| 25 Mar 2022 |
 @janne.hess:helsinki-systems.de | nvm, it just landed on master 👀 | 12:34:40 |
 bobvanderlinden | Awesome! That is very helpful indeed 👍👍 | 21:40:42 |
| 26 Mar 2022 |
|  Andrea Pascal joined the room. | 17:03:13 |
| 27 Mar 2022 |
 flokli | ElvishJerricco: can we get https://github.com/NixOS/nixpkgs/pull/164943 finished up and merged? ;-) | 11:28:17 |
| 28 Mar 2022 |
 Arian | Everything seems resolved no? | 12:49:32 |
| @janne.hess:helsinki-systems.de | I'll read over it later today and give it the checkmark of approval if I think it's done ;) Also the classic "if nobody has reasons I shouldn't merge this until …" ultimatum while I'm at it :P | 12:50:30 |
| Arian | Sounds good to me. Currently on my phone but when I have laptop again I'll also give it a green check | 12:51:02 |
| Arian | (can't find the approval button on the GitHub app lol) | 12:51:12 |
| @janne.hess:helsinki-systems.de | It's under "Reviews" | 12:51:23 |
| bobvanderlinden | btw, i worked on lvm and luks support a bit: https://github.com/NixOS/nixpkgs/compare/master...bobvanderlinden:pr-systemd-initrd-luks-lvm, not ready for a PR yet, but it looked promising with the experiments i've done in qemu | 13:26:00 |
 @elvishjerricco:matrix.org | bobvanderlinden: Awesome. Couple of notes: Why replace coreutils with busybox? IIRC it's not smaller. Also, for LUKS, you don't need to use the full non-minimal systemd. You can do systemdMinimal.override { withCryptsetup = true; } (just probably need to put that in all-packages.nix under some name to make sure it's cached by Hydra). | 18:33:25 |
| Arian | I'll work on Plymouth at ocean sprint I think | 18:36:05 |
| @elvishjerricco:matrix.org | (In fact, I just tried replacing coreutils with busybox, and it's a megabyte larger this way, and I had a collision on the reboot binary) | 18:36:54 |
| bobvanderlinden | In reply to @elvishjerricco:matrix.org
bobvanderlinden: Awesome. Couple of notes: Why replace coreutils with busybox? IIRC it's not smaller. Also, for LUKS, you don't need to use the full non-minimal systemd. You can do systemdMinimal.override { withCryptsetup = true; } (just probably need to put that in all-packages.nix under some name to make sure it's cached by Hydra). For my use I hated how bash worked in qemu in console mode. It felt quite buggy, as if readline was broken. But I placed it in a separate commit, it's optional. | 19:06:49 |
| bobvanderlinden | I haven't optimized on size yet. I first want to see it work, I can optimize once I understand all of the pieces that are needed | 19:07:35 |
| bobvanderlinden | I also tried to apply the changes from Janne Heß, but couldn't yet get the full boot to work | 19:18:19 |
| @janne.hess:helsinki-systems.de | In reply to @bobvanderlinden_:matrix.org
I also tried to apply the changes from Janne Heß, but couldn't yet get the full boot to work understandable, I had to pick like 20 commits from 40 branches :P I hope the situation gets better when my 2 PRs are in. Once the current "main" PR of this project is in, it should be just one more PR for the full handover | 19:19:10 |
| bobvanderlinden | Ah 2 prs? Which ones? | 19:20:09 |
| @janne.hess:helsinki-systems.de | well one isn't really necessary, it just makes the diff smaller for the "real" PR | 19:20:39 |
| @janne.hess:helsinki-systems.de | Important: https://github.com/NixOS/nixpkgs/pull/165453
Not so important: https://github.com/NixOS/nixpkgs/pull/165684 | 19:20:51 |
| 29 Mar 2022 |
| @janne.hess:helsinki-systems.de | ElvishJerricco: https://github.com/NixOS/nixpkgs/commit/e4fd33b8cc04f6cc052c7288b1179aa362271dae
not sure about whether startAt is also available in the initrd but that is what would fix my last issue
| 07:42:41 |
| @elvishjerricco:matrix.org | Janne Heß: Cool. I see no reason not to support startAt. It's just some convenient Nix for generating timer units, which should function in initrd just fine | 07:47:15 |
| @janne.hess:helsinki-systems.de | so it's available? okay, I'll pull it up | 07:47:34 |
| @elvishjerricco:matrix.org | Now, I think it's certainly going to be esoteric to need timer units in initrd, but there's no cost to allowing it :P | 07:47:46 |
| @elvishjerricco:matrix.org | I wouldn't be surprised if we find them useful for networking stuff maybe? Like timeouts and retries and whatnot | 07:48:18 |
| @janne.hess:helsinki-systems.de | Nono, I'm wondering whether the option does what I expect it to do when I set it for the initrd of if it's just unused | 07:48:35 |
| @elvishjerricco:matrix.org | Oh... good point. I forget where the code that translates startAt to timer units lives, and if we'll have to find a way to reuse that for initrd... | 07:49:27 |
| @janne.hess:helsinki-systems.de | or we don't and just add it when someone needs it? | 07:49:42 |
| @elvishjerricco:matrix.org | also fair | 07:49:50 |
| @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 |
