Hi, Amy Pillow <[email protected]> skribis:
> I am confused why we still use and recommend the nix-base32 scheme for > source hashes and store paths. As far as I can tell, it's a completely > unique algorithm that has no related RFC or documentation. It works > fine it seems, it is a matter of professionalism. I believe the encoding > scheme should be fully described and published, with multiple > implementations, and there are of course dozens of alphabets and > algorithms to choose from that meet this criteria, some of which are > already implemented by Guix. nix-base32 has multiple implementations: (guix base32) and the one in C++ inherited from Nix. There are places where it will always be used internally (for store file names in particular) because it’s part of the protocol. And there are places where we could switch to something different, typically in package definitions. But honestly, I don’t find it crucial—certainly less than, say, switching to SHA512, SHA3, or similar for origin hashes. Ludo’.
