Rust quickstart for JavaScript programmers

So, you want to learn Rust, and fast. This guide is for people who already have experience programming, at least in languages like JavaScript or Python. It will make it a bit easier if you have experience with compiled, strongly-typed, or functional languages like C++, Java, or Haskell, but that’s not required.

Why another Rust guide? This is a rough outline that I have used in the past when holding Rust for Beginners workshops. Feel free to contact me if you want me to host a workshop for you.

Rust in a sentence

Rust is a general-purpose, statically-typed, functionally-inspired, high-level programming language that specializes in memory safety and speed and features an algebraic type system.

• General-purpose — Not limited to a particular application or platform, Rust can be used to build a diverse variety of solutions.
• Statically-typed — Data types are known (or computed) and enforced at compile time.
• Functionally-inspired — Many features are imported from, or are derivatives of features from FP-paradigm languages like Ocaml and Haskell.
• High-level — Rust provides a large degree of abstraction from assembler code.
• Memory safety — All safe code (that which does not use the unsafe keyword) is guaranteed by the compiler to not contain memory safety issues such as dangling pointers, null dereferencing, or accidental memory leaks.¹
• Speed — Rust code compiles directly to machine code with zero or near-zero² runtime overhead. It can fairly easily interoperate directly with C code.³ With the absence of a garbage collector, or really any runtime to speak of, memory is managed by the programmer, albeit within the strict bounds enforced by the compiler’s memory safety features, notwithstanding unsafe blocks.
• Algebraic type system — While most modern programming languages include some form of product type, Rust also includes a fully-featured sum type as well, in the form of a discriminated union (enum).

Set up

If you don’t want to install anything, you can just use the Rust playground, and skip this section. Otherwise, if you want to set up your own system, just follow the guide on the Rust website, as it’s not worth duplicating the information here.

Tools

If you follow the website, you’ll end up with a few different tools installed on your system:

• rustup manages your Rust installation(s): what versions you have installed, for which targets, etc.
• rustc is the Rust compiler. Although you can use it directly, it is cumbersome for large and complex projects.
• cargo will probably be the most-used tool of the three mentioned here during your Rust career. It is an all-in-one project manager, dependency installer, linter, formatter, documentation builder, and build tool. (Technically it delegates out most of those commands to dedicated tools like rustc, rustdoc, rustfmt, and clippy, but Cargo brings them all together nicely.)

You can check that everything is installed properly by running cargo version, and you should see something like this:

$ cargo version
cargo 1.66.1 (ad779e08b 2023-01-10)

You can check the latest stable version number on the Releases page on GitHub, and you can upgrade the version currently installed on your computer with this command:

rustup update stable

Creating a project with Cargo

Let’s get up-and-running as quickly as possible. If you’re familiar with Node.js, Cargo may feel a bit similar.

Create a new project in the directory hello-rust:

cargo new --bin hello-rust

The --bin flag means that the new project will create standalone executable binary. The other option is the --lib flag, which will create a library package, which will not compile into an executable application. If neither flag is specified, creating a binary application is the default.

Let’s quickly make sure that everything is working correctly by building and running the default generated “Hello World” project:

cd hello-rust
cargo run

cargo run will implicitly run cargo build to build the application if the existing build files are out-of-date.

If everything worked according to plan, you should see something like the following:

   Compiling hello-rust v0.1.0 (/.../hello-rust)
    Finished dev [unoptimized + debuginfo] target(s) in 0.92s
     Running `target/debug/hello-rust`
Hello, world!

If everything appears to be working correctly, open up src/main.rs in your favorite editor.

The language

The default “Hello, world!” program in Rust looks something like this:

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

If you’re already familiar with languages like C, C++, Java, etc., this sort of boilerplate should feel natural:

• The default program entrypoint is a function called main. It takes no arguments and has no return type.
• The body of the function is enclosed in curly braces.
• The contents of the function are:
  • A function⁴ call to println! with a single argument, the string "Hello, world!".

Simple function definitions typically look like this:

fn function_name(param1: Type1, param2: Type2) -> ReturnType {
    // body
}

where the parameter list (everything inside the round brackets (...)) and return type (everything between the closing parenthesis ) and the opening curly bracket {) are optional, depending on the needs of your function.

From here on, all of the code we’ll be discussing can go directly inside the main function body (although some things, like structs and other function definitions, can, and are usually recommended to be placed outside).

Basic bindings and numeric primitives

The first thing we’ll look at is variable bindings.

let my_variable = 10;

(In addition to being valid Rust, this is also valid JavaScript—how about that!)

Though the syntax here is pretty simple, there’s a bit more going on under the hood.

Remember how Rust is a statically-typed programming language? Well, in this case, the type is not explicit in the code, but the compiler still assigns it one. In this case, the type of this variable is i32, which is the default for otherwise unrestricted number types. i32 means “32-bit signed integer”.

As you may have guessed, there are a bunch of other number types. Generally, the names take the form of <single-letter-prefix><bit-size>.

Examples:

• i8 — signed byte
• u64 — unsigned 64-bit integer
• usize — unsigned word-size integer (platform dependent: 32 bits on a 32-bit machine, 64 bits on a 64-bit machine, etc.)
• f32 — float
• f64 — double

We have three options to explicitly indicate the type of our binding:

1. Copy

   let my_variable: i32 = 10;

   This syntax declares the type of the binding, and the expression is coerced to that type if possible.

2. Copy

   let my_variable = 10 as i32;

   This syntax casts the expression to the given type, which the binding then assumes.

3. Copy

   let my_variable = 10i32;

   This syntax is unique to numeric literals, and it declares the type of the expression, which the binding then assumes.

If the compiler is able to determine the appropriate type for your binding by itself, you do not need to specify it in the source code. Note that type declarations for function parameters and return types are always required (unless the function does not return anything).

Other primitives

Booleans

In addition to numeric types, Rust also has booleans:

// This is a comment; ignored by the compiler
// This type annotation is superfluous, but it is included for clarity.
let happy_to_learn_rust: bool = true;

let ever_going_to_give_you_up = false;

Characters

Single Unicode characters, delimited by single quotes ':

let currency_symbol: char = '$';
// Multi-byte characters allowed!
let kanji_character: char = '字';

Strings

This can be a bit of a tricky topic in Rust, because there are a few different string types to think about. However, you usually only need to worry about two:

• String is a heap-allocated string that can be mutated in-place. This is called an “owned string”. When learning Rust, this is the string type you should probably reach for first, to save on headaches.
• &str is a fixed-length string that cannot be mutated. This can be called a “string slice” or “string reference”.

(Technically this is a bit of an oversimplification, but it will get you 90% of the way there. Once we’ve gone over references and ownership we can come back to this topic.)

For now, you can convert between the two string types fairly easily:

let my_str: &str = "hello";
let my_string: String = my_str.to_string();
let another_str: &str = &my_string;

Arrays

Arrays are fixed-length, homogenous collections delimited by square brackets [] with elements separated by commas ,. Note the type signature takes the form of [<element-type>; <length>].

let my_i32_array: [u32; 4] = [1, 2, 3, 4];
let my_bool_array: [bool; 0] = [];
let my_char_array: [char; 3] = ['a', 'b', 'c'];

// Error: type signature has incorrect length
// let incorrect_array: [char; 100] = [];

Tuples

Tuples are fixed-length, heterogenous collections delimited by round brackets () with elements separated by commas ,.

let my_tuple: (i32, char, [bool; 2]) = (1, 'a', [true, false]);
// It's probably easier to just let the compiler compute the type
let my_tuple = (1, 'a', [true, false]);

More about bindings

Mutability

This code will not compile:

let x = 0;
x = 1;

The let keyword in Rust works a bit like how the const keyword works in JavaScript, in that all identifiers it declares are immutable. In JavaScript, const identifiers simply cannot be reassigned, but its properties can still be changed. In Rust, immutability is total: no reassignments, and no mutation of contents either. You could say it’s like const + Object.freeze in JavaScript.

This seems like a pretty stringent limitation at first. However, there’s an escape hatch: the mut keyword. Actually, if you’ve been following along and trying out the code examples, you’ll already know about it—here’s the error message you’d get if you tried to compile the above example:

error[E0384]: cannot assign twice to immutable variable `x`
 --> src\main.rs:4:5
  |
3 |     let x = 0;
  |         -
  |         |
  |         first assignment to `x`
  |         help: consider making this binding mutable: `mut x`
4 |     x = 1;
  |     ^^^^^ cannot assign twice to immutable variable

For more information about this error, try `rustc --explain E0384`.

Do you see it? The Rust compiler itself suggests a way to fix our code by using the mut keyword!

Having to use a whole additional keyword just to make a variable mutable may seem kind of inconvenient at first. However, immutable-by-default variable declarations facilitate certain compile-time guarantees about your code, which can help detecting bugs early and make code more flexible (e.g. for concurrent programming).

Shadowing

You may have noticed that a previous example re-bound the my_tuple identifier a second time using the let keyword. If we were to do something like this in a language like JavaScript, we’d get an error:

Uncaught SyntaxError: redeclaration of let my_tuple

In Rust, however, this is allowed. It simply defines another (new, discrete) identifier that just happens to shadow (override the name of) another identifier.

This can be useful when you wish to change the type or mutability of an identifier.

Scoping

Identifiers in Rust are block-scoped, meaning that an identifier declared within a set of curly braces {} is inaccessible outside of the braces.

fn main() {
    let x = 0;
    { // `y` only exists in this scope
        let y = 1;
        println!("{} {}", x, y);
    } // scope ends
    println!("{} {}", x, y); // `y` is out of scope
}

Output:

error[E0425]: cannot find value `y` in this scope
 --> src/main.rs:7:26
  |
7 |     println!("{} {}", x, y); // `y` is out of scope
  |                          ^ help: a local variable with a similar name exists: `x`

More on this later, when we talk about ownership.

More data structures

If you wanted to, you could probably do most of your programming with just the primitive data structures described above (arrays, tuples, and primitives). However, Rust still has more to offer.

Structs

To a programmer familiar with an object-oriented or classical style of programming, structs should feel familiar. They look and often feel very similar to a class or dictionary-esque value.

First, we define the type:

struct BlogPost {
    title: String,
    text: String,
    author: String,
    timestamp: u64,
}

Then, to create an instance:

let mut post = BlogPost {
    title: "Rust quickstart for JavaScript programmers".to_string(),
    text: "So, you want to learn Rust, and fast...".to_string(),
    author: "Jacob Lindahl".to_string(),
    timestamp: 1600000000000,
};

Member access:

// read
println!("{}", post.title);

// mutate
post.author = "John Doe".to_string();

Now, really, structs are just a way to give nice labels to pieces of a tuple. This is evident in some of the other ways you can declare a struct type:

struct MyTupleStruct(u8);
struct RGBColor(u8, u8, u8);
struct RGBAColor(u8, u8, u8, u8);

If you’re familiar with the newtype pattern, this is usually how it is implemented in Rust.

If you really want, you can also create completely empty structs:

struct LiterallyNothing;

While these may not initially seem to be useful, implementing some traits on them may change that.

Enums

Rust’s enums are one of its most powerful features. Although languages like Java and TypeScript offer a primitive form of enum, Rust’s are fully-featured sum types, implemented as tagged unions.

If you’re coming from a language like Java, the most basic enum will look pretty familiar:

enum LogLevel {
    Error,
    Warn,
    Info,
    Debug,
}

However, Rust is just getting started. Each variant of an enum can contain data, just like a struct or a tuple.

enum ImageFilter {
    InvertColors,
    Blur(f64),
    HueRotate(f64),
    DropShadow {
        x: f64,
        y: f64,
        blur: f64,
        color: String,
    },
}

The power doesn’t end there: keep reading for more!

Control flow

Rust has most of the control flow expressions you may expect, and a few you may not.

if

Compared to other curly-brace languages (like C++, Java, and JavaScript), Rust’s if statements have two notable differences:

• Lack of the requirement for the condition expression to be enclosed in parentheses.

  Example

  Rust

  Copy

  if a > b {
      // ...
  }

  JavaScript, etc.

  Copy

  if (a > b) {
    // ...
  }

• if structures in Rust can be expressions as well as just normal statements. That is, they can resolve to a value like a ternary expression in JavaScript.

  Example

  Rust

  Copy

  let value = if a > b { a + b } else { a * b };

  JavaScript

  Copy

  let value = a > b ? a + b : a * b;

while & loop

Rust has three looping structures. loop is the simplest: it just loops forever until it hits a break. while, like its namesake in other languages, loops while a condition holds (or until it hits a break). Similarly in style to the if statement, the while condition does not need to be enclosed in parentheses.

let mut x = 0;
loop {
    if x >= 10 {
        break;
    }
    x += 1;
}

let mut x = 0;
while x < 10 {
    x += 1;
}

The third looping structure is the for loop. It operates on any iterable type.

Arrays:

for i in [2, 4, 6, 8] {
    println!("{i}");
}

Output:

2
4
6
8

Ranges:

for i in 0..5 {
    println!("{i}");
}

Output:

0
1
2
3
4

There are other iterable structures, like Vec and HashSet, which you can explore if you wish.

match

Instead of the switch statement found in many other common languages, Rust opted for the more “functional” match construct.

let operator = "*";

match operator {
    "+" => println!("add"),
    "-" => println!("subtract"),
    "*" => println!("multiply"),
    "/" => println!("divide"),
    _ => println!("unknown"), // `_` is the catch-all pattern
}

While matching against input cases using Rust’s pattern matching syntax, you can also extract pieces from the input data using destructuring.

enum MediaType {
    Movie,
    Series { episodes: u32 },
}

let media_type: MediaType = /* ... */;

match media_type {
    MediaType::Movie => println!("It's a movie!"), // single line terminated with comma
    MediaType::Series { episodes } => { // multi-line enclosed in curly braces
        println!("It's a TV show!");
        println!("It has {episodes} episodes!");
    }
}

match expressions, since they are expressions, can also resolve to a value:

let unit_count = match media_type {
    MediaType::Series { episodes } => episodes,
    _ => 1,
};

if let

A special version of the if expression can perform destructuring as well. It’s like a conditional let.

This is equivalent to the previous listing:

let unit_count = if let MediaType::Series { episodes } = media_type {
    episodes
} else {
    1
};

Adding behavior

impl blocks

We can add behavior to individual types with a simple impl block:

struct BasicGreeter {
    greeting: String,
}

impl BasicGreeter {
    fn greet(&self, name: &str) {
        println!("{}, {name}!", self.greeting);
    }
}

let g = BasicGreeter {
    greeting: "Welcome".to_string(),
};
g.greet("John");

Output:

Welcome, John!

The &self parameter is special: it causes a function to be a method, operating on an instance of a type, as opposed to an associated function, which does not necessarily operate on an instance. In object-oriented terms, functions that take a self parameter (or any of the variants) are like instance methods, and functions that do not are like static methods.

Traits

Traits are the primary form of abstraction in Rust. A trait describes a set of behaviors that a type implements. It’s very similar to an interface in a language like Java. Actually, we’ve been using some traits already, in a subtle sort of way.

Remember how to print things to the screen? println!(...)? As you probably noticed by now, we can print a bunch of different things. Numbers, strings, booleans, etc.

println!("{}, {}, {}, {}", "hello", 42, 3.14, true);

This is an example of abstraction: all of the different types all support the behavior of “being printed.”

In Rust, this behavior is described by the Display trait, which is like adding a toString method to a class in Java.⁵

All of these types (String, &str, u32, f32, bool, …) implement Display.

Let’s take a look at writing a trait and implementing it.

// create a trait
trait Greeter {
    fn greet(&self, name: &str);
}

struct MorningGreeter;

// implement the trait on MorningGreeter
impl Greeter for MorningGreeter {
    fn greet(&self, name: &str) {
        println!("Good morning, {name}!");
    }
}

struct EveningGreeter;

// implement the trait on EveningGreeter
impl Greeter for EveningGreeter {
    fn greet(&self, name: &str) {
        println!("Good evening, {name}!");
    }
}

let g1 = MorningGreeter;
g1.greet("Alice"); // -> Good morning, Alice!

let g2 = EveningGreeter;
g2.greet("Bob"); // -> Good evening, Bob!

This isn’t terribly interesting yet. Let’s spice it up some.

fn greet_wizard<G: Greeter>(g: G) {
    g.greet("Gandalf");
}

greet_wizard(EveningGreeter); // -> Good evening, Gandalf!

This uses a generic type parameter to allow us to pass it any parameter that implements the Greeter trait. If you’re familiar with Java, etc., the angle bracket <> syntax might look familiar.

The colon in G: Greeter means “G implements Greeter.” You can specify multiple trait bounds using a +, like so: T: Debug + Display.

If you have too many generic parameters, or if the bounds are too complex, you can move them to a where clause to organize your function signature a little:

fn my_function<T>(t: T) where T: Send + Sync {}

If you don’t need to refer to the generic parameter by name, you can use a shorthand:

fn greet_wizard(g: impl Greeter) {
    g.greet("Gandalf");
}

greet_wizard(EveningGreeter); // -> Good evening, Gandalf!

This code is identical to the previous greet_wizard example, but it’s a little easier to read, since there’s no G generic parameter floating around.

Trait objects

There’s another way to accept parameters based on what traits they implement, as opposed to by concrete type. Using the dyn keyword, we can create a trait object. It’s a bit of a hairy topic if you dive deeply into it, but for now, keep in mind the following properties:

• dyn Trait is unsized, meaning you usually can’t work with it directly, since the compiler cannot know how big it is.
• Therefore, trait objects are usually used behind some form of pointer, either regular (&dyn Trait) or smart (Box<dyn Trait>).
• Trait objects include a vtable, which can make function access a tiny bit slower. Usually, Rust’s monomorphized generics are preferred, since they can be optimized per-type, and also preserve type information across the codebase.

References & ownership

If you’ve worked with a systems programming language like C++ or C before, you’ve probably heard of “pointers.” Rust also has an equivalent construct, also called (raw) pointers. However, as any C++ or C developer would tell you, you need to be careful when dealing with pointers. There are a lot of potential issues that arise when working with pointers:

• Null pointers
• Dangling pointers / use after free
• Double free
• Data races
• Buffer overflow
• etc.

However, because pointers provide a layer of indirection invaluable to programmers, Rust didn’t discard the concept in its pursuit of memory safety. Instead, it vastly improved upon it with its notion of references and the borrow checker.

At runtime, references serve the same purpose as pointers: a layer of indirection to some desired data. However, Rust applies a set of rules to how references can be used in valid Rust code to ensure memory safety and avoid all of the issues mentioned above.

The ownership rules look like this:

• Every value has one single owner.
• When the owner goes out of scope, the value is dropped.

Here’s an example:

{ // begin
    let x = 6;
} // end

In this example, x is the owner of the value 6. Once x goes out of scope (at the line marked // end), the 6 ceases to exist as well.

Ownership of a value can be transferred:

let my_string: String = "Hello, world!".to_string();
let moved_string: String = my_string;

At first, my_string is the owner of the string. Then, the string is moved into moved_string. “Moving” is Rust-speak for transferring ownership.

What happens now that the string has been moved out of my_string and into moved_string? Can we still use my_string? Let’s try:

let my_string = "Hello, world!".to_string();
let moved_string = my_string;
println!("{}", my_string);

Output:

error[E0382]: borrow of moved value: `my_string`
 --> src/main.rs:4:16
  |
2 | let my_string = "Hello, world!".to_string();
  |     --------- move occurs because `my_string` has type `String`, which does not implement the `Copy` trait
3 | let moved_string = my_string;
  |                    --------- value moved here
4 | println!("{}", my_string);
  |                ^^^^^^^^^ value borrowed here after move

Let’s walk through the error message:

• borrow of moved value: `my_string`

  To “borrow” a value is to take a reference to it. This says that we’re trying to create a reference to a value that has been moved away: the container that used to hold the value—my_string—is empty!

• move occurs because `my_string` has type `String`, which does not implement the `Copy` trait

  There are some values that don’t really need the power of move semantics. These are usually small, stack-allocated, statically-sized values like many of the primitives. Instead of getting moved from one owner to another, these values are just copied from one place to another, since it’s so cheap to do. Types like this implement the Copy trait. References are also Copy!

• value moved here

  The Rust compiler shows us exactly where in the code the value was moved out of my_string.

• value borrowed here after move

  The println!(...) macro borrows the values it prints out, since it only needs to read them.

Since a value can only have one owner, references give us a way to pass around a reference to a value that someone else owns, allowing that value to be used in more than one place at a time. A normal & reference is read-only⁶, so holding one does not allow you to mutate the underlying value. The exclusive &mut reference, on the other hand, allows the owner of the reference to change the underlying value without the value’s owner having to give up ownership.

References also have a few special rules to go along with them:

• There can be an unlimited number of normal & references alive at one time, OR
• There can be a maximum of one exclusive &mut reference (also known as a “mutable reference”) alive at one time.

Lifetimes

Let’s write a function that takes two string references and returns the longer of the two:⁷

fn longest(a: &str, b: &str) -> &str {
    if a.len() > b.len() {
        a
    } else {
        b
    }
}

This code doesn’t actually compile!

Output:

error[E0106]: missing lifetime specifier
 --> src/main.rs:2:37
  |
2 |     fn longest(a: &str, b: &str) -> &str {
  |                   ----     ----     ^ expected named lifetime parameter
  |
  = help: this function's return type contains a borrowed value, but the signature does not say whether it is borrowed from `a` or `b`
help: consider introducing a named lifetime parameter
  |
2 |     fn longest<'a>(a: &'a str, b: &'a str) -> &'a str {
  |               ++++     ++          ++          ++

Memory safety

The Rust compiler is your friend. It’s trying to make sure that your code is memory safe (and a lot of other things), so it’s not going to compile code for you that might be memory-unsafe.

This function signature says “I’m taking in two string references as input, and I’m returning a string reference.” Where could the reference that the function returns come from? Functions can access data from three places:

• Values provided in the function arguments.
• Values generated in the body of the function.
• Global values.

So, if a function is returning a reference to a value, the owner of the value must be from one of those three places.

Actually, we can rule out one of the places entirely. Due to the ownership rules that dictate that a value is dropped when its owner goes out of scope, all values that are generated by (and therefore, owned by) the function are dropped when the function ends (unless they are returned or ownership is otherwise transferred). Therefore, a function can never return a reference to a value that is owned by the function, since this would create a dangling pointer.

Therefore, a function that returns a reference can only be referencing data that is either static/global or passed in via the arguments.

In order to prevent use-after-free errors, the Rust compiler also needs to know for how long a reference is valid. If you have a reference to value X and the owner of X goes out of scope, X will be dropped, so Rust needs to ensure that all references to X will not be used after that occurs.

Lifetime annotations

In order to provide this assurance, Rust uses the concept of lifetimes. A value’s lifetime tells you for how long it is OK to hold onto and use that value. Global values and owned values can be held onto for as long as you want, so they have a static lifetime. This is a specially-named lifetime in Rust, and it’s denoted as 'static. It means “you can hold onto this value for as long as you want.”

However, if you receive a reference to a value, you’re not the owner of that value, so it could go out of scope at some point—a point in time that you do not control. So, when a function returns a reference, it needs to tell the Rust compiler how long it is OK to hold onto that reference.

We do this using lifetime annotations. Here’s a super-simple example:

fn str_identity<'a>(s: &'a str) -> &'a str {
    s
}

This function simply returns the exact string reference it was given. But let’s look at the new syntax.

First, we instantiate a lifetime specifier <'a>.⁸ All lifetime specifiers begin with the tick/apostrophe/prime character '. You’ll notice that lifetime specifiers are declared in the same place as generic type parameters, and this is because lifetimes are actually a kind of generic variable: this function can be called on any string reference, not just one specific “lifetime” of string reference. (There are many more powerful ways to create bounds on and with lifetime specifiers, but we’ll skip that.)⁹

Next, we use the lifetime specifier in two places:

• s: &'a str

  s is a reference to a value that has some lifetime, and the 'a lifetime specifier will represent a lifetime that lives no longer than the value that s references. The Rust compiler will try to choose the longest possible lifetime here.¹⁰

• -> &'a str

  This function will return a reference to a string value that is guaranteed to live at least as long as 'a.

Let’s look back at that error message:

error[E0106]: missing lifetime specifier
 --> src/main.rs:2:37
  |
2 |     fn longest(a: &str, b: &str) -> &str {
  |                   ----     ----     ^ expected named lifetime parameter
  |
  = help: this function's return type contains a borrowed value, but the signature does not say whether it is borrowed from `a` or `b`
help: consider introducing a named lifetime parameter
  |
2 |     fn longest<'a>(a: &'a str, b: &'a str) -> &'a str {
  |               ++++     ++          ++          ++

We see that the Rust compiler is suggesting that we add some lifetime specifiers to our code. It’s telling us to add a lifetime specifier to our return value that is bounded by (lasts no longer than) the values referenced by a and b. That tells the compiler that the value returned by longest is guaranteed to be valid while both a and b are valid.

If a struct contains a reference, it is required to use lifetime specifiers:

struct StringWrapper<'a>(&'a str);

Of course, to maintain memory safety, the struct can only be used while 'a is valid (that is, so long as the contained reference refers to a value that is still alive).

Note: Lifetimes is a somewhat stubborn topic that can take a little bit of work to understand, so don’t worry if it doesn’t click the first time around!

Resources

Reading materials

More posts for learning Rust

• Fathomable Rust Macros. Rust procedural macros.
• Nothing in Rust. Different ways of representing nonexistence in Rust.

External resources

• The Rust Book. Everyone should read.
• The Rustonomicon. Advanced and specific topics. Often more useful as a reference than a read-through.
• Common Rust Lifetime Misconceptions. Excellent write-up that may help to dispel some common misunderstandings about how lifetimes work. However, it is a bit technical, so don’t feel like you have to understand everything it says before you can use Rust.

Exercises

• Rust Playground is not actually a source of exercises, but it is a good place to try out some code without having to spin up an editor and new project.
• Rustlings is a repository of all sorts of exercises to help you get comfortable using Rust.
• Exercism provides a variety of online exercises and other resources.

Jacob Lindahl is a graduate student at the Tokyo Institute of Technology.

Connect with Jacob on Twitter, Mastodon, and Farcaster.