Use schematic for configuration

[milesj]

Description

Big fan of this project, looking forward to its progress.

With that said, I see that you're using raw serde for configuration, and I humbly suggest using my schematic (https://moonrepo.github.io/schematic/config/index.html) crate for configuration instead. I've spent a lot of time fine tuning this crate, as it packs a ton of features. Some of these features include layering + merging + extending, default values, built-in validation, environment variables, and more.

All of this is backed by a powerful derive macro, for example:

use schematic::Config;

#[derive(Config)]
struct AppConfig {
    #[setting(default = 3000, env = "PORT")]
    pub port: usize,

    #[setting(default = true)]
    pub secure: bool,

    #[setting(default = vec!["localhost".into()])]
    pub allowed_hosts: Vec<String>,
}

On top of configuration, the crate is also provides schema modeling (https://moonrepo.github.io/schematic/schema/index.html), which is automatically generated when using the Config macro. Schemas enable powerful code generation, for example, generating TypeScript types or JSON schemas!

And lastly, if using miette, you'd get nice error messages like so:

To Reproduce

N/A

Expected Behavior

N/A

Environment:

Additional Context

Open to collaborating on this if you're interested!

[kaplanelad]

Hey @milesj,
Thank you for reaching out and sharing your project! It looks like a very cool.

At the moment, we don’t have any immediate issues with our current configuration setup. After reviewing the project issues/questions, I do not see user problems in this area, as the current configuration approach is working well for us.

Additionally, we aim to minimize the number of dependencies in our project to reduce compilation time, maintenance efforts, and the learning curve. We are already using serde_yaml for our configuration serialization, so for now, we’ll continue using it as it suits our needs.