| 8 Jun 2026 |
 K900 | They run a full NixOS | 12:48:31 |
 Primrose (== Léana) | Is that faster or slower | 12:48:38 |
| K900 | Just no kernel | 12:48:41 |
 hexa | much fastrer | 12:48:45 |
| K900 | Faster than VM tests | 12:48:44 |
| hexa | Redacted or Malformed Event | 12:48:46 |
| hexa | Redacted or Malformed Event | 12:48:47 |
| hexa | Redacted or Malformed Event | 12:48:51 |
| hexa | https://github.com/NixOS/nixpkgs/pull/478109 | 12:49:20 |
| Primrose (== Léana) | is cgroups support also available on lix? | 12:49:51 |
| hexa | yes | 12:50:00 |
| Primrose (== Léana) | awesome | 12:50:04 |
| hexa | nix.settings.auto-allocate-uids = true;
nix.settings.experimental-features = ["auto-allocate-uids" "cgroups"];
nix.settings.extra-system-features = ["uid-range"];
| 12:50:14 |
| hexa | don't pass /dev/net into the sandbox, that's not required | 12:50:32 |
| hexa | I don't think it is exposed in pkgs.testers | 12:53:12 |
| hexa | nixos-lib = import (release.nixpkgs + "/nixos/lib") {
inherit (pkgs) lib;
};
nixos-lib.runTest {
hostPkgs = pkgs;
imports = [ ./test.nix ];
};
| 12:55:17 |
| Primrose (== Léana) | release.nixpkgs is some pinned nixpkgs? | 12:55:43 |
| hexa | and then define them as you woudl in nixos/test | 12:55:48 |
| hexa | oh, yeah | 12:55:51 |
| Primrose (== Léana) | Cool, will do
Though I'll spend the rest of my day enjoy zurihac, I'll probably try it out tomrrow on the train | 12:56:30 |
| Primrose (== Léana) | thanks a lot!! | 12:56:32 |
| hexa | sure, np | 12:56:40 |
 maralorn | That'd be awesome. | 15:15:30 |
| maralorn | All the while I am trying to merge your and Vasily's changes into one coherent fix for the lix build detection situation. | 15:16:35 |
| maralorn | I think I got it. | 20:24:48 |
| maralorn | But now I have tests failing for unrelated reasons. Primrose (== Léana), I am very confused the "golden-test" comment that we have been using to make sure the correct derivations for the files are in store have never included the failing example.^^ Now when I changed a bit of logic suddenly those tests are failing.^^ | 20:26:35 |
| maralorn | I believe that the validations for the failing test until now never went wrong when the associated derivations where missing from the store. | 20:34:39 |
| maralorn | Kinda funny. That oversight happened like 4 years ago.^^ | 20:34:52 |
| maralorn | Okay, everyone. Build finished detection should now work again on lix. Please try it out. | 22:10:47 |
| maralorn | * Okay, everyone. Build finished detection should now work again on lix. On nom main. Please try it out. | 22:11:03 |
