Source/WebCore/ChangeLog

   1 2013-09-04  Andre Moreira Magalhaes   <andre.magalhaes@collabora.co.uk>
   2
   3         [GStreamer] cannot play live streams
   4         https://bugs.webkit.org/show_bug.cgi?id=116831
   5
   6         Reviewed by Philippe Normand.
   7
   8         Fix issues with rtsp streams embedded on <video> not loading.
   9
  10         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
  11         Do not reset pipeline to READY state on STATE_CHANGE_ASYNC when entering PAUSED state for
  12         live streams, otherwise we enter an endless loop of READY->PAUSED->READY->PAUSED when
  13         starting playback.
  14
  15 2013-09-03  Chris Fleizach  <cfleizach@apple.com>
  16
  17         AX: REGRESSION: @title is exposed as AXDescription when label label from contents already exists.
  18         https://bugs.webkit.org/show_bug.cgi?id=120550
  19
  20         Reviewed by Mario Sanchez Prada.
  21
  22         Resolve a FIXME from the accessible name computation refactoring so that alternative text for links do not
  23         show up in the title field and do not duplicate naming when a title tag is used.
  24
  25         Effectively, this means that links no longer use AXTitle for alternative text. They use AXDescription
  26         like all other elements.
  27
  28         Test: platform/mac/accessibility/link-with-title.html
  29
  30         * accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
  31         (-[WebAccessibilityObjectWrapper accessibilityTitle]):
  32         (-[WebAccessibilityObjectWrapper accessibilityDescription]):
  33
  34 2013-09-03  Darin Adler  <darin@apple.com>
  35
  36         Change type of Document::doctype back to a raw pointer
  37         https://bugs.webkit.org/show_bug.cgi?id=120617
  38
  39         Reviewed by Andreas Kling.
  40
  41         * dom/Document.cpp:
  42         (WebCore::Document::doctype): Return a raw pointer.
  43         (WebCore::Document::childrenChanged): Use the raw pointer.
  44         Also added a FIXME about this code that is probably in the wrong place.
  45         * dom/Document.h: More of the same.
  46
  47         * editing/markup.cpp:
  48         (WebCore::documentTypeString): Get rid of local variable entirely,
  49         since null is already handled right by createMarkup, and also remove
  50         the call to get since doctype is just a raw pointer.
  51
  52 2013-09-03  Andreas Kling  <akling@apple.com>
  53
  54         ASSERTION FAILED: frame().view() == this closing a page with SVG or video
  55         <https://webkit.org/b/120645>
  56
  57         Reviewed by Antti Koivisto.
  58
  59         Have RenderSVGResourceContainer check if the document is being destroyed before
  60         triggering any repaints. This replaces the previous check for a null RenderView
  61         which meant basically the same thing.
  62
  63         We could add more and better assertions to catch unnecessary work during tree
  64         teardown, but let's do that separately.
  65
  66         * rendering/svg/RenderSVGResourceContainer.cpp:
  67         (WebCore::RenderSVGResourceContainer::markClientForInvalidation):
  68
  69 2013-09-03  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
  70
  71         Fix uninitialized build warning in make_names.pl
  72         https://bugs.webkit.org/show_bug.cgi?id=120658
  73
  74         Reviewed by Andreas Kling.
  75
  76         No new tests, no behavior change.
  77
  78         * dom/make_names.pl:
  79         (printTypeChecks): Fixed a build warning since r154965.
  80
  81 2013-09-03  Darin Adler  <darin@apple.com>
  82
  83         Fix backwards branch in ~Node from r154967
  84         https://bugs.webkit.org/show_bug.cgi?id=120659
  85
  86         Reviewed by Andreas Kling.
  87
  88         * dom/Node.cpp:
  89         (WebCore::Node::~Node): Fix backwards branch. The old code ran the code only
  90         when we did not remove the item from the table. I removed a ! from this expression
  91         after review; bad idea.
  92
  93 2013-09-03  Enrica Casucci  <enrica@apple.com>
  94
  95         Can't select Katakana word by double-clicking.
  96         <rdar://problem/14654926>
  97
  98         Reviewed by Alexey Proskuryakov and Ryosuke Niwa.
  99
 100         For some languages, like Japanese we need
 101         to use more context for word breaking.
 102
 103         New test: editing/selection/doubleclick-japanese-text.html
 104
 105         * platform/text/TextBoundaries.h:
 106         (WebCore::requiresContextForWordBoundary):
 107
 108 2013-09-03  Mark Lam  <mark.lam@apple.com>
 109
 110         Converting StackIterator to a callback interface.
 111         https://bugs.webkit.org/show_bug.cgi?id=120564.
 112
 113         Reviewed by Filip Pizlo.
 114
 115         No new tests.
 116
 117         * bindings/js/JSXMLHttpRequestCustom.cpp:
 118         (WebCore::SendFunctor::SendFunctor):
 119         (WebCore::SendFunctor::hasViableFrame):
 120         (WebCore::SendFunctor::operator()):
 121         (WebCore::JSXMLHttpRequest::send):
 122         * bindings/js/ScriptCallStackFactory.cpp:
 123         (WebCore::CreateScriptCallStackFunctor::CreateScriptCallStackFunctor):
 124         (WebCore::CreateScriptCallStackFunctor::operator()):
 125         (WebCore::createScriptCallStack):
 126         (WebCore::CreateScriptCallStackForConsoleFunctor::CreateScriptCallStackForConsoleFunctor):
 127         (WebCore::CreateScriptCallStackForConsoleFunctor::operator()):
 128
 129 2013-09-03  Oliver Hunt  <oliver@apple.com>
 130
 131         Support structured clone of Map and Set
 132         https://bugs.webkit.org/show_bug.cgi?id=120654
 133
 134         Reviewed by Simon Fraser.
 135
 136         Add support for cloning Map and Set.  Fairly self explanatory change.
 137         Needed to add Forwarding headers for the JSMap, JSSet and MapData classes.
 138
 139         * ForwardingHeaders/runtime/JSMap.h: Added.
 140         * ForwardingHeaders/runtime/JSSet.h: Added.
 141         * ForwardingHeaders/runtime/MapData.h: Added.
 142         * bindings/js/SerializedScriptValue.cpp:
 143         (WebCore::CloneSerializer::isMap):
 144         (WebCore::CloneSerializer::isSet):
 145         (WebCore::CloneSerializer::startSet):
 146         (WebCore::CloneSerializer::startMap):
 147         (WebCore::CloneSerializer::serialize):
 148         (WebCore::CloneDeserializer::consumeMapDataTerminationIfPossible):
 149         (WebCore::CloneDeserializer::deserialize):
 150
 151 2013-09-03  Bear Travis  <betravis@adobe.com>
 152
 153         [CSS Shapes] Shape's content gets extra left offset when left-border is positive on the content box
 154         https://bugs.webkit.org/show_bug.cgi?id=117573
 155
 156         Reviewed by David Hyatt.
 157
 158         Nested blocks need to take into account their offset from the shape-inside container.
 159         The new code calculates the offset from the shape-inside container, then applies the
 160         offset to the computed segments. The line must be moved down by the offset's height,
 161         and each segment must be moved left by the offset's width.
 162
 163         Test: fast/shapes/shape-inside/shape-inside-offset-block-children.html
 164
 165         * rendering/RenderBlock.cpp:
 166         (WebCore::RenderBlock::relayoutShapeDescendantIfMoved): Relayout a block child if its
 167         new logical left would cause it to rest at a new position within a shape container.
 168         (WebCore::RenderBlock::logicalOffsetFromShapeAncestorContainer): Calculate the logical
 169         offset form a shape inside ancestor container.
 170         (WebCore::RenderBlock::layoutBlockChild): Call relayoutShapeDescendantIfMoved with the
 171         new position offset.
 172         * rendering/RenderBlock.h:
 173         * rendering/RenderBlockLineLayout.cpp:
 174         (WebCore::RenderBlock::updateShapeAndSegmentsForCurrentLine): Use layout offset, rather
 175         than just vertical offset.
 176         (WebCore::RenderBlock::updateShapeAndSegmentsForCurrentLineInFlowThread): Ditto.
 177         (WebCore::RenderBlock::layoutRunsAndFloatsInRange): Ditto.
 178         * rendering/shapes/ShapeInsideInfo.h:
 179         (WebCore::ShapeInsideInfo::computeSegmentsForLine): Shift segments logically left when
 180         there is an inline offset.
 181
 182 2013-09-03  Antoine Quint  <graouts@apple.com>
 183
 184         Web Inspector: exceptions triggered from console evaluation do not pause the debugger
 185         https://bugs.webkit.org/show_bug.cgi?id=120460
 186
 187         Reviewed by Timothy Hatcher.
 188
 189         * inspector/InjectedScriptSource.js:
 190         Explicitly set a sourceURL such that the frontend may identify injected script when
 191         processing call frames in order to hide such code from the debugger.
 192
 193 2013-09-03  Andreas Kling  <akling@apple.com>
 194
 195         Support Vector<Ref<T>>.
 196         <https://webkit.org/b/120637>
 197
 198         Reviewed by Antti Koivisto.
 199
 200         Use Vector<Ref<T>> internally in Page.
 201
 202         * page/Page.cpp:
 203         (WebCore::networkStateChanged):
 204         (WebCore::Page::refreshPlugins):
 205
 206             Clean up these functions and use Vector<Ref<Frame>> to store pointers to frames
 207             since we know they are not going to be null.
 208
 209         (WebCore::Page::pluginViews):
 210
 211             Changed this to return a Vector<Ref<PluginView>> by value instead of passing a
 212             writable vector as an argument. Clean up loops with 'auto'.
 213
 214         (WebCore::Page::storageBlockingStateChanged):
 215         (WebCore::Page::privateBrowsingStateChanged):
 216
 217             Tweaked for pluginViews() returning a Vector now.
 218
 219         (WebCore::Page::setVisibilityState):
 220
 221             Store Documents that need a visibilitychange event in a Vector<Ref<Document>>.
 222
 223 2013-08-28  Sergio Villar Senin  <svillar@igalia.com>
 224
 225         [CSS Grid Layout] Add parsing for named grid lines
 226         https://bugs.webkit.org/show_bug.cgi?id=119540
 227
 228         Reviewed by Andreas Kling.
 229
 230         From Blink r150381,r150587 by <jchaffraix@chromium.org>
 231
 232         Adds parsing support for named grid lines at <grid-line> level,
 233         i.e., inside grid-{row|column}-{start|end}. This change covers
 234         only the parsing, layout changes coming in a follow up patch.
 235
 236         * css/CSSComputedStyleDeclaration.cpp:
 237         (WebCore::valueForGridPosition):
 238         * css/CSSParser.cpp:
 239         (WebCore::CSSParser::parseIntegerOrStringFromGridPosition):
 240         (WebCore::CSSParser::parseGridPosition):
 241         * css/CSSParser.h:
 242         * css/StyleResolver.cpp:
 243         (WebCore::createGridPosition):
 244         * rendering/RenderGrid.cpp:
 245         (WebCore::RenderGrid::resolveGridPositionFromStyle):
 246         * rendering/style/GridPosition.h:
 247         (WebCore::GridPosition::setExplicitPosition):
 248         (WebCore::GridPosition::setSpanPosition):
 249         (WebCore::GridPosition::integerPosition):
 250         (WebCore::GridPosition::namedGridLine):
 251
 252 2013-09-03  Alexey Proskuryakov  <ap@apple.com>
 253
 254         [Mac] Hyphenation respects regional format settings language instead of primary language
 255         https://bugs.webkit.org/show_bug.cgi?id=120641
 256
 257         Reviewed by Dan Bernstein.
 258
 259         Fixes hyphenation tests on my machine with non-English regional format settings.
 260
 261         * platform/text/cf/HyphenationCF.cpp: (createValueForNullKey):
 262         Use primary UI language for hyphenation, not regional settings language.
 263
 264 2013-09-03  Danilo Cesar Lemes de Paula  <danilo.cesar@collabora.co.uk>
 265
 266         [GTK][EFL] include missing localized strings for subtitle auto track
 267         https://bugs.webkit.org/show_bug.cgi?id=120629
 268
 269         those methods are necessary to show the "Auto" track on webkitgtk
 270
 271         Reviewed by Gustavo Noronha Silva.
 272
 273         * platform/efl/LocalizedStringsEfl.cpp:
 274         (WebCore::textTrackAutomaticMenuItemText):
 275         * platform/gtk/LocalizedStringsGtk.cpp:
 276         (WebCore::textTrackAutomaticMenuItemText):
 277
 278 2013-09-03  Daniel Bates  <dabates@apple.com>
 279
 280         Require layout when -webkit-overflow-scrolling changes
 281         https://bugs.webkit.org/show_bug.cgi?id=120535
 282
 283         Reviewed by Darin Adler.
 284
 285         Test: fast/repaint/overflow-scroll-touch-repaint.html
 286
 287         We want to require a layout when the value of -webkit-overflow-scrolling changes
 288         since -webkit-overflow-scrolling creates a stacking context.
 289
 290         * rendering/style/RenderStyle.cpp:
 291         (WebCore::RenderStyle::changeRequiresLayout):
 292
 293 2013-09-02  Ryosuke Niwa  <rniwa@webkit.org>
 294
 295         Support the "json" responseType and JSON response entity in XHR
 296         https://bugs.webkit.org/show_bug.cgi?id=73648
 297
 298         Reviewed by Oliver Hunt.
 299
 300         Based on the patch written by Jarred Nicholls.
 301
 302         Implement 'json' type for XMLHttpRequest.response. We cache the result on JSC side as a cached attribute
 303         unlike other response types like 'document' and 'blob' for which the parsed response object is cached
 304         in XMLHttpRequest itself. In the long run, we should do the same for other types of response types.
 305
 306         Also refactored the various code to share the code.
 307
 308         Tests: fast/xmlhttprequest/xmlhttprequest-responsetype-json-invalid.html
 309                fast/xmlhttprequest/xmlhttprequest-responsetype-json-utf16.html
 310                fast/xmlhttprequest/xmlhttprequest-responsetype-json-valid.html
 311
 312         * ForwardingHeaders/runtime/JSONObject.h: Added.
 313
 314         * bindings/js/JSXMLHttpRequestCustom.cpp:
 315         (WebCore::JSXMLHttpRequest::visitChildren):
 316         (WebCore::JSXMLHttpRequest::response): Use JSONParse to parse the response text and cache the result.
 317         Call didCacheResponseJSON to set the cache status and clear the original response buffer.
 318
 319         * xml/XMLHttpRequest.cpp:
 320         (WebCore::XMLHttpRequest::XMLHttpRequest): Added m_responseCacheIsValid to invalidate the cache of
 321         a json response.
 322         (WebCore::XMLHttpRequest::responseText):
 323         (WebCore::XMLHttpRequest::didCacheResponseJSON): Added; Updates m_responseCacheIsValid and clears the
 324         response buffer to save memory.
 325         (WebCore::XMLHttpRequest::responseXML):
 326         (WebCore::XMLHttpRequest::setResponseType):
 327         (WebCore::XMLHttpRequest::responseType):
 328         (WebCore::XMLHttpRequest::clearResponseBuffers):
 329         (WebCore::XMLHttpRequest::didReceiveData):
 330
 331         * xml/XMLHttpRequest.h:
 332         (WebCore::XMLHttpRequest::doneWithoutErrors): Extracted from responseXML.
 333         (WebCore::XMLHttpRequest::responseTextIgnoringResponseType): Extracted from responseText.
 334         (WebCore::XMLHttpRequest::responseCacheIsValid): Added.
 335         (WebCore::XMLHttpRequest::shouldDecodeResponse): Extracted from didReceiveData.
 336         Also modified to decode when the response type is ResponseTypeJSON.
 337
 338         * xml/XMLHttpRequest.idl: Added CachedAttribute IDL extention on response property. This cache is
 339         used when the response type is 'json'.
 340
 341 2013-09-03  Commit Queue  <commit-queue@webkit.org>
 342
 343         Unreviewed, rolling out r154881.
 344         http://trac.webkit.org/changeset/154881
 345         https://bugs.webkit.org/show_bug.cgi?id=120643
 346
 347         Crashes on macworld.com (Requested by kling on #webkit).
 348
 349         * dom/Element.cpp:
 350         (WebCore::Element::setAttributeInternal):
 351
 352 2013-09-03  Andre Moreira Magalhaes   <andre.magalhaes@collabora.co.uk>
 353
 354         [GStreamer] Don't set state to NULL until element is destroyed
 355         https://bugs.webkit.org/show_bug.cgi?id=117354
 356
 357         Reviewed by Philippe Normand.
 358
 359         Don't set playbin to NULL until it is going to be destroyed or if we stay
 360         for too long on the READY state. Instead only set the state to READY as this
 361         allows much faster state changes to PAUSED/PLAYING again. playbin internally
 362         caches some state that is destroyed when setting it to NULL.
 363         This state is independent of the URI and it is even possible to change the
 364         URI in READY state.
 365
 366         To avoid having resources (e.g. audio devices) open indefinitely,
 367         when setting the state to READY we create a timeout and if the timeout
 368         is reached we reset the pipeline state to NULL to free resources.
 369
 370         Also now all state changes use the changePipelineState method instead of setting
 371         the playbin state directly with gst_element_set_state, so we have a better control
 372         of when we are requesting state changes.
 373
 374         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
 375         (WebCore::mediaPlayerPrivateReadyStateTimeoutCallback):
 376         (WebCore::MediaPlayerPrivateGStreamer::MediaPlayerPrivateGStreamer):
 377         (WebCore::MediaPlayerPrivateGStreamer::~MediaPlayerPrivateGStreamer):
 378         (WebCore::MediaPlayerPrivateGStreamer::commitLoad):
 379         (WebCore::MediaPlayerPrivateGStreamer::changePipelineState):
 380         (WebCore::MediaPlayerPrivateGStreamer::setRate):
 381         (WebCore::MediaPlayerPrivateGStreamer::handlePluginInstallerResult):
 382         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.h:
 383
 384 2013-09-03  peavo@outlook.com  <peavo@outlook.com>
 385
 386         [WinCairo] Unneeded code in method GlyphPage::fill().
 387         https://bugs.webkit.org/show_bug.cgi?id=120634
 388
 389         Reviewed by Andreas Kling.
 390
 391         * platform/graphics/win/GlyphPageTreeNodeCairoWin.cpp:
 392         (WebCore::GlyphPage::fill): Remove unneeded call to GetTextMetrics() function.
 393
 394 2013-09-03  Mihnea Ovidenie  <mihnea@adobe.com>
 395
 396         [CSSRegions] Pseudo-elements as regions should not be exposed to JS
 397         https://bugs.webkit.org/show_bug.cgi?id=120633
 398
 399         Reviewed by Andreas Kling.
 400
 401         Until we properly implement the Region interface (http://dev.w3.org/csswg/css-regions/#the-region-interface)
 402         for pseudo-elements, we should not return these as regions in JS.
 403
 404         Tests: fast/regions/get-regions-by-content-pseudo.html
 405                fast/regions/webkit-named-flow-get-regions-pseudo.html
 406
 407         * dom/WebKitNamedFlow.cpp:
 408         (WebCore::WebKitNamedFlow::firstEmptyRegionIndex): Skip pseudo-elements as regions here too,
 409         otherwise we may get an index that cannot be used with getRegions().
 410         (WebCore::WebKitNamedFlow::getRegionsByContent):
 411         (WebCore::WebKitNamedFlow::getRegions):
 412
 413 2013-09-03  Zan Dobersek  <zdobersek@igalia.com>
 414
 415         REGRESSION(r154967): http appcache tests crashing on WK1
 416         https://bugs.webkit.org/show_bug.cgi?id=120620
 417
 418         Reviewed by Andreas Kling.
 419
 420         * loader/appcache/ApplicationCacheGroup.cpp:
 421         (WebCore::ApplicationCacheGroup::cacheDestroyed): Reintroduce pre-r154967 behavior that returned early in
 422         this method if the passed-in ApplicationCache object was not found in the ApplicationCacheGroup's HashSet
 423         of all the caches. This is now done by checking that the HashSet<T>::remove(T) returns true (meaning the
 424         object was found in the HashSet and removed from it) in addition to that HashSet being subsequently empty
 425         before the method moves on to destroying its ApplicationCacheGroup instance.
 426
 427 2013-09-03  Andre Moreira Magalhaes   <andre.magalhaes@collabora.co.uk>
 428
 429         [gstreamer] Disable HTTP request "Accept-Encoding:" header field on gstreamer source element to avoid receiving the wrong size when retrieving data
 430         https://bugs.webkit.org/show_bug.cgi?id=115354
 431
 432         Reviewed by Philippe Normand.
 433
 434         Also disable Accept-Encoding on ResourceRequest::toSoupMessage accordingly.
 435
 436         * platform/network/soup/ResourceRequestSoup.cpp:
 437         (WebCore::ResourceRequest::toSoupMessage):
 438         Call ResourceRequest::updateSoupMessage from ResourceRequest::toSoupMessage so that the
 439         Accept-Encoding header is also respected.
 440
 441 2013-09-03  Krzysztof Czech  <k.czech@samsung.com>
 442
 443         [AX][ATK] Added support for sort and help attributes.
 444         https://bugs.webkit.org/show_bug.cgi?id=120456
 445
 446         Reviewed by Chris Fleizach.
 447
 448         Added support for aria-sort and aria-help attributes.
 449
 450         Test: accessibility/aria-sort.html
 451
 452         * accessibility/atk/WebKitAccessibleWrapperAtk.cpp:
 453         (webkitAccessibleGetAttributes):
 454
 455 2013-09-03  Allan Sandfeld Jensen  <allan.jensen@digia.com>
 456
 457         [Qt] Remove dead code for QtXmlPatterns
 458         https://bugs.webkit.org/show_bug.cgi?id=120624
 459
 460         Reviewed by Simon Hausmann.
 461
 462         Remove code supporting XSLT using QtXmlPatterns which has been
 463         dead for some time.
 464
 465         * Target.pri:
 466         * WebCore.pri:
 467         * dom/TransformSourceQt.cpp: Removed.
 468         * xml/XSLStyleSheetQt.cpp: Removed.
 469         * xml/XSLTProcessorQt.cpp: Removed.
 470         * xml/parser/XMLDocumentParserQt.cpp:
 471         (WebCore::XMLDocumentParser::doEnd):
 472         (WebCore::XMLDocumentParser::parseProcessingInstruction):
 473
 474 2013-09-03  Mihnea Ovidenie  <mihnea@adobe.com>
 475
 476         [CSS Regions] position: fixed is computed relative to the first region, not the viewport
 477         https://bugs.webkit.org/show_bug.cgi?id=111176
 478
 479         Reviewed by David Hyatt.
 480
 481         Fixed positioned elements inside a named flow should be positioned and sized relative to the viewport,
 482         not on the first region, as described in the spec: http://dev.w3.org/csswg/css-regions/#the-flow-into-property.
 483         While the flow thread will still act as containing block for the fixed positioned elements, the painting and hit
 484         testing for the fixed positioned elements is done by RenderView. The layers for the fixed positioned elements
 485         are collected by the flow thread.
 486
 487         Tests: fast/regions/element-in-named-flow-absolute-from-fixed.html
 488                fast/regions/element-in-named-flow-fixed-from-absolute.html
 489                fast/regions/element-inflow-fixed-from-outflow-static.html
 490                fast/regions/element-outflow-static-from-inflow-fixed.html
 491                fast/regions/fixed-element-transformed-parent.html
 492                fast/regions/fixed-in-named-flow-scroll.html
 493                fast/regions/fixed-inside-fixed-in-named-flow.html
 494                fast/regions/fixed-inside-named-flow-zIndex.html
 495                fast/regions/fixed-pos-elem-in-namedflow-noregions.html
 496                fast/regions/fixed-pos-region-in-nested-flow.html
 497
 498         * rendering/FlowThreadController.cpp:
 499         (WebCore::FlowThreadController::collectFixedPositionedLayers):
 500         Return the list of layers for the fixed positioned elements with named flows as containing blocks.
 501         Used in RenderLayer:: paintFixedLayersInNamedFlows and RenderLayer:: hitTestFixedLayersInNamedFlows.
 502         * rendering/FlowThreadController.h:
 503         * rendering/RenderBox.cpp:
 504         (WebCore::RenderBox::mapLocalToContainer):
 505         (WebCore::RenderBox::containingBlockLogicalWidthForPositioned):
 506         (WebCore::RenderBox::containingBlockLogicalHeightForPositioned):
 507         For fixed positioned elements, width and height are given by the view instead of the first region.
 508         * rendering/RenderLayer.cpp:
 509         (WebCore::accumulateOffsetTowardsAncestor):
 510         Modified for the fixed positioned elements inside named flows with the named flows as containing block
 511         to take into account the view scroll.
 512         (WebCore::compareZIndex):
 513         Moved upwards because it is used in RenderLayer::paintFixedLayersInNamedFlows.
 514         (WebCore::RenderLayer::paintFixedLayersInNamedFlows):
 515         Paint the list of fixed layers directly from the RenderView instead of painting them through regions -> named flows.
 516         (WebCore::RenderLayer::paintLayerContents):
 517         (WebCore::RenderLayer::paintList):
 518         (WebCore::RenderLayer::hitTestFixedLayersInNamedFlows):
 519         Hit test the layers for the fix positioned elements inside named flows from the RenderView layer
 520         instead of the region -> named flow layer.
 521         (WebCore::RenderLayer::hitTestLayer):
 522         (WebCore::RenderLayer::calculateRects):
 523         * rendering/RenderLayer.h:
 524         * rendering/RenderLayerCompositor.cpp:
 525         (WebCore::RenderLayerCompositor::updateLayerTreeGeometry):
 526         We do not support yet the accelerated compositing for elements in named flows,
 527         so bail out early here. We need to revisit fixed elements once we finish accelerated compositing for elements in named flows.
 528         * rendering/RenderObject.cpp:
 529         (WebCore::RenderObject::fixedPositionedWithNamedFlowContainingBlock):
 530         (WebCore::hasFixedPosInNamedFlowContainingBlock):
 531         (WebCore::RenderObject::containerForRepaint):
 532         Changed to take into account that RenderView should be the repaintContainer for the
 533         fixed positioned elements with named flow as the containing block.
 534         * rendering/RenderObject.h:
 535
 536 2013-09-03  Arvid Nilsson  <anilsson@rim.com>
 537
 538         [BlackBerry] Remove LayerData::LayerProgram
 539         https://bugs.webkit.org/show_bug.cgi?id=120601
 540
 541         Reviewed by Anders Carlsson.
 542
 543         JIRA 490672
 544         All layer contents are RGBA now, so there's no need to support BGRA any
 545         more and the LayerData::LayerProgram enum can be removed. Related dead
 546         code was removed.
 547
 548         No new tests, no change in behavior.
 549
 550         * platform/graphics/blackberry/EGLImageLayerWebKitThread.cpp:
 551         (WebCore::EGLImageLayerWebKitThread::EGLImageLayerWebKitThread):
 552         * platform/graphics/blackberry/LayerData.h:
 553         (WebCore::LayerData::LayerData):
 554         * platform/graphics/blackberry/LayerRenderer.cpp:
 555         (WebCore::LayerRenderer::compositeLayersRecursive):
 556         (WebCore::LayerRenderer::createProgram):
 557         * platform/graphics/blackberry/LayerRenderer.h:
 558         * platform/graphics/blackberry/LayerWebKitThread.h:
 559         * platform/graphics/blackberry/PluginLayerWebKitThread.cpp:
 560         (WebCore::PluginLayerWebKitThread::setPluginView):
 561
 562 2013-09-03  Xabier Rodriguez Calvar  <calvaris@igalia.com>
 563
 564         [GStreamer] Video player sets system volume to 100%
 565         https://bugs.webkit.org/show_bug.cgi?id=118974
 566
 567         Reviewed by Philippe Normand.
 568
 569         In order to preserve the system volume we need to keep track of
 570         the volume being initialized in the HTMLMediaElement and then just
 571         setting the volume to the sink when initializing the pipeline if
 572         that volume was changed before.
 573
 574         * html/HTMLMediaElement.cpp:
 575         (WebCore::HTMLMediaElement::HTMLMediaElement): Initialized
 576         attribute to false.
 577         (WebCore::HTMLMediaElement::setVolume): Set the attribute to true
 578         when volume is changed.
 579         (WebCore::HTMLMediaElement::updateVolume): Set the volume only if
 580         volume was initialized.
 581         (WebCore::HTMLMediaElement::mediaPlayerPlatformVolumeConfigurationRequired):
 582         Platform volume configuration is required only if volume was not
 583         initialized before.
 584         * html/HTMLMediaElement.h: Added attribute and interface method.
 585         * platform/graphics/MediaPlayer.h:
 586         (WebCore::MediaPlayerClient::mediaPlayerPlatformVolumeConfigurationRequired):
 587         Declared and added default implementation for the interface method.
 588         (WebCore::MediaPlayer::platformVolumeConfigurationRequired):
 589         Asked the client, meaning the HTMLMediaElement if the platform
 590         volume configuration is required.
 591         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
 592         (WebCore::mediaPlayerPrivateVolumeChangedCallback): Added log.
 593         (WebCore::MediaPlayerPrivateGStreamerBase::setVolume): Added log.
 594         (WebCore::MediaPlayerPrivateGStreamerBase::setStreamVolumeElement):
 595         Set the volume only if not platform volume is required and added log.
 596
 597 2013-09-02  Darin Adler  <darin@apple.com>
 598
 599         * inspector/InspectorProfilerAgent.cpp:
 600         (WebCore::InspectorProfilerAgent::removeProfile): Fix braces here; a review
 601         comment I forgot to address in my last check-in.
 602
 603 2013-09-02  Darin Adler  <darin@apple.com>
 604
 605         Cut down on double hashing and code needlessly using hash table iterators
 606         https://bugs.webkit.org/show_bug.cgi?id=120611
 607
 608         Reviewed by Andreas Kling.
 609
 610         Some of these changes are primarily code cleanup, but others could provide
 611         a small code size and speed improvement by avoiding extra hashing.
 612
 613         * Modules/geolocation/Geolocation.cpp:
 614         (WebCore::Geolocation::Watchers::find): Use get instead of find.
 615         (WebCore::Geolocation::Watchers::remove): Use take instead of find.
 616         (WebCore::Geolocation::makeCachedPositionCallbacks): Use the return
 617         value from remove to avoid hashing twice.
 618
 619         * Modules/webaudio/AudioContext.cpp:
 620         (WebCore::AudioContext::addAutomaticPullNode): Use the return value from
 621         add to avoid hashing twice.
 622         (WebCore::AudioContext::removeAutomaticPullNode): Use the return value
 623         from remove to avoid hashing twice.
 624
 625         * Modules/webaudio/AudioNodeInput.cpp:
 626         (WebCore::AudioNodeInput::connect): Use the return value from add to avoid
 627         hashing twice.
 628         (WebCore::AudioNodeInput::disconnect): Use the return value from remove
 629         to avoid hashing twice.
 630
 631         * Modules/webaudio/AudioParam.cpp:
 632         (WebCore::AudioParam::connect): Use the return value from add to avoid
 633         hashing twice.
 634         (WebCore::AudioParam::disconnect): Use the return value from remove to
 635         avoid hashing twice.
 636
 637         * bridge/NP_jsobject.cpp:
 638         (ObjectMap::remove): Use remove instead of find/remove.
 639
 640         * dom/Node.cpp:
 641         (WebCore::Node::~Node): Use the return value from remove instead of
 642         find/remove.
 643
 644         * inspector/InspectorProfilerAgent.cpp:
 645         (WebCore::InspectorProfilerAgent::removeProfile): Remove needless
 646         calls to contains.
 647
 648         * loader/DocumentLoader.cpp:
 649         (WebCore::DocumentLoader::removeSubresourceLoader): Use the return
 650         value from remove instead of find/remove.
 651
 652         * loader/ResourceLoadScheduler.cpp:
 653         (WebCore::ResourceLoadScheduler::HostInformation::remove): Use the
 654         return value from remove to avoid hashing twice.
 655
 656         * loader/appcache/ApplicationCacheGroup.cpp:
 657         (WebCore::ApplicationCacheGroup::disassociateDocumentLoader): Use
 658         remove instead of find/remove.
 659         (WebCore::ApplicationCacheGroup::cacheDestroyed): Removed a needless
 660         call to contains to avoid hashing twice. It's fine to do the check
 661         for an empty hash table unconditionally.
 662
 663         * page/DOMWindow.cpp:
 664         (WebCore::addUnloadEventListener): Eliminated a local variable for clarity.
 665         (WebCore::removeUnloadEventListener): Ditto. Also use remove instead
 666         of find/remove.
 667         (WebCore::removeAllUnloadEventListeners): Ditto. Also use removeAll instead
 668         of find/removeAll.
 669         (WebCore::addBeforeUnloadEventListener): Ditto.
 670         (WebCore::removeBeforeUnloadEventListener): Ditto.
 671         (WebCore::removeAllBeforeUnloadEventListeners): Ditto.
 672
 673         * page/FrameView.cpp:
 674         (WebCore::FrameView::removeViewportConstrainedObject): Use the return
 675         value from remove to avoid hashing twice.
 676         (WebCore::FrameView::removeScrollableArea): Use the return value from
 677         remove instead of find/remove.
 678         (WebCore::FrameView::containsScrollableArea): Use && instead of an if
 679         statement in a way that is idiomatic for this kind of function.
 680
 681         * page/Page.cpp:
 682         (WebCore::Page::addRelevantRepaintedObject): Use the return value from
 683         remove instead of find/remove.
 684
 685         * page/PageGroup.cpp:
 686         (WebCore::PageGroup::removeUserScriptsFromWorld): Use remove instead
 687         of find/remove.
 688         (WebCore::PageGroup::removeUserStyleSheetsFromWorld): Use the return
 689         value from remove instead of find/remove.
 690
 691         * page/PerformanceUserTiming.cpp:
 692         (WebCore::clearPeformanceEntries): Removed a needless call to contains.
 693
 694         * platform/graphics/DisplayRefreshMonitor.cpp:
 695         (WebCore::DisplayRefreshMonitor::removeClient): Use the return value
 696         from remove instead of find/remove.
 697         (WebCore::DisplayRefreshMonitorManager::displayDidRefresh): Use remove
 698         instead of find/remove.
 699
 700         * platform/graphics/blackberry/LayerRenderer.cpp:
 701         (WebCore::LayerRenderer::removeLayer): Use the return value from remove
 702         instead of find/remove.
 703
 704         * platform/win/WindowMessageBroadcaster.cpp:
 705         (WebCore::WindowMessageBroadcaster::removeListener): Use remove instead
 706         of find/remove. It's fine to do the check for an empty hash table unconditionally.
 707
 708         * plugins/PluginDatabase.cpp:
 709         (WebCore::PluginDatabase::removeDisabledPluginFile): Use the return value
 710         from remove instead of find/remove.
 711
 712         * rendering/style/StyleCustomFilterProgramCache.cpp:
 713         (WebCore::StyleCustomFilterProgramCache::lookup): Use get instead of find.
 714         (WebCore::StyleCustomFilterProgramCache::add): Use contains instead of find
 715         in an assertion.
 716         (WebCore::StyleCustomFilterProgramCache::remove): Use remove instead of
 717         find/remove.
 718
 719         * svg/SVGCursorElement.cpp:
 720         (WebCore::SVGCursorElement::removeClient): Use the return value from remove
 721         instead of find/remove.
 722
 723         * svg/SVGDocumentExtensions.cpp:
 724         (WebCore::SVGDocumentExtensions::removeResource): Removed an unneeded call
 725         to contains.
 726         (WebCore::SVGDocumentExtensions::removeAllTargetReferencesForElement): Use
 727         remove instead of find/remove. It's fine to do the check for an empty hash
 728         table unconditionally.
 729         (WebCore::SVGDocumentExtensions::removeAllElementReferencesForTarget): Use
 730         remove instead of find/remove. Also removed unhelpful assertions. One is
 731         already done by HashMap, and the other is just checking a basic invariant
 732         of every HashMap that doesn't need to be checked.
 733
 734         * svg/graphics/SVGImageCache.cpp:
 735         (WebCore::SVGImageCache::removeClientFromCache): Removed an unneeded call
 736         to contains.
 737
 738         * svg/properties/SVGAnimatedProperty.cpp:
 739         (WebCore::SVGAnimatedProperty::~SVGAnimatedProperty): Use the version of
 740         remove that takes an iterator rather than the one that takes a key, so
 741         we don't need to redo the hashing.
 742
 743 2013-09-02  Andreas Kling  <akling@apple.com>
 744
 745         Generate isFooElement() functions from tagname data.
 746         <https://webkit.org/b/120584>
 747
 748         Reviewed by Antti Koivisto.
 749
 750         Add a "generateTypeChecks" attribute that can be used in HTMLTagNames.in & friends.
 751         If present, isFooElement() style helpers will be added to HTMLElementTypeChecks.h.
 752         This also outputs an isElementOfType<T> check for the Element iterators.
 753
 754         Removed all the hand-written isFooElement() functions that only checked tag name.
 755
 756         * html/HTMLTagNames.in:
 757         * svg/svgtags.in:
 758
 759             Added "generateTypeChecks" attribute as appropriate.
 760
 761         * GNUmakefile.am:
 762         * GNUmakefile.list.am:
 763         * WebCore.vcxproj/WebCore.vcxproj:
 764         * WebCore.xcodeproj/project.pbxproj:
 765
 766             Added to build systems based on how HTMLNames.h was done.
 767             We're just outputting an additional header file in the generated code directory
 768             so I suspect most ports will just pick this up automagically.
 769
 770         * dom/make_names.pl:
 771         (defaultTagPropertyHash):
 772         (printLicenseHeader):
 773         (printTypeChecks):
 774         (printTypeChecksHeaderFile):
 775
 776             Generate a separate file for each namespace with isFooElement() helpers for
 777             elements with "generateTypeChecks" attribute set.
 778
 779 2013-09-02  Darin Adler  <darin@apple.com>
 780
 781         [Mac] No need for HardAutorelease, which is same as CFBridgingRelease
 782         https://bugs.webkit.org/show_bug.cgi?id=120569
 783
 784         Reviewed by Andy Estes.
 785
 786         * accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
 787         (AXTextMarkerRange):
 788         (AXTextMarkerRangeStart):
 789         (AXTextMarkerRangeEnd):
 790         (textMarkerForVisiblePosition):
 791         Use CFBridgingRelease.
 792
 793         * platform/mac/KURLMac.mm:
 794         (WebCore::KURL::operator NSURL *): Use CFBridgingRelease.
 795         (WebCore::KURL::createCFURL): Get rid of needless local variable.
 796
 797         * platform/mac/WebCoreNSURLExtras.mm:
 798         (WebCore::mapHostNameWithRange):
 799         (WebCore::URLWithData):
 800         (WebCore::userVisibleString):
 801         * platform/text/mac/StringImplMac.mm:
 802         (WTF::StringImpl::operator NSString *):
 803         Use CFBridgingRelease.
 804
 805 2013-09-01  Andreas Kling  <akling@apple.com>
 806
 807         Ref: A smart pointer for the reference age.
 808         <https://webkit.org/b/120570>
 809
 810         Reviewed by Antti Koivisto.
 811
 812         Use Ref<T> for various stack guards where null checking isn't needed.
 813
 814 2013-09-02  Andreas Kling  <akling@apple.com>
 815
 816         Simplify DocumentType handling.
 817         <https://webkit.org/b/120529>
 818
 819         Reviewed by Antti Koivisto.
 820
 821         Removed the insertedInto()/removedFrom() handlers from DocumentType.
 822
 823         Document no longer keeps a pointer to its doctype node, it was only used for the
 824         document.doctype DOM API, which now just looks through the list of (<=2) children.
 825
 826         The ENABLE(LEGACY_VIEWPORT_ADAPTION) hunk from Document::setDocType() was moved
 827         into Document::childrenChanged().
 828
 829         We no longer clear the style resolver on doctype insertion/removal since it
 830         doesn't actually affect style anyway.
 831
 832         Also made doctype() return a PassRefPtr<DocumentType> instead of a raw pointer.
 833
 834         * dom/Document.cpp:
 835         (WebCore::Document::dispose):
 836         (WebCore::Document::doctype):
 837         (WebCore::Document::childrenChanged):
 838         * dom/Document.h:
 839         * dom/DocumentType.cpp:
 840         * dom/DocumentType.h:
 841         * editing/markup.cpp:
 842         (WebCore::documentTypeString):
 843
 844 2013-09-02  Anton Obzhirov  <a.obzhirov@samsung.com>
 845
 846         <https://webkit.org/b/98350> [GTK] accessibility/aria-invalid.html times out
 847
 848         Reviewed by Mario Sanchez Prada.
 849
 850         The patch exposes aria-invalid attribute to ATK.
 851
 852         * accessibility/atk/AXObjectCacheAtk.cpp:
 853         (WebCore::AXObjectCache::postPlatformNotification):
 854         Added emitting state-change signal for aria-invalid event.
 855         * accessibility/atk/WebKitAccessibleWrapperAtk.cpp:
 856         (webkitAccessibleGetAttributes):
 857         Added aria-invalid attribute.
 858
 859 2013-09-02  Commit Queue  <commit-queue@webkit.org>
 860
 861         Unreviewed, rolling out r154955.
 862         http://trac.webkit.org/changeset/154955
 863         https://bugs.webkit.org/show_bug.cgi?id=120605
 864
 865         broke xcode4 build :| (Requested by kling on #webkit).
 866
 867         * GNUmakefile.am:
 868         * GNUmakefile.list.am:
 869         * WebCore.vcxproj/WebCore.vcxproj:
 870         * WebCore.xcodeproj/project.pbxproj:
 871         * dom/make_names.pl:
 872         (defaultTagPropertyHash):
 873         (printLicenseHeader):
 874         * html/HTMLAnchorElement.h:
 875         (WebCore::isHTMLAnchorElement):
 876         (WebCore::HTMLAnchorElement):
 877         * html/HTMLAreaElement.h:
 878         (WebCore::isHTMLAreaElement):
 879         (WebCore::HTMLAreaElement):
 880         * html/HTMLAudioElement.h:
 881         (WebCore::isHTMLAudioElement):
 882         * html/HTMLBaseElement.h:
 883         (WebCore::isHTMLBaseElement):
 884         (WebCore::HTMLBaseElement):
 885         * html/HTMLCanvasElement.h:
 886         (WebCore::isHTMLCanvasElement):
 887         * html/HTMLElement.h:
 888         * html/HTMLFieldSetElement.h:
 889         (WebCore::isHTMLFieldSetElement):
 890         (WebCore::HTMLFieldSetElement):
 891         * html/HTMLFormElement.h:
 892         (WebCore::isHTMLFormElement):
 893         * html/HTMLFrameSetElement.h:
 894         (WebCore::isHTMLFrameSetElement):
 895         (WebCore::HTMLFrameSetElement):
 896         * html/HTMLImageElement.h:
 897         (WebCore::isHTMLImageElement):
 898         * html/HTMLInputElement.h:
 899         (WebCore::isHTMLInputElement):
 900         * html/HTMLLabelElement.h:
 901         (WebCore::isHTMLLabelElement):
 902         (WebCore::HTMLLabelElement):
 903         * html/HTMLLegendElement.h:
 904         (WebCore::isHTMLLegendElement):
 905         (WebCore::HTMLLegendElement):
 906         * html/HTMLMapElement.h:
 907         (WebCore::isHTMLMapElement):
 908         * html/HTMLMeterElement.h:
 909         (WebCore::isHTMLMeterElement):
 910         * html/HTMLOptGroupElement.h:
 911         (WebCore::isHTMLOptGroupElement):
 912         * html/HTMLOptionElement.h:
 913         (WebCore::isHTMLOptionElement):
 914         * html/HTMLParamElement.h:
 915         (WebCore::isHTMLParamElement):
 916         (WebCore::HTMLParamElement):
 917         * html/HTMLProgressElement.h:
 918         (WebCore::isHTMLProgressElement):
 919         * html/HTMLScriptElement.h:
 920         (WebCore::isHTMLScriptElement):
 921         * html/HTMLSourceElement.h:
 922         (WebCore::isHTMLSourceElement):
 923         (WebCore::HTMLSourceElement):
 924         * html/HTMLStyleElement.h:
 925         (WebCore::isHTMLStyleElement):
 926         (WebCore::HTMLStyleElement):
 927         * html/HTMLTableElement.h:
 928         (WebCore::isHTMLTableElement):
 929         * html/HTMLTableRowElement.h:
 930         (WebCore::isHTMLTableRowElement):
 931         (WebCore::HTMLTableRowElement):
 932         * html/HTMLTagNames.in:
 933         * html/HTMLTextAreaElement.h:
 934         (WebCore::isHTMLTextAreaElement):
 935         * html/HTMLTitleElement.h:
 936         (WebCore::isHTMLTitleElement):
 937         (WebCore::HTMLTitleElement):
 938         * html/HTMLTrackElement.h:
 939         (WebCore::isHTMLTrackElement):
 940         (WebCore::HTMLTrackElement):
 941         * svg/SVGElement.h:
 942         * svg/SVGFontElement.h:
 943         (WebCore::isSVGFontElement):
 944         * svg/SVGFontFaceElement.h:
 945         (WebCore::isSVGFontFaceElement):
 946         (WebCore::SVGFontFaceElement):
 947         * svg/SVGForeignObjectElement.h:
 948         (WebCore::isSVGForeignObjectElement):
 949         (WebCore::SVGForeignObjectElement):
 950         * svg/SVGImageElement.h:
 951         (WebCore::isSVGImageElement):
 952         * svg/SVGScriptElement.h:
 953         (WebCore::isSVGScriptElement):
 954         * svg/svgtags.in:
 955
 956 2013-09-02  Antti Koivisto  <antti@apple.com>
 957
 958         Clean up ContainerNode::childrenChanged
 959         https://bugs.webkit.org/show_bug.cgi?id=120599
 960
 961         Reviewed by Andreas Kling.
 962
 963         - Make childrenChanged take a single struct argument instead of a long list of arguments.
 964         - Use enum instead of childCountDelta. It was always -1, 0, 1 or the total number of children (in case of removing them all).
 965         - Remove use of Node*, give the change range as Elements.
 966         - Related cleanups.
 967
 968         * dom/Attr.cpp:
 969         (WebCore::Attr::childrenChanged):
 970         * dom/Attr.h:
 971         * dom/CharacterData.cpp:
 972         (WebCore::CharacterData::parserAppendData):
 973         (WebCore::CharacterData::dispatchModifiedEvent):
 974         * dom/ContainerNode.cpp:
 975         (WebCore::ContainerNode::insertBefore):
 976         (WebCore::ContainerNode::notifyChildInserted):
 977         (WebCore::ContainerNode::notifyChildRemoved):
 978
 979             Add private helpers for setting up the struct.
 980
 981         (WebCore::ContainerNode::parserInsertBefore):
 982         (WebCore::ContainerNode::replaceChild):
 983         (WebCore::ContainerNode::removeChild):
 984         (WebCore::ContainerNode::parserRemoveChild):
 985         (WebCore::ContainerNode::removeChildren):
 986         (WebCore::ContainerNode::appendChild):
 987         (WebCore::ContainerNode::parserAppendChild):
 988         (WebCore::ContainerNode::childrenChanged):
 989         (WebCore::ContainerNode::updateTreeAfterInsertion):
 990         * dom/ContainerNode.h:
 991         * dom/Document.cpp:
 992         (WebCore::Document::childrenChanged):
 993         * dom/Document.h:
 994         * dom/Element.cpp:
 995         (WebCore::checkForSiblingStyleChanges):
 996
 997             Clean up and simplify. Since we now get element range automatically we don't need to compute it.
 998
 999         (WebCore::Element::childrenChanged):
1000         (WebCore::Element::finishParsingChildren):
1001         * dom/Element.h:
1002         * dom/ShadowRoot.cpp:
1003         (WebCore::ShadowRoot::childrenChanged):
1004         * dom/ShadowRoot.h:
1005         * html/HTMLElement.cpp:
1006         (WebCore::HTMLElement::childrenChanged):
1007         (WebCore::HTMLElement::adjustDirectionalityIfNeededAfterChildrenChanged):
1008
1009             Try to keep the existing behavior. This code needs  more cleanup to be sane. It shouldn't operate on Nodes
1010             as it only really cares about Elements.
1011
1012         * html/HTMLElement.h:
1013         * html/HTMLFieldSetElement.cpp:
1014         (WebCore::HTMLFieldSetElement::childrenChanged):
1015         * html/HTMLFieldSetElement.h:
1016         * html/HTMLObjectElement.cpp:
1017         (WebCore::HTMLObjectElement::childrenChanged):
1018         * html/HTMLObjectElement.h:
1019         * html/HTMLOptGroupElement.cpp:
1020         (WebCore::HTMLOptGroupElement::childrenChanged):
1021         * html/HTMLOptGroupElement.h:
1022         * html/HTMLOptionElement.cpp:
1023         (WebCore::HTMLOptionElement::childrenChanged):
1024         * html/HTMLOptionElement.h:
1025         * html/HTMLOutputElement.cpp:
1026         (WebCore::HTMLOutputElement::childrenChanged):
1027         * html/HTMLOutputElement.h:
1028         * html/HTMLScriptElement.cpp:
1029         (WebCore::HTMLScriptElement::childrenChanged):
1030         * html/HTMLScriptElement.h:
1031         * html/HTMLSelectElement.cpp:
1032         (WebCore::HTMLSelectElement::childrenChanged):
1033         * html/HTMLSelectElement.h:
1034         * html/HTMLStyleElement.cpp:
1035         (WebCore::HTMLStyleElement::childrenChanged):
1036         * html/HTMLStyleElement.h:
1037         * html/HTMLTextAreaElement.cpp:
1038         (WebCore::HTMLTextAreaElement::childrenChanged):
1039         * html/HTMLTextAreaElement.h:
1040         * html/HTMLTitleElement.cpp:
1041         (WebCore::HTMLTitleElement::childrenChanged):
1042         * html/HTMLTitleElement.h:
1043         * html/shadow/InsertionPoint.cpp:
1044         (WebCore::InsertionPoint::childrenChanged):
1045         * html/shadow/InsertionPoint.h:
1046         * svg/SVGClipPathElement.cpp:
1047         (WebCore::SVGClipPathElement::childrenChanged):
1048         * svg/SVGClipPathElement.h:
1049         * svg/SVGElement.cpp:
1050         (WebCore::SVGElement::childrenChanged):
1051         * svg/SVGElement.h:
1052         * svg/SVGFELightElement.cpp:
1053         (WebCore::SVGFELightElement::childrenChanged):
1054         * svg/SVGFELightElement.h:
1055         * svg/SVGFilterElement.cpp:
1056         (WebCore::SVGFilterElement::childrenChanged):
1057         * svg/SVGFilterElement.h:
1058         * svg/SVGFilterPrimitiveStandardAttributes.cpp:
1059         (WebCore::SVGFilterPrimitiveStandardAttributes::childrenChanged):
1060         * svg/SVGFilterPrimitiveStandardAttributes.h:
1061         * svg/SVGFontFaceElement.cpp:
1062         (WebCore::SVGFontFaceElement::childrenChanged):
1063         * svg/SVGFontFaceElement.h:
1064         * svg/SVGFontFaceFormatElement.cpp:
1065         (WebCore::SVGFontFaceFormatElement::childrenChanged):
1066         * svg/SVGFontFaceFormatElement.h:
1067         * svg/SVGFontFaceSrcElement.cpp:
1068         (WebCore::SVGFontFaceSrcElement::childrenChanged):
1069         * svg/SVGFontFaceSrcElement.h:
1070         * svg/SVGFontFaceUriElement.cpp:
1071         (WebCore::SVGFontFaceUriElement::childrenChanged):
1072         * svg/SVGFontFaceUriElement.h:
1073         * svg/SVGGradientElement.cpp:
1074         (WebCore::SVGGradientElement::childrenChanged):
1075         * svg/SVGGradientElement.h:
1076         * svg/SVGMarkerElement.cpp:
1077         (WebCore::SVGMarkerElement::childrenChanged):
1078         * svg/SVGMarkerElement.h:
1079         * svg/SVGMaskElement.cpp:
1080         (WebCore::SVGMaskElement::childrenChanged):
1081         * svg/SVGMaskElement.h:
1082         * svg/SVGPatternElement.cpp:
1083         (WebCore::SVGPatternElement::childrenChanged):
1084         * svg/SVGPatternElement.h:
1085         * svg/SVGScriptElement.cpp:
1086         (WebCore::SVGScriptElement::childrenChanged):
1087         * svg/SVGScriptElement.h:
1088         * svg/SVGStyleElement.cpp:
1089         (WebCore::SVGStyleElement::childrenChanged):
1090         * svg/SVGStyleElement.h:
1091         * svg/SVGTitleElement.cpp:
1092         (WebCore::SVGTitleElement::childrenChanged):
1093         * svg/SVGTitleElement.h:
1094
1095 2013-09-02  Andreas Kling  <akling@apple.com>
1096
1097         Generate isFooElement() functions from tagname data.
1098         <https://webkit.org/b/120584>
1099
1100         Reviewed by Antti Koivisto.
1101
1102         Add a "generateTypeChecks" attribute that can be used in HTMLTagNames.in & friends.
1103         If present, isFooElement() style helpers will be added to HTMLElementTypeChecks.h.
1104         This also outputs an isElementOfType<T> check for the Element iterators.
1105
1106         Removed all the hand-written isFooElement() functions that only checked tag name.
1107
1108         * html/HTMLTagNames.in:
1109         * svg/svgtags.in:
1110
1111             Added "generateTypeChecks" attribute as appropriate.
1112
1113         * GNUmakefile.am:
1114         * GNUmakefile.list.am:
1115         * WebCore.vcxproj/WebCore.vcxproj:
1116         * WebCore.xcodeproj/project.pbxproj:
1117
1118             Added to build systems based on how HTMLNames.h was done.
1119             We're just outputting an additional header file in the generated code directory
1120             so I suspect most ports will just pick this up automagically.
1121
1122         * dom/make_names.pl:
1123         (defaultTagPropertyHash):
1124         (printLicenseHeader):
1125         (printTypeChecks):
1126         (printTypeChecksHeaderFile):
1127
1128             Generate a separate file for each namespace with isFooElement() helpers for
1129             elements with "generateTypeChecks" attribute set.
1130
1131 2013-09-02  Dirk Schulze  <krit@webkit.org>
1132
1133         Use edgeMode=duplicate for blurring on filter() function
1134         https://bugs.webkit.org/show_bug.cgi?id=120590
1135
1136         Reviewed by Antti Koivisto.
1137
1138         Filters on the CSS Image function filter() are not allowed to extend the
1139         dimension of the input image. This causes weird results on blurring an image,
1140         where the fading on the edges is clipped at the half of the fading.
1141         We shouldn't fade edges at all and use the edgeMode=duplicate instead.
1142         This will duplicate the pixel value on the nearest edge of the input image
1143         instead of taking transparent black and results in nice blurred images with
1144         sharp edges.
1145
1146         Spec: http://dev.w3.org/fxtf/filters/#blurEquivalent
1147
1148         Test: fast/filter-image/filter-image-blur.html
1149
1150         * css/CSSFilterImageValue.cpp: Pass consumer information to the renderer.
1151         (WebCore::CSSFilterImageValue::image):
1152         * rendering/FilterEffectRenderer.cpp: Set edgeMode for feGaussianBlur to
1153             'duplicate' or 'none' depending on the consumer.
1154         (WebCore::FilterEffectRenderer::build):
1155         * rendering/FilterEffectRenderer.h: Add enumeration to differ between the
1156             different consumers of the renderer.
1157         * rendering/RenderLayer.cpp:
1158         (WebCore::RenderLayer::updateOrRemoveFilterEffectRenderer):  Pass consumer
1159             information to the renderer.
1160
1161 2013-09-01  Dirk Schulze  <krit@webkit.org>
1162
1163         Add 'edgeMode' attribute to SVGFEGaussianBlur
1164         https://bugs.webkit.org/show_bug.cgi?id=120582
1165
1166         Add 'edgeMode' attribute to the SVGFEGaussianBlur element. This attribute
1167         allows users to define the behavior on edges with the values 'none' where
1168         pixel values outside the input image are treated as transparent black. (The
1169         current blurring behavior.) 'duplicate' which repeats the values on the
1170         nearest edge and 'warp', which takes the pixel of the opposite site of
1171         the input image.
1172         Beside the attribute, this patch implements the behavior of 'duplicate'.
1173
1174         http://dev.w3.org/fxtf/filters/#feGaussianBlurEdgeModeAttribute
1175
1176         Reviewed by Rob Buis.
1177
1178         Tests: svg/dynamic-updates/SVGFEGaussianBlurElement-dom-edgeMode-attr.html
1179                svg/dynamic-updates/SVGFEGaussianBlurElement-svgdom-edgeMode-prop.html
1180                svg/filters/svg-gaussianblur-edgeMode-duplicate-expected.svg
1181                svg/filters/svg-gaussianblur-edgeMode-duplicate.svg
1182
1183         * platform/graphics/filters/FEGaussianBlur.cpp:
1184         (WebCore::FEGaussianBlur::FEGaussianBlur):
1185         (WebCore::FEGaussianBlur::create):
1186         (WebCore::FEGaussianBlur::edgeMode):
1187         (WebCore::FEGaussianBlur::setEdgeMode):
1188         (WebCore::boxBlur):
1189         (WebCore::FEGaussianBlur::platformApplyGeneric):
1190         (WebCore::FEGaussianBlur::determineAbsolutePaintRect):
1191         * platform/graphics/filters/FEGaussianBlur.h:
1192         * rendering/FilterEffectRenderer.cpp:
1193         (WebCore::FilterEffectRenderer::build):
1194         * svg/SVGFEGaussianBlurElement.cpp:
1195         (WebCore::SVGFEGaussianBlurElement::SVGFEGaussianBlurElement):
1196         (WebCore::SVGFEGaussianBlurElement::isSupportedAttribute):
1197         (WebCore::SVGFEGaussianBlurElement::parseAttribute):
1198         (WebCore::SVGFEGaussianBlurElement::svgAttributeChanged):
1199         (WebCore::SVGFEGaussianBlurElement::build):
1200         * svg/SVGFEGaussianBlurElement.h:
1201         * svg/SVGFEGaussianBlurElement.idl:
1202
1203 2013-09-01  Andreas Kling  <akling@apple.com>
1204
1205         EventHandler::m_frame should be a Frame&.
1206         <https://webkit.org/b/120580>
1207
1208         Reviewed by Darin Adler.
1209
1210         EventHandler is tied to the lifetime of its frame, so let m_frame be a Frame&.
1211         A handful of null checks and assertions removed.
1212
1213 2013-09-01  Darin Adler  <darin@apple.com>
1214
1215         Fix a mistake in my recent pasteboard/editor refactoring that was causing tests to fail.
1216
1217         * editing/mac/EditorMac.mm:
1218         (WebCore::getImage): One place this said cachedImage but it should have said tentativeCachedImage.
1219
1220 2013-09-01  Darin Adler  <darin@apple.com>
1221
1222         HitTestResult should have innerNonSharedElement
1223         https://bugs.webkit.org/show_bug.cgi?id=120579
1224
1225         Reviewed by Andreas Kling.
1226
1227         * editing/Editor.cpp:
1228         (WebCore::Editor::copyImage): Call HitTestResult member function version of
1229         innerNonSharedElement instead of a local function that does it.
1230
1231         * page/Chrome.cpp:
1232         (WebCore::Chrome::setToolTip): Use innerNonSharedElement instead of getting
1233         the node and checking if it's an input element. Also added some missing braces.
1234
1235         * page/EventHandler.cpp:
1236         (WebCore::EventHandler::selectClosestWordFromHitTestResult): Use targetNode for
1237         local variables instead of innerNode to match the HitTestResult function name.
1238         (WebCore::EventHandler::selectClosestWordOrLinkFromMouseEvent): Ditto.
1239         (WebCore::EventHandler::handleMousePressEventTripleClick): Ditto.
1240         (WebCore::EventHandler::handleMousePressEventSingleClick): Ditto.
1241         (WebCore::EventHandler::handleMousePressEvent): Ditto.
1242
1243         * rendering/HitTestResult.cpp:
1244         (WebCore::HitTestResult::innerElement): Rewrote so there there is no loop.
1245         (WebCore::HitTestResult::innerNonSharedElement): Ditto.
1246
1247         * rendering/HitTestResult.h: Added innerNonSharedElement. Generally speaking,
1248         we'd like to avoid using Node unless there is some real need.
1249
1250 2013-09-01  Xabier Rodriguez Calvar  <calvaris@igalia.com>
1251
1252         Volume slider value should be 0 when audio is muted
1253         https://bugs.webkit.org/show_bug.cgi?id=120553
1254
1255         Reviewed by Eric Carlson.
1256
1257         Fixed the problem of showing a non empty slider when audio is
1258         muted.
1259
1260         Test: media/volume-bar-empty-when-muted.html.
1261
1262         * html/shadow/MediaControls.cpp:
1263         (WebCore::MediaControls::reset): Use setSliderVolume.
1264         (WebCore::MediaControls::changedVolume): Use setSliderVolume.
1265         (WebCore::MediaControls::setSliderVolume): Added to set the volume
1266         to 0 when muted and to its value otherwise.
1267         * html/shadow/MediaControls.h: Added setSliderVolume.
1268         * html/shadow/MediaControlsApple.cpp:
1269         (WebCore::MediaControlsApple::reset): Used setSliderVolume and
1270         setFullscreenSliderVolume.
1271         (WebCore::MediaControlsApple::changedVolume): Used
1272         setFullscreenSliderVolume.
1273         (WebCore::MediaControlsApple::setFullscreenSliderVolume): Added to
1274         set the volume to 0 when muted and to its value otherwise.
1275         * html/shadow/MediaControlsApple.h: Added setFullscreenSliderVolume
1276         * html/shadow/MediaControlsBlackBerry.cpp:
1277         (WebCore::MediaControlsBlackBerry::reset): Used setSliderVolume.
1278
1279 2013-09-01  Andreas Kling  <akling@apple.com>
1280
1281         Give EditCommand a protected Frame& getter.
1282         <https://webkit.org/b/120574>
1283
1284         Reviewed by Darin Adler.
1285
1286         EditCommand is only created for documents that are attached to a Frame,
1287         we already ASSERTed as much in the EditCommand constructor.
1288
1289         This patch adds a "Frame& EditCommand::frame()" helper, so EditCommand
1290         and its subclasses don't have to fumble around with pointers.
1291
1292 2013-08-31  Antti Koivisto  <antti@apple.com>
1293
1294         Add element ancestor iterator
1295         https://bugs.webkit.org/show_bug.cgi?id=120563
1296
1297         Reviewed by Andreas Kling.
1298
1299         This patch adds ancestor iterators. They iterate over elements parent chain up to the root.
1300
1301         To iterate over Element ancestors:
1302
1303         auto ancestors = elementAncestors(this);
1304         for (auto it = ancestors.begin(), end = ancestors.end(); it != end; ++it) {
1305             Element& element = *it;
1306             ...
1307
1308         To iterate over Element ancestors including the current element:
1309
1310         auto lineage = elementLineage(this);
1311         for (auto it = lineage.begin(), end = lineage.end(); it != end; ++it) {
1312             Element& element = *it;
1313             ...
1314
1315         To iterate over ancestors of a specific Element subclass:
1316
1317         auto htmlAncestors = ancestorsOfType<HTMLElement>(this);
1318         for (auto it = htmlAncestors.begin(), end = htmlAncestors.end(); it != end; ++it) {
1319             HTMLElement& htmlElement = *it;
1320             ...
1321
1322         To iterate over ancestors of a specific Element subclass including the current element:
1323
1324         auto htmlLineage = lineageOfType<HTMLElement>(this);
1325         for (auto it = htmlLineage.begin(), end = htmlLineage.end(); it != end; ++it) {
1326             HTMLElement& htmlElement = *it;
1327             ...
1328
1329         The patch also uses the new types in a few places.
1330
1331         * WebCore.xcodeproj/project.pbxproj:
1332         * accessibility/AccessibilityNodeObject.cpp:
1333         (WebCore::AccessibilityNodeObject::mouseButtonListener):
1334         (WebCore::AccessibilityNodeObject::labelForElement):
1335         * dom/ElementAncestorIterator.h: Added.
1336         (WebCore::::ElementAncestorIterator):
1337         (WebCore::::operator):
1338         (WebCore::::ElementAncestorConstIterator):
1339         (WebCore::::ElementAncestorIteratorAdapter):
1340         (WebCore::::begin):
1341         (WebCore::::end):
1342         (WebCore::::ElementAncestorConstIteratorAdapter):
1343         (WebCore::elementLineage):
1344         (WebCore::elementAncestors):
1345         (WebCore::ancestorsOfType):
1346         * dom/ElementIterator.h:
1347         (WebCore::findElementAncestorOfType):
1348         (WebCore::::traverseAncestor):
1349         (WebCore::=):
1350         * html/HTMLElement.cpp:
1351         (WebCore::HTMLElement::adjustDirectionalityIfNeededAfterChildAttributeChanged):
1352         * html/HTMLFieldSetElement.h:
1353         (WebCore::isHTMLFieldSetElement):
1354         (WebCore::HTMLFieldSetElement):
1355         * html/HTMLFrameSetElement.cpp:
1356         (WebCore::HTMLFrameSetElement::findContaining):
1357         * html/HTMLFrameSetElement.h:
1358         (WebCore::HTMLFrameSetElement):
1359         * html/HTMLInputElement.h:
1360         (WebCore::isHTMLInputElement):
1361         (WebCore::toHTMLInputElement):
1362         * html/HTMLLegendElement.cpp:
1363         (WebCore::HTMLLegendElement::associatedControl):
1364
1365 2013-08-31  Darin Adler  <darin@apple.com>
1366
1367         Refactor URL and image writing so layer-violating parts are in Editor, not Pasteboard (Mac-only at first)
1368         https://bugs.webkit.org/show_bug.cgi?id=120573
1369
1370         Reviewed by Andreas Kling.
1371
1372         * editing/Editor.cpp:
1373         (WebCore::Editor::copy): Use writeImageToPasteboard instead of Pasteboard::writeImage on Mac.
1374         (WebCore::Editor::copyURL): Use writeURLToPasteboard instead of Pasteboard::writeURL on Mac.
1375         (WebCore::innerNonSharedElement): Added. Used in copyImage.
1376         (WebCore::Editor::copyImage): Use innerNonSharedElement instead of innerNonSharedNode.
1377         Use writeImageToPasteboard instead of Pasteboard::writeImage on Mac.
1378
1379         * editing/Editor.h: Add new functions, writeURLToPasteboard and writeImageToPasteboard,
1380         both Mac-only for now.
1381
1382         * editing/mac/EditorMac.mm:
1383         (WebCore::Editor::pasteWithPasteboard): Removed unneeded "m_frame.editor()" round trip that
1384         was left behind in this function.
1385         (WebCore::getImage): Added. Helper used by writeImageToPasteboard.
1386         (WebCore::Editor::writeURLToPasteboard): Added. Sets up PasteboardURL and then calls
1387         Pasteboard::write with it.
1388         (WebCore::Editor::writeImageToPasteboard): Added. Sets up PasteboardImage and then calls
1389         Pasteboard::write with it.
1390         * page/DragController.cpp:
1391         (WebCore::DragController::startDrag): Use Editor::writeURLToPasteboard instead of
1392         Pasteboard::writeURL on Mac.
1393
1394         * platform/Pasteboard.h: Added PasteboardURL and PasteboardImage structures.
1395         Declare write functions for PasteboardWebContent, PasteboardURL, and PasteboardImage.
1396         Guard writeURL and writeImage with !PLATFORM(MAC).
1397
1398         * platform/mac/PasteboardMac.mm:
1399         (WebCore::writeURLForTypes): Changed this to take a PasteboardURL.
1400         (WebCore::Pasteboard::write): Renamed writeURL and writeImage to this. Refactor both
1401         to take PasteboardURL and PasteboardImage arguments.
1402         (WebCore::fileWrapper): Renamed from fileWrapperForImage since the fact that this is
1403         for an image is now clear from its argument, a PasteboardImage.
1404         (WebCore::writeFileWrapperAsRTFDAttachment): Changed this function to use early return.
1405
1406 2013-08-31  Andreas Kling  <akling@apple.com>
1407
1408         EditCommand constructors should take Document&.
1409         <https://webkit.org/b/120566>
1410
1411         Reviewed by Darin Adler.
1412
1413         Let EditCommand's and all subclass constructors take Document& instead of Document*.
1414         EditCommand::document() now returns a Document&.
1415
1416         Added Editor::document() which returns Editor::frame().document() after asserting
1417         that it's non-null, to make passing a Document& from Editor functions easier.
1418
1419 2013-08-31  Andreas Kling  <akling@apple.com>
1420
1421         Make more use of toRenderFoo(RenderObject&) cast helpers.
1422         <https://webkit.org/b/120565>
1423
1424         Reviewed by Darin Adler.
1425
1426         To make amends for putting ugly toRenderFoo(&renderer) casts everywhere,
1427         here's a patch that switches a bunch of code over to using reference-based casts.
1428
1429         I removed pointer-based casts altogether for these renderers:
1430
1431             - RenderBR
1432             - RenderCombineText
1433             - RenderListMarker
1434             - RenderVideo
1435             - RenderView
1436
1437 2013-08-31  Santosh Mahto  <santosh.ma@samsung.com>
1438
1439         warning: unused parameter ‘renderingMode'  in ImageBufferCairo.cpp:94
1440         https://bugs.webkit.org/show_bug.cgi?id=120543
1441
1442         Reviewed by Darin Adler.
1443
1444         Fixing warning by adding ASSERT_UNUSED.
1445
1446         * platform/graphics/cairo/ImageBufferCairo.cpp:
1447         (WebCore::ImageBuffer::ImageBuffer):
1448
1449 2013-08-31  Andreas Kling  <akling@apple.com>
1450
1451         Fix typo in r154931 that caused assertions in continuation tests.
1452
1453         * rendering/InlineFlowBox.cpp:
1454         (WebCore::InlineFlowBox::paint):
1455
1456 2013-08-31  Rob Buis  <rwlbuis@webkit.org>
1457
1458         XMLSerializer-attribute-namespace-prefix-conflicts can't produce reliable results
1459         https://bugs.webkit.org/show_bug.cgi?id=120490
1460
1461         Reviewed by Ryosuke Niwa.
1462
1463         Retain the key of the namespaces map by using AtomicString as key instead of AtomicStringImpl*.
1464
1465         * editing/MarkupAccumulator.h:
1466
1467 2013-08-31  Andreas Kling  <akling@apple.com>
1468
1469         InlineBox::renderer() and pals should return references.
1470         <https://webkit.org/b/120562>
1471
1472         Reviewed by Antti Koivisto.
1473
1474         Make InlineBox::renderer() and its friends return references instead of pointers.
1475         Every box always has a renderer, so this clears up any ambiguity, and a number of
1476         weird assertions in SVG code.
1477
1478         Made InlineBox::m_renderer private so subclasses have to go through renderer().
1479         The next step will be to replace renderer() with tightly-typed accessors where
1480         applicable (e.g InlineTextBox's renderer is always a RenderText.)
1481
1482         Also added reference versions of these RenderObject casts:
1483
1484             - toRenderBox()
1485             - toRenderBlock()
1486             - toRenderInline()
1487             - toRenderSVGInlineText()
1488
1489 2013-08-31  Andreas Kling  <akling@apple.com>
1490
1491         PostAttachCallbackDisabler should take a ContainerNode&.
1492         <https://webkit.org/b/120560>
1493
1494         Reviewed by Antti Koivisto.
1495
1496         Let PostAttachCallbackDisabler take a reference and remove an ugly assertion.
1497
1498         * dom/ContainerNode.h:
1499         (WebCore::PostAttachCallbackDisabler::PostAttachCallbackDisabler):
1500         (WebCore::PostAttachCallbackDisabler::~PostAttachCallbackDisabler):
1501         * dom/Document.cpp:
1502         (WebCore::Document::recalcStyle):
1503         * style/StyleResolveTree.cpp:
1504         (WebCore::Style::attachRenderTree):
1505
1506 2013-08-31  Antti Koivisto  <antti@apple.com>
1507
1508         Rename ChildIterator -> ElementChildIterator, DescendantIterator -> ElementDescendantIterator
1509         https://bugs.webkit.org/show_bug.cgi?id=120561
1510
1511         Reviewed by Andreas Kling.
1512
1513         New names are more descriptive and consistent. They also put related iterator files close by in file lists.
1514         The types are always used with auto so longer names don't make them any clumsier to use.
1515
1516         For simplicity included all Element iterator types from ElementIterator.h so it is the only file that has
1517         to be included by the clients.
1518
1519         * GNUmakefile.list.am:
1520         * Target.pri:
1521         * WebCore.vcxproj/WebCore.vcxproj:
1522         * WebCore.xcodeproj/project.pbxproj:
1523         * accessibility/AccessibilityNodeObject.cpp:
1524         * accessibility/AccessibilityRenderObject.cpp:
1525         * accessibility/AccessibilityTable.cpp:
1526         * css/CSSFontFaceSource.cpp:
1527         * dom/ChildIterator.h: Removed.
1528         * dom/DescendantIterator.h: Removed.
1529         * dom/Document.cpp:
1530         * dom/ElementChildIterator.h: Copied from Source/WebCore/dom/ChildIterator.h.
1531         (WebCore::::ElementChildIterator):
1532         (WebCore::::operator):
1533         (WebCore::::ElementChildConstIterator):
1534         (WebCore::::ElementChildIteratorAdapter):
1535         (WebCore::::begin):
1536         (WebCore::::end):
1537         (WebCore::::ElementChildConstIteratorAdapter):
1538         (WebCore::elementChildren):
1539         (WebCore::childrenOfType):
1540         * dom/ElementDescendantIterator.h: Copied from Source/WebCore/dom/DescendantIterator.h.
1541         (WebCore::::ElementDescendantIterator):
1542         (WebCore::::operator):
1543         (WebCore::::ElementDescendantConstIterator):
1544         (WebCore::::ElementDescendantIteratorAdapter):
1545         (WebCore::::begin):
1546         (WebCore::::end):
1547         (WebCore::::ElementDescendantConstIteratorAdapter):
1548         (WebCore::elementDescendants):
1549         (WebCore::descendantsOfType):
1550         * dom/ElementIterator.h:
1551         * dom/Node.cpp:
1552         * dom/TreeScope.cpp:
1553         * editing/ApplyStyleCommand.cpp:
1554         * editing/markup.cpp:
1555         * html/HTMLAppletElement.cpp:
1556         * html/HTMLFieldSetElement.cpp:
1557         * html/HTMLLabelElement.cpp:
1558         * html/HTMLMapElement.cpp:
1559         * html/HTMLMediaElement.cpp:
1560         * html/HTMLObjectElement.cpp:
1561         * loader/PlaceholderDocument.cpp:
1562         * rendering/FilterEffectRenderer.cpp:
1563         * style/StyleResolveTree.cpp:
1564         * svg/SVGElement.cpp:
1565         * svg/SVGSVGElement.cpp:
1566         * svg/animation/SMILTimeContainer.cpp:
1567         * svg/graphics/SVGImage.cpp:
1568
1569 2013-08-31  Andreas Kling  <akling@apple.com>
1570
1571         Don't do document style recalc unless there's a RenderView.
1572         <https://webkit.org/b/120558>
1573
1574         Reviewed by Antti Koivisto.
1575
1576         There's no sense in computing style for a Document that has no RenderView.
1577         Checking this before continuing also lets us know that there's a Frame & FrameView
1578         present, simplifying some things later on.
1579
1580         * dom/Document.cpp:
1581         (WebCore::Document::recalcStyle):
1582         * style/StyleResolveForDocument.cpp:
1583         (WebCore::Style::resolveForDocument):
1584
1585 2013-08-31  Antti Koivisto  <antti@apple.com>
1586
1587         Add common base for element iterators
1588         https://bugs.webkit.org/show_bug.cgi?id=120557
1589
1590         Reviewed by Andreas Kling.
1591
1592         Add ElementIterator and use it as the base for child and descendant iterators. The only difference is the definition of operator++.
1593
1594         Also renamed DescendantIteratorAssertions to ElementIteratorAssertions.
1595
1596         * GNUmakefile.list.am:
1597         * Target.pri:
1598         * WebCore.vcxproj/WebCore.vcxproj:
1599         * WebCore.xcodeproj/project.pbxproj:
1600         * dom/ChildIterator.h:
1601         (WebCore::::ChildIterator):
1602         (WebCore::::operator):
1603         (WebCore::::ChildConstIterator):
1604         (WebCore::::begin):
1605         (WebCore::::end):
1606         * dom/DescendantIterator.h:
1607         (WebCore::::DescendantIterator):
1608         (WebCore::::operator):
1609         (WebCore::::DescendantConstIterator):
1610         * dom/DescendantIteratorAssertions.h: Removed.
1611         * dom/ElementIterator.h: Added.
1612         (WebCore::::ElementIterator):
1613         (WebCore::::traverseNext):
1614         (WebCore::::traversePrevious):
1615         (WebCore::::traverseNextSibling):
1616         (WebCore::::traversePreviousSibling):
1617         (WebCore::::operator):
1618         (WebCore::=):
1619         (WebCore::::ElementConstIterator):
1620         * dom/ElementIteratorAssertions.h: Copied from Source/WebCore/dom/DescendantIteratorAssertions.h.
1621         (WebCore::ElementIteratorAssertions::ElementIteratorAssertions):
1622         (WebCore::ElementIteratorAssertions::domTreeHasMutated):
1623         (WebCore::ElementIteratorAssertions::dropEventDispatchAssertion):
1624
1625 2013-08-30  Brent Fulgham  <bfulgham@apple.com>
1626
1627         [Windows] Build fix after r154921.
1628
1629         AVFoundationCF uses an internal wrapper object around the C-API. The call point
1630         in the original patch was inside the wrapper, which did not (yet) provide
1631         a method implementation.
1632
1633         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
1634         (WebCore::MediaPlayerPrivateAVFoundationCF::updateVideoLayerGravity):
1635         (WebCore::AVFWrapper::updateVideoLayerGravity):
1636
1637 2013-08-30  Simon Fraser  <simon.fraser@apple.com>
1638
1639         Video with object-fit: cover can spill outside the box
1640         https://bugs.webkit.org/show_bug.cgi?id=52103
1641
1642         Reviewed by Dean Jackson.
1643
1644         object-fit on renderers which use accelerated compositing needs special
1645         treatment.
1646
1647         For directly composited images, and video, GraphicsLayer needs to know
1648         both the size of the content layer, and also a rectangle at which this
1649         should be clipped (because, for the first time, that content layer can be
1650         larger than the renderer's content box).
1651
1652         AVFoundation would always aspect-ratio fit video by default, so plumb
1653         through MediaPlayer a way to override that when object-fit requires it.
1654
1655         Added a LAYER_TREE_INCLUDES_CONTENT_LAYERS enum to the layerTreeAsText()
1656         flags so we can dump content layers for testing.
1657
1658         Tests: compositing/images/direct-image-object-fit.html
1659                compositing/reflections/direct-image-object-fit-reflected.html
1660                compositing/video/video-object-fit.html
1661
1662         * page/Frame.h: New LayerTreeFlagsIncludeContentLayers flag.
1663         * platform/graphics/GraphicsLayer.h: New flag.
1664         * platform/graphics/MediaPlayer.cpp:
1665         (WebCore::MediaPlayer::shouldMaintainAspectRatio):
1666         (WebCore::MediaPlayer::setShouldMaintainAspectRatio):
1667         * platform/graphics/MediaPlayer.h:
1668         * platform/graphics/MediaPlayerPrivate.h:
1669         (WebCore::MediaPlayerPrivateInterface::shouldMaintainAspectRatio):
1670         (WebCore::MediaPlayerPrivateInterface::setShouldMaintainAspectRatio):
1671         * platform/graphics/avfoundation/MediaPlayerPrivateAVFoundation.cpp:
1672         (WebCore::MediaPlayerPrivateAVFoundation::MediaPlayerPrivateAVFoundation):
1673         (WebCore::MediaPlayerPrivateAVFoundation::setShouldMaintainAspectRatio):
1674         * platform/graphics/avfoundation/MediaPlayerPrivateAVFoundation.h:
1675         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
1676         (WebCore::MediaPlayerPrivateAVFoundationCF::updateVideoLayerGravity):
1677         (WebCore::AVFWrapper::platformLayer):
1678         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.h:
1679         * platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.h:
1680         * platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
1681         (WebCore::MediaPlayerPrivateAVFoundationObjC::createVideoLayer):
1682         (WebCore::MediaPlayerPrivateAVFoundationObjC::updateVideoLayerGravity):
1683         * platform/graphics/ca/GraphicsLayerCA.cpp: We need a new m_contentsClippingLayer to
1684         clip the contents layer, which only gets created when necessary. It has to be cloned
1685         for reflections.
1686         (WebCore::GraphicsLayerCA::willBeDestroyed):
1687         (WebCore::GraphicsLayerCA::setContentsRect):
1688         (WebCore::GraphicsLayerCA::setContentsClippingRect):
1689         (WebCore::GraphicsLayerCA::commitLayerChangesBeforeSublayers):
1690         (WebCore::GraphicsLayerCA::updateSublayerList):
1691         (WebCore::GraphicsLayerCA::updateContentsImage):
1692         (WebCore::GraphicsLayerCA::updateContentsMediaLayer):
1693         (WebCore::GraphicsLayerCA::updateContentsCanvasLayer):
1694         (WebCore::GraphicsLayerCA::updateContentsColorLayer):
1695         (WebCore::GraphicsLayerCA::updateContentsRects):
1696         (WebCore::GraphicsLayerCA::dumpAdditionalProperties):
1697         (WebCore::GraphicsLayerCA::ensureCloneLayers):
1698         (WebCore::GraphicsLayerCA::removeCloneLayers):
1699         (WebCore::GraphicsLayerCA::fetchCloneLayers):
1700         * platform/graphics/ca/GraphicsLayerCA.h:
1701         * rendering/RenderLayerBacking.cpp: Need to push both the contentsRect and
1702         the contentsClippingRect down to the GraphicsLayers. Most of the time they
1703         are the same, unless object-fit makes them different.
1704         (WebCore::RenderLayerBacking::resetContentsRect):
1705         (WebCore::RenderLayerBacking::positionOverflowControlsLayers):
1706         (WebCore::RenderLayerBacking::updateDirectlyCompositedBackgroundColor):
1707         (WebCore::RenderLayerBacking::updateDirectlyCompositedBackgroundImage):
1708         (WebCore::RenderLayerBacking::updateImageContents):
1709         (WebCore::RenderLayerBacking::contentsBox):
1710         * rendering/RenderLayerCompositor.cpp:
1711         (WebCore::RenderLayerCompositor::layerTreeAsText):
1712         * rendering/RenderVideo.cpp:
1713         (WebCore::RenderVideo::updatePlayer):
1714         * testing/Internals.cpp:
1715         (WebCore::Internals::layerTreeAsText):
1716         * testing/Internals.h:
1717         * testing/Internals.idl:
1718
1719 2013-08-30  Brent Fulgham  <bfulgham@apple.com>
1720
1721         [Windows] Update to incorporate additional suggestions
1722         https://bugs.webkit.org/show_bug.cgi?id=120448
1723
1724         Reviewed by Darin Adler
1725
1726         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp: Add
1727         notes as to why AVFoundationCF needs this extra method call.
1728         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.h: Use OVERRIDE
1729         * rendering/RenderVideo.cpp: Replace ternary operator with logical &&.
1730
1731 2013-08-30  Brent Fulgham  <bfulgham@apple.com>
1732
1733         [Windows] Vide element in page always uses non-hw accelerated mode.
1734         https://bugs.webkit.org/show_bug.cgi?id=120448
1735
1736         Reviewed by Darin Adler
1737
1738         This patch is unfortunately larger than my original idea, but seems to make the
1739         layout system happier. Instead of switching into composited mode when building
1740         the media player, we now build the original layout tree with compositing active
1741         if the underlying media element requires it. The AVFoundationCF player needs to
1742         have the compositor available at construction time so it can attach to the
1743         rendering device. Otherwise it falls back to CPU-only mode.
1744
1745         * platform/graphics/MediaPlayer.cpp:
1746         (WebCore::MediaPlayer::requiresImmediateCompositing): Added
1747         * platform/graphics/MediaPlayer.h:
1748         * platform/graphics/MediaPlayerPrivate.h:
1749         (WebCore::MediaPlayerPrivateInterface::requiresImmediateCompositing): Added
1750         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
1751         (WebCore::MediaPlayerPrivateAVFoundationCF::createAVPlayer): Added
1752         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.h:
1753         (WebCore::MediaPlayerPrivateAVFoundationCF::requiresImmediateCompositing):
1754         * rendering/RenderLayerCompositor.cpp:
1755         (WebCore::RenderLayerCompositor::requiresCompositingForVideo): Uses new
1756         'requiresImmediateCompositing' to short-circuit check for whether a
1757         compositor is required.
1758         * rendering/RenderVideo.cpp:
1759         (WebCore::RenderVideo::requiresImmediateCompositing):
1760         * rendering/RenderVideo.h:
1761
1762 2013-08-30  Joseph Pecoraro  <pecoraro@apple.com>
1763
1764         Web Inspector: Breakpoints should have Automatically Continue Option
1765         https://bugs.webkit.org/show_bug.cgi?id=120187
1766
1767         Reviewed by Timothy Hatcher.
1768
1769         Tests: inspector-protocol/debugger/removeBreakpoint.html
1770                inspector-protocol/debugger/setBreakpoint-autoContinue.html
1771                inspector-protocol/debugger/setBreakpoint-column.html
1772                inspector-protocol/debugger/setBreakpoint-condition.html
1773                inspector-protocol/debugger/setBreakpoint.html
1774
1775         * inspector/Inspector.json:
1776         Convert the "condition" argument of setBreakpoint and setBreakpointByUrl
1777         to a BreakpointOptions object with optional properties. One of which
1778         is "condition" and a new option "autoContinue".
1779
1780         * bindings/js/ScriptDebugServer.h:
1781         * bindings/js/ScriptDebugServer.cpp:
1782         (WebCore::ScriptDebugServer::hasBreakpoint):
1783         (WebCore::ScriptDebugServer::pauseIfNeeded):
1784         Automatically continue if the breakpoint was hit and has autoContinue.
1785
1786         * inspector/InspectorDebuggerAgent.h:
1787         * inspector/InspectorDebuggerAgent.cpp:
1788         (WebCore::buildObjectForBreakpointCookie):
1789         (WebCore::InspectorDebuggerAgent::setBreakpointByUrl):
1790         (WebCore::InspectorDebuggerAgent::setBreakpoint):
1791         (WebCore::InspectorDebuggerAgent::continueToLocation):
1792         (WebCore::InspectorDebuggerAgent::didParseSource):
1793         * inspector/ScriptBreakpoint.h:
1794         (WebCore::ScriptBreakpoint::ScriptBreakpoint):
1795         Handle the new incoming BreakpointOptions type, and set the
1796         autoContinue state on ScriptBreakpoints.
1797
1798         * inspector/front-end/DebuggerModel.js:
1799         (WebInspector.DebuggerModel.prototype.setBreakpointByURL):
1800         (WebInspector.DebuggerModel.prototype.setBreakpointBySourceId):
1801         Update old front-end to be compatible with the new API.
1802
1803 2013-08-30  Dean Jackson  <dino@apple.com>
1804
1805         Animations should use double for key values, not floats
1806         https://bugs.webkit.org/show_bug.cgi?id=120547
1807
1808         Reviewed by Simon Fraser.
1809
1810         Merge https://chromium.googlesource.com/chromium/blink/+/71de046541c77120874b9bff82958ee9e0e20c7c
1811
1812         Some files have been renamed in the Blink port, and they have made some
1813         improvements, but I took what applied to us.
1814
1815         All our existing tests passed.
1816
1817         * css/StyleResolver.cpp:
1818         (WebCore::StyleResolver::keyframeStylesForAnimation):
1819         * css/WebKitCSSKeyframeRule.cpp:
1820         (WebCore::StyleKeyframe::parseKeyString):
1821         * css/WebKitCSSKeyframeRule.h:
1822         (WebCore::StyleKeyframe::getKeys):
1823         * platform/graphics/GraphicsLayer.h:
1824         (WebCore::AnimationValue::keyTime):
1825         (WebCore::AnimationValue::AnimationValue):
1826         (WebCore::FloatAnimationValue::create):
1827         (WebCore::FloatAnimationValue::FloatAnimationValue):
1828         (WebCore::TransformAnimationValue::create):
1829         (WebCore::TransformAnimationValue::TransformAnimationValue):
1830         (WebCore::FilterAnimationValue::create):
1831         (WebCore::FilterAnimationValue::FilterAnimationValue):
1832         * rendering/RenderLayerBacking.cpp:
1833         (WebCore::RenderLayerBacking::startAnimation):
1834         * rendering/style/KeyframeList.h:
1835         (WebCore::KeyframeValue::KeyframeValue):
1836         (WebCore::KeyframeValue::key):
1837         (WebCore::KeyframeValue::setKey):
1838
1839 2013-08-30  Brendan Long  <b.long@cablelabs.com>
1840
1841         [GStreamer] support in-band text tracks
1842         https://bugs.webkit.org/show_bug.cgi?id=103771
1843
1844         Reviewed by Eric Carlson.
1845
1846         Tests: New tests added because existing tests were too specific.
1847            media/track/in-band/track-in-band-kate-ogg-cues-added-once.html
1848            media/track/in-band/track-in-band-kate-ogg-kind.html
1849            media/track/in-band/track-in-band-kate-ogg-language.html
1850            media/track/in-band/track-in-band-kate-ogg-mode.html
1851            media/track/in-band/track-in-band-kate-ogg-style.html
1852            media/track/in-band/track-in-band-kate-ogg-track-order.html
1853            media/track/in-band/track-in-band-srt-mkv-cues-added-once.html
1854            media/track/in-band/track-in-band-srt-mkv-kind.html
1855            media/track/in-band/track-in-band-srt-mkv-language.html
1856            media/track/in-band/track-in-band-srt-mkv-mode.html
1857            media/track/in-band/track-in-band-srt-mkv-style.html
1858            media/track/in-band/track-in-band-srt-mkv-track-order.html
1859
1860         * CMakeLists.txt: Add InbandTextTrackPrivateGStreamer, InbandGenericTextTrack, InbandWebVTTTextTrack, and TextCombinerGStreamer files.
1861         * GNUmakefile.list.am: Same.
1862         * PlatformEfl.cmake: Same.
1863         * Target.pri: Same.
1864         * WebCore.vcxproj/WebCore.vcxproj: Same.
1865         * WebCore.vcxproj/WebCore.vcxproj.filters: Same.
1866         * WebCore.xcodeproj/project.pbxproj: Same.
1867         * html/track/InbandGenericTextTrack.cpp: Split out code for handling generic cues.
1868         (WebCore::GenericTextTrackCueMap::GenericTextTrackCueMap): Move from InbandTextTrack.
1869         (WebCore::GenericTextTrackCueMap::~GenericTextTrackCueMap): Same.
1870         (WebCore::GenericTextTrackCueMap::add): Same.
1871         (WebCore::GenericTextTrackCueMap::find): Same.
1872         (WebCore::GenericTextTrackCueMap::remove): Same.
1873         (WebCore::InbandGenericTextTrack::create): Same.
1874         (WebCore::InbandGenericTextTrack::updateCueFromCueData): Same.
1875         (WebCore::InbandGenericTextTrack::addGenericCue): Same.
1876         (WebCore::InbandGenericTextTrack::updateGenericCue): Same.
1877         (WebCore::InbandGenericTextTrack::removeGenericCue): Same.
1878         (WebCore::InbandGenericTextTrack::removeCue): Same.
1879         (WebCore::InbandGenericTextTrack::InbandGenericTextTrack): Empty.
1880         (WebCore::InbandGenericTextTrack::~InbandGenericTextTrack): Empty.
1881         * html/track/InbandGenericTextTrack.h: Copied from Source/WebCore/html/track/InbandTextTrack.h.
1882         The only addition is the ASSERT_NOT_REACHED() for WebVTT cues.
1883         * html/track/InbandTextTrack.cpp: Add label and language changed callbacks.
1884         (WebCore::InbandTextTrack::create): Return a generic or WebVTT text track based on the private CueFormat.
1885         (WebCore::InbandTextTrack::labelChanged): Added.
1886         (WebCore::InbandTextTrack::languageChanged): Added.
1887         * html/track/InbandTextTrack.h: Add label and language changed callbacks.
1888         * html/track/InbandWebVTTTextTrack.cpp: Added, based on InbandTextTrack.
1889         (WebCore::InbandWebVTTTextTrack::create): Same.
1890         (WebCore::InbandWebVTTTextTrack::InbandWebVTTTextTrack): Empty.
1891         (WebCore::InbandWebVTTTextTrack::~InbandWebVTTTextTrack): Empty.
1892         (WebCore::InbandWebVTTTextTrack::parseWebVTTCueData): Sends data to a WebVTTParser.
1893         (WebCore::InbandWebVTTTextTrack::newCuesParsed): Adds cues when WebVTTParser parses them.
1894         (WebCore::InbandWebVTTTextTrack::fileFailedToParse): Prints a warning when WebVTTParser has a problem.
1895         * html/track/InbandWebVTTTextTrack.h: Added.
1896         * platform/graphics/InbandTextTrackPrivate.h:
1897         (WebCore::InbandTextTrackPrivate::cueFormat): For determining if the track will have generic or WebVTT cues.
1898         (WebCore::InbandTextTrackPrivate::InbandTextTrackPrivate): Pass CueFormat in the constructor.
1899         * platform/graphics/InbandTextTrackPrivateClient.h: Same.
1900         * platform/graphics/avfoundation/InbandTextTrackPrivateAVF.cpp:
1901         (WebCore::InbandTextTrackPrivateAVF::InbandTextTrackPrivateAVF): Pass CueFormat (Generic) to InbandTextTrackPrivate.
1902         * platform/graphics/gstreamer/GRefPtrGStreamer.cpp: Add GRefPtr specializations for GstSample and GstEvent.
1903         * platform/graphics/gstreamer/GRefPtrGStreamer.h: Same.
1904         * platform/graphics/gstreamer/GStreamerUtilities.h: Add WARN_MEDIA_MESSAGE.
1905         * platform/graphics/gstreamer/GStreamerVersioning.h: Add a function to check GStreamer version at runtime.
1906         * platform/graphics/gstreamer/InbandTextTrackPrivateGStreamer.cpp: Added.
1907         (WebCore::textTrackPrivateEventCallback): Watches for tag and stream start events.
1908         (WebCore::textTrackPrivateSampleTimeoutCallback): See notifyTrackOfSample();
1909         (WebCore::textTrackPrivateStreamTimeoutCallback): See notifyTrackOfStreamChanged();
1910         (WebCore::textTrackPrivateTagsChangeTimeoutCallback): See notifyTrackOfTagsChanged();
1911         (WebCore::InbandTextTrackPrivateGStreamer::InbandTextTrackPrivateGStreamer): Initializes tags and stream and sets up event callback.
1912         (WebCore::InbandTextTrackPrivateGStreamer::~InbandTextTrackPrivateGStreamer): Calls disconnect:
1913         (WebCore::InbandTextTrackPrivateGStreamer::disconnect): Removes signal handlers and frees some memory.
1914         (WebCore::InbandTextTrackPrivateGStreamer::handleSample): Adds samples to a list and sets up callback.
1915         (WebCore::InbandTextTrackPrivateGStreamer::streamChanged): Sets up callback.
1916         (WebCore::InbandTextTrackPrivateGStreamer::tagsChanged): Same.
1917         (WebCore::InbandTextTrackPrivateGStreamer::notifyTrackOfSample): Parses all queued samples with WebVTTParser.
1918         (WebCore::InbandTextTrackPrivateGStreamer::notifyTrackOfStreamChanged): Keeps track of current stream.
1919         (WebCore::InbandTextTrackPrivateGStreamer::notifyTrackOfTagsChanged): Sets label and language from tags.
1920         * platform/graphics/gstreamer/InbandTextTrackPrivateGStreamer.h: Added.
1921         (WebCore::InbandTextTrackPrivateGStreamer::create): Basic RefPtr create function.
1922         (WebCore::InbandTextTrackPrivateGStreamer::pad): Returns the pad this track is associated with (used
1923         to determine if a playbin text stream has already been associated with a text track).
1924         (WebCore::InbandTextTrackPrivateGStreamer::setIndex): Sets the track index (used for sorting).
1925         (WebCore::InbandTextTrackPrivateGStreamer::streamId): Returns the stream ID (used to handle new samples).
1926         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
1927         (WebCore::mediaPlayerPrivateTextChangedCallback): Called for playbin "text-changed" event. See textChanged().
1928         (WebCore::mediaPlayerPrivateTextChangeTimeoutCallback): See notifyPlayerOfText().
1929         (WebCore::mediaPlayerPrivateNewTextSampleCallback): See newTextSample().
1930         (WebCore::MediaPlayerPrivateGStreamer::MediaPlayerPrivateGStreamer): Initialize m_textTimerHandler.
1931         (WebCore::MediaPlayerPrivateGStreamer::~MediaPlayerPrivateGStreamer): Disconnect text tracks (they don't necessarily
1932         get freed here, since a script could hold a reference).
1933         (WebCore::MediaPlayerPrivateGStreamer::textChanged): Setup callback for notifyPlayerOfText.
1934         (WebCore::MediaPlayerPrivateGStreamer::notifyPlayerOfText): Create text tracks.
1935         (WebCore::MediaPlayerPrivateGStreamer::newTextSample): Handle new samples by giving them to a text track
1936         with a matching stream. This method is syncryonous because we need to get the stream start sticky event
1937         immediately.
1938         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.h:
1939         * platform/graphics/gstreamer/TextCombinerGStreamer.cpp: Added. This element forwards buffers from all
1940         of its input pads, but also converts plain text to WebVTT as needed.
1941         (webkit_text_combiner_init): Setup internal funnel.
1942         (webkitTextCombinerPadEvent): If the caps are plain text make sure we have a webvttenv, otherwise connect
1943         directly to the funnel.
1944         (webkitTextCombinerRequestNewPad): Setup ghostpad and event callback.
1945         (webkitTextCombinerReleasePad): Release pad and optional associated webvttenc.
1946         (webkit_text_combiner_class_init): Setup pad templates and request/release pad functions.
1947         (webkitTextCombinerNew): Returns a new WebKitTextCombiner.
1948         * platform/graphics/gstreamer/TextCombinerGStreamer.h: Added.
1949         * platform/graphics/gstreamer/TextSinkGStreamer.cpp: Added.
1950         (webkit_text_sink_init): Set sync=false.
1951         (webkitTextSinkGetProperty): Ignore sync property.
1952         (webkitTextSinkSetProperty): Same.
1953         (webkitTextSinkQuery): Ignore position and duration queries, forward everything else to appsink.
1954         (webkit_text_sink_class_init): Setup property and query functions.
1955         (webkitTextSinkNew): Return a new WebKitTextSink.
1956         * platform/graphics/gstreamer/TextSinkGStreamer.h: Added.
1957
1958 2013-08-30  Dirk Schulze  <krit@webkit.org>
1959
1960         Animate CSS Image filter() function
1961         https://bugs.webkit.org/show_bug.cgi?id=119938
1962
1963         Reviewed by Simon Fraser.
1964
1965         With this patch, the new introduced CSS Image function filter() can be
1966         animated. According to the spec, just filter functions can be
1967         interpolated.
1968
1969         The patch also prepares StyleImage blending for interpolation of other
1970         generated images like gradients or cross-fade().
1971
1972         http://dev.w3.org/fxtf/filters/#interpolating-filter-image
1973
1974         Test: fast/filter-image/filter-image-animation.html
1975
1976         * css/CSSComputedStyleDeclaration.cpp: Reuse the code that creates a
1977             CSSValueList from ComputeStyle logic.
1978         (WebCore::valueForPixel):
1979             For StyleRules we want to have not-adjusted length values.
1980         (WebCore::ComputedStyleExtractor::valueForShadow):
1981             Add argument to switch between adjusted and not-adjusted length.
1982         (WebCore::ComputedStyleExtractor::valueForFilter):
1983             Ditto.
1984         (WebCore::ComputedStyleExtractor::propertyValue):
1985         * css/CSSComputedStyleDeclaration.h:
1986         * css/CSSFilterImageValue.h: Add helper functions
1987             for animating filters. We need to pass the FilterOperations for
1988             the image generation and the CSSValueList for StyleRule.
1989         (WebCore::CSSFilterImageValue::filterOperations):
1990         (WebCore::CSSFilterImageValue::setFilterOperations):
1991         (WebCore::CSSFilterImageValue::cachedImage):
1992         * page/animation/CSSPropertyAnimation.cpp:
1993             Add animation code to support animations between two filter()
1994             function values.
1995         (WebCore::blendFilterOperations):
1996         (WebCore::blendFunc):
1997         (WebCore::filterBlend):
1998         * rendering/style/StyleGeneratedImage.h: Add helper functions.
1999         (WebCore::CSSFilterImageValue::imageValue):
2000
2001
2002 2013-08-30  Leo Yang  <leoyang@blackberry.com>
2003
2004         Make sure remove CachedResourceClient when destructing IconLoader
2005         https://bugs.webkit.org/show_bug.cgi?id=120541
2006
2007         Reviewed by Darin Adler.
2008
2009         It's a good practice to call CachedResource::removeClient(client)
2010         when the client is being destructed. We need to do this for InconLoader
2011         to prevent m_resource from keeping dangling client in case m_resource
2012         is referenced by someone else in the future.
2013
2014         Found by code inspection. Just a defensive enhancement no new tests.
2015
2016         * loader/icon/IconLoader.cpp:
2017         (WebCore::IconLoader::~IconLoader):
2018
2019 2013-08-30  Hans Muller  <hmuller@adobe.com>
2020
2021         [CSS Shapes] Redefine the ShapeIntervals class as a template
2022         https://bugs.webkit.org/show_bug.cgi?id=120381
2023
2024         Reviewed by Alexandru Chiculita.
2025
2026         No new tests, this was just an internal refactoring.
2027
2028         The existing ShapeIntervals class has been converted into a template whose
2029         type specifies the type of the interval's x1 and x2 horizontal endpoints
2030         (formerly float). There were several other minor changes, all in the realm
2031         of refactoring:
2032
2033         - The original type was a struct with public x1 and x2 fields. It's now a class
2034         with x1 and x2 accessors. ASSERTS are now used to maintain the x2 >= x1 invariant.
2035         In the original code the invariant was not checked.
2036
2037         - The logical comparison operators have been overloaded for ShapeInterval.
2038         This obviates the IntervalX1Comparator class which has been removed.
2039
2040         - The names of the global ShapeInterval Vector set operation methods have been
2041         changed to reflect the fact that they're now members of the template class,
2042         rather than globals.
2043
2044         PolygonShape.cpp depended on the ShapeInterval class. In the one or two places
2045         where an interval's x1 or x1 fields had been set explicitly, ShapeInterval::set()
2046         is now used to set both fields. This also enables the invariant check mentioned
2047         earlier. The other changes to this class are syntatic, to account for the ShapeInterval
2048         class's changes.
2049
2050         * CMakeLists.txt:
2051         * GNUmakefile.list.am:
2052         * WebCore.vcxproj/WebCore.vcxproj:
2053         * WebCore.xcodeproj/project.pbxproj:
2054         * rendering/shapes/PolygonShape.cpp:
2055         (WebCore::appendIntervalX):
2056         (WebCore::computeXIntersections):
2057         (WebCore::computeOverlappingEdgeXProjections):
2058         (WebCore::PolygonShape::getExcludedIntervals):
2059         (WebCore::PolygonShape::getIncludedIntervals):
2060         * rendering/shapes/PolygonShape.h:
2061         * rendering/shapes/ShapeInterval.cpp: Removed.
2062         * rendering/shapes/ShapeInterval.h: The entire implementation is now here.
2063         (WebCore::ShapeInterval::ShapeInterval):
2064         (WebCore::ShapeInterval::x1):
2065         (WebCore::ShapeInterval::x2):
2066         (WebCore::ShapeInterval::set):
2067         (WebCore::ShapeInterval::overlaps):
2068         (WebCore::ShapeInterval::intersect):
2069         (WebCore::ShapeInterval::sortVector):
2070         (WebCore::ShapeInterval::uniteVectors):
2071         (WebCore::ShapeInterval::intersectVectors):
2072         (WebCore::ShapeInterval::subtractVectors):
2073         (WebCore::operator==):
2074         (WebCore::operator!=):
2075         (WebCore::operator< ):
2076         (WebCore::operator> ):
2077         (WebCore::operator<=):
2078         (WebCore::operator>=):
2079
2080 2013-08-30  Antti Koivisto  <antti@apple.com>
2081
2082         Use Element& in StyleResolveTree
2083         https://bugs.webkit.org/show_bug.cgi?id=120540
2084
2085         Reviewed by Andreas Kling.
2086
2087         * dom/ContainerNode.cpp:
2088         (WebCore::attachChild):
2089         (WebCore::detachChild):
2090         * dom/Document.cpp:
2091         (WebCore::Document::recalcStyle):
2092         (WebCore::Document::createRenderTree):
2093         (WebCore::Document::detach):
2094         * dom/Element.cpp:
2095         (WebCore::Element::lazyReattach):
2096         (WebCore::Element::updateExistingPseudoElement):
2097         (WebCore::Element::createPseudoElementIfNeeded):
2098         (WebCore::disconnectPseudoElement):
2099         * dom/ShadowRoot.cpp:
2100         (WebCore::ShadowRoot::setResetStyleInheritance):
2101         * html/HTMLDetailsElement.cpp:
2102         (WebCore::HTMLDetailsElement::parseAttribute):
2103         * html/HTMLInputElement.cpp:
2104         (WebCore::HTMLInputElement::updateType):
2105         (WebCore::HTMLInputElement::parseAttribute):
2106         * html/HTMLObjectElement.cpp:
2107         (WebCore::HTMLObjectElement::renderFallbackContent):
2108         * html/HTMLPlugInImageElement.cpp:
2109         (WebCore::HTMLPlugInImageElement::willRecalcStyle):
2110         (WebCore::HTMLPlugInImageElement::documentWillSuspendForPageCache):
2111         (WebCore::HTMLPlugInImageElement::documentDidResumeFromPageCache):
2112         (WebCore::HTMLPlugInImageElement::restartSnapshottedPlugIn):
2113         * html/HTMLSelectElement.cpp:
2114         (WebCore::HTMLSelectElement::parseAttribute):
2115         (WebCore::HTMLSelectElement::parseMultipleAttribute):
2116         * html/HTMLViewSourceDocument.cpp:
2117         (WebCore::HTMLViewSourceDocument::createContainingTable):
2118         (WebCore::HTMLViewSourceDocument::addSpanWithClassName):
2119         (WebCore::HTMLViewSourceDocument::addLine):
2120         (WebCore::HTMLViewSourceDocument::finishLine):
2121         (WebCore::HTMLViewSourceDocument::addBase):
2122         (WebCore::HTMLViewSourceDocument::addLink):
2123         * html/parser/HTMLConstructionSite.cpp:
2124         (WebCore::executeTask):
2125         * html/parser/HTMLTreeBuilder.cpp:
2126         (WebCore::HTMLTreeBuilder::callTheAdoptionAgency):
2127         * html/shadow/InsertionPoint.cpp:
2128         (WebCore::InsertionPoint::willAttachRenderers):
2129         (WebCore::InsertionPoint::willDetachRenderers):
2130         * loader/PlaceholderDocument.cpp:
2131         (WebCore::PlaceholderDocument::createRenderTree):
2132         * style/StyleResolveTree.cpp:
2133         (WebCore::Style::attachChildren):
2134         (WebCore::Style::attachRenderTree):
2135         (WebCore::Style::detachChildren):
2136         (WebCore::Style::detachRenderTree):
2137         (WebCore::Style::resolveLocal):
2138         (WebCore::Style::updateTextStyle):
2139         (WebCore::Style::resolveShadowTree):
2140         (WebCore::Style::resolveTree):
2141
2142             Documents only ever have one child element. Remove the loop.
2143
2144         (WebCore::Style::detachRenderTreeInReattachMode):
2145         (WebCore::Style::reattachRenderTree):
2146         * style/StyleResolveTree.h:
2147         * svg/SVGTests.cpp:
2148         (WebCore::SVGTests::handleAttributeChange):
2149         * xml/XMLErrors.cpp:
2150         (WebCore::XMLErrors::insertErrorMessageBlock):
2151         * xml/parser/XMLDocumentParserLibxml2.cpp:
2152         (WebCore::XMLDocumentParser::startElementNs):
2153
2154 2013-08-30  Andreas Kling  <akling@apple.com>
2155
2156         Document style resolve should take Document&.
2157         <https://webkit.org/b/120534>
2158
2159         Reviewed by Darin Adler.
2160
2161         Tweak Style::resolveTree(Document*) and Style::resolveForDocument(Document*) to take Document& instead.
2162
2163         * dom/Document.cpp:
2164         (WebCore::Document::recalcStyle):
2165         * html/HTMLLinkElement.cpp:
2166         (WebCore::HTMLLinkElement::process):
2167         * style/StyleResolveForDocument.cpp:
2168         (WebCore::Style::resolveForDocument):
2169         * style/StyleResolveForDocument.h:
2170         * style/StyleResolveTree.cpp:
2171         (WebCore::Style::resolveTree):
2172         * style/StyleResolveTree.h:
2173
2174 2013-08-30  Darin Adler  <darin@apple.com>
2175
2176         [Mac] No need for Pasteboard::getDataSelection
2177         https://bugs.webkit.org/show_bug.cgi?id=120536
2178
2179         Reviewed by Anders Carlsson.
2180
2181         * editing/Editor.h: Added some Mac-only private member functions.
2182
2183         * editing/mac/EditorMac.mm:
2184         (WebCore::Editor::selectionInWebArchiveFormat): Added.
2185         (WebCore::Editor::adjustedSelectionRange): Added.
2186         (WebCore::attributedStringForRange): Added.
2187         (WebCore::dataInRTFDFormat): Added.
2188         (WebCore::dataInRTFFormat): Added.
2189         (WebCore::Editor::dataSelectionForPasteboard): Moved the implementation
2190         of Pasteboard::getDataSelection here, refactoring to share code with the
2191         writeSelectionToPasteboard function.
2192         (WebCore::Editor::writeSelectionToPasteboard): Refactored to share code
2193         with the new function above.
2194
2195         * platform/Pasteboard.h: Removed getDataSelection. One less layering violation.
2196         * platform/mac/PasteboardMac.mm: Ditto.
2197
2198 2013-08-30  Rob Buis  <rwlbuis@webkit.org>
2199
2200         SVG error parsing empty path
2201         https://bugs.webkit.org/show_bug.cgi?id=78980
2202
2203         Reviewed by Darin Adler.
2204
2205         According to the spec (http://www.w3.org/TR/SVG/paths.html#PathData), path 'd' attribute can be empty.
2206         No error should be reported in this case.
2207
2208         * svg/SVGPathParser.cpp:
2209         (WebCore::SVGPathParser::parsePathDataFromSource):
2210         * svg/SVGPathUtilities.cpp:
2211         (WebCore::buildPathFromString):
2212         (WebCore::buildSVGPathByteStreamFromSVGPathSegList):
2213         (WebCore::buildPathFromByteStream):
2214         (WebCore::buildSVGPathSegListFromByteStream):
2215         (WebCore::buildStringFromByteStream):
2216         (WebCore::buildStringFromSVGPathSegList):
2217         (WebCore::buildSVGPathByteStreamFromString):
2218         (WebCore::buildAnimatedSVGPathByteStream):
2219         (WebCore::addToSVGPathByteStream):
2220
2221 2013-08-30  Andreas Kling  <akling@apple.com>
2222
2223         Try to fix CSS_VARIABLES and CSS_DEVICE_ADAPTATION builds.
2224
2225         * css/StyleResolver.cpp:
2226         (WebCore::StyleResolver::StyleResolver):
2227         (WebCore::StyleResolver::resolveVariables):
2228
2229 2013-08-30  Commit Queue  <commit-queue@webkit.org>
2230
2231         Unreviewed, rolling out r154891.
2232         http://trac.webkit.org/changeset/154891
2233         https://bugs.webkit.org/show_bug.cgi?id=120539
2234
2235         broke the mac build (Requested by Ossy on #webkit).
2236
2237         * platform/network/SynchronousLoaderClient.h:
2238
2239 2013-08-30  Csaba Osztrogonác  <ossy@webkit.org>
2240
2241         Add USE(CFNETWORK) guard to SynchronousLoaderClient::didReceiveAuthenticationChallengedidReceiveAuthenticationChallenge()
2242         https://bugs.webkit.org/show_bug.cgi?id=120532
2243
2244         Reviewed by Darin Adler.
2245
2246         * platform/network/SynchronousLoaderClient.h:
2247
2248 2013-08-30  Brent Fulgham  <bfulgham@apple.com>
2249
2250         [Windows] Video inside page always uses non-hardware accelerated playback
2251         https://bugs.webkit.org/show_bug.cgi?id=120448
2252
2253         Reviewed by Eric Carlson.
2254
2255         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
2256         (WebCore::MediaPlayerPrivateAVFoundationCF::createAVPlayer): Check for
2257         D3D device. If it doesn't exist, switch to compositing mode and try again.
2258
2259 2013-08-30  Leo Yang  <leoyang@blackberry.com>
2260
2261         Possible dangling CachedResourceClient of StyleRuleImport and XSLImportRule
2262         https://bugs.webkit.org/show_bug.cgi?id=120479
2263
2264         Reviewed by Darin Adler.
2265
2266         In StyleRuleImport::requestStyleSheet() and XSLImportRule::loadSheet() we
2267         didn't call removeClient() for m_cachedSheet before assigning m_cachedSheet
2268         a new value. This could leave the client as a client of the old cached
2269         sheet and dangling after the client is deleted. Fix them by calling removeClient()
2270         before assigning m_cacheSheet a new value.
2271
2272         Found by code inspection. Seems no way to test it automatically.
2273
2274         * css/StyleRuleImport.cpp:
2275         (WebCore::StyleRuleImport::requestStyleSheet):
2276         * xml/XSLImportRule.cpp:
2277         (WebCore::XSLImportRule::loadSheet):
2278
2279 2013-08-30  Gabor Abraham  <abrhm@inf.u-szeged.hu>
2280
2281         Fix unused parameter warning in graphics/cairo/GraphicsContextCairo.cpp file.
2282         https://bugs.webkit.org/show_bug.cgi?id=120524
2283
2284         Reviewed by Darin Adler.
2285
2286         * platform/graphics/cairo/GraphicsContextCairo.cpp:
2287         (WebCore::GraphicsContext::fillRectWithRoundedHole):
2288
2289 2013-08-30  Andreas Kling  <akling@apple.com>
2290
2291         StyleResolver & friends should use Document&.
2292         <https://webkit.org/b/120527>
2293
2294         Reviewed by Antti Koivisto.
2295
2296         Document&-ify StyleResolver, ElementRuleCollector and SelectorChecker.
2297
2298 2013-08-30  Brent Fulgham  <bfulgham@apple.com>
2299
2300         [Windows] Unreviewed build correction after r154835. Only seen when doing a
2301         full (clean) rebuild.
2302
2303         * DerivedSources.cpp: Remove reference to deleted JSHTMLDialogElement.cpp
2304
2305 2013-08-30  Tamas Czene  <tczene@inf.u-szeged.hu>
2306
2307         Resolve unused parameter warning in FindController.cpp
2308         https://bugs.webkit.org/show_bug.cgi?id=120407
2309
2310         Reviewed by Darin Adler.
2311
2312         * accessibility/atk/WebKitAccessibleInterfaceText.cpp:
2313         (wordAtPositionForAtkBoundary):
2314
2315 2013-08-30  Arpita Bahuguna  <a.bah@samsung.com>
2316
2317         setAttributeNode() does not set the new value to an existing attribute if specified attribute is in a different case.
2318         https://bugs.webkit.org/show_bug.cgi?id=120293
2319
2320         Reviewed by Darin Adler.
2321
2322         setAttributeNode() performs a case-insensitive search for an existing
2323         attribute. If an existing attribute is found, it retrieves the index of
2324         such an attribute. For setting the attribute's new value, we call upon
2325         setAttributeInternal() to which both the index as well as the name of
2326         the attribute is passed.
2327         The name passed to this method is the same as the one passed to the
2328         setAttributeNode() API from the webpage and thus can be in any case.
2329
2330         However, setAttributeInternal() uses this name to get the corresponding
2331         existing attribute node. Since this retrieval is not case-insensitive,
2332         the existing node is not returned and thus the new value is not set on
2333         the existing node.
2334         We should instead use the passed index and use that to retrieve the
2335         existing node.
2336
2337         Note that obtaining the attribute's value using getAttributeNode() would
2338         still return the correct value, i.e. the new one.
2339
2340         Also, this change shall make our behavior similar to that of FF and IE.
2341
2342         Test: fast/dom/Element/setAttributeNode-for-existing-attribute.html
2343
2344         * dom/Element.cpp:
2345         (WebCore::Element::setAttributeInternal):
2346         If the passed index is not equal to attributeNotFound, we use that index
2347         to retrieve the existing attribute.
2348
2349 2013-08-30  Mikhail Pozdnyakov  <mikhail.pozdnyakov@intel.com>
2350
2351         SimpleClassVectorTraits shall be used for RuleData
2352         https://bugs.webkit.org/show_bug.cgi?id=120526
2353
2354         Reviewed by Andreas Kling.
2355
2356         Whereas RuleData is POD type and whereas there are quite a lot of Vector<RuleData> occurrences in the code,
2357         it makes sense to use SimpleClassVectorTraits for RuleData and hence to allow using more efficient mem functions
2358         in vectors.
2359
2360         * css/RuleSet.h:
2361
2362 2013-08-30  Andreas Kling  <akling@apple.com>
2363
2364         Node::document() should return a reference.
2365         <https://webkit.org/b/120496>
2366
2367         Reviewed by Antti Koivisto.
2368
2369         Now that orphan DocumentType nodes also have document pointers, it's no longer
2370         possible to have a null Node::document().
2371
2372         Cement this by making document() return a reference, and remove the various
2373         null checks exposed by this.
2374
2375 2013-08-30  Andrei Parvu  <parvu@adobe.com>
2376
2377         [CSS Masking] -webkit-mask-repeat: space does not work
2378         Added the space option to background-repeat and -webkit-mask-repeat.
2379         With the property value 'space', the background or mask image gets repeated as often as it fits within the background positioning
2380         area. The repeated images are spaced equally to fill the unused area.
2381         https://bugs.webkit.org/show_bug.cgi?id=119324
2382
2383         Reviewed by Dirk Schulze.
2384
2385         Tests: css3/background/background-repeat-space-border.html
2386                css3/background/background-repeat-space-content.html
2387                css3/background/background-repeat-space-padding.html
2388                css3/masking/mask-repeat-space-border.html
2389                css3/masking/mask-repeat-space-content.html
2390                css3/masking/mask-repeat-space-padding.html
2391
2392         * platform/graphics/GeneratorGeneratedImage.cpp:
2393         (WebCore::GeneratorGeneratedImage::drawPattern): Passed the space values to the image buffer.
2394         * platform/graphics/Image.cpp:
2395         (WebCore::Image::drawTiled): Added the space values when computing the location of the tile.
2396         * platform/graphics/Image.h: Added the space property.
2397         (WebCore::Image::spaceSize):
2398         (WebCore::Image::setSpaceSize):
2399         * platform/graphics/ImageBuffer.h: Added the space property.
2400         (WebCore::ImageBuffer::spaceSize):
2401         (WebCore::ImageBuffer::setSpaceSize):
2402         * platform/graphics/cg/ImageBufferCG.cpp: Passed the space values when copying an image.
2403         (WebCore::ImageBuffer::copyImage):
2404         * platform/graphics/cg/ImageCG.cpp: Added the space values when creating a platform pattern.
2405         (WebCore::Image::drawPattern):
2406         * rendering/RenderBoxModelObject.cpp:
2407         (WebCore::RenderBoxModelObject::paintFillLayerExtended): Computed the space values on x and y axis.
2408         (WebCore::getSpace):
2409         (WebCore::RenderBoxModelObject::calculateBackgroundImageGeometry): Pass the space values to the Image class.
2410         * rendering/RenderBoxModelObject.h: Added the space property.
2411         (WebCore::RenderBoxModelObject::BackgroundImageGeometry::spaceSize):
2412         (WebCore::RenderBoxModelObject::BackgroundImageGeometry::setSpaceSize):
2413         * svg/graphics/SVGImage.cpp: Passed the space property to the created image.
2414         (WebCore::SVGImage::drawPatternForContainer):
2415         * svg/graphics/SVGImageForContainer.cpp: Passed the space property to the image property.
2416         (WebCore::SVGImageForContainer::drawPattern):
2417
2418 2013-08-30  Antti Koivisto  <antti@apple.com>
2419
2420         Remove AttachContext
2421         https://bugs.webkit.org/show_bug.cgi?id=120518
2422
2423         Reviewed by Andreas Kling.
2424
2425         This type is not useful anymore. Just pass the precomputed style to attachRenderTree and reattach-or-not flag to detachRenderTree.
2426
2427         * style/StyleResolveTree.cpp:
2428         (WebCore::Style::createRendererIfNeeded):
2429         (WebCore::Style::attachChildren):
2430         (WebCore::Style::attachShadowRoot):
2431         (WebCore::Style::attachRenderTree):
2432         (WebCore::Style::detachChildren):
2433         (WebCore::Style::detachShadowRoot):
2434         (WebCore::Style::detachRenderTree):
2435         (WebCore::Style::reattachRenderTree):
2436         (WebCore::Style::resolveLocal):
2437         * style/StyleResolveTree.h:
2438
2439 2013-08-30  Commit Queue  <commit-queue@webkit.org>
2440
2441         Unreviewed, rolling out r154826.
2442         http://trac.webkit.org/changeset/154826
2443         https://bugs.webkit.org/show_bug.cgi?id=120517
2444
2445         Still breaks icloud.com (Requested by mwenge_ on #webkit).
2446
2447         * rendering/RenderBox.cpp:
2448         (WebCore::RenderBox::computePositionedLogicalWidth):
2449         (WebCore::RenderBox::computePositionedLogicalHeight):
2450         * rendering/RenderBox.h:
2451         (WebCore::RenderBox::intrinsicSize):
2452         * rendering/RenderButton.h:
2453         * rendering/RenderFileUploadControl.cpp:
2454         (WebCore::RenderFileUploadControl::computePreferredLogicalWidths):
2455         * rendering/RenderListBox.cpp:
2456         (WebCore::RenderListBox::RenderListBox):
2457         (WebCore::RenderListBox::computePreferredLogicalWidths):
2458         (WebCore::RenderListBox::computeLogicalHeight):
2459         * rendering/RenderListBox.h:
2460         * rendering/RenderMenuList.h:
2461         * rendering/RenderSlider.cpp:
2462         (WebCore::RenderSlider::computePreferredLogicalWidths):
2463         * rendering/RenderTextControl.cpp:
2464         (WebCore::RenderTextControl::RenderTextControl):
2465         (WebCore::RenderTextControl::computeLogicalHeight):
2466         * rendering/RenderTextControl.h:
2467
2468 2013-08-29  Ryosuke Niwa  <rniwa@webkit.org>
2469
2470         Unduplicate the code to convert between VisiblePosition and index
2471         https://bugs.webkit.org/show_bug.cgi?id=120506
2472
2473         Reviewed by Darin Adler.
2474
2475         Encapsulate the conversion between VisiblePosition and index into indexForVisiblePosition
2476         and visiblePositionForIndexUsingCharacterIterator. It's unfortunate that these two functions
2477         are different from the two other existing functions of similar names but we've at least
2478         confined problems into a single cpp file now.
2479
2480         * accessibility/AccessibilityRenderObject.cpp:
2481         (WebCore::AccessibilityRenderObject::visiblePositionForIndex):
2482         (WebCore::AccessibilityRenderObject::indexForVisiblePosition):
2483         * editing/Editor.cpp:
2484         (WebCore::findFirstMarkable):
2485         * editing/htmlediting.cpp:
2486         (WebCore::indexForVisiblePosition):
2487         (WebCore::visiblePositionForIndexUsingCharacterIterator):
2488         * editing/htmlediting.h:
2489         * html/HTMLTextFormControlElement.cpp:
2490         (WebCore::hasVisibleTextArea):
2491         (WebCore::HTMLTextFormControlElement::setSelectionRange):
2492         (WebCore::HTMLTextFormControlElement::indexForVisiblePosition):
2493         (WebCore::HTMLTextFormControlElement::visiblePositionForIndex):
2494         * html/HTMLTextFormControlElement.h:
2495         * rendering/RenderTextControl.cpp:
2496         * rendering/RenderTextControl.h:
2497
2498 2013-08-29  Ryosuke Niwa  <rniwa@webkit.org>
2499
2500         Avoid Node references from AXObjectCache from leaking
2501         https://bugs.webkit.org/show_bug.cgi?id=120501
2502
2503         Reviewed by Darin Adler.
2504
2505         Merge https://chromium.googlesource.com/chromium/blink/+/454f31497613b6d0fbcfb0df757254b64a177c06
2506         without any tests since we don't have the same infrastructure to detect leaks in WebKit.
2507
2508         A real world example of this would be selecting an <option> item inside frame by keyboard. The node will not be deref()-ed until the topDocument() is detached.
2509
2510         The issue was that AccessibilityMenuListOption is created in childrenChanged()
2511         hook called when its RenderObject is being destroyed. This patch modifies AccessibilityMenuListPopup so it won't create AccessibilityMenuListOption if its
2512         element is already detached.
2513
2514         * accessibility/AccessibilityMenuListPopup.cpp:
2515         (WebCore::AccessibilityMenuListPopup::menuListOptionAccessibilityObject):
2516         * rendering/RenderMenuList.cpp:
2517         (WebCore::RenderMenuList::didUpdateActiveOption):
2518
2519 2013-08-26  Simon Fraser  <simon.fraser@apple.com>
2520
2521         Implement object-fit CSS property
2522         https://bugs.webkit.org/show_bug.cgi?id=52040
2523
2524         Reviewed by Antti Koivisto, Sam Weinig.
2525
2526         Merge object-fit patch from Blink r156535, which started as a patch
2527         by me.
2528
2529         Since then, the spec has gone to CR. This patch is an
2530         implementation of object-fit as described in
2531         http://www.w3.org/TR/2012/CR-css3-images-20120417/#object-fit
2532
2533         Object-fit is used to maintain the aspect ratio of replaced content
2534         within its content box. All object-fit values but the initial one
2535         ('fill') will always ensure that the aspect ratio is retained, in
2536         different ways (fit inside the content box, cover the content box, or
2537         use intrinsic size). Painting is always clipped against the content
2538         box, regardless of the 'overflow' property.
2539
2540         Tests: fast/css/object-fit/object-fit-canvas.html
2541                fast/css/object-fit/object-fit-embed.html
2542                fast/css/object-fit/object-fit-grow-landscape.html
2543                fast/css/object-fit/object-fit-grow-portrait.html
2544                fast/css/object-fit/object-fit-img-svg.html
2545                fast/css/object-fit/object-fit-img-svg2.html
2546                fast/css/object-fit/object-fit-img.html
2547                fast/css/object-fit/object-fit-input-image.html
2548                fast/css/object-fit/object-fit-object.html
2549                fast/css/object-fit/object-fit-shrink.html
2550                fast/css/object-fit/object-fit-video-poster.html
2551                fast/css/parsing-object-fit.html
2552                http/tests/css/object-fit-delayed-img-svg.html
2553                media/video-object-fit-change.html
2554                media/video-object-fit.html
2555
2556         * css/CSSComputedStyleDeclaration.cpp:
2557         (WebCore::ComputedStyleExtractor::propertyValue):
2558         * css/CSSParser.cpp:
2559         (WebCore::isValidKeywordPropertyAndValue):
2560         (WebCore::isKeywordPropertyID):
2561         (WebCore::CSSParser::parseValue):
2562         * css/CSSPrimitiveValueMappings.h:
2563         (WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
2564         (WebCore::CSSPrimitiveValue::operator EObjectFit):
2565         * css/CSSProperty.cpp:
2566         (WebCore::CSSProperty::isInheritedProperty):
2567         * css/CSSPropertyNames.in:
2568         * css/CSSValueKeywords.in:
2569         * css/DeprecatedStyleBuilder.cpp:
2570         (WebCore::DeprecatedStyleBuilder::DeprecatedStyleBuilder):
2571         * css/StyleResolver.cpp:
2572         (WebCore::StyleResolver::applyProperty):
2573         * css/html.css:
2574         (video): Set object-fit to 'contain'. This is how VIDEO elements
2575         work, apparently.
2576         * loader/cache/CachedImage.cpp:
2577         (WebCore::CachedImage::imageSizeForRenderer):
2578         * loader/cache/CachedImage.h:
2579         * platform/graphics/LayoutSize.h:
2580         (WebCore::fitLayoutSizeToAspectRatio): New function to grow or shrink
2581         in one dimension to fit to the aspect ratio.
2582         * rendering/RenderHTMLCanvas.cpp:
2583         (WebCore::RenderHTMLCanvas::paintReplaced): Apply object-fit and
2584         clip if necessary.
2585         * rendering/RenderImage.cpp:
2586         (WebCore::RenderImage::updateInnerContentRect):
2587         (WebCore::RenderImage::imageDimensionsChanged): Update intrinsic
2588         size properly, and recalculate the inner content rectangle (the
2589         exact area occupied by the replaced content) again if appropriate.
2590         (WebCore::RenderImage::paintReplaced): Apply object-fit and clip
2591         if necessary.
2592         (WebCore::RenderImage::foregroundIsKnownToBeOpaqueInRect):
2593         object-fit may leave parts of the content box empty, in which case
2594         it won't be fully obscured.
2595         (WebCore::RenderImage::layout):
2596         * rendering/RenderImage.h:
2597         * rendering/RenderImageResource.cpp:
2598         (WebCore::RenderImageResource::intrinsicSize): Need this to
2599         differentiate between intrinsic and extrinsic size for SVG images.
2600         * rendering/RenderImageResource.h:
2601         * rendering/RenderImageResourceStyleImage.h:
2602         * rendering/RenderReplaced.cpp:
2603         (WebCore::RenderReplaced::replacedContentRect): Return the
2604         rectangle occupied by the replaced content. This will be identical
2605         to the content box if object-fit is 'fill', but will typically be
2606         something else for other values.
2607         * rendering/RenderReplaced.h:
2608         * rendering/RenderVideo.cpp:
2609         (WebCore::RenderVideo::videoBox): Not much left to do here, with
2610         the new RenderReplaced::replacedContentRect() method in place.
2611         (WebCore::RenderVideo::paintReplaced): Apply object-fit and clip
2612         if necessary.
2613         * rendering/style/RenderStyle.cpp:
2614         (WebCore::RenderStyle::changeRequiresRepaint):
2615         * rendering/style/RenderStyle.h:
2616         * rendering/style/RenderStyleConstants.h:
2617         * rendering/style/StyleRareNonInheritedData.cpp:
2618         (WebCore::StyleRareNonInheritedData::StyleRareNonInheritedData):
2619         (WebCore::StyleRareNonInheritedData::operator==):
2620         * rendering/style/StyleRareNonInheritedData.h:
2621
2622 2013-08-29  Tim Horton  <timothy_horton@apple.com>
2623
2624         SVG clipping, masking, and gradients-on-text do not respect the device scale factor
2625         https://bugs.webkit.org/show_bug.cgi?id=120377
2626         <rdar://problem/14777944>
2627
2628         Reviewed by Simon Fraser.
2629
2630         Tests: svg/custom/masking-clipping-hidpi.svg
2631
2632         calculateTransformationToOutermostCoordinateSystem should include the page's
2633         device scale factor in its transform.
2634
2635         * rendering/svg/SVGRenderingContext.cpp:
2636         (WebCore::SVGRenderingContext::calculateTransformationToOutermostCoordinateSystem):
2637
2638 2013-08-29  Chris Fleizach  <cfleizach@apple.com>
2639
2640         AX: Crash when VoiceOver reference a TextMarker from an iframe that has been removed
2641         https://bugs.webkit.org/show_bug.cgi?id=120318
2642
2643         Reviewed by Darin Adler.
2644
2645         These crashes occur because an AXTextMarker is retrieved that reference a Node in an iframe.
2646         The iframe goes away, but when it tries to clean up and remove that Node from the InUse cache,
2647         the document() tree has already been detached, so it never clears the actual InUse cache.
2648
2649         The fix here is to pre-emptively clean up any nodes in the document going away when the frame is about to disconnect.
2650
2651         I'm removing the clearAXObjectCache() at the disconnectOwnerElement because it
2652            1) Cleared the AXObjectCache for the iframe document (which is always empty -- only the top level doc maintains the cache), because
2653                 the document() tree had already been detached... so it essentially did nothing.
2654            2) And if it did work, we wouldn't want this behavior -- that is to say, when an iframe goes away, we don't want to clear the entire cache for all
2655                the other documents (there's even an existing layout test to verify this behavior).
2656
2657         Test: platform/mac/accessibility/stale-textmarker-crash.html
2658
2659         * accessibility/AXObjectCache.cpp:
2660         (WebCore::AXObjectCache::clearTextMarkerNodesInUse):
2661         * accessibility/AXObjectCache.h:
2662         * page/Frame.cpp:
2663         (WebCore::Frame::disconnectOwnerElement):
2664
2665 2013-08-29  Brent Fulgham  <bfulgham@apple.com>
2666
2667         [Windows] More unreviewed gardening of project file.
2668         Move css files inline with related source code.
2669
2670         * WebCore.vcxproj/WebCore.vcxproj:
2671         * WebCore.vcxproj/WebCore.vcxproj.filters:
2672
2673 2013-08-29  Sam Weinig  <sam@webkit.org>
2674
2675         Add ENABLE guards for Promises
2676         https://bugs.webkit.org/show_bug.cgi?id=120488
2677
2678         Reviewed by Andreas Kling.
2679
2680         * Configurations/FeatureDefines.xcconfig:
2681
2682 2013-08-29  Andreas Kling  <akling@apple.com>
2683
2684         Inspector: Use a bit of RefPtr::release() in generated code.
2685         <https://webkit.org/b/120485>
2686
2687         Reviewed by Darin Adler.
2688
2689         Just something I spotted while reading InspectorBackendDispatcher.cpp.
2690         Avoid 3 cases of reference count churnery.
2691
2692         * inspector/CodeGeneratorInspectorStrings.py:
2693
2694 2013-08-29  Christophe Dumez  <ch.dumez@sisa.samsung.com>
2695
2696         According to DOM4, all DocType nodes should have a document
2697         https://bugs.webkit.org/show_bug.cgi?id=99244
2698
2699         Reviewed by Darin Adler.
2700
2701         Doctypes now always have a node document and can be moved across document boundaries as per
2702         the latest DOM4 specification:
2703         http://dom.spec.whatwg.org/#dom-domimplementation-createdocumenttype
2704         http://dom.spec.whatwg.org/#dom-node-ownerdocument
2705
2706         This means that DOMImplementation.createDocumentType() now sets the ownerDocument of the
2707         new DocumentType Node to the associated document of the current "context" object. In
2708         DOM4, all nodes have a document at all times. DocumentType nodes can now be moved across
2709         document boundaries so that the node can be added to a Document after being created.
2710
2711         This means we will no longer need to special case DocumentType nodes in the code and
2712         Node::document() can no longer return NULL, which means that we'll be able to remove
2713         NULL checks in call sites.
2714
2715         Firefox stable and since recently Blink already follow DOM4 here while IE10 does not (yet).
2716
2717         Test: fast/dom/createDocumentType-ownerDocument.html
2718
2719         * dom/ContainerNode.cpp:
2720         (WebCore::checkAcceptChild):
2721         * dom/DOMImplementation.cpp:
2722         (WebCore::DOMImplementation::createDocumentType):
2723         (WebCore::DOMImplementation::createDocument):
2724         * dom/Node.h:
2725         (WebCore::Node::document):
2726
2727 2013-08-29  Darin Adler  <darin@apple.com>
2728
2729         Pasteboard::writeSelection violates layering (first step, fixes it for Mac platform only)
2730         https://bugs.webkit.org/show_bug.cgi?id=120483
2731
2732         Reviewed by Anders Carlsson.
2733
2734         * editing/Editor.cpp:
2735         (WebCore::Editor::cut): Added some comments. Use the new writeSelectionToPasteboard
2736         function on Mac instead of Pasteboard::writeSelection.
2737         (WebCore::Editor::copy): Ditto.
2738
2739         * editing/Editor.h: Removed an old unused Mac-only writeSelectionToPasteboard function
2740         that was a cover that called through to Pasteboard::writeSelectionForTypes. Added a new
2741         Mac-only writeSelectionToPasteboard function that is destined to become cross-platform soon.
2742
2743         * editing/mac/EditorMac.mm:
2744         (WebCore::Editor::writeSelectionToPasteboard): Added. Uses a new pattern where the Editor
2745         puts all the data into a structure called PasteboardWebContent then calls the Pasteboard to
2746         do the work. The platform-specific aspect of PasteboardWebContent is what formats are needed
2747         for each platform.
2748
2749         * page/DragController.cpp:
2750         (WebCore::DragController::startDrag): Added some comments. Use the new writeSelectionToPasteboard
2751         function on Mac instead of Pasteboard::writeSelection.
2752
2753         * platform/Pasteboard.h: Remove some unneeded forward declarations. Added comments for all functions
2754         that don't belong in this class because they are layering violations; this becomes the to do list for
2755         the project we are beginning here. Added the new PasteboardWebContent structure, empty on all platforms
2756         except for Mac for now. Removed writeSelectionForTypes, a Mac-only function that is no longer used.
2757         Added setTypes and writeAfterSettingTypes, the two halves of the future function named writeWebContent.
2758         Put the writeSelection function inside a "not Mac" if statement. Later to be deleted entirely.
2759
2760         * platform/mac/PasteboardMac.mm: Removed now-unneeded selectionPasteboardTypes,
2761         Pasteboard::writeSelectionForTypes, and writeSelection functions.
2762         (WebCore::Pasteboard::setTypes): Added. First half of writing web content to the pasteboard.
2763         (WebCore::Pasteboard::writeAfterSettingTypes): Added. Second half of writing web content to the pasteboard.
2764
2765 2013-08-29  Antti Koivisto  <antti@apple.com>
2766
2767         Remove code behind ENABLE(DIALOG_ELEMENT)
2768         https://bugs.webkit.org/show_bug.cgi?id=120467
2769
2770         Reviewed by Darin Adler.
2771
2772         It is incomplete and no one is building it.
2773
2774         * CMakeLists.txt:
2775         * Configurations/FeatureDefines.xcconfig:
2776         * DerivedSources.make:
2777         * DerivedSources.pri:
2778         * GNUmakefile.list.am:
2779         * Target.pri:
2780         * WebCore.exp.in:
2781         * WebCore.xcodeproj/project.pbxproj:
2782         * bindings/generic/RuntimeEnabledFeatures.cpp:
2783         * bindings/generic/RuntimeEnabledFeatures.h:
2784         * css/StyleResolver.cpp:
2785         (WebCore::StyleResolver::adjustRenderStyle):
2786         * css/html.css:
2787         * dom/ContextFeatures.cpp:
2788         * dom/ContextFeatures.h:
2789         * dom/Document.cpp:
2790         * dom/Document.h:
2791         * dom/Element.cpp:
2792         (WebCore::Element::removedFrom):
2793         (WebCore::Element::setContainsFullScreenElementOnAncestorsCrossingFrameBoundaries):
2794         * dom/Element.h:
2795         (WebCore::Element::isDisabledFormControl):
2796         * dom/ElementRareData.h:
2797         (WebCore::ElementRareData::ElementRareData):
2798         * html/HTMLDialogElement.cpp: Removed.
2799         * html/HTMLDialogElement.h: Removed.
2800         * html/HTMLDialogElement.idl: Removed.
2801         * html/HTMLElementsAllInOne.cpp:
2802         * html/HTMLTagNames.in:
2803         * rendering/RenderDialog.cpp: Removed.
2804         * rendering/RenderDialog.h: Removed.
2805         * rendering/RenderLayer.cpp:
2806         (WebCore::RenderLayer::rebuildZOrderLists):
2807         (WebCore::RenderLayer::collectLayers):
2808         * rendering/RenderLayer.h:
2809         * rendering/RenderObject.h:
2810         (WebCore::RenderObject::isCounter):
2811         * rendering/RenderingAllInOne.cpp:
2812         * testing/InternalSettings.cpp:
2813         (WebCore::InternalSettings::Backup::Backup):
2814         (WebCore::InternalSettings::Backup::restoreTo):
2815         * testing/InternalSettings.h:
2816
2817 2013-08-29  Simon Fraser  <simon.fraser@apple.com>
2818
2819         Fix slider thumb event handling to use local, not absolute coordinates
2820         https://bugs.webkit.org/show_bug.cgi?id=120480
2821
2822         Reviewed by Darin Adler.
2823
2824         SliderThumbElement::setPositionFromPoint() did all of its coordinate
2825         math by mapping renderer rects into absolute coordinates, which was
2826         unnecessary and expensive.
2827
2828         Fix by doing all the math in the coordinate space of the input's
2829         renderer. This simplified the code. Also, currentPosition
2830         was computed but unused, so was removed.
2831
2832         No behavior change. Tested by fast/forms/range/slider-transformed.html
2833
2834         * html/shadow/SliderThumbElement.cpp:
2835         (WebCore::SliderThumbElement::setPositionFromPoint):
2836
2837 2013-08-29  Zan Dobersek  <zdobersek@igalia.com>
2838
2839         [Automake] libWebCoreDOM.la could use a better name
2840         https://bugs.webkit.org/show_bug.cgi?id=120232
2841
2842         Reviewed by Martin Robinson.
2843
2844         * bindings/gobject/GNUmakefile.am: Rename the libWebCoreDOM library to libGObjectDOMBindings.
2845
2846 2013-08-29  Joseph Pecoraro  <pecoraro@apple.com>
2847
2848         Web Inspector: Download Web Archive of Inspected Page
2849         https://bugs.webkit.org/show_bug.cgi?id=119774
2850
2851         Reviewed by Timothy Hatcher.
2852
2853         Add PageAgent.archive which will return a Base-64 encoded web archive
2854         when successful. In order to then allow saving non-string files, extend
2855         InspectorFrontendHost.save to allow for Base-64 encoded data.
2856
2857         * inspector/Inspector.json:
2858         * inspector/InspectorPageAgent.cpp:
2859         (WebCore::InspectorPageAgent::archive):
2860         * inspector/InspectorPageAgent.h:
2861         Introduce and implement PageAgent.archive. Create a Web Archive of the
2862         page's main frame.
2863
2864         * inspector/InspectorFrontendClient.h:
2865         * inspector/InspectorFrontendClientLocal.h:
2866         (WebCore::InspectorFrontendClientLocal::save):
2867         * inspector/InspectorFrontendHost.cpp:
2868         (WebCore::InspectorFrontendHost::save):
2869         * inspector/InspectorFrontendHost.h:
2870         * inspector/InspectorFrontendHost.idl:
2871         Extend InspectorFrontendHost.save to include a base64Encoded param.
2872
2873         * inspector/front-end/FileManager.js:
2874         * inspector/front-end/InspectorFrontendHostStub.js:
2875         (.WebInspector.InspectorFrontendHostStub.prototype.save):
2876         Misc. updates to the old inspector for function changes.
2877
2878 2013-08-29  Robert Hogan  <robert@webkit.org>
2879
2880         Positioned Replaced Elements That Aren't RenderReplaced get Incorrect Width
2881         https://bugs.webkit.org/show_bug.cgi?id=93735
2882
2883         Reviewed by David Hyatt.
2884
2885         Replaced elements that aren't RenderReplaced aren't |isReplaced| and don't have an
2886         intrinsic height or width. This causes them to go down the wrong height and width computation
2887         path in RenderBox when they are absolute positioned.
2888
2889         The notion of |isReplaced| is entwined with the notion of being |isInline| so it isn't really
2890         possible to make them isReplaced without re-wiring a lot of code. So instead use an ad-hoc definition
2891         of isReplacedElement in RenderBox to bring all replaced elements into the height and width calculation.
2892         To make sure we get the right height and width in there, give non-RenderReplaced replaced renderers
2893         the helpers for returning their approximation of intrinsic height and width.
2894
2895         The initial attempt at landing this patch had to be rolled out because it used LayoutUnit() for default
2896         intrinsic height of some replaced elements and this made the layout of the elements unstable in some sites.
2897         The fix for this issue is captured in intrinsic-button-and-input-height.html.
2898
2899         Tests: fast/replaced/intrinsic-button-and-input-height.html
2900                fast/replaced/width-and-height-of-positioned-replaced-elements.html
2901
2902         * rendering/RenderBox.cpp:
2903         (WebCore::isReplacedElement):
2904         (WebCore::RenderBox::computePositionedLogicalWidth):
2905         (WebCore::RenderBox::computePositionedLogicalHeight):
2906         * rendering/RenderBox.h:
2907         (WebCore::RenderBox::intrinsicSize):
2908         * rendering/RenderButton.h:
2909         * rendering/RenderFileUploadControl.cpp:
2910         (WebCore::RenderFileUploadControl::computePreferredLogicalWidths):
2911         * rendering/RenderListBox.cpp:
2912         (WebCore::RenderListBox::RenderListBox):
2913         (WebCore::RenderListBox::computePreferredLogicalWidths):
2914         (WebCore::RenderListBox::computeLogicalHeight):
2915         * rendering/RenderListBox.h:
2916         * rendering/RenderMenuList.h:
2917         * rendering/RenderReplaced.h:
2918         * rendering/RenderSlider.cpp:
2919         (WebCore::RenderSlider::computePreferredLogicalWidths):
2920         * rendering/RenderTextControl.cpp:
2921         (WebCore::RenderTextControl::RenderTextControl):
2922         (WebCore::RenderTextControl::computeLogicalHeight):
2923         * rendering/RenderTextControl.h:
2924
2925 2013-08-29  Brent Fulgham  <bfulgham@apple.com>
2926
2927         [Windows] Unreviewed gardening. Add missing *.css files from project.
2928
2929         * WebCore.vcxproj/WebCore.vcxproj:
2930         * WebCore.vcxproj/WebCore.vcxproj.filters:
2931
2932 2013-08-29  Pratik Solanki  <pratik.solanki@gmail.com>
2933
2934         SharedBuffer m_segments and m_dataArray must be exclusive
2935         https://bugs.webkit.org/show_bug.cgi?id=77715
2936
2937         Reviewed by Benjamin Poulain.
2938
2939         When USE(NETWORK_CFDATA_ARRAY_CALLBACK) is enabled, we use m_dataArray to hold the incoming
2940         data. We do not use m_segments. Since they are exclusive in practice, do not define or use
2941         m_segments when NETWORK_CFDATA_ARRAY_CALLBACK is enabled.
2942
2943         No new tests because no functional changes.
2944
2945         * platform/SharedBuffer.cpp:
2946         (WebCore::SharedBuffer::append):
2947         (WebCore::SharedBuffer::clear):
2948         (WebCore::SharedBuffer::copy):
2949         (WebCore::SharedBuffer::buffer):
2950         (WebCore::SharedBuffer::getSomeData):
2951         * platform/SharedBuffer.h:
2952
2953 2013-08-29  Daniel Bates  <dabates@apple.com>
2954
2955         [iOS] Upstream changes to WebCore/style
2956         https://bugs.webkit.org/show_bug.cgi?id=120173
2957
2958         Reviewed by Darin Adler.
2959
2960         * style/StyleResolveTree.cpp:
2961         (WebCore::Style::elementImplicitVisibility): Added; specific to iOS.
2962         (WebCore::Style::CheckForVisibilityChangeOnRecalcStyle::CheckForVisibilityChangeOnRecalcStyle): Added; specific to iOS.
2963         (WebCore::Style::CheckForVisibilityChangeOnRecalcStyle::~CheckForVisibilityChangeOnRecalcStyle): Added; specific to iOS.
2964         (WebCore::Style::resolveTree): Modified to instantiate CheckForVisibilityChangeOnRecalcStyle when building on iOS.
2965
2966 2013-08-29  Arnaud Renevier  <a.renevier@samsung.com>
2967
2968         [cairo] canvas drawing on itself doesn't work with accelerated canvas
2969         https://bugs.webkit.org/show_bug.cgi?id=118808
2970
2971         Reviewed by Martin Robinson.
2972
2973         When copying an accelerated image, we try to get its dimensions with
2974         cairo_image_surface_get_width/cairo_image_surface_get_height. As
2975         surface is not an image, this returns width and height of 0.
2976
2977         Many other places use cairo_image_surface_get although the surface may
2978         be a gl surface.
2979
2980         This patch fixes those issues by implementing a cairoSurfaceSize
2981         helper that returns the surface size whatever type it is.
2982
2983         It use cairo_surface_create_similar instead of
2984         cairo_image_surface_create in copyCairoImageSurface. It also calls
2985         cairo_paint in encodeImage when a drawing over a black background is
2986         needed.
2987
2988         It copies the surface to an image surface if needed in extractImage.
2989
2990         No new tests. Covered by existing tests.
2991
2992         * platform/graphics/cairo/BitmapImageCairo.cpp:
2993         (WebCore::BitmapImage::BitmapImage):
2994         (WebCore::BitmapImage::draw):
2995         (WebCore::BitmapImage::checkForSolidColor):
2996         * platform/graphics/cairo/CairoUtilities.cpp:
2997         (WebCore::copyCairoImageSurface):
2998         (WebCore::cairoSurfaceSize):
2999         * platform/graphics/cairo/CairoUtilities.h:
3000         * platform/graphics/cairo/GraphicsContext3DCairo.cpp:
3001         (WebCore::GraphicsContext3D::ImageExtractor::extractImage):
3002         * platform/graphics/gtk/GdkCairoUtilities.cpp:
3003         (cairoSurfaceToGdkPixbuf):
3004         * platform/graphics/gtk/GdkCairoUtilities.h:
3005         * platform/graphics/gtk/ImageBufferGtk.cpp:
3006         (WebCore::encodeImage):
3007         * platform/graphics/gtk/ImageGtk.cpp:
3008         (WebCore::BitmapImage::getGdkPixbuf):
3009         * platform/gtk/DragIcon.cpp:
3010         (WebCore::DragIcon::setImage):
3011
3012 2013-08-29  Tamas Czene  <tczene@inf.u-szeged.hu>
3013
3014         Buildfix after r154806 for !ENABLE(CSS_REGIONS) platforms.
3015         https://bugs.webkit.org/show_bug.cgi?id=120461
3016
3017         Reviewed by Csaba Osztrogonác.
3018
3019         moveToFlowThreadIfNeeded() is used only inside of ENABLE(CSS_REGIONS)
3020         block (and only once), so it should be guarded too.
3021
3022         * style/StyleResolveTree.cpp:
3023
3024 2013-08-29  David Kilzer  <ddkilzer@apple.com>
3025
3026         BUILD FIX (r154778): AXSearchFieldCancelButtonText() is not defined on iOS
3027
3028         See: AX: Cancel button in search field not accessible.
3029         <https://webkit.org/b/120322>
3030
3031         Fixes the following build error:
3032
3033             In file included from Source/WebCore/accessibility/AccessibilityAllInOne.cpp:44:
3034             Source/WebCore/accessibility/AccessibilitySearchFieldButtons.cpp:46:12: error: use of undeclared identifier 'AXSearchFieldCancelButtonText'
3035                 return AXSearchFieldCancelButtonText();
3036                        ^
3037             1 error generated.
3038
3039         * accessibility/AccessibilitySearchFieldButtons.cpp:
3040         (WebCore::AccessibilitySearchFieldCancelButton::accessibilityDescription):
3041         Return String() instead of calling AXSearchFieldCancelButtonText() on iOS.
3042
3043 2013-08-29  Antti Koivisto  <antti@apple.com>
3044
3045         Remove NodeRenderingContext
3046         https://bugs.webkit.org/show_bug.cgi?id=120466
3047
3048         Reviewed by Andreas Kling.
3049
3050         Switch the few remaining clients of this class to call the underlying code directly.
3051
3052         * CMakeLists.txt:
3053         * GNUmakefile.list.am:
3054         * Target.pri:
3055         * WebCore.exp.in:
3056         * WebCore.vcxproj/WebCore.vcxproj:
3057         * WebCore.xcodeproj/project.pbxproj:
3058         * css/StyleResolver.cpp:
3059         (WebCore::StyleResolver::State::initForStyleResolve):
3060         * dom/DOMAllInOne.cpp:
3061         * dom/Node.cpp:
3062         * dom/Node.h:
3063
3064             Remove Node::parentNodeForRenderingAndStyle(). Make clients invoke NodeRenderingTraversal::parent() directly.
3065
3066         * dom/NodeRenderingContext.cpp: Removed.
3067         * dom/NodeRenderingContext.h: Removed.
3068         * dom/ShadowRoot.h:
3069         * dom/Text.cpp:
3070         * dom/Text.h:
3071         * dom/TreeScope.cpp:
3072         * dom/TreeScope.h:
3073
3074             Remove virtual TreeScope::resetStyleInheritance. Make clients cast to ShadowRoot when needed.
3075
3076         * html/HTMLSummaryElement.cpp:
3077         (WebCore::HTMLSummaryElement::detailsElement):
3078         * style/StyleResolveTree.cpp:
3079         (WebCore::Style::createTextRendererIfNeeded):
3080
3081             For consistency switch resetStyleInheritance check to use parentNode() instead of renderingParentNode. This matches the NodeRenderingTraversal implementation.
3082             This difference is probably not testable in current trunk.
3083
3084         (WebCore::Style::resolveTree):
3085
3086 2013-08-29  Antti Koivisto  <antti@apple.com>
3087
3088         Move element renderer creation out of NodeRenderingContext
3089         https://bugs.webkit.org/show_bug.cgi?id=120461
3090
3091         Reviewed by Andreas Kling.
3092
3093         Move NodeRenderingContext::createRendererIfNeeded() and the related utility functions to StyleResolveTree.
3094
3095         Tighten typing and constness. Refactor sligthly to be more understandable.
3096
3097         * dom/Element.cpp:
3098         (WebCore::Element::shouldMoveToFlowThread):
3099         * dom/Element.h:
3100         * dom/NodeRenderingContext.cpp:
3101         (WebCore::NodeRenderingContext::NodeRenderingContext):
3102         (WebCore::NodeRenderingContext::nextRenderer):
3103         (WebCore::NodeRenderingContext::previousRenderer):
3104         (WebCore::NodeRenderingContext::parentRenderer):
3105         * dom/NodeRenderingContext.h:
3106         * dom/PseudoElement.h:
3107         * style/StyleResolveTree.cpp:
3108         (WebCore::Style::nextSiblingRenderer):
3109         (WebCore::Style::shouldCreateRenderer):
3110         (WebCore::Style::elementInsideRegionNeedsRenderer):
3111         (WebCore::Style::moveToFlowThreadIfNeeded):
3112         (WebCore::Style::createRendererIfNeeded):
3113         (WebCore::Style::attachRenderTree):
3114         * svg/SVGElement.cpp:
3115         (WebCore::SVGElement::shouldMoveToFlowThread):
3116         * svg/SVGElement.h:
3117
3118 2013-08-28  Chris Fleizach  <cfleizach@apple.com>
3119
3120         AX: WebProcess at com.apple.WebCore: WebCore::AXObjectCache::rootObject + 27
3121         https://bugs.webkit.org/show_bug.cgi?id=120434
3122
3123         Reviewed by Darin Adler.
3124
3125         Crash logs indicate that there's a null pointer access in rootObject. That seems like it could only
3126         happen in Document was null.
3127
3128         Unfortunately, there are no reproducible steps and no other information to construct a test case.
3129
3130         * accessibility/AXObjectCache.cpp:
3131         (WebCore::AXObjectCache::rootObject):
3132
3133 2013-08-28  Ryosuke Niwa  <rniwa@webkit.org>
3134
3135         The code to look for an ancestor form element is duplicated in three different places
3136         https://bugs.webkit.org/show_bug.cgi?id=120391
3137
3138         Reviewed by Darin Adler.
3139
3140         Unduplicated the code by putting a single implementation in HTMLFormElement.cpp.
3141
3142         * WebCore.order:
3143         * html/FormAssociatedElement.cpp:
3144         (WebCore::FormAssociatedElement::findAssociatedForm):
3145         (WebCore::FormAssociatedElement::formAttributeChanged):
3146         * html/HTMLElement.cpp:
3147         (WebCore::HTMLElement::virtualForm):
3148         * html/HTMLElement.h:
3149         * html/HTMLFormControlElement.cpp:
3150         (WebCore::HTMLFormControlElement::HTMLFormControlElement):
3151         * html/HTMLFormElement.cpp:
3152         (WebCore::HTMLFormElement::findClosestFormAncestor):
3153         * html/HTMLFormElement.h:
3154         * html/HTMLImageElement.cpp:
3155         (WebCore::HTMLImageElement::insertedInto):
3156         * html/HTMLObjectElement.cpp:
3157         (WebCore::HTMLObjectElement::HTMLObjectElement):
3158         * html/parser/HTMLTreeBuilder.cpp:
3159         (WebCore::HTMLTreeBuilder::HTMLTreeBuilder):
3160
3161 2013-08-28  Ryosuke Niwa  <rniwa@webkit.org>
3162
3163         Stop throwing DOM exceptions in internal 'XMLHttpRequest' response getters
3164         https://bugs.webkit.org/show_bug.cgi?id=120446
3165
3166         Reviewed by Alexey Proskuryakov.
3167
3168         Merge https://chromium.googlesource.com/chromium/blink/+/c8188c21452501b68950a9fcc1f5cbc7b4de4df5
3169
3170         Unlike 'responseText' and 'responseXML', 'responseBlob' and
3171         'responseArrayBuffer' are not exposed to JavaScript (they don't
3172         appear in the IDL or in the specification). As they are only called from
3173         custom bindings in response to a JavaScript call to the 'response' getter,
3174         we can safely replace the exception-throwing code in the implementation
3175         with an ASSERT that the request type is correct.
3176
3177         * bindings/js/JSXMLHttpRequestCustom.cpp:
3178         (WebCore::JSXMLHttpRequest::response):
3179         * xml/XMLHttpRequest.cpp:
3180         (WebCore::XMLHttpRequest::responseBlob):
3181         (WebCore::XMLHttpRequest::responseArrayBuffer):
3182         * xml/XMLHttpRequest.h:
3183
3184 2013-08-28  Chris Curtis  <chris_curtis@apple.com>
3185
3186         https://bugs.webkit.org/show_bug.cgi?id=119548
3187         Refactoring Exception throws.
3188
3189         Reviewed by Geoffrey Garen.
3190
3191         Gets column information from the error object for reporting exceptions.
3192         * bindings/js/JSDOMBinding.cpp:
3193         (WebCore::reportException):
3194         * bindings/js/ScriptCallStackFactory.cpp:
3195         (WebCore::createScriptCallStackFromException):
3196
3197         Moved setting an exception into the vm, These functions changed to use the new functionality.
3198
3199         * bindings/js/JSAudioBufferSourceNodeCustom.cpp:
3200         (WebCore::JSAudioBufferSourceNode::setBuffer):
3201         * bindings/js/JSBiquadFilterNodeCustom.cpp:
3202         (WebCore::JSBiquadFilterNode::setType):
3203         * bindings/js/JSCryptoCustom.cpp:
3204         (WebCore::JSCrypto::getRandomValues):
3205         * bindings/js/JSDOMBinding.cpp:
3206         (WebCore::setDOMException):
3207         * bindings/js/JSInjectedScriptHostCustom.cpp:
3208         (WebCore::JSInjectedScriptHost::setFunctionVariableValue):
3209         * bindings/js/JSJavaScriptCallFrameCustom.cpp:
3210         (WebCore::JSJavaScriptCallFrame::evaluate):
3211         (WebCore::JSJavaScriptCallFrame::setVariableValue):
3212         * bindings/js/JSNodeFilterCondition.cpp:
3213         (WebCore::JSNodeFilterCondition::acceptNode):
3214         * bindings/js/JSOscillatorNodeCustom.cpp:
3215         (WebCore::JSOscillatorNode::setType):
3216         * bindings/js/JSPannerNodeCustom.cpp:
3217         (WebCore::JSPannerNode::setPanningModel):
3218         (WebCore::JSPannerNode::setDistanceModel):
3219         * bindings/js/JSSVGLengthCustom.cpp:
3220         (WebCore::JSSVGLength::convertToSpecifiedUnits):
3221         * bindings/js/JSWebGLRenderingContextCustom.cpp:
3222         (WebCore::getObjectParameter):
3223         (WebCore::JSWebGLRenderingContext::getAttachedShaders):
3224         (WebCore::JSWebGLRenderingContext::getExtension):
3225         (WebCore::JSWebGLRenderingContext::getFramebufferAttachmentParameter):
3226         (WebCore::JSWebGLRenderingContext::getParameter):
3227         (WebCore::JSWebGLRenderingContext::getProgramParameter):
3228         (WebCore::JSWebGLRenderingContext::getShaderParameter):
3229         (WebCore::JSWebGLRenderingContext::getUniform):
3230         (WebCore::dataFunctionf):
3231         (WebCore::dataFunctioni):
3232         (WebCore::dataFunctionMatrix):
3233         * bindings/js/JSXMLHttpRequestCustom.cpp:
3234         (WebCore::JSXMLHttpRequest::open):
3235         * bindings/js/SerializedScriptValue.cpp:
3236         (WebCore::CloneBase::throwStackOverflow):
3237         (WebCore::CloneDeserializer::throwValidationError):
3238         (WebCore::SerializedScriptValue::maybeThrowExceptionIfSerializationFailed):
3239         * bindings/js/WorkerScriptController.cpp:
3240         (WebCore::WorkerScriptController::evaluate):
3241         (WebCore::WorkerScriptController::setException):
3242         * bridge/c/c_instance.cpp:
3243         (JSC::Bindings::CInstance::moveGlobalExceptionToExecState):
3244         (JSC::Bindings::CInstance::invokeMethod):
3245         (JSC::Bindings::CInstance::invokeDefaultMethod):
3246         (JSC::Bindings::CInstance::invokeConstruct):
3247         (JSC::Bindings::CInstance::toJSPrimitive):
3248         * bridge/objc/objc_instance.mm:
3249         (ObjcInstance::invokeMethod):
3250         * bridge/objc/objc_runtime.mm:
3251         (JSC::Bindings::ObjcArray::setValueAt):
3252         (JSC::Bindings::ObjcArray::valueAt):
3253         * bridge/objc/objc_utility.mm:
3254         (JSC::Bindings::throwError):
3255         * bridge/qt/qt_instance.cpp:
3256         (JSC::Bindings::QtField::valueFromInstance):
3257         (JSC::Bindings::QtField::setValueToInstance):
3258         * bridge/runtime_array.cpp:
3259         (JSC::RuntimeArray::put):
3260         (JSC::RuntimeArray::putByIndex):
3261         * bridge/runtime_object.cpp:
3262         (JSC::Bindings::RuntimeObject::throwInvalidAccessError):
3263
3264 2013-08-28  Alexey Proskuryakov  <ap@apple.com>
3265
3266         Remove an unused data member from Page.
3267
3268         Rubber-stamped by Brady Eidson.
3269
3270         * page/Page.cpp:
3271         * page/Page.h:
3272         Removed m_cookieEnabled. This was completely dead code, long obsoleted by PageSettings.
3273
3274 2013-08-28  Gustavo Noronha Silva  <gns@gnome.org>
3275
3276         Unreviewed build fix - copy/paste failure, copied too much.
3277
3278         * bindings/gobject/WebKitDOMCustom.h:
3279
3280 2013-08-28  Dean Jackson  <dino@apple.com>
3281
3282         [WebGL] CoreGraphics can provide greyscale image data
3283         https://webkit.org/b/120443
3284
3285         Reviewed by Simon Fraser.
3286
3287         CoreGraphics can decode greyscale or greyscale+alpha images
3288         while preserving the format. Our WebGL texture unpacker
3289         was seeing this and assuming it meant the data did not come
3290         from an <img> element. Since that method already special cased
3291         CoreGraphics, the fix was to simply return true for these
3292         extra types.
3293
3294         I also renamed srcFormatComeFromDOMElementOrImageData
3295         to srcFormatComesFromDOMElementOrImageData.
3296
3297         Test: fast/canvas/webgl/tex-image-with-greyscale-image.html
3298
3299         * platform/graphics/GraphicsContext3D.cpp: Call new name.
3300         * platform/graphics/GraphicsContext3D.h:
3301         (WebCore::GraphicsContext3D::srcFormatComesFromDOMElementOrImageData):
3302         Add support for R8, AR8, A8, and RA8 data formats.
3303
3304 2013-08-28  Gustavo Noronha Silva  <gns@gnome.org>
3305
3306         [GTK] HTMLElement lost setID and getID - need to add compatibility symbols
3307         https://bugs.webkit.org/show_bug.cgi?id=120440
3308
3309         Reviewed by Martin Robinson.
3310
3311         No tests, just adding compatibility symbols.
3312
3313         setID and getID were removed, and the parent class (Element) ones should be used instead.
3314         We need to keep our ABI compatible, though, so add compatibility symbols.
3315
3316         * bindings/gobject/WebKitDOMCustom.cpp:
3317         (webkit_dom_html_element_get_id):
3318         (webkit_dom_html_element_set_id):
3319         * bindings/gobject/WebKitDOMCustom.h:
3320
3321 2013-08-28  Simon Fraser  <simon.fraser@apple.com>
3322
3323         Fix compositing layers in columns
3324         https://bugs.webkit.org/show_bug.cgi?id=120436
3325
3326         Reviewed by Dave Hyatt.
3327
3328         Remove the old hack in RenderLayer::updateLayerPosition() for placing
3329         layers in columns, which changed the layer position for composited
3330         layers; this broke hit-testing.
3331
3332         Fix a better way by moving compositing layers to the correct
3333         positions that take column offsets into account, by fixing
3334         RenderLayer::convertToLayerCoords() to optionally apply column
3335         adjustment, and using this in the code which positions compositing layers.
3336
3337         Tests: compositing/columns/ancestor-clipped-in-paginated.html
3338                compositing/columns/clipped-in-paginated.html
3339                compositing/columns/composited-columns-vertical-rl.html
3340                compositing/columns/composited-columns.html
3341                compositing/columns/composited-in-paginated-rl.html
3342                compositing/columns/composited-in-paginated-writing-mode-rl.html
3343                compositing/columns/composited-lr-paginated-repaint.html
3344                compositing/columns/composited-rl-paginated-repaint.html
3345                compositing/columns/hittest-composited-in-paginated.html
3346                compositing/columns/rotated-in-paginated.html
3347                compositing/columns/untransformed-composited-in-paginated.html
3348
3349         * rendering/RenderLayer.cpp:
3350         (WebCore::RenderLayer::updateLayerPosition):
3351         (WebCore::RenderLayer::convertToPixelSnappedLayerCoords):
3352         (WebCore::accumulateOffsetTowardsAncestor):
3353         (WebCore::RenderLayer::convertToLayerCoords):
3354         * rendering/RenderLayer.h:
3355         * rendering/RenderLayerBacking.cpp:
3356         (WebCore::RenderLayerBacking::updateCompositedBounds):
3357         (WebCore::RenderLayerBacking::updateGraphicsLayerGeometry):
3358
3359 2013-08-28  Brent Fulgham  <bfulgham@webkit.org>
3360
3361         [WinCairo] Unreviewed build fix.
3362
3363         * WebCore.vcxproj/WebCore.vcxproj: Don't exclude the full screen
3364         window from the build.
3365         * WebCore.vcxproj/WebCore.vcxproj.filters: Ditto.
3366
3367 2013-08-28  Benjamin Poulain  <benjamin@webkit.org>
3368
3369         Fix the build after r154780
3370
3371         * dom/SpaceSplitString.cpp:
3372         (WebCore::SpaceSplitStringData::create):
3373
3374 2013-08-27  Benjamin Poulain  <bpoulain@apple.com>
3375
3376         Simplify and clean SpaceSplitString
3377         https://bugs.webkit.org/show_bug.cgi?id=120385
3378
3379         Reviewed by Ryosuke Niwa.
3380
3381         Clean up of SpaceSplitString following the cleaning of the DOMTokenList hierarchy.
3382         This brings the following:
3383         - Fix the coding style of the header.
3384         - Remove the concepts of empty and null. The list can now be empty or have something.
3385           There is no null state.
3386         - Put the tokens directly following SpaceSplitStringData instead of using a Vector in between.
3387
3388         * WebCore.exp.in:
3389         * dom/ElementData.h:
3390         (WebCore::ElementData::hasClass):
3391         * dom/SpaceSplitString.cpp:
3392         (WebCore::tokenizeSpaceSplitString):
3393         (WebCore::SpaceSplitStringData::containsAll):
3394         (WebCore::SpaceSplitString::set):
3395         (WebCore::SpaceSplitString::spaceSplitStringContainsValue):
3396         (WebCore::TokenCounterProcessor::TokenCounterProcessor):
3397         (WebCore::TokenCounterProcessor::processToken):
3398         (WebCore::TokenCounterProcessor::tokenCount):
3399         (WebCore::TokenInitializerProcessor::TokenInitializerProcessor):
3400         (WebCore::TokenInitializerProcessor::processToken):
3401         (WebCore::TokenInitializerProcessor::nextMemoryBucket):
3402         (WebCore::SpaceSplitStringData::create):
3403         (WebCore::SpaceSplitStringData::destroy):
3404         * dom/SpaceSplitString.h:
3405         (WebCore::SpaceSplitStringData::contains):
3406         (WebCore::SpaceSplitStringData::size):
3407         (WebCore::SpaceSplitStringData::operator[]):
3408         (WebCore::SpaceSplitStringData::ref):
3409         (WebCore::SpaceSplitStringData::deref):
3410         (WebCore::SpaceSplitStringData::SpaceSplitStringData):
3411         (WebCore::SpaceSplitStringData::~SpaceSplitStringData):
3412         (WebCore::SpaceSplitStringData::tokenArrayStart):
3413         (WebCore::SpaceSplitString::SpaceSplitString):
3414         (WebCore::SpaceSplitString::operator!=):
3415         (WebCore::SpaceSplitString::clear):
3416         (WebCore::SpaceSplitString::contains):
3417         (WebCore::SpaceSplitString::containsAll):
3418         (WebCore::SpaceSplitString::size):
3419         (WebCore::SpaceSplitString::isEmpty):
3420         (WebCore::SpaceSplitString::operator[]):
3421         (WebCore::SpaceSplitString::spaceSplitStringContainsValue):
3422         * html/ClassList.cpp:
3423         (WebCore::ClassList::classNames):
3424         * page/EventHandler.cpp:
3425         (WebCore::findDropZone):
3426
3427 2013-08-28  Benjamin Poulain  <benjamin@webkit.org>
3428
3429         Simplify and clean SpaceSplitString
3430    &nbs