Automatic update from web-platform-tests
multicol: Ensure that clipped out elements don't get painted.
This patch keep track of whether the logical bounding box of an
object was empty in fragmentainer iterator. If it was, then it is
likely to be placed in the first column (since typically empty rects
are 0,0 0x0). The first column doesn't have the top edge clip so
we might end up painting contents outside of the container
(see testcase).
By using the portion rect, we ensure that the fragment clip we put
in place is bounded on all sides.
R=pdr@chromium.org, wangxianzhu@chromium.org
Bug: 1017599
Change-Id: I436802f7410124ab2251bfff13ce7dcff49b335b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1879564
Commit-Queue: vmpstr <vmpstr@chromium.org>
Reviewed-by: Xianzhu Wang <wangxianzhu@chromium.org>
Cr-Commit-Position: refs/heads/master@{#709509}
--
wpt-commits: 9ceef5af8b581b9e02f11b0a877b7dc0c8edafe6
wpt-pr: 19872
Differential Revision: https://phabricator.services.mozilla.com/D53483
Automatic update from web-platform-tests
Fetch Metadata: deprecate `nested-navigate` mode.
As discussed in
https://github.com/w3c/webappsec-fetch-metadata/issues/45,
We have decided to shift the model around nested navigations from
exposure via the request's `mode` to its `destination`. This patch
shift the `nested-navigate` mode to two parts:
- `navigate` for frame and iframe.
- `no-cors` for object.
Change-Id: I09a2be6b41547f6eeb1e308416234a40f0c9b5b3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1862690
Commit-Queue: Yifan Luo <lyf@google.com>
Reviewed-by: Mike West <mkwst@chromium.org>
Cr-Commit-Position: refs/heads/master@{#709478}
--
wpt-commits: 2eebdd2f37758cdd94c0c2bddc70763862ef283f
wpt-pr: 19703
Differential Revision: https://phabricator.services.mozilla.com/D53477
Automatic update from web-platform-tests
Refactor MediaQueryList tests
Fixes#19688 via bdbff2f. Fixes#14429.
Follow-up of #18390.
Related: #18389.
* Add "width" and "height" paramters to `createIFrame` harness
* Reflow <iframe> document in `createIFrame`
* Introduce higher-level `createMQL` harness
* Make `triggerMQLEvent` work with MediaQueryList
* Use `createMQL` in "handleEvent" tests
* Use `promise_rejects_*` and `EventWatcher` in "handleEvent" tests
* Split "matchMedia.xht" and use new harness
* Merge in "matchMediaAddListener.html" and use new harness
* Rename "handleEvent" test
* Merge in empty string test
* Merge in "(min-width: 1px)" test
* Add "use strict" directive
* Remove bogus test
* Remove extra MediaQueryList IDL tests
* Remove extra MediaQueryListEvent IDL tests
--
wpt-commits: 93d54359245275eac0045011572cfb87c65b8009
wpt-pr: 19841
Differential Revision: https://phabricator.services.mozilla.com/D53475
Automatic update from web-platform-tests
Pass outside fetch client settings object for ServiceWorkerContainer#register()
A brief summary of this plumbing is below:
blink::ServiceWorkerContainer::registerServiceWorker()
-> (IPC from a renderer to the browser)
-> content::ServiceworkerContextCore::RegisterServiceWorker()
Start a registration job asynchronously
Create a ServiceWorkerVersion and set the passed settings object
-> content::ServiceWorkerVersion::StartWorker()
-> (IPC from the browser to a renderer)
-> WebEmbeddedWorkerImpl::StartWorkerContext()
-> ServiceWorkerGlobalScope::FetchAndRunClassicScript()
(Note that renderer processes of above don't have to be the same.)
This CL has behavioral changes. The 'Referer' header of the request
triggered by register() in a document will be changed.
Before this CL it was the url of the top-level script but after this CL
it is document's url. This behavior is consistent with the spec.
Bug: 937177
Change-Id: I4c137d10b7d5204151ae7eff8bfd8fdeb34fb365
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1861294
Commit-Queue: Kinuko Yasuda <kinuko@chromium.org>
Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Reviewed-by: Matt Falkenhagen <falken@chromium.org>
Cr-Commit-Position: refs/heads/master@{#709462}
--
wpt-commits: 98f4cb23e0c4e16c821721601f1c94c837803ee8
wpt-pr: 19746
Differential Revision: https://phabricator.services.mozilla.com/D53473
Automatic update from web-platform-tests
Correct a path in update-built-tests.sh.
This was added in #16816, but the test job wasn't run.
--
Make sure the update_built job runs when it needs to.
--
wpt-commits: 411c92baf1b714b3c970a8f3ed7575b6eddf63e9, 715bbdcf499287feaef607a7dc781a0a9115dfac
wpt-pr: 19888
Differential Revision: https://phabricator.services.mozilla.com/D53472
Automatic update from web-platform-tests
taskcluster: Use the mach logger instead of tbpl for stdout logging. (#19814)
* With the mach logger a timestamp of when a test starts or ends
is printed, which is very useful to detect tests taking too much
time to complete or hanging.
* On top of that the mach logger provides a more useful summary at
the end of the suite run because it gives information about how
many tests (and not only subtests) failed, also it summarizes the
cause of the unexpected results (like error or timeout).
* Also, the logs produces by the mach logger are more compact, around
half the size in megabytes is needed for storing a log with match
(instead of tbpl)
* The environment variable TERM=dumb is passed to disable colors to
be printed on the terminal
--
wpt-commits: 8e9016586f4563dfcb4090882177e3c10866abe7
wpt-pr: 19814
Differential Revision: https://phabricator.services.mozilla.com/D53470
Automatic update from web-platform-tests
[NativeFileSystem] Make FileSystemHandle cloneable
Updates postMessage() to clone FileSystemFileHandle and
FileSystemDirectoryHandle objects for same origin targets. Including
a FileSystemHandle object with a cross origin message fails by
dispatching a 'messageerror' event instead of 'message' event.
The change consists of four parts:
(1) Updates V8ScriptValueSerializerForModules to serialize
FileSystemFileHandle and FileSystemDirectoryHandle into
blink::SerializedScriptValue, by following these steps:
* Write a tag for the handle type (file or directory).
* Write the name of the file or directory.
* Creates a mojom::blink::NativeFileSystemTransferTokenPtr by calling
blink:NativeFileSystemHandle::Transfer(). This token informs the
storage::NativeFileSystemManagerImpl that a transfer is in progress.
The NativeFileSystemManagerImpl creates a
NativeFileSystemTransferTokenImpl to store the information required
to clone the handle.
* Stores the token in
blink::SerializedScriptValue::native_file_system_tokens_. This
array tracks all cloned FileSystemFileHandle. The
blink::mojom::CloneableMessage struct is also updated to hold this
array for MessagePort and BroadcastChannels.
* Write the index of the token in the native_file_system_tokens_
array.
(2) Updates V8ScriptValueDeserializerForModules to deserialize
FileSystemFileHandle objects when creating clones for the message
targets. This is the inverse of (1). Deserializing uses
mojom::blink::NativeFileSystemManager to redeem the token, which
creates the mojom::blink::NativeFileSystemFileHandlePtr or
mojom::blink::NativeFileSystemDirectoryHandlePtr using the info
stored by NativeFileSystemTransferTokenImpl.
(3) Updates content::NativeFileSystemManagerImpl to support token
transfers. To redeem a token, NativeFileSystemManagerImpl receives
a mojo message that includes the token as well as a request for a
handle interface like mojom::blink::NativeFileSystemFileHandlePtr.
NativeFileSystemManagerImpl finds the token and then binds the request.
Token redemption does not return any results. Token redemption should
never fail, unless a render process is misbehaving.
NativeFileSystemManagerImpl performs a few sanity checks before binding
the mojo request, including a token existence check, a handle type
check and an origin check. If any of the sanity checks fail,
NativeFileSystemManagerImpl silently fails closing the redeemed
FileHandle's pipe.
(4) Adds a cross origin check to window and message port messaging.
Most message targets, like dedicated workers, are same origin only.
However, both windows and message port messages can go cross origin.
When a cross origin message includes a FileSystemHandle, the message
must fail with a 'messageerror' event to prevent cross origin access
to the FileSystemHandle.
Messaging between windows already included origin information before
this change. This change adds a NativeFileSystem origin check before
dispatching a message event to a window. The message event is
replaced with a message error when a cross origin NativeFileSystem
object exists in the message data.
For message ports, no sender origin information existed before this
change. This change updates the CloneableMessage structs to
include a 'sender_origin' url::Origin property. Message ports use
this property to perform the same cross origin NativeFileSystem
check as the window.
The NativeFileSystemManagerImpl performs an additional origin check
before binding the FileSystemHandle mojo request. The
NativeFileSystemManagerImpl cannot trust the postMessage() origin
check performed in the render process.
Bug: 955192
Change-Id: Ieeb76bd8102067d70c5d7719622ecd4930c3a88f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1791942
Commit-Queue: Steve Becker <stevebe@microsoft.com>
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Reviewed-by: Yuki Shiino <yukishiino@chromium.org>
Reviewed-by: Kinuko Yasuda <kinuko@chromium.org>
Reviewed-by: Marijn Kruisselbrink <mek@chromium.org>
Reviewed-by: Olivier Yiptong <oyiptong@chromium.org>
Cr-Commit-Position: refs/heads/master@{#709407}
--
wpt-commits: 474923949524b5c05a9e6f28ec082fdca87078de
wpt-pr: 18921
Differential Revision: https://phabricator.services.mozilla.com/D53468
Automatic update from web-platform-tests
[css-lists] Add test for list-style-type:<string> with Ahem font
In the other reftests the width of the marker text is not known, so the
references have to rely on some tricks with the ::before pseudo-element.
This patch adds a test that uses the Ahem font, so that the width of the
characters is known. Then the result can be checked with a simpler
reference file.
Note it fails in LayoutNG due to the lack of scrollbar (bug 1012294),
but it passes in legacy.
Bug: 687946
Change-Id: I7fd0856e93d4e489af65b5c767ec81c109804c5c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1879067
Commit-Queue: Oriol Brufau <obrufau@igalia.com>
Reviewed-by: Manuel Rego <rego@igalia.com>
Cr-Commit-Position: refs/heads/master@{#709405}
--
wpt-commits: 52aeca2db20bd3304407bf1a1fca4e5a827db243
wpt-pr: 19868
Differential Revision: https://phabricator.services.mozilla.com/D53467
Automatic update from web-platform-tests
[Azure Pipelines] increase experimental runs to every 3h (#19877)
This uses the new epochs/three_hourly branch just introduced:
https://github.com/web-platform-tests/wpt/pull/19873
The first run of the action created the branch:
https://github.com/web-platform-tests/wpt/runs/274428602
Also decreate the job timeout to 3h to prevent a queue of jobs from
building up if jobs start hanging or taking very long. The typical
running time for both Edge and Safari are now ~1.7h.
--
wpt-commits: 4efaeb295baf76c416c3c32ecf8d81d6c602ad11
wpt-pr: 19877
Differential Revision: https://phabricator.services.mozilla.com/D53466
Automatic update from web-platform-tests
[infra] add a cron workflow to update a new epochs/three_hourly branch (#19873)
The epochs/three_hourly branch doesn't exist yet and nothing will be
triggered by it initially, which makes it a good candidate for
verifying that `./wpt rev-list` combined with GitHub Actions will be a
robust way to maintain the epochs branches.
The immediate motivation for this is to increase the frequency of Edge
and Safari runs on Azure Pipelines from every 6 to every 3 hours:
https://github.com/web-platform-tests/wpt/issues/18669#issuecomment-546062584
--
wpt-commits: b482734e5092c5f5a6aadc66aa1c599742f2dc67
wpt-pr: 19873
Differential Revision: https://phabricator.services.mozilla.com/D53465
Automatic update from web-platform-tests
Move relationship properties out of aria_attributes.idl, remove string version
We want to ship non-relationship aria properties on ElementInternals
first, because the relationship properties are still WIP.
We are also removing the string version of the relationship properties.
Change-Id: Iccd73d1c9b1b31a431764e7989083dac43ed76d0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1874751
Reviewed-by: Kent Tamura <tkent@chromium.org>
Reviewed-by: Meredith Lane <meredithl@chromium.org>
Commit-Queue: Rakina Zata Amni <rakina@chromium.org>
Cr-Commit-Position: refs/heads/master@{#709359}
--
wpt-commits: b7ac8324bfc2ab9f0351056a1ea8381a845f6106
wpt-pr: 19842
Differential Revision: https://phabricator.services.mozilla.com/D53464
Automatic update from web-platform-tests
Don't trigger :focus-visible matching if a meta-key was pressed.
Also, move UpdateHadKeyboardEvent() call into KeyboardEventManager::KeyEvent().
Happy to revert the latter change if folks don't like it, but it seemed logically consistent with e.g. gesture detection.
Bug: 920458
Change-Id: Ia423a0533325314ae4f83e619337b5a81af0f4c4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1871715
Commit-Queue: Alice Boxhall <aboxhall@chromium.org>
Reviewed-by: Kent Tamura <tkent@chromium.org>
Cr-Commit-Position: refs/heads/master@{#709357}
--
wpt-commits: 69646ed085ad3e2295549c3ecf3f5caa9e4e1e9a
wpt-pr: 19817
Differential Revision: https://phabricator.services.mozilla.com/D53463
Automatic update from web-platform-tests
[workers] Correct assertions
This test generates subtests based on the value of a binding named
`unexpected`. Prior to this commit, no such binding was declared by the
test, so it failed to declare any subtests and instead produced a
ReferenceError. Currently, testharness.js interprets this result as a
failing "single-page test," although it will soon be more accurately
reported as a harness error [1].
Update the test to assert the presence of the properties as originally
intended.
[1] https://github.com/web-platform-tests/rfcs/blob/master/rfcs/single_test.md
Co-authored-by: Ms2ger <Ms2ger@gmail.com>
--
wpt-commits: ca2983ebb0a61e076e13ce848ab999293093f4cb
wpt-pr: 19477
Differential Revision: https://phabricator.services.mozilla.com/D53461
Automatic update from web-platform-tests
[fetch] Correct test bugs
Prior to this commit, this test referenced a non-existent binding while
generating subtest titles. This produced an uncaught exception in all
implementations. Currently, testharness.js interprets this result as a
failing "single-page test," although it will soon be more accurately
reported as a harness error [1].
Remove the unresolvable reference and improve the subtest title to more
accurately describe the behavior under test.
[1] https://github.com/web-platform-tests/rfcs/blob/master/rfcs/single_test.md
--
[fetch] Correct syntax error
--
wpt-commits: 285fcd938bc74cf4161c2ff7b84a0f274c0f1320, 7031c8cc58033312878636c8f973589946352385
wpt-pr: 19480
Differential Revision: https://phabricator.services.mozilla.com/D53459
Automatic update from web-platform-tests
css-flexbox: narrow special case for deprecated flex box
https://chromium-review.googlesource.com/c/chromium/src/+/1845506
modified LayoutBox::SizesLogicalWidthToFitContent() to return true
for webkit-blox/webkit-inline-box. This is problematic as the display
type does not necessarily correlate to the LayoutObject that is used.
The right fix is to look at the display type *and* the LayoutObject
used.
BUG=1014001
TEST=external/wpt/compat/webkit-box-fieldset.html
Change-Id: I80cc6d166e02c1e2ec6c2c0c58bcf770793ca41e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1874250
Commit-Queue: Scott Violet <sky@chromium.org>
Reviewed-by: Christian Biesinger <cbiesinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#709205}
--
wpt-commits: bee8fe07402f7f919688a16ac13def147b79a6eb
wpt-pr: 19855
Differential Revision: https://phabricator.services.mozilla.com/D53457
Automatic update from web-platform-tests
[LayoutNG] ComputeReplacedSize edge case fix
Edge case for "have aspect ratio, but no intrinsic size"
If there was no intrinsic size, we treated it as no size at all.
But, if there is a css size, we should use that instead.
Bug: 1015311
Change-Id: Ia8f12bd23cda5c18b9aae318eee91906840475d3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1877595
Commit-Queue: Aleks Totic <atotic@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#709134}
--
wpt-commits: cfffd794f8c086ac20c664658738a4dce18462c1
wpt-pr: 19857
Differential Revision: https://phabricator.services.mozilla.com/D53456
Automatic update from web-platform-tests
Check that enumerateDevices is returning new objects for every call (#19488)
* Check that enumerateDevices is returning new objects for every call
* Update mediacapture-streams/MediaDevices-enumerateDevices-returned-objects.https.html
Co-Authored-By: Philip Jägenstedt <philip@foolip.org>
--
wpt-commits: f2c6436b3496cddaef5340fe32bb71084facfb3e
wpt-pr: 19488
Differential Revision: https://phabricator.services.mozilla.com/D53451
Automatic update from web-platform-tests
[css-grid] Grid align justify stretch (#19809)
* add grid test for align and justify stretch and css support
--
wpt-commits: c27376e1d4d398bcd08d2511c38dcb564a7e48f4
wpt-pr: 19809
Differential Revision: https://phabricator.services.mozilla.com/D53450
Automatic update from web-platform-tests
[LayoutNG] Fix hyphens when rewinding occurs
This patch removes |NGLineInfo::LineEndFragment|, which was
introduced to support hyphens and ellipsis with the single
code. However, ellipsis was changed to use different code,
and |LineEndFragment| ended up with non-optimal way to handle
hyphens.
Although hyphens appear at most once in a line, storing this
to |NGLineInfo| makes harder to keep it in sync with the
hyphenated item. This patch changes it to a field in
|NGInlineItemResult|.
Bug: 1015297
Change-Id: I0706cf54bebfa967661dc40123c6dfd5887ff59c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1868534
Commit-Queue: Koji Ishii <kojii@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#708894}
--
wpt-commits: b529e77cc4fdeb81b249e1dfdce955e2f1fed89a
wpt-pr: 19800
Differential Revision: https://phabricator.services.mozilla.com/D53443
Automatic update from web-platform-tests
Fix test/ref text different and remove expectations.
Remove the remaining test failures for writing-mode propagation to
viewport from TestExpectations. One of the tests had a missing period
in the test text that caused it to fail, and the other one was fixed in
the github repo, now passing.
Bug: 988585
Change-Id: I307eb4453479a71529f6c96d87b299f22a7a2c36
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1875263
Reviewed-by: Stephen Chenney <schenney@chromium.org>
Commit-Queue: Rune Lillesveen <futhark@chromium.org>
Cr-Commit-Position: refs/heads/master@{#708748}
--
wpt-commits: e68120da0fb52f010f206f3ecc63cfa09885b0f4
wpt-pr: 19849
Differential Revision: https://phabricator.services.mozilla.com/D53442
Automatic update from web-platform-tests
[testharness.js] Don't complete after allowed exc. (#19612)
The following tests intentionally produce an uncaught exception and then define
subtests:
- html/semantics/scripting-1/the-script-element/module/error-and-slow-dependency.html
- html/webappapis/scripting/processing-model-2/window-onerror-parse-error.html
- html/webappapis/scripting/processing-model-2/window-onerror-runtime-error.html
- html/webappapis/scripting/processing-model-2/window-onerror-runtime-error-throw.html
testharness.js immediately transitions to "complete" and ignores the
subsequent subtests.
--
wpt-commits: e9465a52d29ca76d30e327db675bb66139f27b8c
wpt-pr: 19612
Differential Revision: https://phabricator.services.mozilla.com/D53439
pyup-bot
moz-wptsync-bot
Aleks Totic
Oriol Brufau
Vladimir Levin
Yifan Luo
Alexey Shvayka
Kenichi Ishibashi
Ms2ger
Philip Jägenstedt
Carlos Alberto Lopez Perez
Steve Becker
Rakina Zata Amni
Alice Boxhall
Mike Pennisi
Scott Violet
autofoolip
youennf
_Ros_
Sebastian Silbermann
Adam Rice
Simon Pieters
Koji Ishii
Rune Lillesveen
Rayan Kanso