Agenda for Oct 30th, 2025

[nairnandu]

Here is the proposed agenda for the Interop team meeting on Oct 23rd, 2025:

• Interop 2026 proposal review
  • Review any open questions on signals
• Review open PRs
• Review test change proposals
• Open question from Agenda for Oct 16th, 2025 #1205 on the threshold for removing a focus area from Interop based on spec changes

[zcorpan]

• Open question from Agenda for Oct 16th, 2025 #1205 on the threshold for removing a focus area from Interop based on spec changes

The team implementing the Navigation API in Gecko have reported a number of spec issues, of which one is a known web compat issue. The web compat issue has been fixed in Gecko (by violating the spec), so it doesn't currently block shipping.

whatwg/html#11786 is another web compat issue that was identified by WebKit. (This has been fixed.)

[nairnandu]

Attendees: @meyerweb, @nairnandu, @fantasai, @YulunWu, @saronyitbarek, @nt1m, @jgraham, @captainbrosset, @jakearchibald, @zcorpan, @foolip, @tantek, @stubbornella

Notes

• Interop 2026 proposal review
  • Review any open questions on signals
  • Proposal rankings might be delayed due to TPAC timelines
  • Proposal ranking changes to Dec 4th
  • Selection process would range from Dec 4th to the 18th - Readme updated.
• Review open PRs
  • [SAA] Fix test failures due to default cookie availability wpt#55018
    • merged
  • [Gecko Bug 1991265] Make push to replace load conversion follow spec. wpt#55399
    • Need follow-up
  • [Gecko Bug 1994673] Adjust try-tactic-alignment to test rendering. wpt#55489
    • merged
  • html: dynamic import entry should not be cached for HTTP error wpt#55505
    • Wont land this year
  • [Gecko Bug 1993997] Test RTCEncodedVideoFrame & RTCEncodedAudioFrame in webrtc-encoded-transform/idlharness. wpt#55588
    • Reviews required
  • Make software_renderer compute backdrop bounds like skia_renderer. wpt#55600
    • merged
  • Fix largest-contentful-paint/image-upscaling.html to use test_driver.… wpt#55609
    • Need further review from WebKit
  • [css-anchor-position-1] position-area normal alignment towards non-auto inset wpt#55722
    • Needs review from Chromium
• Review test change proposals
  • [Anchor positioning] anchors-visible tests should probably not be in interop? #998 - Chris and Elika to review offline with Emilio
  • Fix storage-access-api/requestStorageAccess-cross-origin-iframe-navigation-relax.sub.https.window.html to adhere to the specification #1212, Fix storage-access-api/requestStorageAccess-web-socket.sub.https.window.html to properly respect same-origin policy #1213 - need reviews
  • [Navigation API] several ordering-and-transition/ should expect to signal abort #1214 - need reviews
    • yulunwu: will reach out internally for review
    • nt1m: if there are regressions, we should exclude those tests
    • jgraham: can check internally. Agree, if there are significant changes
• tantek: Open question from Agenda for Oct 16th, 2025 Agenda for Oct 16th, 2025 #1205 on the threshold for removing a focus area from Interop based on spec changes
  • fantasai: the goal is to get agreement on spec and tests as part of the process. Might be hard to define that ahead of time? When the spec changes causes a large enough change in implementation, it would be prudent to exclude those parts of the feature from Interop. changes like the API level adjustments to the Anchor Positioning spec were acceptable because they were agreed upon and manageable within the timeline. A decision should be made on a case-by-case basis, considering whether a solid feature can be achieved by year-end and if the work required for implementations is reasonable
  • nt1m: How do we manage the comms to developers if we drop a focus area
  • fantasai: we could potentially say that we found massive issues and as such we chose to defer the implementation
  • jgraham: this is discussing the option of last resort. We should try and avoid situation where this becomes a question. Practically Interop is not a place to do spec work. We do on occasion take on spec work risk, but that should be an exception. The experience this year will affect how we assess spec risk for next year
  • tantek: there are several elements here - how big the change is, what is the timing of the change being proposed, how we communicate about the change to developers. We can trust developers with the truth of the matter. Interop is about increasing the quality of the platform and we should continue to lean on that principle. We can come up with a list of factors to assess the risk. Another factor is timing that should be considered. In addition, if a change is severe enough that requires re-architecting a feature that should be a clear indication.
  • stubbornella: we have dropped tests and that has worked to eliminate some risk. Could we do that more aggressively as opposed to dropping a focus area. However, it does beg the question if we are providing value to developers
  • tantek: what we dont want to do is to declare victory when developers cant really use it
  • stubbornella: core web vitals was another area where there was some struggle this year to align implementations and spec. The variance in implementations (Chromium) and tests from spec did pose a challenge
  • fantasai: we should only drop an area if its no longer valuable to developers
  • nt1m: when we commit to a focus area, there is a lot of effort that goes into it. Reviewing how we end up in a situation like that should be evaluated
  • jgraham: we have had a lot of mechanisms for mitigating issues like this. As an example, the experience with the Navigation API, where turning it on broke sites that depended on non-specified behavior, making it risky to ship the feature. Existing strategies might not work in this particular case.
  • tantek: agree. I think I'd like to reframe "drop" to "postpone" because in all likelihood we'll be able to fix up an area to succeed with it the following year. This might be something we want to consider for next year
  • zcorpan: an example of issues raised for Navigation API: WHATWG issues. Also see comment above