03:29 <discocaml> <astreamingcomesacrossthesky> What am I missing about `ocaml -I`? I have compiled ocaml libs in a directory I'm passing as the arg to -I but it doesn't work. However, if I add the extra path segment pointing to the directory containing the .cmi files, it does. Am I doing something stupid or does it only look at the top level of the specified directory?

03:48 <discocaml> <astreamingcomesacrossthesky> What am I missing about `ocaml -I`? I have a directory containing directories of compiled ocaml libs. I pass this parent dir as the arg to -I but it doesn't work. However, if I add the extra path segment pointing to the directory containing the .cmi files themselves, it does. Am I doing something stupid or does it only look at the top level of the specified directory?

04:09 <discocaml> <chrisarmstrong> I don't think it works recursively; I believe you need to pass in each directory containing `.cmi` files as a parameter

04:22 <discocaml> <astreamingcomesacrossthesky> Good to know. I'm using a non-opam install of OCaml and I'm trying to wrap my head around how libraries are resolved. I suppose I can just add all the desired dirs to .ocamlinit so I don't have to specify them each time the top level is started

04:32 <discocaml> <chrisarmstrong> If you insist on not using opam, at least use `ocamlfind` to locate your libraries when calling the top-level

08:10 <discocaml> <inline_93060_96044> ocaml -I just works for flat folder structure, i just tried it on msys2 on windows

10:59 <discocaml> <pierrels> Is it possible to use `Alcotest.skip ()` inside of a QCheck test to make it skip the overlying Alcotest test ? Something like :

10:59 <discocaml> <pierrels> ```ocaml

10:59 <discocaml> <pierrels> let qcheck_test = QCheck.Test.make ~count:100 ~name:"test" QCheck.(int) (fun k -> Alcotest.skip ()) in

10:59 <discocaml> <pierrels> let alcotest_test = QCheck_alcotest.to_alcotest qcheck_test in

10:59 <discocaml> <pierrels> Alcotest.run "?" [ "test", [alcotest_test]]

10:59 <discocaml> <pierrels> ```

11:00 <discocaml> <pierrels> I would like it to mark this test as skipped, instead it is marked as a failed test because it raises the Skip exception from Alcotest core.

15:45 <discocaml> <drupyog> Credits to Sebastien Micheland for the idea. Special dédicace to @companion_cube , Fold and simple filters as let-operators:

15:46 <discocaml> <drupyog>

15:46 <discocaml> <drupyog> ```ocaml

15:46 <discocaml> <drupyog> let (let<@) (lst, acc) expr = List.fold_left (fun a b -> expr (b,a)) acc lst

15:46 <discocaml> <drupyog> let (and@) (lst : _ list) acc = (lst, acc)

15:46 <discocaml> <drupyog> let (and+) l l' = List.map2 (fun x y -> x,y) l l'

15:46 <discocaml> <drupyog> let (and*) l l' = List.concat_map (fun x -> List.map (fun y -> x,y) l') l

15:46 <discocaml> <drupyog> let (and?=) l v = List.filter_map (fun x -> if x = v then Some (x,()) else None) l

15:46 <discocaml> <drupyog> ```

15:46 <discocaml> <drupyog>

15:46 <discocaml> <drupyog> ```ocaml

15:46 <discocaml> <drupyog> let<@ n=[1;2;3]

15:46 <discocaml> <drupyog> and=? () = 2

15:46 <discocaml> <drupyog> and* m=["foo";"bar";"baz"]

15:46 <discocaml> <drupyog> and@ acc="stuff: " in acc^m ^ string_of_int n ;;

15:46 <discocaml> <drupyog> - : string = "stuff: foo2bar2baz2"

15:46 <discocaml> <drupyog> ```

15:47 <discocaml> <drupyog> (I might have a bunch of let-operators to write fake-comprehension on Iters and Seqs in *several* different projets)

15:52 <companion_cube> 😂

15:53 <companion_cube> I really use `let@` a lot these days

15:55 <discocaml> <drupyog> What's `let@` ?

15:55 <companion_cube> omg

15:55 <companion_cube> `let (let@) = (@@)`

15:55 <olle> Just go full Forth at that point

15:57 <companion_cube> look at what Gleam did btw

15:57 <companion_cube> to just avoid most of the pain with only the equivalent of `let@`

15:57 <discocaml> <deepspacejohn> oh yeah let@ is great. I use it all the time.

15:58 <discocaml> <drupyog> Wait, how do you us that ?

15:59 <companion_cube> let@ ic = with_open "foo" in

15:59 <discocaml> <drupyog> oh right

15:59 <companion_cube> let@ x = Lock.with protected in

15:59 <discocaml> <drupyog> that explains a lot why I don't see the point

15:59 <discocaml> <drupyog> (I don't write code like that a lot)

15:59 <companion_cube> let@ _sp = Trace.with_span "doing foo" in

15:59 <companion_cube> heh :)

16:00 <discocaml> <drupyog> (however, I do walk over really weird iteration spaces a lot ...)

16:03 <companion_cube> heh :)

16:04 <companion_cube> I suspect `let@` would work nicely with `Iter`

16:04 <companion_cube> already CPS

16:05 <discocaml> <drupyog> Well, it's basically Iter.iter

16:07 <discocaml> <drupyog> @companion_cube jokes aside, the let-operators for fold `let@>` and `and@` would fit nicely in lot's of container's stuff

16:08 <discocaml> <drupyog> (probably with other names)

16:08 <companion_cube> yeah it reads weirdly I guess 😅

16:09 <discocaml> <drupyog> it's actually quite nice, you declare you are doing a fold, and the `and@` denotes the accumulator

16:29 <discocaml> <yawaramin> on a related note, https://github.com/ocaml/ocaml/pull/13698

16:38 <olle> yawaramin, can you add a use-case with and without this op?

16:39 <olle> Wait I have to leave anyway

16:39 <olle> Still tho

16:39 <discocaml> <smondet> The older PR has a bunch of examples https://github.com/ocaml/ocaml/pull/9887

16:40 <olle> smondet cool, thanks!