if the build is failing due to oom.
you can try --jobs num and --cores num flags for nixos-rebuild or nix-cli command, to limit the cpu usage.
also try out zramSwap or a swap file or partition if not enough memory. (also checkout swapspace if going to use a dedicated swap partition)
There will be pre-built cached packages later in the development cycle, notice the Hydra jobset: TBD bullet point, so that’s gonna make things much easier.
Thanks for the hints. I don’t mind compiling - There is enough RAM and disk space available. My problem is that the VM has newest unstable packages but the PR is behind master (does that really matter?):
This means basically everything needs to be compiled. I tried to remove as many unnecessary packages as possible (who needs a browser or a terminal anyways :D) but it is still a lot.
So all I can do is wait for Hydra providing pre-compiled packages?