| 16 Jan 2026 |
 Marie | I also tried LTS kernel, but it also crashes there | 11:47:24 |
 K900 | Try killsnoop from bcc | 11:47:37 |
| K900 | It should at least tell you what's killing it | 11:47:44 |
| Marie | TIME PID COMM SIG TPID RESULT
12:49:03 228374 kill 10 223946 0
12:49:03 228379 kill 10 223946 0
12:49:03 228384 kill 10 223946 0
| 11:50:05 |
| Marie | i'm pretty sure it's killing itself tbh | 11:50:21 |
| Marie | because these pids are dead after it crashes | 11:50:36 |
| K900 | That tracks, possibly | 11:50:56 |
| Marie | i think i found the reason | 12:04:29 |
| Marie | it didn't like fish as my login shell | 12:04:36 |
| Marie | changed it to bash and it started working | 12:04:50 |
| Marie | but why | 12:05:13 |
| K900 | The bigger question to me is why is Flatpak even leaking that | 12:13:00 |
| K900 | Or how for that matter | 12:13:06 |
| K900 | Given there's no fish in the container | 12:13:13 |
| K900 | (or at least there really shouldn't be) | 12:13:19 |
| Marie | switched back to fish from bash and now it's not broken anymore | 13:20:55 |
| Marie | i'm very confused | 13:20:59 |
| Marie | but at least it works now | 13:21:01 |
 Atemu | Well perhaps that's the issue? Because it's set as the user's shell? Though I would have expected that to simply be normalised within the container to prevent stuff like this. Don't know how flatpak generally handles users inside its NS | 16:51:23 |
| K900 | Yeah it should remap | 16:51:48 |
| K900 | I think | 16:51:49 |
