This ensures that we don't pass non-trivially-copiable types through the FFI
boundary.
Differential Revision: https://phabricator.services.mozilla.com/D61626
--HG--
extra : moz-landing-system : lando
intern() has to take a str in Python 2 and 3 (i.e bytes in the former and
unicode in the latter). A recent upstream change changed url_base to unicode
which revealed there was one place we forgot to call ensure_str when interning
in the wpt metadata update code. This was blocking the sync.
Differential Revision: https://phabricator.services.mozilla.com/D62064
Automatic update from web-platform-tests
[TableNG] Fix minor colspan percentage redistribution bug
Legacy code was not enforcing 2 invariants of colspan
redistribution.
- min_width <= max_width
- redistribution should never shrink width.
The fix introduces a part of new colspan redistribution
test suite.
Bug: 958381
Change-Id: I3cc6e55af86dd376ffbc0a4c4710900df8bccce5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2030805
Commit-Queue: Aleks Totic <atotic@chromium.org>
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#738062}
--
wpt-commits: aede77a55588790eca722d2fe5ee948a9f20e216
wpt-pr: 21534
Automatic update from web-platform-tests
Restore pull_request_previews.yml to use secrets.DEPLOY_TOKEN (#21561)
As per https://github.com/web-platform-tests/wpt/pull/21560, the
re-launch of pull_request_previews.yml appears to work using my personal
access token. Switching it back to use DEPLOY_TOKEN to see if that
works.
--
wpt-commits: aa4c61f655811a1a6b2e70c3b5377fb4e34b1aa9
wpt-pr: 21561
Automatic update from web-platform-tests
Re-land pull_request_previews.yml workflow (#21558)
This attempts to re-land the pull_request_previews.yml workflow, to
create deployments for wptpr.live. This was previously reverted because
it was not properly triggering the dependent deployment workflow. I've
tested it on my local fork of wpt with my personal token and can't
reproduce the failure (i.e. it is working), so I'm landing this again to
see if it works now.
This reverts commit 462f3c4680dc9d4c5dd4740742947cd16d6fc1df. I've
changed two main things:
* Use my personal token temporarily, to mimic the local setup.
* Only deploy one PR, to avoid spamming.
--
wpt-commits: 5bfe66a9068ede4f0228ace529c3bc7b1a54c202
wpt-pr: 21558
Automatic update from web-platform-tests
Update chromium log formatter to handle known_intermittent. (#21517)
Update chromium log formatter to handle known_intermittent.
It will combine the expected status and anything in known_intermittent
into a space-separated list of statuses which are output into the
"expected" field in the chromium log.
--
wpt-commits: 9a559c43f8c4469af284d460f89b295eae4edc9b
wpt-pr: 21517
Automatic update from web-platform-tests
Remove instance of 'whitelist' in css/css-pseudo
As per https://whatwg.org/style-guide
--
wpt-commits: 1703644cf1da595b6926d1f6007c17121720a5b3
wpt-pr: 21319
Automatic update from web-platform-tests
[Security] Fix WPTs that are wrong or with undeterministic failure output
This fixes three tests:
- blob-popup.https.html checked for a random string to be empty,
which makes test expectations unreliable, use length instead.
- coop-navigated-popup.https.html had a failure in the cleanup, trying
to use a non-existent object. This change defines the broadcast channel
beforehand and uses the proper variable name.
- historical/coep-navigate-popup-unsafe-inherit.https.html had an
invalid expectation that navigating a popup from a COOP: same-origin
document to a COOP: unsafe-none document should not create a new
browsing context group.
Bug: 922191
Change-Id: Ife24aee4369a3a25bbff32a2fe5e8c29abb2d468
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2030902
Commit-Queue: Pâris Meuleman <pmeuleman@chromium.org>
Auto-Submit: Pâris Meuleman <pmeuleman@chromium.org>
Reviewed-by: Arthur Sonzogni <arthursonzogni@chromium.org>
Reviewed-by: Arthur Hemery <ahemery@chromium.org>
Cr-Commit-Position: refs/heads/master@{#737734}
--
wpt-commits: b49a11955de10ebff809462df4b54d5609b41fb9
wpt-pr: 21547
Automatic update from web-platform-tests
Block the XML parser on parser blocking stylesheets.
The html spec talks about script blocking stylesheets[1], but we do
script blocking in body by pausing the html parser. The APIs for doing
that for the XML parser had empty implementations which meant that
querying style from a script after an external stylesheet resource would
not wait for the stylesheet to load.
[1] https://html.spec.whatwg.org/multipage/semantics.html#interactions-of-styling-and-scripting
Bug: 1043944, 917116
Change-Id: I08422fa0608db8972154a8b19854906e154dfdd2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2012021
Commit-Queue: Rune Lillesveen <futhark@chromium.org>
Reviewed-by: Kouhei Ueno <kouhei@chromium.org>
Reviewed-by: Mason Freed <masonfreed@chromium.org>
Cr-Commit-Position: refs/heads/master@{#737721}
--
wpt-commits: 220c6b1d4bd084c1835cf274f3ed5f8dd8468fea
wpt-pr: 21335
Automatic update from web-platform-tests
[css-pseudo] Fix hit-testing for nested ::marker
Since r731964, if you clicked a ::marker originated by a ::before or
::after, and you read the 'path' property of the event, the first node
in the array might be the ::before or ::after pseudo-element.
Actually it was fine if you clicked the contents (text or image) of the
::marker. But it's possible to make the ::marker taller than its
contents by setting a big 'line-height'. Then you could click inside the
::marker but outside its contents, and reproduce the problem.
That was wrong, as described in pseudo_element.h,
> Pseudo element are not allowed to be the inner node for hit testing.
This patch fixes it by making PseudoElement::InnerNodeForHitTesting
iterate ancestors until it finds one which is not a pseudo-element.
Bug: 1048000, 457718
TEST=external/wpt/css/css-pseudo/marker-hit-testing.html
Some checks fail in legacy because the marker ignores 'line-height'.
Change-Id: I7bed7d0824638b0c7f4b63a744a3ca4978285445
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2030973
Reviewed-by: Rune Lillesveen <futhark@chromium.org>
Commit-Queue: Oriol Brufau <obrufau@igalia.com>
Cr-Commit-Position: refs/heads/master@{#737720}
--
wpt-commits: 738d7d25a53e359ecb83349885a34a61b3944fa5
wpt-pr: 21542
Automatic update from web-platform-tests
IndexedDB: Drop max numbers in parallel-cursors-upgrade.html
The external/wpt/IndexedDB/parallel-cursors-upgrade.html test
exercises 10k cursors in parallel, which an implementation should
support. The test was historically flaky, although it hasn't timed out
on the bots lately. But it is still slow - requiring multiple seconds
on many platforms, and 20-40 seconds on debug builds. Drop the maximum
number of parallel cursors to stop being such a drain on the commons.
Local tests: the time for --iterations=40 dropped from 23s to 5s.
Bug: 869364
Change-Id: I5a0b9de3b37650917e40cc0c34e7473a3870b21f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2033814
Auto-Submit: Joshua Bell <jsbell@chromium.org>
Commit-Queue: Victor Costan <pwnall@chromium.org>
Reviewed-by: Victor Costan <pwnall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#737669}
--
wpt-commits: 3487b250cd7304e7c8e6e7c447a8e3aee2c22082
wpt-pr: 21536
Automatic update from web-platform-tests
Move tests in custom-elements/state/ to external/wpt/custom-elements/state.tentative/
Rename |internals| to |i| to avoid lint errors.
This is a preparation to ship the feature.
Bug: 1012098
Change-Id: I789f44c072b57f05e1ec8ddfbf0855872e144996
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2032632
Reviewed-by: Rakina Zata Amni <rakina@chromium.org>
Commit-Queue: Kent Tamura <tkent@chromium.org>
Cr-Commit-Position: refs/heads/master@{#737659}
--
wpt-commits: 7b464208c72a8cf11236204b7af4bf17170ebe42
wpt-pr: 21541
Automatic update from web-platform-tests
Add some tests for Node IDL members on Attr objects
--
wpt-commits: 70490abb43593ab54c736e2fb6b5dc1cd3b306c0
wpt-pr: 21540
Automatic update from web-platform-tests
Add WPT tests to verify pings do not send a referrer header
Add WPT tests to verify pings do not send a referrer header as
specified in step 3:
https://html.spec.whatwg.org/multipage/links.html#hyperlink-auditing
Bug: 1046392
Change-Id: Ie0f133be6db34084aac26280a29a90cc9e68e707
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2024855
Reviewed-by: Dominic Farolino <dom@chromium.org>
Reviewed-by: Yutaka Hirano <yhirano@chromium.org>
Commit-Queue: Rob Buis <rbuis@igalia.com>
Cr-Commit-Position: refs/heads/master@{#737596}
--
wpt-commits: c67d8f5bdace9762af3ea99417fb5f2140438e14
wpt-pr: 21469
Automatic update from web-platform-tests
[css-flex] More changes to test suite after review of files (#7947)
* Made more changes based on analysis of tests that were <= 5 tests to a given url hash in flex
* Restore assert
Co-authored-by: Greg Whitworth <865244+gregwhitworth@users.noreply.github.com>
Co-authored-by: Stephen McGruer <stephen.mcgruer@gmail.com>
--
wpt-commits: 81a0e680add2d790bc8e0473c2b9407855703b40
wpt-pr: 7947
Automatic update from web-platform-tests
Major new manifest version (v8): path trie edition
This essentially implements https://github.com/web-platform-tests/rfcs/pull/40.
--
wpt-commits: 31c0f5efba38b7d1d7f45ac449bcbc892e8771ce
wpt-pr: 16537
Automatic update from web-platform-tests
[ElementTiming and LargestContentfulPaint]Fix background img sizes
Currently, we use exclusively the size of the LayoutObject to compute
the background image sizes. This is incorrect, for instance when looking
at CSS styles like ::first-letter. This CL fixes this by passing the
image border to ImagePaintTimingDetector and ImageElementTiming.
Bug: 996921
Change-Id: I94e4a2f3c8873be765b144179f7729430c885dac
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2023523
Commit-Queue: Nicolás Peña Moreno <npm@chromium.org>
Reviewed-by: Xianzhu Wang <wangxianzhu@chromium.org>
Cr-Commit-Position: refs/heads/master@{#737289}
--
wpt-commits: 3061f228d3d8b900b3aec8c7642ff58fc04ea011
wpt-pr: 21474
