| 25 Dec 2025 |
 llakala | In reply to @emilazy:matrix.org
I think this is probably LLM slop but as someone who was on the call with GitHub I would not say this is at all a reasonable takeaway ty very much for your nixpkgs core team work btw. you guys are the thankless heroes helping shit actually get better | 19:50:07 |
 leona | (please use channels.nixos.org, not the redirect) | 19:53:13 |
| llakala | I'm a lot happier with my tarball times these days with npins and lix. the cppnix tarball downloads were dark days | 19:54:49 |
| llakala | although a nixpkgs `git pull` is always painful. I wish git did more logging rather than looking like it freezes | 19:55:41 |
 tea | Itd be interesting to explore a scheme where we download definitions for individual files instead of the whole tarball. Not sure how it could be made to actually scale | 22:05:49 |
| tea | the 50ish MB size of nixpkgs is not that big but just big enough to be annoying on a slow connection | 22:06:23 |
 Linux Hackerman | https://github.com/Mic92/fast-flake-update is relevant I guess | 23:15:01 |
| 26 Dec 2025 |
 Arian | https://github.com/NixOS/nix/issues/14715 | 08:33:28 |
 0x4fbb09 it/its ⛯✇ΘΔ | differential updates would be smart here, though it'd depend on how often you update. should be practical to generate diff files between the last week of nixos-unstable and latest | 08:34:04 |
 Emma [it/its] | Git...? | 08:42:56 |
| 0x4fbb09 it/its ⛯✇ΘΔ | how much does git like shallow checkouts (and shallow fetches of a shallow repo?)
iirc there are/were performance problems there
(a static file system would also be easier to host) | 09:19:17 |
| Emma [it/its] | git doesnt need any serverside logic beyond a... plain static http server | 09:21:05 |
| tea | true, that's a good idea | 09:25:28 |
| 0x4fbb09 it/its ⛯✇ΘΔ | need, no, but idk how much slower the dumb transport is | 09:27:08 |
| Emma [it/its] | its reasonably fast | 09:27:41 |
|  isabel changed their profile picture. | 11:37:33 |
|  zitrone 🍋 changed their display name from zitrone to zitrone (39c3 DECT ZITR/9487). | 12:19:27 |
|  Jean 💕 joined the room. | 12:33:54 |
|  eta changed their profile picture. | 14:22:30 |
|  hrmny | 6069 changed their display name from hrmny to hrmny | 6069. | 15:32:03 |
|  jasper changed their display name from jappie to jappie @ 39c3. | 15:49:41 |
|  aktaboot changed their profile picture. | 19:25:49 |
| 27 Dec 2025 |
|  mio joined the room. | 04:17:27 |
| mio | any option to disable cgroup on unsupported environment? I am getting error: expected cgroup directory on nix on droid | 04:18:10 |
 delroth | not from what I can tell - this seems like it must be crashing while trying to detect if cgroups are available, it's a bug | 06:11:45 |
| delroth | cc raitobezarius detectAvailableCgroupFeatures can throw if getLocalHierarchy fails | 06:13:56 |
| delroth | mio: would recommend filing a bug on git.lix.systems with the exact error message, the contents of /proc/mounts on that system, the contents of /proc/self/cgroup in the shell where you're running Lix (I'm guessing there is no lix-daemon involved in this setup?), possibly other relevant info | 06:16:03 |
 aloisw | Why does is even run cgroup things if the experimental feature and setting are not enabled? | 06:16:45 |
| delroth | because the feature detection is not conditional on the experimental feature | 06:17:01 |
| delroth | which arguably is also a separate bug | 06:17:16 |
