Automatic update from web-platform-tests
Pin the six version in wptrunner (#16963)
--
wp5At-commits: 17289de555db2a65533fe745a559ed51dc0db2cd
wpt-pr: 16963
Automatic update from web-platform-tests
Factor out a base class for the wpt metadata compilers
The manifest compilers follow a mostly common design so it seems
reasonable to put the common parts into a base class, and allow
subclasses to override the functionaility where required.
Differential Revision: https://phabricator.services.mozilla.com/D24177
bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1536804
gecko-commit: 7f7a2ccd90510ef9d36b0408627fd6ee39f3add5
gecko-integration-branch: central
gecko-reviewers: ato
--
wp5At-commits: 9cdfad2d6f13799bf35b4d4e0e588e35f5029f5c
wpt-pr: 16129
Automatic update from web-platform-tests
[LayoutNG] Scrollbar size comparasance fix
Also added a DCHECK to make sure that we never
set NeedsLayout flags while in intermediate layout.
Bug: 967361
Change-Id: I078d3884025c0d659cf7a25df694e082d6dd3f1c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1633031
Commit-Queue: Aleks Totic <atotic@chromium.org>
Reviewed-by: Emil A Eklund <eae@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664596}
--
wp5At-commits: 9d79f0ae5ab310264caf81d6340166ba620c4efd
wpt-pr: 17090
Automatic update from web-platform-tests
[PeerConnection] Add RTCRtpSender.setStreams()
This change exposes the setStreams() method of RTCRtpSender, which allows
changing the stream associations of an existing transceiver.
Intent to Implement and Ship:
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/g5Grj_VtF8Q
Drive-by: Fix minor lint issues.
Bug: 844386
Change-Id: I7fe463d6811d5cb74cb297bf81d5bc90040fe86a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1611836
Commit-Queue: Guido Urdaneta <guidou@chromium.org>
Reviewed-by: Philip Jägenstedt <foolip@chromium.org>
Reviewed-by: Henrik Boström <hbos@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664586}
--
wp5At-commits: 3c8213349f55f0be79f0b998edeacd29a4f3fae2
wpt-pr: 16848
Automatic update from web-platform-tests
Fix issue with inline WebXR sessions returning null viewports
Introduction of the "none" eye missed inverting the logic for
determining which viewport to return in the XRWebGLLayer, which caused
it to return the viewport typically associated with the right eye, which
is null for inline sessions.
Bug: 966077
Change-Id: I6ca51c327f7081ea4180e8515cc076b3ed8c4de8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1635731
Reviewed-by: Alexander Cooper <alcooper@chromium.org>
Reviewed-by: Klaus Weidner <klausw@chromium.org>
Commit-Queue: Brandon Jones <bajones@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664576}
--
wp5At-commits: e195fc3dbc8f0af6990c5b997479cc81528142a1
wpt-pr: 17086
Automatic update from web-platform-tests
Freezing on load.
In change https://crrev.com/0daefdcf34ba9a8c21ab907b2159c5652169a45b
caused a dcheck failure occasionally because of an implicit open of a
document inside an implicit close while running the
wpt/html/webappapis/dynamic-markup-insertion/opening-the-input-stream/abort-while-navigating.window.html
wpt test.
See: https://chromium-swarm.appspot.com/task?id=44fb7387f8411e10
for the failure.
https://chromium-review.googlesource.com/c/chromium/src/+/1628499
reverted the original code so this is a reland of some of that code.
Because of the complexity of that value use a simpler
approach and use the Frame::IsLoading attribute instead which doesn't
exhibit these cycles.
BUG=907125
Change-Id: Ifb805cf8ab90b6a228f47d88cb2ee872d9a21239
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1626956
Commit-Queue: Dave Tapuska <dtapuska@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Reviewed-by: David Bokan <bokan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664344}
--
wp5At-commits: 803be94ba72c0c4739fdb234f4f4a6c6ab41735d
wpt-pr: 16998
Automatic update from web-platform-tests
[docs] Generalize desc of Python file handlers
Remove references to PHP since readers may not be familiar with that
language.
--
[docs] Integrate wptserve documentation
Relocate instructions that are specifically geared towards test authors
to the "docs" directory. Persist API documentation in its current
location and configure Sphinx to include relevant aspects in the
generated website.
--
wp5At-commits: fe3019b4c6526136171192a0f7e993b06f59f62b, eeb34a17faace5d8b05abee1f4c56536a44a9cee
wpt-pr: 17039
Automatic update from web-platform-tests
Only allow one full screen per user activation - experimental
When the user activation state is active, we should only allow one full
screen. This is a simply change to do an experiment to see if any
web page breaks.
Intent to ship link is
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/Y58tbs-TSgE
Bug: 852645
Change-Id: Iebef20ba197ecd09e7067986073999e334b07498
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1594805
Reviewed-by: Mustaq Ahmed <mustaq@chromium.org>
Reviewed-by: Philip Jägenstedt <foolip@chromium.org>
Commit-Queue: Lan Wei <lanwei@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664292}
--
wp5At-commits: 7650b8ba52b6a95a8d58804aef50c91f15a68dc4
wpt-pr: 16758
Automatic update from web-platform-tests
[LayoutNG] Eliminate source of self-collapsedness inconsistency.
Remove an unnecessary check for border+padding logical height. No need
to test that separately, as border+padding is part of logical height,
which we would already have found to be positive at this point in that
case.
After re-layout, NGBlockNode::CopyFragmentDataToLayoutBox() would update
self-collapsedness (UpdateIsSelfCollapsing()), but the percentage based
padding would be resolved against the old size of the containing block,
as LayoutNG hasn't gotten around to updating that yet (doesn't happen
until we finish layout of the containing block).
Note that the test included used to pass with NG (if we bypassed the
DCHECK failure, that is).
Bug: 967193
Change-Id: I9f35b598ea251ed2f7e1d5b44c1364f1aca17dbe
Cq-Include-Trybots: luci.chromium.try:linux_layout_tests_layout_ng
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1634691
Commit-Queue: Emil A Eklund <eae@chromium.org>
Reviewed-by: Emil A Eklund <eae@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664283}
--
wp5At-commits: 8071657a98e017760bcbe006af6db863e9fda53e
wpt-pr: 17068
Automatic update from web-platform-tests
[css-logical] Inheritance, initial values (#14948)
Test that CSS Logical Properties and Values properties do not inherit.
Test that initial values match the spec.
https://www.w3.org/TR/css-logical-1/#property-index
Co-authored-by: Eric Willigers <ericwilligers@chromium.org>
--
wp5At-commits: f6a6bd5e4353ec9f236e53b7af1389cab0aae3a9
wpt-pr: 14948
Automatic update from web-platform-tests
Firefox.find_binary_path should return None is the binary isn't found (#17070)
--
wp5At-commits: 03dda932f137a3d8a64cac5302753dd4767de16e
wpt-pr: 17070
Automatic update from web-platform-tests
mozlog: move the capture io class from web-platform/tests to mozlog (bug 1021926)
The ability to capture the parent process' stdio is suggested to be a useful feature
to move from web-platform/tests into mozlog. To do so, I have created a new capture.py
file within mozlog/mozlog. This includes the CaptureIO class and its dependencies,
including the LoggingWrapper and LogThread classes. These have been removed from their
original location, to avoid duplication, and the files depending on them updated
accordingly.
It would be useful to add unittests testing the CaptureIO enter and exit methods, and
the original_stdio, logging_queue and logging_thread properties. I have begun such a
file with test_capture.py in mozlog/tests. This is a work in progress, however I may
need some guidance, please, in regards to creating appropriate mock data to assert.
Differential Revision: https://phabricator.services.mozilla.com/D22166
gecko-commit: f6705b8320496d0e67213299773e0c429f2b63ef
gecko-integration-branch: central
gecko-reviewers: jgraham
--
wp5At-commits: 1b9a791de2970f86b8cb93e6e595a974565c9999
wpt-pr: 15880
Automatic update from web-platform-tests
Pass rebuild through correctly to invalidate cache (#17072)
This makes us always assume the cache and manifest are in-sync,
which they aren't if e.g. the manifest version changes
--
wp5At-commits: 3f832722a2c8971ca3a895d67c1ff00099551a3b
wpt-pr: 17072
Automatic update from web-platform-tests
[ServiceWorker] Reject update() from within installing workers
The service worker spec mandates that update() fails immediately if
called from within service workers that are currently installing
(i.e. if update() is called in the |install|-event-handler).
This change implements this behavior and updates the WPT to match the
spec.
Bug: 895845
Change-Id: Ie07563e5ac8260554ef0947be4cb020124cfeddf
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1626600
Commit-Queue: Yannic Bonenberger <contact@yannic-bonenberger.com>
Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Cr-Commit-Position: refs/heads/master@{#663695}
--
wp5At-commits: cf1859b73fd230edf58da2818b914d34c1258023
wpt-pr: 16983
Automatic update from web-platform-tests
Avoid triggering spurious downloads for no-store responsive images
When responsive images are not cached in the memory cache (due to
`no-store` or devtools settings) and the viewport is being resized,
that triggers spurious downloads.
This CL fixes that, by aligning to the spec (somewhat) and making sure
that viewport changes that don't modify the image URL do not trigger a
resource load.
Bug: 960633
Change-Id: I8105bcf5dbef99767ac704872800f4d417caf8b7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1622168
Reviewed-by: Fredrik Söderquist <fs@opera.com>
Commit-Queue: Yoav Weiss <yoavweiss@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664269}
--
wp5At-commits: 966e9921407cfe4cf1028ecc7200a46cdc76d722
wpt-pr: 17061
Automatic update from web-platform-tests
Ignore all exceptions when checking if requirements are met (#17067)
Worse case, we fall back to needlessly calling pip here. I've
occasionally seen surprising failures due to (possibly broken)
packages, so let's just let pip deal with this.
--
wp5At-commits: 4547dc6e7424c68fac1a86e9cac05f21b8f352bb
wpt-pr: 17067
Automatic update from web-platform-tests
[css-ui] Correct property name
The test for the `appearance` property mistakenly referred to the vendor
prefixed version, `-webkit-appearance`. In addition to contradicting the
test name and metadata, this caused the test intended for
`-webkit-appearance` to reference the non-existent property
`-webkit--webkit-appearance`.
Correct both tests, using the suite's build process to generate the
`-webkit-appearance` test from the corrected version of the `appearance`
test.
--
wp5At-commits: 1ec27af3d596305643cde5c7b2c8104bddbfb7e2
wpt-pr: 17054
Automatic update from web-platform-tests
Get Edge browser version (#17056)
* Get Edge version using PowerShell on Windows
* Search for Edge binary in expected locations if --binary is not specified
--
wp5At-commits: 816924d8a73ac20159d2ec8e2d14dfc2e882cb17
wpt-pr: 17056
Automatic update from web-platform-tests
[wake-lock] Add tests for AbortSignal (#15943)
- Add test for invalid abortSignal and enable abortSignal tests in DedicatedWorker
--
wp5At-commits: b2d43d1fad90cccfd3e61bffcf1cbbcdac18a979
wpt-pr: 15943
Automatic update from web-platform-tests
[LayoutNG] RT elements must fall back to legacy layout.
Normally, an RT element is somewhere inside a RUBY element. We already
force legacy layout for the entire subtree established by RUBY elements,
but we need to do the same for RT elements, in case they don't have a
RUBY ancestor. RT elements create a special legacy layout object,
without necessarily establishing a new formatting context.
Bug: 967255
Change-Id: I35857fdeced55ee10a385b58b117f28007bd7791
Cq-Include-Trybots: luci.chromium.try:linux_layout_tests_layout_ng
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1632261
Commit-Queue: Morten Stenshorne <mstensho@chromium.org>
Commit-Queue: Emil A Eklund <eae@chromium.org>
Reviewed-by: Emil A Eklund <eae@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664014}
--
wp5At-commits: 9297e3a70844b16378370f9cd59e0f340cd99fcb
wpt-pr: 17053
Automatic update from web-platform-tests
Revert "Use promise in wpt pointer event tests to finish the tests"
This reverts commit 42f777c7597c0cc3cac62717729ce10ed2cfd524.
Reason for revert: <INSERT REASONING HERE>
Original change's description:
> Use promise in wpt pointer event tests to finish the tests
>
> Instead of adding a variable to check if the test is finished, we can
> create a promise object and finish the test at its last callback.
>
> Bug: 606367
> Change-Id: Iabf281a1fffca85630683e79436a8b442057b824
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1618002
> Commit-Queue: Lan Wei <lanwei@chromium.org>
> Reviewed-by: Navid Zolghadr <nzolghadr@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#663936}
TBR=lanwei@chromium.org,nzolghadr@chromium.org
Change-Id: Ic907a109cc3c377347b80882318c95e05f2b2389
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 606367
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1631364
Reviewed-by: Lan Wei <lanwei@chromium.org>
Commit-Queue: Lan Wei <lanwei@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664001}
--
wp5At-commits: 3160ff63375b5e2522c4804dd3ddc8732a288d60
wpt-pr: 17055
Automatic update from web-platform-tests
Use promise in wpt pointer event tests to finish the tests
Instead of adding a variable to check if the test is finished, we can
create a promise object and finish the test at its last callback.
Bug: 606367
Change-Id: Iabf281a1fffca85630683e79436a8b442057b824
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1618002
Commit-Queue: Lan Wei <lanwei@chromium.org>
Reviewed-by: Navid Zolghadr <nzolghadr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#663936}
--
wp5At-commits: adee37f7be39fe59c3b356cf7dccb22584765d3a
wpt-pr: 16955
Automatic update from web-platform-tests
CORB should block event-stream, gzip and x-www-form-urlencoded.
This CL adds CORB coverage for:
1) text/event-stream, application/x-www-form-urlencoded, based on the
code review discussion in a previous CL here:
https://chromium-review.googlesource.com/c/chromium/src/+/1604244/4/services/network/cross_origin_read_blocking.cc#227
2) application/gzip, which wasn't mentioned explicitly in the CR
discussion above, but which is ranked #212 in the spreadsheet
mentioned in https://github.com/whatwg/fetch/issues/860#issuecomment-457330454
and therefore probably should have been included in r659671 together
with x-gzip (ranked #54) and zip (ranked #71).
Bug: 802836
Change-Id: I8c10f900110a2cb471437a19425bfd5e38aed2fe
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1628809
Reviewed-by: Charlie Reis <creis@chromium.org>
Commit-Queue: Łukasz Anforowicz <lukasza@chromium.org>
Cr-Commit-Position: refs/heads/master@{#663824}
--
wp5At-commits: 4308b47e147274f6702a0c1add77488fbc75a309
wpt-pr: 17006
Automatic update from web-platform-tests
End XRSession when data provider is disconnected.
There are slightly different code paths to handle the disconnect for
immersive and non-immersive sessions. There can only be one immersive
session, but there can be multiple non-immersive sessions and they all
use the same magic window provider as their data provider.
Bug: 937906
Change-Id: I71f05fa2d854c2654744e818d407e4b94a2fd95e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1628269
Reviewed-by: Bill Orr <billorr@chromium.org>
Reviewed-by: Alexander Cooper <alcooper@chromium.org>
Commit-Queue: Jacob DeWitt <jacde@chromium.org>
Cr-Commit-Position: refs/heads/master@{#663821}
--
wp5At-commits: 060d986f6ced5d596805b23ca9382d0105cb2524
wpt-pr: 17003
Automatic update from web-platform-tests
[LayoutNG] Avoid indefinite CB inline-size when entering legacy child.
The constraint space has several resolution sizes, depending on what
we're to resolve. There's "available size", "percentage resolution size"
(and "replaced percentage resolution size"), and even "percentage
resolution inline-size for parent writing mode". The latter is used when
resolving percentage based margins and padding. This works when staying
within LayoutNG, but when we need to enter legacy layout for some child,
we convert all those values into something that legacy layout
understands, and dumb down the number of values to basically 3 (plus two
optional values); see BoxLayoutExtraInput. Avoid setting an indefinite
containing block inline-size, since we use this value to resolve
percentage padding and margins.
The test included only *crashed* without this fix (it rendered correctly
if DCHECKs were bypassed), but I figured it would be useful to have a
test for correct rendering in this case as well.
Bug: 966795
Change-Id: Ia2a2ede55a257b41dc38236a8bb0ff689e75a310
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1631370
Commit-Queue: Emil A Eklund <eae@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Reviewed-by: Emil A Eklund <eae@chromium.org>
Cr-Commit-Position: refs/heads/master@{#663800}
--
wp5At-commits: a6a3a634c463305c0c8e81d069a2f89be78a3b75
wpt-pr: 17031
Automatic update from web-platform-tests
Change pointer id to differentiate from other pointer type
In WebDriver spec, we assume that the actions with the same id and type
are the same, we should give the pointer actions of different source
types a different id.
Bug: 606367
Change-Id: I1d8f6a08ee8d98c4afd281f077e23067ad915ec3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1626960
Reviewed-by: Navid Zolghadr <nzolghadr@chromium.org>
Commit-Queue: Lan Wei <lanwei@chromium.org>
Cr-Commit-Position: refs/heads/master@{#663758}
--
wp5At-commits: aa49c074c969f2cf54107a0d0e82b9b6ccdf2b0c
wpt-pr: 16978
Automatic update from web-platform-tests
[docs] Add information on virtualenv (#16670)
Many of the subcommands of the WPT CLI require the `virtualenv` utility,
but the installation instructions did not instruct contributors to
install it. Add an installation step describing how to install the
dependency.
--
wp5At-commits: 9954b61de31f317c7a0a3e286aca0731b81e7e5f
wpt-pr: 16670
Automatic update from web-platform-tests
wake-lock: Fix invalid types test in wakelock-type.https.any.js (#17043)
During #17019's review process, `async` was dropped from a promise_test(),
and we ended up not returning a Promise there at all.
Fix the test by wrapping the `invalidTypes.map()` call in `Promise.all()`.
--
wp5At-commits: 13653dfec844625b952a4a06dcc3d0ee62c01dde
wpt-pr: 17043
Automatic update from web-platform-tests
Add test for screen wake lock on worker
which should not be allowed
--
Simplify the code with a simple return
--
wp5At-commits: da26956719c783726113581640bfa52f5390b8d7, a7edbaff02166b61389b03bdbcf1a8134949111f
wpt-pr: 17024
Automatic update from web-platform-tests
Reland "Portals: Add test that portals can't load data, javascript or about URLs."
This is a reland of 53511d46a059446be0913c208f24b313d050c932
Original change's description:
> Portals: Add test that portals can't load data, javascript or about URLs.
>
> Bug: 962500
> Change-Id: Ib5e275a7f31be3a7e24dfc7bf976b6e2866624fb
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1614703
> Commit-Queue: Lucas Gadani <lfg@chromium.org>
> Reviewed-by: Jeremy Roman <jbroman@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#661004}
Bug: 962500
Change-Id: Ia814f166da0daeec46b195bb96c0e20945f0de16
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1631720
Commit-Queue: Jeremy Roman <jbroman@chromium.org>
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Cr-Commit-Position: refs/heads/master@{#663605}
--
wp5At-commits: dfea0b6184416833fe6114d652dc0cd115645cb7
wpt-pr: 17036
Automatic update from web-platform-tests
Check for detached ArrayBuffers in Web Bluetooth writeValue functions
This change adds checks to verify that the ArrayBuffers passed to Web
Bluetooth's characteristic and descriptor writeValue() functions have
not been detached. If so then an InvalidStateError is thrown.
While the previous behavior was safe from use-after-frees it silently
set the characteristic or descriptor to an empty value which would be
difficult to debug.
Change-Id: Iee3d6b9f80d63961199b16a5de90bdd141ed152b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1604307
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Commit-Queue: Reilly Grant <reillyg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#658415}
--
Rewrite some tests using promise_rejects
--
wp5At-commits: d0b28df0fa7f932b05f751fd60d58a7d810bd94f, 164e4bfd899c4f27dd107fd7153f31c620cfef06
wpt-pr: 16763