Florian Angeletti

It is true that it is unfortunate that there is possibly two notions of inclusion at odd when reading ``` Module types do not match: Y does not include X...

Another tentative: ```ocaml Modules do not match: functor (M : Y) -> ... is not included in functor (M : X) -> ... The type of the functor parameter in...

Except that the term subtyping doesn't convey a lot of information beyond the existence of some partial order such that ` ¬(A ≺ B)` (aka calling a cat an eukaryote...

Anoter reproduction case would be: ```ocaml module M: sig val f: unit -> unit [@@alert test "1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111\n\ 2222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222222\n" ] end = struct let f () = () end let _...

I ma planning to investigate this issue after my work on the structured output for the compiler.

This feature wish is worth a discussion from my point of view (cc @goldfirere) in light with the recent proposal from open in type expressions.

In term of timing, this seems more reasonable to me to move the issue to the 5.3 milestone. @kayceesrk , @gadmm , please don't hesitate to tell me if I...

An important point that is missing from the discussion so far is that why should we want to make the `cardinal` function `O(1)` by decreasing the performance of more important...

Beware that the error in the testsuite for `pr7817_bad.ml` is a soundness issue.

After some code jigsawing b478b0b798f4f , I can confirm that this version of the PR never calls function reading the current environment from functions using an explicit level argument. Thus...