Re: [whatwg/dom] Expose an `aborted` promise on AbortSignal? (#946)

FWIW, `signal.race(..)` would serve *some* of my usages (in CAF), but I'm not sure it would serve all of them. Here are a couple of concerns:

1. CAF includes a `signalRace(..)` utility (and a `signalAll(..)`) that allow you to pass an array of multiple signals and do a race (or all) across them. I'm not sure how I would implement those utils with only a `race(..)` on each signal that only accepts a signel promise. Maybe I'm missing it, but I'm not sure how to implement these utils with `signal.race(..)`.

2. CAF "extends" abort signals in that it allows you to pass along a "reason" value for the cancellation (e.g, `token.abort("took too long")`). It uses this value to reject the `signal.pr` promise, so that any userland code doing a `catch(..)` or `try..catch` would see that abort reason.

   If I had a promise primitive (similar to my current `signal.pr`), then I can monkey-patch in the reason extension by adding a `catch(..)` or `finally(..)` on it that then substitutes the abort reason as a resolution value. But if all I have is a black-box `race(..)` method that hides the promise, I don't think I can accomplish that. There would be an inconsistency where the CAF mechanisms would expose this reason value, but if userland code observed the signal itself by calling their own `.race(..)`, they wouldn't see the reason come through. That's unfortunate.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/whatwg/dom/issues/946#issuecomment-845943237