Conversation

This has some good "nerd snipe" potential for programming language designers. In grad school, I was poking around possible solutions to this (though sadly didn't go far enough). A possible direction - modules could expose all of their implied imports to configurability.

Quote Tweet

Jonathan Müller

@foonathan

· Mar 30

I wrote my thoughts about writing a new type vs adding new functions to change behavior: saturating_add vs. saturating_int – new function vs. new type? foonathan.net/2022/03/behavi #cpp #cplusplus

Joshua Barretto

@jsbarretto

Apr 4

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.

Brendan Zabarauskas

@brendanzab

Apr 4

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?

Joshua Barretto

@jsbarretto

Apr 4

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)

Joshua Barretto

@jsbarretto

Apr 4

...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)

Joshua Barretto

@jsbarretto

Apr 4

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?

Replying to

and

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?

12:31 PM · Apr 4, 2022Twitter Web App

Replying to

and

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)

Brendan Zabarauskas

@brendanzab

Apr 4

What do you mean by best effort?

Show replies

Replying to

and

Combining that with staging could be interesting. Not sure about runtime reps - it's something I'm interested in with dependent types. But yeah in general I'd hope that the common case for effects is pretty static.

Brendan Zabarauskas

@brendanzab

Apr 4

Sixty has some cool stuff with being able to combine higher rank polymorphism with unboxed stuff which is interesting, dunno if it is useful at all in this case and can't remember the exact details.