1 Month Learning Rust - Packages, Crates, and Modules

1 Month Learning Rust - Packages, Crates, and Modules

Daily short news for you
  • Since the Lunar New Year holiday has started, I won't be posting anymore. See you all after the holiday! 😁

    » Read more
  • Continuing about jj. I'm wondering if there are any GUI software made for it yet to make it easier to use. There are already so many similar to git that I can't count them all.

    Luckily, the author has compiled them all together in Community-built tools around Jujutsu 🥳

    » Read more
  • Turso announces that they are rewriting SQLite in Rust. This adds another piece of evidence supporting the notion that Rust is "redefining" many things.

    But the deeper reason is more interesting. Why are they doing this? Everyone knows that SQLite is open source, and anyone can create a fork to modify it as they wish. Does the Turso team dislike or distrust C—the language used to build SQLite?

    Let me share a bit of a story. Turso is a provider of database server services based on SQLite; they have made some customizations to a fork of SQLite to serve their purposes, calling it libSQL. They are "generous" in allowing the community to contribute freely.

    Returning to the point that SQLite is open source but not open contribution. There is only a small group of people behind the maintenance of this source code, and they do not accept pull requests from others. This means that any changes or features are created solely by this group. It seems that SQLite is very popular, but the community cannot do what they want, which is to contribute to its development.

    We know that most open source applications usually come with a "tests" directory that contains very strict tests. This makes collaboration in development much easier. If you want to modify or add a new feature, you first need to ensure that the changes pass all the tests. Many reports suggest that SQLite does not publicly share this testing suite. This inadvertently makes it difficult for those who want to modify the source code, as they are uncertain whether their new implementation is compatible with the existing features.

    tursodatabase/limbo is the project rewriting SQLite in Rust mentioned at the beginning of this article. They claim that it is fully compatible with SQLite and completely open source. Limbo is currently in the final stages of development. Let’s wait and see what the results will be in the future. For a detailed article, visit Introducing Limbo: A complete rewrite of SQLite in Rust.

    » Read more

Problem

In every programming language, there is something equally important - the package manager. Take JavaScript for example, it is already familiar with npm - a tool that allows you to download and share many useful libraries. Here, developers contribute and use packages to accelerate the product development process.

In Rust, cargo is a tool that functions similarly to npm, it also allows sharing and downloading packages, but the usage is somewhat different. Therefore, in today's article, we will explore the basic components, modules, and the cargo package manager together.

Packages and Crates

Rust provides two concepts, Crates and Packages, to unify the understanding. Simply put, a Crate is a single source file, while Packages are collections of multiple Crates.

cargo is the package manager integrated into Rust, it has similar functionality to Node's npm. This means that we can easily create, add, and remove packages.

The syntax to create a package:

$ cargo new my-project

A directory my-project is created, let's see what's inside it.

$ ls my-project
Cargo.toml
src

The src directory contains a main.rs file, which is the place for executable code. Cargo.toml stores various information about my-project such as its name, dependencies, similar to package.json in a JS/Node project.

To run the project, we use the command cargo run.

$ cargo run
   Compiling my-project v0.1.0 (/Users/hoaitx/my-project)
    Finished dev [unoptimized + debuginfo] target(s) in 0.81s
     Running `target/debug/my-project`
Hello, world!  

Modules

In JavaScript or Node.js, we can import/export modules using familiar syntax such as import/export or require, separating related logic into separate files for reusability. Rust has a similar concept, but with some significant differences.

To understand it simply, let's consider an .rs file as a module in Rust. When declaring a module mymodule in src/main.rs, it means that we are "declaring" to Rust that I will use this module...

pub mod mymodule;

fn main() {
    ...  
}

In this case, Rust will search for your mymodule in the following locations:

  • Inline, meaning within the same file where you declare the module (here it is main.rs).
  • In the file src/mymodule.rs (for newer Rust versions).
  • In the file src/mymodule/mod.rs (for older versions, still supported).

Suppose we write code for mymodule, let's go through each method.

  1. Inline
pub mod mymodule {
    pub fn hello() {
        println!("Hello, world!");
    }
}

fn main() {
    mymodule::hello();
}
  1. In the file src/mymodule.rs

File src/mymodule.rs:

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

File src/main.rs

pub mod mymodule;

fn main() {
    mymodule::hello();
}

This declaration method may seem unfamiliar compared to regular JS/Node, where you export what you need and import what you use. In Rust, modules are "declared" for use, and Rust will automatically search for modules through the specified locations mentioned above.

We can also declare a module within another module. For example, adding a sub-module mysubmodule to src/mymodule.rs:

pub mod mysubmodule {
    pub fn sub_hello() {
        println!("Hello, world!");
    }
}

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

Then in src/main.rs, we can use:

pub mod mymodule;

fn main() {
    mymodule::hello();
    mymodule::mysubmodule::sub_hello();
}

If you notice, you will see that the syntax to access sub-modules is ::, so the deeper the sub-modules, the longer the path. To simplify the code, Rust provides the use syntax to reduce code duplication.

In the file src/main.rs:

pub mod mymodule;

use mymodule::mysubmodule;

fn main() {
    mymodule::hello();
    mysubmodule::sub_hello();
}

use mymodule::mysubmodule has shortened the syntax for calling sub_hello(), and if you want to further shorten it, you can continue accessing sub-modules inside the use statement.

Modules can be declared using absolute or relative paths. Absolute paths start with crate, while relative paths do not. When using absolute paths, its starting position is from src, while with relative paths, it starts directly from the usage position.

mod front_of_house {
    pub mod hosting {
        fn add_to_waitlist() {}
    }
}

pub fn eat_at_restaurant() {
    // absolute path
    crate::front_of_house::hosting::add_to_waitlist();

    // relative path
    front_of_house::hosting::add_to_waitlist();
}

In the eat_at_restaurant function, there are two ways to call the add_to_waitlist function in the front_of_house modules. While the absolute path requires the additional crate declaration, making the code slightly longer compared to the relative path that starts directly from calling into the front_of_house modules. The choice between absolute and relative paths depends on each individual's workflow.

In relative paths, there is also the super syntax to "step back" and call a function in another module.

fn deliver_order() {}

mod back_of_house {
    fn fix_incorrect_order() {
        cook_order();
        super::deliver_order();
    }

    fn cook_order() {}
}

One interesting thing about modules in Rust is that modules declared from the beginning can be accessed throughout the project in different places. It's different from the import/export module syntax you might have seen in other programming languages. For example, in JavaScript, to call the same module in multiple places or files, we need to import it multiple times. But not in Rust.

To make it easier to understand, I will create another module named myothermodule in src/myothermodule.rs, and immediately we can call a function in src/mymodule.rs using super.

# file src/myothermodule.rs

pub fn call_mymodule() {
  super::mymodule::hello();
}

crates.io is the place to aggregate shared packages. You can search, read documentation, download, and use most shared packages here. For example, I want to use a package named rand above. There are two ways to install it:

  • Use the command cargo add rand.
  • Open Cargo.toml and add the name and version of the package under [dependencies]. Then use the cargo fetch command if your code editor does not automatically download the newly added package.

To use these packages, we don't need to declare modules but can directly use use. For example, calling a function that generates a random number in the range 1-100.

use rand::Rng;

fn main() {
    let secret_number = rand::thread_rng().gen_range(1..=100);
}
Premium
Hello

5 profound lessons

Every product comes with stories. The success of others is an inspiration for many to follow. 5 lessons learned have changed me forever. How about you? Click now!

Every product comes with stories. The success of others is an inspiration for many to follow. 5 lessons learned have changed me forever. How about you? Click now!

View all

Subscribe to receive new article notifications

or
* The summary newsletter is sent every 1-2 weeks, cancel anytime.

Comments (0)

Leave a comment...
Scroll or click to go to the next page