lib/types: init {types.attrsWith}

[hsjobeki]

types.attrsWith combines attrsOf and lazyAttrsOf with a configurable lazy = true/false parameter.

Other parameters are possible and on hold see: #344216

Things done

• 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/)
• 24.11 Release Notes (or backporting 23.11 and 24.05 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.

[infinisil]

@hsjobeki Btw the "Co-Authored-By: @hsjobeki" doesn't appear to work :P

[hsjobeki]

@infinisil Depraction message works now. With/without type merging.

[infinisil]

Looked at this together in a meeting with @hsjobeki, looking good, let's deal with any issues if they come up!

[paparodeo]

this change seems like it broke one of the ofborg eval tests.

nix-build --arg pkgs 'import ./. {}' ./lib/tests/release.nix

https://gist.github.com/GrahamcOfBorg/7d1ce79b712e477612915ee9247188c8

[K900]

So do we revert this or does anyone have a quick fix?

[K900]

Reverted.

[Mic92]

We will add lib-tests soon to github ci.

[hsjobeki]

@infinisil @K900 Please just remove the failing test line. It was only there for testing the "nice" deprecation message. But it seems the test framework cannot handle that yet.

I would need to first fix the test framework. Which is a bit more work.

I am not sure what the convention is for reverting a revert but i opened up this one:

#361391

[Mic92]

The convention is to revert the revert and than potentially fix the issue.

[nixos-discourse]

This pull request has been mentioned on NixOS Discourse. There might be relevant details there:

https://discourse.nixos.org/t/use-cases-of-option-type-internals/57317/1