Editorial: remove unused %AggregateErrorPrototype% intrinsic

[ljharb]

This is non-normative since it is not a function, and nothing uses it in this or any other spec. It was mistakenly added in #2040.

[ljharb]

Just realized why none of these aren't auto-linking - https://tc39.es/ecma262/#sec-nativeerror used to dynamically refer to them, but now it uses the dotted intrinsic form.

We could save this one for #2056, and leave it now for consistency, or merge this now. Up to yall.

[michaelficarra]

I say just throw it in to #2056. In the editor call, I would like to talk about the poor usability of our dynamic definitions of errors and typed arrays.

edit: Result of this discussion: #2135

[ExE-Boss]

I suggested this in #2040 (comment), but it was rejected by @chicoxyzzy because of tc39/proposal-promise-any#37 (comment).

[ljharb]

Indeed :-) I did some grepping and couldn't find uses of this one, so I made the PR; but then I realized it was the NativeError magic.

[caridy]

I'm getting confused by the different comments here. My assumption is that you cannot remove this since it is an undeniable intrinsic. Some objects will be created with __proto__ set to this intrinsic object.

[ljharb]

@caridy i'm not sure why that makes a difference? %AggregateError.prototype% will always refer to this object; the non-dotted form is a legacy way of referencing it. An intrinsic no longer needs to be "well-known" to be accessible via the % syntax.

[caridy]

@ljharb thanks for the clarification. It makes perfect sense now!