Timeline
Apr 23, 2021:
- 11:11 PM Changeset in webkit [276546] by
-
- 2 edits in trunk/Source/WebCore
Crash in IndentOutdentCommand::outdentParagraph attempting to move a paragraph
https://bugs.webkit.org/show_bug.cgi?id=224909
Patch by Ian Gilbert <iang@apple.com> on 2021-04-23
Reviewed by Ryosuke Niwa.
Added check inside IndentOutdentCommand::outdentParagraph to ensure that inserting a
HTMLBRElement was successful before attempting to move a paragraph to that node.
- editing/IndentOutdentCommand.cpp:
(WebCore::IndentOutdentCommand::outdentParagraph):
- 5:47 PM Changeset in webkit [276545] by
-
- 2 edits in trunk/Source/WebKit
PCM: debug mode should send the second report on a 10 second delay after the first
https://bugs.webkit.org/show_bug.cgi?id=225010
<rdar://problem/77092303>
Reviewed by John Wilander.
Fix a bug in PCM debug mode where we don't set the timer for 10_s
after sending the first report for an attribution.
No new tests, this is debug mode only. Non-debug mode behavior is
covered by existing tests, and I tested debug mode manually.
- NetworkProcess/PrivateClickMeasurementManager.cpp:
(WebKit::PrivateClickMeasurementManager::firePendingAttributionRequests):
If the attribution has been sent to only one endpoint, indicated by
a non-null laterTimeToSend value, we should set the timer to be 10
seconds if debug mode is enabled.
Also, change the interval time from 1 minute to 10 seconds, because
there was no good reason that it was 1 minute and we should be
consistent. Also 1 minute is a long time to wait during a test.
- 5:40 PM Changeset in webkit [276544] by
-
- 1 copy in tags/Safari-611.2.4
Tag Safari-611.2.4.
- 5:36 PM Changeset in webkit [276543] by
-
- 1 copy in tags/Safari-611.1.21.2.6
Tag Safari-611.1.21.2.6.
- 5:26 PM Changeset in webkit [276542] by
-
- 2 edits in branches/safari-611.1.21.2-branch/Source/JavaScriptCore
Cherry-pick r276527. rdar://problem/77092686
[YARR Interpreter] Improper backtrack of parentheses with non-zero based greedy quantifiers
https://bugs.webkit.org/show_bug.cgi?id=224983
Reviewed by Mark Lam.
When we backtrack a parentheses with a greedy non zero based quantifier,
we don't properly restore for the case where we hadn't reached the minimum count.
We now save the input position on entry and restore it when we backtrack for
this case. We also properly release the allocated ParenthesesDisjunctionContext's.
- yarr/YarrInterpreter.cpp: (JSC::Yarr::Interpreter::matchParentheses): (JSC::Yarr::Interpreter::backtrackParentheses):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@276527 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 5:26 PM Changeset in webkit [276541] by
-
- 8 edits in branches/safari-611.1.21.2-branch/Source
Versioning.
WebKit-7611.1.21.2.6
- 5:25 PM Changeset in webkit [276540] by
-
- 1 edit2 deletes in trunk/Tools
Delete unused buildbot.css and default.css
https://bugs.webkit.org/show_bug.cgi?id=224997
Reviewed by Alexey Proskuryakov.
- CISupport/build-webkit-org/public_html/buildbot.css: Removed.
- CISupport/build-webkit-org/public_html/default.css: Removed.
- 5:22 PM Changeset in webkit [276539] by
-
- 1 copy in tags/Safari-611.1.21.161.6
Tag Safari-611.1.21.161.6.
- 5:21 PM Changeset in webkit [276538] by
-
- 2 edits in branches/safari-611.1.21.161-branch/Source/JavaScriptCore
Cherry-pick r276527. rdar://problem/77092673
[YARR Interpreter] Improper backtrack of parentheses with non-zero based greedy quantifiers
https://bugs.webkit.org/show_bug.cgi?id=224983
Reviewed by Mark Lam.
When we backtrack a parentheses with a greedy non zero based quantifier,
we don't properly restore for the case where we hadn't reached the minimum count.
We now save the input position on entry and restore it when we backtrack for
this case. We also properly release the allocated ParenthesesDisjunctionContext's.
- yarr/YarrInterpreter.cpp: (JSC::Yarr::Interpreter::matchParentheses): (JSC::Yarr::Interpreter::backtrackParentheses):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@276527 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 5:21 PM Changeset in webkit [276537] by
-
- 2 edits in branches/safari-611.1.21.161-branch/Source/JavaScriptCore
Cherry-pick r276524. rdar://problem/77092702
Fix B3 strength reduction for shl.
https://bugs.webkit.org/show_bug.cgi?id=224913
rdar://76978874
Reviewed by Michael Saboff.
If the operation can potentially either underflow or overflow, then the result
can be any value.
- b3/B3ReduceStrength.cpp:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@276524 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 5:18 PM Changeset in webkit [276536] by
-
- 1 edit2 copies1 add1 delete in trunk/Websites/bugs.webkit.org
Bugzilla needs to linkify identifiers (Part 1)
https://bugs.webkit.org/show_bug.cgi?id=224312
<rdar://problem/75845629>
Reviewed by Dewei Zhu.
- extensions/Commits: Moved from Trac.
- extensions/Commits/Config.pm: Changed Plugin name.
- extensions/Commits/Extension.pm: Ditto.
(_replace_revision): Use commit.webkit.org instead of trac.
- extensions/Trac: Moved to Commits.
- 5:18 PM Changeset in webkit [276535] by
-
- 8 edits in branches/safari-611.1.21.161-branch/Source
Versioning.
WebKit-7611.1.21.161.6
- 5:16 PM Changeset in webkit [276534] by
-
- 1 copy in tags/Safari-611.1.21.0.14
Tag Safari-611.1.21.0.14.
- 5:13 PM Changeset in webkit [276533] by
-
- 2 edits in branches/safari-611.1.21.0-branch/Source/JavaScriptCore
Cherry-pick r276527. rdar://problem/77092655
[YARR Interpreter] Improper backtrack of parentheses with non-zero based greedy quantifiers
https://bugs.webkit.org/show_bug.cgi?id=224983
Reviewed by Mark Lam.
When we backtrack a parentheses with a greedy non zero based quantifier,
we don't properly restore for the case where we hadn't reached the minimum count.
We now save the input position on entry and restore it when we backtrack for
this case. We also properly release the allocated ParenthesesDisjunctionContext's.
- yarr/YarrInterpreter.cpp: (JSC::Yarr::Interpreter::matchParentheses): (JSC::Yarr::Interpreter::backtrackParentheses):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@276527 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 5:13 PM Changeset in webkit [276532] by
-
- 2 edits in branches/safari-611.1.21.0-branch/Source/JavaScriptCore
Cherry-pick r276524. rdar://problem/77092695
Fix B3 strength reduction for shl.
https://bugs.webkit.org/show_bug.cgi?id=224913
rdar://76978874
Reviewed by Michael Saboff.
If the operation can potentially either underflow or overflow, then the result
can be any value.
- b3/B3ReduceStrength.cpp:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@276524 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 5:08 PM Changeset in webkit [276531] by
-
- 8 edits in branches/safari-611.1.21.0-branch/Source
Versioning.
WebKit-7611.1.21.0.14
- 4:45 PM Changeset in webkit [276530] by
-
- 2 edits in trunk/Source/WebCore
Crash in constructCustomElementSynchronously
https://bugs.webkit.org/show_bug.cgi?id=224992
<rdar://66988026>
Reviewed by Tadeu Zagallo.
Exit early when the global object is nullptr although this shouldn't happen.
No new tests since we have no reproductions.
- bindings/js/JSCustomElementInterface.cpp:
(WebCore::JSCustomElementInterface::tryToConstructCustomElement): Added a null check.
- 4:38 PM Changeset in webkit [276529] by
-
- 2 edits in branches/safari-611-branch/Source/JavaScriptCore
Cherry-pick r276527. rdar://problem/77091667
[YARR Interpreter] Improper backtrack of parentheses with non-zero based greedy quantifiers
https://bugs.webkit.org/show_bug.cgi?id=224983
Reviewed by Mark Lam.
When we backtrack a parentheses with a greedy non zero based quantifier,
we don't properly restore for the case where we hadn't reached the minimum count.
We now save the input position on entry and restore it when we backtrack for
this case. We also properly release the allocated ParenthesesDisjunctionContext's.
- yarr/YarrInterpreter.cpp: (JSC::Yarr::Interpreter::matchParentheses): (JSC::Yarr::Interpreter::backtrackParentheses):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@276527 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 4:21 PM Changeset in webkit [276528] by
-
- 3 edits in trunk/Source/WebCore
Add additional guards around USE_ANGLE
https://bugs.webkit.org/show_bug.cgi?id=225001
Reviewed by Fujii Hironori.
When USE(ANGLE) initializeOpenGLShims isn't used or available.
When USE(ANGLE) the ANGLE shader compiler isn't needed.
- platform/graphics/GLContext.cpp:
(WebCore::initializeOpenGLShimsIfNeeded):
- platform/graphics/texmap/GraphicsContextGLTextureMapper.cpp:
- 4:06 PM Changeset in webkit [276527] by
-
- 2 edits in trunk/Source/JavaScriptCore
[YARR Interpreter] Improper backtrack of parentheses with non-zero based greedy quantifiers
https://bugs.webkit.org/show_bug.cgi?id=224983
Reviewed by Mark Lam.
When we backtrack a parentheses with a greedy non zero based quantifier,
we don't properly restore for the case where we hadn't reached the minimum count.
We now save the input position on entry and restore it when we backtrack for
this case. We also properly release the allocated ParenthesesDisjunctionContext's.
- yarr/YarrInterpreter.cpp:
(JSC::Yarr::Interpreter::matchParentheses):
(JSC::Yarr::Interpreter::backtrackParentheses):
- 4:01 PM Changeset in webkit [276526] by
-
- 5 edits1 copy2 adds in trunk/LayoutTests
http/tests/security/referrer-policy-header.html is slow
https://bugs.webkit.org/show_bug.cgi?id=224998
Reviewed by Geoffrey Garen.
The test is a flaky timeout on the bots because it is slow. On my machine, it takes 17 seconds to run
it with a debug build. To speed things up, I am now running the subtests in parallel instead of
sequentially. I also split the test in 2 (regular response vs multipart response). Each tests now takes
about ~5 seconds to run on my machine.
- http/tests/security/referrer-policy-header-expected.txt:
- http/tests/security/referrer-policy-header-multipart-expected.txt: Added.
- http/tests/security/referrer-policy-header-multipart.html: Added.
- http/tests/security/referrer-policy-header-test.js: Copied from LayoutTests/http/tests/security/referrer-policy-header.html.
(printResults):
(onmessage):
(runTests):
- http/tests/security/referrer-policy-header.html:
- http/tests/security/resources/postReferrer.py:
- http/tests/security/resources/serve-referrer-policy-and-test.py:
- 3:46 PM Changeset in webkit [276525] by
-
- 2 edits in branches/safari-611-branch/Source/JavaScriptCore
Cherry-pick r276524. rdar://problem/77089783
Fix B3 strength reduction for shl.
https://bugs.webkit.org/show_bug.cgi?id=224913
rdar://76978874
Reviewed by Michael Saboff.
If the operation can potentially either underflow or overflow, then the result
can be any value.
- b3/B3ReduceStrength.cpp:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@276524 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 3:14 PM Changeset in webkit [276524] by
-
- 2 edits in trunk/Source/JavaScriptCore
Fix B3 strength reduction for shl.
https://bugs.webkit.org/show_bug.cgi?id=224913
rdar://76978874
Reviewed by Michael Saboff.
If the operation can potentially either underflow or overflow, then the result
can be any value.
- b3/B3ReduceStrength.cpp:
- 3:06 PM Changeset in webkit [276523] by
-
- 8 edits in trunk/Source/WebKit
Unreviewed, reverting r275562.
https://bugs.webkit.org/show_bug.cgi?id=225002
Causes issues in iOS contextmenu animations and interactions
with other popovers
Reverted changeset:
"[iOS] contextmenu hints can be clipped by the WKWebView"
https://bugs.webkit.org/show_bug.cgi?id=224204
https://trac.webkit.org/changeset/275562
- 3:04 PM Changeset in webkit [276522] by
-
- 2 edits in trunk/Source/WebKit
REGRESSION (r273154): fast/forms/ios/repeatedly-focus-offscreen-select.html is consistently failing
https://bugs.webkit.org/show_bug.cgi?id=224985
<rdar://problem/77042177>
Reviewed by Wenson Hsieh.
The test is failing after r273154, which made <select> elements present
a UIMenu rather than a popover. The old logic ensured popovers were not
presented when the <select> element was offscreen (see r265117 for more
information on why that behavior was necessary), but the new
presentation omitted that logic.
To achieve the correct behavior, and fix the failing test, <select>
menus should not be presented when the element is offscreen.
- UIProcess/ios/forms/WKFormSelectPicker.mm:
(-[WKSelectPicker controlBeginEditing]):
- 3:00 PM Changeset in webkit [276521] by
-
- 7 edits in trunk
Not computing image aspect ratios from width and height attributes for lazy loaded images
https://bugs.webkit.org/show_bug.cgi?id=224197
Reviewed by Darin Adler.
LayoutTests/imported/w3c:
The test cases for error images and images without src in img-aspect-ratio.html are passed. This patch
doesn't change the behavior of the original aspect ratio case, so it's failed like before.
- web-platform-tests/html/rendering/replaced-elements/attributes-for-embedded-content-and-images/img-aspect-ratio-expected.txt:
Source/WebCore:
This patch supports error images and lazy loaded images (without src attribute) to compute
implicit aspect ratios from width and height attributes. Refactor the code a bit. Added
intrinsicAspectRatioFromWidthHeight() to compute aspect ratio from width and height attributes when
the object is allowed to which is decided by canMapWidthHeightToAspectRatio().
Remove!downcast<RenderImage>(*this).cachedImage()constraint, so that images without src attributes
is allowed. As to error images, compute the aspect ratio when the image shouldDisplayBrokenImageIcon().
- rendering/RenderImage.cpp:
(WebCore::RenderImage::canMapWidthHeightToAspectRatio const): To indicate that the object is allowed
to compute aspect ratio from width and height attributes.
(WebCore::RenderImage::computeIntrinsicRatioInformation const): When shouldDisplayBrokenImageIcon(),
try to compute the aspect ratio from attributes width and height.
- rendering/RenderImage.h:
- rendering/RenderReplaced.cpp:
(WebCore::RenderReplaced::intrinsicAspectRatioFromWidthHeight const): Compute the aspect ratio from attributes width and height.
(WebCore::RenderReplaced::computeIntrinsicRatioInformation const):
- rendering/RenderReplaced.h:
(WebCore::RenderReplaced::canMapWidthHeightToAspectRatio const): Ditto.
- 2:53 PM Changeset in webkit [276520] by
-
- 3 edits in trunk/LayoutTests
imported/w3c/web-platform-tests/html/semantics/scripting-1/the-script-element/css-module/css-module-worker-test.html is a flakey crash and failure
https://bugs.webkit.org/show_bug.cgi?id=222750
Unreviewed test gardening.
Move expectation to a more general location, this isn't limited to macOS.
- TestExpectations:
- platform/mac/TestExpectations:
- 2:38 PM Changeset in webkit [276519] by
-
- 2 edits in trunk/Source/WebKit
Unreviewed, reverting r276451.
https://bugs.webkit.org/show_bug.cgi?id=224999
Broke perf testing
Reverted changeset:
"[iOS] Update sandbox message filter syntax"
https://bugs.webkit.org/show_bug.cgi?id=223384
https://trac.webkit.org/changeset/276451
- 2:32 PM Changeset in webkit [276518] by
-
- 1 edit in branches/safari-611-branch/Source/WebKit/UIProcess/AuxiliaryProcessProxy.cpp
Unreviewed build fix. rdar://problem/76963040
WebKit2/UIProcess/AuxiliaryProcessProxy.cpp:163:10: error: use of undeclared identifier 'isMainRunLoop'
- 2:29 PM Changeset in webkit [276517] by
-
- 13 edits in branches/safari-611-branch/Source
Cherry-pick r276324. rdar://problem/77086404
FullGCActivityCallback should use the percentage of pages uncompressed in RAM to determine deferral.
https://bugs.webkit.org/show_bug.cgi?id=224817
Reviewed by Filip Pizlo.
Source/JavaScriptCore:
Right now we try to determine if too many pages are paged out by
dereferencing them and bailing out of the GC if we go over a
deadline. While this works if the only goal is to avoid causing
extensive thrashing on spinny disks (HDD), it doesn't prevent
thrashing when access to disk is fast (e.g. SSD). This is because
on fast disks the proportional time to load the memory from disk
is much lower. Additionally, on SSDs in particular we don't want
to load the pages into RAM then bail as that will force a
different page onto disk, increasing wear.
This patch switches to asking the OS if each MarkedBlock is paged
out. Then if we are over a threshold we wait until we would have
GC'd anyway. This patch uses the (maxVMGrowthFactor - 1) as the
percentage of "slow" pages (paged out or compressed) needed to
defer the GC. The idea behind that threshold is that if we add
that many pages then the same number of pages would be forced
out of RAM for us to do a GC anyway (in the limit).
- heap/BlockDirectory.cpp: (JSC::BlockDirectory::updatePercentageOfPagedOutPages): (JSC::BlockDirectory::isPagedOut): Deleted.
- heap/BlockDirectory.h:
- heap/FullGCActivityCallback.cpp: (JSC::FullGCActivityCallback::doCollection):
- heap/Heap.cpp: (JSC::Heap::isPagedOut):
- heap/Heap.h:
- heap/MarkedSpace.cpp: (JSC::MarkedSpace::isPagedOut):
- heap/MarkedSpace.h:
- runtime/OptionsList.h:
Source/WebKit:
Add mincore to the acceptable syscall list.
- WebProcess/com.apple.WebProcess.sb.in:
Source/WTF:
Add a noexcept flavor of FunctionTraits. On Linux mincore (and probably other syscalls) are marked noexcept so the existing overloads don't work.
- wtf/FunctionTraits.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@276324 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 2:01 PM Changeset in webkit [276516] by
-
- 2 edits in trunk/Source/JavaScriptCore
[JSC][Win] callOperationNoExceptionCheck() also needs to support operations that return SlowPathReturnType
https://bugs.webkit.org/show_bug.cgi?id=224964
Reviewed by Yusuke Suzuki.
r229989 (Bug 183655) added the x64 Windows support only for
callOperation(), but for callOperationNoExceptionCheck().
callOperationNoExceptionCheck() also needs the x64 Windows
support.
This change is a preparation for Bug 224920 that is going to use
callOperationNoExceptionCheck instead of callOperation.
- jit/JIT.h:
(callOperation): Rewrote by using 'if constexpr' instead of SFINAE.
(callOperationNoExceptionCheck): Added a new implementation for
x64 Windows based on callOperation.
- 1:11 PM Changeset in webkit [276515] by
-
- 3 edits in trunk/Source/WebKit
[iOS] GPU Process sandbox lacks IOMobileFramebufferUserClient method filter
https://bugs.webkit.org/show_bug.cgi?id=224956
<rdar://problem/68227590>
Reviewed by Per Arne Vollan.
This patch updates the GPU process sandbox to match the method filter used in the WebContent process
so that we will retain consistent protections when activating the GPU Process. These are the same rules
governing access to this resource when the GPU process is disabled, so should not create a change in behavior.
We also remove unneeded telemetry to reduce messaging burden and avoid performance issues.
- Resources/SandboxProfiles/ios/com.apple.WebKit.GPU.sb:
- Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb:
- 12:53 PM Changeset in webkit [276514] by
-
- 2 edits in trunk/Source/WebKit
Web Inspector: [Cocoa] WKInspectorResourceURLSchemeHandler needs to serialize cleanup actions
https://bugs.webkit.org/show_bug.cgi?id=224986
<rdar://76768454>
Reviewed by Devin Rousso.
- UIProcess/Inspector/mac/WKInspectorResourceURLSchemeHandler.mm:
(-[WKInspectorResourceURLSchemeHandler webView:startURLSchemeTask:]):
Do cleanup on the main queue so that it can be serialized with reads.
(-[WKInspectorResourceURLSchemeHandler webView:stopURLSchemeTask:]):
Ensure that all removals from the map are processed before doing a lookup.
- 12:51 PM Changeset in webkit [276513] by
-
- 6 edits in trunk/Source/WebCore
Remove virtual function calls in GraphicsLayer destructors
https://bugs.webkit.org/show_bug.cgi?id=180232
Reviewed by Adrian Perez de Castro.
I notice that ~CoordinatedGraphicsLayer makes a virtual function call to
GraphicsLayer::willBeDestroyed, which makes a virtual function call to
CoordinatedGraphicsLayer::removeFromParent. I think that the functions are being called as
intended, because ~CoordinatedGraphicsLayer has not yet been fully destroyed. However, I'm
reminded of Effective C++ item #9: Never call virtual functions during construction or
destruction ("because such calls will never go to a more derived class than that of the
currently executing constructor or destructor"). This code is almost certain to break if
anyone tries in the future to subclass any of the existing subclasses of GraphicsLayer, so
let's refactor it a bit. This doesn't fix anything, but my hope is that it will make the
code a bit harder to break, and not the opposite.
The main risk here is that some reordering of operations is necessary. The derived class
portion of removeFromParent must now be executed before willBeDestroyed. It can't happen
after, because parent would already be unset by that point. It's hard to be certain that
this won't break anything, but I think it should be fine.
- platform/graphics/GraphicsLayer.cpp:
(WebCore::GraphicsLayer::willBeDestroyed):
(WebCore::GraphicsLayer::removeFromParentInternal):
(WebCore::GraphicsLayer::removeFromParent):
- platform/graphics/GraphicsLayer.h:
- platform/graphics/ca/GraphicsLayerCA.cpp:
(WebCore::GraphicsLayerCA::~GraphicsLayerCA):
(WebCore::GraphicsLayerCA::willBeDestroyed): Deleted.
- platform/graphics/ca/GraphicsLayerCA.h:
- platform/graphics/texmap/coordinated/CoordinatedGraphicsLayer.cpp:
(WebCore::CoordinatedGraphicsLayer::~CoordinatedGraphicsLayer):
- 12:34 PM Changeset in webkit [276512] by
-
- 4 edits2 adds in trunk/Tools
Add a conftest.py to run existing webkitpy tests in pytest
https://bugs.webkit.org/show_bug.cgi?id=224687
Reviewed by Jonathan Bedard.
- Scripts/webkitpy/common/system/executive_unittest.py:
(ExecutiveTest.serial_test_run_in_parallel): Deal with the fact that pytest
running the tests might be not be the same version as the autoinstalled version,
and not API compatible.
- Scripts/webkitpy/conftest.py: Added.
(pytest_configure): Define the markers the plugins in conftest use
(pytest_addoption): Add --run-integration to allow them to be disabled by default.
(pytest_pycollect_makeitem): Rename serial/integration tests so pytest finds them.
(pytest_collection_modifyitems): Mark tests as skipped when needed per the above.
- Scripts/webkitpy/pytest.ini: Added.
- Scripts/webkitpy/test/main_unittest.py:
(TestStubs): Stop these from being picked up by pytest as tests.
- Scripts/webkitpy/test/markers.py: Fix this so pytest is technically optional,
even though it is always present because of the autoinstalled copy.
- 12:21 PM Changeset in webkit [276511] by
-
- 1 edit in branches/safari-611-branch/Source/WebKit/UIProcess/Network/NetworkProcessProxy.cpp
Unreviewed build fix.
WebKit2/UIProcess/Network/NetworkProcessProxy.cpp:127:17: error: use of undeclared identifier 'defaultProcess'
- 12:06 PM Changeset in webkit [276510] by
-
- 10 edits in trunk/Source
Remove decoder memory allocations based on untrusted data (sizes) in the stream; related changes
https://bugs.webkit.org/show_bug.cgi?id=224984
Reviewed by Sam Weinig.
Source/WebCore:
- platform/network/cf/CertificateInfoCFNet.cpp:
(WTF::Persistence::decodeCFData): Removed unneeded check for zero size. Removed code that
locally allocates a vector before bufferIsLargeEnoughToContain is called. Instead use
bufferPointerForDirectRead, which makes does the buffer size check, and pass the pointer
directly to CFDataCreate.
Source/WebKit:
- Platform/IPC/ArgumentCoders.h: Remove the calls to
HashMap::reserveInitialCapacity and HashSet::reserveInitialCapacity, based
on number read in from the decoder. This means there will be more wasted
memory in these HashMap and HashSet objects, so we have to test to make
sure this does not create a performance problem. But without this check,
we are trying to allocate memory based on an unstrusted size.
- Shared/Cocoa/WebCoreArgumentCodersCocoa.mm:
(IPC::ArgumentCoder<RefPtr<ApplePayError>>::encode): Removed the coder
for a Vector of these RefPtr, replaced it with a coder for an individual one,
allowing the Vector ArgumentCoder template to handle vector size and construction.
One benefit is that this adds in a shrinkToFit and prevents us from making any
separate mistake about pre-sizing the Vector here since we use shared code.
(IPC::ArgumentCoder<RefPtr<ApplePayError>>::decode): Ditto.
- Shared/WebCoreArgumentCoders.cpp:
(IPC::ArgumentCoder<RefPtr<SecurityOrigin>>::encode): Ditto.
(IPC::ArgumentCoder<RefPtr<SecurityOrigin>>::decode): Ditto.
(IPC::ArgumentCoder<WebCore::CDMInstanceSession::KeyStatusVector>::encode):
(IPC::ArgumentCoder<WebCore::CDMInstanceSession::KeyStatusVector>::decode):
Removed unnecessary specialization for the KeyStatusVector. There is already
an ArgumentCoder for Vector, for std::pair, for Ref<SharedBuffer>, and for
enumerations like CDMKeyStatus, so there's no need to have a specialized
coder for this. This function that we are removing had a call to
reserveInitialCapacity, but the Vector ArgumentCoder template does not.
- Shared/WebCoreArgumentCoders.h: Replaced the
ArgumentCoder<Vector<RefPtr<WebCore::ApplePayError>>> specialization with
ArgumentCoder<RefPtr<WebCore::ApplePayError>>. Removed the
ArgumentCoder<WebCore::CDMInstanceSession::KeyStatusVector> specialization.
Source/WTF:
- wtf/persistence/PersistentDecoder.cpp:
(WTF::Persistence::Decoder::bufferPointerForDirectRead): Added.
(WTF::Persistence::Decoder::decodeFixedLengthData): Refactor to use bufferPointerForDirectRead.
- wtf/persistence/PersistentDecoder.h: Added bufferPointerForDirectRead function for use in the
rare cases where we want to read directly out of the decoder buffer, rather than writing to a
passed-in pointer. Also did a small refactoring of bufferIsLargeEnoughToContain to use &&
rather than an if statement.
- 11:57 AM Changeset in webkit [276509] by
-
- 2 edits in trunk/Source/WebKit
Set proper network logging level for full web browsers
https://bugs.webkit.org/show_bug.cgi?id=224982
<rdar://problem/76903650>
Reviewed by Brent Fulgham.
Full web browsers have different low-level network logging requirements
for ephemeral sessions.
No new tests. There's no automated way to test if network level logs
show up. I tested that logging is correct for web browser and non web
browser apps manually.
- NetworkProcess/cocoa/NetworkSessionCocoa.mm:
(WebKit::configurationForSessionID):
(WebKit::NetworkSessionCocoa::NetworkSessionCocoa):
- 11:54 AM Changeset in webkit [276508] by
-
- 2 edits in trunk/LayoutTests
[macOS] imported/w3c/web-platform-tests/html/semantics/scripting-1/the-script-element/css-module/css-module-worker-test.html is a flakey crash and failure
https://bugs.webkit.org/show_bug.cgi?id=222750
Unreviewed test gardening.
- platform/mac/TestExpectations:
- 11:15 AM Changeset in webkit [276507] by
-
- 3 edits in trunk/Source/ThirdParty/ANGLE
Unreviewed, reverting r276190.
broke a test internally.
Reverted changeset:
"Metal-ANGLE: Shared memory texture tests failing in iOS
Simulator"
https://bugs.webkit.org/show_bug.cgi?id=222685
https://commits.webkit.org/r276190
- 10:44 AM Changeset in webkit [276506] by
-
- 4 edits in trunk/Source
Disable GPUProcess on shipping iOS
https://bugs.webkit.org/show_bug.cgi?id=224897
Reviewed by Eric Carlson.
Source/WebKit:
- Shared/WebPreferencesDefaultValues.cpp:
(WebKit::defaultUseGPUProcessForCanvasRenderingEnabled):
(WebKit::defaultUseGPUProcessForDOMRenderingEnabled):
(WebKit::defaultUseGPUProcessForMediaEnabled):
(WebKit::defaultUseGPUProcessForWebGLEnabled):
(WebKit::defaultCaptureAudioInGPUProcessEnabled):
(WebKit::defaultCaptureVideoInGPUProcessEnabled):
(WebKit::defaultWebRTCCodecsInGPUProcess):
Source/WTF:
- wtf/PlatformEnableCocoa.h:
- 10:31 AM Changeset in webkit [276505] by
-
- 2 edits in trunk/Tools
Make report-non-inclusive-language ignore .db files
https://bugs.webkit.org/show_bug.cgi?id=224979
Reviewed by Darin Adler.
- Scripts/report-non-inclusive-language:
- 10:21 AM Changeset in webkit [276504] by
-
- 7 edits in trunk
Fix KVO for camera/microphone capture state WKWebView API
https://bugs.webkit.org/show_bug.cgi?id=224922
<rdar://problem/77008199>
Reviewed by Eric Carlson.
Source/WebKit:
Add missing willChange observer call and migrate keys to API keys.
Covered by API test.
- UIProcess/Cocoa/PageClientImplCocoa.h:
- UIProcess/Cocoa/PageClientImplCocoa.mm:
(WebKit::PageClientImplCocoa::microphoneCaptureWillChange):
(WebKit::PageClientImplCocoa::cameraCaptureWillChange):
(WebKit::PageClientImplCocoa::microphoneCaptureChanged):
(WebKit::PageClientImplCocoa::cameraCaptureChanged):
- UIProcess/PageClient.h:
(WebKit::PageClient::microphoneCaptureWillChange):
(WebKit::PageClient::cameraCaptureWillChange):
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::updateReportedMediaCaptureState):
Tools:
Add tests for capture state API.
- TestWebKitAPI/Tests/WebKit/GetUserMedia.mm:
(-[MediaCaptureObserver observeValueForKeyPath:ofObject:change:context:]):
(TestWebKitAPI::waitUntilCameraState):
(TestWebKitAPI::waitUntilMicrophoneState):
(TestWebKitAPI::TEST):
- 10:04 AM Changeset in webkit [276503] by
-
- 1 edit in branches/safari-611-branch/Source/WebCore/layout/inlineformatting/InlineContentBreaker.cpp
Apply patch. rdar://problem/76963029
- 10:00 AM WebKitGTK/2.32.x edited by
- (diff)
- 9:55 AM Changeset in webkit [276502] by
-
- 25 edits in trunk/Source
Improve our constructDeletedValue() template specializations
https://bugs.webkit.org/show_bug.cgi?id=224889
Reviewed by Darin Adler.
Source/WebCore:
Improve our constructDeletedValue() template specializations and make them more consistent:
- Use placement-new instead of object assignment since we don't want/need to destroy the existing object (since it is uninitialized).
- Do as little initialization as possible for performance reasons.
- Modules/indexeddb/IDBKeyData.cpp:
(WebCore::IDBKeyData::deletedValue): Deleted.
- Modules/indexeddb/IDBKeyData.h:
(WebCore::IDBKeyDataHashTraits::constructDeletedValue):
(WebCore::IDBKeyDataHashTraits::isDeletedValue):
(WebCore::IDBKeyData::isDeletedValue const): Deleted.
- Modules/indexeddb/shared/IDBResourceIdentifier.cpp:
(WebCore::IDBResourceIdentifier::deletedValue): Deleted.
(WebCore::IDBResourceIdentifier::isHashTableDeletedValue const): Deleted.
- Modules/indexeddb/shared/IDBResourceIdentifier.h:
(WebCore::IDBResourceIdentifierHashTraits::constructDeletedValue):
(WebCore::IDBResourceIdentifierHashTraits::isDeletedValue):
- dom/MessagePortIdentifier.h:
(WTF::HashTraits<WebCore::MessagePortIdentifier>::constructDeletedValue):
(WTF::HashTraits<WebCore::MessagePortIdentifier>::isDeletedValue):
- history/BackForwardItemIdentifier.h:
(WTF::HashTraits<WebCore::BackForwardItemIdentifier>::constructDeletedValue):
(WTF::HashTraits<WebCore::BackForwardItemIdentifier>::isDeletedValue):
- layout/LayoutUnits.h:
(WTF::HashTraits<WebCore::Layout::SlotPosition>::constructDeletedValue):
(WTF::HashTraits<WebCore::Layout::SlotPosition>::isDeletedValue):
- loader/PrivateClickMeasurement.h:
(WebCore::PrivateClickMeasurement::AttributionDestinationSite::matches const):
(WTF::HashTraits<WebCore::PrivateClickMeasurement::SourceSite>::constructDeletedValue):
(WTF::HashTraits<WebCore::PrivateClickMeasurement::SourceSite>::isDeletedValue):
(WTF::HashTraits<WebCore::PrivateClickMeasurement::AttributionDestinationSite>::constructDeletedValue):
(WTF::HashTraits<WebCore::PrivateClickMeasurement::AttributionDestinationSite>::isDeletedValue):
(WebCore::PrivateClickMeasurement::SourceSite::isHashTableDeletedValue const): Deleted.
(WebCore::PrivateClickMeasurement::SourceSite::deletedValue): Deleted.
(WebCore::PrivateClickMeasurement::SourceSite::constructDeletedValue): Deleted.
(WebCore::PrivateClickMeasurement::SourceSite::deleteValue): Deleted.
(WebCore::PrivateClickMeasurement::SourceSite::isDeletedValue const): Deleted.
(WebCore::PrivateClickMeasurement::AttributionDestinationSite::isHashTableDeletedValue const): Deleted.
(WebCore::PrivateClickMeasurement::AttributionDestinationSite::deletedValue): Deleted.
(WebCore::PrivateClickMeasurement::AttributionDestinationSite::constructDeletedValue): Deleted.
(WebCore::PrivateClickMeasurement::AttributionDestinationSite::deleteValue): Deleted.
(WebCore::PrivateClickMeasurement::AttributionDestinationSite::isDeletedValue const): Deleted.
- page/ClientOrigin.h:
(WTF::HashTraits<WebCore::ClientOrigin>::constructDeletedValue):
- page/GlobalWindowIdentifier.h:
(WTF::HashTraits<WebCore::GlobalWindowIdentifier>::constructDeletedValue):
(WTF::HashTraits<WebCore::GlobalWindowIdentifier>::isDeletedValue):
- platform/Cookie.h:
(WTF::HashTraits<WebCore::Cookie>::constructDeletedValue):
(WebCore::Cookie::Cookie): Deleted.
- platform/graphics/FontCache.cpp:
(WebCore::FontPlatformDataCacheKeyHashTraits::constructDeletedValue):
- platform/graphics/FontCache.h:
(WebCore::FontCascadeCacheKeyHashTraits::constructDeletedValue):
- platform/graphics/IntPointHash.h:
(WTF::HashTraits<WebCore::IntPoint>::constructDeletedValue):
(WTF::HashTraits<WebCore::IntPoint>::isDeletedValue):
- rendering/CSSValueKey.h:
(WTF::HashTraits<WebCore::CSSValueKey>::constructDeletedValue):
- workers/service/ServiceWorkerClientIdentifier.h:
(WTF::HashTraits<WebCore::ServiceWorkerClientIdentifier>::constructDeletedValue):
(WTF::HashTraits<WebCore::ServiceWorkerClientIdentifier>::isDeletedValue):
Source/WebCore/PAL:
- pal/SessionID.h:
(WTF::HashTraits<PAL::SessionID>::constructDeletedValue):
Source/WebKit:
Improve our constructDeletedValue() template specializations and make them more consistent:
- Use placement-new instead of object assignment since we don't want/need to destroy the existing object (since it is uninitialized).
- Do as little initialization as possible for performance reasons.
- NetworkProcess/cache/NetworkCache.h:
(WTF::HashTraits<WebKit::NetworkCache::GlobalFrameID>::constructDeletedValue):
(WTF::HashTraits<WebKit::NetworkCache::GlobalFrameID>::isDeletedValue):
- Platform/IPC/StringReference.h:
(WTF::HashTraits<IPC::StringReference>::constructDeletedValue):
- Shared/CallbackID.h:
(WTF::HashTraits<WebKit::CallbackID>::constructDeletedValue):
(WTF::HashTraits<WebKit::CallbackID>::isDeletedValue):
Source/WebKitLegacy/mac:
Improve our constructDeletedValue() template specializations and make them more consistent:
- Use placement-new instead of object assignment since we don't want/need to destroy the existing object (since it is uninitialized).
- Do as little initialization as possible for performance reasons.
- History/BinaryPropertyList.cpp:
(IntegerArray::integers const):
(IntegerArray::size const):
(IntegerArrayHashTraits::constructDeletedValue):
(IntegerArrayHashTraits::isDeletedValue):
(IntegerArray::markDeleted): Deleted.
(IntegerArray::isDeletedValue const): Deleted.
(IntegerArray::deletedValueSize): Deleted.
- 9:53 AM Changeset in webkit [276501] by
-
- 2 edits in branches/safari-611-branch/Source/WebKit
Cherry-pick r276482. rdar://problem/77074513
[Mac] CMBaseClass object pointers can become unaligned on x86
https://bugs.webkit.org/show_bug.cgi?id=224950
<rdar://77020922>
Reviewed by Eric Carlson.
CMBaseClass has a 4-byte version member before its 8-byte pointers on x86. Deal with this
the same way we do with other pointer-bearing, static, CM-type objects: enforce a 4-byte
packing, and prepend the struct with another 4-byte object in order to force the pointers
into 8-byte alignment.
- Shared/mac/MediaFormatReader/CoreMediaWrapped.h: (WebKit::CoreMediaWrapped<Wrapped>::vTable):
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@276482 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 9:31 AM Changeset in webkit [276500] by
-
- 2 edits in trunk/Source/WebCore
[GStreamer][EME] cbcs fixes
https://bugs.webkit.org/show_bug.cgi?id=224976
Reviewed by Philippe Normand.
In cbcs you can get from qtdemux buffers without an iv_size
because it had constant_iv_size. We need to check that.
When decryption bails out because of the buffer being unencrypted,
we should TRACE so.
cbcs buffers could not include the subsample_count, we consider 0
as default.
- platform/graphics/gstreamer/eme/WebKitCommonEncryptionDecryptorGStreamer.cpp:
(transformInPlace):
- 9:15 AM WebKitGTK/2.32.x edited by
- (diff)
- 9:14 AM WebKitGTK/2.32.x edited by
- (diff)
- 9:12 AM WebKitGTK/2.32.x edited by
- (diff)
- 8:40 AM Changeset in webkit [276499] by
-
- 2 edits in trunk/Source/WebCore
Improve HashTableDeletedValue of ServiceWorkerRegistrationKey
https://bugs.webkit.org/show_bug.cgi?id=224975
Patch by Michael Catanzaro <Michael Catanzaro> on 2021-04-23
Reviewed by Chris Dumez.
We construct only the SecurityOriginData, no need to construct a full
ServiceWorkerRegistrationKey or URL. No behavior changes.
- workers/service/ServiceWorkerRegistrationKey.h:
(WTF::HashTraits<WebCore::ServiceWorkerRegistrationKey>::constructDeletedValue):
(WTF::HashTraits<WebCore::ServiceWorkerRegistrationKey>::isDeletedValue):
(WebCore::ServiceWorkerRegistrationKey::isHashTableDeletedValue const): Deleted.
(WebCore::ServiceWorkerRegistrationKey::ServiceWorkerRegistrationKey): Deleted.
- 8:16 AM Changeset in webkit [276498] by
-
- 8 edits in trunk/LayoutTests/imported/w3c
Import the update to "mapping attribute width and height to aspect ratio" tests from WPT
https://bugs.webkit.org/show_bug.cgi?id=224911
Reviewed by Rob Buis.
Import the update to "mapping attribute width and height to aspect ratio" tests from WPT commit 0ae0bb7cab. Split up the test cases, to make sure
each image has its own test case and won't affect testing other images.
- web-platform-tests/html/rendering/replaced-elements/attributes-for-embedded-content-and-images/canvas-aspect-ratio-expected.txt:
- web-platform-tests/html/rendering/replaced-elements/attributes-for-embedded-content-and-images/img-aspect-ratio-expected.txt:
- web-platform-tests/html/rendering/replaced-elements/attributes-for-embedded-content-and-images/img-aspect-ratio.html:
- web-platform-tests/html/rendering/replaced-elements/attributes-for-embedded-content-and-images/picture-aspect-ratio-expected.txt:
- web-platform-tests/html/rendering/replaced-elements/attributes-for-embedded-content-and-images/picture-aspect-ratio.html:
- web-platform-tests/html/rendering/replaced-elements/attributes-for-embedded-content-and-images/resources/aspect-ratio.js:
(test_computed_style_aspect_ratio):
- web-platform-tests/html/rendering/replaced-elements/attributes-for-embedded-content-and-images/video-aspect-ratio-expected.txt:
- 7:32 AM Changeset in webkit [276497] by
-
- 5 edits in trunk/Source
Handle rotation correctly in WebRTC GPUProcess rendering pipeline
https://bugs.webkit.org/show_bug.cgi?id=224927
<rdar://76915991>
Reviewed by Eric Carlson.
Source/WebCore:
Add a routine to update the affine transformn of the root layer.
Manually tested.
- platform/graphics/avfoundation/objc/LocalSampleBufferDisplayLayer.h:
- platform/graphics/avfoundation/objc/LocalSampleBufferDisplayLayer.mm:
(WebCore::LocalSampleBufferDisplayLayer::updateRootLayerAffineTransform):
Source/WebKit:
Instead of updating the inner layer, update the root layer.
- GPUProcess/webrtc/RemoteSampleBufferDisplayLayer.cpp:
(WebKit::RemoteSampleBufferDisplayLayer::updateAffineTransform):
- 7:07 AM Changeset in webkit [276496] by
-
- 2 edits in trunk/Source/JavaScriptCore
Unreviewed, reverting r276486.
https://bugs.webkit.org/show_bug.cgi?id=224973
broke windows build
Reverted changeset:
"[JSC][Win] callOperationNoExceptionCheck() also needs to
support operations that return SlowPathReturnType"
https://bugs.webkit.org/show_bug.cgi?id=224964
https://trac.webkit.org/changeset/276486
- 6:29 AM Changeset in webkit [276495] by
-
- 2 edits in trunk/Source/WebCore
Unreviewed, GStreamer follow-up after r276493
- platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
(WebCore::MediaPlayerPrivateGStreamer::createGSTPlayBin): Remove spurious adoptGRef call, as
this is not transfer-full.
- 6:17 AM Changeset in webkit [276494] by
-
- 2 edits in trunk/LayoutTests
Unreviewed, GTK gardening
- platform/gtk/TestExpectations: Remove test expectation for
imported/w3c/web-platform-tests/mediacapture-fromelement/capture.html which is tracked in
the main TestExpectations already.
- 2:46 AM Changeset in webkit [276493] by
-
- 7 edits in trunk/Source/WebCore
[Media] Allow access to MediaElement id from MediaPlayerPrivate
https://bugs.webkit.org/show_bug.cgi?id=224818
Patch by Philippe Normand <pnormand@igalia.com> on 2021-04-23
Reviewed by Xabier Rodriguez-Calvar.
A new method is added in the MediaPlayer allowing to query the client media element for its
identifier. That could be useful for accurate naming of the internal player/pipeline in the
MediaPlayerPrivate. If no specific id was set on the media element then the id is empty
string and the MediaPlayerPrivate needs to handle that by forging a unique id.
This also lead me to simplify pipeline-related code in the GStreamer player. The player can
handle only one pipeline in its entire life-time so the code handling pipeline "re-loading"
was actually never hit.
- html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::parseAttribute):
- html/HTMLMediaElement.h:
- platform/graphics/MediaPlayer.cpp:
(WebCore::MediaPlayer::elementId const):
- platform/graphics/MediaPlayer.h:
(WebCore::MediaPlayerClient::mediaPlayerElementId const):
- platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:
(WebCore::MediaPlayerPrivateGStreamer::load):
(WebCore::MediaPlayerPrivateGStreamer::createGSTPlayBin):
(WebCore::MediaPlayerPrivateGStreamer::loadFull): Deleted.
(WebCore::MediaPlayerPrivateGStreamer::setPipeline): Deleted.
- platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.h:
- 2:43 AM Changeset in webkit [276492] by
-
- 2 edits in trunk/Source/WebCore
Do not set synchronous scrolling for layers without relevant scrolling scopes
https://bugs.webkit.org/show_bug.cgi?id=221383
Reviewed by Simon Fraser.
Fixed backgrounds could force the layer compositor to set synchronous scrolling because they're slow
to repaint. However that cannot be done if the layer's boxScrollingScope hasn't been previously updated
in RenderLayer::updateLayerPosition.
- rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::updateSynchronousScrollingNodes): Do not call
setSynchronousScrollingReasons() if there is no scrollingScope.
- 2:23 AM Changeset in webkit [276491] by
-
- 12 edits4 adds in trunk
[LibWebRTC][WPE][GTK] Enable vp9 and NO_MAIN_THREAD_WRAPPING
https://bugs.webkit.org/show_bug.cgi?id=222795
Patch by Philippe Normand <pnormand@igalia.com> on 2021-04-23
Reviewed by Xabier Rodriguez-Calvar.
Source/ThirdParty/libwebrtc:
- CMakeLists.txt: Enable NO_MAIN_THREAD_WRAPPING as in the XCode build and harmonize options
between Apple and non-Apple ports.
- Source/webrtc/common_video/h265/h265_pps_parser.h:
Source/WebCore:
Add VP9 encoding and decoding support in the WPE/GTK LibWebRTC backend. The underlying
encoding support is based on LibWebRTC's internal VP9 support. The GStreamer VPx encoders
lack simulcast and high profile support, so can't really be used anyway. The
GStreamerVP8Encoder was removed for this reason as well.
- platform/mediastream/libwebrtc/GStreamerVideoDecoderFactory.cpp:
(WebCore::VP9Decoder::VP9Decoder):
(WebCore::VP9Decoder::Create):
(WebCore::GStreamerVideoDecoderFactory::CreateVideoDecoder):
(WebCore::GStreamerVideoDecoderFactory::GStreamerVideoDecoderFactory):
(WebCore::GStreamerVideoDecoderFactory::GetSupportedFormats const):
- platform/mediastream/libwebrtc/GStreamerVideoDecoderFactory.h:
- platform/mediastream/libwebrtc/GStreamerVideoEncoder.cpp:
- platform/mediastream/libwebrtc/GStreamerVideoEncoderFactory.cpp:
(WebCore::GStreamerVideoEncoder::makeElement):
(WebCore::GStreamerVideoEncoder::createEncoder):
(WebCore::GStreamerVideoEncoder::SetRestrictionCaps):
(WebCore::GStreamerVideoEncoderFactory::CreateVideoEncoder):
(WebCore::GStreamerVideoEncoderFactory::GStreamerVideoEncoderFactory):
(WebCore::GStreamerVideoEncoderFactory::GetSupportedFormats const):
- platform/mediastream/libwebrtc/GStreamerVideoEncoderFactory.h:
- platform/mediastream/libwebrtc/LibWebRTCProviderGStreamer.cpp:
(WebCore::LibWebRTCProviderGStreamer::createDecoderFactory):
(WebCore::LibWebRTCProviderGStreamer::createEncoderFactory):
LayoutTests:
Update GLIB baselines for webrtc tests.
- platform/glib/TestExpectations:
- platform/glib/imported/w3c/web-platform-tests/webrtc/RTCPeerConnection-onnegotiationneeded-expected.txt: Added.
- platform/glib/imported/w3c/web-platform-tests/webrtc/RTCPeerConnection-restartIce.https-expected.txt: Added.
- platform/glib/imported/w3c/web-platform-tests/webrtc/RTCRtpTransceiver-setCodecPreferences-expected.txt: Added.
- 12:16 AM Changeset in webkit [276490] by
-
- 15 edits in trunk/Source/WebKit
[GTK] Implement PageClient::makeViewBlank()
https://bugs.webkit.org/show_bug.cgi?id=224236
Reviewed by Chris Dumez.
Make most of the changes in r275485 cross platform and implement PageClient::makeViewBlank() for the GTK port.
- UIProcess/API/glib/WebKitUIClient.cpp:
- UIProcess/API/gtk/PageClientImpl.cpp:
(WebKit::PageClientImpl::makeViewBlank):
- UIProcess/API/gtk/PageClientImpl.h:
- UIProcess/API/gtk/WebKitWebViewBase.cpp:
(webkitWebViewBaseSnapshot):
(webkitWebViewBaseDraw):
(webkitWebViewBaseMakeBlank):
- UIProcess/API/gtk/WebKitWebViewBasePrivate.h:
- UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::didUpdateRenderingAfterCommittingLoad):
(WebKit::WebPageProxy::stopMakingViewBlankDueToLackOfRenderingUpdate):
(WebKit::WebPageProxy::makeViewBlankIfUnpaintedSinceLastLoadCommit):
(WebKit::WebPageProxy::didCommitLoadForFrame):
- UIProcess/WebPageProxy.h:
- UIProcess/WebPageProxy.messages.in:
- UIProcess/mac/WebPageProxyMac.mm:
(WebKit::WebPageProxy::didUpdateRenderingAfterCommittingLoad): Deleted.
- WebProcess/WebPage/CoordinatedGraphics/CompositingCoordinator.cpp:
(WebKit::CompositingCoordinator::flushPendingLayerChanges):
- WebProcess/WebPage/CoordinatedGraphics/DrawingAreaCoordinatedGraphics.cpp:
(WebKit::DrawingAreaCoordinatedGraphics::updateBackingStoreState):
(WebKit::DrawingAreaCoordinatedGraphics::display):
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::didUpdateRendering):
(WebKit::WebPage::didCommitLoad):
- WebProcess/WebPage/WebPage.h:
- WebProcess/WebPage/mac/WebPageMac.mm:
(WebKit::WebPage::didUpdateRendering): Deleted.
Apr 22, 2021:
- 10:42 PM Changeset in webkit [276489] by
-
- 2 edits1 add in trunk/LayoutTests
[GTK] Unreviewed test gardening. Update baseline after r276448.
- platform/glib/TestExpectations:
- platform/gtk/fast/forms/datalist/input-list-expected.txt: Added.
- 10:25 PM Changeset in webkit [276488] by
-
- 18 edits2 adds in trunk
[css-counter-styles] Parse @counter-style descriptors
https://bugs.webkit.org/show_bug.cgi?id=224718
Patch by Tyler Wilcock <Tyler Wilcock> on 2021-04-22
Reviewed by Darin Adler.
LayoutTests/imported/w3c:
Parsing for all @counter-style descriptors is implemented with this
patch, so mark more tests passing.
You'll notice that some @counter-style descriptors implemented in this
patch did not gain any passing tests (e.g.pad,negative). In all
of these cases, the expected results contain a <string> value, and we
fail only because we incorrectly don't serialize these <string> values
with quotes. I have manually confirmed in all cases that these values
are properly parsed, so it's just the serialization that's incorrect.
These <string> values serialize without quotes because WebKit's representation
of custom identifiers is not a separate type, but instead overloaded onto the
CSS_STRING type. This means that during serialization time, WebKit must guess
whether it is actually serializing a string (and include quotes if so), or if
it's serializing a custom ident (leaving off quotes if so).
Relevant code snippet:
Relevant changelog snippet from David Hyatt, 2016-12-07:
We also overload CSS_STRING primitive value type and have it act as both a string
and a custom identifier. This is lame, since the parser should have made two different
types of objects instead, but since our parser doesn't do that yet, I added a serializeAsStringOrCustomIdent
that preserves our old behavior of "quote the string only if needed." In this case what
that really meant was "Try to guess that we were originally a custom ident and leave off
quotes if so." This function will go away once we properly create CSSStringValues and
CSSCustomIdentValues instead of turning the latter into strings.
- web-platform-tests/css/css-counter-styles/counter-style-fallback-expected.txt:
- web-platform-tests/css/css-counter-styles/counter-style-prefix-suffix-syntax-expected.txt:
- web-platform-tests/css/css-counter-styles/counter-style-range-syntax-expected.txt:
- web-platform-tests/css/css-counter-styles/counter-style-speak-as-syntax-expected.txt:
- web-platform-tests/css/css-counter-styles/counter-style-system-syntax-expected.txt:
Source/WebCore:
Implement parsing and CSSCounterStyleRule IDL interface for @counter-style descriptors.
See spec for full details on all descriptors:
https://drafts.csswg.org/css-counter-styles-3/#the-counter-style-rule
Test: webexposed/counter-style-image-symbols-not-exposed.html and WPTs
- css/CSSComputedStyleDeclaration.cpp:
(WebCore::ComputedStyleExtractor::valueForPropertyInStyle):
Returnnullptrfor new @counter-style descriptor properties.
- css/CSSCounterStyleRule.cpp:
(WebCore::toCounterStyleSystemEnum):
(WebCore::symbolsValidForSystem):
(WebCore::StyleRuleCounterStyle::newValueInvalidOrEqual const):
(WebCore::CSSCounterStyleRule::cssText const):
(WebCore::CSSCounterStyleRule::setName):
(WebCore::CSSCounterStyleRule::setterInternal):
(WebCore::CSSCounterStyleRule::setSystem):
(WebCore::CSSCounterStyleRule::setNegative):
(WebCore::CSSCounterStyleRule::setPrefix):
(WebCore::CSSCounterStyleRule::setSuffix):
(WebCore::CSSCounterStyleRule::setRange):
(WebCore::CSSCounterStyleRule::setPad):
(WebCore::CSSCounterStyleRule::setFallback):
(WebCore::CSSCounterStyleRule::setSymbols):
(WebCore::CSSCounterStyleRule::setAdditiveSymbols):
(WebCore::CSSCounterStyleRule::setSpeakAs):
Implement setters and tangential functionality required by setters.
- css/CSSCounterStyleRule.h:
Replace FIXME with actual descriptor getter and setter
implementations.
- css/CSSProperties.json:
Add @counter-style descriptor properties.
- css/CSSValueKeywords.in:
Add new values required for
systemandspeak-as
@counter-style descriptor properties.
- css/parser/CSSParserContext.cpp:
(WebCore::CSSParserContext::isPropertyRuntimeDisabled const):
Ensure new @counter-style descriptors are disabled at runtime based
on CSSParserContext state.
- css/parser/CSSPropertyParser.cpp:
(WebCore::consumeCounterStyleSystem):
(WebCore::consumeCounterStyleSymbol):
(WebCore::consumeCounterStyleNegative):
(WebCore::consumeCounterStyleRangeBound):
(WebCore::consumeCounterStyleRange):
(WebCore::consumeCounterStylePad):
(WebCore::consumeCounterStyleSymbols):
(WebCore::consumeCounterStyleAdditiveSymbols):
(WebCore::consumeCounterStyleSpeakAs):
(WebCore::CSSPropertyParser::parseCounterStyleDescriptor):
Parse @counter-style descriptors.
Tools:
- DumpRenderTree/TestOptions.cpp:
(WTR::TestOptions::defaults):
Fix typo (missing 's'). CSSCounterStyleAtRulesEnabled, not
CSSCounterStyleAtRuleEnabled.
LayoutTests:
Add test ensuring <image> @counter-style symbol values cannot be
parsed when thecounterStyleAtRuleImageSymbolsEnabledfeature flag
is disabled.
---
This test is skipped on Windows because I haven't been able to get the
required feature flags (CSSCounterStyleAtRulesEnabled and
CSSCounterStyleAtRuleImageSymbolsEnabled) to work properly for that
port.
The code hidden behind these flags is all in the CSS parser, which is not
unique to Windows, so I think we can be confident that if the test passes
on all other platforms, that the behavior is correct on Windows too.
One attempt at implementing the necessary Windows-specific flag functionality is here:
https://bugs.webkit.org/attachment.cgi?id=426371&action=edit
Which failed to compile[1] with this error:
C:\cygwin\home\buildbot\worker\Windows-EWS\build\Tools\DumpRenderTree\win\DumpRenderTree.cpp(834,51): error C2039: 'setCSSCounterStyleAtRulesEnabled': is not a member of 'IWebPreferencesPrivate7' [C:\cygwin\home\buildbot\worker\Windows-EWS\build\WebKitBuild\Release\Tools\DumpRenderTree\DumpRenderTreeLib.vcxproj]
C:\cygwin\home\buildbot\worker\Windows-EWS\build\Tools\DumpRenderTree\win\DumpRenderTree.cpp(835,62): error C2039: 'setCSSCounterStyleAtRuleImageSymbolsEnabled': is not a member of 'IWebPreferencesPrivate7' [C:\cygwin\home\buildbot\worker\Windows-EWS\build\WebKitBuild\Release\Tools\DumpRenderTree\DumpRenderTreeLib.vcxproj]
Those methods are present in
IWebPreferencesPrivate7.idl, and implemented similarly to other
flags in other places (e.g. win/WebPreferences.{h, cpp}, win/WebPreferenceKeysPrivate.h).
I can't reproduce this compilation error on my Windows machine.
I then tried removing the lines that caused the above compilation failure.
Those setters are called in DumpRenderTree::enableExperimentalFeatures, so in
lieu of enabling these flags there I could enable the flag I need via test header.
That patch is: https://bugs.webkit.org/attachment.cgi?id=426509&action=edit
This results in successful compilation, but causes lots (all?) of the
layout tests to fail[2] with a stacktrace that looks like:
- 10:20 PM Changeset in webkit [276487] by
-
- 2 edits in trunk/WebDriverTests
[GTK][WebDriver] Garden keypress failures
Unreviewed test gardening.
- TestExpectations.json:
- 10:16 PM Changeset in webkit [276486] by
-
- 2 edits in trunk/Source/JavaScriptCore
[JSC][Win] callOperationNoExceptionCheck() also needs to support operations that return SlowPathReturnType
https://bugs.webkit.org/show_bug.cgi?id=224964
Reviewed by Yusuke Suzuki.
r229989 (Bug 183655) added the x64 Windows support only for
callOperation(), but for callOperationNoExceptionCheck().
callOperationNoExceptionCheck() also needs the x64 Windows
support.
This change is a preparation for Bug 224920 that is going to use
callOperationNoExceptionCheck instead of callOperation.
- jit/JIT.h:
(callOperation): Rewrote by using 'if constexpr' instead of SFINAE.
(callOperationNoExceptionCheck): Added a new implementation for
x64 Windows based on callOperation.
- 10:06 PM Changeset in webkit [276485] by
-
- 2 edits in trunk/Source/WTF
[WTF] Undefined behavior warning in StringBuilder::allocateBuffer
https://bugs.webkit.org/show_bug.cgi?id=224942
Reviewed by Chris Dumez.
- wtf/text/StringBuilder.cpp:
(WTF::StringBuilder::allocateBuffer): Add check before calling memcpy
- 8:56 PM Changeset in webkit [276484] by
-
- 3 edits2 adds in trunk
Sometimes appHighlights are not painted when created.
https://bugs.webkit.org/show_bug.cgi?id=224904
Reviewed by Antti Koivisto.
Source/WebCore:
In order for highlights to be painted correctly, we need to not use the new line layout code,
so we need to opt out when we are repainting due to a highlight being added or removed.
Test: http/wpt/css/css-highlight-api/highlight-text-dynamic.html
- Modules/highlight/Highlight.cpp:
(WebCore::repaintRange):
LayoutTests:
- http/wpt/css/css-highlight-api/highlight-text-dynamic-expected-mismatch.html: Added.
- http/wpt/css/css-highlight-api/highlight-text-dynamic.html: Added.
- 8:26 PM Changeset in webkit [276483] by
-
- 2 edits in trunk/LayoutTests
[GTK] Unreviewed test gardening. Update test expectations.
- platform/gtk/TestExpectations:
- 8:23 PM Changeset in webkit [276482] by
-
- 2 edits in trunk/Source/WebKit
[Mac] CMBaseClass object pointers can become unaligned on x86
https://bugs.webkit.org/show_bug.cgi?id=224950
<rdar://77020922>
Reviewed by Eric Carlson.
CMBaseClass has a 4-byte version member before its 8-byte pointers on x86. Deal with this
the same way we do with other pointer-bearing, static, CM-type objects: enforce a 4-byte
packing, and prepend the struct with another 4-byte object in order to force the pointers
into 8-byte alignment.
- Shared/mac/MediaFormatReader/CoreMediaWrapped.h:
(WebKit::CoreMediaWrapped<Wrapped>::vTable):
- 7:52 PM Changeset in webkit [276481] by
-
- 7 edits20 adds in trunk/LayoutTests
[GTK] Unreviewed test gardening. Update baselines after r276448.
r276448 turned on editable <input type=date> and <input type=time>,
many tests baselines need to be updated or emitted as result differ
from general baseline.
- platform/gtk/imported/w3c/web-platform-tests/html/dom/reflection-forms-expected.txt:
- platform/gtk/imported/w3c/web-platform-tests/html/rendering/widgets/baseline-alignment-and-overflow.tentative-expected.txt:
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-checkValidity-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-reportValidity-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-validity-badInput-expected.txt:
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-validity-rangeOverflow-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-validity-rangeUnderflow-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-validity-stepMismatch-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-validity-valid-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-validity-valueMissing-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/constraints/form-validation-willValidate-expected.txt:
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/date-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/datetime-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/datetime-local-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/input-seconds-leading-zeroes-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/input-valueasdate-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/input-valueasdate-stepping-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/input-valueasnumber-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/input-valueasnumber-stepping-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/month-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/selection-expected.txt:
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/time-2-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/time-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/valueMode-expected.txt:
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/forms/the-input-element/week-expected.txt: Added.
- platform/gtk/imported/w3c/web-platform-tests/html/semantics/selectors/pseudo-classes/inrange-outofrange-expected.txt: Added.
- 7:33 PM Changeset in webkit [276480] by
-
- 2 edits in trunk/Source/WebKit
Followup to r276397
https://bugs.webkit.org/show_bug.cgi?id=224883
Reviewed by Darin Adler.
Use a range-based for loop when mapping FloatQuads in
sendTapHighlightForNodeIfNecessary.
No change in behavior.
- WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::sendTapHighlightForNodeIfNecessary):
- 6:29 PM Changeset in webkit [276479] by
-
- 2 edits in trunk/LayoutTests
[ BigSur wk2 arm64 ] webrtc/captureCanvas-webrtc.html
https://bugs.webkit.org/show_bug.cgi?id=224963
Uneviewed test gardneing.
- platform/mac-wk2/TestExpectations: Updated test expectations to Pass Timeout for BigSur arm64 only while test is reviewed.
- 5:56 PM Changeset in webkit [276478] by
-
- 3 edits2 adds in trunk
Fix color gamut in WebRTC when using VPx software decoder.
https://bugs.webkit.org/show_bug.cgi?id=220972
<rdar://problem/73616455>
Reviewed by Youenn Fablet.
Source/WebCore:
By default, and how libwebrtc is configuring its encoders and decoder, video range levels are used
(16-235). Video range was assumed to use full range instead (0-255), causing blacks to look grey and
whites to appear washed-out.
It is theoretically possible, when using VP9 that full-range level are used. While the information
related to video levels is available at the VP9 codec level; only the color space information is
being made available in libwebrtc's public image data structure, color range isn't.
For now we can assume the default is sufficient, and a follow-up change will be submitted to libwebrtc
later on.
In memory of Dr. Alex. May he rest in peace.
Test: webrtc/video-vp8-videorange.html
- platform/mediastream/mac/RealtimeIncomingVideoSourceCocoa.mm:
(WebCore::RealtimeIncomingVideoSourceCocoa::pixelBufferPool):
LayoutTests:
- webrtc/video-vp8-videorange-expected.txt: Added.
- webrtc/video-vp8-videorange.html: Added.
- 5:42 PM Changeset in webkit [276477] by
-
- 4 edits in trunk/Source/WebKit
[iOS] Remove useless 'network-outbound' rule from non-networking file rule
https://bugs.webkit.org/show_bug.cgi?id=224940
<rdar://problem/66583476>
Reviewed by Per Arne Vollan.
No new tests. No effect on behavior.
- Resources/SandboxProfiles/ios/com.apple.WebKit.GPU.sb:
- Resources/SandboxProfiles/ios/com.apple.WebKit.Networking.sb:
- Resources/SandboxProfiles/ios/com.apple.WebKit.WebAuthn.sb:
- 5:09 PM Changeset in webkit [276476] by
-
- 14 edits in trunk
[Cocoa] re-enable test case WKInspectorDelegate.InspectorConfiguration
https://bugs.webkit.org/show_bug.cgi?id=224577
<rdar://70505272>
Reviewed by Devin Rousso.
Source/WebInspectorUI:
Adjust the CSP directive list in the <meta> tag to allow for customization
by WebKit ports.
- UserInterface/Main.html:
- Remove
default-src 'self'so as not to block custom scheme loads. - Remove
connect-src * ws:so as not to block custom scheme loads. - Add
object-src 'none'as we have no reason to allow <object>, <embed>, or <applet>.
Source/WebKit:
For the purposes of testing, we want to be able to issue a fetch() that will
hit a custom URL scheme handler registered with _WKInspectorConfiguration.
This is not allowed by the existing <meta> tag CSP directive list because 'connect-src *'
does not allow connecting to arbitrary schemes, just arbitrary domains.
To fix this, relax the 'connect-src' directive in Main.html and apply
a dynamically-computed CSP directive using the 'Content-Security-Policy' HTTP
response header. This is only sent for main resources (Main.html and Test.html)
using the newly added inspector-resource: URL scheme handler.
The dynamically computed directive explicitly allows 'self' and any other registered
custom URL scheme handlers. WebKit ports which have not migrated away from file:///
will only apply the weaker 'connect-src' directive from the <meta> tag after this change.
Progresses an existing API test: WKInspectorDelegate.InspectorConfiguration.
- UIProcess/API/Cocoa/_WKInspectorDelegate.h:
- UIProcess/API/APIInspectorClient.h:
(API::InspectorClient::frontendLoaded):
- UIProcess/API/Cocoa/_WKInspectorPrivateForTesting.h:
- UIProcess/API/Cocoa/_WKInspector.mm:
- UIProcess/API/Cocoa/_WKInspectorTesting.mm:
(JavaScriptSnippetToFetchURL):
(-[_WKInspector _fetchURLForTesting:]):
- UIProcess/Inspector/WebInspectorUIProxy.cpp:
(WebKit::WebInspectorUIProxy::frontendLoaded): Notify the _WKInspectorDelegate
adapter that the frontend has finished loading.
- UIProcess/Inspector/mac/WKInspectorResourceURLSchemeHandler.h:
- UIProcess/Inspector/mac/WKInspectorResourceURLSchemeHandler.mm:
(-[WKInspectorResourceURLSchemeHandler allowedURLSchemesForCSP]):
(-[WKInspectorResourceURLSchemeHandler setAllowedURLSchemesForCSP:]):
(-[WKInspectorResourceURLSchemeHandler webView:startURLSchemeTask:]):
Added. Keep track of allowed custom schemes and allowed main resources.
Apply the CSP directive for main resource requests only.
- UIProcess/Inspector/mac/WKInspectorViewController.mm:
(-[WKInspectorViewController webViewConfiguration]):
Set the allowed URL schemes property so the URL scheme handler can include
the schemes in the dynamically computed 'connect-src' directive.
Tools:
Trigger a fetch of a resource that uses a custom URL scheme handler in
order to test that custom scheme handlers registered in an _WKInspectorConfiguration
are getting used as expected.
Add an -inspectorFrontendLoaded: delegate method to -WKInspectorDelegate.
This is mainly used to prevent tests from progressing with a half-loaded inspector.
To make it possible to do this fetch(), there are some CSP related changes
that were made for Main.html so that fetching from a registered custom scheme is allowed.
Drive-by, per post-commit comments, add a better fix for the memory leak reported in
https://bugs.webkit.org/show_bug.cgi?id=223899. Thanks Joe!
- TestWebKitAPI/Tests/WebKitCocoa/WKInspectorDelegate.mm:
(resetGlobalState):
(-[InspectorDelegate inspectorFrontendLoaded:]):
(TEST):
(-[UIDelegate _webView:didAttachLocalInspector:]):
(-[UIDelegate _webView:willCloseLocalInspector:]):
(-[UIDelegate _webViewDidEnableInspectorBrowserDomain:]):
(-[UIDelegate _webViewDidDisableInspectorBrowserDomain:]):
- 4:04 PM Changeset in webkit [276475] by
-
- 2 edits in trunk/Tools
Unreviewed. Gardening TestSSL flakiness
- TestWebKitAPI/glib/TestExpectations.json:
- 4:04 PM Changeset in webkit [276474] by
-
- 1 copy in tags/Safari-612.1.11.5
Tag Safari-612.1.11.5.
- 3:33 PM Changeset in webkit [276473] by
-
- 9 edits in trunk
Add trace points for WKWebView snapshotting
https://bugs.webkit.org/show_bug.cgi?id=224943
Reviewed by Tim Horton.
Source/WebKit:
Add new trace points to make snapshotting-related issues easier to diagnose. In the UI
process we start/end a trace scope when requested to snapshot, and before calling the
client's completion handler.
Trace CARenderServer snapshots which happen in the UI process, and in the web process, trace
the start and end of the snapshot painting.
- UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView takeSnapshotWithConfiguration:completionHandler:]):
- UIProcess/API/Cocoa/_WKThumbnailView.mm:
(-[_WKThumbnailView requestSnapshot]):
- UIProcess/API/ios/WKWebViewIOS.mm:
(-[WKWebView _snapshotRect:intoImageOfWidth:completionHandler:]):
- WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::paintSnapshotAtSize):
Source/WTF:
New snapshot values.
- wtf/SystemTracing.h:
Tools:
New snapshot descriptions.
- Tracing/SystemTracePoints.plist:
- 3:31 PM Changeset in webkit [276472] by
-
- 2 edits in trunk/Source/WebKit
Unreviewed. Guard doesPageNeedTCCD in platform COCOA after r274213
https://bugs.webkit.org/show_bug.cgi?id=224951
Silence -Wunused-function in non-cocoa platforms.
- UIProcess/UserMediaPermissionRequestManagerProxy.cpp:
- 3:31 PM Changeset in webkit [276471] by
-
- 2 edits in trunk/Source/WebCore
[SOUP] Null pointer dereference in ResourceLoader::didReceiveAuthenticationChallenge
https://bugs.webkit.org/show_bug.cgi?id=224838
Patch by Michael Catanzaro <Michael Catanzaro> on 2021-04-22
Reviewed by Adrian Perez de Castro.
Add a RELEASE_ASSERT_NOT_REACHED() to avoid this warning.
- platform/network/soup/AuthenticationChallenge.h:
- 3:24 PM Changeset in webkit [276470] by
-
- 5 edits in trunk/Source/JavaScriptCore
Unreviewed, reverting r276456.
https://bugs.webkit.org/show_bug.cgi?id=224952
Windows specific crash
Reverted changeset:
"[JSC} Remove exception checks from non-throwing function
calls in Baseline JIT"
https://bugs.webkit.org/show_bug.cgi?id=224920
https://trac.webkit.org/changeset/276456
- 3:23 PM Changeset in webkit [276469] by
-
- 4 edits in trunk/Source
[SOUP] Show more connection information in the inspector
https://bugs.webkit.org/show_bug.cgi?id=224873
Source/WebCore:
Handle SOUP_HTTP_2_0.
Reviewed by Adrian Perez de Castro.
- platform/network/soup/ResourceResponseSoup.cpp:
(WebCore::ResourceResponse::ResourceResponse):
Source/WebKit:
Show the h2 protocol as well as the connection identifier
in the inspector.
Reviewed by Adrian Perez de Castro.
- NetworkProcess/soup/NetworkDataTaskSoup.cpp:
(WebKit::NetworkDataTaskSoup::gotHeadersCallback):
(WebKit::soupHTTPVersionToString):
(WebKit::NetworkDataTaskSoup::didGetHeaders):
- 3:16 PM Changeset in webkit [276468] by
-
- 2 edits in trunk/Source/WebKitLegacy/win
Disable prefixed WebAudio on WebKitLegacy-Windows
https://bugs.webkit.org/show_bug.cgi?id=224944
Reviewed by Eric Carlson.
Disable prefixed WebAudio on WebKitLegacy-Windows. It is already disabled everywhere else.
- WebView.cpp:
(WebView::notifyPreferencesChanged):
- 3:12 PM Changeset in webkit [276467] by
-
- 2 edits in trunk/LayoutTests
REGRESSION (Metal ANGLE): webgl/2.0.0/conformance/glsl/misc/shader-uniform-packing-restrictions.html is constantly timing out.
https://bugs.webkit.org/show_bug.cgi?id=224257
Unreviewed test gardening.
- platform/mac-wk2/TestExpectations: Updating test that's timing out constantly to slow to see if it improves.
- 3:09 PM Changeset in webkit [276466] by
-
- 2 edits in trunk/Source/ThirdParty/ANGLE
[CMake][ANGLE] Silence -Wreturn-type for gcc and clang
https://bugs.webkit.org/show_bug.cgi?id=224949
Reviewed by Don Olmstead.
Avoid "control reaches end of non-void function" in
sh::TIntermRebuild::traverseChildren(...)
- CMakeLists.txt:
- 3:07 PM Changeset in webkit [276465] by
-
- 10 edits in trunk
Add discrete animation support for border-image-repeat
https://bugs.webkit.org/show_bug.cgi?id=224923
Reviewed by Antti Koivisto.
LayoutTests/imported/w3c:
Mark 10 WPT progressions.
- web-platform-tests/web-animations/animation-model/animation-types/accumulation-per-property-001-expected.txt:
- web-platform-tests/web-animations/animation-model/animation-types/addition-per-property-001-expected.txt:
- web-platform-tests/web-animations/animation-model/animation-types/interpolation-per-property-001-expected.txt:
Source/WebCore:
- animation/CSSPropertyAnimation.cpp:
(WebCore::CSSPropertyAnimationWrapperMap::CSSPropertyAnimationWrapperMap):
- rendering/style/NinePieceImage.cpp:
(WebCore::operator<<):
- rendering/style/NinePieceImage.h:
- rendering/style/RenderStyle.cpp:
(WebCore::RenderStyle::setBorderImageHorizontalRule):
(WebCore::RenderStyle::setBorderImageVerticalRule):
- rendering/style/RenderStyle.h:
(WebCore::RenderStyle::borderImageHorizontalRule const):
(WebCore::RenderStyle::borderImageVerticalRule const):
- 3:02 PM Changeset in webkit [276464] by
-
- 3 edits2 adds in trunk
[RenderTreeBuilder] Multi-column spanners are not part of the enclosing multi-column renderer's subtree
https://bugs.webkit.org/show_bug.cgi?id=224914
<rdar://76846946>
Reviewed by Simon Fraser.
Source/WebCore:
This patch handles the case when the newly inserted renderer is a column spanner and checking enclosingFragmentedFlow()
on the spanner's new parent returns the incorrect enclosing multi-column renderer.
This can result in having the spanner in multiple multi-column flows.
(Note that the spanners are directly attached to the RenderBlockFlow that establishes the multi-column context, while non-spanner
renderers are part of the RenderMultiColumnFlow subtree. e.g
RenderBlockFlow <- spanner's parent.
RenderMultiColumnFlow
Non-spanner renderer
spanner renderer
calling enclosingFragmentedFlow() on a spanner's parent (RenderBlockFlow) returns the parent multi-column flow -and
not the one it establishes. e.g.
RenderMultiColumnFlow <- parent multi-column flow -in case of nested columns.
RenderBlockFlow <- spanner's parent.
RenderMultiColumnFlow
Non-spanner renderer
spanner renderer
)
Test: fast/multicol/spanner-inserted-to-multiple-multicolumns.html
- rendering/updating/RenderTreeBuilder.cpp:
(WebCore::RenderTreeBuilder::childFlowStateChangesAndAffectsParentBlock):
LayoutTests:
- fast/multicol/spanner-inserted-to-multiple-multicolumns-expected.txt: Added.
- fast/multicol/spanner-inserted-to-multiple-multicolumns.html: Added.
- 2:56 PM Changeset in webkit [276463] by
-
- 2 edits in trunk/Source/WebCore
[WebXR] Optional button test is incorrect
https://bugs.webkit.org/show_bug.cgi?id=224945
Reviewed by Tim Horton.
Fix a conditional test that was always true.
- testing/WebFakeXRInputController.cpp:
(WebCore::WebFakeXRInputController::getButtonOrPlaceholder const):
(A != B A != C) -> (A != B && A != C) - 2:50 PM Changeset in webkit [276462] by
-
- 2 edits in trunk/Source/WebCore
[WebXR] Closure is capturing this unnecessarily
https://bugs.webkit.org/show_bug.cgi?id=224946
Reviewed by Tim Horton.
Remove the "this" capture.
- Modules/webxr/WebXRInputSourceArray.cpp:
(WebCore::WebXRInputSourceArray::handleRemovedInputSources):
- 2:19 PM Changeset in webkit [276461] by
-
- 2 edits in trunk/Source/WebKit
NowPlayingInfoForGPUManager::clearNowPlayingInfoPrivate() sometimes launches the GPUProcess on page close
https://bugs.webkit.org/show_bug.cgi?id=224939
Reviewed by Simon Fraser.
Make sure NowPlayingInfoForGPUManager::clearNowPlayingInfoPrivate() never launches the GPUProcess. The GPU
process already takes care of clearing now playing info on exit.
- WebProcess/GPU/media/WebMediaStrategy.cpp:
(WebKit::WebMediaStrategy::createNowPlayingManager const):
- 1:52 PM Changeset in webkit [276460] by
-
- 5 edits in trunk/Source/WebKit
Call RemoteMediaEngineConfigurationFactory::registerFactory() in WebProcess::setUseGPUProcessForMedia()
https://bugs.webkit.org/show_bug.cgi?id=224937
Reviewed by Eric Carlson.
Call RemoteMediaEngineConfigurationFactory::registerFactory() in WebProcess::setUseGPUProcessForMedia(),
like we used to do, instead of doing it in the GPUProcessConnection constructor. It is safer to
register the factory before we have a GPUProcessConnection and we wouldn't want things to behave
differently depending on whether or not the GPUProcess is already launched. I made sure that this does
not eagerly launch the GPUProcess.
- WebProcess/GPU/GPUProcessConnection.cpp:
(WebKit::GPUProcessConnection::GPUProcessConnection):
(WebKit::GPUProcessConnection::mediaEngineConfigurationFactory): Deleted.
- WebProcess/GPU/GPUProcessConnection.h:
- WebProcess/WebProcess.cpp:
(WebKit::WebProcess::setUseGPUProcessForMedia):
(WebKit::WebProcess::mediaEngineConfigurationFactory):
- WebProcess/WebProcess.h:
- 1:36 PM Changeset in webkit [276459] by
-
- 2 edits in trunk/Source/WebKit
Fix ENABLE macro around NetworkProcessConnection::connectToRTCDataChannelRemoteSource declaration
https://bugs.webkit.org/show_bug.cgi?id=224938
Patch by Alex Christensen <achristensen@webkit.org> on 2021-04-22
Reviewed by Chris Dumez.
This matches the macro around the definition.
- WebProcess/Network/NetworkProcessConnection.h:
- 1:25 PM Changeset in webkit [276458] by
-
- 4 edits2 deletes in trunk
Unreviewed, reverting r271644.
https://bugs.webkit.org/show_bug.cgi?id=224908
Caused regression
<http://webkit.org/b/221962|webkit.org/b/221962>
Reverted changeset:
"[css-multicol] OOM with 1px height columns"
https://bugs.webkit.org/show_bug.cgi?id=220490
https://trac.webkit.org/changeset/271644
- 1:22 PM Changeset in webkit [276457] by
-
- 2 edits in trunk/Tools
Add "Tim Nguyen" to contributors.json.
https://bugs.webkit.org/show_bug.cgi?id=224936
Patch by Tim Nguyen <ntim@apple.com> on 2021-04-22
Reviewed by Alexey Proskuryakov.
- Scripts/webkitpy/common/config/contributors.json:
- 1:17 PM Changeset in webkit [276456] by
-
- 5 edits in trunk/Source/JavaScriptCore
[JSC} Remove exception checks from non-throwing function calls in Baseline JIT
https://bugs.webkit.org/show_bug.cgi?id=224920
Reviewed by Tadeu Zagallo.
These functions are not taking JSGlobalObject and will not throw an error.
Use callOperationNoExceptionCheck instead to avoid emitting unnecessary exception checks.
- jit/JIT.cpp:
(JSC::JIT::emitEnterOptimizationCheck):
- jit/JITOpcodes.cpp:
(JSC::JIT::emitSlow_op_new_object):
(JSC::JIT::emit_op_catch):
(JSC::JIT::emit_op_switch_imm):
(JSC::JIT::emitSlow_op_loop_hint):
(JSC::JIT::emit_op_profile_type):
- jit/JITOpcodes32_64.cpp:
(JSC::JIT::emitSlow_op_new_object):
(JSC::JIT::emit_op_catch):
(JSC::JIT::emit_op_switch_imm):
(JSC::JIT::emit_op_profile_type):
- jit/JITPropertyAccess.cpp:
(JSC::JIT::emitWriteBarrier):
- 1:01 PM Changeset in webkit [276455] by
-
- 2 edits in trunk/Tools
[Win][WTR] eventSender.leapForward doesn't work as expected
https://bugs.webkit.org/show_bug.cgi?id=224918
Reviewed by Don Olmstead.
fast/events/click-count.html was failing because
eventSender.leapForward didn't work as expected.
Use Sleep() API to implement it as well as DRT does.
- WebKitTestRunner/win/EventSenderProxyWin.cpp:
(WTR::EventSenderProxy::dispatchMessage):
(WTR::EventSenderProxy::EventSenderProxy):
(WTR::EventSenderProxy::leapForward):
- 12:48 PM Changeset in webkit [276454] by
-
- 4 edits in trunk
v2: REGRESSION(r266890): [Cocoa] Fix API::InspectorClient leak
https://bugs.webkit.org/show_bug.cgi?id=223899
<rdar://problem/75249282>
Reviewed by Devin Rousso.
Address post-review feedback.
Source/WebKit:
- UIProcess/API/Cocoa/_WKInspectorTesting.mm:
(-[_WKInspector _openURLExternallyForTesting:useFrontendAPI:]):
Tools:
- TestWebKitAPI/Tests/WebKitCocoa/WKInspectorDelegate.mm:
(TEST):
- 12:32 PM Changeset in webkit [276453] by
-
- 2 edits in trunk/Source/WebKit
Add some new messages.in files to the Mac CMake build
- PlatformMac.cmake:
- 12:21 PM Changeset in webkit [276452] by
-
- 2 edits in trunk/LayoutTests
Unreviewed, reverting r276393.
This expectation is needed so that the test will pass on bots
Reverted changeset:
"Mass removing prior test expectations that don't need to be
set anymore"
https://bugs.webkit.org/show_bug.cgi?id=224898
https://commits.webkit.org/r276393
- 12:19 PM Changeset in webkit [276451] by
-
- 2 edits in trunk/Source/WebKit
[iOS] Update sandbox message filter syntax
https://bugs.webkit.org/show_bug.cgi?id=223384
<rdar://problem/75531614>
Reviewed by Brent Fulgham.
- Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb:
- 11:44 AM Changeset in webkit [276450] by
-
- 61 edits2 adds in trunk
Implement FontFace in Workers for OffscreenCanvas
https://bugs.webkit.org/show_bug.cgi?id=224178
<rdar://problem/76529024>
Reviewed by Darin Adler.
LayoutTests/imported/w3c:
FontFace is now available for Workers, rebaseline related tests.
- web-platform-tests/css/css-font-loading/fontfaceset-load-css-wide-keywords-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.align.center.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.align.end.ltr.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.align.end.rtl.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.align.left.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.align.right.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.align.start.ltr.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.align.start.rtl.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.baseline.alphabetic.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.baseline.bottom.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.baseline.hanging.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.baseline.ideographic.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.baseline.middle.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.baseline.top.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.fill.maxWidth.bound.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.fill.maxWidth.fontface.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.fontface.notinpage.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.fontface.repeat.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.fontface.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.space.basic.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.space.collapse.end.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.space.collapse.nonspace.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.space.collapse.other.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.space.collapse.space.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.draw.space.collapse.start.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.measure.actualBoundingBox.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.measure.advances.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.measure.baselines.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.measure.emHeights.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.measure.fontBoundingBox.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.measure.width.basic.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.measure.width.empty.worker-expected.txt:
- web-platform-tests/html/canvas/offscreen/text/2d.text.measure.width.space.worker-expected.txt:
Source/WebCore:
To comply with spec and complete the OffscreenCanvas implementation,
FontFace needs to be made available on WorkerGlobalScope. This patch
makes it available and fixes crashes using it due to accessing
main-thread only singletons.
No new tests, rebaselined existing tests.
- Headers.cmake:
- Sources.txt:
- WebCore.xcodeproj/project.pbxproj: Add new files WorkerFontLoadRequest.cpp, WorkerFontLoadRequest.h
- css/CSSFontFace.cpp:
(WebCore::CSSFontFace::fontCacheFallingBackToSingleton):
(WebCore::CSSFontFace::font):
- css/CSSFontFace.h: Use the ScriptExecutionContext's FontCache if possible when loading fonts.
- css/CSSFontFaceSet.cpp:
(WebCore::CSSFontFaceSet::ensureLocalFontFacesForFamilyRegistered):
(WebCore::CSSFontFaceSet::matchingFacesExcludingPreinstalledFonts):
Use the correct CSSValuePool and FontCache in
ensureLocalFontFacesForFamilyRegistered. Also clarify a vaguely-named
variable.
- css/CSSFontFaceSource.cpp:
(WebCore::CSSFontFaceSource::CSSFontFaceSource):
Use FontLoadRequest::isPending instead of making assumptions about the
request state on construction.
- css/FontFace.cpp:
(WebCore::FontFace::create):
(WebCore::FontFace::display const):
- css/FontFace.h:
- css/FontFace.idl: Construct with a ScriptExecutionContext instead of a Document and make available for Workers.
- css/FontFaceSet.idl: Expose to Workers.
- loader/FontLoadRequest.h:
(WebCore::FontLoadRequest::isWorkerFontLoadRequest const):
Add isPending function and new type WorkerFontLoadRequest.
- loader/cache/CachedFont.cpp:
(WebCore::CachedFont::createCustomFontData):
Use convertWOFFToSfntIfNecessary().
- loader/cache/CachedFontLoadRequest.h: Implement FontLoadRequest::isPending.
- platform/graphics/Font.cpp:
(WebCore::Font::create):
(WebCore::Font::Font):
(WebCore::Font::systemFallbackFontForCharacter const):
- platform/graphics/Font.h: Add an optional FontCache parameter for use with vertical font data retrieval and add a FontCache parameter to systemFallbackFontForCharacter instead of using the singleton.
- platform/graphics/FontCache.cpp:
(WebCore::FontCache::fontForPlatformData):
Pass FontCache parameter when creating fonts.
- platform/graphics/FontCache.h:
(WebCore::FontCache::fontCacheFallingBackToSingleton):
- platform/graphics/FontCascade.cpp:
(WebCore::FontCascade::update const):
- platform/graphics/cairo/FontCairoHarfbuzzNG.cpp:
(WebCore::FontCascade::fontForCombiningCharacterSequence const):
Rename fontCacheFallbackToSingleton -> fontCacheFallingBackToSingleton
- platform/graphics/FontCascadeFonts.cpp:
(WebCore::FontCascadeFonts::realizeFallbackRangesAt):
(WebCore::FontCascadeFonts::glyphDataForSystemFallback):
Use familyNamesData instead of familyNames and pass the FontCache
parameter when calling glyphDataForSystemFallback to make this safe
to use in a Worker.
- platform/graphics/FontTaggedSettings.h: Add missing wtf/Hasher.h include.
- platform/graphics/WOFFFileFormat.cpp:
(WebCore::convertWOFFToSfntIfNecessary):
- platform/graphics/WOFFFileFormat.h: Add utility function to convert WOFF fonts on non-Cocoa platforms.
- workers/WorkerFontLoadRequest.cpp: Added.
(WebCore::WorkerFontLoadRequest::WorkerFontLoadRequest):
(WebCore::WorkerFontLoadRequest::load):
(WebCore::WorkerFontLoadRequest::ensureCustomFontData):
(WebCore::WorkerFontLoadRequest::createFont):
(WebCore::WorkerFontLoadRequest::setClient):
(WebCore::WorkerFontLoadRequest::didReceiveResponse):
(WebCore::WorkerFontLoadRequest::didReceiveData):
(WebCore::WorkerFontLoadRequest::didFinishLoading):
(WebCore::WorkerFontLoadRequest::didFail):
- workers/WorkerFontLoadRequest.h: Added. Implement FontLoadRequest for Workers.
- workers/WorkerGlobalScope.cpp:
(WebCore::WorkerGlobalScope::fonts):
(WebCore::WorkerGlobalScope::fontLoadRequest):
(WebCore::WorkerGlobalScope::beginLoadingFontSoon):
- workers/WorkerGlobalScope.h: Implement ScriptExecutionContext font-loading related functions.
- workers/WorkerGlobalScope.idl: Include FontFaceSource.
- 11:41 AM Changeset in webkit [276449] by
-
- 5 edits1 add in trunk
[JSC] Baseline should have fast path for switch_imm
https://bugs.webkit.org/show_bug.cgi?id=224521
Reviewed by Tadeu Zagallo.
JSTests:
- stress/switch-imm-baseline.js: Added.
(shouldBe):
(test):
Source/JavaScriptCore:
This patch implements op_switch_imm fast path in Baseline.
We have this fast path in LLInt, DFG, and FTL. So only Baseline lacks this.
- dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::emitSwitchIntJump):
- jit/JITOpcodes.cpp:
(JSC::JIT::emit_op_switch_imm):
- jit/JITOpcodes32_64.cpp:
(JSC::JIT::emit_op_switch_imm):
- 11:09 AM Changeset in webkit [276448] by
-
- 13 edits7 adds in trunk
[GTK] Turn on editable <input type=date> and <input type=time> fields
https://bugs.webkit.org/show_bug.cgi?id=224921
Reviewed by Carlos Garcia Campos.
.:
Turn on editable date and time inputs for GTK
- Source/cmake/OptionsGTK.cmake: Add new source file to the list.
Source/WebCore:
Turn on editable date and time inputs for GTK.
No new tests. This is covered by enabling existing <input type=date>
and <input type=time> tests.
- Headers.cmake: Add headers to the list.
- loader/EmptyClients.cpp: Add missing header include.
Source/WebKit:
Turn on editable date and time inputs for GTK.
- SourcesGTK.txt: Add source file.
- UIProcess/API/gtk/PageClientImpl.cpp:
(WebKit::PageClientImpl::createDateTimePicker): Added.
- UIProcess/API/gtk/PageClientImpl.h: Added method declaration.
- UIProcess/gtk/WebDateTimePickerGtk.cpp: Added. Stub for date/time picker which
doesn't have any user interface. This will allow clients to use the editable shadow
DOM date/time input editable fields.
(WebKit::WebDateTimePickerGtk::create):
(WebKit::WebDateTimePickerGtk::~WebDateTimePickerGtk):
(WebKit::WebDateTimePickerGtk::WebDateTimePickerGtk):
(WebKit::WebDateTimePickerGtk::endPicker):
(WebKit::WebDateTimePickerGtk::showDateTimePicker):
- UIProcess/gtk/WebDateTimePickerGtk.h: Added.
Source/WTF:
Turn on editable date and time inputs for GTK.
- Scripts/Preferences/WebPreferencesInternal.yaml: Set editable date/time inputs
on by default for GTK as well.
LayoutTests:
Turn on editable date and time inputs for GTK.
- platform/gtk/TestExpectations: Enable new forms test for WebKitGTK.
- platform/gtk/fast/forms/date/date-input-rendering-basic-expected.txt: Added.
- platform/gtk/fast/forms/date/date-pseudo-elements-expected.txt: Added.
- platform/gtk/fast/forms/time/time-input-rendering-basic-expected.txt: Added.
- 11:05 AM Changeset in webkit [276447] by
-
- 5 edits in trunk/Source
Fix the tvOS build after r276325
https://bugs.webkit.org/show_bug.cgi?id=224929
<rdar://problem/77003408>
Reviewed by Wenson Hsieh.
Source/WebKit:
- Platform/spi/ios/UIKitSPI.h:
- UIProcess/ios/forms/WKDateTimeInputControl.mm:
[UIColor systemBackgroundColor] is SPI on tvOS.
(-[WKDateTimePickerViewController viewDidLoad]):
UIBlurEffectStyleSystemMaterial is not available on tvOS. Instead, use
a UIView with a system background color as the background view.
Source/WTF:
Fix the tvOS build after r276325
- wtf/PlatformHave.h:
- 11:00 AM Changeset in webkit [276446] by
-
- 2 edits in trunk/Source/WebKit
v2: Web Inspector: exempt API::SharedJSContext from remote inspection and automatic inspection
https://bugs.webkit.org/show_bug.cgi?id=224841
<rdar://69386559>
Reviewed by Devin Rousso.
Use the RemoteInspector C SPI to temporarily turn off "allow remote inspection by default".
- UIProcess/API/Cocoa/APISerializedScriptValueCocoa.mm:
(API::SharedJSContext::ensureContext):
- 10:12 AM Changeset in webkit [276445] by
-
- 4 edits in trunk
[css-grid] WPT Tests css/css-grid/grid-items/grid-item-percentage-sizes-*.html fail
https://bugs.webkit.org/show_bug.cgi?id=209650
Patch by Ziran Sun <Ziran Sun> on 2021-04-22
Reviewed by Sergio Villar Senin.
Source/WebCore:
We need to check child's writing mode in order to pick the correct dimension to resolve
percentage height against. Overriding containing block size should also be chosen in favour
of special code for orthogonal writing mode.
This change is an import of Chromium CL at
https://chromium-review.googlesource.com/c/chromium/src/+/1348055.
Three tests that failed are now passing.
- rendering/RenderBox.cpp:
(WebCore::RenderBox::computePercentageLogicalHeight const):
LayoutTests:
- TestExpectations: grid-item-percentage-sizes-00{1-3}.html are passing and unskipped now.
- 9:52 AM Changeset in webkit [276444] by
-
- 1 copy in tags/Safari-611.1.21.161.5
Tag Safari-611.1.21.161.5.
- 9:44 AM Changeset in webkit [276443] by
-
- 8 edits in branches/safari-611.1.21.161-branch/Source
Versioning.
WebKit-7611.1.21.161.5
- 9:43 AM Changeset in webkit [276442] by
-
- 3 edits in trunk/Tools
Ensure all non-local AutoInstalled libraries specify version
https://bugs.webkit.org/show_bug.cgi?id=224872
Reviewed by Jonathan Bedard.
Currently on Python 3 running test-webkitpy fails due to us attempting to
install flask 2.0.0rc1, recently released. We shouldn't, however, be installing
the latest possible version of flask, but rather specifying one. We do actually
attempt to specify one in webkitflaskpy, but we pass the Version object to the
wrong place.
Now actually specifying the versions, we need to change some of them to the
latest version which supports Python 2 to be able to install them to run
test-webkitpy-python2.
To ensure we don't repeat this, this makes AutoInstall.register raise if you try
to register a non-local package without specifying a version.
Additionally, require AutoInstall.register's package argument be a Package by
passing moving the lookup to AutoInstall.install and requiring it to be
successful.
Finally, while I'm touching this, make registering 'autoinstalled' an outright
failure, and make installing a local package fail if it cannot be found.
- Scripts/libraries/webkitcorepy/webkitcorepy/autoinstall.py:
(AutoInstall.register):
(AutoInstall.install):
- Scripts/libraries/webkitflaskpy/webkitflaskpy/init.py:
- 9:38 AM Changeset in webkit [276441] by
-
- 5 edits1 add in trunk
[iOS] The Look Up text service popover should avoid covering selected text
https://bugs.webkit.org/show_bug.cgi?id=224915
rdar://75891746
Reviewed by Megan Gardner.
Source/WebKit:
Adjust the presentation rect when showing the Look Up text service popup using the callout bar on iOS so that
we use the entire bounds of the selection. Currently, we only use the first selection rect (top to bottom) as
the presentation rect, which may cause the popup to obscure selected text.
This seems to have been derived from iOS legacy web view behavior that dating back to rdar://7475904; while the
exact reasoning for this is effectively lost to time, it's likely due to how there used to only be a single
block selection rect in non-editable content in iOS 5 anyways, so taking only the first selection rect would not
be incorrect. However, we now use character granularity selection everywhere in WebKit2, so this logic is no
longer sensible.
Test: TextServicesTests.LookUpPresentationRect
- UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView _lookupForWebView:]):
(-[WKContentView _translateForWebView:]):
Make a similar adjustment here, since this presentation rect is intended to match that of the Look Up popover.
Tools:
Add a new API test to exercise the change.
- TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
- TestWebKitAPI/Tests/ios/TextServicesTests.mm: Added.
(TestWebKitAPI::handleLookup):
(TestWebKitAPI::TEST):
- TestWebKitAPI/ios/UIKitSPI.h: Add SPI declarations for UIWKTextInteractionAssistant.
- 9:35 AM Changeset in webkit [276440] by
-
- 6 edits in branches/safari-612.1.11-branch/Source
Cherry-pick r275899. rdar://problem/76986496
Revert r275172
https://bugs.webkit.org/show_bug.cgi?id=224499
<rdar://76396241>
Patch by Alex Christensen <achristensen@webkit.org> on 2021-04-13
Reviewed by Tim Horton.
Source/WebKit:
It turns out that terminate_with_reason uses a different syscall than xpc_connection_kill in its implementation,
and the former is not allowed by Messages's sandbox yet. Revert for Messages until they add it.
- Platform/IPC/cocoa/ConnectionCocoa.mm: (IPC::ConnectionTerminationWatchdog::watchdogTimerFired): (IPC::Connection::kill):
- Platform/spi/Cocoa/ReasonSPI.h:
- UIProcess/Launcher/mac/ProcessLauncherMac.mm: (WebKit::ProcessLauncher::platformInvalidate): (WebKit::terminateWithReason):
Source/WTF:
- wtf/spi/darwin/XPCSPI.h:
git-svn-id: https://svn.webkit.org/repository/webkit/trunk@275899 268f45cc-cd09-0410-ab3c-d52691b4dbfc
- 9:34 AM Changeset in webkit [276439] by
-
- 2 edits in trunk/Source/WebKit
[macOS] Re-block access to SYS_setattrlist
https://bugs.webkit.org/show_bug.cgi?id=224892
<rdar://problem/74828532>
Reviewed by Per Arne Vollan.
No new tests. No change in behavior.
- WebProcess/com.apple.WebProcess.sb.in:
- 9:12 AM Changeset in webkit [276438] by
-
- 3 edits in trunk/Tools
Services EWS should run for webkitscmpy and webkitcorepy changes
https://bugs.webkit.org/show_bug.cgi?id=223941
Reviewed by Jonathan Bedard.
- CISupport/ews-build/steps.py:
(CheckPatchRelevance):
- CISupport/ews-build/steps_unittest.py: Added unit-test.
- 9:04 AM Changeset in webkit [276437] by
-
- 5 edits in trunk/Source/WebKit
Call RemoteCDMFactory::registerFactory() in WebProcess::setUseGPUProcessForMedia()
https://bugs.webkit.org/show_bug.cgi?id=224907
Reviewed by Eric Carlson.
Call RemoteCDMFactory::registerFactory() in WebProcess::setUseGPUProcessForMedia() like we
did before r275887, instead of doing it in the GPUProcessConnection constructor. This is
safer as it makes sure the CDMFactory gets registered even if the GPUProcess has not launched
yet. This is also more consistent with what we do for LegacyCDM since Bug 224905.
The MediaKeys API still seems functional after r275887 (unlike the LegacyCDM one) so it does
not seem it caused obvious breakage. That said, it is still safer to restore pre-r275887
behavior. There is also no drawback to restoring the previous behavior since I made sure the
GPUProcess does not get launched eagerly.
- WebProcess/GPU/GPUProcessConnection.cpp:
(WebKit::GPUProcessConnection::GPUProcessConnection):
(WebKit::GPUProcessConnection::dispatchMessage):
- WebProcess/GPU/GPUProcessConnection.h:
- WebProcess/WebProcess.cpp:
(WebKit::WebProcess::setUseGPUProcessForMedia):
(WebKit::WebProcess::cdmFactory):
- WebProcess/WebProcess.h:
- 8:17 AM Changeset in webkit [276436] by
-
- 4 edits1 move1 delete in trunk/Tools
Add an xfail marker for webkitpy's tests; get SCM passing
https://bugs.webkit.org/show_bug.cgi?id=224886
Reviewed by Jonathan Bedard.
Also:
Delete the old, unused skip_if decorator.
Fix our runner to detect unexpected-success and report them as failures.
- Scripts/webkitpy/common/checkout/scm/scm_unittest.py:
(GitSVNTest.test_rebase_in_progress):
(GitTestWithMock.test_create_patch):
- Scripts/webkitpy/test/markers.py:
(xfail):
(xfail.decorator):
- Scripts/webkitpy/test/runner.py:
(Runner.handle):
(_Worker.handle):
(TestResult):
(TestResult.init):
(TestResult.addSuccess):
- Scripts/webkitpy/test/runner_unittest.py:
(FakeTestCase):
(FakeTestCase.init):
(FakeTestCase.id):
(FakeModuleSuite.run):
- Scripts/webkitpy/test/skip.py: Removed.
- Scripts/webkitpy/test/skip_unittest.py: Removed.
- 7:48 AM Changeset in webkit [276435] by
-
- 2 edits in trunk/Tools
server_process_unittest.py's MockFile should all be bytes
https://bugs.webkit.org/show_bug.cgi?id=224877
Reviewed by Jonathan Bedard.
This fixes the test when run under pytest on Python 3.
- Scripts/webkitpy/port/server_process_unittest.py:
(MockFile.read):
- 7:09 AM Changeset in webkit [276434] by
-
- 6 edits in trunk/Tools
Fix potential flakiness when running webkitpy tests
https://bugs.webkit.org/show_bug.cgi?id=224887
Reviewed by Jonathan Bedard.
- Scripts/webkitpy/common/system/executive_unittest.py:
(ExecutiveTest.serial_test_run_in_parallel): Increase the delay to ensure it's
much greater than the VM spawn time, as the current time is a bit marginal when
the system is under load and we're using the spawn multiprocessing start method
- Scripts/webkitpy/test/main.py:
(Tester.run): Allow argv to be passed through for testing
- Scripts/webkitpy/test/main_unittest.py:
(TesterTest.test_no_tests_found): Explicitly pass argv to avoid using sys.argv
- Scripts/webkitpy/w3c/test_parser_unittest.py:
(TestParserTest.test_analyze_pixel_test_all_true): Reset options after test
(TestParserTest.test_analyze_pixel_test_all_false): Reset options after test
- Scripts/webkitpy/xcode/simulated_device_unittest.py:
(SimulatedDeviceTest.reset_simulated_device_manager): _device_identifier_to_name
is a dict, and should remain one
- 7:06 AM Changeset in webkit [276433] by
-
- 33 edits4 copies11 adds in trunk
Implement WebXR Input Sources
https://bugs.webkit.org/show_bug.cgi?id=223257
Reviewed by Youenn Fablet.
.:
Enable WPE Gamepad when WebXR is enabled.
- Source/cmake/OptionsWPE.cmake:
LayoutTests/imported/w3c:
Update WebXR Input Source test expectations.
- web-platform-tests/webxr/events_input_source_recreation.https-expected.txt: Added.
- web-platform-tests/webxr/events_input_sources_change.https-expected.txt: Added.
- web-platform-tests/webxr/events_session_select.https-expected.txt: Added.
- web-platform-tests/webxr/events_session_select_subframe.https-expected.txt: Added.
- web-platform-tests/webxr/events_session_squeeze.https-expected.txt: Added.
- web-platform-tests/webxr/getInputPose_handedness.https-expected.txt: Added.
- web-platform-tests/webxr/xrInputSource_add_remove.https-expected.txt: Added.
- web-platform-tests/webxr/xrInputSource_emulatedPosition.https-expected.txt: Added.
- web-platform-tests/webxr/xrInputSource_profiles.https-expected.txt: Added.
- web-platform-tests/webxr/xrInputSource_sameObject.https-expected.txt: Added.
- web-platform-tests/webxr/xrReferenceSpace_originOffset.https-expected.txt: Added.
Source/WebCore:
This patch implements the platform data definition and the DOM bits required to support WebXR Input Sources, the input mechanism used in WebXR.
Example XR input sources include, but are not limited to, handheld controllers, optically tracked hands, and gaze-based input methods.
More info about the API in:
- https://immersive-web.github.io/webxr/#input
- https://immersive-web.github.io/webxr-gamepads-module/#webxr-device-api-integration
Tested by WebXR WPT.
- Modules/gamepad/Gamepad.h: Add setConnected method.
- Modules/webxr/WebXRGamepad.cpp: Bridge between WebXRInputSource and Gamepad
(WebCore::WebXRGamepad::WebXRGamepad):
- Modules/webxr/WebXRGamepad.h:
- Modules/webxr/WebXRInputSpace.cpp: Instance of WebXRSpace used for WebXRInputSource spaces.
(WebCore::WebXRInputSpace::create):
(WebCore::WebXRInputSpace::WebXRInputSpace):
(WebCore::WebXRInputSpace::nativeOrigin const):
- Modules/webxr/WebXRInputSpace.h:
- Modules/webxr/WebXRFrame.cpp:
(WebCore::WebXRFrame::populatePose): set emulatedPosition based on the spaces.
- Modules/webxr/WebXRInputSource.cpp: Complete WebXRInputSource implementation.
(WebCore::WebXRInputSpace::create):
(WebCore::WebXRInputSpace::WebXRInputSpace):
(WebCore::WebXRInputSpace::nativeOrigin const):
(WebCore::WebXRInputSource::create):
(WebCore::WebXRInputSource::WebXRInputSource):
(WebCore::WebXRInputSource::update):
(WebCore::WebXRInputSource::requiresInputSourceChange):
(WebCore::WebXRInputSource::disconnect):
(WebCore::WebXRInputSource::pollEvents):
(WebCore::WebXRInputSource::createEvent):
- Modules/webxr/WebXRInputSource.h:
(WebCore::WebXRInputSource::handle const):
(WebCore::WebXRInputSource::handedness const):
(WebCore::WebXRInputSource::targetRayMode const):
(WebCore::WebXRInputSource::targetRaySpace const):
(WebCore::WebXRInputSource::gripSpace const):
(WebCore::WebXRInputSource::profiles const):
(WebCore::WebXRInputSource::gamepad const):
- Modules/webxr/WebXRInputSource.idl: Add gamepad attribute.
- Modules/webxr/WebXRInputSourceArray.cpp: Implement input source updates and event dispatching.
(WebCore::WebXRInputSourceArray::create):
(WebCore::WebXRInputSourceArray::WebXRInputSourceArray):
(WebCore::WebXRInputSourceArray::length const):
(WebCore::WebXRInputSourceArray::item const):
(WebCore::WebXRInputSourceArray::clear):
(WebCore::WebXRInputSourceArray::update):
(WebCore::WebXRInputSourceArray::handleRemovedInputSources):
(WebCore::WebXRInputSourceArray::handleAddedOrUpdatedInputSources):
- Modules/webxr/WebXRInputSourceArray.h:
- Modules/webxr/WebXRSession.cpp:
(WebCore::WebXRSession::WebXRSession): Set tracking delegate before initializing tracking and rendering.
(WebCore::WebXRSession::isPositionEmulated const): Add helper method.
(WebCore::WebXRSession::shutdown): Clear WebXRInputSourceArray instance.
(WebCore::WebXRSession::sessionDidInitializeInputSources): Dispatch initial InputSource discovery event.
(WebCore::WebXRSession::onFrame): Update WebXRInputSourceArray instance.
- Modules/webxr/WebXRSession.h:
- Modules/webxr/WebXRSpace.cpp: Add virtual class isPositionEmulated to be used in WebXRFrame.
(WebCore::WebXRSpace::isPositionEmulated const):
- Modules/webxr/WebXRSpace.h:
- Modules/webxr/WebXRSystem.cpp:
(WebCore::WebXRSystem::requestSession): update FIXME comment.
- Modules/webxr/XRHandedness.h: Reuse PlatformXR enum.
- Modules/webxr/XRInputSourceEvent.cpp:
(WebCore::XRInputSourceEvent::XRInputSourceEvent):
(WebCore::XRInputSourceEvent::setFrameActive):
- Modules/webxr/XRInputSourceEvent.h:
- Modules/webxr/XRInputSourcesChangeEvent.h:
- Modules/webxr/XRTargetRayMode.h: Reuse PlatformXR enum.
- platform/gamepad/GamepadConstants.cpp:
(WebCore::xrStandardGamepadMappingString): Add xr-standard gamepad mapping name.
- platform/gamepad/GamepadConstants.h:
- platform/xr/PlatformXR.h: Add Input Source frame data.
- testing/WebFakeXRDevice.cpp: Implement required changes to run and pass WebXR Input Source tests.
(WebCore::SimulatedXRDevice::initializeTrackingAndRendering):
(WebCore::SimulatedXRDevice::frameTimerFired):
(WebCore::WebFakeXRDevice::simulateResetPose):
(WebCore::WebFakeXRDevice::simulateInputSourceConnection):
- testing/WebFakeXRDevice.h:
- testing/WebFakeXRInputController.cpp:
(WebCore::WebFakeXRInputController::create):
(WebCore::WebFakeXRInputController::WebFakeXRInputController):
(WebCore::WebFakeXRInputController::setGripOrigin):
(WebCore::WebFakeXRInputController::setPointerOrigin):
(WebCore::WebFakeXRInputController::disconnect):
(WebCore::WebFakeXRInputController::reconnect):
(WebCore::WebFakeXRInputController::setSupportedButtons):
(WebCore::WebFakeXRInputController::updateButtonState):
(WebCore::WebFakeXRInputController::getFrameData):
(WebCore::WebFakeXRInputController::getButtonOrPlaceholder const):
- testing/WebFakeXRInputController.h:
LayoutTests:
Update WebXR Input Source test expectations.
- platform/wpe/TestExpectations:
- 5:18 AM Changeset in webkit [276432] by
-
- 9 edits in trunk
[SOUP] Add support for preconnect
https://bugs.webkit.org/show_bug.cgi?id=177934
Reviewed by Sergio Villar Senin.
.:
Enable server preconnect support when building with libsoup3.
- Source/cmake/OptionsGTK.cmake:
- Source/cmake/OptionsWPE.cmake:
Source/WebKit:
Use new libsoup API to implement preconnect support.
- NetworkProcess/NetworkDataTask.cpp:
(WebKit::NetworkDataTask::create):
- NetworkProcess/soup/NetworkDataTaskSoup.cpp:
(WebKit::NetworkDataTaskSoup::NetworkDataTaskSoup):
(WebKit::NetworkDataTaskSoup::createRequest):
(WebKit::NetworkDataTaskSoup::resume):
(WebKit::NetworkDataTaskSoup::preconnectCallback):
- NetworkProcess/soup/NetworkDataTaskSoup.h:
LayoutTests:
Remove expectations for tests that are now passing.
- platform/glib/TestExpectations:
- 5:14 AM Changeset in webkit [276431] by
-
- 3 edits in trunk/Source/ThirdParty/ANGLE
Unreviewed, WPE build fix after r276390
Fix namespacing issues in nullptr_t invocations.
- src/compiler/translator/tree_util/IntermRebuild.cpp:
(sh::TIntermRebuild::BaseResult::BaseResult):
(sh::PreResult::PreResult):
(sh::PostResult::PostResult):
- src/compiler/translator/tree_util/IntermRebuild.h:
- 4:52 AM Changeset in webkit [276430] by
-
- 2 edits in trunk/Source/WebKit
[GTK] Build error: Couldn't find include 'Soup-2.4.gir'
https://bugs.webkit.org/show_bug.cgi?id=224926
Reviewed by Philippe Normand.
Use the right soup API version when generating the introspection.
- PlatformGTK.cmake:
- 4:47 AM Changeset in webkit [276429] by
-
- 10 edits in trunk
[GTK][WPE] Bump libsoup3 version to 2.99.4
https://bugs.webkit.org/show_bug.cgi?id=224925
Reviewed by Philippe Normand.
.:
- Source/cmake/OptionsGTK.cmake:
- Source/cmake/OptionsWPE.cmake:
Source/WebCore:
Update to use the new API.
- platform/network/soup/ResourceResponseSoup.cpp:
(WebCore::ResourceResponse::ResourceResponse):
- platform/network/soup/SoupVersioning.h:
(soup_message_get_tls_peer_certificate):
(soup_message_get_tls_peer_certificate_errors):
(soup_message_get_tls_certificate): Deleted.
(soup_message_get_tls_certificate_errors): Deleted.
Tools/buildstream:
- elements/sdk/libsoup3.bst:
LayoutTests:
Remove expectations for tests that should be passing now.
- platform/glib/TestExpectations:
- 1:29 AM Changeset in webkit [276428] by
-
- 9 edits1 add in trunk
[iOS] Web processes recreated after crashing are not created with the right contentSizeCategory
https://bugs.webkit.org/show_bug.cgi?id=224916
Reviewed by Tim Horton.
Source/WebCore:
The constructor of WKWebView sets the contentSizeCategory for the first page.
But if the web process crashes, the new one won't have it set.
This patch simply adds the contentSizeCategory to the WebProcessCreationParameters.
Tests: TextStyleFontSize.Startup
TextStyleFontSize.AfterCrash
- rendering/RenderThemeCocoa.h:
Source/WebKit:
- Shared/WebProcessCreationParameters.cpp:
(WebKit::WebProcessCreationParameters::encode const):
(WebKit::WebProcessCreationParameters::decode):
- Shared/WebProcessCreationParameters.h:
- UIProcess/Cocoa/WebProcessPoolCocoa.mm:
(WebKit::WebProcessPool::platformInitializeWebProcess):
- WebProcess/cocoa/WebProcessCocoa.mm:
(WebKit::WebProcess::platformInitializeWebProcess):
Tools:
- TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
- TestWebKitAPI/Tests/ios/TextStyleFontSize.mm: Added.
(-[TextStyleFontSizeWebView _contentSizeCategory]):
(TEST):
- 1:27 AM Changeset in webkit [276427] by
-
- 18 edits2 adds in trunk
[JSC] DFG / FTL should inline switch_string
https://bugs.webkit.org/show_bug.cgi?id=224578
Reviewed by Mark Lam.
JSTests:
- microbenchmarks/switch-inlining.js: Added.
(inner):
(outer):
- stress/switch-inlining-nested.js: Added.
(shouldBe):
(inner):
(outer):
Source/JavaScriptCore:
Because of r275840 change, we no longer copy StringJumpTable when compiling DFG / FTL code.
Instead we are using a pointer to UnlinkedStringTable stored in UnlinkedCodeBlock.
This allows DFG / FTL to inline CodeBlock which includes op_switch_string. We were previously not able
to do that because we cannot copy StringImpl in DFG / FTL concurrent compiler thread.
- We handle StringJumpTable / UnlinkedStringJumpTable in the same way as SimpleJumpTable / UnlinkedSimpleJumpTable.
- We put m_ctiDefault of StringJumpTable in the last element of m_ctiOffsets vector of StringJumpTable to make sizeof(StringJumpTable) small.
- We use m_indexInTable instead of m_branchOffset in FTL switch generation to make switch table dense.
The microbenchmark shows 30% improvement because of unlocking inlining feature.
ToT Patched
switch-inlining 27.1238+-0.2708 20.2630+-0.1477 definitely 1.3386x faster
<geometric> 27.1238+-0.2708 20.2630+-0.1477 definitely 1.3386x faster
- bytecode/JumpTable.h:
(JSC::StringJumpTable::ensureCTITable):
(JSC::StringJumpTable::ctiForValue const):
(JSC::StringJumpTable::ctiDefault const):
(JSC::StringJumpTable::isEmpty const):
- bytecode/UnlinkedCodeBlock.h:
(JSC::UnlinkedStringJumpTable::indexForValue const):
- dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::parseBlock):
(JSC::DFG::ByteCodeParser::InlineStackEntry::InlineStackEntry):
- dfg/DFGCapabilities.cpp:
(JSC::DFG::capabilityLevel):
- dfg/DFGGraph.h:
- dfg/DFGJITCompiler.cpp:
(JSC::DFG::JITCompiler::link):
- dfg/DFGOperations.cpp:
(JSC::DFG::JSC_DEFINE_JIT_OPERATION):
- dfg/DFGOperations.h:
- dfg/DFGSpeculativeJIT.cpp:
(JSC::DFG::SpeculativeJIT::emitSwitchStringOnString):
- ftl/FTLLowerDFGToB3.cpp:
(JSC::FTL::DFG::LowerDFGToB3::switchStringSlow):
- ftl/FTLOperations.cpp:
(JSC::FTL::JSC_DEFINE_JIT_OPERATION):
- ftl/FTLOperations.h:
- jit/JIT.cpp:
(JSC::JIT::link):
- jit/JITOpcodes.cpp:
(JSC::JIT::emit_op_switch_string):
- jit/JITOpcodes32_64.cpp:
(JSC::JIT::emit_op_switch_string):
- jit/JITOperations.cpp:
(JSC::JSC_DEFINE_JIT_OPERATION):
- 1:25 AM WebKitGTK/2.32.x edited by
- (diff)
- 12:50 AM Changeset in webkit [276426] by
-
- 3 edits in trunk/Source/WebCore
REGRESSION(r275790) [GTK][WPE] Scrolling rendering fails after the second page
https://bugs.webkit.org/show_bug.cgi?id=224824
Patch by Alejandro G. Castro <alex@igalia.com> on 2021-04-22
Reviewed by Simon Fraser.
We are making the implementation more similar to the Apple
scrolling coordinator. Now we reimplement willStartRenderingUpdate
to be able to synchronize the state there. That way we can avoid
the timer and just schedule a rendering update when we want to
commit the tree state.
Test: Fixes the timeout of the overflow-scrollable-after-back.html,
added in the regression.
- page/scrolling/nicosia/ScrollingCoordinatorNicosia.cpp:
(WebCore::ScrollingCoordinatorNicosia::ScrollingCoordinatorNicosia):
(WebCore::ScrollingCoordinatorNicosia::pageDestroyed):
(WebCore::ScrollingCoordinatorNicosia::commitTreeStateIfNeeded):
(WebCore::ScrollingCoordinatorNicosia::scheduleTreeStateCommit):
(WebCore::ScrollingCoordinatorNicosia::willStartRenderingUpdate):
(WebCore::ScrollingCoordinatorNicosia::commitTreeState): Deleted.
- page/scrolling/nicosia/ScrollingCoordinatorNicosia.h:
- 12:40 AM Changeset in webkit [276425] by
-
- 2 edits in trunk/Source/WebCore
GraphicsContextGLOpenGLCocoa should check that ANGLE is loaded in WK1 releaseCurrentContext
https://bugs.webkit.org/show_bug.cgi?id=224532
<rdar://problem/76632149>
Patch by Kimmo Kinnunen <kkinnunen@apple.com> on 2021-04-22
Reviewed by Kenneth Russell.
Avoid crashing in case WK1 is used but ANGLE library is not present.
There are some cases where ANGLE could be omitted for size savings.
- platform/graphics/cocoa/GraphicsContextGLOpenGLCocoa.mm:
(WebCore::GraphicsContextGLOpenGL::releaseCurrentContext):
