Source/WebCore/ChangeLog

   1 2018-11-14  Devin Rousso  <drousso@apple.com>
   2
   3         Web Inspector: Canvas: send a call stack with each action instead of an array of call frames
   4         https://bugs.webkit.org/show_bug.cgi?id=191628
   5
   6         Reviewed by Dean Jackson.
   7
   8         Updated existing test: inspector/model/recording.html
   9
  10         * inspector/InspectorCanvas.h:
  11         * inspector/InspectorCanvas.cpp:
  12         (WebCore::InspectorCanvas::indexForData):
  13         (WebCore::InspectorCanvas::buildInitialState):
  14         (WebCore::InspectorCanvas::buildAction):
  15         Drive-by: prevent de-duplicated objects from being destroyed while recording.
  16
  17 2018-11-14  Stephan Szabo  <stephan.szabo@sony.com>
  18
  19         [Win] Compile Service Worker support
  20         https://bugs.webkit.org/show_bug.cgi?id=191409
  21
  22         Reviewed by Youenn Fablet.
  23
  24         Fix compilation errors when ENABLE(SERVICE_WORKER)
  25         on Windows with clang-cl. Clang on dllexport
  26         platforms does not support specifying the
  27         dllexport on both a class and members of the class
  28         and unistd.h isn't provided but also appeared to
  29         not be used.
  30
  31         No new tests, should be covered by existing tests.
  32
  33         * workers/service/ServiceWorkerProvider.h:
  34         * workers/service/context/SWContextManager.cpp:
  35
  36 2018-11-14  Joseph Pecoraro  <pecoraro@apple.com>
  37
  38         Web Inspector: Keep Web Inspector window alive across process swaps (PSON) (Remote Inspector)
  39         https://bugs.webkit.org/show_bug.cgi?id=191494
  40         <rdar://problem/45469854>
  41
  42         Reviewed by Devin Rousso.
  43
  44         * inspector/InspectorClient.h:
  45         (WebCore::InspectorClient::allowRemoteInspectionToPageDirectly const):
  46         Provide a hook so that a client may wish to allow direct remote inspection of the Page.
  47         This is used by WebKitLegacy only.
  48
  49         * page/Page.cpp:
  50         (Page::Page):
  51         Only enable the PageDebuggable if the client wishes remote inspection of the Page directly.
  52         This is used by WebKitLegacy only.
  53
  54         * inspector/InspectorController.cpp:
  55         (WebCore::InspectorController::connectFrontend):
  56         * inspector/InspectorController.h:
  57         * page/PageDebuggable.cpp:
  58         (WebCore::PageDebuggable::connect):
  59         (WebCore::PageDebuggable::disconnect):
  60         * page/PageDebuggable.h:
  61         When a frontend connects, always enable the developer extras for the Page.
  62         This is pretty much only for the remote path, which allows inspection if developer
  63         extras were not already enabled (iOS). This simplifies the logic, and toggling
  64         developer extras after it was already enabled is not really important.
  65
  66 2018-11-14  Per Arne Vollan  <pvollan@apple.com>
  67
  68         REGRESSION (WEBPROCESS_WINDOWSERVER_BLOCKING): requestAnimationFrame Stops Completing
  69         https://bugs.webkit.org/show_bug.cgi?id=190884
  70
  71         Reviewed by Dean Jackson.
  72
  73         Only notify display refresh monitors with matching display ID.
  74
  75         Test: fast/animation/request-animation-frame-in-two-pages.html
  76
  77         * platform/graphics/DisplayRefreshMonitorManager.cpp:
  78         (WebCore::DisplayRefreshMonitorManager::displayWasUpdated):
  79         * platform/graphics/DisplayRefreshMonitorManager.h:
  80
  81 2018-11-14  Youenn Fablet  <youenn@apple.com>
  82
  83         Convert libwebrtc error types to DOM exceptions
  84         https://bugs.webkit.org/show_bug.cgi?id=191590
  85
  86         Reviewed by Alex Christensen.
  87
  88         Make use of overloaded callback method that provides an error type.
  89         This type is then used to create a DOM exception with the correct type.
  90         Covered by existing tests.
  91
  92         * Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:
  93         (WebCore::LibWebRTCMediaEndpoint::doSetRemoteDescription):
  94         (WebCore::LibWebRTCMediaEndpoint::createSessionDescriptionFailed):
  95         (WebCore::LibWebRTCMediaEndpoint::setLocalSessionDescriptionFailed):
  96         (WebCore::LibWebRTCMediaEndpoint::setRemoteSessionDescriptionFailed):
  97         * Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.h:
  98         * Modules/mediastream/libwebrtc/LibWebRTCObservers.h:
  99         (WebCore::toExceptionCode):
 100
 101 2018-11-14  Youenn Fablet  <youenn@apple.com>
 102
 103         Allow to remove MediaStreamPrivate observers when iterating over observers
 104         https://bugs.webkit.org/show_bug.cgi?id=187256
 105
 106         Reviewed by Eric Carlson.
 107
 108         Migrate the observer list from a Vector to a HashSet.
 109         This is more robust to multiple observing and keeping of order of observers is not required.
 110         Copy the set of observers to a vector before iterating over it.
 111         This allows to remove an observer while iterating, which is now used in UserMediaRequest.
 112
 113         Covered by existing tests.
 114
 115         * Modules/mediastream/UserMediaRequest.cpp:
 116         (WebCore::UserMediaRequest::mediaStreamIsReady):
 117         * platform/mediastream/MediaStreamPrivate.cpp:
 118         (WebCore::MediaStreamPrivate::addObserver):
 119         (WebCore::MediaStreamPrivate::removeObserver):
 120         (WebCore::MediaStreamPrivate::forEachObserver const):
 121         (WebCore::MediaStreamPrivate::updateActiveState):
 122         (WebCore::MediaStreamPrivate::addTrack):
 123         (WebCore::MediaStreamPrivate::removeTrack):
 124         (WebCore::MediaStreamPrivate::characteristicsChanged):
 125         * platform/mediastream/MediaStreamPrivate.h:
 126
 127 2018-11-14  Youenn Fablet  <youenn@apple.com>
 128
 129         Calling removeTrack on different RTCPeerConnection should throw InvalidAccessError
 130         https://bugs.webkit.org/show_bug.cgi?id=191603
 131
 132         Reviewed by Chris Dumez.
 133
 134         Make sure to check that the sender peer connection backend is matching.
 135         Covered by rebased WPT test.
 136
 137         * Modules/mediastream/RTCPeerConnection.cpp:
 138         (WebCore::RTCPeerConnection::removeTrack):
 139         * Modules/mediastream/RTCRtpSender.cpp:
 140         (WebCore::RTCRtpSender::isCreatedBy const):
 141         * Modules/mediastream/RTCRtpSender.h:
 142
 143 2018-11-14  Fujii Hironori  <Hironori.Fujii@sony.com>
 144
 145         [curl] Unify CookieJarCurlDatabase and the abstract class CookieJarCurl
 146         https://bugs.webkit.org/show_bug.cgi?id=191620
 147
 148         Reviewed by Alex Christensen.
 149
 150         Remove a abstract class CookieJarCurl which is not needed anymore.
 151         And, rename CookieJarCurlDatabase to CookieJarCurl.
 152
 153         No new tests because there's no behaviour change in WebCore.
 154
 155         * platform/Curl.cmake: Replaced CookieJarCurlDatabase.cpp with CookieJarCurl.cpp.
 156         * platform/network/curl/CookieJarCurl.cpp: Renamed from Source/WebCore/platform/network/curl/CookieJarCurlDatabase.cpp.
 157         * platform/network/curl/CookieJarCurl.h: Merged CookieJarCurl.h and CookieJarCurlDatabase.h.
 158         * platform/network/curl/CookieJarCurlDatabase.h: Removed.
 159         * platform/network/curl/NetworkStorageSessionCurl.cpp:
 160         (WebCore::NetworkStorageSession::NetworkStorageSession): Replaced CookieJarCurlDatabase with CookieJarCurl.
 161
 162 2018-11-14  Christopher Reid  <chris.reid@sony.com>
 163
 164         [WPE] Remove glib usage in PlatformKeyboardEventWPE.cpp
 165         https://bugs.webkit.org/show_bug.cgi?id=191606
 166
 167         Reviewed by Michael Catanzaro.
 168
 169         No behavior change.
 170
 171         Use StringBuilder::append(UChar32) as a generic way to convert a uint32_t code point to WTFString.
 172
 173         * platform/wpe/PlatformKeyboardEventWPE.cpp:
 174         (WebCore::PlatformKeyboardEvent::keyValueForWPEKeyCode):
 175         (WebCore::PlatformKeyboardEvent::singleCharacterString):
 176
 177 2018-11-13  Zalan Bujtas  <zalan@apple.com>
 178
 179         [LFC][IFC] Construct dedicated runs when the inline element requires it (part 2)
 180         https://bugs.webkit.org/show_bug.cgi?id=191623
 181
 182         Reviewed by Antti Koivisto.
 183
 184         This patch expands the breaking behaviour to support separate start/end breaks.
 185
 186         <span>parent </span><span style="padding: 10px;">start<span> middle </span>end</span><span> parent</span>
 187
 188         input to line breaking -> <parent start middle end parent>
 189         output of line breaking (considering infinite constraint) -> <parent start middle end parent>
 190         due to padding, final runs -> <parent><start middle end><parent>
 191
 192         "parent" -> n/a
 193         "start" -> BreakAtStart
 194         " middle " -> n/a
 195         "end" -> BreakAtEnd
 196         "parent" -> n/a
 197
 198         Another example:
 199         <span>parent </span><span style="padding-right: 10px;">start<span> middle </span>end</span><span> parent</span>
 200
 201         line breaking -> <parent start middle end parent>
 202         due to padding-right, final runs -> <parent start middle end><parent>
 203
 204         "parent" -> n/a
 205         "start" -> n/a
 206         " middle " -> n/a
 207         "end" -> BreakAtEnd
 208         "parent" -> n/a
 209
 210         * layout/inlineformatting/InlineFormattingContext.cpp:
 211         (WebCore::Layout::InlineFormattingContext::splitInlineRunIfNeeded const):
 212         (WebCore::Layout::InlineFormattingContext::collectInlineContent const): Move to a recursive algorithm (which is fine, inline contents don't tend to be too deep)
 213         (WebCore::Layout::InlineFormattingContext::contentRequiresSeparateRun const): Deleted.
 214         * layout/inlineformatting/InlineFormattingContext.h:
 215         * layout/inlineformatting/InlineFormattingState.cpp:
 216         (WebCore::Layout::InlineFormattingState::detachingRules const):
 217         * layout/inlineformatting/InlineFormattingState.h:
 218         (WebCore::Layout::InlineFormattingState::lastInlineItem const):
 219         (WebCore::Layout::InlineFormattingState::addDetachingRule):
 220
 221 2018-11-14  Youenn Fablet  <youenn@apple.com>
 222
 223         Add support for RTCRtpCodecParameters.sdpFmtpLine
 224         https://bugs.webkit.org/show_bug.cgi?id=191591
 225
 226         Reviewed by Eric Carlson.
 227
 228         Covered by rebased test.
 229
 230         * Modules/mediastream/RTCRtpCodecParameters.h:
 231         * Modules/mediastream/RTCRtpCodecParameters.idl:
 232         * Modules/mediastream/libwebrtc/LibWebRTCUtils.cpp:
 233         (WebCore::toRTCCodecParameters):
 234
 235 2018-11-14  Youenn Fablet  <youenn@apple.com>
 236
 237         Add support for transport and peerConnection stats
 238         https://bugs.webkit.org/show_bug.cgi?id=191592
 239
 240         Reviewed by Alex Christensen.
 241
 242         Covered by rebased tests.
 243
 244         * Modules/mediastream/RTCStatsReport.h:
 245         (WebCore::RTCStatsReport::TransportStats::TransportStats):
 246         (WebCore::RTCStatsReport::PeerConnectionStats::PeerConnectionStats):
 247         * Modules/mediastream/RTCStatsReport.idl:
 248         * Modules/mediastream/libwebrtc/LibWebRTCStatsCollector.cpp:
 249         (WebCore::fillRTCTransportStats):
 250         (WebCore::fillRTCPeerConnectionStats):
 251         (WebCore::LibWebRTCStatsCollector::OnStatsDelivered):
 252
 253 2018-11-14  Ali Juma  <ajuma@chromium.org>
 254
 255         Transform of composited layer not updated when layer also needs repaint
 256         https://bugs.webkit.org/show_bug.cgi?id=191598
 257
 258         Reviewed by Simon Fraser.
 259
 260         Trigger a compositing geometry update whenever a RenderLayer's transform changes,
 261         even when other parts of its style have changed in a way that produces a
 262         StyleDifference greater than RecompositeLayer.
 263
 264         Test: compositing/geometry/transform-and-repaint-updates-geometry.html
 265
 266         * rendering/RenderLayerCompositor.cpp:
 267         (WebCore::RenderLayerCompositor::layerStyleChanged):
 268
 269 2018-11-13  Jiewen Tan  <jiewen_tan@apple.com>
 270
 271         [WebAuthN] Support CTAP HID authenticators on macOS
 272         https://bugs.webkit.org/show_bug.cgi?id=188623
 273         <rdar://problem/43353777>
 274
 275         Reviewed by Brent Fulgham and Chris Dumez.
 276
 277         This patch removes AuthenticatorCoordinatorClient::~AuthenticatorCoordinatorClient to ignore
 278         any incompleted CompletionHandlers as calling them in destructors could cause unexpected cyclic
 279         dependency. Also, it adds a hack to temporarily deal with nullable userhandle.
 280
 281         Tests: http/wpt/webauthn/ctap-hid-failure.https.html
 282                http/wpt/webauthn/ctap-hid-success.https.html
 283                http/wpt/webauthn/public-key-credential-create-failure-hid-silent.https.html
 284                http/wpt/webauthn/public-key-credential-create-failure-hid.https.html
 285                http/wpt/webauthn/public-key-credential-create-success-hid.https.html
 286                http/wpt/webauthn/public-key-credential-get-failure-hid-silent.https.html
 287                http/wpt/webauthn/public-key-credential-get-failure-hid.https.html
 288                http/wpt/webauthn/public-key-credential-get-success-hid.https.html
 289
 290         * Modules/webauthn/AuthenticatorCoordinatorClient.cpp:
 291         (WebCore::AuthenticatorCoordinatorClient::~AuthenticatorCoordinatorClient): Deleted.
 292         * Modules/webauthn/AuthenticatorCoordinatorClient.h:
 293         * Modules/webauthn/PublicKeyCredentialCreationOptions.h:
 294         * Modules/webauthn/fido/DeviceResponseConverter.cpp:
 295         (fido::readCTAPGetAssertionResponse):
 296         * Modules/webauthn/fido/FidoConstants.h:
 297
 298 2018-11-13  Ross Kirsling  <ross.kirsling@sony.com>
 299
 300         [WebRTC] Provide default implementation of LibWebRTCProvider
 301         https://bugs.webkit.org/show_bug.cgi?id=191611
 302
 303         Reviewed by Michael Catanzaro.
 304
 305         Refactor LibWebRTCProvider such that platform-specific implementations need not worry about specifying defaults.
 306
 307         * PlatformWin.cmake:
 308         * platform/GStreamer.cmake:
 309         * platform/SourcesGLib.txt:
 310         * platform/mediastream/libwebrtc/LibWebRTCProvider.cpp:
 311         * platform/mediastream/libwebrtc/LibWebRTCProviderCocoa.cpp:
 312         * platform/mediastream/libwebrtc/LibWebRTCProviderGStreamer.cpp: Renamed from Source/WebCore/platform/mediastream/libwebrtc/LibWebRTCProviderGlib.cpp.
 313         * platform/mediastream/libwebrtc/LibWebRTCProviderGStreamer.h: Renamed from Source/WebCore/platform/mediastream/libwebrtc/LibWebRTCProviderGlib.h.
 314         * platform/mediastream/libwebrtc/LibWebRTCProviderWin.cpp: Removed.
 315
 316 2018-11-13  Timothy Hatcher  <timothy@apple.com>
 317
 318         Use a light scrollbar for transparent web views in dark mode.
 319         https://bugs.webkit.org/show_bug.cgi?id=191559
 320         rdar://problem/46000489
 321
 322         Reviewed by Dean Jackson.
 323
 324         Test: css-dark-mode/supported-color-schemes-scrollbar.html
 325
 326         * css/CSSProperties.json: Marked supported-color-schemes as a custom Value.
 327         * css/StyleBuilderCustom.h:
 328         (WebCore::StyleBuilderCustom::applyValueSupportedColorSchemes):
 329         * editing/cocoa/WebContentReaderCocoa.mm: Use FrameView's useDarkAppearance().
 330         (WebCore::createFragment):
 331         * inspector/InspectorOverlay.cpp:
 332         (WebCore::InspectorOverlay::paint): Use FrameView's useDarkAppearance().
 333         * page/FrameView.cpp:
 334         (WebCore::FrameView::recalculateScrollbarOverlayStyle): Use a light scrollbar for
 335         transparent web views in dark mode.
 336         (WebCore::FrameView::rendererForSupportedColorSchemes const): Added.
 337         Return the body for document element renderer.
 338         (WebCore::FrameView::useDarkAppearance const): Use rendererForSupportedColorSchemes.
 339         (WebCore::FrameView::styleColorOptions const): Added. Ditto.
 340         * page/FrameView.h:
 341         * rendering/style/RenderStyle.cpp:
 342         (WebCore::rareInheritedDataChangeRequiresRepaint): Drive-by fix. Added supportedColorSchemes.
 343         * rendering/style/RenderStyle.h:
 344         (WebCore::RenderStyle::setHasExplicitlySetSupportedColorSchemes): Added.
 345         (WebCore::RenderStyle::hasExplicitlySetSupportedColorSchemes const): Added.
 346         (WebCore::RenderStyle::NonInheritedFlags::operator== const): Added supportedColorSchemes.
 347         * svg/graphics/SVGImage.cpp:
 348         (WebCore::SVGImage::draw): Use FrameView's useDarkAppearance().
 349         * testing/Internals.cpp:
 350         (WebCore::Internals::setViewIsTransparent): Added.
 351         (WebCore::Internals::scrollbarOverlayStyle const): Added.
 352         * testing/Internals.h:
 353         * testing/Internals.idl: Added setViewIsTransparent and scrollbarOverlayStyle.
 354
 355 2018-11-13  Ross Kirsling  <ross.kirsling@sony.com>
 356
 357         [AppleWin] Unreviewed build fix after r238108.
 358
 359         * platform/graphics/ca/win/PlatformCALayerWin.cpp:
 360         (printLayer):
 361         (PlatformCALayerWin::embeddedViewID const):
 362         * platform/graphics/ca/win/PlatformCALayerWin.h:
 363
 364 2018-11-13  Youenn Fablet  <youenn@apple.com>
 365
 366         RTCPeerConnection.getTransceivers is not always exposing all transceivers
 367         https://bugs.webkit.org/show_bug.cgi?id=191589
 368
 369         Reviewed by Eric Carlson.
 370
 371         Implement the collect transceiver algorithm using libwebrtc backend.
 372         Call this algorithm everytime transceivers are retrieved from JS.
 373         For Plan B, make this a no-op as this is not supported.
 374         Introduce senders/receivers/transceivers getters where we just look at already created transceivers.
 375
 376         Covered by existing and rebased tests.
 377
 378         * Modules/mediastream/PeerConnectionBackend.h:
 379         (WebCore::PeerConnectionBackend::collectTransceivers):
 380         * Modules/mediastream/RTCPeerConnection.cpp:
 381         (WebCore::RTCPeerConnection::getSenders const):
 382         (WebCore::RTCPeerConnection::getReceivers const):
 383         (WebCore::RTCPeerConnection::getTransceivers const):
 384         * Modules/mediastream/RTCPeerConnection.h:
 385         * Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:
 386         (WebCore::LibWebRTCMediaEndpoint::collectTransceivers):
 387         * Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.h:
 388         * Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.cpp:
 389         (WebCore::LibWebRTCPeerConnectionBackend::addTrack):
 390         (WebCore::LibWebRTCPeerConnectionBackend::existingTransceiver):
 391         (WebCore::LibWebRTCPeerConnectionBackend::collectTransceivers):
 392         (WebCore::LibWebRTCPeerConnectionBackend::applyRotationForOutgoingVideoSources):
 393         (WebCore::LibWebRTCPeerConnectionBackend::shouldOfferAllowToReceive const):
 394         * Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.h:
 395
 396 2018-11-13  Wenson Hsieh  <wenson_hsieh@apple.com>
 397
 398         [iOS] Do not show selection UI for editable elements with opacity near zero
 399         https://bugs.webkit.org/show_bug.cgi?id=191442
 400         <rdar://problem/45958625>
 401
 402         Reviewed by Simon Fraser.
 403
 404         Tests: editing/selection/ios/do-not-zoom-to-focused-hidden-contenteditable.html
 405                editing/selection/ios/hide-selection-after-hiding-contenteditable.html
 406                editing/selection/ios/hide-selection-in-contenteditable-nested-transparency.html
 407                editing/selection/ios/hide-selection-in-hidden-contenteditable-frame.html
 408                editing/selection/ios/hide-selection-in-hidden-contenteditable.html
 409
 410         * rendering/RenderObject.cpp:
 411         (WebCore::RenderObject::isTransparentRespectingParentFrames const):
 412
 413         Add a helper function to determine whether a RenderObject is contained within a transparent layer, taking parent
 414         frames into account. A layer is considered transparent if its opacity is less than a small threshold (i.e. 0.01).
 415         Opacity on ancestor elements is applied multiplicatively.
 416
 417         * rendering/RenderObject.h:
 418
 419 2018-11-13  Eric Carlson  <eric.carlson@apple.com>
 420
 421         [MediaStream] Observer AVCaptureDevice "suspended" property
 422         https://bugs.webkit.org/show_bug.cgi?id=191587
 423         <rdar://problem/46030598>
 424
 425         Reviewed by Youenn Fablet.
 426
 427         No new tests, AVCapture can only be tested manually.
 428
 429         * platform/mediastream/mac/AVCaptureDeviceManager.h:
 430         * platform/mediastream/mac/AVCaptureDeviceManager.mm:
 431         (WebCore::AVCaptureDeviceManager::captureDevicesInternal): Don't notify of devices "changes"
 432         the first time the device list is scanned.
 433         (WebCore::deviceIsAvailable): Don't check for "isInUseByAnotherApplication", it doesn't
 434         change device availability.
 435         (WebCore::AVCaptureDeviceManager::beginObservingDevices): New, observe "suspended" on all
 436         devices and add them to the cached list.
 437         (WebCore::AVCaptureDeviceManager::stopObservingDevices): New, opposite of above.
 438         (WebCore::AVCaptureDeviceManager::refreshCaptureDevices): Watch for changes in the list of
 439         devices.
 440         (WebCore::AVCaptureDeviceManager::~AVCaptureDeviceManager): Stop observing all cached devices.
 441         (WebCore::AVCaptureDeviceManager::registerForDeviceNotifications):
 442         (-[WebCoreAVCaptureDeviceManagerObserver disconnect]):
 443         (-[WebCoreAVCaptureDeviceManagerObserver deviceConnectedDidChange:]):
 444         (-[WebCoreAVCaptureDeviceManagerObserver observeValueForKeyPath:ofObject:change:context:]):
 445         (WebCore::AVCaptureDeviceManager::refreshAVCaptureDevicesOfType): Deleted.
 446         (WebCore::AVCaptureDeviceManager::deviceConnected): Deleted.
 447         (WebCore::AVCaptureDeviceManager::deviceDisconnected): Deleted.
 448         (-[WebCoreAVCaptureDeviceManagerObserver deviceDisconnected:]): Deleted.
 449         (-[WebCoreAVCaptureDeviceManagerObserver deviceConnected:]): Deleted.
 450
 451         * platform/mediastream/mac/AVVideoCaptureSource.h:
 452         * platform/mediastream/mac/AVVideoCaptureSource.mm:
 453         (WebCore::AVVideoCaptureSource::~AVVideoCaptureSource): Stop observing "running" (not "rate")
 454         and "suspended".
 455         (WebCore::AVVideoCaptureSource::setupSession): Observe "running" (not "rate"), and "suspended".
 456         (WebCore::AVVideoCaptureSource::captureDeviceSuspendedDidChange):
 457         (-[WebCoreAVVideoCaptureSourceObserver observeValueForKeyPath:ofObject:change:context:]):
 458
 459 2018-11-13  Devin Rousso  <drousso@apple.com>
 460
 461         Web Inspector: REGRESSION(r238122): fetching the CertificateInfo triggers an ASSERT in workers
 462         https://bugs.webkit.org/show_bug.cgi?id=191597
 463
 464         Reviewed by Joseph Pecoraro.
 465
 466         When WebInspector is open, the `CertificateInfo` for every `ResourceResponse` is now fetched,
 467         meaning that we may try to fetch in situations previously unexpected.
 468
 469         * platform/network/cocoa/ResourceResponseCocoa.mm:
 470         (WebCore::ResourceResponse::platformCertificateInfo const):
 471
 472 2018-11-13  Timothy Hatcher  <timothy@apple.com>
 473
 474         Treat supported-color-schemes as the second highest priority property.
 475         https://bugs.webkit.org/show_bug.cgi?id=191556
 476         rdar://problem/46000076
 477
 478         Reviewed by Dean Jackson.
 479
 480         Test: css-dark-mode/supported-color-schemes-priority.html
 481
 482         * css/CSSProperties.json: Make supported-color-schemes high-priority and add a comment.
 483         * css/StyleResolver.cpp:
 484         (WebCore::StyleResolver::applyMatchedProperties): Manually handle supported-color-schemes
 485         after -webkit-ruby-position, before other properties, so it can affect resolved colors.
 486
 487 2018-11-13  Charlie Turner  <cturner@igalia.com>
 488
 489         [EME][GStreamer] Make CDMInstance's available in decryptors, and factor out some EME utility classes.
 490         https://bugs.webkit.org/show_bug.cgi?id=191316
 491
 492         Reviewed by Xabier Rodriguez-Calvar.
 493
 494         Another preparation in patch getting ready to move the decryption
 495         logic behind the CDMInstance and out of the GStreamer decryptors
 496         themselves. The first step taken here is to arrange for the
 497         instances to always be available in the decryptors when they need
 498         to decrypt content.
 499
 500         In doing so, there were a number of hairy bits of code that could
 501         use some abstraction, so the opportunity was taken to do that as
 502         well.
 503
 504         Covered by tests in media/encrypted-media and
 505         imported/w3c/web-platform-tests/encrypted-media.
 506
 507         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
 508         (WebCore::MediaPlayerPrivateGStreamer::handleMessage): Remove
 509         drm-key-needed since it was not being used anywhere.
 510         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
 511         (WebCore::MediaPlayerPrivateGStreamerBase::handleSyncMessage):
 512         Factor out the parsing of decryption system information from
 513         GStreamer, since it was not clear what that code was doing unless
 514         you squinted pretty hard. Also remove the duplicated
 515         initialization-data-encountered posting.
 516         (WebCore::MediaPlayerPrivateGStreamerBase::initializationDataEncountered):
 517         Refactored to make it a more general method and usable in more
 518         situations. It now has an optional to stop it from eliding init
 519         datas for a different key system. This is required the first time
 520         we post them, since if a CDM instance has already been set, and if
 521         the stream init datas are for different systems, we ended up never
 522         posting an encrypted event.
 523         (WebCore::MediaPlayerPrivateGStreamerBase::attemptToDecryptWithLocalInstance):
 524         Actually send a CDMInstance now when in regular playback mode.
 525         (WebCore::MediaPlayerPrivateGStreamerBase::dispatchDecryptionKey):
 526         Remove m_needToSendCredentials, it was not being used.
 527         (WebCore::MediaPlayerPrivateGStreamerBase::handleProtectionEvent):
 528         Refactored to use the new initializationDataEncountered.
 529         (WebCore::MediaPlayerPrivateGStreamerBase::reportWaitingForKey):
 530         Log the waiting state, since it was currently not clear what that
 531         logging message was even telling you!
 532         (WebCore::extractEventsAndSystemsFromMessage): Deleted.
 533         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.h:
 534         * platform/graphics/gstreamer/eme/GStreamerEMEUtilities.h:
 535         (WebCore::InitData::InitData): New class that encapsulates both
 536         single instantiation and streaming instantiation.
 537         (WebCore::InitData::append): Used for the streaming mode, when you
 538         are concatenating init datas together.
 539         (WebCore::InitData::payload const):
 540         (WebCore::InitData::systemId const):
 541         (WebCore::InitData::payloadContainerType const):
 542         (WebCore::InitData::isFromDifferentContainer):
 543         (WebCore::ProtectionSystemEvents::ProtectionSystemEvents):
 544         (WebCore::ProtectionSystemEvents::events const):
 545         (WebCore::ProtectionSystemEvents::availableSystems const):
 546         * platform/graphics/gstreamer/eme/WebKitCommonEncryptionDecryptorGStreamer.cpp:
 547         (webkitMediaCommonEncryptionDecryptTransformInPlace): If you post
 548         waiting-for-key after requesting a CDM instance, it will flap back
 549         to not waiting for a key almost immediately, didn't make sense
 550         positing after requesting an instance. Also post key-received when
 551         we receive the key.
 552         (webkitMediaCommonEncryptionDecryptSinkEventHandler): It has now
 553         been arranged that a CDMInstance will always be present in an OOB
 554         message, so parse it out here.
 555         * platform/graphics/gstreamer/mse/MediaPlayerPrivateGStreamerMSE.cpp:
 556         (WebCore::MediaPlayerPrivateGStreamerMSE::attemptToDecryptWithInstance):
 557         As above, make sure when posting the OOB that a CDMInstance is present.
 558
 559 2018-11-13  Charlie Turner  <cturner@igalia.com>
 560
 561         Various compiler warnings/errors fixes.
 562         https://bugs.webkit.org/show_bug.cgi?id=191583
 563
 564         Reviewed by Frédéric Wang.
 565
 566         * Modules/indexeddb/server/MemoryIDBBackingStore.cpp:
 567         (WebCore::IDBServer::MemoryIDBBackingStore::clearObjectStore):
 568         ASSERT is only compiled in DEBUG mode, so guarding with
 569         !LOG_DISABLED is wrong.
 570         * rendering/RenderLayerCompositor.cpp:
 571         (WebCore::RenderLayerCompositor::updateCompositingLayers):
 572         showPaintOrderTree is only compiled in ENABLE(TREE_DEBUGGING)
 573         mode, so guarding with !LOG_DISABLED was wrong.
 574         (WebCore::RenderLayerCompositor::computeCompositingRequirements):
 575         Ditto, this time with member .depth.
 576         (WebCore::RenderLayerCompositor::traverseUnchangedSubtree): Ditto.
 577         * rendering/svg/SVGRenderSupport.cpp:
 578         (WebCore::SVGRenderSupport::styleChanged): Add another unused
 579         parameter.
 580
 581 2018-11-12  Antoine Quint  <graouts@apple.com>
 582
 583         [Web Animations] Don't schedule animation frames or update style while an accelerated animation is running
 584         https://bugs.webkit.org/show_bug.cgi?id=191542
 585         <rdar://problem/45356027>
 586
 587         Reviewed by Simon Fraser.
 588
 589         Test: animations/no-style-recalc-during-accelerated-animation.html
 590
 591         In order to be more power-efficient, we stop scheduling calls to updateAnimationsAndSendEvents() when running only accelerated
 592         animations. To do that, we prevent scheduling further animation resolution if we're in the process of updating animations, and
 593         when we are done, call the new DocumentTimeline::scheduleNextTick() method that will check whether we have only accelerated
 594         animations running, and in that case check which of those animations needs an update the soonest and starts a timer scheduled
 595         for that time when we'll schedule animation resolution.
 596
 597         By default, animations compute the time until their natural completion but in the case of CSS Animations, we want to make sure
 598         we also update animations in-flight to dispatch "animationiteration" events.
 599
 600         * animation/AnimationEffect.h: Make the simpleIterationProgress() public so it can be called by WebAnimation::timeToNextTick().
 601         * animation/DocumentTimeline.cpp:
 602         (WebCore::DocumentTimeline::DocumentTimeline): Create the m_tickScheduleTimer and set it up to call scheduleAnimationResolutionIfNeeded().
 603         (WebCore::DocumentTimeline::suspendAnimations): If we don't already have a cached current time, cache the current time.
 604         (WebCore::DocumentTimeline::resumeAnimations): Reset the cached current time to ensure we'll get a fresh one when updating animations next.
 605         (WebCore::DocumentTimeline::liveCurrentTime const): Factor the code to compute the current time out of currentTime() so that we can
 606         cache the current time in suspendAnimations() without also automatically clearing the current time.
 607         (WebCore::DocumentTimeline::currentTime): Use liveCurrentTime() and cacheCurrentTime() since much of the code from this function has been
 608         factored out into those. Additionally, we were failing to clear the current time if called inside an animation frame, which we now do correctly
 609         by virtue of using cacheCurrentTime(). This fixes some flakiness.
 610         (WebCore::DocumentTimeline::cacheCurrentTime): Factor the code to cache the current time out of currentTime().
 611         (WebCore::DocumentTimeline::maybeClearCachedCurrentTime): No need to clear the current time if we get suspended.
 612         (WebCore::DocumentTimeline::scheduleAnimationResolutionIfNeeded): Prevent scheduling an animation update if we're in the middle of one already,
 613         scheduleNextTick() will be called after animations are updated to see if we should schedule an animation update instead.
 614         (WebCore::DocumentTimeline::unscheduleAnimationResolution): Cancel the m_tickScheduleTimer if we need to unschedule animation resolution.
 615         (WebCore::DocumentTimeline::animationResolutionTimerFired): Factor the call to applyPendingAcceleratedAnimations() out of updateAnimationsAndSendEvents()
 616         and call scheduleNextTick().
 617         (WebCore::DocumentTimeline::updateAnimationsAndSendEvents): Set the new m_isUpdatingAnimations member variable to true while this function is running.
 618         (WebCore::DocumentTimeline::scheduleNextTick): Schedule an animation update immediately if we have any relevant animation that is not accelerated.
 619         Otherwise, iterate through all animations to figure out the earliest moment at which we need to update animations.
 620         (WebCore::DocumentTimeline::updateListOfElementsWithRunningAcceleratedAnimationsForElement): Use the new WebAnimation::isRunningAccelerated() function.
 621         * animation/DocumentTimeline.h:
 622         * animation/WebAnimation.cpp:
 623         (WebCore::WebAnimation::isRunningAccelerated const): Since we end up checking if an animation is running with an accelerated effect, we introduce a new
 624         function to get that information directly through the WebAnimation object without bothering about its effect.
 625         (WebCore::WebAnimation::resolve): We should only call updateFinishedState() here since timingDidChange() would also notify the timeline about a potential
 626         change in relevance, which is not necessary and which would schedule an animation frame even for animations that are accelerated.
 627         (WebCore::WebAnimation::timeToNextTick const): Compute the time until our animation completion or, in the case of CSS animations, the next iteration.
 628         * animation/WebAnimation.h:
 629
 630 2018-11-13  Miguel Gomez  <magomez@igalia.com>
 631
 632         [GTK][WPE] Incorrect tile coverage when resizing a layer out of the visible area
 633         https://bugs.webkit.org/show_bug.cgi?id=191545
 634
 635         Reviewed by Žan Doberšek.
 636
 637         Keep track of layer size changes even if they happen when the layer is not in the visible
 638         area, so we can update edge tiles when the layer gets visible.
 639
 640         * platform/graphics/texmap/coordinated/TiledBackingStore.cpp:
 641         (WebCore::TiledBackingStore::createTiles):
 642         * platform/graphics/texmap/coordinated/TiledBackingStore.h:
 643
 644 2018-11-12  Rob Buis  <rbuis@igalia.com>
 645
 646         Content-Type parameter values should allow empty quoted strings
 647         https://bugs.webkit.org/show_bug.cgi?id=191388
 648
 649         Reviewed by Dean Jackson.
 650
 651         According to RFC 2045 and https://mimesniff.spec.whatwg.org/#parsing-a-mime-type empty
 652         quoted strings are acceptable for Content-Type parameter values. They
 653         are accepted by Firefox and Chrome implementations as well.
 654
 655         Test: web-platform-tests/xhr/overridemimetype-blob.html
 656
 657         * platform/network/ParsedContentType.cpp:
 658         (WebCore::parseToken):
 659         (WebCore::parseQuotedString):
 660         (WebCore::parseContentType):
 661         * platform/network/ParsedContentType.h:
 662
 663 2018-11-12  Christopher Reid  <chris.reid@sony.com>
 664
 665         [Curl] Reject entire cookie if the domain fails a tailmatch.
 666         https://bugs.webkit.org/show_bug.cgi?id=191406
 667
 668         Reviewed by Youenn Fablet.
 669
 670         Currently we don't put domain attribute of cookie when it fails a tailmatch. As Firefox
 671         and Chrome do, we are going to reject the entire cookie if the domain fails a tailmatch instead.
 672         Also cleanup Cookie database implementation to make them testable better.
 673
 674         Tests: TestWebKitAPI/Tests/WebCore/curl/Cookies.cpp
 675
 676         * platform/network/curl/CookieJarDB.cpp:
 677         (WebCore::CookieJarDB::canAcceptCookie): Added.
 678         (WebCore::CookieJarDB::setCookie):
 679         * platform/network/curl/CookieUtil.cpp:
 680         (WebCore::CookieUtil::parseCookieAttributes):
 681         (WebCore::CookieUtil::parseCookieHeader):
 682         * platform/network/curl/CookieUtil.h:
 683
 684 2018-11-12  Devin Rousso  <drousso@apple.com>
 685
 686         Web Inspector: Network: show secure certificate details per-request
 687         https://bugs.webkit.org/show_bug.cgi?id=191447
 688         <rdar://problem/30019476>
 689
 690         Reviewed by Joseph Pecoraro.
 691
 692         Test: http/tests/inspector/network/resource-response-security.html
 693
 694         * loader/ResourceLoader.h:
 695         (WebCore::ResourceLoader::shouldIncludeCertificateInfo const):
 696         * loader/ResourceLoader.cpp:
 697         (WebCore::ResourceLoader::shouldIncludeCertificateInfo const): Added.
 698         Always save certificate information when WebInspector is open.
 699
 700         * platform/network/CertificateInfoBase.h: Added.
 701         (WebCore::CertificateInfoBase::containsNonRootSHA1SignedCertificate const):
 702         (WebCore::CertificateInfoBase::summaryInfo const):
 703         (WebCore::CertificateInfoBase::isEmpty const):
 704         * platform/network/cf/CertificateInfo.h:
 705         (WebCore::CertificateInfo::summaryInfo const): Added.
 706         * platform/network/cf/CertificateInfoCFNet.cpp: Renamed from Source/WebCore/platform/network/mac/CertificateInfoMac.mm.
 707         (WebCore::CertificateInfo::containsNonRootSHA1SignedCertificate):
 708         (WebCore::CertificateInfo::summaryInfo const): Added.
 709         * platform/network/curl/CertificateInfo.h:
 710         (WebCore::CertificateInfo::summaryInfo const): Added.
 711         (WebCore::CertificateInfo::isEmpty const): Added.
 712         * platform/network/soup/CertificateInfo.h:
 713         (WebCore::CertificateInfo::summaryInfo const): Added.
 714         (WebCore::CertificateInfo::isEmpty const): Added.
 715         Create base class for `CertificateInfo` so that `InspectorNetworkAgent` doesn't need to have
 716         platform-specific code in its implementation.
 717
 718         * platform/network/cocoa/CertificateInfoCocoa.mm: Renamed from Source/WebCore/platform/network/mac/CertificateInfoMac.mm.
 719         * platform/network/curl/CertificateInfoCFNet.cpp: Renamed from Source/WebCore/platform/network/curl/CertificateInfo.cpp.
 720         * platform/network/soup/CertificateInfoSoup.cpp: Renamed from Source/WebCore/platform/network/soup/CertificateInfo.cpp.
 721
 722         * inspector/NetworkResourcesData.h:
 723         (WebCore::NetworkResourcesData::ResourceData::certificateInfo const): Added.
 724         (WebCore::NetworkResourcesData::ResourceData::setCertificateInfo): Added.
 725         * inspector/NetworkResourcesData.cpp:
 726         (WebCore::NetworkResourcesData::responseReceived):
 727
 728         * inspector/agents/InspectorNetworkAgent.cpp:
 729         (WebCore::InspectorNetworkAgent::buildObjectForResourceResponse):
 730
 731         * PlatformAppleWin.cmake:
 732         * PlatformMac.cmake:
 733         * SourcesCocoa.txt:
 734         * WebCore.xcodeproj/project.pbxproj:
 735         * platform/Curl.cmake:
 736         * platform/SourcesSoup.txt:
 737
 738 2018-11-12  Zalan Bujtas  <zalan@apple.com>
 739
 740         Do not collapse the soon-to-be-parent anon block when we shuffle around the marker item renderer.
 741         https://bugs.webkit.org/show_bug.cgi?id=191554
 742         <rdar://problem/45825265>
 743
 744         Reviewed by Antti Koivisto.
 745
 746         While moving the marker item renderer to its correct subtree, we accidentally remove the soon-to-be parent anonymous block.
 747         Moving a renderer is a 2 step process:
 748         1. Detach the renderer from its current parent
 749         2. Attach it to its new parent.
 750         During step #1, we check if there is a chance to collapse anonymous blocks. In this case the soon-to-be-parent is a sibling anonymous block which, after detaching the marker sibling
 751         is not needed anymore (except we use it as the new parent).
 752
 753         Test: fast/inline/marker-list-item-move-should-not-crash.html
 754
 755         * rendering/updating/RenderTreeBuilder.cpp:
 756         (WebCore::RenderTreeBuilder::detach):
 757         * rendering/updating/RenderTreeBuilder.h:
 758         * rendering/updating/RenderTreeBuilderBlock.cpp:
 759         (WebCore::RenderTreeBuilder::Block::detach):
 760         * rendering/updating/RenderTreeBuilderBlock.h:
 761         * rendering/updating/RenderTreeBuilderList.cpp:
 762         (WebCore::RenderTreeBuilder::List::updateItemMarker):
 763
 764 2018-11-12  Javier Fernandez  <jfernandez@igalia.com>
 765
 766         [css-grid] Refactoring to make more explicit the orthogonal items' pre-layout logic
 767         https://bugs.webkit.org/show_bug.cgi?id=191358
 768
 769         Reviewed by Manuel Rego Casasnovas.
 770
 771         These changes are just a refactoring to ease the integration of the new Baseline Alignment
 772         logic in a follow up patch.
 773
 774         We need to properly estimate the grid area size of orthogonal items so that we can perform
 775         an accurate pre-layout. This is important because orthogonal items will synthesize their baseline
 776         if they participate in any baseline alignment context.
 777
 778         No new tests, since no behavior change has been introduced in this patch.
 779
 780         * rendering/Grid.cpp:
 781         (WebCore::Grid::setNeedsItemsPlacement):
 782         * rendering/Grid.h:
 783         * rendering/GridTrackSizingAlgorithm.cpp:
 784         (WebCore::GridTrackSizingAlgorithm::estimatedGridAreaBreadthForChild const):
 785         (WebCore::GridTrackSizingAlgorithm::gridAreaBreadthForChild const):
 786         (WebCore::GridTrackSizingAlgorithm::isRelativeGridLengthAsAuto const):
 787         (WebCore::GridTrackSizingAlgorithm::isRelativeSizedTrackAsAuto const):
 788         (WebCore::GridTrackSizingAlgorithm::gridTrackSize const):
 789         (WebCore::IndefiniteSizeStrategy::findUsedFlexFraction const):
 790         (WebCore::GridTrackSizingAlgorithm::run):
 791         (WebCore::GridTrackSizingAlgorithm::reset):
 792         * rendering/GridTrackSizingAlgorithm.h:
 793         (WebCore::GridTrackSizingAlgorithmStrategy::gridTrackSize const):
 794         * rendering/RenderGrid.cpp:
 795         (WebCore::RenderGrid::repeatTracksSizingIfNeeded):
 796         (WebCore::RenderGrid::layoutBlock):
 797         (WebCore::RenderGrid::computeIntrinsicLogicalWidths const):
 798         (WebCore::RenderGrid::computeTrackSizesForIndefiniteSize const):
 799         (WebCore::RenderGrid::placeItemsOnGrid const):
 800         (WebCore::RenderGrid::performGridItemsPreLayout const):
 801         (WebCore::overrideSizeChanged):
 802         (WebCore::hasRelativeBlockAxisSize):
 803         (WebCore::RenderGrid::updateGridAreaLogicalSize const):
 804         (WebCore::RenderGrid::layoutGridItems):
 805         * rendering/RenderGrid.h:
 806
 807 2018-11-12  Sihui Liu  <sihui_liu@apple.com>
 808
 809         imported/w3c/web-platform-tests/IndexedDB/keygenerator-explicit.html crashing on iOS device
 810         https://bugs.webkit.org/show_bug.cgi?id=191500
 811
 812         Reviewed by Dean Jackson.
 813
 814         When double value is bigger than maximum unsigned int, converting double to unsigned int has
 815         different behaviors on macOS and iOS. On macOS, the result would be 0 while on iOS it would be
 816         maximum unsigned int.
 817
 818         Covered by existing test.
 819
 820         * Modules/indexeddb/server/SQLiteIDBBackingStore.cpp:
 821         (WebCore::IDBServer::SQLiteIDBBackingStore::generateKeyNumber):
 822         (WebCore::IDBServer::SQLiteIDBBackingStore::maybeUpdateKeyGeneratorNumber):
 823
 824 2018-11-12  Basuke Suzuki  <basuke.suzuki@sony.com>
 825
 826         [Curl] Add API Test for Curl cookie backend.
 827         https://bugs.webkit.org/show_bug.cgi?id=191493
 828
 829         Reviewed by Youenn Fablet.
 830
 831         Refactoring for cookie backend interface.
 832
 833         Tests: TestWebKitAPI/Tests/WebCore/curl/Cookies.cpp
 834
 835         * platform/FileSystem.h:
 836         * platform/network/curl/CookieJarCurlDatabase.cpp:
 837         (WebCore::cookiesForSession):
 838         (WebCore::CookieJarCurlDatabase::setCookiesFromDOM const):
 839         (WebCore::CookieJarCurlDatabase::setCookiesFromHTTPResponse const):
 840         (WebCore::CookieJarCurlDatabase::getRawCookies const):
 841         * platform/network/curl/CookieJarDB.cpp:
 842         (WebCore::CookieJarDB::openDatabase):
 843         (WebCore::CookieJarDB::checkSQLiteReturnCode):
 844         (WebCore::CookieJarDB::isEnabled const):
 845         (WebCore::CookieJarDB::searchCookies):
 846         (WebCore::CookieJarDB::setCookie):
 847         (WebCore::CookieJarDB::deleteCookie):
 848         (WebCore::CookieJarDB::deleteCookieInternal):
 849         (WebCore::CookieJarDB::deleteCookies):
 850         (WebCore::CookieJarDB::deleteAllCookies):
 851         (WebCore::CookieJarDB::executeSimpleSql):
 852         (WebCore::CookieJarDB::isEnabled): Deleted.
 853         * platform/network/curl/CookieJarDB.h:
 854         * platform/network/curl/CookieUtil.cpp:
 855         (WebCore::CookieUtil::parseCookieHeader):
 856         * platform/network/curl/CookieUtil.h:
 857         * platform/win/FileSystemWin.cpp:
 858         (WebCore::FileSystem::generateTemporaryPath):
 859         (WebCore::FileSystem::openTemporaryFile):
 860         (WebCore::FileSystem::createTemporaryDirectory):
 861         (WebCore::FileSystem::deleteNonEmptyDirectory):
 862
 863 2018-11-12  Tim Horton  <timothy_horton@apple.com>
 864
 865         Make it possible to edit images inline
 866         https://bugs.webkit.org/show_bug.cgi?id=191352
 867         <rdar://problem/30107985>
 868
 869         Reviewed by Dean Jackson.
 870
 871         Tests: editing/images/basic-editable-image.html
 872                editing/images/reparent-editable-image-maintains-strokes.html
 873
 874         Add the beginnings of a mechanism to replace images with a special attribute
 875         with a native drawing view in the UI process.
 876
 877         * page/Settings.yaml:
 878         Add a setting to control whether images become natively editable when they
 879         have the x-apple-editable-image attribute.
 880
 881         * html/HTMLImageElement.cpp:
 882         (WebCore::HTMLImageElement::editableImageViewID const):
 883         Lazily generate an EmbeddedViewID and persist it on the <img> element.
 884
 885         * html/HTMLImageElement.h:
 886         Rearrange the service controls methods to sit before the members.
 887         Add m_editableImageViewID and editableImageViewID().
 888
 889         * platform/graphics/GraphicsLayer.cpp:
 890         (WebCore::GraphicsLayer::nextEmbeddedViewID):
 891         * platform/graphics/GraphicsLayer.h:
 892         (WebCore::GraphicsLayer::setContentsToEmbeddedView):
 893         Add a new ContentsLayerPurpose, EmbeddedView, which is only supported
 894         on Cocoa platforms and when using RemoteLayerTree.
 895         Add ContentsLayerEmbeddedViewType, which currently only has the EditableImage type.
 896         Add setContentsToEmbeddedView, which takes a ContentsLayerEmbeddedViewType
 897         and an EmbeddedViewID to uniquely identify and communicate about the
 898         embedded view (which may move between layers, since it is tied to an element).
 899
 900         * platform/graphics/ca/GraphicsLayerCA.cpp:
 901         (WebCore::GraphicsLayerCA::createPlatformCALayerForEmbeddedView):
 902         (WebCore::GraphicsLayerCA::setContentsToEmbeddedView):
 903         When setting GraphicsLayer's contents to an embedded view, we use
 904         a special PlatformCALayer factory that takes the EmbeddedViewID and type.
 905         GraphicsLayerCARemote will override this and make a correctly-initialized
 906         PlatformCALayerRemote that keeps track of the EmbeddedViewID.
 907
 908         * platform/graphics/ca/GraphicsLayerCA.h:
 909         * platform/graphics/ca/PlatformCALayer.cpp:
 910         (WebCore::operator<<):
 911         * platform/graphics/ca/PlatformCALayer.h:
 912         * platform/graphics/ca/cocoa/PlatformCALayerCocoa.h:
 913         * platform/graphics/ca/cocoa/PlatformCALayerCocoa.mm:
 914         (WebCore::PlatformCALayerCocoa::PlatformCALayerCocoa):
 915         (WebCore::PlatformCALayerCocoa::embeddedViewID const):
 916         Add stubs and logging for EmbeddedViewID on PlatformCALayer.
 917         These will be overridden by PlatformCALayerRemote to do more interesting things.
 918
 919         * rendering/RenderImage.cpp:
 920         (WebCore::RenderImage::isEditableImage const):
 921         Add a getter that return true if the setting is enabled and
 922         x-apple-editable-image is empty or true.
 923
 924         (WebCore::RenderImage::requiresLayer const):
 925         RenderImage requires a layer either if RenderReplaced does, or we are an
 926         editable image.
 927
 928         * rendering/RenderImage.h:
 929         * rendering/RenderLayer.cpp:
 930         (WebCore::RenderLayer::shouldBeNormalFlowOnly const):
 931         (WebCore::RenderLayer::calculateClipRects const):
 932         * rendering/RenderLayerBacking.cpp:
 933         (WebCore::RenderLayerBacking::updateConfiguration):
 934         Push the EmbeddedViewID and type down to GraphicsLayer for editable images.
 935
 936         * rendering/RenderLayerCompositor.cpp:
 937         (WebCore::RenderLayerCompositor::requiresCompositingLayer const):
 938         (WebCore::RenderLayerCompositor::requiresOwnBackingStore const):
 939         (WebCore::RenderLayerCompositor::reasonsForCompositing const):
 940         (WebCore::RenderLayerCompositor::requiresCompositingForEditableImage const):
 941         * rendering/RenderLayerCompositor.h:
 942         Make editable images require compositing implicitly.
 943
 944 2018-11-12  Don Olmstead  <don.olmstead@sony.com>
 945
 946         Shipped PNGs include bad profiles: iCCP: known incorrect sRGB profile
 947         https://bugs.webkit.org/show_bug.cgi?id=189230
 948         <rdar://problem/44050379>
 949
 950         Reviewed by Joseph Pecoraro.
 951
 952         Runs all png images through zopflipng. This results in a smaller file
 953         size and takes care of this issue as a byproduct.
 954
 955         * Modules/modern-media-controls/images/iOS/airplay-placard@1x.png:
 956         * Modules/modern-media-controls/images/iOS/airplay-placard@2x.png:
 957         * Modules/modern-media-controls/images/iOS/airplay-placard@3x.png:
 958         * Modules/modern-media-controls/images/iOS/invalid-placard@1x.png:
 959         * Modules/modern-media-controls/images/iOS/invalid-placard@2x.png:
 960         * Modules/modern-media-controls/images/iOS/invalid-placard@3x.png:
 961         * Modules/modern-media-controls/images/iOS/pip-placard@1x.png:
 962         * Modules/modern-media-controls/images/iOS/pip-placard@2x.png:
 963         * Modules/modern-media-controls/images/iOS/pip-placard@3x.png:
 964         * Modules/modern-media-controls/images/macOS/airplay-placard@1x.png:
 965         * Modules/modern-media-controls/images/macOS/airplay-placard@2x.png:
 966         * Modules/modern-media-controls/images/macOS/invalid-placard@1x.png:
 967         * Modules/modern-media-controls/images/macOS/invalid-placard@2x.png:
 968         * Modules/modern-media-controls/images/macOS/pip-placard@1x.png:
 969         * Modules/modern-media-controls/images/macOS/pip-placard@2x.png:
 970         * Resources/AttachmentPlaceholder.png:
 971         * Resources/AttachmentPlaceholder@2x.png:
 972         * Resources/ListButtonArrow.png:
 973         * Resources/ListButtonArrow@2x.png:
 974         * Resources/missingImage.png:
 975         * Resources/missingImage@2x.png:
 976         * Resources/missingImage@3x.png:
 977         * Resources/moveCursor.png:
 978         * Resources/northEastSouthWestResizeCursor.png:
 979         * Resources/northSouthResizeCursor.png:
 980         * Resources/northWestSouthEastResizeCursor.png:
 981         * Resources/nullPlugin.png:
 982         * Resources/nullPlugin@2x.png:
 983         * Resources/panIcon.png:
 984         * Resources/textAreaResizeCorner.png:
 985         * Resources/textAreaResizeCorner@2x.png:
 986         * Resources/urlIcon.png:
 987
 988 2018-11-12  Youenn Fablet  <youenn@apple.com>
 989
 990         RealtimeOutgoing A/V sources should observe their sources only if having a sink
 991         https://bugs.webkit.org/show_bug.cgi?id=191490
 992
 993         Reviewed by Eric Carlson.
 994
 995         Observe the source that generates media based on the sinks:
 996         - Do not observe at creation time
 997         - For first sink, start observing
 998         - When no more sink, stop observing
 999         Apply this principle for both outgoing audio and video sources.
1000         Add locks for the sinks to ensure thread-safety.
1001         Make sinks HashSet which is more robust.
1002
1003         Do some refactoring to better isolate generic outgoing sources from Cocoa/GTK implementations.
1004
1005         Covered by existing tests and updated webrtc/remove-track.html.
1006
1007         * platform/mediastream/RealtimeOutgoingAudioSource.cpp:
1008         (WebCore::RealtimeOutgoingAudioSource::~RealtimeOutgoingAudioSource):
1009         (WebCore::RealtimeOutgoingAudioSource::stop):
1010         (WebCore::RealtimeOutgoingAudioSource::AddSink):
1011         (WebCore::RealtimeOutgoingAudioSource::RemoveSink):
1012         (WebCore::RealtimeOutgoingAudioSource::sendAudioFrames):
1013         * platform/mediastream/RealtimeOutgoingAudioSource.h:
1014         * platform/mediastream/RealtimeOutgoingVideoSource.cpp:
1015         (WebCore::RealtimeOutgoingVideoSource::RealtimeOutgoingVideoSource):
1016         (WebCore::RealtimeOutgoingVideoSource::~RealtimeOutgoingVideoSource):
1017         (WebCore::RealtimeOutgoingVideoSource::observeSource):
1018         (WebCore::RealtimeOutgoingVideoSource::setSource):
1019         (WebCore::RealtimeOutgoingVideoSource::stop):
1020         (WebCore::RealtimeOutgoingVideoSource::AddOrUpdateSink):
1021         (WebCore::RealtimeOutgoingVideoSource::RemoveSink):
1022         * platform/mediastream/RealtimeOutgoingVideoSource.h:
1023         (WebCore::RealtimeOutgoingVideoSource::isSilenced const):
1024         * platform/mediastream/gstreamer/RealtimeOutgoingAudioSourceLibWebRTC.cpp:
1025         (WebCore::RealtimeOutgoingAudioSourceLibWebRTC::pullAudioData):
1026         * platform/mediastream/mac/RealtimeOutgoingAudioSourceCocoa.cpp:
1027         (WebCore::RealtimeOutgoingAudioSourceCocoa::RealtimeOutgoingAudioSourceCocoa):
1028         (WebCore::RealtimeOutgoingAudioSourceCocoa::audioSamplesAvailable):
1029         (WebCore::RealtimeOutgoingAudioSourceCocoa::pullAudioData):
1030         * platform/mediastream/mac/RealtimeOutgoingAudioSourceCocoa.h:
1031         * platform/mediastream/mac/RealtimeOutgoingVideoSourceCocoa.cpp:
1032         (WebCore::RealtimeOutgoingVideoSourceCocoa::sampleBufferUpdated):
1033
1034 2018-11-12  Youenn Fablet  <youenn@apple.com>
1035
1036         Support setting stream ids when adding a transceiver
1037         https://bugs.webkit.org/show_bug.cgi?id=191307
1038
1039         Reviewed by Eric Carlson.
1040
1041         Add support for streams in RTCTransceiverInit.
1042         Add plumbing down to libwebrtc.
1043         Covered by rebased tests.
1044
1045         * Modules/mediastream/RTCPeerConnection.h:
1046         * Modules/mediastream/RTCPeerConnection.idl:
1047         * Modules/mediastream/libwebrtc/LibWebRTCUtils.cpp:
1048         (WebCore::fromRtpTransceiverInit):
1049
1050 2018-11-12  Antti Koivisto  <antti@apple.com>
1051
1052         Support dynamic pseudo-classes on elements with display: contents
1053         https://bugs.webkit.org/show_bug.cgi?id=181640
1054         <rdar://problem/36605415>
1055
1056         Reviewed by Dean Jackson.
1057
1058         The code for :hover and :active style invalidation assumes that only elements with renderer need invalidation.
1059
1060         This patch fixes '.display-content-element:hover span' case but not '.display-content-element:hover' case but
1061         includes tests for both. The latter is not super useful anyway (as it only affects rendering with inherited
1062         text properties).
1063
1064         Test: fast/css/display-contents-hover-active.html
1065
1066         * dom/Document.cpp:
1067         (WebCore::Document::updateHoverActiveState):
1068
1069             Traverse up in composed tree instead of render tree when invalidating. This has the same order as render tree
1070             but also includes display:content elements. This also allows removing the special display:none case.
1071
1072         * dom/Element.cpp:
1073         (WebCore::Element::setActive):
1074         (WebCore::Element::setHovered):
1075
1076             Also look into display:contents style for invalidation checks.
1077
1078         (WebCore::Element::renderOrDisplayContentsStyle const):
1079
1080             Make this helper an Element member.
1081
1082         * dom/Element.h:
1083         * dom/Node.cpp:
1084         (WebCore::Node::parentElementInComposedTree const):
1085
1086             Support starting from a PseudoElement. This is consistent with ComposedTreeAncestorIterator.
1087
1088         * rendering/updating/RenderTreePosition.cpp:
1089         (WebCore::RenderTreePosition::nextSiblingRenderer const):
1090         * style/StyleTreeResolver.cpp:
1091         (WebCore::Style::TreeResolver::resolveElement):
1092         (WebCore::Style::TreeResolver::createAnimatedElementUpdate):
1093         (WebCore::Style::shouldResolveElement):
1094         (WebCore::Style::TreeResolver::resolveComposedTree):
1095         (WebCore::Style::renderOrDisplayContentsStyle): Deleted.
1096
1097             Use the Element::renderOrDisplayContentsStyle() instead.
1098
1099 2018-11-12  Antoine Quint  <graouts@apple.com>
1100
1101         [Web Animations] Turn Web Animations experimental
1102         https://bugs.webkit.org/show_bug.cgi?id=191543
1103
1104         Reviewed by Dean Jackson.
1105
1106         * page/RuntimeEnabledFeatures.h:
1107
1108 2018-11-12  Simon Fraser  <simon.fraser@apple.com>
1109
1110         feFlood with alpha color doesn't work correctly
1111         https://bugs.webkit.org/show_bug.cgi?id=163666
1112
1113         Reviewed by Zalan Bujtas.
1114
1115         FEFlood::platformApplySoftware() erroneously used colorWithOverrideAlpha()
1116         rather than multiplying the flood color with the flood opacity as other browsers do.
1117
1118         Test: svg/filters/feFlood-with-alpha-color.html
1119
1120         * platform/graphics/Color.cpp:
1121         (WebCore::Color::colorWithAlpha const): I tried using colorWithAlphaMultipliedBy() elsewhere,
1122         and it triggered a behavior change, so add a comment.
1123         * platform/graphics/filters/FEFlood.cpp:
1124         (WebCore::FEFlood::platformApplySoftware):
1125         * svg/SVGStopElement.cpp:
1126         (WebCore::SVGStopElement::stopColorIncludingOpacity const):
1127
1128 2018-11-12  Eric Carlson  <eric.carlson@apple.com>
1129
1130         Require <iframe allow="display"> for an iframe to use getDisplayMedia
1131         https://bugs.webkit.org/show_bug.cgi?id=191505
1132         <rdar://problem/45968811>
1133
1134         Reviewed by Jer Noble.
1135
1136         Test: http/tests/media/media-stream/get-display-media-iframe-allow-attribute.html
1137
1138         * Modules/mediastream/MediaDevicesRequest.cpp:
1139         (WebCore::MediaDevicesRequest::start):
1140         * Modules/mediastream/UserMediaController.cpp:
1141         (WebCore::isAllowedToUse):
1142         (WebCore::UserMediaController::canCallGetUserMedia):
1143         (WebCore::UserMediaController::logGetUserMediaDenial):
1144         * Modules/mediastream/UserMediaController.h:
1145         * Modules/mediastream/UserMediaRequest.cpp:
1146         (WebCore::UserMediaRequest::start):
1147
1148 2018-11-12  Simon Fraser  <simon.fraser@apple.com>
1149
1150         Make compositing updates incremental
1151         https://bugs.webkit.org/show_bug.cgi?id=90342
1152
1153         Reviewed by Antti Koivisto.
1154
1155         Previously, updating compositing layers required two full RenderLayer tree traversals,
1156         and all the work was done for every RenderLayer on each composting update. This could be expensive
1157         on pages with lots of RenderLayers.
1158
1159         These changes make compositing updates more incremental. Compositing updates still require
1160         two tree traversals. The first determines which RenderLayers need to be composited (of those which
1161         weren't already made composited at style-change time), because of reasons that can only be determined
1162         post-layout, and indirect reasons including overlap. The second traversal updates the configuration, geometry
1163         and GraphicsLayer tree for the composited layers. Dependencies on both descendant and ancestor state make
1164         it hard to fold these two traversals together.
1165
1166         In order to minimize the work done during these traversals, dirty bits are stored on RenderLayers,
1167         and propagated to ancestor layers in paint order. There are two sets of bits: those related to the first
1168         "compositing requirements" traversal, and those related to the second "update backing and hierarchy" traversal.
1169         When a RenderLayer gets a dirty bit set, bits are propagated to ancestors to indicate that children need
1170         to be visited.
1171
1172         Sadly entire subtrees can't be skipped during the "compositing requirements" traversal because we still have
1173         to accumulate overlap rects, but RenderLayerCompositor::traverseUnchangedSubtree() is used to minimize
1174         work in that case. Subtrees can be skipped in the "update backing and hierarchy" traversal. Entire traversals can
1175         be skipped if no change has triggered the need for that traversal.
1176
1177         These changes fix a correctness issue where transform changes now trigger overlap re-evaluation, which causes
1178         more layer geometry updates than before. This regressed the MotionMark "Focus" test, when geometry updates
1179         triggered layer resizes as the filter blur radius changed, which then triggered repaints. This is fixed by
1180         excluding composited filters from the composited bounds (but still taking them into account for overlap).
1181
1182         Care is taken to avoid triggering traversals in non-composited documents (tested by no-updates-in-non-composited-iframe.html).
1183
1184         Code to set the dirty bits is added in various places that change properties that compositing depends on.
1185
1186         These changes also subsume the patch in 176196; we now never consult properties that rely on layout from the
1187         style change code path, and the only call stack for geometry updates is from the "update backing and hierarchy"
1188         traversal, which is always a pre-order traversal.
1189
1190         Tests: compositing/geometry/stacking-context-change-layer-reparent.html
1191                compositing/layer-creation/change-to-overlap.html
1192                compositing/updates/no-updates-in-non-composited-iframe.html
1193
1194         * html/canvas/WebGLRenderingContextBase.cpp:
1195         (WebCore::WebGLRenderingContextBase::markContextChanged): Need to differentiate between a canvas becoming composited
1196         for the first time, and its pixels changing with a new 'CanvasPixelsChanged' value.
1197         * page/FrameView.cpp:
1198         (WebCore::FrameView::setViewportConstrainedObjectsNeedLayout):
1199         * page/Page.cpp:
1200         (WebCore::Page::setPageScaleFactor):
1201         * platform/graphics/ca/GraphicsLayerCA.cpp:
1202         (WebCore::GraphicsLayerCA::updateBackdropFilters): If we just made a layer for backdrops, we need to update sublayers.
1203         * rendering/RenderBox.cpp:
1204         (WebCore::RenderBox::styleWillChange):
1205         * rendering/RenderLayer.cpp:
1206         (WebCore::RenderLayer::RenderLayer):
1207         (WebCore::RenderLayer::~RenderLayer):
1208         (WebCore::RenderLayer::addChild):
1209         (WebCore::RenderLayer::removeChild):
1210         (WebCore::RenderLayer::shouldBeStackingContext const):
1211         (WebCore::RenderLayer::stackingContext const):
1212         (WebCore::RenderLayer::dirtyZOrderLists):
1213         (WebCore::RenderLayer::dirtyNormalFlowList):
1214         (WebCore::RenderLayer::updateNormalFlowList):
1215         (WebCore::RenderLayer::rebuildZOrderLists):
1216         (WebCore::RenderLayer::setAncestorsHaveCompositingDirtyFlag):
1217         (WebCore::RenderLayer::contentChanged):
1218         (WebCore::RenderLayer::updateLayerPositions):
1219         (WebCore::RenderLayer::updateTransform):
1220         (WebCore::RenderLayer::updateLayerPosition):
1221         (WebCore::RenderLayer::enclosingCompositingLayer const):
1222         (WebCore::RenderLayer::enclosingCompositingLayerForRepaint const):
1223         (WebCore::RenderLayer::clippingRootForPainting const):
1224         (WebCore::RenderLayer::scrollTo):
1225         (WebCore::RenderLayer::updateCompositingLayersAfterScroll):
1226         (WebCore::RenderLayer::updateScrollInfoAfterLayout):
1227         (WebCore::RenderLayer::paintLayerContents):
1228         (WebCore::RenderLayer::hitTest):
1229         (WebCore::RenderLayer::hitTestLayer):
1230         (WebCore::RenderLayer::calculateClipRects const):
1231         (WebCore::outputPaintOrderTreeLegend):
1232         (WebCore::outputPaintOrderTreeRecursive):
1233         (WebCore::compositingContainer): Deleted.
1234         * rendering/RenderLayer.h:
1235         (WebCore::RenderLayer::clearZOrderLists):
1236         (WebCore::RenderLayer::paintOrderParent const):
1237         * rendering/RenderLayerBacking.cpp:
1238         (WebCore::RenderLayerBacking::updateCompositedBounds):
1239         (WebCore::RenderLayerBacking::updateAfterWidgetResize):
1240         (WebCore::RenderLayerBacking::updateAfterLayout):
1241         (WebCore::RenderLayerBacking::updateConfigurationAfterStyleChange):
1242         (WebCore::RenderLayerBacking::updateConfiguration):
1243         (WebCore::RenderLayerBacking::updateGeometry):
1244         (WebCore::RenderLayerBacking::setRequiresBackgroundLayer):
1245         (WebCore::RenderLayerBacking::updateMaskingLayer):
1246         (WebCore::RenderLayerBacking::paintsContent const):
1247         (WebCore::RenderLayerBacking::contentChanged):
1248         (WebCore::RenderLayerBacking::setContentsNeedDisplay):
1249         (WebCore::RenderLayerBacking::setContentsNeedDisplayInRect):
1250         (WebCore::RenderLayerBacking::startAnimation):
1251         (WebCore::RenderLayerBacking::animationFinished):
1252         (WebCore::RenderLayerBacking::startTransition):
1253         (WebCore::RenderLayerBacking::transitionFinished):
1254         (WebCore::RenderLayerBacking::setCompositedBounds):
1255         * rendering/RenderLayerBacking.h:
1256         * rendering/RenderLayerCompositor.cpp:
1257         (WebCore::RenderLayerCompositor::CompositingState::CompositingState):
1258         (WebCore::RenderLayerCompositor::enableCompositingMode):
1259         (WebCore::RenderLayerCompositor::cacheAcceleratedCompositingFlags):
1260         (WebCore::RenderLayerCompositor::cacheAcceleratedCompositingFlagsAfterLayout):
1261         (WebCore::RenderLayerCompositor::willRecalcStyle):
1262         (WebCore::RenderLayerCompositor::didRecalcStyleWithNoPendingLayout):
1263         (WebCore::RenderLayerCompositor::updateCompositingLayers):
1264         (WebCore::RenderLayerCompositor::computeCompositingRequirements):
1265         (WebCore::RenderLayerCompositor::traverseUnchangedSubtree):
1266         (WebCore::RenderLayerCompositor::updateBackingAndHierarchy):
1267         (WebCore::RenderLayerCompositor::appendDocumentOverlayLayers):
1268         (WebCore::RenderLayerCompositor::layerBecameNonComposited):
1269         (WebCore::RenderLayerCompositor::logLayerInfo):
1270         (WebCore::clippingChanged):
1271         (WebCore::styleAffectsLayerGeometry):
1272         (WebCore::RenderLayerCompositor::layerStyleChanged):
1273         (WebCore::RenderLayerCompositor::needsCompositingUpdateForStyleChangeOnNonCompositedLayer const):
1274         (WebCore::RenderLayerCompositor::updateBacking):
1275         (WebCore::RenderLayerCompositor::updateLayerCompositingState):
1276         (WebCore::RenderLayerCompositor::layerWasAdded):
1277         (WebCore::RenderLayerCompositor::layerWillBeRemoved):
1278         (WebCore::RenderLayerCompositor::enclosingNonStackingClippingLayer const):
1279         (WebCore::RenderLayerCompositor::computeExtent const):
1280         (WebCore::RenderLayerCompositor::addToOverlapMap):
1281         (WebCore::RenderLayerCompositor::addToOverlapMapRecursive):
1282         (WebCore::RenderLayerCompositor::rootLayerConfigurationChanged):
1283         (WebCore::RenderLayerCompositor::parentFrameContentLayers):
1284         (WebCore::RenderLayerCompositor::updateRootLayerPosition):
1285         (WebCore::RenderLayerCompositor::needsToBeComposited const):
1286         (WebCore::RenderLayerCompositor::requiresCompositingLayer const):
1287         (WebCore::RenderLayerCompositor::requiresOwnBackingStore const):
1288         (WebCore::RenderLayerCompositor::reasonsForCompositing const):
1289         (WebCore::RenderLayerCompositor::clippedByAncestor const):
1290         (WebCore::RenderLayerCompositor::requiresCompositingForAnimation const):
1291         (WebCore::RenderLayerCompositor::requiresCompositingForTransform const):
1292         (WebCore::RenderLayerCompositor::requiresCompositingForVideo const):
1293         (WebCore::RenderLayerCompositor::requiresCompositingForFilters const):
1294         (WebCore::RenderLayerCompositor::requiresCompositingForWillChange const):
1295         (WebCore::RenderLayerCompositor::requiresCompositingForPlugin const):
1296         (WebCore::RenderLayerCompositor::requiresCompositingForFrame const):
1297         (WebCore::RenderLayerCompositor::requiresCompositingForScrollableFrame const):
1298         (WebCore::RenderLayerCompositor::requiresCompositingForPosition const):
1299         (WebCore::RenderLayerCompositor::requiresCompositingForOverflowScrolling const):
1300         (WebCore::RenderLayerCompositor::styleChangeMayAffectIndirectCompositingReasons):
1301         (WebCore::RenderLayerCompositor::fixedLayerIntersectsViewport const):
1302         (WebCore::RenderLayerCompositor::useCoordinatedScrollingForLayer const):
1303         (WebCore::RenderLayerCompositor::rootOrBodyStyleChanged):
1304         (WebCore::RenderLayerCompositor::rootBackgroundColorOrTransparencyChanged):
1305         (WebCore::operator<<):
1306         (WebCore::RenderLayerCompositor::setCompositingLayersNeedRebuild): Deleted.
1307         (WebCore::checkIfDescendantClippingContextNeedsUpdate): Deleted.
1308         (WebCore::isScrollableOverflow): Deleted.
1309         (WebCore::styleHasTouchScrolling): Deleted.
1310         (WebCore::styleChangeRequiresLayerRebuild): Deleted.
1311         (WebCore::RenderLayerCompositor::rebuildCompositingLayerTree): Deleted.
1312         (WebCore::RenderLayerCompositor::rootFixedBackgroundsChanged): Deleted.
1313         (WebCore::RenderLayerCompositor::updateLayerTreeGeometry): Deleted.
1314         (WebCore::RenderLayerCompositor::updateCompositingDescendantGeometry): Deleted.
1315         * rendering/RenderLayerCompositor.h:
1316         * rendering/RenderTreeAsText.cpp:
1317         (WebCore::writeLayers):
1318
1319 2018-11-12  Rob Buis  <rbuis@igalia.com>
1320
1321         CSSCalcOperation constructor wastes 6KB of Vector capacity on cnn.com
1322         https://bugs.webkit.org/show_bug.cgi?id=190839
1323
1324         Reviewed by Frédéric Wang.
1325
1326         The CSSCalcOperation ctor that takes a leftSide and rightSide parameter
1327         wastes memory since it will always have size 2 but claims the
1328         default Vector size. So make sure to reserve an initial capacity of 2.
1329
1330         * css/CSSCalculationValue.cpp:
1331
1332 2018-11-12  Yusuke Suzuki  <yusukesuzuki@slowstart.org>
1333
1334         WTFMove(xxx) is used in arguments while other arguments touch xxx
1335         https://bugs.webkit.org/show_bug.cgi?id=191544
1336
1337         Reviewed by Alex Christensen.
1338
1339         The order of the evaluation of C++ arguments is undefined. If we use WTFMove(xxx),
1340         xxx should not be touched in the other arguments. This patch fixes such uses in
1341         IDB code.
1342
1343         * Modules/indexeddb/IDBObjectStore.cpp:
1344         (WebCore::IDBObjectStore::deleteIndex):
1345         * Modules/indexeddb/IDBTransaction.cpp:
1346         (WebCore::IDBTransaction::scheduleOperation):
1347         * Modules/indexeddb/server/MemoryIDBBackingStore.cpp:
1348         (WebCore::IDBServer::MemoryIDBBackingStore::registerObjectStore):
1349         * Modules/indexeddb/server/MemoryObjectStore.cpp:
1350         (WebCore::IDBServer::MemoryObjectStore::registerIndex):
1351
1352 2018-11-12  Zalan Bujtas  <zalan@apple.com>
1353
1354         [LFC][IFC] Construct dedicated runs when the inline element requires it.
1355         https://bugs.webkit.org/show_bug.cgi?id=191509
1356
1357         Reviewed by Antti Koivisto.
1358
1359         In certain cases, a run can overlap multiple inline elements like this:
1360
1361         <span>normal text content</span><span style="position: relative; left: 10px;">but this one needs a dedicated run</span><span>end of text</span>
1362
1363         The content above generates one long run <normal text contentbut this one needs dedicated runend of text> <- input to line breaking.
1364         However, since the middle run is positioned, it needs to be moved independently from the rest of the content, hence it needs a dedicated inline run.
1365
1366         * layout/inlineformatting/InlineFormattingContext.cpp:
1367         (WebCore::Layout::InlineFormattingContext::layout const):
1368         (WebCore::Layout::contentRequiresSeparateRun):
1369         (WebCore::Layout::InlineFormattingContext::splitInlineRunIfNeeded const):
1370         (WebCore::Layout::InlineFormattingContext::postProcessInlineRuns const):
1371         (WebCore::Layout::InlineFormattingContext::closeLine const):
1372         (WebCore::Layout::InlineFormattingContext::appendContentToLine const):
1373         (WebCore::Layout::InlineFormattingContext::layoutInlineContent const):
1374         (WebCore::Layout::InlineFormattingContext::instrinsicWidthConstraints const):
1375         * layout/inlineformatting/InlineFormattingContext.h:
1376         (WebCore::Layout::InlineFormattingContext::inlineFormattingState const):
1377         * layout/inlineformatting/InlineLineBreaker.cpp:
1378         (WebCore::Layout::InlineLineBreaker::nextRun): mid-word breaking is not implemented yet.
1379         * layout/inlineformatting/InlineRun.h:
1380         (WebCore::Layout::InlineRun::overlapsMultipleInlineItems const):
1381         * layout/inlineformatting/InlineRunProvider.cpp:
1382         (WebCore::Layout::InlineRunProvider::processInlineTextItem):
1383         * layout/inlineformatting/InlineRunProvider.h:
1384         (WebCore::Layout::InlineRunProvider::Run::TextContext::expand):
1385         (WebCore::Layout::InlineRunProvider::Run::textContext):
1386         (WebCore::Layout::InlineRunProvider::Run::TextContext::setStart): Deleted.
1387         (WebCore::Layout::InlineRunProvider::Run::TextContext::setLength): Deleted.
1388
1389 2018-11-12  Jer Noble  <jer.noble@apple.com>
1390
1391         [MSE] Frame re-ordering can cause iframes to never be enqueued
1392         https://bugs.webkit.org/show_bug.cgi?id=191485
1393
1394         Reviewed by Eric Carlson.
1395
1396         Test: media/media-source/media-source-dropped-iframe.html
1397
1398         Some frame re-ordering techniques result in files where the first frame has a
1399         decode timestamp < 0, but a presentation timestamp >= 0. When appending these
1400         samples to existing content, we can fail to enqueue the first frame because its
1401         DTS overlaps an existing sample, but the presentation timestamp does not.
1402         Rather than try to only enqueue samples whose decode timestamps are > than the
1403         greatest decode end timestamp (minus some fudge factor), allow all frames to be
1404         added to the decode queue if they are strictly ordered greater than the last
1405         enqueued frame.
1406
1407         * Modules/mediasource/SourceBuffer.cpp:
1408         (WebCore::SourceBuffer::TrackBuffer::TrackBuffer):
1409         (WebCore::SourceBuffer::sourceBufferPrivateDidReceiveSample):
1410         (WebCore::SourceBuffer::provideMediaData):
1411         (WebCore::SourceBuffer::reenqueueMediaForTime):
1412
1413 2018-11-12  Yusuke Suzuki  <yusukesuzuki@slowstart.org>
1414
1415         IDBTransaction does not use "RefPtr<IDBTransaction> self"
1416         https://bugs.webkit.org/show_bug.cgi?id=190436
1417
1418         Reviewed by Alex Christensen.
1419
1420         It seems that `RefPtr<IDBTransaction> self;` is not effective since it does not capture anything.
1421         Use `protectedThis = makeRef(*this)` instead.
1422
1423         No behavior change.
1424
1425         * Modules/indexeddb/IDBTransaction.cpp:
1426         (WebCore::IDBTransaction::IDBTransaction):
1427
1428 2018-11-12  Alicia Boya García  <aboya@igalia.com>
1429
1430         [MSE][GStreamer] Introduce AbortableTaskQueue
1431         https://bugs.webkit.org/show_bug.cgi?id=190902
1432
1433         Reviewed by Xabier Rodriguez-Calvar.
1434
1435         A new synchronization primitive is introduced: AbortableTaskQueue,
1436         which allows to send work to the main thread from a background thread
1437         with the option to perform two-phase cancellation (startAborting() and
1438         finishAborting()).
1439
1440         This new primitive has been used to overhaul GstBus messaging in
1441         AppendPipeline. A lot of code made redundant has been deleted in the
1442         process and lots of internal functions were now able to be made
1443         private. As part of the refactor all glib signals in AppendPipeline
1444         now use lambdas. All usages of WTF::isMainThread() in AppendPipeline
1445         have been replaced by isMainThread() for consistency with the rest of
1446         WebKit.
1447
1448         Two-phase cancellation is still not used in AppendPipeline as of this
1449         patch, but it will be used in a future patch that makes use of
1450         GStreamer flushes to implement correct MSE abort semantics. There are
1451         unit tests to ensure it works correctly, even if it's still not used.
1452
1453         * WebCore.xcodeproj/project.pbxproj:
1454         * platform/AbortableTaskQueue.h: Added.
1455         * platform/graphics/gstreamer/mse/AppendPipeline.cpp:
1456         (WebCore::AppendPipeline::dumpAppendState):
1457         (WebCore::AppendPipeline::AppendPipeline):
1458         (WebCore::AppendPipeline::~AppendPipeline):
1459         (WebCore::AppendPipeline::appsrcEndOfAppendCheckerProbe):
1460         (WebCore::AppendPipeline::handleAppsinkNewSampleFromAnyThread):
1461         (WebCore::AppendPipeline::connectDemuxerSrcPadToAppsinkFromAnyThread):
1462         * platform/graphics/gstreamer/mse/AppendPipeline.h:
1463         (WebCore::AppendPipeline::sourceBufferPrivate):
1464         (WebCore::AppendPipeline::appsinkCaps):
1465         (WebCore::AppendPipeline::track):
1466         (WebCore::AppendPipeline::demuxerSrcPadCaps):
1467         (WebCore::AppendPipeline::playerPrivate):
1468
1469 2018-11-12  Xabier Rodriguez Calvar  <calvaris@igalia.com>
1470
1471         [GStreamer][EME] waitingforkey event should consider decryptors' waiting status
1472         https://bugs.webkit.org/show_bug.cgi?id=191459
1473
1474         Reviewed by Carlos Garcia Campos.
1475
1476         The new cross platform architecture to report waitingforkey and
1477         recover from it requires a more accurate knowledge of what is
1478         going on with the decryptors because events are reported only once
1479         (per key exchange run) and crossplatform only continues if we are
1480         actually ready to continue, meaning that no decryptors are
1481         waiting.
1482
1483         * platform/graphics/gstreamer/GUniquePtrGStreamer.h: Added
1484         GstIterator deleter.
1485         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
1486         (WebCore::MediaPlayerPrivateGStreamerBase::setWaitingForKey): Bail
1487         out if we are requested to not wait anymore but there are still
1488         waiting decryptors.
1489         (WebCore::MediaPlayerPrivateGStreamerBase::waitingForKey const):
1490         Query the pipeline, just a query after pipeline is built and
1491         manual inspection during build. The query is optimal but sometimes
1492         we can get this request when the pipeline is under construction so
1493         queries do not arrive at the decryptors and we have to deliver it
1494         by ourselves.
1495         (WebCore::MediaPlayerPrivateGStreamerBase::reportWaitingForKey:
1496         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.h:
1497         (WebCore::MediaPlayerPrivateGStreamerBase::reportWaitingForKey:
1498         Deleted because it is now inlined.
1499         * platform/graphics/gstreamer/eme/WebKitClearKeyDecryptorGStreamer.cpp:
1500         (webKitMediaClearKeyDecryptorDecrypt): Fixed small compiler warning.
1501         * platform/graphics/gstreamer/eme/WebKitCommonEncryptionDecryptorGStreamer.cpp:
1502         (webkit_media_common_encryption_decrypt_class_init): Override
1503         query method.
1504         (webkitMediaCommonEncryptionDecryptTransformInPlace): When the
1505         decryptor is going to block to wait, report before. When the
1506         decryptor receives the key, report it got it.
1507         (webkitMediaCommonEncryptionDecryptSinkEventHandler): Do not
1508         handle waitingforkey here.
1509         (webkitMediaCommonEncryptionDecryptorQueryHandler): Report if the
1510         decryptor is waiting.
1511
1512 2018-11-12  Michael Catanzaro  <mcatanzaro@igalia.com>
1513
1514         [GTK] Silence ATK_XY_PARENT warnings
1515         https://bugs.webkit.org/show_bug.cgi?id=191504
1516
1517         Reviewed by Carlos Garcia Campos.
1518
1519         * accessibility/atk/WebKitAccessibleInterfaceComponent.cpp:
1520         (atkToContents):
1521         * accessibility/atk/WebKitAccessibleInterfaceText.cpp:
1522         (textExtents):
1523         * accessibility/atk/WebKitAccessibleUtil.cpp:
1524         (contentsRelativeToAtkCoordinateType):
1525
1526 2018-11-11  Wenson Hsieh  <wenson_hsieh@apple.com>
1527
1528         Implement a new edit command to change the enclosing list type
1529         https://bugs.webkit.org/show_bug.cgi?id=191487
1530         <rdar://problem/45955922>
1531
1532         Reviewed by Ryosuke Niwa.
1533
1534         Add support for a new edit command that changes the type of the enclosing list element around the selection from
1535         unordered to ordered list and vice versa. This new edit command is exposed only to internal WebKit2 clients, via
1536         SPI on WKWebView (`-_changeListType:`).
1537
1538         This is currently intended for use in Mail compose, but may also be adopted by legacy Notes in the future. As
1539         such, the behavior of this editing command mostly matches shipping behavior in Mail compose (which is currently
1540         implemented entirely by Mail). See below for more details.
1541
1542         Test:   editing/execCommand/change-list-type.html
1543                 WKWebViewEditActions.ChangeListType
1544
1545         * Sources.txt:
1546         * WebCore.xcodeproj/project.pbxproj:
1547         * editing/ChangeListTypeCommand.cpp: Added.
1548         (WebCore::listConversionTypeForSelection):
1549         (WebCore::ChangeListTypeCommand::listConversionType):
1550
1551         Helper that returns a potential list conversion command that may be executed at the given document's selection,
1552         if any exists. We also use existing logic from Mail here to determine which list to change, by walking up the
1553         DOM from the lowest common ancestor container of the current selection until we hit the first list element.
1554
1555         (WebCore::ChangeListTypeCommand::createNewList):
1556
1557         Helper method to create a new list element to replace the given list, and then clone element data from the given
1558         list to the new list. This addresses an existing bug in Mail, wherein changing list type for an enclosing list
1559         which contains inline style properties drops the inline styles, because existing logic in Mail that implements
1560         this editing command only copies the `class` attribute of the old list to the new list.
1561
1562         (WebCore::ChangeListTypeCommand::doApply):
1563
1564         Apply the edit command by running the following steps:
1565         -   Find the enclosing list element, if any (see above).
1566         -   Create a new list element of the opposite type as the enclosing list, and clone over element data from the
1567             list element being replaced.
1568         -   Insert the new list next to the original list.
1569         -   Move all children of the original list to the new list.
1570         -   Remove the original list.
1571         -   Set the selection to the end of the new list.
1572
1573         * editing/ChangeListTypeCommand.h: Added.
1574         * editing/EditAction.h:
1575
1576         Add a pair of new edit actions for conversion from unordered list to ordered list and vice versa.
1577
1578         * editing/Editor.cpp:
1579         (WebCore::Editor::changeSelectionListType):
1580
1581         Implement this by creating and applying a new ChangeListTypeCommand.
1582
1583         (WebCore::Editor::canChangeSelectionListType): Deleted.
1584
1585         Remove this for now, since there's no need for it until full support for edit command validation is implemented.
1586
1587         * editing/Editor.h:
1588         * testing/Internals.cpp:
1589         (WebCore::Internals::changeSelectionListType):
1590         * testing/Internals.h:
1591         * testing/Internals.idl:
1592
1593         Add internal hooks to change list type from layout tests.
1594
1595 2018-11-11  Daniel Bates  <dabates@apple.com>
1596
1597         [iOS] Draw caps lock indicator in password fields
1598         https://bugs.webkit.org/show_bug.cgi?id=190565
1599         <rdar://problem/45262343>
1600
1601         Reviewed by Dean Jackson.
1602
1603         Draw the caps lock indicator in a focused password field on iOS. This makes the behavior of password
1604         fields on iOS more closely match the behavior of password fields on Mac.
1605
1606         The majority of this patch is implementing PlatformKeyboardEvent::currentCapsLockState() for iOS.
1607         In Legacy WebKit, the implementation boils down to calling call -[::WebEvent modifierFlags]. In
1608         Modern WebKit the UIProcess is responsible for -[::WebEvent modifierFlags] and passing it the
1609         WebProcess to store such that invocations of PlatformKeyboardEvent::currentCapsLockState() consult
1610         the store in the WebProcess. A smaller part of this patch is having both the legacy and modern
1611         web views listen for keyboard availability changes so as to update the the caps lock state when
1612         a hardware keyboard is detached or attached.
1613
1614         * WebCore.xcodeproj/project.pbxproj:
1615         * page/EventHandler.cpp:
1616         (WebCore::EventHandler::capsLockStateMayHaveChanged const): Extracted from EventHandler::internalKeyEvent()
1617         so that it can shared between WebCore, Modern WebKit, and Legacy WebKit code.
1618         (WebCore::EventHandler::internalKeyEvent): Modified to call capsLockStateMayHaveChanged().
1619         * page/EventHandler.h:
1620         * platform/cocoa/KeyEventCocoa.mm:
1621         (WebCore::PlatformKeyboardEvent::currentCapsLockState): Moved from KeyEventMac.mm.
1622         (WebCore::PlatformKeyboardEvent::getCurrentModifierState): Moved from KeyEventMac.mm.
1623         * platform/ios/KeyEventIOS.mm:
1624         (WebCore::PlatformKeyboardEvent::currentStateOfModifierKeys): Fetch the current modifier state.
1625         (WebCore::PlatformKeyboardEvent::currentCapsLockState): Deleted; we now use the Cocoa implementation.
1626         (WebCore::PlatformKeyboardEvent::getCurrentModifierState): Deleted; we now use the Cocoa implementation.
1627         * platform/ios/WebEvent.h:
1628         * platform/ios/WebEvent.mm:
1629         (+[WebEvent modifierFlags]): Added.
1630         * platform/mac/KeyEventMac.mm:
1631         (WebCore::PlatformKeyboardEvent::currentCapsLockState): Deleted; moved to KeyEventCocoa.mm to be shared
1632         by both Mac and iOS.
1633         (WebCore::PlatformKeyboardEvent::getCurrentModifierState): Deleted; moved to KeyEventCocoa.mm to be shared
1634         by both Mac and iOS.
1635         * rendering/RenderThemeCocoa.h:
1636         * rendering/RenderThemeCocoa.mm:
1637         (WebCore::RenderThemeCocoa::shouldHaveCapsLockIndicator const): Moved from RenderThemeMac.mm.
1638         * rendering/RenderThemeIOS.h:
1639         * rendering/RenderThemeIOS.mm:
1640         (WebCore::RenderThemeIOS::shouldHaveCapsLockIndicator const): Deleted.
1641         * rendering/RenderThemeMac.h:
1642         * rendering/RenderThemeMac.mm:
1643         (WebCore::RenderThemeMac::shouldHaveCapsLockIndicator const): Deleted; moved to RenderThemeCocoa.mm to be
1644         shared by both Mac and iOS.
1645
1646 2018-11-11  Zalan Bujtas  <zalan@apple.com>
1647
1648         [LFC][BFC] In-flow positioned logic is really formatting context dependent.
1649         https://bugs.webkit.org/show_bug.cgi?id=191512
1650
1651         Reviewed by Simon Fraser.
1652
1653         Move block formatting context specific code from FormattingContext to BlockFormattingContext.
1654
1655         * layout/FormattingContext.cpp:
1656         (WebCore::Layout::FormattingContext::placeInFlowPositionedChildren const): Deleted.
1657         * layout/FormattingContext.h:
1658         * layout/FormattingContextGeometry.cpp:
1659         (WebCore::Layout::FormattingContext::Geometry::inFlowPositionedPositionOffset):
1660         * layout/blockformatting/BlockFormattingContext.cpp:
1661         (WebCore::Layout::BlockFormattingContext::placeInFlowPositionedChildren const):
1662         (WebCore::Layout::BlockFormattingContext::computeInFlowPositionedPosition const): Deleted.
1663         * layout/blockformatting/BlockFormattingContext.h:
1664         * layout/blockformatting/BlockFormattingContextGeometry.cpp:
1665         (WebCore::Layout::BlockFormattingContext::Geometry::inFlowPositionedPosition): Deleted.
1666         * layout/inlineformatting/InlineFormattingContext.cpp:
1667         (WebCore::Layout::InlineFormattingContext::computeInFlowPositionedPosition const): Deleted.
1668         * layout/inlineformatting/InlineFormattingContext.h:
1669
1670 2018-11-11  Myles C. Maxfield  <mmaxfield@apple.com>
1671
1672         Address post-review comments after r237955
1673         https://bugs.webkit.org/show_bug.cgi?id=191496
1674
1675         Reviewed by Darin Adler.
1676
1677         * rendering/TextDecorationPainter.cpp:
1678         (WebCore::TextDecorationPainter::paintTextDecoration):
1679         * style/InlineTextBoxStyle.cpp:
1680         (WebCore::computeUnderlineOffset):
1681         * style/InlineTextBoxStyle.h:
1682
1683 2018-11-11  Benjamin Poulain  <benjamin@webkit.org>
1684
1685         Fix a fixme: rename wtfObjcMsgSend to wtfObjCMsgSend
1686         https://bugs.webkit.org/show_bug.cgi?id=191492
1687
1688         Reviewed by Alex Christensen.
1689
1690         Rename file.
1691
1692         * accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
1693         * page/mac/EventHandlerMac.mm:
1694         * platform/mac/URLMac.mm:
1695         * platform/mac/WebCoreNSURLExtras.mm:
1696         * platform/mac/WebCoreObjCExtras.mm:
1697         * rendering/RenderThemeMac.mm:
1698
1699 2018-11-10  Benjamin Poulain  <benjamin@webkit.org>
1700
1701         Fix a fixme: rename wtfObjcMsgSend to wtfObjCMsgSend
1702         https://bugs.webkit.org/show_bug.cgi?id=191492
1703
1704         Reviewed by Alex Christensen.
1705
1706         * accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
1707         * page/mac/EventHandlerMac.mm:
1708         * platform/mac/URLMac.mm:
1709         * platform/mac/WebCoreNSURLExtras.mm:
1710         * platform/mac/WebCoreObjCExtras.mm:
1711         * rendering/RenderThemeMac.mm:
1712         (WebCore::RenderThemeMac::systemColor const):
1713
1714 2018-11-10  Megan Gardner  <megan_gardner@apple.com>
1715
1716         Fix build for 32bit Mac
1717         https://bugs.webkit.org/show_bug.cgi?id=191511
1718
1719         Unreviewed Build Fix.
1720
1721         Build fix, not tests needed.
1722
1723         Make the apporiate delecrations for 32-bit mac support.
1724
1725         * editing/mac/DictionaryLookup.mm:
1726
1727 2018-11-10  Simon Fraser  <simon.fraser@apple.com>
1728
1729         Remove support for -webkit-svg-shadow
1730         https://bugs.webkit.org/show_bug.cgi?id=187429
1731         <rdar://problem/41920735>
1732
1733         Reviewed by Dean Jackson.
1734
1735         -webkit-svg-shadow was a non-standard hack for online iWork, and they no longer use it,
1736         so remove it. No other browser supports it, and chromestatus say it's used on less than
1737         0.000001% of pages.
1738
1739         * css/CSSComputedStyleDeclaration.cpp:
1740         (WebCore::ComputedStyleExtractor::valueForPropertyinStyle):
1741         * css/CSSProperties.json:
1742         * css/SVGCSSComputedStyleDeclaration.cpp:
1743         (WebCore::ComputedStyleExtractor::svgPropertyValue):
1744         * css/StyleBuilderCustom.h:
1745         (WebCore::StyleBuilderCustom::applyInitialWebkitSvgShadow): Deleted.
1746         (WebCore::StyleBuilderCustom::applyInheritWebkitSvgShadow): Deleted.
1747         (WebCore::StyleBuilderCustom::applyValueWebkitSvgShadow): Deleted.
1748         * css/parser/CSSPropertyParser.cpp:
1749         (WebCore::CSSPropertyParser::parseSingleValue):
1750         * rendering/RenderElement.cpp:
1751         (WebCore::RenderElement::didAttachChild):
1752         * rendering/svg/RenderSVGImage.cpp:
1753         (WebCore::RenderSVGImage::layout):
1754         * rendering/svg/RenderSVGImage.h:
1755         * rendering/svg/RenderSVGModelObject.cpp:
1756         (WebCore::RenderSVGModelObject::RenderSVGModelObject):
1757         * rendering/svg/RenderSVGModelObject.h:
1758         (WebCore::RenderSVGModelObject::repaintRectInLocalCoordinatesExcludingSVGShadow const): Deleted.
1759         (WebCore::RenderSVGModelObject::hasSVGShadow const): Deleted.
1760         (WebCore::RenderSVGModelObject::setHasSVGShadow): Deleted.
1761         * rendering/svg/RenderSVGRoot.cpp:
1762         (WebCore::RenderSVGRoot::RenderSVGRoot):
1763         (WebCore::RenderSVGRoot::updateCachedBoundaries):
1764         * rendering/svg/RenderSVGRoot.h:
1765         * rendering/svg/RenderSVGShape.cpp:
1766         (WebCore::RenderSVGShape::updateRepaintBoundingBox):
1767         * rendering/svg/RenderSVGShape.h:
1768         * rendering/svg/SVGRenderSupport.cpp:
1769         (WebCore::SVGRenderSupport::clippedOverflowRectForRepaint):
1770         (WebCore::SVGRenderSupport::layoutChildren):
1771         (WebCore::SVGRenderSupport::styleChanged):
1772         (WebCore::SVGRenderSupport::repaintRectForRendererInLocalCoordinatesExcludingSVGShadow): Deleted.
1773         (WebCore::SVGRenderSupport::rendererHasSVGShadow): Deleted.
1774         (WebCore::SVGRenderSupport::setRendererHasSVGShadow): Deleted.
1775         (WebCore::SVGRenderSupport::intersectRepaintRectWithShadows): Deleted.
1776         (WebCore::SVGRenderSupport::childAdded): Deleted.
1777         * rendering/svg/SVGRenderSupport.h:
1778
1779 2018-11-10  Ryan Haddad  <ryanhaddad@apple.com>
1780
1781         Unreviewed, rolling out r238065.
1782
1783         Breaks internal builds.
1784
1785         Reverted changeset:
1786
1787         "Make it possible to edit images inline"
1788         https://bugs.webkit.org/show_bug.cgi?id=191352
1789         https://trac.webkit.org/changeset/238065
1790
1791 2018-11-10  Tim Horton  <timothy_horton@apple.com>
1792
1793         Make it possible to edit images inline
1794         https://bugs.webkit.org/show_bug.cgi?id=191352
1795         <rdar://problem/30107985>
1796
1797         Reviewed by Dean Jackson.
1798
1799         Tests: editing/images/basic-editable-image.html
1800                editing/images/reparent-editable-image-maintains-strokes.html
1801
1802         Add the beginnings of a mechanism to replace images with a special attribute
1803         with a native drawing view in the UI process.
1804
1805         * page/Settings.yaml:
1806         Add a setting to control whether images become natively editable when they
1807         have the x-apple-editable-image attribute.
1808
1809         * html/HTMLImageElement.cpp:
1810         (WebCore::HTMLImageElement::editableImageViewID const):
1811         Lazily generate an EmbeddedViewID and persist it on the <img> element.
1812
1813         * html/HTMLImageElement.h:
1814         Rearrange the service controls methods to sit before the members.
1815         Add m_editableImageViewID and editableImageViewID().
1816
1817         * platform/graphics/GraphicsLayer.cpp:
1818         (WebCore::GraphicsLayer::nextEmbeddedViewID):
1819         * platform/graphics/GraphicsLayer.h:
1820         (WebCore::GraphicsLayer::setContentsToEmbeddedView):
1821         Add a new ContentsLayerPurpose, EmbeddedView, which is only supported
1822         on Cocoa platforms and when using RemoteLayerTree.
1823         Add ContentsLayerEmbeddedViewType, which currently only has the EditableImage type.
1824         Add setContentsToEmbeddedView, which takes a ContentsLayerEmbeddedViewType
1825         and an EmbeddedViewID to uniquely identify and communicate about the
1826         embedded view (which may move between layers, since it is tied to an element).
1827
1828         * platform/graphics/ca/GraphicsLayerCA.cpp:
1829         (WebCore::GraphicsLayerCA::createPlatformCALayerForEmbeddedView):
1830         (WebCore::GraphicsLayerCA::setContentsToEmbeddedView):
1831         When setting GraphicsLayer's contents to an embedded view, we use
1832         a special PlatformCALayer factory that takes the EmbeddedViewID and type.
1833         GraphicsLayerCARemote will override this and make a correctly-initialized
1834         PlatformCALayerRemote that keeps track of the EmbeddedViewID.
1835
1836         * platform/graphics/ca/GraphicsLayerCA.h:
1837         * platform/graphics/ca/PlatformCALayer.cpp:
1838         (WebCore::operator<<):
1839         * platform/graphics/ca/PlatformCALayer.h:
1840         * platform/graphics/ca/cocoa/PlatformCALayerCocoa.h:
1841         * platform/graphics/ca/cocoa/PlatformCALayerCocoa.mm:
1842         (WebCore::PlatformCALayerCocoa::PlatformCALayerCocoa):
1843         (WebCore::PlatformCALayerCocoa::embeddedViewID const):
1844         Add stubs and logging for EmbeddedViewID on PlatformCALayer.
1845         These will be overridden by PlatformCALayerRemote to do more interesting things.
1846
1847         * rendering/RenderImage.cpp:
1848         (WebCore::RenderImage::isEditableImage const):
1849         Add a getter that return true if the setting is enabled and
1850         x-apple-editable-image is empty or true.
1851
1852         (WebCore::RenderImage::requiresLayer const):
1853         RenderImage requires a layer either if RenderReplaced does, or we are an
1854         editable image.
1855
1856         * rendering/RenderImage.h:
1857         * rendering/RenderLayer.cpp:
1858         (WebCore::RenderLayer::shouldBeNormalFlowOnly const):
1859         (WebCore::RenderLayer::calculateClipRects const):
1860         * rendering/RenderLayerBacking.cpp:
1861         (WebCore::RenderLayerBacking::updateConfiguration):
1862         Push the EmbeddedViewID and type down to GraphicsLayer for editable images.
1863
1864         * rendering/RenderLayerCompositor.cpp:
1865         (WebCore::RenderLayerCompositor::requiresCompositingLayer const):
1866         (WebCore::RenderLayerCompositor::requiresOwnBackingStore const):
1867         (WebCore::RenderLayerCompositor::reasonsForCompositing const):
1868         (WebCore::RenderLayerCompositor::requiresCompositingForEditableImage const):
1869         * rendering/RenderLayerCompositor.h:
1870         Make editable images require compositing implicitly.
1871
1872 2018-11-09  Zalan Bujtas  <zalan@apple.com>
1873
1874         [iOS] Issue initial paint soon after the visuallyNonEmpty milestone is fired.
1875         https://bugs.webkit.org/show_bug.cgi?id=191078
1876         <rdar://problem/45736178>
1877
1878         Reviewed by Antti Koivisto.
1879
1880         1. Improve visuallyNonEmpty milestone confidence level.
1881             Ignore whitespace and non visible text content.
1882             Parsing the main document should not necessarily fire the milestone. Check if there's any pending scripts/css/font loading.
1883             Check if the html/body is actually visible.
1884
1885         2. Issue initial paint soon after the milestone fires.
1886             Use a 0ms timer to flush the initial paint.
1887             Throttle additional flushes for 500ms and 1.5s (original behaviour).
1888
1889         3. Suspend optional style recalcs and layouts while painting is being throttled.
1890            When parsing yields we initiate a 0ms style recalc/layout timer.
1891            These optional layouts produce content that we have no intention to paint.
1892
1893         * dom/Document.cpp:
1894         (WebCore::Document::scheduleStyleRecalc):
1895         (WebCore::Document::shouldScheduleLayout):
1896         * page/ChromeClient.h:
1897         * page/FrameView.cpp:
1898         (WebCore::FrameView::resetLayoutMilestones):
1899         (WebCore::FrameView::qualifiesAsVisuallyNonEmpty const):
1900         (WebCore::FrameView::updateSignificantRenderedTextMilestoneIfNeeded):
1901         (WebCore::FrameView::updateIsVisuallyNonEmpty):
1902         * page/FrameView.h:
1903         (WebCore::FrameView::incrementVisuallyNonEmptyCharacterCount): Ignore whitespace characters. Some pages start with plenty of whitespace only content.
1904         * platform/graphics/FontCascade.h:
1905         * rendering/RenderText.cpp: Check whether the text is actually visible at this point.
1906         (WebCore::RenderText::RenderText):
1907
1908 2018-11-09  John Wilander  <wilander@apple.com>
1909
1910         Add ability to configure document.cookie lifetime cap through user defaults
1911         https://bugs.webkit.org/show_bug.cgi?id=191480
1912         <rdar://problem/45240871>
1913
1914         Reviewed by Chris Dumez.
1915
1916         No new tests. Existing test makes sure we don't regress.
1917
1918         This change makes the capped lifetime in seconds configurable through
1919         user defaults.
1920
1921         * platform/network/NetworkStorageSession.h:
1922         * platform/network/cf/NetworkStorageSessionCFNet.cpp:
1923         (WebCore::NetworkStorageSession::setAgeCapForClientSideCookies):
1924         (WebCore::NetworkStorageSession::setShouldCapLifetimeForClientSideCookies): Deleted.
1925             Renamed setAgeCapForClientSideCookies().
1926         * platform/network/cocoa/NetworkStorageSessionCocoa.mm:
1927         (WebCore::filterCookies):
1928         (WebCore::NetworkStorageSession::setCookiesFromDOM const):
1929
1930 2018-11-09  Ryan Haddad  <ryanhaddad@apple.com>
1931
1932         Unreviewed, rolling out r238047.
1933
1934         Introduced layout test failures on iOS simulator.
1935
1936         Reverted changeset:
1937
1938         "[iOS] Draw caps lock indicator in password fields"
1939         https://bugs.webkit.org/show_bug.cgi?id=190565
1940         https://trac.webkit.org/changeset/238047
1941
1942 2018-11-09  Tim Horton  <timothy_horton@apple.com>
1943
1944         Normal-flow-only flex items don't correctly respect z-index
1945         https://bugs.webkit.org/show_bug.cgi?id=191486
1946
1947         Reviewed by Simon Fraser.
1948
1949         Test: css3/flexbox/z-index-with-normal-flow-only.html
1950
1951         * rendering/RenderLayer.cpp:
1952         (WebCore::canCreateStackingContext):
1953         r125693 did not ensure that flex items which would otherwise be
1954         normal-flow-only would be put into the z-order tree when necessary.
1955         Fix by respecting the same trigger we use to make layers for flex items;
1956         namely, not having auto z-index.
1957
1958 2018-11-09  Wenson Hsieh  <wenson_hsieh@apple.com>
1959
1960         [Cocoa] Implement SPI on WKWebView to increase and decrease list levels
1961         https://bugs.webkit.org/show_bug.cgi?id=191471
1962         <rdar://problem/45952472>
1963
1964         Reviewed by Tim Horton.
1965
1966         Add new method stubs for changing the list type for the current selection (to be implemented in a future patch).
1967
1968         * editing/Editor.cpp:
1969         (WebCore::Editor::canChangeSelectionListType):
1970         (WebCore::Editor::changeSelectionListType):
1971         * editing/Editor.h:
1972
1973 2018-11-09  Keith Rollin  <krollin@apple.com>
1974
1975         Unreviewed build fix after https://bugs.webkit.org/show_bug.cgi?id=191324
1976
1977         Remove the use of .xcfilelists until their side-effects are better
1978         understood.
1979
1980         * WebCore.xcodeproj/project.pbxproj:
1981
1982 2018-11-09  Jer Noble  <jer.noble@apple.com>
1983
1984         SourceBuffer throws an error when appending a second init segment after changeType().
1985         https://bugs.webkit.org/show_bug.cgi?id=191474
1986
1987         Reviewed by Eric Carlson.
1988
1989         Test: media/media-source/media-source-changetype-second-init.html
1990
1991         When encountering an initialization segment after changeType(), add the parsed codec types
1992         to the list of allowed codecs.
1993
1994         * Modules/mediasource/SourceBuffer.cpp:
1995         (WebCore::SourceBuffer::validateInitializationSegment):
1996         * platform/mock/mediasource/MockMediaPlayerMediaSource.cpp:
1997         (WebCore::MockMediaPlayerMediaSource::supportsType):
1998         * platform/mock/mediasource/MockSourceBufferPrivate.cpp:
1999         (WebCore::MockSourceBufferPrivate::canSwitchToType):
2000         * platform/mock/mediasource/MockSourceBufferPrivate.h:
2001
2002 2018-11-09  Eric Carlson  <eric.carlson@apple.com>
2003
2004         [MediaStream] AVVideoCaptureSource reports incorrect size when frames are scaled
2005         https://bugs.webkit.org/show_bug.cgi?id=191479
2006         <rdar://problem/45952201>
2007
2008         Reviewed by Jer Noble.
2009
2010         No new tests, tested manually.
2011
2012         * platform/mediastream/RealtimeVideoSource.cpp:
2013         (WebCore::RealtimeVideoSource::standardVideoSizes): Drive-by fix: add a few more standard
2014         video frame sizes, correct a typo.
2015         (WebCore::RealtimeVideoSource::bestSupportedSizeAndFrameRate): Drive-by fix: don't consider
2016         rescaled sized when we already have an exact or aspect ratio match because it won't be used.
2017
2018         * platform/mediastream/mac/AVVideoCaptureSource.h:
2019         (WebCore::AVVideoCaptureSource::width const): Deleted.
2020         (WebCore::AVVideoCaptureSource::height const): Deleted.
2021         * platform/mediastream/mac/AVVideoCaptureSource.mm:
2022         (WebCore::AVVideoCaptureSource::setSizeAndFrameRateWithPreset): Delete m_requestedSize.
2023         (WebCore::AVVideoCaptureSource::shutdownCaptureSession): Delete m_width and m_height.
2024         (WebCore::AVVideoCaptureSource::processNewFrame): Don't call setSize with captured size,
2025         the frame may be resized before deliver.
2026
2027 2018-11-09  Ross Kirsling  <ross.kirsling@sony.com>
2028
2029         Unreviewed MSVC build fix after r238039 (and r238046).
2030
2031         * bindings/js/JSWorkerGlobalScopeBase.cpp:
2032         * bindings/js/JSWorkerGlobalScopeBase.h:
2033
2034 2018-11-09  Basuke Suzuki  <basuke.suzuki@sony.com>
2035
2036         [Curl][WebKit] Implement Proxy configuration API.
2037         https://bugs.webkit.org/show_bug.cgi?id=189053
2038
2039         Reviewed by Youenn Fablet.
2040
2041         Added API to set proxy from the app.
2042
2043         No new tests because there's no behaviour change in WebCore.
2044
2045         * platform/network/NetworkStorageSession.h:
2046         * platform/network/curl/CurlContext.h:
2047         (WebCore::CurlContext::setProxySettings):
2048         * platform/network/curl/CurlProxySettings.h:
2049         * platform/network/curl/NetworkStorageSessionCurl.cpp:
2050         (WebCore::NetworkStorageSession::setProxySettings const):
2051
2052 2018-11-09  Antti Koivisto  <antti@apple.com>
2053
2054         Use OptionSet for layout milestones
2055         https://bugs.webkit.org/show_bug.cgi?id=191470
2056
2057         Reviewed by Dean Jackson.
2058
2059         * WebCore.xcodeproj/project.pbxproj:
2060         * loader/EmptyFrameLoaderClient.h:
2061         * loader/FrameLoader.cpp:
2062         (WebCore::FrameLoader::didReachLayoutMilestone):
2063         * loader/FrameLoader.h:
2064         * loader/FrameLoaderClient.h:
2065         * page/FrameView.cpp:
2066         (WebCore::FrameView::FrameView):
2067         (WebCore::FrameView::addPaintPendingMilestones):
2068         (WebCore::FrameView::fireLayoutRelatedMilestonesIfNeeded):
2069         (WebCore::FrameView::firePaintRelatedMilestonesIfNeeded):
2070         * page/FrameView.h:
2071         * page/LayoutMilestone.h: Copied from Source/WebCore/page/LayoutMilestones.h.
2072
2073         Renamed to appease WK2 IPC code generation.
2074
2075         * page/LayoutMilestones.h: Removed.
2076         * page/Page.cpp:
2077         (WebCore::Page::addLayoutMilestones):
2078         (WebCore::Page::removeLayoutMilestones):
2079         (WebCore::Page::isCountingRelevantRepaintedObjects const):
2080         * page/Page.h:
2081         (WebCore::Page::requestedLayoutMilestones const):
2082
2083 2018-11-09  Daniel Bates  <dabates@apple.com>
2084
2085         [iOS] Draw caps lock indicator in password fields
2086         https://bugs.webkit.org/show_bug.cgi?id=190565
2087         <rdar://problem/45262343>
2088
2089         Reviewed by Dean Jackson.
2090
2091         Draw the caps lock indicator in a focused password field on iOS. This makes the behavior of password
2092         fields on iOS more closely match the behavior of password fields on Mac. For now, we only draw the
2093         indicator when caps locks is enabled via the hardware keyboard. We will look to support the software
2094         keyboard in a subsequent commit (see <https://bugs.webkit.org/show_bug.cgi?id=191475>).
2095
2096         The majority of this patch is implementing PlatformKeyboardEvent::currentCapsLockState() for iOS.
2097         In Legacy WebKit, the implementation boils down to calling call -[::WebEvent modifierFlags]. In
2098         Modern WebKit the UIProcess is responsible for -[::WebEvent modifierFlags] and passing it the
2099         WebProcess to store such that invocations of PlatformKeyboardEvent::currentCapsLockState() consult
2100         the store in the WebProcess. A smaller part of this patch is having both the legacy and modern
2101         web views listen for keyboard availability changes so as to update the the caps lock state when
2102         a hardware keyboard is detached or attached.
2103
2104         * WebCore.xcodeproj/project.pbxproj:
2105         * page/EventHandler.cpp:
2106         (WebCore::EventHandler::capsLockStateMayHaveChanged const): Extracted from EventHandler::internalKeyEvent()
2107         so that it can shared between WebCore, Modern WebKit, and Legacy WebKit code.
2108         (WebCore::EventHandler::internalKeyEvent): Modified to call capsLockStateMayHaveChanged().
2109         * page/EventHandler.h:
2110         * platform/cocoa/KeyEventCocoa.mm:
2111         (WebCore::PlatformKeyboardEvent::currentCapsLockState): Moved from KeyEventMac.mm.
2112         (WebCore::PlatformKeyboardEvent::getCurrentModifierState): Moved from KeyEventMac.mm.
2113         * platform/ios/KeyEventIOS.mm:
2114         (WebCore::PlatformKeyboardEvent::currentStateOfModifierKeys): Fetch the current modifier state.
2115         (WebCore::PlatformKeyboardEvent::currentCapsLockState): Deleted; we now use the Cocoa implementation.
2116         (WebCore::PlatformKeyboardEvent::getCurrentModifierState): Deleted; we now use the Cocoa implementation.
2117         * platform/ios/WebEvent.h:
2118         * platform/ios/WebEvent.mm:
2119         (+[WebEvent modifierFlags]): Added.
2120         * platform/mac/KeyEventMac.mm:
2121         (WebCore::PlatformKeyboardEvent::currentCapsLockState): Deleted; moved to KeyEventCocoa.mm to be shared
2122         by both Mac and iOS.
2123         (WebCore::PlatformKeyboardEvent::getCurrentModifierState): Deleted; moved to KeyEventCocoa.mm to be shared
2124         by both Mac and iOS.
2125         * rendering/RenderThemeCocoa.h:
2126         * rendering/RenderThemeCocoa.mm:
2127         (WebCore::RenderThemeCocoa::shouldHaveCapsLockIndicator const): Moved from RenderThemeMac.mm.
2128         * rendering/RenderThemeIOS.h:
2129         * rendering/RenderThemeIOS.mm:
2130         (WebCore::RenderThemeIOS::shouldHaveCapsLockIndicator const): Deleted.
2131         * rendering/RenderThemeMac.h:
2132         * rendering/RenderThemeMac.mm:
2133         (WebCore::RenderThemeMac::shouldHaveCapsLockIndicator const): Deleted; moved to RenderThemeCocoa.mm to be
2134         shared by both Mac and iOS.
2135
2136 2018-11-09  Chris Dumez  <cdumez@apple.com>
2137
2138         Unreviewed attempt to fix WinCairo build after r238039.
2139
2140         * bindings/js/JSWorkerGlobalScopeBase.h:
2141
2142 2018-11-09  Fujii Hironori  <Hironori.Fujii@sony.com>
2143
2144         Extensions3DOpenGLES.h:  warning: 'blitFramebuffer' overrides a member function but is not marked 'override' [-Winconsistent-missing-override]
2145         https://bugs.webkit.org/show_bug.cgi?id=191451
2146
2147         Reviewed by Dean Jackson.
2148
2149         No new tests because there is no behavior change.
2150
2151         * platform/graphics/opengl/Extensions3DOpenGLES.cpp:
2152         (WebCore::Extensions3DOpenGLES::setEXTContextLostCallback): Deleted unused method.
2153         * platform/graphics/opengl/Extensions3DOpenGLES.h: Marked 'override'.
2154
2155 2018-11-09  Andy Estes  <aestes@apple.com>
2156
2157         [Payment Request] canMakePayment() should not consider serialized payment method data
2158         https://bugs.webkit.org/show_bug.cgi?id=191432
2159
2160         Reviewed by Dean Jackson.
2161
2162         In https://github.com/w3c/payment-request/pull/806, we're changing the specification of
2163         canMakePayment() to not consider serialized payment method data when deciding if a payment
2164         method is supported. For Apple Pay, this means we resolve to true for
2165         "https://apple.com/apple-pay", even if an ApplePayRequest is omitted or is missing required
2166         fields.
2167
2168         Added test cases to
2169         http/tests/paymentrequest/payment-request-canmakepayment-method.https.html and
2170         http/tests/paymentrequest/payment-request-show-method.https.html.
2171
2172         * Modules/paymentrequest/PaymentRequest.cpp:
2173         (WebCore::PaymentRequest::canMakePayment):
2174
2175 2018-11-09  Andy Estes  <aestes@apple.com>
2176
2177         [Payment Request] PaymentResponse.details should be updated when the user accepts a retried payment
2178         https://bugs.webkit.org/show_bug.cgi?id=191440
2179
2180         Reviewed by Dean Jackson.
2181
2182         PaymentResponse.details was being initialized in the PaymentResponse constructor and never
2183         updated when the user accepts a retried payment. We need to update it.
2184
2185         Added a test case to http/tests/paymentrequest/payment-response-retry-method.https.html.
2186
2187         * Modules/paymentrequest/PaymentRequest.cpp:
2188         (WebCore::PaymentRequest::accept):
2189         * Modules/paymentrequest/PaymentResponse.cpp:
2190         (WebCore::PaymentResponse::PaymentResponse):
2191         (WebCore::PaymentResponse::setDetailsFunction):
2192         * Modules/paymentrequest/PaymentResponse.h:
2193
2194 2018-11-09  Fujii Hironori  <Hironori.Fujii@sony.com>
2195
2196         MediaPlayerPrivateMediaFoundation.h: warning: 'GetService' overrides a member function but is not marked 'override' [-Winconsistent-missing-override]
2197         https://bugs.webkit.org/show_bug.cgi?id=191453
2198
2199         Reviewed by Per Arne Vollan.
2200
2201         No new tests because there is no behavior change.
2202
2203         * platform/graphics/win/MediaPlayerPrivateMediaFoundation.h: Marked with 'override' and removed 'virtual'.
2204
2205 2018-11-09  Chris Dumez  <cdumez@apple.com>
2206
2207         Unreviewed attempt to fix internal build on macOS.
2208
2209         'Export' is defined in several headers.
2210
2211         * bindings/js/JSDOMGlobalObject.cpp:
2212         * bindings/js/JSDOMGlobalObject.h:
2213         * bridge/jsc/BridgeJSC.cpp:
2214         * bridge/jsc/BridgeJSC.h:
2215
2216 2018-11-09  Chris Dumez  <cdumez@apple.com>
2217
2218         HTML form validation bubble disappears
2219         https://bugs.webkit.org/show_bug.cgi?id=191418
2220
2221         Reviewed by Simon Fraser.
2222
2223         If we validate a form and find an invalid form control, we'll scroll it into view and show
2224         the validation bubble. However, scrolling the element into view may be an asynchronous
2225         operation, in which case it would discard the validation bubble prematurely because scrolling
2226         hides the validation bubble. To address the issue, we now show the validation message
2227         asynchronously after focusing the element (and potentially scrolling it into view).
2228
2229         Test: fast/forms/scroll-into-view-and-show-validation-message.html
2230
2231         * html/HTMLFormControlElement.cpp:
2232         (WebCore::HTMLFormControlElement::focusAndShowValidationMessage):
2233
2234 2018-11-09  Brent Fulgham  <bfulgham@apple.com>
2235
2236         [Windows][DirectX] Be more rigors about BeginFigure/EndFigure and Close operations.
2237         https://bugs.webkit.org/show_bug.cgi?id=191452
2238         <rdar://problem/45933964>
2239
2240         Reviewed by Zalan Bujtas.
2241
2242         Do a better job of balancing the BeginFigure/EndFigure calls in
2243         the PathDirect2D implementation. Failure to do so puts the Geometry sink
2244         into an error state that prevents it from producing drawing output.
2245
2246
2247         * platform/graphics/Path.h:
2248         * platform/graphics/win/GraphicsContextDirect2D.cpp:
2249         (WebCore::GraphicsContext::drawPath): Flush is needed here.
2250         (WebCore::GraphicsContext::fillPath): Ditto.
2251         (WebCore::GraphicsContext::strokePath): Ditto.
2252         * platform/graphics/win/PathDirect2D.cpp:
2253         (WebCore::Path::drawDidComplete):
2254         (WebCore::Path::closeAnyOpenGeometries):
2255         (WebCore::Path::transform):
2256         (WebCore::Path::openFigureAtCurrentPointIfNecessary):
2257         (WebCore::Path::moveTo):
2258         (WebCore::Path::closeSubpath):
2259
2260 2018-11-09  Jer Noble  <jer.noble@apple.com>
2261
2262         [Cocoa] Fix failing imported/w3c/web-platform-tests/media-source/mediasource-changetype-play.html test
2263         https://bugs.webkit.org/show_bug.cgi?id=191396
2264
2265         Reviewed by Eric Carlson.
2266
2267         When changeType() is called, exempt video and text tracks (in addition to just audio tracks)
2268         from "same codec" requirements.
2269
2270         * Modules/mediasource/SourceBuffer.cpp:
2271         (WebCore::SourceBuffer::validateInitializationSegment):
2272
2273 2018-11-09  Carlos Garcia Campos  <cgarcia@igalia.com>
2274
2275         REGRESSION(r236365): [GTK] Many form-related tests are failing
2276         https://bugs.webkit.org/show_bug.cgi?id=189993
2277
2278         Reviewed by Michael Catanzaro.
2279
2280         Only the first form data element is added to the message body due to a return added by mistake in r236365.
2281
2282         * platform/network/soup/ResourceRequestSoup.cpp:
2283         (WebCore::ResourceRequest::updateSoupMessageBody const): Remove return.
2284
2285 2018-11-09  Zalan Bujtas  <zalan@apple.com>
2286
2287         [LFC][IFC] Move some code from InlineFormattingContext::Line to InlineFormattingContext/Geometry
2288         https://bugs.webkit.org/show_bug.cgi?id=191445
2289
2290         Reviewed by Antti Koivisto.
2291
2292         The idea here is that Line should not have to deal with all the post processig tasks like the runs final aligments.
2293         (The line class would eventually turn into a collection of random things).
2294
2295         * layout/inlineformatting/InlineFormattingContext.cpp:
2296         (WebCore::Layout::InlineFormattingContext::closeLine const):
2297         (WebCore::Layout::InlineFormattingContext::appendContentToLine const):
2298         (WebCore::Layout::InlineFormattingContext::layoutInlineContent const):
2299         * layout/inlineformatting/InlineFormattingContext.h:
2300         (WebCore::Layout::InlineFormattingContext::Line::contentLogicalLeft const):
2301         (WebCore::Layout::InlineFormattingContext::Line::lastRunType const):
2302         * layout/inlineformatting/InlineFormattingContextGeometry.cpp:
2303         (WebCore::Layout::InlineFormattingContext::Geometry::adjustedLineLogicalLeft):
2304         (WebCore::Layout::InlineFormattingContext::Geometry::justifyRuns):
2305         (WebCore::Layout::InlineFormattingContext::Geometry::computeExpansionOpportunities):
2306         * layout/inlineformatting/Line.cpp:
2307         (WebCore::Layout::InlineFormattingContext::Line::Line):
2308         (WebCore::Layout::InlineFormattingContext::Line::init):
2309         (WebCore::Layout::InlineFormattingContext::Line::contentLogicalRight const):
2310         (WebCore::Layout::InlineFormattingContext::Line::appendContent):
2311         (WebCore::Layout::InlineFormattingContext::Line::close):
2312         (WebCore::Layout::adjustedLineLogicalLeft): Deleted.
2313         (WebCore::Layout::InlineFormattingContext::Line::contentLogicalRight): Deleted.
2314         (WebCore::Layout::InlineFormattingContext::Line::computeExpansionOpportunities): Deleted.
2315         (WebCore::Layout::InlineFormattingContext::Line::justifyRuns): Deleted.
2316
2317 2018-11-09  Philippe Normand  <pnormand@igalia.com>
2318
2319         Unreviewed, GStreamer build warning fix
2320
2321         * platform/mediastream/libwebrtc/GStreamerVideoEncoderFactory.cpp:
2322         (WebCore::GStreamerVideoEncoder::newSampleCallback): Timesamp()
2323         returns a uint32_t, fix format string accordingly.
2324
2325 2018-11-08  Said Abou-Hallawa  <sabouhallawa@apple.com>
2326
2327         Add an SPI to allow WebView clients to add additional supported image formats
2328         https://bugs.webkit.org/show_bug.cgi?id=190454
2329
2330         Reviewed by Simon Fraser.
2331
2332         Add an SPI to allow additional supported image formats in WebView. These
2333         additional formats can be set in the WKWebViewConfiguration as an NSArray
2334         of NStrings. Each string represents an image source type aka UTI.
2335
2336         The ImageSourceType in the functions' names will be replaced by ImageType.
2337         ImageType in this context is the image UTI (Uniform Type Identifier).
2338
2339         * platform/MIMETypeRegistry.cpp:
2340         (WebCore::MIMETypeRegistry::supportedImageMIMETypes):
2341         (WebCore::MIMETypeRegistry::additionalSupportedImageMIMETypes):
2342         (WebCore::supportedImageMIMETypesForEncoding):
2343         (WebCore::MIMETypeRegistry::isSupportedImageMIMEType):
2344         * platform/MIMETypeRegistry.h:
2345         * platform/graphics/cg/ImageDecoderCG.cpp:
2346         (WebCore::ImageDecoderCG::filenameExtension const):
2347         (WebCore::ImageDecoderCG::encodedDataStatus const):
2348         * platform/graphics/cg/ImageSourceCG.h:
2349         * platform/graphics/cg/ImageSourceCGMac.mm:
2350         (WebCore::MIMETypeForImageType):
2351         (WebCore::preferredExtensionForImageType):
2352         (WebCore::MIMETypeForImageSourceType): Deleted.
2353         (WebCore::preferredExtensionForImageSourceType): Deleted.
2354         * platform/graphics/cg/ImageSourceCGWin.cpp:
2355         (WebCore::MIMETypeForImageType):
2356         (WebCore::preferredExtensionForImageType):
2357         (WebCore::MIMETypeForImageSourceType): Deleted.
2358         (WebCore::preferredExtensionForImageSourceType): Deleted.
2359         * platform/graphics/cg/UTIRegistry.cpp:
2360         (WebCore::defaultSupportedImageTypes):
2361         (WebCore::additionalSupportedImageTypes):
2362         (WebCore::setAdditionalSupportedImageTypes):
2363         (WebCore::isSupportedImageType):
2364         (WebCore::supportedDefaultImageSourceTypes): Deleted.
2365         (WebCore::isSupportImageSourceType): Deleted.
2366         * platform/graphics/cg/UTIRegistry.h:
2367
2368 2018-11-08  Megan Gardner  <megan_gardner@apple.com>
2369
2370         Adopt Reveal Framework to replace Lookup
2371         https://bugs.webkit.org/show_bug.cgi?id=191097
2372
2373         Reviewed by Tim Horton.
2374
2375         Reveal is not currently testable.
2376
2377         The Reveal framework does the same job as Lookup and DataDectors.
2378         In this patch we switch from using Lookup to determine what text
2379         to select and define to using Reveal and RVItems. Since this
2380         code needs to work on older systems, and Reveal is newer, we also need
2381         to keep around the old code for old systems so that they can also
2382         continue to work. Eventually we will remove this code and also likly switch
2383         to passing RVItems across from the web process rather than making them
2384         on each side.
2385
2386         * SourcesCocoa.txt:
2387         * WebCore.xcodeproj/project.pbxproj:
2388         * editing/mac/DictionaryLookup.h:
2389         * editing/mac/DictionaryLookup.mm:
2390
2391         Create a delegate to respond to Reveal and help draw the string or highlight.
2392
2393         (-[WebRevealHighlight initWithHighlightRect:useDefaultHighlight:attributedString:]):
2394         (-[WebRevealHighlight setClearTextIndicator:]):
2395         (-[WebRevealHighlight revealContext:rectsForItem:]):
2396         (-[WebRevealHighlight revealContext:drawRectsForItem:]):
2397         (-[WebRevealHighlight revealContext:shouldUseDefaultHighlightForItem:]):
2398         (-[WebRevealHighlight revealContext:stopHighlightingItem:]):
2399         (WebCore::showPopupOrCreateAnimationController):
2400
2401         Unify the code paths and utalize the Reveal framework to create and/or display the popovers.
2402
2403         (WebCore::DictionaryLookup::showPopup):
2404         (WebCore::DictionaryLookup::animationControllerForPopup):
2405
2406         Pipe the new callback through.
2407
2408         (WebCore::tokenRange): Deleted.
2409
2410         Only used with Lookup
2411
2412         (WebCore::selectionContainsPosition): Deleted.
2413
2414         Only used with Lookup.
2415
2416         * editing/mac/DictionaryLookupLegacy.mm: Copied from Source/WebCore/editing/mac/DictionaryLookup.mm.
2417
2418         Keep a copy of the previous implementation of DictionaryLookup, because Reveal not available on older
2419         system.
2420
2421         (WebCore::tokenRange):
2422         (WebCore::selectionContainsPosition):
2423         (WebCore::expandSelectionByCharacters):
2424         (WebCore::showPopupOrCreateAnimationController):
2425         (WebCore::DictionaryLookup::showPopup):
2426         (WebCore::DictionaryLookup::hidePopup):
2427         (WebCore::DictionaryLookup::animationControllerForPopup):
2428
2429 2018-11-08  Keith Rollin  <krollin@apple.com>
2430
2431         Create .xcfilelist files
2432         https://bugs.webkit.org/show_bug.cgi?id=191324
2433         <rdar://problem/45852819>
2434
2435         Reviewed by Alex Christensen.
2436
2437         As part of preparing for enabling XCBuild, create and use .xcfilelist
2438         files. These files are using during Run Script build phases in an
2439         Xcode project. If a Run Script build phase produces new files that are
2440         used later as inputs to subsequent build phases, XCBuild needs to know
2441         about these files. These files can be either specified in an "output
2442         files" section of the Run Script phase editor, or in .xcfilelist files
2443         that are associated with the Run Script build phase.
2444
2445         This patch takes the second approach. It consists of three sets of changes:
2446
2447         - Modify the DerivedSources.make files to have a
2448           'print_all_generated_files" target that produces a list of the files
2449           they create.
2450
2451         - Create a shell script that produces .xcfilelist files from the
2452           output of the previous step, as well as for the files created in the
2453           Generate Unified Sources build steps.
2454
2455         - Add the new .xcfilelist files to the associated projects.
2456
2457         Note that, with these changes, the Xcode workspace and projects can no
2458         longer be fully loaded into Xcode 9. Xcode will attempt to load the
2459         projects that have .xcfilelist files associated with them, but will
2460         fail and display a placeholder for those projects instead. It's
2461         expected that all developers are using Xcode 10 by now and that not
2462         being able to load into Xcode 9 is not a practical issue. Keep in mind
2463         that this is strictly an IDE issue, and that the projects can still be
2464         built with `xcodebuild`.
2465
2466         Also note that the shell script that creates the .xcfilelist files can
2467         also be used to verify that the set of files that's currently checked
2468         in is up-to-date. This checking can be used as part of a check-in hook
2469         or part of check-webkit-style to sooner catch cases where the
2470         .xcfilelist files need to be regenerated.
2471
2472         No new tests -- no changed functionality.
2473
2474         * DerivedSources.make:
2475         * DerivedSources.xcfilelist: Added.
2476         * UnifiedSources.xcfilelist: Added.
2477         * WebCore.xcodeproj/project.pbxproj:
2478
2479 2018-11-08  Don Olmstead  <don.olmstead@sony.com>
2480
2481         Provide generic implementation of SSLKeyGenerator functions
2482         https://bugs.webkit.org/show_bug.cgi?id=191428
2483
2484         Reviewed by Michael Catanzaro.
2485
2486         No new tests. No change in behavior.
2487
2488         Both SSLKeyGeneratorGLib and SSLKeyGeneratorIOS were stub
2489         implementations. Since <keygen> is deprecated it looks unlikely that
2490         an implementation is actually needed, however an implementation is
2491         needed until support is reomved.
2492
2493         * Sources.txt:
2494         * SourcesCocoa.txt:
2495         * WebCore.xcodeproj/project.pbxproj:
2496         * platform/SSLKeyGenerator.cpp: Added.
2497         (WebCore::getSupportedKeySizes):
2498         (WebCore::signedPublicKeyAndChallengeString):
2499         * platform/SourcesGLib.txt:
2500         * platform/glib/SSLKeyGeneratorGLib.cpp: Removed.
2501         * platform/ios/SSLKeyGeneratorIOS.cpp: Removed.
2502
2503 2018-11-07  Myles C. Maxfield  <mmaxfield@apple.com>
2504
2505         Unprefix text-decoration CSS3 properties
2506         https://bugs.webkit.org/show_bug.cgi?id=127193
2507
2508         Reviewed by Dean Jackson.
2509
2510         The properties are stable, and there is interop.
2511
2512         Test: fast/css3-text/css3-text-decoration/unprefix.html
2513
2514         * css/CSSComputedStyleDeclaration.cpp:
2515         (WebCore::ComputedStyleExtractor::valueForPropertyinStyle):
2516         * css/CSSProperties.json:
2517         * css/StyleResolver.cpp:
2518         (WebCore::shouldApplyPropertyInParseOrder):
2519         (WebCore::isValidVisitedLinkProperty):
2520         * css/parser/CSSParserFastPaths.cpp:
2521         (WebCore::isColorPropertyID):
2522         (WebCore::CSSParserFastPaths::isValidKeywordPropertyAndValue):
2523         (WebCore::CSSParserFastPaths::isKeywordPropertyID):
2524         * css/parser/CSSPropertyParser.cpp:
2525         (WebCore::CSSPropertyParser::parseSingleValue):
2526         * editing/cocoa/DataDetection.mm:
2527         (WebCore::DataDetection::detectContentInRange):
2528         * rendering/TextDecorationPainter.cpp:
2529         (WebCore::decorationColor):
2530         * rendering/style/RenderStyle.cpp:
2531         (WebCore::RenderStyle::colorIncludingFallback const):
2532         (WebCore::RenderStyle::visitedDependentColor const):
2533
2534 2018-11-08  Timothy Hatcher  <timothy@apple.com>
2535
2536         Add experimental support for a `supported-color-schemes` CSS property.
2537         https://bugs.webkit.org/show_bug.cgi?id=191319
2538         rdar://problem/45852261
2539
2540         Reviewed by Dean Jackson.
2541
2542         Tests: css-dark-mode/parse-supported-color-schemes.html
2543                css-dark-mode/supported-color-schemes-css.html
2544
2545         * WebCore.xcodeproj/project.pbxproj:
2546         * css/CSSComputedStyleDeclaration.cpp:
2547         (WebCore::ComputedStyleExtractor::valueForPropertyinStyle):
2548         * css/CSSProperties.json:
2549         * css/CSSValueKeywords.in:
2550         * css/StyleBuilderConverter.h:
2551         (WebCore::StyleBuilderConverter::updateSupportedColorSchemes):
2552         (WebCore::StyleBuilderConverter::convertSupportedColorSchemes):
2553         * css/StyleResolver.cpp:
2554         (WebCore::StyleResolver::colorFromPrimitiveValue const):
2555         * css/parser/CSSPropertyParser.cpp:
2556         (WebCore::consumeSupportedColorSchemes):
2557         (WebCore::CSSPropertyParser::parseSingleValue):
2558         * dom/Document.cpp:
2559         (WebCore::Document::resetLinkColor):
2560         (WebCore::Document::resetVisitedLinkColor):
2561         (WebCore::Document::resetActiveLinkColor):
2562         (WebCore::Document::processSupportedColorSchemes):
2563         (WebCore::Document::useDarkAppearance const):
2564         (WebCore::Document::styleColorOptions const):
2565         * dom/Document.h:
2566         * editing/cocoa/WebContentReaderCocoa.mm:
2567         (WebCore::createFragment):
2568         * html/canvas/CanvasRenderingContext2D.cpp:
2569         (WebCore::CanvasRenderingContext2D::drawFocusIfNeededInternal):
2570         * inspector/InspectorOverlay.cpp:
2571         (WebCore::InspectorOverlay::paint):
2572         * page/FrameView.cpp:
2573         (WebCore::FrameView::useDarkAppearance const):
2574         (WebCore::FrameView::paintScrollCorner):
2575         * platform/mac/DragImageMac.mm:
2576         (WebCore::createDragImageForLink):
2577         * rendering/InlineFlowBox.cpp:
2578         (WebCore::InlineFlowBox::paintBoxDecorations):
2579         * rendering/InlineTextBox.cpp:
2580         (WebCore::InlineTextBox::paintPlatformDocumentMarker):
2581         * rendering/RenderBox.cpp:
2582         (WebCore::RenderBox::paintRootBoxFillLayers):
2583         (WebCore::RenderBox::paintBackground):
2584         * rendering/RenderElement.cpp:
2585         (WebCore::RenderElement::selectionColor const):
2586         (WebCore::RenderElement::selectionBackgroundColor const):
2587         (WebCore::RenderElement::paintFocusRing):
2588         * rendering/RenderImage.cpp:
2589         (WebCore::RenderImage::paintAreaElementFocusRing):
2590         * rendering/RenderLayerBacking.cpp:
2591         (WebCore::RenderLayerBacking::paintContents):
2592         * rendering/RenderLayerCompositor.cpp:
2593         (WebCore::RenderLayerCompositor::paintContents):
2594         * rendering/RenderListBox.cpp:
2595         (WebCore::RenderListBox::paintItemForeground):
2596         (WebCore::RenderListBox::paintItemBackground):
2597         * rendering/RenderObject.cpp:
2598         (WebCore::RenderObject::useDarkAppearance const):
2599         (WebCore::RenderObject::styleColorOptions const):
2600         * rendering/RenderObject.h:
2601         * rendering/RenderTableCell.cpp:
2602         (WebCore::RenderTableCell::paintBackgroundsBehindCell):
2603         * rendering/RenderTheme.cpp:
2604         (WebCore::RenderTheme::paint):
2605         * rendering/RenderTheme.h:
2606         (WebCore::RenderTheme::usingDarkAppearance const): Deleted.
2607         * rendering/RenderThemeMac.h:
2608         * rendering/RenderThemeMac.mm:
2609         (WebCore::RenderThemeMac::documentViewFor const):
2610         (WebCore::RenderThemeMac::adjustMenuListStyle const):
2611         (WebCore::RenderThemeMac::paintCellAndSetFocusedElementNeedsRepaintIfNecessary):
2612         (WebCore::RenderThemeMac::paintSliderThumb):
2613         (WebCore::RenderThemeMac::usingDarkAppearance const): Deleted.
2614         * rendering/style/RenderStyle.h:
2615         (WebCore::RenderStyle::supportedColorSchemes const):
2616         (WebCore::RenderStyle::setSupportedColorSchemes):
2617         (WebCore::RenderStyle::initialSupportedColorSchemes):
2618         * rendering/style/RenderStyleConstants.h:
2619         * rendering/style/StyleRareInheritedData.cpp:
2620         (WebCore::StyleRareInheritedData::StyleRareInheritedData):
2621         (WebCore::StyleRareInheritedData::operator== const):
2622         * rendering/style/StyleRareInheritedData.h:
2623         * rendering/style/StyleSupportedColorSchemes.h: Added.
2624         (WebCore::StyleSupportedColorSchemes::StyleSupportedColorSchemes):
2625         (WebCore::StyleSupportedColorSchemes::operator== const):
2626         (WebCore::StyleSupportedColorSchemes::operator!= const):
2627         (WebCore::StyleSupportedColorSchemes::isAuto const):
2628         (WebCore::StyleSupportedColorSchemes::isOnly const):
2629         (WebCore::StyleSupportedColorSchemes::colorSchemes const):
2630         (WebCore::StyleSupportedColorSchemes::add):
2631         (WebCore::StyleSupportedColorSchemes::contains const):
2632         (WebCore::StyleSupportedColorSchemes::setAllowsTransformations):
2633         (WebCore::StyleSupportedColorSchemes::allowsTransformations const):
2634         * svg/graphics/SVGImage.cpp:
2635         (WebCore::SVGImage::draw):
2636
2637 2018-11-08  Joseph Pecoraro  <pecoraro@apple.com>
2638
2639         Web Inspector: Restrict domains at the target level instead of only at the window level
2640         https://bugs.webkit.org/show_bug.cgi?id=191344
2641
2642         Reviewed by Devin Rousso.
2643
2644         * inspector/WorkerInspectorController.cpp:
2645         (WebCore::WorkerInspectorController::WorkerInspectorController):
2646         Remove Inspector domain from workers. It was unused anyways.
2647
2648 2018-11-08  Dean Jackson  <dino@apple.com>
2649
2650         Add a String literal that returns a String
2651         https://bugs.webkit.org/show_bug.cgi?id=191425
2652         <rdar://problem/45914556>
2653
2654         Reviewed by Sam Weinig.
2655
2656         Use _str where possible.
2657
2658         API Test in WPT.
2659
2660         * Modules/fetch/FetchRequest.cpp:
2661         (WebCore::computeReferrer):
2662         * Modules/indexeddb/IDBKeyPath.cpp:
2663         (WebCore::loggingString):
2664         * Modules/webdatabase/OriginLock.cpp:
2665         (WebCore::OriginLock::lockFileNameForPath):
2666         * css/CSSBasicShapes.cpp:
2667         (WebCore::updateCornerRadiusWidthAndHeight):
2668         * html/canvas/WebGL2RenderingContext.cpp:
2669         (WebCore::WebGL2RenderingContext::getParameter):
2670         * html/canvas/WebGLRenderingContext.cpp:
2671         (WebCore::WebGLRenderingContext::getParameter):
2672         * loader/LinkHeader.cpp:
2673         (WebCore::parseParameterValue):
2674         * loader/LinkLoader.cpp:
2675         (WebCore::LinkLoader::preloadIfNeeded):
2676         * page/NavigatorBase.cpp:
2677         (WebCore::NavigatorBase::platform):
2678         * platform/DateComponents.cpp:
2679         (WebCore::DateComponents::toString const):
2680         * platform/mac/PlatformEventFactoryMac.mm:
2681         (WebCore::keyIdentifierForKeyEvent):
2682         * rendering/RenderListMarker.cpp:
2683         (WebCore::RenderListMarker::suffix const):
2684         * rendering/RenderMenuList.cpp:
2685         (RenderMenuList::setText):
2686         * testing/InternalSettings.cpp:
2687         (WebCore::InternalSettings::userInterfaceDirectionPolicy):
2688         (WebCore::InternalSettings::systemLayoutDirection):
2689         * testing/Internals.cpp:
2690         (WebCore::Internals::shadowRootType const):
2691         (WebCore::Internals::getCurrentCursorInfo):
2692
2693 2018-11-08  Jonathan Hammer  <jonathan@e3software.com>
2694
2695         Plain text drag in contenteditable is always DragOperationCopy, never DragOperationMove
2696         https://bugs.webkit.org/show_bug.cgi?id=191228
2697         <rdar://problem/45786830>
2698
2699         Reviewed by Wenson Hsieh.
2700
2701         DragController::beginDrag should not call cleanupAfterSystemDrag because
2702         the drag is still in progress even after the call to m_client.beginDrag()
2703         returns. This is in contrast to DragController::doSystemDrag, where the
2704         call to cleanupAfterSystemDrag is appropriate because the drag has
2705         concluded by the time m_client.startDrag() returns.
2706
2707         Test: fast/events/drag-and-drop-move-not-copy.html
2708
2709         * page/DragController.cpp:
2710         (WebCore::DragController::beginDrag):
2711
2712 2018-11-08  Zalan Bujtas  <zalan@apple.com>
2713
2714         [BFC][IFC] InlineFormattingContext::Line::alignRuns() should take care of all the alignments.
2715         https://bugs.webkit.org/show_bug.cgi?id=191414
2716
2717         Reviewed by Antti Koivisto.
2718
2719         * layout/inlineformatting/Line.cpp:
2720         (WebCore::Layout::InlineFormattingContext::Line::close):
2721
2722 2018-11-08  Jiewen Tan  <jiewen_tan@apple.com>
2723
2724         [WebAuthN] Import CTAP device request/response converters from Chromium
2725         https://bugs.webkit.org/show_bug.cgi?id=190784
2726         <rdar://problem/45460333>
2727
2728         Reviewed by Brent Fulgham.
2729
2730         This patch imports Chromium's CTAP device request/response converters:
2731         https://fidoalliance.org/specs/fido-v2.0-ps-20170927/fido-client-to-authenticator-protocol-v2.0-ps-20170927.html#message-encoding
2732         1. It directly imports the following files and suit them to WebKit's coding style:
2733         https://cs.chromium.org/chromium/src/device/fido/device_response_converter.cc?l=20&rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2734         https://cs.chromium.org/chromium/src/device/fido/device_response_converter.h?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2735         https://cs.chromium.org/chromium/src/device/fido/authenticator_get_info_response.cc?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2736         https://cs.chromium.org/chromium/src/device/fido/authenticator_get_info_response.h?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2737         https://cs.chromium.org/chromium/src/device/fido/authenticator_supported_options.cc?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2738         https://cs.chromium.org/chromium/src/device/fido/authenticator_supported_options.h?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2739         https://cs.chromium.org/chromium/src/device/fido/ctap_request_unittest.cc?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2740         https://cs.chromium.org/chromium/src/device/fido/ctap_response_unittest.cc?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2741         https://cs.chromium.org/chromium/src/device/fido/fido_test_data.h?l=1&rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2742         2. It gathers the following methods into DeviceRequestConverter:
2743         CtapGetAssertionRequest::EncodeAsCBOR()
2744         CtapMakeCredentialRequest::EncodeAsCBOR()
2745         PublicKeyCredentialDescriptor::ConvertToCBOR()
2746         PublicKeyCredentialParams::ConvertToCBOR()
2747         PublicKeyCredentialRpEntity::ConvertToCBOR()
2748         PublicKeyCredentialUserEntity::ConvertToCBOR()
2749         3. It also apply a patch from Chromium to CBORValue:
2750         https://chromium.googlesource.com/chromium/src/+/7b6fbff35cd8e4d508f08e1cd93b4aa0a0dc402c%5E%21/
2751
2752         Besides importing things from Chromium, it also implements UserVerificationRequirement for both
2753         PublicKeyCredentialCreationOptions and PublicKeyCredentialRequestOptions such that both options
2754         can specify more dimensions of a desired authenticator.
2755
2756         Covered by API tests.
2757
2758         * CMakeLists.txt:
2759         * DerivedSources.make:
2760         * Modules/webauthn/PublicKeyCredentialCreationOptions.h:
2761         (WebCore::PublicKeyCredentialCreationOptions::AuthenticatorSelectionCriteria::encode const):
2762         (WebCore::PublicKeyCredentialCreationOptions::AuthenticatorSelectionCriteria::decode):
2763         * Modules/webauthn/PublicKeyCredentialCreationOptions.idl:
2764         * Modules/webauthn/PublicKeyCredentialRequestOptions.h:
2765         (WebCore::PublicKeyCredentialRequestOptions::encode const):
2766         (WebCore::PublicKeyCredentialRequestOptions::decode):
2767         * Modules/webauthn/PublicKeyCredentialRequestOptions.idl:
2768         * Modules/webauthn/UserVerificationRequirement.h: Copied from Source/WebCore/Modules/webauthn/PublicKeyCredentialRequestOptions.idl.
2769         * Modules/webauthn/UserVerificationRequirement.idl: Copied from Source/WebCore/Modules/webauthn/PublicKeyCredentialRequestOptions.idl.
2770         * Modules/webauthn/cbor/CBORValue.cpp:
2771         (cbor::CBORValue::CBORValue):
2772         (cbor::CBORValue::getBool const):
2773         * Modules/webauthn/cbor/CBORValue.h:
2774         * Modules/webauthn/fido/AuthenticatorGetInfoResponse.cpp: Added.
2775         (fido::toArrayValue):
2776         (fido::AuthenticatorGetInfoResponse::AuthenticatorGetInfoResponse):
2777         (fido::AuthenticatorGetInfoResponse::setMaxMsgSize):
2778         (fido::AuthenticatorGetInfoResponse::setPinProtocols):
2779         (fido::AuthenticatorGetInfoResponse::setExtensions):
2780         (fido::AuthenticatorGetInfoResponse::setOptions):
2781         (fido::encodeAsCBOR):
2782         * Modules/webauthn/fido/AuthenticatorGetInfoResponse.h: Added.
2783         * Modules/webauthn/fido/AuthenticatorSupportedOptions.cpp: Added.
2784         (fido::AuthenticatorSupportedOptions::setSupportsResidentKey):
2785         (fido::AuthenticatorSupportedOptions::setUserVerificationAvailability):
2786         (fido::AuthenticatorSupportedOptions::setUserPresenceRequired):
2787         (fido::AuthenticatorSupportedOptions::setClientPinAvailability):
2788         (fido::AuthenticatorSupportedOptions::setIsPlatformDevice):
2789         (fido::convertToCBOR):
2790         * Modules/webauthn/fido/AuthenticatorSupportedOptions.h: Added.
2791         * Modules/webauthn/fido/DeviceRequestConverter.cpp: Added.
2792         (fido::convertRpEntityToCBOR):
2793         (fido::convertUserEntityToCBOR):
2794         (fido::convertParametersToCBOR):
2795         (fido::convertDescriptorToCBOR):
2796         (fido::encodeMakeCredenitalRequestAsCBOR):
2797         (fido::encodeGetAssertionRequestAsCBOR):
2798         (fido::encodeEmptyAuthenticatorRequest):
2799         * Modules/webauthn/fido/DeviceRequestConverter.h: Copied from Source/WebCore/Modules/webauthn/fido/FidoConstants.h.
2800         * Modules/webauthn/fido/DeviceResponseConverter.cpp: Added.
2801         (fido::convertStringToProtocolVersion):
2802         (fido::getResponseCode):
2803         (fido::getCredentialId):
2804         (fido::readCTAPMakeCredentialResponse):
2805         (fido::readCTAPGetAssertionResponse):
2806         (fido::readCTAPGetInfoResponse):
2807         * Modules/webauthn/fido/DeviceResponseConverter.h: Copied from Source/WebCore/Modules/webauthn/fido/FidoConstants.cpp.
2808         * Modules/webauthn/fido/FidoConstants.cpp:
2809         (fido::isCtapDeviceResponseCode):
2810         (fido::publicKeyCredentialTypeToString):
2811         * Modules/webauthn/fido/FidoConstants.h:
2812         * Sources.txt:
2813         * WebCore.xcodeproj/project.pbxproj:
2814
2815 2018-11-07  Justin Michaud  <justin_michaud@apple.com>
2816
2817         CSS Painting API should pass size, arguments and input properties to paint callback
2818         https://bugs.webkit.org/show_bug.cgi?id=191309
2819
2820         Reviewed by Chris Dumez.
2821
2822         Call paint() callback with input properties and arguments. This patch adds a stub for
2823         the CSS Typed OM StylePropertyMapReadOnly, and passes all the arguments as strings without
2824         any syntax checking to the paint callback.
2825
2826         Test: fast/css-custom-paint/properties.html
2827
2828         * CMakeLists.txt:
2829         * DerivedSources.make:
2830         * Sources.txt:
2831         * WebCore.xcodeproj/project.pbxproj:
2832         * bindings/js/JSCSSStyleValueCustom.cpp: Copied from Source/WebCore/css/CSSPaintCallback.h.
2833         (WebCore::toJSNewlyCreated):
2834         (WebCore::toJS):
2835         * bindings/js/WebCoreBuiltinNames.h:
2836         * css/CSSPaintCallback.h:
2837         * css/CSSPaintCallback.idl:
2838         * css/CSSPaintImageValue.cpp:
2839         (WebCore::CSSPaintImageValue::image):
2840         * css/CSSPaintImageValue.h:
2841         * css/CSSPaintSize.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
2842         (WebCore::CSSPaintSize::create):
2843         (WebCore::CSSPaintSize::width const):
2844         (WebCore::CSSPaintSize::height const):
2845         (WebCore::CSSPaintSize::CSSPaintSize):
2846         * css/CSSPaintSize.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
2847         * css/parser/CSSPropertyParserHelpers.cpp:
2848         (WebCore::CSSPropertyParserHelpers::consumeCustomPaint):
2849         * css/typedom/CSSNumericValue.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
2850         * css/typedom/CSSNumericValue.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
2851         * css/typedom/CSSStyleValue.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
2852         (WebCore::CSSStyleValue::isUnitValue):
2853         (WebCore::CSSStyleValue::isUnparsedValue):
2854         * css/typedom/CSSStyleValue.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
2855         * css/typedom/CSSUnitValue.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
2856         * css/typedom/CSSUnitValue.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
2857         * css/typedom/CSSUnparsedValue.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
2858         * css/typedom/CSSUnparsedValue.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
2859         * css/typedom/StylePropertyMapReadOnly.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
2860         (WebCore::StylePropertyMapReadOnly::create):
2861         (WebCore::StylePropertyMapReadOnly::get):
2862         (WebCore::StylePropertyMapReadOnly::StylePropertyMapReadOnly):
2863         * css/typedom/StylePropertyMapReadOnly.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
2864         * platform/graphics/CustomPaintImage.cpp:
2865         (WebCore::CustomPaintImage::CustomPaintImage):
2866         (WebCore::CustomPaintImage::doCustomPaint):
2867         * platform/graphics/CustomPaintImage.h:
2868
2869 2018-11-07  Brent Fulgham  <bfulgham@apple.com>
2870
2871         Provide better Font fallbacks for DirectX backend
2872         https://bugs.webkit.org/show_bug.cgi?id=191412
2873         <rdar://problem/45899207>
2874
2875         Reviewed by Zalan Bujtas.
2876
2877         Registration of custom fonts through GDI are not always visible
2878         through the DirectWrite/GDI bridging layer. If a font is not located,
2879         locate the closest matching avialable font rather than failing.
2880
2881         * platform/graphics/FontPlatformData.h:
2882         * platform/graphics/win/FontCustomPlatformData.cpp:
2883         (WebCore::FontCustomPlatformData::fontPlatformData): Use new font
2884         fallback logic.
2885         * platform/graphics/win/FontPlatformDataDirect2D.cpp:
2886         (WebCore::FontPlatformData::platformDataInit):
2887         (WebCore::FontPlatformData::platformIsEqual const):
2888         (WebCore::FontPlatformData::createFallbackFont): Added.
2889
2890 2018-11-07  Zalan Bujtas  <zalan@apple.com>
2891
2892         Click and touch event listeners on the body don't work
2893         https://bugs.webkit.org/show_bug.cgi?id=191392
2894         <rdar://problem/5844416>
2895
2896         Reviewed by Simon Fraser.
2897
2898         Remove the old quirk of ignoring onclick handlers on the body and beyond.
2899
2900         Test: fast/events/click-handler-on-body-simple.html
2901
2902         * page/ios/FrameIOS.mm:
2903         (WebCore::Frame::nodeRespondingToClickEvents):
2904
2905 2018-11-07  Brent Fulgham  <bfulgham@apple.com>
2906
2907         [Windows][DirectX] Update canvas code to pass more tests
2908         https://bugs.webkit.org/show_bug.cgi?id=191337
2909         <rdar://problem/45878801>
2910
2911         Reviewed by Dean Jackson.
2912
2913         Update the Direct2D code paths to comply with our canvas tests, improving the
2914         the test results scores to 579/770.
2915
2916         PathDirect2D was updated with an implementation of 'addArcTo' based on work by
2917         Dirk Schulze <vbs85@gmx.de> (see https://hg.mozilla.org/mozilla-central/rev/b116b49459f8).
2918
2919         Tests: canvas/philip/tests
2920
2921         * platform/graphics/ImageBuffer.cpp:
2922         (WebCore::ImageBuffer::createCompatibleBuffer): Direct2D needs access to the graphics
2923         context to create the buffer.
2924         * platform/graphics/ImageBuffer.h:
2925         * platform/graphics/Path.h:
2926         * platform/graphics/win/GraphicsContextDirect2D.cpp:
2927         (WebCore::GraphicsContext::drawPattern): Flush needed.
2928         (WebCore::GraphicsContext::drawRect): Ditto.
2929         (WebCore::GraphicsContextPlatformPrivate::setMiterLimit): Correct for Direct2D definition of miter limit.
2930         (WebCore::GraphicsContextPlatformPrivate::strokeStyleProperties const): Added helper function.
2931         (WebCore::GraphicsContextPlatformPrivate::recomputeStrokeStyle): Update for new helper.
2932         (WebCore::GraphicsContext::drawLine): Ditto.
2933         (WebCore::drawWithShadowHelper): Ditto.
2934         (WebCore::GraphicsContext::fillRect): Add flush.
2935         (WebCore::GraphicsContext::platformFillRoundedRect): Update for helper.
2936         (WebCore::GraphicsContext::clipPath): Add flush.
2937         (WebCore::GraphicsContext::strokeRect): Ditto.
2938         (WebCore::GraphicsContext::drawLineForText): Update for upstream changes.
2939         (WebCore::GraphicsContext::drawLinesForText): Ditto.
2940         * platform/graphics/win/GraphicsContextPlatformPrivateDirect2D.h:
2941         * platform/graphics/win/ImageBufferDirect2D.cpp:
2942         (WebCore::createCroppedImageIfNecessary): Add missing implementations.
2943         (WebCore::createBitmapImageAfterScalingIfNeeded): Ditto.
2944         (WebCore::ImageBuffer::copyImage const): Ditto.
2945         (WebCore::ImageBuffer::sinkIntoImage): Ditto.
2946         (WebCore::ImageBuffer::fastCopyImageMode): Ditto.
2947         (WebCore::ImageBuffer::sinkIntoNativeImage): Ditto.
2948         (WebCore::ImageBuffer::copyNativeImage const): Ditto.
2949         * platform/graphics/win/PathDirect2D.cpp:
2950         (WebCore::Path::operator=):
2951         (WebCore::Path::drawDidComplete): This should never have been const.
2952         It manipulates the path!
2953         (WebCore::Path::transform): Properly transform existing geometries.
2954         (WebCore::Path::openFigureAtCurrentPointIfNecessary): Added.
2955         (WebCore::Path::moveTo):
2956         (WebCore::Path::addLineTo): Make sure figure starts at a valid point.
2957         (WebCore::Path::addQuadCurveTo): Ditto.
2958         (WebCore::Path::addBezierCurveTo): Ditto.
2959         (WebCore::Path::addArcTo): Add implementation.
2960         (WebCore::Path::closeSubpath):
2961         (WebCore::drawArcSection):
2962         (WebCore::Path::addArc): Update to build large arcs out of small arc segments. If the
2963         arc is effectively a complete circle, use the ellipse drawing routines.
2964         (WebCore::Path::addRect): Make sure we start at a valid starting point.
2965         (WebCore::Path::addEllipse): Correct for definition of D2D ellipse.
2966         (WebCore::Path::drawDidComplete const): Deleted.
2967         * platform/graphics/win/SimpleFontDataDirect2D.cpp:
2968         (WebCore::Font::platformWidthForGlyph const):
2969         * rendering/svg/RenderSVGResourceClipper.cpp:
2970         (WebCore::RenderSVGResourceClipper::applyClippingToContext):
2971         * rendering/svg/RenderSVGResourceFilter.cpp:
2972         (WebCore::RenderSVGResourceFilter::applyResource):
2973         * rendering/svg/RenderSVGResourceMasker.cpp:
2974         (WebCore::RenderSVGResourceMasker::applyResource):
2975         * rendering/svg/SVGRenderingContext.cpp:
2976         (WebCore::SVGRenderingContext::createImageBuffer):
2977         * rendering/svg/SVGRenderingContext.h:
2978
2979 2018-11-07  Wenson Hsieh  <wenson_hsieh@apple.com>
2980
2981         Add an editing command for creating and inserting child lists
2982         https://bugs.webkit.org/show_bug.cgi?id=191335
2983         <rdar://problem/45814050>
2984
2985         Reviewed by Ryosuke Niwa.
2986
2987         Currently, insertOrderedList and insertUnorderedList only toggle or change list state (i.e. if the selection is
2988         in an ordered or unordered list, reinserting the same list type removes the current list, and inserting a
2989         different list type changes the enclosing list).
2990
2991         However, for certain internal clients (e.g. Mail), if the start of the selection is enclosed by a list item, we
2992         instead create a new list item and insert it after the enclosing list item, and then create a new list within
2993         that list item. Currently, this logic is implemented in Mail for legacy-WebKit-based Mail compose. This patch
2994         brings this logic into WebKit in the form of a new editing command.
2995
2996         Tests: editing/execCommand/insert-nested-lists-in-table.html
2997                editing/execCommand/insert-nested-lists-with-pre.html
2998                editing/execCommand/insert-nested-lists.html
2999
3000         * Sources.txt:
3001         * WebCore.xcodeproj/project.pbxproj:
3002         * editing/Editor.cpp:
3003         (WebCore::Editor::insertOrderedList):
3004         (WebCore::Editor::insertUnorderedList):
3005         * editing/EditorCommand.cpp:
3006         (WebCore::executeInsertOrderedList):
3007         (WebCore::executeInsertUnorderedList):
3008         (WebCore::executeInsertNestedUnorderedList):
3009         (WebCore::executeInsertNestedOrderedList):
3010         (WebCore::createCommandMap):
3011         * editing/IndentOutdentCommand.cpp:
3012         (WebCore::IndentOutdentCommand::outdentParagraph):
3013         * editing/InsertListCommand.cpp:
3014         (WebCore::InsertListCommand::doApply):
3015         (WebCore::InsertListCommand::editingAction const):
3016         * editing/InsertListCommand.h:
3017
3018         Change a couple of `enum`s into `enum class`es.
3019
3020         * editing/InsertNestedListCommand.cpp: Added.
3021         (WebCore::InsertNestedListCommand::insertUnorderedList):
3022         (WebCore::InsertNestedListCommand::insertOrderedList):
3023         (WebCore::InsertNestedListCommand::doApply):
3024         * editing/InsertNestedListCommand.h: Added.
3025
3026         Add a new edit command to insert a new list (as a child of any containing list). If the start of the selection
3027         is in a list item, we create a new list item, move the selection into the list item, and increment its list
3028         level; otherwise, simply fall back to inserting a list.
3029
3030         * editing/ModifySelectionListLevel.cpp:
3031         (WebCore::IncreaseSelectionListLevelCommand::doApply):
3032         (WebCore::IncreaseSelectionListLevelCommand::increaseSelectionListLevel):
3033         (WebCore::IncreaseSelectionListLevelCommand::increaseSelectionListLevelOrdered):
3034         (WebCore::IncreaseSelectionListLevelCommand::increaseSelectionListLevelUnordered):
3035         * editing/ModifySelectionListLevel.h:
3036
3037         Expose this constructor, allowing other edit commands to change selection list level as a composite edit
3038         command. Also, change an `enum` into an `enum class`.
3039
3040         (WebCore::IncreaseSelectionListLevelCommand::create):
3041
3042 2018-11-07  Chris Dumez  <cdumez@apple.com>
3043
3044         ASSERT(renderer()) under HTMLTextAreaElement::updateValue()
3045         https://bugs.webkit.org/show_bug.cgi?id=191391
3046         <rdar://problem/34219633>
3047
3048         Reviewed by Geoffrey Garen.
3049
3050         Update HTMLTextFormControlElement::didEditInnerTextValue() to not call subtreeHasChanged()
3051         if the element has no renderer, similarly to what is already done in
3052         HTMLTextFormControlElement::setRangeText() and HTMLInputElement::setEditingValue().
3053
3054         Test: editing/inserting/inset-html-textarea-without-renderer.html
3055
3056         * html/HTMLTextFormControlElement.cpp:
3057         (WebCore::HTMLTextFormControlElement::didEditInnerTextValue):
3058
3059 2018-11-07  Youenn Fablet  <youenn@apple.com>
3060
3061         Allow setting RTCRtpTransceiver.direction
3062         https://bugs.webkit.org/show_bug.cgi?id=191346
3063
3064         Reviewed by Eric Carlson.
3065
3066         Remove readonly from the direction attribute.
3067         Keep setDirection for now as it is in use, with an intent to remove it.
3068         Driven-by fix as in some cases, the rtc source track might be changed
3069         and current implementation is not expecting that.
3070         In such a case, stop observing the old track (which should no longer call us)
3071         before observing the new one.
3072
3073         Covered by rebased tests.
3074
3075         * Modules/mediastream/RTCRtpTransceiver.idl:
3076         * platform/mediastream/RealtimeIncomingAudioSource.cpp:
3077         (WebCore::RealtimeIncomingAudioSource::setSourceTrack):
3078         * platform/mediastream/RealtimeIncomingVideoSource.cpp:
3079         (WebCore::RealtimeIncomingVideoSource::setSourceTrack):
3080
3081 2018-11-07  Sihui Liu  <sihui_liu@apple.com>
3082
3083         RELEASE_ASSERT(!m_hardClosedForUserDelete) fails in WebCore::IDBServer::UniqueIDBDatabase::invokeOperationAndTransactionTimer
3084         https://bugs.webkit.org/show_bug.cgi?id=191326
3085         <rdar://problem/45769269>
3086
3087         Reviewed by Geoffrey Garen.
3088
3089         UniqueIDBDatabase should ignore incoming requests or operations when it is already marked as
3090         hardClosedForUserDelete.
3091
3092         Test: IndexedDB.IndexedDBUserDelete.
3093
3094         * Modules/indexeddb/server/UniqueIDBDatabase.cpp:
3095         (WebCore::IDBServer::UniqueIDBDatabase::didPerformActivateTransactionInBackingStore):
3096         (WebCore::IDBServer::UniqueIDBDatabase::immediateCloseForUserDelete):
3097         * Modules/indexeddb/server/UniqueIDBDatabaseConnection.cpp:
3098         (WebCore::IDBServer::UniqueIDBDatabaseConnection::establishTransaction):
3099         (WebCore::IDBServer::UniqueIDBDatabaseConnection::deleteTransaction):
3100         * Modules/indexeddb/server/UniqueIDBDatabaseConnection.h:
3101         * Modules/indexeddb/server/UniqueIDBDatabaseTransaction.cpp:
3102         (WebCore::IDBServer::UniqueIDBDatabaseTransaction::commit):
3103
3104 2018-11-07  Myles C. Maxfield  <mmaxfield@apple.com>
3105
3106         Positioned text underline can look like a strike-through
3107         https://bugs.webkit.org/show_bug.cgi?id=191341
3108
3109         Reviewed by Simon Fraser.
3110
3111         We should just clamp the value so it can't go above the baseline.
3112
3113         We shouldn't do this at parse time because it's totally reasonable for text-underline-position: under to want
3114         a negative text-underline-offset. Instead, we just do it at used value time.
3115
3116         Test: fast/css3-text/css3-text-decoration/text-underline-negative.html
3117
3118         * style/InlineTextBoxStyle.cpp:
3119         (WebCore::computeUnderlineOffset):
3120
3121 2018-11-07  Chris Dumez  <cdumez@apple.com>
3122
3123         Unreviewed, fix iOS build with recent SDKs.
3124
3125         * platform/network/cocoa/ResourceResponseCocoa.mm:
3126         (WebCore::ResourceResponse::platformCertificateInfo const):
3127
3128 2018-11-07  Chris Dumez  <cdumez@apple.com>
3129
3130         Unreviewed, fix iOS build with recent SDKs.
3131
3132         * platform/network/cocoa/ResourceResponseCocoa.mm:
3133         (WebCore::ResourceResponse::platformCertificateInfo const):
3134
3135 2018-11-07  Myles C. Maxfield  <mmaxfield@apple.com>
3136
3137         Dotted underlines that skip descenders are invisible
3138         https://bugs.webkit.org/show_bug.cgi?id=191403
3139
3140         Reviewed by Simon Fraser.
3141
3142         Turns out our underline bounding boxes had negative width. When drawing the full
3143         underline that was fine because it was handled by the 2D graphics engine, but when
3144         we try to split up the box into dots, our "for" loop was taking 0 iterations (because
3145         the end was before the start).
3146
3147         Test: fast/css3-text/css3-text-decoration/text-underline-style.html
3148
3149         * platform/graphics/cairo/GraphicsContextCairo.cpp:
3150         (WebCore::GraphicsContext::drawLineForText):
3151         * platform/graphics/cg/GraphicsContextCG.cpp:
3152         (WebCore::GraphicsContext::drawLineForText):
3153         * platform/graphics/win/GraphicsContextDirect2D.cpp:
3154         (WebCore::GraphicsContext::drawLineForText):
3155
3156 2018-11-07  Andy Estes  <aestes@apple.com>
3157
3158         Crash in WebCore::PaymentRequest::canMakePayment when Apple Pay payment method data is missing
3159         https://bugs.webkit.org/show_bug.cgi?id=191331
3160
3161         Reviewed by Alexey Proskuryakov.
3162
3163         Apple Pay requires merchants specify an ApplePayRequest (which contains several required
3164         fields) as payment method data when constructing a new PaymentRequest. If the
3165         ApplePayRequest is missing required fields, or is missing entirely, canMakePayment() should
3166         resolve to false.
3167
3168         We would properly resolve to false when an ApplePayRequest was specified with missing
3169         required fields, but we would crash when the ApplePayRequest was missing entirely.
3170
3171         This patch fixes the crash by checking for an empty JSValue before trying to convert it to
3172         an ApplePayRequest struct. Because we stringify ApplePayRequests in the PaymentRequest
3173         constructor then parse them again in canMakePayments, an undefined or null payment method
3174         data stringifies to a null String, which then parses to an empty JSValue.
3175
3176         Added test case to http/tests/paymentrequest/payment-request-canmakepayment-method.https.html.
3177
3178         * Modules/applepay/paymentrequest/ApplePayPaymentHandler.cpp:
3179         (WebCore::ApplePayPaymentHandler::convertData):
3180         * Modules/paymentrequest/PaymentRequest.cpp:
3181         (WebCore::PaymentRequest::canMakePayment):
3182
3183 2018-11-07  Simon Fraser  <simon.fraser@apple.com>
3184
3185         Revert 237849: it breaks MotionMark
3186         https://bugs.webkit.org/show_bug.cgi?id=191398
3187
3188         Reviewed by Simon Fraser.
3189
3190         This change broke painting while the test is running.
3191
3192         * rendering/RenderLayerCompositor.cpp:
3193         (WebCore::RenderLayerCompositor::updateScrollCoordinatedStatus):
3194
3195 2018-11-07  Don Olmstead  <don.olmstead@sony.com>
3196
3197         Simplify macros in platform
3198         https://bugs.webkit.org/show_bug.cgi?id=191378
3199
3200         Reviewed by Michael Catanzaro.
3201
3202         No new tests. No change in behavior.
3203
3204         Modifies a few cases where a platform macro could be simplified.
3205
3206         * page/EventHandler.cpp:
3207         * platform/network/DNS.h:
3208         * platform/network/NetworkStateNotifier.h:
3209
3210 2018-11-07  Simon Fraser  <simon.fraser@apple.com>
3211
3212         Some WK1 repaint tests are flakey
3213         https://bugs.webkit.org/show_bug.cgi?id=190627
3214
3215         Reviewed by Zalan Bujtas.
3216
3217         Repaint tracking in GraphicsLayerCA was sensitive to whether there were already dirty
3218         rects on the layer, since tracking happened after checks against existing dirty rects.
3219         This caused some WK1 repaint tests to be flakey, since there's no guarantee that
3220         no repaints happen between the last layer flush and a test calling startTrackingRepaints().
3221
3222         Fix by moving the repaint tracking to before the checks against existing dirty rects.
3223         This is more similar to how repaint tracking on FrameView works.
3224
3225         * platform/graphics/ca/GraphicsLayerCA.cpp:
3226         (WebCore::GraphicsLayerCA::setNeedsDisplayInRect):
3227
3228 2018-11-07  Dean Jackson  <dino@apple.com>
3229
3230         [iOS] WebGL leaks exact GPU type
3231         https://bugs.webkit.org/show_bug.cgi?id=191393
3232
3233         Reviewed by Tim Horton.
3234
3235         The fully exposed renderer info shouldn't be necessary
3236         on iOS, where the GPUs and drivers are consistent
3237         enough that people shouldn't need to write code specifically
3238         for them. Reduce the ability to fingerprint by simply
3239         returning "Apple GPU".
3240
3241         The other option would have been to disable the extension,
3242         but I think it might still be useful to know you're on
3243         an iOS device.
3244
3245         Test: fast/canvas/webgl/hide-some-renderer-info.html
3246
3247         * html/canvas/WebGL2RenderingContext.cpp: Return "Apple GPU"
3248         on iOS.
3249         (WebCore::WebGL2RenderingContext::getParameter):
3250         * html/canvas/WebGLRenderingContext.cpp: Ditto.
3251         (WebCore::WebGLRenderingContext::getParameter):
3252
3253 2018-11-07  Daniel Bates  <dabates@apple.com>
3254
3255         Override +[UIKeyboard isInHardwareKeyboardMode] in WebKitTestRunner and DumpRenderTree
3256         https://bugs.webkit.org/show_bug.cgi?id=190141
3257
3258         Reviewed by Darin Adler.
3259
3260         Remove the runtime application check for WebKitTestRunner as we no longer need this
3261         now that WebKitTestRunner and DumpRenderTree override +[UIKeyboard isInHardwareKeyboardMode].
3262
3263         * platform/RuntimeApplicationChecks.h:
3264         * platform/cocoa/RuntimeApplicationChecksCocoa.mm:
3265         (WebCore::IOSApplication::isWebKitTestRunner): Deleted.
3266
3267 2018-11-07  Joseph Pecoraro  <pecoraro@apple.com>
3268
3269         Web Inspector: Fix "Javascript" => "JavaScript" enum in protocol generated objects
3270         https://bugs.webkit.org/show_bug.cgi?id=191340
3271
3272         Reviewed by Devin Rousso.
3273
3274         * inspector/agents/InspectorMemoryAgent.cpp:
3275         (WebCore::InspectorMemoryAgent::collectSample):
3276         Use new enum name.
3277
3278 2018-11-07  Dan Bernstein  <mitz@apple.com>
3279
3280         Add a pseudoclass that matches img elements that are backed by an attachment
3281         https://bugs.webkit.org/show_bug.cgi?id=191286
3282         <rdar://problem/45823554>
3283
3284         Reviewed by Dean Jackson.
3285
3286         Test: fast/css/has-attachment.html
3287
3288         Added a :has-attachment pseudoclass, enabled only when the attachment element is enabled at
3289         runtime, that matches elements (currently only img elements qualify) that have an attachment
3290         element in their shadow DOM.
3291
3292         * css/CSSSelector.cpp:
3293         (WebCore::CSSSelector::selectorText const): Handle the new PseudoClassHasAttachment value.
3294         * css/CSSSelector.h: Added a new value to the PseudoClassType enum.
3295
3296         * css/SelectorChecker.cpp:
3297         (WebCore::SelectorChecker::checkOne const): Check the new PseudoClassHasAttachment value
3298           using the new hasAttachment test function.
3299
3300         * css/SelectorCheckerTestFunctions.h:
3301         (WebCore::hasAttachment): Added. Tests if the element is an img element that has an
3302           attachment.
3303
3304         * css/SelectorPseudoClassAndCompatibilityElementMap.in: Added has-attachment.
3305
3306         * css/parser/CSSParserContext.cpp:
3307         (WebCore::CSSParserContext::CSSParserContext): Initialize new attachmentEnabled member based
3308           on RuntimeEnabledFeatures.
3309         (WebCore::operator==): Compare new attachmentEnabled member.
3310         * css/parser/CSSParserContext.h:
3311         (WebCore::CSSParserContextHash::hash): Include new attachmentEnabled member.
3312
3313         * css/parser/CSSSelectorParser.cpp:
3314         (WebCore::CSSSelectorParser::consumePseudo): Reject :has-attachment if the attachment
3315           element is not enabled.
3316
3317         * cssjit/SelectorCompiler.cpp:
3318         (WebCore::SelectorCompiler::addPseudoClassType): Handle PseudoClassHasAttachment.
3319
3320 2018-11-07  Don Olmstead  <don.olmstead@sony.com>
3321
3322         Make generic MainThreadSharedTimer implementation
3323         https://bugs.webkit.org/show_bug.cgi?id=191327
3324
3325         Reviewed by Michael Catanzaro.
3326
3327         No new tests. No change in behavior.
3328
3329         Moves generic implementation for MainThreadSharedTimerGlib.cpp into
3330         MainThreadSharedTimer.cpp.
3331
3332         * platform/MainThreadSharedTimer.cpp:
3333         (WebCore::MainThreadSharedTimer::MainThreadSharedTimer):
3334         (WebCore::MainThreadSharedTimer::setFireInterval):
3335         (WebCore::MainThreadSharedTimer::stop):
3336         (WebCore::MainThreadSharedTimer::invalidate):
3337         * platform/MainThreadSharedTimer.h:
3338         * platform/SourcesGLib.txt:
3339         * platform/glib/MainThreadSharedTimerGLib.cpp: Removed.
3340
3341 2018-11-07  Ali Juma  <ajuma@chromium.org>
3342
3343         IntersectionObserverEntry doesn't keep JS wrappers of rects alive
3344         https://bugs.webkit.org/show_bug.cgi?id=191330
3345
3346         Reviewed by Chris Dumez.
3347
3348         Retain wrappers of each rect in an IntersectionObserverEntry as long as the entry's wrapper
3349         is alive, by adding these wrappers as opaque roots.
3350
3351         Test: intersection-observer/intersection-observer-entry-keeps-js-wrappers-of-rects-alive.html
3352
3353         * bindings/js/JSIntersectionObserverEntryCustom.cpp:
3354         (WebCore::JSIntersectionObserverEntry::visitAdditionalChildren):
3355         * dom/DOMRectReadOnly.idl:
3356         * page/IntersectionObserverEntry.h:
3357         (WebCore::IntersectionObserverEntry::rootBounds const): Make this return a raw pointer instead of a RefPtr so that it
3358         can be called in JSIntersectionObserverEntry::visitAdditionalChildren, which can be called from non-main threads.
3359         (WebCore::IntersectionObserverEntry::boundingClientRect const): Ditto.
3360         (WebCore::IntersectionObserverEntry::intersectionRect const): Ditto.
3361
3362 2018-11-07  Simon Fraser  <simon.fraser@apple.com>
3363
3364         TileController::tileSize() should not have side effects
3365         https://bugs.webkit.org/show_bug.cgi?id=191349
3366
3367         Reviewed by Zalan Bujtas.
3368
3369         Calling TileController::tileSize() would recompute a new tile size and set m_tileSizeLocked,
3370         which caused test failures if logging was enabled when running tests.
3371
3372         * platform/graphics/ca/TileController.cpp:
3373         (WebCore::TileController::tileSize const):
3374         (WebCore::TileController::computeTileSize):
3375         * platform/graphics/ca/TileController.h:
3376         * platform/graphics/ca/TileGrid.cpp:
3377         (WebCore::TileGrid::revalidateTiles):
3378
3379 2018-11-07  Charlie Turner  <cturner@igalia.com>
3380
3381         [EME][GStreamer] Ensure key id buffers are present and simplify lifetime management of ClearKey class.
3382         https://bugs.webkit.org/show_bug.cgi?id=191157
3383
3384         Reviewed by Xabier Rodriguez-Calvar.
3385
3386         This is in preparation for moving the clearkey decryptor behind a
3387         new decrypt API in CDMInstance, which will be sent into the