Embedding Default Configuration Files in Rust Binaries with rust-embed

In real‑world development, applications often read configuration files from external sources to increase flexibility. While Spring Boot uses application.yml packaged inside the JAR, this guide shows how to achieve a similar effect in the Rust ecosystem.

By using the third‑party rust-embed library, you can embed a default configuration file into the compiled binary and optionally replace it with a file in the program’s directory.

Define the embedded file location and accessor

In src/resources/embed_resource.rs we declare the embed folder and a function to retrieve the default config:

use rust_embed::RustEmbed;

#[derive(RustEmbed)]
#[folder = "src/embedfiles/"]
struct Asset;

pub fn get_app_default() -> Option<rust_embed::EmbeddedFile> {
    Asset::get("./app_default.yml")
}

The macro tells rust‑embed to embed all files under src/embedfiles/. The accessor returns an EmbeddedFile struct for the requested path.

Use the embedded file

In main.rs we first check whether app.yml exists next to the executable; if it does, we read and print it. Otherwise we fall back to the embedded default configuration:

fn main() {
    if Path::new("./app.yml").exists() {
        let contents = fs::read_to_string("./app.yml").expect("Read file error!");
        println!("{}", contents);
        return;
    }
    let app = get_app_default().unwrap();
    let f = std::str::from_utf8(app.data.as_ref()).unwrap();
    println!("{}", f);
}

This priority logic ensures the external app.yml overrides the embedded default when present.

Testing the behavior

Compile the program and move the binary to /tmp to avoid interference:

cargo build
mv target/debug/embed /tmp

Running the binary without an app.yml prints the contents of the embedded default file. Creating an app.yml in the same directory with custom content and rerunning the binary prints the new file’s contents instead.

Source code: https://github.com/jiashiwen/wenpanrust/tree/main/embed

The example builds and runs successfully on macOS; stay tuned for the next installment.