Nix team and nixpkgs relationship

[zimbatm]

Nix team and nixpkgs relationship

[zimbatm]

Ok. Let me know if you need anything from the infra team.

[thufschmitt]

Fully agree with the issue :)

Not sure about a good solution though.
I don't think having the Nix team as CODEOWNER would help as we do releases when we think they are ready, and we probably wouldn't catch more regression just by doing the Nixpkgs update ourselves.

Something about some better hydra checks would help a lot indeed (related: #9741, #9744, #9743). It wouldn't be perfect since Hydra isn't meant to change the Nix version used, so the only tests that might exert the new Nix will be the tests that run Nix in a VM. But still much better than what we have right now.

[Ericson2314]

@roberth has done a lot of deep diving into this recently.

[thufschmitt]

Just discussed that during the Nix maintainers meeting today:

• @roberth willing to spend some time to improve the situation from the CI side
• Suggestion of having the Nix maintainers do the Nix releases on Nixpkgs:
  • Currently some confusion about the responsibilities of people
  • Maybe we should be more involved?
  • Having a good CI would diminish the need for that a lot
  • To reconsider if improving the CI isn't enough

[zimbatm]

Sweet! By the way, GitHub recently doubled the number of resources on the runners, and also now supports nested virtualization.

[nixos-discourse]

This issue has been mentioned on NixOS Discourse. There might be relevant details there:

https://discourse.nixos.org/t/2024-01-19-nix-team-meeting-minutes-116/38837/1