Post #1234616

@sj good question! The start image is a binary, so in theory, someone could have 'poisoned' it to mess with the reproduction and hide their backdoor. By taking an old image as starting point, that attacker would have to have had the foresight to 'poison' the image back in 2020, made sure this remained undetected until now, and done it in a way that is still effective today. That is much harder / less likely than the same attack in a recent image. Of course even better would be to have some kind of 'bootstrap' from (almost) only source to having nix available in the vm to perform the build. I have not followed https://bootstrappable.org/ enough to know how close we are to that, though. There is https://github.com/oriansj/stage0-posix which I understand is used in https://guix.gnu.org/en/blog/2023/the-full-source-bootstrap-building-from-source-all-the-way-down/ but I have not yet wrapped my head around how this fits together in practice /cc @civodul @janneke