nixos: add key to shared module to allow disabling it

[Enzime]

Description

By adding key, this allows users to disable this shared module or they can choose to not disable this shared module (by filtering by key before disabling)

This means users can disable all shared modules if all modules are paths or attrsets with a key:

configuration.nix:

{ config, ... }:

{
  home-manager.users.enzime = { ... }: {
    disabledModules = config.home-manager.sharedModules;
  };
}

Or disabling just this module specifically:

{ ... }:

{
  home-manager.users.enzime = { ... }: {
    disabledModules = [ { key = "home-manager#nixos-shared-module"; } ];
  };
}

Or disabling all modules when you have modules you can't disable (like lambdas):

{ ... }:

{
  home-manager.users.enzime = { ... }: {
    disabledModules = lib.filter (v: lib.isString v || lib.isPath v || (lib.isAttrs v && v ? key)) config.home-manager.sharedModules;
  };
}

https://nixos.org/manual/nixos/unstable/#sec-replace-modules

Checklist

• Change is backwards compatible.

• Code formatted with ./format.

• Code tested through nix-shell --pure tests -A run.all or nix develop --ignore-environment .#all using Flakes.

• Test cases updated/added. See example.

• Commit messages are formatted like

  {component}: {description}
  
  {long description}
  

  See CONTRIBUTING for more information and recent commit messages for examples.

• If this PR adds a new module

  • Added myself as module maintainer. See example.

Maintainer CC

cc @rycee

[roberth]

The key is good, but I don't understand why sharedModules is a list, and it probably doesn't need to be.
List indexes aren't good identifiers, so I would recommend to change it to coercedTo (listOf raw) (ms: { imports = ms; }) deferredModule.
Useful, stable identifiers for this module / these modules can be provided in the form of flake output attributes, whose values (modules) each have a unique key.

[roberth]

Good first step, and further improvements possible.

[Enzime]

Changing sharedModules is a lot bigger of a refactor than the currently proposed changes, I think it would be best to do that in a different PR

[Enzime]

cc @teto :)

[teto]

I dont know this aspect of the module system but the approval of roberth is good enough for me