Missing super identifier

[nzakas]

In class methods, there should be a super identifier prepopulated in the scope (similar to this and arguments).

[michaelficarra]

That isn't how super works. It must be the target of a call or a member access. It is not a regular identifier, doesn't resolve how other identifiers do in the scope chain, and shouldn't show up in escope's output.

[nzakas]

It needs to show up as a reference somehow to have a complete picture of what's happening. How would you recommend that be done?

[michaelficarra]

Since super resolves through home objects, you should be able to statically resolve these references. I'm not sure what you're asking.

edit: Ah, I think I understand. You can use the __proto__ setter to change the [[Prototype]] value of the class's prototype object, so the reference is unknown until runtime. Is that what you were asking about?

[nzakas]

Not exactly. I suppose what I'm really asking for is for super to act like this in escope. this doesn't exist as a scope variable and this references don't count as references in escope. The problem right now is that super does count as a referenced, and that's throwing things off.

[michaelficarra]

Agreed. 👍

[Constellation]

@nzakas:

LGTM. escope should handle super reference specially.

[Constellation]

@nzakas, @michaelficarra

How super is handled in estree project?
We need to align it.

[nzakas]

super is just represented as an Identifier, same as this.

[Constellation]

I remember that this is represented as ThisExpression, right?
Anyway, I'll implement it with assumption that super is represented as Identifier.

[nzakas]

Oh yeah, there's ThisExpression. So far, there's no SuperExpression. I just filed an issue to get clarification: estree/estree#54

[Constellation]

@nzakas
Cool :)