docs: add more depth to Flake input

At present, if a user copy-pastes the current documentation, a version
of Nixpkgs will be added to their ``flake.lock`` lockfile. This is
probably not what the user wants as Nixpkgs, the tarball, is quite large
& the user likely expects the StevenBlack Flake to be used as a module
to just be using the Nixpkgs of the system.

This change should also help new Nix users know how to pin a Nix version
if there ever comes a time where the Nix :ac:`API` changes—even if ``lib``
tries to be stable & backwards-compatible.

Additionally, provide a comment about where & what the input can be—such
as a mirror or a local fork for development.

Format: text/x-rst

diff --git a/readme_template.md b/readme_template.md
index b9e134fce1afb1af476925bf5ceecdf09a391fc0..af3df41401905086cf96804faa847b73dee020fa 100644 (file)
--- a/readme_template.md
+++ b/readme_template.md
@@ -386,7 +386,14 @@ like this:
 
 ```nix
 {
-  inputs.hosts.url = "github:StevenBlack/hosts";
+  inputs = {
+    nixpkgs.url = "github:NixOS/nixpkgs?ref=$YOUR-REF";
+    hosts = {
+      url = "github:StevenBlack/hosts"; # or a fork/mirror
+      inputs.nixpkgs.follows = "nixpkgs";
+    };
+  };
+
   outputs = { self, nixpkgs, hosts }: {
     nixosConfigurations.my-hostname = {
       system = "<architecture>";