Source/WebCore/ChangeLog

   1 2012-07-05  Robert Hogan  <robert@webkit.org>
   2
   3         CSS 2.1 failure: vertical-align-boxes-001 fails
   4         https://bugs.webkit.org/show_bug.cgi?id=90626
   5
   6         Reviewed by Eric Seidel.
   7
   8         Tests: css2.1/20110323/vertical-align-boxes-001.htm
   9
  10         A percentage value vertical-align is always a percentage of the actual line-height rather than
  11         the margin box per http://www.w3.org/TR/CSS21/visudet.html#propdef-vertical-align: 'Percentages:
  12         refer to the 'line-height' of the element itself'.  Confusingly, RenderBox::lineheight() is a
  13         shorthand into the dimensions of the margin box for replaced elements in the other vertical-align
  14         cases, i.e. where it's the margin box that's relevant rather than the 'line-height'. So rather than patch RenderBox's
  15         lineHeight() to somehow consider the percentage cases, just give percentage vertical-align the full computedLineHeight()
  16         rather than lineHeight()'s margin box.
  17
  18         * rendering/RootInlineBox.cpp:
  19         (WebCore::RootInlineBox::verticalPositionForBox):
  20
  21 2012-07-13  Ryosuke Niwa  <rniwa@webkit.org>
  22
  23         Iterating backwards over HTMLCollection is O(n^2)
  24         https://bugs.webkit.org/show_bug.cgi?id=91306
  25
  26         Reviewed by Anders Carlsson.
  27
  28         Fixed the bug by introducing itemBefore that iterates nodes backwards to complement itemAfter.
  29         Unfortunately, some HTML collections such as HTMLFormCollection and HTMLTableRowsCollection have
  30         its own itemAfter function and writing an equivalent itemBefore is somewhat tricky. For now,
  31         added a new boolean flag indicating whether a given HTML collection supports itemBefore or not,
  32         and left those HTML collections that override itemAfter alone.
  33
  34         This also paves our way to share more code between DynamicNodeList and HTMLCollection.
  35
  36         Test: perf/htmlcollection-backwards-iteration.html
  37
  38         * dom/DynamicNodeList.h:
  39         (WebCore::DynamicNodeListCacheBase::DynamicNodeListCacheBase): Takes ItemBeforeSupportType.
  40         (WebCore::DynamicNodeListCacheBase::supportsItemBefore): Added.
  41         (DynamicNodeListCacheBase):
  42         (WebCore::DynamicNodeListCacheBase::setItemCache): Replaced a FIXME by an assertion now that
  43         we can.
  44         * html/HTMLAllCollection.cpp:
  45         (WebCore::HTMLAllCollection::HTMLAllCollection): Supports itemBefore since it doesn't override
  46         itemAfter.
  47         * html/HTMLCollection.cpp:
  48         (WebCore::HTMLCollection::HTMLCollection):
  49         (WebCore::HTMLCollection::create):
  50         (WebCore::isAcceptableElement): Made it a static local function instead of a static member.
  51         (WebCore::nextNode): Templatized.
  52         (WebCore::itemBeforeOrAfter): Extracted from itemAfter and templatized.
  53         (WebCore::HTMLCollection::itemBefore): Added.
  54         (WebCore::HTMLCollection::itemAfter):
  55         (WebCore::HTMLCollection::shouldSearchFromFirstItem): Added. Determines whether we should reset
  56         the item cache to the first item. We obviously do if the cache is invalid. If the target offset
  57         is after the cached offset, then we shouldn't go back regardless of availability of itemBefore.
  58         Otherwise, we go back to the first item iff itemBefore is not available or the distance from
  59         the cached offset to the target offset is greater than the target offset itself.
  60         (WebCore::HTMLCollection::length):
  61         (WebCore::HTMLCollection::item): Use the term "offset" to match the terminology elsewhere.
  62         (WebCore::HTMLCollection::itemBeforeOrAfterCachedItem): Ditto. Also added the logic to iterate
  63         nodes backwards using itemBefore. Once we're in this branch, we should always find a matching
  64         item since the target offset was less than the cached offset, and offsets are non-negative.
  65         If we had ever reached the end of the loop without finding an item, it indicates that the cache
  66         has been invalid and we have some serious bug elsewhere.
  67         * html/HTMLCollection.h:
  68         (WebCore::HTMLCollectionCacheBase::HTMLCollectionCacheBase):
  69         (HTMLCollection):
  70         * html/HTMLOptionsCollection.cpp:
  71         (WebCore::HTMLOptionsCollection::HTMLOptionsCollection): Supports itemBefore since it doesn't
  72         override itemAfter.
  73         * html/HTMLFormCollection.cpp:
  74         (WebCore::HTMLFormCollection::HTMLFormCollection): Doesn't support itemBefore as it overrides
  75         itemAfter.
  76         * html/HTMLNameCollection.cpp:
  77         (WebCore::HTMLNameCollection::HTMLNameCollection): Ditto.
  78         * html/HTMLPropertiesCollection.cpp:
  79         (WebCore::HTMLPropertiesCollection::HTMLPropertiesCollection):
  80         * html/HTMLTableRowsCollection.cpp:
  81         (WebCore::HTMLTableRowsCollection::HTMLTableRowsCollection):
  82
  83 2012-07-13  Eric Penner  <epenner@google.com>
  84
  85         [chromium] Add 'self-managed' option to CCPrioritizedTexture to enable render-surface and canvas use cases.
  86         https://bugs.webkit.org/show_bug.cgi?id=91177
  87
  88         Reviewed by Adrienne Walker.
  89
  90         This makes the render-surface memory use case generic as 'self-managed' textures,
  91         as this use case is popping up in other places (eg. canvases). It's exactly the
  92         same idea except we can have as many place-holders as we want at arbitrary
  93         priorities.
  94
  95         This already tested by the render surface unit tests which now also use the
  96         generic placeholder.
  97
  98         * platform/graphics/chromium/cc/CCLayerTreeHost.cpp:
  99         (WebCore::CCLayerTreeHost::CCLayerTreeHost):
 100         (WebCore::CCLayerTreeHost::initializeLayerRenderer):
 101         (WebCore::CCLayerTreeHost::updateLayers):
 102         (WebCore::CCLayerTreeHost::setPrioritiesForSurfaces):
 103         (WebCore):
 104         (WebCore::CCLayerTreeHost::setPrioritiesForLayers):
 105         (WebCore::CCLayerTreeHost::prioritizeTextures):
 106         (WebCore::CCLayerTreeHost::calculateMemoryForRenderSurfaces):
 107         (WebCore::CCLayerTreeHost::paintLayerContents):
 108         * platform/graphics/chromium/cc/CCLayerTreeHost.h:
 109         (CCLayerTreeHost):
 110         * platform/graphics/chromium/cc/CCPrioritizedTexture.cpp:
 111         (WebCore::CCPrioritizedTexture::CCPrioritizedTexture):
 112         (WebCore::CCPrioritizedTexture::setToSelfManagedMemoryPlaceholder):
 113         * platform/graphics/chromium/cc/CCPrioritizedTexture.h:
 114         (CCPrioritizedTexture):
 115         (WebCore::CCPrioritizedTexture::setIsSelfManaged):
 116         (WebCore::CCPrioritizedTexture::isSelfManaged):
 117         * platform/graphics/chromium/cc/CCPrioritizedTextureManager.cpp:
 118         (WebCore::CCPrioritizedTextureManager::prioritizeTextures):
 119         (WebCore::CCPrioritizedTextureManager::acquireBackingTextureIfNeeded):
 120         (WebCore::CCPrioritizedTextureManager::destroyBacking):
 121         * platform/graphics/chromium/cc/CCPrioritizedTextureManager.h:
 122         (CCPrioritizedTextureManager):
 123         (WebCore::CCPrioritizedTextureManager::memoryForSelfManagedTextures):
 124
 125 2012-07-13  Kent Tamura  <tkent@chromium.org>
 126
 127         Internals: Clean up the mock PagePopupDriver correctly.
 128         https://bugs.webkit.org/show_bug.cgi?id=91250
 129
 130         Unreviewed, a trivial testing code fix.
 131
 132         * testing/InternalSettings.cpp:
 133         (WebCore::InternalSettings::Backup::restoreTo):
 134         (WebCore::InternalSettings::reset):
 135         Resetting PaePopupDriver here instead of Backup::restoreTo.
 136         Also, close the mock popup before resetting PagePopupDriver by clearing m_pagePopupDriver.
 137         * testing/MockPagePopupDriver.cpp:
 138         (WebCore::MockPagePopupDriver::~MockPagePopupDriver):
 139         Close the popup.
 140
 141 2012-07-13  Tony Payne  <tpayne@chromium.org>
 142
 143         Remove Widget from screenColorProfile
 144         https://bugs.webkit.org/show_bug.cgi?id=91300
 145
 146         Reviewed by Adam Barth.
 147
 148         Chromium, the only platform implementing screenColorProfile, does not
 149         need the Widget, so removing for simplicity.
 150
 151         Covered by existing tests.
 152
 153         * platform/PlatformScreen.h:
 154         (WebCore): Updated comment to remove reference to type param that no
 155         longer exists and removed Widget param.
 156         * platform/blackberry/PlatformScreenBlackBerry.cpp:
 157         (WebCore::screenColorProfile): Removed widget param.
 158         * platform/chromium/PlatformScreenChromium.cpp:
 159         (WebCore::screenColorProfile): Removed widget param.
 160         * platform/efl/PlatformScreenEfl.cpp:
 161         (WebCore::screenColorProfile): Removed widget param.
 162         * platform/gtk/PlatformScreenGtk.cpp:
 163         (WebCore::screenColorProfile): Removed widget param.
 164         * platform/image-decoders/ImageDecoder.h:
 165         (WebCore::ImageDecoder::qcmsOutputDeviceProfile): removed param to
 166         match screenColorProfile()'s new spec.
 167         * platform/mac/PlatformScreenMac.mm:
 168         (WebCore::screenColorProfile): Removed widget param.
 169         * platform/qt/PlatformScreenQt.cpp:
 170         (WebCore::screenColorProfile): Removed widget param.
 171         * platform/win/PlatformScreenWin.cpp:
 172         (WebCore::screenColorProfile): Removed widget param.
 173
 174 2012-07-13  Brian Anderson  <brianderson@chromium.org>
 175
 176         [chromium] Add flushes to CCTextureUpdater::update
 177         https://bugs.webkit.org/show_bug.cgi?id=89035
 178
 179         Reviewed by Adrienne Walker.
 180
 181         Automatic flushes are being removed from the command buffer, so
 182         this moves the flushes into the CCTextureUpdater itself.
 183
 184         CCTextureUpdaterTest added to verify texture upload/flushing patterns.
 185
 186         * platform/graphics/chromium/cc/CCGraphicsContext.h:
 187         (WebCore::CCGraphicsContext::flush):
 188         (CCGraphicsContext):
 189         * platform/graphics/chromium/cc/CCTextureUpdater.cpp:
 190         (WebCore):
 191         (WebCore::CCTextureUpdater::update): Manual flushes added here.
 192
 193 2012-07-13  Kiran Muppala  <cmuppala@apple.com>
 194
 195         REGRESSION: RenderInline boundingBox ignores relative position offset
 196         https://bugs.webkit.org/show_bug.cgi?id=91168
 197
 198         Reviewed by Simon Fraser.
 199
 200         RenderGeometryMap, used for caching the transform to the view,
 201         expects the first mapping pushed, to be that of the view itself.
 202         RenderInline was instead pushing it's own offset first.  Besides
 203         the offset of the view itself was not being pushed.
 204
 205         Relaxed the RenderGeometryMap restriction that the first pushed
 206         step should be of the view.  It is sufficient that the view's mapping
 207         is pushed in the first call to pushMappingsToAncestor.  Modified
 208         RenderInline to push the offset of the view also to the geometry map.
 209
 210         Test: fast/inline/inline-relative-offset-boundingbox.html
 211
 212         * rendering/RenderGeometryMap.cpp:
 213         (WebCore::RenderGeometryMap::pushMappingsToAncestor): Add assertion to
 214         check if mapping to view was pushed in first invocation.
 215         (WebCore::RenderGeometryMap::pushView): Correct assertion that checks
 216         if the view's mapping is the first one to be applied.
 217         (WebCore::RenderGeometryMap::stepInserted): Use isRenderView to check if
 218         a mapping step belongs to a view instead of using mapping size.
 219         (WebCore::RenderGeometryMap::stepRemoved): Ditto.
 220         * rendering/RenderInline.cpp:
 221         (WebCore::(anonymous namespace)::AbsoluteQuadsGeneratorContext::AbsoluteQuadsGeneratorContext):
 222         Push mappings all the way up to and including the view.
 223
 224 2012-07-13  Xianzhu Wang  <wangxianzhu@chromium.org>
 225
 226         Move WebCore/platform/text/Base64 to WTF/wtf/text
 227         https://bugs.webkit.org/show_bug.cgi?id=91162
 228
 229         Reviewed by Adam Barth.
 230
 231         No new tests. Files moving only.
 232
 233         * CMakeLists.txt:
 234         * GNUmakefile.list.am:
 235         * Modules/websockets/WebSocketHandshake.cpp:
 236         (WebCore::generateSecWebSocketKey):
 237         (WebCore::WebSocketHandshake::getExpectedWebSocketAccept):
 238         * Target.pri:
 239         * WebCore.gypi:
 240         * WebCore.order:
 241         * WebCore.vcproj/WebCore.vcproj:
 242         * WebCore.xcodeproj/project.pbxproj:
 243         * fileapi/FileReaderLoader.cpp:
 244         (WebCore::FileReaderLoader::convertToDataURL):
 245         * inspector/DOMPatchSupport.cpp:
 246         (WebCore::DOMPatchSupport::createDigest):
 247         * inspector/InspectorFileSystemAgent.cpp:
 248         (WebCore):
 249         * inspector/InspectorPageAgent.cpp:
 250         (WebCore::InspectorPageAgent::cachedResourceContent):
 251         (WebCore::InspectorPageAgent::sharedBufferContent):
 252         * loader/archive/mhtml/MHTMLArchive.cpp:
 253         (WebCore::MHTMLArchive::generateMHTMLData):
 254         * loader/archive/mhtml/MHTMLParser.cpp:
 255         (WebCore::MHTMLParser::parseNextPart):
 256         * page/DOMWindow.cpp:
 257         (WebCore::DOMWindow::btoa):
 258         (WebCore::DOMWindow::atob):
 259         * page/Page.cpp:
 260         (WebCore::Page::userStyleSheetLocationChanged):
 261         * platform/graphics/cairo/ImageBufferCairo.cpp:
 262         (WebCore::ImageBuffer::toDataURL):
 263         * platform/graphics/cg/ImageBufferCG.cpp:
 264         (WebCore::CGImageToDataURL):
 265         * platform/graphics/gtk/ImageBufferGtk.cpp:
 266         (WebCore::ImageBuffer::toDataURL):
 267         * platform/graphics/skia/FontCustomPlatformData.cpp:
 268         (WebCore::createUniqueFontName):
 269         * platform/graphics/skia/ImageBufferSkia.cpp:
 270         (WebCore::ImageBuffer::toDataURL):
 271         (WebCore::ImageDataToDataURL):
 272         * platform/graphics/win/FontCustomPlatformData.cpp:
 273         (WebCore::createUniqueFontName):
 274         * platform/graphics/wince/FontCustomPlatformData.cpp:
 275         (WebCore::createUniqueFontName):
 276         * platform/graphics/wince/ImageBufferWinCE.cpp:
 277         * platform/graphics/wx/ImageBufferWx.cpp:
 278         * platform/network/DataURL.cpp:
 279         (WebCore::handleDataURL):
 280         * platform/network/cf/ResourceHandleCFNet.cpp:
 281         (WebCore::applyBasicAuthorizationHeader):
 282         * platform/network/mac/ResourceHandleMac.mm:
 283         (WebCore::applyBasicAuthorizationHeader):
 284         * platform/network/soup/ResourceHandleSoup.cpp:
 285         * platform/win/SSLKeyGeneratorWin.cpp:
 286         (WebCore::WebCore::signedPublicKeyAndChallengeString):
 287
 288 2012-07-13  Xianzhu Wang  <wangxianzhu@chromium.org>
 289
 290         [Chromium] Sometimes bottom of text is truncated when page has a fractional scale
 291         https://bugs.webkit.org/show_bug.cgi?id=88684
 292
 293         Reviewed by Tony Chang.
 294
 295         When the page has a fractional scale, the ascent and descent part of the fonts might be fractional.
 296         If the descent part is rounded down, the bottom of the text might be truncated when displayed
 297         when subpixel text positioning is enabled.
 298         To avoid that, borrow one unit from the ascent when possible.
 299
 300         Test: fast/text/descent-clip-in-scaled-page.html
 301
 302         * platform/graphics/harfbuzz/FontPlatformDataHarfBuzz.cpp:
 303         (WebCore::FontPlatformData::setupPaint): Moved NoPreference handling into querySystemForRenderStyle so that fontRenderStyle() can have actual styles without NoPreference.
 304         (WebCore::FontPlatformData::querySystemForRenderStyle): Added NoPreference handling (moved from setupPaint)
 305         * platform/graphics/harfbuzz/FontPlatformDataHarfBuzz.h:
 306         (FontPlatformData):
 307         (WebCore::FontPlatformData::fontRenderStyle): Added to let SimpleFontDataSkia access the font render styles.
 308         * platform/graphics/skia/SimpleFontDataSkia.cpp:
 309         (WebCore::SimpleFontData::platformInit):
 310
 311 2012-07-13  Ryosuke Niwa  <rniwa@webkit.org>
 312
 313         Remove an assertion after r122637.
 314
 315         * dom/DynamicNodeList.h:
 316         (WebCore::DynamicNodeListCacheBase::shouldInvalidateTypeOnAttributeChange):
 317
 318 2012-07-13  Pierre Rossi  <pierre.rossi@gmail.com>
 319
 320         [Qt] Improve the mobile theme slightly
 321         https://bugs.webkit.org/show_bug.cgi?id=90806
 322
 323         Reviewed by Kenneth Rohde Christiansen.
 324
 325         Improve drawing of the mobile theme's controls' background.
 326
 327         Ensure the focus ring never appears with the mobile theme, since it
 328         looks bad in combination with the highlights.
 329
 330         No new tests. The painting code from the mobile theme is still
 331         not covered specifically (it will when we revive pixel tests).
 332
 333         * platform/qt/RenderThemeQtMobile.cpp:
 334         (WebCore):
 335         (WebCore::addPointToOctants): Added. This is simply a helper to avoid
 336             doing too much duplicate work in drawControlBackground.
 337         (WebCore::drawControlBackground): Rely on the octant logic added above
 338             and take the opportunity to increase the granularity.
 339         (WebCore::borderPen):
 340         (WebCore::StylePainterMobile::findLineEdit):
 341         (WebCore::RenderThemeQtMobile::adjustTextFieldStyle):
 342         * platform/qt/RenderThemeQtMobile.h:
 343         (RenderThemeQtMobile):
 344         (WebCore::RenderThemeQtMobile::supportsFocusRing):
 345
 346 2012-07-13  Julien Chaffraix  <jchaffraix@webkit.org>
 347
 348         Remove an always-failing table-wrapping check in RenderObject::addChild
 349         https://bugs.webkit.org/show_bug.cgi?id=91286
 350
 351         Reviewed by Eric Seidel.
 352
 353         Due to the structure of the code, this test is always failing (newChild->isTableCell()
 354         is true to get in the branch).
 355
 356         The changeset adding the code didn't add testing so I poundered adding the mentioned test,
 357         which is passing. However the test would need to be blindly changed to be included in our
 358         test harness. I would also expect this code to be exercised by other table tests anyway.
 359
 360         * rendering/RenderObject.cpp:
 361         (WebCore::RenderObject::addChild):
 362         Removed never-reached branch. While at it, removed a 'what' comment in the same file.
 363
 364 2012-07-13  Emil A Eklund  <eae@chromium.org>
 365
 366         Use LayoutBoxExtent for image outsets
 367         https://bugs.webkit.org/show_bug.cgi?id=91166
 368
 369         Reviewed by Tony Chang.
 370
 371         Change RenderStyle and calling code to use LayoutBoxExtent for image
 372         outsets and remove text direction and writing mode versions of the
 373         outline getters from RenderStyle as LayoutBoxExtent provides the same
 374         functionality.
 375
 376         No new tests, no change in functionality.
 377
 378         * platform/graphics/FractionalLayoutBoxExtent.h:
 379         * platform/graphics/FractionalLayoutBoxExtent.cpp:
 380         (WebCore::FractionalLayoutBoxExtent::logicalTop):
 381         (WebCore::FractionalLayoutBoxExtent::logicalBottom):
 382         Add logicalTop and logicalBottom methods to go with the existing
 383         logicalLeft and logicalRight ones.
 384
 385         * platform/graphics/FractionalLayoutRect.h:
 386         (WebCore::FractionalLayoutRect::expand):
 387         Add FractionalLayoutBoxExtent version of expand method.
 388
 389         * rendering/InlineFlowBox.cpp:
 390         (WebCore::InlineFlowBox::addBorderOutsetVisualOverflow):
 391         Change implementation to use the new FractionalLayoutBoxExtent version of
 392         borderImageOutsets and the logicalTop/Bottom/Left/Right methods.
 393
 394         (WebCore::clipRectForNinePieceImageStrip):
 395         Change implementation to use the new FractionalLayoutBoxExtent version of
 396         borderImageOutsets.
 397
 398         * rendering/RenderBox.cpp:
 399         (WebCore::RenderBox::maskClipRect):
 400         Change implementation to use the new FractionalLayoutBoxExtent version of
 401         borderImageOutsets and the new FractionalLayoutRect::expand method.
 402
 403         (WebCore::RenderBox::addVisualEffectOverflow):
 404         Change implementation to use the new FractionalLayoutBoxExtent version of
 405         borderImageOutsets.
 406
 407         * rendering/RenderBoxModelObject.cpp:
 408         (WebCore::RenderBoxModelObject::paintNinePieceImage):
 409         Change implementation to use the new FractionalLayoutBoxExtent version of
 410         borderImageOutsets and the new FractionalLayoutRect::expand method.
 411
 412         * rendering/style/RenderStyle.h:
 413         * rendering/style/RenderStyle.cpp:
 414         (WebCore::RenderStyle::imageOutsets):
 415         Change getImageOutsets to return a FractionalLayoutBoxExtent object and
 416         rename to imageOutsets to match the webkit naming convention for getters.
 417
 418         Remove getBorderImageHorizontalOutsets, getBorderImageVerticalOutsets,
 419         getBorderImageInlineDirectionOutsets, getImageHorizontalOutsets,
 420         getImageVerticalOutsets and getBorderImageBlockDirectionOutsets methods
 421         as the same functionality is provided by FractionalLayoutBoxExtent.
 422
 423 2012-07-13  David Hyatt  <hyatt@apple.com>
 424
 425         https://bugs.webkit.org/show_bug.cgi?id=91278
 426         Improve block margin estimation function to account for not having a layout and for quirks mode
 427
 428         Reviewed by Simon Fraser.
 429
 430         * rendering/RenderBlock.cpp:
 431         (WebCore::RenderBlock::marginBeforeEstimateForChild):
 432         Revise marginBeforeEstimateForChild so that it computes block margins for the grandchild before
 433         recurring. This includes the quirks margin information as well. This ensures that the margins are
 434         up-to-date when checked, even before the object has had its first layout.
 435
 436         * rendering/RenderBlock.h:
 437         (WebCore::RenderBlock::setMarginStartForChild):
 438         (WebCore::RenderBlock::setMarginEndForChild):
 439         (WebCore::RenderBlock::setMarginBeforeForChild):
 440         (WebCore::RenderBlock::setMarginAfterForChild):
 441         * rendering/RenderBox.cpp:
 442         (WebCore::RenderBox::computeBlockDirectionMargins):
 443         * rendering/RenderBox.h:
 444         (RenderBox):
 445         Add consts in order to compile.
 446
 447 2012-07-13  Ryosuke Niwa  <rniwa@webkit.org>
 448
 449         NodeLists should not invalidate on irreleavnt attribute changes
 450         https://bugs.webkit.org/show_bug.cgi?id=91277
 451
 452         Reviewed by Ojan Vafai.
 453
 454         Explicitely check the invalidation type and the changed attribute in NodeListNodeData::invalidateCaches
 455         and ElementRareData::clearHTMLCollectionCaches to only invalidate node lists affected by the change.
 456
 457         Also merged invalidateNodeListsCacheAfterAttributeChanged and invalidateNodeListsCacheAfterChildrenChanged
 458         as invalidateNodeListCachesInAncestors since they're almost identical after r122498.
 459
 460         In addition, moved shouldInvalidateNodeListForType from Document.cpp to DynamicNodeList.h and renamed it to
 461         shouldInvalidateTypeOnAttributeChange since it needs to called in Node.cpp and ElementRareData.h.
 462
 463         * dom/Attr.cpp:
 464         (WebCore::Attr::setValue):
 465         (WebCore::Attr::childrenChanged):
 466         * dom/ContainerNode.cpp:
 467         (WebCore::ContainerNode::childrenChanged):
 468         * dom/Document.cpp:
 469         (WebCore::Document::registerNodeListCache): Calls isRootedAtDocument() instead of directly comparing
 470         the value of NodeListRootType in order to prepare for the bug 80269.
 471         (WebCore::Document::unregisterNodeListCache): Ditto.
 472         (WebCore): shouldInvalidateNodeListForType is moved to DynamicNodeList.h
 473         (WebCore::Document::shouldInvalidateNodeListCaches):
 474         * dom/DynamicNodeList.h:
 475         (DynamicNodeListCacheBase):
 476         (WebCore::DynamicNodeListCacheBase::shouldInvalidateTypeOnAttributeChange): Moved from Document.cpp.
 477         * dom/Element.cpp:
 478         (WebCore::Element::attributeChanged):
 479         * dom/ElementRareData.h:
 480         (WebCore::ElementRareData::clearHTMLCollectionCaches): Takes const QualifiedName* to compare against
 481         the invalidation type of HTML collections via shouldInvalidateTypeOnAttributeChange.
 482         * dom/Node.cpp:
 483         (WebCore::Node::invalidateNodeListCachesInAncestors): Merged invalidateNodeListCachesInAncestors and
 484         invalidateNodeListsCacheAfterChildrenChanged. Also pass attrName to clearHTMLCollectionCaches.
 485         (WebCore::NodeListsNodeData::invalidateCaches): Compares attrName against the invalidation type of
 486         node lists via shouldInvalidateTypeOnAttributeChange.
 487         (WebCore):
 488         * dom/Node.h:
 489         (Node):
 490         * dom/NodeRareData.h:
 491         (WebCore::NodeRareData::ensureNodeLists): Merged NodeRareData::createNodeLists.
 492         (WebCore::NodeRareData::clearChildNodeListCache): Moved from Node.cpp.
 493         (NodeRareData):
 494         * html/HTMLCollection.h:
 495         (HTMLCollectionCacheBase):
 496
 497 2012-07-13  Arpita Bahuguna  <arpitabahuguna@gmail.com>
 498
 499         Refactor RenderTable to use the section's iteration functions.
 500         https://bugs.webkit.org/show_bug.cgi?id=89751
 501
 502         Reviewed by Julien Chaffraix.
 503
 504         Removing anti-pattern wherever possible from RenderTable code. Also, modifying
 505         RenderTable sections' iterations to use helper functions.
 506
 507         No new tests required for this change since no change in behavior is expected.
 508
 509         * rendering/RenderTable.cpp:
 510         (WebCore::RenderTable::addOverflowFromChildren):
 511         (WebCore::RenderTable::setCellLogicalWidths):
 512         (WebCore::RenderTable::outerBorderStart):
 513         (WebCore::RenderTable::outerBorderEnd):
 514         Removed anti-patterns involving iterations over RenderObjects.
 515
 516         (WebCore::RenderTable::outerBorderAfter):
 517         Modified RenderTable sections' iteration to use helper functions.
 518
 519 2012-07-13  Enrica Casucci  <enrica@apple.com>
 520
 521         Threadsafety issues in WebScriptObject
 522         https://bugs.webkit.org/show_bug.cgi?id=90849
 523
 524         Reviewed by Geoff Garen.
 525
 526         Updated fix for this bug. The JSC API lock needs to be acquired also in JSObject.
 527
 528         * bindings/objc/WebScriptObject.mm:
 529         (-[WebScriptObject JSObject]):
 530
 531 2012-07-13  Raymond Toy  <rtoy@google.com>
 532
 533         DelayNode doesn't work if delayTime.value == delayTime.maxValue
 534         https://bugs.webkit.org/show_bug.cgi?id=90357
 535
 536         Reviewed by Kenneth Russell.
 537
 538         Increase delay buffer size slightly so that the read and write
 539         pointers don't become equal when the delay and the max delay are
 540         the same.
 541
 542         Tests: webaudio/delaynode-max-default-delay.html
 543                webaudio/delaynode-max-nondefault-delay.html
 544
 545         * Modules/webaudio/DelayDSPKernel.cpp:
 546         (WebCore): Moved SmoothingTimeConstant to WebCore namespace.
 547         (WebCore::DelayDSPKernel::DelayDSPKernel): Add some additional checks to prevent crashes; use bufferLengthForDelay to compute buffer length.
 548         (WebCore::DelayDSPKernel::bufferLengthForDelay): New function to compute buffer length.
 549         * Modules/webaudio/DelayDSPKernel.h:
 550         (DelayDSPKernel): Declare bufferLengthForDelay.
 551
 552 2012-07-13  Benjamin Poulain  <bpoulain@apple.com>
 553
 554         Always aggressively preload on iOS
 555         https://bugs.webkit.org/show_bug.cgi?id=91276
 556
 557         Reviewed by Simon Fraser.
 558
 559         * loader/cache/CachedResourceLoader.cpp:
 560         (WebCore::CachedResourceLoader::preload):
 561
 562 2012-07-13  Vineet Chaudhary  <rgf748@motorola.com>
 563
 564         Restructure V8Utilities::extractTransferables() with help of toV8Sequence()
 565         https://bugs.webkit.org/show_bug.cgi?id=91208
 566
 567         Reviewed by Kentaro Hara.
 568
 569         We can remove the specialised check for MessagePort from V8Utilities::extractTransferables()
 570         using toV8Sequence() as it validates the passed object for sequence type per WebIDL spec.
 571
 572         No new test as just refactoring.
 573         Existing tests under fast/dom/Window/window-* fast/dom/events/*
 574         covers tests.
 575
 576         * bindings/v8/V8Utilities.cpp:
 577         (WebCore::extractTransferables):
 578
 579 2012-07-13  Vincent Scheib  <scheib@chromium.org>
 580
 581         Pointer Lock handles disconnected DOM elements
 582         https://bugs.webkit.org/show_bug.cgi?id=77029
 583
 584         Reviewed by Adrienne Walker.
 585
 586         Pointer Lock Controller now checks when elements or documents are
 587         removed, and unlocks if the target element is being removed.
 588
 589         Tests: pointer-lock/locked-element-iframe-removed-from-dom.html
 590                pointer-lock/locked-element-removed-from-dom.html
 591
 592         * dom/Document.cpp:
 593         (WebCore::Document::detach):
 594         * dom/Element.cpp:
 595         (WebCore::Element::removedFrom):
 596         (WebCore::Element::webkitRequestPointerLock):
 597         * page/PointerLockController.cpp:
 598         (WebCore::PointerLockController::requestPointerLock):
 599         (WebCore::PointerLockController::elementRemoved):
 600         (WebCore):
 601         (WebCore::PointerLockController::documentDetached):
 602         (WebCore::PointerLockController::didLosePointerLock):
 603         (WebCore::PointerLockController::enqueueEvent):
 604         * page/PointerLockController.h:
 605         (WebCore):
 606         (PointerLockController):
 607
 608 2012-07-13  Ryosuke Niwa  <rniwa@webkit.org>
 609
 610         HTMLCollection should use DynamicNodeList's invalidation model
 611         https://bugs.webkit.org/show_bug.cgi?id=90326
 612
 613         Reviewed by Anders Carlsson.
 614
 615         Make HTMLCollection invalidated upon attribute and children changes instead of invalidating it on demand
 616         by comparing DOM tree versions. Node that HTMLCollections owned by Document are invalidated with other
 617         document-rooted node lists in m_listsInvalidatedAtDocument for simplicity although this mechanism is
 618         normally used for node lists owned by a non-Document node that contains nodes outside of its subtree.
 619         ItemProperties and FormControls are more "traditional" users of the mechanism.
 620
 621         Also, merged DynamicNodeList::invalidateCache and HTMLCollection::invalidateCache.
 622
 623         * dom/Document.cpp:
 624         (WebCore::Document::registerNodeListCache): Renamed. No longer takes NodeListInvalidationType or
 625         NodeListRootType since they can be obtained from the cache base. Increment the node list counter for
 626         InvalidateOnIdNameAttrChange when a HTMLCollection is passed in since all HTMLCollections need to be
 627         invalidated on id or name content attribute changes due to named getters.
 628         (WebCore::Document::unregisterNodeListCache): Ditto.
 629         (WebCore::shouldInvalidateNodeListForType):
 630         (WebCore::Document::shouldInvalidateNodeListCaches):
 631         (WebCore::Document::clearNodeListCaches):
 632         * dom/Document.h:
 633         (WebCore): Added InvalidateOnIdNameAttrChange, InvalidateOnHRefAttrChange, and InvalidateOnAnyAttrChange.
 634         (Document):
 635         * dom/DynamicNodeList.cpp:
 636         (WebCore::DynamicNodeListCacheBase::invalidateCache): Added. Invalidates caches of both DynamicNodeList
 637         and HTMLCollection. We can't afford to use virtual function calls here because this function is called on
 638         all node lists and HTML collections owned by ancestors of an element under which a node is added, removed,
 639         or its attributes are changed.
 640         (WebCore):
 641         * dom/DynamicNodeList.h:
 642         (WebCore::DynamicNodeListCacheBase::DynamicNodeListCacheBase): Initializes member variables directly
 643         instead of calling clearCache now that DynamicNodeListCacheBase::invalidateCache has become polymorphic.
 644         (DynamicNodeListCacheBase): Increased the number of bits for m_invalidationType since we now have 9
 645         invalidation types.
 646         (WebCore::DynamicSubtreeNodeList::~DynamicSubtreeNodeList):
 647         (WebCore::DynamicSubtreeNodeList::DynamicSubtreeNodeList):
 648         * dom/ElementRareData.h:
 649         (ElementRareData):
 650         (WebCore::ElementRareData::clearHTMLCollectionCaches): Added.
 651         (WebCore::ElementRareData::adoptTreeScope): Added; similar to NodeRareData::adoptTreeScope.
 652         * dom/Node.cpp:
 653         (WebCore::Node::invalidateNodeListsCacheAfterAttributeChanged): Clears HTML collection caches as well as
 654         node list caches.
 655         (WebCore::Node::invalidateNodeListsCacheAfterChildrenChanged): Ditto.
 656         * dom/NodeRareData.h:
 657         (WebCore::NodeListsNodeData::adoptTreeScope):
 658         * dom/TreeScopeAdopter.cpp:
 659         (WebCore::TreeScopeAdopter::moveTreeToNewScope): Calls ElementRareData's adoptTreeScope as well as
 660         NodeRareData's.
 661         * html/HTMLAllCollection.cpp:
 662         (WebCore::HTMLAllCollection::namedItemWithIndex):
 663         * html/HTMLCollection.cpp:
 664         (WebCore::rootTypeFromCollectionType): Added. As mentioned above, treat all Document-owned HTML collection
 665         as if rooted at document for convenience.
 666         (WebCore::invalidationTypeExcludingIdAndNameAttributes): Added. Since all HTML collection requires
 667         invalidation on id and name content attribute changes, which is taken care by the special logic in
 668         Document::registerNodeListCache, exclude those two attributes from consideration.
 669         (WebCore::HTMLCollection::HTMLCollection): Calls Document::registerNodeListCache.
 670         (WebCore::HTMLCollection::~HTMLCollection): Calls Document::unregisterNodeListCache.
 671         (WebCore::HTMLCollection::length):
 672         (WebCore::HTMLCollection::item):
 673         (WebCore::HTMLCollection::namedItem):
 674         (WebCore::HTMLCollection::updateNameCache):
 675         * html/HTMLCollection.h:
 676         (WebCore::HTMLCollectionCacheBase::HTMLCollectionCacheBase):
 677         (HTMLCollectionCacheBase): Removed m_cacheTreeVersion and clearCache since they're no longer used.
 678         (HTMLCollection):
 679         * html/HTMLFormCollection.cpp:
 680         (WebCore::HTMLFormCollection::namedItem):
 681         (WebCore::HTMLFormCollection::updateNameCache):
 682         * html/HTMLOptionsCollection.h:
 683         (HTMLOptionsCollection):
 684         * html/HTMLPropertiesCollection.cpp:
 685         (WebCore::HTMLPropertiesCollection::updateNameCache):
 686         * html/HTMLPropertiesCollection.h:
 687         (WebCore::HTMLPropertiesCollection::invalidateCache):
 688
 689 2012-07-13  Shawn Singh  <shawnsingh@chromium.org>
 690
 691         [chromium] Remove incorrect debug assertion in LayerRendererChromium.cpp
 692         https://bugs.webkit.org/show_bug.cgi?id=91260
 693
 694         Reviewed by Adrienne Walker.
 695
 696         ASSERT(!clipped) was being triggered after skinny almost-degenerate
 697         quads went through anti-aliasing inflation, and then were being
 698         transformed back from device space to local space. It turns out
 699         this assertion is too aggressive, and we don't yet have an obvious
 700         need to change the behavior on the clipped==true case.
 701
 702         No new tests needed, this patch fixes only comments and debug code.
 703
 704         * platform/graphics/chromium/LayerRendererChromium.cpp:
 705         (WebCore::LayerRendererChromium::drawRenderPassQuad):
 706            fixed a comment.
 707
 708         (WebCore::LayerRendererChromium::drawTileQuad):
 709            fixed a similar comment, removed unnecessary assertion.
 710
 711 2012-07-13  Philip Rogers  <pdr@google.com>
 712
 713         Remove assert in localCoordinateSpaceTransform()
 714         https://bugs.webkit.org/show_bug.cgi?id=91189
 715
 716         Reviewed by Nikolas Zimmermann.
 717
 718         The assert in localCoordinateSpaceTransform was added to catch subclasses forgetting
 719         to override the method but it is better to simply return the identity matrix.
 720
 721         This scenario can occur when we break the SVG content model, such as asking for
 722         the CTM of a <g> element inside a <tspan>. This is undefined in the spec because
 723         tspan is not a subclass of SVGLocatable but both Firefox and Opera
 724         implement this by returning the identity matrix.
 725
 726         Test: svg/custom/invalid-ctm.svg
 727
 728         * svg/SVGStyledElement.cpp:
 729         (WebCore::SVGStyledElement::localCoordinateSpaceTransform):
 730
 731 2012-07-13  Kentaro Hara  <haraken@chromium.org>
 732
 733         [V8] String wrappers should be marked Independent
 734         https://bugs.webkit.org/show_bug.cgi?id=91251
 735
 736         Reviewed by Adam Barth.
 737
 738         Currently V8 String wrappers are not marked Independent.
 739         By marking them Independent, they can be reclaimed by the scavenger GC.
 740
 741         I tried to find some cases where this change reduces memory usage,
 742         but couldn't due to sensitive behavior of GC.
 743
 744         No tests. No change in behavior.
 745
 746         * bindings/v8/V8Binding.cpp:
 747         (WebCore::StringCache::v8ExternalStringSlow):
 748
 749 2012-07-13  Peter Beverloo  <peter@chromium.org>
 750
 751         [Chromium] Make the v8 i18n API dependency conditional for Android, disable strict aliasing
 752         https://bugs.webkit.org/show_bug.cgi?id=91240
 753
 754         Reviewed by Adam Barth.
 755
 756         Disable the v8 internationalization API for Chromium Android, as it's
 757         disabled and not always available in checkouts. Furthermore, disable
 758         strict aliasing for the webkit_remaining target, similar to what
 759         x11-based builds are doing (see the webcore_prerequisites target).
 760
 761         * WebCore.gyp/WebCore.gyp:
 762
 763 2012-07-13  Kentaro Hara  <haraken@chromium.org>
 764
 765         [CallWith=XXX] arguments should be placed at the head of method arguments
 766         https://bugs.webkit.org/show_bug.cgi?id=91217
 767
 768         Reviewed by Adam Barth.
 769
 770         The EFL build with the ENABLE_FILE_SYSTEM flag caused a build error,
 771         because CodeGeneratorJS.pm assumes webkitEntries(ScriptExecutionContext*, HTMLInputElement*)
 772         but the actual signature is webkitEntries(HTMLInputElement*, ScriptExecutionContext*) (bug 91185).
 773
 774         Per https://trac.webkit.org/wiki/WebKitIDL#CallWith, [CallWith=XXX] arguments should be placed
 775         at the head of the arguments. (i.e. the behavior of CodeGeneratorJS.pm is correct.)
 776
 777         Thus the correct fix is (1) to change the signature of webkitEntries() and webkitGetAsEntry()
 778         so that ScriptExecutionContext* comes first and (2) to modify CodeGeneratorV8.pm to support the order.
 779
 780         Test: bindings/scripts/test/TestObj.idl
 781
 782         * Modules/filesystem/DataTransferItemFileSystem.h: Placed ScriptExecutionContext* at the head of arguments.
 783         (DataTransferItemFileSystem):
 784         * Modules/filesystem/HTMLInputElementFileSystem.cpp: Ditto.
 785         (WebCore::HTMLInputElementFileSystem::webkitEntries):
 786         * Modules/filesystem/HTMLInputElementFileSystem.h: Ditto.
 787         (HTMLInputElementFileSystem):
 788         * Modules/filesystem/chromium/DataTransferItemFileSystemChromium.cpp: Ditto.
 789         (WebCore::DataTransferItemFileSystem::webkitGetAsEntry):
 790
 791         * bindings/scripts/CodeGeneratorV8.pm: Modified to support the correct order.
 792         (GenerateNormalAttrGetter):
 793         (GenerateNormalAttrSetter):
 794         (GenerateFunctionCallString):
 795
 796         * bindings/scripts/test/V8/V8TestInterface.cpp: Updated run-bindings-tests results.
 797         (WebCore::TestInterfaceV8Internal::supplementalMethod2Callback):
 798
 799 2012-07-13  Mary Wu  <mary.wu@torchmobile.com.cn>
 800
 801         [BlackBerry] Some small changes in network code
 802         https://bugs.webkit.org/show_bug.cgi?id=90974
 803
 804         Reviewed by Rob Buis.
 805
 806         1. Set status in NetworkJob/SocketStreamHandleBlackBerry so that
 807         its wrapped stream can also query the stream result.
 808         2. pass download attribute to NetworkRequest.
 809
 810         RIM PR# 171555
 811         Reviewed internally by Lyon Chen and Joe Mason.
 812
 813         * platform/network/blackberry/NetworkJob.cpp:
 814         (WebCore::NetworkJob::handleNotifyClose):
 815         * platform/network/blackberry/NetworkJob.h:
 816         (WebCore::NetworkJob::status):
 817         * platform/network/blackberry/ResourceRequestBlackBerry.cpp:
 818         (WebCore::platformTargetTypeForRequest):
 819         * platform/network/blackberry/SocketStreamHandle.h:
 820         (WebCore::SocketStreamHandle::status):
 821         (SocketStreamHandle):
 822         * platform/network/blackberry/SocketStreamHandleBlackBerry.cpp:
 823         (WebCore::SocketStreamHandle::notifyStatusReceived):
 824         (WebCore::SocketStreamHandle::notifyClose):
 825
 826 2012-07-13  Vsevolod Vlasov  <vsevik@chromium.org>
 827
 828         Web Inspector: Remove uiSourceCode from Resource.
 829         https://bugs.webkit.org/show_bug.cgi?id=91201
 830
 831         Reviewed by Pavel Feldman.
 832
 833         Removed Resource._uiSourceCode field as it is not used anymore.
 834
 835         * inspector/front-end/Resource.js:
 836         (WebInspector.Resource.prototype.isHidden):
 837         * inspector/front-end/UISourceCode.js:
 838         (WebInspector.UISourceCode):
 839
 840 2012-07-13  Sheriff Bot  <webkit.review.bot@gmail.com>
 841
 842         Unreviewed, rolling out r122450 and r122580.
 843         http://trac.webkit.org/changeset/122450
 844         http://trac.webkit.org/changeset/122580
 845         https://bugs.webkit.org/show_bug.cgi?id=91263
 846
 847         Caused multiple regressions on ClusterFuzz (Requested by
 848         inferno-sec on #webkit).
 849
 850         * bindings/js/ScriptWrappable.h:
 851         (WebCore::ScriptWrappable::reportMemoryUsage):
 852         * bindings/v8/DOMDataStore.cpp:
 853         (WebCore::DOMDataStore::reportMemoryUsage):
 854         * bindings/v8/IntrusiveDOMWrapperMap.h:
 855         (WebCore::ChunkedTable::reportMemoryUsage):
 856         * bindings/v8/ScriptWrappable.h:
 857         (WebCore::ScriptWrappable::reportMemoryUsage):
 858         * bindings/v8/V8Binding.cpp:
 859         (WebCore::V8BindingPerIsolateData::reportMemoryUsage):
 860         (WebCore::StringCache::reportMemoryUsage):
 861         * bindings/v8/V8DOMMap.h:
 862         * css/PropertySetCSSStyleDeclaration.cpp:
 863         (WebCore::InlineCSSStyleDeclaration::ensureMutablePropertySet):
 864         * css/StylePropertySet.cpp:
 865         * css/StylePropertySet.h:
 866         (WebCore::StylePropertySet::reportMemoryUsage):
 867         * dom/CharacterData.cpp:
 868         (WebCore::CharacterData::reportMemoryUsage):
 869         * dom/ContainerNode.h:
 870         (WebCore::ContainerNode::reportMemoryUsage):
 871         * dom/Document.cpp:
 872         (WebCore::Document::reportMemoryUsage):
 873         * dom/Element.cpp:
 874         (WebCore::Element::detachAttribute):
 875         (WebCore::Element::removeAttribute):
 876         (WebCore::Element::attributes):
 877         (WebCore::Element::setAttributeInternal):
 878         (WebCore::Element::parserSetAttributes):
 879         (WebCore::Element::hasEquivalentAttributes):
 880         (WebCore::Element::createAttributeData):
 881         (WebCore):
 882         (WebCore::Element::setAttributeNode):
 883         (WebCore::Element::removeAttributeNode):
 884         (WebCore::Element::getAttributeNode):
 885         (WebCore::Element::getAttributeNodeNS):
 886         (WebCore::Element::hasAttribute):
 887         (WebCore::Element::hasAttributeNS):
 888         (WebCore::Element::normalizeAttributes):
 889         (WebCore::Element::cloneAttributesFromElement):
 890         * dom/Element.h:
 891         (WebCore::Element::attributeData):
 892         (Element):
 893         (WebCore::Element::reportMemoryUsage):
 894         (WebCore::Element::ensureAttributeData):
 895         (WebCore::Element::updatedAttributeData):
 896         (WebCore::Element::ensureUpdatedAttributeData):
 897         * dom/ElementAttributeData.cpp:
 898         (WebCore::ElementAttributeData::attrIfExists):
 899         (WebCore::ElementAttributeData::ensureAttr):
 900         (WebCore::ElementAttributeData::setAttr):
 901         (WebCore::ElementAttributeData::removeAttr):
 902         (WebCore::ElementAttributeData::setClass):
 903         (WebCore):
 904         (WebCore::ElementAttributeData::ensureInlineStyle):
 905         (WebCore::ElementAttributeData::ensureMutableInlineStyle):
 906         (WebCore::ElementAttributeData::destroyInlineStyle):
 907         (WebCore::ElementAttributeData::addAttribute):
 908         (WebCore::ElementAttributeData::removeAttribute):
 909         (WebCore::ElementAttributeData::isEquivalent):
 910         (WebCore::ElementAttributeData::detachAttrObjectsFromElement):
 911         (WebCore::ElementAttributeData::getAttributeItemIndexSlowCase):
 912         (WebCore::ElementAttributeData::cloneDataFrom):
 913         (WebCore::ElementAttributeData::clearAttributes):
 914         (WebCore::ElementAttributeData::replaceAttribute):
 915         (WebCore::ElementAttributeData::getAttributeNode):
 916         * dom/ElementAttributeData.h:
 917         (WebCore::ElementAttributeData::create):
 918         (ElementAttributeData):
 919         (WebCore::ElementAttributeData::setIdForStyleResolution):
 920         (WebCore::ElementAttributeData::setAttributeStyle):
 921         (WebCore::ElementAttributeData::length):
 922         (WebCore::ElementAttributeData::isEmpty):
 923         (WebCore::ElementAttributeData::attributeItem):
 924         (WebCore::ElementAttributeData::getAttributeItem):
 925         (WebCore::ElementAttributeData::reportMemoryUsage):
 926         (WebCore::ElementAttributeData::ElementAttributeData):
 927         (WebCore::ElementAttributeData::attributeVector):
 928         (WebCore::ElementAttributeData::clonedAttributeVector):
 929         (WebCore::ElementAttributeData::removeAttribute):
 930         (WebCore::ElementAttributeData::getAttributeItemIndex):
 931         * dom/MemoryInstrumentation.h:
 932         (MemoryInstrumentation):
 933         (MemoryObjectInfo):
 934         (WebCore::MemoryObjectInfo::reportInstrumentedPointer):
 935         (WebCore::MemoryObjectInfo::reportPointer):
 936         (WebCore::MemoryObjectInfo::reportInstrumentedObject):
 937         (WebCore::MemoryObjectInfo::reportObject):
 938         (WebCore::MemoryObjectInfo::reportObjectInfo):
 939         (WebCore::MemoryObjectInfo::reportHashMap):
 940         (WebCore::MemoryObjectInfo::reportHashSet):
 941         (WebCore::MemoryObjectInfo::reportListHashSet):
 942         (WebCore::MemoryObjectInfo::reportVector):
 943         (WebCore::MemoryObjectInfo::reportString):
 944         (WebCore::MemoryObjectInfo::objectType):
 945         (WebCore::MemoryObjectInfo::objectSize):
 946         (WebCore::MemoryObjectInfo::memoryInstrumentation):
 947         * dom/Node.cpp:
 948         (WebCore::Node::reportMemoryUsage):
 949         * dom/QualifiedName.h:
 950         (WebCore::QualifiedName::QualifiedNameImpl::reportMemoryUsage):
 951         (WebCore::QualifiedName::reportMemoryUsage):
 952         * dom/StyledElement.cpp:
 953         (WebCore::StyledElement::style):
 954         (WebCore::StyledElement::classAttributeChanged):
 955         (WebCore::StyledElement::setInlineStyleProperty):
 956         (WebCore::StyledElement::removeInlineStyleProperty):
 957         (WebCore::StyledElement::addSubresourceAttributeURLs):
 958         * dom/StyledElement.h:
 959         (WebCore::StyledElement::ensureInlineStyle):
 960         * html/parser/HTMLConstructionSite.cpp:
 961         (WebCore::HTMLConstructionSite::createHTMLElementFromSavedElement):
 962         * platform/TreeShared.h:
 963         (WebCore::TreeShared::reportMemoryUsage):
 964         * xml/parser/XMLDocumentParserQt.cpp:
 965         (WebCore::XMLDocumentParser::XMLDocumentParser):
 966
 967 2012-07-13  Huang Dongsung  <luxtella@company100.net>
 968
 969         Remove down-casting to BitmapImage in GraphicsContext::drawImage.
 970         https://bugs.webkit.org/show_bug.cgi?id=90755
 971
 972         Reviewed by Simon Fraser.
 973
 974         Add a BitmapImage draw method which takes RespectImageOrientationEnum enum as
 975         the last argument for CG. Then we can remove the conditional down-casting in
 976         GraphicsContext::drawImage.
 977
 978         This change is needed for parallel image decoders. Because parallel image
 979         decoders use a Bitmap image wrapper class which extends Image (not Bitmap), the
 980         down-casting above causes the loss of RespectImageOrientationEnum which must be
 981         passed to BitmapImage.
 982
 983         No new tests, no behavior change.
 984
 985         * platform/graphics/BitmapImage.cpp:
 986         * platform/graphics/BitmapImage.h:
 987         * platform/graphics/GraphicsContext.cpp:
 988         (WebCore::GraphicsContext::drawImage):
 989         * platform/graphics/Image.cpp:
 990         (WebCore::Image::draw):
 991         (WebCore):
 992         * platform/graphics/Image.h:
 993         (Image):
 994
 995 2012-07-13  Lauro Neto  <lauro.neto@openbossa.org>
 996
 997         Fix QtWebKit build with OpenGLES after GC3D/E3D refactor
 998         https://bugs.webkit.org/show_bug.cgi?id=91156
 999
1000         Reviewed by Noam Rosenthal.
1001
1002         Adds several build fixes.
1003
1004         * platform/graphics/GraphicsContext3D.h:
1005             Use E3DOpenGLES instead of previously removed E3DQt.
1006
1007         * platform/graphics/OpenGLESShims.h:
1008             Enable defines for Qt.
1009
1010         * platform/graphics/opengl/Extensions3DOpenGLES.cpp:
1011         (WebCore::Extensions3DOpenGLES::blitFramebuffer):
1012         (WebCore):
1013         (WebCore::Extensions3DOpenGLES::renderbufferStorageMultisample):
1014         (WebCore::Extensions3DOpenGLES::copyTextureCHROMIUM):
1015             Added pure virtual stubs.
1016
1017         (WebCore::Extensions3DOpenGLES::supportsExtension):
1018             Remove PROC suffix. See bug #91130.
1019
1020         * platform/graphics/opengl/Extensions3DOpenGLES.h:
1021         (Extensions3DOpenGLES):
1022             Added pure virtual stubs.
1023
1024         * platform/graphics/opengl/GraphicsContext3DOpenGLES.cpp:
1025         (WebCore::GraphicsContext3D::reshapeFBOs):
1026             Readded missing function after removed in r122250.
1027
1028         (WebCore):
1029             Use PLATFORM(BLACKBERRY) guard around port-specific include.
1030
1031         * platform/graphics/qt/GraphicsContext3DQt.cpp:
1032             Added USE(OPENGL_ES_2) guard instead of always loading the OpenGL extensions.
1033
1034
1035 2012-07-13  Keishi Hattori  <keishi@webkit.org>
1036
1037         Form of FormAssociatedElement is not updated when id target changes.
1038         https://bugs.webkit.org/show_bug.cgi?id=91042
1039
1040         Reviewed by Kent Tamura.
1041
1042         Test: fast/forms/update-form-attribute-element.html
1043
1044         This patch introduces the IdTargetObserver and IdTargetObserverRegistry class.
1045         They can be used to be notified when the element that an id is pointing to (the id target)
1046         changes.
1047
1048         * CMakeLists.txt: Added IdTargetObserverRegistry.{h,cpp} and IdTargetObserver.{h,cpp}
1049         * GNUmakefile.list.am: Ditto.
1050         * Target.pri: Ditto.
1051         * WebCore.gypi: Ditto.
1052         * WebCore.vcproj/WebCore.vcproj: Ditto.
1053         * WebCore.xcodeproj/project.pbxproj: Ditto.
1054         * dom/DOMAllInOne.cpp:
1055         * dom/IdTargetObserver.cpp: Added. When you want notified of changes to an id target, you should create a new class that inherits this.
1056         (WebCore):
1057         (WebCore::IdTargetObserver::IdTargetObserver):
1058         (WebCore::IdTargetObserver::~IdTargetObserver):
1059         * dom/IdTargetObserver.h: Added.
1060         (WebCore):
1061         (IdTargetObserver):
1062         * dom/IdTargetObserverRegistry.cpp: Added.
1063         (WebCore):
1064         (WebCore::IdTargetObserverRegistry::create):
1065         (WebCore::IdTargetObserverRegistry::addObserver): Register an IdTargetObserver to observe an id target.
1066         (WebCore::IdTargetObserverRegistry::removeObserver): Unregisters an IdTargetObserver from observing.
1067         (WebCore::IdTargetObserverRegistry::notifyObserversInternal):
1068         * dom/IdTargetObserverRegistry.h: Added.
1069         (WebCore):
1070         (IdTargetObserverRegistry):
1071         (WebCore::IdTargetObserverRegistry::IdTargetObserverRegistry):
1072         (WebCore::IdTargetObserverRegistry::notifyObservers): Calls idTargetChanged on all observers for an id. Inlining first part of function for performance.
1073         * dom/TreeScope.cpp:
1074         (WebCore::TreeScope::TreeScope):
1075         (WebCore::TreeScope::addElementById): Calls IdTargetObserverRegistry::notifyObservers because the id target might have changed.
1076         (WebCore::TreeScope::removeElementById): Ditto.
1077         * dom/TreeScope.h:
1078         (WebCore):
1079         (WebCore::TreeScope::idTargetObserverRegistry):
1080         (TreeScope):
1081         * html/FormAssociatedElement.cpp: Observer for id targets defined by the form attribute.
1082         (WebCore::FormAssociatedElement::didMoveToNewDocument):
1083         (WebCore::FormAssociatedElement::insertedInto):
1084         (WebCore::FormAssociatedElement::removedFrom):
1085         (WebCore::FormAssociatedElement::formAttributeChanged):
1086         (WebCore::FormAssociatedElement::resetFormAttributeTargetObserver): Creates and sets up a new FormAttributeTargetObserver.
1087         (WebCore):
1088         (WebCore::FormAssociatedElement::formAttributeTargetChanged):
1089         (WebCore::FormAttributeTargetObserver::create):
1090         (WebCore::FormAttributeTargetObserver::FormAttributeTargetObserver):
1091         (WebCore::FormAttributeTargetObserver::idTargetChanged):
1092         * html/FormAssociatedElement.h:
1093         (FormAssociatedElement):
1094         * html/FormController.cpp:
1095         * html/FormController.h:
1096         (FormController):
1097         * html/HTMLFormElement.cpp:
1098         (WebCore::HTMLFormElement::removedFrom):
1099         (WebCore::HTMLFormElement::formElementIndexWithFormAttribute): Modified to take a range. It
1100         scans the range and returns the index to insert the element in m_associatedElement.
1101         (WebCore::HTMLFormElement::formElementIndex): Modified to only scan the elements in
1102         m_associatedElement that precede and follow the form element.
1103         * html/HTMLFormElement.h:
1104         (HTMLFormElement):
1105
1106 2012-07-13  Gabor Rapcsanyi  <rgabor@webkit.org>
1107
1108         Optimizing blend filter to ARM-NEON with intrinsics
1109         https://bugs.webkit.org/show_bug.cgi?id=90949
1110
1111         Reviewed by Zoltan Herczeg.
1112
1113         The feBlend SVG filter modes can be greatly fasten up with ARM-NEON since
1114         we are able to calculate with 2 pixels (8 channels) at the same time.
1115         The code is written with NEON intrinsics and it doesn't affect the
1116         general - it has the same behaviour as the original algorithm.
1117         With this NEON optimization the calculation is ~4.5 times faster for each mode.
1118
1119         Existing tests cover this issue.
1120
1121         * CMakeLists.txt:
1122         * GNUmakefile.list.am:
1123         * Target.pri:
1124         * WebCore.gypi:
1125         * WebCore.vcproj/WebCore.vcproj:
1126         * WebCore.xcodeproj/project.pbxproj:
1127         * platform/graphics/filters/FEBlend.cpp:
1128         (WebCore::FEBlend::platformApplyGeneric):
1129         (WebCore):
1130         (WebCore::FEBlend::platformApplySoftware):
1131         * platform/graphics/filters/FEBlend.h:
1132         (FEBlend):
1133         * platform/graphics/filters/arm/FEBlendNEON.h: Added.
1134         (WebCore):
1135         (FEBlendUtilitiesNEON):
1136         (WebCore::FEBlendUtilitiesNEON::div255): integer divison with 255
1137         (WebCore::FEBlendUtilitiesNEON::normal): calculate normal mode blending for two pixels
1138         (WebCore::FEBlendUtilitiesNEON::multiply): calculate multiply mode blending for two pixels
1139         (WebCore::FEBlendUtilitiesNEON::screen): calculate screen mode blending for two pixels
1140         (WebCore::FEBlendUtilitiesNEON::darken): calculate darken mode blending for two pixels
1141         (WebCore::FEBlendUtilitiesNEON::lighten): calculate lighten mode blending for two pixels
1142         (WebCore::FEBlend::platformApplyNEON):
1143
1144 2012-07-13  Ilya Tikhonovsky  <loislo@chromium.org>
1145
1146         Web Inspector: native memory instrumentation: extract instrumentation methods into MemoryClassInfo
1147         https://bugs.webkit.org/show_bug.cgi?id=91227
1148
1149         Reviewed by Pavel Feldman.
1150
1151         void Node::reportMemoryUsage(MemoryObjectInfo* memoryObjectInfo) const
1152         {
1153             MemoryClassInfo<Node> info(memoryObjectInfo, this, MemoryInstrumentation::DOM);
1154             info.visitBaseClass<ScriptWrappable>(this);
1155
1156             info.addMember(m_notInstrumentedPointer); // automatically detects poniter/reference
1157             info.addInstrumentedMember(m_next);
1158             info.addHashSet<MemoryInstrumentation::NonClass>(m_aHash);                // NonClass value_type (report only size of internal template structures)
1159             info.addHashSet<MemoryInstrumentation::NotInstrumentedClass>(m_aHashSet); // not instrumented value_type (use sizeof)
1160             info.addHashSet<MemoryInstrumentation::InstrumentedClass>(m_aHashSet);    // instrumented value_type (call visit)
1161         }
1162
1163         The change is covered by existing tests for native memory snapshot.
1164
1165         * bindings/v8/DOMDataStore.cpp:
1166         (WebCore::DOMDataStore::reportMemoryUsage):
1167         * bindings/v8/IntrusiveDOMWrapperMap.h:
1168         (WebCore::ChunkedTable::reportMemoryUsage):
1169         * bindings/v8/ScriptWrappable.h:
1170         (WebCore::ScriptWrappable::reportMemoryUsage):
1171         * bindings/v8/V8Binding.cpp:
1172         (WebCore::V8BindingPerIsolateData::reportMemoryUsage):
1173         (WebCore::StringCache::reportMemoryUsage):
1174         * bindings/v8/V8DOMMap.h:
1175         * css/StylePropertySet.h:
1176         (WebCore::StylePropertySet::reportMemoryUsage):
1177         * dom/CharacterData.cpp:
1178         (WebCore::CharacterData::reportMemoryUsage):
1179         * dom/ContainerNode.h:
1180         (WebCore::ContainerNode::reportMemoryUsage):
1181         * dom/Document.cpp:
1182         (WebCore::Document::reportMemoryUsage):
1183         * dom/Element.h:
1184         (WebCore::Element::reportMemoryUsage):
1185         * dom/ElementAttributeData.h:
1186         (WebCore::ElementAttributeData::reportMemoryUsage):
1187         * dom/MemoryInstrumentation.h:
1188         (MemoryInstrumentation):
1189         (WebCore::MemoryObjectInfo::objectType):
1190         (WebCore::MemoryObjectInfo::objectSize):
1191         (WebCore::MemoryObjectInfo::memoryInstrumentation):
1192         (MemoryObjectInfo):
1193         (WebCore::MemoryObjectInfo::reportObjectInfo):
1194         (WebCore):
1195         (MemoryClassInfo):
1196         (WebCore::MemoryClassInfo::MemoryClassInfo):
1197         (WebCore::MemoryClassInfo::visitBaseClass):
1198         (WebCore::MemoryClassInfo::reportInstrumentedPointer):
1199         (WebCore::MemoryClassInfo::reportInstrumentedObject):
1200         (WebCore::MemoryClassInfo::reportPointer):
1201         (WebCore::MemoryClassInfo::reportObject):
1202         (WebCore::MemoryClassInfo::reportHashMap):
1203         (WebCore::MemoryClassInfo::reportHashSet):
1204         (WebCore::MemoryClassInfo::reportListHashSet):
1205         (WebCore::MemoryClassInfo::reportVector):
1206         (WebCore::MemoryClassInfo::reportString):
1207         * dom/Node.cpp:
1208         (WebCore::Node::reportMemoryUsage):
1209         * dom/QualifiedName.h:
1210         (WebCore::QualifiedName::QualifiedNameImpl::reportMemoryUsage):
1211         (WebCore::QualifiedName::reportMemoryUsage):
1212         * platform/TreeShared.h:
1213         (WebCore::TreeShared::reportMemoryUsage):
1214
1215 2012-07-13  Pavel Feldman  <pfeldman@chromium.org>
1216
1217         Web Inspector: align scope filters
1218         https://bugs.webkit.org/show_bug.cgi?id=91213
1219
1220         Reviewed by Vsevolod Vlasov.
1221
1222         * inspector/front-end/elementsPanel.css:
1223         (.crumbs):
1224         * inspector/front-end/inspector.css:
1225         (.status-bar > div):
1226         (.scope-bar):
1227         (.scope-bar li):
1228         (.scope-bar li.all):
1229         * inspector/front-end/networkLogView.css:
1230
1231 2012-07-13  Peter Rybin  <peter.rybin@gmail.com>
1232
1233         Web Inspector: too many hardcoded strings in InspectorBackendDispatcher.
1234         https://bugs.webkit.org/show_bug.cgi?id=89198
1235
1236         Reviewed by Yury Semikhatsky.
1237
1238         Instead of generating error message string on every call (mostly for nothing),
1239         error message is generated deeper inside the handler and only command name
1240         is passed every time.
1241
1242         * inspector/CodeGeneratorInspector.py:
1243         (Generator.process_command):
1244
1245 2012-07-13  Joshua Netterfield  <jnetterfield@rim.com>
1246
1247         [BlackBerry] Update about:* pages
1248         https://bugs.webkit.org/show_bug.cgi?id=91121
1249
1250         Reviewed by Yong Li.
1251
1252         Update the about:config pages, and improve the aesthetics of the about:build, about:version, about:credits, about:memory, about:config, and similar pages.
1253
1254         No new tests, because there is no new funtionality.
1255
1256         * platform/network/blackberry/NetworkJob.cpp: Update the aesthetics of about:* pages
1257
1258 2012-07-13  Olivier Blin  <olivier.blin@softathome.com>
1259
1260         Fix checking for optional DeviceOrientationEvent.absolute in JSC bindings
1261         https://bugs.webkit.org/show_bug.cgi?id=91225
1262
1263         Reviewed by Steve Block.
1264
1265         This issue comes from r105036
1266
1267         * bindings/js/JSDeviceOrientationEventCustom.cpp:
1268         (WebCore::JSDeviceOrientationEvent::initDeviceOrientationEvent):
1269
1270 2012-07-13  Andrei Bucur  <abucur@adobe.com>
1271         [CSS Regions] Fix build for bug 89000
1272         https://bugs.webkit.org/show_bug.cgi?id=91215
1273
1274         Reviewed by Kentaro Hara.
1275
1276         Remove the unused variable m_state that was a leftover from a previous version of the patch.
1277
1278         Tests: No new tests, build fix.
1279
1280         * dom/WebKitNamedFlow.cpp:
1281         (WebCore::WebKitNamedFlow::WebKitNamedFlow):
1282         * dom/WebKitNamedFlow.h:
1283         (WebKitNamedFlow):
1284
1285 2012-07-13  Kenichi Ishibashi  <bashi@chromium.org>
1286
1287         [Chromium] Fix bugs in HarfBuzzShaper
1288         https://bugs.webkit.org/show_bug.cgi?id=90951
1289
1290         Reviewed by Tony Chang.
1291
1292         The current implementation has following problems:
1293         - Cannot render RTL text if the TextRun is divided into more than two
1294           HarfBuzzRun.
1295         - Script handling in TextRun partitioning is incorrect.
1296         - Inaccurate calculation of selection rect.
1297         - Wrong rendering position when the first glyph of the TextRun have
1298           non-zero offsets in terms of HarfBuzz.
1299
1300         To fix these problems I rewrote HarfBuzzShaper class. Here is the summary:
1301         - Divide the whole range of TextRun first, then shape them in visual
1302           order.
1303         - Divide TextRun in the same way of old-harfbuzz's
1304           hb_utf16_script_run_next().
1305         - Prefer float than int when calculating selection.
1306         - Adjust the drawing point after shaping.
1307
1308         Added tests covers the fix except for the last problem. The last problem will be covered
1309         by fast/text/international/complex-joining-using-gpos.html after chromium linux port switches
1310         to use HarfBuzzShaper.
1311
1312         Tests: fast/text/shaping/shaping-script-order.html
1313                fast/text/shaping/shaping-selection-rect.html
1314
1315         * platform/graphics/harfbuzz/FontHarfBuzz.cpp:
1316         (WebCore::Font::drawComplexText): Adjusts point after shaping.
1317         * platform/graphics/harfbuzz/ng/HarfBuzzShaper.cpp:
1318         (WebCore::HarfBuzzShaper::HarfBuzzRun::HarfBuzzRun):
1319         (WebCore):
1320         (WebCore::HarfBuzzShaper::HarfBuzzRun::applyShapeResult): Added.
1321         (WebCore::HarfBuzzShaper::HarfBuzzRun::setGlyphAndAdvance): Offsets are no longer needed.
1322         (WebCore::HarfBuzzShaper::HarfBuzzRun::xPositionForOffset): Calculates character offset based on advance.
1323         (WebCore::normalizeCharacters): Added.
1324         (WebCore::HarfBuzzShaper::HarfBuzzShaper):
1325         (WebCore::HarfBuzzShaper::~HarfBuzzShaper):
1326         (WebCore::HarfBuzzShaper::shape): Divides TextRun first, then shapes them.
1327         (WebCore::HarfBuzzShaper::adjustStartPoint): Added.
1328         (WebCore::HarfBuzzShaper::collectHarfBuzzRuns): Added.
1329         (WebCore::HarfBuzzShaper::shapeHarfBuzzRuns): Added.
1330         (WebCore::HarfBuzzShaper::setGlyphPositionsForHarfBuzzRun): Followed other changes.
1331         (WebCore::HarfBuzzShaper::selectionRect): Use float for calculating selection.
1332         * platform/graphics/harfbuzz/ng/HarfBuzzShaper.h:
1333         (HarfBuzzShaper): Holds the start index of character. Removed unnecessary variables.
1334         (WebCore::HarfBuzzShaper::HarfBuzzRun::create): Ditto.
1335         (HarfBuzzRun):
1336         (WebCore::HarfBuzzShaper::HarfBuzzRun::fontData): Added.
1337         (WebCore::HarfBuzzShaper::HarfBuzzRun::startIndex): Ditto.
1338         (WebCore::HarfBuzzShaper::HarfBuzzRun::glyphs): Ditto.
1339         (WebCore::HarfBuzzShaper::HarfBuzzRun::advances): Ditto.
1340
1341 2012-07-13  Kentaro Hara  <haraken@chromium.org>
1342
1343         Unreviewed, rolling out r122545.
1344         http://trac.webkit.org/changeset/122545
1345         https://bugs.webkit.org/show_bug.cgi?id=91185
1346
1347         We found that this was a wrong fix
1348
1349         * bindings/scripts/CodeGeneratorJS.pm:
1350         (GenerateImplementation):
1351
1352 2012-07-13  Kentaro Hara  <haraken@chromium.org>
1353
1354         Unreviewed, rolling out r122553.
1355         http://trac.webkit.org/changeset/122553
1356         https://bugs.webkit.org/show_bug.cgi?id=91198
1357
1358         We found that this was a wrong fix
1359
1360         * bindings/scripts/CodeGeneratorJS.pm:
1361         (GenerateImplementation):
1362
1363 2012-07-13  Kent Tamura  <tkent@chromium.org>
1364
1365         Change the timing of form state restore
1366         https://bugs.webkit.org/show_bug.cgi?id=89962
1367
1368         Reviewed by Hajime Morita.
1369
1370         For a preparation to fix a form identification problem (Bug 91209),
1371         restore controls states when the parsing of their owner forms is
1372         completed. For controls without owners, their states are restored when
1373         their parsing is completed as ever.
1374
1375         No new tests. This doesn't change observable behavior.
1376
1377         * html/FormController.cpp:
1378         (WebCore::ownerFormForState):
1379         Added. This code was used in formKey(), and restoreControlState*() use it.
1380         (WebCore::FormKeyGenerator::formKey): Use ownerFormForState(). No behavior change.
1381         (WebCore::FormController::restoreControlStateFor):
1382         Moved some code from HTMLFormControlElementWithState::finishParsingChildren().
1383         The difference is we don't resotre state if this control is owned by a form.
1384         (WebCore::FormController::restoreControlStateIn):
1385         Restore states of associated controls. This is called from
1386         finishParsingChildren() for <form>.
1387         * html/FormController.h:
1388         (FormController):
1389         - Declare restoreControlStateFor() and restoreControlStateIn().
1390         - Make takeStateForFormElement() private.
1391
1392         * html/FormAssociatedElement.cpp:
1393         (WebCore::FormAssociatedElement::isFormControlElementWithState):
1394         Added. The default implementation returns false.
1395         * html/FormAssociatedElement.h:
1396         (FormAssociatedElement):
1397         Added isFormControlElementWithState() for FormController::restoreControlStateIn().
1398         * html/HTMLFormControlElementWithState.cpp:
1399         (WebCore::HTMLFormControlElementWithState::finishParsingChildren):
1400         Some code was moved to FormController:restoreControlStateFor().
1401         (WebCore::HTMLFormControlElementWithState::isFormControlElementWithState):
1402         Added. Returns true.
1403         * html/HTMLFormControlElementWithState.h:
1404         (HTMLFormControlElementWithState): Declare isFormControlElementWithState().
1405         * html/HTMLFormElement.cpp:
1406         (WebCore::HTMLFormElement::finishParsingChildren):
1407         Added. Calls FormController::restoreControlStateIn().
1408         * html/HTMLFormElement.h:
1409         (HTMLFormElement): Declare finishParsingChildren().
1410
1411 2012-07-13  Kent Tamura  <tkent@chromium.org>
1412
1413         Make calendar pickers testable
1414         https://bugs.webkit.org/show_bug.cgi?id=84827
1415
1416         Reviewed by Hajime Morita.
1417
1418         WebCore:
1419         - Add PagePopupDriver, an interface to open/close a PagePopup.
1420         - Add setPagePopupDriver() to ChromeClient in order to inject a
1421         PagePopupDriver for testing.
1422
1423         Internals:
1424         Add MockPagePopupDriver, which creates a MockPagePopup, which creates a
1425         normal <iframe> in the top-level document, and load the popup content on
1426         it. MockPagePopupDriver is enabled by
1427         internals.settings.setEnableMockPagePopup(true).
1428
1429         Test: fast/forms/date/calendar-picker-appearance.html
1430
1431         * WebCore.gypi: Add new files.
1432         * WebCore.xcodeproj/project.pbxproj: Add files to make this buildable.
1433         * page/ChromeClient.h:
1434         (ChromeClient): Add function for PagePopupDriver.
1435         * loader/EmptyClients.h: Add empty implementations for PagePopupDriver functions.
1436         * page/PagePopupDriver.h: Added.
1437         (PagePopupDriver):
1438         (WebCore::PagePopupDriver::~PagePopupDriver):
1439
1440         * testing/InternalSettings.cpp:
1441         (WebCore::InternalSettings::Backup::restoreTo): Reset the mock PagePopupDriver.
1442         (WebCore::InternalSettings::setEnableMockPagePopup):
1443         Register MockPagePopupDriver to ChromeClient.
1444         * testing/InternalSettings.h:
1445         (InternalSettings): Declare setEnableMockPagePopup().
1446         * testing/InternalSettings.idl: ditto.
1447         * testing/MockPagePopupDriver.cpp: Added.
1448         (MockPagePopup): Pseudo PagePopup implementation with the standard <iframe>.
1449         (WebCore::MockPagePopupDriver::MockPagePopupDriver): Added.
1450         (WebCore::MockPagePopupDriver::create): Added.
1451         (WebCore::MockPagePopupDriver::~MockPagePopupDriver): Added.
1452         (WebCore::MockPagePopupDriver::openPagePopup):
1453         Added. An override of PagePopupDriver function. This creates a MockPagePopup.
1454         (WebCore::MockPagePopupDriver::closePagePopup):
1455         Added. An override of PagePopupDriver function. This deletes the MockPagePopup.
1456         * testing/MockPagePopupDriver.h:
1457         (MockPagePopupDriver): Added.
1458
1459         * testing/v8/WebCoreTestSupport.cpp:
1460         (WebCoreTestSupport::injectPagePopupController):
1461         Added. Production code uses per-Page context feature framework. However
1462         MockPagePopup uses the same page as the host page. So we can't use the
1463         framework and need to inject window.pagePopupController manually.
1464         * testing/v8/WebCoreTestSupport.h:
1465         (WebCoreTestSupport): Add injectPagePopupController().
1466
1467 2012-07-12  Pavel Feldman  <pfeldman@chromium.org>
1468
1469         Web Inspector: mute the native looks of the selects in the console.
1470         https://bugs.webkit.org/show_bug.cgi?id=91120
1471
1472         Reviewed by Vsevolod Vlasov.
1473
1474         This is necessary for Mac now that we don't use border images for select.
1475
1476         * inspector/front-end/ConsoleView.js:
1477         (WebInspector.ConsoleView.prototype.get statusBarItems):
1478         (WebInspector.ConsoleView.prototype.addContext):
1479         (WebInspector.ConsoleView.prototype.removeContext):
1480         (WebInspector.ConsoleView.prototype._updateIsolatedWorldSelector):
1481         (WebInspector.ConsoleView.prototype._appendIsolatedContextOption):
1482         (WebInspector.ConsoleView.prototype._currentEvaluationContext):
1483         (WebInspector.ConsoleView.prototype._currentIsolatedContextId):
1484         * inspector/front-end/StatusBarButton.js:
1485         (WebInspector.StatusBarComboBox):
1486         (WebInspector.StatusBarComboBox.prototype.addOption):
1487         (WebInspector.StatusBarComboBox.prototype.removeOption):
1488         (WebInspector.StatusBarComboBox.prototype.removeOptions):
1489         (WebInspector.StatusBarComboBox.prototype.selectedOption):
1490         * inspector/front-end/inspector.css:
1491         (.status-bar-select-container):
1492         (select.status-bar-item):
1493         (.console-context):
1494
1495 2012-07-13  Andrei Bucur  <abucur@adobe.com>
1496
1497         [CSS Regions] Fix the lifecycle for the flow objects and their renderers
1498         https://bugs.webkit.org/show_bug.cgi?id=89000
1499
1500         Reviewed by Eric Seidel.
1501
1502         This patch adds the concept of a NamedFlowCollection, owned by the document, that keeps track of
1503         all the named flows that exist in the Document. This collection contains a ListHashSet of weak references to
1504         all the existing NamedFlows in the document. This is not a managed set because the CREATED flows are referenced from the renderer and
1505         the NULL flows are only cached, they should be destructible.
1506         Two named flows are considered to be equal if they have the same name.
1507         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.
1508         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
1509         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
1510         so when they go back to the CREATED state, the listeners would still be there.
1511
1512         Link to spec: http://www.w3.org/TR/2012/WD-css3-regions-20120503/
1513
1514         Tests: The old tests have been modified to take into account the new behavior
1515
1516         * CMakeLists.txt:
1517         * GNUmakefile.list.am:
1518         * Target.pri:
1519         * WebCore.gypi:
1520         * WebCore.vcproj/WebCore.vcproj:
1521         * WebCore.xcodeproj/project.pbxproj:
1522         * dom/DOMAllInOne.cpp:
1523         * dom/Document.cpp:
1524         (WebCore::Document::~Document):
1525         (WebCore):
1526         (WebCore::Document::webkitGetFlowByName):
1527         (WebCore::Document::namedFlows):
1528         * dom/Document.h:
1529         (WebCore):
1530         (Document):
1531         * dom/WebKitNamedFlow.cpp:
1532         (WebCore::WebKitNamedFlow::WebKitNamedFlow):
1533         (WebCore::WebKitNamedFlow::~WebKitNamedFlow):
1534         (WebCore::WebKitNamedFlow::create):
1535         (WebCore):
1536         (WebCore::WebKitNamedFlow::name):
1537         (WebCore::WebKitNamedFlow::overset):
1538         (WebCore::nodeInFlowThread):
1539         (WebCore::WebKitNamedFlow::getRegionsByContentNode):
1540         (WebCore::WebKitNamedFlow::getContent):
1541         (WebCore::WebKitNamedFlow::setRenderer):
1542         * dom/WebKitNamedFlow.h:
1543         (WebCore):
1544         (WebKitNamedFlow):
1545         (WebCore::WebKitNamedFlow::getFlowState):
1546         (WebCore::WebKitNamedFlow::switchFlowState):
1547         * dom/WebKitNamedFlowCollection.cpp: Added.
1548         (WebCore):
1549         (WebCore::WebKitNamedFlowCollection::WebKitNamedFlowCollection):
1550         (WebCore::WebKitNamedFlowCollection::length): An O(1) operation
1551         (WebCore::WebKitNamedFlowCollection::item): An O(N) operation
1552         (WebCore::WebKitNamedFlowCollection::flowByName): An O(1) operation
1553         (WebCore::WebKitNamedFlowCollection::ensureNamedFlowInCreatedState): An O(1) operation
1554         (WebCore::WebKitNamedFlowCollection::moveNamedFlowToNullState): An O(1) operation
1555         (WebCore::WebKitNamedFlowCollection::discardNamedFlow): An O(1) operation
1556         (WebCore::WebKitNamedFlowCollection::documentDestroyed):
1557         (WebCore::WebKitNamedFlowCollection::NamedFlowHashFunctions::hash):
1558         (WebCore::WebKitNamedFlowCollection::NamedFlowHashFunctions::equal):
1559         (WebKitNamedFlowCollection::NamedFlowHashFunctions):
1560         (WebCore::WebKitNamedFlowCollection::NamedFlowHashTranslator::hash):
1561         (WebCore::WebKitNamedFlowCollection::NamedFlowHashTranslator::equal):
1562         * dom/WebKitNamedFlowCollection.h: Copied from Source/WebCore/dom/WebKitNamedFlow.h.
1563         (WebCore):
1564         (WebKitNamedFlowCollection):
1565         (WebCore::WebKitNamedFlowCollection::create):
1566         (WebCore::WebKitNamedFlowCollection::document):
1567         * rendering/FlowThreadController.cpp:
1568         (WebCore::FlowThreadController::ensureRenderFlowThreadWithName):
1569         (WebCore::FlowThreadController::removeFlowThread):
1570         (WebCore):
1571         * rendering/FlowThreadController.h:
1572         (FlowThreadController):
1573         * rendering/RenderNamedFlowThread.cpp:
1574         (WebCore::RenderNamedFlowThread::RenderNamedFlowThread):
1575         (WebCore):
1576         (WebCore::RenderNamedFlowThread::~RenderNamedFlowThread):
1577         (WebCore::RenderNamedFlowThread::removeRegionFromThread):
1578         (WebCore::RenderNamedFlowThread::unregisterNamedFlowContentNode):
1579         (WebCore::RenderNamedFlowThread::flowThreadName):
1580         (WebCore::RenderNamedFlowThread::willBeDestroyed):
1581         * rendering/RenderNamedFlowThread.h:
1582         (RenderNamedFlowThread):
1583         (WebCore::RenderNamedFlowThread::contentNodes):
1584         (WebCore::RenderNamedFlowThread::canBeDestroyed):
1585
1586 2012-07-13  Vineet Chaudhary  <rgf748@motorola.com>
1587
1588         [V8Bindings] Implement generalised method to validates that the passed object is a sequence type.
1589         https://bugs.webkit.org/show_bug.cgi?id=91056
1590
1591         Reviewed by Kentaro Hara.
1592
1593         Currently the V8 implementation validates that the passed object is a sequence type only
1594         for MessagePort in V8Utilities::extractTransferables().
1595         There should be generalised method for other types too.
1596         Spec URL: http://www.w3.org/TR/2012/WD-WebIDL-20120207/#es-sequence
1597
1598         No new test, Just refactoring. There should be no behavioral changes.
1599
1600         * bindings/v8/V8Binding.h:
1601         (WebCore::toV8Sequence): Added implementation of toV8Sequence().
1602
1603 2012-07-13  Zeno Albisser  <zeno@webkit.org>
1604
1605         [Qt][WK2] Implement GraphicsSurface for Linux/GLX.
1606         https://bugs.webkit.org/show_bug.cgi?id=90881
1607
1608         Add a GLX based GraphicsSurface implementation for Linux.
1609         Native X windows are being used for exchanging textures
1610         with the UIProcess.
1611
1612         Reviewed by Noam Rosenthal.
1613
1614         * Target.pri:
1615         * WebCore.pri:
1616         * platform/graphics/surfaces/GraphicsSurface.cpp:
1617         (WebCore::GraphicsSurface::create):
1618             Move creating GraphicsSurface instance into
1619             platformCreate/platformImport functions to allow
1620             platform specific creation based on the provided flags.
1621         (WebCore::GraphicsSurface::GraphicsSurface):
1622         (WebCore::GraphicsSurface::~GraphicsSurface):
1623             Call platformDestroy when destroying a GraphicsSurface.
1624         (WebCore):
1625         * platform/graphics/surfaces/GraphicsSurface.h:
1626             Make platformCreate/platformImport functions static
1627             to be able to call these from the static create function.
1628             Add Destructor prototype and add GraphicsSurfacePrivate member.
1629         (WebCore):
1630         (GraphicsSurface):
1631         * platform/graphics/surfaces/mac/GraphicsSurfaceMac.cpp:
1632         (WebCore):
1633         (WebCore::GraphicsSurface::platformCreate):
1634         (WebCore::GraphicsSurface::platformImport):
1635             Insert creation of GraphicsSurface instance.
1636             This allows having a platform specific creation mechanism
1637             for GLX.
1638         * platform/graphics/surfaces/qt/GraphicsSurfaceGLX.cpp: Added.
1639         (WebCore):
1640         (OffScreenRootWindow):
1641         (WebCore::OffScreenRootWindow::OffScreenRootWindow):
1642         (WebCore::OffScreenRootWindow::get):
1643         (WebCore::OffScreenRootWindow::~OffScreenRootWindow):
1644             Add an OffScreenRootWindow singelton that is being used
1645             as a parent for all native offscreen windows.
1646         (GraphicsSurfacePrivate):
1647             This class is used to manage all the X related resources
1648             such as opening a display or creating XPixmaps etc.
1649         (WebCore::GraphicsSurfacePrivate::GraphicsSurfacePrivate):
1650             Open a connection to the X server and create a
1651             QOpenGLContext that can be used to resolve GL functions.
1652         (WebCore::GraphicsSurfacePrivate::~GraphicsSurfacePrivate):
1653             Properly cleanup and release all the X resources again.
1654         (WebCore::GraphicsSurfacePrivate::createSurface):
1655             Create a surface that is placed off screen and can be
1656             used as a rendering target by the WebProcess.
1657         (WebCore::GraphicsSurfacePrivate::createPixmap):
1658             Create a GLXPixmap from the XWindow that was previously
1659             redirected by the WebProcess. This GLXPixmap can then be
1660             bound to a texture and being painted on screen by the
1661             UIProcess.
1662         (WebCore::GraphicsSurfacePrivate::makeCurrent):
1663         (WebCore::GraphicsSurfacePrivate::swapBuffers):
1664         (WebCore::GraphicsSurfacePrivate::display):
1665         (WebCore::GraphicsSurfacePrivate::glxPixmap):
1666         (WebCore::GraphicsSurfacePrivate::size):
1667         (WebCore::GraphicsSurfacePrivate::glContext):
1668         (WebCore::resolveGLMethods):
1669             Initialize all the function pointers for the GL functions used.
1670         (WebCore::GraphicsSurface::platformExport):
1671         (WebCore::GraphicsSurface::platformGetTextureID):
1672             Bind the GLXPixmap to a texture.
1673         (WebCore::GraphicsSurface::platformCopyToGLTexture):
1674             Not beeing implemented for GLX.
1675         (WebCore::GraphicsSurface::platformCopyFromFramebuffer):
1676             Blit origin fbo onto the GraphicsSurface's backing.
1677         (WebCore::GraphicsSurface::platformCreate):
1678         (WebCore::GraphicsSurface::platformImport):
1679         (WebCore::GraphicsSurface::platformLock):
1680         (WebCore::GraphicsSurface::platformUnlock):
1681         (WebCore::GraphicsSurface::platformDestroy):
1682
1683 2012-07-13  Dongwoo Im  <dw.im@samsung.com>
1684
1685         CodeGeneratorJS.pm : SetterExpression should use 'push' rather than 'unshift'
1686         https://bugs.webkit.org/show_bug.cgi?id=91198
1687
1688         Reviewed by Kentaro Hara.
1689
1690         'SetterExpression' should use 'push' to make arguments, rather than 'unshift'.
1691
1692         No new tests. Covered by existing tests.
1693
1694         * bindings/scripts/CodeGeneratorJS.pm:
1695         (GenerateImplementation):
1696
1697 2012-07-13  Yoshifumi Inoue  <yosin@chromium.org>
1698
1699         REGRESSION(r119948): [Forms] Spin button Up/Down actions make value to zero for input type "number" when step mismatched
1700         https://bugs.webkit.org/show_bug.cgi?id=91197
1701
1702         Reviewed by Kent Tamura.
1703
1704         This patch fixes implementation of Decimal::ceiling() and floor().
1705         They return wrong value for small fractional numbers.
1706
1707         The bug is occurred when:
1708           - Step-able input type, e.g. number, date, datetime, and so on.
1709           - Current value is step mismatched
1710           - Current value is smaller than step
1711           - Step up/down by spin button
1712         because spin button up/down actions are implemented in InputType::setpUpFromRenderer
1713         which calls Decimal::ceiling() and floor() for step mismatched case.
1714
1715         Tests: fast/forms/number/number-stepup-stepdown-from-renderer.html: Added test cases
1716                WebKit/chromium/tests/DecimalTest.cpp: Added test cases
1717
1718         * platform/Decimal.cpp:
1719         (WebCore::Decimal::ceiling): Changed to return 1 for positive small fractional number.
1720         (WebCore::Decimal::floor): Changed to return -1 for negative small fractional number.
1721
1722 2012-07-13  Dominic Mazzoni  <dmazzoni@google.com>
1723
1724         Should be possible to focus elements within canvas fallback content
1725         https://bugs.webkit.org/show_bug.cgi?id=87898
1726
1727         Reviewed by Chris Fleizach.
1728
1729         Patches isFocusable in dom/Node.cpp and html/HTMLFormControlElement.cpp
1730         to make elements focusable if they're a descendent of a canvas element
1731         if they would otherwise have been focusable but just didn't have
1732         a renderer. Adds a bit to ElementRareData to efficiently keep track
1733         of elements in a canvas subtree.
1734
1735         Test: fast/canvas/fallback-content.html
1736
1737         * dom/Element.cpp:
1738         (WebCore::Element::attach):
1739         (WebCore::Element::detach):
1740         (WebCore::Element::setIsInCanvasSubtree):
1741         (WebCore):
1742         (WebCore::Element::isInCanvasSubtree):
1743         * dom/Element.h:
1744         (Element):
1745         * dom/ElementRareData.h:
1746         (ElementRareData):
1747         (WebCore::ElementRareData::ElementRareData):
1748         * dom/Node.cpp:
1749         (WebCore::Node::isFocusable):
1750         * html/HTMLCanvasElement.cpp:
1751         (WebCore::HTMLCanvasElement::attach):
1752         (WebCore):
1753         * html/HTMLCanvasElement.h:
1754         (HTMLCanvasElement):
1755         * html/HTMLFormControlElement.cpp:
1756         (WebCore::HTMLFormControlElement::isFocusable):
1757
1758 2012-07-12  Carlos Garcia Campos  <cgarcia@igalia.com>
1759
1760         [GTK] Add API to get HTTPS status to WebKit2 GTK+
1761         https://bugs.webkit.org/show_bug.cgi?id=91100
1762
1763         Reviewed by Martin Robinson.
1764
1765         * platform/network/soup/ResourceResponse.h:
1766         (WebCore::ResourceResponse::soupMessageCertificate): Return the
1767         certificate.
1768         (WebCore::ResourceResponse::setSoupMessageCertificate): Set a
1769         certificate.
1770         (WebCore::ResourceResponse::soupMessageTLSErrors): Return the TLS
1771         errors.
1772         (WebCore::ResourceResponse::setSoupMessageTLSErrors): Set TLS
1773         errors.
1774         * platform/network/soup/ResourceResponseSoup.cpp:
1775         (WebCore::ResourceResponse::toSoupMessage): Set the certificate
1776         and TLS errors to the newly created SoupMessage.
1777         (WebCore::ResourceResponse::updateFromSoupMessage): Get the
1778         certificate and TLS errors from the SoupMessage.
1779
1780 2012-07-13  Ryosuke Niwa  <rniwa@webkit.org>
1781
1782         RadioNodeList is not updated upon input type change
1783         https://bugs.webkit.org/show_bug.cgi?id=91178
1784
1785         Reviewed by Alexey Proskuryakov.
1786
1787         Invalidate the radio node lists when type content attribute changes since it excludes
1788         image type input elements.
1789
1790         Test: fast/forms/radionodelist-image-type.html
1791
1792         * dom/Document.cpp:
1793         (WebCore::shouldInvalidateNodeListForType):
1794         * dom/Document.h: Renamed InvalidateOnIdNameForAttrChange to InvalidateOnFormAttrChange
1795         since listing all attribute name isn't useful at this point.
1796         * html/RadioNodeList.cpp:
1797         (WebCore::RadioNodeList::RadioNodeList):
1798
1799 2012-07-12  Dongwoo Im  <dw.im@samsung.com>
1800
1801         CodeGeneratorJS.pm need to handle the attribute which has "CallWith=ScriptExecutionContext" option.
1802         https://bugs.webkit.org/show_bug.cgi?id=91185
1803
1804         Reviewed by Kentaro Hara.
1805
1806         When an attribute has "CallWith=ScriptExecutionContext" option, 'ScriptExecutionContext*' parameter should be the last parameter.
1807
1808         No new tests. Covered by existing tests.
1809
1810         * bindings/scripts/CodeGeneratorJS.pm:
1811         (GenerateImplementation):
1812
1813 2012-07-12  Yoshifumi Inoue  <yosin@chromium.org>
1814
1815         REGRESSION(r117738): [Forms] stepMismatch for input type "time" with large step value always return false.
1816         https://bugs.webkit.org/show_bug.cgi?id=91062
1817
1818         Reviewed by Kent Tamura.
1819
1820         This patch changes value of StepRange::acceptableError to zero for
1821         integer restricted step value.
1822
1823         No new tests, existing test covers (fast/forms/time/ValidityState-stepMismatch-time.html) this change although they are listed in TestExpectation file.
1824
1825         * html/StepRange.cpp:
1826         (WebCore::StepRange::acceptableError): Changed to return 0 if step value is restricted to be an integer.
1827
1828 2012-07-12  Dan Bernstein  <mitz@apple.com>
1829
1830         Pass an option flag to CFStringGetHyphenationLocationBeforeIndex() that tells it to not
1831         terminate the search at the last word boundary before the given index.
1832
1833         Reviewed by Adele Peterson.
1834
1835         No test, because the current version of Core Foundation ignores the options parameter.
1836
1837         * platform/text/cf/HyphenationCF.cpp:
1838         (WebCore::lastHyphenLocation): Changed the options parameter from 0 to 1.
1839
1840 2012-07-12  Eric Seidel  <eric@webkit.org>
1841
1842         Incorrect behaviour calling Range setStart or setEnd with boundary in different document
1843         https://bugs.webkit.org/show_bug.cgi?id=42517
1844
1845         Reviewed by Ojan Vafai.
1846
1847         Added a new static inline "checkForDifferentRootContainer" to share some code
1848         and made setStart/setEnd do the right thing in the x-document case.  I removed
1849         the bogus checks in set*After/set*Before functions, and since they just call
1850         through to setStart/setEnd, they also now do the right thing.
1851
1852         Test: fast/dom/Range/set-wrong-document-err.html
1853
1854         * dom/Range.cpp:
1855         (WebCore::checkForDifferentRootContainer):
1856         (WebCore):
1857         (WebCore::Range::setStart):
1858         (WebCore::Range::setEnd):
1859         (WebCore::Range::setStartAfter):
1860         (WebCore::Range::setEndBefore):
1861         (WebCore::Range::setEndAfter):
1862         (WebCore::Range::setStartBefore):
1863
1864 2012-07-12  Erik Arvidsson  <arv@chromium.org>
1865
1866         [V8] Simplify CodeGeneratorV8 since V8OnProto is only true for DOMWindow
1867         https://bugs.webkit.org/show_bug.cgi?id=91165
1868
1869         Reviewed by Nate Chapin.
1870
1871         The old code was dead code since V8OnProto only ever gets set to 1 for DOMWindow.
1872
1873         No new tests. No change in functionality.
1874
1875         * bindings/scripts/CodeGeneratorV8.pm:
1876         (GenerateNormalAttrSetter):
1877
1878 2012-07-12  Ryosuke Niwa  <rniwa@webkit.org>
1879
1880         Build fix. Initialize unused const member variables to make compilers happy.
1881
1882         * dom/DynamicNodeList.h:
1883         (WebCore::DynamicNodeListCacheBase::DynamicNodeListCacheBase):
1884
1885 2012-07-12  Konrad Piascik  <kpiascik@rim.com>
1886
1887         Web Inspector: Geolocation override
1888         https://bugs.webkit.org/show_bug.cgi?id=89365
1889
1890         Reviewed by Pavel Feldman.
1891
1892         Reverted original patch.
1893         Change has not been reviewed by the right people. It declares
1894         public protocol methods and is wrong in several ways.
1895
1896         * Modules/geolocation/GeolocationController.cpp:
1897         (WebCore::GeolocationController::GeolocationController):
1898         (WebCore::GeolocationController::positionChanged):
1899         * Modules/geolocation/GeolocationController.h:
1900         (GeolocationController):
1901         * inspector/Inspector.json:
1902         * inspector/InspectorInstrumentation.cpp:
1903         (WebCore):
1904         * inspector/InspectorInstrumentation.h:
1905         (WebCore):
1906         (InspectorInstrumentation):
1907         * inspector/InspectorPageAgent.cpp:
1908         (WebCore::InspectorPageAgent::InspectorPageAgent):
1909         * inspector/InspectorPageAgent.h:
1910         * inspector/front-end/Settings.js:
1911         * inspector/front-end/SettingsScreen.js:
1912         (WebInspector.UserAgentSettingsTab):
1913         (WebInspector.UserAgentSettingsTab.prototype._createDeviceMetricsElement):
1914         * inspector/front-end/UserAgentSupport.js:
1915
1916 2012-07-12  Ryosuke Niwa  <rniwa@webkit.org>
1917
1918         Move m_type and m_hasNameCache from HTMLCollectionCacheBase to DynamicNodeListCacheBase for better bit packing
1919         https://bugs.webkit.org/show_bug.cgi?id=91164
1920
1921         Reviewed by Anders Carlsson.
1922
1923         Moved m_type and m_hasNameCache from HTMLCollection and renamed them to m_collectionType and m_isNameCacheValid.
1924
1925         Also renamed shouldIncludeChildren to shouldOnlyIncludeDirectChildren and negated the return value since
1926         all HTMLCollection include children in the collection and the function was meant to tell us whether the collection
1927         should include descendents that are not direct children of base().
1928
1929         In addition, renamed nextNodeOrSibling to nextNode since "or sibling" doesn't seem to add any semantic clarity.
1930
1931         * dom/DynamicNodeList.h:
1932         (WebCore::DynamicNodeListCacheBase::DynamicNodeListCacheBase):
1933         (DynamicNodeListCacheBase):
1934         (WebCore::DynamicNodeListCacheBase::type):
1935         (WebCore::DynamicNodeListCacheBase::hasNameCache):
1936         (WebCore::DynamicNodeListCacheBase::setHasNameCache):
1937         (WebCore::DynamicNodeListCacheBase::clearCache):
1938         * html/HTMLCollection.cpp:
1939         (WebCore::shouldOnlyIncludeDirectChildren):
1940         (WebCore::HTMLCollection::HTMLCollection):
1941         (WebCore::HTMLCollection::isAcceptableElement):
1942         (WebCore::nextNode):
1943         (WebCore::HTMLCollection::itemAfter):
1944         * html/HTMLCollection.h:
1945         (WebCore::HTMLCollectionCacheBase::HTMLCollectionCacheBase):
1946         (HTMLCollectionCacheBase):
1947         (WebCore::HTMLCollectionCacheBase::clearCache):
1948
1949 2012-07-12  Shinya Kawanaka  <shinyak@chromium.org>
1950
1951         Needs callback before AuthorShadowRoot is added.
1952         https://bugs.webkit.org/show_bug.cgi?id=91167
1953
1954         Reviewed by Hajime Morita.
1955
1956         This is a preliminary patch for adding multiple Shadow DOM support for media elements and form elements.
1957         They assume that UserAgentShadowRoot is the oldest, however currently a page author may try to add
1958         AuthorShadowRoot before adding UserAgentShadowRoot.
1959
1960         This patch adds a callback before AuthorShadowRoot is being added, and allow us to add UserAgentShadowRoot
1961         for those elements. See also Bug 77936, Bug 77937, Bug 90532.
1962
1963         * dom/Element.h:
1964         (WebCore::Element::willAddAuthorShadowRoot):
1965         * dom/ElementShadow.cpp:
1966         (WebCore::ElementShadow::addShadowRoot):
1967         * dom/ElementShadow.h:
1968         (ElementShadow):
1969         * dom/ShadowRoot.cpp:
1970         (WebCore::ShadowRoot::create):
1971
1972 2012-07-12  Dana Jansens  <danakj@chromium.org>
1973
1974         [chromium] Remove the RenderPass pointer from RenderPassDrawQuad
1975         https://bugs.webkit.org/show_bug.cgi?id=91023
1976
1977         Reviewed by Adrienne Walker.
1978
1979         Removes the RenderPass pointer and keeps only an integer id in the
1980         quad to refer back to the RenderPass the quad reads from.
1981
1982         Covered by existing tests.
1983
1984         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
1985         (WebCore::CCLayerTreeHostImpl::calculateRenderPasses):
1986         (WebCore::findRenderPassById):
1987         (WebCore):
1988         (WebCore::removeRenderPassesRecursive):
1989         (WebCore::CCLayerTreeHostImpl::CullRenderPassesWithCachedTextures::shouldRemoveRenderPass):
1990         (WebCore::CCLayerTreeHostImpl::CullRenderPassesWithNoQuads::shouldRemoveRenderPass):
1991         (WebCore::CCLayerTreeHostImpl::removeRenderPasses):
1992         (WebCore::CCLayerTreeHostImpl::prepareToDraw):
1993         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.h:
1994         (FrameData):
1995         (CullRenderPassesWithCachedTextures):
1996         (CullRenderPassesWithNoQuads):
1997         (CCLayerTreeHostImpl):
1998         * platform/graphics/chromium/cc/CCRenderPass.cpp:
1999         (WebCore::CCRenderPass::appendQuadsForRenderSurfaceLayer):
2000         * platform/graphics/chromium/cc/CCRenderPass.h:
2001         (WebCore):
2002         * platform/graphics/chromium/cc/CCRenderPassDrawQuad.cpp:
2003         (WebCore::CCRenderPassDrawQuad::create):
2004         (WebCore::CCRenderPassDrawQuad::CCRenderPassDrawQuad):
2005         * platform/graphics/chromium/cc/CCRenderPassDrawQuad.h:
2006         (CCRenderPassDrawQuad):
2007         * platform/graphics/chromium/cc/CCRenderSurface.cpp:
2008         (WebCore::CCRenderSurface::appendQuads):
2009         * platform/graphics/chromium/cc/CCRenderSurface.h:
2010         (CCRenderSurface):
2011
2012 2012-07-12  Elliott Sprehn  <esprehn@gmail.com>
2013
2014         Unneeded tree walking when adding or removing children due to RenderCounter / RenderQuote logic
2015         https://bugs.webkit.org/show_bug.cgi?id=89900
2016
2017         Reviewed by Eric Seidel and Abhishek Arya.
2018
2019         Previously we would walk the all children a renderer whenever adding
2020         or removing a child renderer from its RenderObjectChildList to look for
2021         RenderQuote and RenderCounter instances to update. This patch introduces
2022         a counter in RenderView for the number of RenderQuote and RenderCounter
2023         instances in that document so we can avoid these traversals.
2024
2025         No tests needed since this is just a short circuiting of logic and the existing
2026         tests should cover it.
2027
2028         * rendering/RenderCounter.cpp:
2029         (WebCore::RenderCounter::RenderCounter): Increment instance counter.
2030         (WebCore::RenderCounter::willBeDestroyed): Decrement instance counter.
2031         (WebCore):
2032         (WebCore::RenderCounter::rendererRemovedFromTree): Short circuit when counter is zero.
2033         (WebCore::RenderCounter::rendererSubtreeAttached): Short circuit when counter is zero.
2034         * rendering/RenderCounter.h:
2035         (RenderCounter):
2036         * rendering/RenderObjectChildList.cpp:
2037         (WebCore::RenderObjectChildList::removeChildNode): Short circuit calling into Counter and Quote code when the document is being destroyed.
2038         * rendering/RenderQuote.cpp:
2039         (WebCore::RenderQuote::RenderQuote):
2040         (WebCore::RenderQuote::willBeDestroyed):
2041         (WebCore):
2042         (WebCore::RenderQuote::rendererSubtreeAttached): Increment instance counter.
2043         (WebCore::RenderQuote::rendererRemovedFromTree): Decrement instance counter.
2044         * rendering/RenderQuote.h:
2045         (RenderQuote):
2046         * rendering/RenderView.cpp:
2047         (WebCore::RenderView::RenderView):
2048         * rendering/RenderView.h: Methods for managing the RenderQuote and RenderCounter counts.
2049         (RenderView):
2050         (WebCore::RenderView::addRenderQuote):
2051         (WebCore::RenderView::removeRenderQuote):
2052         (WebCore::RenderView::hasRenderQuotes):
2053         (WebCore::RenderView::addRenderCounter):
2054         (WebCore::RenderView::removeRenderCounter):
2055         (WebCore::RenderView::hasRenderCounters):
2056
2057 2012-07-12  Ryosuke Niwa  <rniwa@webkit.org>
2058
2059         Merge HTMLCollectionWithArrayStorage into HTMLCollection
2060         https://bugs.webkit.org/show_bug.cgi?id=91144
2061
2062         Reviewed by Anders Carlsson.
2063
2064         Merged HTMLCollectionWithArrayStorage::item into HTMLCollection::item and got rid of
2065         HTMLCollectionWithArrayStorage. Also de-virtualized HTMLCollection::item and HTMLCollection::length
2066         and merged itemInArrayAfter and itemAfter.
2067
2068         In addition, improved the algorithm in HTMLCollection::length to take advantage of item cache.
2069         Instead of always computing the length from the beginning, we start the search from the cached item
2070         so that if we're near end of the collection, we avoid significant portion of the node traversal.
2071
2072         Furthermore, made HTMLCollection always call setItemCache with a non-null item and HTMLCollection::item
2073         set the length cache when it reaches the end of the collection to avoid redundant length calculations.
2074
2075         * dom/DynamicNodeList.h:
2076         (WebCore::DynamicNodeListCacheBase::setItemCache): Add a FIXME.
2077         * html/HTMLCollection.cpp:
2078         (WebCore::HTMLCollection::itemAfter): Regular HTMLCollection doesn't have uses elements array so
2079         assert that offsetInArray is always 0.
2080         (WebCore): Removed calcLength.
2081         (WebCore::HTMLCollection::length): Rewritten. The algorithm is as follows:
2082         When there is no item cache, we look for the first item by calling item(0). If item(0) returns null,
2083         then it must have set the length cache so bail out. If the previous step didn't bail out, then
2084         the item cache is valid and not null (see above), so count the number of remaining items in collection
2085         starting at the cached item's offset + 1.
2086         (WebCore::HTMLCollection::item): Avoid calling setItemCache with null. When the first item is null,
2087         set the length cache instead.
2088         (WebCore::HTMLCollection::itemAfterCachedItem): Extracted from HTMLCollectionWithArrayStorage::item.
2089         (WebCore::HTMLCollection::namedItem): Pass in arrayOffset to itemAfter. Note this variable is never
2090         used since only HTMLFormCollection and HTMLPropertiesCollection use array offsets but they override
2091         this function.
2092         (WebCore::HTMLCollection::updateNameCache): Ditto.
2093         * html/HTMLCollection.h:
2094         (HTMLCollection):
2095         (WebCore):
2096         * html/HTMLFormCollection.cpp: Removed calcLength(). Even though this function didn't iterate over
2097         the collection directly, HTMLFormElement::length and HTMLFieldSetElement::length did so we're not
2098         regressing any performance here.
2099         (WebCore::HTMLFormCollection::HTMLFormCollection):
2100         (WebCore::HTMLFormCollection::itemAfter):
2101         * html/HTMLFormCollection.h:
2102         (HTMLFormCollection):
2103         * html/HTMLNameCollection.cpp:
2104         (WebCore::HTMLNameCollection::itemAfter):
2105         * html/HTMLNameCollection.h:
2106         (HTMLNameCollection):
2107         * html/HTMLPropertiesCollection.cpp: Removed calcLength().
2108         (WebCore::HTMLPropertiesCollection::HTMLPropertiesCollection):
2109         (WebCore::HTMLPropertiesCollection::itemAfter):
2110         (WebCore):
2111         * html/HTMLPropertiesCollection.h:
2112         (HTMLPropertiesCollection):
2113         * html/HTMLTableRowsCollection.cpp:
2114         (WebCore::HTMLTableRowsCollection::itemAfter):
2115         * html/HTMLTableRowsCollection.h:
2116         (HTMLTableRowsCollection):
2117
2118 2012-07-12  Pravin D  <pravind.2k4@gmail.com>
2119
2120         Row size/position is wrongly calculated when table having overlapping rowspan cell and colspan cell
2121         https://bugs.webkit.org/show_bug.cgi?id=16811
2122
2123         Reviewed by Julien Chaffraix.
2124
2125         The height of a row is calculated by taking the max height of the cells contained in it. When a row contains
2126         a rowSpan cell and if this row is not the last row of the cell, then its height is max height of other non
2127         rowSpan cells. If the row is the last row of the rowSpan cell, then using the contraint laid by CSS2.1 spec
2128         "For a rowSpan cell, the sum of the row heights involved must be great enough to encompass the cell spanning the rows",
2129         the last remaining height of the rowSpan(cell height minus heights of other involved rows) is taken into consideration
2130         while calculating the height of this row.
2131         Currently when calculating the height of the row we are only using the height of the primary cell at position (row, col).
2132         However when a row has colSpan cell and rowSpan, they might overlap. In such a sitution as only the primary cells height
2133         is considered, the height of the row will be calculated worngly if the other overlapping cell has greater height.
2134         Thus all the overlapping cell at position (row, col) must be considered while calculating the height of a row.
2135
2136         Test: fast/table/last-cell-of-rowspan-overlapping-colspan-cell.html
2137
2138         * rendering/RenderTableSection.cpp:
2139         (WebCore::RenderTableSection::calcRowLogicalHeight):
2140         Fixed function to use all the overlapping cells at position(row, col) to calculate the height/position of rows.
2141
2142 2012-07-12  Joshua Bell  <jsbell@chromium.org>
2143
2144         IndexedDB: Enable IDBFactory.deleteDatabase() and webkitGetDatabaseNames() in Workers
2145         https://bugs.webkit.org/show_bug.cgi?id=90310
2146
2147         Reviewed by Tony Chang.
2148
2149         Simplify Document vs. Worker logic for IDBFactory::open() and hook up the
2150         other two IDBFactory methods for use by workers as well.
2151
2152         Test: storage/indexeddb/factory-basics-workers.html
2153
2154         * Modules/indexeddb/IDBFactory.cpp:
2155         (isContextValid): Helper function consolidating checks that context is usable.
2156         (getIndexedDBDatabasePath): Helper function for accessing group settings.
2157         (WebCore::IDBFactory::getDatabaseNames): Simplify - just pass context through to back end.
2158         (WebCore::IDBFactory::open): Ditto.
2159         (WebCore::IDBFactory::deleteDatabase): Ditto.
2160         (WebCore::IDBFactory::cmp): Whitespace.
2161         * Modules/indexeddb/IDBFactoryBackendImpl.cpp: Obsolete openFromWorker() removed.
2162         (WebCore::IDBFactoryBackendImpl::getDatabaseNames): Signature updated.
2163         (WebCore::IDBFactoryBackendImpl::deleteDatabase): Signature updated.
2164         (WebCore::IDBFactoryBackendImpl::open): Signature updated.
2165         * Modules/indexeddb/IDBFactoryBackendImpl.h:
2166         (IDBFactoryBackendImpl):
2167         * Modules/indexeddb/IDBFactoryBackendInterface.h: Interface methods now take both SecurityOrigin
2168         and ScriptExecutionContext, but not Frame. In the proxy, SecurityOrigin is redundant (can be
2169         accessed from context) but on the real back end the context is null (as Frame was previously).
2170         (IDBFactoryBackendInterface):
2171         * inspector/InspectorIndexedDBAgent.cpp:
2172         (WebCore):
2173         (WebCore::InspectorIndexedDBAgent::requestDatabaseNamesForFrame):
2174         (WebCore::InspectorIndexedDBAgent::requestDatabase):
2175         (WebCore::InspectorIndexedDBAgent::requestData):
2176
2177 2012-07-12  Dana Jansens  <danakj@chromium.org>
2178
2179         [chromium] The root layer should not try create a second RenderSurface for itself
2180         https://bugs.webkit.org/show_bug.cgi?id=91124
2181
2182         Reviewed by Adrienne Walker.
2183
2184         Tests: CCLayerTreeHostImplTest.rootLayerDoesntCreateExtraSurface
2185
2186         * platform/graphics/chromium/cc/CCLayerTreeHostCommon.cpp:
2187         (WebCore::subtreeShouldRenderToSeparateSurface):
2188
2189 2012-07-12  Adam Barth  <abarth@webkit.org>
2190
2191         Regression (r122359) Layout Test html5lib/runner.html is failing
2192         https://bugs.webkit.org/show_bug.cgi?id=91047
2193
2194         Reviewed by Tony Chang.
2195
2196         This ASSERT is bogus because doctypes can be removed from the DOM and
2197         then re-added.
2198
2199         Test: fast/viewport/viewport-legacy-xhtmlmp-remove-and-add.html
2200
2201         * dom/Document.cpp:
2202         (WebCore::Document::setDocType):
2203
2204 2012-07-12  Ojan Vafai  <ojan@chromium.org>
2205
2206         Implied minimum size of flex items is min-content
2207         https://bugs.webkit.org/show_bug.cgi?id=87546
2208
2209         Reviewed by Tony Chang.
2210
2211         http://dev.w3.org/csswg/css3-flexbox/#min-size-auto
2212         In the main axis direction, min-size of auto means min-content.
2213
2214         Test: css3/flexbox/flex-item-min-size.html
2215
2216         * rendering/RenderBox.cpp:
2217         (WebCore::RenderBox::computeLogicalWidthInRegionUsing):
2218         (WebCore::RenderBox::computeContentLogicalHeightUsing):
2219         (WebCore::RenderBox::computeReplacedLogicalWidthUsing):
2220         (WebCore::RenderBox::computeReplacedLogicalHeightUsing):
2221         (WebCore::RenderBox::computePositionedLogicalWidthUsing):
2222         (WebCore::RenderBox::computePositionedLogicalHeightUsing):
2223         It turned out that these FIXMEs are all unnecessary with the changes to RenderFlexibleBox.
2224
2225         * rendering/RenderFlexibleBox.cpp:
2226         (WebCore::RenderFlexibleBox::computeMainAxisPreferredSizes):
2227         (WebCore::RenderFlexibleBox::adjustChildSizeForMinAndMax):
2228
2229 2012-07-12  Eric Penner  <epenner@google.com>
2230
2231         [chromium] Use CCTexture/TextureAllocator and remove TextureManager
2232         https://bugs.webkit.org/show_bug.cgi?id=91001
2233
2234         Reviewed by Adrienne Walker.
2235
2236         Use CCTexture to clean up CCPrioritizedTexture::Backing.
2237         Add TextureAllocator.h and remove remainder of TextureManager.h/cpp.
2238         Minor move/refactoring of link/unlink in CCPrioritizedTexture.
2239         Remove double initialization of the default memory limit.
2240
2241         Covered by existing tests (refactoring).
2242
2243         * WebCore.gypi:
2244         * platform/graphics/chromium/BitmapCanvasLayerTextureUpdater.cpp:
2245         * platform/graphics/chromium/BitmapCanvasLayerTextureUpdater.h:
2246         (WebCore):
2247         * platform/graphics/chromium/BitmapSkPictureCanvasLayerTextureUpdater.cpp:
2248         * platform/graphics/chromium/BitmapSkPictureCanvasLayerTextureUpdater.h:
2249         (WebCore):
2250         * platform/graphics/chromium/ImageLayerChromium.cpp:
2251         * platform/graphics/chromium/LayerRendererChromium.cpp:
2252         (WebCore::LayerRendererChromium::initialize):
2253         * platform/graphics/chromium/ManagedTexture.cpp: Removed.
2254         * platform/graphics/chromium/ManagedTexture.h: Removed.
2255         * platform/graphics/chromium/RenderSurfaceChromium.h:
2256         (WebCore):
2257         * platform/graphics/chromium/TextureAllocator.h: Copied from Source/WebCore/platform/graphics/chromium/cc/CCPriorityCalculator.h.
2258         (WebCore):
2259         (TextureAllocator):
2260         (WebCore::TextureAllocator::~TextureAllocator):
2261         * platform/graphics/chromium/TextureManager.cpp: Removed.
2262         * platform/graphics/chromium/TextureManager.h: Removed.
2263         * platform/graphics/chromium/TiledLayerChromium.cpp:
2264         * platform/graphics/chromium/TrackingTextureAllocator.cpp:
2265         (WebCore::TrackingTextureAllocator::createTexture):
2266         (WebCore::TrackingTextureAllocator::deleteTexture):
2267         * platform/graphics/chromium/TrackingTextureAllocator.h:
2268         * platform/graphics/chromium/cc/CCHeadsUpDisplay.cpp:
2269         * platform/graphics/chromium/cc/CCLayerTreeHost.cpp:
2270         (WebCore::CCLayerTreeHost::prioritizeTextures):
2271         * platform/graphics/chromium/cc/CCLayerTreeHost.h:
2272         (WebCore):
2273         * platform/graphics/chromium/cc/CCLayerTreeHostImpl.cpp:
2274         (WebCore::CCLayerTreeHostImpl::CCLayerTreeHostImpl):
2275         * platform/graphics/chromium/cc/CCPrioritizedTexture.cpp:
2276         (WebCore::CCPrioritizedTexture::CCPrioritizedTexture):
2277         (WebCore::CCPrioritizedTexture::setDimensions):
2278         (WebCore::CCPrioritizedTexture::textureId):
2279         (WebCore::CCPrioritizedTexture::bindTexture):
2280         (WebCore::CCPrioritizedTexture::framebufferTexture2D):
2281         (WebCore::CCPrioritizedTexture::link):
2282         (WebCore):
2283         (WebCore::CCPrioritizedTexture::unlink):
2284         * platform/graphics/chromium/cc/CCPrioritizedTexture.h:
2285         (WebCore):
2286         (WebCore::CCPrioritizedTexture::bytes):
2287         (CCPrioritizedTexture):
2288         (WebCore::CCPrioritizedTexture::haveBackingTexture):
2289         (WebCore::CCPrioritizedTexture::Backing::Backing):
2290         (WebCore::CCPrioritizedTexture::Backing::~Backing):
2291         (WebCore::CCPrioritizedTexture::Backing::owner):
2292         (Backing):
2293         (WebCore::CCPrioritizedTexture::backing):
2294         * platform/graphics/chromium/cc/CCPrioritizedTextureManager.cpp:
2295         (WebCore::CCPrioritizedTextureManager::prioritizeTextures):
2296         (WebCore::CCPrioritizedTextureManager::requestLate):
2297         (WebCore::CCPrioritizedTextureManager::acquireBackingTextureIfNeeded):
2298         (WebCore::CCPrioritizedTextureManager::reduceMemory):
2299         (WebCore::CCPrioritizedTextureManager::clearAllMemory):
2300         (WebCore::CCPrioritizedTextureManager::registerTexture):
2301         (WebCore::CCPrioritizedTextureManager::returnBackingTexture):
2302         (WebCore::CCPrioritizedTextureManager::createBacking):
2303         (WebCore::CCPrioritizedTextureManager::destroyBacking):
2304         (WebCore::CCPrioritizedTextureManager::assertInvariants):
2305         * platform/graphics/chromium/cc/CCPrioritizedTextureManager.h:
2306         (CCPrioritizedTextureManager):
2307         (WebCore::CCPrioritizedTextureManager::defaultMemoryAllocationLimit):
2308         (WebCore::CCPrioritizedTextureManager::compareBackings):
2309         * platform/graphics/chromium/cc/CCPriorityCalculator.h:
2310         * platform/graphics/chromium/cc/CCScopedTexture.cpp:
2311         * platform/graphics/chromium/cc/CCScopedTexture.h:
2312         * platform/graphics/chromium/cc/CCTexture.h:
2313         (WebCore::CCTexture::CCTexture):
2314         * platform/graphics/chromium/cc/CCTiledLayerImpl.cpp:
2315         (WebCore):
2316         * platform/graphics/chromium/cc/CCVideoLayerImpl.cpp:
2317
2318 2012-07-12  Ryosuke Niwa  <rniwa@webkit.org>
2319
2320         Let XCode have its own way.
2321
2322         * WebCore.xcodeproj/project.pbxproj:
2323
2324 2012-07-12  Julien Chaffraix  <jchaffraix@webkit.org>
2325
2326         ASSERT(genChild->isListMarker() || genChild->style()->styleType() == FIRST_LETTER) triggered on flex-box content
2327         https://bugs.webkit.org/show_bug.cgi?id=91003
2328
2329         Reviewed by Abhishek Arya.
2330
2331         Tests: fast/flexbox/assert-generated-deprecated-flexbox.html
2332                fast/flexbox/assert-generated-new-flexbox.html
2333
2334         The issue was that findBeforeAfterParent didn't return the right parent for the flex-box case. This would
2335         make us update the wrong children (and not propagate the style updates properly).
2336
2337         * rendering/RenderObjectChildList.cpp:
2338         (WebCore::findBeforeAfterParent):
2339         Added a check for flex boxes (both deprecated and new).
2340
2341 2012-07-12  Pravin D  <pravind.2k4@gmail.com>
2342
2343         Percentage width replaced element in zero percent/fixed width container block incorrectly rendered.
2344         https://bugs.webkit.org/show_bug.cgi?id=9493
2345
2346         Reviewed by Andy Estes.
2347
2348         When the width of the container is zero percent/fixed value then the width of the replaced element must also be zero.
2349
2350         Test: fast/css/percent-width-img-inside-zero-percent-and-fixed-container.html
2351
2352         * rendering/RenderBox.cpp:
2353         (WebCore::RenderBox::computeReplacedLogicalWidthUsing):
2354           When the containing block's available width is zero there can be two cases.
2355           The containing block is floated/positioned in which case the width of the replaced child element must be its instrinsic width.
2356           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
2357           replaced elment must be zero.
2358
2359 2012-07-11  Ryosuke Niwa  <rniwa@webkit.org>
2360
2361         invalidateNodeListsCacheAfterAttributeChanged should dynamically figure out which attribute needs invalidation
2362         https://bugs.webkit.org/show_bug.cgi?id=91046
2363
2364         Reviewed by Anders Carlsson.
2365
2366         Added an array of counters (m_nodeListCounts) for each set of attributes (NodeListInvalidationType) node lists care about
2367         to Document, and made DynamicSubtreeNodeList's constructor and destructor increment and decrement these counters via
2368         registerDynamicSubtreeNodeList and unregisterDynamicSubtreeNodeList respectively. shouldInvalidateDynamicSubtreeNodeList,
2369         which is called by invalidateNodeListsCacheAfterAttributeChanged, then use these counters to determine whether a given
2370         attribute change should result in node list invalidations.
2371
2372         Also removed m_numNodeListCaches from TreeScope because this counter has now become redundant with m_nodeListCounts.
2373
2374         * dom/ChildNodeList.cpp:
2375         (WebCore::ChildNodeList::ChildNodeList): Do not invalidate on attribute changes.
2376         * dom/ClassNodeList.cpp:
2377         (WebCore::ClassNodeList::ClassNodeList): Invalidate on class attribute changes.
2378         * dom/Document.cpp:
2379         (WebCore::Document::Document): Initialize m_nodeListCounts.
2380         (WebCore::Document::~Document): Add assertions to make sure m_listsInvalidatedAtDocument, m_nodeListCounts, and
2381         m_collections are all empty.
2382         (WebCore::Document::registerDynamicSubtreeNodeList): This function is now called for all DynamicSubtreeNodeLists supposed
2383         to just ones rooted at the document in order to increment the counter for each invalidation type.
2384         (WebCore::Document::unregisterDynamicSubtreeNodeList): Ditto.
2385         (WebCore::shouldInvalidateDynamicSubtreeNodeListForType): Checks the attribute name against NodeListInvalidationType.
2386         (WebCore::Document::shouldInvalidateDynamicSubtreeNodeList): Returns true if the given attribute name matches the invalidation
2387         type of the existing DynamicSubtreeNodeLists in the document. If the attribute name is not given (used when children change),
2388         then it checks for the existence of any DynamicSubtreeNodeLists. Conceptually, this function can be written as a list of
2389         "if" statements that checks values in m_nodeListCounts and the attribute name. We use "for" loop and switch statement instead
2390         to share logic and detect future addition of values to NodeListInvalidationType.
2391         * dom/Document.h:
2392         (Document): Moved RootType and InvalidationType from DynamicNodeListCacheBase and renamed them to NodeListRootType and
2393         NodeListInvalidationType respectively in order to reduce the possibility of future name collisions. Also the invalidation type
2394         now contains 6 values instead of 2.
2395         * dom/DynamicNodeList.cpp:
2396         (WebCore):
2397         * dom/DynamicNodeList.h:
2398         (WebCore::DynamicNodeListCacheBase::DynamicNodeListCacheBase):
2399         (WebCore::DynamicNodeListCacheBase::shouldInvalidateOnAttributeChange):
2400         (WebCore::DynamicNodeListCacheBase::rootType): Added.
2401         (WebCore::DynamicNodeListCacheBase::invalidationType): Added.
2402         (DynamicNodeListCacheBase): Uses 3 bits to store invalidation type now that the number of values have increased from 2 to 6.
2403         (WebCore::DynamicNodeList::DynamicNodeList):
2404         (WebCore::DynamicSubtreeNodeList::~DynamicSubtreeNodeList): Call unregisterDynamicSubtreeNodeList.
2405         (WebCore::DynamicSubtreeNodeList::DynamicSubtreeNodeList): Call registerDynamicSubtreeNodeList.
2406         * dom/MicroDataItemList.cpp:
2407         (WebCore::MicroDataItemList::MicroDataItemList): Invalidate on itemscope, itemprop, and itemtype content attribute changes.
2408         * dom/NameNodeList.cpp:
2409         (WebCore::NameNodeList::NameNodeList): Invalidate on name attribute changes.
2410         * dom/Node.cpp:
2411         (WebCore::Node::clearRareData):
2412         (WebCore::Node::invalidateNodeListsCacheAfterAttributeChanged): Replaced the hard coded check list of attributes, by a call
2413         to shouldInvalidateDynamicSubtreeNodeList.
2414         (WebCore::Node::invalidateNodeListsCacheAfterChildrenChanged): Calls shouldInvalidateDynamicSubtreeNodeList.
2415         (WebCore::Node::getElementsByTagName):
2416         (WebCore::Node::getElementsByTagNameNS):
2417         (WebCore::Node::getElementsByName):
2418         (WebCore::Node::getElementsByClassName):
2419         (WebCore::Node::radioNodeList):
2420         (WebCore::NodeRareData::createNodeLists):
2421         * dom/NodeRareData.h:
2422         (WebCore::NodeListsNodeData::adoptTreeScope): Unregister and register node lists in m_tagNodeListCacheNS since all node lists
2423         need to be accounted in m_nodeListCounts.
2424         (WebCore::NodeRareData::ensureNodeLists):
2425         (NodeRareData):
2426         * dom/TagNodeList.cpp:
2427         (WebCore::TagNodeList::TagNodeList): Do not invalidate on any attribute changes.
2428         * dom/TreeScope.cpp:
2429         (WebCore::TreeScope::TreeScope): No longer initializes m_numNodeListCaches since it has been removed.
2430         * dom/TreeScope.h:
2431         (TreeScope): Removed m_numNodeListCaches.
2432         * dom/TreeScopeAdopter.cpp:
2433         (WebCore::TreeScopeAdopter::moveTreeToNewScope):
2434         * html/HTMLCollection.h:
2435         (WebCore::HTMLCollectionCacheBase::HTMLCollectionCacheBase): Just pass in DoNotInvalidateOnAttributeChanges for now since
2436         it's never used in HTMLCollections.
2437         * html/LabelableElement.cpp:
2438         (WebCore::LabelableElement::labels):
2439         * html/LabelsNodeList.cpp:
2440         (WebCore::LabelsNodeList::LabelsNodeList): Invalidate on for content attribute changes.
2441         (WebCore::LabelsNodeList::~LabelsNodeList):
2442         * html/RadioNodeList.cpp:
2443         (WebCore::RadioNodeList::RadioNodeList): Invalidate on id, name, and for content attribute changes.
2444         (WebCore::RadioNodeList::~RadioNodeList):
2445
2446 2012-07-12  Gavin Barraclough  <barraclough@apple.com>
2447
2448         Threadsafety issues in WebScriptObject
2449         https://bugs.webkit.org/show_bug.cgi?id=90849
2450
2451         Reviewed by Filip Pizlo & Oliver Hunt.
2452
2453         Updated fix for this bug. Taking the JSC API lock from WebScriptObject::release
2454         may not be safe; better to just guard the JSWrapperCache with its own spinlock.
2455
2456         * bindings/objc/WebScriptObject.mm:
2457         (WebCore::getJSWrapper):
2458             - Added spinlock; also retain/autorelease the returned wrapper - it is unsafe
2459               to wait for the caller to do so, due to a race condition vs release removing
2460               the wrapper from the map.
2461         (WebCore::addJSWrapper):
2462             - Take the spinlock guarding the cache.
2463         (WebCore::removeJSWrapper):
2464             - Take the spinlock guarding the cache.
2465         (WebCore::removeJSWrapperIfRetainCountOne):
2466             - Take the spinlock guarding the cache, remove the wrapper if retainCount is one.
2467         (WebCore::createJSWrapper):
2468             - Remove the API lock; this method no longer needs to retain/autorelease (this is
2469               done by getJSWrapper).
2470         (-[WebScriptObject _setImp:originRootObject:rootObject:]):
2471             - Remove the API lock.
2472         (-[WebScriptObject release]):
2473             - Remove the API lock, retainCount check moved into removeJSWrapperIfRetainCountOne.
2474
2475 2012-07-11  David Hyatt  <hyatt@apple.com>
2476
2477         https://bugs.webkit.org/show_bug.cgi?id=91000
2478         REGRESSION (r122244): Overflow elements don't shrink as much as they should.
2479
2480         Reviewed by Simon Fraser.
2481
2482         This is a fix for a a regression from https://bugs.webkit.org/show_bug.cgi?id=90646.
2483
2484         I incorrectly analyzed the issue with Robert Hogan's negative margin patch and fooled myself into putting back
2485         in an incorrect minimum width check from long ago.
2486
2487         What should have happened in the test case I patched is that the overflow element should shrink to 0. The issue
2488         with improving the logical top estimate in the previous patch is it made the clear delta become 0. This in turn
2489         exposed a bug in our clearing algorithm with Robert's changes where you could need a relayout even if you didn't
2490         actually move. This issue only occurs because the floats list is getting changed mid-layout because of negative margins.
2491
2492         The patch changes getClearDelta to call setChildNeedsLayout(true) on children whose widths change even when their
2493         positions do not. In effect this dynamic addition of new floats after you have done a layout on the child already means
2494         that you can need to lay out again despite not actually having to move.
2495
2496         To handle this, the code that does the relayout is now called if the child needs a relayout. This is done even if
2497         the logical top estimate matches the final position.
2498
2499         No new tests required, since the test in fast/block/float is now correctly covering the issue.
2500
2501         * rendering/RenderBlock.cpp:
2502         (WebCore::RenderBlock::layoutBlockChild):
2503         (WebCore::RenderBlock::getClearDelta):
2504         * rendering/RenderBox.cpp:
2505         (WebCore::RenderBox::shrinkLogicalWidthToAvoidFloats):
2506
2507 2012-07-12  James Weatherall  <wez@chromium.org>
2508
2509         storage tests are flaky (crashing) on windows
2510         https://bugs.webkit.org/show_bug.cgi?id=90469
2511
2512         Reviewed by Kentaro Hara.
2513
2514         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.
2515
2516         This patch is intended to resolve flakiness in the storage tests including:
2517           storage/indexeddb/mozilla/indexes.html
2518           storage/indexeddb/mozilla/key-requirements-inline-and-passed.html
2519           storage/websql/multiple-databases-garbage-collection.html
2520
2521         * bindings/v8/NPV8Object.cpp:
2522         (WebCore::disposeUnderlyingV8Object):
2523         Zero the NPObject's underlying rootObject.
2524         (_NPN_EvaluateHelper):
2525         Add check that the underlying V8 object reference is valid.
2526
2527 2012-07-12  No'am Rosenthal  <noam.rosenthal@nokia.com>
2528
2529         Move TextureMapperAnimation and texmap/LayerTransform to platform/graphics
2530         https://bugs.webkit.org/show_bug.cgi?id=91111
2531
2532         Reviewed by Kenneth Rohde Christiansen.
2533
2534         TextureMapperAnimation and LayerTransform are not specific to TextureMapper, and we want
2535         to use them for other purposes as well. Moving them to platform/graphics would make that
2536         more explicit.
2537
2538         No new tests, moving files around.
2539
2540         * GNUmakefile.list.am:
2541         * Target.pri:
2542         * WebCore.gypi:
2543         * platform/graphics/GraphicsLayerAnimation.cpp: Renamed from Source/WebCore/platform/graphics/texmap/TextureMapperAnimation.cpp.
2544         * platform/graphics/GraphicsLayerAnimation.h: Renamed from Source/WebCore/platform/graphics/texmap/TextureMapperAnimation.h.
2545         * platform/graphics/GraphicsLayerTransform.cpp: Renamed from Source/WebCore/platform/graphics/texmap/LayerTransform.cpp.
2546         * platform/graphics/GraphicsLayerTransform.h: Renamed from Source/WebCore/platform/graphics/texmap/LayerTransform.h.
2547         * platform/graphics/texmap/GraphicsLayerTextureMapper.cpp:
2548         (WebCore::GraphicsLayerTextureMapper::addAnimation):
2549         * platform/graphics/texmap/GraphicsLayerTextureMapper.h:
2550         (GraphicsLayerTextureMapper):
2551         * platform/graphics/texmap/TextureMapperLayer.h:
2552         (TextureMapperLayer):
2553         (WebCore::TextureMapperLayer::setAnimatedTransform):
2554         (WebCore::TextureMapperLayer::setAnimatedOpacity):
2555
2556 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2557
2558         Web Inspector: Simplify UISourceCode code after moving revisions support inside it.
2559         https://bugs.webkit.org/show_bug.cgi?id=91118
2560
2561         Reviewed by Pavel Feldman.
2562
2563         * inspector/front-end/UISourceCode.js:
2564         (WebInspector.UISourceCode.prototype._setContent):
2565         (WebInspector.UISourceCode.prototype.revertToOriginal):
2566         (WebInspector.UISourceCode.prototype.revertAndClearHistory):
2567         (WebInspector.UISourceCode.prototype.contentChanged):
2568         (WebInspector.UISourceCode.prototype.commitWorkingCopy):
2569         (WebInspector.Revision.prototype.revertToThis):
2570
2571 2012-07-12  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
2572
2573         Unreviewed. Build fix for r122462.
2574
2575         * platform/qt/MemoryUsageSupportQt.cpp:
2576
2577 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2578
2579         Web Inspector: Remove resources panel edited resources search support.
2580         https://bugs.webkit.org/show_bug.cgi?id=91101
2581
2582         Reviewed by Pavel Feldman.
2583
2584         * inspector/front-end/ResourcesPanel.js:
2585         (WebInspector.ResourcesPanel.prototype.performSearch.callback):
2586         (WebInspector.ResourcesPanel.prototype.performSearch):
2587
2588 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2589
2590         Unreviewed r122460 inspector closure compilation fix follow up.
2591
2592         * inspector/front-end/RevisionHistoryView.js:
2593         (WebInspector.RevisionHistoryView):
2594
2595 2012-07-12  Ilya Tikhonovsky  <loislo@chromium.org>
2596
2597         Web Inspector: fix native memory instrumentation code for the bindings instrumentation.
2598         https://bugs.webkit.org/show_bug.cgi?id=91096
2599
2600         The instrumented class has to have instrumentation method which reports class size and type and
2601         the member objects and pointers.
2602
2603         Sample:
2604         void Node::reportMemoryUsage(MemoryObjectInfo* memoryObjectInfo) const
2605         {
2606             memoryObjectInfo->reportObjectInfo(this, MemoryInstrumentation::DOM); // report object size and type.
2607             TreeShared<Node, ContainerNode>::reportMemoryUsage(memoryObjectInfo); // call base class instrumentation.
2608             ScriptWrappable::reportMemoryUsage(memoryObjectInfo); // call base class instrumentation.
2609             memoryObjectInfo->reportPointer(m_document, MemoryInstrumentation::DOM); // report uninstrumented pointer.
2610             memoryObjectInfo->reportInstrumentedPointer(m_next); // report instrumented pointer.
2611             memoryObjectInfo->reportInstrumentedObject(m_anObject); // report instrumented object.
2612         }
2613
2614         Reviewed by Pavel Feldman.
2615
2616         Existing tests for native memory instrumentation.
2617
2618         * bindings/v8/DOMDataStore.cpp:
2619         (WebCore::DOMDataStore::reportMemoryUsage):
2620         * bindings/v8/DOMDataStore.h:
2621         (WebCore):
2622         (DOMDataStore):
2623         * bindings/v8/IntrusiveDOMWrapperMap.h:
2624         (WebCore::ChunkedTable::reportMemoryUsage):
2625         * bindings/v8/ScriptProfiler.cpp:
2626         (WebCore::ScriptProfiler::collectBindingMemoryInfo):
2627         * bindings/v8/V8Binding.cpp:
2628         (WebCore::V8BindingPerIsolateData::reportMemoryUsage):
2629         (WebCore::StringCache::reportMemoryUsage):
2630         * bindings/v8/V8Binding.h:
2631         (WebCore):
2632         (StringCache):
2633         (V8BindingPerIsolateData):
2634         * bindings/v8/V8DOMMap.h:
2635         (WebCore):
2636         (AbstractWeakReferenceMap):
2637         * inspector/InspectorMemoryAgent.cpp:
2638         (WebCore):
2639
2640 2012-07-12  Pavel Feldman  <pfeldman@chromium.org>
2641
2642         Web Inspector: beautify find bar looks, simplify search update routines.
2643         https://bugs.webkit.org/show_bug.cgi?id=91087
2644
2645         Reviewed by Vsevolod Vlasov.
2646
2647         This change updates the looks to the one on the screenshots and simplifies match count update routines.
2648
2649         * inspector/front-end/SearchController.js:
2650         (WebInspector.SearchController.onMatchesMouseDown):
2651         (WebInspector.SearchController):
2652         (WebInspector.SearchController.prototype.activePanelChanged.performPanelSearch):
2653         (WebInspector.SearchController.prototype.activePanelChanged):
2654         (WebInspector.SearchController.prototype._updateSearchNavigationButtonState):
2655         (WebInspector.SearchController.prototype._updateSearchMatchesCountAndCurrentMatchIndex):
2656         (WebInspector.SearchController.prototype._onKeyDown):
2657         (WebInspector.SearchController.prototype._onInput):
2658         (WebInspector.SearchController.prototype._onNextButtonSearch):
2659         (WebInspector.SearchController.prototype._onPrevButtonSearch):
2660         (WebInspector.SearchController.prototype._performSearch):
2661         * inspector/front-end/inspector.css:
2662         (#search):
2663         (#search:focus):
2664         (.toolbar-search-navigation-controls):
2665         (.toolbar-search-navigation):
2666         (.toolbar-search-navigation.enabled:hover):
2667         (.toolbar-search-navigation.enabled, .toolbar-search-navigation.enabled:active):
2668         (.toolbar-search):
2669         (.toolbar-search-control):
2670         (.search-results-matches):
2671         (.inspector-footer):
2672
2673 2012-07-12  Joshua Bell  <jsbell@chromium.org>
2674
2675         IndexedDB: ASSERT hit calling open from callback in Worker
2676         https://bugs.webkit.org/show_bug.cgi?id=90832
2677
2678         Reviewed by Kentaro Hara.
2679
2680         GroupSettings are used to provide the backing store path in some
2681         ports. Accessing those settings from a Worker was added, but the
2682         access referenced thread startup data that is cleared before the
2683         run loop, so an IDBFactory.open() call executed asynchronously
2684         would dereference a null pointer. Plumb the settings startup
2685         data into the context itself, like all of the other properties.
2686
2687         Test: storage/indexeddb/open-twice-workers.html
2688
2689         * Modules/indexeddb/IDBFactory.cpp:
2690         (WebCore::IDBFactory::open):
2691         * workers/DedicatedWorkerContext.cpp:
2692         (WebCore::DedicatedWorkerContext::DedicatedWorkerContext):
2693         * workers/DedicatedWorkerContext.h:
2694         (WebCore::DedicatedWorkerContext::create):
2695         (DedicatedWorkerContext):
2696         * workers/DedicatedWorkerThread.cpp:
2697         (WebCore::DedicatedWorkerThread::createWorkerContext):
2698         * workers/DedicatedWorkerThread.h:
2699         (DedicatedWorkerThread):
2700         * workers/SharedWorkerContext.cpp:
2701         (WebCore::SharedWorkerContext::SharedWorkerContext):
2702         * workers/SharedWorkerContext.h:
2703         (WebCore::SharedWorkerContext::create):
2704         (SharedWorkerContext):
2705         * workers/SharedWorkerThread.cpp:
2706         (WebCore::SharedWorkerThread::createWorkerContext):
2707         * workers/SharedWorkerThread.h:
2708         (SharedWorkerThread):
2709         * workers/WorkerContext.cpp:
2710         (WebCore::WorkerContext::WorkerContext):
2711         * workers/WorkerContext.h:
2712         (WebCore::WorkerContext::groupSettings):
2713         (WorkerContext):
2714         * workers/WorkerThread.cpp:
2715         (WebCore::WorkerThread::workerThread):
2716         * workers/WorkerThread.h:
2717         (WorkerThread):
2718
2719 2012-07-12  Allan Sandfeld Jensen  <allan.jensen@nokia.com>
2720
2721         [Qt] Implement MemoryUsageSupport
2722         https://bugs.webkit.org/show_bug.cgi?id=91094
2723
2724         Reviewed by Adam Barth.
2725
2726         Implements Qt versions of the memory-usage functions using the information we have available
2727         from the various memory systems used in WebKit.
2728
2729         Also gets rid of a redundant indirection in V8GCController.
2730
2731         * Target.pri:
2732         * bindings/v8/V8GCController.cpp:
2733         (WebCore::V8GCController::gcEpilogue):
2734         (WebCore::V8GCController::checkMemoryUsage):
2735         * platform/qt/MemoryUsageSupportQt.cpp: Added.
2736         (WebCore::mallocMemoryUsage):
2737         (WebCore::memoryUsageKB):
2738         (WebCore::actualMemoryUsageKB):
2739         (WebCore::MemoryUsageSupport::memoryUsageMB):
2740         (WebCore::MemoryUsageSupport::actualMemoryUsageMB):
2741         (WebCore::MemoryUsageSupport::lowMemoryUsageMB):
2742         (WebCore::MemoryUsageSupport::highMemoryUsageMB):
2743         (WebCore::MemoryUsageSupport::highUsageDeltaMB):
2744         (WebCore::MemoryUsageSupport::processMemorySizesInBytes):
2745         * platform/qt/PlatformSupport.h:
2746         (PlatformSupport):
2747
2748 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2749
2750         Web Inspector: Revision history view should be updated when uiSourceCodes are removed or replaced.
2751         https://bugs.webkit.org/show_bug.cgi?id=91095
2752
2753         Reviewed by Pavel Feldman.
2754
2755         RevisionHistoryView is now reset in WorkspaceReset event handler.
2756         UISourceCodes are now removed from RevisionHistoryView when uiSourceCode is removed or replace.
2757
2758         * inspector/front-end/RevisionHistoryView.js:
2759         (WebInspector.RevisionHistoryView):
2760         (WebInspector.RevisionHistoryView.prototype._clearHistory):
2761         (WebInspector.RevisionHistoryView.prototype._uiSourceCodeRemoved):
2762         (WebInspector.RevisionHistoryView.prototype._uiSourceCodeReplaced):
2763         * inspector/front-end/ScriptsPanel.js:
2764         (WebInspector.ScriptsPanel.prototype._reset):
2765
2766 2012-07-12  Dongwoo Im  <dw.im@samsung.com>
2767
2768         InspectorFileSystemAgent.cpp have to include File.h explicitly
2769         https://bugs.webkit.org/show_bug.cgi?id=91078
2770
2771         Reviewed by Vsevolod Vlasov.
2772
2773         Make sure the InspectorFileSystemAgent.cpp include the File.h by including explicitly.
2774
2775         No new tests. Covered by existing tests.
2776
2777         * inspector/InspectorFileSystemAgent.cpp: Include File.h
2778
2779 2012-07-12  Vsevolod Vlasov  <vsevik@chromium.org>
2780
2781         Web Inspector: [Regression] Stale revisions should be removed on navigation.
2782         https://bugs.webkit.org/show_bug.cgi?id=91080
2783
2784         Reviewed by Pavel Feldman.
2785
2786         Moved all the code responsible for script mapping resetting to Workspace.
2787         Stale revisions are removed in Workspace._reset now.
2788
2789         * inspector/front-end/ScriptSnippetModel.js:
2790         (WebInspector.ScriptSnippetModel):
2791         * inspector/front-end/StylesPanel.js:
2792         (WebInspector.StylesUISourceCodeProvider):
2793         (WebInspector.StylesUISourceCodeProvider.prototype._populate):
2794         (WebInspector.StylesUISourceCodeProvider.prototype.reset):
2795         * inspector/front-end/UISourceCode.js:
2796         (WebInspector.UISourceCode.prototype._restoreRevisionHistory):
2797         * inspector/front-end/Workspace.js:
2798         (WebInspector.Workspace):
2799         (WebInspector.Workspace.prototype._reset):
2800
2801 2012-07-12  Andreas Kling  <kling@webkit.org>
2802
2803         Make ElementAttributeData a variable-sized object to reduce memory use.
2804         <http://webkit.org/b/88240>
2805
2806         Reviewed by Antti "Obi-Wan" Koivisto.
2807
2808         Take advantage of the fact that we know at ElementAttributeData construction time how many attributes
2809         it needs to accomodate and allocate exactly as much space as needed instead of using a Vector.
2810         For elements that never have their attribute list mutated (the vast majority), this saves a lot of
2811         memory and removes the indirection to Vector<Attribute>'s heap-allocated storage.
2812
2813         Introduced a mutability flag to ElementAttributeData and sprinkled assertions all over to make sure
2814         that nothing tries to mutate an element with a raw attribute array.
2815
2816         When an Element's attribute(s) are mutated, we reconstruct the ElementAttributeData, this time using
2817         a Vector as backing instead. This is triggered when calling Element::mutableAttributeData().
2818
2819         This reduces memory consumption by 3.2MB when viewing the full HTML5 spec at <http://whatwg.org/c/>.
2820         That is a ~35% reduction in DOM attribute memory use.
2821
2822         Furthermore, that page ends up promoting 27% of the elements to mutable attribute storage due to dynamic
2823         adding of "class" attributes. For more static pages, savings are even greater.
2824
2825         Also did away with ElementAttributeData::removeAttribute(name) and do separate index lookup where
2826         needed. Not a big deal but avoids double lookup of attributes when removing them.
2827
2828         * css/PropertySetCSSStyleDeclaration.cpp:
2829         (WebCore::InlineCSSStyleDeclaration::ensureMutablePropertySet):
2830
2831             Call ensureInlineStyle() on the element so we know we have a StylePropertySet to work with.
2832             May cause the creation (and replacement) of a new underlying StylePropertySet on the Element's
2833             attribute data.
2834
2835         * css/StylePropertySet.cpp:
2836         (WebCore::StylePropertySet::adoptCSSOMWrapperFrom):
2837
2838             Added. Allows switching the underlying StylePropertySet object while retaining CSSOM wrapper
2839             identity (web facing behavior.)
2840
2841         * dom/Element.cpp:
2842         (WebCore::Element::detachAttribute):
2843         (WebCore::Element::removeAttribute):
2844         (WebCore::Element::attributes):
2845         (WebCore::Element::setAttributeInternal):
2846         (WebCore::Element::parserSetAttributes):
2847         (WebCore::Element::hasEquivalentAttributes):
2848         (WebCore::Element::setAttributeNode):
2849         (WebCore::Element::removeAttributeNode):
2850         (WebCore::Element::getAttributeNode):
2851         (WebCore::Element::getAttributeNodeNS):
2852         (WebCore::Element::hasAttribute):
2853         (WebCore::Element::hasAttributeNS):
2854         (WebCore::Element::normalizeAttributes):
2855         (WebCore::Element::cloneAttributesFromElement):
2856         (WebCore::Element::createMutableAttributeData):
2857         * dom/Element.h:
2858         (WebCore::Element::attributeData):
2859         (Element):
2860         (WebCore::Element::updatedAttributeData):
2861         (WebCore::Element::ensureAttributeData):
2862         (WebCore::Element::ensureUpdatedAttributeData):
2863         (WebCore::Element::mutableAttributeData):
2864         (WebCore):
2865         * dom/ElementAttributeData.cpp:
2866         (WebCore::ElementAttributeData::createImmutable):
2867         (WebCore):
2868         (WebCore::ElementAttributeData::ElementAttributeData):
2869         (WebCore::ElementAttributeData::~ElementAttributeData):
2870         (WebCore::ElementAttributeData::attrIfExists):
2871         (WebCore::ElementAttributeData::ensureAttr):
2872         (WebCore::ElementAttributeData::setAttr):
2873         (WebCore::ElementAttributeData::removeAttr):
2874         (WebCore::ElementAttributeData::ensureInlineStyle):
2875         (WebCore::ElementAttributeData::ensureMutableInlineStyle):
2876         (WebCore::ElementAttributeData::destroyInlineStyle):
2877         (WebCore::ElementAttributeData::addAttribute):
2878         (WebCore::ElementAttributeData::removeAttribute):
2879         (WebCore::ElementAttributeData::isEquivalent):
2880         (WebCore::ElementAttributeData::detachAttrObjectsFromElement):
2881         (WebCore::ElementAttributeData::getAttributeItemIndexSlowCase):
2882         (WebCore::ElementAttributeData::cloneDataFrom):
2883         (WebCore::ElementAttributeData::clearAttributes):
2884         (WebCore::ElementAttributeData::replaceAttribute):
2885         (WebCore::ElementAttributeData::getAttributeNode):
2886         * dom/ElementAttributeData.h:
2887         (WebCore::ElementAttributeData::create):
2888         (ElementAttributeData):
2889         (WebCore::ElementAttributeData::setClass):
2890         (WebCore::ElementAttributeData::setIdForStyleResolution):
2891         (WebCore::ElementAttributeData::inlineStyle):
2892         (WebCore::ElementAttributeData::setAttributeStyle):
2893         (WebCore::ElementAttributeData::reportMemoryUsage):
2894         (WebCore::ElementAttributeData::isMutable):
2895         (WebCore::ElementAttributeData::makeMutable):
2896         (WebCore::ElementAttributeData::length):
2897         (WebCore):
2898         (WebCore::ElementAttributeData::isEmpty):
2899         (WebCore::ElementAttributeData::array):
2900         (WebCore::ElementAttributeData::removeAttribute):
2901         (WebCore::ElementAttributeData::getAttributeItem):
2902         (WebCore::ElementAttributeData::getAttributeItemIndex):
2903         (WebCore::ElementAttributeData::attributeItem):
2904         * dom/StyledElement.cpp:
2905         (WebCore::StyledElement::style):
2906         (WebCore::StyledElement::classAttributeChanged):
2907         (WebCore::StyledElement::setInlineStyleProperty):
2908         (WebCore::StyledElement::removeInlineStyleProperty):
2909         (WebCore::StyledElement::addSubresourceAttributeURLs):
2910         * dom/StyledElement.h:
2911         (WebCore::StyledElement::ensureInlineStyle):
2912         * html/parser/HTMLConstructionSite.cpp:
2913         (WebCore::HTMLConstructionSite::createHTMLElementFromSavedElement):
2914         * xml/parser/XMLDocumentParserQt.cpp:
2915         (WebCore::XMLDocumentParser::XMLDocumentParser):
2916
2917 2012-07-12  Florin Malita  <fmalita@chromium.org>
2918
2919         Incorrect handling of chained pending resources in SVGUseElement
2920         https://bugs.webkit.org/show_bug.cgi?id=89686
2921
2922         Reviewed by Nikolas Zimmermann.
2923
2924         Currently SVGUseElement builds the shadow tree when the target first
2925         becomes available. This is normally OK, but if the target itself (or
2926         one of its children) is a <use> element with pending resources, then
2927         the shadow expansion only captures the current state of the tree and
2928         never gets updated when the pending resource becomes available.
2929
2930         In order to support arbitrary <use>-on-<use> constructs, this patch
2931         tracks nested <use> dependencies and rebuilds the dependent trees
2932         whenever the target gets updated.
2933
2934
2935         Tests: svg/custom/use-nested-expected.svg
2936                svg/custom/use-nested.svg
2937
2938         * svg/SVGElement.cpp:
2939         (WebCore::SVGElement::removedFrom): removedFrom needs to be called up the inheritance chain
2940         before invoking removeAllElementReferencesForTarget. Otherwise we could end up finding the
2941         element being removed as a valid target in SVGUseElement::buildInstanceTree because its
2942         InDocument flag is not cleared yet.
2943         * svg/SVGUseElement.cpp:
2944         (WebCore::SVGUseElement::~SVGUseElement):
2945         (WebCore::SVGUseElement::clearResourceReferences):
2946         (WebCore::SVGUseElement::buildPendingResource):
2947         (WebCore::SVGUseElement::buildShadowAndInstanceTree):
2948         (WebCore::SVGUseElement::buildInstanceTree):
2949         * svg/SVGUseElement.h:
2950         (SVGUseElement):
2951         Track <use> -> <use> dependencies using SVGDocumentExtensions'
2952         m_elementDependencies framework and ensure dependent trees are rebuilt
2953         when the target itself gets rebuilt.
2954
2955 2012-07-12  MORITA Hajime  <morrita@google.com>
2956
2957         Typo: ParentTranversalDetails should be ParentTraversalDetails
2958         https://bugs.webkit.org/show_bug.cgi?id=91059
2959
2960         Reviewed by Andreas Kling.
2961
2962         Did a bare rename from ParentTranversalDetails to ParentTraversalDetails
2963
2964         No new tests, just a rename.
2965
2966         * dom/ComposedShadowTreeWalker.cpp:
2967         (WebCore::ComposedShadowTreeWalker::ParentTraversalDetails::didTraverseInsertionPoint):
2968         (WebCore::ComposedShadowTreeWalker::ParentTraversalDetails::didTraverseShadowRoot):
2969         (WebCore::ComposedShadowTreeWalker::ParentTraversalDetails::didFindNode):
2970         (WebCore::ComposedShadowTreeWalker::findParent):
2971         (WebCore::ComposedShadowTreeWalker::traverseNodeEscapingFallbackContents):
2972         (WebCore::ComposedShadowTreeWalker::traverseParent):
2973         (WebCore::ComposedShadowTreeWalker::traverseParentInCurrentTree):
2974         (WebCore::ComposedShadowTreeWalker::traverseParentBackToYoungerShadowRootOrHost):
2975         * dom/ComposedShadowTreeWalker.h:
2976         (WebCore::ComposedShadowTreeWalker::ParentTraversalDetails::ParentTraversalDetails):
2977         (ComposedShadowTreeWalker):
2978         * dom/NodeRenderingContext.h:
2979         (NodeRenderingContext):
2980
2981 2012-07-11  Antti Koivisto  <antti@apple.com>
2982
2983         Failure to dispatch didFinishLoadForFrame if font load fails synchronously
2984         https://bugs.webkit.org/show_bug.cgi?id=91018
2985
2986         Reviewed by Enrica Casucci.
2987
2988         New font loads may be triggered by layout after the document load is complete but before we have dispatched
2989         didFinishLoading for the frame. If the load fails synchronously we might fail to ever invoke
2990         FrameLoader::checkLoadComplete and so fail to dispatch didFinishLoading.
2991
2992         Make sure this doesn't happen by calling FrameLoader::checkLoadComplete explicitly after triggering font loads.
2993
2994         * css/CSSFontSelector.cpp:
2995         (WebCore::CSSFontSelector::beginLoadTimerFired):
2996
2997 2012-07-12  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2998
2999         Register protocol handler files should be in Modules/protocolhandler
3000         https://bugs.webkit.org/show_bug.cgi?id=90766
3001
3002         Reviewed by Hajime Morita.
3003
3004         The register protocol handler files are now self-contained. This patch is moved to the Modules.
3005
3006         No new tests. Covered by existing tests.
3007
3008         * CMakeLists.txt:
3009         * GNUmakefile.am:
3010         * GNUmakefile.list.am:
3011         * Modules/protocolhandler/NavigatorRegisterProtocolHandler.cpp: Renamed from Source/WebCore/page/NavigatorRegisterProtocolHandler.cpp.
3012         (WebCore):
3013         (WebCore::initProtocolHandlerWhitelist):
3014         (WebCore::verifyCustomHandlerURL):
3015         (WebCore::isProtocolWhitelisted):
3016         (WebCore::verifyProtocolHandlerScheme):
3017         (WebCore::NavigatorRegisterProtocolHandler::NavigatorRegisterProtocolHandler):
3018         (WebCore::NavigatorRegisterProtocolHandler::~NavigatorRegisterProtocolHandler):
3019         (WebCore::NavigatorRegisterProtocolHandler::registerProtocolHandler):
3020         (WebCore::customHandlersStateString):
3021         (WebCore::NavigatorRegisterProtocolHandler::isProtocolHandlerRegistered):
3022         (WebCore::NavigatorRegisterProtocolHandler::unregisterProtocolHandler):
3023         * Modules/protocolhandler/NavigatorRegisterProtocolHandler.h: Renamed from Source/WebCore/page/NavigatorRegisterProtocolHandler.h.
3024         (WebCore):
3025         (NavigatorRegisterProtocolHandler):
3026         * Modules/protocolhandler/NavigatorRegisterProtocolHandler.idl: Renamed from Source/WebCore/page/NavigatorRegisterProtocolHandler.idl.
3027         * Target.pri:
3028         * WebCore.gyp/WebCore.gyp:
3029         * WebCore.gypi:
3030         * WebCore.pri:
3031         * WebCore.vcproj/WebCore.vcproj:
3032         * WebCore.vcproj/WebCoreCommon.vsprops:
3033         * WebCore.vcproj/copyForwardingHeaders.cmd:
3034         * WebCore.xcodeproj/project.pbxproj:
3035
3036 2012-07-12  Jonathan Dong  <jonathan.dong@torchmobile.com.cn>
3037
3038         [BlackBerry] Cannot use digest proxy auth and NTLM auth at the same time
3039         https://bugs.webkit.org/show_bug.cgi?id=91054
3040
3041         Reviewed by George Staikos.
3042
3043         Added an interface function syncProxyCredential() in class
3044         PageClientBlackBerry, which is responsible to notify WebPageClient
3045         to synchronize proxy credential to the chrome process.
3046
3047         Internally reviewed by Jason Liu <jason.liu@torchmobile.com.cn>
3048
3049         No new tests since there's no functional change.
3050
3051         * platform/blackberry/PageClientBlackBerry.h:
3052         * platform/network/blackberry/NetworkJob.cpp:
3053         (WebCore::NetworkJob::storeCredentials): Remember the accepted proxy
3054         credential and notify webpage client to synchronize it.
3055
3056 2012-07-12  Xingnan Wang  <xingnan.wang@intel.com>
3057
3058         [IndexedDB] upperOpen set to true in lowerBound()/lowerOpen set to true in upperBound()
3059         https://bugs.webkit.org/show_bug.cgi?id=90867
3060
3061         Reviewed by Tony Chang.
3062
3063         In the latest W3C spec upperOpen/lowerOpen are set to true in lowerBound()/upperBound(), which
3064         we should keep aligned with.
3065
3066         No new tests - updated storage/indexeddb/keyrange.html to match new behavior.
3067
3068         * Modules/indexeddb/IDBKeyRange.cpp:
3069         (WebCore::IDBKeyRange::lowerBound):
3070         (WebCore::IDBKeyRange::upperBound):
3071
3072 2012-07-12  MORITA Hajime  <morrita@google.com>
3073
3074         [Shadow DOM] <video> with <shadow> crashes
3075         https://bugs.webkit.org/show_bug.cgi?id=91055
3076
3077         Reviewed by Kent Tamura.
3078
3079         This is similar to Bug 90480, where an undesired renderer is created by
3080         locating an insertion point on the shadow boundary.
3081
3082         This change adds a guard for such case by cheking whether the
3083         source node of each to-be-created renderer comes from the UA shadow
3084         tree, which is allowed to have a renderer.
3085
3086         Test: fast/dom/shadow/insertion-point-video-crash.html
3087
3088         * html/HTMLMediaElement.cpp:
3089         (WebCore::HTMLMediaElement::childShouldCreateRenderer): Added a check.
3090         (WebCore::HTMLMediaElement::mediaControls): Added const.
3091         (WebCore::HTMLMediaElement::hasMediaControls): Added const.
3092         * html/HTMLMediaElement.h:
3093         (HTMLMediaElement):
3094
3095 2012-07-12  Yoshifumi Inoue  <yosin@chromium.org>
3096
3097         REGRESSION(r122184): LocaleMac::currentLocale should use current locale rather than newly create locale object.
3098         https://bugs.webkit.org/show_bug.cgi?id=91057
3099
3100         Reviewed by Kent Tamura.
3101
3102         This patch changes NSLocale object of LocaleMac::m_locale variable to
3103         current NSLocale object rather than newly created NSLocale object from
3104         locale identifier.
3105
3106         No new tests. We don't have way to change system preferences from
3107         test scripts and restoring them. To test this patch, we need to do so.
3108
3109         * platform/text/mac/LocaleMac.h:
3110         (LocaleMac): Added a constructor which takes NSLocale object.
3111         * platform/text/mac/LocaleMac.mm:
3112         (WebCore::LocaleMac::LocaleMac): Added a constructor which takes NSLocale object.
3113         (WebCore::LocaleMac::currentLocale): Changed to construct LocaleMac object from NSLocale object rather than locale identifier.
3114
3115 2012-07-11  Carlos Garcia Campos  <cgarcia@igalia.com>
3116
3117         [GTK] Add webkit_cookie_manager_set_persistent_storage() to WebKit2 GTK+ API
3118         https://bugs.webkit.org/show_bug.cgi?id=83016
3119
3120         Reviewed by Martin Robinson.
3121
3122         * platform/network/soup/CookieStorageSoup.cpp:
3123         (WebCore::soupCookiesChanged): Do not notify about changes in
3124         other cookie jars than the current one.
3125
3126 2012-07-12  Philip Rogers  <pdr@google.com>
3127
3128         Refactor RenderSVGShape bounding box code
3129         https://bugs.webkit.org/show_bug.cgi?id=90655
3130
3131         Reviewed by Nikolas Zimmermann.
3132
3133         RenderSVGShape::objectBoundingBox worked differently than RenderSVGShape::strokeBoundingBox by
3134         not caching the object bounding box and instead computing it on each call. For consistency and
3135         performance objectBoundingBox has been refactored to return a cached value.
3136
3137         createShape has been renamed updateShapeFromElement for understandability. updateShapeFromElement
3138         now updates the internal state of the shape (bounding boxes, etc) from the associated element.
3139         RenderSVGShape::inflateWithStrokeAndMarkerBounds has been merged into
3140         RenderSVGShape::calculateStrokeBoundingBox which is called from updateShapeFromElement.
3141
3142         After this change all bounding box computation is now handled in updateShapeFromElement. Because
3143         subclasses override updateShapeFromElement it will be easy for them to have custom bounding box
3144         code there (as will happen for RenderSVGPath in a followup patch).
3145
3146         strokeBoundingBox and objectBoundingBox are now able to return their cached values immediately
3147         in RenderSVGRect and RenderSVGEllipse instead of checking their fallback state on each call.
3148
3149         Additionally, to save space RenderSVGEllipse and RenderSVGRect now use the m_fillBoundingBox and
3150         m_strokeBoundingBox of RenderSVGShape instead of having their own.
3151
3152         This patch also removes setStrokeAndMarkerBoundingBox that was previously dead code.
3153
3154         No new tests, just a refactoring.
3155
3156         * rendering/svg/RenderSVGEllipse.cpp:
3157         (WebCore::RenderSVGEllipse::updateShapeFromElement):
3158         (WebCore):
3159         (WebCore::RenderSVGEllipse::fillShape):
3160         (WebCore::RenderSVGEllipse::strokeShape):
3161         (WebCore::RenderSVGEllipse::shapeDependentStrokeContains):
3162         * rendering/svg/RenderSVGEllipse.h:
3163         (RenderSVGEllipse):
3164         (WebCore::RenderSVGEllipse::isEmpty):
3165         * rendering/svg/RenderSVGRect.cpp:
3166         (WebCore::RenderSVGRect::updateShapeFromElement):
3167         (WebCore):
3168         (WebCore::RenderSVGRect::fillShape):
3169         (WebCore::RenderSVGRect::strokeShape):
3170         (WebCore::RenderSVGRect::shapeDependentStrokeContains):
3171         (WebCore::RenderSVGRect::shapeDependentFillContains):
3172         * rendering/svg/RenderSVGRect.h:
3173         (RenderSVGRect):
3174         (WebCore::RenderSVGRect::isEmpty):
3175         * rendering/svg/RenderSVGShape.cpp:
3176         (WebCore::RenderSVGShape::updateShapeFromElement):
3177         (WebCore):
3178         (WebCore::RenderSVGShape::layout):
3179         (WebCore::RenderSVGShape::calculateObjectBoundingBox):
3180         (WebCore::RenderSVGShape::calculateStrokeBoundingBox):
3181         (WebCore::RenderSVGShape::updateRepaintBoundingBox):
3182         * rendering/svg/RenderSVGShape.h:
3183         (RenderSVGShape):
3184         (WebCore::RenderSVGShape::objectBoundingBox):
3185         (WebCore::RenderSVGShape::strokeBoundingBox):
3186
3187 2012-07-12  Kent Tamura  <tkent@chromium.org>
3188
3189         Do not save the form state signature if nothing is saved
3190         https://bugs.webkit.org/show_bug.cgi?id=91050
3191
3192         Reviewed by Hajime Morita.
3193
3194         This change will reduce the size of HistoyItem.
3195
3196         Test: fast/forms/state-restore-empty-state.html
3197
3198         * html/FormController.cpp:
3199         (WebCore::FormController::formElementsState):
3200         If stateVector has only the signature string, clear it.
3201
3202 2012-07-11  Mark Rowe  <mrowe@apple.com>
3203
3204         <http://webkit.org/b/91051> Fix the Qt Mac build after r122400.
3205
3206         Qt on Mac builds with a deployment target of OS X 10.5. However, it was not been setting
3207         BUILDING_ON_LEOPARD / TARGETING_LEOPARD and thus was falling down code paths in DisplaySleepDisabler.cpp
3208         that were specific to Snow Leopard and newer. After the removal of BUILDING_ON_LEOPARD
3209         and TARGETING_LEOPARD it ended up falling down the correct Leopard-compatible code path,
3210         which revealed that the code made assumptions about which headers had already been included.
3211
3212         * platform/mac/DisplaySleepDisabler.cpp: Include CoreServices/CoreServices.h to pull in
3213         a declaration of UpdateSystemActivity when targeting Leopard.
3214
3215 2012-07-11  Matt Falkenhagen  <falken@chromium.org>
3216
3217         Add dialog element feature toggle to InternalSettings
3218         https://bugs.webkit.org/show_bug.cgi?id=90934
3219
3220         Reviewed by Hajime Morita.
3221
3222         This enables layout tests to be written for dialog although the feature is disabled by default.
3223
3224         Tests: fast/dom/HTMLDialogElement/dialog-disabled.html
3225                fast/dom/HTMLDialogElement/dialog-enabled.html
3226
3227         * WebCore.exp.in: Added newly exported symbol.
3228         * testing/InternalSettings.cpp:
3229         (WebCore::InternalSettings::Backup::Backup): Backup dialog feature flag.
3230         (WebCore::InternalSettings::Backup::restoreTo): Restore dialog feature flag.
3231         (WebCore::InternalSettings::setDialogElementEnabled): Added.
3232         (WebCore):
3233         * testing/InternalSettings.h: Added support for dialog.
3234         (Backup):
3235         (InternalSettings):
3236         * testing/InternalSettings.idl: Added support for dialog.
3237
3238 2012-07-11  Kent Tamura  <tkent@chromium.org>
3239
3240         Accessing width or height of a detached image input element causes crash
3241         https://bugs.webkit.org/show_bug.cgi?id=90885
3242
3243         Reviewed by Kentaro Hara.
3244
3245         Test: fast/forms/image/width-and-height-of-detached-input.html
3246
3247         * html/ImageInputType.cpp:
3248         (WebCore::ImageInputType::height): Add null check for m_imageLoader.
3249         (WebCore::ImageInputType::width): ditto.
3250
3251 2012-07-11  Kent Tamura  <tkent@chromium.org>
3252
3253         Do not save state of stateless form controls
3254         https://bugs.webkit.org/show_bug.cgi?id=90964
3255
3256         Reviewed by Hajime Morita.
3257
3258         By Bug 89409, we started to store unmodified form control
3259         state. However we don't need to make such state for the following
3260         types.
3261         - password
3262         - submit
3263         - reset
3264         - button
3265         - keygen
3266
3267         Test: fast/forms/state-restore-skip-stateless.html
3268
3269         * html/BaseButtonInputType.cpp:
3270         (WebCore::BaseButtonInputType::shouldSaveAndRestoreFormControlState):
3271         Added. Disable saving state for submit, reset, and button types.
3272         * html/BaseButtonInputType.h:
3273         (BaseButtonInputType): Add shouldSaveAndRestoreFormControlState.
3274         * html/HTMLFormControlElementWithState.h:
3275         (HTMLFormControlElementWithState): Make shouldSaveAndRestoreFormControlState virtual.
3276         * html/HTMLInputElement.cpp:
3277         (WebCore::HTMLInputElement::shouldSaveAndRestoreFormControlState):
3278         Added. Checks InputType::shouldSaveAndRestoreFormControlState first.
3279         * html/HTMLInputElement.h:
3280         (HTMLInputElement): Override shouldSaveAndRestoreFormControlState.
3281         * html/HTMLKeygenElement.cpp:
3282         (WebCore::HTMLKeygenElement::shouldSaveAndRestoreFormControlState)
3283         Added. Disable saving state for <keygen>.:
3284         * html/HTMLKeygenElement.h: Override shouldSaveAndRestoreFormControlState.
3285         * html/InputType.cpp:
3286         (WebCore::InputType::shouldSaveAndRestoreFormControlState):
3287         Added. Enable saving state for all types by default.
3288         * html/InputType.h:
3289         (InputType): Add shouldSaveAndRestoreFormControlState.
3290         * html/PasswordInputType.cpp:
3291         (WebCore::PasswordInputType::shouldSaveAndRestoreFormControlState):
3292         Added. Disabled saving state.
3293         (WebCore::PasswordInputType::saveFormControlState):
3294         Because shouldSaveAndRestoreFormControlState() returns false,
3295         saveFormControlState should be never called.
3296         * html/PasswordInputType.h:
3297         (PasswordInputType): Override shouldSaveAndRestoreFormControlState.
3298
3299 2012-07-11  No'am Rosenthal  <noam.rosenthal@nokia.com>
3300
3301         Compilation error with GLES2 when using gl2ext.h from ANGLE.
3302         https://bugs.webkit.org/show_bug.cgi?id=91030
3303
3304         Reviewed by Kenneth Russell.
3305
3306         gl2ext.h defines PFNGLRENDERBUFFERSTORAGEMULTISAMPLEIMG and not
3307         PFNGLRENDERBUFFERSTORAGEMULTISAMPLEIMGPROC. This is a bug in the original Khronos header,
3308         but we should work around it until fixed headers are integrated.
3309
3310
3311         * platform/graphics/opengl/Extensions3DOpenGLES.h:
3312         (Extensions3DOpenGLES):
3313             Removed the "PROC" suffix to work around the header bug.
3314
3315 2012-07-11  Mark Rowe  <mrowe@apple.com>
3316
3317         Attempt to fix the Chromium Mac build after r122400.
3318
3319         * platform/text/cf/HyphenationCF.cpp: Ensure that Chromium only compiles the body of this file if
3320         building for a new enough version of OS X.
3321
3322 2012-07-11  Kent Tamura  <tkent@chromium.org>
3323
3324         [Mac] Sort project.pbxproj
3325
3326         * WebCore.xcodeproj/project.pbxproj: Apply Tools/Scripts/sort-Xcode-project-file
3327
3328 2012-07-11  Dan Bernstein  <mitz@apple.com>
3329
3330         When a table row height grows because of pagination, not all cells’ heights are adjusted
3331         https://bugs.webkit.org/show_bug.cgi?id=91043
3332
3333         Reviewed by Sam Weinig.
3334
3335         The fix for in bug <http://webkit.org/b/83595> in <http://trac.webkit.org/r113738> made table
3336         rows grow as necessary to fit cells that grow as a result of pagination. But it had two bad
3337         side effects: earlier cells on the row would not grow by the same amount, and later cells on
3338         the row would factor the existing growth into their intrinsic padding.
3339
3340         Test: fast/multicol/table-row-height-increase.html
3341
3342         * rendering/RenderTableSection.cpp:
3343         (WebCore::RenderTableSection::layoutRows): Now when a cell becomes taller than the row height,
3344         the additional height needed is recorded, and the cell is shrunk back to row height. Then
3345         after finishing the row, all cells occurring on the row (including cells spanning it but not
3346         starting on it) are grown by the same amount.
3347
3348 2012-07-11  Mark Rowe  <mrowe@apple.com>
3349
3350         <http://webkit.org/b/91024> Build against the latest SDK when targeting older OS X versions.
3351
3352         Reviewed by Dan Bernstein.
3353
3354         The deployment target is already set to the version that we're targeting, and it's that setting
3355         which determines which functionality from the SDK is available to us.
3356
3357         * Configurations/Base.xcconfig:
3358
3359 2012-07-11  Mark Rowe  <mrowe@apple.com>
3360
3361         Switch a few cases that care about the SDK over to checking __MAC_OS_X_VERSION_MAX_ALLOWED so that things build.
3362
3363         Part of <http://webkit.org/b/91015> Remove BUILDING_ON / TARGETING macros in favor of system availability macros
3364
3365         Reviewed by Filip Pizlo.
3366
3367         * platform/graphics/cg/GraphicsContextCG.cpp: The SDK we're building against determines whether the constant is available.
3368         * platform/mac/EmptyProtocolDefinitions.h:  The SDK we're building against determines whether the protocols are declared.
3369         * platform/mac/NSScrollerImpDetails.h: The SDK we're building against determines whether the constants and method are
3370         available.
3371         * platform/mac/SharedTimerMac.mm: The SDK we're building against determines whether the function has already been declared.
3372
3373 2012-07-11  Mark Rowe  <mrowe@apple.com>
3374
3375         <http://webkit.org/b/91015> Remove BUILDING_ON / TARGETING macros in favor of system availability macros
3376
3377         This removal was handled by a script that translates the relevant macros in to the equivalent checks
3378         using the system availability macros.
3379
3380         Reviewed by Filip Pizlo.
3381
3382         * WebCore.exp.in:
3383         * accessibility/AccessibilityList.h:
3384         * accessibility/AccessibilityTable.h:
3385         * accessibility/mac/AXObjectCacheMac.mm:
3386         * editing/mac/EditorMac.mm:
3387         * loader/MainResourceLoader.cpp:
3388         * loader/MainResourceLoader.h:
3389         * page/AlternativeTextClient.h:
3390         * page/ContextMenuController.cpp:
3391         * page/mac/SettingsMac.mm:
3392         * platform/LocalizedStrings.cpp:
3393         * platform/MemoryPressureHandler.cpp:
3394         * platform/audio/mac/AudioBusMac.mm:
3395         * platform/graphics/Gradient.h: