1 2019-07-25 Truitt Savell <tsavell@apple.com>
3 Unreviewed, rolling out r247821.
5 Caused two crashing Layout Tests
9 "Make storing cross-origin top-level prefetches in HTTP cache
11 https://bugs.webkit.org/show_bug.cgi?id=199499
12 https://trac.webkit.org/changeset/247821
14 2019-07-25 Chris Dumez <cdumez@apple.com>
16 Avoid UI Process hangs when the WebContent process is showing JS prompts / alerts
17 https://bugs.webkit.org/show_bug.cgi?id=200107
18 <rdar://problem/53034592>
20 Reviewed by Geoffrey Garen.
22 Add testing infrastructure for SendSyncOption::ProcessIncomingSyncMessagesWhenWaitingForSyncReply.
24 Test: fast/misc/testProcessIncomingSyncMessagesWhenWaitingForSyncReply.html
26 * page/ChromeClient.h:
27 * testing/Internals.cpp:
28 (WebCore::Internals::testProcessIncomingSyncMessagesWhenWaitingForSyncReply):
29 * testing/Internals.h:
30 * testing/Internals.idl:
32 2019-07-25 Rob Buis <rbuis@igalia.com>
34 Make storing cross-origin top-level prefetches in HTTP cache optional
35 https://bugs.webkit.org/show_bug.cgi?id=199499
37 Reviewed by Youenn Fablet.
39 Add Internals API for adding an event listener to prefetch load.
41 Test: http/wpt/prefetch/link-prefetch-skip-disk-cache.html
43 * bindings/js/JSEventListener.h:
45 (WebCore::EventTarget::addEventListener):
47 * html/HTMLLinkElement.h:
48 * testing/Internals.cpp:
49 (WebCore::Internals::addPrefetchLoadEventListener):
50 * testing/Internals.h:
51 * testing/Internals.idl:
53 2019-07-24 Yusuke Suzuki <ysuzuki@apple.com>
55 [bmalloc] Add IsoHeap test to ensure that IsoHeap pages are not allocating too large VA
56 https://bugs.webkit.org/show_bug.cgi?id=200103
60 No behavior change in WebCore.
62 * page/MemoryRelease.cpp:
63 * page/ResourceUsageThread.h:
64 (WebCore::TagInfo::TagInfo): Deleted.
65 * page/cocoa/ResourceUsageThreadCocoa.mm:
66 (WebCore::vmPageSize): Deleted.
67 (WebCore::logFootprintComparison): Deleted.
68 (WebCore::displayNameForVMTag): Deleted.
69 (WebCore::pagesPerVMTag): Deleted.
71 2019-07-24 Devin Rousso <drousso@apple.com>
73 Web Inspector: support `console.screenshot` with detached Nodes that are graphical
74 https://bugs.webkit.org/show_bug.cgi?id=199309
76 Reviewed by Joseph Pecoraro.
78 Test: inspector/console/console-screenshot.html
80 * page/PageConsoleClient.cpp:
81 (WebCore::PageConsoleClient::screenshot):
83 2019-07-24 Saam Barati <sbarati@apple.com>
85 [WHLSL] Use bmalloc instead of system malloc for WHLSL data structures
86 https://bugs.webkit.org/show_bug.cgi?id=200063
88 Rubber-stamped by Sam Weinig.
90 This is a 25-30% WHLSL compile time speedup on Justin's textured-cube demo
91 and a ~40% compile time speedup on the hello-cube demo.
93 The phases that speed up from this are:
95 includeStandardLibrary
96 autoInitializeVariables
99 * Modules/webgpu/WHLSL/AST/WHLSLAddressSpace.h:
100 * Modules/webgpu/WHLSL/AST/WHLSLArrayReferenceType.h:
101 * Modules/webgpu/WHLSL/AST/WHLSLArrayType.h:
102 * Modules/webgpu/WHLSL/AST/WHLSLAssignmentExpression.h:
103 * Modules/webgpu/WHLSL/AST/WHLSLBaseFunctionAttribute.h:
104 * Modules/webgpu/WHLSL/AST/WHLSLBaseSemantic.h:
105 * Modules/webgpu/WHLSL/AST/WHLSLBlock.h:
106 * Modules/webgpu/WHLSL/AST/WHLSLBooleanLiteral.h:
107 * Modules/webgpu/WHLSL/AST/WHLSLBreak.h:
108 * Modules/webgpu/WHLSL/AST/WHLSLBuiltInSemantic.h:
109 * Modules/webgpu/WHLSL/AST/WHLSLCallExpression.h:
110 * Modules/webgpu/WHLSL/AST/WHLSLCommaExpression.h:
111 * Modules/webgpu/WHLSL/AST/WHLSLConstantExpression.h:
112 * Modules/webgpu/WHLSL/AST/WHLSLContinue.h:
113 * Modules/webgpu/WHLSL/AST/WHLSLDereferenceExpression.h:
114 * Modules/webgpu/WHLSL/AST/WHLSLDoWhileLoop.h:
115 * Modules/webgpu/WHLSL/AST/WHLSLDotExpression.h:
116 * Modules/webgpu/WHLSL/AST/WHLSLEffectfulExpressionStatement.h:
117 * Modules/webgpu/WHLSL/AST/WHLSLEnumerationDefinition.h:
118 * Modules/webgpu/WHLSL/AST/WHLSLEnumerationMember.h:
119 * Modules/webgpu/WHLSL/AST/WHLSLEnumerationMemberLiteral.h:
120 * Modules/webgpu/WHLSL/AST/WHLSLExpression.h:
121 * Modules/webgpu/WHLSL/AST/WHLSLFallthrough.h:
122 * Modules/webgpu/WHLSL/AST/WHLSLFloatLiteral.h:
123 * Modules/webgpu/WHLSL/AST/WHLSLFloatLiteralType.h:
124 * Modules/webgpu/WHLSL/AST/WHLSLForLoop.h:
125 * Modules/webgpu/WHLSL/AST/WHLSLFunctionDeclaration.h:
126 * Modules/webgpu/WHLSL/AST/WHLSLFunctionDefinition.h:
127 * Modules/webgpu/WHLSL/AST/WHLSLGlobalVariableReference.h:
128 * Modules/webgpu/WHLSL/AST/WHLSLIfStatement.h:
129 * Modules/webgpu/WHLSL/AST/WHLSLIndexExpression.h:
130 * Modules/webgpu/WHLSL/AST/WHLSLIntegerLiteral.h:
131 * Modules/webgpu/WHLSL/AST/WHLSLIntegerLiteralType.h:
132 * Modules/webgpu/WHLSL/AST/WHLSLLogicalExpression.h:
133 * Modules/webgpu/WHLSL/AST/WHLSLLogicalNotExpression.h:
134 * Modules/webgpu/WHLSL/AST/WHLSLMakeArrayReferenceExpression.h:
135 * Modules/webgpu/WHLSL/AST/WHLSLMakePointerExpression.h:
136 * Modules/webgpu/WHLSL/AST/WHLSLNamedType.h:
137 * Modules/webgpu/WHLSL/AST/WHLSLNativeFunctionDeclaration.h:
138 * Modules/webgpu/WHLSL/AST/WHLSLNativeTypeDeclaration.h:
139 * Modules/webgpu/WHLSL/AST/WHLSLNullLiteral.h:
140 * Modules/webgpu/WHLSL/AST/WHLSLNullLiteralType.h:
141 * Modules/webgpu/WHLSL/AST/WHLSLNumThreadsFunctionAttribute.h:
142 * Modules/webgpu/WHLSL/AST/WHLSLPointerType.h:
143 * Modules/webgpu/WHLSL/AST/WHLSLPropertyAccessExpression.h:
144 * Modules/webgpu/WHLSL/AST/WHLSLReadModifyWriteExpression.h:
145 * Modules/webgpu/WHLSL/AST/WHLSLReferenceType.h:
146 * Modules/webgpu/WHLSL/AST/WHLSLResolvableType.h:
147 * Modules/webgpu/WHLSL/AST/WHLSLResourceSemantic.h:
148 * Modules/webgpu/WHLSL/AST/WHLSLReturn.h:
149 * Modules/webgpu/WHLSL/AST/WHLSLSpecializationConstantSemantic.h:
150 * Modules/webgpu/WHLSL/AST/WHLSLStageInOutSemantic.h:
151 * Modules/webgpu/WHLSL/AST/WHLSLStatement.h:
152 * Modules/webgpu/WHLSL/AST/WHLSLStatementList.h:
153 * Modules/webgpu/WHLSL/AST/WHLSLStructureDefinition.h:
154 * Modules/webgpu/WHLSL/AST/WHLSLStructureElement.h:
155 * Modules/webgpu/WHLSL/AST/WHLSLSwitchCase.h:
156 * Modules/webgpu/WHLSL/AST/WHLSLSwitchStatement.h:
157 * Modules/webgpu/WHLSL/AST/WHLSLTernaryExpression.h:
158 * Modules/webgpu/WHLSL/AST/WHLSLType.h:
159 * Modules/webgpu/WHLSL/AST/WHLSLTypeDefinition.h:
160 * Modules/webgpu/WHLSL/AST/WHLSLTypeReference.h:
161 * Modules/webgpu/WHLSL/AST/WHLSLUnnamedType.h:
162 * Modules/webgpu/WHLSL/AST/WHLSLUnsignedIntegerLiteral.h:
163 * Modules/webgpu/WHLSL/AST/WHLSLUnsignedIntegerLiteralType.h:
164 * Modules/webgpu/WHLSL/AST/WHLSLVariableDeclaration.h:
165 * Modules/webgpu/WHLSL/AST/WHLSLVariableDeclarationsStatement.h:
166 * Modules/webgpu/WHLSL/AST/WHLSLVariableReference.h:
167 * Modules/webgpu/WHLSL/AST/WHLSLWhileLoop.h:
168 * Modules/webgpu/WHLSL/Metal/WHLSLTypeNamer.cpp:
170 2019-07-24 Devin Rousso <drousso@apple.com>
172 Web Inspector: support `console.screenshot` with ImageData and ImageBitmap
173 https://bugs.webkit.org/show_bug.cgi?id=199304
175 Reviewed by Joseph Pecoraro.
177 Test: inspector/console/console-screenshot.html
179 * page/PageConsoleClient.cpp:
180 (WebCore::PageConsoleClient::screenshot):
182 2019-07-24 Devin Rousso <drousso@apple.com>
184 Web Inspector: REGRESSION: Overlay: highlighting a ::before/::after shows the node info of the parent instead
185 https://bugs.webkit.org/show_bug.cgi?id=200051
187 Reviewed by Joseph Pecoraro.
189 When highlighting a pseudo-element, only use the parent node for selector information. All
190 rendering data should still come from the pseudo-element itself.
192 * inspector/InspectorOverlay.cpp:
193 (WebCore::drawShapeHighlight):
194 (WebCore::InspectorOverlay::drawElementTitle):
196 2019-07-24 Devin Rousso <drousso@apple.com>
198 Web Inspector: support `console.screenshot` with subclasses of CanvasRenderingContext
199 https://bugs.webkit.org/show_bug.cgi?id=199305
201 Reviewed by Joseph Pecoraro.
203 Test: inspector/console/console-screenshot.html
205 * page/PageConsoleClient.cpp:
206 (WebCore::canvasRenderingContext):
207 (WebCore::PageConsoleClient::record):
208 (WebCore::PageConsoleClient::recordEnd):
209 (WebCore::PageConsoleClient::screenshot):
211 2019-07-24 Youenn Fablet <youenn@apple.com>
213 Remove MediaPlayerPrivate::supportsMuting()
214 https://bugs.webkit.org/show_bug.cgi?id=200005
216 Reviewed by Eric Carlson.
218 Remove no longer needed interface method.
219 No change of behavior.
222 * platform/graphics/MediaPlayer.cpp:
223 (WebCore::MediaPlayer::setVolume):
224 (WebCore::MediaPlayer::setMuted):
225 * platform/graphics/MediaPlayerPrivate.h:
226 * platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.h:
227 * platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.h:
228 * platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaStreamAVFObjC.h:
229 * platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.h:
230 * platform/graphics/win/MediaPlayerPrivateMediaFoundation.cpp:
231 * platform/graphics/win/MediaPlayerPrivateMediaFoundation.h:
233 2019-07-24 Tim Horton <timothy_horton@apple.com>
235 Daring Fireball long press highlights are unnecessarily inflated due to false illegibility
236 https://bugs.webkit.org/show_bug.cgi?id=200064
238 Reviewed by Geoff Garen.
240 If we consider text illegible on the given estimated background color,
241 we bail from doing a tightly fitted selection-only TextIndicator and
242 instead just paint the page without modification into the indicator,
243 causing ugly overlap and an excessively inflated indicator.
245 Change the mechanism we use to determine illegibility to be based on
246 a standard, instead of a constant chosen by hand 13 years ago.
248 Test: fast/text-indicator/text-indicator-with-low-contrast-text.html
250 * platform/graphics/ColorUtilities.cpp:
251 (WebCore::luminance):
254 (WebCore::contrastRatio):
255 Add a function that computes the contrast ratio given two colors using
256 the formula from WCAG.
258 * platform/graphics/ColorUtilities.h:
259 * rendering/TextPaintStyle.cpp:
260 (WebCore::textColorIsLegibleAgainstBackgroundColor):
261 Make use of WCAG's minimum legible contrast ratio instead of an
262 arbitrary color difference cutoff for determining whether we consider
263 text legible. It seems sensible and also considers the text on DF readable
264 (which it seems to be to me!).
266 * testing/Internals.cpp:
267 (WebCore::Internals::TextIndicatorInfo::TextIndicatorInfo):
268 * testing/Internals.h:
269 * testing/Internals.idl:
270 Expose all of the text rects to Internals, not just the bounding rect.
271 Expose some more TextIndicator options to Internals so that we can
272 turn on the legibility mechanism.
274 2019-07-24 Devin Rousso <drousso@apple.com>
276 Web Inspector: print the target of `console.screenshot` last so the target is the closest item to the image
277 https://bugs.webkit.org/show_bug.cgi?id=199308
279 Reviewed by Joseph Pecoraro.
281 Right now, evaluating `console.screenshot(document, "test", 1);` will log a message to the
282 console with `#document`, `"test"`, and `1`, all on different lines (togglable by a
283 disclosure triangle) since `#document` isn't stringifiable.
285 The ideal situation would be to have `"test"` and `1` logged on the same line, and then have
286 `#document` be in a disclosure triangle. This way, you can "label" your images using
287 additional arguments (e.g. `console.screenshot(document.images[1], "second image");`), as
288 well as provide other data.
290 If the only argument was the target, it should print as if it was `console.log(target);`.
292 If there are no arguments, it should print the text "Viewport"` before the image.
294 Test: inspector/console/console-screenshot.html
296 * page/PageConsoleClient.cpp:
297 (WebCore::PageConsoleClient::addMessage):
298 (WebCore::PageConsoleClient::screenshot):
300 2019-07-24 Alicia Boya GarcÃa <aboya@igalia.com>
302 [MSE] Reenqueue after removeCodedFrames()
303 https://bugs.webkit.org/show_bug.cgi?id=199749
305 Reviewed by Xabier Rodriguez-Calvar.
307 Before this patch, SourceBuffer::removeCodedFrames() did not trigger
308 an immediate reenqueue, but rather just set the `needsReenqueuing`
309 flag, deferring it for the next append... but there may not be another
310 append! In that case, the removed frames would still wrongly play.
312 This is the case for instance in tests where a single long media
313 append is done and then "cropped" with SourceBuffer.erase().
315 Test: media/media-source/media-source-erase-after-last-append.html
317 * Modules/mediasource/SourceBuffer.cpp:
318 (WebCore::SourceBuffer::removeCodedFrames):
320 2019-07-24 Jer Noble <jer.noble@apple.com>
322 Add per-document autoplay quirk for netflix.com.
323 https://bugs.webkit.org/show_bug.cgi?id=200058
325 Reviewed by Eric Carlson.
327 Opt netflix.com into the desktop-style per-document autoplay quirk.
330 (WebCore::Quirks::needsPerDocumentAutoplayBehavior const):
332 2019-07-24 Alicia Boya GarcÃa <aboya@igalia.com>
334 [GStreamer] Don't crash with empty video src
335 https://bugs.webkit.org/show_bug.cgi?id=200081
337 When a <video> element is set to load empty or about:blank, a player is still
338 created, but no pipeline is loaded. This patch fixes some assertion errors that
339 manifested in that case.
341 Reviewed by Philippe Normand.
343 Test: imported/w3c/web-platform-tests/html/semantics/embedded-content/the-video-element/video_crash_empty_src.html
345 * platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
346 (WebCore::MediaPlayerPrivateGStreamer::loadFull):
347 (WebCore::MediaPlayerPrivateGStreamer::platformDuration const):
348 (WebCore::MediaPlayerPrivateGStreamer::paused const):
350 2019-07-23 Justin Fan <justin_fan@apple.com>
352 [WebGPU] Implement errors for GPURenderPipeline creation
353 https://bugs.webkit.org/show_bug.cgi?id=200046
355 Reviewed by Myles C. Maxfield.
357 Replacing error logging in GPURenderPipeline creation with GPUError generation.
358 Update GPUErrorScopes to re-use an error message prefix for less boiler-plate.
360 Test: webgpu/render-pipeline-errors.html
362 * Modules/webgpu/WebGPUDevice.cpp:
363 (WebCore::WebGPUDevice::createRenderPipeline const):
364 * Modules/webgpu/WebGPURenderPipelineDescriptor.cpp:
365 (WebCore::WebGPURenderPipelineDescriptor::tryCreateGPURenderPipelineDescriptor const):
366 * Modules/webgpu/WebGPURenderPipelineDescriptor.h:
367 * platform/graphics/gpu/GPUDevice.cpp:
368 (WebCore::GPUDevice::tryCreateRenderPipeline const):
369 * platform/graphics/gpu/GPUDevice.h:
370 * platform/graphics/gpu/GPUErrorScopes.cpp:
371 (WebCore::GPUErrorScopes::generatePrefixedError):
372 * platform/graphics/gpu/GPUErrorScopes.h:
373 (WebCore::GPUErrorScopes::setErrorPrefix):
374 * platform/graphics/gpu/GPURenderPipeline.h:
375 * platform/graphics/gpu/cocoa/GPURenderPipelineMetal.mm:
376 (WebCore::tryCreateMtlDepthStencilState):
377 (WebCore::trySetVertexInput):
378 (WebCore::trySetColorStates):
379 (WebCore::trySetMetalFunctions):
380 (WebCore::trySetFunctions):
381 (WebCore::convertRenderPipelineDescriptor):
382 (WebCore::tryCreateMtlRenderPipelineState):
383 (WebCore::GPURenderPipeline::tryCreate):
384 (WebCore::GPURenderPipeline::GPURenderPipeline):
386 2019-07-23 Konstantin Tokarev <annulen@yandex.ru>
388 ImageDecoderDirect2D::hotSpot() should return WTF::nullopt instead of default constructed value
389 https://bugs.webkit.org/show_bug.cgi?id=200055
391 Reviewed by Brent Fulgham.
393 * platform/graphics/win/ImageDecoderDirect2D.cpp:
394 (WebCore::ImageDecoderDirect2D::hotSpot const):
396 2019-07-23 Tim Horton <timothy_horton@apple.com>
398 Long press hint has half blue shape in Feedly.com
399 https://bugs.webkit.org/show_bug.cgi?id=200053
400 <rdar://problem/53323369>
402 Reviewed by Simon Fraser.
404 No new tests, because TextIndicator snapshots are not yet testable.
406 * rendering/RenderLayer.cpp:
407 (WebCore::RenderLayer::paintForegroundForFragments):
408 Paint ChildBlockBackgrounds when painting selection AND backgrounds;
409 only exclude it when doing a selection-only paint.
411 2019-07-23 Chris Fleizach <cfleizach@apple.com>
413 AX: CrashTracer: com.apple.WebKit.WebContent at WebKit: WebKit::WebSpeechSynthesisClient::speak
414 https://bugs.webkit.org/show_bug.cgi?id=199988
416 Reviewed by Per Arne Vollan.
418 Implement the reset state to cancel current speech jobs.
420 * Modules/speech/SpeechSynthesis.cpp:
421 (WebCore::SpeechSynthesis::startSpeakingImmediately):
422 (WebCore::SpeechSynthesis::cancel):
423 * platform/PlatformSpeechSynthesizer.h:
424 * platform/ios/PlatformSpeechSynthesizerIOS.mm:
425 (WebCore::PlatformSpeechSynthesizer::resetState):
426 * platform/mac/PlatformSpeechSynthesizerMac.mm:
427 (WebCore::PlatformSpeechSynthesizer::resetState):
429 2019-07-23 Zalan Bujtas <zalan@apple.com>
431 [iOS] Tapping the search field on a search results page on zillow.com shows and immediately dismisses the keyboard
432 https://bugs.webkit.org/show_bug.cgi?id=200044
433 <rdar://problem/53103732>
435 Reviewed by Wenson Hsieh.
437 1. The keyboard gets dismissed as the result of scroll event.
438 2. The (horizontal)scroll event is initiated by WebKit as we try to re-center the content.
439 3. The content gets off-centered as the result of the newly constructed drop-down menu which slightly sticks out of the document to the right (layout overflows).
441 It works with shipping version of iOS because _zoomToFocusRect operates on stale viewport information (see r244494 for the progression).
442 This patch applies a site specific quirk to restore shipping behavior.
445 (WebCore::Quirks::shouldAvoidScrollingWhenFocusedContentIsVisible const):
448 2019-07-23 Said Abou-Hallawa <sabouhallawa@apple.com>
450 CanvasRenderingContext2D.setTransfrom() reads only the aliases attributes of DOMMatrix2DInit
451 https://bugs.webkit.org/show_bug.cgi?id=199850
453 Reviewed by Simon Fraser.
455 The validateAndFixup step of DOMMatrix2DInit sets the matrix attributes
456 (m** fields) from the aliases attributes (a, b, c, d, e, f) if the former
457 ones do not exist and the later ones exist. See
458 https://www.w3.org/TR/geometry-1/#dommatrixinit-dictionary and
459 DOMMatrixReadOnly::validateAndFixup(DOMMatrix2DInit& init).
461 Since CanvasRenderingContext2DBase::setTransform() validates and fixes
462 the input DOMMatrix2DInit, it should rely on matrix attributes.
464 Test: fast/canvas/setTransfrom-aliases-transform.html
466 * html/canvas/CanvasRenderingContext2DBase.cpp:
467 (WebCore::CanvasRenderingContext2DBase::setTransform):
469 2019-07-23 Tim Horton <timothy_horton@apple.com>
471 Long press hint of AirPods buy buttons are tall and narrow during animation
472 https://bugs.webkit.org/show_bug.cgi?id=200036
473 <rdar://problem/53145697>
475 Reviewed by Wenson Hsieh.
477 New test: fast/text-indicator/text-indicator-with-tiny-child.html
480 (WebCore::Range::borderAndTextRects const):
482 Add a BoundingRectBehavior that ignores 1x1 and smaller rects.
484 * page/TextIndicator.cpp:
485 (WebCore::absoluteBoundingRectForRange):
486 Enable IgnoreTinyRects.
488 2019-07-23 Wenson Hsieh <wenson_hsieh@apple.com>
490 [macOS 10.15] Web process crashes when attempting to show the font panel via Font > Show Fonts
491 https://bugs.webkit.org/show_bug.cgi?id=200021
492 <rdar://problem/53301325>
494 Reviewed by Ryosuke Niwa.
496 In macOS 10.15, attempting to access ~/Library/FontCollections/ from the web process causes a Sandbox violation
497 and a subsequent crash. This can be triggered by trying to show the font panel via Font > Show Fonts in the
498 context menu after right clicking.
500 In the (near) future, we should fix this by moving logic to show the font panel, color panel and styles panel
501 from the Editor in the web process to the UI process in WebKit2. However, for the time being, we can just work
502 around this by adding a new EditorClient hook to determine whether to allow the font panel to show; this returns
503 true in the legacy macOS WebKit port, and false everywhere else.
505 Test: editing/mac/style/do-not-crash-when-showing-font-panel.html
507 * editing/mac/EditorMac.mm:
508 (WebCore::Editor::showFontPanel):
509 * loader/EmptyClients.cpp:
510 * page/EditorClient.h:
512 2019-07-23 Ryosuke Niwa <rniwa@webkit.org>
514 WebKit should strip away system font names from the pasted content
515 https://bugs.webkit.org/show_bug.cgi?id=199975
516 <rdar://problem/53336353>
518 Reviewed by Darin Adler.
520 Cocoa HTML Writer sometimes generate system font names such as ".AppleSystemUIFont", ".SFUI-Regular", and ".SF UI Mono".
521 We need to strip away these font names upon paste to avoid these font names falling back to Times New Roman.
523 Added the code to strip these font names away in EditingStyle::mergeStyleFromRulesForSerialization, which is used by
524 StylizedMarkupAccumulator to generate HTML during copy. This works because WebContentReader::readWebArchive invokes
525 sanitizeMarkupWithArchive which inserts the pasteboard content into a temporary document then re-serializes back to HTML
526 using StylizedMarkupAccumulator before the actual pasting happens.
528 This approach has a few benefits over stripping away these font names in ReplaceSelectionCommand:
530 1. It would only affect clients that opts-in to copy & paste sanitization. e.g. it won't affect legacy WebKit clients
531 and those that opt out of pasteboard content sanitization.
533 2. It preserves font names such as ".SF Blah" that a website may insert as some kind of house keeping purposes if ever.
534 While we don't have any evidence that there is any such a website but it's a real risk nonetheless. The copy side fix would
535 only affect cross-site and cross-app pasting, which is rare and less likely to affect real user scenarios.
537 3. It avoids exposing bogus .Apple* or .SF* font names to websites that directly use event.clipboardData.getData.
538 Indeed stripping away bogus markup like this is one of the key features / benefit of using copy & paste sanitization.
540 Test: editing/pasteboard/paste-cocoa-writer-markup-with-system-fonts.html
542 * editing/EditingStyle.cpp:
543 (WebCore::usesForbiddenSystemFontAsOnlyFontFamilyName): Added.
544 (WebCore::EditingStyle::mergeStyleFromRulesForSerialization): Added the code to remove font-family property when needed.
545 * platform/graphics/FontCache.h:
546 * platform/graphics/cocoa/FontCacheCoreText.cpp:
547 (WebCore::isSystemFont): Moved.
548 (WebCore::FontCache::isSystemFontForbiddenForEditing): Added.
549 * platform/graphics/freetype/FontCacheFreeType.cpp:
550 (WebCore::FontCache::isSystemFontForbiddenForEditing): Added. Always returns false.
551 * platform/graphics/win/FontCacheWin.cpp:
552 (WebCore::FontCache::isSystemFontForbiddenForEditing): Ditto.
554 2019-07-22 Yusuke Suzuki <ysuzuki@apple.com>
556 [bmalloc] Each IsoPage gets 1MB VA because VMHeap::tryAllocateLargeChunk rounds up
557 https://bugs.webkit.org/show_bug.cgi?id=200024
559 Reviewed by Saam Barati.
561 Changed how we interpret VM tags. Add IsoHeap VM tag support, and rename WebAssembly tag
564 * page/ResourceUsageData.h:
565 * page/ResourceUsageOverlay.h:
566 * page/cocoa/ResourceUsageOverlayCocoa.mm:
567 (WebCore::HistoricResourceUsageData::HistoricResourceUsageData):
568 * page/cocoa/ResourceUsageThreadCocoa.mm:
569 (WebCore::displayNameForVMTag):
570 (WebCore::categoryForVMTag):
572 2019-07-22 Youenn Fablet <youenn@apple.com>
574 Fix inspector override conversion in InspectorPageAgent::overrideSetting
575 https://bugs.webkit.org/show_bug.cgi?id=200017
577 Reviewed by Devin Rousso.
579 Covered by inspector/page/overrideSetting-MockCaptureDevicesEnabled.html
580 which should not fail on bots not having cameras
582 * inspector/agents/InspectorPageAgent.cpp:
583 (WebCore::toOptionalBool):
584 (WebCore::InspectorPageAgent::overrideSetting):
585 Previously, the conversion for the mock capture value was implicit from a bool pointer to an optional.
586 Make an explicit conversion as done for regular settings.
588 2019-07-22 Ross Kirsling <ross.kirsling@sony.com>
590 Unreviewed non-unified build fix.
592 * Modules/webdatabase/DatabaseContext.h:
593 * html/HTMLDialogElement.cpp:
594 * html/HTMLScriptElement.cpp:
595 * inspector/agents/InspectorDatabaseAgent.cpp:
596 * workers/service/server/SWServerWorker.cpp:
597 * workers/service/server/SWServerWorker.h:
598 * worklets/WorkletGlobalScope.cpp:
599 * worklets/WorkletGlobalScope.h:
601 2019-07-22 Wenson Hsieh <wenson_hsieh@apple.com>
603 [iOS] [WK1] UIWebView always jumps to the top left corner when scrolling to reveal the selection
604 https://bugs.webkit.org/show_bug.cgi?id=200013
605 <rdar://problem/52526901>
607 Reviewed by Simon Fraser.
609 After <https://trac.webkit.org/r244141>, we no longer attempt to scroll to reveal the text selection in
610 UIWebView after changing the selection, due to how we use the legacy document view rect in legacy WebKit when
611 computing the visual viewport. This causes the viewRect in RenderLayer::scrollRectToVisible to be the same size
612 as the content size, which then causes us to always scroll to the origin when revealing the selection.
614 To make selection revealing work again in legacy WebKit, conditionally restore the old behavior of using the
615 unobscured content rect as the view rect, only in the case where scrolling is delegated and the platform widget
618 Test: WebKitLegacy.ScrollToRevealSelection
620 * page/FrameView.cpp:
621 (WebCore::FrameView::viewRectExpandedByContentInsets const):
622 (WebCore::FrameView::visualViewportRectExpandedByContentInsets const): Deleted.
624 Additionally rename visualViewportRectExpandedByContentInsets to viewRectExpandedByContentInsets, to reflect the
625 fact that this may either be the visual viewport rect or unobscured content rect.
628 * rendering/RenderLayer.cpp:
629 (WebCore::RenderLayer::scrollRectToVisible):
631 2019-07-22 Brent Fulgham <bfulgham@apple.com>
633 Correct web audio-related crash in seed reports
634 https://bugs.webkit.org/show_bug.cgi?id=200009
635 <rdar://problem/51565203>
637 Reviewed by Per Arne Vollan.
639 Update the 'createMix' method to do proper return value checking so that
640 we can clear the result of MTAudioProcessingTapCreate if the create operation
643 * platform/graphics/avfoundation/AudioSourceProviderAVFObjC.mm:
644 (WebCore::AudioSourceProviderAVFObjC::createMix):
646 2019-07-22 Per Arne Vollan <pvollan@apple.com>
648 Prewarmed font does not find any matches
649 https://bugs.webkit.org/show_bug.cgi?id=200007
651 Reviewed by Brent Fulgham.
653 Prewarming the font name 'SF Pro Text' does not find any matches. The font name ".SF NS Text/Display"
654 should be used instead. Also add 'Lucida Grande' to the list of prewarmed font, since it is used as
655 fallback font in many cases.
657 No new tests, covered by existing tests.
659 * platform/graphics/cocoa/FontCacheCoreText.cpp:
660 (WebCore::fontFamiliesForPrewarming):
662 2019-07-22 Youenn Fablet <youenn@apple.com>
664 Make sure the player underlying volume is correctly set after creating a new player
665 https://bugs.webkit.org/show_bug.cgi?id=199974
666 <rdar://problem/52762559>
668 Reviewed by Eric Carlson.
670 When creating a player before starting to play, the volume might not be correctly initialized.
671 Fix that by updating the volume like done for rate and muted state.
672 A future refactoring should probably move that code in MediaPlayer.
674 Add an internal API to query the volume value from the underlying player.
676 Test: http/tests/media/audio-volume-zero.html
678 * html/HTMLMediaElement.cpp:
679 (WebCore::HTMLMediaElement::updateVolume):
680 (WebCore::HTMLMediaElement::updatePlayState):
681 (WebCore::HTMLMediaElement::effectiveVolume const):
682 (WebCore::HTMLMediaElement::effectiveMuted const):
683 * html/HTMLMediaElement.h:
684 * testing/Internals.cpp:
685 (WebCore::Internals::privatePlayerVolume):
686 * testing/Internals.h:
687 * testing/Internals.idl:
688 * testing/Internals.mm:
689 (WebCore::Internals::privatePlayerVolume):
691 2019-07-22 Devin Rousso <drousso@apple.com>
693 Web Inspector: Page: mark any WebCore::Setting inspector override as UNLIKELY
694 https://bugs.webkit.org/show_bug.cgi?id=199925
696 Reviewed by Joseph Pecoraro.
698 * Scripts/SettingsTemplates/Settings.cpp.erb:
700 2019-07-22 Simon Fraser <simon.fraser@apple.com>
702 Fix crashes in ScrollingStateNode::insertChild()
703 https://bugs.webkit.org/show_bug.cgi?id=200023
704 rdar://problem/53265378
706 Reviewed by Darin Adler.
708 Crash data suggest that ScrollingStateNode::insertChild() can be passed an index that
709 is larger than the size of the vector, causing crashes.
711 Fix defensively by falling back to append() if the passed index is equal to or larger
712 than the size of the children vector.
714 * page/scrolling/ScrollingStateNode.cpp:
715 (WebCore::ScrollingStateNode::insertChild):
717 2019-07-22 Simon Fraser <simon.fraser@apple.com>
719 Make some constructors explicit
720 https://bugs.webkit.org/show_bug.cgi?id=199981
722 Reviewed by Daniel Bates.
724 Make explicit public constructors of objects that take POD or String& arguments,
725 to reduce the changes of the compiler doing implicit conversions.
727 * Modules/indexeddb/server/IndexValueEntry.h:
728 * Modules/indexeddb/server/IndexValueStore.h:
729 * Modules/indexeddb/shared/IDBDatabaseInfo.h:
730 * Modules/webdatabase/ChangeVersionData.h:
731 (WebCore::ChangeVersionData::ChangeVersionData):
732 * Modules/webdatabase/OriginLock.h:
733 * Modules/websockets/WebSocketExtensionParser.h:
734 (WebCore::WebSocketExtensionParser::WebSocketExtensionParser):
735 * bindings/js/ScriptCachedFrameData.h:
736 * contentextensions/URLFilterParser.cpp:
737 (WebCore::ContentExtensions::PatternParser::PatternParser):
738 * css/parser/CSSTokenizer.h:
739 * css/parser/SizesAttributeParser.h:
740 * dom/SpaceSplitString.cpp:
741 (WebCore::TokenIsEqualToCStringTokenProcessor::TokenIsEqualToCStringTokenProcessor):
742 * html/FormController.cpp:
743 (WebCore::SavedFormState::appendControlState):
744 (WebCore::SavedFormState::takeControlState):
745 * loader/ProgressTracker.cpp:
746 (WebCore::ProgressItem::ProgressItem):
747 * page/DragController.cpp:
748 (WebCore::DragController::DragController):
749 (WebCore::DragController::performDragOperation):
750 * page/DragController.h:
751 * page/animation/CSSPropertyAnimation.cpp:
752 (WebCore::AnimationPropertyWrapperBase::AnimationPropertyWrapperBase):
753 * page/ios/DOMTimerHoldingTank.h:
754 (WebCore::DeferDOMTimersForScope::DeferDOMTimersForScope):
755 * platform/CountedUserActivity.h:
756 (WebCore::CountedUserActivity::CountedUserActivity):
757 * platform/audio/DirectConvolver.h:
758 * platform/audio/DownSampler.h:
759 * platform/audio/DynamicsCompressorKernel.h:
760 * platform/audio/EqualPowerPanner.h:
761 * platform/audio/MultiChannelResampler.h:
762 * platform/audio/ReverbAccumulationBuffer.h:
763 * platform/audio/ReverbInputBuffer.h:
764 * platform/audio/SincResampler.cpp:
765 * platform/audio/UpSampler.h:
766 * platform/audio/ios/AudioFileReaderIOS.h:
767 * platform/audio/mac/AudioFileReaderMac.h:
768 * platform/audio/mac/AudioSessionMac.cpp:
769 (WebCore::AudioSessionPrivate::AudioSessionPrivate):
770 * platform/cf/KeyedDecoderCF.h:
771 * platform/graphics/WidthIterator.cpp:
772 (WebCore::OriginalAdvancesForCharacterTreatedAsSpace::OriginalAdvancesForCharacterTreatedAsSpace):
773 * platform/graphics/cocoa/FontCacheCoreText.h:
774 (WebCore::SynthesisPair::SynthesisPair):
775 * platform/graphics/displaylists/DisplayListItems.h:
776 (WebCore::DisplayList::DrawingItem::DrawingItem):
777 * platform/network/cf/ResourceRequest.h:
778 (WebCore::ResourceRequest::ResourceRequest):
779 * platform/text/TextCodecICU.h:
780 * rendering/RenderTableSection.h:
781 (WebCore::CellSpan::CellSpan):
782 * rendering/shapes/RasterShape.cpp:
783 * rendering/shapes/RasterShape.h:
784 (WebCore::RasterShapeIntervals::RasterShapeIntervals):
785 * testing/MockLibWebRTCPeerConnection.h:
786 (WebCore::MockLibWebRTCIceCandidate::MockLibWebRTCIceCandidate):
788 2019-07-22 Rob Buis <rbuis@igalia.com>
790 Remove some unneeded code in MathOperator
791 https://bugs.webkit.org/show_bug.cgi?id=199935
793 Reviewed by Frédéric Wang.
795 Remove unused method.
797 No new tests, not changing behavior.
799 * rendering/mathml/MathOperator.h:
801 2019-07-21 Konstantin Tokarev <annulen@yandex.ru>
803 Add DataListButtonElement.cpp to unified sources
804 https://bugs.webkit.org/show_bug.cgi?id=199989
806 Reviewed by Michael Catanzaro.
809 * WebCore.xcodeproj/project.pbxproj:
811 2019-07-21 Andres Gonzalez <andresg_22@apple.com>
813 Add accessibilityInsertText for text insertion in edit fields.
814 https://bugs.webkit.org/show_bug.cgi?id=199973
816 Reviewed by Chris Fleizach.
818 Renamed accessibilityInsertText to _accessibilityInsertText.
820 * accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
821 (-[WebAccessibilityObjectWrapper _accessibilityInsertText:]):
822 (-[WebAccessibilityObjectWrapper accessibilityInsertText:]): Deleted.
823 * accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
824 (-[WebAccessibilityObjectWrapper _accessibilityInsertText:]):
825 (-[WebAccessibilityObjectWrapper accessibilityInsertText:]): Deleted.
827 2019-07-21 Saam Barati <sbarati@apple.com>
829 [WHLSL] Checker needs to setError() when a property access node can't commit its base type
830 https://bugs.webkit.org/show_bug.cgi?id=199978
832 Reviewed by Dean Jackson.
834 Test: webgpu/whlsl/null-array-property-access.html
836 * Modules/webgpu/WHLSL/WHLSLChecker.cpp:
837 (WebCore::WHLSL::Checker::finishVisiting):
839 2019-07-21 Saam Barati <sbarati@apple.com>
841 [WHLSL] Return the zero-value enum in the enum-from-integer constructor when the integer is not a valid enum value
842 https://bugs.webkit.org/show_bug.cgi?id=199853
844 Reviewed by Dean Jackson.
846 Test: webgpu/whlsl/enum-integer-constructor.html
848 * Modules/webgpu/WHLSL/Metal/WHLSLNativeFunctionWriter.cpp:
849 (WebCore::WHLSL::Metal::writeNativeFunction):
851 2019-07-21 Myles C. Maxfield <mmaxfield@apple.com>
853 [iOS] Fast and complex text codepaths disagree about how to lay out bopomofo with tone marks
854 https://bugs.webkit.org/show_bug.cgi?id=199912
855 <rdar://problem/52182454>
857 Reviewed by Simon Fraser.
859 This is because CoreText has special composition rules for CJK languages, which we don't have in our simple
860 text codepath. Rather than implementing the composition rules in WebKit, we can simply disable them in
863 Test: fast/text/international/system-language/composition.html
865 * platform/graphics/mac/SimpleFontDataCoreText.cpp:
866 (WebCore::Font::getCFStringAttributes const):
868 2019-07-20 Andres Gonzalez <andresg_22@apple.com>
870 Add accessibilityInsertText for text insertion in edit fields.
871 https://bugs.webkit.org/show_bug.cgi?id=199973
873 Reviewed by Chris Fleizach.
875 Tests: accessibility/insert-newline.html
876 accessibility/ios-simulator/insert-newline.html
878 Accessibility clients like VoiceOver and Voice Control were entering
879 text in text fields by replacing the entire content of the field
880 (SetValue) and then setting the insertion point to the appropriate
881 offset (SetSelectedTextRange). accessibilityInsertText gives a simpler
882 interface to clients to insert text at the insertion point location.
883 In addition, this provides a workaround for the issue encountered with
884 the previous method when inserting a linebreak.
886 * accessibility/AccessibilityObject.cpp:
887 (WebCore::AccessibilityObject::insertText):
888 * accessibility/AccessibilityObject.h:
889 * accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
890 (-[WebAccessibilityObjectWrapper accessibilityInsertText:]):
891 * accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
892 (-[WebAccessibilityObjectWrapper accessibilityInsertText:]):
894 2019-07-20 Alexander Mikhaylenko <exalm7659@gmail.com>
896 REGRESSION(r246033/r246496): [GTK] Kinetic scrolling doesn't work
897 https://bugs.webkit.org/show_bug.cgi?id=199322
899 Reviewed by Michael Catanzaro.
901 Check ENABLE(KINETIC_SCROLLING) instead of ENABLE(ASYNC_SCROLLING) for kinetic scrolling.
903 * platform/PlatformWheelEvent.h:
904 * platform/generic/ScrollAnimatorGeneric.cpp:
905 (WebCore::ScrollAnimatorGeneric::handleWheelEvent):
906 * platform/gtk/PlatformWheelEventGtk.cpp:
907 (WebCore::PlatformWheelEvent::PlatformWheelEvent):
909 2019-07-20 Zalan Bujtas <zalan@apple.com>
911 [LFC][TFC] Introduce table formatting context.
912 https://bugs.webkit.org/show_bug.cgi?id=199979
913 <rdar://problem/53346292>
915 Reviewed by Antti Koivisto.
917 Add the basic class structure for the table formatting context.
918 https://www.w3.org/TR/CSS22/tables.html
921 * WebCore.xcodeproj/project.pbxproj:
922 * layout/FormattingState.h:
923 (WebCore::Layout::FormattingState::isTableFormattingState const):
924 * layout/tableformatting/TableFormattingContext.cpp: Added.
925 (WebCore::Layout::TableFormattingContext::TableFormattingContext):
926 (WebCore::Layout::TableFormattingContext::layout const):
927 * layout/tableformatting/TableFormattingContext.h: Added.
928 * layout/tableformatting/TableFormattingState.cpp: Added.
929 (WebCore::Layout::TableFormattingState::TableFormattingState):
930 (WebCore::Layout::TableFormattingState::~TableFormattingState):
931 * layout/tableformatting/TableFormattingState.h: Added.
932 * layout/tableformatting/TableInvalidation.cpp: Added.
933 (WebCore::Layout::TableInvalidation::invalidate):
934 * layout/tableformatting/TableInvalidation.h: Added.
936 2019-07-20 Zalan Bujtas <zalan@apple.com>
938 [LFC][IFC] Add support for mid-word break.
939 https://bugs.webkit.org/show_bug.cgi?id=199976
940 <rdar://problem/53337349>
942 Reviewed by Antti Koivisto.
944 Introduce leading and trailing InlineTextItem to support incoming and outgoing partial content.
945 TextUtil::split is a straight copy of SimpleLineLayout::split.
947 * layout/inlineformatting/InlineFormattingContextLineLayout.cpp:
948 (WebCore::Layout::inlineItemWidth):
949 (WebCore::Layout::LineLayout::close):
950 (WebCore::Layout::LineLayout::placeInlineItem):
951 (WebCore::Layout::LineLayout::layout):
952 * layout/inlineformatting/InlineTextItem.cpp:
953 (WebCore::Layout::InlineTextItem::split const):
954 (WebCore::Layout::InlineTextItem::split): Deleted.
955 * layout/inlineformatting/InlineTextItem.h:
956 (WebCore::Layout::InlineTextItem::inlineBox const):
957 * layout/inlineformatting/text/TextUtil.cpp:
958 (WebCore::Layout::TextUtil::split):
959 * layout/inlineformatting/text/TextUtil.h:
961 2019-07-20 Zalan Bujtas <zalan@apple.com>
963 [Text autosizing] Do not nuke the style on dynamicViewportSizeUpdate
964 https://bugs.webkit.org/show_bug.cgi?id=199718
965 <rdar://problem/53344961>
967 Reviewed by Simon Fraser.
969 When the page scale changes (e.g. as the result of shink to fit mode) we need to visit all the text content on the page and check whether they need to be boosted.
970 Currently we call setNeedsRecalcStyleInAllFrames() to accomplish it. Unfortunatelly setNeedsRecalcStyleInAllFrames destroys all the style information which means that the
971 subsequent styleResolve() needs to start from scratch.
972 This patch addresses this issue by directly adjusting the computed style information when text boosting is required and schedules layout accordingly. We also trigger this style adjusting
973 on a timer so that rapid dynamicViewportSizeUpdate() calls won't trigger redundant layouts.
975 * css/StyleResolver.cpp:
976 (WebCore::hasTextChild):
977 (WebCore::StyleResolver::adjustRenderStyleForTextAutosizing):
978 (WebCore::hasTextChildren): Deleted.
979 * css/StyleResolver.h:
982 (WebCore::Page::invalidateTextAutoSizeInAllFrames):
985 2019-07-20 Saam Barati <sbarati@apple.com>
987 [WHLSL] Make enums work
988 https://bugs.webkit.org/show_bug.cgi?id=199634
990 Reviewed by Robin Morisset.
992 This patch makes enums work. There were a couple changes needed to do this work:
993 1. In the metal codegen, we were emitting code like "EnumType.EnumValue"
994 instead of "EnumType::EnumValue".
996 2. Inside the enumeration declaration AST node, we track enums in a HashMap
997 because we need to look up names of enum values in that map in parts of the
998 compiler. However, HashMap is not ordered but we were relying on the iteration
999 order of it to generate the default values of enums. E.g, in: "enum { Foo = 10, Bar }",
1000 Bar should be 11. However, there was no guarantee that we iterated over Bar
1001 after Foo. Instead, we just track the default values as we parse an enum
1002 inside of the parser. This allows us to continue to keep using HashMap without
1003 also adding an auxiliary data structure (or using more memory in some way) to
1004 track the order of the enum elements.
1006 I've also filed a patch to do follow-up work on implementing the correct
1007 operational behavior when constructing an enum from its underlying type when
1008 the underlying value does not match any enum value:
1009 https://bugs.webkit.org/show_bug.cgi?id=199853
1011 Tests: webgpu/whlsl/enums-2.html
1012 webgpu/whlsl/enums.html
1014 * Modules/webgpu/WHLSL/AST/WHLSLEnumerationMember.h:
1015 (WebCore::WHLSL::AST::EnumerationMember::EnumerationMember):
1016 (WebCore::WHLSL::AST::EnumerationMember::value):
1017 (WebCore::WHLSL::AST::EnumerationMember::setValue): Deleted.
1018 * Modules/webgpu/WHLSL/Metal/WHLSLFunctionWriter.cpp:
1019 (WebCore::WHLSL::Metal::FunctionDefinitionWriter::visit):
1020 (WebCore::WHLSL::Metal::FunctionDefinitionWriter::constantExpressionString):
1021 * Modules/webgpu/WHLSL/Metal/WHLSLTypeNamer.cpp:
1022 (WebCore::WHLSL::Metal::TypeNamer::emitNamedTypeDefinition):
1023 * Modules/webgpu/WHLSL/WHLSLASTDumper.cpp:
1024 (WebCore::WHLSL::ASTDumper::visit):
1025 * Modules/webgpu/WHLSL/WHLSLChecker.cpp:
1026 (WebCore::WHLSL::Checker::visit):
1027 * Modules/webgpu/WHLSL/WHLSLParser.cpp:
1028 (WebCore::WHLSL::Parser::parseEnumerationDefinition):
1029 (WebCore::WHLSL::Parser::parseEnumerationMember):
1030 * Modules/webgpu/WHLSL/WHLSLParser.h:
1031 * Modules/webgpu/WHLSL/WHLSLVisitor.cpp:
1032 (WebCore::WHLSL::Visitor::visit):
1034 2019-07-19 Youenn Fablet <youenn@apple.com>
1036 Remote WebInspector should enable mock capture devices in UIProcess if doing it in WebProcess
1037 https://bugs.webkit.org/show_bug.cgi?id=199924
1038 <rdar://problem/50552067>
1040 Reviewed by Devin Rousso.
1042 Add necessary API to set mock capture devices override.
1043 In case of desynchronization between webprocess and uiprocess, make sure to return early
1044 and fail capture instead of crashing.
1046 * inspector/InspectorClient.h:
1047 (WebCore::InspectorClient::setMockCaptureDevicesEnabled):
1048 * inspector/agents/InspectorPageAgent.cpp:
1049 (WebCore::InspectorPageAgent::disable):
1050 (WebCore::InspectorPageAgent::overrideSetting):
1051 * platform/mock/MockRealtimeMediaSourceCenter.cpp:
1053 2019-07-19 Andy Estes <aestes@apple.com>
1055 [Apple Pay] Log less in PaymentCoordinator::shouldEnableApplePayAPIs()
1056 https://bugs.webkit.org/show_bug.cgi?id=199958
1058 Reviewed by Tim Horton.
1060 Only log when APIs are disabled due to user scripts.
1062 * Modules/applepay/PaymentCoordinator.cpp:
1063 (WebCore::PaymentCoordinator::shouldEnableApplePayAPIs const):
1065 2019-07-19 Dean Jackson <dino@apple.com>
1067 Add a menu item to toggle between showing and hiding link previews
1068 https://bugs.webkit.org/show_bug.cgi?id=199940
1069 <rdar://problem/53032288>
1071 Reviewed by Beth Dakin.
1073 New strings for Show/Hide Link Previews.
1075 * en.lproj/Localizable.strings:
1077 2019-07-19 Antoine Quint <graouts@apple.com>
1079 Links stop working after long-pressing a link (WK1)
1080 https://bugs.webkit.org/show_bug.cgi?id=199952
1081 <rdar://problem/52291860>
1083 Reviewed by Wenson Hsieh.
1085 Test: fast/events/touch/ios/click-after-long-press.html
1087 When EventHandler::tryToBeginDragAtPoint() is called, we must reset m_mousePressed to false so that
1088 WebChromeClientIOS::observedContentChange() is called by EventHandler::mousePressed() when we would
1089 process the next tap after a drag interaction.
1091 * page/ios/EventHandlerIOS.mm:
1092 (WebCore::EventHandler::tryToBeginDragAtPoint):
1094 2019-07-19 Sihui Liu <sihui_liu@apple.com>
1096 IndexedDB: error in starting version change transaction may be neglected
1097 https://bugs.webkit.org/show_bug.cgi?id=199818
1098 <rdar://problem/52925738>
1100 Reviewed by Brady Eidson.
1102 For version change transaction, IDBServer didn't wait the result of beginTransaction on the background thread
1103 before giving the IDBClient the result of open request. In this case, beginTransaction may fail to update the
1104 DatabaseVersion in database file or set m_originalDatabaseInfoBeforeVersionChange, but the transaction was
1105 marked as started. When we later set m_databaseInfo with m_originalDatabaseInfoBeforeVersionChange,
1106 m_databaseInfo could become nullptr.
1108 To write a test for this, we will need to simulate an SQLite error. I manually tested this by crafting the
1109 SQLiteStatement in beginTransaction, making it an invalid statement, and verified that error event, instead of
1110 ungradeneeded event is dispatched to the IDBRequest.
1112 * Modules/indexeddb/server/UniqueIDBDatabase.cpp:
1113 (WebCore::IDBServer::UniqueIDBDatabase::startVersionChangeTransaction):
1114 (WebCore::IDBServer::UniqueIDBDatabase::performStartVersionChangeTransaction):
1115 (WebCore::IDBServer::UniqueIDBDatabase::didPerformStartVersionChangeTransaction):
1116 (WebCore::IDBServer::UniqueIDBDatabase::immediateCloseForUserDelete):
1117 (WebCore::IDBServer::UniqueIDBDatabase::beginTransactionInBackingStore): Deleted.
1118 * Modules/indexeddb/server/UniqueIDBDatabase.h:
1120 2019-07-19 Zalan Bujtas <zalan@apple.com>
1122 [LFC][IFC] Add partial content handling
1123 https://bugs.webkit.org/show_bug.cgi?id=199930
1125 Reviewed by Antti Koivisto.
1127 Move the actual inline item placement to a dedicated functions. This enables us to
1128 construct a dedicated InlineItem for the partial content and pass it in to this function.
1130 <span>The quick brown fox jumps over the lazy dog</span> <- inline content
1131 [The][ ][quick][ ][brown][ ][fox][ ][jumps][ ][over][ ][the][ ][lazy][ ][dog] <- InlineItems
1135 [ick] <- Partial InlineItem (position: 6 length: 3) from [quick] (position: 4 length: 5)
1137 * layout/inlineformatting/InlineFormattingContextLineLayout.cpp:
1138 (WebCore::Layout::LineLayout::close):
1139 (WebCore::Layout::LineLayout::placeInlineItem):
1140 (WebCore::Layout::LineLayout::layout):
1141 (WebCore::Layout::LineInput::LineInput):
1142 (WebCore::Layout::InlineFormattingContext::InlineLayout::layout const):
1143 (WebCore::Layout::InlineFormattingContext::InlineLayout::computedIntrinsicWidth const):
1144 * layout/inlineformatting/InlineTextItem.cpp:
1145 (WebCore::Layout::InlineTextItem::split):
1146 * layout/inlineformatting/InlineTextItem.h:
1148 2019-07-19 Charlie Turner <cturner@igalia.com>
1150 [GStreamer] Flush get_range calls during PAUSED->READY in WebKitWebSource
1151 https://bugs.webkit.org/show_bug.cgi?id=199934
1153 Reviewed by Xabier Rodriguez-Calvar.
1155 Unit testing not applicable.
1157 * platform/graphics/gstreamer/WebKitWebSourceGStreamer.cpp:
1158 (webKitWebSrcChangeState): A well-behaved element should unblock streaming threads
1159 during a PAUSED->READY transition, so do that here.
1161 2019-07-19 Rob Buis <rbuis@igalia.com>
1163 Remove some unneeded code in MathOperator
1164 https://bugs.webkit.org/show_bug.cgi?id=199935
1166 Reviewed by Frédéric Wang.
1168 The initialize method is unneeded since the member
1169 variables all use default initialization.
1171 No new tests, not changing behavior.
1173 * rendering/mathml/MathOperator.cpp:
1174 (WebCore::MathOperator::MathOperator):
1175 (WebCore::MathOperator::GlyphAssemblyData::initialize): Deleted.
1177 2019-07-18 Antoine Quint <graouts@apple.com>
1179 Enable simulated mouse events quirk for soundcloud.com
1180 https://bugs.webkit.org/show_bug.cgi?id=199908
1181 <rdar://problem/52915981>
1183 Reviewed by Brent Fulgham.
1185 Scrubbing on soundcloud.com relies on mouse events alone when displaying on iPad with the desktop UA string.
1186 Opting into the simulated mouse events quirk makes scrubbing work and we also ensure that page scrolling is
1187 disabled as a result of scrubbing by returning true in shouldDispatchedSimulatedMouseEventsAssumeDefaultPrevented().
1189 Also making a drive-by fix in shouldDispatchedSimulatedMouseEventsAssumeDefaultPrevented() to use a reference
1190 instead of a pointer when downcasting a null-checked pointer.
1193 (WebCore::Quirks::shouldDispatchSimulatedMouseEvents const):
1194 (WebCore::Quirks::shouldDispatchedSimulatedMouseEventsAssumeDefaultPrevented const):
1196 2019-07-18 Per Arne Vollan <pvollan@apple.com>
1198 Prewarm font cache with more fonts
1199 https://bugs.webkit.org/show_bug.cgi?id=199283
1201 Reviewed by Myles C. Maxfield.
1203 Initialize the list of families requiring system fallback in the prewarming information struct, since
1204 this was part of the original, approved patch, and is needed for the performance improvement.
1206 No new tests, covered by existing tests.
1208 * platform/graphics/cocoa/FontCacheCoreText.cpp:
1209 (WebCore::fontFamiliesForPrewarming):
1210 (WebCore::FontCache::prewarmGlobally):
1212 2019-07-18 Ryan Haddad <ryanhaddad@apple.com>
1214 Unreviewed, rolling out r247531.
1216 Caused multiple IndexedDB layout test crashes.
1220 "IndexedDB: error in starting version change transaction may
1222 https://bugs.webkit.org/show_bug.cgi?id=199818
1223 https://trac.webkit.org/changeset/247531
1225 2019-07-18 Antoine Quint <graouts@apple.com>
1227 Ensure Quirks::shouldDispatchSimulatedMouseEvents() works for all Google Maps and Amazon domains
1228 https://bugs.webkit.org/show_bug.cgi?id=199904
1229 <rdar://problem/53250104>
1231 Reviewed by Dean Jackson.
1233 Use topPrivatelyControlledDomain() to determine whether the URL is a Google or Amazon domain so as to apply
1234 Google Maps and Amazon quirks to all the various domain names used.
1237 (WebCore::Quirks::isAmazon const):
1238 (WebCore::Quirks::isGoogleMaps const):
1239 (WebCore::Quirks::shouldDispatchSimulatedMouseEvents const):
1240 (WebCore::Quirks::simulatedMouseEventTypeForTarget const):
1243 2019-07-18 Per Arne Vollan <pvollan@apple.com>
1245 Crash under WebPage::boundaryEventOccurred
1246 https://bugs.webkit.org/show_bug.cgi?id=199907
1248 Reviewed by Chris Fleizach.
1250 Add null pointer checks.
1252 No new tests, since I have not been able to reproduce this in a test.
1254 * Modules/speech/SpeechSynthesis.cpp:
1255 (WebCore::SpeechSynthesis::didStartSpeaking):
1256 (WebCore::SpeechSynthesis::didFinishSpeaking):
1257 (WebCore::SpeechSynthesis::didPauseSpeaking):
1258 (WebCore::SpeechSynthesis::didResumeSpeaking):
1259 (WebCore::SpeechSynthesis::speakingErrorOccurred):
1260 (WebCore::SpeechSynthesis::boundaryEventOccurred):
1262 2019-07-18 Antoine Quint <graouts@apple.com>
1264 [Pointer Events] The button and buttons properties are incorrect on iOS
1265 https://bugs.webkit.org/show_bug.cgi?id=199910
1266 <rdar://problem/52778374>
1268 Reviewed by Dean Jackson.
1270 The button and buttons properties were always set to 0 on iOS. We now use the correct values such that
1271 button is always 0 except for "pointermove" where it's -1, and "buttons" is 1 as long as the pointer is
1272 in contact with the touch surface.
1274 Tests: pointerevents/ios/pointer-event-button-and-buttons-pointer-cancel.html
1275 pointerevents/ios/pointer-event-button-and-buttons.html
1277 * dom/ios/PointerEventIOS.cpp:
1278 (WebCore::buttonForType):
1279 (WebCore::buttonsForType):
1280 (WebCore::PointerEvent::PointerEvent):
1282 2019-07-18 Antoine Quint <graouts@apple.com>
1284 REGRESSION: Unable to enable simulated mouse events using the runtime flag
1285 https://bugs.webkit.org/show_bug.cgi?id=199909
1286 <rdar://problem/53254895>
1288 Reviewed by Dean Jackson.
1290 The fix for wkb.ug/199508 (r247152) broke the ability enable simulated mouse events using the runtime flag.
1291 We now add a check for the flag in Quirks::shouldDispatchSimulatedMouseEvents() so that the runtime flag
1292 overrides the content of that function. This is more practical and allows for a quick override using the flag.
1294 Additionally, we need only call shouldDispatchSimulatedMouseEvents() in simulatedMouseEventTypeForTarget() since
1295 that function already calls needsQuirks().
1298 (WebCore::EventNames::isTouchRelatedEventType const):
1300 (WebCore::Node::moveNodeToNewDocument):
1302 (WebCore::Quirks::shouldDispatchSimulatedMouseEvents const):
1303 (WebCore::Quirks::simulatedMouseEventTypeForTarget const):
1305 2019-07-18 Antoine Quint <graouts@apple.com>
1307 REGRESSION: Panning on an Amazon product image scrolls the page on iPadOS
1308 https://bugs.webkit.org/show_bug.cgi?id=199905
1309 <rdar://problem/49124529>
1311 Reviewed by Dean Jackson.
1313 Amazon product pages include images that the user can touch and pan to show zoomed details in a side image. This
1314 currently works on iPadOS thanks to the dispatch of simulated "mousemove" events on the product image, but the site
1315 doesn't call preventDefault() when handling those events as it wasn't necessary for macOS.
1317 We add a new quirk that will indicate that a given element is such a product image.
1320 (WebCore::Quirks::isAmazon const):
1321 (WebCore::Quirks::shouldDispatchSimulatedMouseEvents const):
1322 (WebCore::Quirks::shouldDispatchedSimulatedMouseEventsAssumeDefaultPrevented const):
1323 (WebCore::Quirks::simulatedMouseEventTypeForTarget const):
1326 2019-07-18 Youenn Fablet <youenn@apple.com>
1328 Make sure to set kCTFontFallbackOptionAttribute to kCTFontFallbackOptionSystem for system fonts
1329 https://bugs.webkit.org/show_bug.cgi?id=199769
1330 <rdar://problem/49390297>
1332 Reviewed by Myles C. Maxfield.
1334 When getting a system font, set the appropriate attribute so that it
1335 does not fallback to a user initiated font.
1336 Add an ASSERT that checks that the font in use is not a user font if
1337 policy is to not use user installed fonts.
1339 Tests: fast/text/user-installed-fonts/extended-character-with-user-font.html
1340 fast/text/user-installed-fonts/extended-character.html
1342 * platform/graphics/Font.h:
1343 * platform/graphics/FontCascadeFonts.cpp:
1344 (WebCore::FontCascadeFonts::glyphDataForSystemFallback):
1345 * platform/graphics/cocoa/FontCacheCoreText.cpp:
1346 (WebCore::preparePlatformFont):
1347 (WebCore::FontDatabase::singletonAllowingUserInstalledFonts):
1348 (WebCore::FontDatabase::singletonDisallowingUserInstalledFonts):
1349 (WebCore::addAttributesForInstalledFonts):
1350 (WebCore::addAttributesForWebFonts):
1351 (WebCore::installedFontMandatoryAttributes):
1352 * platform/graphics/mac/SimpleFontDataCoreText.cpp:
1353 (WebCore::Font::isUserInstalledFont const):
1355 2019-07-18 Zalan Bujtas <zalan@apple.com>
1357 [LFC][IFC] InlineFormattingContext::LineLayout::placeInlineItems is getting too complex.
1358 https://bugs.webkit.org/show_bug.cgi?id=199898
1359 <rdar://problem/53235958>
1361 Reviewed by Antti Koivisto.
1363 It's time to restructure LineLayout::placeInlineItems to be able to expand it further.
1364 Introduce the LineLayout class. This class is responsible to place the inline items on the current line.
1365 (Rename InlineFormattingContext::lineLayout -> InlineFormattingContext::InlineLayout and
1366 use Line::InitialConstraints in LineInput)
1368 * layout/inlineformatting/InlineFormattingContext.cpp:
1369 (WebCore::Layout::InlineFormattingContext::layout const):
1370 (WebCore::Layout::InlineFormattingContext::computeIntrinsicWidthConstraints const):
1371 * layout/inlineformatting/InlineFormattingContext.h:
1372 (WebCore::Layout::InlineFormattingContext::LineLayout::layoutState const): Deleted.
1373 * layout/inlineformatting/InlineFormattingContextLineLayout.cpp:
1374 (WebCore::Layout::inlineItemWidth):
1375 (WebCore::Layout::LineLayout::layoutState const):
1376 (WebCore::Layout::LineLayout::UncommittedContent::runs):
1377 (WebCore::Layout::LineLayout::UncommittedContent::isEmpty const):
1378 (WebCore::Layout::LineLayout::UncommittedContent::size const):
1379 (WebCore::Layout::LineLayout::UncommittedContent::width const):
1380 (WebCore::Layout::LineLayout::UncommittedContent::add):
1381 (WebCore::Layout::LineLayout::UncommittedContent::reset):
1382 (WebCore::Layout::LineLayout::LineLayout):
1383 (WebCore::Layout::LineLayout::commitPendingContent):
1384 (WebCore::Layout::LineLayout::close):
1385 (WebCore::Layout::LineLayout::layout):
1386 (WebCore::Layout::LineInput::LineInput):
1387 (WebCore::Layout::InlineFormattingContext::InlineLayout::InlineLayout):
1388 (WebCore::Layout::InlineFormattingContext::InlineLayout::layout const):
1389 (WebCore::Layout::InlineFormattingContext::InlineLayout::computedIntrinsicWidth const):
1390 (WebCore::Layout::InlineFormattingContext::InlineLayout::createDisplayRuns const):
1391 (WebCore::Layout::InlineFormattingContext::InlineLayout::alignRuns const):
1392 (WebCore::Layout::UncommittedContent::runs): Deleted.
1393 (WebCore::Layout::UncommittedContent::isEmpty const): Deleted.
1394 (WebCore::Layout::UncommittedContent::size const): Deleted.
1395 (WebCore::Layout::UncommittedContent::width const): Deleted.
1396 (WebCore::Layout::UncommittedContent::add): Deleted.
1397 (WebCore::Layout::UncommittedContent::reset): Deleted.
1398 (WebCore::Layout::LineInput::HorizontalConstraint::HorizontalConstraint): Deleted.
1399 (WebCore::Layout::InlineFormattingContext::LineLayout::LineLayout): Deleted.
1400 (WebCore::Layout::InlineFormattingContext::LineLayout::placeInlineItems const): Deleted.
1401 (WebCore::Layout::InlineFormattingContext::LineLayout::layout const): Deleted.
1402 (WebCore::Layout::InlineFormattingContext::LineLayout::computedIntrinsicWidth const): Deleted.
1403 (WebCore::Layout::InlineFormattingContext::LineLayout::createDisplayRuns const): Deleted.
1404 (WebCore::Layout::InlineFormattingContext::LineLayout::alignRuns const): Deleted.
1405 * layout/inlineformatting/InlineLine.cpp:
1406 (WebCore::Layout::Line::Line):
1407 * layout/inlineformatting/InlineLine.h:
1409 2019-07-18 Carlos Garcia Campos <cgarcia@igalia.com>
1411 [GTK] MOUSE_CURSOR_SCALE is not implemented
1412 https://bugs.webkit.org/show_bug.cgi?id=109469
1414 Reviewed by Darin Adler.
1416 Use gdk_cursor_new_from_surface() instead of gdk_cursor_new_from_pixbuf() to avoid the surface to pixbuf
1419 * platform/gtk/CursorGtk.cpp:
1420 (WebCore::createCustomCursor):
1422 2019-07-17 Chris Dumez <cdumez@apple.com>
1424 Prewarm local storage in the NetworkProcess to reduce WebContent process hangs
1425 https://bugs.webkit.org/show_bug.cgi?id=199879
1426 <rdar://problem/53217757>
1428 Reviewed by Ryosuke Niwa.
1430 When JS accesses window.localStorage for the first time, we end up doing a
1431 synchronous IPC to the network process to pull in all items in the local
1432 storage for the origin. If the network process does not have this data in
1433 memory, it has to read it from a database on disk, which may take a significant
1434 amount of time and hang the WebContent process during this time.
1436 To alleviate this problem, this patch introduces prewarming on the local storage
1437 in the network process when loading a given origin in the WebContent process.
1438 This way, in most cases, when the JS accesses window.localStorage for the first
1439 time, the synchronous IPC to the network process returns much faster (measured
1440 50-100ms for a very large database, down from 250-300ms), as it only needs to
1441 IPC the data over, without the need to fetch it from disk.
1443 As a safety net to avoid excessive prewarming, we currently prewarm at most 5
1444 security origins per page load.
1446 * loader/DocumentLoader.cpp:
1447 (WebCore::DocumentLoader::commitData):
1448 * page/DOMWindow.cpp:
1449 (WebCore::DOMWindow::prewarmLocalStorageIfNecessary):
1452 (WebCore::Frame::didPrewarmLocalStorage):
1453 (WebCore::Frame::mayPrewarmLocalStorage const):
1455 * storage/Storage.cpp:
1456 (WebCore::Storage::prewarm):
1457 * storage/Storage.h:
1458 * storage/StorageArea.h:
1459 (WebCore::StorageArea::prewarm):
1461 2019-07-17 Robin Morisset <rmorisset@apple.com>
1463 [WHLSL] The lexer should not choke on a single '/'
1464 https://bugs.webkit.org/show_bug.cgi?id=199886
1466 Reviewed by Myles Maxfield.
1468 The bug is trivial: just recognize that a single '/' without a '*' or '/' afterwards is not the beginning of a comment.
1470 While investigating this bug, I also found that part of the parser fails to properly propagate errors, so I fixed it.
1472 New test in LayoutTests/webgpu/whlsl/float-math.html
1473 I also added a new flag to WHLSLPrepare.cpp. It can be used to parse the full standard library (this is how I found this bug in the first place).
1475 * Modules/webgpu/WHLSL/WHLSLLexer.cpp:
1476 (WebCore::WHLSL::Lexer::skipWhitespaceAndComments):
1477 * Modules/webgpu/WHLSL/WHLSLParser.cpp:
1478 (WebCore::WHLSL::Parser::parsePossibleTernaryConditional):
1479 * Modules/webgpu/WHLSL/WHLSLPrepare.cpp:
1480 (WebCore::WHLSL::prepareShared):
1481 * Modules/webgpu/WHLSL/WHLSLStandardLibraryUtilities.cpp:
1482 (WebCore::WHLSL::includeStandardLibrary):
1483 * Modules/webgpu/WHLSL/WHLSLStandardLibraryUtilities.h:
1485 2019-07-17 Simon Fraser <simon.fraser@apple.com>
1487 Unreviewed build fix after r247549.
1489 * Modules/webgpu/WHLSL/WHLSLRecursiveTypeChecker.cpp:
1491 2019-07-17 Robin Morisset <rmorisset@apple.com>
1493 [WHLSL] Remove traps from the compiler
1494 https://bugs.webkit.org/show_bug.cgi?id=195811
1495 <rdar://problem/50746299>
1497 Reviewed by Myles Maxfield.
1499 Remove the Trap statement from the language, see https://github.com/gpuweb/WHLSL/issues/301.
1501 No new tests, as this part of the language was already untested, and we are removing it, not adding anything.
1503 * Modules/webgpu/WHLSL/AST/WHLSLAST.h:
1504 * Modules/webgpu/WHLSL/AST/WHLSLStatement.h:
1505 (WebCore::WHLSL::AST::Statement::isSwitchStatement const):
1506 * Modules/webgpu/WHLSL/AST/WHLSLTrap.h: Removed.
1507 * Modules/webgpu/WHLSL/Metal/WHLSLFunctionWriter.cpp:
1508 * Modules/webgpu/WHLSL/WHLSLASTDumper.cpp:
1509 * Modules/webgpu/WHLSL/WHLSLASTDumper.h:
1510 * Modules/webgpu/WHLSL/WHLSLLexer.cpp:
1511 (WebCore::WHLSL::Token::typeName):
1512 (WebCore::WHLSL::Lexer::recognizeKeyword):
1513 * Modules/webgpu/WHLSL/WHLSLLexer.h:
1514 * Modules/webgpu/WHLSL/WHLSLParser.cpp:
1515 (WebCore::WHLSL::Parser::parseStatement):
1516 * Modules/webgpu/WHLSL/WHLSLParser.h:
1517 * Modules/webgpu/WHLSL/WHLSLStatementBehaviorChecker.cpp:
1518 * Modules/webgpu/WHLSL/WHLSLVisitor.cpp:
1519 (WebCore::WHLSL::Visitor::visit):
1520 * Modules/webgpu/WHLSL/WHLSLVisitor.h:
1521 * WebCore.xcodeproj/project.pbxproj:
1523 2019-07-17 Robin Morisset <rmorisset@apple.com>
1525 [WHLSL] checkRecursiveType should not have exponential complexity.
1526 https://bugs.webkit.org/show_bug.cgi?id=199835
1528 Reviewed by Myles Maxfield.
1530 The change is very similar to that in https://bugs.webkit.org/show_bug.cgi?id=199688.
1531 Just keep track of which types have already been visited, and don't visit them again.
1533 No new tests as there is no intended functional change.
1535 * Modules/webgpu/WHLSL/WHLSLRecursiveTypeChecker.cpp:
1536 (WebCore::WHLSL::RecursiveTypeChecker::visit):
1537 (WebCore::WHLSL::checkRecursiveTypes):
1539 2019-07-17 Carlos Eduardo Ramalho <cadubentzen@gmail.com>
1541 Add missing #include's
1542 https://bugs.webkit.org/show_bug.cgi?id=199856
1544 Reviewed by Simon Fraser.
1546 The missing headers were found during development when unified sources changed.
1548 No new tests required. Adding #include's only.
1550 * Modules/speech/DOMWindowSpeechSynthesis.cpp:
1551 * Modules/webgpu/WHLSL/WHLSLAutoInitializeVariables.cpp:
1552 * Modules/webgpu/WHLSL/WHLSLPreserveVariableLifetimes.cpp:
1553 * Modules/webgpu/WHLSL/WHLSLPropertyResolver.cpp:
1554 * Modules/webgpu/WebGPUCommandEncoder.cpp:
1556 2019-07-17 Ryosuke Niwa <rniwa@webkit.org>
1558 Add inputmode to the feature status page
1559 https://bugs.webkit.org/show_bug.cgi?id=199884
1561 Reviewed by Wenson Hsieh.
1567 2019-07-17 Antoine Quint <graouts@apple.com>
1569 Disable Pointer Events prior to watchOS 6
1570 https://bugs.webkit.org/show_bug.cgi?id=199890
1571 <rdar://problem/53206113>
1573 Reviewed by Dean Jackson.
1575 * Configurations/FeatureDefines.xcconfig:
1577 2019-07-17 Antoine Quint <graouts@apple.com>
1579 Unable to bring up custom media controls on iOS for video.sina.cn
1580 https://bugs.webkit.org/show_bug.cgi?id=199889
1581 <rdar://problem/51883919>
1583 Reviewed by Dean Jackson.
1585 Videos on video.sina.cn by default have the "controls" attribute and are set not to autoplay. This means that the original state
1586 of the media controls are set to show the built-in media controls and also show the prominent play button to begin playback. The
1587 display of the play button also requires a tap gesture recognizer, which calls preventDefault() when the "touchend" is received
1588 to prevent double-tap-to-zoom, but also has the side-effect of preventing a "click" event from being dispatched for a tap.
1590 The video.sina.cn code would eventually remove the "controls" attribute, which would make the built-in media controls not visible,
1591 but still participate in hit-testing because we keep the shadow DOM around in order to potentially show the Airplay or picture-in-picture
1592 placards. Additionally, we wouldn't disable the tap gesture recognizer when the "controls" attribute was removed.
1594 We now ensure that both gesture recognizers used by iOS inline media controls are only enabled when media controls are visible.
1596 Test: media/modern-media-controls/media-controller/ios/media-controller-allows-click-over-video-with-no-controls.html
1598 * Modules/modern-media-controls/controls/ios-inline-media-controls.js:
1599 (IOSInlineMediaControls.prototype.set showsStartButton):
1600 (IOSInlineMediaControls.prototype.get visible):
1601 (IOSInlineMediaControls.prototype.set visible):
1602 (IOSInlineMediaControls.prototype._updateGestureRecognizers):
1603 (IOSInlineMediaControls.prototype._tapGestureRecognizerStateDidChange):
1604 (IOSInlineMediaControls.prototype._pinchGestureRecognizerStateDidChange):
1606 2019-07-17 Christopher Reid <chris.reid@sony.com>
1608 Bytecode cache should use FileSystem
1609 https://bugs.webkit.org/show_bug.cgi?id=199759
1611 Reviewed by Yusuke Suzuki.
1613 * platform/SharedBuffer.cpp:
1615 2019-07-17 Zalan Bujtas <zalan@apple.com>
1617 Unable to tap buttons at top of Wells Fargo app’s Payees screen
1618 https://bugs.webkit.org/show_bug.cgi?id=199846
1619 <rdar://problem/48112220>
1621 Reviewed by Simon Fraser.
1623 This patch restores the stop-before-reaching-the-body heuristics for the approximate node finding feature only.
1624 (WebFrame::approximateNodeAtViewportLocation mistakenly relied on this odd quirk of nodeRespondingToClickEvents.)
1627 * page/ios/FrameIOS.mm:
1628 (WebCore::Frame::approximateNodeAtViewportLocationLegacy):
1630 2019-07-17 Simon Fraser <simon.fraser@apple.com>
1632 [iOS WK2] Avoid lots of compositing backing store for offscreen position:fixed descendants
1633 https://bugs.webkit.org/show_bug.cgi?id=199819
1634 rdar://problem/51977416
1636 Reviewed by Zalan Bujtas.
1638 There was code to ensure that position:fixed layers and their descendants don't have their
1639 backing store detached (thrown away to save memory when offscreen), since that can trigger
1640 flashing when async scrolling moves layers in and out of the visual viewport.
1642 However, some pages have many descendant layers of positon:fixed which entrain backing
1643 store, because, for example, they have a stack of translated-offscreen menu elements inside
1644 a fixed header, and those elements may composite because of overflow:scroll.
1646 To avoid using too much memory on such pages, allow backing store detachment for fixed
1647 layers that are outside the layout viewport. Add a flag to RenderLayer that's set for layers
1648 which are fixed, or descendants of fixed, and consult that flag in
1649 updateAllowsBackingStoreDetaching(). The logic there is similar to
1650 RenderLayerCompositor::requiresCompositingForPosition().
1652 I considered allowing all position:fixed to composite (since this patch would keep most of
1653 the memory saving), but historically we've avoided compositing out-of-viewport
1654 position:fixed because it's quite common to have them with negative z-index, and compositing
1655 those has significant compositing knock-on effects.
1657 GraphicsLayer flushing no longer needs to track the viewport-constrained status of layers,
1658 so remove that code.
1660 This patch removes backing-store pinning for sticky layers.
1661 scrolling/ios/reconcile-layer-position-recursive.html then revealed a bug in
1662 ScrollingStateStickyNode::reconcileLayerPositionForViewportRect(), which was assuming that
1663 the sticky element was scrolled by the page, causing a bad layer position to get sync'd onto
1664 the layer. Fixed by copying code from ScrollingTreeStickyNode that is smarter about
1665 computing layer positions.
1667 This patch fixes jetsams on kmart.com.au in iOS 13 beta.
1669 Test: compositing/layer-creation/fixed-position-descendants-out-of-view.html
1671 * page/scrolling/ScrollingStateStickyNode.cpp:
1672 (WebCore::ScrollingStateStickyNode::computeLayerPosition const):
1673 (WebCore::ScrollingStateStickyNode::reconcileLayerPositionForViewportRect):
1674 * page/scrolling/ScrollingStateStickyNode.h:
1675 * platform/graphics/GraphicsLayer.h:
1676 (WebCore::GraphicsLayer::setAllowsBackingStoreDetaching):
1677 (WebCore::GraphicsLayer::allowsBackingStoreDetaching const):
1678 (WebCore::GraphicsLayer::setIsViewportConstrained): Deleted.
1679 (WebCore::GraphicsLayer::isViewportConstrained const): Deleted.
1680 (WebCore::GraphicsLayer::setCanDetachBackingStore): Deleted.
1681 (WebCore::GraphicsLayer::canDetachBackingStore const): Deleted.
1682 * platform/graphics/ca/GraphicsLayerCA.cpp:
1683 (WebCore::GraphicsLayerCA::GraphicsLayerCA):
1684 (WebCore::GraphicsLayerCA::recursiveVisibleRectChangeRequiresFlush const):
1685 (WebCore::GraphicsLayerCA::setVisibleAndCoverageRects):
1686 (WebCore::GraphicsLayerCA::recursiveCommitChanges):
1687 (WebCore::GraphicsLayerCA::updateCoverage):
1688 (WebCore::GraphicsLayerCA::createTransformAnimationsFromKeyframes):
1689 * platform/graphics/ca/GraphicsLayerCA.h:
1690 * rendering/RenderLayer.cpp:
1691 (WebCore::RenderLayer::RenderLayer):
1692 (WebCore::RenderLayer::updateLayerPositions):
1693 (WebCore::outputPaintOrderTreeLegend):
1694 (WebCore::outputPaintOrderTreeRecursive):
1695 * rendering/RenderLayer.h:
1696 * rendering/RenderLayerBacking.cpp:
1697 (WebCore::RenderLayerBacking::updateAllowsBackingStoreDetaching):
1698 (WebCore::RenderLayerBacking::updateOverflowControlsLayers):
1699 (WebCore::RenderLayerBacking::setIsScrollCoordinatedWithViewportConstrainedRole): Deleted.
1700 * rendering/RenderLayerBacking.h:
1701 * rendering/RenderLayerCompositor.cpp:
1702 (WebCore::RenderLayerCompositor::computeCompositingRequirements):
1703 (WebCore::RenderLayerCompositor::requiresCompositingForPosition const):
1704 (WebCore::RenderLayerCompositor::updateOverflowControlsLayers):
1705 (WebCore::RenderLayerCompositor::updateScrollCoordinationForLayer):
1707 2019-07-17 Chris Dumez <cdumez@apple.com>
1709 Avoid unnecessary copy of monitors under DisplayRefreshMonitorManager::displayWasUpdated()
1710 https://bugs.webkit.org/show_bug.cgi?id=199847
1712 Reviewed by Said Abou-Hallawa.
1714 There is always one one monitor in m_monitors for a given displayID so there is
1715 no need to copy the monitors inside DisplayRefreshMonitorManager::displayWasUpdated()
1716 and no need to iterate over all of them. This patch also factors a bit differently
1717 so that it becomes more obvious.
1719 * platform/graphics/DisplayRefreshMonitorManager.cpp:
1720 (WebCore::DisplayRefreshMonitorManager::createMonitorForClient):
1721 (WebCore::DisplayRefreshMonitorManager::unregisterClient):
1722 (WebCore::DisplayRefreshMonitorManager::displayDidRefresh):
1723 (WebCore::DisplayRefreshMonitorManager::displayWasUpdated):
1724 (WebCore::DisplayRefreshMonitorManager::findMonitorForDisplay const):
1725 (WebCore::DisplayRefreshMonitorManager::monitorForDisplay const):
1726 * platform/graphics/DisplayRefreshMonitorManager.h:
1728 2019-07-17 Sihui Liu <sihui_liu@apple.com>
1730 IndexedDB: error in starting version change transaction may be neglected
1731 https://bugs.webkit.org/show_bug.cgi?id=199818
1732 <rdar://problem/52925738>
1734 Reviewed by Brady Eidson.
1736 For version change transaction, IDBServer didn't wait the result of beginTransaction on the background thread
1737 before giving the IDBClient the result of open request. In this case, beginTransaction may fail to update the
1738 DatabaseVersion in database file or set m_originalDatabaseInfoBeforeVersionChange, but the transaction was
1739 marked as started. When we later set m_databaseInfo with m_originalDatabaseInfoBeforeVersionChange,
1740 m_databaseInfo could become nullptr.
1742 To write a test for this, we will need to simulate an SQLite error. I manually tested this by crafting the
1743 SQLiteStatement in beginTransaction, making it an invalid statement, and verified that error event, instead of
1744 ungradeneeded event is dispatched to the IDBRequest.
1746 * Modules/indexeddb/server/UniqueIDBDatabase.cpp:
1747 (WebCore::IDBServer::UniqueIDBDatabase::startVersionChangeTransaction):
1748 (WebCore::IDBServer::UniqueIDBDatabase::performStartVersionChangeTransaction):
1749 (WebCore::IDBServer::UniqueIDBDatabase::didPerformStartVersionChangeTransaction):
1750 (WebCore::IDBServer::UniqueIDBDatabase::beginTransactionInBackingStore): Deleted.
1751 * Modules/indexeddb/server/UniqueIDBDatabase.h:
1753 2019-07-17 Daniel Bates <dabates@apple.com>
1755 Typing into a cell in a Google Sheet lags behind by one character
1756 https://bugs.webkit.org/show_bug.cgi?id=199587
1757 <rdar://problem/51616845>
1759 Reviewed by Brent Fulgham.
1761 Add a Google Sheets quirk. Put all DOM timers scheduled from keydown and keypress event listeners
1762 into a holding tank. The timers continue to tick, but are barred from executing their action until
1763 the next text insertion or deletion or 32 ms (on device) have elapsed, whichever is sooner. We only
1764 allocate a holding tank once per document, only if the quirk is active, and this allocation is done
1765 when the document schedules a timer on keydown or keypress. The holding tank lives for the lifetime
1768 The story behind the quirk:
1770 On keypress Google Sheets schedules timers and expects that a DOM update will occur (i.e. text
1771 will be inserted or deleted) within the same event loop iteration as the dispatched keypress. The
1772 UI Events spec. [1] makes no such guarantee of when a DOM update must occur in relation to the keypress
1773 event. It could happen in the same event loop iteration as the key press (as Google expects), the
1774 next iteration, 500ms later, 2 minutes later, etc. What the spec does guarantee is that by the time
1775 a DOM input event is dispatched that the DOM will be updated. And this is the solution to the problem
1776 Google Sheets is trying to solve, but is doing so using pre-IE 9 technology (though similar
1777 functionality was available via onpropertychange in IE < 9).
1779 See also <https://github.com/w3c/uievents/issues/238>, which is tracking a spec. text update for
1782 Test: fast/events/ios/dom-update-on-keydown-quirk.html
1784 [1] <https://w3c.github.io/uievents/> (Editor's Draft, 14 October 2018)
1787 * WebCore.xcodeproj/project.pbxproj:
1788 Add some files to the project.
1791 (WebCore::Document::domTimerHoldingTank): Added.
1793 (WebCore::Document::domTimerHoldingTankIfExists): Added.
1795 * page/DOMTimer.cpp:
1796 (WebCore::DOMTimer::install): Put the newly instantiated timer into the holding tank.
1797 (WebCore::DOMTimer::removeById): Remove the timer from the holding tank.
1798 (WebCore::DOMTimer::fired): Check if the timer is in the holding tank. If it is and it is a one-
1799 shot timer then schedule it for the next event loop iteration. If it's a repeating timer just
1800 let it continue ticking. Otherwise, do what we no now and execute the timer's action. The reason
1801 we do not suspend timers in the holding tank is because:
1802 1. Far out timers (Google Sheets registers timers as far out as 5 minutes!) are not penalized.
1803 Though smart supension logic could avoid this. See (3).
1805 2. Empirical observations indicate that the keyboard will perform the insertion or deletion
1806 reasonably quickly (not the same event loop iteration as the keydown, but within two iterations out).
1807 So, the timers in the holding tank are short-lived.
1809 3. Simplifies the code. There is no need to keep additional bookkeeping to track multiple timer
1810 suspension reasons (timers currently can only have one suspension reason) or alternatively defer
1811 scheduling a timer until a later time and computing a new "fair" firing time when scheduled.
1812 * page/EventHandler.cpp:
1813 (WebCore::EventHandler::internalKeyEvent): Place a token on the stack to put all DOM timers
1814 scheduled on keydown and keypress into the holding tank if the quirk is enabled.
1816 (WebCore::Quirks::needsDeferKeyDownAndKeyPressTimersUntilNextEditingCommand const): Added.
1818 * page/Settings.yaml: Added setting so that this quirk can be enabled from a layout test. This setting
1819 also lets us enable the quirk for all sites or for certain third-party apps if desired.
1820 * page/ios/DOMTimerHoldingTank.cpp: Added.
1821 (WebCore::DOMTimerHoldingTank::DOMTimerHoldingTank):
1822 (WebCore::DOMTimerHoldingTank::add):
1823 (WebCore::DOMTimerHoldingTank::remove):
1824 (WebCore::DOMTimerHoldingTank::contains):
1825 (WebCore::DOMTimerHoldingTank::removeAll):
1826 (WebCore::DOMTimerHoldingTank::stopExceededMaximumHoldTimer):
1827 * page/ios/DOMTimerHoldingTank.h: Added.
1828 (WebCore::DeferDOMTimersForScope::DeferDOMTimersForScope):
1829 (WebCore::DeferDOMTimersForScope::~DeferDOMTimersForScope):
1830 (WebCore::DeferDOMTimersForScope::isDeferring):
1832 2019-07-17 Darin Adler <darin@apple.com>
1834 No need for isURLAllowed function in Frame
1835 https://bugs.webkit.org/show_bug.cgi?id=120266
1837 Reviewed by Alex Christensen.
1839 * css/SelectorCheckerTestFunctions.h:
1840 (WebCore::matchesFullScreenPseudoClass): Use is<> instead of
1841 the isFrameElementBase function. Also updated includes and spacing.
1844 (WebCore::Document::resolveStyle): Shortened an old, long comment
1845 that happened to mention HTMLFrameElementBase::isURLAllowed by name.
1847 * dom/Element.h: Deleted isFrameElementBase.
1849 * html/HTMLAppletElement.cpp:
1850 (WebCore::HTMLAppletElement::updateWidget): Tweaked comment.
1852 * html/HTMLBodyElement.cpp:
1853 (WebCore::HTMLBodyElement::insertedIntoAncestor): Moved comment from
1854 here to didFinishInsertingNode, since the code it's talking about is there.
1855 Also got rid of unnecessary local variable.
1856 (WebCore::HTMLBodyElement::didFinishInsertingNode): Moved comment here
1857 and elaborated a bit. Also changed the code to copy margin attributes in a
1858 simpler way without requiring special getter functions. There is a slight
1859 behavior change here, but seems unlikely to affect any websites or tests.
1860 Would be good to add better test coverage, which I think should be done
1861 when fixing the problems described in the FIXME.
1863 * html/HTMLEmbedElement.cpp:
1864 (WebCore::HTMLEmbedElement::updateWidget): allowedToLoadFrameURL -> canLoadURL
1865 Also Tweaked comments.
1867 * html/HTMLFrameElement.cpp:
1868 (WebCore::HTMLFrameElement::rendererIsNeeded): isURLAllowed -> canLoad
1870 * html/HTMLFrameElementBase.cpp:
1871 (WebCore::HTMLFrameElementBase::HTMLFrameElementBase): Removed initial
1872 values for no-longer-needed data members.
1874 (WebCore::HTMLFrameElementBase::canLoadScriptURL const): Renamed from
1875 isURLAllowed. Also refactored so it now calls canLoadURL.
1876 (WebCore::HTMLFrameElementBase::canLoad const): Renamed from isURLAllowed.
1877 Also added FIXME comments about multiple ways that this code incorrectly
1878 handles the empty string. Might inspire some tests and future fixes.
1879 (WebCore::HTMLFrameElementBase::canLoadURL const): Renamed from isURLAllowed.
1880 Added an overload that takes a relative URL string largely to make this
1881 more parallel with the similar code in HTMLPlugInImageElement. Refactored
1882 to call canAddSubframe and isProhibitedSelfReference, removing the need
1883 to call Frame::isURLAllowed.
1884 (WebCore::HTMLFrameElementBase::openURL): isURLAllowed -> canLoad
1885 (WebCore::HTMLFrameElementBase::parseAttribute): Removed unneeded code
1886 to parse marginwidth, marginheight, and scrolling, which are now done
1887 as needed instead. Might have to add code back if we ever add code to
1888 make these cause dynamic updates when changed, but that was not the
1889 behavior of the existing code.
1890 (WebCore::HTMLFrameElementBase::scrollingMode const): Changed to call
1891 attributeWithoutSynchronization rather than reading from a data member.
1893 * html/HTMLFrameElementBase.h: Moved scrollingMode into the .cpp file.
1894 Removed marginWidth and marginHeight since they are no longer called.
1895 Made canContainRangeEndPoint override be private. Renamed isURLAllowed
1896 to canLoad/canLoadScriptURL/canLoadURL. Removed unneeded isFrameElementBase
1897 since is<HTMLFrameElementBase> already works without it. Removed unneeded
1898 data members m_scrolling, m_marginWidth, and m_marginHeight.
1900 * html/HTMLFrameOwnerElement.cpp:
1901 (WebCore::HTMLFrameOwnerElement::HTMLFrameOwnerElement): Moved initialization
1902 of m_contentFrame and m_sandboxFlags to the header.
1903 (WebCore::HTMLFrameOwnerElement::clearContentFrame): 0 -> nullptr
1904 (WebCore::HTMLFrameOwnerElement::canAddSubframe const): Added. Code was moved
1905 here from multiple functions that enforced the subframe count limit.
1906 (WebCore::HTMLFrameOwnerElement::isProhibitedSelfReference const): Added.
1907 Moved code here from Frame::isURLAllowed so it's easy to use as needed.
1909 * html/HTMLFrameOwnerElement.h: Renamed isURLAllowed to canLoadScriptURL.
1910 Added protected canAddSubframe and isProhibitedSelfReference functions.
1911 Initialized the data members here.
1913 * html/HTMLIFrameElement.cpp:
1914 (WebCore::HTMLIFrameElement::sandbox): Added braces for multi-line if statement.
1915 (WebCore::HTMLIFrameElement::parseAttribute): Removed unneeded storage of the
1916 allow attribute value. Added missing code to invalidate m_featurePolicy when
1917 the allow attribute changes. This could change behavior, but unlikely to affect
1918 any websites of tests. Would be good to add a test for this.
1919 (WebCore::HTMLIFrameElement::rendererIsNeeded): isURLAllowed -> canLoad. Also
1920 reordered the code to be more efficient for display: none (why not?).
1921 (WebCore::HTMLIFrameElement::featurePolicy const): Use attributeWithoutSynchronization
1922 to read the allow attribute rather than using a data member.
1924 * html/HTMLIFrameElement.h: Removed the unneeded allow function member and
1925 m_allow data member since DOM reflection takes care of the function and the
1926 featurePolicy function can get the attribute without a data member.
1928 * html/HTMLObjectElement.cpp:
1929 (WebCore::HTMLObjectElement::updateWidget): allowedToLoadFrameURL -> canLoadURL
1930 Also tweaked comments.
1932 * html/HTMLPlugInElement.cpp:
1933 (WebCore::HTMLPlugInElement::canLoadScriptURL const): Added. Code moved here
1934 from HTMLFrameOwnerElement::isURLAllowed. Added comment about why this is not
1935 necessarily correct and might need improvement.
1937 * html/HTMLPlugInElement.h: Added private override of canLoadScriptURL.
1939 * html/HTMLPlugInImageElement.cpp:
1940 (WebCore::HTMLPlugInImageElement::canLoadURL const): Renamed from
1941 allowedToLoadFrameURL and refactored into two functions to match the
1942 structure of the similar function in HTMLFrameElementBase. Also refactored
1943 to use canAddSubframe and isProhibitedSelfReference rather than Frame::isURLAllowed.
1944 (WebCore::HTMLPlugInImageElement::canLoadPlugInContent const): Renamed from
1945 allowedToLoadPluginContent.
1946 (WebCore::HTMLPlugInImageElement::requestObject): allowedToLoadPluginContent
1947 -> canLoadPlugInContent
1949 * html/HTMLPlugInImageElement.h: Renamed functions as mentioned above.
1951 * loader/SubframeLoader.cpp:
1952 (WebCore::SubframeLoader::requestFrame): isURLAllowed -> canLoadScriptURL
1954 * page/FocusController.h: Removed unneeded forward declaration of HTMLFrameOwnerElement.
1957 (WebCore::Frame::isURLAllowed const): Deleted.
1958 * page/Frame.h: Ditto.
1960 * page/FrameView.cpp:
1961 (WebCore::FrameView::init): Removed code to initialize m_margins, since nothing
1962 was looking at it after that.
1963 (WebCore::FrameView::setMarginWidth): Deleted.
1964 (WebCore::FrameView::setMarginHeight): Deleted.
1966 * page/FrameView.h: Removed marginWidth, marginHeight, setMarginWidth,
1967 setMarginHeight, and m_margins because there was no code anywhere calling
1968 marginWidth or marginHeight, and no other side effects of these functions.
1970 2019-07-17 Alex Christensen <achristensen@webkit.org>
1972 Add a runtime-disabled dialog element skeleton
1973 https://bugs.webkit.org/show_bug.cgi?id=199839
1975 Reviewed by Ryosuke Niwa.
1977 Beginning implementation of https://html.spec.whatwg.org/multipage/interactive-elements.html#the-dialog-element
1978 Covered by rebaselining WPT tests.
1981 * DerivedSources-input.xcfilelist:
1982 * DerivedSources-output.xcfilelist:
1983 * DerivedSources.make:
1985 * WebCore.xcodeproj/project.pbxproj:
1986 * bindings/js/WebCoreBuiltinNames.h:
1987 * html/HTMLDialogElement.cpp: Added.
1988 (WebCore::HTMLDialogElement::HTMLDialogElement):
1989 (WebCore::HTMLDialogElement::open):
1990 (WebCore::HTMLDialogElement::setOpen):
1991 (WebCore::HTMLDialogElement::returnValue):
1992 (WebCore::HTMLDialogElement::setReturnValue):
1993 (WebCore::HTMLDialogElement::show):
1994 (WebCore::HTMLDialogElement::showModal):
1995 (WebCore::HTMLDialogElement::close):
1996 * html/HTMLDialogElement.h: Added.
1997 * html/HTMLDialogElement.idl: Added.
1998 * html/HTMLTagNames.in:
1999 * page/RuntimeEnabledFeatures.h:
2000 (WebCore::RuntimeEnabledFeatures::setDialogElementEnabled):
2001 (WebCore::RuntimeEnabledFeatures::dialogElementEnabled const):
2003 2019-07-17 Youenn Fablet <youenn@apple.com>
2005 Hop to the main thread when doing logging in RealtimeIncomingVideoSourceCocoa
2006 https://bugs.webkit.org/show_bug.cgi?id=199865
2008 Reviewed by Darin Adler.
2010 LoggerHelper routines allow logging messages in system console and inspector console.
2011 These routines iterate through a Vector of log observers which is not thread safe.
2012 Document, the main log observer, also expects to be called on the main thread.
2013 Manually tested (writing a layout test for this would require more than 2 seconds).
2015 * platform/mediastream/mac/RealtimeIncomingVideoSourceCocoa.mm:
2016 (WebCore::RealtimeIncomingVideoSourceCocoa::OnFrame):
2018 2019-07-17 Commit Queue <commit-queue@webkit.org>
2020 Unreviewed, rolling out r247505.
2021 https://bugs.webkit.org/show_bug.cgi?id=199871
2023 "Caused failed ASSERT in stress test" (Requested by creid on
2028 "Bytecode cache should use FileSystem"
2029 https://bugs.webkit.org/show_bug.cgi?id=199759
2030 https://trac.webkit.org/changeset/247505
2032 2019-07-17 Youenn Fablet <youenn@apple.com>
2034 Add release logging to AudioSampleDataSource
2035 https://bugs.webkit.org/show_bug.cgi?id=199814
2036 <rdar://problem/53128140>
2038 Reviewed by Eric Carlson.
2040 Use release logging for AudioSampleDataSource in case of errors.
2041 This requires passing a logger and a logger identifier.
2042 For that purpose, pass a MediaStreamTrackPrivate to it constructor
2043 and reuse its logger and log identifier.
2045 For each log, hop to the main thread as release logging requires to be done in the main thread.
2046 Add specific handling for RealtimeOutgoingAudioSource as its MediaStreamTrack can change in case of replaceTrack.
2047 In that case, we change of logger when changing the MediaStreamTrackPrivate.
2049 No observable change of behavior, covered by existing tests.
2051 * platform/audio/mac/AudioSampleDataSource.h:
2052 * platform/audio/mac/AudioSampleDataSource.mm:
2053 (WebCore::AudioSampleDataSource::create):
2054 (WebCore::AudioSampleDataSource::AudioSampleDataSource):
2055 (WebCore::AudioSampleDataSource::setupConverter):
2056 (WebCore::AudioSampleDataSource::pushSamplesInternal):
2057 (WebCore::AudioSampleDataSource::pullSamplesInternal):
2058 (WebCore::AudioSampleDataSource::setLogger):
2059 (WebCore::AudioSampleDataSource::logChannel const):
2060 * platform/mediastream/RealtimeOutgoingAudioSource.cpp:
2061 (WebCore::RealtimeOutgoingAudioSource::setSource):
2062 * platform/mediastream/RealtimeOutgoingAudioSource.h:
2063 (WebCore::RealtimeOutgoingAudioSource::sourceUpdated):
2064 * platform/mediastream/mac/AudioTrackPrivateMediaStreamCocoa.cpp:
2065 (WebCore::AudioTrackPrivateMediaStreamCocoa::audioSamplesAvailable):
2066 * platform/mediastream/mac/AudioTrackPrivateMediaStreamCocoa.h:
2067 Instead of using its own logger/identifier member field, reuse its MediaStreamTrackPrivate.
2068 * platform/mediastream/mac/RealtimeOutgoingAudioSourceCocoa.cpp:
2069 (WebCore::RealtimeOutgoingAudioSourceCocoa::RealtimeOutgoingAudioSourceCocoa):
2070 (WebCore::RealtimeOutgoingAudioSourceCocoa::sourceUpdated):
2071 * platform/mediastream/mac/RealtimeOutgoingAudioSourceCocoa.h:
2072 * platform/mediastream/mac/WebAudioSourceProviderAVFObjC.mm:
2073 (WebCore::WebAudioSourceProviderAVFObjC::prepare):
2075 2019-07-17 Kenneth Russell <kbr@chromium.org>
2077 Make ANGLE work inside WebKit2's sandbox
2078 https://bugs.webkit.org/show_bug.cgi?id=199844
2080 Reviewed by Alex Christensen.
2082 Fix missing #import introduced at the last minute while removing
2083 some debugging code in the last patch.
2085 Tested with San Angeles WebGL demo and some conformance tests.
2087 * platform/graphics/cocoa/WebGLLayer.mm:
2089 2019-07-17 Andres Gonzalez <andresg_22@apple.com>
2091 Hang in WebCore::AccessibilityRenderObject::visiblePositionRangeForLine.
2092 https://bugs.webkit.org/show_bug.cgi?id=199845
2093 <rdar://problem/53121017>
2095 Reviewed by Chris Fleizach.
2097 - Added check for upper bound NSNotFound for parameter to visiblePositionRangeForLine.
2098 - Fixed signed/unsigned issue for AXTextMarkerRangeForLine attribute.
2100 * accessibility/AccessibilityRenderObject.cpp:
2101 (WebCore::AccessibilityRenderObject::visiblePositionRangeForLine const):
2102 * accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
2103 (-[WebAccessibilityObjectWrapper accessibilityAttributeValue:forParameter:]):
2105 2019-07-17 Rob Buis <rbuis@igalia.com>
2107 Add referrerpolicy attribute support for <script> elements
2108 https://bugs.webkit.org/show_bug.cgi?id=185550
2110 Reviewed by Youenn Fablet.
2112 This patch adds 'referrerpolicy' attribute support for script elements.
2113 If set, the value is restricted to the ReferrerPolicy enum, and
2114 if valid it is used for the script fetch.
2115 If not set or invalid, the current behavior is kept.
2117 Tests: http/tests/referrer-policy-script/no-referrer-when-downgrade/cross-origin-http-http.html
2118 http/tests/referrer-policy-script/no-referrer-when-downgrade/cross-origin-http.https.html
2119 http/tests/referrer-policy-script/no-referrer-when-downgrade/same-origin.html
2120 http/tests/referrer-policy-script/no-referrer/cross-origin-http-http.html
2121 http/tests/referrer-policy-script/no-referrer/cross-origin-http.https.html
2122 http/tests/referrer-policy-script/no-referrer/same-origin.html
2123 http/tests/referrer-policy-script/origin-when-cross-origin/cross-origin-http-http.html
2124 http/tests/referrer-policy-script/origin-when-cross-origin/cross-origin-http.https.html
2125 http/tests/referrer-policy-script/origin-when-cross-origin/same-origin.html
2126 http/tests/referrer-policy-script/origin/cross-origin-http-http.html
2127 http/tests/referrer-policy-script/origin/cross-origin-http.https.html
2128 http/tests/referrer-policy-script/origin/same-origin.html
2129 http/tests/referrer-policy-script/same-origin/cross-origin-http-http.html
2130 http/tests/referrer-policy-script/same-origin/cross-origin-http.https.html
2131 http/tests/referrer-policy-script/same-origin/same-origin.html
2132 http/tests/referrer-policy-script/strict-origin-when-cross-origin/cross-origin-http-http.html
2133 http/tests/referrer-policy-script/strict-origin-when-cross-origin/cross-origin-http.https.html
2134 http/tests/referrer-policy-script/strict-origin-when-cross-origin/same-origin.html
2135 http/tests/referrer-policy-script/strict-origin/cross-origin-http-http.html
2136 http/tests/referrer-policy-script/strict-origin/cross-origin-http.https.html
2137 http/tests/referrer-policy-script/strict-origin/same-origin.html
2138 http/tests/referrer-policy-script/unsafe-url/cross-origin-http-http.html
2139 http/tests/referrer-policy-script/unsafe-url/cross-origin-http.https.html
2140 http/tests/referrer-policy-script/unsafe-url/same-origin.html
2142 * bindings/js/CachedScriptFetcher.cpp:
2143 (WebCore::CachedScriptFetcher::requestScriptWithCache const):
2144 * bindings/js/CachedScriptFetcher.h:
2145 (WebCore::CachedScriptFetcher::CachedScriptFetcher):
2146 * dom/InlineClassicScript.h:
2147 * dom/LoadableClassicScript.cpp:
2148 (WebCore::LoadableClassicScript::create):
2149 * dom/LoadableClassicScript.h:
2150 * dom/LoadableModuleScript.cpp:
2151 (WebCore::LoadableModuleScript::create):
2152 (WebCore::LoadableModuleScript::LoadableModuleScript):
2153 * dom/LoadableModuleScript.h:
2154 * dom/LoadableScript.h:
2155 (WebCore::LoadableScript::LoadableScript):
2156 * dom/ScriptElement.cpp:
2157 (WebCore::ScriptElement::requestClassicScript):
2158 (WebCore::ScriptElement::requestModuleScript):
2159 * dom/ScriptElement.h:
2160 * dom/ScriptElementCachedScriptFetcher.h:
2161 (WebCore::ScriptElementCachedScriptFetcher::ScriptElementCachedScriptFetcher):
2162 * html/HTMLIFrameElement.cpp:
2163 (WebCore::HTMLIFrameElement::referrerPolicyForBindings const):
2164 * html/HTMLScriptElement.cpp:
2165 (WebCore::HTMLScriptElement::setReferrerPolicyForBindings):
2166 (WebCore::HTMLScriptElement::referrerPolicyForBindings const):
2167 (WebCore::HTMLScriptElement::referrerPolicy const):
2168 * html/HTMLScriptElement.h:
2169 * html/HTMLScriptElement.idl:
2170 * html/parser/CSSPreloadScanner.cpp:
2171 (WebCore::CSSPreloadScanner::emitRule):
2172 * html/parser/HTMLPreloadScanner.cpp:
2173 (WebCore::TokenPreloadScanner::StartTagScanner::createPreloadRequest):
2174 (WebCore::TokenPreloadScanner::StartTagScanner::processAttribute):
2175 * html/parser/HTMLResourcePreloader.cpp:
2176 (WebCore::PreloadRequest::resourceRequest):
2177 * html/parser/HTMLResourcePreloader.h:
2178 (WebCore::PreloadRequest::PreloadRequest):
2179 * platform/ReferrerPolicy.cpp:
2180 (WebCore::referrerPolicyToString):
2181 * platform/ReferrerPolicy.h:
2182 * svg/SVGScriptElement.h:
2184 2019-07-16 Christopher Reid <chris.reid@sony.com>
2186 Bytecode cache should use FileSystem
2187 https://bugs.webkit.org/show_bug.cgi?id=199759
2189 Reviewed by Yusuke Suzuki.
2191 * platform/SharedBuffer.cpp:
2193 2019-07-16 Simon Fraser <simon.fraser@apple.com>
2195 Fix includes and make a constructor explicit to avoid unified-sources-triggered test failures
2196 https://bugs.webkit.org/show_bug.cgi?id=199849
2198 Reviewed by Wenson Hsieh.
2200 webkit.org/b/199587 triggered some unexpected test failures because of unified sources include
2201 order. The TextStream operator in ScrollingTreeOverflowScrollingNode::dumpProperties() was invoking
2202 the implicit constructor of TransformOperations.
2204 * page/scrolling/ScrollingTreeOverflowScrollingNode.cpp: Include TextStream.h
2205 * platform/graphics/transforms/TransformOperations.h: Make ctor explicit.
2207 2019-07-16 Antoine Quint <graouts@apple.com>
2209 REGRESSION: fast/events/watchos/double-tap-to-zoom* tests crashing in com.apple.WebKit: -[WKContentView(WKInteraction) _singleTapDidReset:]
2210 https://bugs.webkit.org/show_bug.cgi?id=199840
2211 <rdar://problem/52856531>
2213 Reviewed by Wenson Hsieh.
2215 Fix a build error that occured when building for the watchOS Simulator in the debug configuration.
2217 * rendering/RenderLayerCompositor.cpp:
2218 (WebCore::RenderLayerCompositor::updateScrollingNodeForScrollingProxyRole):
2220 2019-07-16 Justin Fan <justin_fan@apple.com>
2222 [WebGPU] Convert GPUComputePipeline LOG statements to generate GPUErrors
2223 https://bugs.webkit.org/show_bug.cgi?id=199773
2225 Reviewed by Myles C. Maxfield.
2227 Generate a GPUError if GPUComputePipeline creation fails.
2228 Rework the WHLSL test harness to take advantage of this.
2230 Test: webgpu/compute-pipeline-errors.html
2232 * Modules/webgpu/WebGPUComputePipelineDescriptor.cpp:
2233 (WebCore::WebGPUComputePipelineDescriptor::tryCreateGPUComputePipelineDescriptor const):
2234 * Modules/webgpu/WebGPUComputePipelineDescriptor.h:
2235 * Modules/webgpu/WebGPUDevice.cpp:
2236 (WebCore::WebGPUDevice::createComputePipeline const):
2237 * platform/graphics/gpu/GPUBuffer.h:
2238 * platform/graphics/gpu/GPUComputePipeline.h:
2239 Now inherits from GPUObjectBase, though this isn't super useful yet
2240 since GPUComputePipeline itself does not expose any operations that can error.
2241 * platform/graphics/gpu/GPUDevice.cpp:
2242 (WebCore::GPUDevice::tryCreateComputePipeline const):
2243 * platform/graphics/gpu/GPUDevice.h:
2244 * platform/graphics/gpu/cocoa/GPUBufferMetal.mm:
2245 (WebCore::GPUBuffer::validateBufferUsage):
2246 * platform/graphics/gpu/cocoa/GPUComputePipelineMetal.mm:
2247 (WebCore::trySetMetalFunctions):
2248 (WebCore::trySetFunctions):
2249 (WebCore::convertComputePipelineDescriptor):
2250 (WebCore::tryCreateMTLComputePipelineState):
2251 (WebCore::GPUComputePipeline::tryCreate):
2252 (WebCore::GPUComputePipeline::GPUComputePipeline):
2254 2019-07-11 Justin Fan <justin_fan@apple.com>
2256 [WebGPU] Move error scopes out of GPUDevice for more portable error generation
2257 https://bugs.webkit.org/show_bug.cgi?id=199740
2259 Reviewed by Myles C. Maxfield.
2261 Move error generation into a separate RefCounted class to allow GPU objects to generate
2262 errors independent of any GPUDevice.
2263 Create GPUObjectBase to delegate error generation and refactor GPUBuffer to inherit from GPUObjectBase.
2265 No behavior change or new tests. Error scopes covered by error-scopes-test.html.
2267 * Modules/webgpu/WebGPUDevice.cpp:
2268 (WebCore::WebGPUDevice::WebGPUDevice): Now creates a GPUErrorGenerator.
2269 (WebCore::WebGPUDevice::createBuffer const): Pass the GPUErrorGenerator to any created GPUBuffer.
2270 (WebCore::WebGPUDevice::createBufferMapped const): Ditto.
2271 (WebCore::WebGPUDevice::popErrorScope): Shouldn't be const. Can just ask for the GPUError rather than passing a lambda.
2272 (WebCore::WebGPUDevice::pushErrorScope const): Deleted.
2273 (WebCore::WebGPUDevice::popErrorScope const): Deleted.
2274 * Modules/webgpu/WebGPUDevice.h:
2275 (WebCore::WebGPUDevice::pushErrorScope):
2277 * WebCore.xcodeproj/project.pbxproj:
2278 * platform/graphics/gpu/GPUBuffer.h: Now inherits from GPUObjectBase for error generation ease.
2279 * platform/graphics/gpu/GPUDevice.cpp:
2280 (WebCore::GPUDevice::tryCreateBuffer): Ensure GPUBuffers reference the GPUErrorGenerator.
2281 (WebCore::GPUDevice::pushErrorScope): Deleted. No longer needed here.
2282 (WebCore::GPUDevice::popErrorScope): Deleted.
2283 (WebCore::GPUDevice::registerError): Deleted.
2284 * platform/graphics/gpu/GPUDevice.h: Move error scope logic out.
2285 * platform/graphics/gpu/GPUErrorGenerator.cpp: Added.
2286 (WebCore::GPUErrorGenerator::pushErrorScope):
2287 (WebCore::GPUErrorGenerator::popErrorScope):
2288 (WebCore::GPUErrorGenerator::generateError):
2289 * platform/graphics/gpu/GPUErrorGenerator.h: Added.
2290 (WebCore::GPUErrorGenerator::create):
2291 * platform/graphics/gpu/GPUObjectBase.h: Added.
2292 (WebCore::GPUObjectBase::generateError):
2293 (WebCore::GPUObjectBase::GPUObjectBase):
2294 * platform/graphics/gpu/cocoa/GPUBufferMetal.mm: Use the GPUErrorGenerator directly during buffer creation.
2295 (WebCore::GPUBuffer::validateBufferUsage):
2296 (WebCore::GPUBuffer::tryCreate):
2297 (WebCore::GPUBuffer::GPUBuffer):
2298 * platform/graphics/gpu/cocoa/GPUQueueMetal.mm:
2299 (WebCore::GPUQueue::submit): Prevent possible null dereference.
2301 2019-07-10 Justin Fan <justin_fan@apple.com>
2303 [WebGPU] Implement GPUError and error scopes
2304 https://bugs.webkit.org/show_bug.cgi?id=199655
2306 Reviewed by Myles C. Maxfield.
2308 Add GPUErrorFilter, GPUError, and the ability to push and pop error scopes on a GPUDevice.
2309 This will allow us to check for WebGPU errors from JavaScript.
2310 Currently, only some GPUBuffer creation errors are reported for testing; more errors will follow in later patches.
2312 Test: webgpu/error-scopes-test.html
2314 * Modules/webgpu/GPUErrorFilter.idl: Added.
2315 * Modules/webgpu/GPUOutOfMemoryError.idl: Added.
2316 * Modules/webgpu/GPUValidationError.idl: Added.
2317 * Modules/webgpu/WebGPUDevice.cpp:
2318 (WebCore::WebGPUDevice::pushErrorScope):
2319 (WebCore::WebGPUDevice::popErrorScope): Resolve the Promise here, since GPUDevice shouldn't worry about DOM layer.
2320 * Modules/webgpu/WebGPUDevice.h:
2321 * Modules/webgpu/WebGPUDeviceErrorScopes.cpp: Added.
2322 (WebCore::WebGPUDeviceErrorScopes::pushErrorScope): Delegates to WebGPUDevice.
2323 (WebCore::WebGPUDeviceErrorScopes::popErrorScope): Ditto.
2324 * Modules/webgpu/WebGPUDeviceErrorScopes.h: Added.
2325 * Modules/webgpu/WebGPUDeviceErrorScopes.idl: Added.
2326 * platform/graphics/gpu/GPUBuffer.h:
2327 * platform/graphics/gpu/GPUDevice.cpp: The actual error stack lives here.
2328 (WebCore::GPUDevice::pushErrorScope):
2329 (WebCore::GPUDevice::popErrorScope): Calls a callback with a GPUError, if any.
2330 (WebCore::GPUDevice::registerError): Actually creates GPUErrors.
2331 * platform/graphics/gpu/GPUDevice.h:
2332 * platform/graphics/gpu/GPUError.cpp: Added.
2333 (WebCore::createError): Factory function for various error types.
2334 * platform/graphics/gpu/GPUError.h: Added.
2335 * platform/graphics/gpu/GPUErrorFilter.h: Added.
2336 * platform/graphics/gpu/GPUOutOfMemoryError.h: Added.
2337 (WebCore::GPUOutOfMemoryError::create):
2338 * platform/graphics/gpu/GPUValidationError.cpp: Added.
2339 (WebCore::GPUValidationError::create):
2340 (WebCore::GPUValidationError::GPUValidationError):
2341 * platform/graphics/gpu/GPUValidationError.h: Added.
2342 (WebCore::GPUValidationError::message const):
2343 * platform/graphics/gpu/cocoa/GPUBufferMetal.mm:
2344 (WebCore::GPUBuffer::validateBufferUsage): Create an actual GPUValidationError :)
2345 (WebCore::GPUBuffer::tryCreate): Create GPUOutOfMemoryErrors where appropriate.
2347 Add file/name references:
2349 * DerivedSources-input.xcfilelist:
2350 * DerivedSources-output.xcfilelist:
2351 * DerivedSources.make:
2353 * WebCore.xcodeproj/project.pbxproj:
2354 * bindings/js/WebCoreBuiltinNames.h:
2357 * Modules/webgpu/WHLSL/WHLSLCheckTextureReferences.cpp:
2359 2019-07-16 Commit Queue <commit-queue@webkit.org>
2361 Unreviewed, rolling out r247493.
2362 https://bugs.webkit.org/show_bug.cgi?id=199841
2364 "The new whlsl-for-loop.html test is failing on the bots"
2365 (Requested by rmorisset on #webkit).
2369 "[WHLSL] Desugar for loops and while loops"
2370 https://bugs.webkit.org/show_bug.cgi?id=199726
2371 https://trac.webkit.org/changeset/247493
2373 2019-07-16 Per Arne Vollan <pvollan@apple.com>
2375 Prewarm font cache with more fonts
2376 https://bugs.webkit.org/show_bug.cgi?id=199283
2378 Reviewed by Myles C. Maxfield.
2380 Use the existing prewarm method in FontCache to prewarm the cache with a set of font family names.
2381 The font list consists of some of the fonts used by the 10 most popular sites on Alexa top sites.
2382 This is a confirmed improvement in page load time.
2384 No new tests, covered by existing tests.
2386 * page/ProcessWarming.cpp:
2387 (WebCore::ProcessWarming::prewarmGlobally):
2388 * platform/graphics/FontCache.cpp:
2389 (WebCore::FontCache::prewarmGlobally):
2390 * platform/graphics/FontCache.h:
2391 * platform/graphics/cocoa/FontCacheCoreText.cpp:
2392 (WebCore::FontCache::prewarmGlobally):
2394 2019-07-16 Robin Morisset <rmorisset@apple.com>
2396 [WHLSL] Desugar for loops and while loops
2397 https://bugs.webkit.org/show_bug.cgi?id=199726
2399 Reviewed by Myles C. Maxfield.
2401 This patch makes loops behave a lot more similarly to the spec.
2402 In particular, while loops are just degenerate for loops, so they can be desugared in the parser.
2403 And for loops can themselves be heavily desugared, into having no initializer, one condition and one increment
2404 by putting any initializer in a block around the loop, putting true in the condition if there is none, and putting any litteral in the increment if there is none.
2406 Debugging this patch revealed an unrelated bug where the parser would generate an empty comma expression on ";", which the checker would then choke on.
2407 The parser now generates an empty block in that case (as specified in the spec), which fixes the problem.
2409 Finally, while updating the NameResolver for the new structure of loops I realized that it was needlessly generating extra NameContext.
2410 They are not expected by the spec, that forbids naked variable declarations outside of blocks anyway.
2411 So I removed this unnecessary work, and fixed the parser to correctly forbid such naked variable declarations.
2413 * Modules/webgpu/WHLSL/AST/WHLSLAST.h:
2414 * Modules/webgpu/WHLSL/AST/WHLSLForLoop.h:
2415 (WebCore::WHLSL::AST::ForLoop::ForLoop):
2416 (WebCore::WHLSL::AST::ForLoop::condition):
2417 (WebCore::WHLSL::AST::ForLoop::increment):
2418 * Modules/webgpu/WHLSL/AST/WHLSLStatement.h:
2419 (WebCore::WHLSL::AST::Statement::isVariableDeclarationsStatement const):
2420 * Modules/webgpu/WHLSL/AST/WHLSLWhileLoop.h: Removed.
2421 * Modules/webgpu/WHLSL/Metal/WHLSLFunctionWriter.cpp:
2422 (WebCore::WHLSL::Metal::FunctionDefinitionWriter::visit):
2423 * Modules/webgpu/WHLSL/WHLSLASTDumper.cpp:
2424 (WebCore::WHLSL::ASTDumper::visit):
2425 * Modules/webgpu/WHLSL/WHLSLASTDumper.h:
2426 * Modules/webgpu/WHLSL/WHLSLChecker.cpp:
2427 (WebCore::WHLSL::Checker::visit):
2428 * Modules/webgpu/WHLSL/WHLSLNameResolver.cpp:
2429 (WebCore::WHLSL::NameResolver::visit):
2430 * Modules/webgpu/WHLSL/WHLSLNameResolver.h:
2431 * Modules/webgpu/WHLSL/WHLSLParser.cpp:
2432 (WebCore::WHLSL::Parser::parseBlockBody):
2433 (WebCore::WHLSL::Parser::parseForLoop):
2434 (WebCore::WHLSL::Parser::parseWhileLoop):
2435 (WebCore::WHLSL::Parser::parseStatement):
2436 (WebCore::WHLSL::Parser::parseEffectfulExpression):
2437 * Modules/webgpu/WHLSL/WHLSLParser.h:
2438 * Modules/webgpu/WHLSL/WHLSLStatementBehaviorChecker.cpp:
2439 * Modules/webgpu/WHLSL/WHLSLVisitor.cpp:
2440 (WebCore::WHLSL::Visitor::visit):
2441 * Modules/webgpu/WHLSL/WHLSLVisitor.h:
2442 * WebCore.xcodeproj/project.pbxproj:
2444 2019-07-16 Tim Horton <timothy_horton@apple.com>
2446 NSTextFinder holes don't scroll with the page
2447 https://bugs.webkit.org/show_bug.cgi?id=199815
2448 <rdar://problem/52280514>
2450 Reviewed by Simon Fraser.
2452 * rendering/ScrollAlignment.h:
2454 2019-07-16 Youenn Fablet <youenn@apple.com>
2456 [iOS] Make sure the first camera device in the list is the front camera
2457 https://bugs.webkit.org/show_bug.cgi?id=199811
2458 <rdar://problem/53125157>
2460 Reviewed by Jer Noble.
2462 Some websites call getUserMedia with a deviceId constraint and theey pick the
2463 first deviceId in the list provided by enumerateDevices.
2464 On iOS, this is the back camera which is often not what is expected by WebRTC applications.
2465 Instead, make sure the first camera device is the front camera.
2469 * platform/mediastream/mac/AVCaptureDeviceManager.mm:
2471 2019-07-16 Chris Dumez <cdumez@apple.com>
2473 Speed up StorageManager::getValues()
2474 https://bugs.webkit.org/show_bug.cgi?id=199812
2476 Reviewed by Alex Christensen.
2478 * storage/StorageMap.cpp:
2479 (WebCore::StorageMap::importItems):
2480 * storage/StorageMap.h:
2482 2019-07-16 Simon Fraser <simon.fraser@apple.com>
2484 Remove "GraphicsContext 0x7ffd8765f6b0 stack changed by" logging
2485 https://bugs.webkit.org/show_bug.cgi?id=199834
2487 Reviewed by Dean Jackson.
2489 I inadvertently left in a couple of GraphicsContextStateStackChecker (used to detect mismatched
2490 GraphicsContext save/restore) in r246725, so remove them.
2492 * rendering/RenderLayer.cpp:
2493 (WebCore::RenderLayer::paintLayerContents):
2494 * rendering/RenderLayerBacking.cpp:
2495 (WebCore::RenderLayerBacking::paintIntoLayer):
2497 2019-07-16 Wenson Hsieh <wenson_hsieh@apple.com>
2499 [Text autosizing] [iPadOS] Paragraph text on the front page of LinkedIn.com is not boosted
2500 https://bugs.webkit.org/show_bug.cgi?id=199827
2501 <rdar://problem/53152660>
2503 Reviewed by Zalan Bujtas.
2505 After r247467, we try to avoid boosting some text that might cause vertical overflow due to line height
2506 increasing inside a container that expects a fixed integer number of lines. However, in the case of
2507 linkedin.com, the line height specified is a fixed value of 26px, which greatly exceeds the specified font size
2508 of 14px. In this case, it's safe to boost font size, since doing so would not affect the line height at all.
2510 To handle this case, don't avoid text autosizing due to the "fixed number of lines" heuristic if the line height
2511 is fixed, and is much larger than the font size.
2513 Test: fast/text-autosizing/ios/idempotentmode/idempotent-autosizing-candidates.html
2515 * rendering/style/TextSizeAdjustment.cpp:
2516 (WebCore::AutosizeStatus::updateStatus):
2518 2019-07-16 Zalan Bujtas <zalan@apple.com>
2520 [ContentChangeObserver] Cancel ongoing content observation when tap is failed/cancelled
2521 https://bugs.webkit.org/show_bug.cgi?id=199828
2522 <rdar://problem/53152696>
2524 Reviewed by Wenson Hsieh.
2526 This patch ensures that we stop the content observation (initiated by touch start) when the tap
2527 is cancelled/failed.
2531 * page/ios/ContentChangeObserver.cpp:
2532 (WebCore::ContentChangeObserver::didCancelPotentialTap):
2533 * page/ios/ContentChangeObserver.h:
2535 2019-07-16 Rob Buis <rbuis@igalia.com>
2537 Remove support for beforeload on link=prefetch
2538 https://bugs.webkit.org/show_bug.cgi?id=199632
2540 Reviewed by Youenn Fablet.
2542 Remove support for beforeload on link=prefetch
2543 by not calling shouldLoadLink for prefetch.
2544 Also make loadLink return void since it always
2547 Tests: http/wpt/prefetch/beforeload-dynamic.html
2548 http/wpt/prefetch/beforeload.html
2550 * html/HTMLLinkElement.cpp:
2551 (WebCore::HTMLLinkElement::process):
2552 * loader/LinkLoader.cpp:
2553 (WebCore::LinkLoader::prefetchIfNeeded):
2554 (WebCore::LinkLoader::loadLink):
2555 * loader/LinkLoader.h:
2557 2019-07-16 Youenn Fablet <youenn@apple.com>
2559 Hop to main thread for release logging in RealtimeIncomingAudioSource
2560 https://bugs.webkit.org/show_bug.cgi?id=199813
2562 Reviewed by Eric Carlson.
2564 LoggerHelper routines allow logging messages in system console and inspector console.
2565 These routines iterate through a Vector of log observers which is not thread safe.
2566 Document, the main log observer, also expects to be called on the main thread.
2567 Manually tested (writing a layout test for this would require more than 2 seconds).
2569 * platform/mediastream/mac/RealtimeIncomingAudioSourceCocoa.cpp:
2570 (WebCore::RealtimeIncomingAudioSourceCocoa::OnData):
2572 2019-07-15 Ryan Haddad <ryanhaddad@apple.com>
2574 Unreviewed, rolling out r247444.
2576 Caused two scrolling tests to fail on iOS Simulator
2580 "Typing into a cell in a Google Sheet lags behind by one
2582 https://bugs.webkit.org/show_bug.cgi?id=199587
2583 https://trac.webkit.org/changeset/247444
2585 2019-07-15 Zalan Bujtas <zalan@apple.com>
2587 outlook.live.com has odd viewport with edge gap
2588 https://bugs.webkit.org/show_bug.cgi?id=199822
2589 <rdar://problem/53029072>
2591 Reviewed by Wenson Hsieh.
2593 r244944 introduced a viewport width quirk because at certain screen size outlook.live.com's flex setup produced a slighly broken layout.
2594 We don't seem to need this quirk anymore (composed column's min-width is removed.)
2597 (WebCore::Quirks::shouldIgnoreShrinkToFitContent const): Deleted.
2598 (WebCore::Quirks::overriddenViewLayoutWidth const): Deleted.
2601 2019-07-15 Saam Barati <sbarati@apple.com>
2603 [WHLSL] Matrix memory layout should match HLSL by laying out columns linearly
2604 https://bugs.webkit.org/show_bug.cgi?id=199215
2606 Reviewed by Myles C. Maxfield.
2608 This patch makes it so that we lay out matrices in memory in the same
2609 way HLSL does. This is by laying out columns linearly in memory. So a float4x4
2610 composed by this series of floats in memory:
2611 [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]
2613 composes this logical matrix:
2619 To implement this, we switch to using an array to represent the memory
2620 contents linear memory layout of a matrix. So the matrix float4x3 will now
2621 be an array<float, 12> in metal. Then, we change the indexed getter and
2622 setter methods for matrices to load and store from and to the correct
2623 memory locations. The memory layout of matrices is observable to WHLSL
2624 when using a matrix as an input/output to a shader.
2626 Test: webgpu/whlsl-matrix-memory-layout.html
2628 * Modules/webgpu/WHLSL/Metal/WHLSLNativeFunctionWriter.cpp:
2629 (WebCore::WHLSL::Metal::writeNativeFunction):
2630 * Modules/webgpu/WHLSL/Metal/WHLSLNativeTypeWriter.cpp:
2631 (WebCore::WHLSL::Metal::writeNativeType):
2633 2019-07-15 Wenson Hsieh <wenson_hsieh@apple.com>
2635 [Text autosizing] [iPadOS] Product label text is clipped in portrait mode on the front page of sephora.com
2636 https://bugs.webkit.org/show_bug.cgi?id=199806
2637 <rdar://problem/52902482>
2639 Reviewed by Zalan Bujtas.
2641 On sephora.com, some product label text is currently boosted by idempotent text autosizing, which causes the
2642 labels be vertically clipped. This patch augments the idempotent text autosizing heuristic to avoid this case by
2643 checking if the element to be boosted has a fixed height or max height, whose value is very close to a small
2644 integer multiple of the line height. In this case, it's likely that the website expects the text to be no more
2645 than a few lines' worth of height, so boosting the text is likely to break the page.
2647 Test: fast/text-autosizing/ios/idempotentmode/idempotent-autosizing-candidates.html
2649 * rendering/style/RenderStyle.cpp:
2650 (WebCore::RenderStyle::isIdempotentTextAutosizingCandidate const):
2651 * rendering/style/TextSizeAdjustment.cpp:
2652 (WebCore::AutosizeStatus::updateStatus):
2653 * rendering/style/TextSizeAdjustment.h:
2655 Rename Fields::DisplayNone to Fields::AvoidSubtree to avoid introducing another bit in RenderStyle's inherited
2658 2019-07-15 Myles C. Maxfield <mmaxfield@apple.com>
2660 New York font erroneously gets synthetic bold
2661 https://bugs.webkit.org/show_bug.cgi?id=199653
2663 Unreviewed watchOS build fix.
2665 * page/ProcessWarming.cpp:
2666 (WebCore::ProcessWarming::prewarmGlobally):
2667 * page/cocoa/MemoryReleaseCocoa.mm:
2668 (WebCore::platformReleaseMemory):
2669 * platform/graphics/FontCascadeDescription.cpp:
2670 * platform/graphics/FontDescription.h:
2671 * platform/graphics/cocoa/FontCacheCoreText.cpp:
2672 (WebCore::invalidateFontCache):
2673 (WebCore::lookupFallbackFont):
2674 * platform/graphics/cocoa/FontDescriptionCocoa.cpp:
2675 (WebCore::FontCascadeDescription::effectiveFamilyAt const):
2676 * platform/graphics/cocoa/FontFamilySpecificationCoreText.cpp:
2677 (WebCore::FontFamilySpecificationCoreText::fontRanges const):
2678 * platform/graphics/cocoa/SystemFontDatabaseCoreText.cpp:
2680 2019-07-15 Commit Queue <commit-queue@webkit.org>
2682 Unreviewed, rolling out r247462.
2683 https://bugs.webkit.org/show_bug.cgi?id=199816
2685 completely messed up the patch (Requested by litherum on
2690 "New York font erroneously gets synthetic bold"
2691 https://bugs.webkit.org/show_bug.cgi?id=199653
2692 https://trac.webkit.org/changeset/247462
2694 2019-07-11 Simon Fraser <simon.fraser@apple.com>
2696 Fix builds where HAVE_DESIGN_SYSTEM_UI_FONTS is not defined.
2698 * platform/graphics/cocoa/FontCacheCoreText.cpp:
2699 (WebCore::fontWithFamilySpecialCase):
2701 2019-07-11 Justin Fan <justin_fan@apple.com>
2703 Unreviewed build fix.
2705 * Modules/webgpu/WHLSL/WHLSLRecursionChecker.cpp: Missing include.
2707 2019-07-11 Myles C. Maxfield <mmaxfield@apple.com>
2709 New York font erroneously gets synthetic bold
2710 https://bugs.webkit.org/show_bug.cgi?id=199653
2711 <rdar://problem/51692592>
2713 Reviewed by Simon Fraser.
2715 This patch adds support for -apple-system-ui-serif, -apple-system-ui-monospaced,
2716 and -apple-system-ui-rounded, behind an SPI that is off-by-default. We don't want
2717 to expose these fonts to the web because we don't a standardization story for them
2718 yet, but we do want some apps to be able to use them.
2720 WebKit clients who want to use these fonts can set
2721 -[WKPreferences _shouldAllowDesignSystemUIFonts] = YES.
2723 The patch generalizes our existing system-ui infrastructure to handle these three
2724 additional fonts. It also explicitly disables the unsupported dot-prefixed names
2725 so they don't leak out into Web content.
2727 Tests: fast/text/design-system-ui-10.html
2728 fast/text/design-system-ui-11.html
2729 fast/text/design-system-ui-12.html
2730 fast/text/design-system-ui-13.html
2731 fast/text/design-system-ui-14.html
2732 fast/text/design-system-ui-15.html
2733 fast/text/design-system-ui-16.html
2734 fast/text/design-system-ui-2.html
2735 fast/text/design-system-ui-3.html
2736 fast/text/design-system-ui-4.html
2737 fast/text/design-system-ui-5.html
2738 fast/text/design-system-ui-6.html
2739 fast/text/design-system-ui-7.html
2740 fast/text/design-system-ui-8.html
2741 fast/text/design-system-ui-9.html
2742 fast/text/design-system-ui.html
2744 * css/CSSFontFace.cpp:
2745 (WebCore::CSSFontFace::shouldAllowDesignSystemUIFonts const):
2746 * css/CSSFontFace.h:
2747 * css/CSSFontFaceSource.cpp:
2748 (WebCore::CSSFontFaceSource::load):
2749 * css/StyleResolver.cpp:
2750 (WebCore::StyleResolver::initializeFontStyle):
2751 * page/Settings.yaml:
2752 * platform/graphics/FontDescription.cpp:
2753 (WebCore::m_shouldAllowDesignSystemUIFonts):
2754 (WebCore::m_shouldAllowUserInstalledFonts): Deleted.
2755 * platform/graphics/FontDescription.h:
2756 (WebCore::FontDescription::shouldAllowDesignSystemUIFonts const):
2757 (WebCore::FontDescription::setShouldAllowDesignSystemUIFonts):
2758 (WebCore::FontDescription::operator== const):
2759 * platform/graphics/cocoa/FontCacheCoreText.cpp:
2760 (WebCore::platformFontLookupWithFamily):
2761 (WebCore::fontWithFamily):
2762 * platform/graphics/cocoa/FontCacheCoreText.h:
2763 * platform/graphics/cocoa/FontDescriptionCocoa.cpp:
2764 (WebCore::systemFontUse):
2765 (WebCore::systemFontCascadeList):
2766 (WebCore::FontCascadeDescription::effectiveFamilyCount const):
2767 (WebCore::FontCascadeDescription::effectiveFamilyAt const):
2768 (WebCore::isSystemFontString): Deleted.
2769 (WebCore::isUIFontTextStyle): Deleted.
2770 * platform/graphics/cocoa/FontPlatformDataCocoa.mm:
2771 (WebCore::FontPlatformData::ctFont const):
2772 * platform/graphics/cocoa/SystemFontDatabaseCoreText.cpp:
2773 (WebCore::SystemFontDatabaseCoreText::createSystemUI):
2774 (WebCore::SystemFontDatabaseCoreText::createDesignSystemUI):
2775 (WebCore::SystemFontDatabaseCoreText::createTextStyle):
2776 (WebCore::SystemFontDatabaseCoreText::cascadeList):
2777 (WebCore::SystemFontDatabaseCoreText::applyWeightItalicsAndFallbackBehavior):
2778 (WebCore::SystemFontDatabaseCoreText::systemFontParameters):
2779 * platform/graphics/cocoa/SystemFontDatabaseCoreText.h:
2780 * platform/graphics/ios/FontCacheIOS.mm:
2781 (WebCore::platformFontWithFamilySpecialCase):
2782 * platform/graphics/mac/FontCacheMac.mm:
2783 (WebCore::platformFontWithFamilySpecialCase):
2784 * style/StyleResolveForDocument.cpp:
2785 (WebCore::Style::resolveForDocument):
2786 * svg/graphics/SVGImage.cpp:
2787 (WebCore::SVGImage::dataChanged):
2788 * testing/InternalSettings.cpp:
2789 (WebCore::InternalSettings::Backup::Backup):
2790 (WebCore::InternalSettings::Backup::restoreTo):
2791 (WebCore::InternalSettings::setShouldAllowDesignSystemUIFonts):
2792 * testing/InternalSettings.h:
2793 * testing/InternalSettings.idl:
2795 2019-07-15 Chris Dumez <cdumez@apple.com>
2797 Crash under DisplayRefreshMonitorManager::displayWasUpdated()
2798 https://bugs.webkit.org/show_bug.cgi?id=199808
2799 <rdar://problem/53070144>
2801 Reviewed by Geoffrey Garen.
2803 Copy m_monitors before iterating over it because the calling displayLinkFired() on the
2804 monitor may end up calling DisplayRefreshMonitorManager::displayDidRefresh() synchronously,
2805 which removes the monitor from m_monitors.
2807 * platform/graphics/DisplayRefreshMonitorManager.cpp:
2808 (WebCore::DisplayRefreshMonitorManager::displayWasUpdated):
2810 2019-07-15 Yusuke Suzuki <ysuzuki@apple.com>
2812 [JSC] Improve wasm wpt test results by fixing miscellaneous issues
2813 https://bugs.webkit.org/show_bug.cgi?id=199783
2815 Reviewed by Mark Lam.
2817 Remove WebAssemblyPrototype.
2819 Tests: imported/w3c/web-platform-tests/wasm/jsapi/constructor/instantiate-bad-imports.any.html
2820 imported/w3c/web-platform-tests/wasm/jsapi/constructor/instantiate-bad-imports.any.worker.html
2821 imported/w3c/web-platform-tests/wasm/jsapi/global/constructor.any.html
2822 imported/w3c/web-platform-tests/wasm/jsapi/global/constructor.any.worker.html
2823 imported/w3c/web-platform-tests/wasm/jsapi/global/toString.any.html
2824 imported/w3c/web-platform-tests/wasm/jsapi/global/toString.any.worker.html
2825 imported/w3c/web-platform-tests/wasm/jsapi/global/value-set.any.html
2826 imported/w3c/web-platform-tests/wasm/jsapi/global/value-set.any.worker.html
2827 imported/w3c/web-platform-tests/wasm/jsapi/global/valueOf.any.html
2828 imported/w3c/web-platform-tests/wasm/jsapi/global/valueOf.any.worker.html
2829 imported/w3c/web-platform-tests/wasm/jsapi/instance/constructor-bad-imports.any.html
2830 imported/w3c/web-platform-tests/wasm/jsapi/instance/constructor-bad-imports.any.worker.html
2831 imported/w3c/web-platform-tests/wasm/jsapi/instance/constructor.any.html
2832 imported/w3c/web-platform-tests/wasm/jsapi/instance/constructor.any.worker.html
2833 imported/w3c/web-platform-tests/wasm/jsapi/instance/exports.any.html
2834 imported/w3c/web-platform-tests/wasm/jsapi/instance/exports.any.worker.html
2835 imported/w3c/web-platform-tests/wasm/jsapi/instance/toString.any.html
2836 imported/w3c/web-platform-tests/wasm/jsapi/instance/toString.any.worker.html
2837 imported/w3c/web-platform-tests/wasm/jsapi/interface.any.html
2838 imported/w3c/web-platform-tests/wasm/jsapi/interface.any.worker.html
2839 imported/w3c/web-platform-tests/wasm/jsapi/memory/buffer.any.html
2840 imported/w3c/web-platform-tests/wasm/jsapi/memory/buffer.any.worker.html
2841 imported/w3c/web-platform-tests/wasm/jsapi/memory/constructor.any.html
2842 imported/w3c/web-platform-tests/wasm/jsapi/memory/constructor.any.worker.html
2843 imported/w3c/web-platform-tests/wasm/jsapi/memory/grow.any.html
2844 imported/w3c/web-platform-tests/wasm/jsapi/memory/grow.any.worker.html
2845 imported/w3c/web-platform-tests/wasm/jsapi/memory/toString.any.html
2846 imported/w3c/web-platform-tests/wasm/jsapi/memory/toString.any.worker.html
2847 imported/w3c/web-platform-tests/wasm/jsapi/module/constructor.any.html
2848 imported/w3c/web-platform-tests/wasm/jsapi/module/constructor.any.worker.html
2849 imported/w3c/web-platform-tests/wasm/jsapi/module/customSections.any.html
2850 imported/w3c/web-platform-tests/wasm/jsapi/module/customSections.any.worker.html
2851 imported/w3c/web-platform-tests/wasm/jsapi/module/exports.any.html
2852 imported/w3c/web-platform-tests/wasm/jsapi/module/exports.any.worker.html
2853 imported/w3c/web-platform-tests/wasm/jsapi/module/imports.any.html
2854 imported/w3c/web-platform-tests/wasm/jsapi/module/imports.any.worker.html
2855 imported/w3c/web-platform-tests/wasm/jsapi/module/toString.any.html
2856 imported/w3c/web-platform-tests/wasm/jsapi/module/toString.any.worker.html
2857 imported/w3c/web-platform-tests/wasm/jsapi/table/constructor.any.html
2858 imported/w3c/web-platform-tests/wasm/jsapi/table/constructor.any.worker.html
2859 imported/w3c/web-platform-tests/wasm/jsapi/table/get-set.any.html
2860 imported/w3c/web-platform-tests/wasm/jsapi/table/get-set.any.worker.html
2861 imported/w3c/web-platform-tests/wasm/jsapi/table/grow.any.html
2862 imported/w3c/web-platform-tests/wasm/jsapi/table/grow.any.worker.html
2863 imported/w3c/web-platform-tests/wasm/jsapi/table/length.any.html
2864 imported/w3c/web-platform-tests/wasm/jsapi/table/length.any.worker.html
2865 imported/w3c/web-platform-tests/wasm/jsapi/table/toString.any.html
2866 imported/w3c/web-platform-tests/wasm/jsapi/table/toString.any.worker.html
2868 * bindings/js/JSDOMWindowBase.cpp:
2869 (WebCore::JSDOMWindowBase::compileStreaming):
2870 (WebCore::JSDOMWindowBase::instantiateStreaming):
2872 2019-07-15 Dean Jackson <dino@apple.com>
2874 [WebGL] Remove software rendering and simplify context creation on macOS
2875 https://bugs.webkit.org/show_bug.cgi?id=199789
2877 Reviewed by Sam Weinig.
2879 We don't ever want to fall-back to the software renderer. We'd be better
2880 off failing to create the context completely.
2882 Also, the number of fall-back attempts we were making before hitting
2883 the software renderer was overkill. All hardware we support should
2884 handle a 32bpp buffer.
2886 Lastly, we don't want to support supersampling - multisampling only.
2888 I lied… there is one more thing - failing to create the context
2889 was causing an ASSERT trying to remove the GC3D from the global list.
2891 * html/canvas/WebGLRenderingContextBase.cpp:
2892 (WebCore::WebGLRenderingContextBase::create):
2893 * page/Settings.yaml: Remove forceSoftwareWebGL setting.
2894 * platform/graphics/GraphicsContext3DAttributes.h:
2895 * platform/graphics/GraphicsContext3DManager.cpp:
2896 (WebCore::GraphicsContext3DManager::addContext):
2897 (WebCore::GraphicsContext3DManager::removeContext):
2898 (WebCore::GraphicsContext3DManager::removeContextRequiringHighPerformance):
2899 * platform/graphics/cocoa/GraphicsContext3DCocoa.mm:
2900 (WebCore::GraphicsContext3D::GraphicsContext3D):
2901 (WebCore::setPixelFormat): Deleted.
2903 2019-07-14 Dean Jackson <dino@apple.com>
2905 Move more WebGL things into unified builds
2906 https://bugs.webkit.org/show_bug.cgi?id=199787
2908 Reviewed by Wenson Hsieh.
2910 Unify as many WebGL files as possible.
2913 * WebCore.xcodeproj/project.pbxproj:
2914 * html/canvas/WebGLFramebuffer.cpp:
2915 * html/canvas/WebGLObject.h: Move a helper function into an inline.
2916 (WebCore::objectOrZero):
2917 * html/canvas/WebGLRenderingContextBase.cpp:
2918 * platform/graphics/cocoa/FontCacheCoreText.cpp:
2919 * platform/mediastream/MediaStreamTrackPrivate.cpp:
2921 2019-07-15 Keith Rollin <krollin@apple.com>
2923 Add missing webgpu includes
2924 https://bugs.webkit.org/show_bug.cgi?id=199802
2925 <rdar://problem/53119120>
2927 Reviewed by Myles C. Maxfield.
2929 WHLSLFunctionWriter.cpp fails to compile on tvOS and watchOS due to
2930 missing includes. This file likely compiles on other platforms due to
2931 the Unified Sources facility that clumps several implementation files
2932 together, exposing those implementation files to the others' includes.
2933 It fails on tvOS and watchOS due to the different ways these
2934 implementation files can get clumped together on different platforms.
2936 No new tests as there is no functional change.
2938 * Modules/webgpu/WHLSL/Metal/WHLSLFunctionWriter.cpp:
2940 2019-07-15 Michael Catanzaro <mcatanzaro@igalia.com>
2942 Unreviewed, rolling out r247440.
2948 "[JSC] Improve wasm wpt test results by fixing miscellaneous
2950 https://bugs.webkit.org/show_bug.cgi?id=199783
2951 https://trac.webkit.org/changeset/247440
2953 2019-07-15 Daniel Bates <dabates@apple.com>
2955 Typing into a cell in a Google Sheet lags behind by one character
2956 https://bugs.webkit.org/show_bug.cgi?id=199587
2957 <rdar://problem/51616845>
2959 Reviewed by Brent Fulgham.
2961 Add a Google Sheets quirk. Put all DOM timers scheduled from keydown and keypress event listeners
2962 into a holding tank. The timers continue to tick, but are barred from executing their action until
2963 the next text insertion or deletion or 32 ms (on device) have elapsed, whichever is sooner. We only
2964 allocate a holding tank once per document, only if the quirk is active, and this allocation is done
2965 when the document schedules a timer on keydown or keypress. The holding tank lives for the lifetime
2968 The story behind the quirk:
2970 On keypress Google Sheets schedules timers and expects that a DOM update will occur (i.e. text
2971 will be inserted or deleted) within the same event loop iteration as the dispatched keypress. The
2972 UI Events spec. [1] makes no such guarantee of when a DOM update must occur in relation to the keypress
2973 event. It could happen in the same event loop iteration as the key press (as Google expects), the
2974 next iteration, 500ms later, 2 minutes later, etc. What the spec does guarantee is that by the time
2975 a DOM input event is dispatched that the DOM will be updated. And this is the solution to the problem
2976 Google Sheets is trying to solve, but is doing so using pre-IE 9 technology (though similar
2977 functionality was available via onpropertychange in IE < 9).
2979 See also <https://github.com/w3c/uievents/issues/238>, which is tracking a spec. text update for
2982 Test: fast/events/ios/dom-update-on-keydown-quirk.html
2984 [1] <https://w3c.github.io/uievents/> (Editor's Draft, 14 October 2018)
2987 * WebCore.xcodeproj/project.pbxproj:
2988 Add some files to the project.
2991 (WebCore::Document::domTimerHoldingTank): Added.
2993 (WebCore::Document::domTimerHoldingTankIfExists): Added.
2995 * page/DOMTimer.cpp:
2996 (WebCore::DOMTimer::install): Put the newly instantiated timer into the holding tank.
2997 (WebCore::DOMTimer::removeById): Remove the timer from the holding tank.
2998 (WebCore::DOMTimer::fired): Check if the timer is in the holding tank. If it is and it is a one-
2999 shot timer then schedule it for the next event loop iteration. If it's a repeating timer just
3000 let it continue ticking. Otherwise, do what we no now and execute the timer's action. The reason
3001 we do not suspend timers in the holding tank is because:
3002 1. Far out timers (Google Sheets registers timers as far out as 5 minutes!) are not penalized.
3003 Though smart supension logic could avoid this. See (3).
3005 2. Empirical observations indicate that the keyboard will perform the insertion or deletion
3006 reasonably quickly (not the same event loop iteration as the keydown, but within two iterations out).
3007 So, the timers in the holding tank are short-lived.
3009 3. Simplifies the code. There is no need to keep additional bookkeeping to track multiple timer
3010 suspension reasons (timers currently can only have one suspension reason) or alternatively defer
3011 scheduling a timer until a later time and computing a new "fair" firing time when scheduled.
3012 * page/EventHandler.cpp:
3013 (WebCore::EventHandler::internalKeyEvent): Place a token on the stack to put all DOM timers
3014 scheduled on keydown and keypress into the holding tank if the quirk is enabled.
3016 (WebCore::Quirks::needsDeferKeyDownAndKeyPressTimersUntilNextEditingCommand const): Added.
3018 * page/Settings.yaml: Added setting so that this quirk can be enabled from a layout test. This setting
3019 also lets us enable the quirk for all sites or for certain third-party apps if desired.
3020 * page/ios/DOMTimerHoldingTank.cpp: Added.
3021 (WebCore::DOMTimerHoldingTank::DOMTimerHoldingTank):
3022 (WebCore::DOMTimerHoldingTank::add):
3023 (WebCore::DOMTimerHoldingTank::remove):
3024 (WebCore::DOMTimerHoldingTank::contains):
3025 (WebCore::DOMTimerHoldingTank::removeAll):
3026 (WebCore::DOMTimerHoldingTank::stopExceededMaximumHoldTimer):
3027 * page/ios/DOMTimerHoldingTank.h: Added.
3028 (WebCore::DeferDOMTimersForScope::DeferDOMTimersForScope):
3029 (WebCore::DeferDOMTimersForScope::~DeferDOMTimersForScope):
3030 (WebCore::DeferDOMTimersForScope::isDeferring):
3032 2019-07-15 Ryan Haddad <ryanhaddad@apple.com>
3034 Unreviewed, rolling out r247397, r247378, r247366.
3036 Broke watchOS builds.
3038 Reverted changesets:
3040 "[WebGPU] Implement GPUError and error scopes"
3041 https://bugs.webkit.org/show_bug.cgi?id=199655
3042 https://trac.webkit.org/changeset/247366/webkit
3044 "[WebGPU] Move error scopes out of GPUDevice for more portable error generation"
3045 https://bugs.webkit.org/show_bug.cgi?id=199740
3046 https://trac.webkit.org/changeset/247397/webkit
3048 2019-07-15 Rob Buis <rbuis@igalia.com>
3050 MathML WPT test for RadicalDegreeBottomRaisePercent fails
3051 https://bugs.webkit.org/show_bug.cgi?id=183631
3053 Reviewed by Frédéric Wang.
3055 When stretching the radical operator also take thickness
3056 and radical gap into account:
3057 https://mathml-refresh.github.io/mathml-core/#radicals-msqrt-mroot
3059 Test: imported/w3c/web-platform-tests/mathml/presentation-markup/radicals/root-parameters-1.html
3061 * rendering/mathml/RenderMathMLRoot.cpp:
3062 (WebCore::RenderMathMLRoot::layoutBlock):
3064 2019-07-15 Yusuke Suzuki <ysuzuki@apple.com>
3066 [JSC] Improve wasm wpt test results by fixing miscellaneous issues
3067 https://bugs.webkit.org/show_bug.cgi?id=199783
3069 Reviewed by Mark Lam.
3071 Remove WebAssemblyPrototype.
3073 Tests: imported/w3c/web-platform-tests/wasm/jsapi/constructor/instantiate-bad-imports.any.html
3074 imported/w3c/web-platform-tests/wasm/jsapi/constructor/instantiate-bad-imports.any.worker.html
3075 imported/w3c/web-platform-tests/wasm/jsapi/global/constructor.any.html
3076 imported/w3c/web-platform-tests/wasm/jsapi/global/constructor.any.worker.html
3077 imported/w3c/web-platform-tests/wasm/jsapi/global/toString.any.html
3078 imported/w3c/web-platform-tests/wasm/jsapi/global/toString.any.worker.html
3079 imported/w3c/web-platform-tests/wasm/jsapi/global/value-set.any.html
3080 imported/w3c/web-platform-tests/wasm/jsapi/global/value-set.any.worker.html
3081 imported/w3c/web-platform-tests/wasm/jsapi/global/valueOf.any.html
3082 imported/w3c/web-platform-tests/wasm/jsapi/global/valueOf.any.worker.html
3083 imported/w3c/web-platform-tests/wasm/jsapi/instance/constructor-bad-imports.any.html
3084 imported/w3c/web-platform-tests/wasm/jsapi/instance/constructor-bad-imports.any.worker.html
3085 imported/w3c/web-platform-tests/wasm/jsapi/instance/constructor.any.html
3086 imported/w3c/web-platform-tests/wasm/jsapi/instance/constructor.any.worker.html
3087 imported/w3c/web-platform-tests/wasm/jsapi/instance/exports.any.html
3088 imported/w3c/web-platform-tests/wasm/jsapi/instance/exports.any.worker.html
3089 imported/w3c/web-platform-tests/wasm/jsapi/instance/toString.any.html
3090 imported/w3c/web-platform-tests/wasm/jsapi/instance/toString.any.worker.html
3091 imported/w3c/web-platform-tests/wasm/jsapi/interface.any.html
3092 imported/w3c/web-platform-tests/wasm/jsapi/interface.any.worker.html
3093 imported/w3c/web-platform-tests/wasm/jsapi/memory/buffer.any.html
3094 imported/w3c/web-platform-tests/wasm/jsapi/memory/buffer.any.worker.html
3095 imported/w3c/web-platform-tests/wasm/jsapi/memory/constructor.any.html
3096 imported/w3c/web-platform-tests/wasm/jsapi/memory/constructor.any.worker.html
3097 imported/w3c/web-platform-tests/wasm/jsapi/memory/grow.any.html
3098 imported/w3c/web-platform-tests/wasm/jsapi/memory/grow.any.worker.html
3099 imported/w3c/web-platform-tests/wasm/jsapi/memory/toString.any.html
3100 imported/w3c/web-platform-tests/wasm/jsapi/memory/toString.any.worker.html
3101 imported/w3c/web-platform-tests/wasm/jsapi/module/constructor.any.html
3102 imported/w3c/web-platform-tests/wasm/jsapi/module/constructor.any.worker.html
3103 imported/w3c/web-platform-tests/wasm/jsapi/module/customSections.any.html
3104 imported/w3c/web-platform-tests/wasm/jsapi/module/customSections.any.worker.html
3105 imported/w3c/web-platform-tests/wasm/jsapi/module/exports.any.html
3106 imported/w3c/web-platform-tests/wasm/jsapi/module/exports.any.worker.html
3107 imported/w3c/web-platform-tests/wasm/jsapi/module/imports.any.html
3108 imported/w3c/web-platform-tests/wasm/jsapi/module/imports.any.worker.html
3109 imported/w3c/web-platform-tests/wasm/jsapi/module/toString.any.html
3110 imported/w3c/web-platform-tests/wasm/jsapi/module/toString.any.worker.html
3111 imported/w3c/web-platform-tests/wasm/jsapi/table/constructor.any.html
3112 imported/w3c/web-platform-tests/wasm/jsapi/table/constructor.any.worker.html
3113 imported/w3c/web-platform-tests/wasm/jsapi/table/get-set.any.html
3114 imported/w3c/web-platform-tests/wasm/jsapi/table/get-set.any.worker.html
3115 imported/w3c/web-platform-tests/wasm/jsapi/table/grow.any.html
3116 imported/w3c/web-platform-tests/wasm/jsapi/table/grow.any.worker.html
3117 imported/w3c/web-platform-tests/wasm/jsapi/table/length.any.html
3118 imported/w3c/web-platform-tests/wasm/jsapi/table/length.any.worker.html
3119 imported/w3c/web-platform-tests/wasm/jsapi/table/toString.any.html
3120 imported/w3c/web-platform-tests/wasm/jsapi/table/toString.any.worker.html
3122 * bindings/js/JSDOMWindowBase.cpp:
3123 (WebCore::JSDOMWindowBase::compileStreaming):
3124 (WebCore::JSDOMWindowBase::instantiateStreaming):
3126 2019-07-15 Youenn Fablet <youenn@apple.com>
3128 Enable a debug WebRTC mode without any encryption
3129 https://bugs.webkit.org/show_bug.cgi?id=199177
3130 <rdar://problem/52074986>
3132 Reviewed by Eric Carlson.
3134 For every RTCPeerConnection, first set whether to use encryption or not
3135 based on page settings.
3136 If encryption is disabled, log it.
3137 Add internals API to toggle the switch from tests.
3138 Test: webrtc/disable-encryption.html
3140 * Modules/mediastream/RTCPeerConnection.cpp:
3141 (WebCore::RTCPeerConnection::RTCPeerConnection):
3142 * Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.cpp:
3143 (WebCore::createLibWebRTCPeerConnectionBackend):
3144 * inspector/agents/InspectorPageAgent.cpp:
3145 * page/Settings.yaml:
3146 * platform/mediastream/libwebrtc/LibWebRTCProvider.cpp:
3147 (WebCore::LibWebRTCProvider::setEnableWebRTCEncryption):
3148 * platform/mediastream/libwebrtc/LibWebRTCProvider.h:
3149 * testing/Internals.cpp:
3150 (WebCore::Internals::resetToConsistentState):
3151 (WebCore::Internals::setEnableWebRTCEncryption):
3152 * testing/Internals.h:
3153 * testing/Internals.idl:
3155 2019-07-15 Sihui Liu <sihui_liu@apple.com>
3157 window.openDatabase is not writable
3158 https://bugs.webkit.org/show_bug.cgi?id=199737
3159 <rdar://problem/52551332>
3161 Reviewed by Chris Dumez.
3163 In r246707 we made openDatabase an undetectable attribute of window, and it was set to be read-only. This broke
3164 some sites which replace window.openDatabase with their own implementation when window.openDatabase does not
3165 exists or WebSQL is not implemented.
3167 This patch removes the readonly property and adds a setter for openDatabase.
3169 * Modules/webdatabase/DOMWindowWebDatabase.idl:
3170 * bindings/js/JSDOMWindowCustom.cpp:
3171 (WebCore::JSDOMWindow::setOpenDatabase):
3173 2019-07-15 Youenn Fablet <youenn@apple.com>
3176 https://bugs.webkit.org/show_bug.cgi?id=199791
3178 Reviewed by Eric Carlson.
3180 As discussed in https://github.com/rtcweb-wg/mdns-ice-candidates/issues/91,
3181 use 0.0.0.0 for c lines when filtering the SDP.
3182 Covered by updated test.
3184 * Modules/mediastream/PeerConnectionBackend.cpp:
3185 (WebCore::PeerConnectionBackend::filterSDP const):
3187 2019-07-15 Zalan Bujtas <zalan@apple.com>
3189 naver.com: Video stops when tapping on the video to use buttons
3190 https://bugs.webkit.org/show_bug.cgi?id=199784
3191 <rdar://problem/48068610>
3193 Reviewed by Maciej Stachowiak.
3195 Opt out of simulated mouse event sending for tv.naver.com.
3196 When tv.naver.com receives the simulated mouseUp at touchEnd, it pauses the video right away and calls preventDefault()
3197 which prevents us from running the hover heuristics.
3200 (WebCore::Quirks::shouldDispatchSimulatedMouseEvents const):
3202 2019-07-15 Michael Catanzaro <mcatanzaro@igalia.com>
3204 [GTK] GitHub breaks on FreeBSD because of "unsupported browser"
3205 https://bugs.webkit.org/show_bug.cgi?id=199745
3207 Reviewed by Carlos Garcia Campos.
3209 It's been a while since I last updated the fake version numbers in our user agent, both for
3210 the user agent quirks for naughty websites and also the Safari version in our standard user
3211 agent. Update them. This should fix github.com on FreeBSD at least. I also noticed some
3212 wonkiness on Google Docs recently that I thought required this update, but I didn't do
3213 anything about it at the time because I wasn't able to reproduce the issue when I tried
3216 This could absolutely break websites, because the web is awful, but that's calculated risk.
3218 * platform/UserAgentQuirks.cpp:
3219 (WebCore::UserAgentQuirks::stringForQuirk):
3220 * platform/glib/UserAgentGLib.cpp:
3221 (WebCore::buildUserAgentString):
3223 2019-07-14 Chris Dumez <cdumez@apple.com>
3225 Add threading assertion to WeakPtrFactory::createWeakPtr()
3226 https://bugs.webkit.org/show_bug.cgi?id=199639
3228 Reviewed by Ryosuke Niwa.
3230 * platform/ScrollableArea.cpp:
3231 * rendering/RenderObject.cpp:
3233 2019-07-14 Dean Jackson <dino@apple.com>
3235 WebGL not supported on WKWebView on UIKit for Mac
3236 https://bugs.webkit.org/show_bug.cgi?id=199785
3237 <rdar://problem/52911449>
3239 Reviewed by Antoine Quint.
3241 UIKit for Mac was not creating a CGLPixelFormatObj because
3242 it wasn't using the code hidden in PLATFORM(MAC). Instead
3243 we should be guarding for USE(OPENGL).
3245 There are still some inconsistencies: <rdar://53062794>
3247 Test: webgl/smell-test.html
3249 * platform/graphics/cocoa/GraphicsContext3DCocoa.mm:
3250 (WebCore::GraphicsContext3D::GraphicsContext3D):
3251 (WebCore::GraphicsContext3D::allowOfflineRenderers const): We have to return
3252 true for this, since we don't have access to the Window Server.
3254 2019-07-13 Wenson Hsieh <wenson_hsieh@apple.com>
3256 [Text autosizing] [iPadOS] Further adjust our heuristics to determine text autosizing candidates
3257 https://bugs.webkit.org/show_bug.cgi?id=199780
3258 <rdar://problem/52289088>
3260 Reviewed by Simon Fraser.
3262 Our current idempotent text autosizing candidate heuristic makes the right judgment call most of the time, but
3263 there is still a large batch of text autosizing bugs left unfixed by the first iteration of the heuristic added
3264 in r246781. This patch attempts to address most of these bugs by adjusting the decision-tree-based heuristic
3265 once again, mostly with improvements to the model generation pipeline.
3267 During the first iteration, I placed emphasis on tuning the max tree depth and min leaf size hyperparameters
3268 when coming up with my decision tree, and didn't consider the inclusion or exclusion of each feature as a
3269 hyperparameters. As such, the trees generated using the pipeline tended to use too many features, and as a
3270 result, tended to have cross-validation overall accuracy scores hovering around 73%.
3272 In this revised model generation pipeline, I now consider the inclusion of each feature (along with max depth
3273 and min leaf size, as before) as a hyperparameter. Since this increases the number of hyperparameters by many
3274 orders of magnitude, a naive grid search (as described in the prior ChangeLog entry) is no longer a tractible
3275 procedure for tuning hyperparameters to the training algorithm.
3277 Instead, I now use a stochastic greedy algorithm to search for good sets of hyperparameters; this process begins
3278 with seeding some number (usually 20-24) of "searchers" with completely randomized sets of hyperparameters (i.e.
3279 random max depth, random leaf size, and random subsets of features). I then evaluate the average performance of
3280 each set of hyperparameters by using them to generate 2000 decision trees over 90% of the training data, and
3281 then cross-validating these trees against the remaining 10%. These cross-validation scores are aggregated into a
3282 single confusion matrix, which is then passed into a loss function that computes a single value indicating how
3283 well training with the set of hyperparameters generalized to cross-validation data. After experimenting with
3284 various loss functions, I settled on the following:
3286 `k(false positive rate)^2 + (false negative rate)^2`
3288 ...where a constant k is chosen to penalize false positives (i.e. broken layout) more harshly than false
3289 negatives (small text). Additionally, squaring the false negative and false positive rates seems to help avoid
3290 converging on solutions that heavily favor reducing only false positives or false negatives, or vice versa.
3292 The stochastic algorithm starts by computing a loss value for the randomly generated configuration. Then, for
3293 an indefinite number of iterations, it randomly mutates the configuration (e.g. by adding or removing features,
3294 or changing min leaf size or max tree depth) and computes a new loss value for the mutated configuration. If the
3295 mutated configuration performs better (i.e. achieves lower loss) than the current configuration, I set the
3296 current configuration to be the mutated configuration. Otherwise, I keep the current (non-mutated) configuration
3297 as-is. The stochastic algorithm then proceeds, ad-infinitum, with this current configuration.
3299 Of course, since each mutation is small, this strategy so far is prone to leaving each searcher stuck in local
3300 optima. To mitigate this, for each searcher, I keep track of a side-table of configurations that have already
3301 been tested; when random mutations would normally lead to testing a configuration that has already been tested,
3302 each searcher instead increases the chance of applying additional mutations. This has the effect of searchers
3303 initially exhausting similar configurations, and expanding to test more and more dissimilar configurations as
3304 the local alternatives all turn out to be worse. This allows searchers to effectively jump out of local optima
3305 after being stuck for a long time.
3307 So, using these strategies, I simultaneously ran a handful of searchers until they all appeared to converge
3308 (a process that takes 8-12 hours on my current dataset). Many of the searchers achieved configurations with
3309 cross-validation scores of 81% and above, up from the 73% of the previous attempt. These additionally have the
3310 added bonus of reducing the number of features, often making the final trees themselves shallower and simpler to
3311 understand than before.
3313 This patch introduces one such decision tree generated using a set of hyperparameters acquired via this
3314 stochasic search algorithm; it appears to simultaneously use fewer features, and achieve better cross-validation
3317 Test: fast/text-autosizing/ios/idempotentmode/idempotent-autosizing-candidates.html
3319 * css/StyleResolver.cpp:
3320 (WebCore::StyleResolver::adjustRenderStyleForTextAutosizing):
3322 Adjust the early return to bail if either (1) the element is a candidate and the computed size is already equal
3323 to the boosted size, or (2) the element is not a candidate and the computed size is already equal to the
3324 specified size. Since the autosizing candidate heuristic depends on styles specified on the element itself (as
3325 opposed to styles on any element in the ancestor chain), a parent may be an autosizing candidate, but a child of
3328 * rendering/style/RenderStyle.cpp:
3329 (WebCore::RenderStyle::isIdempotentTextAutosizingCandidate const):
3331 Revamp the idempotent text autosizing candidate heuristic. See the explanation above for more details.
3333 * rendering/style/RenderStyle.h:
3335 Remove some bits from RenderStyle's autosizeStatus, now that we care about fewer bits of information from the
3338 * rendering/style/TextSizeAdjustment.cpp:
3339 (WebCore::AutosizeStatus::updateStatus):
3340 * rendering/style/TextSizeAdjustment.h:
3342 2019-07-13 Simon Fraser <simon.fraser@apple.com>
3344 Don't do async overflow scrolling for visibility:hidden scrollers
3345 https://bugs.webkit.org/show_bug.cgi?id=199779
3347 Reviewed by Dean Jackson.
3349 An overflow:scroll with visibility:hidden is not scrollable on macOS, even if it has visible
3350 content. So disable async overflow:scroll when the scroller has non-visible visibility (this also
3351 takes visibility on ancestors into account).
3353 visibility:hidden overflow:scroll can be common because some JS libraries use it
3354 (https://github.com/wnr/element-resize-detector).
3356 Test: compositing/scrolling/async-overflow-scrolling/visibility-hidden-scrollers.html
3358 * rendering/RenderLayer.cpp:
3359 (WebCore::RenderLayer::canUseCompositedScrolling const): Don't use hasVisibleContent() because
3360 that's affected by visible children.
3362 2019-07-13 Robin Morisset <rmorisset@apple.com>
3364 [WHLSL] Return statements don't need to keep track of the function they're in
3365 https://bugs.webkit.org/show_bug.cgi?id=199763
3367 Reviewed by Myles C. Maxfield.
3369 Return::m_function is only used in the Checker, and it can easily enough keep track of the current function.
3370 This means we no longer need to keep track of the current function in the NameResolver, and we can save 8 bytes per Return
3372 Since I was touching the NameResolver I also removed a few pointless overrides of Visitor::visit().
3374 No new tests as there is no intended functional change.
3376 * Modules/webgpu/WHLSL/AST/WHLSLReturn.h:
3377 * Modules/webgpu/WHLSL/WHLSLChecker.cpp:
3378 (WebCore::WHLSL::Checker::visit):
3379 * Modules/webgpu/WHLSL/WHLSLNameResolver.cpp:
3380 (WebCore::WHLSL::NameResolver::NameResolver):
3381 (WebCore::WHLSL::resolveTypeNamesInFunctions):
3382 * Modules/webgpu/WHLSL/WHLSLNameResolver.h:
3384 2019-07-13 Andres Gonzalez <andresg_22@apple.com>
3386 Add accessibility support to WKDataListSuggestionsView.
3387 https://bugs.webkit.org/show_bug.cgi?id=199772
3388 <rdar://problem/47095851>
3390 Reviewed by Chris Fleizach.
3392 Localizable strings for accessibility announcements.
3394 * en.lproj/Localizable.strings:
3396 2019-07-13 Chris Dumez <cdumez@apple.com>
3398 Drop non thread-safe usage of WeakPtr in VideoFullscreenInterfaceAVKit
3399 https://bugs.webkit.org/show_bug.cgi?id=199775
3401 Reviewed by Eric Carlson.
3403 The VideoFullscreenInterfaceAVKit constructor was making a weakPtr on the UI Thread
3404 of an WebThread object. The WeakPtr would then be used as a data member throughout
3405 the class on the UIThread. This is not thread-safe.
3407 This patch switches to using a raw pointer instead of a WeakPtr. This is a partial
3408 rollout of r243298, which turned the raw pointer into a WeakPtr for hardening
3409 purposes. For extra safety, this patch updates the VideoFullscreenControllerContext
3410 so that it notifies its clients (i.e. PlaybackSessionInterfaceAVKit) that it is
3411 getting destroyed, so that they can null-out their m_videoFullscreenModel &
3412 m_fullscreenChangeObserver data members. This gives the sames guarantees as WeakPtr
3413 but in a thread-safe way.
3415 This is very similar to the fix that was done for PlaybackSessionInterfaceAVKit in
3418 * platform/cocoa/VideoFullscreenModel.h:
3419 (WebCore::VideoFullscreenModelClient::modelDestroyed):
3420 * platform/ios/VideoFullscreenInterfaceAVKit.h:
3421 * platform/ios/VideoFullscreenInterfaceAVKit.mm:
3422 (VideoFullscreenInterfaceAVKit::setVideoFullscreenModel):
3423 (VideoFullscreenInterfaceAVKit::setVideoFullscreenChangeObserver):
3424 (VideoFullscreenInterfaceAVKit::modelDestroyed):
3425 * platform/ios/WebVideoFullscreenControllerAVKit.mm:
3426 (VideoFullscreenControllerContext::~VideoFullscreenControllerContext):
3428 2019-07-13 Zalan Bujtas <zalan@apple.com>
3430 Cannot bring up custom media controls at all on v.youku.com
3431 https://bugs.webkit.org/show_bug.cgi?id=199699
3432 <rdar://problem/51835327>
3434 Reviewed by Simon Fraser.
3436 The "find the node under the finger" heuristic should only find nodes that are visible to hit-testing.
3438 When the user taps on the screen, we run a "find the best node under the finger" heuristic and use the node's location
3439 to dispatch the associated event (e.g. mousePressed).
3440 Ideally the "best node under the finger" and the final target node for the associated event are the same.
3441 However these two methods configure the hit-testing process differently which could lead to node mismatch.
3442 The "best node" heuristic calls hit-testing with AllowChildFrameContent. This flag allows hit-testing to descend into
3443 subframes even if the subframe is not visible to hit-testing (visibility: hidden).
3444 While event dispatching never descends into subfames through hit-testing, but instead it forwards the dispatching to subframes that are visible to hit-testing.
3446 This patch addresses the mismatching node issue by calling the descending version of hit-testing with a flag that enforces visiblity check before descending into a subframe.
3448 Tests: fast/events/touch/ios/visibility-hidden-iframe-click.html
3449 fast/events/touch/ios/visibility-hidden-nested-iframe-click.html
3451 * page/ios/FrameIOS.mm:
3452 (WebCore::Frame::hitTestResultAtViewportLocation):
3453 * rendering/HitTestRequest.h:
3454 (WebCore::HitTestRequest::skipsChildFrameContentInvisibleToHitTest const):
3455 * rendering/RenderWidget.cpp:
3456 (WebCore::RenderWidget::nodeAtPoint):
3458 2019-07-13 Chris Dumez <cdumez@apple.com>
3460 Fix non thread-safe usage of makeWeakPtr() in MediaPlayerPrivateAVFoundation
3461 https://bugs.webkit.org/show_bug.cgi?id=199777
3463 Reviewed by Eric Carlson.
3465 The code was calling makeWeakPtr() on a main-thread object, from a background thread.
3466 This is not thread safe. To address the issue, this patches creates the WeakPtr ahead
3467 of time, on the main thread.
3469 * platform/graphics/avfoundation/MediaPlayerPrivateAVFoundation.cpp:
3470 (WebCore::MediaPlayerPrivateAVFoundation::MediaPlayerPrivateAVFoundation):
3471 (WebCore::MediaPlayerPrivateAVFoundation::scheduleMainThreadNotification):
3472 (WebCore::MediaPlayerPrivateAVFoundation::dispatchNotification):
3473 * platform/graphics/avfoundation/MediaPlayerPrivateAVFoundation.h:
3475 2019-07-12 Thibault Saunier <tsaunier@igalia.com>
3477 [GStreamer] Mock GStreamer realtime sources should keep a Ref of their mock realtime media sources
3478 https://bugs.webkit.org/show_bug.cgi?id=194326
3480 WrappedMockRealtimeVideoSource is a subclass of RealtimeMediaSource which is refcounted, we can't
3481 use a unique_ptr on those.
3483 Also changed m_wrappedSource type to its actual type so it is cleaner even if needed
3484 to upcast it to RealtimeMediaSource so some method that are made private in the mock
3485 can still be called.
3487 Reviewed by Youenn Fablet.
3489 This fixes MediaStream tests
3491 * platform/mediastream/gstreamer/MockGStreamerAudioCaptureSource.cpp:
3492 (WebCore::WrappedMockRealtimeAudioSource::create):
3493 (WebCore::WrappedMockRealtimeAudioSource::asRealtimeMediaSource):
3494 (WebCore::WrappedMockRealtimeAudioSource::WrappedMockRealtimeAudioSource):
3495 (WebCore::m_wrappedSource):
3496 (WebCore::MockGStreamerAudioCaptureSource::startProducingData):
3497 (WebCore::MockGStreamerAudioCaptureSource::settings):
3498 (WebCore::MockGStreamerAudioCaptureSource::capabilities):
3499 * platform/mediastream/gstreamer/MockGStreamerAudioCaptureSource.h:
3500 * platform/mediastream/gstreamer/MockGStreamerVideoCaptureSource.cpp:
3501 (WebCore::WrappedMockRealtimeVideoSource::create):
3502 (WebCore::WrappedMockRealtimeVideoSource::asRealtimeMediaSource):
3503 (WebCore::WrappedMockRealtimeVideoSource::WrappedMockRealtimeVideoSource):
3504 (WebCore::m_wrappedSource):
3505 (WebCore::MockGStreamerVideoCaptureSource::settings):
3506 (WebCore::MockGStreamerVideoCaptureSource::capabilities):
3507 * platform/mediastream/gstreamer/MockGStreamerVideoCaptureSource.h:
3509 2019-07-12 Alex Christensen <achristensen@webkit.org>
3511 Change RELEASE_ASSERT in DocumentWriter::addData to ASSERT and early return
3512 https://bugs.webkit.org/show_bug.cgi?id=199756
3513 <rdar://problem/51554775>
3515 Reviewed by Brady Eidson.
3517 Attempts to reach this assertion were unsuccessful, but sometimes this assertion crashes.
3518 Let's change it to an early return to prevent crashes.
3520 * loader/DocumentWriter.cpp:
3521 (WebCore::DocumentWriter::addData):
3523 2019-07-12 Justin Fan <justin_fan@apple.com>
3525 [WebGPU] Move error scopes out of GPUDevice for more portable error generation
3526 https://bugs.webkit.org/show_bug.cgi?id=199740
3528 Reviewed by Myles C. Maxfield.
3530 Move error generation into a separate RefCounted class to allow GPU objects to generate
3531 errors independent of any GPUDevice.
3532 Create GPUObjectBase to delegate error generation and refactor GPUBuffer to inherit from GPUObjectBase.
3534 No behavior change or new tests. Error scopes covered by error-scopes-test.html.
3536 * Modules/webgpu/WebGPUDevice.cpp:
3537 (WebCore::WebGPUDevice::WebGPUDevice): Now creates a GPUErrorGenerator.
3538 (WebCore::WebGPUDevice::createBuffer const): Pass the GPUErrorGenerator to any created GPUBuffer.
3539 (WebCore::WebGPUDevice::createBufferMapped const): Ditto.
3540 (WebCore::WebGPUDevice::popErrorScope): Shouldn't be const. Can just ask for the GPUError rather than passing a lambda.
3541 (WebCore::WebGPUDevice::pushErrorScope const): Deleted.
3542 (WebCore::WebGPUDevice::popErrorScope const): Deleted.
3543 * Modules/webgpu/WebGPUDevice.h:
3544 (WebCore::WebGPUDevice::pushErrorScope):
3546 * WebCore.xcodeproj/project.pbxproj:
3547 * platform/graphics/gpu/GPUBuffer.h: Now inherits from GPUObjectBase for error generation ease.
3548 * platform/graphics/gpu/GPUDevice.cpp:
3549 (WebCore::GPUDevice::tryCreateBuffer): Ensure GPUBuffers reference the GPUErrorGenerator.
3550 (WebCore::GPUDevice::pushErrorScope): Deleted. No longer needed here.
3551 (WebCore::GPUDevice::popErrorScope): Deleted.
3552 (WebCore::GPUDevice::registerError): Deleted.
3553 * platform/graphics/gpu/GPUDevice.h: Move error scope logic out.
3554 * platform/graphics/gpu/GPUErrorGenerator.cpp: Added.
3555 (WebCore::GPUErrorGenerator::pushErrorScope):
3556 (WebCore::GPUErrorGenerator::popErrorScope):
3557 (WebCore::GPUErrorGenerator::generateError):
3558 * platform/graphics/gpu/GPUErrorGenerator.h: Added.
3559 (WebCore::GPUErrorGenerator::create):
3560 * platform/graphics/gpu/GPUObjectBase.h: Added.
3561 (WebCore::GPUObjectBase::generateError):
3562 (WebCore::GPUObjectBase::GPUObjectBase):
3563 * platform/graphics/gpu/cocoa/GPUBufferMetal.mm: Use the GPUErrorGenerator directly during buffer creation.
3564 (WebCore::GPUBuffer::validateBufferUsage):
3565 (WebCore::GPUBuffer::tryCreate):
3566 (WebCore::GPUBuffer::GPUBuffer):
3567 * platform/graphics/gpu/cocoa/GPUQueueMetal.mm:
3568 (WebCore::GPUQueue::submit): Prevent possible null dereference.
3570 2019-07-12 Youenn Fablet <youenn@apple.com>
3572 Add release logging for quota checks
3573 https://bugs.webkit.org/show_bug.cgi?id=199697
3575 Reviewed by Alex Christensen.
3577 Log whether a request to extend quota is made and the result of the request.
3578 This logging should happen in the networking process.
3579 No change of behavior.
3581 * platform/Logging.h:
3582 * storage/StorageQuotaManager.cpp:
3583 (WebCore::StorageQuotaManager::askForMoreSpace):
3584 (WebCore::StorageQuotaManager::processPendingRequests):
3586 2019-07-12 Youenn Fablet <youenn@apple.com>
3588 Stopping a cloned MediaStream video track should not stop any other video track
3589 https://bugs.webkit.org/show_bug.cgi?id=199635
3591 Reviewed by Eric Carlson.
3593 In case a track is requesting its source to end, the
3594 RealtimeVideoSource should request its own source to end and not stop it directly.
3596 Also, if a track is removing itself as an observer to a RealtimeVideoSource, we should
3597 stop the underlying source only if this one does not have any other observer.
3598 Covered by updated test.
3600 * platform/mediastream/RealtimeMediaSource.cpp:
3601 (WebCore::RealtimeMediaSource::removeObserver):
3602 * platform/mediastream/RealtimeMediaSource.h:
3603 * platform/mediastream/RealtimeVideoSource.cpp:
3604 (WebCore::RealtimeVideoSource::requestToEnd):
3605 (WebCore::RealtimeVideoSource::stopBeingObserved):
3606 * platform/mediastream/RealtimeVideoSource.h:
3608 2019-07-12 Timothy Hatcher <timothy@apple.com>
3610 Drop DarkModeCSSEnabled as an experimental feature and always enable it.
3611 https://bugs.webkit.org/show_bug.cgi?id=199725
3612 rdar://problem/52970972
3614 Reviewed by Megan Gardner.
3616 Tests: css-dark-mode
3618 * css/CSSComputedStyleDeclaration.cpp:
3619 (WebCore::ComputedStyleExtractor::valueForPropertyInStyle):
3620 * css/MediaQueryEvaluator.cpp:
3621 (WebCore::prefersColorSchemeEvaluate):
3622 * css/MediaQueryExpression.cpp:
3623 (WebCore::featureWithValidIdent):
3624 (WebCore::isFeatureValidWithoutValue):
3625 * css/parser/CSSPropertyParser.cpp:
3626 (WebCore::CSSPropertyParser::parseSingleValue):
3627 * html/HTMLMetaElement.cpp:
3628 (WebCore::HTMLMetaElement::process):
3629 * page/RuntimeEnabledFeatures.h:
3630 (WebCore::RuntimeEnabledFeatures::setDarkModeCSSEnabled): Deleted.
3631 (WebCore::RuntimeEnabledFeatures::darkModeCSSEnabled const): Deleted.
3633 2019-07-12 Carlos Alberto Lopez Perez <clopez@igalia.com>
3635 Refactor ShadowBlur: remove some class variables and use function parameters instead.
3636 https://bugs.webkit.org/show_bug.cgi?id=199511
3638 Reviewed by Said Abou-Hallawa.
3640 On the ShadowBlur class it is confusing to know if the status of the m_layerImage buffer or the
3641 values calculated by calculateLayerBoundingRect() are valid between the different function calls.
3643 To avoid this problem, pass this values as function parameters instead of storing them in the class.
3645 No new tests, no intended behaviour change.
3647 * platform/graphics/ShadowBlur.cpp:
3648 (WebCore::ShadowBlur::ShadowBlur):
3649 (WebCore::ShadowBlur::calculateLayerBoundingRect):
3650 (WebCore::ShadowBlur::drawShadowBuffer):
3651 (WebCore::ShadowBlur::drawRectShadow): Pass the parameters from the callback.
3652 (WebCore::ShadowBlur::drawInsetShadow): Ditto.
3653 (WebCore::ShadowBlur::drawRectShadowWithoutTiling):
3654 (WebCore::ShadowBlur::drawInsetShadowWithoutTiling):
3655 (WebCore::ShadowBlur::drawRectShadowWithTiling):
3656 (WebCore::ShadowBlur::drawInsetShadowWithTiling):
3657 (WebCore::ShadowBlur::drawLayerPieces):
3658 (WebCore::ShadowBlur::drawLayerPiecesAndFillCenter):
3659 (WebCore::ShadowBlur::blurShadowBuffer):
3660 (WebCore::ShadowBlur::blurAndColorShadowBuffer):
3661 (WebCore::ShadowBlur::drawShadowLayer):
3662 * platform/graphics/ShadowBlur.h: Use a struct to pass the values calculated on calculateLayerBoundingRect().
3664 2019-07-12 Adrian Perez de Castro <aperez@igalia.com>
3666 [ATK] Avoid unneeded call to to core(selection) in listObjectForSelection()
3667 https://bugs.webkit.org/show_bug.cgi?id=199748
3668 <rdar://problem/52995908>
3670 Reviewed by Konstantin Tokarev.
3672 No new tests needed.
3674 * accessibility/atk/WebKitAccessibleInterfaceSelection.cpp:
3675 (listObjectForCoreSelection): Renamed from listObjectForSelection() and remove the
3676 unneeded call to core() now that coreSelection is passed directly to the function.
3677 (optionFromList): Change to pass coreSelection directly to listObjectForCoreSelection().
3679 2019-07-11 Chris Dumez <cdumez@apple.com>
3681 Fix non-thread safe use of makeWeakPtr() under MediaPlayerPrivateMediaFoundation::CustomVideoPresenter::processInputNotify()
3682 https://bugs.webkit.org/show_bug.cgi?id=199727
3684 Reviewed by Eric Carlson.
3686 Stop calling makeWeakPtr() on a main thread object from a background thread.
3688 * platform/graphics/win/MediaPlayerPrivateMediaFoundation.cpp:
3689 (WebCore::MediaPlayerPrivateMediaFoundation::CustomVideoPresenter::processInputNotify):
3691 2019-07-11 Youenn Fablet <youenn@apple.com>
3693 Register a MediaStreamTrack as media producer only if it is a capture track
3694 https://bugs.webkit.org/show_bug.cgi?id=199566
3695 <rdar://problem/52761361>
3697 Addressing post-landing comment to simplify setMuted call.
3700 * Modules/mediastream/MediaStreamTrack.cpp:
3701 (WebCore::MediaStreamTrack::MediaStreamTrack):
3703 2019-07-11 Youenn Fablet <youenn@apple.com>
3705 Protect CoreAudioSharedUnit::m_clients for accessing in different threads simultaneously
3706 https://bugs.webkit.org/show_bug.cgi?id=199717
3708 Reviewed by Eric Carlson.
3710 Add a lock whenever accessing to m_clients.
3711 Manual tests show that audio capture still works.
3713 * platform/mediastream/mac/CoreAudioCaptureSource.cpp:
3714 (WebCore::CoreAudioSharedUnit::addClient):
3715 (WebCore::CoreAudioSharedUnit::removeClient):
3716 (WebCore::CoreAudioSharedUnit::forEachClient const):
3717 (WebCore::CoreAudioSharedUnit::processMicrophoneSamples):
3718 (WebCore::CoreAudioSharedUnit::captureFailed):
3720 2019-07-11 Chris Dumez <cdumez@apple.com>
3722 Drop non thread-safe usage of WeakPtr in PlaybackSessionInterfaceAVKit
3723 https://bugs.webkit.org/show_bug.cgi?id=199698
3725 Reviewed by Eric Carlson.
3727 The PlaybackSessionInterfaceAVKit constructor was making a weakPtr on the UI Thread
3728 of an WebThread object. The WeakPtr would then be used as a data member throughout
3729 the class on the UIThread. This is not thread-safe.
3731 This patch switches to using a raw pointer instead of a WeakPtr. This is a partial
3732 rollout of r243337, which turned the raw pointer into a WeakPtr for hardening
3733 purposes. For extra safety, this patch updates the VideoFullscreenControllerContext
3734 so that it notifies its clients (i.e. PlaybackSessionInterfaceAVKit) that it is
3735 getting destroyed, so that they can null-out their m_playbackSessionModel data
3736 member. This gives the sames guarantees than WeakPtr but in a thread-safe way.
3738 * platform/cocoa/PlaybackSessionModel.h:
3739 (WebCore::PlaybackSessionModelClient::modelDestroyed):
3740 * platform/ios/PlaybackSessionInterfaceAVKit.h:
3741 * platform/ios/PlaybackSessionInterfaceAVKit.mm:
3742 (WebCore::PlaybackSessionInterfaceAVKit::PlaybackSessionInterfaceAVKit):
3743 (WebCore::PlaybackSessionInterfaceAVKit::~PlaybackSessionInterfaceAVKit):
3744 (WebCore::PlaybackSessionInterfaceAVKit::playbackSessionModel const):
3745 (WebCore::PlaybackSessionInterfaceAVKit::modelDestroyed):
3746 * platform/ios/WebVideoFullscreenControllerAVKit.mm:
3747 (VideoFullscreenControllerContext::~VideoFullscreenControllerContext):
3748 (VideoFullscreenControllerContext::addClient):
3749 (VideoFullscreenControllerContext::removeClient):
3751 2019-07-11 Simon Fraser <simon.fraser@apple.com>
3753 Fix builds where HAVE_DESIGN_SYSTEM_UI_FONTS is not defined.
3755 * platform/graphics/cocoa/FontCacheCoreText.cpp:
3756 (WebCore::fontWithFamilySpecialCase):
3758 2019-07-11 Justin Fan <justin_fan@apple.com>
3760 Unreviewed build fix.
3762 * Modules/webgpu/WHLSL/WHLSLRecursionChecker.cpp: Missing include.
3764 2019-07-11 Myles C. Maxfield <mmaxfield@apple.com>
3766 New York font erroneously gets synthetic bold
3767 https://bugs.webkit.org/show_bug.cgi?id=199653
3768 <rdar://problem/51692592>
3770 Reviewed by Simon Fraser.
3772 This patch adds support for -apple-system-ui-serif, -apple-system-ui-monospaced,
3773 and -apple-system-ui-rounded, behind an SPI that is off-by-default. We don't want
3774 to expose these fonts to the web because we don't a standardization story for them
3775 yet, but we do want some apps to be able to use them.
3777 WebKit clients who want to use these fonts can set
3778 -[WKPreferences _shouldAllowDesignSystemUIFonts] = YES.
3780 The patch generalizes our existing system-ui infrastructure to handle these three
3781 additional fonts. It also explicitly disables the unsupported dot-prefixed names
3782 so they don't leak out into Web content.
3784 Tests: fast/text/design-system-ui-10.html
3785 fast/text/design-system-ui-11.html
3786 fast/text/design-system-ui-12.html
3787 fast/text/design-system-ui-13.html
3788 fast/text/design-system-ui-14.html
3789 fast/text/design-system-ui-15.html
3790 fast/text/design-system-ui-16.html
3791 fast/text/design-system-ui-2.html
3792 fast/text/design-system-ui-3.html
3793 fast/text/design-system-ui-4.html
3794 fast/text/design-system-ui-5.html
3795 fast/text/design-system-ui-6.html
3796 fast/text/design-system-ui-7.html
3797 fast/text/design-system-ui-8.html
3798 fast/text/design-system-ui-9.html
3799 fast/text/design-system-ui.html
3801 * css/CSSFontFace.cpp:
3802 (WebCore::CSSFontFace::shouldAllowDesignSystemUIFonts const):
3803 * css/CSSFontFace.h:
3804 * css/CSSFontFaceSource.cpp:
3805 (WebCore::CSSFontFaceSource::load):
3806 * css/StyleResolver.cpp:
3807 (WebCore::StyleResolver::initializeFontStyle):
3808 * page/Settings.yaml:
3809 * platform/graphics/FontDescription.cpp:
3810 (WebCore::m_shouldAllowDesignSystemUIFonts):
3811 (WebCore::m_shouldAllowUserInstalledFonts): Deleted.
3812 * platform/graphics/FontDescription.h:
3813 (WebCore::FontDescription::shouldAllowDesignSystemUIFonts const):
3814 (WebCore::FontDescription::setShouldAllowDesignSystemUIFonts):
3815 (WebCore::FontDescription::operator== const):
3816 * platform/graphics/cocoa/FontCacheCoreText.cpp:
3817 (WebCore::platformFontLookupWithFamily):
3818 (WebCore::fontWithFamily):
3819 * platform/graphics/cocoa/FontCacheCoreText.h:
3820 * platform/graphics/cocoa/FontDescriptionCocoa.cpp:
3821 (WebCore::systemFontUse):
3822 (WebCore::systemFontCascadeList):
3823 (WebCore::FontCascadeDescription::effectiveFamilyCount const):
3824 (WebCore::FontCascadeDescription::effectiveFamilyAt const):
3825 (WebCore::isSystemFontString): Deleted.
3826 (WebCore::isUIFontTextStyle): Deleted.
3827 * platform/graphics/cocoa/FontPlatformDataCocoa.mm:
3828 (WebCore::FontPlatformData::ctFont const):
3829 * platform/graphics/cocoa/SystemFontDatabaseCoreText.cpp:
3830 (WebCore::SystemFontDatabaseCoreText::createSystemUI):
3831 (WebCore::SystemFontDatabaseCoreText::createDesignSystemUI):
3832 (WebCore::SystemFontDatabaseCoreText::createTextStyle):
3833 (WebCore::SystemFontDatabaseCoreText::cascadeList):
3834 (WebCore::SystemFontDatabaseCoreText::applyWeightItalicsAndFallbackBehavior):
3835 (WebCore::SystemFontDatabaseCoreText::systemFontParameters):
3836 * platform/graphics/cocoa/SystemFontDatabaseCoreText.h:
3837 * platform/graphics/ios/FontCacheIOS.mm:
3838 (WebCore::platformFontWithFamilySpecialCase):
3839 * platform/graphics/mac/FontCacheMac.mm:
3840 (WebCore::platformFontWithFamilySpecialCase):
3841 * style/StyleResolveForDocument.cpp:
3842 (WebCore::Style::resolveForDocument):
3843 * svg/graphics/SVGImage.cpp:
3844 (WebCore::SVGImage::dataChanged):
3845 * testing/InternalSettings.cpp:
3846 (WebCore::InternalSettings::Backup::Backup):
3847 (WebCore::InternalSettings::Backup::restoreTo):
3848 (WebCore::InternalSettings::setShouldAllowDesignSystemUIFonts):
3849 * testing/InternalSettings.h:
3850 * testing/InternalSettings.idl:
3852 2019-07-11 Pablo Saavedra <psaavedra@igalia.com>
3854 [WPE][GTK] Build failure with ENABLE_ACCESSIBILITY=OFF
3855 https://bugs.webkit.org/show_bug.cgi?id=199625
3857 Added ENABLE(ACCESSIBILITY) and replaced HAVE(ACCESSIBILITY)
3858 with ENABLE(ACCESSIBILITY) in the code.
3860 Additionally, the TestRunner code generator now honors the
3861 Conditional IDL format.
3863 No new tests as there is no intended functional change
3865 Reviewed by Konstantin Tokarev.
3867 * accessibility/AXObjectCache.cpp:
3868 * accessibility/AXObjectCache.h:
3869 * accessibility/AccessibilityObject.cpp:
3870 (WebCore::AccessibilityObject::detach):
3871 (WebCore::AccessibilityObject::isDetached const):
3872 * accessibility/AccessibilityObject.h:
3873 * accessibility/atk/AXObjectCacheAtk.cpp:
3874 * accessibility/atk/AccessibilityObjectAtk.cpp:
3875 * accessibility/atk/WebKitAccessible.cpp:
3876 * accessibility/atk/WebKitAccessible.h:
3877 * accessibility/atk/WebKitAccessibleHyperlink.cpp:
3878 * accessibility/atk/WebKitAccessibleHyperlink.h:
3879 * accessibility/atk/WebKitAccessibleInterfaceAction.cpp:
3880 * accessibility/atk/WebKitAccessibleInterfaceAction.h:
3881 * accessibility/atk/WebKitAccessibleInterfaceComponent.cpp:
3882 * accessibility/atk/WebKitAccessibleInterfaceComponent.h:
3883 * accessibility/atk/WebKitAccessibleInterfaceDocument.cpp:
3884 * accessibility/atk/WebKitAccessibleInterfaceDocument.h:
3885 * accessibility/atk/WebKitAccessibleInterfaceEditableText.cpp:
3886 * accessibility/atk/WebKitAccessibleInterfaceEditableText.h:
3887 * accessibility/atk/WebKitAccessibleInterfaceHyperlinkImpl.cpp:
3888 * accessibility/atk/WebKitAccessibleInterfaceHyperlinkImpl.h:
3889 * accessibility/atk/WebKitAccessibleInterfaceHypertext.cpp:
3890 * accessibility/atk/WebKitAccessibleInterfaceHypertext.h:
3891 * accessibility/atk/WebKitAccessibleInterfaceImage.cpp:
3892 * accessibility/atk/WebKitAccessibleInterfaceImage.h:
3893 * accessibility/atk/WebKitAccessibleInterfaceSelection.cpp:
3894 * accessibility/atk/WebKitAccessibleInterfaceSelection.h:
3895 * accessibility/atk/WebKitAccessibleInterfaceTable.cpp:
3896 * accessibility/atk/WebKitAccessibleInterfaceTable.h:
3897 * accessibility/atk/WebKitAccessibleInterfaceTableCell.cpp:
3898 * accessibility/atk/WebKitAccessibleInterfaceTableCell.h:
3899 * accessibility/atk/WebKitAccessibleInterfaceText.cpp:
3900 * accessibility/atk/WebKitAccessibleInterfaceText.h:
3901 * accessibility/atk/WebKitAccessibleInterfaceValue.cpp:
3902 * accessibility/atk/WebKitAccessibleInterfaceValue.h:
3903 * accessibility/atk/WebKitAccessibleUtil.cpp:
3904 * accessibility/atk/WebKitAccessibleUtil.h:
3905 * accessibility/ios/AXObjectCacheIOS.mm:
3906 * accessibility/ios/AccessibilityObjectIOS.mm:
3907 * accessibility/ios/WebAccessibilityObjectWrapperIOS.h:
3908 * accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
3909 * accessibility/mac/AXObjectCacheMac.mm:
3910 * accessibility/mac/AccessibilityObjectBase.mm:
3911 * accessibility/mac/AccessibilityObjectMac.mm:
3912 * accessibility/mac/WebAccessibilityObjectWrapperBase.mm:
3913 * accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
3914 * accessibility/win/AccessibilityObjectWin.cpp:
3915 * accessibility/win/AccessibilityObjectWrapperWin.cpp:
3917 (WebCore::Document::prepareForDestruction):
3918 * editing/FrameSelection.h:
3919 * editing/atk/FrameSelectionAtk.cpp:
3920 * html/HTMLTextFormControlElement.cpp:
3921 (WebCore::HTMLTextFormControlElement::setInnerTextValue):
3922 * testing/Internals.cpp:
3923 (WebCore::Internals::resetToConsistentState):
3925 2019-07-11 Justin Fan <justin_fan@apple.com>
3927 [WebGPU] Implement GPUError and error scopes
3928 https://bugs.webkit.org/show_bug.cgi?id=199655
3930 Reviewed by Myles C. Maxfield.
3932 Add GPUErrorFilter, GPUError, and the ability to push and pop error scopes on a GPUDevice.
3933 This will allow us to check for WebGPU errors from JavaScript.
3934 Currently, only some GPUBuffer creation errors are reported for testing; more errors will follow in later patches.
3936 Test: webgpu/error-scopes-test.html
3938 * Modules/webgpu/GPUErrorFilter.idl: Added.
3939 * Modules/webgpu/GPUOutOfMemoryError.idl: Added.
3940 * Modules/webgpu/GPUValidationError.idl: Added.
3941 * Modules/webgpu/WebGPUDevice.cpp:
git://git.webkit.org / WebKit-https.git / blob