Source/WebCore/ChangeLog

   1 2012-07-13  Ryosuke Niwa  <rniwa@webkit.org>
   2
   3         NodeLists should not invalidate on irreleavnt attribute changes
   4         https://bugs.webkit.org/show_bug.cgi?id=91277
   5
   6         Reviewed by Ojan Vafai.
   7
   8         Explicitely check the invalidation type and the changed attribute in NodeListNodeData::invalidateCaches
   9         and ElementRareData::clearHTMLCollectionCaches to only invalidate node lists affected by the change.
  10
  11         Also merged invalidateNodeListsCacheAfterAttributeChanged and invalidateNodeListsCacheAfterChildrenChanged
  12         as invalidateNodeListCachesInAncestors since they're almost identical after r122498.
  13
  14         In addition, moved shouldInvalidateNodeListForType from Document.cpp to DynamicNodeList.h and renamed it to
  15         shouldInvalidateTypeOnAttributeChange since it needs to called in Node.cpp and ElementRareData.h.
  16
  17         * dom/Attr.cpp:
  18         (WebCore::Attr::setValue):
  19         (WebCore::Attr::childrenChanged):
  20         * dom/ContainerNode.cpp:
  21         (WebCore::ContainerNode::childrenChanged):
  22         * dom/Document.cpp:
  23         (WebCore::Document::registerNodeListCache): Calls isRootedAtDocument() instead of directly comparing
  24         the value of NodeListRootType in order to prepare for the bug 80269.
  25         (WebCore::Document::unregisterNodeListCache): Ditto.
  26         (WebCore): shouldInvalidateNodeListForType is moved to DynamicNodeList.h
  27         (WebCore::Document::shouldInvalidateNodeListCaches):
  28         * dom/DynamicNodeList.h:
  29         (DynamicNodeListCacheBase):
  30         (WebCore::DynamicNodeListCacheBase::shouldInvalidateTypeOnAttributeChange): Moved from Document.cpp.
  31         * dom/Element.cpp:
  32         (WebCore::Element::attributeChanged):
  33         * dom/ElementRareData.h:
  34         (WebCore::ElementRareData::clearHTMLCollectionCaches): Takes const QualifiedName* to compare against
  35         the invalidation type of HTML collections via shouldInvalidateTypeOnAttributeChange.
  36         * dom/Node.cpp:
  37         (WebCore::Node::invalidateNodeListCachesInAncestors): Merged invalidateNodeListCachesInAncestors and
  38         invalidateNodeListsCacheAfterChildrenChanged. Also pass attrName to clearHTMLCollectionCaches.
  39         (WebCore::NodeListsNodeData::invalidateCaches): Compares attrName against the invalidation type of
  40         node lists via shouldInvalidateTypeOnAttributeChange.
  41         (WebCore):
  42         * dom/Node.h:
  43         (Node):
  44         * dom/NodeRareData.h:
  45         (WebCore::NodeRareData::ensureNodeLists): Merged NodeRareData::createNodeLists.
  46         (WebCore::NodeRareData::clearChildNodeListCache): Moved from Node.cpp.
  47         (NodeRareData):
  48         * html/HTMLCollection.h:
  49         (HTMLCollectionCacheBase):
  50
  51 2012-07-13  Arpita Bahuguna  <arpitabahuguna@gmail.com>
  52
  53         Refactor RenderTable to use the section's iteration functions.
  54         https://bugs.webkit.org/show_bug.cgi?id=89751
  55
  56         Reviewed by Julien Chaffraix.
  57
  58         Removing anti-pattern wherever possible from RenderTable code. Also, modifying
  59         RenderTable sections' iterations to use helper functions.
  60
  61         No new tests required for this change since no change in behavior is expected.
  62
  63         * rendering/RenderTable.cpp:
  64         (WebCore::RenderTable::addOverflowFromChildren):
  65         (WebCore::RenderTable::setCellLogicalWidths):
  66         (WebCore::RenderTable::outerBorderStart):
  67         (WebCore::RenderTable::outerBorderEnd):
  68         Removed anti-patterns involving iterations over RenderObjects.
  69
  70         (WebCore::RenderTable::outerBorderAfter):
  71         Modified RenderTable sections' iteration to use helper functions.
  72
  73 2012-07-13  Enrica Casucci  <enrica@apple.com>
  74
  75         Threadsafety issues in WebScriptObject
  76         https://bugs.webkit.org/show_bug.cgi?id=90849
  77
  78         Reviewed by Geoff Garen.
  79
  80         Updated fix for this bug. The JSC API lock needs to be acquired also in JSObject.
  81
  82         * bindings/objc/WebScriptObject.mm:
  83         (-[WebScriptObject JSObject]):
  84
  85 2012-07-13  Raymond Toy  <rtoy@google.com>
  86
  87         DelayNode doesn't work if delayTime.value == delayTime.maxValue
  88         https://bugs.webkit.org/show_bug.cgi?id=90357
  89
  90         Reviewed by Kenneth Russell.
  91
  92         Increase delay buffer size slightly so that the read and write
  93         pointers don't become equal when the delay and the max delay are
  94         the same.
  95
  96         Tests: webaudio/delaynode-max-default-delay.html
  97                webaudio/delaynode-max-nondefault-delay.html
  98
  99         * Modules/webaudio/DelayDSPKernel.cpp:
 100         (WebCore): Moved SmoothingTimeConstant to WebCore namespace.
 101         (WebCore::DelayDSPKernel::DelayDSPKernel): Add some additional checks to prevent crashes; use bufferLengthForDelay to compute buffer length.
 102         (WebCore::DelayDSPKernel::bufferLengthForDelay): New function to compute buffer length.
 103         * Modules/webaudio/DelayDSPKernel.h:
 104         (DelayDSPKernel): Declare bufferLengthForDelay.
 105
 106 2012-07-13  Benjamin Poulain  <bpoulain@apple.com>
 107
 108         Always aggressively preload on iOS
 109         https://bugs.webkit.org/show_bug.cgi?id=91276
 110
 111         Reviewed by Simon Fraser.
 112
 113         * loader/cache/CachedResourceLoader.cpp:
 114         (WebCore::CachedResourceLoader::preload):
 115
 116 2012-07-13  Vineet Chaudhary  <rgf748@motorola.com>
 117
 118         Restructure V8Utilities::extractTransferables() with help of toV8Sequence()
 119         https://bugs.webkit.org/show_bug.cgi?id=91208
 120
 121         Reviewed by Kentaro Hara.
 122
 123         We can remove the specialised check for MessagePort from V8Utilities::extractTransferables()
 124         using toV8Sequence() as it validates the passed object for sequence type per WebIDL spec.
 125
 126         No new test as just refactoring.
 127         Existing tests under fast/dom/Window/window-* fast/dom/events/*
 128         covers tests.
 129
 130         * bindings/v8/V8Utilities.cpp:
 131         (WebCore::extractTransferables):
 132
 133 2012-07-13  Vincent Scheib  <scheib@chromium.org>
 134
 135         Pointer Lock handles disconnected DOM elements
 136         https://bugs.webkit.org/show_bug.cgi?id=77029
 137
 138         Reviewed by Adrienne Walker.
 139
 140         Pointer Lock Controller now checks when elements or documents are
 141         removed, and unlocks if the target element is being removed.
 142
 143         Tests: pointer-lock/locked-element-iframe-removed-from-dom.html
 144                pointer-lock/locked-element-removed-from-dom.html
 145
 146         * dom/Document.cpp:
 147         (WebCore::Document::detach):
 148         * dom/Element.cpp:
 149         (WebCore::Element::removedFrom):
 150         (WebCore::Element::webkitRequestPointerLock):
 151         * page/PointerLockController.cpp:
 152         (WebCore::PointerLockController::requestPointerLock):
 153         (WebCore::PointerLockController::elementRemoved):
 154         (WebCore):
 155         (WebCore::PointerLockController::documentDetached):
 156         (WebCore::PointerLockController::didLosePointerLock):
 157         (WebCore::PointerLockController::enqueueEvent):
 158         * page/PointerLockController.h:
 159         (WebCore):
 160         (PointerLockController):
 161
 162 2012-07-13  Ryosuke Niwa  <rniwa@webkit.org>
 163
 164         HTMLCollection should use DynamicNodeList's invalidation model
 165         https://bugs.webkit.org/show_bug.cgi?id=90326
 166
 167         Reviewed by Anders Carlsson.
 168
 169         Make HTMLCollection invalidated upon attribute and children changes instead of invalidating it on demand
 170         by comparing DOM tree versions. Node that HTMLCollections owned by Document are invalidated with other
 171         document-rooted node lists in m_listsInvalidatedAtDocument for simplicity although this mechanism is
 172         normally used for node lists owned by a non-Document node that contains nodes outside of its subtree.
 173         ItemProperties and FormControls are more "traditional" users of the mechanism.
 174
 175         Also, merged DynamicNodeList::invalidateCache and HTMLCollection::invalidateCache.
 176
 177         * dom/Document.cpp:
 178         (WebCore::Document::registerNodeListCache): Renamed. No longer takes NodeListInvalidationType or
 179         NodeListRootType since they can be obtained from the cache base. Increment the node list counter for
 180         InvalidateOnIdNameAttrChange when a HTMLCollection is passed in since all HTMLCollections need to be
 181         invalidated on id or name content attribute changes due to named getters.
 182         (WebCore::Document::unregisterNodeListCache): Ditto.
 183         (WebCore::shouldInvalidateNodeListForType):
 184         (WebCore::Document::shouldInvalidateNodeListCaches):
 185         (WebCore::Document::clearNodeListCaches):
 186         * dom/Document.h:
 187         (WebCore): Added InvalidateOnIdNameAttrChange, InvalidateOnHRefAttrChange, and InvalidateOnAnyAttrChange.
 188         (Document):
 189         * dom/DynamicNodeList.cpp:
 190         (WebCore::DynamicNodeListCacheBase::invalidateCache): Added. Invalidates caches of both DynamicNodeList
 191         and HTMLCollection. We can't afford to use virtual function calls here because this function is called on
 192         all node lists and HTML collections owned by ancestors of an element under which a node is added, removed,
 193         or its attributes are changed.
 194         (WebCore):
 195         * dom/DynamicNodeList.h:
 196         (WebCore::DynamicNodeListCacheBase::DynamicNodeListCacheBase): Initializes member variables directly
 197         instead of calling clearCache now that DynamicNodeListCacheBase::invalidateCache has become polymorphic.
 198         (DynamicNodeListCacheBase): Increased the number of bits for m_invalidationType since we now have 9
 199         invalidation types.
 200         (WebCore::DynamicSubtreeNodeList::~DynamicSubtreeNodeList):
 201         (WebCore::DynamicSubtreeNodeList::DynamicSubtreeNodeList):
 202         * dom/ElementRareData.h:
 203         (ElementRareData):
 204         (WebCore::ElementRareData::clearHTMLCollectionCaches): Added.
 205         (WebCore::ElementRareData::adoptTreeScope): Added; similar to NodeRareData::adoptTreeScope.
 206         * dom/Node.cpp:
 207         (WebCore::Node::invalidateNodeListsCacheAfterAttributeChanged): Clears HTML collection caches as well as
 208         node list caches.
 209         (WebCore::Node::invalidateNodeListsCacheAfterChildrenChanged): Ditto.
 210         * dom/NodeRareData.h:
 211         (WebCore::NodeListsNodeData::adoptTreeScope):
 212         * dom/TreeScopeAdopter.cpp:
 213         (WebCore::TreeScopeAdopter::moveTreeToNewScope): Calls ElementRareData's adoptTreeScope as well as
 214         NodeRareData's.
 215         * html/HTMLAllCollection.cpp:
 216         (WebCore::HTMLAllCollection::namedItemWithIndex):
 217         * html/HTMLCollection.cpp:
 218         (WebCore::rootTypeFromCollectionType): Added. As mentioned above, treat all Document-owned HTML collection
 219         as if rooted at document for convenience.
 220         (WebCore::invalidationTypeExcludingIdAndNameAttributes): Added. Since all HTML collection requires
 221         invalidation on id and name content attribute changes, which is taken care by the special logic in
 222         Document::registerNodeListCache, exclude those two attributes from consideration.
 223         (WebCore::HTMLCollection::HTMLCollection): Calls Document::registerNodeListCache.
 224         (WebCore::HTMLCollection::~HTMLCollection): Calls Document::unregisterNodeListCache.
 225         (WebCore::HTMLCollection::length):
 226         (WebCore::HTMLCollection::item):
 227         (WebCore::HTMLCollection::namedItem):
 228         (WebCore::HTMLCollection::updateNameCache):
 229         * html/HTMLCollection.h:
 230         (WebCore::HTMLCollectionCacheBase::HTMLCollectionCacheBase):
 231         (HTMLCollectionCacheBase): Removed m_cacheTreeVersion and clearCache since they're no longer used.
 232         (HTMLCollection):
 233         * html/HTMLFormCollection.cpp:
 234         (WebCore::HTMLFormCollection::namedItem):
 235         (WebCore::HTMLFormCollection::updateNameCache):
 236         * html/HTMLOptionsCollection.h:
 237         (HTMLOptionsCollection):
 238         * html/HTMLPropertiesCollection.cpp:
 239         (WebCore::HTMLPropertiesCollection::updateNameCache):
 240         * html/HTMLPropertiesCollection.h:
 241         (WebCore::HTMLPropertiesCollection::invalidateCache):
 242
 243 2012-07-13  Shawn Singh  <shawnsingh@chromium.org>
 244
 245         [chromium] Remove incorrect debug assertion in LayerRendererChromium.cpp
 246         https://bugs.webkit.org/show_bug.cgi?id=91260
 247
 248         Reviewed by Adrienne Walker.
 249
 250         ASSERT(!clipped) was being triggered after skinny almost-degenerate
 251         quads went through anti-aliasing inflation, and then were being
 252         transformed back from device space to local space. It turns out
 253         this assertion is too aggressive, and we don't yet have an obvious
 254         need to change the behavior on the clipped==true case.
 255
 256         No new tests needed, this patch fixes only comments and debug code.
 257
 258         * platform/graphics/chromium/LayerRendererChromium.cpp:
 259         (WebCore::LayerRendererChromium::drawRenderPassQuad):
 260            fixed a comment.
 261
 262         (WebCore::LayerRendererChromium::drawTileQuad):
 263            fixed a similar comment, removed unnecessary assertion.
 264
 265 2012-07-13  Philip Rogers  <pdr@google.com>
 266
 267         Remove assert in localCoordinateSpaceTransform()
 268         https://bugs.webkit.org/show_bug.cgi?id=91189
 269
 270         Reviewed by Nikolas Zimmermann.
 271
 272         The assert in localCoordinateSpaceTransform was added to catch subclasses forgetting
 273         to override the method but it is better to simply return the identity matrix.
 274
 275         This scenario can occur when we break the SVG content model, such as asking for
 276         the CTM of a <g> element inside a <tspan>. This is undefined in the spec because
 277         tspan is not a subclass of SVGLocatable but both Firefox and Opera
 278         implement this by returning the identity matrix.
 279
 280         Test: svg/custom/invalid-ctm.svg
 281
 282         * svg/SVGStyledElement.cpp:
 283         (WebCore::SVGStyledElement::localCoordinateSpaceTransform):
 284
 285 2012-07-13  Kentaro Hara  <haraken@chromium.org>
 286
 287         [V8] String wrappers should be marked Independent
 288         https://bugs.webkit.org/show_bug.cgi?id=91251
 289
 290         Reviewed by Adam Barth.
 291
 292         Currently V8 String wrappers are not marked Independent.
 293         By marking them Independent, they can be reclaimed by the scavenger GC.
 294
 295         I tried to find some cases where this change reduces memory usage,
 296         but couldn't due to sensitive behavior of GC.
 297
 298         No tests. No change in behavior.
 299
 300         * bindings/v8/V8Binding.cpp:
 301         (WebCore::StringCache::v8ExternalStringSlow):
 302
 303 2012-07-13  Peter Beverloo  <peter@chromium.org>
 304
 305         [Chromium] Make the v8 i18n API dependency conditional for Android, disable strict aliasing
 306         https://bugs.webkit.org/show_bug.cgi?id=91240
 307
 308         Reviewed by Adam Barth.
 309
 310         Disable the v8 internationalization API for Chromium Android, as it's
 311         disabled and not always available in checkouts. Furthermore, disable
 312         strict aliasing for the webkit_remaining target, similar to what
 313         x11-based builds are doing (see the webcore_prerequisites target).
 314
 315         * WebCore.gyp/WebCore.gyp:
 316
 317 2012-07-13  Kentaro Hara  <haraken@chromium.org>
 318
 319         [CallWith=XXX] arguments should be placed at the head of method arguments
 320         https://bugs.webkit.org/show_bug.cgi?id=91217
 321
 322         Reviewed by Adam Barth.
 323
 324         The EFL build with the ENABLE_FILE_SYSTEM flag caused a build error,
 325         because CodeGeneratorJS.pm assumes webkitEntries(ScriptExecutionContext*, HTMLInputElement*)
 326         but the actual signature is webkitEntries(HTMLInputElement*, ScriptExecutionContext*) (bug 91185).
 327
 328         Per https://trac.webkit.org/wiki/WebKitIDL#CallWith, [CallWith=XXX] arguments should be placed
 329         at the head of the arguments. (i.e. the behavior of CodeGeneratorJS.pm is correct.)
 330
 331         Thus the correct fix is (1) to change the signature of webkitEntries() and webkitGetAsEntry()
 332         so that ScriptExecutionContext* comes first and (2) to modify CodeGeneratorV8.pm to support the order.
 333
 334         Test: bindings/scripts/test/TestObj.idl
 335
 336         * Modules/filesystem/DataTransferItemFileSystem.h: Placed ScriptExecutionContext* at the head of arguments.
 337         (DataTransferItemFileSystem):
 338         * Modules/filesystem/HTMLInputElementFileSystem.cpp: Ditto.
 339         (WebCore::HTMLInputElementFileSystem::webkitEntries):
 340         * Modules/filesystem/HTMLInputElementFileSystem.h: Ditto.
 341         (HTMLInputElementFileSystem):
 342         * Modules/filesystem/chromium/DataTransferItemFileSystemChromium.cpp: Ditto.
 343         (WebCore::DataTransferItemFileSystem::webkitGetAsEntry):
 344
 345         * bindings/scripts/CodeGeneratorV8.pm: Modified to support the correct order.
 346         (GenerateNormalAttrGetter):
 347         (GenerateNormalAttrSetter):
 348         (GenerateFunctionCallString):
 349
 350         * bindings/scripts/test/V8/V8TestInterface.cpp: Updated run-bindings-tests results.
 351         (WebCore::TestInterfaceV8Internal::supplementalMethod2Callback):
 352
 353 2012-07-13  Mary Wu  <mary.wu@torchmobile.com.cn>
 354
 355         [BlackBerry] Some small changes in network code
 356         https://bugs.webkit.org/show_bug.cgi?id=90974
 357
 358         Reviewed by Rob Buis.
 359
 360         1. Set status in NetworkJob/SocketStreamHandleBlackBerry so that
 361         its wrapped stream can also query the stream result.
 362         2. pass download attribute to NetworkRequest.
 363
 364         RIM PR# 171555
 365         Reviewed internally by Lyon Chen and Joe Mason.
 366
 367         * platform/network/blackberry/NetworkJob.cpp:
 368         (WebCore::NetworkJob::handleNotifyClose):
 369         * platform/network/blackberry/NetworkJob.h:
 370         (WebCore::NetworkJob::status):
 371         * platform/network/blackberry/ResourceRequestBlackBerry.cpp:
 372         (WebCore::platformTargetTypeForRequest):
 373         * platform/network/blackberry/SocketStreamHandle.h:
 374         (WebCore::SocketStreamHandle::status):
 375         (SocketStreamHandle):
 376         * platform/network/blackberry/SocketStreamHandleBlackBerry.cpp:
 377         (WebCore::SocketStreamHandle::notifyStatusReceived):
 378         (WebCore::SocketStreamHandle::notifyClose):
 379
 380 2012-07-13  Vsevolod Vlasov  <vsevik@chromium.org>
 381
 382         Web Inspector: Remove uiSourceCode from Resource.
 383         https://bugs.webkit.org/show_bug.cgi?id=91201
 384
 385         Reviewed by Pavel Feldman.
 386
 387         Removed Resource._uiSourceCode field as it is not used anymore.
 388
 389         * inspector/front-end/Resource.js:
 390         (WebInspector.Resource.prototype.isHidden):
 391         * inspector/front-end/UISourceCode.js:
 392         (WebInspector.UISourceCode):
 393
 394 2012-07-13  Sheriff Bot  <webkit.review.bot@gmail.com>
 395
 396         Unreviewed, rolling out r122450 and r122580.
 397         http://trac.webkit.org/changeset/122450
 398         http://trac.webkit.org/changeset/122580
 399         https://bugs.webkit.org/show_bug.cgi?id=91263
 400
 401         Caused multiple regressions on ClusterFuzz (Requested by
 402         inferno-sec on #webkit).
 403
 404         * bindings/js/ScriptWrappable.h:
 405         (WebCore::ScriptWrappable::reportMemoryUsage):
 406         * bindings/v8/DOMDataStore.cpp:
 407         (WebCore::DOMDataStore::reportMemoryUsage):
 408         * bindings/v8/IntrusiveDOMWrapperMap.h:
 409         (WebCore::ChunkedTable::reportMemoryUsage):
 410         * bindings/v8/ScriptWrappable.h:
 411         (WebCore::ScriptWrappable::reportMemoryUsage):
 412         * bindings/v8/V8Binding.cpp:
 413         (WebCore::V8BindingPerIsolateData::reportMemoryUsage):
 414         (WebCore::StringCache::reportMemoryUsage):
 415         * bindings/v8/V8DOMMap.h:
 416         * css/PropertySetCSSStyleDeclaration.cpp:
 417         (WebCore::InlineCSSStyleDeclaration::ensureMutablePropertySet):
 418         * css/StylePropertySet.cpp:
 419         * css/StylePropertySet.h:
 420         (WebCore::StylePropertySet::reportMemoryUsage):
 421         * dom/CharacterData.cpp:
 422         (WebCore::CharacterData::reportMemoryUsage):
 423         * dom/ContainerNode.h:
 424         (WebCore::ContainerNode::reportMemoryUsage):
 425         * dom/Document.cpp:
 426         (WebCore::Document::reportMemoryUsage):
 427         * dom/Element.cpp:
 428         (WebCore::Element::detachAttribute):
 429         (WebCore::Element::removeAttribute):
 430         (WebCore::Element::attributes):
 431         (WebCore::Element::setAttributeInternal):
 432         (WebCore::Element::parserSetAttributes):
 433         (WebCore::Element::hasEquivalentAttributes):
 434         (WebCore::Element::createAttributeData):
 435         (WebCore):
 436         (WebCore::Element::setAttributeNode):
 437         (WebCore::Element::removeAttributeNode):
 438         (WebCore::Element::getAttributeNode):
 439         (WebCore::Element::getAttributeNodeNS):
 440         (WebCore::Element::hasAttribute):
 441         (WebCore::Element::hasAttributeNS):
 442         (WebCore::Element::normalizeAttributes):
 443         (WebCore::Element::cloneAttributesFromElement):
 444         * dom/Element.h:
 445         (WebCore::Element::attributeData):
 446         (Element):
 447         (WebCore::Element::reportMemoryUsage):
 448         (WebCore::Element::ensureAttributeData):
 449         (WebCore::Element::updatedAttributeData):
 450         (WebCore::Element::ensureUpdatedAttributeData):
 451         * dom/ElementAttributeData.cpp:
 452         (WebCore::ElementAttributeData::attrIfExists):
 453         (WebCore::ElementAttributeData::ensureAttr):
 454         (WebCore::ElementAttributeData::setAttr):
 455         (WebCore::ElementAttributeData::removeAttr):
 456         (WebCore::ElementAttributeData::setClass):
 457         (WebCore):
 458         (WebCore::ElementAttributeData::ensureInlineStyle):
 459         (WebCore::ElementAttributeData::ensureMutableInlineStyle):
 460         (WebCore::ElementAttributeData::destroyInlineStyle):
 461         (WebCore::ElementAttributeData::addAttribute):
 462         (WebCore::ElementAttributeData::removeAttribute):
 463         (WebCore::ElementAttributeData::isEquivalent):
 464         (WebCore::ElementAttributeData::detachAttrObjectsFromElement):
 465         (WebCore::ElementAttributeData::getAttributeItemIndexSlowCase):
 466         (WebCore::ElementAttributeData::cloneDataFrom):
 467         (WebCore::ElementAttributeData::clearAttributes):
 468         (WebCore::ElementAttributeData::replaceAttribute):
 469         (WebCore::ElementAttributeData::getAttributeNode):
 470         * dom/ElementAttributeData.h:
 471         (WebCore::ElementAttributeData::create):
 472         (ElementAttributeData):
 473         (WebCore::ElementAttributeData::setIdForStyleResolution):
 474         (WebCore::ElementAttributeData::setAttributeStyle):
 475         (WebCore::ElementAttributeData::length):
 476         (WebCore::ElementAttributeData::isEmpty):
 477         (WebCore::ElementAttributeData::attributeItem):
 478         (WebCore::ElementAttributeData::getAttributeItem):
 479         (WebCore::ElementAttributeData::reportMemoryUsage):
 480         (WebCore::ElementAttributeData::ElementAttributeData):
 481         (WebCore::ElementAttributeData::attributeVector):
 482         (WebCore::ElementAttributeData::clonedAttributeVector):
 483         (WebCore::ElementAttributeData::removeAttribute):
 484         (WebCore::ElementAttributeData::getAttributeItemIndex):
 485         * dom/MemoryInstrumentation.h:
 486         (MemoryInstrumentation):
 487         (MemoryObjectInfo):
 488         (WebCore::MemoryObjectInfo::reportInstrumentedPointer):
 489         (WebCore::MemoryObjectInfo::reportPointer):
 490         (WebCore::MemoryObjectInfo::reportInstrumentedObject):
 491         (WebCore::MemoryObjectInfo::reportObject):
 492         (WebCore::MemoryObjectInfo::reportObjectInfo):
 493         (WebCore::MemoryObjectInfo::reportHashMap):
 494         (WebCore::MemoryObjectInfo::reportHashSet):
 495         (WebCore::MemoryObjectInfo::reportListHashSet):
 496         (WebCore::MemoryObjectInfo::reportVector):
 497         (WebCore::MemoryObjectInfo::reportString):
 498         (WebCore::MemoryObjectInfo::objectType):
 499         (WebCore::MemoryObjectInfo::objectSize):
 500         (WebCore::MemoryObjectInfo::memoryInstrumentation):
 501         * dom/Node.cpp:
 502         (WebCore::Node::reportMemoryUsage):
 503         * dom/QualifiedName.h:
 504         (WebCore::QualifiedName::QualifiedNameImpl::reportMemoryUsage):
 505         (WebCore::QualifiedName::reportMemoryUsage):
 506         * dom/StyledElement.cpp:
 507         (WebCore::StyledElement::style):
 508         (WebCore::StyledElement::classAttributeChanged):
 509         (WebCore::StyledElement::setInlineStyleProperty):
 510         (WebCore::StyledElement::removeInlineStyleProperty):
 511         (WebCore::StyledElement::addSubresourceAttributeURLs):
 512         * dom/StyledElement.h:
 513         (WebCore::StyledElement::ensureInlineStyle):
 514         * html/parser/HTMLConstructionSite.cpp:
 515         (WebCore::HTMLConstructionSite::createHTMLElementFromSavedElement):
 516         * platform/TreeShared.h:
 517         (WebCore::TreeShared::reportMemoryUsage):
 518         * xml/parser/XMLDocumentParserQt.cpp:
 519         (WebCore::XMLDocumentParser::XMLDocumentParser):
 520
 521 2012-07-13  Huang Dongsung  <luxtella@company100.net>
 522
 523         Remove down-casting to BitmapImage in GraphicsContext::drawImage.
 524         https://bugs.webkit.org/show_bug.cgi?id=90755
 525
 526         Reviewed by Simon Fraser.
 527
 528         Add a BitmapImage draw method which takes RespectImageOrientationEnum enum as
 529         the last argument for CG. Then we can remove the conditional down-casting in
 530         GraphicsContext::drawImage.
 531
 532         This change is needed for parallel image decoders. Because parallel image
 533         decoders use a Bitmap image wrapper class which extends Image (not Bitmap), the
 534         down-casting above causes the loss of RespectImageOrientationEnum which must be
 535         passed to BitmapImage.
 536
 537         No new tests, no behavior change.
 538
 539         * platform/graphics/BitmapImage.cpp:
 540         * platform/graphics/BitmapImage.h:
 541         * platform/graphics/GraphicsContext.cpp:
 542         (WebCore::GraphicsContext::drawImage):
 543         * platform/graphics/Image.cpp:
 544         (WebCore::Image::draw):
 545         (WebCore):
 546         * platform/graphics/Image.h:
 547         (Image):
 548
 549 2012-07-13  Lauro Neto  <lauro.neto@openbossa.org>
 550
 551         Fix QtWebKit build with OpenGLES after GC3D/E3D refactor
 552         https://bugs.webkit.org/show_bug.cgi?id=91156
 553
 554         Reviewed by Noam Rosenthal.
 555
 556         Adds several build fixes.
 557
 558         * platform/graphics/GraphicsContext3D.h:
 559             Use E3DOpenGLES instead of previously removed E3DQt.
 560
 561         * platform/graphics/OpenGLESShims.h:
 562             Enable defines for Qt.
 563
 564         * platform/graphics/opengl/Extensions3DOpenGLES.cpp:
 565         (WebCore::Extensions3DOpenGLES::blitFramebuffer):
 566         (WebCore):
 567         (WebCore::Extensions3DOpenGLES::renderbufferStorageMultisample):
 568         (WebCore::Extensions3DOpenGLES::copyTextureCHROMIUM):
 569             Added pure virtual stubs.
 570
 571         (WebCore::Extensions3DOpenGLES::supportsExtension):
 572             Remove PROC suffix. See bug #91130.
 573
 574         * platform/graphics/opengl/Extensions3DOpenGLES.h:
 575         (Extensions3DOpenGLES):
 576             Added pure virtual stubs.
 577
 578         * platform/graphics/opengl/GraphicsContext3DOpenGLES.cpp:
 579         (WebCore::GraphicsContext3D::reshapeFBOs):
 580             Readded missing function after removed in r122250.
 581
 582         (WebCore):
 583             Use PLATFORM(BLACKBERRY) guard around port-specific include.
 584
 585         * platform/graphics/qt/GraphicsContext3DQt.cpp:
 586             Added USE(OPENGL_ES_2) guard instead of always loading the OpenGL extensions.
 587
 588
 589 2012-07-13  Keishi Hattori  <keishi@webkit.org>
 590
 591         Form of FormAssociatedElement is not updated when id target changes.
 592         https://bugs.webkit.org/show_bug.cgi?id=91042
 593
 594         Reviewed by Kent Tamura.
 595
 596         Test: fast/forms/update-form-attribute-element.html
 597
 598         This patch introduces the IdTargetObserver and IdTargetObserverRegistry class.
 599         They can be used to be notified when the element that an id is pointing to (the id target)
 600         changes.
 601
 602         * CMakeLists.txt: Added IdTargetObserverRegistry.{h,cpp} and IdTargetObserver.{h,cpp}
 603         * GNUmakefile.list.am: Ditto.
 604         * Target.pri: Ditto.
 605         * WebCore.gypi: Ditto.
 606         * WebCore.vcproj/WebCore.vcproj: Ditto.
 607         * WebCore.xcodeproj/project.pbxproj: Ditto.
 608         * dom/DOMAllInOne.cpp:
 609         * dom/IdTargetObserver.cpp: Added. When you want notified of changes to an id target, you should create a new class that inherits this.
 610         (WebCore):
 611         (WebCore::IdTargetObserver::IdTargetObserver):
 612         (WebCore::IdTargetObserver::~IdTargetObserver):
 613         * dom/IdTargetObserver.h: Added.
 614         (WebCore):
 615         (IdTargetObserver):
 616         * dom/IdTargetObserverRegistry.cpp: Added.
 617         (WebCore):
 618         (WebCore::IdTargetObserverRegistry::create):
 619         (WebCore::IdTargetObserverRegistry::addObserver): Register an IdTargetObserver to observe an id target.
 620         (WebCore::IdTargetObserverRegistry::removeObserver): Unregisters an IdTargetObserver from observing.
 621         (WebCore::IdTargetObserverRegistry::notifyObserversInternal):
 622         * dom/IdTargetObserverRegistry.h: Added.
 623         (WebCore):
 624         (IdTargetObserverRegistry):
 625         (WebCore::IdTargetObserverRegistry::IdTargetObserverRegistry):
 626         (WebCore::IdTargetObserverRegistry::notifyObservers): Calls idTargetChanged on all observers for an id. Inlining first part of function for performance.
 627         * dom/TreeScope.cpp:
 628         (WebCore::TreeScope::TreeScope):
 629         (WebCore::TreeScope::addElementById): Calls IdTargetObserverRegistry::notifyObservers because the id target might have changed.
 630         (WebCore::TreeScope::removeElementById): Ditto.
 631         * dom/TreeScope.h:
 632         (WebCore):
 633         (WebCore::TreeScope::idTargetObserverRegistry):
 634         (TreeScope):
 635         * html/FormAssociatedElement.cpp: Observer for id targets defined by the form attribute.
 636         (WebCore::FormAssociatedElement::didMoveToNewDocument):
 637         (WebCore::FormAssociatedElement::insertedInto):
 638         (WebCore::FormAssociatedElement::removedFrom):
 639         (WebCore::FormAssociatedElement::formAttributeChanged):
 640         (WebCore::FormAssociatedElement::resetFormAttributeTargetObserver): Creates and sets up a new FormAttributeTargetObserver.
 641         (WebCore):
 642         (WebCore::FormAssociatedElement::formAttributeTargetChanged):
 643         (WebCore::FormAttributeTargetObserver::create):
 644         (WebCore::FormAttributeTargetObserver::FormAttributeTargetObserver):
 645         (WebCore::FormAttributeTargetObserver::idTargetChanged):
 646         * html/FormAssociatedElement.h:
 647         (FormAssociatedElement):
 648         * html/FormController.cpp:
 649         * html/FormController.h:
 650         (FormController):
 651         * html/HTMLFormElement.cpp:
 652         (WebCore::HTMLFormElement::removedFrom):
 653         (WebCore::HTMLFormElement::formElementIndexWithFormAttribute): Modified to take a range. It
 654         scans the range and returns the index to insert the element in m_associatedElement.
 655         (WebCore::HTMLFormElement::formElementIndex): Modified to only scan the elements in
 656         m_associatedElement that precede and follow the form element.
 657         * html/HTMLFormElement.h:
 658         (HTMLFormElement):
 659
 660 2012-07-13  Gabor Rapcsanyi  <rgabor@webkit.org>
 661
 662         Optimizing blend filter to ARM-NEON with intrinsics
 663         https://bugs.webkit.org/show_bug.cgi?id=90949
 664
 665         Reviewed by Zoltan Herczeg.
 666
 667         The feBlend SVG filter modes can be greatly fasten up with ARM-NEON since
 668         we are able to calculate with 2 pixels (8 channels) at the same time.
 669         The code is written with NEON intrinsics and it doesn't affect the
 670         general - it has the same behaviour as the original algorithm.
 671         With this NEON optimization the calculation is ~4.5 times faster for each mode.
 672
 673         Existing tests cover this issue.
 674
 675         * CMakeLists.txt:
 676         * GNUmakefile.list.am:
 677         * Target.pri:
 678         * WebCore.gypi:
 679         * WebCore.vcproj/WebCore.vcproj:
 680         * WebCore.xcodeproj/project.pbxproj:
 681         * platform/graphics/filters/FEBlend.cpp:
 682         (WebCore::FEBlend::platformApplyGeneric):
 683         (WebCore):
 684         (WebCore::FEBlend::platformApplySoftware):
 685         * platform/graphics/filters/FEBlend.h:
 686         (FEBlend):
 687         * platform/graphics/filters/arm/FEBlendNEON.h: Added.
 688         (WebCore):
 689         (FEBlendUtilitiesNEON):
 690         (WebCore::FEBlendUtilitiesNEON::div255): integer divison with 255
 691         (WebCore::FEBlendUtilitiesNEON::normal): calculate normal mode blending for two pixels
 692         (WebCore::FEBlendUtilitiesNEON::multiply): calculate multiply mode blending for two pixels
 693         (WebCore::FEBlendUtilitiesNEON::screen): calculate screen mode blending for two pixels
 694         (WebCore::FEBlendUtilitiesNEON::darken): calculate darken mode blending for two pixels
 695         (WebCore::FEBlendUtilitiesNEON::lighten): calculate lighten mode blending for two pixels
 696         (WebCore::FEBlend::platformApplyNEON):
 697
 698 2012-07-13  Ilya Tikhonovsky  <loislo@chromium.org>
 699
 700         Web Inspector: native memory instrumentation: extract instrumentation methods into MemoryClassInfo
 701         https://bugs.webkit.org/show_bug.cgi?id=91227
 702
 703         Reviewed by Pavel Feldman.
 704
 705         void Node::reportMemoryUsage(MemoryObjectInfo* memoryObjectInfo) const
 706         {
 707             MemoryClassInfo<Node> info(memoryObjectInfo, this, MemoryInstrumentation::DOM);
 708             info.visitBaseClass<ScriptWrappable>(this);
 709
 710             info.addMember(m_notInstrumentedPointer); // automatically detects poniter/reference
 711             info.addInstrumentedMember(m_next);
 712             info.addHashSet<MemoryInstrumentation::NonClass>(m_aHash);                // NonClass value_type (report only size of internal template structures)
 713             info.addHashSet<MemoryInstrumentation::NotInstrumentedClass>(m_aHashSet); // not instrumented value_type (use sizeof)
 714             info.addHashSet<MemoryInstrumentation::InstrumentedClass>(m_aHashSet);    // instrumented value_type (call visit)
 715         }
 716
 717         The change is covered by existing tests for native memory snapshot.
 718
 719         * bindings/v8/DOMDataStore.cpp:
 720         (WebCore::DOMDataStore::reportMemoryUsage):
 721         * bindings/v8/IntrusiveDOMWrapperMap.h:
 722         (WebCore::ChunkedTable::reportMemoryUsage):
 723         * bindings/v8/ScriptWrappable.h:
 724         (WebCore::ScriptWrappable::reportMemoryUsage):
 725         * bindings/v8/V8Binding.cpp:
 726         (WebCore::V8BindingPerIsolateData::reportMemoryUsage):
 727         (WebCore::StringCache::reportMemoryUsage):
 728         * bindings/v8/V8DOMMap.h:
 729         * css/StylePropertySet.h:
 730         (WebCore::StylePropertySet::reportMemoryUsage):
 731         * dom/CharacterData.cpp:
 732         (WebCore::CharacterData::reportMemoryUsage):
 733         * dom/ContainerNode.h:
 734         (WebCore::ContainerNode::reportMemoryUsage):
 735         * dom/Document.cpp:
 736         (WebCore::Document::reportMemoryUsage):
 737         * dom/Element.h:
 738         (WebCore::Element::reportMemoryUsage):
 739         * dom/ElementAttributeData.h:
 740         (WebCore::ElementAttributeData::reportMemoryUsage):
 741         * dom/MemoryInstrumentation.h:
 742         (MemoryInstrumentation):
 743         (WebCore::MemoryObjectInfo::objectType):
 744         (WebCore::MemoryObjectInfo::objectSize):
 745         (WebCore::MemoryObjectInfo::memoryInstrumentation):
 746         (MemoryObjectInfo):
 747         (WebCore::MemoryObjectInfo::reportObjectInfo):
 748         (WebCore):
 749         (MemoryClassInfo):
 750         (WebCore::MemoryClassInfo::MemoryClassInfo):
 751         (WebCore::MemoryClassInfo::visitBaseClass):
 752         (WebCore::MemoryClassInfo::reportInstrumentedPointer):
 753         (WebCore::MemoryClassInfo::reportInstrumentedObject):
 754         (WebCore::MemoryClassInfo::reportPointer):
 755         (WebCore::MemoryClassInfo::reportObject):
 756         (WebCore::MemoryClassInfo::reportHashMap):
 757         (WebCore::MemoryClassInfo::reportHashSet):
 758         (WebCore::MemoryClassInfo::reportListHashSet):
 759         (WebCore::MemoryClassInfo::reportVector):
 760         (WebCore::MemoryClassInfo::reportString):
 761         * dom/Node.cpp:
 762         (WebCore::Node::reportMemoryUsage):
 763         * dom/QualifiedName.h:
 764         (WebCore::QualifiedName::QualifiedNameImpl::reportMemoryUsage):
 765         (WebCore::QualifiedName::reportMemoryUsage):
 766         * platform/TreeShared.h:
 767         (WebCore::TreeShared::reportMemoryUsage):
 768
 769 2012-07-13  Pavel Feldman  <pfeldman@chromium.org>
 770
 771         Web Inspector: align scope filters
 772         https://bugs.webkit.org/show_bug.cgi?id=91213
 773
 774         Reviewed by Vsevolod Vlasov.
 775
 776         * inspector/front-end/elementsPanel.css:
 777         (.crumbs):
 778         * inspector/front-end/inspector.css:
 779         (.status-bar > div):
 780         (.scope-bar):
 781         (.scope-bar li):
 782         (.scope-bar li.all):
 783         * inspector/front-end/networkLogView.css:
 784
 785 2012-07-13  Peter Rybin  <peter.rybin@gmail.com>
 786
 787         Web Inspector: too many hardcoded strings in InspectorBackendDispatcher.
 788         https://bugs.webkit.org/show_bug.cgi?id=89198
 789
 790         Reviewed by Yury Semikhatsky.
 791
 792         Instead of generating error message string on every call (mostly for nothing),
 793         error message is generated deeper inside the handler and only command name
 794         is passed every time.
 795
 796         * inspector/CodeGeneratorInspector.py:
 797         (Generator.process_command):
 798
 799 2012-07-13  Joshua Netterfield  <jnetterfield@rim.com>
 800
 801         [BlackBerry] Update about:* pages
 802         https://bugs.webkit.org/show_bug.cgi?id=91121
 803
 804         Reviewed by Yong Li.
 805
 806         Update the about:config pages, and improve the aesthetics of the about:build, about:version, about:credits, about:memory, about:config, and similar pages.
 807
 808         No new tests, because there is no new funtionality.
 809
 810         * platform/network/blackberry/NetworkJob.cpp: Update the aesthetics of about:* pages
 811
 812 2012-07-13  Olivier Blin  <olivier.blin@softathome.com>
 813
 814         Fix checking for optional DeviceOrientationEvent.absolute in JSC bindings
 815         https://bugs.webkit.org/show_bug.cgi?id=91225
 816
 817         Reviewed by Steve Block.
 818
 819         This issue comes from r105036
 820
 821         * bindings/js/JSDeviceOrientationEventCustom.cpp:
 822         (WebCore::JSDeviceOrientationEvent::initDeviceOrientationEvent):
 823
 824 2012-07-13  Andrei Bucur  <abucur@adobe.com>
 825         [CSS Regions] Fix build for bug 89000
 826         https://bugs.webkit.org/show_bug.cgi?id=91215
 827
 828         Reviewed by Kentaro Hara.
 829
 830         Remove the unused variable m_state that was a leftover from a previous version of the patch.
 831
 832         Tests: No new tests, build fix.
 833
 834         * dom/WebKitNamedFlow.cpp:
 835         (WebCore::WebKitNamedFlow::WebKitNamedFlow):
 836         * dom/WebKitNamedFlow.h:
 837         (WebKitNamedFlow):
 838
 839 2012-07-13  Kenichi Ishibashi  <bashi@chromium.org>
 840
 841         [Chromium] Fix bugs in HarfBuzzShaper
 842         https://bugs.webkit.org/show_bug.cgi?id=90951
 843
 844         Reviewed by Tony Chang.
 845
 846         The current implementation has following problems:
 847         - Cannot render RTL text if the TextRun is divided into more than two
 848           HarfBuzzRun.
 849         - Script handling in TextRun partitioning is incorrect.
 850         - Inaccurate calculation of selection rect.
 851         - Wrong rendering position when the first glyph of the TextRun have
 852           non-zero offsets in terms of HarfBuzz.
 853
 854         To fix these problems I rewrote HarfBuzzShaper class. Here is the summary:
 855         - Divide the whole range of TextRun first, then shape them in visual
 856           order.
 857         - Divide TextRun in the same way of old-harfbuzz's
 858           hb_utf16_script_run_next().
 859         - Prefer float than int when calculating selection.
 860         - Adjust the drawing point after shaping.
 861
 862         Added tests covers the fix except for the last problem. The last problem will be covered
 863         by fast/text/international/complex-joining-using-gpos.html after chromium linux port switches
 864         to use HarfBuzzShaper.
 865
 866         Tests: fast/text/shaping/shaping-script-order.html
 867                fast/text/shaping/shaping-selection-rect.html
 868
 869         * platform/graphics/harfbuzz/FontHarfBuzz.cpp:
 870         (WebCore::Font::drawComplexText): Adjusts point after shaping.
 871         * platform/graphics/harfbuzz/ng/HarfBuzzShaper.cpp:
 872         (WebCore::HarfBuzzShaper::HarfBuzzRun::HarfBuzzRun):
 873         (WebCore):
 874         (WebCore::HarfBuzzShaper::HarfBuzzRun::applyShapeResult): Added.
 875         (WebCore::HarfBuzzShaper::HarfBuzzRun::setGlyphAndAdvance): Offsets are no longer needed.
 876         (WebCore::HarfBuzzShaper::HarfBuzzRun::xPositionForOffset): Calculates character offset based on advance.
 877         (WebCore::normalizeCharacters): Added.
 878         (WebCore::HarfBuzzShaper::HarfBuzzShaper):
 879         (WebCore::HarfBuzzShaper::~HarfBuzzShaper):
 880         (WebCore::HarfBuzzShaper::shape): Divides TextRun first, then shapes them.
 881         (WebCore::HarfBuzzShaper::adjustStartPoint): Added.
 882         (WebCore::HarfBuzzShaper::collectHarfBuzzRuns): Added.
 883         (WebCore::HarfBuzzShaper::shapeHarfBuzzRuns): Added.
 884         (WebCore::HarfBuzzShaper::setGlyphPositionsForHarfBuzzRun): Followed other changes.
 885         (WebCore::HarfBuzzShaper::selectionRect): Use float for calculating selection.
 886         * platform/graphics/harfbuzz/ng/HarfBuzzShaper.h:
 887         (HarfBuzzShaper): Holds the start index of character. Removed unnecessary variables.
 888         (WebCore::HarfBuzzShaper::HarfBuzzRun::create): Ditto.
 889         (HarfBuzzRun):
 890         (WebCore::HarfBuzzShaper::HarfBuzzRun::fontData): Added.
 891         (WebCore::HarfBuzzShaper::HarfBuzzRun::startIndex): Ditto.
 892         (WebCore::HarfBuzzShaper::HarfBuzzRun::glyphs): Ditto.
 893         (WebCore::HarfBuzzShaper::HarfBuzzRun::advances): Ditto.
 894
 895 2012-07-13  Kentaro Hara  <haraken@chromium.org>
 896
 897         Unreviewed, rolling out r122545.
 898         http://trac.webkit.org/changeset/122545
 899         https://bugs.webkit.org/show_bug.cgi?id=91185
 900
 901         We found that this was a wrong fix
 902
 903         * bindings/scripts/CodeGeneratorJS.pm:
 904         (GenerateImplementation):
 905
 906 2012-07-13  Kentaro Hara  <haraken@chromium.org>
 907
 908         Unreviewed, rolling out r122553.
 909         http://trac.webkit.org/changeset/122553
 910         https://bugs.webkit.org/show_bug.cgi?id=91198
 911
 912         We found that this was a wrong fix
 913
 914         * bindings/scripts/CodeGeneratorJS.pm:
 915         (GenerateImplementation):
 916
 917 2012-07-13  Kent Tamura  <tkent@chromium.org>
 918
 919         Change the timing of form state restore
 920         https://bugs.webkit.org/show_bug.cgi?id=89962
 921
 922         Reviewed by Hajime Morita.
 923
 924         For a preparation to fix a form identification problem (Bug 91209),
 925         restore controls states when the parsing of their owner forms is
 926         completed. For controls without owners, their states are restored when
 927         their parsing is completed as ever.
 928
 929         No new tests. This doesn't change observable behavior.
 930
 931         * html/FormController.cpp:
 932         (WebCore::ownerFormForState):
 933         Added. This code was used in formKey(), and restoreControlState*() use it.
 934         (WebCore::FormKeyGenerator::formKey): Use ownerFormForState(). No behavior change.
 935         (WebCore::FormController::restoreControlStateFor):
 936         Moved some code from HTMLFormControlElementWithState::finishParsingChildren().
 937         The difference is we don't resotre state if this control is owned by a form.
 938         (WebCore::FormController::restoreControlStateIn):
 939         Restore states of associated controls. This is called from
 940         finishParsingChildren() for <form>.
 941         * html/FormController.h:
 942         (FormController):
 943         - Declare restoreControlStateFor() and restoreControlStateIn().
 944         - Make takeStateForFormElement() private.
 945
 946         * html/FormAssociatedElement.cpp:
 947         (WebCore::FormAssociatedElement::isFormControlElementWithState):
 948         Added. The default implementation returns false.
 949         * html/FormAssociatedElement.h:
 950         (FormAssociatedElement):
 951         Added isFormControlElementWithState() for FormController::restoreControlStateIn().
 952         * html/HTMLFormControlElementWithState.cpp:
 953         (WebCore::HTMLFormControlElementWithState::finishParsingChildren):
 954         Some code was moved to FormController:restoreControlStateFor().
 955         (WebCore::HTMLFormControlElementWithState::isFormControlElementWithState):
 956         Added. Returns true.
 957         * html/HTMLFormControlElementWithState.h:
 958         (HTMLFormControlElementWithState): Declare isFormControlElementWithState().
 959         * html/HTMLFormElement.cpp:
 960         (WebCore::HTMLFormElement::finishParsingChildren):
 961         Added. Calls FormController::restoreControlStateIn().
 962         * html/HTMLFormElement.h:
 963         (HTMLFormElement): Declare finishParsingChildren().
 964
 965 2012-07-13  Kent Tamura  <tkent@chromium.org>
 966
 967         Make calendar pickers testable
 968         https://bugs.webkit.org/show_bug.cgi?id=84827
 969
 970         Reviewed by Hajime Morita.
 971
 972         WebCore:
 973         - Add PagePopupDriver, an interface to open/close a PagePopup.
 974         - Add setPagePopupDriver() to ChromeClient in order to inject a
 975         PagePopupDriver for testing.
 976
 977         Internals:
 978         Add MockPagePopupDriver, which creates a MockPagePopup, which creates a
 979         normal <iframe> in the top-level document, and load the popup content on
 980         it. MockPagePopupDriver is enabled by
 981         internals.settings.setEnableMockPagePopup(true).
 982
 983         Test: fast/forms/date/calendar-picker-appearance.html
 984
 985         * WebCore.gypi: Add new files.
 986         * WebCore.xcodeproj/project.pbxproj: Add files to make this buildable.
 987         * page/ChromeClient.h:
 988         (ChromeClient): Add function for PagePopupDriver.
 989         * loader/EmptyClients.h: Add empty implementations for PagePopupDriver functions.
 990         * page/PagePopupDriver.h: Added.
 991         (PagePopupDriver):
 992         (WebCore::PagePopupDriver::~PagePopupDriver):
 993
 994         * testing/InternalSettings.cpp:
 995         (WebCore::InternalSettings::Backup::restoreTo): Reset the mock PagePopupDriver.
 996         (WebCore::InternalSettings::setEnableMockPagePopup):
 997         Register MockPagePopupDriver to ChromeClient.
 998         * testing/InternalSettings.h:
 999         (InternalSettings): Declare setEnableMockPagePopup().
1000         * testing/InternalSettings.idl: ditto.
1001         * testing/MockPagePopupDriver.cpp: Added.
1002         (MockPagePopup): Pseudo PagePopup implementation with the standard <iframe>.
1003         (WebCore::MockPagePopupDriver::MockPagePopupDriver): Added.
1004         (WebCore::MockPagePopupDriver::create): Added.
1005         (WebCore::MockPagePopupDriver::~MockPagePopupDriver): Added.
1006         (WebCore::MockPagePopupDriver::openPagePopup):
1007         Added. An override of PagePopupDriver function. This creates a MockPagePopup.
1008         (WebCore::MockPagePopupDriver::closePagePopup):
1009         Added. An override of PagePopupDriver function. This deletes the MockPagePopup.
1010         * testing/MockPagePopupDriver.h:
1011         (MockPagePopupDriver): Added.
1012
1013         * testing/v8/WebCoreTestSupport.cpp:
1014         (WebCoreTestSupport::injectPagePopupController):
1015         Added. Production code uses per-Page context feature framework. However
1016         MockPagePopup uses the same page as the host page. So we can't use the
1017         framework and need to inject window.pagePopupController manually.
1018         * testing/v8/WebCoreTestSupport.h:
1019         (WebCoreTestSupport): Add injectPagePopupController().
1020
1021 2012-07-12  Pavel Feldman  <pfeldman@chromium.org>
1022
1023         Web Inspector: mute the native looks of the selects in the console.
1024         https://bugs.webkit.org/show_bug.cgi?id=91120
1025
1026         Reviewed by Vsevolod Vlasov.
1027
1028         This is necessary for Mac now that we don't use border images for select.
1029
1030         * inspector/front-end/ConsoleView.js:
1031         (WebInspector.ConsoleView.prototype.get statusBarItems):
1032         (WebInspector.ConsoleView.prototype.addContext):
1033         (WebInspector.ConsoleView.prototype.removeContext):
1034         (WebInspector.ConsoleView.prototype._updateIsolatedWorldSelector):
1035         (WebInspector.ConsoleView.prototype._appendIsolatedContextOption):
1036         (WebInspector.ConsoleView.prototype._currentEvaluationContext):
1037         (WebInspector.ConsoleView.prototype._currentIsolatedContextId):
1038         * inspector/front-end/StatusBarButton.js:
1039         (WebInspector.StatusBarComboBox):
1040         (WebInspector.StatusBarComboBox.prototype.addOption):
1041         (WebInspector.StatusBarComboBox.prototype.removeOption):
1042         (WebInspector.StatusBarComboBox.prototype.removeOptions):
1043         (WebInspector.StatusBarComboBox.prototype.selectedOption):
1044         * inspector/front-end/inspector.css:
1045         (.status-bar-select-container):
1046         (select.status-bar-item):
1047         (.console-context):
1048
1049 2012-07-13  Andrei Bucur  <abucur@adobe.com>
1050
1051         [CSS Regions] Fix the lifecycle for the flow objects and their renderers
1052         https://bugs.webkit.org/show_bug.cgi?id=89000
1053
1054         Reviewed by Eric Seidel.
1055
1056         This patch adds the concept of a NamedFlowCollection, owned by the document, that keeps track of
1057         all the named flows that exist in the Document. This collection contains a ListHashSet of weak references to
1058         all the existing NamedFlows in the document. This is not a managed set because the CREATED flows are referenced from the renderer and
1059         the NULL flows are only cached, they should be destructible.
1060         Two named flows are considered to be equal if they have the same name.
1061         I've changed the NamedFlow state to depend on the existence of its renderer. A flow thread that has a renderer will also have a NamedFlow object.
1062         A flow thread without a renderer can have a NamedFlow object, but only in the NULL state. It's possible for a NamedFlow object to jump from the
1063         NULL state to the CREATED state if it was not destroyed (e.g. it was referenced from JS). Keeping track of the NULL state flows that have listeners will be important
1064         so when they go back to the CREATED state, the listeners would still be there.
1065
1066         Link to spec: http://www.w3.org/TR/2012/WD-css3-regions-20120503/
1067
1068         Tests: The old tests have been modified to take into account the new behavior
1069
1070         * CMakeLists.txt:
1071         * GNUmakefile.list.am:
1072         * Target.pri:
1073         * WebCore.gypi:
1074         * WebCore.vcproj/WebCore.vcproj:
1075         * WebCore.xcodeproj/project.pbxproj:
1076         * dom/DOMAllInOne.cpp:
1077         * dom/Document.cpp:
1078         (WebCore::Document::~Document):
1079         (WebCore):
1080         (WebCore::Document::webkitGetFlowByName):
1081         (WebCore::Document::namedFlows):
1082         * dom/Document.h:
1083         (WebCore):
1084         (Document):
1085         * dom/WebKitNamedFlow.cpp:
1086         (WebCore::WebKitNamedFlow::WebKitNamedFlow):
1087         (WebCore::WebKitNamedFlow::~WebKitNamedFlow):
1088         (WebCore::WebKitNamedFlow::create):
1089         (WebCore):
1090         (WebCore::WebKitNamedFlow::name):
1091         (WebCore::WebKitNamedFlow::overset):
1092         (WebCore::nodeInFlowThread):
1093         (WebCore::WebKitNamedFlow::getRegionsByContentNode):
1094         (WebCore::WebKitNamedFlow::getContent):
1095         (WebCore::WebKitNamedFlow::setRenderer):
1096         * dom/WebKitNamedFlow.h:
1097         (WebCore):
1098         (WebKitNamedFlow):
1099         (WebCore::WebKitNamedFlow::getFlowState):
1100         (WebCore::WebKitNamedFlow::switchFlowState):
1101         * dom/WebKitNamedFlowCollection.cpp: Added.
1102         (WebCore):
1103         (WebCore::WebKitNamedFlowCollection::WebKitNamedFlowCollection):
1104         (WebCore::WebKitNamedFlowCollection::length): An O(1) operation
1105         (WebCore::WebKitNamedFlowCollection::item): An O(N) operation
1106         (WebCore::WebKitNamedFlowCollection::flowByName): An O(1) operation
1107         (WebCore::WebKitNamedFlowCollection::ensureNamedFlowInCreatedState): An O(1) operation
1108         (WebCore::WebKitNamedFlowCollection::moveNamedFlowToNullState): An O(1) operation
1109         (WebCore::WebKitNamedFlowCollection::discardNamedFlow): An O(1) operation
1110         (WebCore::WebKitNamedFlowCollection::documentDestroyed):
1111         (WebCore::WebKitNamedFlowCollection::NamedFlowHashFunctions::hash):
1112         (WebCore::WebKitNamedFlowCollection::NamedFlowHashFunctions::equal):
1113         (WebKitNamedFlowCollection::NamedFlowHashFunctions):
1114         (WebCore::WebKitNamedFlowCollection::NamedFlowHashTranslator::hash):
1115         (WebCore::WebKitNamedFlowCollection::NamedFlowHashTranslator::equal):
1116         * dom/WebKitNamedFlowCollection.h: Copied from Source/WebCore/dom/WebKitNamedFlow.h.
1117         (WebCore):
1118         (WebKitNamedFlowCollection):
1119         (WebCore::WebKitNamedFlowCollection::create):
1120         (WebCore::WebKitNamedFlowCollection::document):
1121         * rendering/FlowThreadController.cpp:
1122         (WebCore::FlowThreadController::ensureRenderFlowThreadWithName):
1123         (WebCore::FlowThreadController::removeFlowThread):
1124         (WebCore):
1125         * rendering/FlowThreadController.h:
1126         (FlowThreadController):
1127         * rendering/RenderNamedFlowThread.cpp:
1128         (WebCore::RenderNamedFlowThread::RenderNamedFlowThread):
1129         (WebCore):
1130         (WebCore::RenderNamedFlowThread::~RenderNamedFlowThread):
1131         (WebCore::RenderNamedFlowThread::removeRegionFromThread):
1132         (WebCore::RenderNamedFlowThread::unregisterNamedFlowContentNode):
1133         (WebCore::RenderNamedFlowThread::flowThreadName):
1134         (WebCore::RenderNamedFlowThread::willBeDestroyed):
1135         * rendering/RenderNamedFlowThread.h:
1136         (RenderNamedFlowThread):
1137         (WebCore::RenderNamedFlowThread::contentNodes):
1138         (WebCore::RenderNamedFlowThread::canBeDestroyed):
1139
1140 2012-07-13  Vineet Chaudhary  <rgf748@motorola.com>
1141
1142         [V8Bindings] Implement generalised method to validates that the passed object is a sequence type.
1143         https://bugs.webkit.org/show_bug.cgi?id=91056
1144
1145         Reviewed by Kentaro Hara.
1146
1147         Currently the V8 implementation validates that the passed object is a sequence type only
1148         for MessagePort in V8Utilities::extractTransferables().
1149         There should be generalised method for other types too.
1150         Spec URL: http://www.w3.org/TR/2012/WD-WebIDL-20120207/#es-sequence
1151
1152         No new test, Just refactoring. There should be no behavioral changes.
1153
1154         * bindings/v8/V8Binding.h:
1155         (WebCore::toV8Sequence): Added implementation of toV8Sequence().
1156
1157 2012-07-13  Zeno Albisser  <zeno@webkit.org>
1158
1159         [Qt][WK2] Implement GraphicsSurface for Linux/GLX.
1160         https://bugs.webkit.org/show_bug.cgi?id=90881
1161
1162         Add a GLX based GraphicsSurface implementation for Linux.
1163         Native X windows are being used for exchanging textures
1164         with the UIProcess.
1165
1166         Reviewed by Noam Rosenthal.
1167
1168         * Target.pri:
1169         * WebCore.pri:
1170         * platform/graphics/surfaces/GraphicsSurface.cpp:
1171         (WebCore::GraphicsSurface::create):
1172             Move creating GraphicsSurface instance into
1173             platformCreate/platformImport functions to allow
1174             platform specific creation based on the provided flags.
1175         (WebCore::GraphicsSurface::GraphicsSurface):
1176         (WebCore::GraphicsSurface::~GraphicsSurface):
1177             Call platformDestroy when destroying a GraphicsSurface.
1178         (WebCore):
1179         * platform/graphics/surfaces/GraphicsSurface.h:
1180             Make platformCreate/platformImport functions static
1181             to be able to call these from the static create function.
1182             Add Destructor prototype and add GraphicsSurfacePrivate member.
1183         (WebCore):
1184         (GraphicsSurface):
1185         * platform/graphics/surfaces/mac/GraphicsSurfaceMac.cpp:
1186         (WebCore):
1187         (WebCore::GraphicsSurface::platformCreate):
1188         (WebCore::GraphicsSurface::platformImport):
1189             Insert creation of GraphicsSurface instance.
1190             This allows having a platform specific creation mechanism
1191             for GLX.
1192         * platform/graphics/surfaces/qt/GraphicsSurfaceGLX.cpp: Added.
1193         (WebCore):
1194         (OffScreenRootWindow):
1195         (WebCore::OffScreenRootWindow::OffScreenRootWindow):
1196         (WebCore::OffScreenRootWindow::get):
1197         (WebCore::OffScreenRootWindow::~OffScreenRootWindow):
1198             Add an OffScreenRootWindow singelton that is being used
1199             as a parent for all native offscreen windows.
1200         (GraphicsSurfacePrivate):
1201             This class is used to manage all the X related resources
1202             such as opening a display or creating XPixmaps etc.
1203         (WebCore::GraphicsSurfacePrivate::GraphicsSurfacePrivate):
1204             Open a connection to the X server and create a
1205             QOpenGLContext that can be used to resolve GL functions.
1206         (WebCore::GraphicsSurfacePrivate::~GraphicsSurfacePrivate):
1207             Properly cleanup and release all the X resources again.
1208         (WebCore::GraphicsSurfacePrivate::createSurface):
1209             Create a surface that is placed off screen and can be
1210             used as a rendering target by the WebProcess.
1211         (WebCore::GraphicsSurfacePrivate::createPixmap):
1212             Create a GLXPixmap from the XWindow that was previously
1213             redirected by the WebProcess. This GLXPixmap can then be
1214             bound to a texture and being painted on screen by the
1215             UIProcess.
1216         (WebCore::GraphicsSurfacePrivate::makeCurrent):
1217         (WebCore::GraphicsSurfacePrivate::swapBuffers):
1218         (WebCore::GraphicsSurfacePrivate::display):
1219         (WebCore::GraphicsSurfacePrivate::glxPixmap):
1220         (WebCore::GraphicsSurfacePrivate::size):
1221         (WebCore::GraphicsSurfacePrivate::glContext):
1222         (WebCore::resolveGLMethods):
1223             Initialize all the function pointers for the GL functions used.
1224         (WebCore::GraphicsSurface::platformExport):
1225         (WebCore::GraphicsSurface::platformGetTextureID):
1226             Bind the GLXPixmap to a texture.
1227         (WebCore::GraphicsSurface::platformCopyToGLTexture):
1228             Not beeing implemented for GLX.
1229         (WebCore::GraphicsSurface::platformCopyFromFramebuffer):
1230             Blit origin fbo onto the GraphicsSurface's backing.
1231         (WebCore::GraphicsSurface::platformCreate):
1232         (WebCore::GraphicsSurface::platformImport):
1233         (WebCore::GraphicsSurface::platformLock):
1234         (WebCore::GraphicsSurface::platformUnlock):
1235         (WebCore::GraphicsSurface::platformDestroy):
1236
1237 2012-07-13  Dongwoo Im  <dw.im@samsung.com>
1238
1239         CodeGeneratorJS.pm : SetterExpression should use 'push' rather than 'unshift'
1240         https://bugs.webkit.org/show_bug.cgi?id=91198
1241
1242         Reviewed by Kentaro Hara.
1243
1244         'SetterExpression' should use 'push' to make arguments, rather than 'unshift'.
1245
1246         No new tests. Covered by existing tests.
1247
1248         * bindings/scripts/CodeGeneratorJS.pm:
1249         (GenerateImplementation):
1250
1251 2012-07-13  Yoshifumi Inoue  <yosin@chromium.org>
1252
1253         REGRESSION(r119948): [Forms] Spin button Up/Down actions make value to zero for input type "number" when step mismatched
1254         https://bugs.webkit.org/show_bug.cgi?id=91197
1255
1256         Reviewed by Kent Tamura.
1257
1258         This patch fixes implementation of Decimal::ceiling() and floor().
1259         They return wrong value for small fractional numbers.
1260
1261         The bug is occurred when:
1262           - Step-able input type, e.g. number, date, datetime, and so on.
1263           - Current value is step mismatched
1264           - Current value is smaller than step
1265           - Step up/down by spin button
1266         because spin button up/down actions are implemented in InputType::setpUpFromRenderer
1267         which calls Decimal::ceiling() and floor() for step mismatched case.
1268
1269         Tests: fast/forms/number/number-stepup-stepdown-from-renderer.html: Added test cases
1270                WebKit/chromium/tests/DecimalTest.cpp: Added test cases
1271
1272         * platform/Decimal.cpp:
1273         (WebCore::Decimal::ceiling): Changed to return 1 for positive small fractional number.
1274         (WebCore::Decimal::floor): Changed to return -1 for negative small fractional number.
1275
1276 2012-07-13  Dominic Mazzoni  <dmazzoni@google.com>
1277
1278         Should be possible to focus elements within canvas fallback content
1279         https://bugs.webkit.org/show_bug.cgi?id=87898
1280
1281         Reviewed by Chris Fleizach.
1282
1283         Patches isFocusable in dom/Node.cpp and html/HTMLFormControlElement.cpp
1284         to make elements focusable if they're a descendent of a canvas element
1285         if they would otherwise have been focusable but just didn't have
1286         a renderer. Adds a bit to ElementRareData to efficiently keep track
1287         of elements in a canvas subtree.
1288
1289         Test: fast/canvas/fallback-content.html
1290
1291         * dom/Element.cpp:
1292         (WebCore::Element::attach):
1293         (WebCore::Element::detach):
1294         (WebCore::Element::setIsInCanvasSubtree):
1295         (WebCore):
1296         (WebCore::Element::isInCanvasSubtree):
1297         * dom/Element.h:
1298         (Element):
1299         * dom/ElementRareData.h:
1300         (ElementRareData):
1301         (WebCore::ElementRareData::ElementRareData):
1302         * dom/Node.cpp:
1303         (WebCore::Node::isFocusable):
1304         * html/HTMLCanvasElement.cpp:
1305         (WebCore::HTMLCanvasElement::attach):
1306         (WebCore):
1307         * html/HTMLCanvasElement.h:
1308         (HTMLCanvasElement):
1309         * html/HTMLFormControlElement.cpp:
1310         (WebCore::HTMLFormControlElement::isFocusable):
1311
1312 2012-07-12  Carlos Garcia Campos  <cgarcia@igalia.com>
1313
1314         [GTK] Add API to get HTTPS status to WebKit2 GTK+
1315         https://bugs.webkit.org/show_bug.cgi?id=91100
1316
1317         Reviewed by Martin Robinson.
1318
1319         * platform/network/soup/ResourceResponse.h:
1320         (WebCore::ResourceResponse::soupMessageCertificate): Return the
1321         certificate.
1322         (WebCore::ResourceResponse::setSoupMessageCertificate): Set a
1323         certificate.
1324         (WebCore::ResourceResponse::soupMessageTLSErrors): Return the TLS
1325         errors.
1326         (WebCore::ResourceResponse::setSoupMessageTLSErrors): Set TLS
1327         errors.
1328         * platform/network/soup/ResourceResponseSoup.cpp:
1329         (WebCore::ResourceResponse::toSoupMessage): Set the certificate
1330         and TLS errors to the newly created SoupMessage.
1331         (WebCore::ResourceResponse::updateFromSoupMessage): Get the
1332         certificate and TLS errors from the SoupMessage.
1333
1334 2012-07-13  Ryosuke Niwa  <rniwa@webkit.org>
1335
1336         RadioNodeList is not updated upon input type change
1337         https://bugs.webkit.org/show_bug.cgi?id=91178
1338
1339         Reviewed by Alexey Proskuryakov.
1340
1341         Invalidate the radio node lists when type content attribute changes since it excludes
1342         image type input elements.
1343
1344         Test: fast/forms/radionodelist-image-type.html
1345
1346         * dom/Document.cpp:
1347         (WebCore::shouldInvalidateNodeListForType):
1348         * dom/Document.h: Renamed InvalidateOnIdNameForAttrChange to InvalidateOnFormAttrChange
1349         since listing all attribute name isn't useful at this point.
1350         * html/RadioNodeList.cpp:
1351         (WebCore::RadioNodeList::RadioNodeList):
1352
1353 2012-07-12  Dongwoo Im  <dw.im@samsung.com>
1354
1355         CodeGeneratorJS.pm need to handle the attribute which has "CallWith=ScriptExecutionContext" option.
1356         https://bugs.webkit.org/show_bug.cgi?id=91185
1357
1358         Reviewed by Kentaro Hara.
1359
1360         When an attribute has "CallWith=ScriptExecutionContext" option, 'ScriptExecutionContext*' parameter should be the last parameter.
1361
1362         No new tests. Covered by existing tests.
1363
1364         * bindings/scripts/CodeGeneratorJS.pm:
1365         (GenerateImplementation):
1366
1367 2012-07-12  Yoshifumi Inoue  <yosin@chromium.org>
1368
1369         REGRESSION(r117738): [Forms] stepMismatch for input type "time" with large step value always return false.
1370         https://bugs.webkit.org/show_bug.cgi?id=91062
1371
1372         Reviewed by Kent Tamura.
1373
1374         This patch changes value of StepRange::acceptableError to zero for
1375         integer restricted step value.
1376
1377         No new tests, existing test covers (fast/forms/time/ValidityState-stepMismatch-time.html) this change although they are listed in TestExpectation file.
1378
1379         * html/StepRange.cpp:
1380         (WebCore::StepRange::acceptableError): Changed to return 0 if step value is restricted to be an integer.
1381
1382 2012-07-12  Dan Bernstein  <mitz@apple.com>
1383
1384         Pass an option flag to CFStringGetHyphenationLocationBeforeIndex() that tells it to not
1385         terminate the search at the last word boundary before the given index.
1386
1387         Reviewed by Adele Peterson.
1388
1389         No test, because the current version of Core Foundation ignores the options parameter.
1390
1391         * platform/text/cf/HyphenationCF.cpp:
1392         (WebCore::lastHyphenLocation): Changed the options parameter from 0 to 1.
1393
1394 2012-07-12  Eric Seidel  <eric@webkit.org>
1395
1396         Incorrect behaviour calling Range setStart or setEnd with boundary in different document
1397         https://bugs.webkit.org/show_bug.cgi?id=42517
1398
1399         Reviewed by Ojan Vafai.
1400
1401         Added a new static inline "checkForDifferentRootContainer" to share some code
1402         and made setStart/setEnd do the right thing in the x-document case.  I removed
1403         the bogus checks in set*After/set*Before functions, and since they just call
1404         through to setStart/setEnd, they also now do the right thing.
1405
1406         Test: fast/dom/Range/set-wrong-document-err.html
1407
1408         * dom/Range.cpp:
1409         (WebCore::checkForDifferentRootContainer):
1410         (WebCore):
1411         (WebCore::Range::setStart):
1412         (WebCore::Range::setEnd):
1413         (WebCore::Range::setStartAfter):
1414         (WebCore::Range::setEndBefore):
1415         (WebCore::Range::setEndAfter):
1416         (WebCore::Range::setStartBefore):
1417
1418 2012-07-12  Erik Arvidsson  <arv@chromium.org>
1419
1420         [V8] Simplify CodeGeneratorV8 since V8OnProto is only true for DOMWindow
1421         https://bugs.webkit.org/show_bug.cgi?id=91165
1422
1423         Reviewed by Nate Chapin.
1424
1425         The old code was dead code since V8OnProto only ever gets set to 1 for DOMWindow.
1426
1427         No new tests. No change in functionality.
1428
1429         * bindings/scripts/CodeGeneratorV8.pm:
1430         (GenerateNormalAttrSetter):
1431
1432 2012-07-12  Ryosuke Niwa  <rniwa@webkit.org>
1433
1434         Build fix. Initialize unused const member variables to make compilers happy.
1435
1436         * dom/DynamicNodeList.h:
1437         (WebCore::DynamicNodeListCacheBase::DynamicNodeListCacheBase):
1438
1439 2012-07-12  Konrad Piascik  <kpiascik@rim.com>
1440
1441         Web Inspector: Geolocation override
1442         https://bugs.webkit.org/show_bug.cgi?id=89365
1443
1444         Reviewed by Pavel Feldman.
1445
1446         Reverted original patch.
1447         Change has not been reviewed by the right people. It declares
1448         public protocol methods and is wrong in several ways.
1449
1450         * Modules/geolocation/GeolocationController.cpp:
1451         (WebCore::GeolocationController::GeolocationController):
1452         (WebCore::GeolocationController::positionChanged):
1453         * Modules/geolocation/GeolocationController.h:
1454         (GeolocationController):
1455         * inspector/Inspector.json:
1456         * inspector/InspectorInstrumentation.cpp:
1457         (WebCore):
1458         * inspector/InspectorInstrumentation.h:
1459         (WebCore):
1460         (InspectorInstrumentation):
1461         * inspector/InspectorPageAgent.cpp:
1462         (WebCore::InspectorPageAgent::InspectorPageAgent):
1463         * inspector/InspectorPageAgent.h:
1464         * inspector/front-end/Settings.js:
1465         * inspector/front-end/SettingsScreen.js:
1466         (WebInspector.UserAgentSettingsTab):
1467         (WebInspector.UserAgentSettingsTab.prototype._createDeviceMetricsElement):
1468         * inspector/front-end/UserAgentSupport.js:
1469
1470 2012-07-12  Ryosuke Niwa  <rniwa@webkit.org>
1471
1472         Move m_type and m_hasNameCache from HTMLCollectionCacheBase to DynamicNodeListCacheBase for better bit packing
1473         https://bugs.webkit.org/show_bug.cgi?id=91164
1474
1475         Reviewed by Anders Carlsson.
1476
1477         Moved m_type and m_hasNameCache from HTMLCollection and renamed them to m_collectionType and m_isNameCacheValid.
1478
1479         Also renamed shouldIncludeChildren to shouldOnlyIncludeDirectChildren and negated the return value since
1480         all HTMLCollection include children in the collection and the function was meant to tell us whether the collection
1481         should include descendents that are not direct children of base().
1482
1483         In addition, renamed nextNodeOrSibling to nextNode since "or sibling" doesn't seem to add any semantic clarity.
1484
1485         * dom/DynamicNodeList.h:
1486         (WebCore::DynamicNodeListCacheBase::DynamicNodeListCacheBase):
1487         (DynamicNodeListCacheBase):
1488         (WebCore::DynamicNodeListCacheBase::type):
1489         (WebCore::DynamicNodeListCacheBase::hasNameCache):
1490         (WebCore::DynamicNodeListCacheBase::setHasNameCache):
1491         (WebCore::DynamicNodeListCacheBase::clearCache):
1492         * html/HTMLCollection.cpp:
1493         (WebCore::shouldOnlyIncludeDirectChildren):
1494         (WebCore::HTMLCollection::HTMLCollection):
1495         (WebCore::HTMLCollection::isAcceptableElement):
1496         (WebCore::nextNode):
1497         (WebCore::HTMLCollection::itemAfter):
1498         * html/HTMLCollection.h:
1499         (WebCore::HTMLCollectionCacheBase::HTMLCollectionCacheBase):
1500         (HTMLCollectionCacheBase):
1501         (WebCore::HTMLCollectionCacheBase::clearCache):
1502
1503 2012-07-12  Shinya Kawanaka  <shinyak@chromium.org>
1504
1505         Needs callback before AuthorShadowRoot is added.
1506         https://bugs.webkit.org/show_bug.cgi?id=91167
1507
1508         Reviewed by Hajime Morita.
1509
1510         This is a preliminary patch for adding multiple Shadow DOM support for media elements and form elements.
1511         They assume that UserAgentShadowRoot is the oldest, however currently a page author may try to add
1512         AuthorShadowRoot before adding UserAgentShadowRoot.
1513
1514         This patch adds a callback before AuthorShadowRoot is being added, and allow us to add UserAgentShadowRoot
1515         for those elements. See also Bug 77936, Bug 77937, Bug 90532.
1516
1517         * dom/Element.h:
1518         (WebCore::Element::willAddAuthorShadowRoot):
1519         * dom/ElementShadow.cpp:
1520         (WebCore::ElementShadow::addShadowRoot):
1521         * dom/ElementShadow.h:
1522         (ElementShadow):
1523         * dom/ShadowRoot.cpp:
1524         (WebCore::ShadowRoot::create):
1525
1526 2012-07-12  Dana Jansens  <danakj@chromium.org>
1527
1528         [chromium] Remove the RenderPass pointer from RenderPassDrawQuad
1529         https://bugs.webkit.org/show_bug.cgi?id=91023
1530
1531         Reviewed by Adrienne Walker.
1532
1533         Removes the RenderPass pointer and keeps only an integer id in the
1534         quad to refer back to the RenderPass the quad reads from.
1535
1536         Covered by existing tests.
1537
1538         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
1539         (WebCore::CCLayerTreeHostImpl::calculateRenderPasses):
1540         (WebCore::findRenderPassById):
1541         (WebCore):
1542         (WebCore::removeRenderPassesRecursive):
1543         (WebCore::CCLayerTreeHostImpl::CullRenderPassesWithCachedTextures::shouldRemoveRenderPass):
1544         (WebCore::CCLayerTreeHostImpl::CullRenderPassesWithNoQuads::shouldRemoveRenderPass):
1545         (WebCore::CCLayerTreeHostImpl::removeRenderPasses):
1546         (WebCore::CCLayerTreeHostImpl::prepareToDraw):
1547         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.h:
1548         (FrameData):
1549         (CullRenderPassesWithCachedTextures):
1550         (CullRenderPassesWithNoQuads):
1551         (CCLayerTreeHostImpl):
1552         * platform/graphics/chromium/cc/CCRenderPass.cpp:
1553         (WebCore::CCRenderPass::appendQuadsForRenderSurfaceLayer):
1554         * platform/graphics/chromium/cc/CCRenderPass.h:
1555         (WebCore):
1556         * platform/graphics/chromium/cc/CCRenderPassDrawQuad.cpp:
1557         (WebCore::CCRenderPassDrawQuad::create):
1558         (WebCore::CCRenderPassDrawQuad::CCRenderPassDrawQuad):
1559         * platform/graphics/chromium/cc/CCRenderPassDrawQuad.h:
1560         (CCRenderPassDrawQuad):
1561         * platform/graphics/chromium/cc/CCRenderSurface.cpp:
1562         (WebCore::CCRenderSurface::appendQuads):
1563         * platform/graphics/chromium/cc/CCRenderSurface.h:
1564         (CCRenderSurface):
1565
1566 2012-07-12  Elliott Sprehn  <esprehn@gmail.com>
1567
1568         Unneeded tree walking when adding or removing children due to RenderCounter / RenderQuote logic
1569         https://bugs.webkit.org/show_bug.cgi?id=89900
1570
1571         Reviewed by Eric Seidel and Abhishek Arya.
1572
1573         Previously we would walk the all children a renderer whenever adding
1574         or removing a child renderer from its RenderObjectChildList to look for
1575         RenderQuote and RenderCounter instances to update. This patch introduces
1576         a counter in RenderView for the number of RenderQuote and RenderCounter
1577         instances in that document so we can avoid these traversals.
1578
1579         No tests needed since this is just a short circuiting of logic and the existing
1580         tests should cover it.
1581
1582         * rendering/RenderCounter.cpp:
1583         (WebCore::RenderCounter::RenderCounter): Increment instance counter.
1584         (WebCore::RenderCounter::willBeDestroyed): Decrement instance counter.
1585         (WebCore):
1586         (WebCore::RenderCounter::rendererRemovedFromTree): Short circuit when counter is zero.
1587         (WebCore::RenderCounter::rendererSubtreeAttached): Short circuit when counter is zero.
1588         * rendering/RenderCounter.h:
1589         (RenderCounter):
1590         * rendering/RenderObjectChildList.cpp:
1591         (WebCore::RenderObjectChildList::removeChildNode): Short circuit calling into Counter and Quote code when the document is being destroyed.
1592         * rendering/RenderQuote.cpp:
1593         (WebCore::RenderQuote::RenderQuote):
1594         (WebCore::RenderQuote::willBeDestroyed):
1595         (WebCore):
1596         (WebCore::RenderQuote::rendererSubtreeAttached): Increment instance counter.
1597         (WebCore::RenderQuote::rendererRemovedFromTree): Decrement instance counter.
1598         * rendering/RenderQuote.h:
1599         (RenderQuote):
1600         * rendering/RenderView.cpp:
1601         (WebCore::RenderView::RenderView):
1602         * rendering/RenderView.h: Methods for managing the RenderQuote and RenderCounter counts.
1603         (RenderView):
1604         (WebCore::RenderView::addRenderQuote):
1605         (WebCore::RenderView::removeRenderQuote):
1606         (WebCore::RenderView::hasRenderQuotes):
1607         (WebCore::RenderView::addRenderCounter):
1608         (WebCore::RenderView::removeRenderCounter):
1609         (WebCore::RenderView::hasRenderCounters):
1610
1611 2012-07-12  Ryosuke Niwa  <rniwa@webkit.org>
1612
1613         Merge HTMLCollectionWithArrayStorage into HTMLCollection
1614         https://bugs.webkit.org/show_bug.cgi?id=91144
1615
1616         Reviewed by Anders Carlsson.
1617
1618         Merged HTMLCollectionWithArrayStorage::item into HTMLCollection::item and got rid of
1619         HTMLCollectionWithArrayStorage. Also de-virtualized HTMLCollection::item and HTMLCollection::length
1620         and merged itemInArrayAfter and itemAfter.
1621
1622         In addition, improved the algorithm in HTMLCollection::length to take advantage of item cache.
1623         Instead of always computing the length from the beginning, we start the search from the cached item
1624         so that if we're near end of the collection, we avoid significant portion of the node traversal.
1625
1626         Furthermore, made HTMLCollection always call setItemCache with a non-null item and HTMLCollection::item
1627         set the length cache when it reaches the end of the collection to avoid redundant length calculations.
1628
1629         * dom/DynamicNodeList.h:
1630         (WebCore::DynamicNodeListCacheBase::setItemCache): Add a FIXME.
1631         * html/HTMLCollection.cpp:
1632         (WebCore::HTMLCollection::itemAfter): Regular HTMLCollection doesn't have uses elements array so
1633         assert that offsetInArray is always 0.
1634         (WebCore): Removed calcLength.
1635         (WebCore::HTMLCollection::length): Rewritten. The algorithm is as follows:
1636         When there is no item cache, we look for the first item by calling item(0). If item(0) returns null,
1637         then it must have set the length cache so bail out. If the previous step didn't bail out, then
1638         the item cache is valid and not null (see above), so count the number of remaining items in collection
1639         starting at the cached item's offset + 1.
1640         (WebCore::HTMLCollection::item): Avoid calling setItemCache with null. When the first item is null,
1641         set the length cache instead.
1642         (WebCore::HTMLCollection::itemAfterCachedItem): Extracted from HTMLCollectionWithArrayStorage::item.
1643         (WebCore::HTMLCollection::namedItem): Pass in arrayOffset to itemAfter. Note this variable is never
1644         used since only HTMLFormCollection and HTMLPropertiesCollection use array offsets but they override
1645         this function.
1646         (WebCore::HTMLCollection::updateNameCache): Ditto.
1647         * html/HTMLCollection.h:
1648         (HTMLCollection):
1649         (WebCore):
1650         * html/HTMLFormCollection.cpp: Removed calcLength(). Even though this function didn't iterate over
1651         the collection directly, HTMLFormElement::length and HTMLFieldSetElement::length did so we're not
1652         regressing any performance here.
1653         (WebCore::HTMLFormCollection::HTMLFormCollection):
1654         (WebCore::HTMLFormCollection::itemAfter):
1655         * html/HTMLFormCollection.h:
1656         (HTMLFormCollection):
1657         * html/HTMLNameCollection.cpp:
1658         (WebCore::HTMLNameCollection::itemAfter):
1659         * html/HTMLNameCollection.h:
1660         (HTMLNameCollection):
1661         * html/HTMLPropertiesCollection.cpp: Removed calcLength().
1662         (WebCore::HTMLPropertiesCollection::HTMLPropertiesCollection):
1663         (WebCore::HTMLPropertiesCollection::itemAfter):
1664         (WebCore):
1665         * html/HTMLPropertiesCollection.h:
1666         (HTMLPropertiesCollection):
1667         * html/HTMLTableRowsCollection.cpp:
1668         (WebCore::HTMLTableRowsCollection::itemAfter):
1669         * html/HTMLTableRowsCollection.h:
1670         (HTMLTableRowsCollection):
1671
1672 2012-07-12  Pravin D  <pravind.2k4@gmail.com>
1673
1674         Row size/position is wrongly calculated when table having overlapping rowspan cell and colspan cell
1675         https://bugs.webkit.org/show_bug.cgi?id=16811
1676
1677         Reviewed by Julien Chaffraix.
1678
1679         The height of a row is calculated by taking the max height of the cells contained in it. When a row contains
1680         a rowSpan cell and if this row is not the last row of the cell, then its height is max height of other non
1681         rowSpan cells. If the row is the last row of the rowSpan cell, then using the contraint laid by CSS2.1 spec
1682         "For a rowSpan cell, the sum of the row heights involved must be great enough to encompass the cell spanning the rows",
1683         the last remaining height of the rowSpan(cell height minus heights of other involved rows) is taken into consideration
1684         while calculating the height of this row.
1685         Currently when calculating the height of the row we are only using the height of the primary cell at position (row, col).
1686         However when a row has colSpan cell and rowSpan, they might overlap. In such a sitution as only the primary cells height
1687         is considered, the height of the row will be calculated worngly if the other overlapping cell has greater height.
1688         Thus all the overlapping cell at position (row, col) must be considered while calculating the height of a row.
1689
1690         Test: fast/table/last-cell-of-rowspan-overlapping-colspan-cell.html
1691
1692         * rendering/RenderTableSection.cpp:
1693         (WebCore::RenderTableSection::calcRowLogicalHeight):
1694         Fixed function to use all the overlapping cells at position(row, col) to calculate the height/position of rows.
1695
1696 2012-07-12  Joshua Bell  <jsbell@chromium.org>
1697
1698         IndexedDB: Enable IDBFactory.deleteDatabase() and webkitGetDatabaseNames() in Workers
1699         https://bugs.webkit.org/show_bug.cgi?id=90310
1700
1701         Reviewed by Tony Chang.
1702
1703         Simplify Document vs. Worker logic for IDBFactory::open() and hook up the
1704         other two IDBFactory methods for use by workers as well.
1705
1706         Test: storage/indexeddb/factory-basics-workers.html
1707
1708         * Modules/indexeddb/IDBFactory.cpp:
1709         (isContextValid): Helper function consolidating checks that context is usable.
1710         (getIndexedDBDatabasePath): Helper function for accessing group settings.
1711         (WebCore::IDBFactory::getDatabaseNames): Simplify - just pass context through to back end.
1712         (WebCore::IDBFactory::open): Ditto.
1713         (WebCore::IDBFactory::deleteDatabase): Ditto.
1714         (WebCore::IDBFactory::cmp): Whitespace.
1715         * Modules/indexeddb/IDBFactoryBackendImpl.cpp: Obsolete openFromWorker() removed.
1716         (WebCore::IDBFactoryBackendImpl::getDatabaseNames): Signature updated.
1717         (WebCore::IDBFactoryBackendImpl::deleteDatabase): Signature updated.
1718         (WebCore::IDBFactoryBackendImpl::open): Signature updated.
1719         * Modules/indexeddb/IDBFactoryBackendImpl.h:
1720         (IDBFactoryBackendImpl):
1721         * Modules/indexeddb/IDBFactoryBackendInterface.h: Interface methods now take both SecurityOrigin
1722         and ScriptExecutionContext, but not Frame. In the proxy, SecurityOrigin is redundant (can be
1723         accessed from context) but on the real back end the context is null (as Frame was previously).
1724         (IDBFactoryBackendInterface):
1725         * inspector/InspectorIndexedDBAgent.cpp:
1726         (WebCore):
1727         (WebCore::InspectorIndexedDBAgent::requestDatabaseNamesForFrame):
1728         (WebCore::InspectorIndexedDBAgent::requestDatabase):
1729         (WebCore::InspectorIndexedDBAgent::requestData):
1730
1731 2012-07-12  Dana Jansens  <danakj@chromium.org>
1732
1733         [chromium] The root layer should not try create a second RenderSurface for itself
1734         https://bugs.webkit.org/show_bug.cgi?id=91124
1735
1736         Reviewed by Adrienne Walker.
1737
1738         Tests: CCLayerTreeHostImplTest.rootLayerDoesntCreateExtraSurface
1739
1740         * platform/graphics/chromium/cc/CCLayerTreeHostCommon.cpp:
1741         (WebCore::subtreeShouldRenderToSeparateSurface):
1742
1743 2012-07-12  Adam Barth  <abarth@webkit.org>
1744
1745         Regression (r122359) Layout Test html5lib/runner.html is failing
1746         https://bugs.webkit.org/show_bug.cgi?id=91047
1747
1748         Reviewed by Tony Chang.
1749
1750         This ASSERT is bogus because doctypes can be removed from the DOM and
1751         then re-added.
1752
1753         Test: fast/viewport/viewport-legacy-xhtmlmp-remove-and-add.html
1754
1755         * dom/Document.cpp:
1756         (WebCore::Document::setDocType):
1757
1758 2012-07-12  Ojan Vafai  <ojan@chromium.org>
1759
1760         Implied minimum size of flex items is min-content
1761         https://bugs.webkit.org/show_bug.cgi?id=87546
1762
1763         Reviewed by Tony Chang.
1764
1765         http://dev.w3.org/csswg/css3-flexbox/#min-size-auto
1766         In the main axis direction, min-size of auto means min-content.
1767
1768         Test: css3/flexbox/flex-item-min-size.html
1769
1770         * rendering/RenderBox.cpp:
1771         (WebCore::RenderBox::computeLogicalWidthInRegionUsing):
1772         (WebCore::RenderBox::computeContentLogicalHeightUsing):
1773         (WebCore::RenderBox::computeReplacedLogicalWidthUsing):
1774         (WebCore::RenderBox::computeReplacedLogicalHeightUsing):
1775         (WebCore::RenderBox::computePositionedLogicalWidthUsing):
1776         (WebCore::RenderBox::computePositionedLogicalHeightUsing):
1777         It turned out that these FIXMEs are all unnecessary with the changes to RenderFlexibleBox.
1778
1779         * rendering/RenderFlexibleBox.cpp:
1780         (WebCore::RenderFlexibleBox::computeMainAxisPreferredSizes):
1781         (WebCore::RenderFlexibleBox::adjustChildSizeForMinAndMax):
1782
1783 2012-07-12  Eric Penner  <epenner@google.com>
1784
1785         [chromium] Use CCTexture/TextureAllocator and remove TextureManager
1786         https://bugs.webkit.org/show_bug.cgi?id=91001
1787
1788         Reviewed by Adrienne Walker.
1789
1790         Use CCTexture to clean up CCPrioritizedTexture::Backing.
1791         Add TextureAllocator.h and remove remainder of TextureManager.h/cpp.
1792         Minor move/refactoring of link/unlink in CCPrioritizedTexture.
1793         Remove double initialization of the default memory limit.
1794
1795         Covered by existing tests (refactoring).
1796
1797         * WebCore.gypi:
1798         * platform/graphics/chromium/BitmapCanvasLayerTextureUpdater.cpp:
1799         * platform/graphics/chromium/BitmapCanvasLayerTextureUpdater.h:
1800         (WebCore):
1801         * platform/graphics/chromium/BitmapSkPictureCanvasLayerTextureUpdater.cpp:
1802         * platform/graphics/chromium/BitmapSkPictureCanvasLayerTextureUpdater.h:
1803         (WebCore):
1804         * platform/graphics/chromium/ImageLayerChromium.cpp:
1805         * platform/graphics/chromium/LayerRendererChromium.cpp:
1806         (WebCore::LayerRendererChromium::initialize):
1807         * platform/graphics/chromium/ManagedTexture.cpp: Removed.
1808         * platform/graphics/chromium/ManagedTexture.h: Removed.
1809         * platform/graphics/chromium/RenderSurfaceChromium.h:
1810         (WebCore):
1811         * platform/graphics/chromium/TextureAllocator.h: Copied from Source/WebCore/platform/graphics/chromium/cc/CCPriorityCalculator.h.
1812         (WebCore):
1813         (TextureAllocator):
1814         (WebCore::TextureAllocator::~TextureAllocator):
1815         * platform/graphics/chromium/TextureManager.cpp: Removed.
1816         * platform/graphics/chromium/TextureManager.h: Removed.
1817         * platform/graphics/chromium/TiledLayerChromium.cpp:
1818         * platform/graphics/chromium/TrackingTextureAllocator.cpp:
1819         (WebCore::TrackingTextureAllocator::createTexture):
1820         (WebCore::TrackingTextureAllocator::deleteTexture):
1821         * platform/graphics/chromium/TrackingTextureAllocator.h:
1822         * platform/graphics/chromium/cc/CCHeadsUpDisplay.cpp:
1823         * platform/graphics/chromium/cc/CCLayerTreeHost.cpp:
1824         (WebCore::CCLayerTreeHost::prioritizeTextures):
1825         * platform/graphics/chromium/cc/CCLayerTreeHost.h:
1826         (WebCore):
1827         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
1828         (WebCore::CCLayerTreeHostImpl::CCLayerTreeHostImpl):
1829         * platform/graphics/chromium/cc/CCPrioritizedTexture.cpp:
1830         (WebCore::CCPrioritizedTexture::CCPrioritizedTexture):
1831         (WebCore::CCPrioritizedTexture::setDimensions):
1832         (WebCore::CCPrioritizedTexture::textureId):
1833         (WebCore::CCPrioritizedTexture::bindTexture):
1834         (WebCore::CCPrioritizedTexture::framebufferTexture2D):
1835         (WebCore::CCPrioritizedTexture::link):
1836         (WebCore):
1837         (WebCore::CCPrioritizedTexture::unlink):
1838         * platform/graphics/chromium/cc/CCPrioritizedTexture.h:
1839         (WebCore):
1840         (WebCore::CCPrioritizedTexture::bytes):
1841         (CCPrioritizedTexture):
1842         (WebCore::CCPrioritizedTexture::haveBackingTexture):
1843         (WebCore::CCPrioritizedTexture::Backing::Backing):
1844         (WebCore::CCPrioritizedTexture::Backing::~Backing):
1845         (WebCore::CCPrioritizedTexture::Backing::owner):
1846         (Backing):
1847         (WebCore::CCPrioritizedTexture::backing):
1848         * platform/graphics/chromium/cc/CCPrioritizedTextureManager.cpp:
1849         (WebCore::CCPrioritizedTextureManager::prioritizeTextures):
1850         (WebCore::CCPrioritizedTextureManager::requestLate):
1851         (WebCore::CCPrioritizedTextureManager::acquireBackingTextureIfNeeded):
1852         (WebCore::CCPrioritizedTextureManager::reduceMemory):
1853         (WebCore::CCPrioritizedTextureManager::clearAllMemory):
1854         (WebCore::CCPrioritizedTextureManager::registerTexture):
1855         (WebCore::CCPrioritizedTextureManager::returnBackingTexture):
1856         (WebCore::CCPrioritizedTextureManager::createBacking):
1857         (WebCore::CCPrioritizedTextureManager::destroyBacking):
1858         (WebCore::CCPrioritizedTextureManager::assertInvariants):
1859         * platform/graphics/chromium/cc/CCPrioritizedTextureManager.h:
1860         (CCPrioritizedTextureManager):
1861         (WebCore::CCPrioritizedTextureManager::defaultMemoryAllocationLimit):
1862         (WebCore::CCPrioritizedTextureManager::compareBackings):
1863         * platform/graphics/chromium/cc/CCPriorityCalculator.h:
1864         * platform/graphics/chromium/cc/CCScopedTexture.cpp:
1865         * platform/graphics/chromium/cc/CCScopedTexture.h:
1866         * platform/graphics/chromium/cc/CCTexture.h:
1867         (WebCore::CCTexture::CCTexture):
1868         * platform/graphics/chromium/cc/CCTiledLayerImpl.cpp:
1869         (WebCore):
1870         * platform/graphics/chromium/cc/CCVideoLayerImpl.cpp:
1871
1872 2012-07-12  Ryosuke Niwa  <rniwa@webkit.org>
1873
1874         Let XCode have its own way.
1875
1876         * WebCore.xcodeproj/project.pbxproj:
1877
1878 2012-07-12  Julien Chaffraix  <jchaffraix@webkit.org>
1879
1880         ASSERT(genChild->isListMarker() || genChild->style()->styleType() == FIRST_LETTER) triggered on flex-box content
1881         https://bugs.webkit.org/show_bug.cgi?id=91003
1882
1883         Reviewed by Abhishek Arya.
1884
1885         Tests: fast/flexbox/assert-generated-deprecated-flexbox.html
1886                fast/flexbox/assert-generated-new-flexbox.html
1887
1888         The issue was that findBeforeAfterParent didn't return the right parent for the flex-box case. This would
1889         make us update the wrong children (and not propagate the style updates properly).
1890
1891         * rendering/RenderObjectChildList.cpp:
1892         (WebCore::findBeforeAfterParent):
1893         Added a check for flex boxes (both deprecated and new).
1894
1895 2012-07-12  Pravin D  <pravind.2k4@gmail.com>
1896
1897         Percentage width replaced element in zero percent/fixed width container block incorrectly rendered.
1898         https://bugs.webkit.org/show_bug.cgi?id=9493
1899
1900         Reviewed by Andy Estes.
1901
1902         When the width of the container is zero percent/fixed value then the width of the replaced element must also be zero.
1903
1904         Test: fast/css/percent-width-img-inside-zero-percent-and-fixed-container.html
1905
1906         * rendering/RenderBox.cpp:
1907         (WebCore::RenderBox::computeReplacedLogicalWidthUsing):
1908           When the containing block's available width is zero there can be two cases.
1909           The containing block is floated/positioned in which case the width of the replaced child element must be its instrinsic width.
1910           On the other hand if the width of the container is specified to be either zero percent or fixed value then the width of the
1911           replaced elment must be zero.
1912
1913 2012-07-11  Ryosuke Niwa  <rniwa@webkit.org>
1914
1915         invalidateNodeListsCacheAfterAttributeChanged should dynamically figure out which attribute needs invalidation
1916         https://bugs.webkit.org/show_bug.cgi?id=91046
1917
1918         Reviewed by Anders Carlsson.
1919
1920         Added an array of counters (m_nodeListCounts) for each set of attributes (NodeListInvalidationType) node lists care about
1921         to Document, and made DynamicSubtreeNodeList's constructor and destructor increment and decrement these counters via
1922         registerDynamicSubtreeNodeList and unregisterDynamicSubtreeNodeList respectively. shouldInvalidateDynamicSubtreeNodeList,
1923         which is called by invalidateNodeListsCacheAfterAttributeChanged, then use these counters to determine whether a given
1924         attribute change should result in node list invalidations.
1925
1926         Also removed m_numNodeListCaches from TreeScope because this counter has now become redundant with m_nodeListCounts.
1927
1928         * dom/ChildNodeList.cpp:
1929         (WebCore::ChildNodeList::ChildNodeList): Do not invalidate on attribute changes.
1930         * dom/ClassNodeList.cpp:
1931         (WebCore::ClassNodeList::ClassNodeList): Invalidate on class attribute changes.
1932         * dom/Document.cpp:
1933         (WebCore::Document::Document): Initialize m_nodeListCounts.
1934         (WebCore::Document::~Document): Add assertions to make sure m_listsInvalidatedAtDocument, m_nodeListCounts, and
1935         m_collections are all empty.
1936         (WebCore::Document::registerDynamicSubtreeNodeList): This function is now called for all DynamicSubtreeNodeLists supposed
1937         to just ones rooted at the document in order to increment the counter for each invalidation type.
1938         (WebCore::Document::unregisterDynamicSubtreeNodeList): Ditto.
1939         (WebCore::shouldInvalidateDynamicSubtreeNodeListForType): Checks the attribute name against NodeListInvalidationType.
1940         (WebCore::Document::shouldInvalidateDynamicSubtreeNodeList): Returns true if the given attribute name matches the invalidation
1941         type of the existing DynamicSubtreeNodeLists in the document. If the attribute name is not given (used when children change),
1942         then it checks for the existence of any DynamicSubtreeNodeLists. Conceptually, this function can be written as a list of
1943         "if" statements that checks values in m_nodeListCounts and the attribute name. We use "for" loop and switch statement instead
1944         to share logic and detect future addition of values to NodeListInvalidationType.
1945         * dom/Document.h:
1946         (Document): Moved RootType and InvalidationType from DynamicNodeListCacheBase and renamed them to NodeListRootType and
1947         NodeListInvalidationType respectively in order to reduce the possibility of future name collisions. Also the invalidation type
1948         now contains 6 values instead of 2.
1949         * dom/DynamicNodeList.cpp:
1950         (WebCore):
1951         * dom/DynamicNodeList.h:
1952         (WebCore::DynamicNodeListCacheBase::DynamicNodeListCacheBase):
1953         (WebCore::DynamicNodeListCacheBase::shouldInvalidateOnAttributeChange):
1954         (WebCore::DynamicNodeListCacheBase::rootType): Added.
1955         (WebCore::DynamicNodeListCacheBase::invalidationType): Added.
1956         (DynamicNodeListCacheBase): Uses 3 bits to store invalidation type now that the number of values have increased from 2 to 6.
1957         (WebCore::DynamicNodeList::DynamicNodeList):
1958         (WebCore::DynamicSubtreeNodeList::~DynamicSubtreeNodeList): Call unregisterDynamicSubtreeNodeList.
1959         (WebCore::DynamicSubtreeNodeList::DynamicSubtreeNodeList): Call registerDynamicSubtreeNodeList.
1960         * dom/MicroDataItemList.cpp:
1961         (WebCore::MicroDataItemList::MicroDataItemList): Invalidate on itemscope, itemprop, and itemtype content attribute changes.
1962         * dom/NameNodeList.cpp:
1963         (WebCore::NameNodeList::NameNodeList): Invalidate on name attribute changes.
1964         * dom/Node.cpp:
1965         (WebCore::Node::clearRareData):
1966         (WebCore::Node::invalidateNodeListsCacheAfterAttributeChanged): Replaced the hard coded check list of attributes, by a call
1967         to shouldInvalidateDynamicSubtreeNodeList.
1968         (WebCore::Node::invalidateNodeListsCacheAfterChildrenChanged): Calls shouldInvalidateDynamicSubtreeNodeList.
1969         (WebCore::Node::getElementsByTagName):
1970         (WebCore::Node::getElementsByTagNameNS):
1971         (WebCore::Node::getElementsByName):
1972         (WebCore::Node::getElementsByClassName):
1973         (WebCore::Node::radioNodeList):
1974         (WebCore::NodeRareData::createNodeLists):
1975         * dom/NodeRareData.h:
1976         (WebCore::NodeListsNodeData::adoptTreeScope): Unregister and register node lists in m_tagNodeListCacheNS since all node lists
1977         need to be accounted in m_nodeListCounts.
1978         (WebCore::NodeRareData::ensureNodeLists):
1979         (NodeRareData):
1980         * dom/TagNodeList.cpp:
1981         (WebCore::TagNodeList::TagNodeList): Do not invalidate on any attribute changes.
1982         * dom/TreeScope.cpp:
1983         (WebCore::TreeScope::TreeScope): No longer initializes m_numNodeListCaches since it has been removed.
1984         * dom/TreeScope.h:
1985         (TreeScope): Removed m_numNodeListCaches.
1986         * dom/TreeScopeAdopter.cpp:
1987         (WebCore::TreeScopeAdopter::moveTreeToNewScope):
1988         * html/HTMLCollection.h:
1989         (WebCore::HTMLCollectionCacheBase::HTMLCollectionCacheBase): Just pass in DoNotInvalidateOnAttributeChanges for now since
1990         it's never used in HTMLCollections.
1991         * html/LabelableElement.cpp:
1992         (WebCore::LabelableElement::labels):
1993         * html/LabelsNodeList.cpp:
1994         (WebCore::LabelsNodeList::LabelsNodeList): Invalidate on for content attribute changes.
1995         (WebCore::LabelsNodeList::~LabelsNodeList):
1996         * html/RadioNodeList.cpp:
1997         (WebCore::RadioNodeList::RadioNodeList): Invalidate on id, name, and for content attribute changes.
1998         (WebCore::RadioNodeList::~RadioNodeList):
1999
2000 2012-07-12  Gavin Barraclough  <barraclough@apple.com>
2001
2002         Threadsafety issues in WebScriptObject
2003         https://bugs.webkit.org/show_bug.cgi?id=90849
2004
2005         Reviewed by Filip Pizlo & Oliver Hunt.
2006
2007         Updated fix for this bug. Taking the JSC API lock from WebScriptObject::release
2008         may not be safe; better to just guard the JSWrapperCache with its own spinlock.
2009
2010         * bindings/objc/WebScriptObject.mm:
2011         (WebCore::getJSWrapper):
2012             - Added spinlock; also retain/autorelease the returned wrapper - it is unsafe
2013               to wait for the caller to do so, due to a race condition vs release removing
2014               the wrapper from the map.
2015         (WebCore::addJSWrapper):
2016             - Take the spinlock guarding the cache.
2017         (WebCore::removeJSWrapper):
2018             - Take the spinlock guarding the cache.
2019         (WebCore::removeJSWrapperIfRetainCountOne):
2020             - Take the spinlock guarding the cache, remove the wrapper if retainCount is one.
2021         (WebCore::createJSWrapper):
2022             - Remove the API lock; this method no longer needs to retain/autorelease (this is
2023               done by getJSWrapper).
2024         (-[WebScriptObject _setImp:originRootObject:rootObject:]):
2025             - Remove the API lock.
2026         (-[WebScriptObject release]):
2027             - Remove the API lock, retainCount check moved into removeJSWrapperIfRetainCountOne.
2028
2029 2012-07-11  David Hyatt  <hyatt@apple.com>
2030
2031         https://bugs.webkit.org/show_bug.cgi?id=91000
2032         REGRESSION (r122244): Overflow elements don't shrink as much as they should.
2033
2034         Reviewed by Simon Fraser.
2035
2036         This is a fix for a a regression from https://bugs.webkit.org/show_bug.cgi?id=90646.
2037
2038         I incorrectly analyzed the issue with Robert Hogan's negative margin patch and fooled myself into putting back
2039         in an incorrect minimum width check from long ago.
2040
2041         What should have happened in the test case I patched is that the overflow element should shrink to 0. The issue
2042         with improving the logical top estimate in the previous patch is it made the clear delta become 0. This in turn
2043         exposed a bug in our clearing algorithm with Robert's changes where you could need a relayout even if you didn't
2044         actually move. This issue only occurs because the floats list is getting changed mid-layout because of negative margins.
2045
2046         The patch changes getClearDelta to call setChildNeedsLayout(true) on children whose widths change even when their
2047         positions do not. In effect this dynamic addition of new floats after you have done a layout on the child already means
2048         that you can need to lay out again despite not actually having to move.
2049
2050         To handle this, the code that does the relayout is now called if the child needs a relayout. This is done even if
2051         the logical top estimate matches the final position.
2052
2053         No new tests required, since the test in fast/block/float is now correctly covering the issue.
2054
2055         * rendering/RenderBlock.cpp:
2056         (WebCore::RenderBlock::layoutBlockChild):
2057         (WebCore::RenderBlock::getClearDelta):
2058         * rendering/RenderBox.cpp:
2059         (WebCore::RenderBox::shrinkLogicalWidthToAvoidFloats):
2060
2061 2012-07-12  James Weatherall  <wez@chromium.org>
2062
2063         storage tests are flaky (crashing) on windows
2064         https://bugs.webkit.org/show_bug.cgi?id=90469
2065
2066         Reviewed by Kentaro Hara.
2067
2068         Add a missing check that the underlying V8 object reference in a V8 NPObject is valid, and zero the NPObject's rootObject member when disposing it, to ensure that it won't be mistakenly touched after that point.
2069
2070         This patch is intended to resolve flakiness in the storage tests including:
2071           storage/indexeddb/mozilla/indexes.html
2072           storage/indexeddb/mozilla/key-requirements-inline-and-passed.html
2073           storage/websql/multiple-databases-garbage-collection.html
2074
2075         * bindings/v8/NPV8Object.cpp:
2076         (WebCore::disposeUnderlyingV8Object):
2077         Zero the NPObject's underlying rootObject.
2078         (_NPN_EvaluateHelper):
2079         Add check that the underlying V8 object reference is valid.
2080
2081 2012-07-12  No'am Rosenthal  <noam.rosenthal@nokia.com>
2082
2083         Move TextureMapperAnimation and texmap/LayerTransform to platform/graphics
2084         https://bugs.webkit.org/show_bug.cgi?id=91111
2085
2086         Reviewed by Kenneth Rohde Christiansen.
2087
2088         TextureMapperAnimation and LayerTransform are not specific to TextureMapper, and we want
2089         to use them for other purposes as well. Moving them to platform/graphics would make that
2090         more explicit.
2091
2092         No new tests, moving files around.
2093
2094         * GNUmakefile.list.am:
2095         * Target.pri:
2096         * WebCore.gypi:
2097         * platform/graphics/GraphicsLayerAnimation.cpp: Renamed from Source/WebCore/platform/graphics/texmap/TextureMapperAnimation.cpp.
2098         * platform/graphics/GraphicsLayerAnimation.h: Renamed from Source/WebCore/platform/graphics/texmap/TextureMapperAnimation.h.
2099         * platform/graphics/GraphicsLayerTransform.cpp: Renamed from Source/WebCore/platform/graphics/texmap/LayerTransform.cpp.
2100         * platform/graphics/GraphicsLayerTransform.h: Renamed from Source/WebCore/platform/graphics/texmap/LayerTransform.h.
2101         * platform/graphics/texmap/GraphicsLayerTextureMapper.cpp:
2102         (WebCore::GraphicsLayerTextureMapper::addAnimation):
2103         * platform/graphics/texmap/GraphicsLayerTextureMapper.h:
2104         (GraphicsLayerTextureMapper):
2105         * platform/graphics/texmap/TextureMapperLayer.h:
2106         (TextureMapperLayer):
2107         (WebCore::TextureMapperLayer::setAnimatedTransform):
2108         (WebCore::TextureMapperLayer::setAnimatedOpacity):
2109
2110 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2111
2112         Web Inspector: Simplify UISourceCode code after moving revisions support inside it.
2113         https://bugs.webkit.org/show_bug.cgi?id=91118
2114
2115         Reviewed by Pavel Feldman.
2116
2117         * inspector/front-end/UISourceCode.js:
2118         (WebInspector.UISourceCode.prototype._setContent):
2119         (WebInspector.UISourceCode.prototype.revertToOriginal):
2120         (WebInspector.UISourceCode.prototype.revertAndClearHistory):
2121         (WebInspector.UISourceCode.prototype.contentChanged):
2122         (WebInspector.UISourceCode.prototype.commitWorkingCopy):
2123         (WebInspector.Revision.prototype.revertToThis):
2124
2125 2012-07-12  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
2126
2127         Unreviewed. Build fix for r122462.
2128
2129         * platform/qt/MemoryUsageSupportQt.cpp:
2130
2131 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2132
2133         Web Inspector: Remove resources panel edited resources search support.
2134         https://bugs.webkit.org/show_bug.cgi?id=91101
2135
2136         Reviewed by Pavel Feldman.
2137
2138         * inspector/front-end/ResourcesPanel.js:
2139         (WebInspector.ResourcesPanel.prototype.performSearch.callback):
2140         (WebInspector.ResourcesPanel.prototype.performSearch):
2141
2142 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2143
2144         Unreviewed r122460 inspector closure compilation fix follow up.
2145
2146         * inspector/front-end/RevisionHistoryView.js:
2147         (WebInspector.RevisionHistoryView):
2148
2149 2012-07-12  Ilya Tikhonovsky  <loislo@chromium.org>
2150
2151         Web Inspector: fix native memory instrumentation code for the bindings instrumentation.
2152         https://bugs.webkit.org/show_bug.cgi?id=91096
2153
2154         The instrumented class has to have instrumentation method which reports class size and type and
2155         the member objects and pointers.
2156
2157         Sample:
2158         void Node::reportMemoryUsage(MemoryObjectInfo* memoryObjectInfo) const
2159         {
2160             memoryObjectInfo->reportObjectInfo(this, MemoryInstrumentation::DOM); // report object size and type.
2161             TreeShared<Node, ContainerNode>::reportMemoryUsage(memoryObjectInfo); // call base class instrumentation.
2162             ScriptWrappable::reportMemoryUsage(memoryObjectInfo); // call base class instrumentation.
2163             memoryObjectInfo->reportPointer(m_document, MemoryInstrumentation::DOM); // report uninstrumented pointer.
2164             memoryObjectInfo->reportInstrumentedPointer(m_next); // report instrumented pointer.
2165             memoryObjectInfo->reportInstrumentedObject(m_anObject); // report instrumented object.
2166         }
2167
2168         Reviewed by Pavel Feldman.
2169
2170         Existing tests for native memory instrumentation.
2171
2172         * bindings/v8/DOMDataStore.cpp:
2173         (WebCore::DOMDataStore::reportMemoryUsage):
2174         * bindings/v8/DOMDataStore.h:
2175         (WebCore):
2176         (DOMDataStore):
2177         * bindings/v8/IntrusiveDOMWrapperMap.h:
2178         (WebCore::ChunkedTable::reportMemoryUsage):
2179         * bindings/v8/ScriptProfiler.cpp:
2180         (WebCore::ScriptProfiler::collectBindingMemoryInfo):
2181         * bindings/v8/V8Binding.cpp:
2182         (WebCore::V8BindingPerIsolateData::reportMemoryUsage):
2183         (WebCore::StringCache::reportMemoryUsage):
2184         * bindings/v8/V8Binding.h:
2185         (WebCore):
2186         (StringCache):
2187         (V8BindingPerIsolateData):
2188         * bindings/v8/V8DOMMap.h:
2189         (WebCore):
2190         (AbstractWeakReferenceMap):
2191         * inspector/InspectorMemoryAgent.cpp:
2192         (WebCore):
2193
2194 2012-07-12  Pavel Feldman  <pfeldman@chromium.org>
2195
2196         Web Inspector: beautify find bar looks, simplify search update routines.
2197         https://bugs.webkit.org/show_bug.cgi?id=91087
2198
2199         Reviewed by Vsevolod Vlasov.
2200
2201         This change updates the looks to the one on the screenshots and simplifies match count update routines.
2202
2203         * inspector/front-end/SearchController.js:
2204         (WebInspector.SearchController.onMatchesMouseDown):
2205         (WebInspector.SearchController):
2206         (WebInspector.SearchController.prototype.activePanelChanged.performPanelSearch):
2207         (WebInspector.SearchController.prototype.activePanelChanged):
2208         (WebInspector.SearchController.prototype._updateSearchNavigationButtonState):
2209         (WebInspector.SearchController.prototype._updateSearchMatchesCountAndCurrentMatchIndex):
2210         (WebInspector.SearchController.prototype._onKeyDown):
2211         (WebInspector.SearchController.prototype._onInput):
2212         (WebInspector.SearchController.prototype._onNextButtonSearch):
2213         (WebInspector.SearchController.prototype._onPrevButtonSearch):
2214         (WebInspector.SearchController.prototype._performSearch):
2215         * inspector/front-end/inspector.css:
2216         (#search):
2217         (#search:focus):
2218         (.toolbar-search-navigation-controls):
2219         (.toolbar-search-navigation):
2220         (.toolbar-search-navigation.enabled:hover):
2221         (.toolbar-search-navigation.enabled, .toolbar-search-navigation.enabled:active):
2222         (.toolbar-search):
2223         (.toolbar-search-control):
2224         (.search-results-matches):
2225         (.inspector-footer):
2226
2227 2012-07-12  Joshua Bell  <jsbell@chromium.org>
2228
2229         IndexedDB: ASSERT hit calling open from callback in Worker
2230         https://bugs.webkit.org/show_bug.cgi?id=90832
2231
2232         Reviewed by Kentaro Hara.
2233
2234         GroupSettings are used to provide the backing store path in some
2235         ports. Accessing those settings from a Worker was added, but the
2236         access referenced thread startup data that is cleared before the
2237         run loop, so an IDBFactory.open() call executed asynchronously
2238         would dereference a null pointer. Plumb the settings startup
2239         data into the context itself, like all of the other properties.
2240
2241         Test: storage/indexeddb/open-twice-workers.html
2242
2243         * Modules/indexeddb/IDBFactory.cpp:
2244         (WebCore::IDBFactory::open):
2245         * workers/DedicatedWorkerContext.cpp:
2246         (WebCore::DedicatedWorkerContext::DedicatedWorkerContext):
2247         * workers/DedicatedWorkerContext.h:
2248         (WebCore::DedicatedWorkerContext::create):
2249         (DedicatedWorkerContext):
2250         * workers/DedicatedWorkerThread.cpp:
2251         (WebCore::DedicatedWorkerThread::createWorkerContext):
2252         * workers/DedicatedWorkerThread.h:
2253         (DedicatedWorkerThread):
2254         * workers/SharedWorkerContext.cpp:
2255         (WebCore::SharedWorkerContext::SharedWorkerContext):
2256         * workers/SharedWorkerContext.h:
2257         (WebCore::SharedWorkerContext::create):
2258         (SharedWorkerContext):
2259         * workers/SharedWorkerThread.cpp:
2260         (WebCore::SharedWorkerThread::createWorkerContext):
2261         * workers/SharedWorkerThread.h:
2262         (SharedWorkerThread):
2263         * workers/WorkerContext.cpp:
2264         (WebCore::WorkerContext::WorkerContext):
2265         * workers/WorkerContext.h:
2266         (WebCore::WorkerContext::groupSettings):
2267         (WorkerContext):
2268         * workers/WorkerThread.cpp:
2269         (WebCore::WorkerThread::workerThread):
2270         * workers/WorkerThread.h:
2271         (WorkerThread):
2272
2273 2012-07-12  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
2274
2275         [Qt] Implement MemoryUsageSupport
2276         https://bugs.webkit.org/show_bug.cgi?id=91094
2277
2278         Reviewed by Adam Barth.
2279
2280         Implements Qt versions of the memory-usage functions using the information we have available
2281         from the various memory systems used in WebKit.
2282
2283         Also gets rid of a redundant indirection in V8GCController.
2284
2285         * Target.pri:
2286         * bindings/v8/V8GCController.cpp:
2287         (WebCore::V8GCController::gcEpilogue):
2288         (WebCore::V8GCController::checkMemoryUsage):
2289         * platform/qt/MemoryUsageSupportQt.cpp: Added.
2290         (WebCore::mallocMemoryUsage):
2291         (WebCore::memoryUsageKB):
2292         (WebCore::actualMemoryUsageKB):
2293         (WebCore::MemoryUsageSupport::memoryUsageMB):
2294         (WebCore::MemoryUsageSupport::actualMemoryUsageMB):
2295         (WebCore::MemoryUsageSupport::lowMemoryUsageMB):
2296         (WebCore::MemoryUsageSupport::highMemoryUsageMB):
2297         (WebCore::MemoryUsageSupport::highUsageDeltaMB):
2298         (WebCore::MemoryUsageSupport::processMemorySizesInBytes):
2299         * platform/qt/PlatformSupport.h:
2300         (PlatformSupport):
2301
2302 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2303
2304         Web Inspector: Revision history view should be updated when uiSourceCodes are removed or replaced.
2305         https://bugs.webkit.org/show_bug.cgi?id=91095
2306
2307         Reviewed by Pavel Feldman.
2308
2309         RevisionHistoryView is now reset in WorkspaceReset event handler.
2310         UISourceCodes are now removed from RevisionHistoryView when uiSourceCode is removed or replace.
2311
2312         * inspector/front-end/RevisionHistoryView.js:
2313         (WebInspector.RevisionHistoryView):
2314         (WebInspector.RevisionHistoryView.prototype._clearHistory):
2315         (WebInspector.RevisionHistoryView.prototype._uiSourceCodeRemoved):
2316         (WebInspector.RevisionHistoryView.prototype._uiSourceCodeReplaced):
2317         * inspector/front-end/ScriptsPanel.js:
2318         (WebInspector.ScriptsPanel.prototype._reset):
2319
2320 2012-07-12  Dongwoo Im  <dw.im@samsung.com>
2321
2322         InspectorFileSystemAgent.cpp have to include File.h explicitly
2323         https://bugs.webkit.org/show_bug.cgi?id=91078
2324
2325         Reviewed by Vsevolod Vlasov.
2326
2327         Make sure the InspectorFileSystemAgent.cpp include the File.h by including explicitly.
2328
2329         No new tests. Covered by existing tests.
2330
2331         * inspector/InspectorFileSystemAgent.cpp: Include File.h
2332
2333 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2334
2335         Web Inspector: [Regression] Stale revisions should be removed on navigation.
2336         https://bugs.webkit.org/show_bug.cgi?id=91080
2337
2338         Reviewed by Pavel Feldman.
2339
2340         Moved all the code responsible for script mapping resetting to Workspace.
2341         Stale revisions are removed in Workspace._reset now.
2342
2343         * inspector/front-end/ScriptSnippetModel.js:
2344         (WebInspector.ScriptSnippetModel):
2345         * inspector/front-end/StylesPanel.js:
2346         (WebInspector.StylesUISourceCodeProvider):
2347         (WebInspector.StylesUISourceCodeProvider.prototype._populate):
2348         (WebInspector.StylesUISourceCodeProvider.prototype.reset):
2349         * inspector/front-end/UISourceCode.js:
2350         (WebInspector.UISourceCode.prototype._restoreRevisionHistory):
2351         * inspector/front-end/Workspace.js:
2352         (WebInspector.Workspace):
2353         (WebInspector.Workspace.prototype._reset):
2354
2355 2012-07-12  Andreas Kling  <kling@webkit.org>
2356
2357         Make ElementAttributeData a variable-sized object to reduce memory use.
2358         <http://webkit.org/b/88240>
2359
2360         Reviewed by Antti "Obi-Wan" Koivisto.
2361
2362         Take advantage of the fact that we know at ElementAttributeData construction time how many attributes
2363         it needs to accomodate and allocate exactly as much space as needed instead of using a Vector.
2364         For elements that never have their attribute list mutated (the vast majority), this saves a lot of
2365         memory and removes the indirection to Vector<Attribute>'s heap-allocated storage.
2366
2367         Introduced a mutability flag to ElementAttributeData and sprinkled assertions all over to make sure
2368         that nothing tries to mutate an element with a raw attribute array.
2369
2370         When an Element's attribute(s) are mutated, we reconstruct the ElementAttributeData, this time using
2371         a Vector as backing instead. This is triggered when calling Element::mutableAttributeData().
2372
2373         This reduces memory consumption by 3.2MB when viewing the full HTML5 spec at <http://whatwg.org/c/>.
2374         That is a ~35% reduction in DOM attribute memory use.
2375
2376         Furthermore, that page ends up promoting 27% of the elements to mutable attribute storage due to dynamic
2377         adding of "class" attributes. For more static pages, savings are even greater.
2378
2379         Also did away with ElementAttributeData::removeAttribute(name) and do separate index lookup where
2380         needed. Not a big deal but avoids double lookup of attributes when removing them.
2381
2382         * css/PropertySetCSSStyleDeclaration.cpp:
2383         (WebCore::InlineCSSStyleDeclaration::ensureMutablePropertySet):
2384
2385             Call ensureInlineStyle() on the element so we know we have a StylePropertySet to work with.
2386             May cause the creation (and replacement) of a new underlying StylePropertySet on the Element's
2387             attribute data.
2388
2389         * css/StylePropertySet.cpp:
2390         (WebCore::StylePropertySet::adoptCSSOMWrapperFrom):
2391
2392             Added. Allows switching the underlying StylePropertySet object while retaining CSSOM wrapper
2393             identity (web facing behavior.)
2394
2395         * dom/Element.cpp:
2396         (WebCore::Element::detachAttribute):
2397         (WebCore::Element::removeAttribute):
2398         (WebCore::Element::attributes):
2399         (WebCore::Element::setAttributeInternal):
2400         (WebCore::Element::parserSetAttributes):
2401         (WebCore::Element::hasEquivalentAttributes):
2402         (WebCore::Element::setAttributeNode):
2403         (WebCore::Element::removeAttributeNode):
2404         (WebCore::Element::getAttributeNode):
2405         (WebCore::Element::getAttributeNodeNS):
2406         (WebCore::Element::hasAttribute):
2407         (WebCore::Element::hasAttributeNS):
2408         (WebCore::Element::normalizeAttributes):
2409         (WebCore::Element::cloneAttributesFromElement):
2410         (WebCore::Element::createMutableAttributeData):
2411         * dom/Element.h:
2412         (WebCore::Element::attributeData):
2413         (Element):
2414         (WebCore::Element::updatedAttributeData):
2415         (WebCore::Element::ensureAttributeData):
2416         (WebCore::Element::ensureUpdatedAttributeData):
2417         (WebCore::Element::mutableAttributeData):
2418         (WebCore):
2419         * dom/ElementAttributeData.cpp:
2420         (WebCore::ElementAttributeData::createImmutable):
2421         (WebCore):
2422         (WebCore::ElementAttributeData::ElementAttributeData):
2423         (WebCore::ElementAttributeData::~ElementAttributeData):
2424         (WebCore::ElementAttributeData::attrIfExists):
2425         (WebCore::ElementAttributeData::ensureAttr):
2426         (WebCore::ElementAttributeData::setAttr):
2427         (WebCore::ElementAttributeData::removeAttr):
2428         (WebCore::ElementAttributeData::ensureInlineStyle):
2429         (WebCore::ElementAttributeData::ensureMutableInlineStyle):
2430         (WebCore::ElementAttributeData::destroyInlineStyle):
2431         (WebCore::ElementAttributeData::addAttribute):
2432         (WebCore::ElementAttributeData::removeAttribute):
2433         (WebCore::ElementAttributeData::isEquivalent):
2434         (WebCore::ElementAttributeData::detachAttrObjectsFromElement):
2435         (WebCore::ElementAttributeData::getAttributeItemIndexSlowCase):
2436         (WebCore::ElementAttributeData::cloneDataFrom):
2437         (WebCore::ElementAttributeData::clearAttributes):
2438         (WebCore::ElementAttributeData::replaceAttribute):
2439         (WebCore::ElementAttributeData::getAttributeNode):
2440         * dom/ElementAttributeData.h:
2441         (WebCore::ElementAttributeData::create):
2442         (ElementAttributeData):
2443         (WebCore::ElementAttributeData::setClass):
2444         (WebCore::ElementAttributeData::setIdForStyleResolution):
2445         (WebCore::ElementAttributeData::inlineStyle):
2446         (WebCore::ElementAttributeData::setAttributeStyle):
2447         (WebCore::ElementAttributeData::reportMemoryUsage):
2448         (WebCore::ElementAttributeData::isMutable):
2449         (WebCore::ElementAttributeData::makeMutable):
2450         (WebCore::ElementAttributeData::length):
2451         (WebCore):
2452         (WebCore::ElementAttributeData::isEmpty):
2453         (WebCore::ElementAttributeData::array):
2454         (WebCore::ElementAttributeData::removeAttribute):
2455         (WebCore::ElementAttributeData::getAttributeItem):
2456         (WebCore::ElementAttributeData::getAttributeItemIndex):
2457         (WebCore::ElementAttributeData::attributeItem):
2458         * dom/StyledElement.cpp:
2459         (WebCore::StyledElement::style):
2460         (WebCore::StyledElement::classAttributeChanged):
2461         (WebCore::StyledElement::setInlineStyleProperty):
2462         (WebCore::StyledElement::removeInlineStyleProperty):
2463         (WebCore::StyledElement::addSubresourceAttributeURLs):
2464         * dom/StyledElement.h:
2465         (WebCore::StyledElement::ensureInlineStyle):
2466         * html/parser/HTMLConstructionSite.cpp:
2467         (WebCore::HTMLConstructionSite::createHTMLElementFromSavedElement):
2468         * xml/parser/XMLDocumentParserQt.cpp:
2469         (WebCore::XMLDocumentParser::XMLDocumentParser):
2470
2471 2012-07-12  Florin Malita  <fmalita@chromium.org>
2472
2473         Incorrect handling of chained pending resources in SVGUseElement
2474         https://bugs.webkit.org/show_bug.cgi?id=89686
2475
2476         Reviewed by Nikolas Zimmermann.
2477
2478         Currently SVGUseElement builds the shadow tree when the target first
2479         becomes available. This is normally OK, but if the target itself (or
2480         one of its children) is a <use> element with pending resources, then
2481         the shadow expansion only captures the current state of the tree and
2482         never gets updated when the pending resource becomes available.
2483
2484         In order to support arbitrary <use>-on-<use> constructs, this patch
2485         tracks nested <use> dependencies and rebuilds the dependent trees
2486         whenever the target gets updated.
2487
2488
2489         Tests: svg/custom/use-nested-expected.svg
2490                svg/custom/use-nested.svg
2491
2492         * svg/SVGElement.cpp:
2493         (WebCore::SVGElement::removedFrom): removedFrom needs to be called up the inheritance chain
2494         before invoking removeAllElementReferencesForTarget. Otherwise we could end up finding the
2495         element being removed as a valid target in SVGUseElement::buildInstanceTree because its
2496         InDocument flag is not cleared yet.
2497         * svg/SVGUseElement.cpp:
2498         (WebCore::SVGUseElement::~SVGUseElement):
2499         (WebCore::SVGUseElement::clearResourceReferences):
2500         (WebCore::SVGUseElement::buildPendingResource):
2501         (WebCore::SVGUseElement::buildShadowAndInstanceTree):
2502         (WebCore::SVGUseElement::buildInstanceTree):
2503         * svg/SVGUseElement.h:
2504         (SVGUseElement):
2505         Track <use> -> <use> dependencies using SVGDocumentExtensions'
2506         m_elementDependencies framework and ensure dependent trees are rebuilt
2507         when the target itself gets rebuilt.
2508
2509 2012-07-12  MORITA Hajime  <morrita@google.com>
2510
2511         Typo: ParentTranversalDetails should be ParentTraversalDetails
2512         https://bugs.webkit.org/show_bug.cgi?id=91059
2513
2514         Reviewed by Andreas Kling.
2515
2516         Did a bare rename from ParentTranversalDetails to ParentTraversalDetails
2517
2518         No new tests, just a rename.
2519
2520         * dom/ComposedShadowTreeWalker.cpp:
2521         (WebCore::ComposedShadowTreeWalker::ParentTraversalDetails::didTraverseInsertionPoint):
2522         (WebCore::ComposedShadowTreeWalker::ParentTraversalDetails::didTraverseShadowRoot):
2523         (WebCore::ComposedShadowTreeWalker::ParentTraversalDetails::didFindNode):
2524         (WebCore::ComposedShadowTreeWalker::findParent):
2525         (WebCore::ComposedShadowTreeWalker::traverseNodeEscapingFallbackContents):
2526         (WebCore::ComposedShadowTreeWalker::traverseParent):
2527         (WebCore::ComposedShadowTreeWalker::traverseParentInCurrentTree):
2528         (WebCore::ComposedShadowTreeWalker::traverseParentBackToYoungerShadowRootOrHost):
2529         * dom/ComposedShadowTreeWalker.h:
2530         (WebCore::ComposedShadowTreeWalker::ParentTraversalDetails::ParentTraversalDetails):
2531         (ComposedShadowTreeWalker):
2532         * dom/NodeRenderingContext.h:
2533         (NodeRenderingContext):
2534
2535 2012-07-11  Antti Koivisto  <antti@apple.com>
2536
2537         Failure to dispatch didFinishLoadForFrame if font load fails synchronously
2538         https://bugs.webkit.org/show_bug.cgi?id=91018
2539
2540         Reviewed by Enrica Casucci.
2541
2542         New font loads may be triggered by layout after the document load is complete but before we have dispatched
2543         didFinishLoading for the frame. If the load fails synchronously we might fail to ever invoke
2544         FrameLoader::checkLoadComplete and so fail to dispatch didFinishLoading.
2545
2546         Make sure this doesn't happen by calling FrameLoader::checkLoadComplete explicitly after triggering font loads.
2547
2548         * css/CSSFontSelector.cpp:
2549         (WebCore::CSSFontSelector::beginLoadTimerFired):
2550
2551 2012-07-12  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2552
2553         Register protocol handler files should be in Modules/protocolhandler
2554         https://bugs.webkit.org/show_bug.cgi?id=90766
2555
2556         Reviewed by Hajime Morita.
2557
2558         The register protocol handler files are now self-contained. This patch is moved to the Modules.
2559
2560         No new tests. Covered by existing tests.
2561
2562         * CMakeLists.txt:
2563         * GNUmakefile.am:
2564         * GNUmakefile.list.am:
2565         * Modules/protocolhandler/NavigatorRegisterProtocolHandler.cpp: Renamed from Source/WebCore/page/NavigatorRegisterProtocolHandler.cpp.
2566         (WebCore):
2567         (WebCore::initProtocolHandlerWhitelist):
2568         (WebCore::verifyCustomHandlerURL):
2569         (WebCore::isProtocolWhitelisted):
2570         (WebCore::verifyProtocolHandlerScheme):
2571         (WebCore::NavigatorRegisterProtocolHandler::NavigatorRegisterProtocolHandler):
2572         (WebCore::NavigatorRegisterProtocolHandler::~NavigatorRegisterProtocolHandler):
2573         (WebCore::NavigatorRegisterProtocolHandler::registerProtocolHandler):
2574         (WebCore::customHandlersStateString):
2575         (WebCore::NavigatorRegisterProtocolHandler::isProtocolHandlerRegistered):
2576         (WebCore::NavigatorRegisterProtocolHandler::unregisterProtocolHandler):
2577         * Modules/protocolhandler/NavigatorRegisterProtocolHandler.h: Renamed from Source/WebCore/page/NavigatorRegisterProtocolHandler.h.
2578         (WebCore):
2579         (NavigatorRegisterProtocolHandler):
2580         * Modules/protocolhandler/NavigatorRegisterProtocolHandler.idl: Renamed from Source/WebCore/page/NavigatorRegisterProtocolHandler.idl.
2581         * Target.pri:
2582         * WebCore.gyp/WebCore.gyp:
2583         * WebCore.gypi:
2584         * WebCore.pri:
2585         * WebCore.vcproj/WebCore.vcproj:
2586         * WebCore.vcproj/WebCoreCommon.vsprops:
2587         * WebCore.vcproj/copyForwardingHeaders.cmd:
2588         * WebCore.xcodeproj/project.pbxproj:
2589
2590 2012-07-12  Jonathan Dong  <jonathan.dong@torchmobile.com.cn>
2591
2592         [BlackBerry] Cannot use digest proxy auth and NTLM auth at the same time
2593         https://bugs.webkit.org/show_bug.cgi?id=91054
2594
2595         Reviewed by George Staikos.
2596
2597         Added an interface function syncProxyCredential() in class
2598         PageClientBlackBerry, which is responsible to notify WebPageClient
2599         to synchronize proxy credential to the chrome process.
2600
2601         Internally reviewed by Jason Liu <jason.liu@torchmobile.com.cn>
2602
2603         No new tests since there's no functional change.
2604
2605         * platform/blackberry/PageClientBlackBerry.h:
2606         * platform/network/blackberry/NetworkJob.cpp:
2607         (WebCore::NetworkJob::storeCredentials): Remember the accepted proxy
2608         credential and notify webpage client to synchronize it.
2609
2610 2012-07-12  Xingnan Wang  <xingnan.wang@intel.com>
2611
2612         [IndexedDB] upperOpen set to true in lowerBound()/lowerOpen set to true in upperBound()
2613         https://bugs.webkit.org/show_bug.cgi?id=90867
2614
2615         Reviewed by Tony Chang.
2616
2617         In the latest W3C spec upperOpen/lowerOpen are set to true in lowerBound()/upperBound(), which
2618         we should keep aligned with.
2619
2620         No new tests - updated storage/indexeddb/keyrange.html to match new behavior.
2621
2622         * Modules/indexeddb/IDBKeyRange.cpp:
2623         (WebCore::IDBKeyRange::lowerBound):
2624         (WebCore::IDBKeyRange::upperBound):
2625
2626 2012-07-12  MORITA Hajime  <morrita@google.com>
2627
2628         [Shadow DOM] <video> with <shadow> crashes
2629         https://bugs.webkit.org/show_bug.cgi?id=91055
2630
2631         Reviewed by Kent Tamura.
2632
2633         This is similar to Bug 90480, where an undesired renderer is created by
2634         locating an insertion point on the shadow boundary.
2635
2636         This change adds a guard for such case by cheking whether the
2637         source node of each to-be-created renderer comes from the UA shadow
2638         tree, which is allowed to have a renderer.
2639
2640         Test: fast/dom/shadow/insertion-point-video-crash.html
2641
2642         * html/HTMLMediaElement.cpp:
2643         (WebCore::HTMLMediaElement::childShouldCreateRenderer): Added a check.
2644         (WebCore::HTMLMediaElement::mediaControls): Added const.
2645         (WebCore::HTMLMediaElement::hasMediaControls): Added const.
2646         * html/HTMLMediaElement.h:
2647         (HTMLMediaElement):
2648
2649 2012-07-12  Yoshifumi Inoue  <yosin@chromium.org>
2650
2651         REGRESSION(r122184): LocaleMac::currentLocale should use current locale rather than newly create locale object.
2652         https://bugs.webkit.org/show_bug.cgi?id=91057
2653
2654         Reviewed by Kent Tamura.
2655
2656         This patch changes NSLocale object of LocaleMac::m_locale variable to
2657         current NSLocale object rather than newly created NSLocale object from
2658         locale identifier.
2659
2660         No new tests. We don't have way to change system preferences from
2661         test scripts and restoring them. To test this patch, we need to do so.
2662
2663         * platform/text/mac/LocaleMac.h:
2664         (LocaleMac): Added a constructor which takes NSLocale object.
2665         * platform/text/mac/LocaleMac.mm:
2666         (WebCore::LocaleMac::LocaleMac): Added a constructor which takes NSLocale object.
2667         (WebCore::LocaleMac::currentLocale): Changed to construct LocaleMac object from NSLocale object rather than locale identifier.
2668
2669 2012-07-11  Carlos Garcia Campos  <cgarcia@igalia.com>
2670
2671         [GTK] Add webkit_cookie_manager_set_persistent_storage() to WebKit2 GTK+ API
2672         https://bugs.webkit.org/show_bug.cgi?id=83016
2673
2674         Reviewed by Martin Robinson.
2675
2676         * platform/network/soup/CookieStorageSoup.cpp:
2677         (WebCore::soupCookiesChanged): Do not notify about changes in
2678         other cookie jars than the current one.
2679
2680 2012-07-12  Philip Rogers  <pdr@google.com>
2681
2682         Refactor RenderSVGShape bounding box code
2683         https://bugs.webkit.org/show_bug.cgi?id=90655
2684
2685         Reviewed by Nikolas Zimmermann.
2686
2687         RenderSVGShape::objectBoundingBox worked differently than RenderSVGShape::strokeBoundingBox by
2688         not caching the object bounding box and instead computing it on each call. For consistency and
2689         performance objectBoundingBox has been refactored to return a cached value.
2690
2691         createShape has been renamed updateShapeFromElement for understandability. updateShapeFromElement
2692         now updates the internal state of the shape (bounding boxes, etc) from the associated element.
2693         RenderSVGShape::inflateWithStrokeAndMarkerBounds has been merged into
2694         RenderSVGShape::calculateStrokeBoundingBox which is called from updateShapeFromElement.
2695
2696         After this change all bounding box computation is now handled in updateShapeFromElement. Because
2697         subclasses override updateShapeFromElement it will be easy for them to have custom bounding box
2698         code there (as will happen for RenderSVGPath in a followup patch).
2699
2700         strokeBoundingBox and objectBoundingBox are now able to return their cached values immediately
2701         in RenderSVGRect and RenderSVGEllipse instead of checking their fallback state on each call.
2702
2703         Additionally, to save space RenderSVGEllipse and RenderSVGRect now use the m_fillBoundingBox and
2704         m_strokeBoundingBox of RenderSVGShape instead of having their own.
2705
2706         This patch also removes setStrokeAndMarkerBoundingBox that was previously dead code.
2707
2708         No new tests, just a refactoring.
2709
2710         * rendering/svg/RenderSVGEllipse.cpp:
2711         (WebCore::RenderSVGEllipse::updateShapeFromElement):
2712         (WebCore):
2713         (WebCore::RenderSVGEllipse::fillShape):
2714         (WebCore::RenderSVGEllipse::strokeShape):
2715         (WebCore::RenderSVGEllipse::shapeDependentStrokeContains):
2716         * rendering/svg/RenderSVGEllipse.h:
2717         (RenderSVGEllipse):
2718         (WebCore::RenderSVGEllipse::isEmpty):
2719         * rendering/svg/RenderSVGRect.cpp:
2720         (WebCore::RenderSVGRect::updateShapeFromElement):
2721         (WebCore):
2722         (WebCore::RenderSVGRect::fillShape):
2723         (WebCore::RenderSVGRect::strokeShape):
2724         (WebCore::RenderSVGRect::shapeDependentStrokeContains):
2725         (WebCore::RenderSVGRect::shapeDependentFillContains):
2726         * rendering/svg/RenderSVGRect.h:
2727         (RenderSVGRect):
2728         (WebCore::RenderSVGRect::isEmpty):
2729         * rendering/svg/RenderSVGShape.cpp:
2730         (WebCore::RenderSVGShape::updateShapeFromElement):
2731         (WebCore):
2732         (WebCore::RenderSVGShape::layout):
2733         (WebCore::RenderSVGShape::calculateObjectBoundingBox):
2734         (WebCore::RenderSVGShape::calculateStrokeBoundingBox):
2735         (WebCore::RenderSVGShape::updateRepaintBoundingBox):
2736         * rendering/svg/RenderSVGShape.h:
2737         (RenderSVGShape):
2738         (WebCore::RenderSVGShape::objectBoundingBox):
2739         (WebCore::RenderSVGShape::strokeBoundingBox):
2740
2741 2012-07-12  Kent Tamura  <tkent@chromium.org>
2742
2743         Do not save the form state signature if nothing is saved
2744         https://bugs.webkit.org/show_bug.cgi?id=91050
2745
2746         Reviewed by Hajime Morita.
2747
2748         This change will reduce the size of HistoyItem.
2749
2750         Test: fast/forms/state-restore-empty-state.html
2751
2752         * html/FormController.cpp:
2753         (WebCore::FormController::formElementsState):
2754         If stateVector has only the signature string, clear it.
2755
2756 2012-07-11  Mark Rowe  <mrowe@apple.com>
2757
2758         <http://webkit.org/b/91051> Fix the Qt Mac build after r122400.
2759
2760         Qt on Mac builds with a deployment target of OS X 10.5. However, it was not been setting
2761         BUILDING_ON_LEOPARD / TARGETING_LEOPARD and thus was falling down code paths in DisplaySleepDisabler.cpp
2762         that were specific to Snow Leopard and newer. After the removal of BUILDING_ON_LEOPARD
2763         and TARGETING_LEOPARD it ended up falling down the correct Leopard-compatible code path,
2764         which revealed that the code made assumptions about which headers had already been included.
2765
2766         * platform/mac/DisplaySleepDisabler.cpp: Include CoreServices/CoreServices.h to pull in
2767         a declaration of UpdateSystemActivity when targeting Leopard.
2768
2769 2012-07-11  Matt Falkenhagen  <falken@chromium.org>
2770
2771         Add dialog element feature toggle to InternalSettings
2772         https://bugs.webkit.org/show_bug.cgi?id=90934
2773
2774         Reviewed by Hajime Morita.
2775
2776         This enables layout tests to be written for dialog although the feature is disabled by default.
2777
2778         Tests: fast/dom/HTMLDialogElement/dialog-disabled.html
2779                fast/dom/HTMLDialogElement/dialog-enabled.html
2780
2781         * WebCore.exp.in: Added newly exported symbol.
2782         * testing/InternalSettings.cpp:
2783         (WebCore::InternalSettings::Backup::Backup): Backup dialog feature flag.
2784         (WebCore::InternalSettings::Backup::restoreTo): Restore dialog feature flag.
2785         (WebCore::InternalSettings::setDialogElementEnabled): Added.
2786         (WebCore):
2787         * testing/InternalSettings.h: Added support for dialog.
2788         (Backup):
2789         (InternalSettings):
2790         * testing/InternalSettings.idl: Added support for dialog.
2791
2792 2012-07-11  Kent Tamura  <tkent@chromium.org>
2793
2794         Accessing width or height of a detached image input element causes crash
2795         https://bugs.webkit.org/show_bug.cgi?id=90885
2796
2797         Reviewed by Kentaro Hara.
2798
2799         Test: fast/forms/image/width-and-height-of-detached-input.html
2800
2801         * html/ImageInputType.cpp:
2802         (WebCore::ImageInputType::height): Add null check for m_imageLoader.
2803         (WebCore::ImageInputType::width): ditto.
2804
2805 2012-07-11  Kent Tamura  <tkent@chromium.org>
2806
2807         Do not save state of stateless form controls
2808         https://bugs.webkit.org/show_bug.cgi?id=90964
2809
2810         Reviewed by Hajime Morita.
2811
2812         By Bug 89409, we started to store unmodified form control
2813         state. However we don't need to make such state for the following
2814         types.
2815         - password
2816         - submit
2817         - reset
2818         - button
2819         - keygen
2820
2821         Test: fast/forms/state-restore-skip-stateless.html
2822
2823         * html/BaseButtonInputType.cpp:
2824         (WebCore::BaseButtonInputType::shouldSaveAndRestoreFormControlState):
2825         Added. Disable saving state for submit, reset, and button types.
2826         * html/BaseButtonInputType.h:
2827         (BaseButtonInputType): Add shouldSaveAndRestoreFormControlState.
2828         * html/HTMLFormControlElementWithState.h:
2829         (HTMLFormControlElementWithState): Make shouldSaveAndRestoreFormControlState virtual.
2830         * html/HTMLInputElement.cpp:
2831         (WebCore::HTMLInputElement::shouldSaveAndRestoreFormControlState):
2832         Added. Checks InputType::shouldSaveAndRestoreFormControlState first.
2833         * html/HTMLInputElement.h:
2834         (HTMLInputElement): Override shouldSaveAndRestoreFormControlState.
2835         * html/HTMLKeygenElement.cpp:
2836         (WebCore::HTMLKeygenElement::shouldSaveAndRestoreFormControlState)
2837         Added. Disable saving state for <keygen>.:
2838         * html/HTMLKeygenElement.h: Override shouldSaveAndRestoreFormControlState.
2839         * html/InputType.cpp:
2840         (WebCore::InputType::shouldSaveAndRestoreFormControlState):
2841         Added. Enable saving state for all types by default.
2842         * html/InputType.h:
2843         (InputType): Add shouldSaveAndRestoreFormControlState.
2844         * html/PasswordInputType.cpp:
2845         (WebCore::PasswordInputType::shouldSaveAndRestoreFormControlState):
2846         Added. Disabled saving state.
2847         (WebCore::PasswordInputType::saveFormControlState):
2848         Because shouldSaveAndRestoreFormControlState() returns false,
2849         saveFormControlState should be never called.
2850         * html/PasswordInputType.h:
2851         (PasswordInputType): Override shouldSaveAndRestoreFormControlState.
2852
2853 2012-07-11  No'am Rosenthal  <noam.rosenthal@nokia.com>
2854
2855         Compilation error with GLES2 when using gl2ext.h from ANGLE.
2856         https://bugs.webkit.org/show_bug.cgi?id=91030
2857
2858         Reviewed by Kenneth Russell.
2859
2860         gl2ext.h defines PFNGLRENDERBUFFERSTORAGEMULTISAMPLEIMG and not
2861         PFNGLRENDERBUFFERSTORAGEMULTISAMPLEIMGPROC. This is a bug in the original Khronos header,
2862         but we should work around it until fixed headers are integrated.
2863
2864
2865         * platform/graphics/opengl/Extensions3DOpenGLES.h:
2866         (Extensions3DOpenGLES):
2867             Removed the "PROC" suffix to work around the header bug.
2868
2869 2012-07-11  Mark Rowe  <mrowe@apple.com>
2870
2871         Attempt to fix the Chromium Mac build after r122400.
2872
2873         * platform/text/cf/HyphenationCF.cpp: Ensure that Chromium only compiles the body of this file if
2874         building for a new enough version of OS X.
2875
2876 2012-07-11  Kent Tamura  <tkent@chromium.org>
2877
2878         [Mac] Sort project.pbxproj
2879
2880         * WebCore.xcodeproj/project.pbxproj: Apply Tools/Scripts/sort-Xcode-project-file
2881
2882 2012-07-11  Dan Bernstein  <mitz@apple.com>
2883
2884         When a table row height grows because of pagination, not all cells’ heights are adjusted
2885         https://bugs.webkit.org/show_bug.cgi?id=91043
2886
2887         Reviewed by Sam Weinig.
2888
2889         The fix for in bug <http://webkit.org/b/83595> in <http://trac.webkit.org/r113738> made table
2890         rows grow as necessary to fit cells that grow as a result of pagination. But it had two bad
2891         side effects: earlier cells on the row would not grow by the same amount, and later cells on
2892         the row would factor the existing growth into their intrinsic padding.
2893
2894         Test: fast/multicol/table-row-height-increase.html
2895
2896         * rendering/RenderTableSection.cpp:
2897         (WebCore::RenderTableSection::layoutRows): Now when a cell becomes taller than the row height,
2898         the additional height needed is recorded, and the cell is shrunk back to row height. Then
2899         after finishing the row, all cells occurring on the row (including cells spanning it but not
2900         starting on it) are grown by the same amount.
2901
2902 2012-07-11  Mark Rowe  <mrowe@apple.com>
2903
2904         <http://webkit.org/b/91024> Build against the latest SDK when targeting older OS X versions.
2905
2906         Reviewed by Dan Bernstein.
2907
2908         The deployment target is already set to the version that we're targeting, and it's that setting
2909         which determines which functionality from the SDK is available to us.
2910
2911         * Configurations/Base.xcconfig:
2912
2913 2012-07-11  Mark Rowe  <mrowe@apple.com>
2914
2915         Switch a few cases that care about the SDK over to checking __MAC_OS_X_VERSION_MAX_ALLOWED so that things build.
2916
2917         Part of <http://webkit.org/b/91015> Remove BUILDING_ON / TARGETING macros in favor of system availability macros
2918
2919         Reviewed by Filip Pizlo.
2920
2921         * platform/graphics/cg/GraphicsContextCG.cpp: The SDK we're building against determines whether the constant is available.
2922         * platform/mac/EmptyProtocolDefinitions.h:  The SDK we're building against determines whether the protocols are declared.
2923         * platform/mac/NSScrollerImpDetails.h: The SDK we're building against determines whether the constants and method are
2924         available.
2925         * platform/mac/SharedTimerMac.mm: The SDK we're building against determines whether the function has already been declared.
2926
2927 2012-07-11  Mark Rowe  <mrowe@apple.com>
2928
2929         <http://webkit.org/b/91015> Remove BUILDING_ON / TARGETING macros in favor of system availability macros
2930
2931         This removal was handled by a script that translates the relevant macros in to the equivalent checks
2932         using the system availability macros.
2933
2934         Reviewed by Filip Pizlo.
2935
2936         * WebCore.exp.in:
2937         * accessibility/AccessibilityList.h:
2938         * accessibility/AccessibilityTable.h:
2939         * accessibility/mac/AXObjectCacheMac.mm:
2940         * editing/mac/EditorMac.mm:
2941         * loader/MainResourceLoader.cpp:
2942         * loader/MainResourceLoader.h:
2943         * page/AlternativeTextClient.h:
2944         * page/ContextMenuController.cpp:
2945         * page/mac/SettingsMac.mm:
2946         * platform/LocalizedStrings.cpp:
2947         * platform/MemoryPressureHandler.cpp:
2948         * platform/audio/mac/AudioBusMac.mm:
2949         * platform/graphics/Gradient.h:
2950         * platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
2951         * platform/graphics/ca/GraphicsLayerCA.cpp:
2952         * platform/graphics/ca/PlatformCALayer.h:
2953         * platform/graphics/ca/mac/PlatformCALayerMac.mm:
2954         * platform/graphics/ca/mac/TileCache.mm:
2955         * platform/graphics/cg/GraphicsContextCG.cpp:
2956         * platform/graphics/cg/ImageBufferCG.cpp:
2957         * platform/graphics/cg/ImageBufferDataCG.cpp:
2958         * platform/graphics/cg/ImageBufferDataCG.h:
2959         * platform/graphics/cg/ImageCG.cpp:
2960         * platform/graphics/cg/ImageSourceCG.cpp:
2961         * platform/graphics/cg/PathCG.cpp:
2962         * platform/graphics/cocoa/FontPlatformDataCocoa.mm:
2963         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
2964         * platform/graphics/mac/ComplexTextController.cpp:
2965         * platform/graphics/mac/ComplexTextControllerCoreText.mm:
2966         * platform/graphics/mac/FontCacheMac.mm:
2967         * platform/graphics/mac/FontCustomPlatformData.cpp:
2968         * platform/graphics/mac/FontMac.mm:
2969         * platform/graphics/mac/GraphicsContextMac.mm:
2970         * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
2971         * platform/graphics/mac/SimpleFontDataMac.mm:
2972         * platform/graphics/mac/WebLayer.h:
2973         * platform/graphics/mac/WebLayer.mm:
2974         * platform/graphics/opengl/GraphicsContext3DOpenGL.cpp:
2975         * platform/mac/CursorMac.mm:
2976         * platform/mac/DisplaySleepDisabler.cpp:
2977         * platform/mac/DisplaySleepDisabler.h:
2978         * platform/mac/EmptyProtocolDefinitions.h:
2979         * platform/mac/HTMLConverter.h:
2980         * platform/mac/HTMLConverter.mm:
2981         * platform/mac/MemoryPressureHandlerMac.mm:
2982         * platform/mac/NSScrollerImpDetails.h:
2983         * platform/mac/PlatformEventFactoryMac.mm:
2984         * platform/mac/PopupMenuMac.mm:
2985         * platform/mac/ScrollAnimatorMac.mm:
2986         * platform/mac/ScrollElasticityController.mm:
2987         * platform/mac/SharedTimerMac.mm:
2988         * platform/mac/SuddenTermination.mm:
2989         * platform/mac/WebCoreFullScreenWindow.mm:
2990         * platform/mac/WebCoreNSCellExtras.h:
2991         * platform/mac/WebCoreSystemInterface.h:
2992         * platform/mac/WebCoreSystemInterface.mm:
2993         * platform/mac/WebFontCache.mm:
2994         * platform/mac/WebVideoFullscreenController.mm:
2995         * platform/mac/WebVideoFullscreenHUDWindowController.mm:
2996         * platform/network/Credential.h:
2997         * platform/network/ResourceHandle.h:
2998         * platform/network/cf/DNSCFNet.cpp:
2999         * platform/network/cf/ProxyServerCFNet.cpp:
3000         * platform/network/cf/ResourceRequest.h:
3001         * platform/network/cf/ResourceRequestCFNet.cpp:
3002         * platform/network/cf/SocketStreamHandleCFNet.cpp:
3003         * platform/network/mac/AuthenticationMac.mm:
3004         * platform/network/mac/CookieStorageMac.mm:
3005         * platform/network/mac/ResourceHandleMac.mm:
3006         * platform/network/mac/ResourceRequestMac.mm:
3007         * platform/network/mac/WebCoreURLResponse.mm:
3008         * platform/text/TextChecking.h:
3009         * platform/text/cf/HyphenationCF.cpp:
3010         * platform/text/mac/HyphenationMac.mm:
3011         * rendering/RenderLayerBacking.cpp:
3012         * rendering/RenderLayerCompositor.cpp:
3013         * rendering/RenderThemeMac.mm:
3014
3015 2012-07-11  Robert Kroeger  <rjkroege@chromium.org>
3016
3017         Suppress horizontal conversion of PlatformWheelEvents when hasPreciseScrollingDeltas is true
3018         https://bugs.webkit.org/show_bug.cgi?id=89580
3019
3020         WebKit GTK and Chromium Linux force vertical wheel events to
3021         scroll horizontally when over horizontal scroll bars.  This is
3022         undesirable for touchpad scrolling with
3023         hasPreciseScrollingDeltas() == true. Modified shouldTurnVerticalTicksIntoHorizontal
3024         to not perform this conversion for PlatformWheelEvents with preciseScrollingDeltas.
3025
3026         Reviewed by Adam Barth.
3027
3028         Tests: fast/events/touch/gesture/touch-gesture-scroll-sideways.html
3029
3030         * page/EventHandler.cpp:
3031         (WebCore::EventHandler::shouldTurnVerticalTicksIntoHorizontal):
3032         (WebCore::EventHandler::handleWheelEvent):
3033         * page/EventHandler.h:
3034         (EventHandler):
3035         * page/chromium/EventHandlerChromium.cpp:
3036         (WebCore::EventHandler::shouldTurnVerticalTicksIntoHorizontal):
3037         * page/gtk/EventHandlerGtk.cpp:
3038         (WebCore::EventHandler::shouldTurnVerticalTicksIntoHorizontal):
3039
3040 2012-07-11  Hayato Ito  <hayato@chromium.org>
3041
3042         Unreviewed gardening.
3043         Build fix for r122373.
3044
3045         * platform/chromium/support/WebCompositorSharedQuadState.cpp:
3046         (WebKit::WebCompositorSharedQuadState::WebCompositorSharedQuadState):
3047
3048 2012-07-11  Shinya Kawanaka  <shinyak@chromium.org>
3049
3050         SVGUseElement should not use Element::ensureShadowRoot().
3051         https://bugs.webkit.org/show_bug.cgi?id=90938
3052
3053         Reviewed by Hajime Morita.
3054
3055         We would like to remove Element::ensureShadowRoot() because we would like to stabilize the lifecycle of Shadow DOM.
3056         This patch rewrites SVGUseElement::create() not to use Element::ensureShadowRoot(). Since our convention to create
3057         a shadow subtree is to create a method createShadowSubtree(), I obeyed it.
3058
3059         No new tests. Simple refactoring.
3060
3061         * svg/SVGUseElement.cpp:
3062         (WebCore::SVGUseElement::create):
3063         (WebCore::SVGUseElement::createShadowSubtree):
3064         (WebCore):
3065         * svg/SVGUseElement.h:
3066         (SVGUseElement):
3067
3068 2012-07-11  Alexandre Elias  <aelias@google.com>
3069
3070         [chromium] Move compositor quads to Platform/chromium/public
3071         https://bugs.webkit.org/show_bug.cgi?id=90582
3072
3073         Reviewed by Adrienne Walker.
3074
3075         This moves CCSharedQuadState, CCDrawQuad, and all but two CC*DrawQuad
3076         classes to the WebKit namespace, as a first step to pushing them
3077         across the process boundary for the ubercompositor.
3078
3079         - The intent is to serialize the class hierarchy using the same
3080         mechanism as WebInputEvent.  In order to do this, there are three
3081         requirements: pure POD data, a method returning size, and a packing
3082         pragma.
3083
3084         - Private data members are fine with this kind of serializer, and a
3085         default constructor is not needed.  Because of that, we can maintain
3086         the same encapsulation and convenient APIs (behind
3087         WEBKIT_IMPLEMENTATION) as the original classes.  To ease the
3088         transition, the original WebCore headers still exist and typedef to
3089         the new classes.
3090
3091         - However, SharedQuadState will be serialized using the normal
3092         IPC_STRUCT_TRAITS macro, so I made its members public.  A custom
3093         serializer (on quad lists) will maintain the pointers from quads to
3094         SharedQuadStates.
3095
3096         - I converted the Material casting mechanism to materialCast() methods
3097         living in the derived classes.  That way, the WebCompositorQuad header
3098         doesn't need to know about all its derived classes.
3099
3100         - Quad classes not yet transitioned can still be used in
3101         non-ubercompositor mode.  CCRenderPassDrawQuad and CCYUVVideoDrawQuad
3102         are currently non-POD and I left them in their original files.
3103
3104         This approach is the best I've found so far, since it preserves all
3105         WebCore-facing APIs and avoids unnecessary code duplication (new quad
3106         types or members can be added by modifying only one place).  There
3107         also should not be an unreasonable amount of custom serializer code
3108         required.
3109
3110         No new tests. (No-op refactoring.)
3111
3112         * WebCore.gypi:
3113         * platform/chromium/support/WebCompositorQuad.cpp: Added.
3114         (WebKit):
3115         (WebKit::WebCompositorQuad::WebCompositorQuad):
3116         (WebKit::WebCompositorQuad::opaqueRect):
3117         (WebKit::WebCompositorQuad::setQuadVisibleRect):
3118         (WebKit::WebCompositorQuad::size):
3119         * platform/chromium/support/WebCompositorSharedQuadState.cpp: Renamed from Source/WebCore/platform/graphics/chromium/cc/CCSharedQuadState.cpp.
3120         (WebKit):
3121         (WebKit::WebCompositorSharedQuadState::WebCompositorSharedQuadState):
3122         (WebKit::WebCompositorSharedQuadState::create):
3123         (WebKit::WebCompositorSharedQuadState::isLayerAxisAlignedIntRect):
3124         * platform/chromium/support/WebCompositorSolidColorQuad.cpp: Renamed from Source/WebCore/platform/graphics/chromium/cc/CCSolidColorDrawQuad.cpp.
3125         (WebKit):
3126         (WebKit::WebCompositorSolidColorQuad::create):
3127         (WebKit::WebCompositorSolidColorQuad::WebCompositorSolidColorQuad):
3128         (WebKit::WebCompositorSolidColorQuad::materialCast):
3129         * platform/chromium/support/WebCompositorTextureQuad.cpp: Renamed from Source/WebCore/platform/graphics/chromium/cc/CCTextureDrawQuad.cpp.
3130         (WebKit):
3131         (WebKit::WebCompositorTextureQuad::create):
3132         (WebKit::WebCompositorTextureQuad::WebCompositorTextureQuad):
3133         (WebKit::WebCompositorTextureQuad::setNeedsBlending):
3134         (WebKit::WebCompositorTextureQuad::materialCast):
3135         * platform/graphics/chromium/LayerRendererChromium.cpp:
3136         (WebCore::LayerRendererChromium::drawQuad):
3137         * platform/graphics/chromium/LayerRendererChromium.h:
3138         * platform/graphics/chromium/cc/CCCheckerboardDrawQuad.cpp:
3139         (WebCore::CCCheckerboardDrawQuad::create):
3140         (WebCore::CCCheckerboardDrawQuad::CCCheckerboardDrawQuad):
3141         (WebCore::CCCheckerboardDrawQuad::materialCast):
3142         (WebCore):
3143         * platform/graphics/chromium/cc/CCCheckerboardDrawQuad.h:
3144         (CCCheckerboardDrawQuad):
3145         * platform/graphics/chromium/cc/CCDebugBorderDrawQuad.cpp:
3146         (WebCore::CCDebugBorderDrawQuad::create):
3147         (WebCore::CCDebugBorderDrawQuad::CCDebugBorderDrawQuad):
3148         (WebCore::CCDebugBorderDrawQuad::materialCast):
3149         (WebCore):
3150         * platform/graphics/chromium/cc/CCDebugBorderDrawQuad.h:
3151         (CCDebugBorderDrawQuad):
3152         * platform/graphics/chromium/cc/CCDrawQuad.cpp: Removed.
3153         * platform/graphics/chromium/cc/CCDrawQuad.h:
3154         (WebCore):
3155         * platform/graphics/chromium/cc/CCIOSurfaceDrawQuad.cpp:
3156         (WebCore::CCIOSurfaceDrawQuad::create):
3157         (WebCore::CCIOSurfaceDrawQuad::CCIOSurfaceDrawQuad):
3158         (WebCore::CCIOSurfaceDrawQuad::materialCast):
3159         (WebCore):
3160         * platform/graphics/chromium/cc/CCIOSurfaceDrawQuad.h:
3161         (CCIOSurfaceDrawQuad):
3162         * platform/graphics/chromium/cc/CCLayerImpl.h:
3163         (WebCore):
3164         * platform/graphics/chromium/cc/CCRenderPass.h:
3165         (WebCore):
3166         * platform/graphics/chromium/cc/CCRenderPassDrawQuad.cpp:
3167         (WebCore::CCRenderPassDrawQuad::create):
3168         (WebCore::CCRenderPassDrawQuad::CCRenderPassDrawQuad):
3169         (WebCore::CCRenderPassDrawQuad::materialCast):
3170         (WebCore):
3171         * platform/graphics/chromium/cc/CCRenderPassDrawQuad.h:
3172         (CCRenderPassDrawQuad):
3173         * platform/graphics/chromium/cc/CCRenderSurface.h:
3174         (WebCore):
3175         * platform/graphics/chromium/cc/CCSharedQuadState.h:
3176         (WebCore):
3177         * platform/graphics/chromium/cc/CCSolidColorDrawQuad.h:
3178         (WebCore):
3179         * platform/graphics/chromium/cc/CCStreamVideoDrawQuad.cpp:
3180         (WebCore::CCStreamVideoDrawQuad::create):
3181         (WebCore::CCStreamVideoDrawQuad::CCStreamVideoDrawQuad):
3182         (WebCore::CCStreamVideoDrawQuad::materialCast):
3183         (WebCore):
3184         * platform/graphics/chromium/cc/CCStreamVideoDrawQuad.h:
3185         (CCStreamVideoDrawQuad):
3186         * platform/graphics/chromium/cc/CCTextureDrawQuad.h:
3187         (WebCore):
3188         * platform/graphics/chromium/cc/CCTileDrawQuad.cpp:
3189         (WebCore::CCTileDrawQuad::create):
3190         (WebCore::CCTileDrawQuad::CCTileDrawQuad):
3191         (WebCore::CCTileDrawQuad::materialCast):
3192         (WebCore):
3193         * platform/graphics/chromium/cc/CCTileDrawQuad.h:
3194         (CCTileDrawQuad):
3195         * platform/graphics/chromium/cc/CCYUVVideoDrawQuad.cpp:
3196         (WebCore::CCYUVVideoDrawQuad::create):
3197         (WebCore::CCYUVVideoDrawQuad::CCYUVVideoDrawQuad):
3198         (WebCore::CCYUVVideoDrawQuad::materialCast):
3199         (WebCore):
3200         * platform/graphics/chromium/cc/CCYUVVideoDrawQuad.h:
3201         (CCYUVVideoDrawQuad):
3202
3203 2012-07-11  Robert Kroeger  <rjkroege@chromium.org>
3204
3205         [chromium] [regression] Don't use ScrollByPrecisePixels on Chromium Mac.
3206         https://bugs.webkit.org/show_bug.cgi?id=91020
3207
3208         Reviewed by Adam Barth.
3209
3210         A change in https://bugs.webkit.org/show_bug.cgi?id=87535 to
3211         improve the operation of smooth scrolling incorrectly caused
3212         Chromium Mac to use the wrong scroll granularity on
3213         hasPreciseScrollingDeltas() == true wheelevent scrolls.
3214         Exclude the change on the Chromium Mac platform.
3215
3216         * platform/ScrollAnimator.cpp:
3217         (WebCore::ScrollAnimator::handleWheelEvent): Adjusted #ifdef to exclude Chromium
3218         Mac from ScrollByPrecisePixels change.
3219
3220 2012-07-11  Jia Pu  <jpu@apple.com>
3221
3222         On Mac, autocorrection sometimes fails to take place in Safari.
3223         https://bugs.webkit.org/show_bug.cgi?id=89982
3224
3225         Reviewed by Ryosuke Niwa.
3226
3227         We should check the value of shouldCheckForCorrection, not shouldShowCorrectionPanel, to determine if we should
3228         early return in markAndReplaceFor(). Also, since we don't want autocorrection to happen when we only change selection,
3229         not the content, so we shouldn't set TextCheckingTypeCorrection flag in markMisspellingsAndBadGrammar() as this function
3230         is triggered only by selection change.
3231
3232         * editing/Editor.cpp:
3233         (WebCore::Editor::markAndReplaceFor):
3234         (WebCore::Editor::markMisspellingsAndBadGrammar):
3235
3236 2012-07-11  Sheriff Bot  <webkit.review.bot@gmail.com>
3237
3238         Unreviewed, rolling out r122361.
3239         http://trac.webkit.org/changeset/122361
3240         https://bugs.webkit.org/show_bug.cgi?id=91027
3241
3242         Broke Windows build and fast/forms/state-restore-broken-
3243         state.html (Requested by rniwa on #webkit).
3244
3245         * testing/Internals.idl:
3246
3247 2012-07-10  Simon Fraser  <simon.fraser@apple.com>
3248
3249         Assertion ASSERTION FAILED: enclosingIntRect(rendererMappedResult) == enclosingIntRect(FloatQuad(result).boundingBox()) when compositing in paginated mode
3250         https://bugs.webkit.org/show_bug.cgi?id=90919
3251
3252         Reviewed by Antti Koivisto.
3253
3254         r121124 added a fast path for geometry mapping that goes via layers
3255         when possible. However, this broke paginated pages, which put
3256         the root (RenderView) layer into column mode, because it failed
3257         to check for columns on the ancestor layer.
3258
3259         Rather than make a risky change to convertToLayerCoords(), add a local
3260         function canMapViaLayer(), which is like RenderLayer::canUseConvertToLayerCoords(),
3261         but doesn't check for compositing (compositing itself is not a reason
3262         to avoid convertToLayerCoords). Call canMapViaLayer() with the ancestorLayer
3263         to check whether the ancestor has columns, which fixes the bug.
3264
3265         Test: compositing/columns/geometry-map-paginated-assert.html
3266
3267         * rendering/RenderGeometryMap.cpp:
3268         (WebCore::canMapViaLayer):
3269         (WebCore::RenderGeometryMap::pushMappingsToAncestor):
3270
3271 2012-07-11  Dana Jansens  <danakj@chromium.org>
3272
3273         [chromium] Rename layerRect to contentRect for rects that live in content space
3274         https://bugs.webkit.org/show_bug.cgi?id=90843
3275
3276         Reviewed by Adrienne Walker.
3277
3278         Dropped the layerTransform() from CCSharedQuadState, as nothing should be
3279         using it to draw with. RenderPasses need a weird drawTransform right now
3280         which was stored in layerTransform, so moved this to the RenderPass quad.
3281
3282         No new tests. No change in behaviour.
3283
3284         * platform/graphics/chromium/CanvasLayerTextureUpdater.cpp:
3285         (WebCore::CanvasLayerTextureUpdater::paintContents):
3286         * platform/graphics/chromium/ContentLayerChromium.cpp:
3287         (WebCore::ContentLayerChromium::update):
3288         (WebCore::ContentLayerChromium::needMoreUpdates):
3289         * platform/graphics/chromium/ImageLayerChromium.cpp:
3290         (WebCore::ImageLayerChromium::update):
3291         * platform/graphics/chromium/LayerChromium.cpp:
3292         (WebCore::LayerChromium::visibleContentOpaqueRegion):
3293         * platform/graphics/chromium/LayerChromium.h:
3294         (WebCore::LayerChromium::visibleContentRect):
3295         (WebCore::LayerChromium::setVisibleContentRect):
3296         (LayerChromium):
3297         * platform/graphics/chromium/LayerRendererChromium.cpp:
3298         (WebCore::LayerRendererChromium::drawRenderPassQuad):
3299         (WebCore::LayerRendererChromium::drawTileQuad):
3300         (WebCore::LayerRendererChromium::drawYUVVideoQuad):
3301         (WebCore::LayerRendererChromium::drawStreamVideoQuad):
3302         (WebCore::LayerRendererChromium::drawIOSurfaceQuad):
3303         * platform/graphics/chromium/TiledLayerChromium.cpp:
3304         (WebCore::TiledLayerChromium::updateBounds):
3305         (WebCore::TiledLayerChromium::setNeedsDisplayRect):
3306         (WebCore::TiledLayerChromium::invalidateContentRect):
3307         (WebCore::TiledLayerChromium::updateTiles):
3308         (WebCore::TiledLayerChromium::setTexturePriorities):
3309         (WebCore::TiledLayerChromium::setTexturePrioritiesInRect):
3310         (WebCore::TiledLayerChromium::visibleContentOpaqueRegion):
3311         (WebCore::TiledLayerChromium::updateContentRect):
3312         (WebCore::TiledLayerChromium::needsIdlePaint):
3313         (WebCore::TiledLayerChromium::idlePaintRect):
3314         * platform/graphics/chromium/TiledLayerChromium.h:
3315         (TiledLayerChromium):
3316         * platform/graphics/chromium/cc/CCDrawQuad.h:
3317         (WebCore::CCDrawQuad::visibleContentRect):
3318         * platform/graphics/chromium/cc/CCLayerImpl.cpp:
3319         (WebCore::CCLayerImpl::createSharedQuadState):
3320         (WebCore::CCLayerImpl::appendDebugBorderQuad):
3321         (WebCore::CCLayerImpl::visibleContentOpaqueRegion):
3322         * platform/graphics/chromium/cc/CCLayerImpl.h:
3323         (WebCore::CCLayerImpl::visibleContentRect):
3324         (WebCore::CCLayerImpl::setVisibleContentRect):
3325         (CCLayerImpl):
3326         * platform/graphics/chromium/cc/CCLayerTilingData.cpp:
3327         (WebCore::CCLayerTilingData::contentRectToTileIndices):
3328         (WebCore::CCLayerTilingData::opaqueRegionInContentRect):
3329         (WebCore::CCLayerTilingData::setBounds):
3330         * platform/graphics/chromium/cc/CCLayerTilingData.h:
3331         (CCLayerTilingData):
3332         * platform/graphics/chromium/cc/CCLayerTreeHostCommon.cpp:
3333         (WebCore::calculateVisibleContentRect):
3334         (WebCore::layerShouldBeSkipped):
3335         (WebCore):
3336         (WebCore::calculateVisibleAndScissorRectsInternal):
3337         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
3338         (WebCore::CCLayerTreeHostImpl::calculateRenderPasses):
3339         * platform/graphics/chromium/cc/CCOcclusionTracker.cpp:
3340         (WebCore::addOcclusionBehindLayer):
3341         * platform/graphics/chromium/cc/CCRenderPassDrawQuad.cpp:
3342         (WebCore::CCRenderPassDrawQuad::create):
3343         (WebCore::CCRenderPassDrawQuad::CCRenderPassDrawQuad):
3344         * platform/graphics/chromium/cc/CCRenderPassDrawQuad.h:
3345         (CCRenderPassDrawQuad):
3346         (WebCore::CCRenderPassDrawQuad::drawTransform):
3347         * platform/graphics/chromium/cc/CCRenderSurface.cpp:
3348         (WebCore::CCRenderSurface::createSharedQuadState):
3349         (WebCore::CCRenderSurface::createReplicaSharedQuadState):
3350         (WebCore::CCRenderSurface::appendQuads):
3351         * platform/graphics/chromium/cc/CCSharedQuadState.cpp:
3352         (WebCore::CCSharedQuadState::create):
3353         (WebCore::CCSharedQuadState::CCSharedQuadState):
3354         (WebCore::CCSharedQuadState::isLayerAxisAlignedIntRect):
3355         * platform/graphics/chromium/cc/CCSharedQuadState.h:
3356         (CCSharedQuadState):
3357         (WebCore::CCSharedQuadState::visibleContentRect):
3358         * platform/graphics/chromium/cc/CCTiledLayerImpl.cpp:
3359         (WebCore::CCTiledLayerImpl::appendQuads):
3360         (WebCore::CCTiledLayerImpl::visibleContentOpaqueRegion):
3361
3362 2012-07-11  Chris Fleizach  <cfleizach@apple.com>
3363
3364         ARIA spinbutton role incorrectly mapped to ProgressIndicatorRole
3365         https://bugs.webkit.org/show_bug.cgi?id=77298
3366
3367         Reviewed by Anders Carlsson.
3368
3369         * accessibility/AccessibilityObject.cpp:
3370         (WebCore::createARIARoleMap):
3371
3372 2012-07-11  Huang Dongsung  <luxtella@company100.net>
3373
3374         BitmapImage::frameIsCompleteAtIndex() must return false if ImageDecoder is not initialized.
3375         https://bugs.webkit.org/show_bug.cgi?id=90757
3376
3377         Reviewed by Simon Fraser.
3378
3379         The current code fortunately has worked so far because only
3380         BitmapImage::startAnimation calls frameIsCompleteAtIndex, and startAnimation
3381         cannot call frameIsCompleteAtIndex if ImageDecoder is not yet initialized.
3382         startAnimation returns at the first line becase shouldAnimate() always return
3383         false in this case.
3384
3385             if (m_frameTimer || !shouldAnimate() || frameCount() <= 1)
3386                 return;
3387
3388         This change is needed because parallel image decoders call
3389         BitmapImage::frameIsCompleteAtIndex in other places too.
3390
3391         No new tests, manually tested whether the caller exists or not that
3392         calls BitmapImage::frameIsCompleteAtIndex() before ImageDecoder is
3393         initialized.
3394
3395         * platform/graphics/BitmapImage.cpp:
3396         (WebCore::BitmapImage::frameIsCompleteAtIndex):
3397
3398 2012-07-11  Kevin Ollivier  <kevino@theolliviers.com>
3399
3400         [wx] Unreviewed build fix. Update code to use the new constant name.