<Leonidas>
companion_cube: If I had the choice, I'd still prefer alcotest over ounit because ounit is a very java/xunit style library whereas the composable testables in alcotest are much nicer API-wise
<Leonidas>
also it doesn't overload silly operators just because you can write `foo >::: bar`
<d_bot>
<Continuation Calculus> Do you plan to maintain it, or was it just as a demonstration of the approach of keeping one process open?
olle has quit [Remote host closed the connection]
mro has quit [Remote host closed the connection]
mro has joined #ocaml
mro has quit [Remote host closed the connection]
<companion_cube>
Leonidas: the API is bad in both cases imho
<companion_cube>
I find the first class modules of alcotest weird
<Leonidas>
companion_cube: where do you need to use first-class modules?
<companion_cube>
The checkable thing
<companion_cube>
The ounit actual test functions are more straightforward to use
<Leonidas>
Testables?
<d_bot>
<cemerick> The former is the plan
<d_bot>
<cemerick> It's what will keep using e.g. dream feasible long term
<Leonidas>
I find these pretty straightforward, there's combinators and you can implement your own TESTABLES. it is kinda like Cmdliner or Fmt, just with simpler types.
<d_bot>
<undu> It's a barrier for newcomers: 1. Knowing what is a "testable" and how to define it, 2. Knowing what is Fmt.t and how to define one
<d_bot>
<undu> I think this could be eased somewhat
<Leonidas>
Sure, but the most common testables are predefined anyway
<d_bot>
<EduardoRFS> let f: type n a . (n, n s) eq -> a = ...
<d_bot>
<EduardoRFS> ```
<d_bot>
<EduardoRFS> I understand that there is valid cases for n, but from what I understand it's only the case under recursive types, is recursive types supposed to be sound in OCaml?
<d_bot>
<octachron> `let f: ... = fun x -> let rec never x = never x in never x`?
<d_bot>
<octachron> But no, the function `f` can never return.
<d_bot>
<octachron> Recursive types are sound in OCaml.
<d_bot>
<EduardoRFS> sad
<d_bot>
<EduardoRFS> Was having fun doing Coq -> OCaml, but this broke me