Welcome to Comprehensive Rust 🦀

This is a free Rust course developed by the Android team at Google. The course covers the full spectrum of Rust, from basic syntax to advanced topics like generics and error handling.

The latest version of the course can be found at https://google.github.io/comprehensive-rust/. If you are reading somewhere else, please check there for updates.

The course is available in other languages. Select your preferred language in the top right corner of the page or check the Translations page for a list of all available translations.

The course is also available as a PDF.

The goal of the course is to teach you Rust. We assume you don't know anything about Rust and hope to:

• Give you a comprehensive understanding of the Rust syntax and language.
• Enable you to modify existing programs and write new programs in Rust.
• Show you common Rust idioms.

We call the first four course days Rust Fundamentals.

Building on this, you're invited to dive into one or more specialized topics:

• Android: a half-day course on using Rust for Android platform development (AOSP). This includes interoperability with C, C++, and Java.
• Chromium: a half-day course on using Rust within Chromium based browsers. This includes interoperability with C++ and how to include third-party crates in Chromium.
• Bare-metal: a whole-day class on using Rust for bare-metal (embedded) development. Both microcontrollers and application processors are covered.
• Concurrency: a whole-day class on concurrency in Rust. We cover both classical concurrency (preemptively scheduling using threads and mutexes) and async/await concurrency (cooperative multitasking using futures).

Non-Goals

Rust is a large language and we won't be able to cover all of it in a few days. Some non-goals of this course are:

• Learning how to develop macros: please see Chapter 19.5 in the Rust Book and Rust by Example instead.

Assumptions

The course assumes that you already know how to program. Rust is a statically-typed language and we will sometimes make comparisons with C and C++ to better explain or contrast the Rust approach.

If you know how to program in a dynamically-typed language such as Python or JavaScript, then you will be able to follow along just fine too.

Running the Course

This page is for the course instructor.

Here is a bit of background information about how we've been running the course internally at Google.

We typically run classes from 9:00 am to 4:00 pm, with a 1 hour lunch break in the middle. This leaves 3 hours for the morning class and 3 hours for the afternoon class. Both sessions contain multiple breaks and time for students to work on exercises.

Before you run the course, you will want to:

1. Make yourself familiar with the course material. We've included speaker notes to help highlight the key points (please help us by contributing more speaker notes!). When presenting, you should make sure to open the speaker notes in a popup (click the link with a little arrow next to "Speaker Notes"). This way you have a clean screen to present to the class.

2. Decide on the dates. Since the course takes four days, we recommend that you schedule the days over two weeks. Course participants have said that they find it helpful to have a gap in the course since it helps them process all the information we give them.

3. Find a room large enough for your in-person participants. We recommend a class size of 15-25 people. That's small enough that people are comfortable asking questions --- it's also small enough that one instructor will have time to answer the questions. Make sure the room has desks for yourself and for the students: you will all need to be able to sit and work with your laptops. In particular, you will be doing a lot of live-coding as an instructor, so a lectern won't be very helpful for you.

4. On the day of your course, show up to the room a little early to set things up. We recommend presenting directly using mdbook serve running on your laptop (see the installation instructions). This ensures optimal performance with no lag as you change pages. Using your laptop will also allow you to fix typos as you or the course participants spot them.

5. Let people solve the exercises by themselves or in small groups. We typically spend 30-45 minutes on exercises in the morning and in the afternoon (including time to review the solutions). Make sure to ask people if they're stuck or if there is anything you can help with. When you see that several people have the same problem, call it out to the class and offer a solution, e.g., by showing people where to find the relevant information in the standard library.

That is all, good luck running the course! We hope it will be as much fun for you as it has been for us!

Please provide feedback afterwards so that we can keep improving the course. We would love to hear what worked well for you and what can be made better. Your students are also very welcome to send us feedback!

Course Structure

This page is for the course instructor.

Rust Fundamentals

The first four days make up Rust Fundamentals. The days are fast paced and we cover a lot of ground!

Course schedule:

• Day 1 Morning (2 hours and 5 minutes, including breaks)

• Day 1 Afternoon (2 hours and 35 minutes, including breaks)

• Day 2 Morning (2 hours and 10 minutes, including breaks)

• Day 2 Afternoon (3 hours and 15 minutes, including breaks)

• Day 3 Morning (2 hours and 20 minutes, including breaks)

• Day 3 Afternoon (1 hour and 55 minutes, including breaks)

• Day 4 Morning (2 hours and 40 minutes, including breaks)

• Day 4 Afternoon (2 hours and 20 minutes, including breaks)

Deep Dives

In addition to the 4-day class on Rust Fundamentals, we cover some more specialized topics:

Rust in Android

The Rust in Android deep dive is a half-day course on using Rust for Android platform development. This includes interoperability with C, C++, and Java.

You will need an AOSP checkout. Make a checkout of the course repository on the same machine and move the src/android/ directory into the root of your AOSP checkout. This will ensure that the Android build system sees the Android.bp files in src/android/.

Ensure that adb sync works with your emulator or real device and pre-build all Android examples using src/android/build_all.sh. Read the script to see the commands it runs and make sure they work when you run them by hand.

Rust in Chromium

The Rust in Chromium deep dive is a half-day course on using Rust as part of the Chromium browser. It includes using Rust in Chromium's gn build system, bringing in third-party libraries ("crates") and C++ interoperability.

You will need to be able to build Chromium --- a debug, component build is recommended for speed but any build will work. Ensure that you can run the Chromium browser that you've built.

Bare-Metal Rust

The Bare-Metal Rust deep dive is a full day class on using Rust for bare-metal (embedded) development. Both microcontrollers and application processors are covered.

For the microcontroller part, you will need to buy the BBC micro:bit v2 development board ahead of time. Everybody will need to install a number of packages as described on the welcome page.

Concurrency in Rust

The Concurrency in Rust deep dive is a full day class on classical as well as async/await concurrency.

You will need a fresh crate set up and the dependencies downloaded and ready to go. You can then copy/paste the examples into src/main.rs to experiment with them:

cargo init concurrency
cd concurrency
cargo add tokio --features full
cargo run

Course schedule:

• Morning (3 hours and 20 minutes, including breaks)

• Afternoon (3 hours and 20 minutes, including breaks)

Format

The course is meant to be very interactive and we recommend letting the questions drive the exploration of Rust!

Keyboard Shortcuts

There are several useful keyboard shortcuts in mdBook:

• Arrow-Left: Navigate to the previous page.
• Arrow-Right: Navigate to the next page.
• Ctrl + Enter: Execute the code sample that has focus.
• s: Activate the search bar.

Translations

The course has been translated into other languages by a set of wonderful volunteers:

• Brazilian Portuguese by @rastringer, @hugojacob, @joaovicmendes, and @henrif75.
• Chinese (Simplified) by @suetfei, @wnghl, @anlunx, @kongy, @noahdragon, @superwhd, @SketchK, and @nodmp.
• Chinese (Traditional) by @hueich, @victorhsieh, @mingyc, @kuanhungchen, and @johnathan79717.
• Farsi by @DannyRavi, @javad-jafari, @Alix1383, @moaminsharifi , @hamidrezakp and @mehrad77.
• Japanese by @CoinEZ-JPN, @momotaro1105, @HidenoriKobayashi and @kantasv.
• Korean by @keispace, @jiyongp, @jooyunghan, and @namhyung.
• Spanish by @deavid.
• Ukrainian by @git-user-cpp, @yaremam and @reta.
• Farsi by @DannyRavi, @javad-jafari, @Alix1383, @moaminsharifi, @hamidrezakp and @mehrad77.

Use the language picker in the top-right corner to switch between languages.

Incomplete Translations

There is a large number of in-progress translations. We link to the most recently updated translations:

• Arabic by @younies
• Bengali by @raselmandol.
• French by @KookaS, @vcaen and @AdrienBaudemont.
• German by @Throvn and @ronaldfw.
• Italian by @henrythebuilder and @detro.

The full list of translations with their current status is also available either as of their last update or synced to the latest version of the course.

If you want to help with this effort, please see our instructions for how to get going. Translations are coordinated on the issue tracker.

Using Cargo

When you start reading about Rust, you will soon meet Cargo, the standard tool used in the Rust ecosystem to build and run Rust applications. Here we want to give a brief overview of what Cargo is and how it fits into the wider ecosystem and how it fits into this training.

Installation

Please follow the instructions on https://rustup.rs/.

This will give you the Cargo build tool (cargo) and the Rust compiler (rustc). You will also get rustup, a command line utility that you can use to install to different compiler versions.

After installing Rust, you should configure your editor or IDE to work with Rust. Most editors do this by talking to rust-analyzer, which provides auto-completion and jump-to-definition functionality for VS Code, Emacs, Vim/Neovim, and many others. There is also a different IDE available called RustRover.

The Rust Ecosystem

The Rust ecosystem consists of a number of tools, of which the main ones are:

• rustc: the Rust compiler which turns .rs files into binaries and other intermediate formats.

• cargo: the Rust dependency manager and build tool. Cargo knows how to download dependencies, usually hosted on https://crates.io, and it will pass them to rustc when building your project. Cargo also comes with a built-in test runner which is used to execute unit tests.

• rustup: the Rust toolchain installer and updater. This tool is used to install and update rustc and cargo when new versions of Rust are released. In addition, rustup can also download documentation for the standard library. You can have multiple versions of Rust installed at once and rustup will let you switch between them as needed.

Code Samples in This Training

For this training, we will mostly explore the Rust language through examples which can be executed through your browser. This makes the setup much easier and ensures a consistent experience for everyone.

Installing Cargo is still encouraged: it will make it easier for you to do the exercises. On the last day, we will do a larger exercise which shows you how to work with dependencies and for that you need Cargo.

The code blocks in this course are fully interactive:

fn main() {
    println!("Edit me!");
}

You can use Ctrl + Enter to execute the code when focus is in the text box.

Running Code Locally with Cargo

If you want to experiment with the code on your own system, then you will need to first install Rust. Do this by following the instructions in the Rust Book. This should give you a working rustc and cargo. At the time of writing, the latest stable Rust release has these version numbers:

% rustc --version
rustc 1.69.0 (84c898d65 2023-04-16)
% cargo --version
cargo 1.69.0 (6e9a83356 2023-04-12)

You can use any later version too since Rust maintains backwards compatibility.

With this in place, follow these steps to build a Rust binary from one of the examples in this training:

1. Click the "Copy to clipboard" button on the example you want to copy.

2. Use cargo new exercise to create a new exercise/ directory for your code:

   $ cargo new exercise
        Created binary (application) `exercise` package
   

3. Navigate into exercise/ and use cargo run to build and run your binary:

   $ cd exercise
   $ cargo run
      Compiling exercise v0.1.0 (/home/mgeisler/tmp/exercise)
       Finished dev [unoptimized + debuginfo] target(s) in 0.75s
        Running `target/debug/exercise`
   Hello, world!
   

4. Replace the boiler-plate code in src/main.rs with your own code. For example, using the example on the previous page, make src/main.rs look like

   fn main() {
       println!("Edit me!");
   }

5. Use cargo run to build and run your updated binary:

   $ cargo run
      Compiling exercise v0.1.0 (/home/mgeisler/tmp/exercise)
       Finished dev [unoptimized + debuginfo] target(s) in 0.24s
        Running `target/debug/exercise`
   Edit me!
   

6. Use cargo check to quickly check your project for errors, use cargo build to compile it without running it. You will find the output in target/debug/ for a normal debug build. Use cargo build --release to produce an optimized release build in target/release/.

7. You can add dependencies for your project by editing Cargo.toml. When you run cargo commands, it will automatically download and compile missing dependencies for you.

Welcome to Day 1

This is the first day of Rust Fundamentals. We will cover a lot of ground today:

• Basic Rust syntax: variables, scalar and compound types, enums, structs, references, functions, and methods.
• Types and type inference.
• Control flow constructs: loops, conditionals, and so on.
• User-defined types: structs and enums.
• Pattern matching: destructuring enums, structs, and arrays.

Schedule

Including 10 minute breaks, this session should take about 2 hours and 5 minutes. It contains:

Hello, World

This segment should take about 15 minutes. It contains:

What is Rust?

Rust is a new programming language which had its 1.0 release in 2015:

• Rust is a statically compiled language in a similar role as C++
  • rustc uses LLVM as its backend.
• Rust supports many platforms and architectures:
  • x86, ARM, WebAssembly, ...
  • Linux, Mac, Windows, ...
• Rust is used for a wide range of devices:
  • firmware and boot loaders,
  • smart displays,
  • mobile phones,
  • desktops,
  • servers.

Benefits of Rust

Some unique selling points of Rust:

• Compile time memory safety - whole classes of memory bugs are prevented at compile time

  • No uninitialized variables.
  • No double-frees.
  • No use-after-free.
  • No NULL pointers.
  • No forgotten locked mutexes.
  • No data races between threads.
  • No iterator invalidation.

• No undefined runtime behavior - what a Rust statement does is never left unspecified

  • Array access is bounds checked.
  • Integer overflow is defined (panic or wrap-around).

• Modern language features - as expressive and ergonomic as higher-level languages

  • Enums and pattern matching.
  • Generics.
  • No overhead FFI.
  • Zero-cost abstractions.
  • Great compiler errors.
  • Built-in dependency manager.
  • Built-in support for testing.
  • Excellent Language Server Protocol support.

Playground

The Rust Playground provides an easy way to run short Rust programs, and is the basis for the examples and exercises in this course. Try running the "hello-world" program it starts with. It comes with a few handy features:

• Under "Tools", use the rustfmt option to format your code in the "standard" way.

• Rust has two main "profiles" for generating code: Debug (extra runtime checks, less optimization) and Release (fewer runtime checks, lots of optimization). These are accessible under "Debug" at the top.

• If you're interested, use "ASM" under "..." to see the generated assembly code.

Types and Values

This segment should take about 40 minutes. It contains:

Hello, World

Let us jump into the simplest possible Rust program, a classic Hello World program:

fn main() {
    println!("Hello 🌍!");
}

What you see:

• Functions are introduced with fn.
• Blocks are delimited by curly braces like in C and C++.
• The main function is the entry point of the program.
• Rust has hygienic macros, println! is an example of this.
• Rust strings are UTF-8 encoded and can contain any Unicode character.

Variables

Rust provides type safety via static typing. Variable bindings are made with let:

fn main() {
    let x: i32 = 10;
    println!("x: {x}");
    // x = 20;
    // println!("x: {x}");
}

Values

Here are some basic built-in types, and the syntax for literal values of each type.

The types have widths as follows:

• iN, uN, and fN are N bits wide,
• isize and usize are the width of a pointer,
• char is 32 bits wide,
• bool is 8 bits wide.

Arithmetic

fn interproduct(a: i32, b: i32, c: i32) -> i32 {
    return a * b + b * c + c * a;
}

fn main() {
    println!("result: {}", interproduct(120, 100, 248));
}

Type Inference

Rust will look at how the variable is used to determine the type:

fn takes_u32(x: u32) {
    println!("u32: {x}");
}

fn takes_i8(y: i8) {
    println!("i8: {y}");
}

fn main() {
    let x = 10;
    let y = 20;

    takes_u32(x);
    takes_i8(y);
    // takes_u32(y);
}

fn main() {
    let x = 3.14;
    let y = 20;
    assert_eq!(x, y);
    // ERROR: no implementation for `{float} == {integer}`
}

Exercise: Fibonacci

The Fibonacci sequence begins with [0,1]. For n>1, the n'th Fibonacci number is calculated recursively as the sum of the n-1'th and n-2'th Fibonacci numbers.

Write a function fib(n) that calculates the n'th Fibonacci number. When will this function panic?

fn fib(n: u32) -> u32 {
    if n < 2 {
        // The base case.
        todo!("Implement this")
    } else {
        // The recursive case.
        todo!("Implement this")
    }
}

fn main() {
    let n = 20;
    println!("fib({n}) = {}", fib(n));
}

Solution

fn fib(n: u32) -> u32 {
    if n < 2 {
        return n;
    } else {
        return fib(n - 1) + fib(n - 2);
    }
}

fn main() {
    let n = 20;
    println!("fib({n}) = {}", fib(n));
}

Control Flow Basics

This segment should take about 40 minutes. It contains:

if expressions

You use if expressions exactly like if statements in other languages:

fn main() {
    let x = 10;
    if x == 0 {
        println!("zero!");
    } else if x < 100 {
        println!("biggish");
    } else {
        println!("huge");
    }
}

In addition, you can use if as an expression. The last expression of each block becomes the value of the if expression:

fn main() {
    let x = 10;
    let size = if x < 20 { "small" } else { "large" };
    println!("number size: {}", size);
}

Loops

There are three looping keywords in Rust: while, loop, and for:

while

The while keyword works much like in other languages, executing the loop body as long as the condition is true.

fn main() {
    let mut x = 200;
    while x >= 10 {
        x = x / 2;
    }
    println!("Final x: {x}");
}

for

The for loop iterates over ranges of values or the items in a collection:

fn main() {
    for x in 1..5 {
        println!("x: {x}");
    }

    for elem in [1, 2, 3, 4, 5] {
        println!("elem: {elem}");
    }
}

loop

The loop statement just loops forever, until a break.

fn main() {
    let mut i = 0;
    loop {
        i += 1;
        println!("{i}");
        if i > 100 {
            break;
        }
    }
}

break and continue

If you want to immediately start the next iteration use continue.

If you want to exit any kind of loop early, use break. With loop, this can take an optional expression that becomes the value of the loop expression.

fn main() {
    let mut i = 0;
    loop {
        i += 1;
        if i > 5 {
            break;
        }
        if i % 2 == 0 {
            continue;
        }
        println!("{}", i);
    }
}

Labels

Both continue and break can optionally take a label argument which is used to break out of nested loops:

fn main() {
    let s = [[5, 6, 7], [8, 9, 10], [21, 15, 32]];
    let mut elements_searched = 0;
    let target_value = 10;
    'outer: for i in 0..=2 {
        for j in 0..=2 {
            elements_searched += 1;
            if s[i][j] == target_value {
                break 'outer;
            }
        }
    }
    print!("elements searched: {elements_searched}");
}

Blocks and Scopes

Blocks

A block in Rust contains a sequence of expressions, enclosed by braces {}. Each block has a value and a type, which are those of the last expression of the block:

fn main() {
    let z = 13;
    let x = {
        let y = 10;
        println!("y: {y}");
        z - y
    };
    println!("x: {x}");
}

If the last expression ends with ;, then the resulting value and type is ().

Scopes and Shadowing

A variable's scope is limited to the enclosing block.

You can shadow variables, both those from outer scopes and variables from the same scope:

fn main() {
    let a = 10;
    println!("before: {a}");
    {
        let a = "hello";
        println!("inner scope: {a}");

        let a = true;
        println!("shadowed in inner scope: {a}");
    }

    println!("after: {a}");
}

Functions

fn gcd(a: u32, b: u32) -> u32 {
    if b > 0 {
        gcd(b, a % b)
    } else {
        a
    }
}

fn main() {
    println!("gcd: {}", gcd(143, 52));
}

Macros

Macros are expanded into Rust code during compilation, and can take a variable number of arguments. They are distinguished by a ! at the end. The Rust standard library includes an assortment of useful macros.

• println!(format, ..) prints a line to standard output, applying formatting described in std::fmt.
• format!(format, ..) works just like println! but returns the result as a string.
• dbg!(expression) logs the value of the expression and returns it.
• todo!() marks a bit of code as not-yet-implemented. If executed, it will panic.
• unreachable!() marks a bit of code as unreachable. If executed, it will panic.

fn factorial(n: u32) -> u32 {
    let mut product = 1;
    for i in 1..=n {
        product *= dbg!(i);
    }
    product
}

fn fizzbuzz(n: u32) -> u32 {
    todo!()
}

fn main() {
    let n = 4;
    println!("{n}! = {}", factorial(n));
}

Exercise: Collatz Sequence

The Collatz Sequence is defined as follows, for an arbitrary n₁ greater than zero:

• If n_i is 1, then the sequence terminates at n_i.
• If n_i is even, then n_i+1 = n_i / 2.
• If n_i is odd, then n_i+1 = 3 * n_i + 1.

For example, beginning with n₁ = 3:

• 3 is odd, so n₂ = 3 * 3 + 1 = 10;
• 10 is even, so n₃ = 10 / 2 = 5;
• 5 is odd, so n₄ = 3 * 5 + 1 = 16;
• 16 is even, so n₅ = 16 / 2 = 8;
• 8 is even, so n₆ = 8 / 2 = 4;
• 4 is even, so n₇ = 4 / 2 = 2;
• 2 is even, so n₈ = 1; and
• the sequence terminates.

Write a function to calculate the length of the collatz sequence for a given initial n.

/// Determine the length of the collatz sequence beginning at `n`.
fn collatz_length(mut n: i32) -> u32 {
  todo!("Implement this")
}

#[test]
fn test_collatz_length() {
    assert_eq!(collatz_length(11), 15);
}

fn main() {
    println!("Length: {}", collatz_length(11));
}

Solution

/// Determine the length of the collatz sequence beginning at `n`.
fn collatz_length(mut n: i32) -> u32 {
    let mut len = 1;
    while n > 1 {
        n = if n % 2 == 0 { n / 2 } else { 3 * n + 1 };
        len += 1;
    }
    len
}

#[test]
fn test_collatz_length() {
    assert_eq!(collatz_length(11), 15);
}

fn main() {
    println!("Length: {}", collatz_length(11));
}

Welcome Back

Including 10 minute breaks, this session should take about 2 hours and 35 minutes. It contains:

Tuples and Arrays

This segment should take about 35 minutes. It contains:

Arrays

fn main() {
    let mut a: [i8; 10] = [42; 10];
    a[5] = 0;
    println!("a: {a:?}");
}

Tuples

fn main() {
    let t: (i8, bool) = (7, true);
    println!("t.0: {}", t.0);
    println!("t.1: {}", t.1);
}

Array Iteration

The for statement supports iterating over arrays (but not tuples).

fn main() {
    let primes = [2, 3, 5, 7, 11, 13, 17, 19];
    for prime in primes {
        for i in 2..prime {
            assert_ne!(prime % i, 0);
        }
    }
}

Patterns and Destructuring

When working with tuples and other structured values it's common to want to extract the inner values into local variables. This can be done manually by directly accessing the inner values:

fn print_tuple(tuple: (i32, i32)) {
    let left = tuple.0;
    let right = tuple.1;
    println!("left: {left}, right: {right}");
}

However, Rust also supports using pattern matching to destructure a larger value into its constituent parts:

fn print_tuple(tuple: (i32, i32)) {
    let (left, right) = tuple;
    println!("left: {left}, right: {right}");
}

Exercise: Nested Arrays

Arrays can contain other arrays:

#![allow(unused)]
fn main() {
let array = [[1, 2, 3], [4, 5, 6], [7, 8, 9]];
}

What is the type of this variable?

Use an array such as the above to write a function transpose which will transpose a matrix (turn rows into columns):

Copy the code below to https://play.rust-lang.org/ and implement the function. This function only operates on 3x3 matrices.

// TODO: remove this when you're done with your implementation.
#![allow(unused_variables, dead_code)]

fn transpose(matrix: [[i32; 3]; 3]) -> [[i32; 3]; 3] {
    unimplemented!()
}

#[test]
fn test_transpose() {
    let matrix = [
        [101, 102, 103], //
        [201, 202, 203],
        [301, 302, 303],
    ];
    let transposed = transpose(matrix);
    assert_eq!(
        transposed,
        [
            [101, 201, 301], //
            [102, 202, 302],
            [103, 203, 303],
        ]
    );
}

fn main() {
    let matrix = [
        [101, 102, 103], // <-- the comment makes rustfmt add a newline
        [201, 202, 203],
        [301, 302, 303],
    ];

    println!("matrix: {:#?}", matrix);
    let transposed = transpose(matrix);
    println!("transposed: {:#?}", transposed);
}

Solution

fn transpose(matrix: [[i32; 3]; 3]) -> [[i32; 3]; 3] {
    let mut result = [[0; 3]; 3];
    for i in 0..3 {
        for j in 0..3 {
            result[j][i] = matrix[i][j];
        }
    }
    result
}

#[test]
fn test_transpose() {
    let matrix = [
        [101, 102, 103], //
        [201, 202, 203],
        [301, 302, 303],
    ];
    let transposed = transpose(matrix);
    assert_eq!(
        transposed,
        [
            [101, 201, 301], //
            [102, 202, 302],
            [103, 203, 303],
        ]
    );
}

fn main() {
    let matrix = [
        [101, 102, 103], // <-- the comment makes rustfmt add a newline
        [201, 202, 203],
        [301, 302, 303],
    ];

    println!("matrix: {:#?}", matrix);
    let transposed = transpose(matrix);
    println!("transposed: {:#?}", transposed);
}

References

This segment should take about 55 minutes. It contains:

Shared References

A reference provides a way to access another value without taking ownership of the value, and is also called "borrowing". Shared references are read-only, and the referenced data cannot change.

fn main() {
    let a = 'A';
    let b = 'B';
    let mut r: &char = &a;
    println!("r: {}", *r);
    r = &b;
    println!("r: {}", *r);
}

A shared reference to a type T has type &T. A reference value is made with the & operator. The * operator "dereferences" a reference, yielding its value.

Rust will statically forbid dangling references:

fn x_axis(x: &i32) -> &(i32, i32) {
    let point = (*x, 0);
    return &point;
}

Exclusive References

Exclusive references, also known as mutable references, allow changing the value they refer to. They have type &mut T.

fn main() {
    let mut point = (1, 2);
    let x_coord = &mut point.0;
    *x_coord = 20;
    println!("point: {point:?}");
}

Slices

A slice gives you a view into a larger collection:

fn main() {
    let mut a: [i32; 6] = [10, 20, 30, 40, 50, 60];
    println!("a: {a:?}");

    let s: &[i32] = &a[2..4];

    println!("s: {s:?}");
}

• Slices borrow data from the sliced type.

Strings

We can now understand the two string types in Rust:

• &str is a slice of UTF-8 encoded bytes, similar to &[u8].
• String is an owned buffer of UTF-8 encoded bytes, similar to Vec<T>.

fn main() {
    let s1: &str = "World";
    println!("s1: {s1}");

    let mut s2: String = String::from("Hello ");
    println!("s2: {s2}");
    s2.push_str(s1);
    println!("s2: {s2}");

    let s3: &str = &s2[s2.len() - s1.len()..];
    println!("s3: {s3}");
}

Exercise: Geometry

We will create a few utility functions for 3-dimensional geometry, representing a point as [f64;3]. It is up to you to determine the function signatures.

// Calculate the magnitude of a vector by summing the squares of its coordinates
// and taking the square root. Use the `sqrt()` method to calculate the square
// root, like `v.sqrt()`.


fn magnitude(...) -> f64 {
    todo!()
}

// Normalize a vector by calculating its magnitude and dividing all of its
// coordinates by that magnitude.


fn normalize(...) {
    todo!()
}

// Use the following `main` to test your work.

fn main() {
    println!("Magnitude of a unit vector: {}", magnitude(&[0.0, 1.0, 0.0]));

    let mut v = [1.0, 2.0, 9.0];
    println!("Magnitude of {v:?}: {}", magnitude(&v));
    normalize(&mut v);
    println!("Magnitude of {v:?} after normalization: {}", magnitude(&v));
}

Solution

/// Calculate the magnitude of the given vector.
fn magnitude(vector: &[f64; 3]) -> f64 {
    let mut mag_squared = 0.0;
    for coord in vector {
        mag_squared += coord * coord;
    }
    mag_squared.sqrt()
}

/// Change the magnitude of the vector to 1.0 without changing its direction.
fn normalize(vector: &mut [f64; 3]) {
    let mag = magnitude(vector);
    for item in vector {
        *item /= mag;
    }
}

fn main() {
    println!("Magnitude of a unit vector: {}", magnitude(&[0.0, 1.0, 0.0]));

    let mut v = [1.0, 2.0, 9.0];
    println!("Magnitude of {v:?}: {}", magnitude(&v));
    normalize(&mut v);
    println!("Magnitude of {v:?} after normalization: {}", magnitude(&v));
}

User-Defined Types

This segment should take about 50 minutes. It contains:

Named Structs

Like C and C++, Rust has support for custom structs:

struct Person {
    name: String,
    age: u8,
}

fn describe(person: &Person) {
    println!("{} is {} years old", person.name, person.age);
}

fn main() {
    let mut peter = Person { name: String::from("Peter"), age: 27 };
    describe(&peter);

    peter.age = 28;
    describe(&peter);

    let name = String::from("Avery");
    let age = 39;
    let avery = Person { name, age };
    describe(&avery);

    let jackie = Person { name: String::from("Jackie"), ..avery };
    describe(&jackie);
}

Tuple Structs

If the field names are unimportant, you can use a tuple struct:

struct Point(i32, i32);

fn main() {
    let p = Point(17, 23);
    println!("({}, {})", p.0, p.1);
}

This is often used for single-field wrappers (called newtypes):

struct PoundsOfForce(f64);
struct Newtons(f64);

fn compute_thruster_force() -> PoundsOfForce {
    todo!("Ask a rocket scientist at NASA")
}

fn set_thruster_force(force: Newtons) {
    // ...
}

fn main() {
    let force = compute_thruster_force();
    set_thruster_force(force);
}

Enums

The enum keyword allows the creation of a type which has a few different variants:

#[derive(Debug)]
enum Direction {
    Left,
    Right,
}

#[derive(Debug)]
enum PlayerMove {
    Pass,                        // Simple variant
    Run(Direction),              // Tuple variant
    Teleport { x: u32, y: u32 }, // Struct variant
}

fn main() {
    let player_move: PlayerMove = PlayerMove::Run(Direction::Left);
    println!("On this turn: {player_move:?}");
}

const

Constants are evaluated at compile time and their values are inlined wherever they are used:

const DIGEST_SIZE: usize = 3;
const ZERO: Option<u8> = Some(42);

fn compute_digest(text: &str) -> [u8; DIGEST_SIZE] {
    let mut digest = [ZERO.unwrap_or(0); DIGEST_SIZE];
    for (idx, &b) in text.as_bytes().iter().enumerate() {
        digest[idx % DIGEST_SIZE] = digest[idx % DIGEST_SIZE].wrapping_add(b);
    }
    digest
}

fn main() {
    let digest = compute_digest("Hello");
    println!("digest: {digest:?}");
}

According to the Rust RFC Book these are inlined upon use.

Only functions marked const can be called at compile time to generate const values. const functions can however be called at runtime.

static

Static variables will live during the whole execution of the program, and therefore will not move:

static BANNER: &str = "Welcome to RustOS 3.14";

fn main() {
    println!("{BANNER}");
}

As noted in the Rust RFC Book, these are not inlined upon use and have an actual associated memory location. This is useful for unsafe and embedded code, and the variable lives through the entirety of the program execution. When a globally-scoped value does not have a reason to need object identity, const is generally preferred.

Type Aliases

A type alias creates a name for another type. The two types can be used interchangeably.

enum CarryableConcreteItem {
    Left,
    Right,
}

type Item = CarryableConcreteItem;

// Aliases are more useful with long, complex types:
use std::cell::RefCell;
use std::sync::{Arc, RwLock};
type PlayerInventory = RwLock<Vec<Arc<RefCell<Item>>>>;

Exercise: Elevator Events

We will create a data structure to represent an event in an elevator control system. It is up to you to define the types and functions to construct various events. Use #[derive(Debug)] to allow the types to be formatted with {:?}.

This exercise only requires creating and populating data structures so that main runs without errors. The next part of the course will cover getting data out of these structures.

#[derive(Debug)]
/// An event in the elevator system that the controller must react to.
enum Event {
    // TODO: add required variants
}

/// A direction of travel.
#[derive(Debug)]
enum Direction {
    Up,
    Down,
}

/// The car has arrived on the given floor.
fn car_arrived(floor: i32) -> Event {
    todo!()
}

/// The car doors have opened.
fn car_door_opened() -> Event {
    todo!()
}

/// The car doors have closed.
fn car_door_closed() -> Event {
    todo!()
}

/// A directional button was pressed in an elevator lobby on the given floor.
fn lobby_call_button_pressed(floor: i32, dir: Direction) -> Event {
    todo!()
}

/// A floor button was pressed in the elevator car.
fn car_floor_button_pressed(floor: i32) -> Event {
    todo!()
}

fn main() {
    println!(
        "A ground floor passenger has pressed the up button: {:?}",
        lobby_call_button_pressed(0, Direction::Up)
    );
    println!("The car has arrived on the ground floor: {:?}", car_arrived(0));
    println!("The car door opened: {:?}", car_door_opened());
    println!(
        "A passenger has pressed the 3rd floor button: {:?}",
        car_floor_button_pressed(3)
    );
    println!("The car door closed: {:?}", car_door_closed());
    println!("The car has arrived on the 3rd floor: {:?}", car_arrived(3));
}

Solution

#[derive(Debug)]
/// An event in the elevator system that the controller must react to.
enum Event {
    /// A button was pressed.
    ButtonPressed(Button),

    /// The car has arrived at the given floor.
    CarArrived(Floor),

    /// The car's doors have opened.
    CarDoorOpened,

    /// The car's doors have closed.
    CarDoorClosed,
}

/// A floor is represented as an integer.
type Floor = i32;

/// A direction of travel.
#[derive(Debug)]
enum Direction {
    Up,
    Down,
}

/// A user-accessible button.
#[derive(Debug)]
enum Button {
    /// A button in the elevator lobby on the given floor.
    LobbyCall(Direction, Floor),

    /// A floor button within the car.
    CarFloor(Floor),
}

/// The car has arrived on the given floor.
fn car_arrived(floor: i32) -> Event {
    Event::CarArrived(floor)
}

/// The car doors have opened.
fn car_door_opened() -> Event {
    Event::CarDoorOpened
}

/// The car doors have closed.
fn car_door_closed() -> Event {
    Event::CarDoorClosed
}

/// A directional button was pressed in an elevator lobby on the given floor.
fn lobby_call_button_pressed(floor: i32, dir: Direction) -> Event {
    Event::ButtonPressed(Button::LobbyCall(dir, floor))
}

/// A floor button was pressed in the elevator car.
fn car_floor_button_pressed(floor: i32) -> Event {
    Event::ButtonPressed(Button::CarFloor(floor))
}

fn main() {
    println!(
        "A ground floor passenger has pressed the up button: {:?}",
        lobby_call_button_pressed(0, Direction::Up)
    );
    println!("The car has arrived on the ground floor: {:?}", car_arrived(0));
    println!("The car door opened: {:?}", car_door_opened());
    println!(
        "A passenger has pressed the 3rd floor button: {:?}",
        car_floor_button_pressed(3)
    );
    println!("The car door closed: {:?}", car_door_closed());
    println!("The car has arrived on the 3rd floor: {:?}", car_arrived(3));
}

Welcome to Day 2

Now that we have seen a fair amount of Rust, today will focus on Rust's type system:

• Pattern matching: extracting data from structures.
• Methods: associating functions with types.
• Traits: behaviors shared by multiple types.
• Generics: parameterizing types on other types.
• Standard library types and traits: a tour of Rust's rich standard library.

Schedule

Including 10 minute breaks, this session should take about 2 hours and 10 minutes. It contains:

Pattern Matching

This segment should take about 1 hour. It contains:

Matching Values

The match keyword lets you match a value against one or more patterns. The comparisons are done from top to bottom and the first match wins.

The patterns can be simple values, similarly to switch in C and C++:

#[rustfmt::skip]
fn main() {
    let input = 'x';
    match input {
        'q'                       => println!("Quitting"),
        'a' | 's' | 'w' | 'd'     => println!("Moving around"),
        '0'..='9'                 => println!("Number input"),
        key if key.is_lowercase() => println!("Lowercase: {key}"),
        _                         => println!("Something else"),
    }
}

The _ pattern is a wildcard pattern which matches any value. The expressions must be exhaustive, meaning that it covers every possibility, so _ is often used as the final catch-all case.

Match can be used as an expression. Just like if, each match arm must have the same type. The type is the last expression of the block, if any. In the example above, the type is ().

A variable in the pattern (key in this example) will create a binding that can be used within the match arm.

A match guard causes the arm to match only if the condition is true.

Structs

Like tuples, Struct can also be destructured by matching:

struct Foo {
    x: (u32, u32),
    y: u32,
}

#[rustfmt::skip]
fn main() {
    let foo = Foo { x: (1, 2), y: 3 };
    match foo {
        Foo { x: (1, b), y } => println!("x.0 = 1, b = {b}, y = {y}"),
        Foo { y: 2, x: i }   => println!("y = 2, x = {i:?}"),
        Foo { y, .. }        => println!("y = {y}, other fields were ignored"),
    }
}

Enums

Like tuples, enums can also be destructured by matching:

Patterns can also be used to bind variables to parts of your values. This is how you inspect the structure of your types. Let us start with a simple enum type:

enum Result {
    Ok(i32),
    Err(String),
}

fn divide_in_two(n: i32) -> Result {
    if n % 2 == 0 {
        Result::Ok(n / 2)
    } else {
        Result::Err(format!("cannot divide {n} into two equal parts"))
    }
}

fn main() {
    let n = 100;
    match divide_in_two(n) {
        Result::Ok(half) => println!("{n} divided in two is {half}"),
        Result::Err(msg) => println!("sorry, an error happened: {msg}"),
    }
}

Here we have used the arms to destructure the Result value. In the first arm, half is bound to the value inside the Ok variant. In the second arm, msg is bound to the error message.

Let Control Flow

Rust has a few control flow constructs which differ from other languages. They are used for pattern matching:

• if let expressions
• let else expressions
• while let expressions

if let expressions

The if let expression lets you execute different code depending on whether a value matches a pattern:

use std::time::Duration;

fn sleep_for(secs: f32) {
    if let Ok(duration) = Duration::try_from_secs_f32(secs) {
        std::thread::sleep(duration);
        println!("slept for {duration:?}");
    }
}

fn main() {
    sleep_for(-10.0);
    sleep_for(0.8);
}

let else expressions

For the common case of matching a pattern and returning from the function, use let else. The "else" case must diverge (return, break, or panic - anything but falling off the end of the block).

fn hex_or_die_trying(maybe_string: Option<String>) -> Result<u32, String> {
    if let Some(s) = maybe_string {
        if let Some(first_byte_char) = s.chars().next() {
            if let Some(digit) = first_byte_char.to_digit(16) {
                Ok(digit)
            } else {
                return Err(String::from("not a hex digit"));
            }
        } else {
            return Err(String::from("got empty string"));
        }
    } else {
        return Err(String::from("got None"));
    }
}

fn main() {
    println!("result: {:?}", hex_or_die_trying(Some(String::from("foo"))));
}

Like with if let, there is a while let variant which repeatedly tests a value against a pattern:

fn main() {
    let mut name = String::from("Comprehensive Rust 🦀");
    while let Some(c) = name.pop() {
        println!("character: {c}");
    }
    // (There are more efficient ways to reverse a string!)
}

Here String::pop returns Some(c) until the string is empty, after which it will return None. The while let lets us keep iterating through all items.

#![allow(unused)]
fn main() {
fn hex_or_die_trying(maybe_string: Option<String>) -> Result<u32, String> {
    let Some(s) = maybe_string else {
        return Err(String::from("got None"));
    };

    let Some(first_byte_char) = s.chars().next() else {
        return Err(String::from("got empty string"));
    };

    let Some(digit) = first_byte_char.to_digit(16) else {
        return Err(String::from("not a hex digit"));
    };

    return Ok(digit);
}
}

Exercise: Expression Evaluation

Let's write a simple recursive evaluator for arithmetic expressions.

An example of a small arithmetic expression could be 10 + 20, which evaluates to 30. We can represent the expression as a tree:

A bigger and more complex expression would be (10 * 9) + ((3 - 4) * 5), which evaluate to 85. We represent this as a much bigger tree:

In code, we will represent the tree with two types:

#![allow(unused)]
fn main() {
/// An operation to perform on two subexpressions.
#[derive(Debug)]
enum Operation {
    Add,
    Sub,
    Mul,
    Div,
}

/// An expression, in tree form.
#[derive(Debug)]
enum Expression {
    /// An operation on two subexpressions.
    Op { op: Operation, left: Box<Expression>, right: Box<Expression> },

    /// A literal value
    Value(i64),
}
}

The Box type here is a smart pointer, and will be covered in detail later in the course. An expression can be "boxed" with Box::new as seen in the tests. To evaluate a boxed expression, use the deref operator (*) to "unbox" it: eval(*boxed_expr).

Some expressions cannot be evaluated and will return an error. The standard Result<Value, String> type is an enum that represents either a successful value (Ok(Value)) or an error (Err(String)). We will cover this type in detail later.

Copy and paste the code into the Rust playground, and begin implementing eval. The final product should pass the tests. It may be helpful to use todo!() and get the tests to pass one-by-one. You can also skip a test temporarily with #[ignore]:

#[test]
#[ignore]
fn test_value() { .. }

#![allow(unused)]
fn main() {
/// An operation to perform on two subexpressions.
#[derive(Debug)]
enum Operation {
    Add,
    Sub,
    Mul,
    Div,
}

/// An expression, in tree form.
#[derive(Debug)]
enum Expression {
    /// An operation on two subexpressions.
    Op { op: Operation, left: Box<Expression>, right: Box<Expression> },

    /// A literal value
    Value(i64),
}

fn eval(e: Expression) -> Result<i64, String> {
    todo!()
}

#[test]
fn test_value() {
    assert_eq!(eval(Expression::Value(19)), Ok(19));
}

#[test]
fn test_sum() {
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Add,
            left: Box::new(Expression::Value(10)),
            right: Box::new(Expression::Value(20)),
        }),
        Ok(30)
    );
}

#[test]
fn test_recursion() {
    let term1 = Expression::Op {
        op: Operation::Mul,
        left: Box::new(Expression::Value(10)),
        right: Box::new(Expression::Value(9)),
    };
    let term2 = Expression::Op {
        op: Operation::Mul,
        left: Box::new(Expression::Op {
            op: Operation::Sub,
            left: Box::new(Expression::Value(3)),
            right: Box::new(Expression::Value(4)),
        }),
        right: Box::new(Expression::Value(5)),
    };
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Add,
            left: Box::new(term1),
            right: Box::new(term2),
        }),
        Ok(85)
    );
}

#[test]
fn test_zeros() {
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Add,
            left: Box::new(Expression::Value(0)),
            right: Box::new(Expression::Value(0))
        }),
        Ok(0)
    );
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Mul,
            left: Box::new(Expression::Value(0)),
            right: Box::new(Expression::Value(0))
        }),
        Ok(0)
    );
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Sub,
            left: Box::new(Expression::Value(0)),
            right: Box::new(Expression::Value(0))
        }),
        Ok(0)
    );
}

#[test]
fn test_error() {
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Div,
            left: Box::new(Expression::Value(99)),
            right: Box::new(Expression::Value(0)),
        }),
        Err(String::from("division by zero"))
    );
}
}

Solution

/// An operation to perform on two subexpressions.
#[derive(Debug)]
enum Operation {
    Add,
    Sub,
    Mul,
    Div,
}

/// An expression, in tree form.
#[derive(Debug)]
enum Expression {
    /// An operation on two subexpressions.
    Op { op: Operation, left: Box<Expression>, right: Box<Expression> },

    /// A literal value
    Value(i64),
}

fn eval(e: Expression) -> Result<i64, String> {
    match e {
        Expression::Op { op, left, right } => {
            let left = match eval(*left) {
                Ok(v) => v,
                Err(e) => return Err(e),
            };
            let right = match eval(*right) {
                Ok(v) => v,
                Err(e) => return Err(e),
            };
            Ok(match op {
                Operation::Add => left + right,
                Operation::Sub => left - right,
                Operation::Mul => left * right,
                Operation::Div => {
                    if right == 0 {
                        return Err(String::from("division by zero"));
                    } else {
                        left / right
                    }
                }
            })
        }
        Expression::Value(v) => Ok(v),
    }
}

#[test]
fn test_value() {
    assert_eq!(eval(Expression::Value(19)), Ok(19));
}

#[test]
fn test_sum() {
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Add,
            left: Box::new(Expression::Value(10)),
            right: Box::new(Expression::Value(20)),
        }),
        Ok(30)
    );
}

#[test]
fn test_recursion() {
    let term1 = Expression::Op {
        op: Operation::Mul,
        left: Box::new(Expression::Value(10)),
        right: Box::new(Expression::Value(9)),
    };
    let term2 = Expression::Op {
        op: Operation::Mul,
        left: Box::new(Expression::Op {
            op: Operation::Sub,
            left: Box::new(Expression::Value(3)),
            right: Box::new(Expression::Value(4)),
        }),
        right: Box::new(Expression::Value(5)),
    };
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Add,
            left: Box::new(term1),
            right: Box::new(term2),
        }),
        Ok(85)
    );
}

#[test]
fn test_zeros() {
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Add,
            left: Box::new(Expression::Value(0)),
            right: Box::new(Expression::Value(0))
        }),
        Ok(0)
    );
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Mul,
            left: Box::new(Expression::Value(0)),
            right: Box::new(Expression::Value(0))
        }),
        Ok(0)
    );
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Sub,
            left: Box::new(Expression::Value(0)),
            right: Box::new(Expression::Value(0))
        }),
        Ok(0)
    );
}

#[test]
fn test_error() {
    assert_eq!(
        eval(Expression::Op {
            op: Operation::Div,
            left: Box::new(Expression::Value(99)),
            right: Box::new(Expression::Value(0)),
        }),
        Err(String::from("division by zero"))
    );
}

fn main() {
    let expr = Expression::Op {
        op: Operation::Sub,
        left: Box::new(Expression::Value(20)),
        right: Box::new(Expression::Value(10)),
    };
    println!("expr: {expr:?}");
    println!("result: {:?}", eval(expr));
}

Methods and Traits

This segment should take about 50 minutes. It contains:

Methods

Rust allows you to associate functions with your new types. You do this with an impl block:

#[derive(Debug)]
struct Race {
    name: String,
    laps: Vec<i32>,
}

impl Race {
    // No receiver, a static method
    fn new(name: &str) -> Self {
        Self { name: String::from(name), laps: Vec::new() }
    }

    // Exclusive borrowed read-write access to self
    fn add_lap(&mut self, lap: i32) {
        self.laps.push(lap);
    }

    // Shared and read-only borrowed access to self
    fn print_laps(&self) {
        println!("Recorded {} laps for {}:", self.laps.len(), self.name);
        for (idx, lap) in self.laps.iter().enumerate() {
            println!("Lap {idx}: {lap} sec");
        }
    }

    // Exclusive ownership of self (covered later)
    fn finish(self) {
        let total: i32 = self.laps.iter().sum();
        println!("Race {} is finished, total lap time: {}", self.name, total);
    }
}

fn main() {
    let mut race = Race::new("Monaco Grand Prix");
    race.add_lap(70);
    race.add_lap(68);
    race.print_laps();
    race.add_lap(71);
    race.print_laps();
    race.finish();
    // race.add_lap(42);
}

The self arguments specify the "receiver" - the object the method acts on. There are several common receivers for a method:

• &self: borrows the object from the caller using a shared and immutable reference. The object can be used again afterwards.
• &mut self: borrows the object from the caller using a unique and mutable reference. The object can be used again afterwards.
• self: takes ownership of the object and moves it away from the caller. The method becomes the owner of the object. The object will be dropped (deallocated) when the method returns, unless its ownership is explicitly transmitted. Complete ownership does not automatically mean mutability.
• mut self: same as above, but the method can mutate the object.
• No receiver: this becomes a static method on the struct. Typically used to create constructors which are called new by convention.

Traits

Rust lets you abstract over types with traits. They're similar to interfaces:

trait Pet {
    /// Return a sentence from this pet.
    fn talk(&self) -> String;

    /// Print a string to the terminal greeting this pet.
    fn greet(&self);
}

Implementing Traits

trait Pet {
    fn talk(&self) -> String;

    fn greet(&self) {
        println!("Oh you're a cutie! What's your name? {}", self.talk());
    }
}

struct Dog {
    name: String,
    age: i8,
}

impl Pet for Dog {
    fn talk(&self) -> String {
        format!("Woof, my name is {}!", self.name)
    }
}

fn main() {
    let fido = Dog { name: String::from("Fido"), age: 5 };
    fido.greet();
}

Supertraits

A trait can require that types implementing it also implement other traits, called supertraits. Here, any type implementing Pet must implement Animal.

trait Animal {
    fn leg_count(&self) -> u32;
}

trait Pet: Animal {
    fn name(&self) -> String;
}

struct Dog(String);

impl Animal for Dog {
    fn leg_count(&self) -> u32 {
        4
    }
}

impl Pet for Dog {
    fn name(&self) -> String {
        self.0.clone()
    }
}

fn main() {
    let puppy = Dog(String::from("Rex"));
    println!("{} has {} legs", puppy.name(), puppy.leg_count());
}

Associated Types

Associated types are placeholder types which are supplied by the trait implementation.

#[derive(Debug)]
struct Meters(i32);
#[derive(Debug)]
struct MetersSquared(i32);

trait Multiply {
    type Output;
    fn multiply(&self, other: &Self) -> Self::Output;
}

impl Multiply for Meters {
    type Output = MetersSquared;
    fn multiply(&self, other: &Self) -> Self::Output {
        MetersSquared(self.0 * other.0)
    }
}

fn main() {
    println!("{:?}", Meters(10).multiply(&Meters(20)));
}

Deriving

Supported traits can be automatically implemented for your custom types, as follows:

#[derive(Debug, Clone, Default)]
struct Player {
    name: String,
    strength: u8,
    hit_points: u8,
}

fn main() {
    let p1 = Player::default(); // Default trait adds `default` constructor.
    let mut p2 = p1.clone(); // Clone trait adds `clone` method.
    p2.name = String::from("EldurScrollz");
    // Debug trait adds support for printing with `{:?}`.
    println!("{p1:?} vs. {p2:?}");
}

Exercise: Logger Trait

Let's design a simple logging utility, using a trait Logger with a log method. Code which might log its progress can then take an &impl Logger. In testing, this might put messages in the test logfile, while in a production build it would send messages to a log server.

However, the StdoutLogger given below logs all messages, regardless of verbosity. Your task is to write a VerbosityFilter type that will ignore messages above a maximum verbosity.

This is a common pattern: a struct wrapping a trait implementation and implementing that same trait, adding behavior in the process. What other kinds of wrappers might be useful in a logging utility?

pub trait Logger {
    /// Log a message at the given verbosity level.
    fn log(&self, verbosity: u8, message: &str);
}

struct StdoutLogger;

impl Logger for StdoutLogger {
    fn log(&self, verbosity: u8, message: &str) {
        println!("verbosity={verbosity}: {message}");
    }
}

// TODO: Define and implement `VerbosityFilter`.

fn main() {
    let logger = VerbosityFilter { max_verbosity: 3, inner: StdoutLogger };
    logger.log(5, "FYI");
    logger.log(2, "Uhoh");
}

Solution

pub trait Logger {
    /// Log a message at the given verbosity level.
    fn log(&self, verbosity: u8, message: &str);
}

struct StdoutLogger;

impl Logger for StdoutLogger {
    fn log(&self, verbosity: u8, message: &str) {
        println!("verbosity={verbosity}: {message}");
    }
}

/// Only log messages up to the given verbosity level.
struct VerbosityFilter {
    max_verbosity: u8,
    inner: StdoutLogger,
}

impl Logger for VerbosityFilter {
    fn log(&self, verbosity: u8, message: &str) {
        if verbosity <= self.max_verbosity {
            self.inner.log(verbosity, message);
        }
    }
}

fn main() {
    let logger = VerbosityFilter { max_verbosity: 3, inner: StdoutLogger };
    logger.log(5, "FYI");
    logger.log(2, "Uhoh");
}

Welcome Back

Including 10 minute breaks, this session should take about 3 hours and 15 minutes. It contains:

Generics

This segment should take about 45 minutes. It contains:

Generic Functions

Rust supports generics, which lets you abstract algorithms or data structures (such as sorting or a binary tree) over the types used or stored.

/// Pick `even` or `odd` depending on the value of `n`.
fn pick<T>(n: i32, even: T, odd: T) -> T {
    if n % 2 == 0 {
        even
    } else {
        odd
    }
}

fn main() {
    println!("picked a number: {:?}", pick(97, 222, 333));
    println!("picked a string: {:?}", pick(28, "dog", "cat"));
}

Generic Data Types

You can use generics to abstract over the concrete field type:

#[derive(Debug)]
struct Point<T> {
    x: T,
    y: T,
}

impl<T> Point<T> {
    fn coords(&self) -> (&T, &T) {
        (&self.x, &self.y)
    }

    fn set_x(&mut self, x: T) {
        self.x = x;
    }
}

fn main() {
    let integer = Point { x: 5, y: 10 };
    let float = Point { x: 1.0, y: 4.0 };
    println!("{integer:?} and {float:?}");
    println!("coords: {:?}", integer.coords());
}

Generic Traits

Traits can also be generic, just like types and functions. A trait's parameters get concrete types when it is used.

#[derive(Debug)]
struct Foo(String);

impl From<u32> for Foo {
    fn from(from: u32) -> Foo {
        Foo(format!("Converted from integer: {from}"))
    }
}

impl From<bool> for Foo {
    fn from(from: bool) -> Foo {
        Foo(format!("Converted from bool: {from}"))
    }
}

fn main() {
    let from_int = Foo::from(123);
    let from_bool = Foo::from(true);
    println!("{from_int:?}, {from_bool:?}");
}

Trait Bounds

When working with generics, you often want to require the types to implement some trait, so that you can call this trait's methods.

You can do this with T: Trait:

fn duplicate<T: Clone>(a: T) -> (T, T) {
    (a.clone(), a.clone())
}

// struct NotCloneable;

fn main() {
    let foo = String::from("foo");
    let pair = duplicate(foo);
    println!("{pair:?}");
}

impl Trait

Similar to trait bounds, an impl Trait syntax can be used in function arguments and return values:

// Syntactic sugar for:
//   fn add_42_millions<T: Into<i32>>(x: T) -> i32 {
fn add_42_millions(x: impl Into<i32>) -> i32 {
    x.into() + 42_000_000
}

fn pair_of(x: u32) -> impl std::fmt::Debug {
    (x + 1, x - 1)
}

fn main() {
    let many = add_42_millions(42_i8);
    println!("{many}");
    let many_more = add_42_millions(10_000_000);
    println!("{many_more}");
    let debuggable = pair_of(27);
    println!("debuggable: {debuggable:?}");
}

dyn Trait

In addition to using traits for static dispatch via generics, Rust also supports using them for type-erased, dynamic dispatch via trait objects:

struct Dog {
    name: String,
    age: i8,
}
struct Cat {
    lives: i8,
}

trait Pet {
    fn talk(&self) -> String;
}

impl Pet for Dog {
    fn talk(&self) -> String {
        format!("Woof, my name is {}!", self.name)
    }
}

impl Pet for Cat {
    fn talk(&self) -> String {
        String::from("Miau!")
    }
}

// Uses generics and static dispatch.
fn generic(pet: &impl Pet) {
    println!("Hello, who are you? {}", pet.talk());
}

// Uses type-erasure and dynamic dispatch.
fn dynamic(pet: &dyn Pet) {
    println!("Hello, who are you? {}", pet.talk());
}

fn main() {
    let cat = Cat { lives: 9 };
    let dog = Dog { name: String::from("Fido"), age: 5 };

    generic(&cat);
    generic(&dog);

    dynamic(&cat);
    dynamic(&dog);
}

Exercise: Generic min

In this short exercise, you will implement a generic min function that determines the minimum of two values, using the Ord trait.

use std::cmp::Ordering;

// TODO: implement the `min` function used in `main`.

fn main() {
    assert_eq!(min(0, 10), 0);
    assert_eq!(min(500, 123), 123);

    assert_eq!(min('a', 'z'), 'a');
    assert_eq!(min('7', '1'), '1');

    assert_eq!(min("hello", "goodbye"), "goodbye");
    assert_eq!(min("bat", "armadillo"), "armadillo");
}

Solution

use std::cmp::Ordering;

fn min<T: Ord>(l: T, r: T) -> T {
    match l.cmp(&r) {
        Ordering::Less | Ordering::Equal => l,
        Ordering::Greater => r,
    }
}

fn main() {
    assert_eq!(min(0, 10), 0);
    assert_eq!(min(500, 123), 123);

    assert_eq!(min('a', 'z'), 'a');
    assert_eq!(min('7', '1'), '1');

    assert_eq!(min("hello", "goodbye"), "goodbye");
    assert_eq!(min("bat", "armadillo"), "armadillo");
}

Standard Library Types

This segment should take about 1 hour. It contains:

Standard Library

Rust comes with a standard library which helps establish a set of common types used by Rust libraries and programs. This way, two libraries can work together smoothly because they both use the same String type.

In fact, Rust contains several layers of the Standard Library: core, alloc and std.

• core includes the most basic types and functions that don't depend on libc, allocator or even the presence of an operating system.
• alloc includes types which require a global heap allocator, such as Vec, Box and Arc.
• Embedded Rust applications often only use core, and sometimes alloc.

Documentation

Rust comes with extensive documentation. For example:

• All of the details about loops.
• Primitive types like u8.
• Standard library types like Option or BinaryHeap.

Use rustup doc --std or https://std.rs to view the documentation.

In fact, you can document your own code:

/// Determine whether the first argument is divisible by the second argument.
///
/// If the second argument is zero, the result is false.
fn is_divisible_by(lhs: u32, rhs: u32) -> bool {
    if rhs == 0 {
        return false;
    }
    lhs % rhs == 0
}

The contents are treated as Markdown. All published Rust library crates are automatically documented at docs.rs using the rustdoc tool. It is idiomatic to document all public items in an API using this pattern.

To document an item from inside the item (such as inside a module), use //! or /*! .. */, called "inner doc comments":

//! This module contains functionality relating to divisibility of integers.

Option

We have already seen some use of Option<T>. It stores either a value of type T or nothing. For example, String::find returns an Option<usize>.

fn main() {
    let name = "Löwe 老虎 Léopard Gepardi";
    let mut position: Option<usize> = name.find('é');
    println!("find returned {position:?}");
    assert_eq!(position.unwrap(), 14);
    position = name.find('Z');
    println!("find returned {position:?}");
    assert_eq!(position.expect("Character not found"), 0);
}

Result

Result is similar to Option, but indicates the success or failure of an operation, each with a different enum variant. It is generic: Result<T, E> where T is used in the Ok variant and E appears in the Err variant.

use std::fs::File;
use std::io::Read;

fn main() {
    let file: Result<File, std::io::Error> = File::open("diary.txt");
    match file {
        Ok(mut file) => {
            let mut contents = String::new();
            if let Ok(bytes) = file.read_to_string(&mut contents) {
                println!("Dear diary: {contents} ({bytes} bytes)");
            } else {
                println!("Could not read file content");
            }
        }
        Err(err) => {
            println!("The diary could not be opened: {err}");
        }
    }
}

String

String is a growable UTF-8 encoded string:

fn main() {
    let mut s1 = String::new();
    s1.push_str("Hello");
    println!("s1: len = {}, capacity = {}", s1.len(), s1.capacity());

    let mut s2 = String::with_capacity(s1.len() + 1);
    s2.push_str(&s1);
    s2.push('!');
    println!("s2: len = {}, capacity = {}", s2.len(), s2.capacity());

    let s3 = String::from("🇨🇭");
    println!("s3: len = {}, number of chars = {}", s3.len(), s3.chars().count());
}

String implements Deref<Target = str>, which means that you can call all str methods on a String.

Vec

Vec is the standard resizable heap-allocated buffer:

fn main() {
    let mut v1 = Vec::new();
    v1.push(42);
    println!("v1: len = {}, capacity = {}", v1.len(), v1.capacity());

    let mut v2 = Vec::with_capacity(v1.len() + 1);
    v2.extend(v1.iter());
    v2.push(9999);
    println!("v2: len = {}, capacity = {}", v2.len(), v2.capacity());

    // Canonical macro to initialize a vector with elements.
    let mut v3 = vec![0, 0, 1, 2, 3, 4];

    // Retain only the even elements.
    v3.retain(|x| x % 2 == 0);
    println!("{v3:?}");

    // Remove consecutive duplicates.
    v3.dedup();
    println!("{v3:?}");
}

Vec implements Deref<Target = [T]>, which means that you can call slice methods on a Vec.

HashMap

Standard hash map with protection against HashDoS attacks:

use std::collections::HashMap;

fn main() {
    let mut page_counts = HashMap::new();
    page_counts.insert("Adventures of Huckleberry Finn", 207);
    page_counts.insert("Grimms' Fairy Tales", 751);
    page_counts.insert("Pride and Prejudice", 303);

    if !page_counts.contains_key("Les Misérables") {
        println!(
            "We know about {} books, but not Les Misérables.",
            page_counts.len()
        );
    }

    for book in ["Pride and Prejudice", "Alice's Adventure in Wonderland"] {
        match page_counts.get(book) {
            Some(count) => println!("{book}: {count} pages"),
            None => println!("{book} is unknown."),
        }
    }

    // Use the .entry() method to insert a value if nothing is found.
    for book in ["Pride and Prejudice", "Alice's Adventure in Wonderland"] {
        let page_count: &mut i32 = page_counts.entry(book).or_insert(0);
        *page_count += 1;
    }

    println!("{page_counts:#?}");
}

Exercise: Counter

In this exercise you will take a very simple data structure and make it generic. It uses a std::collections::HashMap to keep track of which values have been seen and how many times each one has appeared.

The initial version of Counter is hard coded to only work for u32 values. Make the struct and its methods generic over the type of value being tracked, that way Counter can track any type of value.

If you finish early, try using the entry method to halve the number of hash lookups required to implement the count method.

use std::collections::HashMap;

/// Counter counts the number of times each value of type T has been seen.
struct Counter {
    values: HashMap<u32, u64>,
}

impl Counter {
    /// Create a new Counter.
    fn new() -> Self {
        Counter {
            values: HashMap::new(),
        }
    }

    /// Count an occurrence of the given value.
    fn count(&mut self, value: u32) {
        if self.values.contains_key(&value) {
            *self.values.get_mut(&value).unwrap() += 1;
        } else {
            self.values.insert(value, 1);
        }
    }

    /// Return the number of times the given value has been seen.
    fn times_seen(&self, value: u32) -> u64 {
        self.values.get(&value).copied().unwrap_or_default()
    }
}

fn main() {
    let mut ctr = Counter::new();
    ctr.count(13);
    ctr.count(14);
    ctr.count(16);
    ctr.count(14);
    ctr.count(14);
    ctr.count(11);

    for i in 10..20 {
        println!("saw {} values equal to {}", ctr.times_seen(i), i);
    }

    let mut strctr = Counter::new();
    strctr.count("apple");
    strctr.count("orange");
    strctr.count("apple");
    println!("got {} apples", strctr.times_seen("apple"));
}

Solution

use std::collections::HashMap;
use std::hash::Hash;

/// Counter counts the number of times each value of type T has been seen.
struct Counter<T> {
    values: HashMap<T, u64>,
}

impl<T: Eq + Hash> Counter<T> {
    /// Create a new Counter.
    fn new() -> Self {
        Counter { values: HashMap::new() }
    }

    /// Count an occurrence of the given value.
    fn count(&mut self, value: T) {
        *self.values.entry(value).or_default() += 1;
    }

    /// Return the number of times the given value has been seen.
    fn times_seen(&self, value: T) -> u64 {
        self.values.get(&value).copied().unwrap_or_default()
    }
}

fn main() {
    let mut ctr = Counter::new();
    ctr.count(13);
    ctr.count(14);
    ctr.count(16);
    ctr.count(14);
    ctr.count(14);
    ctr.count(11);

    for i in 10..20 {
        println!("saw {} values equal to {}", ctr.times_seen(i), i);
    }

    let mut strctr = Counter::new();
    strctr.count("apple");
    strctr.count("orange");
    strctr.count("apple");
    println!("got {} apples", strctr.times_seen("apple"));
}

Standard Library Traits

This segment should take about 1 hour and 10 minutes. It contains:

Comparisons

These traits support comparisons between values. All traits can be derived for types containing fields that implement these traits.

PartialEq and Eq

PartialEq is a partial equivalence relation, with required method eq and provided method ne. The == and != operators will call these methods.

struct Key {
    id: u32,
    metadata: Option<String>,
}
impl PartialEq for Key {
    fn eq(&self, other: &Self) -> bool {
        self.id == other.id
    }
}

Eq is a full equivalence relation (reflexive, symmetric, and transitive) and implies PartialEq. Functions that require full equivalence will use Eq as a trait bound.

PartialOrd and Ord

PartialOrd defines a partial ordering, with a partial_cmp method. It is used to implement the <, <=, >=, and > operators.

use std::cmp::Ordering;
#[derive(Eq, PartialEq)]
struct Citation {
    author: String,
    year: u32,
}
impl PartialOrd for Citation {
    fn partial_cmp(&self, other: &Self) -> Option<Ordering> {
        match self.author.partial_cmp(&other.author) {
            Some(Ordering::Equal) => self.year.partial_cmp(&other.year),
            author_ord => author_ord,
        }
    }
}

Ord is a total ordering, with cmp returning Ordering.

struct Key {
    id: u32,
    metadata: Option<String>,
}
impl PartialEq<u32> for Key {
    fn eq(&self, other: &u32) -> bool {
        self.id == *other
    }
}

Operators

Operator overloading is implemented via traits in std::ops:

#[derive(Debug, Copy, Clone)]
struct Point {
    x: i32,
    y: i32,
}

impl std::ops::Add for Point {
    type Output = Self;

    fn add(self, other: Self) -> Self {
        Self { x: self.x + other.x, y: self.y + other.y }
    }
}

fn main() {
    let p1 = Point { x: 10, y: 20 };
    let p2 = Point { x: 100, y: 200 };
    println!("{p1:?} + {p2:?} = {:?}", p1 + p2);
}

From and Into

Types implement From and Into to facilitate type conversions. Unlike as, these traits correspond to lossless, infallible conversions.

fn main() {
    let s = String::from("hello");
    let addr = std::net::Ipv4Addr::from([127, 0, 0, 1]);
    let one = i16::from(true);
    let bigger = i32::from(123_i16);
    println!("{s}, {addr}, {one}, {bigger}");
}

Into is automatically implemented when From is implemented:

fn main() {
    let s: String = "hello".into();
    let addr: std::net::Ipv4Addr = [127, 0, 0, 1].into();
    let one: i16 = true.into();
    let bigger: i32 = 123_i16.into();
    println!("{s}, {addr}, {one}, {bigger}");
}

Casting

Rust has no implicit type conversions, but does support explicit casts with as. These generally follow C semantics where those are defined.

fn main() {
    let value: i64 = 1000;
    println!("as u16: {}", value as u16);
    println!("as i16: {}", value as i16);
    println!("as u8: {}", value as u8);
}

The results of as are always defined in Rust and consistent across platforms. This might not match your intuition for changing sign or casting to a smaller type -- check the docs, and comment for clarity.

Casting with as is a relatively sharp tool that is easy to use incorrectly, and can be a source of subtle bugs as future maintenance work changes the types that are used or the ranges of values in types. Casts are best used only when the intent is to indicate unconditional truncation (e.g. selecting the bottom 32 bits of a u64 with as u32, regardless of what was in the high bits).

For infallible casts (e.g. u32 to u64), prefer using From or Into over as to confirm that the cast is in fact infallible. For fallible casts, TryFrom and TryInto are available when you want to handle casts that fit differently from those that don't.

Read and Write

Using Read and BufRead, you can abstract over u8 sources:

use std::io::{BufRead, BufReader, Read, Result};

fn count_lines<R: Read>(reader: R) -> usize {
    let buf_reader = BufReader::new(reader);
    buf_reader.lines().count()
}

fn main() -> Result<()> {
    let slice: &[u8] = b"foo\nbar\nbaz\n";
    println!("lines in slice: {}", count_lines(slice));

    let file = std::fs::File::open(std::env::current_exe()?)?;
    println!("lines in file: {}", count_lines(file));
    Ok(())
}

Similarly, Write lets you abstract over u8 sinks:

use std::io::{Result, Write};

fn log<W: Write>(writer: &mut W, msg: &str) -> Result<()> {
    writer.write_all(msg.as_bytes())?;
    writer.write_all("\n".as_bytes())
}

fn main() -> Result<()> {
    let mut buffer = Vec::new();
    log(&mut buffer, "Hello")?;
    log(&mut buffer, "World")?;
    println!("Logged: {buffer:?}");
    Ok(())
}

The Default Trait

Default trait produces a default value for a type.

#[derive(Debug, Default)]
struct Derived {
    x: u32,
    y: String,
    z: Implemented,
}

#[derive(Debug)]
struct Implemented(String);

impl Default for Implemented {
    fn default() -> Self {
        Self("John Smith".into())
    }
}

fn main() {
    let default_struct = Derived::default();
    println!("{default_struct:#?}");

    let almost_default_struct =
        Derived { y: "Y is set!".into(), ..Derived::default() };
    println!("{almost_default_struct:#?}");

    let nothing: Option<Derived> = None;
    println!("{:#?}", nothing.unwrap_or_default());
}

Closures

Closures or lambda expressions have types which cannot be named. However, they implement special Fn, FnMut, and FnOnce traits:

fn apply_and_log(func: impl FnOnce(i32) -> i32, func_name: &str, input: i32) {
    println!("Calling {func_name}({input}): {}", func(input))
}

fn main() {
    let n = 3;
    let add_3 = |x| x + n;
    apply_and_log(&add_3, "add_3", 10);
    apply_and_log(&add_3, "add_3", 20);

    let mut v = Vec::new();
    let mut accumulate = |x: i32| {
        v.push(x);
        v.iter().sum::<i32>()
    };
    apply_and_log(&mut accumulate, "accumulate", 4);
    apply_and_log(&mut accumulate, "accumulate", 5);

    let multiply_sum = |x| x * v.into_iter().sum::<i32>();
    apply_and_log(multiply_sum, "multiply_sum", 3);
}

fn make_greeter(prefix: String) -> impl Fn(&str) {
    return move |name| println!("{} {}", prefix, name);
}

fn main() {
    let hi = make_greeter("Hi".to_string());
    hi("Greg");
}

Exercise: ROT13

In this example, you will implement the classic "ROT13" cipher. Copy this code to the playground, and implement the missing bits. Only rotate ASCII alphabetic characters, to ensure the result is still valid UTF-8.

use std::io::Read;

struct RotDecoder<R: Read> {
    input: R,
    rot: u8,
}

// Implement the `Read` trait for `RotDecoder`.

fn main() {
    let mut rot =
        RotDecoder { input: "Gb trg gb gur bgure fvqr!".as_bytes(), rot: 13 };
    let mut result = String::new();
    rot.read_to_string(&mut result).unwrap();
    println!("{}", result);
}

#[cfg(test)]
mod test {
    use super::*;

    #[test]
    fn joke() {
        let mut rot =
            RotDecoder { input: "Gb trg gb gur bgure fvqr!".as_bytes(), rot: 13 };
        let mut result = String::new();
        rot.read_to_string(&mut result).unwrap();
        assert_eq!(&result, "To get to the other side!");
    }

    #[test]
    fn binary() {
        let input: Vec<u8> = (0..=255u8).collect();
        let mut rot = RotDecoder::<&[u8]> { input: input.as_ref(), rot: 13 };
        let mut buf = [0u8; 256];
        assert_eq!(rot.read(&mut buf).unwrap(), 256);
        for i in 0..=255 {
            if input[i] != buf[i] {
                assert!(input[i].is_ascii_alphabetic());
                assert!(buf[i].is_ascii_alphabetic());
            }
        }
    }
}

What happens if you chain two RotDecoder instances together, each rotating by 13 characters?

Solution

use std::io::Read;

struct RotDecoder<R: Read> {
    input: R,
    rot: u8,
}

impl<R: Read> Read for RotDecoder<R> {
    fn read(&mut self, buf: &mut [u8]) -> std::io::Result<usize> {
        let size = self.input.read(buf)?;
        for b in &mut buf[..size] {
            if b.is_ascii_alphabetic() {
                let base = if b.is_ascii_uppercase() { 'A' } else { 'a' } as u8;
                *b = (*b - base + self.rot) % 26 + base;
            }
        }
        Ok(size)
    }
}

fn main() {
    let mut rot =
        RotDecoder { input: "Gb trg gb gur bgure fvqr!".as_bytes(), rot: 13 };
    let mut result = String::new();
    rot.read_to_string(&mut result).unwrap();
    println!("{}", result);
}

#[cfg(test)]
mod test {
    use super::*;

    #[test]
    fn joke() {
        let mut rot =
            RotDecoder { input: "Gb trg gb gur bgure fvqr!".as_bytes(), rot: 13 };
        let mut result = String::new();
        rot.read_to_string(&mut result).unwrap();
        assert_eq!(&result, "To get to the other side!");
    }

    #[test]
    fn binary() {
        let input: Vec<u8> = (0..=255u8).collect();
        let mut rot = RotDecoder::<&[u8]> { input: input.as_ref(), rot: 13 };
        let mut buf = [0u8; 256];
        assert_eq!(rot.read(&mut buf).unwrap(), 256);
        for i in 0..=255 {
            if input[i] != buf[i] {
                assert!(input[i].is_ascii_alphabetic());
                assert!(buf[i].is_ascii_alphabetic());
            }
        }
    }
}

Welcome to Day 3

Today, we will cover:

• Memory management, lifetimes, and the borrow checker: how Rust ensures memory safety.
• Smart pointers: standard library pointer types.

Schedule

Including 10 minute breaks, this session should take about 2 hours and 20 minutes. It contains:

Memory Management

This segment should take about 1 hour. It contains:

Review of Program Memory

Programs allocate memory in two ways:

• Stack: Continuous area of memory for local variables.

  • Values have fixed sizes known at compile time.
  • Extremely fast: just move a stack pointer.
  • Easy to manage: follows function calls.
  • Great memory locality.

• Heap: Storage of values outside of function calls.

  • Values have dynamic sizes determined at runtime.
  • Slightly slower than the stack: some book-keeping needed.
  • No guarantee of memory locality.

Example

Creating a String puts fixed-sized metadata on the stack and dynamically sized data, the actual string, on the heap:

fn main() {
    let s1 = String::from("Hello");
}

fn main() {
    let mut s1 = String::from("Hello");
    s1.push(' ');
    s1.push_str("world");
    // DON'T DO THIS AT HOME! For educational purposes only.
    // String provides no guarantees about its layout, so this could lead to
    // undefined behavior.
    unsafe {
        let (capacity, ptr, len): (usize, usize, usize) = std::mem::transmute(s1);
        println!("capacity = {capacity}, ptr = {ptr:#x}, len = {len}");
    }
}

Approaches to Memory Management

Traditionally, languages have fallen into two broad categories:

• Full control via manual memory management: C, C++, Pascal, ...
  • Programmer decides when to allocate or free heap memory.
  • Programmer must determine whether a pointer still points to valid memory.
  • Studies show, programmers make mistakes.
• Full safety via automatic memory management at runtime: Java, Python, Go, Haskell, ...
  • A runtime system ensures that memory is not freed until it can no longer be referenced.
  • Typically implemented with reference counting or garbage collection.

Rust offers a new mix:

Full control and safety via compile time enforcement of correct memory management.

It does this with an explicit ownership concept.

Ownership

All variable bindings have a scope where they are valid and it is an error to use a variable outside its scope:

struct Point(i32, i32);

fn main() {
    {
        let p = Point(3, 4);
        println!("x: {}", p.0);
    }
    println!("y: {}", p.1);
}

We say that the variable owns the value. Every Rust value has precisely one owner at all times.

At the end of the scope, the variable is dropped and the data is freed. A destructor can run here to free up resources.

Move Semantics

An assignment will transfer ownership between variables:

fn main() {
    let s1: String = String::from("Hello!");
    let s2: String = s1;
    println!("s2: {s2}");
    // println!("s1: {s1}");
}

• The assignment of s1 to s2 transfers ownership.
• When s1 goes out of scope, nothing happens: it does not own anything.
• When s2 goes out of scope, the string data is freed.

Before move to s2:

After move to s2:

When you pass a value to a function, the value is assigned to the function parameter. This transfers ownership:

fn say_hello(name: String) {
    println!("Hello {name}")
}

fn main() {
    let name = String::from("Alice");
    say_hello(name);
    // say_hello(name);
}

std::string s1 = "Cpp";
std::string s2 = s1;  // Duplicate the data in s1.

Clone

Sometimes you want to make a copy of a value. The Clone trait accomplishes this.

fn say_hello(name: String) {
    println!("Hello {name}")
}

fn main() {
    let name = String::from("Alice");
    say_hello(name.clone());
    say_hello(name);
}

Copy Types

While move semantics are the default, certain types are copied by default:

fn main() {
    let x = 42;
    let y = x;
    println!("x: {x}"); // would not be accessible if not Copy
    println!("y: {y}");
}

These types implement the Copy trait.

You can opt-in your own types to use copy semantics:

#[derive(Copy, Clone, Debug)]
struct Point(i32, i32);

fn main() {
    let p1 = Point(3, 4);
    let p2 = p1;
    println!("p1: {p1:?}");
    println!("p2: {p2:?}");
}

• After the assignment, both p1 and p2 own their own data.
• We can also use p1.clone() to explicitly copy the data.

The Drop Trait

Values which implement Drop can specify code to run when they go out of scope:

struct Droppable {
    name: &'static str,
}

impl Drop for Droppable {
    fn drop(&mut self) {
        println!("Dropping {}", self.name);
    }
}

fn main() {
    let a = Droppable { name: "a" };
    {
        let b = Droppable { name: "b" };
        {
            let c = Droppable { name: "c" };
            let d = Droppable { name: "d" };
            println!("Exiting block B");
        }
        println!("Exiting block A");
    }
    drop(a);
    println!("Exiting main");
}

Exercise: Builder Type

In this example, we will implement a complex data type that owns all of its data. We will use the "builder pattern" to support building a new value piece-by-piece, using convenience functions.

Fill in the missing pieces.

#[derive(Debug)]
enum Language {
    Rust,
    Java,
    Perl,
}

#[derive(Clone, Debug)]
struct Dependency {
    name: String,
    version_expression: String,
}

/// A representation of a software package.
#[derive(Debug)]
struct Package {
    name: String,
    version: String,
    authors: Vec<String>,
    dependencies: Vec<Dependency>,
    language: Option<Language>,
}

impl Package {
    /// Return a representation of this package as a dependency, for use in
    /// building other packages.
    fn as_dependency(&self) -> Dependency {
        todo!("1")
    }
}

/// A builder for a Package. Use `build()` to create the `Package` itself.
struct PackageBuilder(Package);

impl PackageBuilder {
    fn new(name: impl Into<String>) -> Self {
        todo!("2")
    }

    /// Set the package version.
    fn version(mut self, version: impl Into<String>) -> Self {
        self.0.version = version.into();
        self
    }

    /// Set the package authors.
    fn authors(mut self, authors: Vec<String>) -> Self {
        todo!("3")
    }

    /// Add an additional dependency.
    fn dependency(mut self, dependency: Dependency) -> Self {
        todo!("4")
    }

    /// Set the language. If not set, language defaults to None.
    fn language(mut self, language: Language) -> Self {
        todo!("5")
    }

    fn build(self) -> Package {
        self.0
    }
}

fn main() {
    let base64 = PackageBuilder::new("base64").version("0.13").build();
    println!("base64: {base64:?}");
    let log =
        PackageBuilder::new("log").version("0.4").language(Language::Rust).build();
    println!("log: {log:?}");
    let serde = PackageBuilder::new("serde")
        .authors(vec!["djmitche".into()])
        .version(String::from("4.0"))
        .dependency(base64.as_dependency())
        .dependency(log.as_dependency())
        .build();
    println!("serde: {serde:?}");
}

Solution

#[derive(Debug)]
enum Language {
    Rust,
    Java,
    Perl,
}

#[derive(Clone, Debug)]
struct Dependency {
    name: String,
    version_expression: String,
}

/// A representation of a software package.
#[derive(Debug)]
struct Package {
    name: String,
    version: String,
    authors: Vec<String>,
    dependencies: Vec<Dependency>,
    language: Option<Language>,
}

impl Package {
    /// Return a representation of this package as a dependency, for use in
    /// building other packages.
    fn as_dependency(&self) -> Dependency {
        Dependency {
            name: self.name.clone(),
            version_expression: self.version.clone(),
        }
    }
}

/// A builder for a Package. Use `build()` to create the `Package` itself.
struct PackageBuilder(Package);

impl PackageBuilder {
    fn new(name: impl Into<String>) -> Self {
        Self(Package {
            name: name.into(),
            version: "0.1".into(),
            authors: vec![],
            dependencies: vec![],
            language: None,
        })
    }

    /// Set the package version.
    fn version(mut self, version: impl Into<String>) -> Self {
        self.0.version = version.into();
        self
    }

    /// Set the package authors.
    fn authors(mut self, authors: Vec<String>) -> Self {
        self.0.authors = authors;
        self
    }

    /// Add an additional dependency.
    fn dependency(mut self, dependency: Dependency) -> Self {
        self.0.dependencies.push(dependency);
        self
    }

    /// Set the language. If not set, language defaults to None.
    fn language(mut self, language: Language) -> Self {
        self.0.language = Some(language);
        self
    }

    fn build(self) -> Package {
        self.0
    }
}

fn main() {
    let base64 = PackageBuilder::new("base64").version("0.13").build();
    println!("base64: {base64:?}");
    let log =
        PackageBuilder::new("log").version("0.4").language(Language::Rust).build();
    println!("log: {log:?}");
    let serde = PackageBuilder::new("serde")
        .authors(vec!["djmitche".into()])
        .version(String::from("4.0"))
        .dependency(base64.as_dependency())
        .dependency(log.as_dependency())
        .build();
    println!("serde: {serde:?}");
}

Smart Pointers

This segment should take about 55 minutes. It contains:

Box<T>

Box is an owned pointer to data on the heap:

fn main() {
    let five = Box::new(5);
    println!("five: {}", *five);
}

Box<T> implements Deref<Target = T>, which means that you can call methods from T directly on a Box<T>.

Recursive data types or data types with dynamic sizes cannot be stored inline without a pointer indirection. Box accomplishes that indirection:

#[derive(Debug)]
enum List<T> {
    /// A non-empty list: first element and the rest of the list.
    Element(T, Box<List<T>>),
    /// An empty list.
    Nil,
}

fn main() {
    let list: List<i32> =
        List::Element(1, Box::new(List::Element(2, Box::new(List::Nil))));
    println!("{list:?}");
}

Rc

Rc is a reference-counted shared pointer. Use this when you need to refer to the same data from multiple places:

use std::rc::Rc;

fn main() {
    let a = Rc::new(10);
    let b = Rc::clone(&a);

    println!("a: {a}");
    println!("b: {b}");
}

• See Arc and Mutex if you are in a multi-threaded context.
• You can downgrade a shared pointer into a Weak pointer to create cycles that will get dropped.

Owned Trait Objects

We previously saw how trait objects can be used with references, e.g &dyn Pet. However, we can also use trait objects with smart pointers like Box to create an owned trait object: Box<dyn Pet>.

struct Dog {
    name: String,
    age: i8,
}
struct Cat {
    lives: i8,
}

trait Pet {
    fn talk(&self) -> String;
}

impl Pet for Dog {
    fn talk(&self) -> String {
        format!("Woof, my name is {}!", self.name)
    }
}

impl Pet for Cat {
    fn talk(&self) -> String {
        String::from("Miau!")
    }
}

fn main() {
    let pets: Vec<Box<dyn Pet>> = vec![
        Box::new(Cat { lives: 9 }),
        Box::new(Dog { name: String::from("Fido"), age: 5 }),
    ];
    for pet in pets {
        println!("Hello, who are you? {}", pet.talk());
    }
}

Memory layout after allocating pets:

Exercise: Binary Tree

A binary tree is a tree-type data structure where every node has two children (left and right). We will create a tree where each node stores a value. For a given node N, all nodes in a N's left subtree contain smaller values, and all nodes in N's right subtree will contain larger values.

Implement the following types, so that the given tests pass.

Extra Credit: implement an iterator over a binary tree that returns the values in order.

/// A node in the binary tree.
#[derive(Debug)]
struct Node<T: Ord> {
    value: T,
    left: Subtree<T>,
    right: Subtree<T>,
}

/// A possibly-empty subtree.
#[derive(Debug)]
struct Subtree<T: Ord>(Option<Box<Node<T>>>);

/// A container storing a set of values, using a binary tree.
///
/// If the same value is added multiple times, it is only stored once.
#[derive(Debug)]
pub struct BinaryTree<T: Ord> {
    root: Subtree<T>,
}

impl<T: Ord> BinaryTree<T> {
    fn new() -> Self {
        Self { root: Subtree::new() }
    }

    fn insert(&mut self, value: T) {
        self.root.insert(value);
    }

    fn has(&self, value: &T) -> bool {
        self.root.has(value)
    }

    fn len(&self) -> usize {
        self.root.len()
    }
}

// Implement `new`, `insert`, `len`, and `has` for `Subtree`.

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn len() {
        let mut tree = BinaryTree::new();
        assert_eq!(tree.len(), 0);
        tree.insert(2);
        assert_eq!(tree.len(), 1);
        tree.insert(1);
        assert_eq!(tree.len(), 2);
        tree.insert(2); // not a unique item
        assert_eq!(tree.len(), 2);
    }

    #[test]
    fn has() {
        let mut tree = BinaryTree::new();
        fn check_has(tree: &BinaryTree<i32>, exp: &[bool]) {
            let got: Vec<bool> =
                (0..exp.len()).map(|i| tree.has(&(i as i32))).collect();
            assert_eq!(&got, exp);
        }

        check_has(&tree, &[false, false, false, false, false]);
        tree.insert(0);
        check_has(&tree, &[true, false, false, false, false]);
        tree.insert(4);
        check_has(&tree, &[true, false, false, false, true]);
        tree.insert(4);
        check_has(&tree, &[true, false, false, false, true]);
        tree.insert(3);
        check_has(&tree, &[true, false, false, true, true]);
    }

    #[test]
    fn unbalanced() {
        let mut tree = BinaryTree::new();
        for i in 0..100 {
            tree.insert(i);
        }
        assert_eq!(tree.len(), 100);
        assert!(tree.has(&50));
    }
}

Solution

use std::cmp::Ordering;

/// A node in the binary tree.
#[derive(Debug)]
struct Node<T: Ord> {
    value: T,
    left: Subtree<T>,
    right: Subtree<T>,
}

/// A possibly-empty subtree.
#[derive(Debug)]
struct Subtree<T: Ord>(Option<Box<Node<T>>>);

/// A container storing a set of values, using a binary tree.
///
/// If the same value is added multiple times, it is only stored once.
#[derive(Debug)]
pub struct BinaryTree<T: Ord> {
    root: Subtree<T>,
}

impl<T: Ord> BinaryTree<T> {
    fn new() -> Self {
        Self { root: Subtree::new() }
    }

    fn insert(&mut self, value: T) {
        self.root.insert(value);
    }

    fn has(&self, value: &T) -> bool {
        self.root.has(value)
    }

    fn len(&self) -> usize {
        self.root.len()
    }
}

impl<T: Ord> Subtree<T> {
    fn new() -> Self {
        Self(None)
    }

    fn insert(&mut self, value: T) {
        match &mut self.0 {
            None => self.0 = Some(Box::new(Node::new(value))),
            Some(n) => match value.cmp(&n.value) {
                Ordering::Less => n.left.insert(value),
                Ordering::Equal => {}
                Ordering::Greater => n.right.insert(value),
            },
        }
    }

    fn has(&self, value: &T) -> bool {
        match &self.0 {
            None => false,
            Some(n) => match value.cmp(&n.value) {
                Ordering::Less => n.left.has(value),
                Ordering::Equal => true,
                Ordering::Greater => n.right.has(value),
            },
        }
    }

    fn len(&self) -> usize {
        match &self.0 {
            None => 0,
            Some(n) => 1 + n.left.len() + n.right.len(),
        }
    }
}

impl<T: Ord> Node<T> {
    fn new(value: T) -> Self {
        Self { value, left: Subtree::new(), right: Subtree::new() }
    }
}

fn main() {
    let mut tree = BinaryTree::new();
    tree.insert("foo");
    assert_eq!(tree.len(), 1);
    tree.insert("bar");
    assert!(tree.has(&"foo"));
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn len() {
        let mut tree = BinaryTree::new();
        assert_eq!(tree.len(), 0);
        tree.insert(2);
        assert_eq!(tree.len(), 1);
        tree.insert(1);
        assert_eq!(tree.len(), 2);
        tree.insert(2); // not a unique item
        assert_eq!(tree.len(), 2);
    }

    #[test]
    fn has() {
        let mut tree = BinaryTree::new();
        fn check_has(tree: &BinaryTree<i32>, exp: &[bool]) {
            let got: Vec<bool> =
                (0..exp.len()).map(|i| tree.has(&(i as i32))).collect();
            assert_eq!(&got, exp);
        }

        check_has(&tree, &[false, false, false, false, false]);
        tree.insert(0);
        check_has(&tree, &[true, false, false, false, false]);
        tree.insert(4);
        check_has(&tree, &[true, false, false, false, true]);
        tree.insert(4);
        check_has(&tree, &[true, false, false, false, true]);
        tree.insert(3);
        check_has(&tree, &[true, false, false, true, true]);
    }

    #[test]
    fn unbalanced() {
        let mut tree = BinaryTree::new();
        for i in 0..100 {
            tree.insert(i);
        }
        assert_eq!(tree.len(), 100);
        assert!(tree.has(&50));
    }
}

Welcome Back

Including 10 minute breaks, this session should take about 1 hour and 55 minutes. It contains:

Borrowing

This segment should take about 55 minutes. It contains:

Borrowing a Value

As we saw before, instead of transferring ownership when calling a function, you can let a function borrow the value:

#[derive(Debug)]
struct Point(i32, i32);

fn add(p1: &Point, p2: &Point) -> Point {
    Point(p1.0 + p2.0, p1.1 + p2.1)
}

fn main() {
    let p1 = Point(3, 4);
    let p2 = Point(10, 20);
    let p3 = add(&p1, &p2);
    println!("{p1:?} + {p2:?} = {p3:?}");
}

• The add function borrows two points and returns a new point.
• The caller retains ownership of the inputs.

Borrow Checking

Rust's borrow checker puts constraints on the ways you can borrow values. For a given value, at any time:

• You can have one or more shared references to the value, or
• You can have exactly one exclusive reference to the value.

fn main() {
    let mut a: i32 = 10;
    let b: &i32 = &a;

    {
        let c: &mut i32 = &mut a;
        *c = 20;
    }

    println!("a: {a}");
    println!("b: {b}");
}

Borrow Errors

As a concrete example of how these borrowing rules prevent memory errors, consider the case of modifying a collection while there are references to its elements:

fn main() {
    let mut vec = vec![1, 2, 3, 4, 5];
    let elem = &vec[2];
    vec.push(6);
    println!("{elem}");
}

Similarly, consider the case of iterator invalidation:

fn main() {
    let mut vec = vec![1, 2, 3, 4, 5];
    for elem in &vec {
        vec.push(elem * 2);
    }
}

Interior Mutability

In some situations, it's necessary to modify data behind a shared (read-only) reference. For example, a shared data structure might have an internal cache, and wish to update that cache from read-only methods.

The "interior mutability" pattern allows exclusive (mutable) access behind a shared reference. The standard library provides several ways to do this, all while still ensuring safety, typically by performing a runtime check.

Cell

Cell wraps a value and allows getting or setting the value using only a shared reference to the Cell. However, it does not allow any references to the inner value. Since there are no references, borrowing rules cannot be broken.

use std::cell::Cell;

fn main() {
    // Note that `cell` is NOT declared as mutable.
    let cell = Cell::new(5);

    cell.set(123);
    println!("{}", cell.get());
}

RefCell

RefCell allows accessing and mutating a wrapped value by providing alternative types Ref and RefMut that emulate &T/&mut T without actually being Rust references.

These types perform dynamic checks using a counter in the RefCell to prevent existence of a RefMut alongside another Ref/RefMut.

By implementing Deref (and DerefMut for RefMut), these types allow calling methods on the inner value without allowing references to escape.

use std::cell::RefCell;

fn main() {
    // Note that `cell` is NOT declared as mutable.
    let cell = RefCell::new(5);

    {
        let mut cell_ref = cell.borrow_mut();
        *cell_ref = 123;

        // This triggers an error at runtime.
        // let other = cell.borrow();
        // println!("{}", *other);
    }

    println!("{cell:?}");
}

Exercise: Health Statistics

You're working on implementing a health-monitoring system. As part of that, you need to keep track of users' health statistics.

You'll start with a stubbed function in an impl block as well as a User struct definition. Your goal is to implement the stubbed out method on the User struct defined in the impl block.

Copy the code below to https://play.rust-lang.org/ and fill in the missing method:

// TODO: remove this when you're done with your implementation.
#![allow(unused_variables, dead_code)]


#![allow(dead_code)]
pub struct User {
    name: String,
    age: u32,
    height: f32,
    visit_count: usize,
    last_blood_pressure: Option<(u32, u32)>,
}

pub struct Measurements {
    height: f32,
    blood_pressure: (u32, u32),
}

pub struct HealthReport<'a> {
    patient_name: &'a str,
    visit_count: u32,
    height_change: f32,
    blood_pressure_change: Option<(i32, i32)>,
}

impl User {
    pub fn new(name: String, age: u32, height: f32) -> Self {
        Self { name, age, height, visit_count: 0, last_blood_pressure: None }
    }

    pub fn visit_doctor(&mut self, measurements: Measurements) -> HealthReport {
        todo!("Update a user's statistics based on measurements from a visit to the doctor")
    }
}

fn main() {
    let bob = User::new(String::from("Bob"), 32, 155.2);
    println!("I'm {} and my age is {}", bob.name, bob.age);
}

#[test]
fn test_visit() {
    let mut bob = User::new(String::from("Bob"), 32, 155.2);
    assert_eq!(bob.visit_count, 0);
    let report =
        bob.visit_doctor(Measurements { height: 156.1, blood_pressure: (120, 80) });
    assert_eq!(report.patient_name, "Bob");
    assert_eq!(report.visit_count, 1);
    assert_eq!(report.blood_pressure_change, None);
    assert!((report.height_change - 0.9).abs() < 0.00001);

    let report =
        bob.visit_doctor(Measurements { height: 156.1, blood_pressure: (115, 76) });

    assert_eq!(report.visit_count, 2);
    assert_eq!(report.blood_pressure_change, Some((-5, -4)));
    assert_eq!(report.height_change, 0.0);
}

Solution

#![allow(dead_code)]
pub struct User {
    name: String,
    age: u32,
    height: f32,
    visit_count: usize,
    last_blood_pressure: Option<(u32, u32)>,
}

pub struct Measurements {
    height: f32,
    blood_pressure: (u32, u32),
}

pub struct HealthReport<'a> {
    patient_name: &'a str,
    visit_count: u32,
    height_change: f32,
    blood_pressure_change: Option<(i32, i32)>,
}

impl User {
    pub fn new(name: String, age: u32, height: f32) -> Self {
        Self { name, age, height, visit_count: 0, last_blood_pressure: None }
    }

    pub fn visit_doctor(&mut self, measurements: Measurements) -> HealthReport {
        self.visit_count += 1;
        let bp = measurements.blood_pressure;
        let report = HealthReport {
            patient_name: &self.name,
            visit_count: self.visit_count as u32,
            height_change: measurements.height - self.height,
            blood_pressure_change: match self.last_blood_pressure {
                Some(lbp) => {
                    Some((bp.0 as i32 - lbp.0 as i32, bp.1 as i32 - lbp.1 as i32))
                }
                None => None,
            },
        };
        self.height = measurements.height;
        self.last_blood_pressure = Some(bp);
        report
    }
}

fn main() {
    let bob = User::new(String::from("Bob"), 32, 155.2);
    println!("I'm {} and my age is {}", bob.name, bob.age);
}

#[test]
fn test_visit() {
    let mut bob = User::new(String::from("Bob"), 32, 155.2);
    assert_eq!(bob.visit_count, 0);
    let report =
        bob.visit_doctor(Measurements { height: 156.1, blood_pressure: (120, 80) });
    assert_eq!(report.patient_name, "Bob");
    assert_eq!(report.visit_count, 1);
    assert_eq!(report.blood_pressure_change, None);
    assert!((report.height_change - 0.9).abs() < 0.00001);

    let report =
        bob.visit_doctor(Measurements { height: 156.1, blood_pressure: (115, 76) });

    assert_eq!(report.visit_count, 2);
    assert_eq!(report.blood_pressure_change, Some((-5, -4)));
    assert_eq!(report.height_change, 0.0);
}

Lifetimes

This segment should take about 50 minutes. It contains:

Lifetime Annotations

A reference has a lifetime, which must not "outlive" the value it refers to. This is verified by the borrow checker.

The lifetime can be implicit - this is what we have seen so far. Lifetimes can also be explicit: &'a Point, &'document str. Lifetimes start with ' and 'a is a typical default name. Read &'a Point as "a borrowed Point which is valid for at least the lifetime a".

Lifetimes are always inferred by the compiler: you cannot assign a lifetime yourself. Explicit lifetime annotations create constraints where there is ambiguity; the compiler verifies that there is a valid solution.

Lifetimes become more complicated when considering passing values to and returning values from functions.

#[derive(Debug)]
struct Point(i32, i32);

fn left_most(p1: &Point, p2: &Point) -> &Point {
    if p1.0 < p2.0 {
        p1
    } else {
        p2
    }
}

fn main() {
    let p1: Point = Point(10, 10);
    let p2: Point = Point(20, 20);
    let p3 = left_most(&p1, &p2); // What is the lifetime of p3?
    println!("p3: {p3:?}");
}

fn left_most<'a>(p1: &'a Point, p2: &'a Point) -> &'a Point {

Lifetimes in Function Calls

Lifetimes for function arguments and return values must be fully specified, but Rust allows lifetimes to be elided in most cases with a few simple rules. This is not inference -- it is just a syntactic shorthand.

• Each argument which does not have a lifetime annotation is given one.
• If there is only one argument lifetime, it is given to all un-annotated return values.
• If there are multiple argument lifetimes, but the first one is for self, that lifetime is given to all un-annotated return values.

#[derive(Debug)]
struct Point(i32, i32);

fn cab_distance(p1: &Point, p2: &Point) -> i32 {
    (p1.0 - p2.0).abs() + (p1.1 - p2.1).abs()
}

fn nearest<'a>(points: &'a [Point], query: &Point) -> Option<&'a Point> {
    let mut nearest = None;
    for p in points {
        if let Some((_, nearest_dist)) = nearest {
            let dist = cab_distance(p, query);
            if dist < nearest_dist {
                nearest = Some((p, dist));
            }
        } else {
            nearest = Some((p, cab_distance(p, query)));
        };
    }
    nearest.map(|(p, _)| p)
}

fn main() {
    let points = &[Point(1, 0), Point(1, 0), Point(-1, 0), Point(0, -1)];
    println!("{:?}", nearest(points, &Point(0, 2)));
}

fn nearest<'a, 'q>(points: &'a [Point], query: &'q Point) -> Option<&'q Point> {

Lifetimes in Data Structures

If a data type stores borrowed data, it must be annotated with a lifetime:

#[derive(Debug)]
struct Highlight<'doc>(&'doc str);

fn erase(text: String) {
    println!("Bye {text}!");
}

fn main() {
    let text = String::from("The quick brown fox jumps over the lazy dog.");
    let fox = Highlight(&text[4..19]);
    let dog = Highlight(&text[35..43]);
    // erase(text);
    println!("{fox:?}");
    println!("{dog:?}");
}

Exercise: Protobuf Parsing

In this exercise, you will build a parser for the protobuf binary encoding. Don't worry, it's simpler than it seems! This illustrates a common parsing pattern, passing slices of data. The underlying data itself is never copied.

Fully parsing a protobuf message requires knowing the types of the fields, indexed by their field numbers. That is typically provided in a proto file. In this exercise, we'll encode that information into match statements in functions that get called for each field.

We'll use the following proto:

message PhoneNumber {
  optional string number = 1;
  optional string type = 2;
}

message Person {
  optional string name = 1;
  optional int32 id = 2;
  repeated PhoneNumber phones = 3;
}

A proto message is encoded as a series of fields, one after the next. Each is implemented as a "tag" followed by the value. The tag contains a field number (e.g., 2 for the id field of a Person message) and a wire type defining how the payload should be determined from the byte stream.

Integers, including the tag, are represented with a variable-length encoding called VARINT. Luckily, parse_varint is defined for you below. The given code also defines callbacks to handle Person and PhoneNumber fields, and to parse a message into a series of calls to those callbacks.

What remains for you is to implement the parse_field function and the ProtoMessage trait for Person and PhoneNumber.

/// A wire type as seen on the wire.
enum WireType {
    /// The Varint WireType indicates the value is a single VARINT.
    Varint,
    /// The I64 WireType indicates that the value is precisely 8 bytes in
    /// little-endian order containing a 64-bit signed integer or double type.
    //I64,  -- not needed for this exercise
    /// The Len WireType indicates that the value is a length represented as a
    /// VARINT followed by exactly that number of bytes.
    Len,
    // The I32 WireType indicates that the value is precisely 4 bytes in
    // little-endian order containing a 32-bit signed integer or float type.
    //I32,  -- not needed for this exercise
}

#[derive(Debug)]
/// A field's value, typed based on the wire type.
enum FieldValue<'a> {
    Varint(u64),
    //I64(i64),  -- not needed for this exercise
    Len(&'a [u8]),
    //I32(i32),  -- not needed for this exercise
}

#[derive(Debug)]
/// A field, containing the field number and its value.
struct Field<'a> {
    field_num: u64,
    value: FieldValue<'a>,
}

trait ProtoMessage<'a>: Default {
    fn add_field(&mut self, field: Field<'a>);
}

impl From<u64> for WireType {
    fn from(value: u64) -> Self {
        match value {
            0 => WireType::Varint,
            //1 => WireType::I64,  -- not needed for this exercise
            2 => WireType::Len,
            //5 => WireType::I32,  -- not needed for this exercise
            _ => panic!("Invalid wire type: {value}"),
        }
    }
}

impl<'a> FieldValue<'a> {
    fn as_str(&self) -> &'a str {
        let FieldValue::Len(data) = self else {
            panic!("Expected string to be a `Len` field");
        };
        std::str::from_utf8(data).expect("Invalid string")
    }

    fn as_bytes(&self) -> &'a [u8] {
        let FieldValue::Len(data) = self else {
            panic!("Expected bytes to be a `Len` field");
        };
        data
    }

    fn as_u64(&self) -> u64 {
        let FieldValue::Varint(value) = self else {
            panic!("Expected `u64` to be a `Varint` field");
        };
        *value
    }
}

/// Parse a VARINT, returning the parsed value and the remaining bytes.
fn parse_varint(data: &[u8]) -> (u64, &[u8]) {
    for i in 0..7 {
        let Some(b) = data.get(i) else {
            panic!("Not enough bytes for varint");
        };
        if b & 0x80 == 0 {
            // This is the last byte of the VARINT, so convert it to
            // a u64 and return it.
            let mut value = 0u64;
            for b in data[..=i].iter().rev() {
                value = (value << 7) | (b & 0x7f) as u64;
            }
            return (value, &data[i + 1..]);
        }
    }

    // More than 7 bytes is invalid.
    panic!("Too many bytes for varint");
}

/// Convert a tag into a field number and a WireType.
fn unpack_tag(tag: u64) -> (u64, WireType) {
    let field_num = tag >> 3;
    let wire_type = WireType::from(tag & 0x7);
    (field_num, wire_type)
}


/// Parse a field, returning the remaining bytes
fn parse_field(data: &[u8]) -> (Field, &[u8]) {
    let (tag, remainder) = parse_varint(data);
    let (field_num, wire_type) = unpack_tag(tag);
    let (fieldvalue, remainder) = match wire_type {
        _ => todo!("Based on the wire type, build a Field, consuming as many bytes as necessary.")
    };
    todo!("Return the field, and any un-consumed bytes.")
}

/// Parse a message in the given data, calling `T::add_field` for each field in
/// the message.
///
/// The entire input is consumed.
fn parse_message<'a, T: ProtoMessage<'a>>(mut data: &'a [u8]) -> T {
    let mut result = T::default();
    while !data.is_empty() {
        let parsed = parse_field(data);
        result.add_field(parsed.0);
        data = parsed.1;
    }
    result
}

#[derive(Debug, Default)]
struct PhoneNumber<'a> {
    number: &'a str,
    type_: &'a str,
}

#[derive(Debug, Default)]
struct Person<'a> {
    name: &'a str,
    id: u64,
    phone: Vec<PhoneNumber<'a>>,
}

// TODO: Implement ProtoMessage for Person and PhoneNumber.

fn main() {
    let person: Person = parse_message(&[
        0x0a, 0x07, 0x6d, 0x61, 0x78, 0x77, 0x65, 0x6c, 0x6c, 0x10, 0x2a, 0x1a,
        0x16, 0x0a, 0x0e, 0x2b, 0x31, 0x32, 0x30, 0x32, 0x2d, 0x35, 0x35, 0x35,
        0x2d, 0x31, 0x32, 0x31, 0x32, 0x12, 0x04, 0x68, 0x6f, 0x6d, 0x65, 0x1a,
        0x18, 0x0a, 0x0e, 0x2b, 0x31, 0x38, 0x30, 0x30, 0x2d, 0x38, 0x36, 0x37,
        0x2d, 0x35, 0x33, 0x30, 0x38, 0x12, 0x06, 0x6d, 0x6f, 0x62, 0x69, 0x6c,
        0x65,
    ]);
    println!("{:#?}", person);
}

Solution

/// A wire type as seen on the wire.
enum WireType {
    /// The Varint WireType indicates the value is a single VARINT.
    Varint,
    /// The I64 WireType indicates that the value is precisely 8 bytes in
    /// little-endian order containing a 64-bit signed integer or double type.
    //I64,  -- not needed for this exercise
    /// The Len WireType indicates that the value is a length represented as a
    /// VARINT followed by exactly that number of bytes.
    Len,
    // The I32 WireType indicates that the value is precisely 4 bytes in
    // little-endian order containing a 32-bit signed integer or float type.
    //I32,  -- not needed for this exercise
}

#[derive(Debug)]
/// A field's value, typed based on the wire type.
enum FieldValue<'a> {
    Varint(u64),
    //I64(i64),  -- not needed for this exercise
    Len(&'a [u8]),
    //I32(i32),  -- not needed for this exercise
}

#[derive(Debug)]
/// A field, containing the field number and its value.
struct Field<'a> {
    field_num: u64,
    value: FieldValue<'a>,
}

trait ProtoMessage<'a>: Default {
    fn add_field(&mut self, field: Field<'a>);
}

impl From<u64> for WireType {
    fn from(value: u64) -> Self {
        match value {
            0 => WireType::Varint,
            //1 => WireType::I64,  -- not needed for this exercise
            2 => WireType::Len,
            //5 => WireType::I32,  -- not needed for this exercise
            _ => panic!("Invalid wire type: {value}"),
        }
    }
}

impl<'a> FieldValue<'a> {
    fn as_str(&self) -> &'a str {
        let FieldValue::Len(data) = self else {
            panic!("Expected string to be a `Len` field");
        };
        std::str::from_utf8(data).expect("Invalid string")
    }

    fn as_bytes(&self) -> &'a [u8] {
        let FieldValue::Len(data) = self else {
            panic!("Expected bytes to be a `Len` field");
        };
        data
    }

    fn as_u64(&self) -> u64 {
        let FieldValue::Varint(value) = self else {
            panic!("Expected `u64` to be a `Varint` field");
        };
        *value
    }
}

/// Parse a VARINT, returning the parsed value and the remaining bytes.
fn parse_varint(data: &[u8]) -> (u64, &[u8]) {
    for i in 0..7 {
        let Some(b) = data.get(i) else {
            panic!("Not enough bytes for varint");
        };
        if b & 0x80 == 0 {
            // This is the last byte of the VARINT, so convert it to
            // a u64 and return it.
            let mut value = 0u64;
            for b in data[..=i].iter().rev() {
                value = (value << 7) | (b & 0x7f) as u64;
            }
            return (value, &data[i + 1..]);
        }
    }

    // More than 7 bytes is invalid.
    panic!("Too many bytes for varint");
}

/// Convert a tag into a field number and a WireType.
fn unpack_tag(tag: u64) -> (u64, WireType) {
    let field_num = tag >> 3;
    let wire_type = WireType::from(tag & 0x7);
    (field_num, wire_type)
}

/// Parse a field, returning the remaining bytes
fn parse_field(data: &[u8]) -> (Field, &[u8]) {
    let (tag, remainder) = parse_varint(data);
    let (field_num, wire_type) = unpack_tag(tag);
    let (fieldvalue, remainder) = match wire_type {
        WireType::Varint => {
            let (value, remainder) = parse_varint(remainder);
            (FieldValue::Varint(value), remainder)
        }
        WireType::Len => {
            let (len, remainder) = parse_varint(remainder);
            let len: usize = len.try_into().expect("len not a valid `usize`");
            if remainder.len() < len {
                panic!("Unexpected EOF");
            }
            let (value, remainder) = remainder.split_at(len);
            (FieldValue::Len(value), remainder)
        }
    };
    (Field { field_num, value: fieldvalue }, remainder)
}

/// Parse a message in the given data, calling `T::add_field` for each field in
/// the message.
///
/// The entire input is consumed.
fn parse_message<'a, T: ProtoMessage<'a>>(mut data: &'a [u8]) -> T {
    let mut result = T::default();
    while !data.is_empty() {
        let parsed = parse_field(data);
        result.add_field(parsed.0);
        data = parsed.1;
    }
    result
}

#[derive(PartialEq)]
#[derive(Debug, Default)]
struct PhoneNumber<'a> {
    number: &'a str,
    type_: &'a str,
}

#[derive(PartialEq)]
#[derive(Debug, Default)]
struct Person<'a> {
    name: &'a str,
    id: u64,
    phone: Vec<PhoneNumber<'a>>,
}

impl<'a> ProtoMessage<'a> for Person<'a> {
    fn add_field(&mut self, field: Field<'a>) {
        match field.field_num {
            1 => self.name = field.value.as_str(),
            2 => self.id = field.value.as_u64(),
            3 => self.phone.push(parse_message(field.value.as_bytes())),
            _ => {} // skip everything else
        }
    }
}

impl<'a> ProtoMessage<'a> for PhoneNumber<'a> {
    fn add_field(&mut self, field: Field<'a>) {
        match field.field_num {
            1 => self.number = field.value.as_str(),
            2 => self.type_ = field.value.as_str(),
            _ => {} // skip everything else
        }
    }
}

fn main() {
    let person: Person = parse_message(&[
        0x0a, 0x07, 0x6d, 0x61, 0x78, 0x77, 0x65, 0x6c, 0x6c, 0x10, 0x2a, 0x1a,
        0x16, 0x0a, 0x0e, 0x2b, 0x31, 0x32, 0x30, 0x32, 0x2d, 0x35, 0x35, 0x35,
        0x2d, 0x31, 0x32, 0x31, 0x32, 0x12, 0x04, 0x68, 0x6f, 0x6d, 0x65, 0x1a,
        0x18, 0x0a, 0x0e, 0x2b, 0x31, 0x38, 0x30, 0x30, 0x2d, 0x38, 0x36, 0x37,
        0x2d, 0x35, 0x33, 0x30, 0x38, 0x12, 0x06, 0x6d, 0x6f, 0x62, 0x69, 0x6c,
        0x65,
    ]);
    println!("{:#?}", person);
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_id() {
        let person_id: Person = parse_message(&[0x10, 0x2a]);
        assert_eq!(person_id, Person { name: "", id: 42, phone: vec![] });
    }

    #[test]
    fn test_name() {
        let person_name: Person = parse_message(&[
            0x0a, 0x0e, 0x62, 0x65, 0x61, 0x75, 0x74, 0x69, 0x66, 0x75, 0x6c, 0x20,
            0x6e, 0x61, 0x6d, 0x65,
        ]);
        assert_eq!(
            person_name,
            Person { name: "beautiful name", id: 0, phone: vec![] }
        );
    }

    #[test]
    fn test_just_person() {
        let person_name_id: Person =
            parse_message(&[0x0a, 0x04, 0x45, 0x76, 0x61, 0x6e, 0x10, 0x16]);
        assert_eq!(person_name_id, Person { name: "Evan", id: 22, phone: vec![] });
    }

    #[test]
    fn test_phone() {
        let phone: Person = parse_message(&[
            0x0a, 0x00, 0x10, 0x00, 0x1a, 0x16, 0x0a, 0x0e, 0x2b, 0x31, 0x32, 0x33,
            0x34, 0x2d, 0x37, 0x37, 0x37, 0x2d, 0x39, 0x30, 0x39, 0x30, 0x12, 0x04,
            0x68, 0x6f, 0x6d, 0x65,
        ]);
        assert_eq!(
            phone,
            Person {
                name: "",
                id: 0,
                phone: vec![PhoneNumber { number: "+1234-777-9090", type_: "home" },],
            }
        );
    }
}

Welcome to Day 4

Today we will cover topics relating to building large-scale software in Rust:

• Iterators: a deep dive on the Iterator trait.
• Modules and visibility.
• Testing.
• Error handling: panics, Result, and the try operator ?.
• Unsafe Rust: the escape hatch when you can't express yourself in safe Rust.

Schedule

Including 10 minute breaks, this session should take about 2 hours and 40 minutes. It contains:

Iterators

This segment should take about 45 minutes. It contains:

Iterator

The Iterator trait supports iterating over values in a collection. It requires a next method and provides lots of methods. Many standard library types implement Iterator, and you can implement it yourself, too:

struct Fibonacci {
    curr: u32,
    next: u32,
}

impl Iterator for Fibonacci {
    type Item = u32;

    fn next(&mut self) -> Option<Self::Item> {
        let new_next = self.curr + self.next;
        self.curr = self.next;
        self.next = new_next;
        Some(self.curr)
    }
}

fn main() {
    let fib = Fibonacci { curr: 0, next: 1 };
    for (i, n) in fib.enumerate().take(5) {
        println!("fib({i}): {n}");
    }
}

IntoIterator

The Iterator trait tells you how to iterate once you have created an iterator. The related trait IntoIterator defines how to create an iterator for a type. It is used automatically by the for loop.

struct Grid {
    x_coords: Vec<u32>,
    y_coords: Vec<u32>,
}

impl IntoIterator for Grid {
    type Item = (u32, u32);
    type IntoIter = GridIter;
    fn into_iter(self) -> GridIter {
        GridIter { grid: self, i: 0, j: 0 }
    }
}

struct GridIter {
    grid: Grid,
    i: usize,
    j: usize,
}

impl Iterator for GridIter {
    type Item = (u32, u32);

    fn next(&mut self) -> Option<(u32, u32)> {
        if self.i >= self.grid.x_coords.len() {
            self.i = 0;
            self.j += 1;
            if self.j >= self.grid.y_coords.len() {
                return None;
            }
        }
        let res = Some((self.grid.x_coords[self.i], self.grid.y_coords[self.j]));
        self.i += 1;
        res
    }
}

fn main() {
    let grid = Grid { x_coords: vec![3, 5, 7, 9], y_coords: vec![10, 20, 30, 40] };
    for (x, y) in grid {
        println!("point = {x}, {y}");
    }
}

FromIterator

FromIterator lets you build a collection from an Iterator.

fn main() {
    let primes = vec![2, 3, 5, 7];
    let prime_squares = primes.into_iter().map(|p| p * p).collect::<Vec<_>>();
    println!("prime_squares: {prime_squares:?}");
}

fn collect<B>(self) -> B
where
    B: FromIterator<Self::Item>,
    Self: Sized

Exercise: Iterator Method Chaining

In this exercise, you will need to find and use some of the provided methods in the Iterator trait to implement a complex calculation.

Copy the following code to https://play.rust-lang.org/ and make the tests pass. Use an iterator expression and collect the result to construct the return value.

#![allow(unused)]
fn main() {
/// Calculate the differences between elements of `values` offset by `offset`,
/// wrapping around from the end of `values` to the beginning.
///
/// Element `n` of the result is `values[(n+offset)%len] - values[n]`.
fn offset_differences(offset: usize, values: Vec<i32>) -> Vec<i32> {
    unimplemented!()
}

#[test]
fn test_offset_one() {
    assert_eq!(offset_differences(1, vec![1, 3, 5, 7]), vec![2, 2, 2, -6]);
    assert_eq!(offset_differences(1, vec![1, 3, 5]), vec![2, 2, -4]);
    assert_eq!(offset_differences(1, vec![1, 3]), vec![2, -2]);
}

#[test]
fn test_larger_offsets() {
    assert_eq!(offset_differences(2, vec![1, 3, 5, 7]), vec![4, 4, -4, -4]);
    assert_eq!(offset_differences(3, vec![1, 3, 5, 7]), vec![6, -2, -2, -2]);
    assert_eq!(offset_differences(4, vec![1, 3, 5, 7]), vec![0, 0, 0, 0]);
    assert_eq!(offset_differences(5, vec![1, 3, 5, 7]), vec![2, 2, 2, -6]);
}

#[test]
fn test_degenerate_cases() {
    assert_eq!(offset_differences(1, vec![0]), vec![0]);
    assert_eq!(offset_differences(1, vec![1]), vec![0]);
    let empty: Vec<i32> = vec![];
    assert_eq!(offset_differences(1, empty), vec![]);
}
}

Solution

/// Calculate the differences between elements of `values` offset by `offset`,
/// wrapping around from the end of `values` to the beginning.
///
/// Element `n` of the result is `values[(n+offset)%len] - values[n]`.
fn offset_differences(offset: usize, values: Vec<i32>) -> Vec<i32> {
    let a = (&values).into_iter();
    let b = (&values).into_iter().cycle().skip(offset);
    a.zip(b).map(|(a, b)| *b - *a).collect()
}

#[test]
fn test_offset_one() {
    assert_eq!(offset_differences(1, vec![1, 3, 5, 7]), vec![2, 2, 2, -6]);
    assert_eq!(offset_differences(1, vec![1, 3, 5]), vec![2, 2, -4]);
    assert_eq!(offset_differences(1, vec![1, 3]), vec![2, -2]);
}

#[test]
fn test_larger_offsets() {
    assert_eq!(offset_differences(2, vec![1, 3, 5, 7]), vec![4, 4, -4, -4]);
    assert_eq!(offset_differences(3, vec![1, 3, 5, 7]), vec![6, -2, -2, -2]);
    assert_eq!(offset_differences(4, vec![1, 3, 5, 7]), vec![0, 0, 0, 0]);
    assert_eq!(offset_differences(5, vec![1, 3, 5, 7]), vec![2, 2, 2, -6]);
}

#[test]
fn test_degenerate_cases() {
    assert_eq!(offset_differences(1, vec![0]), vec![0]);
    assert_eq!(offset_differences(1, vec![1]), vec![0]);
    let empty: Vec<i32> = vec![];
    assert_eq!(offset_differences(1, empty), vec![]);
}

fn main() {}

Modules

This segment should take about 40 minutes. It contains:

Modules

We have seen how impl blocks let us namespace functions to a type.

Similarly, mod lets us namespace types and functions:

mod foo {
    pub fn do_something() {
        println!("In the foo module");
    }
}

mod bar {
    pub fn do_something() {
        println!("In the bar module");
    }
}

fn main() {
    foo::do_something();
    bar::do_something();
}

Filesystem Hierarchy

Omitting the module content will tell Rust to look for it in another file:

mod garden;

This tells Rust that the garden module content is found at src/garden.rs. Similarly, a garden::vegetables module can be found at src/garden/vegetables.rs.

The crate root is in:

• src/lib.rs (for a library crate)
• src/main.rs (for a binary crate)

Modules defined in files can be documented, too, using "inner doc comments". These document the item that contains them -- in this case, a module.

//! This module implements the garden, including a highly performant germination
//! implementation.

// Re-export types from this module.
pub use garden::Garden;
pub use seeds::SeedPacket;

/// Sow the given seed packets.
pub fn sow(seeds: Vec<SeedPacket>) {
    todo!()
}

/// Harvest the produce in the garden that is ready.
pub fn harvest(garden: &mut Garden) {
    todo!()
}

Visibility

Modules are a privacy boundary:

• Module items are private by default (hides implementation details).
• Parent and sibling items are always visible.
• In other words, if an item is visible in module foo, it's visible in all the descendants of foo.

mod outer {
    fn private() {
        println!("outer::private");
    }

    pub fn public() {
        println!("outer::public");
    }

    mod inner {
        fn private() {
            println!("outer::inner::private");
        }

        pub fn public() {
            println!("outer::inner::public");
            super::private();
        }
    }
}

fn main() {
    outer::public();
}

use, super, self

A module can bring symbols from another module into scope with use. You will typically see something like this at the top of each module:

use std::collections::HashSet;
use std::process::abort;

Paths

Paths are resolved as follows:

1. As a relative path:

   • foo or self::foo refers to foo in the current module,
   • super::foo refers to foo in the parent module.

2. As an absolute path:

   • crate::foo refers to foo in the root of the current crate,
   • bar::foo refers to foo in the bar crate.

Exercise: Modules for a GUI Library

In this exercise, you will reorganize a small GUI Library implementation. This library defines a Widget trait and a few implementations of that trait, as well as a main function.

It is typical to put each type or set of closely-related types into its own module, so each widget type should get its own module.

Cargo Setup

The Rust playground only supports one file, so you will need to make a Cargo project on your local filesystem:

cargo init gui-modules
cd gui-modules
cargo run

Edit the resulting src/main.rs to add mod statements, and add additional files in the src directory.

Source

Here's the single-module implementation of the GUI library:

pub trait Widget {
    /// Natural width of `self`.
    fn width(&self) -> usize;

    /// Draw the widget into a buffer.
    fn draw_into(&self, buffer: &mut dyn std::fmt::Write);

    /// Draw the widget on standard output.
    fn draw(&self) {
        let mut buffer = String::new();
        self.draw_into(&mut buffer);
        println!("{buffer}");
    }
}

pub struct Label {
    label: String,
}

impl Label {
    fn new(label: &str) -> Label {
        Label { label: label.to_owned() }
    }
}

pub struct Button {
    label: Label,
}

impl Button {
    fn new(label: &str) -> Button {
        Button { label: Label::new(label) }
    }
}

pub struct Window {
    title: String,
    widgets: Vec<Box<dyn Widget>>,
}

impl Window {
    fn new(title: &str) -> Window {
        Window { title: title.to_owned(), widgets: Vec::new() }
    }

    fn add_widget(&mut self, widget: Box<dyn Widget>) {
        self.widgets.push(widget);
    }

    fn inner_width(&self) -> usize {
        std::cmp::max(
            self.title.chars().count(),
            self.widgets.iter().map(|w| w.width()).max().unwrap_or(0),
        )
    }
}

impl Widget for Window {
    fn width(&self) -> usize {
        // Add 4 paddings for borders
        self.inner_width() + 4
    }

    fn draw_into(&self, buffer: &mut dyn std::fmt::Write) {
        let mut inner = String::new();
        for widget in &self.widgets {
            widget.draw_into(&mut inner);
        }

        let inner_width = self.inner_width();

        // TODO: Change draw_into to return Result<(), std::fmt::Error>. Then use the
        // ?-operator here instead of .unwrap().
        writeln!(buffer, "+-{:-<inner_width$}-+", "").unwrap();
        writeln!(buffer, "| {:^inner_width$} |", &self.title).unwrap();
        writeln!(buffer, "+={:=<inner_width$}=+", "").unwrap();
        for line in inner.lines() {
            writeln!(buffer, "| {:inner_width$} |", line).unwrap();
        }
        writeln!(buffer, "+-{:-<inner_width$}-+", "").unwrap();
    }
}

impl Widget for Button {
    fn width(&self) -> usize {
        self.label.width() + 8 // add a bit of padding
    }

    fn draw_into(&self, buffer: &mut dyn std::fmt::Write) {
        let width = self.width();
        let mut label = String::new();
        self.label.draw_into(&mut label);

        writeln!(buffer, "+{:-<width$}+", "").unwrap();
        for line in label.lines() {
            writeln!(buffer, "|{:^width$}|", &line).unwrap();
        }
        writeln!(buffer, "+{:-<width$}+", "").unwrap();
    }
}

impl Widget for Label {
    fn width(&self) -> usize {
        self.label.lines().map(|line| line.chars().count()).max().unwrap_or(0)
    }

    fn draw_into(&self, buffer: &mut dyn std::fmt::Write) {
        writeln!(buffer, "{}", &self.label).unwrap();
    }
}

fn main() {
    let mut window = Window::new("Rust GUI Demo 1.23");
    window.add_widget(Box::new(Label::new("This is a small text GUI demo.")));
    window.add_widget(Box::new(Button::new("Click me!")));
    window.draw();
}

Solution

src
├── main.rs
├── widgets
│   ├── button.rs
│   ├── label.rs
│   └── window.rs
└── widgets.rs

// ---- src/widgets.rs ----
mod button;
mod label;
mod window;

pub trait Widget {
    /// Natural width of `self`.
    fn width(&self) -> usize;

    /// Draw the widget into a buffer.
    fn draw_into(&self, buffer: &mut dyn std::fmt::Write);

    /// Draw the widget on standard output.
    fn draw(&self) {
        let mut buffer = String::new();
        self.draw_into(&mut buffer);
        println!("{buffer}");
    }
}

pub use button::Button;
pub use label::Label;
pub use window::Window;

// ---- src/widgets/label.rs ----
use super::Widget;

pub struct Label {
    label: String,
}

impl Label {
    pub fn new(label: &str) -> Label {
        Label { label: label.to_owned() }
    }
}

impl Widget for Label {
    fn width(&self) -> usize {
        // ANCHOR_END: Label-width
        self.label.lines().map(|line| line.chars().count()).max().unwrap_or(0)
    }

    // ANCHOR: Label-draw_into
    fn draw_into(&self, buffer: &mut dyn std::fmt::Write) {
        // ANCHOR_END: Label-draw_into
        writeln!(buffer, "{}", &self.label).unwrap();
    }
}

// ---- src/widgets/button.rs ----
use super::{Label, Widget};

pub struct Button {
    label: Label,
}

impl Button {
    pub fn new(label: &str) -> Button {
        Button { label: Label::new(label) }
    }
}

impl Widget for Button {
    fn width(&self) -> usize {
        // ANCHOR_END: Button-width
        self.label.width() + 8 // add a bit of padding
    }

    // ANCHOR: Button-draw_into
    fn draw_into(&self, buffer: &mut dyn std::fmt::Write) {
        // ANCHOR_END: Button-draw_into
        let width = self.width();
        let mut label = String::new();
        self.label.draw_into(&mut label);

        writeln!(buffer, "+{:-<width$}+", "").unwrap();
        for line in label.lines() {
            writeln!(buffer, "|{:^width$}|", &line).unwrap();
        }
        writeln!(buffer, "+{:-<width$}+", "").unwrap();
    }
}

// ---- src/widgets/window.rs ----
use super::Widget;

pub struct Window {
    title: String,
    widgets: Vec<Box<dyn Widget>>,
}

impl Window {
    pub fn new(title: &str) -> Window {
        Window { title: title.to_owned(), widgets: Vec::new() }
    }

    pub fn add_widget(&mut self, widget: Box<dyn Widget>) {
        self.widgets.push(widget);
    }

    fn inner_width(&self) -> usize {
        std::cmp::max(
            self.title.chars().count(),
            self.widgets.iter().map(|w| w.width()).max().unwrap_or(0),
        )
    }
}

impl Widget for Window {
    fn width(&self) -> usize {
        // ANCHOR_END: Window-width
        // Add 4 paddings for borders
        self.inner_width() + 4
    }

    // ANCHOR: Window-draw_into
    fn draw_into(&self, buffer: &mut dyn std::fmt::Write) {
        // ANCHOR_END: Window-draw_into
        let mut inner = String::new();
        for widget in &self.widgets {
            widget.draw_into(&mut inner);
        }

        let inner_width = self.inner_width();

        // TODO: after learning about error handling, you can change
        // draw_into to return Result<(), std::fmt::Error>. Then use
        // the ?-operator here instead of .unwrap().
        writeln!(buffer, "+-{:-<inner_width$}-+", "").unwrap();
        writeln!(buffer, "| {:^inner_width$} |", &self.title).unwrap();
        writeln!(buffer, "+={:=<inner_width$}=+", "").unwrap();
        for line in inner.lines() {
            writeln!(buffer, "| {:inner_width$} |", line).unwrap();
        }
        writeln!(buffer, "+-{:-<inner_width$}-+", "").unwrap();
    }
}

// ---- src/main.rs ----
mod widgets;

use widgets::Widget;

fn main() {
    let mut window = widgets::Window::new("Rust GUI Demo 1.23");
    window
        .add_widget(Box::new(widgets::Label::new("This is a small text GUI demo.")));
    window.add_widget(Box::new(widgets::Button::new("Click me!")));
    window.draw();
}

Testing

This segment should take about 45 minutes. It contains:

Unit Tests

Rust and Cargo come with a simple unit test framework:

• Unit tests are supported throughout your code.

• Integration tests are supported via the tests/ directory.

Tests are marked with #[test]. Unit tests are often put in a nested tests module, using #[cfg(test)] to conditionally compile them only when building tests.

fn first_word(text: &str) -> &str {
    match text.find(' ') {
        Some(idx) => &text[..idx],
        None => &text,
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_empty() {
        assert_eq!(first_word(""), "");
    }

    #[test]
    fn test_single_word() {
        assert_eq!(first_word("Hello"), "Hello");
    }

    #[test]
    fn test_multiple_words() {
        assert_eq!(first_word("Hello World"), "Hello");
    }
}

• This lets you unit test private helpers.
• The #[cfg(test)] attribute is only active when you run cargo test.

Other Types of Tests

Integration Tests

If you want to test your library as a client, use an integration test.

Create a .rs file under tests/:

// tests/my_library.rs
use my_library::init;

#[test]
fn test_init() {
    assert!(init().is_ok());
}

These tests only have access to the public API of your crate.

Documentation Tests

Rust has built-in support for documentation tests:

#![allow(unused)]
fn main() {
/// Shortens a string to the given length.
///
/// ```
/// # use playground::shorten_string;
/// assert_eq!(shorten_string("Hello World", 5), "Hello");
/// assert_eq!(shorten_string("Hello World", 20), "Hello World");
/// ```
pub fn shorten_string(s: &str, length: usize) -> &str {
    &s[..std::cmp::min(length, s.len())]
}
}

• Code blocks in /// comments are automatically seen as Rust code.
• The code will be compiled and executed as part of cargo test.
• Adding # in the code will hide it from the docs, but will still compile/run it.
• Test the above code on the Rust Playground.

Compiler Lints and Clippy

The Rust compiler produces fantastic error messages, as well as helpful built-in lints. Clippy provides even more lints, organized into groups that can be enabled per-project.

#[deny(clippy::cast_possible_truncation)]
fn main() {
    let x = 3;
    while (x < 70000) {
        x *= 2;
    }
    println!("X probably fits in a u16, right? {}", x as u16);
}

Exercise: Luhn Algorithm

The Luhn algorithm is used to validate credit card numbers. The algorithm takes a string as input and does the following to validate the credit card number:

• Ignore all spaces. Reject numbers with fewer than two digits.

• Moving from right to left, double every second digit: for the number 1234, we double 3 and 1. For the number 98765, we double 6 and 8.

• After doubling a digit, sum the digits if the result is greater than 9. So doubling 7 becomes 14 which becomes 1 + 4 = 5.

• Sum all the undoubled and doubled digits.

• The credit card number is valid if the sum ends with 0.

The provided code provides a buggy implementation of the luhn algorithm, along with two basic unit tests that confirm that most of the algorithm is implemented correctly.

Copy the code below to https://play.rust-lang.org/ and write additional tests to uncover bugs in the provided implementation, fixing any bugs you find.

#![allow(unused)]
fn main() {
pub fn luhn(cc_number: &str) -> bool {
    let mut sum = 0;
    let mut double = false;

    for c in cc_number.chars().rev() {
        if let Some(digit) = c.to_digit(10) {
            if double {
                let double_digit = digit * 2;
                sum +=
                    if double_digit > 9 { double_digit - 9 } else { double_digit };
            } else {
                sum += digit;
            }
            double = !double;
        } else {
            continue;
        }
    }

    sum % 10 == 0
}

#[cfg(test)]
mod test {
    use super::*;

    #[test]
    fn test_valid_cc_number() {
        assert!(luhn("4263 9826 4026 9299"));
        assert!(luhn("4539 3195 0343 6467"));
        assert!(luhn("7992 7398 713"));
    }

    #[test]
    fn test_invalid_cc_number() {
        assert!(!luhn("4223 9826 4026 9299"));
        assert!(!luhn("4539 3195 0343 6476"));
        assert!(!luhn("8273 1232 7352 0569"));
    }
}
}

Solution

// This is the buggy version that appears in the problem.
#[cfg(never)]
pub fn luhn(cc_number: &str) -> bool {
    let mut sum = 0;
    let mut double = false;

    for c in cc_number.chars().rev() {
        if let Some(digit) = c.to_digit(10) {
            if double {
                let double_digit = digit * 2;
                sum +=
                    if double_digit > 9 { double_digit - 9 } else { double_digit };
            } else {
                sum += digit;
            }
            double = !double;
        } else {
            continue;
        }
    }

    sum % 10 == 0
}

// This is the solution and passes all of the tests below.
pub fn luhn(cc_number: &str) -> bool {
    let mut sum = 0;
    let mut double = false;
    let mut digits = 0;

    for c in cc_number.chars().rev() {
        if let Some(digit) = c.to_digit(10) {
            digits += 1;
            if double {
                let double_digit = digit * 2;
                sum +=
                    if double_digit > 9 { double_digit - 9 } else { double_digit };
            } else {
                sum += digit;
            }
            double = !double;
        } else if c.is_whitespace() {
            // New: accept whitespace.
            continue;
        } else {
            // New: reject all other characters.
            return false;
        }
    }

    // New: check that we have at least two digits
    digits >= 2 && sum % 10 == 0
}

fn main() {
    let cc_number = "1234 5678 1234 5670";
    println!(
        "Is {cc_number} a valid credit card number? {}",
        if luhn(cc_number) { "yes" } else { "no" }
    );
}

#[cfg(test)]
mod test {
    use super::*;

    #[test]
    fn test_valid_cc_number() {
        assert!(luhn("4263 9826 4026 9299"));
        assert!(luhn("4539 3195 0343 6467"));
        assert!(luhn("7992 7398 713"));
    }

    #[test]
    fn test_invalid_cc_number() {
        assert!(!luhn("4223 9826 4026 9299"));
        assert!(!luhn("4539 3195 0343 6476"));
        assert!(!luhn("8273 1232 7352 0569"));
    }

    #[test]
    fn test_non_digit_cc_number() {
        assert!(!luhn("foo"));
        assert!(!luhn("foo 0 0"));
    }

    #[test]
    fn test_empty_cc_number() {
        assert!(!luhn(""));
        assert!(!luhn(" "));
        assert!(!luhn("  "));
        assert!(!luhn("    "));
    }

    #[test]
    fn test_single_digit_cc_number() {
        assert!(!luhn("0"));
    }

    #[test]
    fn test_two_digit_cc_number() {
        assert!(luhn(" 0 0 "));
    }
}

Welcome Back

Including 10 minute breaks, this session should take about 2 hours and 20 minutes. It contains:

Error Handling

This segment should take about 1 hour and 5 minutes. It contains:

Panics

Rust handles fatal errors with a "panic".

Rust will trigger a panic if a fatal error happens at runtime:

fn main() {
    let v = vec![10, 20, 30];
    println!("v[100]: {}", v[100]);
}

• Panics are for unrecoverable and unexpected errors.
  • Panics are symptoms of bugs in the program.
  • Runtime failures like failed bounds checks can panic
  • Assertions (such as assert!) panic on failure
  • Purpose-specific panics can use the panic! macro.
• A panic will "unwind" the stack, dropping values just as if the functions had returned.
• Use non-panicking APIs (such as Vec::get) if crashing is not acceptable.

use std::panic;

fn main() {
    let result = panic::catch_unwind(|| "No problem here!");
    println!("{result:?}");

    let result = panic::catch_unwind(|| {
        panic!("oh no!");
    });
    println!("{result:?}");
}

Result

Our primary mechanism for error handling in Rust is the Result enum, which we briefly saw when discussing standard library types.

use std::fs::File;
use std::io::Read;

fn main() {
    let file: Result<File, std::io::Error> = File::open("diary.txt");
    match file {
        Ok(mut file) => {
            let mut contents = String::new();
            if let Ok(bytes) = file.read_to_string(&mut contents) {
                println!("Dear diary: {contents} ({bytes} bytes)");
            } else {
                println!("Could not read file content");
            }
        }
        Err(err) => {
            println!("The diary could not be opened: {err}");
        }
    }
}

Try Operator

Runtime errors like connection-refused or file-not-found are handled with the Result type, but matching this type on every call can be cumbersome. The try-operator ? is used to return errors to the caller. It lets you turn the common

match some_expression {
    Ok(value) => value,
    Err(err) => return Err(err),
}

into the much simpler

some_expression?

We can use this to simplify our error handling code:

use std::io::Read;
use std::{fs, io};

fn read_username(path: &str) -> Result<String, io::Error> {
    let username_file_result = fs::File::open(path);
    let mut username_file = match username_file_result {
        Ok(file) => file,
        Err(err) => return Err(err),
    };

    let mut username = String::new();
    match username_file.read_to_string(&mut username) {
        Ok(_) => Ok(username),
        Err(err) => Err(err),
    }
}

fn main() {
    //fs::write("config.dat", "alice").unwrap();
    let username = read_username("config.dat");
    println!("username or error: {username:?}");
}

Try Conversions

The effective expansion of ? is a little more complicated than previously indicated:

expression?

works the same as

match expression {
    Ok(value) => value,
    Err(err)  => return Err(From::from(err)),
}

The From::from call here means we attempt to convert the error type to the type returned by the function. This makes it easy to encapsulate errors into higher-level errors.

Example

use std::error::Error;
use std::io::Read;
use std::{fmt, fs, io};

#[derive(Debug)]
enum ReadUsernameError {
    IoError(io::Error),
    EmptyUsername(String),
}

impl Error for ReadUsernameError {}

impl fmt::Display for ReadUsernameError {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        match self {
            Self::IoError(e) => write!(f, "I/O error: {e}"),
            Self::EmptyUsername(path) => write!(f, "Found no username in {path}"),
        }
    }
}

impl From<io::Error> for ReadUsernameError {
    fn from(err: io::Error) -> Self {
        Self::IoError(err)
    }
}

fn read_username(path: &str) -> Result<String, ReadUsernameError> {
    let mut username = String::with_capacity(100);
    fs::File::open(path)?.read_to_string(&mut username)?;
    if username.is_empty() {
        return Err(ReadUsernameError::EmptyUsername(String::from(path)));
    }
    Ok(username)
}

fn main() {
    //std::fs::write("config.dat", "").unwrap();
    let username = read_username("config.dat");
    println!("username or error: {username:?}");
}

Dynamic Error Types

Sometimes we want to allow any type of error to be returned without writing our own enum covering all the different possibilities. The std::error::Error trait makes it easy to create a trait object that can contain any error.

use std::error::Error;
use std::fs;
use std::io::Read;

fn read_count(path: &str) -> Result<i32, Box<dyn Error>> {
    let mut count_str = String::new();
    fs::File::open(path)?.read_to_string(&mut count_str)?;
    let count: i32 = count_str.parse()?;
    Ok(count)
}

fn main() {
    fs::write("count.dat", "1i3").unwrap();
    match read_count("count.dat") {
        Ok(count) => println!("Count: {count}"),
        Err(err) => println!("Error: {err}"),
    }
}

thiserror

The thiserror crate provides macros to help avoid boilerplate when defining error types. It provides derive macros that assist in implementing From<T>, Display, and the Error trait.

use std::io::Read;
use std::{fs, io};
use thiserror::Error;

#[derive(Debug, Error)]
enum ReadUsernameError {
    #[error("I/O error: {0}")]
    IoError(#[from] io::Error),
    #[error("Found no username in {0}")]
    EmptyUsername(String),
}

fn read_username(path: &str) -> Result<String, ReadUsernameError> {
    let mut username = String::with_capacity(100);
    fs::File::open(path)?.read_to_string(&mut username)?;
    if username.is_empty() {
        return Err(ReadUsernameError::EmptyUsername(String::from(path)));
    }
    Ok(username)
}

fn main() {
    //fs::write("config.dat", "").unwrap();
    match read_username("config.dat") {
        Ok(username) => println!("Username: {username}"),
        Err(err) => println!("Error: {err:?}"),
    }
}

anyhow

The anyhow crate provides a rich error type with support for carrying additional contextual information, which can be used to provide a semantic trace of what the program was doing leading up to the error.

This can be combined with the convenience macros from thiserror to avoid writing out trait impls explicitly for custom error types.

use anyhow::{bail, Context, Result};
use std::fs;
use std::io::Read;
use thiserror::Error;

#[derive(Clone, Debug, Eq, Error, PartialEq)]
#[error("Found no username in {0}")]
struct EmptyUsernameError(String);

fn read_username(path: &str) -> Result<String> {
    let mut username = String::with_capacity(100);
    fs::File::open(path)
        .with_context(|| format!("Failed to open {path}"))?
        .read_to_string(&mut username)
        .context("Failed to read")?;
    if username.is_empty() {
        bail!(EmptyUsernameError(path.to_string()));
    }
    Ok(username)
}

fn main() {
    //fs::write("config.dat", "").unwrap();
    match read_username("config.dat") {
        Ok(username) => println!("Username: {username}"),
        Err(err) => println!("Error: {err:?}"),
    }
}

Exercise: Rewriting with Result

The following implements a very simple parser for an expression language. However, it handles errors by panicking. Rewrite it to instead use idiomatic error handling and propagate errors to a return from main. Feel free to use thiserror and anyhow.

Hint: start by fixing error handling in the parse function. Once that is working correctly, update Tokenizer to implement Iterator<Item=Result<Token, TokenizerError>> and handle that in the parser.

use std::iter::Peekable;
use std::str::Chars;

/// An arithmetic operator.
#[derive(Debug, PartialEq, Clone, Copy)]
enum Op {
    Add,
    Sub,
}

/// A token in the expression language.
#[derive(Debug, PartialEq)]
enum Token {
    Number(String),
    Identifier(String),
    Operator(Op),
}

/// An expression in the expression language.
#[derive(Debug, PartialEq)]
enum Expression {
    /// A reference to a variable.
    Var(String),
    /// A literal number.
    Number(u32),
    /// A binary operation.
    Operation(Box<Expression>, Op, Box<Expression>),
}

fn tokenize(input: &str) -> Tokenizer {
    return Tokenizer(input.chars().peekable());
}

struct Tokenizer<'a>(Peekable<Chars<'a>>);

impl<'a> Tokenizer<'a> {
    fn collect_number(&mut self, first_char: char) -> Token {
        let mut num = String::from(first_char);
        while let Some(&c @ '0'..='9') = self.0.peek() {
            num.push(c);
            self.0.next();
        }
        Token::Number(num)
    }

    fn collect_identifier(&mut self, first_char: char) -> Token {
        let mut ident = String::from(first_char);
        while let Some(&c @ ('a'..='z' | '_' | '0'..='9')) = self.0.peek() {
            ident.push(c);
            self.0.next();
        }
        Token::Identifier(ident)
    }
}

impl<'a> Iterator for Tokenizer<'a> {
    type Item = Token;

    fn next(&mut self) -> Option<Token> {
        let c = self.0.next()?;
        match c {
            '0'..='9' => Some(self.collect_number(c)),
            'a'..='z' => Some(self.collect_identifier(c)),
            '+' => Some(Token::Operator(Op::Add)),
            '-' => Some(Token::Operator(Op::Sub)),
            _ => panic!("Unexpected character {c}"),
        }
    }
}

fn parse(input: &str) -> Expression {
    let mut tokens = tokenize(input);

    fn parse_expr<'a>(tokens: &mut Tokenizer<'a>) -> Expression {
        let Some(tok) = tokens.next() else {
            panic!("Unexpected end of input");
        };
        let expr = match tok {
            Token::Number(num) => {
                let v = num.parse().expect("Invalid 32-bit integer");
                Expression::Number(v)
            }
            Token::Identifier(ident) => Expression::Var(ident),
            Token::Operator(_) => panic!("Unexpected token {tok:?}"),
        };
        // Look ahead to parse a binary operation if present.
        match tokens.next() {
            None => expr,
            Some(Token::Operator(op)) => Expression::Operation(
                Box::new(expr),
                op,
                Box::new(parse_expr(tokens)),
            ),
            Some(tok) => panic!("Unexpected token {tok:?}"),
        }
    }

    parse_expr(&mut tokens)
}

fn main() {
    let expr = parse("10+foo+20-30");
    println!("{expr:?}");
}

Solution

use thiserror::Error;
use std::iter::Peekable;
use std::str::Chars;

/// An arithmetic operator.
#[derive(Debug, PartialEq, Clone, Copy)]
enum Op {
    Add,
    Sub,
}

/// A token in the expression language.
#[derive(Debug, PartialEq)]
enum Token {
    Number(String),
    Identifier(String),
    Operator(Op),
}

/// An expression in the expression language.
#[derive(Debug, PartialEq)]
enum Expression {
    /// A reference to a variable.
    Var(String),
    /// A literal number.
    Number(u32),
    /// A binary operation.
    Operation(Box<Expression>, Op, Box<Expression>),
}

fn tokenize(input: &str) -> Tokenizer {
    return Tokenizer(input.chars().peekable());
}

#[derive(Debug, Error)]
enum TokenizerError {
    #[error("Unexpected character '{0}' in input")]
    UnexpectedCharacter(char),
}

struct Tokenizer<'a>(Peekable<Chars<'a>>);

impl<'a> Tokenizer<'a> {
    fn collect_number(&mut self, first_char: char) -> Token {
        let mut num = String::from(first_char);
        while let Some(&c @ '0'..='9') = self.0.peek() {
            num.push(c);
            self.0.next();
        }
        Token::Number(num)
    }

    fn collect_identifier(&mut self, first_char: char) -> Token {
        let mut ident = String::from(first_char);
        while let Some(&c @ ('a'..='z' | '_' | '0'..='9')) = self.0.peek() {
            ident.push(c);
            self.0.next();
        }
        Token::Identifier(ident)
    }
}

impl<'a> Iterator for Tokenizer<'a> {
    type Item = Result<Token, TokenizerError>;

    fn next(&mut self) -> Option<Result<Token, TokenizerError>> {
        let c = self.0.next()?;
        match c {
            '0'..='9' => Some(Ok(self.collect_number(c))),
            'a'..='z' | '_' => Some(Ok(self.collect_identifier(c))),
            '+' => Some(Ok(Token::Operator(Op::Add))),
            '-' => Some(Ok(Token::Operator(Op::Sub))),
            _ => Some(Err(TokenizerError::UnexpectedCharacter(c))),
        }
    }
}

#[derive(Debug, Error)]
enum ParserError {
    #[error("Tokenizer error: {0}")]
    TokenizerError(#[from] TokenizerError),
    #[error("Unexpected end of input")]
    UnexpectedEOF,
    #[error("Unexpected token {0:?}")]
    UnexpectedToken(Token),
    #[error("Invalid number")]
    InvalidNumber(#[from] std::num::ParseIntError),
}

fn parse(input: &str) -> Result<Expression, ParserError> {
    let mut tokens = tokenize(input);

    fn parse_expr<'a>(
        tokens: &mut Tokenizer<'a>,
    ) -> Result<Expression, ParserError> {
        let tok = tokens.next().ok_or(ParserError::UnexpectedEOF)??;
        let expr = match tok {
            Token::Number(num) => {
                let v = num.parse()?;
                Expression::Number(v)
            }
            Token::Identifier(ident) => Expression::Var(ident),
            Token::Operator(_) => return Err(ParserError::UnexpectedToken(tok)),
        };
        // Look ahead to parse a binary operation if present.
        Ok(match tokens.next() {
            None => expr,
            Some(Ok(Token::Operator(op))) => Expression::Operation(
                Box::new(expr),
                op,
                Box::new(parse_expr(tokens)?),
            ),
            Some(Err(e)) => return Err(e.into()),
            Some(Ok(tok)) => return Err(ParserError::UnexpectedToken(tok)),
        })
    }

    parse_expr(&mut tokens)
}

fn main() -> anyhow::Result<()> {
    let expr = parse("10+foo+20-30")?;
    println!("{expr:?}");
    Ok(())
}

Unsafe Rust

This segment should take about 1 hour and 5 minutes. It contains:

Unsafe Rust

The Rust language has two parts:

• Safe Rust: memory safe, no undefined behavior possible.
• Unsafe Rust: can trigger undefined behavior if preconditions are violated.

We saw mostly safe Rust in this course, but it's important to know what Unsafe Rust is.

Unsafe code is usually small and isolated, and its correctness should be carefully documented. It is usually wrapped in a safe abstraction layer.

Unsafe Rust gives you access to five new capabilities:

• Dereference raw pointers.
• Access or modify mutable static variables.
• Access union fields.
• Call unsafe functions, including extern functions.
• Implement unsafe traits.

We will briefly cover unsafe capabilities next. For full details, please see Chapter 19.1 in the Rust Book and the Rustonomicon.

Dereferencing Raw Pointers

Creating pointers is safe, but dereferencing them requires unsafe:

fn main() {
    let mut s = String::from("careful!");

    let r1 = &raw mut s;
    let r2 = r1 as *const String;

    // SAFETY: r1 and r2 were obtained from references and so are guaranteed to
    // be non-null and properly aligned, the objects underlying the references
    // from which they were obtained are live throughout the whole unsafe
    // block, and they are not accessed either through the references or
    // concurrently through any other pointers.
    unsafe {
        println!("r1 is: {}", *r1);
        *r1 = String::from("uhoh");
        println!("r2 is: {}", *r2);
    }

    // NOT SAFE. DO NOT DO THIS.
    /*
    let r3: &String = unsafe { &*r1 };
    drop(s);
    println!("r3 is: {}", *r3);
    */
}

Mutable Static Variables

It is safe to read an immutable static variable:

static HELLO_WORLD: &str = "Hello, world!";

fn main() {
    println!("HELLO_WORLD: {HELLO_WORLD}");
}

However, since data races can occur, it is unsafe to read and write mutable static variables:

static mut COUNTER: u32 = 0;

fn add_to_counter(inc: u32) {
    // SAFETY: There are no other threads which could be accessing `COUNTER`.
    unsafe {
        COUNTER += inc;
    }
}

fn main() {
    add_to_counter(42);

    // SAFETY: There are no other threads which could be accessing `COUNTER`.
    unsafe {
        println!("COUNTER: {COUNTER}");
    }
}

Unions

Unions are like enums, but you need to track the active field yourself:

#[repr(C)]
union MyUnion {
    i: u8,
    b: bool,
}

fn main() {
    let u = MyUnion { i: 42 };
    println!("int: {}", unsafe { u.i });
    println!("bool: {}", unsafe { u.b }); // Undefined behavior!
}

Unsafe Functions

Calling Unsafe Functions

A function or method can be marked unsafe if it has extra preconditions you must uphold to avoid undefined behaviour:

extern "C" {
    fn abs(input: i32) -> i32;
}

fn main() {
    let emojis = "🗻∈🌏";

    // SAFETY: The indices are in the correct order, within the bounds of the
    // string slice, and lie on UTF-8 sequence boundaries.
    unsafe {
        println!("emoji: {}", emojis.get_unchecked(0..4));
        println!("emoji: {}", emojis.get_unchecked(4..7));
        println!("emoji: {}", emojis.get_unchecked(7..11));
    }

    println!("char count: {}", count_chars(unsafe { emojis.get_unchecked(0..7) }));

    // SAFETY: `abs` doesn't deal with pointers and doesn't have any safety
    // requirements.
    unsafe {
        println!("Absolute value of -3 according to C: {}", abs(-3));
    }

    // Not upholding the UTF-8 encoding requirement breaks memory safety!
    // println!("emoji: {}", unsafe { emojis.get_unchecked(0..3) });
    // println!("char count: {}", count_chars(unsafe {
    // emojis.get_unchecked(0..3) }));
}

fn count_chars(s: &str) -> usize {
    s.chars().count()
}

Writing Unsafe Functions

You can mark your own functions as unsafe if they require particular conditions to avoid undefined behaviour.

/// Swaps the values pointed to by the given pointers.
///
/// # Safety
///
/// The pointers must be valid and properly aligned.
unsafe fn swap(a: *mut u8, b: *mut u8) {
    let temp = *a;
    *a = *b;
    *b = temp;
}

fn main() {
    let mut a = 42;
    let mut b = 66;

    // SAFETY: ...
    unsafe {
        swap(&mut a, &mut b);
    }

    println!("a = {}, b = {}", a, b);
}

Implementing Unsafe Traits

Like with functions, you can mark a trait as unsafe if the implementation must guarantee particular conditions to avoid undefined behaviour.

For example, the zerocopy crate has an unsafe trait that looks something like this:

use std::{mem, slice};

/// ...
/// # Safety
/// The type must have a defined representation and no padding.
pub unsafe trait IntoBytes {
    fn as_bytes(&self) -> &[u8] {
        let len = mem::size_of_val(self);
        unsafe { slice::from_raw_parts((&raw const self).cast::<u8>(), len) }
    }
}

// SAFETY: `u32` has a defined representation and no padding.
unsafe impl IntoBytes for u32 {}

Safe FFI Wrapper

Rust has great support for calling functions through a foreign function interface (FFI). We will use this to build a safe wrapper for the libc functions you would use from C to read the names of files in a directory.

You will want to consult the manual pages:

• opendir(3)
• readdir(3)
• closedir(3)

You will also want to browse the std::ffi module. There you find a number of string types which you need for the exercise:

You will convert between all these types:

• &str to CString: you need to allocate space for a trailing \0 character,
• CString to *const i8: you need a pointer to call C functions,
• *const i8 to &CStr: you need something which can find the trailing \0 character,
• &CStr to &[u8]: a slice of bytes is the universal interface for "some unknown data",
• &[u8] to &OsStr: &OsStr is a step towards OsString, use OsStrExt to create it,
• &OsStr to OsString: you need to clone the data in &OsStr to be able to return it and call readdir again.

The Nomicon also has a very useful chapter about FFI.

Copy the code below to https://play.rust-lang.org/ and fill in the missing functions and methods:

// TODO: remove this when you're done with your implementation.
#![allow(unused_imports, unused_variables, dead_code)]

mod ffi {
    use std::os::raw::{c_char, c_int};
    #[cfg(not(target_os = "macos"))]
    use std::os::raw::{c_long, c_uchar, c_ulong, c_ushort};

    // Opaque type. See https://doc.rust-lang.org/nomicon/ffi.html.
    #[repr(C)]
    pub struct DIR {
        _data: [u8; 0],
        _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
    }

    // Layout according to the Linux man page for readdir(3), where ino_t and
    // off_t are resolved according to the definitions in
    // /usr/include/x86_64-linux-gnu/{sys/types.h, bits/typesizes.h}.
    #[cfg(not(target_os = "macos"))]
    #[repr(C)]
    pub struct dirent {
        pub d_ino: c_ulong,
        pub d_off: c_long,
        pub d_reclen: c_ushort,
        pub d_type: c_uchar,
        pub d_name: [c_char; 256],
    }

    // Layout according to the macOS man page for dir(5).
    #[cfg(all(target_os = "macos"))]
    #[repr(C)]
    pub struct dirent {
        pub d_fileno: u64,
        pub d_seekoff: u64,
        pub d_reclen: u16,
        pub d_namlen: u16,
        pub d_type: u8,
        pub d_name: [c_char; 1024],
    }

    unsafe extern "C" {
        pub unsafe fn opendir(s: *const c_char) -> *mut DIR;

        #[cfg(not(all(target_os = "macos", target_arch = "x86_64")))]
        pub unsafe fn readdir(s: *mut DIR) -> *const dirent;

        // See https://github.com/rust-lang/libc/issues/414 and the section on
        // _DARWIN_FEATURE_64_BIT_INODE in the macOS man page for stat(2).
        //
        // "Platforms that existed before these updates were available" refers
        // to macOS (as opposed to iOS / wearOS / etc.) on Intel and PowerPC.
        #[cfg(all(target_os = "macos", target_arch = "x86_64"))]
        #[link_name = "readdir$INODE64"]
        pub unsafe fn readdir(s: *mut DIR) -> *const dirent;

        pub unsafe fn closedir(s: *mut DIR) -> c_int;
    }
}

use std::ffi::{CStr, CString, OsStr, OsString};
use std::os::unix::ffi::OsStrExt;

#[derive(Debug)]
struct DirectoryIterator {
    path: CString,
    dir: *mut ffi::DIR,
}

impl DirectoryIterator {
    fn new(path: &str) -> Result<DirectoryIterator, String> {
        // Call opendir and return a Ok value if that worked,
        // otherwise return Err with a message.
        unimplemented!()
    }
}

impl Iterator for DirectoryIterator {
    type Item = OsString;
    fn next(&mut self) -> Option<OsString> {
        // Keep calling readdir until we get a NULL pointer back.
        unimplemented!()
    }
}

impl Drop for DirectoryIterator {
    fn drop(&mut self) {
        // Call closedir as needed.
        unimplemented!()
    }
}

fn main() -> Result<(), String> {
    let iter = DirectoryIterator::new(".")?;
    println!("files: {:#?}", iter.collect::<Vec<_>>());
    Ok(())
}

Solution

mod ffi {
    use std::os::raw::{c_char, c_int};
    #[cfg(not(target_os = "macos"))]
    use std::os::raw::{c_long, c_uchar, c_ulong, c_ushort};

    // Opaque type. See https://doc.rust-lang.org/nomicon/ffi.html.
    #[repr(C)]
    pub struct DIR {
        _data: [u8; 0],
        _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
    }

    // Layout according to the Linux man page for readdir(3), where ino_t and
    // off_t are resolved according to the definitions in
    // /usr/include/x86_64-linux-gnu/{sys/types.h, bits/typesizes.h}.
    #[cfg(not(target_os = "macos"))]
    #[repr(C)]
    pub struct dirent {
        pub d_ino: c_ulong,
        pub d_off: c_long,
        pub d_reclen: c_ushort,
        pub d_type: c_uchar,
        pub d_name: [c_char; 256],
    }

    // Layout according to the macOS man page for dir(5).
    #[cfg(all(target_os = "macos"))]
    #[repr(C)]
    pub struct dirent {
        pub d_fileno: u64,
        pub d_seekoff: u64,
        pub d_reclen: u16,
        pub d_namlen: u16,
        pub d_type: u8,
        pub d_name: [c_char; 1024],
    }

    unsafe extern "C" {
        pub unsafe fn opendir(s: *const c_char) -> *mut DIR;

        #[cfg(not(all(target_os = "macos", target_arch = "x86_64")))]
        pub unsafe fn readdir(s: *mut DIR) -> *const dirent;

        // See https://github.com/rust-lang/libc/issues/414 and the section on
        // _DARWIN_FEATURE_64_BIT_INODE in the macOS man page for stat(2).
        //
        // "Platforms that existed before these updates were available" refers
        // to macOS (as opposed to iOS / wearOS / etc.) on Intel and PowerPC.
        #[cfg(all(target_os = "macos", target_arch = "x86_64"))]
        #[link_name = "readdir$INODE64"]
        pub unsafe fn readdir(s: *mut DIR) -> *const dirent;

        pub unsafe fn closedir(s: *mut DIR) -> c_int;
    }
}

use std::ffi::{CStr, CString, OsStr, OsString};
use std::os::unix::ffi::OsStrExt;

#[derive(Debug)]
struct DirectoryIterator {
    path: CString,
    dir: *mut ffi::DIR,
}

impl DirectoryIterator {
    fn new(path: &str) -> Result<DirectoryIterator, String> {
        // Call opendir and return a Ok value if that worked,
        // otherwise return Err with a message.
        let path =
            CString::new(path).map_err(|err| format!("Invalid path: {err}"))?;
        // SAFETY: path.as_ptr() cannot be NULL.
        let dir = unsafe { ffi::opendir(path.as_ptr()) };
        if dir.is_null() {
            Err(format!("Could not open {path:?}"))
        } else {
            Ok(DirectoryIterator { path, dir })
        }
    }
}

impl Iterator for DirectoryIterator {
    type Item = OsString;
    fn next(&mut self) -> Option<OsString> {
        // Keep calling readdir until we get a NULL pointer back.
        // SAFETY: self.dir is never NULL.
        let dirent = unsafe { ffi::readdir(self.dir) };
        if dirent.is_null() {
            // We have reached the end of the directory.
            return None;
        }
        // SAFETY: dirent is not NULL and dirent.d_name is NUL
        // terminated.
        let d_name = unsafe { CStr::from_ptr((*dirent).d_name.as_ptr()) };
        let os_str = OsStr::from_bytes(d_name.to_bytes());
        Some(os_str.to_owned())
    }
}

impl Drop for DirectoryIterator {
    fn drop(&mut self) {
        // Call closedir as needed.
        // SAFETY: self.dir is never NULL.
        if unsafe { ffi::closedir(self.dir) } != 0 {
            panic!("Could not close {:?}", self.path);
        }
    }
}

fn main() -> Result<(), String> {
    let iter = DirectoryIterator::new(".")?;
    println!("files: {:#?}", iter.collect::<Vec<_>>());
    Ok(())
}

#[cfg(test)]
mod tests {
    use super::*;
    use std::error::Error;

    #[test]
    fn test_nonexisting_directory() {
        let iter = DirectoryIterator::new("no-such-directory");
        assert!(iter.is_err());
    }

    #[test]
    fn test_empty_directory() -> Result<(), Box<dyn Error>> {
        let tmp = tempfile::TempDir::new()?;
        let iter = DirectoryIterator::new(
            tmp.path().to_str().ok_or("Non UTF-8 character in path")?,
        )?;
        let mut entries = iter.collect::<Vec<_>>();
        entries.sort();
        assert_eq!(entries, &[".", ".."]);
        Ok(())
    }

    #[test]
    fn test_nonempty_directory() -> Result<(), Box<dyn Error>> {
        let tmp = tempfile::TempDir::new()?;
        std::fs::write(tmp.path().join("foo.txt"), "The Foo Diaries\n")?;
        std::fs::write(tmp.path().join("bar.png"), "<PNG>\n")?;
        std::fs::write(tmp.path().join("crab.rs"), "//! Crab\n")?;
        let iter = DirectoryIterator::new(
            tmp.path().to_str().ok_or("Non UTF-8 character in path")?,
        )?;
        let mut entries = iter.collect::<Vec<_>>();
        entries.sort();
        assert_eq!(entries, &[".", "..", "bar.png", "crab.rs", "foo.txt"]);
        Ok(())
    }
}

Welcome to Rust in Android

Rust is supported for system software on Android. This means that you can write new services, libraries, drivers or even firmware in Rust (or improve existing code as needed).

Setup

We will be using a Cuttlefish Android Virtual Device to test our code. Make sure you have access to one or create a new one with:

source build/envsetup.sh
lunch aosp_cf_x86_64_phone-trunk_staging-userdebug
acloud create

Please see the Android Developer Codelab for details.

The code on the following pages can be found in the src/android/ directory of the course material. Please git clone the repository to follow along.

Build Rules

The Android build system (Soong) supports Rust via a number of modules:

We will look at rust_binary and rust_library next.

Rust Binaries

Let us start with a simple application. At the root of an AOSP checkout, create the following files:

hello_rust/Android.bp:

rust_binary {
    name: "hello_rust",
    crate_name: "hello_rust",
    srcs: ["src/main.rs"],
}

hello_rust/src/main.rs:

//! Rust demo.

/// Prints a greeting to standard output.
fn main() {
    println!("Hello from Rust!");
}

You can now build, push, and run the binary:

m hello_rust
adb push "$ANDROID_PRODUCT_OUT/system/bin/hello_rust" /data/local/tmp
adb shell /data/local/tmp/hello_rust

Hello from Rust!

Rust Libraries

You use rust_library to create a new Rust library for Android.

Here we declare a dependency on two libraries:

• libgreeting, which we define below,
• libtextwrap, which is a crate already vendored in external/rust/crates/.

hello_rust/Android.bp:

rust_binary {
    name: "hello_rust_with_dep",
    crate_name: "hello_rust_with_dep",
    srcs: ["src/main.rs"],
    rustlibs: [
        "libgreetings",
        "libtextwrap",
    ],
    prefer_rlib: true, // Need this to avoid dynamic link error.
}

rust_library {
    name: "libgreetings",
    crate_name: "greetings",
    srcs: ["src/lib.rs"],
}

hello_rust/src/main.rs:

//! Rust demo.

use greetings::greeting;
use textwrap::fill;

/// Prints a greeting to standard output.
fn main() {
    println!("{}", fill(&greeting("Bob"), 24));
}

hello_rust/src/lib.rs:

//! Greeting library.

/// Greet `name`.
pub fn greeting(name: &str) -> String {
    format!("Hello {name}, it is very nice to meet you!")
}

You build, push, and run the binary like before:

m hello_rust_with_dep
adb push "$ANDROID_PRODUCT_OUT/system/bin/hello_rust_with_dep" /data/local/tmp
adb shell /data/local/tmp/hello_rust_with_dep

Hello Bob, it is very
nice to meet you!

AIDL

The Android Interface Definition Language (AIDL) is supported in Rust:

• Rust code can call existing AIDL servers,
• You can create new AIDL servers in Rust.

Birthday Service Tutorial

To illustrate how to use Rust with Binder, we're going to walk through the process of creating a Binder interface. We're then going to both implement the described service and write client code that talks to that service.

AIDL Interfaces

You declare the API of your service using an AIDL interface:

birthday_service/aidl/com/example/birthdayservice/IBirthdayService.aidl:

package com.example.birthdayservice;

/** Birthday service interface. */
interface IBirthdayService {
    /** Generate a Happy Birthday message. */
    String wishHappyBirthday(String name, int years);
}

birthday_service/aidl/Android.bp:

aidl_interface {
    name: "com.example.birthdayservice",
    srcs: ["com/example/birthdayservice/*.aidl"],
    unstable: true,
    backend: {
        rust: { // Rust is not enabled by default
            enabled: true,
        },
    },
}

Generated Service API

Binder generates a trait for each interface definition.

birthday_service/aidl/com/example/birthdayservice/IBirthdayService.aidl:

/** Birthday service interface. */
interface IBirthdayService {
    /** Generate a Happy Birthday message. */
    String wishHappyBirthday(String name, int years);
}

out/soong/.intermediates/.../birthdayservice/IBirthdayService.rs:

trait IBirthdayService {
    fn wishHappyBirthday(&self, name: &str, years: i32) -> binder::Result<String>;
}

Your service will need to implement this trait, and your client will use this trait to talk to the service.

Service Implementation

We can now implement the AIDL service:

birthday_service/src/lib.rs:

use com_example_birthdayservice::aidl::com::example::birthdayservice::IBirthdayService::IBirthdayService;
use com_example_birthdayservice::binder;

/// The `IBirthdayService` implementation.
pub struct BirthdayService;

impl binder::Interface for BirthdayService {}

impl IBirthdayService for BirthdayService {
    fn wishHappyBirthday(&self, name: &str, years: i32) -> binder::Result<String> {
        Ok(format!("Happy Birthday {name}, congratulations with the {years} years!"))
    }
}

birthday_service/Android.bp:

rust_library {
    name: "libbirthdayservice",
    srcs: ["src/lib.rs"],
    crate_name: "birthdayservice",
    rustlibs: [
        "com.example.birthdayservice-rust",
        "libbinder_rs",
    ],
}

AIDL Server

Finally, we can create a server which exposes the service:

birthday_service/src/server.rs:

//! Birthday service.
use birthdayservice::BirthdayService;
use com_example_birthdayservice::aidl::com::example::birthdayservice::IBirthdayService::BnBirthdayService;
use com_example_birthdayservice::binder;

const SERVICE_IDENTIFIER: &str = "birthdayservice";

/// Entry point for birthday service.
fn main() {
    let birthday_service = BirthdayService;
    let birthday_service_binder = BnBirthdayService::new_binder(
        birthday_service,
        binder::BinderFeatures::default(),
    );
    binder::add_service(SERVICE_IDENTIFIER, birthday_service_binder.as_binder())
        .expect("Failed to register service");
    binder::ProcessState::join_thread_pool();
}

birthday_service/Android.bp:

rust_binary {
    name: "birthday_server",
    crate_name: "birthday_server",
    srcs: ["src/server.rs"],
    rustlibs: [
        "com.example.birthdayservice-rust",
        "libbinder_rs",
        "libbirthdayservice",
    ],
    prefer_rlib: true, // To avoid dynamic link error.
}

Deploy

We can now build, push, and start the service:

m birthday_server
adb push "$ANDROID_PRODUCT_OUT/system/bin/birthday_server" /data/local/tmp
adb root
adb shell /data/local/tmp/birthday_server

In another terminal, check that the service runs:

adb shell service check birthdayservice

Service birthdayservice: found

You can also call the service with service call:

adb shell service call birthdayservice 1 s16 Bob i32 24

Result: Parcel(
  0x00000000: 00000000 00000036 00610048 00700070 '....6...H.a.p.p.'
  0x00000010: 00200079 00690042 00740072 00640068 'y. .B.i.r.t.h.d.'
  0x00000020: 00790061 00420020 0062006f 0020002c 'a.y. .B.o.b.,. .'
  0x00000030: 006f0063 0067006e 00610072 00750074 'c.o.n.g.r.a.t.u.'
  0x00000040: 0061006c 00690074 006e006f 00200073 'l.a.t.i.o.n.s. .'
  0x00000050: 00690077 00680074 00740020 00650068 'w.i.t.h. .t.h.e.'
  0x00000060: 00320020 00200034 00650079 00720061 ' .2.4. .y.e.a.r.'
  0x00000070: 00210073 00000000                   's.!.....        ')

AIDL Client

Finally, we can create a Rust client for our new service.

birthday_service/src/client.rs:

use com_example_birthdayservice::aidl::com::example::birthdayservice::IBirthdayService::IBirthdayService;
use com_example_birthdayservice::binder;

const SERVICE_IDENTIFIER: &str = "birthdayservice";

/// Call the birthday service.
fn main() -> Result<(), Box<dyn Error>> {
    let name = std::env::args().nth(1).unwrap_or_else(|| String::from("Bob"));
    let years = std::env::args()
        .nth(2)
        .and_then(|arg| arg.parse::<i32>().ok())
        .unwrap_or(42);

    binder::ProcessState::start_thread_pool();
    let service = binder::get_interface::<dyn IBirthdayService>(SERVICE_IDENTIFIER)
        .map_err(|_| "Failed to connect to BirthdayService")?;

    // Call the service.
    let msg = service.wishHappyBirthday(&name, years)?;
    println!("{msg}");
}

birthday_service/Android.bp:

rust_binary {
    name: "birthday_client",
    crate_name: "birthday_client",
    srcs: ["src/client.rs"],
    rustlibs: [
        "com.example.birthdayservice-rust",
        "libbinder_rs",
    ],
    prefer_rlib: true, // To avoid dynamic link error.
}

Notice that the client does not depend on libbirthdayservice.

Build, push, and run the client on your device:

m birthday_client
adb push "$ANDROID_PRODUCT_OUT/system/bin/birthday_client" /data/local/tmp
adb shell /data/local/tmp/birthday_client Charlie 60

Happy Birthday Charlie, congratulations with the 60 years!

Changing API

Let us extend the API with more functionality: we want to let clients specify a list of lines for the birthday card:

package com.example.birthdayservice;

/** Birthday service interface. */
interface IBirthdayService {
    /** Generate a Happy Birthday message. */
    String wishHappyBirthday(String name, int years, in String[] text);
}

This results in an updated trait definition for IBirthdayService:

trait IBirthdayService {
    fn wishHappyBirthday(
        &self,
        name: &str,
        years: i32,
        text: &[String],
    ) -> binder::Result<String>;
}

Updating Client and Service

Update the client and server code to account for the new API.

birthday_service/src/lib.rs:

impl IBirthdayService for BirthdayService {
    fn wishHappyBirthday(
        &self,
        name: &str,
        years: i32,
        text: &[String],
    ) -> binder::Result<String> {
        let mut msg = format!(
            "Happy Birthday {name}, congratulations with the {years} years!",
        );

        for line in text {
            msg.push('\n');
            msg.push_str(line);
        }

        Ok(msg)
    }
}

birthday_service/src/client.rs:

let msg = service.wishHappyBirthday(
    &name,
    years,
    &[
        String::from("Habby birfday to yuuuuu"),
        String::from("And also: many more"),
    ],
)?;

Working With AIDL Types

AIDL types translate into the appropriate idiomatic Rust type:

• Primitive types map (mostly) to idiomatic Rust types.
• Collection types like slices, Vecs and string types are supported.
• References to AIDL objects and file handles can be sent between clients and services.
• File handles and parcelables are fully supported.

Primitive Types

Primitive types map (mostly) idiomatically:

Array Types

The array types (T[], byte[], and List<T>) get translated to the appropriate Rust array type depending on how they are used in the function signature:

Sending Objects

AIDL objects can be sent either as a concrete AIDL type or as the type-erased IBinder interface:

birthday_service/aidl/com/example/birthdayservice/IBirthdayInfoProvider.aidl:

package com.example.birthdayservice;

interface IBirthdayInfoProvider {
    String name();
    int years();
}

birthday_service/aidl/com/example/birthdayservice/IBirthdayService.aidl:

import com.example.birthdayservice.IBirthdayInfoProvider;

interface IBirthdayService {
    /** The same thing, but using a binder object. */
    String wishWithProvider(IBirthdayInfoProvider provider);

    /** The same thing, but using `IBinder`. */
    String wishWithErasedProvider(IBinder provider);
}

birthday_service/src/client.rs:

/// Rust struct implementing the `IBirthdayInfoProvider` interface.
struct InfoProvider {
    name: String,
    age: u8,
}

impl binder::Interface for InfoProvider {}

impl IBirthdayInfoProvider for InfoProvider {
    fn name(&self) -> binder::Result<String> {
        Ok(self.name.clone())
    }

    fn years(&self) -> binder::Result<i32> {
        Ok(self.age as i32)
    }
}

fn main() {
    binder::ProcessState::start_thread_pool();
    let service = connect().expect("Failed to connect to BirthdayService");

    // Create a binder object for the `IBirthdayInfoProvider` interface.
    let provider = BnBirthdayInfoProvider::new_binder(
        InfoProvider { name: name.clone(), age: years as u8 },
        BinderFeatures::default(),
    );

    // Send the binder object to the service.
    service.wishWithProvider(&provider)?;

    // Perform the same operation but passing the provider as an `SpIBinder`.
    service.wishWithErasedProvider(&provider.as_binder())?;
}

Parcelables

Binder for Rust supports sending parcelables directly:

birthday_service/aidl/com/example/birthdayservice/BirthdayInfo.aidl:

package com.example.birthdayservice;

parcelable BirthdayInfo {
    String name;
    int years;
}

birthday_service/aidl/com/example/birthdayservice/IBirthdayService.aidl:

import com.example.birthdayservice.BirthdayInfo;

interface IBirthdayService {
    /** The same thing, but with a parcelable. */
    String wishWithInfo(in BirthdayInfo info);
}

birthday_service/src/client.rs:

fn main() {
    binder::ProcessState::start_thread_pool();
    let service = connect().expect("Failed to connect to BirthdayService");

    let info = BirthdayInfo { name: "Alice".into(), years: 123 };
    service.wishWithInfo(&info)?;
}

Sending Files

Files can be sent between Binder clients/servers using the ParcelFileDescriptor type:

birthday_service/aidl/com/example/birthdayservice/IBirthdayService.aidl:

interface IBirthdayService {
    /** The same thing, but loads info from a file. */
    String wishFromFile(in ParcelFileDescriptor infoFile);
}

birthday_service/src/client.rs:

fn main() {
    binder::ProcessState::start_thread_pool();
    let service = connect().expect("Failed to connect to BirthdayService");

    // Open a file and put the birthday info in it.
    let mut file = File::create("/data/local/tmp/birthday.info").unwrap();
    writeln!(file, "{name}")?;
    writeln!(file, "{years}")?;

    // Create a `ParcelFileDescriptor` from the file and send it.
    let file = ParcelFileDescriptor::new(file);
    service.wishFromFile(&file)?;
}

birthday_service/src/lib.rs:

impl IBirthdayService for BirthdayService {
    fn wishFromFile(
        &self,
        info_file: &ParcelFileDescriptor,
    ) -> binder::Result<String> {
        // Convert the file descriptor to a `File`. `ParcelFileDescriptor` wraps
        // an `OwnedFd`, which can be cloned and then used to create a `File`
        // object.
        let mut info_file = info_file
            .as_ref()
            .try_clone()
            .map(File::from)
            .expect("Invalid file handle");

        let mut contents = String::new();
        info_file.read_to_string(&mut contents).unwrap();

        let mut lines = contents.lines();
        let name = lines.next().unwrap();
        let years: i32 = lines.next().unwrap().parse().unwrap();

        Ok(format!("Happy Birthday {name}, congratulations with the {years} years!"))
    }
}

Testing in Android

Building on Testing, we will now look at how unit tests work in AOSP. Use the rust_test module for your unit tests:

testing/Android.bp:

rust_library {
    name: "libleftpad",
    crate_name: "leftpad",
    srcs: ["src/lib.rs"],
}

rust_test {
    name: "libleftpad_test",
    crate_name: "leftpad_test",
    srcs: ["src/lib.rs"],
    host_supported: true,
    test_suites: ["general-tests"],
}

testing/src/lib.rs:

#![allow(unused)]
fn main() {
//! Left-padding library.

/// Left-pad `s` to `width`.
pub fn leftpad(s: &str, width: usize) -> String {
    format!("{s:>width$}")
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn short_string() {
        assert_eq!(leftpad("foo", 5), "  foo");
    }

    #[test]
    fn long_string() {
        assert_eq!(leftpad("foobar", 6), "foobar");
    }
}
}

You can now run the test with

atest --host libleftpad_test

The output looks like this:

INFO: Elapsed time: 2.666s, Critical Path: 2.40s
INFO: 3 processes: 2 internal, 1 linux-sandbox.
INFO: Build completed successfully, 3 total actions
//comprehensive-rust-android/testing:libleftpad_test_host            PASSED in 2.3s
    PASSED  libleftpad_test.tests::long_string (0.0s)
    PASSED  libleftpad_test.tests::short_string (0.0s)
Test cases: finished with 2 passing and 0 failing out of 2 test cases

Notice how you only mention the root of the library crate. Tests are found recursively in nested modules.

GoogleTest

The GoogleTest crate allows for flexible test assertions using matchers:

use googletest::prelude::*;

#[googletest::test]
fn test_elements_are() {
    let value = vec!["foo", "bar", "baz"];
    expect_that!(value, elements_are!(eq(&"foo"), lt(&"xyz"), starts_with("b")));
}

If we change the last element to "!", the test fails with a structured error message pin-pointing the error:

---- test_elements_are stdout ----
Value of: value
Expected: has elements:
  0. is equal to "foo"
  1. is less than "xyz"
  2. starts with prefix "!"
Actual: ["foo", "bar", "baz"],
  where element #2 is "baz", which does not start with "!"
  at src/testing/googletest.rs:6:5
Error: See failure output above

#[test]
fn test_multiline_string_diff() {
    let haiku = "Memory safety found,\n\
                 Rust's strong typing guides the way,\n\
                 Secure code you'll write.";
    assert_that!(
        haiku,
        eq("Memory safety found,\n\
            Rust's silly humor guides the way,\n\
            Secure code you'll write.")
    );
}

    Value of: haiku
Expected: is equal to "Memory safety found,\nRust's silly humor guides the way,\nSecure code you'll write."
Actual: "Memory safety found,\nRust's strong typing guides the way,\nSecure code you'll write.",
  which isn't equal to "Memory safety found,\nRust's silly humor guides the way,\nSecure code you'll write."
Difference(-actual / +expected):
 Memory safety found,
-Rust's strong typing guides the way,
+Rust's silly humor guides the way,
 Secure code you'll write.
  at src/testing/googletest.rs:17:5

Mocking

For mocking, Mockall is a widely used library. You need to refactor your code to use traits, which you can then quickly mock:

use std::time::Duration;

#[mockall::automock]
pub trait Pet {
    fn is_hungry(&self, since_last_meal: Duration) -> bool;
}

#[test]
fn test_robot_dog() {
    let mut mock_dog = MockPet::new();
    mock_dog.expect_is_hungry().return_const(true);
    assert_eq!(mock_dog.is_hungry(Duration::from_secs(10)), true);
}

#[test]
fn test_robot_cat() {
    let mut mock_cat = MockPet::new();
    mock_cat
        .expect_is_hungry()
        .with(mockall::predicate::gt(Duration::from_secs(3 * 3600)))
        .return_const(true);
    mock_cat.expect_is_hungry().return_const(false);
    assert_eq!(mock_cat.is_hungry(Duration::from_secs(1 * 3600)), false);
    assert_eq!(mock_cat.is_hungry(Duration::from_secs(5 * 3600)), true);
}

Logging

You should use the log crate to automatically log to logcat (on-device) or stdout (on-host):

hello_rust_logs/Android.bp:

rust_binary {
    name: "hello_rust_logs",
    crate_name: "hello_rust_logs",
    srcs: ["src/main.rs"],
    rustlibs: [
        "liblog_rust",
        "liblogger",
    ],
    host_supported: true,
}

hello_rust_logs/src/main.rs:

//! Rust logging demo.

use log::{debug, error, info};

/// Logs a greeting.
fn main() {
    logger::init(
        logger::Config::default()
            .with_tag_on_device("rust")
            .with_max_level(log::LevelFilter::Trace),
    );
    debug!("Starting program.");
    info!("Things are going fine.");
    error!("Something went wrong!");
}

Build, push, and run the binary on your device:

m hello_rust_logs
adb push "$ANDROID_PRODUCT_OUT/system/bin/hello_rust_logs" /data/local/tmp
adb shell /data/local/tmp/hello_rust_logs

The logs show up in adb logcat:

adb logcat -s rust

09-08 08:38:32.454  2420  2420 D rust: hello_rust_logs: Starting program.
09-08 08:38:32.454  2420  2420 I rust: hello_rust_logs: Things are going fine.
09-08 08:38:32.454  2420  2420 E rust: hello_rust_logs: Something went wrong!

Interoperability

Rust has excellent support for interoperability with other languages. This means that you can:

• Call Rust functions from other languages.
• Call functions written in other languages from Rust.

When you call functions in a foreign language we say that you're using a foreign function interface, also known as FFI.

Interoperability with C

Rust has full support for linking object files with a C calling convention. Similarly, you can export Rust functions and call them from C.

You can do it by hand if you want:

unsafe extern "C" {
    safe fn abs(x: i32) -> i32;
}

fn main() {
    let x = -42;
    let abs_x = abs(x);
    println!("{x}, {abs_x}");
}

We already saw this in the Safe FFI Wrapper exercise.

This assumes full knowledge of the target platform. Not recommended for production.

We will look at better options next.

Using Bindgen

The bindgen tool can auto-generate bindings from a C header file.

First create a small C library:

interoperability/bindgen/libbirthday.h:

typedef struct card {
  const char* name;
  int years;
} card;

void print_card(const card* card);

interoperability/bindgen/libbirthday.c:

#include <stdio.h>
#include "libbirthday.h"

void print_card(const card* card) {
  printf("+--------------\n");
  printf("| Happy Birthday %s!\n", card->name);
  printf("| Congratulations with the %i years!\n", card->years);
  printf("+--------------\n");
}

Add this to your Android.bp file:

interoperability/bindgen/Android.bp:

cc_library {
    name: "libbirthday",
    srcs: ["libbirthday.c"],
}

Create a wrapper header file for the library (not strictly needed in this example):

interoperability/bindgen/libbirthday_wrapper.h:

#include "libbirthday.h"

You can now auto-generate the bindings:

interoperability/bindgen/Android.bp:

rust_bindgen {
    name: "libbirthday_bindgen",
    crate_name: "birthday_bindgen",
    wrapper_src: "libbirthday_wrapper.h",
    source_stem: "bindings",
    static_libs: ["libbirthday"],
}

Finally, we can use the bindings in our Rust program:

interoperability/bindgen/Android.bp:

rust_binary {
    name: "print_birthday_card",
    srcs: ["main.rs"],
    rustlibs: ["libbirthday_bindgen"],
}

interoperability/bindgen/main.rs:

//! Bindgen demo.

use birthday_bindgen::{card, print_card};

fn main() {
    let name = std::ffi::CString::new("Peter").unwrap();
    let card = card { name: name.as_ptr(), years: 42 };
    // SAFETY: The pointer we pass is valid because it came from a Rust
    // reference, and the `name` it contains refers to `name` above which also
    // remains valid. `print_card` doesn't store either pointer to use later
    // after it returns.
    unsafe {
        print_card(&card as *const card);
    }
}

Build, push, and run the binary on your device:

m print_birthday_card
adb push "$ANDROID_PRODUCT_OUT/system/bin/print_birthday_card" /data/local/tmp
adb shell /data/local/tmp/print_birthday_card

Finally, we can run auto-generated tests to ensure the bindings work:

interoperability/bindgen/Android.bp:

rust_test {
    name: "libbirthday_bindgen_test",
    srcs: [":libbirthday_bindgen"],
    crate_name: "libbirthday_bindgen_test",
    test_suites: ["general-tests"],
    auto_gen_config: true,
    clippy_lints: "none", // Generated file, skip linting
    lints: "none",
}

atest libbirthday_bindgen_test

Calling Rust

Exporting Rust functions and types to C is easy:

interoperability/rust/libanalyze/analyze.rs

//! Rust FFI demo.
#![deny(improper_ctypes_definitions)]

use std::os::raw::c_int;

/// Analyze the numbers.
// SAFETY: There is no other global function of this name.
#[unsafe(no_mangle)]
pub extern "C" fn analyze_numbers(x: c_int, y: c_int) {
    if x < y {
        println!("x ({x}) is smallest!");
    } else {
        println!("y ({y}) is probably larger than x ({x})");
    }
}

interoperability/rust/libanalyze/analyze.h

#ifndef ANALYSE_H
#define ANALYSE_H

void analyze_numbers(int x, int y);

#endif

interoperability/rust/libanalyze/Android.bp

rust_ffi {
    name: "libanalyze_ffi",
    crate_name: "analyze_ffi",
    srcs: ["analyze.rs"],
    include_dirs: ["."],
}

We can now call this from a C binary:

interoperability/rust/analyze/main.c

#include "analyze.h"

int main() {
  analyze_numbers(10, 20);
  analyze_numbers(123, 123);
  return 0;
}

interoperability/rust/analyze/Android.bp

cc_binary {
    name: "analyze_numbers",
    srcs: ["main.c"],
    static_libs: ["libanalyze_ffi"],
}

Build, push, and run the binary on your device:

m analyze_numbers
adb push "$ANDROID_PRODUCT_OUT/system/bin/analyze_numbers" /data/local/tmp
adb shell /data/local/tmp/analyze_numbers

With C++

The CXX crate makes it possible to do safe interoperability between Rust and C++.

The overall approach looks like this:

The Bridge Module

CXX relies on a description of the function signatures that will be exposed from each language to the other. You provide this description using extern blocks in a Rust module annotated with the #[cxx::bridge] attribute macro.

#[allow(unsafe_op_in_unsafe_fn)]
#[cxx::bridge(namespace = "org::blobstore")]
mod ffi {
    // Shared structs with fields visible to both languages.
    struct BlobMetadata {
        size: usize,
        tags: Vec<String>,
    }

    // Rust types and signatures exposed to C++.
    extern "Rust" {
        type MultiBuf;

        fn next_chunk(buf: &mut MultiBuf) -> &[u8];
    }

    // C++ types and signatures exposed to Rust.
    unsafe extern "C++" {
        include!("include/blobstore.h");

        type BlobstoreClient;

        fn new_blobstore_client() -> UniquePtr<BlobstoreClient>;
        fn put(self: Pin<&mut BlobstoreClient>, parts: &mut MultiBuf) -> u64;
        fn tag(self: Pin<&mut BlobstoreClient>, blobid: u64, tag: &str);
        fn metadata(&self, blobid: u64) -> BlobMetadata;
    }
}

Rust Bridge Declarations

#[cxx::bridge]
mod ffi {
    extern "Rust" {
        type MyType; // Opaque type
        fn foo(&self); // Method on `MyType`
        fn bar() -> Box<MyType>; // Free function
    }
}

struct MyType(i32);

impl MyType {
    fn foo(&self) {
        println!("{}", self.0);
    }
}

fn bar() -> Box<MyType> {
    Box::new(MyType(123))
}

Generated C++

#[cxx::bridge]
mod ffi {
    // Rust types and signatures exposed to C++.
    extern "Rust" {
        type MultiBuf;

        fn next_chunk(buf: &mut MultiBuf) -> &[u8];
    }
}

Results in (roughly) the following C++:

struct MultiBuf final : public ::rust::Opaque {
  ~MultiBuf() = delete;

private:
  friend ::rust::layout;
  struct layout {
    static ::std::size_t size() noexcept;
    static ::std::size_t align() noexcept;
  };
};

::rust::Slice<::std::uint8_t const> next_chunk(::org::blobstore::MultiBuf &buf) noexcept;

C++ Bridge Declarations

#[cxx::bridge]
mod ffi {
    // C++ types and signatures exposed to Rust.
    unsafe extern "C++" {
        include!("include/blobstore.h");

        type BlobstoreClient;

        fn new_blobstore_client() -> UniquePtr<BlobstoreClient>;
        fn put(self: Pin<&mut BlobstoreClient>, parts: &mut MultiBuf) -> u64;
        fn tag(self: Pin<&mut BlobstoreClient>, blobid: u64, tag: &str);
        fn metadata(&self, blobid: u64) -> BlobMetadata;
    }
}

Results in (roughly) the following Rust:

#[repr(C)]
pub struct BlobstoreClient {
    _private: ::cxx::private::Opaque,
}

pub fn new_blobstore_client() -> ::cxx::UniquePtr<BlobstoreClient> {
    extern "C" {
        #[link_name = "org$blobstore$cxxbridge1$new_blobstore_client"]
        fn __new_blobstore_client() -> *mut BlobstoreClient;
    }
    unsafe { ::cxx::UniquePtr::from_raw(__new_blobstore_client()) }
}

impl BlobstoreClient {
    pub fn put(&self, parts: &mut MultiBuf) -> u64 {
        extern "C" {
            #[link_name = "org$blobstore$cxxbridge1$BlobstoreClient$put"]
            fn __put(
                _: &BlobstoreClient,
                parts: *mut ::cxx::core::ffi::c_void,
            ) -> u64;
        }
        unsafe {
            __put(self, parts as *mut MultiBuf as *mut ::cxx::core::ffi::c_void)
        }
    }
}

// ...

Shared Types

#[cxx::bridge]
mod ffi {
    #[derive(Clone, Debug, Hash)]
    struct PlayingCard {
        suit: Suit,
        value: u8,  // A=1, J=11, Q=12, K=13
    }

    enum Suit {
        Clubs,
        Diamonds,
        Hearts,
        Spades,
    }
}

Shared Enums

#[cxx::bridge]
mod ffi {
    enum Suit {
        Clubs,
        Diamonds,
        Hearts,
        Spades,
    }
}

Generated Rust:

#![allow(unused)]
fn main() {
#[derive(Copy, Clone, PartialEq, Eq)]
#[repr(transparent)]
pub struct Suit {
    pub repr: u8,
}

#[allow(non_upper_case_globals)]
impl Suit {
    pub const Clubs: Self = Suit { repr: 0 };
    pub const Diamonds: Self = Suit { repr: 1 };
    pub const Hearts: Self = Suit { repr: 2 };
    pub const Spades: Self = Suit { repr: 3 };
}
}

Generated C++:

enum class Suit : uint8_t {
  Clubs = 0,
  Diamonds = 1,
  Hearts = 2,
  Spades = 3,
};

Rust Error Handling

#[cxx::bridge]
mod ffi {
    extern "Rust" {
        fn fallible(depth: usize) -> Result<String>;
    }
}

fn fallible(depth: usize) -> anyhow::Result<String> {
    if depth == 0 {
        return Err(anyhow::Error::msg("fallible1 requires depth > 0"));
    }

    Ok("Success!".into())
}

C++ Error Handling

#[cxx::bridge]
mod ffi {
    unsafe extern "C++" {
        include!("example/include/example.h");
        fn fallible(depth: usize) -> Result<String>;
    }
}

fn main() {
    if let Err(err) = ffi::fallible(99) {
        eprintln!("Error: {}", err);
        process::exit(1);
    }
}

Additional Types

Building in Android

Create a cc_library_static to build the C++ library, including the CXX generated header and source file.

cc_library_static {
    name: "libcxx_test_cpp",
    srcs: ["cxx_test.cpp"],
    generated_headers: [
        "cxx-bridge-header",
        "libcxx_test_bridge_header"
    ],
    generated_sources: ["libcxx_test_bridge_code"],
}

Building in Android

Create two genrules: One to generate the CXX header, and one to generate the CXX source file. These are then used as inputs to the cc_library_static.

// Generate a C++ header containing the C++ bindings
// to the Rust exported functions in lib.rs.
genrule {
    name: "libcxx_test_bridge_header",
    tools: ["cxxbridge"],
    cmd: "$(location cxxbridge) $(in) --header > $(out)",
    srcs: ["lib.rs"],
    out: ["lib.rs.h"],
}

// Generate the C++ code that Rust calls into.
genrule {
    name: "libcxx_test_bridge_code",
    tools: ["cxxbridge"],
    cmd: "$(location cxxbridge) $(in) > $(out)",
    srcs: ["lib.rs"],
    out: ["lib.rs.cc"],
}

Building in Android

Create a rust_binary that depends on libcxx and your cc_library_static.

rust_binary {
    name: "cxx_test",
    srcs: ["lib.rs"],
    rustlibs: ["libcxx"],
    static_libs: ["libcxx_test_cpp"],
}

Interoperability with Java

Java can load shared objects via Java Native Interface (JNI). The jni crate allows you to create a compatible library.

First, we create a Rust function to export to Java:

interoperability/java/src/lib.rs:

#![allow(unused)]
fn main() {
//! Rust <-> Java FFI demo.

use jni::objects::{JClass, JString};
use jni::sys::jstring;
use jni::JNIEnv;

/// HelloWorld::hello method implementation.
// SAFETY: There is no other global function of this name.
#[unsafe(no_mangle)]
pub extern "system" fn Java_HelloWorld_hello(
    mut env: JNIEnv,
    _class: JClass,
    name: JString,
) -> jstring {
    let input: String = env.get_string(&name).unwrap().into();
    let greeting = format!("Hello, {input}!");
    let output = env.new_string(greeting).unwrap();
    output.into_raw()
}
}

interoperability/java/Android.bp:

rust_ffi_shared {
    name: "libhello_jni",
    crate_name: "hello_jni",
    srcs: ["src/lib.rs"],
    rustlibs: ["libjni"],
}

We then call this function from Java:

interoperability/java/HelloWorld.java:

class HelloWorld {
    private static native String hello(String name);

    static {
        System.loadLibrary("hello_jni");
    }

    public static void main(String[] args) {
        String output = HelloWorld.hello("Alice");
        System.out.println(output);
    }
}

interoperability/java/Android.bp:

java_binary {
    name: "helloworld_jni",
    srcs: ["HelloWorld.java"],
    main_class: "HelloWorld",
    required: ["libhello_jni"],
}

Finally, you can build, sync, and run the binary:

m helloworld_jni
adb sync  # requires adb root && adb remount
adb shell /system/bin/helloworld_jni

Welcome to Rust in Chromium

Rust is supported for third-party libraries in Chromium, with first-party glue code to connect between Rust and existing Chromium C++ code.

Today, we'll call into Rust to do something silly with strings. If you've got a corner of the code where you're displaying a UTF8 string to the user, feel free to follow this recipe in your part of the codebase instead of the exact part we talk about.

Setup

Make sure you can build and run Chromium. Any platform and set of build flags is OK, so long as your code is relatively recent (commit position 1223636 onwards, corresponding to November 2023):

gn gen out/Debug
autoninja -C out/Debug chrome
out/Debug/chrome # or on Mac, out/Debug/Chromium.app/Contents/MacOS/Chromium

(A component, debug build is recommended for quickest iteration time. This is the default!)

See How to build Chromium if you aren't already at that point. Be warned: setting up to build Chromium takes time.

It's also recommended that you have Visual Studio code installed.

About the exercises

This part of the course has a series of exercises which build on each other. We'll be doing them spread throughout the course instead of just at the end. If you don't have time to complete a certain part, don't worry: you can catch up in the next slot.

Comparing Chromium and Cargo Ecosystems

The Rust community typically uses cargo and libraries from crates.io. Chromium is built using gn and ninja and a curated set of dependencies.

When writing code in Rust, your choices are:

• Use gn and ninja with the help of the templates from //build/rust/*.gni (e.g. rust_static_library that we'll meet later). This uses Chromium's audited toolchain and crates.
• Use cargo, but restrict yourself to Chromium's audited toolchain and crates
• Use cargo, trusting a toolchain and/or crates downloaded from the internet

From here on we'll be focusing on gn and ninja, because this is how Rust code can be built into the Chromium browser. At the same time, Cargo is an important part of the Rust ecosystem and you should keep it in your toolbox.

Mini exercise

Split into small groups and:

• Brainstorm scenarios where cargo may offer an advantage and assess the risk profile of these scenarios.
• Discuss which tools, libraries, and groups of people need to be trusted when using gn and ninja, offline cargo, etc.

Chromium Rust policy

Chromium does not yet allow first-party Rust except in rare cases as approved by Chromium's Area Tech Leads.

Chromium's policy on third party libraries is outlined here - Rust is allowed for third party libraries under various circumstances, including if they're the best option for performance or for security.

Very few Rust libraries directly expose a C/C++ API, so that means that nearly all such libraries will require a small amount of first-party glue code.

First-party Rust glue code for a particular third-party crate should normally be kept in third_party/rust/<crate>/<version>/wrapper.

Because of this, today's course will be heavily focused on:

• Bringing in third-party Rust libraries ("crates")
• Writing glue code to be able to use those crates from Chromium C++.

If this policy changes over time, the course will evolve to keep up.

Build rules

Rust code is usually built using cargo. Chromium builds with gn and ninja for efficiency --- its static rules allow maximum parallelism. Rust is no exception.

Adding Rust code to Chromium

In some existing Chromium BUILD.gn file, declare a rust_static_library:

import("//build/rust/rust_static_library.gni")

rust_static_library("my_rust_lib") {
  crate_root = "lib.rs"
  sources = [ "lib.rs" ]
}

You can also add deps on other Rust targets. Later we'll use this to depend upon third party code.

Including unsafe Rust Code

Unsafe Rust code is forbidden in rust_static_library by default --- it won't compile. If you need unsafe Rust code, add allow_unsafe = true to the gn target. (Later in the course we'll see circumstances where this is necessary.)

import("//build/rust/rust_static_library.gni")

rust_static_library("my_rust_lib") {
  crate_root = "lib.rs"
  sources = [
    "lib.rs",
    "hippopotamus.rs"
  ]
  allow_unsafe = true
}

Depending on Rust Code from Chromium C++

Simply add the above target to the deps of some Chromium C++ target.

import("//build/rust/rust_static_library.gni")

rust_static_library("my_rust_lib") {
  crate_root = "lib.rs"
  sources = [ "lib.rs" ]
}

# or source_set, static_library etc.
component("preexisting_cpp") {
  deps = [ ":my_rust_lib" ]
}

Visual Studio Code

Types are elided in Rust code, which makes a good IDE even more useful than for C++. Visual Studio code works well for Rust in Chromium. To use it,

• Ensure your VSCode has the rust-analyzer extension, not earlier forms of Rust support
• gn gen out/Debug --export-rust-project (or equivalent for your output directory)
• ln -s out/Debug/rust-project.json rust-project.json

Build rules exercise

In your Chromium build, add a new Rust target to //ui/base/BUILD.gn containing:

#![allow(unused)]
fn main() {
// SAFETY: There is no other global function of this name.
#[unsafe(no_mangle)]
pub extern "C" fn hello_from_rust() {
    println!("Hello from Rust!")
}
}

Important: note that no_mangle here is considered a type of unsafety by the Rust compiler, so you'll need to allow unsafe code in your gn target.

Add this new Rust target as a dependency of //ui/base:base. Declare this function at the top of ui/base/resource/resource_bundle.cc (later, we'll see how this can be automated by bindings generation tools):

extern "C" void hello_from_rust();

Call this function from somewhere in ui/base/resource/resource_bundle.cc - we suggest the top of ResourceBundle::MaybeMangleLocalizedString. Build and run Chromium, and ensure that "Hello from Rust!" is printed lots of times.

If you use VSCode, now set up Rust to work well in VSCode. It will be useful in subsequent exercises. If you've succeeded, you will be able to use right-click "Go to definition" on println!.

Where to find help

• The options available to the rust_static_library gn template
• Information about #[unsafe(no_mangle)]
• Information about extern "C"
• Information about gn's --export-rust-project switch
• How to install rust-analyzer in VSCode

Testing

Rust community typically authors unit tests in a module placed in the same source file as the code being tested. This was covered earlier in the course and looks like this:

#![allow(unused)]
fn main() {
#[cfg(test)]
mod tests {
    #[test]
    fn my_test() {
        todo!()
    }
}
}

In Chromium we place unit tests in a separate source file and we continue to follow this practice for Rust --- this makes tests consistently discoverable and helps to avoid rebuilding .rs files a second time (in the test configuration).

This results in the following options for testing Rust code in Chromium:

• Native Rust tests (i.e. #[test]). Discouraged outside of //third_party/rust.
• gtest tests authored in C++ and exercising Rust via FFI calls. Sufficient when Rust code is just a thin FFI layer and the existing unit tests provide sufficient coverage for the feature.
• gtest tests authored in Rust and using the crate under test through its public API (using pub mod for_testing { ... } if needed). This is the subject of the next few slides.

rust_gtest_interop Library

The rust_gtest_interop library provides a way to:

• Use a Rust function as a gtest testcase (using the #[gtest(...)] attribute)
• Use expect_eq! and similar macros (similar to assert_eq! but not panicking and not terminating the test when the assertion fails).

Example:

use rust_gtest_interop::prelude::*;

#[gtest(MyRustTestSuite, MyAdditionTest)]
fn test_addition() {
    expect_eq!(2 + 2, 4);
}

GN Rules for Rust Tests

The simplest way to build Rust gtest tests is to add them to an existing test binary that already contains tests authored in C++. For example:

test("ui_base_unittests") {
  ...
  sources += [ "my_rust_lib_unittest.rs" ]
  deps += [ ":my_rust_lib" ]
}

Authoring Rust tests in a separate static_library also works, but requires manually declaring the dependency on the support libraries:

rust_static_library("my_rust_lib_unittests") {
  testonly = true
  is_gtest_unittests = true
  crate_root = "my_rust_lib_unittest.rs"
  sources = [ "my_rust_lib_unittest.rs" ]
  deps = [
    ":my_rust_lib",
    "//testing/rust_gtest_interop",
  ]
}

test("ui_base_unittests") {
  ...
  deps += [ ":my_rust_lib_unittests" ]
}

chromium::import! Macro

After adding :my_rust_lib to GN deps, we still need to learn how to import and use my_rust_lib from my_rust_lib_unittest.rs. We haven't provided an explicit crate_name for my_rust_lib so its crate name is computed based on the full target path and name. Fortunately we can avoid working with such an unwieldy name by using the chromium::import! macro from the automatically-imported chromium crate:

chromium::import! {
    "//ui/base:my_rust_lib";
}

use my_rust_lib::my_function_under_test;

Under the covers the macro expands to something similar to:

extern crate ui_sbase_cmy_urust_ulib as my_rust_lib;

use my_rust_lib::my_function_under_test;

More information can be found in the doc comment of the chromium::import macro.

Testing exercise

Time for another exercise!

In your Chromium build:

• Add a testable function next to hello_from_rust. Some suggestions: adding two integers received as arguments, computing the nth Fibonacci number, summing integers in a slice, etc.
• Add a separate ..._unittest.rs file with a test for the new function.
• Add the new tests to BUILD.gn.
• Build the tests, run them, and verify that the new test works.

Interoperability with C++

The Rust community offers multiple options for C++/Rust interop, with new tools being developed all the time. At the moment, Chromium uses a tool called CXX.

You describe your whole language boundary in an interface definition language (which looks a lot like Rust) and then CXX tools generate declarations for functions and types in both Rust and C++.

See the CXX tutorial for a full example of using this.

Example Bindings

CXX requires that the whole C++/Rust boundary is declared in cxx::bridge modules inside .rs source code.

#[cxx::bridge]
mod ffi {
    extern "Rust" {
        type MultiBuf;

        fn next_chunk(buf: &mut MultiBuf) -> &[u8];
    }

    unsafe extern "C++" {
        include!("example/include/blobstore.h");

        type BlobstoreClient;

        fn new_blobstore_client() -> UniquePtr<BlobstoreClient>;
        fn put(self: &BlobstoreClient, buf: &mut MultiBuf) -> Result<u64>;
    }
}

// Definitions of Rust types and functions go here

Limitations of CXX

By far the most useful page when using CXX is the type reference.

CXX fundamentally suits cases where:

• Your Rust-C++ interface is sufficiently simple that you can declare all of it.
• You're using only the types natively supported by CXX already, for example std::unique_ptr, std::string, &[u8] etc.

It has many limitations --- for example lack of support for Rust's Option type.

These limitations constrain us to using Rust in Chromium only for well isolated "leaf nodes" rather than for arbitrary Rust-C++ interop. When considering a use-case for Rust in Chromium, a good starting point is to draft the CXX bindings for the language boundary to see if it appears simple enough.

CXX Error Handling

CXX's support for Result<T,E> relies on C++ exceptions, so we can't use that in Chromium. Alternatives:

• The T part of Result<T, E> can be:

  • Returned via out parameters (e.g. via &mut T). This requires that T can be passed across the FFI boundary - for example T has to be:
    • A primitive type (like u32 or usize)
    • A type natively supported by cxx (like UniquePtr<T>) that has a suitable default value to use in a failure case (unlike Box<T>).
  • Retained on the Rust side, and exposed via reference. This may be needed when T is a Rust type, which cannot be passed across the FFI boundary, and cannot be stored in UniquePtr<T>.

• The E part of Result<T, E> can be:

  • Returned as a boolean (e.g. true representing success, and false representing failure)
  • Preserving error details is in theory possible, but so far hasn't been needed in practice.

CXX Error Handling: QR Example

The QR code generator is an example where a boolean is used to communicate success vs failure, and where the successful result can be passed across the FFI boundary:

#[cxx::bridge(namespace = "qr_code_generator")]
mod ffi {
    extern "Rust" {
        fn generate_qr_code_using_rust(
            data: &[u8],
            min_version: i16,
            out_pixels: Pin<&mut CxxVector<u8>>,
            out_qr_size: &mut usize,
        ) -> bool;
    }
}

CXX Error Handling: PNG Example

A prototype of a PNG decoder illustrates what can be done when the successful result cannot be passed across the FFI boundary:

#[cxx::bridge(namespace = "gfx::rust_bindings")]
mod ffi {
    extern "Rust" {
        /// This returns an FFI-friendly equivalent of `Result<PngReader<'a>,
        /// ()>`.
        fn new_png_reader<'a>(input: &'a [u8]) -> Box<ResultOfPngReader<'a>>;

        /// C++ bindings for the `crate::png::ResultOfPngReader` type.
        type ResultOfPngReader<'a>;
        fn is_err(self: &ResultOfPngReader) -> bool;
        fn unwrap_as_mut<'a, 'b>(
            self: &'b mut ResultOfPngReader<'a>,
        ) -> &'b mut PngReader<'a>;

        /// C++ bindings for the `crate::png::PngReader` type.
        type PngReader<'a>;
        fn height(self: &PngReader) -> u32;
        fn width(self: &PngReader) -> u32;
        fn read_rgba8(self: &mut PngReader, output: &mut [u8]) -> bool;
    }
}

Using cxx in Chromium

In Chromium, we define an independent #[cxx::bridge] mod for each leaf-node where we want to use Rust. You'd typically have one for each rust_static_library. Just add