purpur-servers: init 1.14.1 up to latest

[heisfer]

resolves #37

Things Done

• Updated all purpur builds lock file
• Updated Readme
• nix fmt
• nix flake check
• Tested pkgs.purpurServers.purpur in own config

Test code

{ inputs, pkgs, ... }:
{
  imports = [ inputs.nix-minecraft.nixosModules.minecraft-servers ];
  nixpkgs.overlays = [ inputs.nix-minecraft.overlay ];

  services.minecraft-servers = {
    enable = true;
    eula = true;
    servers = {
      purpur = {
        enable = true;
        package = pkgs.purpurServers.purpur;
      };
    };
  };
}

Notes

• Purpur api doesn't store sha256
• Package is modified version of paper package
• nix fmt formatted other unrelated files

[Infinidoge]

Please revert all formatting changes, they cause a lot of noise in an otherwise clean and straight forward PR.

Additionally, is there any reason why the update script is written to be async? I can't really see any reason why it should be, and it adds unnecessary complexity.

[heisfer]

Please revert all formatting changes, they cause a lot of noise in an otherwise clean and straight forward PR.

Sure thing!

Additionally, is there any reason why the update script is written to be async? I can't really see any reason why it should be, and it adds unnecessary complexity.

The reason for that is fastly download old build and generate sha256 hash (since they only support md5 and I believe md5 is deprecated in nix). I'll revert it back to non async since it's no longer needed.

[heisfer]

Things changed:

1. Reverted update.py
2. Added bad.json so we don't request not found builds each time.
3. Rechecked java versions.
4. Added log4j.xml for specific versions.
5. Formated with nixfmt-rfc-style as it says in flake.nix (CONTRIBUTING.MD says nixpkgs-fmt)

I hope commit message is fit for CONTRIBUTING.md

[Infinidoge]

Overall looks good to me, small question with regards to the java version selector