Things done

Remove with lib in meta, proactively, in the whole repo, once and for all

Methodology

Basically, all packages with a with lib in meta will have meta = with lib; in it,
so I just replaced treewide this string to meta = in all **/*.nix.

This naive alteration would break a lot of stuff, like everything, so I had to fix the
broken stuff.

For this there is nil which is a code checker and also
implements the language server protocol. Then I did a simple bash for that iterates over
**/*.nix and applies the LSP to it discarding the output. Files with failure then are
listed and saved in a nil-failures.log. Why log? Log is in the gitignore so no chance
of causing a mess. At least from this part.

To start fixing what is basically thousands and thousands of files, I used substituteInPlace
with some common patterns, not overthinking too much to catch them all. Examples:

• license = licenses → license = lib.licenses
• maintainers = with maintainers → maintainers = with lib.maintainers

And many others, basically following this same pattern of simple text substitution, no regex.

Between the substitutions, I ran that nil loop again to recalculate the failing files and then
decided the next replacement.

After about a dozen iterations, there remained about 800 files to be fixed. The following files
were not changed because of a false positive where __curPos is undefined, but it seemed like an
implementation detail:

• lib/tests/modules/merge-module-with-key.nix
• lib/tests/modules/test-mergeAttrDefinitionsWithPrio.nix
• lib/tests/modules/types-attrTag.nix
• nixos/modules/profiles/nix-builder-vm.nix
• pkgs/applications/emulators/wine/sources.nix
• pkgs/development/compilers/corretto/mk-corretto.nix
• pkgs/top-level/release-cross.nix

By consequence, this PR also fixes issues where version is not defined in meta.changelog, which
in some cases I had to rewrite some mkDerivation using the finalAttrs form. There were rare cases,
about 10, probably.

Also by consequence, nil warned about unused recs and unused arguments, which I removed. Also, there were
unused let variables, which I didn't touch. A remarkable case was a arch something in some ROCm files.

There were some rare cases where passthru were defined after meta, then I moved it up.

TL;DR: this shouldn't have any rebuilds or eval issues.

Review tips

Download the full PR patch and look through the diff from there. Most of the time the diff will be boringly the same.

I'll squash the commits when this is about to get merged

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandboxing enabled in nix.conf? (See Nix manual)
  • sandbox = relaxed
  • sandbox = true
• 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/)
• 25.05 Release Notes (or backporting 24.11 and 25.05 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
• Fits CONTRIBUTING.md.

Add a 👍 reaction to pull requests you find important.