| 5 May 2022 |
 K900 | Nice | 06:30:19 |
 @janne.hess:helsinki-systems.de | In reply to @elvishjerricco:matrix.org
Aw, I wanted to try the plymouth PR but I don't seem to get a graphical boot screen with qemu-vm About that, we should see if the closure size increases when enabling logind without adding more files | 08:43:58 |
| @janne.hess:helsinki-systems.de | I don't find that this would a lot of sense. Compiling in new features, making the initrd larger, just for 2 udev rules | 08:44:31 |
 @arianvp:matrix.org | How about patching the meson.build to just always include those udev rules? | 10:04:54 |
| @arianvp:matrix.org | Maybe it's worth discussing with upstream | 10:05:02 |
|  Ash joined the room. | 12:42:19 |
 @elvishjerricco:matrix.org | In reply to @janne.hess:helsinki-systems.de
I don't find that this would a lot of sense. Compiling in new features, making the initrd larger, just for 2 udev rules Well when I tried it, the initrd didn't get noticeably larger. But then again it might make the size of the systemdStage1 derivation output bigger, which consumes more store space, even if the initrd isn't bigger. | 13:19:02 |
| @janne.hess:helsinki-systems.de | I was also wondering if we could get rid of pkcs11 and merge the networkd support into the regular stage 1 init | 13:21:12 |
| @janne.hess:helsinki-systems.de | But size measurements have to be taken and I currently don't have time for it | 13:21:25 |
| @elvishjerricco:matrix.org | IIRC, my networkd PR did have a much bigger initrd. But I might be remembering some odd experiments with that... | 13:22:10 |
| @janne.hess:helsinki-systems.de | I only talk about building it into the closure, not copying it by default ;) | 13:23:16 |
| @elvishjerricco:matrix.org | Oh fair. | 13:23:58 |
| @elvishjerricco:matrix.org | On that front | 13:24:04 |
| @elvishjerricco:matrix.org | I've also tested just straight up using the stage 2 systemd as the package | 13:24:20 |
| @elvishjerricco:matrix.org | Since we're pretty good about only copying in the necessities, it only got 3M bigger | 13:24:35 |
| @janne.hess:helsinki-systems.de | that's still a lot imo | 13:24:50 |
| @elvishjerricco:matrix.org | Honestly to just keep the number of derivations in the world down, we might consider just eating that cost and doing that | 13:24:53 |
| @elvishjerricco:matrix.org | Yea it's a tough tradeoff | 13:25:01 |
| @elvishjerricco:matrix.org | But maybe we could even figure out why it's so much bigger and cut it back down | 13:25:18 |
| @janne.hess:helsinki-systems.de | I don't find it as tough. building this small systemd is very fast and is probably unnoticable on hydra | 13:25:29 |
| @elvishjerricco:matrix.org | In reply to @janne.hess:helsinki-systems.de
I don't find it as tough. building this small systemd is very fast and is probably unnoticable on hydra Well the issue is more about the amount of store space consumed on users' machines. But I guess it's only like 12M | 13:29:35 |
| @elvishjerricco:matrix.org | In reply to @janne.hess:helsinki-systems.de
I don't find it as tough. building this small systemd is very fast and is probably unnoticable on hydra * Well the issue is more about the amount of store space consumed on users' machines. But I guess it's only like 12M for the systemdStage1 package | 13:29:47 |
 lovesegfault | I'd be happy with merging the plymouth PR as-is, and adding tasks to the project to:
- Evaluate how to include the udev rules in the
systemdStage1 package
- Evaluate whether having a
systemdStage1 is worth it overall
- Adding tests
In anyone opposed?
| 21:35:27 |
| @elvishjerricco:matrix.org | I don't particularly care. It's a small enough hack that we can always come back and change it later trivially. As long as Janne Heß thinks the PR is otherwise ready, I'm good | 21:36:39 |
| @janne.hess:helsinki-systems.de | which one? | 21:36:54 |
| lovesegfault | https://github.com/NixOS/nixpkgs/pull/170136 | 21:37:01 |
| @janne.hess:helsinki-systems.de | yeah I agree with you both | 21:41:01 |
| lovesegfault | Sweet, let me add stuff to the board | 21:41:17 |
| lovesegfault | Done: https://github.com/NixOS/nixpkgs/pull/170136#issuecomment-1119074275 | 21:44:23 |
| lovesegfault | I'm going to review that PR one more time before hitting el button | 21:44:33 |
