Source/WebCore/ChangeLog

   1 2016-12-09  Antoine Quint  <graouts@apple.com>
   2
   3         [Modern Media Controls] Implement the pageScaleFactor property
   4         https://bugs.webkit.org/show_bug.cgi?id=165660
   5
   6         Reviewed by Dean Jackson.
   7
   8         We implement the pageScaleFactor property on MediaController. This property is set
   9         by the HTMLMediaElement when the page scale factor changes, and we only choose to handle
  10         it on iOS by setting controlsDependOnPageScaleFactor to true on the MediaControlsHost.
  11
  12         To do so, we now size the media controls by multiplying the layout size of the media by
  13         the page scale factor, and apply an inverse CSS zoom on the media controls themselves.
  14
  15         Test: media/modern-media-controls/media-controller/media-controller-scale-factor.html
  16
  17         * Modules/modern-media-controls/controls/media-controls.css:
  18         (.media-controls-container,):
  19         * Modules/modern-media-controls/controls/media-controls.js:
  20         (MediaControls.prototype.get scaleFactor):
  21         (MediaControls.prototype.set scaleFactor):
  22         (MediaControls.prototype.commitProperty):
  23         * Modules/modern-media-controls/media/media-controller.js:
  24         (MediaController):
  25         (MediaController.prototype.set pageScaleFactor):
  26         (MediaController.prototype._updateControlsSize):
  27         * html/HTMLMediaElement.cpp:
  28         (WebCore::controllerJSValue):
  29         (WebCore::HTMLMediaElement::setControllerJSProperty):
  30
  31         We no longer hit a JSC assertion when trying to set the pageScaleFactor property before
  32         the JS controller was actually created.
  33
  34 2016-12-09  Daniel Bates  <dabates@apple.com>
  35
  36         Attempt to fix the Mac CMake build following <http://trac.webkit.org/changeset/209549>
  37         (https://bugs.webkit.org/show_bug.cgi?id=165572)
  38
  39         Add directory html/canvas to the list of forwarding headers directories so that
  40         the Legacy WebKit Objective-C DOM binding DOMHTMLCanvasElement.mm can ultimately
  41         include header WebGLContextAttributes.h.
  42
  43         * PlatformMac.cmake:
  44
  45 2016-12-09  Brady Eidson  <beidson@apple.com>
  46
  47         LinkIconCollector refactoring.
  48         https://bugs.webkit.org/show_bug.cgi?id=165657
  49
  50         Reviewed by Dean Jackson.
  51
  52         No new tests (Refactor, no behavior change)
  53
  54         * WebCore.xcodeproj/project.pbxproj:
  55
  56         * html/LinkIconCollector.cpp:
  57         (WebCore::iconSize):
  58         (WebCore::compareIcons):
  59         * html/LinkIconCollector.h:
  60
  61         * platform/LinkIcon.h: Move outside of LinkIconCollector for future expandability.
  62
  63 2016-12-09  Ryan Haddad  <ryanhaddad@apple.com>
  64
  65         Unreviewed, rolling out r209574.
  66
  67         This change broke the Windows build.
  68
  69         Reverted changeset:
  70
  71         "[WebIDL] Remove custom bindings for Geolocation"
  72         https://bugs.webkit.org/show_bug.cgi?id=165625
  73         http://trac.webkit.org/changeset/209574
  74
  75 2016-12-09  Ryosuke Niwa  <rniwa@webkit.org>
  76
  77         Custom Elements from a different document are not customized when created with innerHTML
  78         https://bugs.webkit.org/show_bug.cgi?id=165617
  79
  80         Reviewed by Antti Koivisto.
  81
  82         The bug was caused by a superflous null check on window in createHTMLElementOrFindCustomElementInterface.
  83         Removed the nullcheck to fix the bug.
  84
  85         Test: fast/custom-elements/adopting-from-frameless-document.html
  86
  87         * html/parser/HTMLConstructionSite.cpp:
  88         (WebCore::HTMLConstructionSite::createHTMLElementOrFindCustomElementInterface):
  89
  90 2016-12-09  Daniel Bates  <dabates@apple.com>
  91
  92         [CSP] Policy of window opener not applied to about:blank window
  93         https://bugs.webkit.org/show_bug.cgi?id=165531
  94         <rdar://problem/29426639>
  95
  96         Reviewed by Brent Fulgham.
  97
  98         Fixes an issue where the content security policy of the opener document was not applied to
  99         an about:blank window.
 100
 101         An about:blank window inherits its security origin from its opener document. It should also
 102         copy (inherit) the ContentSecurityPolicy from its opener document. When copying the ContentSecurityPolicy
 103         state from the opener document to the about:blank document we must take care to avoid copying
 104         any upgrade-insecure-request directive because new windows should not inherit it by definition.
 105         With respect to upgrade-insecure-requests, new windows should only inherit the insecure navigation set
 106         from their opener document.
 107
 108         Test: http/tests/security/contentSecurityPolicy/image-blocked-in-about-blank-window.html
 109
 110         * dom/Document.cpp:
 111         (WebCore::Document::initContentSecurityPolicy): Copy the ContentSecurityPolicy state from the
 112         owner document to this document when it inherits its security origin from its owner. An about:blank
 113         window is one example of a document that inherits its security origin from its owner.
 114         * loader/WorkerThreadableLoader.cpp:
 115         (WebCore::WorkerThreadableLoader::MainThreadBridge::MainThreadBridge): Call ContentSecurityPolicy::copyUpgradeInsecureRequestStateFrom()
 116         to copy the upgrade insecure requests state from the owner document to the worker now that
 117         ContentSecurityPolicy::copyStateFrom() no longer does this.
 118         * page/csp/ContentSecurityPolicy.cpp:
 119         (WebCore::ContentSecurityPolicy::copyStateFrom): Do not copy the upgrade insecure request state.
 120         Callers are now responsible for calling ContentSecurityPolicy::copyUpgradeInsecureRequestStateFrom()
 121         to copy this state.
 122         * page/csp/ContentSecurityPolicyDirectiveList.cpp:
 123         (WebCore::ContentSecurityPolicyDirectiveList::parse): Ignore directive upgrade-insecure-requests when
 124         inheriting ContentSecurityPolicy state as this directive as the Upgrade Insecure Requests feature has
 125         its own inheritance semantics that differ from the semantics of copying a ContentSecurityPolicy object.
 126         * xml/XSLTProcessor.cpp:
 127         (WebCore::XSLTProcessor::createDocumentFromSource): Call ContentSecurityPolicy::copyUpgradeInsecureRequestStateFrom()
 128         to copy the upgrade insecure requests state from the original document to the transformed document now
 129         that ContentSecurityPolicy::copyStateFrom() no longer does this.
 130
 131 2016-12-09  Antoine Quint  <graouts@apple.com>
 132
 133         [Modern Media Controls] Remaining time label first appears way to the left
 134         https://bugs.webkit.org/show_bug.cgi?id=165637
 135
 136         Reviewed by Dean Jackson.
 137
 138         We would sometimes see the remaining time label be laid out in an incorrect position
 139         when a video would start playing. This happened because the time label was being committed
 140         from a previous value before, in the same frame, we would call the layout() function
 141         of MacOSInlineMediaControls. This would set the newly computed location for the
 142         remaining time label, but because we would reset the list of dirty properties after
 143         calling all layout functions, the new value set in MacOSInlineMediaControls.layout()
 144         would be disregarded and the wrong, committed value would persist until it was reset
 145         in a much later frame.
 146
 147         We now correctly clear the list of dirty nodes before laying them out, giving all nodes
 148         a chance to become dirty again during layout, and updated again in the next frame.
 149
 150         Test: media/modern-media-controls/layout-node/node-made-dirty-during-layout.html
 151
 152         * Modules/modern-media-controls/controls/layout-node.js:
 153         (performScheduledLayout):
 154
 155 2016-12-09  Per Arne Vollan  <pvollan@apple.com>
 156
 157         Fix compile errors on Windows when building with .proj files.
 158
 159         Unreviewed build fix.
 160
 161         Fix include paths.
 162
 163         * platform/graphics/win/IntPointWin.cpp:
 164         * platform/graphics/win/IntRectWin.cpp:
 165         * platform/graphics/win/IntSizeWin.cpp:
 166
 167 2016-12-09  Csaba Osztrogonác  <ossy@webkit.org>
 168
 169         Unreviewed buildfix after r209570.
 170
 171         * bindings/js/JSDOMGlobalObject.cpp:
 172         (WebCore::JSDOMGlobalObject::addBuiltinGlobals):
 173
 174 2016-12-01  Sergio Villar Senin  <svillar@igalia.com>
 175
 176         [css-grid] Pass Grid as argument to items' placement methods
 177         https://bugs.webkit.org/show_bug.cgi?id=165250
 178
 179         Reviewed by Darin Adler.
 180
 181         In order to constify computeIntrinsicLogicalWidths() it is required to constify
 182         placeItemsOnGrid() first, which is the base method of the grid items' positioning logic. The
 183         first step is to constify all the methods invoked by the latter, which basically means to
 184         pass the Grid as argument to all of them instead of directly using the m_grid attribute from
 185         RenderGrid. As this is an intermediate step, a new const_cast<> was required in the
 186         intrinsic size computation. However it will be promptly removed after the const-ification of
 187         placeItemsOnGrid().
 188
 189         After this, only the methods used by the track sizing algorithm will directly access
 190         m_grid. All those would get a reference to the Grid via GridSizingData, but that's a matter
 191         of a follow up patch.
 192
 193         Apart from that, m_gridIsDirty was removed because it was always too confusing. It was
 194         replaced by Grid's m_needsItemsPlacement which is much more concise. The clearGrid() call
 195         was indeed only forcing a new placement of the grid items.
 196
 197         No new tests as this is a refactoring.
 198
 199         * rendering/RenderGrid.cpp:
 200         (WebCore::RenderGrid::Grid::setNeedsItemsPlacement): New method to flag the Grid whenever
 201         grid items' placement algorithm needs to be run.
 202         (WebCore::RenderGrid::canPerformSimplifiedLayout):
 203         (WebCore::RenderGrid::layoutBlock):
 204         (WebCore::RenderGrid::computeIntrinsicLogicalWidths):
 205         (WebCore::RenderGrid::computeEmptyTracksForAutoRepeat):
 206         (WebCore::RenderGrid::placeItemsOnGrid):
 207         (WebCore::RenderGrid::populateExplicitGridAndOrderIterator): Constified. Got Grid as
 208         argument.
 209         (WebCore::RenderGrid::createEmptyGridAreaAtSpecifiedPositionsOutsideGrid): Ditto.
 210         (WebCore::RenderGrid::placeSpecifiedMajorAxisItemsOnGrid): Ditto.
 211         (WebCore::RenderGrid::placeAutoMajorAxisItemsOnGrid): Ditto.
 212         (WebCore::RenderGrid::placeAutoMajorAxisItemOnGrid): Ditto.
 213         (WebCore::RenderGrid::clearGrid):
 214         (WebCore::RenderGrid::offsetAndBreadthForPositionedChild):
 215         (WebCore::RenderGrid::numTracks):
 216         * rendering/RenderGrid.h:
 217
 218 2016-12-09  Gavin Barraclough  <barraclough@apple.com>
 219
 220         Revert - Add _WKIconLoadingDelegate SPI
 221         https://bugs.webkit.org/show_bug.cgi?id=164894
 222
 223         Unreviewed rollout due to performance regression.
 224
 225         * WebCore.xcodeproj/project.pbxproj:
 226         * dom/Document.cpp:
 227         (WebCore::Document::implicitClose):
 228         * html/LinkIconCollector.cpp:
 229         (WebCore::iconSize):
 230         (WebCore::compareIcons):
 231         * html/LinkIconCollector.h:
 232         * loader/DocumentLoader.cpp:
 233         (WebCore::DocumentLoader::startIconLoading): Deleted.
 234         (WebCore::DocumentLoader::didGetLoadDecisionForIcon): Deleted.
 235         (WebCore::DocumentLoader::finishedLoadingIcon): Deleted.
 236         * loader/DocumentLoader.h:
 237         * loader/FrameLoaderClient.h:
 238         * loader/icon/IconLoader.cpp:
 239         (WebCore::IconLoader::IconLoader):
 240         (WebCore::IconLoader::startLoading):
 241         (WebCore::IconLoader::notifyFinished):
 242         * loader/icon/IconLoader.h:
 243         * platform/LinkIcon.h: Removed.
 244
 245 2016-12-08  Jiewen Tan  <jiewen_tan@apple.com>
 246
 247         [WebCrypto] Remove NoInterfaceObject attribute from SubtleCrypto Interface
 248         https://bugs.webkit.org/show_bug.cgi?id=165629
 249         <rdar://problem/29586203>
 250
 251         Reviewed by Chris Dumez.
 252
 253         Covered by existing tests.
 254
 255         * crypto/SubtleCrypto.idl:
 256
 257 2016-12-08  Keith Rollin  <krollin@apple.com>
 258
 259         FileHandle::printf doesn't build on Windows
 260         https://bugs.webkit.org/show_bug.cgi?id=165642
 261
 262         Reviewed by Ryosuke Niwa.
 263
 264         Fix the build -- vasprintf does not exist on Windows, so make
 265         FileHandle::printf a no-op on that platform.
 266
 267         No new tests -- affected facility is currently only used in testing
 268         code and is not user reachable.
 269
 270         * platform/FileHandle.cpp:
 271         (WebCore::FileHandle::printf):
 272
 273 2016-12-08  Antoine Quint  <graouts@apple.com>
 274
 275         [Modern Media Controls] Exiting fullscreen with paused media shows the start button
 276         https://bugs.webkit.org/show_bug.cgi?id=165640
 277
 278         Reviewed by Dean Jackson.
 279
 280         Like all supporting objects, the StartSupport instance gets recreated when creating controls
 281         for new layout traits, which happens when going from fullscreen to inline. Due to this, the
 282         "_isPlayed" property is not longer set to true, even if the media has played prior to being
 283         paused in fullscreen. We now also check the media's played range to identify whether media
 284         has been played.
 285
 286         * Modules/modern-media-controls/media/start-support.js:
 287         (StartSupport.prototype._shouldShowStartButton):
 288         (StartSupport):
 289
 290 2016-12-08  Myles C. Maxfield  <mmaxfield@apple.com>
 291
 292         ASSERTION FAILED: locale in WebCore::lastHyphenLocation
 293         https://bugs.webkit.org/show_bug.cgi?id=164182
 294
 295         Reviewed by Zalan Bujtas.
 296
 297         Our hyphenation code assumes the locale is in a small set of known locales.
 298         However, web content can put whatever string it wants in the locale field.
 299
 300         Test: fast/text/hyphenation-unknown-locale.html
 301
 302         * platform/text/cf/HyphenationCF.cpp:
 303         (WebCore::lastHyphenLocation):
 304         * rendering/RenderText.cpp:
 305         (WebCore::maxWordFragmentWidth): When computing the minimum preferred width of
 306         an element, we want to know the size of the widest nonbreakable unit. When
 307         hyphenation is enabled, we have to measure the width of each fragment of each
 308         word (where "fragment" is determined by a substring between two successive
 309         hyphenation points, appended with a hyphen character). A hyphenation point at
 310         string offset 0 is irrelevant for this purpose, because it would lead to an
 311         empty fragment. Therefore, skip this situation.
 312
 313 2016-12-08  David Kilzer  <ddkilzer@apple.com>
 314
 315         Always check the return value of pthread_key_create()
 316         <https://webkit.org/b/165274>
 317
 318         Reviewed by Darin Adler.
 319
 320         * platform/ios/wak/WebCoreThread.mm:
 321         (InitThreadContextKey): Call CRASH() if pthread_key_create()
 322         returns an error.  The error code will be stored in a register
 323         available in a crash log, so no need to log the value
 324         explicitly.
 325
 326 2016-12-08  Alex Christensen  <achristensen@webkit.org>
 327
 328         Reduce PassRefPtr use in platform/graphics
 329         https://bugs.webkit.org/show_bug.cgi?id=165636
 330
 331         Reviewed by Dean Jackson.
 332
 333         No change in behavior.
 334
 335         * platform/graphics/GraphicsContext3D.h:
 336         * platform/graphics/mac/DisplayRefreshMonitorMac.h:
 337         (WebCore::DisplayRefreshMonitorMac::create):
 338         * platform/graphics/opengl/GraphicsContext3DOpenGLCommon.cpp:
 339         (WebCore::GraphicsContext3D::createForCurrentGLContext):
 340         (WebCore::GraphicsContext3D::paintRenderingResultsToImageData):
 341
 342 2016-12-08  Keith Rollin  <krollin@apple.com>
 343
 344         Move FileHandle to WebCore FileHandle.h
 345         https://bugs.webkit.org/show_bug.cgi?id=165562
 346
 347         Reviewed by Alex Christensen.
 348
 349         Add FileHandle, a small wrapper around FileSystem facilities to manage
 350         the lifetime of an open file and give the facilities a C++ interface.
 351
 352         No new tests -- affected facility is currently only used in testing
 353         code and is not user reachable.
 354
 355         * CMakeLists.txt:
 356         * WebCore.xcodeproj/project.pbxproj:
 357         * platform/FileHandle.cpp: Added.
 358         (WebCore::FileHandle::FileHandle):
 359         (WebCore::FileHandle::~FileHandle):
 360         (WebCore::FileHandle::operator=):
 361         (WebCore::FileHandle::operator bool):
 362         (WebCore::FileHandle::open):
 363         (WebCore::FileHandle::read):
 364         (WebCore::FileHandle::write):
 365         (WebCore::FileHandle::printf):
 366         (WebCore::FileHandle::close):
 367         * platform/FileHandle.h: Added.
 368
 369 2016-12-07  Ryosuke Niwa  <rniwa@webkit.org>
 370
 371         ASSERTION FAILED: m_items.isEmpty() in CustomElementReactionQueue destructor
 372         https://bugs.webkit.org/show_bug.cgi?id=162029
 373         <rdar://problem/28945851>
 374
 375         Reviewed by Chris Dumez.
 376
 377         The bug was caused by Document::removedLastRef enqueuing disconnectedCallback during a tear down.
 378         Don't enqueue a disconnectedCallback while a document is getting torn down since that should not be
 379         observable to author scripts. The connected, adopted, and attributeChanged callbacks are immune from
 380         this problem since they don't happen during a document destruction.
 381
 382         Note that this was also the case prior to this patch since the disconnectedCallback would have been
 383         added to the current CustomElementReactionQueue which will be destructed without invoking callbacks
 384         (or hit a release assertion added in r208785 and r209426 for now).
 385
 386         Tests: fast/custom-elements/disconnected-callback-in-detached-iframe.html
 387                fast/custom-elements/element-queue-during-document-destruction.html
 388
 389         * dom/CustomElementReactionQueue.cpp:
 390         (WebCore::CustomElementReactionQueue::enqueueConnectedCallbackIfNeeded): Added an assertion that
 391         document's refCount hasn't reached zero yet.
 392         (WebCore::CustomElementReactionQueue::enqueueDisconnectedCallbackIfNeeded): Fixed the bug.
 393         (WebCore::CustomElementReactionQueue::enqueueAdoptedCallbackIfNeeded): Added the same assertion.
 394         (WebCore::CustomElementReactionQueue::enqueueAttributeChangedCallbackIfNeeded): Ditto.
 395
 396 2016-12-08  Daniel Bates  <dabates@apple.com>
 397
 398         Add Strict Mixed Content Checking and Upgrade Insecure Requests to WebKit Feature Status dashboard
 399         https://bugs.webkit.org/show_bug.cgi?id=165635
 400
 401         Rubber-stamped by Ryosuke Niwa.
 402
 403         * features.json:
 404
 405 2016-12-08  Daniel Bates  <dabates@apple.com>
 406
 407         Implement Strict Mixed Content Checking
 408         https://bugs.webkit.org/show_bug.cgi?id=165438
 409         <rdar://problem/26103867>
 410
 411         Reviewed by Brent Fulgham and Andy Estes.
 412
 413         Add support for the CSP directive block-all-mixed-content to enable strict mixed content checking
 414         as per <https://www.w3.org/TR/2016/CR-mixed-content-20160802/#strict-checking> (2 August 2016).
 415
 416         Currently WebKit only blocks blockable content as such content can contaminate the security origin
 417         that loaded it. Optionally-blockable content, including images, would be allowed to load as mixed
 418         content. When strict mixed content checking is enabled all mixed content is blocked. That is, both
 419         blockable and optionally-blockable content will be blocked. A web site can opt into strict mixed
 420         content checking by adding the directive block-all-mixed-content to their content security policy.
 421
 422         Tests: http/tests/security/contentSecurityPolicy/block-all-mixed-content/data-url-iframe-in-main-frame.html
 423                http/tests/security/contentSecurityPolicy/block-all-mixed-content/duplicate-directive.html
 424                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-css-in-iframe-report-only.html
 425                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-css-in-iframe.html
 426                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-css-in-main-frame.html
 427                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-iframe-in-iframe.html
 428                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-iframe-in-main-frame.html
 429                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-blob-url-iframe-in-iframe.html
 430                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-iframe-with-enforced-and-report-policies.html
 431                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-iframe-with-inherited-policy.html
 432                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-iframe.html
 433                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-javascript-url-iframe-in-iframe.html
 434                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-main-frame.html
 435                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-xslt-document-in-iframe-with-inherited-policy.html
 436                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-plugin-in-iframe.html
 437                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-plugin-in-main-frame.html
 438                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-script-in-iframe-with-inherited-policy.html
 439                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-script-in-iframe.html
 440                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-script-in-main-frame.html
 441                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-xhr-asynchronous-in-iframe.html
 442                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-xhr-asynchronous-in-main-frame.html
 443                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-xhr-synchronous-in-iframe.html
 444                http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-xhr-synchronous-in-main-frame.html
 445                http/tests/security/contentSecurityPolicy/block-all-mixed-content/secure-image-after-upgrade-in-iframe.html
 446                http/tests/security/contentSecurityPolicy/block-all-mixed-content/secure-image-after-upgrade-redirect-in-iframe.html
 447
 448         * dom/Document.cpp:
 449         (WebCore::Document::initSecurityContext): Inherit strict mixed content checking mode from parent document.
 450         * dom/SecurityContext.h:
 451         (WebCore::SecurityContext::isStrictMixedContentMode): Added.
 452         (WebCore::SecurityContext::setStrictMixedContentMode): Added.
 453         * loader/DocumentLoader.cpp:
 454         (WebCore::DocumentLoader::willSendRequest): Check mixed content policy with respect to the current frame.
 455         The document in the current frame may have opted into strict mixed content checking or inherited it from
 456         its parent document.
 457         * loader/DocumentWriter.cpp:
 458         (WebCore::DocumentWriter::begin): Inherit the strict mixed content checking mode from the owner document
 459         when loading a JavaScript URL in a frame (e.g. <iframe src="javascript:...">) because such URLs inherit
 460         the security origin of their parent document.
 461         * loader/MixedContentChecker.cpp:
 462         (WebCore::MixedContentChecker::canDisplayInsecureContent): Check the content security policy of the document
 463         and the strict mixed content checking mode bit on the document (in that order) to determine if we are in
 464         strict mode. Block display of insecure content when in strict mode. Modified to take enum AlwaysDisplayInNonStrictMode (defaults
 465         to AlwaysDisplayInNonStrictMode::No) as to whether to allow our current relaxed behavior of displaying insecure
 466         content in non-strict mode.
 467         (WebCore::MixedContentChecker::canRunInsecureContent): Check the content security policy of the document
 468         and the strict mixed content checking mode bit on the document (in that order) to determine if we are in
 469         strict mode. Block running of insecure content when in strict mode.
 470         * loader/MixedContentChecker.h:
 471         * loader/cache/CachedResourceLoader.cpp:
 472         (WebCore::CachedResourceLoader::checkInsecureContent): Always check mixed content policy with respect to
 473         the current frame. The document in the current frame may have opted into strict mixed content checking or
 474         inherited it from its parent document. Also renamed a local variable f to frame to better describe its
 475         purpose.
 476         * page/csp/ContentSecurityPolicy.cpp:
 477         (WebCore::ContentSecurityPolicy::allowRunningOrDisplayingInsecureContent): Added. Iterate through all the
 478         policies and report violations with respect to policies that have directive block-all-mixed-content.
 479         (WebCore::ContentSecurityPolicy::didReceiveHeader): Move logic to set eval() error message from here...
 480         (WebCore::ContentSecurityPolicy::applyPolicyToScriptExecutionContext): ...to here so that we only perform
 481         it once we are ready to apply the CSP policy to the script execution context. Additionally, enable
 482         strict mixed content checking on the script execution context if applicable.
 483         (WebCore::ContentSecurityPolicy::reportViolation): Added overrides that take a string and a directive list
 484         object (ContentSecurityPolicyDirectiveList) for the effective violated directive and its associated directive
 485         list, respectively. We make use of these overrides so as to support reporting block-all-mixed-content
 486         violations, which are not implemented using a ContentSecurityPolicyDirective object as it seemed sufficient
 487         to implement it as a boolean on ContentSecurityPolicyDirectiveList.
 488         * page/csp/ContentSecurityPolicy.h:
 489         * page/csp/ContentSecurityPolicyDirectiveList.cpp:
 490         (WebCore::ContentSecurityPolicyDirectiveList::setBlockAllMixedContentEnabled): Added.
 491         (WebCore::ContentSecurityPolicyDirectiveList::addDirective): Parse the directive block-all-mixed-content.
 492         * page/csp/ContentSecurityPolicyDirectiveList.h:
 493         (WebCore::ContentSecurityPolicyDirectiveList::hasBlockAllMixedContentDirective): Added.
 494         * page/csp/ContentSecurityPolicyDirectiveNames.cpp:
 495         * page/csp/ContentSecurityPolicyDirectiveNames.h: Add constant for "block-all-mixed-content".
 496
 497 2016-12-08  Sam Weinig  <sam@webkit.org>
 498
 499         [WebIDL] Remove custom bindings for Geolocation
 500         https://bugs.webkit.org/show_bug.cgi?id=165625
 501
 502         Reviewed by Alex Christensen.
 503
 504         * CMakeLists.txt:
 505         * DerivedSources.cpp:
 506         * DerivedSources.make:
 507         * WebCore.xcodeproj/project.pbxproj:
 508         * bindings/js/JSBindingsAllInOne.cpp:
 509         Add/remove files.
 510
 511         * Modules/geolocation/GeoNotifier.cpp:
 512         (WebCore::GeoNotifier::GeoNotifier):
 513         (WebCore::GeoNotifier::hasZeroTimeout):
 514         (WebCore::GeoNotifier::startTimerIfNeeded):
 515         * Modules/geolocation/GeoNotifier.h:
 516         (WebCore::GeoNotifier::create):
 517         (WebCore::GeoNotifier::options):
 518         Update to store PositionOptions as a value, and pass the PositionCallback
 519         as a Ref, rather than a RefPtr, since it is not optional.
 520
 521         * Modules/geolocation/Geolocation.cpp:
 522         (WebCore::createGeoposition):
 523         Use auto, to get the good type for Coordinates (Ref).
 524
 525         (WebCore::Geolocation::getCurrentPosition):
 526         (WebCore::Geolocation::watchPosition):
 527         (WebCore::Geolocation::haveSuitableCachedPosition):
 528         (WebCore::Geolocation::startUpdating):
 529         * Modules/geolocation/Geolocation.h:
 530         Update to pass PositionOptions itself now that it is a plain struct.
 531
 532         * Modules/geolocation/Geolocation.idl:
 533         Remove [Custom] extended attribute and add FIXME about nullable annotation.
 534
 535         * Modules/geolocation/Geoposition.h:
 536         Pass/store Coordinates as a Ref.
 537
 538         * Modules/geolocation/PositionOptions.h:
 539         Convert to be a simple struct.
 540
 541         * Modules/geolocation/PositionOptions.idl:
 542         Added.
 543
 544         * bindings/js/JSGeolocationCustom.cpp:
 545         Removed.
 546
 547 2016-12-08  Alex Christensen  <achristensen@webkit.org>
 548
 549         REGRESSION (URL parser): Relative URLs aren’t resolved correctly when the base URL is an applewebdata: URL
 550         https://bugs.webkit.org/show_bug.cgi?id=165621
 551
 552         Reviewed by Dan Bernstein.
 553
 554         Covered by new API tests.
 555
 556         * platform/URLParser.cpp:
 557         (WebCore::URLParser::parse):
 558         URLs with nonspecial schemes and no slash after the host get no slash as the path to maintain compatibility with all browsers.
 559         This was proposed to the URL spec in https://github.com/whatwg/url/issues/148
 560         When such as URL is used as a base URL with a relative path, in order to maintain compatibility with URL::parse we need to prepend
 561         a slash to the path.  For completeness I added tests with a relative path, a relative query, a relative fragment, and a relative empty string,
 562         and because the fate of the spec is unclear in this case, I decided to maintain compatibility with URL::parse in all these cases.
 563
 564 2016-12-06  Filip Pizlo  <fpizlo@apple.com>
 565
 566         Concurrent GC should be stable enough to land enabled on X86_64
 567         https://bugs.webkit.org/show_bug.cgi?id=164990
 568
 569         Reviewed by Geoffrey Garen.
 570
 571         Made WebCore down with concurrent marking by adding some locking and adapting to some new API.
 572
 573         This has new test modes in run-sjc-stress-tests. Also, the way that LayoutTests run is already
 574         a fantastic GC test.
 575
 576         * ForwardingHeaders/heap/DeleteAllCodeEffort.h: Added.
 577         * ForwardingHeaders/heap/LockDuringMarking.h: Added.
 578         * bindings/js/GCController.cpp:
 579         (WebCore::GCController::deleteAllCode):
 580         (WebCore::GCController::deleteAllLinkedCode):
 581         * bindings/js/GCController.h:
 582         * bindings/js/JSDOMBinding.cpp:
 583         (WebCore::getCachedDOMStructure):
 584         (WebCore::cacheDOMStructure):
 585         * bindings/js/JSDOMGlobalObject.cpp:
 586         (WebCore::JSDOMGlobalObject::addBuiltinGlobals):
 587         (WebCore::JSDOMGlobalObject::visitChildren):
 588         * bindings/js/JSDOMGlobalObject.h:
 589         (WebCore::getDOMConstructor):
 590         * bindings/js/JSDOMPromise.cpp:
 591         (WebCore::DeferredPromise::DeferredPromise):
 592         (WebCore::DeferredPromise::clear):
 593         * bindings/js/JSXPathResultCustom.cpp:
 594         (WebCore::JSXPathResult::visitAdditionalChildren):
 595         * dom/EventListenerMap.cpp:
 596         (WebCore::EventListenerMap::clear):
 597         (WebCore::EventListenerMap::replace):
 598         (WebCore::EventListenerMap::add):
 599         (WebCore::EventListenerMap::remove):
 600         (WebCore::EventListenerMap::find):
 601         (WebCore::EventListenerMap::removeFirstEventListenerCreatedFromMarkup):
 602         (WebCore::EventListenerMap::copyEventListenersNotCreatedFromMarkupToTarget):
 603         (WebCore::EventListenerIterator::EventListenerIterator):
 604         * dom/EventListenerMap.h:
 605         (WebCore::EventListenerMap::lock):
 606         * dom/EventTarget.cpp:
 607         (WebCore::EventTarget::visitJSEventListeners):
 608         * dom/EventTarget.h:
 609         (WebCore::EventTarget::visitJSEventListeners): Deleted.
 610         * dom/Node.cpp:
 611         (WebCore::Node::eventTargetDataConcurrently):
 612         (WebCore::Node::ensureEventTargetData):
 613         (WebCore::Node::clearEventTargetData):
 614         * dom/Node.h:
 615         * page/MemoryRelease.cpp:
 616         (WebCore::releaseCriticalMemory):
 617         * page/cocoa/MemoryReleaseCocoa.mm:
 618         (WebCore::jettisonExpensiveObjectsOnTopLevelNavigation):
 619         (WebCore::registerMemoryReleaseNotifyCallbacks):
 620
 621 2016-12-08  Dave Hyatt  <hyatt@apple.com>
 622
 623         [CSS Parser] Add strict checking for right parens to selector functions like :matches, :not etc.
 624         https://bugs.webkit.org/show_bug.cgi?id=165624
 625
 626         Reviewed by Dean Jackson.
 627
 628         Unskipped fast/selectors/invalid-functional-pseudo-class.html.
 629
 630         * css/parser/CSSSelectorParser.cpp:
 631         (WebCore::CSSSelectorParser::consumePseudo):
 632         Bail if we don't actually consume a right parenthesis.
 633
 634 2016-12-08  Jeremy Jones  <jeremyj@apple.com>
 635
 636         Exit fullscreen on escape key down, not up.
 637         https://bugs.webkit.org/show_bug.cgi?id=165618
 638         rdar://problem/29581688
 639
 640         This problem was introduced by https://bugs.webkit.org/show_bug.cgi?id=165416
 641         The fix to only exit fullscreen on key down, not key up.
 642
 643         Reviewed by Jer Noble.
 644
 645         * page/EventHandler.cpp:
 646         (WebCore::EventHandler::keyEvent):
 647
 648 2016-12-08  Dave Hyatt  <hyatt@apple.com>
 649
 650         [CSS Parser] Unskip background parsing tests
 651         https://bugs.webkit.org/show_bug.cgi?id=165619
 652
 653         Reviewed by Zalan Bujtas.
 654
 655         Change background-position so that the origin being set is tracked
 656         on a per-x/y basis instead of as a single boolean. The single boolean
 657         caused lousy computed style dumping, since we would switch into verbose
 658         mode for both x and y, even if only one of them had a pair set for it.
 659
 660         * css/CSSComputedStyleDeclaration.cpp:
 661         (WebCore::createPositionListForLayer):
 662         * css/StyleBuilderConverter.h:
 663         (WebCore::StyleBuilderConverter::convertPositionComponent):
 664         * rendering/style/FillLayer.cpp:
 665         (WebCore::FillLayer::FillLayer):
 666         (WebCore::FillLayer::operator=):
 667         (WebCore::FillLayer::fillUnsetProperties):
 668         * rendering/style/FillLayer.h:
 669         (WebCore::FillLayer::isBackgroundXOriginSet):
 670         (WebCore::FillLayer::isBackgroundYOriginSet):
 671         (WebCore::FillLayer::setBackgroundXOrigin):
 672         (WebCore::FillLayer::setBackgroundYOrigin):
 673         (WebCore::FillLayer::clearXPosition):
 674         (WebCore::FillLayer::clearYPosition):
 675         (WebCore::FillLayer::isBackgroundOriginSet): Deleted.
 676
 677 2016-12-08  Jeremy Jones  <jeremyj@apple.com>
 678
 679         pointer lock should exit when a js alert is shown
 680         https://bugs.webkit.org/show_bug.cgi?id=165423
 681         rdar://problem/29430803
 682
 683         Reviewed by Jer Noble.
 684
 685         Test: pointer-lock/lock-lost-on-alert.html
 686
 687         * page/DOMWindow.cpp:
 688         (WebCore::DOMWindow::alert):
 689         (WebCore::DOMWindow::confirm):
 690         (WebCore::DOMWindow::prompt):
 691
 692 2016-12-07  Sam Weinig  <sam@webkit.org>
 693
 694         [WebIDL] WebGLContextAttributes should be a dictionary
 695         https://bugs.webkit.org/show_bug.cgi?id=165572
 696
 697         Reviewed by Dean Jackson.
 698
 699         * CMakeLists.txt:
 700         * WebCore.xcodeproj/project.pbxproj:
 701         Remove unused files.
 702
 703         * bindings/js/JSHTMLCanvasElementCustom.cpp:
 704         (WebCore::JSHTMLCanvasElement::getContext):
 705         (WebCore::attributesFor3DContext): Deleted.
 706         Re-work getContext to use JSDOMConvert and use the WebGLContextAttributes dictionary.
 707
 708         * bindings/scripts/CodeGeneratorJS.pm:
 709         (NativeToJSValueDOMConvertNeedsState):
 710         (NativeToJSValueDOMConvertNeedsGlobalObject):
 711         Add missing inclusion of dictionaries.
 712
 713         * html/HTMLCanvasElement.cpp:
 714         (WebCore::maxActivePixelMemory):
 715         (WebCore::HTMLCanvasElement::getContext):
 716         (WebCore::HTMLCanvasElement::is2dType):
 717         (WebCore::HTMLCanvasElement::getContext2d):
 718         (WebCore::shouldEnableWebGL):
 719         (WebCore::HTMLCanvasElement::is3dType):
 720         (WebCore::HTMLCanvasElement::getContextWebGL):
 721         * html/HTMLCanvasElement.h:
 722         Refactor context creation of 2d and WebGL contexts into their own functions. Make the
 723         WebGL one take a WebGLContextAttributes struct.
 724
 725         * html/HTMLCanvasElement.idl:
 726         No functionality change. Make IDL match the latest spec.
 727
 728         * html/canvas/CanvasContextAttributes.cpp: Removed.
 729         * html/canvas/CanvasContextAttributes.h: Removed.
 730         Remove unneeded base class.
 731
 732         * html/canvas/WebGLContextAttributes.cpp: Removed.
 733         * html/canvas/WebGLContextAttributes.h:
 734         Instead of a class, we can use a type alias of GraphicsContext3D::Attributes for
 735         WebGLContextAttributes.
 736
 737         * html/canvas/WebGLContextAttributes.idl:
 738         Convert to a dictionary.
 739
 740         * html/canvas/WebGL2RenderingContext.cpp:
 741         * html/canvas/WebGL2RenderingContext.h:
 742         * html/canvas/WebGLRenderingContextBase.cpp:
 743         (WebCore::WebGLRenderingContextBase::create):
 744         (WebCore::WebGLRenderingContextBase::WebGLRenderingContextBase):
 745         (WebCore::WebGLRenderingContextBase::clearIfComposited):
 746         (WebCore::WebGLRenderingContextBase::getContextAttributes):
 747         (WebCore::WebGLRenderingContextBase::applyStencilTest):
 748         * html/canvas/WebGLRenderingContextBase.h:
 749         Update to WebGLContextAttributes by value, since it is a simple struct now. Update getContextAttributes
 750         to return a std::optional<WebGLContextAttributes>.
 751
 752         * html/canvas/WebGLRenderingContextBase.idl:
 753         Make getContextAttributes return a nullable WebGLContextAttributes, as per spec.
 754
 755         * platform/graphics/GraphicsContext3D.h:
 756         Sort attributes into categories, one that match WebGLContextAttributes, and the remainder.
 757
 758 2016-12-08  Zalan Bujtas  <zalan@apple.com>
 759
 760         Do not create multicolumn context for certain type of renderers.
 761         https://bugs.webkit.org/show_bug.cgi?id=135601
 762         <rdar://problem/27686300>
 763
 764         Reviewed by David Hyatt.
 765
 766         Certain type of renderers including forms should not initialize multicolumn context for
 767         their inner (shadow) content. It just does not makes sense and is inline with what other browsers do.
 768
 769         Test: fast/multicol/no-multicol-for-textareas.html
 770
 771         * rendering/RenderBlockFlow.cpp:
 772         (WebCore::RenderBlockFlow::willCreateColumns):
 773         (WebCore::RenderBlockFlow::requiresColumns):
 774         * rendering/RenderBlockFlow.h:
 775
 776 2016-12-08  Dave Hyatt  <hyatt@apple.com>
 777
 778         [CSS Parser] REGRESSION: Values of 0 should not be allowed for -webkit-aspect-ratio
 779         https://bugs.webkit.org/show_bug.cgi?id=165607
 780
 781         Reviewed by Simon Fraser.
 782
 783         Unskipped fast/css/aspect-ratio-parsing-tests.html.
 784
 785         * css/parser/CSSPropertyParser.cpp:
 786         (WebCore::consumeWebkitAspectRatio):
 787         Don't allow 0 as either the left or right value of the ratio.
 788
 789 2016-12-08  Antti Koivisto  <antti@apple.com>
 790
 791         ::after and ::before don't work on :host
 792         https://bugs.webkit.org/show_bug.cgi?id=164675
 793         <rdar://problem/29231874>
 794
 795         Reviewed by Andreas Kling.
 796
 797         We didn't allow combining :host with any other selectors. Pseudo elements should be allowed.
 798
 799         Test: fast/shadow-dom/shadow-host-with-before-after.html
 800
 801         * css/ElementRuleCollector.cpp:
 802         (WebCore::ElementRuleCollector::matchAuthorRules):
 803
 804             Allow pseudo element matching.
 805
 806         (WebCore::ElementRuleCollector::matchHostPseudoClassRules):
 807
 808             Use the normal collectMatchingRulesForList path instead of a direct call to SelectorChecker::matchHostPseudoClass.
 809             This path supports pseudo elements.
 810
 811             Pass the information that we are matching :host rules with m_isMatchingHostPseudoClass bit
 812             similarly to how this is done with slots.
 813
 814         (WebCore::ElementRuleCollector::ruleMatches):
 815         * css/ElementRuleCollector.h:
 816         * css/RuleSet.cpp:
 817
 818             Collect :host rules similarly to other rules.
 819
 820         (WebCore::RuleSet::addRule):
 821         * css/SelectorChecker.cpp:
 822
 823             Rename didMoveToShadowHost->mayMatchHostPseudoClass for clarity.
 824
 825         (WebCore::SelectorChecker::match):
 826
 827             Enable :host matching in the current context immediately if needed.
 828
 829         (WebCore::SelectorChecker::matchHostPseudoClass):
 830
 831             The check for illegal combinations with :host is now done in checkOne().
 832
 833         (WebCore::localContextForParent):
 834         (WebCore::SelectorChecker::checkOne):
 835
 836             In context where :host matches nothing else can match, except pseudo elements.
 837
 838         * css/SelectorChecker.h:
 839
 840 2016-12-08  Alex Christensen  <achristensen@webkit.org>
 841
 842         Fix CMake build.
 843
 844         * PlatformMac.cmake:
 845         Add PerformanceLoggingCocoa.mm after r209181.
 846
 847 2016-12-08  Per Arne Vollan  <pvollan@apple.com>
 848
 849         [Win] Some versions of Cygwin Perl generate incorrect C++ code.
 850         https://bugs.webkit.org/show_bug.cgi?id=165524
 851         rdar://problem/29508043
 852
 853         Reviewed by Daniel Bates.
 854
 855         Perl strings with variables seems to be handled differently, depending on which version of
 856         Cygwin Perl you are running. To make sure they are handled identically, specify variables
 857         in strings as "${var}" instead of "$var".
 858
 859         * bindings/scripts/CodeGeneratorJS.pm:
 860         (GenerateHeader):
 861         (GenerateImplementation):
 862
 863 2016-12-08  Alex Christensen  <achristensen@webkit.org>
 864
 865         Fix EFL build after r209526
 866         https://bugs.webkit.org/show_bug.cgi?id=165594
 867
 868         * css/parser/CSSPropertyParserHelpers.cpp:
 869         (WebCore::CSSPropertyParserHelpers::consumePositiveIntegerRaw):
 870         nullptr is being automatically converted to false everywhere but EFL.
 871         It should be false.  This function returns a bool.
 872
 873 2016-12-07  Dave Hyatt  <hyatt@apple.com>
 874
 875         REGRESSION: font shorthand parsing is broken
 876         https://bugs.webkit.org/show_bug.cgi?id=165594
 877
 878         Reviewed by Zalan Bujtas.
 879
 880         Unskipping fast/css/font-calculated-value.html.
 881
 882         * css/parser/CSSPropertyParser.cpp:
 883         (WebCore::consumeFontWeight):
 884         The new parser did not support calc() in the font-weight property.
 885         In order to support a positive integer, add new functions to consume
 886         and handle positive integers without making a CSSPrimitiveValue out of
 887         the number. These correspond to consumeNumberRaw but are limited to positive
 888         integers.
 889
 890         (WebCore::CSSPropertyParser::consumeFont):
 891         The implicit bool was incorrectly set to true when properties were
 892         defined because releaseNonNull() cleared the variables being null
 893         checked. Add separate booleans to avoid this.
 894
 895         * css/parser/CSSPropertyParserHelpers.cpp:
 896         (WebCore::CSSPropertyParserHelpers::CalcParser::consumePositiveIntegerRaw):
 897         (WebCore::CSSPropertyParserHelpers::consumePositiveIntegerRaw):
 898         * css/parser/CSSPropertyParserHelpers.h:
 899         Add the new helpers that consume positive integers without creating
 900         CSSPrimitiveValues out of them.
 901
 902 2016-12-07  Jeremy Jones  <jeremyj@apple.com>
 903
 904         Exit pointer lock when page goes into page cache.
 905         https://bugs.webkit.org/show_bug.cgi?id=165425
 906         rdar://problem/29430834
 907
 908         Reviewed by Jer Noble.
 909
 910         Test: pointer-lock/lock-lost-on-navigation.html
 911
 912         Cancel pointer lock when page goes into page cache and add methods to Internals
 913         so it can be tested.
 914
 915         * dom/Document.cpp:
 916         (WebCore::Document::setPageCacheState):
 917         * page/PointerLockController.h:
 918         * testing/Internals.cpp:
 919         (WebCore::Internals::pageHasPendingPointerLock):
 920         (WebCore::Internals::pageHasPointerLock):
 921         * testing/Internals.h:
 922         * testing/Internals.idl:
 923
 924 2016-12-07  Antoine Quint  <graouts@apple.com>
 925
 926         [Modern Media Controls] Tracks panel does not display in the right location in fullscreen and cannot be dismissed
 927         https://bugs.webkit.org/show_bug.cgi?id=165575
 928
 929         Reviewed by Dean Jackson.
 930
 931         We position the tracks panel programmatically based on the current bounds of the fullscreen button in fullscreen.
 932         This allows for the tracks panel to be shown in the right spot even after dragging the controls bar. We also use
 933         the fullscreen media controls as the target for tracking "mousedown" events when the tracks panel is shown in a
 934         fullscreen presentation.
 935
 936         Tests: media/modern-media-controls/tracks-support/tracks-support-show-panel-after-dragging-controls.html
 937                media/modern-media-controls/tracks-support/tracks-support-show-panel-fullscreen.html
 938
 939         * Modules/modern-media-controls/controls/macos-fullscreen-media-controls.css:
 940         (.media-controls.mac.fullscreen .scrubber):
 941         (.media-controls.mac.fullscreen .tracks-panel): Deleted.
 942         * Modules/modern-media-controls/controls/macos-fullscreen-media-controls.js:
 943         (MacOSFullscreenMediaControls.prototype.showTracksPanel):
 944         * Modules/modern-media-controls/controls/tracks-panel.js:
 945         (TracksPanel.prototype.presentInParent):
 946         (TracksPanel.prototype.hide):
 947         (TracksPanel.prototype.get bottomY):
 948         (TracksPanel.prototype.set bottomY):
 949         (TracksPanel.prototype.commitProperty):
 950         (TracksPanel.prototype._mousedownTarget):
 951
 952 2016-12-07  Antoine Quint  <graouts@apple.com>
 953
 954         [Modern Media Controls] Use a small picture-in-picture button in fullscreen
 955         https://bugs.webkit.org/show_bug.cgi?id=165587
 956
 957         Reviewed by Dean Jackson.
 958
 959         Use a newer, smaller asset for picture-in-picture in fullscreen.
 960
 961         * Modules/modern-media-controls/controls/icon-service.js:
 962         * Modules/modern-media-controls/images/macOS/pip-in-fullscreen@1x.png: Added.
 963         * Modules/modern-media-controls/images/macOS/pip-in-fullscreen@2x.png: Added.
 964
 965 2016-12-07  Dean Jackson  <dino@apple.com>
 966
 967         Remove runtime toggle for pointer-lock
 968         https://bugs.webkit.org/show_bug.cgi?id=165577
 969         <rdar://problems/29566996>
 970
 971         Reviewed by Jon Lee.
 972
 973         Remove any runtime calls to check if pointer-lock is enabled. It's
 974         either compiled in or out.
 975
 976         Covered by existing tests.
 977
 978         * bindings/generic/RuntimeEnabledFeatures.h:
 979         (WebCore::RuntimeEnabledFeatures::setPointerLockEnabled): Deleted.
 980         (WebCore::RuntimeEnabledFeatures::pointerLockEnabled): Deleted.
 981         * dom/Document.idl:
 982         * dom/Element.idl:
 983         * page/Page.h:
 984         * page/PointerLockController.cpp:
 985         (WebCore::PointerLockController::requestPointerLock):
 986         (WebCore::PointerLockController::requestPointerUnlock):
 987         (WebCore::PointerLockController::requestPointerUnlockAndForceCursorVisible):
 988
 989 2016-12-07  John Wilander  <wilander@apple.com>
 990
 991         Allow commas in Accept, Accept-Language, and Content-Language request headers for simple CORS
 992         https://bugs.webkit.org/show_bug.cgi?id=165566
 993         <rdar://problem/29560983>
 994
 995         Reviewed by Alex Christensen.
 996
 997         Updated the existing tests.
 998
 999         * platform/network/HTTPParsers.cpp:
1000         (WebCore::isValidAcceptHeaderValue):
1001             Now also accepts ','.
1002         (WebCore::isValidLanguageHeaderValue):
1003             Ditto.
1004
1005 2016-12-07  Dave Hyatt  <hyatt@apple.com>
1006
1007         [CSS Parser] Stop skipping tests in the css3 LayoutTests subdirectory
1008         https://bugs.webkit.org/show_bug.cgi?id=165574
1009
1010         Reviewed by Dean Jackson.
1011
1012         * css/CSSGroupingRule.cpp:
1013         (WebCore::CSSGroupingRule::insertRule):
1014         The old parser didn't support the creation of @namespace rules in the
1015         CSS OM, but the new parser does. We need to ensure that just like for
1016         @import rules, that the insertion of a @namespace rule inside a
1017         @supports rule is not allowed. This test passed in the old parser because
1018         the rule just didn't get created, but in the new parser, we need to
1019         add a check along with the @import rule check to ensure it isn't allowed.
1020
1021 2016-12-07  Antoine Quint  <graouts@apple.com>
1022
1023         [Modern Media Controls] Remove aspect ratio button
1024         https://bugs.webkit.org/show_bug.cgi?id=165571
1025
1026         Reviewed by Dean Jackson.
1027
1028         Since we do not have a way to toggle letterboxing on an HTMLMediaElement, we remove the
1029         corresponding button in the media controls.
1030
1031         * Modules/modern-media-controls/controls/aspect-ratio-button.js: Removed.
1032         * Modules/modern-media-controls/controls/macos-fullscreen-media-controls.js:
1033         * Modules/modern-media-controls/js-files:
1034
1035 2016-12-07  Jeremy Jones  <jeremyj@apple.com>
1036
1037         One esc to exit fullscreen and pointer lock
1038         https://bugs.webkit.org/show_bug.cgi?id=165416
1039         rdar://problem/29430711
1040
1041         Reviewed by Jer Noble.
1042
1043         Test: pointer-lock/lock-lost-on-esc-in-fullscreen.html
1044
1045         When handling escape, also exit fullscreen on escape.
1046         This also means that escape can exit fullscreen in TestRunner.
1047
1048         * page/EventHandler.cpp:
1049         (WebCore::EventHandler::keyEvent):
1050
1051 2016-12-07  Yusuke Suzuki  <utatane.tea@gmail.com>
1052
1053         [JSC] Drop translate phase in module loader
1054         https://bugs.webkit.org/show_bug.cgi?id=164861
1055
1056         Reviewed by Saam Barati.
1057
1058         * bindings/js/JSDOMWindowBase.cpp:
1059         * bindings/js/JSWorkerGlobalScopeBase.cpp:
1060
1061 2016-12-07  Keith Rollin  <krollin@apple.com>
1062
1063         Network event record/replay
1064         https://bugs.webkit.org/show_bug.cgi?id=164527
1065         <rdar://problem/29168157>
1066
1067         Reviewed by Alex Christensen.
1068
1069         Export parseURLEncodedForm.
1070
1071         No new tests -- no functionality added, changed, or removed.
1072
1073         * platform/URLParser.h:
1074
1075 2016-12-07  Dave Hyatt  <hyatt@apple.com>
1076
1077         [CSS Parser] Consolidate string/ident/url serialization functions
1078         https://bugs.webkit.org/show_bug.cgi?id=165552
1079
1080         Reviewed by Zalan Bujtas.
1081
1082         Right now CSSParser has string, ident and url serialization functions
1083         called quoteCSStringIfNeeded (which actually serializes both strings and
1084         identifiers), as well as quoteCSSURLIfNeeded.
1085
1086         CSSMarkup already has serialization functions that exist outside of the
1087         CSSParser and that handle serialization of strings, idents and URLs. This
1088         patch eliminates the CSSParser functions and consolidates all of the
1089         serialization to use CSSMarkup's functions.
1090
1091         Note that we are not spec-compliant at all here, and so I had to amend
1092         the functions to support our non-spec-compliant serialization. The goal
1093         of this patch is consolidation and not to fix our broken serialization.
1094
1095         Notable changes include parameterizing string serialization so that
1096         both single and double quotes are supported, since in the existing code
1097         we're sometimes spec-compliant (CSSSelectors) and sometimes not
1098         (CSSPrimitiveValue).
1099
1100         We also overload CSS_STRING primitive value type and have it act as both
1101         a string and a custom identifier. This is lame, since the parser should
1102         have made two different types of objects instead, but since our parser
1103         doesn't do that yet, I added a serializeAsStringOrCustomIdent that
1104         preserves our old behavior of "quote the string only if needed." In this
1105         case what that really meant was "Try to guess that we were originally a
1106         custom ident and leave off quotes if so." This function will go away
1107         once we properly create CSSStringValues and CSSCustomIdentValues instead
1108         of turning the latter into strings.
1109
1110         * css/CSSBasicShapes.cpp:
1111         (WebCore::buildPathString):
1112         * css/CSSImageValue.cpp:
1113         (WebCore::CSSImageValue::customCSSText):
1114         * css/CSSMarkup.cpp:
1115         (WebCore::isCSSTokenizerURL):
1116         (WebCore::serializeString):
1117         (WebCore::serializeURL):
1118         (WebCore::serializeAsStringOrCustomIdent):
1119         (WebCore::serializeURI): Deleted.
1120         * css/CSSMarkup.h:
1121         * css/CSSPrimitiveValue.cpp:
1122         (WebCore::CSSPrimitiveValue::formatNumberForCustomCSSText):
1123         * css/CSSSelector.cpp:
1124         (WebCore::CSSSelector::selectorText):
1125         * css/parser/CSSParser.cpp:
1126         (WebCore::isCSSTokenizerIdent): Deleted.
1127         (WebCore::isCSSTokenizerURL): Deleted.
1128         (WebCore::quoteCSSStringInternal): Deleted.
1129         (WebCore::quoteCSSString): Deleted.
1130         (WebCore::quoteCSSStringIfNeeded): Deleted.
1131         (WebCore::quoteCSSURLIfNeeded): Deleted.
1132         * css/parser/CSSParser.h:
1133         * html/HTMLElement.cpp:
1134         (WebCore::HTMLElement::mapLanguageAttributeToLocale):
1135
1136 2016-12-07  Dean Jackson  <dino@apple.com>
1137
1138         Expose internal API to detect media documents
1139         https://bugs.webkit.org/show_bug.cgi?id=165559
1140         <rdar://problems/29564054>
1141
1142         Reviewed by Antoine Quint.
1143
1144         Expose via MediaControlsHost whether or not we're in
1145         a MediaDocument.
1146
1147         Since MediaControlsHost is only exposed to our
1148         internal controls, it's not possible to test this
1149         directly. However, a subsequent test for media documents
1150         with audio-only content will cover this change.
1151
1152         * Modules/mediacontrols/MediaControlsHost.cpp: Add isInMediaDocument and put const on
1153         various methods that should have it.
1154         (WebCore::MediaControlsHost::captionDisplayMode):
1155         (WebCore::MediaControlsHost::supportsFullscreen):
1156         (WebCore::MediaControlsHost::isVideoLayerInline):
1157         (WebCore::MediaControlsHost::isInMediaDocument):
1158         * Modules/mediacontrols/MediaControlsHost.h:
1159         * Modules/mediacontrols/MediaControlsHost.idl: Add isInMediaDocument.
1160
1161 2016-12-07  Zalan Bujtas  <zalan@apple.com>
1162
1163         ASSERTION FAILED: count >= 1 in WebCore::RenderMultiColumnSet::columnCount
1164         https://bugs.webkit.org/show_bug.cgi?id=136365
1165         <rdar://problem/27686727>
1166
1167         Reviewed by Simon Fraser.
1168
1169         Margin value overflow ends up confusing the column count computing logic.
1170
1171         Test: fast/multicol/assert-on-column-count-when-zoomed-in.html
1172
1173         * css/StyleResolver.cpp:
1174         (WebCore::addIntrinsicMargins):
1175
1176 2016-12-07  Ryosuke Niwa  <rniwa@webkit.org>
1177
1178         document.caretRangeFromPoint doesn't retarget the resultant Range correctly.
1179         https://bugs.webkit.org/show_bug.cgi?id=165146
1180
1181         Reviewed by Sam Weinig.
1182
1183         The bug was caused by caretRangeFromPoint not retargeting the resultant Range correctly.
1184         Namely, it's possible for RenderObject::positionForPoint to move across shadow boundary
1185         even if node was identically equal to ancestorInThisScope(node).
1186
1187         Fixed the bug by directly retargeting the range's container node and its offset as done
1188         for elementFromPoint in r206795.
1189
1190         Test: fast/shadow-dom/caret-range-from-point-in-shadow-tree.html
1191
1192         * dom/Document.cpp:
1193         (WebCore::Document::caretRangeFromPoint):
1194
1195 2016-12-07  Antoine Quint  <graouts@apple.com>
1196
1197         [Modern Media Controls] Entering fullscreen and returning to inline shows fullscreen controls
1198         https://bugs.webkit.org/show_bug.cgi?id=165536
1199
1200         Reviewed by Tim Horton.
1201
1202         We regressed when we implemented the fix for webkit.org/b/165494 and we started to rely solely
1203         on the "webkitpresentationmodechanged" event to identify presention mode changes. As it turns out,
1204         when the "webkitpresentationmodechanged" event is dispatched when exiting fullscreen and returning
1205         to the inline presentation mode, querying the "webkitPresentationMode" property says "inline" while
1206         "webkitDisplayingFullscreen" still returns true (raised as webkit.org/b/165538).
1207
1208         We now use the "webkitPresentationMode" property when that property is supported and we're using the
1209         "webkitpresentationmodechanged" event, and we use the "webkitDisplayingFullscreen" property otherwise.
1210
1211         Test: media/modern-media-controls/media-controller/media-controller-inline-to-fullscreen-to-inline.html
1212
1213         * Modules/modern-media-controls/media/media-controller.js:
1214         (MediaController.prototype.get layoutTraits):
1215
1216 2016-12-07  Wenson Hsieh  <wenson_hsieh@apple.com>
1217
1218         Scroll position jumps to the origin when scrolling without momentum at the end of a scroll snapping container
1219         https://bugs.webkit.org/show_bug.cgi?id=165474
1220         <rdar://problem/29534305>
1221
1222         Reviewed by Simon Fraser.
1223
1224         When initializing an AppKit _NSScrollingMomentumCalculator, if the initial and target positions are the same and
1225         the initial velocity is (0, 0), the momentum calculator will output (0, 0) as the animated scroll position when
1226         animating. This causes the scroll position to jump to the top left in some cases when scrolling in scroll snap
1227         containers. To fix this, we teach the ScrollingMomentumCalculatorMac to return an animation duration of 0 and
1228         an animated scroll position equal to the final scroll position when this is the case.
1229
1230         Test: tiled-drawing/scrolling/scroll-snap/scrolling-jumps-to-top.html
1231
1232         * page/scrolling/mac/ScrollingMomentumCalculatorMac.h:
1233         * page/scrolling/mac/ScrollingMomentumCalculatorMac.mm:
1234         (WebCore::ScrollingMomentumCalculatorMac::ScrollingMomentumCalculatorMac):
1235         (WebCore::ScrollingMomentumCalculatorMac::scrollOffsetAfterElapsedTime):
1236         (WebCore::ScrollingMomentumCalculatorMac::animationDuration):
1237
1238 2016-12-07  Nan Wang  <n_wang@apple.com>
1239
1240         AX: menu type toolbar should be mapped correctly on Mac
1241         https://bugs.webkit.org/show_bug.cgi?id=165537
1242
1243         Reviewed by Chris Fleizach.
1244
1245         The menu tag with a toolbar type should have the corresponding role.
1246
1247         Test: accessibility/mac/menu-type-toolbar.html
1248
1249         * accessibility/AccessibilityRenderObject.cpp:
1250         (WebCore::AccessibilityRenderObject::determineAccessibilityRole):
1251
1252 2016-12-07  Wenson Hsieh  <wenson_hsieh@apple.com>
1253
1254         Add a new project for recording and playing back editing commands in editable web content
1255         https://bugs.webkit.org/show_bug.cgi?id=165114
1256         <rdar://problem/29408135>
1257
1258         Reviewed by Beth Dakin.
1259
1260         Adds new scripts used to record and play back editing, as well as a new Xcode Copy files phase that pushes these
1261         scripts to the internal system directory when installing. See the Tools ChangeLog and individual comments below
1262         for more details. Covered by 3 new unit tests in the EditingHistory project.
1263
1264         * InternalScripts/DumpEditingHistory.js: Added.
1265         (beginProcessingTopLevelUpdate):
1266         (endProcessingTopLevelUpdate):
1267         (appendDOMUpdatesFromRecords):
1268         (appendSelectionUpdateIfNecessary):
1269
1270         Adds new entries into the top-level list of DOM updates captured when editing. Respectively, these are input
1271         events and selection changes.
1272
1273         (EditingHistory.getEditingHistoryAsJSONString):
1274         * InternalScripts/EditingHistoryUtil.js: Added.
1275         (prototype._scramble):
1276         (prototype.applyToText):
1277         (prototype.applyToFilename):
1278         (prototype._scrambedNumberIndexForCode):
1279         (prototype._scrambedLowercaseIndexForCode):
1280         (prototype._scrambedUppercaseIndexForCode):
1281
1282         Naive implementation of an obfuscator. Currently, this only affects alphanumeric characters. Obfuscation is off
1283         by default, but can be toggled on in JavaScript.
1284
1285         (elementFromMarkdown):
1286         (GlobalNodeMap):
1287         (GlobalNodeMap.prototype.nodesForGUIDs):
1288         (GlobalNodeMap.prototype.guidsForTNodes):
1289         (GlobalNodeMap.prototype.nodeForGUID):
1290         (GlobalNodeMap.prototype.guidForNode):
1291         (GlobalNodeMap.prototype.hasGUIDForNode):
1292         (GlobalNodeMap.prototype.nodes):
1293         (GlobalNodeMap.prototype.toObject):
1294         (GlobalNodeMap.fromObject):
1295         (GlobalNodeMap.dataForNode):
1296         (GlobalNodeMap.elementFromTagName):
1297         (GlobalNodeMap.nodeAttributesToObject):
1298         (GlobalNodeMap.prototype.descriptionHTMLForGUID):
1299         (GlobalNodeMap.prototype.descriptionHTMLForNode):
1300
1301         The GlobalNodeMap keeps track of every node that has appeared in the DOM, assigning each node a globally unique
1302         identifier (GUID). This GUID is used when reconstructing the DOM, as well as unapplying or applying editing.
1303
1304         (SelectionState):
1305         (SelectionState.prototype.isEqual):
1306         (SelectionState.prototype.applyToSelection):
1307         (SelectionState.fromSelection):
1308         (SelectionState.prototype.toObject):
1309         (SelectionState.fromObject):
1310
1311         Represents a snapshot of the Selection state (determined by getSelection()).
1312
1313         (DOMUpdate):
1314         (DOMUpdate.prototype.apply):
1315         (DOMUpdate.prototype.unapply):
1316         (DOMUpdate.prototype.targetNode):
1317         (DOMUpdate.prototype.detailsElement):
1318         (DOMUpdate.ofType):
1319         (DOMUpdate.fromRecords):
1320
1321         A DOMUpdate is an abstract object representing a change in the DOM that may be applied and unapplied. These are
1322         also serializable as hashes, which may then be converted to JSON when generating editing history data.
1323
1324         (ChildListUpdate):
1325         (ChildListUpdate.prototype.apply):
1326         (ChildListUpdate.prototype.unapply):
1327         (ChildListUpdate.prototype._nextSibling):
1328         (ChildListUpdate.prototype._removedNodes):
1329         (ChildListUpdate.prototype._addedNodes):
1330         (ChildListUpdate.prototype.toObject):
1331         (ChildListUpdate.prototype.detailsElement):
1332         (ChildListUpdate.fromObject):
1333
1334         These three update types correspond to the three types of DOM mutations. These may appear as top-level updates
1335         if they are not captured during an input event, but for the majority of user-input-driven changes, they will be
1336         children of an input event.
1337
1338         (CharacterDataUpdate):
1339         (CharacterDataUpdate.prototype.apply):
1340         (CharacterDataUpdate.prototype.unapply):
1341         (CharacterDataUpdate.prototype.detailsElement):
1342         (CharacterDataUpdate.prototype.toObject):
1343         (CharacterDataUpdate.fromObject):
1344         (AttributeUpdate):
1345         (AttributeUpdate.prototype.apply):
1346         (AttributeUpdate.prototype.unapply):
1347         (AttributeUpdate.prototype.detailsElement):
1348         (AttributeUpdate.prototype.toObject):
1349         (AttributeUpdate.fromObject):
1350         (SelectionUpdate):
1351         (SelectionUpdate.prototype.apply):
1352         (SelectionUpdate.prototype.unapply):
1353         (SelectionUpdate.prototype.toObject):
1354         (SelectionUpdate.fromObject):
1355         (SelectionUpdate.prototype._rangeDescriptionHTML):
1356         (SelectionUpdate.prototype._anchorDescriptionHTML):
1357         (SelectionUpdate.prototype._focusDescriptionHTML):
1358         (SelectionUpdate.prototype.detailsElement):
1359
1360         Represents a change in the Selection. While no changes to the DOM structure occur as a result of a
1361         SelectionUpdate, the information contained in these updates is used to determine where the selection should be
1362         when rewinding or playing back the editing history.
1363
1364         (InputEventUpdate):
1365         (InputEventUpdate.prototype._obfuscatedData):
1366         (InputEventUpdate.prototype.apply):
1367         (InputEventUpdate.prototype.unapply):
1368         (InputEventUpdate.prototype.toObject):
1369         (InputEventUpdate.fromObject):
1370         (InputEventUpdate.prototype.detailsElement):
1371
1372         Represents an update due to user input, which consists of some number of child DOM mutation updates.
1373
1374         * WebCore.xcodeproj/project.pbxproj:
1375
1376 2016-12-07  Jer Noble  <jer.noble@apple.com>
1377
1378         ASSERT crash while running media-source/mediasource-activesourcebuffers.html under Stress GC bot.
1379         https://bugs.webkit.org/show_bug.cgi?id=165514
1380
1381         Reviewed by Eric Carlson.
1382
1383         If a track associated with MSE is disabled after a SourceBuffer begins parsing a queued
1384         append operation, SourceBuffer can get into a state where it asks SourceBufferPrivateAVFObjC
1385         if it is ready to accept data for that disabled track. This causes an ASSERT_NOT_REACHED in
1386         isReadyForMoreData().
1387
1388         However, this seems to be a valid condition; we can safely just return "false" from
1389         isReadyForMoreSamples() (and also bail from notifyClientWhenReadyForMoreSamples()) when asked
1390         about a disabled track.
1391
1392         * platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:
1393         (WebCore::SourceBufferPrivateAVFObjC::isReadyForMoreSamples):
1394         (WebCore::SourceBufferPrivateAVFObjC::notifyClientWhenReadyForMoreSamples):
1395
1396 2016-12-07  Dave Hyatt  <hyatt@apple.com>
1397
1398         [CSS Parser] Eliminate the Scope class and fold it into CSSTokenizer
1399         https://bugs.webkit.org/show_bug.cgi?id=165532
1400
1401         Reviewed by Dean Jackson.
1402
1403         The Scope class, nested inside CSSTokenizer, is both poorly named and serves
1404         no purpose. It's especially weird that it was the entry point into tokenization
1405         instead of CSSTokenizer.
1406
1407         Eliminate the class, fold its members and functions into CSSTokenizer, and have
1408         callers just make a CSSTokenizer instead.
1409
1410         * css/parser/CSSParser.cpp:
1411         (WebCore::CSSParser::parseSupportsCondition):
1412         (WebCore::CSSParser::parseSelector):
1413         * css/parser/CSSParserImpl.cpp:
1414         (WebCore::CSSParserImpl::parseValue):
1415         (WebCore::CSSParserImpl::parseCustomPropertyValue):
1416         (WebCore::CSSParserImpl::parseInlineStyleDeclaration):
1417         (WebCore::CSSParserImpl::parseDeclarationList):
1418         (WebCore::CSSParserImpl::parseRule):
1419         (WebCore::CSSParserImpl::parseStyleSheet):
1420         (WebCore::CSSParserImpl::parseKeyframeKeyList):
1421         (WebCore::CSSParserImpl::parseDeclarationListForInspector):
1422         (WebCore::CSSParserImpl::parseStyleSheetForInspector):
1423         * css/parser/CSSTokenizer.cpp:
1424         (WebCore::CSSTokenizer::CSSTokenizer):
1425         (WebCore::CSSTokenizer::tokenRange):
1426         (WebCore::CSSTokenizer::tokenCount):
1427         (WebCore::CSSTokenizer::registerString):
1428         (WebCore::CSSTokenizer::Scope::Scope): Deleted.
1429         (WebCore::CSSTokenizer::Scope::tokenRange): Deleted.
1430         (WebCore::CSSTokenizer::Scope::tokenCount): Deleted.
1431         * css/parser/CSSTokenizer.h:
1432         (WebCore::CSSTokenizer::Scope::storeString): Deleted.
1433         * css/parser/CSSTokenizerInputStream.cpp:
1434         (WebCore::CSSTokenizerInputStream::CSSTokenizerInputStream):
1435         * css/parser/CSSTokenizerInputStream.h:
1436         * css/parser/MediaQueryParser.cpp:
1437         (WebCore::MediaQueryParser::parseMediaQuerySet):
1438         * css/parser/SizesAttributeParser.cpp:
1439         (WebCore::SizesAttributeParser::SizesAttributeParser):
1440
1441 2016-12-07  Joseph Pecoraro  <pecoraro@apple.com>
1442
1443         Web Inspector: Remove unused and mostly untested Page domain commands and events
1444         https://bugs.webkit.org/show_bug.cgi?id=165507
1445
1446         Reviewed by Brian Burg.
1447
1448         Remove any code associated with the commands/events being removed.
1449
1450         * inspector/DOMPatchSupport.cpp:
1451         * inspector/DOMPatchSupport.h:
1452         * inspector/InspectorClient.h:
1453         (WebCore::InspectorClient::handleJavaScriptDialog): Deleted.
1454         * inspector/InspectorInstrumentation.cpp:
1455         (WebCore::InspectorInstrumentation::scriptsEnabledImpl): Deleted.
1456         (WebCore::InspectorInstrumentation::willRunJavaScriptDialogImpl): Deleted.
1457         (WebCore::InspectorInstrumentation::didRunJavaScriptDialogImpl): Deleted.
1458         * inspector/InspectorInstrumentation.h:
1459         (WebCore::InspectorInstrumentation::scriptsEnabled): Deleted.
1460         (WebCore::InspectorInstrumentation::willRunJavaScriptDialog): Deleted.
1461         (WebCore::InspectorInstrumentation::didRunJavaScriptDialog): Deleted.
1462         * inspector/InspectorPageAgent.cpp:
1463         (WebCore::InspectorPageAgent::disable):
1464         (WebCore::InspectorPageAgent::didClearWindowObjectInWorld):
1465         (WebCore::InspectorPageAgent::didPaint):
1466         (WebCore::InspectorPageAgent::buildObjectForFrame):
1467         (WebCore::InspectorPageAgent::archive):
1468         (WebCore::createXHRTextDecoder): Deleted.
1469         (WebCore::InspectorPageAgent::getScriptExecutionStatus): Deleted.
1470         (WebCore::InspectorPageAgent::setScriptExecutionDisabled): Deleted.
1471         (WebCore::InspectorPageAgent::willRunJavaScriptDialog): Deleted.
1472         (WebCore::InspectorPageAgent::didRunJavaScriptDialog): Deleted.
1473         (WebCore::InspectorPageAgent::scriptsEnabled): Deleted.
1474         (WebCore::InspectorPageAgent::handleJavaScriptDialog): Deleted.
1475         * inspector/InspectorPageAgent.h:
1476         * page/Chrome.cpp:
1477         (WebCore::Chrome::runBeforeUnloadConfirmPanel):
1478         (WebCore::Chrome::runJavaScriptAlert):
1479         (WebCore::Chrome::runJavaScriptConfirm):
1480         (WebCore::Chrome::runJavaScriptPrompt):
1481         * page/Settings.cpp:
1482         (WebCore::Settings::setScriptEnabled):
1483
1484 2016-12-07  Jer Noble  <jer.noble@apple.com>
1485
1486         [pointer-lock] Cursor should become visible when exiting pointer-lock via ESC key.
1487         https://bugs.webkit.org/show_bug.cgi?id=165377
1488
1489         Reviewed by Eric Carlson.
1490
1491         Follow up to previous patch; clear the m_forceCursorVisibleUponUnlock flag upon losing
1492         pointer lock, so that subsequent unlocks don't erroneously cause the cursor to become
1493         visible.
1494
1495         * page/PointerLockController.cpp:
1496         (WebCore::PointerLockController::didLosePointerLock):
1497
1498 2016-12-06  Geoffrey Garen  <ggaren@apple.com>
1499
1500         performance.now() should truncate to 100us
1501         https://bugs.webkit.org/show_bug.cgi?id=165503
1502         <rdar://problem/29544531>
1503
1504         Reviewed by Mark Lam.
1505
1506         * page/Performance.cpp:
1507         (WebCore::Performance::reduceTimeResolution):
1508
1509 2016-12-07  Dave Hyatt  <hyatt@apple.com>
1510
1511         [CSS Parser] Turn back on a bunch of layout tests
1512         https://bugs.webkit.org/show_bug.cgi?id=165529
1513
1514         Reviewed by Zalan Bujtas.
1515
1516         Fixes fast/css/transform-origin-parsing.html.
1517
1518         * css/parser/CSSPropertyParser.cpp:
1519         (WebCore::CSSPropertyParser::consumeTransformOrigin):
1520         The implicit check was relying on a null check of the z value, but
1521         we called releaseNonNull on it already. This meant all z properties were
1522         marked as implicit incorrectly. Fix by storing whether or not we had a z
1523         value in a local variable.
1524
1525 2016-12-07  Andreas Kling  <akling@apple.com>
1526
1527         [iOS] Remove bitrotted FrameLoader attempt to limit page caching under memory pressure.
1528         <https://webkit.org/b/165527>
1529
1530         Reviewed by Antti Koivisto.
1531
1532         This code no longer does anything, as the page cache is hard capped
1533         at 0 entries while under memory pressure.
1534
1535         * loader/FrameLoader.cpp:
1536         (WebCore::FrameLoader::commitProvisionalLoad):
1537
1538 2016-12-06  Antoine Quint  <graouts@apple.com>
1539
1540         [Modern Media Controls] Use tap and pinch gestures on iOS
1541         https://bugs.webkit.org/show_bug.cgi?id=165518
1542
1543         Reviewed by Dean Jackson.
1544
1545         We now use touch events to identify taps and pinches on media controls to respond immediately
1546         to user interaction on buttons and enter fullscreen when pinching out.
1547
1548         * Modules/modern-media-controls/controls/button.css:
1549         (button):
1550         * Modules/modern-media-controls/controls/button.js:
1551         (Button):
1552         (Button.prototype.handleEvent):
1553         (Button.prototype.gestureRecognizerStateDidChange):
1554         (Button.prototype._notifyDelegateOfActivation):
1555         (Button.prototype._handleUIEvent): Deleted.
1556         * Modules/modern-media-controls/controls/controls-bar.js:
1557         (ControlsBar.prototype.set fadesWhileIdle):
1558         (ControlsBar.prototype.gestureRecognizerStateDidChange):
1559         * Modules/modern-media-controls/controls/icon-button.js:
1560         (IconButton.prototype._updateImage):
1561         (IconButton):
1562         * Modules/modern-media-controls/controls/ios-inline-media-controls.js:
1563         (IOSInlineMediaControls.prototype.gestureRecognizerStateDidChange):
1564         * Modules/modern-media-controls/gesture-recognizers/gesture-recognizer.js: Added.
1565         (GestureRecognizer):
1566         (GestureRecognizer.prototype.get state):
1567         (GestureRecognizer.prototype.set state):
1568         (GestureRecognizer.prototype.get target):
1569         (GestureRecognizer.prototype.set target):
1570         (GestureRecognizer.prototype.get numberOfTouches):
1571         (GestureRecognizer.prototype.get enabled):
1572         (GestureRecognizer.prototype.set enabled):
1573         (GestureRecognizer.prototype.reset):
1574         (GestureRecognizer.prototype.locationInElement):
1575         (GestureRecognizer.prototype.locationInClient):
1576         (GestureRecognizer.prototype.locationOfTouchInElement):
1577         (GestureRecognizer.prototype.touchesBegan):
1578         (GestureRecognizer.prototype.touchesMoved):
1579         (GestureRecognizer.prototype.touchesEnded):
1580         (GestureRecognizer.prototype.touchesCancelled):
1581         (GestureRecognizer.prototype.gestureBegan):
1582         (GestureRecognizer.prototype.gestureChanged):
1583         (GestureRecognizer.prototype.gestureEnded):
1584         (GestureRecognizer.prototype.enterPossibleState):
1585         (GestureRecognizer.prototype.enterBeganState):
1586         (GestureRecognizer.prototype.enterEndedState):
1587         (GestureRecognizer.prototype.enterCancelledState):
1588         (GestureRecognizer.prototype.enterFailedState):
1589         (GestureRecognizer.prototype.enterChangedState):
1590         (GestureRecognizer.prototype.enterRecognizedState):
1591         (GestureRecognizer.prototype.handleEvent):
1592         (GestureRecognizer.prototype._initRecognizer):
1593         (GestureRecognizer.prototype._updateBaseListeners):
1594         (GestureRecognizer.prototype._removeTrackingListeners):
1595         (GestureRecognizer.prototype._updateTargetTouches):
1596         (GestureRecognizer.prototype._updateKeyboardModifiers):
1597         * Modules/modern-media-controls/gesture-recognizers/pinch.js: Added.
1598         (PinchGestureRecognizer):
1599         (PinchGestureRecognizer.prototype.get velocity):
1600         (PinchGestureRecognizer.prototype.touchesBegan):
1601         (PinchGestureRecognizer.prototype.touchesMoved):
1602         (PinchGestureRecognizer.prototype.touchesEnded):
1603         (PinchGestureRecognizer.prototype.gestureBegan):
1604         (PinchGestureRecognizer.prototype.gestureChanged):
1605         (PinchGestureRecognizer.prototype.gestureEnded):
1606         (PinchGestureRecognizer.prototype.reset):
1607         (PinchGestureRecognizer.prototype._recordGesture):
1608         (PinchGestureRecognizer.prototype._updateStateWithEvent):
1609         (PinchGestureRecognizer.prototype._distance):
1610         * Modules/modern-media-controls/gesture-recognizers/tap.js: Added.
1611         (TapGestureRecognizer):
1612         (TapGestureRecognizer.prototype.touchesBegan):
1613         (TapGestureRecognizer.prototype.touchesMoved):
1614         (TapGestureRecognizer.prototype.touchesEnded):
1615         (TapGestureRecognizer.prototype.reset):
1616         (TapGestureRecognizer.prototype.locationInElement):
1617         (TapGestureRecognizer.prototype.locationInClient):
1618         (TapGestureRecognizer.prototype._clearTimer):
1619         (TapGestureRecognizer.prototype._rewindTimer):
1620         (TapGestureRecognizer.prototype._timerFired):
1621         * Modules/modern-media-controls/js-files:
1622         * Modules/modern-media-controls/media/airplay-support.js:
1623         (AirplaySupport.prototype.buttonWasClicked): Deleted.
1624         * Modules/modern-media-controls/media/controls-visibility-support.js:
1625         (ControlsVisibilitySupport):
1626         (ControlsVisibilitySupport.prototype.handleEvent):
1627         (ControlsVisibilitySupport.prototype.syncControl): Deleted.
1628         * Modules/modern-media-controls/media/fullscreen-support.js:
1629         (FullscreenSupport):
1630         (FullscreenSupport.prototype.iOSInlineMediaControlsRecognizedPinchInGesture):
1631         (FullscreenSupport.prototype.buttonWasClicked): Deleted.
1632         * Modules/modern-media-controls/media/media-controller-support.js:
1633         (MediaControllerSupport):
1634         (MediaControllerSupport.prototype.handleEvent):
1635         (MediaControllerSupport.prototype.buttonWasClicked): Deleted.
1636         * Modules/modern-media-controls/media/mute-support.js:
1637         (MuteSupport.prototype.buttonWasClicked): Deleted.
1638         * Modules/modern-media-controls/media/pip-support.js:
1639         (PiPSupport.prototype.buttonWasClicked): Deleted.
1640         * Modules/modern-media-controls/media/playback-support.js:
1641         (PlaybackSupport.prototype.buttonWasClicked): Deleted.
1642         * Modules/modern-media-controls/media/skip-back-support.js:
1643         (SkipBackSupport.prototype.buttonWasClicked): Deleted.
1644         * Modules/modern-media-controls/media/start-support.js:
1645         (StartSupport.prototype.buttonWasClicked): Deleted.
1646         * Modules/modern-media-controls/media/tracks-support.js:
1647         (TracksSupport.prototype.buttonWasClicked): Deleted.
1648
1649 2016-12-06  Dean Jackson  <dino@apple.com>
1650
1651         Apply styling to media documents with modern controls
1652         https://bugs.webkit.org/show_bug.cgi?id=165499
1653         <rdar://problems/29543847>
1654
1655         Reviewed by Antoine Quint.
1656
1657         Forgot to commit most of the changes in MediaDocument.cpp :(
1658
1659         * html/MediaDocument.cpp:
1660         (WebCore::MediaDocumentParser::createDocumentStructure):
1661
1662 2016-12-06  Simon Fraser  <simon.fraser@apple.com>
1663
1664         Two tiled drawing tests failing with visual viewports enabled.
1665         https://bugs.webkit.org/show_bug.cgi?id=165489
1666
1667         Reviewed by Dean Jackson.
1668
1669         computeLayoutViewportOrigin() gets called for iframes when the layout viewport
1670         is zero-sized, but the visual viewport is non-zero. It doesn't really make sense
1671         to compute a layout viewport when the visual viewport is larger than the layout
1672         viewport, but if this happens just anchor the layout viewport at the origin
1673         of the visual viewport.
1674
1675         * page/FrameView.cpp:
1676         (WebCore::FrameView::computeLayoutViewportOrigin):
1677
1678 2016-12-04  Darin Adler  <darin@apple.com>
1679
1680         Remove various stray uses of WebCore::Dictionary
1681         https://bugs.webkit.org/show_bug.cgi?id=165358
1682
1683         Reviewed by Sam Weinig.
1684
1685         I discovered three completely unused classes in the MediaStream code:
1686         MediaTrackConstraint, MediaTrackConstraintSet, and MediaTrackConstraints.
1687         While there is some chance we may implement these at some point, they are
1688         likely to be dictionaries or some other kind of data structure rather than
1689         clases, and there is no reason to keep the current classes.
1690
1691         * CMakeLists.txt: Removed MediaTrackConstraint, MediaTrackConstraintSet,
1692         and MediaTrackConstraints.
1693         * DerivedSources.cpp: Ditto.
1694         * DerivedSources.make: Ditto.
1695
1696         * Modules/indexeddb/IDBDatabase.h: Removed include of Dictionary.h.
1697         * Modules/indexeddb/IDBObjectStore.cpp: Added include of HeapInlines.h.
1698         * Modules/indexeddb/IDBRequest.cpp: Added include of StrongInlines.h.
1699         * Modules/indexeddb/IDBTransaction.cpp: Added include of SerializedScriptValue.h.
1700
1701         * Modules/mediastream/MediaStreamTrack.cpp:
1702         (WebCore::MediaStreamTrack::getConstraints): Deleted.
1703         This function is implemented entirely in the bindings layer.
1704         * Modules/mediastream/MediaStreamTrack.h: Removed getConstraints.
1705
1706         * Modules/mediastream/MediaTrackConstraint.cpp: Removed.
1707         * Modules/mediastream/MediaTrackConstraint.h: Removed.
1708         * Modules/mediastream/MediaTrackConstraint.idl: Removed.
1709         * Modules/mediastream/MediaTrackConstraintSet.cpp: Removed.
1710         * Modules/mediastream/MediaTrackConstraintSet.h: Removed.
1711         * Modules/mediastream/MediaTrackConstraintSet.idl: Removed.
1712         * Modules/mediastream/MediaTrackConstraints.cpp: Removed.
1713         * Modules/mediastream/MediaTrackConstraints.h: Removed.
1714         * Modules/mediastream/MediaTrackConstraints.idl: Removed.
1715
1716         * Modules/mediastream/NavigatorUserMedia.idl: Use the type names from the latest
1717         getUserMedia specification for the arguments to webkitGetUserMedia; not sure they
1718         are precisely correct for this legacy function, but since this is a JSBuiltin,
1719         it's all ignored anyway. Just important not to call anything "Dictionary".
1720
1721         * WebCore.xcodeproj/project.pbxproj: Removed MediaTrackConstraint, MediaTrackConstraintSet,
1722         and MediaTrackConstraints.
1723
1724         * bindings/js/JSApplePaySessionCustom.cpp:
1725         (WebCore::JSApplePaySession::completeShippingMethodSelection): Use uncheckedArgument, since
1726         the code explicitly checks the number of arguments.
1727         (WebCore::JSApplePaySession::completeShippingContactSelection): Ditto.
1728         (WebCore::JSApplePaySession::completePaymentMethodSelection): Ditto.
1729
1730         * bindings/js/JSFontFaceCustom.cpp: Removed include of Dictionary.h.
1731         * dom/Document.cpp: Ditto.
1732         * dom/Element.cpp: Ditto.
1733
1734         * dom/Element.h: Removed forward declaration of Dictionary.
1735
1736 2016-12-06  Wenson Hsieh  <wenson_hsieh@apple.com>
1737
1738         After preventing a beforeinput event, an input event is fired when formatting rich text
1739         https://bugs.webkit.org/show_bug.cgi?id=165435
1740         <rdar://problem/29522314>
1741
1742         Reviewed by Ryosuke Niwa.
1743
1744         This regressed after I refactored some input event event dispatch logic when formatting text in r208461.
1745         I moved the logic for dispatching input events when applying styles into Editor::applyStyle, but left out an
1746         early return after firing the beforeinput event if the default behavior was prevented, which caused us to
1747         continue on and dispatch an input event. The fix is to bail from applyStyle if default was prevented.
1748
1749         Augmented an existing layout test to cover this case.
1750
1751         * editing/Editor.cpp:
1752         (WebCore::Editor::applyStyle):
1753         (WebCore::Editor::applyParagraphStyle):
1754
1755 2016-12-06  Antoine Quint  <graouts@apple.com>
1756
1757         [Modern Media Controls] Instantiate iOS media controls
1758         https://bugs.webkit.org/show_bug.cgi?id=165498
1759
1760         Reviewed by Dean Jackson.
1761
1762         Add support for the modern media controls runtime flags in RenderThemeIOS and, when instantiating iOS controls,
1763         return an IOSInlineMediaControls class.
1764
1765         * Modules/modern-media-controls/media/media-controller.js:
1766         (MediaController.prototype._controlsClass):
1767         (MediaController):
1768         * rendering/RenderThemeIOS.h:
1769         * rendering/RenderThemeIOS.mm:
1770         (WebCore::RenderThemeIOS::mediaControlsStyleSheet):
1771         (WebCore::RenderThemeIOS::mediaControlsScript):
1772         (WebCore::RenderThemeIOS::mediaControlsBase64StringForIconAndPlatform):
1773
1774 2016-12-06  Dean Jackson  <dino@apple.com>
1775
1776         Apply styling to media documents with modern controls
1777         https://bugs.webkit.org/show_bug.cgi?id=165499
1778         <rdar://problems/29543847>
1779
1780         Reviewed by Antoine Quint.
1781
1782         Existing MediaDocuments got styling from the injected style
1783         sheet, which had rules for such documents. The modern media controls
1784         use a scoped style rule, so we can't touch the document style. Instead,
1785         inline the style directly from MediaDocument.
1786
1787         Test: media/modern-media-controls/media-documents/background-color-and-centering.html
1788
1789         * html/MediaDocument.cpp:
1790         (WebCore::MediaDocumentParser::createDocumentStructure):
1791
1792 2016-12-06  Zalan Bujtas  <zalan@apple.com>
1793
1794         Move RenderElement::rendererForRootBackground to RenderView.
1795         https://bugs.webkit.org/show_bug.cgi?id=165454
1796
1797         Reviewed by Simon Fraser.
1798
1799         Currently RenderElement::rendererForRootBackground can only be called on document element's renderer.
1800         This patch removes this limitation and moves the functionality from RenderElement to RenderView.
1801
1802         No change in functionality.
1803
1804         * page/FrameView.cpp:
1805         (WebCore::FrameView::calculateExtendedBackgroundMode):
1806         * rendering/RenderBox.cpp:
1807         (WebCore::RenderBox::paintRootBoxFillLayers):
1808         * rendering/RenderElement.cpp:
1809         (WebCore::RenderElement::rendererForRootBackground): Deleted.
1810         * rendering/RenderElement.h:
1811         * rendering/RenderLayerBacking.cpp:
1812         (WebCore::RenderLayerBacking::rendererBackgroundColor):
1813         * rendering/RenderView.cpp:
1814         (WebCore::RenderView::rendererForRootBackground):
1815         (WebCore::rendererObscuresBackground):
1816         (WebCore::RenderView::paintBoxDecorations):
1817         (WebCore::RenderView::rootBackgroundIsEntirelyFixed):
1818         * rendering/RenderView.h:
1819
1820 2016-12-05  Jiewen Tan  <jiewen_tan@apple.com>
1821
1822         [Part 1 of 2] Add support for SPKI format while doing SubtleCrypto.importKey/exportKey
1823         https://bugs.webkit.org/show_bug.cgi?id=129978
1824         <rdar://problem/21799829>
1825
1826         This is part 1 of Bug 129978. In this patch, it adds the SPKI format support for
1827         SubtleCrypto.importKey/exportKey. Currently support algorithms are RSAES-PKCS1-v1_5,
1828         RSASSA-PKCS1-v1_5 and RSA-OAEP.
1829
1830         Reviewed by Brent Fulgham.
1831
1832         Tests: crypto/subtle/rsa-import-jwk-key-export-spki-key.html
1833                crypto/subtle/rsa-import-spki-key-export-jwk-key.html
1834                crypto/subtle/rsa-import-spki-key-export-spki-key.html
1835                crypto/subtle/rsa-import-spki-small-key.html
1836                crypto/subtle/rsa-oaep-generate-export-key-spki.html
1837                crypto/subtle/rsa-oaep-import-spki-key.html
1838                crypto/subtle/rsaes-pkcs1-v1_5-generate-export-key-spki.html
1839                crypto/subtle/rsaes-pkcs1-v1_5-import-spki-key.html
1840                crypto/subtle/rsassa-pkcs1-v1_5-generate-export-key-spki.html
1841                crypto/subtle/rsassa-pkcs1-v1_5-import-spki-key.html
1842                crypto/workers/subtle/rsa-export-spki-key.html
1843                crypto/workers/subtle/rsa-import-spki-key.html
1844
1845         * bindings/js/JSSubtleCryptoCustom.cpp:
1846         (WebCore::toKeyData):
1847         * crypto/CommonCryptoUtilities.h:
1848         * crypto/algorithms/CryptoAlgorithmRSAES_PKCS1_v1_5.cpp:
1849         (WebCore::CryptoAlgorithmRSAES_PKCS1_v1_5::importKey):
1850         (WebCore::CryptoAlgorithmRSAES_PKCS1_v1_5::exportKey):
1851         * crypto/algorithms/CryptoAlgorithmRSASSA_PKCS1_v1_5.cpp:
1852         (WebCore::CryptoAlgorithmRSASSA_PKCS1_v1_5::importKey):
1853         (WebCore::CryptoAlgorithmRSASSA_PKCS1_v1_5::exportKey):
1854         * crypto/algorithms/CryptoAlgorithmRSA_OAEP.cpp:
1855         (WebCore::CryptoAlgorithmRSA_OAEP::importKey):
1856         (WebCore::CryptoAlgorithmRSA_OAEP::exportKey):
1857         * crypto/gnutls/CryptoKeyRSAGnuTLS.cpp:
1858         (WebCore::CryptoKeyRSA::importSpki):
1859         (WebCore::CryptoKeyRSA::exportSpki):
1860         * crypto/keys/CryptoKeyRSA.h:
1861         * crypto/mac/CryptoKeyRSAMac.cpp:
1862         (WebCore::bytesUsedToEncodedLength):
1863         (WebCore::bytesNeededForEncodedLength):
1864         (WebCore::addEncodedASN1Length):
1865         (WebCore::CryptoKeyRSA::importSpki):
1866         (WebCore::CryptoKeyRSA::exportSpki):
1867
1868 2016-12-06  Simon Fraser  <simon.fraser@apple.com>
1869
1870         Fix editing text flakiness introduced in r209411
1871         https://bugs.webkit.org/show_bug.cgi?id=165500
1872
1873         Reviewed by Dean Jackson.
1874
1875         The "insideFixed" out param was left uninitialized sometimes. This affected FrameSelection's
1876         m_caretInsidePositionFixed, which caused problems in these tests.
1877
1878         Tested by editing tests in WK1.
1879
1880         * editing/htmlediting.cpp:
1881         (WebCore::absoluteBoundsForLocalCaretRect):
1882
1883 2016-12-06  Alexey Proskuryakov  <ap@apple.com>
1884
1885         Correct SDKROOT values in xcconfig files
1886         https://bugs.webkit.org/show_bug.cgi?id=165487
1887         rdar://problem/29539209
1888
1889         Reviewed by Dan Bernstein.
1890
1891         Fix suggested by Dan Bernstein.
1892
1893         * Configurations/DebugRelease.xcconfig:
1894
1895 2016-12-06  Saam Barati  <sbarati@apple.com>
1896
1897         Remove old Wasm object model
1898         https://bugs.webkit.org/show_bug.cgi?id=165481
1899
1900         Reviewed by Keith Miller and Mark Lam.
1901
1902         * testing/Internals.cpp:
1903         (WebCore::Internals::parserMetaData):
1904
1905 2016-12-06  Antoine Quint  <graouts@apple.com>
1906
1907         [Modern Media Controls] Media controls use the fullscreen layout after going from inline to fullscreen to PiP to inline
1908         https://bugs.webkit.org/show_bug.cgi?id=165494
1909
1910         Reviewed by Dean Jackson.
1911
1912         We would only call _updateControlsIfNeeded() when entering or leaving fullscreen, so we going from fullscreen to PiP to
1913         inline would retain fullscreen controls since we would not check for the event that indicates we went back from PiP to
1914         inline. On platforms that support it, we listen to the "webkitpresentationmodechanged" event
1915         to update the controls, and only "webkitfullscreenchange" on others (ie. Yosemite).
1916
1917         Test: media/modern-media-controls/media-controller/media-controller-inline-to-fullscreen-to-pip-to-inline.html
1918
1919         * Modules/modern-media-controls/media/media-controller.js:
1920         (MediaController):
1921         (MediaController.prototype.handleEvent):
1922         (MediaController.prototype._returnMediaLayerToInlineIfNeeded):
1923
1924 2016-12-06  Antoine Quint  <graouts@apple.com>
1925
1926         [Modern Media Controls] Rendering issues with controls bar when captions are on
1927         https://bugs.webkit.org/show_bug.cgi?id=165390
1928
1929         Reviewed by Dean Jackson.
1930
1931         We would face some layout issues with captions due to RenderImage::layoutShadowControls()
1932         expecting a single RenderBox in the media controls shadow root, which was the case with
1933         legacy media controls, but no longer the case with modern media controls. We now host
1934         both the captions and the media controls elements under a single container, and add
1935         an asertion in RenderImage to check that a single RenderBox child exists.
1936
1937         Test: media/modern-media-controls/media-controller/media-controller-single-container.html
1938
1939         * Modules/modern-media-controls/controls/media-controls.css:
1940         (.media-controls-container):
1941         (.media-controls-container,):
1942         (.media-controls-container > *):
1943         (.media-controls):
1944         * Modules/modern-media-controls/controls/text-tracks.css:
1945         (video::-webkit-media-text-track-container):
1946         * Modules/modern-media-controls/media/media-controller.js:
1947         (MediaController):
1948         (MediaController.prototype._updateControlsIfNeeded):
1949         * rendering/RenderImage.cpp:
1950         (WebCore::RenderImage::layoutShadowControls):
1951
1952 2016-12-06  Antoine Quint  <graouts@apple.com>
1953
1954         [Modern Media Controls] Automatically hide the controls bar when the mouse is idle
1955         https://bugs.webkit.org/show_bug.cgi?id=165492
1956
1957         Reviewed by Dean Jackson.
1958
1959         We now automatically hide the controls bar. When the media is no longer paused, the controls remain
1960         visible for 4 seconds, regardless of where the mouse pointer is located. When the user mouses over the
1961         media, the controls become visible and automatically hide 4 seconds after the last time the user has
1962         moved his mouse over the media. When the user mouses out of the media, the controls automatically hide.
1963         When the mouse is over the controls bar, it remains visible. When the media is paused, the controls bar
1964         remain visible regardless of the mouse position.
1965
1966         Tests: media/modern-media-controls/media-controller/media-controller-auto-hide-mouse-enter-and-mouse-leave.html
1967                media/modern-media-controls/media-controller/media-controller-auto-hide-mouse-enter-over-controls-bar.html
1968                media/modern-media-controls/media-controller/media-controller-auto-hide-mouse-leave-after-play.html
1969                media/modern-media-controls/media-controller/media-controller-auto-hide-pause.html
1970                media/modern-media-controls/media-controller/media-controller-auto-hide-rewind-with-mouse-enter.html
1971                media/modern-media-controls/media-controller/media-controller-auto-hide.html
1972
1973         * Modules/modern-media-controls/controls/controls-bar.css: Added.
1974         (.controls-bar):
1975         (.controls-bar.faded):
1976         * Modules/modern-media-controls/controls/controls-bar.js:
1977         (ControlsBar.prototype.get userInteractionEnabled):
1978         (ControlsBar.prototype.set userInteractionEnabled):
1979         (ControlsBar.prototype.get fadesWhileIdle):
1980         (ControlsBar.prototype.set fadesWhileIdle):
1981         (ControlsBar.prototype.get visible):
1982         (ControlsBar.prototype.set visible):
1983         (ControlsBar.prototype.get faded):
1984         (ControlsBar.prototype.set faded):
1985         (ControlsBar.prototype.handleEvent):
1986         (ControlsBar.prototype.commitProperty):
1987         (ControlsBar.prototype._cancelAutoHideTimer):
1988         (ControlsBar.prototype._rewindAutoHideTimer):
1989         (ControlsBar.prototype._autoHideTimerFired):
1990         * Modules/modern-media-controls/controls/media-controls.js:
1991         * Modules/modern-media-controls/media/controls-visibility-support.js:
1992         (ControlsVisibilitySupport.prototype.get mediaEvents):
1993         (ControlsVisibilitySupport.prototype.syncControl):
1994         (ControlsVisibilitySupport):
1995         * Modules/modern-media-controls/media/media-controller.js:
1996         (MediaController.prototype._updateControlsIfNeeded):
1997
1998 2016-12-06  Zalan Bujtas  <zalan@apple.com>
1999
2000         Can not select whole line when using flexbox
2001         https://bugs.webkit.org/show_bug.cgi?id=165299
2002
2003         Reviewed by David Hyatt.
2004
2005         RootInlineBox::selectionTopAdjustedForPrecedingBlock assumes that the preceding block is
2006         always above the current line. However in certain layout contexts (flex as an example) the block before
2007         could just be on the same line as the current one.
2008
2009         This patch checks if we actually need to adjust the selection top to avoid vertical selection overlap.
2010
2011         Test: fast/flexbox/flexbox-fail-to-select-same-line.html
2012
2013         * rendering/RenderBlock.cpp:
2014         (WebCore::RenderBlock::blockBeforeWithinSelectionRoot): Deleted.
2015         * rendering/RenderBlock.h:
2016         * rendering/RootInlineBox.cpp:
2017         (WebCore::blockBeforeWithinSelectionRoot):
2018         (WebCore::RootInlineBox::selectionTopAdjustedForPrecedingBlock):
2019
2020 2016-12-06  Ryosuke Niwa  <rniwa@webkit.org>
2021
2022         Add more assertions to ElementQueue diagnose a bug
2023         https://bugs.webkit.org/show_bug.cgi?id=164814
2024
2025         Reviewed by Alexey Proskuryakov.
2026
2027         Set m_invoking in release builds now that assertions are also run in release builds.
2028
2029         * dom/CustomElementReactionQueue.cpp:
2030         (WebCore::CustomElementReactionStack::ElementQueue::invokeAll):
2031
2032 2016-12-06  Jer Noble  <jer.noble@apple.com>
2033
2034         YouTube sometimes pauses when switching tabs
2035         https://bugs.webkit.org/show_bug.cgi?id=165468
2036
2037         Reviewed by Eric Carlson.
2038
2039         Test: media/media-session-restrictions.html
2040
2041         The bitfield which represents the restrictions for "VideoAudio" media type was unititialized,
2042         leading to what was effectively a random set of restrictions at runtime.
2043
2044         * platform/audio/PlatformMediaSessionManager.cpp:
2045         (WebCore::PlatformMediaSessionManager::resetRestrictions):
2046         * testing/Internals.cpp:
2047         (WebCore::mediaTypeFromString):
2048         (WebCore::Internals::setMediaSessionRestrictions):
2049         (WebCore::Internals::mediaSessionRestrictions):
2050         * testing/Internals.h:
2051         * testing/Internals.idl:
2052
2053 2016-12-06  Mark Lam  <mark.lam@apple.com>
2054
2055         Introduce the concept of Immutable Prototype Exotic Objects to comply with the spec.
2056         https://bugs.webkit.org/show_bug.cgi?id=165227
2057         <rdar://problem/29442665>
2058
2059         Reviewed by Saam Barati.
2060
2061         Make all objects in window.__proto__'s prototype chain immutable prototype exotic
2062         objects.  This gives us roughly equivalent behavior to other browsers.
2063
2064         Firefox's behavior differ slightly in that Firefox will fail any attempted
2065         assignment their __proto__, while the immutable prototype exotic objects will
2066         only fail if the assignment is of a different value.  See
2067         https://tc39.github.io/ecma262/#sec-immutable-prototype-exotic-objects.
2068
2069         Chrome differs in that assignment to window.__proto__ is also handled like an
2070         immutable prototype exotic object.  Instead we adhere to the current HTML spec
2071         that says that the assignment should fail unconditionally.  See
2072         https://html.spec.whatwg.org/#the-windowproxy-exotic-object and
2073         https://html.spec.whatwg.org/#windowproxy-setprototypeof.
2074
2075         If the HTML spec is changed to make the WindowProxy and Location objects into
2076         immutable prototype exotic objects later, we can update to match the spec then.
2077
2078         Test: js/prototype-assignment.html
2079
2080         * bindings/js/JSDOMWindowProperties.h:
2081         * bindings/scripts/CodeGeneratorJS.pm:
2082         (GenerateHeader):
2083         (GeneratePrototypeDeclaration):
2084         * bindings/scripts/IDLAttributes.txt:
2085         * dom/EventTarget.idl:
2086         * page/DOMWindow.idl:
2087
2088 2016-12-06  Andreas Kling  <akling@apple.com>
2089
2090         [Cocoa] Add some memory usage related information to sysdiagnose state dumps
2091         <https://webkit.org/b/165375>
2092         <rdar://problem/29057243>
2093
2094         Reviewed by Darin Adler.
2095
2096         Follow-up to address two more review comments.
2097
2098         * page/PerformanceLogging.cpp:
2099         (WebCore::PerformanceLogging::javaScriptObjectCounts): Use WTFMove() when returning the
2100         HashCountedSet to avoid making an extra copy.
2101
2102 2016-12-06  Dave Hyatt  <hyatt@apple.com>
2103
2104         REGRESSION (Safari 10): Scrolling not working inside height 100% table
2105         https://bugs.webkit.org/show_bug.cgi?id=164366
2106         <rdar://problem/29095535>
2107
2108         Reviewed by Zalan Bujtas.
2109
2110         Added fast/table/overflow-percent-height-regression.html
2111
2112         * rendering/RenderBox.cpp:
2113         (WebCore::RenderBox::computePercentageLogicalHeight):
2114         This patch wasn't merged correctly from Blink. The value returned needed
2115         to be 0, not unset.
2116
2117 2016-12-06  Sam Weinig  <sam@webkit.org>
2118
2119         REGRESSION: media/track LayoutTests are flaky failures
2120         https://bugs.webkit.org/show_bug.cgi?id=165432
2121
2122         Reviewed by Antoine Quint.
2123
2124         RenderThemeMac was caching the media controls script / style sheets
2125         and not being invalidated when the RuntimeEnabledFeatures changed. So,
2126         we can just cache both.
2127
2128         * bindings/generic/RuntimeEnabledFeatures.h:
2129         Initialize m_areModernMediaControlsEnabled to false.
2130
2131         * rendering/RenderThemeMac.h:
2132         * rendering/RenderThemeMac.mm:
2133         (WebCore::RenderThemeMac::mediaControlsStyleSheet):
2134         (WebCore::RenderThemeMac::mediaControlsScript):
2135         Have separate caches for the modern and legacy media control scripts/style sheets.
2136
2137 2016-12-06  Sam Weinig  <sam@webkit.org>
2138
2139         DumpRenderTree crashed in com.apple.WebCore: WebCore::MockMediaEndpoint::generateDtlsInfo const + 23
2140         https://bugs.webkit.org/show_bug.cgi?id=165486
2141         <rdar://problem/29149588>
2142
2143         Reviewed by Tim Horton.
2144
2145         The MockMediaEndpoint can be destroyed before the task created in MockMediaEndpoint::generateDtlsInfo()
2146         gets a chance to run, so guard it with a weak pointer.
2147
2148         * platform/mock/MockMediaEndpoint.cpp:
2149         (WebCore::MockMediaEndpoint::MockMediaEndpoint):
2150         (WebCore::MockMediaEndpoint::generateDtlsInfo):
2151         * platform/mock/MockMediaEndpoint.h:
2152
2153 2016-12-06  Chris Dumez  <cdumez@apple.com>
2154
2155         Add HTML interactive form validation blacklist for some WebKit-specific sites
2156         https://bugs.webkit.org/show_bug.cgi?id=165470
2157         <rdar://problem/29509424>
2158
2159         Reviewed by Simon Fraser.
2160
2161         Add HTML interactive form validation blacklist for some WebKit-specific sites
2162         expecting our old behavior, until they get a chance to be updated.
2163
2164         * html/HTMLFormElement.cpp:
2165         (WebCore::isURLBlacklistedForInteractiveFormValidation):
2166         (WebCore::HTMLFormElement::prepareForSubmission):
2167
2168 2016-12-06  Chris Dumez  <cdumez@apple.com>
2169
2170         Unreviewed, rollout r209050.
2171
2172         This change may impact other popovers than the HTML validation one.
2173         The previous code is safer.
2174
2175         * platform/ValidationBubble.h:
2176         * platform/ios/ValidationBubbleIOS.mm:
2177         (-[WebValidationBubbleDelegate adaptivePresentationStyleForPresentationController:traitCollection:]):
2178         (WebCore::ValidationBubble::setAnchorRect):
2179         * platform/spi/ios/UIKitSPI.h:
2180
2181 2016-12-05  Simon Fraser  <simon.fraser@apple.com>
2182
2183         Improve the behavior of scroll-into-view when the target is inside position:fixed
2184         https://bugs.webkit.org/show_bug.cgi?id=165354
2185
2186         Reviewed by Zalan Bujtas.
2187
2188         The existing RenderLayer::scrollRectToVisible() code paid no heed to whether the
2189         target was inside position:fixed, resulting in unwanted scrolls.
2190
2191         Fix this by plumbing through from the call sites a "insideFixed" flag which we get
2192         when we call localToAbsolute(), and use this flag to avoid scrolling at all if
2193         unzoomed.
2194
2195         If zoomed and we're focussing something inside position:fixed, and if visual viewports
2196         are enabled, we can compute the visual viewport required to reveal the target rect,
2197         which gives us the ideal scroll position.
2198
2199         Fix a bug on non-iOS platforms when zoomed, which is to scale the viewRect since
2200         frameView.visibleContentRect() gives an unscaled rect on those platforms.
2201
2202         Not all callers of scrollRectToVisible() are fixed, but those that are not will get
2203         the current behavior.
2204
2205         Tests: fast/overflow/scroll-anchor-in-position-fixed.html
2206                fast/visual-viewport/zoomed-scroll-into-view-fixed.html
2207                fast/visual-viewport/zoomed-scroll-to-anchor-in-position-fixed.html
2208
2209         * dom/Element.cpp:
2210         (WebCore::Element::scrollIntoView):
2211         (WebCore::Element::scrollIntoViewIfNeeded):
2212         (WebCore::Element::scrollIntoViewIfNotVisible):
2213         (WebCore::Element::updateFocusAppearance):
2214         * editing/FrameSelection.cpp:
2215         (WebCore::FrameSelection::FrameSelection):
2216         (WebCore::FrameSelection::absoluteCaretBounds):
2217         (WebCore::FrameSelection::recomputeCaretRect):
2218         (WebCore::FrameSelection::revealSelection):
2219         * editing/FrameSelection.h:
2220         * editing/VisiblePosition.cpp:
2221         (WebCore::VisiblePosition::absoluteCaretBounds):
2222         * editing/VisiblePosition.h:
2223         * editing/htmlediting.cpp:
2224         (WebCore::absoluteBoundsForLocalCaretRect):
2225         * editing/htmlediting.h:
2226         * page/FrameView.cpp:
2227         (WebCore::FrameView::scrollElementToRect):
2228         (WebCore::FrameView::scrollToAnchor):
2229         * page/PrintContext.cpp:
2230         (WebCore::PrintContext::outputLinkedDestinations):
2231         * rendering/RenderElement.cpp:
2232         (WebCore::RenderElement::getLeadingCorner):
2233         (WebCore::RenderElement::getTrailingCorner):
2234         (WebCore::RenderElement::absoluteAnchorRect):
2235         (WebCore::RenderElement::anchorRect): Deleted.
2236         * rendering/RenderElement.h:
2237         * rendering/RenderLayer.cpp:
2238         (WebCore::RenderLayer::scrollRectToVisible):
2239         (WebCore::RenderLayer::getRectToExpose):
2240         (WebCore::RenderLayer::autoscroll):
2241         * rendering/RenderLayer.h:
2242         * rendering/RenderObject.cpp:
2243         (WebCore::RenderObject::scrollRectToVisible):
2244         * rendering/RenderObject.h:
2245
2246 2016-12-06  Myles C. Maxfield  <mmaxfield@apple.com>
2247
2248         [Cocoa] REGRESSION(r205396): Intermediate CTRuns with initial advances get double counted when glyph origins are enabled
2249         https://bugs.webkit.org/show_bug.cgi?id=165084
2250
2251         Reviewed by Simon Fraser.
2252
2253         When glyph origins are not enabled, an intermediate CTRun's initial advance is simply added
2254         to the previous glyph's advance. However, when glyph origins are enabled, this shouldn't
2255         occur.
2256
2257         Test: fast/text/initial-advance-in-intermediate-run-complex.html
2258
2259         * platform/graphics/mac/ComplexTextController.cpp:
2260         (WebCore::ComplexTextController::adjustGlyphsAndAdvances):
2261
2262 2016-12-06  Simon Fraser  <simon.fraser@apple.com>
2263
2264         Enable visual viewports by default on Mac, and iOS Wk2
2265         https://bugs.webkit.org/show_bug.cgi?id=165452
2266
2267         Reviewed by Zalan Bujtas.
2268
2269         Some tests revealed an issue where, during scrollbar updates, the visual viewport would be
2270         15px wider than the layout viewport, which results in the layout viewport shifting 15px to the
2271         right. Prevent this by disallowing viewport updates during the AdjustViewSize phase; we'll
2272         be called again in postLayoutTasks.
2273
2274         * page/FrameView.cpp:
2275         (WebCore::FrameView::updateLayoutViewport):
2276
2277 2016-12-06  Daniel Bates  <dabates@apple.com>
2278
2279         Cleanup: Rename some HTTP 0.9-specific functions to conform to WebKit Code Style Guidelines
2280         https://bugs.webkit.org/show_bug.cgi?id=165451
2281
2282         Reviewed by Alex Christensen.
2283
2284         Rename Document::shouldEnforceHTTP0_9Sandbox() and ResourceResponseBase::isHttpVersion0_9()
2285         to Document::shouldEnforceHTTP09Sandbox() and ResourceResponseBase::isHTTP09, respectively,
2286         to conform to the WebKit Code Style Guidelines.
2287
2288         * dom/Document.cpp:
2289         (WebCore::Document::initSecurityContext):
2290         (WebCore::Document::shouldEnforceHTTP09Sandbox): Renamed; formerly named shouldEnforceHTTP0_9Sandbox
2291         * dom/Document.h:
2292         * platform/network/ResourceHandle.cpp:
2293         (WebCore::ResourceHandle::didReceiveResponse):
2294         * platform/network/ResourceResponseBase.cpp:
2295         (WebCore::ResourceResponseBase::isHTTP09): Renamed; formerly named isHttpVersion0_9.
2296         * platform/network/ResourceResponseBase.h:
2297
2298 2016-12-06  Daniel Bates  <dabates@apple.com>
2299
2300         Use Vector::uncheckedAppend() in more places
2301         https://bugs.webkit.org/show_bug.cgi?id=164952
2302
2303         Reviewed by Darin Adler.
2304
2305         We can use Vector::uncheckedAppend() whenever the number of items that will be appended
2306         to a vector is equal to or less than its capacity. Using Vector::uncheckedAppend() is
2307         more efficient than Vector::append() as it avoids checking the capacity of the vector
2308         before appending a value to the end of it.
2309
2310         * bindings/js/SerializedScriptValue.cpp:
2311         (WebCore::SerializedScriptValue::SerializedScriptValue):
2312         * contentextensions/DFABytecodeCompiler.cpp:
2313         (WebCore::ContentExtensions::DFABytecodeCompiler::extractJumpTable):
2314         * contentextensions/DFAMinimizer.cpp:
2315         * css/StyleProperties.cpp:
2316         (WebCore::StyleProperties::copyPropertiesInSet): Use Vector::uncheckedAppend() and inline
2317         the assignment of the temporary variable value into the if condition to limit its scope
2318         as it is referenced exactly once in the loop body.
2319         * css/StyleRule.cpp:
2320         (WebCore::StyleRuleGroup::StyleRuleGroup): Use Vector::uncheckedAppend() and write for-loop
2321         using a C++11 range-based for-loop.
2322         * css/parser/CSSParserValues.cpp:
2323         (WebCore::CSSParserSelector::setLangArgumentList):
2324         * cssjit/SelectorCompiler.cpp:
2325         (WebCore::SelectorCompiler::SelectorCodeGenerator::generateElementIsNthLastChildOf):
2326         * dom/DocumentMarkerController.cpp:
2327         (WebCore::updateRenderedRectsForMarker):
2328         * page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
2329         (WebCore::convertToLayoutUnits):
2330         * platform/audio/AudioBus.cpp:
2331         (WebCore::AudioBus::AudioBus):
2332         * platform/graphics/FontCache.cpp:
2333         (WebCore::FontCache::purgeInactiveFontData):
2334         * platform/graphics/PathUtilities.cpp:
2335         (WebCore::polygonsForRect):
2336         * platform/mediastream/CaptureDeviceManager.cpp:
2337         (CaptureDeviceManager::bestSourcesForTypeAndConstraints):
2338         * platform/mediastream/RealtimeMediaSource.cpp:
2339         (WebCore::RealtimeMediaSource::fitnessDistance):
2340         * platform/network/cf/FormDataStreamCFNet.cpp:
2341         (WebCore::formCreate):
2342
2343 2016-12-04  Darin Adler  <darin@apple.com>
2344
2345         Use ASCIICType more, and improve it a little bit
2346         https://bugs.webkit.org/show_bug.cgi?id=165360
2347
2348         Reviewed by Sam Weinig.
2349
2350         * css/CSSGrammar.y.in: Use isASCIIDigit.
2351
2352         * css/parser/CSSParser.cpp:
2353         (WebCore::CSSParser::parseFontFaceUnicodeRange): Use isASCIIHexDigit and
2354         toASCIIHexValue.
2355         (WebCore::isEqualToCSSIdentifier): Use isASCIILower.
2356         * html/FormController.cpp:
2357         (WebCore::isNotFormControlTypeCharacter): Use isASCIILower.
2358         * html/parser/CSSPreloadScanner.cpp:
2359         (WebCore::CSSPreloadScanner::tokenize): Use isASCIIAlpha.
2360         * platform/Decimal.cpp:
2361         (WebCore::Decimal::fromString): Use isASCIIDigit.
2362         * platform/FileSystem.cpp:
2363         (WebCore::decodeFromFilename): Use isASCIIHexDigit and toASCIIHexValue.
2364         * platform/URL.cpp:
2365         (WebCore::isLetterMatchIgnoringCase): Deleted.
2366         (WebCore::isSchemeCharacterMatchIgnoringCase): Deleted.
2367         (WebCore::assertProtocolIsGood): Use isASCIIUpper.
2368         (WebCore::URL::protocolIs): Use isASCIIAlphaCaselessEqual.
2369         (WebCore::URL::parse): Ditto.
2370         (WebCore::protocolIs): Ditto.
2371         (WebCore::protocolIsInHTTPFamily): Ditto.
2372         * platform/URLParser.cpp:
2373         (WeCore::URLParser::parseIPv4Piece): Use isASCIIDigit.
2374
2375         * platform/mac/WebCoreNSURLExtras.mm:
2376         (WebCore::isRussianDomainNameCharacter): Use isASCIIDigit.
2377         (WebCore::allCharactersAllowedByTLDRules): Ditto.
2378         (WebCore::dataWithUserTypedString): Use upperNibbleToASCIIHexDigit and
2379         lowerNibbleToASCIIHexDigit.
2380         (WebCore::dataForURLComponentType): Ditto.
2381         (WebCore::createStringWithEscapedUnsafeCharacters): Ditto.
2382         (WebCore::userVisibleString): Use isASCIIHexDigit, toASCIIHexValue,
2383         upperNibbleToASCIIHexDigit, and lowerNibbleToASCIIHexDigit.
2384         (WebCore::isUserVisibleURL): Use isASCIIHexDigit and toASCIIHexValue.
2385
2386         * platform/network/FormDataBuilder.cpp:
2387         (WebCore::FormDataBuilder::encodeStringAsFormData): Use isASCIIAlphanumeric.
2388
2389         * rendering/mathml/RenderMathMLToken.cpp:
2390         (WebCore::mathVariant): Use isASCIIUpper, isASCIILower, and isASCIIDigit.
2391
2392         * svg/SVGParserUtilities.cpp:
2393         (WebCore::genericParseNumber): Use isASCIIDigit.
2394         * svg/SVGPathStringSource.cpp:
2395         (WebCore::nextCommandHelper): Ditto.
2396         * xml/XPathParser.cpp:
2397         (WebCore::XPath::Parser::lexNumber): Ditto.
2398         (WebCore::XPath::Parser::nextTokenInternal): Ditto.
2399
2400 2016-12-06  Ryan Haddad  <ryanhaddad@apple.com>
2401
2402         Rebaseline bindings tests after r209390.
2403
2404         Unreviewed test gardening.
2405
2406         * bindings/scripts/test/JS/JSTestObj.cpp:
2407         (WebCore::convertDictionaryToJS):
2408         * bindings/scripts/test/JS/JSTestObj.h:
2409
2410 2016-12-06  Tim Horton  <timothy_horton@apple.com>
2411
2412         PDF markup/annotations visible in Mail on macOS but not on iOS
2413         https://bugs.webkit.org/show_bug.cgi?id=165444
2414         <rdar://problem/28942896>
2415
2416         Reviewed by Simon Fraser.
2417
2418         No new tests, but enabled an existing test for this feature.
2419
2420         * platform/graphics/cg/PDFDocumentImage.cpp:
2421         (WebCore::PDFDocumentImage::drawPDFPage):
2422         * platform/spi/cg/CoreGraphicsSPI.h:
2423         Make use of CGContextDrawPDFPageWithAnnotations where available,
2424         to paint the page with annotations.
2425
2426 2016-12-06  Dave Hyatt  <hyatt@apple.com>
2427
2428         [CSS Parser] Turn on the new CSS parser
2429         https://bugs.webkit.org/show_bug.cgi?id=165213
2430
2431         Reviewed by Zalan Bujtas.
2432
2433         * page/Settings.in:
2434
2435 2016-12-06  Dan Bernstein  <mitz@apple.com>
2436
2437         Tried to fix the USE(APPLE_INTERNAL_SDK) build after r209385.
2438
2439         * platform/spi/cocoa/NSTouchBarSPI.h:
2440
2441 2016-12-05  Jer Noble  <jer.noble@apple.com>
2442
2443         [pointer-lock] Cursor should become visible when exiting pointer-lock via ESC key.
2444         https://bugs.webkit.org/show_bug.cgi?id=165377
2445
2446         Reviewed by Darin Adler.
2447
2448         Introduce the concept of "force cursor to become visible" when requesting the pointer be
2449         unlocked.  ESC events will cause the pointer to become visible, while normal key events
2450         will not.
2451
2452         * page/EventHandler.cpp:
2453         (WebCore::EventHandler::keyEvent):
2454         * page/PointerLockController.cpp:
2455         (WebCore::PointerLockController::requestPointerUnlock):
2456         (WebCore::PointerLockController::requestPointerUnlockAndForceCursorVisible):
2457         (WebCore::PointerLockController::documentDetached):
2458         (WebCore::PointerLockController::didLosePointerLock):
2459         * page/PointerLockController.h:
2460
2461         Drive-by fix: apply Darin's feedback to existing functions:
2462
2463         * dom/Document.cpp:
2464         (WebCore::Document::prepareForDestruction):
2465         * dom/Element.cpp:
2466         (WebCore::Element::removedFrom):
2467         * page/PointerLockController.cpp:
2468         (WebCore::PointerLockController::elementRemoved):
2469         (WebCore::PointerLockController::documentDetached):
2470
2471
2472 2016-12-06  Sam Weinig  <sam@webkit.org>
2473
2474         [WebIDL] Add support for converting dictionaries to JS
2475         https://bugs.webkit.org/show_bug.cgi?id=165367
2476
2477         Reviewed by Darin Adler and Alex Christensen.
2478
2479         - Adds support for converting dictionary types to JSValues via JSDOMConvert.
2480         - Adopts that functionality to correct the CryptoKeyPair implementation, which
2481           is supposed to be a dictionary.
2482           (While doing this, I also update places that were passing both a CryptoKey
2483           and CryptoKeyPair to use a Variant, since they always only wanted one.)
2484         - Re-works DOMPromise and DeferredPromise to be based on JSDOMConvert and IDLTypes.
2485
2486         * CMakeLists.txt:
2487         * PlatformEfl.cmake:
2488         * PlatformGTK.cmake:
2489         * PlatformMac.cmake:
2490         * WebCore.xcodeproj/project.pbxproj:
2491         Remove JSCryptoKeyPairCustom.cpp and CryptoKeyPair.cpp.
2492
2493         * Modules/encryptedmedia/legacy/WebKitMediaKeys.cpp:
2494         * Modules/mediasource/MediaSource.cpp:
2495         * Modules/mediasource/SourceBuffer.cpp:
2496         * bindings/js/JSMediaDevicesCustom.cpp:
2497         * html/MediaController.cpp:
2498         * html/track/TextTrack.cpp:
2499         Add some missing includes.
2500
2501         * Modules/applepay/ApplePaySession.cpp:
2502         (WebCore::ApplePaySession::canMakePaymentsWithActiveCard):
2503         (WebCore::ApplePaySession::openPaymentSetup):
2504         * Modules/fetch/FetchBody.cpp:
2505         (WebCore::FetchBody::text):
2506         (WebCore::FetchBody::consume):
2507         (WebCore::FetchBody::loadingFailed):
2508         * Modules/fetch/FetchBodyConsumer.cpp:
2509         (WebCore::FetchBodyConsumer::resolveWithData):
2510         (WebCore::FetchBodyConsumer::resolve):
2511         * Modules/fetch/FetchBodyOwner.cpp:
2512         (WebCore::FetchBodyOwner::blob):
2513         (WebCore::FetchBodyOwner::formData):
2514         (WebCore::FetchBodyOwner::text):
2515         * Modules/fetch/FetchResponse.h:
2516         * Modules/mediastream/MediaDevices.h:
2517         * Modules/mediastream/MediaEndpointPeerConnection.cpp:
2518         (WebCore::MediaEndpointPeerConnection::replaceTrack):
2519         (WebCore::MediaEndpointPeerConnection::replaceTrackTask):
2520         * Modules/mediastream/MediaEndpointPeerConnection.h:
2521         * Modules/mediastream/MediaStreamTrack.cpp:
2522         (WebCore::MediaStreamTrack::applyConstraints):
2523         * Modules/mediastream/PeerConnectionBackend.cpp:
2524         (WebCore::PeerConnectionBackend::setLocalDescription):
2525         (WebCore::PeerConnectionBackend::setRemoteDescription):
2526         (WebCore::PeerConnectionBackend::addIceCandidate):
2527         * Modules/mediastream/PeerConnectionBackend.h:
2528         * Modules/mediastream/RTCPeerConnection.cpp:
2529         (WebCore::RTCPeerConnection::queuedSetLocalDescription):
2530         (WebCore::RTCPeerConnection::queuedSetRemoteDescription):
2531         (WebCore::RTCPeerConnection::queuedAddIceCandidate):
2532         (WebCore::RTCPeerConnection::replaceTrack):
2533         * Modules/mediastream/RTCPeerConnection.h:
2534         * Modules/mediastream/RTCRtpSender.cpp:
2535         (WebCore::RTCRtpSender::replaceTrack):
2536         * Modules/mediastream/RTCRtpSender.h:
2537         * Modules/mediastream/UserMediaRequest.cpp:
2538         (WebCore::UserMediaRequest::deny):
2539         * Modules/webaudio/AudioContext.cpp:
2540         (WebCore::AudioContext::suspend):
2541         (WebCore::AudioContext::resume):
2542         * bindings/js/JSCustomElementRegistryCustom.cpp:
2543         (WebCore::whenDefinedPromise):
2544         * bindings/js/ScriptModuleLoader.cpp:
2545         (WebCore::ScriptModuleLoader::resolve):
2546         (WebCore::ScriptModuleLoader::notifyFinished):
2547         * css/FontFace.h:
2548         * css/FontFaceSet.cpp:
2549         (WebCore::FontFaceSet::load):
2550         * css/FontFaceSet.h:
2551         * dom/CustomElementRegistry.cpp:
2552         (WebCore::CustomElementRegistry::addElementDefinition):
2553         * html/HTMLMediaElement.cpp:
2554         (WebCore::HTMLMediaElement::rejectPendingPlayPromises):
2555         * html/HTMLMediaElement.h:
2556         Update to use new IDLType based promises.
2557
2558         * bindings/generic/IDLTypes.h:
2559         Add a new type, ParameterType, to use as the parameter to
2560         DOMPromise/DeferredPromise.
2561
2562         * bindings/js/JSCryptoKeyPairCustom.cpp:
2563         Removed.
2564
2565         * bindings/js/JSDOMConvert.h:
2566         (WebCore::JSConverter<IDLDictionary<T>>::convert):
2567         Add JSConverter specialization for IDLDictionary. Have it simply forward to a generated
2568         convertDictionaryToJS function, following the convention set in place by IDLEnumeration.
2569
2570         * bindings/js/JSDOMPromise.cpp:
2571         * bindings/js/JSDOMPromise.h:
2572         Re-write to be based on IDLTypes and JSDOMConvert, simplifying the implementation.
2573
2574         * bindings/js/JSSubtleCryptoCustom.cpp:
2575         * bindings/js/JSWebKitSubtleCryptoCustom.cpp:
2576         Update for variant based KeyOrKeyPair.
2577
2578         * bindings/scripts/CodeGeneratorJS.pm:
2579         (GenerateDictionaryHeaderContent):
2580         (GenerateDictionaryImplementationContent):
2581         Add generation of the convertDictionaryToJS function. I made it require the JSGenerateToJSObject
2582         extended attribute for now, as the majority of dictionaries do not need this code generated for them.
2583
2584         * bindings/scripts/test/JS/JSTestObj.cpp:
2585         (WebCore::convertDictionaryToJS):
2586         * bindings/scripts/test/JS/JSTestObj.h:
2587         * bindings/scripts/test/TestObj.idl:
2588         Add JSGenerateToJSObject to a dictionary to test generation in bindings tests.
2589
2590         * crypto/CryptoAlgorithm.h:
2591         Change KeyOrKeyPairCallback to take a Variant<RefPtr<CryptoKey>, CryptoKeyPair>, rather
2592         than two separate arguments.
2593
2594         * crypto/CryptoKeyPair.cpp:
2595         Removed.
2596
2597         * crypto/CryptoKeyPair.h:
2598         Convert to a struct.
2599
2600         * crypto/CryptoKeyPair.idl:
2601         Convert to a dictionary.
2602
2603         * crypto/algorithms/CryptoAlgorithmAES_CBC.cpp:
2604         (WebCore::CryptoAlgorithmAES_CBC::generateKey):
2605         * crypto/algorithms/CryptoAlgorithmAES_KW.cpp:
2606         (WebCore::CryptoAlgorithmAES_KW::generateKey):
2607         * crypto/algorithms/CryptoAlgorithmHMAC.cpp:
2608         (WebCore::CryptoAlgorithmHMAC::generateKey):
2609         * crypto/algorithms/CryptoAlgorithmRSAES_PKCS1_v1_5.cpp:
2610         (WebCore::CryptoAlgorithmRSAES_PKCS1_v1_5::generateKey):
2611         * crypto/algorithms/CryptoAlgorithmRSASSA_PKCS1_v1_5.cpp:
2612         (WebCore::CryptoAlgorithmRSASSA_PKCS1_v1_5::generateKey):
2613         * crypto/algorithms/CryptoAlgorithmRSA_OAEP.cpp:
2614         (WebCore::CryptoAlgorithmRSA_OAEP::generateKey):
2615         * crypto/keys/CryptoKeyRSA.h:
2616         * crypto/mac/CryptoKeyRSAMac.cpp:
2617         (WebCore::CryptoKeyRSA::generatePair):
2618         Update for new signature of KeyOrKeyPairCallback.
2619
2620 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2621
2622         [CSS Parser] Reject invalid hex colors on the fast path
2623         https://bugs.webkit.org/show_bug.cgi?id=165461
2624
2625         Reviewed by Zalan Bujtas.
2626
2627         * css/parser/CSSParserFastPaths.cpp:
2628         (WebCore::fastParseColorInternal):
2629         Make sure to check for success. There was even a FIXME in the code about this!
2630         Matches our old parser's behavior.
2631
2632 2016-12-05  Antoine Quint  <graouts@apple.com>
2633
2634         [Modern Media Controls] Controls bar in fullscreen cannot be dragged
2635         https://bugs.webkit.org/show_bug.cgi?id=165448
2636
2637         Reviewed by Simon Fraser.
2638
2639         We weren't listening to mousemove and mouseup events on the right event target, window in the context
2640         of a fullscreen shadow root makes no sense. We now listen to those events on the media controls, which
2641         we size to fit the whole of the media element's bounds. We then apply a transform to the controls bar
2642         to apply the dragging distance.
2643
2644         Test: media/modern-media-controls/macos-fullscreen-media-controls/macos-fullscreen-media-controls-drag.html
2645
2646         * Modules/modern-media-controls/controls/controls-bar.js:
2647         (ControlsBar.prototype.get translation):
2648         (ControlsBar.prototype.set translation):
2649         (ControlsBar.prototype.commitProperty):
2650         (ControlsBar): Deleted.
2651         * Modules/modern-media-controls/controls/macos-fullscreen-media-controls.css:
2652         (.media-controls.mac.fullscreen):
2653         (.media-controls.mac.fullscreen > .controls-bar):
2654         * Modules/modern-media-controls/controls/macos-fullscreen-media-controls.js:
2655         (MacOSFullscreenMediaControls.prototype._handleMousedown):
2656         (MacOSFullscreenMediaControls.prototype._handleMousemove):
2657         (MacOSFullscreenMediaControls.prototype._handleMouseup):
2658
2659 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2660
2661         [CSS Parser] Allow calc in SVG attributes
2662         https://bugs.webkit.org/show_bug.cgi?id=165459
2663
2664         Reviewed by Zalan Bujtas.
2665
2666         * css/parser/CSSPropertyParserHelpers.cpp:
2667         (WebCore::CSSPropertyParserHelpers::consumeLength):
2668
2669 2016-12-05  Brady Eidson  <beidson@apple.com>
2670
2671         Build fix followup to r209379:
2672
2673         STP 19 fails to launch on 16B255
2674         https://bugs.webkit.org/show_bug.cgi?id=165388
2675         -and corresponding-
2676         rdar://problem/29514476
2677
2678         Rubberstamped by Tim Horton.
2679
2680         No new tests (No behavior change).
2681
2682         * platform/spi/cocoa/NSTouchBarSPI.h: Strategically relocate NS_ASSUME_NONNULL* macros.
2683
2684 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2685
2686         [CSS Parser] Make sure the SVGPaint URI and currentColor style still paints using the URI
2687         https://bugs.webkit.org/show_bug.cgi?id=165457
2688
2689         Reviewed by Zalan Bujtas.
2690
2691         * rendering/svg/SVGResources.cpp:
2692         (WebCore::paintingResourceFromSVGPaint):
2693         Add support for this paint type. The old parser doesn't use it.
2694
2695 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2696
2697         [CSS Parser] Fix dashboard region parsing
2698         https://bugs.webkit.org/show_bug.cgi?id=165456
2699
2700         Reviewed by Zalan Bujtas.
2701
2702         * css/parser/CSSPropertyParser.cpp:
2703         (WebCore::consumeWebkitDashboardRegion):
2704         Make sure to reject whitespace by itself as well as unclosed dashboard-region functions.
2705
2706 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2707
2708         [CSS Parser] Properly reject large numeric values
2709         https://bugs.webkit.org/show_bug.cgi?id=165455
2710
2711         Reviewed by Zalan Bujtas.
2712
2713         The new parser clamped numeric values in both the slow and fast paths to the max
2714         and min float values. The old parser simply allowed the values to be inf, and then
2715         had std::isinf checks to reject.
2716
2717         Blink rejects also even though it clamps, but I could not discern the mechanism by
2718         which they did so. Therefore I am changing the new parser to exactly match the old
2719         parser. Numeric values are no longer clamped, but instead are allowed to be inf, and
2720         isinf checks now exist in the new parser in the same places they do in the old parser.
2721
2722         * css/parser/CSSParserFastPaths.cpp:
2723         (WebCore::parseSimpleLength):
2724         (WebCore::parseSimpleLengthValue):
2725         * css/parser/CSSParserToken.cpp:
2726         (WebCore::CSSParserToken::CSSParserToken):
2727         * css/parser/CSSPropertyParserHelpers.cpp:
2728         (WebCore::CSSPropertyParserHelpers::consumeLength):
2729         (WebCore::CSSPropertyParserHelpers::consumePercent):
2730
2731 2016-12-05  Ricky Mondello  <rmondello@apple.com>
2732
2733         STP 19 fails to launch on 16B255
2734         https://bugs.webkit.org/show_bug.cgi?id=165388
2735         -and corresponding-
2736         rdar://problem/29514476
2737
2738         Reviewed by Tim Horton.
2739
2740         * platform/spi/cocoa/NSTouchBarSPI.h: Re-declare the SPI symbols as weak. The cited crash itself is for a
2741             symbol we weren't handling here, but is part of the same group. Re-declare all four of these symbols to
2742             be safe.
2743
2744 2016-12-05  Antoine Quint  <graouts@apple.com>
2745
2746         ERROR: post-layout: dirty renderer(s) - Encountered with LayoutTest media/modern-media-controls/media-controller/media-controller-fullscreen-ltr.html
2747         https://bugs.webkit.org/show_bug.cgi?id=165312
2748
2749         Reviewed by Simon Fraser.
2750
2751         Reverting part of the code added in https://bugs.webkit.org/show_bug.cgi?id=165287 that triggered
2752         an assertion. We go back to removing previous media controls as we add new ones when the fullscreen
2753         status changes, and simply hide the controls during the animated transition using a CSS pseudo-class.
2754         This also fixes an issue where we wouldn't have removed the previous controls should we have entered
2755         fullscreen in a different way than clicking on the fullscreen button in the media controls.
2756
2757         We restore testing coverage that was fixed due to this assertion.
2758
2759         * Modules/modern-media-controls/controls/media-controls.css:
2760         (:host(:-webkit-animating-full-screen-transition) .media-controls):
2761         * Modules/modern-media-controls/controls/media-controls.js:
2762         (MediaControls.prototype.fadeIn):
2763         (MediaControls.prototype.presentInElement): Deleted.
2764         * Modules/modern-media-controls/media/fullscreen-support.js:
2765         (FullscreenSupport.prototype.buttonWasClicked):
2766         * Modules/modern-media-controls/media/media-controller.js:
2767         (MediaController.prototype._updateControlsIfNeeded):
2768
2769 2016-12-05  Dean Jackson  <dino@apple.com>
2770
2771         MediaDocuments crash with modern media controls
2772         https://bugs.webkit.org/show_bug.cgi?id=165446
2773         <rdar://problem/29524959>
2774
2775         Reviewed by Antoine Quint.
2776
2777         The modern media controls inject a <style> element into
2778         the document, which exposed a bug when used in MediaDocuments.
2779         Such documents were not getting a charset, and the hash table
2780         for the CSSParserContexts was crashing.
2781
2782         Test: media/modern-media-controls/media-documents/insert-style-should-not-crash.html
2783
2784         * css/parser/CSSParserMode.h: Guard against an empty charset.
2785         (WebCore::CSSParserContextHash::hash):
2786         * dom/InlineStyleSheetOwner.cpp: Ask for the charset with fallback.
2787         (WebCore::parserContextForElement):
2788
2789 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2790
2791         [CSS Parser] shape-rendering supports crispEdges rather than crisp-edges
2792         https://bugs.webkit.org/show_bug.cgi?id=165443
2793
2794         Reviewed by Tim Horton.
2795
2796         * css/parser/CSSParserFastPaths.cpp:
2797         (WebCore::CSSParserFastPaths::isValidKeywordPropertyAndValue):
2798
2799 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2800
2801         [CSS Parser] Support setting of custom properties from the CSS OM
2802         https://bugs.webkit.org/show_bug.cgi?id=165442
2803
2804         Reviewed by Tim Horton.
2805
2806         * css/parser/CSSParser.cpp:
2807         (WebCore::CSSParser::parseCustomPropertyValue):
2808         Call in to a new function in CSSParserImpl that can handle
2809         custom property parsing.
2810
2811         * css/parser/CSSParserImpl.cpp:
2812         (WebCore::CSSParserImpl::parseCustomPropertyValue):
2813         New function that is similar to parseValue, but handling custom
2814         properties.
2815
2816         (WebCore::CSSParserImpl::consumeDeclaration):
2817         (WebCore::CSSParserImpl::consumeCustomPropertyValue):
2818         (WebCore::CSSParserImpl::consumeVariableValue): Deleted.
2819         * css/parser/CSSParserImpl.h:
2820         Rename consumeVariableValue to consumeCustomPropertyValue, since that
2821         is more clear. consumeVariableValue sounds like you might be resolving
2822         a variable reference rather than parsing a custom property's value.
2823
2824 2016-12-05  Tim Horton  <timothy_horton@apple.com>
2825
2826         Adopt CAMachPort-as-layer-contents
2827         https://bugs.webkit.org/show_bug.cgi?id=141687
2828         <rdar://problem/19393233>
2829
2830         Reviewed by Darin Adler.
2831
2832         No new tests, just a performance bump.
2833
2834         * platform/spi/cocoa/QuartzCoreSPI.h:
2835         Add some SPI.
2836
2837 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2838
2839         [CSS Parser] calcs on column-width that resolve to 0 should be discarded
2840         https://bugs.webkit.org/show_bug.cgi?id=165439
2841
2842         Reviewed by Dean Jackson.
2843
2844         * css/parser/CSSPropertyParser.cpp:
2845         (WebCore::consumeColumnWidth):
2846
2847 2016-12-05  Dean Jackson  <dino@apple.com>
2848
2849         pointer lock needs to be feature detectable
2850         https://bugs.webkit.org/show_bug.cgi?id=165426
2851         <rdar://problem/29486715>
2852
2853         Reviewed by Antoine Quint.
2854
2855         Annotate the public-facing API for pointer-lock, so
2856         that it is hidden when not available.
2857
2858         Also move the Setting to a RuntimeEnabledFeature, since
2859         pointer-lock is exposed from WebKit as a feature.
2860
2861         Tests: pointer-lock/pointerlock-interface-disabled.html
2862                pointer-lock/pointerlock-interface.html
2863
2864         * bindings/generic/RuntimeEnabledFeatures.h:
2865         (WebCore::RuntimeEnabledFeatures::setPointerLockEnabled):
2866         (WebCore::RuntimeEnabledFeatures::pointerLockEnabled):
2867         * dom/Document.idl:
2868         * dom/Element.idl:
2869         * page/PointerLockController.cpp:
2870         (WebCore::PointerLockController::requestPointerLock):
2871         (WebCore::PointerLockController::requestPointerUnlock):
2872         * page/Settings.in:
2873
2874 2016-12-05  Keith Rollin  <krollin@apple.com>
2875
2876         Reduce number of platformMemoryUsage calls
2877         https://bugs.webkit.org/show_bug.cgi?id=164375
2878
2879         Reviewed by Darin Adler.
2880
2881         Removed the calls to WTF::releaseFastMallocFreeMemory (it's already
2882         called elsewhere in the free-all-memory pipeline) and
2883         malloc_zone_pressure_relief (it should be called by the OS on its own
2884         terms).
2885
2886         No new tests -- no new or changed features.
2887
2888         * platform/cocoa/MemoryPressureHandlerCocoa.mm:
2889         (WebCore::MemoryPressureHandler::ReliefLogger::platformMemoryUsage):
2890
2891 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2892
2893         [CSS Parser] Leave the Animation type alone when the property is invalid
2894         https://bugs.webkit.org/show_bug.cgi?id=165418
2895
2896         Reviewed by Dean Jackson.
2897
2898         * css/CSSToStyleMap.cpp:
2899         (WebCore::CSSToStyleMap::mapAnimationProperty):
2900
2901 2016-12-05  Antti Koivisto  <antti@apple.com>
2902
2903         keyframes do not work when defined inside a style in a shadowRoot
2904         https://bugs.webkit.org/show_bug.cgi?id=164608
2905         <rdar://problem/29210251>
2906
2907         Reviewed by Darin Adler.
2908
2909         With :host and ::slotted rules a keyframe animation affecting an element can be
2910         defined in a style scope different from the element's own scope. Style resolver
2911         loses the scope information when building the RenderStyle so there is no way
2912         to find out the correct scope.
2913
2914         Fix by passing style scope through to style builder and including a scope association
2915         with the animation name. Find the correct scope when resolving keyframes.
2916
2917         Test: fast/shadow-dom/shadow-host-animation.html
2918
2919         * css/CSSToStyleMap.cpp:
2920         (WebCore::CSSToStyleMap::mapAnimationName):
2921
2922             Include scope with the name.
2923
2924         * css/ElementRuleCollector.cpp:
2925         (WebCore::MatchRequest::MatchRequest):
2926         (WebCore::ElementRuleCollector::addMatchedRule):
2927         (WebCore::ElementRuleCollector::sortAndTransferMatchedRules):
2928         (WebCore::ElementRuleCollector::matchAuthorRules):
2929         (WebCore::ElementRuleCollector::matchAuthorShadowPseudoElementRules):
2930         (WebCore::ElementRuleCollector::matchHostPseudoClassRules):
2931         (WebCore::ElementRuleCollector::matchSlottedPseudoElementRules):
2932         (WebCore::ElementRuleCollector::collectMatchingRulesForList):
2933
2934             Replace treeContextOrdinal int with Style::ScopeOrdinal enum carrying the same information.
2935             Simplify the code removing unnecessary use of MatchRequest struct.
2936
2937         (WebCore::compareRules):
2938         * css/ElementRuleCollector.h:
2939         * css/StyleResolver.cpp:
2940         (WebCore::StyleResolver::MatchResult::addMatchedProperties):
2941         (WebCore::StyleResolver::CascadedProperties::setPropertyInternal):
2942         (WebCore::StyleResolver::CascadedProperties::set):
2943         (WebCore::StyleResolver::CascadedProperties::setDeferred):
2944
2945             Pass styleScopeOrdinal through the cascade mechanism
2946
2947         (WebCore::cascadeLevelForIndex):
2948         (WebCore::StyleResolver::CascadedProperties::addMatch):
2949         (WebCore::StyleResolver::CascadedProperties::addImportantMatches):
2950         (WebCore::StyleResolver::CascadedProperties::Property::apply):
2951
2952             Set styleScopeOrdinal in State when applying style.
2953
2954         (WebCore::StyleResolver::CascadedProperties::addStyleProperties): Deleted.
2955
2956             Move the code to the only caller.
2957
2958         * css/StyleResolver.h:
2959         (WebCore::StyleResolver::State::styleScopeOrdinal):
2960         (WebCore::StyleResolver::State::setStyleScopeOrdinal):
2961         * page/animation/CompositeAnimation.cpp:
2962         (WebCore::KeyframeAnimation::KeyframeAnimation):
2963         (WebCore::KeyframeAnimation::resolveKeyframeStyles):
2964
2965             Find the correct scope for resolving keyframes based on the scope ordinal.
2966
2967         * platform/animation/Animation.cpp:
2968         * platform/animation/Animation.h:
2969
2970             Add m_nameStyleScopeOrdinal that tells the scope where the name is defined.
2971
2972         * style/StyleScope.cpp:
2973         (WebCore::Style::Scope::forOrdinal):
2974
2975             Find the scope for ordinal.
2976
2977         * style/StyleScope.h:
2978
2979             Define ScopeOrdinal types.
2980
2981         (WebCore::Style::operator++):
2982
2983 2016-12-05  Dave Hyatt  <hyatt@apple.com>
2984
2985         [CSS Parser] Support glyph-orientation-horizontal and glyph-orientation-vertical
2986         https://bugs.webkit.org/show_bug.cgi?id=165414
2987
2988         Reviewed by Zalan Bujtas.
2989
2990         * css/parser/CSSPropertyParser.cpp:
2991         (WebCore::consumeGlyphOrientation):
2992         (WebCore::CSSPropertyParser::parseSingleValue):
2993         Add support for the glyph-orientation-horizontal and glyph-orientation-vertical
2994         properties. They take an angle and allow unitless values.
2995
2996         * css/parser/CSSPropertyParserHelpers.cpp:
2997         (WebCore::CSSPropertyParserHelpers::consumeAngle):
2998         Fix a bug in the handling of unitless values for angles. Make sure to
2999         actually pass in the value instead of always doing 0. Blink does not
3000         accept unitless values for angles at all, so this is another difference
3001         to investigate for SVG in the future.
3002
3003 2016-12-05  Ryan Haddad  <ryanhaddad@apple.com>
3004
3005         Unreviewed, rolling out r209299.
3006
3007         This change appears to have caused LayoutTest failures on
3008         Sierra WK1.
3009
3010         Reverted changeset:
3011
3012         "Improve the behavior of scroll-into-view when the target is
3013         inside position:fixed"
3014         https://bugs.webkit.org/show_bug.cgi?id=165354
3015         http://trac.webkit.org/changeset/209299
3016
3017 2016-12-05  Andreas Kling  <akling@apple.com>
3018
3019         [Cocoa] Add some memory usage related information to sysdiagnose state dumps
3020         <https://webkit.org/b/165375>
3021         <rdar://problem/29057243>
3022
3023         Reviewed by Darin Adler.
3024
3025         Add a flag to memoryUsageStatistics() to allow gathering of slightly more expensive information.
3026         This mode is used when capturing a state dump for sysdiagnose.
3027
3028         The more expensive information added in this patch relates to information about live objects
3029         and memory on the JavaScript heap.
3030
3031         * WebCore.xcodeproj/project.pbxproj:
3032         * page/PerformanceLogging.cpp:
3033         (WebCore::PerformanceLogging::memoryUsageStatistics):
3034         (WebCore::PerformanceLogging::javaScriptObjectCounts):
3035         (WebCore::PerformanceLogging::didReachPointOfInterest):
3036         * page/PerformanceLogging.h:
3037
3038 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3039
3040         [CSS Parser] Make sure the transform fast path uses WebKitCSSTransformValue
3041         https://bugs.webkit.org/show_bug.cgi?id=165399
3042
3043         Reviewed by Dean Jackson.
3044
3045         * css/parser/CSSParserFastPaths.cpp:
3046         (WebCore::parseTransformTranslateArguments):
3047         (WebCore::parseTransformNumberArguments):
3048         (WebCore::parseSimpleTransformValue):
3049         (WebCore::parseSimpleTransformList):
3050
3051 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3052
3053         [CSS Parser] Fix calc() with -webkit-line-clamp
3054         https://bugs.webkit.org/show_bug.cgi?id=165398
3055
3056         Reviewed by Zalan Bujtas.
3057
3058         Remove the aggressive token type checking up front, since it was causing calc() to not
3059         be allowed.
3060
3061         * css/parser/CSSPropertyParser.cpp:
3062         (WebCore::consumeLineClamp):
3063
3064 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3065
3066         [CSS Parser] Support -webkit-text-decoration
3067         https://bugs.webkit.org/show_bug.cgi?id=165391
3068
3069         Reviewed by Dean Jackson.
3070
3071         * css/parser/CSSPropertyParser.cpp:
3072         (WebCore::CSSPropertyParser::parseShorthand):
3073
3074 2016-12-05  Beth Dakin  <bdakin@apple.com>
3075
3076         STP 19 fails to launch on 16B255
3077         https://bugs.webkit.org/show_bug.cgi?id=165388
3078         -and corresponding-
3079         rdar://problem/29514476
3080
3081         Rubber-stamped by Brady Eidson.
3082
3083         * platform/spi/cocoa/NSTouchBarSPI.h:
3084
3085 2016-12-05  Beth Dakin  <bdakin@apple.com>
3086
3087         STP 19 fails to launch on 16B255
3088         https://bugs.webkit.org/show_bug.cgi?id=165388
3089         -and corresponding-
3090         rdar://problem/29514476
3091
3092         Reviewed by Tim Horton.
3093
3094         Speculative fix.
3095         * platform/spi/cocoa/NSTouchBarSPI.h:
3096
3097 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3098
3099         [CSS Parser] Support the 'alphabetic' keyword for text-underline-position
3100         https://bugs.webkit.org/show_bug.cgi?id=165387
3101
3102         Reviewed by Simon Fraser.
3103
3104         Fixes fast/css3-text/css3-text-decoration/getComputedStyle/getComputedStyle-text-underline-position.html
3105
3106         * css/parser/CSSPropertyParser.cpp:
3107         (WebCore::CSSPropertyParser::parseSingleValue):
3108
3109 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3110
3111         [CSS Parser] Fix assert when unknown properties are encountered.
3112         https://bugs.webkit.org/show_bug.cgi?id=165385
3113
3114         Reviewed by Zalan Bujtas.
3115
3116         * css/CSSToStyleMap.cpp:
3117         (WebCore::CSSToStyleMap::mapAnimationProperty):
3118
3119 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3120
3121         [CSS Parser] Fix rx and ry parsing
3122         https://bugs.webkit.org/show_bug.cgi?id=165383
3123
3124         Reviewed by Dean Jackson.
3125
3126         * css/parser/CSSParserFastPaths.cpp:
3127         (WebCore::isSimpleLengthPropertyID):
3128         rx and ry can be negative in the slow path, so make sure they can be negative in the
3129         fast path too.
3130
3131         * css/parser/CSSPropertyParser.cpp:
3132         (WebCore::consumeRxOrRy):
3133         Disallow auto as a value since we are not equipped to handle it, and it's not clear if it's
3134         even valid.
3135
3136 2016-12-05  Konstantin Tokarev  <annulen@yandex.ru>
3137
3138         Add __STDC_FORMAT_MACROS before inttypes.h is included
3139         https://bugs.webkit.org/show_bug.cgi?id=165374
3140
3141         We need formatting macros like PRIu64 to be available in all places where
3142         inttypes.h header is used. All these usages get inttypes.h definitions
3143         via wtf/Assertions.h header, except SQLiteFileSystem.cpp where formatting
3144         macros are not used anymore since r185129.
3145
3146         This patch fixes multiple build errors with MinGW and reduces number of
3147         independent __STDC_FORMAT_MACROS uses in the code base.
3148
3149         Reviewed by Darin Adler.
3150
3151         No new tests needed.
3152
3153         * platform/sql/SQLiteFileSystem.cpp: Removed unused inttypes.h
3154         inclusion.
3155
3156 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3157
3158         [CSS Parser] Properly fail on bad values for -webkit-clip-path
3159         https://bugs.webkit.org/show_bug.cgi?id=165382
3160
3161         Reviewed by Dean Jackson.
3162
3163         * css/parser/CSSPropertyParser.cpp:
3164         (WebCore::consumeBasicShapeOrBox):
3165
3166 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3167
3168         [CSS Parser] The page-break-* properties are only keyword props for old parser.
3169         https://bugs.webkit.org/show_bug.cgi?id=165381
3170
3171         Reviewed by Dean Jackson.
3172
3173         * css/parser/CSSParser.cpp:
3174         (WebCore::isKeywordPropertyID):
3175         * css/parser/CSSParserFastPaths.cpp:
3176         (WebCore::CSSParserFastPaths::isKeywordPropertyID):
3177
3178 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3179
3180         [CSS Parser] Don't use CSS_PARSER_INTEGER unit for resolved integer calcs.
3181         https://bugs.webkit.org/show_bug.cgi?id=165379
3182
3183         Reviewed by Dean Jackson.
3184
3185         * css/parser/CSSPropertyParserHelpers.cpp:
3186         (WebCore::CSSPropertyParserHelpers::CalcParser::consumeNumber):
3187
3188 2016-12-05  Dave Hyatt  <hyatt@apple.com>
3189
3190         [CSS Parser] Add parseValue support to new parser. Use new parser for UA sheet too if useNewParser is set.
3191         https://bugs.webkit.org/show_bug.cgi?id=165376
3192
3193         Reviewed by Zalan Bujtas.
3194
3195         * css/StyleColor.cpp:
3196         (WebCore::StyleColor::isColorKeyword):
3197         Include system colors when using the fast parseValue path.
3198
3199         * css/parser/CSSParser.cpp:
3200         (WebCore::CSSParser::setupParser):
3201         Add an assert to catch any code path using the old parser when the new parser flag is set.
3202
3203         (WebCore::CSSParser::parseSheet):
3204         (WebCore::CSSParser::parseRule):
3205         (WebCore::CSSParser::parseKeyframeRule):
3206         (WebCore::CSSParser::parseSupportsCondition):
3207         (WebCore::CSSParser::parseValue):
3208         (WebCore::CSSParser::parseSelector):
3209         (WebCore::CSSParser::parseDeclaration):
3210         Patched to use the new parser in UASheetMode as well as other modes when the new parser flag
3211         is set. parseValue is patched to use the new parser's fast paths and to call into CSSParserImpl's
3212         parseValue.
3213
3214         * css/parser/CSSParserFastPaths.cpp:
3215         (WebCore::isSimpleLengthPropertyID):
3216         Support CSSPropertyShapeMargin in the fast path since the old parser did in its fast path.
3217
3218         (WebCore::parseSimpleLengthValue):
3219         Don't ever return unitless numbers. If we accept a unitless number, convert the unit to PX still.
3220
3221         (WebCore::CSSParserFastPaths::parseColor):
3222         Use the CSSValuePool when creating colors on the fast path.
3223
3224         * css/parser/CSSParserImpl.cpp:
3225         (WebCore::CSSParserImpl::parseValue):
3226         * css/parser/CSSParserImpl.h:
3227         Change the return type to be compatible with the old parser's ParseResult flag.
3228
3229 2016-12-05  Konstantin Tokarev  <annulen@yandex.ru>
3230
3231         Removed MediaPlayerPrivateTaskTimer
3232         https://bugs.webkit.org/show_bug.cgi?id=165373
3233
3234         Reviewed by Sam Weinig.
3235
3236         It is not used anywhere since QTSDK removal in r165476.
3237
3238         No new tests needed.
3239
3240         * platform/graphics/win/MediaPlayerPrivateTaskTimer.cpp: Removed.
3241         * platform/graphics/win/MediaPlayerPrivateTaskTimer.h: Removed.
3242
3243 2016-12-04  Dave Hyatt  <hyatt@apple.com>
3244
3245         [CSS Parser] Eliminate in-place lowercasing in the parser.
3246         https://bugs.webkit.org/show_bug.cgi?id=165368
3247
3248         Reviewed by Darin Adler.
3249
3250         Replace the in-place lowercasing that the parser does with new
3251         mechanisms. In-place lowercasing ruins serialization and doesn't
3252         work on CSS parsed from static strings. It also has the side effect
3253         of mutating strings passed in from JavaScript like for querySelectorAll.
3254
3255         For class/id selectors, we now check if the string is lowercase or not.
3256         If it contains uppercase ASCII characters, then we allocate the RareData
3257         for the selector. RareData now has two fields instead of one for the value,
3258         a matching value (all lowercase in quirks mode), and a serializing value (the
3259         original string). Because this is done at the CSSSelector level, the old
3260         parser has been patched as well for these cases.
3261
3262         In addition, in-place lowercasing was done for pseudo-elements, for
3263         media query features, and for attr(). In all of these cases we do
3264         lowercase converting by first checking if it's needed. Serialization will
3265         not retain the original string in these cases, so we may want to revisit
3266         these cases in the future and apply a solution similar to what we did for
3267         selectors.
3268
3269         * css/CSSGrammar.y.in:
3270         * css/CSSSelector.cpp:
3271         (WebCore::CSSSelector::createRareData):
3272         (WebCore::CSSSelector::selectorText):
3273         (WebCore::CSSSelector::RareData::RareData):
3274         (WebCore::CSSSelector::RareData::~RareData):
3275         * css/CSSSelector.h:
3276         (WebCore::CSSSelector::RareData::create):
3277         (WebCore::CSSSelector::setValue):
3278         (WebCore::CSSSelector::value):
3279         (WebCore::CSSSelector::serializingValue):
3280         * css/MediaQueryExp.cpp:
3281         (WebCore::MediaQueryExpression::MediaQueryExpression):
3282         * css/parser/CSSParserToken.cpp:
3283         (WebCore::convertToASCIILowercaseInPlace): Deleted.
3284         (WebCore::CSSParserToken::convertToASCIILowercaseInPlace): Deleted.
3285         * css/parser/CSSParserToken.h:
3286         * css/parser/CSSParserValues.h:
3287         (WebCore::CSSParserSelector::setValue):
3288         * css/parser/CSSPropertyParser.cpp:
3289         (WebCore::consumeAttr):
3290         * css/parser/CSSSelectorParser.cpp:
3291         (WebCore::CSSSelectorParser::consumeId):
3292         (WebCore::CSSSelectorParser::consumeClass):
3293         (WebCore::CSSSelectorParser::consumePseudo):
3294         * css/parser/MediaQueryParser.cpp:
3295         (WebCore::MediaQueryParser::readFeature):
3296
3297 2016-12-04  Gyuyoung Kim  <gyuyoung.kim@webkit.org>
3298
3299         Fix a build break on EFL since r209303.
3300
3301         Unreviewed build fix.
3302
3303         * bindings/js/SerializedScriptValue.cpp:
3304         (WebCore::exceptionForSerializationFailure): Add a return in the end of function.
3305
3306 2016-12-03  Dan Bernstein  <mitz@apple.com>
3307
3308         [Cocoa] Expose InjectedBundlePageEditorClient via the Objective-C bundle SPI
3309         https://bugs.webkit.org/show_bug.cgi?id=165276
3310         <rdar://problem/29467040>
3311
3312         Reviewed by Darin Adler.
3313
3314         * platform/ios/PlatformPasteboardIOS.mm:
3315         (WebCore::PlatformPasteboard::write): Make sure to put the client data on the pasteboard,
3316           like we do on macOS.
3317
3318 2016-12-03  Dave Hyatt  <hyatt@apple.com>
3319
3320         [CSS Parser] Remove line numbers from StyleRule.
3321         https://bugs.webkit.org/show_bug.cgi?id=165361
3322
3323         Reviewed by Simon Fraser.
3324
3325         StyleRules have a concept of a source line that is eventually passed
3326         to the inspector. This was only ever used by normal rules, i.e., ones with
3327         selectors, and set to 0 for all other rules. This line was set to the line number
3328         at which the end of the selector text occurred.
3329
3330         Because Inspector already computes the start and end range for the selector
3331         text, storing a source line on StyleRule ends up being redundant. This patch
3332         gets rid of the source line and uses the end line of the selector text
3333         instead.
3334
3335         * css/CSSGrammar.y.in:
3336         Remove the code that updates the last seen selector line.
3337
3338         * css/CSSKeyframeRule.cpp:
3339         (WebCore::StyleKeyframe::StyleKeyframe):
3340         * css/CSSKeyframesRule.cpp:
3341         (WebCore::StyleRuleKeyframes::StyleRuleKeyframes):
3342         No longer need to pass in a 0 line number.
3343
3344         * css/StyleRule.cpp:
3345         (WebCore::StyleRule::StyleRule):
3346         (WebCore::StyleRule::create):
3347         (WebCore::StyleRule::splitIntoMultipleRulesWithMaximumSelectorComponentCount):
3348         (WebCore::StyleRuleFontFace::StyleRuleFontFace):
3349         (WebCore::StyleRuleGroup::StyleRuleGroup):
3350         (WebCore::StyleRuleCharset::StyleRuleCharset):
3351         (WebCore::StyleRuleNamespace::StyleRuleNamespace):
3352         * css/StyleRule.h:
3353         (WebCore::StyleRuleBase::StyleRuleBase):
3354         (WebCore::StyleRuleBase::sourceLine): Deleted.
3355         * css/StyleRuleImport.cpp:
3356         (WebCore::StyleRuleImport::StyleRuleImport):
3357         Remove m_sourceLine and change the create methods and constructors to not
3358         require a line number.
3359
3360         * css/parser/CSSParser.cpp:
3361         (WebCore::CSSParser::createStyleRule):
3362         Line number no longer needed.
3363
3364         (WebCore::CSSParser::updateLastSelectorLineAndPosition): Deleted.
3365         * css/parser/CSSParser.h:
3366         Got rid of the function that tracks and updates the last seen selector line.
3367
3368         * css/parser/CSSParserImpl.cpp:
3369         (WebCore::CSSParserImpl::consumeStyleRule):
3370         Fix the rule creation in the new parser to not pass in a 0 line number.
3371
3372         * inspector/InspectorStyleSheet.cpp:
3373         (WebCore::buildSourceRangeObject):
3374         (WebCore::InspectorStyleSheet::buildObjectForSelectorList):
3375         (WebCore::InspectorStyleSheet::buildObjectForRule):
3376         * inspector/InspectorStyleSheet.h:
3377         Patch the methods that build up the selector range to return the end line
3378         information for selector text so that it can be set as the source line
3379         for the rule (thus eliminating the need to store the line number on the style
3380         rule itself).
3381
3382 2016-12-02  Sam Weinig  <sam@webkit.org>
3383
3384         optional sequence values not handled correctly by binding generator
3385         https://bugs.webkit.org/show_bug.cgi?id=142562
3386
3387         Also fixes:
3388             Remove non-standard postMessage overload
3389             https://bugs.webkit.org/show_bug.cgi?id=161911
3390         and
3391             Wrong argument order in window.postMessage
3392             https://bugs.webkit.org/show_bug.cgi?id=63141
3393
3394         Reviewed by Darin Adler.
3395
3396         * WebCore.xcodeproj/project.pbxproj:
3397         Remove no longer needed files.
3398
3399         * bindings/generic/IDLTypes.h:
3400         * bindings/js/JSDOMConvert.h:
3401         (WebCore::Converter<IDLObject>::convert):
3402         Add support for the WebIDL object type.
3403
3404         * bindings/js/JSDOMBinding.cpp:
3405         (WebCore::createDOMException):
3406         Add support for throwing stack overflow errors.
3407
3408         * bindings/js/JSDOMWindowCustom.cpp:
3409         (WebCore::handlePostMessage): Deleted.
3410         (WebCore::JSDOMWindow::postMessage): Deleted.
3411         * bindings/js/JSDedicatedWorkerGlobalScopeCustom.cpp: Removed.
3412         * bindings/js/JSMessagePortCustom.cpp:
3413         (WebCore::JSMessagePort::postMessage): Deleted.
3414         (WebCore::extractTransferables): Deleted.
3415         * bindings/js/JSMessagePortCustom.h: Removed.
3416         * bindings/js/JSWorkerCustom.cpp:
3417         (WebCore::JSWorker::postMessage): Deleted.
3418         Remove custom bindings for postMessage.
3419
3420         * bindings/js/SerializedScriptValue.h:
3421         Switch to using enum class.
3422
3423         * bindings/js/SerializedScriptValue.cpp:
3424         (WebCore::SerializedScriptValue::transferArrayBuffers):
3425         (WebCore::SerializedScriptValue::create):
3426         Add new create function that takes the transfer list, processes it, and returns
3427         MessagePorts and SerializedScriptValue / exception.
3428
3429         (WebCore::CloneBase::throwStackOverflow): Deleted.
3430         (WebCore::CloneDeserializer::throwValidationError): Deleted.
3431         Remove uncalled functions.
3432
3433         * bindings/scripts/CodeGenerator.pm:
3434         (IsRefPtrType):
3435         (IsBuiltinType):
3436         * bindings/scripts/CodeGeneratorJS.pm:
3437         (AddToIncludesForIDLType):
3438         (GetBaseIDLType):
3439         Add support for the WebIDL 'object' type.
3440
3441         * bindings/scripts/test/JS/JSTestObj.cpp:
3442         * bindings/scripts/test/TestObj.idl:
3443         Add tests for 'object'.
3444
3445         * dom/ExceptionCode.h:
3446         Add two new ExceptionCodes:
3447           - ExistingExceptionError, to indicate that implementation code threw a JS exception.
3448           - StackOverflowError, to indicate that a stack overflow exception should be thrown.
3449
3450         * dom/MessagePort.cpp:
3451         (WebCore::MessagePort::postMessage):
3452         * dom/MessagePort.h:
3453         * dom/MessagePort.idl:
3454         * page/DOMWindow.cpp:
3455         (WebCore::DOMWindow::postMessage):
3456         * page/DOMWindow.h: