WebKit-https.git
7 months agoMove XPCService main to a shared file calling a C function
achristensen@apple.com [Tue, 5 Feb 2019 01:31:31 +0000 (01:31 +0000)]
Move XPCService main to a shared file calling a C function
https://bugs.webkit.org/show_bug.cgi?id=194256

Reviewed by Brady Eidson.

This not only reduces a few kilobytes of duplicated binary from the XPCService executables,
It will allow me to introduce a new kind of executable which will need to parse argv and call a different C function.
This is similar to r236075 but more general.

* Shared/API/Cocoa/WKMain.h: Added.
* Shared/API/Cocoa/WKMain.mm: Added.
(WKXPCServiceMain):
* Shared/EntryPointUtilities/Cocoa/AuxiliaryProcessMain.cpp: Added.
(main):
* Shared/EntryPointUtilities/Cocoa/XPCService/XPCServiceMain.mm:
(main): Deleted.
* SourcesCocoa.txt:
* WebKit.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240956 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoMake MessageSender functions const
achristensen@apple.com [Tue, 5 Feb 2019 01:30:24 +0000 (01:30 +0000)]
Make MessageSender functions const
https://bugs.webkit.org/show_bug.cgi?id=194247

Reviewed by Brady Eidson.

* NetworkProcess/Downloads/Download.cpp:
(WebKit::Download::messageSenderConnection const):
(WebKit::Download::messageSenderDestinationID const):
(WebKit::Download::messageSenderConnection): Deleted.
(WebKit::Download::messageSenderDestinationID): Deleted.
* NetworkProcess/Downloads/Download.h:
* NetworkProcess/Downloads/PendingDownload.cpp:
(WebKit::PendingDownload::messageSenderConnection const):
(WebKit::PendingDownload::messageSenderDestinationID const):
(WebKit::PendingDownload::messageSenderConnection): Deleted.
(WebKit::PendingDownload::messageSenderDestinationID): Deleted.
* NetworkProcess/Downloads/PendingDownload.h:
* NetworkProcess/IndexedDB/WebIDBConnectionToClient.cpp:
(WebKit::WebIDBConnectionToClient::messageSenderConnection const):
(WebKit::WebIDBConnectionToClient::messageSenderConnection): Deleted.
* NetworkProcess/IndexedDB/WebIDBConnectionToClient.h:
* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::messageSenderConnection const):
(WebKit::NetworkResourceLoader::messageSenderConnection): Deleted.
* NetworkProcess/NetworkResourceLoader.h:
* NetworkProcess/NetworkSocketStream.cpp:
(WebKit::NetworkSocketStream::messageSenderConnection const):
(WebKit::NetworkSocketStream::messageSenderDestinationID const):
(WebKit::NetworkSocketStream::messageSenderConnection): Deleted.
(WebKit::NetworkSocketStream::messageSenderDestinationID): Deleted.
* NetworkProcess/NetworkSocketStream.h:
* NetworkProcess/ServiceWorker/WebSWServerConnection.h:
* NetworkProcess/ServiceWorker/WebSWServerToContextConnection.cpp:
(WebKit::WebSWServerToContextConnection::messageSenderConnection const):
(WebKit::WebSWServerToContextConnection::messageSenderDestinationID const):
(WebKit::WebSWServerToContextConnection::messageSenderConnection): Deleted.
(WebKit::WebSWServerToContextConnection::messageSenderDestinationID): Deleted.
* NetworkProcess/ServiceWorker/WebSWServerToContextConnection.h:
* Platform/IPC/MessageSender.h:
* Shared/AuxiliaryProcess.cpp:
(WebKit::AuxiliaryProcess::messageSenderConnection const):
(WebKit::AuxiliaryProcess::messageSenderDestinationID const):
(WebKit::AuxiliaryProcess::messageSenderConnection): Deleted.
(WebKit::AuxiliaryProcess::messageSenderDestinationID): Deleted.
* Shared/AuxiliaryProcess.h:
* UIProcess/WebConnectionToWebProcess.cpp:
(WebKit::WebConnectionToWebProcess::messageSenderConnection const):
(WebKit::WebConnectionToWebProcess::messageSenderDestinationID const):
(WebKit::WebConnectionToWebProcess::messageSenderConnection): Deleted.
(WebKit::WebConnectionToWebProcess::messageSenderDestinationID): Deleted.
* UIProcess/WebConnectionToWebProcess.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::messageSenderConnection const):
(WebKit::WebPageProxy::messageSenderDestinationID const):
(WebKit::WebPageProxy::messageSenderConnection): Deleted.
(WebKit::WebPageProxy::messageSenderDestinationID): Deleted.
* UIProcess/WebPageProxy.h:
* WebProcess/Databases/IndexedDB/WebIDBConnectionToServer.cpp:
(WebKit::WebIDBConnectionToServer::messageSenderConnection const):
(WebKit::WebIDBConnectionToServer::messageSenderConnection): Deleted.
* WebProcess/Databases/IndexedDB/WebIDBConnectionToServer.h:
* WebProcess/Network/WebResourceLoader.cpp:
(WebKit::WebResourceLoader::messageSenderConnection const):
(WebKit::WebResourceLoader::messageSenderDestinationID const):
(WebKit::WebResourceLoader::messageSenderConnection): Deleted.
(WebKit::WebResourceLoader::messageSenderDestinationID): Deleted.
* WebProcess/Network/WebResourceLoader.h:
* WebProcess/Network/WebSocketStream.cpp:
(WebKit::WebSocketStream::messageSenderConnection const):
(WebKit::WebSocketStream::messageSenderDestinationID const):
(WebKit::WebSocketStream::messageSenderConnection): Deleted.
(WebKit::WebSocketStream::messageSenderDestinationID): Deleted.
* WebProcess/Network/WebSocketStream.h:
* WebProcess/Storage/WebSWClientConnection.h:
* WebProcess/WebConnectionToUIProcess.cpp:
(WebKit::WebConnectionToUIProcess::messageSenderConnection const):
(WebKit::WebConnectionToUIProcess::messageSenderDestinationID const):
(WebKit::WebConnectionToUIProcess::messageSenderConnection): Deleted.
(WebKit::WebConnectionToUIProcess::messageSenderDestinationID): Deleted.
* WebProcess/WebConnectionToUIProcess.h:
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::messageSenderConnection const):
(WebKit::WebPage::messageSenderDestinationID const):
(WebKit::WebPage::messageSenderConnection): Deleted.
(WebKit::WebPage::messageSenderDestinationID): Deleted.
* WebProcess/WebPage/WebPage.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240955 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoTake additional process assertion while downloading.
beidson@apple.com [Tue, 5 Feb 2019 01:27:23 +0000 (01:27 +0000)]
Take additional process assertion while downloading.
<rdar://problem/47741356> and https://bugs.webkit.org/show_bug.cgi?id=194239

Reviewed by Chris Dumez.

When the first download starts, grab this new assertion.
When the last download ends, release it.

* Configurations/Network-iOS.entitlements:

* NetworkProcess/Downloads/DownloadManager.cpp:
(WebKit::DownloadManager::dataTaskBecameDownloadTask):
(WebKit::DownloadManager::downloadFinished):
* NetworkProcess/Downloads/DownloadManager.h:

* Platform/spi/ios/AssertionServicesSPI.h:

* UIProcess/ProcessAssertion.cpp:
(WebKit::ProcessAssertion::ProcessAssertion):
* UIProcess/ProcessAssertion.h:
(WebKit::ProcessAssertion::ProcessAssertion):

* UIProcess/WebProcessProxy.cpp:
(WebKit::WebProcessProxy::didSetAssertionState):

* UIProcess/ios/ProcessAssertionIOS.mm:
(WebKit::flagsForState):
(WebKit::reasonForState):
(WebKit::ProcessAssertion::ProcessAssertion):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240954 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agowebkitpy: Precedence of booted devices should match precedence in DEFAULT_DEVICE_TYPES
jbedard@apple.com [Tue, 5 Feb 2019 01:16:28 +0000 (01:16 +0000)]
webkitpy: Precedence of booted devices should match precedence in DEFAULT_DEVICE_TYPES
https://bugs.webkit.org/show_bug.cgi?id=194158
<rdar://problem/47737268>

Rubber-stamped by Aakash Jain.

* Scripts/webkitpy/port/device_port.py:
(DevicePort.supported_device_types): Sort supported device types by the DEFAULT_DEVICE_TYPES.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240953 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Resources: missing resource data for document on reload
drousso@apple.com [Tue, 5 Feb 2019 01:11:56 +0000 (01:11 +0000)]
Web Inspector: Resources: missing resource data for document on reload
https://bugs.webkit.org/show_bug.cgi?id=194243
<rdar://problem/47559021>

Reviewed by Joseph Pecoraro.

* UserInterface/Views/ResourceDetailsSidebarPanel.js:
(WI.ResourceDetailsSidebarPanel.prototype.set resource):
(WI.ResourceDetailsSidebarPanel.prototype._applyResourceEventListeners):

* UserInterface/Base/Object.js:
(WI.Object.removeEventListener):
Drive-by: there's no need to iterate over the entire table to check if `thisObject` exists,
as that is handled for us by `ListMultimap`, which we later call anyways.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240952 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[JSC] ExecutableToCodeBlockEdge should be smaller
ysuzuki@apple.com [Tue, 5 Feb 2019 01:09:57 +0000 (01:09 +0000)]
[JSC] ExecutableToCodeBlockEdge should be smaller
https://bugs.webkit.org/show_bug.cgi?id=194244

Reviewed by Michael Saboff.

ExecutableToCodeBlockEdge is allocated so many times. However its memory layout is not efficient.
sizeof(ExecutableToCodeBlockEdge) is 24bytes, but it discards 7bytes due to one bool m_isActive flag.
Because our size classes are rounded by 16bytes, ExecutableToCodeBlockEdge takes 32bytes. So, half of
it is wasted. We should fit it into 16bytes so that we can efficiently allocate it.

In this patch, we leverages TypeInfoMayBePrototype bit in JSTypeInfo. It is a bit special TypeInfo bit
since this is per-cell bit. We rename this to TypeInfoPerCellBit, and use it as a `m_isActive` mark in
ExecutableToCodeBlockEdge. In JSObject subclasses, we use it as MayBePrototype flag.

Since this flag is not changed in CAS style, we must not change this in concurrent threads. This is OK
for ExecutableToCodeBlockEdge's m_isActive flag since this is touched on the main thread (ScriptExecutable::installCode
does not touch it if it is called in non-main threads).

* bytecode/ExecutableToCodeBlockEdge.cpp:
(JSC::ExecutableToCodeBlockEdge::finishCreation):
(JSC::ExecutableToCodeBlockEdge::visitChildren):
(JSC::ExecutableToCodeBlockEdge::activate):
(JSC::ExecutableToCodeBlockEdge::deactivate):
(JSC::ExecutableToCodeBlockEdge::isActive const):
* bytecode/ExecutableToCodeBlockEdge.h:
* runtime/JSCell.h:
* runtime/JSCellInlines.h:
(JSC::JSCell::perCellBit const):
(JSC::JSCell::setPerCellBit):
(JSC::JSCell::mayBePrototype const): Deleted.
(JSC::JSCell::didBecomePrototype): Deleted.
* runtime/JSObject.cpp:
(JSC::JSObject::setPrototypeDirect):
* runtime/JSObject.h:
* runtime/JSObjectInlines.h:
(JSC::JSObject::mayBePrototype const):
(JSC::JSObject::didBecomePrototype):
* runtime/JSTypeInfo.h:
(JSC::TypeInfo::perCellBit):
(JSC::TypeInfo::mergeInlineTypeFlags):
(JSC::TypeInfo::mayBePrototype): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240951 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: REGRESSION: Resources: WI.TreeOutline assertions when refreshing the...
mattbaker@apple.com [Tue, 5 Feb 2019 01:08:40 +0000 (01:08 +0000)]
Web Inspector: REGRESSION: Resources: WI.TreeOutline assertions when refreshing the page
https://bugs.webkit.org/show_bug.cgi?id=194242
<rdar://problem/47802027>

Reviewed by Joseph Pecoraro.

* UserInterface/Views/FolderizedTreeElement.js:
(WI.FolderizedTreeElement.prototype.removeChildren):
FolderTreeElements are removed by the base class call to removeChildren.
Calling `removeChildren` for detached TreeElements is unnecessary.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240950 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[CG] Enable setAdditionalSupportedImageTypes for WK1
said@apple.com [Tue, 5 Feb 2019 00:47:17 +0000 (00:47 +0000)]
[CG] Enable setAdditionalSupportedImageTypes for WK1
https://bugs.webkit.org/show_bug.cgi?id=194190

Reviewed by Tim Horton.

Source/WebCore:

Move the function webCoreStringVectorFromNSStringArray from WebKit to
WebCore so it can be used by both WebKit and WebKitLegacy.

* platform/mac/StringUtilities.h:
* platform/mac/StringUtilities.mm:
(WebCore::webCoreStringVectorFromNSStringArray):

Source/WebKit:

Move the function webCoreStringVectorFromNSStringArray from WebKit to
WebCore so it can be used by both WebKit and WebKitLegacy.

* Platform/mac/StringUtilities.h:
* Platform/mac/StringUtilities.mm:
(WebKit::webCoreStringVectorFromNSStringArray): Deleted.
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]):

Source/WebKitLegacy/mac:

1. Define the preferences key AdditionalSupportedImageTypes.
2. Add additionalSupportedImageTypes as a private preference property.
3. Define the setter and the getter of the property.
4. Add function to convert from an id to an array of strings.
5. Call WebCore::setAdditionalSupportedImageTypes when preferences change.

* WebView/WebPreferenceKeysPrivate.h:
* WebView/WebPreferences.mm:
(-[WebPreferences _stringArrayValueForKey:]):
(-[WebPreferences _setStringArrayValueForKey:forKey:]):
(-[WebPreferences setAdditionalSupportedImageTypes:]):
(-[WebPreferences additionalSupportedImageTypes]):
* WebView/WebPreferencesPrivate.h:
* WebView/WebView.mm:
(-[WebView _preferencesChanged:]):

Tools:

Add a test similar to the one under WebKit Cocoa but make it create a
WebView instead of a WKWebView.

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/mac/100x100-red.tga: Added.
* TestWebKitAPI/Tests/mac/AdditionalSupportedImageTypes.html: Added.
* TestWebKitAPI/Tests/mac/AdditionalSupportedImageTypes.mm: Added.
(-[AdditionalSupportedImageTypesTest webView:didFinishLoadForFrame:]):
(TestWebKitAPI::runTest):
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240949 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoAdd WebView SPI to temporarily force light or dark appearance on a page.
timothy@apple.com [Tue, 5 Feb 2019 00:28:24 +0000 (00:28 +0000)]
Add WebView SPI to temporarily force light or dark appearance on a page.
https://bugs.webkit.org/show_bug.cgi?id=194230

Reviewed by Tim Horton.

* WebView/WebView.mm:
(-[WebView _useDarkAppearance:]): Added.
(-[WebView _setUseDarkAppearance:]): Added.
* WebView/WebViewPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240948 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: REGRESSION: clicking a selected call frame doesn't re-scroll
mattbaker@apple.com [Mon, 4 Feb 2019 23:32:46 +0000 (23:32 +0000)]
Web Inspector: REGRESSION: clicking a selected call frame doesn't re-scroll
https://bugs.webkit.org/show_bug.cgi?id=194169
<rdar://problem/47743864>

Reviewed by Devin Rousso.

* UserInterface/Views/TreeOutline.js:
(WI.TreeOutline.prototype._handleMouseDown):
Add a special case for a single-selection TreeOutline with
allowsRepeatSelection enabled. Since the element is already
selected, bypass the SelectionCongroller and dispatch an
event with event.data.selectedByUser set to true.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240947 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Styles: fix race conditions when editing
nvasilyev@apple.com [Mon, 4 Feb 2019 23:30:49 +0000 (23:30 +0000)]
Web Inspector: Styles: fix race conditions when editing
https://bugs.webkit.org/show_bug.cgi?id=192739
<rdar://problem/46752925>

Reviewed by Devin Rousso.

Source/WebInspectorUI:

Editing CSS property in the style editor syncronously updates CSSStyleDeclaration on the front-end
and asyncronously updates the backend by calling CSSAgent.setStyleText. After the new style text is applied
on the backend, CSSStyleDeclaration (on the front-end) gets updated.

Unsure there's no race conditions by introducing `_updatesInProgressCount`:

  - Increment it before calling CSSAgent.setStyleText.
  - Decrement it after CSSAgent.setStyleText is finished.

Prevent updates of CSSStyleDeclaration when _updatesInProgressCount isn't 0.

* UserInterface/Models/CSSProperty.js:
(WI.CSSProperty.prototype._updateOwnerStyleText):
* UserInterface/Models/CSSStyleDeclaration.js:
(WI.CSSStyleDeclaration):
(WI.CSSStyleDeclaration.prototype.set text): Removed.
(WI.CSSStyleDeclaration.prototype.setText): Added.
Change the setter to a method since it has side effects including an asynchronous backend call.

* UserInterface/Models/DOMNodeStyles.js:
(WI.DOMNodeStyles.prototype.changeStyleText):

* UserInterface/Views/SpreadsheetStyleProperty.js:
(WI.SpreadsheetStyleProperty.prototype.get nameTextField): Removed.
(WI.SpreadsheetStyleProperty.prototype.get valueTextField): Removed.
Drive-by: remove unused code.
LayoutTests:

* inspector/css/modify-css-property-expected.txt:
* inspector/css/modify-css-property-race-expected.txt: Added.
* inspector/css/modify-css-property-race.html: Added.
* inspector/css/modify-css-property.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240946 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[Web GPU] Code clean-up for RenderPipeline backend
justin_fan@apple.com [Mon, 4 Feb 2019 23:12:44 +0000 (23:12 +0000)]
[Web GPU] Code clean-up for RenderPipeline backend
https://bugs.webkit.org/show_bug.cgi?id=194238

Reviewed by Dean Jackson.

Replace dot syntax setters with calls to setter methods, and remove unnecessary setter calls for
the input state's descriptor arrays.

Covered by existing tests; no change in behavior.

* platform/graphics/gpu/cocoa/GPURenderPipelineMetal.mm:
(WebCore::tryCreateMtlDepthStencilState): Refactor to use implicit setters rather than dot syntax.
(WebCore::setInputStateForPipelineDescriptor): Ditto, and remove unnecessary setter calls on array objects.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240945 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUse deferrable timer to restart the Responsiveness Timer on each wheel event
benjamin@webkit.org [Mon, 4 Feb 2019 22:30:31 +0000 (22:30 +0000)]
Use deferrable timer to restart the Responsiveness Timer on each wheel event
https://bugs.webkit.org/show_bug.cgi?id=194135
Source/WebCore:

<rdar://problem/47724099>

Reviewed by Simon Fraser.

The original DeferrableOneShotTimer was not really deferrable.
What it allows is to restart the count down from scratch after
firing.

For this optimization, I want to keep the correct timing but avoid
starting a real timer every time.

I renamed DeferrableOneShotTimer to ResettableOneShotTimer and
created a real DeferrableOneShotTimer that support deadlines.

* css/CSSImageGeneratorValue.cpp:
* html/HTMLPlugInImageElement.h:
* loader/cache/CachedResource.h:
* platform/Timer.cpp:
(WebCore::DeferrableOneShotTimer::startOneShot):
(WebCore::DeferrableOneShotTimer::fired):
* platform/Timer.h:
(WebCore::TimerBase::nextFireTime const):
(WebCore::ResettableOneShotTimer::ResettableOneShotTimer):
(WebCore::DeferrableOneShotTimer::DeferrableOneShotTimer):
(WebCore::DeferrableOneShotTimer::stop):
(WebCore::DeferrableOneShotTimer::restart): Deleted.
* platform/graphics/ca/TileController.h:
* platform/graphics/cg/SubimageCacheWithTimer.h:

Source/WebKit:

Reviewed by Simon Fraser.

Simon Fraser suggested a neat improvement over my previous optimization
of ResponsivenessTimer.

Instead of reseting the deadline with every event, we can let the timer
fire and add the missing time from the last start.

I implemented that behavior in the new Deferrable Timer class and use
it from ResponsivenessTimer.

* NetworkProcess/watchos/NetworkProximityAssertion.h:
* UIProcess/ResponsivenessTimer.h:
* WebProcess/Plugins/PluginView.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240944 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUse constants for pointer types
graouts@webkit.org [Mon, 4 Feb 2019 22:28:32 +0000 (22:28 +0000)]
Use constants for pointer types
https://bugs.webkit.org/show_bug.cgi?id=194232

Reviewed by Dean Jackson.

We cannot use an enum for the pointer type since a custom pointer type can be created by developers when creating a
pointer event using JavaScript, but we can at least used string constants for the ones created internally.

* dom/PointerEvent.cpp:
(WebCore::PointerEvent::mousePointerType):
(WebCore::PointerEvent::penPointerType):
(WebCore::PointerEvent::touchPointerType):
* dom/PointerEvent.h:
* dom/ios/PointerEventIOS.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240943 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[First paint] Adjust "finishedParsingMainDocument" flag by taking deferred and async...
zalan@apple.com [Mon, 4 Feb 2019 22:19:41 +0000 (22:19 +0000)]
[First paint] Adjust "finishedParsingMainDocument" flag by taking deferred and async scripts into account.
https://bugs.webkit.org/show_bug.cgi?id=194168

Reviewed by Simon Fraser.

Source/WebCore:

First paint should not be blocked by async or deferred scripts.

* page/FrameView.cpp:
(WebCore::FrameView::qualifiesAsVisuallyNonEmpty const):

Tools:

Test if the firstVisuallyNoneEmpty milestone fires before the deferred script's post-message.

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKit/FirstVisuallyNonEmptyMilestoneWithDeferredScript.mm: Added.
(-[FirstPaintMessageHandler userContentController:didReceiveScriptMessage:]):
(-[RenderingProgressNavigationDelegate _webView:renderingProgressDidChange:]):
(TEST):
* TestWebKitAPI/Tests/WebKit/deferred-script-load.html: Added.
* TestWebKitAPI/Tests/WebKit/deferred-script.js: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240942 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoAsync overflow scroll with border-radius renders incorrectly
simon.fraser@apple.com [Mon, 4 Feb 2019 21:51:32 +0000 (21:51 +0000)]
Async overflow scroll with border-radius renders incorrectly
https://bugs.webkit.org/show_bug.cgi?id=194205
<rdar://problem/47771668>

Reviewed by Zalan Bujtas.

Source/WebCore:

When an element has composited overflow:scroll and border-radius, we need to make a layer
to clip to the inside of the border radius if necessary.

Existing code simply turned off needsDescendantsClippingLayer for composited scrolling
layers, but now we check to see if the inner border is rounded. If we have both a m_childContainmentLayer
and scrolling layers, we need to adjust the location of the scrolling layers (which are parented
in m_childContainmentLayer).

Also fix offsetFromRenderer for these layers; it's positive for layers inset from the top left
of the border box.

Tests: compositing/clipping/border-radius-async-overflow-clipping-layer.html
       compositing/clipping/border-radius-async-overflow-non-stacking.html
       compositing/clipping/border-radius-async-overflow-stacking.html

* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::updateConfiguration):
(WebCore::RenderLayerBacking::updateGeometry):
(WebCore::RenderLayerBacking::updateChildClippingStrategy): Layout is always up-to-date now, so remove the comment.

LayoutTests:

New baselines, mostly correcting offsetFromRenderer.

* compositing/clipping/border-radius-async-overflow-clipping-layer-expected.txt: Added.
* compositing/clipping/border-radius-async-overflow-clipping-layer.html: Added.
* compositing/clipping/border-radius-async-overflow-non-stacking-expected.html: Added.
* compositing/clipping/border-radius-async-overflow-non-stacking.html: Added.
* compositing/clipping/border-radius-async-overflow-stacking-expected.html: Added.
* compositing/clipping/border-radius-async-overflow-stacking.html: Added.
* compositing/scrolling/overflow-scrolling-layers-are-self-painting-expected.txt:
* platform/ios/compositing/overflow/scrolling-without-painting-expected.txt:
* platform/ios/compositing/overflow/textarea-scroll-touch-expected.txt: html.css specifies a border-radius on <textarea> for iOS, so we make additional
clipping layers.
* platform/ios/compositing/rtl/rtl-scrolling-with-transformed-descendants-expected.txt:
* platform/ios/compositing/scrolling/overflow-scrolling-layers-are-self-painting-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240941 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoPageOverlayController's layers should be created lazily
simon.fraser@apple.com [Mon, 4 Feb 2019 21:16:22 +0000 (21:16 +0000)]
PageOverlayController's layers should be created lazily
https://bugs.webkit.org/show_bug.cgi?id=194199
Source/WebCore:

Reviewed by Tim Horton.

Expose PageOverlayController::hasDocumentOverlays() and hasViewOverlays()
and use them to only parent the overlay-hosting layers when necessary.

For document overlays, RenderLayerCompositor::appendDocumentOverlayLayers() can
simply do nothing if there are none. Updates are triggered via Page::installedPageOverlaysChanged(),
which calls FrameView::setNeedsCompositingConfigurationUpdate() to trigger the root layer
compositing updates that parents the layerWithDocumentOverlays().

View overlays are added to the layer tree via the DrawingArea. When we go between having
none and some view overlays, Page::installedPageOverlaysChanged() calls attachViewOverlayGraphicsLayer()
on the ChromeClient, and the DrawingArea responds by calling updateRootLayers() and scheduling a
compositing flush (this has to be done manually because view overlay layers are outside the
subtree managed by RenderLayerCompositor).

Now that GraphicsLayers are ref-counted, we can let the DrawingArea simply retain its m_viewOverlayRootLayer;
there is no need for RenderLayerCompositor::attachRootLayer()/detachRootLayer() to do anything with view
overlay layers. This implies that a page can navigate (new FrameView) and view overlays will persist, without
having to be manually removed and re-added. We can also remove the Frame argument to attachViewOverlayGraphicsLayer().

* loader/EmptyClients.h:
* page/ChromeClient.h:
* page/FrameView.cpp:
(WebCore::FrameView::setNeedsCompositingConfigurationUpdate): These functions need to schedule a compositing flush
because there may be nothing else that does.
(WebCore::FrameView::setNeedsCompositingGeometryUpdate):
* page/Page.cpp:
(WebCore::Page::installedPageOverlaysChanged):
* page/Page.h:
* page/PageOverlayController.cpp:
(WebCore::PageOverlayController::hasDocumentOverlays const):
(WebCore::PageOverlayController::hasViewOverlays const):
(WebCore::PageOverlayController::attachViewOverlayLayers): PageOverlayController has the Page so it
might as well be the one to call through the ChromeClient.
(WebCore::PageOverlayController::detachViewOverlayLayers):
(WebCore::PageOverlayController::installPageOverlay):
(WebCore::PageOverlayController::uninstallPageOverlay):
* page/PageOverlayController.h:
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::updateCompositingLayers): isFullUpdate is always true; remove it.
(WebCore::RenderLayerCompositor::appendDocumentOverlayLayers):
(WebCore::RenderLayerCompositor::attachRootLayer):
(WebCore::RenderLayerCompositor::detachRootLayer):

Source/WebKit:

rdar://problem/46571593

Reviewed by Tim Horton.

Expose PageOverlayController::hasDocumentOverlays() and hasViewOverlays()
and use them to only parent the overlay-hosting layers when necessary.

For document overlays, RenderLayerCompositor::appendDocumentOverlayLayers() can
simply do nothing if there are none. Updates are triggered via Page::installedPageOverlaysChanged(),
which calls FrameView::setNeedsCompositingConfigurationUpdate() to trigger the root layer
compositing updates that parents the layerWithDocumentOverlays().

View overlays are added to the layer tree via the DrawingArea. When we go between having
none and some view overlays, Page::installedPageOverlaysChanged() calls attachViewOverlayGraphicsLayer()
on the ChromeClient, and the DrawingArea responds by calling updateRootLayers() and scheduling a
compositing flush (this has to be done manually because view overlay layers are outside the
subtree managed by RenderLayerCompositor).

Now that GraphicsLayers are ref-counted, we can let the DrawingArea simply retain its m_viewOverlayRootLayer;
there is no need for RenderLayerCompositor::attachRootLayer()/detachRootLayer() to do anything with view
overlay layers. This implies that a page can navigate (new FrameView) and view overlays will persist, without
having to be manually removed and re-added. We can also remove the Frame argument to attachViewOverlayGraphicsLayer().

* WebProcess/WebCoreSupport/WebChromeClient.cpp:
(WebKit::WebChromeClient::attachViewOverlayGraphicsLayer):
* WebProcess/WebCoreSupport/WebChromeClient.h:
* WebProcess/WebPage/AcceleratedDrawingArea.cpp:
(WebKit::AcceleratedDrawingArea::attachViewOverlayGraphicsLayer):
* WebProcess/WebPage/AcceleratedDrawingArea.h:
* WebProcess/WebPage/DrawingArea.h:
(WebKit::DrawingArea::attachViewOverlayGraphicsLayer):
* WebProcess/WebPage/RemoteLayerTree/RemoteLayerTreeDrawingArea.h:
* WebProcess/WebPage/RemoteLayerTree/RemoteLayerTreeDrawingArea.mm:
(WebKit::RemoteLayerTreeDrawingArea::attachViewOverlayGraphicsLayer):
* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.h:
* WebProcess/WebPage/mac/TiledCoreAnimationDrawingArea.mm:
(WebKit::TiledCoreAnimationDrawingArea::attachViewOverlayGraphicsLayer):
(WebKit::TiledCoreAnimationDrawingArea::mainFrameContentSizeChanged):

Source/WebKitLegacy/mac:

rdar://problem/46571593

Reviewed by Tim Horton.

* WebCoreSupport/WebChromeClient.h:
* WebCoreSupport/WebChromeClient.mm:
(WebChromeClient::attachViewOverlayGraphicsLayer):

Source/WebKitLegacy/win:

rdar://problem/46571593

Reviewed by Tim Horton.

Expose PageOverlayController::hasDocumentOverlays() and hasViewOverlays()
and use them to only parent the overlay-hosting layers when necessary.

For document overlays, RenderLayerCompositor::appendDocumentOverlayLayers() can
simply do nothing if there are none. Updates are triggered via Page::installedPageOverlaysChanged(),
which calls FrameView::setNeedsCompositingConfigurationUpdate() to trigger the root layer
compositing updates that parents the layerWithDocumentOverlays().

View overlays are added to the layer tree via the DrawingArea. When we go between having
none and some view overlays, Page::installedPageOverlaysChanged() calls attachViewOverlayGraphicsLayer()
on the ChromeClient, and the DrawingArea responds by calling updateRootLayers() and scheduling a
compositing flush (this has to be done manually because view overlay layers are outside the
subtree managed by RenderLayerCompositor).

Now that GraphicsLayers are ref-counted, we can let the DrawingArea simply retain its m_viewOverlayRootLayer;
there is no need for RenderLayerCompositor::attachRootLayer()/detachRootLayer() to do anything with view
overlay layers. This implies that a page can navigate (new FrameView) and view overlays will persist, without
having to be manually removed and re-added. We can also remove the Frame argument to attachViewOverlayGraphicsLayer().

* WebCoreSupport/WebChromeClient.cpp:
(WebChromeClient::attachViewOverlayGraphicsLayer):
* WebCoreSupport/WebChromeClient.h:

LayoutTests:

rdar://problem/46571593

Reviewed by Tim Horton.

* pageoverlay/overlay-remove-reinsert-view-expected.txt: We no longer unparent the overlays
on view removal, so new results.
* platform/ios-wk2/TestExpectations: Unskip some iOS tests.
* platform/ios-wk2/pageoverlay/overlay-installation-expected.txt: Added.
* platform/ios-wk2/pageoverlay/overlay-large-document-expected.txt: Added.
* platform/ios-wk2/pageoverlay/overlay-large-document-scrolled-expected.txt: Added.
* platform/ios/TestExpectations: Unskip some iOS tests.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240940 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoSource/WebCore:
commit-queue@webkit.org [Mon, 4 Feb 2019 21:03:20 +0000 (21:03 +0000)]
Source/WebCore:
When performing Increment or Decrement on sliders, check to see if the slider is disabled.
https://bugs.webkit.org/show_bug.cgi?id=173497

Patch by Eric Liang <ericliang@apple.com> on 2019-02-04
Reviewed by Chris Fleizach.

Test: accessibility/set-value-not-work-for-disabled-sliders.html

* accessibility/AccessibilityNodeObject.cpp:
(WebCore::AccessibilityNodeObject::alterSliderValue):

LayoutTests:
Check if slider value changed after calling AX Increment or Decrement on disabled sliders.
https://bugs.webkit.org/show_bug.cgi?id=193497

Patch by Eric Liang <ericliang@apple.com> on 2019-02-04
Reviewed by Chris Fleizach.

* accessibility/set-value-not-work-for-disabled-sliders.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240939 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[JSC] Shrink size of FunctionExecutable
ysuzuki@apple.com [Mon, 4 Feb 2019 21:02:27 +0000 (21:02 +0000)]
[JSC] Shrink size of FunctionExecutable
https://bugs.webkit.org/show_bug.cgi?id=194191

Reviewed by Michael Saboff.

This patch reduces the size of FunctionExecutable. Since it is allocated in IsoSubspace, reducing the size directly
improves the allocation efficiency.

1. ScriptExecutable (base class of FunctionExecutable) has several members, but it is meaningful only in FunctionExecutable.
   We remove this from ScriptExecutable, and move it to FunctionExecutable.

2. FunctionExecutable has several data which are rarely used. One for FunctionOverrides functionality, which is typically
   used for JSC debugging purpose, and another is TypeSet and offsets for type profiler. We move them to RareData and reduce
   the size of FunctionExecutable in the common case.

This patch changes the size of FunctionExecutable from 176 to 144.

* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpSource):
(JSC::CodeBlock::finishCreation):
* dfg/DFGNode.h:
(JSC::DFG::Node::OpInfoWrapper::as const):
* interpreter/StackVisitor.cpp:
(JSC::StackVisitor::Frame::computeLineAndColumn const):
* runtime/ExecutableBase.h:
* runtime/FunctionExecutable.cpp:
(JSC::FunctionExecutable::FunctionExecutable):
(JSC::FunctionExecutable::ensureRareDataSlow):
* runtime/FunctionExecutable.h:
* runtime/Intrinsic.h:
* runtime/ModuleProgramExecutable.cpp:
(JSC::ModuleProgramExecutable::ModuleProgramExecutable):
* runtime/ProgramExecutable.cpp:
(JSC::ProgramExecutable::ProgramExecutable):
* runtime/ScriptExecutable.cpp:
(JSC::ScriptExecutable::ScriptExecutable):
(JSC::ScriptExecutable::overrideLineNumber const):
(JSC::ScriptExecutable::typeProfilingStartOffset const):
(JSC::ScriptExecutable::typeProfilingEndOffset const):
* runtime/ScriptExecutable.h:
(JSC::ScriptExecutable::firstLine const):
(JSC::ScriptExecutable::setOverrideLineNumber): Deleted.
(JSC::ScriptExecutable::hasOverrideLineNumber const): Deleted.
(JSC::ScriptExecutable::overrideLineNumber const): Deleted.
(JSC::ScriptExecutable::typeProfilingStartOffset const): Deleted.
(JSC::ScriptExecutable::typeProfilingEndOffset const): Deleted.
* runtime/StackFrame.cpp:
(JSC::StackFrame::computeLineAndColumn const):
* tools/JSDollarVM.cpp:
(JSC::functionReturnTypeFor):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240938 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUnreviewed, improve preprocessor guard
mcatanzaro@igalia.com [Mon, 4 Feb 2019 20:49:37 +0000 (20:49 +0000)]
Unreviewed, improve preprocessor guard
https://bugs.webkit.org/show_bug.cgi?id=194166
<rdar://problem/47694328>

AC mode is only optional in GTK port. Requested by Zan.

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::updatePreferences):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240937 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoREGRESSION (r237269): webkit-patch what-broke is broken
dean_johnson@apple.com [Mon, 4 Feb 2019 19:49:08 +0000 (19:49 +0000)]
REGRESSION (r237269): webkit-patch what-broke is broken
https://bugs.webkit.org/show_bug.cgi?id=193025

Reviewed by Aakash Jain.

* Scripts/webkitpy/common/host.py:
(Host.buildbot):
(Host.bugzilla): Renamed to 'buildbot'.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240936 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[iOS] API test WKAttachmentTests.InsertAndRemoveDuplicateAttachment is failing after...
wenson_hsieh@apple.com [Mon, 4 Feb 2019 18:59:40 +0000 (18:59 +0000)]
[iOS] API test WKAttachmentTests.InsertAndRemoveDuplicateAttachment is failing after r240902
https://bugs.webkit.org/show_bug.cgi?id=194207

Reviewed by Tim Horton.

Fix this API test by making it robust against smart paste on iOS. Currently, this test pastes an attachment
element and assumes that the pasted attachment is adjacent to the original one. However, after enabling smart
paste, we now insert an extra space in between. Instead of executing "DeleteBackward" twice to test removing the
attachment elements, we delete backwards once to remove the pasted attachment, and move the selection to the
start and delete forwards to remove the original attachment.

* TestWebKitAPI/Tests/WebKitCocoa/WKAttachmentTests.mm:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240934 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoAdd demos for Intersection Observer API
jond@apple.com [Mon, 4 Feb 2019 18:55:31 +0000 (18:55 +0000)]
Add demos for Intersection Observer API
https://bugs.webkit.org/show_bug.cgi?id=194219

Reviewed by Simon Fraser.

* demos/intersection-observer/lazy-image-load/index.html: Added.
* demos/intersection-observer/lazy-image-load/scripts.js: Added.
(ImageLoader):
(ImageLoader.prototype.findImages):
(ImageLoader.prototype.smallURLForImage):
(ImageLoader.prototype.largeURLForImage):
(ImageLoader.prototype.setupObserver):
(ImageLoader.prototype.intersectionsChanged):
* demos/intersection-observer/lazy-image-load/style.css: Added.
(body):
(section):
(header):
(.main):
(.main img):
(.stories):
(.world, .local):
(.world):
(.local):
(.ad):
(.contents):
(.sidebar):
(.links):
(.main-story):
(.main-story img):
(.top-story):
(.top-story img):
(.mid-story img):
(.link-group):
(.link-group img):
(.link-group ul):
(.link-group li):
(footer):
(footer ul):
(footer a):
(.footer-col):
* demos/intersection-observer/simple/script.js: Added.
(addToLog):
(clearLog):
(stringFromRect):
(intersectedCallback):
* demos/intersection-observer/simple/simple-observer-iframe.html: Added.
* demos/intersection-observer/simple/simple-observer-root-margin.html: Added.
* demos/intersection-observer/simple/simple-observer.html: Added.
* demos/intersection-observer/simple/styles.css: Added.
(body):
(.results):
(.results button):
(h2):
(#logging):
* demos/intersection-observer/triggered-animation/index.html: Added.
* demos/intersection-observer/triggered-animation/script.js: Added.
(AnimationManager):
(AnimationManager.prototype.setupObserver):
(AnimationManager.prototype.intersectionsChanged):
* demos/intersection-observer/triggered-animation/style.css: Added.
(body):
(p):
(.animation-container h1):
(.animation-container):
(.animation-container .box):
(.slide.box):
(.animation-container.visible > .slide.box):
(.slide.animation-container div:nth-of-type(1)):
(.slide.animation-container div:nth-of-type(2)):
(.slide.animation-container div:nth-of-type(3)):
(.spin):
(.animation-container.visible .spin):
(@keyframes spin):
(to):
(.animation-container.svg):
(.animation-container.svg path):
(.animation-container.svg.visible):
(@keyframes stroke-move):
(.animation-container.svg circle):
(.animation-container.svg.visible circle):
(.animation-container.svg circle:nth-of-type(2)):
(.animation-container.svg circle:nth-of-type(3)):
(.animation-container.svg circle:nth-of-type(4)):
(.animation-container.svg circle:nth-of-type(5)):
(@keyframes circle-scale):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240932 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoIndexedDB: leak WebIDBConnectionToServer in layout tests
sihui_liu@apple.com [Mon, 4 Feb 2019 18:51:19 +0000 (18:51 +0000)]
IndexedDB: leak WebIDBConnectionToServer in layout tests
https://bugs.webkit.org/show_bug.cgi?id=193688
<rdar://problem/47353263>

Reviewed by Geoffrey Garen.

Source/WebCore:

Let IDBConnectionToServer keep a WeakPtr of IDBConnectionToServerDelegate.

* Modules/indexeddb/client/IDBConnectionToServer.cpp:
(WebCore::IDBClient::IDBConnectionToServer::IDBConnectionToServer):
* Modules/indexeddb/client/IDBConnectionToServer.h:
* Modules/indexeddb/client/IDBConnectionToServerDelegate.h:

Source/WebKit:

* NetworkProcess/NetworkConnectionToWebProcess.cpp:
(WebKit::NetworkConnectionToWebProcess::didClose):
(WebKit::NetworkConnectionToWebProcess::establishIDBConnectionToServer):
(WebKit::NetworkConnectionToWebProcess::removeIDBConnectionToServer): Deleted.
* NetworkProcess/NetworkConnectionToWebProcess.h:
* NetworkProcess/NetworkConnectionToWebProcess.messages.in:
* WebProcess/Databases/IndexedDB/WebIDBConnectionToServer.cpp:
(WebKit::WebIDBConnectionToServer::~WebIDBConnectionToServer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240931 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoMake sure to remove the device observer in AVVideoCaptureSource
youenn@apple.com [Mon, 4 Feb 2019 18:47:23 +0000 (18:47 +0000)]
Make sure to remove the device observer in AVVideoCaptureSource
https://bugs.webkit.org/show_bug.cgi?id=194181
<rdar://problem/47739247>

Reviewed by Eric Carlson.

Make sure to remove the device observer when the observer is destroyed.
To simplify things, add the observer in AVVideoCaptureSource constructor and remove it in the destructor.

Make also sure the session observer is also removed whenever the session is released by AVVideoCaptureSource.

Covered by manual test.

       * platform/mediastream/mac/AVVideoCaptureSource.h:
* platform/mediastream/mac/AVVideoCaptureSource.mm:
(WebCore::AVVideoCaptureSource::AVVideoCaptureSource):
(WebCore::AVVideoCaptureSource::~AVVideoCaptureSource):
(WebCore::AVVideoCaptureSource::initializeSession):
(WebCore::AVVideoCaptureSource::clearSession):
(WebCore::AVVideoCaptureSource::stopProducingData):
(WebCore::AVVideoCaptureSource::setupSession):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240930 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago<rdar://problem/47788802>
graouts@webkit.org [Mon, 4 Feb 2019 18:19:10 +0000 (18:19 +0000)]
<rdar://problem/47788802>

Unreviewed build fix.

* Platform/spi/ios/UIKitSPI.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240929 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoCapture state should be managed consistently when doing process swapping
youenn@apple.com [Mon, 4 Feb 2019 17:37:01 +0000 (17:37 +0000)]
Capture state should be managed consistently when doing process swapping
https://bugs.webkit.org/show_bug.cgi?id=194122
<rdar://problem/47609293>

Reviewed by Eric Carlson.

Source/WebKit:

When doing PSON, WebPageProxy::resetState is called.
It resets the media state, but does not call the client delegates.
Instead of directly updating the media state, call the routine used to update it so that client delegates are called.

Covered by new API test and layout test.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _mediaCaptureState]):
* UIProcess/API/Cocoa/WKWebViewPrivate.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::resetState):
(WebKit::WebPageProxy::isPlayingMediaDidChange):
(WebKit::WebPageProxy::updatePlayingMediaDidChange):
* UIProcess/WebPageProxy.h:

Tools:

* TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:
(-[GetUserMediaUIDelegate _webView:requestUserMediaAuthorizationForDevices:url:mainFrameURL:decisionHandler:]):
(-[GetUserMediaUIDelegate _webView:checkUserMediaPermissionForURL:mainFrameURL:frameIdentifier:decisionHandler:]):
(-[GetUserMediaUIDelegate _webView:mediaCaptureStateDidChange:]):
* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::isDoingMediaCapture const):
* WebKitTestRunner/InjectedBundle/TestRunner.h:
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::isDoingMediaCapture const):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
* WebKitTestRunner/cocoa/TestControllerCocoa.mm:
(WTR::TestController::isDoingMediaCapture const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240928 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUse a dedicated type instead of int32_t for pointer identifiers
commit-queue@webkit.org [Mon, 4 Feb 2019 16:22:49 +0000 (16:22 +0000)]
Use a dedicated type instead of int32_t for pointer identifiers
https://bugs.webkit.org/show_bug.cgi?id=194217

Patch by Antoine Quint <graouts@apple.com> on 2019-02-04
Reviewed by Antti Koivisto.

Source/WebCore:

* WebCore.xcodeproj/project.pbxproj:
* dom/PointerEvent.h:
* dom/PointerID.h: Added.
* page/PointerCaptureController.cpp:
(WebCore::PointerCaptureController::setPointerCapture):
(WebCore::PointerCaptureController::releasePointerCapture):
(WebCore::PointerCaptureController::hasPointerCapture):
(WebCore::PointerCaptureController::touchEndedOrWasCancelledForIdentifier):
(WebCore::PointerCaptureController::hasCancelledPointerEventForIdentifier):
(WebCore::PointerCaptureController::cancelPointer):
* page/PointerCaptureController.h:

Source/WebKit:

* Scripts/webkit/messages.py:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::cancelPointer):
* UIProcess/WebPageProxy.h:
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::cancelPointer):
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240927 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[iOS] Unable to make a selection in jsfiddle.net using arrow keys when requesting...
wenson_hsieh@apple.com [Mon, 4 Feb 2019 16:05:18 +0000 (16:05 +0000)]
[iOS] Unable to make a selection in jsfiddle.net using arrow keys when requesting desktop site
Followup to https://bugs.webkit.org/show_bug.cgi?id=193758

Reviewed by Daniel Bates.

Put the iOS-specific behavior behind an EditingBehavior check, rather than a compile-time guard. No change in
behavior.

* editing/EditingBehavior.h:
(WebCore::EditingBehavior::shouldMoveSelectionToEndWhenFocusingTextInput const):
* html/HTMLInputElement.cpp:
(WebCore::HTMLInputElement::setDefaultSelectionAfterFocus):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240926 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUnreviewed, avoid -Wswitch warnings introduced in r240880
mcatanzaro@igalia.com [Mon, 4 Feb 2019 16:02:28 +0000 (16:02 +0000)]
Unreviewed, avoid -Wswitch warnings introduced in r240880
https://bugs.webkit.org/show_bug.cgi?id=193740
<rdar://problem/47527267>

* NetworkProcess/soup/NetworkDataTaskSoup.cpp:
(WebKit::NetworkDataTaskSoup::dispatchDidReceiveResponse):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240925 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[LFC][IFC] Make InlineFormattingContext::collectInlineContent non-recursive.
zalan@apple.com [Mon, 4 Feb 2019 15:49:24 +0000 (15:49 +0000)]
[LFC][IFC] Make InlineFormattingContext::collectInlineContent non-recursive.
https://bugs.webkit.org/show_bug.cgi?id=194210

Reviewed by Antti Koivisto.

Use iterative algorithm to collect inline content (and add breaking rules).
This is in preparation for fixing the inline preferred width computation.

* layout/Verification.cpp:
(WebCore::Layout::resolveForRelativePositionIfNeeded):
* layout/inlineformatting/InlineFormattingContext.cpp:
(WebCore::Layout::addDetachingRules):
(WebCore::Layout::createAndAppendInlineItem):
(WebCore::Layout::InlineFormattingContext::collectInlineContent const):
(WebCore::Layout::InlineFormattingContext::collectInlineContentForSubtree const): Deleted.
* layout/inlineformatting/InlineFormattingContext.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240924 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUnreviewed, fix unused variable warnings introduced in r240912
mcatanzaro@igalia.com [Mon, 4 Feb 2019 15:48:44 +0000 (15:48 +0000)]
Unreviewed, fix unused variable warnings introduced in r240912
https://bugs.webkit.org/show_bug.cgi?id=194198
<rdar://problem/47776051>

* page/FrameView.cpp:
(WebCore::FrameView::setNeedsCompositingConfigurationUpdate):
(WebCore::FrameView::setNeedsCompositingGeometryUpdate):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240923 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[GTK] Allow pinch zoom on touchpad
commit-queue@webkit.org [Mon, 4 Feb 2019 15:47:07 +0000 (15:47 +0000)]
[GTK] Allow pinch zoom on touchpad
https://bugs.webkit.org/show_bug.cgi?id=194201

Patch by Alexander Mikhaylenko <exalm7659@gmail.com> on 2019-02-04
Reviewed by Michael Catanzaro.

Enable touchpad events for WebkitWebViewBase, then feed touchpad
pinch events to GestureController.

* UIProcess/API/gtk/WebKitWebViewBase.cpp:
(webkitWebViewBaseRealize): Added GDK_TOUCHPAD_GESTURE_MASK to event mask.
(webkitWebViewBaseEvent): Added.
(webkit_web_view_base_class_init): Override event vfunc.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240922 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[css-scroll-snap] scroll-snap-align not honored on child with non-visible overflow
commit-queue@webkit.org [Mon, 4 Feb 2019 15:39:48 +0000 (15:39 +0000)]
[css-scroll-snap] scroll-snap-align not honored on child with non-visible overflow
https://bugs.webkit.org/show_bug.cgi?id=191816

Patch by Frederic Wang <fwang@igalia.com> on 2019-02-04
Reviewed by Wenson Hsieh.

Source/WebCore:

This patch fixes a bug that prevents children of a scroll container to create snap positions
when they have non-visible overflow. This happens because for such a child, the function
RenderBox::findEnclosingScrollableContainer() will return the child itself rather than the
scroll container. To address that issue, we introduce a new
RenderObject::enclosingScrollableContainerForSnapping() helper function that ensures that
a real RenderBox ancestor is returned.

Test: css3/scroll-snap/scroll-snap-children-with-overflow.html

* page/scrolling/AxisScrollSnapOffsets.cpp:
(WebCore::updateSnapOffsetsForScrollableArea): Use enclosingScrollableContainerForSnapping()
so that we don't skip children with non-visible overflow.
* rendering/RenderLayerModelObject.cpp:
(WebCore::RenderLayerModelObject::styleDidChange): Ditto. The new function calls
enclosingBox().
* rendering/RenderObject.cpp:
(WebCore::RenderObject::enclosingScrollableContainerForSnapping const): Return
the scrollable container of the enclosing box. If it is actually the render object itself
then start the search from the parent box instead.
* rendering/RenderObject.h: Declare enclosingScrollableContainerForSnapping().

LayoutTests:

Add a test to verify that children with non-visible overflow create snap offsets.

* css3/scroll-snap/scroll-snap-children-with-overflow-expected.txt: Added.
* css3/scroll-snap/scroll-snap-children-with-overflow.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240921 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUnreviewed, only force settings.acceleratedCompositingEnabled() to true for Cocoa...
mcatanzaro@igalia.com [Mon, 4 Feb 2019 15:36:14 +0000 (15:36 +0000)]
Unreviewed, only force settings.acceleratedCompositingEnabled() to true for Cocoa ports
https://bugs.webkit.org/show_bug.cgi?id=194166
<rdar://problem/47694328>

Accelerated compositing is broken for me at least. We're a long way from being able to force
it.

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::updatePreferences):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240920 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[GTK] Remove last python2 dependency from CMake build process
mcatanzaro@igalia.com [Mon, 4 Feb 2019 15:00:12 +0000 (15:00 +0000)]
[GTK] Remove last python2 dependency from CMake build process
https://bugs.webkit.org/show_bug.cgi?id=194218

Patch by Abderrahim Kitouni <akitouni@gnome.org> on 2019-02-04
Reviewed by Michael Catanzaro.

* InspectorGResources.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240919 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoSource/WebCore:
antti@apple.com [Mon, 4 Feb 2019 12:48:21 +0000 (12:48 +0000)]
Source/WebCore:
Rename GraphicsLayer and PlatformCALayer scrolling layer type enum values to be less ambiguous
https://bugs.webkit.org/show_bug.cgi?id=194215

Reviewed by Frédéric Wang.

GraphicsLayer::Type::Scrolling -> GraphicsLayer::Type::ScrollContainer
PlatformCALayer::LayerTypeScrollingLayer -> PlatformCALayer::LayerTypeScrollContainerLayer

* platform/graphics/GraphicsLayer.cpp:
(WebCore::GraphicsLayer::supportsLayerType):
* platform/graphics/GraphicsLayer.h:
* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayer::supportsLayerType):
(WebCore::GraphicsLayerCA::initialize):
* platform/graphics/ca/PlatformCALayer.cpp:
(WebCore::operator<<):
* platform/graphics/ca/PlatformCALayer.h:
* platform/graphics/ca/cocoa/PlatformCALayerCocoa.mm:
(WebCore::PlatformCALayerCocoa::PlatformCALayerCocoa):
(WebCore::PlatformCALayerCocoa::commonInit):
* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::updateScrollingLayers):
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::ensureRootLayer):

Source/WebKit:
Rename GraphicsLayer and PlatformCALayer type enum values to match "scroll container layer" convention
https://bugs.webkit.org/show_bug.cgi?id=194215

Reviewed by Frédéric Wang.

* Shared/RemoteLayerTree/RemoteLayerBackingStore.mm:
(WebKit::RemoteLayerBackingStore::drawInContext):
* UIProcess/RemoteLayerTree/RemoteLayerTreeHost.mm:
(WebKit::RemoteLayerTreeHost::makeNode):
* UIProcess/RemoteLayerTree/ios/RemoteLayerTreeHostIOS.mm:
(WebKit::RemoteLayerTreeHost::makeNode):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240918 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoDFG's doesGC() is incorrect about the SameValue node's behavior.
mark.lam@apple.com [Mon, 4 Feb 2019 08:15:46 +0000 (08:15 +0000)]
DFG's doesGC() is incorrect about the SameValue node's behavior.
https://bugs.webkit.org/show_bug.cgi?id=194211
<rdar://problem/47608913>

Reviewed by Saam Barati.

Only the DoubleRepUse case is guaranteed to not GC.  The other case may GC because
it calls operationSameValue() which may allocate memory for resolving ropes.

* dfg/DFGDoesGC.cpp:
(JSC::DFG::doesGC):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240917 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[iOS] Tiles not created in large scrollable iframes
antti@apple.com [Mon, 4 Feb 2019 07:49:33 +0000 (07:49 +0000)]
[iOS] Tiles not created in large scrollable iframes
https://bugs.webkit.org/show_bug.cgi?id=193665

Reviewed by Simon Fraser.

Source/WebCore:

We are not syncing scroll position back to the graphics layer tree correctly.

Test by Frédéric Wang.

* page/scrolling/AsyncScrollingCoordinator.cpp:
(WebCore::AsyncScrollingCoordinator::frameViewRootLayerDidChange):
(WebCore::AsyncScrollingCoordinator::reconcileScrollingState):
(WebCore::AsyncScrollingCoordinator::reconcileScrollPosition):

Factor setting and syncing scrolling layer positions into a function.
Use bounds.origin scrolling mechanic when scrollContainerLayer is present.

(WebCore::AsyncScrollingCoordinator::scrollableAreaScrollbarLayerDidChange):
(WebCore::AsyncScrollingCoordinator::setSynchronousScrollingReasons):
(WebCore::AsyncScrollingCoordinator::updateScrollLayerPosition): Deleted.
* page/scrolling/AsyncScrollingCoordinator.h:
* page/scrolling/ScrollingCoordinator.cpp:
(WebCore::ScrollingCoordinator::scrollContainerLayerForFrameView):
(WebCore::ScrollingCoordinator::scrolledContentsLayerForFrameView):
(WebCore::ScrollingCoordinator::scrollLayerForFrameView): Deleted.
* page/scrolling/ScrollingCoordinator.h:
* rendering/RenderLayerCompositor.cpp:

Rename scrollLayer to scrolledContentsLayer according to out preferred naming scheme and use it in that role only.
Add scrollContainerLayer as a separate layer. It is only constructed when using async scrolling on iOS.

(WebCore::RenderLayerCompositor::~RenderLayerCompositor):
(WebCore::RenderLayerCompositor::customPositionForVisibleRectComputation const):
(WebCore::RenderLayerCompositor::visibleRectForLayerFlushing const):
(WebCore::RenderLayerCompositor::didChangePlatformLayerForLayer):
(WebCore::RenderLayerCompositor::frameViewDidChangeSize):
(WebCore::RenderLayerCompositor::updateScrollLayerPosition):
(WebCore::RenderLayerCompositor::frameViewDidScroll):
(WebCore::RenderLayerCompositor::updateLayerForTopOverhangArea):
(WebCore::RenderLayerCompositor::updateLayerForBottomOverhangArea):
(WebCore::RenderLayerCompositor::updateLayerForHeader):
(WebCore::RenderLayerCompositor::updateLayerForFooter):
(WebCore::RenderLayerCompositor::updateOverflowControlsLayers):
(WebCore::RenderLayerCompositor::ensureRootLayer):
(WebCore::RenderLayerCompositor::destroyRootLayer):
(WebCore::RenderLayerCompositor::updateScrollingNodeForScrollingRole):
* rendering/RenderLayerCompositor.h:

Source/WebKit:

* UIProcess/RemoteLayerTree/ios/ScrollingTreeFrameScrollingNodeRemoteIOS.mm:
(WebKit::ScrollingTreeFrameScrollingNodeRemoteIOS::commitStateBeforeChildren):
* UIProcess/RemoteLayerTree/ios/ScrollingTreeScrollingNodeDelegateIOS.mm:
(WebKit::ScrollingTreeScrollingNodeDelegateIOS::commitStateBeforeChildren):

We now use scrollContainerLayer consistently so remove the special cases.

LayoutTests:

Test by Frédéric Wang.

* fast/scrolling/ios/scroll-iframe-expected.html:
* fast/scrolling/ios/scroll-iframe.html:
* platform/ios-wk2/compositing/iframes/scrolling-iframe-expected.txt:
* platform/ios-wk2/compositing/tiling/tiled-drawing-async-frame-scrolling-expected.txt:
* platform/ios-wk2/scrollingcoordinator/scrolling-tree/fixed-inside-frame-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240916 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[JSC] UnlinkedMetadataTable assumes that MetadataTable is destroyed before it is...
ysuzuki@apple.com [Mon, 4 Feb 2019 07:13:00 +0000 (07:13 +0000)]
[JSC] UnlinkedMetadataTable assumes that MetadataTable is destroyed before it is destructed, but order of destruction of JS heap cells are not guaranteed
https://bugs.webkit.org/show_bug.cgi?id=194031

Reviewed by Saam Barati.

UnlinkedMetadataTable assumes that MetadataTable linked against this UnlinkedMetadataTable is already destroyed when UnlinkedMetadataTable is destroyed.
This means that UnlinkedCodeBlock is destroyed after all the linked CodeBlocks are destroyed. But this assumption is not valid since GC's finalizer
sweeps objects without considering the dependencies among swept objects. UnlinkedMetadataTable can be destroyed even before linked MetadataTable is
destroyed if UnlinkedCodeBlock is destroyed before linked CodeBlock is destroyed.

To make the above assumption valid, we make UnlinkedMetadataTable RefCounted object, and make MetadataTable hold the strong ref to UnlinkedMetadataTable.
This ensures that UnlinkedMetadataTable is destroyed after all the linked MetadataTables are destroyed.

* bytecode/MetadataTable.cpp:
(JSC::MetadataTable::MetadataTable):
(JSC::MetadataTable::~MetadataTable):
* bytecode/UnlinkedCodeBlock.cpp:
(JSC::UnlinkedCodeBlock::UnlinkedCodeBlock):
(JSC::UnlinkedCodeBlock::visitChildren):
(JSC::UnlinkedCodeBlock::estimatedSize):
(JSC::UnlinkedCodeBlock::setInstructions):
* bytecode/UnlinkedCodeBlock.h:
(JSC::UnlinkedCodeBlock::metadata):
(JSC::UnlinkedCodeBlock::metadataSizeInBytes):
* bytecode/UnlinkedMetadataTable.h:
(JSC::UnlinkedMetadataTable::create):
* bytecode/UnlinkedMetadataTableInlines.h:
(JSC::UnlinkedMetadataTable::UnlinkedMetadataTable):
* runtime/CachedTypes.cpp:
(JSC::CachedMetadataTable::decode const):
(JSC::CachedCodeBlock::metadata const):
(JSC::UnlinkedCodeBlock::UnlinkedCodeBlock):
(JSC::CachedCodeBlock<CodeBlockType>::decode const):
(JSC::CachedCodeBlock<CodeBlockType>::encode):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240915 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[curl] [WebKit] Assertion failures of missing networkStorageSession for storage/index...
Hironori.Fujii@sony.com [Mon, 4 Feb 2019 02:15:55 +0000 (02:15 +0000)]
[curl] [WebKit] Assertion failures of missing networkStorageSession for storage/indexeddb tests
https://bugs.webkit.org/show_bug.cgi?id=194141

Reviewed by Ross Kirsling.

Implemented WebKit::NetworkProcess::switchToNewTestingSession and
WebKit::NetworkProcess::ensureSession for curl port.

* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::switchToNewTestingSession):
(WebKit::NetworkProcess::ensureSession):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240914 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[Win] WebKitTestRunners is failing to create the IndexedDB directory.
Hironori.Fujii@sony.com [Mon, 4 Feb 2019 02:11:32 +0000 (02:11 +0000)]
[Win] WebKitTestRunners is failing to create the IndexedDB directory.
https://bugs.webkit.org/show_bug.cgi?id=194142

Reviewed by Ross Kirsling.

There three 'separator' variables in TestController.cpp. Renamed
it 'pathSeparator', and unified them.

* WebKitTestRunner/TestController.cpp:
(WTR::TestController::generateContextConfiguration const):
(WTR::createTestURL):
(WTR::TestController::platformAdjustContext):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240913 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoMake setNeedsLayout on the root more explicitly about triggering its side-effects
simon.fraser@apple.com [Mon, 4 Feb 2019 01:48:13 +0000 (01:48 +0000)]
Make setNeedsLayout on the root more explicitly about triggering its side-effects
https://bugs.webkit.org/show_bug.cgi?id=194198

Reviewed by Antti Koivisto.

Source/WebCore:

Calling setNeedsLayout() on the FrameView or RenderView is an odd concept; the render tree
generally manages its own dirty state.

Most callers of setNeedsLayout() on the root are really trying to trigger the side-effects
of layout, like compositing updates, which are required when view configuration state, like
headers, footers and transparency, change. These dependencies are currently implicit and
poorly defined.

Renaming "setNeedsLayout" on FrameView is a step towards being more explicit about pure
rendering updates, vs updates of downstream data strutures like compositing. It's now called
setNeedsLayoutAfterViewConfigurationChange(). In addition, expose
setNeedsCompositingConfigurationUpdate() and setNeedsCompositingGeometryUpdate() so callers
can trigger the appropriate types of compositing updates on the root layer.

In addition, FrameViewLayoutContext::setNeedsLayoutAfterViewConfigurationChange() schedules a
layout. Withtout this, some callers would dirty the RenderView's layout but rely on some
other trigger to make the layout happen.

This cleanup was prompted by noticing that FrameView::setHeaderHeight() dirtied layout
but never scheduled it, making banner insertion in MiniBrowser unreliable.

This patch also removes the aliasing of headerHeight/footerHeight between Page and
FrameView. Banners are a property of Page, so FrameView fetches the banner heights
from Page.

* page/FrameView.cpp:
(WebCore::FrameView::headerHeight const):
(WebCore::FrameView::footerHeight const):
(WebCore::FrameView::availableContentSizeChanged):
(WebCore::FrameView::setNeedsLayoutAfterViewConfigurationChange):
(WebCore::FrameView::setNeedsCompositingConfigurationUpdate):
(WebCore::FrameView::setNeedsCompositingGeometryUpdate):
(WebCore::FrameView::scheduleSelectionUpdate):
(WebCore::FrameView::setTransparent):
(WebCore::FrameView::setBaseBackgroundColor):
(WebCore::FrameView::setAutoSizeFixedMinimumHeight):
(WebCore::FrameView::enableAutoSizeMode):
(WebCore::FrameView::setHeaderHeight): Deleted.
(WebCore::FrameView::setFooterHeight): Deleted.
(WebCore::FrameView::setNeedsLayout): Deleted.
* page/FrameView.h:
* page/FrameViewLayoutContext.cpp:
(WebCore::FrameViewLayoutContext::setNeedsLayoutAfterViewConfigurationChange):
(WebCore::FrameViewLayoutContext::setNeedsLayout): Deleted.
* page/FrameViewLayoutContext.h:
* page/Page.cpp:
(WebCore::Page::setPageScaleFactor):
(WebCore::Page::setHeaderHeight):
(WebCore::Page::setFooterHeight):
(WebCore::Page::addHeaderWithHeight): Deleted.
(WebCore::Page::addFooterWithHeight): Deleted.
* page/Page.h:
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::updateBacking):
* testing/Internals.cpp:
(WebCore::Internals::resetToConsistentState):
(WebCore::Internals::setHeaderHeight):
(WebCore::Internals::setFooterHeight):

Source/WebKit:

Call the newly named functions.

* WebProcess/Plugins/PDF/PDFPlugin.mm:
(WebKit::PDFPlugin::updateScrollbars):
* WebProcess/Plugins/PluginView.cpp:
(WebKit::PluginView::didInitializePlugin):
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::setHeaderBannerHeightForTesting):
(WebKit::WebPage::setFooterBannerHeightForTesting):
* WebProcess/WebPage/mac/PageBannerMac.mm:
(WebKit::PageBanner::addToPage):
(WebKit::PageBanner::detachFromPage):
(WebKit::PageBanner::hide):

Source/WebKitLegacy/mac:

Call the newly named functions.

* WebView/WebFrame.mm:
(-[WebFrame setNeedsLayout]):
* WebView/WebHTMLView.mm:
(-[WebHTMLView setNeedsLayout:]):

Tools:

No need to set the banner heights on navigation now, since Page stores them.

* MiniBrowser/mac/WK2BrowserWindowController.m:
(-[WK2BrowserWindowController webView:didFinishNavigation:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240912 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoParse and handle Ad Click Attribution attributes in HTMLAnchorElement::handleClick()
wilander@apple.com [Mon, 4 Feb 2019 00:44:07 +0000 (00:44 +0000)]
Parse and handle Ad Click Attribution attributes in HTMLAnchorElement::handleClick()
https://bugs.webkit.org/show_bug.cgi?id=194104
<rdar://problem/47649991>

Reviewed by Chris Dumez, Daniel Bates, and Darin Adler.

Source/WebCore:

Test: http/tests/adClickAttribution/anchor-tag-attributes-validation.html

This patch adds parsing and validation of the two new Ad Click Attribution
attributes in anchor elements: adcampaignid and addestination. The data is
not yet forwarded into the loader.

* html/HTMLAnchorElement.cpp:
(WebCore::HTMLAnchorElement::parseAdClickAttribution const):
(WebCore::HTMLAnchorElement::handleClick):
    Now calls HTMLAnchorElement::parseAdClickAttribution().
* html/HTMLAnchorElement.h:
* loader/AdClickAttribution.h:
    Made WebCore::AdClickAttribution copyable since it's needed to have it be
    WTF::Optional. Also made AdClickAttribution::MaxEntropy public. Changed
    numeric types from unsigned short to uint32_t.
(WebCore::AdClickAttribution::Campaign::isValid const):
(WebCore::AdClickAttribution::Conversion::isValid const):

Tools:

* TestWebKitAPI/Tests/WebCore/AdClickAttribution.cpp:
(TestWebKitAPI::TEST):
    Changed numeric types from unsigned short to uint32_t.

LayoutTests:

This test case makes sure invalid data triggers console warnings.

* http/tests/adClickAttribution/anchor-tag-attributes-validation-expected.txt: Added.
* http/tests/adClickAttribution/anchor-tag-attributes-validation.html: Added.
* platform/ios-wk2/http/tests/adClickAttribution/anchor-tag-attributes-validation-expected.txt: Added.
    Console line numbers are not emitted when running iOS tests so this -expected.txt file doesn't have them.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240911 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoValidate navigation policy decisions to avoid crashes in continueLoadAfterNavigationP...
rniwa@webkit.org [Sun, 3 Feb 2019 22:48:22 +0000 (22:48 +0000)]
Validate navigation policy decisions to avoid crashes in continueLoadAfterNavigationPolicy
https://bugs.webkit.org/show_bug.cgi?id=194189

Reviewed by Geoffrey Garen.

Source/WebCore:

Introduced PolicyCheckIdentifier to pair each navigation policy check request with a decision,
and deployed it in PolicyChecker. The identifier is passed from WebContent process to UI process
in WebKit2, and passed it back with the policy decision.

Because PolicyCheckIdentifier embeds the process identifier from which a navigation policy is checked,
we would be able to detect when UI process had sent the decision to a wrong WebContent process.

This patch also adds release assertions to make sure history().provisionalItem() is set whenever
we're requesting a navigation policy check.

These code changes should either:
1. Fix crashes in FrameLoader::continueLoadAfterNavigationPolicy where isBackForwardLoadType would
   return true yet history().provisionalItem() is null.
2. Detect a bug that UI process can send a navigation policy decision to a wrong WebContent process.
3. Rule out the possibility that (2) exists.

* loader/DocumentLoader.cpp:
(WebCore::DocumentLoader::willSendRequest):
(WebCore::DocumentLoader::responseReceived):
* loader/EmptyClients.cpp:
(WebCore::EmptyFrameLoaderClient::dispatchDecidePolicyForNewWindowAction):
(WebCore::EmptyFrameLoaderClient::dispatchDecidePolicyForNavigationAction):
* loader/EmptyFrameLoaderClient.h:
* loader/FrameLoader.cpp:
(WebCore::FrameLoader::checkContentPolicy):
(WebCore::FrameLoader::loadURL):
(WebCore::FrameLoader::load):
(WebCore::FrameLoader::loadWithDocumentLoader):
(WebCore::FrameLoader::loadPostRequest):
* loader/FrameLoader.h:
* loader/FrameLoaderClient.h:
* loader/FrameLoaderTypes.h:
(WebCore::PolicyCheckIdentifier): Added.
(WebCore::PolicyCheckIdentifier::operator== const): Added.
(WebCore::PolicyCheckIdentifier::PolicyCheckIdentifier): Added.
(WebCore::PolicyCheckIdentifier::encode const): Added.
(WebCore::PolicyCheckIdentifier::decode): Added.
* loader/PolicyChecker.cpp:
(WebCore::PolicyCheckIdentifier::generate):
(WebCore::PolicyCheckIdentifier::isValidFor): Returns true if the identifer matches. Also release asserts
that the process ID is same, and that m_check is always not zero (meaning it's a generated value).
The failure of these release assertions would indicate that there is a bug in UI process, which results in
a policy decision response being sent to a wrong Web process.
(WebCore::PolicyChecker::checkNavigationPolicy): Exit early if isValidFor fails.
(WebCore::PolicyChecker::checkNewWindowPolicy):

Source/WebKit:

Pass the policy check identifier around functions and store it in PolicyDecisionSender
so that we can send it back to WebCore with the navigation policy decision.

We also store it in WebFrame in the case the policy decision had to be invalidated
before the decision was received (via WebFrame::invalidatePolicyListener).

* Scripts/webkit/messages.py:
* UIProcess/ProvisionalPageProxy.cpp:
(WebKit::ProvisionalPageProxy::decidePolicyForNavigationActionAsync):
(WebKit::ProvisionalPageProxy::decidePolicyForResponse):
* UIProcess/ProvisionalPageProxy.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::PolicyDecisionSender): Added PolicyCheckIdentifier as a member.
(WebKit::WebPageProxy::PolicyDecisionSender::create):
(WebKit::WebPageProxy::PolicyDecisionSender::send):
(WebKit::WebPageProxy::PolicyDecisionSender::PolicyDecisionSender):
(WebKit::WebPageProxy::receivedNavigationPolicyDecision):
(WebKit::WebPageProxy::decidePolicyForNavigationActionAsync):
(WebKit::WebPageProxy::decidePolicyForNavigationActionAsyncShared):
(WebKit::WebPageProxy::decidePolicyForNavigationAction):
(WebKit::WebPageProxy::decidePolicyForNavigationActionSync):
(WebKit::WebPageProxy::decidePolicyForNewWindowAction):
(WebKit::WebPageProxy::decidePolicyForResponse):
(WebKit::WebPageProxy::decidePolicyForResponseShared):
* UIProcess/WebPageProxy.h:
* UIProcess/WebPageProxy.messages.in:
* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::dispatchDecidePolicyForResponse):
(WebKit::WebFrameLoaderClient::dispatchDecidePolicyForNewWindowAction):
(WebKit::WebFrameLoaderClient::dispatchDecidePolicyForNavigationAction):
* WebProcess/WebCoreSupport/WebFrameLoaderClient.h:
* WebProcess/WebPage/WebFrame.cpp:
(WebKit::WebFrame::setUpPolicyListener):
(WebKit::WebFrame::invalidatePolicyListener):
(WebKit::WebFrame::didReceivePolicyDecision):
* WebProcess/WebPage/WebFrame.h:
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::didReceivePolicyDecision):
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

Source/WebKitLegacy/mac:

Pass the policy check identifier around functions and store it in WebFramePolicyListener
so that we can send it back to WebCore with the navigation policy decision.

* WebCoreSupport/WebFrameLoaderClient.h:
* WebCoreSupport/WebFrameLoaderClient.mm:
(WebFrameLoaderClient::dispatchDecidePolicyForResponse):
(WebFrameLoaderClient::dispatchDecidePolicyForNewWindowAction):
(WebFrameLoaderClient::dispatchDecidePolicyForNavigationAction):
(WebFrameLoaderClient::dispatchWillSubmitForm):
(WebFrameLoaderClient::setUpPolicyListener):
(-[WebFramePolicyListener initWithFrame:identifier:policyFunction:defaultPolicy:]):
(-[WebFramePolicyListener initWithFrame:identifier:policyFunction:defaultPolicy:appLinkURL:]):
(-[WebFramePolicyListener invalidate]):
(-[WebFramePolicyListener dealloc]):
(-[WebFramePolicyListener receivedPolicyDecision:]):
(-[WebFramePolicyListener initWithFrame:policyFunction:defaultPolicy:]): Deleted.
(-[WebFramePolicyListener initWithFrame:policyFunction:defaultPolicy:appLinkURL:]): Deleted.

Source/WebKitLegacy/win:

Pass the policy check identifier around functions and store it in WebFramePolicyListener
so that we can send it back to WebCore with the navigation policy decision.

* WebCoreSupport/WebFrameLoaderClient.cpp:
(WebFrameLoaderClient::dispatchDecidePolicyForResponse):
(WebFrameLoaderClient::dispatchDecidePolicyForNewWindowAction):
(WebFrameLoaderClient::dispatchDecidePolicyForNavigationAction):
(WebFrameLoaderClient::dispatchWillSubmitForm):
(WebFrameLoaderClient::setUpPolicyListener):
* WebCoreSupport/WebFrameLoaderClient.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240909 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoDon't include ScrollCoordinator.h from Element.h
antti@apple.com [Sun, 3 Feb 2019 21:02:54 +0000 (21:02 +0000)]
Don't include ScrollCoordinator.h from Element.h
https://bugs.webkit.org/show_bug.cgi?id=194206

Reviewed by Daniel Bates.

* dom/Element.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240906 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUnable to move selection into editable roots with 0 height
wenson_hsieh@apple.com [Sun, 3 Feb 2019 20:52:58 +0000 (20:52 +0000)]
Unable to move selection into editable roots with 0 height
https://bugs.webkit.org/show_bug.cgi?id=194143
<rdar://problem/47767284>

Reviewed by Ryosuke Niwa.

Source/WebCore:

Currently, positions inside editable elements of height 0 are not considered to be candidates when
canonicalizing a position to its visible counterpart. This prevents us from moving the selection into these
editable roots at all. To fix this, we relax this constraint by allowing positions anchored by root editable
elements to be candidates.

Test: editing/selection/insert-text-in-empty-content-editable.html

* dom/Position.cpp:
(WebCore::Position::isCandidate const):

LayoutTests:

Add a new layout test that executes editing commands in a contenteditable element of height 0, and adjust an
existing layout test that copies and pastes an image element to wait until the image has loaded.

* editing/pasteboard/styled-element-markup.html:
* editing/selection/insert-text-in-empty-content-editable-expected.txt: Added.
* editing/selection/insert-text-in-empty-content-editable.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240905 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoTidyup of Pagination and FrameView m_mediaType initialization
simon.fraser@apple.com [Sun, 3 Feb 2019 19:46:31 +0000 (19:46 +0000)]
Tidyup of Pagination and FrameView m_mediaType initialization
https://bugs.webkit.org/show_bug.cgi?id=194203

Reviewed by Darin Adler.

Fix post-commit feedback on lines around code changed in r240901.

* page/FrameView.cpp:
(WebCore::FrameView::FrameView):
* page/FrameView.h:
* rendering/Pagination.h:
(WebCore::Pagination::operator!= const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240904 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUnreviewed, rolling out r240896.
commit-queue@webkit.org [Sun, 3 Feb 2019 13:53:54 +0000 (13:53 +0000)]
Unreviewed, rolling out r240896.
https://bugs.webkit.org/show_bug.cgi?id=194202

"Fixes leaks, but is probably not the correct fix." (Requested
by ddkilzer on #webkit).

Reverted changeset:

"Leak of WTF::StringImpl under SymbolImpl::createNullSymbol()
(48 bytes) in com.apple.WebKit.WebContent running layout
tests"
https://bugs.webkit.org/show_bug.cgi?id=193291
https://trac.webkit.org/changeset/240896

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240903 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoTurn on Smart Paste
megan_gardner@apple.com [Sun, 3 Feb 2019 09:05:21 +0000 (09:05 +0000)]
Turn on Smart Paste
https://bugs.webkit.org/show_bug.cgi?id=193786

Reviewed by Ryosuke Niwa.

Source/WebCore:

Turned on a modified tests:
LayoutTests/editing/pasteboard/smart-paste-001.html
LayoutTests/editing/pasteboard/smart-paste-002.html
LayoutTests/editing/pasteboard/smart-paste-003.html
LayoutTests/editing/pasteboard/smart-paste-004.html
LayoutTests/editing/pasteboard/smart-paste-005.html
LayoutTests/editing/pasteboard/smart-paste-006.html
LayoutTests/editing/pasteboard/smart-paste-007.html
LayoutTests/editing/pasteboard/smart-paste-008.html

* platform/ios/PasteboardIOS.mm:
(WebCore::Pasteboard::canSmartReplace):
Turn on smart replace.

Source/WebKit:

* Platform/spi/ios/UIKitSPI.h:
* UIProcess/ios/TextCheckerIOS.mm:
(WebKit::TextChecker::isSmartInsertDeleteEnabled):
Use the flag from UIKit to determine if smart copy paste should be
turned on.

LayoutTests:

Rebaselined many tests and/or added iOS specific expectations to
correctly reflect smart copy paste behavior.

* editing/editing.js:
(runEditingTestWithCallbackLogging):
(runDumpAsTextEditingTest):
Added ability to handle async functions.
* editing/pasteboard/smart-paste-001.html:
* editing/pasteboard/smart-paste-002.html:
* editing/pasteboard/smart-paste-003.html:
* editing/pasteboard/smart-paste-004.html:
* editing/pasteboard/smart-paste-005.html:
* editing/pasteboard/smart-paste-006.html:
* editing/pasteboard/smart-paste-007.html:
* editing/pasteboard/smart-paste-008.html:
* platform/ios-wk2/editing/pasteboard/smart-paste-007-expected.txt:
* platform/ios-wk2/editing/pasteboard/smart-paste-008-expected.txt:
* platform/ios/TestExpectations:
* platform/ios/editing/pasteboard/smart-paste-001-expected.txt: Added.
* platform/ios/editing/pasteboard/smart-paste-002-expected.txt: Added.
* platform/ios/editing/pasteboard/smart-paste-003-expected.txt: Added.
* platform/ios/editing/pasteboard/smart-paste-004-expected.txt: Added.
* platform/ios/editing/pasteboard/smart-paste-005-expected.txt: Added.
* platform/ios/editing/pasteboard/smart-paste-006-expected.txt: Added.
* resources/ui-helper.js:
(window.UIHelper.doubleClickAt):
(window.UIHelper.async.selectWordByDoubleTapOrClick):
Added helper functions to allow for selecting a word by double tap or click.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240902 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoTidy up data memebers of FrameView and related classes to shrink class sizes
simon.fraser@apple.com [Sun, 3 Feb 2019 00:16:44 +0000 (00:16 +0000)]
Tidy up data memebers of FrameView and related classes to shrink class sizes
https://bugs.webkit.org/show_bug.cgi?id=194197

Reviewed by Zalan Bujtas.

Source/WebCore:

Make various enums 8-bit. Re-order data members of FrameView, FrameViewLayoutContext, Widget
and ScrollView to optimize padding.

Use more data member initializers.

This shrinks FrameView from 1168 (144 padding bytes) to 1096 (96 padding bytes).

* dom/Element.h:
* page/FrameView.cpp:
(WebCore::FrameView::FrameView):
* page/FrameView.h:
* page/FrameViewLayoutContext.h:
* platform/ScrollTypes.h:
* platform/ScrollView.h: Data members should come after member functions.
* platform/Widget.h:
* rendering/Pagination.h:
(WebCore::Pagination::Pagination): Deleted.

Source/WebKit:

* UIProcess/PageClient.h: Declare enum size.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240901 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoLeak of NSArray (4.25 Kbytes) in com.apple.WebKit.WebContent running WebKit layout...
ddkilzer@apple.com [Sat, 2 Feb 2019 20:20:36 +0000 (20:20 +0000)]
Leak of NSArray (4.25 Kbytes) in com.apple.WebKit.WebContent running WebKit layout tests on iOS Simulator
<https://webkit.org/b/194178>
<rdar://problem/47753473>

Reviewed by Alexey Proskuryakov.

* WebKitTestRunner/InjectedBundle/cocoa/ActivateFontsCocoa.mm:
(WTR::uninstallFakeHelvetica): Pass nullptr into
CTFontManagerUnregisterFontsForURLs() to ignore any errors.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240900 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoRename "scrollingLayer" in RenderLayerBacking to "scrollContainerLayer" for clarity
simon.fraser@apple.com [Sat, 2 Feb 2019 20:05:22 +0000 (20:05 +0000)]
Rename "scrollingLayer" in RenderLayerBacking to "scrollContainerLayer" for clarity
https://bugs.webkit.org/show_bug.cgi?id=194194

Rubber-stampted by Antti Koivisto.

Our desired terminology is "scrollContainerLayer" for the outer, clipping layer
for scrolling, and "scrolledContentsLayer" for the contents that move when you scroll.

* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::updateDebugIndicators):
(WebCore::RenderLayerBacking::destroyGraphicsLayers):
(WebCore::RenderLayerBacking::updateGeometry):
(WebCore::RenderLayerBacking::setLocationOfScrolledContents):
(WebCore::RenderLayerBacking::updateAfterDescendants):
(WebCore::RenderLayerBacking::updateInternalHierarchy):
(WebCore::RenderLayerBacking::updateDrawsContent):
(WebCore::RenderLayerBacking::updateScrollingLayers):
* rendering/RenderLayerBacking.h:
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::didChangePlatformLayerForLayer):
(WebCore::RenderLayerCompositor::updateScrollingNodeForScrollingRole):
(WebCore::LegacyWebKitScrollingLayerCoordinator::registerAllViewportConstrainedLayers):
(WebCore::LegacyWebKitScrollingLayerCoordinator::updateScrollingLayer):
(WebCore::LegacyWebKitScrollingLayerCoordinator::unregisterAllScrollingLayers):
(WebCore::LegacyWebKitScrollingLayerCoordinator::removeScrollingLayer):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240899 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoSource/WebCore:
justin_fan@apple.com [Sat, 2 Feb 2019 19:53:55 +0000 (19:53 +0000)]
Source/WebCore:
[Web GPU] Fix GPURenderPassEncoder::setVertexBuffers and allow overlapping indices with GPUBindGroups
https://bugs.webkit.org/show_bug.cgi?id=194125

Reviewed by Myles C. Maxfield.

GPURenderPassEncoder::setVertexBuffers is now actually written to set all buffers provided. In addition,
shift vertex input buffer indices so that any resource bindings can bind vertex buffer resources to the same indices.

Existing tests cover setVertexBuffers. Updated buffer-resource-triangles to assign bind groups and vertex buffers to the same index.

* Modules/webgpu/WHLSL/Metal/WHLSLVertexBufferIndexCalculator.cpp: Added.
(WebCore::WHLSL::Metal::calculateVertexBufferIndex): Simple shifting function for vertex input buffer indices.
* Modules/webgpu/WHLSL/Metal/WHLSLVertexBufferIndexCalculator.h: Added.
* Modules/webgpu/WebGPUProgrammablePassEncoder.cpp:
(WebCore::WebGPUProgrammablePassEncoder::setBindGroup const): Limit maximum bind group indices to 0 to 3.
* Modules/webgpu/WebGPURenderPassEncoder.cpp:
(WebCore::WebGPURenderPassEncoder::setVertexBuffers): Limit vertex input indices to 0 to 15.
* Modules/webgpu/WebGPURenderPassEncoder.h: Move IDL/bindings bug note to IDL file.
* Modules/webgpu/WebGPURenderPassEncoder.idl: Ditto.
* platform/graphics/gpu/GPULimits.h: Added. Home for Web GPU limits constants shared between files.
* platform/graphics/gpu/GPURenderPassEncoder.h: Change IDL/bindings bug workaround to unsigned long long to prevent narrowing compared to spec.
* platform/graphics/gpu/cocoa/GPURenderPassEncoderMetal.mm:
(WebCore::GPURenderPassEncoder::setVertexBuffers): Now properly calls Metal's setVertexBuffers.
* platform/graphics/gpu/cocoa/GPURenderPipelineMetal.mm:
(WebCore::setInputStateForPipelineDescriptor): Fix validation checks for vertex attribute numbers and vertex buffer indices.

Add symbols to project:
* Sources.txt:
* WebCore.xcodeproj/project.pbxproj:

LayoutTests:
[WebGPU] Fix GPURenderPassEncoder::setVertexBuffers and allow overlapping indices with GPUBindGroups
https://bugs.webkit.org/show_bug.cgi?id=194125

Reviewed by Myles C. Maxfield.

Updated buffer-resource-triangles to assign bind groups and vertex buffers to the same index.

* webgpu/buffer-resource-triangles.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240898 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoAsync overflow scroll is jumpy on macOS if the main thread is busy
simon.fraser@apple.com [Sat, 2 Feb 2019 19:09:18 +0000 (19:09 +0000)]
Async overflow scroll is jumpy on macOS if the main thread is busy
https://bugs.webkit.org/show_bug.cgi?id=194184
<rdar://problem/47758655>

Reviewed by Antti Koivisto.

This change extends to macOS some existing overflow-scroll functionality for iOS.
When an async scroll is in process in the scroll thread (or UI process), we periodically
message back to the web process main thread with scroll position updates. These
can trigger post-scroll compositing updates, but we need to ensure that this update
doesn't clobber the scroll position of the native layers, which would trigger
stutters.

To do this we have the notion of a scroll position "sync" (ScrollingLayerPositionAction::Sync) which
pokes the new value into the GraphicsLayer (hence making visible rect computations work), but doesn't
propagate it to the platform layer. This patch wires up syncs for macOS during async overflow scrolling,
coming out of AsyncScrollingCoordinator::updateScrollPositionAfterAsyncScroll().

In RenderLayerBacking, m_scrollingContentsLayer is renamed to m_scrolledContentsLayer, and I added
updateScrollOffset() and setLocationOfScrolledContents() to handle the set vs. sync, and to keep
the iOS vs macOS differences in one function. This allows for more code sharing in RenderLayerBacking::updateGeometry().

There's a confusing bit in the m_childClippingMaskLayer code (trac.webkit.org/178029) where the setOffsetFromRenderer()
just looks wrong; it should match m_scrollingLayer. This code is never hit for Cocoa, which never has m_childClippingMaskLayer.

* page/scrolling/mac/ScrollingTreeOverflowScrollingNodeMac.mm:
(WebCore::ScrollingTreeOverflowScrollingNodeMac::setScrollPosition): Logging
(WebCore::ScrollingTreeOverflowScrollingNodeMac::setScrollLayerPosition): Logging
* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::RenderLayer):
(WebCore::RenderLayer::scrollTo):
* rendering/RenderLayer.h: Rename m_requiresScrollBoundsOriginUpdate to m_requiresScrollPositionReconciliation
and make it available on all platforms. Just reorder m_adjustForIOSCaretWhenScrolling to reduce #ifdef nesting confusion.
* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::updateDebugIndicators):
(WebCore::RenderLayerBacking::destroyGraphicsLayers):
(WebCore::RenderLayerBacking::updateGeometry):
(WebCore::RenderLayerBacking::setLocationOfScrolledContents):
(WebCore::RenderLayerBacking::updateScrollOffset):
(WebCore::RenderLayerBacking::updateDrawsContent):
(WebCore::RenderLayerBacking::updateScrollingLayers):
(WebCore::RenderLayerBacking::paintingPhaseForPrimaryLayer const):
(WebCore::RenderLayerBacking::parentForSublayers const):
(WebCore::RenderLayerBacking::setContentsNeedDisplay):
(WebCore::RenderLayerBacking::setContentsNeedDisplayInRect):
(WebCore::RenderLayerBacking::paintContents):
(WebCore::RenderLayerBacking::backingStoreMemoryEstimate const):
* rendering/RenderLayerBacking.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240897 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoLeak of WTF::StringImpl under SymbolImpl::createNullSymbol() (48 bytes) in com.apple...
ddkilzer@apple.com [Sat, 2 Feb 2019 18:35:42 +0000 (18:35 +0000)]
Leak of WTF::StringImpl under SymbolImpl::createNullSymbol() (48 bytes) in com.apple.WebKit.WebContent running layout tests
<https://webkit.org/b/193291>
<rdar://problem/46655953>

Reviewed by Keith Miller.

* wtf/text/SymbolImpl.h:
(WTF::SymbolImpl::~SymbolImpl): Fix the leak by implementing the
class destructor that calls StringImpl::deref() on `m_owner`.
Two of the three constructors leak the StringImpl when setting
`m_owner`, so we need to balance that by manually calling
deref().

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240896 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[LFC] Initialize ICB's style with fixed width/height.
zalan@apple.com [Sat, 2 Feb 2019 16:32:05 +0000 (16:32 +0000)]
[LFC] Initialize ICB's style with fixed width/height.
https://bugs.webkit.org/show_bug.cgi?id=194188

Reviewed by Antti Koivisto.

Source/WebCore:

Let's set ICB's logical width/height to Fixed so that it's available when we try to resolve a box's height
in FormattingContext::Geometry::computedHeightValue() by using the containing block's height (ICB in this case).

* layout/LayoutState.cpp:
(WebCore::Layout::LayoutState::LayoutState):
* layout/LayoutState.h:
* layout/layouttree/LayoutTreeBuilder.cpp:
(WebCore::Layout::TreeBuilder::createLayoutTree):

Tools:

* LayoutReloaded/misc/LFC-passing-tests.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240895 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[LFC] Add missing case to out-of-flow non-replaced horizontal used margin value compu...
zalan@apple.com [Sat, 2 Feb 2019 16:26:42 +0000 (16:26 +0000)]
[LFC] Add missing case to out-of-flow non-replaced horizontal used margin value computation
https://bugs.webkit.org/show_bug.cgi?id=194185

Reviewed by Antti Koivisto.

Source/WebCore:

If none of horizontal values (left, right, width, margin-left/right) are auto, then usedHorizontalMarginValues = computedHorizontalMarginValues.

Test: fast/block/block-only/absolute-position-left-right-margin.html

* layout/FormattingContextGeometry.cpp:
(WebCore::Layout::FormattingContext::Geometry::outOfFlowNonReplacedHorizontalGeometry):

Tools:

* LayoutReloaded/misc/LFC-passing-tests.txt:

LayoutTests:

* fast/block/block-only/absolute-position-left-right-margin-expected.html: Added.
* fast/block/block-only/absolute-position-left-right-margin.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240894 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[JSC] Decouple JIT related data from CodeBlock
ysuzuki@apple.com [Sat, 2 Feb 2019 07:15:47 +0000 (07:15 +0000)]
[JSC] Decouple JIT related data from CodeBlock
https://bugs.webkit.org/show_bug.cgi?id=194187

Patch by Yusuke Suzuki <ysuzuki@apple.com> on 2019-02-01
Reviewed by Saam Barati.

CodeBlock holds bunch of data which is only used after JIT starts compiling it.
We have three types of data in CodeBlock.

1. The data which is always used. CodeBlock needs to hold it.
2. The data which is touched even in LLInt, but it is only meaningful in JIT tiers. The example is profiling.
3. The data which is used after the JIT compiler starts running for the given CodeBlock.

This patch decouples (3) from CodeBlock as CodeBlock::JITData. Even if we have bunch of CodeBlocks, only small
number of them gets JIT compilation. Always allocating (3) data enlarges the size of CodeBlock, leading to the
memory waste. Potentially we can decouple (2) in another data structure, but we first do (3) since (3) is beneficial
in both non-JIT and *JIT* modes.

JITData is created only when JIT compiler wants to use it. So it can be concurrently created and used, so it is guarded
by the lock of CodeBlock.

The size of CodeBlock is reduced from 512 to 352.

This patch improves memory footprint and gets 1.1% improvement in RAMification.

    Footprint geomean: 36696503 (34.997 MB)
    Peak Footprint geomean: 38595988 (36.808 MB)
    Score: 37634263 (35.891 MB)

    Footprint geomean: 37172768 (35.451 MB)
    Peak Footprint geomean: 38978288 (37.173 MB)
    Score: 38064824 (36.301 MB)

* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::~CodeBlock):
(JSC::CodeBlock::propagateTransitions):
(JSC::CodeBlock::ensureJITDataSlow):
(JSC::CodeBlock::finalizeBaselineJITInlineCaches):
(JSC::CodeBlock::getICStatusMap):
(JSC::CodeBlock::addStubInfo):
(JSC::CodeBlock::addJITAddIC):
(JSC::CodeBlock::addJITMulIC):
(JSC::CodeBlock::addJITSubIC):
(JSC::CodeBlock::addJITNegIC):
(JSC::CodeBlock::findStubInfo):
(JSC::CodeBlock::addByValInfo):
(JSC::CodeBlock::addCallLinkInfo):
(JSC::CodeBlock::getCallLinkInfoForBytecodeIndex):
(JSC::CodeBlock::addRareCaseProfile):
(JSC::CodeBlock::rareCaseProfileForBytecodeOffset):
(JSC::CodeBlock::rareCaseProfileCountForBytecodeOffset):
(JSC::CodeBlock::resetJITData):
(JSC::CodeBlock::stronglyVisitStrongReferences):
(JSC::CodeBlock::shrinkToFit):
(JSC::CodeBlock::linkIncomingCall):
(JSC::CodeBlock::linkIncomingPolymorphicCall):
(JSC::CodeBlock::unlinkIncomingCalls):
(JSC::CodeBlock::updateAllPredictionsAndCountLiveness):
(JSC::CodeBlock::dumpValueProfiles):
(JSC::CodeBlock::setPCToCodeOriginMap):
(JSC::CodeBlock::findPC):
(JSC::CodeBlock::dumpMathICStats):
* bytecode/CodeBlock.h:
(JSC::CodeBlock::ensureJITData):
(JSC::CodeBlock::setJITCodeMap):
(JSC::CodeBlock::jitCodeMap):
(JSC::CodeBlock::likelyToTakeSlowCase):
(JSC::CodeBlock::couldTakeSlowCase):
(JSC::CodeBlock::lazyOperandValueProfiles):
(JSC::CodeBlock::stubInfoBegin): Deleted.
(JSC::CodeBlock::stubInfoEnd): Deleted.
(JSC::CodeBlock::callLinkInfosBegin): Deleted.
(JSC::CodeBlock::callLinkInfosEnd): Deleted.
(JSC::CodeBlock::jitCodeMap const): Deleted.
(JSC::CodeBlock::numberOfRareCaseProfiles): Deleted.
* bytecode/MethodOfGettingAValueProfile.cpp:
(JSC::MethodOfGettingAValueProfile::emitReportValue const):
(JSC::MethodOfGettingAValueProfile::reportValue):
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::InlineStackEntry::InlineStackEntry):
* jit/JIT.h:
* jit/JITOperations.cpp:
(JSC::tryGetByValOptimize):
* jit/JITPropertyAccess.cpp:
(JSC::JIT::privateCompileGetByVal):
(JSC::JIT::privateCompilePutByVal):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240893 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoConvert additional String::format clients to alternative approaches
darin@apple.com [Sat, 2 Feb 2019 04:05:55 +0000 (04:05 +0000)]
Convert additional String::format clients to alternative approaches
https://bugs.webkit.org/show_bug.cgi?id=192746

Reviewed by Alexey Proskuryakov.

Source/JavaScriptCore:

* inspector/agents/InspectorConsoleAgent.cpp:
(Inspector::InspectorConsoleAgent::stopTiming): Use makeString
and FormattedNumber::fixedWidth.

Source/WebCore:

This round of conversions covers less-trivial cases such as floating
point numerals and hexadecimal. Not yet taking on pointer serialization
("%p") or padding with spaces of zero digits, so call sites using those
have been left untouched.

In some cases these new idioms are a bit clumsy, and we could follow up
with additional convenience functions to make them more elegant.

* Modules/indexeddb/IDBKeyData.cpp:
(WebCore::IDBKeyData::loggingString const): Use more ASCIILiteral and
more appendLiteral for efficiency. Use upperNibbleToLowercaseASCIIHexDigit,
lowerNibbleToLowercaseASCIIHexDigit, and also makeString and FormattedNumber.

* css/MediaQueryEvaluator.cpp:
(WebCore::aspectRatioValueAsString): Use makeString and FormattedNumber.
Doing it this way makes it a little clearer that we have an unpleasant
use of fixed 6-digit precision here.

* html/FTPDirectoryDocument.cpp:
(WebCore::processFilesizeString): Use makeString and FormattedNumber.
* html/track/VTTCue.cpp:
(WebCore::VTTCueBox::applyCSSProperties): Ditto.
* page/CaptionUserPreferencesMediaAF.cpp:
(WebCore::CaptionUserPreferencesMediaAF::windowRoundedCornerRadiusCSS const): Ditto.
* page/History.cpp:
(WebCore::History::stateObjectAdded): Ditto.
* page/cocoa/ResourceUsageOverlayCocoa.mm:
(WebCore::formatByteNumber): Ditto.
(WebCore::gcTimerString): Use String::number.
(WebCore::ResourceUsageOverlay::platformDraw): Use makeString and FormattedNumber.

* page/scrolling/AxisScrollSnapOffsets.cpp:
(WebCore::snapOffsetsToString): Removed some unnecessary copying in the for loop,
use appendLiteral, and use appendFixedWidthNumber.
(WebCore::snapOffsetRangesToString): Ditto.
(WebCore::snapPortOrAreaToString): Use makeString and FormattedNumber.

* platform/animation/TimingFunction.cpp:
(WebCore::TimingFunction::cssText const): Use makeString.

* platform/cocoa/KeyEventCocoa.mm:
(WebCore::keyIdentifierForCharCode): Use makeString and the ASCIIHexDigit
functions.
* platform/graphics/Color.cpp:
(WebCore::Color::nameForRenderTreeAsText const): Ditto.

* platform/mock/MockRealtimeVideoSource.cpp:
(WebCore::MockRealtimeVideoSource::drawText): Use makeString and FormattedNumber.

* platform/text/PlatformLocale.cpp:
(WebCore::DateTimeStringBuilder::visitField): Use String::numberToStringFixedWidth.

Source/WebCore/PAL:

* pal/FileSizeFormatter.cpp:
(fileSizeDescription): Use makeString and FormattedNumber.

Source/WebKit:

* UIProcess/Cocoa/ViewGestureController.cpp:
(WebKit::ViewGestureController::SnapshotRemovalTracker::startWatchdog):
Use makeString and FormattedNumber.

* UIProcess/WebAuthentication/Cocoa/LocalAuthenticator.mm:
(WebKit::LocalAuthenticator::getAssertion): Added a comment about an
incorrect format specifier and left this code as is for now.

Source/WTF:

* wtf/JSONValues.cpp:
(WTF::appendDoubleQuotedStringEscapedCharacter): Renamed from
escapeChar and reordered arguments to make sense as an append function.
(WTF::appendDoubleQuotedString): Renamed from doubleQuoteString,
reordered arguments to make sense as an append function, take a
StringView instead of a String, used early exit to make the code
a bit easier to read. Use the ASCIIHexDigit functions to construct
a hex number a nibble at a time rather than using String::format.
(WTF::JSONImpl::Value::writeJSON const): Update for name change.
(WTF::JSONImpl::ObjectBase::writeJSON const): Ditto.

Tools:

* WebKitTestRunner/TestController.cpp:
(WTR::originUserVisibleName): Use makeString and reduce the use of std::string
as an intermediate in code that ultimately constructs a WTF::String.
(WTR::userMediaOriginHash): Use makeString.
(WTR::TestController::didNavigateWithNavigationData): More of the same.
(WTR::TestController::didPerformClientRedirect): Ditto.
(WTR::TestController::didPerformServerRedirect): Ditto.
(WTR::TestController::didUpdateHistoryTitle): Ditto.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240892 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoRemove the unused layerForScrolling()
simon.fraser@apple.com [Sat, 2 Feb 2019 02:30:53 +0000 (02:30 +0000)]
Remove the unused layerForScrolling()
https://bugs.webkit.org/show_bug.cgi?id=194180

Reviewed by Zalan Bujtas.

Remove ScrollableArea::layerForScrolling() and derivations. This was unused.

* page/FrameView.cpp:
(WebCore::FrameView::layerForScrolling const): Deleted.
* page/FrameView.h:
* page/scrolling/ScrollingCoordinator.cpp:
(WebCore::ScrollingCoordinator::scrollLayerForScrollableArea): Deleted.
* page/scrolling/ScrollingCoordinator.h:
* platform/ScrollableArea.h:
(WebCore::ScrollableArea::layerForScrolling const): Deleted.
* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::calculateClipRects const):
* rendering/RenderLayer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240891 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[ews-build] Add unit test to verify builder keys
aakash_jain@apple.com [Sat, 2 Feb 2019 02:19:43 +0000 (02:19 +0000)]
[ews-build] Add unit test to verify builder keys
https://bugs.webkit.org/show_bug.cgi?id=194152

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/loadConfig_unittest.py:
(ConfigDotJSONTest.test_builder_keys): Added unit-test.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240890 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoMove soft-linking of TelephonyUtilities.framework out of TUCallSPI.h
ddkilzer@apple.com [Sat, 2 Feb 2019 00:45:51 +0000 (00:45 +0000)]
Move soft-linking of TelephonyUtilities.framework out of TUCallSPI.h
<https://webkit.org/b/193866>

Reviewed by Alex Christensen.

Source/WebCore/PAL:

* PAL.xcodeproj/project.pbxproj:
* pal/PlatformMac.cmake:
- Update for rename of TUCallSPI.h to TelephonyUtilitiesSPI.h.

* pal/spi/mac/TelephonyUtilitiesSPI.h: Rename from Source/WebCore/PAL/pal/spi/mac/TUCallSPI.h.
- Remove soft-linking code from header.

Source/WebKit:

* Platform/mac/MenuUtilities.mm:
- Update for rename of TUCallSPI.h to TelephonyUtilitiesSPI.h.
- Add soft-linking code formerly in TUCallSPI.h.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240889 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[JSC] Remove some of IsoSubspaces for JSFunction subclasses
ysuzuki@apple.com [Sat, 2 Feb 2019 00:24:47 +0000 (00:24 +0000)]
[JSC] Remove some of IsoSubspaces for JSFunction subclasses
https://bugs.webkit.org/show_bug.cgi?id=194177

Reviewed by Saam Barati.

JSGeneratorFunction, JSAsyncFunction, and JSAsyncGeneratorFunction do not add any fields / classInfo methods.
We can share the IsoSubspace for JSFunction.

* runtime/JSAsyncFunction.h:
* runtime/JSAsyncGeneratorFunction.h:
* runtime/JSGeneratorFunction.h:
* runtime/VM.cpp:
(JSC::VM::VM):
* runtime/VM.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240888 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[LFC] Expand tests coverage (12 new tests -> 787)
zalan@apple.com [Fri, 1 Feb 2019 23:59:56 +0000 (23:59 +0000)]
[LFC] Expand tests coverage (12 new tests -> 787)

Unreviewed test gardening.

* LayoutReloaded/misc/LFC-passing-tests.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240887 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[LFC] Fix statically positioned replaced out-of-flow horizontal geometry
zalan@apple.com [Fri, 1 Feb 2019 23:54:22 +0000 (23:54 +0000)]
[LFC] Fix statically positioned replaced out-of-flow horizontal geometry
https://bugs.webkit.org/show_bug.cgi?id=194163

Reviewed by Simon Fraser.

Source/WebCore:

Fix a typo.

* layout/FormattingContextGeometry.cpp:
(WebCore::Layout::FormattingContext::Geometry::outOfFlowReplacedHorizontalGeometry):

Tools:

775

* LayoutReloaded/misc/LFC-passing-tests.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240886 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoMove XPCService entry points from mac directory to new Cocoa directory
achristensen@apple.com [Fri, 1 Feb 2019 23:32:42 +0000 (23:32 +0000)]
Move XPCService entry points from mac directory to new Cocoa directory
https://bugs.webkit.org/show_bug.cgi?id=194129

Reviewed by Chris Dumez.

* Configurations/NetworkService.xcconfig:
* Configurations/PluginService.64.xcconfig:
* Configurations/WebContentService.xcconfig:
* NetworkProcess/EntryPoint/Cocoa: Copied from Source/WebKit/NetworkProcess/EntryPoint/mac.
* NetworkProcess/EntryPoint/mac: Removed.
* PluginProcess/EntryPoint/Cocoa: Copied from Source/WebKit/PluginProcess/EntryPoint/mac.
* PluginProcess/EntryPoint/mac: Removed.
* Shared/EntryPointUtilities/Cocoa: Copied from Source/WebKit/Shared/EntryPointUtilities/mac.
* Shared/EntryPointUtilities/Cocoa/XPCService/XPCServiceEntryPoint.h:
* Shared/EntryPointUtilities/Cocoa/XPCService/XPCServiceEntryPoint.mm:
* Shared/EntryPointUtilities/Cocoa/XPCService/XPCServiceMain.mm:
* Shared/EntryPointUtilities/mac: Removed.
* SourcesCocoa.txt:
* WebKit.xcodeproj/project.pbxproj:
* WebProcess/EntryPoint/Cocoa: Copied from Source/WebKit/WebProcess/EntryPoint/mac.
* WebProcess/EntryPoint/mac: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240885 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: create icons for media event types instead of using a blue circle
drousso@apple.com [Fri, 1 Feb 2019 23:29:36 +0000 (23:29 +0000)]
Web Inspector: create icons for media event types instead of using a blue circle
https://bugs.webkit.org/show_bug.cgi?id=190381
<rdar://problem/45507995>

Reviewed by Brian Burg.

* UserInterface/Models/DOMNode.js:
(WI.DOMNode.isPlayEvent): Added.
(WI.DOMNode.isPauseEvent): Added.
(WI.DOMNode.isStopEvent): Added.

* UserInterface/Views/DOMEventsBreakdownView.js:
(WI.DOMEventsBreakdownView.prototype.layout):
* UserInterface/Views/DOMEventsBreakdownView.css:
(.dom-events-breakdown tr > :matches(th, td)):
(.dom-events-breakdown .graph):
(.dom-events-breakdown .graph > :matches(img, .area)): Added.
(.dom-events-breakdown .graph > img): Added.
(.dom-events-breakdown .inherited > .name, .dom-events-breakdown .inherited > .graph > img): Added.
(.dom-events-breakdown .graph > :matches(.point, .area)): Deleted.
(.dom-events-breakdown .graph > .point): Deleted.
(.dom-events-breakdown .inherited > .name, .dom-events-breakdown .inherited > .graph > .point): Deleted.

* UserInterface/Views/NetworkTableContentView.js:
(WI.NetworkTableContentView.prototype._populateWaterfallGraph.createDOMEventLine):

* UserInterface/Images/EventPause.svg: Added.
* UserInterface/Images/EventPlay.svg: Added.
* UserInterface/Images/EventProcessing.svg: Added.
* UserInterface/Images/EventStop.svg: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240884 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoRewrite and rebaseline smart copy paste tests to work on iOS and MacOS
megan_gardner@apple.com [Fri, 1 Feb 2019 23:27:30 +0000 (23:27 +0000)]
Rewrite and rebaseline smart copy paste tests to work on iOS and MacOS
https://bugs.webkit.org/show_bug.cgi?id=194175

Reviewed by Megan Gardner.

Using the functions committed in 194070 Ryosuke wrote, make the
smart copy paste tests run on both iOS and MacOS.
These tests will be turned on in 193786.

* editing/editing.js:
(runEditingTestWithCallbackLogging):
(runDumpAsTextEditingTest):
* editing/pasteboard/smart-paste-001.html:
* editing/pasteboard/smart-paste-002-expected.txt:
* editing/pasteboard/smart-paste-002.html:
* editing/pasteboard/smart-paste-003.html:
* editing/pasteboard/smart-paste-004.html:
* editing/pasteboard/smart-paste-005-expected.txt:
* editing/pasteboard/smart-paste-005.html:
* editing/pasteboard/smart-paste-006-expected.txt:
* editing/pasteboard/smart-paste-006.html:
* editing/pasteboard/smart-paste-007.html:
* editing/pasteboard/smart-paste-008.html:
* platform/mac/editing/pasteboard/smart-paste-008-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240883 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[iOS] Consistent 1 sec hang when triggering modal alerts while handling synchronous...
wenson_hsieh@apple.com [Fri, 1 Feb 2019 23:12:14 +0000 (23:12 +0000)]
[iOS] Consistent 1 sec hang when triggering modal alerts while handling synchronous touch events
https://bugs.webkit.org/show_bug.cgi?id=194140
<rdar://problem/47728098>

Reviewed by Tim Horton.

Source/WebKit:

Currently, the UI process hangs when attempting to synchronously present modal UI from the web process while the
UI process is waiting for sync IPC in the web process. While we have logic to generally mitigate IPC deadlock in
this scenario by dispatching the web process' sync IPC immediately with the intention of allowing the web
process to finish processing sync IPC (and consequently unblock the UI process), this fails in the case where
the sync IPC message from the web process to the UI process requires main thread execution for an arbitrary
amount of time (for instance, modal alert dialogs). In this case, we'll end up in a state where we've handled
the web process' sync IPC in the UI process, yet we can't resume execution since the web process is still
blocked.

By far the most egregious scenario in which this manifests is during synchronous gesture recognizer IPC, i.e.
grabbing position information from the UI process, and handling touch events synchronously. Luckily, these are
also cases where (1) we know sync IPC may safely time out, and (2) the presentation of modal UI from the web
process should cause the gesture recognizers to fail anyways. As such, we can mitigate these scenarios in the
web process by responding to the these pending sync IPC messages *before* sending our own sync IPC to the UI
process.

Test: fast/events/touch/ios/show-modal-alert-during-touch-start.html

* Shared/ios/InteractionInformationAtPosition.h:
(WebKit::InteractionInformationAtPosition::invalidInformation):
* Shared/ios/InteractionInformationAtPosition.mm:
(WebKit::InteractionInformationAtPosition::encode const):
(WebKit::InteractionInformationAtPosition::decode):

Add a new flag to indicate whether an interaction information response can be valid. Interaction information
cannot be valid in the case where the interaction information request was interrupted by certain sync IPC
messages from the web process.

* UIProcess/API/C/WKContextConfigurationRef.cpp:
(WKContextConfigurationIgnoreSynchronousMessagingTimeoutsForTesting):
(WKContextConfigurationSetIgnoreSynchronousMessagingTimeoutsForTesting):

Add some testing SPI to ignore sync IPC timeouts, for the purposes of testing. Rather than use the existing
Objective-C SPI in WKWebProcessPoolConfiguration, I decided to add C API plumbing for this flag, so that other
non-Cocoa ports may also support the new layout test option to ignore sync IPC timeouts.

* UIProcess/API/C/WKContextConfigurationRef.h:
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView ensurePositionInformationIsUpToDate:]):
(-[WKContentView _positionInformationDidChange:]):
* WebProcess/WebCoreSupport/WebChromeClient.cpp:
(WebKit::WebChromeClient::runBeforeUnloadConfirmPanel):
(WebKit::WebChromeClient::runJavaScriptAlert):
(WebKit::WebChromeClient::runJavaScriptConfirm):
(WebKit::WebChromeClient::runJavaScriptPrompt):
(WebKit::WebChromeClient::print):
(WebKit::WebChromeClient::exceededDatabaseQuota):
(WebKit::WebChromeClient::reachedApplicationCacheOriginQuota):

Cancel any pending sync IPC replies prior to sending sync IPC messages to the UI process which may result in
sync IPC deadlock, by using the new helper method, sendSyncWithDelayedReply, instead of just sendSync.

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::cancelGesturesBlockedOnSynchronousReplies):

Add a helper to cancel pending sync messages coming in from the UI process that are being called from within
gesture recognizer delegate hooks.

(WebKit::WebPage::touchEventSync):
* WebProcess/WebPage/WebPage.h:

Add a new helper, sendSyncWithDelayedReply, to be used when sending a sync IPC message to the UI process that
cannot be immediately completed upon arrival. Importantly, this cancels pending sync replies, and also passes
IPC::SendSyncOption::InformPlatformProcessWillSuspend.

* WebProcess/WebPage/WebPage.messages.in:

Change these from LegacySync to Delayed messages.

* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::getPositionInformation):

Make this sync IPC handler (as well as WebPage::touchEventSync) store the IPC reply during the scope of the
method, and invoke the stored reply at the end of the method if it wasn't interrupted due to something calling
cancelGesturesBlockedOnSynchronousReplies().

(WebKit::WebPage::positionInformation):

Refactor getPositionInformation by pulling out the logic for building an InteractionInformationAtPosition into
a separate helper.

(WebKit::WebPage::requestPositionInformation):

Tools:

* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::setShouldDismissJavaScriptAlertsAsynchronously):

Add a new TestRunner hook to make modal JavaScript alerts dismiss asynchronously. This is used by the new layout
test to induce an IPC deadlock when presenting a modal alert during touch start.

* WebKitTestRunner/InjectedBundle/TestRunner.h:
* WebKitTestRunner/TestController.cpp:
(WTR::runJavaScriptAlert):

Add a client callback function for running JavaScript alerts.

(WTR::TestController::createOtherPage):
(WTR::TestController::generateContextConfiguration const):

Add a test option to disable IPC timeouts for a layout test. This forces the test to reliably time out without
the fix in this patch.

(WTR::TestController::createWebViewWithOptions):

Plumb TestOptions to generateContextConfiguration.

(WTR::TestController::resetPreferencesToConsistentValues):
(WTR::TestController::resetStateToConsistentValues):
(WTR::updateTestOptionsFromTestHeader):
(WTR::TestController::setShouldDismissJavaScriptAlertsAsynchronously):
(WTR::TestController::handleJavaScriptAlert):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
* WebKitTestRunner/TestOptions.h:
(WTR::TestOptions::hasSameInitializationOptions const):

LayoutTests:

Add a test that induces sync IPC deadlock by presenting a modal alert while handling touch start. This test
forces sync IPC timeouts to be disabled, and passes if we do not time out while handling a touch.

* fast/events/touch/ios/show-modal-alert-during-touch-start-expected.txt: Added.
* fast/events/touch/ios/show-modal-alert-during-touch-start.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240882 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoNetwork Process crash when resuming downloads: '-[__NSDictionaryI setObject:forKey...
david_quesada@apple.com [Fri, 1 Feb 2019 23:08:48 +0000 (23:08 +0000)]
Network Process crash when resuming downloads: '-[__NSDictionaryI setObject:forKey:]: unrecognized selector sent to instance %p'
https://bugs.webkit.org/show_bug.cgi?id=194144
rdar://problem/47553456

Reviewed by Geoffrey Garen.

* NetworkProcess/Downloads/cocoa/DownloadCocoa.mm:
(WebKit::Download::resume):
    Make a mutable copy of the root object decoded from the resume data.
    It might have been originally encoded as an immutable dictionary.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240881 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoREGRESSION: Flaky ASSERTION FAILED: m_uncommittedState.state == State::Committed...
cdumez@apple.com [Fri, 1 Feb 2019 22:52:58 +0000 (22:52 +0000)]
REGRESSION: Flaky ASSERTION FAILED: m_uncommittedState.state == State::Committed on http/tests/cookies/same-site/fetch-after-top-level-navigation-initiated-from-iframe-in-cross-origin-page.html
https://bugs.webkit.org/show_bug.cgi?id=193740
<rdar://problem/47527267>

Reviewed by Alex Christensen.

Source/WebCore:

* loader/DocumentLoader.cpp:
(WebCore::DocumentLoader::willSendRequest):
(WebCore::DocumentLoader::continueAfterContentPolicy):
* loader/FrameLoader.cpp:
(WebCore::FrameLoader::loadURL):
(WebCore::FrameLoader::loadWithDocumentLoader):
(WebCore::FrameLoader::continueLoadAfterNavigationPolicy):
* loader/FrameLoader.h:
* loader/FrameLoaderTypes.h:
* loader/PolicyChecker.cpp:
(WebCore::PolicyChecker::checkNavigationPolicy):
(WebCore::PolicyChecker::checkNewWindowPolicy):
* loader/PolicyChecker.h:

Source/WebKit:

The issue was happening when the page is triggering a cross-site navigation while in the middle of parsing. This would cause us to
start a new provisional load in a new process before the previous process sends the DidFinishLoadForFrame() IPC to the UIProcess.
Getting such IPC after a provisional load has started would mess up our state machine and trip assertions.

This patch restores non-PSON behavior which is that the previous load in the old process now gets stopped so that no DidFinishLoadForFrame()
/ DidFailLoadForFrame() gets sent. To achieve this behavior, I introduced a new "StopAllLoads" PolicyAction that we now send the old
process when the load is continuing in a new process, instead of sending it "Ignore".

* NetworkProcess/NetworkDataTaskBlob.cpp:
(WebKit::NetworkDataTaskBlob::dispatchDidReceiveResponse):
* NetworkProcess/cocoa/NetworkSessionCocoa.mm:
(toNSURLSessionResponseDisposition):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::receivedNavigationPolicyDecision):

Tools:

Add API test coverage.

* TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240880 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUnreviewed build fix; Respect the 'condition:' field in the WebPreferences.yaml when...
jer.noble@apple.com [Fri, 1 Feb 2019 22:49:46 +0000 (22:49 +0000)]
Unreviewed build fix; Respect the 'condition:' field in the WebPreferences.yaml when generating
WebPreferencesStoreDefaultsMap.cpp.

* page/SettingsBase.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240879 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoRemove invalid assertion in DFG's compileDoubleRep().
mark.lam@apple.com [Fri, 1 Feb 2019 22:47:19 +0000 (22:47 +0000)]
Remove invalid assertion in DFG's compileDoubleRep().
https://bugs.webkit.org/show_bug.cgi?id=194130
<rdar://problem/47699474>

Reviewed by Saam Barati.

JSTests:

* stress/constant-fold-double-rep-into-double-constant.js: Added.

Source/JavaScriptCore:

* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileDoubleRep):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240878 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoDispatch pointercancel events when content is panned or zoomed on iOS
graouts@webkit.org [Fri, 1 Feb 2019 21:53:40 +0000 (21:53 +0000)]
Dispatch pointercancel events when content is panned or zoomed on iOS
https://bugs.webkit.org/show_bug.cgi?id=193962
<rdar://problem/47629134>

Reviewed by Dean Jackson.

Source/WebCore:

Expose two new methods on PointerCaptureController so that, given a pointer id, it can be established whether this pointer
has been cancelled, which is important because a cancelled pointer should no longer dispatch any further pointer events, and
to cancel a pointer.

Tests: pointerevents/ios/touch-action-pointercancel-pan-x.html
       pointerevents/ios/touch-action-pointercancel-pan-y.html
       pointerevents/ios/touch-action-pointercancel-pinch-zoom.html

* WebCore.xcodeproj/project.pbxproj: Make PointerCaptureController.h Private so that it can be imported from WebKit.
* dom/PointerEvent.h: Remove an unnecessary #if ENABLE(POINTER_EVENTS) since the entire file is already contained in one.
Then we add a new create() method that takes an event type, a pointer id and a pointer type (touch vs. pen) that we use
to create pointercancel events in PointerCaptureController::cancelPointer().
* page/Page.cpp:
(WebCore::Page::Page): Pass the Page as a parameter when creating the PointerCaptureController.
* page/PointerCaptureController.cpp:
(WebCore::PointerCaptureController::PointerCaptureController): Add a Page reference to the constructor since we'll need
the page to access its main frame's EventHandler to perform hit testing in case we do not have a capture target override
in cancelPointer().
(WebCore::PointerCaptureController::releasePointerCapture): Drive-by, remove the the implicit parameter since on iOS we
don't need to differentiate. We'll bring this back for the macOS work.
(WebCore::PointerCaptureController::hasCancelledPointerEventForIdentifier): New method we'll use when dispatching pointer
events to identify whether a pointer id has already been cancelled which will allow for _not_ dispatching any further
pointer events for this pointer id.
(WebCore::PointerCaptureController::pointerEventWillBeDispatched): Keep track of the pointer type so we can preserve it
when dispatching pointercancel events for a given pointer id.
(WebCore::PointerCaptureController::cancelPointer): Dispatch a pointercancel for the provided pointer id, using the capture
target override as the event's target, if there is one, and otherwise hit-testing at the provided location to figure out
what the target should be.
* page/PointerCaptureController.h: Switch the target overrides from Element* to RefPtr<Element> to ensure it may not be
deleted while we still need them. Existing code already ensures these get set to nullptr.

Source/WebKit:

When a user-agent-provided interaction, such as panning or zooming on iOS, uses a set of touches, we should dispatch a pointercancel
event for the pointer ids of the touches involved. To facilitate this, we add a new method on WKContentView to cancel all the pointers
matching active touches for a provided UIGestureRecognizer through an async IPC call into the Web process using the new method
PointerCaptureController::cancelPointer().

* Platform/spi/ios/UIKitSPI.h: Add the necessary forward declaration for a necessary UIKit SPI allowing us to get the set of last-seen
UITouches by the identifier generated for the matching WebKit touch.
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView scrollViewWillBeginZooming:withView:]): Dispatch touchcancel events for all pointers involved in a pinch gesture on the
top-level UIScrollView.
(-[WKWebView _scrollView:adjustedOffsetForOffset:translation:startPoint:locationInView:horizontalVelocity:verticalVelocity:]): Dispatch
touchcancel events for all pointers involved in a pan gesture on the top-level UIScrollView. We can infer this by looking at whether the
adjusted content offset, after accounting for the permitted touch actions, is different from the original content offset.
* UIProcess/PageClient.h: Expose a new virtual cancelPointersForGestureRecognizer() method which will allow the iOS implementation to
forward the call to WKContentViewInteraction.
(WebKit::PageClient::cancelPointersForGestureRecognizer):
* UIProcess/RemoteLayerTree/RemoteScrollingCoordinatorProxy.h: Expose the WebPageProxy such that we may access it to cancel pointers for
a given gesture recognizer from within ScrollingTreeScrollingNodeDelegateIOS.
(WebKit::RemoteScrollingCoordinatorProxy::webPageProxy const):
* UIProcess/RemoteLayerTree/ios/ScrollingTreeScrollingNodeDelegateIOS.h:
* UIProcess/RemoteLayerTree/ios/ScrollingTreeScrollingNodeDelegateIOS.mm:
(-[WKScrollingNodeScrollViewDelegate _scrollView:adjustedOffsetForOffset:translation:startPoint:locationInView:horizontalVelocity:verticalVelocity:]):
Dispatch touchcancel events for all pointers involved in a pan gesture on a nested UIScrollView. We can infer this by looking at
whether the adjusted content offset, after accounting for the permitted touch actions, is different from the original content offset.
(-[WKScrollingNodeScrollViewDelegate scrollViewWillBeginZooming:withView:]): Dispatch touchcancel events for all pointers involved in a
pinch gesture on a nested UIScrollView.
(-[WKScrollingNodeScrollViewDelegate cancelPointersForGestureRecognizer:]):
(WebKit::ScrollingTreeScrollingNodeDelegateIOS::cancelPointersForGestureRecognizer):
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::cancelPointer):
* UIProcess/WebPageProxy.h:
* UIProcess/ios/PageClientImplIOS.h:
* UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::cancelPointersForGestureRecognizer):
* UIProcess/ios/WKContentViewInteraction.h:
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView cancelPointersForGestureRecognizer:]): Obtain all active UITouch objects for the view and dispatch a pointercancel event,
through the WebPageProxy, for all touches associated with the provided gesture recognizer.
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::cancelPointer):
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

LayoutTests:

Adding a few tests for "pointercancel" and adding "touch-action: none" on tests that would now be affected by canceling pointers. We also unflake a few tests.

* pointerevents/ios/pointer-events-implicit-capture.html:
* pointerevents/ios/pointer-events-is-primary.html:
* pointerevents/ios/touch-action-pan-x-pan-y.html:
* pointerevents/ios/touch-action-pan-x.html:
* pointerevents/ios/touch-action-pan-y-expected.txt:
* pointerevents/ios/touch-action-pan-y.html:
* pointerevents/ios/touch-action-pinch-zoom-allows-zooming.html:
* pointerevents/ios/touch-action-pointercancel-pan-x-expected.txt: Added.
* pointerevents/ios/touch-action-pointercancel-pan-x.html: Added.
* pointerevents/ios/touch-action-pointercancel-pan-y-expected.txt: Added.
* pointerevents/ios/touch-action-pointercancel-pan-y.html: Added.
* pointerevents/ios/touch-action-pointercancel-pinch-zoom-expected.txt: Added.
* pointerevents/ios/touch-action-pointercancel-pinch-zoom.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240875 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoMake the WebKit default for media source based on the WebCore default.
jer.noble@apple.com [Fri, 1 Feb 2019 21:48:53 +0000 (21:48 +0000)]
Make the WebKit default for media source based on the WebCore default.
https://bugs.webkit.org/show_bug.cgi?id=194172

Reviewed by Eric Carlson.

Source/WebCore:

* page/SettingsBase.h:

Source/WebKit:

* Shared/WebPreferences.yaml:
* UIProcess/API/Cocoa/WKPreferences.mm:
(-[WKPreferences _setMediaSourceEnabled:]):
(-[WKPreferences _mediaSourceEnabled]):
* UIProcess/API/Cocoa/WKPreferencesPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240874 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Make WI.ColumnChart a WI.View subclass
joepeck@webkit.org [Fri, 1 Feb 2019 21:40:25 +0000 (21:40 +0000)]
Web Inspector: Make WI.ColumnChart a WI.View subclass
https://bugs.webkit.org/show_bug.cgi?id=194171

Rubber-stamped by Devin Rousso.

* UserInterface/Views/CPUTimelineOverviewGraph.js:
(WI.CPUTimelineOverviewGraph):
* UserInterface/Views/ColumnChart.js:
(WI.ColumnChart):
(WI.ColumnChart.prototype.set size):
(WI.ColumnChart.prototype.layout):
(WI.ColumnChart.prototype.get element): Deleted.
(WI.ColumnChart.prototype.needsLayout): Deleted.
(WI.ColumnChart.prototype.updateLayout): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240872 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Make WI.StackedLineChart a WI.View subclass
joepeck@webkit.org [Fri, 1 Feb 2019 21:24:20 +0000 (21:24 +0000)]
Web Inspector: Make WI.StackedLineChart a WI.View subclass
https://bugs.webkit.org/show_bug.cgi?id=194119

Rubber-stamped by Devin Rousso.

* UserInterface/Views/LineChart.js:
(WI.LineChart.prototype.layout):
(WI.LineChart):
* UserInterface/Views/MemoryTimelineOverviewGraph.js:
(WI.MemoryTimelineOverviewGraph):
* UserInterface/Views/StackedLineChart.js:
(WI.StackedLineChart):
(WI.StackedLineChart.prototype.set size):
(WI.StackedLineChart.prototype.layout):
(WI.StackedLineChart.prototype.get element): Deleted.
(WI.StackedLineChart.prototype.get points): Deleted.
(WI.StackedLineChart.prototype.needsLayout): Deleted.
(WI.StackedLineChart.prototype.updateLayout): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240871 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Make WI.CircleChart a WI.View subclass
joepeck@webkit.org [Fri, 1 Feb 2019 21:24:18 +0000 (21:24 +0000)]
Web Inspector: Make WI.CircleChart a WI.View subclass
https://bugs.webkit.org/show_bug.cgi?id=194118

Reviewed by Matt Baker.

* UserInterface/Views/CircleChart.js:
(WI.CircleChart.prototype.get centerElement):
(WI.CircleChart.prototype.layout):
(WI.CircleChart.prototype.get element): Deleted.
(WI.CircleChart.prototype.needsLayout): Deleted.
(WI.CircleChart.prototype.updateLayout): Deleted.
* UserInterface/Views/MemoryTimelineView.js:
(WI.MemoryTimelineView):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240870 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Timeline Detail Views do not reset properly when new time range select...
joepeck@webkit.org [Fri, 1 Feb 2019 21:24:16 +0000 (21:24 +0000)]
Web Inspector: Timeline Detail Views do not reset properly when new time range selection contains nothing
https://bugs.webkit.org/show_bug.cgi?id=194115
<rdar://problem/47716693>

Rubber-stamped by Devin Rousso.

* UserInterface/Views/CPUTimelineView.js:
(WI.CPUTimelineView.prototype.reset):
(WI.CPUTimelineView.prototype.clear):
(WI.CPUTimelineView.prototype.layout):
* UserInterface/Views/MemoryTimelineView.js:
(WI.MemoryTimelineView.prototype.reset):
(WI.MemoryTimelineView.prototype.clear):
(WI.MemoryTimelineView.prototype.layout):
When there are no visible records in the selected range clear our UI.
Introduce a `clear` method that clears the UI but keeps
non-range-specific values (e.g. maximums).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240869 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Memory timeline starts from zero when both CPU and Memory timeline...
joepeck@webkit.org [Fri, 1 Feb 2019 21:24:14 +0000 (21:24 +0000)]
Web Inspector: Memory timeline starts from zero when both CPU and Memory timeline are enabled
https://bugs.webkit.org/show_bug.cgi?id=194111
<rdar://problem/47714555>

Rubber-stamped by Devin Rousso.

* page/ResourceUsageThread.cpp:
(WebCore::ResourceUsageThread::waitUntilObservers):
(WebCore::ResourceUsageThread::threadBody):
Wait a short period of time before the first listener registers
and we start sampling. This will allow multiple listeners to
register, each that wants different data.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240868 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Timeline graphs have drawing issues with multiple discontinuities
joepeck@webkit.org [Fri, 1 Feb 2019 21:24:12 +0000 (21:24 +0000)]
Web Inspector: Timeline graphs have drawing issues with multiple discontinuities
https://bugs.webkit.org/show_bug.cgi?id=194110
<rdar://problem/47714356>

Reviewed by Devin Rousso.

* UserInterface/Views/CPUTimelineView.js:
(WI.CPUTimelineView):
(WI.CPUTimelineView.prototype.layout.xScale): Deleted.
* UserInterface/Views/MemoryTimelineOverviewGraph.js:
(WI.MemoryTimelineOverviewGraph.prototype.layout.insertDiscontinuity):
(WI.MemoryTimelineOverviewGraph.prototype.layout):
* UserInterface/Views/MemoryTimelineView.js:
(WI.MemoryTimelineView.prototype.layout.xScale):
(WI.MemoryTimelineView.prototype.layout.yScale):
(WI.MemoryTimelineView.prototype.layout):
Handle if multiple discontinuities exist between records.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240867 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Timeline time range selection should show duration alongside start...
joepeck@webkit.org [Fri, 1 Feb 2019 21:24:09 +0000 (21:24 +0000)]
Web Inspector: Timeline time range selection should show duration alongside start and end
https://bugs.webkit.org/show_bug.cgi?id=194109
<rdar://problem/47714279>

Reviewed by Devin Rousso.

* Localizations/en.lproj/localizedStrings.js:
* UserInterface/Views/TimelineRecordingContentView.js:
(WI.TimelineRecordingContentView.prototype._updateTimeRangePathComponents):
Include the duration when not obvious.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240866 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoWeb Inspector: Improve API and documentation of ColumnChart
joepeck@webkit.org [Fri, 1 Feb 2019 21:23:51 +0000 (21:23 +0000)]
Web Inspector: Improve API and documentation of ColumnChart
https://bugs.webkit.org/show_bug.cgi?id=193982

Reviewed by Devin Rousso.

This used to be named "BarChart". Convert remaining instances
of "bar" to "column" and clean up related things.

* UserInterface/Views/CPUTimelineOverviewGraph.css:
(body[dir=rtl] .timeline-overview-graph.cpu > .column-chart):
(.timeline-overview-graph.cpu > .column-chart > svg > rect):
(body[dir=rtl] .timeline-overview-graph.cpu > .bar-chart): Deleted.
(.timeline-overview-graph.cpu > .bar-chart > svg > rect): Deleted.
* UserInterface/Views/CPUTimelineOverviewGraph.js:
(WI.CPUTimelineOverviewGraph.prototype.layout):
* UserInterface/Views/ColumnChart.js:
(WI.ColumnChart):
(WI.ColumnChart.prototype.get columns):
(WI.ColumnChart.prototype.addColumn):
(WI.ColumnChart.prototype.clear):
(WI.ColumnChart.prototype.updateLayout):
(WI.ColumnChart.prototype.get bars): Deleted.
(WI.ColumnChart.prototype.addBar): Deleted.
* UserInterface/Views/StackedLineChart.js:
(WI.StackedLineChart.prototype.get element):
(WI.StackedLineChart.prototype.get points):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240865 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[JSC] Unify CodeBlock IsoSubspaces
ysuzuki@apple.com [Fri, 1 Feb 2019 21:14:19 +0000 (21:14 +0000)]
[JSC] Unify CodeBlock IsoSubspaces
https://bugs.webkit.org/show_bug.cgi?id=194167

Reviewed by Saam Barati.

When we move CodeBlock into its IsoSubspace, we create IsoSubspaces for each subclass of CodeBlock.
But this is not necessary since,

1. They do not override the classInfo methods.
2. sizeof(ProgramCodeBlock etc.) == sizeof(CodeBlock) since subclasses adds no additional fields.

Creating IsoSubspace for each subclass is costly in terms of memory. Especially, IsoSubspace for
ProgramCodeBlock is. We typically create only one ProgramCodeBlock, and it means the rest of the
MarkedBlock (16KB - sizeof(footer) - sizeof(ProgramCodeBlock)) is just wasted.

This patch unifies these IsoSubspaces into one.

* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::destroy):
* bytecode/CodeBlock.h:
* bytecode/EvalCodeBlock.cpp:
(JSC::EvalCodeBlock::destroy): Deleted.
* bytecode/EvalCodeBlock.h: We drop some utility functions in EvalCodeBlock and use UnlinkedEvalCodeBlock's one directly.
* bytecode/FunctionCodeBlock.cpp:
(JSC::FunctionCodeBlock::destroy): Deleted.
* bytecode/FunctionCodeBlock.h:
* bytecode/GlobalCodeBlock.h:
* bytecode/ModuleProgramCodeBlock.cpp:
(JSC::ModuleProgramCodeBlock::destroy): Deleted.
* bytecode/ModuleProgramCodeBlock.h:
* bytecode/ProgramCodeBlock.cpp:
(JSC::ProgramCodeBlock::destroy): Deleted.
* bytecode/ProgramCodeBlock.h:
* interpreter/Interpreter.cpp:
(JSC::Interpreter::execute):
* runtime/VM.cpp:
(JSC::VM::VM):
* runtime/VM.h:
(JSC::VM::forEachCodeBlockSpace):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240864 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[ews-app] Add display name field to BuilderMapping model
aakash_jain@apple.com [Fri, 1 Feb 2019 20:22:51 +0000 (20:22 +0000)]
[ews-app] Add display name field to BuilderMapping model
https://bugs.webkit.org/show_bug.cgi?id=194153

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-app/ews/models/buildermapping.py:
(BuilderMapping):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240863 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoForce settings.acceleratedCompositingEnabled() to true for WebKit
simon.fraser@apple.com [Fri, 1 Feb 2019 20:20:26 +0000 (20:20 +0000)]
Force settings.acceleratedCompositingEnabled() to true for WebKit
https://bugs.webkit.org/show_bug.cgi?id=194166
rdar://problem/47694328

Reviewed by Tim Horton.

WebKit relies on accelerated compositing for its drawing areas to function correctly,
and sometimes clients (*cough* Safari *cough*) have bugs where they clobber the settings,
so force this setting to be always on (and release log if this happens).

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::updatePreferences):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240862 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoDon't use base layer() as the scroll layer in scrolling tree.
antti@apple.com [Fri, 1 Feb 2019 20:10:23 +0000 (20:10 +0000)]
Don't use base layer() as the scroll layer in scrolling tree.
https://bugs.webkit.org/show_bug.cgi?id=194160

Reviewed by Simon Fraser.

Source/WebCore:

Maintain scrollContainerLayer() and scrolledContentsLayer() separately in ScrollingTreeScrollingNode.

* page/scrolling/AsyncScrollingCoordinator.cpp:
(WebCore::AsyncScrollingCoordinator::frameViewRootLayerDidChange):
(WebCore::AsyncScrollingCoordinator::setNodeLayers):
* page/scrolling/AsyncScrollingCoordinator.h:
* page/scrolling/ScrollingCoordinator.h:
(WebCore::ScrollingCoordinator::setNodeLayers):

Turn the long layer parameter list into a struct.

* page/scrolling/ScrollingStateNode.cpp:
(WebCore::ScrollingStateNode::ScrollingStateNode):
(WebCore::ScrollingStateNode::setAllPropertiesChanged):
(WebCore::ScrollingStateNode::setLayer):
* page/scrolling/ScrollingStateNode.h:
* page/scrolling/ScrollingStateScrollingNode.cpp:
(WebCore::ScrollingStateScrollingNode::ScrollingStateScrollingNode):
(WebCore::ScrollingStateScrollingNode::setAllPropertiesChanged):
(WebCore::ScrollingStateScrollingNode::setScrollContainerLayer):
(WebCore::ScrollingStateScrollingNode::dumpProperties const):
* page/scrolling/ScrollingStateScrollingNode.h:
(WebCore::ScrollingStateScrollingNode::scrollContainerLayer const):
* page/scrolling/ScrollingTree.cpp:
(WebCore::ScrollingTree::commitTreeState):
* page/scrolling/ScrollingTreeScrollingNode.cpp:
(WebCore::ScrollingTreeScrollingNode::commitStateBeforeChildren):
* page/scrolling/ScrollingTreeScrollingNode.h:
(WebCore::ScrollingTreeScrollingNode::scrollContainerLayer const):
(WebCore::ScrollingTreeScrollingNode::scrolledContentsLayer const):
* page/scrolling/ios/ScrollingTreeFrameScrollingNodeIOS.h:
* page/scrolling/ios/ScrollingTreeFrameScrollingNodeIOS.mm:
(WebCore::ScrollingTreeFrameScrollingNodeIOS::commitStateBeforeChildren):
(WebCore::ScrollingTreeFrameScrollingNodeIOS::scrollPosition const):
(WebCore::ScrollingTreeFrameScrollingNodeIOS::setScrollLayerPosition):
(WebCore::ScrollingTreeFrameScrollingNodeIOS::scrollLayer const): Deleted.
* page/scrolling/mac/ScrollingTreeFixedNode.mm:
(WebCore::ScrollingTreeFixedNode::commitStateBeforeChildren):
* page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.h:
* page/scrolling/mac/ScrollingTreeFrameScrollingNodeMac.mm:
(WebCore::ScrollingTreeFrameScrollingNodeMac::commitStateBeforeChildren):
(WebCore::ScrollingTreeFrameScrollingNodeMac::commitStateAfterChildren):
(WebCore::ScrollingTreeFrameScrollingNodeMac::scrollPosition const):
(WebCore::ScrollingTreeFrameScrollingNodeMac::setScrollLayerPosition):
(WebCore::ScrollingTreeFrameScrollingNodeMac::exposedUnfilledArea const):
* page/scrolling/mac/ScrollingTreeOverflowScrollingNodeMac.h:
* page/scrolling/mac/ScrollingTreeOverflowScrollingNodeMac.mm:
(WebCore::ScrollingTreeOverflowScrollingNodeMac::commitStateBeforeChildren):
(WebCore::ScrollingTreeOverflowScrollingNodeMac::scrollPosition const):
(WebCore::ScrollingTreeOverflowScrollingNodeMac::setScrollLayerPosition):
* page/scrolling/mac/ScrollingTreeStickyNode.mm:
(WebCore::ScrollingTreeStickyNode::commitStateBeforeChildren):
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::didChangePlatformLayerForLayer):
(WebCore::RenderLayerCompositor::updateScrollingNodeForViewportConstrainedRole):
(WebCore::RenderLayerCompositor::updateScrollingNodeForScrollingRole):
(WebCore::RenderLayerCompositor::updateScrollingNodeForFrameHostingRole):

Source/WebKit:

* Shared/RemoteLayerTree/RemoteScrollingCoordinatorTransaction.cpp:
(ArgumentCoder<ScrollingStateNode>::encode):
(ArgumentCoder<ScrollingStateNode>::decode):
(ArgumentCoder<ScrollingStateScrollingNode>::encode):
(ArgumentCoder<ScrollingStateScrollingNode>::decode):
(WebKit::dump):
* UIProcess/RemoteLayerTree/RemoteScrollingCoordinatorProxy.cpp:
(WebKit::RemoteScrollingCoordinatorProxy::connectStateNodeLayers):
* UIProcess/RemoteLayerTree/ios/RemoteScrollingCoordinatorProxyIOS.mm:
(WebKit::RemoteScrollingCoordinatorProxy::connectStateNodeLayers):
* UIProcess/RemoteLayerTree/ios/ScrollingTreeFrameScrollingNodeRemoteIOS.mm:
(WebKit::ScrollingTreeFrameScrollingNodeRemoteIOS::commitStateBeforeChildren):
* UIProcess/RemoteLayerTree/ios/ScrollingTreeOverflowScrollingNodeIOS.mm:
(WebKit::ScrollingTreeOverflowScrollingNodeIOS::commitStateBeforeChildren):
* UIProcess/RemoteLayerTree/ios/ScrollingTreeScrollingNodeDelegateIOS.mm:
(WebKit::ScrollingTreeScrollingNodeDelegateIOS::commitStateBeforeChildren):
(WebKit::ScrollingTreeScrollingNodeDelegateIOS::commitStateAfterChildren):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240861 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoUnreviewed, follow-up after r240859
ysuzuki@apple.com [Fri, 1 Feb 2019 19:33:21 +0000 (19:33 +0000)]
Unreviewed, follow-up after r240859
https://bugs.webkit.org/show_bug.cgi?id=194145

Replace OOB HeapCellType with cellHeapCellType since they are completely the same.
And rename cellDangerousBitsSpace back to cellSpace.

* runtime/JSCellInlines.h:
(JSC::JSCell::subspaceFor):
* runtime/VM.cpp:
(JSC::VM::VM):
* runtime/VM.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240860 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[JSC] Remove cellJSValueOOBSpace
ysuzuki@apple.com [Fri, 1 Feb 2019 18:59:55 +0000 (18:59 +0000)]
[JSC] Remove cellJSValueOOBSpace
https://bugs.webkit.org/show_bug.cgi?id=194145

Reviewed by Mark Lam.

* runtime/JSObject.h:
(JSC::JSObject::subspaceFor): Deleted.
* runtime/VM.cpp:
(JSC::VM::VM):
* runtime/VM.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240859 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[SOUP] Improve use of SoupCookiePersistentStorageType
mcatanzaro@igalia.com [Fri, 1 Feb 2019 17:49:10 +0000 (17:49 +0000)]
[SOUP] Improve use of SoupCookiePersistentStorageType
https://bugs.webkit.org/show_bug.cgi?id=194103

Reviewed by Carlos Garcia Campos.

Turn it into an enum class, and stop casting it to and from uint32_t.

* NetworkProcess/Cookies/WebCookieManager.h:
* NetworkProcess/Cookies/WebCookieManager.messages.in:
* NetworkProcess/Cookies/soup/WebCookieManagerSoup.cpp:
(WebKit::WebCookieManager::setCookiePersistentStorage):
* NetworkProcess/NetworkSessionCreationParameters.cpp:
(WebKit::NetworkSessionCreationParameters::privateSessionParameters):
(WebKit::NetworkSessionCreationParameters::decode):
* NetworkProcess/NetworkSessionCreationParameters.h:
* Shared/soup/SoupCookiePersistentStorageType.h:
(): Deleted.
* UIProcess/API/glib/WebKitCookieManager.cpp:
* UIProcess/WebCookieManagerProxy.h:
* UIProcess/soup/WebCookieManagerProxySoup.cpp:
(WebKit::WebCookieManagerProxy::setCookiePersistentStorage):
(WebKit::WebCookieManagerProxy::getCookiePersistentStorage const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240858 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[macOS] Revert parts of r240811, since it causes kernel panics.
pvollan@apple.com [Fri, 1 Feb 2019 17:47:10 +0000 (17:47 +0000)]
[macOS] Revert parts of r240811, since it causes kernel panics.
https://bugs.webkit.org/show_bug.cgi?id=194061

Unreviewed crash fix.

* WebProcess/com.apple.WebProcess.sb.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240857 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoAPI Test broken: TestWebKitAPI.WebKit2.GetUserMediaReprompt
youenn@apple.com [Fri, 1 Feb 2019 17:45:22 +0000 (17:45 +0000)]
API Test broken: TestWebKitAPI.WebKit2.GetUserMediaReprompt
https://bugs.webkit.org/show_bug.cgi?id=194155

Reviewed by Eric Carlson.

* TestWebKitAPI/Tests/WebKit/getUserMedia.html:
Revert use of async functions as it seems it cannot be evaluated from TestWebKitAPI.
Keep calling enumerateDevices to continue cover https://bugs.webkit.org/show_bug.cgi?id=194106.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240856 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[SOUP] Improve use of PAL::SessionID in WebKitCookieManager
mcatanzaro@igalia.com [Fri, 1 Feb 2019 16:51:14 +0000 (16:51 +0000)]
[SOUP] Improve use of PAL::SessionID in WebKitCookieManager
https://bugs.webkit.org/show_bug.cgi?id=194092

Reviewed by Carlos Garcia Campos.

Always get it from the data store instead of assuming the default session ID, and add a
helper function for this.

* UIProcess/API/glib/WebKitCookieManager.cpp:
(_WebKitCookieManagerPrivate::sessionID const):
(_WebKitCookieManagerPrivate::~_WebKitCookieManagerPrivate):
(webkitCookieManagerCreate):
(webkit_cookie_manager_set_persistent_storage):
(webkit_cookie_manager_set_accept_policy):
(webkit_cookie_manager_get_accept_policy):
(webkit_cookie_manager_add_cookie):
(webkit_cookie_manager_get_cookies):
(webkit_cookie_manager_delete_cookie):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240850 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoREGRESSION (r240698): fast/scrolling/sticky-to-fixed.html can cause a crash
simon.fraser@apple.com [Fri, 1 Feb 2019 16:45:15 +0000 (16:45 +0000)]
REGRESSION (r240698): fast/scrolling/sticky-to-fixed.html can cause a crash
https://bugs.webkit.org/show_bug.cgi?id=194134
rdar://problem/47721210

Reviewed by Daniel Bates.

fast/scrolling/sticky-to-fixed.html changes the scrolling node type, which causes
scrollingCoordinator->insertNode() to return a different ScrollingNodeID to the one
passed in. We have to handle this, removing the node for the nodeID and unregistering
the layer with the old nodeID.

* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::attachScrollingNode):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoREGRESSION(r239915): css3/font-feature-font-face-local.html failing on WPE
carlosgc@webkit.org [Fri, 1 Feb 2019 14:36:31 +0000 (14:36 +0000)]
REGRESSION(r239915): css3/font-feature-font-face-local.html failing on WPE
https://bugs.webkit.org/show_bug.cgi?id=194008

Reviewed by Žan Doberšek.

Source/WebCore:

We need to replace control characters with zero width space too.

* platform/graphics/freetype/GlyphPageTreeNodeFreeType.cpp:
(WebCore::GlyphPage::fill):
* platform/text/CharacterProperties.h:
(WebCore::isControlCharacter):

LayoutTests:

Remove test expectation.

* platform/wpe/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240848 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[WPE] Enable font variations
carlosgc@webkit.org [Fri, 1 Feb 2019 14:35:08 +0000 (14:35 +0000)]
[WPE] Enable font variations
https://bugs.webkit.org/show_bug.cgi?id=194148

Reviewed by Žan Doberšek.

.:

Enable variation fonts when required dependencies are available.

* Source/cmake/OptionsWPE.cmake:

LayoutTests:

* platform/wpe/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240847 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months agoComplexText: Test fast/text/word-space-nbsp.html fails when using complex text path
carlosgc@webkit.org [Fri, 1 Feb 2019 14:33:15 +0000 (14:33 +0000)]
ComplexText: Test fast/text/word-space-nbsp.html fails when using complex text path
https://bugs.webkit.org/show_bug.cgi?id=193279

Reviewed by Michael Catanzaro.

Source/WebCore:

The fix applied in r135884 to WidthIterator should also be applied to ComplexTextController.

* platform/graphics/ComplexTextController.cpp:
(WebCore::ComplexTextController::adjustGlyphsAndAdvances):

LayoutTests:

Remove test expectation.

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240846 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[LFC] Adjust replaced element's intrinsic ratio
zalan@apple.com [Fri, 1 Feb 2019 14:31:04 +0000 (14:31 +0000)]
[LFC] Adjust replaced element's intrinsic ratio
https://bugs.webkit.org/show_bug.cgi?id=194154

Reviewed by Antti Koivisto.

Source/WebCore:

Decouple image and iframe replaced types and set intrinsic ratio accordingly.

* layout/layouttree/LayoutBox.cpp:
(WebCore::Layout::Box::Box):
* layout/layouttree/LayoutBox.h:
(WebCore::Layout::Box::isReplaced const):
(WebCore::Layout::Box::isIFrame const):
(WebCore::Layout::Box::isImage const):
* layout/layouttree/LayoutReplaced.cpp:
(WebCore::Layout::Replaced::hasIntrinsicRatio const):
(WebCore::Layout::Replaced::intrinsicRatio const):
(WebCore::Layout::Replaced::hasAspectRatio const):
* layout/layouttree/LayoutReplaced.h:
* layout/layouttree/LayoutTreeBuilder.cpp:
(WebCore::Layout::TreeBuilder::createSubTree):

Tools:

761

* LayoutReloaded/misc/LFC-passing-tests.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240845 268f45cc-cd09-0410-ab3c-d52691b4dbfc

7 months ago[WPE] Show test results in minibrowser
carlosgc@webkit.org [Fri, 1 Feb 2019 14:28:57 +0000 (14:28 +0000)]
[WPE] Show test results in minibrowser
https://bugs.webkit.org/show_bug.cgi?id=194150

Reviewed by Žan Doberšek.

* Scripts/webkitpy/port/wpe.py:
(WPEPort.show_results_html_file):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@240844 268f45cc-cd09-0410-ab3c-d52691b4dbfc