Source/WebCore/ChangeLog

   1 2014-07-14  Tim Horton  <timothy_horton@apple.com>
   2
   3         Fix the build.
   4
   5         * platform/audio/MediaSessionManager.h:
   6
   7 2014-07-14  Eric Carlson  <eric.carlson@apple.com>
   8
   9         [Mac] don't enable low power audio mode on external output devices
  10         https://bugs.webkit.org/show_bug.cgi?id=134877
  11
  12         Reviewed by Sam Weinig.
  13
  14         No new tests, this deals with changes to the audio hardware at runtime.
  15
  16         * WebCore.xcodeproj/project.pbxproj: Remove AudioSessionListener.h.
  17         * WebCore.vcxproj/WebCore.vcxproj: Ditto.
  18
  19         * platform/audio/AudioHardwareListener.cpp:
  20         (WebCore::AudioHardwareListener::AudioHardwareListener): Initialize m_outputDeviceSupportsLowPowerMode
  21             to true on iOS.
  22         * platform/audio/AudioHardwareListener.h:
  23         (WebCore::AudioHardwareListener::outputDeviceSupportsLowPowerMode): New accessor.
  24         (WebCore::AudioHardwareListener::setHardwareActivity): New setter for derived classes.
  25         (WebCore::AudioHardwareListener::setOutputDeviceSupportsLowPowerMode): Ditto.
  26
  27         Remove AudioSessionListener interface, it wasn't being used.
  28         * platform/audio/AudioSession.cpp:
  29         (WebCore::AudioSession::addListener): Deleted.
  30         (WebCore::AudioSession::removeListener): Deleted.
  31         (WebCore::AudioSession::beganAudioInterruption): Deleted.
  32         (WebCore::AudioSession::endedAudioInterruption): Deleted.
  33         * platform/audio/AudioSession.h:
  34         * platform/audio/AudioSessionListener.h: Removed.
  35
  36         * platform/audio/MediaSessionManager.cpp:
  37         (WebCore::MediaSessionManager::addSession): Allocate the AudioHardwareListener if necessary.
  38         (WebCore::MediaSessionManager::removeSession): Free the AudioHardwareListener if necessary.
  39         (WebCore::MediaSessionManager::audioOutputDeviceChanged): AudioHardwareListener client interface
  40             called when the output device changes, call updateSessionState to make sure we are using
  41             the correct buffer size.
  42         * platform/audio/MediaSessionManager.h:
  43
  44         * platform/audio/ios/AudioDestinationIOS.h:
  45         * platform/audio/ios/AudioSessionIOS.mm:
  46         (WebCore::AudioSessionPrivate::AudioSessionPrivate): Drive-by cleanup, remove ObjC helper object
  47             that was used to listen for OS notifications, it is no longer used.
  48         (SOFT_LINK_POINTER): Deleted.
  49         (-[WebAudioSessionHelper initWithCallback:]): Deleted.
  50         (-[WebAudioSessionHelper dealloc]): Deleted.
  51         (-[WebAudioSessionHelper interruption:]): Deleted.
  52
  53         * platform/audio/mac/AudioHardwareListenerMac.cpp:
  54         (WebCore::currentDeviceSupportsLowPowerBufferSize): New, return true only if using build-in
  55             transport device.
  56         (WebCore::processIsRunningPropertyDescriptor): Return reference to static AudioObjectPropertyAddress
  57             for kAudioHardwarePropertyProcessIsRunning instead of declaring one in every method
  58             that needs one.
  59         (WebCore::outputDevicePropertyDescriptor): Return reference to static AudioObjectPropertyAddress
  60             for kAudioHardwarePropertyDefaultOutputDevice.
  61         (WebCore::AudioHardwareListenerMac::AudioHardwareListenerMac): Restructure and add audio object
  62             listener for default output device.
  63         (WebCore::AudioHardwareListenerMac::~AudioHardwareListenerMac): *Remove* listener audio object
  64             property listener instead of *Adding* a new one. Remove new listener.
  65         (WebCore::AudioHardwareListenerMac::propertyChanged): Enumerate the properties that changed,
  66             call appropriate method.
  67         (WebCore::AudioHardwareListenerMac::processIsRunningChanged): Renamed from setHardwareActive,
  68             cleanup.
  69         (WebCore::AudioHardwareListenerMac::outputDeviceChanged): New, call client.audioHardwareOutputDeviceChanged.
  70         (WebCore::AudioHardwareListenerMac::setHardwareActive): Deleted, renamed processIsRunningChanged.
  71         * platform/audio/mac/AudioHardwareListenerMac.h:
  72
  73         * platform/audio/mac/MediaSessionManagerMac.cpp:
  74         (MediaSessionManager::updateSessionState): Only set the output buffer size to 4K when hardware
  75             supports it.
  76
  77 2014-07-13  Benjamin Poulain  <benjamin@webkit.org>
  78
  79         Remove SelectorCheckerFastPath from the style resolution algorithm
  80         https://bugs.webkit.org/show_bug.cgi?id=134866
  81
  82         Reviewed by Antti Koivisto.
  83
  84         SelectorCheckerFastPath is now pure overhead because it can almost never match
  85         if the CSS JIT was unable to compile.
  86
  87         * css/ElementRuleCollector.cpp:
  88         (WebCore::ElementRuleCollector::ruleMatches):
  89         The "pre-filter" behind fastCheckableSelector had two parts:
  90         1) Filtering the pseudoID.
  91         2) Filtering on the rule hash.
  92
  93         The first part has been generalized (RuleDatacanMatchPseudoElement())
  94         and moved to collectMatchingRulesForList().
  95
  96         (WebCore::ElementRuleCollector::collectMatchingRulesForList):
  97         * css/RuleSet.cpp:
  98         (WebCore::selectorCanMatchPseudoElement):
  99         (WebCore::RuleData::RuleData):
 100         (WebCore::RuleSet::addRegionRule):
 101         (WebCore::RuleSet::addRulesFromSheet):
 102         * css/RuleSet.h:
 103         (WebCore::RuleData::canMatchPseudoElement):
 104         (WebCore::RuleData::hasFastCheckableSelector): Deleted.
 105         * css/StyleResolver.cpp:
 106
 107 2014-07-13  Benjamin Poulain  <benjamin@webkit.org>
 108
 109         Remove an useless check from SelectorChecker
 110         https://bugs.webkit.org/show_bug.cgi?id=134868
 111
 112         Reviewed by Darin Adler.
 113
 114         * css/SelectorChecker.cpp:
 115         (WebCore::SelectorChecker::matchRecursively):
 116         The condition of this if() branch can never be met for the mode "QueryingRules".
 117
 118         The next condition in that if() is "dynamicPseudo != NOPSEUDO", which implies
 119         a pseudo element was matched prior to the current context/simple selector.
 120         This cannot happen with QueryingRules, since we never match pseudo elements for
 121         SelectorQuery.
 122
 123 2014-07-12  Nikos Andronikos  <nikos.andronikos-webkit@cisra.canon.com.au>
 124
 125         Elements with rendering disabled due to dimensions should not contribute to parent bounding box
 126         https://bugs.webkit.org/show_bug.cgi?id=134184
 127
 128         Reviewed by Dirk Schulze.
 129
 130         SVG elements that have rendering disabled should not contribute to any ancestor elements bounding box.
 131         Examples of elements with rendering disabled:
 132         - basic shape with width <= 0 or height <= 0
 133         - path with no path data (d attribute missing or empty)
 134         - polyline or polygon element with no point data (points attribute missing or empty)
 135
 136         To achieve this a method (isRenderingDisabled) was added to RenderSVGShape and it's derived classes.
 137         This is used to determine if an element is included when creating the union of child bounding boxes
 138         in a container element.
 139
 140         Tests: svg/custom/GetBBox-path-nodata.html
 141                svg/custom/GetBBox-polygon-nodata.html
 142                svg/custom/GetBBox-polyline-nodata.html
 143                svg/custom/getBBox-container-hiddenchild.html
 144
 145         * rendering/svg/RenderSVGEllipse.cpp:
 146         (WebCore::RenderSVGEllipse::isRenderingDisabled):
 147         New method added. Checks bounding box to determine if rendering is disabled.
 148         * rendering/svg/RenderSVGEllipse.h:
 149         * rendering/svg/RenderSVGPath.cpp:
 150         (WebCore::RenderSVGPath::isRenderingDisabled):
 151         New method added. Checks bounding box to determine if rendering is disabled.
 152         * rendering/svg/RenderSVGPath.h:
 153         * rendering/svg/RenderSVGRect.cpp:
 154         (WebCore::RenderSVGRect::isRenderingDisabled):
 155         New method added. Checks bounding box to determine if rendering is disabled.
 156         * rendering/svg/RenderSVGRect.h:
 157         * rendering/svg/RenderSVGShape.h:
 158         (WebCore::RenderSVGShape::isRenderingDisabled):
 159         New method added. Always returns false so that derived classes that do not
 160         implement this method retain the existing behaviour.
 161         * rendering/svg/SVGRenderSupport.cpp:
 162         (WebCore::SVGRenderSupport::computeContainerBoundingBoxes):
 163         For each element potentially being included in the unioned bounding box of
 164         a container, check isRenderingDisabled and skip that element if true.
 165
 166         * rendering/svg/RenderSVGEllipse.cpp:
 167         (WebCore::RenderSVGEllipse::isRenderingDisabled):
 168         * rendering/svg/RenderSVGEllipse.h:
 169         * rendering/svg/RenderSVGPath.cpp:
 170         (WebCore::RenderSVGPath::isRenderingDisabled):
 171         * rendering/svg/RenderSVGPath.h:
 172         * rendering/svg/RenderSVGRect.cpp:
 173         (WebCore::RenderSVGRect::isRenderingDisabled):
 174         * rendering/svg/RenderSVGRect.h:
 175         * rendering/svg/RenderSVGShape.h:
 176         * rendering/svg/SVGRenderSupport.cpp:
 177         (WebCore::SVGRenderSupport::computeContainerBoundingBoxes):
 178
 179 2014-07-12  Benjamin Poulain  <benjamin@webkit.org>
 180
 181         Rename selectorListContainsUncommonAttributeSelector() to selectorListContainsAttributeSelector()
 182         https://bugs.webkit.org/show_bug.cgi?id=134862
 183
 184         Reviewed by Sam Weinig.
 185
 186         Unlike containsUncommonAttributeSelector(), selectorListContainsUncommonAttributeSelector() does not
 187         evaluate the attribute for "uncommon" types.
 188
 189         It would be possible to change the function instead to evaluate common attributes based
 190         on the match type and the pseudo class type. Such change would be more risky
 191         and we would get very little benefit from it, I leave that for later if that ever becomes useful.
 192
 193         * css/RuleSet.cpp:
 194         (WebCore::selectorListContainsAttributeSelector):
 195         (WebCore::containsUncommonAttributeSelector):
 196         (WebCore::selectorListContainsUncommonAttributeSelector): Deleted.
 197
 198 2014-07-12  Zan Dobersek  <zdobersek@igalia.com>
 199
 200         Use braced-init-list to create one-item Vector object in insertPerformanceEntry
 201         https://bugs.webkit.org/show_bug.cgi?id=133675
 202
 203         Reviewed by Darin Adler.
 204
 205         * page/PerformanceUserTiming.cpp:
 206         (WebCore::insertPerformanceEntry): Avoid wasting three lines for creating a
 207         Vector object with just one item that in the end gets copied. Use the
 208         braced-init-list syntax instead, leveraging Vector's std::initializer_list
 209         constructor.
 210
 211 2014-07-12  Allan Sandfeld Jensen  <allan.jensen@digia.com>
 212
 213         Memory leaks with autoLoadImages off
 214         https://bugs.webkit.org/show_bug.cgi?id=124411
 215
 216         Reviewed by Darin Adler.
 217
 218         Do not emit notifyFinished for images with deferred load,
 219         and allow deferred loads to be cancelled.
 220
 221         * loader/cache/CachedResource.cpp:
 222         (WebCore::CachedResource::checkNotify):
 223         (WebCore::CachedResource::cancelLoad):
 224
 225 2014-07-11  Jer Noble  <jer.noble@apple.com>
 226
 227         [MSE] http/tests/media/media-source/mediasource-duration.html is failing.
 228         https://bugs.webkit.org/show_bug.cgi?id=134852
 229
 230         Reviewed by Eric Carlson.
 231
 232         Fixes the following tests:
 233         http/tests/media/media-source/mediasource-config-change-mp4-a-bitrate.html
 234         http/tests/media/media-source/mediasource-config-change-mp4-av-audio-bitrate.html
 235         http/tests/media/media-source/mediasource-config-change-mp4-av-video-bitrate.html
 236         http/tests/media/media-source/mediasource-config-change-mp4-v-bitrate.html
 237         http/tests/media/media-source/mediasource-config-change-mp4-v-framerate.html
 238         http/tests/media/media-source/mediasource-duration.html
 239         http/tests/media/media-source/mediasource-play.html
 240
 241         The primary change necessary to fix the mediasource-duration.html test was to add support
 242         for delaying the completion of a seek operation until the HTMLMediaElement's readyState
 243         rises to > HAVE_CURRENT_DATA. This is accomplished by modifying MediaSourcePrivate to have
 244         waitForSeekCompleted() and seekCompleted() virtual methods. These are called by MediaSource
 245         when a seek operation results in the current time moving outside the currently buffered time
 246         ranges, and when an append operation results in the readyState changing, respectively.
 247
 248         A number of other drive-by fixes were necessary to get this test fully passing, as noted
 249         below.
 250
 251         Make the MediaSource the primary owner of the media's duration, rather than the MediaSourcePrivate.
 252         Move the MediaSourcePrivateClient pointer to the MediaSourcePrivate from the MediaPlayerPrivate, so
 253         the MediaSource's duration can be retrieved.  While we're at it, do the same thing for buffered.
 254
 255         * Modules/mediasource/MediaSource.cpp:
 256         (WebCore::MediaSource::MediaSource): Initialize m_duration.
 257         (WebCore::MediaSource::duration): Simple accessor.
 258         (WebCore::MediaSource::setDurationInternal): Bring 'duration change algorithm' up to spec.
 259         (WebCore::MediaSource::setReadyState): Reset m_duration on close.
 260         * Modules/mediasource/MediaSource.h:
 261         * platform/graphics/MediaSourcePrivate.h:
 262         * platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.mm:
 263         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::load): Do not call setPrivateAndOpen().
 264         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::durationDouble): Pass through to MediaSourcePrivateAVFObjC.
 265         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::buffered): Ditto.
 266         * platform/graphics/avfoundation/objc/MediaSourcePrivateAVFObjC.h:
 267         * platform/graphics/avfoundation/objc/MediaSourcePrivateAVFObjC.mm:
 268         (WebCore::MediaSourcePrivateAVFObjC::create): Call setPrivateAndOpen().
 269         (WebCore::MediaSourcePrivateAVFObjC::MediaSourcePrivateAVFObjC): Set m_client.
 270         (WebCore::MediaSourcePrivateAVFObjC::duration): Pass through to MediaSourcePrivateClient.
 271         (WebCore::MediaSourcePrivateAVFObjC::buffered): Ditto.
 272         (WebCore::MediaSourcePrivateAVFObjC::durationChanged): Pass through to MediaPlayerPrivateMediaSourceAVFObjC.
 273         (WebCore::MediaSourcePrivateAVFObjC::setDuration): Deleted.
 274         * platform/graphics/gstreamer/MediaSourceGStreamer.cpp:
 275         (WebCore::MediaSourceGStreamer::open): Pass in MediaSourcePrivateClient.
 276         (WebCore::MediaSourceGStreamer::MediaSourceGStreamer): Initialize m_mediaSource.
 277         (WebCore::MediaSourceGStreamer::durationChanged): Retrieve the duration from MediaSourcePrivateClient.
 278         (WebCore::MediaSourceGStreamer::markEndOfStream): Remove unnecssary ASSERT.
 279         (WebCore::MediaSourceGStreamer::unmarkEndOfStream): Ditto.
 280         (WebCore::MediaSourceGStreamer::setDuration): Deleted.
 281         * platform/graphics/gstreamer/MediaSourceGStreamer.h:
 282         * platform/mock/mediasource/MockMediaPlayerMediaSource.cpp:
 283         (WebCore::MockMediaPlayerMediaSource::load): Do not call setPrivateAndOpen().
 284         (WebCore::MockMediaPlayerMediaSource::buffered): Pass through to MockMediaSourcePrivate.
 285         (WebCore::MockMediaPlayerMediaSource::durationDouble): Ditto.
 286         (WebCore::MockMediaPlayerMediaSource::advanceCurrentTime): Ditto.
 287         * platform/mock/mediasource/MockMediaSourcePrivate.cpp:
 288         (WebCore::MockMediaSourcePrivate::create): Call setPrivateAndOpen().
 289         (WebCore::MockMediaSourcePrivate::MockMediaSourcePrivate): Set m_client.
 290         (WebCore::MockMediaSourcePrivate::duration): Pass through to MediaSourcePrivateClient.
 291         (WebCore::MockMediaSourcePrivate::buffered): Ditto.
 292         (WebCore::MockMediaSourcePrivate::durationChanged): Pass thorugh to MockMediaPlayerMediaSource.
 293         (WebCore::MockMediaSourcePrivate::setDuration): Deleted.
 294
 295         Route seekToTime through MediaSource, rather than through MediaSourcePrivate, so that
 296         the time can be compared against the buffered ranges, and trigger the delay of the seek
 297         operation if necessary. Add a seekTimer to MediaPlayerPrivateMediaSourceAVFObjC, as this
 298         guarantees the order of asynchronous operations, rather than callOnMainThread, which can
 299         cause async operations to occur out of order.
 300
 301         * Modules/mediasource/MediaSource.cpp:
 302         (WebCore::MediaSource::seekToTime): Bring up to spec.
 303         (WebCore::MediaSource::completeSeek): Ditto.
 304         (WebCore::MediaSource::monitorSourceBuffers): Call completeSeek() when appropriate.
 305         * Modules/mediasource/SourceBuffer.cpp:
 306         (WebCore::SourceBuffer::sourceBufferPrivateSeekToTime): Deleted.
 307         (WebCore::SourceBuffer::seekToTime): Renamed from sourceBufferPrivateSeekToTime().
 308         * platform/graphics/MediaSourcePrivate.h:
 309         * platform/graphics/MediaSourcePrivateClient.h:
 310         * platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.h:
 311         * platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.mm:
 312         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::MediaPlayerPrivateMediaSourceAVFObjC): Add seekTimer. Only
 313             call timeChanged() if no longer seeking, thereby triggering a 'seeked' event.
 314         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::~MediaPlayerPrivateMediaSourceAVFObjC): Clear m_seekTimer.
 315         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::seekWithTolerance): Use m_seekTimer.
 316         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::seekTimerFired): Call seekInternal.
 317         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::seekInternal): Add logging.
 318         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::waitForSeekCompleted): Added.
 319         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::seekCompleted): Added; trigger 'seeked'.
 320         (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::setReadyState): No longer attempt to finish seek when
 321             readyState changes here; this has been moved up to MediaSource.cpp.
 322         * platform/graphics/avfoundation/objc/MediaSourcePrivateAVFObjC.h:
 323         * platform/graphics/avfoundation/objc/MediaSourcePrivateAVFObjC.mm:
 324         (WebCore::MediaSourcePrivateAVFObjC::waitForSeekCompleted): Pass through to MediaPlayerPrivateMediaSourceAVFObjC.
 325         (WebCore::MediaSourcePrivateAVFObjC::seekCompleted): Ditto.
 326         (WebCore::MediaSourcePrivateAVFObjC::seekToTime): Pass through to MediaSourcePrivateClient.
 327         (WebCore::MediaSourcePrivateAVFObjC::fastSeekTimeForMediaTime): Ditto.
 328         * platform/mock/mediasource/MockMediaPlayerMediaSource.cpp:
 329         (WebCore::MockMediaPlayerMediaSource::MockMediaPlayerMediaSource): Initialize m_seekCompleted.
 330         (WebCore::MockMediaPlayerMediaSource::seeking): Check for an uncompleted seek operation.
 331         (WebCore::MockMediaPlayerMediaSource::seekWithTolerance): Ditto.
 332         (WebCore::MockMediaPlayerMediaSource::waitForSeekCompleted): Added.
 333         (WebCore::MockMediaPlayerMediaSource::seekCompleted): Added; trigger 'seeked'.
 334         * platform/mock/mediasource/MockMediaPlayerMediaSource.h:
 335         * platform/mock/mediasource/MockMediaSourcePrivate.cpp:
 336         (WebCore::MockMediaSourcePrivate::waitForSeekCompleted): Pass through to MockMediaPlayerMediaSource.
 337         (WebCore::MockMediaSourcePrivate::seekCompleted): Ditto.
 338         * platform/mock/mediasource/MockMediaSourcePrivate.h:
 339
 340         Drive-by fixes.
 341
 342         * Modules/mediasource/MediaSource.cpp:
 343         (WebCore::MediaSource::streamEndedWithError): Re-order the steps in streamEndedWithError()
 344             to avoid the MediaSource being closed and re-opened by the resulting duration change
 345             operation.
 346         * Modules/mediasource/MediaSource.h:
 347         * Modules/mediasource/SourceBuffer.cpp:
 348         (WebCore::SourceBuffer::remove): Added logging.
 349         (WebCore::SourceBuffer::removeCodedFrames): Ditto.
 350         (WebCore::SourceBuffer::hasFutureTime): Swap an ASSERT for an early-return; it's possible
 351             for currentTime() to be outside of a buffered area.
 352         * Modules/mediasource/SourceBuffer.h:
 353         * html/HTMLMediaElement.cpp:
 354         (WebCore::HTMLMediaElement::parseAttribute): Do not issue an additional 'timeupdate' event
 355             after finishSeek() issues one of its own.
 356         * platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:
 357         (WebCore::globalDataParserQueue): Allow parsing operations to happen concurrently on
 358             background queues.
 359
 360 2014-07-12  Eric Carlson  <eric.carlson@apple.com>
 361
 362         [iOS] update control type when playback state changes
 363         https://bugs.webkit.org/show_bug.cgi?id=134856
 364
 365         Reviewed by Dean Jackson.
 366
 367         Check to make sure the correct type of media controls are showing when playback state
 368         changes so we don't end up with the wrong type of controls during playback.
 369
 370         * Modules/mediacontrols/mediaControlsApple.js:
 371         (Controller.prototype.HandledVideoEvents): Drive-by fix, call handlePlay on 'playing' instead
 372             of 'play' so controls don't hide too early.
 373         * Modules/mediacontrols/mediaControlsiOS.js:
 374         (ControllerIOS.prototype.setPlaying): New, call updateControls to make sure inline
 375             controls are shown when appropriate.
 376
 377 2014-07-12  Dean Jackson  <dino@apple.com>
 378
 379         [iOS Media] Start playback button should indicate when it can't play
 380         https://bugs.webkit.org/show_bug.cgi?id=134851
 381
 382         Post-commit review with Eric Carlson on IRC.
 383
 384         * Modules/mediacontrols/mediaControlsiOS.js:
 385         (ControllerIOS.prototype.updateStatusDisplay): Remove the line that sets the
 386         class on the inline play button.
 387
 388 2014-07-11  Dean Jackson  <dino@apple.com>
 389
 390         [iOS Media] Start playback button should indicate when it can't play
 391         https://bugs.webkit.org/show_bug.cgi?id=134851
 392
 393         Reviewed by Sam Weinig.
 394
 395         If a video element has errors and cannot play, then the start
 396         playback button should use the "broken" icon.
 397
 398         * Modules/mediacontrols/mediaControlsApple.js: Add a new "failed" class.
 399         * Modules/mediacontrols/mediaControlsiOS.js:
 400         (ControllerIOS.prototype.shouldHaveStartPlaybackButton): We need a playback button
 401         even when there is an error.
 402         (ControllerIOS.prototype.handleWirelessPickerButtonTouchStart): Don't offer Airplay
 403         if we're in an error state.
 404         (ControllerIOS.prototype.updateStatusDisplay): Set the "failed" class on
 405         the playback buttons if we are in an error state.
 406
 407 2014-07-11  Benjamin Poulain  <benjamin@webkit.org>
 408
 409         Partition the CSS rules based on the most specific filter of the rightmost fragment
 410         https://bugs.webkit.org/show_bug.cgi?id=134828
 411
 412         Reviewed by Andreas Kling.
 413
 414         Previously, RuleSet was partitioning each rule based on the rightmost filter.
 415         While fast, this had the side effect of putting many selectors with ID match in the class
 416         bucket (because the selectors are generally written starting with the ID).
 417
 418         This patch replace the code of findBestRuleSetAndAdd() by a simple loop going over all
 419         the simple selectors in the rightmost fragment to find the best bucket.
 420
 421         * css/ElementRuleCollector.cpp:
 422         (WebCore::ElementRuleCollector::ruleMatches):
 423         * css/RuleSet.cpp:
 424         (WebCore::isSelectorMatchingHTMLBasedOnRuleHash):
 425         I unified ruleData.hasRightmostSelectorMatchingHTMLBasedOnRuleHash() and hasMultipartSelector().
 426
 427         (WebCore::RuleData::RuleData):
 428         (WebCore::rulesCountForName):
 429         (WebCore::RuleSet::addRule):
 430         I removed the recursive part of findBestRuleSetAndAdd() (which was wrong anyway). The function
 431         was useless so I just moved the algorithm to addRule() directly.
 432
 433         We first loop over all the CSSSelectors related by SubSelector, this correspond to the rightmost fragment.
 434         If a filter with high specificity is found, we add the rule immediately and end there.
 435         If a filter that is not very specific is found, we keep a pointer to the selector to use it later.
 436
 437         (WebCore::RuleSet::findBestRuleSetAndAdd): Deleted.
 438         * css/RuleSet.h:
 439         (WebCore::RuleData::hasMultipartSelector): Deleted.
 440
 441 2014-07-11  Alex Christensen  <achristensen@webkit.org>
 442
 443         [WinCairo] Unreviewed build fix after r170937.
 444
 445         * WebCore.vcxproj/WebCoreGeneratedWinCairo.make:
 446         Update and apply changes applied to WebCoreGenerated.make in r170937.
 447
 448 2014-07-11  Joseph Pecoraro  <pecoraro@apple.com>
 449
 450         Web Inspector: Crash when using a stale InspectableNode Node
 451         https://bugs.webkit.org/show_bug.cgi?id=134849
 452
 453         Reviewed by Timothy Hatcher.
 454
 455         * inspector/PageConsoleAgent.cpp:
 456
 457 2014-07-11  Jer Noble  <jer.noble@apple.com>
 458
 459         [MSE] Separate MediaSource logging into its own log channel
 460         https://bugs.webkit.org/show_bug.cgi?id=134809
 461
 462         Reviewed by Eric Carlson.
 463
 464         Separate out MediaSource logging (which can be quite verbose) into its own log channel,
 465         distinct from the Media log channel. Add some per-sample logging and further sequester that
 466         logging into its own MediaSourceSamples log channel.
 467
 468         Add the MediaSource and MediaSourceSamples channels:
 469         * platform/Logging.h:
 470
 471         Move LOG(Media) -> LOG(MediaSource):
 472         * Modules/mediasource/MediaSource.cpp:
 473         (WebCore::MediaSource::MediaSource):
 474         (WebCore::MediaSource::~MediaSource):
 475         (WebCore::MediaSource::setReadyState):
 476         (WebCore::MediaSource::addSourceBuffer):
 477         (WebCore::MediaSource::removeSourceBuffer):
 478         (WebCore::MediaSource::isTypeSupported):
 479         * Modules/mediasource/SourceBuffer.cpp:
 480         (WebCore::SourceBuffer::sourceBufferPrivateSeekToTime):
 481         (WebCore::SourceBuffer::sourceBufferPrivateAppendComplete):
 482         (WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample):
 483         (WebCore::SourceBuffer::sourceBufferPrivateDidBecomeReadyForMoreSamples):
 484         (WebCore::SourceBuffer::provideMediaData):
 485         (WebCore::SourceBuffer::monitorBufferingRate):
 486         (WebCore::SourceBuffer::textTrackAddCues):
 487         (WebCore::SourceBuffer::hasFutureTime):
 488         * platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:
 489         (WebCore::SourceBufferPrivateAVFObjC::didParseStreamDataAsAsset):
 490         (WebCore::SourceBufferPrivateAVFObjC::didFailToParseStreamDataWithError):
 491         (WebCore::SourceBufferPrivateAVFObjC::processCodedFrame):
 492         (WebCore::SourceBufferPrivateAVFObjC::didProvideContentKeyRequestInitializationDataForTrackID):
 493         (WebCore::SourceBufferPrivateAVFObjC::append):
 494         (WebCore::SourceBufferPrivateAVFObjC::layerDidReceiveError):
 495         (WebCore::SourceBufferPrivateAVFObjC::rendererDidReceiveError):
 496         (WebCore::SourceBufferPrivateAVFObjC::flushAndEnqueueNonDisplayingSamples):
 497         (WebCore::SourceBufferPrivateAVFObjC::enqueueSample):
 498
 499 2014-07-11  Brady Eidson  <beidson@apple.com>
 500
 501         Phone numbers that span two lines are not detected.
 502         <rdar://problem/17601146> and https://bugs.webkit.org/show_bug.cgi?id=134808
 503
 504         Reviewed by Tim Horton.
 505
 506         * editing/Editor.cpp:
 507         (WebCore::Editor::scanSelectionForTelephoneNumbers): After scanning a range from the TextIterator,
 508             create an "edge range" window around the end of the TextIterator range, and scan it.
 509             Also make sure to not accumulate duplicate ranges that might have showed up in both the
 510             TextIterator range and the edge window range.
 511
 512 2014-07-11  Enrica Casucci  <enrica@apple.com>
 513
 514         Implement textStylingAtPosition in WK2.
 515         https://bugs.webkit.org/show_bug.cgi?id=134843
 516         <rdar://problem/17614981>
 517
 518         Reviewed by Benjamin Poulain.
 519
 520         Adding some exports and making styleForSelectionStart public.
 521
 522         * WebCore.exp.in:
 523         * editing/Editor.h:
 524
 525 2014-07-11  Zalan Bujtas  <zalan@apple.com>
 526
 527         REGRESSION (r168868): eBay 'see all' links fail due to different JS bindings conversion behavior.
 528         https://bugs.webkit.org/show_bug.cgi?id=134841
 529         <rdar://problem/17577563>
 530
 531         Reviewed by Simon Fraser.
 532
 533         Revert back to type int for Element.scroll* APIs in order to preserve exception handling
 534         behavior.
 535
 536         Test: cssom/non-subpixel-scroll-top-left-values.html
 537
 538         * dom/Element.cpp:
 539         (WebCore::Element::scrollLeft):
 540         (WebCore::Element::scrollTop):
 541         (WebCore::Element::setScrollLeft):
 542         (WebCore::Element::setScrollTop):
 543         (WebCore::Element::scrollWidth):
 544         (WebCore::Element::scrollHeight):
 545         * dom/Element.h:
 546         * dom/Element.idl:
 547         * html/HTMLBodyElement.cpp:
 548         (WebCore::adjustForZoom):
 549         (WebCore::HTMLBodyElement::scrollLeft):
 550         (WebCore::HTMLBodyElement::setScrollLeft):
 551         (WebCore::HTMLBodyElement::scrollTop):
 552         (WebCore::HTMLBodyElement::setScrollTop):
 553         (WebCore::HTMLBodyElement::scrollHeight):
 554         (WebCore::HTMLBodyElement::scrollWidth):
 555         * html/HTMLBodyElement.h:
 556
 557 2014-07-11  Joseph Pecoraro  <pecoraro@apple.com>
 558
 559         Web Inspector: Debugger Pause button does not work
 560         https://bugs.webkit.org/show_bug.cgi?id=134785
 561
 562         Reviewed by Timothy Hatcher.
 563
 564         * CMakeLists.txt:
 565         * DerivedSources.make:
 566         * inspector/CommandLineAPIModuleSource.js:
 567         Minification strips the sourceURL command. Add it back with minification.
 568
 569 2014-07-11  Andreas Kling  <akling@apple.com>
 570
 571         Use the bare minimum tile coverage rect when under memory pressure.
 572         <https://webkit.org/b/134837>
 573
 574         When the browser is under critical memory pressure, don't generate any
 575         more tiles than are needed to cover the exposed viewport rect.
 576
 577         Reviewed by Pratik Solanki.
 578
 579         * page/FrameView.cpp:
 580         (WebCore::FrameView::computeCoverageRect):
 581
 582 2014-07-11  Javier Fernandez  <jfernandez@igalia.com>
 583
 584         [CSS Grid Layout] Implement justify-self css property
 585         https://bugs.webkit.org/show_bug.cgi?id=134419
 586
 587         Reviewed by Dean Jackson.
 588
 589         This change adds the justify-self property from CSS 3 Box Alignment
 590         and implements the parsing.
 591
 592         From Blink r164685 by <jchaffraix@chromium.org>
 593
 594         Test: fast/css/parse-justify-self.html
 595
 596         * css/CSSComputedStyleDeclaration.cpp:
 597         (WebCore::ComputedStyleExtractor::propertyValue):
 598         * css/CSSParser.cpp:
 599         (WebCore::isValidKeywordPropertyAndValue):
 600         (WebCore::CSSParser::parseValue):
 601         (WebCore::isItemPositionKeyword):
 602         (WebCore::CSSParser::parseJustifySelf):
 603         * css/CSSParser.h:
 604         * css/CSSPrimitiveValueMappings.h:
 605         (WebCore::CSSPrimitiveValue::CSSPrimitiveValue):
 606         (WebCore::CSSPrimitiveValue::operator EJustifySelf):
 607         (WebCore::CSSPrimitiveValue::operator EJustifySelfOverflowAlignment):
 608         * css/CSSPropertyNames.in:
 609         * css/CSSValueKeywords.in:
 610         * css/DeprecatedStyleBuilder.cpp:
 611         (WebCore::DeprecatedStyleBuilder::DeprecatedStyleBuilder):
 612         * css/StyleResolver.cpp:
 613         (WebCore::StyleResolver::applyProperty):
 614         * rendering/style/RenderStyle.h:
 615         * rendering/style/RenderStyleConstants.h:
 616         * rendering/style/StyleRareNonInheritedData.cpp:
 617         (WebCore::StyleRareNonInheritedData::StyleRareNonInheritedData):
 618         (WebCore::StyleRareNonInheritedData::operator==):
 619         * rendering/style/StyleRareNonInheritedData.h:
 620
 621 2014-07-11  Bear Travis  <betravis@adobe.com>
 622
 623         [Feature Queries] Feature Query CSS Grammar Productions Should Return a Value
 624         https://bugs.webkit.org/show_bug.cgi?id=134810
 625
 626         Reviewed by Antti Koivisto.
 627
 628         The parsing return values are used to determine if a feature query is valid. Two
 629         grammar productions have been updated to correctly return a value.
 630
 631         Tests exist for this feature but will not be run until the feature is turned on.
 632         See https://bugs.webkit.org/show_bug.cgi?id=134404.
 633
 634         * css/CSSGrammar.y.in: Adding return values for two valid feature query productions.
 635
 636 2014-07-10  Brent Fulgham  <bfulgham@apple.com>
 637
 638         Use a separate backdrop element to allow cues to have highlight and background color
 639         https://bugs.webkit.org/show_bug.cgi?id=134821
 640         <rdar://problem/15999721>
 641
 642         Reviewed by Eric Carlson.
 643
 644         Add a new <div> element wrapping the existing cue <span>. This allows
 645         us to have a highlight on the cue (in the <span> background), as well
 646         as an overall background color.
 647
 648         * Modules/mediacontrols/mediaControlsApple.css:
 649         (video::-webkit-media-text-track-display-backdrop): New markup for
 650         the backdrop element of the caption.
 651         * html/track/VTTCue.cpp:
 652         (WebCore::VTTCue::cueBackdropShadowPseudoId): Added to
 653         allow user customization of the cue backdrop.
 654         (WebCore::VTTCue::initialize): Rename the old "m_cueBackgroundBox" to
 655         "m_cueHighlightBox" and add a new "m_cueBackdropBox" member.
 656         (WebCore::VTTCue::updateDisplayTree): Update for m_cueHighlightBox.
 657         (WebCore::VTTCue::getDisplayTree): Make m_cueHighlightBox a child
 658         of the new m_cueBackdropBox element, and add m_cueBackdropBox to
 659         the display tree.
 660         * html/track/VTTCue.h:
 661         (WebCore::VTTCue::element):
 662         * page/CaptionUserPreferencesMediaAF.cpp:
 663         (WebCore::CaptionUserPreferencesMediaAF::setInterestedInCaptionPreferenceChanges):
 664         Fix for missing caption style updates. Even if we are already
 665         listening for caption changes, we still want to update the new
 666         instance's style sheet to match.
 667         (WebCore::CaptionUserPreferencesMediaAF::captionsStyleSheetOverride):
 668         * rendering/RenderVTTCue.cpp:
 669         (WebCore::RenderVTTCue::initializeLayoutParameters): Take the new
 670         <div> into consideration when looking for the Cue text element.
 671
 672 2014-07-11  Zalan Bujtas  <zalan@apple.com>
 673
 674         Subpixel layout: return integral results for offset*, client*, scroll* by default.
 675         https://bugs.webkit.org/show_bug.cgi?id=134651
 676
 677         Reviewed by Simon Fraser.
 678
 679         Revert to returning integral values for Element.offset* client* scroll* by default.
 680         Fractional values break number of sites(tight design) and JS frameworks(fail to handle fractional values).
 681
 682         Since snapped dimension depends on both the original point and the width/height of the box,
 683         we need to call RenderBoxModelObject::pixelSnapped*() helpers, instead of round().
 684
 685         Covered by existing tests
 686
 687         * dom/Element.cpp:
 688         (WebCore::subpixelMetricsEnabled):
 689         (WebCore::convertToNonSubpixelValueIfNeeded):
 690         (WebCore::Element::offsetLeft):
 691         (WebCore::Element::offsetTop):
 692         (WebCore::Element::offsetWidth):
 693         (WebCore::Element::offsetHeight):
 694         (WebCore::Element::clientLeft):
 695         (WebCore::Element::clientTop):
 696         (WebCore::Element::clientWidth):
 697         (WebCore::Element::clientHeight):
 698         * page/Settings.in:
 699
 700 2014-07-11  Zalan Bujtas  <zalan@apple.com>
 701
 702         Subpixel rendering: icloud.com password arrow has clipped circle at some window sizes.
 703         https://bugs.webkit.org/show_bug.cgi?id=134824
 704         <rdar://problem/17612729>
 705
 706         Reviewed by Simon Fraser.
 707
 708         Transparent layer clipping needs device pixel offset adjusting.
 709
 710         Test: fast/layers/wrong-clipping-semi-transparent-compositing-layer-on-subpixel-position.html
 711
 712         * rendering/RenderLayer.cpp:
 713         (WebCore::RenderLayer::beginTransparencyLayers):
 714         (WebCore::RenderLayer::paintLayer):
 715         (WebCore::RenderLayer::paintLayerContents):
 716         (WebCore::RenderLayer::paintBackgroundForFragments):
 717         (WebCore::RenderLayer::paintForegroundForFragments):
 718         * rendering/RenderLayer.h:
 719
 720 2014-07-11  Carlos Garcia Campos  <cgarcia@igalia.com>
 721
 722         Unreviewed. Remove unused file.
 723
 724         * html/shadow/MediaControlsGtk.h: Removed.
 725
 726 2014-07-11  Manuel Rego Casasnovas  <rego@igalia.com>
 727
 728         [CSS Grid Layout] Update grid-auto-flow to the new syntax
 729         https://bugs.webkit.org/show_bug.cgi?id=134057
 730
 731         Reviewed by Sergio Villar Senin.
 732
 733         In last versions of the spec grid-auto-flow syntax has changed. New
 734         syntax is:
 735         [ row | column ] && dense? | stack && [ row | column ]?
 736
 737         Implemented parsing for new syntax and added/modified test cases in
 738         current layout tests.
 739
 740         For the moment, the implementation keeps working in the same way, but
 741         using "stack" value for grid-auto-flow property instead of "none". This
 742         should be fixed in a follow-up patch once "stack" is properly
 743         implemented.
 744
 745         Also "dense" needs to be reviewed. Right now auto-placement algorithm is
 746         always "dense" and never "sparse".
 747
 748         No new tests, update current tests to add new cases.
 749
 750         * css/CSSComputedStyleDeclaration.cpp:
 751         (WebCore::ComputedStyleExtractor::propertyValue): Adapt to new syntax.
 752         * css/CSSParser.cpp:
 753         (WebCore::isValidKeywordPropertyAndValue): Removed grid-auto-flow as it
 754         is not a keyword anymore.
 755         (WebCore::isKeywordPropertyID): Ditto.
 756         (WebCore::CSSParser::parseValue): Add specific method for parsing
 757         grid-auto-flow.
 758         (WebCore::CSSParser::parseGridShorthand): Adapt parsing for
 759         grid-auto-flow property inside the shorthand.
 760         (WebCore::isValidGridAutoFlowId): Helper method to check if the CSSValue
 761         id is a valid keyword for grid-auto-flow property.
 762         (WebCore::CSSParser::parseGridAutoFlow): Implement new parsing for
 763         grid-auto-flow.
 764         * css/CSSParser.h: Method header for grid-auto-flow parsing.
 765         * css/CSSPrimitiveValueMappings.h:
 766         (WebCore::CSSPrimitiveValue::operator GridAutoFlow): Deleted.
 767         (WebCore::CSSPrimitiveValue::CSSPrimitiveValue): Deleted.
 768         * css/CSSValueKeywords.in: Add new keywords required by grid-auto-flow:
 769         "dense" and "stack".
 770         * css/DeprecatedStyleBuilder.cpp:
 771         (WebCore::DeprecatedStyleBuilder::DeprecatedStyleBuilder): Remove
 772         default handler as grid-auto-flow is not a keyword now.
 773         * css/StyleResolver.cpp:
 774         (WebCore::StyleResolver::applyProperty): Implement particular handler
 775         for new grid-auto-flow syntax.
 776         * rendering/RenderGrid.cpp:
 777         (WebCore::RenderGrid::placeItemsOnGrid): Keep old "none" behavior as
 778         "stack" behavior for the moment.
 779         (WebCore::RenderGrid::autoPlacementMajorAxisDirection): Use the new
 780         helper methods in RenderStyle.
 781         (WebCore::RenderGrid::autoPlacementMinorAxisDirection): Ditto.
 782         * rendering/style/GridResolvedPosition.cpp:
 783         (WebCore::GridResolvedPosition::resolveGridPositionsFromStyle): Again
 784         keep old "none" behavior for "stack".
 785         * rendering/style/RenderStyle.h: Add new helper methods to know the
 786         direction and algorithm of grid-auto-flow property.
 787         * rendering/style/RenderStyleConstants.h: Redefine GridAutoFlow enum
 788         using flags for algorithm and direction.
 789         * rendering/style/StyleGridData.h: Change m_gridAutoFlow type to
 790         unsigned.
 791
 792 2014-07-10  Myles C. Maxfield  <mmaxfield@apple.com>
 793
 794         [iOS] Pasting rich content does not perform a two-step paste
 795         https://bugs.webkit.org/show_bug.cgi?id=134795
 796
 797         Reviewed by Benjamin Poulain.
 798
 799         UIKit requires writing something (anything!) to a particular pasteboard
 800         during copy in order to flag the content as rich. When a paste occurs,
 801         UIKit will read this flag and perform a two-step paste.
 802
 803         Test: platform/ios-sim/editing/pasteboard/two-step-paste.html
 804
 805         * platform/ios/PlatformPasteboardIOS.mm:
 806         (WebCore::PlatformPasteboard::write):
 807
 808 2014-07-09  Andy Estes  <aestes@apple.com>
 809
 810         [iOS] Some QuickLook documents are not displayed as previews
 811         https://bugs.webkit.org/show_bug.cgi?id=134788
 812         <rdar://problem/17278194>
 813
 814         Reviewed by Pratik Solanki.
 815
 816         ResourceRequest::isMainResourceRequest() is not reliable (e.g. ResourceRequests reconstituted from an
 817         NSURLRequest have no way to set this bit properly) and should probably be removed. Instead of using
 818         isMainResourceRequest() to check if a QuickLookHandle should be created when the Network Process is in use,
 819         just check if the current ResourceLoader is the same as the DocumentLoader's ResourceLoader.
 820
 821         No new tests. QuickLook is not testable from WebKit.
 822
 823         * platform/network/ios/QuickLook.mm:
 824         (WebCore::QuickLookHandle::create):
 825
 826 2014-07-10  Andreas Kling  <akling@apple.com>
 827
 828         [iOS WebKit2] Don't uninstall the memory pressure listener after first signal.
 829         <https://webkit.org/b/134814>
 830
 831         The call to uninstall() should not be made on iOS since iOS doesn't use the
 832         hold-off mechanism in MemoryPressureHandler.
 833
 834         Reviewed by Benjamin Poulain.
 835
 836         * platform/cocoa/MemoryPressureHandlerCocoa.mm:
 837         (WebCore::MemoryPressureHandler::respondToMemoryPressure):
 838
 839 2014-07-10  Jer Noble  <jer.noble@apple.com>
 840
 841         [MSE] Overlapping appended ranges must cause the decoder to flush and re-enqueue.
 842         https://bugs.webkit.org/show_bug.cgi?id=134805
 843
 844         Reviewed by Eric Carlson.
 845
 846         When appending a range of media data which overlaps with an existing range, SourceBuffer
 847         must cause the decoder to flush and re-enqueue samples. Those samples removed by appending
 848         the overlapping range may have already been enqueued for display, and unless the decode
 849         queue is flushed, corruption or decode errors may occur as the new samples are enqueued for
 850         display.
 851
 852         Add a boolean flag onto TrackBuffer to indicate whether the decoder needs to be flushed and
 853         re-enqueued the next time an append operation completes. Set this flag whenever samples are
 854         removed due to overlapping or an explicit call to removeCodedFrames(). Move the contents of
 855         sourceBufferPrivateSeekToTime() (which previously did flushing and re-enqueueing) into a new
 856         function, reenqueueMediaForTime(), which can be called from sourceBufferPrivateAppendComplete().
 857
 858         Drive-by fix: findSyncSampleAfterDecodeIterator() would return the passed in iterator if that
 859         sample is a sync-sample. Fix this to correctly return the next sync sample.
 860
 861         Drive-by fix: Use a SampleMap, rather than a DecodeOrderSampleMap, to track erased samples
 862         so that the erasedSamples are correctly accounted for in both presentation and decode orders.
 863
 864         * Modules/mediasource/SampleMap.cpp:
 865         (WebCore::SampleMap::empty): Add convenience method.
 866         (WebCore::DecodeOrderSampleMap::findSyncSampleAfterDecodeIterator): Drive-by fix.
 867         * Modules/mediasource/SampleMap.h:
 868         (WebCore::SampleMap::decodeOrder): Added const accessor.
 869         (WebCore::SampleMap::presentationOrder): Ditto.
 870         (WebCore::SampleMap::addRange): Added.
 871         * Modules/mediasource/SourceBuffer.cpp:
 872         (WebCore::SourceBuffer::TrackBuffer::TrackBuffer): Add needsReenqueeing flag.
 873         (WebCore::SourceBuffer::sourceBufferPrivateSeekToTime): Move contents into reenqueueMediaForTime().
 874         (WebCore::SourceBuffer::sourceBufferPrivateAppendComplete): Call reenqueMediaForTime() if necessary.
 875         (WebCore::SourceBuffer::removeCodedFrames): Set needsReenqueing.
 876         (WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample): Ditto.
 877         (WebCore::SourceBuffer::provideMediaData): Drive-by fix.
 878         (WebCore::SourceBuffer::reenqueueMediaForTime): Moved from sourceBufferPrivateSeekToTime().
 879         * Modules/mediasource/SourceBuffer.h:
 880
 881 2014-07-10  Pratik Solanki  <psolanki@apple.com>
 882
 883         ASSERT in SharedBuffer::maybeAppendDataArray() on MobileSafari launch
 884         https://bugs.webkit.org/show_bug.cgi?id=134812
 885         <rdar://problem/17628434>
 886
 887         Reviewed by Joseph Pecoraro.
 888
 889         Fix bug in my fix in r170930. Initialize the badly named m_shouldUsePurgeableMemory field to
 890         false. This field indicates when it is okay to use purgeable memory and is set to true once
 891         the resource is finished loading. By setting it to true in the constructor we were creating
 892         purgeable memory while the resource was still being loaded and this triggered the assert.
 893
 894         No new tests. Should be covered by existing tests.
 895
 896         * platform/cf/SharedBufferCF.cpp:
 897         (WebCore::SharedBuffer::SharedBuffer):
 898
 899 2014-07-10  Andreas Kling  <akling@apple.com>
 900
 901         [iOS WebKit2] Some memory pressure relief tweaks.
 902         <https://webkit.org/b/134811>
 903
 904         Split memory pressure relief into critical and non-critical sections.
 905         Non-critical relief is for clearing out things that are really not
 906         essential, e.g unused font data, text measurement caches, etc.
 907
 908         On iOS, only flip the "WebKit is under memory pressure" flag when we
 909         are under *critical* memroy pressure, rather than doing it early on
 910         and gimping ourselves because other processes are too big.
 911
 912         Also added logging for when we transition in/out of system pressure.
 913
 914         Reviewed by Geoffrey Garen.
 915
 916         * platform/MemoryPressureHandler.cpp:
 917         (WebCore::MemoryPressureHandler::releaseNoncriticalMemory):
 918         (WebCore::MemoryPressureHandler::releaseCriticalMemory):
 919         (WebCore::MemoryPressureHandler::releaseMemory):
 920         * platform/MemoryPressureHandler.h:
 921         (WebCore::MemoryPressureHandler::ReliefLogger::loggingEnabled):
 922         * platform/cocoa/MemoryPressureHandlerCocoa.mm:
 923         (WebCore::MemoryPressureHandler::install):
 924
 925 2014-07-10  Beth Dakin  <bdakin@apple.com>
 926
 927         Need Setting/WKPreference that allows clients to prevent scrollbars from drawing
 928         on a secondary thread
 929         https://bugs.webkit.org/show_bug.cgi?id=134778
 930         -and corresponding-
 931         <rdar://problem/17595333>
 932
 933         Reviewed by Tim Horton.
 934
 935         This is a requirement for some types of performance tests. The patch adds a new
 936         virtual function to ScrollableArea that forces subclasses to indicate the value of
 937         the Setting. This is required because Scrollbar and ScrollableArea can’t get to
 938         Settings on their own.
 939
 940         * page/FrameView.cpp:
 941         (WebCore::FrameView::forceUpdateScrollbarsOnMainThreadForPerformanceTesting):
 942         * page/FrameView.h:
 943         * page/Settings.in:
 944         * platform/ScrollableArea.h:
 945         * platform/Scrollbar.cpp:
 946         (WebCore::Scrollbar::supportsUpdateOnSecondaryThread):
 947         * platform/win/PopupMenuWin.h:
 948         * rendering/RenderLayer.cpp:
 949         (WebCore::RenderLayer::forceUpdateScrollbarsOnMainThreadForPerformanceTesting):
 950         * rendering/RenderLayer.h:
 951         * rendering/RenderListBox.cpp:
 952         (WebCore::RenderListBox::forceUpdateScrollbarsOnMainThreadForPerformanceTesting):
 953         * rendering/RenderListBox.h:
 954
 955 2014-07-10  Brady Eidson  <beidson@apple.com>
 956
 957         Phone number highlights should always be visible if the mouse hovers over.
 958         <rdar://problem/17527476> and https://bugs.webkit.org/show_bug.cgi?id=134784
 959
 960         Reviewed by Tim Horton.
 961
 962         * WebCore.exp.in:
 963
 964 2014-07-10  Timothy Horton  <timothy_horton@apple.com>
 965
 966         Assertions or crashes under _takeViewSnapshot when restoring windows
 967         https://bugs.webkit.org/show_bug.cgi?id=134792
 968
 969         Reviewed by Simon Fraser.
 970
 971         * platform/graphics/cocoa/IOSurface.mm:
 972         (IOSurface::createFromImage):
 973
 974 2014-07-10  Youenn Fablet  <youenn.fablet@crf.canon.fr>
 975
 976         [XHR] overrideMimeType() should be able to change encoding in HEADERS RECEIVED state
 977         https://bugs.webkit.org/show_bug.cgi?id=128968
 978
 979         Reviewed by Alexey Proskuryakov.
 980
 981         Moved response encoding computation from didReceiveResponse to didReceiveData, just before the decoder is instantiated.
 982         This allows overrideMimeType to be changed within readystatechange event callback and have an impact on selected encoding.
 983
 984         Test: http/tests/xmlhttprequest/overridemimetype-headers-received-state-force-shiftjis.html
 985
 986         * xml/XMLHttpRequest.cpp:
 987         (WebCore::XMLHttpRequest::didReceiveResponse): Removed m_responseEncoding computation from response and mimeTypeOverride.
 988         (WebCore::XMLHttpRequest::didReceiveData): Added m_responseEncoding computation from response and mimeTypeOverride.
 989
 990 2014-07-09  Brady Eidson  <beidson@apple.com>
 991
 992         Fire connected/disconnected events for Gamepads.
 993         https://bugs.webkit.org/show_bug.cgi?id=134386
 994
 995         Reviewed by Dean Jackson.
 996
 997         No new tests (No effect in a currently tested config)
 998
 999         * Modules/gamepad/GamepadManager.cpp:
1000         (WebCore::navigatorGamepadFromDOMWindow): Handling converting a possibly-null Navigator into
1001             a possibly null NavigatorGamepad.
1002         (WebCore::GamepadManager::platformGamepadConnected): Notify blind Navigator/DOMWindows of all
1003             previously attached Gamepads, then notify everybody of this new gamepad.
1004         (WebCore::GamepadManager::platformGamepadDisconnected): Handle dispatching the disconnected
1005             event to all registered DOMWindows.
1006         (WebCore::GamepadManager::platformGamepadInputActivity): Notify blind Navigator/DOMWindows of all
1007             attached Gamepads.
1008         (WebCore::GamepadManager::makeGamepadVisible): Handles notifying setting up a new gamepads
1009             with all NavigatorGamepads as well as dispatching the connected even to DOMWindows.
1010         (WebCore::GamepadManager::registerDOMWindow):
1011         (WebCore::GamepadManager::unregisterDOMWindow):
1012         (WebCore::GamepadManager::makeGamepadsVisibileToBlindNavigators): Deleted.
1013         * Modules/gamepad/GamepadManager.h:
1014
1015         * Modules/gamepad/NavigatorGamepad.cpp:
1016         (WebCore::NavigatorGamepad::gamepadAtIndex):
1017         * Modules/gamepad/NavigatorGamepad.h:
1018
1019 2014-07-09  Benjamin Poulain  <bpoulain@apple.com>
1020
1021         [iOS][WK2] Disable text quantization while actively changing the page's scale factor
1022         https://bugs.webkit.org/show_bug.cgi?id=134781
1023
1024         Reviewed by Tim Horton and Myles C. Maxfield.
1025
1026         Query the chrome client to setup quantization on each layers.
1027
1028         * page/ChromeClient.h:
1029         (WebCore::ChromeClient::hasStablePageScaleFactor):
1030         * platform/graphics/mac/FontMac.mm:
1031         (WebCore::Font::drawGlyphs):
1032         * rendering/RenderLayer.cpp:
1033         (WebCore::RenderLayer::setupFontSubpixelQuantization):
1034
1035 2014-07-09  peavo@outlook.com  <peavo@outlook.com>
1036
1037         [Curl] Cache entry invalidated too early.
1038         https://bugs.webkit.org/show_bug.cgi?id=134681
1039
1040         Reviewed by Alex Christensen.
1041
1042         When a cache entry has expired, it is invalidated when a request for the corresponding url is started.
1043         This is too early, since the resource is possibly not modified (even though it has expired),
1044         and the server might respond with a 304 (not modified) response.
1045         When we then receive a 304 response, the cache entry is deleted, and we have no cached response/data to provide.
1046         This can be solved by not invalidating the entry when a request for that url is started,
1047         but instead invalidate when a 200 OK response is received (which means the resource has been modified).
1048
1049         * platform/network/curl/CurlCacheManager.cpp:
1050         (WebCore::CurlCacheManager::isCached): Avoid invalidating cache entry when it has expired, wait until we know the server response.
1051         * platform/network/curl/CurlCacheManager.h: Added const modifier.
1052
1053 2014-07-09  Myles C. Maxfield  <mmaxfield@apple.com>
1054
1055         Revert r170413 and r170390
1056         https://bugs.webkit.org/show_bug.cgi?id=134741
1057
1058         Reviewed by Benjamin Poulain.
1059
1060         These patches caused a variety of failures. I'm going to break them up into
1061         smaller chunks and commit them separately at some point in the future.
1062
1063         Requesting a review from Ben Poulain because there were some conflicts with
1064         r170561 in InlineTextBox.cpp and InlineTextBox.h.
1065
1066         * WebCore.exp.in:
1067         * platform/graphics/Font.cpp:
1068         (WebCore::Font::drawText):
1069         (WebCore::Font::drawEmphasisMarks):
1070         (WebCore::Font::adjustSelectionRectForText):
1071         (WebCore::computeUnderlineType):
1072         * platform/graphics/Font.h:
1073         * platform/graphics/FontFastPath.cpp:
1074         (WebCore::Font::getGlyphsAndAdvancesForSimpleText):
1075         (WebCore::Font::drawSimpleText):
1076         (WebCore::Font::drawEmphasisMarksForSimpleText):
1077         (WebCore::Font::drawGlyphBuffer):
1078         (WebCore::Font::drawEmphasisMarks):
1079         (WebCore::Font::adjustSelectionRectForSimpleText):
1080         * platform/graphics/GlyphBuffer.h:
1081         (WebCore::GlyphBuffer::size):
1082         (WebCore::GlyphBuffer::glyphs):
1083         (WebCore::GlyphBuffer::advances):
1084         (WebCore::GlyphBuffer::fontDataAt):
1085         (WebCore::GlyphBuffer::advanceAt):
1086         (WebCore::GlyphBuffer::offsetAt):
1087         (WebCore::GlyphBuffer::reverse):
1088         (WebCore::GlyphBuffer::offsetInString):
1089         (WebCore::GlyphBuffer::swap):
1090         * platform/graphics/GraphicsContext.cpp:
1091         * platform/graphics/Latin1TextIterator.h:
1092         (WebCore::Latin1TextIterator::Latin1TextIterator):
1093         (WebCore::Latin1TextIterator::currentCharacter):
1094         * platform/graphics/SimpleFontData.h:
1095         * platform/graphics/SurrogatePairAwareTextIterator.cpp:
1096         (WebCore::SurrogatePairAwareTextIterator::SurrogatePairAwareTextIterator):
1097         * platform/graphics/SurrogatePairAwareTextIterator.h:
1098         (WebCore::SurrogatePairAwareTextIterator::currentCharacter):
1099         * platform/graphics/TextRun.h:
1100         (WebCore::TextRun::length):
1101         (WebCore::TextRun::charactersLength):
1102         * platform/graphics/WidthIterator.cpp:
1103         (WebCore::WidthIterator::glyphDataForCharacter):
1104         (WebCore::applyFontTransforms):
1105         (WebCore::WidthIterator::advanceInternal):
1106         (WebCore::WidthIterator::advance):
1107         (WebCore::WidthIterator::advanceOneCharacter):
1108         * platform/graphics/WidthIterator.h:
1109         * platform/graphics/cairo/FontCairo.cpp:
1110         (WebCore::Font::drawGlyphs):
1111         * platform/graphics/cairo/FontCairoHarfbuzzNG.cpp:
1112         (WebCore::Font::drawComplexText):
1113         (WebCore::Font::drawEmphasisMarksForComplexText):
1114         (WebCore::Font::adjustSelectionRectForComplexText):
1115         * platform/graphics/harfbuzz/HarfBuzzShaper.cpp:
1116         (WebCore::HarfBuzzShaper::setNormalizedBuffer):
1117         * platform/graphics/mac/FontComplexTextMac.cpp:
1118         (WebCore::Font::adjustSelectionRectForComplexText):
1119         (WebCore::Font::getGlyphsAndAdvancesForComplexText):
1120         (WebCore::Font::drawComplexText):
1121         (WebCore::Font::drawEmphasisMarksForComplexText):
1122         * platform/graphics/mac/FontMac.mm:
1123         (WebCore::Font::drawGlyphs):
1124         * platform/graphics/win/FontCGWin.cpp:
1125         (WebCore::Font::drawGlyphs):
1126         * platform/graphics/win/FontWin.cpp:
1127         (WebCore::Font::adjustSelectionRectForComplexText):
1128         (WebCore::Font::getGlyphsAndAdvancesForComplexText):
1129         (WebCore::Font::drawComplexText):
1130         (WebCore::Font::drawEmphasisMarksForComplexText):
1131         * platform/graphics/wince/FontWinCE.cpp:
1132         (WebCore::Font::drawGlyphs):
1133         (WebCore::Font::drawComplexText):
1134         (WebCore::Font::drawEmphasisMarksForComplexText):
1135         (WebCore::Font::selectionRectForComplexText):
1136         * rendering/InlineTextBox.cpp:
1137         (WebCore::InlineTextBox::isSelected):
1138         (WebCore::InlineTextBox::selectionState):
1139         (WebCore::InlineTextBox::localSelectionRect):
1140         (WebCore::InlineTextBox::paint):
1141         (WebCore::InlineTextBox::selectionStartEnd):
1142         (WebCore::InlineTextBox::paintSelection):
1143         (WebCore::InlineTextBox::paintCompositionBackground):
1144         (WebCore::InlineTextBox::paintDocumentMarker):
1145         (WebCore::InlineTextBox::paintTextMatchMarker):
1146         (WebCore::InlineTextBox::computeRectForReplacementMarker):
1147         (WebCore::InlineTextBox::paintCompositionUnderline):
1148         (WebCore::InlineTextBox::positionForOffset):
1149         (WebCore::InlineTextBox::constructTextRun):
1150         * rendering/InlineTextBox.h:
1151         (WebCore::InlineTextBox::truncation):
1152         * rendering/RenderCombineText.cpp:
1153         (WebCore::RenderCombineText::getStringToRender):
1154         * rendering/RenderCombineText.h:
1155         * rendering/RenderTextLineBoxes.cpp:
1156         (WebCore::ellipsisRectForBox):
1157         * rendering/svg/SVGInlineFlowBox.cpp:
1158         (WebCore::SVGInlineFlowBox::computeTextMatchMarkerRectForRenderer):
1159         * rendering/svg/SVGInlineTextBox.cpp:
1160         (WebCore::SVGInlineTextBox::positionForOffset):
1161         (WebCore::SVGInlineTextBox::localSelectionRect):
1162         (WebCore::SVGInlineTextBox::paintSelectionBackground):
1163         (WebCore::SVGInlineTextBox::mapStartEndPositionsIntoFragmentCoordinates):
1164         (WebCore::SVGInlineTextBox::paintText):
1165         * rendering/svg/SVGInlineTextBox.h:
1166         * rendering/svg/SVGTextMetricsBuilder.cpp:
1167         (WebCore::SVGTextMetricsBuilder::currentCharacterStartsSurrogatePair):
1168         (WebCore::SVGTextMetricsBuilder::advance):
1169         * rendering/svg/SVGTextQuery.cpp:
1170         (WebCore::SVGTextQuery::mapStartEndPositionsIntoFragmentCoordinates):
1171         (WebCore::SVGTextQuery::modifyStartEndPositionsRespectingLigatures):
1172         (WebCore::SVGTextQuery::subStringLengthCallback):
1173         (WebCore::SVGTextQuery::startPositionOfCharacterCallback):
1174         (WebCore::SVGTextQuery::endPositionOfCharacterCallback):
1175         (WebCore::SVGTextQuery::rotationOfCharacterCallback):
1176         (WebCore::SVGTextQuery::extentOfCharacterCallback):
1177         (WebCore::SVGTextQuery::characterNumberAtPositionCallback):
1178         * rendering/svg/SVGTextQuery.h:
1179         * rendering/svg/SVGTextRunRenderingContext.cpp:
1180         (WebCore::SVGGlyphToPathTranslator::SVGGlyphToPathTranslator):
1181         (WebCore::SVGTextRunRenderingContext::glyphDataForCharacter):
1182         * rendering/svg/SVGTextRunRenderingContext.h:
1183         * svg/SVGFontData.cpp:
1184         (WebCore::SVGFontData::applySVGGlyphSelection):
1185         * svg/SVGFontData.h:
1186
1187 2014-07-08  Jeffrey Pfau  <jpfau@apple.com>
1188
1189         Fix flaky loading tests
1190         https://bugs.webkit.org/show_bug.cgi?id=133696
1191
1192         Reviewed by Ryosuke Niwa.
1193
1194         In some cases, the Frame may be destroyed while the loader is
1195         initializing. Make sure we don't delete it too early.
1196
1197         * loader/FrameLoader.cpp:
1198         (WebCore::FrameLoader::init):
1199
1200 2014-07-09  Simon Fraser  <simon.fraser@apple.com>
1201
1202         Make IndirectCompositingReason an enum class
1203         https://bugs.webkit.org/show_bug.cgi?id=134789
1204
1205         Reviewed by Tim Horton.
1206
1207         Make RenderLayer::IndirectCompositingReason an enum class, and change some code
1208         in RenderLayerCompositor::reasonsForCompositing() to use a switch.
1209
1210         * rendering/RenderLayer.cpp:
1211         (WebCore::RenderLayer::RenderLayer):
1212         * rendering/RenderLayer.h:
1213         * rendering/RenderLayerCompositor.cpp:
1214         (WebCore::RenderLayerCompositor::computeCompositingRequirements):
1215         (WebCore::RenderLayerCompositor::requiresOwnBackingStore):
1216         (WebCore::RenderLayerCompositor::reasonsForCompositing):
1217         (WebCore::RenderLayerCompositor::requiresCompositingForIndirectReason):
1218
1219 2014-07-09  Pratik Solanki  <psolanki@apple.com>
1220
1221         Make SharedBuffer::append(SharedBuffer*) be smarter about CFData and data arrays
1222         https://bugs.webkit.org/show_bug.cgi?id=134731
1223
1224         Reviewed by Antti Koivisto.
1225
1226         If the target SharedBuffer has a CFDataRef or a data array then we can simply retain that
1227         CFDataRef or data array elements in the SharedBuffer being appended to. This avoids
1228         unnecessary copying.
1229
1230         No new tests because no functional changes.
1231
1232         * platform/SharedBuffer.cpp:
1233         (WebCore::SharedBuffer::append):
1234         (WebCore::SharedBuffer::maybeAppendPlatformData):
1235         * platform/SharedBuffer.h:
1236         * platform/cf/SharedBufferCF.cpp:
1237         (WebCore::SharedBuffer::maybeAppendPlatformData):
1238         (WebCore::SharedBuffer::maybeAppendDataArray):
1239         * platform/soup/SharedBufferSoup.cpp:
1240         (WebCore::SharedBuffer::maybeAppendPlatformData):
1241
1242 2014-07-09  Brent Fulgham  <bfulgham@apple.com>
1243
1244         [Win] Remove uses of 'bash' in build system
1245         https://bugs.webkit.org/show_bug.cgi?id=134782
1246         <rdar://problem/17615533>
1247
1248         Reviewed by Dean Jackson.
1249
1250         Remove uses of 'bash' by replacing Windows-specific bash scripts
1251         with Perl equivalents.
1252
1253         * WebCore.vcxproj/WebCoreGenerated.make:
1254         * WebCore.vcxproj/WebCoreGenerated.vcxproj:
1255         * WebCore.vcxproj/WebCoreGenerated.vcxproj.filters:
1256         * WebCore.vcxproj/WebCorePreBuild.cmd:
1257         * WebCore.vcxproj/build-generated-files.pl: Copied from Source/WebCore/WebCore.vcxproj/build-generated-files.sh.
1258         * WebCore.vcxproj/build-generated-files.sh: Removed.
1259         * WebCore.vcxproj/migrate-scripts.pl: Copied from Source/WebCore/WebCore.vcxproj/migrate-scripts.sh.
1260         * WebCore.vcxproj/migrate-scripts.sh: Removed.
1261
1262 2014-07-09  Anders Carlsson  <andersca@apple.com>
1263
1264         Support transparent WKWebViews
1265         https://bugs.webkit.org/show_bug.cgi?id=134779
1266         <rdar://problem/17351058>
1267
1268         Reviewed by Tim Horton.
1269
1270         Schedule rebuilding the compositing layers if a FrameView's transparency changes.
1271
1272         * page/FrameView.cpp:
1273         (WebCore::FrameView::setTransparent):
1274
1275 2014-07-09  Javier Fernandez  <jfernandez@igalia.com>
1276         CSS canvas color parsing accepts invalid color identifiers
1277         https://bugs.webkit.org/show_bug.cgi?id=134661
1278
1279         Reviewed by Benjamin Poulain.
1280
1281         Current implementation of the CSSParser::parseSystemColor assumes
1282         that if a valid cssValueKeywordID is got then it has to be a valid
1283         color. Such assumption is wrong and lead to many bugs and layout
1284         test failures.
1285
1286         The parseSystemFunction determines now whether the parsed color is
1287         valid or not.
1288
1289         Addtionally, a new method has been added to share the logic of
1290         determining whether a CSSValueID is a valid primitive values for
1291         colors or not. Generally, we should avoid passing invalid color
1292         identifiers to the theming API.
1293
1294         No new tests, but added additional cases to the
1295         canvas-color-serialization.html, test-setting-canvas-color and
1296         rgb-color-parse test.
1297
1298         * css/CSSParser.cpp:
1299         (WebCore::validPrimitiveValueColor): Added.
1300         (WebCore::parseColorValue):
1301         (WebCore::CSSParser::parseSystemColor):
1302
1303 2014-06-28  Jer Noble  <jer.noble@apple.com>
1304
1305         [MSE] http/tests/media/media-source/mediasource-remove.html is failing
1306         https://bugs.webkit.org/show_bug.cgi?id=134768
1307
1308         Reviewed by Eric Carlson.
1309
1310         Fix multiple bugs causing mediasource-remove.html to fail:
1311
1312         Separate out setDuration() into setDurationInternal() so that steps which require
1313         us to run the "duration change algorithm" don't bail out if the SoureBuffer is
1314         inside updating().
1315
1316         * Modules/mediasource/MediaSource.cpp:
1317         (WebCore::MediaSource::setDuration):
1318         (WebCore::MediaSource::setDurationInternal):
1319         * Modules/mediasource/MediaSource.h:
1320         * Modules/mediasource/SourceBuffer.cpp:
1321         (WebCore::SourceBuffer::sourceBufferPrivateDidReceiveInitializationSegment):
1322         (WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample):
1323
1324         When removing coded frames, do so starting from the first sample in the range in
1325         decode order, so that frames dependant on removed frames are themselves removed. Add
1326         a convenience method in SampleMap findSampleWithPresentationTime(), and rename
1327         findSampleAfterPresentationTime() to findSampleOnOrAfterPresentationTime() to correctly
1328         reflect what the method does, and simplify its implementation by searching the map's keys
1329         directly.
1330
1331         * Modules/mediasource/SampleMap.cpp:
1332         (WebCore::PresentationOrderSampleMap::findSampleWithPresentationTime):
1333         (WebCore::PresentationOrderSampleMap::findSampleOnOrAfterPresentationTime):
1334         (WebCore::DecodeOrderSampleMap::findSyncSampleAfterPresentationTime):
1335         (WebCore::PresentationOrderSampleMap::findSampleAfterPresentationTime): Deleted.
1336         * Modules/mediasource/SampleMap.h:
1337         * Modules/mediasource/SourceBuffer.cpp:
1338         (WebCore::decodeTimeComparator):
1339         (WebCore::SourceBuffer::removeCodedFrames):
1340
1341         Throw the correct exception (INVALID_STATE_ERR) from SourceBuffer::remove().
1342
1343         * Modules/mediasource/SourceBuffer.cpp:
1344         (WebCore::SourceBuffer::remove):
1345
1346 2014-07-09  Pratik Solanki  <psolanki@apple.com>
1347
1348         Add SharedBuffer::wrapCFDataArray() and use it
1349         https://bugs.webkit.org/show_bug.cgi?id=134733
1350
1351         Reviewed by Antti Koivisto.
1352
1353         No new tests. Should be covered by existing tests.
1354
1355         * platform/SharedBuffer.h:
1356         * platform/cf/SharedBufferCF.cpp:
1357         (WebCore::SharedBuffer::wrapCFDataArray):
1358         (WebCore::SharedBuffer::SharedBuffer):
1359         * platform/network/ResourceHandle.h:
1360         * platform/network/cf/ResourceHandleCFNet.cpp:
1361         (WebCore::ResourceHandle::handleDataArray): Deleted.
1362         * platform/network/cf/ResourceHandleCFURLConnectionDelegateWithOperationQueue.cpp:
1363         (WebCore::ResourceHandleCFURLConnectionDelegateWithOperationQueue::didReceiveDataArray):
1364         * platform/network/cf/SynchronousResourceHandleCFURLConnectionDelegate.cpp:
1365         (WebCore::SynchronousResourceHandleCFURLConnectionDelegate::didReceiveDataArray):
1366         * platform/network/mac/WebCoreResourceHandleAsDelegate.mm:
1367         (-[WebCoreResourceHandleAsDelegate connection:didReceiveDataArray:]):
1368         * platform/network/mac/WebCoreResourceHandleAsOperationQueueDelegate.mm:
1369         (-[WebCoreResourceHandleAsOperationQueueDelegate connection:didReceiveDataArray:]):
1370
1371 2014-07-09  Pratik Solanki  <psolanki@apple.com>
1372
1373         Move resource buffering from SynchronousNetworkLoaderClient to NetworkResourceLoader
1374         https://bugs.webkit.org/show_bug.cgi?id=134732
1375
1376         Reviewed by Darin Adler.
1377
1378         No new tests because no functional changes.
1379
1380         * WebCore.exp.in:
1381
1382 2014-07-09  Tim Horton  <timothy_horton@apple.com>
1383
1384         Fix the !USE(IOSURFACE) build.
1385
1386         * WebCore.exp.in:
1387
1388 2014-07-09  Tim Horton  <timothy_horton@apple.com>
1389
1390         Use IOSurface ViewSnapshots everywhere on Mac, remove JPEG encoding path
1391         https://bugs.webkit.org/show_bug.cgi?id=134773
1392
1393         Reviewed by Anders Carlsson.
1394
1395         * WebCore.exp.in:
1396         * platform/graphics/cocoa/IOSurface.h:
1397         * platform/graphics/cocoa/IOSurface.mm:
1398         (IOSurface::createFromImage):
1399         Move make-an-IOSurface-from-a-CGImageRef into WebCore::IOSurface.
1400
1401 2014-07-09  Enrica Casucci  <enrica@apple.com>
1402
1403         Implement Editor::fontAttributesForSelectionStart() or iOS.
1404         https://bugs.webkit.org/show_bug.cgi?id=134771
1405         <rdar://problem/16167838>
1406
1407         Reviewed by Ryosuke Niwa.
1408
1409         Implement for iOS the equivalent function we have for OS X.
1410
1411         * editing/ios/EditorIOS.mm:
1412         (WebCore::Editor::fontAttributesForSelectionStart):
1413
1414 2014-07-09  Brent Fulgham  <bfulgham@apple.com>
1415
1416         [Win] Remove use of 'grep' in build steps
1417         https://bugs.webkit.org/show_bug.cgi?id=134770
1418         <rdar://problem/17608783>
1419
1420         Reviewed by Tim Horton.
1421
1422         Replace uses of the grep command in Windows builds with the equivalent
1423         Perl program.
1424
1425         * WebCore.vcxproj/WebCoreGenerated.make:
1426         * WebCore.vcxproj/WebCorePreBuild.cmd:
1427         * platform/graphics/avfoundation/cf/WebCoreAVCFResourceLoader.cpp:
1428         (WebCore::WebCoreAVCFResourceLoader::notifyFinished):
1429
1430 2014-07-09  Alex Christensen  <achristensen@webkit.org>
1431
1432         Added css jit profiler, disabled by default.
1433         https://bugs.webkit.org/show_bug.cgi?id=134695
1434
1435         Reviewed by Benjamin Poulain.
1436
1437         * css/ElementRuleCollector.cpp:
1438         (WebCore::ElementRuleCollector::ruleMatches):
1439         Tell the RuleData when its compiled selector is used.
1440         * css/RuleSet.cpp:
1441         (WebCore::RuleData::RuleData):
1442         Initialize the compiled selector use count to 0.
1443         * css/RuleSet.h:
1444         (WebCore::RuleData::~RuleData):
1445         (WebCore::RuleData::compiledSelectorUsed):
1446         Count and log the number of times a compiled selector is used.
1447         * cssjit/SelectorCompiler.h:
1448         Added option for CSS_SELECTOR_JIT_PROFILING set to 0 by default.
1449         * dom/SelectorQuery.cpp:
1450         (WebCore::SelectorDataList::executeCompiledSimpleSelectorChecker):
1451         (WebCore::SelectorDataList::execute):
1452         Tell the SelectorData when its compiled selector is used.
1453         * dom/SelectorQuery.h:
1454         (WebCore::SelectorDataList::SelectorData::SelectorData):
1455         (WebCore::SelectorDataList::SelectorData::~SelectorData):
1456         (WebCore::SelectorDataList::SelectorData::compiledSelectorUsed):
1457         Count and log the number of times a compiled selector is used.
1458
1459 2014-07-09  Eric Carlson  <eric.carlson@apple.com>
1460
1461         [iOS] caption size is sometimes incorrect in fullscreen
1462         https://bugs.webkit.org/show_bug.cgi?id=134740
1463
1464         Reviewed by Jer Noble.
1465
1466         Captions on iOS are displayed in fullscreen with a TextTrackRepresentation object. Because
1467         the fullscreen video presentation is controlled by code in the UI process running on the
1468         UI thread, WebCore is notified of changes to fullscreen state asynchronously. This resulted
1469         in the TextTrackRepresentation object being created and/or destroyed too late some of the
1470         time, which caused us to sometimes display captions incorrectly. Fix this by setting up and
1471         tearing down the TextTrackRepresentation object when WebCore's 'webkitfullscreenchange'
1472         event fires.
1473
1474         * Modules/mediacontrols/MediaControlsHost.cpp:
1475         (WebCore::MediaControlsHost::enteredFullscreen): Notify text track container.
1476         (WebCore::MediaControlsHost::exitedFullscreen): Ditto.
1477         * Modules/mediacontrols/MediaControlsHost.h:
1478         * Modules/mediacontrols/MediaControlsHost.idl:
1479
1480         * Modules/mediacontrols/mediaControlsApple.js:
1481         (Controller.prototype.handleFullscreenChange): Notify host of fullscreen change.
1482
1483         * html/HTMLMediaElement.cpp:
1484         (WebCore::HTMLMediaElement::requiresTextTrackRepresentation): Only return true when in
1485             fullscreen.
1486         (WebCore::HTMLMediaElement::setVideoFullscreenLayer): Call updateTextTrackDisplay.
1487
1488         * html/shadow/MediaControlElements.cpp:
1489         (WebCore::MediaControlTextTrackContainerElement::MediaControlTextTrackContainerElement): Initialize
1490             m_updateTextTrackRepresentationStyle to false.
1491         (WebCore::MediaControlTextTrackContainerElement::updateDisplay): Move logic for creating
1492             TextTrackRepresentation to updateTextTrackRepresentation.
1493         (WebCore::MediaControlTextTrackContainerElement::updateActiveCuesFontSize): New, split out of updateTimerFired.
1494         (WebCore::MediaControlTextTrackContainerElement::updateTimerFired): Move code to force immediate
1495             font size change to updateActiveCuesFontSize.
1496         (WebCore::MediaControlTextTrackContainerElement::updateTextTrackRepresentation): New, update
1497             text track representation, creating first if necessary.
1498         (WebCore::MediaControlTextTrackContainerElement::clearTextTrackRepresentation): Add an early
1499             return if we don't have a text track representation.
1500         (WebCore::MediaControlTextTrackContainerElement::updateStyleForTextTrackRepresentation): Early
1501             return if there is nothing to be done.
1502         (WebCore::MediaControlTextTrackContainerElement::enteredFullscreen): Force a caption update
1503             if there are visible captions.
1504         (WebCore::MediaControlTextTrackContainerElement::updateSizes): Set m_updateTextTrackRepresentationStyle
1505             to true.
1506         (WebCore::MediaControlTextTrackContainerElement::textTrackRepresentationBoundsChanged):  Force a
1507             caption update if there are visible captions.
1508         * html/shadow/MediaControlElements.h:
1509
1510 2014-07-09  Jer Noble  <jer.noble@apple.com>
1511
1512         [MSE] http/tests/media/media-source/mediasource-endofstream-invaliderror.html is failing.
1513         https://bugs.webkit.org/show_bug.cgi?id=134413
1514
1515         Reviewed by Eric Carlson.
1516
1517         Update MediaSource.endOfStream() to take an optional enum instead of an optional String.
1518
1519         * Modules/mediasource/MediaSource.cpp:
1520         (WebCore::MediaSource::endOfStream):
1521         * Modules/mediasource/MediaSource.h:
1522         * Modules/mediasource/MediaSource.idl:
1523
1524 2014-07-08  Brady Eidson  <beidson@apple.com>
1525
1526         DOMWindows should register themselves with the GamepadManager when they have event listeners
1527         https://bugs.webkit.org/show_bug.cgi?id=134669
1528
1529         Reviewed by Dean Jackson.
1530
1531         No new tests (No effect in a currently tested config)
1532
1533         By having DOMWindows register with the GamepadManager when they have/don’t have gamepad event
1534         listeners, we accomplish two things:
1535         1 - It’s simple for GamepadManager to dispatch gamepad events directly to the DOMWindows interested.
1536         2 - We know when to start monitoring for gamepad connections.
1537
1538         * Modules/gamepad/GamepadManager.cpp:
1539         (WebCore::GamepadManager::GamepadManager):
1540         (WebCore::GamepadManager::registerNavigator):
1541         (WebCore::GamepadManager::unregisterNavigator):
1542         (WebCore::GamepadManager::registerDOMWindow):
1543         (WebCore::GamepadManager::unregisterDOMWindow):
1544         (WebCore::GamepadManager::maybeStartMonitoringGamepads): Start monitoring gamepads if there are any
1545           registered DOMWindows or NavigatorGamepads, and we aren’t already monitoring them.
1546         (WebCore::GamepadManager::maybeStopMonitoringGamepads): Stop monitoring gamepads if there aren’t any
1547           registered DOMWindows or NavigatorGameads, and we have been monitoring them before now.
1548         * Modules/gamepad/GamepadManager.h:
1549
1550         * dom/EventNames.h:
1551         (WebCore::EventNames::isGamepadEventType):
1552
1553         * page/DOMWindow.cpp:
1554         (WebCore::DOMWindow::DOMWindow):
1555         (WebCore::DOMWindow::~DOMWindow): If registered for gamepad events, unregister now.
1556         (WebCore::DOMWindow::incrementGamepadEventListenerCount): If this is the first gamepad event
1557           listener, register with the GamepadManager now.
1558         (WebCore::DOMWindow::decrementGamepadEventListenerCount): If this was the last gamepad event
1559           listener, unregister from the GamepadManager now.
1560         (WebCore::DOMWindow::addEventListener):
1561         (WebCore::DOMWindow::removeEventListener):
1562         * page/DOMWindow.h:
1563
1564 2014-07-08  Sun-woo Nam  <sunny.nam@samsung.com>
1565
1566         [EFL] Fix the build break when ENABLE_ENCRYPTED_MEDIA_V2 is turned on.
1567         https://bugs.webkit.org/show_bug.cgi?id=134750
1568
1569         Reviewed by Gyuyoung Kim.
1570
1571         When turning ENABLE_ENCRYPTED_MEDIA_V2 on, build break occurs.
1572         Some source files need to be included in CMakeLists.txt file for build.
1573         And wrap the included header file which is built only on mac and wince.
1574
1575         * CMakeLists.txt:
1576         * Modules/encryptedmedia/CDMPrivateMediaPlayer.cpp:
1577
1578 2014-07-08  Eric Carlson  <eric.carlson@apple.com>
1579
1580         [iOS] user caption preferences not applied
1581         https://bugs.webkit.org/show_bug.cgi?id=134599
1582
1583         Reviewed by Sam Weinig.
1584
1585         * page/CaptionUserPreferences.cpp:
1586         (WebCore::CaptionUserPreferences::updateCaptionStyleSheetOveride): All Pages in a PageGroup
1587         do not necessarily use the same PageUserContentController, so apply and remove caption user
1588         the style sheet directly to each page rather than to the page group.
1589
1590 2014-07-08  Commit Queue  <commit-queue@webkit.org>
1591
1592         Unreviewed, rolling out r170894.
1593         https://bugs.webkit.org/show_bug.cgi?id=134749
1594
1595         "It broke the build on mountainlion release 32bit" (Requested
1596         by mmirman on #webkit).
1597
1598         Reverted changeset:
1599
1600         "Added css jit profiler, disabled by default."
1601         https://bugs.webkit.org/show_bug.cgi?id=134695
1602         http://trac.webkit.org/changeset/170894
1603
1604 2014-07-08  Jeremy Jones  <jeremyj@apple.com>
1605
1606         WebVideoFullscreenInterfaceAVKit should lazily create m_playerController.
1607         https://bugs.webkit.org/show_bug.cgi?id=134706
1608
1609         Reviewed by Simon Fraser.
1610
1611         * platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
1612         (WebVideoFullscreenInterfaceAVKit::setWebVideoFullscreenModel):
1613         Don't set the delegate if m_playerController isn't created.
1614
1615 2014-07-08  Bem Jones-Bey  <bjonesbe@adobe.com>
1616
1617         Ensure we compute the height of replaced elements to 'auto' when appropriate.
1618         https://bugs.webkit.org/show_bug.cgi?id=134700
1619
1620         Reviewed by David Hyatt.
1621
1622         This fixes two issues:
1623
1624         1) If a replaced element has a percentage height specified then its
1625         height should compute to 'auto' when its containing block does not
1626         have a height 'specified explicitly'. We were taking this to mean when
1627         its containing block's specified height value is 'auto' - in fact it
1628         means when the containing block's computed height is auto.
1629
1630         2) Top and bottom on the containing block should only affect the
1631         height of the block if that block has absolute or fixed position.
1632
1633         This brings us into line with IE, Firefox, and Chrome.
1634
1635         This is a port of Blink patches by Rob Hogan and David Vest.
1636
1637         Tests: css2.1/20110323/height-percentage-005.htm
1638                fast/css/replaced-element-ignore-top-bottom.html
1639
1640         * rendering/RenderBoxModelObject.cpp:
1641         (WebCore::RenderBoxModelObject::hasAutoHeightOrContainingBlockWithAutoHeight):
1642
1643 2014-07-08  Alex Christensen  <achristensen@webkit.org>
1644
1645         Added css jit profiler, disabled by default.
1646         https://bugs.webkit.org/show_bug.cgi?id=134695
1647
1648         Reviewed by Benjamin Poulain.
1649
1650         * css/ElementRuleCollector.cpp:
1651         (WebCore::ElementRuleCollector::ruleMatches):
1652         Tell the RuleData when its compiled selector is used.
1653         * css/RuleSet.cpp:
1654         (WebCore::RuleData::RuleData):
1655         Initialize the compiled selector use count to 0.
1656         * css/RuleSet.h:
1657         (WebCore::RuleData::~RuleData):
1658         (WebCore::RuleData::compiledSelectorUsed):
1659         Count and log the number of times a compiled selector is used.
1660         * cssjit/SelectorCompiler.h:
1661         Added option for CSS_SELECTOR_JIT_PROFILING set to 0 by default.
1662         * dom/SelectorQuery.cpp:
1663         (WebCore::SelectorDataList::executeCompiledSimpleSelectorChecker):
1664         (WebCore::SelectorDataList::execute):
1665         Tell the SelectorData when its compiled selector is used.
1666         * dom/SelectorQuery.h:
1667         (WebCore::SelectorDataList::SelectorData::SelectorData):
1668         (WebCore::SelectorDataList::SelectorData::~SelectorData):
1669         (WebCore::SelectorDataList::SelectorData::compiledSelectorUsed):
1670         Count and log the number of times a compiled selector is used.
1671
1672 2014-07-08  Carlos Garcia Campos  <cgarcia@igalia.com>
1673
1674         Unreviewed. Fix GObject DOM bindings API breaks test after r170700.
1675
1676         Add missing WebKitDOMXPathNSResolver.symbols that I forgot to add
1677         in r170700.
1678
1679         * bindings/gobject/WebKitDOMXPathNSResolver.symbols: Added.
1680
1681 2014-07-07  Pratik Solanki  <psolanki@apple.com>
1682
1683         Unreviewed. iOS build fix after r170871.
1684
1685         * rendering/RenderThemeIOS.mm:
1686         (WebCore::adjustInputElementButtonStyle):
1687
1688 2014-07-07  Zalan Bujtas  <zalan@apple.com>
1689
1690         Subpixel rendering: icloud.com password arrow has clipped circle at some window sizes.
1691         https://bugs.webkit.org/show_bug.cgi?id=134710
1692         <rdar://problem/17545291>
1693
1694         Reviewed by Simon Fraser.
1695
1696         Transparent layer clipping needs device pixel snapping.
1697
1698         Test: fast/layers/wrong-clipping-semi-transparent-layer-on-subpixel-position.html
1699
1700         * rendering/RenderLayer.cpp:
1701         (WebCore::RenderLayer::beginTransparencyLayers):
1702
1703 2014-07-07  Benjamin Poulain  <benjamin@webkit.org>
1704
1705         LinkBuffer should not keep a reference to the MacroAssembler
1706         https://bugs.webkit.org/show_bug.cgi?id=134668
1707
1708         Reviewed by Geoffrey Garen.
1709
1710         * cssjit/SelectorCompiler.cpp:
1711         (WebCore::SelectorCompiler::SelectorCodeGenerator::compile):
1712
1713 2014-07-07  Zalan Bujtas  <zalan@apple.com>
1714
1715         Subpixel rendering: Inline box decoration rounds to integral.
1716         https://bugs.webkit.org/show_bug.cgi?id=134523
1717         <rdar://problem/17530298>
1718
1719         Reviewed by Darin Adler.
1720
1721         This patch removes 2 integral roundings from InlineFlowBox:
1722         1. Border and padding sizes are implicitly integral truncated by the 'int' return type
1723            of borderLogicalLeft/Right()/paddingLogicalLeft/Right(). It results in losing
1724            fractional border/padding values.
1725         2. Painting rectangle is explicitly rounded which pushes border and
1726            other decoration elements to odd device pixel positions on retina displays.
1727         These values get pixel snapped right before calling in to GraphicsContext::*.
1728
1729         Test: fast/inline/hidpi-inline-text-decoration-with-subpixel-value.html
1730
1731         * rendering/InlineBox.h:
1732         (WebCore::InlineBox::frameRect):
1733         * rendering/InlineFlowBox.cpp:
1734         (WebCore::InlineFlowBox::nodeAtPoint):
1735         (WebCore::InlineFlowBox::paintBoxDecorations):
1736         (WebCore::InlineFlowBox::paintMask):
1737         (WebCore::InlineFlowBox::roundedFrameRect): Deleted.
1738         * rendering/InlineFlowBox.h:
1739         (WebCore::InlineFlowBox::borderLogicalLeft):
1740         (WebCore::InlineFlowBox::borderLogicalRight):
1741         (WebCore::InlineFlowBox::paddingLogicalLeft):
1742         (WebCore::InlineFlowBox::paddingLogicalRight):
1743
1744 2014-07-07  Zalan Bujtas  <zalan@apple.com>
1745
1746         Pass RenderLayer reference instead of pointer to RenderLayer::paintingExtent().
1747         https://bugs.webkit.org/show_bug.cgi?id=134714
1748
1749         Reviewed by Simon Fraser.
1750
1751         No change in functionality.
1752
1753         * rendering/RenderLayer.cpp: * -> &
1754         (WebCore::expandClipRectForRegionAndReflection):
1755         (WebCore::expandClipRectForDescendantsAndReflection):
1756         (WebCore::transparencyClipBox):
1757         (WebCore::paintingExtent):
1758         (WebCore::RenderLayer::beginTransparencyLayers):
1759         (WebCore::RenderLayer::paintTransformedLayerIntoFragments):
1760         (WebCore::RenderLayer::hitTestTransformedLayerInFragments):
1761         (WebCore::RenderLayer::paintingExtent): Deleted.
1762         * rendering/RenderLayer.h:
1763
1764 2014-07-07  Myles C. Maxfield  <mmaxfield@apple.com>
1765
1766         Remove misleadingly-named Font::isSVGFont()
1767         https://bugs.webkit.org/show_bug.cgi?id=134697
1768
1769         Reviewed by Dean Jackson.
1770
1771         Being a SVG font is a property of the FontData object, not
1772         the Font object. Providing an isSVGFont() function on the
1773         Font class will likely lead to incorrect code and non-obvious
1774         bugs.
1775
1776         No new tests because there is no behavior change.
1777
1778         * platform/graphics/Font.h:
1779         (WebCore::Font::isSVGFont): Deleted.
1780         * platform/graphics/mac/FontMac.mm:
1781         (WebCore::Font::primaryFontDataIsSystemFont):
1782         * rendering/InlineTextBox.cpp:
1783         (WebCore::InlineTextBox::constructTextRun):
1784         * rendering/RenderBlock.cpp:
1785         (WebCore::constructTextRunInternal):
1786         * rendering/SimpleLineLayout.cpp:
1787         (WebCore::SimpleLineLayout::canUseFor):
1788         * rendering/svg/SVGInlineTextBox.cpp:
1789         (WebCore::SVGInlineTextBox::constructTextRun):
1790         * rendering/svg/SVGTextMetrics.cpp:
1791         (WebCore::SVGTextMetrics::constructTextRun):
1792         (WebCore::SVGTextMetrics::SVGTextMetrics):
1793
1794 2014-07-07  Brady Eidson  <beidson@apple.com>
1795
1796         Move Gamepad index down into PlatformGamepad.
1797         https://bugs.webkit.org/show_bug.cgi?id=134670
1798
1799         Reviewed by Darin Adler.
1800
1801         PlatformGamepads always have an inherent index anyways.
1802         This will make things easier in all upcoming work.
1803
1804         * Modules/gamepad/Gamepad.cpp:
1805         (WebCore::Gamepad::Gamepad):
1806         * Modules/gamepad/Gamepad.h:
1807
1808         * Modules/gamepad/GamepadManager.cpp:
1809         (WebCore::GamepadManager::platformGamepadConnected):
1810         (WebCore::GamepadManager::platformGamepadDisconnected):
1811         (WebCore::GamepadManager::makeGamepadsVisibileToBlindNavigators):
1812         * Modules/gamepad/GamepadManager.h:
1813
1814         * Modules/gamepad/NavigatorGamepad.cpp:
1815         (WebCore::NavigatorGamepad::gamepadsBecameVisible):
1816         (WebCore::NavigatorGamepad::gamepadConnected):
1817         (WebCore::NavigatorGamepad::gamepadDisconnected):
1818         * Modules/gamepad/NavigatorGamepad.h:
1819
1820         * platform/GamepadProviderClient.h:
1821
1822         * platform/PlatformGamepad.h:
1823         (WebCore::PlatformGamepad::index):
1824         (WebCore::PlatformGamepad::PlatformGamepad):
1825
1826         * platform/mac/HIDGamepad.cpp:
1827         (WebCore::HIDGamepad::HIDGamepad):
1828         * platform/mac/HIDGamepad.h:
1829
1830         * platform/mac/HIDGamepadProvider.cpp:
1831         (WebCore::HIDGamepadProvider::deviceAdded):
1832         (WebCore::HIDGamepadProvider::deviceRemoved):
1833         (WebCore::HIDGamepadProvider::removeGamepadForDevice):
1834         * platform/mac/HIDGamepadProvider.h:
1835
1836 2014-07-07  Simon Fraser  <simon.fraser@apple.com>
1837
1838         [UI-side compositing] Support reflections on custom layers like video
1839         https://bugs.webkit.org/show_bug.cgi?id=134701
1840
1841         Reviewed by Tim Horton.
1842
1843         Add a LayerTypeWebGLLayer layer type, and support casting to PlatformCALayerRemoteCustom.
1844
1845         * WebCore.exp.in: Export PlatformCALayerMac::layerTypeForPlatformLayer(CALayer*), and sort.
1846         * platform/graphics/ca/PlatformCALayer.h:
1847         (WebCore::PlatformCALayer::isPlatformCALayerRemoteCustom):
1848         * platform/graphics/ca/mac/PlatformCALayerMac.h:
1849         * platform/graphics/ca/mac/PlatformCALayerMac.mm:
1850         (PlatformCALayerMac::layerTypeForPlatformLayer):
1851         (PlatformCALayerMac::PlatformCALayerMac):
1852         (PlatformCALayerMac::commonInit):
1853
1854 2014-07-07  Simon Fraser  <simon.fraser@apple.com>
1855
1856         [UI-side compositing] Crash when starting a filter transition on a reflected layer
1857         https://bugs.webkit.org/show_bug.cgi?id=134694
1858
1859         Reviewed by Tim Horton.
1860
1861         Don't call the owner if we failed to find the animation key (which actually
1862         isn't used by PlatformCALayerMac anyway).
1863
1864         * platform/graphics/ca/mac/PlatformCALayerMac.mm:
1865         (-[WebAnimationDelegate animationDidStart:]):
1866
1867 2014-07-07  Alex Christensen  <achristensen@webkit.org>
1868
1869         [iOS WebGL] Fix crash with too many nested glsl functions.
1870         https://bugs.webkit.org/show_bug.cgi?id=134683
1871         <rdar://problem/17541529>
1872
1873         Reviewed by Geoffrey Garen.
1874
1875         This fixes the Khronos WebGL 1.0.3 test nested-functions-should-not-crash.html.
1876
1877         * platform/graphics/opengl/Extensions3DOpenGLCommon.cpp:
1878         (WebCore::Extensions3DOpenGLCommon::getTranslatedShaderSourceANGLE):
1879         Add SH_LIMIT_CALL_STACK_DEPTH to extraCompileOptions.
1880
1881 2014-07-07  Ion Rosca  <rosca@adobe.com>
1882
1883         [CSS Blending]The background images set on the root element will blend on an initial white backdrop.
1884         https://bugs.webkit.org/show_bug.cgi?id=134342
1885
1886         Reviewed by Dean Jackson.
1887
1888         Tests: css3/blending/background-blend-mode-body-image.html
1889                css3/blending/background-blend-mode-body-transparent-color-and-image.html
1890                css3/blending/background-blend-mode-body-transparent-image.html
1891
1892         * rendering/RenderBox.cpp:
1893         (WebCore::RenderBox::paintFillLayers):
1894         If the background layers have blend modes, we paint the base background color first,
1895         before starting a new transparency layer. After creating the transparency layer,
1896         the base background will be skipped, and the background layers will blend on top of
1897         a transparent backdrop.
1898         (WebCore::RenderBox::paintFillLayer):
1899         * rendering/RenderBox.h:
1900         * rendering/RenderBoxModelObject.cpp:
1901         (WebCore::RenderBoxModelObject::paintFillLayerExtended):
1902         This method is able to paint only the base background color (BaseBackgroundColorOnly),
1903         skipping the background-color and background-images. This option is available only
1904         on bottom layer of the root renderer and only if the background color is not opaque,
1905         otherwise the method will return early.
1906         Another option is to paint the root background without using the base background color
1907         (BaseBackgroundColorSkip).
1908         * rendering/RenderBoxModelObject.h:
1909         Adding the BaseBackgroundColorUsage enum.
1910
1911 2014-07-06  Yoav Weiss  <yoav@yoav.ws>
1912
1913         Turn on img@sizes compile flag
1914         https://bugs.webkit.org/show_bug.cgi?id=134634
1915
1916         Reviewed by Benjamin Poulain.
1917
1918         * Configurations/FeatureDefines.xcconfig: Moved compile flag to alphabetical order.
1919
1920 2014-07-06  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
1921
1922         Generate isMathMLFoo() toMathMLFoo() by using NODE_TYPE_CASTS macro and generateTypeHelpers
1923         https://bugs.webkit.org/show_bug.cgi?id=134647
1924
1925         Reviewed by Andreas Kling.
1926
1927         Use type cast macro and generation tag for isFoo() and toFoo() in some MathML classes.
1928
1929         No new tests, no behavior change.
1930
1931         * mathml/MathMLElement.h: Place MathMLElementTypeHelpers.h to the bottom of the
1932         file so that functions from that file can use the Element|Node classes and other things defined in this file.
1933         * mathml/MathMLMencloseElement.h:
1934         (WebCore::toMathMLMencloseElement): Deleted.
1935         * mathml/mathtags.in:
1936
1937 2014-07-06  Yusuke Suzuki  <utatane.tea@gmail.com>
1938
1939         CSS JIT: Ensure resolvingMode size is 1 byte
1940         https://bugs.webkit.org/show_bug.cgi?id=134662
1941
1942         Reviewed by Benjamin Poulain.
1943
1944         Since selector compiler uses MacroAssembler::branch8 for SelectorChecker::Mode,
1945         we should ensure sizeof(SelectorChecker::Mode) == 1.
1946         Old implementation works only in little endianness environments.
1947
1948         * css/ElementRuleCollector.cpp:
1949         (WebCore::ElementRuleCollector::matchedResult):
1950         (WebCore::ElementRuleCollector::matchedRuleList):
1951         (WebCore::ElementRuleCollector::collectMatchingRules):
1952         (WebCore::ElementRuleCollector::sortAndTransferMatchedRules):
1953         (WebCore::ElementRuleCollector::hasAnyMatchingRules):
1954         * css/ElementRuleCollector.h:
1955         (WebCore::ElementRuleCollector::ElementRuleCollector):
1956         * css/SelectorChecker.cpp:
1957         (WebCore::SelectorChecker::match):
1958         (WebCore::SelectorChecker::matchRecursively):
1959         (WebCore::SelectorChecker::checkOne):
1960         * css/SelectorChecker.h:
1961         * css/StyleInvalidationAnalysis.cpp:
1962         (WebCore::invalidateStyleRecursively):
1963         * css/StyleResolver.cpp:
1964         (WebCore::StyleResolver::pseudoStyleRulesForElement):
1965         * css/StyleResolver.h:
1966         (WebCore::checkRegionSelector):
1967         * cssjit/SelectorCompiler.cpp:
1968         (WebCore::SelectorCompiler::SelectorCodeGenerator::jumpIfNotResolvingStyle):
1969         (WebCore::SelectorCompiler::elementIsActiveForStyleResolution):
1970         (WebCore::SelectorCompiler::elementIsHoveredForStyleResolution):
1971         * dom/SelectorQuery.cpp:
1972         (WebCore::SelectorDataList::selectorMatches):
1973
1974 2014-07-06  Antti Koivisto  <antti@apple.com>
1975
1976         Don't throttle layer flushes when the main resource is a GIF
1977         https://bugs.webkit.org/show_bug.cgi?id=134650
1978         <rdar://problem/17490712>
1979
1980         Reviewed by Simon Fraser.
1981
1982         Avoid throttling big image animations.
1983
1984         * page/FrameView.cpp:
1985         (WebCore::determineLayerFlushThrottleState):
1986
1987             Disable for image documents.
1988
1989         (WebCore::FrameView::disableLayerFlushThrottlingTemporarilyForInteraction):
1990         (WebCore::FrameView::updateLayerFlushThrottling):
1991
1992             Refactor a bit.
1993
1994         * page/LayerFlushThrottleState.h:
1995
1996             Rename the flag.
1997
1998 2014-07-06  Andreas Kling  <akling@apple.com>
1999
2000         Inline NodeListsNodeData helpers more aggressively.
2001         <https://webkit.org/b/134654>
2002
2003         Get rid of call overhead costs when retrieving NodeLists and
2004         HTMLCollections from Document. These functions have very few clients,
2005         so there's no real reason to keep them out of line.
2006
2007         Reviewed by Antti Koivisto.
2008
2009         * dom/NodeRareData.h:
2010         (WebCore::NodeListsNodeData::addCacheWithAtomicName):
2011         (WebCore::NodeListsNodeData::addCacheWithName):
2012
2013 2014-07-06  Andreas Kling  <akling@apple.com>
2014
2015         Inline Node's rare data getters.
2016         <https://webkit.org/b/134652>
2017
2018         There was a fair bit of overhead when calling Node::ensureRareData()
2019         from Document's collection getters. This patch inlines the case where
2020         we've already materialized the rare data.
2021
2022         Reviewed by Antti Koivisto.
2023
2024         * dom/Node.cpp:
2025         (WebCore::Node::materializeRareData):
2026         (WebCore::Node::rareData): Deleted.
2027         (WebCore::Node::ensureRareData): Deleted.
2028         * dom/Node.h:
2029         * dom/NodeRareData.h:
2030         (WebCore::Node::rareData):
2031         (WebCore::Node::ensureRareData):
2032
2033 2014-07-05  Dean Jackson  <dino@apple.com>
2034
2035         [iOS] MediaDocument should set a viewport
2036         https://bugs.webkit.org/show_bug.cgi?id=134658
2037         <rdar://problem/17552339>
2038
2039         Reviewed by Zalan Bujtas.
2040
2041         A MediaDocument should fill the viewport, so set its
2042         initial-scale to 1 and make it non-zoomable.
2043
2044         * html/MediaDocument.cpp:
2045         (WebCore::MediaDocumentParser::createDocumentStructure): Create
2046         a <head> element and a <meta> element with a 1:1 viewport.
2047
2048 2014-07-05  Commit Queue  <commit-queue@webkit.org>
2049
2050         Unreviewed, rolling out r170825.
2051         https://bugs.webkit.org/show_bug.cgi?id=134653
2052
2053         broke debug builds (Requested by kling on #webkit).
2054
2055         Reverted changeset:
2056
2057         "Generate isMathMLFoo() toMathMLFoo() by using NODE_TYPE_CASTS
2058         macro and generateTypeHelpers"
2059         https://bugs.webkit.org/show_bug.cgi?id=134647
2060         http://trac.webkit.org/changeset/170825
2061
2062 2014-07-05  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2063
2064         Generate isMathMLFoo() toMathMLFoo() by using NODE_TYPE_CASTS macro and generateTypeHelpers
2065         https://bugs.webkit.org/show_bug.cgi?id=134647
2066
2067         Reviewed by Andreas Kling.
2068
2069         Use type cast macro and generation tag for isFoo() and toFoo() in some MathML classes.
2070
2071         No new tests, no behavior change.
2072
2073         * mathml/MathMLElement.h: Place MathMLElementTypeHelpers.h to the bottom of the
2074         file so that functions from that file can use the Element|Node classes and other things defined in this file.
2075         (WebCore::isMathMLElement): Deleted.
2076         * mathml/MathMLMencloseElement.h:
2077         (WebCore::toMathMLMencloseElement): Deleted.
2078         * mathml/mathtags.in:
2079
2080 2014-07-04  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
2081
2082         [EFL] Support Geolocation
2083         https://bugs.webkit.org/show_bug.cgi?id=134439
2084
2085         Reviewed by Antonio Gomes.
2086
2087         Add geoclue files in order to enable geolocation for EFL port.
2088
2089         * PlatformEfl.cmake:
2090
2091 2014-07-04  Benjamin Poulain  <benjamin@webkit.org>
2092
2093         CSS JIT: compile :in-range and :out-of-range
2094         https://bugs.webkit.org/show_bug.cgi?id=134639
2095
2096         Reviewed by Andreas Kling.
2097
2098         The pseudo class :in-range and :out-of-range are trivial non-inlined selectors.
2099
2100         * css/SelectorChecker.cpp:
2101         (WebCore::SelectorChecker::checkOne):
2102         * css/SelectorCheckerTestFunctions.h:
2103         (WebCore::isInRange):
2104         (WebCore::isOutOfRange):
2105         * cssjit/SelectorCompiler.cpp:
2106         (WebCore::SelectorCompiler::addPseudoClassType):
2107
2108 2014-07-04  Andreas Kling  <akling@apple.com>
2109
2110         Fast path for jsStringWithCache() when asked for the same string repeatedly.
2111         <https://webkit.org/b/134635>
2112
2113         Reviewed by Darin Adler.
2114
2115         * WebCore.exp.in:
2116         * bindings/js/JSDOMBinding.cpp:
2117         (WebCore::jsStringWithCache): Deleted.
2118         * bindings/js/JSDOMBinding.h:
2119         (WebCore::JSValueTraits<String>::arrayJSValue):
2120         (WebCore::jsStringWithCache): Deleted.
2121         * bridge/c/c_utility.cpp:
2122         (JSC::Bindings::convertNPVariantToValue):
2123         * loader/cache/CachedResourceHandle.h:
2124
2125 2014-07-04  Zalan Bujtas  <zalan@apple.com>
2126
2127         Subpixel rendering: ebay.com rotating billboard on the main page has cut off buttons.
2128         https://bugs.webkit.org/show_bug.cgi?id=134636
2129         <rdar://problem/17529112>
2130
2131         Reviewed by Darin Adler.
2132
2133         Missing pixel snapping when clipping the transparency layer for BackgroundBleedUseTransparencyLayer.
2134         Clipping needs to take the same snapping strategy as the rest of the painting code.
2135
2136         Test: fast/borders/border-radius-on-subpixel-position-non-hidpi.html
2137
2138         * rendering/RenderBox.cpp:
2139         (WebCore::RenderBox::paintBoxDecorations):
2140
2141 2014-05-10  Andreas Kling  <akling@apple.com>
2142
2143         QualifiedName should use RefPtr<QualifiedNameImpl> internally.
2144         <https://webkit.org/b/132796>
2145
2146         We were not inlining ~QualifiedName() and it was showing up in some
2147         profiles (over 1% on dom-attr.html.)
2148
2149         This patch modernizes QualifiedName a bit by using RefPtr to store
2150         the QualifiedNameImpl instead of doing manual ref() and deref()
2151         in out-of-line functions.
2152
2153         Reviewed by Darin Adler.
2154
2155         * WebCore.exp.in:
2156         * dom/QualifiedName.cpp:
2157         (WebCore::QNameComponentsTranslator::translate):
2158         (WebCore::QualifiedName::QualifiedName):
2159         (WebCore::QualifiedName::~QualifiedName): Deleted.
2160         (WebCore::QualifiedName::deref): Deleted.
2161         * dom/QualifiedName.h:
2162         (WebCore::QualifiedName::QualifiedNameImpl::create):
2163         (WebCore::QualifiedName::QualifiedName):
2164         (WebCore::QualifiedName::isHashTableDeletedValue):
2165         (WebCore::QualifiedName::~QualifiedName):
2166         (WebCore::QualifiedName::operator=):
2167         (WebCore::QualifiedName::impl):
2168         (WebCore::QualifiedName::ref): Deleted.
2169
2170 2014-07-04  Dean Jackson  <dino@apple.com>
2171
2172         [iOS] [Media] Prevent page zoom if it started in the video element
2173         https://bugs.webkit.org/show_bug.cgi?id=134640
2174         <rdar://problem/16630794>
2175
2176         Reviewed by Sam Weinig.
2177
2178         In r170572 I attempted to prevent the page zoom from happening
2179         when the user was attempting to fullscreen the video. That fix was
2180         not complete, because it relied on the touchstart event being
2181         sent before a gesture event. It turns out this doesn't always happen,
2182         so we need to apply the same check in the gesture event - if the
2183         gesture began with two fingers on the media, don't do a page zoom.
2184         The exception is zooming out, if you've managed to make the video
2185         cover the entire screen (also, the pinch-out-of-fullscreen gesture
2186         does not get handled by WebKit).
2187
2188         * Modules/mediacontrols/mediaControlsiOS.js:
2189         (ControllerIOS.prototype.handleBaseGestureChange): Add a check to
2190         make sure our pinch didn't start in the video element.
2191
2192 2014-07-04  Hunseop Jeong  <hs85.jeong@samsung.com>
2193
2194         [CoordinatedGraphics] Use the C++11 syntax to clean-up for loops
2195         https://bugs.webkit.org/show_bug.cgi?id=134594
2196
2197         Reviewed by Darin Adler.
2198
2199         Convert a couple of loops to range-based for
2200
2201         No change in functionality, no new tests.
2202
2203         * platform/graphics/TiledBackingStore.cpp:
2204         (WebCore::TiledBackingStore::updateTileBuffers):
2205         (WebCore::TiledBackingStore::resizeEdgeTiles):
2206         (WebCore::TiledBackingStore::setKeepRect):
2207         * platform/graphics/texmap/coordinated/CompositingCoordinator.cpp:
2208         (WebCore::CompositingCoordinator::~CompositingCoordinator):
2209         (WebCore::CompositingCoordinator::flushPendingImageBackingChanges):
2210         (WebCore::CompositingCoordinator::setVisibleContentsRect):
2211         (WebCore::CompositingCoordinator::purgeBackingStores):
2212         * platform/graphics/texmap/coordinated/CoordinatedBackingStore.cpp:
2213         (WebCore::CoordinatedBackingStore::removeAllTiles):
2214         (WebCore::CoordinatedBackingStore::paintToTextureMapper):
2215         (WebCore::CoordinatedBackingStore::drawBorder):
2216         (WebCore::CoordinatedBackingStore::drawRepaintCounter):
2217         (WebCore::CoordinatedBackingStore::commitTileOperations):
2218         * platform/graphics/texmap/coordinated/CoordinatedGraphicsScene.cpp:
2219         (WebCore::CoordinatedGraphicsScene::adjustPositionForFixedLayers):
2220         (WebCore::CoordinatedGraphicsScene::commitPendingBackingStoreOperations):
2221
2222 2014-07-04  Andreas Kling  <akling@apple.com>
2223
2224         CTTE: ScriptElement always has a HTTPScriptElement or SVGScriptElement.
2225         <https://webkit.org/b/134637>
2226
2227         Have ScriptElement store the pointer to the more specific subclass
2228         element as a reference. Also made the constructor protected since
2229         nobody should instantiate ScriptElement directly.
2230
2231         Reviewed by Antti Koivisto.
2232
2233         * dom/ScriptElement.cpp:
2234         (WebCore::ScriptElement::ScriptElement):
2235         (WebCore::ScriptElement::childrenChanged):
2236         (WebCore::ScriptElement::dispatchErrorEvent):
2237         (WebCore::ScriptElement::prepareScript):
2238         (WebCore::ScriptElement::requestScript):
2239         (WebCore::ScriptElement::executeScript):
2240         (WebCore::ScriptElement::notifyFinished):
2241         (WebCore::ScriptElement::ignoresLoadRequest):
2242         (WebCore::ScriptElement::scriptContent):
2243         * dom/ScriptElement.h:
2244         (WebCore::ScriptElement::element):
2245         * dom/ScriptRunner.cpp:
2246         (WebCore::ScriptRunner::queueScriptForExecution):
2247         * html/HTMLScriptElement.cpp:
2248         (WebCore::HTMLScriptElement::HTMLScriptElement):
2249         * svg/SVGScriptElement.cpp:
2250         (WebCore::SVGScriptElement::SVGScriptElement):
2251
2252 2014-07-04  Julien Quint  <pom@graougraou.com>
2253
2254         input type=range element should only fire change events after committing a  value
2255         https://bugs.webkit.org/show_bug.cgi?id=134545
2256
2257         Reviewed by Dean Jackson.
2258
2259         A "change" event was fired every time the slider thumb element was dragged
2260         by the user. The "change" event is now fired only after the thumb
2261         element has stopped moving; previously, both "input" and "change" events
2262         where dispatched while changes were being made. This new behavior is
2263         consistent with the specification (cf.
2264         http://www.whatwg.org/specs/web-apps/current-work/multipage/common-input-element-attributes.html#event-input-change),
2265         as well as other implementations such as Firefox and Chrome.
2266
2267         * Modules/mediacontrols/mediaControlsApple.js:
2268         (Controller.prototype.createControls): Listen to the "input" event
2269         rather than the "change" event for the timeline control in order to
2270         keep track of value changes when the user is dragging the thumb.
2271         * accessibility/AccessibilitySlider.cpp:
2272         (WebCore::AccessibilitySlider::setValue): Dispatch "change" event while
2273         setting the new value rather than dispatching later, since setting the
2274         value now clears the change flag.
2275         * html/RangeInputType.cpp:
2276         (WebCore::RangeInputType::setValue): Update the text value of the
2277         control in the case when no event is to be dispatched, so that this
2278         value can be checked the next time a "change" event dispatch is
2279         requested.
2280         * html/shadow/SliderThumbElement.cpp:
2281         (WebCore::SliderThumbElement::setPositionFromPoint): Removed the
2282         dispatch of the "change" event, and no longer track the text value of
2283         the element as a result of dispatching a "change" event.
2284         (WebCore::SliderThumbElement::stopDragging): Dispatch the "change" event
2285         on completing the drag.
2286
2287 2014-07-04  Andreas Kling  <akling@apple.com>
2288
2289         CTTE: ApplicationCacheHost always has a DocumentLoader.
2290         <https://webkit.org/b/b/128435>
2291
2292         Have ApplicationCacheHost store the backpointer to its owning
2293         DocumentLoader as a reference.
2294
2295         Reviewed by Sam Weinig.
2296
2297         * loader/DocumentLoader.cpp:
2298         (WebCore::DocumentLoader::DocumentLoader):
2299         (WebCore::DocumentLoader::startLoadingMainResource):
2300         * loader/appcache/ApplicationCacheHost.cpp:
2301         (WebCore::ApplicationCacheHost::ApplicationCacheHost):
2302         (WebCore::ApplicationCacheHost::~ApplicationCacheHost):
2303         (WebCore::ApplicationCacheHost::selectCacheWithoutManifest):
2304         (WebCore::ApplicationCacheHost::selectCacheWithManifest):
2305         (WebCore::ApplicationCacheHost::maybeLoadMainResource):
2306         (WebCore::ApplicationCacheHost::maybeLoadFallbackForMainResponse):
2307         (WebCore::ApplicationCacheHost::maybeLoadFallbackForMainError):
2308         (WebCore::ApplicationCacheHost::failedLoadingMainResource):
2309         (WebCore::ApplicationCacheHost::finishedLoadingMainResource):
2310         (WebCore::ApplicationCacheHost::maybeLoadResource):
2311         (WebCore::ApplicationCacheHost::maybeLoadFallbackForError):
2312         (WebCore::ApplicationCacheHost::maybeLoadSynchronously):
2313         (WebCore::ApplicationCacheHost::notifyDOMApplicationCache):
2314         (WebCore::ApplicationCacheHost::stopDeferringEvents):
2315         (WebCore::ApplicationCacheHost::scheduleLoadFallbackResourceFromApplicationCache):
2316         (WebCore::ApplicationCacheHost::update):
2317         (WebCore::ApplicationCacheHost::swapCache):
2318         (WebCore::ApplicationCacheHost::abort):
2319         (WebCore::ApplicationCacheHost::isApplicationCacheEnabled):
2320         * loader/appcache/ApplicationCacheHost.h:
2321         (WebCore::ApplicationCacheHost::documentLoader): Deleted.
2322
2323 2014-07-04  Antti Koivisto  <antti@apple.com>
2324
2325         REGRESSION(r159345): Lines are clipped between pages when printing web content from Safari
2326         https://bugs.webkit.org/show_bug.cgi?id=134633
2327         <rdar://problem/17088936>
2328
2329         Reviewed by Zalan Bujtas.
2330
2331         Test: printing/simple-lines-break.html
2332
2333         * rendering/SimpleLineLayout.cpp:
2334         (WebCore::SimpleLineLayout::canUseFor):
2335
2336             Disable simple lines if global pagination is enabled.
2337             Printing does not currently use flow threads which we already test for.
2338
2339 2014-07-04  Zan Dobersek  <zdobersek@igalia.com>
2340
2341         Unreviewed. Fixing the build after r170795.
2342
2343         Mac didn't like that change as it seems to properly typedef long long to int64_t.
2344         That's not the case on 64-bit Linux (for some reason), so I'm changing the type of
2345         the two long long variables in FormDataElement to int64_t to avoid any further trouble.
2346
2347         * platform/network/FormData.h:
2348
2349 2014-07-03  Brady Eidson  <beidson@apple.com>
2350
2351         When showing the selection menu, include menu options for all selected phone numbers.
2352         <rdar://problem/16983434>, <rdar://problem/16874568>, and https://bugs.webkit.org/show_bug.cgi?id=134613
2353
2354         Reviewed by Tim Horton.
2355
2356         * English.lproj/Localizable.strings:
2357
2358         * dom/Range.cpp:
2359         (WebCore::rangesOverlap): Fix reversed arguments that caused some types of overlaps to not be detected.
2360
2361 2014-07-03  Daniel Bates  <dabates@apple.com>
2362
2363         Add WTF::move()
2364         https://bugs.webkit.org/show_bug.cgi?id=134500
2365
2366         Rubber-stamped by Anders Carlsson.
2367
2368         Substitute WTF::move() for std::move().
2369
2370         * Modules/battery/NavigatorBattery.cpp:
2371         * Modules/encryptedmedia/MediaKeys.cpp:
2372         * Modules/gamepad/NavigatorGamepad.cpp:
2373         * Modules/gamepad/deprecated/NavigatorGamepad.cpp:
2374         * Modules/geolocation/NavigatorGeolocation.cpp:
2375         * Modules/indexeddb/DOMWindowIndexedDatabase.cpp:
2376         * Modules/indexeddb/IDBDatabaseBackend.cpp:
2377         * Modules/indexeddb/IDBDatabaseBackend.h:
2378         [...]
2379
2380 2014-07-03  Benjamin Poulain  <bpoulain@apple.com>
2381
2382         [iOS][WK2] Fix small bugs of dynamicViewportSizeUpdate that were causing inaccuracies in the computed target
2383         https://bugs.webkit.org/show_bug.cgi?id=134582
2384
2385         Reviewed by Tim Horton.
2386
2387         * WebCore.exp.in:
2388         * platform/ScrollView.h:
2389         (WebCore::ScrollView::unobscuredContentSize):
2390         * platform/ios/ScrollViewIOS.mm:
2391         (WebCore::ScrollView::unobscuredContentRect):
2392         (WebCore::ScrollView::setUnobscuredContentSize):
2393
2394 2014-07-02  Brent Fulgham  <bfulgham@apple.com>
2395
2396         [Mac] WebKit1 WebView iframe not responding to scroll gestures
2397         https://bugs.webkit.org/show_bug.cgi?id=134569
2398         <rdar://problem/17309008>
2399
2400         Reviewed by Simon Fraser.
2401
2402         New test: platform/mac/fast/scrolling/scroll-iframe-fragment.html
2403
2404         WK1 scrollable areas are implemented on top of PlatformWidget (i.e., NSView) objects. They only
2405         perform a scroll operation if the EventHandler::handleWheelEvent method returns 'false', indicating
2406         that the Cocoa layer should do scrolling work. This is in contrast to WK2, where the scrolling is
2407         handled separately and we return 'true' from EventHandler::handleWheelEvent to indicate that the
2408         scroll event has been handled and that no other work is needed. This mismatch in the meaning of
2409         EventHandler::handleWheelEvent means that WK1 views do not properly handle scroll wheel events.
2410
2411         To correct this, we need to see if we are acting on an Element backed by a PlatformWidget. When we
2412         are (the WK1 case) we want to return 'false' to instruct the Cocoa object to do the scroll operation.
2413         To implement proper latching behavior in WK1, we must return 'true' in the case where we are latched
2414         to a PlatformWidget, and that PlatformWidget did not start the scroll gesture at the edge of
2415         the region in the dominant scroll direction.
2416
2417         Finally, additional logic was needed to determine whether the PlatformWidget is scrolled to the
2418         edge in the dominant scroll direction.
2419
2420         * page/EventHandler.cpp:
2421         (WebCore::EventHandler::platformCompleteWheelEvent): Added stub.
2422         (WebCore::EventHandler::platformCompletePlatformWidgetWheelEvent): Added stub.
2423         (WebCore::EventHandler::handleWheelEvent): Call new platformCompletePlatformWidgetWheelEvent if the
2424         event was handled for a PlatformWidget.
2425         * page/EventHandler.h:
2426         * page/mac/EventHandlerMac.mm:
2427         (WebCore::widgetForEventTarget): Added helper function.
2428         (WebCore::scrollViewForEventTarget): Added helper function.
2429         (WebCore::eventTargetIsPlatformWidget): Added helper function.
2430         (WebCore::EventHandler::platformPrepareForWheelEvents): Check whether we are at a scrolled edge
2431         for the NSView (PlatformWidget) case.
2432         (WebCore::EventHandler::platformCompleteWheelEvent): Return 'false' when handling WK1-style scrolls.
2433         (WebCore::EventHandler::platformCompletePlatformWidgetWheelEvent): Added. We generally want to
2434         return false so the PlatformWidget handles scrolling, except in the case where we are latched to
2435         a PlatformWidget, which has already performed its scroll, and we want to avoid propegating a
2436         scroll event to an enclosing region.
2437
2438 2014-07-03  Pratik Solanki  <psolanki@apple.com>
2439
2440         Preserve old behavior of creating an NSURLRequest of the exact same type as passed to us
2441         https://bugs.webkit.org/show_bug.cgi?id=134605
2442         <rdar://problem/17544641>
2443
2444         Reviewed by Andreas Kling.
2445
2446         We have client code that passes us a subclass of NSURLRequest to load resource. Later when
2447         we call willSendRequest, they test to make sure they get an object of the same type. My
2448         optimization in r170642 broke this path when I cleared out the NSURLRequest object. We
2449         already had code in updateNSURLRequest() that was taking this quirk into account but I broke
2450         that. Fix it by reverting to old behavior for such clients.
2451
2452         No new tests though we need one. I will add it later.
2453
2454         * platform/network/cf/ResourceRequest.h:
2455         * platform/network/cf/ResourceRequestCFNet.cpp:
2456         (WebCore::ResourceRequest::doUpdatePlatformRequest):
2457         (WebCore::ResourceRequest::doUpdatePlatformHTTPBody):
2458         (WebCore::ResourceRequest::setStorageSession):
2459         * platform/network/ios/ResourceRequestIOS.mm:
2460         (WebCore::ResourceRequest::updateNSURLRequest):
2461         (WebCore::ResourceRequest::clearOrUpdateNSURLRequest):
2462
2463 2014-07-02  Anders Carlsson  <andersca@apple.com>
2464
2465         Stop using EncoderAdapter/DecoderAdapter for FormData
2466         https://bugs.webkit.org/show_bug.cgi?id=134571
2467
2468         Reviewed by Andreas Kling.
2469
2470         * WebCore.exp.in:
2471         * platform/network/FormData.cpp:
2472         (WebCore::encodeElement): Deleted.
2473         (WebCore::decodeElement): Deleted.
2474         (WebCore::FormData::encode): Deleted.
2475         (WebCore::FormData::decode): Deleted.
2476         * platform/network/FormData.h:
2477         (WebCore::FormDataElement::encode):
2478         (WebCore::FormDataElement::decode):
2479         (WebCore::FormData::encode):
2480         (WebCore::FormData::decode):
2481
2482 2014-07-03  Brady Eidson  <beidson@apple.com>
2483
2484         Selection rects sent to ServicesOverlayController are wrong.
2485         <rdar://problem/16727796> and https://bugs.webkit.org/show_bug.cgi?id=134568
2486
2487         Reviewed by Darin Adler (and Tim Horton and Ryosuke Niwa).
2488
2489         * WebCore.exp.in:
2490
2491         Update the gatherer to keep GapRects separate from LayoutRects:
2492         * editing/SelectionRectGatherer.cpp:
2493         (WebCore::SelectionRectGatherer::addRects):
2494         (WebCore::SelectionRectGatherer::Notifier::~Notifier):
2495         (WebCore::SelectionRectGatherer::clearAndCreateNotifier):
2496         * editing/SelectionRectGatherer.h:
2497
2498         * page/EditorClient.h:
2499         (WebCore::EditorClient::selectionRectsDidChange): Updated to take LayoutRects and GapRects separately.
2500
2501         Change RenderSelectionInfo to also hang on to the individual rects that formed the final bounding rect:
2502         * rendering/RenderSelectionInfo.h:
2503         (WebCore::RenderSelectionInfo::RenderSelectionInfo): If the RenderObject is a RenderText, then call
2504           collectSelectionRectsForLineBoxes instead of selectionRectForRepaint.
2505         (WebCore::RenderSelectionInfo::rects):
2506
2507         * rendering/RenderText.cpp:
2508         (WebCore::RenderText::collectSelectionRectsForLineBoxes): Added
2509         (WebCore::RenderText::selectionRectForRepaint):
2510         * rendering/RenderText.h:
2511
2512         * rendering/RenderTextLineBoxes.cpp:
2513         (WebCore::RenderTextLineBoxes::collectSelectionRectsForRange): Added
2514         * rendering/RenderTextLineBoxes.h:
2515
2516         * rendering/RenderView.cpp:
2517         (WebCore::RenderView::setSubtreeSelection): Add the list of rects to the gatherer instead of just
2518           the bounding rect.
2519
2520 2014-07-03  Brady Eidson  <beidson@apple.com>
2521
2522         Possible crash in IconDatabase in WebCore::IconDatabase::dispatchDidRemoveAllIconsOnMainThread
2523         <rdar://problem/17437687> and https://bugs.webkit.org/show_bug.cgi?id=134517
2524
2525         Reviewed by Eric Carlson.
2526
2527         Since WebCore::IconDatabase is not RefCounted there’s no obvious way to keep it alive until it is
2528         truly no longer needed.
2529
2530         This isn’t generally a problem because they are usually a singleton that lasts the lifetime of the process.
2531
2532         In the WebKit2 case, WebCore::IconDatabases can come and go as their owning WebIconDatabases come and go.
2533
2534         So we can rely on WebIconDatabase to handle the lifetime appropriately.
2535
2536         * loader/icon/IconDatabase.cpp:
2537         (WebCore::IconDatabase::close): If the database is actually closed, notify the client.
2538         (WebCore::IconDatabase::IconDatabase):
2539         (WebCore::IconDatabase::isOpen): Take into account whether there’s any main thread callbacks, plus what
2540           is covered by isOpenBesidesMainThreadCallbacks.
2541         (WebCore::IconDatabase::isOpenBesidesMainThreadCallbacks): Take into account whether the sync thread
2542           is still running and whether or not the database file is still open.
2543         (WebCore::IconDatabase::checkClosedAfterMainThreadCallback): Checks to see if the database has just
2544           become 100% closed and - if so - notify the client.
2545         (WebCore::IconDatabase::dispatchDidImportIconURLForPageURLOnMainThread): Updated to increment the main thread
2546           callback count and to call checkClosedAfterMainThreadCallback when done.
2547         (WebCore::IconDatabase::dispatchDidImportIconDataForPageURLOnMainThread): Ditto.
2548         (WebCore::IconDatabase::dispatchDidRemoveAllIconsOnMainThread): Ditto.
2549         (WebCore::IconDatabase::dispatchDidFinishURLImportOnMainThread): Ditto.
2550         * loader/icon/IconDatabase.h:
2551
2552         * loader/icon/IconDatabaseClient.h:
2553         (WebCore::IconDatabaseClient::didClose): Added. For the IconDatabase to tell its client it is 100% closed.
2554
2555 2014-07-03  Chris Fleizach  <cfleizach@apple.com>
2556
2557         AX: VoiceOver does not read aria-expanded attribute on controls in Safari
2558         https://bugs.webkit.org/show_bug.cgi?id=134527
2559
2560         Reviewed by Mario Sanchez Prada.
2561
2562         Expose whether the ARIA expanded property is present or not.
2563
2564         * accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
2565         (-[WebAccessibilityObjectWrapper accessibilitySupportsARIAExpanded]):
2566
2567 2014-07-02  Jon Honeycutt  <jhoneycutt@apple.com>
2568
2569         Removing an element with CSS -webkit-overflow-scrolling: touch breaks
2570         subsequent touch events
2571
2572         <https://bugs.webkit.org/show_bug.cgi?id=134584>
2573         <rdar://problem/17202021>
2574
2575         This issue is caused by the order of teardown for RenderLayers.
2576
2577         RenderLayer clears its m_backing in its destructor, calling the
2578         RenderLayerBacking destructor. The RenderLayerBacking destructor calls
2579         RenderLayerCompositor::willRemoveScrollingLayer(), which tries to
2580         access the RenderLayer's backing that has already been cleared. This
2581         causes us to skip the call to ChromeClient::removeScrollingLayer(),
2582         resulting in a stale view that can block touch events.
2583
2584         Reviewed by Simon Fraser.
2585
2586         * rendering/RenderLayerBacking.cpp:
2587         (WebCore::RenderLayerBacking::updateScrollingLayers):
2588         When calling willRemoveScrollingLayerWithBacking(), pass both the layer
2589         and the RenderLayerBacking.
2590
2591         * rendering/RenderLayerCompositor.cpp:
2592         (WebCore::RenderLayerCompositor::willRemoveScrollingLayerWithBacking):
2593         Renamed. Use the passed RenderLayerBacking, as the RenderLayer's
2594         backing might have already been cleared.
2595         (WebCore::RenderLayerCompositor::willRemoveScrollingLayer): Deleted.
2596
2597         * rendering/RenderLayerCompositor.h:
2598         Renamed willRemoveScrollingLayer() to
2599         willRemoveScrollingLayerWithBacking() and added a parameter.
2600
2601 2014-07-01  Mark Rowe  <mrowe@apple.com>
2602
2603         Ensure that the WebKit bundle version in the user agent string continues to match the current format.
2604         <https://webkit.org/b/134524> / <rdar://problem/17447771>
2605
2606         Reviewed by Simon Fraser.
2607
2608         * page/cocoa/UserAgent.h:
2609         * page/cocoa/UserAgent.mm:
2610         (WebCore::userVisibleWebKitBundleVersionFromFullVersion): Updated to take an NSString now that it's internal
2611         to the file.
2612         (WebCore::userAgentBundleVersionFromFullVersionString): Limit the bundle version included in the user agent
2613         string to three components.
2614         * page/ios/UserAgentIOS.mm:
2615         (WebCore::standardUserAgentWithApplicationName): Update to call userAgentBundleVersionFromFullVersionString.
2616         * page/mac/UserAgentMac.mm:
2617         (WebCore::standardUserAgentWithApplicationName): Ditto.
2618
2619 2014-07-01  Mark Rowe  <mrowe@apple.com>
2620
2621         <https://webkit.org/b/134522> Remove duplication in code that prepares the user agent string on Mac and iOS
2622
2623         Reviewed by Simon Fraser.
2624
2625         * page/cocoa/UserAgent.h:
2626         * page/cocoa/UserAgent.mm:
2627         (WebCore::userVisibleWebKitBundleVersionFromFullVersion): Moved from WebKit2.
2628         * page/ios/UserAgentIOS.mm:
2629         (WebCore::standardUserAgentWithApplicationName): Pass the WebKit bundle version through userVisibleWebKitBundleVersionFromFullVersion
2630         before including it in the user agent string.
2631         * page/mac/UserAgentMac.mm:
2632         (WebCore::standardUserAgentWithApplicationName): Ditto.
2633
2634 2014-07-02  Mark Rowe  <mrowe@apple.com>
2635
2636         <https://webkit.org/b/134521> iOS should use shared code to determine the system marketing version
2637
2638         Reviewed by Simon Fraser.
2639
2640         * WebCore.xcodeproj/project.pbxproj: Add the new files, and sort the groups they're in.
2641         * page/cocoa/UserAgent.h: Copied from Source/WebCore/page/mac/UserAgent.h.
2642         * page/cocoa/UserAgent.mm: Renamed from Source/WebCore/page/mac/UserAgent.h.
2643         Move systemMarketingVersionForUserAgentString to a location where it can be shared between
2644         Mac and iOS.
2645         * page/ios/UserAgentIOS.mm:
2646         (WebCore::standardUserAgentWithApplicationName): Switch to systemMarketingVersionForUserAgentString.
2647         * page/mac/UserAgentMac.mm:
2648         * platform/cocoa/SystemVersion.h: Renamed from Source/WebCore/platform/mac/SystemVersionMac.h.
2649         * platform/cocoa/SystemVersion.mm: Renamed from Source/WebCore/platform/mac/SystemVersionMac.mm.
2650         Move to a location that makes it clear this is shared between Mac and iOS. Enable the modern Mac
2651         codepath for iOS as well.
2652
2653 2014-07-02  Anders Carlsson  <andersca@apple.com>
2654
2655         Remove keyed coding from FormData
2656         https://bugs.webkit.org/show_bug.cgi?id=134565
2657
2658         Reviewed by Tim Horton.
2659
2660         * platform/network/FormData.cpp:
2661         (WebCore::encodeElement): Deleted.
2662         (WebCore::decodeElement): Deleted.
2663         (WebCore::FormData::encode): Deleted.
2664         (WebCore::FormData::decode): Deleted.
2665         * platform/network/FormData.h:
2666
2667 2014-07-02  Anders Carlsson  <andersca@apple.com>
2668
2669         Get rid of the HistoryItem encoding/decoding in WebCore
2670         https://bugs.webkit.org/show_bug.cgi?id=134564
2671
2672         Reviewed by Beth Dakin.
2673
2674         * WebCore.exp.in:
2675         * history/HistoryItem.cpp:
2676         (WebCore::HistoryItem::encodeBackForwardTree): Deleted.
2677         (WebCore::encodeRect): Deleted.
2678         (WebCore::encodeSize): Deleted.
2679         (WebCore::HistoryItem::encodeBackForwardTreeNode): Deleted.
2680         (WebCore::DecodeRecursionStackElement::DecodeRecursionStackElement): Deleted.
2681         (WebCore::decodeRect): Deleted.
2682         (WebCore::decodeSize): Deleted.
2683         (WebCore::HistoryItem::decodeBackForwardTree): Deleted.
2684         * history/HistoryItem.h:
2685
2686 2014-07-02  Alex Christensen  <achristensen@webkit.org>
2687
2688         [WebGL] Fix crash when glsl expressions are too complex.
2689         https://bugs.webkit.org/show_bug.cgi?id=134554
2690         <rdar://problem/17540462>
2691
2692         Reviewed by Dean Jackson.
2693
2694         This fixes the Khronos test conformance/glsl/bugs/long-expressions-should-not-crash.html.
2695
2696         * platform/graphics/opengl/Extensions3DOpenGLCommon.cpp:
2697         (WebCore::Extensions3DOpenGLCommon::getTranslatedShaderSourceANGLE):
2698         Add SH_LIMIT_EXPRESSION_COMPLEXITY to the ANGLE compile options.
2699
2700 2014-07-02  Benjamin Poulain  <benjamin@webkit.org>
2701
2702         WebCore JIT: rename registerCount to something less generic and add new types for lists of registers and lists of stack references
2703         https://bugs.webkit.org/show_bug.cgi?id=134552
2704
2705         Reviewed by Alex Christensen.
2706
2707         Little cleanup:
2708         -The name registerCount was a little too generic. Rename that to "maximumRegisterCount" to avoid confusion.
2709         -Add a new type RegisterVector for any vector holding registers. This is just to avoid repeating the inline
2710          size everywhere, no functional change.
2711         -Same idea for the stack: welcome StackReferenceVector!
2712
2713         * cssjit/FunctionCall.h:
2714         (WebCore::FunctionCall::saveAllocatedCallerSavedRegisters):
2715         Remove the appendVector here. It was unnecessarily cautious, StackAllocator already protect us
2716         from mistakes.
2717
2718         * cssjit/RegisterAllocator.h:
2719         (WebCore::RegisterAllocator::allocatedRegisters):
2720         * cssjit/SelectorCompiler.cpp:
2721         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateSelectorChecker):
2722         Changing from the count of "calleeSavedRegisterCount" to "maximumRegisterCount" will cause
2723         calleeSavedRegisterStackReferences to always overallocate.
2724         The code generator is never on the heap, so that should not change anything.
2725
2726         * cssjit/StackAllocator.h:
2727         (WebCore::StackAllocator::push):
2728         (WebCore::StackAllocator::pop):
2729
2730 2014-07-02  Dan Bernstein  <mitz@apple.com>
2731
2732         Remove some code only needed for versions of Safari that are no longer supported
2733         https://bugs.webkit.org/show_bug.cgi?id=134537
2734
2735         Reviewed by Anders Carlsson.
2736
2737         * bindings/objc/DOMHTML.mm:
2738         (-[DOMHTMLInputElement _rectOnScreen]): Deleted.
2739         (-[DOMHTMLInputElement _replaceCharactersInRange:withString:selectingFromIndex:]): Deleted.
2740         (-[DOMHTMLInputElement _selectedRange]): Deleted.
2741         * bindings/objc/DOMPrivate.h:
2742         * editing/EditorCommand.cpp:
2743         (WebCore::createCommandMap):
2744
2745 2014-07-02  Carlos Garcia Campos  <cgarcia@igalia.com>
2746
2747         [GStreamer] MediaPlayerPrivateGStreamer.cpp fails to build when VIDEO_TRACK is enabled and using GSTREAMER_MPEGTS
2748         https://bugs.webkit.org/show_bug.cgi?id=134548
2749
2750         Reviewed by Philippe Normand.
2751
2752         Include AtomicStringHash.h when ENABLE(VIDEO_TRACK) && USE(GSTREAMER_MPEGTS)
2753         since it's used by HashMap<AtomicString, RefPtr<InbandMetadataTextTrackPrivateGStreamer>>
2754
2755         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.h:
2756
2757 2014-07-02  Carlos Garcia Campos  <cgarcia@igalia.com>
2758
2759         [GTK] Unstable headers should include webkitdomdefines-unstable.h
2760         https://bugs.webkit.org/show_bug.cgi?id=134540
2761
2762         Reviewed by Martin Robinson.
2763
2764         Instead of the class header that can only be included from webkitdom.h.
2765
2766         * bindings/scripts/CodeGeneratorGObject.pm:
2767         (WriteData):
2768
2769 2014-07-02  Mario Sanchez Prada  <mario.prada@samsung.com>
2770
2771         REGRESSION(r170008): [GTK] Layout Tests fast/forms/option-constructor-selected.html and fast/forms/select-live-pseudo-selectors.html crash.
2772         https://bugs.webkit.org/show_bug.cgi?id=134399
2773
2774         Reviewed by Chris Fleizach.
2775
2776         Check that the value of optionIndex passed as a parameter is valid
2777         before calling didUpdateActiveOption() over the item of the menu
2778         list, to prevent crashes in platforms where renderers (and therefore
2779         accessibility objects) might not be created for items at the time
2780         of calling AccessibilityMenuList::didUpdateActiveOption().
2781
2782         * accessibility/AccessibilityMenuList.cpp:
2783         (WebCore::AccessibilityMenuList::didUpdateActiveOption): Added check.
2784
2785 2014-07-02  Carlos Garcia Campos  <cgarcia@igalia.com>
2786
2787         [GTK] Do not add forward declaration of class struct for interfaces
2788         https://bugs.webkit.org/show_bug.cgi?id=134297
2789
2790         Reviewed by Martin Robinson.
2791
2792         * PlatformGTK.cmake: Make files generated by
2793         gobject-generate-headers.pl script depend on it.
2794         * bindings/gobject/WebKitDOMEventTarget.h: Remove forward declaration.
2795         * bindings/gobject/WebKitDOMNodeFilter.h: Ditto
2796         * bindings/gobject/WebKitDOMXPathNSResolver.h: Ditto.
2797         * bindings/scripts/gobject-generate-headers.pl: Use Iface instead
2798         of Class suffix for interfaces.
2799
2800 2014-07-01  Carlos Garcia Campos  <cgarcia@igalia.com>
2801
2802         [GTK] XPathNSResolver should be exposed as an interface instead of as a class
2803         https://bugs.webkit.org/show_bug.cgi?id=134028
2804
2805         Reviewed by Gustavo Noronha Silva.
2806
2807         Add custom implementation of WebKitDOMXPathNSResolver to expose it
2808         as an interface instead of a class. It also includes an internal
2809         default implementation of the interface to support methods that
2810         create a XPathNSResolver using the native implementation.
2811
2812         * PlatformGTK.cmake: Add new files to compilation.
2813         * bindings/gobject/GObjectXPathNSResolver.cpp: Added.
2814         (WebCore::GObjectXPathNSResolver::~GObjectXPathNSResolver):
2815         (WebCore::GObjectXPathNSResolver::lookupNamespaceURI): Call
2816         webkit_dom_xpath_ns_resolver_lookup_namespace_uri().
2817         * bindings/gobject/GObjectXPathNSResolver.h: Added.
2818         (WebCore::GObjectXPathNSResolver::create):
2819         (WebCore::GObjectXPathNSResolver::GObjectXPathNSResolver): Create
2820         a new GObjectXPathNSResolver for the given WebKitDOMXPathNSResolver.
2821         * bindings/gobject/WebKitDOMXPathNSResolver.cpp: Added.
2822         (webkit_dom_xpath_ns_resolver_default_init):
2823         (webkit_dom_xpath_ns_resolver_lookup_namespace_uri):
2824         (webkitDOMNativeXPathNSResolverFinalize):
2825         (webkitDOMNativeXPathNSResolverConstructor):
2826         (webkit_dom_native_xpath_ns_resolver_init):
2827         (webkit_dom_native_xpath_ns_resolver_class_init):
2828         (webkitDOMNativeXPathNSResolverLookupNamespaceURI):
2829         (webkitDOMXPathNSResolverIfaceInit):
2830         (WebKit::core): Return the native XPathNSResolver when using the
2831         internal implementation or create a custom XPathNSResolver when
2832         using a user provided WebKitDOMXPathNSResolver.
2833         (WebKit::kit):
2834         * bindings/gobject/WebKitDOMXPathNSResolver.h: Added.
2835         * bindings/gobject/WebKitDOMXPathNSResolverPrivate.h: Added.
2836         * bindings/scripts/CodeGeneratorGObject.pm:
2837         (GenerateFunction): Add exceptions for XPathNSResolver since
2838         the core method returns a PassRefPtr.
2839
2840 2014-07-01  Frédéric Wang  <fred.wang@free.fr>
2841
2842         The OpenType MATH table is not read on Mac
2843         https://bugs.webkit.org/show_bug.cgi?id=133569
2844
2845         Reviewed by Chris Fleizach.
2846
2847         An incorrect tag format is used to load the OpenType MATH table on Mac.
2848         This commit makes the format consistent with what is used for other tables on Mac.
2849         This will allow to load the MATH table and to use it for the MathML rendering.
2850
2851         No new tests. platform/graphics/opentype already contains tests for the OpenType MATH support.
2852
2853         * platform/graphics/opentype/OpenTypeMathData.cpp:
2854
2855 2014-07-01  Alex Christensen  <achristensen@webkit.org>
2856
2857         [Win64] Unreviewed build fix.
2858
2859         * WebCore.vcxproj/WebCore.vcxproj:
2860         Exclude CA files from 32 and 64 bit WinCairo builds.
2861         * WebCore.vcxproj/WebCoreTestSupport.vcxproj:
2862         Don't use a precompiled header in 32 or 63 bit builds of MockCDM.cpp.
2863
2864 2014-07-01  Alex Christensen  <achristensen@webkit.org>
2865
2866         Reduce dynamic memory allocation in css jit.
2867         https://bugs.webkit.org/show_bug.cgi?id=134416
2868
2869         Reviewed by Benjamin Poulain.
2870
2871         * cssjit/FunctionCall.h:
2872         (WebCore::FunctionCall::saveAllocatedRegisters):
2873         * cssjit/RegisterAllocator.h:
2874         * cssjit/SelectorCompiler.cpp:
2875         (WebCore::SelectorCompiler::addPseudoClassType):
2876         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateSelectorChecker):
2877         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsNthChild):
2878         * cssjit/StackAllocator.h:
2879         (WebCore::StackAllocator::push):
2880         (WebCore::StackAllocator::pop):
2881         Use inlineCapacity template parameter to reduce malloc calls.
2882         I use 32 as the inline capacity for non-register related vectors because that
2883         is probably big enough for the selector compiler, and if it is not then a malloc
2884         call will not be significant because the selector is very complex.
2885
2886 2014-07-01  Eric Carlson  <eric.carlson@apple.com>
2887
2888         WebVTT percentage value can be a float
2889         https://bugs.webkit.org/show_bug.cgi?id=134511
2890
2891         Reviewed by Brent Fulgham.
2892
2893         No new tests, existing tests modified.
2894
2895         * html/track/TextTrackCueGeneric.cpp:
2896         (WebCore::TextTrackCueGeneric::setLine): int -> double.
2897         (WebCore::TextTrackCueGeneric::setPosition): Ditto.
2898         * html/track/TextTrackCueGeneric.h:
2899
2900         * html/track/VTTCue.cpp:
2901         (WebCore::VTTCue::setLine): Ditto.
2902         (WebCore::VTTCue::setPosition):  Ditto.
2903         (WebCore::VTTCue::setCueSettings): Support floating point values for line, position, and size.
2904         (WebCore::scanPercentage): Deleted.
2905         * html/track/VTTCue.h:
2906
2907         * html/track/VTTScanner.cpp:
2908         (WebCore::VTTScanner::scanFloat): Support negative values.
2909         * html/track/VTTScanner.h:
2910
2911         * html/track/WebVTTParser.cpp: Include parseFloatPercentageValue in non-WEBVTT_REGIONS builds.
2912         * html/track/WebVTTParser.h:
2913
2914         * platform/graphics/avfoundation/InbandTextTrackPrivateAVF.cpp:
2915         (WebCore::InbandTextTrackPrivateAVF::processCueAttributes): Drive by change to take a cue
2916             reference rather than a pointer.
2917         (WebCore::InbandTextTrackPrivateAVF::processAttributedStrings): Ditto.
2918         (WebCore::InbandTextTrackPrivateAVF::removeCompletedCues): Loop over all cues, don't LOG that
2919             we are asking the client to delete the cue because the client will have already deleted
2920             it when status when to GenericCueData::Complete.
2921         * platform/graphics/avfoundation/InbandTextTrackPrivateAVF.h:
2922
2923 2014-07-01  Simon Fraser  <simon.fraser@apple.com>
2924
2925         Improve behavior of position:sticky on zoomed pages
2926         https://bugs.webkit.org/show_bug.cgi?id=134535
2927         <rdar://problem/12898829>
2928
2929         Reviewed by Dean Jackson.
2930
2931         Despite r138036, position:sticky still had incorrect behavior in zoomed pages.
2932
2933         Fix by removing the scaling of the rect from viewportConstrainedVisibleContentRect(),
2934         since all the other math happens in unscaled coordinates.
2935
2936         * rendering/RenderBoxModelObject.cpp:
2937         (WebCore::RenderBoxModelObject::constrainingRectForStickyPosition):
2938
2939 2014-07-01  Alex Christensen  <achristensen@webkit.org>
2940
2941         [WebGL] Fix PVRTC extension.
2942         https://bugs.webkit.org/show_bug.cgi?id=134525
2943
2944         Reviewed by Dean Jackson.
2945
2946         * html/canvas/WebGLRenderingContext.cpp:
2947         (WebCore::WebGLRenderingContext::validateCompressedTexDimensions):
2948         Remove the return false which I mistakenly committed in r170128
2949         and make the gl error match the S3TC error and the khronos test.
2950
2951 2014-07-01  Chris Fleizach  <cfleizach@apple.com>
2952
2953         AX: iOS does not expose ARIA toggle buttons at all (not even as a regular button)
2954         https://bugs.webkit.org/show_bug.cgi?id=133875
2955
2956         Reviewed by Enrica Casucci.
2957
2958         Support the toggle button type in iOS.
2959
2960         Test: platform/ios-sim/accessibility/toggle-button.html
2961
2962         * accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
2963         (-[WebAccessibilityObjectWrapper accessibilityCanFuzzyHitTest]):
2964         (-[WebAccessibilityObjectWrapper accessibilityTraits]):
2965         (-[WebAccessibilityObjectWrapper determineIsAccessibilityElement]):
2966
2967 2014-07-01  Dean Jackson  <dino@apple.com>
2968
2969         [iOS] Subsampled JPEG images do not draw correctly via the canvas APIs
2970         https://bugs.webkit.org/show_bug.cgi?id=134513
2971         <rdar://problem/12078860>
2972         <rdar://problem/16745393>
2973
2974         Reviewed by Tim Horton.
2975
2976         Subsampled images (e.g. JPEG) were not consistently using
2977         their original dimensions and subsampled dimensions. This caused
2978         things like texImage2D to pack the pixels incorrectly, or drawImage
2979         to squish the rendering.
2980
2981         Renamed m_scale to m_subsamplingScale on FrameData.
2982
2983         Tests: fast/canvas/image-potential-subsample.html
2984                fast/canvas/webgl/tex-image-and-sub-image-2d-with-potentially-subsampled-image.html
2985
2986         * platform/graphics/BitmapImage.cpp:
2987         (WebCore::BitmapImage::cacheFrame): Rename to m_subsamplingScale.
2988         (WebCore::BitmapImage::frameAtIndex): Ditto.
2989         * platform/graphics/BitmapImage.h:
2990         (WebCore::FrameData::FrameData): Ditto.
2991         * platform/graphics/cg/BitmapImageCG.cpp:
2992         (WebCore::FrameData::clear): Ditto.
2993         (WebCore::BitmapImage::BitmapImage): Ditto.
2994         (WebCore::BitmapImage::draw): Use a scaledSrcRect that reflects the subsampled size,
2995         rather than assuming the srcRect accurately reflects how many pixels we have
2996         in the Bitmap.
2997         (WebCore::BitmapImage::copyUnscaledFrameAtIndex):
2998         * platform/graphics/cg/GraphicsContext3DCG.cpp:
2999         (WebCore::GraphicsContext3D::ImageExtractor::extractImage): Similar fix, although this
3000         time we just ask the image decoder to take into account the subsampled size
3001         when it is "generating" a frame, causing it to use the bitmap data it has already
3002         decoded.
3003
3004 2014-07-01  Joseph Pecoraro  <pecoraro@apple.com>
3005
3006         Web Inspector: Selected DOM element highlights invisible near bottom of the viewport (topContentInset?)
3007         https://bugs.webkit.org/show_bug.cgi?id=133818
3008
3009         Reviewed by Beth Dakin.
3010
3011         * inspector/InspectorOverlay.cpp:
3012         (WebCore::InspectorOverlay::update):
3013         Take the topContentInset into account when sizing the InspectorOverlay's FrameView.
3014
3015 2014-07-01  Chris Fleizach  <cfleizach@apple.com>
3016
3017         AX: HTML indeterminate IDL attribute not mapped to checkbox value=2
3018         https://bugs.webkit.org/show_bug.cgi?id=134492
3019
3020         Reviewed by Andreas Kling.
3021
3022         Support the indeterminate attribute in AX code.
3023
3024         Test: Update existing test: accessibility/aria-checked-mixed-value
3025
3026         * accessibility/AccessibilityObject.cpp:
3027         (WebCore::AccessibilityObject::checkboxOrRadioValue):
3028
3029 2014-07-01  Myles C. Maxfield  <mmaxfield@apple.com>
3030
3031         Typing an automatic text replacement phrase after a br in contenteditable is not rendered as expected
3032         https://bugs.webkit.org/show_bug.cgi?id=133883
3033
3034         Reviewed by Enrica Casucci.
3035
3036         enclosingDeletableElement() makes sure that the element's container is editable, but not
3037         that it is capable of having content inside it (like how a <br> can't).
3038
3039         No new tests. I don't think this is testable because it requires setting some state in System Preferences.
3040
3041         * editing/DeleteButtonController.cpp:
3042         (WebCore::enclosingDeletableElement):
3043
3044 2014-07-01  Alex Christensen  <achristensen@webkit.org>
3045
3046         [iOS] Unreviewed build fix after r170640.
3047
3048         * page/EditorClient.h:
3049         (WebCore::EditorClient::selectedTelephoneNumberRangesChanged):
3050         Add ifdefs to make everything compile and link.
3051
3052 2014-07-01  Daniel Bates  <dabates@apple.com>
3053
3054         Remove unnecessary calls to std::move()
3055         https://bugs.webkit.org/show_bug.cgi?id=134493
3056
3057         Reviewed by Anders Carlsson.
3058
3059         * Modules/encryptedmedia/CDM.cpp:
3060         (WebCore::CDM::CDM):
3061         * css/CSSGrammar.y.in:
3062         * css/CSSPrimitiveValue.cpp:
3063         (WebCore::CSSPrimitiveValue::formatNumberValue):
3064         * css/MediaQuery.cpp:
3065         (WebCore::MediaQuery::MediaQuery):
3066         * platform/graphics/mac/FontMac.mm:
3067         (WebCore::Font::dashesForIntersectionsWithRect):
3068
3069 2014-07-01  Zalan Bujtas  <zalan@apple.com>
3070
3071         Subpixel rendering: Pixel crack in breadcrumbs at devforums.apple.com.
3072         https://bugs.webkit.org/show_bug.cgi?id=134491
3073
3074         Reviewed by Simon Fraser.
3075
3076         Do not early round geometry values, while computing background image position. It changes the final
3077         subpixel values and that could lead to wrong snap positions.
3078         pixelSnapBackgroundImageGeometryForPainting() takes care of geometry snapping.
3079
3080         Test: fast/backgrounds/background-image-size-double-rounding.html
3081
3082         * rendering/RenderBoxModelObject.cpp:
3083         (WebCore::RenderBoxModelObject::calculateBackgroundImageGeometry):
3084
3085 2014-07-01  Pratik Solanki  <psolanki@apple.com>
3086
3087         Create NSURLRequest lazily when USE(CFNETWORK) is enabled
3088         https://bugs.webkit.org/show_bug.cgi?id=134441
3089
3090         Reviewed by Andreas Kling.
3091
3092         No new tests. Should be covered by existing tests.
3093
3094         * platform/network/cf/ResourceRequest.h:
3095         (WebCore::ResourceRequest::ResourceRequest):
3096         (WebCore::ResourceRequest::encodingRequiresPlatformData):
3097         * platform/network/cf/ResourceRequestCFNet.cpp:
3098         (WebCore::ResourceRequest::doUpdatePlatformRequest):
3099         (WebCore::ResourceRequest::doUpdatePlatformHTTPBody):
3100         (WebCore::ResourceRequest::setStorageSession):
3101         * platform/network/cocoa/ResourceRequestCocoa.mm:
3102         (WebCore::ResourceRequest::nsURLRequest):
3103
3104 2014-07-01  Brady Eidson  <beidson@apple.com>
3105
3106         Combine the Telephone and Selection overlay controllers, updating UI behavior.
3107         https://bugs.webkit.org/show_bug.cgi?id=134461
3108
3109         Reviewed by Tim Horton.
3110
3111         * WebCore.exp.in:
3112
3113         * dom/Range.cpp:
3114         (WebCore::Range::contains): Returns true if the Range completely contains the passed-in Range.
3115         * dom/Range.h:
3116
3117 2014-07-01  Antti Koivisto  <antti@apple.com>
3118
3119         REGRESSION(160908): vube.com video won't play after going into and out of fullscreen
3120         https://bugs.webkit.org/show_bug.cgi?id=134489
3121
3122         Reviewed by Zalan Bujtas.
3123
3124         Test: fullscreen/full-screen-plugin.html
3125
3126         It is difficult to restore the render tree correctly in all cases after removing a full screen
3127         renderer from the tree. r160908 avoided dealing with this by simply always reconstructing the subtree.
3128         Unfortunately plugin lifetime is currently tied to its renderer so this would cause the plugin to restart.
3129
3130         With this patch we avoid reconstruction in normal cases and only force it if the render tree is complicated.
3131
3132         * dom/Document.cpp:
3133         (WebCore::unwrapFullScreenRenderer):
3134
3135             Force reconstruction conditionally.
3136
3137         (WebCore::Document::webkitWillEnterFullScreenForElement):
3138         (WebCore::Document::webkitDidExitFullScreenForElement):
3139         * rendering/RenderFullScreen.cpp:
3140         (WebCore::RenderFullScreen::wrapRenderer):
3141         (WebCore::RenderFullScreen::unwrapRenderer):
3142
3143             Deal with the simple case of single child, possibly in anonymous wrapper.
3144             In other cases request reconstruction.
3145             This is covered by the existing fullscreen tests.
3146
3147         * rendering/RenderFullScreen.h:
3148
3149 2014-07-01  Zan Dobersek  <zdobersek@igalia.com>
3150
3151         Remove remaining Vector<> copies in WebCore accessibility code
3152         https://bugs.webkit.org/show_bug.cgi?id=133263
3153
3154         Reviewed by Darin Adler.
3155
3156         * accessibility/AXObjectCache.cpp:
3157         (WebCore::AXObjectCache::notificationPostTimerFired): Don't copy the Vector member and
3158         then clear it -- move it into the local variable instead.
3159         * accessibility/AccessibilityNodeObject.cpp:
3160         (WebCore::AccessibilityNodeObject::ariaLabeledByText): Move the axElements Vector into
3161         the AccessibilityText constructor.
3162         (WebCore::AccessibilityNodeObject::stringValue): Create a const reference to the Vector of
3163         HTMLSelectElement's list items instead of copying it.
3164         * accessibility/AccessibilityObject.h:
3165         (WebCore::AccessibilityText::AccessibilityText): Take the Vector parameter by non-const
3166         value and move it into the member variable.
3167         * accessibility/AccessibilityRenderObject.cpp:
3168         (WebCore::AccessibilityRenderObject::stringValue): Create a const reference to the Vector
3169         of HTMLSelectElement's list items instead of copying it.
3170         (WebCore::AccessibilityRenderObject::ariaSelectedRows): Wrap the AccessibilityChildrenVector
3171         iteration in a lambda. Use it to iterate over either the newly-constructed Vector (in case
3172         the object has the tree role) or the reference to the vector returned by
3173         AccessibilityTable::rows() (in case the object is an AccessibilityTable instance).
3174
3175 2014-06-30  Commit Queue  <commit-queue@webkit.org>
3176
3177         Unreviewed, rolling out r170605.
3178         https://bugs.webkit.org/show_bug.cgi?id=134484
3179
3180         The changes cause crashes on ARM64 (Requested by benjaminp on
3181         #webkit).
3182
3183         Reverted changeset:
3184
3185         "Reduce dynamic memory allocation in css jit."
3186         https://bugs.webkit.org/show_bug.cgi?id=134416
3187         http://trac.webkit.org/changeset/170605
3188
3189 2014-06-30  Anders Carlsson  <andersca@apple.com>
3190
3191         Change the AddBackForwardItem message to take a page state object
3192         https://bugs.webkit.org/show_bug.cgi?id=134475
3193
3194         Reviewed by Andreas Kling.
3195
3196         * WebCore.exp.in:
3197
3198 2014-06-30  Benjamin Poulain  <benjamin@webkit.org>
3199
3200         [iOS][WK2] Do not put tap highlight on images that are not in links, and on applet/embed/object
3201         https://bugs.webkit.org/show_bug.cgi?id=134471
3202         <rdar://problem/17513542>
3203
3204         Reviewed by Enrica Casucci.
3205
3206         UIKit has some special code to avoid tap highlight on certain type of object. This is not very flexible
3207         since the page has no way of overriding that behavior.
3208
3209         This patch implement the same feature by changing the default stylesheet. The tap highlight is disabled
3210         by default for those object, and enabled for images that are in a link.
3211
3212         * css/html.css:
3213         (applet, embed, object, img):
3214         (:-webkit-any-link img):
3215
3216 2014-06-30  Myles C. Maxfield  <mmaxfield@apple.com>
3217
3218         All Indic text is rendered as boxes on iOS
3219         https://bugs.webkit.org/show_bug.cgi?id=134464
3220
3221         Reviewed by Simon Fraser.
3222
3223         I made a typo in r170207.
3224
3225         Test: fast/text/indic.html
3226
3227         * platform/graphics/ios/FontCacheIOS.mm:
3228         (WebCore::languageSpecificFallbackFont):
3229
3230 2014-06-30  Myles C. Maxfield  <mmaxfield@apple.com>
3231
3232         [iOS] languageSpecificFallbackFont() is not consistent
3233         https://bugs.webkit.org/show_bug.cgi?id=134473
3234
3235         Reviewed by Darin Adler.
3236
3237         It can use "<" throughout the function as well as keeping the fallthrough behavior consistent.
3238
3239         No tests because there is no behavior change.
3240
3241         * platform/graphics/ios/FontCacheIOS.mm:
3242         (WebCore::languageSpecificFallbackFont):
3243
3244 2014-06-30  Brent Fulgham  <bfulgham@apple.com>
3245
3246         [Win] Unreviewed gardening.
3247
3248         * WebCore.vcxproj/WebCore.vcxproj: Do not compile JSFileException
3249         and JSFileReaderSync multiple times.
3250
3251 2014-06-30  Daniel Bates  <dabates@apple.com>
3252
3253         Avoid copying function object in lambda function in ScrollingThread::dispatchBarrier(); actually use move semantics
3254         https://bugs.webkit.org/show_bug.cgi?id=134470
3255
3256         Reviewed by Anders Carlsson.
3257
3258         Currently we always copy construct the captured std::function object when calling callOnMainThread()
3259         in the lambda function created in ScrollingThread::dispatchBarrier() because captured variables are
3260         const in a lambda expression's body by default. That is, the std::function object is captured as const.
3261         Instead, we should mark this lambda expression as mutable so the captured std::function object is
3262         non-const and hence we can use move semantics when passing it to callOnMainThread().
3263
3264         * page/scrolling/ScrollingThread.cpp:
3265         (WebCore::ScrollingThread::dispatchBarrier):
3266
3267 2014-06-30  Anders Carlsson  <andersca@apple.com>
3268
3269         Adopt the legacy session decoding inside WebPage::restoreSession for now
3270         https://bugs.webkit.org/show_bug.cgi?id=134465
3271
3272         Reviewed by Sam Weinig.
3273
3274         * WebCore.exp.in:
3275
3276 2014-06-30  Commit Queue  <commit-queue@webkit.org>
3277
3278         Unreviewed, rolling out r170578.
3279         https://bugs.webkit.org/show_bug.cgi?id=134462
3280
3281         causes assertions loading nytimes.com on iOS (Requested by
3282         smfr on #webkit).
3283
3284         Reverted changeset:
3285
3286         "Create NSURLRequest lazily when USE(CFNETWORK) is enabled"
3287         https://bugs.webkit.org/show_bug.cgi?id=134441
3288         http://trac.webkit.org/changeset/170578
3289
3290 2014-06-30  Alex Christensen  <achristensen@webkit.org>
3291
3292         Reduce dynamic memory allocation in css jit.
3293         https://bugs.webkit.org/show_bug.cgi?id=134416
3294
3295         Reviewed by Benjamin Poulain.
3296
3297         * cssjit/FunctionCall.h:
3298         (WebCore::FunctionCall::FunctionCall):
3299         (WebCore::FunctionCall::saveAllocatedCallerSavedRegisters):
3300         * cssjit/RegisterAllocator.h:
3301         * cssjit/SelectorCompiler.cpp:
3302         (WebCore::SelectorCompiler::addPseudoClassType):
3303         (WebCore::SelectorCompiler::SelectorCodeGenerator::generatePrologue):
3304         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateSelectorChecker):
3305         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsNthChild):
3306         Use inlineCapacity template parameter to reduce malloc calls.
3307         I use 16 and 32 as the inline capacity for non-register related vectors because that
3308         is probably big enough for the selector compiler, and if it is not then a malloc
3309         call will not be significant because the selector is very complex.
3310         * cssjit/StackAllocator.h:
3311         (WebCore::StackAllocator::push):
3312         Pass a vector of stack references to match calls to pop and to avoid a call to appendVector.
3313
3314 2014-06-30  Alex Christensen  <achristensen@webkit.org>
3315
3316         Use non-thumb registers in armv7 css jit.
3317         https://bugs.webkit.org/show_bug.cgi?id=134450
3318
3319         Reviewed by Geoff Garen.
3320
3321         * cssjit/RegisterAllocator.h:
3322         (WebCore::RegisterAllocator::isValidRegister):
3323         (WebCore::RegisterAllocator::isCallerSavedRegister):
3324         Use additional general purpose registers.
3325         * cssjit/SelectorCompiler.cpp:
3326         (WebCore::SelectorCompiler::SelectorCodeGenerator::compile):
3327         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateSelectorChecker):
3328         Always succeed when compiling css selectors.
3329
3330 2014-06-30  Myles C. Maxfield  <mmaxfield@apple.com>
3331
3332         Tiles on bottom of screen are not always allocated when necessary
3333         https://bugs.webkit.org/show_bug.cgi?id=134272
3334
3335         Reviewed by Simon Fraser.
3336
3337         The initial visibleRect is in the coordinate of the root layer, so its origin
3338         is at the top left of the view. The initial rect we were using doesn't
3339         include the contents inset, so it was too short, which was causing tiles near
3340         the bottom of the screen to not always be allocated if the tile threshold was
3341         close to the bottom of the view. Instead, we want to include the contents
3342         inset size so the visible rect includes the entire view.
3343         GraphicsLayerCA::recursiveCommitChanges() takes care of mapping and cropping
3344         the visible rect into the tiled layer's coordinate system, at which point it
3345         is used for visible tile logic.
3346
3347         Test: platform/mac-wk2/tiled-drawing/visible-rect-content-inset.html
3348
3349         * platform/graphics/ca/TileCoverageMap.cpp:
3350         (WebCore::TileCoverageMap::update):
3351         * rendering/RenderLayerCompositor.cpp:
3352         (WebCore::RenderLayerCompositor::flushPendingLayerChanges):
3353
3354 2014-06-30  Brent Fulgham  <bfulgham@apple.com>
3355
3356         [Win] Unreviewed project gardening.
3357
3358         * WebCore.vcxproj/WebCore.vcxproj: Clean up path information so
3359         files appear in appropriate places in project view.
3360         * WebCore.vcxproj/WebCore.vcxproj.filters: Ditto.
3361
3362 2014-06-30  Zalan Bujtas  <zalan@apple.com>
3363
3364         Remove redundant repaintCompositedLayers() parameter and its dependencies.
3365         https://bugs.webkit.org/show_bug.cgi?id=134431
3366
3367         Reviewed by Darin Adler.
3368
3369         No change in functionality.
3370
3371         * rendering/RenderLayer.cpp:
3372         (WebCore::RenderLayer::convertToPixelSnappedLayerCoords): Deleted.
3373         (WebCore::RenderLayer::convertToLayerCoords): Deleted.
3374         * rendering/RenderLayer.h:
3375         * rendering/RenderLayerCompositor.cpp:
3376         (WebCore::RenderLayerCompositor::repaintCompositedLayers):
3377         (WebCore::RenderLayerCompositor::recursiveRepaintLayer):
3378         * rendering/RenderLayerCompositor.h:
3379
3380 2014-06-30  Alex Christensen  <achristensen@webkit.org>
3381
3382         Remove webkit prefix from pointer lock.
3383         https://bugs.webkit.org/show_bug.cgi?id=134312
3384
3385         Reviewed by Dean Jackson.
3386
3387         * dom/Document.cpp:
3388         (WebCore::Document::exitPointerLock):
3389         (WebCore::Document::pointerLockElement):
3390         (WebCore::Document::webkitExitPointerLock): Deleted.
3391         (WebCore::Document::webkitPointerLockElement): Deleted.
3392         * dom/Document.h:
3393         * dom/Document.idl:
3394         * dom/Element.cpp:
3395         (WebCore::Element::requestPointerLock):
3396         (WebCore::Element::webkitRequestPointerLock): Deleted.
3397         * dom/Element.h:
3398         * dom/Element.idl:
3399         * dom/EventNames.h:
3400         * dom/MouseEvent.idl:
3401         * dom/MouseRelatedEvent.h:
3402         (WebCore::MouseRelatedEvent::movementX):
3403         (WebCore::MouseRelatedEvent::movementY):
3404         (WebCore::MouseRelatedEvent::webkitMovementX): Deleted.
3405         (WebCore::MouseRelatedEvent::webkitMovementY): Deleted.
3406         * page/PointerLockController.cpp:
3407         (WebCore::PointerLockController::PointerLockController):
3408         Initialize all data members.
3409         (WebCore::PointerLockController::requestPointerLock):
3410         (WebCore::PointerLockController::didAcquirePointerLock):
3411         (WebCore::PointerLockController::didNotAcquirePointerLock):
3412         (WebCore::PointerLockController::didLosePointerLock):
3413
3414 2014-06-30  Myles C. Maxfield  <mmaxfield@apple.com>
3415
3416         U16_IS_SURROGATE should not be called with a UChar32
3417         https://bugs.webkit.org/show_bug.cgi?id=134440
3418
3419         Reviewed by Darin Adler.
3420
3421         No new tests because there is no behavior change.
3422
3423         * platform/graphics/mac/ComplexTextController.cpp:
3424         (WebCore::advanceByCombiningCharacterSequence):
3425
3426 2014-06-30  Ryuan Choi  <ryuan.choi@samsung.com>
3427
3428         [EFL] Remove m_evasObject from Widget
3429         https://bugs.webkit.org/show_bug.cgi?id=134442
3430
3431         Reviewed by Gyuyoung Kim.
3432
3433         Since WebKit1/EFL was dropped, we don't need to keep the m_evasObject
3434         in the Widget.cpp because it's not used for WebKit2/Efl.
3435
3436         * platform/Widget.h:
3437         (WebCore::Widget::evasObject): Deleted.
3438         * platform/efl/ScrollbarEfl.cpp:
3439         Removed WebKit1/Efl specific code which uses m_evasObject.
3440         It will be a dummy until WebKit2/Efl support proper scrollbar implementation.
3441         (WebCore::ScrollbarEfl::ScrollbarEfl):