| 28 Dec 2025 |
 delroth | have you looked at the code that's presumably throwing? it can't run if no cgroupv2 fs is found, so clearly there should be a cgroupv2 fs? | 10:32:40 |
 raitobezarius | not yet | 11:25:02 |
| raitobezarius | ok, read the code, got it | 11:27:22 |
| raitobezarius | there's a cgroup v2 fs, but there's no cgroup associated to the process running | 11:27:34 |
| delroth | or selinux prevents the /proc read | 11:59:44 |
| delroth | * or selinux prevents the /proc/self/cgroup read | 11:59:55 |
|  Grimmauld (any/all) joined the room. | 12:04:56 |
| Grimmauld (any/all) | Todays horror: If your /etc/gitconfig is malformed/invalid, flakes refuse to build and git fetch builtins also break. Even if the config is valid, it is taken into consideration. Including aliases, making git fetchers wildly impure..... | 12:06:36 |
| raitobezarius | In reply to @delroth:delroth.net
or selinux prevents the /proc/self/cgroup read Good point | 12:36:01 |
| raitobezarius | In reply to @grimmauld:m.grimmauld.de
Todays horror: If your /etc/gitconfig is malformed/invalid, flakes refuse to build and git fetch builtins also break. Even if the config is valid, it is taken into consideration. Including aliases, making git fetchers wildly impure..... Except that secrets are inherently impure alas :p | 12:36:30 |
 aloisw | Honestly I think that's kinda not so bad? Like you still need the valid commit or NAR hash. | 14:48:14 |
| aloisw | Of course, in combination with smudge filters things start becoming weird, but that kinda requires explicit collaboration between the config and the repo. | 14:49:00 |
