attrsets: clarify that mapAttrs maps over *leaf* attrs

[ghost]

Description of changes

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandbox = true set in nix.conf? (See Nix manual)
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 23.05 Release Notes (or backporting 22.11 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
  • (Release notes changes) Ran nixos/doc/manual/md-to-db.sh to update generated release notes
• Fits CONTRIBUTING.md.

[ghost]

@roberth, related to your suggestion, it appears that we do not have any kind of mapAttrs that applies the function to every node (i.e. no leaf/interior distinction). Is that correct? I went looking for one when I wrote #212186 but didn't find anything.

Is there a reason it was omitted? Everybody who needs it seems to write their own recursion.

[roberth]

I think that's correct. It may have to do with the lack of static or nominal types, which means that we can't consistently distinguish between an attrset that's part of a tree structure and an attrset that represents something else. We do have those predicate functions (Attrs -> Bool, List String -> Attrs -> Bool) for that, but they add a lot of noise. At that point you might as well write the algorithm yourself, with a more easily understandable end result without the indirection.

Everybody who needs it seems to write their own recursion.

In a language like Haskell, recursion schemes are kind of nice, but Nix is a different language, where different things are useful, and different things are idiomatic.

You could write a library that represents trees in a safe way, supporting a nice function like foldTree :: (AttrSet a -> a) -> Tree -> a which can hold any type a safely without a predicate function, but I think users wouldn't like the Tree representation.