| 11 Jan 2022 |
 Daniel Phan on conduit | Oops, nevermind. https://github.com/divnix/devos/blob/main/profiles/core/default.nix#L149 wasn't true by default, so you had to clear the iptables rules. | 10:37:10 |
 @worldofgeese:one.ems.host | I can make a PR for the docs to specify that openFirewall must be set to true when creating an ISO if remote access is desired | 10:40:43 |
| Daniel Phan on conduit | Yes, that will do. Just set services.openssh.openFirewall = true; in hosts/bootstrap.nix and recreate the ISo | 10:41:58 |
| Daniel Phan on conduit | * Yes, that will do. Just set services.openssh.openFirewall = true; in hosts/bootstrap.nix and recreate the ISO | 10:42:00 |
| @worldofgeese:one.ems.host | Well, I almost succeeded in the initial install but was bitten by https://github.com/NixOS/nixpkgs/issues/73404.
The suggested solution is to `nixos-enter` and try again but that doesn't work because `nixos-install --flake .#host` expects `/mnt`
I have logs at https://gist.github.com/worldofgeese/76446406a2c970a8908167bad8482596 | 20:03:00 |
| Daniel Phan on conduit | What about nixos-rebuild boot .#host? | 20:06:28 |
| @worldofgeese:one.ems.host | From inside the chroot? | 20:14:09 |
| Daniel Phan on conduit | Yep | 20:14:13 |
| @worldofgeese:one.ems.host | Let me give it a shot | 20:14:28 |
| @worldofgeese:one.ems.host | Redacted or Malformed Event | 20:24:14 |
| @worldofgeese:one.ems.host | Well, maybe? I think all that's left is to try a reboot and see | 20:26:56 |
| @worldofgeese:one.ems.host | There's no output apart from warning that my git tree is dirty and some cachix saved settings | 20:27:28 |
| @worldofgeese:one.ems.host | I'm going to do a systemctl reboot and pray 🙏🏼 | 20:27:53 |
| @worldofgeese:one.ems.host | Hard stop at GRUB. 🤔 And I did have systemd-boot enabled. I'll give this another try tomorrow, thanks again Daniel | 20:34:42 |
| @worldofgeese:one.ems.host | Pushed my forked config to https://github.com/worldofgeese/devos | 20:53:11 |
| 13 Jan 2022 |
 @gytis-ivaskevicius:matrix.org | Whats exactly the reason why flakes are not expected to support nested package sets in the future? | 14:35:24 |
| @gytis-ivaskevicius:matrix.org | is it due to extra eval? | 14:35:42 |
 Pacman99 | I think the simplest reason is that it hurts the performance of nix search | 18:12:36 |
| Pacman99 | But looking at the decision with the end goal of flakes in mind, it makes sense that a flake should advertise a set of package around the topic of that flake. A nested package set indicates that the flake might need to be seperated. | 18:13:52 |
| Pacman99 | * But looking at the decision with the end goal of flakes in mind, it makes sense that a flake should advertise a set of packages around the topic of that flake. A nested package set indicates that the flake might need to be seperated. | 18:15:14 |
| Pacman99 | * But looking at the decision with the end goal of flakes in mind, it makes sense that a flake should advertise a set of packages around the topic of that flake. A nested package set indicates that a flake should be separated. | 18:15:25 |
| @gytis-ivaskevicius:matrix.org | Do yeah, about all this | 19:43:54 |
| @gytis-ivaskevicius:matrix.org | I had a thought | 19:43:57 |
| @gytis-ivaskevicius:matrix.org | What if stuff like 'nix shell nixpkgs#gnome3.nautilus' were to point to 'subflake' named 'gnome3', package 'nautilus' | 19:45:09 |
| @gytis-ivaskevicius:matrix.org | And su flakes should be able to reference each other | 19:45:41 |
| @gytis-ivaskevicius:matrix.org | This way you could not only solve this issue but also gain improvements in terms of parallelism/memory usage/caching | 19:46:43 |
| @gytis-ivaskevicius:matrix.org | Just a thought | 19:46:52 |
| @gytis-ivaskevicius:matrix.org | Also possibly nix evaluation would no longer take 30gb of ram 😅 | 19:47:21 |
| Pacman99 | Well I imagine that some of those subsystems could just get their own flake, like imagine if you could instead to github:nixos/gnome#nautilus | 21:26:16 |
| Pacman99 | I think in general the vision is to eventually start removing stuff from nixpkgs, so those subsystems get their own flake and github repo to work with. Which is basically circling back to the pro/con monorepo debate. | 21:27:30 |
