Source/WebCore/ChangeLog

   1 2014-05-27  Myles C. Maxfield  <mmaxfield@apple.com>
   2
   3         Lists styled with SVG fonts are not rendered as expected
   4         https://bugs.webkit.org/show_bug.cgi?id=133198
   5
   6         Reviewed by Simon Fraser.
   7
   8         There were a couple places in RenderListMarker and RenderMenuList that were
   9         implicitly creating TextRuns by passing a String to a function which
  10         expected a TextRun. Because TextRun has a constructor which takes a single
  11         String and isn't marked explicit, TextRuns were being created without
  12         any of the associated code that initializes the TextRun (such as creating
  13         a RenderingContext if necessary).
  14
  15         There is currently one more client of this code that needs to be fixed.
  16         This is in RenderThemeIOS::adjustMenuListButtonStyle(). However, fixing
  17         this will be a somewhat large change, and is out of scope for this patch.
  18         I will submit a follow-up patch that fixes this permanently. Once this is
  19         fixed, I can mark TextRun's constructors as explicit.
  20
  21         Test: svg/custom/list-items-with-svg-font-family.html
  22
  23         * platform/graphics/TextRun.h:
  24         (WebCore::TextRun::TextRun): Mark constructors as explicit.
  25         * rendering/RenderListMarker.cpp:
  26         (WebCore::RenderListMarker::computePreferredLogicalWidths): Call
  27         RenderBlock::constructTextRun.
  28         (WebCore::RenderListMarker::getRelativeMarkerRect): Ditto.
  29         * rendering/RenderMenuList.cpp:
  30         (RenderMenuList::updateOptionsWidth): Ditto.
  31         * rendering/RenderThemeIOS.mm:
  32         (WebCore::adjustInputElementButtonStyle): Instead of crashing, mark
  33         form controls rendered with SVG fonts as having 0 width. This isn't
  34         perfect, but fixing this is out of scope for this patch.
  35
  36 2014-06-04  Andreas Kling  <akling@apple.com>
  37
  38         WebKit should adopt journal_mode=wal for all SQLite databases.
  39         <https://webkit.org/b/133496>
  40         <rdar://problem/15600186>
  41
  42         We already had code to enable write-ahead logging, but it didn't work
  43         since calling SQLiteStatement::step() without first calling prepare()
  44         is a no-op.
  45
  46         Converted it to use into a simple executeCommand() instead.
  47
  48         Reviewed by Brady Eidson.
  49
  50         * platform/sql/SQLiteDatabase.cpp:
  51         (WebCore::SQLiteDatabase::open):
  52
  53 2014-06-04  Radu Stavila  <stavila@adobe.com>
  54
  55         REGRESSION: [CSS Regions] Regions with overflow:auto generate scrollbars even if the content flows into the following regions and as such, should not be scrolled
  56         https://bugs.webkit.org/show_bug.cgi?id=133481
  57
  58         Reviewed by Antti Koivisto.
  59
  60         When enlarging the layout overflow of a scrollable region, the portion of the content that flows
  61         in the region must be taken into consideration. If the content continues to flow into the next
  62         region, the layout overflow of the current region must not be enlarged.
  63
  64         Test: fast/regions/scrollable-first-region.html
  65
  66         * rendering/RenderFlowThread.cpp:
  67         (WebCore::RenderFlowThread::addRegionsOverflowFromChild):
  68
  69 2014-06-03  Benjamin Poulain  <benjamin@webkit.org>
  70
  71         Fix some corner cases of :hover and :active in quirks mode
  72         https://bugs.webkit.org/show_bug.cgi?id=133493
  73
  74         Reviewed by Andreas Kling.
  75
  76         Previously, the quirks mode of :hover and :active were evaluated from the SelectorFragment
  77         in fragmentOnlyMatchesLinksInQuirksMode(). There were two problems with that approach:
  78         1) The FragmentGenerator optimizes out some filter when they always match or can never match.
  79            Those fragment should still qualify :hover and :active, but they were lost
  80            for fragmentOnlyMatchesLinksInQuirksMode().
  81         2) Explicitely listing every attribute of SelectorFragment in fragmentOnlyMatchesLinksInQuirksMode()
  82            was extremly fragile. When :not() and :-webkit-any() were added, we forgot to update the quirks
  83            exceptions.
  84
  85         This patch moves the quirks mode exception code from the code generator to the Fragment generator.
  86         The useless filters are no longer ignored for the quirks, and we don't need to update the quirks mode detection
  87         with every change of SelectorFragment.
  88
  89         It is inefficient and ugly, but that is how that quirks is defined. It is only a problem for compilation,
  90         execution time remains unchanged.
  91
  92         * cssjit/SelectorCompiler.cpp:
  93         (WebCore::SelectorCompiler::SelectorFragment::SelectorFragment):
  94         (WebCore::SelectorCompiler::addPseudoClassType):
  95         (WebCore::SelectorCompiler::SelectorCodeGenerator::SelectorCodeGenerator):
  96         (WebCore::SelectorCompiler::pseudoClassOnlyMatchesLinksInQuirksMode):
  97         (WebCore::SelectorCompiler::constructFragments):
  98         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateSpecialFailureInQuirksModeForActiveAndHoverIfNeeded):
  99         (WebCore::SelectorCompiler::fragmentOnlyMatchesLinksInQuirksMode): Deleted.
 100
 101 2014-06-03  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
 102
 103         Need to check invalid scheme in navigator content utils
 104         https://bugs.webkit.org/show_bug.cgi?id=133296
 105
 106         Reviewed by Darin Adler.
 107
 108         * Modules/navigatorcontentutils/NavigatorContentUtils.cpp:
 109         (WebCore::verifyProtocolHandlerScheme): Just fix a wrong typo.
 110
 111 2014-06-03  Timothy Horton  <timothy_horton@apple.com>
 112
 113         WebKit2 View Gestures (Zoom): Can show unpainted regions when zoom-pinching-out
 114         https://bugs.webkit.org/show_bug.cgi?id=133472
 115         <rdar://problem/15686848>
 116
 117         Reviewed by Simon Fraser.
 118
 119         * WebCore.exp.in:
 120
 121 2014-06-03  Yoav Weiss  <yoav@yoav.ws>
 122
 123         Refactor the srcset parser into its own file
 124         https://bugs.webkit.org/show_bug.cgi?id=133480
 125
 126         Reviewed by Andreas Kling.
 127
 128         No new tests, since this patch doesn't change the parser's behavior.
 129
 130         This patch moves the srcset parser into its own file,
 131         in order to prepare the code for alignment with recent spec changes.
 132
 133         * CMakeLists.txt:
 134         * html/HTMLImageElement.cpp:
 135         * html/parser/HTMLParserIdioms.cpp:
 136         (WebCore::compareByScaleFactor): Deleted.
 137         (WebCore::parseDescriptors): Deleted.
 138         (WebCore::parseImagesWithScaleFromSrcsetAttribute): Deleted.
 139         (WebCore::bestFitSourceForImageAttributes): Deleted.
 140         * html/parser/HTMLParserIdioms.h:
 141         (WebCore::ImageWithScale::ImageWithScale): Deleted.
 142         (WebCore::ImageWithScale::imageURL): Deleted.
 143         (WebCore::ImageWithScale::scaleFactor): Deleted.
 144         * html/parser/HTMLPreloadScanner.cpp:
 145         * html/parser/HTMLSrcsetParser.cpp: Added.
 146         (WebCore::compareByScaleFactor):
 147         (WebCore::parseDescriptors):
 148         (WebCore::parseImagesWithScaleFromSrcsetAttribute):
 149         (WebCore::bestFitSourceForImageAttributes):
 150         * html/parser/HTMLSrcsetParser.h: Added.
 151         (WebCore::ImageWithScale::ImageWithScale):
 152         (WebCore::ImageWithScale::imageURL):
 153         (WebCore::ImageWithScale::scaleFactor):
 154
 155 2014-06-03  Mario Sanchez Prada  <mario.prada@samsung.com>
 156
 157         [ATK] accessibility/textarea-selected-text-range.html is failing
 158         https://bugs.webkit.org/show_bug.cgi?id=133479
 159
 160         Reviewed by Chris Fleizach.
 161
 162         Implement atk_text_set_caret_offset() in terms of
 163         atk_text_set_selection() for zero-length selections.
 164
 165         No new tests needed, the following test should just start
 166         passing again now: accessibility/textarea-selected-text-range.html.
 167
 168         * accessibility/atk/WebKitAccessibleInterfaceText.cpp:
 169         (webkitAccessibleTextSetCaretOffset): Reimplemented this function
 170         in terms of webkitAccessibleTextSetSelection().
 171
 172 2014-06-03  Alex Christensen  <achristensen@webkit.org>
 173
 174         Corrected arm64 assertion in css jit.
 175
 176         Rubber stamped by Benjamin Poulain.
 177
 178         * cssjit/StackAllocator.h:
 179         (WebCore::StackAllocator::pop):
 180         Changed assertion from incorrectly matching x86_64 to using the correct minimum offset,
 181         taking into account that a stackUnitInBytes is used for two registers on arm64
 182         except the last stack unit, which could only be used for one register if there are
 183         an odd number of registers.
 184
 185 2014-06-03  Benjamin Poulain  <bpoulain@apple.com>
 186
 187         [iOS] Start ignoring common bogus viewport parameters
 188         https://bugs.webkit.org/show_bug.cgi?id=133466
 189         <rdar://problem/16203396>
 190
 191         Reviewed by Sam Weinig.
 192
 193         Incorrect viewports are spreading really quickly lately. We have to start ignoring
 194         some combinations of arguments as they are almost always used incorrectly, causing
 195         a poor user experience on those websites.
 196
 197         * dom/ViewportArguments.cpp:
 198         (WebCore::finalizeViewportArguments):
 199
 200 2014-06-03  Yusuke Suzuki  <utatane.tea@gmail.com>
 201
 202         CSS JIT: add support for the "any" pseudo class
 203         https://bugs.webkit.org/show_bug.cgi?id=133473
 204
 205         Reviewed by Benjamin Poulain.
 206
 207         Add support for the :any pseudo class.
 208
 209         Test: fast/selectors/pseudo-class-any.html
 210
 211         * cssjit/SelectorCompiler.cpp:
 212         (WebCore::SelectorCompiler::addPseudoClassType):
 213         (WebCore::SelectorCompiler::minimumRegisterRequirements):
 214         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatching):
 215         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatchesAnyPseudoClass):
 216
 217 2014-06-02  Jer Noble  <jer.noble@apple.com>
 218
 219         Update PlatformTimeRanges to use MediaTime rather than doubles for time values.
 220         https://bugs.webkit.org/show_bug.cgi?id=133454
 221
 222         Reviewed by Eric Carlson.
 223
 224         PlatformTimeRanges will now use MediaTime, so operations at the platform level won't
 225         lose precision when converting to and from doubles and rational time values like CMTime.
 226
 227         * Modules/mediasource/SourceBuffer.cpp:
 228         (WebCore::currentTimeFudgeFactor):
 229         (WebCore::SourceBuffer::hasCurrentTime):
 230         (WebCore::SourceBuffer::hasFutureTime):
 231         (WebCore::SourceBuffer::canPlayThrough):
 232         * html/HTMLMediaElement.cpp:
 233         (WebCore::HTMLMediaElement::parseAttribute):
 234         * html/TimeRanges.cpp:
 235         (WebCore::TimeRanges::TimeRanges):
 236         (WebCore::TimeRanges::start):
 237         (WebCore::TimeRanges::end):
 238         (WebCore::TimeRanges::add):
 239         (WebCore::TimeRanges::contain):
 240         (WebCore::TimeRanges::find):
 241         (WebCore::TimeRanges::nearest):
 242         (WebCore::TimeRanges::totalDuration):
 243         * platform/graphics/MediaPlayerPrivate.h:
 244         (WebCore::MediaPlayerPrivateInterface::seekable):
 245         * platform/graphics/PlatformTimeRanges.cpp:
 246         (WebCore::PlatformTimeRanges::create):
 247         (WebCore::PlatformTimeRanges::PlatformTimeRanges):
 248         (WebCore::PlatformTimeRanges::invert):
 249         (WebCore::PlatformTimeRanges::start):
 250         (WebCore::PlatformTimeRanges::end):
 251         (WebCore::PlatformTimeRanges::add):
 252         (WebCore::PlatformTimeRanges::contain):
 253         (WebCore::PlatformTimeRanges::find):
 254         (WebCore::PlatformTimeRanges::nearest):
 255         (WebCore::PlatformTimeRanges::totalDuration):
 256         * platform/graphics/PlatformTimeRanges.h:
 257         (WebCore::PlatformTimeRanges::Range::Range):
 258         (WebCore::PlatformTimeRanges::Range::isPointInRange):
 259         * platform/graphics/avfoundation/MediaPlayerPrivateAVFoundation.cpp:
 260         (WebCore::MediaPlayerPrivateAVFoundation::extraMemoryCost):
 261         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
 262         (WebCore::MediaPlayerPrivateAVFoundationCF::platformBufferedTimeRanges):
 263         * platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
 264         (WebCore::MediaPlayerPrivateAVFoundationObjC::platformBufferedTimeRanges):
 265         * platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.mm:
 266         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::seekable):
 267         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::buffered): Deleted.
 268         * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
 269         (WebCore::MediaPlayerPrivateQTKit::buffered):
 270         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
 271         (WebCore::MediaPlayerPrivateGStreamer::buffered):
 272         * platform/mock/mediasource/MockMediaPlayerMediaSource.cpp:
 273         (WebCore::MockMediaPlayerMediaSource::advanceCurrentTime):
 274
 275 2014-06-03  Benjamin Poulain  <bpoulain@apple.com>
 276
 277         [iOS][WK2] Align the highlight rects with device pixels
 278         https://bugs.webkit.org/show_bug.cgi?id=133464
 279
 280         Reviewed by Enrica Casucci.
 281
 282         Add utility functions to pixel align points and rects.
 283
 284         * WebCore.exp.in:
 285         * platform/graphics/FloatPoint.h:
 286         (WebCore::flooredToDevicePixels):
 287         (WebCore::ceiledToDevicePixels):
 288         * platform/graphics/FloatRect.cpp:
 289         (WebCore::enclosingRectExtendedToDevicePixels):
 290         * platform/graphics/FloatRect.h:
 291
 292 2014-06-03  Timothy Horton  <timothy_horton@apple.com>
 293
 294         Prepopulated tiles may never be unparented from TileGrids with aggressive tile retention on
 295         https://bugs.webkit.org/show_bug.cgi?id=133471
 296
 297         Reviewed by Simon Fraser.
 298
 299         * platform/graphics/ca/mac/TileController.h:
 300         * platform/graphics/ca/mac/TileGrid.h:
 301         (WebCore::TileGrid::TileCohortInfo::TileCohortInfo):
 302         * platform/graphics/ca/mac/TileGrid.mm:
 303         (WebCore::TileGrid::TileCohortInfo::timeUntilExpiration):
 304         (WebCore::TileGrid::cohortRemovalTimerFired):
 305         Factor timeUntilExpiration() out of cohortRemovalTimerFired().
 306
 307         (WebCore::TileGrid::revalidateTiles):
 308         Ensure that secondary tiles that never touch the primary coverage rect
 309         are still unparented eventually (after 2 seconds, the same lifetime we use
 310         for cohorts in non-aggressive retention).
 311         Schedule another revalidateTiles if we left parented secondary tiles behind.
 312
 313 2014-06-03  Timothy Horton  <timothy_horton@apple.com>
 314
 315         TileGrid::prepopulateRect populates the wrong tiles if scale != 1
 316         https://bugs.webkit.org/show_bug.cgi?id=133470
 317
 318         Reviewed by Simon Fraser.
 319
 320         * page/FrameView.cpp:
 321         (WebCore::FrameView::requestScrollPositionUpdate):
 322         * platform/graphics/ca/mac/TileGrid.mm:
 323         (WebCore::TileGrid::prepopulateRect):
 324         (WebCore::TileGrid::revalidateTiles):
 325         prepopulateRect takes rects in "content" coordinates, but when later passed to
 326         ensureTilesForRect, they're expected to be in the same coordinate space as the layer's
 327         visible rect, and are scaled up. So, apply the inverse scale to counter that.
 328
 329 2014-06-03  Timothy Horton  <timothy_horton@apple.com>
 330
 331         Minor TileCoverageMap refinements
 332         https://bugs.webkit.org/show_bug.cgi?id=133468
 333
 334         Reviewed by Simon Fraser.
 335
 336         * platform/graphics/ca/mac/TileController.mm:
 337         (WebCore::TileController::setContentsScale):
 338         * platform/graphics/ca/mac/TileCoverageMap.h:
 339         * platform/graphics/ca/mac/TileCoverageMap.mm:
 340         (WebCore::TileCoverageMap::TileCoverageMap):
 341         (WebCore::TileCoverageMap::setDeviceScaleFactor):
 342         Keep the TileCoverageMap's contentsScale at the device scale factor, so that it
 343         looks sharp on non-1x displays.
 344
 345         (WebCore::TileCoverageMap::update):
 346         Decrease the height available for the TileCoverageMap by the topContentInset;
 347         it's already being vertically offset by the topContentInset, but the height was
 348         not adjusted, so it would run off the bottom of the window.
 349
 350         * platform/graphics/ca/mac/TileGrid.mm:
 351         (WebCore::TileGrid::drawTileMapContents):
 352         If using aggressive tile retention (and thus not coloring tiles based on age),
 353         dim tiles that aren't in the visible cohort.
 354
 355 2014-06-03  Timothy Horton  <timothy_horton@apple.com>
 356
 357         Fix a typo, TileGrid::updateTilerLayerProperties
 358         https://bugs.webkit.org/show_bug.cgi?id=133474
 359
 360         Reviewed by Anders Carlsson.
 361
 362         * platform/graphics/ca/mac/TileController.mm:
 363         (WebCore::TileController::setAcceleratesDrawing):
 364         (WebCore::TileController::setTilesOpaque):
 365         (WebCore::TileController::setTileDebugBorderWidth):
 366         (WebCore::TileController::setTileDebugBorderColor):
 367         * platform/graphics/ca/mac/TileGrid.h:
 368         * platform/graphics/ca/mac/TileGrid.mm:
 369         (WebCore::TileGrid::updateTileLayerProperties):
 370         (WebCore::TileGrid::updateTilerLayerProperties):
 371         s/updateTilerLayerProperties/updateTileLayerProperties/g
 372
 373 2014-06-03  Carlos Garcia Campos  <cgarcia@igalia.com>
 374
 375         [SOUP] ResourceLoadTiming.responseStart is always 0
 376         https://bugs.webkit.org/show_bug.cgi?id=133469
 377
 378         Reviewed by Sergio Villar Senin.
 379
 380         * platform/network/soup/ResourceHandleSoup.cpp:
 381         (WebCore::sendRequestCallback): Record the responseStart time.
 382
 383 2014-06-02  David Kilzer  <ddkilzer@apple.com>
 384
 385         [iOS] WebCore fails to build: platform/network/mac/ResourceHandleMac.mm:729:62: error: use of undeclared identifier '_CFURLConnectionCopyTimingData'
 386         <http://webkit.org/b/133445>
 387         <rdar://problem/17090035>
 388
 389         Reviewed by Alexey Proskuryakov.
 390
 391         * platform/network/mac/ResourceHandleMac.mm: Attempt to fix the
 392         build by including a private header and local declaration.
 393
 394 2014-05-31  Anders Carlsson  <andersca@apple.com>
 395
 396         Change DeferrableOneShotTimer to use std::function instead of being a class template
 397         https://bugs.webkit.org/show_bug.cgi?id=133423
 398
 399         Reviewed by Antti Koivisto.
 400
 401         Also, add a new constructor to DeferrableOneShotTimer that takes an std::function. Change the callback
 402         to not pass a DeferrableOneShotTimer& anymore. Change the delay parameter to use std::chrono::milliseconds.
 403
 404         * css/CSSImageGeneratorValue.cpp:
 405         (WebCore::CSSImageGeneratorValue::CachedGeneratedImage::CachedGeneratedImage):
 406         (WebCore::CSSImageGeneratorValue::CachedGeneratedImage::evictionTimerFired):
 407         * css/CSSImageGeneratorValue.h:
 408         * html/HTMLPlugInImageElement.cpp:
 409         (WebCore::HTMLPlugInImageElement::simulatedMouseClickTimerFired):
 410         * html/HTMLPlugInImageElement.h:
 411         * loader/cache/CachedResource.cpp:
 412         (WebCore::deadDecodedDataDeletionIntervalForResourceType):
 413         (WebCore::CachedResource::destroyDecodedDataIfNeeded):
 414         (WebCore::CachedResource::decodedDataDeletionTimerFired):
 415         * loader/cache/CachedResource.h:
 416         * loader/cache/MemoryCache.cpp:
 417         (WebCore::MemoryCache::MemoryCache):
 418         * loader/cache/MemoryCache.h:
 419         (WebCore::MemoryCache::setDeadDecodedDataDeletionInterval):
 420         (WebCore::MemoryCache::deadDecodedDataDeletionInterval):
 421         * platform/Timer.h:
 422         (WebCore::Timer::Timer):
 423         (WebCore::DeferrableOneShotTimer::DeferrableOneShotTimer):
 424         * platform/graphics/cg/SubimageCacheWithTimer.cpp:
 425         (WebCore::SubimageCacheWithTimer::invalidateCacheTimerFired):
 426         * platform/graphics/cg/SubimageCacheWithTimer.h:
 427
 428 2014-06-02  Benjamin Poulain  <benjamin@webkit.org>
 429
 430         Make pseudo element matching for style resolution more JIT friendly
 431         https://bugs.webkit.org/show_bug.cgi?id=133441
 432
 433         Reviewed by Antti Koivisto.
 434
 435         Sharing pseudo elements is a complicated mess shared between the SelectorChecker
 436         and the RuleCollector.
 437
 438         For compiling pseudo element matching, it would be best if all the complexity could
 439         be moved to compile time instead of runtime. It is also preferable to keep the compiler
 440         independant of the RuleCollector implementation's detail.
 441
 442         This patch is a first step toward that goal. The main objective here is to make
 443         "dynamicPseudo" completely internal to the SelectorChecker.
 444         This is mostly a mechanical change, it is likely PseudoElement matching could be simplified
 445         further but that is not attempted here.
 446
 447         Test: fast/css/style-sharing-empty-rule-modification.html
 448
 449         * css/ElementRuleCollector.cpp:
 450         (WebCore::ElementRuleCollector::ruleMatches):
 451         The condition regarding pseudo element matching has been moved into SelectorChecker::match().
 452
 453         This creates an extra cost for querySelector(), but that should be fine since the vast majority
 454         of selectors are JIT compiled and never use SelectorChecker in the first place.
 455
 456         (WebCore::ElementRuleCollector::collectMatchingRulesForList):
 457         The loop is changed such that the tree marking (Style::setHasPseudoStyle) could be moved
 458         to SelectorChecker.
 459
 460         The conditions about the security origin for getMatchedCSSRules() and the empty style rules
 461         are moved prior to the call to ElementRuleCollector::ruleMatches(). This ensure we keep the same
 462         behavior of not adding the pseudo style for empty rules (this could also help with tree marking
 463         in general but empty rules are probably uncommon).
 464
 465         There is one behavior change from making this loop simpler: the SharingRules and StyleInvalidation
 466         modes now update their ruleRange. It is useless but it is simpler.
 467
 468         * css/ElementRuleCollector.h:
 469         * css/SelectorChecker.cpp:
 470         (WebCore::SelectorChecker::matchRecursively):
 471         The tree marking has been moved here.
 472
 473         * css/SelectorChecker.h:
 474         (WebCore::SelectorChecker::match):
 475         The pseudo element matching has been moved here.
 476
 477         * css/StyleResolver.h:
 478         (WebCore::checkRegionSelector):
 479         * dom/SelectorQuery.cpp:
 480         (WebCore::SelectorDataList::selectorMatches):
 481
 482 2014-06-02  Jeremy Jones  <jeremyj@apple.com>
 483
 484         Send external playback properties to fullscreen.
 485         https://bugs.webkit.org/show_bug.cgi?id=133366
 486
 487         Reviewed by Simon Fraser.
 488
 489         Refactor property updates in WebVideoFullscreenModelMediaElement and update
 490         setExternalPlayback on webkitcurrentplaybacktargetiswirelesschangedEvent.
 491
 492         * WebCore.exp.in: Add export for setExternalPlayback.
 493         * platform/ios/WebVideoFullscreenInterface.h: add setExternalPlayback
 494         * platform/ios/WebVideoFullscreenInterfaceAVKit.h: add setExternalPlayback
 495         * platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
 496         (-[WebAVPlayerController isPlayingOnExternalScreen]): added
 497         (+[WebAVPlayerController keyPathsForValuesAffectingPlayingOnExternalScreen]): added
 498         (WebVideoFullscreenInterfaceAVKit::setExternalPlayback): added
 499         (WebVideoFullscreenInterfaceAVKit::enterFullscreen): added
 500         * platform/ios/WebVideoFullscreenModelMediaElement.h: add updateForEventName
 501         * platform/ios/WebVideoFullscreenModelMediaElement.mm:
 502         (WebVideoFullscreenModelMediaElement::setMediaElement): refactor property changes into updateForEventName()
 503         (WebVideoFullscreenModelMediaElement::handleEvent): refactor property changes into updateForEventName()
 504         (WebVideoFullscreenModelMediaElement::updateForEventName): remove duplicate code, implement for setExternalPlayback
 505
 506 2014-06-02  Zsolt Borbely  <zsborbely.u-szeged@partner.samsung.com>
 507
 508         WebCore build fix for webreplay
 509         https://bugs.webkit.org/show_bug.cgi?id=133401
 510
 511         Reviewed by Darin Adler.
 512
 513         * replay/MemoizedDOMResult.cpp:
 514         (WebCore::MemoizedDOMResultBase::createFromEncodedResult):
 515         Add a default nullptr return in order to avoid a gcc warning.
 516         The warning is because of a non-void method that has a case which doesn't return.
 517         * replay/ReplayController.cpp:
 518         (WebCore::ReplayController::setForceDeterministicSettings):
 519         Guard the ScrollingCoordinator with the appropriate #ifdef.
 520         * replay/SerializationMethods.cpp:
 521         Guard the KeypressCommand with COCOA as it is only present there.
 522
 523 2014-06-02  Jer Noble  <jer.noble@apple.com>
 524
 525         Further unreviewed GTK build fix after r169536.
 526
 527         Update the m_duration contructor in MediaSourceGStreamer.
 528
 529         * platform/graphics/gstreamer/MediaSourceGStreamer.cpp:
 530         (WebCore::MediaSourceGStreamer::MediaSourceGStreamer):
 531
 532 2014-06-02  Jer Noble  <jer.noble@apple.com>
 533
 534         Unreviewed GTK build fix after r169536.
 535
 536         Update MediaSourceGStreamer to match new duration parameters.
 537
 538         * platform/graphics/gstreamer/MediaSourceGStreamer.cpp:
 539         (WebCore::MediaSourceGStreamer::setDuration): Ditto.
 540         * platform/graphics/gstreamer/MediaSourceGStreamer.h:
 541         * platform/mock/mediasource/MockMediaPlayerMediaSource.cpp:
 542
 543 2014-06-02  Jer Noble  <jer.noble@apple.com>
 544
 545         [MSE][Mac] Media does not generate 'ended' event when playing to duration.
 546         https://bugs.webkit.org/show_bug.cgi?id=133452
 547
 548         Reviewed by Eric Carlson.
 549
 550         Add a boundaryTimeObserver to the AVSampleBufferRenderSynchronizer to detect when
 551         playback crosses the duration boundary, and pause in that event. Also refuse to begin
 552         playback if the currentTime is beyond the duration. Also, pause playback if duration
 553         becomes less than currentTime.
 554
 555         * platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.h:
 556         * platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.mm:
 557         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::~MediaPlayerPrivateMediaSourceAVFObjC):
 558         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::playInternal):
 559         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::currentMediaTime):
 560         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::currentTimeDouble):
 561         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::durationChanged):
 562
 563         Update MediaSource::duration() and ::setDuration() to take and give MediaTimes rather than doubles:
 564         * Modules/mediasource/MediaSource.cpp:
 565         (WebCore::MediaSource::duration): double -> MediaTime.
 566         (WebCore::MediaSource::setDuration): Ditto.
 567         (WebCore::MediaSource::streamEndedWithError): Ditto.
 568         * platform/graphics/MediaSourcePrivate.h: Ditto.
 569         * platform/graphics/avfoundation/objc/MediaSourcePrivateAVFObjC.h:
 570         * platform/graphics/avfoundation/objc/MediaSourcePrivateAVFObjC.mm:
 571         (WebCore::MediaSourcePrivateAVFObjC::MediaSourcePrivateAVFObjC): Ditto.
 572         (WebCore::MediaSourcePrivateAVFObjC::duration): Ditto.
 573         (WebCore::MediaSourcePrivateAVFObjC::setDuration): Ditto.
 574         * platform/mock/mediasource/MockMediaPlayerMediaSource.cpp:
 575         (WebCore::MockMediaPlayerMediaSource::MockMediaPlayerMediaSource): Ditto.
 576         (WebCore::MockMediaPlayerMediaSource::maxTimeSeekableDouble): Ditto.
 577         (WebCore::MockMediaPlayerMediaSource::durationDouble): Ditto.
 578         (WebCore::MockMediaPlayerMediaSource::advanceCurrentTime): Ditto.
 579         (WebCore::MockMediaPlayerMediaSource::updateDuration): Ditto.
 580         * platform/mock/mediasource/MockMediaPlayerMediaSource.h:
 581         * platform/mock/mediasource/MockMediaSourcePrivate.cpp:
 582         (WebCore::MockMediaSourcePrivate::MockMediaSourcePrivate): Ditto.
 583         (WebCore::MockMediaSourcePrivate::duration): Ditto.
 584         (WebCore::MockMediaSourcePrivate::setDuration): Ditto.
 585         * platform/mock/mediasource/MockMediaSourcePrivate.h:
 586
 587 2014-06-02  Zalan Bujtas  <zalan@apple.com>
 588
 589         Rename ImagePaintingContext to ImagePaintingOptions.
 590         https://bugs.webkit.org/show_bug.cgi?id=133451
 591
 592         Reviewed by Simon Fraser.
 593
 594         No change in functionality.
 595
 596         * WebCore.exp.in:
 597         * html/HTMLCanvasElement.cpp:
 598         (WebCore::HTMLCanvasElement::paint):
 599         * html/canvas/CanvasRenderingContext2D.cpp:
 600         (WebCore::CanvasRenderingContext2D::drawImage):
 601         * platform/graphics/GraphicsContext.cpp:
 602         (WebCore::GraphicsContext::drawImage):
 603         (WebCore::GraphicsContext::drawTiledImage):
 604         (WebCore::GraphicsContext::drawImageBuffer):
 605         * platform/graphics/GraphicsContext.h:
 606         (WebCore::ImagePaintingOptions::ImagePaintingOptions):
 607         (WebCore::ImagePaintingContext::ImagePaintingContext): Deleted.
 608         * rendering/RenderBoxModelObject.cpp:
 609         (WebCore::RenderBoxModelObject::paintFillLayerExtended):
 610         * rendering/RenderEmbeddedObject.cpp:
 611         (WebCore::RenderEmbeddedObject::paintSnapshotImage):
 612         * rendering/RenderImage.cpp:
 613         (WebCore::RenderImage::paintIntoRect):
 614         * rendering/RenderSnapshottedPlugIn.cpp:
 615         (WebCore::RenderSnapshottedPlugIn::paintSnapshot):
 616
 617 2014-05-31  Brent Fulgham  <bfulgham@apple.com>
 618
 619         REGRESSION (r154622): Borders disappear when hovering over cells in table
 620         https://bugs.webkit.org/show_bug.cgi?id=133057
 621         <rdar://problem/15922359>
 622
 623         Reviewed by Tim Horton.
 624
 625         * rendering/RenderTableCell.cpp:
 626         (WebCore::RenderTableCell::paintCollapsedBorders): Revert recent
 627         non-edge collapsed border optimization due to drawing
 628         regressions.
 629
 630 2014-06-02  Zalan Bujtas  <zalan@apple.com>
 631
 632         Provide better encapsulation for image related painting properties.
 633         https://bugs.webkit.org/show_bug.cgi?id=133434
 634
 635         Reviewed by Darin Adler.
 636
 637         Image related drawing property list keeps growing for the functions below and results in
 638         multiple overloadings with seemingly random default parameters.
 639
 640         GraphicsContext::drawImage()
 641         GraphicsContext::drawTiledImage()
 642         GraphicsContext::drawImageBuffer()
 643
 644         This patch provides encapsulation for the image drawing properties. Encapsulation ensures that
 645         introducing a new property does not result in yet another overloaded function.
 646
 647         No change in functionality.
 648
 649         * WebCore.exp.in:
 650         * html/HTMLCanvasElement.cpp:
 651         (WebCore::HTMLCanvasElement::paint):
 652         * html/canvas/CanvasRenderingContext2D.cpp:
 653         (WebCore::CanvasRenderingContext2D::drawImage):
 654         (WebCore::CanvasRenderingContext2D::compositeBuffer):
 655         * platform/graphics/GraphicsContext.cpp:
 656         (WebCore::GraphicsContext::drawImage):
 657         (WebCore::GraphicsContext::drawTiledImage):
 658         (WebCore::GraphicsContext::drawImageBuffer):
 659         * platform/graphics/GraphicsContext.h:
 660         (WebCore::ImagePaintingContext::ImagePaintingContext):
 661         * platform/graphics/ShadowBlur.cpp:
 662         (WebCore::ShadowBlur::endShadowLayer):
 663         * platform/graphics/cairo/ImageBufferCairo.cpp:
 664         (WebCore::ImageBuffer::draw):
 665         * platform/graphics/filters/FEComposite.cpp:
 666         (WebCore::FEComposite::platformApplySoftware):
 667         * rendering/FilterEffectRenderer.cpp:
 668         (WebCore::FilterEffectRendererHelper::applyFilterEffect):
 669         * rendering/RenderBoxModelObject.cpp:
 670         (WebCore::RenderBoxModelObject::paintFillLayerExtended):
 671         * rendering/RenderEmbeddedObject.cpp:
 672         (WebCore::RenderEmbeddedObject::paintSnapshotImage):
 673         * rendering/RenderImage.cpp:
 674         (WebCore::RenderImage::paintReplaced):
 675         (WebCore::RenderImage::paintIntoRect):
 676         * rendering/RenderSnapshottedPlugIn.cpp:
 677         (WebCore::RenderSnapshottedPlugIn::paintSnapshot):
 678
 679 2014-06-02  Commit Queue  <commit-queue@webkit.org>
 680
 681         Unreviewed, rolling out r169529.
 682         https://bugs.webkit.org/show_bug.cgi?id=133447
 683
 684         caused assert in drawImage() (Requested by zalan on #webkit).
 685
 686         Reverted changeset:
 687
 688         "Provide better encapsulation for image related painting
 689         properties."
 690         https://bugs.webkit.org/show_bug.cgi?id=133434
 691         http://trac.webkit.org/changeset/169529
 692
 693 2014-06-02  Zalan Bujtas  <zalan@apple.com>
 694
 695         Provide better encapsulation for image related painting properties.
 696         https://bugs.webkit.org/show_bug.cgi?id=133434
 697
 698         Reviewed by Darin Adler.
 699
 700         Image related drawing property list keeps growing for the functions below and results in
 701         multiple overloadings with seemingly random default parameters.
 702
 703         GraphicsContext::drawImage()
 704         GraphicsContext::drawTiledImage()
 705         GraphicsContext::drawImageBuffer()
 706
 707         This patch provides encapsulation for the image drawing properties. Encapsulation ensures that
 708         introducing a new property does not result in yet another overloaded function.
 709
 710         No change in functionality.
 711
 712         * WebCore.exp.in:
 713         * html/HTMLCanvasElement.cpp:
 714         (WebCore::HTMLCanvasElement::paint):
 715         * html/canvas/CanvasRenderingContext2D.cpp:
 716         (WebCore::CanvasRenderingContext2D::drawImage):
 717         (WebCore::CanvasRenderingContext2D::compositeBuffer):
 718         * platform/graphics/GraphicsContext.cpp:
 719         (WebCore::GraphicsContext::drawImage):
 720         (WebCore::GraphicsContext::drawTiledImage):
 721         (WebCore::GraphicsContext::drawImageBuffer):
 722         * platform/graphics/GraphicsContext.h:
 723         (WebCore::ImagePaintingContext::ImagePaintingContext):
 724         * platform/graphics/ShadowBlur.cpp:
 725         (WebCore::ShadowBlur::endShadowLayer):
 726         * platform/graphics/cairo/ImageBufferCairo.cpp:
 727         (WebCore::ImageBuffer::draw):
 728         * platform/graphics/filters/FEComposite.cpp:
 729         (WebCore::FEComposite::platformApplySoftware):
 730         * rendering/FilterEffectRenderer.cpp:
 731         (WebCore::FilterEffectRendererHelper::applyFilterEffect):
 732         * rendering/RenderBoxModelObject.cpp:
 733         (WebCore::RenderBoxModelObject::paintFillLayerExtended):
 734         * rendering/RenderEmbeddedObject.cpp:
 735         (WebCore::RenderEmbeddedObject::paintSnapshotImage):
 736         * rendering/RenderImage.cpp:
 737         (WebCore::RenderImage::paintReplaced):
 738         (WebCore::RenderImage::paintIntoRect):
 739         * rendering/RenderSnapshottedPlugIn.cpp:
 740         (WebCore::RenderSnapshottedPlugIn::paintSnapshot):
 741
 742 2014-06-02  Kiran  <kiran.guduru@samsung.com>
 743
 744         Add mock DTMFSender support.
 745         https://bugs.webkit.org/show_bug.cgi?id=133298
 746
 747         Reviewed by Eric Carlson.
 748
 749         Implemented mock implementation for RTCPeerConnectionHandler. Now creating
 750         RTCDTMFSenderHandlerMock and returning while creating DTMFSender instead of
 751         returning null.
 752
 753         Test: fast/mediastream/RTCPeerConnection-dtmf.html.
 754
 755         * CMakeLists.txt:
 756         * platform/mock/RTCDTMFSenderHandlerMock.cpp: Added.
 757         (WebCore::RTCDTMFSenderHandlerMock::RTCDTMFSenderHandlerMock):
 758         (WebCore::RTCDTMFSenderHandlerMock::setClient):
 759         (WebCore::RTCDTMFSenderHandlerMock::insertDTMF):
 760         * platform/mock/RTCDTMFSenderHandlerMock.h: Added.
 761         * platform/mock/RTCPeerConnectionHandlerMock.cpp:
 762         (WebCore::RTCPeerConnectionHandlerMock::createDTMFSender):
 763
 764 2014-06-02  Xabier Rodriguez Calvar  <calvaris@igalia.com>
 765
 766         [GTK] media controls should prevent 'click' events from reaching the page
 767         https://bugs.webkit.org/show_bug.cgi?id=133402
 768
 769         Reviewed by Eric Carlson.
 770
 771         When handling events, we events we need to return true in order
 772         that the superclass gets the value and prevents the execution of
 773         the default event handler in the page.
 774
 775         Test: media/media-controls-cancel-events.html
 776
 777         * Modules/mediacontrols/mediaControlsGtk.js:
 778         (ControllerGtk.prototype.handleMuteButtonMouseOver):
 779         (ControllerGtk.prototype.handleVolumeBoxMouseOut):
 780         (ControllerGtk.prototype.handleCaptionButtonClicked):
 781         (ControllerGtk.prototype.handleCaptionButtonMouseOver):
 782         (ControllerGtk.prototype.handleCaptionButtonShowMenu):
 783         (ControllerGtk.prototype.handleCaptionButtonMouseOut):
 784         (ControllerGtk.prototype.handleCaptionMouseOut): Return true to
 785         prevent forwarding the event execution.
 786
 787 2014-06-02  Yusuke Suzuki  <utatane.tea@gmail.com>
 788
 789         CSS JIT: add support for the "not" pseudo class
 790         https://bugs.webkit.org/show_bug.cgi?id=133314
 791
 792         Reviewed by Benjamin Poulain.
 793
 794         Add :not pseudo class support to CSS JIT. Current implementation
 795         doesn't support visitedMatchType. So when link pseudo class appears in
 796         :not sub fragment, fallback to the non-JIT implementation.
 797
 798         Tests: fast/selectors/not-active-hover-quirks.html
 799                fast/selectors/not-active-hover-strict.html
 800                fast/selectors/pseudo-class-not.html
 801
 802         * cssjit/SelectorCompiler.cpp:
 803         (WebCore::SelectorCompiler::addPseudoClassType):
 804         (WebCore::SelectorCompiler::SelectorCodeGenerator::SelectorCodeGenerator):
 805         (WebCore::SelectorCompiler::constructFragments):
 806         (WebCore::SelectorCompiler::minimumRegisterRequirements):
 807         (WebCore::SelectorCompiler::computeBacktrackingInformation):
 808         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateSelectorChecker):
 809         (WebCore::SelectorCompiler::fragmentOnlyMatchesLinksInQuirksMode):
 810         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatching):
 811         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementDataMatching):
 812         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsNthChild):
 813         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatchesNotPseudoClass):
 814         (WebCore::SelectorCompiler::SelectorCodeGenerator::computeBacktrackingInformation): Deleted.
 815
 816 2014-06-01  Jer Noble  <jer.noble@apple.com>
 817
 818         [MSE] Appends of overlapping sample data do not clear existing samples properly.
 819         https://bugs.webkit.org/show_bug.cgi?id=133435
 820
 821         Reviewed by Darin Adler.
 822
 823         Test: media/media-source/media-source-overlapping-append.html
 824
 825         The MSE spec has a spec error in how it removes overlapping frames. The intention seems to be
 826         to check for overlapping samples only when the incoming sample increases the highest presentation time
 827         field. But due to a mismatch in the range of a sample, defined as [start, end), and the check, which
 828         treats the end as inclusive, the overlapping check is almost never run.
 829
 830         Add the sample comparison logic to match [start, end) ranges, and rename the existing check to differentiate
 831         it from the new one:
 832         * Modules/mediasource/SampleMap.cpp:
 833         (WebCore::SamplePresentationTimeIsInsideRangeComparator::operator()):
 834         (WebCore::SamplePresentationTimeIsWithinRangeComparator::operator()):
 835         (WebCore::SampleMap::findSamplesBetweenPresentationTimes):
 836         (WebCore::SampleMap::findSamplesWithinPresentationRange):
 837         * Modules/mediasource/SampleMap.h:
 838
 839         Update the overlapping check to occur whenever the highest presentation time increases, and update the
 840         logic within that check to catch all overlapping frames.
 841         * Modules/mediasource/SourceBuffer.cpp:
 842         (WebCore::SourceBuffer::sourceBufferPrivateDidReceiveInitializationSegment):
 843         (WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample):
 844         (WebCore::SourceBuffer::bufferedSamplesForTrackID):
 845         * Modules/mediasource/SourceBuffer.h:
 846         * WebCore.exp.in:
 847
 848         Add a dump method to MediaSample so that samples can be easily converted to a string for testing.
 849         * platform/MediaSample.h:
 850         * platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:
 851         (WebCore::MediaSampleAVFObjC::presentationSize):
 852         (WebCore::MediaSampleAVFObjC::dump):
 853
 854         Update the Mock MSE implementation to pass along a "generation" field, to aid in testing.
 855         * platform/mock/mediasource/MockBox.cpp:
 856         (WebCore::MockSampleBox::MockSampleBox):
 857         * platform/mock/mediasource/MockBox.h:
 858         * platform/mock/mediasource/MockSourceBufferPrivate.cpp:
 859         (WebCore::MockMediaSample::dump):
 860
 861         Add a method to internals to dump the buffered samples to string.
 862         * Modules/mediasource/SourceBuffer.cpp:
 863         (WebCore::SourceBuffer::bufferedSamplesForTrackID):
 864         * Modules/mediasource/SourceBuffer.h:
 865         * testing/Internals.cpp:
 866         (WebCore::Internals::bufferedSamplesForTrackID):
 867         * testing/Internals.h:
 868         * testing/Internals.idl:
 869
 870 2014-06-01  Nikos Andronikos  <nikos.andronikos-webkit@cisra.canon.com.au>
 871
 872         getBBox() returns (0,0) when width or height is zero.
 873         https://bugs.webkit.org/show_bug.cgi?id=93290
 874
 875         Reviewed by Philip Rogers.
 876
 877         The SVG spec states that a width or height of zero for  rects, circles and
 878         ellipses is valid but disables rendering of that shape.
 879         In this case, getBBox() should return the actual bounding box of the element.
 880         Previously, WebKit always returned (0,0,0,0) rather than the
 881         correct bounding box. This patch fixes that.
 882
 883         Rects and ellipses both now follow the same behaviour where
 884         the fall-back code path is only used when the shape will render (e.g. not
 885         when a dimension <= 0).
 886
 887         This necessitates calling calculateRadiiAndCenter for ellipses even
 888         when the calculated values will be discarded in the fall back code path.
 889         But calculateRadiiAndCenter is cheap and this avoids making changes
 890         elsewhere to stop CG drawing a zero width (or height) ellipse.
 891
 892         Tests: svg/custom/getBBox-js-circle-zerodimension.html
 893                svg/custom/getBBox-js-ellipse-zerodimension.html
 894                svg/custom/getBBox-js-rect-zerodimension.html
 895
 896         * rendering/svg/RenderSVGEllipse.cpp:
 897         (WebCore::RenderSVGEllipse::updateShapeFromElement):
 898         Only follow fall-back code path when shape renders.
 899         width||height==0 is not an error case so calculate bounding box.
 900
 901         * rendering/svg/RenderSVGRect.cpp:
 902         (WebCore::RenderSVGRect::updateShapeFromElement):
 903         Only follow fall-back code path when shape renders.
 904         rx||ry==0 is not an error case so calculate bounding box.
 905
 906 2014-06-01  Simon Fraser  <simon.fraser@apple.com>
 907
 908         Don't scroll to non-integral scroll offsets
 909         https://bugs.webkit.org/show_bug.cgi?id=133436
 910         <rdar://problem/17088336>
 911
 912         Reviewed by Darin Adler.
 913
 914         With the gradual migration of scroll offsets to floating point, we ended
 915         up with a fully floating point code path been event scroll deltas and
 916         setting scroll positions in the scrolling tree.
 917
 918         With some input devices (e.g. Logitech scrollwheel mice), the scroll deltas
 919         come in with non-integral values, and that caused the scroll offset to be non-integral,
 920         resulting in fuzzy layer contents. So round the scroll position.
 921
 922         * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
 923         (WebCore::ScrollingTreeFrameScrollingNodeMac::setScrollPosition):
 924
 925 2014-05-31  Anders Carlsson  <andersca@apple.com>
 926
 927         Add a LazyNeverDestroyed class template and use it
 928         https://bugs.webkit.org/show_bug.cgi?id=133425
 929
 930         Reviewed by Darin Adler.
 931
 932         * Modules/webdatabase/DatabaseBackendBase.cpp:
 933         (WebCore::guidMutex):
 934         * crypto/CryptoAlgorithmRegistry.cpp:
 935         (WebCore::registryMutex):
 936         * inspector/WorkerDebuggerAgent.cpp:
 937         * loader/CrossOriginAccessControl.cpp:
 938         (WebCore::isOnAccessControlResponseHeaderWhitelist):
 939         * platform/network/NetworkStateNotifier.cpp:
 940         (WebCore::networkStateNotifier):
 941         * workers/DefaultSharedWorkerRepository.cpp:
 942         (WebCore::DefaultSharedWorkerRepository::instance):
 943         * workers/DefaultSharedWorkerRepository.h:
 944         * workers/WorkerThread.cpp:
 945         (WebCore::threadSetMutex):
 946         * xml/XMLHttpRequest.cpp:
 947         (WebCore::staticData):
 948
 949 2014-06-01  Zalan Bujtas  <zalan@apple.com>
 950
 951         Subpixel rendering: Selection gap produces a pixel line overlap on inline boxes.
 952         https://bugs.webkit.org/show_bug.cgi?id=133429
 953
 954         Reviewed by Simon Fraser.
 955
 956         Do not floor logical right/left values when selection gap is calculated for RenderBlocks.
 957         It produces overlapping line/gap because the floored value is intersecting/too far from the
 958         non-floored content selection rect.
 959
 960         Tests: fast/inline/hidpi-selection-gap-and-inline-selection-have-gap-rtl.html
 961                fast/inline/hidpi-selection-gap-overlaps-inline-selection.html
 962
 963         * rendering/RenderBlock.cpp:
 964         (WebCore::RenderBlock::logicalLeftSelectionGap):
 965         (WebCore::RenderBlock::logicalRightSelectionGap):
 966
 967 2014-05-31  Timothy Horton  <timothy_horton@apple.com>
 968
 969         REGRESSION (r169412): Garbage pixels when vertical rubber-banding on http://yvoschaap.com/pmarca/
 970         https://bugs.webkit.org/show_bug.cgi?id=133418
 971         <rdar://problem/17065494>
 972
 973         Reviewed by Sam Weinig.
 974
 975         * WebCore.exp.in:
 976         * platform/graphics/FloatPoint.h:
 977         (WebCore::toFloatPoint):
 978         * platform/graphics/Image.cpp:
 979         (WebCore::Image::drawTiled):
 980         The manual tiling code previously adjusted the initial
 981         tiling source rect so that if the pattern was out of phase,
 982         the first row and column would paint only the necessary part
 983         of the tile. However, it did not also adjust the destination rect,
 984         so the tile was painted in entirely the wrong place.
 985         Additional testing (by enabling manual tiling mode for all patterned
 986         images and running all of the layout tests) exposed another bug
 987         where the last row and column would paint too much, which
 988         was resolved by clipping to the outer image's destination rect.
 989
 990 2014-05-31  Benjamin Poulain  <benjamin@webkit.org>
 991
 992         CSS JIT cleanup: move two utility functions out of the properties-matcher section
 993         https://bugs.webkit.org/show_bug.cgi?id=133427
 994
 995         Reviewed by Andreas Kling.
 996
 997         Just move the code, no other modification.
 998
 999         * cssjit/SelectorCompiler.cpp:
1000         (WebCore::SelectorCompiler::SelectorCodeGenerator::jumpIfNoPreviousAdjacentElement):
1001         (WebCore::SelectorCompiler::SelectorCodeGenerator::jumpIfNoNextAdjacentElement):
1002
1003 2014-05-31  Darin Adler  <darin@apple.com>
1004
1005         Viewport percentage tests that resize the viewport are flaky
1006         https://bugs.webkit.org/show_bug.cgi?id=133351
1007
1008         Reviewed by Anders Carlsson.
1009
1010         * page/FrameView.cpp:
1011         (WebCore::FrameView::setFrameRect): Removed code that explicitly sends a resize event here.
1012         The resize event will be sent as a side effect of layout instead.
1013         (WebCore::FrameView::sendResizeEventIfNeeded): Don't send a resize event until layout is done,
1014         since this is what propagates the new geometry in so the JavaScript code can detect it. Also
1015         simplified the logic and streamlined a little, and added a FIXME about the bad idea here where
1016         the code enqueues an event for an unknown time in the future. Removed the unneeded code that
1017         passes in the DOM window as the target to dispatchWindowEvent, since that gets filled in as the
1018         target by default.
1019
1020 2014-05-31  Anders Carlsson  <andersca@apple.com>
1021
1022         Don't use std::mutex::try_lock to check if we own a mutex
1023         https://bugs.webkit.org/show_bug.cgi?id=133421
1024
1025         Reviewed by Dan Bernstein.
1026
1027         According the C++ standard, calling try_lock on a mutex that's already owned by the current thread
1028         is undefined behavior, so don't do that.
1029
1030         * Modules/webdatabase/DatabaseBackendBase.cpp:
1031         (WebCore::guidToVersionMap):
1032         (WebCore::updateGuidVersionMap):
1033         (WebCore::guidToDatabaseMap):
1034         (WebCore::guidForOriginAndName):
1035
1036 2014-05-30  Enrica Casucci  <enrica@apple.com>
1037
1038         REGRESSION (WebKit2): space space to insert period doesn't work in web forms.
1039         https://bugs.webkit.org/show_bug.cgi?id=133413
1040         <rdar://problem/16948333>
1041
1042         Reviewed by Benjamin Poulain.
1043
1044         Adding utility functions and their exports.
1045
1046         * WebCore.exp.in:
1047         * editing/TextIterator.cpp:
1048         (WebCore::plainTextReplacingNoBreakSpace):
1049         * editing/TextIterator.h:
1050         * editing/VisibleUnits.cpp:
1051         (WebCore::charactersAroundPosition):
1052         (WebCore::characterBeforePosition): Deleted.
1053         * editing/VisibleUnits.h:
1054
1055 2014-05-30  Zalan Bujtas  <zalan@apple.com>
1056
1057         Cleanup temporary setting of image interpolation quality in GraphicsContext::drawImage*
1058         https://bugs.webkit.org/show_bug.cgi?id=133407
1059
1060         Reviewed by Darin Adler.
1061
1062         Add helper class to set/reset image interpolation quality value.
1063
1064         No change in functionality.
1065
1066         * platform/graphics/GraphicsContext.cpp:
1067         (WebCore::InterpolationQualityMaintainer::InterpolationQualityMaintainer):
1068         (WebCore::InterpolationQualityMaintainer::~InterpolationQualityMaintainer):
1069         (WebCore::GraphicsContext::drawImage):
1070         (WebCore::GraphicsContext::drawTiledImage):
1071         (WebCore::GraphicsContext::drawImageBuffer):
1072
1073 2014-05-30  Myles C. Maxfield  <mmaxfield@apple.com>
1074
1075         Support -apple-system-font on OS X
1076         https://bugs.webkit.org/show_bug.cgi?id=133382
1077
1078         Addressing post-review comments.
1079
1080         * platform/mac/WebFontCache.mm:
1081         (+[WebFontCache internalFontWithFamily:traits:weight:size:]):
1082
1083 2014-05-30  Prashant Hiremath  <hiremathprashants@gmail.com>
1084
1085         Remove repetitive header in HTMLInputElement.cpp
1086         https://bugs.webkit.org/show_bug.cgi?id=133376
1087
1088         Reviewed by Anders Carlsson.
1089
1090         It is just removal repeative headers hence no test needed.
1091
1092         * html/HTMLInputElement.cpp:
1093
1094 2014-05-30  Zalan Bujtas  <zalan@apple.com>
1095
1096         Remove redundant GraphicsContext::drawImage() function.
1097         https://bugs.webkit.org/show_bug.cgi?id=133347
1098
1099         Reviewed by Simon Fraser.
1100
1101         There are 2 GraphicsContext::drawImage functions with very similar parameter list. This looks
1102         to be a source of confusion and results in passing unneeded parameters.
1103
1104         No change in functionality.
1105
1106         * html/canvas/CanvasRenderingContext2D.cpp:
1107         (WebCore::CanvasRenderingContext2D::drawImage):
1108         (WebCore::drawImageToContext):
1109         * platform/graphics/GraphicsContext.cpp:
1110         (WebCore::GraphicsContext::drawImage):
1111         * platform/graphics/GraphicsContext.h:
1112         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
1113         (WebCore::MediaPlayerPrivateGStreamerBase::paint):
1114         * platform/graphics/texmap/TextureMapperImageBuffer.cpp:
1115         (WebCore::BitmapTextureImageBuffer::updateContents):
1116         * rendering/RenderBoxModelObject.cpp:
1117         (WebCore::RenderBoxModelObject::paintNinePieceImage):
1118
1119 2014-05-29  Alex Christensen  <achristensen@webkit.org>
1120
1121         Prepare css jit for arm64 and other architectures.
1122         https://bugs.webkit.org/show_bug.cgi?id=133394
1123
1124         Reviewed by Benjamin Poulain.
1125
1126         * cssjit/FunctionCall.h:
1127         (WebCore::FunctionCall::swapArguments):
1128         * cssjit/RegisterAllocator.h:
1129         Use a designated temporary register on arm64 to prevent using a deallocated argument register to swap argument registers in.
1130         * cssjit/SelectorCompiler.cpp:
1131         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateSelectorChecker):
1132         Added assertion to make sure there are enough registers.
1133
1134 2014-05-29  Benjamin Poulain  <benjamin@webkit.org>
1135
1136         CSS JIT: add support for the pseudo classes :hover and :active
1137         https://bugs.webkit.org/show_bug.cgi?id=133295
1138
1139         Reviewed by Andreas Kling.
1140
1141         This patch adds compilation support for :hover and :active. The code generation of both of them is trivial,
1142         most of the patch is improving the infrastructure and safety.
1143
1144         The document compatibility mode is moved to its own typed enum. This ensure the values are exclusives and
1145         the native type is known (unsigned char here). The values are changes to binary flags to take advantage of
1146         test-and-branch on ARM64 (not used explicitely, it is automatically generated by the MacroAssembler).
1147
1148         The other important refactoring is the code updating the flags on RenderStyle::NonInheritedFlags.
1149         The code was duplicated in some places so it is move into its own function addFlagToElementStyleFromContext().
1150
1151         The generators themself are very simple. First we check the quirks-mode exceptions, that is generally excluded
1152         at compile time because the conditions are almost never met, even in strict mode.
1153         Then we have two cases: style resolution and query selector. Almost everything is just function calls.
1154
1155         * cssjit/SelectorCompiler.cpp:
1156         (WebCore::SelectorCompiler::SelectorFragment::SelectorFragment):
1157         (WebCore::SelectorCompiler::addPseudoClassType):
1158         (WebCore::SelectorCompiler::SelectorCodeGenerator::addFlagToElementStyleFromContext):
1159         (WebCore::SelectorCompiler::fragmentOnlyMatchesLinksInQuirksMode):
1160         (WebCore::SelectorCompiler::getDocument):
1161         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateSpecialFailureInQuirksModeForActiveAndHoverIfNeeded):
1162         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementMatching):
1163         (WebCore::SelectorCompiler::elementIsActive):
1164         (WebCore::SelectorCompiler::elementIsActiveForStyleResolution):
1165         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsActive):
1166         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsFirstChild):
1167         (WebCore::SelectorCompiler::elementIsHovered):
1168         (WebCore::SelectorCompiler::elementIsHoveredForStyleResolution):
1169         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsHovered):
1170         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsLastChild):
1171         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsOnlyChild):
1172         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsNthChild):
1173         * dom/Document.cpp:
1174         (WebCore::Document::Document):
1175         (WebCore::Document::setCompatibilityMode):
1176         (WebCore::Document::implicitOpen):
1177         (WebCore::Document::cloneDataFromDocument):
1178         * dom/Document.h:
1179         (WebCore::Document::compatibilityModeMemoryOffset):
1180         (WebCore::Document::inQuirksMode):
1181         (WebCore::Document::inLimitedQuirksMode):
1182         (WebCore::Document::inNoQuirksMode):
1183         (WebCore::Document::compatibilityMode): Deleted.
1184         * html/ImageDocument.cpp:
1185         (WebCore::ImageDocument::ImageDocument):
1186         * html/MediaDocument.cpp:
1187         (WebCore::MediaDocument::MediaDocument):
1188         * html/PluginDocument.cpp:
1189         (WebCore::PluginDocument::PluginDocument):
1190         * html/TextDocument.cpp:
1191         (WebCore::TextDocument::TextDocument):
1192         * html/parser/HTMLConstructionSite.cpp:
1193         (WebCore::HTMLConstructionSite::setDefaultCompatibilityMode):
1194         (WebCore::HTMLConstructionSite::setCompatibilityMode):
1195         (WebCore::HTMLConstructionSite::setCompatibilityModeFromDoctype):
1196         (WebCore::HTMLConstructionSite::insertDoctype):
1197         * html/parser/HTMLConstructionSite.h:
1198         * loader/DocumentWriter.cpp:
1199         (WebCore::DocumentWriter::replaceDocument):
1200         * loader/SinkDocument.cpp:
1201         (WebCore::SinkDocument::SinkDocument):
1202         * page/ios/FrameIOS.mm:
1203         (WebCore::Frame::initWithSimpleHTMLDocument):
1204         * rendering/style/RenderStyle.h:
1205
1206 2014-05-28  Jon Honeycutt  <jhoneycutt@apple.com>
1207
1208         Roll r168668 back in.
1209
1210         <https://bugs.webkit.org/show_bug.cgi?id=132621>
1211
1212         Reviewed by Brent Fulgham.
1213
1214         Tests: fast/events/beforeload-assertion.html
1215                fast/events/beforeload-iframe-crash.html
1216                fast/events/beforeload-input-time-crash.html
1217
1218         * WebCore.exp.in:
1219         * dom/Document.cpp:
1220         (WebCore::Document::updateLayoutIgnorePendingStylesheets):
1221         * dom/Document.h:
1222         * html/HTMLAppletElement.cpp:
1223         (WebCore::HTMLAppletElement::renderWidgetForJSBindings):
1224         * html/HTMLEmbedElement.cpp:
1225         (WebCore::HTMLEmbedElement::renderWidgetForJSBindings):
1226         * html/HTMLObjectElement.cpp:
1227         (WebCore::HTMLObjectElement::renderWidgetForJSBindings):
1228         * page/FrameView.cpp:
1229         (WebCore::FrameView::FrameView):
1230         (WebCore::FrameView::reset):
1231         (WebCore::FrameView::updateEmbeddedObjectsTimerFired):
1232         (WebCore::FrameView::flushAnyPendingPostLayoutTasks):
1233         (WebCore::FrameView::performPostLayoutTasks):
1234         * page/FrameView.h:
1235         * testing/Internals.cpp:
1236         (WebCore::Internals::updateLayoutIgnorePendingStylesheetsAndRunPostLayoutTasks):
1237         * testing/Internals.h:
1238         * testing/Internals.idl:
1239
1240 2014-05-29  Myles C. Maxfield  <mmaxfield@apple.com>
1241
1242         Support -apple-system-font on OS X
1243         https://bugs.webkit.org/show_bug.cgi?id=133382
1244
1245         Reviewed by Simon Fraser.
1246
1247         Guard on both family names: "-apple-system-font" and "-webkit-system-font"
1248
1249         Test: platform/mac/fast/text/systemFont.html
1250
1251         * platform/mac/WebFontCache.mm:
1252         (+[WebFontCache internalFontWithFamily:traits:weight:size:]):
1253
1254 2014-05-29  Alexey Proskuryakov  <ap@apple.com>
1255
1256         Loading <object> from WebArchive crashes
1257         https://bugs.webkit.org/show_bug.cgi?id=133386
1258         <rdar://problem/13345509>
1259
1260         Reviewed by Brady Eidson.
1261
1262         Test: webarchive/loading/object.html
1263
1264         This (a) fixes the crash, and (b) avoids the crash.
1265
1266         * loader/DocumentLoader.cpp: (WebCore::DocumentLoader::continueAfterContentPolicy):
1267         Some types of substitute data - such as WebArchive - don't contain HTTP result codes,
1268         so let's not drop to <object> fallback content when status is 0.
1269         And if the load somehow failed anyway, don't crash by trying to deliver substitute data
1270         to a finished loader.
1271
1272 2014-05-29  Alex Christensen  <achristensen@webkit.org>
1273
1274         [iOS WebGL] Fix depth buffer clearing issue.
1275         https://bugs.webkit.org/show_bug.cgi?id=133388
1276
1277         Reviewed by Dean Jackson.
1278
1279         No new tests, but this fixes webgl/1.0.1/conformance/canvas/buffer-preserve-test.html.
1280
1281         * platform/graphics/mac/WebGLLayer.mm:
1282         (-[WebGLLayer display]):
1283         * platform/graphics/opengl/GraphicsContext3DOpenGLCommon.cpp:
1284         (WebCore::GraphicsContext3D::prepareTexture):
1285         Moved marking the layer as composited from the mac-only prepareTexture to
1286         the end of [WebGLLayer display] which is called by mac and iOS.
1287
1288 2014-05-29  Jeremy Jones  <jeremyj@apple.com>
1289
1290         Perserve caption selection in fullscreen.
1291         https://bugs.webkit.org/show_bug.cgi?id=133350
1292
1293         Reviewed by Eric Carlson.
1294
1295         Use the logic from the inline player to calculate the selected caption index.
1296
1297         * html/HTMLMediaElement.h:
1298         (WebCore::HTMLMediaElement::mediaControlsHost):
1299         Expose mediaControlsHost() for use by WebVideoFullscreenModelMediaElement.
1300
1301         * platform/ios/WebVideoFullscreenModelMediaElement.mm:
1302         (WebVideoFullscreenModelMediaElement::updateLegibleOptions):
1303         Use logic from buildCaptionsMenu from mediaControlsApple.js.
1304
1305 2014-05-29  Jeremy Jones  <jeremyj@apple.com>
1306
1307         Use correct version of requiresFullscreenForVideoPlayback().
1308         https://bugs.webkit.org/show_bug.cgi?id=133331
1309
1310         Reviewed by Eric Carlson.
1311
1312         Test: media/video-fullscreeen-only-controls.html
1313
1314         Update the last place not yet using m_mediaSession->requiresFullscreenForVideoPlayback.
1315
1316         * html/HTMLMediaElement.cpp:
1317         (WebCore::HTMLMediaElement::parseAttribute):
1318         switch to media session version for requriesFullscreenForVideoPlayback.
1319
1320 2014-05-29  Carlos Garcia Campos  <cgarcia@igalia.com>
1321
1322         [GTK] Add support for HiDPI icons
1323         https://bugs.webkit.org/show_bug.cgi?id=133377
1324
1325         Reviewed by Martin Robinson.
1326
1327         Use the GTK+ missing icon at 16 or 32 size depending on the device
1328         scale factor.
1329
1330         * platform/graphics/gtk/ImageGtk.cpp:
1331         (WebCore::loadMissingImageIconFromTheme):
1332         (WebCore::Image::loadPlatformResource):
1333
1334 2014-05-29  Owen Taylor  <otaylor@redhat.com>
1335
1336         [GTK] Add HighDPI support for non-accelerated compositing contents
1337         https://bugs.webkit.org/show_bug.cgi?id=131562
1338
1339         Reviewed by Anders Carlsson.
1340
1341         No new tests. This will be tested once we have the proper dependencies in the WebKit testing
1342         JHBuild.
1343
1344         * platform/cairo/WidgetBackingStore.h:
1345         (WebCore::WidgetBackingStore::WidgetBackingStore): Accept a device scale argument.
1346         * platform/cairo/WidgetBackingStoreCairo.cpp: Use the device scale argument to make the surface the proper size and set the surface device scale.
1347         * platform/cairo/WidgetBackingStoreCairo.h: Accept a device scale argument.
1348         * platform/graphics/cairo/CairoUtilities.cpp: Add a new helper to set the device scale if Cairo built against is new enough.
1349         * platform/graphics/cairo/CairoUtilities.h:
1350         * platform/gtk/GtkVersioning.h: Add the HAVE_GTK_SCALE_FACTOR macro.
1351         * platform/gtk/WidgetBackingStoreGtkX11.cpp: Use the device scale argument to make the surface the proper size and set the surface device scale.
1352         * platform/gtk/WidgetBackingStoreGtkX11.h: Accept a device scale argument.
1353
1354 2014-05-28  Brent Fulgham  <bfulgham@apple.com>
1355
1356         Captions during HLS playback flash/are difficult to readily
1357         https://bugs.webkit.org/show_bug.cgi?id=133365
1358         <rdar://problem/16584132>
1359
1360         Reviewed by Eric Carlson.
1361
1362         Correct two problems:
1363         1. Add new logic to handle cues that are identical to existing cues, except
1364            that they cover a slightly later moment in playback. For example, a cue
1365            encoding device might emit a cue at 24 fps, with many of the cues containing
1366            the same content but covering only 1/24th of a second.
1367
1368            Rather than hiding and displaying our cues at the cue encoding rate, we should
1369            simply continue to display the same cue as long as the content is unchanged.
1370         2. The cue "isEqual" logic for TextTrackCueGeneric improperly short-circuited the
1371            VTTCue isEqual logic. This was done to avoid a false failure due to the cue
1372            type being 'generic' rather than 'vtt', but incorrectly treated cues with
1373            non-matching text as equal.
1374
1375         * html/track/DataCue.cpp:
1376         (WebCore::DataCue::cueContentsMatch): Move cue comparison code to new method that
1377         doesn't check that the 'type' of the cue is identical.
1378         (WebCore::DataCue::isEqual): Modified to use new 'cueContentsMatch' method.
1379         (WebCore::DataCue::isStartOfCue): Confirms cues match except for start and end times.
1380         * html/track/DataCue.h:
1381         * html/track/TextTrack.cpp:
1382         (WebCore::TextTrack::hasCue): Modify to check if a new cue is simply a continuation
1383         of an existing cue.
1384         * html/track/TextTrackCue.cpp:
1385         (WebCore::TextTrackCue::cueContentsMatch): New method containing the equality code
1386         that used to live in 'isEqual'.
1387         (WebCore::TextTrackCue::isEqual): Modified to use new 'cueContentsMatch' method.
1388         (WebCore::TextTrackCue::isStartOfCue): Confirms cues match except for start and
1389         end times.
1390         * html/track/TextTrackCue.h:
1391         * html/track/TextTrackCueGeneric.cpp:
1392         (WebCore::TextTrackCueGeneric::cueContentsMatch): New method containing the equality
1393         code that used to live in 'isEqual'.
1394         (WebCore::TextTrackCueGeneric::isEqual): Modified to use new 'cueContentsMatch' method.
1395         (WebCore::TextTrackCueGeneric::isStartOfCue): Confirms cues match except for start and
1396         end times.
1397         * html/track/TextTrackCueGeneric.h:
1398         * html/track/VTTCue.cpp:
1399         (WebCore::VTTCue::cueContentsMatch): New method containing the equality code that
1400         used to live in 'isEqual'.
1401         (WebCore::VTTCue::isEqual): Modified to use new 'cueContentsMatch' method.
1402         (WebCore::VTTCue::isStartOfCue): Confirms cues match except for start and end times.
1403         * html/track/VTTCue.h:
1404
1405 2014-05-28  Oliver Hunt  <oliver@apple.com>
1406
1407         Whoops, update bindings tests results.
1408
1409 2014-05-28  Oliver Hunt  <oliver@apple.com>
1410
1411         Make DOM properties exposed as instance properties use the base object instead of |this|
1412         https://bugs.webkit.org/show_bug.cgi?id=133353
1413
1414         Reviewed by Geoffrey Garen.
1415
1416         Make the bindings generator emit code that uses the base object for
1417         a property, rather than the this object. This means that attributes
1418         that we need to be treated as instance properties will match their old
1419         instance-based behaviour.
1420
1421         * bindings/scripts/CodeGeneratorJS.pm:
1422         (InterfaceRequiresAttributesOnInstanceForCompatibility):
1423         (InterfaceRequiresAttributesOnInstance):
1424         (GetCastingHelperForBaseObject):
1425         (GenerateImplementation):
1426         * bindings/scripts/test/JS/JSTestActiveDOMObject.cpp:
1427         (WebCore::jsTestActiveDOMObjectExcitingAttr):
1428         * bindings/scripts/test/JS/JSTestException.cpp:
1429         (WebCore::jsTestExceptionName):
1430         * bindings/scripts/test/JS/JSTestInterface.cpp:
1431         (WebCore::jsTestInterfaceImplementsStr3):
1432         (WebCore::jsTestInterfaceSupplementalStr3):
1433         (WebCore::setJSTestInterfaceConstructorImplementsStaticAttr):
1434         (WebCore::setJSTestInterfaceImplementsStr2):
1435         (WebCore::setJSTestInterfaceImplementsStr3):
1436         (WebCore::setJSTestInterfaceImplementsNode):
1437         (WebCore::setJSTestInterfaceConstructorSupplementalStaticAttr):
1438         (WebCore::setJSTestInterfaceSupplementalStr2):
1439         (WebCore::setJSTestInterfaceSupplementalStr3):
1440         (WebCore::setJSTestInterfaceSupplementalNode):
1441         * bindings/scripts/test/JS/JSTestNondeterministic.cpp:
1442         (WebCore::setJSTestNondeterministicNondeterministicWriteableAttr):
1443         (WebCore::setJSTestNondeterministicNondeterministicExceptionAttr):
1444         (WebCore::setJSTestNondeterministicNondeterministicGetterExceptionAttr):
1445         (WebCore::setJSTestNondeterministicNondeterministicSetterExceptionAttr):
1446         (WebCore::setJSTestSerializedScriptValueInterfaceCachedValue):
1447         * bindings/scripts/test/JS/JSTestTypedefs.cpp:
1448         (WebCore::jsTestTypedefsConstructorTestSubObj):
1449         (WebCore::setJSTestTypedefsUnsignedLongLongAttr):
1450         (WebCore::setJSTestTypedefsImmutableSerializedScriptValue):
1451         (WebCore::setJSTestTypedefsAttrWithGetterException):
1452         (WebCore::setJSTestTypedefsAttrWithSetterException):
1453         (WebCore::setJSTestTypedefsStringAttrWithGetterException):
1454         (WebCore::setJSTestTypedefsStringAttrWithSetterException):
1455
1456 2014-05-28  Simon Fraser  <simon.fraser@apple.com>
1457
1458         [iOS WK2] Improve behavior of position:fixed inside accelerated overflow-scroll
1459         https://bugs.webkit.org/show_bug.cgi?id=133352
1460
1461         Reviewed by Tim Horton.
1462
1463         When adjusting layers inside accelerated overflow-scroll, pass a delta
1464         down to descendant nodes. Fix behavior of sticky nodes inside fixed nodes
1465         inside overflow scroll (sticky should only be affected if the parent node
1466         is a scrolling node).
1467
1468         * page/scrolling/mac/ScrollingTreeFixedNode.mm:
1469         (WebCore::ScrollingTreeFixedNode::updateLayersAfterAncestorChange): Just remove
1470         some .get()
1471         * page/scrolling/mac/ScrollingTreeStickyNode.mm:
1472         (WebCore::ScrollingTreeStickyNode::updateLayersAfterAncestorChange): Remove .get()s.
1473         We never care about the cumulative delta here, because sticky nodes only respond to
1474         changes in their direct scrolling ancestors. Only adjust the layer if the parent
1475         is a scrolling node.
1476
1477 2014-05-27  Chris Fleizach  <cfleizach@apple.com>
1478
1479         AX: WebKit does not recognize ARIA 1.1 tables
1480         https://bugs.webkit.org/show_bug.cgi?id=133163
1481
1482         Reviewed by Darin Adler.
1483
1484         The AXObjectCache code that determines which object to create based on the role needs to be token fallback aware.
1485
1486         Test: accessibility/table-fallback-roles-expose-element-attributes.html
1487
1488         * accessibility/AXObjectCache.cpp:
1489         (WebCore::nodeHasRole):
1490
1491 2014-05-28  Radu Stavila  <stavila@adobe.com>
1492
1493         REGRESSION (r168046): Invalid layout in multicol
1494         https://bugs.webkit.org/show_bug.cgi?id=133322
1495
1496         Reviewed by Antti Koivisto.
1497
1498         In some situations, deleting the line boxes of a multicol element causes the line to region
1499         link to become invalid.
1500
1501         Test: fast/multicol/newmulticol/multicol-inside-multicol.html
1502
1503         * rendering/RenderFlowThread.cpp:
1504         (WebCore::RenderFlowThread::clearLinesToRegionMap):
1505         (WebCore::RenderFlowThread::deleteLines):
1506         (WebCore::RenderFlowThread::willBeDestroyed):
1507         * rendering/RenderFlowThread.h:
1508
1509 2014-05-28  Ryuan Choi  <ryuan.choi@samsung.com>
1510
1511         Remove DocumentThreadableLoaderClient.h
1512         https://bugs.webkit.org/show_bug.cgi?id=133341
1513
1514         Reviewed by Gyuyoung Kim.
1515
1516         DocumentThreadableLoaderClient is not used anymore.
1517
1518         * WebCore.order:
1519         * loader/DocumentThreadableLoader.cpp:
1520         (WebCore::DocumentThreadableLoader::redirectReceived):
1521         * loader/DocumentThreadableLoaderClient.h: Removed.
1522         * loader/ThreadableLoaderClient.h:
1523         (WebCore::ThreadableLoaderClient::isDocumentThreadableLoaderClient): Deleted.
1524
1525 2014-05-28  Carlos Garcia Campos  <cgarcia@igalia.com>
1526
1527         Unreviewed. Update GObject DOM bindings symbols file after r169017.
1528
1529         * bindings/gobject/webkitdom.symbols: Add webkit_dom_navigator_get_hardware_concurrency.
1530
1531 2014-05-28  Carlos Garcia Campos  <cgarcia@igalia.com>
1532
1533         REGRESSION(r168868): [GTK] Element.offset* client* scroll* API changed
1534         https://bugs.webkit.org/show_bug.cgi?id=133313
1535
1536         Reviewed by Philippe Normand.
1537
1538         Keep old API that uses long also for GObject bindings.
1539
1540         * dom/Element.idl:
1541
1542 2014-05-27  Commit Queue  <commit-queue@webkit.org>
1543
1544         Unreviewed, rolling out r169400.
1545         https://bugs.webkit.org/show_bug.cgi?id=133338
1546
1547         Broke the iOS build in
1548         LocalizedDateCache::calculateMaximumWidth() (Requested by smfr
1549         on #webkit).
1550
1551         Reverted changeset:
1552
1553         "SVG TextRuns do not always get RenderingContexts"
1554         https://bugs.webkit.org/show_bug.cgi?id=133198
1555         http://trac.webkit.org/changeset/169400
1556
1557 2014-05-27  Timothy Horton  <timothy_horton@apple.com>
1558
1559         REGRESSION (Margin tiles): Background gradient on dom.spec.whatwg.org is very slow to render
1560         https://bugs.webkit.org/show_bug.cgi?id=133335
1561         <rdar://problem/17011392>
1562
1563         Reviewed by Simon Fraser.
1564
1565         * platform/graphics/Image.cpp:
1566         (WebCore::Image::drawTiled):
1567         Make cross-platform and use some iOS code which avoids patterned drawing when the pattern buffer
1568         would be very large. Instead, it calls draw() repeatedly over the pattern space, avoiding the allocation
1569         of a huge (in the case of bug 133335, nearly 2GB) buffer.
1570
1571 2014-05-27  Simon Fraser  <simon.fraser@apple.com>
1572
1573         [iOS WK2] Fix behavior of position:sticky inside accelerated overflow-scroll
1574         https://bugs.webkit.org/show_bug.cgi?id=133334
1575         <rdar://problem/16462535>
1576
1577         Reviewed by Tim Horton.
1578
1579         When the scroll position changes in an accelerated overflow-scroll element, we have
1580         to update child nodes in the scrolling tree for position:sticky. That requires a
1581         more generic ability to update the scrolling tree after some arbitrary zoom or
1582         scroll. To do this, we need to know the current fixed position rect, rather than
1583         having it passed in.
1584
1585         So make the fixed position rect available from ScrollingTree, and make it possible
1586         to get the current scrollPosition() from any ScrollingTreeScrollingNode.
1587
1588         Also, implement updateLayersAfterDelegatedScroll() in ScrollingTreeOverflowScrollingNodeIOS,
1589         and have it update descendant layers.
1590
1591         Finally, fix ScrollingTreeOverflowScrollingNode to use the correct rectangle for its
1592         constraints math, using the scroll position of the parent node if appropriate.
1593
1594         * page/scrolling/ScrollingTree.h:
1595         * page/scrolling/ScrollingTreeScrollingNode.h:
1596         * page/scrolling/ios/ScrollingTreeFrameScrollingNodeIOS.mm:
1597         (WebCore::ScrollingTreeFrameScrollingNodeIOS::updateChildNodesAfterScroll):
1598         * page/scrolling/ios/ScrollingTreeIOS.cpp:
1599         (WebCore::ScrollingTreeIOS::fixedPositionRect):
1600         * page/scrolling/ios/ScrollingTreeIOS.h:
1601         * page/scrolling/mac/ScrollingTreeStickyNode.mm:
1602         (WebCore::ScrollingTreeStickyNode::updateLayersAfterAncestorChange):
1603
1604 2014-05-27  Simon Fraser  <simon.fraser@apple.com>
1605
1606         Do some renaming in the scrolling tree
1607         https://bugs.webkit.org/show_bug.cgi?id=133333
1608
1609         Reviewed by Tim Horton.
1610
1611         Rename ScrollingTreeNode::parentScrollPositionDidChange() to updateLayersAfterAncestorChange()
1612         since this goes deep and to indicate that it's about layer updating.
1613
1614         Also rename the "viewportRect" parameter to "fixedPositionRect", since on iOS
1615         this is the custom fixed-position rect that changes on zooming.
1616
1617         * page/scrolling/ScrollingTree.cpp:
1618         (WebCore::ScrollingTree::viewportChangedViaDelegatedScrolling):
1619         * page/scrolling/ScrollingTree.h:
1620         * page/scrolling/ScrollingTreeFrameScrollingNode.h:
1621         * page/scrolling/ScrollingTreeNode.h:
1622         * page/scrolling/ScrollingTreeScrollingNode.h:
1623         * page/scrolling/ios/ScrollingTreeFrameScrollingNodeIOS.h:
1624         * page/scrolling/ios/ScrollingTreeFrameScrollingNodeIOS.mm:
1625         (WebCore::ScrollingTreeFrameScrollingNodeIOS::updateLayersAfterViewportChange):
1626         (WebCore::ScrollingTreeFrameScrollingNodeIOS::updateChildNodesAfterScroll):
1627         * page/scrolling/mac/ScrollingTreeFixedNode.h:
1628         * page/scrolling/mac/ScrollingTreeFixedNode.mm:
1629         (WebCore::ScrollingTreeFixedNode::updateLayersAfterAncestorChange):
1630         (WebCore::ScrollingTreeFixedNode::parentScrollPositionDidChange): Deleted.
1631         * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.h:
1632         * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
1633         (WebCore::ScrollingTreeFrameScrollingNodeMac::setScrollLayerPosition):
1634         * page/scrolling/mac/ScrollingTreeStickyNode.h:
1635         * page/scrolling/mac/ScrollingTreeStickyNode.mm:
1636         (WebCore::ScrollingTreeStickyNode::updateLayersAfterAncestorChange):
1637         (WebCore::ScrollingTreeStickyNode::parentScrollPositionDidChange): Deleted.
1638
1639 2014-05-27  Simon Fraser  <simon.fraser@apple.com>
1640
1641         Instantiate a ScrollingTreeMac on Mac
1642         https://bugs.webkit.org/show_bug.cgi?id=133332
1643
1644         Reviewed by Tim Horton.
1645
1646         Mac previously instantiated a ThreadedScrollingTree, but this has no platform
1647         knowledge. To support future expansion, create a ScrollingTreeMac subclass and
1648         instantiate that on Mac.
1649
1650         Also protect ScrollingCoordinatorMac in PLATFORM(MAC) #ifdefs and remove it from the
1651         EXCLUDED_SOURCE_FILE_NAMES list in the xcconfig file, following general trends.
1652
1653         * Configurations/WebCore.xcconfig:
1654         * WebCore.xcodeproj/project.pbxproj:
1655         * page/scrolling/ThreadedScrollingTree.cpp:
1656         (WebCore::ThreadedScrollingTree::create): Deleted.
1657         * page/scrolling/ThreadedScrollingTree.h:
1658         * page/scrolling/mac/ScrollingCoordinatorMac.h:
1659         * page/scrolling/mac/ScrollingCoordinatorMac.mm:
1660         (WebCore::ScrollingCoordinatorMac::ScrollingCoordinatorMac):
1661         * page/scrolling/mac/ScrollingTreeMac.cpp: Copied from Source/WebCore/page/scrolling/mac/ScrollingCoordinatorMac.h.
1662         (ScrollingTreeMac::create):
1663         (ScrollingTreeMac::ScrollingTreeMac):
1664         * page/scrolling/mac/ScrollingTreeMac.h: Copied from Source/WebCore/page/scrolling/mac/ScrollingCoordinatorMac.h.
1665
1666 2014-05-27  Simon Fraser  <simon.fraser@apple.com>
1667
1668         Prepare for position:sticky support inside accelerated overflow-scroll with WK2
1669         https://bugs.webkit.org/show_bug.cgi?id=133329
1670
1671         Reviewed by Tim Horton.
1672
1673         RenderLayerCompositor::computeStickyViewportConstraints() was always using
1674         the viewport rect as the constraining rect for sticky position. This is not
1675         correct when inside overflow:scroll.
1676
1677         Refactor code in RenderBoxModelObject to make getting the rect easy, and use
1678         it to compute the constraints.
1679
1680         * rendering/RenderBoxModelObject.cpp:
1681         (WebCore::RenderBoxModelObject::constrainingRectForStickyPosition):
1682         (WebCore::RenderBoxModelObject::stickyPositionOffset):
1683         * rendering/RenderBoxModelObject.h:
1684         * rendering/RenderLayerCompositor.cpp:
1685         (WebCore::RenderLayerCompositor::computeStickyViewportConstraints):
1686
1687 2014-05-27  Bem Jones-Bey  <bjonesbe@adobe.com>
1688
1689         vw/vh units used as font/line-height values don't scale with the viewport
1690         https://bugs.webkit.org/show_bug.cgi?id=87846
1691
1692         Reviewed by Darin Adler.
1693
1694         This patch moves the resolution of viewport units to style recalc
1695         time. Currently viewport units are left unresolved during style
1696         recalcs, which leads to many problems with viewport units. Moving the
1697         resolution will fix these problems, as well as reduce the plumbing
1698         that goes on.
1699
1700         This patch touches a lot of files since the valueForLength functions
1701         no longer need a RenderView. The interesting changes are in:
1702
1703         - CSSToLengthConversionData -> CSSPrimitiveValue: for moving
1704             resolution to style recalc time.
1705         - Length / LengthFunctions: no longer needs to know about viewport
1706             units.
1707         - FrameView -> Document -> StyleResolver: for scheduling style recalcs
1708             upon resize
1709
1710         Note that getComputedStyle will now return pixel values when viewport
1711         units are used. This behavior matches Firefox and the css3-cascade
1712         spec.
1713
1714         This is based on a Blink patch by timloh@chromium.org.
1715
1716         Tests: css3/viewport-percentage-lengths/viewport-percentage-lengths-anonymous-block.html
1717                css3/viewport-percentage-lengths/viewport-percentage-lengths-calc.html
1718                css3/viewport-percentage-lengths/viewport-percentage-lengths-percent-size-child.html
1719                css3/viewport-percentage-lengths/viewport-percentage-lengths-relative-font-size.html
1720                css3/viewport-percentage-lengths/viewport-percentage-lengths-resize.html
1721
1722         * WebCore.exp.in: Remove RenderView argument to floatValueForLength.
1723         * accessibility/atk/WebKitAccessibleInterfaceText.cpp:
1724         (getAttributeSetForAccessibilityObject): Remove RenderView argument.
1725         * css/BasicShapeFunctions.cpp:
1726         (WebCore::convertToLength): Ditto.
1727         (WebCore::floatValueForCenterCoordinate): Ditto.
1728         * css/BasicShapeFunctions.h: Ditto.
1729         * css/CSSCalculationValue.cpp:
1730         (WebCore::unitCategory): Remove special handling for viewport units,
1731             as they get resolved to pixels.
1732         (WebCore::createCSS): Ditto.
1733         * css/CSSComputedStyleDeclaration.cpp:
1734         (WebCore::positionOffsetValue): Remove RendewView argument.
1735         (WebCore::getBorderRadiusCornerValues): Remove handling of viewport
1736             units, as they are already resolve to pixels here.
1737         (WebCore::getBorderRadiusCornerValue): Remove RenderView argument.
1738         (WebCore::getBorderRadiusShorthandValue): Ditto.
1739         (WebCore::specifiedValueForGridTrackBreadth): Remove handling of
1740             viewport units, as they are already resolved to pixels here.
1741         (WebCore::specifiedValueForGridTrackSize): Remove RenderView argument.
1742         (WebCore::valueForGridTrackList): Ditto.
1743         (WebCore::lineHeightFromStyle): Ditto.
1744         (WebCore::ComputedStyleExtractor::propertyValue): Ditto.
1745         * css/CSSGradientValue.cpp:
1746         (WebCore::CSSLinearGradientValue::createGradient): Pass RenderView to
1747             CSSToLengthConversionData constructor.
1748         (WebCore::CSSRadialGradientValue::createGradient): Ditto.
1749         * css/CSSPrimitiveValue.cpp:
1750         (WebCore::CSSPrimitiveValue::unitCategory): Remove handling of
1751             viewport units, as they get resolved to pixels.
1752         (WebCore::CSSPrimitiveValue::CSSPrimitiveValue): Ditto.
1753         (WebCore::CSSPrimitiveValue::init): Ditto.
1754         (WebCore::CSSPrimitiveValue::computeLengthDouble): Resolve viewport
1755             units to pixels.
1756         (WebCore::CSSPrimitiveValue::canonicalUnitTypeForCategory): Remove
1757         handling of viewport units, since they get resolved to pixels.
1758         (WebCore::CSSPrimitiveValue::viewportPercentageLength): Deleted.
1759         * css/CSSPrimitiveValue.h:
1760         (WebCore::CSSPrimitiveValue::isLength): Remove special handling of
1761             viewport units, since they get resolved to pixels.
1762         * css/CSSPrimitiveValueMappings.h:
1763         (WebCore::CSSPrimitiveValue::convertToLength): Ditto.
1764         * css/CSSToLengthConversionData.cpp:
1765         (WebCore::CSSToLengthConversionData::viewportWidthFactor): Caclulate
1766             conversion factor for viewport units to pixels. Note this does not
1767             set hasViewportUnits on the RenderStyle if computing the font
1768             size, because in the font size case, the RenderStyle is the
1769             parent's style, not the current renderer's style.
1770         (WebCore::CSSToLengthConversionData::viewportHeightFactor): Ditto.
1771         (WebCore::CSSToLengthConversionData::viewportMinFactor): Ditto.
1772         (WebCore::CSSToLengthConversionData::viewportMaxFactor): Ditto.
1773         * css/CSSToLengthConversionData.h:
1774         (WebCore::CSSToLengthConversionData::CSSToLengthConversionData): Add
1775             RenderView parameter, and make RenderStyle non-const so that
1776             hasViewportUnits can be set on the Style..
1777         (WebCore::CSSToLengthConversionData::style): style is now non-const.
1778         (WebCore::CSSToLengthConversionData::copyWithAdjustedZoom): Handle
1779             RenderView argument.
1780         * css/CSSToStyleMap.cpp:
1781         (WebCore::CSSToStyleMap::mapFillXPosition): Remove handling of
1782             viewport units, as they are already resolved to pixels here.
1783         (WebCore::CSSToStyleMap::mapFillYPosition): Ditto.
1784         * css/DeprecatedStyleBuilder.cpp:
1785         (WebCore::ApplyPropertyLength::applyValue): Ditto.
1786         (WebCore::ApplyPropertyBorderRadius::applyValue): Ditto.
1787         (WebCore::ApplyPropertyComputeLength::applyValue): Ditto.
1788         (WebCore::ApplyPropertyFontSize::applyValue): Ditto.
1789         (WebCore::ApplyPropertyLineHeight::applyValue): Ditto.
1790         (WebCore::ApplyPropertyLineHeightForIOSTextAutosizing::applyValue): Ditto.
1791         (WebCore::ApplyPropertyWordSpacing::applyValue): Ditto.
1792         (WebCore::ApplyPropertyVerticalAlign::applyValue): Ditto.
1793         (WebCore::ApplyPropertyTextIndent::applyValue): Ditto.
1794         * css/LengthFunctions.cpp:
1795         (WebCore::minimumIntValueForLength): Remove RenderView argument.
1796         (WebCore::intValueForLength): Remove RenderView argument and
1797             roundPecentages.
1798         (WebCore::minimumValueForLength): Remove RenderView argument, and
1799             special handling of viewport units.
1800         (WebCore::valueForLength): Ditto.
1801         (WebCore::floatValueForLength): Ditto.
1802         (WebCore::floatSizeForLengthSize): Remove RenderView argument.
1803         * css/LengthFunctions.h: Ditto.
1804         * css/MediaQueryEvaluator.cpp:
1805         (WebCore::MediaQueryEvaluator::eval): Add RenderView argument for
1806             CSSToLengthConversionData constuctor.
1807         * css/StyleResolver.cpp:
1808         (WebCore::StyleResolver::State::clear): Change to 0 arg constructor
1809             for CSSToLengthConversionData.
1810         (WebCore::StyleResolver::State::updateConversionData): In order to get
1811             the RenderView, we need to have a valid Element set on the state.
1812             Since this means that we need to set the conversion data in more
1813             places, move the contruction of conversion data into this helper
1814             method.
1815         (WebCore::StyleResolver::State::initElement): Make sure conversion
1816             data is up to date.
1817         (WebCore::StyleResolver::State::initForStyleResolve): Ditto.
1818         (WebCore::StyleResolver::State::setStyle): When the style is set, we
1819             need to make sure to update the conversion data.
1820         (WebCore::StyleResolver::styleForElement): If the style has viewport
1821             units, flage the document so that we can make sure to recalc the
1822             viewport unit values when the viewport is resized.
1823         (WebCore::StyleResolver::pseudoStyleForElement): Ditto.
1824         (WebCore::StyleResolver::updateFont): Make sure the font having
1825             viewport units updates the style.
1826         (WebCore::StyleResolver::convertToIntLength): Remove handling of
1827         viewport units, as they are resolved to pixels.
1828         (WebCore::StyleResolver::convertToFloatLength): Ditto.
1829         (WebCore::StyleResolver::clearCachedPropertiesAffectedByViewportUnits):
1830             Invalidate the matched properties cache for styles with viewport
1831             units when the viewport is resized.
1832         (WebCore::createGridTrackBreadth): Remove handling of viewport units,
1833             as they are resolved to pixels.
1834         (WebCore::StyleResolver::applyProperty): Ditto.
1835         (WebCore::StyleResolver::hasMediaQueriesAffectedByViewportChange):
1836             Rename to be less confusing, since it has nothing to do with
1837             viewport units.
1838         (WebCore::StyleResolver::createFilterOperations): Remove handling of
1839             viewport units, as they are resolved to pixels.
1840         (WebCore::StyleResolver::affectedByViewportChange): Deleted.
1841         (WebCore::StyleResolver::viewportPercentageValue): Deleted.
1842         * css/StyleResolver.h:
1843         (WebCore::StyleResolver::State::State): C++11 cleanup: 0 -> nullptr.
1844         (WebCore::StyleResolver::State::setFontSizeHasViewportUnits): Set if
1845             the font size is specified in viewport units.
1846         (WebCore::StyleResolver::State::fontSizeHasViewportUnits): Ditto.
1847         (WebCore::StyleResolver::State::setStyle): Deleted.
1848         * css/TransformFunctions.cpp:
1849         (WebCore::convertToFloatLength): Remove handling of viewport units,
1850             since they are resolved to pixels.
1851         * dom/Document.cpp:
1852         (WebCore::Document::Document): Add flag to determine if some style in
1853             the document has viewport units.
1854         (WebCore::Document::pageSizeAndMarginsInPixels): Remove RenderView
1855             argument.
1856         (WebCore::Document::updateViewportUnitsOnResize): Mark elements with
1857             viewport units for style recalc when the viewport size has
1858             changed.
1859         * dom/Document.h:
1860         (WebCore::Document::setHasStyleWithViewportUnits): Flag to determine
1861             if some style in the document has viewport units.
1862         (WebCore::Document::hasStyleWithViewportUnits): Ditto.
1863         * html/HTMLAreaElement.cpp:
1864         (WebCore::HTMLAreaElement::getRegion): Remove RenderView argument.
1865         * page/FrameView.cpp:
1866         (WebCore::FrameView::layout): When the viewport is resized, call
1867             updateViewportUnitsOnResize on the document.
1868         * platform/Length.h:
1869         (WebCore::Length::isSpecified): Remove handling of viewport units,
1870             since they are now resolved to pixels.
1871         (WebCore::Length::viewportPercentageLength): Deleted.
1872         (WebCore::Length::isViewportPercentage): Deleted.
1873         * rendering/ClipPathOperation.h:
1874         (WebCore::ShapeClipPathOperation::pathForReferenceRect): Remove
1875             RenderView argument.
1876         (WebCore::BoxClipPathOperation::pathForReferenceRect): Ditto.
1877         * rendering/RenderBlock.cpp:
1878         (WebCore::RenderBlock::updateBlockChildDirtyBitsBeforeLayout): Remove
1879             handling of viewport units, since they get updated by
1880             updateViewportUnitsOnResize when the viewport is resized.
1881         (WebCore::RenderBlock::nodeAtPoint): Remove RenderView argument.
1882         (WebCore::RenderBlock::lineHeight): Ditto.
1883         * rendering/RenderBox.cpp:
1884         (WebCore::RenderBox::paintBoxDecorations): Remove RenderView argument.
1885         (WebCore::RenderBox::computeContentAndScrollbarLogicalHeightUsing):
1886             Remove handling of viewport units, as they are resolved to pixels.
1887         (WebCore::RenderBox::computePercentageLogicalHeight): Ditto.
1888         (WebCore::RenderBox::computeReplacedLogicalWidthUsing): Ditto.
1889         (WebCore::RenderBox::computeReplacedLogicalHeightUsing): Ditto.
1890         (WebCore::RenderBox::hasViewportPercentageLogicalHeight): Deleted.
1891         * rendering/RenderBox.h:
1892         * rendering/RenderBoxModelObject.cpp:
1893         (WebCore::RenderBoxModelObject::getBackgroundRoundedRect): Remove
1894             RenderView argument.
1895         (WebCore::RenderBoxModelObject::calculateFillTileSize): Remove
1896             handling of viewport units, as they are resolved to pixels.
1897         (WebCore::computeBorderImageSide): Remove RenderView argument.
1898         (WebCore::RenderBoxModelObject::paintNinePieceImage): Ditto.
1899         (WebCore::RenderBoxModelObject::paintBorder): Ditto.
1900         (WebCore::RenderBoxModelObject::paintBoxShadow): Ditto.
1901         * rendering/RenderElement.cpp:
1902         (WebCore::RenderElement::repaintAfterLayoutIfNeeded): Ditto.
1903         * rendering/RenderElement.h:
1904         (WebCore::RenderElement::valueForLength): Remove unused
1905             roundPercentages argument.
1906         (WebCore::RenderElement::minimumValueForLength): Remove unused
1907             RoundPercentages and RenderView arguments.
1908         * rendering/RenderGrid.cpp:
1909         (WebCore::RenderGrid::layoutGridItems): Remove handling of viewport
1910             units as they are resolved to pixels.
1911         * rendering/RenderInline.cpp:
1912         (WebCore::computeMargin): Ditto.
1913         (WebCore::RenderInline::lineHeight): Remove RenderView argument.
1914         * rendering/RenderLayer.cpp:
1915         (WebCore::RenderLayer::setupClipPath): Ditto.
1916         * rendering/RenderLineBreak.cpp:
1917         (WebCore::RenderLineBreak::lineHeight): Ditto.
1918         * rendering/RenderScrollbarPart.cpp:
1919         (WebCore::calcScrollbarThicknessUsing): Ditto.
1920         (WebCore::RenderScrollbarPart::computeScrollbarWidth): Ditto.
1921         (WebCore::RenderScrollbarPart::computeScrollbarHeight): Ditto.
1922         * rendering/RenderTable.cpp:
1923         (WebCore::RenderTable::convertStyleLogicalHeightToComputedHeight):
1924             Remove handling of viewport units, since they are resolved to
1925             pixels.
1926         (WebCore::RenderTable::computePreferredLogicalWidths): Now that
1927             viewport unit values are resolved to pixels at style recalc time,
1928             no special checking is needed to handle them, so update the
1929             comment to reflect that.
1930         * rendering/RenderThemeIOS.mm:
1931         (WebCore::applyCommonButtonPaddingToStyle): Add RenderView argument to
1932             CSSToLengthConversionData constructor.
1933         (WebCore::RenderThemeIOS::adjustButtonStyle): Ditto.
1934         * rendering/RenderThemeMac.mm:
1935         (WebCore::RenderThemeMac::paintMenuListButtonGradients): Remove
1936             RenderView argument.
1937         * rendering/RenderView.cpp:
1938         (WebCore::RenderView::layout): Remove handling of viewport units,
1939             since they are resolved to pixels.
1940         * rendering/RootInlineBox.cpp:
1941         (WebCore::RootInlineBox::ascentAndDescentForBox): Remove RenderView
1942             argument.
1943         * rendering/shapes/Shape.cpp:
1944         (WebCore::Shape::createShape): Remove RenderView argument.
1945         * rendering/shapes/Shape.h:
1946         * rendering/shapes/ShapeOutsideInfo.cpp:
1947         (WebCore::ShapeOutsideInfo::computedShape): Ditto.
1948         * rendering/style/BasicShapes.cpp:
1949         (WebCore::BasicShapeCircle::floatValueForRadiusInBox): Ditto.
1950         (WebCore::BasicShapeCircle::path): Ditto.
1951         (WebCore::BasicShapeEllipse::floatValueForRadiusInBox): Ditto.
1952         (WebCore::BasicShapeEllipse::path): Ditto.
1953         (WebCore::BasicShapePolygon::path): Ditto.
1954         (WebCore::floatSizeForLengthSize): Ditto.
1955         (WebCore::BasicShapeInset::path): Ditto.
1956         * rendering/style/BasicShapes.h:
1957         * rendering/style/RenderStyle.cpp:
1958         (WebCore::calcRadiiFor): Ditto.
1959         (WebCore::RenderStyle::getRoundedBorderFor): Ditto.
1960         (WebCore::RenderStyle::computedLineHeight): Remove handling of
1961             viewport units since they are resolved to pixels.
1962         * rendering/style/RenderStyle.h:
1963         (WebCore::RenderStyle::setHasViewportUnits): Set if this style has
1964         a property set with viewport units.
1965         (WebCore::RenderStyle::hasViewportUnits): Ditto.
1966         * rendering/svg/RenderSVGRoot.cpp:
1967         (WebCore::resolveLengthAttributeForSVG): Remove RenderView argument.
1968         (WebCore::RenderSVGRoot::computeReplacedLogicalWidth): Ditto.
1969         (WebCore::RenderSVGRoot::computeReplacedLogicalHeight): Ditto.
1970         * rendering/svg/SVGRenderingContext.cpp:
1971         (WebCore::SVGRenderingContext::prepareToRenderSVGContent): Ditto.
1972
1973 2014-05-27  Zoltan Horvath  <zoltan@webkit.org>
1974
1975         [CSS Shapes] serialization of the computed value should omit the default radii
1976         https://bugs.webkit.org/show_bug.cgi?id=132747
1977
1978         Reviewed by Darin Adler.
1979
1980         We should omit closest-side value, whereever is possible. Specification:
1981         http://www.w3.org/TR/2014/CR-css-shapes-1-20140320/#basic-shape-computed-values
1982         This change aligns the behavior to the specification.
1983
1984         Added new test cases for the shape-outside parsing tests.
1985
1986         * css/CSSBasicShapes.cpp:
1987         (WebCore::buildCircleString):
1988         (WebCore::CSSBasicShapeCircle::cssText):
1989         (WebCore::CSSBasicShapeEllipse::cssText):
1990
1991 2014-05-27  Commit Queue  <commit-queue@webkit.org>
1992
1993         Unreviewed, rolling out r169399.
1994         https://bugs.webkit.org/show_bug.cgi?id=133326
1995
1996         Causes assertion failures in tests.  See
1997         http://build.webkit.org/results/Apple%20Mavericks%20Debug%20WK1%20(Tests)/r169399%20(5622)/results.html
1998         (Requested by mlam on #webkit).
1999
2000         Reverted changeset:
2001
2002         "Caret's screen position does not update during an overflow
2003         scroll"
2004         https://bugs.webkit.org/show_bug.cgi?id=133256
2005         http://trac.webkit.org/changeset/169399
2006
2007 2014-05-27  Myles C. Maxfield  <mmaxfield@apple.com>
2008
2009         Lists styled with SVG fonts are not rendered as expected
2010         https://bugs.webkit.org/show_bug.cgi?id=133198
2011
2012         Reviewed by Simon Fraser.
2013
2014         There were a couple places in RenderListMarker and RenderMenuList that were
2015         implicitly creating TextRuns by passing a String to a function which
2016         expected a TextRun. Because TextRun has a constructor which takes a single
2017         String and isn't marked explicit, TextRuns were being created without
2018         any of the associated code that initializes the TextRun (such as creating
2019         a RenderingContext if necessary).
2020
2021         This patch marks the aforementioned constructor as "explicit" to discourage
2022         such behavior in the future.
2023
2024         Test: svg/custom/list-items-with-svg-font-family.html
2025
2026         * platform/graphics/TextRun.h:
2027         (WebCore::TextRun::TextRun): Mark constructors as explicit.
2028         * rendering/RenderListMarker.cpp:
2029         (WebCore::RenderListMarker::computePreferredLogicalWidths): Call
2030         RenderBlock::constructTextRun.
2031         (WebCore::RenderListMarker::getRelativeMarkerRect): Ditto.
2032         * rendering/RenderMenuList.cpp:
2033         (RenderMenuList::updateOptionsWidth): Ditto.
2034
2035 2014-05-23  Myles C. Maxfield  <mmaxfield@apple.com>
2036
2037         Caret's screen position does not update during an overflow scroll
2038         https://bugs.webkit.org/show_bug.cgi?id=133256
2039
2040         Reviewed by Simon Fraser.
2041
2042         During a scroll, we set a flag which caret position is gated on, but we never
2043         actually invalidate the caret's position. This patch does so.
2044
2045         Test: editing/input/caret-position-during-scroll.html
2046
2047         * editing/FrameSelection.h:
2048         * rendering/RenderLayer.cpp:
2049         (WebCore::RenderLayer::scrollTo):
2050
2051 2014-05-27  Timothy Horton  <timothy_horton@apple.com>
2052
2053         Turn on the LayerPool for iOS
2054         https://bugs.webkit.org/show_bug.cgi?id=133299
2055
2056         Reviewed by Antti Koivisto.
2057
2058         * Configurations/WebCore.xcconfig:
2059         * platform/cocoa/MemoryPressureHandlerCocoa.mm:
2060         (WebCore::MemoryPressureHandler::platformReleaseMemory):
2061         * platform/graphics/ca/mac/TileController.mm:
2062         (WebCore::TileController::createTileLayer):
2063         * platform/graphics/ca/mac/TileGrid.mm:
2064         (WebCore::TileGrid::removeTiles):
2065         Turn on the LayerPool for iOS.
2066         We've used this on Mac for a long time, and the WebKit1
2067         ("legacy") iOS TileCache also had a similar pool.
2068         This helps a good bit with flashiness while scrolling.
2069
2070 2014-05-27  David Hyatt  <hyatt@apple.com>
2071
2072         REGRESSION: Audio volume scrubber does not show volume level thumb
2073         <rdar://problem/16902004>
2074         https://bugs.webkit.org/show_bug.cgi?id=133310
2075
2076         Reviewed by Dean Jackson.
2077
2078         This was just a silly mistake on my part with the setting of the pagination
2079         layers. Layers inside transforms are supposed to clear out their pagination
2080         status based off their ancestor containing block layers, and not off the
2081         enclosing pagination layer.
2082
2083         Added fast/multicol/pagination/nested-transforms.html
2084
2085         * rendering/RenderLayer.cpp:
2086         (WebCore::RenderLayer::updatePagination):
2087
2088 2014-05-27  Eric Carlson  <eric.carlson@apple.com>
2089
2090         [Mac] media controls should prevent 'click' events from reaching the page
2091         https://bugs.webkit.org/show_bug.cgi?id=133268
2092
2093         Reviewed by Jer Noble.
2094
2095         Test: media/media-controls-cancel-events.html
2096
2097         * Modules/mediacontrols/mediaControlsApple.js:
2098         (Controller.prototype.handleEvent): Allow event handler to cancel event default action.
2099         (Controller.prototype.handleRewindButtonClicked): Cancel event default.
2100         (Controller.prototype.handlePlayButtonClicked): Ditto.
2101         (Controller.prototype.handleMuteButtonClicked): Ditto.
2102         (Controller.prototype.handleMinButtonClicked): Ditto.
2103         (Controller.prototype.handleCaptionButtonClicked): Ditto.
2104         (Controller.prototype.handleFullscreenButtonClicked): Ditto.
2105         * Modules/mediacontrols/mediaControlsiOS.js:
2106         (ControllerIOS.prototype.handlePlayButtonTouchEnd): Ditto.
2107         (ControllerIOS.prototype.handlePlayButtonTouchCancel): Ditto.
2108         (ControllerIOS.prototype.handleFullscreenTouchEnd): Ditto.
2109         (ControllerIOS.prototype.handleFullscreenTouchCancel): Ditto.
2110         (ControllerIOS.prototype.handleStartPlaybackButtonTouchEnd): Ditto.
2111         (ControllerIOS.prototype.handleStartPlaybackButtonTouchCancel): Ditto.
2112         (ControllerIOS.prototype.handleWirelessPickerButtonClicked): Ditto.
2113
2114 2014-05-27  Jon Lee  <jonlee@apple.com>
2115
2116         Fix Mavericks bots from r169383.
2117         https://bugs.webkit.org/show_bug.cgi?id=133311
2118
2119         Reviewed by Eric Carlson.
2120
2121         * platform/graphics/avfoundation/objc/CDMSessionMediaSourceAVFObjC.h: Update #ifdef to check for ENABLE(MEDIA_SOURCE).
2122         * platform/graphics/avfoundation/objc/CDMSessionMediaSourceAVFObjC.mm: Ditto.
2123
2124 2014-05-27  Radu Stavila  <stavila@adobe.com>
2125
2126         REGRESSION (r168046): Incorrect handling of multicol spanner
2127         https://bugs.webkit.org/show_bug.cgi?id=133249
2128
2129         Reviewed by David Hyatt.
2130
2131         When removing the child of an object, the removeChild method should return the next element
2132         following the one being deleted. This will ensure that proper iteration and destruction of
2133         an object's children is possible.
2134
2135         No new tests added. This is because the test that triggers this issue also triggers
2136         another assertion, in RenderMultiColumnSet::requiresBalancing(). However, that is a totally
2137         separate problem and will be handled in a separate bug: https://bugs.webkit.org/show_bug.cgi?id=133309
2138
2139         * rendering/RenderBlock.cpp:
2140         (WebCore::RenderBlock::removeChild):
2141         * rendering/RenderBlock.h:
2142         * rendering/RenderBlockFlow.cpp:
2143         (WebCore::RenderBlockFlow::removeChild):
2144         * rendering/RenderBlockFlow.h:
2145         * rendering/RenderButton.cpp:
2146         (WebCore::RenderButton::removeChild):
2147         * rendering/RenderButton.h:
2148         * rendering/RenderElement.cpp:
2149         (WebCore::RenderElement::removeChild):
2150         (WebCore::RenderElement::removeChildInternal):
2151         * rendering/RenderElement.h:
2152         * rendering/RenderMenuList.cpp:
2153         (RenderMenuList::removeChild):
2154         * rendering/RenderMenuList.h:
2155         * rendering/RenderMultiColumnFlowThread.cpp:
2156         (WebCore::RenderMultiColumnFlowThread::flowThreadDescendantInserted):
2157         * rendering/RenderRuby.cpp:
2158         (WebCore::RenderRubyAsInline::removeChild):
2159         (WebCore::RenderRubyAsBlock::removeChild):
2160         * rendering/RenderRuby.h:
2161         * rendering/RenderRubyRun.cpp:
2162         (WebCore::RenderRubyRun::removeChild):
2163         * rendering/RenderRubyRun.h:
2164         * rendering/mathml/RenderMathMLScripts.cpp:
2165         (WebCore::RenderMathMLScripts::removeChildInternal):
2166         (WebCore::RenderMathMLScripts::removeChild):
2167         (WebCore::RenderMathMLScriptsWrapper::removeChildInternal):
2168         (WebCore::RenderMathMLScriptsWrapper::removeChild):
2169         * rendering/mathml/RenderMathMLScripts.h:
2170         * rendering/svg/RenderSVGContainer.cpp:
2171         (WebCore::RenderSVGContainer::removeChild):
2172         * rendering/svg/RenderSVGContainer.h:
2173         * rendering/svg/RenderSVGInline.cpp:
2174         (WebCore::RenderSVGInline::removeChild):
2175         * rendering/svg/RenderSVGInline.h:
2176         * rendering/svg/RenderSVGRoot.cpp:
2177         (WebCore::RenderSVGRoot::removeChild):
2178         * rendering/svg/RenderSVGRoot.h:
2179         * rendering/svg/RenderSVGText.cpp:
2180         (WebCore::RenderSVGText::removeChild):
2181         * rendering/svg/RenderSVGText.h:
2182
2183 2014-05-27  Jon Lee  <jonlee@apple.com>
2184
2185         Update ENABLE(MEDIA_SOURCE) on Mac
2186         https://bugs.webkit.org/show_bug.cgi?id=133141
2187
2188         Reviewed by Darin Adler.
2189
2190         * Configurations/FeatureDefines.xcconfig:
2191
2192 2014-05-27  Tibor Meszaros  <tmeszaros.u-szeged@partner.samsung.com>
2193
2194         Remove BLOB guards
2195         https://bugs.webkit.org/show_bug.cgi?id=132863
2196
2197         Reviewed by Csaba Osztrogonác.
2198
2199         * Configurations/FeatureDefines.xcconfig:
2200         * Modules/websockets/WebSocketChannel.cpp:
2201         (WebCore::WebSocketChannel::WebSocketChannel):
2202         (WebCore::WebSocketChannel::didFail):
2203         (WebCore::WebSocketChannel::processOutgoingFrameQueue):
2204         (WebCore::WebSocketChannel::abortOutgoingFrameQueue):
2205         * Modules/websockets/WebSocketChannel.h:
2206         * WebCore.exp.in:
2207         * WebCore.vcxproj/WebCore.vcxproj:
2208         * bindings/js/JSBlobCustom.cpp:
2209         (WebCore::JSBlobConstructor::constructJSBlob):
2210         * bindings/js/JSFileReaderCustom.cpp:
2211         * dom/DOMExceptions.in:
2212         * dom/EventTargetFactory.in:
2213         * dom/ScriptExecutionContext.cpp:
2214         (WebCore::ScriptExecutionContext::publicURLManager):
2215         * dom/ScriptExecutionContext.h:
2216         * fileapi/AsyncFileStream.cpp:
2217         * fileapi/AsyncFileStream.h:
2218         * fileapi/Blob.cpp:
2219         (WebCore::Blob::Blob):
2220         * fileapi/Blob.h:
2221         (WebCore::Blob::slice):
2222         * fileapi/Blob.idl:
2223         * fileapi/FileError.h:
2224         * fileapi/FileError.idl:
2225         * fileapi/FileException.cpp:
2226         * fileapi/FileException.h:
2227         * fileapi/FileException.idl:
2228         * fileapi/FileReader.cpp:
2229         * fileapi/FileReader.h:
2230         * fileapi/FileReader.idl:
2231         * fileapi/FileReaderLoader.cpp:
2232         * fileapi/FileReaderLoader.h:
2233         * fileapi/FileReaderLoaderClient.h:
2234         * fileapi/FileReaderSync.cpp:
2235         * fileapi/FileReaderSync.h:
2236         * fileapi/FileReaderSync.idl:
2237         * fileapi/FileThread.cpp:
2238         * fileapi/FileThread.h:
2239         * fileapi/ThreadableBlobRegistry.cpp:
2240         (WebCore::ThreadableBlobRegistry::registerFileBlobURL): Deleted.
2241         (WebCore::ThreadableBlobRegistry::registerBlobURL): Deleted.
2242         (WebCore::ThreadableBlobRegistry::registerBlobURLForSlice): Deleted.
2243         (WebCore::ThreadableBlobRegistry::unregisterBlobURL): Deleted.
2244         (WebCore::ThreadableBlobRegistry::blobSize): Deleted.
2245         (WebCore::ThreadableBlobRegistry::getCachedOrigin): Deleted.
2246         * fileapi/WebKitBlobBuilder.cpp:
2247         (WebCore::BlobBuilder::append):
2248         * fileapi/WebKitBlobBuilder.h:
2249         * html/DOMURL.cpp:
2250         * html/DOMURL.h:
2251         * html/DOMURL.idl:
2252         * html/PublicURLManager.cpp:
2253         * html/PublicURLManager.h:
2254         * loader/LoaderStrategy.cpp:
2255         (WebCore::LoaderStrategy::createBlobRegistry):
2256         * loader/LoaderStrategy.h:
2257         * page/DOMWindow.idl:
2258         * page/SecurityOrigin.cpp:
2259         (WebCore::SecurityOrigin::shouldUseInnerURL):
2260         (WebCore::getCachedOrigin):
2261         * platform/FileStream.cpp:
2262         * platform/FileStream.h:
2263         * platform/FileStreamClient.h:
2264         * platform/SchemeRegistry.cpp:
2265         (WebCore::canDisplayOnlyIfCanRequestSchemes):
2266         * platform/network/BlobRegistry.cpp:
2267         * platform/network/BlobRegistryImpl.cpp:
2268         * platform/network/BlobResourceHandle.cpp:
2269         * platform/network/BlobResourceHandle.h:
2270         * platform/network/FormData.cpp:
2271         (WebCore::FormData::deepCopy):
2272         (WebCore::FormData::appendFile):
2273         (WebCore::FormData::appendBlob):
2274         (WebCore::FormData::appendKeyValuePairItems):
2275         (WebCore::FormData::resolveBlobReferences):
2276         (WebCore::encodeElement):
2277         (WebCore::decodeElement):
2278         * platform/network/FormData.h:
2279         (WebCore::FormDataElement::FormDataElement):
2280         (WebCore::operator==):
2281         * platform/network/cf/FormDataStreamCFNet.cpp:
2282         (WebCore::closeCurrentStream):
2283         (WebCore::advanceCurrentStream):
2284         (WebCore::formCreate):
2285         (WebCore::formRead):
2286         (WebCore::setHTTPBody):
2287         * platform/network/curl/ResourceHandleManager.cpp:
2288         (WebCore::getFormElementsCount):
2289         * platform/network/soup/ResourceHandleSoup.cpp:
2290         (WebCore::addEncodedBlobToSoupMessageBody):
2291         (WebCore::addFormElementsToSoupMessage):
2292         * workers/WorkerGlobalScope.idl:
2293         * xml/XMLHttpRequest.cpp:
2294         (WebCore::XMLHttpRequest::send):
2295         (WebCore::XMLHttpRequest::createRequest):
2296
2297 2014-05-27  Zsolt Borbely  <zsborbely.u-szeged@partner.samsung.com>
2298
2299         Allow building CMake based ports with WEB_REPLAY
2300         https://bugs.webkit.org/show_bug.cgi?id=133154
2301
2302         Reviewed by Csaba Osztrogonác.
2303
2304         * CMakeLists.txt:
2305
2306 2014-05-27  Adrian Perez de Castro  <aperez@igalia.com>
2307
2308         [GStreamer] Add missing <wtf/MathExtras.h> header
2309         https://bugs.webkit.org/show_bug.cgi?id=133236
2310
2311         Reviewed by Philippe Normand.
2312
2313         In some configurations, not including <wtf/MathExtras.h> will cause
2314         errors about undefined functions. In particular, this happens in
2315         Linux with glibc 2.19 and compiling with Clang. Having the header
2316         included is the correct thing to do anyway.
2317
2318         * platform/graphics/gstreamer/GStreamerUtilities.cpp:
2319         Include <wtf/MathExtras.h>
2320
2321 2014-05-27  Manuel Rego Casasnovas  <rego@igalia.com>
2322
2323         Setting up OrderIterator shouldn't require an extra Vector
2324         https://bugs.webkit.org/show_bug.cgi?id=119061
2325
2326         Reviewed by Darin Adler.
2327
2328         From Blink r153971 by <jchaffraix@chromium.org>
2329
2330         This patches reuse the OrderIterator's Vector. It provides a helper class OrderIteratorPopulator, used for
2331         manipulating the Vector directly. Which allows to consolidate the code into a single implementation across
2332         flexbox and grid.
2333
2334         No new tests, already covered by current tests.
2335
2336         * rendering/OrderIterator.cpp:
2337         (WebCore::OrderIteratorPopulator::~OrderIteratorPopulator): Reset OrderIterator and call
2338         removeDuplicatedOrderValues().
2339         (WebCore::OrderIteratorPopulator::removeDuplicatedOrderValues): Sorts the Vector and removes the duplicated
2340         order values.
2341         (WebCore::OrderIteratorPopulator::collectChild): Collect order value information from child.
2342         (WebCore::OrderIterator::setOrderValues): Deleted.
2343         * rendering/OrderIterator.h:
2344         (WebCore::OrderIteratorPopulator::OrderIteratorPopulator): Add helper class to manipulate OrderValues Vector.
2345         * rendering/RenderFlexibleBox.cpp:
2346         (WebCore::RenderFlexibleBox::layoutBlock): Remove unneeded code related to old OrderValues vector.
2347         (WebCore::RenderFlexibleBox::prepareOrderIteratorAndMargins): Populate OrderIterator using collectChild().
2348         (WebCore::RenderFlexibleBox::computeMainAxisPreferredSizes): Deleted.
2349         * rendering/RenderFlexibleBox.h: Rename computeMainAxisPreferredSizes() to prepareOrderIteratorAndMargins().
2350         * rendering/RenderGrid.cpp:
2351         (WebCore::RenderGrid::populateExplicitGridAndOrderIterator): Populate OrderIterator using collectChild().
2352
2353 2014-05-26  Tim Horton  <timothy_horton@apple.com>
2354
2355         [wk2] RemoteLayerBackingStore front buffers should be purgeable when unparented
2356         https://bugs.webkit.org/show_bug.cgi?id=133020
2357         <rdar://problem/16521736>
2358
2359         Reviewed by Simon Fraser.
2360
2361         * WebCore.exp.in:
2362         * platform/graphics/cg/IOSurfacePool.cpp:
2363         (WebCore::IOSurfacePool::willAddSurface):
2364         * platform/graphics/cocoa/IOSurface.h:
2365         * platform/graphics/cocoa/IOSurface.mm:
2366         (IOSurface::releaseGraphicsContext):
2367         Rename clearGraphicsContext to releaseGraphicsContext for clarity.
2368
2369 2014-05-26  Philip Rogers  <pdr@google.com>
2370
2371         Remove special case for transparent SVG root layers
2372         https://bugs.webkit.org/show_bug.cgi?id=116856
2373
2374         Reviewed by Dirk Schulze.
2375
2376         This patch removes a special case from RenderLayer for root SVG layers
2377         with opacity. Instead of checking whether a composited layer exists in
2378         SVGRenderingContext::prepareToRenderSVGContent we now exclude the root
2379         SVG renderobject from the opacity checks in prepareToRenderSVGContent.
2380
2381         This is a merge of https://src.chromium.org/viewvc/blink?view=rev&revision=174102
2382         by Philip Rogers <pdr@chromium.org>.
2383
2384         No new tests as this is covered by svg/custom/composited-svg-with-opacity.html.
2385
2386         * rendering/RenderLayer.cpp:
2387         (WebCore::RenderLayer::isTransparent): Deleted.
2388         * rendering/RenderLayer.h:
2389         * rendering/svg/SVGRenderingContext.cpp:
2390         (WebCore::SVGRenderingContext::prepareToRenderSVGContent):
2391             The opacity case for root svg renderers is now handled by
2392             RenderLayer.
2393
2394 2014-05-26  Yusuke Suzuki  <utatane.tea@gmail.com>
2395
2396         CSS JIT: Fix build error raised when CSS_SELECTOR_JIT_DEBUGGING is 1
2397         https://bugs.webkit.org/show_bug.cgi?id=133266
2398
2399         Reviewed by Benjamin Poulain.
2400
2401         When CSS_SELECTOR_JIT_DEBUGGING is 1, build erros occur on x86_64.
2402
2403         * cssjit/SelectorCompiler.cpp:
2404         (WebCore::SelectorCompiler::SelectorCodeGenerator::generatePrologue):
2405         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateEpilogue):
2406
2407 2014-05-26  Benjamin Poulain  <benjamin@webkit.org>
2408
2409         Fix the quirks mode selector matching of the pseudo classes :hover and :active
2410         https://bugs.webkit.org/show_bug.cgi?id=133063
2411
2412         Reviewed by Antti Koivisto.
2413
2414         Our implementation of the quirks mode of :active and :hover was very wrong. The only
2415         thing it was doing is verify the pseudo class is not the first selector of a fragment
2416         (which was conveniently the only thing that was tested :)).
2417
2418         Since those pseudo class were only checking for the order of the filters, something like
2419             #target:hover
2420         would succeed because :hover is not the first simple selector, while
2421             :hover#target
2422         would fail.
2423
2424         That behavior is also a problem for the CSS JIT as it is an implementation detail of SelectorChecker
2425         and compiling something like that with our out-of-order matching would be nonsense.
2426
2427         This patch update the implementation to follow http://quirks.spec.whatwg.org/#the-:active-and-:hover-quirk
2428         Basically, the only cases that do not work in quirks mode are selectors composed only of "*, :hover and :active".
2429
2430         To implement this behavior, I needed to be able to inspect a complete selector fragment, including
2431         what is before and after :hover/:active.
2432         To do that, I replaced the boolean isSubSelector by a pointer to the first selector of the fragment.
2433         When we need to match :active/:hover in quirks mode, we just go over all the selectors in the fragment
2434         to find one of the qualifying match type.
2435
2436         Tests: fast/selectors/active-hover-quirks.html
2437                fast/selectors/active-quirks.html
2438                fast/selectors/hover-quirks.html
2439
2440         * css/SelectorChecker.cpp:
2441         (WebCore::SelectorChecker::matchRecursively):
2442         (WebCore::canMatchHoverOrActiveInQuirksMode):
2443         (WebCore::SelectorChecker::checkOne):
2444         * css/SelectorChecker.h:
2445         (WebCore::SelectorChecker::SelectorCheckingContext::SelectorCheckingContext):
2446
2447 2014-05-26  Zalan Bujtas  <zalan@apple.com>
2448
2449         Split the call to adjustForLocalZoom out into a separate expression.
2450         https://bugs.webkit.org/show_bug.cgi?id=133286
2451
2452         Reviewed by Darin Adler.
2453
2454         This patch ensures that adjustForLocalZoom() always gets called before we look at
2455         zoomFactor's value. (and not rely on the undefined behavior of the order of function argument evaluation)
2456
2457         * dom/Element.cpp:
2458         (WebCore::Element::offsetLeft):
2459         (WebCore::Element::offsetTop):
2460
2461 2014-05-26  Darin Adler  <darin@apple.com>
2462
2463         Class name matching should use ASCII case-insensitive matching, not Unicode case folding
2464         https://bugs.webkit.org/show_bug.cgi?id=133292
2465
2466         Reviewed by Anders Carlsson.
2467
2468         Tests: fast/dom/getElementsByClassName/ASCII-case-insensitive.html
2469                fast/dom/getElementsByClassName/case-sensitive.html
2470
2471         * dom/SpaceSplitString.cpp:
2472         (WebCore::hasNonASCIIOrUpper): Deleted.
2473         (WebCore::tokenizeSpaceSplitString): Use a for loop instead of while.
2474         (WebCore::spaceSplitStringTable): Renamed from sharedDataMap; the new name is supposed
2475         to help us see the analogy with the atomic string table.
2476         (WebCore::SpaceSplitString::set): Removed unneeded special case for null and preflight
2477         since AtomicString::convertToASCIILowercase now handles both of those. Changed to call
2478         convertToASCIILowercase instead of foldCase, since we don't want to fold non-ASCII.
2479         (WebCore::SpaceSplitString::spaceSplitStringContainsValue): Ditto.
2480         (WebCore::SpaceSplitStringData::create): Marked this inline since it's only called in
2481         one place and that place is in this file. Also used auto a bit and used get instead of
2482         find since the value type is a simple pointer.
2483         (WebCore::SpaceSplitStringData::destroy): Removed unneeded check for null. We never
2484         create any SpaceSplitStringData with empty strings, and a null is a kind of empty string.
2485
2486         * dom/SpaceSplitString.h: Removed some unneeded includes and some unneeded uses of the
2487         inline keyword. Changed types from size_t to unsigned in a couple places; we had a mix
2488         of the types and there was no reason to use size_t there.
2489
2490 2014-05-26  Javier Fernandez  <jfernandez@igalia.com>
2491
2492         [CSS Grid Layout] Implementation of the "grid" shorthand.
2493         https://bugs.webkit.org/show_bug.cgi?id=132122
2494
2495         Reviewed by Darin Adler.
2496
2497         The grid property is a shorthand that sets all of the explicit
2498         grid properties (grid-template-rows, grid-template-columns, and
2499         grid-template-areas) as well as all the implicit grid properties
2500         (grid-auto-rows, grid-auto-columns, and grid-auto-flow) in a
2501         single declaration
2502
2503         Notice that either explicit or implicit grid can be specified,
2504         assigning the initial values to the omitted properties.
2505
2506         Test: fast/css-grid-layout/grid-shorthand-get-set.html
2507
2508         * css/CSSComputedStyleDeclaration.cpp:
2509         (WebCore::ComputedStyleExtractor::propertyValue):
2510         * css/CSSParser.cpp:
2511         (WebCore::CSSParser::parseValue):
2512         (WebCore::CSSParser::parseGridShorthand):
2513         * css/CSSParser.h:
2514         * css/CSSPropertyNames.in:
2515         * css/StylePropertyShorthand.cpp:
2516         (WebCore::webkitGridShorthand):
2517         * css/StylePropertyShorthand.h:
2518
2519 2014-05-26  Zalan Bujtas  <zalan@apple.com>
2520
2521         Subpixel layout: Legacy Element.offset* client* return values are invalid in certain cases.
2522         https://bugs.webkit.org/show_bug.cgi?id=133272
2523
2524         Reviewed by Simon Fraser.
2525
2526         Element.offset* client* functions applied various rounding strategies on the return values before r168868.
2527         (for example, offsetLeft/Top either floored through implicit integer arithmetics or rounded explicitly depending
2528         whether the zoom scale was 1. see http://trac.webkit.org/changeset/168868/trunk/Source/WebCore/dom/Element.cpp for details)
2529         This patch mimics the legacy behavior by either rounding or flooring the return value.
2530
2531         Not testable (legacy switch is not exposed)
2532
2533         * dom/Element.cpp:
2534         (WebCore::adjustForLocalZoom):
2535         (WebCore::convertToNonSubpixelValueIfNeeded):
2536         (WebCore::Element::offsetLeft):
2537         (WebCore::Element::offsetTop):
2538
2539 2014-05-26  Michał Pakuła vel Rutka  <m.pakula@samsung.com>
2540
2541         [EFL] Input fields and text areas are not rendered correctly after r167771
2542         https://bugs.webkit.org/show_bug.cgi?id=133181
2543
2544         Reviewed by Gyuyoung Kim.
2545
2546         r167771 changed arguments of RenderTheme two virtual method arguments which were not reflected
2547         in changes in derived classes, which caused parent method to be called. Arguments in derived class
2548         were updated, also 'override' specifiers were added to all virtual methods to prevent similar errors.
2549
2550         Already covered by fast/forms/textarea-placeholder-wrapping.html
2551
2552         * platform/efl/RenderThemeEfl.cpp:
2553         (WebCore::RenderThemeEfl::paintTextField):
2554         (WebCore::RenderThemeEfl::paintTextArea):
2555         * platform/efl/RenderThemeEfl.h:
2556         (WebCore::RenderThemeEfl::supportsHover):
2557         (WebCore::RenderThemeEfl::supportsControlTints):
2558
2559 2014-05-20  Sergio Villar Senin  <svillar@igalia.com>
2560
2561         [CSS Grid Layout] Children of grid containers must be turned into grid items
2562         https://bugs.webkit.org/show_bug.cgi?id=132991
2563
2564         Reviewed by Darin Adler.
2565
2566         From Blink r150472 by <jchaffraix@chromium.org>
2567
2568         According to specs each child of a grid container must become a
2569         grid item meaning that grid items are grid level boxes and thus,
2570         do not participate in the block formatting context but in the grid
2571         formatting one.
2572
2573         This change updates the grid items' 'display' property after style
2574         resolution so that we match the specification (see section 4. Grid
2575         Items). The spec basically instructs us to compute the value of
2576         'display' by applying the table in CSS2.1 Chapter 9.7
2577         http://www.w3.org/TR/CSS2/visuren.html#dis-pos-flo
2578
2579         Test: fast/css-grid-layout/grid-item-display.html
2580
2581         * css/StyleResolver.cpp:
2582         (WebCore::isDisplayGridBox):
2583         (WebCore::StyleResolver::adjustRenderStyle):
2584
2585 2014-05-26  Zan Dobersek  <zdobersek@igalia.com>
2586
2587         Remove Vector copies in ShorthandPropertyWrapper implementation and use
2588         https://bugs.webkit.org/show_bug.cgi?id=133265
2589
2590         Reviewed by Simon Fraser.
2591
2592         * page/animation/CSSPropertyAnimation.cpp:
2593         (WebCore::ShorthandPropertyWrapper::ShorthandPropertyWrapper): Move the passed-in Vector
2594         into the member variable instead of using Vector<>::swap().
2595         (WebCore::ShorthandPropertyWrapper::propertyWrappers): Return a const reference of the member
2596         variable instead of a const value.
2597         (WebCore::CSSPropertyAnimationWrapperMap::CSSPropertyAnimationWrapperMap): Move the Vector
2598         object into the ShorthandProperthyWrapper constructor.
2599         (WebCore::gatherEnclosingShorthandProperties): Deploy a range-based for-loop.
2600
2601 2014-05-26  Tanay C  <tanay.c@samsung.com>
2602
2603         [EFL] Fix build error in blob.cpp after r168435
2604         https://bugs.webkit.org/show_bug.cgi?id=132678
2605
2606         Reviewed by Alexey Proskuryakov.
2607
2608         * fileapi/Blob.cpp:
2609         (WebCore::Blob::size): Using isInBound to check range of local var
2610         actualsize to resolve build error
2611
2612 2014-05-25  Ryuan Choi  <ryuan.choi@samsung.com>
2613
2614         [EFL] Remove TileCairo and TiledBackingStoreBackendCairo
2615         https://bugs.webkit.org/show_bug.cgi?id=133274
2616
2617         Reviewed by Gyuyoung Kim.
2618
2619         Now, TiledBackingStore is only used for CoordinatedGraphics and the EFL port.
2620         CoordinatedGraphics does not use TileCairo and TiledBackingStoreBackendCairo.
2621
2622         This patch removed them and refactored related files.
2623
2624         * PlatformEfl.cmake:
2625         * PlatformGTK.cmake:
2626         * WebCore.vcxproj/WebCore.vcxproj:
2627         * WebCore.vcxproj/WebCore.vcxproj.filters:
2628         * platform/graphics/TiledBackingStore.cpp:
2629         (WebCore::TiledBackingStore::TiledBackingStore): Deleted.
2630         * platform/graphics/TiledBackingStore.h:
2631         * platform/graphics/TiledBackingStoreBackend.h:
2632         * platform/graphics/cairo/TileCairo.cpp: Removed.
2633         * platform/graphics/cairo/TiledBackingStoreBackendCairo.cpp: Removed.
2634
2635 2014-05-25  Jinwoo Song  <jinwoo7.song@samsung.com>
2636
2637         setData() of DataTransfer has a void return type
2638         https://bugs.webkit.org/show_bug.cgi?id=133108
2639
2640         Reviewed by Alexey Proskuryakov.
2641
2642         According to HTML5 spec, setData() of DataTranfer interface does not return value.
2643         http://www.w3.org/TR/html/editing.html#the-datatransfer-interface
2644
2645         Test: editing/pasteboard/set_data_typeof_return.html
2646
2647         * dom/DataTransfer.cpp:
2648         (WebCore::DataTransfer::setData):
2649         * dom/DataTransfer.h:
2650         * dom/DataTransfer.idl:
2651         * platform/Pasteboard.h:
2652         * platform/efl/PasteboardEfl.cpp:
2653         (WebCore::Pasteboard::writeString):
2654         * platform/gtk/PasteboardGtk.cpp:
2655         (WebCore::Pasteboard::writeString):
2656         * platform/ios/PasteboardIOS.mm:
2657         (WebCore::Pasteboard::writeString):
2658         * platform/mac/PasteboardMac.mm:
2659         (WebCore::Pasteboard::writeString):
2660         * platform/win/PasteboardWin.cpp:
2661         (WebCore::Pasteboard::writeString):
2662
2663 2014-05-25  Benjamin Poulain  <bpoulain@apple.com>
2664
2665         [iOS][WK2] Use ScrollView's scrollOffset as the unobscuredContentRect
2666         https://bugs.webkit.org/show_bug.cgi?id=133262
2667
2668         Reviewed by Simon Fraser.
2669
2670         Since VisibleContentRect was fixed, we were no longer sending scroll events when updating the scrollOffset
2671         when updating the visible content rects. The reason is that the scrollOffset was defined as the top left of the
2672         VisibleContentRect, and as such was already at the end position after updating the unobscured rect.
2673
2674         This patch split the unobscuredContentRect in unobscuredContentSize (updated live on zoom) and the position defined
2675         by the ScrollView's scrollOffset (updated when scrolling).
2676
2677         * WebCore.exp.in:
2678         * platform/ScrollView.h:
2679         * platform/ios/ScrollViewIOS.mm:
2680         (WebCore::ScrollView::unobscuredContentRect):
2681         (WebCore::ScrollView::setUnobscuredContentSize):
2682         (WebCore::ScrollView::setUnobscuredContentRect): Deleted.
2683
2684 2014-05-25  David Kilzer  <ddkilzer@apple.com>
2685
2686         Add type-checked casts for TransformOperations
2687         <http://webkit.org/b/133217>
2688
2689         Reviewed by Simon Fraser.
2690
2691         * platform/graphics/GraphicsLayer.cpp:
2692         (WebCore::GraphicsLayer::validateTransformOperations):
2693         * platform/graphics/ca/GraphicsLayerCA.cpp:
2694         (WebCore::getTransformFunctionValue):
2695         - Switch to type-checked casts.
2696
2697         * platform/graphics/transforms/IdentityTransformOperation.h:
2698         * platform/graphics/transforms/Matrix3DTransformOperation.h:
2699         * platform/graphics/transforms/MatrixTransformOperation.h:
2700         * platform/graphics/transforms/PerspectiveTransformOperation.h:
2701         * platform/graphics/transforms/RotateTransformOperation.h:
2702         * platform/graphics/transforms/ScaleTransformOperation.h:
2703         * platform/graphics/transforms/SkewTransformOperation.h:
2704         * platform/graphics/transforms/TranslateTransformOperation.h:
2705         - Add 'final' to class declaration.
2706         - Add 'override' to overridden methods.
2707         - Add type-checked cast via TRANSFORMOPERATION_TYPE_CASTS macro.
2708         - Move implementation of operator==(const TransformOperation&)
2709           from header to source file so it is able to use a type-checked
2710           cast, and switch back to using a reference instead of a
2711           pointer.
2712         - Add or update ASSERT in private constructor to check for
2713           correct OperationType in classes that represent multiple
2714           types.
2715
2716         * platform/graphics/transforms/Matrix3DTransformOperation.cpp:
2717         (WebCore::Matrix3DTransformOperation::operator==): Added.
2718
2719         * platform/graphics/transforms/MatrixTransformOperation.cpp:
2720         (WebCore::Matrix3DTransformOperation::operator==): Added.
2721         (WebCore::MatrixTransformOperation::blend):
2722         - Switch to type-checked casts and use a reference.
2723
2724         * platform/graphics/transforms/PerspectiveTransformOperation.cpp:
2725         (WebCore::Matrix3DTransformOperation::operator==): Added.
2726         (WebCore::PerspectiveTransformOperation::blend):
2727         * platform/graphics/transforms/RotateTransformOperation.cpp:
2728         (WebCore::Matrix3DTransformOperation::operator==): Added.
2729         (WebCore::RotateTransformOperation::blend):
2730         * platform/graphics/transforms/ScaleTransformOperation.cpp:
2731         (WebCore::Matrix3DTransformOperation::operator==): Added.
2732         (WebCore::ScaleTransformOperation::blend):
2733         * platform/graphics/transforms/SkewTransformOperation.cpp:
2734         (WebCore::Matrix3DTransformOperation::operator==): Added.
2735         (WebCore::SkewTransformOperation::blend):
2736         * platform/graphics/transforms/TranslateTransformOperation.cpp:
2737         (WebCore::Matrix3DTransformOperation::operator==): Added.
2738         (WebCore::TranslateTransformOperation::blend):
2739         - Switch to type-checked casts.
2740
2741         * platform/graphics/transforms/TransformOperation.h:
2742         (WebCore::TransformOperation::isRotateTransformOperationType):
2743         (WebCore::TransformOperation::isScaleTransformOperationType):
2744         (WebCore::TransformOperation::isSkewTransformOperationType):
2745         (WebCore::TransformOperation::isTranslateTransformOperationType):
2746         - Add type-checking methods used in constructors and type-checked
2747           casts.
2748         - Define TRANSFORMOPERATION_TYPE_CASTS macro used by subclasses.
2749
2750 2014-05-25  David Kilzer  <ddkilzer@apple.com>
2751
2752         Crash in WebCore::TextResourceDecoder::checkForCSSCharset
2753         <http://webkit.org/b/133257>
2754         <rdar://problem/17027109>
2755
2756         Reviewed by Alexey Proskuryakov.
2757
2758         Test: fast/encoding/css-charset-missing-semi-colon-and-newline.html
2759
2760         * loader/TextResourceDecoder.cpp:
2761         (WebCore::TextResourceDecoder::checkForCSSCharset): Add early
2762         return.
2763
2764 2014-05-24  Eric Carlson  <eric.carlson@apple.com>
2765
2766         [Mac] preload AVAssetTrack properties before asking for them
2767         https://bugs.webkit.org/show_bug.cgi?id=133240
2768
2769         Reviewed by Jer Noble.
2770
2771         * platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
2772         (WebCore::assetTrackMetadataKeyNames): Add preferredTransform and naturalSize to the list
2773             of properties we preload.
2774
2775 2014-05-24  Zsolt Borbely  <zsborbely.u-szeged@partner.samsung.com>
2776
2777         Remove one of the CSSProperty constructor
2778         https://bugs.webkit.org/show_bug.cgi?id=131094
2779
2780         Reviewed by Darin Adler.
2781
2782         Remove one of the CSSProperty constructor, because it is obsolete.
2783
2784         * css/CSSProperty.h:
2785         (WebCore::CSSProperty::CSSProperty): Deleted.
2786         * css/StyleProperties.h:
2787         (WebCore::StyleProperties::PropertyReference::toCSSProperty):
2788
2789 2014-05-23  Simon Fraser  <simon.fraser@apple.com>
2790
2791         Rename ScrollingTreeScrollingNode's m_scrollPosition to make it clear that it's the value committed from the state tree
2792         https://bugs.webkit.org/show_bug.cgi?id=133254
2793
2794         Reviewed by Tim Horton.
2795
2796         Make ScrollingTreeScrollingNode::scrollPosition() pure virtual, and rename
2797         the member variable and associated getter to make it clear that they relate
2798         to the last committed scroll position.
2799
2800         * page/scrolling/ScrollingTreeScrollingNode.cpp:
2801         (WebCore::ScrollingTreeScrollingNode::updateBeforeChildren):
2802         * page/scrolling/ScrollingTreeScrollingNode.h:
2803         (WebCore::ScrollingTreeScrollingNode::lastCommittedScrollPosition):
2804         (WebCore::ScrollingTreeScrollingNode::scrollPosition): Deleted.
2805
2806 2014-05-23  Simon Fraser  <simon.fraser@apple.com>
2807
2808         Share some more ScrollingTreeScrollingNode code
2809         https://bugs.webkit.org/show_bug.cgi?id=133248
2810
2811         Reviewed by Sam Weinig.
2812
2813         Push scrollBy(), scrollByWithoutContentEdgeConstraints() and setScrollPosition()
2814         down to ScrollingTreeFrameScrollingNode.
2815
2816         This requires that scrollPosition() return the right thing for each class, so make
2817         it virtual. Future patches will reduce the confusion between the committed scroll
2818         position and the one derived from layers.
2819
2820         * page/scrolling/ScrollingTreeFrameScrollingNode.cpp:
2821         (WebCore::ScrollingTreeFrameScrollingNode::scrollBy):
2822         (WebCore::ScrollingTreeFrameScrollingNode::scrollByWithoutContentEdgeConstraints):
2823         (WebCore::ScrollingTreeFrameScrollingNode::setScrollPosition):
2824         * page/scrolling/ScrollingTreeFrameScrollingNode.h:
2825         * page/scrolling/ScrollingTreeScrollingNode.h:
2826         (WebCore::ScrollingTreeScrollingNode::scrollPosition):
2827         * page/scrolling/ios/ScrollingTreeFrameScrollingNodeIOS.h:
2828         * page/scrolling/ios/ScrollingTreeFrameScrollingNodeIOS.mm:
2829         (WebCore::ScrollingTreeFrameScrollingNodeIOS::setScrollPosition): Deleted.
2830         (WebCore::ScrollingTreeFrameScrollingNodeIOS::scrollBy): Deleted.
2831         (WebCore::ScrollingTreeFrameScrollingNodeIOS::scrollByWithoutContentEdgeConstraints): Deleted.
2832         * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.h:
2833         * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
2834         (WebCore::ScrollingTreeFrameScrollingNodeMac::updateBeforeChildren):
2835         (WebCore::ScrollingTreeFrameScrollingNodeMac::immediateScrollBy):
2836         (WebCore::ScrollingTreeFrameScrollingNodeMac::immediateScrollByWithoutContentEdgeConstraints):
2837         (WebCore::ScrollingTreeFrameScrollingNodeMac::scrollPosition):
2838         (WebCore::ScrollingTreeFrameScrollingNodeMac::setScrollPosition):
2839         (WebCore::ScrollingTreeFrameScrollingNodeMac::scrollBy): Deleted.
2840         (WebCore::ScrollingTreeFrameScrollingNodeMac::scrollByWithoutContentEdgeConstraints): Deleted.
2841
2842 2014-05-24  Chris Fleizach  <cfleizach@apple.com>
2843
2844         AX: fix coordinate mapping for iOS accessibility
2845         https://bugs.webkit.org/show_bug.cgi?id=133188
2846
2847         Reviewed by Sam Weinig.
2848
2849         Make WebCore aware of the accessibility point/rect conversion methods.
2850
2851         * accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
2852         (-[WebAccessibilityObjectWrapper convertPointToScreenSpace:]):
2853         (-[WebAccessibilityObjectWrapper convertRectToScreenSpace:]):
2854         * loader/EmptyClients.h:
2855         * page/Chrome.cpp:
2856         (WebCore::Chrome::accessibilityScreenToRootView):
2857         (WebCore::Chrome::rootViewToAccessibilityScreen):
2858         * page/Chrome.h:
2859         * page/ChromeClient.h:
2860         * platform/HostWindow.h:
2861
2862 2014-05-24  Zalan Bujtas  <zalan@apple.com>
2863
2864         Subpixel rendering: Non-compositing transforms with subpixel coordinates paint to wrong position.
2865         https://bugs.webkit.org/show_bug.cgi?id=133184
2866         <rdar://problem/16745606>
2867
2868         Reviewed by Simon Fraser.
2869
2870         Snapping relative negative coordinate values should produce the same position as if they were
2871         positive absolute coordinates.
2872         When a child box gets positioned off of its containers towards top/left, its relative coordinates
2873         become negative. Pixel snapping those negative values should produce the same
2874         final painting position as if the child box was fixed positioned with positive coordinates.
2875         Since halfway values always round away from zero, negative and positive halfway values
2876         produce opposite rounding direction.
2877         This patch ensures that negative halfway values round to the direction as if they were positive.
2878
2879         Test: fast/layers/hidpi-floor-negative-coordinate-values-to-maintain-rounding-direction.html
2880
2881         * platform/LayoutUnit.h:
2882         (WebCore::roundToDevicePixel):
2883         * rendering/RenderLayer.cpp:
2884         (WebCore::RenderLayer::paintLayerByApplyingTransform):
2885
2886 2014-05-24  Frédéric Wang  <fred.wang@free.fr>
2887
2888         Use size variants and glyph assembly from the MATH data.
2889         https://bugs.webkit.org/show_bug.cgi?id=130322
2890
2891         Reviewed by Chris Fleizach.
2892
2893         This patch modifies the RenderMathMLOperator code to use the MATH table
2894         when one is provided in the current font on the <math> tag. More
2895         precisely, the MathVariants table is used to draw a size variant or
2896         a glyph assembly. The displaystyle attribute is not supported yet, so
2897         for now large operators are always assumed to be in display style. The
2898         MATH support does not work well with all platforms+fonts, so at the
2899         moment the default font-family on the <math> is not changed.
2900
2901         Tests: mathml/opentype/large-operators-LatinModern.html
2902                mathml/opentype/opentype-stretchy.html
2903                mathml/opentype/vertical-LatinModern.html
2904
2905         * css/mathml.css: We only specify the default font-family on the math root, so that people can easily style the mathematics.
2906         For now, old fonts without the MATH table are still used as the default.
2907         (math):
2908         (math, mfenced > *): Deleted.
2909         (mo, mfenced): Deleted.
2910         * platform/graphics/SimpleFontData.cpp: don't return the math data if the font is loading.
2911         (WebCore::SimpleFontData::mathData):
2912         * platform/graphics/opentype/OpenTypeMathData.cpp: update #ifdef so that disabling ENABLE_OPENTYPE_MATH won't lead to errors with unused parameters.
2913         (WebCore::OpenTypeMathData::OpenTypeMathData):
2914         (WebCore::OpenTypeMathData::getMathConstant):
2915         (WebCore::OpenTypeMathData::getItalicCorrection):
2916         (WebCore::OpenTypeMathData::getMathVariants):
2917         * rendering/mathml/RenderMathMLOperator.cpp:
2918         (WebCore::RenderMathMLOperator::boundsForGlyph):
2919         (WebCore::RenderMathMLOperator::heightForGlyph):
2920         (WebCore::RenderMathMLOperator::advanceForGlyph):
2921         (WebCore::RenderMathMLOperator::computePreferredLogicalWidths): We handle preferred width of size variants.
2922         (WebCore::RenderMathMLOperator::shouldAllowStretching): This function now only returns whether the operator will stretch and no longer has side effect.
2923         (WebCore::RenderMathMLOperator::getGlyphAssemblyFallBack): We add a function to convert from the MathVariant table data to the format supported by RenderMathMLOperator.
2924         (WebCore::RenderMathMLOperator::getDisplayStyleLargeOperator): We add a function to get the glyph that will be used for large operators in display style.
2925         (WebCore::RenderMathMLOperator::findStretchyData): We make this function handle size variants.
2926         (WebCore::RenderMathMLOperator::updateStyle): We handle size variants.
2927         (WebCore::RenderMathMLOperator::paint): We handle size variants.
2928         * rendering/mathml/RenderMathMLOperator.h:
2929
2930 2014-05-23  Tim Horton  <timothy_horton@apple.com>
2931
2932         REGRESSION (iOS WebKit2): requestAnimationFrame fires more than once between layer tree commits
2933         https://bugs.webkit.org/show_bug.cgi?id=132794
2934         <rdar://problem/16877909>
2935
2936         Reviewed by Simon Fraser.
2937
2938         Virtualize DisplayRefreshMonitor so that WebKit2 can implement its own DisplayRefreshMonitor for UI-side compositing views.
2939         This allows the synchronization of requestAnimationFrame callbacks with UI-side compositing painting.
2940
2941         * WebCore.exp.in:
2942         * WebCore.xcodeproj/project.pbxproj:
2943
2944         * page/ChromeClient.h:
2945         (WebCore::ChromeClient::graphicsLayerFactory):
2946         (WebCore::ChromeClient::createDisplayRefreshMonitor):
2947         * platform/graphics/DisplayRefreshMonitorFactory.h: Added.
2948         * platform/graphics/GraphicsLayerUpdater.cpp:
2949         (WebCore::GraphicsLayerUpdater::createDisplayRefreshMonitor):
2950         * platform/graphics/GraphicsLayerUpdater.h:
2951         * rendering/RenderLayerCompositor.cpp:
2952         (WebCore::RenderLayerCompositor::createDisplayRefreshMonitor):
2953         * rendering/RenderLayerCompositor.h:
2954         Give the ChromeClient a chance to provide us with a custom DisplayRefreshMonitor.
2955         If it does not, we'll fall back to making a DisplayRefreshMonitorMac or
2956         DisplayRefreshMonitorIOS depending on the platform.
2957
2958         * dom/ScriptedAnimationController.cpp:
2959         * dom/ScriptedAnimationController.h:
2960         * platform/graphics/DisplayRefreshMonitor.cpp:
2961         * platform/graphics/DisplayRefreshMonitor.h:
2962         * platform/graphics/DisplayRefreshMonitorClient.cpp: Added.
2963         * platform/graphics/DisplayRefreshMonitorClient.h: Added.
2964         * platform/graphics/DisplayRefreshMonitorManager.cpp: Added.
2965         * platform/graphics/DisplayRefreshMonitorManager.h: Added.
2966         Split DisplayRefreshMonitor.cpp into one file per class.
2967         Use references in a few places.
2968         Remove some useless comments.
2969
2970         * platform/graphics/DisplayRefreshMonitor.cpp:
2971         (WebCore::DisplayRefreshMonitor::displayDidRefresh):
2972         Use takeAny() and remove a FIXME about it not existing.
2973
2974         * platform/graphics/DisplayRefreshMonitorClient.cpp: Added.
2975         (WebCore::DisplayRefreshMonitorClient::fireDisplayRefreshIfNeeded):
2976         Use an early return instead of having the whole function body in an if().
2977
2978         * platform/graphics/ios/DisplayRefreshMonitorIOS.h: Added.
2979         * platform/graphics/ios/DisplayRefreshMonitorIOS.mm:
2980         * platform/graphics/mac/DisplayRefreshMonitorMac.cpp:
2981         * platform/graphics/mac/DisplayRefreshMonitorMac.h: Added.
2982         Virtualize DisplayRefreshMonitorIOS and DisplayRefreshMonitorMac,
2983         and move things specific to each of them out of DisplayRefreshMonitor itself.
2984
2985 2014-05-23  Yusuke Suzuki  <utatane.tea@gmail.com>
2986
2987         CSS JIT: Apply backtracking optimization to adjacent backtracking
2988         https://bugs.webkit.org/show_bug.cgi?id=132951
2989
2990         Reviewed by Benjamin Poulain.
2991
2992         Apply the backtracking optimization to the adjacent backtracking.
2993         This optimization is already done for the descendant backtracking.
2994         We apply this to the adjacent backtracking similarly.
2995
2996         Test: fast/selectors/backtracking-adjacent-optimized.html
2997
2998         * cssjit/SelectorCompiler.cpp:
2999         (WebCore::SelectorCompiler::SelectorFragment::SelectorFragment):
3000         (WebCore::SelectorCompiler::solveAdjacentBacktrackingActionForDirectAdjacent):
3001         (WebCore::SelectorCompiler::solveBacktrackingAction):
3002         (WebCore::SelectorCompiler::computeBacktrackingStartOffsetInChain):
3003         (WebCore::SelectorCompiler::computeBacktrackingHeightFromDescendant):
3004         (WebCore::SelectorCompiler::computeBacktrackingWidthFromIndirectAdjacent):
3005         (WebCore::SelectorCompiler::SelectorCodeGenerator::computeBacktrackingInformation):
3006         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateIndirectAdjacentTreeWalker):
3007         (WebCore::SelectorCompiler::SelectorCodeGenerator::linkFailures):
3008         (WebCore::SelectorCompiler::computeBacktrackingStartHeightFromDescendant): Deleted.
3009
3010 2014-05-23  Alex Christensen  <achristensen@webkit.org>
3011
3012         Make CSS JIT run on ARM64.
3013         https://bugs.webkit.org/show_bug.cgi?id=133156
3014
3015         Reviewed by Benjamin Poulain.
3016
3017         * cssjit/FunctionCall.h:
3018         (WebCore::FunctionCall::saveAllocatedRegisters):
3019         (WebCore::FunctionCall::restoreAllocatedRegisters):
3020         Use StackAllocator's new push and pop functions to push and pop a vector instead of iterating it.
3021         * cssjit/RegisterAllocator.h:
3022         (WebCore::RegisterAllocator::reserveCalleeSavedRegisters):
3023         (WebCore::RegisterAllocator::restoreCalleeSavedRegisters):
3024         Return a vector of registers to allocate instead of doing the allocation to make the RegisterAllocator
3025         not need to know about the StackAllocator and to use the new vector push and pop functions.
3026         (WebCore::RegisterAllocator::~RegisterAllocator):
3027         Store RegisterIDs instead of StackReferences to avoid needing to know about the stack.
3028         * cssjit/SelectorCompiler.cpp:
3029         (WebCore::SelectorCompiler::SelectorCodeGenerator::compile):
3030         Removed the requirement for assert to be disabled to print disassembly when debugging css jit.
3031         (WebCore::SelectorCompiler::SelectorCodeGenerator::generatePrologue):
3032         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateEpilogue):
3033         Added to correctly push the link register and frame pointer.
3034         This is required if the jit code calls a function on arm64 and helpful for debugging tools on x86_64.
3035         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateSelectorChecker):
3036         Generate the prologue and epilogue which respectively push and pop
3037         the link register, frame pointer, and callee saved registers if needed.
3038         * cssjit/StackAllocator.h:
3039         (WebCore::StackAllocator::push):
3040         (WebCore::StackAllocator::pop):
3041         Added new vector push and pop functions to use stp and ldb instructions on arm64.
3042
3043 2014-05-23  Jeremy Jones  <jeremyj@apple.com>
3044
3045         Hide fullscreen immediately when switching tabs.
3046         https://bugs.webkit.org/show_bug.cgi?id=133239
3047
3048         Reviewed by Eric Carlson.
3049
3050         This change allows fullscreen to disappear immediately while still cleaning
3051         up fullscreen normally.
3052
3053         * WebCore.exp.in:
3054         * platform/ios/WebVideoFullscreenControllerAVKit.h:
3055         renamed function to requestHideAndExitFullscreen
3056
3057         * platform/ios/WebVideoFullscreenControllerAVKit.mm:
3058         renamed function to requestHideAndExitFullscreen
3059
3060         (-[WebVideoFullscreenController requestHideAndExitFullscreen]):
3061         rename of requestExitFullscreen that also hides.
3062
3063         (-[WebVideoFullscreenController requestExitFullscreen]): Deleted.
3064         * platform/ios/WebVideoFullscreenInterfaceAVKit.h: declare requestHideAndExitFullscreen()
3065         * platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
3066         (WebVideoFullscreenInterfaceAVKit::requestHideAndExitFullscreen):
3067         this hides the window before requesting exitfullscreen from the model.
3068
3069 2014-05-23  Tim Horton  <timothy_horton@apple.com>
3070
3071         [iOS] WKPDFView should have a page indicator
3072         https://bugs.webkit.org/show_bug.cgi?id=133109
3073         <rdar://problem/16383003>
3074
3075         Reviewed by Anders Carlsson.
3076
3077         * English.lproj/Localizable.strings:
3078
3079 2014-05-23  Enrica Casucci  <enrica@apple.com>
3080
3081         Remove invalid ASSERT in wordRangeForPosition.
3082         https://bugs.webkit.org/show_bug.cgi?id=133232
3083
3084         Reviewed by Ryosuke Niwa.
3085
3086         In WK2 for iOS this function is called on every selection change.
3087         The selection could be set by Javascript in an element that is not visible.
3088         In that case we won't be able to create a VisiblePosition therefore wordRangeForPosition
3089         should not ASSERT that the VisiblePosition is not null, but rather
3090         perform an early return if it is null.
3091
3092         * editing/VisibleUnits.cpp:
3093         (WebCore::wordRangeFromPosition):
3094
3095 2014-05-23  Manuel Rego Casasnovas  <rego@igalia.com>
3096
3097         [CSS Grid Layout] Use gridRowCount() and gridColumnCount() when possible
3098         https://bugs.webkit.org/show_bug.cgi?id=133213
3099
3100         Reviewed by Sergio Villar Senin.
3101
3102         Use gridRowCount() and gridColumnCount() instead of m_grid.size() and
3103         m_grid[0].size() to make the code more descriptive and easier to
3104         understand.
3105
3106         No new tests (No change in behavior).
3107
3108         * rendering/RenderGrid.cpp:
3109         (WebCore::RenderGrid::growGrid):
3110         (WebCore::RenderGrid::placeAutoMajorAxisItemOnGrid):
3111
3112 2014-05-23  Alexey Proskuryakov  <ap@apple.com>
3113
3114         REGRESSION (r153917): Chart for the Category Explorer within Quicken Essentials
3115         does not draw until you force a redraw
3116         https://bugs.webkit.org/show_bug.cgi?id=133228
3117         <rdar://problem/15221231>
3118
3119         Reviewed by Simon Fraser.
3120
3121         * platform/RuntimeApplicationChecks.cpp: (WebCore::applicationIsQuickenEssentials):
3122         * platform/RuntimeApplicationChecks.h:
3123         Added a check for Quicken Essentials.
3124
3125         * platform/mac/WidgetMac.mm: (WebCore::Widget::paint): Don't do an early return
3126         for Quicken Essentials, it depends on drawing to update its layer.
3127
3128 2014-05-23  Mihnea Ovidenie  <mihnea@adobe.com>
3129
3130         REGRESSION(r169105): fast/regions/cssom/region-range-for-box-crash.html is more crashy than before
3131         https://bugs.webkit.org/show_bug.cgi?id=133124
3132
3133         Reviewed by David Hyatt.
3134
3135         When adjusting the end points of a subtree selection in the case of a document
3136         with named flows, we have to make sure that we pass the computed end
3137         points to the visible selection algorithm only if the end points are
3138         part of the same subtree, otherwise assume a selection clear behavior.
3139
3140         The test fast/regions/cssom/region-range-for-box-crash.html is used to
3141         test this behavior.
3142
3143         * rendering/SelectionSubtreeRoot.cpp:
3144         (WebCore::SelectionSubtreeRoot::adjustForVisibleSelection):
3145         (WebCore::SelectionSubtreeRoot::selectionStartEndPositions): Moved to header file.
3146         * rendering/SelectionSubtreeRoot.h:
3147         (WebCore::SelectionSubtreeRoot::selectionStartEndPositions):
3148
3149 2014-05-23  Jeremy Jones  <jeremyj@apple.com>
3150
3151         reset m_havePreparedToPlay when changing media engines.
3152         https://bugs.webkit.org/show_bug.cgi?id=133199
3153
3154         Reviewed by Eric Carlson.
3155
3156         If a play request comes in before the media player is created, m_havePreparedToPlay
3157         is set to true. But when the player is created, this prevents the new player from
3158         getting the prepare call.
3159
3160         * html/HTMLMediaElement.cpp:
3161         (WebCore::HTMLMediaElement::mediaPlayerEngineUpdated): set m_havePreparedToPlay false.
3162
3163 2014-05-23  Oliver Hunt  <oliver@apple.com>
3164
3165         Navigator object needs to have properties directly on the instance object
3166         https://bugs.webkit.org/show_bug.cgi?id=133221
3167
3168         Reviewed by Mark Lam.
3169
3170         Flag the Navigator object as requiring properties to be on the instance
3171         as there were a few compatibility issues when on the prototype.
3172
3173         * bindings/scripts/CodeGeneratorJS.pm:
3174         (InterfaceRequiresAttributesOnInstance):
3175
3176 2014-05-23  Simon Fraser  <simon.fraser@apple.com>
3177
3178         Switch CSSGradientValue to use more references
3179         https://bugs.webkit.org/show_bug.cgi?id=133206
3180
3181         Reviewed by Andreas Kling.
3182
3183         Switch from pointers to references in various places.
3184
3185         * css/CSSGradientValue.cpp:
3186         (WebCore::CSSGradientValue::image):
3187         (WebCore::CSSGradientValue::addStops):
3188         (WebCore::positionFromValue):
3189         (WebCore::CSSGradientValue::computeEndPoint):
3190         (WebCore::CSSLinearGradientValue::createGradient):
3191         (WebCore::CSSRadialGradientValue::resolveRadius):
3192         (WebCore::CSSRadialGradientValue::createGradient):
3193         * css/CSSGradientValue.h:
3194
3195 2014-05-23  Zalan Bujtas  <zalan@apple.com>
3196
3197         Subpixel rendering: Make webkit-box-shadow painting subpixel aware.
3198         https://bugs.webkit.org/show_bug.cgi?id=133201
3199         <rdar://problem/16072830>
3200
3201         Reviewed by Simon Fraser.
3202
3203         This patch enables webkit-box-shadow to be painted on a subpixel position. However, we
3204         still truncate -webkit-box-shadow property values. Tracked here: http://webkit.org/b/133202
3205
3206         Test: fast/box-shadow/hidpi-webkit-box-shadow-subpixel-position.html
3207
3208         * platform/graphics/FloatRoundedRect.cpp:
3209         (WebCore::FloatRoundedRect::inflateWithRadii): same as in from RoundedRect.
3210         (WebCore::FloatRoundedRect::adjustRadii): same as in from RoundedRect.
3211         * platform/graphics/FloatRoundedRect.h:
3212         * rendering/RenderBoxModelObject.cpp:
3213         (WebCore::RenderBoxModelObject::paintBoxShadow):
3214
3215 2014-05-22  peavo@outlook.com  <peavo@outlook.com>
3216
3217         [Curl] Crash when exceeding maximum cache limit.
3218         https://bugs.webkit.org/show_bug.cgi?id=133185
3219
3220         Reviewed by Brent Fulgham.
3221
3222         When the maximum cache limit is exceeded, I get a crash.
3223         This happens when deleting cache entries, because a reference to the url string object in the LRU list
3224         is used as a parameter to invalidateCacheEntry(), when called from makeRoomForNewEntry().
3225         When the string is removed from the LRU list in makeRoomForNewEntry(), the string is deleted.
3226         Next, the string is accessed again to remove the url from the index, and we crash.
3227
3228         This can be fixed by removing the string from the LRU list after it is removed from the index.
3229
3230         Fixing the crash also revealed an infinite loop problem.
3231         If the url for some reason only exist in the LRU list, and not in the index,
3232         we will inifitely loop in makeRoomForNewEntry(), trying to remove this url from the cache, but never succeeding.
3233         This can be fixed by removing the url from the LRU list, also when it's not in the index.
3234
3235         * platform/network/curl/CurlCacheManager.cpp:
3236         (WebCore::CurlCacheManager::makeRoomForNewEntry): Avoid infinite loop by checking if there are more cache entries to remove.
3237         (WebCore::CurlCacheManager::invalidateCacheEntry): Avoid crash and infinite loop by removing url from LRU list last.
3238
3239 2014-05-22  Simon Fraser  <simon.fraser@apple.com>
3240
3241         Make viewport units work in CSS gradients
3242         https://bugs.webkit.org/show_bug.cgi?id=133204
3243         <rdar://problem/17012259>
3244
3245         Reviewed by Tim Horton.
3246
3247         Make viewport percentage lengths work in gradients.
3248
3249         Test: fast/gradients/viewport-units-gradient.html
3250
3251         * css/CSSGradientValue.cpp:
3252         (WebCore::CSSGradientValue::addStops):
3253         (WebCore::CSSLinearGradientValue::createGradient):
3254         (WebCore::CSSRadialGradientValue::createGradient):
3255         * css/CSSGradientValue.h:
3256
3257 2014-05-22  Benjamin Poulain  <bpoulain@apple.com>
3258
3259         [iOS][WK2] Add support for minimal-ui viewports
3260         https://bugs.webkit.org/show_bug.cgi?id=133162
3261
3262         Reviewed by Simon Fraser.
3263
3264         Add support for minimal-ui directly into the Viewport configuration.
3265
3266         ViewportConfiguration knows about two kinds of layout size in scrollview coordinates:
3267         -normal size
3268         -minimal ui size.
3269
3270         Initially, the page is initialized with resetMinimalUI() and starts with a normal layout.
3271
3272         At any time before the main resource is loaded, the Web page can update its viewport meta tag
3273         to include minimal-ui. The viewport configuration is updated immediately to reflect the minimal-ui
3274         layout size. Any layout after that takes minimal-ui into account.
3275
3276         * WebCore.exp.in:
3277         * css/LengthFunctions.cpp:
3278         (WebCore::minimumValueForLength):
3279         (WebCore::floatValueForLength):
3280         * css/StyleResolver.cpp:
3281         (WebCore::StyleResolver::viewportPercentageValue):
3282         * dom/ViewportArguments.h:
3283         (WebCore::ViewportArguments::operator==):
3284         * page/FrameView.cpp:
3285         (WebCore::FrameView::setViewportSizeForCSSViewportUnits):
3286         (WebCore::FrameView::viewportSizeForCSSViewportUnits):
3287         (WebCore::FrameView::setViewportSize): Deleted.
3288         (WebCore::FrameView::viewportSize): Deleted.
3289         * page/FrameView.h:
3290         * page/ViewportConfiguration.cpp:
3291         (WebCore::ViewportConfiguration::ViewportConfiguration):
3292         (WebCore::ViewportConfiguration::setMinimumLayoutSize):
3293         (WebCore::ViewportConfiguration::setMinimumLayoutSizeForMinimalUI):
3294         (WebCore::ViewportConfiguration::activeMinimumLayoutSizeInScrollViewCoordinates):
3295         (WebCore::ViewportConfiguration::resetMinimalUI):
3296         (WebCore::ViewportConfiguration::pageWillRenderFirstFrame):
3297         (WebCore::ViewportConfiguration::initialScale):
3298         (WebCore::ViewportConfiguration::minimumScale):
3299         (WebCore::ViewportConfiguration::updateConfiguration):
3300         (WebCore::ViewportConfiguration::layoutWidth):
3301         (WebCore::ViewportConfiguration::layoutHeight):
3302         (WebCore::ViewportConfiguration::description):
3303         * page/ViewportConfiguration.h:
3304         (WebCore::ViewportConfiguration::minimumLayoutSizeForMinimalUI):
3305         (WebCore::ViewportConfiguration::usesMinimalUI):
3306         * rendering/RenderView.cpp:
3307         (WebCore::RenderView::viewportSizeForCSSViewportUnits):
3308         (WebCore::RenderView::viewportSize): Deleted.
3309         * rendering/RenderView.h:
3310
3311 2014-05-22  Ryosuke Niwa  <rniwa@webkit.org>
3312
3313         Can't type in status in facebook.com on iOS Safari because keyboard disappears
3314         https://bugs.webkit.org/show_bug.cgi?id=133196
3315
3316         Reviewed by Geoffrey Garen.
3317
3318         Fixed the bug by rolling out r156252.
3319
3320         When the user taps on the status on facebook.com, the page focuses a textarea,
3321         which is then made momentarily invisible by setting display:none.
3322         The page then removes display:none later and expects the textarea to be still focused.
3323
3324         With r156252, the focus is removed after the page sets display:none and the keyboard disappears.
3325         Since the focus is never reset on the textarea, the user can never type in anything.
3326
3327         The specification may need to change here given that this (rather odd) behavior/expectation exists
3328         on one of the most popular websites on the Web.
3329
3330         * dom/Document.cpp:
3331         (WebCore::Document::Document):
3332         (WebCore::Document::recalcStyle):
3333         (WebCore::Document::updateLayout):
3334         (WebCore::Document::resetHiddenFocusElementSoon): Deleted.
3335         (WebCore::Document::resetHiddenFocusElementTimer): Deleted.
3336         * dom/Document.h:
3337
3338 2014-05-22  Myles C. Maxfield  <mmaxfield@apple.com>
3339
3340         http/tests/security/xss-DENIED-xsl-document-redirect.xml fails with NetworkProcess
3341         https://bugs.webkit.org/show_bug.cgi?id=132523
3342
3343         Reviewed by Alexey Proskuryakov.
3344
3345         We should not manufacture a response in the event of an error during a
3346         synchronous XHR. In addition, this test removes two places that are
3347         sensitive to such a manufactured response.
3348
3349         Updates test expectations.
3350
3351         * loader/DocumentThreadableLoader.cpp:
3352         (WebCore::DocumentThreadableLoader::loadRequest): Don't inspect a
3353         loader response if there is an error
3354         * platform/network/cf/ResourceHandleCFNet.cpp:
3355         (WebCore::ResourceHandle::platformLoadResourceSynchronously): Do not
3356         manufacture a response
3357         * platform/network/mac/ResourceHandleMac.mm:
3358         (WebCore::ResourceHandle::platformLoadResourceSynchronously): Do not
3359         manufacture a response
3360         * xml/XSLTProcessorLibxslt.cpp:
3361         (WebCore::docLoaderFunc): Only use the response's URL if no error
3362         occurred.
3363
3364 2014-05-22  Myles C. Maxfield  <litherum@gmail.com>
3365
3366         Unreviewed typo fix.
3367
3368         Reviewed by NOBODY.
3369
3370         No new tests.
3371
3372         * platform/graphics/mac/FontMac.mm:
3373         (WebCore::Font::dashesForIntersectionsWithRect): Accidentally said run.length() instead of glyphBuffer.size().
3374
3375 2014-05-22  Andreas Kling  <akling@apple.com>
3376
3377         [iOS WebKit2] Web process should try to shrink its memory footprint when going into background.
3378         <https://webkit.org/b/133197>
3379         <rdar://problem/17011561>
3380
3381         Make releaseMemory() public so we can call it from the process-will-suspend callback.
3382
3383         Reviewed by Gavin Barraclough.
3384
3385         * WebCore.exp.in:
3386         * platform/MemoryPressureHandler.h:
3387
3388 2014-05-22  Andreas Kling  <akling@apple.com>
3389
3390         Hook up a setting for showing detailed logging during memory pressure relief.
3391         <https://webkit.org/b/133194>
3392
3393         Make the pressure relief logger opt-in. Also make it flush memory back to the
3394         OS (for both malloc and FastMalloc) to get more accurate numbers at each step.
3395
3396         Reviewed by Gavin Barraclough.
3397
3398         * WebCore.exp.in:
3399         * platform/MemoryPressureHandler.cpp:
3400         * platform/MemoryPressureHandler.h:
3401         (WebCore::MemoryPressureHandler::ReliefLogger::ReliefLogger):
3402         (WebCore::MemoryPressureHandler::ReliefLogger::~ReliefLogger):
3403         (WebCore::MemoryPressureHandler::ReliefLogger::setLoggingEnabled):
3404         * platform/cocoa/MemoryPressureHandlerCocoa.mm:
3405         (WebCore::MemoryPressureHandler::ReliefLogger::platformMemoryUsage):
3406
3407 2014-05-22  Brady Eidson  <beidson@apple.com>
3408
3409         Don't scan for phone numbers in editable regions
3410         <rdar://problem/16949846> and https://bugs.webkit.org/show_bug.cgi?id=133192
3411
3412         Reviewed by