WebKit-https.git
10 months agoWeb Inspector: Restrict domains at the target level instead of only at the window...
joepeck@webkit.org [Thu, 8 Nov 2018 20:27:04 +0000 (20:27 +0000)]
Web Inspector: Restrict domains at the target level instead of only at the window level
https://bugs.webkit.org/show_bug.cgi?id=191344

Reviewed by Devin Rousso.

Source/JavaScriptCore:

* inspector/protocol/Console.json:
* inspector/protocol/Debugger.json:
* inspector/protocol/Heap.json:
* inspector/protocol/Runtime.json:
Remove workerSupported as it is now no longer necessary. It is implied
by availability being empty (meaning it is supported everywhere).

* inspector/protocol/Inspector.json:
* inspector/protocol/ScriptProfiler.json:
Restrict to "javascript" and "web" debuggables, not available in workers.

* inspector/protocol/Worker.json:
Cleanup, remove empty types list.

* inspector/protocol/Recording.json:
Cleanup, only expose this in the "web" domain for now.

* inspector/scripts/codegen/generate_js_backend_commands.py:
(JSBackendCommandsGenerator.generate_domain):
* inspector/scripts/codegen/models.py:
(Protocol.parse_domain):
Allow a list of debuggable types. Add "worker" even though it is unused
since that is a type we would want to allow or consider.

(Domain.__init__):
(Domains):
Remove now unnecessary workerSupported code.
Allow availability on a domain with only types.

* inspector/scripts/tests/generic/expected/worker-supported-domains.json-result: Removed.
* inspector/scripts/tests/generic/worker-supported-domains.json: Removed.

Source/WebCore:

* inspector/WorkerInspectorController.cpp:
(WebCore::WorkerInspectorController::WorkerInspectorController):
Remove Inspector domain from workers. It was unused anyways.

Source/WebInspectorUI:

With this change the list of agents available on the target should accurately
reflect the domains that target supports. We should start moving in the direction
that if we are checking for, or using, a feature, we should ideally be using
code that looks like `target.FooAgent.method` and not `window.FooAgent` or
an unprefixed `FooAgent`. More changes to follow.

* UserInterface/Controllers/AppControllerBase.js:
* UserInterface/Base/DebuggableType.js: Added.
Extract debuggable type enum out into its own file loaded early on.

* UserInterface/Main.html:
* UserInterface/Test.html:
Load DebuggableType.js.

* UserInterface/Controllers/NetworkManager.js:
(WI.NetworkManager.prototype.initializeTarget):
With these changes target.ServiceWorker is only exposed on a ServiceWorker target.

* UserInterface/Protocol/Connection.js:
(InspectorBackend.WorkerConnection):
With these changes Target.js restricts the domains exposed on the Target so just
clone all of the agents (we don't need to do anything special).

* UserInterface/Protocol/InspectorBackend.js:
(InspectorBackendClass.prototype.activateDomain):
Update supported domains per-type. If there is no explicit availability
then the domain is supported on all types.

(InspectorBackendClass.prototype.supportedDomainsForDebuggableType):
Get supported domains per-type.

(InspectorBackendClass.prototype.get workerSupportedDomains): Deleted.
(InspectorBackendClass.prototype.workerSupportedDomain): Deleted.
workerSupportedDomains is now unnecessary.

* UserInterface/Protocol/Target.js:
(WI.Target.prototype._supportedDomainsForTargetType):
Restrict supported domains for this target based on its type.

* UserInterface/Protocol/Legacy/10.3/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/11.0/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/11.3/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/12.0/InspectorBackendCommands.js:
* Versions/Inspector-iOS-10.3.json:
* Versions/Inspector-iOS-11.0.json:
* Versions/Inspector-iOS-11.3.json:
* Versions/Inspector-iOS-12.0.json:
Regenerate protocol files now that workerSupportedDomains is unnecessary
and explicit availability has been added to other domains.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237997 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[Sierra Release WK2] Layout Test media/track/track-cue-css.html is a flaky failure
pvollan@apple.com [Thu, 8 Nov 2018 20:08:50 +0000 (20:08 +0000)]
[Sierra Release WK2] Layout Test media/track/track-cue-css.html is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=191200

Unreviewed, speculative fix for test flakiness.

* media/track/track-cue-css-expected.html:
* media/track/track-cue-css.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237996 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[Payment Request] Update web platform tests
aestes@apple.com [Thu, 8 Nov 2018 19:53:53 +0000 (19:53 +0000)]
[Payment Request] Update web platform tests
https://bugs.webkit.org/show_bug.cgi?id=191426

Updated web-platform-tests/payment-request/ to match wpt as of commit 577284e.

Reviewed by Jiewen Tan.

* resources/import-expectations.json:
* web-platform-tests/payment-request/META.yml:
* web-platform-tests/payment-request/payment-request-abort-method.https.html:
* web-platform-tests/payment-request/payment-request-canmakepayment-method-protection.https-expected.txt: Added.
* web-platform-tests/payment-request/payment-request-canmakepayment-method-protection.https.html: Added.
* web-platform-tests/payment-request/payment-request-canmakepayment-method.https.html:
* web-platform-tests/payment-request/payment-request-constructor.https.html:
* web-platform-tests/payment-request/payment-request-id-attribute.https.html:
* web-platform-tests/payment-request/payment-response/retry-method-manual.https.html:
* web-platform-tests/payment-request/rejects_if_not_active.https.html:
* web-platform-tests/payment-request/w3c-import.log:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237995 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoRemoved mbmalloc target from WTF
ggaren@apple.com [Thu, 8 Nov 2018 19:49:01 +0000 (19:49 +0000)]
Removed mbmalloc target from WTF
https://bugs.webkit.org/show_bug.cgi?id=191313

Reviewed by Saam Barati.

For benchmarking, WTF::fastMalloc is no longer meaningfully different
from bmalloc. (And bmalloc has its own mbmalloc target.)

* Configurations/mbmalloc.xcconfig: Removed.
* WTF.xcodeproj/project.pbxproj:
* wtf/mbmalloc.cpp: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237994 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoUnreviewed, a quick build fix for r237983
jiewen_tan@apple.com [Thu, 8 Nov 2018 19:41:42 +0000 (19:41 +0000)]
Unreviewed, a quick build fix for r237983

Temporarily make CtapRequestTest.cpp/CtapResponseTest.cpp to mac only to avoid iOS build failures.

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj
* TestWebKitAPI/Tests/WebCore/mac/CtapRequestTest.cpp: Renamed from Tools/TestWebKitAPI/Tests/WebCore/CtapRequestTest.cpp.
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebCore/mac/CtapResponseTest.cpp: Renamed from Tools/TestWebKitAPI/Tests/WebCore/CtapResponseTest.cpp.
(TestWebKitAPI::getTestCorruptedSignResponse):
(TestWebKitAPI::getTestCredentialRawIdBytes):
(TestWebKitAPI::convertToVector):
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237993 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoRe-add PLATFORM(IOS), now with the strict meaning
ap@apple.com [Thu, 8 Nov 2018 19:40:29 +0000 (19:40 +0000)]
Re-add PLATFORM(IOS), now with the strict meaning
https://bugs.webkit.org/show_bug.cgi?id=191281

Reviewed by Tim Horton.

* wtf/Platform.h: No change in behavior. Some of the macros look a bit weird
when expanded, it might be that the values are incorrect for some flavors of iOS family.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237992 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoFix steps_unittest.py failures related to recent Win10/Win7 changes
ryanhaddad@apple.com [Thu, 8 Nov 2018 19:25:23 +0000 (19:25 +0000)]
Fix steps_unittest.py failures related to recent Win10/Win7 changes
https://bugs.webkit.org/show_bug.cgi?id=191290

Reviewed by Alexey Proskuryakov.

* BuildSlaveSupport/build.webkit.org-config/steps_unittest.py: Update expected results.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237991 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAdd a String literal that returns a String
dino@apple.com [Thu, 8 Nov 2018 18:51:59 +0000 (18:51 +0000)]
Add a String literal that returns a String
https://bugs.webkit.org/show_bug.cgi?id=191425
<rdar://problem/45914556>

Reviewed by Sam Weinig.

Source/WebCore:

Use _str where possible.

API Test in WPT.

* Modules/fetch/FetchRequest.cpp:
(WebCore::computeReferrer):
* Modules/indexeddb/IDBKeyPath.cpp:
(WebCore::loggingString):
* Modules/webdatabase/OriginLock.cpp:
(WebCore::OriginLock::lockFileNameForPath):
* css/CSSBasicShapes.cpp:
(WebCore::updateCornerRadiusWidthAndHeight):
* html/canvas/WebGL2RenderingContext.cpp:
(WebCore::WebGL2RenderingContext::getParameter):
* html/canvas/WebGLRenderingContext.cpp:
(WebCore::WebGLRenderingContext::getParameter):
* loader/LinkHeader.cpp:
(WebCore::parseParameterValue):
* loader/LinkLoader.cpp:
(WebCore::LinkLoader::preloadIfNeeded):
* page/NavigatorBase.cpp:
(WebCore::NavigatorBase::platform):
* platform/DateComponents.cpp:
(WebCore::DateComponents::toString const):
* platform/mac/PlatformEventFactoryMac.mm:
(WebCore::keyIdentifierForKeyEvent):
* rendering/RenderListMarker.cpp:
(WebCore::RenderListMarker::suffix const):
* rendering/RenderMenuList.cpp:
(RenderMenuList::setText):
* testing/InternalSettings.cpp:
(WebCore::InternalSettings::userInterfaceDirectionPolicy):
(WebCore::InternalSettings::systemLayoutDirection):
* testing/Internals.cpp:
(WebCore::Internals::shadowRootType const):
(WebCore::Internals::getCurrentCursorInfo):

Source/WTF:

Add a new String literal, _str, that will return a String type.
This is useful when ""_s won't work, such as for things that
don't take an ASCIILiteral directly e.g. ExceptionOr<String>
or Variants.

* wtf/text/WTFString.h:
(WTF::StringLiterals::operator _str): Added.

Tools:

Test _str.

* TestWebKitAPI/Tests/WTF/WTFString.cpp:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237990 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agowebkitpy: Generalize trailing SDK specifier (Part 2)
jbedard@apple.com [Thu, 8 Nov 2018 16:16:18 +0000 (16:16 +0000)]
webkitpy: Generalize trailing SDK specifier (Part 2)
https://bugs.webkit.org/show_bug.cgi?id=191275
<rdar://problem/45703660>

Rubber-stamped by Aakash Jain.

Re-landing this change, configuration issues have been resolved.

* Scripts/webkitpy/port/ios_device.py:
(IOSDevicePort): Use normalized SDK.
* Scripts/webkitpy/port/ios_simulator.py:
(IOSSimulatorPort): Use normalized SDK.
(IOSSimulatorPort._build_driver_flags): Use platform SDK in build driver flags.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237989 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[MediaStream] Use DeviceIdHashSaltStorage to generate device ID hash salt
eric.carlson@apple.com [Thu, 8 Nov 2018 15:53:43 +0000 (15:53 +0000)]
[MediaStream] Use DeviceIdHashSaltStorage to generate device ID hash salt
https://bugs.webkit.org/show_bug.cgi?id=190825
<rdar://problem/45486085>

Source/WebKit:

Reviewed by Youenn Fablet.

* UIProcess/API/C/WKUserMediaPermissionCheck.cpp:
(WKUserMediaPermissionCheckSetUserMediaAccessInfo):
* UIProcess/API/glib/WebKitDeviceInfoPermissionRequest.cpp:
(webkitDeviceInfoPermissionRequestAllow):
(webkitDeviceInfoPermissionRequestDeny):
* UIProcess/Cocoa/UIDelegate.mm:
(WebKit::UIDelegate::UIClient::checkUserMediaPermissionForOrigin):
* UIProcess/DeviceIdHashSaltStorage.cpp:
(WebKit::DeviceIdHashSaltStorage::regenerateDeviceIdHashSaltForOrigin):
(WebKit::DeviceIdHashSaltStorage::deleteDeviceIdHashSaltForOrigin):
* UIProcess/DeviceIdHashSaltStorage.h:
* UIProcess/UserMediaPermissionCheckProxy.cpp:
(WebKit::UserMediaPermissionCheckProxy::setUserMediaAccessInfo):
* UIProcess/UserMediaPermissionCheckProxy.h:
* UIProcess/UserMediaPermissionRequestManagerProxy.cpp:
(WebKit::UserMediaPermissionRequestManagerProxy::createPermissionRequest):
(WebKit::UserMediaPermissionRequestManagerProxy::userMediaAccessWasGranted):
(WebKit::UserMediaPermissionRequestManagerProxy::requestUserMediaPermissionForFrame):
(WebKit::UserMediaPermissionRequestManagerProxy::getUserMediaPermissionInfo):
(WebKit::UserMediaPermissionRequestManagerProxy::updateDeviceIDHashSalt):
(WebKit::UserMediaPermissionRequestManagerProxy::enumerateMediaDevicesForFrame):
* UIProcess/UserMediaPermissionRequestManagerProxy.h:
* UIProcess/UserMediaPermissionRequestProxy.cpp:
(WebKit::UserMediaPermissionRequestProxy::UserMediaPermissionRequestProxy):
* UIProcess/UserMediaPermissionRequestProxy.h:
(WebKit::UserMediaPermissionRequestProxy::create):
(WebKit::UserMediaPermissionRequestProxy::deviceIdentifierHashSalt const): Deleted.

LayoutTests:

Update enumerate-devices-source-id to expect the same device to have the same ID in all
iframes, remove enumerate-devices-source-id-persistent because the logic was faulty.

Reviewed by Youenn Fablet.

* TestExpectations: Unskip http/tests/media/media-stream/enumerate-devices-source-id.html
* http/tests/media/media-stream/enumerate-devices-source-id-expected.txt:
* http/tests/media/media-stream/enumerate-devices-source-id-persistent-expected.txt: Removed.
* http/tests/media/media-stream/enumerate-devices-source-id-persistent.html: Removed.
* http/tests/media/media-stream/enumerate-devices-source-id.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237988 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoConsider removing double load for accessing the MetadataTable from LLInt
yusukesuzuki@slowstart.org [Thu, 8 Nov 2018 15:42:10 +0000 (15:42 +0000)]
Consider removing double load for accessing the MetadataTable from LLInt
https://bugs.webkit.org/show_bug.cgi?id=190933

Reviewed by Keith Miller.

This patch removes double load for accesses to MetadataTable from LLInt.
MetadataTable is now specially RefCounted class, which has interesting memory layout.
When refcount becomes 0, MetadataTable asks UnlinkedMetadataTable to destroy itself.

* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::finishCreation):
(JSC::CodeBlock::estimatedSize):
(JSC::CodeBlock::visitChildren):
* bytecode/CodeBlock.h:
(JSC::CodeBlock::metadata):
* bytecode/CodeBlockInlines.h:
(JSC::CodeBlock::forEachValueProfile):
(JSC::CodeBlock::forEachArrayProfile):
(JSC::CodeBlock::forEachArrayAllocationProfile):
(JSC::CodeBlock::forEachObjectAllocationProfile):
(JSC::CodeBlock::forEachLLIntCallLinkInfo):
* bytecode/MetadataTable.cpp:
(JSC::MetadataTable::MetadataTable):
(JSC::MetadataTable::~MetadataTable):
(JSC::MetadataTable::sizeInBytes):
* bytecode/MetadataTable.h:
(JSC::MetadataTable::get):
(JSC::MetadataTable::forEach):
(JSC::MetadataTable::ref const):
(JSC::MetadataTable::deref const):
(JSC::MetadataTable::refCount const):
(JSC::MetadataTable::hasOneRef const):
(JSC::MetadataTable::buffer):
(JSC::MetadataTable::linkingData const):
(JSC::MetadataTable::getImpl):
* bytecode/UnlinkedMetadataTable.h:
(JSC::UnlinkedMetadataTable::buffer const):
* bytecode/UnlinkedMetadataTableInlines.h:
(JSC::UnlinkedMetadataTable::UnlinkedMetadataTable):
(JSC::UnlinkedMetadataTable::~UnlinkedMetadataTable):
(JSC::UnlinkedMetadataTable::addEntry):
(JSC::UnlinkedMetadataTable::sizeInBytes):
(JSC::UnlinkedMetadataTable::finalize):
(JSC::UnlinkedMetadataTable::link):
(JSC::UnlinkedMetadataTable::unlink):
* llint/LowLevelInterpreter.asm:
* llint/LowLevelInterpreter32_64.asm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237987 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoPlain text drag in contenteditable is always DragOperationCopy, never DragOperationMove
commit-queue@webkit.org [Thu, 8 Nov 2018 15:36:05 +0000 (15:36 +0000)]
Plain text drag in contenteditable is always DragOperationCopy, never DragOperationMove
https://bugs.webkit.org/show_bug.cgi?id=191228
<rdar://problem/45786830>

Patch by Jonathan Hammer <jonathan@e3software.com> on 2018-11-08
Reviewed by Wenson Hsieh.

Source/WebCore:

DragController::beginDrag should not call cleanupAfterSystemDrag because
the drag is still in progress even after the call to m_client.beginDrag()
returns. This is in contrast to DragController::doSystemDrag, where the
call to cleanupAfterSystemDrag is appropriate because the drag has
concluded by the time m_client.startDrag() returns.

Test: fast/events/drag-and-drop-move-not-copy.html

* page/DragController.cpp:
(WebCore::DragController::beginDrag):

LayoutTests:

A test that checks that dragging plain text in a contenteditable element results
in the text being moved and not copied. Only applies to WK1. Test skipped on
WK2 because WK2 EventSender does not support drag events. Test also skipped on
AppleWin because drag and drop doesn't work correctly in DRT (according to the
TestExpectations file).

* fast/events/drag-and-drop-move-not-copy-expected.txt: Added.
* fast/events/drag-and-drop-move-not-copy.html: Added.
* platform/win/TestExpectations:
* platform/wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237986 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[BFC][IFC] InlineFormattingContext::Line::alignRuns() should take care of all the...
zalan@apple.com [Thu, 8 Nov 2018 15:20:29 +0000 (15:20 +0000)]
[BFC][IFC] InlineFormattingContext::Line::alignRuns() should take care of all the alignments.
https://bugs.webkit.org/show_bug.cgi?id=191414

Reviewed by Antti Koivisto.

* layout/inlineformatting/Line.cpp:
(WebCore::Layout::InlineFormattingContext::Line::close):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237985 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[GTK] Unreviewed test gardening
aboya@igalia.com [Thu, 8 Nov 2018 11:21:58 +0000 (11:21 +0000)]
[GTK] Unreviewed test gardening
https://bugs.webkit.org/show_bug.cgi?id=191404

* platform/gtk/TestExpectations:
* platform/gtk/fast/block/basic/height-percentage-simple-expected.txt: Added.
* platform/gtk/fast/inline/simple-inline-block-expected.txt: Added.
* platform/gtk/fast/inline/simple-intruding-float1-expected.txt: Added.
* platform/gtk/fast/inline/simple-intruding-floats2-expected.txt: Added.
* platform/gtk/fast/inline/simple-shrink-to-fit-inline-block-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237984 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[WebAuthN] Import CTAP device request/response converters from Chromium
jiewen_tan@apple.com [Thu, 8 Nov 2018 10:29:58 +0000 (10:29 +0000)]
[WebAuthN] Import CTAP device request/response converters from Chromium
https://bugs.webkit.org/show_bug.cgi?id=190784
<rdar://problem/45460333>

Reviewed by Brent Fulgham.

Source/WebCore:

This patch imports Chromium's CTAP device request/response converters:
https://fidoalliance.org/specs/fido-v2.0-ps-20170927/fido-client-to-authenticator-protocol-v2.0-ps-20170927.html#message-encoding
1. It directly imports the following files and suit them to WebKit's coding style:
https://cs.chromium.org/chromium/src/device/fido/device_response_converter.cc?l=20&rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
https://cs.chromium.org/chromium/src/device/fido/device_response_converter.h?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
https://cs.chromium.org/chromium/src/device/fido/authenticator_get_info_response.cc?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
https://cs.chromium.org/chromium/src/device/fido/authenticator_get_info_response.h?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
https://cs.chromium.org/chromium/src/device/fido/authenticator_supported_options.cc?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
https://cs.chromium.org/chromium/src/device/fido/authenticator_supported_options.h?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
https://cs.chromium.org/chromium/src/device/fido/ctap_request_unittest.cc?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
https://cs.chromium.org/chromium/src/device/fido/ctap_response_unittest.cc?rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
https://cs.chromium.org/chromium/src/device/fido/fido_test_data.h?l=1&rcl=098dfd90850ffa84c27a884ab75edd2d99c4ec45
2. It gathers the following methods into DeviceRequestConverter:
CtapGetAssertionRequest::EncodeAsCBOR()
CtapMakeCredentialRequest::EncodeAsCBOR()
PublicKeyCredentialDescriptor::ConvertToCBOR()
PublicKeyCredentialParams::ConvertToCBOR()
PublicKeyCredentialRpEntity::ConvertToCBOR()
PublicKeyCredentialUserEntity::ConvertToCBOR()
3. It also apply a patch from Chromium to CBORValue:
https://chromium.googlesource.com/chromium/src/+/7b6fbff35cd8e4d508f08e1cd93b4aa0a0dc402c%5E%21/

Besides importing things from Chromium, it also implements UserVerificationRequirement for both
PublicKeyCredentialCreationOptions and PublicKeyCredentialRequestOptions such that both options
can specify more dimensions of a desired authenticator.

Covered by API tests.

* CMakeLists.txt:
* DerivedSources.make:
* Modules/webauthn/PublicKeyCredentialCreationOptions.h:
(WebCore::PublicKeyCredentialCreationOptions::AuthenticatorSelectionCriteria::encode const):
(WebCore::PublicKeyCredentialCreationOptions::AuthenticatorSelectionCriteria::decode):
* Modules/webauthn/PublicKeyCredentialCreationOptions.idl:
* Modules/webauthn/PublicKeyCredentialRequestOptions.h:
(WebCore::PublicKeyCredentialRequestOptions::encode const):
(WebCore::PublicKeyCredentialRequestOptions::decode):
* Modules/webauthn/PublicKeyCredentialRequestOptions.idl:
* Modules/webauthn/UserVerificationRequirement.h: Copied from Source/WebCore/Modules/webauthn/PublicKeyCredentialRequestOptions.idl.
* Modules/webauthn/UserVerificationRequirement.idl: Copied from Source/WebCore/Modules/webauthn/PublicKeyCredentialRequestOptions.idl.
* Modules/webauthn/cbor/CBORValue.cpp:
(cbor::CBORValue::CBORValue):
(cbor::CBORValue::getBool const):
* Modules/webauthn/cbor/CBORValue.h:
* Modules/webauthn/fido/AuthenticatorGetInfoResponse.cpp: Added.
(fido::toArrayValue):
(fido::AuthenticatorGetInfoResponse::AuthenticatorGetInfoResponse):
(fido::AuthenticatorGetInfoResponse::setMaxMsgSize):
(fido::AuthenticatorGetInfoResponse::setPinProtocols):
(fido::AuthenticatorGetInfoResponse::setExtensions):
(fido::AuthenticatorGetInfoResponse::setOptions):
(fido::encodeAsCBOR):
* Modules/webauthn/fido/AuthenticatorGetInfoResponse.h: Added.
* Modules/webauthn/fido/AuthenticatorSupportedOptions.cpp: Added.
(fido::AuthenticatorSupportedOptions::setSupportsResidentKey):
(fido::AuthenticatorSupportedOptions::setUserVerificationAvailability):
(fido::AuthenticatorSupportedOptions::setUserPresenceRequired):
(fido::AuthenticatorSupportedOptions::setClientPinAvailability):
(fido::AuthenticatorSupportedOptions::setIsPlatformDevice):
(fido::convertToCBOR):
* Modules/webauthn/fido/AuthenticatorSupportedOptions.h: Added.
* Modules/webauthn/fido/DeviceRequestConverter.cpp: Added.
(fido::convertRpEntityToCBOR):
(fido::convertUserEntityToCBOR):
(fido::convertParametersToCBOR):
(fido::convertDescriptorToCBOR):
(fido::encodeMakeCredenitalRequestAsCBOR):
(fido::encodeGetAssertionRequestAsCBOR):
(fido::encodeEmptyAuthenticatorRequest):
* Modules/webauthn/fido/DeviceRequestConverter.h: Copied from Source/WebCore/Modules/webauthn/fido/FidoConstants.h.
* Modules/webauthn/fido/DeviceResponseConverter.cpp: Added.
(fido::convertStringToProtocolVersion):
(fido::getResponseCode):
(fido::getCredentialId):
(fido::readCTAPMakeCredentialResponse):
(fido::readCTAPGetAssertionResponse):
(fido::readCTAPGetInfoResponse):
* Modules/webauthn/fido/DeviceResponseConverter.h: Copied from Source/WebCore/Modules/webauthn/fido/FidoConstants.cpp.
* Modules/webauthn/fido/FidoConstants.cpp:
(fido::isCtapDeviceResponseCode):
(fido::publicKeyCredentialTypeToString):
* Modules/webauthn/fido/FidoConstants.h:
* Sources.txt:
* WebCore.xcodeproj/project.pbxproj:

Tools:

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebCore/CBORValueTest.cpp:
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebCore/CtapRequestTest.cpp: Added.
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebCore/CtapResponseTest.cpp: Added.
(TestWebKitAPI::getTestCorruptedSignResponse):
(TestWebKitAPI::getTestCredentialRawIdBytes):
(TestWebKitAPI::convertToVector):
(TestWebKitAPI::TEST):
* TestWebKitAPI/Tests/WebCore/FidoTestData.h: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237983 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[GTK] Crash when running with sandbox enabled
carlosgc@webkit.org [Thu, 8 Nov 2018 08:28:56 +0000 (08:28 +0000)]
[GTK] Crash when running with sandbox enabled
https://bugs.webkit.org/show_bug.cgi?id=191353

Reviewed by Michael Catanzaro.

g_key_file_set_string_list() expects a null-terminated array as parameter.

* UIProcess/Launcher/glib/BubblewrapLauncher.cpp:
(WebKit::createFlatpakInfo): Append nullptr to sharedPermissions array.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237982 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoCSS Painting API should pass size, arguments and input properties to paint callback
justin_michaud@apple.com [Thu, 8 Nov 2018 05:29:59 +0000 (05:29 +0000)]
CSS Painting API should pass size, arguments and input properties to paint callback
https://bugs.webkit.org/show_bug.cgi?id=191309

Reviewed by Chris Dumez.

Source/WebCore:

Call paint() callback with input properties and arguments. This patch adds a stub for
the CSS Typed OM StylePropertyMapReadOnly, and passes all the arguments as strings without
any syntax checking to the paint callback.

Test: fast/css-custom-paint/properties.html

* CMakeLists.txt:
* DerivedSources.make:
* Sources.txt:
* WebCore.xcodeproj/project.pbxproj:
* bindings/js/JSCSSStyleValueCustom.cpp: Copied from Source/WebCore/css/CSSPaintCallback.h.
(WebCore::toJSNewlyCreated):
(WebCore::toJS):
* bindings/js/WebCoreBuiltinNames.h:
* css/CSSPaintCallback.h:
* css/CSSPaintCallback.idl:
* css/CSSPaintImageValue.cpp:
(WebCore::CSSPaintImageValue::image):
* css/CSSPaintImageValue.h:
* css/CSSPaintSize.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
(WebCore::CSSPaintSize::create):
(WebCore::CSSPaintSize::width const):
(WebCore::CSSPaintSize::height const):
(WebCore::CSSPaintSize::CSSPaintSize):
* css/CSSPaintSize.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
* css/parser/CSSPropertyParserHelpers.cpp:
(WebCore::CSSPropertyParserHelpers::consumeCustomPaint):
* css/typedom/CSSNumericValue.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
* css/typedom/CSSNumericValue.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
* css/typedom/CSSStyleValue.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
(WebCore::CSSStyleValue::isUnitValue):
(WebCore::CSSStyleValue::isUnparsedValue):
* css/typedom/CSSStyleValue.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
* css/typedom/CSSUnitValue.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
* css/typedom/CSSUnitValue.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
* css/typedom/CSSUnparsedValue.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
* css/typedom/CSSUnparsedValue.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
* css/typedom/StylePropertyMapReadOnly.h: Copied from Source/WebCore/css/CSSPaintCallback.h.
(WebCore::StylePropertyMapReadOnly::create):
(WebCore::StylePropertyMapReadOnly::get):
(WebCore::StylePropertyMapReadOnly::StylePropertyMapReadOnly):
* css/typedom/StylePropertyMapReadOnly.idl: Copied from Source/WebCore/css/CSSPaintCallback.idl.
* platform/graphics/CustomPaintImage.cpp:
(WebCore::CustomPaintImage::CustomPaintImage):
(WebCore::CustomPaintImage::doCustomPaint):
* platform/graphics/CustomPaintImage.h:

LayoutTests:

* fast/css-custom-paint/properties-expected.html: Added.
* fast/css-custom-paint/properties.html: Added.
* fast/css-custom-paint/worklet.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237981 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoProvide better Font fallbacks for DirectX backend
bfulgham@apple.com [Thu, 8 Nov 2018 05:15:02 +0000 (05:15 +0000)]
Provide better Font fallbacks for DirectX backend
https://bugs.webkit.org/show_bug.cgi?id=191412
<rdar://problem/45899207>

Reviewed by Zalan Bujtas.

Registration of custom fonts through GDI are not always visible
through the DirectWrite/GDI bridging layer. If a font is not located,
locate the closest matching avialable font rather than failing.

* platform/graphics/FontPlatformData.h:
* platform/graphics/win/FontCustomPlatformData.cpp:
(WebCore::FontCustomPlatformData::fontPlatformData): Use new font
fallback logic.
* platform/graphics/win/FontPlatformDataDirect2D.cpp:
(WebCore::FontPlatformData::platformDataInit):
(WebCore::FontPlatformData::platformIsEqual const):
(WebCore::FontPlatformData::createFallbackFont): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237980 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoClick and touch event listeners on the body don't work
zalan@apple.com [Thu, 8 Nov 2018 03:33:30 +0000 (03:33 +0000)]
Click and touch event listeners on the body don't work
https://bugs.webkit.org/show_bug.cgi?id=191392
<rdar://problem/5844416>

Reviewed by Simon Fraser.

Source/WebCore:

Remove the old quirk of ignoring onclick handlers on the body and beyond.

Test: fast/events/click-handler-on-body-simple.html

* page/ios/FrameIOS.mm:
(WebCore::Frame::nodeRespondingToClickEvents):

LayoutTests:

* fast/events/click-handler-on-body-simple-expected.txt: Added.
* fast/events/click-handler-on-body-simple.html: Added.
* fast/events/touch/touch-handler-on-body-simple-expected.txt: Added.
* fast/events/touch/touch-handler-on-body-simple.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237978 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[Windows][DirectX] Update canvas code to pass more tests
bfulgham@apple.com [Thu, 8 Nov 2018 03:30:36 +0000 (03:30 +0000)]
[Windows][DirectX] Update canvas code to pass more tests
https://bugs.webkit.org/show_bug.cgi?id=191337
<rdar://problem/45878801>

Reviewed by Dean Jackson.

Update the Direct2D code paths to comply with our canvas tests, improving the
the test results scores to 579/770.

PathDirect2D was updated with an implementation of 'addArcTo' based on work by
Dirk Schulze <vbs85@gmx.de> (see https://hg.mozilla.org/mozilla-central/rev/b116b49459f8).

Tests: canvas/philip/tests

* platform/graphics/ImageBuffer.cpp:
(WebCore::ImageBuffer::createCompatibleBuffer): Direct2D needs access to the graphics
context to create the buffer.
* platform/graphics/ImageBuffer.h:
* platform/graphics/Path.h:
* platform/graphics/win/GraphicsContextDirect2D.cpp:
(WebCore::GraphicsContext::drawPattern): Flush needed.
(WebCore::GraphicsContext::drawRect): Ditto.
(WebCore::GraphicsContextPlatformPrivate::setMiterLimit): Correct for Direct2D definition of miter limit.
(WebCore::GraphicsContextPlatformPrivate::strokeStyleProperties const): Added helper function.
(WebCore::GraphicsContextPlatformPrivate::recomputeStrokeStyle): Update for new helper.
(WebCore::GraphicsContext::drawLine): Ditto.
(WebCore::drawWithShadowHelper): Ditto.
(WebCore::GraphicsContext::fillRect): Add flush.
(WebCore::GraphicsContext::platformFillRoundedRect): Update for helper.
(WebCore::GraphicsContext::clipPath): Add flush.
(WebCore::GraphicsContext::strokeRect): Ditto.
(WebCore::GraphicsContext::drawLineForText): Update for upstream changes.
(WebCore::GraphicsContext::drawLinesForText): Ditto.
* platform/graphics/win/GraphicsContextPlatformPrivateDirect2D.h:
* platform/graphics/win/ImageBufferDirect2D.cpp:
(WebCore::createCroppedImageIfNecessary): Add missing implementations.
(WebCore::createBitmapImageAfterScalingIfNeeded): Ditto.
(WebCore::ImageBuffer::copyImage const): Ditto.
(WebCore::ImageBuffer::sinkIntoImage): Ditto.
(WebCore::ImageBuffer::fastCopyImageMode): Ditto.
(WebCore::ImageBuffer::sinkIntoNativeImage): Ditto.
(WebCore::ImageBuffer::copyNativeImage const): Ditto.
* platform/graphics/win/PathDirect2D.cpp:
(WebCore::Path::operator=):
(WebCore::Path::drawDidComplete): This should never have been const.
It manipulates the path!
(WebCore::Path::transform): Properly transform existing geometries.
(WebCore::Path::openFigureAtCurrentPointIfNecessary): Added.
(WebCore::Path::moveTo):
(WebCore::Path::addLineTo): Make sure figure starts at a valid point.
(WebCore::Path::addQuadCurveTo): Ditto.
(WebCore::Path::addBezierCurveTo): Ditto.
(WebCore::Path::addArcTo): Add implementation.
(WebCore::Path::closeSubpath):
(WebCore::drawArcSection):
(WebCore::Path::addArc): Update to build large arcs out of small arc segments. If the
arc is effectively a complete circle, use the ellipse drawing routines.
(WebCore::Path::addRect): Make sure we start at a valid starting point.
(WebCore::Path::addEllipse): Correct for definition of D2D ellipse.
(WebCore::Path::drawDidComplete const): Deleted.
* platform/graphics/win/SimpleFontDataDirect2D.cpp:
(WebCore::Font::platformWidthForGlyph const):
* rendering/svg/RenderSVGResourceClipper.cpp:
(WebCore::RenderSVGResourceClipper::applyClippingToContext):
* rendering/svg/RenderSVGResourceFilter.cpp:
(WebCore::RenderSVGResourceFilter::applyResource):
* rendering/svg/RenderSVGResourceMasker.cpp:
(WebCore::RenderSVGResourceMasker::applyResource):
* rendering/svg/SVGRenderingContext.cpp:
(WebCore::SVGRenderingContext::createImageBuffer):
* rendering/svg/SVGRenderingContext.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237977 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAdd an editing command for creating and inserting child lists
wenson_hsieh@apple.com [Thu, 8 Nov 2018 03:14:37 +0000 (03:14 +0000)]
Add an editing command for creating and inserting child lists
https://bugs.webkit.org/show_bug.cgi?id=191335
<rdar://problem/45814050>

Reviewed by Ryosuke Niwa.

Source/WebCore:

Currently, insertOrderedList and insertUnorderedList only toggle or change list state (i.e. if the selection is
in an ordered or unordered list, reinserting the same list type removes the current list, and inserting a
different list type changes the enclosing list).

However, for certain internal clients (e.g. Mail), if the start of the selection is enclosed by a list item, we
instead create a new list item and insert it after the enclosing list item, and then create a new list within
that list item. Currently, this logic is implemented in Mail for legacy-WebKit-based Mail compose. This patch
brings this logic into WebKit in the form of a new editing command.

Tests: editing/execCommand/insert-nested-lists-in-table.html
       editing/execCommand/insert-nested-lists-with-pre.html
       editing/execCommand/insert-nested-lists.html

* Sources.txt:
* WebCore.xcodeproj/project.pbxproj:
* editing/Editor.cpp:
(WebCore::Editor::insertOrderedList):
(WebCore::Editor::insertUnorderedList):
* editing/EditorCommand.cpp:
(WebCore::executeInsertOrderedList):
(WebCore::executeInsertUnorderedList):
(WebCore::executeInsertNestedUnorderedList):
(WebCore::executeInsertNestedOrderedList):
(WebCore::createCommandMap):
* editing/IndentOutdentCommand.cpp:
(WebCore::IndentOutdentCommand::outdentParagraph):
* editing/InsertListCommand.cpp:
(WebCore::InsertListCommand::doApply):
(WebCore::InsertListCommand::editingAction const):
* editing/InsertListCommand.h:

Change a couple of `enum`s into `enum class`es.

* editing/InsertNestedListCommand.cpp: Added.
(WebCore::InsertNestedListCommand::insertUnorderedList):
(WebCore::InsertNestedListCommand::insertOrderedList):
(WebCore::InsertNestedListCommand::doApply):
* editing/InsertNestedListCommand.h: Added.

Add a new edit command to insert a new list (as a child of any containing list). If the start of the selection
is in a list item, we create a new list item, move the selection into the list item, and increment its list
level; otherwise, simply fall back to inserting a list.

* editing/ModifySelectionListLevel.cpp:
(WebCore::IncreaseSelectionListLevelCommand::doApply):
(WebCore::IncreaseSelectionListLevelCommand::increaseSelectionListLevel):
(WebCore::IncreaseSelectionListLevelCommand::increaseSelectionListLevelOrdered):
(WebCore::IncreaseSelectionListLevelCommand::increaseSelectionListLevelUnordered):
* editing/ModifySelectionListLevel.h:

Expose this constructor, allowing other edit commands to change selection list level as a composite edit
command. Also, change an `enum` into an `enum class`.

(WebCore::IncreaseSelectionListLevelCommand::create):

LayoutTests:

Add a new layout tests that exercise the "InsertNested(Un)orderedList" editing commands in several scenarios
including undo, redo, executing the edit command with a ranged selection, outdenting to decrease list level,
inserting lists in and around tables and table cells, and inserting lists in and around pre elements.

* editing/execCommand/insert-nested-lists-expected.txt: Added.
* editing/execCommand/insert-nested-lists-in-table-expected.txt: Added.
* editing/execCommand/insert-nested-lists-in-table.html: Added.
* editing/execCommand/insert-nested-lists-with-pre-expected.txt: Added.
* editing/execCommand/insert-nested-lists-with-pre.html: Added.
* editing/execCommand/insert-nested-lists.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237976 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoASSERT(renderer()) under HTMLTextAreaElement::updateValue()
cdumez@apple.com [Thu, 8 Nov 2018 03:11:33 +0000 (03:11 +0000)]
ASSERT(renderer()) under HTMLTextAreaElement::updateValue()
https://bugs.webkit.org/show_bug.cgi?id=191391
<rdar://problem/34219633>

Reviewed by Geoffrey Garen.

Source/WebCore:

Update HTMLTextFormControlElement::didEditInnerTextValue() to not call subtreeHasChanged()
if the element has no renderer, similarly to what is already done in
HTMLTextFormControlElement::setRangeText() and HTMLInputElement::setEditingValue().

Test: editing/inserting/inset-html-textarea-without-renderer.html

* html/HTMLTextFormControlElement.cpp:
(WebCore::HTMLTextFormControlElement::didEditInnerTextValue):

LayoutTests:

Add layout test coverage. I have verified that this test passes in Firefox and Chrome.
It used to crash before this change in debug.

* editing/inserting/inset-html-textarea-without-renderer-expected.txt: Added.
* editing/inserting/inset-html-textarea-without-renderer.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237975 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAllow setting RTCRtpTransceiver.direction
youenn@apple.com [Thu, 8 Nov 2018 02:12:32 +0000 (02:12 +0000)]
Allow setting RTCRtpTransceiver.direction
https://bugs.webkit.org/show_bug.cgi?id=191346

Reviewed by Eric Carlson.

LayoutTests/imported/w3c:

* web-platform-tests/webrtc/RTCPeerConnection-remote-track-mute.https-expected.txt:
* web-platform-tests/webrtc/RTCPeerConnection-transceivers.https-expected.txt:
* web-platform-tests/webrtc/RTCRtpTransceiver.https-expected.txt:

Source/WebCore:

Remove readonly from the direction attribute.
Keep setDirection for now as it is in use, with an intent to remove it.
Driven-by fix as in some cases, the rtc source track might be changed
and current implementation is not expecting that.
In such a case, stop observing the old track (which should no longer call us)
before observing the new one.

Covered by rebased tests.

* Modules/mediastream/RTCRtpTransceiver.idl:
* platform/mediastream/RealtimeIncomingAudioSource.cpp:
(WebCore::RealtimeIncomingAudioSource::setSourceTrack):
* platform/mediastream/RealtimeIncomingVideoSource.cpp:
(WebCore::RealtimeIncomingVideoSource::setSourceTrack):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237974 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoREGRESSION(r236705): Web Inspector: wrong text color for selected DataGrid row
nvasilyev@apple.com [Thu, 8 Nov 2018 02:00:49 +0000 (02:00 +0000)]
REGRESSION(r236705): Web Inspector: wrong text color for selected DataGrid row
https://bugs.webkit.org/show_bug.cgi?id=191405
<rdar://problem/45893283>

Reviewed by Matt Baker.

Selection background color for both Dark Mode mode and Light Mode is dark blue.
The text color for both cases should be white.

* UserInterface/Views/DataGrid.css:
(body:not(.window-inactive, .window-docked-inactive) .data-grid:focus tr.editable.selected .cell-content > input):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237973 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[BigInt] Add support to BigInt into ValueAdd
ticaiolima@gmail.com [Thu, 8 Nov 2018 01:47:27 +0000 (01:47 +0000)]
[BigInt] Add support to BigInt into ValueAdd
https://bugs.webkit.org/show_bug.cgi?id=186177

Reviewed by Keith Miller.

JSTests:

* stress/big-int-negate-jit.js:
* stress/value-add-big-int-and-string.js: Added.
* stress/value-add-big-int-prediction-propagation.js: Added.
* stress/value-add-big-int-untyped.js: Added.

PerformanceTests:

The idea of BigIntBench is to provide a set of microbenchmarks and
benchmarks to evaluate how fast BigInt computations are happening on
JSC implementation.

Now, we are adding microbenchmarks in this set,
but the plan is to move these tests to "JSTest/microbenchmarks" when
BigInt is enabled by default. After that, the focus of Bigint bench is
to provide a set of tests that represents real use cases of BigInt in
JS programs.

* BigIntBench/big-int-add-prediction-propagation.js: Added.
* BigIntBench/big-int-simple-add.js: Added.
* BigIntBench/big-int-simple-sub.js: Added.

Source/JavaScriptCore:

We are adding a very primitive specialization case of BigInts into ValueAdd.
When compiling a speculated version of this node to BigInt, we are currently
calling 'operationAddBigInt', a function that expects only BigInts as
parameter and effectly add numbers using JSBigInt::add. To properly
speculate BigInt operands, we changed ArithProfile to observe when
its result is a BigInt. With this new observation, we are able to identify
when ValueAdd results into a String or BigInt.

Here are some numbers for this specialization running
microbenchmarks:

big-int-simple-add                   21.5411+-1.1096  ^  15.3502+-0.7027  ^ definitely 1.4033x faster
big-int-add-prediction-propagation   13.7762+-0.5578  ^  10.8117+-0.5330  ^ definitely 1.2742x faster

* bytecode/ArithProfile.cpp:
(JSC::ArithProfile::emitObserveResult):
(JSC::ArithProfile::shouldEmitSetNonNumeric const):
(JSC::ArithProfile::shouldEmitSetBigInt const):
(JSC::ArithProfile::emitSetNonNumeric const):
(JSC::ArithProfile::emitSetBigInt const):
(WTF::printInternal):
(JSC::ArithProfile::shouldEmitSetNonNumber const): Deleted.
(JSC::ArithProfile::emitSetNonNumber const): Deleted.
* bytecode/ArithProfile.h:
(JSC::ArithProfile::observedUnaryInt):
(JSC::ArithProfile::observedUnaryNumber):
(JSC::ArithProfile::observedBinaryIntInt):
(JSC::ArithProfile::observedBinaryNumberInt):
(JSC::ArithProfile::observedBinaryIntNumber):
(JSC::ArithProfile::observedBinaryNumberNumber):
(JSC::ArithProfile::didObserveNonInt32 const):
(JSC::ArithProfile::didObserveNonNumeric const):
(JSC::ArithProfile::didObserveBigInt const):
(JSC::ArithProfile::setObservedNonNumeric):
(JSC::ArithProfile::setObservedBigInt):
(JSC::ArithProfile::observeResult):
(JSC::ArithProfile::didObserveNonNumber const): Deleted.
(JSC::ArithProfile::setObservedNonNumber): Deleted.
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::makeSafe):
* dfg/DFGFixupPhase.cpp:
(JSC::DFG::FixupPhase::fixupNode):
* dfg/DFGNode.h:
(JSC::DFG::Node::mayHaveNonNumericResult):
(JSC::DFG::Node::mayHaveBigIntResult):
(JSC::DFG::Node::mayHaveNonNumberResult): Deleted.
* dfg/DFGNodeFlags.cpp:
(JSC::DFG::dumpNodeFlags):
* dfg/DFGNodeFlags.h:
* dfg/DFGOperations.cpp:
* dfg/DFGOperations.h:
* dfg/DFGPredictionPropagationPhase.cpp:
* dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::compileValueAdd):
* ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::LowerDFGToB3::compileValueAdd):
* runtime/CommonSlowPaths.cpp:
(JSC::updateArithProfileForUnaryArithOp):
(JSC::updateArithProfileForBinaryArithOp):

Tools:

* Scripts/run-jsc-benchmarks:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237972 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agowebrtc/datachannel/basic-tcp.html will crash with an invalid crash
youenn@apple.com [Thu, 8 Nov 2018 01:34:31 +0000 (01:34 +0000)]
webrtc/datachannel/basic-tcp.html will crash with an invalid crash
https://bugs.webkit.org/show_bug.cgi?id=178285
<rdar://problem/34985374>

Reviewed by Eric Carlson.

Reintroduce change made to libwebrtc and erroneously removed when refreshing libwebrtc.

* Source/webrtc/rtc_base/physicalsocketserver.cc:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237970 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoRELEASE_ASSERT(!m_hardClosedForUserDelete) fails in WebCore::IDBServer::UniqueIDBData...
sihui_liu@apple.com [Thu, 8 Nov 2018 01:31:39 +0000 (01:31 +0000)]
RELEASE_ASSERT(!m_hardClosedForUserDelete) fails in WebCore::IDBServer::UniqueIDBDatabase::invokeOperationAndTransactionTimer
https://bugs.webkit.org/show_bug.cgi?id=191326
<rdar://problem/45769269>

Reviewed by Geoffrey Garen.

Source/WebCore:

UniqueIDBDatabase should ignore incoming requests or operations when it is already marked as
hardClosedForUserDelete.

Test: IndexedDB.IndexedDBUserDelete.

* Modules/indexeddb/server/UniqueIDBDatabase.cpp:
(WebCore::IDBServer::UniqueIDBDatabase::didPerformActivateTransactionInBackingStore):
(WebCore::IDBServer::UniqueIDBDatabase::immediateCloseForUserDelete):
* Modules/indexeddb/server/UniqueIDBDatabaseConnection.cpp:
(WebCore::IDBServer::UniqueIDBDatabaseConnection::establishTransaction):
(WebCore::IDBServer::UniqueIDBDatabaseConnection::deleteTransaction):
* Modules/indexeddb/server/UniqueIDBDatabaseConnection.h:
* Modules/indexeddb/server/UniqueIDBDatabaseTransaction.cpp:
(WebCore::IDBServer::UniqueIDBDatabaseTransaction::commit):

Tools:

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKitCocoa/IndexedDBUserDelete.html: Added.
* TestWebKitAPI/Tests/WebKitCocoa/IndexedDBUserDelete.mm: Added.
(-[IndexedDBUserDeleteMessageHandler userContentController:didReceiveScriptMessage:]):
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237969 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoPositioned text underline can look like a strike-through
mmaxfield@apple.com [Thu, 8 Nov 2018 01:08:33 +0000 (01:08 +0000)]
Positioned text underline can look like a strike-through
https://bugs.webkit.org/show_bug.cgi?id=191341

Reviewed by Simon Fraser.

Source/WebCore:

We should just clamp the value so it can't go above the baseline.

We shouldn't do this at parse time because it's totally reasonable for text-underline-position: under to want
a negative text-underline-offset. Instead, we just do it at used value time.

Test: fast/css3-text/css3-text-decoration/text-underline-negative.html

* style/InlineTextBoxStyle.cpp:
(WebCore::computeUnderlineOffset):

LayoutTests:

* fast/css3-text/css3-text-decoration/text-underline-negative-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-underline-negative.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237955 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago"/api/report" does not check commit time correctly.
dewei_zhu@apple.com [Thu, 8 Nov 2018 00:39:38 +0000 (00:39 +0000)]
"/api/report" does not check commit time correctly.
https://bugs.webkit.org/show_bug.cgi?id=191351

Reviewed by Ryosuke Niwa.

Test result report API does not convert formated time string to timestamp correctly
which result in not checking commit time correctly.

* public/include/report-processor.php: Use 'strtotime' instead of 'floatval'.
Accepts the time delta within 1 seconds.
* server-tests/api-report-tests.js: Added unit tests.
(reportWitMismatchingCommitTime):
(reportWithOneSecondCommitTimeDifference):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237953 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoUnreviewed, fix iOS build with recent SDKs.
cdumez@apple.com [Thu, 8 Nov 2018 00:36:47 +0000 (00:36 +0000)]
Unreviewed, fix iOS build with recent SDKs.

Source/WebCore:

* platform/network/cocoa/ResourceResponseCocoa.mm:
(WebCore::ResourceResponse::platformCertificateInfo const):

Source/WebKit:

* UIProcess/ios/fullscreen/WKFullScreenWindowControllerIOS.mm:
(-[WKFullScreenWindowController _EVOrganizationName]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237952 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoUnreviewed, fix iOS build with recent SDKs.
cdumez@apple.com [Thu, 8 Nov 2018 00:32:21 +0000 (00:32 +0000)]
Unreviewed, fix iOS build with recent SDKs.

* platform/network/cocoa/ResourceResponseCocoa.mm:
(WebCore::ResourceResponse::platformCertificateInfo const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237951 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoCustom test group form should use commit set map before customization as the behavior...
dewei_zhu@apple.com [Thu, 8 Nov 2018 00:30:15 +0000 (00:30 +0000)]
Custom test group form should use commit set map before customization as the behavior of radio buttons.
https://bugs.webkit.org/show_bug.cgi?id=191347

Reviewed by Ryosuke Niwa.

The radio button behavior should always set the same revision while editing the revision input.
That means we should not use the intermediate commit set map but use the commit set map before
"Customize" link is clicked.

* browser-tests/customizable-test-group-form-tests.js: Added a unit test for this bug.
* public/v3/components/customizable-test-group-form.js: Pass uncustomized commit set so that the radio button
behavoir preserves.
(CustomizableTestGroupForm):
(CustomizableTestGroupForm.prototype.setCommitSetMap):
(CustomizableTestGroupForm.prototype.didConstructShadowTree):
(CustomizableTestGroupForm.prototype.render):
(CustomizableTestGroupForm.prototype._renderCustomRevisionTable):
(CustomizableTestGroupForm.prototype._constructTableBodyList):
(CustomizableTestGroupForm.prototype._constructTableRowForCommitsWithoutOwner):
(CustomizableTestGroupForm.prototype._constructTableRowForCommitsWithOwner):
(CustomizableTestGroupForm.prototype._constructRevisionRadioButtons):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237950 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoResource Load Statistics: Enable cap on partitioned cache max age
wilander@apple.com [Thu, 8 Nov 2018 00:15:59 +0000 (00:15 +0000)]
Resource Load Statistics: Enable cap on partitioned cache max age
https://bugs.webkit.org/show_bug.cgi?id=191395
<rdar://problem/45885977>

Reviewed by Chris Dumez.

This change does not bump the network cache revision number.
It's not needed since it was done just a week ago.

* UIProcess/ResourceLoadStatisticsMemoryStore.cpp:
(WebKit::ResourceLoadStatisticsMemoryStore::setCacheMaxAgeCap):
(WebKit::ResourceLoadStatisticsMemoryStore::updateCacheMaxAgeCap):
* UIProcess/ResourceLoadStatisticsMemoryStore.h:
* UIProcess/WebResourceLoadStatisticsStore.cpp:
(WebKit::WebResourceLoadStatisticsStore::didCreateNetworkProcess):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237949 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoDotted underlines that skip descenders are invisible
mmaxfield@apple.com [Thu, 8 Nov 2018 00:10:20 +0000 (00:10 +0000)]
Dotted underlines that skip descenders are invisible
https://bugs.webkit.org/show_bug.cgi?id=191403

Reviewed by Simon Fraser.

Source/WebCore:

Turns out our underline bounding boxes had negative width. When drawing the full
underline that was fine because it was handled by the 2D graphics engine, but when
we try to split up the box into dots, our "for" loop was taking 0 iterations (because
the end was before the start).

Test: fast/css3-text/css3-text-decoration/text-underline-style.html

* platform/graphics/cairo/GraphicsContextCairo.cpp:
(WebCore::GraphicsContext::drawLineForText):
* platform/graphics/cg/GraphicsContextCG.cpp:
(WebCore::GraphicsContext::drawLineForText):
* platform/graphics/win/GraphicsContextDirect2D.cpp:
(WebCore::GraphicsContext::drawLineForText):

LayoutTests:

* fast/css3-text/css3-text-decoration/text-underline-style-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-underline-style.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237948 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoCrash in WebCore::PaymentRequest::canMakePayment when Apple Pay payment method data...
aestes@apple.com [Wed, 7 Nov 2018 23:49:29 +0000 (23:49 +0000)]
Crash in WebCore::PaymentRequest::canMakePayment when Apple Pay payment method data is missing
https://bugs.webkit.org/show_bug.cgi?id=191331

Reviewed by Alexey Proskuryakov.

Source/WebCore:

Apple Pay requires merchants specify an ApplePayRequest (which contains several required
fields) as payment method data when constructing a new PaymentRequest. If the
ApplePayRequest is missing required fields, or is missing entirely, canMakePayment() should
resolve to false.

We would properly resolve to false when an ApplePayRequest was specified with missing
required fields, but we would crash when the ApplePayRequest was missing entirely.

This patch fixes the crash by checking for an empty JSValue before trying to convert it to
an ApplePayRequest struct. Because we stringify ApplePayRequests in the PaymentRequest
constructor then parse them again in canMakePayments, an undefined or null payment method
data stringifies to a null String, which then parses to an empty JSValue.

Added test case to http/tests/paymentrequest/payment-request-canmakepayment-method.https.html.

* Modules/applepay/paymentrequest/ApplePayPaymentHandler.cpp:
(WebCore::ApplePayPaymentHandler::convertData):
* Modules/paymentrequest/PaymentRequest.cpp:
(WebCore::PaymentRequest::canMakePayment):

LayoutTests:

* http/tests/paymentrequest/payment-request-canmakepayment-method.https.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237947 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoRevert 237849: it breaks MotionMark
simon.fraser@apple.com [Wed, 7 Nov 2018 23:02:37 +0000 (23:02 +0000)]
Revert 237849: it breaks MotionMark
https://bugs.webkit.org/show_bug.cgi?id=191398

Reviewed by Simon Fraser.

Source/WebCore:

This change broke painting while the test is running.

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

LayoutTests:

* fast/scrolling/ios/update-scroll-coordinated-status-expected.txt: Removed.
* fast/scrolling/ios/update-scroll-coordinated-status.html: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237946 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[iOS] Support deleting to the start of the line and to the end of the line
dbates@webkit.org [Wed, 7 Nov 2018 22:14:21 +0000 (22:14 +0000)]
[iOS] Support deleting to the start of the line and to the end of the line
https://bugs.webkit.org/show_bug.cgi?id=191389
<rdar://problem/45883972>

Reviewed by Wenson Hsieh.

Add support for the key commands Command + Delete and Control + K to delete to the start
of the line and to delete to the end of the line, respectively.

* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _deleteToStartOfLine]): Added.
(-[WKContentView _deleteToEndOfLine]): Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237945 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoSimplify macros in platform
don.olmstead@sony.com [Wed, 7 Nov 2018 22:12:42 +0000 (22:12 +0000)]
Simplify macros in platform
https://bugs.webkit.org/show_bug.cgi?id=191378

Reviewed by Michael Catanzaro.

No new tests. No change in behavior.

Modifies a few cases where a platform macro could be simplified.

* page/EventHandler.cpp:
* platform/network/DNS.h:
* platform/network/NetworkStateNotifier.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237944 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[iOS] Option + Delete should delete previous word
dbates@webkit.org [Wed, 7 Nov 2018 22:11:59 +0000 (22:11 +0000)]
[iOS] Option + Delete should delete previous word
https://bugs.webkit.org/show_bug.cgi?id=191381
<rdar://problem/23681351>

Reviewed by Wenson Hsieh.

Implement support for pressing Option + Delete to delete the previous word. This makes text editing
in WebKit more closely match the platform conventions.

* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _deleteByWord]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237943 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoSome WK1 repaint tests are flakey
simon.fraser@apple.com [Wed, 7 Nov 2018 21:46:19 +0000 (21:46 +0000)]
Some WK1 repaint tests are flakey
https://bugs.webkit.org/show_bug.cgi?id=190627

Reviewed by Zalan Bujtas.
Source/WebCore:

Repaint tracking in GraphicsLayerCA was sensitive to whether there were already dirty
rects on the layer, since tracking happened after checks against existing dirty rects.
This caused some WK1 repaint tests to be flakey, since there's no guarantee that
no repaints happen between the last layer flush and a test calling startTrackingRepaints().

Fix by moving the repaint tracking to before the checks against existing dirty rects.
This is more similar to how repaint tracking on FrameView works.

* platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::setNeedsDisplayInRect):

LayoutTests:

New baseline.

* compositing/masks/compositing-clip-path-change-no-repaint-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237942 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[WinCairo] Unreviewed layout test gardening.
ross.kirsling@sony.com [Wed, 7 Nov 2018 21:23:34 +0000 (21:23 +0000)]
[WinCairo] Unreviewed layout test gardening.

* platform/wincairo/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237941 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[iOS] WebGL leaks exact GPU type
dino@apple.com [Wed, 7 Nov 2018 21:12:55 +0000 (21:12 +0000)]
[iOS] WebGL leaks exact GPU type
https://bugs.webkit.org/show_bug.cgi?id=191393
Source/WebCore:

Reviewed by Tim Horton.

The fully exposed renderer info shouldn't be necessary
on iOS, where the GPUs and drivers are consistent
enough that people shouldn't need to write code specifically
for them. Reduce the ability to fingerprint by simply
returning "Apple GPU".

The other option would have been to disable the extension,
but I think it might still be useful to know you're on
an iOS device.

Test: fast/canvas/webgl/hide-some-renderer-info.html

* html/canvas/WebGL2RenderingContext.cpp: Return "Apple GPU"
on iOS.
(WebCore::WebGL2RenderingContext::getParameter):
* html/canvas/WebGLRenderingContext.cpp: Ditto.
(WebCore::WebGLRenderingContext::getParameter):

LayoutTests:

<rdar://problem/45788297>

Reviewed by Tim Horton.

A test that checks that when the GPU vendor is Apple we're
returning a generic renderer name.

* fast/canvas/webgl/hide-some-renderer-info-expected.txt: Added.
* fast/canvas/webgl/hide-some-renderer-info.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237940 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoIntersectionObserver GC-dependent tests are flaky
ajuma@chromium.org [Wed, 7 Nov 2018 21:00:34 +0000 (21:00 +0000)]
IntersectionObserver GC-dependent tests are flaky
https://bugs.webkit.org/show_bug.cgi?id=191387

Reviewed by Chris Dumez.

Make these tests more robust by triggering GC repeatedly inside a setInterval(). Having only a single call to
GC is potentially flaky, since that call will early-out and return immediately if a GC is already underway.

* intersection-observer/root-element-deleted.html:
* intersection-observer/target-deleted.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237939 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoVersioning.
kocsen_chung@apple.com [Wed, 7 Nov 2018 20:59:15 +0000 (20:59 +0000)]
Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237938 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoOverride +[UIKeyboard isInHardwareKeyboardMode] in WebKitTestRunner and DumpRenderTree
dbates@webkit.org [Wed, 7 Nov 2018 20:12:49 +0000 (20:12 +0000)]
Override +[UIKeyboard isInHardwareKeyboardMode] in WebKitTestRunner and DumpRenderTree
https://bugs.webkit.org/show_bug.cgi?id=190141

Reviewed by Darin Adler.

Source/WebCore:

Remove the runtime application check for WebKitTestRunner as we no longer need this
now that WebKitTestRunner and DumpRenderTree override +[UIKeyboard isInHardwareKeyboardMode].

* platform/RuntimeApplicationChecks.h:
* platform/cocoa/RuntimeApplicationChecksCocoa.mm:
(WebCore::IOSApplication::isWebKitTestRunner): Deleted.

Source/WebKit:

Directly use +[UIKeyboard isInHardwareKeyboardMode]. We will override this class method
in DumpRenderTree and WebKitTestRunner to always return NO so as to make test runs
deterministic regardless of whether a hardware keyboard is attached.

* Shared/NativeWebKeyboardEvent.h:
* Shared/ios/NativeWebKeyboardEventIOS.mm:
(WebKit::isInHardwareKeyboardMode): Deleted.
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _startAssistingNode:userIsInteracting:blurPreviousNode:changingActivityState:userObject:]):
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::platformEditorState const):

Tools:

Override +[UIKeyboard isInHardwareKeyboardMode] to always return NO so as to make test
runs deterministic regardless of whether a hardware keyboard is attached.

* DumpRenderTree/mac/DumpRenderTree.mm:
(overrideIsInHardwareKeyboardMode): Added.
(prepareConsistentTestingEnvironment):
* TestRunnerShared/spi/UIKitTestSPI.h: Forward declare SPI.
* WebKitTestRunner/ios/TestControllerIOS.mm:
(overrideIsInHardwareKeyboardMode): Added.
(WTR::TestController::platformInitialize):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237935 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoWeb Inspector: Fix "Javascript" => "JavaScript" enum in protocol generated objects
commit-queue@webkit.org [Wed, 7 Nov 2018 19:48:04 +0000 (19:48 +0000)]
Web Inspector: Fix "Javascript" => "JavaScript" enum in protocol generated objects
https://bugs.webkit.org/show_bug.cgi?id=191340

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2018-11-07
Reviewed by Devin Rousso.

Source/JavaScriptCore:

* inspector/ConsoleMessage.cpp:
(Inspector::messageSourceValue):
Use new enum name.

* inspector/scripts/codegen/generator.py:
Correct the casing of "JavaScript".

Source/WebCore:

* inspector/agents/InspectorMemoryAgent.cpp:
(WebCore::InspectorMemoryAgent::collectSample):
Use new enum name.

Source/WebInspectorUI:

* UserInterface/Models/Canvas.js:
(WI.Canvas.fromPayload):
Insert a missing break statement for the WebGPU case.

* UserInterface/Models/MemoryTimelineRecord.js:
(WI.MemoryTimelineRecord.memoryCategoriesFromProtocol):
Use new enum name.

* UserInterface/Protocol/Legacy/10.0/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/10.3/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/11.0/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/11.3/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/12.0/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/8.0/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/9.0/InspectorBackendCommands.js:
* UserInterface/Protocol/Legacy/9.3/InspectorBackendCommands.js:
Update enums with new generator.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237934 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAlign wide opcodes in the instruction stream
tzagallo@apple.com [Wed, 7 Nov 2018 19:38:14 +0000 (19:38 +0000)]
Align wide opcodes in the instruction stream
https://bugs.webkit.org/show_bug.cgi?id=191254

Reviewed by Keith Miller.

Pad the bytecode with nops to ensure that wide opcodes are 4-byte
aligned on platforms that don't like unaligned memory access.

For that, add a new type to represent jump targets, BoundLabel, which
delays computing the offset in case we need to emit nops for padding.
Extra padding is also emitted before op_yield and at the of each
BytecodeWriter fragment, to ensure that the bytecode remains aligned
after the rewriting.

As a side effect, we can longer guarantee that the point immediately
before emitting an opcode is the start of that opcode, since nops
might be emitted in between if the opcode needs to be wide. To fix
that, we only take the offset of opcodes after they have been emitted,
using `m_lastInstruction.offset()`.

* bytecode/BytecodeDumper.h:
(JSC::BytecodeDumper::dumpValue):
* bytecode/BytecodeGeneratorification.cpp:
(JSC::BytecodeGeneratorification::run):
* bytecode/BytecodeList.rb:
* bytecode/BytecodeRewriter.h:
(JSC::BytecodeRewriter::Fragment::align):
(JSC::BytecodeRewriter::insertFragmentBefore):
(JSC::BytecodeRewriter::insertFragmentAfter):
* bytecode/Fits.h:
* bytecode/InstructionStream.h:
(JSC::InstructionStreamWriter::ref):
* bytecode/PreciseJumpTargetsInlines.h:
(JSC::updateStoredJumpTargetsForInstruction):
* bytecompiler/BytecodeGenerator.cpp:
(JSC::Label::setLocation):
(JSC::BoundLabel::target):
(JSC::BoundLabel::saveTarget):
(JSC::BoundLabel::commitTarget):
(JSC::BytecodeGenerator::generate):
(JSC::BytecodeGenerator::recordOpcode):
(JSC::BytecodeGenerator::alignWideOpcode):
(JSC::BytecodeGenerator::emitProfileControlFlow):
(JSC::BytecodeGenerator::emitResolveScope):
(JSC::BytecodeGenerator::emitGetFromScope):
(JSC::BytecodeGenerator::emitPutToScope):
(JSC::BytecodeGenerator::emitGetById):
(JSC::BytecodeGenerator::emitDirectGetById):
(JSC::BytecodeGenerator::emitPutById):
(JSC::BytecodeGenerator::emitDirectPutById):
(JSC::BytecodeGenerator::emitGetByVal):
(JSC::BytecodeGenerator::emitCreateThis):
(JSC::BytecodeGenerator::beginSwitch):
(JSC::BytecodeGenerator::endSwitch):
(JSC::BytecodeGenerator::emitRequireObjectCoercible):
(JSC::BytecodeGenerator::emitYieldPoint):
(JSC::BytecodeGenerator::emitToThis):
(JSC::Label::bind): Deleted.
* bytecompiler/BytecodeGenerator.h:
(JSC::BytecodeGenerator::recordOpcode): Deleted.
* bytecompiler/Label.h:
(JSC::BoundLabel::BoundLabel):
(JSC::BoundLabel::operator int):
(JSC::Label::bind):
* generator/Opcode.rb:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237933 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAdd a pseudoclass that matches img elements that are backed by an attachment
mitz@apple.com [Wed, 7 Nov 2018 19:33:54 +0000 (19:33 +0000)]
Add a pseudoclass that matches img elements that are backed by an attachment
https://bugs.webkit.org/show_bug.cgi?id=191286
<rdar://problem/45823554>

Reviewed by Dean Jackson.

Source/WebCore:

Test: fast/css/has-attachment.html

Added a :has-attachment pseudoclass, enabled only when the attachment element is enabled at
runtime, that matches elements (currently only img elements qualify) that have an attachment
element in their shadow DOM.

* css/CSSSelector.cpp:
(WebCore::CSSSelector::selectorText const): Handle the new PseudoClassHasAttachment value.
* css/CSSSelector.h: Added a new value to the PseudoClassType enum.

* css/SelectorChecker.cpp:
(WebCore::SelectorChecker::checkOne const): Check the new PseudoClassHasAttachment value
  using the new hasAttachment test function.

* css/SelectorCheckerTestFunctions.h:
(WebCore::hasAttachment): Added. Tests if the element is an img element that has an
  attachment.

* css/SelectorPseudoClassAndCompatibilityElementMap.in: Added has-attachment.

* css/parser/CSSParserContext.cpp:
(WebCore::CSSParserContext::CSSParserContext): Initialize new attachmentEnabled member based
  on RuntimeEnabledFeatures.
(WebCore::operator==): Compare new attachmentEnabled member.
* css/parser/CSSParserContext.h:
(WebCore::CSSParserContextHash::hash): Include new attachmentEnabled member.

* css/parser/CSSSelectorParser.cpp:
(WebCore::CSSSelectorParser::consumePseudo): Reject :has-attachment if the attachment
  element is not enabled.

* cssjit/SelectorCompiler.cpp:
(WebCore::SelectorCompiler::addPseudoClassType): Handle PseudoClassHasAttachment.

LayoutTests:

* fast/css/has-attachment-expected.txt: Added.
* fast/css/has-attachment.html: Added.
* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237932 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoUnreviewed, rolling out r237913.
jbedard@apple.com [Wed, 7 Nov 2018 19:25:39 +0000 (19:25 +0000)]
Unreviewed, rolling out r237913.

Breaks running tests on Mac without iOS SDK installed

Reverted changeset:

"webkitpy: Generalize trailing SDK specifier (Part 2)"
https://bugs.webkit.org/show_bug.cgi?id=191275
https://trac.webkit.org/changeset/237913

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237931 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoMake generic MainThreadSharedTimer implementation
don.olmstead@sony.com [Wed, 7 Nov 2018 18:49:12 +0000 (18:49 +0000)]
Make generic MainThreadSharedTimer implementation
https://bugs.webkit.org/show_bug.cgi?id=191327

Reviewed by Michael Catanzaro.

No new tests. No change in behavior.

Moves generic implementation for MainThreadSharedTimerGlib.cpp into
MainThreadSharedTimer.cpp.

* platform/MainThreadSharedTimer.cpp:
(WebCore::MainThreadSharedTimer::MainThreadSharedTimer):
(WebCore::MainThreadSharedTimer::setFireInterval):
(WebCore::MainThreadSharedTimer::stop):
(WebCore::MainThreadSharedTimer::invalidate):
* platform/MainThreadSharedTimer.h:
* platform/SourcesGLib.txt:
* platform/glib/MainThreadSharedTimerGLib.cpp: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237930 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoIntersectionObserverEntry doesn't keep JS wrappers of rects alive
ajuma@chromium.org [Wed, 7 Nov 2018 18:30:30 +0000 (18:30 +0000)]
IntersectionObserverEntry doesn't keep JS wrappers of rects alive
https://bugs.webkit.org/show_bug.cgi?id=191330

Reviewed by Chris Dumez.

Source/WebCore:

Retain wrappers of each rect in an IntersectionObserverEntry as long as the entry's wrapper
is alive, by adding these wrappers as opaque roots.

Test: intersection-observer/intersection-observer-entry-keeps-js-wrappers-of-rects-alive.html

* bindings/js/JSIntersectionObserverEntryCustom.cpp:
(WebCore::JSIntersectionObserverEntry::visitAdditionalChildren):
* dom/DOMRectReadOnly.idl:
* page/IntersectionObserverEntry.h:
(WebCore::IntersectionObserverEntry::rootBounds const): Make this return a raw pointer instead of a RefPtr so that it
can be called in JSIntersectionObserverEntry::visitAdditionalChildren, which can be called from non-main threads.
(WebCore::IntersectionObserverEntry::boundingClientRect const): Ditto.
(WebCore::IntersectionObserverEntry::intersectionRect const): Ditto.

LayoutTests:

Add test coverage.

* intersection-observer/intersection-observer-entry-keeps-js-wrappers-of-rects-alive-expected.txt: Added.
* intersection-observer/intersection-observer-entry-keeps-js-wrappers-of-rects-alive.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237929 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoUnreviewed, update test expectations for fast/events/pointer.
ryanhaddad@apple.com [Wed, 7 Nov 2018 17:30:39 +0000 (17:30 +0000)]
Unreviewed, update test expectations for fast/events/pointer.

* TestExpectations: Skip this directory.
* platform/ios/TestExpectations: Remove skip since it is now skipped globally.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237928 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoLayout Test fast/text/international/khmer-selection.html is crashing
pvollan@apple.com [Wed, 7 Nov 2018 17:04:37 +0000 (17:04 +0000)]
Layout Test fast/text/international/khmer-selection.html is crashing
https://bugs.webkit.org/show_bug.cgi?id=191368

Unreviewed test gardening.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237927 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoLayout Test fast/block/basic/quirk-mode-percent-height.html is failing
pvollan@apple.com [Wed, 7 Nov 2018 16:57:39 +0000 (16:57 +0000)]
Layout Test fast/block/basic/quirk-mode-percent-height.html is failing
https://bugs.webkit.org/show_bug.cgi?id=191366

Unreviewed test gardening.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237926 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoTileController::tileSize() should not have side effects
simon.fraser@apple.com [Wed, 7 Nov 2018 15:21:03 +0000 (15:21 +0000)]
TileController::tileSize() should not have side effects
https://bugs.webkit.org/show_bug.cgi?id=191349

Reviewed by Zalan Bujtas.

Calling TileController::tileSize() would recompute a new tile size and set m_tileSizeLocked,
which caused test failures if logging was enabled when running tests.

* platform/graphics/ca/TileController.cpp:
(WebCore::TileController::tileSize const):
(WebCore::TileController::computeTileSize):
* platform/graphics/ca/TileController.h:
* platform/graphics/ca/TileGrid.cpp:
(WebCore::TileGrid::revalidateTiles):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237925 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoSource/WebKit:
commit-queue@webkit.org [Wed, 7 Nov 2018 14:57:40 +0000 (14:57 +0000)]
Source/WebKit:
Force a gregorian calendar to show for credit card expiration date inputs
(autocomplete='cc-exp'*) regardless of default system settings.
https://bugs.webkit.org/show_bug.cgi?id=191096
rdar://problem/42640256

Patch by Zamiul Haque <zhaque@apple.com> on 2018-11-07
Reviewed by Tim Horton.

Added some plumbing code to expose the calendar identifier of the calendar used by a
presented date picker. Added shouldPresentGregorianCalendar:, to be used for
determining what property values a date input control must present a Gregorian
calendar for.

* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView formInputPeripheral]):
* UIProcess/ios/forms/WKFormInputControl.mm:

Tools:
Force a gregorian calendar to show for credit card expiration date inputs
(autocomplete='cc-exp'*) regardless of default system settings.
https://bugs.webkit.org/show_bug.cgi?id=191096
rdar://problem/42640256

Patch by Zamiul Haque <zhaque@apple.com> on 2018-11-07
Reviewed by Tim Horton.

Implemented simulateForeignDefaultCalendar and calendarType to be used
for changing the default calendar returned by the system (ie. [NSCalendar
currentCalendar]) and getting the calendar type identifier of a presented date
picker.

* DumpRenderTree/ios/UIScriptControllerIOS.mm:
(WTR::UIScriptController::isShowingDataListSuggestions const):
(WTR::UIScriptController::calendarType const):
(WTR::UIScriptController::setDefaultCalendarType):
* DumpRenderTree/mac/UIScriptControllerMac.mm:
(WTR::UIScriptController::calendarType const):
(WTR::UIScriptController::setDefaultCalendarType):
* TestRunnerShared/UIScriptContext/UIScriptController.cpp:
(WTR::UIScriptController::selectionEndGrabberViewRect const):
(WTR::UIScriptController::calendarType const): Deleted.
* WebKitTestRunner/TestController.cpp:
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/UIScriptControllerCocoa.mm:
(WTR::UIScriptController::calendarType const):
* WebKitTestRunner/cocoa/TestControllerCocoa.mm:
(WTR::swizzledCalendar):
(WTR::TestController::getOverriddenCalendarIdentifier const):
(WTR::TestController::setDefaultCalendarType):
(WTR::TestController::cocoaResetStateToConsistentValues):
* WebKitTestRunner/ios/UIScriptControllerIOS.mm:
(WTR::UIScriptController::selectionEndGrabberViewRect const):
(WTR::UIScriptController::calendarType const): Deleted.

LayoutTests:
Created a layout test to ensure that date controls marked as credit card expiry
fields present a Gregorian calendar regardless of default system settings.
https://bugs.webkit.org/show_bug.cgi?id=191096
rdar://problem/42640256

Patch by Zamiul Haque <zhaque@apple.com> on 2018-11-07
Reviewed by Tim Horton.

Added two new methods to UIHelper, helping to facilitate this. calendarType
returns the calendar identifier of the NSCalendar instance used by the presented
date picker and setDefaultCalendarType accepts a calendar identifier as an
argument for changing the default system settings.

* fast/forms/ios/force-gregorian-calendar-for-credit-card-expiry.html:
* platform/win/TestExpectations:
* resources/ui-helper.js:
(window.UIHelper.calendarType):
(window.UIHelper.setDefaultCalendarType):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237924 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoUnreviewed. Fix GTK/WPE build when memfd is not available after r237922.
carlosgc@webkit.org [Wed, 7 Nov 2018 14:49:17 +0000 (14:49 +0000)]
Unreviewed. Fix GTK/WPE build when memfd is not available after r237922.

* Platform/unix/SharedMemoryUnix.cpp:
(WebKit::createSharedMemory):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237923 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[Linux] Use memfd_create when available in SharedMemory implementation
carlosgc@webkit.org [Wed, 7 Nov 2018 14:24:09 +0000 (14:24 +0000)]
[Linux] Use memfd_create when available in SharedMemory implementation
https://bugs.webkit.org/show_bug.cgi?id=189741

Reviewed by Michael Catanzaro.

.:

Add include check for linux/memfd.h header.

* Source/cmake/OptionsCommon.cmake:

Source/WebKit:

If memfd is available, use it instead of shm_open.

* Platform/unix/SharedMemoryUnix.cpp:
(WebKit::createSharedMemory): Helper to create the shared memory, trying first with memfd and falling back to
shm if it's not available.
(WebKit::SharedMemory::create): Use createSharedMemory() helper.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237922 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[EME][GStreamer] Ensure key id buffers are present and simplify lifetime management...
cturner@igalia.com [Wed, 7 Nov 2018 11:50:30 +0000 (11:50 +0000)]
[EME][GStreamer] Ensure key id buffers are present and simplify lifetime management of ClearKey class.
https://bugs.webkit.org/show_bug.cgi?id=191157

Reviewed by Xabier Rodriguez-Calvar.

This is in preparation for moving the clearkey decryptor behind a
new decrypt API in CDMInstance, which will be sent into the
pipeline to handle key management and decryption. This is for a
later patch.

Covered by existing clear key tests in media/encrypted-media.

* platform/graphics/gstreamer/GStreamerCommon.h:
(WebCore::GstMappedBuffer::data const): Add a const data accessor,
since we are now providing operator=='s on const objects of this
class that need const access to the data pointer.
(WebCore::GstMappedBuffer::operator==): Add a swap of the new
equality operator so you don't have to remember to have the
GstBuffer on the RHS of the equality all the time.
(WebCore::operator==): Define an equality operator between Gst
buffers and WebCore's mapped buffers. Gst creates a ref and a
separate read view under the covers in the memcmp call, so we do
not need to map the buffer ourselves.
* platform/graphics/gstreamer/eme/WebKitClearKeyDecryptorGStreamer.cpp:
(webkit_media_clear_key_decrypt_class_init): Remove setup/release
bindings.
(webkit_media_clear_key_decrypt_init): Initialize gcrypt cipher
here once instead of for every buffer to be decrypted.
(webKitMediaClearKeyDecryptorFinalize): And destroy the cipher
context when the decryptor is destroyed.
(webKitMediaClearKeyDecryptorFindAndSetKey): Factor out the key
retrieval and context setting in this method, call it for each
sample.
(webKitMediaClearKeyDecryptorDecrypt): Base key id buffer into
this function, and remove cipher creation / destroy methods.
* platform/graphics/gstreamer/eme/WebKitCommonEncryptionDecryptorGStreamer.cpp:
(webkit_media_common_encryption_decrypt_class_init): Remove
setup/release bindings.
(webkitMediaCommonEncryptionDecryptTransformInPlace): Ensure a key
id is present and pass it to the decrypt class method.
(webKitMediaCommonEncryptionDecryptDefaultSetupCipher): Deleted.
(webKitMediaCommonEncryptionDecryptDefaultReleaseCipher): Deleted.
* platform/graphics/gstreamer/eme/WebKitCommonEncryptionDecryptorGStreamer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237921 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[Cairo] Move state change operations from GraphicsContextCairo to CairoOperations
fred.wang@free.fr [Wed, 7 Nov 2018 11:29:27 +0000 (11:29 +0000)]
[Cairo] Move state change operations from GraphicsContextCairo to CairoOperations
https://bugs.webkit.org/show_bug.cgi?id=179610

Unreviewed build fix.

Patch by Frederic Wang <fwang@igalia.com> on 2018-11-07

* platform/graphics/cairo/GraphicsContextCairo.cpp: Add missing include.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237920 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoREGRESSION(r237547): Test failures on 32-bit JSC since the JIT was disabled
tzagallo@apple.com [Wed, 7 Nov 2018 09:05:15 +0000 (09:05 +0000)]
REGRESSION(r237547): Test failures on 32-bit JSC since the JIT was disabled
https://bugs.webkit.org/show_bug.cgi?id=191184

Reviewed by Saam Barati.

JSTests:

Most tests were failing due to timeouts, since they are too slow to
run on CLoop. The exceptions are:

proxy-get-set-correct-receiver.js: Had to reduce the recursion depth not to overflow on CLoop
dont-crash-on-stack-overflow-when-parsing-builtin.js and
dont-crash-on-stack-overflow-when-parsing-default-constructor.js: had
to change the stack size since CLoop requires it to be page aligned.

* microbenchmarks/array-push-1.js:
* microbenchmarks/array-push-2.js:
* microbenchmarks/elidable-new-object-dag.js:
* microbenchmarks/elidable-new-object-roflcopter.js:
* microbenchmarks/elidable-new-object-tree.js:
* microbenchmarks/getter-richards.js:
* microbenchmarks/sinkable-new-object-dag.js:
* microbenchmarks/string-concat-long-convert.js:
* microbenchmarks/typed-array-get-set-by-val-profiling.js:
* slowMicrobenchmarks/array-push-3.js:
* slowMicrobenchmarks/large-map-iteration-with-additions.js:
* slowMicrobenchmarks/spread-small-array.js:
* slowMicrobenchmarks/undefined-property-access.js:
* stress/activation-sink-default-value-tdz-error.js:
* stress/activation-sink-default-value.js:
* stress/activation-sink-osrexit-default-value-tdz-error.js:
* stress/activation-sink-osrexit-default-value.js:
* stress/activation-sink-osrexit.js:
* stress/activation-sink.js:
* stress/allow-math-ic-b3-code-duplication.js:
* stress/array-push-multiple-int32.js:
* stress/arrowfunction-activation-sink-osrexit-default-value-tdz-error.js:
* stress/arrowfunction-lexical-this-activation-sink-osrexit.js:
* stress/arrowfunction-lexical-this-activation-sink.js:
* stress/dont-crash-on-stack-overflow-when-parsing-builtin.js:
* stress/dont-crash-on-stack-overflow-when-parsing-default-constructor.js:
* stress/elide-new-object-dag-then-exit.js:
* stress/materialize-regexp-cyclic.js:
* stress/new-regex-inline.js:
* stress/op_add.js:
* stress/op_bitand.js:
* stress/op_bitor.js:
* stress/op_bitxor.js:
* stress/op_div-ConstVar.js:
* stress/op_div-VarConst.js:
* stress/op_div-VarVar.js:
* stress/op_lshift-ConstVar.js:
* stress/op_lshift-VarConst.js:
* stress/op_lshift-VarVar.js:
* stress/op_mod-ConstVar.js:
* stress/op_mod-VarConst.js:
* stress/op_mod-VarVar.js:
* stress/op_mul-ConstVar.js:
* stress/op_mul-VarConst.js:
* stress/op_mul-VarVar.js:
* stress/op_rshift-ConstVar.js:
* stress/op_rshift-VarConst.js:
* stress/op_rshift-VarVar.js:
* stress/op_sub-ConstVar.js:
* stress/op_sub-VarConst.js:
* stress/op_sub-VarVar.js:
* stress/op_urshift-ConstVar.js:
* stress/op_urshift-VarConst.js:
* stress/op_urshift-VarVar.js:
* stress/proxy-get-set-correct-receiver.js:
* stress/regress-179562.js:
* stress/rest-parameter-many-arguments.js:
* stress/sampling-profiler-richards.js:
* stress/splay-flash-access-1ms.js:
* stress/tailCallForwardArguments.js:
* stress/typed-array-get-by-val-profiling.js:
* typeProfiler/getter-richards.js:

PerformanceTests:

Skip test when the JIT is disabled, it takes too long to run on CLoop.

* JetStream/cdjs/main.js:

Source/JavaScriptCore:

Fix API test on CLoop: we can only disable the LLInt when the JIT is enabled.

* API/tests/PingPongStackOverflowTest.cpp:
(testPingPongStackOverflow):

LayoutTests:

Skip test when the JIT is disabled since it's too slow to run on CLoop.

* js/script-tests/regress-139548.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237919 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoReenable imported/w3c/web-platform-tests/webrtc/RTCPeerConnection-ontrack.https.html
youenn@apple.com [Wed, 7 Nov 2018 07:43:22 +0000 (07:43 +0000)]
Reenable imported/w3c/web-platform-tests/webrtc/RTCPeerConnection-ontrack.https.html
https://bugs.webkit.org/show_bug.cgi?id=191343

Unreviewed.

LayoutTests/imported/w3c:

* web-platform-tests/webrtc/RTCPeerConnection-ontrack.https-expected.txt:

LayoutTests:

* TestExpectations: Unskipping test.
* platform/ios-wk2/imported/w3c/web-platform-tests/webrtc/RTCPeerConnection-ontrack.https-expected.txt: Removed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237918 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoServiceWorker Inspector: Uncaught Exception: null is not an object (evaluating 'resou...
commit-queue@webkit.org [Wed, 7 Nov 2018 07:32:45 +0000 (07:32 +0000)]
ServiceWorker Inspector: Uncaught Exception: null is not an object (evaluating 'resource.target.addResource')
https://bugs.webkit.org/show_bug.cgi?id=191339

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2018-11-06
Reviewed by Matt Baker.

* workers/service/ServiceWorkerJob.cpp:
(WebCore::ServiceWorkerJob::fetchScriptWithContext):
Use the Service Worker's identifier, not this static but otherwise unknown identifier.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237917 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agosender.replaceTrack() fails with InvalidStateError if the transceiver.direction is...
youenn@apple.com [Wed, 7 Nov 2018 06:05:04 +0000 (06:05 +0000)]
sender.replaceTrack() fails with InvalidStateError if the transceiver.direction is "inactive"
https://bugs.webkit.org/show_bug.cgi?id=191202

Reviewed by Eric Carlson.

LayoutTests/imported/w3c:

Changes made upstreamed.

* web-platform-tests/webrtc/RTCPeerConnection-setRemoteDescription-replaceTrack.https-expected.txt:
* web-platform-tests/webrtc/RTCPeerConnection-setRemoteDescription-replaceTrack.https.html:

Source/WebCore:

Covered by updated test.

* Modules/mediastream/RTCPeerConnection.cpp:
(WebCore::RTCPeerConnection::removeTrack):
Update as per spec, in particular make sure to not stop the sender when removing the track.

LayoutTests:

* TestExpectations: skipping a timing out related test.
It is already timing out but is also flaky.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237916 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoCustomizable test group form should not reset manually edited commit value sometimes.
dewei_zhu@apple.com [Wed, 7 Nov 2018 05:58:15 +0000 (05:58 +0000)]
Customizable test group form should not reset manually edited commit value sometimes.
https://bugs.webkit.org/show_bug.cgi?id=190863

Reviewed by Ryosuke Niwa.

After changing the radio button and manually editing the commit value, commit value should not be reset
while changing the name of the test group.
Add the logic to prompt warning when manually typed commit does not exist.

* browser-tests/customizable-test-group-form-tests.js: Added a unit test for this bug.
* browser-tests/index.html:
* public/v3/components/customizable-test-group-form.js: Should always update commit set as long as
the repository of that row does not have a owner repository.
(CustomizableTestGroupForm.prototype._constructTableRowForCommitsWithoutOwner):
(CustomizableTestGroupForm.prototype._constructTableRowForCommitsWithOwner):
(CustomizableTestGroupForm.prototype._constructRevisionRadioButtons):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237915 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agowebkitpy: Boot a single worker for custom devices
jbedard@apple.com [Wed, 7 Nov 2018 05:06:33 +0000 (05:06 +0000)]
webkitpy: Boot a single worker for custom devices
https://bugs.webkit.org/show_bug.cgi?id=191312
<rdar://problem/45848717>

Rubber-stamped by Aakash Jain.

* Scripts/webkitpy/layout_tests/controllers/manager.py:
(Manager.run): Custom devices only run a hand-full of tests. Booting multiple
simulators is expensive and unneeded for these tests.
* Scripts/webkitpy/port/ios.py:
(IOSPort. child_processes): No reason to memoize this call, and it prevents
overriding the number of child processes when booting custom devices.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237914 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agowebkitpy: Generalize trailing SDK specifier (Part 2)
jbedard@apple.com [Wed, 7 Nov 2018 05:01:30 +0000 (05:01 +0000)]
webkitpy: Generalize trailing SDK specifier (Part 2)
https://bugs.webkit.org/show_bug.cgi?id=191275
<rdar://problem/45703660>

Reviewed by Aakash Jain.

* Scripts/webkitpy/port/ios_device.py:
(IOSDevicePort): Use normalized SDK.
* Scripts/webkitpy/port/ios_simulator.py:
(IOSSimulatorPort): Use normalized SDK.
(IOSSimulatorPort._build_driver_flags): Use platform SDK in build driver flags.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237913 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[WebGPU] Experimental prototype for WebGPURenderPipeline and WebGPUSwapChain
justin_fan@apple.com [Wed, 7 Nov 2018 04:37:04 +0000 (04:37 +0000)]
[WebGPU] Experimental prototype for WebGPURenderPipeline and WebGPUSwapChain
https://bugs.webkit.org/show_bug.cgi?id=191291

Reviewed by Myles Maxfield.

Source/JavaScriptCore:

Properly disable WEBGPU on all non-Metal platforms for now.

* Configurations/FeatureDefines.xcconfig:

Source/WebCore:

Test: webgpu/render-pipelines.html
Also update test webgpu/webgpu-basics.html to create a WebGPURenderPipeline.

Begin implementation of WebGPURenderPipeline and WebGPUSwapChain. A WebGPURenderPipeline is backed by a
GPURenderPipeline, created lazily using the properties of the passed-in WebGPURenderPipelineDescriptor.
On Metal-supported systems, GPURenderPipeline is an interface to a MTLRenderPipelineState.
The MTLRenderPipelineState is created with the WebGPUDevice currently configured on the WebGPURenderingContext.

* CMakeLists.txt:
* Configurations/FeatureDefines.xcconfig:
* DerivedSources.make:
* Modules/webgpu/GPUDevice.cpp:
(WebCore::GPUDevice::createRenderPipeline const):
* Modules/webgpu/GPUDevice.h:
* Modules/webgpu/GPUPipelineDescriptorBase.h:
* Modules/webgpu/GPUPipelineStageDescriptor.h:
* Modules/webgpu/GPURenderPipelineDescriptor.h:
(WebCore::GPURenderPipelineDescriptor::GPURenderPipelineDescriptor):
(WebCore::GPURenderPipelineDescriptor::primitiveTopology):
* Modules/webgpu/WebGPUDevice.cpp:
(WebCore::WebGPUDevice::createRenderPipeline const):
* Modules/webgpu/WebGPUDevice.h:
(WebCore::WebGPUDevice::adapter const):
(WebCore::WebGPUDevice::device const):
* Modules/webgpu/WebGPUDevice.idl:
* Modules/webgpu/WebGPUPipelineDescriptorBase.h:
* Modules/webgpu/WebGPUPipelineDescriptorBase.idl:
* Modules/webgpu/WebGPUPipelineStageDescriptor.h:
* Modules/webgpu/WebGPUPipelineStageDescriptor.idl:
* Modules/webgpu/WebGPURenderPipeline.cpp:
(WebCore::WebGPURenderPipeline::create):
(WebCore::WebGPURenderPipeline::WebGPURenderPipeline):
* Modules/webgpu/WebGPURenderPipeline.h:
* Modules/webgpu/WebGPURenderPipeline.idl:
* Modules/webgpu/WebGPURenderPipelineDescriptor.h:
* Modules/webgpu/WebGPURenderPipelineDescriptor.idl:
* Modules/webgpu/WebGPURenderingContext.cpp:
(WebCore::WebGPURenderingContext::create):
(WebCore::WebGPURenderingContext::WebGPURenderingContext):
* Modules/webgpu/WebGPURenderingContext.h:
* Modules/webgpu/WebGPUShaderModule.h:
(WebCore::WebGPUShaderModule::module const):
* Modules/webgpu/WebGPUShaderStage.h:
* Modules/webgpu/WebGPUShaderStage.idl:
* Modules/webgpu/WebGPUSwapChain.cpp:
(WebCore::WebGPUSwapChain::configure):
(WebCore::WebGPUSwapChain::reshape):
(WebCore::WebGPUSwapChain::markLayerComposited):
* Modules/webgpu/WebGPUSwapChain.h:
(WebCore::WebGPUSwapChain::WebGPUSwapChain):
* Modules/webgpu/WebGPUSwapChain.idl:
* Modules/webgpu/cocoa/GPURenderPipeline.h:
(WebCore::GPURenderPipeline::platformRenderPipeline const):
* Modules/webgpu/cocoa/GPURenderPipelineMetal.mm: Added.
(WebCore::setFunctionsForPipelineDescriptor):
(WebCore::GPURenderPipeline::create):
(WebCore::GPURenderPipeline::GPURenderPipeline):
* Modules/webgpu/cocoa/GPUSwapChain.h:
(WebCore::GPUSwapChain::platformLayer const):
* Modules/webgpu/cocoa/GPUSwapChainMetal.mm:
(WebCore::GPUSwapChain::create):
(WebCore::GPUSwapChain::GPUSwapChain):
(WebCore::GPUSwapChain::setDevice):
(WebCore::GPUSwapChain::reshape):
(WebCore::GPUSwapChain::present):
* Sources.txt:
* SourcesCocoa.txt:
* WebCore.xcodeproj/project.pbxproj:
* bindings/js/WebCoreBuiltinNames.h:

Source/WebCore/PAL:

Properly disable WEBGPU on all non-Metal platforms for now.

* Configurations/FeatureDefines.xcconfig:

Source/WebKit:

Properly disable WEBGPU on all non-Metal platforms for now.

* Configurations/FeatureDefines.xcconfig:

Source/WebKitLegacy/mac:

Properly disable WEBGPU on all non-Metal platforms for now.

* Configurations/FeatureDefines.xcconfig:

Source/WTF:

Properly disable WEBGPU on all non-Metal platforms for now.

* wtf/Platform.h:

Tools:

Properly disable WEBGPU on all non-Metal platforms for now.

* TestWebKitAPI/Configurations/FeatureDefines.xcconfig:

LayoutTests:

* webgpu/js/set-up-webgpu-contexts.js: Added helper script for running different webgpu tests.
(runWebGPUTests):
(async.runWebGPUTestsOnCanvas):
(async.setUpContexts):
* webgpu/render-pipelines-expected.txt: Added.
* webgpu/render-pipelines.html: Added.
* webgpu/shader-modules-expected.txt:
* webgpu/shader-modules.html: Updated to use functions in set-up-webgpu-contexts.js.
* webgpu/webgpu-basics-expected.txt:
* webgpu/webgpu-basics.html: Updated through creating a WebGPURenderPipeline.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237912 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAdd support for sender/receiver getCapabilities
youenn@apple.com [Wed, 7 Nov 2018 04:01:49 +0000 (04:01 +0000)]
Add support for sender/receiver getCapabilities
https://bugs.webkit.org/show_bug.cgi?id=191192

Reviewed by Eric Carlson.

LayoutTests/imported/w3c:

* web-platform-tests/webrtc/RTCRtpCapabilities-helper.js:
(string_appeared_here.validateRtpCapabilities): Change already made upstream.
* web-platform-tests/webrtc/RTCRtpReceiver-getCapabilities-expected.txt:
* web-platform-tests/webrtc/RTCRtpSender-getCapabilities-expected.txt:
* web-platform-tests/webrtc/RTCRtpTransceiver-setCodecPreferences-expected.txt:

Source/WebCore:

Expose sender/receiver RTCRtpCapabilities to JS.
Add corresponding IDL and plumbing down to libwebrtc peer connection factory.
Covered by rebased tests.

* CMakeLists.txt:
* DerivedSources.make:
* Modules/mediastream/PeerConnectionBackend.cpp:
(WebCore::PeerConnectionBackend::receiverCapabilities):
(WebCore::PeerConnectionBackend::senderCapabilities):
* Modules/mediastream/PeerConnectionBackend.h:
* Modules/mediastream/RTCRtpCapabilities.idl: Added.
* Modules/mediastream/RTCRtpReceiver.cpp:
(WebCore::RTCRtpReceiver::getCapabilities):
* Modules/mediastream/RTCRtpReceiver.h:
* Modules/mediastream/RTCRtpReceiver.idl:
* Modules/mediastream/RTCRtpSender.cpp:
(WebCore::RTCRtpSender::getCapabilities):
* Modules/mediastream/RTCRtpSender.h:
* Modules/mediastream/RTCRtpSender.idl:
* Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.cpp:
(WebCore::PeerConnectionBackend::receiverCapabilities):
(WebCore::PeerConnectionBackend::senderCapabilities):
* Sources.txt:
* WebCore.xcodeproj/project.pbxproj:
* platform/mediastream/RTCPMuxPolicy.h:
* platform/mediastream/RTCRtpCapabilities.h: Added.
* platform/mediastream/libwebrtc/LibWebRTCProvider.cpp:
(WebCore::typeFromKind):
(WebCore::fromStdString):
(WebCore::toChannels):
(WebCore::toRTCRtpCapabilities):
(WebCore::LibWebRTCProvider::receiverCapabilities):
(WebCore::LibWebRTCProvider::senderCapabilities):
* platform/mediastream/libwebrtc/LibWebRTCProvider.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237911 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoCalling sender.replaceTrack() twice produces a new transceiver and its corresponding...
youenn@apple.com [Wed, 7 Nov 2018 03:23:59 +0000 (03:23 +0000)]
Calling sender.replaceTrack() twice produces a new transceiver and its corresponding m= section
https://bugs.webkit.org/show_bug.cgi?id=191261

Reviewed by Eric Carlson.

Source/WebCore:

Handle the case of replacing a track in a sender that has no track.
In particular, do not create a new m-section as was implied by plan B implementation.
Instead, set the track directly on the rtc sender.
Covered by webrtc/video-addTransceiver.html.

* Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:
(WebCore::LibWebRTCMediaEndpoint::createSourceAndRTCTrack):
(WebCore::LibWebRTCMediaEndpoint::addTransceiver):
(WebCore::LibWebRTCMediaEndpoint::setSenderSourceFromTrack):
* Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.h:
* Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.cpp:
(WebCore::LibWebRTCPeerConnectionBackend::setSenderSourceFromTrack):
* Modules/mediastream/libwebrtc/LibWebRTCPeerConnectionBackend.h:
* Modules/mediastream/libwebrtc/LibWebRTCRtpSenderBackend.cpp:
(WebCore::LibWebRTCRtpSenderBackend::replaceTrack):

LayoutTests:

* webrtc/video-addTransceiver-expected.txt:
* webrtc/video-addTransceiver.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237910 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoPost too much text to iFrame could crash webkit
cdumez@apple.com [Wed, 7 Nov 2018 03:12:18 +0000 (03:12 +0000)]
Post too much text to iFrame could crash webkit
https://bugs.webkit.org/show_bug.cgi?id=190947
<rdar://problem/45678231>

Reviewed by Geoffrey Garen.

Optimize SuffixTree (Which is used by XSSAuditor) to stop storing each Node's
children as a static array of 128 pointers and use a dynamic array (vector)
instead. This uses way less memory. Also make SuffixTree and SuffixTree::Node
as fast allocated for performance. This part of the change is based on the
following Blink change:
- https://chromium.googlesource.com/chromium/src.git/+/6ca590e1c7edaa7c56cac9e3e3c39cf398ca8d4d

Also update the XSSAuditor to construct the SuffixTree lazily since there are
many cases (including the one in this bug) where we were spending a significant
amount of time building the SuffixTree and then never querying it.

* html/parser/XSSAuditor.cpp:
(WebCore::XSSAuditor::init):
(WebCore::XSSAuditor::decodedHTTPBodySuffixTree):
(WebCore::XSSAuditor::isContainedInRequest):
* html/parser/XSSAuditor.h:
* platform/text/SuffixTree.h:
(WebCore::SuffixTree::mightContain):
(WebCore::SuffixTree::Node::Node):
(WebCore::SuffixTree::Node::~Node):
(WebCore::SuffixTree::Node::find):
(WebCore::SuffixTree::Node::end):
(WebCore::SuffixTree::build):
(WebCore::SuffixTree<Codebook>::Node::childAt):
(WebCore::SuffixTree::Node::at): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237909 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoSupport onremovetrack for RTCPeerConnection removed tracks
youenn@apple.com [Wed, 7 Nov 2018 02:31:30 +0000 (02:31 +0000)]
Support onremovetrack for RTCPeerConnection removed tracks
https://bugs.webkit.org/show_bug.cgi?id=191299

Reviewed by Eric Carlson.

LayoutTests/imported/w3c:

* web-platform-tests/webrtc/RTCPeerConnection-setRemoteDescription-tracks.https-expected.txt:

Source/WebCore:

When applying SDP, LibWebRTCMediaEndpoint gets notified of a removed track.
In that case, make sure to remove it from its remote stream(s) so as
to notify the application of the changes.
Work around the receiver missing the list of streams by storing in a map
the list of the remote streams for a given remote track.

Covered by rebased test.

* Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.cpp:
(WebCore::LibWebRTCMediaEndpoint::mediaStreamFromRTCStream):
(WebCore::LibWebRTCMediaEndpoint::removeRemoteTrack):
(WebCore::LibWebRTCMediaEndpoint::removeRemoteStream):
(WebCore::LibWebRTCMediaEndpoint::stop):
* Modules/mediastream/libwebrtc/LibWebRTCMediaEndpoint.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237908 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[ews-app] Add URL configuration for OpenSource EWS Django app
aakash_jain@apple.com [Wed, 7 Nov 2018 02:08:11 +0000 (02:08 +0000)]
[ews-app] Add URL configuration for OpenSource EWS Django app
https://bugs.webkit.org/show_bug.cgi?id=191132

Rubber-stamped by Lucas Forschler.

* BuildSlaveSupport/ews-app/ews-app/urls.py: Added entry for ews.urls.
* BuildSlaveSupport/ews-app/ews/urls.py: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237907 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoREGRESSION (r237878): css-dark-mode/supported-color-schemes.html is failing on Mojave
timothy@apple.com [Wed, 7 Nov 2018 01:59:52 +0000 (01:59 +0000)]
REGRESSION (r237878): css-dark-mode/supported-color-schemes.html is failing on Mojave
https://bugs.webkit.org/show_bug.cgi?id=191338

Reviewed by Dean Jackson.

* page/Page.cpp:
(WebCore::Page::appearanceDidChange): Add back call to didChangeStyleSheetEnvironment().
It was dropped by mistake when I moved code to appearanceDidChange(). This is needed to
update styles for semantic colors and from controls when the media prefers-color-scheme
media query is not used on the page.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237906 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoMake mDNS ICE Candidate an experimental flag again
youenn@apple.com [Wed, 7 Nov 2018 01:57:39 +0000 (01:57 +0000)]
Make mDNS ICE Candidate an experimental flag again
https://bugs.webkit.org/show_bug.cgi?id=191262

Reviewed by Dean Jackson.

Source/WebCore:

Rename mdnsICECandidatesEnabled to webRTCICECandidatesEnabled.
This allows grouping the WebRTC runtime flags.
No change of behavior.

* Modules/mediastream/PeerConnectionBackend.cpp:
(WebCore::PeerConnectionBackend::newICECandidate):
* page/RuntimeEnabledFeatures.h:
(WebCore::RuntimeEnabledFeatures::webRTCMDNSICECandidatesEnabled const):
(WebCore::RuntimeEnabledFeatures::setWebRTCMDNSICECandidatesEnabled):
(WebCore::RuntimeEnabledFeatures::mdnsICECandidatesEnabled const): Deleted.
(WebCore::RuntimeEnabledFeatures::setMDNSICECandidatesEnabled): Deleted.

Source/WebKit:

* Shared/WebPreferences.yaml:
* UIProcess/API/C/WKPreferences.cpp:
(WKPreferencesSetWebRTCMDNSICECandidatesEnabled):
(WKPreferencesGetWebRTCMDNSICECandidatesEnabled):
* WebProcess/InjectedBundle/InjectedBundle.cpp:
(WebKit::InjectedBundle::overrideBoolPreferenceForTestRunner):

Tools:

* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::setWebRTCMDNSICECandidatesEnabled):
(WTR::TestRunner::setMDNSICECandidatesEnabled): Deleted.
* WebKitTestRunner/InjectedBundle/TestRunner.h:

LayoutTests:

* webrtc/datachannel/mdns-ice-candidates.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237905 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[ews-app] Add views for OpenSource EWS Django app
aakash_jain@apple.com [Wed, 7 Nov 2018 01:44:46 +0000 (01:44 +0000)]
[ews-app] Add views for OpenSource EWS Django app
https://bugs.webkit.org/show_bug.cgi?id=191131

Rubber-stamped by Lucas Forschler.

* BuildSlaveSupport/ews-app/ews/views/index.py: Added.
* BuildSlaveSupport/ews-app/ews/views/patch.py:
* BuildSlaveSupport/ews-app/ews/views/results.py: Added.
* BuildSlaveSupport/ews-app/ews/views/statusbubble.py: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237904 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoImplement text-underline-offset and text-decoration-thickness
mmaxfield@apple.com [Wed, 7 Nov 2018 01:22:41 +0000 (01:22 +0000)]
Implement text-underline-offset and text-decoration-thickness
https://bugs.webkit.org/show_bug.cgi?id=190774

Reviewed by Dean Jackson.

Source/WebCore:

Because of our existing infrastructure for text decorations, adding support for these
is fairly simple. This patch updates our existing functions to handle thick & placed
underlines, as well as updating our repaint code to correcly handle repainting them.
It also handles animations.

Tests: animations/text-decoration-thickness.html
       animations/text-underline-offset.html
       fast/css3-text/css3-text-decoration/text-decoration-offset-2.html
       fast/css3-text/css3-text-decoration/text-decoration-offset-3.html
       fast/css3-text/css3-text-decoration/text-decoration-offset-auto-length.html
       fast/css3-text/css3-text-decoration/text-decoration-offset-baseline.html
       fast/css3-text/css3-text-decoration/text-decoration-offset-from-font-auto.html
       fast/css3-text/css3-text-decoration/text-decoration-offset-from-font-length.html
       fast/css3-text/css3-text-decoration/text-decoration-offset-repaint.html
       fast/css3-text/css3-text-decoration/text-decoration-offset-under-auto.html
       fast/css3-text/css3-text-decoration/text-decoration-offset-under-length.html
       fast/css3-text/css3-text-decoration/text-decoration-offset.html
       fast/css3-text/css3-text-decoration/text-decoration-thickness-length.html
       fast/css3-text/css3-text-decoration/text-decoration-thickness-repaint.html

* page/animation/CSSPropertyAnimation.cpp:
(WebCore::blendFunc):
(WebCore::CSSPropertyAnimationWrapperMap::CSSPropertyAnimationWrapperMap):
* platform/graphics/FontMetrics.h:
(WebCore::FontMetrics::underlinePosition const):
(WebCore::FontMetrics::setUnderlinePosition):
(WebCore::FontMetrics::underlineThickness const):
(WebCore::FontMetrics::setUnderlineThickness):
* platform/graphics/cocoa/FontCocoa.mm:
(WebCore::Font::platformInit):
* rendering/InlineFlowBox.cpp:
(WebCore::InlineFlowBox::addToLine):
* rendering/SimpleLineLayout.cpp:
(WebCore::SimpleLineLayout::canUseForStyle):
* rendering/TextDecorationPainter.cpp:
(WebCore::TextDecorationPainter::paintTextDecoration):
* rendering/style/RenderStyle.cpp:
(WebCore::RenderStyle::changeAffectsVisualOverflow const):
* rendering/style/TextDecorationThickness.h:
(WebCore::TextDecorationThickness::resolve const):
* style/InlineTextBoxStyle.cpp:
(WebCore::computeUnderlineOffset):
(WebCore::visualOverflowForDecorations):
* style/InlineTextBoxStyle.h:
(WebCore::textDecorationStrokeThickness): Deleted.

LayoutTests:

* animations/text-decoration-thickness-expected.txt: Added.
* animations/text-decoration-thickness.html: Added.
* animations/text-underline-offset-expected.txt: Added.
* animations/text-underline-offset.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-2-expected-mismatch.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-2.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-3-expected-mismatch.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-3.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-auto-length-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-auto-length.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-baseline-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-baseline.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-expected-mismatch.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-from-font-auto-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-from-font-auto.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-from-font-length-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-from-font-length.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-repaint-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-repaint.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-under-auto-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-under-auto.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-under-length-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset-under-length.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-offset.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-thickness-length-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-thickness-length.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-thickness-repaint-expected.html: Added.
* fast/css3-text/css3-text-decoration/text-decoration-thickness-repaint.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237903 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAdd some basic pointer events tests
dino@apple.com [Wed, 7 Nov 2018 00:46:34 +0000 (00:46 +0000)]
Add some basic pointer events tests
https://bugs.webkit.org/show_bug.cgi?id=191336
<rdar://problem/45858574>

Reviewed by Wenson Hsieh.

Generate some fake touches for pointer events testing.

* fast/events/pointer/ios/drag-gives-pointerdown-pointermove-pointerup-expected.txt: Added.
* fast/events/pointer/ios/drag-gives-pointerdown-pointermove-pointerup.html: Added.
* fast/events/pointer/ios/tap-gives-pointerdown-pointerup-expected.txt: Added.
* fast/events/pointer/ios/tap-gives-pointerdown-pointerup.html: Added.
* platform/ios/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237902 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoDon’t use the main queue to create an XPC connection.
commit-queue@webkit.org [Wed, 7 Nov 2018 00:19:14 +0000 (00:19 +0000)]
Don’t use the main queue to create an XPC connection.
https://bugs.webkit.org/show_bug.cgi?id=191160
<rdar://problem/45736262>

Patch by Suresh Koppisetty <skoppisetty@apple.com> on 2018-11-06
Reviewed by Geoffrey Garen.

Don't use the main queue to create an XPC connection. xpc_connection_set_bootstrap does
dispatch_mach_send_barrier_f on this queue which delays the sending of the subsequent
bootstrap message (sent to launchd for launching a new target process) when the main queue is busy.

Now that the XPC connection runs on the default concurrent queue, errorHandler could be invoked
by any thread. The job of errorHandler is to dispatch errorHandlerImpl to the main thread,
where errorHandlerImpl is responsible for the actual handling of errors if any.

* UIProcess/Launcher/mac/ProcessLauncherMac.mm:
(WebKit::ProcessLauncher::launchProcess):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237901 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[ews-app] Add models for OpenSource EWS Django app
aakash_jain@apple.com [Wed, 7 Nov 2018 00:07:19 +0000 (00:07 +0000)]
[ews-app] Add models for OpenSource EWS Django app
https://bugs.webkit.org/show_bug.cgi?id=191123

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-app/ews/models/__init__.py:
* BuildSlaveSupport/ews-app/ews/models/buildermappings.py: Added.
* BuildSlaveSupport/ews-app/ews/models/builds.py: Added.
* BuildSlaveSupport/ews-app/ews/models/patch.py:
* BuildSlaveSupport/ews-app/ews/models/steps.py: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237900 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoSkip css-dark-mode tests on older macOS platforms
timothy@apple.com [Tue, 6 Nov 2018 23:57:39 +0000 (23:57 +0000)]
Skip css-dark-mode tests on older macOS platforms
https://bugs.webkit.org/show_bug.cgi?id=191334

Reviewed by Brian Burg.

* platform/mac-highsierra/css-dark-mode/prefers-color-scheme-expected.txt: Removed.
* platform/mac-highsierra/css-dark-mode/supported-color-schemes-expected.txt: Removed.
* platform/mac-sierra/css-dark-mode/prefers-color-scheme-expected.txt: Removed.
* platform/mac-sierra/css-dark-mode/supported-color-schemes-expected.txt: Removed.
* platform/mac/TestExpectations: Skip css-dark-mode for Sierra and HighSierra.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237899 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAdd missing iOS availability macro to _WKAutomationSessionConfiguration
bburg@apple.com [Tue, 6 Nov 2018 23:54:00 +0000 (23:54 +0000)]
Add missing iOS availability macro to _WKAutomationSessionConfiguration
https://bugs.webkit.org/show_bug.cgi?id=191320

Reviewed by Tim Horton.

* UIProcess/API/Cocoa/_WKAutomationSessionConfiguration.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237898 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[ews-app] Fix Invalid HTTP_HOST header
aakash_jain@apple.com [Tue, 6 Nov 2018 23:24:12 +0000 (23:24 +0000)]
[ews-app] Fix Invalid HTTP_HOST header
https://bugs.webkit.org/show_bug.cgi?id=191325

Reviewed by Lucas Forschler.

Changing ALLOWED_HOSTS to * as the landing page should be publicly accessible.

* BuildSlaveSupport/ews-app/ews-app/settings.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237897 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[ews-app] Rename EWS_app to ews-app - part 2
aakash_jain@apple.com [Tue, 6 Nov 2018 23:16:38 +0000 (23:16 +0000)]
[ews-app] Rename EWS_app to ews-app - part 2
https://bugs.webkit.org/show_bug.cgi?id=191317

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-app/EWS_app: Removed.
* BuildSlaveSupport/ews-app/ews-app: Moved from BuildSlaveSupport/ews-app/EWS_app.
* BuildSlaveSupport/ews-app/ews-app/settings.py:
* BuildSlaveSupport/ews-app/ews-app/urls.py:
* BuildSlaveSupport/ews-app/ews-app/wsgi.py:
* BuildSlaveSupport/ews-app/manage.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237896 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoResource Load Statistics: Remove cap on partitioned cache max age if it matches a...
wilander@apple.com [Tue, 6 Nov 2018 23:10:36 +0000 (23:10 +0000)]
Resource Load Statistics: Remove cap on partitioned cache max age if it matches a network reload (redirect-only)
https://bugs.webkit.org/show_bug.cgi?id=189760
<rdar://problem/44612242>

Reviewed by Youenn Fablet and Antti Koivisto.

Source/WebCore:

No new tests. Existing test fleshed out.

* platform/network/ResourceResponseBase.cpp:
(WebCore::ResourceResponseBase::isRedirection const):
* platform/network/ResourceResponseBase.h:
(WebCore::ResourceResponseBase::isRedirection const): Deleted.
    Moved to the implementation file so that I can export it without warning.

Source/WebKit:

When a redirect cache entry for a prevalent resource reaches its max
age cap, it is loaded again from the network and the network response
is compared with what's in the cache. If it's a match, the cache entry
is fully accepted. If not, a new capped cache entry is created.

This feature is still off by default.

* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::canUseCachedRedirect const):
    Now checks the new member m_cacheEntryForMaxAgeCapValidation.
(WebKit::NetworkResourceLoader::retrieveCacheEntry):
    New case handled for entry->hasReachedPrevalentResourceAgeCap().
(WebKit::NetworkResourceLoader::validateCacheEntryForMaxAgeCapValidation):
    This is where the new logic compares the incoming redirect with the
    cached one. If they match, an uncapped entry is allowed to be stored.
    If they don't match, a new capped entry will be stored.
(WebKit::NetworkResourceLoader::willSendRedirectedRequest):
    Now calls the new
    NetworkResourceLoader::validateCacheEntryForMaxAgeCapValidation()
    function.
* NetworkProcess/NetworkResourceLoader.h:
    Added the m_cacheEntryForMaxAgeCapValidation member.
* NetworkProcess/cache/NetworkCache.cpp:
(WebKit::NetworkCache::makeUseDecision):
    Restored to previous behavior which means
    UseDecision::NoDueToPrevalentResourceAgeCap was removed.
(WebKit::NetworkCache::Cache::retrieve):
    Restored to previous behavior.
(WebKit::NetworkCache::Cache::storeRedirect):
    Now takes the optional maxAgeCap and caps the entry's max age if it's set.
(WebKit::NetworkCache::hasReachedPrevalentResourceAgeCap): Deleted.
* NetworkProcess/cache/NetworkCache.h:
* NetworkProcess/cache/NetworkCacheEntry.cpp:
(WebKit::NetworkCache::Entry::encodeAsStorageRecord const):
    Now encodes m_maxAgeCap.
(WebKit::NetworkCache::Entry::decodeStorageRecord):
    Now decodes m_maxAgeCap.
(WebKit::NetworkCache::Entry::hasReachedPrevalentResourceAgeCap const):
    Added.
(WebKit::NetworkCache::Entry::capMaxAge):
    Added.
* NetworkProcess/cache/NetworkCacheEntry.h:
    Added the optional m_maxAgeCap.
* NetworkProcess/cache/NetworkCacheSpeculativeLoad.cpp:
(WebKit::NetworkCache::SpeculativeLoad::willSendRedirectedRequest):
    Now gets the optional max age cap from the NetworkStorageSession and sends it
    to Cache::storeRedirect().
* NetworkProcess/cache/NetworkCacheStatistics.cpp:
(WebKit::NetworkCache::cachedEntryReuseFailureToDiagnosticKey):
    Removed handling of UseDecision::NoDueToPrevalentResourceAgeCap.

LayoutTests:

* http/tests/resourceLoadStatistics/cap-cache-max-age-for-prevalent-resource-expected.txt:
* http/tests/resourceLoadStatistics/cap-cache-max-age-for-prevalent-resource.html:
    The test now performs:
    - an initial Fetch request with a permanent redirect to ID 1234. Capped max age.
    - a second Fetch request with a permanent redirect to ID 4321. Capped max age.
    - a third Fetch request with a permanent redirect to and empty ID. Capped max age.
    - a fourth Fetch request with a permanent redirect to and empty ID. Max age cap removed.
    - a fifth Fetch request with a permanent redirect to and empty ID. Regular cache hit.
* http/tests/resourceLoadStatistics/resources/cached-permanent-redirect.php:
    Now returns an HTTP 500 if the request contains validation headers.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237895 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[ews-app] Rename EWS_app to ews-app - part 1
aakash_jain@apple.com [Tue, 6 Nov 2018 23:08:51 +0000 (23:08 +0000)]
[ews-app] Rename EWS_app to ews-app - part 1
https://bugs.webkit.org/show_bug.cgi?id=191317

Reviewed by Lucas Forschler.

* BuildSlaveSupport/EWS_app: Removed.
* BuildSlaveSupport/ews-app: Moved from BuildSlaveSupport/EWS_app.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237894 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoSpelling dots are drawn in the wrong place
mmaxfield@apple.com [Tue, 6 Nov 2018 23:06:37 +0000 (23:06 +0000)]
Spelling dots are drawn in the wrong place
https://bugs.webkit.org/show_bug.cgi?id=190764

Reviewed by Dean Jackson.

Source/WebCore:

- Dots should not be clipped.
- Dots should be horizontally centered.
- Dots should be drawn behind the text.
- Distance from the baseline to the top of the dot should be 11.035% of font size.
- Dot diameter should be 13.247% of the font size.
- Distance between the dots (right side of the left dot to left side of the right dot) should be 9.457% of the font size.
- The "font size" used in these calculations should be clamped so it's 10px <= font size <= 40px.

Tests: editing/spelling/spelling-dots-position-2.html
       editing/spelling/spelling-dots-position-3.html
       editing/spelling/spelling-dots-position.html
       editing/spelling/spelling-dots-repaint.html

* platform/graphics/cocoa/GraphicsContextCocoa.mm:
(WebCore::colorForMarkerLineStyle): Align iOS and macOS implementations.
(WebCore::GraphicsContext::drawDotsForDocumentMarker): Place the dots correctly.
* rendering/InlineFlowBox.cpp:
(WebCore::InlineFlowBox::addToLine): The KnownToHaveNoOverflow flag should be cleared if the element has spelling dots,
    because there is no guarantee the spelling dots will lie inside the layout rect of the element.
(WebCore::InlineFlowBox::addTextBoxVisualOverflow): Update the repaint rects to include splling dot positions.
* rendering/InlineFlowBox.h: Comments should explain why, not say what.
* rendering/InlineTextBox.cpp:
(WebCore::InlineTextBox::paint): Draw the dots behind the text.
(WebCore::InlineTextBox::hasMarkers const): Convenience.
(WebCore::InlineTextBox::paintPlatformDocumentMarkers): Refactor bounds information into a helper function.
(WebCore::InlineTextBox::calculateUnionOfAllDocumentMarkerBounds const): Use for repaint rect calculation.
(WebCore::InlineTextBox::calculateDocumentMarkerBounds const): Place the dots correctly.
(WebCore::InlineTextBox::paintPlatformDocumentMarker): Call the helper method.
(WebCore::InlineTextBox::collectMarkedTextsForDocumentMarkers const):
(WebCore::InlineTextBox::collectMarkedTextsForDocumentMarkers): Deleted.
* rendering/InlineTextBox.h: Declare the helper methods.
* rendering/SimpleLineLayout.cpp: Simple line layout doesn't know how to paint spelling dots, so make the presence of
    spelling dots opt us out of SLL.
(WebCore::SimpleLineLayout::canUseForWithReason):
* rendering/SimpleLineLayoutCoverage.cpp:
(WebCore::SimpleLineLayout::printReason):
* rendering/SimpleLineLayoutCoverage.h: Add a new opt-out reason.

Tools:

Previously, it was impossible for WebKitTestRunner to draw spelling dots. This patch adds support for a header
at the top of test files, of the form <!-- webkit-test-runner [ spellCheckingDots=true ] --> which will cause
dots to be drawn.

* WebKitTestRunner/TestController.cpp:
(WTR::TestController::resetStateToConsistentValues):
(WTR::updateTestOptionsFromTestHeader):
(WTR::TestController::platformResetStateToConsistentValues):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/TestOptions.h:
(WTR::TestOptions::hasSameInitializationOptions const):
* WebKitTestRunner/cocoa/TestControllerCocoa.mm:
(WTR::TestController::cocoaResetStateToConsistentValues):
* WebKitTestRunner/cocoa/TestRunnerWKWebView.h:
* WebKitTestRunner/ios/TestControllerIOS.mm:
(WTR::TestController::platformResetStateToConsistentValues):
* WebKitTestRunner/mac/TestControllerMac.mm:
(WTR::TestController::platformResetStateToConsistentValues):

LayoutTests:

* editing/spelling/resources/VerySmallDescentAhem.ttf: Added. In order to test repaint rects,
      this is a version of Ahem with a very small descent, so that the dots don't intersect
      with the text itself.
* editing/spelling/spelling-dots-position-2-expected-mismatch.html: Added.
* editing/spelling/spelling-dots-position-2.html: Added. Make sure dots are painted in the
      correct vertical place.
* editing/spelling/spelling-dots-position-3-expected-mismatch.html: Added.
* editing/spelling/spelling-dots-position-3.html: Added. Make sure dots are painted in the
      correct vertical place.
* editing/spelling/spelling-dots-position-expected.html: Added.
* editing/spelling/spelling-dots-position.html: Added. Make sure dots are not painted in
      the wrong place.
* editing/spelling/spelling-dots-repaint-expected.html: Added.
* editing/spelling/spelling-dots-repaint.html: Added. Test repaint by drawing an element
      with spelling dots, and then deleting the element from the document. The spelling
      dots should be removed too.
* fast/writing-mode/english-bt-text-with-spelling-marker-expected.html:
* fast/writing-mode/english-bt-text-with-spelling-marker.html: Update the test to compensate
      for new spelling dot positions.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237893 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[WinCairo] Unreviewed layout test gardening.
ross.kirsling@sony.com [Tue, 6 Nov 2018 22:49:01 +0000 (22:49 +0000)]
[WinCairo] Unreviewed layout test gardening.

* platform/wincairo/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237891 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoFeature flags not present for Visual Studio when using build-webkit
don.olmstead@sony.com [Tue, 6 Nov 2018 22:10:30 +0000 (22:10 +0000)]
Feature flags not present for Visual Studio when using build-webkit
https://bugs.webkit.org/show_bug.cgi?id=191288

Reviewed by Fujii Hironori.

Passes feature flags to Visual studio builds within build-webkit.
Removes cmakeBasedPortArguments as it is no longer in use.

* Scripts/build-jsc:
* Scripts/build-webkit:
* Scripts/webkitdirs.pm:
(cmakeBasedPortArguments): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237887 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoREGRESSION (r230523): Caps lock indicator not shown in password field
pvollan@apple.com [Tue, 6 Nov 2018 21:58:29 +0000 (21:58 +0000)]
REGRESSION (r230523): Caps lock indicator not shown in password field
https://bugs.webkit.org/show_bug.cgi?id=190056

Reviewed by Ryosuke Niwa.

Source/WebCore:

When WindowServer access is blocked, GetCurrentModifiers() always returns 0. Instead of calling
GetCurrentModifiers(), store the current modifiers from the key event argument in the method
WebKit::WebPage::keyEvent, and use the stored value to detect if Caps lock is on. Additionally,
the modifiers needs to be updated when the window becomes active.

Test: fast/events/detect-caps-lock.html

* Sources.txt:
* platform/PlatformKeyboardEvent.h:
* platform/graphics/FontTaggedSettings.cpp:
* platform/mac/KeyEventMac.mm:
(WebCore::PlatformKeyboardEvent::currentCapsLockState):
(WebCore::PlatformKeyboardEvent::getCurrentModifierState):
* testing/Internals.cpp:
(WebCore::Internals::capsLockIsOn):
* testing/Internals.h:
* testing/Internals.idl:

Source/WebKit:

Update cached modifier state in the WebProcess when WebKit::WebPage::keyEvent is called, and
when the window becomes active.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::dispatchActivityStateChange):
(WebKit::WebPageProxy::updateCurrentModifierState):
* UIProcess/WebPageProxy.h:
* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::keyEvent):
(WebKit::WebPage::updateCurrentModifierState):
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

Tools:

* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::toggleCapsLock):
* WebKitTestRunner/InjectedBundle/TestRunner.h:
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
* WebKitTestRunner/cocoa/TestControllerCocoa.mm:
(WTR::TestController::toggleCapsLock):

LayoutTests:

* TestExpectations:
* fast/events/detect-caps-lock-expected.txt: Added.
* fast/events/detect-caps-lock.html: Added.
* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237886 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAdd ObjC SPI equivalent of WKContextHandlesSafeBrowsing
achristensen@apple.com [Tue, 6 Nov 2018 21:53:21 +0000 (21:53 +0000)]
Add ObjC SPI equivalent of WKContextHandlesSafeBrowsing
https://bugs.webkit.org/show_bug.cgi?id=191329

Reviewed by Tim Horton.

Like WKContextHandlesSafeBrowsing, this is so I can make one change in WebKit and enable safe browsing
in WKWebView and disable it in Safari at the same time without broken revisions between landing two changes
in two repositories.

* UIProcess/API/Cocoa/WKWebView.mm:
(+[WKWebView handlesSafeBrowsing]):
* UIProcess/API/Cocoa/WKWebViewPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237885 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoCSS grid elements with justify-content: space-around have extra whitespace, sometimes...
jfernandez@igalia.com [Tue, 6 Nov 2018 21:45:40 +0000 (21:45 +0000)]
CSS grid elements with justify-content: space-around have extra whitespace, sometimes a lot
https://bugs.webkit.org/show_bug.cgi?id=191308

Reviewed by Dean Jackson.

LayoutTests/imported/w3c:

Imported WPT to cover the behavior changes added in this patch.

* resources/import-expectations.json:
* web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-001-expected.txt: Added.
* web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-001.html: Added.
* web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-002-expected.txt: Added.
* web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-002.html: Added.
* web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-003-expected.txt: Added.
* web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-003.html: Added.
* web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-004-expected.txt: Added.
* web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-004.html: Added.
* web-platform-tests/css/css-grid/layout-algorithm/w3c-import.log:

Source/WebCore:

The CSS WG resolved [1] that Content Alignment should account to the
track sizing algorithm.

The sizing algorithm has been modified so that two new steps (1.5
and 2.5) were added to compute the Content Alignment offsets after
resolving the columns' and rows' sizes respectively.

This change decouples the Content Alignment logic from the tracks
position, so that we can use it as part of the track sizing algorithm.

I also had to store the whole ContentAlignmentData structure in two
class attributes. We need both, position and distribution offsets, to
be used in different parts of the layout logic.

[1] https://github.com/w3c/csswg-drafts/issues/2557

Tests: imported/w3c/web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-001.html
       imported/w3c/web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-002.html
       imported/w3c/web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-003.html
       imported/w3c/web-platform-tests/css/css-grid/layout-algorithm/grid-content-distribution-must-account-for-track-sizing-004.html
       imported/w3c/web-platform-tests/css/css-grid/layout-algorithm/grid-percent-cols-filled-shrinkwrap-001.html
       imported/w3c/web-platform-tests/css/css-grid/layout-algorithm/grid-percent-cols-spanned-shrinkwrap-001.html
       imported/w3c/web-platform-tests/css/css-grid/layout-algorithm/grid-percent-rows-filled-shrinkwrap-001.html
       imported/w3c/web-platform-tests/css/css-grid/layout-algorithm/grid-percent-rows-spanned-shrinkwrap-001.html

* rendering/GridTrackSizingAlgorithm.cpp:
(WebCore::GridTrackSizingAlgorithm::gridAreaBreadthForChild const):
* rendering/RenderGrid.cpp:
(WebCore::RenderGrid::repeatTracksSizingIfNeeded):
(WebCore::RenderGrid::layoutBlock):
(WebCore::RenderGrid::gridItemOffset const):
(WebCore::RenderGrid::trackSizesForComputedStyle const):
(WebCore::RenderGrid::populateGridPositionsForDirection):
(WebCore::RenderGrid::gridAreaBreadthForOutOfFlowChild):
(WebCore::contentDistributionOffset):
(WebCore::RenderGrid::computeContentPositionAndDistributionOffset):
(WebCore::RenderGrid::nonCollapsedTracks const):
* rendering/RenderGrid.h:
(WebCore::ContentAlignmentData::isValid):
(WebCore::ContentAlignmentData::defaultOffsets):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237884 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoREGRESSION: Web Inspector: Popovers drawing incorrectly in Network Tab (clicking...
commit-queue@webkit.org [Tue, 6 Nov 2018 21:34:27 +0000 (21:34 +0000)]
REGRESSION: Web Inspector: Popovers drawing incorrectly in Network Tab (clicking between a few items)
https://bugs.webkit.org/show_bug.cgi?id=191323

Patch by Joseph Pecoraro <pecoraro@apple.com> on 2018-11-06
Reviewed by Devin Rousso.

* UserInterface/Views/Popover.js:
(WI.Popover.prototype._drawBackground):
Save and restore the context state between updates. Something
was affecting the clearRect, making it so that it didn't actually
clear the entire rect each time.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237883 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoIndexedDB: WAL file keeps growing
sihui_liu@apple.com [Tue, 6 Nov 2018 20:56:54 +0000 (20:56 +0000)]
IndexedDB: WAL file keeps growing
https://bugs.webkit.org/show_bug.cgi?id=191294
<rdar://problem/41333493>

Reviewed by Chris Dumez.

Source/WebCore:

When we quit apps, the database connection may not be shut down properly, and WAL file will be retained on disk.
On the next open of database connection, new logs will be appended to the original WAL file, which keeps
increasing size of the WAL file. We should do a manual checkpoint when we open a IndexedDB to make sure previous
log is written to database and truncate WAL file.

Test: IndexedDB.IndexedDBTempFileSize

* platform/sql/SQLiteDatabase.cpp:
(WebCore::SQLiteDatabase::open):

Tools:

* TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
* TestWebKitAPI/Tests/WebKitCocoa/IndexedDBTempFileSize-1.html: Added.
* TestWebKitAPI/Tests/WebKitCocoa/IndexedDBTempFileSize-2.html: Added.
* TestWebKitAPI/Tests/WebKitCocoa/IndexedDBTempFileSize.mm: Added.
(-[IndexedDBFileSizeMessageHandler userContentController:didReceiveScriptMessage:]):
(TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237882 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoAdjust handling of Include paths that need quoting
krollin@apple.com [Tue, 6 Nov 2018 20:54:55 +0000 (20:54 +0000)]
Adjust handling of Include paths that need quoting
https://bugs.webkit.org/show_bug.cgi?id=191314
<rdar://problem/45849143>

Reviewed by Dan Bernstein.

There are several places in the JavaScriptCore Xcode project where the
paths defined in HEADER_SEARCH_PATHS are quoted. That is, the
definitions look like:

    HEADER_SEARCH_PATHS = (
        "\"${BUILT_PRODUCTS_DIR}/DerivedSources/JavaScriptCore\"",
        "\"${BUILT_PRODUCTS_DIR}/LLIntOffsets/${ARCHS}\"",
        "\"$(JAVASCRIPTCORE_FRAMEWORKS_DIR)/JavaScriptCore.framework/PrivateHeaders\"",
        "$(inherited)",
    );

The idea here is presumably to have the resulting $(CPP) command have
-I options where the associated paths are themselves quoted,
protecting against space characters in the paths.

This approach to quote management can break under Xcode 9. If
.xcfilelist files are added to the project, the 'objectVersion' value
in the Xcode project file is changed from 46 to 51. If a project with
objectVersion=51 is presented to Xcode 9 (as can happen when we build
for older OS's), it produces build lines where the quotes are escaped,
thereby becoming part of the path. The build then fails because a
search for a file normally found in a directory called "Foo" will be
looked for in "\"Foo\"", which doesn't exist.

Simply removing the escaped quotes from the HEADER_SEARCH_PATHS
definition doesn't work, leading to paths that need quoting due to
space characters but that don't get this quoting (the part of the path
after the space appears to simply go missing).

Removing the escaped quotes from the HEADER_SEARCH_PATHS and moving
the definitions to the .xcconfig fixes this problem.

* Configurations/ToolExecutable.xcconfig:
* JavaScriptCore.xcodeproj/project.pbxproj:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237881 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoIntersectionObserver doesn't keep target's JS wrapper alive
ajuma@chromium.org [Tue, 6 Nov 2018 20:52:51 +0000 (20:52 +0000)]
IntersectionObserver doesn't keep target's JS wrapper alive
https://bugs.webkit.org/show_bug.cgi?id=190235

Reviewed by Ryosuke Niwa.

Source/WebCore:

Retain JS wrappers of targets in queued entries using a vector of GCReachableRef owned by
IntersectionObserver, which gets cleared after the entries have been delivered.

Make IntersectionObserver::takeRecords return a struct which has both the vector of GCReachableRefs
for targets and the vector of intersection observer entries, so that the GCReachableRefs survive
until the creation of JS wrappers for the entries.

Modify IntersectionObserver::hasPendingActivity to keep the observer alive while it has
entries to deliver.

Tests: intersection-observer/intersection-observer-entry-keeps-js-wrapper-of-target-alive.html
       intersection-observer/intersection-observer-keeps-js-wrapper-of-target-alive.html
       intersection-observer/target-deleted.html

* Sources.txt:
* WebCore.xcodeproj/project.pbxproj:
* bindings/js/JSIntersectionObserverEntryCustom.cpp:
(WebCore::JSIntersectionObserverEntry::visitAdditionalChildren): Keep the target element's wrapper alive while the
IntersectionObserverEntry's wrapper is alive.
* page/IntersectionObserver.cpp:
(WebCore::IntersectionObserver::takeRecords): Change return type to include GCReachableRefs for each record's target, so that
each target can be kept until a JS wrapper is constructed for its IntersectionObserverEntry.
(WebCore::IntersectionObserver::appendQueuedEntry):
(WebCore::IntersectionObserver::notify): Erase GCReachableRefs for targets after delivering the corresponding records.
(WebCore::IntersectionObserver::hasPendingActivity const): Keep the IntersectionObserver alive until queued entries are delivered.
(WebCore::IntersectionObserver::stop):
* page/IntersectionObserver.h:
* page/IntersectionObserver.idl:
* page/IntersectionObserverEntry.h:
(WebCore::IntersectionObserverEntry::target const): Make this return a raw pointer instead of a RefPtr so that it
can be called in JSIntersectionObserverEntry::visitAdditionalChildren, which runs on the GC thread (it's illegal to ref a Node
on a non-main thread).
* page/IntersectionObserverEntry.idl:

LayoutTests:

Add test coverage.

Update test that depends on a target getting GC'd to call takeRecords() since
targets with pending entries are no logner GC'd.

* intersection-observer/intersection-observer-entry-keeps-js-wrapper-of-target-alive-expected.txt: Added.
* intersection-observer/intersection-observer-entry-keeps-js-wrapper-of-target-alive.html: Added.
* intersection-observer/intersection-observer-keeps-js-wrapper-of-target-alive-expected.txt: Added.
* intersection-observer/intersection-observer-keeps-js-wrapper-of-target-alive.html: Added.
* intersection-observer/root-element-deleted.html:
* intersection-observer/target-deleted-expected.txt: Added.
* intersection-observer/target-deleted.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237880 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago[MediaStream] User should not be prompted again after denying getDisplayMedia request
eric.carlson@apple.com [Tue, 6 Nov 2018 20:12:41 +0000 (20:12 +0000)]
[MediaStream] User should not be prompted again after denying getDisplayMedia request
https://bugs.webkit.org/show_bug.cgi?id=191227
<rdar://problem/45784512>

Reviewed by Youenn Fablet.

Source/WebKit:

* UIProcess/UserMediaPermissionRequestManagerProxy.cpp:
(WebKit::UserMediaPermissionRequestManagerProxy::userMediaAccessWasDenied):
(WebKit::UserMediaPermissionRequestManagerProxy::wasRequestDenied):
(WebKit::UserMediaPermissionRequestManagerProxy::requestUserMediaPermissionForFrame):
* UIProcess/UserMediaPermissionRequestManagerProxy.h:

LayoutTests:

* TestWebKitAPI/Tests/WebKitCocoa/GetDisplayMedia.mm:
(-[GetDisplayMediaUIDelegate _webView:requestUserMediaAuthorizationForDevices:url:mainFrameURL:decisionHandler:]):
(TestWebKitAPI::GetDisplayMediaTest::promptForCapture):
(TestWebKitAPI::TEST_F):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237879 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months ago<picture> container doesn't update when prefers-color-scheme media query changes
timothy@apple.com [Tue, 6 Nov 2018 19:53:13 +0000 (19:53 +0000)]
<picture> container doesn't update when prefers-color-scheme media query changes
https://bugs.webkit.org/show_bug.cgi?id=190913
rdar://problem/45608456

Reviewed by Dean Jackson.

Source/WebCore:

Test: css-dark-mode/prefers-color-scheme-picture-element.html

* css/MediaQueryEvaluator.cpp:
(WebCore::isAppearanceDependent): Added.
(WebCore::MediaQueryEvaluator::evaluate const): Keep track of appearanceDependentResults.
* css/MediaQueryEvaluator.h:
* css/StyleResolver.cpp:
(WebCore::StyleResolver::addAppearanceDependentMediaQueryResult): Added.
(WebCore::StyleResolver::hasMediaQueriesAffectedByAppearanceChange const): Added.
* css/StyleResolver.h:
(WebCore::StyleResolver::hasAppearanceDependentMediaQueries const): Added.
* dom/Document.cpp:
(WebCore::Document::evaluateMediaQueryList): Call checkAppearanceDependentPictures.
(WebCore::Document::checkAppearanceDependentPictures): Added.
(WebCore::Document::addAppearanceDependentPicture): Added.
(WebCore::Document::removeAppearanceDependentPicture): Added.
* dom/Document.h:
* html/HTMLImageElement.cpp:
(WebCore::HTMLImageElement::bestFitSourceFromPictureElement): Call addAppearanceDependentPicture.
* html/HTMLPictureElement.cpp:
(WebCore::HTMLPictureElement::~HTMLPictureElement): Call removeAppearanceDependentPicture.
(WebCore::HTMLPictureElement::didMoveToNewDocument): Ditto.
(WebCore::HTMLPictureElement::appearanceChangeAffectedPicture const): Added.
* html/HTMLPictureElement.h:
* page/Page.cpp:
(WebCore::Page::appearanceDidChange): Added.
(WebCore::Page::setUseSystemAppearance): Call appearanceDidChange.
(WebCore::Page::setUseDarkAppearance): Call appearanceDidChange.
* page/Page.h:
* style/StyleScope.cpp:
(WebCore::Style::Scope::evaluateMediaQueriesForAppearanceChange): Added.
* style/StyleScope.h:

LayoutTests:

* css-dark-mode/prefers-color-scheme-picture-element-expected.txt: Added.
* css-dark-mode/prefers-color-scheme-picture-element.html: Added.
* platform/mac-highsierra/css-dark-mode/prefers-color-scheme-picture-element-expected.txt: Added.
* platform/mac-sierra/css-dark-mode/prefers-color-scheme-picture-element-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237878 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoMultiple stress/regexp-compile-oom.js tests are failing on High Sierra Debug and...
msaboff@apple.com [Tue, 6 Nov 2018 19:19:06 +0000 (19:19 +0000)]
Multiple stress/regexp-compile-oom.js tests are failing on High Sierra Debug and Release JSC testers.
https://bugs.webkit.org/show_bug.cgi?id=191271

Reviewed by Saam Barati.

JSTests:

Added more test cases and made all test cases run with the same deeply recursive stack
instead of finding that same point for each test case.

* stress/regexp-compile-oom.js:
(prototype.runTest):
(recurseAndTest):
(testList.push.new.TestAndExpectedException):

Source/JavaScriptCore:

Fixed use of ThrowScope my adding release() calls.  Found a few places where we needed
RETURN_IF_EXCEPTION().  After some code inspections determined that we need to cover the
exception bubbling for String.match() with a global RegExp as well as String.replace()
and String.search().

* runtime/RegExpObjectInlines.h:
(JSC::RegExpObject::matchInline):
(JSC::collectMatches):
* runtime/RegExpPrototype.cpp:
(JSC::regExpProtoFuncSearchFast):
* runtime/StringPrototype.cpp:
(JSC::removeUsingRegExpSearch):
(JSC::replaceUsingRegExpSearch):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237877 268f45cc-cd09-0410-ab3c-d52691b4dbfc

10 months agoFix tests after r237863
achristensen@apple.com [Tue, 6 Nov 2018 19:03:45 +0000 (19:03 +0000)]
Fix tests after r237863
https://bugs.webkit.org/show_bug.cgi?id=191315

Reviewed by Chris Dumez.

Tools:

* TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:
With the timing change of r237863, the WKWebView progressed further than before when loading main.html so it completed before main2.html.
This test was meant to test that main2.html does eventually load, and it doesn't matter if main.html loads before it.  I updated the test
to verify that main2.html eventually loads, making this test less timing dependent.

LayoutTests:

* platform/mac-wk2/http/tests/navigation: Added.
* platform/mac-wk2/http/tests/navigation/keyboard-events-during-provisional-navigation-expected.txt: Copied from LayoutTests/http/tests/navigation/keyboard-events-during-provisional-navigation-expected.txt.
This test is timing related, and I need to update its WK2 expectations every time I enable or disable safe browsing checks during navigation.
See r237149.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@237876 268f45cc-cd09-0410-ab3c-d52691b4dbfc