Update all examples, and the lt for the type, to the new underscore'd token

[tabatkins]

This changes the async_iterable<> type to async_sequence<>, to avoid confusion with the interface declaration, and change the async iterable<> interface declaration to async_iterable<>, to avoid future LL(1) issues.

• At least two implementers are interested (and none opposed):
  • Apple per @annevk approval
  • Mozilla per @saschanaz approval (in Rename the async iterable declaration to async_iterable to match the type name. #1504)
• Tests are written and can be reviewed and commented upon at:
  • N/A, this is a grammar change in WebIDL itself
• Implementation bugs are filed:
  • Chromium: https://g-issues.chromium.org/issues/356891478#comment3 and https://g-issues.chromium.org/issues/433299826
  • Gecko: https://bugzilla.mozilla.org/show_bug.cgi?id=1911295#c2 and https://bugzilla.mozilla.org/show_bug.cgi?id=1978444
  • WebKit: https://bugs.webkit.org/show_bug.cgi?id=277548#c3 and https://bugs.webkit.org/show_bug.cgi?id=296291
  • Deno: async iterable should be converted to async_iterable denoland/deno#30020
  • Node.js: Implement support for async_iterable<T> WebIDL type nodejs/node#54175 (comment)
  • webidl2.js: feat: add support for parsing async_sequence<T> type w3c/webidl2.js#775 (comment)
  • widlparser: TODO
• MDN issue is filed: N/A
• The top of this comment includes a clear commit message to use.

---

Preview | Diff

[saschanaz]

Thanks!

[tabatkins]

Thanks for catching that I'd accidentally caught some of the iterable declarations. Reverted back.

Changed over to async sequence<>, per discussion in #1504. Edits got a lot bigger, since I changed terminology across the spec to match.

[tabatkins]

Ughhhh, no, this still violates LL(1). I misread the restriction - it's not that an entire production can't be a prefix to another production, it's that no two productions in an alternative can share a prefix.

Guess it'll have to be async_sequence<>, then. :(

[annevk]

@domenic @MattiasBuelens any final thoughts? Changes look reasonable to me.

[MattiasBuelens]

LGTM.

So we landed on async_sequence<T> for the type (with underscore) and async iterable<T> for the declaration (with space)?

[annevk]

Ah, I suggested making both use underscores in #1504. I don't think @tabatkins has had a chance to see that yet.

[tabatkins]

All right, updated to make both use underscores - async_iterable<> for the interface declarations, and async_sequence<> for the type.

[annevk]

We should probably include the PR template now. I suspect we can mostly reuse #1489 except if bugs have already been fixed.

There's also a build failure remaining.

[tabatkins]

Right, now that this PR isn't just a typo fix but an actual change, let's fill in the boilerplate properly. ^_^ Done.

The build failure will also go away if you kick it; Bikeshed updated its data files to prevent this issue.