| 23 Nov 2025 |
 whispers (it/fae) | * if you use one of the packages in the set, then you want to use the lixPackageSets version. all it does is give the lix version of that package, which is sometimes just overriding the nix version to be lix instead of cppnix and sometimes a full fork (as in the case of nix-eval-jobs). otherwise, each of those will use cppnix internally
(aiui the exceptions are aws-sdk-cpp, boehmgc, and editline, since those are internal to lix) | 22:52:35 |
 Sofie š³ļøāā§ļø (she/her) | Does anyone have a ddc/ic kernel driver for Linux 6.17+ ?
I do not want to patch a kernel driver which used the legacy I2C_CLASS_DDC to use the newer i2c dev scan implementation š
| 23:35:51 |
| Sofie š³ļøāā§ļø (she/her) | I see ,,,, you have to nowadays hook into DRM helpers to do the equilevevnt š | 23:41:00 |
| Sofie š³ļøāā§ļø (she/her) | I mean, I can implement it | 23:53:33 |
| 24 Nov 2025 |
 K900 | Lix 2.94 is in nixos-unstable-small now | 10:03:21 |
| K900 | Possibly also nixos-unstable but I have not checked | 10:03:28 |
|  Aijokey joined the room. | 10:19:53 |
 Coca | Not yet in nixos-unstable | 11:09:58 |
 Linux Hackerman | is it intentional that there's no tag in https://git.lix.systems/lix-project/nixos-module for 2.94.0? | 12:06:13 |
 niklaskorz | I think the module isn't meant to be used with stable versions anymore | 13:29:05 |
| niklaskorz | as mentioned on https://lix.systems/add-to-config/
Using the Lix NixOS module (only recommended for bleeding edge builds)
| 13:30:05 |
 goldstein | I found out that this actually isnāt enough
e.g. in builtins.fetchTree (builtins.parseFlakeRef "git+https://example.org/repo?dir=foo";) parseFlakeRef will include dir as an attr in a resolved flake ref, which fetchTree canāt deal with:
error: unsupported input attribute 'dir' for the 'git' scheme
| 19:17:26 |
| goldstein | I can mention it in the issue ig, although Iām not sure thatās the same issue | 19:17:41 |
| goldstein | cppnix has the same behaviour here | 19:18:06 |
| goldstein | so you need a āsanitizeā step in between that removes dir and maybe narHash | 19:18:55 |
| goldstein | * so you need a āsanitizeā step in between that removes dir ~and maybe narHash~ | 19:20:59 |
| goldstein | * so you need a āsanitizeā step in between that removes dir ~and maybe narHash~ | 19:21:05 |
| goldstein | * so you need a āsanitizeā step in between that removes dir <s>and maybe narHash</s> | 19:21:14 |
| goldstein | * so you need a āsanitizeā step in between that removes dir and maybe narHash | 19:21:35 |
| goldstein | aside: is there a good source of truth for what are the allowed attributes in a flakeref? nix.dev certainly misses some, and Iām not sure if itās intentional (i.e. some attrs are internal-only) or just a docs bug | 19:26:58 |
 helle (just a stray cat girl) | so erm | 19:41:17 |
| helle (just a stray cat girl) | that is a tad cursed | 19:41:20 |
| helle (just a stray cat girl) | the docs for it | 19:41:22 |
| helle (just a stray cat girl) | https://docs.lix.systems/manual/lix/2.94/language/builtins.html#builtins-fetchTree | 19:41:46 |
| helle (just a stray cat girl) | uh, yeah, mention how incomplete they are | 19:41:54 |
| helle (just a stray cat girl) | * uh, yeah, mentions how incomplete they are | 19:43:33 |
| helle (just a stray cat girl) | this technically is it's own bug independent of all the other doc bugs and should probably exist | 19:43:50 |
| goldstein | thanks for the link anyway! | 19:43:51 |
| goldstein | I think these docs are similar to nix.dev ones | 19:44:00 |
| helle (just a stray cat girl) | that part is still the same, not really been touched | 19:44:12 |
