Skip to main content

NanoForge

This Nanoservices package manager is built on top of Docker which means it is supported by all major CI pipelines. Authentication, caching and version control is handled by Docker, and all major cloud providers support private registries.

Prerequisites​

  1. Docker must be installed and running
  2. Rust must be installed because the Rust tool cargo is needed for compilation
  3. python3 must be installed

Installation​

To install the NanoForge, run the following command:

Windows Users​

There is no windows install script yet, but windows developers can clone the GitHub repository and build from source. Add the binary into an appropriate directory in your $PATH.

On *NIX machines, you can run the installer from wget:

wget -qO- https://raw.githubusercontent.com/nanoservicesforge/NanoForge/main/scripts/install.py | python3

Or instead, if you have curl installed, then run:

curl -s https://raw.githubusercontent.com/nanoservicesforge/NanoForge/main/scripts/install.py | python3

On *NIX machines, the installation process uses the sudo command to insert the compiled tool into your bin path, so you might have to input your password to complete the installation process.

Once the tool is installed, you will need to create a Rust project upon which this tool can act.

Quick Start​

To quick start a project:

cargo new nanoforge-test

Inside the Cargo.toml file, we declare the following nanoservice:

[nanoservices.nan-one]
dev_image = "maxwellflitton/nan-one"
prod_image = "maxwellflitton/nan-one"
entrypoint = "."

We can then download the docker package and configure the Cargo.toml file by running the following command:

nanoforge prep

The nanoforge tool now performs the following steps:

  1. It pulls the Docker images (just maxwellflitton/nan-one in this case)
  2. Unpacks and decompresses the image files, then stores them in the .nanoservices_cache directory
  3. Calculates the relative path of both the Cargo.toml file declaring the nanoservice and the downloaded code, then adds this to the dependencies section of the Cargo.toml

This results in the following update to the [dependencies.nan-one] section of Cargo.toml:

[dependencies.nan-one]
path = "../.nanoservices_cache/domain_services/nanoservices/maxwellflitton_nan-one/."

The entrypoint is added to the end of the relative path so you can have multiple workspaces inside your package and you can just point to the right one you want in the build with the entrypoint.

To test to see if our build works, we can add the following code to the src/main.rs file:

use nan_one::hello;

fn main() {
    println!("Hello, world!");
    hello();
}

We can then perform a cargo run command to see if the code runs correctly.