Timeline



May 3, 2019:

11:58 PM Changeset in webkit [244947] by Simon Fraser
  • 3 edits
    4 adds in trunk

[macOS] Fix programmatic scrolling with async overflow scroll
https://bugs.webkit.org/show_bug.cgi?id=197590

Reviewed by Sam Weinig.
Source/WebCore:

ScrollingTreeOverflowScrollingNodeMac needs to handle RequestedScrollPosition.

Tests: scrollingcoordinator/mac/programmatic-frame-scroll.html

scrollingcoordinator/mac/programmatic-overflow-scroll.html

  • page/scrolling/mac/ScrollingTreeOverflowScrollingNodeMac.mm:

(WebCore::ScrollingTreeOverflowScrollingNodeMac::commitStateAfterChildren):

LayoutTests:

Tests for programmatic scrolling of overflow and iframes.

  • scrollingcoordinator/mac/programmatic-frame-scroll-expected.html: Added.
  • scrollingcoordinator/mac/programmatic-frame-scroll.html: Added.
  • scrollingcoordinator/mac/programmatic-overflow-scroll-expected.html: Added.
  • scrollingcoordinator/mac/programmatic-overflow-scroll.html: Added.
7:45 PM Changeset in webkit [244946] by Devin Rousso
  • 6 edits in trunk/Source/WebInspectorUI

Web Inspector: DOM: dragging a node to the console should log the node
https://bugs.webkit.org/show_bug.cgi?id=197330

Reviewed by Timothy Hatcher.

  • UserInterface/Views/DOMTreeOutline.js:

(WI.DOMTreeOutline.prototype._ondragstart):

  • UserInterface/Views/LogContentView.js:

(WI.LogContentView):
(WI.LogContentView.prototype._handleDragOver): Added.
(WI.LogContentView.prototype._handleDrop): Added.

  • UserInterface/Views/QuickConsole.js:

(WI.QuickConsole):
(WI.QuickConsole.prototype._handleDragOver): Added.
(WI.QuickConsole.prototype._handleDrop): Added.

  • UserInterface/Views/GeneralStyleDetailsSidebarPanel.js:

Drive-by: update the format to be more unique.

  • Localizations/en.lproj/localizedStrings.js:
7:40 PM Changeset in webkit [244945] by Wenson Hsieh
  • 6 edits in trunk

REGRESSION (r244897): Caret may appear wider than normal after zooming to focus an editable element
https://bugs.webkit.org/show_bug.cgi?id=197579

Reviewed by Tim Horton.

Source/WebKit:

Fixes a couple of flaky tests (CaretSelectionRectAfterRestoringFirstResponderWithRetainActiveFocusedState and
CaretSelectionRectAfterRestoringFirstResponder) that began failing after r244897. These tests both begin by
focusing an editable element, which causes the web view to zoom in. The tests subsequently check that the caret
rect is {{ 16, 13 }, { 2, 15 }}. While the specified caret rect (coming from EditorState) is {{ 16, 13 }, { 3,
15 }}, the narrower caret rect is used because we take measures to preserve the width of the caret relative to
the view (see the inverse scaling logic in -[WKContentView selectedTextRange] for more details).

See below for more details.

  • UIProcess/ios/WKContentViewInteraction.h:

Remove _isZoomingToRevealFocusedElement, now that we don't need it anymore (see -observeValueForKeyPath:).

  • UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView cleanupInteraction]):
(-[WKContentView observeValueForKeyPath:ofObject:change:context:]):

Stop bailing from a selection update when changing scale, while zooming to reveal the focused element. This
check was added in r239441 to prevent UIWKTextInteractionAssistant's selection scrolling logic from interfering
with WKContentView-driven logic for zooming to the focused element. However, since r244141, this is no longer
necessary since selection scrolling is only driven by code in the web process.

This new update while zooming to reveal the focused element ensures that the WKTextRange returned by
-selectedTextRange after zooming will have a width that is inversely scaled using the content view's current
scale, such that it has a consistent width (relative to the web view) across different scales.

(-[WKContentView _zoomToRevealFocusedElement]):

  • UIProcess/ios/WebPageProxyIOS.mm:

(WebKit::WebPageProxy::layerTreeCommitComplete):

Remove any attempt here to notify PageClient about editor states after focus. This logic was actually incorrect,
since it didn't ensure that the layer tree commit that is being completed actually contained an editor state; as
such, the "editor state" received here could be stale.

Tools:

Fixes a couple of flaky layout tests (ModifyInputAssistantItemBarButtonGroups and
OverrideInputAssistantItemBarButtonGroups) by programmatically blurring focused elements and waiting for the
input session to change, rather than relying on -resignFirstResponder and -waitForNextPresentationUpdate to
ensure that the the focused element has been blurred.

  • TestWebKitAPI/Tests/ios/KeyboardInputTestsIOS.mm:
7:25 PM Changeset in webkit [244944] by Alan Bujtas
  • 7 edits in trunk/Source

[iOS] outlook.live.com: Compose email frame not fully visible and not scrollable
https://bugs.webkit.org/show_bug.cgi?id=197573
<rdar://problem/48008441>

Reviewed by Wenson Hsieh.

Source/WebCore:

The outlook mail view's flex column setup produces a somewhat unfortunate layout at certain viewport widths.
This patch addresses the issue by ensuring that we never fall into that range.

  • page/Quirks.cpp:

(WebCore::Quirks::shouldIgnoreShrinkToFitContent const):
(WebCore::Quirks::overriddenViewportLayoutWidth const):

  • page/Quirks.h:

Source/WebKit:

  • WebProcess/WebPage/WebPage.h:
  • WebProcess/WebPage/ios/WebPageIOS.mm:

(WebKit::WebPage::adjustLayoutSizeIfNeeded):
(WebKit::WebPage::setViewportConfigurationViewLayoutSize):
(WebKit::WebPage::dynamicViewportSizeUpdate):
(WebKit::WebPage::immediatelyShrinkToFitContent):

6:33 PM Changeset in webkit [244943] by ap@apple.com
  • 1 edit
    1 add in trunk/Tools

Add a tool to block spammer accounts
https://bugs.webkit.org/show_bug.cgi?id=197537

Reviewed by Darin Adler.

Usage: block-spammers email1 email2 ...

The tool finds and hides all bugs and comments from this account. Before any work
is done, these bugs and comments are displayed for confirmation.

  • Scripts/block-spammers: Added.

(get_bugzilla_token):
(invalidate_bugzilla_token):
(get_user_info):
(get_user_info_self):
(can_use_this_tool):
(disable_user):
(get_bugs_created_by_user):
(get_bugs_commented_on_by_user):
(get_comments):
(hide_bug):
(hide_comments):
(ask_yes_no):
(sanitized_string):
(main):

6:03 PM Changeset in webkit [244942] by Jonathan Bedard
  • 3 edits in trunk/Tools

run-apit-tests: Add --iterations and --repeat-each
https://bugs.webkit.org/show_bug.cgi?id=197563

Reviewed by Aakash Jain.

  • Scripts/webkitpy/api_tests/manager.py:

(Manager.run): Add --iterations and --repeat-each/

  • Scripts/webkitpy/api_tests/run_api_tests.py:

(parse_args): Duplicate the test list based on --iterations and --repeat-each.

6:01 PM Changeset in webkit [244941] by dino@apple.com
  • 3 edits in trunk/Source/WebKit

Make imageForElementActionType a class method on _WKElementAction
https://bugs.webkit.org/show_bug.cgi?id=197588
<rdar://problem/50463157>

Reviewed by Wenson Hsieh.

Make this helper function a class method, so that it can
be more easily detected.

  • UIProcess/API/Cocoa/_WKElementAction.h:
  • UIProcess/API/Cocoa/_WKElementAction.mm:

(+[_WKElementAction imageForElementActionType:]):
(_WKUIImageForElementActionType): Deleted.

5:17 PM Changeset in webkit [244940] by aakash_jain@apple.com
  • 2 edits in trunk/Tools

New EWS: Clicking on white bubble navigates to page with only bubbles
https://bugs.webkit.org/show_bug.cgi?id=197520

Reviewed by Lucas Forschler.

  • BuildSlaveSupport/ews-app/ews/templates/statusbubble.html: Disable clicking if bubble doesn't have any url.
4:57 PM Changeset in webkit [244939] by ysuzuki@apple.com
  • 4 edits
    1 add in trunk

[JSC] Need to emit SetLocal if we emit MovHint in DFGByteCodeParser
https://bugs.webkit.org/show_bug.cgi?id=197584

Reviewed by Saam Barati.

JSTests:

  • stress/adhoc-setter-frame-should-emit-setlocal-again.js: Added.

(X):
(foo):

Source/JavaScriptCore:

In r244864, we emit MovHint for adhocly created GetterCall/SetterCall frame locals in the callee side to make OSR availability analysis's pruning correct.
However, we just emit MovHint, and we do not emit SetLocal since we ensured that these locals are already flushed in the same place before. However, MovHint
and SetLocal are needed to be a pair in DFGByteCodeParser because we rely on this assumption in SSA conversion phase. SSA conversion phase always emit KillStack
just before MovHint's target location even if the MovHint's target is the same to the previously emitted MovHint and SetLocal.
This patch emits SetLocal too when emitting MovHint for GetterCall/SetterCall frame locals.

The example is like this.

a: SomeValueNode

: MovHint(@a, loc10)

b: SetLocal(@a, loc10)

...

c: MovHint(@a, loc10)

Then, this will be converted to the style in SSA conversion.

a: SomeValueNode

: KillStack(loc10)

b: PutStack(@a, loc10)

...

c: KillStack(loc10)

Then, @b will be removed later since @c kills it.

  • dfg/DFGByteCodeParser.cpp:

(JSC::DFG::ByteCodeParser::inlineCall):

  • heap/MarkedBlock.cpp:

(JSC::MarkedBlock::MarkedBlock):
(JSC::MarkedBlock::Handle::stopAllocating):
(JSC::MarkedBlock::Handle::resumeAllocating):
(JSC::MarkedBlock::aboutToMarkSlow):
(JSC::MarkedBlock::Handle::didConsumeFreeList):

4:41 PM Changeset in webkit [244938] by jiewen_tan@apple.com
  • 7 edits
    1 copy
    1 add in trunk

[WebAuthN] A focused document should be required
https://bugs.webkit.org/show_bug.cgi?id=197543
<rdar://problem/50430989>

Reviewed by Brent Fulgham.

Source/WebCore:

This patch adds a check to see if the invoking document is focused before
calling into WebAuthN. This patch also removes some out-to-dated comments.

Test: http/wpt/webauthn/public-key-credential-unfocused-document.https.html

  • Modules/credentialmanagement/CredentialsContainer.cpp:

(WebCore::CredentialsContainer::get):
(WebCore::CredentialsContainer::isCreate):

  • Modules/webauthn/AuthenticatorCoordinator.cpp:

(WebCore::AuthenticatorCoordinator::create const):
(WebCore::AuthenticatorCoordinator::discoverFromExternalSource const):

LayoutTests:

  • http/wpt/webauthn/public-key-credential-same-origin-with-ancestors.https.html:
  • http/wpt/webauthn/public-key-credential-unfocused-document.https-expected.txt: Added.
  • http/wpt/webauthn/public-key-credential-unfocused-document.https.html: Copied from LayoutTests/http/wpt/webauthn/public-key-credential-same-origin-with-ancestors.https.html.
  • http/wpt/webauthn/resources/last-layer-frame.https.html:
  • http/wpt/webauthn/resources/second-layer-frame.https.html:
4:34 PM Changeset in webkit [244937] by msaboff@apple.com
  • 2 edits in trunk/JSTests

iOS JSC tests frequently exiting with execption after stress/json-stringify-string-builder-overflow.js.no-cjit-validate-phases
https://bugs.webkit.org/show_bug.cgi?id=197586

Reviewed by Keith Miller.

We should only run one config of this test and only when we think we'll have the memory.

  • stress/json-stringify-string-builder-overflow.js:
4:23 PM Changeset in webkit [244936] by Chris Dumez
  • 3 edits in trunk/Source/WebKit

Unresponsive Service Worker processes should get killed
https://bugs.webkit.org/show_bug.cgi?id=197580
<rdar://problem/50458239>

Reviewed by Geoffrey Garen.

Enable the background responsiveness logic for service worker processes so that they get
killed if they becomes unresponsive. The logic was previously not enabling the timer for
service worker processes because it required having a page.

  • UIProcess/BackgroundProcessResponsivenessTimer.cpp:

(WebKit::BackgroundProcessResponsivenessTimer::shouldBeActive const):

  • UIProcess/WebProcessProxy.cpp:

(WebKit::WebProcessProxy::didBecomeUnresponsive):
(WebKit::WebProcessProxy::didFinishLaunching):

4:13 PM Changeset in webkit [244935] by youenn@apple.com
  • 4 edits in trunk

[iOS] set the default maximum camera count to 1 for enumerateDevices
https://bugs.webkit.org/show_bug.cgi?id=197566

Reviewed by Eric Carlson.

Source/WebKit:

The previous number for iOS was 2 as there is a front and a back camera.
Since we are by default setting deviceIds to "", it is better to just have one,
as it might confuse applications to have two devices with the same ID.

Covered by updated test.

  • UIProcess/UserMediaPermissionRequestManagerProxy.cpp:

(WebKit::UserMediaPermissionRequestManagerProxy::computeFilteredDeviceList):

LayoutTests:

  • fast/mediastream/get-user-media-device-id.html:
4:13 PM Changeset in webkit [244934] by dbates@webkit.org
  • 2 edits in trunk/Tools

[lldb-webkit] Support adding pretty-printing for qualified types
https://bugs.webkit.org/show_bug.cgi?id=197518

Reviewed by Darin Adler.

Remove all "::"s from the name of the specified type when computing the name for the Python classes
that will provide summary and synthetic details as LLDB is not happy about registering such classes
when they contain :: in their name.

  • lldb/lldb_webkit.py:

(addSummaryAndSyntheticFormattersForRawBitmaskType):

4:08 PM Changeset in webkit [244933] by Devin Rousso
  • 20 edits
    1 add
    1 delete in trunk/Source

Web Inspector: DOM: rename "low power" to "display composited"
https://bugs.webkit.org/show_bug.cgi?id=197296

Reviewed by Joseph Pecoraro.

Removed specific ChangeLog entries since it is almost entirely mechanical changes.

Source/JavaScriptCore:

  • inspector/protocol/DOM.json:

Source/WebCore:

  • inspector/agents/InspectorDOMAgent.h:
  • inspector/agents/InspectorDOMAgent.cpp:

Source/WebInspectorUI:

  • UserInterface/Protocol/DOMObserver.js:
  • UserInterface/Controllers/DOMManager.js:
  • UserInterface/Controllers/TimelineManager.js:
  • UserInterface/Models/DOMNode.js:
  • UserInterface/Models/MediaInstrument.js:
  • UserInterface/Models/MediaTimelineRecord.js:
  • UserInterface/Models/TimelineRecording.js:
  • UserInterface/Views/DOMEventsBreakdownView.js:
  • UserInterface/Views/DOMEventsBreakdownView.css:
  • UserInterface/Views/NetworkTableContentView.js:
  • UserInterface/Views/NetworkTableContentView.css:
  • UserInterface/Views/TimelineIcons.css:
  • UserInterface/Views/TimelineTabContentView.js:
  • UserInterface/Images/PowerEfficientPlaybackStateChanged.svg: Added.
  • UserInterface/Images/LowPower.svg: Removed.
  • Localizations/en.lproj/localizedStrings.js:
4:04 PM Changeset in webkit [244932] by dbates@webkit.org
  • 23 edits in trunk/Source

Pass KeyboardEvent by reference in more places
https://bugs.webkit.org/show_bug.cgi?id=197480

Reviewed by Wenson Hsieh.

Source/WebCore:

  • editing/Editor.cpp:

(WebCore::Editor::handleKeyboardEvent):
(WebCore::Editor::handleInputMethodKeydown):

  • loader/EmptyClients.cpp:
  • page/EditorClient.h:

Source/WebKit:

  • WebProcess/WebCoreSupport/WebEditorClient.cpp:

(WebKit::WebEditorClient::handleKeyboardEvent):
(WebKit::WebEditorClient::handleInputMethodKeydown):

  • WebProcess/WebCoreSupport/WebEditorClient.h:
  • WebProcess/WebCoreSupport/gtk/WebEditorClientGtk.cpp:

(WebKit::WebEditorClient::handleKeyboardEvent):
(WebKit::WebEditorClient::handleInputMethodKeydown):

  • WebProcess/WebCoreSupport/ios/WebEditorClientIOS.mm:

(WebKit::WebEditorClient::handleKeyboardEvent):
(WebKit::WebEditorClient::handleInputMethodKeydown):

  • WebProcess/WebCoreSupport/mac/WebEditorClientMac.mm:

(WebKit::WebEditorClient::handleKeyboardEvent):
(WebKit::WebEditorClient::handleInputMethodKeydown):

  • WebProcess/WebCoreSupport/wpe/WebEditorClientWPE.cpp:

(WebKit::WebEditorClient::handleKeyboardEvent):
(WebKit::WebEditorClient::handleInputMethodKeydown):

  • WebProcess/WebPage/WebPage.h:
  • WebProcess/WebPage/ios/WebPageIOS.mm:

(WebKit::WebPage::handleEditingKeyboardEvent):

  • WebProcess/WebPage/mac/WebPageMac.mm:

(WebKit::WebPage::handleEditingKeyboardEvent):

  • WebProcess/WebPage/win/WebPageWin.cpp:

(WebKit::WebPage::handleEditingKeyboardEvent):

Source/WebKitLegacy/mac:

  • WebCoreSupport/WebEditorClient.h:
  • WebCoreSupport/WebEditorClient.mm:

(WebEditorClient::handleKeyboardEvent):
(WebEditorClient::handleInputMethodKeydown):

Source/WebKitLegacy/win:

  • WebCoreSupport/WebEditorClient.cpp:

(WebEditorClient::handleKeyboardEvent):
(WebEditorClient::handleInputMethodKeydown):

  • WebCoreSupport/WebEditorClient.h:
  • WebView.cpp:

(WebView::handleEditingKeyboardEvent):

  • WebView.h:
3:43 PM Changeset in webkit [244931] by youenn@apple.com
  • 2 edits in trunk/Source/WebKit

Enable Fetch Keep Alive by default
https://bugs.webkit.org/show_bug.cgi?id=197331

Reviewed by Geoffrey Garen.

  • Shared/WebPreferences.yaml:
3:06 PM Changeset in webkit [244930] by aakash_jain@apple.com
  • 2 edits in trunk/Tools

New EWS: patches on recently added queues listed as #1 for older bugs
https://bugs.webkit.org/show_bug.cgi?id=197496

Reviewed by Lucas Forschler.

  • BuildSlaveSupport/ews-app/ews/views/statusbubble.py:

(StatusBubble._build_bubble):
(StatusBubble._queue_position):

3:06 PM Changeset in webkit [244929] by Alan Coon
  • 7 edits in tags/Safari-608.1.20.0.1/Source/WebCore

Cherry-pick r244922. rdar://problem/50447841

AVFoundation framework isn't always installed
https://bugs.webkit.org/show_bug.cgi?id=197577
<rdar://problem/50447841>

Reviewed by Jer Noble.

Use PAL::isAVFoundationFrameworkAvailable() to check to see if AVFoundation is
installed, not PAL::AVFoundationLibrary().

  • platform/graphics/avfoundation/objc/AVFoundationMIMETypeCache.mm: (WebCore::AVFoundationMIMETypeCache::isAvailable const): (WebCore::AVFoundationMIMETypeCache::loadMIMETypes):
  • platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm: (WebCore::MediaPlayerPrivateAVFoundationObjC::isAvailable): (WebCore::MediaPlayerPrivateAVFoundationObjC::wirelessPlaybackTargetType const): (WebCore::exernalDeviceDisplayNameForPlayer):
  • platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.mm: (WebCore::MediaPlayerPrivateMediaSourceAVFObjC::isAvailable):
  • platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaStreamAVFObjC.mm: (WebCore::MediaPlayerPrivateMediaStreamAVFObjC::isAvailable):
  • platform/graphics/cocoa/HEVCUtilitiesCocoa.mm: (WebCore::validateHEVCParameters):
  • platform/mediastream/mac/AVCaptureDeviceManager.mm: (WebCore::AVCaptureDeviceManager::isAvailable):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244922 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3:03 PM Changeset in webkit [244928] by Alan Coon
  • 7 edits in tags/Safari-608.1.20.0.1/Source

Versioning.

3:03 PM Changeset in webkit [244927] by Chris Dumez
  • 8 edits in trunk/Source/WebCore

[iOS Sim Debug] ASSERTION FAILED The atomic string comes from an other thread! Layout Test imported/w3c/web-platform-tests/workers/WorkerNavigator_appName.htm is a flaky crash
https://bugs.webkit.org/show_bug.cgi?id=197530
<rdar://problem/50448285>

Reviewed by Geoffrey Garen.

The issue is that NavigatorBase::platform() was not thread safe but was called by both Navigator on
the main thread and WorkerNavigator on worker threads.

No new tests, covered by existing tests.

  • page/Navigator.cpp:

(WebCore::Navigator::platform const):

  • page/Navigator.h:
  • page/NavigatorBase.cpp:

(WebCore::NavigatorBase::platform const):

  • page/NavigatorBase.h:

Make NavigatorBase::platform() thread safe.

  • platform/ios/Device.cpp:

(WebCore::deviceName):

  • platform/ios/Device.h:

Make WebCore::deviceName() thread safe.

  • platform/ios/UserAgentIOS.mm:

(WebCore::deviceNameForUserAgent):
Cache value returned by WebCore::deviceName() for performance.

3:01 PM Changeset in webkit [244926] by Chris Dumez
  • 3 edits in trunk/Source/WebCore

Use WeakPtr for JSLazyEventListener::m_originalNode for safety
https://bugs.webkit.org/show_bug.cgi?id=197576
<rdar://problem/24314027>

Reviewed by Alex Christensen.

  • bindings/js/JSLazyEventListener.cpp:

(WebCore::JSLazyEventListener::JSLazyEventListener):
(WebCore::JSLazyEventListener::create):

  • bindings/js/JSLazyEventListener.h:
3:00 PM Changeset in webkit [244925] by Alan Coon
  • 1 copy in tags/Safari-608.1.20.0.1

New tag.

2:59 PM Changeset in webkit [244924] by Devin Rousso
  • 8 edits in trunk/Source/WebKit

REGRESSION(r230367): a "." is prepended before cookies set without an explicit domain
https://bugs.webkit.org/show_bug.cgi?id=197268
<rdar://problem/49767847>

Reviewed by Alex Christensen.

RFC 2965 states "If an explicitly specified value does not start with a dot, the user agent
supplies a leading dot.", so we should only be adding a leading "." if a domain is provided.

In order to not regress r230367, also delete all cookies for the non-"."-prefixed domain.

  • UIProcess/Automation/WebAutomationSession.cpp:

(WebKit::WebAutomationSession::addSingleCookie):
(WebKit::WebAutomationSession::deleteAllCookies):

  • UIProcess/WebCookieManagerProxy.h:
  • UIProcess/WebCookieManagerProxy.cpp:

(WebKit::WebCookieManagerProxy::deleteCookiesForHostnames): Added.
(WebKit::WebCookieManagerProxy::deleteCookiesForHostname): Deleted.

  • UIProcess/API/C/WKCookieManager.cpp:

(WKCookieManagerDeleteCookiesForHostname):

  • NetworkProcess/Cookies/WebCookieManager.messages.in:
  • NetworkProcess/Cookies/WebCookieManager.h:
  • NetworkProcess/Cookies/WebCookieManager.cpp:

(WebKit::WebCookieManager::deleteCookiesForHostnames): Added.
(WebKit::WebCookieManager::deleteCookiesForHostname): Deleted.
Since the underlying NetworkStorageSession accepts a list of domains, extend that support
to allow the IPC to send a list of domains as well, but still keep the C API the same.

2:58 PM Changeset in webkit [244923] by aakash_jain@apple.com
  • 10 edits
    3 adds in trunk/Tools

webkit-patch --no-review upload does not submit patch to New EWS
https://bugs.webkit.org/show_bug.cgi?id=197519
<rdar://problem/50424887>

Reviewed by Lucas Forschler.

  • Scripts/webkitpy/tool/steps/submittoews.py:

(SubmitToEWS.run): Submit to both old and new EWS.

  • Scripts/webkitpy/common/config/urls.py: Added url for new EWS server.
  • Scripts/webkitpy/common/net/ewsserver.py: Added.

(EWSServer._server_url): Method to return server url.
(EWSServer._post_patch_to_ews): Method to post patch to ews.
(EWSServer.submit_to_ews): Method to submit the patch to ews using NetworkTransaction.

  • Scripts/webkitpy/common/net/ewsserver_mock.py: Added Mock EWS Server.
  • Scripts/webkitpy/common/net/ewsserver_unittest.py: Added unit-test for EWS Server.
  • Scripts/webkitpy/common/net/statusserver_mock.py:

(MockStatusServer.submit_to_ews): Updated the log text.

  • Scripts/webkitpy/tool/commands/queues_unittest.py: Updated unit-tests.
  • Scripts/webkitpy/tool/commands/upload_unittest.py: Ditto.
  • Scripts/webkitpy/tool/main.py:

(WebKitPatch.init): Initialize ews_server.

  • Scripts/webkitpy/tool/mocktool.py:

(MockTool.init): Ditto.

2:47 PM Changeset in webkit [244922] by eric.carlson@apple.com
  • 7 edits in trunk/Source/WebCore

AVFoundation framework isn't always installed
https://bugs.webkit.org/show_bug.cgi?id=197577
<rdar://problem/50447841>

Reviewed by Jer Noble.

Use PAL::isAVFoundationFrameworkAvailable() to check to see if AVFoundation is
installed, not PAL::AVFoundationLibrary().

  • platform/graphics/avfoundation/objc/AVFoundationMIMETypeCache.mm:

(WebCore::AVFoundationMIMETypeCache::isAvailable const):
(WebCore::AVFoundationMIMETypeCache::loadMIMETypes):

  • platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:

(WebCore::MediaPlayerPrivateAVFoundationObjC::isAvailable):
(WebCore::MediaPlayerPrivateAVFoundationObjC::wirelessPlaybackTargetType const):
(WebCore::exernalDeviceDisplayNameForPlayer):

  • platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.mm:

(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::isAvailable):

  • platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaStreamAVFObjC.mm:

(WebCore::MediaPlayerPrivateMediaStreamAVFObjC::isAvailable):

  • platform/graphics/cocoa/HEVCUtilitiesCocoa.mm:

(WebCore::validateHEVCParameters):

  • platform/mediastream/mac/AVCaptureDeviceManager.mm:

(WebCore::AVCaptureDeviceManager::isAvailable):

2:24 PM Changeset in webkit [244921] by sihui_liu@apple.com
  • 12 edits
    1 delete in trunk/Source

Add assertion to check whether shm files have maximum FileProtection of CompleteUnlessOpen
https://bugs.webkit.org/show_bug.cgi?id=197390
<rdar://problem/42685773>

Reviewed by Geoffrey Garen.

Source/WebCore:

We have seen crashes about accessing database files after device is locked. We are suspecting this is because
shm files have wrong data protection class, but shm files should not have Complete class protection when it
is created. It is likely the protection class is changed later. Add an assertion to verify our guess. If the
crash signature changes after this patch, we probably need to change database implementation. If it is not, we
have other problem than data protection.

  • platform/sql/SQLiteDatabase.cpp:

(WebCore::SQLiteDatabase::open):

Source/WebKit:

Move data protection check to WebCore so it can be applied to database files.

  • NetworkProcess/cache/NetworkCacheBlobStorage.cpp:

(WebKit::NetworkCache::BlobStorage::add):

  • NetworkProcess/cache/NetworkCacheFileSystem.cpp:

(WebKit::NetworkCache::makeSafeToUseMemoryMapForPath): Deleted.

  • NetworkProcess/cache/NetworkCacheFileSystem.h:
  • NetworkProcess/cache/NetworkCacheFileSystemCocoa.mm: Removed.
  • SourcesCocoa.txt:
  • UIProcess/API/APIContentRuleListStore.cpp:

(API::openAndMapOrCopyContentRuleList):
(API::compiledToFile):

Source/WTF:

  • wtf/FileSystem.cpp:

(WTF::FileSystemImpl::isSafeToUseMemoryMapForPath):
(WTF::FileSystemImpl::makeSafeToUseMemoryMapForPath):

  • wtf/FileSystem.h:
  • wtf/cocoa/FileSystemCocoa.mm:

(WTF::FileSystemImpl::isSafeToUseMemoryMapForPath):
(WTF::FileSystemImpl::makeSafeToUseMemoryMapForPath):

2:22 PM Changeset in webkit [244920] by Chris Dumez
  • 4 edits in trunk

ASSERTION FAILED: [weakThis->m_view window] == weakThis->m_targetWindowForMovePreparation
https://bugs.webkit.org/show_bug.cgi?id=197574
<rdar://problem/50453181>

Reviewed by Tim Horton.

Source/WebKit:

Tweak the assertion so that it holds even if the WebView gets closed because the completion
handler has been called.

  • UIProcess/Cocoa/WebViewImpl.mm:

(WebKit::WebViewImpl::prepareForMoveToWindow):

Tools:

Add API test coverage.

  • TestWebKitAPI/Tests/WebKitCocoa/PrepareForMoveToWindow.mm:

(TEST):

2:03 PM Changeset in webkit [244919] by Basuke Suzuki
  • 15 edits
    2 adds in trunk

[WinCairo] Implement and enable RemoteInspector Server.
https://bugs.webkit.org/show_bug.cgi?id=197432

Reviewed by Ross Kirsling.

.:

Add new build flag USE_INSPECTOR_SOCKET_SERVER to indicate using Socket implementation for RemoteInspector
protocol. Currently PlayStation is the only platform which uses this. WinCairo is the second one.

  • Source/cmake/OptionsPlayStation.cmake:
  • Source/cmake/OptionsWin.cmake:
  • Source/cmake/OptionsWinCairo.cmake:

Source/JavaScriptCore:

Implement Windows implementation for Socket Backend of RemoteInspector and enable it on WinCairo
for experimental feature.

Also add listener interface for connection between RemoteInspector and RemoteInspectorServer
for flexible configuration.

  • PlatformWin.cmake:
  • inspector/remote/RemoteInspector.h:
  • inspector/remote/socket/RemoteInspectorConnectionClient.h:

(Inspector::RemoteInspectorConnectionClient::didAccept):

  • inspector/remote/socket/RemoteInspectorServer.cpp:

(Inspector::RemoteInspectorServer::connect):
(Inspector::RemoteInspectorServer::listenForTargets):
(Inspector::RemoteInspectorServer::didAccept):
(Inspector::RemoteInspectorServer::dispatchMap):
(Inspector::RemoteInspectorServer::start):
(Inspector::RemoteInspectorServer::addServerConnection): Deleted.

  • inspector/remote/socket/RemoteInspectorServer.h:

(Inspector::RemoteInspectorServer::RemoteInspectorServer):

  • inspector/remote/socket/RemoteInspectorSocket.cpp:

(Inspector::RemoteInspector::RemoteInspector):
(Inspector::RemoteInspector::dispatchMap):
(Inspector::RemoteInspector::start):
(Inspector::RemoteInspector::stopInternal):
(Inspector::RemoteInspector::setServerPort):

  • inspector/remote/socket/RemoteInspectorSocket.h:
  • inspector/remote/socket/RemoteInspectorSocketEndpoint.cpp:

(Inspector::RemoteInspectorSocketEndpoint::listenInet):
(Inspector::RemoteInspectorSocketEndpoint::getPort const):
(Inspector::RemoteInspectorSocketEndpoint::acceptInetSocketIfEnabled):

  • inspector/remote/socket/RemoteInspectorSocketEndpoint.h:
  • inspector/remote/socket/posix/RemoteInspectorSocketPOSIX.cpp:

(Inspector::Socket::init): Added.
(Inspector::Socket::listen): Signature changed.
(Inspector::Socket::getPort): Added.

  • inspector/remote/socket/win/RemoteInspectorSocketWin.cpp: Added.

(Inspector::Socket::init):
(Inspector::Socket::Socket::Socket):
(Inspector::Socket::Socket::~Socket):
(Inspector::Socket::Socket::close):
(Inspector::Socket::Socket::operator PlatformSocketType const):
(Inspector::Socket::Socket::operator bool const):
(Inspector::Socket::Socket::leak):
(Inspector::Socket::Socket::create):
(Inspector::Socket::setOpt):
(Inspector::Socket::setOptEnabled):
(Inspector::Socket::enableOpt):
(Inspector::Socket::connectTo):
(Inspector::Socket::bindAndListen):
(Inspector::Socket::connect):
(Inspector::Socket::listen):
(Inspector::Socket::accept):
(Inspector::Socket::createPair):
(Inspector::Socket::setup):
(Inspector::Socket::isValid):
(Inspector::Socket::isListening):
(Inspector::Socket::getPort):
(Inspector::Socket::read):
(Inspector::Socket::write):
(Inspector::Socket::close):
(Inspector::Socket::preparePolling):
(Inspector::Socket::poll):
(Inspector::Socket::isReadable):
(Inspector::Socket::isWritable):
(Inspector::Socket::markWaitingWritable):
(Inspector::Socket::clearWaitingWritable):

1:36 PM Changeset in webkit [244918] by youenn@apple.com
  • 7 edits
    3 adds in trunk

Cache.add and Cache.addAll should compute a correct response body size
https://bugs.webkit.org/show_bug.cgi?id=197464

Reviewed by Chris Dumez.

Source/WebCore:

Compute the response body size as we do for regular Cache.put

Test: http/wpt/cache-storage/cache-quota-add.any.html

  • Modules/cache/CacheStorageConnection.cpp:

(WebCore::CacheStorageConnection::computeRecordBodySize):

  • Modules/cache/CacheStorageConnection.h:
  • Modules/cache/DOMCache.cpp:

(WebCore::FetchTasksHandler::addResponseBody):
(WebCore::DOMCache::addAll):
Compute the response body size requires getting access to the connection.
'this' is added to the lambda which is fine since taskHandler keeps a
Ref to 'this' in its completion handler.
(WebCore::DOMCache::toConnectionRecord):

  • Modules/fetch/FetchResponse.h:

LayoutTests:

  • http/wpt/cache-storage/cache-quota-add.any-expected.txt: Added.
  • http/wpt/cache-storage/cache-quota-add.any.html: Added.
  • http/wpt/cache-storage/cache-quota-add.any.js: Added.
12:42 PM Changeset in webkit [244917] by Brent Fulgham
  • 2 edits in trunk/Source/WebKit

Use more efficient path resolution logic
https://bugs.webkit.org/show_bug.cgi?id=197389
<rdar://problem/50268491>

Reviewed by Maciej Stachowiak.

The code in SandboxExtensionsCocoa.mm 'resolveSymlinksInPath' is pretty inefficient, and tries to reproduce (badly)
logic that is already provided by the operating system.

To make matters worse, 'resolvePathForSandboxExtension' was effectively performing the work of fully resolving
symlinks twice, since NSString's 'stringByStandardizingPath' method does some of this already.

Instead, we should just use NSString's 'stringByResolvingSymlinksInPath', which does the symlink resolution
using more efficient logic than our 'resolveSymlinksInPath' code.

  • Shared/Cocoa/SandboxExtensionCocoa.mm:

(WebKit::resolveSymlinksInPath): Removed.
(WebKit::resolvePathForSandboxExtension): Remove redundant call to 'resolveSymlinksInPath', and switches from
'stringByStandardizingPath' to 'stringByResolvingSymlinksInPath', which can take the place of both calls.
(WebKit::stringByResolvingSymlinksInPath): Switch to call 'stringByResolvingSymlinksInPath'.

12:35 PM Changeset in webkit [244916] by dino@apple.com
  • 2 edits in trunk/Source/WebKit

Need additional UIPreviewAction information in WKImagePreviewViewController
https://bugs.webkit.org/show_bug.cgi?id=197540
<rdar://problem/50430510>

Reviewed by Wenson Hsieh.

Include a WebKitAdditions file to provide a macro for
additional UIPreviewAction information.

While here, take the opportunity to move from the deprecated
UIViewControllerPreviewAction to the newer UIPreviewAction.

  • UIProcess/WKImagePreviewViewController.mm:

(-[WKImagePreviewViewController previewActionItems]):

11:54 AM Changeset in webkit [244915] by ysuzuki@apple.com
  • 33 edits
    4 adds in trunk

[JSC] Generator CodeBlock generation should be idempotent
https://bugs.webkit.org/show_bug.cgi?id=197552

Reviewed by Keith Miller.

JSTests:

Add complex.yaml, which controls how to run JSC shell more.
We split test files into two to run macro task between them which allows debugger to be attached to VM.

  • complex.yaml: Added.
  • complex/generator-regeneration-after.js: Added.
  • complex/generator-regeneration.js: Added.

(gen):

Source/JavaScriptCore:

ES6 Generator saves and resumes the current execution state. Since ES6 generator can save the execution state at expression
granularity (not statement granularity), the saved state involves locals. But if the underlying CodeBlock is jettisoned and
recompiled with different code generation option (like, debugger, type profiler etc.), the generated instructions can be largely
different and it does not have the same state previously used. If we resume the previously created generator with the newly
generator function, resuming is messed up.

function* gen () { ... }
var g = gen();
g.next();

CodeBlock is destroyed & Debugger is enabled.

g.next();

In this patch,

  1. In generatorification, we use index Identifier (localN => Identifier("N")) instead of private symbols to generate the same instructions every time we regenerate the CodeBlock.
  1. We decouple the options which can affect on the generated code (Debugger, TypeProfiler, ControlFlowProfiler) from the BytecodeGenerator, and pass them as a parameter, OptionSet<CodeGeneratorMode>.
  1. Generator ScriptExecutable remembers the previous CodeGeneratorMode and reuses this parameter to regenerate CodeBlock. It means that, even if the debugger is enabled, previously created generators are not debuggable. But newly created generators are debuggable.
  • bytecode/BytecodeGeneratorification.cpp:

(JSC::BytecodeGeneratorification::storageForGeneratorLocal):
(JSC::BytecodeGeneratorification::run):

  • bytecode/CodeBlock.cpp:

(JSC::CodeBlock::finishCreation):
(JSC::CodeBlock::setConstantRegisters):

  • bytecode/UnlinkedCodeBlock.cpp:

(JSC::UnlinkedCodeBlock::UnlinkedCodeBlock):

  • bytecode/UnlinkedCodeBlock.h:

(JSC::UnlinkedCodeBlock::wasCompiledWithDebuggingOpcodes const):
(JSC::UnlinkedCodeBlock::wasCompiledWithTypeProfilerOpcodes const):
(JSC::UnlinkedCodeBlock::wasCompiledWithControlFlowProfilerOpcodes const):
(JSC::UnlinkedCodeBlock::codeGenerationMode const):

  • bytecode/UnlinkedEvalCodeBlock.h:
  • bytecode/UnlinkedFunctionCodeBlock.h:
  • bytecode/UnlinkedFunctionExecutable.cpp:

(JSC::generateUnlinkedFunctionCodeBlock):
(JSC::UnlinkedFunctionExecutable::fromGlobalCode):
(JSC::UnlinkedFunctionExecutable::unlinkedCodeBlockFor):

  • bytecode/UnlinkedFunctionExecutable.h:
  • bytecode/UnlinkedGlobalCodeBlock.h:

(JSC::UnlinkedGlobalCodeBlock::UnlinkedGlobalCodeBlock):

  • bytecode/UnlinkedModuleProgramCodeBlock.h:
  • bytecode/UnlinkedProgramCodeBlock.h:
  • bytecompiler/BytecodeGenerator.cpp:

(JSC::BytecodeGenerator::BytecodeGenerator):
(JSC::BytecodeGenerator::emitTypeProfilerExpressionInfo):
(JSC::BytecodeGenerator::emitProfileType):
(JSC::BytecodeGenerator::emitProfileControlFlow):
(JSC::BytecodeGenerator::pushLexicalScopeInternal):
(JSC::BytecodeGenerator::popLexicalScopeInternal):
(JSC::BytecodeGenerator::prepareLexicalScopeForNextForLoopIteration):
(JSC::BytecodeGenerator::emitCall):
(JSC::BytecodeGenerator::emitCallVarargs):
(JSC::BytecodeGenerator::emitLogShadowChickenPrologueIfNecessary):
(JSC::BytecodeGenerator::emitLogShadowChickenTailIfNecessary):
(JSC::BytecodeGenerator::emitDebugHook):

  • bytecompiler/BytecodeGenerator.h:

(JSC::BytecodeGenerator::generate):
(JSC::BytecodeGenerator::shouldEmitDebugHooks const):
(JSC::BytecodeGenerator::shouldEmitTypeProfilerHooks const):
(JSC::BytecodeGenerator::shouldEmitControlFlowProfilerHooks const):

  • bytecompiler/NodesCodegen.cpp:

(JSC::PrefixNode::emitResolve):
(JSC::EmptyVarExpression::emitBytecode):
(JSC::ReturnNode::emitBytecode):
(JSC::FunctionNode::emitBytecode):

  • parser/ParserModes.h:

(): Deleted.

  • parser/SourceCodeKey.h:

(JSC::SourceCodeFlags::SourceCodeFlags):
(JSC::SourceCodeKey::SourceCodeKey):

  • runtime/CachedTypes.cpp:

(JSC::CachedCodeBlock::isClassContext const):
(JSC::CachedCodeBlock::codeGenerationMode const):
(JSC::UnlinkedCodeBlock::UnlinkedCodeBlock):
(JSC::CachedCodeBlock<CodeBlockType>::encode):
(JSC::CachedCodeBlock::wasCompiledWithDebuggingOpcodes const): Deleted.

  • runtime/CodeCache.cpp:

(JSC::CodeCache::getUnlinkedGlobalCodeBlock):
(JSC::CodeCache::getUnlinkedProgramCodeBlock):
(JSC::CodeCache::getUnlinkedEvalCodeBlock):
(JSC::CodeCache::getUnlinkedModuleProgramCodeBlock):
(JSC::CodeCache::getUnlinkedGlobalFunctionExecutable):
(JSC::generateUnlinkedCodeBlockForFunctions):
(JSC::sourceCodeKeyForSerializedBytecode):
(JSC::sourceCodeKeyForSerializedProgram):
(JSC::sourceCodeKeyForSerializedModule):
(JSC::serializeBytecode):

  • runtime/CodeCache.h:

(JSC::generateUnlinkedCodeBlockImpl):
(JSC::generateUnlinkedCodeBlock):

  • runtime/Completion.cpp:

(JSC::generateProgramBytecode):
(JSC::generateModuleBytecode):

  • runtime/DirectEvalExecutable.cpp:

(JSC::DirectEvalExecutable::create):

  • runtime/IndirectEvalExecutable.cpp:

(JSC::IndirectEvalExecutable::create):

  • runtime/JSGlobalObject.h:

(JSC::JSGlobalObject::defaultCodeGenerationMode const):

  • runtime/ModuleProgramExecutable.cpp:

(JSC::ModuleProgramExecutable::create):

  • runtime/ProgramExecutable.cpp:

(JSC::ProgramExecutable::initializeGlobalProperties):

  • runtime/ScriptExecutable.cpp:

(JSC::ScriptExecutable::ScriptExecutable):
(JSC::ScriptExecutable::newCodeBlockFor):

  • runtime/ScriptExecutable.h:
  • tools/JSDollarVM.cpp:

(JSC::changeDebuggerModeWhenIdle):
(JSC::functionEnableDebuggerModeWhenIdle):
(JSC::functionDisableDebuggerModeWhenIdle):

Tools:

  • Scripts/run-javascriptcore-tests:

(runJSCStressTests):

  • Scripts/run-jsc-stress-tests:
11:52 AM Changeset in webkit [244914] by dino@apple.com
  • 3 edits in trunk/Source/WebKit

Provide UIImages for element actions
https://bugs.webkit.org/show_bug.cgi?id=197539
<rdar://problem/50430363>

Reviewed by Wenson Hsieh.

Use an additions include to export a function that
relates a UIImage to a _WKElementAction.

  • UIProcess/API/Cocoa/_WKElementAction.h:
  • UIProcess/API/Cocoa/_WKElementAction.mm:

(_WKUIImageForElementActionType):

11:31 AM Changeset in webkit [244913] by commit-queue@webkit.org
  • 4 edits in trunk/Source/WebCore

[Cairo] Improve ShadowBlur performance using tiling optimization
https://bugs.webkit.org/show_bug.cgi?id=197308
Patch by Tomoki Imai <Tomoki Imai> on 2019-05-03
Reviewed by Žan Doberšek.

Enable tiling tiling-based optimization for drawRectShadow() and drawInsetShadow().
Since r228776, cairo ports doesn't have tiling-based optimization.

For AppleWin, this patch refactors code and it shares almost same code as cairo port.
Only the difference is that AppleWin uses ScratchBuffer, but cairo ports doesn't.
This should avoid a performance regression for AppleWin.

No new tests, covered by existing tests.

  • platform/graphics/ShadowBlur.cpp:

(WebCore::calculateLobes):
Fix stylecheck errors

(WebCore::ShadowBlur::blurLayerImage):
Fix stylecheck errors

(WebCore::ShadowBlur::calculateLayerBoundingRect):
We don't use position of m_sourceRect, so change the type to FloatSize.

(WebCore::ShadowBlur::drawShadowBuffer):
Use m_layerSize instead of m_shadowedResultSize to fillRect, as m_layerSize is always smaller than m_shadowedResultSize.
It's because in m_layerSize is equal to m_shadowedResultSize if it's not clipped.
Clipping doesn't increase size of m_layerSize, so m_layerSize is always smaller than or equal to m_shadowedResultSize.

(WebCore::ShadowBlur::templateSize const):
Fix stylecheck errors

(WebCore::ShadowBlur::drawRectShadow):
(WebCore::ShadowBlur::drawInsetShadow):
(WebCore::ShadowBlur::drawRectShadowWithoutTiling):
(WebCore::ShadowBlur::drawInsetShadowWithoutTiling):
(WebCore::ShadowBlur::drawRectShadowWithTiling):
(WebCore::ShadowBlur::drawInsetShadowWithTiling):
Incorporate tile-based drawing.
To accomplish it, this patch abstracts GraphicsContext::drawImageBuffer to ShadowBlur::DrawImageCallback,
GraphicsContext::fillRect to ShadowBlur::FillRectCallback, drawing rect with hole to ShadowBlur::FillRectWithHoleCallback.

Variants which takes GraphicsContext as parameter now just calls another drawRectShadow.

(WebCore::ShadowBlur::drawLayerPieces):
Instead of graphicsContext.drawImageBuffer, call corresponding callback.

(WebCore::ShadowBlur::drawLayerPiecesAndFillCenter):
This function calls drawLayerPieces and fill center for outer shadow.
Drawing outer shadow requires another callback for graphicsContext.fillRect.

(WebCore::ShadowBlur::drawShadowLayer):
Use m_layerSize instead of m_shadowedResultSize to fillRect,
as m_layerSize is always smaller than m_shadowedResultSize.

  • platform/graphics/ShadowBlur.h:

Rename m_sourceRect to m_shadowedResultSize, and change it to FloatSize from FloatRect.
Remove GraphicsContext usage as much as possible and replace them by corresponding callbacks.

  • platform/graphics/cairo/CairoOperations.cpp:

(WebCore::Cairo::drawShadowImage):
This function corresponds to ShadowBlur::DrawImageCallback.

(WebCore::Cairo::fillShadowBuffer):
Erase sourceRect, as it's always bigger than layerSize.

(WebCore::Cairo::drawPathShadow):
(WebCore::Cairo::drawGlyphsShadow):
Erase unused parameter.

(WebCore::Cairo::fillRect):
(WebCore::Cairo::fillRoundedRect):
(WebCore::Cairo::fillRectWithRoundedHole):
For tile-based optimization, add extra arguments to drawRectShadow.

(WebCore::Cairo::drawSurface):
Erase unused parameter.

11:18 AM Changeset in webkit [244912] by Antti Koivisto
  • 4 edits in trunk/Source/WebCore

Add a quirk to make youtube navigation bar scrollable without mouse hover on iOS
https://bugs.webkit.org/show_bug.cgi?id=197555
<rdar://problem/49582231>

Reviewed by Brent Fulgham.

  • css/StyleResolver.cpp:

(WebCore::StyleResolver::adjustRenderStyleForSiteSpecificQuirks):

Turn 'overflow:hidden' into 'overflow:auto' on element with id="guide-inner-content".

  • page/Quirks.cpp:

(WebCore::Quirks::needsYouTubeOverflowScrollQuirk const):

  • page/Quirks.h:
10:34 AM Changeset in webkit [244911] by commit-queue@webkit.org
  • 4 edits in trunk/Source/WebKit

Expose full screen controller SPI to check if full screen is open and close full screen
https://bugs.webkit.org/show_bug.cgi?id=197263

Patch by Jay Mulani <jmulani@apple.com> on 2019-05-03
Reviewed by Jer Noble.

  • UIProcess/API/Cocoa/WKWebView.mm:

(-[WKWebView closeFullScreenWindowController]):
(-[WKWebView fullScreenWindowController]):

  • UIProcess/API/Cocoa/WKWebViewInternal.h:
  • UIProcess/API/Cocoa/WKWebViewPrivate.h:
10:32 AM Changeset in webkit [244910] by youenn@apple.com
  • 8 edits in trunk

LayoutTest imported/w3c/web-platform-tests/xhr/event-upload-progress-crossorigin.htm is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=182849
<rdar://problem/50353372>

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

Modernize the test and add an assertion that the body is transmitted after a redirection.

  • web-platform-tests/xhr/event-upload-progress-crossorigin.htm:

Source/WebKit:

In case of restarting the load after a redirection, which now happens in case
of following cross origin redirections with same-origin credentials, make sure to
update the request as would do NetworkLoad.
This is in particular important to preserve the request body.

Covered by WPT updated test.

  • NetworkProcess/NetworkLoad.cpp:

(WebKit::updateRequest):
(WebKit::NetworkLoad::updateRequestAfterRedirection const):
(WebKit::NetworkLoad::continueWillSendRequest):

  • NetworkProcess/NetworkLoad.h:
  • NetworkProcess/NetworkResourceLoader.cpp:

(WebKit::NetworkResourceLoader::continueWillSendRequest):

LayoutTests:

  • platform/mac-wk2/TestExpectations:
9:39 AM Changeset in webkit [244909] by youenn@apple.com
  • 2 edits in trunk/Source/ThirdParty/libwebrtc

Do not require log_to_stderr for WebRTC logging through WebKit
https://bugs.webkit.org/show_bug.cgi?id=197560

Reviewed by Eric Carlson.

  • Source/webrtc/rtc_base/logging.cc:
9:36 AM Changeset in webkit [244908] by Devin Rousso
  • 26 edits
    4 adds in trunk

Web Inspector: Record actions performed on WebGL2RenderingContext
https://bugs.webkit.org/show_bug.cgi?id=176008
<rdar://problem/34213884>

Reviewed by Joseph Pecoraro.

Source/JavaScriptCore:

  • inspector/protocol/Recording.json:
  • inspector/scripts/codegen/generator.py:

Add canvas-webgl2 as a Type.

Source/WebCore:

Tests: inspector/canvas/recording-webgl2.html

inspector/canvas/recording-webgl2-snapshots.html

  • html/canvas/WebGL2RenderingContext.idl:
  • bindings/js/CallTracerTypes.h:
  • inspector/RecordingSwizzleTypes.h:
  • inspector/InspectorCanvas.h:
  • inspector/InspectorCanvas.cpp:

(WebCore::InspectorCanvas::canvasChanged): Added.
(WebCore::InspectorCanvas::resetRecordingData):
(WebCore::shouldSnapshotWebGL2Action): Added.
(WebCore::InspectorCanvas::recordAction):
(WebCore::InspectorCanvas::releaseObjectForRecording):
(WebCore::InspectorCanvas::appendActionSnapshotIfNeeded):
(WebCore::InspectorCanvas::buildAction):

  • inspector/agents/InspectorCanvasAgent.h:
  • inspector/agents/InspectorCanvasAgent.cpp:

(WebCore::InspectorCanvasAgent::canvasChanged): Added.
(WebCore::InspectorCanvasAgent::startRecording):
Provide an actual implementation of CanvasObserver::canvasChanged since it can be used to
determine whether or not an action needs a snapshot.

  • page/PageConsoleClient.cpp:

(WebCore::canvasRenderingContext):

Source/WebInspectorUI:

  • UserInterface/Models/Recording.js:

(WI.Recording.fromPayload):
(WI.Recording.displayNameForSwizzleType):
(WI.Recording.prototype.async swizzle):
(WI.Recording.prototype.createContext):

  • UserInterface/Models/RecordingAction.js:

(WI.RecordingAction.constantNameForParameter):
(WI.RecordingAction._prototypeForType):
(WI.RecordingAction.prototype.async swizzle):
Treat all actions with a snapshot as being a "visible" action.

  • UserInterface/Views/CanvasTabContentView.css:

(.content-view.tab.canvas .navigation-bar > .item .canvas:matches(.webgl, .webgl2, .webgpu, .webmetal) .icon): Added.
(.content-view.tab.canvas .navigation-bar > .item .canvas.webgl): Deleted.

  • UserInterface/Views/CanvasContentView.js:

(WI.CanvasContentView.prototype.initialLayout):

  • UserInterface/Views/CanvasSidebarPanel.js:

(WI.CanvasSidebarPanel.prototype.shown):
(WI.CanvasSidebarPanel.prototype._canvasChanged):
(WI.CanvasSidebarPanel.prototype._updateRecordNavigationItem):

  • UserInterface/Views/RecordingContentView.js:

(WI.RecordingContentView):
(WI.RecordingContentView.prototype.get navigationItems):
(WI.RecordingContentView.prototype.shown):
(WI.RecordingContentView.prototype.initialLayout):

  • UserInterface/Views/RecordingActionTreeElement.js:

(WI.RecordingActionTreeElement._generateDOM.createParameterElement):

LayoutTests:

  • inspector/canvas/recording-webgl2.html: Added.
  • inspector/canvas/recording-webgl2-expected.txt: Added.
  • inspector/canvas/recording-webgl2-snapshots.html: Added.
  • inspector/canvas/recording-webgl2-snapshots-expected.txt: Added.
  • TestExpectations:
  • platform/gtk/TestExpectations:
  • platform/mac/TestExpectations:
  • platform/win/TestExpectations:
  • platform/wincairo/TestExpectations:
7:32 AM Changeset in webkit [244907] by commit-queue@webkit.org
  • 18 edits in trunk

Unreviewed, rolling out r244881.
https://bugs.webkit.org/show_bug.cgi?id=197559

Breaks compilation of jsconly on linux, breaking compilation
for jsc-i386-ews, jsc-mips-ews and jsc-armv7-ews (Requested by
guijemont on #webkit).

Reverted changeset:

"[CMake] Refactor WEBKIT_MAKE_FORWARDING_HEADERS into
WEBKIT_COPY_FILES"
https://bugs.webkit.org/show_bug.cgi?id=197174
https://trac.webkit.org/changeset/244881

6:19 AM Changeset in webkit [244906] by commit-queue@webkit.org
  • 8 edits
    2 copies
    2 adds in trunk

Resolve the percentage values of inset properties against proper box.
https://bugs.webkit.org/show_bug.cgi?id=189549

Patch by Joonghun Park <jh718.park@samsung.com> on 2019-05-03
Reviewed by Antti Koivisto.

Before this CL, sticky element's layout was executed relative to
a box's overflow container,
but the value returned by getComputedStyle was resolved against
its containing block.

So, the computed value and the actual value used in layout
was different before this change.

LayoutTests/imported/w3c:

  • web-platform-tests/css/cssom/getComputedStyle-insets-sticky-container-for-abspos-expected.txt: Copied from LayoutTests/imported/w3c/web-platform-tests/css/cssom/getComputedStyle-insets-sticky-expected.txt.
  • web-platform-tests/css/cssom/getComputedStyle-insets-sticky-container-for-abspos.html: Copied from LayoutTests/imported/w3c/web-platform-tests/css/cssom/getComputedStyle-insets-sticky.html.
  • web-platform-tests/css/cssom/getComputedStyle-insets-sticky-expected.txt:
  • web-platform-tests/css/cssom/getComputedStyle-insets-sticky.html:
  • web-platform-tests/css/cssom/getComputedStyle-sticky-pos-percent-expected.txt: Added.
  • web-platform-tests/css/cssom/getComputedStyle-sticky-pos-percent.html: Added.
  • web-platform-tests/css/cssom/support/getComputedStyle-insets.js:

(runTestsWithWM):

Source/WebCore:

Tests: imported/w3c/web-platform-tests/css/cssom/getComputedStyle-insets-sticky-container-for-abspos.html

imported/w3c/web-platform-tests/css/cssom/getComputedStyle-sticky-pos-percent.html

  • css/CSSComputedStyleDeclaration.cpp:

(WebCore::positionOffsetValue):

  • rendering/RenderBox.cpp:

(WebCore::RenderBox::enclosingScrollportBox const):

  • rendering/RenderBox.h:
1:42 AM Changeset in webkit [244905] by zandobersek@gmail.com
  • 2 edits in trunk/Tools

Unreviewed WPE build fix after r244898.

  • WebKitTestRunner/wpe/TestControllerWPE.cpp:

(WTR::TestController::abortModal): Add the missing (no-op) definition.

May 2, 2019:

11:48 PM Changeset in webkit [244904] by Antti Koivisto
  • 5 edits in trunk/Source/WebCore

Add a quirk to make gmail navigation bar scrollable without mouse hover on iOS
https://bugs.webkit.org/show_bug.cgi?id=197529
<rdar://problem/49403416>

Reviewed by Simon Fraser.

  • css/StyleResolver.cpp:

(WebCore::StyleResolver::adjustRenderStyle):
(WebCore::StyleResolver::adjustRenderStyleForSiteSpecificQuirks):

Turn 'overflow:hidden' to 'overflow:auto' on element with role="navigation".
This should be both reasonably targeted and robust.

  • css/StyleResolver.h:
  • page/Quirks.cpp:

(WebCore::Quirks::needsGMailOverflowScrollQuirk const):

  • page/Quirks.h:
11:23 PM Changeset in webkit [244903] by commit-queue@webkit.org
  • 4 edits in trunk

Mark U+01C0 as a homograph of U+006C
https://bugs.webkit.org/show_bug.cgi?id=197526
<rdar://problem/50301904>

Patch by Alex Christensen <achristensen@webkit.org> on 2019-05-02
Reviewed by Tim Horton.

Source/WTF:

  • wtf/URLHelpers.cpp:

(WTF::URLHelpers::isLookalikeCharacter):

Tools:

  • TestWebKitAPI/Tests/WTF/cocoa/URLExtras.mm:

(TestWebKitAPI::TEST):

11:17 PM Changeset in webkit [244902] by Nikita Vasilyev
  • 2 edits in trunk/Source/WebInspectorUI

Web Inspector: Dark Mode: borders between section should be lighter than background
https://bugs.webkit.org/show_bug.cgi?id=197551

Reviewed by Joseph Pecoraro.

  • UserInterface/Views/DetailsSection.css:

(.details-section .details-section):

10:02 PM Changeset in webkit [244901] by rniwa@webkit.org
  • 2 edits in trunk/LayoutTests

Make focusing-element-with-tabindex-by-tap-or-click.html more robust on iOS
https://bugs.webkit.org/show_bug.cgi?id=197548

Reviewed by Wenson Hsieh.

There was a race between the time the editable element was focused and the keyboard was brought up
to when the next non-editable element was focused. Made the test explicitly wait for the keyboard
to come up and go down between test cases to make it more robust.

  • fast/events/focusing-element-with-tabindex-by-tap-or-click.html:
10:01 PM Changeset in webkit [244900] by Simon Fraser
  • 6 edits in trunk/Source/WebCore

Add logging for RenderLayer clip rects
https://bugs.webkit.org/show_bug.cgi?id=197547

Reviewed by Zalan Bujtas.

Add a ClipRects log channel, and stream output for ClipRect and ClipRects.

  • platform/Logging.h:
  • rendering/ClipRect.cpp:

(WebCore::operator<<):

  • rendering/ClipRect.h:
  • rendering/RenderLayer.cpp:

(WebCore::operator<<):
(WebCore::RenderLayer::calculateClipRects const):

  • rendering/RenderLayer.h:
9:28 PM Changeset in webkit [244899] by youenn@apple.com
  • 11 edits in trunk/Source/WebCore

Make AudioContext::scriptExecutionContext() private
https://bugs.webkit.org/show_bug.cgi?id=197512

Reviewed by Eric Carlson.

Refactor code to make audio nodes not rely on AudioContext::scriptExecutionContext.
Instead, let AudioContext provide the necessary API for its nodes.
Covered by existing tests.

  • Modules/webaudio/AudioBufferSourceNode.cpp:

(WebCore::AudioBufferSourceNode::looping):
(WebCore::AudioBufferSourceNode::setLooping):

  • Modules/webaudio/AudioContext.cpp:

(WebCore::AudioContext::postTask):
(WebCore::AudioContext::origin const):
(WebCore::AudioContext::addConsoleMessage):

  • Modules/webaudio/AudioContext.h:

(WebCore::AudioContext::isStopped const):

  • Modules/webaudio/AudioNode.cpp:

(WebCore::AudioNode::scriptExecutionContext const):

  • Modules/webaudio/AudioNode.h:
  • Modules/webaudio/AudioScheduledSourceNode.cpp:

(WebCore::AudioScheduledSourceNode::finish):

  • Modules/webaudio/DefaultAudioDestinationNode.cpp:

(WebCore::DefaultAudioDestinationNode::resume):
(WebCore::DefaultAudioDestinationNode::suspend):
(WebCore::DefaultAudioDestinationNode::close):

  • Modules/webaudio/MediaElementAudioSourceNode.cpp:

(WebCore::MediaElementAudioSourceNode::wouldTaintOrigin):

  • Modules/webaudio/MediaStreamAudioDestinationNode.cpp:

(WebCore::MediaStreamAudioDestinationNode::MediaStreamAudioDestinationNode):

  • Modules/webaudio/ScriptProcessorNode.cpp:

(WebCore::ScriptProcessorNode::fireProcessEvent):

6:19 PM Changeset in webkit [244898] by Chris Dumez
  • 12 edits
    2 adds in trunk

Add test coverage for <rdar://problem/49731231>
https://bugs.webkit.org/show_bug.cgi?id=196730
<rdar://problem/49731231>

Reviewed by Ryosuke Niwa.

Tools:

Add support for testRunner.abortModal() in macOS's WebKitTestRunner in order
to match DumpRenderTree.

  • WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
  • WebKitTestRunner/InjectedBundle/TestRunner.cpp:

(WTR::TestRunner::abortModal):

  • WebKitTestRunner/InjectedBundle/TestRunner.h:
  • WebKitTestRunner/TestController.cpp:

(WTR::TestController::runModal):

  • WebKitTestRunner/TestController.h:
  • WebKitTestRunner/TestInvocation.cpp:

(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):

  • WebKitTestRunner/gtk/TestControllerGtk.cpp:

(WTR::TestController::abortModal):

  • WebKitTestRunner/ios/TestControllerIOS.mm:

(WTR::TestController::abortModal):

  • WebKitTestRunner/mac/TestControllerMac.mm:

(WTR::TestController::abortModal):

  • WebKitTestRunner/win/TestControllerWin.cpp:

(WTR::TestController::abortModal):

LayoutTests:

  • http/tests/security/showModalDialog-sync-cross-origin-page-load-expected.txt: Added.
  • http/tests/security/showModalDialog-sync-cross-origin-page-load.html: Added.
6:05 PM Changeset in webkit [244897] by Wenson Hsieh
  • 12 edits in trunk/Source/WebKit

REGRESSION: Layout test editing/selection/ios/selection-after-changing-text-with-callout-menu.html is failing
https://bugs.webkit.org/show_bug.cgi?id=197532
<rdar://problem/50177144>

Reviewed by Ryosuke Niwa.

Fixes layout tests that began failing after r244546. See below for details.

  • UIProcess/RemoteLayerTree/RemoteLayerTreeDrawingAreaProxy.mm:

(WebKit::RemoteLayerTreeDrawingAreaProxy::commitLayerTree):

Partially reverts a change in r244546, after which we commit the layer tree and scroll before updating the
page's editor state. The purpose of this change was to ensure that UI process-side element focus scrolling logic
would not conflict with web-process-driven scrolling logic.

Instead, we split the existing logic in WebPageProxy::editorStateChanged into two pieces: one that updates the
editor state (by setting m_editorState), and a second that dispatches updates to PageClient when the first
editor state is received after focusing an element. During a layer tree commit in the UI process, we first
update the editor state, then commit the layer tree and apply scroll position changes, and finally scroll to
reveal the focused element if necessary.

When an editor state is delivered to the UI process in an out-of-band update (i.e. not in a layer tree commit),
simply dispatch the initial editor state for a focused element immediately.

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::scheduleFullEditorStateUpdate):

Add an IPC message to schedule an editor state update in the next remote layer tree commit. See below for more
details.

(WebKit::WebPageProxy::editorStateChanged):
(WebKit::WebPageProxy::dispatchDidReceiveEditorStateAfterFocus):

  • UIProcess/WebPageProxy.h:
  • UIProcess/gtk/WebPageProxyGtk.cpp:

(WebKit::WebPageProxy::updateEditorState):
(WebKit::WebPageProxy::editorStateChanged): Deleted.

  • UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView willFinishIgnoringCalloutBarFadeAfterPerformingAction]):

Additionally ensure that an editor state update is scheduled. This addresses a potential source of flakiness in
the layout test editing/selection/ios/selection-after-changing-text-with-callout-menu.html, where an editor
state update may only be scheduled after the next layout timer fires (this is the case in custom callout menu
actions that change the DOM but do not otherwise trigger any editing commands).

In the problematic scenario, the client could make a change that triggers layout soon; but before the layout
timer fires, the timer for the next remote layer tree commit could fire, such that the next layer tree commit
would not contain the relevant editor state.

This extra step ensures that we always *schedule* an editor state update when performing a callout menu action
that does not automatically dismiss, so that we can prevent the callout bar from dismissing during the correct
scope.

  • UIProcess/ios/WebPageProxyIOS.mm:

(WebKit::WebPageProxy::layerTreeCommitComplete):
(WebKit::WebPageProxy::updateEditorState):

Rename editorStateChanged to updateEditorState, and make the editorStateChanged codepath only executed when an
out-of-band editor state update is delivered to the UI process.

(WebKit::WebPageProxy::dispatchDidReceiveEditorStateAfterFocus):

Notify the UI process that the initial editor state has been received; this prompts us to scroll to reveal the
focused element, if needed.

(WebKit::WebPageProxy::editorStateChanged): Deleted.

  • UIProcess/mac/WebPageProxyMac.mm:

(WebKit::WebPageProxy::updateEditorState):
(WebKit::WebPageProxy::editorStateChanged): Deleted.

  • UIProcess/win/WebPageProxyWin.cpp:

(WebKit::WebPageProxy::updateEditorState):
(WebKit::WebPageProxy::editorStateChanged): Deleted.

  • UIProcess/wpe/WebPageProxyWPE.cpp:

(WebKit::WebPageProxy::updateEditorState):
(WebKit::WebPageProxy::editorStateChanged): Deleted.

  • WebProcess/WebPage/WebPage.h:
  • WebProcess/WebPage/WebPage.messages.in:
5:12 PM Changeset in webkit [244896] by Kocsen Chung
  • 1 copy in tags/Safari-607.3.1

Tag Safari-607.3.1.

4:39 PM Changeset in webkit [244895] by chris.reid@sony.com
  • 6 edits in trunk

[CMake] Add support for LTO builds
https://bugs.webkit.org/show_bug.cgi?id=188986

Reviewed by Don Olmstead.

.:

Add LTO_MODE cmake variable to enable flto in clang builds.
Add support to use ld.lld to build with LTO.

  • Source/cmake/OptionsCommon.cmake:
  • Source/cmake/WebKitCompilerFlags.cmake:

Tools:

Add support to configure --lto-mode with cmake builds.

  • Scripts/build-webkit:
  • Scripts/webkitdirs.pm:

(generateBuildSystemFromCMakeProject):

4:31 PM Changeset in webkit [244894] by rniwa@webkit.org
  • 4 edits in trunk/Source/WebCore

Disable software keyboard for a math field textarea on desmos.com
https://bugs.webkit.org/show_bug.cgi?id=197488

Reviewed by Wenson Hsieh.

Treat a textarea inside a math field span as if it had inputmode content attribute set to none to suppress
the software keyboard on desmos.com as it interferes with website's own UI.

  • html/HTMLElement.cpp:

(WebCore::HTMLElement::canonicalInputMode const):

  • page/Quirks.cpp:

(WebCore::Quirks::needsInputModeNoneImplicitly const):

  • page/Quirks.h:
4:22 PM Changeset in webkit [244893] by timothy@apple.com
  • 5 edits in trunk/Source

NSAttributedString conversion in a loop returns nil and WKUnknownError every other time.
https://bugs.webkit.org/show_bug.cgi?id=197523

Reviewed by Darin Adler.

Source/WebCore:

  • editing/cocoa/HTMLConverter.mm:

(HTMLConverter::convert): Don't return early if m_dataSource is nil. This is already null
checked later and only needed in specific cases, it shouldn't fail the whole conversion.

Source/WebKit:

Caching the WKWebView was loading about:blank to unload the previous content.
This was causing subsequent rapid conversions to fail since the blank load
would be confused with the real content loading. Loading a blank page wasn't
really needed, it just helped keep the cached view in a cleaner state. Instead
of adding complexity to track the extra navigation, we can eliminate the blank load.
Ultimately a process swap will likely happen on the next navigation, and unused
cached views are closed quickly -- so stale content isn't held around too long.
This also avoids extra work and speeds up conversions a bit.

  • UIProcess/API/Cocoa/NSAttributedString.mm:

(+[_WKAttributedStringWebViewCache cacheWebView:]): Don't load about:blank when caching.

  • WebProcess/WebPage/Cocoa/WebPageCocoa.mm:

(WebKit::WebPage::getContentsAsAttributedString): Use rangeOfContents() for a fail
safe way to get the range needed.

3:24 PM Changeset in webkit [244892] by Chris Dumez
  • 31 edits
    2 adds in trunk

Setting a frame's src to a javascript URL should not run it synchronously
https://bugs.webkit.org/show_bug.cgi?id=197466

Reviewed by Darin Adler.

Source/WebCore:

When an iframe's src attribute is set to a javascript URL, whether when parsing
or later on via JS, we now execute the URL's JavaScript asynchronously. We used
to execute it synchronously, which was a source of bugs and also did not match
other browsers.

I have verified that our new behavior is aligned with both Firefox and Chrome.

Note that for backward-compatibility and interoperability with Blink
(https://bugs.chromium.org/p/chromium/issues/detail?id=923585), the
"javascript:" URL will still run synchronously. We should consider dropping
this quirk at some point.

Test: fast/dom/frame-src-javascript-url-async.html

  • loader/NavigationScheduler.cpp:

(WebCore::ScheduledLocationChange::ScheduledLocationChange):
(WebCore::ScheduledLocationChange::~ScheduledLocationChange):
(WebCore::NavigationScheduler::scheduleLocationChange):

  • loader/NavigationScheduler.h:

(WebCore::NavigationScheduler::scheduleLocationChange):

  • loader/SubframeLoader.cpp:

(WebCore::SubframeLoader::requestFrame):

LayoutTests:

  • fast/dom/frame-src-javascript-url-async-expected.txt: Added.
  • fast/dom/frame-src-javascript-url-async.html: Added.

Add layout test coverage for the fact that the javascript URL is executed asynchronously
whether set during parsing or later via JS. Also makes sure that executing the javascript
URL asynchronously does not replace the frame's window. This test passes in both Chrome
and Firefox.

  • imported/blink/fast/frames/navigation-in-pagehide.html:

Re-sync this test from the Blink repository.

  • fast/dom/Element/id-in-frameset-expected.txt:
  • fast/dom/Element/id-in-frameset.html:
  • fast/dom/insertedIntoDocument-iframe-expected.txt:
  • fast/dom/javascript-url-exception-isolation-expected.txt:
  • fast/dom/javascript-url-exception-isolation.html:
  • fast/dom/no-assert-for-malformed-js-url-attribute-expected.txt:
  • fast/dom/resources/javascript-url-crash-function-iframe.html:
  • fast/frames/adopt-from-created-document.html:
  • fast/frames/out-of-document-iframe-has-child-frame.html:
  • fast/loader/javascript-url-iframe-remove-on-navigate-async-delegate.html:
  • fast/loader/javascript-url-iframe-remove-on-navigate.html:
  • fast/loader/unload-mutation-crash.html:
  • fast/parser/resources/set-parent-to-javascript-url.html:
  • fast/parser/xml-error-adopted.xml:
  • http/tests/navigation/lockedhistory-iframe-expected.txt:
  • http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-javascript-url-iframe-in-iframe-expected.txt:
  • http/tests/security/contentSecurityPolicy/javascript-url-allowed-expected.txt:
  • http/tests/security/contentSecurityPolicy/javascript-url-blocked-by-default-src-star-expected.txt:
  • http/tests/security/contentSecurityPolicy/javascript-url-blocked-expected.txt:
  • http/tests/security/javascriptURL/xss-ALLOWED-from-javascript-url-sub-frame-2-level.html:
  • http/tests/security/javascriptURL/xss-ALLOWED-from-javascript-url-sub-frame.html:
  • http/tests/security/javascriptURL/xss-ALLOWED-to-javascript-url-from-javscript-url.html:
  • imported/blink/loader/iframe-sync-loads-expected.txt:
  • js/dom/call-base-resolution.html:
  • platform/wk2/http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-javascript-url-iframe-in-iframe-expected.txt:

Update / Rebaseline existing tests to reflect behavior change. I ran those tests in Firefox and Chrome to confirm that our behavior
is indeed aligned.

2:43 PM Changeset in webkit [244891] by commit-queue@webkit.org
  • 6 edits
    1 delete in trunk

Source/WebCore:
WebVTT: fix vertical cue alignment.
https://bugs.webkit.org/show_bug.cgi?id=136627.
<rdar://problem/49725538>

Patch by Gary Katsevman <git@gkatsev.com> on 2019-05-02
Reviewed by Eric Carlson.

Updated existing test results.

  • html/track/VTTCue.cpp:

(WebCore::VTTCueBox::applyCSSProperties):

LayoutTests:
WebVTT: vertical cue text alignment is the wrong way around
https://bugs.webkit.org/show_bug.cgi?id=136627
<rdar://problem/49725538>

Patch by Gary Katsevman <git@gkatsev.com> on 2019-05-02
Reviewed by Eric Carlson.

  • platform/ios/media/track/track-cue-rendering-vertical-expected.txt:
  • platform/mac/TestExpectations: Unskipped test.
  • platform/mac/media/track/track-cue-rendering-vertical-expected.png: Removed.
  • platform/mac/media/track/track-cue-rendering-vertical-expected.txt:
2:24 PM Changeset in webkit [244890] by Alan Coon
  • 1 copy in tags/Safari-607.2.6.1.2

Tag Safari-607.2.6.1.2.

2:22 PM Changeset in webkit [244889] by Alan Coon
  • 7 edits in branches/safari-607.2.6.1-branch/Source

Versioning.

2:20 PM Changeset in webkit [244888] by Truitt Savell
  • 2 edits in trunk/LayoutTests

[ Sierra Release WK2 ] Layout Test fast/workers/worker-cloneport.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=191644

Unreviewed test gardening.

  • platform/mac-wk2/TestExpectations: Remove expectation.
2:08 PM Changeset in webkit [244887] by Alan Coon
  • 5 edits
    6 adds in tags/Safari-608.1.20.1

Cherry-pick r244851. rdar://problem/49784711

[iOS] Element::focus and Element::scrollIntoView do not clamp scroll positions
https://bugs.webkit.org/show_bug.cgi?id=197211

Reviewed by Simon Fraser.

Source/WebCore:

Fixed the bug that Element::focus and Element::scrollIntoView were not clamping scroll offsets,
which causes scrollTop etc... to return a bogus negative scrolling offset.

Unfortunately, we can't just use FrameView's ScrollableArea::constrainScrollPosition since
scrollRectToVisible relies on the visible rect being expanded by the content insets in order to scroll to
a position within the content insets of UIScrollView; e.g. revealing the top of the page as the center.
We manually expand minimumScrollPosition() and maximumScrollPosition() by the content insets instead.

Tests: fast/scrolling/ios/programmatic-scroll-via-focus-should-clamp-top.html

fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-inside-iframe-should-clamp-top.html
fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-should-clamp-top.html

  • rendering/RenderLayer.cpp: (WebCore::RenderLayer::scrollRectToVisible):

LayoutTests:

Added tests for scrolling via Element.prototype.focus and Element.prototype.scrollIntoView.

  • fast/scrolling/ios/programmatic-scroll-via-focus-should-clamp-top-expected.txt: Added.
  • fast/scrolling/ios/programmatic-scroll-via-focus-should-clamp-top.html: Added.
  • fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-inside-iframe-should-clamp-top-expected.txt: Added.
  • fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-inside-iframe-should-clamp-top.html: Added.
  • fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-should-clamp-top-expected.txt: Added.
  • fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-should-clamp-top.html: Added.
  • platform/ios-wk2/fast/dom/focus-contenteditable-expected.txt: Rebaselined the output now that the scroll position is properly clamped.
  • platform/ios-wk2/imported/w3c/web-platform-tests/css/cssom-view/elementFromPoint-expected.txt: Rebaselined as one more test case is passing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244851 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2:08 PM Changeset in webkit [244886] by Alan Coon
  • 17 edits
    10 adds in tags/Safari-608.1.20.1

Cherry-pick r244849. rdar://problem/50063091

[iOS] Add a version of viewport shrink-to-fit heuristics that preserves page layout
https://bugs.webkit.org/show_bug.cgi?id=197342
<rdar://problem/50063091>

Reviewed by Tim Horton.

Source/WebCore:

Adds support for a new shrink-to-fit heuristic that attempts to lay out the contents of the page at a larger
width in order to shrink content to fit the viewport. See WebKit ChangeLog for more details.

Tests: fast/viewport/ios/shrink-to-fit-content-constant-width.html

fast/viewport/ios/shrink-to-fit-content-large-width-breakpoint.html
fast/viewport/ios/shrink-to-fit-content-no-viewport.html
fast/viewport/ios/shrink-to-fit-content-responsive-viewport-with-horizontal-overflow.html
fast/viewport/ios/shrink-to-fit-content-temporary-overflow.html

  • page/ViewportConfiguration.cpp: (WebCore::ViewportConfiguration::setMinimumEffectiveDeviceWidth): (WebCore::ViewportConfiguration::setIsKnownToLayOutWiderThanViewport): (WebCore::ViewportConfiguration::description const):
  • page/ViewportConfiguration.h: (WebCore::ViewportConfiguration::canIgnoreScalingConstraints const): (WebCore::ViewportConfiguration::minimumEffectiveDeviceWidth const):

Add several new getters and setters in ViewportConfiguration.

(WebCore::ViewportConfiguration::isKnownToLayOutWiderThanViewport const):
(WebCore::ViewportConfiguration::shouldIgnoreMinimumEffectiveDeviceWidth const):

Importantly, only allow ignoring the minimum effective device width in webpages with responsive viewports, if
they also have *not* laid out wider than the viewport.

(WebCore::ViewportConfiguration::setForceAlwaysUserScalable):

Source/WebKit:

This patch introduces a new shrink-to-fit heuristic that attempts to lay out the contents of the page at a
larger width in order to shrink content to fit the viewport. This is similar to existing shrink-to-fit behaviors
used for viewport sizing in multitasking mode, except that it not only scales the view, but additionally expands
the layout size, such that the overall layout of the page is preserved. In fact, the reason we ended up
reverting the existing flavor of shrink-to-fit in all cases except for multitasking was that page layout was not
preserved, which caused elements that poke out of the viewport to make the rest of the page look out of
proportion — see <rdar://problem/23818102> and related radars.

Covered by 5 new layout tests, and by adjusting a couple of existing layout tests. See comments below for more
details.

  • Platform/Logging.h:

Add a new ViewportSizing logging channel. This will only log on pages that overflow the viewport and shrink to
fit as a result.

  • Shared/WebPreferences.yaml:

Turn IgnoreViewportScalingConstraints off by default. This preference currently controls whether we allow
shrink-to-fit behaviors, and is only used by Safari when it is in multitasking mode. The value of this
preference is currenly *on* by default, and is turned off almost immediately during every page load after the
first visible content rect update, wherein visibleContentRectUpdateInfo.allowShrinkToFit() is false.

However, this sometimes causes a brief jitter during page load; to fix this, make the default value for
IgnoreViewportScalingConstraints false, and change the logic in WebPage::updateVisibleContentRects to
setCanIgnoreScalingConstraints to true if either the IgnoreViewportScalingConstraints preference (not only
affected by an internal debug switch) is true, or WKWebView SPI is used to enable the behavior.

  • WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp: (WebKit::WebFrameLoaderClient::dispatchDidFinishDocumentLoad): (WebKit::WebFrameLoaderClient::dispatchDidFinishLoad):

Add a new hook for WebFrameLoaderClient to call into WebPage when document load finishes. Also, tweak
dispatchDidFinishLoad to take a WebFrame& instead of a WebFrame* in a drive-by fix (the frame is assumed to be
non-null anyways).

  • WebProcess/WebPage/WebPage.cpp: (WebKit::WebPage::didCommitLoad): (WebKit::WebPage::didFinishDocumentLoad): (WebKit::WebPage::didFinishLoad):

When finishing document load or finishing the overall load, kick off the shrink-to-fit timer; when committing a
load, cancel the timer.

  • WebProcess/WebPage/WebPage.h:
  • WebProcess/WebPage/ios/WebPageIOS.mm: (WebKit::WebPage::setViewportConfigurationViewLayoutSize):

Don't allow the minimum effective device width from the client to stomp over any minimum effective device width
set as a result of the new shrink-to-fit heuristic; on some pages that load quickly, this can result in a race
where the minimum effective device width (i.e. a value that lower-bounds the minimum layout width) is first set
by the shrink-to-fit heuristic, and then set to an incorrect value by the client.

In the near future, web view SPI used to set the minimum effective device width should actually be removed
altogether, since the new shrink-to-fit heuristic supersedes any need for the client to fiddle with the minimum
effective device width.

(WebKit::WebPage::dynamicViewportSizeUpdate):

When performing a dynamic viewport size update, additionally re-run the shrink-to-fit heuristic. This allows
the minimum layout size of the viewport to be updated, if necessary. An example of where this matters is when a
web page is *below* a tablet/desktop layout breakpoint in portrait device orientation, but then exceeds this
layout breakpoint in landscape orientation. In this scenario, rotating the device should swap between these two
page layouts.

(WebKit::WebPage::resetViewportDefaultConfiguration):
(WebKit::WebPage::scheduleShrinkToFitContent):
(WebKit::WebPage::shrinkToFitContentTimerFired):
(WebKit::WebPage::immediatelyShrinkToFitContent):

Leverage the existing capability for a viewport to have a "minimum effective device width" to grant the viewport
a larger layout size than it would normally have, and then scale down to fit within the bounds of the view. One
challenge with this overall approach is that laying out at a larger width may cause the page to lay out even
wider in response, which may actually worsen horizontal scrolling. To mitigate this, we only attempt to lay out
at the current content width once; if laying out at this width reduced the amount of horizontal scrolling by any
amount, then proceed with this layout width; otherwise, revert to the previous layout width.

(WebKit::WebPage::shouldIgnoreMetaViewport const):

Pull some common logic out into a readonly getter.

(WebKit::WebPage::updateVisibleContentRects):

See the comment below WebPreferences.yaml, above.

LayoutTests:

Introduces new layout tests, and adjusts some existing tests. See comments below.

  • fast/viewport/ios/shrink-to-fit-content-constant-width-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-constant-width.html: Added.

Add a new layout test to exercise the scenario where a constant width viewport narrower than the view is used.

  • fast/viewport/ios/shrink-to-fit-content-large-width-breakpoint-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-large-width-breakpoint.html: Added.

Add a new layout test to exercise the scenario where a responsive website that lays out larger than the view
width ends up with even more horizontal scrolling when laying out at the initial content width. In this
scenario, we shouldn't try to expand the viewport to try and encompass the content width, since that would only
induce even worse horizontal scrolling.

  • fast/viewport/ios/shrink-to-fit-content-no-viewport-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-no-viewport.html: Added.

Add a new layout test for the case where there is no viewport, but content lays out wider than the view.

  • fast/viewport/ios/shrink-to-fit-content-responsive-viewport-with-horizontal-overflow-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-responsive-viewport-with-horizontal-overflow.html: Added.

Add a new layout test for the case where the page has opted for a responsive viewport (device-width, initial
scale 1), but has laid out wider than the viewport anyways. In this case, we want to shrink the contents down to
fit inside the view.

  • fast/viewport/ios/shrink-to-fit-content-temporary-overflow-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-temporary-overflow.html: Added.

Add a new layout test to exercise the case where, during page load, content width temporarily increases, and
then decreases such that it once again fits within the viewport. In this case, we don't want to expand the
viewport to be as wide as the large temporary width of the page.

  • fast/viewport/ios/width-is-device-width-overflowing-body-overflow-hidden-expected.txt:
  • fast/viewport/ios/width-is-device-width-overflowing-body-overflow-hidden.html:
  • fast/viewport/ios/width-is-device-width-overflowing-expected.txt:
  • fast/viewport/ios/width-is-device-width-overflowing.html:

Tweak these 2 existing layout tests to include "shrink-to-fit=no", to prevent the new heuristics from shrinking
the page to fit on device classes that use native viewports by default.

  • platform/ipad/fast/viewport/ios/width-is-device-width-overflowing-body-overflow-hidden-expected.txt:
  • platform/ipad/fast/viewport/ios/width-is-device-width-overflowing-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2:08 PM Changeset in webkit [244885] by Alan Coon
  • 15 edits
    2 adds in tags/Safari-608.1.20.1

Cherry-pick r244823. rdar://problem/46677392

XMLHttpRequest should propagate user gestures for media playback
https://bugs.webkit.org/show_bug.cgi?id=197428
<rdar://problem/46677392>

Reviewed by Jer Noble.

Source/WebCore:

A user gesture the would allow media state change in effect when XMLHttpRequest.send is
called should be active when the event handlers fire after the transaction completes successfully.

Test: http/tests/media/user-gesture-preserved-across-xmlhttprequest.html

  • dom/UserGestureIndicator.cpp: (WebCore::UserGestureIndicator::UserGestureIndicator): Add a 'scope' parameter to potentially limit the scope of the gesture to just media. (WebCore::UserGestureIndicator::~UserGestureIndicator): Clear the scope.
  • dom/UserGestureIndicator.h: (WebCore::UserGestureToken::processingUserGesture const): (WebCore::UserGestureToken::setScope): (WebCore::UserGestureToken::resetScope): (WebCore::UserGestureToken::hasExpired const):
  • page/DOMTimer.cpp: (WebCore::DOMTimerFireState::DOMTimerFireState): Don't need to store the nested timer interval, UserGestureIndicator knows when it started. (WebCore::DOMTimer::DOMTimer): Ditto. (WebCore::DOMTimer::fired): Ditto. (WebCore::DOMTimerFireState::nestedTimerInterval const): Deleted. (WebCore::shouldForwardUserGesture): Deleted. (WebCore::userGestureTokenToForward): Deleted. (WebCore::currentNestedTimerInterval): Deleted.
  • page/DOMTimer.h:
  • testing/Internals.cpp: (WebCore::Internals::setXHRMaximumIntervalForUserGestureForwarding): Override the maximum user gesture interval for testing.
  • testing/Internals.h:
  • testing/Internals.idl:
  • xml/XMLHttpRequest.cpp: (WebCore::XMLHttpRequest::XMLHttpRequest): (WebCore::XMLHttpRequest::send): Stash the user gesture token. (WebCore::XMLHttpRequest::dispatchEvent): Clear user gesture token if it has expired. If still valid, activate it.
  • xml/XMLHttpRequest.h:

LayoutTests:

  • fast/events/popup-blocking-timers5-expected.txt:
  • fast/events/popup-blocking-timers5.html: Decrease the timer interval from 1000 to 900 because the user gesture is invalidated based on wall clock time.
  • fast/events/popup-blocking-timers6-expected.txt:
  • fast/events/popup-blocking-timers6.html: Increase the timer interval from 1001 to 1100 because the user gesture is invalidated based on wall clock time.
  • http/tests/media/user-gesture-preserved-across-xmlhttprequest-expected.txt: Added.
  • http/tests/media/user-gesture-preserved-across-xmlhttprequest.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244823 268f45cc-cd09-0410-ab3c-d52691b4dbfc

1:55 PM Changeset in webkit [244884] by Chris Dumez
  • 4 edits in trunk

Make both filterForRegistrableDomains() in WebKit::NetworkProcess use WebCore::RegistrableDomain::uncheckedCreateFromHost()
https://bugs.webkit.org/show_bug.cgi?id=197521
<rdar://problem/49651503>

Patch by John Wilander <wilander@apple.com> on 2019-05-02
Reviewed by Brent Fulgham.

Source/WebKit:

  • NetworkProcess/NetworkProcess.cpp:

(WebKit::filterForRegistrableDomains):

Tools:

  • TestWebKitAPI/Tests/WebCore/RegistrableDomain.cpp:

(TestWebKitAPI::TEST):

Added a test for host strings with leading dots since they are
often used for the domain attribute in cookies.

1:50 PM Changeset in webkit [244883] by pvollan@apple.com
  • 8 edits in trunk

-[WKWebsiteDataStore removeDataOfTypes:forDataRecords:completionHandler:] doesn't delete _WKWebsiteDataTypeCredentials
https://bugs.webkit.org/show_bug.cgi?id=197510
Source/WebKit:

<rdar://problem/50372338>

Reviewed by Alex Christensen.

When asked to delete persistent credentials for a set of origins, send a message to the network process with the
list of origins. The network process will then delete all persistent credentials from these origins from
NSURLCredentialStorage.

  • NetworkProcess/NetworkProcess.cpp:

(WebKit::NetworkProcess::removeCredentialsWithOrigins):

  • NetworkProcess/NetworkProcess.h:
  • NetworkProcess/NetworkProcess.messages.in:
  • NetworkProcess/cocoa/NetworkProcessCocoa.mm:

(WebKit::NetworkProcess::removeCredentialsWithOrigins):

  • UIProcess/WebsiteData/WebsiteDataStore.cpp:

(WebKit::WebsiteDataStore::removeData):

Tools:

Reviewed by Alex Christensen.

  • TestWebKitAPI/Tests/WebKitCocoa/WKWebsiteDatastore.mm:

(TestWebKitAPI::TEST):

1:46 PM Changeset in webkit [244882] by Alan Coon
  • 7 edits in tags/Safari-608.1.20.1/Source

Versioning.

1:37 PM Changeset in webkit [244881] by don.olmstead@sony.com
  • 18 edits in trunk

[CMake] Refactor WEBKIT_MAKE_FORWARDING_HEADERS into WEBKIT_COPY_FILES
https://bugs.webkit.org/show_bug.cgi?id=197174

Reviewed by Alex Christensen.

.:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES which is a general
purpose macro that copies files to a destination. WEBKIT_COPY_FILES removes the
requirement of a target being passed in so there is no longer any implicit dependency
between the target and the copying. Instead the dependencies need to be explicit.

This opens the macro up for use by third party libraries and for copying other files
to a destination.

  • Source/cmake/WebKitMacros.cmake:

Source/JavaScriptCore:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES and make dependencies
for framework headers explicit.

  • CMakeLists.txt:

Source/WebCore:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES and make dependencies
for framework headers explicit.

  • CMakeLists.txt:

Source/WebCore/PAL:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES.

  • pal/CMakeLists.txt:

Source/WebKit:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES and make dependencies
for framework headers explicit.

  • CMakeLists.txt:
  • PlatformWin.cmake:

Source/WebKitLegacy:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES and make dependencies
for framework headers explicit.

  • PlatformWin.cmake:

Source/WTF:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES.

  • wtf/CMakeLists.txt:

Tools:

Make TestJSC dependent on JavaScriptCorePrivateFrameworkHeaders.

  • TestWebKitAPI/PlatformGTK.cmake:
  • TestWebKitAPI/PlatformWPE.cmake:
1:29 PM Changeset in webkit [244880] by Alan Coon
  • 1 copy in tags/Safari-608.1.20.1

New tag.

12:15 PM Changeset in webkit [244879] by jiewen_tan@apple.com
  • 2 edits in trunk/Source/WebCore

[WebAuthN] Add a quirk for google.com when processing AppID extension
https://bugs.webkit.org/show_bug.cgi?id=196046
<rdar://problem/49088479>

Reviewed by Brent Fulgham.

Relaxing the same site restriction on AppID while in google.com and any
of its subdomains to allow two www.gstatic.com AppIDs to slip in.

Covered by manual tests on Google.com.

  • Modules/webauthn/AuthenticatorCoordinator.cpp:

(WebCore::AuthenticatorCoordinatorInternal::needsAppIdQuirks):
(WebCore::AuthenticatorCoordinatorInternal::processAppIdExtension):

11:40 AM Changeset in webkit [244878] by Alan Coon
  • 7 edits in branches/safari-607-branch/Source

Versioning.

11:09 AM Changeset in webkit [244877] by pvollan@apple.com
  • 2 edits in trunk/Source/WebKit

[macOS] Fix syscall sandbox violation
https://bugs.webkit.org/show_bug.cgi?id=197517
<rdar://problem/49945031>

Reviewed by Brent Fulgham.

A syscall needs to be added to the sandbox on macOS.

  • WebProcess/com.apple.WebProcess.sb.in:
10:59 AM Changeset in webkit [244876] by Ross Kirsling
  • 2 edits in trunk/Source/WebCore

Unreviewed fix for non-unified build after r244853.

  • page/SecurityOrigin.cpp:
10:38 AM Changeset in webkit [244875] by achristensen@apple.com
  • 3 edits in trunk/Source/WebKit

Safe browsing warning should update colors when a user switches between light and dark appearance
https://bugs.webkit.org/show_bug.cgi?id=197443
<rdar://problem/49883917>

Reviewed by Tim Horton.

We draw the WKSafeBrowsingExclamationPoint ourselves, so we need to call setNeedsDisplay when viewDidChangeEffectiveAppearance is called.
Instead of setting NSView.layer.backgroundColor we need to make an NSView subclass I call WKSafeBrowsingBox and we need to set its layer's
backgroundColor in updateLayer, otherwise the CGColor isn't updated from the NSColor.

  • UIProcess/Cocoa/WKSafeBrowsingWarning.h:
  • UIProcess/Cocoa/WKSafeBrowsingWarning.mm:

(colorForItem):
(-[WKSafeBrowsingExclamationPoint viewDidChangeEffectiveAppearance]):
(-[WKSafeBrowsingBox setSafeBrowsingBackgroundColor:]):
(-[WKSafeBrowsingBox updateLayer]):
(-[WKSafeBrowsingWarning initWithFrame:safeBrowsingWarning:completionHandler:]):
(-[WKSafeBrowsingWarning addContent]):
(-[WKSafeBrowsingWarning showDetailsClicked]):
(setBackground): Deleted.

10:35 AM Changeset in webkit [244874] by achristensen@apple.com
  • 2 edits in trunk/Tools

Build fix for internal branch.

  • TestWebKitAPI/Tests/WebKitCocoa/Challenge.mm:

RetainPtr.h was not explicitly included.

10:01 AM Changeset in webkit [244873] by Jonathan Bedard
  • 2 edits in trunk/Tools

webkitpy: Ignore errors when shutting down an already shutdown simulator
<https://bugs.webkit.org/show_bug.cgi?id=197514>
<rdar://problem/50390247>

Reviewed by Lucas Forschler.

  • Scripts/webkitpy/xcode/simulated_device.py:

(SimulatedDevice._shut_down):

9:17 AM Changeset in webkit [244872] by msaboff@apple.com
  • 4 edits in trunk

Unreviewed rollout of r244862.

8:21 AM Changeset in webkit [244871] by magomez@igalia.com
  • 2 edits in trunk/LayoutTests

Unreviewed GTK+ gardening after r244868.

  • platform/gtk/TestExpectations:
7:57 AM Changeset in webkit [244870] by pvollan@apple.com
  • 2 edits in trunk/LayoutTests

[Win10] Some tests are failing only on specific machines
https://bugs.webkit.org/show_bug.cgi?id=197509

Unreviewed test gardening.

  • platform/win/TestExpectations:
3:12 AM Changeset in webkit [244869] by commit-queue@webkit.org
  • 23 edits in trunk

[GTK][WPE] Disable "thin", "thick", "medium" values of mfrac@linethickness at runtime
https://bugs.webkit.org/show_bug.cgi?id=196142

This patch introduces some experimental runtime flag to let users
disable MathML features that are removed from MathML Core [1]. For now,
LayoutTests/imported/w3c:

these features are only disabled on GTK and WPE ports. This patch also
adds a condition to disable "thin", "thick", "medium" values of
mfrac@linethickness at runtime as agreed in [2].

[1] https://mathml-refresh.github.io/mathml-core/
[2] https://github.com/mathml-refresh/mathml/issues/4

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

  • web-platform-tests/mathml/presentation-markup/fractions/frac-linethickness-001.html:

Explicitly enable deprecated features in this test. It should be moved
back to WebKit's test suite if/when it is updated in the upstream WPT
repository so also add a reminder for the next synchronization.

Source/WebCore:

these features are only disabled on GTK and WPE ports. This patch also
adds a condition to disable "thin", "thick", "medium" values of
mfrac@linethickness at runtime as agreed in [2].

[1] https://mathml-refresh.github.io/mathml-core/
[2] https://github.com/mathml-refresh/mathml/issues/4

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

No new tests, covered by frac-linethickness-0001.html

  • mathml/MathMLFractionElement.cpp:

(WebCore::MathMLFractionElement::lineThickness): Just do standard
parsing for MathML lengths when non-core MathML features are disabled.

  • page/Settings.yaml: Add WebCore setting.

Source/WebKit:

these features are only disabled on GTK and WPE ports.

[1] https://mathml-refresh.github.io/mathml-core/

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

  • Shared/WebPreferences.yaml:
  • Shared/WebPreferencesDefaultValues.h:

Source/WebKitLegacy/mac:

these features are only disabled on GTK and WPE ports.

[1] https://mathml-refresh.github.io/mathml-core/

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

  • WebView/WebPreferenceKeysPrivate.h:
  • WebView/WebPreferences.mm:

(+[WebPreferences initialize]):
(-[WebPreferences coreMathMLEnabled]):
(-[WebPreferences setCoreMathMLEnabled:]):

  • WebView/WebPreferencesPrivate.h:
  • WebView/WebView.mm:

(-[WebView _preferencesChanged:]):

Source/WebKitLegacy/win:

these features are only disabled on GTK and WPE ports.

[1] https://mathml-refresh.github.io/mathml-core/

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

  • Interfaces/IWebPreferencesPrivate.idl:
  • WebPreferenceKeysPrivate.h:
  • WebPreferences.cpp:

(WebPreferences::initializeDefaultSettings):
(WebPreferences::coreMathMLEnabled):
(WebPreferences::setCoreMathMLEnabled):

  • WebPreferences.h:
  • WebView.cpp:

(WebView::notifyPreferencesChanged):

Tools:

these features are only disabled on GTK and WPE ports.

[1] https://mathml-refresh.github.io/mathml-core/

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

  • DumpRenderTree/TestOptions.cpp:

(TestOptions::TestOptions):

  • DumpRenderTree/TestOptions.h:
  • DumpRenderTree/mac/DumpRenderTree.mm:

(setWebPreferencesForTestOptions):

May 1, 2019:

11:27 PM Changeset in webkit [244868] by Devin Rousso
  • 3 edits in trunk/LayoutTests

Unreviewed, fix test failures after r242992.
<rdar://problem/50383600>

  • inspector/console/console-screenshot.html:
  • inspector/console/console-screenshot-expected.txt:

Don't print the dataURL in the expectation file.

10:01 PM Changeset in webkit [244867] by mitz@apple.com
  • 9 copies
    1 add in releases/Apple/Safari Technology Preview 81

Added a tag for Safari Technology Preview release 81.

9:20 PM Changeset in webkit [244866] by Wenson Hsieh
  • 4 edits in trunk/Source/WebKit

Add the ability to import WebKitAdditions snippets in WebKit.apinotes
https://bugs.webkit.org/show_bug.cgi?id=197487
<rdar://problem/50389085>

Reviewed by Tim Horton.

  • Shared/API/Cocoa/WebKit.apinotes: Import a new WebKitAdditions file.
  • WebKit.xcodeproj/project.pbxproj: Check all files in Headers/ and PrivateHeaders/ that need replacement, not

simply anything ending with ".h".

  • mac/replace-webkit-additions-includes.py:

(main): Teach the replacement script to handle files in Headers/ or PrivateHeaders/ that don't end in ".h".

8:10 PM Changeset in webkit [244865] by sbarati@apple.com
  • 3 edits
    1 add in trunk

Baseline JIT should do argument value profiling after checking for stack overflow
https://bugs.webkit.org/show_bug.cgi?id=197052
<rdar://problem/50009602>

Reviewed by Yusuke Suzuki.

JSTests:

  • stress/check-stack-overflow-before-value-profiling-arguments.js: Added.

Source/JavaScriptCore:

Otherwise, we may do value profiling without running a write barrier, which
is against the rules of how we do value profiling.

  • jit/JIT.cpp:

(JSC::JIT::compileWithoutLinking):

7:40 PM Changeset in webkit [244864] by ysuzuki@apple.com
  • 4 edits
    1 add in trunk

[JSC] Inlining Getter/Setter should care availability of ad-hocly constructed frame
https://bugs.webkit.org/show_bug.cgi?id=197405

Reviewed by Saam Barati.

JSTests:

  • stress/getter-setter-inlining-should-emit-movhint.js: Added.

(foo):
(test):
(i.o.get f):
(i.o.set f):

Source/JavaScriptCore:

When inlining getter and setter calls, we setup a stack frame which does not appear in the bytecode.
Because Inlining can switch on executable, we could have a graph like this.

BB#0

...
30: GetSetter
31: MovHint(loc10)
32: SetLocal(loc10)
33: MovHint(loc9)
34: SetLocal(loc9)
...
37: GetExecutable(@30)
...
41: Switch(@37)

BB#2

42: GetLocal(loc12, bc#7 of caller)
...
--> callee: loc9 and loc10 are arguments of callee.

...
<HERE, exit to callee, loc9 and loc10 are required in the bytecode>

When we prune OSR availability at the beginning of BB#2 (bc#7 in the caller), we prune loc9 and loc10's liveness because the caller does not actually have loc9 and loc10.
However, when we begin executing the callee, we need OSR exit to be aware of where it can recover the arguments to the setter, loc9 and loc10.

This patch inserts MovHint at the beginning of callee for a getter / setter stack frame to make arguments (loc9 and loc10 in the above example) recoverable from OSR exit.
We also move arity fixup DFG nodes from the caller to the callee, since moved arguments are not live in the caller too.

Interestingly, this fix also reveals the existing issue in LiveCatchVariablePreservationPhase. We emitted Flush for |this| of InlineCallFrame blindly if we saw InlineCallFrame
inside a block which is covered by catch handler. But this is wrong because inlined function can finish its execution within the block, and |this| is completely unrelated to
the catch handler if the catch handler is in the outer callee. We already collect all the live locals at the catch handler. And this locals must include arguments too if the
catch handler is in inlined function. So, we should not emit Flush for each |this| of seen InlineCallFrame. This emitted Flush may connect unrelated locals in the catch handler
to the locals that is only defined and used in the inlined function, and it leads to the results like DFG says the local is live while the bytecode says the local is dead.
This results in reading and using garbage in OSR entry because DFG OSR entry needs to fill live DFG values from the stack.

  • dfg/DFGByteCodeParser.cpp:

(JSC::DFG::ByteCodeParser::inlineCall):
(JSC::DFG::ByteCodeParser::handleGetById):
(JSC::DFG::ByteCodeParser::handlePutById):

  • dfg/DFGLiveCatchVariablePreservationPhase.cpp:

(JSC::DFG::LiveCatchVariablePreservationPhase::handleBlockForTryCatch):

7:20 PM Changeset in webkit [244863] by jiewen_tan@apple.com
  • 5 edits
    3 adds in trunk

[WebAuthN] Adopt SecurityOrigin::isMatchingRegistrableDomainSuffix()
https://bugs.webkit.org/show_bug.cgi?id=197481

Reviewed by Brent Fulgham.

Source/WebCore:

This patch implements Step 6-7 from:
https://www.w3.org/TR/webauthn/#createCredential,
https://www.w3.org/TR/webauthn/#discover-from-external-source.

Test: http/wpt/webauthn/public-key-credential-ip-address.html

  • Modules/webauthn/AuthenticatorCoordinator.cpp:

(WebCore::AuthenticatorCoordinator::create const):
(WebCore::AuthenticatorCoordinator::discoverFromExternalSource const):

LayoutTests:

  • http/wpt/webauthn/public-key-credential-create-failure.https.html:
  • http/wpt/webauthn/public-key-credential-get-failure.https.html:
  • http/wpt/webauthn/public-key-credential-ip-address-expected.txt: Added.
  • http/wpt/webauthn/public-key-credential-ip-address.html: Added.
  • http/wpt/webauthn/resources/public-key-credential-ip-address.https.html: Added.
7:04 PM Changeset in webkit [244862] by msaboff@apple.com
  • 3 edits
    1 add in trunk

ASSERTION FAILED: !m_needExceptionCheck with --validateExceptionChecks=1; ProxyObject.getOwnPropertySlotCommon/JSFunction.callerGetter
https://bugs.webkit.org/show_bug.cgi?id=197485

Reviewed by Saam Barati.

JSTests:

New test.

  • stress/proxy-getOwnPropertySlots-exceptionChecks.js: Added.

(foo):

Source/JavaScriptCore:

Added an EXCEPTION_ASSERT after call to getOwnPropertySlot().

  • runtime/JSObject.cpp:

(JSC::JSObject::getOwnPropertyDescriptor):

5:25 PM Changeset in webkit [244861] by Ross Kirsling
  • 6 edits in trunk/Source/JavaScriptCore

RemoteInspector::updateAutomaticInspectionCandidate should have a default implementation.
https://bugs.webkit.org/show_bug.cgi?id=197439

Reviewed by Devin Rousso.

On non-Cocoa platforms, automatic inspection is not currently implemented,
so updateAutomaticInspectionCandidate falls back to the logic of updateTarget.
This logic already existed in three places, so refactor it into a common private method
and allow our websocket-based RWI implementation to make use of it too.

  • inspector/remote/RemoteInspector.cpp:

(Inspector::RemoteInspector::updateTarget):
(Inspector::RemoteInspector::updateTargetMap):
(Inspector::RemoteInspector::updateAutomaticInspectionCandidate):

  • inspector/remote/RemoteInspector.h:
  • inspector/remote/cocoa/RemoteInspectorCocoa.mm:

(Inspector::RemoteInspector::updateAutomaticInspectionCandidate):

  • inspector/remote/glib/RemoteInspectorGlib.cpp:

(Inspector::RemoteInspector::updateAutomaticInspectionCandidate): Deleted.

  • inspector/remote/socket/RemoteInspectorSocket.cpp:

(Inspector::RemoteInspector::updateAutomaticInspectionCandidate): Deleted.

5:20 PM Changeset in webkit [244860] by youenn@apple.com
  • 3 edits in trunk/Source/WebCore

Add back hasNullReferences() assert in Document::updateIsPlayingMedia
https://bugs.webkit.org/show_bug.cgi?id=197477

Reviewed by Eric Carlson.

AudioContext had a special scriptExecutionContext() getter
that was returning nullptr when AudioContext is stopped.
Instead, make this getter behave as all other scriptExecutionContext() getters
and make sure existing call sites of the getter are fine with the change.

Covered by existing tests.

  • Modules/webaudio/AudioContext.cpp:

(WebCore::AudioContext::AudioContext):
No need to check whether offline since this constructor is for non offline AudioContext.
(WebCore::AudioContext::scriptExecutionContext const):
(WebCore::AudioContext::fireCompletionEvent):

  • dom/Document.cpp:

(WebCore::Document::updateIsPlayingMedia):

5:09 PM Changeset in webkit [244859] by stephan.szabo@sony.com
  • 2 edits in trunk

[PlayStation] Update port for WEBKIT_EXECUTABLE changes
https://bugs.webkit.org/show_bug.cgi?id=197483

Reviewed by Don Olmstead.

  • Source/cmake/OptionsPlayStation.cmake:

Update wrapping for configuring executables' playstation
specific data to new WEBKIT_EXECUTABLE macro.

5:00 PM Changeset in webkit [244858] by Devin Rousso
  • 3 edits in trunk/LayoutTests

Unreviewed, fix test failures after r242809.
<rdar://problem/50383600>

  • inspector/canvas/recording-html-2d.html:
  • inspector/canvas/recording-html-2d-expected.txt:

Don't print the dataURL in the expectation file.

4:47 PM Changeset in webkit [244857] by don.olmstead@sony.com
  • 16 edits in trunk/Tools

TestWebKitAPI config.h should be aware of what suite is being built
https://bugs.webkit.org/show_bug.cgi?id=196583

Reviewed by Konstantin Tokarev.

Migrate to WEBKIT_EXECUTABLE to define the Test* executables. As an added
benefit the macro defines BUILDING_* for the executable which allows the includes
of the config.h to be tailored to the test suite being built. Because of that the
dependencies of the Test* executable are specific to the test suite which may
speed up the build.

Some source files were including "PlatformUtilities.h" which includes WebKit headers
when "Utilities.h" is what was needed.

  • TestWebKitAPI/CMakeLists.txt:
  • TestWebKitAPI/PlatformGTK.cmake:
  • TestWebKitAPI/PlatformJSCOnly.cmake:
  • TestWebKitAPI/PlatformMac.cmake:
  • TestWebKitAPI/PlatformPlayStation.cmake:
  • TestWebKitAPI/PlatformUtilities.h:
  • TestWebKitAPI/PlatformWPE.cmake:
  • TestWebKitAPI/PlatformWin.cmake:
  • TestWebKitAPI/Tests/WTF/RefPtr.cpp:
  • TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:
  • TestWebKitAPI/Tests/WebCore/FileMonitor.cpp:
  • TestWebKitAPI/Tests/WebCore/LineBreaking.mm:
  • TestWebKitAPI/Tests/WebCore/cocoa/SharedBuffer.mm:
  • TestWebKitAPI/Tests/WebCore/cocoa/WebCoreNSURLSession.mm:
  • TestWebKitAPI/config.h:
4:21 PM Changeset in webkit [244856] by Justin Fan
  • 35 edits
    1 delete in trunk

Update WebGPU class names based on sketch.idl
https://bugs.webkit.org/show_bug.cgi?id=194260

Reviewed by Dean Jackson.

Update all exposed Web GPU interface names to GPU* prefix.

Source/WebCore:

Existing Web GPU tests updated to expect new names.

  • Modules/webgpu/WebGPU.idl:
  • Modules/webgpu/WebGPUAdapter.idl:
  • Modules/webgpu/WebGPUBindGroup.idl:
  • Modules/webgpu/WebGPUBindGroupLayout.idl:
  • Modules/webgpu/WebGPUBuffer.idl:
  • Modules/webgpu/WebGPUDevice.idl:
  • Modules/webgpu/WebGPUInputStepMode.h: Removed.
  • Modules/webgpu/WebGPUPipelineLayout.idl:
  • Modules/webgpu/WebGPUProgrammablePassEncoder.idl:
  • Modules/webgpu/WebGPUQueue.idl:
  • Modules/webgpu/WebGPURenderPassEncoder.idl:
  • Modules/webgpu/WebGPURenderPipeline.idl:
  • Modules/webgpu/WebGPUSampler.idl:
  • Modules/webgpu/WebGPUTexture.idl:
  • Modules/webgpu/WebGPUTextureView.idl:
  • WebCore.xcodeproj/project.pbxproj:
  • bindings/js/WebCoreBuiltinNames.h:

LayoutTests:

  • webgpu/adapter-options.html:
  • webgpu/bind-groups-expected.txt:
  • webgpu/bind-groups.html:
  • webgpu/buffer-resource-triangles.html:
  • webgpu/js/webgpu-functions.js:

(beginBasicRenderPass):

  • webgpu/map-read-buffers.html:
  • webgpu/map-write-buffers.html:
  • webgpu/pipeline-layouts-expected.txt:
  • webgpu/pipeline-layouts.html:
  • webgpu/queue-creation-expected.txt:
  • webgpu/queue-creation.html:
  • webgpu/render-command-encoding-expected.txt:
  • webgpu/render-command-encoding.html:
  • webgpu/render-pipelines-expected.txt:
  • webgpu/render-pipelines.html:
  • webgpu/textures-textureviews.html:
  • webgpu/webgpu-enabled.html:
4:07 PM Changeset in webkit [244855] by Alan Bujtas
  • 6 edits in trunk/LayoutTests

[iOS] Unreviewed test gardening.

  • platform/ios-wk2/fast/block/margin-collapse/103-expected.txt:
  • platform/ios/compositing/overflow/theme-affects-visual-overflow-expected.txt:
  • platform/ios/fast/block/margin-collapse/103-expected.txt:
  • platform/ios/fast/parser/bad-xml-slash-expected.txt:
  • platform/ios/tables/mozilla/bugs/bug4527-expected.txt:
3:34 PM Changeset in webkit [244854] by Wenson Hsieh
  • 2 edits in trunk/LayoutTests

REGRESSION(r244494): editing/pasteboard/5761530-1.html is flaky due to a simple line layout bug
https://bugs.webkit.org/show_bug.cgi?id=197438
<rdar://problem/50348091>

Reviewed by Zalan Bujtas.

Work around the simple line layout bug in the flaky layout test by disabling simple line layout.

  • editing/pasteboard/5761530-1.html:
3:08 PM Changeset in webkit [244853] by jiewen_tan@apple.com
  • 7 edits in trunk

Move Document::domainIsRegisterable to SecurityOrigin::isMatchingRegistrableDomainSuffix
https://bugs.webkit.org/show_bug.cgi?id=181950
<rdar://problem/43357371>

Reviewed by Brent Fulgham.

Source/WebCore:

This patch moves Document::domainIsRegisterable to SecurityOrigin::isMatchingRegistrableDomainSuffix
to be more aligned with the HTML standard:
https://html.spec.whatwg.org/multipage/origin.html#is-a-registrable-domain-suffix-of-or-is-equal-to.
Besides that, it also removes redundant codes within the original method that is also done in
OriginAccessEntry::matchesOrigin.

Covered by new API tests.

  • dom/Document.cpp:

(WebCore::Document::setDomain):
(WebCore::Document::domainIsRegisterable const): Deleted.

  • dom/Document.h:
  • page/SecurityOrigin.cpp:

(WebCore::SecurityOrigin::isMatchingRegistrableDomainSuffix const):

  • page/SecurityOrigin.h:

Tools:

  • TestWebKitAPI/Tests/WebCore/SecurityOrigin.cpp:

(TestWebKitAPI::TEST_F):

2:40 PM Changeset in webkit [244852] by aakash_jain@apple.com
  • 2 edits in trunk/Tools

Remove webkitpy and bindings EWS queues from dashboard
https://bugs.webkit.org/show_bug.cgi?id=197475

Reviewed by Alexey Proskuryakov.

  • BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BubbleQueueServer.js:
2:38 PM Changeset in webkit [244851] by rniwa@webkit.org
  • 5 edits
    6 adds in trunk

[iOS] Element::focus and Element::scrollIntoView do not clamp scroll positions
https://bugs.webkit.org/show_bug.cgi?id=197211

Reviewed by Simon Fraser.

Source/WebCore:

Fixed the bug that Element::focus and Element::scrollIntoView were not clamping scroll offsets,
which causes scrollTop etc... to return a bogus negative scrolling offset.

Unfortunately, we can't just use FrameView's ScrollableArea::constrainScrollPosition since
scrollRectToVisible relies on the visible rect being expanded by the content insets in order to scroll to
a position within the content insets of UIScrollView; e.g. revealing the top of the page as the center.
We manually expand minimumScrollPosition() and maximumScrollPosition() by the content insets instead.

Tests: fast/scrolling/ios/programmatic-scroll-via-focus-should-clamp-top.html

fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-inside-iframe-should-clamp-top.html
fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-should-clamp-top.html

  • rendering/RenderLayer.cpp:

(WebCore::RenderLayer::scrollRectToVisible):

LayoutTests:

Added tests for scrolling via Element.prototype.focus and Element.prototype.scrollIntoView.

  • fast/scrolling/ios/programmatic-scroll-via-focus-should-clamp-top-expected.txt: Added.
  • fast/scrolling/ios/programmatic-scroll-via-focus-should-clamp-top.html: Added.
  • fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-inside-iframe-should-clamp-top-expected.txt: Added.
  • fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-inside-iframe-should-clamp-top.html: Added.
  • fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-should-clamp-top-expected.txt: Added.
  • fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-should-clamp-top.html: Added.
  • platform/ios-wk2/fast/dom/focus-contenteditable-expected.txt: Rebaselined the output now that the scroll position

is properly clamped.

  • platform/ios-wk2/imported/w3c/web-platform-tests/css/cssom-view/elementFromPoint-expected.txt: Rebaselined as one

more test case is passing.

2:19 PM Changeset in webkit [244850] by achristensen@apple.com
  • 2 edits in trunk/Source/WebKit

Protect against null crash in fetchDiskCacheEntries
https://bugs.webkit.org/show_bug.cgi?id=197399
<rdar://problem/47759337>

Reviewed by Antti Koivisto.

If the call to Entry::decodeStorageRecord returns nullptr in Storage::traverse, do not call the traverseHandler, which expects
that when it is called with nullptr that is the end of the traversal.

  • NetworkProcess/cache/NetworkCache.cpp:

(WebKit::NetworkCache::Cache::traverse):

2:08 PM Changeset in webkit [244849] by Wenson Hsieh
  • 17 edits
    10 adds in trunk

[iOS] Add a version of viewport shrink-to-fit heuristics that preserves page layout
https://bugs.webkit.org/show_bug.cgi?id=197342
<rdar://problem/50063091>

Reviewed by Tim Horton.

Source/WebCore:

Adds support for a new shrink-to-fit heuristic that attempts to lay out the contents of the page at a larger
width in order to shrink content to fit the viewport. See WebKit ChangeLog for more details.

Tests: fast/viewport/ios/shrink-to-fit-content-constant-width.html

fast/viewport/ios/shrink-to-fit-content-large-width-breakpoint.html
fast/viewport/ios/shrink-to-fit-content-no-viewport.html
fast/viewport/ios/shrink-to-fit-content-responsive-viewport-with-horizontal-overflow.html
fast/viewport/ios/shrink-to-fit-content-temporary-overflow.html

  • page/ViewportConfiguration.cpp:

(WebCore::ViewportConfiguration::setMinimumEffectiveDeviceWidth):
(WebCore::ViewportConfiguration::setIsKnownToLayOutWiderThanViewport):
(WebCore::ViewportConfiguration::description const):

  • page/ViewportConfiguration.h:

(WebCore::ViewportConfiguration::canIgnoreScalingConstraints const):
(WebCore::ViewportConfiguration::minimumEffectiveDeviceWidth const):

Add several new getters and setters in ViewportConfiguration.

(WebCore::ViewportConfiguration::isKnownToLayOutWiderThanViewport const):
(WebCore::ViewportConfiguration::shouldIgnoreMinimumEffectiveDeviceWidth const):

Importantly, only allow ignoring the minimum effective device width in webpages with responsive viewports, if
they also have *not* laid out wider than the viewport.

(WebCore::ViewportConfiguration::setForceAlwaysUserScalable):

Source/WebKit:

This patch introduces a new shrink-to-fit heuristic that attempts to lay out the contents of the page at a
larger width in order to shrink content to fit the viewport. This is similar to existing shrink-to-fit behaviors
used for viewport sizing in multitasking mode, except that it not only scales the view, but additionally expands
the layout size, such that the overall layout of the page is preserved. In fact, the reason we ended up
reverting the existing flavor of shrink-to-fit in all cases except for multitasking was that page layout was not
preserved, which caused elements that poke out of the viewport to make the rest of the page look out of
proportion — see <rdar://problem/23818102> and related radars.

Covered by 5 new layout tests, and by adjusting a couple of existing layout tests. See comments below for more
details.

  • Platform/Logging.h:

Add a new ViewportSizing logging channel. This will only log on pages that overflow the viewport and shrink to
fit as a result.

  • Shared/WebPreferences.yaml:

Turn IgnoreViewportScalingConstraints off by default. This preference currently controls whether we allow
shrink-to-fit behaviors, and is only used by Safari when it is in multitasking mode. The value of this
preference is currenly *on* by default, and is turned off almost immediately during every page load after the
first visible content rect update, wherein visibleContentRectUpdateInfo.allowShrinkToFit() is false.

However, this sometimes causes a brief jitter during page load; to fix this, make the default value for
IgnoreViewportScalingConstraints false, and change the logic in WebPage::updateVisibleContentRects to
setCanIgnoreScalingConstraints to true if either the IgnoreViewportScalingConstraints preference (not only
affected by an internal debug switch) is true, or WKWebView SPI is used to enable the behavior.

  • WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:

(WebKit::WebFrameLoaderClient::dispatchDidFinishDocumentLoad):
(WebKit::WebFrameLoaderClient::dispatchDidFinishLoad):

Add a new hook for WebFrameLoaderClient to call into WebPage when document load finishes. Also, tweak
dispatchDidFinishLoad to take a WebFrame& instead of a WebFrame* in a drive-by fix (the frame is assumed to be
non-null anyways).

  • WebProcess/WebPage/WebPage.cpp:

(WebKit::WebPage::didCommitLoad):
(WebKit::WebPage::didFinishDocumentLoad):
(WebKit::WebPage::didFinishLoad):

When finishing document load or finishing the overall load, kick off the shrink-to-fit timer; when committing a
load, cancel the timer.

  • WebProcess/WebPage/WebPage.h:
  • WebProcess/WebPage/ios/WebPageIOS.mm:

(WebKit::WebPage::setViewportConfigurationViewLayoutSize):

Don't allow the minimum effective device width from the client to stomp over any minimum effective device width
set as a result of the new shrink-to-fit heuristic; on some pages that load quickly, this can result in a race
where the minimum effective device width (i.e. a value that lower-bounds the minimum layout width) is first set
by the shrink-to-fit heuristic, and then set to an incorrect value by the client.

In the near future, web view SPI used to set the minimum effective device width should actually be removed
altogether, since the new shrink-to-fit heuristic supersedes any need for the client to fiddle with the minimum
effective device width.

(WebKit::WebPage::dynamicViewportSizeUpdate):

When performing a dynamic viewport size update, additionally re-run the shrink-to-fit heuristic. This allows
the minimum layout size of the viewport to be updated, if necessary. An example of where this matters is when a
web page is *below* a tablet/desktop layout breakpoint in portrait device orientation, but then exceeds this
layout breakpoint in landscape orientation. In this scenario, rotating the device should swap between these two
page layouts.

(WebKit::WebPage::resetViewportDefaultConfiguration):
(WebKit::WebPage::scheduleShrinkToFitContent):
(WebKit::WebPage::shrinkToFitContentTimerFired):
(WebKit::WebPage::immediatelyShrinkToFitContent):

Leverage the existing capability for a viewport to have a "minimum effective device width" to grant the viewport
a larger layout size than it would normally have, and then scale down to fit within the bounds of the view. One
challenge with this overall approach is that laying out at a larger width may cause the page to lay out even
wider in response, which may actually worsen horizontal scrolling. To mitigate this, we only attempt to lay out
at the current content width once; if laying out at this width reduced the amount of horizontal scrolling by any
amount, then proceed with this layout width; otherwise, revert to the previous layout width.

(WebKit::WebPage::shouldIgnoreMetaViewport const):

Pull some common logic out into a readonly getter.

(WebKit::WebPage::updateVisibleContentRects):

See the comment below WebPreferences.yaml, above.

LayoutTests:

Introduces new layout tests, and adjusts some existing tests. See comments below.

  • fast/viewport/ios/shrink-to-fit-content-constant-width-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-constant-width.html: Added.

Add a new layout test to exercise the scenario where a constant width viewport narrower than the view is used.

  • fast/viewport/ios/shrink-to-fit-content-large-width-breakpoint-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-large-width-breakpoint.html: Added.

Add a new layout test to exercise the scenario where a responsive website that lays out larger than the view
width ends up with even more horizontal scrolling when laying out at the initial content width. In this
scenario, we shouldn't try to expand the viewport to try and encompass the content width, since that would only
induce even worse horizontal scrolling.

  • fast/viewport/ios/shrink-to-fit-content-no-viewport-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-no-viewport.html: Added.

Add a new layout test for the case where there is no viewport, but content lays out wider than the view.

  • fast/viewport/ios/shrink-to-fit-content-responsive-viewport-with-horizontal-overflow-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-responsive-viewport-with-horizontal-overflow.html: Added.

Add a new layout test for the case where the page has opted for a responsive viewport (device-width, initial
scale 1), but has laid out wider than the viewport anyways. In this case, we want to shrink the contents down to
fit inside the view.

  • fast/viewport/ios/shrink-to-fit-content-temporary-overflow-expected.txt: Added.
  • fast/viewport/ios/shrink-to-fit-content-temporary-overflow.html: Added.

Add a new layout test to exercise the case where, during page load, content width temporarily increases, and
then decreases such that it once again fits within the viewport. In this case, we don't want to expand the
viewport to be as wide as the large temporary width of the page.

  • fast/viewport/ios/width-is-device-width-overflowing-body-overflow-hidden-expected.txt:
  • fast/viewport/ios/width-is-device-width-overflowing-body-overflow-hidden.html:
  • fast/viewport/ios/width-is-device-width-overflowing-expected.txt:
  • fast/viewport/ios/width-is-device-width-overflowing.html:

Tweak these 2 existing layout tests to include "shrink-to-fit=no", to prevent the new heuristics from shrinking
the page to fit on device classes that use native viewports by default.

  • platform/ipad/fast/viewport/ios/width-is-device-width-overflowing-body-overflow-hidden-expected.txt:
  • platform/ipad/fast/viewport/ios/width-is-device-width-overflowing-expected.txt:
1:46 PM Changeset in webkit [244848] by Alan Bujtas
  • 4 edits
    2 adds in trunk

[iOS] Star rating is covered with a black circle when writing a review on Yelp
https://bugs.webkit.org/show_bug.cgi?id=197469
<rdar://problem/48094446>

Reviewed by Dean Jackson.

Source/WebCore:

This patch moves the background painting of the radio/checkbox form controls in checked state to RenderTheme.
It enables content authors to disable default appearance using -webkit-appearance: none (it is also inline with what we do on macOS).

Test: fast/forms/radio-and-checkbox-checked-with-no-appearance.html

  • css/html.css:

(input:matches([type="checkbox"], [type="radio"]):checked):

  • rendering/RenderThemeIOS.mm:

(WebCore::RenderThemeIOS::paintCheckboxDecorations):
(WebCore::RenderThemeIOS::paintRadioDecorations):

LayoutTests:

  • fast/forms/radio-and-checkbox-checked-with-no-appearance-expected.html: Added.
  • fast/forms/radio-and-checkbox-checked-with-no-appearance.html: Added.
1:44 PM Changeset in webkit [244847] by Alan Coon
  • 7 edits in tags/Safari-608.1.19.101/Source

Versioning.

1:42 PM Changeset in webkit [244846] by Alan Coon
  • 1 copy in tags/Safari-608.1.19.101

New tag.

1:36 PM Changeset in webkit [244845] by Alan Coon
  • 7 edits in tags/Safari-608.1.19.100/Source

Versioning.

1:35 PM Changeset in webkit [244844] by Ryan Haddad
  • 2 edits in trunk/LayoutTests

REGRESSION: Layout test imported/w3c/web-platform-tests/resource-timing/resource-timing-level1.sub.html is frequently failing on Mac-ews
https://bugs.webkit.org/show_bug.cgi?id=197473

Unreviewed test gardening.

  • platform/mac/TestExpectations:
1:34 PM Changeset in webkit [244843] by Alan Coon
  • 7 edits in tags/Safari-608.1.19.200/Source

Versioning.

1:31 PM Changeset in webkit [244842] by Alan Coon
  • 1 copy in tags/Safari-608.1.19.200

New tag.

1:30 PM Changeset in webkit [244841] by Alan Coon
  • 1 copy in tags/Safari-608.1.19.100

New tag.

1:10 PM Changeset in webkit [244840] by aakash_jain@apple.com
  • 2 edits in trunk/Tools

Remove webkitpy and bindings test queues from old EWS
https://bugs.webkit.org/show_bug.cgi?id=197468

Reviewed by Alexey Proskuryakov.

  • QueueStatusServer/config/queues.py:
12:40 PM Changeset in webkit [244839] by Truitt Savell
  • 2 edits in trunk/LayoutTests

Win10 failing or flakey tests need to be given expectations to allow for new EWS
https://bugs.webkit.org/show_bug.cgi?id=197471

Unreviewed test Gardening

  • platform/win/TestExpectations:
12:39 PM Changeset in webkit [244838] by Kocsen Chung
  • 7 edits in tags/Safari-608.1.20/Source

Versioning.

12:31 PM Changeset in webkit [244837] by commit-queue@webkit.org
  • 4 edits in trunk/Source/WebCore

REGRESSION (r244182): RenderingUpdate should not be scheduled for invisible pages
https://bugs.webkit.org/show_bug.cgi?id=197451

Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2019-05-01
Reviewed by Simon Fraser.

Before r244182, some web pages never need to schedule a RenderingUpdate.
Only pages with rAF callbacks, web animations, intersection and resize
observers needed to do so. After r244182, all pages have to schedule a
RenderingUpdate when a page rendering update is required.

When Safari opens, it create a 'blank' web page. The blank page will not
be visible unless the user selects to show the 'Empty page' in the new
tab. Although the blank page is not visible, the loader needs to resolveStyle()
which requires to scheduleLayerFlushNow().

We need to optimize this case: calling scheduleLayerFlushNow() for invisible
pages. We do that by checking if the page is visible before scheduling
the RenderingUpdate.

Also we need to change or get rid of scheduleLayerFlushNow() since its name
has become confusing. It suggests that it is going to schedule flushing
the layer 'now'. But after r244182, it does scheduleRenderingUpdate() first.
And when it fires, scheduleCompositingLayerFlush() will be called.

  • page/RenderingUpdateScheduler.cpp:

(WebCore::RenderingUpdateScheduler::scheduleRenderingUpdate):

  • rendering/RenderLayerCompositor.cpp:

(WebCore::RenderLayerCompositor::scheduleLayerFlush):
(WebCore::RenderLayerCompositor::didChangeVisibleRect):
(WebCore::RenderLayerCompositor::frameViewDidScroll):
(WebCore::RenderLayerCompositor::attachRootLayer):
(WebCore::RenderLayerCompositor::setLayerFlushThrottlingEnabled):
(WebCore::RenderLayerCompositor::layerFlushTimerFired):
(WebCore::RenderLayerCompositor::scheduleLayerFlushNow): Deleted.

  • rendering/RenderLayerCompositor.h:
12:17 PM Changeset in webkit [244836] by dino@apple.com
  • 2 edits in trunk/Source/WebKit

Link Previews that use WKImagePreviewViewController are not always scaled correctly
https://bugs.webkit.org/show_bug.cgi?id=197450
<rdar://problem/50357695>

Reviewed by Wenson Hsieh.

Tell the UIImageView that we create in WKImagePreviewViewController to use
a filling scaling method.

  • UIProcess/WKImagePreviewViewController.mm:

(-[WKImagePreviewViewController initWithCGImage:defaultActions:elementInfo:]):

11:26 AM Changeset in webkit [244835] by Kocsen Chung
  • 7 edits in trunk/Source

Versioning.

11:20 AM Changeset in webkit [244834] by Kocsen Chung
  • 1 copy in tags/Safari-608.1.20

Tag Safari-608.1.20.

10:48 AM Changeset in webkit [244833] by Ross Kirsling
  • 2 edits in trunk/JSTests

Unreviewed correction to Test262 expectations following r244828.

  • test262/expectations.yaml:
10:42 AM Changeset in webkit [244832] by stephan.szabo@sony.com
  • 4 edits in trunk/JSTests

Add memory-limited skipping to some tests generating very large strings
https://bugs.webkit.org/show_bug.cgi?id=197437

Reviewed by Ross Kirsling.

  • stress/StringObject-define-length-getter-rope-string-oom.js:
  • stress/create-error-out-of-memory-rope-string.js:
  • stress/string-16bit-repeat-overflow.js:
10:36 AM Changeset in webkit [244831] by Alan Coon
  • 2 edits in branches/safari-607-branch/Source/WebInspectorUI

Cherry-pick r244358. rdar://problem/50352725

REGRESSION(r238168) Web Inspector: <iframe src=...> request does not show up in Network Tab
https://bugs.webkit.org/show_bug.cgi?id=193505
<rdar://problem/47325957>

Reviewed by Devin Rousso.

  • UserInterface/Views/NetworkTableContentView.js: (WI.NetworkTableContentView): (WI.NetworkTableContentView.prototype._handleFrameWasAdded): Handle new frames by adding the frame's main resource.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244358 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10:36 AM Changeset in webkit [244830] by Alan Coon
  • 2 edits in branches/safari-607-branch/Source/WebKit

Cherry-pick r244326. rdar://problem/50352719

Web Inspector: CRASH when reopening tab with docked inspector on crashed page
https://bugs.webkit.org/show_bug.cgi?id=196954
<rdar://problem/48716433>

Reviewed by Ryosuke Niwa.

  • UIProcess/mac/WebInspectorProxyMac.mm: (WebKit::WebInspectorProxy::inspectedViewFrameDidChange):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244326 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10:36 AM Changeset in webkit [244829] by Alan Coon
  • 2 edits in branches/safari-607-branch/Source/WebInspectorUI

Cherry-pick r242772. rdar://problem/50352729

Web Inspector: REGRESSION: Network Cookies Table does not load
https://bugs.webkit.org/show_bug.cgi?id=195599

Patch by Joseph Pecoraro <Joseph Pecoraro> on 2019-03-11
Reviewed by Devin Rousso.

  • UserInterface/Views/ResourceCookiesContentView.js: (WI.ResourceCookiesContentView.prototype.tableIndexForRepresentedObject): (WI.ResourceCookiesContentView.prototype.tableRepresentedObjectForIndex): Include needed delegate methods.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@242772 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10:33 AM Changeset in webkit [244828] by Ryan Haddad
  • 30 edits
    1 add in trunk

WebKit has too much of its own UTF-8 code and should rely more on ICU's UTF-8 support
https://bugs.webkit.org/show_bug.cgi?id=195535

Patch by Darin Adler <Darin Adler> on 2019-05-01
Reviewed by Alexey Proskuryakov.

LayoutTests/imported/w3c:

  • web-platform-tests/encoding/textdecoder-utf16-surrogates-expected.txt:

Updated expected results to have the Unicode replacement character in cases where the
text contains unpaired surrogates. The tests are still doing the same operations, and
still getting the same results, but the text output no longer includes illegal UTF-8.

Source/JavaScriptCore:

  • API/JSClassRef.cpp: Removed uneeded include of UTF8Conversion.h.
  • API/JSStringRef.cpp:

(JSStringCreateWithUTF8CString): Updated for changes to convertUTF8ToUTF16.
(JSStringGetUTF8CString): Updated for changes to convertLatin1ToUTF8.
Removed unneeded "true" to get the strict version of convertUTF16ToUTF8,
since that is the default. Also updated for changes to CompletionResult.

  • runtime/JSGlobalObjectFunctions.cpp:

(JSC::decode): Stop using UTF8SequenceLength, and instead use U8_COUNT_TRAIL_BYTES
and U8_MAX_LENGTH. Instead of decodeUTF8Sequence, use U8_NEXT. Also use U_IS_BMP,
U_IS_SUPPLEMENTARY, U16_LEAD, U16_TRAIL, and U_IS_SURROGATE instead of our own
equivalents, since these macros from ICU are correct and efficient.

  • wasm/WasmParser.h:

(JSC::Wasm::Parser<SuccessType>::consumeUTF8String): Updated for changes to
convertUTF8ToUTF16.

Source/WebCore:

  • platform/SharedBuffer.cpp:

(WebCore::utf8Buffer): Removed unnecessary "strict" argument to convertUTF16ToUTF8 since
that is the default behavior. Also updated for changes to return values.

  • xml/XSLTProcessorLibxslt.cpp:

(WebCore::writeToStringBuilder): Removed unnecessary use of StringBuffer for a temporary
buffer for characters. Rewrote to use U8_NEXT and U16_APPEND directly.

  • xml/parser/XMLDocumentParserLibxml2.cpp:

(WebCore::convertUTF16EntityToUTF8): Updated for changes to CompletionResult.

Source/WebKit:

  • Shared/API/APIString.h: Removed uneeded includes and also switched to #pragma once.
  • Shared/API/c/WKString.cpp: Moved include of UTF8Conversion.h here.

(WKStringGetUTF8CStringImpl): Updated for changes to return values.

Source/WTF:

  • wtf/text/AtomicString.cpp:

(WTF::AtomicString::fromUTF8Internal): Added code to compute string length when the
end is nullptr; this behavior used to be implemented inside the
calculateStringHashAndLengthFromUTF8MaskingTop8Bits function.

  • wtf/text/AtomicStringImpl.cpp:

(WTF::HashAndUTF8CharactersTranslator::translate): Updated for change to
convertUTF8ToUTF16.

  • wtf/text/AtomicStringImpl.h: Took the WTF_EXPORT_PRIVATE off of the

AtomicStringImpl::addUTF8 function. This is used only inside a non-inlined function in
the AtomicString class and its behavior changed subtly in this patch; it's helpful
to document that it's not exported.

  • wtf/text/StringImpl.cpp:

(WTF::StringImpl::utf8Impl): Don't pass "true" for strictness to convertUTF16ToUTF8
since strict is the default. Also updated for changes to ConversionResult.
(WTF::StringImpl::utf8ForCharacters): Updated for change to convertLatin1ToUTF8.
(WTF::StringImpl::tryGetUtf8ForRange const): Ditto.

  • wtf/text/StringView.cpp: Removed uneeded include of UTF8Conversion.h.
  • wtf/text/WTFString.cpp:

(WTF::String::fromUTF8): Updated for change to convertUTF8ToUTF16.

  • wtf/unicode/UTF8Conversion.cpp:

(WTF::Unicode::inlineUTF8SequenceLengthNonASCII): Deleted.
(WTF::Unicode::inlineUTF8SequenceLength): Deleted.
(WTF::Unicode::UTF8SequenceLength): Deleted.
(WTF::Unicode::decodeUTF8Sequence): Deleted.
(WTF::Unicode::convertLatin1ToUTF8): Use U8_APPEND, enabling us to remove
almost everything in the function. Also changed resturn value to be a boolean
to indicate success since there is only one possible failure (target exhausted).
There is room for further simplification, since most callers have lengths rather
than end pointers for the source buffer, and all but one caller supplies a buffer
size known to be sufficient, so those don't need a return value, nor do they need
to pass an end of buffer pointer.
(WTF::Unicode::convertUTF16ToUTF8): Use U_IS_LEAD, U_IS_TRAIL,
U16_GET_SUPPLEMENTARY, U_IS_SURROGATE, and U8_APPEND. Also changed behavior
for non-strict mode so that unpaired surrogates will be turned into the
replacement character instead of invalid UTF-8 sequences, because U8_APPEND
won't create an invalid UTF-8 sequence, and because we don't need to do that
for any good reason at any call site.
(WTF::Unicode::isLegalUTF8): Deleted.
(WTF::Unicode::readUTF8Sequence): Deleted.
(WTF::Unicode::convertUTF8ToUTF16): Use U8_NEXT instead of
inlineUTF8SequenceLength, isLegalUTF8, and readUTF8Sequence. Use
U16_APPEND instead of lots of code that does the same thing. There is
room for further simplification since most callers don't need the "all ASCII"
feature and could probably pass the arguments in a more natural way.
(WTF::Unicode::calculateStringHashAndLengthFromUTF8MaskingTop8Bits):
Use U8_NEXT instead of isLegalUTF8, readUTF8Sequence, and various
error handling checks for things that are handled by U8_NEXT. Also removed
support for passing nullptr for end to specify a null-terminated string.
(WTF::Unicode::equalUTF16WithUTF8): Ditto.

  • wtf/unicode/UTF8Conversion.h: Removed UTF8SequenceLength and

decodeUTF8Sequence. Changed the ConversionResult to match WebKit coding
style, with an eye toward perhaps removing it in the future. Changed
the convertUTF8ToUTF16 return value to a boolean and removed the "strict"
argument since no caller was passing false. Changed the convertLatin1ToUTF8
return value to a boolean. Tweaked comments.

LayoutTests:

  • css3/escape-dom-api-expected.txt:
  • fast/text/dangling-surrogates-expected.txt:
  • js/dom/webidl-type-mapping-expected.txt:
  • js/invalid-utf8-in-syntax-error-expected.txt:

Updated expected results to have the Unicode replacement character in cases where the
text contains unpaired surrogates. The tests are still doing the same operations, and
still getting the same results, but the text output no longer includes illegal UTF-8.

  • js/invalid-utf8-in-syntax-error.html: Added. Before adding this, the test was

run, but unlike the rest of the tests in this directory, was only run as part of
run-javascriptcore-tests. There are two reasons for adding this. One is to be
consistent with the rest of the tests here and run a second time as part of the
broader WebKit tests. The second is that we can now use "--reset-results" to generate
new expected results, something that run-webkit-tests has but run-javascriptcore-tests
does not have.

10:13 AM Changeset in webkit [244827] by Shawn Roberts
  • 28 edits
    1 delete in trunk

Unreviewed, rolling out r244821.

LayoutTests/imported/w3c:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

Source/JavaScriptCore:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

Source/WebCore:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

Source/WebKit:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

Source/WTF:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

LayoutTests:

Causing 4 Test262 failures on JSC Release and Debug

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

10:12 AM Changeset in webkit [244826] by Shawn Roberts
  • 5 edits in trunk/Source

Unreviewed, rolling out r244822.

Source/WebCore:

Causing

Reverted changeset:

https://trac.webkit.org/changeset/244822

Source/WTF:

Causing 4 Test262 failures on JSC Release and Debug

Reverted changeset:

https://trac.webkit.org/changeset/244822 https://trac.webkit.org/changeset/244821

10:09 AM Changeset in webkit [244825] by youenn@apple.com
  • 6 edits
    5 adds in trunk

Reject/throw when calling AudioContext methods on a stopped AudioContext
https://bugs.webkit.org/show_bug.cgi?id=197391

Reviewed by Eric Carlson.

Source/WebCore:

Return InvalidStateError in that case.
ASSERT that we do not call lazyInitialize after being stopped
since this would mean we are doing unneeded processing.

Test: http/wpt/webaudio/audiocontext-stopped.html

  • Modules/webaudio/AudioContext.cpp:

(WebCore::AudioContext::lazyInitialize):
(WebCore::AudioContext::createBufferSource):
(WebCore::AudioContext::createMediaElementSource):
(WebCore::AudioContext::createMediaStreamSource):
(WebCore::AudioContext::createMediaStreamDestination):
(WebCore::AudioContext::createScriptProcessor):
(WebCore::AudioContext::createBiquadFilter):
(WebCore::AudioContext::createWaveShaper):
(WebCore::AudioContext::createPanner):
(WebCore::AudioContext::createConvolver):
(WebCore::AudioContext::createDynamicsCompressor):
(WebCore::AudioContext::createAnalyser):
(WebCore::AudioContext::createGain):
(WebCore::AudioContext::createDelay):
(WebCore::AudioContext::createChannelSplitter):
(WebCore::AudioContext::createChannelMerger):
(WebCore::AudioContext::createOscillator):
(WebCore::AudioContext::createPeriodicWave):
(WebCore::AudioContext::startRendering):
(WebCore::AudioContext::suspend):
(WebCore::AudioContext::resume):
(WebCore::AudioContext::close):

  • Modules/webaudio/AudioContext.h:
  • Modules/webaudio/AudioContext.idl:

LayoutTests:

  • http/wpt/webaudio/audiocontext-stopped-expected.txt: Added.
  • http/wpt/webaudio/audiocontext-stopped.html: Added.
  • http/wpt/webaudio/resources/audiocontext-stopped-iframe.html: Added.
  • platform/win/TestExpectations: Skip test for win.
9:34 AM Changeset in webkit [244824] by youenn@apple.com
  • 5 edits
    1 copy
    1 add in trunk

LayoutTests/imported/w3c:
Kept alive loaders should use the redirected request in case of redirections
https://bugs.webkit.org/show_bug.cgi?id=197337

Reviewed by Alex Christensen.

  • web-platform-tests/fetch/api/basic/keepalive-expected.txt:

Source/WebKit:
Kept alive loaders should use the redirected request in case of redirections
https://bugs.webkit.org/show_bug.cgi?id=197337

Reviewed by Alex Christensen.

Instead of using the request, the redirected request should be used in case of redirection.

  • NetworkProcess/NetworkResourceLoader.cpp:

(WebKit::m_shouldCaptureExtraNetworkLoadMetrics):
(WebKit::NetworkResourceLoader::continueWillSendRedirectedRequest):

LayoutTests:
Enable Fetch Keep Alive by default
https://bugs.webkit.org/show_bug.cgi?id=197331

Reviewed by Alex Christensen.

  • platform/mac-wk1/imported/w3c/web-platform-tests/fetch/api/basic/keepalive-expected.txt: Added.
8:56 AM Changeset in webkit [244823] by eric.carlson@apple.com
  • 15 edits
    2 adds in trunk

XMLHttpRequest should propagate user gestures for media playback
https://bugs.webkit.org/show_bug.cgi?id=197428
<rdar://problem/46677392>

Reviewed by Jer Noble.

Source/WebCore:

A user gesture the would allow media state change in effect when XMLHttpRequest.send is
called should be active when the event handlers fire after the transaction completes successfully.

Test: http/tests/media/user-gesture-preserved-across-xmlhttprequest.html

  • dom/UserGestureIndicator.cpp:

(WebCore::UserGestureIndicator::UserGestureIndicator): Add a 'scope' parameter to potentially
limit the scope of the gesture to just media.
(WebCore::UserGestureIndicator::~UserGestureIndicator): Clear the scope.

  • dom/UserGestureIndicator.h:

(WebCore::UserGestureToken::processingUserGesture const):
(WebCore::UserGestureToken::setScope):
(WebCore::UserGestureToken::resetScope):
(WebCore::UserGestureToken::hasExpired const):

  • page/DOMTimer.cpp:

(WebCore::DOMTimerFireState::DOMTimerFireState): Don't need to store the nested timer interval,
UserGestureIndicator knows when it started.
(WebCore::DOMTimer::DOMTimer): Ditto.
(WebCore::DOMTimer::fired): Ditto.
(WebCore::DOMTimerFireState::nestedTimerInterval const): Deleted.
(WebCore::shouldForwardUserGesture): Deleted.
(WebCore::userGestureTokenToForward): Deleted.
(WebCore::currentNestedTimerInterval): Deleted.

  • page/DOMTimer.h:
  • testing/Internals.cpp:

(WebCore::Internals::setXHRMaximumIntervalForUserGestureForwarding): Override the maximum
user gesture interval for testing.

  • testing/Internals.h:
  • testing/Internals.idl:
  • xml/XMLHttpRequest.cpp:

(WebCore::XMLHttpRequest::XMLHttpRequest):
(WebCore::XMLHttpRequest::send): Stash the user gesture token.
(WebCore::XMLHttpRequest::dispatchEvent): Clear user gesture token if it has expired. If still
valid, activate it.

  • xml/XMLHttpRequest.h:

LayoutTests:

  • fast/events/popup-blocking-timers5-expected.txt:
  • fast/events/popup-blocking-timers5.html: Decrease the timer interval from 1000 to 900

because the user gesture is invalidated based on wall clock time.

  • fast/events/popup-blocking-timers6-expected.txt:
  • fast/events/popup-blocking-timers6.html: Increase the timer interval from 1001 to 1100

because the user gesture is invalidated based on wall clock time.

  • http/tests/media/user-gesture-preserved-across-xmlhttprequest-expected.txt: Added.
  • http/tests/media/user-gesture-preserved-across-xmlhttprequest.html: Added.
8:55 AM Changeset in webkit [244822] by Darin Adler
  • 3 edits in trunk/Source

Fix typos: "convering" to "converting".

8:52 AM Changeset in webkit [244821] by Darin Adler
  • 28 edits
    1 add in trunk

WebKit has too much of its own UTF-8 code and should rely more on ICU's UTF-8 support
https://bugs.webkit.org/show_bug.cgi?id=195535

Reviewed by Alexey Proskuryakov.

LayoutTests/imported/w3c:

  • web-platform-tests/encoding/textdecoder-utf16-surrogates-expected.txt:

Updated expected results to have the Unicode replacement character in cases where the
text contains unpaired surrogates. The tests are still doing the same operations, and
still getting the same results, but the text output no longer includes illegal UTF-8.

Source/JavaScriptCore:

  • API/JSClassRef.cpp: Removed uneeded include of UTF8Conversion.h.
  • API/JSStringRef.cpp:

(JSStringCreateWithUTF8CString): Updated for changes to convertUTF8ToUTF16.
(JSStringGetUTF8CString): Updated for changes to convertLatin1ToUTF8.
Removed unneeded "true" to get the strict version of convertUTF16ToUTF8,
since that is the default. Also updated for changes to CompletionResult.

  • runtime/JSGlobalObjectFunctions.cpp:

(JSC::decode): Stop using UTF8SequenceLength, and instead use U8_COUNT_TRAIL_BYTES
and U8_MAX_LENGTH. Instead of decodeUTF8Sequence, use U8_NEXT. Also use U_IS_BMP,
U_IS_SUPPLEMENTARY, U16_LEAD, U16_TRAIL, and U_IS_SURROGATE instead of our own
equivalents, since these macros from ICU are correct and efficient.

  • wasm/WasmParser.h:

(JSC::Wasm::Parser<SuccessType>::consumeUTF8String): Updated for changes to
convertUTF8ToUTF16.

Source/WebCore:

  • platform/SharedBuffer.cpp:

(WebCore::utf8Buffer): Removed unnecessary "strict" argument to convertUTF16ToUTF8 since
that is the default behavior. Also updated for changes to return values.

  • xml/XSLTProcessorLibxslt.cpp:

(WebCore::writeToStringBuilder): Removed unnecessary use of StringBuffer for a temporary
buffer for characters. Rewrote to use U8_NEXT and U16_APPEND directly.

  • xml/parser/XMLDocumentParserLibxml2.cpp:

(WebCore::convertUTF16EntityToUTF8): Updated for changes to CompletionResult.

Source/WebKit:

  • Shared/API/APIString.h: Removed uneeded includes and also switched to #pragma once.
  • Shared/API/c/WKString.cpp: Moved include of UTF8Conversion.h here.

(WKStringGetUTF8CStringImpl): Updated for changes to return values.

Source/WTF:

  • wtf/text/AtomicString.cpp:

(WTF::AtomicString::fromUTF8Internal): Added code to compute string length when the
end is nullptr; this behavior used to be implemented inside the
calculateStringHashAndLengthFromUTF8MaskingTop8Bits function.

  • wtf/text/AtomicStringImpl.cpp:

(WTF::HashAndUTF8CharactersTranslator::translate): Updated for change to
convertUTF8ToUTF16.

  • wtf/text/AtomicStringImpl.h: Took the WTF_EXPORT_PRIVATE off of the

AtomicStringImpl::addUTF8 function. This is used only inside a non-inlined function in
the AtomicString class and its behavior changed subtly in this patch; it's helpful
to document that it's not exported.

  • wtf/text/StringImpl.cpp:

(WTF::StringImpl::utf8Impl): Don't pass "true" for strictness to convertUTF16ToUTF8
since strict is the default. Also updated for changes to ConversionResult.
(WTF::StringImpl::utf8ForCharacters): Updated for change to convertLatin1ToUTF8.
(WTF::StringImpl::tryGetUtf8ForRange const): Ditto.

  • wtf/text/StringView.cpp: Removed uneeded include of UTF8Conversion.h.
  • wtf/text/WTFString.cpp:

(WTF::String::fromUTF8): Updated for change to convertUTF8ToUTF16.

  • wtf/unicode/UTF8Conversion.cpp:

(WTF::Unicode::inlineUTF8SequenceLengthNonASCII): Deleted.
(WTF::Unicode::inlineUTF8SequenceLength): Deleted.
(WTF::Unicode::UTF8SequenceLength): Deleted.
(WTF::Unicode::decodeUTF8Sequence): Deleted.
(WTF::Unicode::convertLatin1ToUTF8): Use U8_APPEND, enabling us to remove
almost everything in the function. Also changed resturn value to be a boolean
to indicate success since there is only one possible failure (target exhausted).
There is room for further simplification, since most callers have lengths rather
than end pointers for the source buffer, and all but one caller supplies a buffer
size known to be sufficient, so those don't need a return value, nor do they need
to pass an end of buffer pointer.
(WTF::Unicode::convertUTF16ToUTF8): Use U_IS_LEAD, U_IS_TRAIL,
U16_GET_SUPPLEMENTARY, U_IS_SURROGATE, and U8_APPEND. Also changed behavior
for non-strict mode so that unpaired surrogates will be turned into the
replacement character instead of invalid UTF-8 sequences, because U8_APPEND
won't create an invalid UTF-8 sequence, and because we don't need to do that
for any good reason at any call site.
(WTF::Unicode::isLegalUTF8): Deleted.
(WTF::Unicode::readUTF8Sequence): Deleted.
(WTF::Unicode::convertUTF8ToUTF16): Use U8_NEXT instead of
inlineUTF8SequenceLength, isLegalUTF8, and readUTF8Sequence. Use
U16_APPEND instead of lots of code that does the same thing. There is
room for further simplification since most callers don't need the "all ASCII"
feature and could probably pass the arguments in a more natural way.
(WTF::Unicode::calculateStringHashAndLengthFromUTF8MaskingTop8Bits):
Use U8_NEXT instead of isLegalUTF8, readUTF8Sequence, and various
error handling checks for things that are handled by U8_NEXT. Also removed
support for passing nullptr for end to specify a null-terminated string.
(WTF::Unicode::equalUTF16WithUTF8): Ditto.

  • wtf/unicode/UTF8Conversion.h: Removed UTF8SequenceLength and

decodeUTF8Sequence. Changed the ConversionResult to match WebKit coding
style, with an eye toward perhaps removing it in the future. Changed
the convertUTF8ToUTF16 return value to a boolean and removed the "strict"
argument since no caller was passing false. Changed the convertLatin1ToUTF8
return value to a boolean. Tweaked comments.

LayoutTests:

  • css3/escape-dom-api-expected.txt:
  • fast/text/dangling-surrogates-expected.txt:
  • js/dom/webidl-type-mapping-expected.txt:
  • js/invalid-utf8-in-syntax-error-expected.txt:

Updated expected results to have the Unicode replacement character in cases where the
text contains unpaired surrogates. The tests are still doing the same operations, and
still getting the same results, but the text output no longer includes illegal UTF-8
because the WTF changes affect the code path that DumpRenderTree and WebKitTestRunner
use to produce the text output.

  • js/invalid-utf8-in-syntax-error.html: Added. Before adding this, the test was

run, but unlike the rest of the tests in this directory, was only run as part of
run-javascriptcore-tests. There are two reasons for adding this. One is to be
consistent with the rest of the tests here and run a second time as part of the
broader WebKit tests. The second is that we can now use "--reset-results" to generate
new expected results, something that run-webkit-tests has but run-javascriptcore-tests
does not have.

7:20 AM WebKitGTK/2.24.x edited by Michael Catanzaro
(diff)

Apr 30, 2019:

11:30 PM Changeset in webkit [244820] by timothy_horton@apple.com
  • 3 edits in trunk/Tools

Fix the WebKitTestRunner build
https://bugs.webkit.org/show_bug.cgi?id=197449
<rdar://problem/50334169>

Reviewed by Alexey Proskuryakov.

  • WebKitTestRunner/mac/EventSenderProxy.mm:

(-[EventSenderSyntheticEvent initPressureEventAtLocation:globalLocation:stage:pressure:stageTransition:phase:time:eventNumber:window:]):
(WTR::EventSenderProxy::mouseForceChanged):
Use some SPI instead of IPI.

7:45 PM Changeset in webkit [244819] by Chris Dumez
  • 8 edits in trunk

Regression(PSON) URL scheme handlers can no longer respond asynchronously
https://bugs.webkit.org/show_bug.cgi?id=197426
<rdar://problem/50256169>

Reviewed by Brady Eidson.

Source/WebKit:

The issue was that when committing the provisional process, we would call WebPageProxy::processDidTerminate()
which would call WebPageProxy::stopAllURLSchemeTasks(). This would terminate all URL scheme tasks associated
with the page, including the one associated with the provisisional page / process.

To address the issue, pass m_process to stopAllURLSchemeTasks() in processDidTerminate() and only stop the
tasks associated with the m_process (which is the process we're about to swap away from).

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::processDidTerminate):
(WebKit::WebPageProxy::stopAllURLSchemeTasks):

  • UIProcess/WebPageProxy.h:
  • UIProcess/WebURLSchemeHandler.cpp:

(WebKit::WebURLSchemeHandler::processForTaskIdentifier):
(WebKit::WebURLSchemeHandler::stopAllTasksForPage):

  • UIProcess/WebURLSchemeHandler.h:
  • UIProcess/WebURLSchemeTask.h:

(WebKit::WebURLSchemeTask::process const):

Tools:

Add API test coverage.

  • TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

(-[PSONScheme setShouldRespondAsynchronously:]):
(-[PSONScheme webView:startURLSchemeTask:]):
(-[PSONScheme webView:stopURLSchemeTask:]):

7:28 PM Changeset in webkit [244818] by wilander@apple.com
  • 19 edits in trunk/Source

Add logging of Ad Click Attribution errors and events to a dedicated channel
https://bugs.webkit.org/show_bug.cgi?id=197332
<rdar://problem/49918800>

Reviewed by Youenn Fablet.

Source/WebCore:

This patch adds an experimental Ad Click Attribution debug mode which
logs information.

No new tests.

  • loader/AdClickAttribution.cpp:

(WebCore::AdClickAttribution::parseConversionRequest):
(WebCore::AdClickAttribution::debugModeEnabled):

  • loader/AdClickAttribution.h:
  • page/RuntimeEnabledFeatures.h:

(WebCore::RuntimeEnabledFeatures::adClickAttributionDebugModeEnabled const):
(WebCore::RuntimeEnabledFeatures::setAdClickAttributionDebugModeEnabled):

  • platform/Logging.h:

Source/WebKit:

This patch adds an experimental Ad Click Attribution debug mode which
logs information. Most changes are just log output in the various
functions in WebKit::AdClickAttributionManager.

The constructor to WebKit::AdClickAttributionManager now takes a
PAL::SessionID so that the log functions can make sure they don't
output anything in ephemeral sessions.

WebProcessPool::platformInitializeNetworkProcess() now picks up the
debug mode setting from the incoming
WebKit::NetworkProcessCreationParameters object.

NetworkResourceLoader::handleAdClickAttributionConversion() was
moved to AdClickAttributionManager::handleConversion() to keep all
the logging in one file.

  • NetworkProcess/AdClickAttributionManager.cpp:

(WebKit::AdClickAttributionManager::storeUnconverted):
(WebKit::AdClickAttributionManager::handleConversion):
(WebKit::AdClickAttributionManager::convert):
(WebKit::AdClickAttributionManager::fireConversionRequest):
(WebKit::AdClickAttributionManager::firePendingConversionRequests):
(WebKit::AdClickAttributionManager::clearExpired):
(WebKit::AdClickAttributionManager::debugModeEnabled const):

  • NetworkProcess/AdClickAttributionManager.h:

(WebKit::AdClickAttributionManager::AdClickAttributionManager):
(WebKit::AdClickAttributionManager::m_sessionID):

  • NetworkProcess/NetworkProcess.cpp:

(WebKit::NetworkProcess::initializeNetworkProcess):

  • NetworkProcess/NetworkProcessCreationParameters.cpp:

(WebKit::NetworkProcessCreationParameters::encode const):
(WebKit::NetworkProcessCreationParameters::decode):

  • NetworkProcess/NetworkProcessCreationParameters.h:
  • NetworkProcess/NetworkResourceLoader.cpp:

(WebKit::NetworkResourceLoader::continueWillSendRedirectedRequest):
(WebKit::NetworkResourceLoader::handleAdClickAttributionConversion): Deleted.

  • NetworkProcess/NetworkSession.cpp:

(WebKit::NetworkSession::NetworkSession):
(WebKit::NetworkSession::handleAdClickAttributionConversion):
(WebKit::NetworkSession::convertAdClickAttribution): Deleted.

  • NetworkProcess/NetworkSession.h:
  • Platform/Logging.h:
  • Shared/WebPreferences.yaml:
  • UIProcess/Cocoa/WebProcessPoolCocoa.mm:

(WebKit::WebProcessPool::platformInitializeNetworkProcess):

Source/WTF:

Added missing RELEASE_LOG_INFO and RELEASE_LOG_INFO_IF dummies
for RELEASE_LOG_DISABLED.

  • wtf/Assertions.h:
6:14 PM Changeset in webkit [244817] by mmaxfield@apple.com
  • 4 edits
    2 adds in trunk

font-weight: 1000 is not parsed successfully
https://bugs.webkit.org/show_bug.cgi?id=197427

Reviewed by Dean Jackson.

Source/WebCore:

The spec says:
"Only values greater than or equal to 1, and less than or equal to 1000, are valid"

This change brings us in-line with all the other browsers.

Test: fast/text/font-weight-1-1000.html

  • css/parser/CSSPropertyParserHelpers.cpp:

(WebCore::CSSPropertyParserHelpers::consumeFontWeightNumber):

LayoutTests:

  • platform/win/TestExpectations: Windows disables variation fonts, so requires values to be divisible by 100
  • fast/text/font-weight-1-1000-expected.txt: Added.
  • fast/text/font-weight-1-1000.html: Added.
5:32 PM Changeset in webkit [244816] by commit-queue@webkit.org
  • 10 edits
    1 delete in trunk

Unreviewed, rolling out r244806.
https://bugs.webkit.org/show_bug.cgi?id=197446

Causing Test262 and JSC test failures on multiple builds
(Requested by ShawnRoberts on #webkit).

Reverted changeset:

"TypeArrays should not store properties that are canonical
numeric indices"
https://bugs.webkit.org/show_bug.cgi?id=197228
https://trac.webkit.org/changeset/244806

5:13 PM Changeset in webkit [244815] by youenn@apple.com
  • 9 edits in trunk/Source

Make Document audio producers use WeakPtr
https://bugs.webkit.org/show_bug.cgi?id=197382

Reviewed by Eric Carlson.

Source/WebCore:

Move from a hash set of raw pointers to a hash set of weak pointers.
This helps make the code cleaner.
No observable change of behavior.

  • Modules/mediastream/MediaStreamTrack.h:
  • dom/Document.cpp:

(WebCore::Document::addAudioProducer):
(WebCore::Document::removeAudioProducer):
(WebCore::Document::updateIsPlayingMedia):
(WebCore::Document::pageMutedStateDidChange):

  • dom/Document.h:
  • html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::updateActiveTextTrackCues):

  • html/HTMLMediaElement.h:
  • page/MediaProducer.h:

Source/WTF:

  • wtf/WeakHashSet.h:

(WTF::WeakHashSet::hasNullReferences const):

5:12 PM Changeset in webkit [244814] by Devin Rousso
  • 3 edits in trunk/Source/WebKit

Crash when running test wpt/tests/element_click/bubbling.py::test_element_disappears_during_click
https://bugs.webkit.org/show_bug.cgi?id=197361
<rdar://problem/49861407>

Reviewed by Brian Burg.

Don't assume that all elements have client rects (e.g. getClientRects). If the container
element isn't visible, then the child won't be either.

  • UIProcess/Automation/SimulatedInputDispatcher.cpp:

(WebKit::SimulatedInputDispatcher::resolveLocation):
(WebKit::SimulatedInputDispatcher::transitionInputSourceToState):

  • UIProcess/Automation/WebAutomationSession.cpp:

(WebKit::WebAutomationSession::viewportInViewCenterPointOfElement):

4:56 PM Changeset in webkit [244813] by youenn@apple.com
  • 6 edits in trunk

[macOS WK1] ASSERTION FAILED: formData in WebCore::ResourceRequest::doUpdateResourceHTTPBody()
https://bugs.webkit.org/show_bug.cgi?id=196864
<rdar://problem/49854497>

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

  • web-platform-tests/xhr/send-redirect-post-upload-expected.txt:

Source/WebCore:

In case of redirection, it is sometimes not possible to retrieve the form data
from its NSInputStream in case of redirections.
To handle this case, reuse the first request form data if the new request has a body.
We also clear the HTTP content type in such a case if the original request has no content type.

Covered by re-enabled tests.

  • platform/network/mac/WebCoreResourceHandleAsOperationQueueDelegate.mm:

(-[WebCoreResourceHandleAsOperationQueueDelegate connection:willSendRequest:redirectResponse:]):

LayoutTests:

  • platform/mac-wk1/TestExpectations:
4:56 PM Changeset in webkit [244812] by Chris Dumez
  • 8 edits in trunk

Unreviewed, rolling out r244802.

Caused an API test failure

Reverted changeset:

"Regression(PSON) URL scheme handlers can no longer respond
asynchronously"
https://bugs.webkit.org/show_bug.cgi?id=197426
https://trac.webkit.org/changeset/244802

4:37 PM Changeset in webkit [244811] by sbarati@apple.com
  • 19 edits in trunk/Source/JavaScriptCore

CodeBlock::m_instructionCount is wrong
https://bugs.webkit.org/show_bug.cgi?id=197304

Reviewed by Yusuke Suzuki.

What we were calling instructionCount() was wrong, as evidenced by
us using it incorrectly both in the sampling profiler and when we
dumped bytecode for a given CodeBlock. Prior to the bytecode rewrite,
instructionCount() was probably valid to do bounds checks against.
However, this is no longer the case. This patch renames what we called
instructionCount() to bytecodeCost(). It is now only used to make decisions
about inlining and tier up heuristics. I've also named options related to
this appropriately.

This patch also introduces instructionsSize(). The result of this method
is valid to do bounds checks against.

  • bytecode/CodeBlock.cpp:

(JSC::CodeBlock::dumpAssumingJITType const):
(JSC::CodeBlock::CodeBlock):
(JSC::CodeBlock::finishCreation):
(JSC::CodeBlock::optimizationThresholdScalingFactor):
(JSC::CodeBlock::predictedMachineCodeSize):

  • bytecode/CodeBlock.h:

(JSC::CodeBlock::instructionsSize const):
(JSC::CodeBlock::bytecodeCost const):
(JSC::CodeBlock::instructionCount const): Deleted.

  • dfg/DFGByteCodeParser.cpp:

(JSC::DFG::ByteCodeParser::inliningCost):
(JSC::DFG::ByteCodeParser::getInliningBalance):

  • dfg/DFGCapabilities.cpp:

(JSC::DFG::mightCompileEval):
(JSC::DFG::mightCompileProgram):
(JSC::DFG::mightCompileFunctionForCall):
(JSC::DFG::mightCompileFunctionForConstruct):
(JSC::DFG::mightInlineFunctionForCall):
(JSC::DFG::mightInlineFunctionForClosureCall):
(JSC::DFG::mightInlineFunctionForConstruct):

  • dfg/DFGCapabilities.h:

(JSC::DFG::isSmallEnoughToInlineCodeInto):

  • dfg/DFGDisassembler.cpp:

(JSC::DFG::Disassembler::dumpHeader):

  • dfg/DFGDriver.cpp:

(JSC::DFG::compileImpl):

  • dfg/DFGPlan.cpp:

(JSC::DFG::Plan::compileInThread):

  • dfg/DFGTierUpCheckInjectionPhase.cpp:

(JSC::DFG::TierUpCheckInjectionPhase::run):

  • ftl/FTLCapabilities.cpp:

(JSC::FTL::canCompile):

  • ftl/FTLCompile.cpp:

(JSC::FTL::compile):

  • ftl/FTLLink.cpp:

(JSC::FTL::link):

  • jit/JIT.cpp:

(JSC::JIT::link):

  • jit/JITDisassembler.cpp:

(JSC::JITDisassembler::dumpHeader):

  • llint/LLIntSlowPaths.cpp:

(JSC::LLInt::shouldJIT):

  • profiler/ProfilerBytecodes.cpp:

(JSC::Profiler::Bytecodes::Bytecodes):

  • runtime/Options.h:
  • runtime/SamplingProfiler.cpp:

(JSC::tryGetBytecodeIndex):
(JSC::SamplingProfiler::processUnverifiedStackTraces):

4:18 PM Changeset in webkit [244810] by Alan Coon
  • 7 edits in tags/Safari-608.1.19.5/Source

Versioning.

4:00 PM Changeset in webkit [244809] by Alan Coon
  • 1 copy in tags/Safari-608.1.19.5

New tag.

3:49 PM Changeset in webkit [244808] by Shawn Roberts
  • 2 edits in trunk/LayoutTests

imported/w3c/web-platform-tests/xhr/event-upload-progress-crossorigin.htm is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=182849

Unreviewed test gardening.

  • platform/ios-simulator-wk2/TestExpectations: Updating expectations for flaky failure on iOS
3:40 PM Changeset in webkit [244807] by Alan Coon
  • 1 copy in tags/Safari-607.2.6.1.1

Tag Safari-607.2.6.1.1.

3:25 PM Changeset in webkit [244806] by Tadeu Zagallo
  • 10 edits
    1 add in trunk

TypeArrays should not store properties that are canonical numeric indices
https://bugs.webkit.org/show_bug.cgi?id=197228
<rdar://problem/49557381>

Reviewed by Darin Adler.

JSTests:

  • stress/typed-array-canonical-numeric-index-string.js: Added.

(makeTest.assert):
(makeTest):
(const.testInvalidIndices.makeTest.set assert):
(const.testInvalidIndices.makeTest):
(const.testValidIndices.makeTest.set assert):
(const.testValidIndices.makeTest):

Source/JavaScriptCore:

According to the spec[1], TypedArrays should not perform an ordinary GetOwnProperty/SetOwnProperty
if the index is a CanonicalNumericIndexString, but invalid according toIntegerIndexedElementGet
and similar functions. I.e., there are a few properties that should not be set in a TypedArray,
like NaN, Infinity and -0.

[1]: https://www.ecma-international.org/ecma-262/9.0/index.html#sec-integer-indexed-exotic-objects-defineownproperty-p-desc

  • CMakeLists.txt:
  • JavaScriptCore.xcodeproj/project.pbxproj:
  • runtime/JSGenericTypedArrayViewInlines.h:

(JSC::JSGenericTypedArrayView<Adaptor>::getOwnPropertySlot):
(JSC::JSGenericTypedArrayView<Adaptor>::put):
(JSC::JSGenericTypedArrayView<Adaptor>::defineOwnProperty):
(JSC::JSGenericTypedArrayView<Adaptor>::getOwnPropertySlotByIndex):
(JSC::JSGenericTypedArrayView<Adaptor>::putByIndex):

  • runtime/JSTypedArrays.cpp:
  • runtime/PropertyName.h:

(JSC::canonicalNumericIndexString):

LayoutTests:

  • fast/canvas/canvas-ImageData-behaviour-expected.txt:
  • fast/canvas/canvas-ImageData-behaviour.js:
3:12 PM Changeset in webkit [244805] by Michael Catanzaro
  • 2 edits in trunk/Source/WebCore/platform/gtk/po

Update Spanish Translation
https://bugs.webkit.org/show_bug.cgi?id=197407

Patch by Álvaro Torralba <donfrutosgomez@gmail.com> on 2019-04-30
Rubber-stamped by Michael Catanzaro.

  • es.po:
2:24 PM Changeset in webkit [244804] by Alan Coon
  • 2 edits in branches/safari-607-branch/Source/WebCore

Cherry-pick r244727. rdar://problem/50344746

[Apple Pay] Increment the API version from 6 to 7
https://bugs.webkit.org/show_bug.cgi?id=197041
<rdar://problem/49986625>

Reviewed by Geoffrey Garen.

  • Modules/applepay/PaymentCoordinatorClient.cpp: (WebCore::PaymentCoordinatorClient::supportsVersion):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244727 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2:24 PM Changeset in webkit [244803] by Alan Coon
  • 2 edits in branches/safari-607-branch/Source/WTF

Cherry-pick r243641. rdar://problem/50232877

Unreviewed build fix.

  • wtf/CMakeLists.txt: Added SpanningTree.h to WTF_PUBLIC_HEADERS.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@243641 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2:11 PM Changeset in webkit [244802] by Chris Dumez
  • 8 edits in trunk

Regression(PSON) URL scheme handlers can no longer respond asynchronously
https://bugs.webkit.org/show_bug.cgi?id=197426
<rdar://problem/50256169>

Reviewed by Brady Eidson.

Source/WebKit:

The issue was that when committing the provisional process, we would call WebPageProxy::processDidTerminate()
which would call WebPageProxy::stopAllURLSchemeTasks(). This would terminate all URL scheme tasks associated
with the page, including the one associated with the provisisional page / process.

To address the issue, pass m_process to stopAllURLSchemeTasks() in processDidTerminate() and only stop the
tasks associated with the m_process (which is the process we're about to swap away from).

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::processDidTerminate):
(WebKit::WebPageProxy::stopAllURLSchemeTasks):

  • UIProcess/WebPageProxy.h:
  • UIProcess/WebURLSchemeHandler.cpp:

(WebKit::WebURLSchemeHandler::processForTaskIdentifier):
(WebKit::WebURLSchemeHandler::stopAllTasksForPage):

  • UIProcess/WebURLSchemeHandler.h:
  • UIProcess/WebURLSchemeTask.h:

(WebKit::WebURLSchemeTask::process const):

Tools:

Add API test coverage.

  • TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

(-[PSONScheme setShouldRespondAsynchronously:]):
(-[PSONScheme webView:startURLSchemeTask:]):
(-[PSONScheme webView:stopURLSchemeTask:]):

1:21 PM Changeset in webkit [244801] by commit-queue@webkit.org
  • 9 edits in trunk/Source

Unreviewed, rolling out r244773.
https://bugs.webkit.org/show_bug.cgi?id=197436

Causing assertion failures on debug queues (Requested by
ShawnRoberts on #webkit).

Reverted changeset:

"Make Document audio producers use WeakPtr"
https://bugs.webkit.org/show_bug.cgi?id=197382
https://trac.webkit.org/changeset/244773

1:20 PM Changeset in webkit [244800] by Simon Fraser
  • 46 edits
    3 adds in trunk

Transform is sometimes left in a bad state after an animation
https://bugs.webkit.org/show_bug.cgi?id=197401
rdar://problem/48179186

Reviewed by Dean Jackson.

Source/WebCore:

In some more complex compositing scenarios, at the end of an animation we'd
fail to push a new transform onto a layer, because updateGeometry() would
think there's an animation running (which there is, but in the "Ending" state).

It's simpler in this code to just always push transform and opacity to the layer;
they will get overridden by the animation while it's running. The current code
dates from the first landing of the file, and the reason for the if (!isRunningAcceleratedTransformAnimation)
check is lost in the sands of time.

I was not able to get a reliable ref or layer tree test for this, because the next compositing update
fixes it, and WTR seems to trigger one. But the added test does show the bug
in Safari, and is a good test to have.

Test: compositing/animation/transform-after-animation.html

  • rendering/RenderLayerBacking.cpp:

(WebCore::RenderLayerBacking::updateGeometry):

LayoutTests:

Share code between all the overlap tests that work by creating dot matrices, and strip
out the transforms from the layer tree dumps, because they can vary with timing in these
tests.

  • compositing/animation/transform-after-animation-expected.html: Added.
  • compositing/animation/transform-after-animation.html: Added.
  • compositing/backing/backing-store-attachment-empty-keyframe-expected.txt:
  • compositing/layer-creation/animation-overlap-with-children.html:
  • compositing/layer-creation/mismatched-rotated-transform-animation-overlap-expected.txt:
  • compositing/layer-creation/mismatched-rotated-transform-animation-overlap.html:
  • compositing/layer-creation/mismatched-rotated-transform-transition-overlap.html:
  • compositing/layer-creation/mismatched-transform-transition-overlap.html:
  • compositing/layer-creation/multiple-keyframes-animation-overlap-expected.txt:
  • compositing/layer-creation/multiple-keyframes-animation-overlap.html:
  • compositing/layer-creation/resources/compositing-overlap-utils.js: Added.

(makeDots):
(layerTreeWithoutTransforms):
(dumpLayers):

  • compositing/layer-creation/scale-rotation-animation-overlap-expected.txt:
  • compositing/layer-creation/scale-rotation-animation-overlap.html:
  • compositing/layer-creation/scale-rotation-transition-overlap.html:
  • compositing/layer-creation/translate-animation-overlap-expected.txt:
  • compositing/layer-creation/translate-animation-overlap.html:
  • compositing/layer-creation/translate-scale-animation-overlap-expected.txt:
  • compositing/layer-creation/translate-scale-animation-overlap.html:
  • compositing/layer-creation/translate-scale-transition-overlap.html:
  • compositing/layer-creation/translate-transition-overlap.html:
  • legacy-animation-engine/compositing/backing/backing-store-attachment-empty-keyframe-expected.txt:
  • legacy-animation-engine/compositing/backing/backing-store-attachment-empty-keyframe.html:
  • legacy-animation-engine/compositing/backing/transform-transition-from-outside-view-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/animation-overlap-with-children.html:
  • legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-animation-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-animation-overlap.html:
  • legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-transition-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-transition-overlap.html:
  • legacy-animation-engine/compositing/layer-creation/mismatched-transform-transition-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/mismatched-transform-transition-overlap.html:
  • legacy-animation-engine/compositing/layer-creation/multiple-keyframes-animation-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/multiple-keyframes-animation-overlap.html:
  • legacy-animation-engine/compositing/layer-creation/scale-rotation-animation-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/scale-rotation-animation-overlap.html:
  • legacy-animation-engine/compositing/layer-creation/scale-rotation-transition-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/scale-rotation-transition-overlap.html:
  • legacy-animation-engine/compositing/layer-creation/translate-animation-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/translate-animation-overlap.html:
  • legacy-animation-engine/compositing/layer-creation/translate-scale-animation-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/translate-scale-animation-overlap.html:
  • legacy-animation-engine/compositing/layer-creation/translate-scale-transition-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/translate-scale-transition-overlap.html:
  • legacy-animation-engine/compositing/layer-creation/translate-transition-overlap-expected.txt:
  • legacy-animation-engine/compositing/layer-creation/translate-transition-overlap.html:
  • platform/ios/compositing/layer-creation/scale-rotation-animation-overlap-expected.txt:
1:20 PM Changeset in webkit [244799] by Chris Dumez
  • 5 edits in trunk

Only use a related page's process if that page has not been closed yet
https://bugs.webkit.org/show_bug.cgi?id=197393
<rdar://problem/50302423>

Reviewed by Tim Horton.

Source/WebKit:

We should not attempt to use a related page's process if that related page has already been closed.
Once closed, a page's process is invalid and trying to launch a new process for the closed page
leads to crashes such as the one in the radar.

  • UIProcess/WebPageProxy.cpp:

(WebKit::WebPageProxy::launchProcess):

  • UIProcess/WebProcessPool.cpp:

(WebKit::WebProcessPool::createWebPage):

Tools:

Add API test coverage.

  • TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:
1:13 PM Changeset in webkit [244798] by Alan Coon
  • 38 edits
    4 adds
    4 deletes in branches/safari-607-branch

Revert "Versioning."
Revert "Cherry-pick r244632. rdar://problem/50344188"

This reverts r244788 and r244632 to reset the state of the 607 branch.

1:09 PM Changeset in webkit [244797] by commit-queue@webkit.org
  • 6 edits
    1 delete in trunk

Unreviewed, rolling out r244774.
https://bugs.webkit.org/show_bug.cgi?id=197431

Causing assertion failures on debug queues (Requested by
ShawnRoberts on #webkit).

Reverted changeset:

"Reject/throw when calling AudioContext methods on a stopped
AudioContext"
https://bugs.webkit.org/show_bug.cgi?id=197391
https://trac.webkit.org/changeset/244774

1:08 PM Changeset in webkit [244796] by Alan Coon
  • 2 edits in branches/safari-607.2.6.1-branch/Source/WebCore

Cherry-pick r244632. rdar://problem/50344188

Do not restart WebRTC stats timer if backend is stopped
https://bugs.webkit.org/show_bug.cgi?id=197257
<rdar://problem/50095879>

Reviewed by Eric Carlson.

We used to stop and reschedule the stat gathering timer in case the
gathering delay is changing. Timer should not be rescheduled if the backend is stopped.

  • Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp: (WebCore::LibWebRTCMediaEndpoint::OnStatsDelivered):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244632 268f45cc-cd09-0410-ab3c-d52691b4dbfc

1:05 PM Changeset in webkit [244795] by Alan Coon
  • 7 edits in branches/safari-607.2.6.1-branch/Source

Versioning.

12:42 PM Changeset in webkit [244794] by achristensen@apple.com
  • 9 edits
    6 copies in trunk

Add WKContentRuleList ping resource-type
https://bugs.webkit.org/show_bug.cgi?id=197325
<rdar://problem/49841404>

Reviewed by Geoff Garen.

Source/WebCore:

Tests: http/tests/contentextensions/block-ping-resource-type-ping.html and http/tests/contentextensions/block-ping-resource-type-raw.html

  • contentextensions/ContentExtensionsBackend.cpp:

(WebCore::ContentExtensions::ContentExtensionsBackend::processContentRuleListsForLoad):

  • contentextensions/ContentExtensionsBackend.h:
  • loader/PingLoader.cpp:

(WebCore::processContentRuleListsForLoad):
(WebCore::PingLoader::sendPing):

  • loader/ResourceLoadInfo.cpp:

(WebCore::ContentExtensions::readResourceType):
(WebCore::ContentExtensions::ResourceLoadInfo::getResourceFlags const):

  • loader/ResourceLoadInfo.h:
  • page/UserContentProvider.cpp:

(WebCore::UserContentProvider::processContentRuleListsForLoad):

  • page/UserContentProvider.h:

LayoutTests:

  • http/tests/contentextensions/block-ping-resource-type-raw--expected.txt: Copied from LayoutTests/http/tests/contentextensions/block-ping-expected.txt.
  • http/tests/contentextensions/block-ping-resource-type-raw.html: Copied from LayoutTests/http/tests/contentextensions/block-ping.html.
  • http/tests/contentextensions/block-ping-resource-type-raw.html.json: Copied from LayoutTests/http/tests/contentextensions/block-ping.html.json.
  • http/tests/contentextensions/block-ping-resource-type-ping-expected.txt: Copied from LayoutTests/http/tests/contentextensions/block-ping-expected.txt.
  • http/tests/contentextensions/block-ping-resource-type-ping.html: Copied from LayoutTests/http/tests/contentextensions/block-ping.html.
  • http/tests/contentextensions/block-ping-resource-type-ping.html.json: Copied from LayoutTests/http/tests/contentextensions/block-ping.html.json.
12:29 PM Changeset in webkit [244793] by Alan Coon
  • 2 edits in branches/safari-607-branch/Source/WebCore

Cherry-pick r244632. rdar://problem/50344188

Do not restart WebRTC stats timer if backend is stopped
https://bugs.webkit.org/show_bug.cgi?id=197257
<rdar://problem/50095879>

Reviewed by Eric Carlson.

We used to stop and reschedule the stat gathering timer in case the
gathering delay is changing. Timer should not be rescheduled if the backend is stopped.

  • Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp: (WebCore::LibWebRTCMediaEndpoint::OnStatsDelivered):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244632 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11:42 AM Changeset in webkit [244792] by aakash_jain@apple.com
  • 3 edits in trunk/Tools

[ews-build] Enable Bindings tests queue on new EWS
https://bugs.webkit.org/show_bug.cgi?id=197424

Reviewed by Lucas Forschler.

  • BuildSlaveSupport/ews-app/ews/views/statusbubble.py:

(StatusBubble):

  • BuildSlaveSupport/ews-build/config.json:
11:38 AM Changeset in webkit [244791] by BJ Burg
  • 3 edits in trunk/Source/JavaScriptCore

Web Automation: use a more informative key to indicate automation availability
https://bugs.webkit.org/show_bug.cgi?id=197377
<rdar://problem/50258069>

Reviewed by Devin Rousso.

The existing WIRAutomationEnabledKey does not encode uncertainty.
Add a new key that provides an 'Unknown' state, and prefer to use it.

Since an application's initial listing is sent from a background dispatch queue
on Cocoa platforms, this can race with main thread initialization that sets up
RemoteInspector::Client. Therefore, the initial listing may not properly represent
the client's capabilites because the client is not yet available. Allowing for
an "Unknown" state that is later upgraded to Available or Not Available makes it
possible to work around this potential race.

  • inspector/remote/RemoteInspectorConstants.h:
  • inspector/remote/cocoa/RemoteInspectorCocoa.mm:

(Inspector::RemoteInspector::pushListingsNow):

11:35 AM Changeset in webkit [244790] by aakash_jain@apple.com
  • 2 edits in trunk/Tools

[ews-build] Configure ews151 for running various tests
https://bugs.webkit.org/show_bug.cgi?id=197419

Reviewed by Lucas Forschler.

  • BuildSlaveSupport/ews-build/config.json:
11:33 AM Changeset in webkit [244789] by aakash_jain@apple.com
  • 3 edits in trunk/Tools

[ews-build] Parse and display bindings test failures
https://bugs.webkit.org/show_bug.cgi?id=197423

Reviewed by Lucas Forschler.

  • BuildSlaveSupport/ews-build/steps.py:

(RunBindingsTests.init): Set timeout of 5 minutes.
(RunBindingsTests.start): Initialize log_observer for json output.
(RunBindingsTests.getResultSummary): Update step and build summary based on bindings test results.
(RunBindingsTests._addToLog): Method to add message to log.

  • BuildSlaveSupport/ews-build/steps_unittest.py: Updated unit-test accordingly.
11:29 AM Changeset in webkit [244788] by Alan Coon
  • 7 edits in branches/safari-607-branch/Source

Versioning.

11:27 AM Changeset in webkit [244787] by Shawn Roberts
  • 2 edits in trunk/LayoutTests

Fixing typo in https://trac.webkit.org/changeset/244780/webkit.

Unreviewed test gardening.

  • platform/mac/TestExpectations:
11:18 AM Changeset in webkit [244786] by Shawn Roberts
  • 2 edits in trunk/LayoutTests

scrollingcoordinator/scrolling-tree/scrolling-tree-includes-frame.html is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=197425

Unreviewed test gardening.

  • platform/mac-wk2/TestExpectations: Updating expectations for flaky failure
11:09 AM Changeset in webkit [244785] by Alan Coon
  • 7 edits in tags/Safari-608.1.19.4/Source

Versioning.

11:07 AM Changeset in webkit [244784] by Alan Coon
  • 1 copy in tags/Safari-608.1.19.4

New tag.

10:57 AM Changeset in webkit [244783] by keith_miller@apple.com
  • 3 edits in trunk/Source/JavaScriptCore

Fix failing ARM64E wasm tests
https://bugs.webkit.org/show_bug.cgi?id=197420

Reviewed by Saam Barati.

This patch fixes a bug in the slow path of our JS->Wasm IC bridge
where we wouldn't untag the link register before tail calling.

Additionally, this patch fixes a broken assert when using setting
Options::useTailCalls=false.

  • bytecompiler/BytecodeGenerator.cpp:

(JSC::BytecodeGenerator::emitCallForwardArgumentsInTailPosition):

  • wasm/js/WebAssemblyFunction.cpp:

(JSC::WebAssemblyFunction::jsCallEntrypointSlow):

10:51 AM Changeset in webkit [244782] by aestes@apple.com
  • 2 edits in trunk/Source/WebCore

Fix the build after r244777.

  • DerivedSources-input.xcfilelist:
  • DerivedSources-output.xcfilelist:
10:37 AM Changeset in webkit [244781] by timothy_horton@apple.com
  • 2 edits in trunk/Source/WebKit

Long-standing rare crash under -[WKWebView _web_immediateActionAnimationControllerForHitTestResultInternal...]
https://bugs.webkit.org/show_bug.cgi?id=197404
<rdar://problem/24589331>

Reviewed by Wenson Hsieh.

  • UIProcess/mac/WKImmediateActionController.mm:

(-[WKImmediateActionController _updateImmediateActionItem]):
(-[WKImmediateActionController menuItem:previewItemAtPoint:]):
(-[WKImmediateActionController menuItem:itemFrameForPoint:]):
(-[WKImmediateActionController _animationControllerForDataDetectedLink]):
Add some null checks; _webHitTestResult can be null if we (somehow) get
an immediate action gesture without having previously gotten a
mouseDidMoveOverElement from the Web Content process. Cover all our bases.

10:08 AM Changeset in webkit [244780] by Shawn Roberts
  • 3 edits in trunk/LayoutTests

animations/play-state-in-shorthand.html is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=192501

Unreviewed test gardening.

  • platform/ios-simulator-wk2/TestExpectations: Updating expectations for flaky failure
  • platform/mac/TestExpectations: Ditto
9:40 AM Changeset in webkit [244779] by pvollan@apple.com
  • 2 edits in trunk/Source/WebKit

[iOS] Fix crash caused by sandbox violation
https://bugs.webkit.org/show_bug.cgi?id=197416
<rdar://problem/50266257>

Reviewed by Brent Fulgham.

Add syscall to the iOS sandbox.

  • Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb:
9:34 AM Changeset in webkit [244778] by Pablo Saavedra
  • 2 edits in trunk/Tools

Unreviewed. Add myself as a committer.

  • Scripts/webkitpy/common/config/contributors.json
9:33 AM Changeset in webkit [244777] by Justin Fan
  • 14 edits
    3 copies
    3 moves
    3 adds in trunk

[WebGPU] Move 'gpu' API entry point from DOMWindow to Navigator
https://bugs.webkit.org/show_bug.cgi?id=197348

Reviewed by Myles C. Maxfield.

Source/WebCore:

Latest API provides 'gpu' through Navigator instead of DOMWindow. Replace DOMWindowWebGPU with NavigatorGPU.

Existing tests updated to match. Add test: webgpu-enabled-in-worker.html to ensure workers can access WebGPU.

  • CMakeLists.txt:
  • DerivedSources.make:
  • Modules/webgpu/NavigatorGPU.cpp:

(WebCore::NavigatorGPU::from):
(WebCore::NavigatorGPU::supplementName):
(WebCore::NavigatorGPU::gpu):
(WebCore::NavigatorGPU::gpu const):

  • Modules/webgpu/NavigatorGPU.h:
  • Modules/webgpu/NavigatorGPU.idl:
  • Modules/webgpu/WorkerNavigatorGPU.cpp:

(WebCore::WorkerNavigatorGPU::from):
(WebCore::WorkerNavigatorGPU::supplementName):
(WebCore::WorkerNavigatorGPU::gpu):
(WebCore::WorkerNavigatorGPU::gpu const):

  • Modules/webgpu/WorkerNavigatorGPU.h:
  • Modules/webgpu/WorkerNavigatorGPU.idl:
  • Sources.txt:
  • WebCore.xcodeproj/project.pbxproj:

LayoutTests:

WebGPU is now accessed through navigator.gpu instead of window.gpu.
Add webgpu-enabled-in-worker to ensure workers can access WebGPU.

  • webgpu/adapter-options.html:
  • webgpu/js/create-worker-device.js: Added.

(async.tryCreateDevice):

  • webgpu/js/webgpu-functions.js:

(async.getBasicDevice):

  • webgpu/queue-creation.html:
  • webgpu/webgpu-enabled-in-worker-expected.txt: Added.
  • webgpu/webgpu-enabled-in-worker.html: Added.
  • webgpu/webgpu-enabled.html:
  • webgpu/whlsl.html:

Add navigator.gpu to expectations for navigator-detached-no-crash.

  • fast/dom/navigator-detached-no-crash-expected.txt:
  • platform/mac-highsierra/fast/dom/navigator-detached-no-crash-expected.txt:
  • platform/mac-wk2/fast/dom/navigator-detached-no-crash-expected.txt:
9:23 AM Changeset in webkit [244776] by youenn@apple.com
  • 5 edits
    1 move
    1 add
    1 delete in trunk/LayoutTests

Update WPT service-worker resource-timing test to use hosts[alt]
https://bugs.webkit.org/show_bug.cgi?id=197329

Reviewed by Chris Dumez.

LayoutTests/imported/w3c:

Set alt server as 127.0.0.1.
Use hosts[alt][] instead of domains[www] to trigger a cross origin load.

  • resources/config.json:
  • web-platform-tests/service-workers/service-worker/resource-timing.https-expected.txt: Removed.
  • web-platform-tests/service-workers/service-worker/resource-timing.sub.https-expected.txt: Added.
  • web-platform-tests/service-workers/service-worker/resource-timing.sub.https.html: Renamed from LayoutTests/imported/w3c/web-platform-tests/service-workers/service-worker/resource-timing.https.html.
  • web-platform-tests/service-workers/service-worker/resources/resource-timing-iframe.sub.html:

LayoutTests:

9:07 AM Changeset in webkit [244775] by Alan Bujtas
  • 12 edits
    6 adds in trunk

Source/WebCore:
Double-tapping a post to like doesn't work on Instagram.com (needs 'dblclick' event)
https://bugs.webkit.org/show_bug.cgi?id=197347
<rdar://problem/49393423>

Reviewed by Wenson Hsieh.

Tests: fast/events/touch/ios/double-tap-for-double-click1.html

fast/events/touch/ios/double-tap-for-double-click2.html

  • page/Frame.h:
  • page/ios/FrameIOS.mm:

(WebCore::Frame::nodeRespondingToDoubleClickEvent):

Source/WebKit:
[iOS] Double-tapping a post to like doesn't work on Instagram.com (needs 'dblclick' event)
https://bugs.webkit.org/show_bug.cgi?id=197347
<rdar://problem/49393423>

Reviewed by Wenson Hsieh.

This patch adds support for dispatching dblclick event as the result of the double tap gesture.

  1. The new double tap gesture recognizer always fires regardless of what type of action the other, existing double gesture recognizers (double-tap to zoom) trigger.
  2. The dblclick event is dispatched on the node with registered dblclick event handler unless the node prohibits double tap gesture (touch-action != auto).
  • UIProcess/WebPageProxy.h:
  • UIProcess/ios/WKContentViewInteraction.h:
  • UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView setupInteraction]):
(-[WKContentView cleanupInteraction]):
(-[WKContentView _removeDefaultGestureRecognizers]):
(-[WKContentView _addDefaultGestureRecognizers]):
(-[WKContentView gestureRecognizer:shouldRecognizeSimultaneouslyWithGestureRecognizer:]):
(-[WKContentView gestureRecognizerShouldBegin:]):
(-[WKContentView _doubleTapRecognizedForDoubleClick:]):

  • UIProcess/ios/WebPageProxyIOS.mm:

(WebKit::WebPageProxy::handleDoubleTapForDoubleClickAtPoint):

  • WebProcess/WebPage/WebPage.h:
  • WebProcess/WebPage/WebPage.messages.in:
  • WebProcess/WebPage/ios/WebPageIOS.mm:

(WebKit::WebPage::handleDoubleTapForDoubleClickAtPoint):

LayoutTests:
Double-tapping a post to like doesn't work on Instagram.com (needs 'dblclick' event)
https://bugs.webkit.org/show_bug.cgi?id=197347
<rdar://problem/49393423>

Reviewed by Wenson Hsieh.

  • fast/events/touch/ios/double-tap-for-double-click1-expected.txt: Added.
  • fast/events/touch/ios/double-tap-for-double-click1.html: Added.
  • fast/events/touch/ios/double-tap-for-double-click2-expected.txt: Added.
  • fast/events/touch/ios/double-tap-for-double-click2.html: Added.
8:17 AM Changeset in webkit [244774] by youenn@apple.com
  • 6 edits
    5 adds in trunk

Reject/throw when calling AudioContext methods on a stopped AudioContext
https://bugs.webkit.org/show_bug.cgi?id=197391

Reviewed by Eric Carlson.

Source/WebCore:

Return InvalidStateError in that case.
ASSERT that we do not call lazyInitialize after being stopped
since this would mean we are doing unneeded processing.

Test: http/wpt/webaudio/audiocontext-stopped.html

  • Modules/webaudio/AudioContext.cpp:

(WebCore::AudioContext::lazyInitialize):
(WebCore::AudioContext::createBufferSource):
(WebCore::AudioContext::createMediaElementSource):
(WebCore::AudioContext::createMediaStreamSource):
(WebCore::AudioContext::createMediaStreamDestination):
(WebCore::AudioContext::createScriptProcessor):
(WebCore::AudioContext::createBiquadFilter):
(WebCore::AudioContext::createWaveShaper):
(WebCore::AudioContext::createPanner):
(WebCore::AudioContext::createConvolver):
(WebCore::AudioContext::createDynamicsCompressor):
(WebCore::AudioContext::createAnalyser):
(WebCore::AudioContext::createGain):
(WebCore::AudioContext::createDelay):
(WebCore::AudioContext::createChannelSplitter):
(WebCore::AudioContext::createChannelMerger):
(WebCore::AudioContext::createOscillator):
(WebCore::AudioContext::createPeriodicWave):
(WebCore::AudioContext::startRendering):
(WebCore::AudioContext::suspend):
(WebCore::AudioContext::resume):
(WebCore::AudioContext::close):

  • Modules/webaudio/AudioContext.h:
  • Modules/webaudio/AudioContext.idl:

LayoutTests:

  • http/wpt/webaudio/audiocontext-stopped-expected.txt: Added.
  • http/wpt/webaudio/audiocontext-stopped.html: Added.
  • http/wpt/webaudio/resources/audiocontext-stopped-iframe.html: Added.
  • platform/win/TestExpectations: Skip test for win.
8:16 AM Changeset in webkit [244773] by youenn@apple.com
  • 9 edits in trunk/Source

Make Document audio producers use WeakPtr
https://bugs.webkit.org/show_bug.cgi?id=197382

Reviewed by Eric Carlson.

Source/WebCore:

Move from a hash set of raw pointers to a hash set of weak pointers.
This helps make the code cleaner.
No observable change of behavior.

  • Modules/mediastream/MediaStreamTrack.h:
  • dom/Document.cpp:

(WebCore::Document::addAudioProducer):
(WebCore::Document::removeAudioProducer):
(WebCore::Document::updateIsPlayingMedia):
(WebCore::Document::pageMutedStateDidChange):

  • dom/Document.h:
  • html/HTMLMediaElement.cpp:

(WebCore::HTMLMediaElement::updateActiveTextTrackCues):

  • html/HTMLMediaElement.h:
  • page/MediaProducer.h:

Source/WTF:

  • wtf/WeakHashSet.h:

(WTF::WeakHashSet::hasNullReferences const):

8:12 AM Changeset in webkit [244772] by Antti Koivisto
  • 5 edits in trunk/Source

Tighten type of ScrollingTree:rootNode() to ScrollingTreeFrameScrollingNode
https://bugs.webkit.org/show_bug.cgi?id=197414

Reviewed by Frédéric Wang.

Source/WebCore:

  • page/scrolling/ScrollingTree.cpp:

(WebCore::ScrollingTree::shouldHandleWheelEventSynchronously):
(WebCore::ScrollingTree::handleWheelEvent):
(WebCore::ScrollingTree::mainFrameViewportChangedViaDelegatedScrolling):
(WebCore::ScrollingTree::updateTreeFromStateNode):

  • page/scrolling/ScrollingTree.h:

(WebCore::ScrollingTree::rootNode const):

Source/WebKit:

  • UIProcess/RemoteLayerTree/RemoteScrollingCoordinatorProxy.cpp:
8:03 AM Changeset in webkit [244771] by youenn@apple.com
  • 2 edits in trunk/Source/WebCore

Refactor AudioContext to register/unregister itself at construction/destruction time
https://bugs.webkit.org/show_bug.cgi?id=197383

Reviewed by Eric Carlson.

Registering/Unregistering is cheap.
Instead of registering/unregistering in initialize/uninitialize,
move this code to constructor/destructor.
No observable change of behavior.

  • Modules/webaudio/AudioContext.cpp:

(WebCore::AudioContext::AudioContext):
(WebCore::AudioContext::~AudioContext):
(WebCore::AudioContext::lazyInitialize):
(WebCore::AudioContext::uninitialize):
(WebCore::AudioContext::visibilityStateChanged):

8:01 AM Changeset in webkit [244770] by aakash_jain@apple.com
  • 3 edits in trunk/Tools

[ews-build] Enable webkitpy queue on new EWS
https://bugs.webkit.org/show_bug.cgi?id=197396

Reviewed by Lucas Forschler.

  • BuildSlaveSupport/ews-app/ews/views/statusbubble.py:

(StatusBubble):

  • BuildSlaveSupport/ews-build/config.json:
7:51 AM Changeset in webkit [244769] by aakash_jain@apple.com
  • 3 edits in trunk/Tools

[ews-build] Parse and display webkitpy failures
https://bugs.webkit.org/show_bug.cgi?id=197395

Reviewed by Lucas Forschler.

  • BuildSlaveSupport/ews-build/steps.py:

(RunWebKitPyTests.start): Initialize log_observer for json output.
(RunWebKitPyTests.getResultSummary): Update step and build summary based on webkitpy results.
(RunWebKitPyTests._addToLog): Method to add message to log.

  • BuildSlaveSupport/ews-build/steps_unittest.py: Updated unit-test accordingly.
7:32 AM Changeset in webkit [244768] by Michael Catanzaro
  • 2 edits in trunk/Source/WebCore

WebCore::StyleColorScheme should not have explicitly-declared copy constructor
https://bugs.webkit.org/show_bug.cgi?id=197412

Reviewed by Don Olmstead.

Either we need to explicitly declare a copy assignment operator here, or the copy
constructor needs to be removed. Having one without the other causes a huge warning spam
with GCC 9. In this case, the copy constructor is redundant because it's identical to an
implicitly-declared copy constructor, so let's just remove it.

  • rendering/style/StyleColorScheme.h:
4:00 AM Changeset in webkit [244767] by Carlos Garcia Campos
  • 2 edits in trunk/Source/WebKit

REGRESSION(r244750): [GTK][WPE] Network process is crashing in all layout tests
https://bugs.webkit.org/show_bug.cgi?id=197406

Reviewed by Žan Doberšek.

Do not trasnfer ownership of cachePath to traverseDirectory() lambda, since it's passed also as first argument
and it can be passed as a reference.

  • NetworkProcess/cache/NetworkCacheStorage.cpp:

(WebKit::NetworkCache::Storage::deleteOldVersions):

2:06 AM Changeset in webkit [244766] by Carlos Garcia Campos
  • 8 edits
    8 adds in trunk

[GTK] Support prefers-color-scheme media query
https://bugs.webkit.org/show_bug.cgi?id=196685

Reviewed by Michael Catanzaro.

.:

Enable DARK_MODE_CSS. Supported color schemes feature is not supported for now. I'm not sure we will be able to
suport it, because it requires to change the theme too many times, which is very slow in GTK.

  • Source/cmake/OptionsGTK.cmake:

Source/WebCore:

Change the gtk-application-prefer-dark-theme setting when tests change the useDarkModeAppearance setting.

  • PlatformGTK.cmake:
  • testing/InternalSettings.cpp:

(WebCore::InternalSettings::resetToConsistentState):
(WebCore::InternalSettings::setUseDarkAppearanceInternal):
(WebCore::InternalSettings::setUseDarkAppearance):

  • testing/InternalSettings.h:

LayoutTests:

Unskip css-dark-mode tests and add platform specific results for some of the tests using the supported color
scheme feature.

  • platform/gtk/TestExpectations:
  • platform/gtk/css-dark-mode/color-scheme-css-expected.txt: Added.
  • platform/gtk/css-dark-mode/color-scheme-meta-expected.txt: Added.
  • platform/gtk/css-dark-mode/default-colors-expected.txt: Added.
  • platform/gtk/css-dark-mode/older-syntax/supported-color-schemes-css-expected.txt: Added.
  • platform/gtk/css-dark-mode/older-syntax/supported-color-schemes-meta-expected.txt: Added.
  • platform/gtk/css-dark-mode/prefers-color-scheme-expected.txt: Added.
2:04 AM Changeset in webkit [244765] by Carlos Garcia Campos
  • 29 edits in trunk/LayoutTests

Unreviewed GTK gardening. Rebaseline tests after r244635.

  • platform/gtk/css3/selectors3/html/css3-modsel-24-expected.png:
  • platform/gtk/css3/selectors3/html/css3-modsel-24-expected.txt:
  • platform/gtk/css3/selectors3/html/css3-modsel-68-expected.png:
  • platform/gtk/css3/selectors3/html/css3-modsel-68-expected.txt:
  • platform/gtk/css3/selectors3/xhtml/css3-modsel-24-expected.png:
  • platform/gtk/css3/selectors3/xhtml/css3-modsel-24-expected.txt:
  • platform/gtk/css3/selectors3/xhtml/css3-modsel-68-expected.png:
  • platform/gtk/css3/selectors3/xhtml/css3-modsel-68-expected.txt:
  • platform/gtk/css3/selectors3/xml/css3-modsel-24-expected.png:
  • platform/gtk/css3/selectors3/xml/css3-modsel-24-expected.txt:
  • platform/gtk/css3/selectors3/xml/css3-modsel-68-expected.png:
  • platform/gtk/css3/selectors3/xml/css3-modsel-68-expected.txt:
  • platform/gtk/fast/forms/basic-inputs-expected.png:
  • platform/gtk/fast/forms/basic-inputs-expected.txt:
  • platform/gtk/fast/forms/basic-textareas-expected.png:
  • platform/gtk/fast/forms/basic-textareas-expected.txt:
  • platform/gtk/fast/forms/basic-textareas-quirks-expected.png:
  • platform/gtk/fast/forms/basic-textareas-quirks-expected.txt:
  • platform/gtk/fast/forms/disabled-select-change-index-expected.png:
  • platform/gtk/fast/forms/disabled-select-change-index-expected.txt:
  • platform/gtk/fast/forms/file/file-input-disabled-expected.png:
  • platform/gtk/fast/forms/file/file-input-disabled-expected.txt:
  • platform/gtk/fast/forms/input-appearance-disabled-expected.png:
  • platform/gtk/fast/forms/input-appearance-disabled-expected.txt:
  • platform/gtk/fast/forms/input-disabled-color-expected.png:
  • platform/gtk/fast/forms/input-disabled-color-expected.txt:
  • platform/gtk/fast/forms/placeholder-pseudo-style-expected.txt:
  • platform/gtk/fast/forms/textarea-placeholder-pseudo-style-expected.txt:

Apr 29, 2019:

8:27 PM Changeset in webkit [244764] by sbarati@apple.com
  • 47 edits in trunk/Source/JavaScriptCore

Make JITType an enum class
https://bugs.webkit.org/show_bug.cgi?id=197394

Reviewed by Yusuke Suzuki.

This makes the code more easily searchable.

  • bytecode/CallLinkStatus.cpp:

(JSC::CallLinkStatus::computeFor):

  • bytecode/CodeBlock.cpp:

(JSC::CodeBlock::dumpAssumingJITType const):
(JSC::CodeBlock::specialOSREntryBlockOrNull):
(JSC::timeToLive):
(JSC::CodeBlock::propagateTransitions):
(JSC::CodeBlock::baselineAlternative):
(JSC::CodeBlock::baselineVersion):
(JSC::CodeBlock::hasOptimizedReplacement):
(JSC::CodeBlock::noticeIncomingCall):
(JSC::CodeBlock::setOptimizationThresholdBasedOnCompilationResult):
(JSC::CodeBlock::tallyFrequentExitSites):
(JSC::CodeBlock::frameRegisterCount):
(JSC::CodeBlock::bytecodeOffsetFromCallSiteIndex):

  • bytecode/CodeBlock.h:

(JSC::CodeBlock::jitType const):
(JSC::CodeBlock::hasBaselineJITProfiling const):

  • bytecode/CodeBlockWithJITType.h:

(JSC::CodeBlockWithJITType::CodeBlockWithJITType):

  • bytecode/DeferredSourceDump.cpp:

(JSC::DeferredSourceDump::DeferredSourceDump):

  • bytecode/DeferredSourceDump.h:
  • bytecode/ExitingJITType.h:

(JSC::exitingJITTypeFor):

  • bytecode/InlineCallFrame.h:

(JSC::baselineCodeBlockForOriginAndBaselineCodeBlock):

  • dfg/DFGByteCodeParser.cpp:

(JSC::DFG::ByteCodeParser::parseCodeBlock):

  • dfg/DFGDisassembler.cpp:

(JSC::DFG::Disassembler::dumpHeader):

  • dfg/DFGDriver.cpp:

(JSC::DFG::compileImpl):

  • dfg/DFGGraph.cpp:

(JSC::DFG::Graph::dump):

  • dfg/DFGJITCode.cpp:

(JSC::DFG::JITCode::JITCode):
(JSC::DFG::JITCode::checkIfOptimizationThresholdReached):
(JSC::DFG::JITCode::optimizeNextInvocation):
(JSC::DFG::JITCode::dontOptimizeAnytimeSoon):
(JSC::DFG::JITCode::optimizeAfterWarmUp):
(JSC::DFG::JITCode::optimizeSoon):
(JSC::DFG::JITCode::forceOptimizationSlowPathConcurrently):
(JSC::DFG::JITCode::setOptimizationThresholdBasedOnCompilationResult):

  • dfg/DFGJITFinalizer.cpp:

(JSC::DFG::JITFinalizer::finalize):
(JSC::DFG::JITFinalizer::finalizeFunction):

  • dfg/DFGOSREntry.cpp:

(JSC::DFG::prepareOSREntry):
(JSC::DFG::prepareCatchOSREntry):

  • dfg/DFGOSRExit.cpp:

(JSC::DFG::OSRExit::executeOSRExit):
(JSC::DFG::reifyInlinedCallFrames):
(JSC::DFG::OSRExit::compileOSRExit):

  • dfg/DFGOSRExitCompilerCommon.cpp:

(JSC::DFG::handleExitCounts):
(JSC::DFG::reifyInlinedCallFrames):
(JSC::DFG::adjustAndJumpToTarget):

  • dfg/DFGOSRExitCompilerCommon.h:

(JSC::DFG::adjustFrameAndStackInOSRExitCompilerThunk):

  • dfg/DFGOperations.cpp:
  • dfg/DFGThunks.cpp:

(JSC::DFG::osrExitGenerationThunkGenerator):

  • dfg/DFGVariableEventStream.cpp:

(JSC::DFG::VariableEventStream::reconstruct const):

  • ftl/FTLCompile.cpp:

(JSC::FTL::compile):

  • ftl/FTLJITCode.cpp:

(JSC::FTL::JITCode::JITCode):

  • ftl/FTLJITFinalizer.cpp:

(JSC::FTL::JITFinalizer::finalizeCommon):

  • ftl/FTLLink.cpp:

(JSC::FTL::link):

  • ftl/FTLOSRExitCompiler.cpp:

(JSC::FTL::compileFTLOSRExit):

  • ftl/FTLThunks.cpp:

(JSC::FTL::genericGenerationThunkGenerator):

  • interpreter/CallFrame.cpp:

(JSC::CallFrame::callSiteBitsAreBytecodeOffset const):
(JSC::CallFrame::callSiteBitsAreCodeOriginIndex const):

  • interpreter/StackVisitor.cpp:

(JSC::StackVisitor::Frame::dump const):

  • jit/AssemblyHelpers.h:

(JSC::AssemblyHelpers::AssemblyHelpers):

  • jit/JIT.cpp:

(JSC::JIT::link):

  • jit/JITCode.cpp:

(JSC::JITCode::typeName):
(WTF::printInternal):

  • jit/JITCode.h:

(JSC::JITCode::bottomTierJIT):
(JSC::JITCode::topTierJIT):
(JSC::JITCode::nextTierJIT):
(JSC::JITCode::isExecutableScript):
(JSC::JITCode::couldBeInterpreted):
(JSC::JITCode::isJIT):
(JSC::JITCode::isOptimizingJIT):
(JSC::JITCode::isBaselineCode):
(JSC::JITCode::jitTypeFor):

  • jit/JITDisassembler.cpp:

(JSC::JITDisassembler::dumpHeader):

  • jit/JITOperations.cpp:
  • jit/JITThunks.cpp:

(JSC::JITThunks::hostFunctionStub):

  • jit/JITToDFGDeferredCompilationCallback.cpp:

(JSC::JITToDFGDeferredCompilationCallback::compilationDidBecomeReadyAsynchronously):
(JSC::JITToDFGDeferredCompilationCallback::compilationDidComplete):

  • jit/JITWorklist.cpp:

(JSC::JITWorklist::compileLater):
(JSC::JITWorklist::compileNow):

  • jit/Repatch.cpp:

(JSC::readPutICCallTarget):
(JSC::ftlThunkAwareRepatchCall):

  • llint/LLIntEntrypoint.cpp:

(JSC::LLInt::setFunctionEntrypoint):
(JSC::LLInt::setEvalEntrypoint):
(JSC::LLInt::setProgramEntrypoint):
(JSC::LLInt::setModuleProgramEntrypoint):

  • llint/LLIntSlowPaths.cpp:

(JSC::LLInt::jitCompileAndSetHeuristics):
(JSC::LLInt::LLINT_SLOW_PATH_DECL):

  • runtime/SamplingProfiler.cpp:

(JSC::SamplingProfiler::processUnverifiedStackTraces):

  • runtime/SamplingProfiler.h:
  • runtime/VM.cpp:

(JSC::jitCodeForCallTrampoline):
(JSC::jitCodeForConstructTrampoline):

  • tools/CodeProfile.cpp:

(JSC::CodeProfile::sample):

  • tools/JSDollarVM.cpp:

(JSC::CallerFrameJITTypeFunctor::CallerFrameJITTypeFunctor):
(JSC::CallerFrameJITTypeFunctor::jitType):
(JSC::functionLLintTrue):
(JSC::functionJITTrue):

7:54 PM Changeset in webkit [244763] by ysuzuki@apple.com
  • 2 edits in trunk/Source/JavaScriptCore

Unreivewed, fix FTL implementation of r244760
https://bugs.webkit.org/show_bug.cgi?id=197362

Reviewed by Saam Barati.

Looked with Saam. ValueFromBlock from double case block was overridden by NaN thing now.

  • ftl/FTLLowerDFGToB3.cpp:

(JSC::FTL::DFG::LowerDFGToB3::compileNormalizeMapKey):

4:55 PM Changeset in webkit [244762] by Truitt Savell
  • 12 edits
    6 deletes in trunk

Unreviewed, rolling out r244755.

Casued several test failures on iOS

Reverted changeset:

"Double-tapping a post to like doesn't work on Instagram.com
(needs 'dblclick' event)"
https://bugs.webkit.org/show_bug.cgi?id=197347
https://trac.webkit.org/changeset/244755

3:40 PM Changeset in webkit [244761] by Chris Dumez
  • 2 edits in trunk/Source/WebKit

[iOS] The UIProcess may get killed for trying to stay runnable in the background for more than 30 seconds
https://bugs.webkit.org/show_bug.cgi?id=197385
<rdar://problem/50001505>

Reviewed by Geoffrey Garen.

If the UIProcess holds a background assertion for itself for 30 seconds, the assertion's invalidation handler
will get called and it is our responsibility to release this assertion or the UIProcess will get killed by the
system. The logic in ProcessAssertion would normally do that but it would also happily try and re-take another
background process assertion shortly after the previous one expired (and before the UIProcess got suspended).
When doing so, the new background assertion would expire right away and we would get killed without its
invalidation handler getting called.

To address the issue, the logic in ProcessAssertion will now prevent taking a new background assertion after
one expires and until the application becomes foreground again.

  • UIProcess/ios/ProcessAssertionIOS.mm:

(-[WKProcessAssertionBackgroundTaskManager init]):
(-[WKProcessAssertionBackgroundTaskManager _updateBackgroundTask]):

3:25 PM Changeset in webkit [244760] by ysuzuki@apple.com
  • 5 edits
    2 adds in trunk

normalizeMapKey should normalize NaN to one PureNaN bit pattern to make MapHash same
https://bugs.webkit.org/show_bug.cgi?id=197362

Reviewed by Saam Barati.

JSTests:

  • stress/map-with-nan.js: Added.

(shouldBe):
(div):
(NaN1):
(NaN2):
(NaN3):
(NaN4):
(NaN1NoInline):
(NaN2NoInline):
(NaN3NoInline):
(NaN4NoInline):
(test1):
(test2):
(test3):
(test4):

  • stress/set-with-nan.js: Added.

(shouldBe):
(div):
(NaN1):
(NaN2):
(NaN3):
(NaN4):
(NaN1NoInline):
(NaN2NoInline):
(NaN3NoInline):
(NaN4NoInline):
(test2):
(test4):

Source/JavaScriptCore:

Our Map/Set's hash algorithm relies on the bit pattern of JSValue. So our Map/Set has
normalization of the key, which normalizes Int32 / Double etc. But we did not normalize
pure NaNs into one canonicalized pure NaN. So we end up having multiple different pure NaNs
in one Map/Set. This patch normalizes NaN into one jsNaN(), which uses PNaN for the representation.

  • dfg/DFGSpeculativeJIT.cpp:

(JSC::DFG::SpeculativeJIT::compileNormalizeMapKey):

  • ftl/FTLLowerDFGToB3.cpp:

(JSC::FTL::DFG::LowerDFGToB3::compileNormalizeMapKey):

  • runtime/HashMapImpl.h:

(JSC::normalizeMapKey):

3:06 PM Changeset in webkit [244759] by Alan Coon
  • 7 edits in tags/Safari-608.1.19.3/Source

Versioning.

3:04 PM Changeset in webkit [244758] by Alan Coon
  • 1 copy in tags/Safari-608.1.19.3

New tag.

2:51 PM Changeset in webkit [244757] by Alan Coon
  • 29 edits in tags/Safari-608.1.19.2

Cherry-pick r244756. rdar://problem/50299396

<rdar://problem/50299396> Fix internal High Sierra build
https://bugs.webkit.org/show_bug.cgi?id=197388

Source/bmalloc:

  • Configurations/Base.xcconfig:

Source/JavaScriptCore:

  • Configurations/Base.xcconfig:

Source/ThirdParty:

  • gtest/xcode/Config/General.xcconfig:

Source/ThirdParty/ANGLE:

  • Configurations/Base.xcconfig:

Source/ThirdParty/libwebrtc:

  • Configurations/Base.xcconfig:

Source/WebCore:

  • Configurations/Base.xcconfig:

Source/WebCore/PAL:

  • Configurations/Base.xcconfig:

Source/WebInspectorUI:

  • Configurations/Base.xcconfig:

Source/WebKit:

  • Configurations/Base.xcconfig:

Source/WebKitLegacy/mac:

  • Configurations/Base.xcconfig:
  • Configurations/WebKitLegacy.xcconfig:

Source/WTF:

  • Configurations/Base.xcconfig:

Tools:

  • DumpRenderTree/mac/Configurations/Base.xcconfig:
  • MiniBrowser/Configurations/Base.xcconfig:
  • MobileMiniBrowser/Configurations/Base.xcconfig:
  • TestWebKitAPI/Configurations/Base.xcconfig:
  • WebKitTestRunner/Configurations/Base.xcconfig:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244756 268f45cc-cd09-0410-ab3c-d52691b4dbfc

2:48 PM Changeset in webkit [244756] by achristensen@apple.com
  • 29 edits in trunk

<rdar://problem/50299396> Fix internal High Sierra build
https://bugs.webkit.org/show_bug.cgi?id=197388

Source/bmalloc:

  • Configurations/Base.xcconfig:

Source/JavaScriptCore:

  • Configurations/Base.xcconfig:

Source/ThirdParty:

  • gtest/xcode/Config/General.xcconfig:

Source/ThirdParty/ANGLE:

  • Configurations/Base.xcconfig:

Source/ThirdParty/libwebrtc:

  • Configurations/Base.xcconfig:

Source/WebCore:

  • Configurations/Base.xcconfig:

Source/WebCore/PAL:

  • Configurations/Base.xcconfig:

Source/WebInspectorUI:

  • Configurations/Base.xcconfig:

Source/WebKit:

  • Configurations/Base.xcconfig:

Source/WebKitLegacy/mac:

  • Configurations/Base.xcconfig:
  • Configurations/WebKitLegacy.xcconfig:

Source/WTF:

  • Configurations/Base.xcconfig:

Tools:

  • DumpRenderTree/mac/Configurations/Base.xcconfig:
  • MiniBrowser/Configurations/Base.xcconfig:
  • MobileMiniBrowser/Configurations/Base.xcconfig:
  • TestWebKitAPI/Configurations/Base.xcconfig:
  • WebKitTestRunner/Configurations/Base.xcconfig:
2:31 PM Changeset in webkit [244755] by Alan Bujtas
  • 12 edits
    6 adds in trunk

Source/WebCore:
Double-tapping a post to like doesn't work on Instagram.com (needs 'dblclick' event)
https://bugs.webkit.org/show_bug.cgi?id=197347
<rdar://problem/49393423>

Reviewed by Wenson Hsieh.

Tests: fast/events/touch/ios/double-tap-for-double-click1.html

fast/events/touch/ios/double-tap-for-double-click2.html

  • page/Frame.h:
  • page/ios/FrameIOS.mm:

(WebCore::Frame::nodeRespondingToDoubleClickEvent):

Source/WebKit:
[iOS] Double-tapping a post to like doesn't work on Instagram.com (needs 'dblclick' event)
https://bugs.webkit.org/show_bug.cgi?id=197347
<rdar://problem/49393423>

Reviewed by Wenson Hsieh.

This patch adds support for dispatching dblclick event as the result of the double tap gesture.

  1. The new double tap gesture recognizer always fires regardless of what type of action the other, existing double gesture recognizers (double-tap to zoom) trigger.
  2. The dblclick event is dispatched on the node with registered dblclick event handler unless the node prohibits double tap gesture (touch-action != auto).
  • UIProcess/WebPageProxy.h:
  • UIProcess/ios/WKContentViewInteraction.h:
  • UIProcess/ios/WKContentViewInteraction.mm:

(-[WKContentView setupInteraction]):
(-[WKContentView cleanupInteraction]):
(-[WKContentView _removeDefaultGestureRecognizers]):
(-[WKContentView _addDefaultGestureRecognizers]):
(-[WKContentView gestureRecognizer:shouldRecognizeSimultaneouslyWithGestureRecognizer:]):
(-[WKContentView gestureRecognizerShouldBegin:]):
(-[WKContentView _doubleTapRecognizedForDoubleClick:]):

  • UIProcess/ios/WebPageProxyIOS.mm:

(WebKit::WebPageProxy::handleDoubleTapForDoubleClickAtPoint):

  • WebProcess/WebPage/WebPage.h:
  • WebProcess/WebPage/WebPage.messages.in:
  • WebProcess/WebPage/ios/WebPageIOS.mm:

(WebKit::WebPage::handleDoubleTapForDoubleClickAtPoint):

LayoutTests:
Double-tapping a post to like doesn't work on Instagram.com (needs 'dblclick' event)
https://bugs.webkit.org/show_bug.cgi?id=197347
<rdar://problem/49393423>

Reviewed by Wenson Hsieh.

  • fast/events/touch/ios/double-tap-for-double-click1-expected.txt: Added.
  • fast/events/touch/ios/double-tap-for-double-click1.html: Added.
  • fast/events/touch/ios/double-tap-for-double-click2-expected.txt: Added.
  • fast/events/touch/ios/double-tap-for-double-click2.html: Added.
2:05 PM Changeset in webkit [244754] by timothy_horton@apple.com
  • 2 edits in trunk/Source/WebKit

REGRESSION (r244142): Fandango accounts tab crashes under Color()
https://bugs.webkit.org/show_bug.cgi?id=197380
<rdar://problem/50186175>

Reviewed by Megan Gardner.

  • UIProcess/API/Cocoa/WKWebView.mm:

(baseScrollViewBackgroundColor):
We now call -_updateScrollViewBackground from -setBackgroundColor:.
-setBackgroundColor: can be called from -[UIView initWithCoder:]; if it
is, WKWebView's _page is not yet initialized, so we crash. Return an invalid color.
Then, the caller falls back to WKContentView's backgroundColor. But WKContentView
isn't initialized yet either. So add a further fallback.

1:50 PM Changeset in webkit [244753] by youenn@apple.com
  • 2 edits in trunk/Tools

Remove spurious GVA printf logging
https://bugs.webkit.org/show_bug.cgi?id=197368
<rdar://problem/35975256>

Reviewed by Alexey Proskuryakov.

  • Scripts/webkitpy/port/mac.py:

(MacPort.logging_patterns_to_strip):

1:45 PM Changeset in webkit [244752] by Simon Fraser
  • 3 edits
    2 adds in trunk

REGRESSION (r238090): animation on https://www.robotodyssey.online gets stuck; site broken
https://bugs.webkit.org/show_bug.cgi?id=197381
Source/WebCore:

Reviewed by Zalan Bujtas.

When -webkit-clip-path changes on a composited layer, we need to trigger a backing geometry update
to push the changes to GraphicsLayers.

Test: compositing/style-change/clip-path-change.html

  • rendering/RenderLayerCompositor.cpp:

(WebCore::recompositeChangeRequiresGeometryUpdate):

LayoutTests:

<rdar://problem/50308128>

Reviewed by Zalan Bujtas.

  • compositing/style-change/clip-path-change-expected.html: Added.
  • compositing/style-change/clip-path-change.html: Added.
1:34 PM Changeset in webkit [244751] by dino@apple.com
  • 2 edits in trunk/Source/WebKit

Add InteractionIsHappening to SuppressSelectionAssistantReason
https://bugs.webkit.org/show_bug.cgi?id=197375
<rdar://problem/50304986>

Reviewed by Megan Gardner.

Add a new reason for suppressing the selection assistant, to
be used in a follow-up patch.

  • UIProcess/ios/WKContentViewInteraction.h: Add InteractionIsHappening.
1:32 PM Changeset in webkit [244750] by achristensen@apple.com
  • 2 edits in trunk/Source/WebKit

Storage::deleteOldVersions does not need to retain Storage
https://bugs.webkit.org/show_bug.cgi?id=197336
<rdar://problem/48679972>

Reviewed by Darin Adler.

  • NetworkProcess/cache/NetworkCacheStorage.cpp:

(WebKit::NetworkCache::Storage::deleteOldVersions):
Something is wrong with the lifetime of Storage, but we only need a String, not the whole Storage.

1:04 PM Changeset in webkit [244749] by youenn@apple.com
  • 17 edits
    2 adds in trunk

getDisplayMedia should be called on user gesture
https://bugs.webkit.org/show_bug.cgi?id=197356
Source/WebCore:

Reviewed by Eric Carlson.

Allow getDisplayMedia on user gesture only.
Otherwise reject the promise.
Minor refactoring to align getDisplayMedia, getUserMedia and
enumerateDevices when called with no document.

Test: fast/mediastream/screencapture-user-gesture.html

  • Modules/mediastream/MediaDevices.cpp:

(WebCore::MediaDevices::getUserMedia const):

  • Modules/mediastream/MediaDevices.h:
  • Modules/mediastream/NavigatorMediaDevices.h:
  • page/DOMWindow.h:
  • testing/Internals.cpp:

(WebCore::Internals::setDisableGetDisplayMediaUserGestureConstraint):

  • testing/Internals.h:
  • testing/Internals.idl:

Tools:

<rdar://problem/50296074>

Reviewed by Eric Carlson.

Update test configuration to inject internals
so that we can use it to simulate a user click to call getDisplayMedia.

  • TestWebKitAPI/Tests/WebKit/getDisplayMedia.html:
  • TestWebKitAPI/Tests/WebKitCocoa/GetDisplayMedia.mm:

(TestWebKitAPI::GetDisplayMediaTest::SetUp):

LayoutTests:

Reviewed by Eric Carlson.

  • fast/mediastream/media-stream-page-muted.html:
  • fast/mediastream/screencapture-user-gesture-expected.txt: Added.
  • fast/mediastream/screencapture-user-gesture.html: Added.
  • http/tests/media/media-stream/get-display-media-iframe-allow-attribute-expected.txt:
  • http/tests/media/media-stream/get-display-media-iframe-allow-attribute.html:
  • http/tests/media/media-stream/get-display-media-prompt.html:
  • http/tests/media/media-stream/resources/get-display-media-devices-iframe.html:
  • resources/testharnessreport.js:
12:56 PM Changeset in webkit [244748] by jfernandez@igalia.com
  • 8 edits
    26 adds in trunk

line should not be broken before the first space after a word
https://bugs.webkit.org/show_bug.cgi?id=197278

Reviewed by Myles C. Maxfield.

LayoutTests/imported/w3c:

Added new Web Platform Tests to cover the cases discussed [1] with the CSS WG.

[1] https://github.com/w3c/csswg-drafts/issues/3701

  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-001-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-001.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-002-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-002.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-003-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-003.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-004-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-004.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-005-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-005.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-006-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-006.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-007-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-007.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-008-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-008.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-009-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-009.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-010-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-010.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-011-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-011.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-012-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-012.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-013-expected.html: Added.
  • web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-013.html: Added.

Source/WebCore:

The 'white-space: break-spaces' only adds breaking opportunities after
a white space character. However, it's possible to break before the
first space after a word when the feature is used in combination with
other properties, like overflow-wrap.

However, breaking before the first space should not be allowed if
there are previous opportunities. We wrongly assumed that we had to
consider these previous breaking opportunities if the proper combination
of line breaking properties is being used, so that breaking before the
first space after a word is allowed.

This wrong assumption caused several issues, like the one described in
the bug, that lead to incorrectly break before the first space even
though there are previous opportunities, either white spaces or between
letters.

Theses issues have been analyzed [1] by the CSS WG and finally agreed on a
expected behavior, represented in the Web Platform tests added in this
patch.

For the later case, of considering previous opportunities between
letters, we have a seperated issue #952254, so the tests covering such
cases will be added to the TestExpecations as Failure entries.

[1] https://github.com/w3c/csswg-drafts/issues/3701

Tests: imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-001.html

imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-002.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-003.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-004.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-005.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-006.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-007.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-008.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-009.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-010.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-011.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-012.html
imported/w3c/web-platform-tests/css/css-text/white-space/break-spaces-before-first-char-013.html

  • rendering/line/BreakingContext.h:

(WebCore::BreakingContext::handleText):
(WebCore::BreakingContext::trailingSpacesHang):

LayoutTests:

Added a few Skip (due to unimplemented features) and Failure (break-all related
issues) entries for some of the tests added by this patch.

12:52 PM Changeset in webkit [244747] by Brent Fulgham
  • 3 edits
    4 moves in trunk/Source/WebKit

[Cocoa] Move common sandbox files from Shared/mac to Shared/Cocoa
https://bugs.webkit.org/show_bug.cgi?id=197376
<rdar://problem/50305272>

Reviewed by Dean Jackson.

The Apple sandboxing logic currently lives in the inappropriate 'Shared/mac' directory. This is
confusing because the code is used by all Apple ports, not just macOS.

This patch just moves the relevant files from 'Shared/mac' to 'Shared/Cocoa' to avoid this confusion.

  • Shared/Cocoa/SandboxExtensionCocoa.mm: Renamed from Source/WebKit/Shared/mac/SandboxExtensionMac.mm.
  • Shared/Cocoa/SandboxInitialiationParametersCocoa.mm: Renamed from Source/WebKit/Shared/mac/SandboxInitialiationParametersMac.mm.
  • Shared/Cocoa/SandboxUtilities.h: Renamed from Source/WebKit/Shared/mac/SandboxUtilities.h.
  • Shared/Cocoa/SandboxUtilities.mm: Renamed from Source/WebKit/Shared/mac/SandboxUtilities.mm.
  • SourcesCocoa.txt:
  • WebKit.xcodeproj/project.pbxproj:
12:36 PM Changeset in webkit [244746] by don.olmstead@sony.com
  • 2 edits in trunk

[CMake] Add WEBKIT_EXECUTABLE macro
https://bugs.webkit.org/show_bug.cgi?id=197206

Unreviewed build fix.

Calls to add_dependencies were not happening.

  • Source/cmake/WebKitMacros.cmake:
12:21 PM Changeset in webkit [244745] by ysuzuki@apple.com
  • 6 edits in trunk/Source

JITStubRoutineSet wastes 180KB of HashTable capacity on can.com
https://bugs.webkit.org/show_bug.cgi?id=186732

Reviewed by Saam Barati.

Source/JavaScriptCore:

Our current mechanism of JITStubRoutineSet consumes more memory than needed. Basically we have HashMap<uintptr_t, StubRoutine*> and register
each executable address by 16 byte to this entry. So if your StubRoutine has 128bytes, it just adds 8 entries to this hash table.
In Gmail, we see a ~2MB table size.

Instead, this patch uses Vector<pair<uintptr_t, StubRoutine*>> and performs binary search onto this sorted vector. Before conservative
scanning, we sort this vector. And doing binary search with the sorted vector to find executing stub routines from the conservative roots.
This vector includes uintptr_t startAddress to make binary searching fast.

Large amount of conservative scan should be filtered by range check, so I think binary search here is OK, but we can decide based on what the
performance bots say.

  • heap/Heap.cpp:

(JSC::Heap::addCoreConstraints):

  • heap/JITStubRoutineSet.cpp:

(JSC::JITStubRoutineSet::~JITStubRoutineSet):
(JSC::JITStubRoutineSet::add):
(JSC::JITStubRoutineSet::prepareForConservativeScan):
(JSC::JITStubRoutineSet::clearMarks):
(JSC::JITStubRoutineSet::markSlow):
(JSC::JITStubRoutineSet::deleteUnmarkedJettisonedStubRoutines):
(JSC::JITStubRoutineSet::traceMarkedStubRoutines):

  • heap/JITStubRoutineSet.h:

(JSC::JITStubRoutineSet::mark):
(JSC::JITStubRoutineSet::prepareForConservativeScan):
(JSC::JITStubRoutineSet::size const): Deleted.
(JSC::JITStubRoutineSet::at const): Deleted.

Source/WTF:

  • wtf/Range.h:

(WTF::Range::contains const):

12:08 PM Changeset in webkit [244744] by commit-queue@webkit.org
  • 5 edits in trunk/Source/WebKit

[GTK] Back/forward gesture snapshot always times out
https://bugs.webkit.org/show_bug.cgi?id=197233

Patch by Alexander Mikhaylenko <exalm7659@gmail.com> on 2019-04-29
Reviewed by Michael Catanzaro.

Delaying web process launch caused a regression where we create ViewGestureController when the
web process doesn't yet exist. The controller immediately tries to connect to it and fails,
and because of that never receives DidHitRenderTreeSizeThreshold() message, so navigation
snapshot always stays until timeout after performing the gesture.

To prevent this, create the controller in webkitWebViewBaseDidRelaunchWebProcess() instead of
webkitWebViewBaseCreateWebPage(). Additionally, since settings are now created earlier than
ViewGestureController, store the value of whether swipe gesture is enabled in WebKitWebViewBase
and immediately apply it when creating the controller.

Since there is now a point where controller is null, make webkitWebViewBaseViewGestureController()
return null and add null checks everywhere.

  • UIProcess/API/glib/WebKitWebView.cpp:

(enableBackForwardNavigationGesturesChanged):
Move the logic into webkitWebViewBaseSetEnableBackForwardNavigationGesture().

  • UIProcess/API/gtk/PageClientImpl.cpp:

(WebKit::PageClientImpl::wheelEventWasNotHandledByWebCore): Add a null check.

  • UIProcess/API/gtk/WebKitWebViewBase.cpp:

(webkitWebViewBaseDraw): Ditto.
(webkitWebViewBaseScrollEvent): Ditto.
(webkitWebViewBaseSetEnableBackForwardNavigationGesture): Added. In addition to what was in
WebKitWebViewBase::enableBackForwardNavigationGesturesChanged(), store the value in a field
for the case ViewGestureController doesn't exist yet.
(webkitWebViewBaseViewGestureController): Return a pointer instead of reference.
(webkitWebViewBaseCreateWebPage): Stop creating ViewGestureController.
(webkitWebViewBaseDidRelaunchWebProcess): Move creating ViewGestureController here. Also
immediately call setSwipeGestureEnabled() with the stored value.
(webkitWebViewBaseDidStartProvisionalLoadForMainFrame): Add a null check.
(webkitWebViewBaseDidFirstVisuallyNonEmptyLayoutForMainFrame):Ditto.
(webkitWebViewBaseDidFinishLoadForMainFrame): Ditto.
(webkitWebViewBaseDidFailLoadForMainFrame): Ditto.
(webkitWebViewBaseDidSameDocumentNavigationForMainFrame): Ditto.
(webkitWebViewBaseDidRestoreScrollPosition): Ditto.

  • UIProcess/API/gtk/WebKitWebViewBasePrivate.h:
11:40 AM Changeset in webkit [244743] by Alan Coon
  • 7 edits in tags/Safari-608.1.19.2/Source

Versioning.

11:22 AM Changeset in webkit [244742] by Chris Dumez
  • 5 edits in trunk/Source

User-facing strings should use curly quotes instead of straight
https://bugs.webkit.org/show_bug.cgi?id=197370

Reviewed by Geoffrey Garen.

Source/WebCore:

Update localizable strings.

  • en.lproj/Localizable.strings:

Source/WebKit:

Update localizable alert text that was recently added to use curly quotes instead of
straight ones, as per our guidelines.

  • UIProcess/Cocoa/WKOrientationAccessAlert.mm:

(WebKit::presentOrientationAccessAlert):

  • UIProcess/Cocoa/WKStorageAccessAlert.mm:

(WebKit::presentStorageAccessAlert):

11:18 AM Changeset in webkit [244741] by Basuke Suzuki
  • 8 edits in trunk

[Win] Add flag to enable version information stamping and disable by default.
https://bugs.webkit.org/show_bug.cgi?id=197249
<rdar://problem/50224412>

Reviewed by Ross Kirsling.

.:

This feature is only used in AppleWin port. Add flag for this task and make it OFF by default.
Then enable it by default on AppleWin.

  • Source/cmake/OptionsWin.cmake:

Source/JavaScriptCore:

This feature is only used in AppleWin port. Add flag for this task and make it OFF by default.
Then enable it by default on AppleWin.

  • CMakeLists.txt:

Source/WTF:

This feature is only used in AppleWin port. Add flag for this task and make it OFF by default.
Then enable it by default on AppleWin.

  • wtf/CMakeLists.txt:

WebKitLibraries:

Modify script not to invoke SVN command if the directory for SVN doesn't exists.

  • win/tools/scripts/auto-version.pl:
11:18 AM Changeset in webkit [244740] by Ross Kirsling
  • 2 edits in trunk/Source/WebCore

Unreviewed fix for non-unified build after r244687.

  • Modules/indexeddb/server/UniqueIDBDatabaseTransaction.h:
11:16 AM Changeset in webkit [244739] by Alan Coon
  • 4 edits in tags/Safari-608.1.19.2/Source

Cherry-pick r244696. rdar://problem/50299396

Fix internal High Sierra build after r244653
https://bugs.webkit.org/show_bug.cgi?id=197131

  • DerivedSources.make: -std=gnu++17 didn't exist yet. -std=gnu++1z did.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244696 268f45cc-cd09-0410-ab3c-d52691b4dbfc

11:16 AM Changeset in webkit [244738] by Alan Coon
  • 4 edits in tags/Safari-608.1.19.2/Source

Cherry-pick r244692. rdar://problem/50299396

Fix an internal High Sierra build after r244653

https://bugs.webkit.org/show_bug.cgi?id=197131

  • DerivedSources.make: Apparently we can't use gnu++17 when preprocessing Platform.h in the makefile.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244692 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10:45 AM Changeset in webkit [244737] by Alan Coon
  • 1 copy in tags/Safari-608.1.19.2

New tag.

9:24 AM Changeset in webkit [244736] by youenn@apple.com
  • 8 edits in trunk

RTCTrackEvent should be delayed until the whole remote description is set
https://bugs.webkit.org/show_bug.cgi?id=196808
<rdar://problem/49802649>

Reviewed by Eric Carlson.

Source/WebCore:

As per https://w3c.github.io/webrtc-pc/#set-description,
fire events just before resolving the setRemoteDescription promise.
This ensures that the exposed stream has all necessary tracks from the beginning.
Pending track events are created in LibWebRTCMediaEndpoint and stored in PeerConnectionBackend.

Covered by updated test.

  • Modules/mediastream/PeerConnectionBackend.cpp:

(WebCore::PeerConnectionBackend::setRemoteDescriptionSucceeded):
(WebCore::PeerConnectionBackend::setRemoteDescriptionFailed):
(WebCore::PeerConnectionBackend::addPendingTrackEvent):
(WebCore::PeerConnectionBackend::stop):

  • Modules/mediastream/PeerConnectionBackend.h:
  • Modules/mediastream/RTCPeerConnection.cpp:
  • Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:

(WebCore::LibWebRTCMediaEndpoint::addRemoteTrack):
(WebCore::LibWebRTCMediaEndpoint::addPendingTrackEvent):
(WebCore::LibWebRTCMediaEndpoint::newTransceiver):

  • Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.h:

LayoutTests:

  • webrtc/video-addTrack.html:
7:10 AM Changeset in webkit [244735] by Michael Catanzaro
  • 2 edits in trunk/Source/WebCore/platform/gtk/po

[GTK] [l10n] Updated Ukrainian translation of WebKitGTK+
https://bugs.webkit.org/show_bug.cgi?id=197364

Patch by Yuri Chornoivan <yurchor@ukr.net> on 2019-04-29
Rubber-stamped by Michael Catanzaro.

  • uk.po:
7:08 AM Changeset in webkit [244734] by Michael Catanzaro
  • 1 edit
    1 add in trunk/Source/WebCore/platform/gtk/po

Add Danish translation
https://bugs.webkit.org/show_bug.cgi?id=194850

Patch by scootergrisen <scootergrisen@gmail.com> on 2019-04-29
Rubber-stamped by Michael Catanzaro.

  • da.po: Added.
7:02 AM Changeset in webkit [244733] by achristensen@apple.com
  • 2 edits in trunk/Source/WebKit

Build fix.

  • UIProcess/ios/WKActionSheet.mm:

(-[WKActionSheet updateSheetPosition]):

5:24 AM Changeset in webkit [244732] by commit-queue@webkit.org
  • 4 edits in trunk/Source/WebKit

Unreviewed, rolling out r244648.
https://bugs.webkit.org/show_bug.cgi?id=197363

Lots of tests are crashing (Requested by KaL on #webkit).

Reverted changeset:

"[GTK] Back/forward gesture snapshot always times out"
https://bugs.webkit.org/show_bug.cgi?id=197233
https://trac.webkit.org/changeset/244648

3:14 AM Changeset in webkit [244731] by Carlos Garcia Campos
  • 6 edits in trunk/Source/WebCore

REGRESSION(r244635): [GTK] Wrong background color used in non-dark mode
https://bugs.webkit.org/show_bug.cgi?id=197276

Reviewed by Michael Catanzaro.

Since r244635, we are now getting the frame view background color from the theme. That's correct for dark mode,
but in non-dark mode we still want to use white backgrounds by default. This made a lot of tests to fail.

  • css/CSSValueKeywords.in: Add -webkit-control-background when HAVE(OS_DARK_MODE_SUPPORT).
  • css/html.css: Use -webkit-control-background instead of -apple-system-control-background.
  • page/FrameView.cpp:

(WebCore::FrameView::updateBackgroundRecursively): Use CSSValueWindow instead of CSSValueWindowframe.

  • rendering/RenderThemeGtk.cpp:

(WebCore::RenderThemeGtk::systemColor const): Only get the window background from the theme in dark mode. Handle
also CSSValueWebkitControlBackground.

  • rendering/RenderThemeMac.mm:

(WebCore::RenderThemeMac::systemColor const): Handle CSSValueWebkitControlBackground when HAVE(OS_DARK_MODE_SUPPORT).

2:14 AM Changeset in webkit [244730] by Stephanie Lewis
  • 5 edits in trunk/Tools

run-benchmarks should have an intial prep and restore env call for tasks that are too expensive to do for every iteration
https://bugs.webkit.org/show_bug.cgi?id=197339

Reviewed by Dewei Zhu.

Add two new functions prepare_inital_env and restore_env_after_all_testing around the entire
test run for setup steps that are too expensive to do every iteration.

  • Scripts/webkitpy/benchmark_runner/benchmark_runner.py:

(BenchmarkRunner._run_benchmark):

  • Scripts/webkitpy/benchmark_runner/browser_driver/browser_driver.py:

(BrowserDriver):
(BrowserDriver.prepare_initial_env):
(BrowserDriver.restore_env_after_all_testing):

  • Scripts/webkitpy/benchmark_runner/browser_driver/linux_browser_driver.py:

(LinuxBrowserDriver.prepare_initial_env):
(LinuxBrowserDriver.restore_env_after_all_testing):

  • Scripts/webkitpy/benchmark_runner/browser_driver/osx_browser_driver.py:

(OSXBrowserDriver.prepare_initial_env):
(OSXBrowserDriver.restore_env_after_all_testing):

1:36 AM Changeset in webkit [244729] by Michael Catanzaro
  • 2 edits in trunk/Source/WebKit

[WPE][GTK] window-object-cleared documentation should reference webkit_frame_get_js_context_for_script_world()
https://bugs.webkit.org/show_bug.cgi?id=197357

Reviewed by Carlos Garcia Campos.

We should recommend using the non-deprecated replacement for
webkit_frame_get_javascript_context_for_script_world().

  • WebProcess/InjectedBundle/API/glib/WebKitScriptWorld.cpp:
1:08 AM Changeset in webkit [244728] by jfernandez@igalia.com
  • 31 edits in trunk/LayoutTests

Update the CSS Text WPT test suite
https://bugs.webkit.org/show_bug.cgi?id=197326

Unreviewed test gardening.

LayoutTests/imported/w3c:

  • web-platform-tests/css/css-text/overflow-wrap/overflow-wrap-break-word-007.html:
  • web-platform-tests/css/css-text/overflow-wrap/w3c-import.log:
  • web-platform-tests/css/css-text/parsing/text-transform-valid.html:
  • web-platform-tests/css/css-text/parsing/word-break-invalid-expected.txt:
  • web-platform-tests/css/css-text/parsing/word-break-invalid.html:
  • web-platform-tests/css/css-text/parsing/word-break-valid-expected.txt:
  • web-platform-tests/css/css-text/parsing/word-break-valid.html:
  • web-platform-tests/css/css-text/text-transform/w3c-import.log:
  • web-platform-tests/css/css-text/white-space/break-spaces-003-expected.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-003.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-004.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-005-expected.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-005.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-006-expected.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-006.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-007-expected.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-007.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-008-expected.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-008.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-009-expected.html:
  • web-platform-tests/css/css-text/white-space/break-spaces-009.html:
  • web-platform-tests/css/css-text/white-space/pre-wrap-008-expected.html:
  • web-platform-tests/css/css-text/white-space/pre-wrap-008.html:
  • web-platform-tests/css/css-text/white-space/pre-wrap-015.html:
  • web-platform-tests/css/css-text/white-space/pre-wrap-016.html:
  • web-platform-tests/css/css-text/white-space/textarea-break-spaces-001-expected.html:
  • web-platform-tests/css/css-text/white-space/w3c-import.log:

LayoutTests:

Note: See TracTimeline for information about the timeline view.