Source/WebCore/ChangeLog

   1 2015-11-29  Simon Fraser  <simon.fraser@apple.com>
   2
   3         Use SVGTransform::SVGTransformType instead of an unsigned short
   4         https://bugs.webkit.org/show_bug.cgi?id=151637
   5
   6         Reviewed by Brady Eidson.
   7
   8         Make 'type' more strongly typed.
   9
  10         * svg/SVGTransformable.cpp:
  11         (WebCore::SVGTransformable::parseTransformValue):
  12         (WebCore::parseAndSkipType):
  13         (WebCore::SVGTransformable::parseTransformType):
  14         (WebCore::SVGTransformable::parseTransformAttribute):
  15         * svg/SVGTransformable.h:
  16
  17 2015-11-27  Brady Eidson  <beidson@apple.com>
  18
  19         Modern IDB: Class-ify IDBGetResult making it impossible to get the data members wrong.
  20         https://bugs.webkit.org/show_bug.cgi?id=151627
  21
  22         Reviewed by Alexey Proskuryakov.
  23
  24         No new tests (No change in behavior).
  25
  26         * CMakeLists.txt:
  27         * WebCore.xcodeproj/project.pbxproj:
  28
  29         * Modules/indexeddb/IDBGetResult.cpp: Added.
  30         (WebCore::IDBGetResult::dataFromBuffer):
  31         (WebCore::IDBGetResult::isolatedCopy):
  32         * Modules/indexeddb/IDBGetResult.h:
  33         (WebCore::IDBGetResult::IDBGetResult):
  34         (WebCore::IDBGetResult::valueBuffer):
  35         (WebCore::IDBGetResult::keyData):
  36         (WebCore::IDBGetResult::primaryKeyData):
  37         (WebCore::IDBGetResult::keyPath):
  38         (WebCore::IDBGetResult::setValueBuffer):
  39         (WebCore::IDBGetResult::setKeyData):
  40         (WebCore::IDBGetResult::setPrimaryKeyData):
  41         (WebCore::IDBGetResult::setKeyPath):
  42         (WebCore::IDBGetResult::dataFromBuffer): Deleted.
  43         (WebCore::IDBGetResult::isolatedCopy): Deleted.
  44
  45         * Modules/indexeddb/client/IDBCursorImpl.cpp:
  46         (WebCore::IDBClient::IDBCursor::setGetResult):
  47         * Modules/indexeddb/client/IDBCursorImpl.h:
  48
  49         * Modules/indexeddb/client/IDBTransactionImpl.cpp:
  50         (WebCore::IDBClient::IDBTransaction::didGetRecordOnServer):
  51
  52         * Modules/indexeddb/legacy/IDBTransactionBackendOperations.cpp:
  53         (WebCore::GetOperation::perform):
  54
  55         * Modules/indexeddb/server/IDBBackingStore.h:
  56
  57         * Modules/indexeddb/server/MemoryCursor.h:
  58
  59         * Modules/indexeddb/server/MemoryIndexCursor.cpp:
  60         (WebCore::IDBServer::MemoryIndexCursor::currentData):
  61
  62         * Modules/indexeddb/server/MemoryObjectStoreCursor.cpp:
  63         (WebCore::IDBServer::MemoryObjectStoreCursor::currentData):
  64
  65         * platform/CrossThreadCopier.h:
  66
  67 2015-11-27  Carlos Garcia Campos  <cgarcia@igalia.com>
  68
  69         [GTK] Do not use the WebCore garbage collector timer
  70         https://bugs.webkit.org/show_bug.cgi?id=151623
  71
  72         Reviewed by Martin Robinson.
  73
  74         Now that garbage collector timers have been implemented in
  75         JavaScriptCore for glib, we don't need to use another Timer in WebCore.
  76
  77         * bindings/js/GCController.cpp:
  78         (WebCore::GCController::garbageCollectSoon):
  79         (WebCore::GCController::garbageCollectNowIfNotDoneRecently):
  80
  81 2015-11-18  Andy Estes  <aestes@apple.com>
  82
  83         [Content Filtering] Crash in DocumentLoader::notifyFinished() when allowing a media document to load
  84         https://bugs.webkit.org/show_bug.cgi?id=151433
  85         rdar://problem/23506594
  86
  87         Reviewed by Alexey Proskuryakov.
  88
  89         When the main resource of a media document commits, WebKit cancels its load since the plug-in or media engine
  90         will do its own loading. If content filtering is enabled, and the filter waits allow the load until the entire
  91         resource is downloaded, then ContentFilter will attempt to call DocumentLoader::notifyFinished() immediately
  92         after delivering the buffered resource data to DocumentLoader. However, delivering the data will have nulled out
  93         DocumentLoader's m_mainResource when the load was cancelled, leading to a crash in notifyFinished().
  94
  95         To resolve this, add a new Stopped state to ContentFilter. Set this state if DocumentLoader clears its main
  96         resource or detaches from its frame. If ContentFilter is in the Stopped state after calling
  97         DocumentLoader::dataReceived(), do not proceed to call DocumentLoader::notifyFinished().
  98
  99         Test: contentfiltering/allow-media-document.html
 100
 101         * loader/ContentFilter.cpp:
 102         (WebCore::ContentFilter::stopFilteringMainResource): Set m_state to Stopped. If m_mainResource is non-null,
 103         removed ContentFilter as a client and set m_mainResource to null.
 104         (WebCore::ContentFilter::notifyFinished): Stopped calling DocumentLoader::notifyFinished() if m_state is Stopped
 105         after calling DocumentLoader::dataReceived().
 106         * loader/ContentFilter.h:
 107         * loader/DocumentLoader.cpp:
 108         (WebCore::DocumentLoader::detachFromFrame): Called ContentFilter::stopFilteringMainResource() instead of setting
 109         m_contentFilter to null.
 110         (WebCore::DocumentLoader::clearMainResource): Ditto.
 111
 112 2015-11-25  Pranjal Jumde  <pjumde@apple.com>
 113
 114         Checks for buffer-overflows when reading characters from textRun
 115         https://bugs.webkit.org/show_bug.cgi?id=151055
 116         <rdar://problem/23251789>
 117
 118         Reviewed by Myles C. Maxfield.
 119
 120         Prevents an off by one error when adding the last font data to the GlyphBuffer.
 121
 122         * Source/WebCore/platform/graphics/WidthIterator.cpp:
 123         * Source/WebCore/platform/graphics/FontCascade.cpp:
 124
 125 2015-11-22  Andy Estes  <aestes@apple.com>
 126
 127         Teach MiniBrowser how to enable the mock content filter
 128         https://bugs.webkit.org/show_bug.cgi?id=151540
 129
 130         Reviewed by Andreas Kling.
 131
 132         Moved the implementation of MockContentFilterEnabler from TestWebKitAPI to here, renamed it to
 133         WebMockContentFilterEnabler, and made it compatible with the legacy Objective-C runtime. Renamed Decision and
 134         DecisionPoint to WebMockContentFilterDecision and WebMockContentFilterDecisionPoint, and changed them from enum
 135         classes to CF_ENUMs so that they can be used by both C++ and Objective-C source files.
 136
 137         * WebCore.xcodeproj/project.pbxproj:
 138         * bindings/js/JSMockContentFilterSettingsCustom.cpp:
 139         (WebCore::JSMockContentFilterSettings::decisionPoint):
 140         (WebCore::JSMockContentFilterSettings::setDecisionPoint):
 141         (WebCore::toDecision):
 142         (WebCore::JSMockContentFilterSettings::decision):
 143         (WebCore::JSMockContentFilterSettings::setDecision):
 144         (WebCore::JSMockContentFilterSettings::unblockRequestDecision):
 145         (WebCore::JSMockContentFilterSettings::setUnblockRequestDecision):
 146         (WebCore::toJSValue): Deleted.
 147         * testing/MockContentFilter.cpp:
 148         (WebCore::MockContentFilter::willSendRequest):
 149         (WebCore::MockContentFilter::responseReceived):
 150         (WebCore::MockContentFilter::addData):
 151         (WebCore::MockContentFilter::finishedAddingData):
 152         (WebCore::MockContentFilter::unblockHandler):
 153         (WebCore::MockContentFilter::maybeDetermineStatus):
 154         * testing/MockContentFilter.h:
 155         * testing/MockContentFilterEnabler.h: Added.
 156         * testing/MockContentFilterEnabler.mm: Added.
 157         (-[WebMockContentFilterEnabler initWithDecision:decisionPoint:blockedString:]):
 158         (-[WebMockContentFilterEnabler initWithCoder:]):
 159         (-[WebMockContentFilterEnabler encodeWithCoder:]):
 160         (-[WebMockContentFilterEnabler enable]):
 161         (-[WebMockContentFilterEnabler dealloc]):
 162         (+[WebMockContentFilterEnabler supportsSecureCoding]):
 163         (-[WebMockContentFilterEnabler copyWithZone:]):
 164         * testing/MockContentFilterSettings.h:
 165         (WebCore::MockContentFilterSettings::decisionPoint):
 166         (WebCore::MockContentFilterSettings::setDecisionPoint):
 167         (WebCore::MockContentFilterSettings::decision):
 168         (WebCore::MockContentFilterSettings::setDecision):
 169         (WebCore::MockContentFilterSettings::unblockRequestDecision):
 170         (WebCore::MockContentFilterSettings::setUnblockRequestDecision):
 171
 172 2015-11-24  Xabier Rodriguez Calvar  <calvaris@igalia.com>
 173
 174         [Streams API] Implement pipeTo method in readable Stream
 175         https://bugs.webkit.org/show_bug.cgi?id=151588
 176
 177         Reviewed by Darin Adler.
 178
 179         Implemented pipeTo method according to the reference implementation in the spec as the spec is not written
 180         yet. It can be found at https://github.com/whatwg/streams/blob/632b26a05f3106650b1ec91239ad5b012e6c64af/reference-implementation/lib/readable-stream.js#L75.
 181
 182         Tests: streams/pipe-to.html
 183                streams/reference-implementation/brand-checks.html
 184                streams/reference-implementation/pipe-through.html
 185                streams/reference-implementation/pipe-to.html
 186                streams/reference-implementation/pipe-to-options.html
 187                streams/reference-implementation/readable-stream-templated
 188
 189         * Modules/streams/ReadableStream.js:
 190         (doPipe): Internal function of pipeTo.
 191         (closeDest): Internal function of pipeTo.
 192         (abortDest): Internal function of pipeTo.
 193         (pipeTo): Implemented as per spec with some other internal functions as helpers.
 194
 195 2015-11-24  Antti Koivisto  <antti@apple.com>
 196
 197         REGRESSION (r190983): Non-element, non-text nodes should not be distributed to slots
 198         https://bugs.webkit.org/show_bug.cgi?id=151566
 199         rdar://problem/23430177
 200
 201         Reviewed by Zalan Bujtas.
 202
 203         We don't invalidate slot assignments except for text or element children. Fix by not
 204         not assigning other nodes to slots as it is not useful.
 205
 206         Test: fast/html/details-comment-crash.html
 207
 208         * dom/SlotAssignment.cpp:
 209         (WebCore::slotNameFromSlotAttribute):
 210         (WebCore::SlotAssignment::findAssignedSlot):
 211         (WebCore::SlotAssignment::assignSlots):
 212
 213 2015-11-23  David Kilzer  <ddkilzer@apple.com>
 214
 215         Hardening against CSSSelector double frees
 216         <http://webkit.org/b/56124>
 217         <rdar://problem/9119036>
 218
 219         Reviewed by Antti Koivisto.
 220
 221         Add some security assertions to catch this issue if it ever
 222         happens in Debug builds, and make changes in
 223         CSSSelector::~CSSSelector() and
 224         CSSSelectorList::deleteSelectors() to prevent obvious issues if
 225         they're ever called twice in Release builds.
 226
 227         No new tests because we don't know how to reproduce this.
 228
 229         * css/CSSSelector.cpp:
 230         (WebCore::CSSSelector::CSSSelector): Initialize
 231         m_destructorHasBeenCalled.
 232         * css/CSSSelector.h:
 233         (WebCore::CSSSelector::m_destructorHasBeenCalled): Add bitfield.
 234         (WebCore::CSSSelector::CSSSelector): Initialize
 235         m_destructorHasBeenCalled.
 236         (WebCore::CSSSelector::~CSSSelector): Add security assertion
 237         that this is never called twice.  Clear out any fields that
 238         would have caused us to dereference an object twice.
 239
 240         * css/CSSSelectorList.cpp:
 241         (WebCore::CSSSelectorList::deleteSelectors): Clear
 242         m_selectorArray when freeing the memory to which it was
 243         pointing.  This prevents re-entrancy issues or calling this
 244         method twice on the same thread.  Also restructure the for()
 245         loop to prevent calling CSSSelector::isLastInSelectorList()
 246         after CSSSelector::~CSSSelector() has been called (via CRBug
 247         241892).
 248
 249 2015-11-23  Brian Burg  <bburg@apple.com>
 250
 251         Web Inspector: inspector settings should not be shared between different inspection levels
 252         https://bugs.webkit.org/show_bug.cgi?id=151151
 253
 254         Reviewed by Timothy Hatcher.
 255
 256         InspectorController and InspectorFrontendClient should know how to compute their
 257         inspection levels, so that this level can be exposed to the Inspector frontend for
 258         preference disambiguation by inspection level.
 259
 260          - A Page with normal web content has an inspection level of 0.
 261          - The first Web Inspector has an inspection level of 1.
 262          - The second Web Inspector that inspects the Inspector has an inspection level of 2.
 263          - And so forth...
 264
 265         For local frontend clients (WK1 and the ProtocolTestStub), the inspection level of
 266         a Page computed as follows:
 267
 268         - If the page's inspector controller has no frontend client, then the Page
 269         is normal web content (level 0).
 270         - If the page's inspector controller has a frontend client, ask for its inspection level.
 271         A frontend client looks at the inspection level reported by inspected page's inspector
 272         controller and increments by one to account for the crossed inspection boundary.
 273
 274         No new tests, this is blocked by test infrastructure fixes (https://webkit.org/b/151573).
 275
 276         * inspector/InspectorController.cpp:
 277         (WebCore::InspectorController::inspectionLevel): Added.
 278         (WebCore::InspectorController::hasInspectorFrontendClient): Deleted.
 279         * inspector/InspectorController.h:
 280         * inspector/InspectorFrontendClient.h:
 281         * inspector/InspectorFrontendClientLocal.cpp:
 282         (WebCore::InspectorFrontendClientLocal::canAttachWindow): Check inspectionLevel().
 283         (WebCore::InspectorFrontendClientLocal::inspectionLevel): Added.
 284         * inspector/InspectorFrontendClientLocal.h:
 285         * inspector/InspectorFrontendHost.cpp:
 286         (WebCore::InspectorFrontendHost::inspectionLevel): Added.
 287         * inspector/InspectorFrontendHost.h:
 288         * inspector/InspectorFrontendHost.idl:
 289         * page/ContextMenuController.cpp:
 290         (WebCore::ContextMenuController::populate): Check inspectionLevel().
 291
 292 2015-11-20  Brian Burg  <bburg@apple.com>
 293
 294         Web Inspector: RemoteInspector should track targets and connections for remote automation
 295         https://bugs.webkit.org/show_bug.cgi?id=151042
 296
 297         Reviewed by Joseph Pecoraro.
 298
 299         Use the new RemoteControllableTarget API method names.
 300
 301         No new tests, no behavior change.
 302
 303         * page/PageDebuggable.cpp:
 304         (WebCore::PageDebuggable::connect):
 305         (WebCore::PageDebuggable::dispatchMessageFromRemote):
 306         (WebCore::PageDebuggable::dispatchMessageFromRemoteFrontend): Deleted.
 307         * page/PageDebuggable.h:
 308
 309 2015-11-23  Brady Eidson  <beidson@apple.com>
 310
 311         Modern IDB: Unskip all indexeddb/mozilla tests that pass.
 312         https://bugs.webkit.org/show_bug.cgi?id=151568
 313
 314         Reviewed by Alex Christensen.
 315
 316         No new tests (Lots of skipped existing tests now pass).
 317
 318         * Modules/indexeddb/shared/IDBError.h:
 319
 320 2015-11-23  Nikos Andronikos <nikos.andronikos-webkit@cisra.canon.com.au>
 321
 322         Add WebCore namespace for generated bindings for supplemental method calls
 323         https://bugs.webkit.org/show_bug.cgi?id=151431
 324
 325         Reviewed by Alex Christensen.
 326
 327         Address build failures for generated supplemental method calls which are in the WebCore namespace.
 328         The bindings are outside the WebCore namespace - so add WebCore namespace specifier where required.
 329
 330         * bindings/scripts/CodeGeneratorJS.pm:
 331         (GenerateHeader):
 332         (GenerateImplementation):
 333         (GenerateParametersCheck):
 334         * bindings/scripts/CodeGeneratorObjC.pm:
 335         (GenerateImplementation):
 336         * bindings/scripts/test/JS/JSTestInterface.cpp:
 337         (WebCore::jsTestInterfaceConstructorSupplementalStaticReadOnlyAttr):
 338         (WebCore::jsTestInterfaceConstructorSupplementalStaticAttr):
 339         (WebCore::jsTestInterfaceSupplementalStr1):
 340         (WebCore::jsTestInterfaceSupplementalStr2):
 341         (WebCore::jsTestInterfaceSupplementalNode):
 342         (WebCore::setJSTestInterfaceConstructorSupplementalStaticAttr):
 343         (WebCore::setJSTestInterfaceSupplementalStr2):
 344         (WebCore::setJSTestInterfaceSupplementalNode):
 345         (WebCore::jsTestInterfacePrototypeFunctionSupplementalMethod1):
 346         (WebCore::jsTestInterfacePrototypeFunctionSupplementalMethod2):
 347         (WebCore::jsTestInterfaceConstructorFunctionSupplementalMethod4):
 348         * bindings/scripts/test/ObjC/DOMTestInterface.mm:
 349         (-[DOMTestInterface supplementalStr1]):
 350         (-[DOMTestInterface supplementalStr2]):
 351         (-[DOMTestInterface setSupplementalStr2:]):
 352         (-[DOMTestInterface supplementalStr3]):
 353         (-[DOMTestInterface setSupplementalStr3:]):
 354         (-[DOMTestInterface supplementalNode]):
 355         (-[DOMTestInterface setSupplementalNode:]):
 356         (-[DOMTestInterface builtinAttribute]):
 357         (-[DOMTestInterface setBuiltinAttribute:]):
 358         (-[DOMTestInterface supplementalMethod1]):
 359         (-[DOMTestInterface supplementalMethod2:objArg:]):
 360         (-[DOMTestInterface supplementalMethod3]):
 361         (-[DOMTestInterface supplementalMethod4]):
 362         (-[DOMTestInterface builtinFunction]):
 363
 364 2015-11-23  Brady Eidson  <beidson@apple.com>
 365
 366         Modern IDB: When a transaction is aborted, call onerror handlers for all in-progress requests.
 367         https://bugs.webkit.org/show_bug.cgi?id=151550
 368
 369         Reviewed by Alex Christensen.
 370
 371         Test: storage/indexeddb/modern/abort-requests-cancelled.html
 372               storage/indexeddb/modern/idbtransaction-objectstore-failures.html (with changes)
 373               storage/indexeddb/modern/index-5.html (with changes)
 374               Various (currently skipped) legacy IDB tests.
 375
 376         * Modules/indexeddb/client/IDBDatabaseImpl.cpp:
 377         (WebCore::IDBClient::IDBDatabase::transaction):
 378         (WebCore::IDBClient::IDBDatabase::startVersionChangeTransaction):
 379         (WebCore::IDBClient::IDBDatabase::didStartTransaction):
 380         (WebCore::IDBClient::IDBDatabase::willCommitTransaction):
 381         (WebCore::IDBClient::IDBDatabase::didCommitTransaction):
 382         (WebCore::IDBClient::IDBDatabase::willAbortTransaction):
 383         (WebCore::IDBClient::IDBDatabase::didAbortTransaction):
 384         (WebCore::IDBClient::IDBDatabase::didCommitOrAbortTransaction):
 385
 386         * Modules/indexeddb/client/IDBTransactionImpl.cpp:
 387         (WebCore::IDBClient::IDBTransaction::abort):
 388         (WebCore::IDBClient::IDBTransaction::abortOnServerAndCancelRequests):
 389         (WebCore::IDBClient::IDBTransaction::didCreateObjectStoreOnServer):
 390         (WebCore::IDBClient::IDBTransaction::didCreateIndexOnServer):
 391         (WebCore::IDBClient::IDBTransaction::didGetRecordOnServer):
 392         (WebCore::IDBClient::IDBTransaction::didDeleteObjectStoreOnServer):
 393         (WebCore::IDBClient::IDBTransaction::didDeleteIndexOnServer):
 394         (WebCore::IDBClient::IDBTransaction::immediateAbort): Deleted.
 395         (WebCore::IDBClient::IDBTransaction::abortOnServer): Deleted.
 396         * Modules/indexeddb/client/IDBTransactionImpl.h:
 397
 398         * Modules/indexeddb/server/MemoryIDBBackingStore.cpp:
 399         (WebCore::IDBServer::MemoryIDBBackingStore::abortTransaction):
 400         (WebCore::IDBServer::MemoryIDBBackingStore::commitTransaction):
 401
 402         * Modules/indexeddb/shared/IDBError.cpp:
 403         (WebCore::idbErrorName):
 404         (WebCore::idbErrorDescription):
 405         * Modules/indexeddb/shared/IDBError.h:
 406
 407         * Modules/indexeddb/shared/IDBResourceIdentifier.cpp:
 408         (WebCore::IDBResourceIdentifier::loggingString):
 409         * Modules/indexeddb/shared/IDBResourceIdentifier.h:
 410
 411 2015-11-23  Brady Eidson  <beidson@apple.com>
 412
 413         Modern IDB: Unskip storage/indexeddb/mozilla/global-data.html.
 414         https://bugs.webkit.org/show_bug.cgi?id=151557
 415
 416         Reviewed by Alex Christensen.
 417
 418         No new tests (Unskipping existing test storage/indexeddb/mozilla/global-data.html).
 419
 420         - Reworking some invalid ASSERTS
 421         - Actually opening pending open-database-requests after a version change transaction completes
 422         - Allow starting new transactions when the version change transaction has *started* finishing,
 423           but before it finishes finishing.
 424
 425         * Modules/indexeddb/client/IDBDatabaseImpl.cpp:
 426         (WebCore::IDBClient::IDBDatabase::transaction):
 427
 428         * Modules/indexeddb/client/IDBTransactionImpl.h:
 429
 430         * Modules/indexeddb/server/MemoryObjectStore.cpp:
 431         (WebCore::IDBServer::MemoryObjectStore::~MemoryObjectStore):
 432
 433         * Modules/indexeddb/server/MemoryObjectStoreCursor.cpp:
 434         (WebCore::IDBServer::MemoryObjectStoreCursor::keyAdded): Deleted.
 435
 436         * Modules/indexeddb/server/UniqueIDBDatabase.cpp:
 437         (WebCore::IDBServer::UniqueIDBDatabase::UniqueIDBDatabase):
 438         (WebCore::IDBServer::UniqueIDBDatabase::handleOpenDatabaseOperations):
 439         (WebCore::IDBServer::UniqueIDBDatabase::commitTransaction):
 440         * Modules/indexeddb/server/UniqueIDBDatabase.h:
 441
 442 2015-11-23  Youenn Fablet  <youenn.fablet@crf.canon.fr>
 443
 444         Remove DOMPromiseWithCallback
 445         https://bugs.webkit.org/show_bug.cgi?id=151565
 446
 447         Reviewed by Eric Carlson.
 448
 449         DOMPromiseWithCallback is no longer used as WebCore JS built-ins are used to the cases where DOMPromiseWithCallback made sense.
 450
 451         No change in behavior.
 452
 453         * Modules/mediastream/MediaDevices.h: Moved from DOMPromiseWithCallback to DOMPromise since the API is now called by JS built-ins.
 454         * bindings/js/JSDOMPromise.h:
 455         (WebCore::DOMPromiseWithCallback::DOMPromiseWithCallback): Deleted.
 456         (WebCore::DOMPromiseIteratorWithCallback::DOMPromiseIteratorWithCallback): Deleted.
 457         (WebCore::Error>::resolve): Deleted.
 458         (WebCore::Error>::reject): Deleted.
 459         (WebCore::Error>::resolveEnd): Deleted.
 460
 461 2015-11-23  Brian Burg  <bburg@apple.com>
 462
 463         Web Inspector: when inspecting the inspector, add the inspection level to the title bar
 464         https://bugs.webkit.org/show_bug.cgi?id=151555
 465
 466         Reviewed by Timothy Hatcher.
 467
 468         * English.lproj/Localizable.strings: add new localized string for alternate inspector title.
 469
 470 2015-11-23  Zan Dobersek  <zdobersek@igalia.com>
 471
 472         [GStreamer] No need to assert the pipeline's bus presence in MediaPlayerPrivateGStreamerBase dtor
 473         https://bugs.webkit.org/show_bug.cgi?id=151558
 474
 475         Reviewed by Carlos Garcia Campos.
 476
 477         In the MediaPlayerPrivateGStreamerBase destructor, there's some leftover code
 478         that acquires a reference to the GStreamer pipeline's bus object and then
 479         just asserts that it's not null. Not very useful, so remove it.
 480
 481         Also use nullptr to null out the m_player member.
 482
 483         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
 484         (WebCore::MediaPlayerPrivateGStreamerBase::~MediaPlayerPrivateGStreamerBase):
 485
 486 2015-11-17  Sergio Villar Senin  <svillar@igalia.com>
 487
 488         ASSERTION FAILED: freeSpace >= 0 in WebCore::RenderGrid::computeTrackSizesForDirection
 489         https://bugs.webkit.org/show_bug.cgi?id=151254
 490
 491         Reviewed by Darin Adler.
 492
 493         When in quirks mode, abnormally huge margins could lead to
 494         negative computations of available logical sizes. We could add
 495         an !document.inQuirksMode() check in the assertion but since
 496         negative freeSpace values are valid (i.e. the condition is not
 497         really part of the contract and everything will work as
 498         expected) I've decided to simply remove it.
 499
 500         Test: fast/css-grid-layout/grid-quirks-mode-huge-margin-crash.html
 501
 502         * rendering/RenderGrid.cpp:
 503         (WebCore::RenderGrid::computeTrackSizesForDirection): Deleted.
 504
 505 2015-11-22  Myles C. Maxfield  <mmaxfield@apple.com>
 506
 507         Font selection should not consult font-variant property
 508         https://bugs.webkit.org/show_bug.cgi?id=151537
 509
 510         Reviewed by Simon Fraser.
 511
 512         In section 4.7 of the CSS Fonts Level 3 spec, it says "[The font-variant and
 513         font-feature-settings] do not affect font selection."
 514
 515         All the other browsers (Chrome, Firefox, and Edge) all obey the spec here. We
 516         are the only one who misbehaves. This patch aligns our behavior with the other
 517         browsers.
 518
 519         Test: fast/text/font-selection-font-variant.html
 520
 521         * css/CSSFontSelector.cpp:
 522         (WebCore::computeTraitsMask): Deleted.
 523         (WebCore::compareFontFaces): Deleted.
 524         (WebCore::CSSFontSelector::getFontFace): Deleted.
 525         * css/CSSParser.cpp:
 526         (WebCore::isValidKeywordPropertyAndValue):
 527         (WebCore::isKeywordPropertyID):
 528         (WebCore::CSSParser::createFontFaceRule):
 529         (WebCore::CSSParser::CSSParser): Deleted.
 530         (WebCore::CSSParser::parseValue): Deleted.
 531         (WebCore::CSSParser::parseDeclaration): Deleted.
 532         (WebCore::CSSParser::clearProperties): Deleted.
 533         (WebCore::CSSParser::parseFontVariant): Deleted.
 534         (WebCore::CSSParser::createStyleRule): Deleted.
 535         (WebCore::CSSParser::deleteFontFaceOnlyValues): Deleted.
 536         * css/CSSParser.h:
 537         * platform/graphics/FontDescription.cpp:
 538         (WebCore::FontDescription::traitsMask): Deleted.
 539         * platform/graphics/win/FontCacheWin.cpp:
 540         (WebCore::traitsInFamilyEnumProc):
 541         * platform/graphics/cocoa/FontCacheCoreText.cpp:
 542         (WebCore::toTraitsMask):
 543         * platform/text/TextFlags.h:
 544
 545 2015-11-22  Carlos Garcia Campos  <cgarcia@igalia.com>
 546
 547         Unreviewed, rolling out r192727.
 548
 549         It made the selections transparent again and broke
 550         /webkit2/WebKitWebView/snapshot
 551
 552         Reverted changeset:
 553
 554         "[GTK] RenderThemeGtk::platformActiveSelectionBackgroundColor,
 555         et. al. should not clobber state of cached GtkStyleContexts"
 556         https://bugs.webkit.org/show_bug.cgi?id=151533
 557         http://trac.webkit.org/changeset/192727
 558
 559 2015-11-22  Michael Catanzaro  <mcatanzaro@igalia.com>
 560
 561         [GTK] RenderThemeGtk::platformActiveSelectionBackgroundColor, et. al. should not clobber state of cached GtkStyleContexts
 562         https://bugs.webkit.org/show_bug.cgi?id=151533
 563
 564         Reviewed by Carlos Garcia Campos.
 565
 566         platformActiveSelectionBackgroundColor(), platformInactiveSelectionBackgroundColor(), etc.
 567         are const functions intended only to return a color used for painting, but since r174929
 568         they also change the state of the cached style contexts we use for GTK_TYPE_ENTRY and
 569         GTK_TYPE_TREE_VIEW. That's wrong; those style contexts should not have any state set. This
 570         could cause theme colors returned by those GtkStyleContexts to change unexpectedly,
 571         depending on whether the state is explicitly set before each use, or whether the theme
 572         actually uses the states.
 573
 574         This didn't cause any regression only because every place using these style contexts
 575         explicitly sets the state of the style contexts before use. In fact, the GtkTreeView style
 576         context is not used anywhere else, and the GtkEntry style context is only used in
 577         paintTextField, which does set the state before use (and then reverts it using
 578         save/restore), so this cannot have broken anything in practice. But it's a landmine waiting
 579         for the next programmer to trip it.
 580
 581         Fix this with a gtk_style_context_save()/gtk_style_context_restore() pair.
 582
 583         * rendering/RenderThemeGtk.cpp:
 584         (WebCore::styleColor):
 585
 586 2015-11-21  Myles C. Maxfield  <mmaxfield@apple.com>
 587
 588         Tiny cleanup in ComplexTextController::collectComplexTextRuns()
 589         https://bugs.webkit.org/show_bug.cgi?id=151534
 590
 591         Reviewed by Zalan Bujtas.
 592
 593         The isMissingGlyph boolean is completely unnecessary. Its entire
 594         responsiblity is duplicated by the "font" pointer.
 595
 596         No new tests because there is no behavior change.
 597
 598         * platform/graphics/mac/ComplexTextController.cpp:
 599         (WebCore::ComplexTextController::collectComplexTextRuns):
 600
 601 == Rolled over to ChangeLog-2015-11-21 ==