| 22 Feb 2026 |
 K900 | I've had to opt out like four packages | 16:02:04 |
 leo60228 | i have a very weird but consistently reproducible issue | 17:05:53 |
| leo60228 | when using plasma-login-manager plasmashell fails to connect to kwin window management and so task manager plasmoids don't work ("The PlasmaWindowManagement protocol hasn't activated in time. The client possibly got denied by kwin? Check kwin output." in plasma-plasmashell.service journal) | 17:06:46 |
| leo60228 | sddm is fine | 17:06:53 |
| leo60228 | happens with fresh plasmashell config files too | 17:07:03 |
| leo60228 | (using the wayland session for both) | 17:07:16 |
| K900 | Hmm I wonder if something is losing a race somehow | 17:24:06 |
| leo60228 | restarting plasmashell and/or kwin doesn't fix it so i feel like it must be something about the environment....? | 17:36:05 |
| leo60228 | there could be a race somewhere else i'm not thinking of though | 17:36:20 |
| K900 | And nothing in kwin logs either? | 17:40:42 |
| K900 | OK I need actual reviews on this | 18:09:44 |
| K900 | https://github.com/NixOS/nixpkgs/pull/318190 | 18:09:45 |
| leo60228 | nothing jumps out, here are logs and environment variables for kwin and plasmashell on each https://gist.github.com/leo60228/49ff2ecee8dd8e1adcdbe6247945d3f9 | 21:37:25 |
| leo60228 | oh wait | 21:43:39 |
| leo60228 | it's incompatible with the workaround from https://github.com/NixOS/nixpkgs/issues/126590 | 21:44:14 |
| leo60228 | that's a really bizarre way for that to manifest but fair enough i guess?? | 21:44:27 |
|  problems changed their profile picture. | 22:16:15 |
| 23 Feb 2026 |
| K900 | Oh my god please don't do that | 06:31:46 |
| K900 | That's SO bad | 06:31:52 |
| K900 | And the environment variables should be a lot better now | 06:31:58 |
| 26 Feb 2026 |
 Sandro 🐧 | still kinda feels slucky unfortunately ... | 15:46:24 |
| Sandro 🐧 | We could not just symlink join all directories from systempackages and add them in one variable in the front? | 15:46:41 |
| K900 | We really shouldn't | 16:20:07 |
| Sandro 🐧 | but would it work? 😥 | 16:55:06 |
| K900 | To some extent | 16:56:34 |
| K900 | It'll very much fuck things up for things running in nix-shells | 16:56:44 |
| Sandro 🐧 | only set it once per shell and not in a new subshell? | 17:14:37 |
| K900 | Not like that | 17:15:53 |
| K900 | They'll just start reading the merged files | 17:15:57 |
|  @hexa:lossy.network left the room. | 17:17:14 |
