00:08 <dh`> figure out the mode restore, write a script that does it

00:08 <dh`> but also, for a roguelike it doesn't really make sense to run fullscreen I'd say so there's a large aspect of "don't worry about it"

00:09 <dh`> although maybe that's rubbish, I guess I normally expect roguelikes to appear in terminal windows

02:07 <twobitsprite> dh`: I'm using graphical tilesets, with IBM's code page 437 as the default

05:00 <discocaml> <jobhdez> Is ocaml better suited for high performance computing than Haskell?

05:06 <companion_cube> I'd say neither are great but oh well :D

05:48 <dh`> twobitsprite: right, that's become common

12:53 <discocaml> <zamzam0127_51933> for /f "tokens=*" %i in ('opam env') do @%i to update the current Shell enviroment. What mean is this?

12:55 <discocaml> <Kali> for every command output by the `opam env`, execute it

12:58 <discocaml> <zamzam0127_51933> What is command to execute that?

13:06 <discocaml> <Kali> this one

13:06 <discocaml> <Kali> that's what it means

13:16 <discocaml> <zamzam0127_51933> for /f "tokens=*" %i in ('opam env') do @%i . this one?

13:31 <discocaml> <Kali> yes

13:31 <discocaml> <Kali> that is what it does

17:15 <deavmi> Heyo, not sure who to report this to but the code on https://ocaml.org/docs/higher-order-functions (the email-domain splitting one). Doesn't run. I was eye balling it for some time and it didn't even make sense to begin with, an optional appearing out of no where to be consumed by Option.map.

17:18 <discocaml> <Kali> it probably mean nth_opt

17:18 <discocaml> <Kali> meant*

17:18 <deavmi> That would make sense, if that were the case. Good suggestion

17:18 <deavmi> thanks

17:19 <discocaml> <Kali> this page seems to be broken in a lot of places actually

17:20 <discocaml> <Kali> https://cdn.discordapp.com/attachments/436568060288172044/1270068837714493582/Screenshot_20240805-131952.png?ex=66b25b47&is=66b109c7&hm=8d2448e5d7350e01370103e8d6798985185dd3a37b6b512efc82e3d07e59526b&

17:20 <discocaml> <Kali> https://cdn.discordapp.com/attachments/436568060288172044/1270069006648479805/Screenshot_20240805-132032.png?ex=66b25b6f&is=66b109ef&hm=cef5c4eafc829714aba1254fced84a1599f827558c3a82383c58127f274d4078&

17:20 <discocaml> <gooby_clown> So with effects, I have two `unit Effect.t` effects, but when I do

17:20 <discocaml> <gooby_clown> ```ocaml

17:20 <discocaml> <gooby_clown> effc = fun (type a) (e: a Effect.t) ->

17:20 <discocaml> <gooby_clown> ...

17:20 <discocaml> <gooby_clown> match e with

17:21 <discocaml> <gooby_clown> | Effect1 | Effect2 as e -> Some (fun k -> ... continue k ())

17:21 <deavmi> (Couldn't open the link) but I do see some TODO_esque notes here and there. Either way, I'm enjoying lerning OCaml, the docs thus far have been great

17:21 <discocaml> <gooby_clown> ```

17:21 <discocaml> <gooby_clown> I omitted some bits, but my point is that it seems the type checker can't figure out that `a` in each of the branches is `unit`, so the call to `continue` fails because it's expecting value of type `a`. I can solve this by splitting up the branches and duplicating the code, but is there a way to tell the type checker that `a_fresh_1 = a_fresh_2 = unit`?

17:22 <discocaml> <Kali> this would be a better question for #advanced-help but anyway no, there are no or-patterns for combining locally abstract types at the moment

17:22 <discocaml> <Kali> you just have to repeat in separate branches

17:23 <discocaml> <gooby_clown> Ah mb about the channel, I'll go there in the future

17:23 <discocaml> <gooby_clown> But thanks for the answer 🙏🏽

17:23 <discocaml> <Kali> 👍

20:14 <twobitsp2ite> what's a good, simple configuration file format for ocaml applications? I know YAML exists, but is there something built into the ocaml standard library that's easy to use? (something like Python's built-in ini file module)

20:15 <twobitsprite> goddammit, every time my laptop sleeps, it reconnects with an alternate nick

20:49 <discocaml> <deepspacejohn> There’s nothing for special formats built into the ocaml stdlib. If you want something very simple that maps to command line arguments, then there’s Arg.read_arg that reads newline separated arguments from a file

20:59 <discocaml> <ada2k> sexp and yaml are the two i see

21:18 <twobitsprite> yeah, I think I'm just going to do with YAML

21:28 <discocaml> <Kali> i use toml sometimes

21:47 <twobitsprite> oh yeah, I forgot toml is a thing... that might be better actually

21:55 <dh`> is this a config file for human editing, or a config file for internal use only?

22:12 <discocaml> <yawaramin> i always use environment variables for internal configs

22:15 <discocaml> <ada2k> i prefer json for internal use, unless it’s only a few values

22:29 <dh`> for human-readable config, from what I've seen toml > yaml > json > xml

22:30 <dh`> when it's for automated use only it doesn't matter, use the least annoying library

22:30 <dh`> my experience with the yaml library (which uses the C libyaml under the covers) has been spotty, so all else being equal I'd suggest something else

22:31 <dh`> (spotty as in I had to write my own code to write out yaml files as I was getting binary trash in the output at unpredictable times)

22:34 <discocaml> <ada2k> yuck

22:34 <discocaml> <ada2k> maybe i should check out toml, i like yaml but it gets on my nerves sometimes

22:35 <discocaml> <ada2k> sexps don’t seem too bad

22:35 <discocaml> <ada2k> i’ve only used them with dune

23:25 <twobitsprite> Does tsdl not include the SDL image library? I don't see it anywhere in the docs

23:30 <twobitsprite> I've even browsed the src and don't any anything about the IMG library

23:51 <twobitsprite> I give up... You either write game code in C, or you write your own C interfaces......

23:51 <dh`> my understanding of toml is limited to 'it appears to have section headings like .ini files, which is a step up for readability'

23:51 <dh`> SDL_image is its own library, so it's not surprising it would need its own set of bindings

23:56 <twobitsprite> dh`: it's literally part of SDL... /usr/include/SDL2/SDL_image.h ... it just uses IMG_ as a prefix instead of SDL_