09:55 <discocaml> <Ada> what’s the difference between the function based and class based interfaces to eio? are functions Eio.Stdenv.net still using classes behind the scenes

10:29 <famubu> Is it possible to have nested lists in odoc?

10:41 <famubu> Is odoc meant to supercede ocamldoc?

12:37 <discocaml> <emiletrotignon> yes odoc is supposed to be a better and newer ocamldoc

14:29 <discocaml> <Ada> read the code, it is just a wrapper around classes

14:30 <discocaml> <Ada> still don't understand the avoidance around classes, they're very useful in moderation

14:35 <discocaml> <Kali> mostly performance concerns

14:43 <companion_cube> They're nice in some situations yeah

14:43 <discocaml> <Kali> i must admit i have never needed to use classes

14:43 <discocaml> <Kali> really goes to show how nice the module system is

14:46 <companion_cube> It's still the most flexible way to get dynamic dispatch

14:51 <discocaml> <lukstafi> I'd try `dune-release` and cross my fingers. But if you have a Dune rule to upload docs already and like doing releases with the GitHub web UI, then `opam-publish` does the rest (it has a much greater chance of working since it does a simple thing).

14:52 <discocaml> <Ada> @ilo_kali the performance concerns are understandable, but there are still some situations where they're applicable, i've done gui stuff with classes where the performance hit is acceptable for creating layouts

14:53 <discocaml> <Ada> wrapping classes with a module system just seems silly

14:57 <companion_cube> if you want inheritance or mixins, then the OO system is the most convenient

15:08 <discocaml> <Ada> eio seems to have a valid usecase for it and call methods infrequently enough for performance to be irrellevant

15:12 <discocaml> <Ada> i’m biased, i still use ruby as my glue language

15:16 <discocaml> <Ada> i’m biased, i still use ruby

16:47 <discocaml> <geoff> famubu: you can beat them by using the curly braces syntax instead of the sugared way

16:48 <discocaml> <geoff> n'est*

16:48 <discocaml> <geoff> Nest* damnit

17:10 <discocaml> <lecondorduplateau> t'es cramé le français

17:35 <companion_cube> another day of being mad at channels not being extensible

17:56 <adrien> you want BatIO? :P

18:07 <companion_cube> yes but well done

18:07 <companion_cube> that said I haven't looked in years, let me take another look

18:08 <companion_cube> oh yeah, no, all that weaks et and global stuff

18:30 <discocaml> <Ada> CCio when

18:38 <companion_cube> https://github.com/c-cube/ocaml-containers/blob/main/CHANGELOG.md#06

18:39 <companion_cube> 0.6 of containers :)

21:20 <discocaml> <diligentclerk> Let's go!!!!

21:21 <companion_cube> I do wish the *standard* library did it tho

21:22 <companion_cube> like give me the ability to make my own channels you cowards!

21:33 <adrien> just a bit of Obj should suffice, right?

21:44 <companion_cube> sadly no, they're full of C stubs

23:45 <discocaml> <regularspatula> Dang, I just hit the too many non-constant contructors error while compiling some code generated by ocaml-protoc (on the pg_qurey.proto file from libpg_query). Is there a way around that?

23:58 <discocaml> <regularspatula> Ah nvm I converted it to polymorphic variant and it was fine. But other errors popped up after that