Using nix and nix-related code for package and depedency management

[bjgavilanes]

We want to use biome (https://biomejs.dev/) as a solution, and alternative, for Prettier and ESLint.

Biome is

a [performant linter (https://github.com/biomejs/biome/tree/main/benchmark#linting) for JavaScript, TypeScript, JSX, CSS and GraphQL that features 300 rules from ESLint, TypeScript ESLint, and other sources.

We got two main ways to integrate linter to this infraestructure for this project:

• By using node-related tools -- npm i -D --save-exact @biomejs/biome
• Using system's package manager

I would prefer the second one. npm, and related, invocations would be a bottleneck for the local-first CI pipeline; however, I do know that package.json related workflow is a common practice.

If it we decide about the second option, I would implement nix as a package and dependency management: the developers would even need to install pnpm, they would just install nix for their packages and dependencies to resolve as indicated in the shell.nix

I would some to disscuss this situation.

[hyduez]

Do u guys even know how to use Nix? Nix is a powerful tool, but I don't think that adding something that no ones knows about will work. However, to use Nix, at least, you need to run it on a Unix-like/POSIX system, like GNU/Linux, BSD-like systems, MacOS or in Windows Subsystem for Linux (WSL2 with systemd if u will install it with multi-user daemon). By the way, shell.nix is ambiguous at this time, flake.nix is what it is using by everyone because it let you modify and add features to packages or even to the flake itself.