I spent four days learning #rust-adjacent toolchains (of which there are four, which are doing some things that are the same and some things that are different: #naersk, #fenix, #rustPlatform, #nixpkgs.#rust).

Throughout these four days I have seen at least a dozen of different error messages. Not a single error message had **anything whatsoever** to do with the error site. Reading traces helped, sometimes, but not always.

If you are a company who has excess money and you want to make the world a better place, please have your R&D department pick up #purenix, I urge you.

I ended up downgrading my whole *large* codebase to #rustStable and simplifying my #flake to:

```nix
{
description = "Flake for building rootrunner with standard nixpkgs Rust and OpenSSL support";

inputs = {
nixpkgs.url = "github:NixOS/nixpkgs";
flake-utils = {
url = "github:numtide/flake-utils";
inputs.nixpkgs.follows = "nixpkgs";
};
};

outputs = { self, nixpkgs, flake-utils, ... }:
flake-utils.lib.eachSystem [ "x86_64-linux" ] (system:
let
pkgs = import nixpkgs { inherit system; };
in
{
devShells.default = pkgs.mkShell {
buildInputs = with pkgs; [
rustc
cargo
rustfmt
clippy
rust-analyzer
pkg-config
sqlx-cli
postgresql
sqlite
nodejs
pnpm
zip
unzip
rsync
openssl.dev
openssl
];
};

devShell = self.devShells.${system}.default;
}
);
}
```

#Nix UX is such crap, it's unreal.

I love the tech, and the language is fun (funny), but it doesn't have to be like this.

I wish I had money to self-fund R&D beyond my people's work on #ghc

I would totally invest into #purenix.

If you want to help me reach my goals of bettering the industry beyond what #Serokell is doing, please consider asking your executives to have a look at #ZeroHR for their tech skill assessment needs.

https://zerohr.io

∅hr: Tech hiring. Streamlined.

∅hr is a platform that streamlines tech hiring.

∅hr

Just as I had to got rid of #haskell tag (in favour of #lean), sadly, I feel like it's time to get rid of #elixir tag in my bio.

I'm doing a #rust rewrite of #ZeroHR backend now and while I'm not ruminating too much over the time and money spent on #elixir backend, I find it so much cheaper and faster to write in well-typed languages.

I really hope to become profitable so that I can have a look at #PureNix + #PureErl + #PureScript combo, but I think I'm years away from such a luxury.

@CorbaTheGeek @lisp_discussions

But, finally, I do have to say that nixlisp is only _a_ way, it’s not _the_ way. I believe #purenix (https://github.com/purenix-org/purenix) is a better approach, where you can compile PureScript to Nix. It still requires the codegen step, but probably much closer to being production-ready and you have access to the existing PureScript ecosystem.

Also an honorary mention to #nickel which is supposed to be the next-generation Nix-language, but not yet finished.

GitHub - purenix-org/purenix: Nix backend for PureScript. Transpile PureScript code to Nix.

Nix backend for PureScript. Transpile PureScript code to Nix. - purenix-org/purenix

GitHub