[RFC] Move home-manager support to separate module repo?

[accelbread]

This is not a final decision, but proposing moving home-manger support to a separate flake like flakelight-rust, etc.

This would have the following advantages:

• It would be able to have a flake dependency on home-manager, and set a default value for inputs.home-manager
• The auto-update script could test with latest home-manager

Downsides:

• People are using current built in support.
• People would need to add an input, and add the module, likely adding 2 lines to their config.

In order to do this, the migration will have to be easy for users configuring home-manager with flakelight. Perhaps that could look like having an module flake that allows either using module or not to work, and adding a warning for deprecation period for those using homeConfigurations but not the module.

Also see https://github.com/nix-community/flakelight/pull/13 for related Darwin discussion.

[cmacrae]

Perhaps that could look like having an module flake that allows either using module or not to work, and adding a warning for deprecation period for those using homeConfigurations but not the module.

I like this idea, personally. I think it's a good idea to move home-manager support out into its own module and your idea for a deprecation period would be a pretty acceptable transition, I feel