Timeline
Dec 9, 2018:
- 11:54 PM Changeset in webkit [239025] by
-
- 2 edits in trunk/Tools
[iOS] DragAndDropTests.RespectsExternalSourceFidelityRankings is flaky
https://bugs.webkit.org/show_bug.cgi?id=192524
Reviewed by Ryosuke Niwa.
This test started failing on certain builds of iOS where UIKit now attempts to reveal the selection by scrolling
(see <rdar://problem/28300343>). This API test simulates two drops from external sources, both from the point
(300, 400) to (100, 300) in window coordinates. On these aforementioned builds of iOS, what happens is that the
first drop inserts text that overflows the editable region by an enormous amount, which creates an extremely
wide selection rect and increases the width of the document; revealing this selection causes the web view to
scroll horizontally, such that the second attempt to simulate a drag from (300, 400) to (100, 300) fails, since
these window coordinates no longer correspond to the drop destination element in the document.
To fix this test, we simply mark the drop destination as
overflow: hiddento ensure that we don't try and
scroll horizontally after the first drop.
- TestWebKitAPI/Tests/WebKitCocoa/autofocus-contenteditable.html:
- 9:30 PM Changeset in webkit [239024] by
-
- 21 edits in trunk/Source
Move capture manager from RealtimeMediaSourceCenter to capture factory
https://bugs.webkit.org/show_bug.cgi?id=192542
Reviewed by Eric Carlson.
Source/WebCore:
We should be able to run mock captures in wither UIProcess or WebProcess.
Currently, mock capture is only done in WebProcess.
This patch is a first step towards that goal.
It also simplifies RealtimeMediaSourceCenter implementation by starting to remove virtual methods.
Further refactoring will remove the need to subclass RealtimeMediaSourceCenter.
Instead, remaining virtual methods will become non virtual and their
implementation will become port specific.
Removed a JS internal method that is not longer used to further simplify RealtimeMediaSourceCenter.
No visible change of behavior.
Covered by existing tests.
- platform/mediastream/CaptureDeviceManager.h:
- platform/mediastream/RealtimeMediaSourceCenter.cpp:
(WebCore::RealtimeMediaSourceCenter::singleton):
(WebCore::RealtimeMediaSourceCenter::getMediaStreamDevices):
(WebCore::RealtimeMediaSourceCenter::getDisplayMediaDevices):
(WebCore::RealtimeMediaSourceCenter::getUserMediaDevices):
(WebCore::RealtimeMediaSourceCenter::captureDeviceWithPersistentID):
(WebCore::RealtimeMediaSourceCenter::unsetAudioFactory):
(WebCore::RealtimeMediaSourceCenter::unsetVideoFactory):
(WebCore::RealtimeMediaSourceCenter::unsetDisplayCaptureFactory):
- platform/mediastream/RealtimeMediaSourceCenter.h:
- platform/mediastream/RealtimeMediaSourceFactory.h:
- platform/mediastream/gstreamer/GStreamerAudioCaptureSource.cpp:
- platform/mediastream/gstreamer/GStreamerVideoCaptureSource.cpp:
- platform/mediastream/gstreamer/RealtimeMediaSourceCenterLibWebRTC.cpp:
- platform/mediastream/gstreamer/RealtimeMediaSourceCenterLibWebRTC.h:
- platform/mediastream/mac/CoreAudioCaptureSource.cpp:
(WebCore::CoreAudioCaptureSourceFactory::audioCaptureDeviceManager):
- platform/mediastream/mac/CoreAudioCaptureSource.h:
- platform/mediastream/mac/RealtimeMediaSourceCenterMac.cpp:
- platform/mediastream/mac/RealtimeMediaSourceCenterMac.h:
- platform/mock/MockRealtimeMediaSourceCenter.cpp:
(WebCore::MockRealtimeMediaSourceCenter::setMockRealtimeMediaSourceCenterEnabled):
(WebCore::MockRealtimeMediaSourceCenter::audioFactory):
(WebCore::MockRealtimeMediaSourceCenter::videoFactory):
(WebCore::MockRealtimeMediaSourceCenter::displayCaptureFactory):
- platform/mock/MockRealtimeMediaSourceCenter.h:
(WebCore::MockRealtimeMediaSourceCenter::audioCaptureDeviceManager):
(WebCore::MockRealtimeMediaSourceCenter::videoCaptureDeviceManager):
(WebCore::MockRealtimeMediaSourceCenter::displayCaptureDeviceManager):
- testing/Internals.cpp:
- testing/Internals.h:
- testing/Internals.idl:
Source/WebKit:
Make sure to unregister for video capture as done for other capture types.
Implement manager which should be doing nothing as the manager is used for enumeration of devices
which should only happen in UIProcess.
Given that mock capture is interfering with UserMediaCaptureManager,
switching on/off mock capture will require creating a new process.
This shortcoming should be fixed once mock capture will be fully happening in
the same process as regular capture.
- WebProcess/cocoa/UserMediaCaptureManager.cpp:
(WebKit::UserMediaCaptureManager::~UserMediaCaptureManager):
- WebProcess/cocoa/UserMediaCaptureManager.h:
- 5:56 PM Changeset in webkit [239023] by
-
- 15 edits in trunk/Source
Enable HTTP and HTTPS proxies on iOS and make it a property of the NSURLSession
https://bugs.webkit.org/show_bug.cgi?id=192374
<rdar://problem/46506286>
Reviewed by Alex Christensen.
Source/WebCore/PAL:
- pal/spi/cf/CFNetworkSPI.h:
Remove the now-unused SPI declaration.
Source/WebKit:
This patch makes it so that we can use HTTP/HTTPS proxies on iOS as well.
To enable on iOS, you can do something like:
$ defaults write -g WebKit2HTTPProxy -string "http://localhost:8080"
$ defaults write -g WebKit2HTTPSProxy -string "http://localhost:8080"
This patch also changes the Proxy to be enabled on a per NSURLSession
basis instead of a per process basis.
- NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::initializeNetworkProcess):
- NetworkProcess/NetworkSessionCreationParameters.cpp:
(WebKit::NetworkSessionCreationParameters::privateSessionParameters):
(WebKit::NetworkSessionCreationParameters::encode const):
(WebKit::NetworkSessionCreationParameters::decode):
- NetworkProcess/NetworkSessionCreationParameters.h:
- NetworkProcess/cocoa/NetworkSessionCocoa.mm:
(WebKit::proxyDictionary):
(WebKit::NetworkSessionCocoa::NetworkSessionCocoa):
- NetworkProcess/mac/NetworkProcessMac.mm:
(WebKit::NetworkProcess::platformInitializeNetworkProcess):
(WebKit::overrideSystemProxies): Deleted.
- UIProcess/API/Cocoa/WKWebsiteDataStore.mm:
(-[WKWebsiteDataStore _initWithConfiguration:]):
- UIProcess/API/Cocoa/_WKWebsiteDataStoreConfiguration.h:
- UIProcess/API/Cocoa/_WKWebsiteDataStoreConfiguration.mm:
(-[_WKWebsiteDataStoreConfiguration httpProxy]):
(-[_WKWebsiteDataStoreConfiguration setHTTPProxy:]):
(-[_WKWebsiteDataStoreConfiguration httpsProxy]):
(-[_WKWebsiteDataStoreConfiguration setHTTPSProxy:]):
- UIProcess/Cocoa/WebProcessPoolCocoa.mm:
(WebKit::WebProcessPool::platformInitializeNetworkProcess):
- UIProcess/WebsiteData/Cocoa/WebsiteDataStoreCocoa.mm:
(WebKit::WebsiteDataStore::parameters):
- UIProcess/WebsiteData/WebsiteDataStoreConfiguration.cpp:
(WebKit::WebsiteDataStoreConfiguration::copy):
- UIProcess/WebsiteData/WebsiteDataStoreConfiguration.h:
(WebKit::WebsiteDataStoreConfiguration::httpProxy const):
(WebKit::WebsiteDataStoreConfiguration::setHTTPProxy):
(WebKit::WebsiteDataStoreConfiguration::httpsProxy const):
(WebKit::WebsiteDataStoreConfiguration::setHTTPSProxy):
- 12:17 PM Changeset in webkit [239022] by
-
- 2 edits in trunk/Source/WebKit
Build failure due to missing include of APIWebsiteDataStore.h
https://bugs.webkit.org/show_bug.cgi?id=192541
Reviewed by Youenn Fablet.
- UIProcess/WebsiteData/WebsiteDataStoreConfiguration.cpp: Add missing
include of APIWebsiteDataStore.h
- 8:43 AM Changeset in webkit [239021] by
-
- 2 edits in trunk/Source/WTF
Unreviewed, fix build failure on GCC 8.2, part 2
Add RefCountedArray::assign, and use it instead of operator= internally.
We should have operator=(const RefCountedArray&) since it will be automatically generated
if we do not have correct implementation here.
- wtf/RefCountedArray.h:
(WTF::RefCountedArray::operator=):
(WTF::RefCountedArray::assign):
- 8:09 AM Changeset in webkit [239020] by
-
- 2 edits in trunk/Source/WTF
Unreviewed, fix build failure on GCC 8.2
We remove operator=<PtrTraits> call since it is not necessary.
This is a workaround. It seems that GCC 8.2 fails to parse this specialization.
- wtf/RefCountedArray.h:
- 1:05 AM Changeset in webkit [239019] by
-
- 2 edits in trunk/Source/WebKit
[iOS device] Crash when attempting to call -[_WKAttachment info] for an editable image
https://bugs.webkit.org/show_bug.cgi?id=192538
Reviewed by Tim Horton.
Fix the crash by ensuring that the result of
-renderedDrawingsurvives long enough to be wrapped by a
RetainPtrin-PNGRepresentation. This isn't currently testable, since this codepath isn't supported on the
iOS simulator; instead, I manually verified using MobileAttachments.app.
- UIProcess/ios/WKDrawingView.mm:
(-[WKDrawingView renderedDrawing]):
- 12:00 AM Changeset in webkit [239018] by
-
- 12 edits in trunk
Unreviewed, rolling out r239010.
https://bugs.webkit.org/show_bug.cgi?id=192537
Breaks fast/visual-viewport/tiled-drawing/zoomed-fixed-
scrolling-layers-state.html again (Requested by ap on
#webkit).
Reverted changeset:
"Allow control over child order when adding nodes to the
scrolling tree"
https://bugs.webkit.org/show_bug.cgi?id=176914
https://trac.webkit.org/changeset/239010
Dec 8, 2018:
- 11:09 PM Changeset in webkit [239017] by
-
- 2 edits in trunk/Source/WTF
Fix stray-semicolon warning seen with a new version of clang in Xcode
https://bugs.webkit.org/show_bug.cgi?id=192534
Reviewed by Alexey Proskuryakov.
- wtf/Lock.h: Removed an unneeded semicolon.
- 9:36 PM Changeset in webkit [239016] by
-
- 7 edits2 adds in trunk
[MediaStream] Scaled video frames should be resized in letterbox mode
https://bugs.webkit.org/show_bug.cgi?id=192528
<rdar://problem/46576638>
Reviewed by Darin Adler.
Source/WebCore:
Test: fast/mediastream/resize-letterbox.html
- platform/graphics/cv/ImageTransferSessionVT.mm:
(WebCore::ImageTransferSessionVT::ImageTransferSessionVT): Use letterbox resize mode, not trim.
- platform/mock/MockRealtimeVideoSource.cpp:
(WebCore::MockRealtimeVideoSource::captureSize const): "Capture" at the preset size, not
necessarily at the requested size to be more like a physical camera.
(WebCore::MockRealtimeVideoSource::settingsDidChange):
(WebCore::MockRealtimeVideoSource::drawAnimation):
(WebCore::MockRealtimeVideoSource::drawBoxes):
(WebCore::MockRealtimeVideoSource::drawText):
(WebCore::MockRealtimeVideoSource::generateFrame):
(WebCore::MockRealtimeVideoSource::imageBuffer const):
- platform/mock/MockRealtimeVideoSource.h:
LayoutTests:
- fast/mediastream/resize-letterbox-expected.txt: Added.
- fast/mediastream/resize-letterbox.html: Added.
- 8:11 PM Changeset in webkit [239015] by
-
- 4 edits in trunk/Source
Don't programmatically capitalize safe browsing warning buttons
https://bugs.webkit.org/show_bug.cgi?id=192531
<rdar://problem/46417791>
Reviewed by Darin Adler.
Source/WebCore:
This doesn't work so well in other languages.
Capitalize the source strings in English instead.
- en.lproj/Localizable.strings:
Source/WebKit:
- UIProcess/Cocoa/WKSafeBrowsingWarning.mm:
(makeButton):
- 5:55 PM Changeset in webkit [239014] by
-
- 2 edits in trunk/Source/WebKit
Safe browsing warning should respect text size from settings
https://bugs.webkit.org/show_bug.cgi?id=192205
Reviewed by Tim Horton.
- UIProcess/Cocoa/WKSafeBrowsingWarning.mm:
(fontWithSize):
(makeButton):
(-[WKSafeBrowsingWarning addContent]):
(-[WKSafeBrowsingWarning showDetailsClicked]):
- 3:53 PM Changeset in webkit [239013] by
-
- 3 edits in trunk/Source/JavaScriptCore
Reduce size of PropertySlot and PutPropertySlot.
https://bugs.webkit.org/show_bug.cgi?id=192526
Reviewed by Keith Miller.
With some minor adjustments, we can reduce the size of PropertySlot from 80 bytes
(19 padding bytes) to 64 bytes (3 padding bytes), and PutPropertySlot from 40
bytes (4 padding bytes) to 32 bytes (0 padding bytes but with 6 unused bits).
These measurements are for a 64-bit build.
- runtime/PropertySlot.h:
- runtime/PutPropertySlot.h:
(JSC::PutPropertySlot::PutPropertySlot):
- 2:10 PM Changeset in webkit [239012] by
-
- 2 edits in trunk/Source/WTF
[WTF] Debug build fails due conflicting abort() method
https://bugs.webkit.org/show_bug.cgi?id=192491
Reviewed by Michael Catanzaro.
- wtf/Assertions.h: Use namespaced std::abort() insted of plain abort() to avoid clashes
inside classes which have an ::abort() method, but only when cplusplus is defined to
allow inclusion of the header in plain C sources.
- 1:53 PM Changeset in webkit [239011] by
-
- 2 edits in trunk/Tools
Fix lldb_webkit.py following r238771
(https://bugs.webkit.org/show_bug.cgi?id=190234)
Substitute WTFURLProvider for WebCoreURLProvider in one call site that was inadvertently
missed in r238771.
- lldb/lldb_webkit.py:
(WebCoreDocumentProvider.url):
- 1:08 PM Changeset in webkit [239010] by
-
- 12 edits in trunk
Allow control over child order when adding nodes to the scrolling tree
https://bugs.webkit.org/show_bug.cgi?id=176914
<rdar://problem/46542237>
Source/WebCore:
Patch by Frederic Wang <fwang@igalia.com> on 2018-12-08
Reviewed by Simon Fraser.
Based on an earlier patch by Simon Fraser.
Previously ScrollingCoordinator just allowed nodes to be "attached" with a given parent,
but with no control over sibling order. To allow for correct hit-testing overflow and
frame scrolling nodes, we have to build the scrolling tree in z-order.
This patch adds a 'childIndex' parameter to attachNode() which gives control over
sibling order. For now, RenderLayerCompositor always uses the default 'notFound' value
for childIndex so the current behavior (appending new nodes at the end of child list) is
preserved.
One test marked as flakey, since scrolling tree order is currently dependent on HashSet
traversal order.
- page/scrolling/AsyncScrollingCoordinator.cpp:
(WebCore::AsyncScrollingCoordinator::attachToStateTree):
(WebCore::AsyncScrollingCoordinator::ensureRootStateNodeForFrameView):
- page/scrolling/AsyncScrollingCoordinator.h:
- page/scrolling/ScrollingCoordinator.h:
(WebCore::ScrollingCoordinator::attachToStateTree):
- page/scrolling/ScrollingStateNode.cpp:
(WebCore::ScrollingStateNode::insertChild):
(WebCore::ScrollingStateNode::indexOfChild const):
- page/scrolling/ScrollingStateNode.h:
- page/scrolling/ScrollingStateTree.cpp:
(WebCore::ScrollingStateTree::nodeTypeAndParentMatch const):
(WebCore::ScrollingStateTree::attachNode):
- page/scrolling/ScrollingStateTree.h:
Source/WebKit:
Patch by Frederic Wang <fwang@igalia.com> on 2018-12-08
Reviewed by Simon Fraser.
Based on an earlier patch by Simon Fraser.
- Shared/RemoteLayerTree/RemoteScrollingCoordinatorTransaction.cpp:
(WebKit::RemoteScrollingCoordinatorTransaction::decode):
LayoutTests:
Reviewed by Simon Fraser.
- platform/mac-wk2/TestExpectations: Mark fast/visual-viewport/tiled-drawing/zoomed-fixed-scrolling-layers-state.html
as flakey, which it will be until we attach in z-order.
- 6:57 AM Changeset in webkit [239009] by
-
- 4 edits in trunk/Source/JavaScriptCore
Record right offset with aligned wide instructions
https://bugs.webkit.org/show_bug.cgi?id=192006
Reviewed by Yusuke Suzuki.
Aligning bytecode instructions inserts nops into the instruction stream.
Emitting an instruction did not record the actual start of the instruction with
aligned instructions, but the nop just before the actual instruction. This was
problematic with the StaticPropertyAnalyzer that used the wrong instruction offset.
- bytecode/InstructionStream.h:
(JSC::InstructionStream::MutableRef::clone):
- bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::alignWideOpcode):
(JSC::BytecodeGenerator::emitCreateThis):
(JSC::BytecodeGenerator::emitNewObject):
- generator/Opcode.rb:
- 5:02 AM Changeset in webkit [239008] by
-
- 2 edits in trunk/Tools
[ews-build] check-patch-relevance failure should not mark the overall build as failure
https://bugs.webkit.org/show_bug.cgi?id=192510
Reviewed by Lucas Forschler.
- BuildSlaveSupport/ews-build/steps.py:
(CheckPatchRelevance.start):
