Editorial: remove redundant contains check in AbortSignal.any()

[vinhill]

Step 4.2.2. of create a dependent abort signal prevents duplicated signals in source signals for example in the following snippet

const controller = new AbortController();
const signal = AbortSignal.any([ controller.signal, AbortSignal.any([controller.signal]) ]);

so maybe the same should hold for

const signal2 = AbortSignal.any([ AbortSignal.any([controller.signal]), controller.signal ]);
const signal2 = AbortSignal.any([ controller.signal, controller.signal ]);

To my knowledge, duplicated signals in source signals should not have any impact due to step 1 of signal abort ensuring signals are only aborted once. So this change should cause no behavior change.

PR Questions

• At least two implementers are interested (and none opposed):
  • …
• Tests:
  • There already is https://github.com/web-platform-tests/wpt/blob/2fcb60d7e9f294400c6274d964428621ca6cd3d2/dom/abort/resources/abort-signal-any-tests.js#L81C1-L88C72
• Implementation bugs are filed:
  • Chromium: …
  • Gecko: …
  • WebKit: …
  • Deno (only for aborting and events): …
  • Node.js (only for aborting and events): …
• MDN issue is filed: …
• The top of this comment includes a clear commit message to use.

---

Preview | Diff

[annevk]

Isn't this what https://infra.spec.whatwg.org/#sets already takes care of?

[vinhill]

My bad, it does. Would it make sense then to remove step 4.2.2.?

[annevk]

@shaseley might know. Given that it's followed by two appends it's not immediately clear at least.

[shaseley]

4.2.2 should be safe to remove since both 4.2.3 and 4.2.4 would be no-ops for a duplicate source. I don't think the difference was intentional (and I remember purposely omitting the check in 4.1 because of set behavior); there might have been a previous version of the any() PR where skipping the steps mattered, but that's not the case now -- so removing to match 4.1 SGTM.

[vinhill]

Is it fine to do the change in this pull request, or should I open a new one? Can I omit implementation bugs given this change should have no impact?

[annevk]

It's fine here. And yes, no need for bugs on editorial changes.

We'll have to prefix it with "Editorial: ", but I can do that when landing.