| 3 Aug 2022 |
 K900 | https://github.com/NixOS/nixpkgs/actions/runs/2791100158 | 16:59:42 |
| K900 | Huh | 16:59:43 |
| K900 | New Actions failure mode | 16:59:56 |
| K900 | Fun | 16:59:58 |
 Winter (she/her) | it's fine | 17:00:08 |
| Winter (she/her) | it passed before | 17:00:12 |
 @elvishjerricco:matrix.org | I'm guessing that's just because it was closed so quickly? | 17:00:13 |
| Winter (she/her) | who knows honestly | 17:00:20 |
| K900 | I know it is, it's the exact same commit | 17:00:23 |
| K900 | And yeah probably | 17:00:25 |
| K900 | But also lol | 17:00:27 |
| @elvishjerricco:matrix.org | that was a bad 5 minutes for github lol | 17:00:41 |
| Winter (she/her) | as if every 5 minutes isn't /s | 17:01:11 |
| @elvishjerricco:matrix.org | lol | 17:01:17 |
| @elvishjerricco:matrix.org | ok, back to focusing on networkd / goofing off... | 17:01:35 |
| @elvishjerricco:matrix.org | Uh, I'm confused by the check error the networkd PR currently has. It's complaining that I have type = options.environment.etc.type, but... that's not what the code says. That's the old commit. The new one doesn't do that. | 17:23:10 |
| @elvishjerricco:matrix.org | Oh, there's a merge conflict because of the goblin thing and the fact that I very slightly modified the makeInitrdNG nix function | 17:25:35 |
| 4 Aug 2022 |
| @elvishjerricco:matrix.org | Created a solution to our concerns about make-initrd-ng gaining a mind of its own and causing destructive problems during shutdown: https://github.com/NixOS/nixpkgs/pull/185085 | 01:02:23 |
| Winter (she/her) | https://github.com/NixOS/nixpkgs/pull/185089 | 02:35:13 |
| Winter (she/her) | saw this was opened ^^ | 02:35:26 |
| @elvishjerricco:matrix.org | Oh, wow yea that's probably better than the scripted initrd's way of handling it: libc::ioctl | 02:37:50 |
| @elvishjerricco:matrix.org | * Oh, wow yea that's probably better than the scripted initrd's way of handling it: https://github.com/NixOS/nixpkgs/blob/master/nixos/modules/system/boot/stage-1-init.sh#L512-L532 | 02:37:55 |
| Winter (she/her) | In reply to @elvishjerricco:matrix.org
Created a solution to our concerns about make-initrd-ng gaining a mind of its own and causing destructive problems during shutdown: https://github.com/NixOS/nixpkgs/pull/185085 Is that /run path hardcoded in systemd or something? How does that get configured to run? I don't see anything that mentions that path in Nixpkgs other than the generation. | 02:39:22 |
| @elvishjerricco:matrix.org | In reply to @winterqt:nixos.dev
Is that /run path hardcoded in systemd or something? How does that get configured to run? I don't see anything that mentions that path in Nixpkgs other than the generation. It's hardcoded in systemd IIUC | 02:39:42 |
| @elvishjerricco:matrix.org | Though, looking at systemd/src/shutdown/shutdown.c, I'm not entirely sure the mount is needed... I dunno | 02:42:11 |
| @elvishjerricco:matrix.org | * Though, looking at systemd/src/shutdown/shutdown.c, I'm not entirely sure the mount is needed, as long as the dir exists... I dunno | 02:42:21 |
| Winter (she/her) | Why is the shutdown ramfs even a thing? Honestly the first time I've ever heard of it. | 02:53:35 |
| @elvishjerricco:matrix.org | Winter (she/her): Because without it you can't properly unmount file systems like /. | 02:54:31 |
| @elvishjerricco:matrix.org | And there are some configurations where that can legitimately matter. Also, it's just a good place to do things like zpool sync, where you know that all IO is done, and all that's left is to ensure it's sync'd | 02:55:38 |
| @elvishjerricco:matrix.org | Another example is closing device mapper devices. IIRC there can be issues with stuff like mdraid on rare occasion if your root FS is stored on it | 02:56:37 |
