Timeline
Apr 28, 2019:
- 10:17 PM Changeset in webkit [244727] by
-
- 2 edits in trunk/Source/WebCore
[Apple Pay] Increment the API version from 6 to 7
https://bugs.webkit.org/show_bug.cgi?id=197041
<rdar://problem/49986625>
Reviewed by Geoffrey Garen.
- Modules/applepay/PaymentCoordinatorClient.cpp:
(WebCore::PaymentCoordinatorClient::supportsVersion):
- 12:28 PM Changeset in webkit [244726] by
-
- 10 edits in trunk
Fix the watchOS engineering build.
.:
- Source/Makefile: Moved libwebrtc logic into Source/ThirdParty/Makefile.
Source/ThirdParty:
- Makefile: Started building libwebrtc on watchOS.
Source/ThirdParty/libwebrtc:
- Makefile: Set OTHER_OPTIONS to build libwebrtc's boringssl target on watchOS, which is a
dependency for TestWebKitAPI's TCPServer.
Source/WebCore:
- Modules/webgpu/WebGPUComputePassEncoder.cpp: Included Logging.h.
Tools:
- TestWebKitAPI/Tests/WebCore/cocoa/AVFoundationSoftLinkTest.mm:
(TestWebKitAPI::TEST): AVVideoPerformanceMetrics does not exist on watchOS.
- 12:06 PM Changeset in webkit [244725] by
-
- 3 edits in trunk/LayoutTests/imported/w3c
Fix imported/w3c/web-platform-tests/fetch/api/request/request-keepalive.html assert_throws call
https://bugs.webkit.org/show_bug.cgi?id=197338
Reviewed by Darin Adler.
- 12:02 PM Changeset in webkit [244724] by
-
- 3 edits in trunk/Source/WebCore
Remove no longer needed mDNS ICE candidate resolution code
https://bugs.webkit.org/show_bug.cgi?id=197315
Reviewed by Eric Carlson.
No change of behavior.
Removed code is no longer exercised as mDNS resolution happens inside libwebrtc
using the same resolution mechanism as for TURN/STUN server names.
- Modules/mediastream/PeerConnectionBackend.cpp:
(WebCore::PeerConnectionBackend::addIceCandidateSucceeded):
(WebCore::PeerConnectionBackend::addIceCandidateFailed):
- Modules/mediastream/PeerConnectionBackend.h:
Apr 27, 2019:
- 3:45 PM Changeset in webkit [244723] by
-
- 11 edits in trunk/Tools
[WKTR] Move test timeout handling to the UIProcess
https://bugs.webkit.org/show_bug.cgi?id=197333
Reviewed by Darin Adler.
Move test timeout handling in WebKitTestRunner to the UIProcess to play nicely with PSON. Previously,
we'd start the timeout timer in the InjectedBundle, which would fail to account of the time spent in
every WebContent process in the case of swapping.
Also, because of process caching, the timeout timer would sometime fire in a cached process and it
would lead to crashes when firing the timer.
- WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
(WTR::InjectedBundle::didReceiveMessageToPage):
(WTR::InjectedBundle::done):
- WebKitTestRunner/InjectedBundle/InjectedBundle.h:
(WTR::InjectedBundle::shouldDumpPixels const):
- WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::TestRunner):
(WTR::TestRunner::waitUntilDone):
(WTR::TestRunner::setWaitUntilDone):
- WebKitTestRunner/InjectedBundle/TestRunner.h:
- WebKitTestRunner/InjectedBundle/gtk/TestRunnerGtk.cpp:
- WebKitTestRunner/InjectedBundle/mac/TestRunnerMac.mm:
- WebKitTestRunner/InjectedBundle/win/TestRunnerWin.cpp:
- WebKitTestRunner/InjectedBundle/wpe/TestRunnerWPE.cpp:
- WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::TestInvocation):
(WTR::TestInvocation::createTestSettingsDictionary):
(WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
(WTR::TestInvocation::initializeWaitToDumpWatchdogTimerIfNeeded):
(WTR::TestInvocation::invalidateWaitToDumpWatchdogTimer):
(WTR::TestInvocation::waitToDumpWatchdogTimerFired):
(WTR::TestInvocation::setWaitUntilDone):
(WTR::TestInvocation::done):
- WebKitTestRunner/TestInvocation.h:
- 12:31 PM Changeset in webkit [244722] by
-
- 3 edits in trunk/Source/WebCore
Move some Compositing logging to the Layers log channel
https://bugs.webkit.org/show_bug.cgi?id=197345
Reviewed by Sam Weinig.
Make Compositing logging a bit less verbose by moving the GraphicsLayer tree dump
to the Layers log channel. Also log GraphicsLayers after flushing, when we'll have
accurate visible rects.
- platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::updateCoverage):
- rendering/RenderLayerCompositor.cpp:
(WebCore::layersLogEnabled):
(WebCore::RenderLayerCompositor::flushPendingLayerChanges):
(WebCore::RenderLayerCompositor::updateCompositingLayers):
- 10:09 AM Changeset in webkit [244721] by
-
- 4 edits in trunk/Source/WebKit
Improve safety of MachMessage class
https://bugs.webkit.org/show_bug.cgi?id=197323
<rdar://problem/44291920>
Reviewed by Darin Adler.
Improve safety of MachMessage class and clean things up a bit.
- Platform/IPC/mac/ConnectionMac.mm:
(IPC::Connection::sendOutgoingMessage):
- Pass MessageReceiverName / MessageName when constructing the MachMessage rather than setting them afterwards since they never change for a given MachMessage.
- Set header->msgh_id to the right value right away instead of setting it first to inlineBodyMessageID and then later fixing it to be outOfLineBodyMessageID when the body is out of line.
- When messageBodyIsOOL was true, we would call getDescriptorAndSkip() which would advance the pointer by sizeof(mach_msg_port_descriptor_t), even though the descriptor type is mach_msg_ool_descriptor_t. This would not matter in the end because we would not use the messageData pointer after this but still.
- Platform/IPC/mac/MachMessage.cpp:
(IPC::MachMessage::create):
Use fastZeroedMalloc() instead of fastMalloc() for safety, given that this class
has a mach_msg_header_t flexible array member. This is what is recommended by the
mach documentation. It is much safer because it otherwize relies on the user
(Connection::sendOutgoingMessage()) to initialize ALL the message members
correctly. I suspect this was the cause of <rdar://problem/44291920> because
Connection::sendOutgoingMessage() would fail to initialize header->msgh_voucher_port
and the MachMessage destructor would then call mach_msg_destroy(header()), which
would mach_msg_destroy_port(header->msgh_voucher_port).
(IPC::MachMessage::MachMessage):
Pass MessageReceiverName / MessageName when constructing the MachMessage rather
than setting them afterwards since they never change for a given MachMessage.
(IPC::MachMessage::messageSize):
Drop if checks for portDescriptorCount and memoryDescriptorCount since the logic
will do the right thing even if they are 0.
- Platform/IPC/mac/MachMessage.h:
(IPC::MachMessage::header):
(IPC::MachMessage::messageReceiverName const):
(IPC::MachMessage::messageName const):
- 9:17 AM Changeset in webkit [244720] by
-
- 7 edits in tags/Safari-608.1.19.1/Source
Versioning.
- 9:14 AM Changeset in webkit [244719] by
-
- 1 copy in tags/Safari-608.1.19.1
New tag.
- 7:41 AM Changeset in webkit [244718] by
-
- 2 edits in trunk/Source/WebCore
Lookup only looking up the first word in selection
https://bugs.webkit.org/show_bug.cgi?id=197341
<rdar://problem/48221414>
Reviewed by Wenson Hsieh.
Lookup is not testable.
Reveal needs the full range in order to correctly create the item for the popover.
- editing/cocoa/DictionaryLookup.mm:
(WebCore::showPopupOrCreateAnimationController):
