iOS WebKitLegacy should match WebKit2's viewport behavior
https://bugs.webkit.org/show_bug.cgi?id=146564
<rdar://problem/20574017>

Reviewed by Brent Fulgham.

* WebCoreSupport/WebChromeClientIOS.mm:
(dictionaryForViewportArguments):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186245 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WebKit should use 80% white background for PiP indicator
https://bugs.webkit.org/show_bug.cgi?id=146444
<rdar://problem/21555726>

Reviewed by Brent Fulgham.

The designers wanted 80% white, not 80% black (r186104).
I've also reverted the AirPlay background back to black.

* Modules/mediacontrols/mediaControlsiOS.css:
(audio::-webkit-media-controls-wireless-playback-status):
(audio::-webkit-media-controls-wireless-playback-status.small):
(audio::-webkit-media-controls-wireless-playback-status.picture-in-picture):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186244 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Gardening: fix broken 32-bit build..
https://bugs.webkit.org/show_bug.cgi?id=146416

Not reviewed.

* UIProcess/API/Cocoa/APISerializedScriptValueCocoa.mm:
(API::SerializedScriptValue::deserialize):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186243 268f45cc-cd09-0410-ab3c-d52691b4dbfc

bmalloc: Shrink the super chunk size
https://bugs.webkit.org/show_bug.cgi?id=146519

Reviewed by Andreas Kling.

We have lots of reports of crashing due to failed VM allocation on iOS.
(This VM limit on iOS is usually 1GB-2GB, and has been as low as 256MB.)

Shrink the super chunk size in case fragmentation is the reason for
VM allocation failure.

This has the downside that >= 2MB allocations will now be super slow,
but they are also super rare (as in never on most websites), so this
is probably an OK tradeoff.

* bmalloc/Sizes.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186242 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Correct *another* typo in r186238.

* UIProcess/API/Cocoa/WKUIDelegatePrivate.h: Fix parameter name scrolView -> scrollView.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186241 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the build.

Rubber-stamped by Brady Eidson.

* WebView/WebView.mm:
(-[WebView initSimpleHTMLDocumentWithStyle:frame:preferences:groupName:]):
Http, not HTTP, for settings generator reasons.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186240 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Correct typo in r186238.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView scrollViewWillEndDragging:withVelocity:targetContentOffset:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186239 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Mac] Clear current AirPlay source before setting another
https://bugs.webkit.org/show_bug.cgi?id=145959

Reviewed by Brent Fulgham.

* Modules/mediasession/WebMediaSessionManager.cpp:
(WebCore::WebMediaSessionManager::clientStateDidChange): Don't make a client begin playing
  to the target just because it has paused.
(WebCore::WebMediaSessionManager::configurePlaybackTargetClients): Always select a client that
  requrested the picker, and never consider paused clients when choosing a client to begin
  playing to the targer. Call setShouldPlayToPlaybackTarget(true) after all of the other
  clients have been told to stop playing to the target.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186238 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Provide delegate SPI for clients to notify WebKit about content inset changes
https://bugs.webkit.org/show_bug.cgi?id=146482
<rdar://problem/21602741>

Reviewed by Darin Adler.

Provide a new SPI for WebKit client applications to use to inform WebKit of planned
adjustments to the edge insets of the view.

* UIProcess/API/Cocoa/WKUIDelegatePrivate.h: Add new optional delegate method.
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView scrollViewWillEndDragging:withVelocity:targetContentOffset:]): If the new delegate
method exists, use the specified content insets, rather than the current state of the view.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186237 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Allow the UIDelegate to customize an image preview
https://bugs.webkit.org/show_bug.cgi?id=146557
-and corresponding-
rdar://problem/21657424

Reviewed by Tim Horton.

Source/WebCore:

Add WEBCORE_EXPORT to use this in WK2.
* loader/cache/CachedResource.h:
(WebCore::CachedResource::url):

Source/WebKit2:

Add imageURL to InteractionInformationAtPosition.
* Shared/InteractionInformationAtPosition.cpp:
(WebKit::InteractionInformationAtPosition::encode):
(WebKit::InteractionInformationAtPosition::decode):
* Shared/InteractionInformationAtPosition.h:

Two new delegate methods.
* UIProcess/API/Cocoa/WKUIDelegatePrivate.h:

Call the delegate methods at the appropriate time.
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView previewViewControllerForPosition:inSourceView:]):
(-[WKContentView commitPreviewViewController:]):

Include the imageURL in InteractionInformationAtPosition.
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::getPositionInformation):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186236 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Instead of requiring ToT UIKIt, check that UI_WEB_TOUCH_EVENT_HAS_IS_POTENTIAL_TAP is defined to 1
https://bugs.webkit.org/show_bug.cgi?id=146558

Reviewed by Beth Dakin.

* Shared/ios/NativeWebTouchEventIOS.mm:
(WebKit::NativeWebTouchEvent::NativeWebTouchEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186235 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove change I accidentally committed in r186232

* UIProcess/WebsiteData/WebsiteDataStore.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186234 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix EWS build after r186182.

* WebProcess/WebPage/WebPage.h:
(WebKit::WebPage::userMediaPermissionRequestManager):
Don't export a function defined in a header, or it will generate a weak external symbol.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186233 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Add preference to disable all http-equiv.
<rdar://problem/9091261> and https://bugs.webkit.org/show_bug.cgi?id=146553

Reviewed by Sam Weinig.
Source/WebCore:

No new tests (Covered by existing test)

* dom/Document.cpp:
(WebCore::Document::processHttpEquiv):
* page/Settings.in:

Source/WebKit/mac:

* WebView/WebPreferenceKeysPrivate.h:

* WebView/WebPreferences.mm:
(+[WebPreferences initialize]):
(-[WebPreferences setMetaRefreshEnabled:]): Still expose this SPI, but have it affect
  the "httpEquivEnabled" setting.
(-[WebPreferences metaRefreshEnabled]): Ditto.
(-[WebPreferences setHTTPEquivEnabled:]):
(-[WebPreferences httpEquivEnabled]):
* WebView/WebPreferencesPrivate.h:

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

Source/WebKit2:

* Shared/WebPreferencesDefinitions.h:

* UIProcess/API/C/WKPreferences.cpp:
(WKPreferencesSetMetaRefreshEnabled): Still expose this SPI, but have it affect
  the "httpEquivEnabled" setting.
(WKPreferencesGetMetaRefreshEnabled): Ditto
(WKPreferencesSetHTTPEquivEnabled):
(WKPreferencesGetHTTPEquivEnabled):
* UIProcess/API/C/WKPreferencesRefPrivate.h:

* WebProcess/InjectedBundle/InjectedBundle.cpp:
(WebKit::InjectedBundle::overrideBoolPreferenceForTestRunner):

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

LayoutTests:

* loader/meta-refresh-disabled.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186232 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Streams API] Remove ReadableStreamController.enqueue() custom binding
https://bugs.webkit.org/show_bug.cgi?id=146455

Reviewed by Darin Adler.

Made enqueue not custom. Updated error to use Default=Undefined in IDL.

No change in behavior.

* Modules/streams/ReadableStreamController.h:
(WebCore::ReadableStreamController::error): Removed unneeded variation of error.
(WebCore::ReadableStreamController::enqueue): Calling ReadableJSStream enqueue method.
* Modules/streams/ReadableStreamController.idl: Updated error with Default=Undefined and made enqueue not custom.
* bindings/js/JSDOMBinding.cpp:
(WebCore::createDOMException): Adding support for RangeError exceptions.
* bindings/js/JSReadableStreamControllerCustom.cpp:
(WebCore::constructJSReadableStreamController): Deleted.
* bindings/js/ReadableJSStream.cpp:
(WebCore::ReadableJSStream::error): Updated to pass error value as parameter.
(WebCore::ReadableJSStream::enqueue): Added exception throwing through ExceptionCode.
(WebCore::ReadableJSStream::retrieveChunkSize): Ditto.
* bindings/js/ReadableJSStream.h:
* dom/ExceptionCode.h: Adding RangeError.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186231 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Build fix for Win EWS bot.
https://bugs.webkit.org/show_bug.cgi?id=146551

Not reviewed.

* tools/JSDollarVMPrototype.cpp:
(JSC::functionCrash):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186230 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WKWebView evaluateJavaScript:completionHandler: should reuse its JSContext instance.
https://bugs.webkit.org/show_bug.cgi?id=146416

Reviewed by Anders Carlsson.

The core issue here is that we need to be able to deserialized an
API::SerializedScriptValue instance into an ObjC object graph.  We'll solve this
by introducing a API::SerializedScriptValue::deserialize() functions that does
the job.  The deserialize() function will will use a SharedJSContext to provide
the needed JSContext for doing this work.

Also updated ScriptMessageHandlerDelegate::didPostMessage() to use this new
deserialize() function.

* Shared/API/APISerializedScriptValue.h:
(API::SerializedScriptValue::deserialize):
(API::SerializedScriptValue::internalRepresentation):
- Changed to return a WebCore::SerializedScriptValue*.  This makes it friendlier
  to use in the 2 places that need it.

* UIProcess/API/Cocoa/APISerializedScriptValueCocoa.mm: Added.
(API::SharedJSContext::SharedJSContext):
(API::SharedJSContext::ensureContext):
(API::SharedJSContext::releaseContext):
(API::SerializedScriptValue::deserialize):
* UIProcess/API/Cocoa/WKUserContentController.mm:
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView evaluateJavaScript:completionHandler:]):
* UIProcess/API/gtk/WebKitWebView.cpp:
(webkitWebViewRunJavaScriptCallback):
- Removed a now unnecessary cast.
* WebKit2.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186229 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Content Extensions] Block synchronous XMLHTTPRequest.
<rdar://problem/21573006> and https://bugs.webkit.org/show_bug.cgi?id=146271

Reviewed by Alex Christensen.

Source/WebCore:

Test: http/tests/contentextensions/sync-xhr-blocked.html

* loader/FrameLoader.cpp:
(WebCore::FrameLoader::loadResourceSynchronously): If content blocked, set up an error, clear
  the response, and clear the response data.

LayoutTests:

* http/tests/contentextensions/sync-xhr-blocked-expected.txt: Added.
* http/tests/contentextensions/sync-xhr-blocked.html: Added.
* http/tests/contentextensions/sync-xhr-blocked.html.json: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186228 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: Show suggest popover on Tab press even if it wasn't showing before
https://bugs.webkit.org/show_bug.cgi?id=146496

Reviewed by Timothy Hatcher.

* UserInterface/Views/CSSStyleDeclarationTextEditor.js:
(WebInspector.CSSStyleDeclarationTextEditor.prototype._handleTabKey): Pressing tab will try to autocomplete before trying
to add an ending colon/semicolon or highlight the next section of text.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186227 268f45cc-cd09-0410-ab3c-d52691b4dbfc

<rdar://problem/21429613> [iOS] Stop making symlinks from PrivateFrameworks to Frameworks
https://bugs.webkit.org/show_bug.cgi?id=146542

Reviewed by Sam Weinig.

Source/JavaScriptCore:

* JavaScriptCore.xcodeproj/project.pbxproj: Removed the build phase that makes the symlink.

Source/WebKit2:

* WebKit2.xcodeproj/project.pbxproj: Removed the build phase that makes the directory of
symlinks.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186226 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] WebSQL doesn't work because openDatabase always fails with DOM Exception 18
https://bugs.webkit.org/show_bug.cgi?id=146234

Reviewed by Sergio Villar Senin.

Source/WebKit2:

This is because we don't provide any quota, and 0 is used by
default, so there's never enough quota and openDatabase fails. We
should expose this in the API, but for now, we could use a default
quota of 5MB like WTR does.

* UIProcess/API/gtk/WebKitUIClient.cpp: Override
exceededDatabaseQuota and return always the default quota.
* UIProcess/gtk/WebInspectorProxyGtk.cpp:
(WebKit::exceededDatabaseQuota): Return the quota based on the
expected usage and current database usabe like mac does.
(WebKit::WebInspectorProxy::platformCreateInspectorPage): Add
custom UI client to implement exceededDatabaseQuota.

Tools:

Enable the WebSQL directory test that was disabled because of this bug.

* TestWebKitAPI/Tests/WebKit2Gtk/TestWebKitWebContext.cpp:
(testWebContextConfiguration):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186225 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, rolling out r185704.
https://bugs.webkit.org/show_bug.cgi?id=146544

Didn't fix the problem (Requested by anttik on #webkit).

Reverted changeset:

"~4% Membuster regression after WebKit r185452"
https://bugs.webkit.org/show_bug.cgi?id=146112
http://trac.webkit.org/changeset/185704

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186224 268f45cc-cd09-0410-ab3c-d52691b4dbfc

iOS WebKitLegacy should match WebKit2's viewport behavior
https://bugs.webkit.org/show_bug.cgi?id=146540
<rdar://problem/20574017>

Reviewed by Dan Bernstein.

A number of changes have been made to WebKit2's viewport behavior
to make it possible to scale "responsive" sites that misbehave (lay out
in a non-responsive way).

* History/WebHistoryItem.mm:
* History/WebHistoryItemPrivate.h:
Add the plumbing required to implement these new behaviors in WebKit1.

* WebKit.iOS.exp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186223 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Skip new web timing test on platforms without web timing.

* platform/ios-sim-deprecated/Skipped:
* platform/mac/TestExpectations:
* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186222 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] Accelerated Compositing stops working after a web process crash
https://bugs.webkit.org/show_bug.cgi?id=146508

Reviewed by Martin Robinson.

The problem is that we don't send the window ID again to the new
web process.

* UIProcess/API/gtk/PageClientImpl.cpp:
(WebKit::PageClientImpl::didRelaunchProcess): Call
webkitWebViewBaseDidRelaunchWebProcess().
* UIProcess/API/gtk/WebKitWebViewBase.cpp:
(webkitWebViewBaseDidRelaunchWebProcess): Set the window ID to
the new drawing area.
* UIProcess/API/gtk/WebKitWebViewBasePrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186221 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Source/ThirdParty/ANGLE:
Re-enable WebGL on WinCairo
https://bugs.webkit.org/show_bug.cgi?id=146537

Reviewed by Csaba Osztrogonác.

* ANGLE.vcxproj/DirectX32.props:
* ANGLE.vcxproj/DirectX64.props:
* ANGLE.vcxproj/libGLESv2Common.props:
Update DirectX directories to use the Windows SDK instead of the June 2010 DirectX SDK.

Source/WebKit:
Re-enable WebGL on WinCairo.
https://bugs.webkit.org/show_bug.cgi?id=146537

Reviewed by Csaba Osztrogonác.

* WebKit.vcxproj/WebKit.sln:
Build the ANGLE projects on WinCairo again.

Source/WTF:
Re-enable WebGL on WinCairo
https://bugs.webkit.org/show_bug.cgi?id=146537

Reviewed by Csaba Osztrogonác.

* wtf/FeatureDefines.h:
Enable WebGL on WinCairo again.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186220 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix test from r186208 and r186216.

* http/tests/misc/webtiming-resolution-expected.txt:
* http/tests/misc/webtiming-resolution.html:
Properly and explicitly test if the time delta is close to a multiple of the expected resolution.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186219 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: Aggregate profile call information on the backend to drastically reduce profile sizes
https://bugs.webkit.org/show_bug.cgi?id=146536

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-07-01
Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

* inspector/protocol/Timeline.json:
Change a CPUProfile from sending a required "calls" param to sending a required
"callInfo" param which includes aggregated information about the calls.

Source/WebCore:

* inspector/TimelineRecordFactory.cpp:
(WebCore::buildAggregateCallInfoInspectorObject):
(WebCore::buildInspectorObject):
Replace the array of Call objects with a single aggregated call info object.

Source/WebInspectorUI:

Since we still support legacy backends, the frontend documents where
it is handling legacy backends with compatibility comments.

* UserInterface/Models/ProfileNode.js:
(WebInspector.ProfileNode):
(WebInspector.ProfileNode.prototype.get callInfo):
Handle a ProfileNode created with callInfo or calls. They are mutually exclusive.

* UserInterface/Models/ScriptTimelineRecord.js:
(WebInspector.ScriptTimelineRecord.prototype._initializeProfileFromPayload.profileNodeFromPayload):
If the profile has per-call information, construct ProfileNodeCall objects, otherwise
just construct the ProfileNode with the aggregate callInfo value.

* UserInterface/Views/ScriptTimelineDataGridNode.js:
(WebInspector.ScriptTimelineDataGridNode.prototype.get data):
When we have aggregate call information we cannot easily partition a script,
so we instead treat the entire script as one large atomic unit in the timeline.
If the timeline range has any portion of the script, show the entire script.
Users used to be able to select a portion of a script and view the relevant
functions called in just that sliver, but this doesn't appear to be a well
known feature or commonly used. In fact, given the small ranges of time it
could be confusing for users.

* UserInterface/Views/TimelineRecordingContentView.js:
(WebInspector.TimelineRecordingContentView.prototype.matchTreeElementAgainstCustomFilters):
Treat as a discrete unit.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186218 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: When autocompleting, pressing tab twice shouldn't insert a tab character
https://bugs.webkit.org/show_bug.cgi?id=145885

Reviewed by Timothy Hatcher.

* UserInterface/Controllers/CodeMirrorCompletionController.js:
(WebInspector.CodeMirrorCompletionController):
(WebInspector.CodeMirrorCompletionController.prototype.updateCompletions): Resolves the promise as having completions.
(WebInspector.CodeMirrorCompletionController.prototype.hideCompletions): Resolves the promise as not having completions.
(WebInspector.CodeMirrorCompletionController.prototype.completeAtCurrentPositionIfNeeded): Returns a WrappedPromise that allows
callers of this function to determine if the autocomplete had any values or was instead not shown.
(WebInspector.CodeMirrorCompletionController.prototype._resolveUpdatePromise):
* UserInterface/Main.html: Added WrappedPromise class.
* UserInterface/Models/WrappedPromise.js: Added WrappedPromise object to expose resolve and reject functions.
* UserInterface/Views/ConsolePrompt.js:
(WebInspector.ConsolePrompt.prototype._handleTabKey): Attempts to find completions for current text.  If there are none, beep.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186217 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reduce resolution of performance.now
https://bugs.webkit.org/show_bug.cgi?id=146531

Reviewed by Simon Fraser.

Make layout test introduced in r186208.

* http/tests/misc/webtiming-resolution-expected.txt:
* http/tests/misc/webtiming-resolution.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186216 268f45cc-cd09-0410-ab3c-d52691b4dbfc

DFG::freezeFragile should register the frozen value's structure
https://bugs.webkit.org/show_bug.cgi?id=136055
rdar://problem/21042120

Reviewed by Mark Lam and Geoffrey Garen.

This fixes weird concurrency bugs where the constant folding phase tries to convert
something to a constant but then crashes because the constant's structure wasn't
registered. The AI was registering the structure of any value it saw, but constant folding
wasn't - and that's fine so long as there ain't no concurrency.

The best fix is to just make it impossible to introduce a constant into the IR without
registering its structure. That's what this change does. This is not only a great
concurrency fix - it also makes the compiler somewhat easier to hack on because it's one
less case of structure registering that you have to remember about.

* dfg/DFGAbstractValue.cpp:
(JSC::DFG::AbstractValue::setOSREntryValue): No need to register.
(JSC::DFG::AbstractValue::set): We still call register, but just to get the watchpoint state.
* dfg/DFGGraph.cpp:
(JSC::DFG::Graph::freezeFragile): Register the structure.
* dfg/DFGStructureRegistrationPhase.cpp:
(JSC::DFG::StructureRegistrationPhase::run): Assert that these are all registered.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186215 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: [Mac] InspectorFrontendHost.beep() doesn't play a sound, sandbox warnings
https://bugs.webkit.org/show_bug.cgi?id=146525

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-07-01
Reviewed by Timothy Hatcher.

* WebProcess/com.apple.WebProcess.sb.in:
Get NSBeep working without any sandbox warnings.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186214 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Make the first click on a rule section create a newline for easy property addition
https://bugs.webkit.org/show_bug.cgi?id=146490

Reviewed by Timothy Hatcher.

* UserInterface/Views/CSSStyleDeclarationTextEditor.js:
(WebInspector.CSSStyleDeclarationTextEditor):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._highlightNextNameOrValue):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._handleEnterKey): Inserts a semicolon if the line is missing one.
(WebInspector.CSSStyleDeclarationTextEditor.prototype._handleMouseDown): If the user clicks on a property with the editor being
unfocused, the name/value containing the cursor will be highlighted.  If instead the user clicks at the end of a line, the
cursor's position is saved for mouseUp.
(WebInspector.CSSStyleDeclarationTextEditor.prototype._handleMouseUp): If the mouseDown cursor position was saved and is equal
to the current cursor's position (the user did not drag), add a newline after the current line and place the cursor on that line.
(WebInspector.CSSStyleDeclarationTextEditor.prototype._handleTabKey):
(WebInspector.CSSStyleDeclarationTextEditor.prototype._handleTabKey.highlightNextNameOrValue): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186212 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, rebaseline js/dom/global-constructors-attributes.html after r186198.

* platform/mac-mavericks/js/dom/global-constructors-attributes-expected.txt:
* platform/mac-yosemite/js/dom/global-constructors-attributes-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186211 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Win] REGRESSION (r185124) CACFLayer handling broken
https://bugs.webkit.org/show_bug.cgi?id=146530
<rdar://problem/21642241>

Reviewed by Tim Horton.

At some point in the past, the set of LayerChange flags
overflowed the default MSVC enum type of 'int'. This caused
Windows rendering code to not receive various update notifications.

Zalan's change moved the DebugIndicatorsChanged flag into that
overflow set, which made it obvious that something was wrong.

The fix is to tell the compiler to use a compatible base type
for the enum.

* platform/graphics/ca/GraphicsLayerCA.h: Prevent overflow.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186210 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Reduce resolution of performance.now.
https://bugs.webkit.org/show_bug.cgi?id=146531
rdar://problem/20116796

Reviewed by Simon Fraser.

Source/WebCore:

Test: http/tests/misc/webtiming-resolution.html

* page/Performance.cpp:
(WebCore::Performance::now):
Floor the time returned by performance.now to the nearest 5 microseconds.

LayoutTests:

* http/tests/misc/webtiming-resolution-expected.txt: Added.
* http/tests/misc/webtiming-resolution.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186208 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Ignore Visual Studio warning in SegmentedVector
https://bugs.webkit.org/show_bug.cgi?id=146514

Patch by Alex Christensen <achristensen@webkit.org> on 2015-07-01
Reviewed by Andreas Kling.

* wtf/SegmentedVector.h:
(WTF::SegmentedVector::deleteAllSegments):
Use pragmas to ignore the 0 element array warning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186207 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[iOS] Build fix

Patch by Eric Carlson <eric.carlson@apple.com> on 2015-07-01
Reviewed by Brent Fulgham.

* platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
(-[WebAVVideoLayer enterPIPModeRedirectingVideoToLayer:]): Renamed from enterOptimizedFullScreenModeRedirectingVideoToLayer.
(-[WebAVVideoLayer leavePIPMode]): Renamed from leaveOptimizedFullScreenMode.
(-[WebAVVideoLayer enterOptimizedFullScreenModeRedirectingVideoToLayer:]): Deleted.
(-[WebAVVideoLayer leaveOptimizedFullScreenMode]): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186206 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[iOS] Support bold and thin italicized system fonts
https://bugs.webkit.org/show_bug.cgi?id=146463
<rdar://problem/20948885>

Reviewed by Darin Adler.

Source/WebCore:

Add the italicized attribute to font descriptors.

Test: fast/text/weighted-italicized-system-font.html

* platform/graphics/ios/FontCacheIOS.mm:
(WebCore::createCTFontWithFamilyNameAndWeight):
* platform/graphics/mac/FontCacheMac.mm:
(WebCore::fontWithFamilySpecialCase):
(WebCore::fontWithFamily):
* platform/spi/cocoa/CoreTextSPI.h:

LayoutTests:

* fast/text/weighted-italicized-system-font-expected.html: Added.
* fast/text/weighted-italicized-system-font.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186205 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Remove code to recompute rounded window corners
https://bugs.webkit.org/show_bug.cgi?id=146534

Reviewed by Simon Fraser.

This code is no longer needed since we're always layer-backed now.

* UIProcess/API/mac/WKView.mm:
(-[WKView displayIfNeeded]): Deleted.
(-[WKView _cacheWindowBottomCornerRect]): Deleted.
* UIProcess/API/mac/WKViewInternal.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186204 268f45cc-cd09-0410-ab3c-d52691b4dbfc

bmalloc: realloc of an XLarge range can unmap adjacent VM ranges
https://bugs.webkit.org/show_bug.cgi?id=146535

Reviewed by Anders Carlsson.

This bug causes a crash when running fast/css/large-list-of-rules-crash.html
with the fix applied for https://bugs.webkit.org/show_bug.cgi?id=146519.

* bmalloc/Allocator.cpp:
(bmalloc::Allocator::reallocate): Start at object + newSize since starting
at object + oldSize means deleting the adjacent VM range.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186203 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, rolling out r185889
https://bugs.webkit.org/show_bug.cgi?id=146528
rdar://problem/21573959

Patch breaks chromeexperiments.com

Reverted changeset:

Source/JavaScriptCore:

* CMakeLists.txt:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj:
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj.filters:
* JavaScriptCore.xcodeproj/project.pbxproj:
* inspector/InjectedScriptSource.js:
(.):
* runtime/JSBoundSlotBaseFunction.cpp: Removed.
* runtime/JSBoundSlotBaseFunction.h: Removed.
* runtime/JSGlobalObject.cpp:
(JSC::JSGlobalObject::init): Deleted.
(JSC::JSGlobalObject::visitChildren): Deleted.
* runtime/JSGlobalObject.h:
(JSC::JSGlobalObject::boundSlotBaseFunctionStructure): Deleted.
* runtime/JSObject.cpp:
(JSC::JSObject::getOwnPropertyDescriptor):
(JSC::getBoundSlotBaseFunctionForGetterSetter): Deleted.
* runtime/VM.cpp:
(JSC::VM::VM): Deleted.
* runtime/VM.h:

LayoutTests:

* inspector-protocol/runtime/getProperties-expected.txt:
* js/dom/native-bindings-descriptors-expected.txt: Removed.
* js/dom/native-bindings-descriptors.html: Removed.
* js/dom/script-tests/native-bindings-descriptors.js: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186202 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix ANGLE Windows build after r186169.
https://bugs.webkit.org/show_bug.cgi?id=146532

Reviewed by Brent Fulgham.

Source/ThirdParty/ANGLE:

* ANGLE.vcxproj/libEGL.vcxproj:
* ANGLE.vcxproj/libEGL.vcxproj.filters:
* ANGLE.vcxproj/libGLESv2.vcxproj:
* ANGLE.vcxproj/libGLESv2.vcxproj.filters:
* ANGLE.vcxproj/libGLESv2Common.props:
* ANGLE.vcxproj/translator_common.vcxproj:
* ANGLE.vcxproj/translator_common.vcxproj.filters:
* ANGLE.vcxproj/translator_glsl.vcxproj:
* ANGLE.vcxproj/translator_glsl.vcxproj.filters:
* ANGLE.vcxproj/translator_hlsl.vcxproj:
* ANGLE.vcxproj/translator_hlsl.vcxproj.filters:
Update Windows build to build new files.

Source/WebCore:

* WebCore.vcxproj/WebCore.vcxproj:
* WebCore.vcxproj/WebCore.vcxproj.filters:
Don't compile OpenGLShims.cpp or Extensions3DOpenGL.cpp on Windows
because Windows uses OpenGLES through ANGLE, not OpenGL.
* platform/graphics/ANGLEWebKitBridge.h:
Use OpenGLESShims.h on Windows instead of OpenGLShims.h.
* platform/graphics/OpenGLESShims.h:
Added needed definitions from OpenGL to compile successfully.
* platform/graphics/win/GL/glext.h: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186201 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web page doesn't update its loading state when web process becomes suspended if there are pending network requests (XHR).
https://bugs.webkit.org/show_bug.cgi?id=146439

When web page is loading, we hold a background activity token in NavigationState and we release the token when the page done
loading. A web page can start loading subresources (like XHR) after the page is loaded, and WebPageProxy will update its
_networkRequestsInProgress state which will notify the client there is network actvity on-going. Since we don't hold
the background activity token after page is loaded, the WebContent process could become suspended without notifying the client
and the client will lose track of the loading state.

The first thought was to hold another background activity when _networkRequestsInProgress flag is set in NavigationState.
However, this could result in side-effect that a WebContent process can stay active for too long if there is a long lasting
XHR resquest.

This patch fix this by notifying the client that network activity is stopped when WebContent process is going to be suspended.
We also cache the networkRequestsInProgress state for suspended process, so that when it becomes foreground, we can tell
the client the correct state of network activity.

Patch by Yongjun Zhang <yongjun_zhang@apple.com> on 2015-07-01
Reviewed by Dan Bernstein.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::WebPageProxy): Initialize m_hasNetworkRequestsOnSuspended.
(WebKit::WebPageProxy::processWillBecomeSuspended): If there is pending requests on suspend, cache the network activity state
    and notify the client.
(WebKit::WebPageProxy::processWillBecomeForground): Restore the network activity state when the process becomes foreground.
(WebKit::WebPageProxy::resetState):
* UIProcess/WebPageProxy.h: Add a data member m_hasNetworkRequestsOnSuspended to cache the network activity state on process
    suspend.
* UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::didSetAssertionState):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186200 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Turn off mediastreamaudiosourcenode.html test
https://bugs.webkit.org/show_bug.cgi?id=146527
<rdar://problem/21641223>

Patch by Matthew Daiter <mdaiter@apple.com> on 2015-07-01
Reviewed by Simon Fraser.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186199 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Disable the experimental WebGL2 implementation
https://bugs.webkit.org/show_bug.cgi?id=146526
<rdar://problem/21641235>

Reviewed by Myles Maxfield.

Source/JavaScriptCore:

Add (and disable) an ENABLE_WEBGL2 flag.

* Configurations/FeatureDefines.xcconfig:

Source/WebCore:

Add (and disable) an ENABLE_WEBGL2 flag. Also protect
anything that is specific to WebGL2.

Covered by running the WebGL 1.0.2 conformance suite
and our LayoutTests.

* Configurations/FeatureDefines.xcconfig:
* bindings/js/JSCanvasRenderingContextCustom.cpp:
(WebCore::toJS):
* bindings/js/JSWebGL2RenderingContextCustom.cpp:
* bindings/js/JSWebGLRenderingContextBaseCustom.cpp:
(WebCore::toJS):
* html/HTMLCanvasElement.cpp:
(WebCore::HTMLCanvasElement::is3dType):
* html/canvas/WebGL2RenderingContext.cpp:
* html/canvas/WebGL2RenderingContext.h:
* html/canvas/WebGL2RenderingContext.idl:
* html/canvas/WebGLFramebuffer.cpp:
(WebCore::WebGLFramebuffer::drawBuffersIfNecessary):
* html/canvas/WebGLGetInfo.cpp:
* html/canvas/WebGLGetInfo.h:
* html/canvas/WebGLRenderingContextBase.cpp:
(WebCore::WebGLRenderingContextBase::create):
* html/canvas/WebGLVertexArrayObject.cpp:
* html/canvas/WebGLVertexArrayObject.h:
* html/canvas/WebGLVertexArrayObject.idl:

Source/WebKit/mac:

Add (and disable) an ENABLE_WEBGL2 flag.

* Configurations/FeatureDefines.xcconfig:

Source/WebKit2:

Add (and disable) an ENABLE_WEBGL2 flag.

* Configurations/FeatureDefines.xcconfig:

Source/WTF:

Add (and disable) an ENABLE_WEBGL2 flag.

* wtf/FeatureDefines.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186198 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix missing space.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186197 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Never try to pass URLs to LaunchServices when they're the same domain as the current URL.
<rdar://problem/20731429> and https://bugs.webkit.org/show_bug.cgi?id=146521

Reviewed by Alex Christensen.

* UIProcess/Cocoa/NavigationState.mm:
(WebKit::tryAppLink): Compare the proposed URL to the current URL to see if they’re in the same domain.
(WebKit::NavigationState::NavigationClient::decidePolicyForNavigationAction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186196 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Regression(183998): Disqus comments take a very long time to load
https://bugs.webkit.org/show_bug.cgi?id=146522
<rdar://problem/21590601>

Reviewed by Simon Fraser.

Source/WebCore:

Stop throttling requestAnimationFrame() in iframes that are not visible
due to them being zero-sized or display:none. Those are usually utility
iframes and throttling them is risky.

Se still throttle requestAnimationFrame() in iframes that would be
visible in theory but are currently not noticeable because they are
outside the viewport.

Test:
- fast/animation/request-animation-frame-throttle-subframe.html
- fast/animation/request-animation-frame-throttle-subframe-display-none.html
- fast/animation/request-animation-frame-throttle-subframe-zero-size.html

* page/FrameView.cpp:
(WebCore::FrameView::updateScriptedAnimationsAndTimersThrottlingState):

LayoutTests:

* fast/animation/request-animation-frame-throttle-subframe-display-none-expected.txt:
* fast/animation/request-animation-frame-throttle-subframe-display-none.html:
Update layout test as we changed behavior here. We no longer throttle display:none
iframes.

* fast/animation/request-animation-frame-throttle-subframe-zero-size-expected.txt: Added.
* fast/animation/request-animation-frame-throttle-subframe-zero-size.html: Added.
Add layout test to check that we don't throttle RaF in iframes that are zero-sized.

* fast/animation/resources/requestAnimationFrame-frame-2.html:
* fast/animation/resources/requestAnimationFrame-frame.html:
Update use of requestAnimationFrame() to be more representative of real world usage.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186195 268f45cc-cd09-0410-ab3c-d52691b4dbfc

WKWebView snapshots have the wrong scale after rotation
https://bugs.webkit.org/show_bug.cgi?id=146476
<rdar://problem/18345247>

Reviewed by Anders Carlsson.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _didCommitLayerTree:]):
"Coordinate" -> "Coordinates".

(-[WKWebView _snapshotRect:intoImageOfWidth:completionHandler:]):
Defer any snapshotting that happens during a resize until after
the resize completes. This will ensure that (in the case of an IOSurface
snapshot) the tiles are up to date, and (in the case of a software snapshot)
that our understanding of the scale and scroll offset of the WKContentView
are up to date, so that we can correctly convert from view to content coordinates
in the UI process.

(-[WKWebView _endAnimatedResize]):
Perform the deferred snapshotting after the resize completes.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186194 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix the iOS build.

* WebCoreSupport/WebUserMediaClient.mm:
(-[WebUserMediaPolicyListener denyOnlyThisRequest]): Remove ASSERT_NOT_REACHED,
because it caused the compiler to warn about noreturn.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186193 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Mark compositing/masks/compositing-clip-path-origin.html as being image-flakey.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186192 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (r179168): Characters overlap after resizing the font on the copy-pasted Japanese text
https://bugs.webkit.org/show_bug.cgi?id=146492

Reviewed by Darin Adler.

Source/WebCore:

The bug was caused by WebKit serializing the used line-height size (e.g. 18px) in the copied content
instead of string "normal" and removeStyleFromRulesAndContext failing to strip it down when text with
a font that influences the line height got pasted. This is because the used value of line-height
property of the context and the pasted content doesn't match when the context doesn't use the same font.

Fixed the bug by not considering line-height as a list of editing properties we try to preserve. This is
fine because we don't provide editing operations to directly manipulate line-height.

Test: editing/pasteboard/cjk-line-height.html

* editing/EditingStyle.cpp:
(WebCore::editingProperties): Removed CSSPropertyLineHeight.

LayoutTests:

Added a regression test. Also reverted the bad rebaseline in r179168:
http://trac.webkit.org/changeset/179168/trunk/LayoutTests/editing/pasteboard/simplfiying-markup-should-not-strip-content-expected.txt

* editing/pasteboard/cjk-line-height-expected.txt: Added.
* editing/pasteboard/cjk-line-height.html: Added.
* editing/pasteboard/simplfiying-markup-should-not-strip-content-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186191 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (r184296): View keeps scrolling upward
https://bugs.webkit.org/show_bug.cgi?id=146497
<rdar://problem/21524942>

Reviewed by Darin Adler.

Avoid improperly triggering the ScrollController wheel event handling
logic when the wheel event deltaX/deltaY are zero. On certain sites,
this caused a programmatic JavaScript scroll to be triggered unexpectedly,
scrolling the page back to some initial state.

This bug was introduced while trying to make sure scrollbars were notified
when the wheel event had come to an end. Revise that change so that we still
follow the right code path for non-stretchable regions. However, make sure
that for zero-delta wheel events we make sure to properly handle the wheel
event phase.

* platform/mac/ScrollAnimatorMac.h:
* platform/mac/ScrollAnimatorMac.mm:
(WebCore::ScrollAnimatorMac::shouldForwardWheelEventsToParent): New helper function
to reduce the complexity of the logic in handleWheelEvent.
(WebCore::ScrollAnimatorMac::handleWheelEvent): When wheel events should be forwarded
to the parent scroll view, if the event was handled or has no change in position
trigger the 'handleWheelEventPhase' logic so that scrollbars are hidden, etc.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186185 268f45cc-cd09-0410-ab3c-d52691b4dbfc

This is a quick follow-on to trac.webkit.org/changeset/186132

Rubber-stamped by Darin Adler.

Make this a one-liner.
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::getPositionInformation):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186184 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Hide the inline controls when going into PiP
https://bugs.webkit.org/show_bug.cgi?id=146487
<rdar://problem/19881159>

Reviewed by Eric Carlson.

When the presentation mode is PiP, toggle a class
on the container element so that we can completely
hide the control toolbar. The placard should still
remain visible.

* Modules/mediacontrols/mediaControlsiOS.css:
(video::-webkit-media-controls-panel-container.picture-in-picture): Add
a rule that hides the toolbar when PiP is active.
* Modules/mediacontrols/mediaControlsiOS.js:
(ControllerIOS.prototype.showControls):
(ControllerIOS.prototype.handlePresentationModeChange): Toggle a
class. Also fix a bug I noticed where exiting from PiP was
not auto-hiding the controls until the user tapped.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186183 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Source/JavaScriptCore:
Enable MEDIA_STREAM flag
https://bugs.webkit.org/show_bug.cgi?id=145947
<rdar://problem/21365829>

Patch by Matthew Daiter <mdaiter@apple.com> on 2015-07-01
Reviewed by Eric Carlson.

* Configurations/FeatureDefines.xcconfig: Added MEDIA_STREAM flag

Source/WebCore:
Enable MEDIA_STREAM flag
https://bugs.webkit.org/show_bug.cgi?id=145947
<rdar://problem/21365829>

Patch by Matthew Daiter <mdaiter@apple.com> on 2015-07-01
Reviewed by Eric Carlson.

* Configurations/FeatureDefines.xcconfig: Added MEDIA_STREAM flag
* Modules/mediastream/MediaStreamTrack.h:
* WebCore.xcodeproj/project.pbxproj: Changed project headers to
private headers
* platform/graphics/avfoundation/MediaPlayerPrivateAVFoundation.h:
* platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.h:
* platform/graphics/mac/MediaPlayerPrivateQTKit.h:
* platform/mediastream/RealtimeMediaSource.h:
* platform/mediastream/mac/AVAudioCaptureSource.mm:
* platform/mediastream/mac/AVCaptureDeviceManager.mm:
(WebCore::refreshCaptureDeviceList): Deleted.
* platform/mediastream/mac/AVMediaCaptureSource.mm:
(WebCore::AVMediaCaptureSource::startProducingData):
(WebCore::AVMediaCaptureSource::stopProducingData):

Source/WebKit/mac:
Enable MEDIA_STREAM flag
https://bugs.webkit.org/show_bug.cgi?id=145947
<rdar://problem/21365829>

Patch by Matthew Daiter <mdaiter@apple.com> on 2015-07-01
Reviewed by Eric Carlson.

* Configurations/FeatureDefines.xcconfig: Added MEDIA_STREAM flag
* WebCoreSupport/WebUserMediaClient.h: Changed signatures
* WebCoreSupport/WebUserMediaClient.mm: Changed signatures
(WebUserMediaClient::requestPermission):
(WebUserMediaClient::cancelRequest):
(-[WebUserMediaPolicyListener denyOnlyThisRequest]):
(-[WebUserMediaPolicyListener shouldClearCache]):

Source/WebKit2:
Enable MEDIA_STREAM flag
https://bugs.webkit.org/show_bug.cgi?id=145947
<rdar://problem/21365829>

Patch by Matthew Daiter <mdaiter@apple.com> on 2015-07-01
Reviewed by Eric Carlson.

* Configurations/FeatureDefines.xcconfig: Add MEDIA_STREAM flag
* WebKit2.xcodeproj/project.pbxproj: Changed some project to private
declarations of headers
* WebProcess/WebCoreSupport/WebUserMediaClient.cpp:
* WebProcess/WebPage/WebPage.h:
(WebKit::WebPage::userMediaPermissionRequestManager):

Tools:
Enabled MEDIA_STREAM flag
https://bugs.webkit.org/show_bug.cgi?id=145947
<rdar://problem/21365829>

Patch by Matthew Daiter <mdaiter@apple.com> on 2015-07-01
Reviewed by Eric Carlson.

* TestWebKitAPI/Tests/WebKit2/UserMedia.cpp:

LayoutTests:
Enable MEDIA_STREAM flag
https://bugs.webkit.org/show_bug.cgi?id=145947
<rdar://problem/21365829>

Patch by Matthew Daiter <mdaiter@apple.com> on 2015-07-01
Reviewed by Eric Carlson.

*
platform/mac-mavericks/js/dom/global-constructors-attributes-expected.txt:
Added attributes to tests
*
platform/mac-yosemite/js/dom/global-constructors-attributes-expected.txt:
Ditto
* platform/mac/js/dom/global-constructors-attributes-expected.txt:
Ditto

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186182 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (r185016): Intermittent crash in WebCore::TextTrackList::remove
https://bugs.webkit.org/show_bug.cgi?id=146493
<rdar://problem/21511122>

Reviewed by Eric Carlson.

The m_textTracks member is frequently null checked during other operations, but
was not checked during track removal. This needs to be corrected.

* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::removeTextTrack): Check that m_textTracks is
not null before using it during track removal.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186181 268f45cc-cd09-0410-ab3c-d52691b4dbfc

PNG mask images are loaded with Accept:image/svg+xml
https://bugs.webkit.org/show_bug.cgi?id=146509
Source/WebCore:

rdar://problem/21584740

Reviewed by Simon Fraser.

For some strange reason MaskImageOperation code loads all mask images, including non-SVG ones
using CachedSVGDocument. Resulting bad accept header may cause server to reject the request.

This is far from ideal but as a quick fix we can override the accept header for mask images to
allow any image type.

Test: http/tests/misc/mask-image-accept.html

* loader/cache/CachedResourceLoader.cpp:
(WebCore::CachedResourceLoader::requestResource):
* loader/cache/CachedResourceRequest.h:
(WebCore::CachedResourceRequest::acceptOverride):
(WebCore::CachedResourceRequest::setAcceptOverride):
* loader/cache/CachedSVGDocumentReference.cpp:
(WebCore::CachedSVGDocumentReference::load):
* loader/cache/CachedSVGDocumentReference.h:
(WebCore::CachedSVGDocumentReference::loadRequested):
(WebCore::CachedSVGDocumentReference::setAcceptsAnyImageType):
(WebCore::CachedSVGDocumentReference::document):
* platform/graphics/MaskImageOperation.cpp:
(WebCore::MaskImageOperation::ensureCachedSVGDocumentReference):

LayoutTests:

Reviewed by Simon Fraser.

* http/tests/misc/mask-image-accept-expected.html: Added.
* http/tests/misc/mask-image-accept.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186180 268f45cc-cd09-0410-ab3c-d52691b4dbfc

REGRESSION (185319): Reproducible crash in WebHistoryItem launching FluidApp.
<rdar://problem/21598293> and https://bugs.webkit.org/show_bug.cgi?id=146494

Reviewed by Darin Adler.

* History/WebHistoryItem.mm:
(-[WebHistoryItem initFromDictionaryRepresentation:]): Don’t reference members of a Vector
  that don’t exist.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186179 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Check shouldOpenExternalURLs instead of isProcessingUserGesture before passing URL to LaunchServices.
<rdar://problem/20577859> and https://bugs.webkit.org/show_bug.cgi?id=146513

Reviewed by Darin Adler.

* UIProcess/Cocoa/NavigationState.mm:
(WebKit::tryAppLink): The expected value of isProcessingUserGesture is lost over redirects.
  shouldOpenExternalURLs actually fits this role perfectly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186178 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[MSE] Failures on W3C media-source tests regarding MIME types
https://bugs.webkit.org/show_bug.cgi?id=146499

Reviewed by Eric Carlson.

Source/WebCore:

Tests: http/tests/media/media-source/SourceBuffer-abort-readyState.html
       http/tests/media/media-source/SourceBuffer-abort-removed.html
       http/tests/media/media-source/SourceBuffer-abort-updating.html
       http/tests/media/media-source/SourceBuffer-abort.html

Multiple failures in the W3C media-source test suite due to two failures in isTypeSupported MIME type handling:
- MIME types without codec strings were being rejected.
- MIME types with codec strings which are rejected by the system are being reported as supported.

For the first, stop rejecting MIME types with non-existent codec strings. For MIME types which do have non-empty
codec strings, treat a "Maybe" response as unsupported.

* Modules/mediasource/MediaSource.cpp:
(WebCore::MediaSource::isTypeSupported):

LayoutTests:

Import four new tests from the w3c test suite <https://github.com/w3c/web-platform-tests.git>.

* http/tests/media/media-source/SourceBuffer-abort-readyState.html: Added.
* http/tests/media/media-source/SourceBuffer-abort-removed.html: Added.
* http/tests/media/media-source/SourceBuffer-abort-updating.html: Added.
* http/tests/media/media-source/SourceBuffer-abort.html: Added.
* http/tests/media/white.mp4: Added.
* http/tests/media/white.webm: Added.
* platform/mac/TestExpectations:
* platform/mac/http/tests/media/media-source/SourceBuffer-abort-expected.txt: Added.
* platform/mac/http/tests/media/media-source/SourceBuffer-abort-readyState-expected.txt: Added.
* platform/mac/http/tests/media/media-source/SourceBuffer-abort-removed-expected.txt: Added.
* platform/mac/http/tests/media/media-source/SourceBuffer-abort-updating-expected.txt: Added.
* platform/mac/http/tests/media/media-source/mediasource-is-type-supported-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186177 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] [Wayland] Build broken with ENABLE_WAYLAND_TARGET=ON
https://bugs.webkit.org/show_bug.cgi?id=146512

Unreviewed, fix build with ENABLE_WAYLAND_TARGET=ON after r185949.

* UIProcess/API/gtk/WebKitWebViewBase.cpp:
(webkitWebViewBaseCreateWebPage):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186176 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] REGRESSION(r183936): Test /webkit2/WebKitWebContext/spell-checker fails since r183936
https://bugs.webkit.org/show_bug.cgi?id=144828

Reviewed by Carlos Garcia Campos.

Source/WebKit2:

Instead of setting spell checking languages unconditionally when building int
developer mode, add some C API that allows setting them in WebKitTestRunner.

* PlatformGTK.cmake: Add the new C API file to the source list.
* UIProcess/API/C/gtk/WKTextCheckerGtk.cpp:
(WKTextCheckerSetSpellCheckingLanguages): Added this function which allows settings
the spell checking languages from the C API.
* UIProcess/API/C/gtk/WKTextCheckerGtk.h: Added.
* UIProcess/gtk/TextCheckerGtk.cpp:
(WebKit::enchantTextChecker): Remove the hack which sets the spell checking languages automatically
for development builds.

Tools:

* WebKitTestRunner/gtk/main.cpp:
(main): Use the new WKTextChecker API to set the spell checking languages
when the process is starting up.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186175 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Mac] Numerous CGColor leaks in swipe gestures
https://bugs.webkit.org/show_bug.cgi?id=146501

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-07-01
Reviewed by Tim Horton.

* UIProcess/mac/ViewGestureControllerMac.mm:
(WebKit::ViewGestureController::beginSwipeGesture):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186174 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Mac] Confusing RetainPtr use in TiledCoreAnimationDrawingArea
https://bugs.webkit.org/show_bug.cgi?id=146503

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-07-01
Reviewed by Tim Horton.

* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::commitTransientZoom):
Simplify adoption into a new RetainPtr.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186173 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[Win] Unreviewed build fix after r186169.

Source/WebKit:

* WebKit.vcxproj/WebKit.sln:
Temporarily don't build ANGLE projects on Windows.

Source/WTF:

* wtf/FeatureDefines.h:
Temporarily disable WebGL on WinCairo. It was already temporarily disabled on AppleWin.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186172 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186171 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Update ANGLE to b11e2483742db884bd0af41f78f528240577356b.
https://bugs.webkit.org/show_bug.cgi?id=145010

Reviewed by Dean Jackson.

Source/ThirdParty/ANGLE:

Huge list of source changes omitted.
ROLL DEPS!

Source/WebCore:

* CMakeLists.txt:
* platform/graphics/ANGLEWebKitBridge.cpp:
(WebCore::appendSymbol):
(WebCore::getStructInfo):
(WebCore::getSymbolInfo):
(WebCore::ANGLEWebKitBridge::cleanupCompilers):
(WebCore::ANGLEWebKitBridge::compileShaderSource):
(WebCore::getValidationResultValue): Deleted.
* platform/graphics/ANGLEWebKitBridge.h:
(WebCore::ANGLEShaderSymbol::isSampler):
* platform/graphics/GraphicsContext3D.h:
(WebCore::GraphicsContext3D::SymbolInfo::SymbolInfo):
* platform/graphics/mac/GraphicsContext3DMac.mm:
* platform/graphics/opengl/Extensions3DOpenGL.cpp:
* platform/graphics/opengl/GraphicsContext3DOpenGLCommon.cpp:
(WebCore::GraphicsContext3D::checkVaryingsPacking):
(WebCore::GraphicsContext3D::precisionsMatch):
Update WebCore to reflect changes in ANGLE.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186169 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[EFL] Unreviewed gardening.

Some tests are flaky or timeout after bumping Ubuntu version from 14.04 to 15.04.
Mark the tests to flaky or timeout.

* platform/efl/TestExpectations:
* platform/efl/compositing/tiling/rotated-tiled-clamped-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186167 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, skip this test for now. We can unskip it when this feature is enabled.

* js/script-tests/intl.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186166 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Frame flattening: Hit-testing an iframe could end up destroying the associated inline tree context.
https://bugs.webkit.org/show_bug.cgi?id=146447
rdar://problem/20613501

Reviewed by Simon Fraser.

This patch ensures that the render tree associated with the document on which
the hit-test is initiated does not get laid out, unless it was directly mutated prior to the hittest.

Hit-test requirements:
1. A clean the render tree before hit-testing gets propagated to the renderers.
Document::updateLayout() ensures it by calling both updateStyleIfNeeded() and layout() not only on the current tree, but also
on the ancestors if needed.

2. No render tree mutation while hit-testing the renderers.

When an iframe is being hit-tested, this hit-test could bubble down to the child frame's render view.
In order to ensure #1, we call Document::updateLayout() on the current (subframe) document.
If updateStyleIfNeeded() mutates the render tree, we mark it dirty for layout(). However frame flattening also
marks the parent renderer (RenderIFrame) dirty.
While calling layout() to clean the current render tree, we end up laying out the parent tree too.
Laying out the parent tree could end up destroying the inline tree context from where the
hittest just bubbled down. (InlineFlowBox -> RenderWidget -> RenderView).

This patch protects the render tree from such unintentional inline tree mutation during hittesting.
After the initial layout we set a layout disallow flag on the frame view to defer subsequent layouts.
This patch only changes behavior when frame flattening is enabled, but in future we may always want to enable this.

Source/WebCore:

Test: fast/frames/flattening/hittest-iframe-while-style-changes-crash.html

* page/FrameView.cpp:
(WebCore::FrameView::layout):
(WebCore::FrameView::startLayoutAtMainFrameViewIfNeeded): Deleted. -> Assertion in no longer valid.
* page/FrameView.h:
* rendering/RenderView.cpp:
(WebCore::FrameFlatteningLayoutDisallower::FrameFlatteningLayoutDisallower):
(WebCore::FrameFlatteningLayoutDisallower::~FrameFlatteningLayoutDisallower):
(WebCore::RenderView::hitTest): Protect the render tree from subsequent layouts.

LayoutTests:

* fast/frames/flattening/hittest-iframe-while-style-changes-crash-expected.txt: Added.
* fast/frames/flattening/hittest-iframe-while-style-changes-crash.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186165 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix CoordinatedGraphics.

* WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.cpp:
(WebKit::CoordinatedLayerTreeHost::didFlushRootLayer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186164 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: add " = $0" hint after selected element in main DOMTreeOutline
https://bugs.webkit.org/show_bug.cgi?id=145739

Patch by Devin Rousso <drousso@apple.com> on 2015-06-30
Reviewed by Timothy Hatcher.

* UserInterface/Views/DOMTreeOutline.css: Adds a semi-opaque " = $0" to the selected element in the DOM tree.
(.dom-tree-outline li.selected > span::after):
(.dom-tree-outline:focus li.selected > span::after):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186163 268f45cc-cd09-0410-ab3c-d52691b4dbfc

More Gtk/EFL fixing.

* WebProcess/WebPage/CoordinatedGraphics/CoordinatedLayerTreeHost.cpp:
(WebKit::CoordinatedLayerTreeHost::didFlushRootLayer):
* WebProcess/WebPage/gtk/LayerTreeHostGtk.cpp:
(WebKit::LayerTreeHostGtk::flushPendingLayerChanges):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186162 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Implement ECMAScript Internationalization API
https://bugs.webkit.org/show_bug.cgi?id=90906

Patch by Andy VanWagoner <thetalecrafter@gmail.com> on 2015-06-30
Reviewed by Benjamin Poulain.

.:

Begin implementing the Intl apis behind ENABLE_INTL flag.
Create the base Intl namespace object.

* Source/cmake/WebKitFeatures.cmake: add ENABLE_INTL flag

Source/JavaScriptCore:

* CMakeLists.txt: add IntlObject.cpp
* Configurations/FeatureDefines.xcconfig: add ENABLE_INTL flag
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj: add IntlObject
* JavaScriptCore.vcxproj/JavaScriptCore.vcxproj.filters: add IntlObject
* JavaScriptCore.xcodeproj/project.pbxproj: add IntlObject
* runtime/CommonIdentifiers.h: add "Intl" name
* runtime/IntlObject.cpp: Added.
(JSC::IntlObject::IntlObject):
(JSC::IntlObject::create):
(JSC::IntlObject::finishCreation):
(JSC::IntlObject::createStructure):
* runtime/IntlObject.h: Added.
* runtime/JSGlobalObject.cpp: Add global Intl
(JSC::JSGlobalObject::init):

Source/WebCore:

Test: js/intl.html

* Configurations/FeatureDefines.xcconfig: add ENABLE_INTL flag

Source/WebKit/mac:

* Configurations/FeatureDefines.xcconfig: add ENABLE_INTL flag

Source/WebKit2:

* Configurations/FeatureDefines.xcconfig: add ENABLE_INTL flag

Source/WTF:

* wtf/FeatureDefines.h: add ENABLE_INTL flag

Tools:

* Scripts/webkitperl/FeatureList.pm: add ENABLE_INTL flag

WebKitLibraries:

* win/tools/vsprops/FeatureDefines.props: add ENABLE_INTL flag
* win/tools/vsprops/FeatureDefinesCairo.props: add ENABLE_INTL flag

LayoutTests:

* TestExpectations: Skip intl until enabled
* js/intl-expected.txt: Added.
* js/intl.html: Added.
* js/script-tests/intl.js: Added.
(global):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186161 268f45cc-cd09-0410-ab3c-d52691b4dbfc

scroll-snap-points do not work very well with mechanical scroll wheel events
https://bugs.webkit.org/show_bug.cgi?id=142501
<rdar://problem/20093511>

Patch by Wenson Hsieh <whsieh@berkeley.edu> on 2015-06-30
Reviewed by Brent Fulgham.

Source/WebCore:

Stateless scroll events generated by scrolling with a mechanical mouse wheel now trigger scroll
snapping after a fixed delay.

Test: platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-overflow-stateless.html

* platform/cocoa/ScrollController.h: Added member to track stateless scrolling.
* platform/cocoa/ScrollController.mm: Added constant for delay for stateless scroll snapping.
(WebCore::ScrollController::processWheelEventForScrollSnapOnAxis): Triggers timer upon stateless scroll events.
(WebCore::ScrollController::horizontalScrollSnapTimerFired): Timer now handles stateless scrolling updates.
(WebCore::ScrollController::verticalScrollSnapTimerFired): See above.
(WebCore::ScrollController::beginScrollSnapAnimation): Handles stateless scroll snapping.

LayoutTests:

* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-overflow-stateless-expected.txt: Added.
* platform/mac-wk2/tiled-drawing/scrolling/scroll-snap/scroll-snap-mandatory-overflow-stateless.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186160 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Fix Mac and Windows builds.

Source/WebKit/win:

* WebCoreSupport/AcceleratedCompositingContext.cpp:
(AcceleratedCompositingContext::flushPendingLayerChanges):
* WebView.cpp:
(WebView::flushPendingGraphicsLayerChanges):

Source/WebKit2:

* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::flushLayers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186159 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Try to fix Gtk and EFL builds.

* platform/graphics/texmap/GraphicsLayerTextureMapper.cpp:
(WebCore::GraphicsLayerTextureMapper::flushCompositingStateForThisLayerOnly):
(WebCore::GraphicsLayerTextureMapper::flushCompositingState):
* platform/graphics/texmap/GraphicsLayerTextureMapper.h:
* platform/graphics/texmap/coordinated/CompositingCoordinator.cpp:
(WebCore::CompositingCoordinator::flushPendingLayerChanges):
* platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.cpp:
(WebCore::CoordinatedGraphicsLayer::flushCompositingState):
(WebCore::CoordinatedGraphicsLayer::flushCompositingStateForThisLayerOnly):
* platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186156 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[iOS] Missing tiles inside position:fixed on scrolling
https://bugs.webkit.org/show_bug.cgi?id=146485
rdar://problem/21226861

Reviewed by Tim Horton.

Layer flushing adjusts the coverage rect for tiled layers, but does so at times
when position:fixed layers are moved around by the scrolling tree. The computed
coverage rect then doesn't reflect the layer's on-screen position, causing missing
tiles.

Fix by pushing the notion of being in a "stable state" onto FrameView, and passing
that state into the layer flush. When not in a stable state, flushing doesn't change
the visible and coverage rects for layers that are viewport-constrained.

Source/WebCore:

* page/FrameView.cpp:
(WebCore::FrameView::reset):
* page/FrameView.h: Remove some velocity-related data members that were unused.
* platform/graphics/GraphicsLayer.h:
(WebCore::GraphicsLayer::flushCompositingState):
(WebCore::GraphicsLayer::flushCompositingStateForThisLayerOnly):
* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::flushCompositingState):
(WebCore::GraphicsLayerCA::flushCompositingStateForThisLayerOnly):
(WebCore::GraphicsLayerCA::setVisibleAndCoverageRects): If this is a viewport-constrained
layer, and the viewport is not stable, don't touch the rects.
(WebCore::GraphicsLayerCA::recursiveCommitChanges):
* platform/graphics/ca/GraphicsLayerCA.h:
(WebCore::GraphicsLayerCA::CommitState::CommitState):
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::flushPendingLayerChanges):

Source/WebKit2:

* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::updateVisibleContentRects):
* WebProcess/WebPage/mac/RemoteLayerTreeDrawingArea.mm:
(WebKit::RemoteLayerTreeDrawingArea::flushLayers):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186155 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rename GraphicsLayer's allowsBackingStoreDetachment to isViewportConstrained
https://bugs.webkit.org/show_bug.cgi?id=146483

Reviewed by Tim Horton.

What GraphicsLayer really needs to know is whether some other thread/process
is moving its platform layers around behind its back, and this is is better
expressed as "isViewportConstrained" rather than "allowsBackingStoreDetachment".

The sense of the flag is flipped, and boolean logic adjusted accordingly.

* platform/graphics/GraphicsLayer.h:
(WebCore::GraphicsLayer::setIsViewportConstrained):
(WebCore::GraphicsLayer::isViewportConstrained):
(WebCore::GraphicsLayer::setAllowsBackingStoreDetachment): Deleted.
(WebCore::GraphicsLayer::allowsBackingStoreDetachment): Deleted.
* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::GraphicsLayerCA):
(WebCore::GraphicsLayerCA::setVisibleAndCoverageRects):
(WebCore::GraphicsLayerCA::recursiveCommitChanges):
* platform/graphics/ca/GraphicsLayerCA.h:
* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::setIsScrollCoordinatedWithViewportConstrainedRole):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186154 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[EFL] Implement ewk_favicon_database_clear() API.
https://bugs.webkit.org/show_bug.cgi?id=146337

Patch by Hyungwook Lee <hyungwook.lee@navercorp.com> on 2015-06-30
Reviewed by Gyuyoung Kim.

We need to provide way to clear favicon data.

* UIProcess/API/efl/ewk_favicon_database.cpp:
(EwkFaviconDatabase::getIconSurfaceSynchronously):
(EwkFaviconDatabase::clearFaviconDatabase):
(EwkFaviconDatabase::iconDataReadyForPageURL):
(ewk_favicon_database_icon_change_callback_del):
(ewk_favicon_database_clear):
* UIProcess/API/efl/ewk_favicon_database.h:
* UIProcess/API/efl/ewk_favicon_database_private.h:
* UIProcess/API/efl/tests/test_ewk2_favicon_database.cpp:
(TEST_F):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186153 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[iOS] Ignore viewport rect changes during zoom bouncing.
<https://webkit.org/b/146480>

Reviewed by Benjamin Poulain.

When the user has pinch-zoomed past the maximum zoom scale and lets go of the screen,
we animate a bounce back to the maximum scale.

Suppress viewport rect updates during this bounce animation since we don't need them
anyway. This fixes an issue where we'd sometimes get a weird rect during the animation
and if could send that to the web process, causing it to generate tiles for a part of
the page that's not actually exposed.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _updateVisibleContentRects]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186152 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Errors in read() are not handled in WTF::cryptographicallyRandomValuesFromOS.
https://bugs.webkit.org/show_bug.cgi?id=146473

Patch by Keith Miller <keith_miller@apple.com> on 2015-06-30
Reviewed by Filip Pizlo.

We were not checking if errors occurred in WTF::cryptographicallyRandomValuesFromOS.
We now buffer the data until enough bits of entropy exist to fill the buffer
rather than crash. Additionally, added two crash functions so we can distinguish
between the two reasons why we crashed in traces.

* wtf/OSRandomSource.cpp:
(WTF::crashUnableToOpenFD):
(WTF::crashUnableToReadFromFD):
(WTF::cryptographicallyRandomValuesFromOS):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186151 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, rebaseline editing/style/unbold-in-bold.html on WK2.

* platform/mac-wk2/editing/style/unbold-in-bold-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186150 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, revert bad wk2 baseline done in r186106.

* platform/mac-wk2/editing/style: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186149 268f45cc-cd09-0410-ab3c-d52691b4dbfc

<rdar://problem/20655729> WebKit should heuristically exclude scrolling touch events from the user gesture category

Reviewed by Benjamin Poulain.

* Platform/spi/ios/UIKitSPI.h:
* Shared/WebEvent.h:
(WebKit::WebTouchEvent::WebTouchEvent):
(WebKit::WebTouchEvent::position):
(WebKit::WebTouchEvent::isPotentialTap):
(WebKit::WebTouchEvent::isGesture):
(WebKit::WebTouchEvent::gestureScale):
(WebKit::WebTouchEvent::gestureRotation):
* Shared/WebEventConversion.cpp:
(WebKit::WebKit2PlatformTouchEvent::WebKit2PlatformTouchEvent):
* Shared/ios/NativeWebTouchEventIOS.mm:
(WebKit::NativeWebTouchEvent::NativeWebTouchEvent):
* Shared/ios/WebTouchEventIOS.cpp:
(WebKit::WebTouchEvent::encode):
(WebKit::WebTouchEvent::decode):
Pipe isPotentialTap through.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186148 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Update the JetStream documentation to reflect the recent changes
https://bugs.webkit.org/show_bug.cgi?id=146474

Reviewed by Geoffrey Garen.

* JetStream/create.rb: Bump the version.
* JetStream/in-depth-TEMPLATE.html: Add cdjs documentation. Remove cordic documentation. Change documentation for splay and mandreel.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186142 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: ProfileNode cleanup
https://bugs.webkit.org/show_bug.cgi?id=146472

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-06-30
Reviewed by Timothy Hatcher.

* UserInterface/Models/ProfileNode.js:
(WebInspector.ProfileNode):
Use Array.prototype.every instead of reduce.
Correct the name of a function to be what callers expect!

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186141 268f45cc-cd09-0410-ab3c-d52691b4dbfc

[GTK] [EFL] Unreviewed build fix after r186118.

Add if PLATFORM(COCOA) back.

* WebProcess/WebProcess.cpp:
(WebKit::WebProcess::nonVisibleProcessCleanupTimerFired):
* WebProcess/WebProcess.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186140 268f45cc-cd09-0410-ab3c-d52691b4dbfc

CABackdropFilter should set windowServerAware to false
https://bugs.webkit.org/show_bug.cgi?id=146469
<rdar://problem/21618614>

Reviewed by Simon Fraser.

* platform/graphics/ca/cocoa/PlatformCALayerCocoa.mm:
(PlatformCALayerCocoa::PlatformCALayerCocoa): Set windowServerAware
to false on Mac (it's not available on iOS).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186139 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: iOS 8: Profile Nodes not showing in JavaScript Timeline
https://bugs.webkit.org/show_bug.cgi?id=146471

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2015-06-30
Reviewed by Brian Burg.

* UserInterface/Models/ScriptTimelineRecord.js:
(WebInspector.ScriptTimelineRecord.prototype._initializeProfileFromPayload.profileNodeCallFromPayload):
For legacy backends, convert the startTime to the proper timestamp type.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186138 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Rolling out r175171
https://bugs.webkit.org/show_bug.cgi?id=146470
<rdar://problem/21349934>

Unreviewed, rolling out <http://trac.webkit.org/changeset/175171>. It
wasn't a huge win and Brady says it caused crash at
<rdar://problem/21349934>.

* platform/network/cf/ResourceRequest.h:
* platform/network/cocoa/ResourceRequestCocoa.mm:
(WebCore::ResourceRequest::doUpdatePlatformRequest):
(WebCore::ResourceRequest::doUpdatePlatformHTTPBody):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186137 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Allow object allocation sinking through GetScope, GetExecutable and SkipScope nodes
https://bugs.webkit.org/show_bug.cgi?id=146431

Reviewed by Filip Pizlo.

* dfg/DFGNode.h:
(JSC::DFG::Node::isFunctionAllocation):
(JSC::DFG::Node::isPhantomFunctionAllocation):
* dfg/DFGObjectAllocationSinkingPhase.cpp:
(JSC::DFG::ObjectAllocationSinkingPhase::handleNode):
* dfg/DFGPromoteHeapAccess.h:
(JSC::DFG::promoteHeapAccess):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186136 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Unreviewed, revert bad WK1 rebaseline done in r186106.

* platform/mac/editing/style/unbold-in-bold-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186135 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Crash under _layoutForNodeHighlight: when inspecting espn.com
https://bugs.webkit.org/show_bug.cgi?id=146422

Reviewed by Joseph Pecoraro.
Source/WebKit/mac:

When inspecting espn.com, MobileSafari sometimes crashes, because
highlight.quads is empty. I was not able to quickly determine why.

Fix by bailing in this case.

* WebInspector/WebNodeHighlightView.mm:
(-[WebNodeHighlightView _layoutForNodeHighlight:parent:]):

Source/WebKit2:

When inspecting espn.com, MobileSafari sometimes crashes, because
highlight.quads is empty. I was not able to quickly determine why.

Fix by bailing in this case.

* UIProcess/WKInspectorHighlightView.mm:
(-[WKInspectorHighlightView _layoutForNodeHighlight:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186134 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Web Inspector: Reduce rendering frames "Other" time by instrumenting compositing
https://bugs.webkit.org/show_bug.cgi?id=146168

Reviewed by Brian Burg.

Source/JavaScriptCore:

* inspector/protocol/Timeline.json:
New timeline record type for compositing events.

Source/WebCore:

Added Inspector instrumentation for measuring CoreAnimation compositing time. We mark the start of a composite
event when the LayerFlushScheduler triggers a scheduled layer flush. InspectorController now exports a function
for marking the end of the composite event, which should be called during the CA transaction post-commit phase
(based on platform support). Lacking platform support, the event is considered complete after CoreAnimation
runloop observers have run.

* inspector/InspectorController.cpp:
(WebCore::InspectorController::didComposite):
* inspector/InspectorController.h:
New export for instrumentation in WebKit2.

* inspector/InspectorInstrumentation.cpp:
(WebCore::InspectorInstrumentation::willCompositeImpl):
(WebCore::InspectorInstrumentation::didCompositeImpl):
* inspector/InspectorInstrumentation.h:
(WebCore::InspectorInstrumentation::willComposite):
(WebCore::InspectorInstrumentation::didComposite):
Plumbing for new instrumentation.

* inspector/InspectorTimelineAgent.cpp:
(WebCore::InspectorTimelineAgent::internalStart):
(WebCore::InspectorTimelineAgent::internalStop):
(WebCore::InspectorTimelineAgent::willComposite):
(WebCore::InspectorTimelineAgent::didComposite):
(WebCore::toProtocol):
* inspector/InspectorTimelineAgent.h:
New Composite event type and instrumentation.

* page/FrameView.cpp:
(WebCore::FrameView::flushCompositingStateIncludingSubframes):
Hook for start of compositing.

* platform/spi/cocoa/QuartzCoreSPI.h:
New header include and interface declaration.

Source/WebInspectorUI:

* Localizations/en.lproj/localizedStrings.js:
* UserInterface/Controllers/TimelineManager.js:
(WebInspector.TimelineManager.prototype._processRecord):
Added handling for new Composite record type. Paint records with a parent Composite record
are flagged to simplify processing after the event hierarchy is unpacked.

* UserInterface/Images/TimelineRecordComposite.svg: Added.
New composite record icon.

* UserInterface/Models/LayoutTimelineRecord.js:
(WebInspector.LayoutTimelineRecord):
(WebInspector.LayoutTimelineRecord.displayNameForEventType):
(WebInspector.LayoutTimelineRecord.prototype.get duringComposite): Added
Composite record support.

* UserInterface/Models/RenderingFrameTimelineRecord.js:
(WebInspector.RenderingFrameTimelineRecord.prototype.durationForTask.get validRecordForTaskType):
Add compositing time when bucketing runloop tasks, ignoring Paint events that are
contained within a Composite event.

* UserInterface/Views/LayoutTimelineOverviewGraph.js:
(WebInspector.LayoutTimelineOverviewGraph.prototype._layoutTimelineRecordAdded):
Add Composite records to paint timeline row.

* UserInterface/Views/RenderingFrameTimelineView.js:
(WebInspector.RenderingFrameTimelineView.prototype._processPendingRecords):
Small unrelated fix.

* UserInterface/Views/TimelineIcons.css:
(.composite-record .icon):
* UserInterface/Views/TimelineRecordBar.css:
(.timeline-record-bar.timeline-record-type-layout.layout-timeline-record-composite > .segment):
* UserInterface/Views/TimelineRecordTreeElement.js:
(WebInspector.TimelineRecordTreeElement):
New styles and tree element icon.

Source/WebKit2:

* WebProcess/WebPage/mac/RemoteLayerTreeDrawingArea.mm:
(WebKit::RemoteLayerTreeDrawingArea::flushLayers):
* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::flushLayers):
Added CA transaction post-commit handlers to instrument end of composite.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186133 268f45cc-cd09-0410-ab3c-d52691b4dbfc

The bounds on InteractionInformationAtPosition should be more precise
https://bugs.webkit.org/show_bug.cgi?id=146468
-and corresponding-
rdar://problem/20739834

Reviewed by Enrica Casucci and Simon Fraser.

Source/WebCore:

Export absoluteContentQuad().
* rendering/RenderBox.h:

Source/WebKit2:

For links, get the TextQuads from the Range, and for images, look at the
absoluteContentQuad(). All other items can continue to use absoluteBoundingBox.
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::getPositionInformation):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186132 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Added myself, Jonathan Davis, as a committer.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186131 268f45cc-cd09-0410-ab3c-d52691b4dbfc

Added WebKit feature status page to navigation.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@186130 268f45cc-cd09-0410-ab3c-d52691b4dbfc