00:24 <companion_cube> rgrinberg: would it make sense ot have a "let-destruct" action in ocaml-lsp-server?

00:25 <companion_cube> for irrefutable destruct using let

00:26 <rgrinberg> So what would that do? Take an expression and produce let expr in expr?

00:27 <companion_cube> let {x; y; z} = expr in

00:27 <companion_cube> ah yes, sorry. I use that pattern almost exclusively with records.

00:27 <companion_cube> (it's defensive, when you need to use all the fields)

00:29 <rgrinberg> Yeah, I suppose this should be the default behavior for destruct on records and on types that are only a single constructor

00:32 <rgrinberg> companion_cube make a feature request in the merlin repo. destruct is implemented there, I only wrap it in lsp

00:32 <companion_cube> 👍

00:33 <companion_cube> ah! https://github.com/ocaml/merlin/issues/1162

04:59 <d_bot> <Dorissa> Hello, I'm Doris an outreachy intern applicant

07:58 <Anarchos> Hello

10:31 <d_bot> <andreypopp> hm... odig 0.7 PR to opam repo was merged 5 days ago but `opam update` doesn't surface it yet... is https://opam.ocaml.org repo cache that stale?

11:25 <d_bot> <undu> Yea, the server that updates it seems to beon it's last lef

11:25 <d_bot> <undu> I think people were looking to replace it

13:07 <d_bot> <saysayo> https://discord.gg/cFvB6mXX, you should check out the outreachy channel for more info

13:59 <smondet[m]> Emacs users: do you have a setting for tuareg or caml-mode to completely disable all emacs-based indentation?

14:47 <d_bot> <ggole> I use ocp-indent, which doesn't use emacs' indentation machinery iirc

15:04 <smondet[m]> yes, and I use ocamlformat everywhere now, my question is how to get Emacs' one out of the way

15:16 <hornhack> hi, how do I print without newline?

15:24 <d_bot> <darrenldl> Printf.printf?

15:27 <d_bot> <RegularSpatula> You mean you want to disable the indenting that happens when you hit return/enter? (Like the basic `C-j` but when you sit return?)

15:29 <Serpent7776> I've just realised polymorphic (<) magically worked for my custom record type without me realizing it. In general it could be subtly wrong though.

16:30 <d_bot> <hcarty> hornhack: `print_string` is another option

16:30 <d_bot> <Bluddy> It's super important not to rely on polymorphic comparison. I think the solution adopted by Base and Containers of disabling it is probably for the best.

16:31 <d_bot> <Bluddy> You should only use polymorphic comparison explicitly when you think it's appropriate.

16:32 <hornhack> ah thx

16:33 <hornhack> you still can do Poly. to compare with Base, right?

16:37 <d_bot> <NULL> I mean, as long as you consider your type constructors to be injective (i.e. you never identify two different values as equivalent), I don't see the issue with polymorphic comparisons

17:26 <Serpent7776> I find the polymorphic behaviour rather surprising (at least for record types), because it might not do the right thing. It's just something one needs to be aware of.

18:52 <companion_cube> A good practice is to expose an equal function for your types, even it's just an alias to =

18:53 <companion_cube> This way you can change it later

18:57 <Serpent7776> I guess I need to learn to use modules, for now I'm just writing scripts

20:31 <d_bot> <andreypopp> I've used menhir for parsing but never used it with error recovery and hence the question — how good is it vs GLR based parsers (tree sitter, lezer)? Considering the case I want to implement LSP (completion, go to def) which should work on invalid/incomplete sources. I know ocamllsp relies on merlin which uses OCaml grammar (which I think is defined used menhir now) but not sure how much work is done specific to OCaml to hav

23:32 <hornhack> hi, I have a question about menhir

23:32 <hornhack> I just want to parse a file into a list of words

23:32 <hornhack> is this correct?

23:33 <hornhack> file : | {[]} | f = file; w = word {w::f};

23:33 <hornhack> menhir wants me symbol file is never accepted

23:33 <hornhack> warns*

23:38 <hornhack> hmm it doesnt warn that anymore

23:38 <hornhack> something I have to do with lexer

23:39 <hornhack> now it just runs forever