Automatic update from web-platform-tests
Allow cross-origin subframes to send automatic beacons.
Fenced frames can send out reporting beacons if an event occurs.
However, only same-origin documents are allowed to send automatic
beacons. This restriction is problematic when a subframe from a
third-party ad server is embedded in an ad loaded with a script from
that server. One such use case is when an ad frame embeds a subframe
that loads the actual contents of an ad. This is known as Third-party ad
serving, or 3PAS.
The solution to allow cross-origin subframes to send automatic beacons
involves four parts: subframes will opt in through a header, the main ad
frame will opt in to allow cross-origin subframes to use its automatic
beacon data when sending automatic beacons, the subframe will then be
able to use the automatic beacon data that is stored in its frame tree's
FencedFrameConfig, and the automatic beacon data will be stored
per-document rather than per-fenced frame config.
The opt-in mechanism for the main ad frame's data is a new
"crossOriginExposed" parameter in
`setReportEventDataForAutomaticBeacons()`, which defaults to false.
The opt-in mechanism for the cross-origin subframe is done through a new
"Allow-Fenced-Frame-Automatic-Beacons" response header.
The reason for storing the beacon data per-document is a security one.
With the current setup, a same-origin subframe of an ad frame could set
automatic beacon data, and a different cross-origin subframe could then
use that data. Since those two frames are not direct descendants of each
other, we want to prevent that behavior, and instead only allow
documents to use opted in data set by direct ancestors.
This CL also modifies the automatic beacon web platform tests to
properly handle multiple beacons being sent with the same data. The
automatic beacon store endpoint now accepts a "beacon type" parameter
that is added to the hash when storing/retrieving a beacon. This will
prevent collisions if two beacons with different types are sent with
the same data.
Change-Id: I4af9d7ef34732dcd56c4f6bcf677f48956f7968c
Bug: 1504306
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5018876
Reviewed-by: Dominic Farolino <dom@chromium.org>
Reviewed-by: Garrett Tanzer <gtanzer@chromium.org>
Commit-Queue: Liam Brady <lbrady@google.com>
Reviewed-by: Arthur Sonzogni <arthursonzogni@chromium.org>
Reviewed-by: Alexander Timin <altimin@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1234109}
--
wpt-commits: 458d51295af838f04608bda0dc912aaf88bc1de3
wpt-pr: 43125
Automatic update from web-platform-tests
wpt: Update some security-features scripts to run under Python 3.
In Python 3, dict.keys() returns an object which doesn't have
a remove() method. Instead, this is converted to a list so that
this kind of manipulation is possible.
basestring no longer exists in Python 3, so str is used instead.
Technically basestring also includes bytes, but as far as I can
tell it's not reasonable for bytes to end up here.
Finally, message is not guaranteed to exist, so the error object
itself is printed instead.
Change-Id: Iad38779f8cfcfe1f464b73e8cbabef923d90f6ec
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5090427
Reviewed-by: Hiroshige Hayashizaki <hiroshige@chromium.org>
Commit-Queue: Jeremy Roman <jbroman@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1234071}
--
wpt-commits: 8b2ff8f91555ee3cda84ad9bca665d960c4769e7
wpt-pr: 43526
Automatic update from web-platform-tests
Fix memory leak in mousemove_prevent_default_action.
The leak was likely the result of the following:
1. Promise.race only ensuring that one of the promises got resolved.
2. test-driver unable to complete the drag 'n drop operation.
Event sequence stalls at dragenter.
Resolve by cancelling the drag 'n drop operation during drag start.
The purpose of the test is not to ensure the successful completion of
a drag 'n drop operation, but rather to ensure that a preventDefault
on mousemove does not block a drag operation from starting.
Bug: 1508026
Change-Id: I9c94780b478a8a0dbb7cc577d0fb53fd7719ea62
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5089155
Reviewed-by: Mustaq Ahmed <mustaq@chromium.org>
Commit-Queue: Kevin Ellis <kevers@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1233999}
--
wpt-commits: 730060196d22566dd233cdbfda9c0c807564ff16
wpt-pr: 43527
Automatic update from web-platform-tests
Revert "[css-images-4] Implement unprefixed cross-fade()."
This reverts commit b15018bba29189526bd341b796f37599a7f319de.
Reason for revert: The CL is likely the cause of crbug.com/1508861.
Original change's description:
> [css-images-4] Implement unprefixed cross-fade().
>
> This is similar to -webkit-cross-fade(), but differs in that it
> has a different syntax, and supports more (or less) than two images.
> It also supports cross-fading against a solid color, which we implement
> internally as a gradient for simplicity. (This may not be optimal for
> memory or performance, but it should do for now.)
>
> We do not implement parts of the spec that are not yet written
> but will be, in particular simplification at computed value time.
>
> The functionality is behind a new runtime flag CSSCrossFade,
> which is set to experimental as usual.
>
> Change-Id: I4d29e14b215a65223b98dcaee8ba99b8f62d94dc
> Bug: 614906
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5055552
> Commit-Queue: Anders Hartvoll Ruud <andruud@chromium.org>
> Reviewed-by: Anders Hartvoll Ruud <andruud@chromium.org>
> Auto-Submit: Steinar H Gunderson <sesse@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#1233870}
Bug: 614906
Change-Id: I77cfd8e9d833a5d000589068e4f1397c3afaad06
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5094054
Owners-Override: Felipe Andrade <fsandrade@chromium.org>
Auto-Submit: Felipe Andrade <fsandrade@chromium.org>
Commit-Queue: Felipe Andrade <fsandrade@chromium.org>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#1233911}
--
wpt-commits: 99b9bc1e616081ad91397f7168d2f17e9992db93
wpt-pr: 43537
Automatic update from web-platform-tests
[css-images-4] Implement unprefixed cross-fade().
This is similar to -webkit-cross-fade(), but differs in that it
has a different syntax, and supports more (or less) than two images.
It also supports cross-fading against a solid color, which we implement
internally as a gradient for simplicity. (This may not be optimal for
memory or performance, but it should do for now.)
We do not implement parts of the spec that are not yet written
but will be, in particular simplification at computed value time.
The functionality is behind a new runtime flag CSSCrossFade,
which is set to experimental as usual.
Change-Id: I4d29e14b215a65223b98dcaee8ba99b8f62d94dc
Bug: 614906
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5055552
Commit-Queue: Anders Hartvoll Ruud <andruud@chromium.org>
Reviewed-by: Anders Hartvoll Ruud <andruud@chromium.org>
Auto-Submit: Steinar H Gunderson <sesse@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1233870}
--
wpt-commits: 2f00e95d38dbb8969039840432c5fb42f3c23255
wpt-pr: 43417
Automatic update from web-platform-tests
Unload: reject promise on pagehide instead of unload
This test fails in Chrome because it depends on the promise triggered
by unload event. This CL changes it to pagehide handler so it would be
triggered as expected even after unload is deprecated in Chrome.
In other browsers, pagehide is also triggered for the case that is
tested by the test (detecting navigation), and is more reliable than
unload.
Bug: 1488371
Change-Id: If44b78e5ff89b1d5ba8901be48cb9ad118a67650
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5070653
Reviewed-by: Rakina Zata Amni <rakina@chromium.org>
Reviewed-by: Kent Tamura <tkent@chromium.org>
Commit-Queue: Nonoka Muraki <murakinonoka@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1233805}
--
wpt-commits: 746807e7a50fda7a427489a9c8f81552edd6c0c7
wpt-pr: 43412
Automatic update from web-platform-tests
Speculation Rules: Drive-by test fixes
Noticed a couple of things that needed fixing while working on another
CL.
Change-Id: I45bab97ce74a3119064a9133c0450be62d29dadc
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5090807
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Commit-Queue: Adithya Srinivasan <adithyas@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1233630}
--
wpt-commits: 82dbabc409af6754fe71aa14b3d7331b04374bef
wpt-pr: 43528
Automatic update from web-platform-tests
IndexedDB: fix structured-clone.any.html
There was a race between setting up the transaction event watcher and
the transaction auto-committing.
Unfortunately I don't see a great way to make this testing pattern more
full-proof, e.g. by checking whether the transaction is already finished
at the time of setting up the observer, since there's no way to check
the transaction's current state.
This is a good example of the pain of mixing IDB's unique flavor of
async with promises.
Bug: 1506270
Change-Id: I9cd630f9e2732532471c04841d4fbf8ddd722d66
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5086039
Commit-Queue: Evan Stade <estade@chromium.org>
Reviewed-by: Nathan Memmott <memmott@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1233503}
--
wpt-commits: c0fdc0225d2294c8a5002f5684f00012609eb9ae
wpt-pr: 43520
Automatic update from web-platform-tests
[soft-navigations] Fix crash when user interaction precedes softnav (#43502)
[1] ended up causing crashes in cases a non-soft-navigation user
interaction precedes a soft navigation.
This CL fixes that by:
* Making sure we're not emitting soft nav entries before their time was
set.
* Emitting softnavs when the time is set, if all conditions were met
before then.
[1] https://chromium-review.googlesource.com/c/chromium/src/+/5071732
Bug: 1507750
Change-Id: I9e26cebfcab2313b425261004a0e1eeaa8517842
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5079439
Commit-Queue: Yoav Weiss <yoavweiss@chromium.org>
Reviewed-by: Ian Clelland <iclelland@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1232691}
Co-authored-by: Yoav Weiss <yoavweiss@chromium.org>
--
wpt-commits: 565993e57b6cceae7f3a2ae7954929d9ab76b809
wpt-pr: 43502
Automatic update from web-platform-tests
DOM: Observable subscription uses Callback or Dictionary union
This CL addresses https://github.com/WICG/observable/issues/71, which
calls for an ergonomic design improvement whereby you can subscribe with
either a full-fledged `Observer` dictionary, or just a single callback
which is used as the subscription's next handler.
Subscribing with a single callback is identical to subscribing with an
`Observer` dictionary whose `next` handler is provided and nothing else.
R=masonf@chromium.org
Bug: 1485981
Change-Id: I5781a6cf404d97ac2de2bc88a303bcd522f63ec5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5071992
Commit-Queue: Dominic Farolino <dom@chromium.org>
Reviewed-by: Mason Freed <masonf@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1233432}
--
wpt-commits: f5ebc4aeab4da3039eb2b169342783747acb4e7d
wpt-pr: 43420
Automatic update from web-platform-tests
Mark two test cases from changing-scroll-snap-align.html as tentative.
These were added for nested snap work on crbug.com/1467300, but the
behavior is subject to an open spec issue as discussed at:
https://github.com/web-platform-tests/interop/issues/602
Bug: 1505893
Change-Id: Iaf9783f7220acc5eeb19387e57f6ec3f88bc19e2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5085089
Reviewed-by: David Awogbemila <awogbemila@chromium.org>
Commit-Queue: Steve Kobes <skobes@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1233355}
--
wpt-commits: 7db499bf01c922eacee27edeb21db75356ee8178
wpt-pr: 43506
Automatic update from web-platform-tests
Integrate RTCRtpEncodingParameters maxFramerate test into main webrtc folder (#43173)
per https://github.com/w3c/webrtc-pc/pull/2785
--
wpt-commits: 37e24fab108ec20ee6091f05266875026609f393
wpt-pr: 43173
Automatic update from web-platform-tests
Add test for w3c/webrtc-pc#2687 (#43171)
* Add test for w3c/webrtc-pc#2687
Check that a new data channel to an unbundling endpoint makes the caller connectionState go through 'connecting'
* Fix lint
--
wpt-commits: ade2b4b05020a2a1d4c0504f343571fb86d12668
wpt-pr: 43171
Automatic update from web-platform-tests
Fix application of positioning area offset in BackgroundImageGeometry
The offset was applied before converting from edge-relative form to
absolute form.
Fixed: 1507850
Change-Id: Ia4f0f3143dc70b3e090e5f61ae518d0d88aff754
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5077504
Commit-Queue: Fredrik Söderquist <fs@opera.com>
Reviewed-by: Stephen Chenney <schenney@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1233313}
--
wpt-commits: 83691d9659c5da00b729bdea4943001c5adee81f
wpt-pr: 43516
Automatic update from web-platform-tests
[FLEDGE]: Add allSlotsRequestedSizes to AuctionConfig.
This new optional field will be sent to InterestGroups that request it
when requesting trustedBiddingSignals.
This CL adds some WPT tests for requestedSize and
allSlotsRequestedSizes in the parse failure case, but no tests for
the successful parse case. I'll add those once the feature does
something.
This also introduces a new feature to enable/disable the new
argument.
Bug: 1506238
Change-Id: I9c7706c69f8bbfafd52fcefadc390c70174a762f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5073440
Reviewed-by: Brendon Tiszka <tiszka@chromium.org>
Reviewed-by: Mike Taylor <miketaylr@chromium.org>
Commit-Queue: Matt Menke <mmenke@chromium.org>
Reviewed-by: Russ Hamilton <behamilton@google.com>
Cr-Commit-Position: refs/heads/main@{#1233000}
--
wpt-commits: 02e3e674eb4b8003b4458e8877e862bf910a9395
wpt-pr: 43456
Automatic update from web-platform-tests
[navigation API] Fix flaky scroll restoration test
after-transition-intercept-handler-modifies.html sometimes fails
because the scroll restoration is being overridden if the layout
changes from the previous same document navigation have not yet been
applied. This becomes more consistent when same document navigation IPCs
are prioritized (see crrev.com/c/4981177).
This CL changes the test to ensure rendering happens between the forward
and back navigations, which in turn ensures the new layout and scroll
position were updated before navigating back.
Bug: 1042118
Change-Id: Ia6a3768f8780f9ddaff80f7fd877ddca5c6184b7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5080962
Commit-Queue: Scott Haseley <shaseley@chromium.org>
Reviewed-by: Nate Chapin <japhet@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1232997}
--
wpt-commits: ff2032f656cec2aa3904e6270925f3fd1306371a
wpt-pr: 43487
Automatic update from web-platform-tests
LoAF: remove desired{Execution|Render}Start
Feedback from the OT was that this attribute is a bit advanced and not
actionable, so removing in favor of keeping the shipped API more nimble.
Bug: 1507960
Change-Id: Ida2fd8b0defbcd1bdb3a12e2ed9b0c5be15ab243
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5078588
Commit-Queue: Noam Rosenthal <nrosenthal@chromium.org>
Reviewed-by: Scott Haseley <shaseley@chromium.org>
Reviewed-by: Michal Mocny <mmocny@chromium.org>
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1232876}
--
wpt-commits: b88b515ae656e4d1e5fad72ef2ab196984f515cb
wpt-pr: 43470
Automatic update from web-platform-tests
[soft navigations] Ignore non-attributed LCP before softnav detection
This CL ignores LCPs that are not attributed to the soft navigation
task after a user interaction and before a soft navigation was
detected, as discussed in [1].
It also fixes a bug in soft navigation detection in the case where all
paints arrive before the URL was changed. Currently in such cases, a
soft navigation entry is not emitted. This CL fixes that.
[1] https://github.com/WICG/soft-navigations/issues/29#issuecomment-1831386977
Bug: 1505994
Change-Id: I444ff03c0b64663b913b4e73a7c820b7bab11047
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5071732
Commit-Queue: Yoav Weiss <yoavweiss@chromium.org>
Reviewed-by: Michal Mocny <mmocny@chromium.org>
Reviewed-by: Ian Clelland <iclelland@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1232451}
--
wpt-commits: 143e58acd4ca2f1164359fbb63c7772007ce251c
wpt-pr: 43451
Automatic update from web-platform-tests
Fix handling invalid date for Expires header value "0" as expired
This CL changes handling the invalid Expires header value "0" as a time in the
past and stale, as described in section 5.3 [1] of RFC 9111.
> A cache recipient MUST interpret invalid date formats, especially the value
> "0", as representing a time in the past (i.e., "already expired").
[1] https://www.rfc-editor.org/rfc/rfc9111.html#section-5.3
Bug: 853508
Change-Id: I853b2972c4420936fda8373e292d2d90fdb7988a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4021895
Reviewed-by: Takashi Toyoshima <toyoshim@chromium.org>
Reviewed-by: Hayato Ito <hayato@chromium.org>
Commit-Queue: Adam Rice <ricea@chromium.org>
Reviewed-by: Adam Rice <ricea@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1232768}
--
wpt-commits: e8d57f52b64b733a0927503ed31b8e4e9c7100a7
wpt-pr: 42675
Automatic update from web-platform-tests
[WPT] [CSP] Rewrite flaky inline-style-blocked
The test was depending on the order in which the
securitypolicyviolation was being received, and was clearly flaky.
This CL rewrites the test completely.
Bug: 1370215
Change-Id: I829c97a885eefa0cc7a49f0952ea6779e1281a6c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5079209
Reviewed-by: Jonathan Hao <phao@chromium.org>
Auto-Submit: Antonio Sartori <antoniosartori@chromium.org>
Commit-Queue: Jonathan Hao <phao@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1232715}
--
wpt-commits: a83bf9a725f275854b37d31dc1708d08c5dd5670
wpt-pr: 43496
Automatic update from web-platform-tests
compute-pressure: Ensure clean test exit
It can happen on certain bots that the pressure updates are still sent
before disconnect() is processed in the mockPressureService, which can
lead in certain cases to receive more updates than allowed in the test.
This patch adds a `gotPenalty` flag, to make sure that at least one
penalty was applied when reaching the maximum allowed updates.
Change-Id: I535b8c21370d14f13f0d6140d1eb2cc043cbeac2
Bug: 1501324
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5076963
Reviewed-by: Fr <beaufort.francois@gmail.com>
Commit-Queue: Arnaud Mandy <arnaud.mandy@intel.com>
Reviewed-by: Raphael Kubo Da Costa <raphael.kubo.da.costa@intel.com>
Cr-Commit-Position: refs/heads/main@{#1232652}
--
wpt-commits: da0a56526aa23f752fc0186b8cf22fa9f05e2fc8
wpt-pr: 43459
Automatic update from web-platform-tests
Fix progress function calculation
We can simplify the progress function when either all the operands
resolve to the same canonical unit or all the operands have the same type.
Currently `progress(10em from 10em to 100em)` crashes as we try to compute the canonical value in either case of simplification.
This CL changes this behavior to distinguish which simplification path
to use, so that in the mentioned progress() example we just take the
double values when all the operands are of the same type.
Bug: 1503730
Change-Id: I361477ba2a03fd6063861d1d1a82d4b77fa77f2f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5067615
Commit-Queue: Daniil Sakhapov <sakhapov@chromium.org>
Reviewed-by: Anders Hartvoll Ruud <andruud@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1232633}
--
wpt-commits: 94c20e8f17b32ff2adfefd87a1fea07189456be8
wpt-pr: 43391
Automatic update from web-platform-tests
Never break at a resumed parallel block flow inline node.
What I did in CL:5063008 caused trouble, because incoming child break
tokens could be re-propagated twice; once in BoxFragmentBuilder::
AddBreakBeforeChild(), and then again when walking through the
unprocessed incoming inline break tokens in BlockLayoutAlgorithm::
Layout().
Revert the code changes from that CL, and replace it with something
better (and simpler): Never consider breaking before a resumed inline
node in a parallel block flow. This is similar to how we already handle
interleaved block formatting context roots. They may also be interrupted
by floats that got pushed from an earlier fragmentainer, and thus not be
able to resume and place anything in one fragmentainer, and try again in
the next. See e.g. web_tests/external/wpt/css/css-break/float-005.html
Bug: 1506393
Change-Id: If580b5a5f33a8b4ddf86677f79d1a9f06819bc51
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5077847
Auto-Submit: Morten Stenshorne <mstensho@chromium.org>
Reviewed-by: Kent Tamura <tkent@chromium.org>
Commit-Queue: Kent Tamura <tkent@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1232515}
--
wpt-commits: cce26e3757dcfd21629f3334072ce9c1292c4d9f
wpt-pr: 43474
Automatic update from web-platform-tests
[wptrunner] Display reruns starting from `1/N` to `N/N` (#43155)
Instead of `0/N` to `(N-1)/N`, which looks incomplete.
--
wpt-commits: dd47ca67f55ac9df45235b2335200fb885dd1357
wpt-pr: 43155
