac/ServicesOverlayController.mm:
(WebKit::ServicesOverlayController::Highlight::createForSelection):
Hold on to the selection's Range so we can use it to compare Highlights later.

(WebKit::ServicesOverlayController::Highlight::Highlight):
(WebKit::ServicesOverlayController::Highlight::setDDHighlight):
Factor the code to set up and paint the highlight out, so that we can
set a new DDHighlightRef on a Highlight and the layer moves/reshapes/repaints.

(WebKit::ServicesOverlayController::buildPhoneNumberHighlights):
(WebKit::ServicesOverlayController::buildSelectionHighlight):
(WebKit::ServicesOverlayController::replaceHighlightsOfTypePreservingEquivalentHighlights):
Factor replaceHighlightsOfTypePreservingEquivalentHighlights out
so that we can use it for buildSelectionHighlight as well.

Steal the DDHighlightRef from the new Highlight when re-using an old one
so that the newly computed rects are used instead of the old ones.

(WebKit::ServicesOverlayController::highlightsAreEquivalent):
We will always have a Range now, so we can always check equivalence using it.

(WebKit::ServicesOverlayController::didScrollFrame):
Rebuild all highlights upon subframe scroll, as they might have moved.
We could optimize this in the future, but for now it's cheap enough
and rare enough that it doesn't matter.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172636 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoREGRESSION (WebKit2 Gestures): White flash when swiping back to cnn.com's homepage...
timothy_horton@apple.com [Fri, 15 Aug 2014 18:42:24 +0000 (18:42 +0000)]
REGRESSION (WebKit2 Gestures): White flash when swiping back to cnn.com's homepage from an article
https://bugs.webkit.org/show_bug.cgi?id=135951
<rdar://problem/18006149>

Reviewed by Simon Fraser.

Wait for (the first visually non-empty layout AND the render tree size threshold to be hit),
OR didFinishLoadForFrame, whichever comes first. Once we've done the first visually non-empty layout,
we'll start the watchdog and tear down the snapshot in three seconds no matter what.
Also, force a repaint so we can asynchronously wait for the Web Process to paint and return to us
before removing the snapshot, which improves our chances that something is actually on the screen.

* UIProcess/API/mac/WKView.mm:
(-[WKView _didFirstVisuallyNonEmptyLayoutForMainFrame]):
(-[WKView _didFinishLoadForMainFrame]):
(-[WKView _removeNavigationGestureSnapshot]):
* UIProcess/API/mac/WKViewInternal.h:
* UIProcess/PageClient.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::didFinishLoadForFrame):
(WebKit::WebPageProxy::didFirstVisuallyNonEmptyLayoutForFrame):
(WebKit::WebPageProxy::removeNavigationGestureSnapshot):
* UIProcess/WebPageProxy.h:
* UIProcess/ios/PageClientImplIOS.h:
* UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::didFirstVisuallyNonEmptyLayoutForMainFrame):
(WebKit::PageClientImpl::didFinishLoadForMainFrame):
* UIProcess/mac/PageClientImpl.h:
* UIProcess/mac/PageClientImpl.mm:
(WebKit::PageClientImpl::didFirstVisuallyNonEmptyLayoutForMainFrame):
(WebKit::PageClientImpl::didFinishLoadForMainFrame):
(WebKit::PageClientImpl::removeNavigationGestureSnapshot):
Plumb didFirstVisuallyNonEmptyLayoutForMainFrame and didFinishLoadForMainFrame
through to ViewGestureController from WebPageProxy via the PageClient, etc.

Ditto for removeNavigationGestureSnapshot, though it is called from a
VoidCallback in ViewGestureController instead of from WebFrameLoaderClient and friends.

* UIProcess/mac/ViewGestureController.h:
* UIProcess/mac/ViewGestureControllerMac.mm:
(WebKit::ViewGestureController::ViewGestureController):
(WebKit::ViewGestureController::endSwipeGesture):
When finishing a swipe, we want to wait for both the first visually non-empty layout
and the render tree size threshold being hit.

(WebKit::ViewGestureController::didHitRenderTreeSizeThreshold):
(WebKit::ViewGestureController::didFirstVisuallyNonEmptyLayoutForMainFrame):
When both of these things have happened, remove the swipe snapshot (after forcing a repaint).
For didFirstVisuallyNonEmptyLayoutForMainFrame, we will also start a watchdog
ensuring that we remove the snapshot in three seconds.

(WebKit::ViewGestureController::didFinishLoadForMainFrame):
When didFinishLoadForMainFrame happens, remove the swipe snapshot (after forcing a repaint).

(WebKit::ViewGestureController::swipeSnapshotWatchdogTimerFired):
If the watchdog timer fires, remove the swipe snapshot (after forcing a repaint).

(WebKit::ViewGestureController::removeSwipeSnapshotAfterRepaint):
Force a repaint and wait for the async callback before removing the snapshot.
It is safe to hold on to the WebPageProxy here because it will always
call all of its callbacks before it is destroyed.
Avoid enqueuing multiple force-repaints.

(WebKit::ViewGestureController::removeSwipeSnapshot):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172635 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoFix plugin visibility initialization
barraclough@apple.com [Fri, 15 Aug 2014 18:26:34 +0000 (18:26 +0000)]
Fix plugin visibility initialization
https://bugs.webkit.org/show_bug.cgi?id=135985

Reviewed by Geoffrey Garen.

* WebProcess/Plugins/PluginView.cpp:
(WebKit::PluginView::didInitializePlugin):
    - We're not currently initializing the plugin visibility state,
      and we're passing the wrong value for window visibility.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172634 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[Cocoa] Add migrate-headers.sh and postprocess-headers.sh to WebKit.xcodeproj
aestes@apple.com [Fri, 15 Aug 2014 17:38:47 +0000 (17:38 +0000)]
[Cocoa] Add migrate-headers.sh and postprocess-headers.sh to WebKit.xcodeproj
https://bugs.webkit.org/show_bug.cgi?id=135984

Reviewed by Daniel Bates.

It's nice to be able to look at these files in Xcode.

* WebKit.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172631 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoSimplify visibility activity accounting for plugins
barraclough@apple.com [Fri, 15 Aug 2014 17:33:22 +0000 (17:33 +0000)]
Simplify visibility activity accounting for plugins
https://bugs.webkit.org/show_bug.cgi?id=135981

Reviewed by Dan Bernstein.

Just take one per connection, rather than implementing a counting mechanism.

* PluginProcess/PluginControllerProxy.cpp:
(WebKit::PluginControllerProxy::PluginControllerProxy):
* PluginProcess/PluginControllerProxy.h:
    - added m_visiblityActivity.
* PluginProcess/PluginProcess.cpp:
(WebKit::PluginProcess::PluginProcess):
(WebKit::PluginProcess::pluginsForWebProcessDidBecomeVisible): Deleted.
(WebKit::PluginProcess::pluginsForWebProcessDidBecomeHidden): Deleted.
* PluginProcess/PluginProcess.h:
    - removed pluginsForWebProcessDidBecomeVisible/Hidden.
* PluginProcess/WebProcessConnection.cpp:
(WebKit::WebProcessConnection::removePluginControllerProxy):
(WebKit::WebProcessConnection::pluginDidBecomeVisible): Deleted.
(WebKit::WebProcessConnection::pluginDidBecomeHidden): Deleted.
* PluginProcess/WebProcessConnection.h:
    - removed pluginDidBecomeVisible/Hidden.
* PluginProcess/mac/PluginControllerProxyMac.mm:
(WebKit::PluginControllerProxy::windowVisibilityChanged):
    - windowVisibilityChanged uses m_visiblityActivity, rather than calling to PluginControllerProxy.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172629 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago<https://webkit.org/b/135979> WebKit1 Plug-in test failures in clean builds after...
mrowe@apple.com [Fri, 15 Aug 2014 16:49:37 +0000 (16:49 +0000)]
<https://webkit.org/b/135979> WebKit1 Plug-in test failures in clean builds after r172595

Reviewed by Dan Bernstein.

* WebKit.xcodeproj/project.pbxproj: Add a WebKitPluginHost.app symlink at the top level of
WebKitLegacy.framework that points into Versions/Current. This enables -[NSBundle pathForAuxiliaryExecutable:]
to work correctly. Adding this symlink requires deleting any content that may already exist
at that path, since prior to r172595 it may have contained either a symlink or an
application bundle.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172627 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoFix typo in Path comment
cavalcantii@gmail.com [Fri, 15 Aug 2014 16:38:35 +0000 (16:38 +0000)]
Fix typo in Path comment
https://bugs.webkit.org/show_bug.cgi?id=135960

Reviewed by Alexey Proskuryakov.

No new tests, no change on behavior.

* platform/graphics/Path.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172626 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoUnreviewed. Update OptionsGTK.cmake and NEWS for 2.5.3 release.
carlosgc@webkit.org [Fri, 15 Aug 2014 12:23:07 +0000 (12:23 +0000)]
Unreviewed. Update OptionsGTK.cmake and NEWS for 2.5.3 release.

.:

* Source/cmake/OptionsGTK.cmake: Bump version numbers.

Source/WebKit2:

* gtk/NEWS: Add release notes for 2.5.3.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172624 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[GTK] Disable memory sampler on non-Linux system
carlosgc@webkit.org [Fri, 15 Aug 2014 11:30:15 +0000 (11:30 +0000)]
[GTK] Disable memory sampler on non-Linux system
https://bugs.webkit.org/show_bug.cgi?id=134483

Patch by Ting-Wei Lan <lantw44@gmail.com> on 2014-08-15
Reviewed by Philippe Normand.

Source/WebKit2/Shared/linux/WebMemorySamplerLinux.cpp only works on
Linux because it uses many Linux-specific features, so we should disable
memory sampler on other systems by default.

* Source/cmake/OptionsGTK.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172623 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[GTK] HTML API documentation should also be installed versioned
carlosgc@webkit.org [Fri, 15 Aug 2014 09:16:07 +0000 (09:16 +0000)]
[GTK] HTML API documentation should also be installed versioned
https://bugs.webkit.org/show_bug.cgi?id=135970

Reviewed by Philippe Normand.

* Source/PlatformGTK.cmake: Append -${WEBKITGTK_API_VERSION} to
the directory name.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172622 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoRemove two unused Range functions.
akling@apple.com [Fri, 15 Aug 2014 04:24:33 +0000 (04:24 +0000)]
Remove two unused Range functions.
<https://webkit.org/b/135944>

Reviewed by Geoffrey Garen.

* dom/Range.cpp:
(WebCore::Range::maxStartOffset): Deleted.
(WebCore::Range::maxEndOffset): Deleted.
* dom/Range.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172621 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoPreserve editing style for -webkit-tap-highlight-color on all ports that enable ENABL...
dbates@webkit.org [Fri, 15 Aug 2014 04:10:26 +0000 (04:10 +0000)]
Preserve editing style for -webkit-tap-highlight-color on all ports that enable ENABLE_TOUCH_EVENTS
https://bugs.webkit.org/show_bug.cgi?id=135963

Reviewed by Andy Estes.

Currently we only preserve the CSS property -webkit-tap-highlight-color during editing operations
for iOS. Instead we should preserve it for all ports that enable ENABLE_TOUCH_EVENTS as this CSS
property isn't specific to iOS.

Also added FIXME comment with regards to the CSS property -webkit-composition-fill-color.

* editing/EditingStyle.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172620 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoCSS JIT: compile the :empty pseudo class
benjamin@webkit.org [Fri, 15 Aug 2014 04:06:27 +0000 (04:06 +0000)]
CSS JIT: compile the :empty pseudo class
https://bugs.webkit.org/show_bug.cgi?id=135958

Reviewed by Andreas Kling.

Source/WebCore:

Compile :empty, nothing fancy.

Tests: fast/selectors/empty-basics.html
       fast/selectors/empty-long-adjacent-backtracking.html
       fast/selectors/empty-adjacent-style-update.html
       fast/selectors/empty-style-update.html

* cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::addPseudoClassType):
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatching):
(WebCore::SelectorCompiler::jumpIfElementIsNotEmpty):
(WebCore::SelectorCompiler::setElementStyleIsAffectedByEmpty):
(WebCore::SelectorCompiler::setElementStyleFromContextIsAffectedByEmptyAndUpdateRenderStyleIfNecessary):
(WebCore::SelectorCompiler::setElementStyleIsAffectedByEmptyAndUpdateRenderStyleIfNecessary):
(WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsEmpty):
* dom/CharacterData.h:
(WebCore::CharacterData::dataMemoryOffset):
* dom/ContainerNode.h:
(WebCore::ContainerNode::firstChildMemoryOffset):
* dom/Node.h:
(WebCore::Node::flagIsText):

Source/WTF:

* wtf/text/StringImpl.h:
(WTF::StringImpl::lengthMemoryOffset):

LayoutTests:

Add some test coverage, :empty had very little testing.

The test empty-adjacent-style-update expose some problems with style update,
I will look into them separately.

* fast/selectors/empty-adjacent-style-update-expected.txt: Added.
* fast/selectors/empty-adjacent-style-update.html: Added.
* fast/selectors/empty-basics-expected.txt: Added.
* fast/selectors/empty-basics.html: Added.
* fast/selectors/empty-long-adjacent-backtracking-expected.txt: Added.
* fast/selectors/empty-long-adjacent-backtracking.html: Added.
* fast/selectors/empty-style-update-expected.txt: Added.
* fast/selectors/empty-style-update.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172619 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoAllocate the whole RegExpMatchesArray backing store up front.
akling@apple.com [Fri, 15 Aug 2014 03:22:01 +0000 (03:22 +0000)]
Allocate the whole RegExpMatchesArray backing store up front.
<https://webkit.org/b/135217>

We were using the generic array backing store allocation path for
RegExpMatchesArray which meant starting with 4 slots and then growing
it dynamically as we append. Since we always know the final number of
entries up front, allocate a perfectly-sized backing store right away.

~2% progression on Octane/regexp.

Reviewed by Geoffrey Garen.

* runtime/JSArray.h:
(JSC::createArrayButterflyWithExactLength):
* runtime/RegExpMatchesArray.cpp:
(JSC::RegExpMatchesArray::create):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172618 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Dragging Undocked inspector by window title has broken behavior
commit-queue@webkit.org [Fri, 15 Aug 2014 00:30:38 +0000 (00:30 +0000)]
Web Inspector: Dragging Undocked inspector by window title has broken behavior
https://bugs.webkit.org/show_bug.cgi?id=135950

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-08-14
Reviewed by Timothy Hatcher.

* UserInterface/Base/Main.js:
On Mac ports where we the inspector content goes everywhere, window dragging behavior
still exists for the top 22px of the window. So don't do our JavaScript based window
dragging for this special region.

* UserInterface/Models/KeyboardShortcut.js:
(WebInspector.KeyboardShortcut.Modifier.get CommandOrControl):
(WebInspector.KeyboardShortcut.prototype.get displayName):
Instead of calling a host function, use the value already in the frontend.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172617 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoImplement scroll snapping animations on Mac
commit-queue@webkit.org [Fri, 15 Aug 2014 00:21:25 +0000 (00:21 +0000)]
Implement scroll snapping animations on Mac
https://bugs.webkit.org/show_bug.cgi?id=135768

Patch by Wenson Hsieh <wenson_hsieh@apple.com> on 2014-08-14
Reviewed by Beth Dakin.

Implementing the scroll snap animations required for snapping in both overflow and mainframe areas on Mac.
Since we receive a series of discrete wheel events, we need to predict the distance we need to traverse,
compute the appropriate snap point, and then compute an animation curve to that location. The snap animations
are split into two types: snapping, which handles letting go of the trackpad with zero velocity, and gliding,
which handles flick gestures. In both cases, sinusoidal curves are used to ease animation to the target
location. In the former case, the initial velocity is low, and increases to a maximum value during the middle
of the animation before decreasing to 0. In the latter case, the curve is computed such that the initial
velocity matches the user's scroll velocity, and the final velocity matches a lower final velocity that is
somewhat arbitrarily computed. (See the FIXME in AxisScrollSnapOffsets::initializeGlideParameters). How the
equations and constants were chosen is described in greater detail in the comments above
AxisScrollSnapOffsets::computeGlideDelta and AxisScrollSnapOffsets::computeSnapDelta. Note that the final
velocity should ideally be equal to 0. However, with this particular curve, this caused the animation to feel
too slow near the snap point.

No tests, since there is no observable change in behavior. Tests will be included when iOS and Mac scrolling
areas hook into this code.

* PlatformMac.cmake:
* WebCore.xcodeproj/project.pbxproj:
* page/scrolling/AxisScrollSnapOffsets.cpp:
(WebCore::closestSnapOffset): Computes the closest snap offset given velocity and prdicted destination.
* page/scrolling/AxisScrollSnapOffsets.h:
* platform/mac/AxisScrollSnapAnimator.h: Added.
(WebCore::AxisScrollSnapAnimatorClient::~AxisScrollSnapAnimatorClient):
* platform/mac/AxisScrollSnapAnimator.mm: Added.
(WebCore::toWheelEventStatus): Converts a pair of PlatformWheelEventPhases to a WheelEventStatus.
(WebCore::projectedInertialScrollDistance): Attempts to predict the distance covered by the inertial scrolling phase.
(WebCore::AxisScrollSnapAnimator::AxisScrollSnapAnimator):
(WebCore::AxisScrollSnapAnimator::handleWheelEvent): Updates the internal state of the AxisScrollSnapAnimator, given a PlatformWheelEvent
(WebCore::AxisScrollSnapAnimator::shouldOverrideWheelEvent): Determines whether or not we should override a wheel event given the animator's internal state.
(WebCore::AxisScrollSnapAnimator::scrollSnapAnimationUpdate): Updates a single loop of the scroll snapping animation.
(WebCore::AxisScrollSnapAnimator::beginScrollSnapAnimation):
(WebCore::AxisScrollSnapAnimator::endScrollSnapAnimation):
(WebCore::AxisScrollSnapAnimator::computeSnapDelta): See comments for more information.
(WebCore::AxisScrollSnapAnimator::computeGlideDelta): See comments for more information.
(WebCore::AxisScrollSnapAnimator::initializeGlideParameters):
(WebCore::AxisScrollSnapAnimator::pushInitialWheelDelta):
(WebCore::AxisScrollSnapAnimator::averageInitialWheelDelta):
(WebCore::AxisScrollSnapAnimator::clearInitialWheelDeltaWindow):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172616 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Split console toggle button wrong - Images/Images/SplitToggleUp.svg...
commit-queue@webkit.org [Fri, 15 Aug 2014 00:18:25 +0000 (00:18 +0000)]
Web Inspector: Split console toggle button wrong - Images/Images/SplitToggleUp.svg does not exist
https://bugs.webkit.org/show_bug.cgi?id=135957

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-08-14
Reviewed by Timothy Hatcher.

* UserInterface/Views/LogContentView.js:
(WebInspector.LogContentView):
Wrong image path passed to platformImagePath.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172615 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoAllow high fidelity type profiling to be enabled and disabled.
commit-queue@webkit.org [Thu, 14 Aug 2014 23:59:44 +0000 (23:59 +0000)]
Allow high fidelity type profiling to be enabled and disabled.
https://bugs.webkit.org/show_bug.cgi?id=135423

Patch by Saam Barati <sbarati@apple.com> on 2014-08-14
Reviewed by Geoffrey Garen.

Source/JavaScriptCore:

- Merged op_put_to_scope_with_profile and op_get_from_scope_with_profile into
  op_profile_types_with_high_fidelity by adding extra arguments to the opcode.
- Altered SymbolTable to use less memory by adding a rare data structure for
  type profiling.
- Created an interface to turn on and off type profiling from the Web
  Inspector.
- Refactored how entries are written to HighFidelityLog to make it
  easier to inline when generating machine code.
- Implemented op_profile_types_with_high_fidelity in the baseline JIT
  by inlining the process of writing to the log and doing a small amount
  of type inference optimizations.

* bytecode/BytecodeList.json:
* bytecode/BytecodeUseDef.h:
(JSC::computeUsesForBytecodeOffset):
(JSC::computeDefsForBytecodeOffset):
* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpBytecode):
(JSC::CodeBlock::CodeBlock):
(JSC::CodeBlock::finalizeUnconditionally):
(JSC::CodeBlock::scopeDependentProfile): Deleted.
* bytecode/CodeBlock.h:
* bytecode/TypeLocation.h:
(JSC::TypeLocation::TypeLocation):
* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::generate):
(JSC::BytecodeGenerator::emitMove):
(JSC::BytecodeGenerator::emitProfileTypesWithHighFidelity):
(JSC::BytecodeGenerator::emitGetFromScopeWithProfile): Deleted.
(JSC::BytecodeGenerator::emitPutToScopeWithProfile): Deleted.
* bytecompiler/BytecodeGenerator.h:
* bytecompiler/NodesCodegen.cpp:
(JSC::ThisNode::emitBytecode):
(JSC::ResolveNode::emitBytecode):
(JSC::BracketAccessorNode::emitBytecode):
(JSC::DotAccessorNode::emitBytecode):
(JSC::FunctionCallValueNode::emitBytecode):
(JSC::FunctionCallResolveNode::emitBytecode):
(JSC::FunctionCallBracketNode::emitBytecode):
(JSC::FunctionCallDotNode::emitBytecode):
(JSC::CallFunctionCallDotNode::emitBytecode):
(JSC::ApplyFunctionCallDotNode::emitBytecode):
(JSC::PostfixNode::emitResolve):
(JSC::PostfixNode::emitBracket):
(JSC::PostfixNode::emitDot):
(JSC::PrefixNode::emitResolve):
(JSC::PrefixNode::emitBracket):
(JSC::PrefixNode::emitDot):
(JSC::ReadModifyResolveNode::emitBytecode):
(JSC::AssignResolveNode::emitBytecode):
(JSC::AssignDotNode::emitBytecode):
(JSC::ReadModifyDotNode::emitBytecode):
(JSC::AssignBracketNode::emitBytecode):
(JSC::ReadModifyBracketNode::emitBytecode):
(JSC::ReturnNode::emitBytecode):
(JSC::FunctionBodyNode::emitBytecode):
* inspector/agents/InspectorRuntimeAgent.cpp:
(Inspector::InspectorRuntimeAgent::InspectorRuntimeAgent):
(Inspector::InspectorRuntimeAgent::getRuntimeTypesForVariablesAtOffsets):
(Inspector::TypeRecompiler::operator()):
(Inspector::recompileAllJSFunctionsForTypeProfiling):
(Inspector::InspectorRuntimeAgent::willDestroyFrontendAndBackend):
(Inspector::InspectorRuntimeAgent::enableHighFidelityTypeProfiling):
(Inspector::InspectorRuntimeAgent::disableHighFidelityTypeProfiling):
(Inspector::InspectorRuntimeAgent::setHighFidelityTypeProfilingEnabledState):
* inspector/agents/InspectorRuntimeAgent.h:
* inspector/agents/JSGlobalObjectRuntimeAgent.cpp:
(Inspector::JSGlobalObjectRuntimeAgent::willDestroyFrontendAndBackend):
* inspector/protocol/Runtime.json:
* jit/JIT.cpp:
(JSC::JIT::privateCompileMainPass):
(JSC::JIT::privateCompile):
* jit/JIT.h:
* jit/JITOpcodes.cpp:
(JSC::JIT::emit_op_profile_types_with_high_fidelity):
* jit/JITOpcodes32_64.cpp:
(JSC::JIT::emit_op_profile_types_with_high_fidelity):
* jit/JITOperations.cpp:
* jit/JITOperations.h:
* llint/LLIntSlowPaths.cpp:
(JSC::LLInt::LLINT_SLOW_PATH_DECL):
(JSC::LLInt::getFromScopeCommon): Deleted.
(JSC::LLInt::putToScopeCommon): Deleted.
* llint/LLIntSlowPaths.h:
* llint/LowLevelInterpreter.asm:
* runtime/CodeCache.cpp:
(JSC::CodeCache::getGlobalCodeBlock):
* runtime/CommonSlowPaths.cpp:
(JSC::SLOW_PATH_DECL):
* runtime/CommonSlowPaths.h:
* runtime/HighFidelityLog.cpp:
(JSC::HighFidelityLog::initializeHighFidelityLog):
(JSC::HighFidelityLog::~HighFidelityLog):
(JSC::HighFidelityLog::processHighFidelityLog):
* runtime/HighFidelityLog.h:
(JSC::HighFidelityLog::LogEntry::structureIDOffset):
(JSC::HighFidelityLog::LogEntry::valueOffset):
(JSC::HighFidelityLog::LogEntry::locationOffset):
(JSC::HighFidelityLog::recordTypeInformationForLocation):
(JSC::HighFidelityLog::logEndPtr):
(JSC::HighFidelityLog::logStartOffset):
(JSC::HighFidelityLog::currentLogEntryOffset):
* runtime/HighFidelityTypeProfiler.cpp:
(JSC::HighFidelityTypeProfiler::logTypesForTypeLocation):
(JSC::descriptorMatchesTypeLocation):
* runtime/HighFidelityTypeProfiler.h:
* runtime/SymbolTable.cpp:
(JSC::SymbolTable::SymbolTable):
(JSC::SymbolTable::cloneCapturedNames):
(JSC::SymbolTable::prepareForHighFidelityTypeProfiling):
(JSC::SymbolTable::uniqueIDForVariable):
(JSC::SymbolTable::uniqueIDForRegister):
(JSC::SymbolTable::globalTypeSetForRegister):
(JSC::SymbolTable::globalTypeSetForVariable):
* runtime/SymbolTable.h:
(JSC::SymbolTable::add):
(JSC::SymbolTable::set):
* runtime/TypeLocationCache.cpp:
(JSC::TypeLocationCache::getTypeLocation):
* runtime/TypeSet.cpp:
(JSC::TypeSet::getRuntimeTypeForValue):
(JSC::TypeSet::addTypeInformation):
(JSC::TypeSet::allPrimitiveTypeNames):
(JSC::TypeSet::addTypeForValue): Deleted.
* runtime/TypeSet.h:
* runtime/VM.cpp:
(JSC::VM::VM):
(JSC::VM::nextTypeLocation):
(JSC::VM::enableHighFidelityTypeProfiling):
(JSC::VM::disableHighFidelityTypeProfiling):
(JSC::VM::dumpHighFidelityProfilingTypes):
* runtime/VM.h:
(JSC::VM::nextLocation): Deleted.

Source/WebCore:

PageRuntimeAgent and WorkerRuntimeAgent now call their super
class's (InspectorRuntimeAgent) implementation of willDestroyFrontendAndBackend
to give InspectorRuntimeAgent a chance to recompile all JavaScript
functions, if necessary, for type profiling.

* inspector/PageRuntimeAgent.cpp:
(WebCore::PageRuntimeAgent::willDestroyFrontendAndBackend):
* inspector/WorkerRuntimeAgent.cpp:
(WebCore::WorkerRuntimeAgent::willDestroyFrontendAndBackend):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172614 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[Win] Correct build when DerivedSources location not supplied by build environment
bfulgham@apple.com [Thu, 14 Aug 2014 22:22:02 +0000 (22:22 +0000)]
[Win] Correct build when DerivedSources location not supplied by build environment
https://bugs.webkit.org/show_bug.cgi?id=135954
<rdar://problem/18023620>

Reviewed by Joseph Pecoraro.

The DerivedSourcesDirectory is not always defined, depending on what
solution is being used to build this project. Instead, just define
it directly, since we already know where it needs to go.

Also:
1. Stop redirecting the copy output to NUL so I can see if this fails in the future.
2. Make sure the target directory exists before copying to it.

* WebCore.vcxproj/copyForwardingHeaders.cmd: Properly specify
the directories for copying the JS file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172609 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoFixed a bmalloc crash seen on the EWS bot
ggaren@apple.com [Thu, 14 Aug 2014 21:39:47 +0000 (21:39 +0000)]
Fixed a bmalloc crash seen on the EWS bot
https://bugs.webkit.org/show_bug.cgi?id=135955

Reviewed by Andreas Kling.

* bmalloc/Syscall.h: Some CG APIs vm_copy their input buffers. If the
input buffer is a malloc region, that region will get marked Copy-On-Write
by the kernel. Calls to madvise() for COW regions fail and return EINVAL
on older OS X's. In 10.10, they still fail, but they do not return
EINVAL.

So, we can only ASSERT that our syscalls succeed starting with 10.10.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172608 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[CSS3-Text] Cosmetics on RenderBlockFlow::textAlignmentForLine
zoltan@webkit.org [Thu, 14 Aug 2014 21:36:40 +0000 (21:36 +0000)]
[CSS3-Text] Cosmetics on RenderBlockFlow::textAlignmentForLine
https://bugs.webkit.org/show_bug.cgi?id=135940

Reviewed by Darin Adler.

Processing the not CSS3_TEXT case first, and removing an extra whitespace after r172524.

* rendering/RenderBlockLineLayout.cpp:
(WebCore::RenderBlockFlow::textAlignmentForLine):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172607 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Avoid re-binding the same function over and over
commit-queue@webkit.org [Thu, 14 Aug 2014 21:34:36 +0000 (21:34 +0000)]
Web Inspector: Avoid re-binding the same function over and over
https://bugs.webkit.org/show_bug.cgi?id=135949

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-08-14
Reviewed by Timothy Hatcher.

* UserInterface/Controllers/TimelineManager.js:
(WebInspector.TimelineManager.prototype._startAutoCapturing):
(WebInspector.TimelineManager.prototype._stopAutoRecordingSoon):
(WebInspector.TimelineManager.prototype._resetAutoRecordingDeadTimeTimeout):
Stash a bound copy of the function once, and reuse it.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172606 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoFixed the bmalloc build
ggaren@apple.com [Thu, 14 Aug 2014 21:17:36 +0000 (21:17 +0000)]
Fixed the bmalloc build
https://bugs.webkit.org/show_bug.cgi?id=135953

Reviewed by Andreas Kling.

* bmalloc.xcodeproj/project.pbxproj: Marked a few headers as private.
These headers are used, so they must be available outside the project.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172605 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[Win] Extend auto-version and version-stamp for RC_ProjectBuildVersion
bfulgham@apple.com [Thu, 14 Aug 2014 21:09:15 +0000 (21:09 +0000)]
[Win] Extend auto-version and version-stamp for RC_ProjectBuildVersion
https://bugs.webkit.org/show_bug.cgi?id=135948
<rdar://problem/18022221>

Reviewed by David Kilzer.

Tools:

* Scripts/webkitperl/auto-version_unittest/autoVersionTests.pl: Add tests
for RC_ProjectBuildVersion.

WebKitLibraries:

* win/tools/scripts/auto-version.pl:
(splitVersion): Use RC_ProjectBuildVersion (or RC_PROJECTBUILDVERSION) if set
in the environment as the 'Build Version', unless the multi-tuple version
supplied by RC_ProjectSourceVersion already has a fourth component.
* win/tools/scripts/version-stamp.pl: Check for RC_ProjectBuildVersion and
use it in preference to RC_PROJECTBUILDVERSION.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172604 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoIndexedDB should respect SchemeRegistry's database access setting.
mhock@apple.com [Thu, 14 Aug 2014 21:04:45 +0000 (21:04 +0000)]
IndexedDB should respect SchemeRegistry's database access setting.
https://bugs.webkit.org/show_bug.cgi?id=135890
<rdar://problem/17989469>

Reviewed by Alexey Proskuryakov.

Source/WebCore:

Test: storage/indexeddb/open-db-private-browsing.html

* Modules/indexeddb/IDBFactory.cpp:

LayoutTests:

* storage/indexeddb/open-db-private-browsing-expected.txt: Added.
* storage/indexeddb/open-db-private-browsing.html: Added. Simple test to open database in private browsing mode.
* storage/indexeddb/resources/open-db-private-browsing.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172603 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[iOS] run-webkit-tests runs webkit-build-directory on every test
dfarler@apple.com [Thu, 14 Aug 2014 20:58:18 +0000 (20:58 +0000)]
[iOS] run-webkit-tests runs webkit-build-directory on every test
https://bugs.webkit.org/show_bug.cgi?id=135409

Reviewed by Daniel Bates.

* Scripts/webkitpy/port/ios.py:
(IOSSimulatorPort.__init__): Cache Mac build directory.
(IOSSimulatorPort.relay_path): Use cached build directory.
(IOSSimulatorPort._path_to_image_diff): Use cached build directory.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172602 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Timelines can get caught in requestAnimationFrame loop when not recording
commit-queue@webkit.org [Thu, 14 Aug 2014 20:40:17 +0000 (20:40 +0000)]
Web Inspector: Timelines can get caught in requestAnimationFrame loop when not recording
https://bugs.webkit.org/show_bug.cgi?id=135946

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2014-08-14
Reviewed by Timothy Hatcher.

Each TimelineContentView was listening for the TimelineManager's start/stop
capturing events. However, only the active recording could possibly update.
Likewise, if a pre-existing recording started and stopped we did not properly
stop in the case where we were waiting for an event to update the current time,
which would result in an infinite rAF loop of no updates.

* UserInterface/Models/TimelineRecording.js:
(WebInspector.TimelineRecording.prototype.unloaded):
When a timeline recording is unloaded trigger an event.

* UserInterface/Views/TimelineContentView.js:
(WebInspector.TimelineContentView.prototype._startUpdatingCurrentTime):
Assert we were not already waiting.

(WebInspector.TimelineContentView.prototype._stopUpdatingCurrentTime):
Properly handle the case of starting/stoping an already existing recording
without a new event coming in.

(WebInspector.TimelineContentView.prototype._recordingTimesUpdated):
(WebInspector.TimelineContentView.prototype._recordingReset):
Convert to a boolean instead of deleting state.

(WebInspector.TimelineContentView.prototype._recordingUnloaded):
Stop listening to timeline manager events that no longer affect unloaded recordings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172601 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agosrcset with w descriptor doesn't behave as expected when sizes is not supported.
yoav@yoav.ws [Thu, 14 Aug 2014 19:46:43 +0000 (19:46 +0000)]
srcset with w descriptor doesn't behave as expected when sizes is not supported.
https://bugs.webkit.org/show_bug.cgi?id=135935

Reviewed by Dean Jackson.

In current implementation, if sizes is not supported, when authors would write
markup such as `<img src=fallback.jpg srcset="100px.jpg 100w, 400px.jpg 400w">`
the first candidate in srcset would be picked, regardless of its dimensions
That is likely to be confusing for authors.
Dropping these "not yet supported" candidates is likely to be less confusing,
and will result in the "fallback.jpg" candidate being picked.

No new tests since this change only concerns builds that are built
with the PICTURE_SIZES compile flag turned off.

* html/parser/HTMLSrcsetParser.cpp:
(WebCore::parseDescriptors):
Drop candidates that include either 'w' or 'h' descriptors when the
sizes feature is not supported.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172600 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoMake inherited style and parent renderer references
antti@apple.com [Thu, 14 Aug 2014 19:38:37 +0000 (19:38 +0000)]
Make inherited style and parent renderer references
https://bugs.webkit.org/show_bug.cgi?id=135857

Reviewed by Andreas Kling.

We don't recurse into non-rendered subtrees anymore so they can't be null.
Also stop using Element::renderStyle() as it differs from e->renderer().style() in
some rare cases (that should be fixed).

* style/StyleResolveTree.cpp:
(WebCore::Style::RenderTreePosition::parent):
(WebCore::Style::RenderTreePosition::RenderTreePosition):
(WebCore::Style::RenderTreePosition::canInsert):
(WebCore::Style::RenderTreePosition::insert):
(WebCore::Style::RenderTreePosition::computeNextSibling):
(WebCore::Style::shouldCreateRenderer):
(WebCore::Style::styleForElement):
(WebCore::Style::createRendererIfNeeded):
(WebCore::Style::textRendererIsNeeded):
(WebCore::Style::createTextRendererIfNeeded):
(WebCore::Style::updateTextRendererAfterContentChange):
(WebCore::Style::attachChildren):
(WebCore::Style::attachDistributedChildren):
(WebCore::Style::attachShadowRoot):
(WebCore::Style::resetStyleForNonRenderedDescendants):
(WebCore::Style::attachBeforeOrAfterPseudoElementIfNeeded):
(WebCore::Style::attachRenderTree):
(WebCore::Style::resolveLocal):
(WebCore::Style::resolveShadowTree):
(WebCore::Style::updateBeforeOrAfterPseudoElement):
(WebCore::Style::resolveTree):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172599 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoUpdate scope resolution to assume that the parent activation is always there
oliver@apple.com [Thu, 14 Aug 2014 19:28:55 +0000 (19:28 +0000)]
Update scope resolution to assume that the parent activation is always there
https://bugs.webkit.org/show_bug.cgi?id=135947

Reviewed by Andreas Kling.

Another incremental step in removing the idea of lazily created
activations.

* dfg/DFGSpeculativeJIT32_64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* dfg/DFGSpeculativeJIT64.cpp:
(JSC::DFG::SpeculativeJIT::compile):
* jit/JITPropertyAccess.cpp:
(JSC::JIT::emitResolveClosure):
* jit/JITPropertyAccess32_64.cpp:
(JSC::JIT::emitResolveClosure):
* llint/LowLevelInterpreter32_64.asm:
* llint/LowLevelInterpreter64.asm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172598 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[Forms] We should share RenderStyle object for optgroup and option element
antti@apple.com [Thu, 14 Aug 2014 18:54:56 +0000 (18:54 +0000)]
[Forms] We should share RenderStyle object for optgroup and option element
https://bugs.webkit.org/show_bug.cgi?id=88405

Reviewed by Andreas Kling.

We can remove this special case now.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172597 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[Services with UI] Selection services don't work inside <iframes>.
enrica@apple.com [Thu, 14 Aug 2014 18:36:10 +0000 (18:36 +0000)]
[Services with UI] Selection services don't work inside <iframes>.
https://bugs.webkit.org/show_bug.cgi?id=135941
<rdar://problem/17957690>

Reviewed by Tim Horton.

Need to map the selection rectangles using the correct FrameView.
When handling the click, we must use the selection from the focused frame.

* WebProcess/WebPage/mac/ServicesOverlayController.mm:
(WebKit::ServicesOverlayController::buildSelectionHighlight):
(WebKit::ServicesOverlayController::handleClick):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172596 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago<https://webkit.org/b/135909> Move helper applications out of the root of the framework.
mrowe@apple.com [Thu, 14 Aug 2014 18:17:10 +0000 (18:17 +0000)]
<https://webkit.org/b/135909> Move helper applications out of the root of the framework.

As described in <https://developer.apple.com/library/mac/technotes/tn2206/>, for bundles containing
a Versions directory there may be no other content at the top level of the bundle other than symlinks.
Upcoming changes to code signing will prevent bundles that violate this rule from being signed.

Reviewed by Sam Weinig.

Source/WebKit:

* WebKit.xcodeproj/project.pbxproj: Add the symlink to WebKitPluginHost.app in the Versions/A
directory of the framework rather than at the top level.

Source/WebKit2:

* Configurations/Base.xcconfig: Define a configuration setting that points to the content directory
of the framework. On OS X this is Versions/A. On iOS, where frameworks are shallow, this is the top level.
* Configurations/BaseLegacyProcess.xcconfig: Install the legacy processes in the content directory
of the framework.
* WebKit2.xcodeproj/project.pbxproj: Copy the legacy processes into the content directory of the
framework during engineering builds. Generate symlinks for the legacy processes to their locations
in Versions/Current. This is necessary because -[NSBundle pathForAuxiliaryExecutable:] only looks
at the top level of the framework wrapper.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172595 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoCreate activations eagerly
oliver@apple.com [Thu, 14 Aug 2014 18:07:00 +0000 (18:07 +0000)]
Create activations eagerly
https://bugs.webkit.org/show_bug.cgi?id=135942

Reviewed by Geoffrey Garen.

Prepare to rewrite activation objects into a more
sane implementation. Step 1 is reverting to eager
creation of the activation object. This results in
a 1.35x regression in earley, but otherwise has a
minimal performance impact.

The earley regression is being tracked by bug #135943

* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::BytecodeGenerator):
(JSC::BytecodeGenerator::emitNewFunctionInternal):
(JSC::BytecodeGenerator::emitNewFunctionExpression):
(JSC::BytecodeGenerator::emitCallEval):
(JSC::BytecodeGenerator::emitPushWithScope):
(JSC::BytecodeGenerator::emitPushCatchScope):
(JSC::BytecodeGenerator::createActivationIfNecessary): Deleted.
* bytecompiler/BytecodeGenerator.h:
* jit/JITOpcodes.cpp:
(JSC::JIT::emit_op_create_activation):
* jit/JITOpcodes32_64.cpp:
(JSC::JIT::emit_op_create_activation):
* llint/LowLevelInterpreter32_64.asm:
* llint/LowLevelInterpreter64.asm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172594 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoUnreviewed. Removing empty directories.
achristensen@apple.com [Thu, 14 Aug 2014 17:59:53 +0000 (17:59 +0000)]
Unreviewed.  Removing empty directories.

* qt: Removed.
* qt/declarative: Removed.
* qt/docs: Removed.
* qt/docs/webkitsnippets: Removed.
* qt/examples: Removed.
* qt/tests: Removed.
* qt/tests/MIMESniffing: Removed.
* qt/tests/benchmarks: Removed.
* qt/tests/qgraphicswebview: Removed.
* qt/tests/qwebelement: Removed.
* qt/tests/qwebframe: Removed.
* qt/tests/qwebhistory: Removed.
* qt/tests/qwebpage: Removed.
* qt/tests/qwebsecurityorigin: Removed.
* qt/tests/qwebview: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172593 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWebKit2GTK - WebKitWebProcess assertion fails when dragging and dropping a file into...
commit-queue@webkit.org [Thu, 14 Aug 2014 17:52:35 +0000 (17:52 +0000)]
WebKit2GTK - WebKitWebProcess assertion fails when dragging and dropping a file into the view
https://bugs.webkit.org/show_bug.cgi?id=127576

Patch by Michael Catanzaro <mcatanzaro@igalia.com> on 2014-08-14
Reviewed by Carlos Garcia Campos.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::performDragControllerAction): Assume read access
to any file that has been dragged into the web view when compiling for
GTK, since we don't support sandbox extensions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172592 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoRemove nonRendererStyle
antti@apple.com [Thu, 14 Aug 2014 17:49:55 +0000 (17:49 +0000)]
Remove nonRendererStyle
https://bugs.webkit.org/show_bug.cgi?id=135938

Reviewed by Andreas Kling.

It is used to support styling of non-renderer option elements. We have a better
mechanism for this, Element::computedStyle().

* dom/Element.cpp:
(WebCore::Element::resetComputedStyle):
(WebCore::Element::willResetComputedStyle):

    Add a virtual callback for computed style reset.

* dom/Element.h:
* dom/Node.h:
(WebCore::Node::nonRendererStyle): Deleted.
* dom/NodeRenderStyle.h:
(WebCore::Node::renderStyle):

    renderStyle() now always matches renderer()->style()

* html/HTMLOptGroupElement.cpp:
(WebCore::HTMLOptGroupElement::HTMLOptGroupElement):
(WebCore::HTMLOptGroupElement::isFocusable):
(WebCore::HTMLOptGroupElement::didAttachRenderers): Deleted.
(WebCore::HTMLOptGroupElement::willDetachRenderers): Deleted.
(WebCore::HTMLOptGroupElement::updateNonRenderStyle): Deleted.
(WebCore::HTMLOptGroupElement::nonRendererStyle): Deleted.
(WebCore::HTMLOptGroupElement::customStyleForRenderer): Deleted.
* html/HTMLOptGroupElement.h:
* html/HTMLOptionElement.cpp:
(WebCore::HTMLOptionElement::isFocusable):

    Use computedStyle.

(WebCore::HTMLOptionElement::willResetComputedStyle):
(WebCore::HTMLOptionElement::didAttachRenderers): Deleted.
(WebCore::HTMLOptionElement::willDetachRenderers): Deleted.
(WebCore::HTMLOptionElement::updateNonRenderStyle): Deleted.
(WebCore::HTMLOptionElement::nonRendererStyle): Deleted.
(WebCore::HTMLOptionElement::customStyleForRenderer): Deleted.
(WebCore::HTMLOptionElement::didRecalcStyle): Deleted.
* html/HTMLOptionElement.h:
* rendering/RenderMenuList.cpp:
(RenderMenuList::updateOptionsWidth):
(RenderMenuList::setTextFromOption):

    Use computedStyle.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172591 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoAdd support for ppc, ppc64, ppc64le, s390, s390x into the CMake build
commit-queue@webkit.org [Thu, 14 Aug 2014 16:21:06 +0000 (16:21 +0000)]
Add support for ppc, ppc64, ppc64le, s390, s390x into the CMake build
https://bugs.webkit.org/show_bug.cgi?id=135937

Patch by Tomas Popela <tpopela@redhat.com> on 2014-08-14
Reviewed by Carlos Garcia Campos.

* CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172587 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoInitialize a separate variable (TestWTF_SOURCES) with the list of files
zandobersek@gmail.com [Thu, 14 Aug 2014 16:08:24 +0000 (16:08 +0000)]
Initialize a separate variable (TestWTF_SOURCES) with the list of files
that are built into the TestWTF executable and move it above the platform-specific
config file inclusion. This makes it possible to add platform-specific files to that
list (e.g. GUniquePtr.cpp for the GTK port).

Rubber-stamped by Carlos Garcia Campos.

* TestWebKitAPI/CMakeLists.txt:
* TestWebKitAPI/PlatformGTK.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172586 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[GTK] Unreviewed GTK gardening.
clopez@igalia.com [Thu, 14 Aug 2014 15:58:28 +0000 (15:58 +0000)]
[GTK] Unreviewed GTK gardening.

* platform/gtk/TestExpectations: Update expectations with new identified failures and flakiness.
* platform/gtk/fast/ruby/ruby-base-merge-block-children-crash-2-expected.txt: Rebaseline after r172517.
* platform/gtk/fast/ruby/ruby/ruby-base-merge-block-children-crash-2-expected.png: Removed. File was in wrong path.
* platform/gtk/fast/ruby/ruby/ruby-base-merge-block-children-crash-2-expected.txt: Removed. File was in wrong path.
* platform/gtk/inspector-protocol/dom/getAccessibilityPropertiesForNode-expected.txt: Rebaseline after r172136.
* platform/gtk/scrollbars/scrollbar-selectors-expected.txt: Added. Rebaseline after r172220.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172585 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[MSE] Altering the quality of a YouTube video will cause the video to distort and...
jer.noble@apple.com [Thu, 14 Aug 2014 15:13:03 +0000 (15:13 +0000)]
[MSE] Altering the quality of a YouTube video will cause the video to distort and display an error message
https://bugs.webkit.org/show_bug.cgi?id=135931

Reviewed by Eric Carlson.

When removing samples from the TrackBuffer's sample map, also remove those samples from the
TrackBuffer's decode queue. Otherwise, removed samples may persist in the decode queue and
either break sync-sample dependencies or cause decoding artifacts.

Pull the code which removes samples from a track buffer into its own utility function, and
use this function both from removeCodedFrames(), and also when samples are removed due to
overlapping appends in sourceBufferPrivateDidReceiveSample(). In order to reference
TrackBuffers outside of SourceBuffer (and in the static removeSamplesFromTrackBuffer()
function), make TrackBuffer a public forward declaration.

* Modules/mediasource/SourceBuffer.cpp:
(WebCore::removeSamplesFromTrackBuffer):
(WebCore::SourceBuffer::removeCodedFrames):
(WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample):
* Modules/mediasource/SourceBuffer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172584 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[GTK] REGRESSION(r171598, CMake): Introspection files are installed at the wrong...
carlosgc@webkit.org [Thu, 14 Aug 2014 13:10:37 +0000 (13:10 +0000)]
[GTK] REGRESSION(r171598, CMake): Introspection files are installed at the wrong path
https://bugs.webkit.org/show_bug.cgi?id=135934

Reviewed by Gustavo Noronha Silva.

Since r171598 gir files are installed under $datadir/webkitgtk-4.0/gir-1.0/.
The gir files should installed in $datadir/gir-1.0/ together with all other
gir files. We don't need to install introspection files in a versioned
directory because their filenames already contain the binary version. But before
r171598, the files were only installed to the right directory if the
gobject-instrospection pkg-config file was in the same prefix than the one we
wanted to install, because the gir and typelibs directories were extracted from
the pkg-config file. We should define INTROSPECTION_INSTALL_GIRDIR and
INTROSPECTION_INSTALL_TYPELIBDIR in OptionsGTK.cmake based on $datadir and $libdir
like we do in the autotools build.

* Source/cmake/FindGObjectIntrospection.cmake: Do not define
INTROSPECTION_INSTALL_GIRDIR and INTROSPECTION_INSTALL_TYPELIBDIR.
* Source/cmake/OptionsGTK.cmake: Define
INTROSPECTION_INSTALL_GIRDIR and INTROSPECTION_INSTALL_TYPELIBDIR.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172583 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[CSSRegions] Move svg tests into fast/regions/svg
mihnea@adobe.com [Thu, 14 Aug 2014 11:13:02 +0000 (11:13 +0000)]
[CSSRegions] Move svg tests into fast/regions/svg
https://bugs.webkit.org/show_bug.cgi?id=135883

Reviewed by Andrei Bucur.

Move tests and adjust paths accordingly.

* fast/regions/svg/svg-doc-fragment-not-collected-expected.html: Renamed from LayoutTests/fast/regions/svg-doc-fragment-not-collected-expected.html.
* fast/regions/svg/svg-doc-fragment-not-collected.html: Renamed from LayoutTests/fast/regions/svg-doc-fragment-not-collected.html.
* fast/regions/svg/svg-element-not-collected-expected.html: Renamed from LayoutTests/fast/regions/svg-element-not-collected-expected.html.
* fast/regions/svg/svg-element-not-collected.html: Renamed from LayoutTests/fast/regions/svg-element-not-collected.html.
* fast/regions/svg/svg-root-element-collected-expected.txt: Renamed from LayoutTests/fast/regions/svg-root-element-collected-expected.txt.
* fast/regions/svg/svg-root-element-collected.html: Renamed from LayoutTests/fast/regions/svg-root-element-collected.html.
* fast/regions/svg/symbol-in-named-flow-crash-expected.txt: Renamed from LayoutTests/fast/regions/symbol-in-named-flow-crash-expected.txt.
* fast/regions/svg/symbol-in-named-flow-crash.svg: Renamed from LayoutTests/fast/regions/symbol-in-named-flow-crash.svg.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172582 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[GTK] WebKit DOM documentation (webkitdomgtk) not installed after make install
carlosgc@webkit.org [Thu, 14 Aug 2014 10:31:39 +0000 (10:31 +0000)]
[GTK] WebKit DOM documentation (webkitdomgtk) not installed after make install
https://bugs.webkit.org/show_bug.cgi?id=135836

Reviewed by Philippe Normand.

* Source/PlatformGTK.cmake: Add install command to also install
the GObject DOM bindings API docs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172581 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoImageBufferDataCairo.h is missing header guards
zandobersek@gmail.com [Thu, 14 Aug 2014 09:35:00 +0000 (09:35 +0000)]
ImageBufferDataCairo.h is missing header guards
https://bugs.webkit.org/show_bug.cgi?id=135933

Reviewed by Philippe Normand.

* platform/graphics/cairo/ImageBufferDataCairo.h:
Add the missing head guards to avoid problems when re-inclusion occurs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172580 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoUnreviewed. In r165709 I mistakenly appended the -fno-exceptions
zandobersek@gmail.com [Thu, 14 Aug 2014 09:33:58 +0000 (09:33 +0000)]
Unreviewed. In r165709 I mistakenly appended the -fno-exceptions
and -fno-strict-aliasing compiler flags to CMAKE_CXX_FLAGS and then
re-set CMAKE_C_FLAGS with the new string. The two flags should really
be appended to CMAKE_C_FLAGS and the same variable re-set with the
new string.

* Source/cmake/OptionsCommon.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172579 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoFix JSC::ARM64Assembler::LinkRecord::RealTypes
commit-queue@webkit.org [Thu, 14 Aug 2014 08:49:48 +0000 (08:49 +0000)]
Fix JSC::ARM64Assembler::LinkRecord::RealTypes
https://bugs.webkit.org/show_bug.cgi?id=135906

Patch by Akos Kiss <akiss@inf.u-szeged.hu> on 2014-08-14
Reviewed by Michael Saboff.

JSC::ARM64Assembler::LinkRecord::RealTypes::m_compareRegister is defined
to occupy 5 bits but JSC::ARM64Assembler::RegisterID needs 6 bits. So,
increase the size of the bit field and also reorganize the struct to
better align with word boundaries.

* assembler/ARM64Assembler.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172578 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoAttempt to fix the build following <http://trac.webkit.org/changeset/172576>
dbates@webkit.org [Thu, 14 Aug 2014 05:09:32 +0000 (05:09 +0000)]
Attempt to fix the build following <trac.webkit.org/changeset/172576>
(https://bugs.webkit.org/show_bug.cgi?id=135895)

Substitute PerThreadStorage<T>::initSharedKeyIfNeeded() for initSharedKeyIfNeeded() in
implementation of PerThread<T>::getFastCase().

* bmalloc/PerThread.h:
(bmalloc::PerThread<T>::getFastCase):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172577 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoMake bmalloc::PerThread work without C++ thread local storage
dbates@webkit.org [Thu, 14 Aug 2014 05:00:22 +0000 (05:00 +0000)]
Make bmalloc::PerThread work without C++ thread local storage
https://bugs.webkit.org/show_bug.cgi?id=135895

Reviewed by Geoffrey Garen.

Implement support for building bmalloc without C++ thread local storage.

* bmalloc/BPlatform.h: Remove macro define BPLATFORM_IOS_SIMULATOR. Added macro function
BCOMPILER_SUPPORTS() and macro define BCOMPILER_SUPPORTS_CXX_THREAD_LOCAL that can be used
to determine whether the compiler supports C++ thread local storage.
* bmalloc/PerThread.h:
(bmalloc::PerThreadStorage::get): Modified to call pthread_getspecific() when building
without C++ thread local storage.
(bmalloc::PerThreadStorage::initSharedKeyIfNeeded): Added.
(bmalloc::PerThreadStorage::init): Moved logic to initialize shared Pthread key from here to
PerThreadStorage::initSharedKeyIfNeeded().
(bmalloc::PerThread<T>::getFastCase): Modified to call PerThreadStorage::initSharedKeyIfNeeded()
before querying PerThreadStorage::get() when building without C++ thread local storage so as to
ensure that the shared key has been initialized.
(_pthread_setspecific_direct): Deleted.
(_pthread_getspecific_direct): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172576 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoAdd ARM64 support to CMake-based builds
commit-queue@webkit.org [Thu, 14 Aug 2014 02:53:35 +0000 (02:53 +0000)]
Add ARM64 support to CMake-based builds
https://bugs.webkit.org/show_bug.cgi?id=135912

Patch by Akos Kiss <akiss@inf.u-szeged.hu> on 2014-08-13
Reviewed by Gyuyoung Kim.

This patch ensures that CMake does not fail with Unknown CPU error when
building for ARM64.

* CMakeLists.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172550 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[EFL] Bump up and patch dependencies to allow building for aarch64
commit-queue@webkit.org [Thu, 14 Aug 2014 02:52:22 +0000 (02:52 +0000)]
[EFL] Bump up and patch dependencies to allow building for aarch64
https://bugs.webkit.org/show_bug.cgi?id=135885

Patch by Akos Kiss <akiss@inf.u-szeged.hu> on 2014-08-13
Reviewed by Gyuyoung Kim.

The currently used versions of pixman, libffi, glib, glib-networking,
and fontconfig fail to configure for aarch64, so they are bumped up to
the latest stable version.

Gstreamer-plugins-base configures for aarch64 but incorrectly recogizes
ARM NEON assembly support, so it is patched with a change that has been
accepted by upstream.

* efl/jhbuild.modules:
* efl/patches/gst-prevent-neon-check-in-configure-from-passing-under-aarch64.patch: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172549 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoEnable CSS_SCROLL_SNAP for iOS
commit-queue@webkit.org [Thu, 14 Aug 2014 01:22:58 +0000 (01:22 +0000)]
Enable CSS_SCROLL_SNAP for iOS
https://bugs.webkit.org/show_bug.cgi?id=135915

Turn on CSS_SCROLL_SNAP for iOS and the iOS simulator.

Patch by Wenson Hsieh <wenson_hsieh@apple.com> on 2014-08-13
Reviewed by Tim Horton.

* Configurations/FeatureDefines.xcconfig:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172548 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: eslint configuration should be stored as .eslintrc
jonowells@apple.com [Thu, 14 Aug 2014 01:13:15 +0000 (01:13 +0000)]
Web Inspector: eslint configuration should be stored as .eslintrc
https://bugs.webkit.org/show_bug.cgi?id=135916

.eslintrc file added to allow for ESLint solutions, within editors and otherwise,
to be automatically configured based on the WebKit JavaScript coding style. The
initial file is incomplete.

Reviewed by Joseph Pecoraro.

* .eslintrc: Added.
* External/.eslintrc: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172547 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago[Services with UI] Use a longer delay duration for editable content.
enrica@apple.com [Thu, 14 Aug 2014 00:39:28 +0000 (00:39 +0000)]
[Services with UI] Use a longer delay duration for editable content.
https://bugs.webkit.org/show_bug.cgi?id=135918
<rdar://problem/17998929>

Reviewed by Tim Horton.

Use a 1 second delay for selections in editable content.

* WebProcess/WebPage/mac/ServicesOverlayController.mm:
(WebKit::ServicesOverlayController::remainingTimeUntilHighlightShouldBeShown):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172546 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoLand correct results for compositing/fixed-image-loading.html on WK1 and WK2.
simon.fraser@apple.com [Thu, 14 Aug 2014 00:23:58 +0000 (00:23 +0000)]
Land correct results for compositing/fixed-image-loading.html on WK1 and WK2.

* compositing/fixed-image-loading-expected.txt:
* platform/mac-wk2/compositing/fixed-image-loading-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172545 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoWeb Inspector: Missing semicolon after console.assert in TimelineSidebarPanel.js
timothy@apple.com [Thu, 14 Aug 2014 00:23:11 +0000 (00:23 +0000)]
Web Inspector: Missing semicolon after console.assert in TimelineSidebarPanel.js
https://bugs.webkit.org/show_bug.cgi?id=135917

Reviewed by Joseph Pecoraro.

* UserInterface/Views/TimelineSidebarPanel.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172544 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years agoCompute and store snap point positions
commit-queue@webkit.org [Thu, 14 Aug 2014 00:16:03 +0000 (00:16 +0000)]
Compute and store snap point positions
https://bugs.webkit.org/show_bug.cgi?id=135268

Patch by Wenson Hsieh <wenson_hsieh@apple.com> on 2014-08-13
Reviewed by Simon Fraser.

Working towards implementing a way to extract snap positions (i.e. LayoutUnits) from the parsed snap point properties from CSS,
as well as a way of storing the information. Supports both mainframe and overflow scrolling.

No changes in behavior. I will write tests when I implement snapping for iOS and Mac.

* WebCore.xcodeproj/project.pbxproj:
* page/FrameView.cpp:
(WebCore::FrameView::updateSnapOffsets): Invokes the static routine in AxisScrollSnapOffsets to update snap offsets for mainframe.
* page/FrameView.h:
* page/scrolling/AxisScrollSnapOffsets.cpp: Added.
(WebCore::appendAndSortChildSnapOffsets):
(WebCore::updateFromStyle):
(WebCore::updateSnapOffsetsForScrollableArea): Update the ScrollableArea's snap axes.
* page/scrolling/AxisScrollSnapOffsets.h: Added.
* platform/ScrollableArea.cpp:
(WebCore::ScrollableArea::setHorizontalSnapOffsets):
(WebCore::ScrollableArea::setVerticalSnapOffsets):
(WebCore::ScrollableArea::clearHorizontalSnapOffsets): Resets the unique_ptr for horizontal snap points.
(WebCore::ScrollableArea::clearVerticalSnapOffsets): Resets the unique_ptr for vertical snap points.
* platform/ScrollableArea.h:
(WebCore::ScrollableArea::horizontalSnapOffsets): Returns a pointer to the ScrollableArea's horizontal snap points.
(WebCore::ScrollableArea::verticalSnapOffsets): Returns a pointer to the ScrollableArea's vertical snap points.
(WebCore::ScrollableArea::updateSnapOffsets):
* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::updateSnapOffsets): Invokes the static routine in AxisScrollSnapOffsets to update snap offsets for an overflow area.
* rendering/RenderLayer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172543 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 years ago<https://webkit.org/b/135911> WebKit should build on Yosemite with the public SDK.
mrowe@apple.com [Wed, 13 Aug 2014 23:55:10 +0000 (23:55 +0000)]
<https://webkit.org/b/135911> WebKit should build on Yosemite with the public SDK.

Reviewed by Darin Adler.

Source/WebCore:

* rendering/RenderThemeMac.mm: Fix the forward-declaration of NSServicesRolloverButtonCell.

Source/WebKit/mac:

* Misc/WebSharingServicePickerController.mm: Forward-declare some details related to NSSharingServicePicker.

Source/WebKit2:

* Platform/IPC/mac/ImportanceAssertion.h: Forward-declare the new assertion functions we use.
* UIProcess/mac/WebContextMenuProxyMac.mm: Forward-declare some details related to NSSharingServicePicker.

Tools:

* DumpRenderTree/mac/TextInputController.m: Don't use extern "C" in a non-C++ file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@172542 268f45cc-cd09-0410-ab3c-d52691b4dbfc