Source/WebCore/ChangeLog

   1 2017-07-22  Dan Bernstein  <mitz@apple.com>
   2
   3         Fixed building for macOS 10.12 with the macOS 10.13 SDK after r219644.
   4
   5         * platform/graphics/avfoundation/objc/CDMSessionAVStreamSession.mm:
   6         (WebCore::CDMSessionAVStreamSession::update): Suppress the objc-literal-conversion warning
   7           in one place where we are calling a class method on a Class and the compiler can’t tell
   8           which class it is, so it chooses wrong and emits a warning.
   9
  10 2017-07-22  Chris Dumez  <cdumez@apple.com>
  11
  12         Merge ExceptionBase into DOMException and drop custom toString
  13         https://bugs.webkit.org/show_bug.cgi?id=174755
  14
  15         Reviewed by Sam Weinig.
  16
  17         Merge ExceptionBase into DOMException and drop custom toString, now that DOMException
  18         is the only remaining exception type.
  19
  20         * CMakeLists.txt:
  21         * Modules/mediastream/MediaStream.h:
  22         * WebCore.xcodeproj/project.pbxproj:
  23         * bindings/js/JSDOMExceptionHandling.cpp:
  24         (WebCore::retrieveErrorMessage):
  25         * dom/DOMAllInOne.cpp:
  26         * dom/DOMException.cpp:
  27         (WebCore::DOMException::DOMException):
  28         (WebCore::DOMException::initializeDescription):
  29         * dom/DOMException.h:
  30         (WebCore::DOMException::code):
  31         (WebCore::DOMException::name):
  32         (WebCore::DOMException::message):
  33         * dom/ExceptionBase.cpp: Removed.
  34         * dom/ExceptionBase.h: Removed.
  35
  36 2017-07-14  Filip Pizlo  <fpizlo@apple.com>
  37
  38         It should be easy to decide how WebKit yields
  39         https://bugs.webkit.org/show_bug.cgi?id=174298
  40
  41         Reviewed by Saam Barati.
  42
  43         No new tests because the WebCore change is just a change to how we #include things.
  44
  45         * inspector/InspectorPageAgent.h:
  46         * inspector/TimelineRecordFactory.h:
  47         * workers/Worker.h:
  48         * workers/WorkerGlobalScopeProxy.h:
  49         * workers/WorkerMessagingProxy.h:
  50
  51 2017-07-22  Said Abou-Hallawa  <sabouhallawa@apple.com>
  52
  53         REGRESSION(r219045): A partially loaded image may not be repainted when its complete frame finishes decoding
  54         https://bugs.webkit.org/show_bug.cgi?id=174230
  55
  56         Reviewed by Simon Fraser.
  57
  58         Because of r219045, we now only repaint the CachedImageClinets which tried
  59         to draw the image but they could not because they have to wait for the image
  60         decoding to finish. This was done by keeping a HashSet of these clients
  61         and make CachedImage own it. This HashSet is cleared once the image frame
  62         finishes decoding and all the waited clients are repainted.
  63
  64         But Multiple asynchronous image decoding requests are allowed for the same
  65         frame if new data is added to the image source. If we tried to draw the
  66         same image twice before it finishes decoding the first request, we will
  67         not be to record this second request since the HashSet will not add the
  68         same client twice. When he second request finishes decoding, CachedImage
  69         will not repaint any client since its HashSet is empty.
  70
  71         To fix this problem we can do the following. When an image frame finishes
  72         decoding, CachedImage will keep its HashSet of pending drawing clients as
  73         long as the image frame is a partially loaded frame.
  74
  75         * loader/cache/CachedImage.cpp:
  76         (WebCore::CachedImage::CachedImageObserver::imageFrameAvailable):
  77         (WebCore::CachedImage::imageFrameAvailable):
  78         * loader/cache/CachedImage.h:
  79         * platform/graphics/BitmapImage.cpp:
  80         (WebCore::BitmapImage::destroyDecodedData):
  81         (WebCore::BitmapImage::dataChanged):
  82         (WebCore::BitmapImage::setCurrentFrameDecodingStatusIfNecessary):
  83         (WebCore::BitmapImage::draw):
  84         (WebCore::BitmapImage::internalStartAnimation):
  85         (WebCore::BitmapImage::internalAdvanceAnimation):
  86         (WebCore::BitmapImage::imageFrameAvailableAtIndex):
  87         * platform/graphics/BitmapImage.h:
  88         * platform/graphics/ImageFrame.cpp:
  89         (WebCore::ImageFrame::decodingStatus):
  90         * platform/graphics/ImageFrame.h: Move DecodingStatus out of this class
  91         to ImageTypes.h to avoid adding other header files to ImageObvsever.h
  92         * platform/graphics/ImageFrameCache.cpp:
  93         (WebCore::ImageFrameCache::setNativeImage):
  94         (WebCore::ImageFrameCache::cacheMetadataAtIndex):
  95         (WebCore::ImageFrameCache::cacheNativeImageAtIndex):
  96         (WebCore::ImageFrameCache::cacheNativeImageAtIndexAsync):
  97         (WebCore::ImageFrameCache::requestFrameAsyncDecodingAtIndex):
  98         (WebCore::ImageFrameCache::frameDecodingStatusAtIndex):
  99         * platform/graphics/ImageFrameCache.h:
 100         * platform/graphics/ImageObserver.h:
 101         * platform/graphics/ImageSource.h:
 102         (WebCore::ImageSource::frameDecodingStatusAtIndex):
 103         * platform/graphics/ImageTypes.h:
 104         * platform/image-decoders/bmp/BMPImageReader.cpp:
 105         (WebCore::BMPImageReader::decodeBMP):
 106         * platform/image-decoders/gif/GIFImageDecoder.cpp:
 107         (WebCore::GIFImageDecoder::frameComplete):
 108         (WebCore::GIFImageDecoder::initFrameBuffer):
 109         * platform/image-decoders/jpeg/JPEGImageDecoder.cpp:
 110         (WebCore::JPEGImageDecoder::outputScanlines):
 111         (WebCore::JPEGImageDecoder::jpegComplete):
 112         * platform/image-decoders/png/PNGImageDecoder.cpp:
 113         (WebCore::PNGImageDecoder::rowAvailable):
 114         (WebCore::PNGImageDecoder::pngComplete):
 115         (WebCore::PNGImageDecoder::frameComplete):
 116         * platform/image-decoders/webp/WEBPImageDecoder.cpp:
 117         (WebCore::WEBPImageDecoder::decode):
 118         * rendering/RenderBoxModelObject.cpp:
 119         (WebCore::RenderBoxModelObject::decodingModeForImageDraw):
 120
 121 2017-07-22  Timothy Horton  <timothy_horton@apple.com>
 122
 123         Drag and Drop preview image for Twitter link is the wrong shape
 124         https://bugs.webkit.org/show_bug.cgi?id=174731
 125         <rdar://problem/33335616>
 126
 127         * dom/Range.cpp:
 128         (WebCore::Range::absoluteRectsForRangeInText):
 129         * page/TextIndicator.cpp:
 130         (WebCore::initializeIndicator):
 131         Apply some post-landing review feedback for r219756.
 132
 133 2017-07-21  Chris Dumez  <cdumez@apple.com>
 134
 135         Drop IDBDatabaseException class
 136         https://bugs.webkit.org/show_bug.cgi?id=174743
 137
 138         Reviewed by Darin Adler.
 139
 140         Drop IDBDatabaseException class and use DOMException directly. IDBDatabaseException was
 141         only an internal concept and was already exposed to the Web as a DOMException. This
 142         simplifies the code a bit and will allow for further simplification in a follow-up patch.
 143
 144         No new tests, no Web-facing behavior change, except for some exception messages
 145         being slightly different.
 146
 147         * CMakeLists.txt:
 148         * Modules/indexeddb/IDBCursor.cpp:
 149         (WebCore::IDBCursor::update):
 150         (WebCore::IDBCursor::advance):
 151         (WebCore::IDBCursor::continuePrimaryKey):
 152         (WebCore::IDBCursor::continueFunction):
 153         (WebCore::IDBCursor::deleteFunction):
 154         * Modules/indexeddb/IDBDatabase.cpp:
 155         (WebCore::IDBDatabase::createObjectStore):
 156         (WebCore::IDBDatabase::transaction):
 157         (WebCore::IDBDatabase::deleteObjectStore):
 158         * Modules/indexeddb/IDBDatabaseError.h: Removed.
 159         * Modules/indexeddb/IDBDatabaseException.cpp: Removed.
 160         * Modules/indexeddb/IDBDatabaseException.h: Removed.
 161         * Modules/indexeddb/IDBFactory.cpp:
 162         (WebCore::IDBFactory::cmp):
 163         * Modules/indexeddb/IDBIndex.cpp:
 164         (WebCore::IDBIndex::setName):
 165         (WebCore::IDBIndex::openCursor):
 166         (WebCore::IDBIndex::openKeyCursor):
 167         (WebCore::IDBIndex::count):
 168         (WebCore::IDBIndex::doCount):
 169         (WebCore::IDBIndex::get):
 170         (WebCore::IDBIndex::doGet):
 171         (WebCore::IDBIndex::getKey):
 172         (WebCore::IDBIndex::doGetKey):
 173         (WebCore::IDBIndex::getAll):
 174         (WebCore::IDBIndex::getAllKeys):
 175         * Modules/indexeddb/IDBKeyRange.cpp:
 176         (WebCore::IDBKeyRange::only):
 177         (WebCore::IDBKeyRange::lowerBound):
 178         (WebCore::IDBKeyRange::upperBound):
 179         (WebCore::IDBKeyRange::bound):
 180         (WebCore::IDBKeyRange::includes):
 181         * Modules/indexeddb/IDBObjectStore.cpp:
 182         (WebCore::IDBObjectStore::setName):
 183         (WebCore::IDBObjectStore::openCursor):
 184         (WebCore::IDBObjectStore::openKeyCursor):
 185         (WebCore::IDBObjectStore::get):
 186         (WebCore::IDBObjectStore::getKey):
 187         (WebCore::IDBObjectStore::putOrAdd):
 188         (WebCore::IDBObjectStore::doDelete):
 189         (WebCore::IDBObjectStore::deleteFunction):
 190         (WebCore::IDBObjectStore::clear):
 191         (WebCore::IDBObjectStore::createIndex):
 192         (WebCore::IDBObjectStore::index):
 193         (WebCore::IDBObjectStore::deleteIndex):
 194         (WebCore::IDBObjectStore::count):
 195         (WebCore::IDBObjectStore::doCount):
 196         (WebCore::IDBObjectStore::getAll):
 197         (WebCore::IDBObjectStore::getAllKeys):
 198         * Modules/indexeddb/IDBOpenDBRequest.cpp:
 199         (WebCore::IDBOpenDBRequest::fireErrorAfterVersionChangeCompletion):
 200         * Modules/indexeddb/IDBRequest.cpp:
 201         (WebCore::IDBRequest::result):
 202         (WebCore::IDBRequest::error):
 203         (WebCore::IDBRequest::uncaughtExceptionInEventHandler):
 204         * Modules/indexeddb/IDBTransaction.cpp:
 205         (WebCore::IDBTransaction::objectStore):
 206         (WebCore::IDBTransaction::abort):
 207         (WebCore::IDBTransaction::abortOnServerAndCancelRequests):
 208         (WebCore::IDBTransaction::didCreateIndexOnServer):
 209         (WebCore::IDBTransaction::putOrAddOnServer):
 210         * Modules/indexeddb/server/IndexValueStore.cpp:
 211         (WebCore::IDBServer::IndexValueStore::addRecord):
 212         * Modules/indexeddb/server/MemoryIDBBackingStore.cpp:
 213         (WebCore::IDBServer::MemoryIDBBackingStore::beginTransaction):
 214         (WebCore::IDBServer::MemoryIDBBackingStore::abortTransaction):
 215         (WebCore::IDBServer::MemoryIDBBackingStore::commitTransaction):
 216         (WebCore::IDBServer::MemoryIDBBackingStore::createObjectStore):
 217         (WebCore::IDBServer::MemoryIDBBackingStore::deleteObjectStore):
 218         (WebCore::IDBServer::MemoryIDBBackingStore::renameObjectStore):
 219         (WebCore::IDBServer::MemoryIDBBackingStore::clearObjectStore):
 220         (WebCore::IDBServer::MemoryIDBBackingStore::createIndex):
 221         (WebCore::IDBServer::MemoryIDBBackingStore::deleteIndex):
 222         (WebCore::IDBServer::MemoryIDBBackingStore::renameIndex):
 223         (WebCore::IDBServer::MemoryIDBBackingStore::deleteRange):
 224         (WebCore::IDBServer::MemoryIDBBackingStore::addRecord):
 225         (WebCore::IDBServer::MemoryIDBBackingStore::getRecord):
 226         (WebCore::IDBServer::MemoryIDBBackingStore::getAllRecords):
 227         (WebCore::IDBServer::MemoryIDBBackingStore::getIndexRecord):
 228         (WebCore::IDBServer::MemoryIDBBackingStore::getCount):
 229         (WebCore::IDBServer::MemoryIDBBackingStore::generateKeyNumber):
 230         (WebCore::IDBServer::MemoryIDBBackingStore::openCursor):
 231         (WebCore::IDBServer::MemoryIDBBackingStore::iterateCursor):
 232         * Modules/indexeddb/server/MemoryIndex.cpp:
 233         (WebCore::IDBServer::MemoryIndex::putIndexKey):
 234         * Modules/indexeddb/server/MemoryObjectStore.cpp:
 235         (WebCore::IDBServer::MemoryObjectStore::createIndex):
 236         (WebCore::IDBServer::MemoryObjectStore::deleteIndex):
 237         * Modules/indexeddb/server/SQLiteIDBBackingStore.cpp:
 238         (WebCore::IDBServer::SQLiteIDBBackingStore::getOrEstablishDatabaseInfo):
 239         (WebCore::IDBServer::SQLiteIDBBackingStore::beginTransaction):
 240         (WebCore::IDBServer::SQLiteIDBBackingStore::abortTransaction):
 241         (WebCore::IDBServer::SQLiteIDBBackingStore::commitTransaction):
 242         (WebCore::IDBServer::SQLiteIDBBackingStore::createObjectStore):
 243         (WebCore::IDBServer::SQLiteIDBBackingStore::deleteObjectStore):
 244         (WebCore::IDBServer::SQLiteIDBBackingStore::renameObjectStore):
 245         (WebCore::IDBServer::SQLiteIDBBackingStore::clearObjectStore):
 246         (WebCore::IDBServer::SQLiteIDBBackingStore::createIndex):
 247         (WebCore::IDBServer::SQLiteIDBBackingStore::uncheckedHasIndexRecord):
 248         (WebCore::IDBServer::SQLiteIDBBackingStore::uncheckedPutIndexKey):
 249         (WebCore::IDBServer::SQLiteIDBBackingStore::uncheckedPutIndexRecord):
 250         (WebCore::IDBServer::SQLiteIDBBackingStore::deleteIndex):
 251         (WebCore::IDBServer::SQLiteIDBBackingStore::renameIndex):
 252         (WebCore::IDBServer::SQLiteIDBBackingStore::keyExistsInObjectStore):
 253         (WebCore::IDBServer::SQLiteIDBBackingStore::deleteUnusedBlobFileRecords):
 254         (WebCore::IDBServer::SQLiteIDBBackingStore::deleteRecord):
 255         (WebCore::IDBServer::SQLiteIDBBackingStore::deleteRange):
 256         (WebCore::IDBServer::SQLiteIDBBackingStore::updateAllIndexesForAddRecord):
 257         (WebCore::IDBServer::SQLiteIDBBackingStore::addRecord):
 258         (WebCore::IDBServer::SQLiteIDBBackingStore::getBlobRecordsForObjectStoreRecord):
 259         (WebCore::IDBServer::SQLiteIDBBackingStore::getRecord):
 260         (WebCore::IDBServer::SQLiteIDBBackingStore::getAllObjectStoreRecords):
 261         (WebCore::IDBServer::SQLiteIDBBackingStore::getAllIndexRecords):
 262         (WebCore::IDBServer::SQLiteIDBBackingStore::getIndexRecord):
 263         (WebCore::IDBServer::SQLiteIDBBackingStore::uncheckedGetIndexRecordForOneKey):
 264         (WebCore::IDBServer::SQLiteIDBBackingStore::getCount):
 265         (WebCore::IDBServer::SQLiteIDBBackingStore::uncheckedGetKeyGeneratorValue):
 266         (WebCore::IDBServer::SQLiteIDBBackingStore::uncheckedSetKeyGeneratorValue):
 267         (WebCore::IDBServer::SQLiteIDBBackingStore::generateKeyNumber):
 268         (WebCore::IDBServer::SQLiteIDBBackingStore::revertGeneratedKeyNumber):
 269         (WebCore::IDBServer::SQLiteIDBBackingStore::maybeUpdateKeyGeneratorNumber):
 270         (WebCore::IDBServer::SQLiteIDBBackingStore::openCursor):
 271         (WebCore::IDBServer::SQLiteIDBBackingStore::iterateCursor):
 272         * Modules/indexeddb/server/SQLiteIDBTransaction.cpp:
 273         (WebCore::IDBServer::SQLiteIDBTransaction::begin):
 274         (WebCore::IDBServer::SQLiteIDBTransaction::commit):
 275         (WebCore::IDBServer::SQLiteIDBTransaction::abort):
 276         * Modules/indexeddb/server/UniqueIDBDatabase.cpp:
 277         (WebCore::IDBServer::UniqueIDBDatabase::performCurrentOpenOperation):
 278         (WebCore::IDBServer::UniqueIDBDatabase::deleteObjectStore):
 279         (WebCore::IDBServer::UniqueIDBDatabase::renameObjectStore):
 280         (WebCore::IDBServer::UniqueIDBDatabase::deleteIndex):
 281         (WebCore::IDBServer::UniqueIDBDatabase::renameIndex):
 282         (WebCore::IDBServer::UniqueIDBDatabase::performPutOrAdd):
 283         (WebCore::IDBServer::UniqueIDBDatabase::commitTransaction):
 284         (WebCore::IDBServer::UniqueIDBDatabase::abortTransaction):
 285         * Modules/indexeddb/shared/IDBError.cpp:
 286         (WebCore::IDBError::name):
 287         (WebCore::IDBError::message):
 288         (WebCore::IDBError::toDOMError):
 289         * Modules/indexeddb/shared/IDBError.h:
 290         (WebCore::IDBError::userDeleteError):
 291         (WebCore::IDBError::isNull):
 292         * WebCore.xcodeproj/project.pbxproj:
 293         * bindings/js/JSDOMExceptionHandling.cpp:
 294         (WebCore::createDOMException):
 295
 296         * dom/DOMException.cpp:
 297         (WebCore::DOMException::initializeDescription):
 298         Initialize description->code from using the actual error codes in the
 299         coreExceptions table, instead of initializing it using |ec|. |ec| is
 300         the index in the ExceptionCode enumeration and it matches the error
 301         code only for legacy exceptions. New exceptions such as DataError
 302         should have a code of 0. Without this change, w3c for IDB would start
 303         failing.
 304
 305         * dom/ExceptionCodeDescription.cpp:
 306         (WebCore::ExceptionCodeDescription::ExceptionCodeDescription):
 307         * dom/ExceptionCodeDescription.h:
 308
 309 2017-07-21  Sam Weinig  <sam@webkit.org>
 310
 311         [WebIDL] Simplify XMLHttpRequest bindings by removing built-ins
 312         https://bugs.webkit.org/show_bug.cgi?id=174740
 313
 314         Reviewed by Chris Dumez.
 315
 316         The behavior we were getting from XMLHttpRequest.js can be implemented
 317         with the [CachedAttribute] extended attribute. This reduces the number
 318         of JS -> C++ jumps required and makes the code more similar to other
 319         bindings that need caching of an IDL any value.
 320
 321         * CMakeLists.txt:
 322         * DerivedSources.make:
 323         * xml/XMLHttpRequest.js: Removed.
 324         Remove XMLHttpRequest.js.
 325
 326         * bindings/js/JSXMLHttpRequestCustom.cpp:
 327         (WebCore::JSXMLHttpRequest::response):
 328         (WebCore::JSXMLHttpRequest::retrieveResponse): Deleted.
 329         * xml/XMLHttpRequest.idl:
 330         Replace builtin with a [CachedAttribute] and [CustomGetter]
 331         for the response property.
 332
 333 2017-07-21  Timothy Horton  <timothy_horton@apple.com>
 334
 335         Drag and Drop preview image for Twitter link is the wrong shape
 336         https://bugs.webkit.org/show_bug.cgi?id=174731
 337         <rdar://problem/33335616>
 338
 339         Reviewed by Zalan Bujtas.
 340
 341         TextIndicator uses Range::borderAndTextQuads and ::absoluteTextRects
 342         in order to get the rects of the indicated text. Currently, these
 343         functions do not respect clipping, so clipped-out text (e.g. as seen
 344         inside links on Twitter) generates lots of meaningless indicated rects.
 345
 346         * page/TextIndicator.cpp:
 347         (WebCore::estimatedBackgroundColorForRange):
 348         (WebCore::hasAnyIllegibleColors):
 349         Change adjustTextIndicatorDataOptionsForEstimatedColorsIfNecessary
 350         to instead be hasAnyIllegibleColors, and referred to in the same place
 351         as hasNonInlineOrReplacedElements, so that it follows the same
 352         upgrade path (leaving textRects empty, so that it is later filled in
 353         with the absoluteBoundingRect). This was a mistake in r219033, which
 354         instead would end up painting all content, but filling in textRects
 355         with the actual individual text rects.
 356
 357         This alone changes the behavior on Twitter from lots of jagged misplaced
 358         rects to a too-large bounding rect. Combined with the following changes,
 359         the bounding rect is reduced to the right size:
 360
 361         (WebCore::initializeIndicator):
 362         Adopt the new Range::borderAndTextQuads and ::absoluteTextRects parameter
 363         and opt-in to respecting clipping for text rects.
 364
 365         * dom/DOMRectList.cpp:
 366         (WebCore::DOMRectList::DOMRectList):
 367         * dom/DOMRectList.h:
 368         (WebCore::DOMRectList::create):
 369         Add a DOMRectList constructor and create() that take FloatRects, similar
 370         to the one that takes FloatQuads, but without the boundingRect() calls.
 371
 372         * dom/Document.h:
 373         * dom/Document.cpp:
 374         (WebCore::Document::convertAbsoluteToClientRects):
 375         Add convertAbsoluteToClientRects, similar to covertAbsoluteToClientQuads,
 376         except acting on rects instead of quads.
 377
 378         * dom/Range.cpp:
 379         (WebCore::Range::absoluteRectsForRangeInText):
 380         (WebCore::Range::absoluteTextRects):
 381         (WebCore::Range::getClientRects):
 382         (WebCore::Range::borderAndTextRects):
 383         (WebCore::Range::boundingRect):
 384         (WebCore::Range::absoluteBoundingRect):
 385         (WebCore::Range::borderAndTextQuads): Deleted.
 386         * dom/Range.h:
 387         Replace borderAndTextQuads with borderAndTextRects, because all callers
 388         just ended up calling boundingBox() on the quads.
 389
 390         Factor absoluteRectsForRangeInText out of absoluteTextRects and
 391         borderAndTextQuads, and teach it to optionally intersect the text rects
 392         with their renderer's absoluteClippedOverflowRect.
 393
 394 2017-07-21  Per Arne Vollan  <pvollan@apple.com>
 395
 396         Round-tripping stroke-width styles through getComputedStyle cause the text to gain a stroke.
 397         https://bugs.webkit.org/show_bug.cgi?id=174701
 398         <rdar://problem/32903679>
 399
 400         Reviewed by Simon Fraser.
 401
 402         The initial value of stroke-color should be transparent, see https://www.w3.org/TR/fill-stroke-3/#stroke-color.
 403         Also, there is no need to set the graphics context in text stroke mode, if the stroke color is not visible.
 404
 405         Test: fast/css/round-trip-stroke-width-using-computed-style.html
 406
 407         * rendering/TextPaintStyle.cpp:
 408         (WebCore::updateGraphicsContext):
 409         * rendering/style/RenderStyle.h:
 410         (WebCore::RenderStyle::initialStrokeWidth):
 411         (WebCore::RenderStyle::initialStrokeColor):
 412         * rendering/style/StyleRareInheritedData.cpp:
 413         (WebCore::StyleRareInheritedData::StyleRareInheritedData):
 414
 415 2017-07-21  Timothy Horton  <timothy_horton@apple.com>
 416
 417         Remove an inaccurate comment from RenderTextLineBoxes
 418
 419         * rendering/RenderTextLineBoxes.cpp:
 420         (WebCore::RenderTextLineBoxes::absoluteRectsForRange):
 421         This comment has been wrong for at least 4 years.
 422
 423 2017-07-21  Timothy Horton  <timothy_horton@apple.com>
 424
 425         TextIndicator::estimatedTextColorsForRange asserts inside HashSet code (inserting reserved value)
 426         https://bugs.webkit.org/show_bug.cgi?id=174733
 427
 428         Reviewed by Wenson Hsieh.
 429
 430         * page/TextIndicator.cpp:
 431         (WebCore::estimatedTextColorsForRange):
 432         (WebCore::adjustTextIndicatorDataOptionsForEstimatedColorsIfNecessary):
 433         RGBA32 isn't a valid hash key, because we have no traits that define the
 434         empty or deleted values, nor do we have any bits we could feasibly
 435         use -- the full range of RGBA32 is easy to reach with various colors.
 436
 437         Instead, hash Color directly.
 438
 439 2017-07-21  Nan Wang  <n_wang@apple.com>
 440
 441         AX: Expose form validation on iOS as hint
 442         https://bugs.webkit.org/show_bug.cgi?id=174722
 443         <rdar://problem/33459761>
 444
 445         Reviewed by Chris Fleizach.
 446
 447         Adding the validation message to the hint of the form control element.
 448
 449         Test: accessibility/ios-simulator/form-control-validation-message.html
 450
 451         * accessibility/AccessibilityObject.cpp:
 452         (WebCore::AccessibilityObject::isShowingValidationMessage):
 453         (WebCore::AccessibilityObject::validationMessage):
 454         * accessibility/AccessibilityObject.h:
 455         * accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
 456         (-[WebAccessibilityObjectWrapper accessibilityHint]):
 457         (-[WebAccessibilityObjectWrapper accessibilityIsShowingValidationMessage]):
 458         * html/HTMLFormControlElement.cpp:
 459         (WebCore::HTMLFormControlElement::isShowingValidationMessage):
 460         * html/HTMLFormControlElement.h:
 461
 462 2017-07-21  Konstantin Tokarev  <annulen@yandex.ru>
 463
 464         [cmake][Mac] Unreviewed, fix linking WebKitSystemInterface
 465
 466         After r219560 link_directories() in PlatformXXX.cmake files does not
 467         affect corresponding targets.
 468
 469         * PlatformMac.cmake:
 470
 471 2017-07-21  Sam Weinig  <sam@webkit.org>
 472
 473         [WebIDL] Remove the need for the side map of DeprecatedCSSOMValue roots
 474         https://bugs.webkit.org/show_bug.cgi?id=174713
 475
 476         Reviewed by Andreas Kling.
 477
 478         We used to not want to bloat the size of CSSValues with a pointer to the
 479         owner CSSStyleDeclaration, so, for GC purposes we need kept a mapping of
 480         them on DOMWrapperWorld that only got set if a wrapper was made. Now that
 481         we have DeprecatedCSSOMValue, a wrapper for CSSValue that only gets created
 482         for the bindings, we no longer need the side table, and simply have the
 483         DeprecatedCSSOMValue maintain a reference to the CSSStyleDeclaration.
 484
 485         * bindings/js/DOMWrapperWorld.h:
 486         (WebCore::DOMWrapperWorld::wrappers):
 487         Remove m_deprecatedCSSOMValueRoots and make m_wrappers private with a getter.
 488
 489         * bindings/js/JSCSSStyleDeclarationCustom.cpp:
 490         (WebCore::JSCSSStyleDeclaration::getPropertyCSSValue): Deleted.
 491         Remove custom implementation of getPropertyCSSValue, now that it no longer
 492         needs to update m_deprecatedCSSOMValueRoots.
 493
 494         * bindings/js/JSDOMWindowBase.cpp:
 495         (WebCore::JSDOMWindowBase::fireFrameClearedWatchpointsForWindow):
 496         * bindings/js/JSDOMWrapperCache.h:
 497         (WebCore::getCachedWrapper):
 498         (WebCore::cacheWrapper):
 499         (WebCore::uncacheWrapper):
 500         * bindings/js/JSNodeCustom.cpp:
 501         (WebCore::getOutOfLineCachedWrapper):
 502         Use wrappers getter rather than direct member variable.
 503
 504         * bindings/js/JSDeprecatedCSSOMValueCustom.cpp:
 505         (WebCore::JSDeprecatedCSSOMValueOwner::isReachableFromOpaqueRoots):
 506         Rather than checking the side table for roots, now get the root
 507         via the reference CSSStyleDeclaration.
 508
 509         (WebCore::JSDeprecatedCSSOMValueOwner::finalize): Deleted.
 510         Remove custom finalize now that we don't have to remove ourself
 511         from m_deprecatedCSSOMValueRoots.
 512
 513         * css/CSSStyleDeclaration.idl:
 514         Remove [Custom] from getPropertyCSSValue.
 515
 516         * css/DeprecatedCSSOMValue.idl:
 517         Remove [JSCustomFinalize].
 518
 519         * css/CSSComputedStyleDeclaration.cpp:
 520         (WebCore::CSSComputedStyleDeclaration::getPropertyCSSValue):
 521         * css/CSSImageValue.cpp:
 522         (WebCore::CSSImageValue::createDeprecatedCSSOMWrapper):
 523         * css/CSSImageValue.h:
 524         * css/CSSPrimitiveValue.cpp:
 525         (WebCore::CSSPrimitiveValue::createDeprecatedCSSOMPrimitiveWrapper):
 526         * css/CSSPrimitiveValue.h:
 527         * css/CSSStyleDeclaration.h:
 528         * css/CSSValue.cpp:
 529         (WebCore::CSSValue::createDeprecatedCSSOMWrapper):
 530         * css/CSSValue.h:
 531         * css/DeprecatedCSSOMCounter.h:
 532         * css/DeprecatedCSSOMPrimitiveValue.cpp:
 533         (WebCore::DeprecatedCSSOMPrimitiveValue::getCounterValue):
 534         (WebCore::DeprecatedCSSOMPrimitiveValue::getRectValue):
 535         (WebCore::DeprecatedCSSOMPrimitiveValue::getRGBColorValue):
 536         * css/DeprecatedCSSOMPrimitiveValue.h:
 537         (WebCore::DeprecatedCSSOMPrimitiveValue::create):
 538         (WebCore::DeprecatedCSSOMPrimitiveValue::DeprecatedCSSOMPrimitiveValue):
 539         * css/DeprecatedCSSOMRGBColor.h:
 540         * css/DeprecatedCSSOMRect.h:
 541         * css/DeprecatedCSSOMValue.h:
 542         (WebCore::DeprecatedCSSOMValue::owner):
 543         (WebCore::DeprecatedCSSOMValue::DeprecatedCSSOMValue):
 544         (WebCore::DeprecatedCSSOMComplexValue::create):
 545         (WebCore::DeprecatedCSSOMComplexValue::DeprecatedCSSOMComplexValue):
 546         * css/DeprecatedCSSOMValueList.h:
 547         (WebCore::DeprecatedCSSOMValueList::create):
 548         (WebCore::DeprecatedCSSOMValueList::DeprecatedCSSOMValueList):
 549         * css/PropertySetCSSStyleDeclaration.cpp:
 550         (WebCore::PropertySetCSSStyleDeclaration::wrapForDeprecatedCSSOM):
 551         * css/PropertySetCSSStyleDeclaration.h:
 552         * svg/SVGElement.cpp:
 553         (WebCore::SVGElement::getPresentationAttribute):
 554         Pass owning CSSStyleDeclaration to DeprecatedCSSOMValue and derived classes.
 555
 556 2017-07-21  Andreas Kling  <akling@apple.com>
 557
 558         Use more references in event dispatch code
 559         https://bugs.webkit.org/show_bug.cgi?id=174681
 560
 561         Reviewed by Geoffrey Garen.
 562
 563         Make EventListener::handleEvent() take a ScriptExecutionContext&, Event&.
 564         Then make everything build again. Various null checks and assertions
 565         disappear in the process.
 566
 567         * bindings/js/JSErrorHandler.cpp:
 568         (WebCore::JSErrorHandler::handleEvent):
 569         * bindings/js/JSErrorHandler.h:
 570         * bindings/js/JSEventListener.cpp:
 571         (WebCore::JSEventListener::initializeJSFunction):
 572         (WebCore::JSEventListener::handleEvent):
 573         (WebCore::eventHandlerAttribute):
 574         * bindings/js/JSEventListener.h:
 575         (WebCore::JSEventListener::jsFunction):
 576         * bindings/js/JSLazyEventListener.cpp:
 577         (WebCore::JSLazyEventListener::initializeJSFunction):
 578         * bindings/js/JSLazyEventListener.h:
 579         * dom/EventListener.h:
 580         * dom/EventTarget.cpp:
 581         (WebCore::EventTarget::fireEventListeners):
 582         * html/HTMLPlugInElement.cpp:
 583         (WebCore::HTMLPlugInElement::defaultEventHandler):
 584         * html/ImageDocument.cpp:
 585         (WebCore::ImageEventListener::handleEvent):
 586         * inspector/CommandLineAPIHost.cpp:
 587         (WebCore::listenerEntriesFromListenerInfo):
 588         * inspector/InspectorDOMAgent.cpp:
 589         (WebCore::InspectorDOMAgent::buildObjectForEventListener):
 590         * inspector/InspectorIndexedDBAgent.cpp:
 591         * platform/Widget.h:
 592         (WebCore::Widget::handleEvent):
 593         * platform/cocoa/WebPlaybackSessionModelMediaElement.h:
 594         * platform/cocoa/WebPlaybackSessionModelMediaElement.mm:
 595         (WebCore::WebPlaybackSessionModelMediaElement::handleEvent):
 596         * platform/cocoa/WebVideoFullscreenModelVideoElement.h:
 597         * platform/cocoa/WebVideoFullscreenModelVideoElement.mm:
 598         (WebVideoFullscreenModelVideoElement::handleEvent):
 599         * svg/SVGTRefElement.cpp:
 600         (WebCore::SVGTRefTargetEventListener::handleEvent):
 601         * svg/animation/SVGSMILElement.cpp:
 602         (WebCore::ConditionEventListener::handleEvent):
 603         (WebCore::SVGSMILElement::handleConditionEvent):
 604         * svg/animation/SVGSMILElement.h:
 605
 606 2017-07-21  Zalan Bujtas  <zalan@apple.com>
 607
 608         iBooks: Overlapping/missing content at beginning/end of paragraph.
 609         https://bugs.webkit.org/show_bug.cgi?id=174717
 610         <rdar://problem/33117912>
 611
 612         Reviewed by Simon Fraser.
 613
 614         By definiton simple and normal line layout should always produce the exact same lineboxes. It enables us not
 615         to force repaint while swapping between these 2 line layouts.
 616         However in certain cases (font size pixel rounding as an example) they don't agree on how much content fits the line and
 617         that could result in missing/overlapping content due to the lack of repaint.
 618
 619         Unable to test.
 620
 621         * rendering/RenderBlockFlow.cpp:
 622         (WebCore::RenderBlockFlow::ensureLineBoxes):
 623
 624 2017-07-21  Chris Dumez  <cdumez@apple.com>
 625
 626         Drop make_dom_exceptions.pl script that we only have DOMException
 627         https://bugs.webkit.org/show_bug.cgi?id=174699
 628
 629         Reviewed by Darin Adler.
 630
 631         Drop make_dom_exceptions.pl script that we only have DOMException.
 632
 633         * CMakeLists.txt:
 634         * DerivedSources.make:
 635         * WebCore.xcodeproj/project.pbxproj:
 636         * WebCoreMacros.cmake:
 637         * bindings/js/JSDOMExceptionHandling.cpp:
 638         * css/FontFaceSet.cpp:
 639         * dom/DOMAllInOne.cpp:
 640         * dom/DOMExceptions.in: Removed.
 641         * dom/ExceptionCodeDescription.cpp: Added.
 642         (WebCore::ExceptionCodeDescription::ExceptionCodeDescription):
 643         * dom/ExceptionCodeDescription.h: Added.
 644         * dom/make_dom_exceptions.pl: Removed.
 645         * inspector/WebInjectedScriptHost.cpp:
 646         (WebCore::WebInjectedScriptHost::subtype):
 647
 648 2017-07-21  Said Abou-Hallawa  <sabouhallawa@apple.com>
 649
 650         [CG] An image should not invoke many system calls before confirming its format is supported
 651         https://bugs.webkit.org/show_bug.cgi?id=174692
 652
 653         Reviewed by Tim Horton.
 654
 655         We should be careful when invoking system calls before confirming that the
 656         image type is available and it is one of the whitelist formats. Otherwise
 657         we will be calling the parsers of the unsupported formats.
 658
 659         * loader/cache/CachedImage.cpp:
 660         (WebCore::CachedImage::setImageDataBuffer): The check isAllowedImageUTI()
 661         is now done in ImageDecoder::encodedDataStatus() which will return Error
 662         if there is an error in the data or "isAllowedImageUTI() returns false."
 663
 664         * platform/graphics/BitmapImage.cpp:
 665         (WebCore::BitmapImage::dataChanged): Avoid calling canUseAsyncDecodingForLargeImages()
 666         before confirming the image type is available and it's supported by WebKit.
 667         canUseAsyncDecodingForLargeImages() tries to cache the first frame of the
 668         image to know its size. Asking the ImageFrameCache to destroy its decoded
 669         frames is not needed unless ImageFrameCache::decodedSize() is not zero.
 670
 671         * platform/graphics/cg/ImageDecoderCG.cpp:
 672         (WebCore::ImageDecoder::encodedDataStatus): Avoid calling CGImageSourceGetStatus()
 673         before knowing the UTI of the image. When knowing it, we call CGImageSourceGetStatus()
 674         and if it returns kCGImageStatusIncomplete or kCGImageStatusComplete, we
 675         check whether isAllowedImageUTI() or not. If isAllowedImageUTI() returns
 676         false, return Error which will make the CachedImage cancel loading the
 677         rest of the image.
 678
 679 2017-07-21  Jeremy Jones  <jeremyj@apple.com>
 680
 681         AudioTrackPrivateMediaStreamCocoa shouldn't set AudioSession::setPreferredBufferSize
 682         https://bugs.webkit.org/show_bug.cgi?id=174707
 683         rdar://problem/33446809
 684
 685         Reviewed by Eric Carlson.
 686
 687         Manually tested for audio side effects.
 688
 689         AudioChannel::copyFrom fails when AudioChannel lengths don't match.
 690
 691         This happens because PlatformMediaSessionManager::updateSessionState() owns and sets
 692         AudioSession::setPreferredBufferSize().
 693
 694         However, AudioTrackPrivateMediaStreamCocoa::createAudioUnit, when it creates an input
 695         audio unit is setting AudioSession::setPreferredBufferSize() directly to its own arbitrary value.
 696
 697         AudioSession::setPreferredBufferSize() should be managed by the higher level
 698         PlatformMediaSessionManager, and not modified by audio unit creation, in order to keep harmony
 699         within the audio pipeline.
 700
 701         * platform/mediastream/mac/AudioTrackPrivateMediaStreamCocoa.cpp:
 702         (WebCore::AudioTrackPrivateMediaStreamCocoa::createAudioUnit):
 703
 704 2017-07-21  Jeremy Jones  <jeremyj@apple.com>
 705
 706         Noise when AudioChannel lengths don't match.
 707         https://bugs.webkit.org/show_bug.cgi?id=174706
 708         rdar://problem/33389856
 709
 710         Reviewed by Eric Carlson.
 711
 712         When AudioChannel lengths don't match, copyFrom() returns early leaving uninitialized data in the audio buffer.
 713         This change zeros out the data, so there isn't objectionable noise sent to the speaker.
 714
 715         * platform/audio/AudioChannel.cpp:
 716         (WebCore::AudioChannel::copyFrom):
 717
 718 2017-07-21  Sam Weinig  <sam@webkit.org>
 719
 720         [WebIDL] Make a few parameters non-nullable in inspector IDL files
 721         https://bugs.webkit.org/show_bug.cgi?id=174719
 722
 723         Reviewed by Joseph Pecoraro.
 724
 725         * inspector/CommandLineAPIHost.cpp:
 726         (WebCore::CommandLineAPIHost::databaseId):
 727         (WebCore::CommandLineAPIHost::storageId):
 728         * inspector/CommandLineAPIHost.h:
 729         * inspector/CommandLineAPIHost.idl:
 730         * inspector/InspectorDOMStorageAgent.cpp:
 731         (WebCore::InspectorDOMStorageAgent::storageId):
 732         * inspector/InspectorDOMStorageAgent.h:
 733         * inspector/InspectorDatabaseAgent.cpp:
 734         (WebCore::InspectorDatabaseAgent::databaseId):
 735         * inspector/InspectorDatabaseAgent.h:
 736         * inspector/InspectorFrontendHost.cpp:
 737         (WebCore::InspectorFrontendHost::showContextMenu):
 738         (WebCore::InspectorFrontendHost::dispatchEventAsContextMenuEvent):
 739         * inspector/InspectorFrontendHost.h:
 740         * inspector/InspectorFrontendHost.idl:
 741         Remove nullability from arguments that should never be null.
 742
 743 2017-07-21  Brady Eidson  <beidson@apple.com>
 744
 745         Get rid of WebCore IconDatabase code.
 746         https://bugs.webkit.org/show_bug.cgi?id=174700
 747
 748         Reviewed by Tim Horton.
 749
 750         No new tests (No behavior change).
 751
 752         Tons of red.
 753         Very little green.
 754         No further explanation.
 755
 756         * CMakeLists.txt:
 757         * WebCore.xcodeproj/project.pbxproj:
 758         * dom/Document.cpp:
 759         (WebCore::Document::implicitClose):
 760         * history/HistoryItem.cpp:
 761         (WebCore::HistoryItem::HistoryItem):
 762         (WebCore::HistoryItem::~HistoryItem):
 763         (WebCore::HistoryItem::reset):
 764         (WebCore::HistoryItem::setURLString):
 765         * inspector/InspectorNetworkAgent.cpp:
 766         (WebCore::InspectorNetworkAgent::willSendRequest):
 767         * loader/DocumentLoader.cpp:
 768         (WebCore::DocumentLoader::~DocumentLoader):
 769         (WebCore::DocumentLoader::startIconLoading):
 770         (WebCore::DocumentLoader::iconLoadDecisionAvailable): Deleted.
 771         (WebCore::iconLoadDecisionCallback): Deleted.
 772         (WebCore::DocumentLoader::getIconLoadDecisionForIconURL): Deleted.
 773         (WebCore::DocumentLoader::continueIconLoadWithDecision): Deleted.
 774         (WebCore::iconDataCallback): Deleted.
 775         (WebCore::DocumentLoader::getIconDataForIconURL): Deleted.
 776         * loader/DocumentLoader.h:
 777         (WebCore::DocumentLoader::linkIcons):
 778         * loader/FrameLoader.cpp:
 779         (WebCore::FrameLoader::FrameLoader):
 780         (WebCore::FrameLoader::stop):
 781         * loader/FrameLoader.h:
 782         (WebCore::FrameLoader::subframeLoader):
 783         (WebCore::FrameLoader::icon): Deleted.
 784         * loader/FrameLoaderClient.h:
 785         * loader/archive/cf/LegacyWebArchive.cpp:
 786         (WebCore::LegacyWebArchive::create):
 787         * loader/icon/IconController.cpp: Removed.
 788         * loader/icon/IconController.h: Removed.
 789         * loader/icon/IconDatabase.cpp: Removed.
 790         * loader/icon/IconDatabase.h: Removed.
 791         * loader/icon/IconDatabaseBase.cpp: Removed.
 792         * loader/icon/IconDatabaseBase.h: Removed.
 793         * loader/icon/IconDatabaseClient.h: Removed.
 794         * loader/icon/IconLoader.cpp:
 795         (WebCore::IconLoader::IconLoader):
 796         (WebCore::IconLoader::startLoading):
 797         (WebCore::IconLoader::notifyFinished):
 798         * loader/icon/IconLoader.h:
 799         * loader/icon/IconRecord.cpp: Removed.
 800         * loader/icon/IconRecord.h: Removed.
 801         * loader/icon/PageURLRecord.cpp: Removed.
 802         * loader/icon/PageURLRecord.h: Removed.
 803         * testing/Internals.cpp:
 804         (WebCore::Internals::shortcutIconURLs):
 805
 806 2017-07-21  Chris Dumez  <cdumez@apple.com>
 807
 808         Rename DOMCoreException class to DOMException
 809         https://bugs.webkit.org/show_bug.cgi?id=174698
 810
 811         Reviewed by Darin Adler.
 812
 813         Rename DOMCoreException class to DOMException to match its Web-exposed name.
 814
 815         * CMakeLists.txt:
 816         * DerivedSources.cpp:
 817         * DerivedSources.make:
 818         * Modules/quota/StorageErrorCallback.cpp:
 819         (WebCore::StorageErrorCallback::CallbackTask::CallbackTask):
 820         * Modules/quota/StorageErrorCallback.h:
 821         * Modules/quota/StorageErrorCallback.idl:
 822         * WebCore.xcodeproj/project.pbxproj:
 823         * bindings/js/JSDOMExceptionHandling.cpp:
 824         (WebCore::retrieveErrorMessage):
 825         (WebCore::createDOMException):
 826         * bindings/scripts/CodeGeneratorJS.pm:
 827         (GenerateOverloadDispatcher):
 828         * css/FontFaceSet.cpp:
 829         * css/FontFaceSet.h:
 830         * dom/DOMAllInOne.cpp:
 831         * dom/DOMException.cpp: Renamed from Source/WebCore/dom/DOMCoreException.cpp.
 832         (WebCore::DOMException::create):
 833         (WebCore::DOMException::DOMException):
 834         (WebCore::DOMException::initializeDescription):
 835         * dom/DOMException.h: Renamed from Source/WebCore/dom/DOMCoreException.h.
 836         (WebCore::DOMException::create):
 837         (WebCore::DOMException::DOMException):
 838         * dom/DOMException.idl: Renamed from Source/WebCore/dom/DOMCoreException.idl.
 839         * dom/DOMExceptions.in:
 840         * dom/make_dom_exceptions.pl:
 841         (generateImplementation):
 842
 843 2017-07-21  Sam Weinig  <sam@webkit.org>
 844
 845         Follow up to https://bugs.webkit.org/show_bug.cgi?id=174659.
 846
 847         * inspector/InspectorFrontendHost.idl:
 848         Remove unnecessary JSGenerateToJSObject noticed in review.
 849
 850 2017-07-20  Chris Dumez  <cdumez@apple.com>
 851
 852         Drop legacy SVGException type
 853         https://bugs.webkit.org/show_bug.cgi?id=174695
 854
 855         Reviewed by Darin Adler.
 856
 857         Drop legacy SVGException type and use DOMException instead, as per the latest
 858         SVG specification. Both Chrome and Firefox no longer expose the SVGException
 859         type.
 860
 861         * CMakeLists.txt:
 862         * DerivedSources.cpp:
 863         * DerivedSources.make:
 864         * WebCore.xcodeproj/project.pbxproj:
 865         * bindings/js/JSDOMExceptionHandling.cpp:
 866         (WebCore::createDOMException):
 867         * bindings/js/JSExceptionBase.cpp:
 868         (WebCore::toExceptionBase):
 869         * dom/DOMExceptions.in:
 870         * svg/SVGAllInOne.cpp:
 871         * svg/SVGCircleElement.cpp:
 872         * svg/SVGException.cpp: Removed.
 873         * svg/SVGException.h: Removed.
 874         * svg/SVGException.idl: Removed.
 875         * svg/SVGGraphicsElement.idl:
 876         * svg/SVGLocatable.cpp:
 877         (WebCore::SVGLocatable::getTransformToElement):
 878         * svg/SVGMatrixValue.h:
 879         * svg/SVGPathSegList.cpp:
 880         (WebCore::SVGPathSegList::getItem):
 881         (WebCore::SVGPathSegList::replaceItem):
 882         (WebCore::SVGPathSegList::removeItem):
 883         * svg/SVGPathSegList.h:
 884         * svg/SVGPathSegList.idl:
 885         * svg/properties/SVGAnimatedEnumerationPropertyTearOff.h:
 886         * svg/properties/SVGListProperty.h:
 887
 888 2017-07-20  Chris Dumez  <cdumez@apple.com>
 889
 890         Hook up ITP quirks to the needsSiteSpecificQuirks setting
 891         https://bugs.webkit.org/show_bug.cgi?id=174691
 892
 893         Reviewed by Darin Adler.
 894
 895         Hook up ITP quirks to the needsSiteSpecificQuirks setting to make it easier for
 896         Web-developers to test their fixes.
 897
 898         * loader/ResourceLoadObserver.cpp:
 899         (WebCore::shouldEnableSiteSpecificQuirks):
 900         (WebCore::areDomainsAssociated):
 901         (WebCore::ResourceLoadObserver::logFrameNavigation):
 902         (WebCore::resourceNeedsSSOQuirk):
 903         (WebCore::ResourceLoadObserver::logSubresourceLoading):
 904         (WebCore::ResourceLoadObserver::logWebSocketLoading):
 905
 906 2017-07-20  Matt Lewis  <jlewis3@apple.com>
 907
 908         Unreviewed, rolling out r219700.
 909
 910         This revision caused consistent timouts on iOS.
 911
 912         Reverted changeset:
 913
 914         "Turn tests at media/modern-media-controls/start-support back
 915         on"
 916         https://bugs.webkit.org/show_bug.cgi?id=174683
 917         http://trac.webkit.org/changeset/219700
 918
 919 2017-07-20  David Quesada  <david_quesada@apple.com>
 920
 921         Add SPI to notify WKNavigationDelegate about client redirects
 922         https://bugs.webkit.org/show_bug.cgi?id=174680
 923         rdar://problem/33184886
 924
 925         Reviewed by Brady Eidson.
 926
 927         * loader/FrameLoader.cpp:
 928         (WebCore::FrameLoader::performClientRedirect):
 929         * loader/FrameLoader.h:
 930         Add a convenience method for NavigationScheduler that handles a FrameLoadRequest
 931         as a client redirect. Currently this means loading the request and informing the
 932         client about it.
 933
 934         * loader/FrameLoaderClient.h:
 935         Add FrameLoaderClient::dispatchDidPerformClientRedirect() to inform the client when
 936         a client redirect occurs.
 937
 938         * loader/NavigationScheduler.cpp:
 939         Removed ScheduledURLNavigation::fire(). This class was never instantiated directly,
 940         and all subclasses override fire(), so this was unused code.
 941         For ScheduledRedirects and ScheduledLocationChange, use FrameLoader's new method to
 942         load the request as a client redirect.
 943
 944 2017-07-20  Chris Dumez  <cdumez@apple.com>
 945
 946         Drop legacy XPathException type
 947         https://bugs.webkit.org/show_bug.cgi?id=174679
 948
 949         Reviewed by Sam Weinig.
 950
 951         Drop legacy XPathException type and use DOMException instead. Both Firefox and Chrome
 952         no longer expose XPathException.
 953
 954         Test: fast/dom/DOMException/XPathException-obsolete.html
 955
 956         * CMakeLists.txt:
 957         * DerivedSources.cpp:
 958         * DerivedSources.make:
 959         * WebCore.xcodeproj/project.pbxproj:
 960         * bindings/js/JSDOMExceptionHandling.cpp:
 961         (WebCore::createDOMException):
 962         * bindings/js/JSExceptionBase.cpp:
 963         (WebCore::toExceptionBase):
 964         * dom/DOMExceptions.in:
 965         * xml/XPathException.cpp: Removed.
 966         * xml/XPathException.h: Removed.
 967         * xml/XPathException.idl: Removed.
 968         * xml/XPathExpression.cpp:
 969         (WebCore::XPathExpression::evaluate):
 970         * xml/XPathParser.cpp:
 971         (WebCore::XPath::Parser::parseStatement):
 972         * xml/XPathResult.cpp:
 973         (WebCore::XPathResult::convertTo):
 974         (WebCore::XPathResult::numberValue):
 975         (WebCore::XPathResult::stringValue):
 976         (WebCore::XPathResult::booleanValue):
 977         (WebCore::XPathResult::singleNodeValue):
 978         (WebCore::XPathResult::snapshotLength):
 979         (WebCore::XPathResult::iterateNext):
 980         (WebCore::XPathResult::snapshotItem):
 981
 982 2017-07-20  Chris Dumez  <cdumez@apple.com>
 983
 984         Unreviewed, rolling out r219706.
 985
 986         Broke iOS build
 987
 988         Reverted changeset:
 989
 990         "Drop legacy XPathException type"
 991         https://bugs.webkit.org/show_bug.cgi?id=174679
 992         http://trac.webkit.org/changeset/219706
 993
 994 2017-07-20  Chris Dumez  <cdumez@apple.com>
 995
 996         Drop legacy XPathException type
 997         https://bugs.webkit.org/show_bug.cgi?id=174679
 998
 999         Reviewed by Sam Weinig.
1000
1001         Drop legacy XPathException type and use DOMException instead. Both Firefox and Chrome
1002         no longer expose XPathException.
1003
1004         Test: fast/dom/DOMException/XPathException-obsolete.html
1005
1006         * CMakeLists.txt:
1007         * DerivedSources.cpp:
1008         * DerivedSources.make:
1009         * WebCore.xcodeproj/project.pbxproj:
1010         * bindings/js/JSDOMExceptionHandling.cpp:
1011         (WebCore::createDOMException):
1012         * bindings/js/JSExceptionBase.cpp:
1013         (WebCore::toExceptionBase):
1014         * dom/DOMExceptions.in:
1015         * xml/XPathException.cpp: Removed.
1016         * xml/XPathException.h: Removed.
1017         * xml/XPathException.idl: Removed.
1018         * xml/XPathExpression.cpp:
1019         (WebCore::XPathExpression::evaluate):
1020         * xml/XPathParser.cpp:
1021         (WebCore::XPath::Parser::parseStatement):
1022         * xml/XPathResult.cpp:
1023         (WebCore::XPathResult::convertTo):
1024         (WebCore::XPathResult::numberValue):
1025         (WebCore::XPathResult::stringValue):
1026         (WebCore::XPathResult::booleanValue):
1027         (WebCore::XPathResult::singleNodeValue):
1028         (WebCore::XPathResult::snapshotLength):
1029         (WebCore::XPathResult::iterateNext):
1030         (WebCore::XPathResult::snapshotItem):
1031
1032 2017-07-20  Chris Dumez  <cdumez@apple.com>
1033
1034         Regression(ITP): May get frequently logged out of wsj.com
1035         https://bugs.webkit.org/show_bug.cgi?id=174661
1036         <rdar://problem/32343256>
1037
1038         Reviewed by Geoffrey Garen.
1039
1040         Add the concept of associated domains in the ResourceLoadObserver. We
1041         previously ignore loads to and from the same domains. We now do the same
1042         if the to and from domains are associated (i.e. owned by the same entity).
1043
1044         For now, only add domains owned by Dow Jones & Company, Inc. to the list,
1045         to address login issues on wsj.com.
1046
1047         No new tests, verified manually on wsj.com.
1048
1049         * loader/ResourceLoadObserver.cpp:
1050         (WebCore::areDomainsAssociated):
1051         (WebCore::ResourceLoadObserver::logFrameNavigation):
1052         (WebCore::ResourceLoadObserver::logSubresourceLoading):
1053         (WebCore::ResourceLoadObserver::logWebSocketLoading):
1054
1055 2017-07-20  Chris Dumez  <cdumez@apple.com>
1056
1057         Replace calls to Vector::resize() with calls to more efficient shrink() / grow() when applicable
1058         https://bugs.webkit.org/show_bug.cgi?id=174660
1059
1060         Reviewed by Geoffrey Garen.
1061
1062         Replace calls to Vector::resize() with calls to more efficient shrink() / grow() when applicable.
1063         This essentially replaces a branch to figure out if the new size is less or greater than the
1064         current size by an assertion.
1065
1066         * Modules/gamepad/Gamepad.cpp:
1067         (WebCore::Gamepad::Gamepad):
1068         * Modules/webaudio/AudioContext.cpp:
1069         (WebCore::AudioContext::addReaction):
1070         * Modules/websockets/WebSocketChannel.cpp:
1071         (WebCore::WebSocketChannel::skipBuffer):
1072         * Modules/websockets/WebSocketDeflater.cpp:
1073         (WebCore::WebSocketDeflater::finish):
1074         * contentextensions/ContentExtensionCompiler.cpp:
1075         (WebCore::ContentExtensions::serializeSelector):
1076         * contentextensions/DFABytecodeCompiler.cpp:
1077         (WebCore::ContentExtensions::append):
1078         * crypto/gcrypt/CryptoAlgorithmAES_CBCGCrypt.cpp:
1079         (WebCore::gcryptEncrypt):
1080         (WebCore::gcryptDecrypt):
1081         * crypto/gcrypt/CryptoAlgorithmECDHGCrypt.cpp:
1082         (WebCore::gcryptDerive):
1083         * platform/gamepad/cocoa/GameControllerGamepadProvider.mm:
1084         (WebCore::GameControllerGamepadProvider::controllerDidConnect):
1085         * platform/gamepad/mac/HIDGamepadProvider.cpp:
1086         (WebCore::HIDGamepadProvider::deviceAdded):
1087         * platform/graphics/ImageBackingStore.h:
1088         (WebCore::ImageBackingStore::setSize):
1089         * platform/graphics/WOFFFileFormat.cpp:
1090         * platform/graphics/avfoundation/InbandMetadataTextTrackPrivateAVF.cpp:
1091         (WebCore::InbandMetadataTextTrackPrivateAVF::updatePendingCueEndTimes):
1092         (WebCore::InbandMetadataTextTrackPrivateAVF::flushPartialCues):
1093         * platform/graphics/avfoundation/InbandTextTrackPrivateAVF.cpp:
1094         (WebCore::InbandTextTrackPrivateAVF::resetCueValues):
1095         (WebCore::InbandTextTrackPrivateAVF::readNativeSampleBuffer):
1096         * platform/graphics/avfoundation/cf/InbandTextTrackPrivateAVCF.cpp:
1097         (WebCore::InbandTextTrackPrivateAVCF::readNativeSampleBuffer):
1098         * platform/graphics/cg/ImageBufferCG.cpp:
1099         (WebCore::cfData):
1100         * platform/image-decoders/bmp/BMPImageDecoder.cpp:
1101         (WebCore::BMPImageDecoder::frameBufferAtIndex):
1102         * platform/image-decoders/ico/ICOImageDecoder.cpp:
1103         (WebCore::ICOImageDecoder::decode):
1104         * platform/image-decoders/jpeg/JPEGImageDecoder.cpp:
1105         (WebCore::JPEGImageDecoder::frameBufferAtIndex):
1106         * platform/image-decoders/png/PNGImageDecoder.cpp:
1107         (WebCore::PNGImageDecoder::frameBufferAtIndex):
1108         (WebCore::PNGImageDecoder::readChunks):
1109         * platform/image-decoders/webp/WEBPImageDecoder.cpp:
1110         (WebCore::WEBPImageDecoder::frameBufferAtIndex):
1111         * platform/image-encoders/JPEGImageEncoder.cpp:
1112         (WebCore::compressRGBABigEndianToJPEG):
1113         * platform/text/DecodeEscapeSequences.h:
1114         (WebCore::URLEscapeSequence::decodeRun):
1115         * platform/text/SuffixTree.h:
1116         (WebCore::SuffixTree::Node::Node):
1117         * rendering/Grid.cpp:
1118         (WebCore::Grid::setNeedsItemsPlacement):
1119         * rendering/RenderTable.cpp:
1120         (WebCore::RenderTable::invalidateCachedColumns):
1121
1122 2017-07-20  Chris Dumez  <cdumez@apple.com>
1123
1124         Regression(ITP): Can no longer log in on abc.go.com
1125         https://bugs.webkit.org/show_bug.cgi?id=174533
1126         <rdar://problem/33325881>
1127
1128         Reviewed by Geoffrey Garen.
1129
1130         Add quirk for sp.auth.adobe.com which is used for SSO by web sites such as
1131         abc.go.com. This would otherwise cause adobe.com to be identified as a
1132         tracker and log in on abc.go.com would break.
1133
1134         No new tests, tested manually on abc.go.com.
1135
1136         * loader/ResourceLoadObserver.cpp:
1137         (WebCore::resourceNeedsSSOQuirk):
1138         (WebCore::ResourceLoadObserver::logSubresourceLoading):
1139
1140 2017-07-20  Antoine Quint  <graouts@apple.com>
1141
1142         Turn tests at media/modern-media-controls/start-support back on
1143         https://bugs.webkit.org/show_bug.cgi?id=174683
1144
1145         Reviewed by Dean Jackson.
1146
1147         Turning those tests back on revealed a small bug that is unlikely to really affect content
1148         on the Web. In the case where the size of the video is known right away, without the need
1149         for loading its metadata, as is the case in the start-support-click-to-start.html test with
1150         a local media resource, all queued layouts are flushed at once and we may call the layout()
1151         method of the left ButtonsContainer which originally is set to contain the play/pause button,
1152         which would remove the play/pause button from the center of the media. So before we potentially
1153         set the play/pause button as the central button, we first assign the default button set for
1154         the two ButtonsContainer instances and only add the play/pause button when we're not showing
1155         the prominent play/pause button.
1156
1157         * Modules/modern-media-controls/controls/inline-media-controls.js:
1158         (InlineMediaControls.prototype.layout):
1159         (InlineMediaControls.prototype._leftContainerButtons):
1160
1161 2017-07-20  Chris Dumez  <cdumez@apple.com>
1162
1163         Drop legacy FileException type
1164         https://bugs.webkit.org/show_bug.cgi?id=174676
1165
1166         Reviewed by Alex Christensen.
1167
1168         Drop legacy FileException type and use DOMException instead, as per the latest
1169         File API specification:
1170         - https://w3c.github.io/FileAPI/#ErrorAndException
1171
1172         Both Firefox and Chrome no longer expose FileException.
1173
1174         * CMakeLists.txt:
1175         * DerivedSources.cpp:
1176         * DerivedSources.make:
1177         * WebCore.xcodeproj/project.pbxproj:
1178         * bindings/js/JSDOMExceptionHandling.cpp:
1179         (WebCore::createDOMException):
1180         * dom/DOMExceptions.in:
1181         * fileapi/FileException.cpp: Removed.
1182         * fileapi/FileException.h: Removed.
1183         * fileapi/FileException.idl: Removed.
1184         * fileapi/FileReaderLoader.cpp:
1185         (WebCore::FileReaderLoader::FileReaderLoader):
1186         (WebCore::FileReaderLoader::failed):
1187         * fileapi/FileReaderLoader.h:
1188         (WebCore::FileReaderLoader::errorCode):
1189         * fileapi/FileReaderSync.cpp:
1190         (WebCore::errorCodeToExceptionCode):
1191         (WebCore::FileReaderSync::startLoading):
1192
1193 2017-07-20  Sam Weinig  <sam@webkit.org>
1194
1195         [WebIDL] Remove custom bindings for InspectorFrontendHost
1196         https://bugs.webkit.org/show_bug.cgi?id=174659
1197
1198         Reviewed by Chris Dumez.
1199
1200         * CMakeLists.txt:
1201         * WebCore.xcodeproj/project.pbxproj:
1202         * bindings/js/JSBindingsAllInOne.cpp:
1203         * bindings/js/JSInspectorFrontendHostCustom.cpp: Removed.
1204         Remove JSInspectorFrontendHostCustom.cpp.
1205
1206         * inspector/InspectorFrontendHost.cpp:
1207         (WebCore::populateContextMenu):
1208         (WebCore::InspectorFrontendHost::showContextMenu):
1209         * inspector/InspectorFrontendHost.h:
1210         * inspector/InspectorFrontendHost.idl:
1211         Re-implement showContextMenu by using an IDL defined ContextMenuItem dictionary
1212         and converting from that to a WebCore::ContextMenu in the implementation.
1213
1214 2017-07-20  Charlie Turner  <cturner@igalia.com>
1215
1216         [GTK] Layout Test http/tests/media/hls/hls-progress.html is failing
1217         https://bugs.webkit.org/show_bug.cgi?id=141469
1218
1219         Reviewed by Xabier Rodriguez-Calvar.
1220
1221         A work-around for getting progress information in a timely manner from
1222         live pipelines. We cannot rely on getting BUFFERING messages within
1223         the stalled time window (3s), so we have to poll for progress
1224         information somehow.
1225
1226         Makes the following pass reliably
1227          http/tests/media/hls/hls-audio-tracks-has-audio.html passed unexpectedly
1228          http/tests/media/hls/hls-audio-tracks.html passed unexpectedly
1229          http/tests/media/hls/hls-progress.html passed unexpectedly
1230
1231         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
1232         (WebCore::findHLSQueue):
1233         (WebCore::isHLSProgressing):
1234         (WebCore::MediaPlayerPrivateGStreamer::fillTimerFired):
1235
1236 2017-07-20  Andreas Kling  <akling@apple.com>
1237
1238         Minor Node->Element type tightening fixes
1239         https://bugs.webkit.org/show_bug.cgi?id=174651
1240
1241         Reviewed by Sam Weinig.
1242
1243         Tighten up some Node* into Element* or even better, Element&.
1244
1245         * dom/Document.cpp:
1246         (WebCore::acceptsEditingFocus): Take a const Element&.
1247         (WebCore::Document::setFocusedElement):
1248         * editing/Editor.cpp:
1249         (WebCore::Editor::findEventTargetFrom): Return an Element*.
1250         (WebCore::Editor::findEventTargetFromSelection): Ditto.
1251         (WebCore::Editor::pasteAsPlainText):
1252         (WebCore::Editor::pasteAsFragment):
1253         (WebCore::Editor::dispatchCPPEvent):
1254         * editing/Editor.h:
1255         * page/DragController.cpp:
1256         (WebCore::DragController::dispatchTextInputEventFor):
1257
1258 2017-07-20  Brady Eidson  <beidson@apple.com>
1259
1260         WKHTTPCookieStore API tests fail on High Sierra.
1261         <rdar://problem/33410271> and https://bugs.webkit.org/show_bug.cgi?id=174666
1262
1263         Reviewed by Andy Estes.
1264
1265         Covered by existing API tests.
1266
1267         In r219567 I'd moved cookie storage observation off of NSHTTPCookieStorage and NSNotificationCenter
1268         to CFHTTPCookieStorage observation callbacks.
1269
1270         This is because notifications were only sent for the default [NSHTTPCookieStorage sharedHTTPCookieStorage]
1271         and not any of the other ones we keep in flight.
1272
1273         Unfortunately that SPI has been disabled in High Sierra.
1274         Fortunately we found a way we can get non-shared NSHTTPCookieStorages to send notifications that works everywhere.
1275
1276         * platform/network/cocoa/CookieStorageObserver.h:
1277         * platform/network/cocoa/CookieStorageObserver.mm:
1278         (-[WebCookieObserverAdapter initWithObserver:]):
1279         (-[WebCookieObserverAdapter cookiesChangedNotificationHandler:]):
1280         (WebCore::CookieStorageObserver::create):
1281         (WebCore::CookieStorageObserver::CookieStorageObserver):
1282         (WebCore::CookieStorageObserver::~CookieStorageObserver):
1283         (WebCore::CookieStorageObserver::startObserving): Use a trick to call some SPI on non-shared NSHTTPCookieStorages
1284           to get them to send notifications.
1285         (WebCore::CookieStorageObserver::stopObserving):
1286         (WebCore::cookiesChanged): Deleted.
1287
1288         * platform/network/cocoa/NetworkStorageSessionCocoa.mm:
1289         (WebCore::NetworkStorageSession::cookieStorageObserver):
1290
1291 2017-07-20  Miguel Gomez  <magomez@igalia.com>
1292
1293         [GStreamer] Some layout tests issue "g_mutex_clear() called on uninitialised or locked mutex" and flaky crash in ~MediaPlayerPrivateGStreamerBase
1294         https://bugs.webkit.org/show_bug.cgi?id=173952
1295
1296         Reviewed by Carlos Garcia Campos.
1297
1298         Adjust MediaPlayerPrivateGStreamerBase to avoid concurrence problems with the GStreamer thread when
1299         destroying the object.
1300
1301         Covered by existent tests.
1302
1303         * platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.cpp:
1304         (WebCore::MediaPlayerPrivateGStreamerBase::~MediaPlayerPrivateGStreamerBase):
1305         (WebCore::MediaPlayerPrivateGStreamerBase::repaint):
1306         (WebCore::MediaPlayerPrivateGStreamerBase::cancelRepaint):
1307
1308 2017-07-19  Zan Dobersek  <zdobersek@igalia.com>
1309
1310         [EME] Push CDMInstance, CDMPrivate and associated types into the Platform layer
1311         https://bugs.webkit.org/show_bug.cgi?id=174496
1312
1313         Reviewed by Xabier Rodriguez-Calvar.
1314
1315         Move the CDMInstance and CDMPrivate interfaces into the platform layer, enabling
1316         implementations of specific key systems at that level, as well as integration with
1317         the MediaPlayerPrivate stack without any layering violations.
1318
1319         While the platform layer still uses the WebCore namespace, the two interface files
1320         have now been moved under the platform/encryptedmedia/ directory and out of the
1321         Modules/encryptedmedia/ directory where the Web-facing API implementation resides.
1322
1323         The two interfaces integrated closely with the Web-facing enumerations and
1324         dictionaries: MediaKeyMessageType, MediaKeySessionType, MediaKeyStatus,
1325         MediaKeySystemConfiguration, MediaKeySystemCapability, MediaKeysRequirement
1326         and MediaKeysRestrictions. Definitions of these types are also moved to the
1327         platform layer without any changes (apart from renaming), and the previous
1328         types now alias against the types now located in the platform layer.
1329
1330         No new tests -- no change in behavior.
1331
1332         * CMakeLists.txt:
1333         * Modules/encryptedmedia/CDM.cpp:
1334         * Modules/encryptedmedia/CDM.h:
1335         * Modules/encryptedmedia/MediaKeyMessageType.h:
1336         * Modules/encryptedmedia/MediaKeySessionType.h:
1337         * Modules/encryptedmedia/MediaKeyStatus.h:
1338         * Modules/encryptedmedia/MediaKeySystemAccess.cpp:
1339         * Modules/encryptedmedia/MediaKeySystemAccess.h:
1340         * Modules/encryptedmedia/MediaKeySystemConfiguration.h:
1341         * Modules/encryptedmedia/MediaKeySystemMediaCapability.h:
1342         * Modules/encryptedmedia/MediaKeysRequirement.h:
1343         * Modules/encryptedmedia/MediaKeysRestrictions.h:
1344         * platform/encryptedmedia/CDMInstance.h: Renamed from Source/WebCore/Modules/encryptedmedia/CDMInstance.h.
1345         * platform/encryptedmedia/CDMKeyStatus.h: Copied from Source/WebCore/Modules/encryptedmedia/MediaKeyStatus.h.
1346         * platform/encryptedmedia/CDMKeySystemConfiguration.h: Copied from Source/WebCore/Modules/encryptedmedia/MediaKeySystemConfiguration.h.
1347         * platform/encryptedmedia/CDMMediaCapability.h: Copied from Source/WebCore/Modules/encryptedmedia/MediaKeySystemMediaCapability.h.
1348         * platform/encryptedmedia/CDMMessageType.h: Copied from Source/WebCore/Modules/encryptedmedia/MediaKeyMessageType.h.
1349         * platform/encryptedmedia/CDMPrivate.h: Renamed from Source/WebCore/Modules/encryptedmedia/CDMPrivate.h.
1350         * platform/encryptedmedia/CDMRequirement.h: Copied from Source/WebCore/Modules/encryptedmedia/MediaKeysRequirement.h.
1351         * platform/encryptedmedia/CDMRestrictions.h: Copied from Source/WebCore/Modules/encryptedmedia/MediaKeysRestrictions.h.
1352         * platform/encryptedmedia/CDMSessionType.h: Copied from Source/WebCore/Modules/encryptedmedia/MediaKeySessionType.h.
1353
1354 2017-07-19  Zan Dobersek  <zdobersek@igalia.com>
1355
1356         Unreviewed follow-up to r219674.
1357
1358         * page/scrolling/coordinatedgraphics/ScrollingCoordinatorCoordinatedGraphics.cpp:
1359         Remove the AsyncScrollingCoordinator.h header include that was added by accident.
1360
1361 2017-07-19  Zan Dobersek  <zdobersek@igalia.com>
1362
1363         [CMake] Clean up Web Crypto build targets
1364         https://bugs.webkit.org/show_bug.cgi?id=174253
1365
1366         Reviewed by Alex Christensen.
1367
1368         Gather the common WebCrypto source files in CMakeLists.txt, including them
1369         in the build unconditionally and instead relying on ENABLE_SUBTLE_CRYPTO
1370         build guards to exclude the code from compilation if the feature is disabled.
1371
1372         PlatformGTK.cmake, PlatformMac.cmake and PlatformWPE.cmake can then remove
1373         duplicated build targets. PlatformMac.cmake still lists all the Mac-specific
1374         Web Crypto build targets.
1375
1376         PlatformGTK.cmake and PlatformWPE.cmake now include GCrypt.cmake if the
1377         USE_GCRYPT variable is enabled. Both ports at the moment enable that variable
1378         though as they by default leverage a libgcrypt-based CrpytoDigest implementation
1379         in the PAL library.
1380
1381         The new GCrypt.cmake file adds the libgcrypt-specific Web Crypto build targets to
1382         the build and also sets up libgcrypt include directiories and libraries.
1383
1384         No new tests -- no change in behavior.
1385
1386         * CMakeLists.txt:
1387         * PlatformGTK.cmake:
1388         * PlatformMac.cmake:
1389         * PlatformWPE.cmake:
1390         * platform/GCrypt.cmake: Added.
1391
1392 2017-07-19  Sam Weinig  <sam@webkit.org>
1393
1394         [WebIDL] Remove custom bindings for CommandLineAPIHost
1395         https://bugs.webkit.org/show_bug.cgi?id=174642
1396
1397         Reviewed by Joseph Pecoraro.
1398
1399         * CMakeLists.txt:
1400         * WebCore.xcodeproj/project.pbxproj:
1401         * bindings/js/JSBindingsAllInOne.cpp:
1402         * bindings/js/JSCommandLineAPIHostCustom.cpp: Removed.
1403         Remove JSCommandLineAPIHostCustom.cpp.
1404
1405         * bindings/js/JSDOMConvertObject.h:
1406         (WebCore::JSConverter<IDLObject>::convert):
1407         Add missing JSConverter specialization for IDLObject.
1408
1409         * inspector/CommandLineAPIHost.cpp:
1410         (WebCore::CommandLineAPIHost::inspect):
1411         (WebCore::listenerEntriesFromListenerInfo):
1412         (WebCore::CommandLineAPIHost::inspectedObject):
1413         (WebCore::CommandLineAPIHost::databaseId):
1414         (WebCore::CommandLineAPIHost::storageId):
1415         (WebCore::CommandLineAPIHost::inspectImpl): Deleted.
1416         (WebCore::CommandLineAPIHost::getEventListenersImpl): Deleted.
1417         (WebCore::CommandLineAPIHost::databaseIdImpl): Deleted.
1418         (WebCore::CommandLineAPIHost::storageIdImpl): Deleted.
1419         * inspector/CommandLineAPIHost.h:
1420         * inspector/CommandLineAPIHost.idl:
1421         Replace custom bindings with modern bindings affordances: dictionaries, records and sequences.
1422
1423 2017-07-19  Zan Dobersek  <zdobersek@igalia.com>
1424
1425         Unreviewed CoordGraphics build fix.
1426
1427         * page/scrolling/coordinatedgraphics/ScrollingCoordinatorCoordinatedGraphics.cpp:
1428         (WebCore::ScrollingCoordinatorCoordinatedGraphics::updateNodeViewportConstraints):
1429         This should be defined for the ScrollingCoordinatorCoordinatedGraphics class.
1430         Also fix the setFixedToViewport() call, fishing out the GraphicsLayer object from
1431         the ScrollingStateNode's LayerRepresentation.
1432         (WebCore::AsyncScrollingCoordinator::updateNodeViewportConstraints): Deleted.
1433
1434 2017-07-19  Chris Dumez  <cdumez@apple.com>
1435
1436         Drop SQLException type
1437         https://bugs.webkit.org/show_bug.cgi?id=174665
1438
1439         Reviewed by Sam Weinig.
1440
1441         Drop SQLException type. We never throw such exception. Also, Firefox
1442         and Chrome no longer expose this type.
1443
1444         * CMakeLists.txt:
1445         * DerivedSources.cpp:
1446         * DerivedSources.make:
1447         * Modules/webdatabase/SQLException.cpp: Removed.
1448         * Modules/webdatabase/SQLException.h: Removed.
1449         * Modules/webdatabase/SQLException.idl: Removed.
1450         * WebCore.xcodeproj/project.pbxproj:
1451         * bindings/js/JSDOMExceptionHandling.cpp:
1452         (WebCore::createDOMException):
1453         * bindings/js/JSExceptionBase.cpp:
1454         (WebCore::toExceptionBase):
1455         * dom/DOMExceptions.in:
1456
1457 2017-07-19  Wenson Hsieh  <wenson_hsieh@apple.com>
1458
1459         Unreviewed, tweak a comment above TemporarySelectionOptionIgnoreSelectionChanges in Editor.cpp.
1460
1461         The comment references the UI process, but it describes an entity in WebCore. This should have been generalized
1462         to instead refer to the client layer in general (i.e. WebKit1 or WebKit2).
1463
1464         * editing/Editor.h:
1465
1466 2017-07-19  Simon Fraser  <simon.fraser@apple.com>
1467
1468         getBoundingClientRects not updated for programmatic scrolls
1469         https://bugs.webkit.org/show_bug.cgi?id=174538
1470         rdar://problem/33049012
1471
1472         Reviewed by Tim Horton.
1473
1474         Baidu.com has two event handlers on its <input>, and both query the input location with getBoundingClientRect()
1475         and the current pageYOffset (via jQuery), then try to scroll the input to the top of the screen. The bug is that
1476         programmatic scrolls did not immediately update the layout viewport rect, so the second call to
1477         getBoundingClientRect() would return stale coordinates, triggering an extra scroll.
1478
1479         To fix this, undo the fix for r219320 which tried to keep getBoundingClientRect() current during unstable scroll
1480         updates by adding a shadow layout viewport rect. Instead, almost always update the layout viewport rect on
1481         FrameView, even during unstable visible rect updates, but not if content insets are being changed interactively,
1482         since changing viewport heights cause problems with bottom-fixed elements. Also, we need to compute a new layout
1483         viewport rect in FrameView::updateLayoutViewport() for programmatic scrolls.
1484
1485         However, always updating the layout viewport triggered issues with the scrolling tree. The scrolling state tree
1486         fossilizes layer positions relative to a specific viewport rect, and that relationship has to be maintained.
1487         There are code paths that recompute fixed/sticky viewport constraints when the layout viewport has changed but
1488         we haven't done layout or recomputed layer positions (e.g. updating viewport-constrained layers via
1489         updateScrollCoordinatedLayersAfterFlush()) and in these cases using a new layout viewport for those computations
1490         results in an inconsistent scrolling tree.
1491
1492         Fix this by not updating scrolling constraints every time we have to re-register scrolling nodes.
1493         updateScrollCoordinatedLayersAfterFlush() only needs to update the layer on the scrolling node (to handle
1494         tiled/non-tiled switches), so make updateScrollCoordinatedLayer() a little more fine-grained, and only update
1495         constraints when we've just computed layer geometry. This allows for different scrolling nodes to have
1496         constraints computed at different times, with different layout viewports, which happens.
1497
1498         Two additional fixes were required to make bottom-fixed bars behave correctly.
1499
1500         First, FrameView::computeLayoutViewportOrigin() had a bug where rounding of half-pixel values would cause it to
1501         fall into the if (visualViewport.height() > layoutViewport.height()) clause, but then fail to clamp for
1502         rubber-banding.
1503
1504         Second, the FrameView::unscaledMaximumScrollPosition() was wrong after zooming on iOS, since it uses visibleSize()
1505         which is affected by page scale on iOS only (and the function wants scale-independent values). Fix with a hack that
1506         should be cleaned up via webkit.org/b/174648.
1507
1508         Tested by existing tests.
1509
1510         * page/FrameView.cpp:
1511         (WebCore::FrameView::computeUpdatedLayoutViewportRect):
1512         (WebCore::FrameView::computeLayoutViewportOrigin):
1513         (WebCore::FrameView::setLayoutViewportOverrideRect):
1514         (WebCore::FrameView::updateLayoutViewport):
1515         (WebCore::FrameView::unscaledMaximumScrollPosition):
1516         (WebCore::FrameView::documentToClientOffset):
1517         (WebCore::FrameView::setUnstableLayoutViewportRect): Deleted.
1518         * page/FrameView.h:
1519         * page/scrolling/AsyncScrollingCoordinator.cpp:
1520         (WebCore::AsyncScrollingCoordinator::updateScrollPositionAfterAsyncScroll):
1521         (WebCore::AsyncScrollingCoordinator::reconcileScrollingState):
1522         (WebCore::AsyncScrollingCoordinator::reconcileViewportConstrainedLayerPositions):
1523         (WebCore::AsyncScrollingCoordinator::updateNodeLayer):
1524         (WebCore::AsyncScrollingCoordinator::updateNodeViewportConstraints):
1525         (WebCore::AsyncScrollingCoordinator::updateViewportConstrainedNode): Deleted.
1526         * page/scrolling/AsyncScrollingCoordinator.h:
1527         * page/scrolling/ScrollingCoordinator.cpp:
1528         (WebCore::operator<<):
1529         * page/scrolling/ScrollingCoordinator.h:
1530         (WebCore::ScrollingCoordinator::reconcileScrollingState):
1531         (WebCore::ScrollingCoordinator::updateNodeLayer):
1532         (WebCore::ScrollingCoordinator::updateNodeViewportConstraints):
1533         (WebCore::ScrollingCoordinator::updateViewportConstrainedNode): Deleted.
1534         * page/scrolling/ScrollingStateFixedNode.cpp:
1535         (WebCore::ScrollingStateFixedNode::updateConstraints):
1536         * page/scrolling/ScrollingStateStickyNode.cpp:
1537         (WebCore::ScrollingStateStickyNode::updateConstraints):
1538         (WebCore::ScrollingStateStickyNode::reconcileLayerPositionForViewportRect):
1539         * page/scrolling/ScrollingTree.cpp:
1540         (WebCore::ScrollingTree::commitTreeState):
1541         * page/scrolling/mac/ScrollingTreeFixedNode.mm:
1542         (WebCore::ScrollingTreeFixedNode::updateLayersAfterAncestorChange):
1543         * page/scrolling/mac/ScrollingTreeStickyNode.mm:
1544         (WebCore::ScrollingTreeStickyNode::updateLayersAfterAncestorChange):
1545         * rendering/RenderLayerBacking.cpp:
1546         (WebCore::RenderLayerBacking::updateGeometry):
1547         * rendering/RenderLayerCompositor.cpp:
1548         (WebCore::RenderLayerCompositor::updateScrollCoordinatedLayersAfterFlush):
1549         (WebCore::RenderLayerCompositor::updateBacking):
1550         (WebCore::RenderLayerCompositor::fixedRootBackgroundLayerChanged):
1551         (WebCore::RenderLayerCompositor::requiresCompositingForPosition):
1552         (WebCore::RenderLayerCompositor::updateScrollCoordinatedStatus):
1553         (WebCore::RenderLayerCompositor::computeFixedViewportConstraints):
1554         (WebCore::RenderLayerCompositor::computeStickyViewportConstraints):
1555         (WebCore::RenderLayerCompositor::updateScrollCoordinatedLayer):
1556         (WebCore::RenderLayerCompositor::didAddScrollingLayer):
1557         * rendering/RenderLayerCompositor.h:
1558
1559 2017-07-19  Megan Gardner  <megan_gardner@apple.com>
1560
1561         Don't write file URLs to iOS Pasteboard
1562         https://bugs.webkit.org/show_bug.cgi?id=174647
1563         <rdar://problem/33199730>
1564
1565         Reviewed by Wenson Hsieh.
1566
1567         Tests updated to reflect the changes. We are no longer vending file URLs in Drag & Drop and Copy/Paste.
1568
1569         * editing/ios/EditorIOS.mm:
1570         (WebCore::Editor::writeImageToPasteboard):
1571         * platform/ios/PlatformPasteboardIOS.mm:
1572         (WebCore::PlatformPasteboard::write):
1573
1574 2017-07-19  Myles C. Maxfield  <mmaxfield@apple.com>
1575
1576         Setting the minimum font size preference doesn’t affect absolute line-height values, so lines overlap
1577         https://bugs.webkit.org/show_bug.cgi?id=174406
1578
1579         Reviewed by Simon Fraser.
1580         <rdar://problem/10139227>
1581
1582         Reviewed by NOBODY.
1583
1584         When the minimumFontSize API preference is set, we will increase the font size without increasing
1585         the line height. If the content specifies line-height as an absolute value, there can be two
1586         unfortunate results:
1587
1588         - Adjacent lines in a paragraph can overlap
1589         - If the paragraph (or containin block or whatever) uses overflow: hidden, the first and last lines
1590         can be cut off and potentially indecipherable.
1591
1592         Instead, we should use the minimum font size preference as a signal that we should increase the
1593         line-height as well as the font-size. Eventually, we will want to increase it by an amount
1594         proportional to the increase in font-size (which can be due to minimumFontSize, minimumLogicalFontSize,
1595         text autosizing, etc.). However, because minimumLogicalFontSize is on by default, this would cause
1596         a behavior change on many webpages which use small text, so such a change would be too risky right now.
1597         Instead, we can pretend that minimumFontSize is the only cause that text increases, and use this as the
1598         only signal to boost the corresponding line-height.
1599
1600         Tests: fast/text/line-height-minimumFontSize-text-zoom.html
1601                fast/text/line-height-minimumFontSize-visual.html
1602                fast/text/line-height-minimumFontSize-zoom.html
1603                fast/text/line-height-minimumFontSize.html
1604                fast/text/line-height-minimumFontSize-autosize.html
1605
1606         * css/StyleBuilderCustom.h:
1607         (WebCore::computeBaseSpecifiedFontSize):
1608         (WebCore::computeLineHeightMultiplierDueToFontSize):
1609         (WebCore::StyleBuilderCustom::applyValueLineHeight):
1610         (WebCore::StyleBuilderCustom::applyValueFill):
1611         (WebCore::StyleBuilderCustom::applyValueStroke):
1612         (WebCore::StyleBuilderCustom::applyValueContent):
1613         * rendering/TextAutoSizing.cpp:
1614
1615 2017-07-18  Myles C. Maxfield  <mmaxfield@apple.com>
1616
1617         Setting the minimum font size preference doesn’t affect absolute line-height values, so lines overlap
1618         https://bugs.webkit.org/show_bug.cgi?id=174406
1619         <rdar://problem/10139227>
1620
1621         Reviewed by Simon Fraser.
1622
1623         When the minimumFontSize API preference is set, we will increase the font size without increasing
1624         the line height. If the content specifies line-height as an absolute value, there can be two
1625         unfortunate results:
1626
1627         - Adjacent lines in a paragraph can overlap
1628         - If the paragraph (or containin block or whatever) uses overflow: hidden, the first and last lines
1629         can be cut off and potentially indecipherable.
1630
1631         Instead, we should use the minimum font size preference as a signal that we should increase the
1632         line-height as well as the font-size. Eventually, we will want to increase it by an amount
1633         proportional to the increase in font-size (which can be due to minimumFontSize, minimumLogicalFontSize,
1634         text autosizing, etc.). However, because minimumLogicalFontSize is on by default, this would cause
1635         a behavior change on many webpages which use small text, so such a change would be too risky right now.
1636         Instead, we can pretend that minimumFontSize is the only cause that text increases, and use this as the
1637         only signal to boost the corresponding line-height.
1638
1639         Tests: fast/text/line-height-minimumFontSize-text-zoom.html
1640                fast/text/line-height-minimumFontSize-visual.html
1641                fast/text/line-height-minimumFontSize-zoom.html
1642                fast/text/line-height-minimumFontSize.html
1643                fast/text/line-height-minimumFontSize-autosize.html
1644
1645         * css/StyleBuilderCustom.h:
1646         (WebCore::computeBaseSpecifiedFontSize):
1647         (WebCore::computeLineHeightMultiplierDueToFontSize):
1648         (WebCore::StyleBuilderCustom::applyValueLineHeight):
1649         (WebCore::StyleBuilderCustom::applyValueFill):
1650         (WebCore::StyleBuilderCustom::applyValueStroke):
1651         (WebCore::StyleBuilderCustom::applyValueContent):
1652         * rendering/TextAutoSizing.cpp:
1653
1654 2017-07-19  Chris Dumez  <cdumez@apple.com>
1655
1656         DOMException should not have its own toString()
1657         https://bugs.webkit.org/show_bug.cgi?id=174630
1658
1659         Reviewed by Sam Weinig.
1660
1661         DOMException should not have its own toString() and should use the one from Error.prototype.toString()
1662         instead:
1663         - https://heycam.github.io/webidl/#idl-DOMException
1664
1665         No new tests, rebaselined existing tests.
1666
1667         * dom/DOMCoreException.idl:
1668
1669 2017-07-19  Antoine Quint  <graouts@apple.com>
1670
1671         [iOS] REGRESSION: Scrubbing media using built-in controls does not pause media
1672         https://bugs.webkit.org/show_bug.cgi?id=174650
1673         <rdar://problem/33401877>
1674
1675         Reviewed by Dean Jackson.
1676
1677         We would only pause when scrubbing on macOS because we only listened to "mousedown" events on the
1678         scrubber's backing <input> element to identify that the user had started interacting with the
1679         scrubber.
1680
1681         Implementing the same technique on iOS required a little more work than just listening to "touchstart"
1682         events on the same element. On top of that, we needed to make sure that we would only respond to
1683         "touchstart" events on the slider's thumb, and not on the track, since only on macOS should the user
1684         be able to click anywhere on the track to scrub. So we turn off pointer-events for the <input> on iOS
1685         only, and turn them back on specifically for the thumb.
1686
1687         There is also some finessing when dealing with touch events where we need to track the identifier of
1688         the touch that started the user interaction. So we keep track of it in an ivar and wait until we get
1689         a "touchend" event where the changedTouches list contains a touch with that same identifier to ensure
1690         the same touch that initiates and ends the scrubbing interaction.
1691
1692         Finally, we fix another issue that was uncovered while turning back on the ScrubbingSupport tests
1693         where we would not trash the cached _value ivar when we initiated scrubbing, which was important since
1694         we would mistakenly use the pre-srubbing value during a scrub.
1695
1696         * Modules/modern-media-controls/controls/slider.css:
1697         (.ios .slider > input):
1698         (.slider > input::-webkit-slider-thumb):
1699         * Modules/modern-media-controls/controls/slider.js:
1700         (Slider.prototype.handleEvent):
1701         (Slider.prototype._handleMousedownEvent):
1702         (Slider.prototype._interactionEndTarget):
1703         (Slider.prototype._handleTouchstartEvent):
1704         (Slider.prototype._valueWillStartChanging):
1705         (Slider.prototype._valueDidStopChanging):
1706         (Slider.prototype._handleMouseupEvent):
1707         (Slider.prototype._handleTouchendEvent):
1708
1709 2017-07-19  Nan Wang  <n_wang@apple.com>
1710
1711         AX: Web page reloaded when a node is labelling multiple childnodes
1712         https://bugs.webkit.org/show_bug.cgi?id=174655
1713
1714         Reviewed by Chris Fleizach.
1715
1716         When we are asking for the aria-labelledby attribute of a node and its
1717         sibling is also labelled by the same node, we get into an infinite loop
1718         in textUnderElement since we only ignore one child. Added checks for
1719         siblings to avoid such loop.
1720
1721         Test: accessibility/mac/aria-labelledby-multiple-child-crash.html
1722
1723         * accessibility/AccessibilityNodeObject.cpp:
1724         (WebCore::AccessibilityNodeObject::textUnderElement):
1725
1726 2017-07-19  Andy Estes  <aestes@apple.com>
1727
1728         Use a cast to work around clang's false -Wobjc-literal-conversion warnings
1729         https://bugs.webkit.org/show_bug.cgi?id=174631
1730
1731         Reviewed by Dan Bernstein.
1732
1733         Instead of ignoring -Wobjc-literal-conversion, use a cast to work around rdar://problem/33383354.
1734
1735         * platform/ios/PlatformPasteboardIOS.mm:
1736         (WebCore::PlatformPasteboard::write):
1737
1738 2017-07-19  Chris Dumez  <cdumez@apple.com>
1739
1740         Make cross-origin properties enumerable
1741         https://bugs.webkit.org/show_bug.cgi?id=174576
1742
1743         Reviewed by Darin Adler.
1744
1745         Makes cross-origin properties enumerable on Window and Location objects
1746         as per:
1747         - https://github.com/whatwg/html/pull/2777
1748
1749         This simplifies our code quite a bit.
1750
1751         No new tests, updated existing tests.
1752
1753         * bindings/js/JSDOMWindowCustom.cpp:
1754         (WebCore::jsDOMWindowGetOwnPropertySlotRestrictedAccess):
1755         (WebCore::JSDOMWindow::getOwnPropertySlotByIndex):
1756         (WebCore::JSDOMWindow::getOwnPropertyNames):
1757         * bindings/js/JSLocationCustom.cpp:
1758         (WebCore::getOwnPropertySlotCommon):
1759         (WebCore::JSLocation::getOwnPropertyNames):
1760         * bindings/scripts/CodeGeneratorJS.pm:
1761         (GenerateHeader):
1762
1763 2017-07-18  Carlos Alberto Lopez Perez  <clopez@igalia.com>
1764
1765         [EME] Build failure with Clang-3.8 on InitDataRegistry.cpp
1766         https://bugs.webkit.org/show_bug.cgi?id=174628
1767
1768         Reviewed by Žan Doberšek.
1769
1770         * Modules/encryptedmedia/InitDataRegistry.cpp:
1771         (WebCore::extractKeyIDsKeyids): Add an explicit cast to unsigned.
1772
1773 2017-07-19  Matt Lewis  <jlewis3@apple.com>
1774
1775         Unreviewed, rolling out r219646.
1776
1777         The test added are failing on all platforms
1778
1779         Reverted changeset:
1780
1781         "Setting the minimum font size preference doesn’t affect
1782         absolute line-height values, so lines overlap"
1783         https://bugs.webkit.org/show_bug.cgi?id=174406
1784         http://trac.webkit.org/changeset/219646
1785
1786 2017-07-19  Yusuke Suzuki  <utatane.tea@gmail.com>
1787
1788         [WTF] Implement WTF::ThreadGroup
1789         https://bugs.webkit.org/show_bug.cgi?id=174081
1790
1791         Reviewed by Mark Lam.
1792
1793         * page/ResourceUsageThread.h:
1794
1795 2017-07-18  Andy Estes  <aestes@apple.com>
1796
1797         [Xcode] Enable CLANG_WARN_RANGE_LOOP_ANALYSIS
1798         https://bugs.webkit.org/show_bug.cgi?id=174631
1799
1800         Reviewed by Tim Horton.
1801
1802         * Configurations/Base.xcconfig:
1803         * rendering/Grid.cpp:
1804         (WebCore::Grid::insert):
1805         * rendering/RenderTreeAsText.cpp:
1806         (WebCore::write):
1807         * rendering/SimpleLineLayoutCoverage.cpp:
1808         (WebCore::SimpleLineLayout::printSimpleLineLayoutCoverage):
1809
1810 2017-07-18  Yusuke Suzuki  <utatane.tea@gmail.com>
1811
1812         WTF::Thread should have the threads stack bounds.
1813         https://bugs.webkit.org/show_bug.cgi?id=173975
1814
1815         Reviewed by Mark Lam.
1816
1817         When creating WebThread, we first allocate WebCore::ThreadGlobalData in UI thread
1818         and share it with WebThread.
1819         The problem is that WebCore::ThreadGlobalData has CachedResourceRequestInitiators.
1820         It allocates AtomicString, which requires WTFThreadData.
1821
1822         In this patch, we call WTF::initializeThreading() before allocating WebCore::ThreadGlobalData.
1823         And we also call AtomicString::init() before calling WebCore::ThreadGlobalData since
1824         WebCore::ThreadGlobalData allocates AtomicString.
1825
1826         * platform/ios/wak/WebCoreThread.mm:
1827         (StartWebThread):
1828
1829 2017-07-18  Myles C. Maxfield  <mmaxfield@apple.com>
1830
1831         Setting the minimum font size preference doesn’t affect absolute line-height values, so lines overlap
1832         https://bugs.webkit.org/show_bug.cgi?id=174406
1833         <rdar://problem/10139227>
1834
1835         Reviewed by Simon Fraser.
1836
1837         When the minimumFontSize API preference is set, we will increase the font size without increasing
1838         the line height. If the content specifies line-height as an absolute value, there can be two
1839         unfortunate results:
1840
1841         - Adjacent lines in a paragraph can overlap
1842         - If the paragraph (or containin block or whatever) uses overflow: hidden, the first and last lines
1843         can be cut off and potentially indecipherable.
1844
1845         Instead, we should use the minimum font size preference as a signal that we should increase the
1846         line-height as well as the font-size. Eventually, we will want to increase it by an amount
1847         proportional to the increase in font-size (which can be due to minimumFontSize, minimumLogicalFontSize,
1848         text autosizing, etc.). However, because minimumLogicalFontSize is on by default, this would cause
1849         a behavior change on many webpages which use small text, so such a change would be too risky right now.
1850         Instead, we can pretend that minimumFontSize is the only cause that text increases, and use this as the
1851         only signal to boost the corresponding line-height.
1852
1853         Tests: fast/text/line-height-minimumFontSize-text-zoom.html
1854                fast/text/line-height-minimumFontSize-visual.html
1855                fast/text/line-height-minimumFontSize-zoom.html
1856                fast/text/line-height-minimumFontSize.html
1857                fast/text/line-height-minimumFontSize-autosize.html
1858
1859         * css/StyleBuilderCustom.h:
1860         (WebCore::computeBaseSpecifiedFontSize):
1861         (WebCore::computeLineHeightMultiplierDueToFontSize):
1862         (WebCore::StyleBuilderCustom::applyValueLineHeight):
1863         (WebCore::StyleBuilderCustom::applyValueFill):
1864         (WebCore::StyleBuilderCustom::applyValueStroke):
1865         (WebCore::StyleBuilderCustom::applyValueContent):
1866         * rendering/TextAutoSizing.cpp:
1867
1868 2017-07-18  Zalan Bujtas  <zalan@apple.com>
1869
1870         Media controls are missing content in fullscreen when document has scroll offset.
1871         https://bugs.webkit.org/show_bug.cgi?id=174644
1872         <rdar://problem/32415323>
1873
1874         Reviewed by Simon Fraser.
1875
1876         If a non-user initiated scrolling (result of resize for example) is processed asynchronously, it might
1877         leapfrog other, programatic scrollings and trigger unintentional scroll offsets (and turn into unwanted clippings).
1878         This patch ensures that both resize and top content inset change are translated into programatic scrolling.
1879
1880         Unable to test full screen video.
1881
1882         * page/FrameView.cpp:
1883         (WebCore::FrameView::setFrameRect):
1884         (WebCore::FrameView::topContentInsetDidChange):
1885
1886 2017-07-18  Andy Estes  <aestes@apple.com>
1887
1888         [Xcode] Enable CLANG_WARN_OBJC_LITERAL_CONVERSION
1889         https://bugs.webkit.org/show_bug.cgi?id=174631
1890
1891         Reviewed by Sam Weinig.
1892
1893         * Configurations/Base.xcconfig:
1894         * platform/ios/PlatformPasteboardIOS.mm:
1895         (WebCore::PlatformPasteboard::write):
1896
1897 2017-07-18  Ali Juma  <ajuma@chromium.org>
1898
1899         Align quirky number parsing with other browsers
1900         https://bugs.webkit.org/show_bug.cgi?id=155874
1901
1902         Reviewed by Simon Fraser.
1903
1904         Parse unit-less non-zero values for animation-duration and transition-duration
1905         as invalid time values in quirks mode.
1906
1907         Test: fast/css/animation-transition-duration-quirksmode.html
1908
1909         * css/parser/CSSPropertyParser.cpp:
1910         (WebCore::consumeAnimationValue):
1911
1912 2017-07-18  Chris Dumez  <cdumez@apple.com>
1913
1914         HysteresisActivity cannot be used in the UIProcess
1915         https://bugs.webkit.org/show_bug.cgi?id=174643
1916         <rdar://problem/33086442>
1917
1918         Reviewed by Tim Horton.
1919
1920         Port HysteresisActivity to RunLoop::Timer so that it can safely be used in
1921         the UIProcess as well.
1922
1923         * platform/HysteresisActivity.h:
1924
1925 2017-07-18  Joseph Pecoraro  <pecoraro@apple.com>
1926
1927         Web Inspector: Modernize InjectedScriptSource
1928         https://bugs.webkit.org/show_bug.cgi?id=173890
1929
1930         Reviewed by Brian Burg.
1931
1932         Covered by existing tests.
1933
1934         * inspector/CommandLineAPIModuleSource.js:
1935         (CommandLineAPIImpl.prototype.copy):
1936         (CommandLineAPIImpl.prototype._inspect):
1937         Use RemoteObject, a new parameter.
1938
1939 2017-07-18  Ryosuke Niwa  <rniwa@webkit.org>
1940
1941         REGRESSION(r218910): Crash when password field changes to text field
1942         https://bugs.webkit.org/show_bug.cgi?id=174560
1943
1944         Reviewed by Zalan Bujtas.
1945
1946         The crash was caused by textMarkerDataForFirstPositionInTextControl accessing a nullptr returned by getOrCreate.
1947         Unfortunately, in order to this fix bug while preserving the exact behavior would require synchronously creating
1948         a renderer for the editing host when the input type changed since we can't create an accessbility object out of
1949         a renderer-less node.
1950
1951         Instead, revert back to pre-r218910 behavior of always using the text control element's axID when notifying
1952         the value change. While this is inconsistent with the way editing commands report content changes, I've since
1953         learned that VoiceOver has code to deal with this exact situation.
1954
1955         Test: accessibility/mac/input-type-change-crash-2.html
1956
1957         * accessibility/AXObjectCache.cpp:
1958         (WebCore::AXObjectCache::textMarkerDataForFirstPositionInTextControl):
1959
1960 2017-07-18  Matt Baker  <mattbaker@apple.com>
1961
1962         Web Inspector: Refactoring: replace InspectorCanvasAgent::CanvasEntry with a helper class
1963         https://bugs.webkit.org/show_bug.cgi?id=174311
1964
1965         Reviewed by Devin Rousso.
1966
1967         This patch adds a helper class for tracking canvases and their data. The
1968         current approach of defining a helper structure in the agent's header is
1969         awkward to use and maintain, and won't scale well as more canvas instrumentation
1970         points and data are added.
1971
1972         * CMakeLists.txt:
1973         * WebCore.xcodeproj/project.pbxproj:
1974         New file.
1975
1976         * inspector/InspectorCanvas.cpp: Added.
1977         (WebCore::InspectorCanvas::create):
1978         (WebCore::InspectorCanvas::InspectorCanvas):
1979         (WebCore::InspectorCanvas::buildObjectForCanvas):
1980         * inspector/InspectorCanvas.h: Added.
1981
1982         * inspector/InspectorCanvasAgent.cpp:
1983         (WebCore::InspectorCanvasAgent::InspectorCanvasAgent):
1984         (WebCore::InspectorCanvasAgent::enable):
1985         (WebCore::InspectorCanvasAgent::requestNode):
1986         (WebCore::InspectorCanvasAgent::requestContent):
1987         (WebCore::InspectorCanvasAgent::requestCSSCanvasClientNodes):
1988         (WebCore::InspectorCanvasAgent::resolveCanvasContext):
1989         (WebCore::InspectorCanvasAgent::frameNavigated):
1990         (WebCore::InspectorCanvasAgent::didCreateCSSCanvas):
1991         (WebCore::InspectorCanvasAgent::didChangeCSSCanvasClientNodes):
1992         (WebCore::InspectorCanvasAgent::didCreateCanvasRenderingContext):
1993         (WebCore::InspectorCanvasAgent::didChangeCanvasMemory):
1994         (WebCore::InspectorCanvasAgent::canvasDestroyed):
1995         (WebCore::InspectorCanvasAgent::canvasDestroyedTimerFired):
1996         (WebCore::InspectorCanvasAgent::clearCanvasData):
1997         (WebCore::InspectorCanvasAgent::unbindCanvas):
1998         (WebCore::InspectorCanvasAgent::assertInspectorCanvas):
1999         (WebCore::InspectorCanvasAgent::findInspectorCanvas):
2000         (WebCore::InspectorCanvasAgent::getCanvasEntry): Deleted.
2001         (WebCore::InspectorCanvasAgent::buildObjectForCanvas): Deleted.
2002         * inspector/InspectorCanvasAgent.h:
2003
2004         * inspector/InspectorController.cpp:
2005         (WebCore::InspectorController::InspectorController):
2006         Remove unused constructor argument.
2007
2008 2017-07-18  Basuke Suzuki  <Basuke.Suzuki@sony.com>
2009
2010         [Curl] Move detail implementation from ResourceHandle to ResourceHandleInternal
2011         https://bugs.webkit.org/show_bug.cgi?id=174641
2012
2013         After moving stuff from ResourceHandleManager to ResourceHandle on
2014         https://bugs.webkit.org/show_bug.cgi?id=173630, still there're many violation
2015         remains between ResourceHandle and ResourceHandleInternal classes.
2016         Many of implementation detail should be move to ResourceHandleInternal
2017         to improve build speed.
2018
2019         Reviewed by Alex Christensen.
2020
2021         * platform/network/ResourceHandle.h:
2022         * platform/network/ResourceHandleInternal.h:
2023         (WebCore::ResourceHandleInternal::ResourceHandleInternal):
2024         * platform/network/curl/ResourceHandleCurl.cpp:
2025         (WebCore::ResourceHandleInternal::~ResourceHandleInternal):
2026         (WebCore::ResourceHandle::start):
2027         (WebCore::ResourceHandle::cancel):
2028         (WebCore::ResourceHandleInternal::initialize):
2029         (WebCore::ResourceHandleInternal::applyAuthentication):
2030         (WebCore::ResourceHandleInternal::setupPUT):
2031         (WebCore::ResourceHandleInternal::setupPOST):
2032         (WebCore::ResourceHandleInternal::setupFormData):
2033         (WebCore::ResourceHandle::platformSetDefersLoading):
2034         (WebCore::ResourceHandleInternal::didFinish):
2035         (WebCore::ResourceHandleInternal::didFail):
2036         (WebCore::ResourceHandleInternal::calculateWebTimingInformations):
2037         (WebCore::ResourceHandleInternal::handleLocalReceiveResponse):
2038         (WebCore::ResourceHandleInternal::willPrepareSendData):
2039         (WebCore::ResourceHandleInternal::didReceiveHeaderLine):
2040         (WebCore::ResourceHandleInternal::didReceiveAllHeaders):
2041         (WebCore::ResourceHandleInternal::didReceiveContentData):
2042         (WebCore::ResourceHandleInternal::readCallback):
2043         (WebCore::ResourceHandleInternal::headerCallback):
2044         (WebCore::ResourceHandleInternal::writeCallback):
2045         (WebCore::ResourceHandle::platformLoadResourceSynchronously):
2046         (WebCore::ResourceHandleInternal::dispatchSynchronousJob):
2047         (WebCore::ResourceHandleInternal::handleDataURL):
2048         (WebCore::ResourceHandle::initialize): Deleted.
2049         (WebCore::ResourceHandle::applyAuthentication): Deleted.
2050         (WebCore::ResourceHandle::setupPUT): Deleted.
2051         (WebCore::ResourceHandle::setupPOST): Deleted.
2052         (WebCore::ResourceHandle::setupFormData): Deleted.
2053         (WebCore::ResourceHandle::didFinish): Deleted.
2054         (WebCore::ResourceHandle::didFail): Deleted.
2055         (WebCore::ResourceHandle::calculateWebTimingInformations): Deleted.
2056         (WebCore::ResourceHandle::handleLocalReceiveResponse): Deleted.
2057         (WebCore::ResourceHandle::willPrepareSendData): Deleted.
2058         (WebCore::ResourceHandle::didReceiveHeaderLine): Deleted.
2059         (WebCore::ResourceHandle::didReceiveAllHeaders): Deleted.
2060         (WebCore::ResourceHandle::didReceiveContentData): Deleted.
2061         (WebCore::ResourceHandle::readCallback): Deleted.
2062         (WebCore::ResourceHandle::headerCallback): Deleted.
2063         (WebCore::ResourceHandle::writeCallback): Deleted.
2064         (WebCore::ResourceHandle::dispatchSynchronousJob): Deleted.
2065         (WebCore::ResourceHandle::handleDataURL): Deleted.
2066
2067 2017-07-18  Andy Estes  <aestes@apple.com>
2068
2069         [Xcode] Enable CLANG_WARN_NON_LITERAL_NULL_CONVERSION
2070         https://bugs.webkit.org/show_bug.cgi?id=174631
2071
2072         Reviewed by Dan Bernstein.
2073
2074         * Configurations/Base.xcconfig:
2075
2076 2017-07-18  Antoine Quint  <graouts@apple.com>
2077
2078         [macOS] Mouse pointer does not hide during fullscreen playback
2079         https://bugs.webkit.org/show_bug.cgi?id=174638
2080         <rdar://problem/33244399>
2081
2082         Reviewed by Dean Jackson.
2083
2084         Test: media/modern-media-controls/css/webkit-cursor-visibility-auto-hide.html
2085
2086         The user-agent stylesheet sets the "-webkit-cursor-visibility" to "auto-hide" for fullscreen <video>
2087         elements. Since we reset the page styles, including UA styles, on .media-controls-container, we need
2088         to explicitly inherit this style property from the page to ensure the mouse pointer automatically
2089         hides in fullscreen.
2090
2091         * Modules/modern-media-controls/controls/media-controls.css:
2092         (.media-controls-container):
2093
2094 2017-07-18  Sam Weinig  <sam@webkit.org>
2095
2096         [WebIDL] Remove custom bindings for MockContentFilterSettings
2097         https://bugs.webkit.org/show_bug.cgi?id=174606
2098
2099         Reviewed by Andy Estes.
2100
2101         * PlatformMac.cmake:
2102         * WebCore.xcodeproj/project.pbxproj:
2103         * bindings/js/JSMockContentFilterSettingsCustom.cpp: Removed.
2104         Remove JSMockContentFilterSettingsCustom.
2105
2106         * testing/MockContentFilterSettings.h:
2107         * testing/MockContentFilterSettings.idl:
2108         Switch from constants to IDL enums to make the bindings and tests simpler.
2109
2110 2017-07-18  Sam Weinig  <sam@webkit.org>
2111
2112         [WebIDL] Replace some custom bindings code in JSCSSStyleDeclarationCustom.cpp with named getters/setters
2113         https://bugs.webkit.org/show_bug.cgi?id=174529
2114
2115         Reviewed by Chris Dumez.
2116
2117         * CMakeLists.txt:
2118         * WebCore.xcodeproj/project.pbxproj:
2119         * css/CSSAllInOne.cpp:
2120         Add CSSStyleDeclaration.cpp
2121
2122         * bindings/js/JSCSSStyleDeclarationCustom.cpp:
2123         Move getter / setter / getOwnProperties logic from here to CSSStyleDeclaration.cpp.
2124
2125         * bindings/scripts/CodeGeneratorJS.pm:
2126         (GenerateGetOwnPropertySlot):
2127         (GenerateGetOwnPropertySlotByIndex):
2128         (InstanceOverridesGetOwnPropertySlot):
2129         (GenerateHeader):
2130         Remove support for no longer needed [CustomGetOwnPropertySlotAndDescriptor].
2131         Add support for [PutOnlyForSupportedProperties], which allows named setters to behave
2132         a bit more like normal setters by allowing the implementation to note which properties
2133         are supported which in turn, allows the bindings to continue down the normal put path
2134         for unsupported properties. This is necessary to mimic the behavior of CSSStyleDeclaration
2135         which is using named setters in place hundreds of additional properties.
2136
2137         (InstanceOverridesDefineOwnProperty):
2138         Add [DefaultDefineOwnProperty] to allow disabling defineOwnProperty overriding that comes
2139         automatically with named setters. We need this for CSSStyleDeclaration since the named
2140         setter usage modeling real properties, not a dictionary style setter. We may want to
2141         revisit this, and add support for DefineOwnProperty here, but this allows us to maintain
2142         the status quo.
2143
2144         (GenerateDefineOwnProperty):
2145         Fix the case when you have a named setter, an indexed getter, but no indexed setter. In that
2146         case, we would get a compile error, due to the index variable being unused.
2147
2148         * bindings/scripts/IDLAttributes.json:
2149         Remove support for no longer needed [CustomGetOwnPropertySlotAndDescriptor], add
2150         [PutOnlyForSupportedProperties] and [DefaultDefineOwnProperty].
2151
2152         * css/CSSStyleDeclaration.h:
2153         * css/CSSStyleDeclaration.cpp: Added.
2154         (WebCore::CSSStyleDeclaration::namedItem):
2155         (WebCore::CSSStyleDeclaration::setNamedItem):
2156         (WebCore::CSSStyleDeclaration::isSupportedPropertyName):
2157         (WebCore::CSSStyleDeclaration::supportedPropertyNames):
2158         Use standard getter / setter functions to implement CSS property name getting and setting.
2159         Convert from Identifier to AtomicString, which are now bridgeable.
2160
2161         * css/CSSStyleDeclaration.idl:
2162         Add named getter and setter. Remove [CustomGetOwnPropertyNames], [CustomGetOwnPropertySlotAndDescriptor]
2163         and [CustomPut], add [DefaultDefineOwnProperty].
2164
2165         * bindings/scripts/test/JS/JSTestNamedSetterWithIndexedGetter.cpp: Added.
2166         * bindings/scripts/test/JS/JSTestNamedSetterWithIndexedGetter.h: Added.
2167         * bindings/scripts/test/JS/JSTestNamedSetterWithIndexedGetterAndSetter.cpp: Added.
2168         * bindings/scripts/test/JS/JSTestNamedSetterWithIndexedGetterAndSetter.h: Added.
2169         * bindings/scripts/test/TestNamedSetterWithIndexedGetter.idl: Added.
2170         * bindings/scripts/test/TestNamedSetterWithIndexedGetterAndSetter.idl: Added.
2171         Add tests for improved behavior when using named setters with variations of indexed getters and setters as well.
2172
2173 2017-07-18  Antoine Quint  <graouts@apple.com>
2174
2175         [iOS] WebKit media controls are sometimes shown after exiting full screen on vimeo.com
2176         https://bugs.webkit.org/show_bug.cgi?id=174627
2177         <rdar://problem/33301005>
2178
2179         Reviewed by Dean Jackson, provisionally reviewed by Jeremy Jones.
2180
2181         On iOS 11, both the WebKit media controls and the Vimeo custom controls would appear sometimes when exiting
2182         from fullscreen when the video was playing and the user would tap the X button, which would pause the video
2183         as well as exit fullscreen.
2184
2185         The reason this happens is that the ControlsVisibilitySupport object, which governs whether the WebKit media
2186         controls should be displayed for a given video, woud listen to "pause" and "webkitfullscreenchange" events
2187         and determine whether to show the WebKit media controls. We listen to the "pause" event because when media
2188         pauses, and the video has the "controls" attribute set, we should show the controls and suspend the controls
2189         auto-hide timer. And we're interested in knowing when we enter and exit fullscreen because we want to override
2190         the "controls" attribute not being set when we enter fullscreen.
2191
2192         However, on iOS 11, it appears that the "webkitfullscreenchange" event is not reliably fired as the user enters
2193         and exits fullscreen, which is tracked by webkit.org/b/174626. So, when the user exits fullscreen, we would be
2194         informed of the video being paused via a "pause" event, but not of the video exiting fullscreen. And because
2195         media events are asynchronous, the "pause" event would sometimes be fired before we exited fullscreen, and when
2196         the _updateControls() would run, we would sometimes determine that we are in fullscreen still and determine
2197         that the WebKit media controls should be shown.
2198
2199         Of course, on iOS, the WebKit media controls are not shown and instead we delegate to AVKit to display media controls.
2200         So we could simply disregard this whole logic in iOS. But we choose to instead use the "webkitpresentationmodechanged"
2201         when the presentation mode API is supported, as is the case on iOS 11, to determine changes of media fullscreen state.
2202         This way, should we ever choose to support fullscreen media controls provided by WebKit on iOS, this logic is already
2203         correct and we write less platform-specific code.
2204
2205         This, alas, cannot be tested since we can't force the X button to be tapped within the AVKit fullscreen controls.
2206
2207         * Modules/modern-media-controls/media/controls-visibility-support.js:
2208         (ControlsVisibilitySupport.prototype.get mediaEvents):
2209         * Modules/modern-media-controls/media/media-controller.js:
2210         (MediaController):
2211         * Modules/modern-media-controls/media/start-support.js:
2212         (StartSupport.prototype.get mediaEvents):
2213
2214 2017-07-18  Matt Lewis  <jlewis3@apple.com>
2215
2216         Unreviewed, rolling out r219610.
2217
2218         This caused an api failure on all platforms for the test
2219         SnapshotImageLargeAsyncDecoding
2220
2221         Reverted changeset:
2222
2223         "Async image decoding for large images should be disabled
2224         after the first time a tile is painted"
2225         https://bugs.webkit.org/show_bug.cgi?id=174451
2226         http://trac.webkit.org/changeset/219610
2227
2228 2017-07-18  Devin Rousso  <drousso@apple.com>
2229
2230         Add CanvasRenderingContext2D::getTransform
2231         https://bugs.webkit.org/show_bug.cgi?id=174278
2232
2233         Reviewed by Dean Jackson.
2234
2235         Tests: fast/canvas/2d.getTransform.modification.html
2236                fast/canvas/2d.getTransform.newobject.html
2237                fast/canvas/2d.setTransform.matrix.html
2238
2239         * css/DOMMatrixReadOnly.h:
2240         Make DOMMatrixReadOnly::validateAndFixup public so that values of DOMMatrixInit are still
2241         usable without having to construct a DOMMatrixReadOnly. This is beneficial in the case that
2242         an exception is thrown, as the validateAndFixup check can happen without any allocations.
2243
2244         * html/canvas/CanvasRenderingContext2D.idl:
2245         * html/canvas/CanvasRenderingContext2D.h:
2246         * html/canvas/CanvasRenderingContext2D.cpp:
2247         (WebCore::CanvasRenderingContext2D::getTransform):
2248         (WebCore::CanvasRenderingContext2D::setTransform):
2249
2250 2017-07-18  Andy Estes  <aestes@apple.com>
2251
2252         [Xcode] Enable CLANG_WARN_BLOCK_CAPTURE_AUTORELEASING
2253         https://bugs.webkit.org/show_bug.cgi?id=174631
2254
2255         Reviewed by Darin Adler.
2256
2257         * Configurations/Base.xcconfig:
2258
2259 2017-07-18  Jonathan Bedard  <jbedard@apple.com>
2260
2261         Unreviewed build fix after r219595
2262
2263         * platform/graphics/cocoa/FontDescriptionCocoa.cpp:
2264         (WebCore::makeNeverDestroyed): Remove duplicate makeNeverDestroyed definition.
2265         * platform/ios/DragImageIOS.mm:
2266         (WebCore::cascadeForSystemFont): Return correct FontCascade object.
2267         (WebCore::createDragImageForLink): Fix compiler error.
2268
2269 2017-07-18  Daniel Bates  <dabates@apple.com>
2270
2271         Cleanup: Remove AlternativeTextInfo and use Variant to represent alternative text info details
2272         https://bugs.webkit.org/show_bug.cgi?id=174604
2273
2274         I inadvertently did not commit some of my local changes.
2275
2276         * editing/AlternativeTextController.cpp:
2277         (WebCore::AlternativeTextController::startAlternativeTextUITimer):
2278         (WebCore::AlternativeTextController::stopAlternativeTextUITimer):
2279         (WebCore::AlternativeTextController::applyPendingCorrection):
2280         (WebCore::AlternativeTextController::hasPendingCorrection):
2281         (WebCore::AlternativeTextController::show):
2282         (WebCore::AlternativeTextController::handleCancelOperation):
2283         (WebCore::AlternativeTextController::dismiss):
2284         (WebCore::AlternativeTextController::dismissSoon):
2285         (WebCore::AlternativeTextController::applyAutocorrectionBeforeTypingIfAppropriate):
2286         (WebCore::AlternativeTextController::timerFired):
2287         (WebCore::AlternativeTextController::handleAlternativeTextUIResult):
2288         (WebCore::AlternativeTextController::respondToMarkerAtEndOfWord):
2289         (WebCore::AlternativeTextController::markerDescriptionForAppliedAlternativeText):
2290
2291 2017-07-17  Daniel Bates  <dabates@apple.com>
2292
2293         Cleanup: Remove AlternativeTextInfo and use Variant to represent alternative text info details
2294         https://bugs.webkit.org/show_bug.cgi?id=174604
2295
2296         Reviewed by Darin Adler.
2297
2298         Remove struct AlternativeTextInfo and represent the alternative text info as instance variables
2299         using Variant to hold either an autocorrection replacement string or the alternative dictation
2300         context.
2301
2302         Not functionality changed. So, no new tests.
2303
2304         * editing/AlternativeTextController.cpp: Remove header Event.h as it is ultimately included by TextEvent.h.
2305         (WebCore::AlternativeTextController::show): Update code as needed.
2306         (WebCore::AlternativeTextController::timerFired): Ditto.
2307         (WebCore::AlternativeTextController::alternativeTextClient): Substitute nullptr for 0.
2308         (WebCore::AlternativeTextController::respondToMarkerAtEndOfWord): Update code as needed.
2309         (WebCore::AutocorrectionAlternativeDetails::create): Deleted.
2310         (WebCore::AutocorrectionAlternativeDetails::replacementString): Deleted.
2311         (WebCore::AutocorrectionAlternativeDetails::AutocorrectionAlternativeDetails): Deleted.
2312         (WebCore::DictationAlternativeDetails::create): Deleted.
2313         (WebCore::DictationAlternativeDetails::dictationContext): Deleted.
2314         (WebCore::DictationAlternativeDetails::DictationAlternativeDetails): Deleted.
2315         * editing/AlternativeTextController.h: Remove header Range.h and forward declare Range. Include header
2316         Position.h and wtf/Variant.h. Sort forward declarations.
2317         (WebCore::AlternativeTextDetails::~AlternativeTextDetails): Deleted.
2318
2319 2017-07-18  Said Abou-Hallawa  <sabouhallawa@apple.com>
2320
2321         Async image decoding for large images should be disabled after the first time a tile is painted
2322         https://bugs.webkit.org/show_bug.cgi?id=174451
2323
2324         Reviewed by Simon Fraser.
2325
2326         Flashing because of DOM mutation can be fixed by disabling the asynchronous
2327         image decoding after the first time a tile was painted.
2328
2329         We can detect this by consulting the tile repaintCount. If it is zero, then
2330         it is safe to use asynchronous image decoded. If the tile repaintCount is
2331         greater than zero, we are not sure if the renderer rectangle has an image
2332         drawn in it already or not. In this case we have to use the synchronous
2333         image decoding to avoid causing a flash.
2334
2335         Tests: fast/images/async-image-background-change.html
2336                fast/images/async-image-src-change.html
2337                http/tests/multipart/multipart-async-image.html
2338
2339         * page/PageOverlayController.cpp:
2340         (WebCore::PageOverlayController::paintContents):
2341         * page/PageOverlayController.h:
2342         * page/linux/ResourceUsageOverlayLinux.cpp:
2343         * page/mac/ServicesOverlayController.h:
2344         * page/mac/ServicesOverlayController.mm:
2345         (WebCore::ServicesOverlayController::Highlight::paintContents):
2346         * platform/graphics/BitmapImage.h:
2347         * platform/graphics/GraphicsLayer.cpp:
2348         (WebCore::GraphicsLayer::paintGraphicsLayerContents):
2349         * platform/graphics/GraphicsLayer.h:
2350         * platform/graphics/GraphicsLayerClient.h:
2351         (WebCore::GraphicsLayerClient::paintContents):
2352         * platform/graphics/avfoundation/cf/MediaPlayerPrivateAVFoundationCF.cpp:
2353         (WebCore::LayerClient::platformCALayerPaintContents):
2354         * platform/graphics/ca/GraphicsLayerCA.cpp:
2355         (WebCore::GraphicsLayerCA::platformCALayerPaintContents):
2356         * platform/graphics/ca/GraphicsLayerCA.h:
2357         * platform/graphics/ca/PlatformCALayer.h:
2358         * platform/graphics/ca/PlatformCALayerClient.h:
2359         (WebCore::PlatformCALayerClient::platformCALayerRepaintCount):
2360         * platform/graphics/ca/TileCoverageMap.cpp:
2361         (WebCore::TileCoverageMap::platformCALayerPaintContents):
2362         * platform/graphics/ca/TileCoverageMap.h:
2363         * platform/graphics/ca/TileGrid.cpp:
2364         (WebCore::TileGrid::platformCALayerPaintContents):
2365         (WebCore::TileGrid::platformCALayerRepaintCount):
2366         * platform/graphics/ca/TileGrid.h:
2367         * platform/graphics/ca/cocoa/PlatformCALayerCocoa.mm:
2368         (PlatformCALayer::drawLayerContents):
2369         * platform/graphics/ca/win/PlatformCALayerWin.cpp:
2370         (PlatformCALayer::drawLayerContents):
2371         * platform/graphics/ca/win/PlatformCALayerWinInternal.cpp:
2372         (PlatformCALayerWinInternal::displayCallback):
2373         * platform/graphics/ca/win/WebTiledBackingLayerWin.cpp:
2374         (WebTiledBackingLayerWin::displayCallback):
2375         * platform/graphics/mac/WebLayer.mm:
2376         (-[WebLayer drawInContext:]):
2377         (-[WebSimpleLayer drawInContext:]):
2378         * rendering/RenderBoxModelObject.cpp:
2379         (WebCore::RenderBoxModelObject::decodingModeForImageDraw):
2380         * rendering/RenderElement.h:
2381         * rendering/RenderLayerBacking.cpp:
2382         (WebCore::RenderLayerBacking::paintContents):
2383         * rendering/RenderLayerBacking.h:
2384         * rendering/RenderLayerCompositor.cpp:
2385         (WebCore::RenderLayerCompositor::paintContents):
2386         * rendering/RenderLayerCompositor.h:
2387         * testing/Internals.cpp:
2388         (WebCore::imageFromImageElement):
2389         (WebCore::bitmapImageFromImageElement):
2390         (WebCore::Internals::imageFrameIndex):
2391         (WebCore::Internals::setImageFrameDecodingDuration):
2392         (WebCore::Internals::resetImageAnimation):
2393         (WebCore::Internals::isImageAnimating):
2394         (WebCore::Internals::setClearDecoderAfterAsyncFrameRequestForTesting):
2395         (WebCore::Internals::imageDecodeCount):
2396         (WebCore::Internals::setLargeImageAsyncDecodingEnabledForTesting):
2397         * testing/Internals.h:
2398         * testing/Internals.idl:
2399
2400 2017-07-18  Chris Dumez  <cdumez@apple.com>
2401
2402         DOMException should have its properties on the prototype
2403         https://bugs.webkit.org/show_bug.cgi?id=174597
2404
2405         Reviewed by Alex Christensen.
2406
2407         DOMException should have its properties on the prototype, as other interfaces:
2408         - https://heycam.github.io/webidl/#idl-DOMException
2409
2410         No new tests, rebaselined existing tests.
2411
2412         * bindings/scripts/CodeGeneratorJS.pm:
2413         (InterfaceRequiresAttributesOnInstance):
2414
2415 2017-07-18  Basuke Suzuki  <Basuke.Suzuki@sony.com>
2416
2417         [Curl] Unify ResourceHandleManager into CurlJobManager.
2418         https://bugs.webkit.org/show_bug.cgi?id=173991
2419
2420         Use CurlJobManager to make ResourceHandle run in background.
2421         CurlJobManager::start() now accept callback to manage life cycle
2422         of the resource easily in a same context.
2423         Actual management tasks of CurlJobManager is now private class
2424         in CurlJobManager which is better thread safeness.
2425         ResourceHandleManager is deleted at all.
2426
2427         Reviewed by Alex Christensen.
2428
2429         * platform/Curl.cmake:
2430         * platform/network/ResourceHandle.h:
2431         * platform/network/curl/CookieJarCurl.cpp:
2432         (WebCore::cookiesForSession):
2433         * platform/network/curl/CurlContext.cpp:
2434         (WebCore::CurlHandle::CurlHandle):
2435         (WebCore::CurlHandle::~CurlHandle):
2436         (WebCore::CurlHandle::errorDescription):
2437         (WebCore::CurlHandle::perform):
2438         (WebCore::CurlHandle::pause):
2439         (WebCore::CurlHandle::appendRequestHeaders):
2440         (WebCore::CurlHandle::appendRequestHeader):
2441         (WebCore::CurlHandle::enableRequestHeaders):
2442         (WebCore::CurlHandle::fetchCookieList):
2443         (WebCore::CurlHandle::getEffectiveURL):
2444         (WebCore::CurlHandle::setPrivateData): Deleted.
2445         (WebCore::CurlHandle::clearRequestHeaders): Deleted.
2446         (WebCore::CurlHandle::getCookieList): Deleted.
2447         (WebCore::CurlHandle::clearCookieList): Deleted.
2448         * platform/network/curl/CurlContext.h:
2449         (WebCore::CurlSList::CurlSList):
2450         (WebCore::CurlSList::~CurlSList):
2451         (WebCore::CurlSList::operator struct curl_slist** ):
2452         (WebCore::CurlSList::head):
2453         (WebCore::CurlSList::isEmpty):
2454         (WebCore::CurlSList::clear):
2455         (WebCore::CurlSList::append):
2456         (WebCore::CurlHandle::errorCode):
2457         (WebCore::CurlHandle::setErrorCode):
2458         (WebCore::CurlHandle::privateData):
2459         (WebCore::CurlHandle::setPrivateData):
2460         * platform/network/curl/CurlDownload.cpp:
2461         (WebCore::CurlDownload::start):
2462         (WebCore::CurlDownload::cancel):
2463         (WebCore::CurlDownload::addHeaders):
2464         (WebCore::CurlDownload::didReceiveData):
2465         (WebCore::CurlDownload::didFail):
2466         (WebCore::CurlDownload::handleCurlMsg): Deleted.
2467         * platform/network/curl/CurlDownload.h:
2468         * platform/network/curl/CurlJobManager.cpp:
2469         (WebCore::CurlJobList::append):
2470         (WebCore::CurlJobList::cancel):
2471         (WebCore::CurlJobList::complete):
2472         (WebCore::CurlJobList::isEmpty):
2473         (WebCore::CurlJobList::withJob):
2474         (WebCore::CurlJobList::withCurlHandle):
2475         (WebCore::CurlJob::invoke):
2476         (WebCore::CurlJobManager::add):
2477         (WebCore::CurlJobManager::cancel):
2478         (WebCore::CurlJobManager::callOnJobThread):
2479         (WebCore::CurlJobManager::startThreadIfNeeded):
2480         (WebCore::CurlJobManager::stopThreadIfNoMoreJobRunning):
2481         (WebCore::CurlJobManager::stopThread):
2482         (WebCore::CurlJobManager::updateJobs):
2483         (WebCore::CurlJobManager::workerThread):
2484         (WebCore::CurlJobManager::CurlJobManager): Deleted.
2485         (WebCore::CurlJobManager::~CurlJobManager): Deleted.
2486         (WebCore::CurlJobManager::remove): Deleted.
2487         (WebCore::CurlJobManager::getActiveCount): Deleted.
2488         (WebCore::CurlJobManager::getPendingCount): Deleted.
2489         (WebCore::CurlJobManager::stopThreadIfIdle): Deleted.
2490         (WebCore::CurlJobManager::updateHandleList): Deleted.
2491         (WebCore::CurlJobManager::addToCurl): Deleted.
2492         (WebCore::CurlJobManager::removeFromCurl): Deleted.
2493         * platform/network/curl/CurlJobManager.h:
2494         (WebCore::CurlJob::CurlJob):
2495         (WebCore::CurlJob::~CurlJob):
2496         (WebCore::CurlJob::operator=):
2497         (WebCore::CurlJob::curlHandle):
2498         (WebCore::CurlJob::ticket):
2499         (WebCore::CurlJob::finished):
2500         (WebCore::CurlJob::error):
2501         (WebCore::CurlJob::cancel):
2502         (WebCore::CurlJobManager::singleton):
2503         (WebCore::CurlJobManager::~CurlJobManager):
2504         (WebCore::CurlJobManager::isActiveJob):
2505         (WebCore::CurlJobManager::runThread): Deleted.
2506         (WebCore::CurlJobManager::setRunThread): Deleted.
2507         * platform/network/curl/ResourceError.h:
2508         (WebCore::ResourceError::ResourceError):
2509         * platform/network/curl/ResourceHandleCurl.cpp:
2510         (WebCore::ResourceHandle::~ResourceHandle):
2511         (WebCore::ResourceHandle::start):
2512         (WebCore::ResourceHandle::cancel):
2513         (WebCore::ResourceHandle::initialize):
2514         (WebCore::ResourceHandle::applyAuthentication):
2515         (WebCore::getFormElementsCount):
2516         (WebCore::ResourceHandle::setupPUT):
2517         (WebCore::ResourceHandle::setupPOST):
2518         (WebCore::ResourceHandle::setupFormData):
2519         (WebCore::ResourceHandle::setHostAllowsAnyHTTPSCertificate):
2520         (WebCore::ResourceHandle::setClientCertificateInfo):
2521         (WebCore::ResourceHandle::platformSetDefersLoading):
2522         (WebCore::ResourceHandle::didFinish):
2523         (WebCore::ResourceHandle::didFail):
2524         (WebCore::ResourceHandle::shouldUseCredentialStorage):
2525         (WebCore::ResourceHandle::didReceiveAuthenticationChallenge):
2526         (WebCore::ResourceHandle::receivedCredential):
2527         (WebCore::ResourceHandle::receivedRequestToContinueWithoutCredential):
2528         (WebCore::ResourceHandle::receivedCancellation):
2529         (WebCore::ResourceHandle::calculateWebTimingInformations):
2530         (WebCore::ResourceHandle::handleLocalReceiveResponse):
2531         (WebCore::ResourceHandle::willPrepareSendData):
2532         (WebCore::ResourceHandle::didReceiveHeaderLine):
2533         (WebCore::ResourceHandle::didReceiveAllHeaders):
2534         (WebCore::ResourceHandle::didReceiveContentData):
2535         (WebCore::ResourceHandle::readCallback):
2536         (WebCore::ResourceHandle::headerCallback):
2537         (WebCore::ResourceHandle::writeCallback):
2538         (WebCore::ResourceHandle::platformLoadResourceSynchronously):
2539         (WebCore::ResourceHandle::dispatchSynchronousJob):
2540         (WebCore::calculateWebTimingInformations): Deleted.
2541         (WebCore::handleLocalReceiveResponse): Deleted.
2542         (WebCore::writeCallback): Deleted.
2543         (WebCore::headerCallback): Deleted.
2544         (WebCore::readCallback): Deleted.
2545         (WebCore::setupFormData): Deleted.
2546         (WebCore::ResourceHandle::handleCurlMsg): Deleted.
2547         * platform/network/curl/ResourceHandleManager.cpp: Removed.
2548         * platform/network/curl/ResourceHandleManager.h: Removed.
2549
2550 2017-07-17  Konstantin Tokarev  <annulen@yandex.ru>
2551
2552         Unreviewed fix to Mac CMake build after r219474.
2553
2554         * PlatformMac.cmake:
2555
2556 2017-07-17  Konstantin Tokarev  <annulen@yandex.ru>
2557
2558         [cmake] Set library types before their targets are created
2559         https://bugs.webkit.org/show_bug.cgi?id=174600
2560
2561         Reviewed by Michael Catanzaro.
2562
2563         Since r219560 library targets are created before PlatformXXX.cmake
2564         files are processed, however library type must be passed in
2565         add_library() call and cannot be changed afterwards. Set these
2566         variables in OptionsXXX.cmake.
2567
2568         No new tests needed.
2569
2570         * PlatformMac.cmake:
2571
2572 2017-07-17  Yoshiaki Jitsukawa  <Yoshiaki.Jitsukawa@sony.com>
2573
2574         Move USE_AVFOUNDATION definition on Windows to wtf/Platform.h
2575         https://bugs.webkit.org/show_bug.cgi?id=174356
2576
2577         Reviewed by Brent Fulgham.
2578
2579         Move the definition of USE_AVFOUNDATION on Windows to wtf/platform.h
2580
2581         Rename WebCoreHeaderDetection.h to AVFoundationHeaderDetection.h
2582
2583         * AVFoundationSupport.py: Moved to Source/WTF/AVFoundationSupport.py.
2584         * DerivedSources.make:
2585         * PlatformWin.cmake:
2586         * config.h:
2587         * platform/graphics/ca/win/PlatformCALayerWin.cpp:
2588         * platform/graphics/ca/win/PlatformCALayerWinInternal.cpp:
2589
2590 2017-07-17  Sam Weinig  <sam@webkit.org>
2591
2592         [WebIDL] Remove custom bindings that require non-caching JS strings
2593         https://bugs.webkit.org/show_bug.cgi?id=174552
2594
2595         Reviewed by Darin Adler.
2596
2597         Adds two new types, UncachedString and OwnedString (both of which
2598         are simple structs that wrap a String) to communicate to the bindings
2599         layer which function to use when converting to a JS string.
2600
2601         * bindings/IDLTypes.h:
2602         (WebCore::IDLString::isNullValue):
2603         Add overloads of isNullValue for UncachedString and OwnedString.
2604
2605         * bindings/js/JSDOMConvertStrings.h:
2606         (WebCore::JSConverter<IDLDOMString>::convert):
2607         (WebCore::JSConverter<IDLByteString>::convert):
2608         (WebCore::JSConverter<IDLUSVString>::convert):
2609         Add overloads of convert for UncachedString that uses JSC::jsString
2610         and for OwnedString that uses JSC::jsOwnedString.
2611
2612         * bindings/js/JSHTMLCanvasElementCustom.cpp:
2613         (WebCore::JSHTMLCanvasElement::toDataURL): Deleted.
2614         Remove custom binding for toDataURL.
2615
2616         * bindings/js/JSXMLHttpRequestCustom.cpp:
2617         (WebCore::JSXMLHttpRequest::retrieveResponse):
2618         (WebCore::JSXMLHttpRequest::responseText): Deleted.
2619         Remove custom binding for responseText. Replace the caller of the
2620         binding function with a simple conversion that will do the same thing.
2621
2622         * bindings/js/StringAdaptors.h: Added.
2623         Add UncachedString and OwnedString adaptors.
2624
2625         * html/HTMLCanvasElement.cpp:
2626         (WebCore::HTMLCanvasElement::toDataURL):
2627         Move quality conversion here, matching toBlob and the spec. Return a UncachedString
2628         to instruct the bindings to use JSC::jsString and not JSC::jsStringWithCache.
2629
2630         (WebCore::HTMLCanvasElement::toBlob):
2631         Use asNumber() rather than toNumber(), since we just checked that it is a number.
2632
2633         * html/HTMLCanvasElement.h:
2634         Update header to account for returning an UncachedString.
2635
2636         * html/HTMLCanvasElement.idl:
2637         Remove [Custom] and make the signature of toDataURL match the spec (and what we have been doing).
2638
2639         * inspector/InspectorCanvasAgent.cpp:
2640         (WebCore::InspectorCanvasAgent::requestContent):
2641         Update to account for UncachedString.
2642
2643         * xml/XMLHttpRequest.cpp:
2644         (WebCore::XMLHttpRequest::responseText):
2645         * xml/XMLHttpRequest.h:
2646         Update responseText to return OwnedString. OwnedString instruct the bindings to
2647         use JSC::jsOwnedString and not JSC::jsStringWithCache.
2648
2649         * xml/XMLHttpRequest.idl:
2650         Remove [CustomGetter].
2651
2652 2017-07-17  Daniel Bates  <dabates@apple.com>
2653
2654         Cleanup: Use OptionSet to represent marker types
2655         https://bugs.webkit.org/show_bug.cgi?id=174594
2656
2657         Reviewed by Darin Adler.
2658
2659         Remove class DocumentMarker::MarkerTypes that duplicates most of functionality of OptionSet
2660         and use OptionSet directly to represent a set of marker types.
2661
2662         No functionality changed. So, no new tests.
2663
2664         * dom/Document.cpp:
2665         (WebCore::Document::updateLayout):
2666         * dom/DocumentMarker.h:
2667         (WebCore::DocumentMarker::MarkerTypes::MarkerTypes): Deleted.
2668         (WebCore::DocumentMarker::MarkerTypes::contains): Deleted.
2669         (WebCore::DocumentMarker::MarkerTypes::intersects): Deleted.
2670         (WebCore::DocumentMarker::MarkerTypes::operator==): Deleted.
2671         (WebCore::DocumentMarker::MarkerTypes::add): Deleted.
2672         (WebCore::DocumentMarker::MarkerTypes::remove): Deleted.
2673         (WebCore::DocumentMarker::AllMarkers::AllMarkers): Deleted.
2674         * dom/DocumentMarkerController.cpp:
2675         (WebCore::DocumentMarkerController::possiblyHasMarkers):
2676         (WebCore::DocumentMarkerController::detach):
2677         (WebCore::DocumentMarkerController::removeMarkers):
2678         (WebCore::DocumentMarkerController::addMarker):
2679         (WebCore::DocumentMarkerController::copyMarkers):
2680         (WebCore::DocumentMarkerController::markersFor):
2681         (WebCore::DocumentMarkerController::markersInRange):
2682         (WebCore::DocumentMarkerController::removeMarkersFromList):
2683         (WebCore::DocumentMarkerController::repaintMarkers):
2684         (WebCore::DocumentMarkerController::shiftMarkers):
2685         (DocumentMarkerController::setMarkersActive):
2686         (DocumentMarkerController::hasMarkers):
2687         (DocumentMarkerController::clearDescriptionOnMarkersIntersectingRange):
2688         Update code as needed.
2689
2690         * dom/DocumentMarkerController.h:
2691         (WebCore::DocumentMarkerController::hasMarkers):
2692         Update code as needed.
2693
2694         * editing/AlternativeTextController.cpp:
2695         (WebCore::markerTypesForAutocorrection):
2696         (WebCore::markerTypesForReplacement):
2697         (WebCore::markerTypesForAppliedDictationAlternative):
2698         Marked as inline and returns an OptionSet<DocumentMarker::MarkerType>. It is unnecessary to use NeverDestroyed
2699         in these functions as constructing an OptionSet and copying/moving it is very efficient. Such operations are
2700         effectively equivalent to an integral assignment and copy of an integral value, respectively.
2701
2702         (WebCore::AlternativeTextController::applyAlternativeTextToRange):
2703         (WebCore::AlternativeTextController::respondToUnappliedSpellCorrection):
2704         (WebCore::AlternativeTextController::markCorrection):
2705         Update code as needed.
2706
2707         * editing/AlternativeTextController.h: While I am here, remove some unnecessary #includes and group
2708         forward declarations of structs.
2709         * editing/CompositeEditCommand.cpp:
2710         (WebCore::CompositeEditCommand::replaceTextInNodePreservingMarkers): Update code as needed.
2711         * editing/Editor.cpp:
2712         (WebCore::Editor::updateMarkersForWordsAffectedByEditing): Ditto.
2713         * editing/SpellChecker.cpp:
2714         (WebCore::SpellChecker::didCheckSucceed): Ditto.
2715         * page/ios/FrameIOS.mm:
2716         (WebCore::Frame::interpretationsForCurrentRoot): Ditto.
2717         * testing/Internals.cpp:
2718         (WebCore::markerTypesFrom): Ditto.
2719         (WebCore::Internals::markerCountForNode): Ditto.
2720         (WebCore::Internals::markerAt): Ditto.
2721
2722 2017-07-17  Devin Rousso  <drousso@apple.com>
2723
2724         Web Inspector: overlay page highlight doesn't disappear when a page is constantly updating
2725         https://bugs.webkit.org/show_bug.cgi?id=174468
2726
2727         Reviewed by Simon Fraser.
2728
2729         Do not allow the PageOverlay to start another fade animation of the same type if one has
2730         already started. As an example, if the PageOverlay is fading out, startFadeOutAnimation
2731         should just return.
2732
2733         * page/PageOverlay.cpp:
2734         (WebCore::PageOverlay::startFadeInAnimation):
2735         (WebCore::PageOverlay::startFadeOutAnimation):
2736
2737 2017-07-17  Darin Adler  <darin@apple.com>
2738
2739         Improve use of NeverDestroyed
2740         https://bugs.webkit.org/show_bug.cgi?id=174348
2741
2742         Reviewed by Sam Weinig.
2743
2744         * Modules/encryptedmedia/MediaKeySession.cpp: Removed unneeded include of
2745         NeverDestroyed.h.
2746
2747         * Modules/encryptedmedia/legacy/LegacyCDM.cpp:
2748         (WebCore::installedCDMFactories): Use makeNeverDestroyed and a lambda rather
2749         than a separate boolean for initialization.
2750
2751         * Modules/encryptedmedia/legacy/LegacyCDMSessionClearKey.cpp:
2752         (WebCore::clearKeyVM): Removed unneeded use of NeverDestroyed to hold a
2753         RefPtr. Simple to just use leakRef instead.
2754
2755         * Modules/gamepad/GamepadManager.cpp: Added an include of NeverDestroyed.h.
2756
2757         * Modules/indexeddb/IDBTransaction.cpp: Removed unneeded include of
2758         NeverDestroyed.h.
2759         * Modules/indexeddb/server/MemoryObjectStore.cpp: Ditto.
2760
2761         * Modules/mediasession/MediaSessionManager.cpp: Moved include of
2762         NeverDestroyed.h here ...
2763         * Modules/mediasession/MediaSessionManager.h: ... from here.
2764
2765         * Modules/mediasource/MediaSourceRegistry.cpp: Moved include of
2766         NeverDestroyed.h here ...
2767         * Modules/mediasource/MediaSourceRegistry.h: ... from here.
2768
2769         * Modules/mediasource/SourceBuffer.cpp: Removed unneeded include of
2770         NeverDestroyed.h.
2771
2772         * Modules/plugins/QuickTimePluginReplacement.h: Initialize
2773         m_scriptObject in the class definition.
2774
2775         * Modules/plugins/QuickTimePluginReplacement.mm:
2776         (WebCore::QuickTimePluginReplacement::supportsMimeType): Use
2777         makeNeverDestroyed and the HashSet constructor instead of a loop.
2778         (WebCore::QuickTimePluginReplacement::supportsFileExtension): Ditto.
2779         (WebCore::QuickTimePluginReplacement::QuickTimePluginReplacement): Did a
2780         slight cleanup of the initializers.
2781         (WebCore::QuickTimePluginReplacement::~QuickTimePluginReplacement): Added
2782         a FIXME; unclear why there is code here at all.
2783
2784         * accessibility/AccessibilityRenderObject.cpp:
2785         (WebCore::AccessibilityRenderObject::inheritsPresentationalRole): Use
2786         makeNeverDestroyed and the Vector constructor instead of using empty
2787         vector checks and Vector::add. Use std::any_of rather than Vector::contains
2788         since we now are matching against pointers.
2789
2790         * bindings/scripts/CodeGeneratorJS.pm:
2791         (GenerateEnumerationImplementationContent): Use const NeverDestroyed.
2792
2793         * bindings/scripts/test/JS/JSTestCallbackInterface.cpp:
2794         * bindings/scripts/test/JS/JSTestObj.cpp:
2795         * bindings/scripts/test/JS/JSTestStandaloneDictionary.cpp:
2796         * bindings/scripts/test/JS/JSTestStandaloneEnumeration.cpp:
2797         Regenerated.
2798
2799         * css/CSSProperty.cpp:
2800         (WebCore::borderDirections): Removed unnecesssary use of NeverDestroyed
2801         for a global object that has a trivial destructor.
2802
2803         * css/ElementRuleCollector.cpp:
2804         (WebCore::leftToRightDeclaration): Removed unneeded use of NeverDestroyed to hold a
2805         Ref. Simple to just use leakRef instead. Also use a lambda instead of an explicit
2806         isEmpty check each time this is called.
2807         (WebCore::rightToLeftDeclaration): Ditto.
2808
2809         * css/makeprop.pl: Removed unneeded include of NeverDestroyed.h and the
2810         uneeded global emptyShorthand. Constructing an empty StylePropertyShorthand
2811         is no less efficient than copying a global empty one was.
2812
2813         * dom/CustomElementReactionQueue.cpp: Added an include of NeverDestroyed.h.
2814
2815         * dom/DOMImplementation.cpp: Removed unneeded include of NeverDestroyed.h.
2816         * dom/InputEvent.cpp: Ditto.
2817
2818         * dom/Microtasks.cpp: Moved include of NeverDestroyed.h here ...
2819         * dom/Microtasks.h: ... from here.
2820
2821         * dom/MutationObserver.cpp: Added an include of NeverDestroyed.h.
2822
2823         * dom/ScopedEventQueue.cpp: Moved include of NeverDestroyed.h here ...
2824         * dom/ScopedEventQueue.h: ... from here. Added Forward.h.
2825
2826         * dom/ScriptElement.cpp:
2827         (WebCore::isLegacySupportedJavaScriptLanguage): Use makeNeverDestroyed
2828         and the HashSet constructor rather than an isEmpty check and a lot of
2829         add function calls. Also removed comments that don't have value any more,
2830         with dubious no longer relevant claims about the behavior of old web browsers.
2831         The function that calls this one already has sufficient comments about why we
2832         hope some day this function can be eliminated.
2833
2834         * dom/SecurityContext.cpp: Removed unneeded include of NeverDestroyed.h.
2835
2836         * dom/make_names.pl:
2837         (printFactoryCppFile): Instead of using a "populate" function, use a function
2838         that creates as HashMap. Then use const auto and makeNeverDestroyed.
2839         (printWrapperFactoryCppFile): Ditto.
2840
2841         * editing/AlternativeTextController.cpp:
2842         (WebCore::markerTypesForAutocorrection): Use makeNeverDestroyed
2843         and the Vector constructor rather than an isEmpty check and a lot of
2844         append function calls.
2845         (WebCore::markerTypesForReplacement): Ditto.
2846         (WebCore::markerTypesForAppliedDictationAlternative): Ditto.
2847
2848         * editing/EditingStyle.cpp:
2849         (WebCore::htmlElementEquivalents): Use const auto, makeNeverDestroyed,
2850         the Vector constructor, and new rather than make_unique. Changed return
2851         type to use const pointers rather than unique_ptr with non-const type.
2852         (WebCore::EditingStyle::conflictsWithImplicitStyleOfElement): Updated
2853         for the above change to htmlElementEquivalents.
2854         (WebCore::htmlAttributeEquivalents): Same approach as above.
2855         (WebCore::EditingStyle::conflictsWithImplicitStyleOfAttributes): Ditto.
2856         (WebCore::elementMatchesAndPropertyIsNotInInlineStyleDecl): Take
2857         references instead of pointers.
2858         (WebCore::EditingStyle::mergeInlineAndImplicitStyleOfElement): Updated
2859         for the above.
2860
2861         * editing/FormatBlockCommand.cpp:
2862         (WebCore::isElementForFormatBlock): Use const auto, makeNeverDestroyed,
2863         and the HashSet constructor rather than isEmpty and a lot of calls to add.
2864         * editing/RemoveFormatCommand.cpp:
2865         (WebCore::isElementForRemoveFormatCommand): Ditto.
2866
2867         * editing/ReplaceSelectionCommand.cpp:
2868         (WebCore::isProhibitedParagraphChild): Use const auto, makeNeverDestroyed,
2869         and a lambda, rather than isEmpty and a lot of calls to add.
2870
2871         * html/Autofill.cpp:
2872         (WebCore::fieldNameMap): Changed return type to be const. Use const auto,
2873         makeNeverDestroyed, a lambda, and an array of values rather than isEmpty
2874         and a lot of calls to add. Stopped doing the ConstructFromLiteral
2875         optimization here. (Easy to add it back if that is a mistake.)
2876
2877         * html/HTMLObjectElement.cpp:
2878         (WebCore::isRecognizedTagName): Use const auto, makeNeverDestroyed, and
2879         a lambda rather than isEmpty.
2880
2881         * html/HTMLStyleElement.cpp: Added include of NeverDestroyed.h.
2882
2883         * html/HTMLVideoElement.cpp: Removed uneeded include of NeverDestroyed.h.
2884
2885         * html/InputType.cpp:
2886         (WebCore::createInputTypeFactoryMap): Replaced the populate function
2887         with this create function.
2888         (WebCore::InputType::create): Use const auto and makeNeverDestroyed
2889         istead of isEmpty. Also put the map right where it is used so it's
2890         not initialized in code path where not needed.
2891
2892         * html/parser/HTMLParserIdioms.cpp: Removed uneeded include of
2893         NeverDestroyed.h.
2894
2895         * inspector/InspectorApplicationCacheAgent.cpp:
2896         (WebCore::InspectorApplicationCacheAgent::networkStateChanged):
2897         Merged into a single line.
2898
2899         * loader/ContentFilter.cpp:
2900         (WebCore::blockedPageURL): Use const auto and makeNeverDestroyed
2901         rather than std::call_once. Since this is a URL and has a non-thread-safe
2902         reference count, this was not thread safe before, so no need to use the
2903         more roundabout and less efficient idiom for its thread safety.
2904
2905         * loader/CrossOriginAccessControl.cpp:
2906         (WebCore::isOnAccessControlResponseHeaderWhitelist): Deleted.
2907         * loader/CrossOriginAccessControl.h: Deleted unused function
2908         isOnAccessControlResponseHeaderWhitelist.
2909
2910         * loader/EmptyClients.cpp: Use leakRef instead of NeverDestroyed<Ref>.
2911
2912         * loader/appcache/ApplicationCacheStorage.cpp: Removed unneeded include of
2913         NeverDestroyed.h.
2914         * page/CaptionUserPreferences.cpp: Ditto.
2915
2916         * page/DebugPageOverlays.cpp:
2917         (WebCore::touchEventRegionColors): Changed the return type to const, since
2918         the callers do not modify the map. Use const auto, makeNeverDestroyed, and
2919         a lambda rather than isEmpty and repeated calls to add.
2920
2921         * page/MainFrame.cpp: Removed unneeded include of NeverDestroyed.h.
2922
2923         * page/MemoryRelease.cpp:
2924         (WebCore::releaseMemory): Use a function directly instead of calling it
2925         inside a lambda.
2926
2927         * page/NavigatorBase.cpp:
2928         (WebCore::NavigatorBase::onLine): Updated to call
2929         NetworkStateNotifier::singleton.
2930
2931         * page/Page.cpp:
2932         (WebCore::allPages): Made this a function instead of a global to use the
2933         normal idiom for such globals.
2934         (WebCore::Page::forEachPage): Updated for change to allPages.
2935         (WebCore::networkStateChanged): Ditto. Also removed a bit of unnecessary
2936         churn by using const AtomicString& instead of AtomicString.
2937         (WebCore::Page::Page): Moved initialization of most data members to the
2938         class definition. Removed initialiation of allPages. Updated the use of
2939         NetworkStateNotifier for its new slightly changed  interface.
2940         (WebCore::Page::~Page): Updated for change to allPages.
2941         (WebCore::Page::clearPreviousItemFromAllPages): Ditto.
2942         (WebCore::Page::updateStyleForAllPagesAfterGlobalChangeInEnvironment): Ditto.
2943         (WebCore::Page::refreshPlugins): Ditto.
2944
2945         * page/Page.h: Moved initialization of most data members here.
2946
2947         * page/PerformanceUserTiming.cpp: Removed many unneded includes.
2948         (WebCore::restrictedMarkFunction): Use const auto, makeNeverDestroyed, and
2949         a lambda instead of isEmpty. Also use an array rather than std::array and
2950         got rid fo the unneeded use of anonymous namespace since this is fine in
2951         the top level WebCore namespace.
2952         (WebCore::clearPerformanceEntries): Tweaked coding style.
2953         (WebCore::UserTiming::mark): Ditto.
2954         (WebCore::UserTiming::findExistingMarkStartTime): Got rid of double hash
2955         table lookup by using find instead of using contains followed by get.
2956         removed unneeded explicit cast to double. Tweaked coding style.
2957         (WebCore::convertToEntrySequence): Tweaked coding style.
2958         (WebCore::getEntrySequenceByName): Deleted.
2959         (WebCore::UserTiming::getMarks): Call get directly instead of getEntrySequenceByName.
2960         (WebCore::UserTiming::getMeasures): Ditto.
2961
2962         * page/RuntimeEnabledFeatures.cpp: Moved include of NeverDestroyed.h here ...
2963         * page/RuntimeEnabledFeatures.h: ... from here. Added Forward.h and Noncopyable.h.
2964
2965         * platform/LocalizedStrings.cpp: Removed unneeded include of NeverDestroyed.h.
2966
2967         * platform/MIMETypeRegistry.cpp:
2968         (WebCore::initializeSupportedImageMIMETypes): Added a missing const for a global
2969         constant array.
2970         (WebCore::initializeSupportedJavaScriptMIMETypes): Ditto.
2971         (WebCore::initializeSupportedNonImageMimeTypes): Ditto.
2972         (WebCore::typesForCommonExtension): Replaced the old mediaMIMETypeMap function
2973         with this one. Moved the common media types array here since it's only used here.
2974         Use const auto, makeNeverDestroyed, and a lambda instead of using an isEmpty check.
2975         Iterate the array using a mdoern for loop. Use HashMap::ensure to avoid the
2976         double hashing that the old code was doing. And moved the code to find an entry
2977         in the map in here from the client functions.
2978         (WebCore::MIMETypeRegistry::getMediaMIMETypeForExtension): Updated to use the
2979         typesForCommonExtension, and changed logic to use this first, and only call
2980         getMIMETypeForExtension for extensions not in the map. This gives the same
2981         result but should be slightly more efficient.
2982         (WebCore::MIMETypeRegistry::getMediaMIMETypesForExtension): Updated to use the
2983         typesForCommonExtension function. Comment about strategy is now in there.
2984         (WebCore::initializeUnsupportedTextMIMETypes): Added a missing const for a global
2985         constant array. Also started using ASCIILiteral.
2986
2987         * platform/MainThreadSharedTimer.cpp: Moved include of NeverDestroyed.h here ...
2988         * platform/MainThreadSharedTimer.h: ... from here. Added Forward.h.
2989
2990         * platform/SchemeRegistry.cpp:
2991         (WebCore::add): Added helper functions to keep code below simpler.
2992         (WebCore::makeNeverDestroyedSchemeSet): Ditto.
2993         (WebCore::allBuiltinSchemes): Use const auto, makeNeverDestroyed, a lambda,
2994         and the helper functions above, rather than isEmpty.
2995         (WebCore::builtinLocalURLSchemes): Use const auto, makeNeverDestroyed, and
2996         the HashSet constructor instead of isEmpty and add calls. Also changed the
2997         return type to be const.
2998         (WebCore::localURLSchemes): Copy the map using assignment instead of an
2999         isEmpty function and a loop with calls to add.
3000         (WebCore::builtinSecureSchemes): Use const auto, makeNeverDestroyed, and
3001         the Vector constructor rather than isEmpty, repeated calls to append,
3002         and shrinkToFit.
3003         (WebCore::secureSchemes): Use auto and makeNeverDestroyedSchemeSet rather
3004         than isEmpty and repeated calls to add.
3005         (WebCore::builtinSchemesWithUniqueOrigins): More of the same.
3006         (WebCore::schemesWithUniqueOrigins): Ditto.
3007         (WebCore::builtinEmptyDocumentSchemes): Ditto.
3008         (WebCore::emptyDocumentSchemes): Ditto.
3009         (WebCore::schemesForbiddenFromDomainRelaxation): Ditto.
3010         (WebCore::builtinCanDisplayOnlyIfCanRequestSchemes): Ditto.
3011         (WebCore::canDisplayOnlyIfCanRequestSchemes): Ditto.
3012         (WebCore::builtinCORSEnabledSchemes): Ditto.
3013         (WebCore::CORSEnabledSchemes): Ditto.
3014         (WebCore::SchemeRegistry::shouldTreatURLSchemeAsLocal): Replaced check for
3015         empty string with check for null string, since that's slightly more efficient
3016         for non-null, non-empty strings, and the hash table can handle empty strings
3017         just fine.
3018         (WebCore::SchemeRegistry::registerURLSchemeAsNoAccess): Added null check here.
3019         (WebCore::SchemeRegistry::shouldTreatURLSchemeAsNoAccess): Ditto.
3020         (WebCore::SchemeRegistry::registerURLSchemeAsDisplayIsolated): Ditto.
3021         (WebCore::SchemeRegistry::shouldTreatURLSchemeAsDisplayIsolated): Ditto.
3022         (WebCore::SchemeRegistry::registerURLSchemeAsSecure): Ditto.
3023         (WebCore::SchemeRegistry::shouldTreatURLSchemeAsSecure): Ditto.
3024         (WebCore::SchemeRegistry::registerURLSchemeAsEmptyDocument): Ditto.
3025         (WebCore::SchemeRegistry::shouldLoadURLSchemeAsEmptyDocument): Ditto.
3026         (WebCore::SchemeRegistry::setDomainRelaxationForbiddenForURLScheme): Ditto.
3027         (WebCore::SchemeRegistry::isDomainRelaxationForbiddenForURLScheme): Ditto.
3028         (WebCore::SchemeRegistry::canDisplayOnlyIfCanRequest): Ditto.
3029         (WebCore::SchemeRegistry::registerAsCanDisplayOnlyIfCanRequest): Ditto.
3030         (WebCore::SchemeRegistry::registerURLSchemeAsNotAllowingJavascriptURLs): Ditto.
3031         (WebCore::SchemeRegistry::shouldTreatURLSchemeAsNotAllowingJavascriptURLs): Ditto.
3032         (WebCore::SchemeRegistry::registerURLSchemeAsAllowingLocalStorageAccessInPrivateBrowsing): Ditto.
3033         (WebCore::SchemeRegistry::allowsLocalStorageAccessInPrivateBrowsing): Ditto.
3034         (WebCore::SchemeRegistry::registerURLSchemeAsAllowingDatabaseAccessInPrivateBrowsing): Ditto.
3035         (WebCore::SchemeRegistry::allowsDatabaseAccessInPrivateBrowsing): Ditto.
3036         (WebCore::SchemeRegistry::registerURLSchemeAsCORSEnabled): Ditto.
3037         (WebCore::SchemeRegistry::shouldTreatURLSchemeAsCORSEnabled): Ditto.
3038         (WebCore::SchemeRegistry::registerURLSchemeAsBypassingContentSecurityPolicy): Ditto.
3039         (WebCore::SchemeRegistry::removeURLSchemeRegisteredAsBypassingContentSecurityPolicy): Ditto.
3040         (WebCore::SchemeRegistry::schemeShouldBypassContentSecurityPolicy): Ditto.
3041         (WebCore::SchemeRegistry::registerURLSchemeAsAlwaysRevalidated): Ditto.
3042         (WebCore::SchemeRegistry::shouldAlwaysRevalidateURLScheme): Ditto.
3043         (WebCore::SchemeRegistry::registerURLSchemeAsCachePartitioned): Ditto.
3044         (WebCore::SchemeRegistry::shouldPartitionCacheForURLScheme): Ditto.
3045         (WebCore::SchemeRegistry::isUserExtensionScheme): Tweaked #if a bit.
3046         (WebCore::SchemeRegistry::isBuiltinScheme): Added null check.
3047
3048         * platform/cocoa/WebVideoFullscreenModelVideoElement.mm:
3049         (WebVideoFullscreenModelVideoElement::observedEventNames): Use const auto,
3050         makeNeverDestroyed, and the Vector constructor rather than a call to size
3051         and then append eacn time this funciton is called.
3052
3053         * platform/gamepad/cocoa/GameControllerGamepadProvider.h: Moved include of
3054         NeverDestroyed.h from here ...
3055         * platform/gamepad/cocoa/GameControllerGamepadProvider.mm: ... to here.
3056
3057         * platform/gamepad/mac/HIDGamepadProvider.cpp: Moved include of
3058         NeverDestroyed.h here ...
3059         * platform/gamepad/mac/HIDGamepadProvider.h: ... from here.
3060
3061         * platform/graphics/FontCascade.cpp:
3062         (WebCore::FontCascade::hasValidAverageCharWidth): Use const auto,
3063         makeNeverDestroyed, and the HashSet constructor.
3064
3065         * platform/graphics/ImageBuffer.cpp:
3066         (WebCore::ImageBuffer::transformColorSpace): Removed unneeded use of
3067         NeverDestroyed<Vector<int>> here, instead using const std::array.
3068
3069         * platform/graphics/ImageBuffer.h: Changed platformTransformColorSpace to take
3070         const std::array<uint8_t, 256>& instead of const Vector<int>&, since the whole
3071         pointer of the argument is that it's a table to map bytes onto other bytes.
3072
3073         * platform/graphics/MediaPlaybackTarget.h: Removed unneeded include of
3074         NeverDestroyed.h, unneeded unused noMediaPlaybackTargetContext function,
3075         and unneeded include of MediaPlaybackTargetContext.h, forward declaring instead.
3076         Made most functions in this class pure virtual instead of having default
3077         implementations.
3078
3079         * platform/graphics/avfoundation/MediaPlaybackTargetMac.h: Added include
3080         of MediaPlaybackTargetContext.h now that it was removed above.
3081
3082         * platform/graphics/avfoundation/MediaPlayerPrivateAVFoundation.cpp:
3083         (WebCore::MediaPlayerPrivateAVFoundation::staticMIMETypeList): Use const auto,
3084         makeNeverDestroyed, and the HashSet constructor.
3085
3086         * platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaStreamAVFObjC.mm:
3087         (WebCore::MediaPlayerPrivateMediaStreamAVFObjC::isAvailable): Reversed the sense
3088         of the boolean logic here to make this simpler and easier to read.
3089         (WebCore::MediaPlayerPrivateMediaStreamAVFObjC::getSupportedTypes): Changed
3090         the implementation here, which empties out the passed in hash set, to use
3091         the clear function rather than assignment from a global empty hash set. Added
3092         a FIXME because it seems peculiar that this would be considered OK.
3093         (WebCore::MediaPlayerPrivateMediaStreamAVFObjC::supportsType): Used ? : to make
3094         this a little more terse; I think it's clearer.
3095
3096         * platform/graphics/cairo/ImageBufferCairo.cpp:
3097         (WebCore::ImageBuffer::platformTransformColorSpace): Updated for change to
3098         argument type.
3099
3100         * platform/graphics/cg/ImageDecoderCG.cpp: Removed unneeded include of
3101         NeverDestroyed.h.
3102         (WebCore::appendImageSourceOption): Streamlined code using auto with
3103         adoptCF; there is no need to write out the type of the RetainPtr.
3104         (WebCore::imageSourceOptions): Instead of NeverDestroyed<RetainPtr>, just use
3105         const auto and leakRef.
3106         (WebCore::imageSourceAsyncOptions): Ditto.
3107
3108         * platform/graphics/gtk/GdkCairoUtilities.cpp:
3109         (WebCore::getDefaultCairoFontOptions): Instead of using LazyNeverDestroyed
3110         on a pointer, just use a global pointer. There is no need to use any kind
3111         of NeverDestroyed on a type with a trivial destructor such as a pointer.
3112
3113         * platform/graphics/ios/FontCacheIOS.mm:
3114         (WebCore::platformFontWithFamilySpecialCase): Instead of using
3115         NeverDestroyed<RetainPtr> just use a raw pointer global.
3116         * platform/graphics/mac/FontCacheMac.mm:
3117         (WebCore::platformFontWithFamilySpecialCase): Ditto.
3118
3119         * platform/graphics/mac/MediaPlayerPrivateQTKit.mm:
3120         (WebCore::createFileTypesSet): Return the HashSet instead of adding to it.
3121         Use Objective-C for/in syntax for an NSArray, not C++ for syntax, although
3122         it seems the C++ syntax was working fine.
3123         (WebCore::mimeCommonTypesCache): Fixed return type so we won't copy the
3124         HashMap every time this function is called. Use const auto and makeNeverDestroyed
3125         instead of using a boolean to do one time initialization.
3126         (WebCore::mimeModernTypesCache): Ditto.
3127         (WebCore::MediaPlayerPrivateQTKit::disableUnsupportedTracks): Marked a
3128         constant array const.
3129
3130         * platform/ios/Device.cpp:
3131         (WebCore::deviceClass): Use a lambda to initialize a variable here instead
3132         of using std::call_once. This function does not need to be thread safe.
3133         (WebCore::deviceName): Use const NeverDestroyed instead of LazyNeverDestroyed
3134         and std::call_once. This function does not need to be thread safe.
3135
3136         * platform/ios/DragImageIOS.mm: Make defaultLinkIndicatorOptions be a
3137         constant instead of a variable.
3138         (WebCore::cascadeForSystemFont): Added helper function.
3139         (WebCore::createDragImageForLink): Use const auto and makeNeverDestroyed
3140         instead of LazyNeverDestroyed and dispatch_once. This code does not need
3141         to be thread safe since its arguments include, for example, a DOM element.
3142
3143         * platform/ios/LegacyTileLayerPool.h: Removed unneeded include of
3144         NeverDestroyed.h.
3145
3146         * platform/ios/QuickLook.mm:
3147         (WebCore::QLPreviewGetSupportedMIMETypesSet): Use a global raw pointer
3148         instead of NeverDestroyed<RetainPtr<NSSet>>.
3149         (WebCore::createQLPreviewProtocol): Deleted.
3150         (WebCore::QLPreviewProtocol): Use a global raw pointer instead of
3151         a NeverDestroyed<Vector<char>>. It's cleaner to use fastStrdup instead
3152         of appending to an array, and less wasteful of memory too.
3153
3154         * platform/ios/WebCoreMotionManager.mm: Removed unneeded include of
3155         NeverDestroyed.h.
3156         * platform/ios/WebSQLiteDatabaseTrackerClient.h: Ditto.
3157         * platform/mac/DragImageMac.mm: Ditto.
3158         * platform/mediastream/CaptureDeviceManager.cpp: Ditto.
3159         * platform/mediastream/RealtimeMediaSourceSupportedConstraints.cpp: Ditto.
3160
3161         * platform/mock/MockRealtimeMediaSource.cpp:
3162         (WebCore::MockRealtimeMediaSource::audioDevices): Use auto,
3163         makeNeverDestroyed, and a lambda instead of checking size each time.
3164         (WebCore::MockRealtimeMediaSource::videoDevices): Ditto.
3165
3166         * platform/mock/mediasource/MockMediaPlayerMediaSource.cpp:
3167         (WebCore::mimeTypeCache): Use const auto, makeNeverDestroyed, and the
3168         HashSet constructor instead of an explicit boolean to initialize.
3169
3170         * platform/network/BlobRegistryImpl.cpp:
3171         (WebCore::blobUtilityQueue): Use a global with a raw reference instead
3172         of a NeverDestroyed<Ref>.
3173
3174         * platform/network/NetworkStateNotifier.cpp:
3175         (WebCore::NetworkStateNotifier::singleton): Renamed this from
3176         networkStateNotifier. We must have missed this when we changed to use the
3177         singleton idiom consistently. Also changed to use normal NeverDestroyed
3178         rather than using LazyNeverDestroyed and call_once, because this is not
3179         used from multiple threads and has no special thread safety requirements.
3180         (WebCore::NetworkStateNotifier::NetworkStateNotifier): Moved the constructor
3181         here and construct the timer. Before the timer was used only on the Mac
3182         platform but now it is there for all platforms.
3183         (WebCore::NetworkStateNotifier::onLine): Moved function here from the header.
3184         It's a littlel less trivial than before because it now triggers a call to
3185         updateState as needed.
3186         (WebCore::NetworkStateNotifier::addListener): Renamed from
3187         addNetworkStateChangeListener since there is only one kind of listener and
3188         the old name was too wordy. Changed to call the new startObserving function,
3189         and guaranteed to call it onlh once. The platform-specific logic for iOS is
3190         now inside the iOS version of that function instead of here.
3191         (WebCore::NetworkStateNotifier::updateState): Moved here now that it is the
3192         same for all platforms. The actual state updating is done in a function named
3193         updateStateWithoutNotifying, which is implemented for each platform.
3194         (WebCore::NetworkStateNotifier::updateStateSoon): New function which takes
3195         advantage of the timer to coalesce network state updates.
3196         (WebCore::NetworkStateNotifier::updateStateWithoutNotifying): Empty stub
3197         version of this function.
3198         (WebCore::NetworkStateNotifier::startObserving): Ditto.
3199         (WebCore::NetworkStateNotifier::notifyNetworkStateChange): Deleted. Code from
3200         this has moved into updateState.
3201
3202         * platform/network/NetworkStateNotifier.h: Greatly cut down what part of this
3203         class is platform-specific.
3204
3205         * platform/network/ios/NetworkStateNotifierIOS.mm:
3206         (-[WebNetworkStateObserver initWithBlock:]): Simplified this class so it
3207         takes a block, rather than using a pointer to the C++ notifier and doing
3208         WebThreadRun.
3209         (-[WebNetworkStateObserver dealloc]): Ditto.
3210         (-[WebNetworkStateObserver networkStateChanged:]): Ditto.
3211         (WebCore::NetworkStateNotifier::updateStateWithoutNotifying):
3212         This now calls isNetworkReachable.
3213         (WebCore::NetworkStateNotifier::startObserving): Check the
3214         Settings::shouldOptOutOfNetworkStateObservation function, and if it's OK
3215         to observe, then create the observer. This code now takes care of the
3216         WebThreadRun and calls updateStateSoon.
3217         (WebCore::NetworkStateNotifier::NetworkStateNotifier): Deleted. The
3218         constructor is now platform independent.
3219         (WebCore::NetworkStateNotifier::~NetworkStateNotifier): Deleted. This
3220         object is never destroyed so we should not write a destructor for it.
3221         (WebCore::NetworkStateNotifier::registerObserverIfNecessary): Deleted.
3222         Replaced by startObserving.
3223         (WebCore::NetworkStateNotifier::onLine): Deleted. This function is now
3224         platform-independent.
3225         (WebCore::setOnLine): Deleted. The logic from this is now in
3226         updateState and updateStateWithoutNotifying.
3227
3228         * platform/network/mac/NetworkStateNotifierMac.cpp:
3229         (WebCore::NetworkStateNotifier::updateStateWithoutNotifying): Renamed
3230         this from updateState and tightened up the code a bit. Also changed
3231         to not assume anything. If there is an error, we leave m_isOnLine alone.
3232         Default behavior is now in the platform-independent code, and the default
3233         is to treat things as on-line if updateStateWithoutNotifying was called
3234         and it was never able to update the state even once.
3235         (WebCore::NetworkStateNotifier::startObserving): Moved most of the code
3236         from the constructor into this function.
3237         (WebCore::NetworkStateNotifier::dynamicStoreCallback): Deleted. This is
3238         now a lambda inside startObserving.
3239         (WebCore::NetworkStateNotifier::networkStateChangeTimerFired): Deleted.
3240         This function is now platform-independent and is named updateState.
3241
3242         * platform/network/win/NetworkStateNotifierWin.cpp:
3243         (WebCore::NetworkStateNotifier::updateStateWithoutNotifying): Renamed
3244         from updateState and tightened the code up a bit. Also changed
3245         to not assume anything. If there is an error, we leave m_isOnLine alone.
3246         Default behavior is now in the platform-independent code, and the default
3247         is to treat things as on-line if updateStateWithoutNotifying was called
3248         and it was never able to update the state even once.
3249         (WebCore::NetworkStateNotifier::addressChanged): Deleted.
3250         This function is now platform-independent and is named updateState.
3251         (WebCore::NetworkStateNotifier::addressChangeCallback): Renamed from
3252         addrChangeCallback. Simplified by using the singleton function instead
3253         of relying on the context pointer.
3254         (WebCore::NetworkStateNotifier::startObserving): Moved most of the code
3255         from the constructor here.
3256
3257         * platform/text/LocaleToScriptMappingDefault.cpp:
3258         (WebCore::scriptNameToCode): Use makeNeverDestroyed.
3259         (WebCore::localeToScriptCodeForFontSelection): Ditto.
3260
3261         * platform/text/ios/LocalizedDateCache.h: Moved include of
3262         NeverDestroyed.h from here ...
3263         * platform/text/ios/LocalizedDateCache.mm: ... to here.
3264
3265         * platform/wpe/RenderThemeWPE.cpp:
3266         (WebCore::RenderTheme::singleton): Use a theme that is never destroyed
3267         rather than Ref that is never destroyed.
3268
3269         * platform/wpe/RenderThemeWPE.h: Use friend NeverDestroyed rather than
3270         a create function since RenderTheme is no longer reference counted.
3271
3272         * rendering/RenderTheme.h: Made RenderTheme no longer reference counted.
3273
3274         * rendering/RenderThemeGtk.cpp:
3275         (WebCore::RenderThemeGtk::create): Deleted.
3276         (WebCore::RenderTheme::singleton): Use a theme that is never destroyed
3277         rather than Ref that is never destroyed.
3278
3279         * rendering/RenderThemeGtk.h: Use friend NeverDestroyed rather than
3280         a create function since RenderTheme is no longer reference counted.
3281         * rendering/RenderThemeIOS.h: Ditto.
3282
3283         * rendering/RenderThemeIOS.mm:
3284         (WebCore::RenderThemeIOS::create): Deleted.
3285         (WebCore::RenderTheme::singleton): Use a theme that is never destroyed
3286         rather than Ref that is never destroyed.
3287
3288         * rendering/RenderThemeMac.h: Use friend NeverDestroyed rather than
3289         a create function since RenderTheme is no longer reference counted.
3290         Also marked a lot of functions final rather than override and initialized
3291         data members in the class deifnition.
3292
3293         * rendering/RenderThemeMac.mm: Moved some SPI interfaces to the top of the
3294         file. They should really go into SPI headers.
3295         (-[WebCoreRenderThemeNotificationObserver init]): Removed the theme argument
3296         here because the theme is a singleton and we don't need a pointer to it.
3297         Add the observer here instead of doing it in the caller.
3298         (-[WebCoreRenderThemeNotificationObserver systemColorsDidChange:]): Call
3299         platformColorsDidChange through the singleton rather than using a pointer.
3300         (-[WebCoreTextFieldCell _coreUIDrawOptionsWithFrame:inView:includeFocus:]):
3301         Tightened up the code a bit by using CFAutorelease instead of using a
3302         combination of NSMakeCollectable, autorelease, and a typecast to accomplish
3303         the same thing.
3304         (WebCore::RenderTheme::singleton): Use a theme that is never destroyed
3305         rather than Ref that is never destroyed.
3306         (WebCore::RenderThemeMac::create): Deleted.
3307         (WebCore::RenderThemeMac::RenderThemeMac): Moved initialization into the
3308         class definition and observing into the WebCoreRenderThemeNotificationObserver
3309         init function.
3310         (WebCore::RenderThemeMac::~RenderThemeMac): Deleted. No need for this since
3311         the object will never be destroyed.
3312         (WebCore::RenderThemeMac::mediaControlsBase64StringForIconNameAndType):
3313         Removed an unnecessary conversion from C string literal to WTF::String and
3314         then to NSString. Instead, just use an NSString literal.
3315
3316         * rendering/RenderThemeWin.cpp:
3317         (WebCore::RenderThemeWin::create): Deleted.
3318         (WebCore::RenderTheme::singleton): Use a theme that is never destroyed
3319         rather than Ref that is never destroyed.
3320
3321         * rendering/RenderThemeWin.h: Use friend NeverDestroyed rather than
3322         a create function since RenderTheme is no longer reference counted.
3323
3324         * rendering/svg/RenderSVGResource.cpp:
3325         (WebCore::removeFromCacheAndInvalidateDependencies): Moved a global into
3326         the code where it is used.
3327
3328         * svg/SVGAnimatedLength.cpp: Removed unneeded include of NeverDestroyed.h.
3329
3330         * svg/SVGAnimationElement.cpp:
3331         (WebCore::SVGAnimationElement::isSupportedAttribute): Use const auto,
3332         makeNeverDestroyed, a lambda, and the new HashSet::add overload instead of
3333         an isEmpty check and lots of separate add function calls.
3334         * svg/SVGClipPathElement.cpp:
3335         (WebCore::SVGClipPathElement::isSupportedAttribute): Ditto.
3336         * svg/SVGComponentTransferFunctionElement.cpp:
3337         (WebCore::SVGComponentTransferFunctionElement::isSupportedAttribute): Ditto.
3338         * svg/SVGCursorElement.cpp:
3339         (WebCore::SVGCursorElement::isSupportedAttribute): Ditto.
3340
3341         * svg/SVGElement.cpp:
3342         (WebCore::createAttributeNameToCSSPropertyIDMap): Replaced the populate
3343         function with this create function.
3344         (WebCore::createAttributeNameToAnimatedPropertyTypeMap): Ditto.
3345         (WebCore::attributeNameToAnimatedPropertyTypeMap): Changed return type to
3346         const reference. Use const auto and makeNeverDestroyed instead of using
3347         isEmpty and a populate function.
3348         (WebCore::createCSSPropertyWithSVGDOMNameToAnimatedPropertyTypeMap):
3349         More of the same.
3350         (WebCore::cssPropertyWithSVGDOMNameToAnimatedPropertyTypeMap): Ditto.
3351         (WebCore::SVGElement::childShouldCreateRenderer): Use an array instead of
3352         a HashSet since we are checking against only 3 or 4 values.
3353         (WebCore::addQualifiedName): Deleted.
3354         (WebCore::SVGElement::animatableAttributeForName): Use const auto,
3355         makeNeverDestroyed, and a lambda to efficiently build the map in this function.
3356         (WebCore::SVGElement::cssPropertyIdForSVGAttributeName): Use const auto and
3357         makeNeverDestroyed instead of using isEmpty and a populate function.
3358
3359         * svg/SVGFilterElement.cpp:
3360         (WebCore::SVGFilterElement::isSupportedAttribute): Use const auto,
3361         makeNeverDestroyed, a lambda, and the new HashSet::add overload instead of
3362         an isEmpty check and lots of separate add function calls.
3363         * svg/SVGFilterPrimitiveStandardAttributes.cpp:
3364         (WebCore::SVGFilterPrimitiveStandardAttributes::isSupportedAttribute): Ditto.
3365         * svg/SVGForeignObjectElement.cpp:
3366         (WebCore::SVGForeignObjectElement::isSupportedAttribute): Ditto.
3367         * svg/SVGGElement.cpp:
3368         (WebCore::SVGGElement::isSupportedAttribute): Ditto.
3369         * svg/SVGGradientElement.cpp:
3370         (WebCore::SVGGradientElement::isSupportedAttribute): Ditto.
3371         * svg/SVGGraphicsElement.cpp:
3372         (WebCore::SVGGraphicsElement::isSupportedAttribute): Ditto.
3373         * svg/SVGImageElement.cpp:
3374         (WebCore::SVGImageElement::isSupportedAttribute): Ditto.
3375         * svg/SVGLangSpace.cpp:
3376         (WebCore::addWithAndWithoutXMLPrefix): Added helper. Uses xmlAtom instead
3377         of making our own NeverDestroyed local xmlPrefix.
3378         (WebCore::SVGLangSpace::addSupportedAttributes): Rewrote to use helper.
3379         * svg/SVGLineElement.cpp:
3380         (WebCore::SVGLineElement::isSupportedAttribute): Use const auto,
3381         makeNeverDestroyed, a lambda, and the new HashSet::add overload instead of
3382         an isEmpty check and lots of separate add function calls.
3383         * svg/SVGLinearGradientElement.cpp:
3384         (WebCore::SVGLinearGradientElement::isSupportedAttribute): Ditto.
3385         * svg/SVGMarkerElement.cpp:
3386         (WebCore::SVGMarkerElement::isSupportedAttribute): Ditto.
3387         * svg/SVGMaskElement.cpp:
3388         (WebCore::SVGMaskElement::isSupportedAttribute): Ditto.
3389         * svg/SVGPathElement.cpp:
3390         (WebCore::SVGPathElement::isSupportedAttribute): Ditto.
3391         * svg/SVGPatternElement.cpp:
3392         (WebCore::SVGPatternElement::isSupportedAttribute): Ditto.
3393         * svg/SVGRadialGradientElement.cpp:
3394         (WebCore::SVGRadialGradientElement::isSupportedAttribute): Ditto.
3395         * svg/SVGTextContentElement.cpp:
3396         (WebCore::SVGTextContentElement::isSupportedAttribute): Ditto.
3397         * svg/SVGTextPathElement.cpp:
3398         (WebCore::SVGTextPathElement::isSupportedAttribute): Ditto.
3399         * svg/animation/SVGSMILElement.cpp:
3400         (WebCore::SVGSMILElement::isSupportedAttribute): Ditto.
3401
3402         * testing/MockContentFilterSettings.h: Removed unneeded include of
3403         NeverDestroyed.h.
3404
3405         * testing/MockGamepadProvider.cpp: Moved include of
3406         NeverDestroyed.h here ...
3407         * testing/MockGamepadProvider.h: ... from here.
3408
3409         * testing/MockPreviewLoaderClient.h: Removed unneeded include of
3410         NeverDestroyed.h.
3411
3412         * workers/Worker.cpp:
3413         (WebCore::allWorkers): Changed this from a global to a function.
3414         (WebCore::Worker::networkStateChanged): Updated for above change and
3415         made this a member function.
3416         (WebCore::Worker::Worker): Updated for above change and
3417         change to NetworkStateNotifier.
3418         (WebCore::Worker::~Worker): Ditto.
3419         * workers/Worker.h: Made networkStateChanged a static member instead of
3420         a friend since that is a cleaner way to give it access to private member
3421         functions.
3422
3423         * xml/XPathFunctions.cpp:
3424         (WebCore::XPath::createFunctionMap): Use create instead of populate style.
3425         (WebCore::XPath::Function::create): Use const auto and makeNeverDestroyed
3426         instead of an isEmpty check.
3427
3428         * xml/XPathParser.cpp:
3429         (WebCore::XPath::createAxisNamesMap): Use create instead of populate style.
3430         (WebCore::XPath::parseAxisName): Use const auto and makeNeverDestroyed
3431         instead of an isEmpty check.
3432
3433 2017-07-17  Brady Eidson  <beidson@apple.com>
3434
3435         REGRESSION(r219298): imported/w3c/IndexedDB-private-browsing/idbfactory_open.html is crashing occassionaly (UniqueIDBDatabase being taken from the IDBServer set twice).
3436         <rdar://problem/33294987> and https://bugs.webkit.org/show_bug.cgi?id=174354
3437
3438         Reviewed by Alex Christensen.
3439
3440         No new tests (Covered by existing tests).
3441
3442         * Modules/indexeddb/server/IDBServer.cpp:
3443         (WebCore::IDBServer::IDBServer::postDatabaseTaskReply): Remove a now invalid ASSERT
3444
3445         * Modules/indexeddb/server/UniqueIDBDatabase.cpp:
3446         (WebCore::IDBServer::UniqueIDBDatabase::scheduleShutdownForClose): Add a RELEASE_ASSERT.
3447         (WebCore::IDBServer::UniqueIDBDatabase::didDeleteBackingStore): Instead of an ad-hoc main thread dispatch, use the "schedule task reply" system
3448           to keep dispatch ordering in tact.
3449         (WebCore::IDBServer::UniqueIDBDatabase::postDatabaseTaskReply): Remove a now invalid ASSERT
3450         (WebCore::IDBServer::UniqueIDBDatabase::immediateCloseForUserDelete): Only take the owning pointer if the object doesn't already own itself.
3451
3452 2017-07-17  Wenson Hsieh  <wenson_hsieh@apple.com>
3453
3454         [iOS DnD] Web process uses too much memory when beginning a drag on a very large image
3455         https://bugs.webkit.org/show_bug.cgi?id=174585
3456         <rdar://problem/33302541>
3457
3458         Reviewed by Tim Horton.
3459
3460         Currently, attempting to drag a very large image fails, either due to us telling CoreGraphics to create an image
3461         buffer that is too large, or because the web process exceeds its memory limit and gets jetsamed. There are two
3462         places where we can optimize our memory use during the drag initialization sequence, and this patch improves
3463         both.
3464
3465         First, on iOS, we attempt to encode and send over a WebCore::Image in the PasteboardImage when writing to the
3466         item providers upon starting a drag. Currently, this Image is only used in the drag and drop codepath, in
3467         PlatformPasteboard::writeObjectRepresentations, to grab the size of the image being written for the purpose of
3468         specifying estimated display size. Serializing and deserializing an Image calls into Image::nativeImage, which
3469         attempts to draw the contents of the image into a buffer so that it can be shipped across to the UI process.
3470         Instead, we can simply compute the size in the web process while we already have the Image, and simply send that
3471         across. For copy/paste, this doesn't result in any behavior change, since we don't use the PasteboardImage's
3472         image in the first place.
3473
3474         Secondly, when starting a drag, we try to allocate create an image buffer the size of the WebCore::Image for the
3475         purpose of generating the drag preview. Instead, this patch establishes a limit on the size of this drag preview
3476         image, such that if the Image's size is larger, we'll scale down the drag preview image to be the maximum
3477         allowed size.
3478
3479         Test: DataInteractionTests.CanStartDragOnEnormousImage.
3480
3481         * editing/ios/EditorIOS.mm:
3482         (WebCore::Editor::writeImageToPasteboard):
3483         * platform/Pasteboard.h:
3484         * platform/graphics/GeometryUtilities.cpp:
3485         (WebCore::sizeWithAreaAndAspectRatio):
3486
3487         Introduce a new helper function to compute a size with the given aspect ratio and area.
3488
3489         * platform/graphics/GeometryUtilities.h:
3490         * platform/ios/DragImageIOS.mm:
3491         (WebCore::createDragImageFromImage):
3492         * platform/ios/PlatformPasteboardIOS.mm:
3493         (WebCore::PlatformPasteboard::writeObjectRepresentations):
3494
3495 2017-07-17  Michael Catanzaro  <mcatanzaro@igalia.com>
3496
3497         [CMake] Include most CMake modules from WebKitCommon.cmake
3498         https://bugs.webkit.org/show_bug.cgi?id=174546
3499
3500         Reviewed by Konstantin Tokarev.
3501
3502         * WebCoreMacros.cmake:
3503
3504 2017-07-17  Michael Catanzaro  <mcatanzaro@igalia.com>
3505
3506         [CMake] Macros in WebKitMacros.cmake should be prefixed with WEBKIT_ namespace
3507         https://bugs.webkit.org/show_bug.cgi?id=174547
3508
3509         Reviewed by Alex Christensen.
3510
3511         * CMakeLists.txt:
3512         * WebCoreMacros.cmake:
3513
3514 2017-07-17  Alex Christensen  <achristensen@webkit.org>
3515
3516         Modernize content extension code
3517         https://bugs.webkit.org/show_bug.cgi?id=174588
3518
3519         Reviewed by Sam Weinig.
3520
3521         No change in behavior.  Just use Ref instead of RefPtr where possible.
3522
3523         * contentextensions/ContentExtension.cpp:
3524         (WebCore::ContentExtensions::ContentExtension::create):
3525         * contentextensions/ContentExtension.h:
3526         * contentextensions/ContentExtensionsBackend.cpp:
3527         (WebCore::ContentExtensions::ContentExtensionsBackend::addContentExtension):
3528         (WebCore::ContentExtensions::ContentExtensionsBackend::actionsForResourceLoad):
3529         * contentextensions/ContentExtensionsBackend.h:
3530
3531 2017-07-17  Simon Fraser  <simon.fraser@apple.com>
3532
3533         clientX/clientY on TouchEvent.touches are wrong
3534         https://bugs.webkit.org/show_bug.cgi?id=174561
3535         rdar://problem/33336041
3536
3537         Reviewed by Tim Horton.
3538
3539         Do some refactoring so that WebKitAdditions code that computes Touch coordinates can use
3540         the same code that MouseRelatedEvent uses.
3541
3542         There is no behavior change in this patch, but the test exercises a behavior change in
3543         WebKitAdditions code.
3544
3545         Test: fast/events/touch/ios/touches-client-coords-after-zoom.html
3546
3547         * dom/MouseRelatedEvent.cpp:
3548         (WebCore::MouseRelatedEvent::init):
3549         (WebCore::MouseRelatedEvent::frameViewFromDOMWindow):
3550         (WebCore::MouseRelatedEvent::pagePointToClientPoint):
3551         (WebCore::MouseRelatedEvent::pagePointToAbsolutePoint):
3552         (WebCore::MouseRelatedEvent::initCoordinates):
3553         (WebCore::MouseRelatedEvent::documentToAbsoluteScaleFactor):
3554         (WebCore::MouseRelatedEvent::computePageLocation):
3555         (WebCore::MouseRelatedEvent::locationInRootViewCoordinates):
3556         (WebCore::MouseRelatedEvent::frameView): Deleted.
3557         * dom/MouseRelatedEvent.h:
3558
3559 2017-07-17  Jeremy Jones  <jeremyj@apple.com>
3560
3561         Add video fullscreen transition logging.
3562         https://bugs.webkit.org/show_bug.cgi?id=174474
3563
3564         Reviewed by Jer Noble.
3565
3566         No new tests, because there is no functional change.
3567
3568         This change just adds logging.
3569
3570         * html/HTMLVideoElement.cpp:
3571         (WebCore::HTMLVideoElement::webkitEnterFullscreen):
3572         (WebCore::HTMLVideoElement::webkitExitFullscreen):
3573         (WebCore::HTMLVideoElement::webkitSetPresentationMode):
3574
3575 2017-07-17  Chris Dumez  <cdumez@apple.com>
3576
3577         click event does not dispatch to parent when child target stops hit testing after mousedown
3578         https://bugs.webkit.org/show_bug.cgi?id=174564
3579         <rdar://problem/33340234>
3580
3581         Reviewed by Simon Fraser.
3582
3583         As per [1], if the mouse down node and the mouse release node differ, then we are supposed to
3584         fire the click event at their common ancestor, if such node exists. This patch implements this
3585         logic. This also aligns our behavior with Blink.
3586
3587         [1] https://w3c.github.io/uievents/#events-mouseevent-event-order
3588
3589         Test: fast/events/mouse-click-different-mouseDown-mouseUp-nodes.html
3590
3591         * page/EventHandler.cpp:
3592         (WebCore::targetNodeForClickEvent):
3593
3594 2017-07-17  Brady Eidson  <beidson@apple.com>
3595
3596         WKHTTPCookieStore observing only works on the default cookie store.
3597         <rdar://problem/33330724> and https://bugs.webkit.org/show_bug.cgi?id=174580
3598
3599         Reviewed by Sam Weinig.
3600
3601         Covered by new API tests.
3602
3603         startObservingCookieChanges and stopObservingCookieChanges are passed a NetworkStorageSession to observe.
3604         On Mac/iOS, the passed-in storage session was ignored and the shared cookie storage was assumed.
3605         Let's fix that.
3606
3607         Also, since using NSNotification based observing only works reliably for the shared cookie storage,
3608         switch to direct CFHTTPCookieStorageRef observing.
3609
3610         * WebCore.xcodeproj/project.pbxproj:
3611
3612         * platform/network/NetworkStorageSession.h:
3613
3614         * platform/network/cocoa/CookieStorageObserver.h: Added.
3615         * platform/network/cocoa/CookieStorageObserver.mm: Added.
3616         (WebCore::cookiesChanged):
3617         (WebCore::CookieStorageObserver::create):
3618         (WebCore::CookieStorageObserver::CookieStorageObserver):
3619         (WebCore::CookieStorageObserver::~CookieStorageObserver):
3620         (WebCore::CookieStorageObserver::startObserving):
3621         (WebCore::CookieStorageObserver::stopObserving):
3622         (WebCore::CookieStorageObserver::cookiesDidChange):
3623
3624         * platform/network/cocoa/NetworkStorageSessionCocoa.mm:
3625         (WebCore::NetworkStorageSession::cookieStorageObserver):
3626
3627         * platform/network/mac/CookieStorageMac.mm:
3628         (WebCore::startObservingCookieChanges):
3629         (WebCore::stopObservingCookieChanges):
3630         (-[WebCookieStorageObjCAdapter notifyCookiesChangedOnMainThread]): Deleted.
3631         (-[WebCookieStorageObjCAdapter cookiesChangedNotificationHandler:]): Deleted.
3632         (-[WebCookieStorageObjCAdapter startListeningForCookieChangeNotificationsWithCallback:]): Deleted.
3633         (-[WebCookieStorageObjCAdapter stopListeningForCookieChangeNotifications]): Deleted.
3634
3635         * platform/spi/cf/CFNetworkSPI.h:
3636
3637 2017-07-17  Sam Weinig  <sam@webkit.org>
3638
3639         [WebIDL] Rename JSCSSValueCustom.cpp to JSDeprecatedCSSOMValueCustom.cpp to match the underlying class
3640         https://bugs.webkit.org/show_bug.cgi?id=174550
3641
3642         Reviewed by Brady Eidson.
3643
3644         * CMakeLists.txt:
3645         * WebCore.xcodeproj/project.pbxproj:
3646         * bindings/js/JSBindingsAllInOne.cpp:
3647         * bindings/js/JSCSSValueCustom.cpp: Removed.
3648         * bindings/js/JSDeprecatedCSSOMValueCustom.cpp: Renamed from bindings/js/JSCSSValueCustom.cpp.
3649         Rename file and update references.
3650
3651 2017-07-17  Antoine Quint  <graouts@apple.com>
3652
3653         REGRESSION: order of AirPlay and volume controls is inconsistent between <audio> and <video>
3654         https://bugs.webkit.org/show_bug.cgi?id=174581
3655         <rdar://problem/33297519>
3656
3657         Reviewed by Sam Weinig.
3658
3659         We had an inconsistency between <audio> and <video> controls for the relative order of the
3660         volume and AirPlay buttons. The <video> layout was correct (volume first and AirPlay after)
3661         and the <audio> layout now is the same.
3662
3663         * Modules/modern-media-controls/controls/inline-media-controls.js:
3664         (InlineMediaControls.prototype._rightContainerButtons):
3665
3666 2017-07-17  Konstantin Tokarev  <annulen@yandex.ru>
3667
3668         [CMake] Create targets before WEBKIT_INCLUDE_CONFIG_FILES_IF_EXISTS is called
3669         https://bugs.webkit.org/show_bug.cgi?id=174557
3670
3671         Reviewed by Michael Catanzaro.
3672
3673         No new tests needed.
3674
3675         * CMakeLists.txt:
3676
3677 2017-07-17  Michael Catanzaro  <mcatanzaro@igalia.com>
3678
3679         Unreviewed, rolling out r219556.
3680
3681         Broke build without WebCrypto
3682
3683         Reverted changeset:
3684
3685         "[CMake] Clean up Web Crypto build targets"
3686         https://bugs.webkit.org/show_bug.cgi?id=174253
3687         http://trac.webkit.org/changeset/219556
3688
3689 2017-07-17  Antoine Quint  <graouts@apple.com>
3690
3691         Media controls draw behind captions
3692         https://bugs.webkit.org/show_bug.cgi?id=174579
3693         <rdar://problem/33295427>
3694
3695         Reviewed by Dean Jackson.
3696
3697         Ensure the captions container is added as a previous sibling to the controls container.
3698
3699         Test: media/modern-media-controls/media-controls/media-controls-display-above-captions.html
3700
3701         * Modules/modern-media-controls/media/media-controller.js:
3702         (MediaController):
3703
3704 2017-07-17  Zan Dobersek  <zdobersek@igalia.com>
3705
3706         [CMake] Clean up Web Crypto build targets
3707         https://bugs.webkit.org/show_bug.cgi?id=174253
3708
3709         Reviewed by Michael Catanzaro.
3710
3711         Gather the common WebCrypto source files in CMakeLists.txt, including them
3712         in the build unconditionally and instead relying on ENABLE_SUBTLE_CRYPTO
3713         build guards to exclude the code from compilation if the feature is disabled.
3714
3715         PlatformGTK.cmake, PlatformMac.cmake and PlatformWPE.cmake can then remove
3716         duplicated build targets. PlatformMac.cmake still lists all the Mac-specific
3717         Web Crypto build targets.
3718
3719         PlatformGTK.cmake and PlatformWPE.cmake now include GCrypt.cmake if the build
3720         was configured to enable the use of libgcrypt. The new CMake file adds the
3721         libgcrypt-specific Web Crypto build targets to the build if the feature was
3722         enabled, and also sets up libgcrypt include directiories and libraries.
3723
3724         No new tests -- no change in behavior.
3725
3726         * CMakeLists.txt:
3727         * PlatformGTK.cmake:
3728         * PlatformMac.cmake:
3729         * PlatformWPE.cmake:
3730         * platform/GCrypt.cmake: Added.
3731
3732 2017-07-16  Michael Catanzaro  <mcatanzaro@igalia.com>
3733
3734         -Wreorder warning caused by GraphicsContext3D cleanup
3735         https://bugs.webkit.org/show_bug.cgi?id=174511
3736
3737         Reviewed by Carlos Garcia Campos.
3738
3739         * platform/graphics/cairo/GraphicsContext3DCairo.cpp:
3740         (WebCore::GraphicsContext3D::GraphicsContext3D):
3741         * platform/graphics/opengl/GraphicsContext3DOpenGLES.cpp:
3742         (WebCore::GraphicsContext3D::GraphicsContext3D):
3743
3744 2017-07-16  Antoine Quint  <graouts@apple.com>
3745
3746         Dismissing the captions panel using the mouse is too eager to remove the captions panel and media controls
3747         https://bugs.webkit.org/show_bug.cgi?id=174571
3748         <rdar://problem/33294968>
3749
3750         Reviewed by Eric Carlson.
3751
3752         We did several things wrong when dismissing the tracks panel:
3753
3754             - we did not check whether we were hosted in a shadow root when figuring if a click was on the tracks panel
3755             - we did not check whether we clicked over the media when dismissing the tracks panel
3756             - we did not check whether auto-hide was on before fading the media controls out when we clicked outside
3757               the media controls bounds
3758
3759         We now correctly account for all of those cases and implement the following behavior when clickng as the tracks
3760         panel is presented:
3761
3762             - dismiss the panel if the click is outside of the panel
3763             - dismiss the panel and the media controls if the click is outside the video and the media controls have
3764               auto-hide on (ie. media is playing)
3765             - dismiss the panel and the media controls after the track selection animation is finished if a track is selected
3766
3767         Tests: media/modern-media-controls/tracks-panel/tracks-panel-up-click-outside-media-does-not-dimiss-media-controls-when-media-is-paused.html
3768                media/modern-media-controls/tracks-panel/tracks-panel-up-click-over-media-does-not-dimiss-media-controls-when-media-is-playing.html
3769
3770         * Modules/modern-media-controls/controls/media-controls.js:
3771         (MediaControls.prototype.hideTracksPanel): Only hide the media controls if we clicked outside of the media
3772         controls bounds and if we have auto-hide on when idle (ie. the media is playing).
3773         (MediaControls.prototype.isPointInControls): Add an option to specify whether the container should be
3774         considered when checking if a point is contained within the media controls bounds.
3775         * Modules/modern-media-controls/controls/tracks-panel.js:
3776         (TracksPanel.prototype._handleMousedown):
3777         (TracksPanel.prototype._isPointInTracksPanel): Correctly check whether the element that we started pressing
3778         on is contained within the tracks panel, accounting for the case where we are presented within a shadow root
3779         (ie. always when runing inside a Web page).
3780
3781 2017-07-16  Ali Juma  <ajuma@chromium.org>
3782
3783         DisallowUserAgentShadowContent moves out of non-UA shadow roots
3784         https://bugs.webkit.org/show_bug.cgi?id=165647
3785
3786         Reviewed by Ryosuke Niwa.
3787
3788         Make rect-based hit-testing include nodes in non-UA shadow trees when the
3789         HitTestRequest has type DisallowUserAgentShadowContent.
3790
3791         Test: fast/dom/nodesFromRect/nodesFromRect-shadow.html
3792
3793         * rendering/HitTestResult.cpp:
3794         (WebCore::HitTestResult::addNodeToRectBasedTestResult):
3795         * testing/Internals.cpp:
3796         (WebCore::Internals::nodesFromRect):
3797         * testing/Internals.h:
3798
3799 2017-07-16  Antoine Quint  <graouts@apple.com>
3800
3801         Clicking edges of media control buttons changes visual state of button (pressed) but doesn't execute action
3802         https://bugs.webkit.org/show_bug.cgi?id=174565
3803         <rdar://problem/33294833>
3804
3805         Reviewed by Dean Jackson.
3806
3807         WebCore doesn't dispatch a "click" event to a parent element when a child that was the original target when
3808         the "mousedown" occured is no longer hit-testing at the location where the mouse pointer is at when the "mouseup"
3809         occurs (see webkit.org/b/174564). Since button icons, which are a <picture> element that is a child of the
3810         <button> element for media controls buttons, shrink to 89% of their size when the ":active" pseudo-class matches,
3811         clicking on the edges of the media controls buttons would not trigger the expected action.
3812
3813         Test: media/modern-media-controls/button/button-click-on-edges.html
3814
3815         * Modules/modern-media-controls/controls/button.css:
3816         (button > picture):
3817
3818 2017-07-16  Said Abou-Hallawa  <sabouhallawa@apple.com>
3819
3820         Make the decision for asynchronously decoding an image be in one place
3821         https://bugs.webkit.org/show_bug.cgi?id=174479
3822
3823         Reviewed by Tim Horton.
3824
3825         Move all the logic of whether a large image should be asynchronously decoded
3826         or not be in one place: RenderBoxModelObject::decodingModeForImageDraw().
3827
3828         * loader/cache/CachedImage.cpp:
3829         (WebCore::CachedImage::addPendingImageDrawingClient): Fixing unrelated
3830         spelling error.
3831         * platform/RuntimeApplicationChecks.h:
3832         * platform/cocoa/RuntimeApplicationChecksCocoa.mm:
3833         (WebCore::IOSApplication::isIBooks):
3834         (WebCore::IOSApplication::isIBooksStorytime):
3835         * platform/graphics/BitmapImage.cpp:
3836         (WebCore::BitmapImage::updateFromSettings): Remove reading the setting
3837         largeImageAsyncDecodingEnabled from this function because it will be read
3838         by RenderBoxModelObject::decodingModeForImageDraw().
3839         (WebCore::BitmapImage::dataChanged):
3840         (WebCore::BitmapImage::draw):
3841         (WebCore::BitmapImage::shouldAnimate):
3842         (WebCore::BitmapImage::canAnimate):
3843         (WebCore::BitmapImage::canUseAsyncDecodingForLargeImages):
3844         (WebCore::BitmapImage::shouldUseAsyncDecodingForAnimatedImages):
3845         (WebCore::BitmapImage::canDestroyDecodedData):
3846         (WebCore::BitmapImage::shouldUseAsyncDecodingForLargeImages): Deleted.
3847         * platform/graphics/BitmapImage.h:
3848         * platform/graphics/ImageSource.cpp:
3849         (WebCore::ImageSource::canUseAsyncDecoding): It is okay to keep the
3850         decoded frame if canUseAsyncDecodingForLargeImages() is true by the setting
3851         largeImageAsyncDecodingEnabled is false.
3852         (WebCore::ImageSource::shouldUseAsyncDecoding): Deleted.
3853         * platform/graphics/ImageSource.h:
3854         * rendering/RenderBoxModelObject.cpp:
3855         (WebCore::RenderBoxModelObject::decodingModeForImageDraw): The plan is to
3856         add a new Internal settings to force asynchronous image decoding regardless
3857         of the image size and the settings.
3858         (WebCore::RenderBoxModelObject::paintFillLayerExtended):
3859         * rendering/RenderBoxModelObject.h:
3860         * rendering/RenderImage.cpp:
3861         (WebCore::RenderImage::paintIntoRect):
3862
3863 2017-07-16  Michael Catanzaro  <mcatanzaro@igalia.com>
3864
3865         [CMake] Raise minimum CMake requirement
3866         https://bugs.webkit.org/show_bug.cgi?id=174545
3867
3868         Reviewed by Konstantin Tokarev.
3869
3870         * WebCoreMacros.cmake:
3871
3872 2017-07-15  Brady Eidson  <beidson@apple.com>
3873
3874         Make sure all CFHTTPCookieStorageRefs we create are scheduled.
3875         <rdar://problem/33221110> and https://bugs.webkit.org/show_bug.cgi?id=174513
3876
3877         Reviewed by Tim Horton.
3878
3879         * platform/spi/cf/CFNetworkSPI.h:
3880
3881 2017-07-15  Myles C. Maxfield  <mmaxfield@apple.com>
3882
3883         Rename RenderStyle::fontSize() to RenderStyle::computedFontPixelSize()
3884         https://bugs.webkit.org/show_bug.cgi?id=174509
3885
3886         Reviewed by Simon Fraser.
3887
3888         We have three font size functions:
3889         - computedFontSize(): returns a float
3890         - specifiedFontSize(): also returns a float
3891         - fontSize(): returns the rounded computedFontSize()
3892
3893         FontDescription uses the convention of labelling rounded values as "pixel",
3894         so a better name font fontSize() is computedFontPixelSize().
3895
3896         Also, because font sizes can never be negative, switch the type from an int
3897         to an unsigned.
3898
3899         No new tests because there is no behavior change.
3900
3901         * accessibility/atk/WebKitAccessibleInterfaceText.cpp:
3902         (getAttributeSetForAccessibilityObject):
3903         * css/StyleResolver.cpp:
3904         (WebCore::StyleResolver::adjustRenderStyle):
3905         * platform/graphics/FontDescription.h:
3906         (WebCore::FontDescription::computedPixelSize):
3907         (WebCore::FontDescription::fontSelectionRequest):
3908         * rendering/RenderBlock.cpp:
3909         (WebCore::styleForFirstLetter):
3910         * rendering/RenderRubyRun.cpp:
3911         (WebCore::shouldOverhang):
3912         (WebCore::RenderRubyRun::getOverhang):
3913         * rendering/RenderRubyText.cpp:
3914         (WebCore::RenderRubyText::adjustInlineDirectionLineBounds):
3915         * rendering/RenderThemeGtk.cpp:
3916         (WebCore::adjustSearchFieldIconStyle):
3917         (WebCore::paintSearchFieldIcon):
3918         * rendering/RenderThemeIOS.mm:
3919         (WebCore::RenderThemeIOS::adjustCheckboxStyle):
3920         (WebCore::RenderThemeIOS::adjustRadioStyle):
3921         * rendering/RenderThemeMac.mm:
3922         (WebCore::RenderThemeMac::controlSizeForFont):
3923         (WebCore::RenderThemeMac::controlSizeForSystemFont):
3924         (WebCore::RenderThemeMac::paintMenuListButtonDecorations):
3925         (WebCore::RenderThemeMac::popupInternalPaddingBox):
3926         (WebCore::RenderThemeMac::adjustMenuListButtonStyle):
3927         * rendering/RenderThemeWin.cpp:
3928         (WebCore::RenderThemeWin::adjustSearchFieldCancelButtonStyle):
3929         (WebCore::RenderThemeWin::adjustSearchFieldResultsDecorationPartStyle):
3930         (WebCore::RenderThemeWin::adjustSearchFieldResultsButtonStyle):
3931         * rendering/TextDecorationPainter.cpp:
3932         (WebCore::TextDecorationPainter::paintTextDecoration):
3933         * rendering/mathml/RenderMathMLRow.cpp:
3934         (WebCore::RenderMathMLRow::computeLineVerticalStretch):
3935         * rendering/style/RenderStyle.cpp:
3936         (WebCore::RenderStyle::computedFontPixelSize):
3937         (WebCore::RenderStyle::computedLineHeight):
3938         (WebCore::RenderStyle::fontSize): Deleted.
3939         * rendering/style/RenderStyle.h:
3940         * style/InlineTextBoxStyle.cpp:
3941         (WebCore::visualOverflowForDecorations):
3942         * style/StyleTreeResolver.cpp:
3943         (WebCore::Style::TreeResolver::resolveElement):
3944         * svg/SVGLengthContext.cpp:
3945         (WebCore::SVGLengthContext::convertValueFromUserUnitsToEMS):
3946