| 9 Apr 2026 |
 lord_fomo | yeah i haven't dug in enough yet to provide any other info 🫣
do report back with solns if you get them though 🙏 | 00:22:50 |
 Neil Mayhew | Yes, Haskell is my day job | 00:22:53 |
| lord_fomo | lucky | 00:22:57 |
| lord_fomo | i think everyone should have to learn at least a bit of that lang | 00:23:10 |
| lord_fomo | heh, i wonder what you'd think of coconut | 00:23:27 |
| lord_fomo | transpiler for python | 00:23:32 |
| Neil Mayhew | It reshapes your brain 😄 | 00:23:35 |
| lord_fomo | been meaning to embrace more | 00:23:37 |
| lord_fomo | imo you can more or less guess if a dev has used haskell before based on their code in any lang 😛 | 00:24:06 |
| lord_fomo | starting to get offtopic, but if you get a chance leer at this,
This approach is more robust, as it uses an overlay to rewire other tools that depend on Nix, ensuring consistency across your setup. Tools like colmena or nix-eval-jobs, which often hardcode a reference to Nix, will now use the version provided by Lix. | 00:24:47 |
| lord_fomo | * starting to get offtopic, but if you get a chance leer at this,
https://coconut-lang.org/ | 00:24:57 |
| lord_fomo | i think it's a decent alt for getting peeps more into fp | 00:25:17 |
| Neil Mayhew | We need a good on-ramp, for sure, and Haskell is probably never going to be that. However, I do tell people that if you already know you want to learn FP, Haskell is probably the best way to learn because it forces you into good habits from the start. | 00:26:34 |
|  @illusaen:matrix.org left the room. | 03:27:21 |
 K900 | Lix and cppnix are not API compatible | 04:32:19 |
| K900 | So you can't just drop in Lix things into applications that expect cppnix | 04:32:31 |
| K900 | The .pc files were intentionally renamed to reflect that | 04:32:40 |
 niko | Even if you get through the trivial changes like renames or small restructures, Lix has an async executor now which means FFI usecases which used to work (and still work with cppnix) no longer work with Lix. In some cases requiring a major refactor since you can’t just call Lix from any thread at any point | 08:03:32 |
 splurged | I'm not using flakes and I don't plan to. Is there a way to make -f "<nixpkgs>" the default for nix run? | 09:17:27 |
 zoë (she/her) | apart from an alias i can't think of anything; however, if your <nixpkgs> is managed with npins (or some other pinning mechanism) through nixos or home-manager, you can use the trick outlined at the end of jade's post about npins and flakes | 09:28:21 |
| zoë (she/her) | if you don't manage your channels using any pinning, you could always write a small wrapper around nix-channel update that writes to the local flake registry to set it to point to the update nixpkgs path afterwards | 09:29:13 |
 raitobezarius | In reply to @splurged:matrix.org
I'm not using flakes and I don't plan to. Is there a way to make -f "<nixpkgs>" the default for nix run? There is if you use nix-run I think | 09:29:33 |
| splurged | Thanks for the tip. I do manage my nixpkgs with npins but for that to work I would need to enable flakes. :) | 09:30:46 |
| zoë (she/her) | * apart from an alias i can't think of anything; however, if your <nixpkgs> is managed with npins (or some other pinning mechanism) through nixos or home-manager, you can use the trick outlined at the end of jade's post about npins and flakes
(see my home-manager config)
| 09:31:08 |
| zoë (she/her) | * apart from an alias i can't think of anything; however, if your <nixpkgs> is managed with npins (or some other pinning mechanism) through nixos or home-manager, you can use the trick outlined at the end of jade's post about npins and flakes
(see my home-manager config for an example)
| 09:31:24 |
| zoë (she/her) | ah yep, that's a caveat. personally i have it enabled for convenience even though i avoid flakes; it's both for this trick as well as for when i need to work on others' projects that use flakes | 09:32:51 |
| zoë (she/her) | * | 09:33:01 |
| zoë (she/her) | i don't think it changes the behaviour of anything i use, so it doesn't really matter | 09:33:31 |
| zoë (she/her) | * i don't think it changes the behaviour of anything i use, so it doesn't really matter to me | 09:33:37 |
| splurged | Yeah. I totally understand. It would be the pragmatic approach. I'm trying to keep my experimental-features at a minimum. I'll keep that in mind when I get tired of typing -f "<nixpkgs>". | 09:38:46 |
