Source/WebCore/ChangeLog

   1 2012-04-05  Hans Muller  <hmuller@adobe.com>
   2
   3         CSS Exclusions polygon shape arguments should be comma separated
   4         https://bugs.webkit.org/show_bug.cgi?id=82368
   5
   6         Reviewed by Ryosuke Niwa.
   7
   8         Changed the CSS Parser to accept a conventional comma separated argument list for the
   9         polygon exclusion shape. The syntax had used spaces to separate x,y coordinates, like:
  10         polygon(10px,20px 30px,40px).  Now commas separate points: polygon(10px 20px, 30px 40px).
  11         This change is per the draft exclusions spec, http://dev.w3.org/csswg/css3-exclusions.
  12         Additional relevant information about CSS argument list syntax can be found
  13         here: http://dev.w3.org/csswg/css3-values/#component-whitespace.
  14
  15         Factored comma recognition idiom in CSSParser.cpp into isComma() utility function.
  16
  17         The existing tests have been updated.
  18
  19         * css/CSSParser.cpp:
  20         (WebCore::isComma)
  21         (WebCore::CSSParser::parseFillPosition)
  22         (WebCore::CSSParser::parseFillRepeat)
  23         (WebCore::CSSParser::parseFillProperty)
  24         (WebCore::CSSParser::parseCubicBezierTimingFunctionValue)
  25         (WebCore::CSSParser::parseAnimationTimingFunction)
  26         (WebCore::CSSParser::parseAnimationProperty)
  27         (WebCore::CSSParser::parseExclusionShapePolygon)
  28         (WebCore::CSSParser::parseDeprecatedGradient)
  29         (WebCore::CSSParser::parseRadialGradient)
  30         (WebCore::CSSParser::parseGradientColorStops)
  31         (WebCore::CSSParser::parseImageSet)
  32         (WebCore::filterInfoForName)
  33         (WebCore::CSSParser::parseCustomFilter)
  34         (WebCore::CSSParser::parseFontFeatureSettings)
  35
  36         * css/CSSWrapShapes.cpp:
  37         (WebCore::CSSWrapShapePolygon::cssText):
  38
  39 2012-04-05  Joshua Bell  <jsbell@chromium.org>
  40
  41         IndexedDB: Support string.length in keyPaths
  42         https://bugs.webkit.org/show_bug.cgi?id=83221
  43
  44         Special case in the IDB spec - keyPaths can reference the |length| property
  45         of string values. Other instrinsic properties (|length| of Array, etc) are
  46         handled automagically. Relevant section of the updated spec is:
  47         http://dvcs.w3.org/hg/IndexedDB/raw-file/tip/Overview.html#key-path-construct
  48
  49         Reviewed by Kentaro Hara.
  50
  51         Test: storage/indexeddb/keypath-intrinsic-properties.html
  52
  53         * bindings/v8/IDBBindingUtilities.cpp:
  54         (WebCore):
  55
  56 2012-04-05  Arvid Nilsson  <anilsson@rim.com>
  57
  58         [BlackBerry] Update the InstrumentedPlatformCanvas after rebasing Skia
  59         https://bugs.webkit.org/show_bug.cgi?id=83314
  60
  61         Reviewed by George Staikos.
  62
  63         RIM PR: 143771
  64         One new virtual method was added to the SkCanvas, to draw a nine piece
  65         image. Override it and mark output as not being a solid color anymore.
  66
  67         * platform/graphics/blackberry/InstrumentedPlatformCanvas.h:
  68         (WebCore::InstrumentedPlatformCanvas::drawBitmapNine):
  69         (InstrumentedPlatformCanvas):
  70
  71 2012-04-05  Oliver Hunt  <oliver@apple.com>
  72
  73         Make WebCore use jsCast rather than static_cast for casting JSC objects
  74         https://bugs.webkit.org/show_bug.cgi?id=83320
  75
  76         Reviewed by Stephanie Lewis.
  77
  78         Mechanically replace static_cast with jsCast where ever we can.
  79
  80         * WebCore.exp.in:
  81         * bindings/js/DOMWrapperWorld.cpp:
  82         (WebCore::JSStringOwner::finalize):
  83         * bindings/js/DOMWrapperWorld.h:
  84         (WebCore::currentWorld):
  85         * bindings/js/JSArrayBufferCustom.cpp:
  86         (WebCore::JSArrayBufferConstructor::constructJSArrayBuffer):
  87         * bindings/js/JSAudioContextCustom.cpp:
  88         (WebCore::JSAudioContextConstructor::constructJSAudioContext):
  89         * bindings/js/JSCSSRuleListCustom.cpp:
  90         (WebCore::JSCSSRuleListOwner::isReachableFromOpaqueRoots):
  91         * bindings/js/JSCSSStyleDeclarationCustom.cpp:
  92         (WebCore::cssPropertyGetterPixelOrPosPrefixCallback):
  93         (WebCore::cssPropertyGetterCallback):
  94         * bindings/js/JSCSSValueCustom.cpp:
  95         (WebCore::JSCSSValueOwner::isReachableFromOpaqueRoots):
  96         (WebCore::JSCSSValueOwner::finalize):
  97         * bindings/js/JSDOMBinding.cpp:
  98         (WebCore::reportException):
  99         * bindings/js/JSDOMBinding.h:
 100         (WebCore::deprecatedGlobalObjectForPrototype):
 101         (WebCore::getDOMPrototype):
 102         * bindings/js/JSDOMFormDataCustom.cpp:
 103         (WebCore::toHTMLFormElement):
 104         (WebCore::JSDOMFormDataConstructor::constructJSDOMFormData):
 105         * bindings/js/JSDOMMimeTypeArrayCustom.cpp:
 106         (WebCore::JSDOMMimeTypeArray::nameGetter):
 107         * bindings/js/JSDOMPluginArrayCustom.cpp:
 108         (WebCore::JSDOMPluginArray::nameGetter):
 109         * bindings/js/JSDOMPluginCustom.cpp:
 110         (WebCore::JSDOMPlugin::nameGetter):
 111         * bindings/js/JSDOMStringMapCustom.cpp:
 112         (WebCore::JSDOMStringMap::nameGetter):
 113         * bindings/js/JSDOMWindowBase.cpp:
 114         (WebCore::toJSDOMWindow):
 115         * bindings/js/JSDOMWindowCustom.cpp:
 116         (WebCore::childFrameGetter):
 117         (WebCore::indexGetter):
 118         (WebCore::namedItemGetter):
 119         (WebCore::toDOMWindow):
 120         * bindings/js/JSDOMWindowCustom.h:
 121         (WebCore::asJSDOMWindow):
 122         * bindings/js/JSDOMWindowShell.h:
 123         (WebCore::JSDOMWindowShell::window):
 124         * bindings/js/JSDOMWrapper.h:
 125         (WebCore::JSDOMWrapper::globalObject):
 126         * bindings/js/JSDataViewCustom.cpp:
 127         (WebCore::JSDataViewConstructor::constructJSDataView):
 128         * bindings/js/JSEventListener.cpp:
 129         (WebCore::JSEventListener::handleEvent):
 130         * bindings/js/JSEventTarget.cpp:
 131         (WebCore):
 132         (WebCore::toEventTarget):
 133         * bindings/js/JSFloat32ArrayCustom.cpp:
 134         (WebCore::JSFloat32ArrayConstructor::constructJSFloat32Array):
 135         * bindings/js/JSFloat64ArrayCustom.cpp:
 136         (WebCore::JSFloat64ArrayConstructor::constructJSFloat64Array):
 137         * bindings/js/JSGeolocationCustom.cpp:
 138         (WebCore::JSGeolocation::getCurrentPosition):
 139         (WebCore::JSGeolocation::watchPosition):
 140         * bindings/js/JSHTMLAllCollectionCustom.cpp:
 141         (WebCore::callHTMLAllCollection):
 142         (WebCore::JSHTMLAllCollection::nameGetter):
 143         * bindings/js/JSHTMLCollectionCustom.cpp:
 144         (WebCore::JSHTMLCollection::nameGetter):
 145         * bindings/js/JSHTMLDocumentCustom.cpp:
 146         (WebCore::JSHTMLDocument::nameGetter):
 147         * bindings/js/JSHTMLFormElementCustom.cpp:
 148         (WebCore::JSHTMLFormElement::nameGetter):
 149         * bindings/js/JSHTMLFrameSetElementCustom.cpp:
 150         (WebCore::JSHTMLFrameSetElement::nameGetter):
 151         * bindings/js/JSHTMLOptionsCollectionCustom.cpp:
 152         (WebCore::JSHTMLOptionsCollection::remove):
 153         * bindings/js/JSImageConstructor.cpp:
 154         (WebCore::constructImage):
 155         * bindings/js/JSInjectedScriptManager.cpp:
 156         (WebCore::InjectedScriptManager::createInjectedScript):
 157         (WebCore::InjectedScriptManager::discardInjectedScript):
 158         (WebCore::InjectedScriptManager::injectedScriptFor):
 159         * bindings/js/JSInt16ArrayCustom.cpp:
 160         (WebCore::JSInt16ArrayConstructor::constructJSInt16Array):
 161         * bindings/js/JSInt32ArrayCustom.cpp:
 162         (WebCore::JSInt32ArrayConstructor::constructJSInt32Array):
 163         * bindings/js/JSInt8ArrayCustom.cpp:
 164         (WebCore::JSInt8ArrayConstructor::constructJSInt8Array):
 165         * bindings/js/JSLazyEventListener.cpp:
 166         (WebCore::JSLazyEventListener::initializeJSFunction):
 167         * bindings/js/JSNamedNodeMapCustom.cpp:
 168         (WebCore::JSNamedNodeMap::nameGetter):
 169         * bindings/js/JSNodeCustom.cpp:
 170         (WebCore::JSNodeOwner::isReachableFromOpaqueRoots):
 171         (WebCore::JSNodeOwner::finalize):
 172         * bindings/js/JSNodeCustom.h:
 173         (WebCore::toJS):
 174         * bindings/js/JSNodeFilterCustom.cpp:
 175         (WebCore::toNodeFilter):
 176         * bindings/js/JSNodeListCustom.cpp:
 177         (WebCore::JSNodeListOwner::isReachableFromOpaqueRoots):
 178         (WebCore::JSNodeList::nameGetter):
 179         * bindings/js/JSPluginElementFunctions.cpp:
 180         (WebCore::runtimeObjectPropertyGetter):
 181         (WebCore::callPlugin):
 182         * bindings/js/JSPopStateEventCustom.cpp:
 183         (WebCore::JSPopStateEvent::state):
 184         * bindings/js/JSSQLTransactionCustom.cpp:
 185         (WebCore::JSSQLTransaction::executeSql):
 186         * bindings/js/JSSharedWorkerCustom.cpp:
 187         (WebCore::JSSharedWorkerConstructor::constructJSSharedWorker):
 188         * bindings/js/JSStorageCustom.cpp:
 189         (WebCore::JSStorage::nameGetter):
 190         * bindings/js/JSStyleSheetListCustom.cpp:
 191         (WebCore::JSStyleSheetList::nameGetter):
 192         * bindings/js/JSTextTrackCueCustom.cpp:
 193         (WebCore::JSTextTrackCueOwner::isReachableFromOpaqueRoots):
 194         * bindings/js/JSTextTrackCustom.cpp:
 195         (WebCore::JSTextTrackOwner::isReachableFromOpaqueRoots):
 196         * bindings/js/JSTextTrackListCustom.cpp:
 197         (WebCore::JSTextTrackListOwner::isReachableFromOpaqueRoots):
 198         * bindings/js/JSTrackCustom.cpp:
 199         (WebCore::toTrack):
 200         * bindings/js/JSUint16ArrayCustom.cpp:
 201         (WebCore::JSUint16ArrayConstructor::constructJSUint16Array):
 202         * bindings/js/JSUint32ArrayCustom.cpp:
 203         (WebCore::JSUint32ArrayConstructor::constructJSUint32Array):
 204         * bindings/js/JSUint8ArrayCustom.cpp:
 205         (WebCore::JSUint8ArrayConstructor::constructJSUint8Array):
 206         * bindings/js/JSUint8ClampedArrayCustom.cpp:
 207         (WebCore::JSUint8ClampedArrayConstructor::constructJSUint8ClampedArray):
 208         * bindings/js/JSWebKitMutationObserverCustom.cpp:
 209         (WebCore::JSWebKitMutationObserverConstructor::constructJSWebKitMutationObserver):
 210         * bindings/js/JSWebKitPointCustom.cpp:
 211         (WebCore::JSWebKitPointConstructor::constructJSWebKitPoint):
 212         * bindings/js/JSWebSocketCustom.cpp:
 213         (WebCore::JSWebSocketConstructor::constructJSWebSocket):
 214         * bindings/js/JSWorkerContextBase.cpp:
 215         (WebCore::toJSDedicatedWorkerContext):
 216         (WebCore::toJSSharedWorkerContext):
 217         * bindings/js/JSWorkerCustom.cpp:
 218         (WebCore::JSWorkerConstructor::constructJSWorker):
 219         * bindings/js/JSXSLTProcessorCustom.cpp:
 220         (WebCore::JSXSLTProcessor::importStylesheet):
 221         (WebCore::JSXSLTProcessor::transformToFragment):
 222         (WebCore::JSXSLTProcessor::transformToDocument):
 223         * bindings/js/ScriptCallStackFactory.cpp:
 224         (WebCore::createScriptCallStackForInspector):
 225         * bindings/js/ScriptControllerMac.mm:
 226         (WebCore::updateStyleIfNeededForBindings):
 227         * bindings/js/ScriptDebugServer.cpp:
 228         (WebCore::ScriptDebugServer::dispatchDidPause):
 229         * bindings/js/ScriptObject.cpp:
 230         (WebCore::ScriptGlobalObject::set):
 231         * bindings/js/ScriptState.cpp:
 232         (WebCore::domWindowFromScriptState):
 233         (WebCore::scriptExecutionContextFromScriptState):
 234         * bindings/js/SerializedScriptValue.cpp:
 235         (WebCore::CloneSerializer::fillTransferMap):
 236         (WebCore::CloneSerializer::dumpArrayBufferView):
 237         (WebCore::CloneDeserializer::getJSValue):
 238         (WebCore::CloneDeserializer::readTerminal):
 239         * bindings/objc/WebScriptObject.mm:
 240         (-[WebScriptObject _isSafeScript]):
 241         (+[WebScriptObject _convertValueToObjcValue:originRootObject:rootObject:]):
 242         * bindings/scripts/CodeGeneratorJS.pm:
 243         (GenerateGetOwnPropertySlotBody):
 244         (GenerateGetOwnPropertyDescriptorBody):
 245         (GenerateImplementation):
 246         (GenerateCallWith):
 247         (NativeToJSValue):
 248         (GenerateConstructorDefinition):
 249         * bridge/jni/jni_jsobject.h:
 250         * bridge/jni/jni_jsobject.mm:
 251         (JavaJSObject::convertJObjectToValue):
 252         * bridge/jni/jsc/JavaInstanceJSC.cpp:
 253         (JavaInstance::invokeMethod):
 254         * bridge/objc/objc_utility.mm:
 255         (JSC::Bindings::convertValueToObjcValue):
 256         * bridge/runtime_method.cpp:
 257         (JSC::callRuntimeMethod):
 258         * bridge/runtime_object.cpp:
 259         (JSC::Bindings::callRuntimeConstructor):
 260         * testing/js/WebCoreTestSupport.cpp:
 261         (WebCoreTestSupport::injectInternalsObject):
 262         (WebCoreTestSupport::resetInternalsObject):
 263
 264 2012-04-05  Martin Robinson  <mrobinson@igalia.com>
 265
 266         [GTK] Scrolling some iframes that are partially out of the viewport leads to repaint errors
 267         https://bugs.webkit.org/show_bug.cgi?id=83309
 268
 269         Reviewed by Gustavo Noronha Silva.
 270
 271         Test: platform/gtk/fast/frames/scrolling-iframe-out-of-viewport.html
 272
 273         The X11 backing store was not properly trimming the scroll region when it
 274         was only a portion of the screen. This was hidden by subsequent repaints.
 275
 276         * platform/gtk/GtkWidgetBackingStoreX11.cpp:
 277         (WebCore::WidgetBackingStore::scroll): Fix the calculation of the scrolling region.
 278
 279 2012-04-05  Sheriff Bot  <webkit.review.bot@gmail.com>
 280
 281         Unreviewed, rolling out r113299.
 282         http://trac.webkit.org/changeset/113299
 283         https://bugs.webkit.org/show_bug.cgi?id=83297
 284
 285         The patch broke 11 tests on Lion. (Requested by jonlee on
 286         #webkit).
 287
 288         * xml/parser/XMLDocumentParser.h:
 289         (XMLDocumentParser):
 290         * xml/parser/XMLDocumentParserLibxml2.cpp:
 291         (WebCore::XMLDocumentParser::XMLDocumentParser):
 292         (WebCore::XMLDocumentParser::startElementNs):
 293         (WebCore::XMLDocumentParser::endElementNs):
 294         (WebCore::getEntityHandler):
 295         (WebCore::XMLDocumentParser::initializeParserContext):
 296
 297 2012-04-05  Justin Novosad  <junov@chromium.org>
 298
 299         [Chromium] Correct misleading trace event names in Canvas2DLayerChromium
 300         https://bugs.webkit.org/show_bug.cgi?id=83310
 301
 302         Reviewed by Adrienne Walker.
 303
 304         * platform/graphics/chromium/Canvas2DLayerChromium.cpp:
 305         (WebCore::Canvas2DLayerChromium::paintContentsIfDirty):
 306
 307 2012-04-05  Dean Jackson  <dino@apple.com>
 308
 309         [mac] requestAnimationFrame sometimes stuck when page loads in a background tab
 310         https://bugs.webkit.org/show_bug.cgi?id=76105
 311
 312         Reviewed by Simon Fraser.
 313
 314         Fix three issues with requestAnimationFrame:
 315         - It's possible for the call to rAF to come to the document before there
 316           is a page associated. Added a guard for this.
 317         - A page may try to suspend the scripted animations before the
 318           ScriptedAnimationController exists, in which case we need to
 319           suspend it immediately after it is created. Do this by keeping
 320           track of the state in Page. Otherwise rAF would be busy looping
 321           on hidden pages until they are brought to the front and hidden again.
 322         - A page created in the background (from WebKit1) does not get
 323           informed it is not visible. This can mean that resume() is called
 324           more times than suspend() and we get into a state where the number
 325           of suspensions becomes -1, and thus fails truthiness tests. Clamp it
 326           to values >= 0.
 327
 328         No new tests, since this is not automatically testable. The most reliable test
 329         is to open a page with rAF in a background tab within Safari.
 330
 331         * dom/Document.cpp:
 332         (WebCore::Document::webkitRequestAnimationFrame):
 333         * dom/ScriptedAnimationController.cpp:
 334         (WebCore::ScriptedAnimationController::resume):
 335         * page/Page.cpp:
 336         (WebCore::Page::Page):
 337         (WebCore::Page::suspendScriptedAnimations):
 338         (WebCore::Page::resumeScriptedAnimations):
 339         * page/Page.h:
 340         (WebCore::Page::scriptedAnimationsSuspended):
 341         (Page):
 342
 343 2012-04-05  Brady Eidson  <beidson@apple.com>
 344
 345         <rdar://problem/9359029> and https://bugs.webkit.org/show_bug.cgi?id=83311
 346         Crashes in WebProcess at WebCore::HistoryController::recursiveSetProvisionalItem when restoring previous session
 347
 348         Reviewed by Sam Weinig.
 349
 350         It's possible to hit a race condition between the UIProcess and the WebProcess where the UIProcess records for a
 351         page have been cleared out but the WebProcess is still trying to perform a history navigation within that page.
 352
 353         In this situation HistoryController code that expects there to always be a current history item in the back/forward
 354         controller is wrong.
 355
 356         No new tests. (The race conditions involved have proven making a test impractical)
 357
 358         * loader/HistoryController.cpp:
 359         (WebCore::HistoryController::recursiveSetProvisionalItem): Don't ASSERT the fromItem. We now know there might not be one.
 360         (WebCore::HistoryController::recursiveGoToItem): Ditto
 361         (WebCore::HistoryController::itemsAreClones): Always return false if either item is null, as a null item and a non-null
 362           item cannot possible be clones of each other.
 363
 364 2012-04-05  Adam Klein  <adamk@chromium.org>
 365
 366         Crash in MutationObservers due to an invalid HashSet iterator
 367         https://bugs.webkit.org/show_bug.cgi?id=83304
 368
 369         Reviewed by Ojan Vafai.
 370
 371         If the observed node has been GCed when we clear transient observers
 372         from it, the HashSet iterator in WebKitMutationObserver::deliver would
 373         be invalidated. This patch fixes that behavior by copying the relevant
 374         registrations into a seperate vector first and operating on the copy.
 375
 376         This patch also fixes a bug: transient observers should be cleared
 377         after every microtask, not just when delivering.
 378
 379         Tests: fast/mutation/clear-transient-without-delivery.html
 380                fast/mutation/transient-gc-crash.html
 381
 382         * dom/MutationObserverRegistration.cpp:
 383         (WebCore::MutationObserverRegistration::observedSubtreeNodeWillDetach):
 384         Notify the observer that it has a transient registration so it can be properly cleared.
 385         * dom/MutationObserverRegistration.h:
 386         (WebCore::MutationObserverRegistration::hasTransientRegistrations):
 387         Add an accessor for use when deliver() creates its vector of registrations.
 388         * dom/WebKitMutationObserver.cpp:
 389         (WebCore::WebKitMutationObserver::setHasTransientRegistration): Add this to the active observer set
 390         to allow transient registrations to be cleared appropriately.
 391         (WebCore::WebKitMutationObserver::deliver): Avoid modifying m_registrations while iterating over it.
 392         Clear registrations before checking for a lack of records to deliver.
 393         * dom/WebKitMutationObserver.h:
 394
 395 2012-04-05  Adam Klein  <adamk@chromium.org>
 396
 397         Rebaseline binding tests after r113272.
 398
 399         * bindings/scripts/test/V8/V8TestInterface.cpp:
 400         (WebCore::V8TestInterface::constructorCallback):
 401         * bindings/scripts/test/V8/V8TestNamedConstructor.cpp:
 402         (WebCore::V8TestNamedConstructorConstructorCallback):
 403         * bindings/scripts/test/V8/V8TestObj.cpp:
 404         (WebCore::V8TestObj::constructorCallback):
 405         * bindings/scripts/test/V8/V8TestSerializedScriptValueInterface.cpp:
 406         (WebCore::V8TestSerializedScriptValueInterface::constructorCallback):
 407
 408 2012-04-05  Stephen White  <senorblanco@chromium.org>
 409
 410         [chromium] Drawing an accelerated canvas onto itself is slow.
 411         https://bugs.webkit.org/show_bug.cgi?id=83295
 412
 413         Fixed by using a GPU-side deepCopy() in skia, so we don't trigger a
 414         readback and re-upload.
 415
 416         Reviewed by James Robinson.
 417
 418         Correctness is covered by existing tests. Performance is covered
 419         by the test case attached to the bug.
 420
 421         * platform/graphics/skia/ImageSkia.cpp:
 422         (WebCore::BitmapImageSingleFrameSkia::create):
 423
 424 2012-04-05  Dana Jansens  <danakj@chromium.org>
 425
 426         [chromium] Record the number of tiles paint-culled instead of the incorrect number of pixels
 427         https://bugs.webkit.org/show_bug.cgi?id=82423
 428
 429         Reviewed by Adrienne Walker.
 430
 431         We currently attempt to record the number of pixels we save from uploading,
 432         however the number is wrong in two ways.
 433         1) An early-out if everything is culled results in nothing being reported.
 434         2) With threaded-compositor, the number of pixels uploaded in one tile can be
 435         changed by culling in another tile, making the number far too difficult to
 436         compute to be worthwhile. If a tile was going to be partial-updated, but is
 437         culled, now a new tile gets to be partial-updated instead, which affects the
 438         numbers.
 439
 440         This patch breaks up CCOverdrawMetrics to make the methods/variables less
 441         overloaded and more clear. This way they don't have to mirror each other
 442         so closely on paint/draw sides.
 443
 444         Then we record the number of tiles for which we prevented upload via paint
 445         culling. This will still over-report (a dirty culled tile stays dirty and
 446         will be culled again each commit), but it seems more reasonable as it
 447         clearly does not try to mimic a performance metric in the way that a pixel
 448         count does, but still gives a rough estimation of the amount of paint
 449         culling going on in a page.
 450
 451         Covered by existing tests, modified where needed.
 452
 453         * platform/graphics/chromium/TiledLayerChromium.cpp:
 454         (WebCore::TiledLayerChromium::prepareToUpdateTiles):
 455         * platform/graphics/chromium/cc/CCOverdrawMetrics.cpp:
 456         (WebCore::CCOverdrawMetrics::CCOverdrawMetrics):
 457         (WebCore::CCOverdrawMetrics::didCullTileForUpload):
 458         (WebCore):
 459         (WebCore::CCOverdrawMetrics::didUpload):
 460         (WebCore::CCOverdrawMetrics::didCullForDrawing):
 461         (WebCore::CCOverdrawMetrics::recordMetrics):
 462         (WebCore::CCOverdrawMetrics::recordMetricsInternal):
 463         * platform/graphics/chromium/cc/CCOverdrawMetrics.h:
 464         (CCOverdrawMetrics):
 465         (WebCore::CCOverdrawMetrics::pixelsCulledForDrawing):
 466         (WebCore::CCOverdrawMetrics::pixelsUploadedOpaque):
 467         (WebCore::CCOverdrawMetrics::pixelsUploadedTranslucent):
 468         (WebCore::CCOverdrawMetrics::tilesCulledForUpload):
 469         * platform/graphics/chromium/cc/CCQuadCuller.cpp:
 470         (WebCore::CCQuadCuller::append):
 471
 472 2012-04-05  Enrica Casucci  <enrica@apple.com>
 473
 474         Provide a separate editing command to cleanup of redundant markup.
 475         https://bugs.webkit.org/show_bug.cgi?id=83240
 476
 477         Reviewed by Ryosuke Niwa.
 478
 479         This patch moves the logic originally performed by the private
 480         method removeRedundantMarkup inside ReplaceSelectionCommand into
 481         a separate command. This way the markup cleanup can be exposed as
 482         a separate command, independent from ReplaceSelectionCommand.
 483
 484         No new tests. No change in functionality.
 485
 486         * CMakeLists.txt:
 487         * GNUmakefile.list.am:
 488         * Target.pri:
 489         * WebCore.gypi:
 490         * WebCore.vcproj/WebCore.vcproj:
 491         * WebCore.xcodeproj/project.pbxproj:
 492         Added new file to project files.
 493         * editing/EditorAllInOne.cpp: Added SimplifyMarkupCommand.
 494         * editing/ReplaceSelectionCommand.cpp:
 495         (WebCore::ReplaceSelectionCommand::doApply): Now uses SimplifyMarkupCommand.
 496         * editing/ReplaceSelectionCommand.h:
 497         (ReplaceSelectionCommand): Removed removeRedundantMarkup.
 498         * editing/SimplifyMarkupCommand.cpp: Added.
 499         (WebCore::SimplifyMarkupCommand::SimplifyMarkupCommand):
 500         (WebCore::SimplifyMarkupCommand::doApply):
 501         * editing/SimplifyMarkupCommand.h: Added.
 502         (WebCore::SimplifyMarkupCommand::create):
 503
 504 2012-04-05  Shawn Singh  <shawnsingh@chromium.org>
 505
 506         [chromium] Need to clip to homogeneous w=0 plane when applying transforms.
 507         https://bugs.webkit.org/show_bug.cgi?id=80806
 508
 509         Reviewed by Adrienne Walker.
 510
 511         Unit tests added to CCLayerTreeHostCommon. This change is also
 512         covered by other existing unit tests and layout tests.
 513
 514         WebCore TransformationMatrix mapRect / mapQuad / projectQuad do
 515         not properly handle the case where a surface is oriented partially
 516         behind the camera, with a perspective projection. In this case,
 517         projected points may appear to be valid in cartesian coordinates,
 518         but they are indeed not valid, and this problem can only be
 519         detected in homogeneous coordinates after applying the transform,
 520         before the divide-by-w step.
 521
 522         The correct solution is to clip geometry where w < 0. This patch
 523         makes this change local to chromium only, to fix rendering bugs
 524         that arise from this problem. The primary fix is to correct
 525         calculateVisibleLayerRect(), but other ancillary locations are
 526         also fixed, in particular, the antialiasing code path is simply
 527         skipped when this case arises.
 528
 529         Eventually this math needs to be merged into TransformationMatrix,
 530         to fix hit-testing bugs that occur in both Chromium and Safari.
 531
 532         * WebCore.gypi:
 533         * platform/graphics/chromium/LayerRendererChromium.cpp:
 534         (WebCore::findTileProgramUniforms):
 535         (WebCore::LayerRendererChromium::drawTileQuad):
 536         * platform/graphics/chromium/cc/CCLayerTreeHostCommon.cpp:
 537         (WebCore::CCLayerTreeHostCommon::calculateVisibleRect):
 538         (WebCore::isScaleOrTranslation):
 539         (WebCore::calculateDrawTransformsAndVisibilityInternal):
 540         * platform/graphics/chromium/cc/CCMathUtil.cpp: Added.
 541         (WebCore):
 542         (WebCore::HomogeneousCoordinate::HomogeneousCoordinate):
 543         (HomogeneousCoordinate):
 544         (WebCore::HomogeneousCoordinate::shouldBeClipped):
 545         (WebCore::HomogeneousCoordinate::cartesianPoint2d):
 546         (WebCore::projectPoint):
 547         (WebCore::mapPoint):
 548         (WebCore::computeClippedPointForEdge):
 549         (WebCore::expandBoundsToIncludePoint):
 550         (WebCore::computeEnclosingRectOfClippedQuad):
 551         (WebCore::computeEnclosingRect):
 552         (WebCore::CCMathUtil::mapClippedRect):
 553         (WebCore::CCMathUtil::projectClippedRect):
 554         (WebCore::CCMathUtil::mapQuad):
 555         (WebCore::CCMathUtil::projectQuad):
 556         * platform/graphics/chromium/cc/CCMathUtil.h: Added.
 557         (WebCore):
 558         (CCMathUtil):
 559         * platform/graphics/chromium/cc/CCOcclusionTracker.cpp:
 560         (WebCore::computeUnoccludedContentRect):
 561
 562 2012-04-05  Patrick Gansterer  <paroga@webkit.org>
 563
 564         [Qt] Correct <wtf/*.h> include paths.
 565         https://bugs.webkit.org/show_bug.cgi?id=83270
 566
 567         Reviewed by Eric Seidel.
 568
 569         Modify the #include declerations so that the
 570         wtf types are included using the full path.
 571
 572         * platform/graphics/TiledBackingStoreBackend.h:
 573
 574 2012-04-05  Antonio Gomes  <agomes@rim.com>
 575
 576         Unreviewed bit left over from rebasing the origin patch before landing it.
 577
 578         Complementary to r113329.
 579
 580         * rendering/HitTestResult.cpp:
 581         (WebCore::HitTestResult::operator=):
 582
 583 2012-04-05  Sheriff Bot  <webkit.review.bot@gmail.com>
 584
 585         Unreviewed, rolling out r113241.
 586         http://trac.webkit.org/changeset/113241
 587         https://bugs.webkit.org/show_bug.cgi?id=83293
 588
 589         Requested by sievers@chromium.org (Requested by zhenyao on
 590         #webkit).
 591
 592         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
 593         (WebCore::CCLayerTreeHostImpl::canDraw):
 594         (WebCore::CCLayerTreeHostImpl::prepareToDraw):
 595         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.h:
 596         (CCLayerTreeHostImpl):
 597         * platform/graphics/chromium/cc/CCThreadProxy.cpp:
 598         (WebCore::CCThreadProxy::scheduledActionDrawAndSwapInternal):
 599
 600 2012-04-05  Sheriff Bot  <webkit.review.bot@gmail.com>
 601
 602         Unreviewed, rolling out r113341.
 603         http://trac.webkit.org/changeset/113341
 604         https://bugs.webkit.org/show_bug.cgi?id=83299
 605
 606         bots caught some build errors (Requested by shawnsingh on
 607         #webkit).
 608
 609         * WebCore.gypi:
 610         * platform/graphics/chromium/LayerRendererChromium.cpp:
 611         (WebCore::findTileProgramUniforms):
 612         (WebCore::LayerRendererChromium::drawTileQuad):
 613         * platform/graphics/chromium/cc/CCLayerTreeHostCommon.cpp:
 614         (WebCore::CCLayerTreeHostCommon::calculateVisibleRect):
 615         (WebCore::isScaleOrTranslation):
 616         (WebCore::calculateDrawTransformsAndVisibilityInternal):
 617         * platform/graphics/chromium/cc/CCMathUtil.cpp: Removed.
 618         * platform/graphics/chromium/cc/CCMathUtil.h: Removed.
 619         * platform/graphics/chromium/cc/CCOcclusionTracker.cpp:
 620         (WebCore::projectQuad):
 621         (WebCore):
 622         (WebCore::computeUnoccludedContentRect):
 623
 624 2012-04-05  Justin Novosad  <junov@chromium.org>
 625
 626         [Chromium] With the skia port, setting LCD text filtering is causing
 627         texture cache invalidations of gpu canvas backing store
 628         https://bugs.webkit.org/show_bug.cgi?id=74183
 629
 630         Reviewed by Stephen White.
 631
 632         Replacing unnecessary usage of SkCanvas::LayerIter with calls to
 633         SkCanvas::isDrawingToLayer(). Same results, lower overhead, and no
 634         GPU texture invalidation.
 635
 636         * platform/graphics/chromium/FontChromiumWin.cpp:
 637         (WebCore):
 638         (WebCore::TransparencyAwareFontPainter::initializeForGDI):
 639         * platform/graphics/harfbuzz/FontHarfBuzz.cpp:
 640         (WebCore):
 641         (WebCore::adjustTextRenderMode):
 642         * platform/graphics/skia/FontSkia.cpp:
 643         (WebCore):
 644         (WebCore::adjustTextRenderMode):
 645         * platform/graphics/skia/SkiaFontWin.cpp:
 646         (WebCore):
 647         (WebCore::disableTextLCD):
 648         * rendering/RenderThemeChromiumWin.cpp:
 649         (WebCore):
 650
 651 2012-04-05  Zan Dobersek  <zandobersek@gmail.com>
 652
 653         media/track/track-webvtt-tc004-magic-header.html flakily times out
 654         https://bugs.webkit.org/show_bug.cgi?id=72279
 655
 656         Reviewed by Eric Carlson.
 657
 658         Move all the checking for the WebVTT file identifier from TextTrackLoader
 659         to WebVTTParser, notifying the parser clients of parsing failures through
 660         a new method. Remove the WebVTTParser::fileIdentifierMaximumLength method
 661         and make WebVTTParser::hasRequiredFileIdentifier private. The latter now
 662         operates on a vector holding the identifier data to which data is added
 663         until there's enough of it to possibly hold the optional BOM character
 664         and the required WEBVTT character sequence.
 665
 666         No new tests. Fixes an existing test.
 667
 668         * html/track/WebVTTParser.cpp: Fix a typo throughout the file:
 669         fileIdentiferLength -> fileIdentifierLength
 670         (WebCore):
 671         (WebCore::WebVTTParser::parseBytes): Only continue with searcing for
 672         the WEBVTT identifier if there's enough data to possibly contain the
 673         identifier. Report that the file failed to parse if the identifier
 674         was then not found.
 675         (WebCore::WebVTTParser::hasRequiredFileIdentifier): Refactor to operate
 676         on the vector containing identifier data and merge with the
 677         hasLongWebVTTIdentifier function.
 678         * html/track/WebVTTParser.h:
 679         (WebVTTParserClient): Add a new method.
 680         (WebVTTParser):
 681         * loader/TextTrackLoader.cpp:
 682         (WebCore::TextTrackLoader::processNewCueData): Parser is now created
 683         immediately regardless of the mime type the response holds.
 684         (WebCore::TextTrackLoader::fileFailedToParse): Log the error, set the
 685         state to failed, start the cue load timer and cancel the load if the file
 686         was not parsed successfully.
 687         (WebCore):
 688         * loader/TextTrackLoader.h:
 689         (TextTrackLoader):
 690
 691 2012-04-05  Shawn Singh  <shawnsingh@chromium.org>
 692
 693         [chromium] Need to clip to homogeneous w=0 plane when applying transforms.
 694         https://bugs.webkit.org/show_bug.cgi?id=80806
 695
 696         Reviewed by Adrienne Walker.
 697
 698         Unit tests added to CCLayerTreeHostCommon. This change is also
 699         covered by other existing unit tests and layout tests.
 700
 701         WebCore TransformationMatrix mapRect / mapQuad / projectQuad do
 702         not properly handle the case where a surface is oriented partially
 703         behind the camera, with a perspective projection. In this case,
 704         projected points may appear to be valid in cartesian coordinates,
 705         but they are indeed not valid, and this problem can only be
 706         detected in homogeneous coordinates after applying the transform,
 707         before the divide-by-w step.
 708
 709         The correct solution is to clip geometry where w < 0. This patch
 710         makes this change local to chromium only, to fix rendering bugs
 711         that arise from this problem. The primary fix is to correct
 712         calculateVisibleLayerRect(), but other ancillary locations are
 713         also fixed, in particular, the antialiasing code path is simply
 714         skipped when this case arises.
 715
 716         Eventually this math needs to be merged into TransformationMatrix,
 717         to fix hit-testing bugs that occur in both Chromium and Safari.
 718
 719         * WebCore.gypi:
 720         * platform/graphics/chromium/LayerRendererChromium.cpp:
 721         (WebCore::findTileProgramUniforms):
 722         (WebCore::LayerRendererChromium::drawTileQuad):
 723         * platform/graphics/chromium/cc/CCLayerTreeHostCommon.cpp:
 724         (WebCore::CCLayerTreeHostCommon::calculateVisibleRect):
 725         (WebCore::isScaleOrTranslation):
 726         (WebCore::calculateDrawTransformsAndVisibilityInternal):
 727         * platform/graphics/chromium/cc/CCMathUtil.cpp: Added.
 728         (WebCore):
 729         (WebCore::HomogeneousCoordinate::HomogeneousCoordinate):
 730         (HomogeneousCoordinate):
 731         (WebCore::HomogeneousCoordinate::shouldBeClipped):
 732         (WebCore::HomogeneousCoordinate::cartesianPoint2d):
 733         (WebCore::projectPoint):
 734         (WebCore::mapPoint):
 735         (WebCore::computeClippedPointForEdge):
 736         (WebCore::expandBoundsToIncludePoint):
 737         (WebCore::computeEnclosingRectOfClippedQuad):
 738         (WebCore::computeEnclosingRect):
 739         (WebCore::CCMathUtil::mapClippedRect):
 740         (WebCore::CCMathUtil::projectClippedRect):
 741         (WebCore::CCMathUtil::mapQuad):
 742         (WebCore::CCMathUtil::projectQuad):
 743         * platform/graphics/chromium/cc/CCMathUtil.h: Added.
 744         (WebCore):
 745         (CCMathUtil):
 746         * platform/graphics/chromium/cc/CCOcclusionTracker.cpp:
 747         (WebCore::computeUnoccludedContentRect):
 748
 749 2012-04-05  Jia Pu  <jpu@apple.com>
 750
 751         Move correction panel related functions from EditorClient into separated AlternativeTextClient class.
 752         https://bugs.webkit.org/show_bug.cgi?id=82970
 753
 754         Reviewed by Enrica Casucci.
 755
 756         No new tests, since there's no change of functionality.
 757
 758         The existing correction panel related functions in EditorClient are currently used only by OS X.
 759         We'd like to move them into separate AlternativeTextClient class to avoid using #ifdef's. This
 760         new client class will also hold dictation alternative related interface that we will add soon for
 761         bug 82503.
 762
 763         * WebCore.vcproj/WebCore.vcproj:
 764         * WebCore.xcodeproj/project.pbxproj:
 765         * editing/AlternativeTextController.cpp: Changes in this file are all for calling functions in
 766           AlternativeTextClient instead of those in EditoClient.
 767         (WebCore::AlternativeTextController::show):
 768         (WebCore::AlternativeTextController::dismiss):
 769         (WebCore::AlternativeTextController::dismissSoon):
 770         (WebCore::AlternativeTextController::respondToUnappliedSpellCorrection):
 771         (WebCore::AlternativeTextController::timerFired):
 772         (WebCore::AlternativeTextController::isAutomaticSpellingCorrectionEnabled):
 773         (WebCore::AlternativeTextController::alternativeTextClient):
 774         (WebCore):
 775         (WebCore::AlternativeTextController::editorClient):
 776         (WebCore::AlternativeTextController::textChecker):
 777         (WebCore::AlternativeTextController::recordAutocorrectionResponseReversed):
 778         (WebCore::AlternativeTextController::recordSpellcheckerResponseForModifiedCorrection):
 779         * editing/AlternativeTextController.h:
 780         (WebCore):
 781         (AlternativeTextController):
 782         * editing/TextCheckingHelper.h:
 783         (WebCore):
 784         * loader/EmptyClients.h:
 785         (EmptyEditorClient):
 786         * page/AlternativeTextClient.h: Added.
 787         (WebCore):
 788         (AlternativeTextClient):
 789         (WebCore::AlternativeTextClient::~AlternativeTextClient):
 790         * page/EditorClient.h:
 791         (EditorClient):
 792         * page/Page.cpp:
 793         (WebCore::Page::Page):
 794         (WebCore::Page::PageClients::PageClients):
 795         * page/Page.h:
 796         (WebCore):
 797         (PageClients):
 798         (Page):
 799         (WebCore::Page::alternativeTextClient):
 800
 801 2012-04-05  Min Qin  <qinmin@google.com>
 802
 803         Fix fullscreen video button for android
 804         https://bugs.webkit.org/show_bug.cgi?id=83087
 805
 806         Reviewed by Dimitri Glazkov.
 807
 808         https://bugs.webkit.org/show_bug.cgi?id=70437 changed css keyword for fullscreen video button.
 809         It is changed from media-fullscreen-button to media-enter-fullscreen-button.
 810         However, it does not change the value in mediaControlChromiumAndroid.css.
 811         No behavior change.
 812
 813         * css/mediaControlsChromiumAndroid.css:
 814         (video::-webkit-media-controls-fullscreen-button):
 815
 816 2012-04-05  Pavel Feldman  <pfeldman@chromium.org>
 817
 818         Web Inspector: CSS file revisions are not restored upon front-end reopen.
 819         https://bugs.webkit.org/show_bug.cgi?id=83292
 820
 821         Reviewed by Yury Semikhatsky.
 822
 823         Populating suggestions before the panel creation now.
 824
 825         * inspector/front-end/inspector.js:
 826         (WebInspector._doLoadedDoneWithCapabilities.get if):
 827
 828 2012-04-04  Simon Fraser  <simon.fraser@apple.com>
 829
 830         Add assertions to check for mutation of RenderLayer z-order lists during enumeration
 831         https://bugs.webkit.org/show_bug.cgi?id=83242
 832
 833         Reviewed by James Robinson.
 834
 835         Add debug-only code that detects whether RenderLayer's z-order and
 836         normal flow lists are being cleared or otherwise mutated while we're
 837         enumerating over them.
 838
 839         * rendering/RenderLayer.cpp:
 840         (WebCore::RenderLayer::RenderLayer):
 841         (WebCore::RenderLayer::paintList):
 842         (WebCore::RenderLayer::calculateLayerBounds):
 843         (WebCore::RenderLayer::dirtyZOrderLists):
 844         (WebCore::RenderLayer::dirtyNormalFlowList):
 845         (WebCore::RenderLayer::updateZOrderListsSlowCase):
 846         (WebCore::RenderLayer::updateNormalFlowList):
 847         * rendering/RenderLayer.h:
 848         (RenderLayer):
 849         (WebCore::RenderLayer::layerListMutationAllowed):
 850         (WebCore::RenderLayer::setLayerListMutationAllowed):
 851         (WebCore):
 852         (LayerListMutationDetector):
 853         (WebCore::LayerListMutationDetector::LayerListMutationDetector):
 854         (WebCore::LayerListMutationDetector::~LayerListMutationDetector):
 855         * rendering/RenderLayerBacking.cpp:
 856         (WebCore::RenderLayerBacking::hasVisibleNonCompositingDescendantLayers):
 857         * rendering/RenderLayerCompositor.cpp:
 858         (WebCore::RenderLayerCompositor::addToOverlapMapRecursive):
 859         (WebCore::RenderLayerCompositor::computeCompositingRequirements):
 860         (WebCore::RenderLayerCompositor::rebuildCompositingLayerTree):
 861         (WebCore::RenderLayerCompositor::updateLayerTreeGeometry):
 862         (WebCore::RenderLayerCompositor::updateCompositingDescendantGeometry):
 863         (WebCore::RenderLayerCompositor::recursiveRepaintLayerRect):
 864         (WebCore::RenderLayerCompositor::layerHas3DContent):
 865
 866 2012-04-05  Sheriff Bot  <webkit.review.bot@gmail.com>
 867
 868         Unreviewed, rolling out r113254.
 869         http://trac.webkit.org/changeset/113254
 870         https://bugs.webkit.org/show_bug.cgi?id=83291
 871
 872         Requested by sievers@chromium.org (Requested by zhenyao on
 873         #webkit).
 874
 875         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
 876         (WebCore::CCLayerTreeHostImpl::canDraw):
 877
 878 2012-04-05  Sheriff Bot  <webkit.review.bot@gmail.com>
 879
 880         Unreviewed, rolling out r113307.
 881         http://trac.webkit.org/changeset/113307
 882         https://bugs.webkit.org/show_bug.cgi?id=83288
 883
 884         Requested by sievers@chromium.org (Requested by zhenyao on
 885         #webkit).
 886
 887         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
 888         (WebCore::CCLayerTreeHostImpl::canDraw):
 889
 890 2012-03-14  Antonio Gomes  <agomes@rim.com>
 891
 892         Extend HitTestResult to support not "discarding" shadow content in favor of its DOM ancestor
 893         https://bugs.webkit.org/show_bug.cgi?id=80847
 894
 895         Reviewed by David Hyatt.
 896
 897         In order to improve touch precision on shadow tree content, patch
 898         extends the rect-hittest'ing system to allow shadow content
 899         to be stored as the rect-hittest result node list.
 900
 901         Primary use cases are for this extension is making it easier to
 902         improve the clickability of default controls of <video> and <audio>
 903         elements.
 904
 905         For the callee site, the use is simple: if shadow content in the rect-hittest
 906         result is desired, just add an extra 'AllowShadowContent' parameter at the creation
 907         of the HitTestResult object.
 908
 909         Test: fast/dom/nodesFromRect-shadowContent.html
 910
 911         * dom/Document.cpp:
 912         (WebCore::Document::nodesFromRect):
 913         * dom/Document.h:
 914         (Document):
 915         * page/EventHandler.cpp:
 916         (WebCore::EventHandler::hitTestResultAtPoint):
 917         * rendering/HitTestResult.cpp:
 918         (WebCore::HitTestResult::HitTestResult):
 919         (WebCore::HitTestResult::addNodeToRectBasedTestResult):
 920         * rendering/HitTestResult.h:
 921         (HitTestResult):
 922         (WebCore::HitTestResult::shadowContentFilterPolicy):
 923         * rendering/RenderImage.cpp:
 924         (WebCore::RenderImage::nodeAtPoint):
 925         * rendering/RenderLayer.cpp:
 926         (WebCore::RenderLayer::hitTestLayer):
 927         (WebCore::RenderLayer::hitTestList):
 928         * testing/Internals.cpp:
 929         (WebCore::Internals::nodesFromRect):
 930         * testing/Internals.h:
 931         (Internals):
 932         * testing/Internals.idl:
 933
 934 2012-04-05  Hans Muller  <hmuller@adobe.com>
 935
 936         Rename CSS Exclusions "rect" value to "rectangle"
 937         https://bugs.webkit.org/show_bug.cgi?id=82367
 938
 939         Reviewed by Ryosuke Niwa.
 940
 941         Changed the CSS exclusion shape called "rect" to "rectangle" to match
 942         the current draft spec - http://dev.w3.org/csswg/css3-exclusions.
 943
 944         The meeting minutes where the decision to rename the exclusion "rect" function
 945         to "rectangle" are: http://lists.w3.org/Archives/Public/www-style/2012Feb/0325.html
 946
 947         No new tests were created; just updated the existing ones.
 948
 949         * css/CSSParser.cpp:
 950         (WebCore::CSSParser::parseWrapShapeRectangle):
 951         (WebCore::CSSParser::parseWrapShape):
 952         * css/CSSParser.h:
 953         * css/CSSWrapShapes.cpp:
 954         (WebCore::CSSWrapShapeRectangle::cssText):
 955         * css/CSSWrapShapes.h:
 956         (WebCore::CSSWrapShapeRectangle::create):
 957         (WebCore::CSSWrapShapeRectangle::type):
 958         (WebCore::CSSWrapShapeRectangle::CSSWrapShapeRectangle):
 959
 960 2012-04-05  Arvid Nilsson  <anilsson@rim.com>
 961
 962         [BlackBerry] Disable paintingGoesToWindow() for the root layer when forced compositing mode is active
 963         https://bugs.webkit.org/show_bug.cgi?id=83127
 964
 965         Reviewed by Rob Buis.
 966
 967         RIM PR: 136381
 968         The BlackBerry port will use accelerated compositing code path to
 969         render the root layer in some situations.
 970
 971         The plan is to do this by activating the "force compositing mode"
 972         setting, but that in itself doesn't get us all the way, since painting
 973         will still go to window for the root layer.
 974
 975         Fixed by adding a special clause in
 976         RenderLayerBacking::paintingGoesToWindow() for PLATFORM(BLACKBERRY).
 977
 978         No new tests, this is currently not testable with the BlackBerry WebKit
 979         testing infrastructure.
 980
 981         * rendering/RenderLayerBacking.cpp:
 982         (WebCore::RenderLayerBacking::paintingGoesToWindow):
 983         * rendering/RenderLayerCompositor.h:
 984         (RenderLayerCompositor):
 985         (WebCore::RenderLayerCompositor::inForcedCompositingMode): Added.
 986
 987 2012-04-05  Stephen Chenney  <schenney@chromium.org>
 988
 989         REGRESSION(99539): Infinite repaint loop with SVGImage and deferred repaint timers
 990         https://bugs.webkit.org/show_bug.cgi?id=78315
 991
 992         Reviewed by Dimitri Glazkov.
 993
 994         The existing fix for this issue was failing to check if the frameView object
 995         was currently _in_ layout, in addition to whether it needs layout. Calling the
 996         redraw method while in layout leads to a debug assertion and potential infinite
 997         layout loops. Now we check whether we need layout or are in layout. We also add
 998         a check when the repaint timer fires to ensure we do not call redraw during layout
 999         at that point.
1000
1001         This patch was tested with tens of thousands of runs on layout test cases that
1002         previously crashed at a rate of about 1 in 25. Now we see no crashes and no test
1003         failures.
1004
1005         No new tests, as this exists to fix flaky existing tests.
1006
1007         * svg/graphics/SVGImageCache.cpp:
1008         (WebCore::SVGImageCache::imageContentChanged):
1009         (WebCore::SVGImageCache::redrawTimerFired):
1010
1011 2012-04-05  Keishi Hattori  <keishi@webkit.org>
1012
1013         Hide datalist element
1014         https://bugs.webkit.org/show_bug.cgi?id=82874
1015
1016         The datalist element should be hidden when ENABLE_DATALIST is on.
1017
1018         Reviewed by Kent Tamura.
1019
1020         Tests: fast/forms/datalist/datalist-fallback-content-expected.html
1021                fast/forms/datalist/datalist-fallback-content.html
1022
1023         * css/html.css:
1024         (datalist): Added display:none.
1025
1026 2012-04-05  W. James MacLean  <wjmaclean@chromium.org>
1027
1028         Adjust max bisection search iterations for TouchpadFlingGestureCurve.
1029         https://bugs.webkit.org/show_bug.cgi?id=83166
1030
1031         Reviewed by James Robinson.
1032
1033         Covered by existing tests.
1034
1035         * platform/TouchpadFlingPlatformGestureCurve.cpp:
1036         (WebCore):
1037
1038 2012-04-05  Peter Rybin  <peter.rybin@gmail.com>
1039
1040         Web Inspector: CodeGeneratorInspector.py: switch FileSystem, Profiler and Worker domains to typed API
1041         https://bugs.webkit.org/show_bug.cgi?id=83069
1042
1043         Reviewed by Yury Semikhatsky.
1044
1045         Domains are switched to 'strict' mode. However, hardly any changes were needed in C++ code because
1046         only basic types are currently used in Inspector.json.
1047
1048         * inspector/CodeGeneratorInspector.py:
1049         * inspector/InspectorProfilerAgent.cpp:
1050         (WebCore::InspectorProfilerAgent::getProfileHeaders):
1051         (WebCore):
1052         (WebCore::InspectorProfilerAgent::getObjectByHeapObjectId):
1053         * inspector/InspectorProfilerAgent.h:
1054         (InspectorProfilerAgent):
1055
1056 2012-04-05  Levi Weintraub  <leviw@chromium.org>
1057
1058         DOM measurement APIs should operate on pixel snapped values.
1059         https://bugs.webkit.org/show_bug.cgi?id=83138
1060
1061         Reviewed by Julien Chaffraix.
1062
1063         DOM measurement APIs return unscaled integers. In an effort to return the best possible values (and
1064         most consistent with current behavior) when we use sub-pixel values for Layout, we want to adjust
1065         the actual on-screen position and size values for scaling, not the sub-pixel values.
1066
1067         No new tests. No change in behavior.
1068
1069         * dom/Element.cpp:
1070         (WebCore::adjustForLocalZoom): Needs to return an integer, since this is what the DOM API calls for.
1071         (WebCore::Element::offsetWidth): Using pixel snapped values for all the following functions.
1072         (WebCore::Element::offsetHeight):
1073         (WebCore::Element::clientLeft):
1074         (WebCore::Element::clientTop):
1075         (WebCore::Element::clientWidth):
1076         (WebCore::Element::clientHeight):
1077         * html/HTMLImageElement.cpp:
1078         (WebCore::HTMLImageElement::width):
1079         (WebCore::HTMLImageElement::height):
1080
1081 2012-04-05  Pavel Feldman  <pfeldman@chromium.org>
1082
1083         Web Inspector: add ability to copy resource URL from web inspector's resources page.
1084         https://bugs.webkit.org/show_bug.cgi?id=83175
1085
1086         Reviewed by Yury Semikhatsky.
1087
1088         Added requested context menu item.
1089
1090         * inspector/front-end/ResourcesPanel.js:
1091         (WebInspector.FrameResourceTreeElement.prototype._handleContextMenuEvent):
1092
1093 2012-04-04  Andrey Kosyakov  <caseq@chromium.org>
1094
1095         Web Inspector: display frame duration instead of time ruler in Timeline panel's frame mode
1096         https://bugs.webkit.org/show_bug.cgi?id=83184
1097
1098         Reviewed by Pavel Feldman.
1099
1100         * inspector/front-end/TimelineFrameController.js:
1101         (WebInspector.TimelineFrameController): Maintain reference to TimelinePresentationModel
1102         (WebInspector.TimelineFrameController.prototype._flushFrame): Report frames to TimelinePresentationModel (only real ones)
1103         * inspector/front-end/TimelineGrid.js:
1104         (WebInspector.TimelineGrid.prototype.removeDividers): remove time marks/labels from upper time ruler
1105         * inspector/front-end/TimelineOverviewPane.js:
1106         (WebInspector.TimelineOverviewPane.prototype._updateEventDividers): Updated call site of createEventDivider
1107         * inspector/front-end/TimelinePanel.js:
1108         (WebInspector.TimelinePanel.prototype._updateEventDividers):
1109         (WebInspector.TimelinePanel.prototype._shouldShowFrames):
1110         (WebInspector.TimelinePanel.prototype._updateFrames): Frame bars rendering logic.
1111         (WebInspector.TimelinePanel.prototype._innerAddRecordToTimeline.addTimestampRecords): Do not keep frames in the panel any more.
1112         (WebInspector.TimelinePanel.prototype._innerAddRecordToTimeline):
1113         (WebInspector.TimelinePanel.prototype._refresh): Moved updateDividers() call here...
1114         (WebInspector.TimelinePanel.prototype._refreshRecords):  ... from here.
1115         * inspector/front-end/TimelinePresentationModel.js:
1116         (WebInspector.TimelinePresentationModel.createEventDivider): Use type & title, not entire record.
1117         (WebInspector.TimelinePresentationModel.prototype.frames): Maintain frames in the TimelinePresentationModel.
1118         (WebInspector.TimelinePresentationModel.prototype.reset): ditto.
1119         (WebInspector.TimelinePresentationModel.prototype.addFrame): ditto.
1120         (WebInspector.TimelinePresentationModel.Record): Fix title for timestamp events (drive by).
1121         * inspector/front-end/timelinePanel.css:
1122         (.timeline-frame-strip):
1123
1124 2012-04-05  MORITA Hajime  <morrita@google.com>
1125
1126         IsCommentNode flags is a waste.
1127         https://bugs.webkit.org/show_bug.cgi?id=83251
1128
1129         Reviewed by Ryosuke Niwa.
1130
1131         Effectively nobody uses it. To save a precious NodeFlags bit, this
1132         change removes NodeFlags::IsCommentFlag and its companion mehtod
1133         Node::isCommentNode().
1134
1135         No new tests. No behaviour change.
1136
1137         * dom/CharacterData.h:
1138         (WebCore::CharacterData::CharacterData):
1139         * dom/Comment.cpp:
1140         (WebCore::Comment::Comment):
1141         * dom/Node.h:
1142         * rendering/RenderTreeAsText.cpp:
1143         (WebCore::getTagName):
1144         This was the only client of Node::isCommentNode(). Replaced it with nodeType() call.
1145
1146 2012-04-05  Hans Wennborg  <hans@chromium.org>
1147
1148         webkit_unit_tests build fix.
1149         https://bugs.webkit.org/show_bug.cgi?id=83261
1150
1151         Unreviewed, webkit_unit_tests buildfix.
1152
1153         The CCLayerTreeHostTestEmptyContentsShouldNotDraw.runMultiThread test
1154         stopped working after r113254.
1155
1156         Disable it for now.
1157
1158         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
1159         (WebCore::CCLayerTreeHostImpl::canDraw):
1160
1161 2012-04-05  Sudarsana Nagineni  <sudarsana.nagineni@linux.intel.com>
1162
1163         [EFL] keycode translation is wrong for function keys
1164         https://bugs.webkit.org/show_bug.cgi?id=82738
1165
1166         This patch fixes keycode translation for function keys.
1167
1168         Reviewed by Philippe Normand.
1169
1170         Test: fast/events/keydown-function-keys.html
1171
1172         * platform/efl/EflKeyboardUtilities.cpp:
1173         (WebCore::createWindowsKeyMap):
1174
1175 2012-04-01  Nikolas Zimmermann  <nzimmermann@rim.com>
1176
1177         Work around an entity parsing bug in libxml2 2.7.3 (supplied with Lion) and unskip tests
1178         https://bugs.webkit.org/show_bug.cgi?id=82577
1179
1180         Reviewed by Filip Pizlo.
1181
1182         Work-around entity expansion bug that affects several SVG tests on Lion.
1183
1184         Sample test document which is currently broken:
1185         <?xml version="1.0" encoding="UTF-8"?>
1186         <!DOCTYPE svg PUBLIC "-//W3C//DTD SVG 1.1 Basic//EN" "http://www.w3.org/Graphics/SVG/1.1/DTD/svg11-basic.dtd" [
1187         <!ENTITY Smile "<rect x='.5' y='.5' width='29' height='39' fill='black' stroke='red'/>">
1188         ]>
1189
1190         <svg xmlns="http://www.w3.org/2000/svg">&Smile;</svg>
1191
1192         The expanded rect carries no namespace, thus an Element will be created for it, instead of a SVGRectElement.
1193         libxml2 2.7.4 fixed this bug (https://bugzilla.gnome.org/show_bug.cgi?id=502960) in 2009 already, but Lion
1194         still ships with 2.7.3, so we need to find a work-around for the problem. It works like this:
1195         - When an entity is requested (getEntityHandler) determine whether the entity is being declared (while the <!ENTITY.. parses)
1196           or wheter its references (when the &Smile; is parsed). If its referenced, record the current depth of the libxml2 parser.
1197         - When startElementNs is called while we're expanding entities, be sure to transfer the namespace of the parent node
1198           to the new node, but only do this if the current depth() is greater than the depth() at the time where entity expansion started.
1199           This way we only apply our workaround for elements inside entities, that get expanded at the insertion point.
1200         - When endElementNs is called, and our current depth() is less than our equal to the depth() where entity expansion started,
1201           clear the recorded detph(), and stop executing the workaround.
1202
1203         It requires storing an extra integer & boolean in XMLDocumentParser, which is only used for this work-around.
1204
1205         * xml/parser/XMLDocumentParser.h:
1206         (XMLDocumentParser):
1207         (WebCore::XMLDocumentParser::isParsingEntityDeclaration):
1208         (WebCore::XMLDocumentParser::setIsParsingEntityDeclaration):
1209         (WebCore::XMLDocumentParser::depthTriggeringEntityExpansion):
1210         (WebCore::XMLDocumentParser::setDepthTriggeringEntityExpansion):
1211         * xml/parser/XMLDocumentParserLibxml2.cpp:
1212         (WebCore::XMLDocumentParser::XMLDocumentParser):
1213         (WebCore::hackAroundLibXMLEntityParsingBug):
1214         (WebCore::XMLDocumentParser::startElementNs):
1215         (WebCore::XMLDocumentParser::endElementNs):
1216         (WebCore::entityDeclarationHandler):
1217         (WebCore::getEntityHandler):
1218         (WebCore::XMLDocumentParser::initializeParserContext):
1219
1220 2012-04-04  Kent Tamura  <tkent@chromium.org>
1221
1222         Add JavaScript and CSS code for the calendar picker implementation
1223         https://bugs.webkit.org/show_bug.cgi?id=83011
1224
1225         Reviewed by Hajime Morita.
1226
1227         Add calendarPicker.js and calendarPicker.css, and add a build rule to
1228         generate a C++ file. This change doesn't make any behavior change
1229         because the code is wrapped by ENABLE(CALENDAR_PICKER).
1230
1231         Tests: ManualTests/forms/calendar-picker.html
1232
1233         * Resources/calendarPicker.css: Added.
1234         * Resources/calendarPicker.js: Added.
1235         * WebCore.gyp/WebCore.gyp: Generate CalendarPicker.{cpp,h} from calendarPicker.{css,js}.
1236
1237 2012-04-05  Kinuko Yasuda  <kinuko@chromium.org>
1238
1239         Expose DataTransferItem.getAsEntry() to allow users access dropped files as FileEntry
1240         https://bugs.webkit.org/show_bug.cgi?id=82592
1241
1242         Reviewed by David Levin.
1243
1244         For now the method is prefixed thus it is to be exposed as 'webkitGetAsEntry'.
1245
1246         The API is proposed and discussed in the following whatwg thread:
1247         http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2011-November/033814.html
1248
1249         Add DataTransferItemFilesystem under Modules/filesystem and implemented the bridging part for chromium.
1250
1251         Test: editing/pasteboard/data-transfer-items-drag-drop-entry.html
1252
1253         * Modules/filesystem/DataTransferItemFileSystem.h: Added.
1254         (DataTransferItemFileSystem):
1255         * Modules/filesystem/DataTransferItemFileSystem.idl: Added for DataTransferItem.getAsEntry which is only exposed if filesystem is enabled.
1256         * Modules/filesystem/chromium/DataTransferItemFileSystemChromium.cpp: Added for chromium implementation.
1257         * Modules/filesystem/chromium/DataTransferItemFileSystemChromium.h: Added.
1258         * Modules/filesystem/chromium/DraggedIsolatedFileSystem.cpp: Added.
1259         * Modules/filesystem/chromium/DraggedIsolatedFileSystem.h: Added.
1260         * WebCore.gypi:
1261         * platform/chromium/ChromiumDataObject.cpp:
1262         * platform/chromium/ChromiumDataObject.h:
1263         (WebCore::ChromiumDataObject::filesystemId): Added.
1264         (WebCore::ChromiumDataObject::setFilesystemId): Added.
1265         * platform/chromium/ClipboardChromium.cpp:
1266
1267 2012-04-04  Patrick Gansterer  <paroga@webkit.org>
1268
1269         Add WTF::getCurrentLocalTime()
1270         https://bugs.webkit.org/show_bug.cgi?id=83164
1271
1272         Reviewed by Alexey Proskuryakov.
1273
1274         Replace the calls to WTF::getLocalTime() with time(0) with the new function.
1275         This allows us to use Win32 API on windows to get the same result in a next step.
1276
1277         * html/FTPDirectoryDocument.cpp:
1278         (WebCore::processFileDateString):
1279         * loader/archive/mhtml/MHTMLArchive.cpp:
1280         (WebCore::MHTMLArchive::generateMHTMLData):
1281
1282 2012-04-04  Hayato Ito  <hayato@chromium.org>
1283
1284         Remove ReifiedTreeTraversal.
1285         https://bugs.webkit.org/show_bug.cgi?id=83110
1286
1287         Reviewed by Dimitri Glazkov.
1288
1289         We can now remove ReifiedTreeTraversal. All clients have switched to use ComposedShadowTreeWalker.
1290
1291         No new tests. No change in functionality.
1292
1293         * CMakeLists.txt:
1294         * GNUmakefile.list.am:
1295         * Target.pri:
1296         * WebCore.exp.in:
1297         * WebCore.gypi:
1298         * WebCore.vcproj/WebCore.vcproj:
1299         * WebCore.xcodeproj/project.pbxproj:
1300         * dom/ReifiedTreeTraversal.cpp: Removed.
1301         * dom/ReifiedTreeTraversal.h: Removed.
1302         * testing/Internals.cpp:
1303         * testing/Internals.h:
1304         (Internals):
1305         * testing/Internals.idl:
1306
1307 2012-04-04  Dmitry Lomov  <dslomov@google.com>
1308
1309         WorkerEventQueue::close might access deleted WorkerEventQueue::EventDispatcherTask.
1310         https://bugs.webkit.org/show_bug.cgi?id=83202
1311
1312         On closing the event queue, WorkerEventQueue cancels all the tasks associated with events.
1313         The tasks in their turn delete themselves from the map whenever task gets executed.
1314         However if shutdown occurs when task is in queue but before task gets executed, the task will be deleted without execution.
1315         This patch makes sure that no deleted tasks stay in WorkerEventQueue, by task removing itself in destructor.
1316
1317         Reviewed by David Levin.
1318
1319         Covered by existing tests.
1320
1321         * workers/WorkerEventQueue.cpp:
1322         (WebCore::WorkerEventQueue::EventDispatcherTask::~EventDispatcherTask):
1323         (WorkerEventQueue::EventDispatcherTask):
1324         (WebCore::WorkerEventQueue::EventDispatcherTask::performTask):
1325
1326 2012-04-04  Julien Chaffraix  <jchaffraix@webkit.org>
1327
1328         RenderLayer scrollbars' updates should be split between layout induced and style change induced
1329         https://bugs.webkit.org/show_bug.cgi?id=83213
1330
1331         Reviewed by Simon Fraser.
1332
1333         Refactoring only, no change in behavior.
1334
1335         This patches splits up the 2 reasons for modifying the scrollbars:
1336         - style updates, handled in updateScrollbarsAfterStyleChange.
1337         - layout time, handled in updateScrollbarsAfterLayout.
1338
1339         * rendering/RenderLayer.cpp:
1340         (WebCore::RenderLayer::contentsSize):
1341         Removed now unneeded const-casts.
1342
1343         (WebCore::RenderLayer::setHasHorizontalScrollbar):
1344         (WebCore::RenderLayer::setHasVerticalScrollbar):
1345         Updated to use hasHorizontalScrollbar / hasVerticalScrollbar.
1346
1347         (WebCore::RenderLayer::scrollWidth):
1348         (WebCore::RenderLayer::scrollHeight):
1349         Made those functions |const|.
1350
1351         (WebCore::RenderLayer::computeScrollDimensions):
1352         Removed the unneeded booleans and move the do-we-have-overflow-logic
1353         into hasHorizontalOverflow and hasVerticalOverflow.
1354
1355         (WebCore::RenderLayer::hasHorizontalOverflow):
1356         (WebCore::RenderLayer::hasVerticalOverflow):
1357         Added those new helper functions.
1358
1359         (WebCore::RenderLayer::updateScrollbarsAfterLayout):
1360         (WebCore::RenderLayer::updateScrollInfoAfterLayout):
1361         Updated the latter to call the former.
1362
1363         (WebCore::RenderLayer::updateScrollbarsAfterStyleChange):
1364         (WebCore::RenderLayer::styleChanged):
1365         Ditto.
1366
1367         (WebCore::overflowCanHaveAScrollbar):
1368         Add this helper function for updateScrollbarsAfterStyleChange.
1369
1370         * rendering/RenderLayer.h:
1371         (WebCore::RenderLayer::hasHorizontalScrollbar):
1372         (WebCore::RenderLayer::hasVerticalScrollbar):
1373         Added those 2 new helper functions.
1374
1375 2012-04-04  Andrei Burago  <aburago@chromium.org>
1376
1377         Auto-size may not work on first load
1378         https://bugs.webkit.org/show_bug.cgi?id=82989
1379
1380         Reviewed by David Levin.
1381
1382         No new tests. The only known repro scenario involves using chrome with a third-party extension,
1383         which makes making a test out of it problematic.
1384
1385         * dom/Document.cpp:
1386         (WebCore::Document::implicitClose):
1387         * page/FrameView.cpp:
1388         (WebCore::FrameView::autoSizeIfEnabled):
1389
1390 2012-04-04  Adam Barth  <abarth@webkit.org>
1391
1392         XSSAuditor doesn't catch injected srcdoc attributes
1393         https://bugs.webkit.org/show_bug.cgi?id=83238
1394
1395         Reviewed by Daniel Bates.
1396
1397         <iframe srcdoc> is an XSS injection vector because the srcdoc inherits
1398         the security origin of the parent page.  This patch updates the XSS
1399         auditor to check whether the attribute is injected using the same
1400         hueristics we use for inline event handlers.
1401
1402         Test: http/tests/security/xssAuditor/iframe-srcdoc.html
1403
1404         * html/parser/XSSAuditor.cpp:
1405         (WebCore::XSSAuditor::filterIframeToken):
1406         (WebCore::XSSAuditor::eraseDangerousAttributesIfInjected):
1407         (WebCore::XSSAuditor::eraseAttributeIfInjected):
1408         (WebCore::XSSAuditor::decodedSnippetForAttribute):
1409         * html/parser/XSSAuditor.h:
1410
1411 2012-04-04  Leo Yang  <leo.yang@torchmobile.com.cn>
1412
1413         [BlackBerry] Add local: protocol support in KURL for blackberry
1414         https://bugs.webkit.org/show_bug.cgi?id=82695
1415
1416         Reviewed by Rob Buis.
1417
1418         BlackBerry porting is using local: protocol just as file: protocol
1419         with the exception that it sandboxes the path to the application private space.
1420
1421         * platform/KURL.cpp:
1422         (WebCore::KURL::parse):
1423         (WebCore::portAllowed):
1424
1425 2012-04-04  Simon Fraser  <simon.fraser@apple.com>
1426
1427         https://bugs.webkit.org/show_bug.cgi?id=82994
1428
1429         Reviewed by James Robinson.
1430
1431         Fix an issue when removing elements with reflections from the document.
1432
1433         Test: compositing/reflections/remove-reflection.html
1434
1435         * platform/graphics/GraphicsLayer.cpp:
1436         (WebCore::GraphicsLayer::willBeDestroyed):
1437
1438 2012-04-04  Charles Wei  <charles.wei@torchmobile.com.cn>
1439
1440         Need using namespace std for system calls in MIMESniffing.cpp
1441         https://bugs.webkit.org/show_bug.cgi?id=82706
1442
1443         Reviewed by Antonio Gomes.
1444
1445         No new tests, just fix the compile error for BlackBerry porting.
1446
1447         * platform/network/MIMESniffing.cpp:
1448
1449 2012-04-04  Greg Billock  <gbillock@google.com>
1450
1451         Switch web intents API to be vendor-prefixed
1452         https://bugs.webkit.org/show_bug.cgi?id=83172
1453
1454         Reviewed by Adam Barth.
1455
1456         * Modules/intents/DOMWindowIntents.idl:
1457         * Modules/intents/NavigatorIntents.cpp:
1458         (WebCore::NavigatorIntents::webkitStartActivity):
1459         * Modules/intents/NavigatorIntents.h:
1460         (NavigatorIntents):
1461         * Modules/intents/NavigatorIntents.idl:
1462
1463 2012-04-04  Alexis Menard  <alexis.menard@openbossa.org>
1464
1465         [Part 4] We should use CSSPropertyID rather than integers when manipulating CSS property ids.
1466         https://bugs.webkit.org/show_bug.cgi?id=83224
1467
1468         Reviewed by Tony Chang.
1469
1470         CSSPropertyID enum holds all the CSS property ids but many parts of WebKit treat the ids
1471         as integers. While it's not incorrect it is nicer to use the enum as a parameter of
1472         functions manipulating property ids, as we ensure that the value passed will be an
1473         existing value. This patch clean up some remaining part of code.
1474
1475         No new tests : There should be no behavior change in this patch.
1476
1477         * css/CSSStyleSelector.cpp:
1478         (WebCore::CSSStyleSelector::collectMatchingRulesForList):
1479         * page/animation/AnimationBase.h:
1480         (WebCore::AnimationBase::affectsProperty):
1481         (WebCore::AnimationBase::isAnimatingProperty):
1482         * page/animation/CompositeAnimation.cpp:
1483         (WebCore::CompositeAnimation::getAnimationForProperty):
1484         (WebCore::CompositeAnimation::overrideImplicitAnimations):
1485         (WebCore::CompositeAnimation::resumeOverriddenImplicitAnimations):
1486         (WebCore::CompositeAnimation::isAnimatingProperty):
1487         * page/animation/CompositeAnimation.h:
1488         (CompositeAnimation):
1489         * page/animation/ImplicitAnimation.cpp:
1490         (WebCore::ImplicitAnimation::affectsProperty):
1491         * page/animation/ImplicitAnimation.h:
1492         (WebCore::ImplicitAnimation::transitionProperty):
1493         (WebCore::ImplicitAnimation::animatingProperty):
1494         (ImplicitAnimation):
1495         * page/animation/KeyframeAnimation.cpp:
1496         (WebCore::KeyframeAnimation::fetchIntervalEndpointsForProperty):
1497         (WebCore::KeyframeAnimation::hasAnimationForProperty):
1498         (WebCore::KeyframeAnimation::overrideAnimations):
1499         (WebCore::KeyframeAnimation::resumeOverriddenAnimations):
1500         (WebCore::KeyframeAnimation::affectsProperty):
1501         * page/animation/KeyframeAnimation.h:
1502         (KeyframeAnimation):
1503         * rendering/RenderLayerBacking.cpp:
1504         (WebCore::RenderLayerBacking::transitionPaused):
1505         (WebCore::RenderLayerBacking::transitionFinished):
1506         (WebCore::RenderLayerBacking::graphicsLayerToCSSProperty):
1507         (WebCore::RenderLayerBacking::cssToGraphicsLayerProperty):
1508         * rendering/RenderLayerBacking.h:
1509         (RenderLayerBacking):
1510
1511 2012-04-04  Rafael Weinstein  <rafaelw@chromium.org>
1512
1513         [MutationObservers] implement takeRecords()
1514         https://bugs.webkit.org/show_bug.cgi?id=83218
1515
1516         Reviewed by Ojan Vafai.
1517
1518         This patch implements MutationObserver.takeRecords per the DOM4 spec.
1519         takeRecords retrieves and clears any pending mutation records for
1520         the observer.
1521
1522         Test: fast/mutation/takeRecords.html
1523
1524         * dom/WebKitMutationObserver.cpp:
1525         (WebCore::WebKitMutationObserver::takeRecords):
1526         (WebCore):
1527         (WebCore::WebKitMutationObserver::deliver):
1528         * dom/WebKitMutationObserver.h:
1529         * dom/WebKitMutationObserver.idl:
1530
1531 2012-04-04  Shinya Kawanaka  <shinyak@chromium.org>
1532
1533         Shadow DOM is exposed in JS.
1534         https://bugs.webkit.org/show_bug.cgi?id=82607
1535
1536         Reviewed by Hajime Morita.
1537
1538         DOMSelection didn't consider nested shadow trees. This patch makes DOMSelection
1539         take nested shadow trees into account.
1540
1541         To test that the element is not in a shadow tree, Internals has a treeScopeRootNode method
1542         which returns the root node of the belonging tree scope.
1543
1544         Test: fast/dom/shadow/selection-shouldnt-expose-shadow-dom.html
1545
1546         * WebCore.exp.in:
1547         * page/DOMSelection.cpp:
1548         (WebCore::selectionShadowAncestor):
1549         (WebCore):
1550         * testing/Internals.cpp:
1551         (WebCore::Internals::treeScopeRootNode):
1552         (WebCore):
1553         * testing/Internals.h:
1554         (Internals):
1555         * testing/Internals.idl:
1556
1557 2012-04-04  Luke Macpherson  <macpherson@chromium.org>
1558
1559         Replace further usage of int with CSSPropertyID.
1560         https://bugs.webkit.org/show_bug.cgi?id=83119
1561
1562         Reviewed by Simon Fraser.
1563
1564         No new tests / no functionality changed.
1565
1566         * css/CSSComputedStyleDeclaration.cpp:
1567         (WebCore::logUnimplementedPropertyID):
1568         (WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):
1569         * css/CSSGrammar.y:
1570         * css/CSSParser.cpp:
1571         (WebCore::CSSParser::CSSParser):
1572         * css/CSSParser.h:
1573         (CSSParser):
1574         (WebCore::ShorthandScope::~ShorthandScope):
1575         * css/CSSProperty.h:
1576         (WebCore::CSSProperty::CSSProperty):
1577         * css/SVGCSSComputedStyleDeclaration.cpp:
1578         (WebCore::CSSComputedStyleDeclaration::getSVGPropertyCSSValue):
1579         * css/StylePropertySet.h:
1580         (StylePropertySet):
1581         * css/makeprop.pl:
1582         * dom/StyledElement.h:
1583         (StyledElement):
1584         * editing/EditingStyle.h:
1585         (EditingStyle):
1586         * html/HTMLElement.h:
1587         (HTMLElement):
1588
1589 2012-04-04  Adam Klein  <adamk@chromium.org>
1590
1591         Use PassRefPtr in V8DOMWrapper interface to avoid explicit ref() calls
1592         https://bugs.webkit.org/show_bug.cgi?id=82238
1593
1594         Reviewed by Adam Barth.
1595
1596         Relanding r112163 without modification, as it still seems valid.
1597         Will watch Chrome Canaries closely for any memory issues.
1598
1599         The setJSWrapper* methods previously featured a comment that asked
1600         callers to ref the objects before passing them in. This change makes
1601         that contract explicit (and allows the removal of the comment).
1602
1603         In addition, for ConstructorCallbacks, this change slightly reduces
1604         refcount churn by passing on the initial ref via RefPtr::release().
1605
1606         No new tests, no change in behavior.
1607
1608         * bindings/scripts/CodeGeneratorV8.pm:
1609         (GenerateConstructorCallback): Use RefPtr::release() to avoid refcount churn and remove explicit ref() call.
1610         (GenerateNamedConstructorCallback): ditto.
1611         * bindings/v8/V8DOMWindowShell.cpp:
1612         (WebCore::V8DOMWindowShell::installDOMWindow): Cast to a PassRefPtr and remove explicit ref call.
1613         * bindings/v8/V8DOMWrapper.cpp:
1614         (WebCore::V8DOMWrapper::setJSWrapperForDOMNode): Pass leaked refs into the DOMNodeMaps.
1615         * bindings/v8/V8DOMWrapper.h:
1616         (V8DOMWrapper): Make the setJSWrapperFor* methods take PassRefPtr<T>.
1617         (WebCore::V8DOMWrapper::setJSWrapperForDOMObject): Pass leaked ref into the DOMObjectMap.
1618         (WebCore::V8DOMWrapper::setJSWrapperForActiveDOMObject): Pass leaked ref into the ActiveDOMObjectMap.
1619         * bindings/v8/V8Proxy.h:
1620         (WebCore::toV8): Remove explicit ref.
1621         * bindings/v8/WorkerContextExecutionProxy.cpp:
1622         (WebCore::WorkerContextExecutionProxy::initContextIfNeeded): Cast to a PassRefPTr and remove explicit ref call.
1623         * bindings/v8/custom/V8HTMLImageElementConstructor.cpp:
1624         (WebCore::v8HTMLImageElementConstructorCallback): Use RefPtr::release() to avoid refcount churn and remove explicit ref.
1625         * bindings/v8/custom/V8WebKitMutationObserverCustom.cpp:
1626         (WebCore::V8WebKitMutationObserver::constructorCallback): ditto.
1627         * bindings/v8/custom/V8WebSocketCustom.cpp:
1628         (WebCore::V8WebSocket::constructorCallback): ditto.
1629         * bindings/v8/custom/V8XMLHttpRequestConstructor.cpp:
1630         (WebCore::V8XMLHttpRequest::constructorCallback): ditto.
1631
1632 2012-04-04  Chris Rogers  <crogers@google.com>
1633
1634         WaveTable::waveDataForFundamentalFrequency() should properly interpret negative frequency
1635         https://bugs.webkit.org/show_bug.cgi?id=83228
1636
1637         Reviewed by Kenneth Russell.
1638
1639         * Modules/webaudio/WaveTable.cpp:
1640         (WebCore::WaveTable::waveDataForFundamentalFrequency):
1641
1642 2012-04-04  Adam Klein  <adamk@chromium.org>
1643
1644         Delay post-insertion notifications until new DOM tree is complete
1645         https://bugs.webkit.org/show_bug.cgi?id=82631
1646
1647         Reviewed by Ojan Vafai.
1648
1649         When inserting a DocumentFragment, WebKit previously would update both
1650         internal WebCore state and mutation event listeners after each node
1651         was inserted. This is inconsistent not only with DOM4, but also
1652         with (at least) Firefox and IE. Given the many bugs over the years in
1653         WebKit due to this behavior, it seems better to delay notification
1654         until the fragment is completely inserted.
1655
1656         The changes to the three core mutation methods below are similar:
1657         the only logic remaining in the loop is checking that insertion is
1658         possible and taking care of that insertion. The entire loop is then
1659         wrapped in forbidEventDispatch/allowEventDispatch, effectively
1660         asserting that none of the code inside will have side effects.
1661
1662         The one bit of logic added to the loop is resizing the targets
1663         vector down to the set of nodes actually inserted as part of the
1664         loop. This makes it possible to simply pass the vector on to
1665         notifyChildrenInserted without having to also pass along a count of
1666         actually-inserted nodes.
1667
1668         As for the code that used to live inside the loop that could have
1669         side-effects, or depended on those side-effects, it has been moved
1670         out, either above (the check that the refChild is still valid in
1671         insertBefore) or after (the calls to notifyChildrenInserted).
1672
1673         Finally, it was necessary to retrofit ChildListMutationScope to take a
1674         vector of added nodes instead of a single node at a time, due to the
1675         assertions in isAddedNodeInOrder (now renamed to be plural). Note that
1676         there is now a single call to ChildListMutationScope::childrenAdded,
1677         inside notifyChildrenInserted.
1678
1679         Test: fast/events/domnodeinserted-entire-fragment.html
1680
1681         * dom/ChildListMutationScope.cpp:
1682         (ChildListMutationScope::MutationAccumulator): Renamed method to be plural.
1683         (WebCore::ChildListMutationScope::MutationAccumulator::areAddedNodesInOrder): Handle a NodeVector instead of a Node.
1684         (WebCore::ChildListMutationScope::MutationAccumulator::childrenAdded): Handle adding a NodeVector instead of a Node.
1685         (WebCore::ChildListMutationScope::MutationAccumulationRouter::childrenAdded): Renamed to be plural, pass NodeVector through.
1686         * dom/ChildListMutationScope.h:
1687         (WebCore::ChildListMutationScope::childrenAdded): ditto.
1688         (MutationAccumulationRouter):
1689         * dom/ContainerNode.cpp:
1690         (WebCore): Renamed updateTreeAfterInsertion to notifyChildrenInserted.
1691         (WebCore::ContainerNode::insertBefore): See main ChangeLog explanation.
1692         (WebCore::ContainerNode::replaceChild): ditto.
1693         (WebCore::ContainerNode::appendChild): ditto.
1694         (WebCore::dispatchChildInsertionEvents): Remove MutationObserver handling.
1695         (WebCore::notifyChildrenInserted): Handle a NodeVector of all inserted children,
1696         and take on responsiblity for MutationObserver handling as well as dispatchSubtreeModifiedEvent.
1697
1698 2012-04-04  Chris Rogers  <crogers@google.com>
1699
1700         RealtimeAnalyserNode should support smaller analysis sizes
1701         https://bugs.webkit.org/show_bug.cgi?id=83215
1702
1703         Reviewed by Kenneth Russell.
1704
1705         RealtimeAnalyserNode analysis size currently goes no lower than 128.
1706         It is useful to support lower power-of-two sizes.
1707         Decrease this limit to 32.
1708
1709         Updated webaudio/realtimeanalyser-fft-sizing-expected.txt
1710
1711         * Modules/webaudio/RealtimeAnalyser.cpp:
1712         (WebCore):
1713
1714 2012-04-04  Tony Chang  <tony@chromium.org>
1715
1716         CSS transitions should work on the flex property
1717         https://bugs.webkit.org/show_bug.cgi?id=75915
1718
1719         Reviewed by Dean Jackson.
1720
1721         Test: transitions/flex-transitions.html
1722
1723         * page/animation/AnimationBase.cpp:
1724         (WebCore::PropertyWrapperGetter::equals): Fix indention.
1725         (PropertyWrapperFlex): Add a wrapper for -webkit-flex which has 3 values (two floats and
1726         a length), but is not a shorthand.
1727         (WebCore::PropertyWrapperFlex::PropertyWrapperFlex):
1728         (WebCore::PropertyWrapperFlex::equals):
1729         (WebCore::PropertyWrapperFlex::blend):
1730         (WebCore):
1731         (WebCore::AnimationBase::ensurePropertyMap):
1732
1733 2012-04-04  Tom Sepez  <tsepez@chromium.org>
1734
1735         XSSAuditor bypass through HTTP Parameter Pollution.
1736         https://bugs.webkit.org/show_bug.cgi?id=81283
1737
1738         Reviewed by Adam Barth.
1739
1740         Deal with concatenation of multiple parameters via comma-splicing that
1741         is common to some webservers. We can no longer trust that all of the
1742         attributes of a reflected script tag, nor the reflected script itself,
1743         came from the same single URL parameter. The fix is to take commas into
1744         account when trucating the snippet used for matching.
1745
1746         Test: http/tests/security/xssAuditor/script-tag-with-comma.html
1747
1748         * html/parser/XSSAuditor.cpp:
1749         (WebCore::XSSAuditor::filterScriptToken):
1750         (WebCore):
1751         (WebCore::XSSAuditor::decodedSnippetForName):
1752         (WebCore::XSSAuditor::decodedSnippetForJavaScript):
1753
1754 2012-04-04  Dan Bernstein  <mitz@apple.com>
1755
1756         Paginated webviews render nothing in their gutters
1757         https://bugs.webkit.org/show_bug.cgi?id=83231
1758
1759         Reviewed by Adele Peterson.
1760
1761         When the RenderView has columns, nothing draws in the gutters.
1762
1763         * page/FrameView.cpp:
1764         (WebCore::FrameView::paintContents): Paint the background color behind everything when
1765         paginated.
1766
1767 2012-04-04  Mark Pilgrim  <pilgrim@chromium.org>
1768
1769         Call histogramEnumeration directly
1770         https://bugs.webkit.org/show_bug.cgi?id=83106
1771
1772         Reviewed by Adam Barth.
1773
1774         Part of a refactoring project to remove the PlatformSupport
1775         abstraction from some functions. See bug 82948.
1776
1777         * bindings/v8/custom/V8InspectorFrontendHostCustom.cpp:
1778         (WebCore::histogramEnumeration):
1779         * platform/chromium/HistogramSupportChromium.cpp:
1780         (WebCore::HistogramSupport::histogramEnumeration):
1781         * platform/chromium/PlatformSupport.h:
1782         (PlatformSupport):
1783
1784 2012-04-04  Zhenyao Mo  <zmo@google.com>
1785
1786         Unreviewed, GPU bots build fix.
1787
1788         Commited on behalf of sievers@chromium.org.
1789
1790         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
1791         (WebCore::CCLayerTreeHostImpl::canDraw):
1792
1793 2012-04-04  Abhishek Arya  <inferno@chromium.org>
1794
1795         Add helpers to create anonymous table parts.
1796         https://bugs.webkit.org/show_bug.cgi?id=83116
1797
1798         Reviewed by Julien Chaffraix.
1799
1800         The patch introduces helpers to create anonymous table parts by
1801         introducing a new static function createAnonymousWithParentRenderer.
1802         The function builds a new anonymous wrapper of the same type as the class,
1803         inheriting style properties from parent and sets a display based on
1804         argument/default values. Also we streamline the RenderBlock functions
1805         to match this naming convention.
1806
1807         * rendering/RenderBlock.cpp:
1808         (WebCore::RenderBlock::createAnonymousBlockWithSameTypeAs):
1809         (WebCore::RenderBlock::createAnonymousWithParentRendererAndDisplay):
1810         (WebCore):
1811         (WebCore::RenderBlock::createAnonymousColumnsWithParentRendererAndDisplay):
1812         (WebCore::RenderBlock::createAnonymousColumnSpanWithParentRendererAndDisplay):
1813         * rendering/RenderBlock.h:
1814         (RenderBlock):
1815         (WebCore::RenderBlock::createAnonymousBlock):
1816         (WebCore::RenderBlock::createAnonymousColumnsBlock):
1817         (WebCore::RenderBlock::createAnonymousColumnSpanBlock):
1818         * rendering/RenderButton.cpp:
1819         (WebCore::RenderButton::addChild):
1820         * rendering/RenderObject.cpp:
1821         (WebCore::RenderObject::addChild):
1822         * rendering/RenderTable.cpp:
1823         (WebCore::RenderTable::addChild):
1824         (WebCore::RenderTable::createAnonymousWithParentRendererAndDisplay):
1825         (WebCore):
1826         * rendering/RenderTable.h:
1827         (RenderTable):
1828         * rendering/RenderTableCell.cpp:
1829         (WebCore::RenderTableCell::createAnonymousWithParentRendererAndDisplay):
1830         (WebCore):
1831         * rendering/RenderTableCell.h:
1832         (RenderTableCell):
1833         * rendering/RenderTableRow.cpp:
1834         (WebCore::RenderTableRow::addChild):
1835         (WebCore::RenderTableRow::createAnonymousWithParentRendererAndDisplay):
1836         (WebCore):
1837         * rendering/RenderTableRow.h:
1838         (RenderTableRow):
1839         * rendering/RenderTableSection.cpp:
1840         (WebCore::RenderTableSection::addChild):
1841         (WebCore::RenderTableSection::createAnonymousWithParentRendererAndDisplay):
1842         (WebCore):
1843         * rendering/RenderTableSection.h:
1844         (RenderTableSection):
1845
1846 2012-04-04  Erik Arvidsson  <arv@chromium.org>
1847
1848         [V8] Add a per context data store and use that for caching boiler plates as well as constructor functions
1849         https://bugs.webkit.org/show_bug.cgi?id=83093
1850
1851         Reviewed by Adam Barth.
1852
1853         This introduces a V8BindingPerContextData class. After creating a new v8::Context we also create a new
1854         per context data instance. This instance maintains the boiler plate and constructor caches. V8DOMWrapper
1855         now delegates the caching to this new class. Previously the caching was only done in V8DOMWindowShell which
1856         made workers and isolated worlds behave slightly different.
1857
1858         Tests: fast/dom/constructor-proto.html
1859                fast/workers/constructor-proto.html
1860
1861         * WebCore.gypi:
1862         * bindings/scripts/CodeGeneratorV8.pm:
1863         (GenerateConstructorGetter): Renamed getConstructor to constructorForType.
1864         * bindings/v8/V8BindingPerContextData.cpp: Added.
1865         (WebCore):
1866         (WebCore::V8BindingPerContextData::dispose): This clears the persistent handles stored in the maps.
1867         (WebCore::V8BindingPerContextData::init): This installs the hidden prototype which is used as the
1868         [[Prototype]] for the constructor functions (for HotMail compat).
1869         (WebCore::V8BindingPerContextData::createWrapperFromCacheSlowCase): Instantiates a new object and
1870         stores it in the boilerplate map, and returns a clone.
1871         (WebCore::V8BindingPerContextData::constructorForTypeSlowCase): Creates the function for the constructor
1872         and stores it in the constructor map.
1873         * bindings/v8/V8BindingPerContextData.h: Added.
1874         (WebCore):
1875         (V8BindingPerContextData):
1876         (WebCore::V8BindingPerContextData::create):
1877         (WebCore::V8BindingPerContextData::~V8BindingPerContextData):
1878         (WebCore::V8BindingPerContextData::createWrapperFromCache): Checks the cache and clones the wrapper in
1879         the cache. If no boiler plate exists one is created and cached.
1880         (WebCore::V8BindingPerContextData::constructorForType): If the constructor already exists the cached
1881         constructor function is returned, otherwise a new constructor function is created and cached.
1882         (WebCore::V8BindingPerContextData::V8BindingPerContextData):
1883         * bindings/v8/V8DOMWindowShell.cpp:
1884         (WebCore::V8DOMWindowShell::disposeContextHandles): Clear the per context data when disposing the handles.
1885         (WebCore::V8DOMWindowShell::initContextIfNeeded): Create and initialize the per context data.
1886         (WebCore::V8DOMWindowShell::installDOMWindow):
1887         * bindings/v8/V8DOMWindowShell.h:
1888         (V8DOMWindowShell):
1889         (WebCore::V8DOMWindowShell::perContextData):
1890         * bindings/v8/V8DOMWrapper.cpp:
1891         (WebCore::V8DOMWrapper::constructorForType): Static method to get the constructor. The per context data
1892         is determined based on the type of the parameters.
1893         (WebCore):
1894         (WebCore::V8DOMWrapper::perContextData):
1895         (WebCore::V8DOMWrapper::instantiateV8Object): Use the per context data to unify the code paths.
1896         * bindings/v8/V8DOMWrapper.h:
1897         (WebCore):
1898         (V8DOMWrapper):
1899         * bindings/v8/V8IsolatedContext.cpp:
1900         (WebCore::V8IsolatedContext::V8IsolatedContext): Create and initialize the per context data.
1901         (WebCore::V8IsolatedContext::destroy): Clear the per context data.
1902         * bindings/v8/V8IsolatedContext.h:
1903         (WebCore):
1904         (WebCore::V8IsolatedContext::perContextData):
1905         (V8IsolatedContext):
1906         * bindings/v8/V8Proxy.cpp:
1907         (WebCore::V8Proxy::retrievePerContextData):
1908         (WebCore):
1909         * bindings/v8/V8Proxy.h:
1910         (WebCore):
1911         (V8Proxy):
1912         * bindings/v8/WorkerContextExecutionProxy.cpp:
1913         (WebCore::WorkerContextExecutionProxy::dispose): Clear the per context data
1914         (WebCore::WorkerContextExecutionProxy::initContextIfNeeded): Create and initialize the per context data.
1915         * bindings/v8/WorkerContextExecutionProxy.h:
1916         (WebCore):
1917         (WebCore::WorkerContextExecutionProxy::perContextData):
1918         (WorkerContextExecutionProxy):
1919
1920 2012-04-04  Shawn Singh  <shawnsingh@chromium.org>
1921
1922         [chromium] Move recursive renderSurface clearing to CCLayerTreeHostImpl
1923         https://bugs.webkit.org/show_bug.cgi?id=82091
1924
1925         Reviewed by James Robinson.
1926
1927         No new tests needed, minor refactoring covered by existing tests.
1928
1929         This patch is just a minor cleanup, moving clearRenderSurfacesOnCCLayerImplRecursive()
1930         from LayerRendererChromium to CCLayerTreeHostImpl. It makes more sense to place the code
1931         there, so that LayerRendererChromium is more like a blind utility for drawing things
1932         while CCLayerTreeHostImpl actually manages the state of the resources.
1933
1934         * platform/graphics/chromium/LayerRendererChromium.cpp:
1935         (WebCore::LayerRendererChromium::beginDrawingFrame):
1936         * platform/graphics/chromium/LayerRendererChromium.h:
1937         (LayerRendererChromium):
1938         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
1939         (WebCore::CCLayerTreeHostImpl::~CCLayerTreeHostImpl):
1940         (WebCore::CCLayerTreeHostImpl::initializeLayerRenderer):
1941         (WebCore::CCLayerTreeHostImpl::sendDidLoseContextRecursive):
1942         (WebCore::CCLayerTreeHostImpl::clearRenderSurfacesOnCCLayerImplRecursive):
1943         (WebCore):
1944         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.h:
1945         (WebCore::CCLayerTreeHostImpl::rootLayer):
1946         (CCLayerTreeHostImpl):
1947
1948 2012-04-04  Chris Rogers  <crogers@google.com>
1949
1950         Web Audio should use MutexTryLocker class
1951         https://bugs.webkit.org/show_bug.cgi?id=83194
1952
1953         Reviewed by Kenneth Russell.
1954
1955         Switch existing Web Audio code from directly calling tryLock() on a Mutex to use a MutexTryLocker.
1956         No new tests since it is a low-level threading primitive and is difficult to test.
1957         Existing Web Audio tests continue to test the process() methods affected.
1958
1959         * Modules/webaudio/AudioBufferSourceNode.cpp:
1960         (WebCore::AudioBufferSourceNode::process):
1961         * Modules/webaudio/AudioParamTimeline.cpp:
1962         (WebCore::AudioParamTimeline::valuesForTimeRange):
1963         * Modules/webaudio/ConvolverNode.cpp:
1964         (WebCore::ConvolverNode::process):
1965         * Modules/webaudio/MediaElementAudioSourceNode.cpp:
1966         (WebCore::MediaElementAudioSourceNode::process):
1967         * Modules/webaudio/Oscillator.cpp:
1968         (WebCore::Oscillator::process):
1969         * Modules/webaudio/WaveShaperProcessor.cpp:
1970         (WebCore::WaveShaperProcessor::process):
1971
1972 2012-04-04  Raphael Kubo da Costa  <rakuco@webkit.org>
1973
1974         [CSS] Make makevalues.pl and makeprop.pl ignore '#'s.
1975         https://bugs.webkit.org/show_bug.cgi?id=83212
1976
1977         Reviewed by Tony Chang.
1978
1979         This patch is a side-effect of my intentions to make the
1980         CMake-based ports work with GCC 4.7 without changing the
1981         parameters currently passed to the preprocessor. "-P" is always
1982         being passed to the preprocessor, but this breaks
1983         dom/make_names.pl due to GCC not outputting empty lines with "-P"
1984         anymore; if I switch to never passing "-P" (my aim), the CSS
1985         scripts currently break because of the additional lines output by
1986         the preprocessor. InFilesParser.pm and make-css-file-array.pl
1987         already skip lines which start with '#', for example.
1988
1989         No new tests, tools plumbing.
1990
1991         * css/makeprop.pl:
1992         * css/makevalues.pl:
1993
1994 2012-04-04  Min Qin  <qinmin@google.com>
1995
1996         Disable fancy upsampling and dithering for decoding jpeg on android
1997         https://bugs.webkit.org/show_bug.cgi?id=83196
1998
1999         Reviewed by Kenneth Russell.
2000
2001         Dithering and fancy upsampling are currently disabled for chrome on android.
2002         This gives us about 20% performance improvement.
2003         Since the screen of mobile devices is small, impact on image quality is limited.
2004         This change does not introduce any changes on other platforms.
2005
2006         * platform/image-decoders/jpeg/JPEGImageDecoder.cpp:
2007         (ditherMode):
2008         (doFancyUpsampling):
2009         (WebCore::JPEGImageReader::decode):
2010
2011 2012-04-04  Daniel Sievers  <sievers@chromium.org>
2012
2013         [Chromium] Always skip draw and readback if there is nothing
2014         to draw.
2015         https://bugs.webkit.org/show_bug.cgi?id=82680
2016
2017         This avoids corruption from pushing frames that have no valid
2018         content drawn into them.
2019         Also in addition to checking for non-existing root layers, check
2020         for root layers with no content bounds. It's possible to see those
2021         with kForceCompositing mode for empty documents.
2022
2023         Reviewed by James Robinson.
2024
2025         Added CCLayerTreeHostTestEmptyContentsShouldNotDraw.
2026
2027         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
2028         (WebCore::CCLayerTreeHostImpl::canDraw):
2029         (WebCore::CCLayerTreeHostImpl::prepareToDraw):
2030         * platform/graphics/chromium/cc/CCThreadProxy.cpp:
2031         (WebCore::CCThreadProxy::scheduledActionDrawAndSwapInternal):
2032
2033 2012-03-15  Jer Noble  <jer.noble@apple.com>
2034
2035         Full Screen mode should cancel before navigation.
2036         https://bugs.webkit.org/show_bug.cgi?id=81295
2037
2038         Reviewed by Anders Carlsson.
2039
2040         No new tests; protect against speculative crasher when a bad client calls the below
2041         functions at inopportune times.
2042
2043         Check that the document is not either detached or in the page cache, and if so, bail out
2044         early:
2045         * dom/Document.cpp:
2046         (WebCore::Document::webkitWillEnterFullScreenForElement):
2047         (WebCore::Document::webkitDidEnterFullScreenForElement):
2048         (WebCore::Document::webkitWillExitFullScreenForElement):
2049         (WebCore::Document::webkitDidExitFullScreenForElement):
2050
2051 2012-04-04  Emil A Eklund  <eae@chromium.org>
2052
2053         Fix getFilterOutsets parameter types in RenderLayer
2054         https://bugs.webkit.org/show_bug.cgi?id=83041
2055
2056         Reviewed by Julien Chaffraix.
2057
2058         The getFilterOutsets method was changed to take integers in r112475, update
2059         RenderLayer to use the right data type for the outset arguments.
2060
2061         No new tests, no change in functionality.
2062
2063         * rendering/RenderLayer.cpp:
2064         (WebCore::RenderLayer::setFilterBackendNeedsRepaintingInRect):
2065         (WebCore::RenderLayer::calculateLayerBounds):
2066
2067 2012-04-04  Erik Arvidsson  <arv@chromium.org>
2068
2069         Rebaseline binding tests.
2070
2071         Unreviewed.
2072
2073         * bindings/scripts/test/V8/V8TestInterface.cpp:
2074         (WebCore::V8TestInterface::constructorCallback):
2075         * bindings/scripts/test/V8/V8TestNamedConstructor.cpp:
2076         (WebCore::V8TestNamedConstructorConstructorCallback):
2077         * bindings/scripts/test/V8/V8TestObj.cpp:
2078         (WebCore::V8TestObj::constructorCallback):
2079         * bindings/scripts/test/V8/V8TestSerializedScriptValueInterface.cpp:
2080         (WebCore::V8TestSerializedScriptValueInterface::constructorCallback):
2081
2082 2012-04-04  Dmitry Lomov  <dslomov@google.com>
2083
2084         [JSC] ArrayBufferView and its ArrayBuffer are appended to object pool in wrong order
2085         https://bugs.webkit.org/show_bug.cgi?id=82090
2086         The implementation of structured cloning algorithm (http://www.w3.org/TR/html5/common-dom-interfaces.html#internal-structured-cloning-algorithm)
2087         in SerializedScriptValue.cpp assigns numerical identifiers to encontered objects as it traverses
2088         the cloned object during serialization.
2089         When the cloning encounters an already seen object, it transfers the assigned numerical id
2090         instead of cloning the object again. Deserialization process then repeats the process in
2091         the mirror fashion, i.e. on deserializing the object it assigns deserialized object a numeric id and if it
2092         deserializes the id it substitutes the perviously deserialized objects. It is critical that serialization and deserialization
2093         assigns numeric ids in the same order.
2094
2095         The bug (discovered by Yong Li) is that when serializing ArrayBufferView, the ids were assigned first to
2096         the ArrayBufferView and then to underlying ArrayBuffer; however on deserialization the ids were assigned another way round.
2097
2098         This patch fixes that by assigning the id first to ArrayBuffer and then to ArrayBufferView, and adds corresponding test cases.
2099
2100         Reviewed by Kenneth Russell.
2101
2102         New test cases added to fast/canvas/web-gl/array-message-passing.html.
2103
2104         * bindings/js/SerializedScriptValue.cpp:
2105         (WebCore::CloneSerializer::checkForDuplicate):
2106         (CloneSerializer):
2107         (WebCore::CloneSerializer::recordObject):
2108         (WebCore::CloneSerializer::startObjectInternal):
2109         (WebCore::CloneSerializer::dumpIfTerminal):
2110
2111 2012-04-04  Ian Vollick  <vollick@chromium.org>
2112
2113         [chromium] When setting animation started events, should check the root layer
2114         https://bugs.webkit.org/show_bug.cgi?id=83060
2115
2116         Reviewed by Adrienne Walker.
2117
2118         No new tests.
2119
2120         * platform/graphics/chromium/cc/CCLayerTreeHost.cpp:
2121         (WebCore::CCLayerTreeHost::animateLayers):
2122         (WebCore::CCLayerTreeHost::animateLayersRecursive):
2123         (WebCore::CCLayerTreeHost::setAnimationEventsRecursive):
2124
2125 2012-04-04  Adam Klein  <adamk@chromium.org>
2126
2127         Web Inspector: break on DOM node insertion only once per operation, not once per inserted node
2128         https://bugs.webkit.org/show_bug.cgi?id=82967
2129
2130         Reviewed by Ojan Vafai.
2131
2132         This change affects the case where a DocumentFragment is inserted,
2133         rather than a single node. This is most common when using innerHTML:
2134         the effect of the change is that inserting, e.g., '<input><input>',
2135         the SubtreeModified breakpoint will be hit once, rather than twice
2136         (once for each input element). Given that the particular node being
2137         inserted wasn't exposed as part of the breakpoint, this seems strictly
2138         better.
2139
2140         Now that the call to willInsertDOMNode is outside the loop, there's
2141         not an obvious node to pass in as the new child. Luckily, InspectorDOMDebuggerAgent
2142         already ignored that argument, so it's simply been removed from the signature.
2143
2144         This changes paves the way to do only tree-modification work, and no
2145         external notifications, inside the loops in appendChild/insertBefore/replaceChild.
2146
2147         * dom/ContainerNode.cpp:
2148         (WebCore::ContainerNode::insertBefore): Hoisted call to willInsertDOMNode out of loop.
2149         (WebCore::ContainerNode::replaceChild): ditto.
2150         (WebCore::ContainerNode::appendChild): ditto.
2151         * inspector/InspectorDOMDebuggerAgent.cpp:
2152         (WebCore::InspectorDOMDebuggerAgent::willInsertDOMNode): Removed first argument (now takes only the parent).
2153         * inspector/InspectorDOMDebuggerAgent.h:
2154         (InspectorDOMDebuggerAgent):
2155         * inspector/InspectorInstrumentation.cpp:
2156         (WebCore::InspectorInstrumentation::willInsertDOMNodeImpl): Removed second argument.
2157         * inspector/InspectorInstrumentation.h:
2158         (InspectorInstrumentation):
2159         (WebCore::InspectorInstrumentation::willInsertDOMNode): Removed second argument.
2160
2161 2012-04-04  Jeffrey Pfau  <jpfau@apple.com>
2162
2163         Move pending sheet removal from ~HTMLLinkElement to removal from document.
2164         https://bugs.webkit.org/show_bug.cgi?id=69184
2165
2166         Reviewed by Adam Barth.
2167
2168         Test: fast/html/pending-stylesheet-crash.html
2169
2170         * html/HTMLLinkElement.cpp:
2171         (WebCore::HTMLLinkElement::~HTMLLinkElement):
2172         (WebCore::HTMLLinkElement::removedFromDocument):
2173
2174 2012-04-04  Mark Pilgrim  <pilgrim@chromium.org>
2175
2176         Call histogramCustomCounts directly
2177         https://bugs.webkit.org/show_bug.cgi?id=83112
2178
2179         Reviewed by Adam Barth.
2180
2181         * platform/chromium/PlatformSupport.h:
2182         (PlatformSupport):
2183         * platform/graphics/chromium/ContentLayerChromium.cpp:
2184         (WebCore::ContentLayerPainter::paint):
2185         * platform/graphics/chromium/cc/CCOverdrawMetrics.cpp:
2186         (WebCore::CCOverdrawMetrics::recordMetricsInternal):
2187
2188 2012-04-04  Alexis Menard  <alexis.menard@openbossa.org>
2189
2190         Animation related classes should use CSSPropertyID rather than integers when manipulating CSS property ids.
2191         https://bugs.webkit.org/show_bug.cgi?id=83050
2192
2193         Reviewed by Simon Fraser.
2194
2195         Split the concept of animate none and animate all into a separate function rather than using the property member
2196         with some magic value set on it. Make also sure that we use CSSPropertyID all over the place rather than integers.
2197         We still have some place where convert from int to CSSPropertyID (mostly due to the fact that we need to make Hash*
2198         related classes to handle CSSPropertyID). While refactoring I also removed some dead code.
2199
2200         No new tests : not behaviour changes intented.
2201
2202         * css/CSSComputedStyleDeclaration.cpp:
2203         (WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):
2204         * css/CSSStyleApplyProperty.cpp:
2205         (WebCore::ApplyPropertyAnimation::applyInheritValue):
2206         (WebCore::CSSStyleApplyProperty::CSSStyleApplyProperty):
2207         * css/CSSStyleSelector.cpp:
2208         (WebCore::CSSStyleSelector::collectMatchingRulesForList):
2209         * page/animation/AnimationBase.cpp:
2210         (WebCore):
2211         (WebCore::PropertyWrapperBase::PropertyWrapperBase):
2212         (WebCore::PropertyWrapperBase::property):
2213         (PropertyWrapperBase):
2214         (WebCore::PropertyWrapperGetter::PropertyWrapperGetter):
2215         (WebCore::PropertyWrapper::PropertyWrapper):
2216         (WebCore::RefCountedPropertyWrapper::RefCountedPropertyWrapper):
2217         (WebCore::StyleImagePropertyWrapper::StyleImagePropertyWrapper):
2218         (WebCore::PropertyWrapperColor::PropertyWrapperColor):
2219         (WebCore::PropertyWrapperShadow::PropertyWrapperShadow):
2220         (WebCore::PropertyWrapperMaybeInvalidColor::PropertyWrapperMaybeInvalidColor):
2221         (WebCore::PropertyWrapperVisitedAffectedColor::PropertyWrapperVisitedAffectedColor):
2222         (WebCore::FillLayersPropertyWrapper::FillLayersPropertyWrapper):
2223         (WebCore::ShorthandPropertyWrapper::ShorthandPropertyWrapper):
2224         (WebCore::PropertyWrapperSVGPaint::PropertyWrapperSVGPaint):
2225         (WebCore::wrapperForProperty):
2226         (WebCore::AnimationBase::propertiesEqual):
2227         The code removed here could never be reached. propertiesEqual is used only in CompositeAnimation::updateTransitions and through
2228         ImplicitAnimation::isTargetPropertyEqual which is also called in CompositeAnimation. Both calls are protected by an ASSERT in
2229         CompositeAnimation:116 where we make sure the property id we're dealing with is a valid property id, therefore prop can't be equal to
2230         the old value cAnimateAll or even the new mode concept introduced in that patch can't be AnimateAll.
2231         (WebCore::AnimationBase::getPropertyAtIndex):
2232         (WebCore::AnimationBase::blendProperties):
2233         (WebCore::AnimationBase::animationOfPropertyIsAccelerated):
2234         (WebCore::gatherEnclosingShorthandProperties):
2235         (WebCore::AnimationBase::animatableShorthandsAffectingProperty):
2236         * page/animation/AnimationBase.h:
2237         (AnimationBase):
2238         * page/animation/CompositeAnimation.cpp:
2239         (WebCore::CompositeAnimation::updateTransitions):
2240         (WebCore::CompositeAnimation::pauseTransitionAtTime):
2241         * page/animation/CompositeAnimation.h:
2242         (CompositeAnimation):
2243         * page/animation/ImplicitAnimation.cpp:
2244         (WebCore::ImplicitAnimation::ImplicitAnimation):
2245         (WebCore::ImplicitAnimation::sendTransitionEvent):
2246         The check removed is useless as ImplicitAnimation constructor ensure via an ASSERT that the animatingProperty value is a correct and valid
2247         property id. Even the old code was ensuring the value of animatingProperty was different from cAnimateAll at construction time. As stated also
2248         in CompositeAnimation::updateTransitions and cover by the same ASSERT explained earlier, ImplicitAnimation are constructed for valid property ids.
2249         (WebCore::ImplicitAnimation::isTargetPropertyEqual):
2250         (WebCore::ImplicitAnimation::blendPropertyValueInStyle):
2251         * page/animation/ImplicitAnimation.h:
2252         (WebCore::ImplicitAnimation::create):
2253         (ImplicitAnimation):
2254         * page/animation/KeyframeAnimation.cpp:
2255         (WebCore::KeyframeAnimation::animate):
2256         (WebCore::KeyframeAnimation::getAnimatedStyle):
2257         (WebCore::KeyframeAnimation::timeToNextService):
2258         * platform/animation/Animation.cpp:
2259         (WebCore::Animation::Animation):
2260         (WebCore::Animation::operator=):
2261         (WebCore::Animation::animationsMatch):
2262         * platform/animation/Animation.h:
2263         (WebCore::Animation::property):
2264         (WebCore::Animation::animationMode):
2265         (WebCore::Animation::setProperty):
2266         (WebCore::Animation::setAnimationMode):
2267         (Animation):
2268         (WebCore::Animation::initialAnimationProperty):
2269         * rendering/RenderLayerBacking.cpp:
2270         (WebCore::RenderLayerBacking::startTransition):
2271         * rendering/RenderLayerBacking.h:
2272         (RenderLayerBacking):
2273         * rendering/style/RenderStyle.cpp:
2274         (WebCore::RenderStyle::transitionForProperty):
2275         * rendering/style/RenderStyle.h:
2276
2277 2012-04-04  Emil A Eklund  <eae@chromium.org>
2278
2279         Fix usage of LayoutUnits and snapping for scrolling in RenderBox
2280         https://bugs.webkit.org/show_bug.cgi?id=83073
2281
2282         Reviewed by Eric Seidel.
2283
2284         Fix usage of LayoutUnits and snapping/rounding logic in RenderBox.
2285
2286         No new tests, no change in functionality.
2287
2288         * rendering/RenderBox.cpp:
2289         (WebCore::RenderBox::scrollWidth):
2290         Change scrollWidth to properly pixel snap values.
2291
2292         (WebCore::RenderBox::scrollHeight):
2293         Change scrollHeight to properly pixel snap values.
2294
2295         (WebCore::RenderBox::scrolledContentOffset):
2296         Change to return IntSize to match function definition.
2297
2298         (WebCore::RenderBox::cachedSizeForOverflowClip):
2299         Changed cachedSizeForOverflowClip to LayoutSize as the overflow and clip
2300         rects all have subpixel precision.
2301
2302         * rendering/RenderBox.h:
2303         (WebCore::RenderBox::pixelSnappedWidth):
2304         (WebCore::RenderBox::pixelSnappedHeight):
2305         Removed FIXME comment as the implementation
2306
2307         (WebCore::RenderBox::minYLayoutOverflow):
2308         (WebCore::RenderBox::maxYLayoutOverflow):
2309         (WebCore::RenderBox::minXLayoutOverflow):
2310         (WebCore::RenderBox::maxXLayoutOverflow):
2311         Added static_cast for border values.
2312
2313         (WebCore::RenderBox::hasVisualOverflow):
2314         Changed to compare two pixel snapped values as we don't want to indicate
2315         overflow in cases where the the size is rounded down resulting in no
2316         visual overflow.
2317
2318 2012-04-04  Emil A Eklund  <eae@chromium.org>
2319
2320         Fix types for location, size and rect calculations for render objects
2321         https://bugs.webkit.org/show_bug.cgi?id=83089
2322
2323         Reviewed by Eric Seidel.
2324
2325         Fix usage of LayoutUnits and rounding for a couple of different render
2326         object classes.
2327
2328         No new tests, no change in functionality.
2329
2330         * rendering/RenderDetailsMarker.cpp:
2331         (WebCore::RenderDetailsMarker::getPath):
2332         Change getPath to take a LayoutPoint as the transform has subpixel
2333         precision already.
2334
2335         * rendering/RenderFlowThread.cpp:
2336         (WebCore::RenderFlowThread::computeLogicalHeight):
2337         Change logicalHeight to LayoutUnit as it is computed from subpixel
2338         values.
2339
2340         * rendering/RenderInputSpeech.cpp:
2341         (WebCore::RenderInputSpeech::paintInputFieldSpeechButton):
2342         Change button rect computation to LayoutRect and pixel snap just before
2343         painting to preserve precision.
2344
2345         * rendering/RenderLineBoxList.cpp:
2346         (WebCore::RenderLineBoxList::rangeIntersectsRect):
2347         (WebCore::RenderLineBoxList::lineIntersectsDirtyRect):
2348         (WebCore::RenderLineBoxList::paint):
2349         Change range calculations to LayoutUnits to preserve precision.
2350
2351         * rendering/RenderMarquee.cpp:
2352         (WebCore::RenderMarquee::computePosition):
2353         Change width calculations to LayoutUnits to preserve precision.
2354
2355         * rendering/RenderTable.cpp:
2356         (WebCore::RenderTable::layoutCaption):
2357         Change table captions to LayoutUnits as the values are computed from
2358         subpixel componenets.
2359
2360         * rendering/style/RenderStyle.cpp:
2361         (WebCore::RenderStyle::getRoundedBorderFor):
2362         Snap border rect as RoundedRects use ints for crisp rendering.
2363
2364         * rendering/style/ShadowData.cpp:
2365         (WebCore::ShadowData::adjustRectForShadow):
2366         Change adjustRectForShadow to take a LayoutRect as it already uses
2367         LayoutUnits.
2368
2369 2012-04-04  Emil A Eklund  <eae@chromium.org>
2370
2371         Convert RootInlineBox to LayoutUnits in preparation for turning on subpixel layout
2372         https://bugs.webkit.org/show_bug.cgi?id=83054
2373
2374         Reviewed by Eric Seidel.
2375
2376         Convert RootInlineBox over to LayoutUnits, this mostly involves updating
2377         the alignment and adjustment code to be subpixel aware.
2378
2379         No new tests, no change in functionality.
2380
2381         * rendering/RootInlineBox.cpp:
2382         (WebCore::RootInlineBox::alignBoxesInBlockDirection):
2383         Change beforeAnnotationsAdjustment to LayoutUnit.
2384
2385         (WebCore::RootInlineBox::beforeAnnotationsAdjustment):
2386         Change method to return LayoutUnit as it is computed from values with
2387         subpixel precision.
2388
2389         (WebCore::RootInlineBox::lineSnapAdjustment):
2390         Round values before computing remainder.
2391
2392         (WebCore::RootInlineBox::ascentAndDescentForBox):
2393         Change ascent and decent calculation to use LayoutUnits as they are
2394         computed from values with subpixel precision.
2395
2396         (WebCore::RootInlineBox::verticalPositionForBox):
2397         Change verticalPosition to LayoutUnit.
2398
2399         * rendering/RootInlineBox.h:
2400         (RootInlineBox):
2401
2402 2012-04-05  Joseph Pecoraro  <pecoraro@apple.com>
2403
2404         <http://webkit.org/b/83108> Web Inspector: JSC Crash inspecting node with object event listener
2405
2406         Reviewed by Geoff Garen.
2407
2408         * bindings/js/ScriptEventListener.cpp:
2409         (WebCore::eventListenerHandlerLocation):
2410         Use jsDynamicCast to bail when the JSObject is not a JSFunction.
2411
2412 2012-04-04  Mark Pilgrim  <pilgrim@chromium.org>
2413
2414         Call suddenTerminationChanged directly
2415         https://bugs.webkit.org/show_bug.cgi?id=83114
2416
2417         Reviewed by Adam Barth.
2418
2419         * platform/chromium/PlatformSupport.h:
2420         (PlatformSupport):
2421         * platform/chromium/SuddenTerminationChromium.cpp:
2422         (WebCore::disableSuddenTermination):
2423         (WebCore::enableSuddenTermination):
2424
2425 2012-04-04  Adam Barth  <abarth@webkit.org>
2426
2427         figure out how to export webcore symbols from webkit.dll properly
2428         https://bugs.webkit.org/show_bug.cgi?id=83105
2429
2430         Reviewed by Dimitri Glazkov.
2431
2432         * WebCore.gypi:
2433             - Separate the platform support files into a new GYP variable so
2434               that they can be build as part of webkit.dll rather than in
2435               webcore_platform.lib. Building them as part of webkit.dll ensure
2436               that the symbols marked for export will actually be in webkit.dll
2437               (and hence exported), even if nothing else in webkit.dll
2438               references the obj file that contains the symbol.
2439         * platform/chromium/support/WebMediaStreamSourcesRequest.cpp:
2440             - Unwind our ugly, ugly hack to make the Windows component build
2441               work.
2442
2443 2012-04-04  Mark Pilgrim  <pilgrim@chromium.org>
2444
2445         Call sandboxEnabled directly
2446         https://bugs.webkit.org/show_bug.cgi?id=83113
2447
2448         Reviewed by Adam Barth.
2449
2450         * platform/chromium/PlatformSupport.h:
2451         (PlatformSupport):
2452
2453 2012-04-04  Peter Rybin  <peter.rybin@gmail.com>
2454
2455         Web Inspector: CodeGeneratorInspector.py: switch IndexedDB, DOMStorage, ApplicationCache domains to typed API
2456         https://bugs.webkit.org/show_bug.cgi?id=83039
2457
2458         Reviewed by Pavel Feldman.
2459
2460         Client code is switched to typed API (all InspectorObject and InspectorArray types are
2461         replaced with generated types from TypeBuilder according to Inspector.json).
2462
2463         Missing array of array specialization is added. Inspector.json is fixed to comform
2464         with actual behavior of InspectorDOMStorageAgent.
2465
2466         Output parameter initialization is assured.
2467
2468         * inspector/CodeGeneratorInspector.py:
2469         * inspector/Inspector.json:
2470         * inspector/InspectorApplicationCacheAgent.cpp:
2471         (WebCore::InspectorApplicationCacheAgent::getFramesWithManifests):
2472         (WebCore::InspectorApplicationCacheAgent::getApplicationCacheForFrame):
2473         (WebCore::InspectorApplicationCacheAgent::buildObjectForApplicationCache):
2474         * inspector/InspectorApplicationCacheAgent.h:
2475         (InspectorApplicationCacheAgent):
2476         * inspector/InspectorDOMStorageAgent.cpp:
2477         (WebCore::InspectorDOMStorageAgent::getDOMStorageEntries):
2478         (WebCore::InspectorDOMStorageAgent::setDOMStorageItem):
2479         (WebCore::InspectorDOMStorageAgent::removeDOMStorageItem):
2480         * inspector/InspectorDOMStorageAgent.h:
2481         (InspectorDOMStorageAgent):
2482         * inspector/InspectorDOMStorageResource.cpp:
2483         (WebCore::InspectorDOMStorageResource::bind):
2484
2485 2012-04-04  Tony Chang  <tony@chromium.org>
2486
2487         Remove flex function css parsing
2488         https://bugs.webkit.org/show_bug.cgi?id=83083
2489
2490         Reviewed by Ojan Vafai.
2491
2492         The flexbox spec switched from using a flex() function on the width
2493         and height to a css property. Now that the -webkit-flex: has been
2494         implemented, we can remove the flex() function code.
2495
2496         No new tests. Removing an old test that used to test the flex() function.
2497
2498         * CMakeLists.txt:
2499         * GNUmakefile.list.am:
2500         * Target.pri:
2501         * WebCore.gypi:
2502         * WebCore.vcproj/WebCore.vcproj:
2503         * WebCore.xcodeproj/project.pbxproj:
2504         * css/CSSAllInOne.cpp:
2505         * css/CSSFlexValue.cpp: Removed.
2506         * css/CSSFlexValue.h: Removed.
2507         * css/CSSParser.cpp:
2508         (WebCore::CSSParser::parseValue):
2509         (WebCore::CSSParser::parseFlex):
2510         * css/CSSParser.h:
2511         (WebCore):
2512         * css/CSSStyleApplyProperty.cpp:
2513         (WebCore::ApplyPropertyLength::applyValue):
2514         (WebCore::CSSStyleApplyProperty::CSSStyleApplyProperty):
2515         * css/CSSValue.cpp:
2516         (WebCore::CSSValue::cssText):
2517         (WebCore::CSSValue::destroy):
2518         * css/CSSValue.h:
2519         * rendering/style/RenderStyle.h:
2520         * rendering/style/StyleFlexibleBoxData.cpp:
2521         (WebCore::StyleFlexibleBoxData::StyleFlexibleBoxData):
2522         (WebCore::StyleFlexibleBoxData::operator==):
2523         * rendering/style/StyleFlexibleBoxData.h:
2524
2525 2012-04-04  Kausalya Madhusudhanan  <kmadhusu@chromium.org>
2526
2527         [Coverity] Address some uninit constructor values.
2528         https://bugs.webkit.org/show_bug.cgi?id=82424
2529
2530         Reviewed by Stephen White.
2531
2532         New tests are not required since I did not modify any code behavior. I just initialized the class member variables in the constructor.
2533
2534         * Modules/gamepad/Gamepad.cpp:
2535         (WebCore::Gamepad::Gamepad):
2536         * accessibility/AccessibilitySpinButton.cpp:
2537         (WebCore::AccessibilitySpinButtonPart::AccessibilitySpinButtonPart):
2538         * dom/WheelEvent.cpp:
2539         (WebCore::WheelEvent::WheelEvent):
2540         * html/canvas/WebGLContextObject.h:
2541         (WebGLContextObject):
2542         * platform/PlatformTouchPoint.h:
2543         (WebCore::PlatformTouchPoint::PlatformTouchPoint):
2544         * platform/graphics/chromium/cc/CCPageScaleAnimation.cpp:
2545         (WebCore::CCPageScaleAnimation::CCPageScaleAnimation):
2546         * platform/graphics/chromium/cc/CCRenderSurface.cpp:
2547         (WebCore::CCRenderSurface::CCRenderSurface):
2548         * xml/parser/MarkupTokenBase.h:
2549         (WebCore::AtomicMarkupTokenBase::AtomicMarkupTokenBase):
2550
2551 2012-04-04  Pavel Feldman  <pfeldman@chromium.org>
2552
2553         Web Inspector: expose savedURL, setToolbarColors and toggleSearchingForNode in InspectorFrontendAPI
2554         https://bugs.webkit.org/show_bug.cgi?id=83170
2555
2556         Reviewed by Yury Semikhatsky.
2557
2558         Chromium port is calling these from the embedder layer and they all are perfect candidates
2559         for the InspectorFrontendAPI. Adding them there.
2560
2561         * inspector/front-end/InspectorFrontendAPI.js:
2562         (InspectorFrontendAPI.enterInspectElementMode):
2563         (InspectorFrontendAPI.savedURL):
2564         (InspectorFrontendAPI.setToolbarColors):
2565
2566 2012-04-04  Ilya Tikhonovsky  <loislo@chromium.org>
2567
2568         Web Inspector: linearise aggregate's retaining size calculation.
2569         https://bugs.webkit.org/show_bug.cgi?id=83125
2570
2571         This version is twice as fast as the original and it is non-recursive.
2572
2573         Reviewed by Yury Semikhatsky.
2574
2575         * inspector/front-end/HeapSnapshot.js:
2576         (WebInspector.HeapSnapshot.prototype._getDominatedIndex): was moved closer to it's usage
2577         (WebInspector.HeapSnapshot.prototype._calculateClassesRetainedSize): it was _buildAggregates' inner function forDominatedNodes.
2578         it was:
2579         a) extracted from _buildAggregates;
2580         b) made non-recursive;
2581         c) many getters were inlined;
2582         d) subarray of dominating nodes were inlined too.
2583         (WebInspector.HeapSnapshot.prototype._buildAggregates): many getters were inlined.
2584         (WebInspector.HeapSnapshot.prototype.aggregates):
2585
2586 2012-04-04  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
2587
2588         Best clickable node might return non "clickable" node.
2589         https://bugs.webkit.org/show_bug.cgi?id=83153
2590
2591         Reviewed by Kenneth Rohde Christiansen.
2592
2593         Do not return non-clickable nodes as a best clickable node.
2594
2595         * page/EventHandler.cpp:
2596         (WebCore::EventHandler::bestClickableNodeForTouchPoint):
2597
2598 2012-04-04  Pavel Feldman  <pfeldman@chromium.org>
2599
2600         Web Inspector: migrate from InspectorFrontendHost.saveAs to InspectorFrontendHost.save.
2601         https://bugs.webkit.org/show_bug.cgi?id=83162
2602
2603         Reviewed by Yury Semikhatsky.
2604
2605         This change adds Save item into the context menus for inspector. It also makes
2606         saveAs parameter optional in the InspectorFrontendHost.save so that we could
2607         skip the dialog for the paths that were already saved. Embedder tells the front-end
2608         what urls were successfully saved so that front-end could issue subsequent save commands
2609         for those files.
2610
2611         * English.lproj/localizedStrings.js:
2612         * inspector/InspectorFrontendClient.h:
2613         (InspectorFrontendClient):
2614         * inspector/InspectorFrontendClientLocal.h:
2615         (WebCore::InspectorFrontendClientLocal::canSave):
2616         (WebCore::InspectorFrontendClientLocal::save):
2617         * inspector/InspectorFrontendHost.cpp:
2618         (WebCore::InspectorFrontendHost::canSave):
2619         (WebCore::InspectorFrontendHost::save):
2620         * inspector/InspectorFrontendHost.h:
2621         (InspectorFrontendHost):
2622         * inspector/InspectorFrontendHost.idl:
2623         * inspector/front-end/InspectorFrontendHostStub.js:
2624         (.WebInspector.InspectorFrontendHostStub.prototype.canSave):
2625         * inspector/front-end/JavaScriptSourceFrame.js:
2626         * inspector/front-end/NetworkPanel.js:
2627         (WebInspector.NetworkLogView.prototype._contextMenu):
2628         (WebInspector.NetworkLogView.prototype._exportAll):
2629         (WebInspector.NetworkLogView.prototype._exportResource):
2630         * inspector/front-end/ResourceView.js:
2631         * inspector/front-end/ResourcesPanel.js:
2632         (WebInspector.FrameResourceTreeElement.prototype._appendSaveAsAction.doSave):
2633         (WebInspector.FrameResourceTreeElement.prototype._appendSaveAsAction.save):
2634         (WebInspector.FrameResourceTreeElement.prototype._appendSaveAsAction):
2635         (WebInspector.ResourceRevisionTreeElement.prototype._handleContextMenuEvent.doSave):
2636         (WebInspector.ResourceRevisionTreeElement.prototype._handleContextMenuEvent.save):
2637         (WebInspector.ResourceRevisionTreeElement.prototype._handleContextMenuEvent):
2638         * inspector/front-end/Settings.js:
2639         * inspector/front-end/SourceFrame.js:
2640         (WebInspector.TextViewerDelegateForSourceFrame.prototype.populateTextAreaContextMenu):
2641         * inspector/front-end/TextViewer.js:
2642         (WebInspector.TextViewer.prototype._contextMenu):
2643         (WebInspector.TextViewer.prototype._commitEditing):
2644         (WebInspector.TextViewerDelegate.prototype.populateTextAreaContextMenu):
2645         * inspector/front-end/TimelineModel.js:
2646         (WebInspector.TimelineModel.prototype.saveToFile):
2647         * inspector/front-end/TimelinePanel.js:
2648         (WebInspector.TimelinePanel.prototype._registerShortcuts):
2649         (WebInspector.TimelinePanel.prototype._contextMenu):
2650         * inspector/front-end/inspector.js:
2651
2652 2012-04-04  Andrey Kosyakov  <caseq@chromium.org>
2653
2654         Web Inspector: time dividers are incorrectly rendered on empty timeline grid
2655         https://bugs.webkit.org/show_bug.cgi?id=83165
2656
2657         Reviewed by Pavel Feldman.
2658
2659         * inspector/front-end/NetworkPanel.js:
2660         (WebInspector.NetworkBaseCalculator.prototype.setDisplayWindow):
2661         * inspector/front-end/TimelineGrid.js:
2662         (WebInspector.TimelineGrid.prototype.updateDividers):
2663
2664 2012-04-04  Sergio Villar Senin  <svillar@igalia.com>
2665
2666         [GTK] ASSERT in SocketStreamHandleSoup.cpp
2667         https://bugs.webkit.org/show_bug.cgi?id=83123
2668
2669         Reviewed by Martin Robinson.
2670
2671         Do not try to reuse a GOwnPtr as calling outPtr() causes an
2672         assertion if the pointer is already valid. Also do not try to
2673         close the IOStream if it was not created.
2674
2675         This patch fixes
2676         http/tests/websocket/tests/hybi/workers/worker-reload.html but I
2677         am leaving it skipped until webkit.org/b/83124 is fixed.
2678
2679         * platform/network/soup/SocketStreamHandleSoup.cpp:
2680         (WebCore::connectedCallback):
2681
2682 2012-04-04  Alexis Menard  <alexis.menard@openbossa.org>
2683
2684         Re-add variable names in CSS related headers to help understanding the meaning of them.
2685         https://bugs.webkit.org/show_bug.cgi?id=83151
2686
2687         Reviewed by Andreas Kling.
2688
2689         r112952, r113031, r113042 removed variable names in some places but at the end it
2690         was better before as it helps understanding what the variables represent. This patch
2691         address this issue and add back the names where it makes sense.
2692
2693         No new tests : cosmetic changes, no behavior change here.
2694
2695         * css/CSSParser.h:
2696         (CSSParser):
2697         * css/StylePropertySet.h:
2698         (StylePropertySet):
2699         * dom/StyledElement.h:
2700         (StyledElement):
2701         * editing/EditingStyle.h:
2702         (EditingStyle):
2703         * editing/Editor.h:
2704         (Editor):
2705         * html/HTMLElement.h:
2706         (HTMLElement):
2707
2708 2012-04-04  Nikolas Zimmermann  <nzimmermann@rim.com>
2709
2710         Implement animatedPathSegList support for SVGPathElement
2711         https://bugs.webkit.org/show_bug.cgi?id=83140
2712
2713         Reviewed by Zoltan Herczeg.
2714
2715         Add the last missing piece of animVal support: animations of the SVGPathElement's 'd' attribute
2716         should be reflected in the 'animatedPathSegList' not the 'pathSegList'.
2717
2718         The SVGAnimatedPathAnimator is special as it doesn't operate on the SVGPathSegList types
2719         directly, but on the SVGPathByteStreams for performance reasons. The SVGPathSegLists
2720         are refcounted and exposed to JS, thus require lots of memory, compared to a simple
2721         byte stream, which SVGPathByteStream is. Only build an 'animatedPathSegList' while
2722         animating if its actually observed by the bindings, as this operation is heavy.
2723
2724         This marks the finish of the animVal patch series: no more string round trips during
2725         animation. Previously paths were animated as byte streams, then converted to a String
2726         passed to pathElement->setAttribute("d", "newString"), parsed again to a byte stream, etc.
2727
2728         A follow-up patch will remove the last crufts of direct setAttribute() animation.
2729
2730         Extend all existing animated SVGPathElement tests to cover animatedPathSegList.
2731
2732         * svg/SVGAnimatedColor.h:
2733         (WebCore::SVGAnimatedColorAnimator::startAnimValAnimation):
2734         (WebCore::SVGAnimatedColorAnimator::stopAnimValAnimation):
2735         (WebCore::SVGAnimatedColorAnimator::resetAnimValToBaseVal):
2736         (WebCore::SVGAnimatedColorAnimator::animValWillChange):
2737         (WebCore::SVGAnimatedColorAnimator::animValDidChange):
2738         (SVGAnimatedColorAnimator):
2739         * svg/SVGAnimatedPath.cpp:
2740         (WebCore::SVGAnimatedPathAnimator::constructFromString):
2741         (WebCore::SVGAnimatedPathAnimator::startAnimValAnimation):
2742         (WebCore):
2743         (WebCore::SVGAnimatedPathAnimator::stopAnimValAnimation):
2744         (WebCore::SVGAnimatedPathAnimator::resetAnimValToBaseVal):
2745         (WebCore::SVGAnimatedPathAnimator::animValWillChange):
2746         (WebCore::SVGAnimatedPathAnimator::animValDidChange):
2747         (WebCore::SVGAnimatedPathAnimator::calculateAnimatedValue):
2748         * svg/SVGAnimatedPath.h:
2749         (SVGAnimatedPathAnimator):
2750         * svg/SVGAnimatedType.cpp:
2751         (WebCore::SVGAnimatedType::valueAsString):
2752         (WebCore::SVGAnimatedType::setValueAsString):
2753         (WebCore::SVGAnimatedType::supportsAnimVal):
2754         * svg/SVGAnimatedTypeAnimator.h:
2755         (SVGAnimatedTypeAnimator):
2756         (WebCore::SVGAnimatedTypeAnimator::findAnimatedPropertiesForAttributeName):
2757         * svg/SVGPathElement.cpp:
2758         (WebCore::SVGPathElement::SVGPathElement):
2759         (WebCore::SVGPathElement::getTotalLength):
2760         (WebCore::SVGPathElement::getPointAtLength):
2761         (WebCore::SVGPathElement::getPathSegAtLength):
2762         (WebCore::SVGPathElement::parseAttribute):
2763         (WebCore::SVGPathElement::svgAttributeChanged):
2764         (WebCore::SVGPathElement::pathByteStream):
2765         (WebCore):
2766         (WebCore::SVGPathElement::lookupOrCreateDWrapper):
2767         (WebCore::SVGPathElement::pathSegList):
2768         (WebCore::SVGPathElement::animatedPathSegList):
2769         (WebCore::SVGPathElement::pathSegListChanged):
2770         * svg/SVGPathElement.h:
2771         (WebCore):
2772         (SVGPathElement):
2773         (WebCore::SVGPathElement::isAnimValObserved):
2774         * svg/SVGPathParserFactory.cpp:
2775         (WebCore::SVGPathParserFactory::buildSVGPathByteStreamFromSVGPathSegList):
2776         (WebCore::SVGPathParserFactory::buildSVGPathByteStreamFromString):
2777         (WebCore::SVGPathParserFactory::buildAnimatedSVGPathByteStream):
2778         * svg/SVGPathParserFactory.h:
2779         (SVGPathParserFactory):
2780         * svg/SVGPathSegWithContext.h:
2781         * svg/properties/SVGAnimatedListPropertyTearOff.h:
2782         (SVGAnimatedListPropertyTearOff):
2783         (WebCore::SVGAnimatedListPropertyTearOff::baseVal):
2784         (WebCore::SVGAnimatedListPropertyTearOff::animVal):
2785         (WebCore::SVGAnimatedListPropertyTearOff::removeItemFromList):
2786         (WebCore::SVGAnimatedListPropertyTearOff::detachListWrappers):
2787         (WebCore::SVGAnimatedListPropertyTearOff::currentAnimatedValue):
2788         (WebCore::SVGAnimatedListPropertyTearOff::animationStarted):
2789         (WebCore::SVGAnimatedListPropertyTearOff::animationEnded):
2790         (WebCore::SVGAnimatedListPropertyTearOff::synchronizeWrappersIfNeeded):
2791         * svg/properties/SVGAnimatedPathSegListPropertyTearOff.h:
2792         (WebCore::SVGAnimatedPathSegListPropertyTearOff::baseVal):
2793         (WebCore::SVGAnimatedPathSegListPropertyTearOff::animVal):
2794         (WebCore::SVGAnimatedPathSegListPropertyTearOff::removeItemFromList):
2795         (SVGAnimatedPathSegListPropertyTearOff):
2796         (WebCore::SVGAnimatedPathSegListPropertyTearOff::animationStarted):
2797         (WebCore::SVGAnimatedPathSegListPropertyTearOff::animationEnded):
2798         (WebCore::SVGAnimatedPathSegListPropertyTearOff::animValDidChange):
2799         (WebCore::SVGAnimatedPathSegListPropertyTearOff::animatedPathByteStream):
2800         (WebCore::SVGAnimatedPathSegListPropertyTearOff::SVGAnimatedPathSegListPropertyTearOff):
2801         * svg/properties/SVGAnimatedTransformListPropertyTearOff.h:
2802         (WebCore::SVGAnimatedTransformListPropertyTearOff::baseVal):
2803         (WebCore::SVGAnimatedTransformListPropertyTearOff::animVal):
2804         * svg/properties/SVGListProperty.h:
2805         (WebCore::SVGListProperty::setValuesAndWrappers):
2806         (WebCore::SVGListProperty::SVGListProperty):
2807         (WebCore::SVGListProperty::~SVGListProperty):
2808         (SVGListProperty):
2809
2810 2012-04-04  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
2811
2812         Best zoomable area does not balance intersection with target area.
2813         https://bugs.webkit.org/show_bug.cgi?id=83148
2814
2815         Reviewed by Kenneth Rohde Christiansen.
2816
2817         Change the two step sort to simply picking the best quotient of touch intersection with target area.
2818
2819         Test: touchadjustment/zoom-fatfinger.html
2820
2821         * page/TouchAdjustment.cpp:
2822         (WebCore::TouchAdjustment::distanceSquaredToTargetCenterLine):
2823         (TouchAdjustment):
2824         (WebCore::TouchAdjustment::zoomableIntersectionQuotient):
2825         (WebCore::TouchAdjustment::findNodeWithLowestDistanceMetric):
2826         (WebCore::findBestClickableCandidate):
2827         (WebCore::findBestZoomableArea):
2828
2829 2012-04-04  Andrey Kosyakov  <caseq@chromium.org>
2830
2831         Web Inspector: make padding and client window width part of timeline calculator's state
2832         https://bugs.webkit.org/show_bug.cgi?id=83122
2833
2834         Reviewed by Pavel Feldman.
2835
2836         - preserve window width and padding within TimelineCalculator
2837         - rely on calculator to add expansion arrow padding
2838         - add TimelineCalculator::computePosition() for computing single position (not entire bar dimensions)
2839
2840         * inspector/front-end/MemoryStatistics.js:
2841         (WebInspector.MemoryStatistics.prototype._refreshDividers):
2842         * inspector/front-end/NetworkPanel.js:
2843         (WebInspector.NetworkLogView.prototype._updateDividersIfNeeded):
2844         (WebInspector.NetworkLogView.prototype._reset):
2845         (WebInspector.NetworkBaseCalculator.prototype.computePosition):
2846         (WebInspector.NetworkBaseCalculator.prototype.formatTime):
2847         (WebInspector.NetworkBaseCalculator.prototype.setDisplayWindow):
2848         * inspector/front-end/TimelineGrid.js:
2849         (WebInspector.TimelineGrid.prototype.updateDividers):
2850         (WebInspector.TimelineGrid.prototype.addEventDividers):
2851         * inspector/front-end/TimelineOverviewPane.js:
2852         (WebInspector.TimelineOverviewPane.prototype._update):
2853         (WebInspector.TimelineOverviewPane.prototype._reset):
2854         (WebInspector.TimelineOverviewCalculator.prototype.computePosition):
2855         (WebInspector.TimelineOverviewCalculator.prototype.setDisplayWindow):
2856         * inspector/front-end/TimelinePanel.js:
2857         (WebInspector.TimelinePanel.prototype._updateEventDividers):
2858         (WebInspector.TimelinePanel.prototype._refresh):
2859         (WebInspector.TimelinePanel.prototype._refreshRecords):
2860         (WebInspector.TimelineCalculator.prototype.computePosition):
2861         (WebInspector.TimelineCalculator.prototype.computeBarGraphWindowPosition):
2862         (WebInspector.TimelineCalculator.prototype.setDisplayWindow):
2863         (WebInspector.TimelineRecordGraphRow.prototype.update):
2864         (WebInspector.TimelineExpandableElement.prototype._update):
2865
2866 2012-04-04  Noel Gordon  <noel.gordon@gmail.com>
2867
2868         [FileSystem] Forward declare File in FileCallback.h
2869         https://bugs.webkit.org/show_bug.cgi?id=83128
2870
2871         Reviewed by Kentaro Hara.
2872
2873         No new tests. Covered by existing tests.
2874
2875         * Modules/filesystem/FileCallback.h: forward declare File class to make
2876         #include of "File.h" redundant.
2877         (WebCore):
2878
2879 2012-04-04  Sheriff Bot  <webkit.review.bot@gmail.com>
2880
2881         Unreviewed, rolling out r113156.
2882         http://trac.webkit.org/changeset/113156
2883         https://bugs.webkit.org/show_bug.cgi?id=83146
2884
2885         Breaks inspector layout tests (Requested by pfeldman on
2886         #webkit).
2887
2888         * inspector/front-end/MemoryStatistics.js:
2889         (WebInspector.MemoryStatistics.prototype._refreshDividers):
2890         * inspector/front-end/TimelineGrid.js:
2891         (WebInspector.TimelineGrid.prototype.updateDividers):
2892         (WebInspector.TimelineGrid.prototype.addEventDividers):
2893         * inspector/front-end/TimelineOverviewPane.js:
2894         (WebInspector.TimelineOverviewPane.prototype._update):
2895         (WebInspector.TimelineOverviewPane.prototype._reset):
2896         * inspector/front-end/TimelinePanel.js:
2897         (WebInspector.TimelinePanel.prototype._updateEventDividers):
2898         (WebInspector.TimelinePanel.prototype._refresh):
2899         (WebInspector.TimelinePanel.prototype._refreshRecords):
2900         (WebInspector.TimelinePanel.prototype.get timelinePaddingLeft):
2901         (WebInspector.TimelineCalculator.prototype.computeBarGraphWindowPosition):
2902         (WebInspector.TimelineRecordGraphRow.prototype.update):
2903         (WebInspector.TimelineExpandableElement.prototype._update):
2904
2905 2012-04-04  Pavel Feldman  <pfeldman@chromium.org>
2906
2907         Web Inspector: front-end should accept raw web socket address for the remote connection.
2908         https://bugs.webkit.org/show_bug.cgi?id=83134
2909
2910         Reviewed by Yury Semikhatsky.
2911
2912         Migrated from the /devtools/page web socket schema to the generic one where remote target address
2913         can be specified as index.html?ws=localhost:9222/page1. Left the old host/page way for backwards
2914         compatibility.
2915
2916         * inspector/front-end/inspector.js:
2917
2918 2012-04-04  Kent Tamura  <tkent@chromium.org>
2919
2920         Add localization functions for the calendar picker
2921         https://bugs.webkit.org/show_bug.cgi?id=83129
2922
2923         Reviewed by Hajime Morita.
2924
2925         No behavior changes yet.
2926
2927         * WebCore.gypi: Add LocalizedCalendar.h and LocalizedCalendarICU.cpp.
2928         * platform/LocalizedStrings.h: Add calendarTodayText() and calendarClearText().
2929         * platform/text/LocalizedCalendar.h:
2930         Added. This provides monthLabels(), weekDayShortLabels(), and firstDayOfWeek().
2931         * platform/text/LocalizedCalendarICU.cpp: Added. ICU implementations of the above functions.
2932         (WebCore::ScopedDateFormat): A wrapper for UDateFormat*.
2933         (WebCore::ScopedDateFormat::ScopedDateFormat):
2934         (WebCore::ScopedDateFormat::~ScopedDateFormat):
2935         (WebCore::ScopedDateFormat::get):
2936         (createFallbackMonthLabels): Creates fallback month labels in English.
2937         (createLabelVector):
2938         A helper for createMonthLabels() and createWeekDayShortLabels().
2939         (createMonthLabels):
2940         Creates month labels using createLabelVector() or createFallbackMonthLabels().
2941         (WebCore::monthLabels):
2942         (createFallbackWeekDayShortLabels): Creates fallback week labels in English.
2943         (createWeekDayShortLabels):
2944         Creates month labels using createLabelVector() or createFallbackWeekDayShortLabels().
2945         (WebCore::weekDayShortLabels):
2946         (getFirstDayOfWeek):
2947         (WebCore::firstDayOfWeek):
2948
2949 2012-04-03  Hans Wennborg  <hans@chromium.org>
2950
2951         Speech JavaScript API: Plumbing for Chromium
2952         https://bugs.webkit.org/show_bug.cgi?id=81667
2953
2954         Reviewed by Darin Fisher.
2955
2956         SpeechGrammar::src() is called with a ScriptExecutionContext, because
2957         the attribute setter needs it. Provide a src() method that doesn't
2958         take a ScriptExecutionContext that can be used internally.
2959
2960         Rename SpeechRecognition::audioStartCallback(), etc. to use event style
2961         names, i.e. didStartAudio(), etc.
2962
2963         Remove SpeechRecognitionClient::notifyVisibilityHidden(). The embedder
2964         can keep track of the visibility.
2965
2966         Remove SpeechRecognitionClient::unregisterSpeechRecognition,
2967         it is not needed.
2968
2969         No new tests, just minor tweaks.
2970
2971         * Modules/speech/SpeechGrammar.h:
2972         (WebCore::SpeechGrammar::src):
2973         * Modules/speech/SpeechRecognition.cpp:
2974         (WebCore::SpeechRecognition::didStartAudio):
2975         (WebCore::SpeechRecognition::didStartSound):
2976         (WebCore::SpeechRecognition::didStartSpeech):
2977         (WebCore::SpeechRecognition::didEndSpeech):
2978         (WebCore::SpeechRecognition::didEndSound):
2979         (WebCore):
2980         (WebCore::SpeechRecognition::didEndAudio):
2981         (WebCore::SpeechRecognition::didReceiveResult):
2982         (WebCore::SpeechRecognition::didReceiveNoMatch):
2983         (WebCore::SpeechRecognition::didDeleteResult):
2984         (WebCore::SpeechRecognition::didReceiveError):
2985         (WebCore::SpeechRecognition::didStart):
2986         (WebCore::SpeechRecognition::didEnd):
2987         * Modules/speech/SpeechRecognition.h:
2988         (SpeechRecognition):
2989         * Modules/speech/SpeechRecognitionClient.h:
2990         (SpeechRecognitionClient):
2991         * Modules/speech/SpeechRecognitionController.h:
2992
2993 2012-04-03  Levi Weintraub  <leviw@chromium.org>
2994
2995         Switch baseline values to LayoutUnits in RenderTableSection.
2996         https://bugs.webkit.org/show_bug.cgi?id=83017
2997
2998         Reviewed by Julien Chaffraix.
2999
3000         We initially intend to keep table layout on integers post switching the render tree to primarily
3001         use sub-pixel precision. Baseline positions switch to sub-pixel units, and in table layout code are
3002         combined with padding (also sub-pixel), which means we ideally won't truncate their values until
3003         after they're added together. Converting baseline values in RenderTableSection to LayoutUnits.
3004
3005         No new tests. No change in behavior.
3006
3007         * rendering/RenderTableSection.cpp:
3008         (WebCore::RenderTableSection::calcRowLogicalHeight):
3009         (WebCore::RenderTableSection::layoutRows):
3010         * rendering/RenderTableSection.h:
3011         (RowStruct):
3012
3013 2012-04-04  Zan Dobersek  <zandobersek@gmail.com>
3014
3015         [GTK] Scrolling doesn't work in WebKit2 since r110185
3016         https://bugs.webkit.org/show_bug.cgi?id=81779
3017
3018         Reviewed by James Robinson.
3019
3020         Bring back functionality of the ScrollAnimatorNone for ports that
3021         use a fallback timer for animation frames.
3022
3023         No new tests - only restoring previous functionality.
3024
3025         * platform/ScrollAnimatorNone.cpp:
3026         (WebCore::ScrollAnimatorNone::ScrollAnimatorNone):
3027         (WebCore::ScrollAnimatorNone::fireUpAnAnimation):
3028         (WebCore):
3029         (WebCore::ScrollAnimatorNone::animationTimerFired):
3030         (WebCore::ScrollAnimatorNone::startNextTimer):
3031         (WebCore::ScrollAnimatorNone::animationTimerActive):
3032         (WebCore::ScrollAnimatorNone::stopAnimationTimerIfNeeded):
3033         * platform/ScrollAnimatorNone.h:
3034         (ScrollAnimatorNone):
3035
3036 2012-04-02  Peter Rybin  <peter.rybin@gmail.com>
3037
3038         Web Inspector: CodeGeneratorInspector.py: switch Inspector, Memory and Database domains to typed API
3039         https://bugs.webkit.org/show_bug.cgi?id=82958
3040
3041         Reviewed by Yury Semikhatsky.
3042
3043         Client code is switched to typed API (all InspectorObject and InspectorArray types are replaced with
3044         generated types from TypeBuilder according to Inspector.json).
3045
3046         Missing array of InspectorValues specialization is added.
3047
3048         * inspector/CodeGeneratorInspector.py:
3049         * inspector/InjectedScriptHost.cpp:
3050         (WebCore::InjectedScriptHost::inspectImpl):
3051         * inspector/InspectorAgent.cpp:
3052         (WebCore::InspectorAgent::inspect):
3053         * inspector/InspectorAgent.h:
3054         (InspectorAgent):
3055         * inspector/InspectorDatabaseAgent.cpp:
3056         (WebCore):
3057         (WebCore::InspectorDatabaseAgent::getDatabaseTableNames):
3058         * inspector/InspectorDatabaseAgent.h:
3059         (InspectorDatabaseAgent):
3060         * inspector/InspectorDatabaseResource.cpp:
3061         (WebCore::InspectorDatabaseResource::bind):
3062         * inspector/InspectorMemoryAgent.cpp:
3063         (WebCore::InspectorMemoryAgent::getDOMNodeCount):
3064         * inspector/InspectorMemoryAgent.h:
3065         (InspectorMemoryAgent):
3066
3067 2012-04-04  Andrey Kosyakov  <caseq@chromium.org>
3068
3069         Web Inspector: make padding and client window width part of timeline calculator's state
3070         https://bugs.webkit.org/show_bug.cgi?id=83122
3071
3072         Reviewed by Pavel Feldman.
3073
3074         - preserve window width and padding within TimelineCalculator
3075         - rely on calculator to add expansion arrow padding
3076         - add TimelineCalculator::computePosition() for computing single position (not entire bar dimensions)
3077
3078         * inspector/front-end/MemoryStatistics.js:
3079         (WebInspector.MemoryStatistics.prototype._refreshDividers):
3080         * inspector/front-end/TimelineGrid.js:
3081         (WebInspector.TimelineGrid.prototype.updateDividers):
3082         * inspector/front-end/TimelineOverviewPane.js:
3083         (WebInspector.TimelineOverviewPane.prototype._update):
3084         (WebInspector.TimelineOverviewPane.prototype._reset):
3085         (WebInspector.TimelineOverviewCalculator.prototype.computePosition):
3086         (WebInspector.TimelineOverviewCalculator.prototype.setDisplayWindow):
3087         * inspector/front-end/TimelinePanel.js:
3088         (WebInspector.TimelinePanel.prototype._updateEventDividers):
3089         (WebInspector.TimelinePanel.prototype._refresh):
3090         (WebInspector.TimelinePanel.prototype._refreshRecords):
3091         (WebInspector.TimelineCalculator):
3092         (WebInspector.TimelineCalculator.prototype.computePosition):
3093         (WebInspector.TimelineCalculator.prototype.computeBarGraphWindowPosition):
3094         (WebInspector.TimelineCalculator.prototype.setDisplayWindow):
3095         (WebInspector.TimelineRecordGraphRow.prototype.update):
3096         (WebInspector.TimelineExpandableElement.prototype._update):
3097
3098 2012-04-04  Sheriff Bot  <webkit.review.bot@gmail.com>
3099
3100         Unreviewed, rolling out r113149.
3101         http://trac.webkit.org/changeset/113149
3102         https://bugs.webkit.org/show_bug.cgi?id=83126
3103
3104         for breaking Chromium builders (Requested by pfeldman on
3105         #webkit).
3106
3107         * Modules/speech/SpeechGrammar.h:
3108         * Modules/speech/SpeechRecognition.cpp:
3109         (WebCore::SpeechRecognition::audioStartCallback):
3110         (WebCore::SpeechRecognition::soundStartCallback):
3111         (WebCore::SpeechRecognition::speechStartCallback):
3112         (WebCore::SpeechRecognition::speechEndCallback):
3113         (WebCore::SpeechRecognition::audioEndCallback):
3114         (WebCore::SpeechRecognition::resultCallback):
3115         (WebCore::SpeechRecognition::noMatchCallback):
3116         (WebCore::SpeechRecognition::resultDeletedCallback):
3117         (WebCore::SpeechRecognition::errorCallback):
3118         (WebCore::SpeechRecognition::startCallback):
3119         (WebCore::SpeechRecognition::endCallback):
3120         * Modules/speech/SpeechRecognition.h:
3121         (SpeechRecognition):
3122         * Modules/speech/SpeechRecognitionClient.h:
3123         (SpeechRecognitionClient):
3124         * Modules/speech/SpeechRecognitionController.h:
3125         (WebCore::SpeechRecognitionController::visibilityHidden):
3126         (WebCore::SpeechRecognitionController::unregisterSpeechRecognition):
3127
3128 2012-04-03  Andrey Kosyakov  <caseq@chromium.org>
3129
3130         Web Inspector: event marks are missing in the timeline overview
3131         https://bugs.webkit.org/show_bug.cgi?id=83043
3132
3133         Reviewed by Pavel Feldman.
3134
3135         - invoke overview event dividers update as part of overview update, not lower pane update;
3136         - store raw records for event dividers within overview;
3137         - moved createEventDividers to a class method of TimelinePresentationModel for reuse.
3138         - ditto for forAllRecords
3139
3140         * inspector/front-end/TimelineOverviewPane.js:
3141         (WebInspector.TimelineOverviewPane):
3142         (WebInspector.TimelineOverviewPane.prototype._update):
3143         (WebInspector.TimelineOverviewPane.prototype._updateCategoryStrips):
3144         (WebInspector.TimelineOverviewPane.prototype._updateEventDividers):
3145         (WebInspector.TimelineOverviewPane.prototype._onRecordAdded):
3146         (WebInspector.TimelineOverviewPane.prototype._reset):
3147         (WebInspector.HeapGraph.prototype.update):
3148         * inspector/front-end/TimelinePanel.js:
3149         (WebInspector.TimelinePanel.prototype._updateEventDividers):
3150         (WebInspector.TimelinePanel.prototype._innerAddRecordToTimeline.addTimestampRecords):
3151         (WebInspector.TimelinePanel.prototype._innerAddRecordToTimeline):
3152         * inspector/front-end/TimelinePresentationModel.js:
3153         (WebInspector.TimelinePresentationModel.isEventDivider):
3154         (WebInspector.TimelinePresentationModel.forAllRecords):
3155         (WebInspector.TimelinePresentationModel.createEventDivider):
3156
3157 2012-04-03  Hans Wennborg  <hans@chromium.org>
3158
3159         Speech JavaScript API: Plumbing for Chromium
3160         https://bugs.webkit.org/show_bug.cgi?id=81667
3161
3162         Reviewed by Darin Fisher.
3163
3164         SpeechGrammar::src() is called with a ScriptExecutionContext, because
3165         the attribute setter needs it. Provide a src() method that doesn't
3166         take a ScriptExecutionContext that can be used internally.
3167
3168         Rename SpeechRecognition::audioStartCallback(), etc. to use event style
3169         names, i.e. didStartAudio(), etc.
3170
3171         Remove SpeechRecognitionClient::notifyVisibilityHidden(). The embedder
3172         can keep track of the visibility.
3173
3174         Remove SpeechRecognitionClient::unregisterSpeechRecognition,
3175         it is not needed.
3176
3177         No new tests, just minor tweaks.
3178
3179         * Modules/speech/SpeechGrammar.h:
3180         (WebCore::SpeechGrammar::src):
3181         * Modules/speech/SpeechRecognition.cpp:
3182         (WebCore::SpeechRecognition::didStartAudio):
3183         (WebCore::SpeechRecognition::didStartSound):
3184         (WebCore::SpeechRecognition::didStartSpeech):
3185         (WebCore::SpeechRecognition::didEndSpeech):
3186         (WebCore::SpeechRecognition::didEndSound):
3187         (WebCore):
3188         (WebCore::SpeechRecognition::didEndAudio):
3189         (WebCore::SpeechRecognition::didReceiveResult):
3190         (WebCore::SpeechRecognition::didReceiveNoMatch):
3191         (WebCore::SpeechRecognition::didDeleteResult):
3192         (WebCore::SpeechRecognition::didReceiveError):
3193         (WebCore::SpeechRecognition::didStart):
3194         (WebCore::SpeechRecognition::didEnd):
3195         * Modules/speech/SpeechRecognition.h:
3196         (SpeechRecognition):
3197         * Modules/speech/SpeechRecognitionClient.h:
3198         (SpeechRecognitionClient):
3199         * Modules/speech/SpeechRecognitionController.h:
3200
3201 2012-04-04  Mariusz Grzegorczyk  <mariusz.g@samsung.com>
3202
3203         Fix build break when CONTEXT_MENUS is disabled.
3204         https://bugs.webkit.org/show_bug.cgi?id=82342
3205
3206         Reviewed by Andreas Kling.
3207
3208         Fixes build break in WebKit-Gtk, and WebKit-EFL ports when CONTEXT_MENUS macro is disabled.
3209
3210         * page/ContextMenuClient.h:
3211         * page/ContextMenuController.h:
3212         * page/Page.cpp:
3213         (WebCore::Page::PageClients::PageClients):
3214         * page/Page.h:
3215         (WebCore):
3216         (PageClients):
3217         * platform/ContextMenu.cpp:
3218         * platform/ContextMenu.h:
3219         * platform/ContextMenuItem.cpp:
3220         * platform/ContextMenuItem.h:
3221         * platform/efl/ContextMenuEfl.cpp:
3222         * platform/efl/ContextMenuItemEfl.cpp:
3223         * platform/gtk/ContextMenuGtk.cpp:
3224         * platform/gtk/ContextMenuItemGtk.cpp:
3225
3226 2012-04-03  Adam Barth  <abarth@webkit.org>
3227
3228         Implement <iframe srcdoc>
3229         https://bugs.webkit.org/show_bug.cgi?id=82991
3230
3231         Reviewed by Sam Weinig.
3232
3233         This patch implements the <iframe srcdoc> feature. This feature allows
3234         authors to easily supply the contents of an iframe without needing to
3235         round-trip to the server. Using srcdoc is more convenient than using
3236         data URLs because we set a bunch of properties of the child document
3237         sensibly. For example, the child inherits the base URL of the parent
3238         and the child uses standards mode by default.
3239
3240         This feature is specified in
3241         <http://www.whatwg.org/specs/web-apps/current-work/#attr-iframe-srcdoc>.
3242         Although the feature has been in the spec for a while, I'm not aware of
3243         any other implementations in major browsers. The srcdoc feature works
3244         especially well with the sandbox and seamless attributes. We already
3245         implement sandbox and will likely implement seamless soon.
3246
3247         The srcdoc feature was announced on the webkit-dev mailing list on March 30:
3248         https://lists.webkit.org/pipermail/webkit-dev/2012-March/020161.html
3249
3250         This patch approaches the implementation using SubstituteData, which is
3251         a mechanism previously used for error messages and the like. Using
3252         SubstituteData has the advantage of not needing to modify the loading
3253         or history pipelines at all, making the integration of srcdoc with the
3254         rest of WebCore quite smooth.
3255
3256         This patch encodes the contents of the srcdoc attribute to and from
3257         UTF-8 when round-tripping the contents through the loader. In a future
3258         patch, I plan to experiment with whether using UTF-16 (or perhaps
3259         another encoding) can improve performance. There might also be a way to
3260         avoid the memcpy entirely, but these optimizations are best left to
3261         followup patches as this patch focuses on the observable behavior of
3262         the feature.
3263
3264         Tests: fast/frames/srcdoc/reloading-a-srcdoc-document-loads-it-again.html
3265                fast/frames/srcdoc/setting-src-does-nothing.html
3266                fast/frames/srcdoc/setting-srcdoc-reloads-document.html
3267                fast/frames/srcdoc/srcdoc-beats-src-dom.html
3268                fast/frames/srcdoc/srcdoc-beats-src.html
3269                fast/frames/srcdoc/srcdoc-can-be-in-qurks-mode.html
3270                fast/frames/srcdoc/srcdoc-can-navigate.html
3271                fast/frames/srcdoc/srcdoc-defaults-to-standards-mode.html
3272                fast/frames/srcdoc/srcdoc-loads-content.html
3273                fast/frames/srcdoc/srcdoc-urls.html
3274                http/tests/security/srcdoc-can-access-parent.html
3275                http/tests/security/srcdoc-in-sandbox-cannot-access-parent.html
3276                http/tests/security/srcdoc-inherits-referrer-for-forms.html
3277                http/tests/security/srcdoc-inherits-referrer.html
3278
3279         * dom/Document.cpp:
3280         (WebCore::Document::Document):
3281         (WebCore::Document::initSecurityContext):
3282             - srcdoc documents need to inherit their security contexts from
3283               their parents. We pick this initialization point to inherit the
3284               base URL and to set the "srcdoc Document" bit so that everything
3285               gets initialized atomically and from precisely the same owner
3286               frame.
3287         * dom/Document.h:
3288         (Document):
3289         (WebCore::Document::isSrcdocDocument):
3290             - This bit of state is present in the HTML5 spec and is referred to
3291               by a number of different requirements in the spec.
3292         * html/HTMLAttributeNames.in:
3293         * html/HTMLFrameElementBase.cpp:
3294         (WebCore::HTMLFrameElementBase::parseAttribute):
3295         (WebCore::HTMLFrameElementBase::location):
3296             - These functions implement the requirement that the srcdoc
3297               attribute takes precedence over the src attribute and triggers a
3298               load of a document with the URL about:srcdoc.
3299         * html/HTMLIFrameElement.idl:
3300             - Expose the srcdoc property as a reflection of the DOM attribute.
3301         * html/parser/HTMLTreeBuilder.cpp:
3302         (WebCore::HTMLTreeBuilder::defaultForInitial):
3303             - This tweak allows srcdoc documents to use standards mode by
3304               default, saving authors from having to include a doctype in each
3305               srcdoc document.
3306         * loader/FrameLoader.cpp:
3307         (WebCore::FrameLoader::defaultSubstituteDataForURL):
3308             - This function transfers the contents of the srcdoc attribute into
3309               the loading pipeline. If the client supplies other
3310               SubstituteData, that takes precendence over our "default"
3311               SubstituteData.
3312         (WebCore::FrameLoader::outgoingReferrer):
3313             - This function implements the requirement from the HTML5 spec that
3314               srcdoc documents inherit their referrer from their parent
3315               document. A recursive implementation might have been more
3316               aesthetic but the iterative implementation seemed like a better
3317               choice.
3318         (WebCore::FrameLoader::shouldTreatURLAsSrcdocDocument):
3319             - An about:srcdoc URL only has special meaning when loaded in an
3320               iframe with a srcdoc attribute. Otherwise, it's just a normal
3321               "about" URL, meaning a blank page.
3322         (WebCore::FrameLoader::urlSelected):
3323         (WebCore::FrameLoader::submitForm):
3324         (WebCore::FrameLoader::loadFrameRequest):
3325         (WebCore::FrameLoader::load):
3326         (WebCore::FrameLoader::loadWithNavigationAction):
3327         (WebCore::FrameLoader::reloadWithOverrideEncoding):
3328         (WebCore::FrameLoader::reload):
3329         (WebCore::FrameLoader::loadResourceSynchronously):
3330             - Update these call sites to call defaultSubstituteDataForURL and
3331               outgoingReferrer consistently.
3332         * loader/FrameLoader.h:
3333         (FrameLoader):
3334
3335 2012-03-29  Geoffrey Garen  <ggaren@apple.com>
3336
3337         First step toward incremental Weak<T> finalization
3338         https://bugs.webkit.org/show_bug.cgi?id=82670
3339
3340         Reviewed by Filip Pizlo.
3341
3342         Updated WebCore for Weak<T> API changes.
3343
3344         * bindings/js/DOMWrapperWorld.cpp:
3345         (WebCore::JSStringOwner::finalize): We're not allowed to get() a dead Weak<T>
3346         anymore, so use the debug-only was() helper function instead.
3347
3348         * bindings/js/JSDOMBinding.h:
3349         (WebCore::uncacheWrapper): Ditto.
3350
3351         * bindings/js/JSNodeCustom.h:
3352         (WebCore::setInlineCachedWrapper):
3353         (WebCore::clearInlineCachedWrapper): We're not allowed to get() a dead
3354         Weak<T>, so I had to push down these ASSERTs into ScriptWrappable.
3355
3356         * bindings/js/JSNodeFilterCondition.cpp:
3357         (WebCore::JSNodeFilterCondition::acceptNode): Updated for non-Handle-ness
3358         of Weak<T>.
3359
3360         * bindings/js/ScriptWrappable.h:
3361         (WebCore::ScriptWrappable::setWrapper):
3362         (WebCore::ScriptWrappable::clearWrapper): Use was(), as above.
3363
3364 2012-04-03  Luke Macpherson  <macpherson@chromium.org>
3365
3366         Don't parse "show" and "hide" as valid values for display property.
3367         https://bugs.webkit.org/show_bug.cgi?id=83115
3368
3369         Reviewed by Adam Barth.
3370
3371         No new tests.
3372
3373         * css/CSSParser.cpp:
3374         (WebCore::isValidKeywordPropertyAndValue):
3375
3376 2012-04-03  Yuta Kitamura  <yutak@chromium.org>
3377
3378         Crash in WebCore::WorkerThreadableWebSocketChannel::Bridge::mainThreadCreateWebSocketChannel
3379         https://bugs.webkit.org/show_bug.cgi?id=82873
3380
3381         Reviewed by David Levin.
3382
3383         WorkerThreadableWebSocketChannel::Bridge should properly handle the cases where inter-thread
3384         callback is not called due to the termination of the worker run loop. Specifically, the bridge
3385         should not send its "this" pointer to the main thread, because the bridge object may be freed
3386         in the worker thread before the main thread starts to process.
3387
3388         Test: http/tests/websocket/tests/hybi/workers/worker-reload.html
3389
3390         * Modules/websockets/ThreadableWebSocketChannelClientWrapper.cpp:
3391         (WebCore::ThreadableWebSocketChannelClientWrapper::ThreadableWebSocketChannelClientWrapper):
3392         (WebCore::ThreadableWebSocketChannelClientWrapper::peer):
3393         (WebCore::ThreadableWebSocketChannelClientWrapper::didCreateWebSocketChannel):
3394         Renamed from setUseHixie76Protocol, as this funtion now also sets m_peer.
3395         Sets m_syncMethodDone to true, because this function is called in the end of
3396         synchronous wait of Bridge::initialize().
3397         (WebCore::ThreadableWebSocketChannelClientWrapper::clearPeer):
3398         (WebCore::ThreadableWebSocketChannelClientWrapper::useHixie76Protocol):
3399         * Modules/websockets/ThreadableWebSocketChannelClientWrapper.h:
3400         Add WorkerThreadableWebSocketChannel::Peer which is initialized after the creation of
3401         WebSocketChannel in the main thread.
3402         (ThreadableWebSocketChannelClientWrapper):
3403         * Modules/websockets/WorkerThreadableWebSocketChannel.cpp:
3404         (WebCore::WorkerThreadableWebSocketChannel::WorkerThreadableWebSocketChannel):
3405         Don't do synchronous wait in the constructor, as a member function may be called
3406         during the wait before the constructor finishes. The meat of the constructor has
3407         moved to initialize() function.
3408         (WebCore::WorkerThreadableWebSocketChannel::Bridge::Bridge):
3409         (WebCore::WorkerThreadableWebSocketChannel::Bridge::~Bridge):
3410         (WorkerContextDidInitializeTask):
3411         (WebCore::WorkerContextDidInitializeTask::create):
3412         (WebCore::WorkerContextDidInitializeTask::~WorkerContextDidInitializeTask):
3413         (WebCore::WorkerContextDidInitializeTask::WorkerContextDidInitializeTask):
3414         (WebCore::WorkerThreadableWebSocketChannel::Bridge::mainThreadInitialize):
3415         (WebCore::WorkerThreadableWebSocketChannel::Bridge::initialize):
3416         Don't pass "this" object to the main thread. Receive the pointer to the peer object
3417         via ThreadableWebSocketChannelClientWrapper which is ThreadSafeRefCounted<>.
3418         (WebCore::WorkerThreadableWebSocketChannel::Bridge::connect):
3419         m_peer may be NULL, and we should not do anything in that case.
3420         (WebCore::WorkerThreadableWebSocketChannel::Bridge::send):
3421         (WebCore::WorkerThreadableWebSocketChannel::Bridge::bufferedAmount):
3422         (WebCore::WorkerThreadableWebSocketChannel::mainThreadClose):
3423         (WebCore::WorkerThreadableWebSocketChannel::Bridge::close):
3424         (WebCore::WorkerThreadableWebSocketChannel::Bridge::fail):
3425         (WebCore::WorkerThreadableWebSocketChannel::Bridge::suspend):
3426         (WebCore::WorkerThreadableWebSocketChannel::Bridge::resume):
3427         * Modules/websockets/WorkerThreadableWebSocketChannel.h:
3428         (WorkerThreadableWebSocketChannel):
3429         (WebCore::WorkerThreadableWebSocketChannel::refThreadableWebSocketChannel):
3430         (WebCore::WorkerThreadableWebSocketChannel::derefThreadableWebSocketChannel):
3431         (Bridge):
3432  &