Editorial: use idiomatic phrasing in CleanupFinalizationRegistry

[michaelficarra]

No description provided.

[bakkot]

I find this more awkward to read FWIW.

[michaelficarra]

@bakkot How about this?

[github-actions]

The rendered spec preview for this PR is available as a single page at https://tc39.es/ecma262/pr/3829 and as multiple pages at https://tc39.es/ecma262/pr/3829/multipage .

[bakkot]

Still just as awkward. Compare the before and after as prose rather than pretending it's a computer program. The earlier prose is much easier to understand.

[michaelficarra]

The step Choose any such _cell_. is what bothers me about the before. How would you phrase that to not use an alias that should only be visible to the "contains"?

[bakkot]

I don't understand what bothers you about it. That is a totally normal phrasing.

[michaelficarra]

The phrasing is fine. It's just in this case it's referring to an alias that I don't think it should be able to refer to. I don't consider the form in the while step to be declaring an alias that is visible to later steps.

[bakkot]

Well, that is how prose works.

[nicolo-ribaudo]

The phrasing is fine. It's just in this case it's referring to an alias that I don't think it should be able to refer to. I don't consider the form in the while step to be declaring an alias that is visible to later steps.

Instead of Choose any such _cell_ we could use Let _cell_ be any of such Records, so that it's declaring _cell_ and not referencing the one "declared" in the While prose that has not one value but many possible values.

[nicolo-ribaudo]

Suggested change

	1. Let _candidates_ be a List whose elements are the elements of _finalizationRegistry_.[[Cells]] whose [[WeakRefTarget]] is ~empty~.
	1. If _candidates_ is empty, return ~unused~.
	1. Let _cell_ be either ~none~ or one of the elements of _finalizationRegistry_.[[Cells]] whose [[WeakRefTarget]] is ~empty~, chosen in an implementation-defined manner.
	1. If _cell_ is ~none~, return ~unused~.
	1. Remove _cell_ from _finalizationRegistry_.[[Cells]].
	1. Perform ? HostCallJobCallback(_callback_, *undefined*, « _cell_.[[HeldValue]] »).

Would this make it easier to follow, with a similar readability as the original prose?