Sync interfaces/ with @webref/idl 3.33.0

[github-actions]

This automated pull request updates interfaces/*.idl from @webref/idl 3.33.0.

Before merging, please check that any tests that depend on the updated IDL files still work. In particular, check for gray boxes as the wpt.fyi check status, which usually means that some tests have regressed.

If additional changes are needed, please manually create another PR based on this one.

See the workflow for how this pull request was created, and the README for how the IDL files in this directory are used.

[foolip]

This will require getting rid of the tentative test for popover. @josepharhar @mfreed7 can either of you take a look at doing that in a separate PR that takes the changes to html.idl from this PR and makes the additional changes required to test popover as part of the big HTML idlharness test?

[mfreed7]

This will require getting rid of the tentative test for popover. @josepharhar @mfreed7 can either of you take a look at doing that in a separate PR that takes the changes to html.idl from this PR and makes the additional changes required to test popover as part of the big HTML idlharness test?

So I put up a CL to do the .tentative.html -> .html rename. As for the idlharness test, what's required? Note that there's already an idlharness test specific to popover:

https://wpt.fyi/results/html/semantics/popovers/idlharness.tentative.html?label=master&label=experimental&aligned&view=subtest&q=popovers

Perhaps that should just get merged somewhere?

[foolip]

@mfreed7 this PR brings in the IDL for popover from the HTML spec, so I'm guessing the dedicated test isn't needed any longer?

It is a bit unfortunate that it all ends up in one big test of course, but I don't know how to improve on that unless IDL is grouped by feature in specs or something like that.

[miketaylr]

OK, so the web-animations idl regression seems intentional - AnimationTimeline was moved from web-animations.idl to scroll-animations.idl

For Safari, the change in webrtc-stats/supported-stats.html looks related to 61989c5, so that might have revealed a bug?

[foolip]

The change in webrtc-stats/supported-stats.https.html seems weird. The only change to the IDL shouldn't have affected any tests. Maybe that test is already flaky?

[miketaylr]

The only change to the IDL shouldn't have affected any tests. Maybe that test is already flaky?

Yeah, I confirmed locally that checking out the idl before that change doesn't affect test results. I re-ran the test about 25 times here, but wasn't able to repro a flake, but it's the most likely solution.

[miketaylr]

I think we can land this as-is, but let's let CI finish running.

[miketaylr]

@jgraham is it normal for wpt-firefox-nightly-stability to take more than 8 hours?

[miketaylr]

All the Firefox failures relate to webtransport, which isn't supported in Firefox (according to MDN), so let's go ahead and land this.

[miketaylr]

(but probably need @foolip or @jgraham to force merge it!)