Conversation

I wonder whether you could do this through an effect system. You could have an 'overflow' effect that can be handled in a manner easily and cleanly chosen by the caller.

1

Brendan Zabarauskas

@brendanzab

Oh, good idea! I guess in practice you'd need an effect system lightweight enough in terms of runtime overhead… which could be a tough ask at the moment? But could be fun to experiment with?

1

So this is something I've been thinking about with my effect system which in the process of being born (preliminary codegen & typing works). Right now, I implement it as stack-walking but one of my goals is to monomorphise *everything* (no runtime HKTs) a consequence of... (1/n)

1

...this is that generic effects (i.e: where generic parameters aren't specified in an effect type, such as async/await) codegen doesn't really compose and can only be done at a best-effort level. I am very unhappy with this, so I've been thinking of changing effect types to (2/n)

1

be more akin to impl Trait / existential types, thereby statically promising that monomorphisation (& also inlining) is possible. I need to do a lot more thinking about this though because I think it's going to have far-reaching consequences for effects in general. Any thoughts?

1

Brendan Zabarauskas

@brendanzab

Hmm not sure. Pretty interested in what Effekt does with this stuff - I think they compile to a representation where handlers are injected as parameters?

2

12:37 PM · Apr 4, 2022Twitter Web App

I think a problem with this is that, again, monomorphisation can only occur on a best-effort basis (unless these 'parameters' are generic, as in Rust's Fn traits)

1

Replying to

and

What do you mean by best effort?

Replying to

and

Ahh I see the tweet above

Quote Tweet

· Apr 4

Replying to @jsbarretto @brendanzab and @jntrnr

...this is that generic effects (i.e: where generic parameters aren't specified in an effect type, such as async/await) codegen doesn't really compose and can only be done at a best-effort level. I am very unhappy with this, so I've been thinking of changing effect types to (2/n)

Replying to

and

As in: only when the compiler can infer that it's permitted to do so. I don't think this generalises very well because the compiler needs to keep track of the entire call stack, even in the presence of callbacks and conditionals and other weirdness like like that

2