Source/WebCore/ChangeLog

   1 2015-03-17  Zalan Bujtas  <zalan@apple.com>
   2
   3         Simple line layout: Change FlowContents::segmentForPosition() to segmentForRun().
   4         https://bugs.webkit.org/show_bug.cgi?id=142785
   5
   6         Reviewed by Antti Koivisto.
   7
   8         This is in transition to support <br>. A particular position could point to multiple
   9         segments when <br> is directly followed by text.
  10
  11         No change in functionality.
  12
  13         * rendering/SimpleLineLayoutFlowContents.cpp:
  14         (WebCore::SimpleLineLayout::FlowContents::segmentIndexForRunSlow):
  15         (WebCore::SimpleLineLayout::FlowContents::segmentIndexForPositionSlow): Deleted.
  16         * rendering/SimpleLineLayoutFlowContents.h:
  17         (WebCore::SimpleLineLayout::FlowContents::segmentForRun):
  18         (WebCore::SimpleLineLayout::FlowContents::segmentIndexForPosition): Deleted.
  19         (WebCore::SimpleLineLayout::FlowContents::segmentForPosition): Deleted.
  20         * rendering/SimpleLineLayoutResolver.cpp:
  21         (WebCore::SimpleLineLayout::RunResolver::Run::text):
  22
  23 2015-03-17  Chris Dumez  <cdumez@apple.com>
  24
  25         [Mac][iOS] setSharedTimerFireInterval() / stopSharedTimer() are expensive
  26         https://bugs.webkit.org/show_bug.cgi?id=142752
  27         <rdar://problem/20176731>
  28
  29         Reviewed by Antti Koivisto.
  30
  31         setSharedTimerFireInterval() / stopSharedTimer() are expensive on Mac
  32         and iOS on pages using a lot of timers.
  33
  34         For example, on bing.com / iOS, ~15.4% of the CPU time is spent in
  35         setSharedTimerFireInterval() and ~14.7% of the CPU time is spent in
  36         stopSharedTimer(). The expensive calls are CFRunLoopAddTimer (11.4%),
  37         CFRunLoopTimerInvalidate (14.1%), CFRunLoopTimerCreate (3.3%).
  38
  39         The issue is that we keep creating, adding to run loop modes, and then
  40         destroying the sharedTimer for each firing event. This is very
  41         expensive. In such case, the CFRunLoopTimerRef documentation advises to
  42         """
  43         ... create a repeating timer with an initial firing time in the distant
  44         future (or the initial firing time) and a very large repeat interval—on
  45         the order of decades or more—and add it to all the necessary run loop
  46         modes. Then, when you know when the timer should fire next, you reset
  47         the firing time with CFRunLoopTimerSetNextFireDate, perhaps from the
  48         timer’s own callback function. This technique effectively produces a
  49         reusable, asynchronous timer.
  50         """ [1].
  51
  52         Doing so greatly decreases CPU time spend in:
  53         - setSharedTimerFireInterval(): 15.4% -> 4.6%
  54         - stopSharedTimer(): 14.6% -> 8.6%
  55
  56         Overall CPU time spent on bing.com in timerFired() goes down from
  57         ~61.2% to ~49.5%.
  58
  59         This patch also refactors the SharedTimer code to share as much as
  60         possible between Mac and iOS.
  61
  62         This patch is based in part on the following patch:
  63         http://trac.webkit.org/changeset/143210
  64
  65         [1] https://developer.apple.com/library/prerelease/ios/documentation/CoreFoundation/Reference/CFRunLoopTimerRef/index.html#//apple_ref/c/func/CFRunLoopTimerSetNextFireDate
  66
  67         * WebCore.xcodeproj/project.pbxproj:
  68         * platform/SharedTimer.h:
  69         (WebCore::SharedTimer::invalidate):
  70         (WebCore::MainThreadSharedTimer::setFiredFunction): Deleted.
  71         (WebCore::MainThreadSharedTimer::setFireInterval): Deleted.
  72         (WebCore::MainThreadSharedTimer::stop): Deleted.
  73         * platform/ThreadTimers.cpp:
  74         (WebCore::ThreadTimers::fireTimersInNestedEventLoop):
  75         * platform/cf/SharedTimerCF.mm: Added.
  76         (WebCore::applicationDidBecomeActive):
  77         (WebCore::setupPowerObserver):
  78         (WebCore::setSharedTimerFiredFunction):
  79         (WebCore::timerFired):
  80         (WebCore::restartSharedTimer):
  81         (WebCore::invalidateSharedTimer):
  82         (WebCore::setSharedTimerFireInterval):
  83         (WebCore::stopSharedTimer):
  84         * platform/efl/SharedTimerEfl.cpp:
  85         (WebCore::invalidateSharedTimer):
  86         * platform/gtk/SharedTimerGtk.cpp:
  87         (WebCore::invalidateSharedTimer):
  88         * platform/ios/SharedTimerIOS.mm: Removed.
  89         * platform/mac/PowerObserverMac.h: Copied from Source/WebCore/platform/efl/SharedTimerEfl.cpp.
  90         * platform/mac/PowerObserverMac.mm: Renamed from Source/WebCore/platform/mac/SharedTimerMac.mm.
  91         (WebCore::PowerObserver::PowerObserver):
  92         (WebCore::PowerObserver::~PowerObserver):
  93         (WebCore::PowerObserver::didReceiveSystemPowerNotification):
  94         * platform/win/SharedTimerWin.cpp:
  95         (WebCore::removeSharedTimer):
  96
  97 2015-03-17  Tim Horton  <timothy_horton@apple.com>
  98
  99         Cannot invoke action menus anymore
 100         https://bugs.webkit.org/show_bug.cgi?id=142797
 101         <rdar://problem/20032670>
 102
 103         Reviewed by Beth Dakin.
 104
 105         * platform/spi/mac/NSMenuSPI.h:
 106         Add additional NSMenu SPI.
 107
 108 2015-03-17  Zalan Bujtas  <zalan@apple.com>
 109
 110         Simple line layout: Split fragments on renderer boundary on the fly.
 111         https://bugs.webkit.org/show_bug.cgi?id=142579
 112
 113         Reviewed by Antti Koivisto.
 114
 115         Fragment splitting at renderers' boundary at the end of the line is no longer needed.
 116         This patch ensures that TextFragmentIterator::nextTextFragment() does not concatenate
 117         fragments across renderer boundary.
 118
 119         Test: fast/text/simple-line-with-multiple-renderers.html
 120
 121         * rendering/SimpleLineLayout.cpp:
 122         (WebCore::SimpleLineLayout::revertRuns):
 123         (WebCore::SimpleLineLayout::LineState::isEmpty):
 124         (WebCore::SimpleLineLayout::LineState::appendFragmentAndCreateRunIfNeeded): Check if fragments need a new run and
 125         whether neighboring fragments need collapsing.
 126         (WebCore::SimpleLineLayout::LineState::revertUncommitted): Revert fragments over multiple renderers if they form
 127         a fragment continuation.
 128         (WebCore::SimpleLineLayout::LineState::removeTrailingWhitespace):
 129         (WebCore::SimpleLineLayout::forceFragmentToLine): Ensure that if a fragment has continuation across multiple renderers,
 130         they all get added to the current line.
 131         (WebCore::SimpleLineLayout::createLineRuns):
 132         (WebCore::SimpleLineLayout::LineState::appendFragment): Deleted.
 133         (WebCore::SimpleLineLayout::FragmentForwardIterator::FragmentForwardIterator): Deleted.
 134         (WebCore::SimpleLineLayout::FragmentForwardIterator::operator++): Deleted.
 135         (WebCore::SimpleLineLayout::FragmentForwardIterator::operator!=): Deleted.
 136         (WebCore::SimpleLineLayout::FragmentForwardIterator::operator*): Deleted.
 137         (WebCore::SimpleLineLayout::begin): Deleted.
 138         (WebCore::SimpleLineLayout::end): Deleted.
 139         (WebCore::SimpleLineLayout::preWrap): Deleted.
 140         (WebCore::SimpleLineLayout::removeTrailingWhitespace): Deleted.
 141         (WebCore::SimpleLineLayout::updateLineConstrains): Deleted.
 142         (WebCore::SimpleLineLayout::splitRunsAtRendererBoundary): Deleted.
 143         (WebCore::SimpleLineLayout::createTextRuns): Deleted.
 144         (WebCore::SimpleLineLayout::create): Deleted.
 145         (WebCore::SimpleLineLayout::Layout::create): Deleted.
 146         * rendering/SimpleLineLayoutFlowContents.h:
 147         (WebCore::SimpleLineLayout::FlowContents::Iterator::operator+):
 148         (WebCore::SimpleLineLayout::FlowContents::Iterator::operator*):
 149         (WebCore::SimpleLineLayout::FlowContents::Iterator::operator->):
 150         * rendering/SimpleLineLayoutFunctions.cpp:
 151         (WebCore::SimpleLineLayout::paintFlow): Unrelated fix on non-collapsed tab position.
 152         * rendering/SimpleLineLayoutTextFragmentIterator.cpp:
 153         (WebCore::SimpleLineLayout::TextFragmentIterator::TextFragmentIterator):
 154         (WebCore::SimpleLineLayout::TextFragmentIterator::nextTextFragment):
 155         (WebCore::SimpleLineLayout::TextFragmentIterator::revertToFragment):
 156         (WebCore::SimpleLineLayout::TextFragmentIterator::textWidth):
 157         (WebCore::SimpleLineLayout::TextFragmentIterator::skipToNextPosition):
 158         * rendering/SimpleLineLayoutTextFragmentIterator.h:
 159         (WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::TextFragment):
 160         (WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::overlapsToNextRenderer):
 161         (WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::isCollapsible):
 162         (WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::operator==):
 163         (WebCore::SimpleLineLayout::TextFragmentIterator::isLineBreak):
 164         (WebCore::SimpleLineLayout::TextFragmentIterator::TextFragment::split): Deleted.
 165
 166 2015-03-17  Jeremy Jones  <jeremyj@apple.com>
 167
 168         When tab hides, pause fullscreen and exit normally.
 169         https://bugs.webkit.org/show_bug.cgi?id=142685
 170
 171         Reviewed by Eric Carlson.
 172
 173         Pause fullscreen playback when switching tabs. Exit fullscreen is not necessary and even prevents
 174         the normal flow of teardown. This allows the normal exit fullscreen call to succeed and call its callback.
 175
 176         * platform/ios/WebVideoFullscreenInterfaceAVKit.mm:
 177         (WebVideoFullscreenInterfaceAVKit::requestHideAndExitFullscreen):
 178
 179 2015-03-17  Per Arne Vollan  <peavo@outlook.com>
 180
 181         [WinCairo] Video position is incorrect when located inside a frame.
 182         https://bugs.webkit.org/show_bug.cgi?id=142784
 183
 184         Reviewed by Brent Fulgham.
 185
 186         We need to take the enclosing frame's position into account, when finding the video position.
 187
 188         * platform/graphics/win/MediaPlayerPrivateMediaFoundation.cpp:
 189         (WebCore::MediaPlayerPrivateMediaFoundation::setSize):
 190
 191 2015-03-17  Benjamin Poulain  <bpoulain@apple.com>
 192
 193         Compile character ranges targeting the same state as range check in the bytecode
 194         https://bugs.webkit.org/show_bug.cgi?id=142759
 195
 196         Reviewed by Alex Christensen.
 197
 198         Previously, character ranges would be compiled as many individual character checks.
 199         For example, a transition on "[a-z]" would do 26 character checks + jump, which leads
 200         to enormous matchines.
 201
 202         With this patch, we find the ranges at lowering time and generate a single instruction
 203         for them: "CheckValueRange". This helps making the machine denser when the input
 204         use character sets.
 205
 206         The second part of this patch goes further in the case where the transitions out of
 207         a state cover the entire alphabet. In that case, we create a fallback transition
 208         on the fly and remove all the ranges made useless.
 209         That case is common when ranges are used with inverse character set (e.g. [^a]+a).
 210
 211         * contentextensions/DFABytecode.h:
 212         (WebCore::ContentExtensions::instructionSizeWithArguments):
 213         * contentextensions/DFABytecodeCompiler.cpp:
 214         (WebCore::ContentExtensions::DFABytecodeCompiler::emitCheckValueRange):
 215         (WebCore::ContentExtensions::DFABytecodeCompiler::compileNode):
 216         (WebCore::ContentExtensions::DFABytecodeCompiler::compileNodeTransitions):
 217         (WebCore::ContentExtensions::DFABytecodeCompiler::compileCheckForRange):
 218         * contentextensions/DFABytecodeCompiler.h:
 219         Extend the compiler to detect ranges and lower them as CheckValueRange.
 220
 221         * contentextensions/DFABytecodeInterpreter.cpp:
 222         (WebCore::ContentExtensions::DFABytecodeInterpreter::interpret):
 223         Range checks in the interpreter.
 224
 225         * contentextensions/NFA.cpp:
 226         (WebCore::ContentExtensions::NFA::setFinal):
 227         This assertion does not make sense with the current codebase. Actions are "compressed",
 228         it is possible to have two patterns with the same action.
 229
 230         * contentextensions/NFAToDFA.cpp:
 231         (WebCore::ContentExtensions::simplifyTransitions):
 232         A very simple DFA optimization function: it only reduce the strength of ranges.
 233
 234         (WebCore::ContentExtensions::NFAToDFA::convert):
 235
 236 2015-03-17  Jer Noble  <jer.noble@apple.com>
 237
 238         REGRESSION (r181423): Crash @ generatedcontent.org at com.apple.WebCore: WebCore::MediaPlayer::maximumDurationToCacheMediaTime const + 4
 239         https://bugs.webkit.org/show_bug.cgi?id=142787
 240
 241         Reviewed by Eric Carlson.
 242
 243         Null check m_player before derefencing.
 244
 245         * html/HTMLMediaElement.cpp:
 246         (WebCore::HTMLMediaElement::parseAttribute):
 247
 248 2015-03-17  Beth Dakin  <bdakin@apple.com>
 249
 250         DOM mouse events have weird timing for force clickable elements in Safari 8.0.3 on
 251         10.10.2
 252         https://bugs.webkit.org/show_bug.cgi?id=142700
 253         -and corresponding-
 254         rdar://problem/20165168
 255
 256         Reviewed by Tim Horton.
 257
 258         This patch adds a new enum and member variable so that EventHandler can keep track
 259         of the current immediate action state.
 260         * page/EventHandler.cpp:
 261         (WebCore::EventHandler::EventHandler):
 262
 263         A new mouse press even is starting. We can re-set m_immediateActionStage to none
 264         unless a Hit Test has already been performed.
 265         (WebCore::EventHandler::handleMousePressEvent):
 266
 267         If an immediate action was completed, then send mouse to the DOM and return early.
 268         This will prevent us from doing our own normal mouseup behaviors such as
 269         navigating to a link that was clicked — we only want to do that if the click was
 270         not used to perform an immediate action.
 271         (WebCore::EventHandler::handleMouseReleaseEvent):
 272         * page/EventHandler.h:
 273         (WebCore::EventHandler::setImmediateActionStage):
 274
 275 2015-03-17  Joseph Pecoraro  <pecoraro@apple.com>
 276
 277         Use a better parameter name for Document.getElementsByClassName
 278         https://bugs.webkit.org/show_bug.cgi?id=142771
 279
 280         Reviewed by Chris Dumez.
 281
 282         * bindings/objc/PublicDOMInterfaces.h:
 283         * dom/Document.idl:
 284
 285 2015-03-17  Timothy Horton  <timothy_horton@apple.com>
 286
 287         Reproducible null deref under ScriptedAnimationController::createDisplayRefreshMonitor
 288         https://bugs.webkit.org/show_bug.cgi?id=142776
 289         <rdar://problem/18921338>
 290
 291         Reviewed by Alexey Proskuryakov.
 292
 293         Test: fast/animation/request-animation-frame-unparented-iframe-crash.html
 294
 295         In some cases (like the new test), we can end up trying to start
 296         requestAnimationFrame on a Document that has no Page. Most paths null-checked
 297         the Page and did the right thing, but one failed to do so. In addition,
 298         the current fallback (when Page is null) can result in us constructing
 299         the wrong kind of DisplayRefreshMonitor, which could lead to trouble
 300         down the road when it's reused. Instead, just completely avoid making a
 301         DisplayRefreshMonitor in the null-page case.
 302
 303         * dom/ScriptedAnimationController.cpp:
 304         (WebCore::ScriptedAnimationController::createDisplayRefreshMonitor):
 305         If the page is null, bail.
 306
 307         * dom/ScriptedAnimationController.h:
 308         * platform/graphics/DisplayRefreshMonitor.cpp:
 309         (WebCore::DisplayRefreshMonitor::create):
 310         Use Optional<> to make it easy to distinguish between ChromeClient
 311         being unreachable (because we don't have a Page for some reason) and
 312         ChromeClient declaring that it doesn't want to override the type of
 313         DisplayRefreshMonitor that is created.
 314
 315         If ChromeClient was unreachable for some reason, we'll get back an engaged
 316         nullptr and return it (instead of creating a DisplayRefreshMonitor based
 317         on the platform). This avoids creating the wrong type of DisplayRefreshMonitor
 318         in the rare case where we can't reach the ChromeClient (e.g. a freshly unparented
 319         IFrame).
 320
 321         If instead the client returns a disengaged Nullopt, we'll interpret that as
 322         "construct the default type", which falls back on the platform #ifdefs to
 323         decide what to make.
 324
 325         * platform/graphics/DisplayRefreshMonitorManager.cpp:
 326         (WebCore::DisplayRefreshMonitorManager::ensureMonitorForClient):
 327         (WebCore::DisplayRefreshMonitorManager::scheduleAnimation):
 328         Silently handle the case where we failed to make a DisplayRefreshMonitor.
 329
 330         * platform/graphics/DisplayRefreshMonitor.h:
 331         * platform/graphics/DisplayRefreshMonitorClient.h:
 332         * platform/graphics/GraphicsLayerUpdater.cpp:
 333         (WebCore::GraphicsLayerUpdater::createDisplayRefreshMonitor):
 334         * platform/graphics/GraphicsLayerUpdater.h:
 335         * rendering/RenderLayerCompositor.cpp:
 336         (WebCore::RenderLayerCompositor::createDisplayRefreshMonitor):
 337         * rendering/RenderLayerCompositor.h:
 338         Adjust to the new signature of createDisplayRefreshMonitor, and return
 339         an engaged (nullptr) Optional if we can't get to ChromeClient for any reason.
 340
 341         * page/ChromeClient.h:
 342         Return Nullopt (indicating a lack of override) by default.
 343
 344 2015-03-17  Dean Jackson  <dino@apple.com>
 345
 346         Implement Scroll Container Animation Triggers
 347         https://bugs.webkit.org/show_bug.cgi?id=142732
 348
 349         Attempt to fix the build.
 350
 351         * css/CSSComputedStyleDeclaration.cpp: Add an #if ENABLE guard.
 352
 353 2015-03-17  Dean Jackson  <dino@apple.com>
 354
 355         Implement Scroll Container Animation Triggers
 356         https://bugs.webkit.org/show_bug.cgi?id=142732
 357
 358         Reviewed by Simon Fraser.
 359
 360         Test: animations/trigger-container-scroll-simple.html
 361
 362         Basic implementation of container-scroll. It only checks
 363         the page scroll position for trigger values (not the scrolling
 364         container in an overflow).
 365
 366         * css/CSSComputedStyleDeclaration.cpp: Add CSSPropertyWebkitAnimationTrigger
 367         so that this property will appear in the inspector.
 368
 369         * page/FrameView.cpp:
 370         (WebCore::FrameView::sendScrollEvent): If the page has scrolled, let the animation
 371         controller know about it.
 372
 373         * page/animation/AnimationBase.cpp:
 374         (WebCore::AnimationBase::updateStateMachine): Whitespace fix.
 375         (WebCore::AnimationBase::fireAnimationEventsIfNeeded): If there is a trigger,
 376         and the scroll position is past it, then tell the state machine that
 377         we should start.
 378         (WebCore::AnimationBase::timeToNextService): Use the scroll position as
 379         an input to the update timer if a trigger is involved.
 380
 381         * page/animation/AnimationController.cpp:
 382         (WebCore::AnimationControllerPrivate::ensureCompositeAnimation): Add whitespace.
 383         (WebCore::AnimationControllerPrivate::scrollWasUpdated): Call updateAnimations.
 384         (WebCore::AnimationController::scrollWasUpdated): Call into AnimationControllerPrivate.
 385         * page/animation/AnimationController.h:
 386         * page/animation/AnimationControllerPrivate.h:
 387
 388         * page/animation/CompositeAnimation.cpp: Keep a record of whether we have a scroll
 389         triggered animation.
 390         (WebCore::CompositeAnimation::CompositeAnimation):
 391         (WebCore::CompositeAnimation::updateKeyframeAnimations):
 392         * page/animation/CompositeAnimation.h:
 393         (WebCore::CompositeAnimation::hasScrollTriggeredAnimation):
 394         * platform/animation/Animation.cpp:
 395         (WebCore::Animation::operator=):
 396
 397 2015-03-17  Simon Fraser  <simon.fraser@apple.com>
 398
 399         Move some code from LogicalSelectionOffsetCaches into RenderElement
 400         https://bugs.webkit.org/show_bug.cgi?id=142758
 401
 402         Reviewed by Myles C. Maxfield.
 403
 404         LogicalSelectionOffsetCaches had some useful code regarding containing blocks etc
 405         that should be used in more places, so move it into RenderElement.
 406
 407         No behavior change.
 408
 409         * rendering/LogicalSelectionOffsetCaches.h:
 410         (WebCore::LogicalSelectionOffsetCaches::LogicalSelectionOffsetCaches):
 411         (WebCore::isContainingBlockCandidateForAbsolutelyPositionedObject): Deleted.
 412         (WebCore::isNonRenderBlockInline): Deleted.
 413         (WebCore::containingBlockForFixedPosition): Deleted.
 414         (WebCore::containingBlockForAbsolutePosition): Deleted.
 415         (WebCore::containingBlockForObjectInFlow): Deleted.
 416         * rendering/RenderBlock.cpp: No need to initialize static data.
 417         (WebCore::RenderBlock::positionedObjects): nullptr
 418         * rendering/RenderElement.cpp:
 419         (WebCore::RenderElement::containingBlockForFixedPosition):
 420         (WebCore::RenderElement::containingBlockForAbsolutePosition):
 421         (WebCore::isNonRenderBlockInline):
 422         (WebCore::RenderElement::containingBlockForObjectInFlow):
 423         * rendering/RenderElement.h:
 424         (WebCore::RenderElement::canContainAbsolutelyPositionedObjects):
 425         * rendering/RenderLayer.cpp:
 426         (WebCore::isContainerForPositioned):
 427         * rendering/RenderObject.cpp:
 428         (WebCore::RenderObject::containingBlock):
 429
 430 2015-03-17  Alex Christensen  <achristensen@webkit.org>
 431
 432         [WinCairo] Unreviewed build fix after r181640.
 433
 434         * platform/graphics/cairo/GraphicsContext3DCairo.cpp:
 435         Tell Windows how to find ShaderLang.h.
 436
 437 2015-03-17  Chris Dumez  <cdumez@apple.com>
 438
 439         'pageLoaded' diagnostic logging is too verbose
 440         https://bugs.webkit.org/show_bug.cgi?id=142727
 441         <rdar://problem/18937048>
 442
 443         Reviewed by Eric Carlson.
 444
 445         Make 'pageLoaded' diagnostic logging less verbose:
 446         - Log once per main frame instead of once for subframe
 447         - Apply sampling
 448
 449         * loader/FrameLoader.cpp:
 450         (WebCore::FrameLoader::checkLoadCompleteForThisFrame):
 451
 452 2015-03-17  Carlos Garcia Campos  <cgarcia@igalia.com>
 453
 454         [GTK] Wrong transfer annotations used in GObject DOM bindings
 455         https://bugs.webkit.org/show_bug.cgi?id=142780
 456
 457         Reviewed by Gustavo Noronha Silva.
 458
 459         We are using transfer none for all methods returning a GObject DOM
 460         Object. That's not true. Only objects derived from Node are
 461         automatically released by the DOM object cache and can be transfer
 462         none. All other objects are added to the cache only to avoid
 463         creating the same wrapper twice for the same core object, but
 464         caller should release the returned reference.
 465
 466         * bindings/gobject/WebKitDOMCustomUnstable.h:
 467         * bindings/scripts/CodeGeneratorGObject.pm:
 468         (GetTransferTypeForReturnType):
 469         (GenerateFunction):
 470
 471 2015-03-17  Carlos Garcia Campos  <cgarcia@igalia.com>
 472
 473         [GTK] WebKitDOM objects leaking
 474         https://bugs.webkit.org/show_bug.cgi?id=118788
 475
 476         Reviewed by Darin Adler and Sergio Villar Senin.
 477
 478         Use a DOMwindowObserver class, derived from DOMWindowProperty to
 479         be notified when the window object is detached from the frame to
 480         clear the DOM objects associated to that frame in that case too.
 481
 482         * bindings/gobject/DOMObjectCache.cpp:
 483
 484 2015-03-17  Zan Dobersek  <zdobersek@igalia.com>
 485
 486         [CMake] Use a forwarding header for ANGLE's ShaderLang.h to avoid picking up ANGLE's EGL headers
 487         https://bugs.webkit.org/show_bug.cgi?id=142530
 488
 489         Reviewed by Darin Adler.
 490
 491         Include the ANGLE's ShaderLang.h through the new forwarding header. This allows
 492         us to not list Source/ThirdParty/ANGLE/include in the list of inclusion directories
 493         and thus avoid ANGLE's EGL and GLES2/GLES3 headers, defaulting to the system-provided
 494         headers instead.
 495
 496         Source/ThirdParty/ANGLE/include/KHR is still used because ANGLE's khrplatform.h is
 497         required by the ShaderLang.h header. Source/ThirdParty/ANGLE/src is not used for the
 498         whole WebCore library anymore, only the ANGLESupport library.
 499
 500         * CMakeLists.txt:
 501         * platform/graphics/ANGLEWebKitBridge.h:
 502         * platform/graphics/cairo/GraphicsContext3DCairo.cpp:
 503
 504 2015-03-17  Matt Baker  <mattbaker@apple.com>
 505
 506         Web Inspector: Show rendering frames (and FPS) in Layout and Rendering timeline
 507         https://bugs.webkit.org/show_bug.cgi?id=142029
 508
 509         Reviewed by Timothy Hatcher.
 510
 511         Add new functionality to the Inspector timelines backend to add runloop data to timeline recordings.
 512
 513         * inspector/InspectorTimelineAgent.cpp:
 514         (WebCore::currentRunLoop):
 515         (WebCore::InspectorTimelineAgent::internalStart):
 516         (WebCore::InspectorTimelineAgent::internalStop):
 517         (WebCore::toProtocol):
 518         (WebCore::InspectorTimelineAgent::InspectorTimelineAgent):
 519         Install observers for the begining and end of the runloop when recording begins. All other
 520         instrumented timeline events get added as children of the current runloop record, which is
 521         sent to the frontend once the runloop completes.
 522
 523         * inspector/InspectorTimelineAgent.h:
 524
 525         * platform/cf/RunLoopObserver.cpp:
 526         (WebCore::RunLoopObserver::schedule):
 527         Wrapper changed to allow observing arbitrary runloop activities.
 528
 529         * platform/cf/RunLoopObserver.h:
 530
 531 2015-03-17  Philippe Normand  <pnormand@igalia.com>
 532
 533         [GTK] basic OpenWebRTC build support
 534         https://bugs.webkit.org/show_bug.cgi?id=142393
 535
 536         Reviewed by Carlos Garcia Campos.
 537
 538         * PlatformGTK.cmake: Add OpenWebRTC CFLAGS and LDFLAGS.
 539
 540 2015-03-17  Joseph Pecoraro  <pecoraro@apple.com>
 541
 542         Remove never used "useLayerOnTop" bindings generator argument
 543         https://bugs.webkit.org/show_bug.cgi?id=142773
 544
 545         Reviewed by Darin Adler.
 546
 547         * bindings/scripts/CodeGenerator.pm:
 548         * bindings/scripts/CodeGeneratorJS.pm:
 549         * bindings/scripts/CodeGeneratorObjC.pm:
 550         * bindings/scripts/generate-bindings.pl:
 551
 552 2015-03-16  Ryosuke Niwa  <rniwa@webkit.org>
 553
 554         Enable ES6 classes by default
 555         https://bugs.webkit.org/show_bug.cgi?id=142774
 556
 557         Reviewed by Gavin Barraclough.
 558
 559         * Configurations/FeatureDefines.xcconfig:
 560
 561 2015-03-16  Simon Fraser  <simon.fraser@apple.com>
 562
 563         Update the debug overlays after layout
 564         https://bugs.webkit.org/show_bug.cgi?id=142768
 565
 566         Reviewed by Zalan Bujtas.
 567
 568         The debug overlays need to be updated after layout, in case elements with wheel event handlers
 569         moved around.
 570
 571         DebugPageOverlays::didLayout() is cheap if there are no overlays.
 572
 573         Call DebugPageOverlays::didLayout() for all frames, not just the main frame, since subframes can contribute
 574         to the main frame's event handler region.
 575
 576         * page/DebugPageOverlays.cpp:
 577         (WebCore::DebugPageOverlays::regionChanged):
 578         * page/DebugPageOverlays.h:
 579         (WebCore::DebugPageOverlays::didLayout):
 580         (WebCore::DebugPageOverlays::didChangeEventHandlers):
 581         * page/FrameView.cpp:
 582         (WebCore::FrameView::layout):
 583
 584 2015-03-16  Jon Lee  <jonlee@apple.com>
 585
 586         Unreviewed fix for crash after r181608.
 587
 588         * WebCore.xcodeproj/project.pbxproj: Include it in the WebCore framework.
 589
 590 2015-03-16  Simon Fraser  <simon.fraser@apple.com>
 591
 592         Generalize the Document code that maintains a set of nodes with event handlers
 593         https://bugs.webkit.org/show_bug.cgi?id=142762
 594
 595         Reviewed by Zalan Bujtas, Darin Adler.
 596
 597         Document.h defines a TouchEventTargetSet type, which will in future be used for
 598         other event types too (wheel events), so rename it to EventTargetSet.
 599
 600         Take didRemoveEventTargetNode() outside the TOUCH_EVENTS #ifdef, and change it
 601         to take a reference.
 602
 603         * dom/Document.cpp:
 604         (WebCore::Document::prepareForDestruction): References
 605         (WebCore::Document::didAddTouchEventHandler): Ditto.
 606         (WebCore::Document::didRemoveEventTargetNode): Ditto.
 607         * dom/Document.h:
 608         (WebCore::Document::touchEventTargets): No-one calls this; just remove the non
 609         #ideffed one.
 610         * html/HTMLInputElement.cpp:
 611         (WebCore::HTMLInputElement::~HTMLInputElement): Pass a ref.
 612         (WebCore::HTMLInputElement::didMoveToNewDocument): Pass a ref.
 613         * page/DOMWindow.cpp:
 614         (WebCore::DOMWindow::removeAllEventListeners): Pass a ref.
 615
 616 2015-03-16  Joseph Pecoraro  <pecoraro@apple.com>
 617
 618         Web Inspector: Object Previews in Indexed DB tables
 619         https://bugs.webkit.org/show_bug.cgi?id=140813
 620
 621         Reviewed by Timothy Hatcher.
 622
 623         * inspector/InspectorIndexedDBAgent.cpp:
 624         Include previews with object store objects.
 625
 626 2015-03-16  Jer Noble  <jer.noble@apple.com>
 627
 628         [Mac] Update missing image UI
 629         https://bugs.webkit.org/show_bug.cgi?id=142592
 630
 631         Reviewed by Darin Adler.
 632
 633         Update the broken image glyphs with new art (including a 3x image).
 634
 635         Test: fast/hidpi/broken-image-icon-very-hidpi.html
 636
 637         * Resources/missingImage.png:
 638         * Resources/missingImage@2x.png:
 639         * Resources/missingImage@3x.png: Added.
 640         * WebCore.xcodeproj/project.pbxproj:
 641         * loader/cache/CachedImage.cpp:
 642         (WebCore::CachedImage::brokenImage):
 643
 644 2015-03-16  Dean Jackson  <dino@apple.com>
 645
 646         Parsing and Style Resolution of Container-based Animation Triggers
 647         https://bugs.webkit.org/show_bug.cgi?id=142687
 648         <rdar://problem/20170007>
 649
 650         Reviewed by Simon Fraser.
 651
 652         Take 2 after the previous patch was rolled out.
 653
 654         This is the beginning of a prototype implementation of
 655         CSS Animation Triggers, as described by
 656         https://lists.w3.org/Archives/Public/www-style/2014Sep/0135.html
 657
 658         In this patch we parse and resolve the value of a new
 659         CSS property "-webkit-animation-trigger". At the moment it
 660         only accepts one function value "container-scroll", which
 661         will trigger the animation at an absolute position within
 662         an element's scrolling container. We expect the syntax to
 663         change in the near future, as the spec is written.
 664
 665         Tests: animations/trigger-computed-style.html
 666                animations/trigger-parsing.html
 667
 668         * WebCore.xcodeproj/project.pbxproj: Add the new files.
 669
 670         * css/CSSAnimationTriggerScrollValue.cpp: Added.
 671         (WebCore::CSSAnimationTriggerScrollValue::customCSSText): Output text for computed style.
 672         (WebCore::CSSAnimationTriggerScrollValue::equals): Compare two values.
 673         * css/CSSAnimationTriggerScrollValue.h: Added. This holds the CSS side of the
 674         scroll trigger. This name may change in the future to better represent the
 675         type of trigger, but it is good enough for now.
 676         (WebCore::CSSAnimationTriggerScrollValue::create):
 677         (WebCore::CSSAnimationTriggerScrollValue::startValue):
 678         (WebCore::CSSAnimationTriggerScrollValue::endValue):
 679         (WebCore::CSSAnimationTriggerScrollValue::CSSAnimationTriggerScrollValue):
 680
 681         * css/CSSComputedStyleDeclaration.cpp:
 682         (WebCore::createAnimationTriggerValue): Maps an AnimationTrigger into a CSSValue.
 683         (WebCore::getAnimationTriggerValue): Gets the current computed style.
 684         (WebCore::ComputedStyleExtractor::propertyValue):
 685
 686         * css/CSSParser.cpp:
 687         (WebCore::CSSParser::parseValue):
 688         (WebCore::CSSParser::parseAnimationTrigger): Parse the "container-scroll" function
 689         and record the value as a CSSAnimationTriggerScrollValue.
 690         (WebCore::CSSParser::parseAnimationProperty): Handle the new property.
 691         * css/CSSParser.h:
 692
 693         * css/CSSPropertyNames.in: Add "-webkit-animation-trigger".
 694
 695         * css/CSSToStyleMap.cpp:
 696         (WebCore::CSSToStyleMap::mapAnimationTrigger): Map a CSSValue into a trigger value on
 697         an Animation object.
 698         * css/CSSToStyleMap.h:
 699
 700         * css/CSSValue.cpp: Handle the new CSSValue type.
 701         (WebCore::CSSValue::equals):
 702         (WebCore::CSSValue::cssText):
 703         (WebCore::CSSValue::destroy):
 704         * css/CSSValue.h:
 705         (WebCore::CSSValue::isAnimationTriggerScrollValue):
 706
 707         * platform/animation/Animation.cpp: Make sure to initialise m_trigger and m_triggerSet,
 708         and use them in the operator==.
 709         * platform/animation/Animation.h: Add AnimationTrigger as a new field.
 710         (WebCore::Animation::isTriggerSet):
 711         (WebCore::Animation::isEmpty):
 712
 713         * platform/animation/AnimationTrigger.h: Added. New base class and subclasses for
 714         "auto" and the scrolling trigger.
 715         (WebCore::AnimationTrigger::~AnimationTrigger):
 716         (WebCore::AnimationTrigger::type):
 717         (WebCore::AnimationTrigger::isAutoAnimationTrigger):
 718         (WebCore::AnimationTrigger::isScrollAnimationTrigger):
 719         (WebCore::AnimationTrigger::AnimationTrigger):
 720         (WebCore::AutoAnimationTrigger::create):
 721         (WebCore::AutoAnimationTrigger::~AutoAnimationTrigger):
 722         (WebCore::AutoAnimationTrigger::AutoAnimationTrigger):
 723         (WebCore::ScrollAnimationTrigger::create):
 724         (WebCore::ScrollAnimationTrigger::~ScrollAnimationTrigger):
 725         (WebCore::ScrollAnimationTrigger::startValue):
 726         (WebCore::ScrollAnimationTrigger::setStartValue):
 727         (WebCore::ScrollAnimationTrigger::endValue):
 728         (WebCore::ScrollAnimationTrigger::setEndValue):
 729         (WebCore::ScrollAnimationTrigger::hasEndValue):
 730         (WebCore::ScrollAnimationTrigger::setHasEndValue):
 731         (WebCore::ScrollAnimationTrigger::ScrollAnimationTrigger):
 732
 733 2015-03-16  Alex Christensen  <achristensen@webkit.org>
 734
 735         Progress towards CMake on Mac
 736         https://bugs.webkit.org/show_bug.cgi?id=142747
 737
 738         Reviewed by Chris Dumez.
 739
 740         * CMakeLists.txt:
 741         * PlatformMac.cmake:
 742         Added more directories, interfaces, and forwarding headers.
 743         Temporarily disabled the generating of ObjC bindings in CMake builds.
 744         * platform/mac/PasteboardMac.mm:
 745         Removed unused include.
 746
 747 2015-03-16  Joanmarie Diggs  <jdiggs@igalia.com>
 748
 749         AX: Crash viewing http://www.last.fm/
 750         https://bugs.webkit.org/show_bug.cgi?id=142309
 751
 752         Reviewed by Chris Fleizach.
 753
 754         The crash occurs when a not-yet-rendered object emits a children-changed
 755         signal. If an assistive technology is listening, AT-SPI2 will attempt to
 756         create and cache the state set for the child being added and the creation
 757         of the state set assumes a rendered object.
 758
 759         Test: platform/gtk/accessibility/no-notification-for-unrendered-iframe-children.html
 760
 761         * accessibility/atk/AXObjectCacheAtk.cpp:
 762         (WebCore::AXObjectCache::attachWrapper):
 763
 764 2015-03-16  Commit Queue  <commit-queue@webkit.org>
 765
 766         Unreviewed, rolling out r181492.
 767         https://bugs.webkit.org/show_bug.cgi?id=142756
 768
 769         May have regressed PLT (Requested by anttik on #webkit).
 770
 771         Reverted changeset:
 772
 773         "Cache glyph widths to GlyphPages"
 774         https://bugs.webkit.org/show_bug.cgi?id=142028
 775         http://trac.webkit.org/changeset/181492
 776
 777 2015-03-16  Roger Fong  <roger_fong@apple.com>
 778
 779         [WebGL2] Instancing draw calls.
 780         https://bugs.webkit.org/show_bug.cgi?id=126939.
 781         <rdar://problem/15002379>
 782
 783         Reviewed by Dean Jackson.
 784
 785         Tested by a modified version of the 1.0.3 conformance tests:
 786         conformance/extensions/angle-instanced-arrays.html
 787         conformance/extensions/angle-instanced-arrays-out-of-bounds.html
 788
 789         These tests will be landed along with other modified extension conformance tests
 790         once approval from Khronos is received.
 791
 792         * html/canvas/WebGL2RenderingContext.cpp:
 793         (WebCore::WebGL2RenderingContext::clear): Generate error if clearing an integer color buffer.
 794         (WebCore::WebGL2RenderingContext::vertexAttribDivisor): Call method from base class.
 795         (WebCore::WebGL2RenderingContext::drawArraysInstanced): Ditto.
 796         (WebCore::WebGL2RenderingContext::drawElementsInstanced): Ditto.
 797         (WebCore::WebGL2RenderingContext::isIntegerFormat): Ditto.
 798         (WebCore::WebGL2RenderingContext::validateDrawElements): Deleted. Move back to base class.
 799         * html/canvas/WebGL2RenderingContext.h:
 800         * html/canvas/WebGL2RenderingContext.idl: Add a missing enum.
 801         * html/canvas/WebGLRenderingContext.cpp:
 802         (WebCore::WebGLRenderingContext::clear): Copied from WebGLRenderingContextBase.
 803         (WebCore::WebGLRenderingContext::validateDrawElements): Deleted. Move back to base class.
 804         * html/canvas/WebGLRenderingContext.h:
 805         * html/canvas/WebGLRenderingContextBase.cpp:
 806         (WebCore::WebGLRenderingContextBase::clear): Deleted. Moved to WebGLRenderingContext.
 807         (WebCore::WebGLRenderingContextBase::getVertexAttrib): Check for WebGL2 context.
 808         (WebCore::WebGLRenderingContext::validateDrawElements): Ditto.
 809         * html/canvas/WebGLRenderingContextBase.h:
 810
 811 2015-03-16  Commit Queue  <commit-queue@webkit.org>
 812
 813         Unreviewed, rolling out r181572.
 814         https://bugs.webkit.org/show_bug.cgi?id=142755
 815
 816         Caused weird test failures in transitions and animations
 817         (Requested by dino on #webkit).
 818
 819         Reverted changeset:
 820
 821         "Parsing and Style Resolution of Container-based Animation
 822         Triggers"
 823         https://bugs.webkit.org/show_bug.cgi?id=142687
 824         http://trac.webkit.org/changeset/181572
 825
 826 2015-03-16  Yoav Weiss  <yoav@yoav.ws>
 827
 828         Remove setCachedImage from HTMLImageElement since it is not used
 829         https://bugs.webkit.org/show_bug.cgi?id=142740
 830
 831         Reviewed by Chris Dumez.
 832
 833         No new tests since this patch is just removing dead code.
 834
 835         HTMLImageElement::setCachedImage is not being called by anyone.
 836         This patch removes it, since it's dead code.
 837
 838         * html/HTMLImageElement.h:
 839         (WebCore::HTMLImageElement::setCachedImage): Deleted.
 840
 841 2015-03-16  Brent Fulgham  <bfulgham@apple.com>
 842
 843         WebKit1 Clients Are Not Reliably Repainted
 844         https://bugs.webkit.org/show_bug.cgi?id=142750
 845         <rdar://problem/20042453>
 846
 847         Reviewed by Simon Fraser.
 848
 849         * page/FrameView.cpp:
 850         (WebCore::FrameView::paintContents): Move "Red Rect" debug painting before
 851         the early return so we can see when this happening in debug builds.
 852         * page/FrameView.h:
 853         (WebCore::FrameView::inPaintableState): Added.
 854
 855 2015-03-16  Chris Dumez  <cdumez@apple.com>
 856
 857         Make DatabaseContext suspendable if there is no pending database activity
 858         https://bugs.webkit.org/show_bug.cgi?id=142716
 859         <rdar://problem/19923085>
 860
 861         Reviewed by Andreas Kling.
 862
 863         Make DatabaseContext suspendable if there is no pending database
 864         activity, i.e:
 865         - No pending Database creation JS callback
 866         - No pending transaction(s)
 867
 868         Suspending is safe in this case because we are not going to interrupt
 869         any database activity, nor fire any JS event.
 870
 871         This greatly increases the likelihood of pages using websql to enter
 872         the PageCache.
 873
 874         Tests: fast/history/page-cache-webdatabase-no-transaction-db.html
 875                fast/history/page-cache-webdatabase-pending-transaction.html
 876
 877         * Modules/webdatabase/Database.cpp:
 878         (WebCore::Database::hasPendingTransaction):
 879         * Modules/webdatabase/Database.h:
 880         * Modules/webdatabase/DatabaseContext.cpp:
 881         (WebCore::DatabaseContext::canSuspend):
 882         * Modules/webdatabase/DatabaseManager.cpp:
 883         (WebCore::DatabaseManager::openDatabase):
 884         * Modules/webdatabase/DatabaseThread.cpp:
 885         (WebCore::DatabaseThread::hasPendingDatabaseActivity):
 886         * Modules/webdatabase/DatabaseThread.h:
 887
 888 2015-03-16  Brady Eidson  <beidson@apple.com>
 889
 890         Addressing additional review feedback after http://trac.webkit.org/changeset/181565
 891         https://bugs.webkit.org/show_bug.cgi?id=142733
 892
 893         Reviewed by Darin Adler.
 894
 895         * loader/icon/IconController.cpp:
 896         (WebCore::IconController::startLoader): Null check page()
 897
 898 2015-03-16  Roger Fong  <roger_fong@apple.com>
 899
 900         [Win] Unreviewed build fix attempt after r181571.
 901
 902         * WebCore.vcxproj/WebCore.vcxproj:
 903
 904 2015-03-16  Roger Fong  <roger_fong@apple.com>
 905
 906         [WebGL2] Multiple Render Targets.
 907         https://bugs.webkit.org/show_bug.cgi?id=126994.
 908         <rdar://problem/15815766>
 909
 910         Reviewed by Dean Jackson.
 911
 912         Tested by a modified version of the 1.0.3 conformance test:
 913         conformance/extensions/webgl-draw-buffers.html
 914         This test will be landed along with other modified extension conformance tests
 915         once approval from Khronos is received.
 916
 917         * html/canvas/WebGL2RenderingContext.cpp:
 918         (WebCore::WebGL2RenderingContext::drawBuffers): Mostly the same as WebGLDrawBuffers::drawBuffersWEBGL.
 919         Returns different error messages, uses non EXT enums.
 920         (WebCore::WebGL2RenderingContext::clearBufferiv): This actually does nothing for now but the validation has been implemented.
 921         (WebCore::WebGL2RenderingContext::clearBufferuiv): Ditto.
 922         (WebCore::WebGL2RenderingContext::clearBufferfv): Ditto.
 923         (WebCore::WebGL2RenderingContext::clearBufferfi): Ditto.
 924         (WebCore::WebGL2RenderingContext::validateFramebufferFuncParameters): Removes the extension object check.
 925         (WebCore::WebGL2RenderingContext::getMaxDrawBuffers):
 926         (WebCore::WebGL2RenderingContext::getMaxColorAttachments): Must return the same thing as getMaxDrawBuffers.
 927         (WebCore::WebGL2RenderingContext::getParameter): Remove extension object checks.
 928         * html/canvas/WebGL2RenderingContext.h:
 929         * html/canvas/WebGLFramebuffer.cpp:
 930         (WebCore::WebGLFramebuffer::drawBuffersIfNecessary):
 931         * html/canvas/WebGLRenderingContext.cpp:
 932         (WebCore::WebGLRenderingContext::validateFramebufferFuncParameters): Copied from WebGLRenderingContextBase.
 933         (WebCore::WebGLRenderingContext::getMaxDrawBuffers): Ditto.
 934         (WebCore::WebGLRenderingContext::getMaxColorAttachments): Ditto.
 935         * html/canvas/WebGLRenderingContext.h:
 936         * html/canvas/WebGLRenderingContextBase.cpp:
 937         (WebCore::WebGLRenderingContextBase::validateFramebufferFuncParameters): Deleted.
 938         * html/canvas/WebGLRenderingContextBase.h:
 939
 940 2015-03-16  Dean Jackson  <dino@apple.com>
 941
 942         Parsing and Style Resolution of Container-based Animation Triggers
 943         https://bugs.webkit.org/show_bug.cgi?id=142687
 944         <rdar://problem/20170007>
 945
 946         Reviewed by Simon Fraser.
 947
 948         This is the beginning of a prototype implementation of
 949         CSS Animation Triggers, as described by
 950         https://lists.w3.org/Archives/Public/www-style/2014Sep/0135.html
 951
 952         In this patch we parse and resolve the value of a new
 953         CSS property "-webkit-animation-trigger". At the moment it
 954         only accepts one function value "container-scroll", which
 955         will trigger the animation at an absolute position within
 956         an element's scrolling container. We expect the syntax to
 957         change in the near future, as the spec is written.
 958
 959         Tests: animations/trigger-computed-style.html
 960                animations/trigger-parsing.html
 961
 962         * WebCore.xcodeproj/project.pbxproj: Add the new files.
 963
 964         * css/CSSAnimationTriggerScrollValue.cpp: Added.
 965         (WebCore::CSSAnimationTriggerScrollValue::customCSSText): Output text for computed style.
 966         (WebCore::CSSAnimationTriggerScrollValue::equals): Compare two values.
 967         * css/CSSAnimationTriggerScrollValue.h: Added. This holds the CSS side of the
 968         scroll trigger. This name may change in the future to better represent the
 969         type of trigger, but it is good enough for now.
 970         (WebCore::CSSAnimationTriggerScrollValue::create):
 971         (WebCore::CSSAnimationTriggerScrollValue::startValue):
 972         (WebCore::CSSAnimationTriggerScrollValue::endValue):
 973         (WebCore::CSSAnimationTriggerScrollValue::CSSAnimationTriggerScrollValue):
 974
 975         * css/CSSComputedStyleDeclaration.cpp:
 976         (WebCore::createAnimationTriggerValue): Maps an AnimationTrigger into a CSSValue.
 977         (WebCore::getAnimationTriggerValue): Gets the current computed style.
 978         (WebCore::ComputedStyleExtractor::propertyValue):
 979
 980         * css/CSSParser.cpp:
 981         (WebCore::CSSParser::parseValue):
 982         (WebCore::CSSParser::parseAnimationTrigger): Parse the "container-scroll" function
 983         and record the value as a CSSAnimationTriggerScrollValue.
 984         (WebCore::CSSParser::parseAnimationProperty): Handle the new property.
 985         * css/CSSParser.h:
 986
 987         * css/CSSPropertyNames.in: Add "-webkit-animation-trigger".
 988
 989         * css/CSSToStyleMap.cpp:
 990         (WebCore::CSSToStyleMap::mapAnimationTrigger): Map a CSSValue into a trigger value on
 991         an Animation object.
 992         * css/CSSToStyleMap.h:
 993
 994         * css/CSSValue.cpp: Handle the new CSSValue type.
 995         (WebCore::CSSValue::equals):
 996         (WebCore::CSSValue::cssText):
 997         (WebCore::CSSValue::destroy):
 998         * css/CSSValue.h:
 999         (WebCore::CSSValue::isAnimationTriggerScrollValue):
1000
1001         * platform/animation/Animation.h: Add AnimationTrigger as a new field.
1002         (WebCore::Animation::isTriggerSet):
1003         (WebCore::Animation::isEmpty):
1004
1005         * platform/animation/AnimationTrigger.h: Added. New base class and subclasses for
1006         "auto" and the scrolling trigger.
1007         (WebCore::AnimationTrigger::~AnimationTrigger):
1008         (WebCore::AnimationTrigger::type):
1009         (WebCore::AnimationTrigger::isAutoAnimationTrigger):
1010         (WebCore::AnimationTrigger::isScrollAnimationTrigger):
1011         (WebCore::AnimationTrigger::AnimationTrigger):
1012         (WebCore::AutoAnimationTrigger::create):
1013         (WebCore::AutoAnimationTrigger::~AutoAnimationTrigger):
1014         (WebCore::AutoAnimationTrigger::AutoAnimationTrigger):
1015         (WebCore::ScrollAnimationTrigger::create):
1016         (WebCore::ScrollAnimationTrigger::~ScrollAnimationTrigger):
1017         (WebCore::ScrollAnimationTrigger::startValue):
1018         (WebCore::ScrollAnimationTrigger::setStartValue):
1019         (WebCore::ScrollAnimationTrigger::endValue):
1020         (WebCore::ScrollAnimationTrigger::setEndValue):
1021         (WebCore::ScrollAnimationTrigger::hasEndValue):
1022         (WebCore::ScrollAnimationTrigger::setHasEndValue):
1023         (WebCore::ScrollAnimationTrigger::ScrollAnimationTrigger):
1024
1025 2015-03-15  Roger Fong  <roger_fong@apple.com>
1026
1027         [WebGL2] Vertex Array Objects.
1028         https://bugs.webkit.org/show_bug.cgi?id=126944.
1029         <rdar://problem/15002455>
1030
1031         Reviewed by Dean Jackson.
1032
1033         Tested by a modified version of the 1.0.3 conformance test:
1034         conformance/extensions/oes-vertex-array-object.html
1035         This test will be landed along with other modified extension conformance tests
1036         once approval from Khronos is received.
1037
1038         * html/canvas/OESVertexArrayObject.cpp: Associate extension calls only with WebGLRenderingContext.
1039         (WebCore::OESVertexArrayObject::deleteVertexArrayOES):
1040         (WebCore::OESVertexArrayObject::bindVertexArrayOES):
1041         * html/canvas/WebGLRenderingContext.cpp:
1042         (WebCore::WebGLRenderingContext::initializeVertexArrayObjects):
1043         (WebCore::WebGLRenderingContext::getParameter):
1044         * html/canvas/WebGL2RenderingContext.cpp: Implement vertex array object calls as part WebGL2 context.
1045         (WebCore::WebGL2RenderingContext::createVertexArray):
1046         (WebCore::WebGL2RenderingContext::deleteVertexArray):
1047         (WebCore::WebGL2RenderingContext::isVertexArray):
1048         (WebCore::WebGL2RenderingContext::bindVertexArray):
1049         (WebCore::WebGL2RenderingContext::initializeVertexArrayObjects):
1050         (WebCore::WebGL2RenderingContext::getParameter):
1051
1052         Duplicate extension object vertex array object extension methods in GraphicsContext3D.
1053         Implementation may change after we upgrade to newer GL headers and profile.
1054         * platform/graphics/GraphicsContext3D.h:
1055         * platform/graphics/opengl/GraphicsContext3DOpenGLCommon.cpp:
1056         (WebCore::GraphicsContext3D::createVertexArray):
1057         (WebCore::GraphicsContext3D::deleteVertexArray):
1058         (WebCore::GraphicsContext3D::isVertexArray):
1059         (WebCore::GraphicsContext3D::bindVertexArray):
1060
1061         Have WebGLVertexArrayObjectOES and WebGLVertexArrayObject inherit from WebGLRenderingContextBase.
1062         * CMakeLists.txt:
1063         * WebCore.xcodeproj/project.pbxproj:
1064         * html/canvas/WebGLVertexArrayObject.cpp:
1065         (WebCore::WebGLVertexArrayObject::create):
1066         (WebCore::WebGLVertexArrayObject::WebGLVertexArrayObject): Use GraphicsContext3D implementation of createVertexArray.
1067         (WebCore::WebGLVertexArrayObject::deleteObjectImpl):
1068         * html/canvas/WebGLVertexArrayObject.h:
1069         * html/canvas/WebGLVertexArrayObjectBase.cpp: Added.
1070         (WebCore::WebGLVertexArrayObjectBase::WebGLVertexArrayObjectBase):
1071         (WebCore::WebGLVertexArrayObjectBase::setElementArrayBuffer):
1072         (WebCore::WebGLVertexArrayObjectBase::setVertexAttribState):
1073         (WebCore::WebGLVertexArrayObjectBase::unbindBuffer):
1074         (WebCore::WebGLVertexArrayObjectBase::setVertexAttribDivisor):
1075         * html/canvas/WebGLVertexArrayObjectBase.h: Added.
1076         (WebCore::WebGLVertexArrayObjectBase::~WebGLVertexArrayObjectBase):
1077         (WebCore::WebGLVertexArrayObjectBase::VertexAttribState::VertexAttribState):
1078         (WebCore::WebGLVertexArrayObjectBase::VertexAttribState::isBound):
1079         (WebCore::WebGLVertexArrayObjectBase::VertexAttribState::validateBinding):
1080         (WebCore::WebGLVertexArrayObjectBase::isDefaultObject):
1081         (WebCore::WebGLVertexArrayObjectBase::hasEverBeenBound):
1082         (WebCore::WebGLVertexArrayObjectBase::setHasEverBeenBound):
1083         (WebCore::WebGLVertexArrayObjectBase::getElementArrayBuffer):
1084         (WebCore::WebGLVertexArrayObjectBase::getVertexAttribState):
1085         * html/canvas/WebGLVertexArrayObjectOES.cpp:
1086         (WebCore::WebGLVertexArrayObjectOES::WebGLVertexArrayObjectOES):
1087         (WebCore::WebGLVertexArrayObjectOES::deleteObjectImpl):
1088         (WebCore::WebGLVertexArrayObjectOES::setElementArrayBuffer): Deleted.
1089         (WebCore::WebGLVertexArrayObjectOES::setVertexAttribState): Deleted.
1090         (WebCore::WebGLVertexArrayObjectOES::unbindBuffer): Deleted.
1091         (WebCore::WebGLVertexArrayObjectOES::setVertexAttribDivisor): Deleted.
1092         * html/canvas/WebGLVertexArrayObjectOES.h:
1093
1094         * html/canvas/WebGLRenderingContextBase.cpp: Use WebGLVertexArrayObjectBase class.
1095         (WebCore::WebGLRenderingContextBase::initializeNewContext):
1096         (WebCore::WebGLRenderingContextBase::disableVertexAttribArray):
1097         (WebCore::WebGLRenderingContextBase::validateVertexAttributes):
1098         (WebCore::WebGLRenderingContextBase::enableVertexAttribArray):
1099         (WebCore::WebGLRenderingContextBase::getVertexAttrib):
1100         (WebCore::WebGLRenderingContextBase::initVertexAttrib0):
1101         (WebCore::WebGLRenderingContextBase::simulateVertexAttrib0):
1102         (WebCore::WebGLRenderingContextBase::restoreStatesAfterVertexAttrib0Simulation):
1103         * html/canvas/WebGLRenderingContextBase.h:
1104         (WebCore::WebGLRenderingContextBase::setBoundVertexArrayObject):
1105
1106         Handle construction of WebGLGetInfo using WebGLVertexArrayObject.
1107         * bindings/js/JSWebGLRenderingContextBaseCustom.cpp:
1108         (WebCore::toJS):
1109         * html/canvas/WebGLGetInfo.cpp:
1110         (WebCore::WebGLGetInfo::WebGLGetInfo):
1111         (WebCore::WebGLGetInfo::getWebGLVertexArrayObjectOES):
1112         (WebCore::WebGLGetInfo::getWebGLVertexArrayObject):
1113         * html/canvas/WebGLGetInfo.h:
1114
1115 2015-03-16  Brent Fulgham  <bfulgham@apple.com>
1116
1117         Potentially uninitialized Inspector values
1118         https://bugs.webkit.org/show_bug.cgi?id=142730
1119
1120         Reviewed by Joseph Pecoraro.
1121
1122         * inspector/InspectorDOMAgent.cpp:
1123         (WebCore::parseColor): Make sure color values are given an initial value.
1124
1125 2015-03-16  Brady Eidson  <beidson@apple.com>
1126
1127         URLs visited during private browsing show up in WebpageIcons.db
1128         rdar://problem/11254910 and https://bugs.webkit.org/show_bug.cgi?id=142733
1129
1130         Patch by Sam Weinig. Reviewed by Brady Eidson.
1131
1132         * loader/icon/IconController.cpp:
1133         (WebCore::IconController::startLoader): Bail early here if the page is using an ephemeral session.
1134         (WebCore::IconController::continueLoadWithDecision): Instead of here.
1135
1136 2015-03-16  Conrad Shultz  <conrad_shultz@apple.com>
1137
1138         Allow clients to selectively disable plug-ins
1139         https://bugs.webkit.org/show_bug.cgi?id=142506
1140
1141         Reviewed by Anders Carlsson.
1142
1143         Add new functionality allow clients to declaratively disable individual plug-ins (in a manner that conceals them
1144         from the page). As part of this:
1145
1146         1) Introduce the concept of web-visible plug-ins and related concepts, as distinct from the real underlying
1147            set of plug-ins.
1148
1149         2) Where applicable, plumb additional information about plug-ins (specifically, bundle identification) deeper
1150            down the stack.
1151
1152         3) Add generic functionality to PluginStrategy to support filtering plug-ins for visibility and introduce
1153            a concrete implementation thereof in WebPlatformStrategies in WebKit2.
1154
1155         4) Add messaging infrastructure to allow clients to set and clear plug-in policies.
1156
1157         While currently only used in a very limited manner, the new declarative plug-in policies are written generically
1158         so that they could be easily used in the future to reduce synchronous messaging to the client when loading plug-ins.
1159
1160         * dom/DOMImplementation.cpp:
1161         (WebCore::DOMImplementation::createDocument):
1162         Update to reflect function rename.
1163
1164         * loader/SubframeLoader.cpp:
1165         (WebCore::findPluginMIMETypeFromURL):
1166         Adopt getWebVisibleMimesAndPluginIndices().
1167         (WebCore::logPluginRequest):
1168         Update to reflect function rename.
1169         (WebCore::SubframeLoader::shouldUsePlugin):
1170         Ditto.
1171
1172         * platform/PlatformStrategies.h:
1173         Export platformStrategies(), since it is now used in WebProcess.cpp.
1174
1175         * plugins/DOMMimeType.cpp:
1176         (WebCore::DOMMimeType::type):
1177         (WebCore::DOMMimeType::description):
1178         (WebCore::DOMMimeType::mimeClassInfo):
1179         Adopt getWebVisibleMimesAndPluginIndices().
1180         (WebCore::DOMMimeType::enabledPlugin):
1181         Ditto.
1182
1183         * plugins/DOMMimeType.h:
1184         Don't return references in a few places where it is no longer safe to do so.
1185         (WebCore::DOMMimeType::mimeClassInfo): Deleted.
1186
1187         * plugins/DOMMimeTypeArray.cpp:
1188         (WebCore::DOMMimeTypeArray::length):
1189         Adopt getWebVisibleMimesAndPluginIndices().
1190         (WebCore::DOMMimeTypeArray::item):
1191         Ditto.
1192         (WebCore::DOMMimeTypeArray::canGetItemsForName):
1193         Ditto.
1194         (WebCore::DOMMimeTypeArray::namedItem):
1195         Ditto.
1196
1197         * plugins/DOMPlugin.cpp:
1198         (WebCore::DOMPlugin::pluginInfo):
1199         Adopt getWebVisiblePlugins().
1200         (WebCore::DOMPlugin::item):
1201         Adopt getWebVisibleMimesAndPluginIndices().
1202         (WebCore::DOMPlugin::canGetItemsForName):
1203         Ditto.
1204         (WebCore::DOMPlugin::namedItem):
1205         Ditto.
1206
1207         * plugins/DOMPlugin.h:
1208         (WebCore::DOMPlugin::pluginInfo): Deleted.
1209
1210         * plugins/DOMPluginArray.cpp:
1211         (WebCore::DOMPluginArray::length):
1212         Adopt getWebVisiblePlugins().
1213         (WebCore::DOMPluginArray::item):
1214         Ditto.
1215         (WebCore::DOMPluginArray::canGetItemsForName):
1216         Ditto.
1217         (WebCore::DOMPluginArray::namedItem):
1218         Ditto.
1219
1220         * plugins/PluginData.cpp:
1221         (WebCore::PluginData::PluginData):
1222         Stash the passed-in Page and call initPlugins().
1223         (WebCore::PluginData::getWebVisiblePlugins):
1224         New member function; call through to PluginStrategy::getWebVisiblePluginInfo().
1225         (WebCore::PluginData::getWebVisibleMimesAndPluginIndices):
1226         New member function; build up the mimes and mimePluginIndices vectors in the same manner as before, but
1227         limited to the web-visible plug-ins.
1228         (WebCore::PluginData::supportsWebVisibleMimeType):
1229         Renamed from supportsMimeType(); update to work in terms of web-visible plug-ins.
1230         (WebCore::PluginData::pluginInfoForWebVisibleMimeType):
1231         Renamed from pluginInfoForMimeType(); ditto.
1232         (WebCore::PluginData::pluginNameForWebVisibleMimeType):
1233         Renamed from pluginNameForMimeType(); ditto.
1234         (WebCore::PluginData::pluginFileForWebVisibleMimeType):
1235         Renamed from pluginFileForMimeType(); ditto.
1236         (WebCore::PluginData::initPlugins):
1237         (WebCore::PluginData::supportsMimeType): Deleted.
1238         (WebCore::PluginData::pluginInfoForMimeType): Deleted.
1239         (WebCore::PluginData::pluginNameForMimeType): Deleted.
1240         (WebCore::PluginData::pluginFileForMimeType): Deleted.
1241
1242         * plugins/PluginData.h:
1243         Add a member variable for the associate Page; declare the PluginLoadClientPolicy enumeration; add
1244         new members to PluginInfo for the clientLoadPolicy and bundle information.
1245         (WebCore::PluginData::PluginData):
1246         Replace some member functions with new ones that will hide plug-ins upon request from the client;
1247         (WebCore::PluginData::mimes): Deleted.
1248         (WebCore::PluginData::mimePluginIndices): Deleted.
1249
1250         * plugins/PluginStrategy.h:
1251         Declare new member functions for retrieving web-visible plug-ins and setting/clearing plug-in policies.
1252
1253         * replay/SerializationMethods.cpp:
1254         (JSC::EncodingTraits<PluginData>::encodeValue):
1255         Remove now-obsolete code for handling MIME types and add a FIXME.
1256         (JSC::DeserializedPluginData::DeserializedPluginData):
1257         (JSC::EncodingTraits<PluginData>::decodeValue):
1258         Ditto.
1259         (JSC::EncodingTraits<PluginInfo>::encodeValue):
1260         Handle the new members in PluginInfo.
1261         (JSC::EncodingTraits<PluginInfo>::decodeValue):
1262         Ditto.
1263
1264         * replay/WebInputs.json:
1265         Teach Replay about PluginLoadClientPolicy.
1266
1267 2015-03-16  Max Stepin  <maxstepin@gmail.com>
1268
1269         Add APNG support
1270         https://bugs.webkit.org/show_bug.cgi?id=17022
1271
1272         Reviewed by Carlos Garcia Campos.
1273
1274         Test: fast/images/animated-png.html
1275
1276         * platform/image-decoders/ImageDecoder.h:
1277         (WebCore::ImageFrame::divide255):
1278         (WebCore::ImageFrame::overRGBA):
1279         * platform/image-decoders/png/PNGImageDecoder.cpp:
1280         (WebCore::frameHeader):
1281         (WebCore::readChunks):
1282         (WebCore::PNGImageReader::PNGImageReader):
1283         (WebCore::PNGImageDecoder::PNGImageDecoder):
1284         (WebCore::PNGImageDecoder::frameBufferAtIndex):
1285         (WebCore::PNGImageDecoder::headerAvailable):
1286         (WebCore::PNGImageDecoder::rowAvailable):
1287         (WebCore::PNGImageDecoder::pngComplete):
1288         (WebCore::PNGImageDecoder::readChunks):
1289         (WebCore::PNGImageDecoder::frameHeader):
1290         (WebCore::PNGImageDecoder::init):
1291         (WebCore::PNGImageDecoder::clearFrameBufferCache):
1292         (WebCore::PNGImageDecoder::initFrameBuffer):
1293         (WebCore::PNGImageDecoder::frameComplete):
1294         (WebCore::PNGImageDecoder::processingStart):
1295         (WebCore::PNGImageDecoder::processingFinish):
1296         (WebCore::PNGImageDecoder::fallbackNotAnimated):
1297         * platform/image-decoders/png/PNGImageDecoder.h:
1298         (WebCore::PNGImageDecoder::frameCount):
1299         (WebCore::PNGImageDecoder::repetitionCount):
1300         (WebCore::PNGImageDecoder::isComplete):
1301
1302 2015-03-15  Benjamin Poulain  <benjamin@webkit.org>
1303
1304         CSS: fix the case-insensitive matching of the attribute selectors Begin, End and Hyphen
1305         https://bugs.webkit.org/show_bug.cgi?id=142715
1306
1307         Reviewed by Brent Fulgham.
1308
1309         Fix attribute matching with:
1310         -Begin: [a^=b].
1311         -End: [a$=b].
1312         -Hyphen: [a|=b].
1313
1314         Tests: fast/selectors/attribute-endswith-value-matching-is-ascii-case-insensitive.html
1315                fast/selectors/attribute-hyphen-value-matching-is-ascii-case-insensitive.html
1316                fast/selectors/attribute-startswith-value-matching-is-ascii-case-insensitive.html
1317
1318         * css/SelectorChecker.cpp:
1319         (WebCore::attributeValueMatches):
1320         I forgot to change CSSSelector::Exact in my last patch.
1321         The tests could not catch that since we use the CSS JIT almost everywhere.
1322
1323         * cssjit/SelectorCompiler.cpp:
1324         (WebCore::SelectorCompiler::attributeValueBeginsWith):
1325         (WebCore::SelectorCompiler::attributeValueEndsWith):
1326         (WebCore::SelectorCompiler::attributeValueMatchHyphenRule):
1327
1328 2015-03-15  Dan Bernstein  <mitz@apple.com>
1329
1330         Fixed the iOS build after r181522.
1331
1332         * page/FrameView.cpp:
1333         (WebCore::FrameView::performPostLayoutTasks):
1334
1335 2015-03-15  Andy Estes  <aestes@apple.com>
1336
1337         [Content Filtering] Adopt new NEFilterSource SPI
1338         https://bugs.webkit.org/show_bug.cgi?id=142710
1339         rdar://problem/19023855
1340
1341         Reviewed by Dan Bernstein.
1342
1343         Teach NetworkExtensionContentFilter to use a new, alternate NEFilterSource SPI on platforms where it is available.
1344
1345         * platform/ContentFilter.cpp:
1346         (WebCore::ContentFilter::types): Renamed HAVE(NE_FILTER_SOURCE) to HAVE(NETWORK_EXTENSION).
1347         * platform/cocoa/NetworkExtensionContentFilter.h: Renamed member variables to remove redundancy, forward-declared NEFilterSourceStatus,
1348         added a dispatch_semaphore member variable to avoid creating and destroying multiple semaphores, and made m_originalData a SharedBuffer.
1349         * platform/cocoa/NetworkExtensionContentFilter.mm:
1350         (decisionInfoReplacementData): Returned the replacement data from a decision handler info dictionary.
1351         (WebCore::createNEFilterSource): Created either an old-style or new-style NEFilterSource object.
1352         (WebCore::NetworkExtensionContentFilter::NetworkExtensionContentFilter): Called receivedResponse:decisionHandler: when using the new SPI.
1353         (WebCore::NetworkExtensionContentFilter::~NetworkExtensionContentFilter): Released the dispatch_semaphore.
1354         (WebCore::NetworkExtensionContentFilter::addData): Appended the copied NSData to m_originalData, avoiding an additional copy previously
1355         being made by NSMutableData. Used the new receivedData:decisionHandler: SPI when appropriate.
1356         (WebCore::NetworkExtensionContentFilter::finishedAddingData): Used the new finishedLoadingWithDecisionHandler: SPI when appropriate.
1357         (WebCore::NetworkExtensionContentFilter::needsMoreData): Changed m_neFilterSourceStatus to m_status.
1358         (WebCore::NetworkExtensionContentFilter::didBlockData): Ditto.
1359         (WebCore::NetworkExtensionContentFilter::getReplacementData): Returned the replacement data from NEFilterSource if the load was blocked.
1360         Otherwise, returned the original data.
1361         (WebCore::NetworkExtensionContentFilter::handleDecision): Added a helper to set m_status and m_replacementData, and to signal m_semaphore.
1362         * platform/spi/cocoa/NEFilterSourceSPI.h: Declared the new NEFilterSource SPI on platforms that support it.
1363
1364 2015-03-15  Brent Fulgham  <bfulgham@apple.com>
1365
1366         Scroll snap points are not supported on iframe content
1367         https://bugs.webkit.org/show_bug.cgi?id=142582
1368         <rdar://problem/20121319>
1369
1370         Tested by css3/scroll-snap/scroll-snap-iframe.html
1371
1372         Reviewed by Simon Fraser.
1373
1374         The scroll snap points were not being applied to the iframe contents because the code
1375         that sets up the scroll snap point content is not called for iframes.
1376
1377         To correct this, we need to make sure the snap offsets are set during post-frame layout
1378         for iframes. We also need to make sure (on Mac) that the scroll animator and timers are updated.
1379
1380         * page/FrameView.cpp:
1381         (WebCore::FrameView::performPostLayoutTasks): Call 'updateSnapOffsets' if the frame is not a
1382         MainFrame. Also notify scroll animators they need to update their snap point settings.
1383
1384 2015-03-15  Simon Fraser  <simon.fraser@apple.com>
1385
1386         Add the same is<RenderBox> test to KeyframeAnimation::computeExtentOfTransformAnimation()
1387         that ImplicitAnimation::computeExtentOfTransformAnimation() has, and change the latter
1388         to the more canonical is<RenderBox>() form.
1389
1390         Fixes an assertion in animations/animation-on-inline-crash.html
1391
1392         * page/animation/ImplicitAnimation.cpp:
1393         (WebCore::ImplicitAnimation::computeExtentOfTransformAnimation):
1394         * page/animation/KeyframeAnimation.cpp:
1395         (WebCore::KeyframeAnimation::computeExtentOfTransformAnimation):
1396
1397 2015-03-15  Simon Fraser  <simon.fraser@apple.com>
1398
1399         And another fix. Thanks to Hunseop Jeong for the fix.
1400
1401         * html/HTMLInputElement.cpp:
1402         (WebCore::HTMLInputElement::runPostTypeUpdateTasks):
1403         (WebCore::HTMLInputElement::didMoveToNewDocument):
1404
1405 2015-03-15  Simon Fraser  <simon.fraser@apple.com>
1406
1407         Fix the touch-event build.
1408
1409         * dom/Document.cpp:
1410         (WebCore::Document::didAddTouchEventHandler):
1411         (WebCore::Document::didRemoveTouchEventHandler):
1412
1413 2015-03-15  Simon Fraser  <simon.fraser@apple.com>
1414
1415         Reduce the side-effects of animations turning off overlap testing
1416         https://bugs.webkit.org/show_bug.cgi?id=92791
1417
1418         Reviewed by Dean Jackson.
1419
1420         When a layer is running a transition or animation of the transform property,
1421         we would simply disable overlap testing for later layers, which had the side-effect
1422         of promoting lots of unrelated elements into layers temporarily.
1423
1424         Fix by maintaining overlap, but computing an overlap extent that takes the animation
1425         into account.
1426
1427         Rotations are currently treated as full rotations. If an extent for the overlap is
1428         hard to compute (e.g. 3d transforms, or matrix animations with a rotation component),
1429         then we fall back to the current behavior.
1430
1431         Tests: compositing/layer-creation/mismatched-rotated-transform-animation-overlap.html
1432                compositing/layer-creation/mismatched-rotated-transform-transition-overlap.html
1433                compositing/layer-creation/mismatched-transform-transition-overlap.html
1434                compositing/layer-creation/multiple-keyframes-animation-overlap.html
1435                compositing/layer-creation/scale-rotation-animation-overlap.html
1436                compositing/layer-creation/scale-rotation-transition-overlap.html
1437                compositing/layer-creation/translate-animation-overlap.html
1438                compositing/layer-creation/translate-scale-animation-overlap.html
1439                compositing/layer-creation/translate-scale-transition-overlap.html
1440                compositing/layer-creation/translate-transition-overlap.html
1441
1442         * page/animation/AnimationBase.cpp:
1443         (WebCore::containsRotation):
1444         (WebCore::AnimationBase::computeTransformedExtentViaTransformList): When we have matched
1445         transform lists, we can map a rectangle through the various operations. Transform-origin
1446         is used to shift the origin of the box first, and then unshift after. If we encounter
1447         a rotation, for now assume it's a full rotation (a future patch could tighten this up).
1448         (WebCore::AnimationBase::computeTransformedExtentViaMatrix): If we're using matrix
1449         interpolation, we have to decompose the matrix to see if there's any rotation component,
1450         and, if there is, fall back to current behavior.
1451         * page/animation/AnimationBase.h:
1452         * page/animation/AnimationController.cpp:
1453         (WebCore::AnimationControllerPrivate::computeExtentOfAnimation):
1454         (WebCore::AnimationController::computeExtentOfAnimation):
1455         * page/animation/AnimationController.h:
1456         * page/animation/AnimationControllerPrivate.h:
1457         * page/animation/CompositeAnimation.cpp:
1458         (WebCore::CompositeAnimation::computeExtentOfTransformAnimation): Ask active keyframe
1459         animations and transitions to compute the bounds extent.
1460         * page/animation/CompositeAnimation.h:
1461         * page/animation/ImplicitAnimation.cpp:
1462         (WebCore::ImplicitAnimation::computeExtentOfTransformAnimation): Compute the extent
1463         of the start and end transforms, and union them.
1464         * page/animation/ImplicitAnimation.h:
1465         * page/animation/KeyframeAnimation.cpp:
1466         (WebCore::KeyframeAnimation::animate):
1467         (WebCore::KeyframeAnimation::getAnimatedStyle): Some nullptr goodness.
1468         (WebCore::KeyframeAnimation::computeExtentOfTransformAnimation): Compute an extent
1469         for each keyframe, and take their union.
1470         * page/animation/KeyframeAnimation.h:
1471         * platform/graphics/GeometryUtilities.cpp:
1472         (WebCore::euclidianDistance): Use Pythagoras to compute a distance.
1473         (WebCore::boundsOfRotatingRect): Given a rect whose location is relative
1474         to the rotation origin, compute a bounds for the rotated rect by computing
1475         the furthest corner from the origin, and sweeping out a circle.
1476         * platform/graphics/GeometryUtilities.h:
1477         * platform/graphics/transforms/Matrix3DTransformOperation.h:
1478         * platform/graphics/transforms/MatrixTransformOperation.h:
1479         * platform/graphics/transforms/PerspectiveTransformOperation.h:
1480         * platform/graphics/transforms/RotateTransformOperation.h:
1481         * platform/graphics/transforms/ScaleTransformOperation.h:
1482         * platform/graphics/transforms/SkewTransformOperation.h:
1483         * platform/graphics/transforms/TransformOperation.h:
1484         (WebCore::TransformOperation::isAffectedByTransformOrigin):
1485         * platform/graphics/transforms/TransformOperations.cpp:
1486         (WebCore::TransformOperations::affectedByTransformOrigin): Ask all the operations if
1487         they are affected by transform-origin.
1488         (WebCore::TransformOperations::blendByMatchingOperations): nullptr.
1489         * platform/graphics/transforms/TransformOperations.h:
1490         * rendering/RenderBox.cpp:
1491         (WebCore::RenderBox::pushMappingToContainer): Comment fix. Only take transforms into account
1492         if the geometry map says so (which is most of the time).
1493         * rendering/RenderGeometryMap.cpp:
1494         (WebCore::RenderGeometryMap::mapToContainer): RenderLayerCompositor is now using the
1495         geometry map in a way that is incompatible with this assertion; it deliberately ignores
1496         transforms sometimes, so we can't easily verify that the mapping matches mapping through
1497         renderers.
1498         (WebCore::RenderGeometryMap::pushMappingsToAncestor): Save and restore the UseTransforms
1499         bit.
1500         * rendering/RenderGeometryMap.h:
1501         * rendering/RenderLayer.cpp:
1502         (WebCore::RenderLayer::boundingBox): Whitespace.
1503         (WebCore::RenderLayer::getOverlapBoundsIncludingChildrenAccountingForTransformAnimations): Helper
1504         function to get the bounds of a layer, including descendants, when a transform animation is running.
1505         * rendering/RenderLayer.h:
1506         * rendering/RenderLayerCompositor.cpp:
1507         (WebCore::RenderLayerCompositor::CompositingState::CompositingState): Add a ancestorHasTransformAnimation
1508         flag to detect nested animated transforms.
1509         (WebCore::RenderLayerCompositor::OverlapExtent::knownToBeHaveExtentUncertainty): This returns true when
1510         the layer is animating transform, and the transition/animation is such that we can't easily compute the
1511         bounds of the animation.
1512         (WebCore::RenderLayerCompositor::computeExtent): const RenderLayer&.
1513         Compute the animated bounds if there's a transform animation running.
1514         (WebCore::RenderLayerCompositor::addToOverlapMap): const RenderLayer&
1515         (WebCore::RenderLayerCompositor::addToOverlapMapRecursive): const RenderLayer&
1516         (WebCore::RenderLayerCompositor::computeCompositingRequirements):
1517         Delay the call to pushMappingsToAncestor() until knowing if there's a transform animation running, and
1518         if there is, push the mapping while ignoring transforms (since the transform is implicitly taken into account
1519         for overlap via the computed animated bounds).
1520         If this layer is running a transform animation, set the childState.ancestorHasTransformAnimation flag so
1521         that descendants will know (nested transform animations fall back to current behavior).
1522         The if (.... && isRunningAcceleratedTransformAnimation()) is what previously caused us to turn off overlap
1523         testing in the face of animations. That now only happens if we were unable to easily compute the animation bounds.
1524         (WebCore::RenderLayerCompositor::isRunningTransformAnimation): This previously tested whether an accelerated animation
1525         was running, but that's timing sensitive; AnimationController can start the transform animation, but it's not yet
1526         considered accelerated until we get an async callback from GraphicsLayer, yet this code needed to know if the
1527         animation was running.
1528         Since transform animations are always accelerated, we can just test for a running transform animation.
1529         (WebCore::RenderLayerCompositor::isRunningAcceleratedTransformAnimation): Deleted.
1530         * rendering/RenderLayerCompositor.h:
1531         * rendering/style/RenderStyle.cpp:
1532         (WebCore::requireTransformOrigin): Some FIXME comments.
1533
1534 2015-03-15  Simon Fraser  <simon.fraser@apple.com>
1535
1536         Clean up related to wheelEvent names
1537         https://bugs.webkit.org/show_bug.cgi?id=142713
1538
1539         Reviewed by Anders Carlsson.
1540
1541         Add EventNames::isWheelEventType() and use it in places that test for the
1542         two wheel event names.
1543
1544         * dom/Document.cpp:
1545         (WebCore::Document::didAddWheelEventHandler):
1546         (WebCore::Document::didRemoveWheelEventHandler):
1547         * dom/Document.h: No need for exports (I grepped). Pass the Node*, which
1548         will be used in a later patch.
1549         * dom/EventNames.h:
1550         (WebCore::EventNames::isWheelEventType):
1551         * dom/Node.cpp:
1552         (WebCore::Node::didMoveToNewDocument):
1553         (WebCore::tryAddEventListener):
1554         (WebCore::tryRemoveEventListener):
1555         (WebCore::Node::defaultEventHandler):
1556         * html/shadow/MediaControlsApple.cpp:
1557         (WebCore::MediaControlsAppleEventListener::handleEvent):
1558         * page/DOMWindow.cpp:
1559         (WebCore::DOMWindow::addEventListener):
1560         (WebCore::DOMWindow::removeEventListener):
1561         * page/ios/FrameIOS.mm:
1562         (WebCore::ancestorRespondingToScrollWheelEvents): Remove dead code.
1563
1564 2015-03-15  Simon Fraser  <simon.fraser@apple.com>
1565
1566         Remove a redundant repaint when a layer becomes composited
1567         https://bugs.webkit.org/show_bug.cgi?id=142711
1568
1569         Reviewed by Anders Carlsson.
1570
1571         RenderLayerCompositor::computeCompositingRequirements() doesn't need to call
1572         repaintOnCompositingChange() when a layer is going to become composited,
1573         because updateBacking() does exactly the same thing. I used an assertion
1574         and ran the tests to ensure this wasn't a behavior change.
1575
1576         * rendering/RenderLayerCompositor.cpp:
1577         (WebCore::RenderLayerCompositor::computeCompositingRequirements):
1578
1579 2015-03-15  Benjamin Poulain  <bpoulain@apple.com>
1580
1581         Change the exact attribute matching to be ASCII case-insensitive
1582         https://bugs.webkit.org/show_bug.cgi?id=142609
1583
1584         Reviewed by Darin Adler.
1585
1586         In CSS, testing attribute values should be ASCII case-insensitive,
1587         previously we were using full unicode case conversion.
1588
1589         Test: fast/selectors/attribute-exact-value-match-is-ascii-case-insensitive.html
1590
1591         * css/CSSParser.cpp:
1592         (WebCore::CSSParser::parseKeyframeSelector):
1593         The CSS parser has its own fast version for ASCII case insensitive.
1594         This code was using the general equalIgnoringASCIICase() which was causing name conflicts,
1595         change that to the normal CSS parser version.
1596
1597         * css/SelectorCheckerTestFunctions.h:
1598         (WebCore::equalIgnoringASCIICase): Deleted.
1599         * cssjit/SelectorCompiler.cpp:
1600         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementAttributeValueExactMatching):
1601
1602 2015-03-15  Brent Fulgham  <bfulgham@apple.com>
1603
1604         scroll snap points do not properly account for zoomed pages
1605         https://bugs.webkit.org/show_bug.cgi?id=142706
1606         <rdar://problem/20165771>
1607
1608         Reviewed by Anders Carlsson.
1609
1610         When a WebView is zoomed (such that it has a non-unity pageScaleFactor), we need to account for this
1611         scaling value when selecting our correct scroll snap point target, as well as when specifying the
1612         pixel location for our animation to target.
1613
1614         * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.h:
1615         * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
1616         (WebCore::ScrollingTreeFrameScrollingNodeMac::pageScaleFactor): Added new delegate method.
1617         * platform/cocoa/ScrollController.h:
1618         (WebCore::ScrollControllerClient::pageScaleFactor): Added new default delegate.
1619         * platform/cocoa/ScrollController.mm:
1620         (WebCore::ScrollController::beginScrollSnapAnimation): Calculate the correct scroll target
1621         based on the page scale factor.
1622
1623 2015-03-15  Csaba Osztrogonác  <ossy@webkit.org>
1624
1625         Fix run-bindings-tests on the WinCairo bot
1626         https://bugs.webkit.org/show_bug.cgi?id=142588
1627
1628         Reviewed by Alex Christensen.
1629
1630         * bindings/scripts/test/JS/JSFloat64Array.cpp: Added property svn:eol-style.
1631         * bindings/scripts/test/JS/JSFloat64Array.h: Added property svn:eol-style.
1632         * bindings/scripts/test/JS/JSTestActiveDOMObject.cpp: Added property svn:eol-style.
1633         * bindings/scripts/test/JS/JSTestActiveDOMObject.h: Modified property svn:eol-style.
1634         * bindings/scripts/test/JS/JSTestCallback.cpp: Modified property svn:eol-style.
1635         * bindings/scripts/test/JS/JSTestCallback.h: Modified property svn:eol-style.
1636         * bindings/scripts/test/JS/JSTestCustomNamedGetter.cpp: Added property svn:eol-style.
1637         * bindings/scripts/test/JS/JSTestCustomNamedGetter.h: Modified property svn:eol-style.
1638         * bindings/scripts/test/JS/JSTestEventConstructor.cpp: Added property svn:eol-style.
1639         * bindings/scripts/test/JS/JSTestEventConstructor.h: Added property svn:eol-style.
1640         * bindings/scripts/test/JS/JSTestEventTarget.cpp: Added property svn:eol-style.
1641         * bindings/scripts/test/JS/JSTestEventTarget.h: Modified property svn:eol-style.
1642         * bindings/scripts/test/JS/JSTestException.cpp: Added property svn:eol-style.
1643         * bindings/scripts/test/JS/JSTestException.h: Added property svn:eol-style.
1644         * bindings/scripts/test/JS/JSTestGenerateIsReachable.cpp: Added property svn:eol-style.
1645         * bindings/scripts/test/JS/JSTestGenerateIsReachable.h: Added property svn:eol-style.
1646         * bindings/scripts/test/JS/JSTestImplements.cpp: Added property svn:eol-style.
1647         * bindings/scripts/test/JS/JSTestImplements.h: Added property svn:eol-style.
1648         * bindings/scripts/test/JS/JSTestInterface.cpp: Modified property svn:eol-style.
1649         * bindings/scripts/test/JS/JSTestInterface.h: Modified property svn:eol-style.
1650         * bindings/scripts/test/JS/JSTestMediaQueryListListener.cpp: Modified property svn:eol-style.
1651         * bindings/scripts/test/JS/JSTestMediaQueryListListener.h: Modified property svn:eol-style.
1652         * bindings/scripts/test/JS/JSTestNamedConstructor.cpp: Added property svn:eol-style.
1653         * bindings/scripts/test/JS/JSTestNamedConstructor.h: Added property svn:eol-style.
1654         * bindings/scripts/test/JS/JSTestNondeterministic.cpp: Added property svn:eol-style.
1655         * bindings/scripts/test/JS/JSTestNondeterministic.h: Added property svn:eol-style.
1656         * bindings/scripts/test/JS/JSTestObj.cpp: Modified property svn:eol-style.
1657         * bindings/scripts/test/JS/JSTestObj.h: Modified property svn:eol-style.
1658         * bindings/scripts/test/JS/JSTestOverloadedConstructors.cpp: Added property svn:eol-style.
1659         * bindings/scripts/test/JS/JSTestOverloadedConstructors.h: Added property svn:eol-style.
1660         * bindings/scripts/test/JS/JSTestSerializedScriptValueInterface.cpp: Modified property svn:eol-style.
1661         * bindings/scripts/test/JS/JSTestSerializedScriptValueInterface.h: Modified property svn:eol-style.
1662         * bindings/scripts/test/JS/JSTestSupplemental.cpp: Added property svn:eol-style.
1663         * bindings/scripts/test/JS/JSTestSupplemental.h: Added property svn:eol-style.
1664         * bindings/scripts/test/JS/JSTestTypedefs.cpp: Added property svn:eol-style.
1665         * bindings/scripts/test/JS/JSTestTypedefs.h: Added property svn:eol-style.
1666         * bindings/scripts/test/JS/JSattribute.cpp: Added property svn:eol-style.
1667         * bindings/scripts/test/JS/JSattribute.h: Added property svn:eol-style.
1668         * bindings/scripts/test/JS/JSreadonly.cpp: Added property svn:eol-style.
1669         * bindings/scripts/test/JS/JSreadonly.h: Added property svn:eol-style.
1670
1671 2015-03-14  Darin Adler  <darin@apple.com>
1672
1673         More event handler improvements
1674         https://bugs.webkit.org/show_bug.cgi?id=142701
1675
1676         Reviewed by Anders Carlsson.
1677
1678         These are the improvements:
1679
1680         - Use EventHandler rather than EventListener as the the type for event handler
1681           attributes. This matches the HTML specification, and also makes sense, since
1682           EventListener means something else (and we use it to mean that!). Also renamed
1683           JSWindowEventListener to WindowEventHandler. Even though this only affects the
1684           JavaScript code generated, it's not really a JavaScript-specific flag.
1685
1686         - Tweak formatting on addEventListener/removeEventListener/dispatchEvent in
1687           all the IDL files where the appear. This includes changing the spelling from
1688           "evt" to "event". Some day we should fix this so these functions only need to
1689           appear in EventTarget.idl.
1690
1691         - Tweak formatting a bit on the IDL files we had to modify anyway.
1692
1693         - Use [Conditional] more often and #if less often in IDL files.
1694
1695         - Added a new [DocumentEventHandler] attribute for the selectionchange event.
1696           This involved adding new event handler attribute functions to JSEventListener.h
1697           for use by the JavaScript bindings.
1698
1699         - Removed a little unused code from the JavaScript code bindings generator.
1700
1701         - Improved the mechanism used by HTMLElement and SVGElement to share the list of
1702           content attributes that are event handlers so there is only one map rather than
1703           two. Made a similar mechanism so that HTMLBodyElement and HTMLFrameSetElement
1704           can share the list of window event handlers.
1705
1706         - Followed the HTML specification by putting all the event handler support in
1707           the HTMLElement class rather than having event handlers apply only a the
1708           particular element that uses those events. We already did this for most of
1709           our event handlers, but we are now doing it for all of them.
1710
1711         * Modules/battery/BatteryManager.idl: Use EventHandler instead of EventListener
1712         as appropriate. Also reformatted addEventListener/removeEventListener/dispatchEvent.
1713         * Modules/encryptedmedia/MediaKeySession.idl: Ditto.
1714         * Modules/indexeddb/IDBDatabase.idl: Ditto.
1715         * Modules/indexeddb/IDBOpenDBRequest.idl: Ditto.
1716         * Modules/indexeddb/IDBRequest.idl: Ditto.
1717         * Modules/indexeddb/IDBTransaction.idl: Ditto.
1718         * Modules/mediastream/MediaStream.idl: Ditto.
1719         * Modules/mediastream/MediaStreamTrack.idl: Ditto.
1720         * Modules/mediastream/RTCDTMFSender.idl: Ditto.
1721         * Modules/mediastream/RTCDataChannel.idl: Ditto.
1722         * Modules/mediastream/RTCPeerConnection.idl: Ditto.
1723         * Modules/speech/SpeechSynthesisUtterance.idl: Ditto.
1724         * Modules/webaudio/AudioBufferSourceNode.idl: Ditto.
1725         * Modules/webaudio/AudioContext.idl: Ditto.
1726         * Modules/webaudio/OscillatorNode.idl: Ditto.
1727         * Modules/webaudio/ScriptProcessorNode.idl: Ditto.
1728         * Modules/websockets/WebSocket.idl: Ditto.
1729         * css/FontLoader.idl: Ditto.
1730         * dom/EventListener.idl: Ditto.
1731         * dom/EventTarget.idl: Ditto.
1732         * dom/MessagePort.idl: Ditto.
1733         * dom/Node.idl: Ditto.
1734         * dom/WebKitNamedFlow.idl: Ditto.
1735         * fileapi/FileReader.idl: Ditto.
1736         * html/MediaController.idl: Ditto.
1737         * html/track/AudioTrackList.idl: Ditto.
1738         * html/track/TextTrackCue.idl: Ditto.
1739         * html/track/TextTrackList.idl: Ditto.
1740         * html/track/VideoTrackList.idl: Ditto.
1741         * loader/appcache/DOMApplicationCache.idl: Ditto.
1742         * page/EventSource.idl: Ditto.
1743         * page/Performance.idl: Ditto.
1744         * workers/AbstractWorker.idl: Ditto.
1745         * workers/DedicatedWorkerGlobalScope.idl: Ditto.
1746         * workers/Worker.idl: Ditto.
1747         * workers/WorkerGlobalScope.idl: Ditto.
1748         * xml/XMLHttpRequest.idl: Ditto.
1749         * xml/XMLHttpRequestUpload.idl: Ditto.
1750
1751         * Modules/notifications/Notification.idl: Did the above, but also used
1752         [Conditional] instead of #if throughout.
1753         * html/track/TextTrack.idl: Ditto.
1754
1755         * Modules/webaudio/AudioNode.idl: Tweaked paragraphing of this file.
1756
1757         * bindings/js/JSEventListener.cpp:
1758         (WebCore::windowEventHandlerAttribute): Renamed to take the word "forwarded"
1759         out of this. More closely matches the terminology used in IDL files and the
1760         HTML specification.
1761         (WebCore::setWindowEventHandlerAttribute): Ditto.
1762         (WebCore::documentEventHandlerAttribute): Added.
1763         (WebCore::setDocumentEventHandlerAttribute): Added.
1764         * bindings/js/JSEventListener.h: Updated for above changes.
1765
1766         * bindings/objc/PublicDOMInterfaces.h: Renamed argument from "evt" to "event".
1767
1768         * bindings/scripts/CodeGeneratorGObject.pm:
1769         (SkipAttribute): Skip attributes of type "EventHandler" rather than attributes
1770         of type "EventListener".
1771         * bindings/scripts/CodeGeneratorObjC.pm:
1772         (SkipAttribute): Ditto.
1773
1774         * bindings/scripts/CodeGeneratorJS.pm:
1775         (GenerateImplementation): Look for the type EventHandler instead of the type
1776         EventListener for event handler attributes. Also added code to handle the new
1777         DocumentEventHandler, and use the name WindowEventHandler instead of the name
1778         JSWindowEventListener. Removed unneeded preflight check to see if we have
1779         writable attributes; it was not doing us any good. (That caused a lot of code
1780         to be un-indented and makes the diff hard to read.)
1781
1782         * bindings/scripts/IDLAttributes.txt: Removed JSWindowEventListener, and added
1783         WindowEventHandler and DocumentEventHandler.
1784
1785         * bindings/scripts/test/TestObj.idl: Use the type EventHandler instead of the
1786         type EventListener. The test output is unchanged, though.
1787
1788         * dom/Document.idl: Got rid of the conditionals and merged all the event handler
1789         attributes into a single sorted, unconditional list. Added some that were missing,
1790         as detected by the event-handler-attributes.html test.
1791
1792         * dom/Element.idl: Ditto. This includes moving attributes here that were formerly
1793         only in certain element classes. Note that the script event handler attribute
1794         support is still here, even though it should be in HTMLElement and SVGElement
1795         instead. There's a FIXME about that, but no real urgency in fixing it.
1796
1797         * html/HTMLAttributeNames.in: Added onmessage. Previously, the support for this
1798         was from the script attribute only, not the content attribute.
1799
1800         * html/HTMLBodyElement.cpp:
1801         (WebCore::HTMLBodyElement::createWindowEventHandlerNameMap): Added. This
1802         contains the list of all the window event handlers that can be set on a body
1803         or frameset element and which are triggered by events on the window.
1804         (WebCore::HTMLBodyElement::eventNameForWindowEventHandlerAttribute): Added.
1805         This is the function to call to use the map above. Given an attribute, it
1806         returns either null if it is not a window event handler attribute, or the
1807         event type if it is one.
1808         (WebCore::HTMLBodyElement::parseAttribute): Updated to use the new
1809         functions above. Handling of these attributes is now unconditional, but
1810         also we don't have to keep the nested if statements here up to date, just
1811         the list of event handler attributes names in the map create function above.
1812
1813         * html/HTMLBodyElement.h: Added public eventNameForWindowEventHandlerAttribute
1814         and private createWindowEventHandlerNameMap functions.
1815
1816         * html/HTMLBodyElement.idl: Updated to use WindowEventHandler, DocumentEventHandler,
1817         and EventHandler. Also made everything unconditional. Also filled out the list here
1818         to match the list handled as content attributes. All covered by the test.
1819
1820         * html/HTMLElement.cpp:
1821         (WebCore::HTMLElement::createEventHandlerNameMap): Added. Replaces the old
1822         populate function. Changed the list of event handler attributes to be a bit more
1823         complete, and to be entirely unconditional. Also refactored this function to use
1824         a new populateEventHandlerNameMap helper, shared with HTMLBodyElement.
1825         (WebCore::HTMLElement::populateEventHandlerNameMap): Added. Factors out the code
1826         that both this class and HTMLBodyElement use to build event handler name maps.
1827         (WebCore::HTMLElement::eventNameForEventHandlerAttribute): Added. This is the
1828         function that call to use the map above. Given an attribute it returns either
1829         null if it is not an event handler attribute, or the event type if it is one.
1830         This is actually two functions. One is a protected function for use by both this
1831         class and HTMLBodyElement so they can share things like the optimization to look
1832         for the "on" prefix. The other is the public function that we actually use in
1833         HTMLElement and SVGElement.
1834         (WebCore::HTMLElement::editabilityFromContentEditableAttr): Tweaked and refactored
1835         to use lineageOfType. Would have been even simpler if this took an element instead
1836         of a node. Unrelated to the event handler changes.
1837         (WebCore::HTMLElement::parseAttribute): Removed long-obsolete code that decided
1838         whether to call through to the base class. The base class function is empty and
1839         never needs to be called, and in any case there is no value in doing work to
1840         decide whether to call through to an empty function. Changed the style of the
1841         function to use early return instead of else. Worth considering whether we want
1842         to return early or call through to base class in this family of functions. It's
1843         more efficient to return early, but doesn't work well if both the derived class
1844         and base class want to respond to changes to the same attribute. The new logic
1845         for event handler attributes is more straightforward than the old, since the
1846         eventNameForEventHandlerAttribute has the logic.
1847         (WebCore::HTMLElement::textToFragment): Tweaked and refactored a bit, and also
1848         changed to return a Ref since this never fails and needs to return null.
1849
1850         * html/HTMLElement.h: Updated for above changes. This includes a template version
1851         of populateEventHandlerNameMap that extracts the array size at compile time and
1852         passes it to the non-template function that does the work.
1853
1854         * html/HTMLFrameElementBase.cpp:
1855         (WebCore::HTMLFrameElementBase::parseAttribute): Removed unneeded code to handle
1856         event handler attributes handled by HTMLElement.
1857         * html/HTMLImageElement.cpp:
1858         (WebCore::HTMLImageElement::parseAttribute): Ditto.
1859         * html/HTMLLinkElement.cpp:
1860         (WebCore::HTMLLinkElement::parseAttribute): Ditto.
1861         * html/HTMLObjectElement.cpp:
1862         (WebCore::HTMLObjectElement::parseAttribute): Ditto.
1863         * html/HTMLScriptElement.cpp:
1864         (WebCore::HTMLScriptElement::parseAttribute): Ditto.
1865
1866         * html/HTMLFrameSetElement.cpp:
1867         (WebCore::HTMLFrameSetElement::parseAttribute): Changed function to early return
1868         style, and added FIXMEs about the many problems in the attribute handling code.
1869         Replaced all the code to handle window event handlers with a new bit of code that
1870         calls HTMLBodyElement::eventNameForWindowEventHandlerAttribute.
1871
1872         * html/HTMLFrameSetElement.idl: Changed to match the list of window event handlers
1873         in HTMLBodyElement.idl, although I did not add the document event handler here.
1874         As in the various other cases, having some extra event handlers does not seem to
1875         do harm and this is covered by the event-handler-attributes.html test.
1876
1877         * html/HTMLMarqueeElement.idl: Renamed EventListener to EventHandler in comment.
1878
1879         * page/DOMWindow.idl: As with Element and Document, removed conditionals, and
1880         filled out the list of event handlers so all the tests in
1881         event-handler-attributes.html will pass.
1882
1883         * svg/SVGElement.cpp:
1884         (WebCore::SVGElement::parseAttribute): Changed code to the early return style,
1885         and replaced the event handler attribute code with new much simpler code that
1886         uses the new HTMLElement::eventNameForEventHandlerAttribute. Also changed the
1887         way we call through to base classes. Just call through to the
1888         SVGLangSpace::parseAttribute function unconditionally, and don't try to use
1889         early return style to arbitrate among base classes. We should make this
1890         simplification throughout the SVG code; there's no need for the complexity
1891         that was there before just to cut down slightly on calls through to base
1892         class parseAttribute functions.
1893
1894         * svg/SVGSVGElement.cpp:
1895         (WebCore::SVGSVGElement::parseAttribute): Changed some of this code to the
1896         early return style and corrected some comments about the window event handler
1897         attributes here. These could use some further testing and might later need to be
1898         properly supported when the attributes are set in script, not just in content.
1899
1900         * svg/SVGScriptElement.cpp:
1901         (WebCore::SVGScriptElement::isSupportedAttribute): Deleted.
1902         (WebCore::SVGScriptElement::parseAttribute): Changed this function to use the
1903         early return style and also to call through to all three base classes. This is
1904         a pattern we should follow elsewhere in SVG to simplify the code. There is no
1905         need for the supportedAttributes sets like the one in this calass, and the code
1906         is unnecessarily complex, perhaps in an attempt to optimize performance. I'm
1907         pretty sure the old code was slower than this new code will be. No need for the
1908         extra hash table lookup every time. Also removed handling of event handler
1909         attribute which is taken care of by SVGElement now.
1910         (WebCore::SVGScriptElement::svgAttributeChanged): Made similar changes for
1911         the same reason as in parseAttribute. This function really needs a new name:
1912         It's the same as parseAttribute, except it's also used when implementing
1913         changes due to SVG animation.
1914
1915         * svg/SVGScriptElement.h: Removed isSupportedAttribute.
1916
1917 2015-03-14  Simon Fraser  <simon.fraser@apple.com>
1918
1919         Clean up use of flags in localToContainer-type functions
1920         https://bugs.webkit.org/show_bug.cgi?id=142704
1921
1922         Reviewed by Alexey Proskuryakov.
1923
1924         RenderObject::localToContainerQuad() had the questionable behavior of always
1925         enforcing UseTransforms in the flags. However, a future patch will need to call
1926         localToContainerQuad() without this flag.
1927
1928         Fix by requiring callers of localToAbsoluteQuad(), localToContainerQuad(), localToContainerPoint(),
1929         and absoluteToLocalQuad() to pass the UseTransforms flag in, providing it as a default argument.
1930
1931         The default value of the MapCoordinatesFlags parameter to mapLocalToContainer(),
1932         a lower-level function, is removed.
1933
1934         Sprinkle a few more nullptrs around.
1935
1936         No behavior change.
1937
1938         * rendering/RenderBlock.cpp:
1939         (WebCore::RenderBlock::selectionGapRectsForRepaint): call the higher-level localToContainerPoint()
1940         instead of mapLocalToContainer().
1941         (WebCore::RenderBlock::absoluteQuads): Pass UseTransforms.
1942         * rendering/RenderBox.cpp:
1943         (WebCore::RenderBox::absoluteQuads): Ditto.
1944         * rendering/RenderBox.h:
1945         * rendering/RenderFlowThread.h:
1946         * rendering/RenderGeometryMap.h:
1947         (WebCore::RenderGeometryMap::absolutePoint):
1948         (WebCore::RenderGeometryMap::absoluteRect):
1949         * rendering/RenderImage.cpp:
1950         (WebCore::RenderImage::collectSelectionRects): This function was erroneously passing
1951         'false' as flags. Pass UseTransforms instead (but no behavior change since
1952         UseTransforms was forced on lower down).
1953         * rendering/RenderInline.h:
1954         * rendering/RenderLayer.cpp:
1955         (WebCore::RenderLayer::scrollRectToVisible): UseTransforms is the parameter default, remove it.
1956         * rendering/RenderLineBreak.cpp:
1957         (WebCore::RenderLineBreak::absoluteQuads):
1958         (WebCore::RenderLineBreak::collectSelectionRects): Another erroneous 'false'.
1959         * rendering/RenderNamedFlowFragment.cpp:
1960         (WebCore::RenderNamedFlowFragment::absoluteQuadsForBoxInRegion):
1961         * rendering/RenderObject.cpp:
1962         (WebCore::RenderObject::localToContainerQuad): Here's where we no longer force
1963         the UseTransforms bit.
1964         (WebCore::RenderObject::localToContainerPoint): Also here.
1965         * rendering/RenderObject.h: I prefer bit flags lined up. Makes it easier to spot errors.
1966         (WebCore::RenderObject::localToAbsoluteQuad):
1967         * rendering/RenderText.cpp:
1968         (WebCore::RenderText::collectSelectionRects): Another bad 'false'.
1969         * rendering/RenderTextLineBoxes.cpp:
1970         (WebCore::RenderTextLineBoxes::absoluteRectsForRange):
1971         (WebCore::RenderTextLineBoxes::absoluteQuads):
1972         (WebCore::RenderTextLineBoxes::absoluteQuadsForRange):
1973         * rendering/RenderView.h:
1974         * rendering/SimpleLineLayoutFunctions.cpp:
1975         (WebCore::SimpleLineLayout::collectAbsoluteQuads):
1976         * rendering/svg/RenderSVGForeignObject.h:
1977         * rendering/svg/RenderSVGInline.cpp:
1978         (WebCore::RenderSVGInline::absoluteQuads): Another bad 'false'.
1979         * rendering/svg/RenderSVGInline.h:
1980         * rendering/svg/RenderSVGModelObject.cpp:
1981         (WebCore::RenderSVGModelObject::absoluteQuads):
1982         * rendering/svg/RenderSVGModelObject.h:
1983         * rendering/svg/RenderSVGRoot.h:
1984         * rendering/svg/RenderSVGText.cpp:
1985         (WebCore::RenderSVGText::absoluteQuads):
1986         * rendering/svg/RenderSVGText.h:
1987         * rendering/svg/SVGRenderSupport.h:
1988
1989 2015-03-14  Brent Fulgham  <bfulgham@apple.com>
1990
1991         [iOS] scroll snap points are animating to the wrong positions...
1992         https://bugs.webkit.org/show_bug.cgi?id=142705
1993         <rdar://problem/20136946>
1994
1995         Reviewed by Simon Fraser.
1996
1997         Avoid adding an extra '0' snap point to our set. We always start with one zero; this
1998         extra append just forces us to do more steps in our search for nearest snap point.
1999
2000         * page/scrolling/AxisScrollSnapOffsets.cpp:
2001         (WebCore::updateFromStyle): Remove extra '0' appended to offsets.
2002
2003 2015-03-14  Dean Jackson  <dino@apple.com>
2004
2005         Feature flag for Animations Level 2
2006         https://bugs.webkit.org/show_bug.cgi?id=142699
2007         <rdar://problem/20165097>
2008
2009         Reviewed by Brent Fulgham.
2010
2011         Add ENABLE_CSS_ANIMATIONS_LEVEL_2 and a runtime flag animationTriggersEnabled.
2012
2013         * Configurations/FeatureDefines.xcconfig:
2014         * bindings/generic/RuntimeEnabledFeatures.cpp:
2015         (WebCore::RuntimeEnabledFeatures::RuntimeEnabledFeatures):
2016         * bindings/generic/RuntimeEnabledFeatures.h:
2017         (WebCore::RuntimeEnabledFeatures::setAnimationTriggersEnabled):
2018         (WebCore::RuntimeEnabledFeatures::animationTriggersEnabled):
2019
2020 2015-03-14  Adenilson Cavalcanti  <cavalcantii@gmail.com>
2021
2022         RenderBlock::imageChange() calling const methods on exit
2023         https://bugs.webkit.org/show_bug.cgi?id=142648
2024
2025         Reviewed by Brent Fulgham.
2026
2027         No new tests, no change on behavior.
2028
2029         * rendering/RenderBlock.cpp:
2030         (WebCore::RenderBlock::imageChanged): Deleted.
2031         * rendering/RenderBlock.h:
2032
2033 2015-03-14  Víctor Manuel Jáquez Leal  <vjaquez@igalia.com>
2034
2035         [GStreamer] share GL context in pipeline
2036         https://bugs.webkit.org/show_bug.cgi?id=142693
2037
2038         Reviewed by Philippe Normand.
2039
2040         GstGL elements in a pipeline need to be aware of the application's
2041         display and its GL context. This information is shared through context
2042         messages between the pipeline and the browser.
2043
2044         This patch shares this context through a GStreamer's synchronous
2045         message, using the GL information held in the web process.
2046
2047         This patch is based on the work of Philippe Normand for Bug 138562.
2048
2049         No new tests because this is platform specific and it depends in the
2050         run-time availability and configurations of GstGL elements.
2051
2052         * PlatformGTK.cmake: appends the GstGL header files in the include
2053         directories. Also its library directory is appended.
2054         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
2055         (WebCore::mediaPlayerPrivateSyncMessageCallback): New callback function.
2056         (WebCore::MediaPlayerPrivateGStreamer::MediaPlayerPrivateGStreamer):
2057         Initialize the new class attributes.
2058         (WebCore::MediaPlayerPrivateGStreamer::handleSyncMessage): New method
2059         for handling synchronous messages from the pipeline. This method
2060         currently only handles the GL context sharing.
2061         (WebCore::MediaPlayerPrivateGStreamer::createGSTPlayBin): Configures
2062         the pipeline's bus to handle the synchronous messages.
2063         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.h: Add new
2064         class methods and attributes.
2065
2066 2015-03-13  Alex Christensen  <achristensen@webkit.org>
2067
2068         Progress towards CMake on Mac.
2069         https://bugs.webkit.org/show_bug.cgi?id=142680
2070
2071         Reviewed by Gyuyoung Kim.
2072
2073         * PlatformMac.cmake:
2074         Added new include directories and removed old source files.
2075
2076 2015-03-13  Commit Queue  <commit-queue@webkit.org>
2077
2078         Unreviewed, rolling out r181483.
2079         https://bugs.webkit.org/show_bug.cgi?id=142688
2080
2081         Caused use-after-free on many tests (Requested by ap on
2082         #webkit).
2083
2084         Reverted changeset:
2085
2086         "Allow clients to selectively disable plug-ins"
2087         https://bugs.webkit.org/show_bug.cgi?id=142506
2088         http://trac.webkit.org/changeset/181483
2089
2090 2015-03-13  Antti Koivisto  <antti@apple.com>
2091
2092         Cache glyph widths to GlyphPages
2093         https://bugs.webkit.org/show_bug.cgi?id=142028
2094
2095         Reviewed by Andreas Kling.
2096
2097         Currently we have a separate cache in Font for glyph widths. In practice we always need
2098         the widths so we can just cache them in GlyphPages. This simplifies the code and removes
2099         a per-character hash lookup from WidthIterator.
2100
2101         * platform/graphics/Font.cpp:
2102         (WebCore::Font::Font):
2103         (WebCore::Font::initCharWidths):
2104         (WebCore::Font::platformGlyphInit):
2105         (WebCore::createAndFillGlyphPage):
2106         (WebCore::Font::computeWidthForGlyph):
2107
2108             Rename to make it clear this doesn't cache.
2109
2110         (WebCore::GlyphPage::setGlyphDataForIndex):
2111
2112             Initialize the width.
2113             This could go to GlyphPage.cpp if we had one.
2114
2115         * platform/graphics/Font.h:
2116         (WebCore::Font::glyphZeroWidth):
2117         (WebCore::Font::isZeroWidthSpaceGlyph):
2118         (WebCore::Font::zeroGlyph): Deleted.
2119         (WebCore::Font::setZeroGlyph): Deleted.
2120         (WebCore::Font::widthForGlyph): Deleted.
2121         * platform/graphics/FontCascade.cpp:
2122         (WebCore::offsetToMiddleOfGlyph):
2123         * platform/graphics/FontCascadeFonts.cpp:
2124         (WebCore::FontCascadeFonts::glyphDataForCharacter):
2125         * platform/graphics/GlyphPage.h:
2126         (WebCore::GlyphData::GlyphData):
2127
2128             Return width too as part of GlyphData.
2129
2130         (WebCore::GlyphPage::glyphDataForIndex):
2131         (WebCore::GlyphPage::setGlyphDataForCharacter):
2132         (WebCore::GlyphPage::setGlyphDataForIndex):
2133         (WebCore::GlyphPage::GlyphPage):
2134         * platform/graphics/WidthIterator.cpp:
2135         (WebCore::WidthIterator::advanceInternal):
2136
2137             No need to lookup width separately now.
2138
2139         * platform/graphics/mac/ComplexTextController.cpp:
2140         (WebCore::ComplexTextController::adjustGlyphsAndAdvances):
2141         * platform/graphics/mac/ComplexTextControllerCoreText.mm:
2142         (WebCore::ComplexTextController::ComplexTextRun::ComplexTextRun):
2143         * rendering/mathml/RenderMathMLOperator.cpp:
2144         (WebCore::RenderMathMLOperator::advanceForGlyph):
2145         * rendering/svg/SVGTextRunRenderingContext.cpp:
2146         (WebCore::missingGlyphForFont):
2147         * svg/SVGFontData.cpp:
2148         (WebCore::SVGFontData::initializeFont):
2149
2150 2015-03-13  Eric Carlson  <eric.carlson@apple.com>
2151
2152         [Mac] Enable WIRELESS_PLAYBACK_TARGET
2153         https://bugs.webkit.org/show_bug.cgi?id=142635
2154
2155         Reviewed by Darin Adler.
2156
2157         * Configurations/FeatureDefines.xcconfig:
2158
2159 2015-03-13  Jeremy Jones  <jeremyj@apple.com>
2160
2161         Fix typo restoreUserInterfaceForOptimizedFullscreeStopWithCompletionHandler.
2162         https://bugs.webkit.org/show_bug.cgi?id=142678
2163
2164         Reviewed by Eric Carlson.
2165
2166         Add the missing 'n' in "fullscreen".
2167
2168         * platform/spi/ios/AVKitSPI.h:
2169
2170 2015-03-13  Timothy Horton  <timothy_horton@apple.com>
2171
2172         Sites that use a device-width viewport but don't have enough height to fill the view are scaled up
2173         https://bugs.webkit.org/show_bug.cgi?id=142664
2174         <rdar://problem/18859470>
2175
2176         Reviewed by Benjamin Poulain.
2177
2178         * page/ViewportConfiguration.cpp:
2179         (WebCore::ViewportConfiguration::shouldIgnoreHorizontalScalingConstraints):
2180         (WebCore::ViewportConfiguration::shouldIgnoreVerticalScalingConstraints):
2181         (WebCore::ViewportConfiguration::shouldIgnoreScalingConstraints):
2182         Split shouldIgnoreScalingConstraints into one for each dimension.
2183
2184         (WebCore::ViewportConfiguration::initialScale):
2185         (WebCore::ViewportConfiguration::minimumScale):
2186         Don't force the initial and minimum scales to cover the whole view if the
2187         page claims to want to lay out to device width but then lays out too big.
2188         This will allow pages that misbehave in this way to scale down further
2189         than they previously could, but will result in a region of empty background
2190         color being exposed at the initial/minimum scale.
2191
2192         (WebCore::ViewportConfiguration::description):
2193         Update the logging to show each dimension separately.
2194
2195         * page/ViewportConfiguration.h:
2196
2197 2015-03-13  Mark Lam  <mark.lam@apple.com>
2198
2199         Replace TCSpinLock with a new WTF::SpinLock based on WTF::Atomic.
2200         <https://webkit.org/b/142674>
2201
2202         Reviewed by Filip Pizlo.
2203
2204         No new tests because there is no new functionality. This is a refactoring effort.
2205
2206         * bindings/objc/WebScriptObject.mm:
2207         * platform/ios/wak/WAKWindow.mm:
2208         (-[WAKWindow initWithLayer:]):
2209         (-[WAKWindow initWithFrame:]):
2210
2211 2015-03-13  Doug Russell  <d_russell@apple.com>
2212
2213         AX: Provide API for assistive tech to ignore DOM key event handlers
2214         https://bugs.webkit.org/show_bug.cgi?id=142059
2215
2216         Reviewed by Beth Dakin.
2217
2218         Assistive technology applications on the desktop are heavily dependent on keyboard navigation being reliable. This is greatly hindered by sites that handle key events without updating keyboard selection and then consume the event. It is important for assistive technology apps to allow users to decide to ignore these handlers that are incorrect for their purposes.
2219
2220         This can be fixed by exposing, via a new accessibility attribute, a way to decide, for a given WebCore::Frame, to pre-empt DOM dispatch and instead let accessibility caret browsing take place.
2221
2222         Test: platform/mac/accessibility/prevent-keyboard-event-dispatch.html
2223
2224         * accessibility/AccessibilityObject.cpp:
2225         (WebCore::AccessibilityObject::preventKeyboardDOMEventDispatch):
2226         (WebCore::AccessibilityObject::setPreventKeyboardDOMEventDispatch):
2227         * accessibility/AccessibilityObject.h:
2228         * accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
2229         (-[WebAccessibilityObjectWrapper accessibilityAttributeNames]):
2230         (-[WebAccessibilityObjectWrapper accessibilityAttributeValue:]):
2231         (-[WebAccessibilityObjectWrapper accessibilityIsAttributeSettable:]):
2232         (-[WebAccessibilityObjectWrapper _accessibilitySetValue:forAttribute:]):
2233         * dom/Element.cpp:
2234         (WebCore::Element::dispatchKeyEvent):
2235         * page/EventHandler.cpp:
2236         (WebCore::EventHandler::keyEvent):
2237         (WebCore::handleKeyboardSelectionMovement):
2238         (WebCore::EventHandler::handleKeyboardSelectionMovementForAccessibility):
2239         * page/EventHandler.h:
2240         * page/Settings.in:
2241
2242 2015-03-09  Conrad Shultz  <conrad_shultz@apple.com>
2243
2244         Allow clients to selectively disable plug-ins
2245         https://bugs.webkit.org/show_bug.cgi?id=142506
2246
2247         Reviewed by Anders Carlsson.
2248
2249         Add new functionality allow clients to declaratively disable individual plug-ins (in a manner that conceals them
2250         from the page). As part of this:
2251
2252         1) Introduce the concept of web-visible plug-ins and related concepts, as distinct from the real underlying
2253            set of plug-ins.
2254
2255         2) Where applicable, plumb additional information about plug-ins (specifically, bundle identification) deeper
2256            down the stack.
2257
2258         3) Add generic functionality to PluginStrategy to support filtering plug-ins for visibility and introduce
2259            a concrete implementation thereof in WebPlatformStrategies in WebKit2.
2260
2261         4) Add messaging infrastructure to allow clients to set and clear plug-in policies.
2262
2263         While currently only used in a very limited manner, the new declarative plug-in policies are written generically
2264         so that they could be easily used in the future to reduce synchronous messaging to the client when loading plug-ins.
2265
2266         * dom/DOMImplementation.cpp:
2267         (WebCore::DOMImplementation::createDocument):
2268         Update to reflect function rename.
2269
2270         * loader/SubframeLoader.cpp:
2271         (WebCore::findPluginMIMETypeFromURL):
2272         Adopt getWebVisibleMimesAndPluginIndices().
2273         (WebCore::logPluginRequest):
2274         Update to reflect function rename.
2275         (WebCore::SubframeLoader::shouldUsePlugin):
2276         Ditto.
2277
2278         * platform/PlatformStrategies.h:
2279         Export platformStrategies(), since it is now used in WebProcess.cpp.
2280
2281         * plugins/DOMMimeType.cpp:
2282         (WebCore::DOMMimeType::type):
2283         (WebCore::DOMMimeType::description):
2284         (WebCore::DOMMimeType::mimeClassInfo):
2285         Adopt getWebVisibleMimesAndPluginIndices().
2286         (WebCore::DOMMimeType::enabledPlugin):
2287         Ditto.
2288
2289         * plugins/DOMMimeType.h:
2290         Don't return references in a few places where it is no longer safe to do so.
2291         (WebCore::DOMMimeType::mimeClassInfo): Deleted.
2292
2293         * plugins/DOMMimeTypeArray.cpp:
2294         (WebCore::DOMMimeTypeArray::length):
2295         Adopt getWebVisibleMimesAndPluginIndices().
2296         (WebCore::DOMMimeTypeArray::item):
2297         Ditto.
2298         (WebCore::DOMMimeTypeArray::canGetItemsForName):
2299         Ditto.
2300         (WebCore::DOMMimeTypeArray::namedItem):
2301         Ditto.
2302
2303         * plugins/DOMPlugin.cpp:
2304         (WebCore::DOMPlugin::pluginInfo):
2305         Adopt getWebVisiblePlugins().
2306         (WebCore::DOMPlugin::item):
2307         Adopt getWebVisibleMimesAndPluginIndices().
2308         (WebCore::DOMPlugin::canGetItemsForName):
2309         Ditto.
2310         (WebCore::DOMPlugin::namedItem):
2311         Ditto.
2312
2313         * plugins/DOMPlugin.h:
2314         (WebCore::DOMPlugin::pluginInfo): Deleted.
2315
2316         * plugins/DOMPluginArray.cpp:
2317         (WebCore::DOMPluginArray::length):
2318         Adopt getWebVisiblePlugins().
2319         (WebCore::DOMPluginArray::item):
2320         Ditto.
2321         (WebCore::DOMPluginArray::canGetItemsForName):
2322         Ditto.
2323         (WebCore::DOMPluginArray::namedItem):
2324         Ditto.
2325
2326         * plugins/PluginData.cpp:
2327         (WebCore::PluginData::PluginData):
2328         Stash the passed-in Page and call initPlugins().
2329         (WebCore::PluginData::getWebVisiblePlugins):
2330         New member function; call through to PluginStrategy::getWebVisiblePluginInfo().
2331         (WebCore::PluginData::getWebVisibleMimesAndPluginIndices):
2332         New member function; build up the mimes and mimePluginIndices vectors in the same manner as before, but
2333         limited to the web-visible plug-ins.
2334         (WebCore::PluginData::supportsWebVisibleMimeType):
2335         Renamed from supportsMimeType(); update to work in terms of web-visible plug-ins.
2336         (WebCore::PluginData::pluginInfoForWebVisibleMimeType):
2337         Renamed from pluginInfoForMimeType(); ditto.
2338         (WebCore::PluginData::pluginNameForWebVisibleMimeType):
2339         Renamed from pluginNameForMimeType(); ditto.
2340         (WebCore::PluginData::pluginFileForWebVisibleMimeType):
2341         Renamed from pluginFileForMimeType(); ditto.
2342         (WebCore::PluginData::initPlugins):
2343         (WebCore::PluginData::supportsMimeType): Deleted.
2344         (WebCore::PluginData::pluginInfoForMimeType): Deleted.
2345         (WebCore::PluginData::pluginNameForMimeType): Deleted.
2346         (WebCore::PluginData::pluginFileForMimeType): Deleted.
2347
2348         * plugins/PluginData.h:
2349         Add a member variable for the associate Page; declare the PluginLoadClientPolicy enumeration; add
2350         new members to PluginInfo for the clientLoadPolicy and bundle information.
2351         (WebCore::PluginData::PluginData):
2352         Replace some member functions with new ones that will hide plug-ins upon request from the client;
2353         (WebCore::PluginData::mimes): Deleted.
2354         (WebCore::PluginData::mimePluginIndices): Deleted.
2355
2356         * plugins/PluginStrategy.h:
2357         Declare new member functions for retrieving web-visible plug-ins and setting/clearing plug-in policies.
2358
2359         * replay/SerializationMethods.cpp:
2360         (JSC::EncodingTraits<PluginData>::encodeValue):
2361         Remove now-obsolete code for handling MIME types and add a FIXME.
2362         (JSC::DeserializedPluginData::DeserializedPluginData):
2363         (JSC::EncodingTraits<PluginData>::decodeValue):
2364         Ditto.
2365         (JSC::EncodingTraits<PluginInfo>::encodeValue):
2366         Handle the new members in PluginInfo.
2367         (JSC::EncodingTraits<PluginInfo>::decodeValue):
2368         Ditto.
2369
2370         * replay/WebInputs.json:
2371         Teach Replay about PluginLoadClientPolicy.
2372
2373 2015-03-13  Chris Dumez  <cdumez@apple.com>
2374
2375         XMLHttpRequests should not prevent a page from entering PageCache
2376         https://bugs.webkit.org/show_bug.cgi?id=142612
2377         <rdar://problem/19923085>
2378
2379         Reviewed by Alexey Proskuryakov.
2380
2381         Make XMLHttpRequest ActiveDOMObjects suspendable in most cases to
2382         drastically improve the likelihood of pages using them to enter
2383         PageCache. XMLHttpRequest used to be only suspendable when not
2384         loading. After this patch, if the XMLHttpRequest is loading when
2385         navigating away from the page, it will be aborted and the page
2386         will enter the PageCache. Upon restoring the page from PageCache,
2387         the XMLHttpRequests' error handlers will be executed to give them
2388         a chance to reload if they want to.
2389
2390         Test: http/tests/navigation/page-cache-xhr.html
2391
2392         * history/PageCache.cpp:
2393         (WebCore::logCanCacheFrameDecision):
2394         (WebCore::PageCache::canCachePageContainingThisFrame):
2395         Do not prevent a page to enter the page cache ff the main document has
2396         an error that is a cancellation and all remaining subresource loaders
2397         are for XHR. We extend the pre-existing mechanism used on iOS, which
2398         allowed PageCaching if the remaining resource loads are for images.
2399
2400         * loader/DocumentLoader.cpp:
2401         (WebCore::areAllLoadersPageCacheAcceptable):
2402         Mark XHR loaders as PageCache acceptable.
2403
2404         * loader/DocumentThreadableLoader.cpp:
2405         (WebCore::DocumentThreadableLoader::isXMLHttpRequest):
2406         * loader/DocumentThreadableLoader.h:
2407         * loader/ThreadableLoader.h:
2408         * loader/cache/CachedResource.cpp:
2409         (WebCore::CachedResource::areAllClientsXMLHttpRequests):
2410         * loader/cache/CachedResource.h:
2411         * loader/cache/CachedResourceClient.h:
2412         (WebCore::CachedResourceClient::isXMLHttpRequest):
2413         * xml/XMLHttpRequest.cpp:
2414         (WebCore::XMLHttpRequest::XMLHttpRequest):
2415         (WebCore::XMLHttpRequest::createRequest):
2416         (WebCore::XMLHttpRequest::canSuspend):
2417         Report that we can suspend XMLHttpRequests as long as the window load
2418         event has already fired. If the window load event has not fired yet,
2419         it would be unsafe to cancel the load in suspend() as it would
2420         potentially cause arbitrary JS execution while suspending.
2421
2422         (WebCore::XMLHttpRequest::suspend):
2423         If suspending for PageCache and the request is currently loading, abort
2424         the load and mark that we should fire the error event upon restoring
2425         from PageCache.
2426
2427         (WebCore::XMLHttpRequest::resume):
2428         (WebCore::XMLHttpRequest::resumeTimerFired):
2429         Upon resuming, fire the error event in a timer if the load was aborted
2430         for suspending. We need to do this in a timer because we are not allowed
2431         to execute arbitrary JS inside resume().
2432
2433         (WebCore::XMLHttpRequest::stop):
2434         Add a assertion to make sure we are not firing event inside stop() as
2435         this would potentially cause arbitrary JS execution and it would be
2436         unsafe. It seems to me that our code is currently unsafe but the
2437         assertion does not seem to be hit by our current layout tests. I am
2438         adding the assertion as it would make it clear we have a bug and we
2439         need to fix it.
2440
2441         * xml/XMLHttpRequest.h:
2442
2443 2015-03-13  Joonghun Park  <jh718.park@samsung.com>
2444
2445         Fix Debug build error 'comparison is always true due to limited range of data type [-Werror=type-limits]'
2446         https://bugs.webkit.org/show_bug.cgi?id=142652
2447
2448         Reviewed by Csaba Osztrogonác.
2449
2450         No new tests, no behavior changes.
2451
2452         Now CSSPropertyID type is uint16_t, so propertyID >= 0 check is needed no more.
2453
2454         * css/CSSPrimitiveValue.cpp:
2455         (WebCore::propertyName):
2456         * css/makeprop.pl:
2457
2458 2015-03-12  Zan Dobersek  <zdobersek@igalia.com>
2459
2460         Remove DrawingBuffer
2461         https://bugs.webkit.org/show_bug.cgi?id=142641
2462
2463         Reviewed by Darin Adler.
2464
2465         Remove the DrawingBuffer class. Objects of this type were only held in the
2466         WebGLRenderingContext (later renamed to WebGLRenderingContextBase) on the
2467         Chromium port, with the relevant code removed in r147888. Since then, the
2468         m_drawingBuffer member variable has always been null.
2469
2470         * CMakeLists.txt:
2471         * PlatformEfl.cmake:
2472         * PlatformGTK.cmake:
2473         * WebCore.vcxproj/WebCore.vcxproj:
2474         * WebCore.vcxproj/WebCore.vcxproj.filters:
2475         * WebCore.xcodeproj/project.pbxproj:
2476         * html/canvas/WebGL2RenderingContext.cpp:
2477         (WebCore::WebGL2RenderingContext::copyTexImage2D):
2478         * html/canvas/WebGLRenderingContext.cpp:
2479         (WebCore::WebGLRenderingContext::copyTexImage2D):
2480         * html/canvas/WebGLRenderingContextBase.cpp:
2481         (WebCore::WebGLRenderingContextBase::WebGLRenderingContextBase):
2482         (WebCore::WebGLRenderingContextBase::initializeNewContext):
2483         (WebCore::WebGLRenderingContextBase::destroyGraphicsContext3D):
2484         (WebCore::WebGLRenderingContextBase::markContextChanged):
2485         (WebCore::WebGLRenderingContextBase::clearIfComposited):
2486         (WebCore::WebGLRenderingContextBase::paintRenderingResultsToCanvas):
2487         (WebCore::WebGLRenderingContextBase::paintRenderingResultsToImageData):
2488         (WebCore::WebGLRenderingContextBase::reshape):
2489         (WebCore::WebGLRenderingContextBase::drawingBufferWidth):
2490         (WebCore::WebGLRenderingContextBase::drawingBufferHeight):
2491         (WebCore::WebGLRenderingContextBase::activeTexture):
2492         (WebCore::WebGLRenderingContextBase::bindFramebuffer):
2493         (WebCore::WebGLRenderingContextBase::bindTexture):
2494         (WebCore::WebGLRenderingContextBase::copyTexSubImage2D):
2495         (WebCore::WebGLRenderingContextBase::deleteFramebuffer):
2496         (WebCore::WebGLRenderingContextBase::disable):
2497         (WebCore::WebGLRenderingContextBase::enable):
2498         (WebCore::WebGLRenderingContextBase::getContextAttributes):
2499         (WebCore::WebGLRenderingContextBase::readPixels):
2500         (WebCore::WebGLRenderingContextBase::loseContextImpl):
2501         (WebCore::WebGLRenderingContextBase::getBoundFramebufferWidth):
2502         (WebCore::WebGLRenderingContextBase::getBoundFramebufferHeight):
2503         (WebCore::WebGLRenderingContextBase::maybeRestoreContext):
2504         * html/canvas/WebGLRenderingContextBase.h:
2505         (WebCore::ScopedDrawingBufferBinder::ScopedDrawingBufferBinder): Deleted.
2506         (WebCore::ScopedDrawingBufferBinder::~ScopedDrawingBufferBinder): Deleted.
2507         * platform/graphics/GraphicsContext.h:
2508         * platform/graphics/GraphicsContext3D.h:
2509         * platform/graphics/cairo/DrawingBufferCairo.cpp: Removed.
2510         * platform/graphics/gpu/DrawingBuffer.cpp: Removed.
2511         * platform/graphics/gpu/DrawingBuffer.h: Removed.
2512         * platform/graphics/gpu/mac/DrawingBufferMac.mm: Removed.
2513         * platform/graphics/opengl/GraphicsContext3DOpenGLCommon.cpp:
2514         (WebCore::GraphicsContext3D::paintRenderingResultsToCanvas):
2515         (WebCore::GraphicsContext3D::paintRenderingResultsToImageData):
2516
2517 2015-03-12  Ryosuke Niwa  <rniwa@webkit.org>
2518
2519         REGRESSION(r180726): Removing an empty line at the end of textarea clears the entire texture
2520         https://bugs.webkit.org/show_bug.cgi?id=142646
2521
2522         Reviewed by Darin Adler.
2523
2524         The bug was caused by TypingCommand::deleteKeyPressed erroneously determining the editable root to be empty because
2525         Position::atStartOfTree returns true when it's anchored at a BR that is immediately below the root editable element.
2526
2527         Fixed the bug by replacing the use of the deprecated atFirstEditingPositionForNode by a code that understands modern
2528         position types such as PositionIsBeforeAnchor in atStartOfTree and atEndOfTree. These two functions will no longer
2529         return true when anchored before or after BR after this patch.
2530
2531         Test: editing/deleting/delete-empty-line-breaks-at-end-of-textarea.html
2532
2533         * dom/Position.cpp:
2534         (WebCore::Position::atStartOfTree):
2535         (WebCore::Position::atEndOfTree):
2536
2537 2015-03-12  Yusuke Suzuki  <utatane.tea@gmail.com>
2538
2539         Integrate MapData into JSMap and JSSet
2540         https://bugs.webkit.org/show_bug.cgi?id=142556
2541
2542         Reviewed by Filip Pizlo.
2543
2544         Now Set has SetData and it's different from MapData.
2545         And MapData/SetData are completely integrated into JSSet and JSMap.
2546         Structured-cloning algorithm need to be aware of these changes.
2547         And in the case of JSSet, since JSSet doesn't need dummy value for construction,
2548         Structured-cloning only serialize the keys in JSSet.
2549
2550         * ForwardingHeaders/runtime/MapDataInlines.h: Added.
2551         * bindings/js/SerializedScriptValue.cpp:
2552         (WebCore::CloneSerializer::serialize):
2553         (WebCore::CloneDeserializer::consumeCollectionDataTerminationIfPossible):
2554         (WebCore::CloneDeserializer::deserialize):
2555         (WebCore::CloneDeserializer::consumeMapDataTerminationIfPossible): Deleted.
2556
2557 2015-03-12  Dan Bernstein  <mitz@apple.com>
2558
2559         Finish up <rdar://problem/20086546> [Cocoa] Add an option to treat certificate chains with SHA1-signed certificates as insecure
2560
2561         Added back OS X bits that I couldn’t land initially in r181317 or had to remove in r181327.
2562
2563         * platform/network/mac/CertificateInfoMac.mm:
2564         (WebCore::CertificateInfo::containsNonRootSHA1SignedCertificate):
2565         * platform/spi/cocoa/SecuritySPI.h:
2566
2567 2015-03-12  Geoffrey Garen  <ggaren@apple.com>
2568
2569         REGRESSION: Crash under Heap::reportExtraMemoryAllocatedSlowCase for media element
2570         https://bugs.webkit.org/show_bug.cgi?id=142636
2571
2572         Reviewed by Mark Hahnenberg.
2573
2574         This was a pre-existing bug that I made a lot worse in
2575         <https://trac.webkit.org/changeset/181411>.
2576
2577         * html/HTMLMediaElement.cpp:
2578         (WebCore::HTMLMediaElement::parseAttribute): Compare size before
2579         subtracting rather than subtracting and then comparing to zero. The
2580         latter technique is not valid for unsigned integers, which will happily
2581         underflow into giant numbers.
2582
2583         * Modules/mediasource/SourceBuffer.cpp:
2584         (WebCore::SourceBuffer::reportExtraMemoryAllocated): This code was
2585          technically correct, but I took the opportunity to clean it up a bit.
2586          There's no need to do two checks here, and it smells bad to check for
2587          a negative unsigned integer.
2588
2589 2015-03-12  Sebastian Dröge  <sebastian@centricular.com>
2590
2591         Stop using single-include headers that are only available since GStreamer >= 1.2.
2592
2593         https://bugs.webkit.org/show_bug.cgi?id=142537
2594
2595         Reviewed by Philippe Normand.
2596
2597         * platform/audio/gstreamer/AudioDestinationGStreamer.cpp:
2598         * platform/audio/gstreamer/AudioFileReaderGStreamer.cpp:
2599         * platform/audio/gstreamer/AudioSourceProviderGStreamer.cpp:
2600         * platform/audio/gstreamer/WebKitWebAudioSourceGStreamer.cpp:
2601         * platform/graphics/gstreamer/GStreamerUtilities.cpp:
2602         * platform/graphics/gstreamer/GStreamerUtilities.h:
2603         * platform/graphics/gstreamer/ImageGStreamer.h:
2604         Instead of using single-include headers for the GStreamer libraries,
2605         directly include the headers we need. The single-include headers were
2606         only added in 1.2, and this would be the only reason why we would
2607         depend on 1.2.
2608
2609 2015-03-12  Eric Carlson  <eric.carlson@apple.com>
2610
2611         [Mac] Update AirPlay handling
2612         https://bugs.webkit.org/show_bug.cgi?id=142541
2613
2614         Unreviewed, respond to post-review comments.
2615
2616         * dom/Document.cpp:
2617         (WebCore::Document::didChoosePlaybackTarget):
2618         * page/ChromeClient.h:
2619         * page/Page.cpp:
2620         (WebCore::Page::showPlaybackTargetPicker):
2621         (WebCore::Page::didChoosePlaybackTarget):
2622         (WebCore::Page::configurePlaybackTargetMonitoring):
2623
2624 2015-03-12  Csaba Osztrogonác  <ossy@webkit.org>
2625
2626         Fix the !ENABLE(PICTURE_SIZES) build
2627         https://bugs.webkit.org/show_bug.cgi?id=142617
2628
2629         Reviewed by Darin Adler.
2630
2631         * html/parser/HTMLPreloadScanner.cpp:
2632         (WebCore::TokenPreloadScanner::StartTagScanner::processAttributes):
2633
2634 2015-03-11  Joseph Pecoraro  <pecoraro@apple.com>
2635
2636         Unreviewed follow-up fix to r181426. Initialize TextPosition with zeros in case it gets used.
2637
2638         Address ASSERT in LayoutTests/printing/page-format-data.html.
2639
2640         * dom/InlineStyleSheetOwner.cpp:
2641         (WebCore::InlineStyleSheetOwner::InlineStyleSheetOwner):
2642         In case the TextPosition gets used because it wasn't created by a parser,
2643         zeros are more realistic values then beforeFirst.
2644
2645 2015-03-11  Roger Fong  <roger_fong@apple.com>
2646
2647         Media element time no longer updates while scrubbing following r181279.
2648         https://bugs.webkit.org/show_bug.cgi?id=142606.
2649         <rdar://problem/20131014>
2650
2651         Reviewed by Eric Carlson.
2652
2653         * Modules/mediacontrols/mediaControlsApple.js:
2654         (Controller.prototype.handleWrapperMouseMove):
2655         Update time as a result of a mouse move if we are scrubbing.
2656
2657 2015-03-11  Joseph Pecoraro  <pecoraro@apple.com>
2658
2659         Web Inspector: CSS parser errors in the console should include column numbers
2660         https://bugs.webkit.org/show_bug.cgi?id=114313
2661
2662         Reviewed by Darin Adler.
2663
2664         Test: inspector-protocol/console/warnings-errors.html
2665
2666         * css/CSSParser.h:
2667         (WebCore::CSSParser::currentCharacterOffset):
2668         Get the current character offset depending on the source type.
2669         Add instance variables to track column position and start
2670         line / column for inline stylesheets.
2671
2672         * css/CSSParser.cpp:
2673         (WebCore::CSSParser::CSSParser):
2674         (WebCore::CSSParser::parseSheet):
2675         Initialize new instance variables.
2676
2677         (WebCore::CSSParser::currentLocation):
2678         Update to include column information for the token. Also, if we are on the
2679         first line we may need to take into account a start column offset as well.
2680
2681         (WebCore::CSSParser::realLex):
2682         Set the token's start column.
2683         When bumping the line number, reset the column offset for the next
2684         line with the next character.
2685
2686         (WebCore::CSSParser::syntaxError):
2687         (WebCore::CSSParser::logError):
2688         Include column information.
2689
2690         * css/StyleSheetContents.cpp:
2691         (WebCore::StyleSheetContents::parseAuthorStyleSheet):
2692         (WebCore::StyleSheetContents::parseString):
2693         (WebCore::StyleSheetContents::parseStringAtPosition):
2694         Include column information.
2695
2696         * css/StyleSheetContents.h:
2697         * dom/InlineStyleSheetOwner.cpp:
2698         (WebCore::InlineStyleSheetOwner::InlineStyleSheetOwner):
2699         (WebCore::InlineStyleSheetOwner::createSheet):
2700         Save and use column information later on.
2701
2702         * dom/InlineStyleSheetOwner.h:
2703         * inspector/InspectorStyleSheet.cpp:
2704         (WebCore::InspectorStyleSheet::ensureSourceData):
2705         Updated parser signature needs starting column and no longer has optional parameters.
2706
2707 2015-03-11  Eric Carlson  <eric.carlson@apple.com>
2708
2709         [Mac] Update AirPlay handling
2710         https://bugs.webkit.org/show_bug.cgi?id=142541
2711
2712         Reviewed by Sam Weinig.
2713
2714         * WebCore.xcodeproj/project.pbxproj:
2715         * dom/Document.cpp:
2716         (WebCore::Document::showPlaybackTargetPicker):
2717         (WebCore::Document::addPlaybackTargetPickerClient):
2718         (WebCore::Document::removePlaybackTargetPickerClient):
2719         (WebCore::Document::configurePlaybackTargetMonitoring):
2720         (WebCore::Document::requiresPlaybackTargetRouteMonitoring):
2721         (WebCore::Document::playbackTargetAvailabilityDidChange):
2722         (WebCore::Document::didChoosePlaybackTarget):
2723         * dom/Document.h:
2724         * html/HTMLMediaElement.cpp:
2725         (WebCore::HTMLMediaElement::registerWithDocument):
2726         (WebCore::HTMLMediaElement::unregisterWithDocument):
2727         (WebCore::HTMLMediaElement::parseAttribute):
2728         * html/HTMLMediaElement.h:
2729         * html/HTMLMediaSession.cpp:
2730         (WebCore::HTMLMediaSession::HTMLMediaSession):
2731         (WebCore::HTMLMediaSession::registerWithDocument):
2732         (WebCore::HTMLMediaSession::unregisterWithDocument):
2733         (WebCore::HTMLMediaSession::showPlaybackTargetPicker):
2734         (WebCore::HTMLMediaSession::hasWirelessPlaybackTargets):
2735         (WebCore::HTMLMediaSession::setHasPlaybackTargetAvailabilityListeners):
2736         (WebCore::HTMLMediaSession::didChoosePlaybackTarget):
2737         (WebCore::HTMLMediaSession::externalOutputDeviceAvailableDidChange):
2738         (WebCore::HTMLMediaSession::requiresPlaybackTargetRouteMonitoring):
2739         * html/HTMLMediaSession.h:
2740         * page/ChromeClient.h:
2741         * page/Page.cpp:
2742         (WebCore::Page::didChoosePlaybackTarget):
2743         (WebCore::Page::playbackTargetAvailabilityDidChange):
2744         (WebCore::Page::configurePlaybackTargetMonitoring):
2745         * page/Page.h:
2746         (WebCore::Page::hasWirelessPlaybackTarget):
2747         (WebCore::Page::playbackTarget):
2748         * platform/audio/MediaSession.cpp:
2749         (WebCore::MediaSession::clientDataBufferingTimerFired):
2750         (WebCore::MediaSession::wirelessRoutesAvailableDidChange): Deleted.
2751         * platform/audio/MediaSession.h:
2752         (WebCore::MediaSession::didChoosePlaybackTarget):
2753         (WebCore::MediaSession::externalOutputDeviceAvailableDidChange):
2754         (WebCore::MediaSession::requiresPlaybackTargetRouteMonitoring):
2755         (WebCore::MediaSessionClient::setWirelessPlaybackTarget):
2756         * platform/audio/MediaSessionManager.cpp:
2757         (WebCore::MediaSessionManager::wirelessRoutesAvailableChanged): Deleted.
2758         * platform/audio/MediaSessionManager.h:
2759         (WebCore::MediaSessionManager::configureWireLessTargetMonitoring):
2760         * platform/audio/ios/MediaSessionManagerIOS.h:
2761         * platform/audio/ios/MediaSessionManagerIOS.mm:
2762         (WebCore::MediaSessionManageriOS::externalOutputDeviceAvailableDidChange):
2763         (-[WebMediaSessionHelper wirelessRoutesAvailableDidChange:]):
2764         * platform/graphics/AVPlaybackTarget.h: Added.
2765         (WebCore::AVPlaybackTarget::~AVPlaybackTarget):
2766         (WebCore::AVPlaybackTarget::AVPlaybackTarget):
2767         (WebCore::AVPlaybackTarget::setDevicePickerContext):
2768         (WebCore::AVPlaybackTarget::devicePickerContext):
2769         * platform/graphics/AVPlaybackTargetPickerClient.h: Added.
2770         (WebCore::AVPlaybackTargetPickerClient::~AVPlaybackTargetPickerClient):
2771         * platform/graphics/MediaPlayer.cpp:
2772         (WebCore::MediaPlayer::setWirelessPlaybackTarget):
2773         * platform/graphics/MediaPlayer.h:
2774         * platform/graphics/MediaPlayerPrivate.h:
2775         (WebCore::MediaPlayerPrivateInterface::setWirelessPlaybackTarget):
2776         * platform/graphics/avfoundation/AVPlaybackTargetMac.mm: Added.
2777         (WebCore::AVPlaybackTarget::encode):
2778         (WebCore::AVPlaybackTarget::decode):
2779         * platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.h:
2780         * platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
2781         (WebCore::MediaPlayerPrivateAVFoundationObjC::cancelLoad):
2782         (WebCore::MediaPlayerPrivateAVFoundationObjC::createAVPlayer):
2783         (WebCore::MediaPlayerPrivateAVFoundationObjC::setVideoFullscreenLayer):
2784         (WebCore::MediaPlayerPrivateAVFoundationObjC::isCurrentPlaybackTargetWireless):
2785         (WebCore::MediaPlayerPrivateAVFoundationObjC::wirelessPlaybackTargetType):
2786         (WebCore::MediaPlayerPrivateAVFoundationObjC::wirelessVideoPlaybackDisabled):
2787         (WebCore::MediaPlayerPrivateAVFoundationObjC::setWirelessVideoPlaybackDisabled):
2788         (WebCore::MediaPlayerPrivateAVFoundationObjC::setWirelessPlaybackTarget):
2789         (WebCore::MediaPlayerPrivateAVFoundationObjC::updateDisableExternalPlayback):
2790         (-[WebCoreAVFMovieObserver observeValueForKeyPath:ofObject:change:context:]):
2791
2792 2015-03-11  Alex Christensen  <achristensen@webkit.org>
2793
2794         [Content Extensions] Add resource type and load type triggers.
2795         https://bugs.webkit.org/show_bug.cgi?id=142422
2796
2797         Reviewed by Benjamin Poulain.
2798
2799         New API tests.
2800
2801         * WebCore.xcodeproj/project.pbxproj:
2802         * contentextensions/ContentExtensionCompiler.cpp:
2803         (WebCore::ContentExtensions::compileRuleList):
2804         * contentextensions/ContentExtensionParser.cpp:
2805         (WebCore::ContentExtensions::getTypeFlags):
2806         (WebCore::ContentExtensions::loadTrigger):
2807         * contentextensions/ContentExtensionRule.h:
2808         * contentextensions/ContentExtensionsBackend.cpp:
2809         (WebCore::ContentExtensions::ContentExtensionsBackend::actionsForResourceLoad):
2810         (WebCore::ContentExtensions::ContentExtensionsBackend::actionsForURL): Deleted.
2811         * contentextensions/ContentExtensionsBackend.h:
2812         * contentextensions/DFABytecode.h:
2813         (WebCore::ContentExtensions::instructionSizeWithArguments):
2814         * contentextensions/DFABytecodeCompiler.cpp:
2815         (WebCore::ContentExtensions::DFABytecodeCompiler::emitAppendConditionalAction):
2816         (WebCore::ContentExtensions::DFABytecodeCompiler::compileNode):
2817         * contentextensions/DFABytecodeCompiler.h:
2818         * contentextensions/DFABytecodeInterpreter.cpp:
2819         (WebCore::ContentExtensions::DFABytecodeInterpreter::interpret):
2820         * contentextensions/DFABytecodeInterpreter.h:
2821         * loader/ResourceLoadInfo.cpp: Added.
2822         (WebCore::toResourceType):
2823         (WebCore::readResourceType):
2824         (WebCore::readLoadType):
2825         (WebCore::ResourceLoadInfo::isThirdParty):
2826         (WebCore::ResourceLoadInfo::getResourceFlags):
2827         * loader/ResourceLoadInfo.h: Added.
2828         * loader/cache/CachedResourceLoader.cpp:
2829         (WebCore::CachedResourceLoader::requestResource):
2830         * page/UserContentController.cpp:
2831         (WebCore::UserContentController::actionsForResourceLoad):
2832         (WebCore::UserContentController::actionsForURL): Deleted.
2833         * page/UserContentController.h:
2834
2835 2015-03-11  Tim Horton  <timothy_horton@apple.com>
2836
2837         Fix the build.
2838
2839         * page/EventHandler.cpp:
2840         (WebCore::EventHandler::selectClosestWordFromHitTestResult):
2841         (WebCore::EventHandler::selectClosestWordOrLinkFromMouseEvent):
2842         (WebCore::EventHandler::handleMousePressEventTripleClick):
2843         (WebCore::EventHandler::handleMousePressEventSingleClick):
2844
2845 2015-03-11  Timothy Horton  <timothy_horton@apple.com>
2846
2847         <attachment> shouldn't use "user-select: all"
2848         https://bugs.webkit.org/show_bug.cgi?id=142453
2849
2850         Reviewed by Darin Adler.
2851
2852         It turns out that "user-select: all" is rife with bugs; in lieu of fixing them
2853         all (at least for now), let's not use "user-select: all" in the default stylesheet
2854         for <attachment>. It's really overkill anyway, since <attachment> can't have children.
2855         The only "user-select: all" behavior we actually want is select-on-click.
2856         So, we'll implement that in a slightly different way.
2857
2858         Tests: fast/attachment/attachment-select-on-click-inside-user-select-all.html
2859                fast/attachment/attachment-select-on-click.html
2860
2861         * css/html.css:
2862         (attachment):
2863         No more "user-select: all".
2864
2865         (attachment:focus): Deleted.
2866         We stopped using attachment focus a while back and forgot to remove this.
2867
2868         * dom/Node.h:
2869         (WebCore::Node::shouldSelectOnMouseDown):
2870         Add a virtual function that Node subclasses can override to indicate they
2871         should be selected on mouse down.
2872
2873         * html/HTMLAttachmentElement.h:
2874         Override the aforementioned virtual function; <attachment> should always
2875         be selected on mouse down.
2876
2877         * page/EventHandler.cpp:
2878         (WebCore::nodeToSelectOnMouseDownForNode):
2879         Determine which node should be selected when a mousedown hits the given node.
2880         If there's any "user-select: all", we go with the outermost "user-select: all".
2881         Otherwise, we give the node a chance to say that it wants to be selected itself.
2882
2883         (WebCore::expandSelectionToRespectSelectOnMouseDown):
2884         Rename this function, it's not just about "user-select: all" anymore.
2885         Make use of nodeToSelectOnMouseDownForNode.
2886
2887         (WebCore::EventHandler::selectClosestWordFromHitTestResult):
2888         (WebCore::EventHandler::selectClosestWordOrLinkFromMouseEvent):
2889         (WebCore::EventHandler::handleMousePressEventTripleClick):
2890         (WebCore::EventHandler::handleMousePressEventSingleClick):
2891         (WebCore::expandSelectionToRespectUserSelectAll): Deleted.
2892         Adjust to the new names.
2893
2894 2015-03-11  Geoffrey Garen  <ggaren@apple.com>
2895
2896         Users of Heap::deprecatedReportExtraMemory should switch to reportExtraMemoryAllocated+reportExtraMemoryVisited
2897         https://bugs.webkit.org/show_bug.cgi?id=142595
2898
2899         Reviewed by Andreas Kling.
2900
2901         Fixed this bug for canvas.
2902
2903         * html/HTMLCanvasElement.cpp:
2904         (WebCore::HTMLCanvasElement::memoryCost): Factored out the helper function
2905         required by our IDL generator.
2906
2907         (WebCore::HTMLCanvasElement::createImageBuffer): Use
2908         reportExtraMemoryAllocated.
2909
2910         * html/HTMLCanvasElement.h:
2911
2912         * html/HTMLCanvasElement.idl: Adopt the IDL for reporting cost in the
2913         right way during GC. This will match our reportExtraMemoryAllocated
2914         with a reportExtraMemoryVisited during GC.
2915
2916 2015-03-11  Roger Fong  <roger_fong@apple.com>
2917
2918         A number of minor edits to the media controls on OSX.
2919         https://bugs.webkit.org/show_bug.cgi?id=142551.
2920         <rdar://problem/20114707>
2921
2922         Reviewed by Darin Adler.
2923
2924         This covers a slew of minor edits to the new media controls. They are as follows.
2925         Small vertical placements adjustments to inline control elements.
2926         Make sure buttons have no focus outlines.
2927         Expand height of mute box that triggers the volume panel appearing.
2928         Turn all button colors into an slightly transparent white.
2929         Center status display text in fullscreen mode.
2930         Lower position of captions container in fullscreen mode.
2931         Show the controls on when done loading of the video an status display is hidden.
2932
2933         * Modules/mediacontrols/mediaControlsApple.css:
2934         (audio::-webkit-media-controls-panel):
2935         (audio::-webkit-media-controls-panel button:focus):
2936         (audio::-webkit-media-controls-rewind-button):
2937         (audio::-webkit-media-controls-play-button):
2938         (audio::-webkit-media-controls-play-button.paused):
2939         (audio::-webkit-media-controls-panel .mute-box):
2940         (video::-webkit-media-controls-volume-max-button):
2941         (audio::-webkit-media-controls-panel .volume-box):
2942         (video::-webkit-media-controls-volume-min-button):
2943         (audio::-webkit-media-controls-wireless-playback-picker-button):
2944         (audio::-webkit-media-controls-toggle-closed-captions-button):
2945         (audio::-webkit-media-controls-closed-captions-container li.selected:hover::before):
2946         (audio::-webkit-media-controls-fullscreen-button):
2947         (audio::-webkit-media-controls-fullscreen-button.exit):
2948         (audio::-webkit-media-controls-status-display):
2949         (audio::-webkit-media-controls-time-remaining-display):
2950         (video:-webkit-full-screen::-webkit-media-controls-panel .volume-box):
2951         (video:-webkit-full-screen::-webkit-media-controls-volume-max-button):
2952         (video:-webkit-full-screen::-webkit-media-controls-volume-min-button):
2953         (video:-webkit-full-screen::-webkit-media-controls-play-button):
2954         (video:-webkit-full-screen::-webkit-media-controls-play-button.paused):
2955         (video:-webkit-full-screen::-webkit-media-controls-seek-back-button):
2956         (video:-webkit-full-screen::-webkit-media-controls-seek-forward-button):
2957         (video:-webkit-full-screen::-webkit-media-controls-status-display):
2958         (video:-webkit-full-screen::-webkit-media-controls-closed-captions-container):
2959         (audio::-webkit-media-controls-panel button:active): Deleted.
2960         * Modules/mediacontrols/mediaControlsApple.js:
2961         (Controller.prototype.setStatusHidden):
2962
2963 2015-03-11  Commit Queue  <commit-queue@webkit.org>
2964
2965         Unreviewed, rolling out r179340 and r179344.
2966         https://bugs.webkit.org/show_bug.cgi?id=142598
2967
2968         Caused images to stay alive forever when navigating away from
2969         the page before they finish loading. (Requested by kling on
2970         #webkit).
2971
2972         Reverted changesets:
2973
2974         "CachedImage: ensure clients overrides imageChanged instead of
2975         notifyFinished"
2976         https://bugs.webkit.org/show_bug.cgi?id=140722
2977         http://trac.webkit.org/changeset/179340
2978
2979         "HTMLImageLoader: fix build failure on assert condition after
2980         r179340"
2981         https://bugs.webkit.org/show_bug.cgi?id=140722
2982         http://trac.webkit.org/changeset/179344
2983
2984 2015-03-11  Geoffrey Garen  <ggaren@apple.com>
2985
2986         Many users of Heap::reportExtraMemory* are wrong, causing lots of memory growth
2987         https://bugs.webkit.org/show_bug.cgi?id=142593
2988
2989         Reviewed by Andreas Kling.
2990
2991         Adopt deprecatedReportExtraMemory as a short-term fix for runaway
2992         memory growth in these cases where we have not adopted
2993         reportExtraMemoryVisited.
2994
2995         Long-term, we should use reportExtraMemoryAllocated+reportExtraMemoryVisited.
2996         That's tracked by https://bugs.webkit.org/show_bug.cgi?id=142595.
2997
2998         Using IOSDebug, I can see that the canvas stress test @ http://jsfiddle.net/fvyw4ba0/,
2999         which used to keep > 1000 1MB NonVolatile GPU allocations live, now keeps about 10 live.
3000
3001         * Modules/mediasource/SourceBuffer.cpp:
3002         (WebCore::SourceBuffer::reportExtraMemoryAllocated):
3003         * bindings/js/JSDocumentCustom.cpp:
3004         (WebCore::toJS):
3005         * bindings/js/JSImageDataCustom.cpp:
3006         (WebCore::toJS):
3007         * bindings/js/JSNodeListCustom.cpp:
3008         (WebCore::createWrapper):
3009         * dom/CollectionIndexCache.cpp:
3010         (WebCore::reportExtraMemoryAllocatedForCollectionIndexCache):
3011         * html/HTMLCanvasElement.cpp:
3012         (WebCore::HTMLCanvasElement::createImageBuffer):
3013         * html/HTMLImageLoader.cpp:
3014         (WebCore::HTMLImageLoader::imageChanged):
3015         * html/HTMLMediaElement.cpp:
3016         (WebCore::HTMLMediaElement::parseAttribute):
3017         * xml/XMLHttpRequest.cpp:
3018         (WebCore::XMLHttpRequest::dropProtection):
3019
3020 2015-03-10  Andy Estes  <aestes@apple.com>
3021
3022         REGRESSION (r180985): contentfiltering/block-after-add-data.html crashes with GuardMalloc
3023         https://bugs.webkit.org/show_bug.cgi?id=142526
3024
3025         Reviewed by Darin Adler.
3026
3027         * loader/DocumentLoader.cpp:
3028         (WebCore::DocumentLoader::dataReceived): Don't delete m_contentFilter until after we're done using its replacement data.
3029
3030 2015-03-10  Sam Weinig  <sam@webkit.org>
3031
3032         Allow adding a button in input elements for auto fill related functionality
3033         <rdar://problem/19782066>
3034         https://bugs.webkit.org/show_bug.cgi?id=142564
3035
3036         Reviewed by Anders Carlsson.
3037
3038         Test: fast/forms/input-auto-fill-button.html
3039
3040         - Adds a new button that can be shown in <input> elements - AutoFillButtonElement.
3041         - Makes the spelling of AutoFill consistent throughout WebCore and WebKit (except
3042           where not feasible due to exported API/SPI).
3043
3044         * CMakeLists.txt:
3045         * WebCore.vcxproj/WebCore.vcxproj:
3046         * WebCore.vcxproj/WebCore.vcxproj.filters:
3047         * WebCore.xcodeproj/project.pbxproj:
3048         Add new files.
3049
3050         * css/html.css:
3051         (input::-webkit-auto-fill-button):
3052         (input::-webkit-auto-fill-button:hover):
3053         (input::-webkit-auto-fill-button:active):
3054         Add default style rules for the AutoFill button based on the ones
3055         used for caps lock indicator.
3056
3057         * html/HTMLInputElement.cpp:
3058         (WebCore::HTMLInputElement::HTMLInputElement):
3059         (WebCore::HTMLInputElement::autoFillButtonElement):
3060         (WebCore::HTMLInputElement::reset):
3061         (WebCore::HTMLInputElement::setValueFromRenderer):
3062         (WebCore::HTMLInputElement::setAutoFilled):
3063         (WebCore::HTMLInputElement::setShowAutoFillButton):
3064         (WebCore::HTMLInputElement::setAutofilled): Deleted.
3065         * html/HTMLInputElement.h:
3066         (WebCore::HTMLInputElement::isAutoFilled):
3067         (WebCore::HTMLInputElement::showAutoFillButton):
3068         (WebCore::HTMLInputElement::isAutofilled): Deleted.
3069         Add new bit to represent whether the AutoFill button should be shown
3070         or not. By default it is not shown, and the client must enable it.
3071
3072         * html/InputType.cpp:
3073         (WebCore::InputType::updateAutoFillButton):
3074         * html/InputType.h:
3075         (WebCore::InputType::autoFillButtonElement):
3076         Add new virtual methods for updating and accessing the AutoFill button.
3077
3078         * html/TextFieldInputType.cpp:
3079         (WebCore::TextFieldInputType::createShadowSubtree):
3080         (WebCore::TextFieldInputType::autoFillButtonElement):
3081         (WebCore::TextFieldInputType::destroyShadowSubtree):
3082         (WebCore::TextFieldInputType::updatePlaceholderText):
3083         (WebCore::TextFieldInputType::shouldDrawAutoFillButton):
3084         (WebCore::TextFieldInputType::autoFillButtonElementWasClicked):
3085         (WebCore::TextFieldInputType::createContainer):
3086         (WebCore::TextFieldInputType::createAutoFillButton):
3087         (WebCore::TextFieldInputType::updateAutoFillButton):
3088         * html/TextFieldInputType.h:
3089         Add support for adding the AutoFill to the shadow DOM of textfields. The implementation
3090         is slightly different than for the caps lock indicator, because I didn't want to force
3091         the creation of a container for all <input> elements just in case an AutoFill button was
3092         added. Instead, if an AutoFill button is added, the container is created on the fly and
3093         the existing DOM is updated to move into it. Once a container is created, it is never
3094         removed.
3095
3096         * html/shadow/AutoFillButtonElement.cpp: Added.
3097         (WebCore::AutoFillButtonElement::create):
3098         (WebCore::AutoFillButtonElement::AutoFillButtonElement):
3099         (WebCore::AutoFillButtonElement::defaultEventHandler):
3100         * html/shadow/AutoFillButtonElement.h: Added.
3101         Add div subclass that swallows the click event and forwards it to the ChromeClient.
3102
3103         * page/ChromeClient.h:
3104         Add handleAutoFillButtonClick client function to inform WebKit that the AutoFill
3105         button was clicked.
3106
3107         * testing/Internals.cpp:
3108         (WebCore::Internals::setAutofilled):
3109         (WebCore::Internals::setShowAutoFillButton):
3110         * testing/Internals.h:
3111         * testing/Internals.idl:
3112         Expose a new internals.setShowAutoFillButton() function to allow testing
3113         of the AutoFill button from layout tests.
3114
3115         * accessibility/AccessibilityObject.cpp:
3116         (WebCore::AccessibilityObject::isValueAutofilled):
3117         * css/SelectorCheckerTestFunctions.h:
3118         (WebCore::isAutofilled):
3119         * css/StyleResolver.cpp:
3120         (WebCore::StyleResolver::canShareStyleWithControl):
3121         Update for new spelling of AutoFill.
3122
3123 2015-03-11  Geoffrey Garen  <ggaren@apple.com>
3124
3125         Refactored the JSC::Heap extra cost API for clarity and to make some known bugs more obvious
3126         https://bugs.webkit.org/show_bug.cgi?id=142589
3127
3128         Reviewed by Andreas Kling.
3129
3130         Updated for renames to JSC extra cost APIs.
3131
3132         Added FIXMEs to our 10 use cases that are currently wrong, including
3133         canvas, which is the cause of https://bugs.webkit.org/show_bug.cgi?id=142457.
3134
3135         * Modules/mediasource/SourceBuffer.cpp:
3136         (WebCore::SourceBuffer::appendBufferInternal):
3137         (WebCore::SourceBuffer::sourceBufferPrivateAppendComplete):
3138         (WebCore::SourceBuffer::reportExtraMemoryAllocated):
3139         (WebCore::SourceBuffer::reportExtraMemoryCost): Deleted.
3140         * Modules/mediasource/SourceBuffer.h:
3141         * bindings/js/JSDocumentCustom.cpp:
3142         (WebCore::toJS):
3143         * bindings/js/JSImageDataCustom.cpp:
3144         (WebCore::toJS):
3145         * bindings/js/JSNodeListCustom.cpp:
3146         (WebCore::createWrapper):
3147         * bindings/scripts/CodeGeneratorJS.pm:
3148         (GenerateImplementation):
3149         * dom/CollectionIndexCache.cpp:
3150         (WebCore::reportExtraMemoryAllocatedForCollectionIndexCache):
3151         (WebCore::reportExtraMemoryCostForCollectionIndexCache): Deleted.
3152         * dom/CollectionIndexCache.h:
3153         (WebCore::Iterator>::computeNodeCountUpdatingListCache):
3154         * html/HTMLCanvasElement.cpp:
3155         (WebCore::HTMLCanvasElement::createImageBuffer):
3156         * html/HTMLCollection.h:
3157         (WebCore::CollectionNamedElementCache::didPopulate):
3158         * html/HTMLImageLoader.cpp:
3159         (WebCore::HTMLImageLoader::imageChanged):
3160         * html/HTMLMediaElement.cpp:
3161         (WebCore::HTMLMediaElement::parseAttribute):
3162         * xml/XMLHttpRequest.cpp:
3163         (WebCore::XMLHttpRequest::dropProtection):
3164
3165 2015-03-11  Benjamin Poulain  <bpoulain@apple.com>
3166
3167         Add basic support for BOL and EOL assertions to the URL Filter parser
3168         https://bugs.webkit.org/show_bug.cgi?id=142568
3169
3170         Reviewed by Alex Christensen.
3171
3172         This patch adds heavily restricted support for BOL and EOL to the URL filter parser.
3173
3174         Both assertions must be the first/last term of their pattern. Any advanced combination
3175         results in a parsing error.
3176
3177         The BOL assertion is easy to represent: currently, any pattern starts at the beginning
3178         of a line and the NFA are generated accordingly.
3179
3180         I had two options to represent the EOL assertion:
3181         1) Add a new special transition on EOL.
3182         2) Add a new vector of actions to the states, conditional to the EOL input.
3183
3184         I picked the first option to avoid growing every state by a vector
3185         that would be empty in the vast majority of cases.
3186
3187
3188         On the matching side, the interpreter was modified to support transitions on '\0'.
3189         DFABytecodeInstruction::CheckValue now stops when running on a character after
3190         the end of the string.
3191
3192         DFABytecodeInstruction::Jump gets two fixes: First we now account for the index
3193         to avoid going past the end of the input. Second, stop on '\0' too... the reason
3194         is that the unconditional jump is only used for fallback edges of the DFA, fallback
3195         edge are not supposed to accept '\0'.
3196
3197         * contentextensions/DFA.cpp:
3198         (WebCore::ContentExtensions::printTransitions):
3199         * contentextensions/DFABytecodeInterpreter.cpp:
3200         (WebCore::ContentExtensions::DFABytecodeInterpreter::interpret):
3201         * contentextensions/DFANode.h:
3202         * contentextensions/NFA.cpp:
3203         (WebCore::ContentExtensions::NFA::addTransition):
3204         (WebCore::ContentExtensions::NFA::addEpsilonTransition):
3205         (WebCore::ContentExtensions::printTransitions):
3206         * contentextensions/NFANode.h:
3207         * contentextensions/NFAToDFA.cpp:
3208         (WebCore::ContentExtensions::populateTransitions):
3209         (WebCore::ContentExtensions::NFAToDFA::convert):
3210         * contentextensions/URLFilterParser.cpp:
3211         (WebCore::ContentExtensions::Term::Term):
3212         (WebCore::ContentExtensions::Term::isEndOfLineAssertion):
3213         (WebCore::ContentExtensions::GraphBuilder::assertionBOL):
3214         (WebCore::ContentExtensions::GraphBuilder::assertionEOL):
3215         (WebCore::ContentExtensions::GraphBuilder::sinkFloatingTermIfNecessary):
3216
3217 2015-03-11  Jer Noble  <jer.noble@apple.com>
3218
3219         [Mac] Update fullscreen placeholder UI to use Vibrancy.
3220         https://bugs.webkit.org/show_bug.cgi?id=142586
3221
3222         Reviewed by Eric Carlson.
3223
3224         Update the fullscreen placeholder with a translucent vibrant appearance
3225         using NSVisualEffectView. Since NSVisuaEffectView is only available for
3226         OS X 10.10 and above, wrap the new implementation in a version check and
3227         retain the old implementation.
3228
3229         Drive-by: Update the strings for the placeholder view with new HI guidance
3230         as well.
3231
3232         * English.lproj/Localizable.strings:
3233         * platform/LocalizedStrings.cpp:
3234         (WebCore::clickToExitFullScreenText):
3235         * platform/mac/WebCoreFullScreenPlaceholderView.h:
3236         * platform/mac/WebCoreFullScreenPlaceholderView.mm:
3237         (-[WebCoreFullScreenPlaceholderView setExitWarningVisible:]):
3238
3239 2015-03-11  Timothy Horton  <timothy_horton@apple.com>
3240
3241         Make it possible to zoom on pages that claim to lay out to device size and then fail to do so
3242         https://bugs.webkit.org/show_bug.cgi?id=142549
3243
3244         Reviewed by Simon Fraser.
3245
3246         * page/ViewportConfiguration.cpp:
3247         (WebCore::ViewportConfiguration::ViewportConfiguration):
3248         Rename m_ignoreScalingConstraints to m_canIgnoreScalingConstraints,
3249         because it being true does not guarantee that we will ignore scaling constraints,
3250         but it being false does guarantee that we won't.
3251
3252         (WebCore::ViewportConfiguration::shouldIgnoreScalingConstraints):
3253         Ignore scaling constraints if the page:
3254             a) claimed to want to lay out to device-width and then laid out too wide
3255             b) claimed to want to lay out to device-height and then laid out too tall
3256             c) claimed to want to lay out with initialScale=1 and then laid out too wide
3257
3258         (WebCore::ViewportConfiguration::initialScale):
3259         (WebCore::ViewportConfiguration::minimumScale):
3260         (WebCore::ViewportConfiguration::allowsUserScaling):
3261         Call shouldIgnoreScalingConstraints() instead of looking at the local, so we can
3262         have some more logic here (as above).
3263
3264         (WebCore::ViewportConfiguration::description):
3265         Dump whether we're ignoring scaling constraints.
3266
3267         (WebCore::ViewportConfiguration::dump):
3268         Use WTFLogAlways so that the output goes to various other logging mechanisms
3269         instead of just stderr.
3270
3271         * page/ViewportConfiguration.h:
3272         (WebCore::ViewportConfiguration::setCanIgnoreScalingConstraints):
3273         (WebCore::ViewportConfiguration::setIgnoreScalingConstraints): Deleted.
3274
3275 2015-03-11  Myles C. Maxfield  <mmaxfield@apple.com>
3276
3277         Use out-of-band messaging for RenderBox::firstLineBaseline() and RenderBox::inlineBlockBaseline()
3278         https://bugs.webkit.org/show_bug.cgi?id=142569
3279
3280         Reviewed by David Hyatt.
3281
3282         Currently, RenderBox::firstLineBaseline() and RenderBox::inlineBlockBaseline() return -1 to mean
3283         that its baseline should be skipped. Instead of using this sentinel value, this patch changes the
3284         return type from int to Optional<int>.
3285
3286         No new tests because there is no behavior change.
3287
3288         * rendering/RenderBlock.cpp:
3289         (WebCore::RenderBlock::baselinePosition):
3290         (WebCore::RenderBlock::firstLineBaseline):
3291         (WebCore::RenderBlock::inlineBlockBaseline):
3292         * rendering/RenderBlock.h:
3293         * rendering/RenderBlockFlow.cpp:
3294         (WebCore::RenderBlockFlow::firstLineBaseline):
3295         (WebCore::RenderBlockFlow::inlineBlockBaseline):
3296         * rendering/RenderBlockFlow.h:
3297         * rendering/RenderBox.h:
3298         (WebCore::RenderBox::firstLineBaseline):
3299         (WebCore::RenderBox::inlineBlockBaseline):
3300         * rendering/RenderDeprecatedFlexibleBox.cpp:
3301         (WebCore::RenderDeprecatedFlexibleBox::layoutHorizontalBox):
3302         * rendering/RenderFlexibleBox.cpp:
3303         (WebCore::RenderFlexibleBox::baselinePosition):
3304         (WebCore::RenderFlexibleBox::firstLineBaseline):
3305         (WebCore::RenderFlexibleBox::inlineBlockBaseline):
3306         (WebCore::RenderFlexibleBox::marginBoxAscentForChild):
3307         * rendering/RenderFlexibleBox.h:
3308         * rendering/RenderMenuList.h:
3309         * rendering/RenderTable.cpp:
3310         (WebCore::RenderTable::cellAbove):
3311         (WebCore::RenderTable::cellBelow):
3312         (WebCore::RenderTable::cellBefore):
3313         (WebCore::RenderTable::cellAfter):
3314         (WebCore::RenderTable::firstLineBlock):
3315         (WebCore::RenderTable::baselinePosition):
3316         (WebCore::RenderTable::inlineBlockBaseline):
3317         (WebCore::RenderTable::firstLineBaseline):
3318         * rendering/RenderTable.h:
3319         * rendering/RenderTableCell.cpp:
3320         (WebCore::RenderTableCell::cellBaselinePosition):
3321         * rendering/RenderTableSection.cpp:
3322         (WebCore::RenderTableSection::firstLineBaseline):
3323         * rendering/RenderTableSection.h:
3324         * rendering/RenderTextControl.h:
3325         * rendering/mathml/RenderMathMLBlock.cpp:
3326         (WebCore::RenderMathMLBlock::baselinePosition):
3327         (WebCore::RenderMathMLTable::firstLineBaseline):
3328         * rendering/mathml/RenderMathMLBlock.h:
3329         * rendering/mathml/RenderMathMLFraction.cpp:
3330         (WebCore::RenderMathMLFraction::firstLineBaseline):
3331         * rendering/mathml/RenderMathMLFraction.h:
3332         * rendering/mathml/RenderMathMLOperator.cpp:
3333         (WebCore::RenderMathMLOperator::firstLineBaseline):
3334         * rendering/mathml/RenderMathMLOperator.h:
3335         * rendering/mathml/RenderMathMLRoot.cpp:
3336         (WebCore::RenderMathMLRoot::firstLineBaseline):
3337         (WebCore::RenderMathMLRoot::layout):
3338         * rendering/mathml/RenderMathMLRoot.h:
3339         * rendering/mathml/RenderMathMLRow.cpp:
3340         (WebCore::RenderMathMLRow::layout):
3341         * rendering/mathml/RenderMathMLScripts.cpp:
3342         (WebCore::RenderMathMLScripts::layout):
3343         (WebCore::RenderMathMLScripts::firstLineBaseline):
3344         * rendering/mathml/RenderMathMLScripts.h:
3345         * rendering/mathml/RenderMathMLSpace.cpp:
3346         (WebCore::RenderMathMLSpace::firstLineBaseline):
3347         * rendering/mathml/RenderMathMLSpace.h:
3348         * rendering/mathml/RenderMathMLUnderOver.cpp:
3349         (WebCore::RenderMathMLUnderOver::firstLineBaseline):
3350         * rendering/mathml/RenderMathMLUnderOver.h:
3351
3352 2015-03-11  Timothy Horton  <timothy_horton@apple.com>
3353
3354         <attachment>s should be created when dropping files onto contentEditable areas
3355         https://bugs.webkit.org/show_bug.cgi?id=142494
3356         <rdar://problem/19982553>
3357
3358         Reviewed by Anders Carlsson.
3359
3360         Covered by existing tests.
3361
3362         * editing/mac/EditorMac.mm:
3363         (WebCore::Editor::WebContentReader::readFilenames):
3364         Instead of inserting the dropped URLs as strings, make an <attachment>
3365         for each.
3366
3367 2015-03-11  David Hyatt  <hyatt@apple.com>
3368
3369         Optimize offsetWidth and offsetHeight to avoid doing layouts.
3370         https://bugs.webkit.org/show_bug.cgi?id=142544
3371
3372         Reviewed by Beth Dakin.
3373
3374         * dom/Document.cpp:
3375         (WebCore::Document::updateLayoutIfDimensionsOutOfDate):
3376         * dom/Document.h:
3377         Added a new method that only updates layout if it determines that the desired dimensions are out
3378         of date.
3379
3380         * dom/Element.cpp:
3381         (WebCore::Element::offsetWidth):
3382         (WebCore::Element::offsetHeight):
3383         Patch offsetWidth and offsetHeight to call the new method rather than updateLayoutIgnorePendingStylesheets.
3384
3385 2015-03-11  Commit Queue  <commit-queue@webkit.org>
3386
3387         Unreviewed, rolling out r181367.
3388         https://bugs.webkit.org/show_bug.cgi?id=142581
3389
3390         Caused crashes on the debug bots (Requested by cdumez on
3391         #webkit).
3392
3393         Reverted changeset:
3394
3395         "Web Inspector: CSS parser errors in the console should
3396         include column numbers"
3397         https://bugs.webkit.org/show_bug.cgi?id=114313
3398         http://trac.webkit.org/changeset/181367
3399
3400 2015-03-11  Myles C. Maxfield  <mmaxfield@apple.com>
3401
3402         Inline block children do not have correct baselines if their children are also block elements
3403         https://bugs.webkit.org/show_bug.cgi?id=142559
3404
3405         Reviewed by Darin Adler.
3406
3407         Perform the same computation on child block elements as child inline elements.
3408
3409         Test: fast/text/baseline-inline-block-block-children.html
3410
3411         * rendering/RenderBlockFlow.cpp:
3412         (WebCore::RenderBlockFlow::inlineBlockBaseline):
3413
3414 2015-03-11  Carlos Alberto Lopez Perez  <clopez@igalia.com>
3415
3416         [CMake][GStreamer] Building EFL or GTK with ENABLE_VIDEO and without ENABLE_WEB_AUDIO is broken.
3417         https://bugs.webkit.org/show_bug.cgi?id=142577
3418
3419         Reviewed by Carlos Garcia Campos.
3420
3421         No new tests, this is a build fix.
3422
3423         * PlatformEfl.cmake: Include GSTREAMER_AUDIO_LIBRARIES on the link step both for ENABLE_VIDEO and ENABLE_WEB_AUDIO.
3424         * PlatformGTK.cmake: Idem.
3425
3426 2015-03-10  Philippe Normand  <pnormand@igalia.com>
3427
3428         Rename MediaStreamCenter to RealtimeMediaSourceCenter
3429         https://bugs.webkit.org/show_bug.cgi?id=142535
3430
3431         Reviewed by Eric Carlson.
3432
3433         As per bug title, rename MediaStreamCenter to
3434         RealtimeMediaSourceCenter, this class manages
3435         RealtimeMediaSources, not MediaStreams. Some un-needed includes of
3436         the header were also removed.
3437
3438 2015-03-11  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
3439
3440         Remove unnecessary create() factory functions
3441         https://bugs.webkit.org/show_bug.cgi?id=142558
3442
3443         Reviewed by Darin Adler.
3444
3445         create() function which just returns new instance can be replaced with std::make_unique<>.
3446
3447         No new tests, no behavior changes.
3448
3449         * bindings/js/JSCryptoKeySerializationJWK.cpp:
3450         (WebCore::JSCryptoKeySerializationJWK::keyDataOctetSequence):
3451         * crypto/keys/CryptoKeyAES.cpp:
3452         (WebCore::CryptoKeyAES::exportData):
3453         * crypto/keys/CryptoKeyDataOctetSequence.h:
3454         * crypto/keys/CryptoKeyHMAC.cpp:
3455         (WebCore::CryptoKeyHMAC::exportData):
3456         * crypto/keys/CryptoKeySerializationRaw.cpp:
3457         (WebCore::CryptoKeySerializationRaw::keyData):
3458         * platform/audio/mac/CARingBuffer.cpp:
3459         (WebCore::CARingBuffer::create): Deleted.
3460         * platform/audio/mac/CARingBuffer.h:
3461         * platform/graphics/avfoundation/AudioSourceProviderAVFObjC.mm:
3462         (WebCore::AudioSourceProviderAVFObjC::prepare):
3463         * testing/Internals.cpp:
3464         (WebCore::Internals::queueMicroTask):
3465         * testing/MicroTaskTest.cpp:
3466         (WebCore::MicroTaskTest::create): Deleted.
3467         * testing/MicroTaskTest.h:
3468         (WebCore::MicroTaskTest::MicroTaskTest):
3469
3470 2015-03-10  Joseph Pecoraro  <pecoraro@apple.com>
3471
3472         Web Inspector: CSS parser errors in the console should include column numbers
3473         https://bugs.webkit.org/show_bug.cgi?id=114313
3474
3475         Reviewed by Benjamin Poulain.
3476
3477         Test: inspector-protocol/console/warnings-errors.html
3478
3479         * css/CSSParser.h:
3480         (WebCore::CSSParser::currentCharacterOffset):
3481         Get the current character offset depending on the source type.
3482         Add instance variables to track column position and start
3483         line / column for inline stylesheets.
3484
3485         * css/CSSParser.cpp:
3486         (WebCore::CSSParser::CSSParser):
3487         (WebCore::CSSParser::parseSheet):
3488         Initialize new instance variables.
3489
3490         (WebCore::CSSParser::currentLocation):
3491         Update to include column information. Also, if we are on the first line
3492         we may need to take into account a start column offset as well.
3493
3494         (WebCore::CSSParser::realLex):
3495         When bumping the line number, reset the column offset for the next
3496         line with the next character.
3497
3498         (WebCore::CSSParser::syntaxError):
3499         (WebCore::CSSParser::logError):
3500         Include column information.
3501
3502         * css/StyleSheetContents.cpp:
3503         (WebCore::StyleSheetContents::parseAuthorStyleSheet):
3504         (WebCore::StyleSheetContents::parseString):
3505         (WebCore::StyleSheetContents::parseStringAtLineAndColumn):
3506         Include column information.
3507
3508         * css/StyleSheetContents.h:
3509         * dom/InlineStyleSheetOwner.cpp:
3510         (WebCore::InlineStyleSheetOwner::InlineStyleSheetOwner):
3511         (WebCore::InlineStyleSheetOwner::createSheet):
3512         Save and use column information later on.
3513
3514         * dom/InlineStyleSheetOwner.h:
3515         * inspector/InspectorStyleSheet.cpp:
3516         (WebCore::InspectorStyleSheet::ensureSourceData):
3517         Updated parser signature needs starting column and no longer has optional parameters.
3518
3519 2015-03-10  Darin Adler  <darin@apple.com>
3520
3521         Try to fix the GTK build.
3522
3523         * html/HTMLVideoElement.idl: Work around gobject bindings generator limitation
3524         by putting in a LANGUAGE_GOBJECT conditional.
3525
3526 2015-03-10  Darin Adler  <darin@apple.com>
3527
3528         Some event handler fixes
3529         https://bugs.webkit.org/show_bug.cgi?id=142474
3530
3531         Reviewed by Anders Carlsson.
3532
3533         * bindings/scripts/CodeGenerator.pm:
3534         (GenerateConditionalStringFromAttributeValue): Improved the algorithm here to
3535         handle combinations of & and | in conditional expressions; that's coming up when
3536         combining conditionals for includes in NavigatorContentUtils.idl.
3537
3538         * bindings/scripts/CodeGeneratorJS.pm:
3539         (AddToImplIncludes): Removed the rudimentary attempt to split and merge
3540         conditionals involving | here; instead we rely on the rules in the
3541         GenerateConditionalStringFromAttributeValue to handle this.
3542         (GenerateImplementation): Passed new "conditional" argument to JSValueToNative.
3543         (GenerateParametersCheck): Pass "conditional" argument to AddToImplIncludes and
3544         JSValueToNative.
3545         (JSValueToNative): Changed to take "conditional" as an argument, since getting
3546         it from the signature won't work for a method parameter.
3547         (WriteData): Merge duplicates based on the result of the
3548         GenerateConditionalStringFromAttributeValue function rather than on the values
3549         passed into that function, since that function converts conditionals into a
3550         canonical form and can make two strings equal that don't start out that way.
3551
3552         * bindings/scripts/CodeGeneratorObjC.pm:
3553         (SkipAttribute): Added code to guarantee we won't make Objective-C bindings
3554         for event handlers. We will rename EventListener to EventHandler in a
3555         subsequent patch.
3556
3557         * bindings/scripts/test/JS/JSTestInterface.cpp: Updated for a progression caused
3558         by the better logic for conditional includes.
3559
3560         * bindings/scripts/test/JS/JSTestObj.cpp: Updated for change to the test below.
3561         * bindings/scripts/test/TestObj.idl: Added test of an event handler attribute.
3562
3563         * dom/Document.idl: Removed unneeded language #if around event handler
3564         attributes. Removed all the commented out event handler attributes.
3565         Sorted event handler attributes into a single section rather than separating
3566         "standard" from "extensions". Sorted the conditional event handler attributes
3567         into paragraphs at the bottom. We will probably make them all unconditional
3568         in a subsequent patch.
3569
3570         * dom/Element.idl: Removed unneeded language #if around event handler
3571         attributes. Removed all the commented out event handler attributes.
3572         Moved event handler attributes here from all derived element classes to
3573         match the approach from the HTML standard. Also resorted the attributes
3574         as in Document above.
3575
3576         * html/HTMLBodyElement.idl: Removed unneeded language #if around event handler
3577         attributes. Removed all the commented out event handler attributes.
3578         Sorted event handler attributes into a single section rather than separating
3579         "standard" from "overrides".
3580
3581         * html/HTMLElement.cpp:
3582         (WebCore::HTMLElement::populateEventNameForAttributeLocalNameMap):
3583         Added a few missing attribute names, and re-sorted a bit. Moved any from
3584         element subclasses here.
3585
3586         * html/HTMLFormElement.cpp:
3587         (WebCore::HTMLFormElement::parseAttribute): Moved code to handle
3588         oncomplete and oncompleteerror to Element and HTMLElement.
3589         * html/HTMLFormElement.idl: Ditto.
3590
3591         * html/HTMLFrameSetElement.idl: Removed #if and resorted as above.
3592
3593         * html/HTMLInputElement.cpp:
3594         (WebCore::HTMLInputElement::parseAttribute): Moved code to handle
3595         onsearch to HTMLElement.
3596
3597         * html/HTMLMediaElement.cpp:
3598         (WebCore::HTMLMediaElement::parseAttribute): Moved code to handle
3599         all the event handler attributes to HTMLElement.
3600         (WebCore::HTMLMediaElement::updateCaptionContainer): Removed a redundant
3601         if statement I noticed while auditing all calls to JSC::call.
3602         (WebCore::HTMLMediaElement::didAddUserAgentShadowRoot): Added clearException.
3603         I noticed this was missing because a flawed earlier version of my patch was
3604         causing the script to fail, leaving an exception behind that caused an
3605         assertion later. So I audited all calls to JSC::call looking for this mistake.
3606
3607         * html/HTMLMediaElement.idl: Moved all the event handlers from here to
3608         Element. Also changed everything possible to use Conditional instead of #if.
3609
3610         * html/HTMLPlugInImageElement.cpp:
3611         (WebCore::HTMLPlugInImageElement::didAddUserAgentShadowRoot): Added
3612         clearException. Noticed it was missing while auditing all calls to JSC::call.
3613
3614         * html/HTMLVideoElement.cpp:
3615         (WebCore::HTMLVideoElement::parseAttribute): Moved event handler code to
3616         HTMLElement.
3617
3618         * html/HTMLVideoElement.idl: Moved event handler to Element.idl and use
3619         Conditional instead of #if.
3620
3621         * page/DOMWindow.idl: Use Conditional instead of #if and tidy up the event
3622         handlers section.
3623
3624         * svg/svgattrs.in: Remove six unused attribute names. Presumably used in
3625         some older scheme to implement event handlers but no longer used at all.
3626
3627 2015-03-10  Brent Fulgham  <bfulgham@apple.com>
3628
3629         CSS scroll-snap-destination and scroll-snap-coordinate are not honoring position values
3630         https://bugs.webkit.org/show_bug.cgi?id=142411
3631
3632         Reviewed by Simon Fraser.
3633
3634         Tested by css3/scroll-snap/scroll-snap-position-values.html.
3635
3636         Revise the CSSParser to recognize that scroll-snap-coordinates and scroll-snap-destination
3637         may be specified as positions, therefore allowing 'top', 'bottom', and 'center' for the Y axis,
3638         and 'left', 'right', and 'center' for the X axis.
3639
3640         Correct implementation to support calculated values for Scroll Snap Point markup. This required the
3641         Scroll Snap Point-specific LengthRepeat class to change its internal representation from a CSSPrimitiveValue
3642         to a regular CSSValue.
3643
3644         Add tests that these position labels, as well as combinations with percentages and pixel offsets
3645         are parsed properly.
3646
3647         * css/CSSComputedStyleDeclaration.cpp:
3648         (WebCore::scrollSnapDestination): Switch from 'percentageOrZoomAdjustedValue' to 'zoomAdjustedPixelValueForLength'
3649         when working with Length values. This is necessary to allow calculated results to be based on the proper default
3650         page dimensions.
3651         (WebCore::scrollSnapPoints): Ditto.
3652         (WebCore::scrollSnapCoordinates): Ditto.
3653         * css/CSSParser.cpp:
3654         (WebCore::CSSParser::parseScrollSnapPositions): Consolidated code for dealing with snap point
3655         positions.
3656         (WebCore::CSSParser::parseScrollSnapDestination): Revise to call new helper function.
3657         (WebCore::CSSParser::parseScrollSnapCoordinate): Ditto.
3658         (WebCore::CSSParser::parseFillPositionX): Rename as parsePositionX.
3659         (WebCore::CSSParser::parseFillPositionY): Rename as parsePositionY.
3660         (WebCore::CSSParser::parseFillProperty): Update to call renamed parsePosition{X|Y} methods.
3661         (WebCore::CSSParser::parseTransformOrigin): Ditto.
3662         (WebCore::CSSParser::parsePerspectiveOrigin): Ditto.
3663         * css/CSSParser.h:
3664         * css/LengthRepeat.h: Revise class to use a CSSValue, rather than a CSSPrimitiveValue, so that we can represent
3665         repeat values as calculations.
3666
3667 2015-03-10  Enrica Casucci  <enrica@apple.com>
3668
3669         Add support for more emoji with variation.
3670         https://bugs.webkit.org/show_bug.cgi?id=142548
3671         rdar://problem/20105008
3672
3673         Reviewed by Tim Horton.
3674
3675         Update ICU rules to support new emoji with variation.
3676
3677         Test: editing/selection/extend-by-character-007.html
3678
3679         * platform/text/TextBreakIterator.cpp:
3680         (WebCore::cursorMovementIterator):
3681
3682 2015-03-10  Alex Christensen  <achristensen@webkit.org>
3683
3684         Use unsigned for HashSet size.
3685         https://bugs.webkit.org/show_bug.cgi?id=142518
3686
3687         Reviewed by Benjamin Poulain.
3688
3689         * Modules/websockets/WebSocketDeflateFramer.cpp:
3690         (WebCore::WebSocketExtensionDeflateFrame::processResponse):
3691         * bindings/js/SerializedScriptValue.cpp:
3692         (WebCore::CloneSerializer::checkForDuplicate):
3693         (WebCore::CloneSerializer::writeConstantPoolIndex):
3694         * dom/ScriptRunner.cpp:
3695         (WebCore::ScriptRunner::~ScriptRunner):
3696         * loader/ResourceLoadScheduler.h:
3697         * platform/graphics/Font.cpp:
3698         (WebCore::Font::systemFallbackFontForCharacter):
3699         * platform/graphics/FontCache.cpp:
3700         (WebCore::FontCache::purgeInactiveFontDataIfNeeded):
3701         (WebCore::FontCache::purgeInactiveFontData):
3702         * platform/graphics/FontCache.h:
3703         * platform/graphics/freetype/FontCacheFreeType.cpp:
3704         (WebCore::FontCache::systemFallbackForCharacters):
3705         * platform/graphics/ios/FontCacheIOS.mm:
3706         (WebCore::FontCache::getSystemFontFallbackForCharacters):
3707         (WebCore::FontCache::systemFallbackForCharacters):
3708         (WebCore::FontCache::similarFont):
3709         * platform/graphics/mac/FontCacheMac.mm:
3710         (WebCore::shouldAutoActivateFontIfNeeded):
3711         (WebCore::FontCache::systemFallbackForCharacters):
3712         (WebCore::FontCache::similarFont):
3713         * platform/graphics/win/FontCacheWin.cpp:
3714         (WebCore::FontCache::systemFallbackForCharacters):
3715         * rendering/RenderDeprecatedFlexibleBox.cpp:
3716         (WebCore::FlexBoxIterator::next):
3717         * rendering/RenderTableSection.cpp:
3718         (WebCore::RenderTableSection::computeOverflowFromCells):
3719
3720 2015-03-10  Eric Carlson  <eric.carlson@apple.com>
3721
3722         [Mac] Refactor media controls code
3723         https://bugs.webkit.org/show_bug.cgi?id=142455
3724
3725         Reviewed by Dean Jackson.
3726
3727         No new tests, updated tests/media/hls/video-controls-live-stream-expected.txt.
3728
3729         * Modules/mediacontrols/mediaControlsApple.js:
3730         (Controller):
3731         (Controller.prototype.addVideoListeners):
3732         (Controller.prototype.removeVideoListeners):
3733         (Controller.prototype.createControls):
3734         (Controller.prototype.setControlsType):
3735         (Controller.prototype.configureInlineControls):
3736         (Controller.prototype.updateStatusDisplay):
3737         (Controller.prototype.handlePanelMouseDown):
3738         (Controller.prototype.showControls):
3739         (Controller.prototype.hideControls):
3740         (Controller.prototype.handleOptimizedFullscreenButtonClicked):
3741         (Controller.prototype.currentPlaybackTargetIsWireless):
3742         (Controller.prototype.updateShouldListenForPlaybackTargetAvailabilityEvent):
3743         (Controller.prototype.updateWirelessPlaybackStatus):
3744         (Controller.prototype.updateWirelessTargetAvailable):
3745         (Controller.prototype.handleWirelessPickerButtonClicked):
3746         (Controller.prototype.handleWirelessPlaybackChange):
3747         (Controller.prototype.handleWirelessTargetAvailableChange):
3748         (Controller.prototype.setShouldListenForPlaybackTargetAvailabilityEvent):
3749         * Modules/mediacontrols/mediaControlsiOS.js:
3750         (ControllerIOS):
3751         (ControllerIOS.prototype.addVideoListeners):
3752         (ControllerIOS.prototype.removeVideoListeners):
3753         (ControllerIOS.prototype.createControls):
3754         (ControllerIOS.prototype.setControlsType):
3755         (ControllerIOS.prototype.configureInlineControls):
3756         (ControllerIOS.prototype.showControls):
3757         (ControllerIOS.prototype.handleWirelessPickerButtonTouchEnd):
3758         (ControllerIOS.prototype.updateShouldListenForPlaybackTargetAvailabilityEvent):
3759         (ControllerIOS.prototype.updateStatusDisplay):
3760         (ControllerIOS.prototype.setShouldListenForPlaybackTargetAvailabilityEvent):
3761         (ControllerIOS.prototype.currentPlaybackTargetIsWireless): Deleted.
3762         (ControllerIOS.prototype.updateWirelessPlaybackStatus): Deleted.
3763         (ControllerIOS.prototype.updateWirelessTargetAvailable): Deleted.
3764         (ControllerIOS.prototype.hideControls): Deleted.
3765         (ControllerIOS.prototype.handleWirelessPlaybackChange): Deleted.
3766         (ControllerIOS.prototype.handleWirelessTargetAvailableChange): Deleted.
3767
3768 2015-03-10  Said Abou-Hallawa  <sabouhallawa@apple.com>
3769
3770         Remove PassRefPtr from svg/properties classes.
3771         https://bugs.webkit.org/show_bug.cgi?id=142063.
3772
3773         Reviewed by Darin Adler.
3774
3775         Remove PassRefPtr from svg/properties classes. All the return types and
3776         the local variables should be RefPtr. If we are sure the pointer can't
3777         be null, we use Ref instead.
3778
3779         * svg/SVGMarkerElement.cpp:
3780         (WebCore::SVGMarkerElement::lookupOrCreateOrientTypeWrapper):
3781         (WebCore::SVGMarkerElement::orientTypeAnimated):
3782         * svg/SVGMarkerElement.h:
3783         * svg/SVGPathElement.cpp:
3784         (WebCore::SVGPathElement::lookupOrCreateDWrapper):
3785         (WebCore::SVGPathElement::pathSegList):
3786         (WebCore::SVGPathElement::animatedPathSegList):
3787         * svg/SVGPathElement.h:
3788         * svg/SVGPolyElement.cpp:
3789         (WebCore::SVGPolyElement::lookupOrCreatePointsWrapper):
3790         (WebCore::SVGPolyElement::points):
3791         (WebCore::SVGPolyElement::animatedPoints):
3792         * svg/SVGPolyElement.h:
3793         * svg/SVGTextContentElement.cpp:
3794         (WebCore::SVGTextContentElement::lookupOrCreateTextLengthWrapper):
3795         (WebCore::SVGTextContentElement::textLengthAnimated):
3796         * svg/SVGTextContentElement.h:
3797         * svg/SVGViewSpec.cpp:
3798         (WebCore::SVGViewSpec::transform):
3799         (WebCore::SVGViewSpec::viewBoxAnimated):
3800         (WebCore::SVGViewSpec::preserveAspectRatioAnimated):
3801         (WebCore::SVGViewSpec::lookupOrCreateViewBoxWrapper):
3802         (WebCore::SVGViewSpec::lookupOrCreatePreserveAspectRatioWrapper):
3803         (WebCore::SVGViewSpec::lookupOrCreateTransformWrapper):
3804         * svg/SVGViewSpec.h:
3805         * svg/properties/SVGAnimatedEnumerationPropertyTearOff.h:
3806         (WebCore::SVGAnimatedEnumerationPropertyTearOff::create):
3807         * svg/properties/SVGAnimatedListPropertyTearOff.h:
3808         (WebCore::SVGAnimatedListPropertyTearOff::create):
3809         * svg/properties/SVGAnimatedPathSegListPropertyTearOff.h:
3810         (WebCore::SVGAnimatedPathSegListPropertyTearOff::create):
3811         * svg/properties/SVGAnimatedProperty.h:
3812         (WebCore::SVGAnimatedProperty::lookupOrCreateWrapper):
3813         * svg/properties/SVGAnimatedPropertyMacros.h:
3814         * svg/properties/SVGAnimatedPropertyTearOff.h:
3815         (WebCore::SVGAnimatedPropertyTearOff::create):
3816         * svg/properties/SVGAnimatedStaticPropertyTearOff.h:
3817         (WebCore::SVGAnimatedStaticPropertyTearOff::create):
3818         * svg/properties/SVGAnimatedTransformListPropertyTearOff.h:
3819         (WebCore::SVGAnimatedTransformListPropertyTearOff::create):
3820         * svg/properties/SVGListProperty.h:
3821         (WebCore::SVGListProperty::initializeValuesAndWrappers):
3822         (WebCore::SVGListProperty::getItemValuesAndWrappers):
3823         (WebCore::SVGListProperty::insertItemBeforeValuesAndWrappers):
3824         (WebCore::SVGListProperty::replaceItemValuesAndWrappers):
3825         (WebCore::SVGListProperty::removeItemValuesAndWrappers):
3826         (WebCore::SVGListProperty::appendItemValuesAndWrappers):
3827         * svg/properties/SVGListPropertyTearOff.h:
3828         (WebCore::SVGListPropertyTearOff::create):
3829         (WebCore::SVGListPropertyTearOff::initialize):
3830         (WebCore::SVGListPropertyTearOff::getItem):
3831         (WebCore::SVGListPropertyTearOff::insertItemBefore):
3832         (WebCore::SVGListPropertyTearOff::replaceItem):
3833         (WebCore::SVGListPropertyTearOff::removeItem):
3834         (WebCore::SVGListPropertyTearOff::appendItem):
3835         * svg/properties/SVGPathSegListPropertyTearOff.cpp:
3836         (WebCore::SVGPathSegListPropertyTearOff::getItem):
3837         (WebCore::SVGPathSegListPropertyTearOff::replaceItem):
3838         (WebCore::SVGPathSegListPropertyTearOff::removeItem):
3839         * svg/properties/SVGPathSegListPropertyTearOff.h:
3840         (WebCore::SVGPathSegListPropertyTearOff::create):
3841         (WebCore::SVGPathSegListPropertyTearOff::initialize):
3842         (WebCore::SVGPathSegListPropertyTearOff::insertItemBefore):
3843         (WebCore::SVGPathSegListPropertyTearOff::appendItem):
3844         * svg/properties/SVGPropertyInfo.h:
3845         * svg/properties/SVGStaticListPropertyTearOff.h:
3846         (WebCore::SVGStaticListPropertyTearOff::create):
3847         * svg/properties/SVGStaticPropertyTearOff.h:
3848         (WebCore::SVGStaticPropertyTearOff::create):
3849         * svg/properties/SVGTransformListPropertyTearOff.h:
3850         (WebCore::SVGTransformListPropertyTearOff::create):
3851         (WebCore::SVGTransformListPropertyTearOff::createSVGTransformFromMatrix):
3852         (WebCore::SVGTransformListPropertyTearOff::consolidate):
3853
3854 2015-03-10  Timothy Horton  <timothy_horton@apple.com>
3855
3856         Fix the build
3857
3858         * platform/network/cf/ResourceHandleCFNet.cpp:
3859         (WebCore::ResourceHandle::createCFURLConnection):
3860
3861 2015-03-10  Benjamin Poulain  <bpoulain@apple.com>
3862
3863         Extend URL filter's Term definition to support groups/subpatterns
3864         https://bugs.webkit.org/show_bug.cgi?id=142519
3865
3866         Reviewed by Alex Christensen.
3867
3868         Pretty simple extension: Term is extended to support holding
3869         a Vector of Term. The quantifier of the Term applies to its
3870         Vector of term as a whole.
3871
3872         To avoid exposing too much internal in the API of Term, I moved
3873         graph generation from GraphBuilder to Term.
3874
3875         Sinking a CharacterSet works as usual. Sinking a Group is done
3876         by sinking each of its Terms one by one and then apply the quantifier
3877         on the whole subgraph. This is done by recursively calling into
3878         Term::generateGraph().
3879
3880         Since groups could be nested, the groups make a stack with the latest
3881         open group on top.
3882         When sinking a floating Term, it is sunk to the latest open group. If there is no open
3883         group, we use the prefix tree and sink the whole subpattern to the graph.
3884
3885         * contentextensions/URLFilterParser.cpp:
3886         (WebCore::ContentExtensions::Term::Term):
3887         (WebCore::ContentExtensions::Term::extendGroupSubpattern):
3888         (WebCore::ContentExtensions::Term::generateGraph):
3889         (WebCore::ContentExtensions::Term::operator==):
3890         (WebCore::ContentExtensions::Term::hash):
3891         (WebCore::ContentExtensions::Term::isUniversalTransition):
3892         (WebCore::ContentExtensions::Term::generateSubgraphForAtom):
3893         (WebCore::ContentExtensions::Term::destroy):
3894         (WebCore::ContentExtensions::Term::Group::operator==):
3895         (WebCore::ContentExtensions::Term::Group::hash):
3896         (WebCore::ContentExtensions::GraphBuilder::finalize):
3897         (WebCore::ContentExtensions::GraphBuilder::atomParenthesesSubpatternBegin):
3898         (WebCore::ContentExtensions::GraphBuilder::atomParenthesesEnd):
3899         (WebCore::ContentExtensions::GraphBuilder::sinkFloatingTermIfNecessary):
3900         (WebCore::ContentExtensions::Term::quantifier): Deleted.
3901         (WebCore::ContentExtensions::Term::visitSimpleTransitions): Deleted.
3902         (WebCore::ContentExtensions::GraphBuilder::addTransitions): Deleted.
3903         (WebCore::ContentExtensions::GraphBuilder::sinkFloatingTerm): Deleted.
3904
3905 2015-03-10  Roger Fong  <roger_fong@apple.com>
3906
3907         Adjustments to media control fonts.
3908         https://bugs.webkit.org/show_bug.cgi?id=142509.
3909         <rdar://problem/20085228>
3910
3911         Reviewed by Darin Adler.
3912
3913         Use system font for controls. Adjust surrounding divs accordingly.
3914         Apply subpixel antialiasing to all media control text.
3915         * Modules/mediacontrols/mediaControlsApple.css:
3916         (::-webkit-media-controls):
3917         (audio::-webkit-media-controls-closed-captions-container .list):
3918         (audio::-webkit-media-controls-status-display):
3919         (audio::-webkit-media-controls-time-remaining-display):
3920         (audio::-webkit-media-controls-current-time-display):
3921         (audio::-webkit-media-controls-time-remaining-display.five-digit-time):
3922         (audio::-webkit-media-controls-current-time-display.five-digit-time):
3923         (audio::-webkit-media-controls-time-remaining-display.six-digit-time):
3924         (audio::-webkit-media-controls-current-time-display.six-digit-time):
3925
3926 2015-03-10  Dan Bernstein  <mitz@apple.com>
3927
3928         Fixed a crash in CertificateInfo::containsNonRootSHA1SignedCertificate when loading a non-HTTPS webapge.
3929
3930         Reviewed by Anders Carlsson.
3931
3932         * platform/network/mac/CertificateInfoMac.mm:
3933         (WebCore::CertificateInfo::containsNonRootSHA1SignedCertificate): Null-check the certificate chain.
3934
3935 2015-03-10  Dan Bernstein  <mitz@apple.com>
3936
3937         Fixed the build.
3938
3939         * platform/spi/cocoa/SecuritySPI.h:
3940
3941 2015-03-10  Zalan Bujtas  <zalan@apple.com>
3942
3943         Simple line layout: Merge TextFragmentIterator::findNextBreakablePosition() and TextFragmentIterator::findNextNonWhitespacePosition().
3944         https://bugs.webkit.org/show_bug.cgi?id=142344
3945
3946         Reviewed by Antti Koivisto.
3947
3948         This patch merges findNextBreakablePosition() and findNextNonWhitespacePosition() so that
3949         the segment looping and position handling logic are no longer duplicated. It also unifies
3950         the static next*() functions' signature.
3951
3952         No change in functionality.
3953
3954         * rendering/SimpleLineLayoutTextFragmentIterator.cpp:
3955         (WebCore::SimpleLineLayout::TextFragmentIterator::nextTextFragment):
3956         (WebCore::SimpleLineLayout::nextBreakablePosition):
3957         (WebCore::SimpleLineLayout::nextNonWhitespacePosition):
3958         (WebCore::SimpleLineLayout::TextFragmentIterator::skipToNextPosition):
3959         (WebCore::SimpleLineLayout::TextFragmen