I'm just a bystander, but: Awesome!!! Thank you so much for this work.

For anyone following along, I believe that this can be tested from the comfort of your couch by:

nix-build -I nixpkgs=https://github.com/Gskartwii/nixpkgs/archive/bootstrap-x86_64-linux.tar.gz "<nixpkgs>" -A make-minimal-bootstrap-sources
nix-build -I nixpkgs=https://github.com/Gskartwii/nixpkgs/archive/bootstrap-x86_64-linux.tar.gz "<nixpkgs>" -A minimal-bootstrap.test

I am doing this now (and expect no problems). There was one quickly fixed issue in the beginning, I got a hash mismatch for nyacc, probably because I had already tested an earlier version? (I resolved this by appending a do-nothing question mark to the nyacc URL, prompting a redownload, for which the hash successfully validated. But I don't fully understand the root cause of this issue. On a machine with a fresh cache, this issue did not arise.)

How hard would it be to drop pkgs/stdenv/linux/bootstrap-files (for x86_64)?

I'm just a bystander, but: Awesome!!! Thank you so much for this work.

For anyone following along, I believe that this can be tested from the comfort of your couch by:

nix-build -I nixpkgs=https://github.com/Gskartwii/nixpkgs/archive/bootstrap-x86_64-linux.tar.gz "<nixpkgs>" -A make-minimal-bootstrap-sources
nix-build -I nixpkgs=https://github.com/Gskartwii/nixpkgs/archive/bootstrap-x86_64-linux.tar.gz "<nixpkgs>" -A minimal-bootstrap.test

I am doing this now (and expect no problems). There was one quickly fixed issue in the beginning, I got a hash mismatch for nyacc, probably because I had already tested an earlier version? (I resolved this by appending a do-nothing question mark to the nyacc URL, prompting a redownload, for which the hash successfully validated. But I don't fully understand the root cause of this issue. On a machine with a fresh cache, this issue did not arise.)

How hard would it be to drop pkgs/stdenv/linux/bootstrap-files (for x86_64)?

Edit. The test successfully concluded as expected.

Thank you, I appreciate the testing. Credits to to dish and GNU Mes folks among others. I do want to check if I can reproduce the hash mismatch locally. I've seen issues like that, where they only appear after a gc cycle.

Very exciting :D