!9IQChSjwSHXPPWTa:lix.systems

Lix

736 Members
Lix user channel. Feel free to discuss on-topic issues here and give each other help. For matrix.to links to the rest of the Lix channels, see: https://wiki.lix.systems/books/lix-organisation/page/matrix-rooms225 Servers

Load older messages


SenderMessageTime
20 Dec 2024
@k900:0upti.meK900I don't know if there is really any specific thing 13:27:13
@k900:0upti.meK900If you just want to have a thing to throw compute at 13:27:24
@samrose:matrix.orgsamrose

well that was an example, but also

What would be the most helpful thing to do for contributing ongoing to lix for those of us who cannot carve out development time?

13:28:07
@k900:0upti.meK900Again, I'm not sure 13:29:14
@k900:0upti.meK900 I don't really know what you can actually do that would help except throwing compute at stuff 13:29:34
@k900:0upti.meK900Because most other things require person-hours13:29:59
21 Dec 2024
@federicodschonborn:matrix.orgLEGO® Worm™ modules = [ lix-module.nixosModules.default ] goes in your NixOS configuration 00:32:36
@federicodschonborn:matrix.orgLEGO® Worm™
nixosConfigurations.whatever = nixpkgs.lib.nixosSystem {
  modules = [
    lix-module.nixosModules.default
    # ./configuration.nix and others
  ];
};
00:33:57
@federicodschonborn:matrix.orgLEGO® Worm™Ah you're using Colmena00:34:36
@federicodschonborn:matrix.orgLEGO® Worm™
{inputs, ...}:

{
  imports = [
    inputs.lix-module.nixosModules.default
  ];
}

Should work, I don't use Colmena but it should have an equivalent of nixosSystem's specialArgs you can pass inputs into

00:35:41
@dmiskovic:matrix.org@dmiskovic:matrix.org left the room.05:14:08
@tammi:greyseal.euTammi (ey/em) ohhh thanks ^^ 05:43:14
@tammi:greyseal.euTammi (ey/em)will try05:43:21
@tammi:greyseal.euTammi (ey/em) hmm 05:44:22
@tammi:greyseal.euTammi (ey/em) also this would probably not have dot inputs in the beginning 05:45:29
@tammi:greyseal.euTammi (ey/em)if i put it into the inputs of a machine configuration05:45:43
@tammi:greyseal.euTammi (ey/em) ohhh i guess that may be my issue here: drop the input. 05:46:21
@srtcd424:auxolotl.orgtc424 (Steve D)OT, but could someone with a brain tell me if https://github.com/nvmd/nixos-meshtastic/pull/1 looks like a sensible way to access a flake's outputs from one of its supplied nixosModules? I can't help thinking there's something really simple I'm missing ..23:49:37
@antifuchs:asf.computerantifuchsI've seen it done (and done it) another way, have the flake load the nixos module wrapped in a function that takes the flake as a parameter23:53:52
@antifuchs:asf.computerantifuchse.g. https://github.com/boinkor-net/tsnsrv/blob/main/nixos/default.nix#L1 and https://github.com/boinkor-net/tsnsrv/blob/main/flake.nix#L8023:54:25
@antifuchs:asf.computerantifuchsthat means your nixos module is now incompatible with the non-flake way of doing nixos modules, so strict downside. another way would be to mandate users include an overlay where you provide the package23:55:17
@srtcd424:auxolotl.orgtc424 (Steve D)Ah, OK, so no obvious, clean answer then? Feels like another weird missing thing from flakes ..23:56:40
@srtcd424:auxolotl.orgtc424 (Steve D)OTOH makes me feel like slightly less of a dunce!23:56:56
@antifuchs:asf.computerantifuchsyeah, I agree, it's kinda a weird missing link. took me some time to accept that there's no really good way to do it either (:23:59:18
@antifuchs:asf.computerantifuchsmaybe the intent was that people write all their stuff in flake.nix (where this is easier, and maximally backwards-incompatible)? so unclear23:59:44
22 Dec 2024
@srtcd424:auxolotl.orgtc424 (Steve D)I'm still new enough to Nix that I don't feel I should comment on the flakes controversy, but the more I actually try to use them for what I feel should be fairly simple things, the more they feel very unfinished, at best :(00:00:47
@antifuchs:asf.computerantifuchsthat's the general vibe, yep. some un-thought-through consequences left in there, some wishful thinking, etc.00:04:40
@cpli:fairydust.spacecpli ☎️8754 changed their display name from cpli to cpli ☎️8754.00:27:33
@kuruczgy:matrix.orgkuruczgy

What's the status of sha256 git object format support for flakes in Lix? I get an error like this when trying to evaluate something from a flake (with a clean working :
error: hash 'ae3c3c3b9dd70146d51991f2c5b9cdb4af7e65104f07e8376ac70ab2807250b0' has wrong length for hash type 'sha1'

Is there some easy workaround?

I found this CppNix issue: https://github.com/NixOS/nix/issues/9980
To me it seems like it should not be very hard to fix this, the only thing nix should need from the repo is the commit hash, and based on the error message it can clearly already query that...

00:54:13
@kuruczgy:matrix.orgkuruczgy *

What's the status of sha256 git object format support for flakes in Lix? I get an error like this when trying to evaluate something from a flake (with a clean working directory) :
error: hash 'ae3c3c3b9dd70146d51991f2c5b9cdb4af7e65104f07e8376ac70ab2807250b0' has wrong length for hash type 'sha1'

Is there some easy workaround?

I found this CppNix issue: https://github.com/NixOS/nix/issues/9980
To me it seems like it should not be very hard to fix this, the only thing nix should need from the repo is the commit hash, and based on the error message it can clearly already query that...

00:56:31

Show newer messages


Back to Room ListRoom Version: 10