WebKit-https.git
3 months ago[Mac Debug] Layout Test resize-observer/observe-element-from-other-frame.html is...
commit-queue@webkit.org [Wed, 24 Apr 2019 16:21:31 +0000 (16:21 +0000)]
[Mac Debug] Layout Test resize-observer/observe-element-from-other-frame.html is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=197063

Patch by Cathie Chen <cathiechen@igalia.com> on 2019-04-24
Reviewed by Ryosuke Niwa.

RenderingUpdate timer might be fired after ResizeTestHelper's. This would produce a flaky failure.
In order to reduce this failure, extend ResizeTestHelper.TIMEOUT like in bug 177484.

* resize-observer/resources/resizeTestHelper.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244593 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRebaseline js/dom/dom-static-property-for-in-iteration.html after r244581
tsavell@apple.com [Wed, 24 Apr 2019 15:52:19 +0000 (15:52 +0000)]
Rebaseline js/dom/dom-static-property-for-in-iteration.html after r244581
https://bugs.webkit.org/show_bug.cgi?id=197218

Unreviewed test gardening.

* platform/mac-wk2/js/dom/dom-static-property-for-in-iteration-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244592 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoFix -Wc++11-narrowing on unsigned char platforms like FreeBSD/aarch64
commit-queue@webkit.org [Wed, 24 Apr 2019 15:51:39 +0000 (15:51 +0000)]
Fix -Wc++11-narrowing on unsigned char platforms like FreeBSD/aarch64
https://bugs.webkit.org/show_bug.cgi?id=197148

Patch by Greg V <greg@unrelenting.technology> on 2019-04-24
Reviewed by Alex Christensen.

* contentextensions/DFACombiner.cpp:
* contentextensions/NFAToDFA.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244591 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoURL set by document.open() is communicated with the WebPageProxy but not the WebFrame...
cdumez@apple.com [Wed, 24 Apr 2019 15:44:28 +0000 (15:44 +0000)]
URL set by document.open() is communicated with the WebPageProxy but not the WebFrameProxy
https://bugs.webkit.org/show_bug.cgi?id=197214
<rdar://problem/49237544>

Reviewed by Alex Christensen.

Source/WebKit:

URL set by document.open() is communicated with the WebPageProxy (since r244361) but not the
WebFrameProxy. This patch fixes this.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _mainFrameURL]):
* UIProcess/API/Cocoa/WKWebViewPrivate.h:
* UIProcess/FrameLoadState.cpp:
(WebKit::FrameLoadState::didExplicitOpen):
* UIProcess/FrameLoadState.h:
* UIProcess/WebFrameProxy.cpp:
(WebKit::WebFrameProxy::didExplicitOpen):
* UIProcess/WebFrameProxy.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::didExplicitOpenForFrame):

Tools:

Extend API test coverage.

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

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244590 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoX-Frame-Options header should be ignored when frame-ancestors CSP directive is present
cdumez@apple.com [Wed, 24 Apr 2019 15:43:09 +0000 (15:43 +0000)]
X-Frame-Options header should be ignored when frame-ancestors CSP directive is present
https://bugs.webkit.org/show_bug.cgi?id=197226
<rdar://problem/50155649>

Reviewed by Alex Christensen.

Source/WebCore:

X-Frame-Options header should be ignored when frame-ancestors CSP directive is present:
- https://www.w3.org/TR/CSP3/#frame-ancestors-and-frame-options

Specification says:
"""
In order to allow backwards-compatible deployment, the frame-ancestors directive _obsoletes_ the
X-Frame-Options header. If a resource is delivered with an policy that includes a directive named
frame-ancestors and whose disposition is "enforce", then the X-Frame-Options header MUST be ignored.
"""

Gecko and Blink follow the specification, WebKit does not. As a result, page [1] is broken with
WebKit-only on Schwab.com. The page height is wrong and you cannot see all the ETFs as a result.

[1] https://www.schwab.com/public/schwab/investing/investment_help/investment_research/etf_research/etfs.html?&path=/Prospect/Research/etfs/overview/oneSourceETFs.asp

Test: http/tests/security/contentSecurityPolicy/1.1/frame-ancestors/frame-ancestors-overrides-X-Frames-Options.html

* loader/DocumentLoader.cpp:
(WebCore::DocumentLoader::responseReceived):
* page/csp/ContentSecurityPolicy.cpp:
(WebCore::ContentSecurityPolicy::overridesXFrameOptions const):
* page/csp/ContentSecurityPolicy.h:
* page/csp/ContentSecurityPolicyDirectiveList.h:
(WebCore::ContentSecurityPolicyDirectiveList::hasFrameAncestorsDirective const):

Source/WebKit:

* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::shouldInterruptLoadForCSPFrameAncestorsOrXFrameOptions):

LayoutTests:

Add layout test coverage.

* http/tests/security/contentSecurityPolicy/1.1/frame-ancestors/frame-ancestors-overrides-X-Frames-Options-expected.txt: Added.
* http/tests/security/contentSecurityPolicy/1.1/frame-ancestors/frame-ancestors-overrides-X-Frames-Options.html: Added.
* http/tests/security/contentSecurityPolicy/resources/frame-ancestors-self-x-frame-options-deny.pl: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244589 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ContentChangeObserver] Do not use the global _WKContentChange in WebKitLegacy
zalan@apple.com [Wed, 24 Apr 2019 15:38:23 +0000 (15:38 +0000)]
[ContentChangeObserver] Do not use the global _WKContentChange in WebKitLegacy
https://bugs.webkit.org/show_bug.cgi?id=196286
<rdar://problem/49364417>

Reviewed by Simon Fraser.

By reporting WKContentIndeterminateChange in sendMouseMoveEvent enables us to remove the global _WKContentChange state.
Using _WKContentChange is fine as long as only the observed frame reports content change during the synthetic click event.
In case of multiple frames, we should really consult the local state instead.
Unfortunately sendMouseMoveEvent has no access to the observed Document object so we can't really identify the observed content change.
WKContentIndeterminateChange triggers asynchronous decision making at the callsite and in the callback we have access
to the active Document/ContentChangeObverver object and can report the correct state.
This is inline with current WebKit(WK2) behaviour.

Manually tested with a WebKitLegacy test app.

* SourcesCocoa.txt:
* WebCore.xcodeproj/project.pbxproj:
* page/ios/ContentChangeObserver.h:
(WebCore::ContentChangeObserver::setHasNoChangeState):
(WebCore::ContentChangeObserver::setHasIndeterminateState):
(WebCore::ContentChangeObserver::setHasVisibleChangeState):
(WebCore::ContentChangeObserver::setObservedContentState): Deleted.
* page/ios/EventHandlerIOS.mm:
(WebCore::EventHandler::mouseMoved):
* platform/ios/wak/WAKWindow.mm:
(-[WAKWindow sendMouseMoveEvent:contentChange:]):
* platform/ios/wak/WKContentObservation.cpp: Removed.
* platform/ios/wak/WKContentObservation.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244588 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[GTK][GStreamer] Flaky ASSERTION FAILED: m_lock.isHeld() in TextureMapperPlatformLaye...
philn@webkit.org [Wed, 24 Apr 2019 14:47:33 +0000 (14:47 +0000)]
[GTK][GStreamer] Flaky ASSERTION FAILED: m_lock.isHeld() in TextureMapperPlatformLayerProxy
https://bugs.webkit.org/show_bug.cgi?id=196739

Reviewed by Xabier Rodriguez-Calvar.

The crash was triggered because m_videoDecoderPlatform not being
explicitely set, its value would be inferred as one of the enum
class values. Making it Optional avoids this issue.

* platform/graphics/gstreamer/MediaPlayerPrivateGStreamerBase.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244587 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRename _highlightLongPressCanClick and only add gesture recognizer when necessary
dino@apple.com [Wed, 24 Apr 2019 12:23:20 +0000 (12:23 +0000)]
Rename _highlightLongPressCanClick and only add gesture recognizer when necessary
https://bugs.webkit.org/show_bug.cgi?id=197231
<rdar://problem/50164234>

Reviewed by Antoine Quint.

Rename _highlightLongPressCanClick to _longPressCanClick since it will be
used in other places.

Only attach the _highlightLongPressGestureRecognizer when we're not
using long presses for preview. This might revert in the future, if we
can set up an appropriate gesture resolution between the two.

* UIProcess/ios/WKContentViewInteraction.h:
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView setupInteraction]):
(-[WKContentView _webTouchEvent:preventsNativeGestures:]):
(-[WKContentView _highlightLongPressRecognized:]):
(-[WKContentView _interactionShouldBeginFromPreviewItemController:forPosition:]):
(-[WKContentView _presentedViewControllerForPreviewItemController:]):
(-[WKContentView _previewItemControllerDidCancelPreview:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244586 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[Build][GStreamer] Update to 1.16.0
philn@webkit.org [Wed, 24 Apr 2019 12:18:16 +0000 (12:18 +0000)]
[Build][GStreamer] Update to 1.16.0
https://bugs.webkit.org/show_bug.cgi?id=197157

Reviewed by Xabier Rodriguez-Calvar.

The upstreamed patches were removed and a new one was added for a
regression detected in the scaletempo GStreamer element. The GNOME
Flatpak was migrated to 3.32 and I switched a few modules to Meson
on the way, as it is the preferred build system now in GStreamer
1.16.

* flatpak/files/httpd-autogen.sh:
* flatpak/flatpakutils.py:
(FlatpakPackage.install):
(WebkitFlatpak.run_in_sandbox):
* flatpak/org.webkit.CommonModules.yaml:
* flatpak/org.webkit.WebKit.yaml:
* gstreamer/jhbuild.modules:
* gstreamer/patches/gst-plugins-bad-0001-aomenc-Add-support-for-10-12bit-decoding.patch: Removed.
* gstreamer/patches/gst-plugins-bad-0002-aomenc-Handle-8-bit_depth-images-with-AOM_IMG_FMT_HI.patch: Removed.
* gstreamer/patches/gst-plugins-bad-0003-fdkaacenc-Remove-MODE_2_1.patch: Removed.
* gstreamer/patches/gst-plugins-bad-0004-fdkaacdec-Use-WAV-channel-mapping-instead-of-interleave-setting.patch: Removed.
* gstreamer/patches/gst-plugins-bad-0005-mssdemux-fix-protection-data-double-free.patch: Removed.
* gstreamer/patches/gst-plugins-base-0001-parsebin-Post-STREAM_COLLECTION-on-EVENT_STREAM_COLL.patch: Removed.
* gstreamer/patches/gst-plugins-good-0001-qtdemux-Detect-and-expose-CEA-608-708-Closed-Caption.patch: Removed.
* gstreamer/patches/gst-plugins-good-0001-qtdemux-Do-not-run-the-preferred-decryptor-context-q.patch: Removed.
* gstreamer/patches/gst-plugins-good-0002-qtdemux-Do-not-unref-a-NULL-stream_tags.patch: Removed.
* gstreamer/patches/gst-plugins-good-0002-qtdemux-add-context-for-a-preferred-protection.patch: Removed.
* gstreamer/patches/gst-plugins-good-0003-qtdemux-Clarify-field-name-about-stream-encryption-s.patch: Removed.
* gstreamer/patches/gst-plugins-good-0003-qtdemux-also-push-buffers-without-encryption-info-in.patch: Removed.
* gstreamer/patches/gst-plugins-good-0004-qtdemux-Add-initial-support-for-AV1-demuxing.patch: Removed.
* gstreamer/patches/gst-plugins-good-0005-qtdemux-Extract-AV1-codec_data-and-put-it-in-the-cap.patch: Removed.
* gstreamer/patches/gst-plugins-good-0006-qtdemux-Recognize-more-AV1-atoms.patch: Removed.
* gstreamer/patches/gst-plugins-good-0011-matroska-Add-the-WebM-encrypted-content-support-in-m.patch: Removed.
* gstreamer/patches/gst-plugins-good-0012-matroskdemux-do-not-use-MapInfo.data-after-unmapping.patch: Removed.
* gstreamer/patches/gst-plugins-good-0013-Avoid-warning-when-reporting-about-decryptors.patch: Removed.
* gstreamer/patches/gst-plugins-good-0014-pulse-Mark-default-devices-as-default.patch: Removed.
* gstreamer/patches/gst-plugins-good-scaletempo-Advertise-interleaved-layout-in-caps-temp.patch: Added.
* gstreamer/patches/gstreamer-0001-protection-Add-a-new-definition-for-unspecified-syst.patch: Removed.
* gstreamer/patches/gstreamer-0002-protection-Fix-the-string-to-define-unspecified-syst.patch: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244585 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[GStreamer] Crash in AudioTrackPrivate with playbin3 enabled
philn@webkit.org [Wed, 24 Apr 2019 12:11:50 +0000 (12:11 +0000)]
[GStreamer] Crash in AudioTrackPrivate with playbin3 enabled
https://bugs.webkit.org/show_bug.cgi?id=196913

Reviewed by Xabier Rodriguez-Calvar.

The crash was due to a playbin3 code path being triggered during
MSE playback, which is not supposed to work in playbin3 anyway.
The problem is that setting the USE_PLAYBIN3 environment variable
to "1" makes the GStreamer playback plugin register the playbin3
element under the playbin name. So that leads to playbin3 being
used everywhere in WebKit where we assume the playbin element is
used. So the proposed solution is to:

- use a WebKit-specific environment variable instead of the
GStreamer USE_PLAYBIN3 variable.
- emit a warning if the USE_PLAYBIN3 environment variable is
detected. We can't unset it ourselves for security reasons.

The patch also includes a code cleanup of the player method
handling the pipeline creation. The previous code had a bug
leading to playbin3 being used for MSE.

* platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
(WebCore::MediaPlayerPrivateGStreamer::createGSTPlayBin):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244584 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[GTK] MiniBrowser: also set the passed bg-color when receiving arguments
carlosgc@webkit.org [Wed, 24 Apr 2019 07:53:58 +0000 (07:53 +0000)]
[GTK] MiniBrowser: also set the passed bg-color when receiving arguments
https://bugs.webkit.org/show_bug.cgi?id=197156

Reviewed by Michael Catanzaro.

The background color is only set when MiniBrowser is launched without arguments. This regressed when tabs
support was added.

* MiniBrowser/gtk/BrowserTab.c:
(browser_tab_set_background_color): Set the passed in color as web view background color.
* MiniBrowser/gtk/BrowserTab.h:
* MiniBrowser/gtk/BrowserWindow.c:
(browser_window_init): Initialize backgroundColor.
(browser_window_append_view): Call browser_tab_set_background_color().
(browser_window_set_background_color): Save the passed in color. This function should now be called before tabs
are added.
* MiniBrowser/gtk/main.c:
(main): Call browser_window_set_background_color() before creating the tabs.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244583 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAX: Remove deprecated Accessibility Object Model events
cfleizach@apple.com [Wed, 24 Apr 2019 07:46:08 +0000 (07:46 +0000)]
AX: Remove deprecated Accessibility Object Model events
https://bugs.webkit.org/show_bug.cgi?id=197073
<rdar://problem/50027819>

Reviewed by Ryosuke Niwa.

Source/WebCore:

Test: accessibility/mac/replace-text-with-range.html

* DerivedSources.make:
* Sources.txt:
* WebCore.xcodeproj/project.pbxproj:
* accessibility/AccessibilityListBoxOption.cpp:
(WebCore::AccessibilityListBoxOption::setSelected):
* accessibility/AccessibilityMediaObject.cpp:
(WebCore::AccessibilityMediaObject::increment):
(WebCore::AccessibilityMediaObject::decrement):
* accessibility/AccessibilityMenuListOption.cpp:
(WebCore::AccessibilityMenuListOption::setSelected):
* accessibility/AccessibilityNodeObject.cpp:
(WebCore::AccessibilityNodeObject::increment):
(WebCore::AccessibilityNodeObject::decrement):
* accessibility/AccessibilityObject.cpp:
(WebCore::AccessibilityObject::press):
(WebCore::AccessibilityObject::replaceTextInRange):
(WebCore::AccessibilityObject::scrollToMakeVisible const):
(WebCore::AccessibilityObject::shouldDispatchAccessibilityEvent const): Deleted.
(WebCore::AccessibilityObject::dispatchAccessibilityEvent const): Deleted.
(WebCore::AccessibilityObject::dispatchAccessibilityEventWithType const): Deleted.
(WebCore::AccessibilityObject::dispatchAccessibleSetValueEvent const): Deleted.
* accessibility/AccessibilityObject.h:
* accessibility/AccessibilityRenderObject.cpp:
(WebCore::AccessibilityRenderObject::setFocused):
(WebCore::AccessibilityRenderObject::setValue):
* accessibility/AccessibilityScrollbar.cpp:
(WebCore::AccessibilityScrollbar::setValue):
* accessibility/AccessibilitySlider.cpp:
(WebCore::AccessibilitySlider::setValue):
* accessibility/ios/AccessibilityObjectIOS.mm:
(WebCore::AccessibilityObject::accessibilityPlatformIncludesObject const):
(WebCore::AccessibilityObject::hasAccessibleDismissEventListener const): Deleted.
* accessibility/ios/WebAccessibilityObjectWrapperIOS.mm:
(-[WebAccessibilityObjectWrapper accessibilityPerformEscape]): Deleted.
(-[WebAccessibilityObjectWrapper accessibilityElementDidBecomeFocused]): Deleted.
* accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(-[WebAccessibilityObjectWrapper accessibilityPerformShowMenuAction]):
* dom/Element.idl:
* dom/EventNames.h:
* dom/EventNames.in:
* html/HTMLAttributeNames.in:
* html/HTMLElement.cpp:
(WebCore::HTMLElement::createEventHandlerNameMap):
* page/Settings.yaml:
* testing/InternalSettings.cpp:
(WebCore::InternalSettings::Backup::Backup):
(WebCore::InternalSettings::Backup::restoreTo):
(WebCore::InternalSettings::setAccessibilityEventsEnabled): Deleted.
* testing/InternalSettings.h:
* testing/InternalSettings.idl:

Source/WebKit:

* Platform/spi/ios/AccessibilitySupportSPI.h:
* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView _initializeWithConfiguration:]):
(-[WKWebView dealloc]):
(accessibilityEventsEnabledChangedCallback): Deleted.
(-[WKWebView _updateAccessibilityEventsEnabled]): Deleted.
* UIProcess/API/Cocoa/WKWebViewInternal.h:
* UIProcess/Cocoa/WKFullKeyboardAccessWatcher.mm:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::updateAccessibilityEventsEnabled): Deleted.
* UIProcess/WebPageProxy.h:
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::updateAccessibilityEventsEnabled): Deleted.
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

Source/WTF:

* wtf/Platform.h:

LayoutTests:

* accessibility/mac/AOM-event-accessiblesetvalue-expected.txt: Removed.
* accessibility/mac/AOM-event-accessiblesetvalue.html: Removed.
* accessibility/mac/AOM-events-all-expected.txt: Removed.
* accessibility/mac/AOM-events-all.html: Removed.
* accessibility/mac/AOM-events-expected.txt: Removed.
* accessibility/mac/AOM-events-webarea-crash-expected.txt: Removed.
* accessibility/mac/AOM-events-webarea-crash.html: Removed.
* accessibility/mac/AOM-events.html: Removed.
* accessibility/mac/replace-text-with-range-expected.txt: Replaced.
* accessibility/mac/replace-text-with-range.html: Replaced.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244582 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoMove Ad Click Attribution from internal feature to experimental feature
wilander@apple.com [Wed, 24 Apr 2019 04:16:30 +0000 (04:16 +0000)]
Move Ad Click Attribution from internal feature to experimental feature
https://bugs.webkit.org/show_bug.cgi?id=197218
<rdar://problem/47651691>

Reviewed by Brent Fulgham.

Source/WebKit:

* Shared/WebPreferences.yaml:

Tools:

* DumpRenderTree/TestOptions.cpp:
(TestOptions::TestOptions):

LayoutTests:

This patch removes the TestRunner instruction for turning on the internal
feature.

* http/tests/adClickAttribution/anchor-tag-attributes-validation.html:
* http/tests/adClickAttribution/attribution-conversion-through-image-redirect-without-priority.html:
* http/tests/adClickAttribution/conversion-disabled-in-ephemeral-session.html:
* http/tests/adClickAttribution/second-attribution-converted-with-higher-priority.html:
* http/tests/adClickAttribution/second-attribution-converted-with-lower-priority.html:
* http/tests/adClickAttribution/second-conversion-with-higher-priority.html:
* http/tests/adClickAttribution/second-conversion-with-lower-priority.html:
* http/tests/adClickAttribution/send-attribution-conversion-request.html:
* http/tests/adClickAttribution/store-ad-click-attribution.html:
* http/tests/adClickAttribution/store-disabled-in-ephemeral-session.html:
* http/tests/contentextensions/block-ad-click-attribution.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244581 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[JSC] Shrink DFG::MinifiedNode
ysuzuki@apple.com [Wed, 24 Apr 2019 04:06:08 +0000 (04:06 +0000)]
[JSC] Shrink DFG::MinifiedNode
https://bugs.webkit.org/show_bug.cgi?id=197224

Reviewed by Filip Pizlo.

Since it is kept alive with compiled DFG code, we should shrink it to save memory.
If it is effective, we should consider minimizing these OSR exit data more aggressively.

* dfg/DFGMinifiedNode.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244580 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoLICM incorrectly assumes it'll never insert a node which provably OSR exits
sbarati@apple.com [Wed, 24 Apr 2019 02:14:23 +0000 (02:14 +0000)]
LICM incorrectly assumes it'll never insert a node which provably OSR exits
https://bugs.webkit.org/show_bug.cgi?id=196721
<rdar://problem/49556479>

Reviewed by Filip Pizlo.

JSTests:

* stress/licm-should-handle-if-a-hoist-causes-a-provable-osr-exit.js: Added.
(foo):

Source/JavaScriptCore:

Previously, we assumed LICM could never hoist code that caused us
to provably OSR exit. This is a bad assumption, as we may very well
hoist such code. Obviously hoisting such code is not ideal. We shouldn't
hoist something we provably know will OSR exit. However, this is super rare,
and the phase is written in such a way where it's easier to gracefully
handle this case than to prevent us from hoisting such code.

If we wanted to ensure we never hoisted code that would provably exit, we'd
have to teach the phase to know when it inserted code that provably exits. I
saw two ways to do that:
1: Save and restore the AI state before actually hoisting.
2: Write an analysis that can determine if such a node would exit.

(1) is bad because it costs in memory and compile time. (2) will inevitably
have bugs as running into this condition is rare.

So instead of (1) or (2), I opted to have LICM gracefully handle when
it causes a provable exit. When we encounter this, we mark all blocks
in the loop as !cfaHasVisited and !cfaDidFinish.

* dfg/DFGLICMPhase.cpp:
(JSC::DFG::LICMPhase::attemptHoist):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244579 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[JSC] Use node index as DFG::MinifiedID
ysuzuki@apple.com [Wed, 24 Apr 2019 01:50:50 +0000 (01:50 +0000)]
[JSC] Use node index as DFG::MinifiedID
https://bugs.webkit.org/show_bug.cgi?id=197186

Reviewed by Saam Barati.

DFG Nodes can be identified with index if the graph is given. We should use unsigned index as a DFG::MinifiedID's underlying
source instead of Node* to reduce the size of VariableEvent from 16 to 12. Vector<VariableEvent> is the main data in DFG's OSR
tracking. It is kept after DFG compilation is done to make OSR work. We saw that this is allocated with large size in GMail.

* JavaScriptCore.xcodeproj/project.pbxproj:
* bytecode/DataFormat.h:
* bytecode/ValueRecovery.h:
* dfg/DFGGenerationInfo.h:
* dfg/DFGMinifiedID.h:
(JSC::DFG::MinifiedID::MinifiedID):
(JSC::DFG::MinifiedID::operator! const):
(JSC::DFG::MinifiedID::operator== const):
(JSC::DFG::MinifiedID::operator!= const):
(JSC::DFG::MinifiedID::operator< const):
(JSC::DFG::MinifiedID::operator> const):
(JSC::DFG::MinifiedID::operator<= const):
(JSC::DFG::MinifiedID::operator>= const):
(JSC::DFG::MinifiedID::hash const):
(JSC::DFG::MinifiedID::dump const):
(JSC::DFG::MinifiedID::isHashTableDeletedValue const):
(JSC::DFG::MinifiedID::fromBits):
(JSC::DFG::MinifiedID::bits const):
(JSC::DFG::MinifiedID::invalidIndex):
(JSC::DFG::MinifiedID::otherInvalidIndex):
(JSC::DFG::MinifiedID::node const): Deleted.
(JSC::DFG::MinifiedID::invalidID): Deleted.
(JSC::DFG::MinifiedID::otherInvalidID): Deleted.
* dfg/DFGMinifiedIDInlines.h: Copied from Source/JavaScriptCore/dfg/DFGMinifiedNode.cpp.
(JSC::DFG::MinifiedID::MinifiedID):
* dfg/DFGMinifiedNode.cpp:
* dfg/DFGValueSource.h:
(JSC::DFG::ValueSource::ValueSource):
* dfg/DFGVariableEvent.h:
(JSC::DFG::VariableEvent::dataFormat const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244578 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Inspector: REGRESSION(r238168): Network: preserve log doesn't show main frame...
drousso@apple.com [Wed, 24 Apr 2019 01:28:49 +0000 (01:28 +0000)]
Web Inspector: REGRESSION(r238168): Network: preserve log doesn't show main frame main resource changes
https://bugs.webkit.org/show_bug.cgi?id=197220
<rdar://problem/49960464>

Reviewed by Timothy Hatcher.

* UserInterface/Views/NetworkTableContentView.js:
(WI.NetworkTableContentView.prototype.reset):
(WI.NetworkTableContentView.prototype._addCollection):
(WI.NetworkTableContentView.prototype._resetCollection): Added.
(WI.NetworkTableContentView.prototype._changeCollection):
(WI.NetworkTableContentView.prototype._mainResourceDidChange):
We want to add the frame's main resource in the case that it's the main frame, but we also
want to clear the main collection of data (e.g. if viewing an imported HAR).
Drive-by: `_groupByDOMNodeNavigationItem` was renamed in a previous patch.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244577 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Inspector: Network: support drag/drop for importing
drousso@apple.com [Wed, 24 Apr 2019 01:20:32 +0000 (01:20 +0000)]
Web Inspector: Network: support drag/drop for importing
https://bugs.webkit.org/show_bug.cgi?id=197221

Reviewed by Timothy Hatcher.

* UserInterface/Views/NetworkTabContentView.js:
(WI.NetworkTabContentView.prototype.async.handleFileDrop): Added.
* UserInterface/Views/NetworkTableContentView.js:
(WI.NetworkTableContentView.prototype.processHAR): Added.
(WI.NetworkTableContentView.prototype._importHAR):

* UserInterface/Base/FileUtilities.js:
(WI.FileUtilities.async readText):
Only `await` if the result is a `Promise`.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244576 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd Xcode version check for Header post-processing scripts
krollin@apple.com [Wed, 24 Apr 2019 00:52:41 +0000 (00:52 +0000)]
Add Xcode version check for Header post-processing scripts
https://bugs.webkit.org/show_bug.cgi?id=197116
<rdar://problem/50058968>

Reviewed by Brent Fulgham.

There are several places in our Xcode projects that post-process
header files after they've been exported. Because of XCBuild, we're
moving to a model where the post-processing is performed at the same
time the header files are exported, rather than as a distinct
post-processing step. This patch disables the distinct step when the
inline processing is available.

In practice, this means prefixing appropriate post-processing Custom
Build phases with:

if [ "${XCODE_VERSION_MAJOR}" -ge "1100" -a "${USE_NEW_BUILD_SYSTEM}" = "YES" ]; then
    # In this configuration, post-processing is performed at the same time as copying in the postprocess-header-rule script, so there's no need for this separate step.
    exit 0
fi

Source/JavaScriptCore:

* JavaScriptCore.xcodeproj/project.pbxproj:

Source/WebKit:

* WebKit.xcodeproj/project.pbxproj:

Source/WebKitLegacy:

* WebKitLegacy.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244575 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244558.
commit-queue@webkit.org [Wed, 24 Apr 2019 00:29:55 +0000 (00:29 +0000)]
Unreviewed, rolling out r244558.
https://bugs.webkit.org/show_bug.cgi?id=197219

Causing crashes on iOS Sim Release and Debug (Requested by
ShawnRoberts on #webkit).

Reverted changeset:

"Remove DeprecatedOptional"
https://bugs.webkit.org/show_bug.cgi?id=197161
https://trac.webkit.org/changeset/244558

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244574 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[iOS] QuickLook documents loaded from file: URLs should be allowed to perform same...
aestes@apple.com [Wed, 24 Apr 2019 00:24:27 +0000 (00:24 +0000)]
[iOS] QuickLook documents loaded from file: URLs should be allowed to perform same-document navigations
https://bugs.webkit.org/show_bug.cgi?id=196749
<rdar://problem/35773454>

Reviewed by Daniel Bates.

Source/WebCore:

QuickLook previews are in a non-local origin defined by a unique x-apple-ql-id: URL, which
isolates the origin that hosted the document from the document preview itself. When a
QuickLook document is loaded as a file: URL, SecurityOrigin's protections against loading
local resources from non-local origins prevented navigations like location.reload() and
fragment navigations.

To allow reloads and same-document navigations in QuickLook documents loaded from file: URLs,
we should grant the QuickLook document's SecurityOrigin access to the file path that loaded
the preview.

Added a new API test.

* dom/Document.cpp:
(WebCore::Document::applyQuickLookSandbox):
* page/SecurityOrigin.cpp:
(WebCore::SecurityOrigin::createNonLocalWithAllowedFilePath):
(WebCore::SecurityOrigin::canDisplay const):
* page/SecurityOrigin.h:

Tools:

Added a new QuickLook API test and added new expectations to existing QuickLook tests.

* TestWebKitAPI/Tests/WebKitCocoa/QuickLook.mm:
(-[QuickLookDelegate webView:didStartProvisionalNavigation:]):
(-[QuickLookDelegate webView:didFinishNavigation:]):
(-[QuickLookDelegate _webView:didFailNavigation:withError:userInfo:]):
(-[QuickLookDelegate webView:didFailProvisionalNavigation:withError:]):
(runTest):
(runTestDecideBeforeLoading):
(runTestDecideAfterLoading):
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244573 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRemove Ad Click Attribution data when removing website data
wilander@apple.com [Wed, 24 Apr 2019 00:13:00 +0000 (00:13 +0000)]
Remove Ad Click Attribution data when removing website data
https://bugs.webkit.org/show_bug.cgi?id=197215
<rdar://problem/47668988>

Reviewed by Chris Dumez.

Source/WebKit:

This patch adds a new WebsiteDataType called AdClickAttributions and flags
it as owned by the network process.

The new website data type is added to the Cocoa API layer for
website data records management.

When either of the two WebsiteDataStore::removeData() functions calls
NetworkProcess::deleteWebsiteData() or
NetworkProcess::deleteWebsiteDataForOrigins() over IPC, the network
process now calls into AdClickAttributionManager::clear() and
AdClickAttributionManager::clearForRegistrableDomain() respectively,
if the website datatypes include WebsiteDataType::AdClickAttributions.

The patch removes a couple of completion handlers in NetworkSession function
calls are because those are not asynchronous.

* NetworkProcess/AdClickAttributionManager.cpp:
(WebKit::AdClickAttributionManager::clear):
(WebKit::AdClickAttributionManager::clearForRegistrableDomain):
(WebKit::AdClickAttributionManager::toString const):
* NetworkProcess/AdClickAttributionManager.h:
* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::deleteWebsiteData):
(WebKit::NetworkProcess::deleteWebsiteDataForOrigins):
(WebKit::NetworkProcess::clearAdClickAttribution):
* NetworkProcess/NetworkSession.cpp:
(WebKit::NetworkSession::clearAdClickAttribution):
(WebKit::NetworkSession::clearAdClickAttributionForRegistrableDomain):
* NetworkProcess/NetworkSession.h:
* Shared/WebsiteData/WebsiteData.cpp:
(WebKit::WebsiteData::ownerProcess):
* Shared/WebsiteData/WebsiteDataType.h:
* UIProcess/API/C/WKWebsiteDataStoreRef.cpp:
(WKWebsiteDataStoreClearAdClickAttributionsThroughWebsiteDataRemoval):
* UIProcess/API/C/WKWebsiteDataStoreRef.h:
* UIProcess/API/Cocoa/WKWebsiteDataRecord.mm:
(dataTypesToString):
* UIProcess/API/Cocoa/WKWebsiteDataRecordInternal.h:
(WebKit::toWebsiteDataType):
(WebKit::toWKWebsiteDataTypes):
* UIProcess/API/Cocoa/WKWebsiteDataRecordPrivate.h:
* UIProcess/API/Cocoa/WKWebsiteDataStore.mm:
(+[WKWebsiteDataStore _allWebsiteDataTypesIncludingPrivate]):

Tools:

This patch adds test infrastructure to delete Ad Click Attribution data
through WebsiteDataStore::removeData().

* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::clearAdClickAttribution):
(WTR::TestRunner::clearAdClickAttributionsThroughWebsiteDataRemoval):
* WebKitTestRunner/InjectedBundle/TestRunner.h:
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::clearAdClickAttributionsThroughWebsiteDataRemoval):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):

LayoutTests:

* http/tests/adClickAttribution/clear-through-website-data-removal-expected.txt: Added.
* http/tests/adClickAttribution/clear-through-website-data-removal.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244572 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agowebkitpy: Kill unresponsive worker processes
jbedard@apple.com [Tue, 23 Apr 2019 23:52:43 +0000 (23:52 +0000)]
webkitpy: Kill unresponsive worker processes
https://bugs.webkit.org/show_bug.cgi?id=197205
<rdar://problem/50104388>

Reviewed by Lucas Forschler.

* Scripts/webkitpy/common/message_pool.py:
(_MessagePool.__init__): Add close timeout.
(_MessagePool._close): Send termination before joining workers, kill unresponsive workers.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244571 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoReturn annotated text checking strings via UIWKDocumentContext
timothy_horton@apple.com [Tue, 23 Apr 2019 23:46:46 +0000 (23:46 +0000)]
Return annotated text checking strings via UIWKDocumentContext
https://bugs.webkit.org/show_bug.cgi?id=197177
<rdar://problem/49064839>

Reviewed by Ryosuke Niwa.

* WebProcess/WebPage/Cocoa/TextCheckingControllerProxy.h:
* WebProcess/WebPage/Cocoa/TextCheckingControllerProxy.mm:
(WebKit::TextCheckingControllerProxy::annotatedSubstringBetweenPositions):
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::requestDocumentEditingContext):
Respect the UIWKDocumentRequestAnnotation flag, returning an attributed
string containing the platform text checking annotations.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244570 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Inspector: Debugger: remove ASSERT_NOT_REACHED where it's possible to reach
drousso@apple.com [Tue, 23 Apr 2019 22:30:06 +0000 (22:30 +0000)]
Web Inspector: Debugger: remove ASSERT_NOT_REACHED where it's possible to reach
https://bugs.webkit.org/show_bug.cgi?id=197210
<rdar://problem/48462912>

Reviewed by Joseph Pecoraro.

* inspector/agents/page/PageDebuggerAgent.cpp:
(WebCore::PageDebuggerAgent::didAddEventListener):
(WebCore::PageDebuggerAgent::didPostMessage):

* inspector/InspectorInstrumentation.cpp:
(WebCore::InspectorInstrumentation::consoleAgentEnabled):
(WebCore::InspectorInstrumentation::timelineAgentEnabled):
Drive-by: add additional `FAST_RETURN_IF_NO_FRONTENDS`.
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244569 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd unit tests for WKWebView.serverTrust
achristensen@apple.com [Tue, 23 Apr 2019 21:46:46 +0000 (21:46 +0000)]
Add unit tests for WKWebView.serverTrust
https://bugs.webkit.org/show_bug.cgi?id=197202

Source/ThirdParty/libwebrtc:

Reviewed by Youenn Fablet.

* libwebrtc.xcodeproj/project.pbxproj:
Move boringssl files from libwebrtc target to boringssl target.
Also, add pkcs7 files to boringssl static library.

Tools:

Reviewed by Youenn Fablet

* TestWebKitAPI/TCPServer.cpp:
(TestWebKitAPI::deleter<BIO>::operator()):
(TestWebKitAPI::deleter<SSL>::operator()):
(TestWebKitAPI::deleter<X509>::operator()):
(TestWebKitAPI::deleter<SSL_CTX>::operator()):
(TestWebKitAPI::deleter<EVP_PKEY>::operator()):
(TestWebKitAPI::TCPServer::TCPServer):
(TestWebKitAPI::TCPServer::listenForConnections):
* TestWebKitAPI/TCPServer.h:
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKitCocoa/Challenge.mm:
(TestWebKitAPI::TEST):
(verifyCertificateAndPublicKey):
(-[ServerTrustDelegate webView:didFinishNavigation:]):
(-[ServerTrustDelegate webView:didReceiveAuthenticationChallenge:completionHandler:]):
* TestWebKitAPI/Tests/WebKitCocoa/Download.mm:
(TestWebKitAPI::downloadAtRate):
* TestWebKitAPI/Tests/WebKitCocoa/PDFLinkReferrer.mm:
(TEST):
* TestWebKitAPI/Tests/WebKitCocoa/WKNavigationResponse.mm:
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244568 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244556.
commit-queue@webkit.org [Tue, 23 Apr 2019 21:38:10 +0000 (21:38 +0000)]
Unreviewed, rolling out r244556.
https://bugs.webkit.org/show_bug.cgi?id=197212

Causing build failures on multiple builders (Requested by
ShawnRoberts on #webkit).

Reverted changeset:

"Create AVFoundationSoftLink.{h,mm} to reduce duplicate code"
https://bugs.webkit.org/show_bug.cgi?id=197171
https://trac.webkit.org/changeset/244556

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244567 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Inspector: Uncaught Exception: null is not an object (evaluating 'this.ownerDocum...
drousso@apple.com [Tue, 23 Apr 2019 21:06:43 +0000 (21:06 +0000)]
Web Inspector: Uncaught Exception: null is not an object (evaluating 'this.ownerDocument.frameIdentifier')
https://bugs.webkit.org/show_bug.cgi?id=196420
<rdar://problem/49444205>

Reviewed by Timothy Hatcher.

Source/JavaScriptCore:

* inspector/protocol/DOM.json:
Modify the existing `frameId` to represent the owner frame of the node, rather than the
frame it holds (in the case of an `<iframe>`).

Source/WebCore:

Modify the existing `frameId` to represent the owner frame of the node, rather than the
frame it holds (in the case of an `<iframe>`).

* inspector/agents/InspectorDOMAgent.cpp:
(WebCore::InspectorDOMAgent::buildObjectForNode):

Source/WebInspectorUI:

* UserInterface/Models/DOMNode.js:
(WI.DOMNode):
(WI.DOMNode.prototype.get frame): Added.
(WI.DOMNode.prototype.get frameIdentifier): Deleted.
Modify the existing `frameId` to represent the owner frame of the node, rather than the
frame it holds (in the case of an `<iframe>`).

* UserInterface/Controllers/DOMDebuggerManager.js:
(WI.DOMDebuggerManager.prototype.domBreakpointsForNode):
(WI.DOMDebuggerManager.prototype._detachDOMBreakpoint):
(WI.DOMDebuggerManager.prototype._resolveDOMBreakpoint):
(WI.DOMDebuggerManager.prototype._nodeInserted):
(WI.DOMDebuggerManager.prototype._nodeRemoved):
* UserInterface/Views/DOMTreeElement.js:
(WI.DOMTreeElement.prototype._populateTagContextMenu):
(WI.DOMTreeElement.prototype._buildAttributeDOM):
* UserInterface/Views/QuickConsole.js:
(WI.QuickConsole.prototype._selectExecutionContext):
* UserInterface/Views/SpreadsheetCSSStyleDeclarationSection.js:
(WI.SpreadsheetCSSStyleDeclarationSection.prototype._highlightNodesWithSelector):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244566 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoFix iOS build.
achristensen@apple.com [Tue, 23 Apr 2019 20:52:31 +0000 (20:52 +0000)]
Fix iOS build.
https://bugs.webkit.org/show_bug.cgi?id=195537

* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::setOpenPanelFilesMediaIcon):
AdoptWK constructor is private now.  Use adoptWK instead.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244565 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoBuild fix after r244545.
commit-queue@webkit.org [Tue, 23 Apr 2019 20:47:43 +0000 (20:47 +0000)]
Build fix after r244545.

Patch by Remy Demarest <rdemarest@apple.com> on 2019-04-23
Reviewed by Tim Horton.

* Platform/spi/mac/AppKitSPI.h:
The SPI was an IPI in High Sierra.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244564 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoContentSecurityPolicy::logToConsole should include line/column number and source...
drousso@apple.com [Tue, 23 Apr 2019 20:40:58 +0000 (20:40 +0000)]
ContentSecurityPolicy::logToConsole should include line/column number and source location
https://bugs.webkit.org/show_bug.cgi?id=114317
<rdar://problem/13614617>

Reviewed by Timothy Hatcher.

Source/WebCore:

No change in functionality.

* page/csp/ContentSecurityPolicy.h:
* page/csp/ContentSecurityPolicy.cpp:
(WebCore::ContentSecurityPolicy::reportViolation const):
(WebCore::ContentSecurityPolicy::logToConsole const):

LayoutTests:

* inspector/debugger/csp-exceptions.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244563 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Inspector: Canvas: support recording TypedOMCSSImageValue
drousso@apple.com [Tue, 23 Apr 2019 20:32:39 +0000 (20:32 +0000)]
Web Inspector: Canvas: support recording TypedOMCSSImageValue
https://bugs.webkit.org/show_bug.cgi?id=192609

Reviewed by Timothy Hatcher.

* inspector/InspectorCanvas.h:
* inspector/InspectorCanvas.cpp:
(WebCore::InspectorCanvas::indexForData):
(WebCore::InspectorCanvas::buildAction):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244562 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAccessibility text search and selection API enhancements.
commit-queue@webkit.org [Tue, 23 Apr 2019 20:31:51 +0000 (20:31 +0000)]
Accessibility text search and selection API enhancements.
https://bugs.webkit.org/show_bug.cgi?id=197095
<rdar://problem/48181791>

Patch by Andres Gonzalez <andresg_22@apple.com> on 2019-04-23
Reviewed by Chris Fleizach.

Source/WebCore:

- Split the existing SelectTextWithCriteria API into two: search text API (SearchTextWithCriteria) and a text operation API (TextOperation: select, replace, capitalize...).
- This allows for more flexibility and extensibility.
- Added the ability to retrieve text markers for multiple search hits.
- Various code clean up and consolidation.
- Added LayoutTest for search API.
- Previous API is marked with "To be deprecated", and is implemented with new implementation. May be removed in a future change.

Test: accessibility/mac/search-text/search-text.html

* accessibility/AccessibilityObject.cpp:
(WebCore::rangeClosestToRange):
(WebCore::AccessibilityObject::rangeOfStringClosestToRangeInDirection const):
(WebCore::AccessibilityObject::findTextRange const):
(WebCore::AccessibilityObject::findTextRanges const):
(WebCore::AccessibilityObject::performTextOperation):
(WebCore::AccessibilityObject::frame const):
(WebCore::AccessibilityObject::selectText): Deleted.
* accessibility/AccessibilityObject.h:
(WebCore::AccessibilitySearchTextCriteria::AccessibilitySearchTextCriteria):
(WebCore::AccessibilityTextOperation::AccessibilityTextOperation):
(WebCore::AccessibilitySelectTextCriteria::AccessibilitySelectTextCriteria): Deleted.
* accessibility/mac/WebAccessibilityObjectWrapperMac.mm:
(accessibilityTextCriteriaForParameterizedAttribute):
(accessibilitySearchTextCriteriaForParameterizedAttribute):
(accessibilityTextOperationForParameterizedAttribute):
(-[WebAccessibilityObjectWrapper IGNORE_WARNINGS_END]):
(-[WebAccessibilityObjectWrapper accessibilityAttributeValue:forParameter:]):
(accessibilitySelectTextCriteriaForCriteriaParameterizedAttribute): Deleted.

Tools:

Added new API JS binding code for searchTextWithCriteria to both WTR and DRT.

* DumpRenderTree/AccessibilityTextMarker.h:
* DumpRenderTree/AccessibilityUIElement.cpp:
(searchTextWithCriteriaCallback):
(AccessibilityUIElement::getJSClass):
* DumpRenderTree/AccessibilityUIElement.h:
* DumpRenderTree/mac/AccessibilityUIElementMac.mm:
(convertVectorToObjectArray):
(convertNSArrayToVector):
(searchTextParameterizedAttributeForCriteria):
(AccessibilityUIElement::getLinkedUIElements):
(AccessibilityUIElement::getDocumentLinks):
(AccessibilityUIElement::getChildren):
(AccessibilityUIElement::getChildrenWithRange):
(AccessibilityUIElement::rowHeaders const):
(AccessibilityUIElement::columnHeaders const):
(AccessibilityUIElement::uiElementArrayAttributeValue const):
(AccessibilityUIElement::searchTextWithCriteria):
(AccessibilityUIElement::attributesOfColumnHeaders):
(AccessibilityUIElement::attributesOfRowHeaders):
(AccessibilityUIElement::attributesOfColumns):
(AccessibilityUIElement::attributesOfRows):
(AccessibilityUIElement::attributesOfVisibleCells):
* WebKitTestRunner/InjectedBundle/AccessibilityUIElement.h:
* WebKitTestRunner/InjectedBundle/Bindings/AccessibilityUIElement.idl:
* WebKitTestRunner/InjectedBundle/mac/AccessibilityUIElementMac.mm:
(WTR::convertVectorToObjectArray):
(WTR::convertNSArrayToVector):
(WTR::searchTextParameterizedAttributeForCriteria):
(WTR::AccessibilityUIElement::getLinkedUIElements):
(WTR::AccessibilityUIElement::getDocumentLinks):
(WTR::AccessibilityUIElement::getUIElementsWithAttribute const):
(WTR::AccessibilityUIElement::getChildren):
(WTR::AccessibilityUIElement::getChildrenWithRange):
(WTR::AccessibilityUIElement::rowHeaders const):
(WTR::AccessibilityUIElement::columnHeaders const):
(WTR::AccessibilityUIElement::uiElementArrayAttributeValue const):
(WTR::AccessibilityUIElement::searchTextWithCriteria):
(WTR::AccessibilityUIElement::attributesOfColumnHeaders):
(WTR::AccessibilityUIElement::attributesOfRowHeaders):
(WTR::AccessibilityUIElement::attributesOfColumns):
(WTR::AccessibilityUIElement::attributesOfRows):
(WTR::AccessibilityUIElement::attributesOfVisibleCells):
(WTR::convertElementsToObjectArray): Deleted.

LayoutTests:

- Added new test for AccessibilitySearchTextWithCriteria API.
- Updated bounds-for-range expected file that includes a list of available APIs.

* accessibility/mac/bounds-for-range-expected.txt:
* accessibility/mac/search-text/search-text-expected.txt: Added.
* accessibility/mac/search-text/search-text.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244561 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Inspector: Canvas: no obvious way of switching to the overview when viewing an...
drousso@apple.com [Tue, 23 Apr 2019 20:31:02 +0000 (20:31 +0000)]
Web Inspector: Canvas: no obvious way of switching to the overview when viewing an imported recording
https://bugs.webkit.org/show_bug.cgi?id=197178
<rdar://problem/50106641>

Reviewed by Timothy Hatcher.

Expand the idea of "imported" recordings to include "saved" recordings, ones whose
underlying <canvas> was destroyed.

* UserInterface/Views/CanvasTabContentView.js:
(WI.CanvasTabContentView):
(WI.CanvasTabContentView.prototype.attached):
(WI.CanvasTabContentView.prototype._removeCanvas):
(WI.CanvasTabContentView.prototype._addRecording):
(WI.CanvasTabContentView.prototype._handleRecordingSavedOrStopped): Added.
(WI.CanvasTabContentView.prototype.initialLayout): Deleted.
(WI.CanvasTabContentView.prototype._recordingImportedOrStopped): Deleted.
Make the "Saved Recordings" folder a child of the "Overview" so that the "Overview" path
component is always visible/clickable.

* UserInterface/Views/CanvasOverviewContentView.js:
(WI.CanvasOverviewContentView):
(WI.CanvasOverviewContentView.prototype.contentViewAdded):
(WI.CanvasOverviewContentView.prototype.attached):
(WI.CanvasOverviewContentView.prototype.detached):
(WI.CanvasOverviewContentView.prototype._addSavedRecording): Added.
(WI.CanvasOverviewContentView.prototype._handleRecordingSaved): Added.
(WI.CanvasOverviewContentView.prototype._handleSavedRecordingClicked): Added.
* UserInterface/Views/CanvasOverviewContentView.css:
(.content-view.canvas-overview .content-view.canvas.saved-recordings): Added.
(.content-view.canvas-overview .content-view.canvas.saved-recordings .tree-outline): Added.
(.content-view.canvas-overview .content-view.canvas.saved-recordings .tree-outline > .item.recording > .icon): Added.
Add a card for "Saved Recordings" that mimics the style of regular canvas cards.

* UserInterface/Controllers/CanvasManager.js:
(WI.CanvasManager):
(WI.CanvasManager.prototype.get savedRecordings): Added.
(WI.CanvasManager.prototype.async processJSON):
(WI.CanvasManager.prototype.disable):
(WI.CanvasManager.prototype._removeCanvas):
(WI.CanvasManager.prototype.get importedRecordings): Deleted.

* Localizations/en.lproj/localizedStrings.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244560 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[iOS] Input field on ddg.gg is auto focused when url is entered with the software...
pvollan@apple.com [Tue, 23 Apr 2019 20:27:24 +0000 (20:27 +0000)]
[iOS] Input field on ddg.gg is auto focused when url is entered with the software keyboard
https://bugs.webkit.org/show_bug.cgi?id=196740

Reviewed by Megan Gardner.

When an url for a page with an autofocused input field  is entered with the software keyboard,
the input field is auto selected, and the software keyboard reappears. This does not happen
when picking the url from favorites. After using the software keyboard to enter the url, the
activity state is being changed to focused. The method '_elementDidFocus' checks whether the
activity state changed, and allows the software keyboard to be shown in this case. To avoid
showing the software keyboard in this case, send the changing activity state bitfield to the
UI process, and check if the focus flag is the only flag set.

* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _elementDidFocus:userIsInteracting:blurPreviousNode:changingActivityState:userObject:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244559 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRemove DeprecatedOptional
commit-queue@webkit.org [Tue, 23 Apr 2019 20:25:59 +0000 (20:25 +0000)]
Remove DeprecatedOptional
https://bugs.webkit.org/show_bug.cgi?id=197161

Patch by Alex Christensen <achristensen@webkit.org> on 2019-04-23
Reviewed by Darin Adler.

Source/JavaScriptCore:

* inspector/InspectorBackendDispatcher.cpp:
* inspector/InspectorBackendDispatcher.h:

Source/WTF:

This was added in r209326 to be compatible with a shipping version of Safari.
We have released several versions of Safari since then, so do what the comments say and remove it.
The existence of this std::optional makes migrating to C++17 harder, and there's no reason to keep it.

* WTF.xcodeproj/project.pbxproj:
* wtf/CMakeLists.txt:
* wtf/DeprecatedOptional.h: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244558 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoMultiple File Input Icon Set Regardless of File List
commit-queue@webkit.org [Tue, 23 Apr 2019 20:18:05 +0000 (20:18 +0000)]
Multiple File Input Icon Set Regardless of File List
https://bugs.webkit.org/show_bug.cgi?id=195537

Source/WebCore:

Patch by Guy Lewin <guy@lewin.co.il> on 2019-04-23
Reviewed by Alexey Proskuryakov.

File input elements display icon with an empty file list after
resetting the file list in 'change' event handler - on iOS

Test: fast/forms/file/file-reset-in-change-using-open-panel-with-icon.html

* html/FileInputType.cpp:
(WebCore::FileInputType::filesChosen):

Source/WebKit:

Patch by Guy Lewin <guy@lewin.co.il> on 2019-04-23
Reviewed by Alexey Proskuryakov.

Add WKOpenPanelResultListenerChooseMediaFiles() to choose files with
icon and display string on iOS file open panels

* UIProcess/API/C/WKOpenPanelResultListener.cpp:
(WKOpenPanelResultListenerChooseMediaFiles):
* UIProcess/API/C/WKOpenPanelResultListener.h:

Tools:

Patch by Guy Lewin <guy@lewin.co.il> on 2019-04-23
Reviewed by Alexey Proskuryakov.

Add setOpenPanelFilesMediaIcon to TestRunner in order to test icon and
display string in iOS file inputs

* DumpRenderTree/TestRunner.cpp:
(SetOpenPanelFilesMediaIconCallback):
(TestRunner::staticFunctions):
(TestRunner::setOpenPanelFilesMediaIcon):
* DumpRenderTree/TestRunner.h:
(TestRunner::openPanelFilesMediaIcon const):
* DumpRenderTree/mac/UIDelegate.mm:
(-[UIDelegate webView:runOpenPanelForFileButtonWithResultListener:allowMultipleFiles:]):
* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::setOpenPanelFilesMediaIcon):
* WebKitTestRunner/InjectedBundle/TestRunner.h:
* WebKitTestRunner/TestController.cpp:
(WTR::runOpenPanel):
(WTR::TestController::resetStateToConsistentValues):
* WebKitTestRunner/TestController.h:
(WTR::TestController::openPanelFileURLsMediaIcon const):
(WTR::TestController::setOpenPanelFileURLsMediaIcon):
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveMessageFromInjectedBundle):

LayoutTests:

Add a layout test to verify file input icon resets on change event (iOS)
Change expectations according to
file-reset-in-change-using-open-panel.html

Patch by Guy Lewin <guy@lewin.co.il> on 2019-04-23
Reviewed by Alexey Proskuryakov.

* fast/forms/file/file-reset-in-change-using-open-panel-with-icon-expected.html: Added.
* fast/forms/file/file-reset-in-change-using-open-panel-with-icon.html: Added.
* platform/win/TestExpectations:
* platform/wincairo/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244557 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoCreate AVFoundationSoftLink.{h,mm} to reduce duplicate code
eric.carlson@apple.com [Tue, 23 Apr 2019 20:13:40 +0000 (20:13 +0000)]
Create AVFoundationSoftLink.{h,mm} to reduce duplicate code
https://bugs.webkit.org/show_bug.cgi?id=197171
<rdar://problem/47454979>

Reviewed by Youenn Fablet.

Source/WebCore:

Tests: TestWebKitAPI/Tests/WebCore/cocoa/AVFoundationSoftLinkTest.mm

* Modules/plugins/QuickTimePluginReplacement.mm:
(WebCore::jsValueWithValueInContext):
(WebCore::jsValueWithAVMetadataItemInContext):
* WebCore.xcodeproj/project.pbxproj:
* platform/audio/ios/AudioSessionIOS.mm:
(WebCore::AudioSession::setCategory):
(WebCore::AudioSession::category const):
(WebCore::AudioSession::routeSharingPolicy const):
(WebCore::AudioSession::routingContextUID const):
(WebCore::AudioSession::sampleRate const):
(WebCore::AudioSession::bufferSize const):
(WebCore::AudioSession::numberOfOutputChannels const):
(WebCore::AudioSession::tryToSetActiveInternal):
(WebCore::AudioSession::preferredBufferSize const):
(WebCore::AudioSession::setPreferredBufferSize):
* platform/audio/ios/MediaSessionManagerIOS.mm:
(-[WebMediaSessionHelper initWithCallback:]):
(-[WebMediaSessionHelper startMonitoringAirPlayRoutes]):
* platform/graphics/avfoundation/AVTrackPrivateAVFObjCImpl.mm:
(WebCore::AVTrackPrivateAVFObjCImpl::audioKind const):
(WebCore::AVTrackPrivateAVFObjCImpl::videoKind const):
(WebCore::AVTrackPrivateAVFObjCImpl::label const):
* platform/graphics/avfoundation/AudioSourceProviderAVFObjC.mm:
(WebCore::AudioSourceProviderAVFObjC::createMix):
* platform/graphics/avfoundation/MediaPlaybackTargetMac.mm:
* platform/graphics/avfoundation/MediaSelectionGroupAVFObjC.mm:
(WebCore::MediaSelectionGroupAVFObjC::updateOptions):
* platform/graphics/avfoundation/objc/AVFoundationMIMETypeCache.mm:
(WebCore::AVFoundationMIMETypeCache::canDecodeType):
(WebCore::AVFoundationMIMETypeCache::loadMIMETypes):
* platform/graphics/avfoundation/objc/CDMInstanceFairPlayStreamingAVFObjC.mm:
(WebCore::CDMInstanceFairPlayStreamingAVFObjC::supportsPersistableState):
(WebCore::CDMInstanceFairPlayStreamingAVFObjC::supportsPersistentKeys):
(WebCore::CDMInstanceFairPlayStreamingAVFObjC::supportsMediaCapability):
(WebCore::CDMInstanceFairPlayStreamingAVFObjC::initializeWithConfiguration):
(WebCore::CDMInstanceSessionFairPlayStreamingAVFObjC::updateLicense):
(WebCore::CDMInstanceSessionFairPlayStreamingAVFObjC::loadSession):
(WebCore::CDMInstanceSessionFairPlayStreamingAVFObjC::removeSessionData):
(WebCore::CDMInstanceSessionFairPlayStreamingAVFObjC::ensureSession):
* platform/graphics/avfoundation/objc/CDMSessionAVContentKeySession.mm:
(WebCore::CDMSessionAVContentKeySession::isAvailable):
(WebCore::CDMSessionAVContentKeySession::releaseKeys):
(WebCore::CDMSessionAVContentKeySession::update):
(WebCore::CDMSessionAVContentKeySession::generateKeyReleaseMessage):
(WebCore::CDMSessionAVContentKeySession::contentKeySession):
* platform/graphics/avfoundation/objc/CDMSessionAVFoundationObjC.mm:
* platform/graphics/avfoundation/objc/CDMSessionAVStreamSession.mm:
(WebCore::CDMSessionAVStreamSession::releaseKeys):
(WebCore::CDMSessionAVStreamSession::update):
(WebCore::CDMSessionAVStreamSession::setStreamSession):
(WebCore::CDMSessionAVStreamSession::generateKeyReleaseMessage):
* platform/graphics/avfoundation/objc/ImageDecoderAVFObjC.mm:
(WebCore::imageDecoderAssetOptions):
(WebCore::ImageDecoderAVFObjC::ImageDecoderAVFObjC):
(WebCore::ImageDecoderAVFObjC::firstEnabledTrack):
(WebCore::ImageDecoderAVFObjC::readSamples):
(SOFT_LINK_CONSTANT_MAY_FAIL): Deleted.
* platform/graphics/avfoundation/objc/InbandTextTrackPrivateAVFObjC.mm:
(WebCore::InbandTextTrackPrivateAVFObjC::label const):
* platform/graphics/avfoundation/objc/InbandTextTrackPrivateLegacyAVFObjC.mm:
(WebCore::InbandTextTrackPrivateLegacyAVFObjC::label const):
* platform/graphics/avfoundation/objc/MediaPlaybackTargetPickerMac.mm:
(WebCore::MediaPlaybackTargetPickerMac::devicePicker):
* platform/graphics/avfoundation/objc/MediaPlayerPrivateAVFoundationObjC.mm:
(WebCore::assetCacheForPath):
(WebCore::MediaPlayerPrivateAVFoundationObjC::clearMediaCache):
(WebCore::MediaPlayerPrivateAVFoundationObjC::clearMediaCacheForOrigins):
(WebCore::MediaPlayerPrivateAVFoundationObjC::cancelLoad):
(WebCore::MediaPlayerPrivateAVFoundationObjC::createImageGenerator):
(WebCore::MediaPlayerPrivateAVFoundationObjC::createAVPlayerLayer):
(WebCore::MediaPlayerPrivateAVFoundationObjC::synchronizeTextTrackState):
(WebCore::MediaPlayerPrivateAVFoundationObjC::createAVAssetForURL):
(WebCore::MediaPlayerPrivateAVFoundationObjC::setAVPlayerItem):
(WebCore::MediaPlayerPrivateAVFoundationObjC::createAVPlayer):
(WebCore::MediaPlayerPrivateAVFoundationObjC::createAVPlayerItem):
(WebCore::MediaPlayerPrivateAVFoundationObjC::supportsType):
(WebCore::MediaPlayerPrivateAVFoundationObjC::isAvailable):
(WebCore::MediaPlayerPrivateAVFoundationObjC::tracksChanged):
(WebCore::determineChangedTracksFromNewTracksAndOldItems):
(WebCore::MediaPlayerPrivateAVFoundationObjC::updateAudioTracks):
(WebCore::MediaPlayerPrivateAVFoundationObjC::updateVideoTracks):
(WebCore::MediaPlayerPrivateAVFoundationObjC::createVideoOutput):
(WebCore::MediaPlayerPrivateAVFoundationObjC::outputMediaDataWillChange):
(WebCore::MediaPlayerPrivateAVFoundationObjC::safeMediaSelectionGroupForLegibleMedia):
(WebCore::MediaPlayerPrivateAVFoundationObjC::safeMediaSelectionGroupForAudibleMedia):
(WebCore::MediaPlayerPrivateAVFoundationObjC::safeMediaSelectionGroupForVisualMedia):
(WebCore::MediaPlayerPrivateAVFoundationObjC::processMediaSelectionOptions):
(WebCore::MediaPlayerPrivateAVFoundationObjC::setCurrentTextTrack):
(WebCore::MediaPlayerPrivateAVFoundationObjC::languageOfPrimaryAudioTrack const):
(WebCore::MediaPlayerPrivateAVFoundationObjC::wirelessPlaybackTargetType const):
(WebCore::exernalDeviceDisplayNameForPlayer):
(WebCore::metadataType):
(WebCore::MediaPlayerPrivateAVFoundationObjC::metadataDidArrive):
(-[WebCoreAVFMovieObserver observeValueForKeyPath:ofObject:change:context:]):
(-[WebCoreAVFPullDelegate outputMediaDataWillChange:]):
(-[WebCoreAVFPullDelegate outputSequenceWasFlushed:]):
(WebCore::MediaPlayerPrivateAVFoundationObjC::processLegacyClosedCaptionsTracks): Deleted.
* platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaSourceAVFObjC.mm:
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::MediaPlayerPrivateMediaSourceAVFObjC):
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::isAvailable):
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::supportsType):
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::ensureLayer):
(WebCore::MediaPlayerPrivateMediaSourceAVFObjC::streamSession):
* platform/graphics/avfoundation/objc/MediaPlayerPrivateMediaStreamAVFObjC.mm:
(-[WebAVSampleBufferStatusChangeListener observeValueForKeyPath:ofObject:change:context:]):
(WebCore::MediaPlayerPrivateMediaStreamAVFObjC::isAvailable):
(WebCore::MediaPlayerPrivateMediaStreamAVFObjC::ensureLayers):
* platform/graphics/avfoundation/objc/SourceBufferPrivateAVFObjC.mm:
(-[WebAVSampleBufferErrorListener observeValueForKeyPath:ofObject:change:context:]):
(WebCore::SourceBufferPrivateAVFObjC::SourceBufferPrivateAVFObjC):
(WebCore::SourceBufferPrivateAVFObjC::~SourceBufferPrivateAVFObjC):
(WebCore::SourceBufferPrivateAVFObjC::trackDidChangeEnabled):
(WebCore::SourceBufferPrivateAVFObjC::enqueueSample):
* platform/graphics/ca/cocoa/PlatformCALayerCocoa.mm:
(WebCore::PlatformCALayerCocoa::layerTypeForPlatformLayer):
(WebCore::PlatformCALayerCocoa::PlatformCALayerCocoa):
(WebCore::PlatformCALayerCocoa::clone const):
(WebCore::PlatformCALayerCocoa::avPlayerLayer const):
* platform/graphics/cocoa/HEVCUtilitiesCocoa.mm:
(WebCore::validateHEVCParameters):
* platform/ios/PlatformSpeechSynthesizerIOS.mm:
(getAVSpeechUtteranceDefaultSpeechRate):
(getAVSpeechUtteranceMaximumSpeechRate):
(-[WebSpeechSynthesisWrapper speakUtterance:]):
(WebCore::PlatformSpeechSynthesizer::initializeVoiceList):
(SOFT_LINK_CONSTANT): Deleted.
* platform/ios/VideoFullscreenInterfaceAVKit.mm:
(-[WebAVPlayerLayer init]):
(-[WebAVPlayerLayer layoutSublayers]):
(-[WebAVPlayerLayer setVideoGravity:]):
(-[WebAVPlayerLayer videoRect]):
(WebAVPlayerLayerView_startRoutingVideoToPictureInPicturePlayerLayerView):
* platform/mac/SerializedPlatformRepresentationMac.mm:
(WebCore::jsValueWithValueInContext):
(WebCore::jsValueWithAVMetadataItemInContext):
* platform/mediarecorder/cocoa/MediaRecorderPrivateWriterCocoa.mm:
(WebCore::getAVFormatIDKeyWithFallback):
(WebCore::getAVNumberOfChannelsKeyWithFallback):
(WebCore::getAVSampleRateKeyWithFallback):
(WebCore::getAVEncoderBitRateKeyWithFallback):
(WebCore::MediaRecorderPrivateWriter::create):
(WebCore::MediaRecorderPrivateWriter::setVideoInput):
(WebCore::MediaRecorderPrivateWriter::setAudioInput):
* platform/mediastream/RealtimeVideoSource.h:
* platform/mediastream/VideoPreset.h:
* platform/mediastream/ios/AVAudioSessionCaptureDeviceManager.mm:
(WebCore::AVAudioSessionCaptureDeviceManager::refreshAudioCaptureDevices):
* platform/mediastream/ios/CoreAudioCaptureSourceIOS.mm:
(-[WebCoreAudioCaptureSourceIOSListener initWithCallback:]):
(-[WebCoreAudioCaptureSourceIOSListener handleInterruption:]):
* platform/mediastream/mac/AVCaptureDeviceManager.mm:
(WebCore::deviceIsAvailable):
(WebCore::AVCaptureDeviceManager::updateCachedAVCaptureDevices):
(WebCore::AVCaptureDeviceManager::refreshCaptureDevices):
(WebCore::AVCaptureDeviceManager::isAvailable):
(WebCore::AVCaptureDeviceManager::~AVCaptureDeviceManager):
* platform/mediastream/mac/AVVideoCaptureSource.mm:
(WebCore::AVVideoPreset::create):
(WebCore::AVVideoPreset::AVVideoPreset):
(WebCore::AVVideoCaptureSource::create):
(WebCore::AVVideoCaptureSource::AVVideoCaptureSource):
(WebCore::AVVideoCaptureSource::capabilities):
(WebCore::sensorOrientationFromVideoOutput):
(WebCore::AVVideoCaptureSource::setupSession):
(WebCore::AVVideoCaptureSource::frameDurationForFrameRate):
(WebCore::AVVideoCaptureSource::setupCaptureSession):
(WebCore::AVVideoCaptureSource::captureOutputDidOutputSampleBufferFromConnection):
(WebCore::AVVideoCaptureSource::generatePresets):
(-[WebCoreAVVideoCaptureSourceObserver addNotificationObservers]):
(-[WebCoreAVVideoCaptureSourceObserver captureOutput:didOutputSampleBuffer:fromConnection:]):

Source/WebCore/PAL:

* PAL.xcodeproj/project.pbxproj:
* pal/cocoa/AVFoundationSoftLink.h: Added.
* pal/cocoa/AVFoundationSoftLink.mm: Added.

Source/WebKit:

* Shared/ios/WebIconUtilities.mm:
(WebKit::iconForVideoFile):
* Shared/mac/WebCoreArgumentCodersMac.mm:
(IPC::ArgumentCoder<WebCore::MediaPlaybackTargetContext>::encodePlatformData):
(IPC::ArgumentCoder<WebCore::MediaPlaybackTargetContext>::decodePlatformData):
* UIProcess/Cocoa/UIDelegate.mm:
(WebKit::UIDelegate::UIClient::decidePolicyForUserMediaPermissionRequest):
* WebProcess/WebPage/RemoteLayerTree/PlatformCALayerRemoteCustom.mm:
(WebKit::PlatformCALayerRemoteCustom::clone const):

Source/WebKitLegacy/mac:

* WebView/WebVideoFullscreenController.mm:
(-[WebVideoFullscreenController setVideoElement:]):
(-[WebVideoFullscreenController windowDidExitFullscreen]):
(SOFT_LINK_CLASS): Deleted.

Tools:

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebCore/cocoa/AVFoundationSoftLinkTest.mm: Added.
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244556 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUse additional members and protocols from WebKitAdditions in WKContentViewInteraction
dino@apple.com [Tue, 23 Apr 2019 19:35:30 +0000 (19:35 +0000)]
Use additional members and protocols from WebKitAdditions in WKContentViewInteraction
https://bugs.webkit.org/show_bug.cgi?id=197184
<rdar://problem/50113848>

Reviewed by Wenson Hsieh.

WebKitAdditions defines some macros to include additional members
and protocols for WKContentViewInteraction.

It also defines some new functions. Provide empty version
of those functions when WebKitAdditions is not available.

* UIProcess/ios/WKContentViewInteraction.h:
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _registerPreview]):
(-[WKContentView _unregisterPreview]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244555 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoSpeed up RenderTheme::systemColor on Speedometer2.
timothy@apple.com [Tue, 23 Apr 2019 19:13:48 +0000 (19:13 +0000)]
Speed up RenderTheme::systemColor on Speedometer2.
https://bugs.webkit.org/show_bug.cgi?id=197203
rdar://problem/50056756

Reviewed by Tim Horton.

* rendering/RenderThemeIOS.mm:
(WebCore::RenderThemeIOS::systemColor const): Remove some unused code. And fetch the cache after an early return.
* rendering/RenderThemeMac.mm:
(WebCore::RenderThemeMac::systemColor const): Avoid some allocations in LocalDefaultSystemAppearance
when a CSS color is in the system color cache.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244554 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAction sheet shares a stringified URL instead of a URL object
timothy_horton@apple.com [Tue, 23 Apr 2019 18:15:39 +0000 (18:15 +0000)]
Action sheet shares a stringified URL instead of a URL object
https://bugs.webkit.org/show_bug.cgi?id=197185
<rdar://problem/49962249>

Reviewed by Darin Adler.

* UIProcess/Cocoa/WKShareSheet.h:
* UIProcess/Cocoa/WKShareSheet.mm:
(-[WKShareSheet presentWithParameters:inRect:completionHandler:]):
(-[WKShareSheet presentWithParameters:completionHandler:]): Deleted.
* UIProcess/Cocoa/WebViewImpl.mm:
(WebKit::WebViewImpl::showShareSheet):
* UIProcess/ios/PageClientImplIOS.mm:
(WebKit::PageClientImpl::showShareSheet):
Make it possible to optionally provide WKShareSheet with
a rect to present relative to.

* UIProcess/ios/WKContentViewInteraction.h:
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _showShareSheet:inRect:completionHandler:]):
(-[WKContentView actionSheetAssistant:shareElementWithURL:rect:]):
(-[WKContentView _showShareSheet:completionHandler:]): Deleted.
Instead of stringifying the URL and using the text selection assistant's
share method, hand WKShareSheet a proper URL.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244553 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, speculative fix for failing inspector/canvas/ tests.
drousso@apple.com [Tue, 23 Apr 2019 18:06:29 +0000 (18:06 +0000)]
Unreviewed, speculative fix for failing inspector/canvas/ tests.
<rdar://problem/48248697>

Reviewed by NOBODY (OOPS!).

* inspector/canvas/resources/recording-utilities.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244552 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Add unit tests for UnApplyPatchIfRequired
aakash_jain@apple.com [Tue, 23 Apr 2019 18:02:55 +0000 (18:02 +0000)]
[ews-build] Add unit tests for UnApplyPatchIfRequired
https://bugs.webkit.org/show_bug.cgi?id=197197

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/steps_unittest.py:
(TestUnApplyPatchIfRequired):
(TestUnApplyPatchIfRequired.test_success):
(TestUnApplyPatchIfRequired.test_failure):
(TestUnApplyPatchIfRequired.test_skip):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244551 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Add unit tests for CleanWorkingDirectory
aakash_jain@apple.com [Tue, 23 Apr 2019 17:58:21 +0000 (17:58 +0000)]
[ews-build] Add unit tests for CleanWorkingDirectory
https://bugs.webkit.org/show_bug.cgi?id=197195

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/steps_unittest.py:
(TestCleanWorkingDirectory):
(TestCleanWorkingDirectory.test_success):
(TestCleanWorkingDirectory.test_failure):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244550 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Use explicit imports instead of wildcard imports
aakash_jain@apple.com [Tue, 23 Apr 2019 17:54:35 +0000 (17:54 +0000)]
[ews-build] Use explicit imports instead of wildcard imports
https://bugs.webkit.org/show_bug.cgi?id=197194

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/factories.py:
* BuildSlaveSupport/ews-build/loadConfig.py:
* BuildSlaveSupport/ews-build/steps_unittest.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244549 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Add unit tests for DownloadBuiltProduct
aakash_jain@apple.com [Tue, 23 Apr 2019 17:53:26 +0000 (17:53 +0000)]
[ews-build] Add unit tests for DownloadBuiltProduct
https://bugs.webkit.org/show_bug.cgi?id=197193

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/steps_unittest.py:
(TestDownloadBuiltProduct):
(TestDownloadBuiltProduct.setUp):
(TestDownloadBuiltProduct.tearDown):
(TestDownloadBuiltProduct.test_success):
(TestDownloadBuiltProduct.test_failure):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244548 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-app] Exception while loading status-bubble when no build step has started
aakash_jain@apple.com [Tue, 23 Apr 2019 17:51:55 +0000 (17:51 +0000)]
[ews-app] Exception while loading status-bubble when no build step has started
https://bugs.webkit.org/show_bug.cgi?id=196676

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-app/ews/views/statusbubble.py:
(StatusBubble.get_build_timestamp): Check if build step is present before accessing its timestamp.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244547 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[iOS] element.focus() sometimes fails to reveal the focused element when it becomes...
rniwa@webkit.org [Tue, 23 Apr 2019 17:10:10 +0000 (17:10 +0000)]
[iOS] element.focus() sometimes fails to reveal the focused element when it becomes editable dynamically
https://bugs.webkit.org/show_bug.cgi?id=197188

Reviewed by Wenson Hsieh.

Source/WebCore:

The bug was caused by the scroll-to-reveal code triggered by Element::updateFocusAppearance updating
the scroll position via scrolling tree update in a layer tree commit which happens after
_zoomToRevealFocusedElement in WKContentView had already scrolled the frame view.

To fix this problem, we need to defer the editor state update until the layer commit (see r244494),
and update the scrolling tree before invoking WebPageProxy::editorStateChanged which brings up
the keyboard and scroll-to-reveal the caret in the UI process side.

We also avoid revealing the focus for the second time via Document::scheduleScrollToFocusedElement
in Element::updateFocusAppearance as this timer based scrolling also happens after we had already
revealed the caret in _zoomToRevealFocusedElement. This is a bit hacky but works for most cases since
we wouldn't bring up a keyboard if the focused element was not editable anyway.

Test: editing/selection/ios/scrolling-to-focused-element-inside-iframe.html

* dom/Element.cpp:
(WebCore::Element::updateFocusAppearance): Avoid scheduling a timer based reavel of the focused element
when we're already revealing the element via selection change.

Source/WebKit:

Commit the scroll tree update before revealing the keyboard via editor state update.

* UIProcess/RemoteLayerTree/RemoteLayerTreeDrawingAreaProxy.mm:
(WebKit::RemoteLayerTreeDrawingAreaProxy::commitLayerTree):

LayoutTests:

Added a regression test.

* editing/selection/ios/scrolling-to-focused-element-inside-iframe-expected.txt: Added.
* editing/selection/ios/scrolling-to-focused-element-inside-iframe.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244546 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoFix layout issues occuring when entering full screen mode.
commit-queue@webkit.org [Tue, 23 Apr 2019 16:25:30 +0000 (16:25 +0000)]
Fix layout issues occuring when entering full screen mode.
https://bugs.webkit.org/show_bug.cgi?id=197086
<rdar://problem/47733671>.

Patch by Remy Demarest <rdemarest@apple.com> on 2019-04-23
Reviewed by Darin Adler.

This issue is the result of changing the style mask of the window after entering
full screen mode. Safari adds an invisible toolbar to display the URL of the page
which ends up breaking the layout. Having that window use a style that includes a
titlebar fixes the bug.

Source/WebCore:

* platform/mac/WebCoreFullScreenWindow.mm:
(-[WebCoreFullScreenWindow constrainFrameRect:toScreen:]): Ensure that the window
can fill the entire screen including the underlapping the menu bar, so that the
window does not resize when the animation is done.
(-[WebCoreFullScreenWindow canBecomeMainWindow]): Borderless windows cannot become
main by default, adding the titlebar allows it to become main, prevent this from
happening at all to preserve the existing behavior.

Source/WebKit:

* Platform/spi/mac/AppKitSPI.h:
Declare an SPI to be used in WKFullScreenWindowController.

* UIProcess/Cocoa/WebViewImpl.mm:
(WebKit::WebViewImpl::fullScreenWindow): Make the full screen window show a titlebar
and make the content view underlap the titlebar to match the current behavior.
Remove NSWindowStyleMaskBorderless which has no effects since it is equal to zero.

* UIProcess/mac/WKFullScreenWindowController.mm:
(-[WKFullScreenWindowController initWithWindow:webView:page:]): Hide the titlebar
before beginning the animation to full screen.
(-[WKFullScreenWindowController finishedEnterFullScreenAnimation:]): Show the title
bar when in full screen instead of a blank bar.
(-[WKFullScreenWindowController finishedExitFullScreenAnimation:]): Hide the title
bar to restore the initial setting.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244545 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAd Click Attribution redirects to well-known location should not trigger a conversion...
wilander@apple.com [Tue, 23 Apr 2019 16:14:05 +0000 (16:14 +0000)]
Ad Click Attribution redirects to well-known location should not trigger a conversion if they are blocked by content blockers
https://bugs.webkit.org/show_bug.cgi?id=197183
<rdar://problem/47763188>

Reviewed by Alex Christensen.

Source/WebKit:

Ad Click Attribution conversions are picked up in the redirect handler
in WebKit::NetworkResourceLoader. Content blocking typically happens in
the continued redirect request handling in the web content process and
a blocked request comes back empty.

We need to call the WebKit::NetworkLoadChecker in the network process
for these specific redirects, just like we do for Ping.

The change makes use of the existing function
NetworkLoadChecker::enableContentExtensionsCheck() for this purpose.

In essence, this change makes it possible to block all conversions made
to a "/.well-known/ad-click-attribution/" URL.

* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::handleAdClickAttributionConversion):
    New convenience function.
(WebKit::NetworkResourceLoader::willSendRedirectedRequest):
    Now calls NetworkLoadChecker::enableContentExtensionsCheck() if
    an Ad Click Attribution conversion was found in the redirect URL.
(WebKit::NetworkResourceLoader::continueWillSendRedirectedRequest):
    If the request was not blocked, it will store any found conversion here.
* NetworkProcess/NetworkResourceLoader.h:

LayoutTests:

* http/tests/contentextensions/block-ad-click-attribution-expected.txt: Added.
* http/tests/contentextensions/block-ad-click-attribution.html: Added.
* http/tests/contentextensions/block-ad-click-attribution.html.json: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244544 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[PlayStation] Support running of JSC tests on remote playstation devices
stephan.szabo@sony.com [Tue, 23 Apr 2019 16:10:53 +0000 (16:10 +0000)]
[PlayStation] Support running of JSC tests on remote playstation devices
https://bugs.webkit.org/show_bug.cgi?id=197170

Reviewed by Ross Kirsling.

Script changes

* Scripts/run-javascriptcore-tests:
(runJSCStressTests): Pass arguments for playstation to
run-jsc-stress-tests
* Scripts/run-jsc-stress-tests:
Handle $hostOS=="playstation" for determineArchitecture, turning
off $isFTLPlatform and runProfiler and defaulting test writer.
* Scripts/webkitdirs.pm:
(jscPath): PlayStation WebKit executables have .elf extension
(executableProductDir): PlayStation builds on Windows hosts but
uses bin directory rather than bin64/bin32 for executable output
location.
* Scripts/webkitruby/jsc-stress-test-writer-playstation.rb: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244543 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agofast/selectors/matches-backtracking.html is a flaky timeout
sroberts@apple.com [Tue, 23 Apr 2019 16:06:46 +0000 (16:06 +0000)]
fast/selectors/matches-backtracking.html is a flaky timeout
https://bugs.webkit.org/show_bug.cgi?id=197068

Unreviewed test gardening.

* platform/mac-wk2/TestExpectations: Updating expectations for flaky test

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244542 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[CMake][Win] Use target oriented design for WebKit
don.olmstead@sony.com [Tue, 23 Apr 2019 15:55:00 +0000 (15:55 +0000)]
[CMake][Win] Use target oriented design for WebKit
https://bugs.webkit.org/show_bug.cgi?id=197173

Reviewed by Alex Christensen.

.:

Override WebKit_DERIVED_SOUCES_DIR within WinCairo in support of a target
oriented design.

* Source/cmake/OptionsWinCairo.cmake:

Source/WebKit:

Enumerate public framework headers for WinCairo's WebKit implementation and copy them.
Migrate to use WebKit_DERIVED_SOURCES_DIR.

* CMakeLists.txt:
* PlatformGTK.cmake:
* PlatformWPE.cmake:
* PlatformWin.cmake:

Tools:

* TestWebKitAPI/PlatformWin.cmake: Add WebKitFrameworkHeaders as a dependency.
* TestWebKitAPI/Tests/WebKit/curl/Certificates.cpp: Remove unused header.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244541 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[Process-Swap-On-Navigation] WebKit hangs when going back to a form submission's...
cdumez@apple.com [Tue, 23 Apr 2019 15:38:11 +0000 (15:38 +0000)]
[Process-Swap-On-Navigation] WebKit hangs when going back to a form submission's page due to Process-Swap-On-Navigation on iOS 12.2 and higher
https://bugs.webkit.org/show_bug.cgi?id=197097
<rdar://problem/50048318>

Reviewed by Alex Christensen.

Source/WebCore:

* loader/EmptyFrameLoaderClient.h:
* loader/FrameLoader.cpp:
(WebCore::FrameLoader::checkLoadCompleteForThisFrame):
If we still have a provisional item (new load has been started or we did not clear it because we're about to retry),
then let the client know the load will continue by passing an extra flag with the didFailProvisionalLoad.

* loader/FrameLoaderClient.h:
* loader/FrameLoaderTypes.h:

Source/WebKit:

The issue is that when doing a history navigation with form data, we try to resubmit the form.
We initially use the ReturnCacheDataDontLoad cache policy without prompting the user to get
cached data. If this fails, we sent a DidFailProvisionalLoadForFrame IPC to the UIProcess but
the WebContent process actually retries the load with ReturnCacheDataElseLoad cache policy.
This new load triggers a new decidePolicyForNavigationAction.

This would cause trouble with process-swapping because we normally destroy the
ProvisionalPageProxy as soon as we receive a DidFailProvisionalLoadForFrame IPC from the
provisional process. As a result, the provisional process would not be able to retry.

To address the issue, we add pass a flag with the DidFailProvisionalLoadForFrame IPC to
if the load will continue or not. When this flag is set, the UIProcess does not destroy
the provisional page.

* Scripts/webkit/messages.py:
Minor IPC message handler build fix.

* UIProcess/API/Cocoa/WKURLSchemeTask.mm:
(-[WKURLSchemeTaskImpl _onlyIfCached]):
* UIProcess/API/Cocoa/WKURLSchemeTaskPrivate.h:
Add new _onlyIfCached SPI to WKURLSchemeTaskPrivate to allow for API testing.

* UIProcess/ProvisionalPageProxy.cpp:
(WebKit::ProvisionalPageProxy::cancel):
Pass WillContinueLoading flag.

(WebKit::ProvisionalPageProxy::validateInput):
Consider IPC with a navigationID of 0 as valid. When the navigation is triggered by the
WebContent process, it sends us a DecidePolicyForNavigationActionAsync IPC with a navigationID
of 0, until the UIProcess can generate a valid identifier.

(WebKit::ProvisionalPageProxy::didFailProvisionalLoadForFrame):

(WebKit::ProvisionalPageProxy::decidePolicyForNavigationActionAsync):
When the WebPageProxy has generated a navigationID for the new WebContent process-initiated
navigation, update m_navigationID so that follow-up IPC is considered valid.

* UIProcess/ProvisionalPageProxy.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::didFailProvisionalLoadForFrame):
(WebKit::WebPageProxy::didFailProvisionalLoadForFrameShared):
(WebKit::WebPageProxy::decidePolicyForNavigationActionAsyncShared):
(WebKit::WebPageProxy::decidePolicyForNavigationAction):
* UIProcess/WebPageProxy.h:
* UIProcess/WebPageProxy.messages.in:
* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::dispatchDidFailProvisionalLoad):
* WebProcess/WebCoreSupport/WebFrameLoaderClient.h:

Source/WebKitLegacy/mac:

* WebCoreSupport/WebFrameLoaderClient.h:
* WebCoreSupport/WebFrameLoaderClient.mm:
(WebFrameLoaderClient::dispatchDidFailProvisionalLoad):

Source/WebKitLegacy/win:

* WebCoreSupport/WebFrameLoaderClient.cpp:
(WebFrameLoaderClient::dispatchDidFailProvisionalLoad):
* WebCoreSupport/WebFrameLoaderClient.h:

Tools:

Add API test coverage.

* TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:
(-[PSONScheme webView:startURLSchemeTask:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244540 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoREGRESSION (r240727) [ Mac iOS ] Layout Test http/tests/workers/service/basic-registe...
youenn@apple.com [Tue, 23 Apr 2019 15:36:58 +0000 (15:36 +0000)]
REGRESSION (r240727) [ Mac iOS ] Layout Test http/tests/workers/service/basic-register-exceptions.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=194368
<rdar://problem/47937467>

Reviewed by Alex Christensen.

Test was exiting sooner than the last registration was rejected.
Make sure that the test is run until the end in a more ordered way.

* http/tests/workers/service/basic-register-exceptions-expected.txt:
* http/tests/workers/service/resources/basic-register-exceptions.js:
(async.doTest):
(then): Deleted.
(catch): Deleted.
* platform/mac-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244539 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ATK] Implement AtkComponentIface scroll_to methods
commit-queue@webkit.org [Tue, 23 Apr 2019 10:47:45 +0000 (10:47 +0000)]
[ATK] Implement AtkComponentIface scroll_to methods
https://bugs.webkit.org/show_bug.cgi?id=196856

Patch by Carlos Garcia Campos <cgarcia@igalia.com> on 2019-04-23
Reviewed by Michael Catanzaro.

Source/WebCore:

Implement scroll_to and scroll_to_point when ATK >= 2.30.

Fixes: accessibility/scroll-to-global-point-iframe-nested.html
       accessibility/scroll-to-global-point-iframe.html
       accessibility/scroll-to-global-point-main-window.html
       accessibility/scroll-to-global-point-nested.html
       accessibility/scroll-to-make-visible-div-overflow.html
       accessibility/scroll-to-make-visible-iframe.html
       accessibility/scroll-to-make-visible-nested-2.html
       accessibility/scroll-to-make-visible-nested.html

* accessibility/AccessibilityObject.cpp:
(WebCore::AccessibilityObject::scrollToMakeVisible const): Add new method receiving the
ScrollRectToVisibleOptions since ATK interface has a parameter to decide how to scroll.
* accessibility/AccessibilityObject.h:
* accessibility/atk/WebKitAccessibleInterfaceComponent.cpp:
(atkToContents):
(webkitAccessibleComponentRefAccessibleAtPoint):
(webkitAccessibleComponentGetExtents):
(webkitAccessibleComponentGrabFocus):
(webkitAccessibleComponentScrollTo):
(webkitAccessibleComponentScrollToPoint):
(webkitAccessibleComponentInterfaceInit):
(core): Deleted.

Tools:

* WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:
(WTR::AccessibilityUIElement::scrollToMakeVisible): Call atk_component_scroll_to().
(WTR::AccessibilityUIElement::scrollToGlobalPoint): Call atk_component_scroll_to_point().
* gtk/jhbuild.modules: Bump atk, at-spi2-core and at-spi2-atk to their latest stable version.

LayoutTests:

Remove expectations for tests that are passing now.

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244538 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ContentChangeObserver] Some dropdown menus may close without user gesture on america...
zalan@apple.com [Tue, 23 Apr 2019 01:36:39 +0000 (01:36 +0000)]
[ContentChangeObserver] Some dropdown menus may close without user gesture on americanexpress.com
https://bugs.webkit.org/show_bug.cgi?id=197175
<rdar://problem/49613013>

Reviewed by Simon Fraser.

Source/WebKit:

Do not generate additional synthetic mouse events (e.g. mouseout in this case) when the content handles the click event. This helps cases when the synthetic mouseout ended up dismissing the dropdown menus.
However it won't regress cases like youtube.com, where sending mouseout is required to have the control bar dismissed on play.

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

LayoutTests:

* fast/events/touch/ios/content-observation/mouse-out-event-should-not-fire-on-click-when-default-prevented-expected.txt: Added.
* fast/events/touch/ios/content-observation/mouse-out-event-should-not-fire-on-click-when-default-prevented.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244537 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Improve summary for UploadBuiltProduct step
aakash_jain@apple.com [Tue, 23 Apr 2019 00:22:06 +0000 (00:22 +0000)]
[ews-build] Improve summary for UploadBuiltProduct step
https://bugs.webkit.org/show_bug.cgi?id=197182

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/steps.py:
(UploadBuiltProduct): Added descriptionDone.
(UploadBuiltProduct.getResultSummary): Overrided getResultSummary.
* BuildSlaveSupport/ews-build/steps_unittest.py:
(TestUploadBuiltProduct.test_success): Updated unit-test.
(TestUploadBuiltProduct.test_failure): Added unit-test for failure case.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244533 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Trigger queues only after uploading the archive
aakash_jain@apple.com [Tue, 23 Apr 2019 00:16:48 +0000 (00:16 +0000)]
[ews-build] Trigger queues only after uploading the archive
https://bugs.webkit.org/show_bug.cgi?id=197180

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/factories.py:
(Factory.__init__):
(BuildFactory.__init__): Remove trigger step from factories.
(BuildFactory.propertiesToPassToTriggers): Deleted.
* BuildSlaveSupport/ews-build/steps.py:
(ConfigureBuild.__init__): Confiure triggers property.
(Trigger): Step to trigger other queues.
(Trigger.propertiesToPassToTriggers): Method to construct properties to pass to triggerred builds.
(UploadBuiltProduct.finished): Trigger builds after upload is successful.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244532 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoDeprecate WKContextCreateWithInjectedBundlePath
commit-queue@webkit.org [Tue, 23 Apr 2019 00:15:03 +0000 (00:15 +0000)]
Deprecate WKContextCreateWithInjectedBundlePath
https://bugs.webkit.org/show_bug.cgi?id=197169

Patch by Alex Christensen <achristensen@webkit.org> on 2019-04-22
Reviewed by Youenn Fablet.

Source/WebKit:

* UIProcess/API/C/WKContext.h:

Tools:

* TestWebKitAPI/PlatformUtilities.cpp:
(TestWebKitAPI::Util::createContextWithInjectedBundle):
Use WKContextCreateWithConfiguration instead of WKContextCreateWithInjectedBundlePath after having set its injected bundle path.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244529 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoREGRESSION(r230681) Do not use stored credentials if WKBundlePageResourceLoadClient...
achristensen@apple.com [Mon, 22 Apr 2019 23:05:48 +0000 (23:05 +0000)]
REGRESSION(r230681) Do not use stored credentials if WKBundlePageResourceLoadClient.shouldUseCredentialStorage returns false
https://bugs.webkit.org/show_bug.cgi?id=197093
<rdar://problem/49708268>

Reviewed by Chris Dumez.

Source/WebKit:

Only get the StoredCredentialsPolicy from the NetworkLoadChecker if we haven't already been told not to use credentials.
Also add some test infrastructure for clearing persistent credentials added by the test.

* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::removeCredential):
* NetworkProcess/NetworkProcess.h:
* NetworkProcess/NetworkProcess.messages.in:
* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::startNetworkLoad):
* NetworkProcess/cocoa/NetworkProcessCocoa.mm:
(WebKit::NetworkProcess::removeCredential):
* UIProcess/API/Cocoa/WKProcessPool.mm:
(-[WKProcessPool _removeCredential:forProtectionSpace:completionHandler:]):
* UIProcess/API/Cocoa/WKProcessPoolPrivate.h:
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::removeCredential):
* UIProcess/WebProcessPool.h:

Tools:

Add a test that does two loads.  The first load shouldUseCredentialStorage returns true and we provide a persistent credential.
The second load shouldUseCredentialStorage returns false and we verify that a challenge is received with no suggested credential.
We also need to make the TCPServer able to handle more than one connection because we need these two loads to come from the same protection space,
and our current Cocoa implementation of NetworkSession uses two NSURLSessions that don't share a connection cache, one for loads with credentials
and one for loads without credentials, so there are two TCP connections to the same server in this test.

* TestWebKitAPI/TCPServer.cpp:
(TestWebKitAPI::TCPServer::TCPServer):
(TestWebKitAPI::TCPServer::~TCPServer):
(TestWebKitAPI::TCPServer::socketBindListen):
(TestWebKitAPI::TCPServer::waitForAndReplyToRequests): Deleted.
* TestWebKitAPI/TCPServer.h:
* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKitCocoa/BasicProposedCredentialPlugIn.mm: Added.
(-[BasicProposedCredentialPlugIn webProcessPlugIn:didCreateBrowserContextController:]):
* TestWebKitAPI/Tests/WebKitCocoa/Challenge.mm:
(respondWithChallengeThenOK):
(TestWebKitAPI::TEST):
(-[ProposedCredentialDelegate webView:didFinishNavigation:]):
(-[ProposedCredentialDelegate webView:didReceiveAuthenticationChallenge:completionHandler:]):
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244521 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUpdate libwebrtc logging when WebCore WebRTC logging is updated
youenn@apple.com [Mon, 22 Apr 2019 22:29:38 +0000 (22:29 +0000)]
Update libwebrtc logging when WebCore WebRTC logging is updated
https://bugs.webkit.org/show_bug.cgi?id=197166
<rdar://problem/50107696>

Unreviewed.

Build fix after https://trac.webkit.org/changeset/244511.

* page/Page.cpp:
(WebCore::Page::configureLoggingChannel):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244517 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoDelayed WebProcessLaunch may break the _relatedWebView SPI
cdumez@apple.com [Mon, 22 Apr 2019 22:17:22 +0000 (22:17 +0000)]
Delayed WebProcessLaunch may break the _relatedWebView SPI
https://bugs.webkit.org/show_bug.cgi?id=197160

Reviewed by Alex Christensen.

Source/WebKit:

Delayed WebProcessLaunch may break the _relatedWebView SPI. The breakage would happen if the client
would relate a WebView to another which has not launched its initial process yet.

To address the issue, when we need a running process for a WebView which has a related view, we need
to make sure the related view has a running process and use that process. Previously, we would share
the "dummy" process instead.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::launchProcess):
(WebKit::WebPageProxy::ensureRunningProcess):
* UIProcess/WebPageProxy.h:
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::createWebPage):

Tools:

Add API test coverage.

* TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244516 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoCache API should return Abort error in case of putting an aborted fetch
youenn@apple.com [Mon, 22 Apr 2019 22:12:03 +0000 (22:12 +0000)]
Cache API should return Abort error in case of putting an aborted fetch
https://bugs.webkit.org/show_bug.cgi?id=196757

Reviewed by Darin Adler.

LayoutTests/imported/w3c:

* web-platform-tests/service-workers/cache-storage/serviceworker/cache-abort.https-expected.txt:
* web-platform-tests/service-workers/cache-storage/window/cache-abort.https-expected.txt:
* web-platform-tests/service-workers/cache-storage/worker/cache-abort.https-expected.txt:

Source/WebCore:

In case of an aborted fetch, call consume callback with an AbortError.
Update the code that handles load cancelling as the loader callback is called.
Covered by rebased tests.

* Modules/fetch/FetchResponse.cpp:
(WebCore::FetchResponse::addAbortSteps):
(WebCore::FetchResponse::BodyLoader::didFail):
(WebCore::FetchResponse::BodyLoader::BodyLoader):
(WebCore::FetchResponse::BodyLoader::~BodyLoader):
(WebCore::FetchResponse::stop):
* Modules/fetch/FetchResponse.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244515 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Inspector: REGRESSION: Sources: auto-continue breakpoint triangle is incorrectly...
drousso@apple.com [Mon, 22 Apr 2019 22:11:14 +0000 (22:11 +0000)]
Web Inspector: REGRESSION: Sources: auto-continue breakpoint triangle is incorrectly positioned
https://bugs.webkit.org/show_bug.cgi?id=197179

Reviewed by Joseph Pecoraro.

* UserInterface/Views/BreakpointTreeElement.css:
(.item.breakpoint .status): Added.
(.item.breakpoint .status > .status-image.auto-continue::after):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244514 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUse curl to download packages for webkitpy autoinstaller
dean_johnson@apple.com [Mon, 22 Apr 2019 22:07:47 +0000 (22:07 +0000)]
Use curl to download packages for webkitpy autoinstaller
https://bugs.webkit.org/show_bug.cgi?id=197164

Reviewed by Darin Adler.

* Scripts/webkitpy/common/system/autoinstall.py:
(AutoInstaller._download): Use curl(1) to download packages necessary for
webkitpy/thirdparty/autoinstalled since Python2's urllib2 module can result in unnecessary
errors which are not present when using curl(1).

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244513 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRemove linked-on-or-after check for download attribute handling
david_quesada@apple.com [Mon, 22 Apr 2019 21:54:58 +0000 (21:54 +0000)]
Remove linked-on-or-after check for download attribute handling
https://bugs.webkit.org/show_bug.cgi?id=197176
rdar://problem/48459714

Reviewed by Alex Christensen.

There are no apps linked prior to this SDK version that are using the _WKDownload SPI,
so it doesn't make sense to preserve the old behavior of forcing download-attribute
links to start downloads.

* UIProcess/Cocoa/VersionChecks.h:
* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::receivedNavigationPolicyDecision):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244512 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUpdate libwebrtc logging when WebCore WebRTC logging is updated
youenn@apple.com [Mon, 22 Apr 2019 21:29:12 +0000 (21:29 +0000)]
Update libwebrtc logging when WebCore WebRTC logging is updated
https://bugs.webkit.org/show_bug.cgi?id=197166

Reviewed by Eric Carlson.

When updating WebRTC logging from Web Inspector, update libwebrtc logging so that no page reload is required.
Manually tested.

* page/Page.cpp:
(WebCore::Page::configureLoggingChannel):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244511 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244502.
cdumez@apple.com [Mon, 22 Apr 2019 21:08:07 +0000 (21:08 +0000)]
Unreviewed, rolling out r244502.

Caused crashes on the bots

Reverted changeset:

"Delayed WebProcessLaunch may break the _relatedWebView SPI"
https://bugs.webkit.org/show_bug.cgi?id=197160
https://trac.webkit.org/changeset/244502

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244510 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoIntroduce the concept of "opportunistic" stacking contexts
simon.fraser@apple.com [Mon, 22 Apr 2019 19:57:58 +0000 (19:57 +0000)]
Introduce the concept of "opportunistic" stacking contexts
https://bugs.webkit.org/show_bug.cgi?id=197077

Reviewed by Zalan Bujtas.

Bring back a variant of some code removed in r236424, which allows a RenderLayer
to be stacking context for painting, without actually being on in terms of CSS.

Internally, RenderLayer will call setIsOpportunisticStackingContext() to make a layer
into a stacking context for painting. External callers deal with isStackingContext()
or isCSSStackingContext().

Sadly we can't make m_forcedStackingContext (set on media element layers) trigger a
non-CSS stacking context; media controls use mix-blend-mode, and rely on the fake-stacking
media layer to be the "isolateBlending" ancestor.

No code uses this yet.

* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::RenderLayer):
(WebCore::RenderLayer::shouldBeCSSStackingContext const):
(WebCore::RenderLayer::isStackingContextChanged):
(WebCore::RenderLayer::setIsOpportunisticStackingContext):
(WebCore::RenderLayer::setIsCSSStackingContext):
(WebCore::RenderLayer::updateAncestorChainHasBlendingDescendants):
(WebCore::RenderLayer::dirtyAncestorChainHasBlendingDescendants):
(WebCore::RenderLayer::beginTransparencyLayers):
(WebCore::RenderLayer::calculateClipRects const):
(WebCore::outputPaintOrderTreeLegend):
(WebCore::outputPaintOrderTreeRecursive):
(WebCore::RenderLayer::shouldBeStackingContext const): Deleted.
(WebCore::RenderLayer::setIsStackingContext): Deleted.
* rendering/RenderLayer.h:
* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::compositingOpacity const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244509 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244437.
ryanhaddad@apple.com [Mon, 22 Apr 2019 19:51:38 +0000 (19:51 +0000)]
Unreviewed, rolling out r244437.

Still breaks internal tests.

Reverted changeset:

"InjectedBundle parameters often need initialization function
called before unarchiving"
https://bugs.webkit.org/show_bug.cgi?id=189709
https://trac.webkit.org/changeset/244437

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244508 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[WebGPU] Move swap chain methods from GPUDevice to GPUCanvasContext
justin_fan@apple.com [Mon, 22 Apr 2019 18:41:16 +0000 (18:41 +0000)]
[WebGPU] Move swap chain methods from GPUDevice to GPUCanvasContext
https://bugs.webkit.org/show_bug.cgi?id=197126

Reviewed by Dean Jackson.

Source/WebCore:

GPUSwapChains are now configured via GPUCanvasContext instead of GPUDevice. Covers WebGPU API
pull request #262.

Existing WebGPU tests updated to match.

* Modules/webgpu/GPUCanvasContext.cpp:
(WebCore::GPUCanvasContext::configureSwapChain):
(WebCore::GPUCanvasContext::replaceSwapChain): Deleted.
* Modules/webgpu/GPUCanvasContext.h:
* Modules/webgpu/GPUCanvasContext.idl:
* Modules/webgpu/WebGPUDevice.cpp:
(WebCore::WebGPUDevice::createSwapChain const): Deleted.
* Modules/webgpu/WebGPUDevice.h:
(WebCore::WebGPUDevice::device):
* Modules/webgpu/WebGPUDevice.idl:
* Modules/webgpu/WebGPUSwapChainDescriptor.cpp: Copied from Source/WebCore/Modules/webgpu/WebGPUSwapChainDescriptor.h.
(WebCore::WebGPUSwapChainDescriptor::asGPUSwapChainDescriptor const):
* Modules/webgpu/WebGPUSwapChainDescriptor.h:
* Sources.txt:
* WebCore.xcodeproj/project.pbxproj:
* platform/graphics/gpu/GPUDevice.cpp:
(WebCore::GPUDevice::setSwapChain):
(WebCore::GPUDevice::tryCreateSwapChain const): Deleted.
* platform/graphics/gpu/GPUDevice.h:
* platform/graphics/gpu/GPUSwapChain.h:
* platform/graphics/gpu/GPUSwapChainDescriptor.h:
(WebCore::GPUSwapChainDescriptor::GPUSwapChainDescriptor):
* platform/graphics/gpu/cocoa/GPUSwapChainMetal.mm:
(WebCore::GPUSwapChain::tryCreate):

LayoutTests:

GPUSwapChains are now configured via GPUCanvasContext instead of GPUDevice.

* webgpu/blend-triangle-strip.html:
* webgpu/js/webgpu-functions.js:
(createBasicSwapChain):
* webgpu/whlsl.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244507 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[JSC] Use volatile load to populate backing page in MarkedBlock::Footer instead of...
ysuzuki@apple.com [Mon, 22 Apr 2019 18:35:04 +0000 (18:35 +0000)]
[JSC] Use volatile load to populate backing page in MarkedBlock::Footer instead of using holdLock
https://bugs.webkit.org/show_bug.cgi?id=197152

Reviewed by Saam Barati.

Emit volatile load instead of using holdLock to populate backing page in MarkedBlock::Footer.

* heap/BlockDirectory.cpp:
(JSC::BlockDirectory::isPagedOut):
* heap/MarkedBlock.h:
(JSC::MarkedBlock::populatePage const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244506 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[JSC] useJIT should subsume useRegExpJIT
ysuzuki@apple.com [Mon, 22 Apr 2019 18:30:16 +0000 (18:30 +0000)]
[JSC] useJIT should subsume useRegExpJIT
https://bugs.webkit.org/show_bug.cgi?id=197153

Reviewed by Alex Christensen.

useJIT should subsume useRegExpJIT. We should immediately disable JIT feature if useJIT = false,
even if useRegExpJIT is true.

* dfg/DFGCapabilities.cpp:
(JSC::DFG::isSupported):
* runtime/Options.cpp:
(JSC::recomputeDependentOptions):
* runtime/RegExp.cpp:
(JSC::RegExp::compile):
(JSC::RegExp::compileMatchOnly):
* runtime/VM.cpp:
(JSC::enableAssembler):
(JSC::VM::canUseRegExpJIT): Deleted.
* runtime/VM.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244505 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Use singular verb in failure description in case of single api test failure
aakash_jain@apple.com [Mon, 22 Apr 2019 18:29:40 +0000 (18:29 +0000)]
[ews-build] Use singular verb in failure description in case of single api test failure
https://bugs.webkit.org/show_bug.cgi?id=197013

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/steps.py:
(AnalyzeAPITestsResults.analyzeResults):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244504 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoMark SVGStringList properties '[SameObject]' in the IDL files
said@apple.com [Mon, 22 Apr 2019 18:17:33 +0000 (18:17 +0000)]
Mark SVGStringList properties '[SameObject]' in the IDL files
Followup to https://bugs.webkit.org/show_bug.cgi?id=197137

Patch by Said Abou-Hallawa <said@apple.com> on 2019-04-22
Reviewed by Darin Adler.

The SVG elements do not create tear-off wrappers for SVGStrigList DOM
objects anymore. Instead they return Ref pointers to the same RefCounted
objects. So they should be marked '[SameObject]' in their IDL files.

* svg/SVGTests.idl:
* svg/SVGViewElement.idl:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244503 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoDelayed WebProcessLaunch may break the _relatedWebView SPI
cdumez@apple.com [Mon, 22 Apr 2019 18:03:47 +0000 (18:03 +0000)]
Delayed WebProcessLaunch may break the _relatedWebView SPI
https://bugs.webkit.org/show_bug.cgi?id=197160

Reviewed by Alex Christensen.

Source/WebKit:

Delayed WebProcessLaunch may break the _relatedWebView SPI. The breakage would happen if the client
would relate a WebView to another which has not launched its initial process yet.

To address the issue, when we need a running process for a WebView which has a related view, we need
to make sure the related view has a running process and use that process. Previously, we would share
the "dummy" process instead.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::launchProcess):
(WebKit::WebPageProxy::ensureRunningProcess):
* UIProcess/WebPageProxy.h:
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::createWebPage):

Tools:

Add API test coverage.

* TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244502 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[PlayStation] Restructuring Remote Inspector classes to support multiple platform.
basuke.suzuki@sony.com [Mon, 22 Apr 2019 16:38:11 +0000 (16:38 +0000)]
[PlayStation] Restructuring Remote Inspector classes to support multiple platform.
https://bugs.webkit.org/show_bug.cgi?id=197030

Reviewed by Don Olmstead.

Restructuring the PlayStation's RemoteInspector backend which uses native socket for the communication to be ready for WinCairo.

What we did is basically:
- Renamed `remote/playstation/` to `remote/socket/`. This directory is now platform independent implementation of socket backend.
- Renamed `RemoteInspectorSocket` class to `RemoteInspectorSocketEndpoint`. This class is platform independent and core of the backend.
- Merged `RemoteInspectorSocket{Client|Server}` classes into `RemoteInspectorSocketEndpoint` class because the differences are little.
- Defined a new interface functions in `Inspector::Socket` (new) namespace.
- Moved POSIX socket implementation into `posix\RemoteInspectorSocketPOSIX.{h|cpp}`.

* PlatformPlayStation.cmake:
* inspector/remote/RemoteInspector.h:
* inspector/remote/playstation/RemoteInspectorSocketClient.h: Merged into RemoteInspectorSocketEndpoint.
* inspector/remote/playstation/RemoteInspectorSocketClientPlayStation.cpp: Merged into RemoteInspectorSocketEndpoint.
* inspector/remote/playstation/RemoteInspectorSocketPlayStation.cpp: Removed.
* inspector/remote/playstation/RemoteInspectorSocketServer.h: Merged into RemoteInspectorSocketEndpoint.
* inspector/remote/playstation/RemoteInspectorSocketServerPlayStation.cpp: Merged into RemoteInspectorSocketEndpoint.
* inspector/remote/socket/RemoteInspectorConnectionClient.cpp: Renamed from inspector\remote\playstation\RemoteInspectorConnectionClientPlayStation.cpp.
* inspector/remote/socket/RemoteInspectorConnectionClient.h: Renamed from inspector\remote\playstation\RemoteInspectorConnectionClient.h.
(Inspector::RemoteInspectorConnectionClient::didAccept):
* inspector/remote/socket/RemoteInspectorMessageParser.cpp: Renamed from inspector\remote\playstation\RemoteInspectorMessageParserPlayStation.cpp.
* inspector/remote/socket/RemoteInspectorMessageParser.h: Renamed from inspector\remote\playstation\RemoteInspectorMessageParser.h.
* inspector/remote/socket/RemoteInspectorServer.cpp: Renamed from inspector\remote\playstation\RemoteInspectorServerPlayStation.cpp.
(Inspector::RemoteInspectorServer::didAccept):
(Inspector::RemoteInspectorServer::start):
* inspector/remote/socket/RemoteInspectorServer.h: Renamed from inspector\remote\playstation\RemoteInspectorServer.h.
* inspector/remote/socket/RemoteInspectorSocket.cpp: Renamed from inspector\remote\playstation\RemoteInspectorPlayStation.cpp.
(Inspector::RemoteInspector::start):
* inspector/remote/socket/RemoteInspectorSocket.h: Copied from inspector\remote\playstation\RemoteInspectorSocket.h.
* inspector/remote/socket/RemoteInspectorSocketEndpoint.cpp: Added.
(Inspector::RemoteInspectorSocketEndpoint::RemoteInspectorSocketEndpoint):
(Inspector::RemoteInspectorSocketEndpoint::~RemoteInspectorSocketEndpoint):
(Inspector::RemoteInspectorSocketEndpoint::wakeupWorkerThread):
(Inspector::RemoteInspectorSocketEndpoint::connectInet):
(Inspector::RemoteInspectorSocketEndpoint::listenInet):
(Inspector::RemoteInspectorSocketEndpoint::isListening):
(Inspector::RemoteInspectorSocketEndpoint::workerThread):
(Inspector::RemoteInspectorSocketEndpoint::createClient):
(Inspector::RemoteInspectorSocketEndpoint::recvIfEnabled):
(Inspector::RemoteInspectorSocketEndpoint::sendIfEnabled):
(Inspector::RemoteInspectorSocketEndpoint::send):
(Inspector::RemoteInspectorSocketEndpoint::acceptInetSocketIfEnabled):
* inspector/remote/socket/RemoteInspectorSocketEndpoint.h: Renamed from inspector\remote\playstation\RemoteInspectorSocket.h.
* inspector/remote/socket/posix/RemoteInspectorSocketPOSIX.cpp: Added.
(Inspector::Socket::connect):
(Inspector::Socket::listen):
(Inspector::Socket::accept):
(Inspector::Socket::createPair):
(Inspector::Socket::setup):
(Inspector::Socket::isValid):
(Inspector::Socket::isListening):
(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):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244501 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[GTK] fix gtk_style_context_set_background deprecation
commit-queue@webkit.org [Mon, 22 Apr 2019 16:31:41 +0000 (16:31 +0000)]
[GTK] fix gtk_style_context_set_background deprecation
https://bugs.webkit.org/show_bug.cgi?id=196912

Patch by Ludovico de Nittis <ludovico.denittis@collabora.com> on 2019-04-22
Reviewed by Carlos Garcia Campos.

Since gtk 3.18 gtk_style_context_set_background has been deprecated.
Backgrounds are not rendered in the UI process, so it can be
simply removed.

* UIProcess/API/gtk/WebKitWebViewBase.cpp:
(webkitWebViewBaseRealize):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244500 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244495.
commit-queue@webkit.org [Mon, 22 Apr 2019 15:50:55 +0000 (15:50 +0000)]
Unreviewed, rolling out r244495.
https://bugs.webkit.org/show_bug.cgi?id=197159

Causing build failures on OpenSource and Internal bots
(Requested by ShawnRoberts on #webkit).

Reverted changeset:

"WHLSLPrepare.cpp always recompiles, even if nothing was
changed"
https://bugs.webkit.org/show_bug.cgi?id=197151
https://trac.webkit.org/changeset/244495

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244499 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoREGRESSION(r241289): [GTK] accessibility/removed-continuation-element-causes-crash...
carlosgc@webkit.org [Mon, 22 Apr 2019 07:58:11 +0000 (07:58 +0000)]
REGRESSION(r241289): [GTK] accessibility/removed-continuation-element-causes-crash.html and accessibility/removed-anonymous-block-child-causes-crash.html crashes
https://bugs.webkit.org/show_bug.cgi?id=194630

Reviewed by Michael Catanzaro.

Source/WebCore:

Do not assume core object always has a wrapper in webkitAccessibleRefChild().

Fixes: accessibility/removed-continuation-element-causes-crash.html
       accessibility/removed-anonymous-block-child-causes-crash.html

* accessibility/atk/WebKitAccessible.cpp:
(webkitAccessibleRefChild): Return early if wrapper is nullptr.

Tools:

Return early from getAttributeSet() is passed in accessible is nullptr.

* WebKitTestRunner/InjectedBundle/atk/AccessibilityUIElementAtk.cpp:

LayoutTests:

Remove expectations for tests that are passing now.

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244498 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[bmalloc] Use StaticPerProcess' mutex as bmalloc::Heap does with PerProcess
ysuzuki@apple.com [Mon, 22 Apr 2019 06:44:25 +0000 (06:44 +0000)]
[bmalloc] Use StaticPerProcess' mutex as bmalloc::Heap does with PerProcess
https://bugs.webkit.org/show_bug.cgi?id=197135

Reviewed by Darin Adler.

This patch leverages StaticPerProcess::mutex() for per process instance's lock in various classes,
as Heap does with PerProcess::mutex().

* bmalloc/AllIsoHeaps.cpp:
(bmalloc::AllIsoHeaps::add):
(bmalloc::AllIsoHeaps::head):
* bmalloc/AllIsoHeaps.h:
* bmalloc/CryptoRandom.cpp:
(bmalloc::ARC4RandomNumberGenerator::randomValues):
* bmalloc/DebugHeap.cpp:
(bmalloc::DebugHeap::memalignLarge):
(bmalloc::DebugHeap::freeLarge):
* bmalloc/DebugHeap.h:
* bmalloc/Scavenger.cpp:
(bmalloc::Scavenger::run):
(bmalloc::Scavenger::runSoon):
(bmalloc::Scavenger::scheduleIfUnderMemoryPressure):
(bmalloc::Scavenger::schedule):
(bmalloc::Scavenger::timeSinceLastFullScavenge):
(bmalloc::Scavenger::scavenge):
(bmalloc::Scavenger::threadRunLoop):
* bmalloc/Scavenger.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244497 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[WTF] Generic memoryFootprint() implementation should use bmalloc on Linux
zandobersek@gmail.com [Mon, 22 Apr 2019 06:22:59 +0000 (06:22 +0000)]
[WTF] Generic memoryFootprint() implementation should use bmalloc on Linux
https://bugs.webkit.org/show_bug.cgi?id=196963

Reviewed by Don Olmstead.

Have the generic memoryFootprint() implementation use bmalloc's
memoryFootprint() API on Linux, whenever the system malloc option is
not enabled. Limitation to Linux platforms is due to the bmalloc
implementation being limited to those configurations (excluding iOS
which doesn't use MemoryFootprintGeneric.cpp).

* wtf/PlatformWPE.cmake: Switch to building MemoryFootprintGeneric.cpp.
* wtf/generic/MemoryFootprintGeneric.cpp:
(WTF::memoryFootprint):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244496 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoWHLSLPrepare.cpp always recompiles, even if nothing was changed
darin@apple.com [Mon, 22 Apr 2019 00:24:03 +0000 (00:24 +0000)]
WHLSLPrepare.cpp always recompiles, even if nothing was changed
https://bugs.webkit.org/show_bug.cgi?id=197151

Reviewed by Dan Bernstein.

* DerivedSources-input.xcfilelist: Script updated this automatically after
DerivedSources.make was corrected.
* DerivedSources-output.xcfilelist: Ditto, although I had to manually remove
one bogus leftover reference to WHLSLStandardLibrary.cpp.

* DerivedSources.make: Updated the rule that builds WHSLStandardLibrary.h to
no longer refer to nonexistent WHLSLStandardLibrary.cpp. Because the dependency
was on a file that was never created, the rule to regenerate WHSLStandardLibrary.h
was running on every build, instead of only when one of the dependencies changed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244495 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoDefer EditorState updates until the next layer tree flush in a few additional circums...
wenson_hsieh@apple.com [Sun, 21 Apr 2019 21:10:31 +0000 (21:10 +0000)]
Defer EditorState updates until the next layer tree flush in a few additional circumstances
https://bugs.webkit.org/show_bug.cgi?id=197145
<rdar://problem/50078170>

Reviewed by Darin Adler.

Source/WebKit:

Gets rid of sendPartialEditorStateAndSchedulePostLayoutUpdate(), in favor of always scheduling a full editor
state update after the next compositing flush.

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::updateEditorStateAfterLayoutIfEditabilityChanged):
(WebKit::WebPage::setNeedsFontAttributes):
(WebKit::WebPage::didChangeOverflowScrollPosition):
(WebKit::WebPage::didChangeSelection):
(WebKit::WebPage::didChangeSelectionOrOverflowScrollPosition):
(WebKit::WebPage::sendPartialEditorStateAndSchedulePostLayoutUpdate): Deleted.
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::platformEditorState const):
(WebKit::WebPage::updateVisibleContentRects):

Tools:

Adjusts several editing API tests to wait for a presentation update following a selection change, programmatic
focus, or showing the font manager.

* TestWebKitAPI/Tests/WebKitCocoa/EditorStateTests.mm:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebKitCocoa/WKContentViewTargetForAction.mm:
(TEST):
* TestWebKitAPI/Tests/mac/FontManagerTests.mm:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244494 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoRegression(r243767) WebFrame::m_navigationIsContinuingInAnotherProcess flag is never...
cdumez@apple.com [Sun, 21 Apr 2019 20:14:04 +0000 (20:14 +0000)]
Regression(r243767) WebFrame::m_navigationIsContinuingInAnotherProcess flag is never reset
https://bugs.webkit.org/show_bug.cgi?id=197144

Reviewed by Darin Adler.

WebFrame::m_navigationIsContinuingInAnotherProcess flag is never reset since it was introduced in
r243767. This leads to leaking Navigation objects in the UIProcess when reusing a previously
suspended process because such process will no longer send the DidDestroyNavigation IPC.

It turns out that resetting the flags causes API tests such as ProcessSwap.QuickBackForwardNavigationWithPSON
to ASSERT. This is because when the UIProcess quickly navigate back and forth without waiting for policy
decisions, we may end up getting the policy decision for a particular navigation *after* we've sent the
DidDestroyNavigation.

As a result, this patch reverts r243767 and fixes in the assertion in http/tests/adClickAttribution/store-ad-click-attribution.html
another way. We initially assumed that the logic in WebPageProxy::didDestroyNavigation() was failing to
ignore the DidDestroyNavigation from the previous process after a swap due to a race, maybe because it was
sometimes received too late and m_provisionalPage was already cleared. However, this would not make sense
since the test is crashing consistently and the page would no longer be able to receive IPC from the
previous process *after* we've committed the provisional process/page.

The real issue was that the DidDestroyNavigation IPC was received *before* we could construct the
provisional page, which is why the logic in WebPageProxy::didDestroyNavigation() was failing to ignore
the bad IPC. In WebPageProxy::receivedNavigationPolicyDecision(), we were calling receivedPolicyDecision()
(which would send the DidReceivePolicyDecision to the previous WebProcess) and then continueNavigationInNewProcess()
in order to construct the provisional page. I personally did not expect we could receive IPC between the
calls to receivedNavigationPolicyDecision() and receivedPolicyDecision(), since we are not yielding and since
the DidReceivePolicyDecision IPC is asynchronous. However, this is exactly what was happening in the context
of this test. The reason is that the DidReceivePolicyDecision IPC was getting wrapped in a synchronous message
and sent as synchronous message due to the Connection::m_inDispatchMessageMarkedToUseFullySynchronousModeForTesting
flag which seems to get set in the test due to some EventSender IPC. I believe this is because the test uses
EventSender to do a click on a link which triggers the navigation.

To address the issue, I now call receivedNavigationPolicyDecision() *after* continueNavigationInNewProcess()
to make sure that we always start the provisional load in the new process before we tell the previous process
to stop loading. This way, there is no way we get IPC from the previous process about the current navigation
before we have a provisional page.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::receivedNavigationPolicyDecision):
(WebKit::WebPageProxy::didDestroyNavigation):
* WebProcess/WebPage/WebFrame.cpp:
(WebKit::WebFrame::didReceivePolicyDecision):
(WebKit::WebFrame::documentLoaderDetached):
* WebProcess/WebPage/WebFrame.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244493 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoWebDriver: add hooks to make it possible to easily run WPT WebDriver web server by...
bburg@apple.com [Sun, 21 Apr 2019 19:48:46 +0000 (19:48 +0000)]
WebDriver: add hooks to make it possible to easily run WPT WebDriver web server by itself
https://bugs.webkit.org/show_bug.cgi?id=197125
<rdar://problem/50061468>

Reviewed by Darin Adler.

Sometimes I want to run wptserve independently of tests. Make it easier to script that
using this class rather than making our own ./wpt serve invocation.

* Scripts/webkitpy/webdriver_tests/webdriver_w3c_web_server.py:
(WebDriverW3CWebServer):
(WebDriverW3CWebServer.document_root): Expose this to clients.
(WebDriverW3CWebServer.wait): Add a method that blocks until the server dies.

(WebDriverW3CWebServer.__enter__):
(WebDriverW3CWebServer.__exit__):
Make it possible to use the web server with a 'with' statement.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244492 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoREGRESSION (r243137): SVGViewElement.viewTarget should not return a new object
said@apple.com [Sun, 21 Apr 2019 00:03:44 +0000 (00:03 +0000)]
REGRESSION (r243137): SVGViewElement.viewTarget should not return a new object
https://bugs.webkit.org/show_bug.cgi?id=197137

Reviewed by Darin Adler.

Source/WebCore:

All the DOM objects accessing the viewTarget of the same SVGViewElement
should hold a Ref pointer to the same SVGStringList property.

Test: svg/dom/SVGViewElement-viewTarget.html

* svg/SVGViewElement.idl:

LayoutTests:

* svg/dom/SVGViewElement-viewTarget-expected.txt: Added.
* svg/dom/SVGViewElement-viewTarget.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244491 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoREGRESSION(r243958): Unnecessary deactivation of AudioSession (PLT Regression)
jer.noble@apple.com [Sat, 20 Apr 2019 23:46:25 +0000 (23:46 +0000)]
REGRESSION(r243958): Unnecessary deactivation of AudioSession (PLT Regression)
https://bugs.webkit.org/show_bug.cgi?id=197123
<rdar://problem/49783264>

Reviewed by Per Arne Vollan.

Only set m_becameActive if we actually activated the AudioSession before starting playback. This
avoids unnecessarily deactivating the AudioSession in processWillSuspend().

* platform/audio/PlatformMediaSessionManager.cpp:
(WebCore::PlatformMediaSessionManager::sessionWillBeginPlayback):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244490 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoHashTable::removeIf always shrinks the hash table by half even if there is nothing...
rniwa@webkit.org [Sat, 20 Apr 2019 19:16:22 +0000 (19:16 +0000)]
HashTable::removeIf always shrinks the hash table by half even if there is nothing left
https://bugs.webkit.org/show_bug.cgi?id=196681
<rdar://problem/49917764>

Reviewed by Darin Adler.

Address Darin's comments by removing the explicit type from std::max.

* wtf/HashTable.h:
(WTF::KeyTraits>::computeBestTableSize):
(WTF::KeyTraits>::shrinkToBestSize):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244489 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[CMake][Win] Use target oriented design for WebKitLegacy
don.olmstead@sony.com [Sat, 20 Apr 2019 19:00:15 +0000 (19:00 +0000)]
[CMake][Win] Use target oriented design for WebKitLegacy
https://bugs.webkit.org/show_bug.cgi?id=197112

Reviewed by Konstantin Tokarev.

.:

Override WebKitLegacy_DERIVED_SOURCES_DIR within WinCairo in support of a target
oriented design.

* Source/cmake/OptionsWinCairo.cmake:

Source/WebKitLegacy:

Enumerate the public framework headers for WebKitLegacy and copy them. Include
directories should be PRIVATE. Migrate to using WebKitLegacy_DERIVED_SOURCES_DIR.

* CMakeLists.txt:
* PlatformMac.cmake:
* PlatformWin.cmake:

Source/WebKitLegacy/win:

Fix an include directory.

* WebKitQuartzCoreAdditions/CMakeLists.txt:

Tools:

Fix includes and libraries for DumpRenderTree after changes to WebKitLegacy targets.

* DumpRenderTree/CMakeLists.txt:
* DumpRenderTree/PlatformWin.cmake:
* DumpRenderTree/win/EventSender.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244488 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUpdate text for old EWS submit button
aakash_jain@apple.com [Sat, 20 Apr 2019 18:44:53 +0000 (18:44 +0000)]
Update text for old EWS submit button
https://bugs.webkit.org/show_bug.cgi?id=197140

Reviewed by Alexey Proskuryakov.

* QueueStatusServer/templates/statusbubble.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244487 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ews-app] New EWS should poll bugzilla more frequently
aakash_jain@apple.com [Sat, 20 Apr 2019 18:31:48 +0000 (18:31 +0000)]
[ews-app] New EWS should poll bugzilla more frequently
https://bugs.webkit.org/show_bug.cgi?id=197138

Reviewed by Alexey Proskuryakov.

* BuildSlaveSupport/ews-app/ews/fetcher.py:
(FetchLoop.__init__):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244486 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[ews-app] Submit to EWS buttons are not aligned properly with status-bubbles
aakash_jain@apple.com [Sat, 20 Apr 2019 18:30:28 +0000 (18:30 +0000)]
[ews-app] Submit to EWS buttons are not aligned properly with status-bubbles
https://bugs.webkit.org/show_bug.cgi?id=197139

Reviewed by Alexey Proskuryakov.

* BuildSlaveSupport/ews-app/ews/templates/statusbubble.html:
* QueueStatusServer/templates/statusbubble.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244485 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUnreviewed, fix iOS build with recent SDKs.
cdumez@apple.com [Sat, 20 Apr 2019 16:44:50 +0000 (16:44 +0000)]
Unreviewed, fix iOS build with recent SDKs.

Source/WebKit:

* UIProcess/API/Cocoa/WKWebView.mm:
(deviceOrientation):
* UIProcess/ios/fullscreen/WKFullScreenViewController.mm:
(-[WKFullScreenViewController initWithWebView:]):
(-[WKFullScreenViewController viewWillTransitionToSize:withTransitionCoordinator:]):
(-[WKFullScreenViewController _statusBarFrameDidChange:]):

Tools:

* TestWebKitAPI/ios/DragAndDropSimulatorIOS.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244484 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoUnreviewed, suppress warnings in non Darwin environments
ysuzuki@apple.com [Sat, 20 Apr 2019 07:08:19 +0000 (07:08 +0000)]
Unreviewed, suppress warnings in non Darwin environments

* jit/ExecutableAllocator.cpp:
(JSC::dumpJITMemory):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244483 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoFix Long Press Image Test
megan_gardner@apple.com [Sat, 20 Apr 2019 03:57:42 +0000 (03:57 +0000)]
Fix Long Press Image Test
https://bugs.webkit.org/show_bug.cgi?id=197130
<rdar://problem/50058818>

Reviewed by Wenson Hsieh.

Fixing up test to reflect changes in https://bugs.webkit.org/show_bug.cgi?id=196833

* fast/events/touch/ios/long-press-on-image-expected.txt:
* fast/events/touch/ios/long-press-on-image.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244482 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months ago[bmalloc] IsoHeap should have lower tier using shared IsoPage
ysuzuki@apple.com [Sat, 20 Apr 2019 03:29:40 +0000 (03:29 +0000)]
[bmalloc] IsoHeap should have lower tier using shared IsoPage
https://bugs.webkit.org/show_bug.cgi?id=196837

Reviewed by Filip Pizlo.

IsoHeap had a scalability problem. Once one instance is allocated from IsoHeap, it immediately allocates 16KB page for this type.
But some types allocate only a few instances. It leads to memory wastage, and it also limits the scalability of IsoHeap since
we need to carefully select classes which will be confined in IsoHeap due to this characteristics. If we can remove this wastage,
we can apply IsoHeap more aggressively without causing memory regression, this is the goal of this patch.

In this patch, we introduce a slow tier to IsoHeap allocation. Initially, the allocator for a certain type allocates instances from
a shared page with the other allocators, and eventually, the allocator tiers up and gets dedicated pages if instances of the type
are allocated a lot. This "shared" tier is slow, but it is totally OK because we will tier up to the normal fast tier if allocation
frequently happens. Even the instance is allocated from pages shared with the other allocators, we still make the allocated memory
region dedicated to the specific type: once a memory region is allocated for a certain type from a shared page, this region continues
being used only for this type even after this memory is freed. To summarize the changes:

1. We introduce "shared" tier to IsoHeap allocation. Up to N (N = 8 for now, but we can pick any power-of-two numbers up to 32) allocations,
   we continue using this tier. We allocate memory from shared pages so that we do not waste 16KB pages for types which only allocates a few instances.

2. We eventually tier up to the "fast" tier, and eventually tier down to the "shared" tier too. We measure the period between slow paths,
   and switch the appropriate tier for the type. Currently, we use 1 seconds as heuristics. We also count # of allocations per cycle to
   avoid pathological slow downs.

3. Shared page mechanism must keep the characteristics of IsoHeap. Once a memory region is allocated for a certain type, this memory region
   must be dedicated to this type. We keep track the allocated memory regions from shared pages in IsoHeapImpl, and ensure that we never
   reuse a memory region for a different type.

This patch improves PLUM2 by 1.4% (128.4MB v.s. 126.62MB), and early Speedometer2 results are performance-neutral.

* CMakeLists.txt:
* bmalloc.xcodeproj/project.pbxproj:
* bmalloc/Algorithm.h:
(bmalloc::roundUpToMultipleOfImpl):
(bmalloc::roundUpToMultipleOf):
* bmalloc/BCompiler.h:
* bmalloc/BExport.h:
* bmalloc/FreeList.h:
* bmalloc/IsoAllocator.h:
* bmalloc/IsoAllocatorInlines.h:
(bmalloc::IsoAllocator<Config>::allocateSlow):
* bmalloc/IsoDeallocator.h:
* bmalloc/IsoDeallocatorInlines.h:
(bmalloc::IsoDeallocator<Config>::deallocate):
* bmalloc/IsoHeapImpl.h:
* bmalloc/IsoHeapImplInlines.h:
(bmalloc::IsoHeapImpl<Config>::scavenge):
(bmalloc::IsoHeapImpl<Config>::forEachLiveObject):
(bmalloc::IsoHeapImpl<Config>::updateAllocationMode):
(bmalloc::IsoHeapImpl<Config>::allocateFromShared):
* bmalloc/IsoPage.h:
(bmalloc::IsoPageBase::IsoPageBase):
(bmalloc::IsoPageBase::isShared const):
* bmalloc/IsoPageInlines.h:
(bmalloc::IsoPage<Config>::IsoPage):
(bmalloc::IsoPageBase::pageFor):
(bmalloc::IsoPage<Config>::pageFor):
(bmalloc::IsoPage<Config>::free):
* bmalloc/IsoSharedConfig.h: Copied from Source/bmalloc/bmalloc/BExport.h.
* bmalloc/IsoSharedHeap.cpp: Copied from Source/bmalloc/bmalloc/BExport.h.
* bmalloc/IsoSharedHeap.h: Copied from Source/bmalloc/bmalloc/IsoAllocator.h.
(bmalloc::VariadicBumpAllocator::VariadicBumpAllocator):
(bmalloc::IsoSharedHeap::IsoSharedHeap):
* bmalloc/IsoSharedHeapInlines.h: Added.
(bmalloc::VariadicBumpAllocator::allocate):
(bmalloc::IsoSharedHeap::allocateNew):
(bmalloc::IsoSharedHeap::allocateSlow):
* bmalloc/IsoSharedPage.cpp: Copied from Source/bmalloc/bmalloc/BExport.h.
(bmalloc::IsoSharedPage::tryCreate):
* bmalloc/IsoSharedPage.h: Copied from Source/bmalloc/bmalloc/IsoDeallocator.h.
(bmalloc::IsoSharedPage::IsoSharedPage):
(bmalloc::indexSlotFor):
* bmalloc/IsoSharedPageInlines.h: Added.
(bmalloc::IsoSharedPage::free):
(bmalloc::IsoSharedPage::startAllocating):
(bmalloc::IsoSharedPage::stopAllocating):
* bmalloc/IsoTLS.h:
* bmalloc/IsoTLSInlines.h:
(bmalloc::IsoTLS::deallocateImpl):
(bmalloc::IsoTLS::deallocateFast):
(bmalloc::IsoTLS::deallocateSlow):
* bmalloc/StdLibExtras.h:
(bmalloc::bitwise_cast):
* test/testbmalloc.cpp:
(testIsoMallocAndFreeFast):
(run):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244481 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoAbstractValue can represent more than int52
sbarati@apple.com [Sat, 20 Apr 2019 00:37:43 +0000 (00:37 +0000)]
AbstractValue can represent more than int52
https://bugs.webkit.org/show_bug.cgi?id=197118
<rdar://problem/49969960>

Reviewed by Michael Saboff.

JSTests:

* stress/abstract-value-can-include-int52.js: Added.
(foo):
(index.index.8.index.60.index.65.index.1234.index.1234.parseInt.string_appeared_here.String.fromCharCode):

Source/JavaScriptCore:

Let's analyze this control flow diamond:

#0
branch #1, #2

#1:
PutStack(JSValue, loc42)
Jump #3

#2:
PutStack(Int52, loc42)
Jump #3

#3:
...

Our abstract value for loc42 at the head of #3 will contain an abstract
value that us the union of Int52 with other things. Obviously in the
above program, a GetStack for loc42 would be inavlid, since it might
be loading either JSValue or Int52. However, the abstract interpreter
just tracks what the value could be, and it could be Int52 or JSValue.

When I did the Int52 refactoring, I expected such things to never happen,
but it turns out it does. We should just allow for this instead of asserting
against it since it's valid IR to do the above.

* bytecode/SpeculatedType.cpp:
(JSC::dumpSpeculation):
* dfg/DFGAbstractValue.cpp:
(JSC::DFG::AbstractValue::checkConsistency const):
* dfg/DFGAbstractValue.h:
(JSC::DFG::AbstractValue::validateTypeAcceptingBoxedInt52 const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244480 268f45cc-cd09-0410-ab3c-d52691b4dbfc

4 months agoWeb Inspector: Timelines: CPU: remove extra border under last section
drousso@apple.com [Sat, 20 Apr 2019 00:33:53 +0000 (00:33 +0000)]
Web Inspector: Timelines: CPU: remove extra border under last section
https://bugs.webkit.org/show_bug.cgi?id=197120

Reviewed by Joseph Pecoraro.

* UserInterface/Views/CPUTimelineView.css:
(.timeline-view.cpu > .content > .overview):
(.timeline-view.cpu > .content > .overview:not(:last-child)): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244479 268f45cc-cd09-0410-ab3c-d52691b4dbfc