Consider splitting out typed function pointers

[aykevl]

Reading through the overview, I saw that there is a proposal for typed function pointers.

Is there anything that typed function pointers provide that is not already covered by anyfunc? I have a feeling that the part of the spec that describes typed function pointers is leaning a bit too much towards supporting the full Java type system. While I think that typed function pointers can be very useful for performance reasons, I don't see how they are related to GC except for making the type system even more complex. Falling back to anyfunc (with a small performance hit) and simply requiring the language to perform its own language-dependent downcasts if needed seems like a simpler solution right now.

Possibly related: #32

[titzer]

AFAICT, there are no downcasts from anyfunc at the moment. What would they downcast to, if there are no typed function pointers? Or do you propose to write any anyfunc into a table and then do a call_indirect on it? That is pretty slow.

[rossberg]

I actually just proposed splitting out typed function references from the GC proposal at last weeks CG meeting. See this branch for a tentative sketch of such a standalone proposal.

[aykevl]

Or do you propose to write any anyfunc into a table and then do a call_indirect on it? That is pretty slow.

That's what I would propose, yes. Yes, it will be slow. But similar to interior pointers in #59, I don't think this should be part of a MVP. (And if this is part of the MVP, why not interior pointers?)

I actually just proposed splitting out typed function references from the GC proposal at last weeks CG meeting.

Great! And good to hear there is ongoing work!

[binji]

Typed function references has been moved to its own repo, here: https://github.com/webassembly/function-references. Closing.