This is especially interesting to me I have long been interested on how we get more “major upstream developers” switching to Nix.
Maybe this is just how paradigm shifts go, but we historically have gotten more regular users and/or developers newish to running Linux (or just very young people in general) to make the jump, and a lot of the “old guard”, so to speak, keeps on using the traditional distros.
The developers of upstream software themselves use NixOS, the easier things will get. E.g. imagine if systemd was developed in the majority by NixOS users…
I’m not a systemd maintainer so no idea how I should even shape the question but basically “Is there any trouble with how things are at the moment?” What would improve in case systemd developers started using Nix [more]?
I am not trying to avoid looking for problems, but rather trying to go for good things.
For example
init : daemons that run arbitrarily long :: Nix daemon : batch jobs that terminate
is a fairly tight analogy. A lot of the sandboxing stuff that Nix does is playing catch up to systemd’s tricks. If systemd developers were all big Nix heads they might just…implement Nix build sandboxing as a first party thing? That would instantly get us better resource management stuff (cgroups, other systemd container features), and also make it really nice to monitor running builds and services side-by-side.