| 29 Mar 2022 |
 @elvishjerricco:matrix.org | though in the same vein, the script style options don't work right now either, though I certainly intend for them to be fixed | 07:50:34 |
 @janne.hess:helsinki-systems.de | okay so I'll remove these as well and we can re-add them when they work? | 07:51:22 |
| @elvishjerricco:matrix.org | Sure | 07:51:59 |
| @janne.hess:helsinki-systems.de | They should work though | 08:41:41 |
| @janne.hess:helsinki-systems.de | you call mkServiceConfig which evaluates the options | 08:41:54 |
| @janne.hess:helsinki-systems.de | yeah this works:
boot.initrd.systemd.services.test = {
wantedBy = [ "initrd.target" ];
script = ''
echo lal
'';
};
| 08:51:54 |
| @janne.hess:helsinki-systems.de | Updated commit: https://github.com/helsinki-systems/nixpkgs/commit/62d4d08e9fd7f0a0d32349ef64731e31a841cb9c | 09:16:10 |
| @elvishjerricco:matrix.org | In reply to @janne.hess:helsinki-systems.de
yeah this works:
boot.initrd.systemd.services.test = {
wantedBy = [ "initrd.target" ];
script = ''
echo lal
'';
};
wait how does that work... I never wrote any code that actually causes the generated script to be added to the initrd... It should produce unit file that fails to actually start because the script doesn't exist. | 17:54:27 |
| @janne.hess:helsinki-systems.de | hmm maybe systemd just said "OK" and it wasn't okay at all? | 17:55:04 |
| @elvishjerricco:matrix.org | In reply to @janne.hess:helsinki-systems.de
Updated commit: https://github.com/helsinki-systems/nixpkgs/commit/62d4d08e9fd7f0a0d32349ef64731e31a841cb9c And aww I had really hoped to avoid having a whole second set of things specially for initrd, but oh well. If that's the best way to accomplish this then so be it | 17:56:41 |
| @elvishjerricco:matrix.org | In reply to @janne.hess:helsinki-systems.de
hmm maybe systemd just said "OK" and it wasn't okay at all? I'm guessing the unit failed to start, but systemd continued with the boot anyway. It'll just go as long as switch-root can succeed and nothing does anything like systemctl emergency or whatever | 17:57:59 |
| @elvishjerricco:matrix.org | I thought about setting /etc/systemd/system/service.conf so that any unit that fails causes it to enter emergency mode, but I figure if the boot can proceed with a failed initrd unit, then maybe we should just let it? | 17:59:04 |
| @janne.hess:helsinki-systems.de | In reply to @elvishjerricco:matrix.org
I thought about setting /etc/systemd/system/service.conf so that any unit that fails causes it to enter emergency mode, but I figure if the boot can proceed with a failed initrd unit, then maybe we should just let it? Yeah, better have an incomplete boot than no boot at all | 18:00:07 |
| @janne.hess:helsinki-systems.de | machine # [ OK ] Started test.service.
machine # initrd.target: AssertPathExists=/etc/initrd-release succeeded.
machine # test.service: Failed to locate executable /nix/store/4lv9l9rgrsy8q5l29wycmdyx37wx0h3p-unit-script-test-start/bin/test-start: No such file or directory
machine # test.service: Failed at step EXEC spawning /nix/store/4lv9l9rgrsy8q5l29wycmdyx37wx0h3p-unit-script-test-start/bin/test-start: No such file or directory
| 18:03:09 |
| @janne.hess:helsinki-systems.de | you are right :( | 18:03:12 |
| @janne.hess:helsinki-systems.de | is that important enough that we have to include it in the MVP? | 18:03:34 |
| @elvishjerricco:matrix.org | In reply to @janne.hess:helsinki-systems.de
is that important enough that we have to include it in the MVP? No | 18:07:29 |
| @elvishjerricco:matrix.org | It's just very nice to have :P | 18:07:42 |
| @janne.hess:helsinki-systems.de | okay so I'll exclude these options as well | 18:07:45 |
| @elvishjerricco:matrix.org | sure | 18:07:52 |
| @janne.hess:helsinki-systems.de | Here you go: https://github.com/helsinki-systems/nixpkgs/commit/0ba59f624e67d4ff44fa4ea63fe89cfd095e9ac4 | 18:10:35 |
| @elvishjerricco:matrix.org | Janne Heß: Looks like roberth has some suggestions on the PR. He doesn't like the mkServiceConfig thing; he found a cleaner way to do it with nixos modules. So I think he might have similar gripes with your forInitrd stuff... | 18:17:10 |
| @elvishjerricco:matrix.org | * Janne Heß: Looks like roberth has some suggestions on the PR. He doesn't like the mkServiceConfig thing; he found a cleaner way to do it with module composition instead of functions that generate modules. So I think he might have similar gripes with your forInitrd stuff... | 18:18:40 |
| @janne.hess:helsinki-systems.de | You go first, I'll fix my code afterwards ;) | 18:22:37 |
| 30 Mar 2022 |
| @janne.hess:helsinki-systems.de | ElvishJerricco: Ping? :) Do you have time for the rebase somewhat soonish? | 18:39:28 |
| 1 Apr 2022 |
| @janne.hess:helsinki-systems.de | I pushed https://github.com/NixOS/nixpkgs/pull/164943/commits/5653209ed95985034bbb43bb63e04190639477a0 to make roberth happy and to have a more extensible way of specifying files | 11:28:14 |
| @janne.hess:helsinki-systems.de | I'm sorry about that, bobvanderlinden. Your lvm attempts will now need to be refactored to the new options :/ (don't do it until we are sure we don't have to change it again!) | 11:28:53 |
| @elvishjerricco:matrix.org | Janne Heß: Sorry I haven't been able to get to systemd-stage1 stuff this week. But I just looked at your new set of changes and I really like them! Thanks! | 17:30:17 |
| @janne.hess:helsinki-systems.de | In reply to @elvishjerricco:matrix.org
Janne Heß: Sorry I haven't been able to get to systemd-stage1 stuff this week. But I just looked at your new set of changes and I really like them! Thanks! Great, pleased to hear that. I hope the biggest reviews are over now | 17:31:09 |
| @elvishjerricco:matrix.org | I really like the environment.etc-like contents option. It's a lot cleaner. | 17:31:46 |
