Advanced Traits

We first covered traits in the “Traits: Defining Shared Behavior” section of Chapter 10, but we didn’t discuss the more advanced details. Now that you know more about Rust, we can get into the nitty-gritty.

Specifying Placeholder Types in Trait Definitions with Associated Types

Associated types connect a type placeholder with a trait such that the trait method definitions can use these placeholder types in their signatures. The implementor of a trait will specify the concrete type to be used instead of the placeholder type for the particular implementation. That way, we can define a trait that uses some types without needing to know exactly what those types are until the trait is implemented.

We’ve described most of the advanced features in this chapter as being rarely needed. Associated types are somewhere in the middle: they’re used more rarely than features explained in the rest of the book but more commonly than many of the other features discussed in this chapter.

One example of a trait with an associated type is the Iterator trait that the standard library provides. The associated type is named Item and stands in for the type of the values the type implementing the Iterator trait is iterating over. The definition of the Iterator trait is as shown in Listing 19-12.

  1. pub trait Iterator {
  2.     type Item;
  4.     fn next(&mut self) -> Option<Self::Item>;
  5. }

Listing 19-12: The definition of the Iterator trait that has an associated type Item

The type Item is a placeholder, and the next method’s definition shows that it will return values of type Option. Implementors of the Iterator trait will specify the concrete type for Item, and the next method will return an Option containing a value of that concrete type.

Associated types might seem like a similar concept to generics, in that the latter allow us to define a function without specifying what types it can handle. To examine the difference between the two concepts, we’ll look at an implementation of the Iterator trait on a type named Counter that specifies the Item type is u32:

Filename: src/lib.rs

  1. struct Counter {
  2.     count: u32,
  3. }
  5. impl Counter {
  6.     fn new() -> Counter {
  7.         Counter { count: 0 }
  8.     }
  9. }
  11. impl Iterator for Counter {
  12.     type Item = u32;
  14.     fn next(&mut self) -> Option<Self::Item> {
  15.         // --snip--
  16.         if self.count < 5 {
  17.             self.count += 1;
  18.             Some(self.count)
  19.         } else {
  20.             None
  21.         }
  22.     }
  23. }

This syntax seems comparable to that of generics. So why not just define the Iterator trait with generics, as shown in Listing 19-13?

  1. pub trait Iterator<T> {
  2.     fn next(&mut self) -> Option<T>;
  3. }

Listing 19-13: A hypothetical definition of the Iterator trait using generics

The difference is that when using generics, as in Listing 19-13, we must annotate the types in each implementation; because we can also implement Iterator for Counter or any other type, we could have multiple implementations of Iterator for Counter. In other words, when a trait has a generic parameter, it can be implemented for a type multiple times, changing the concrete types of the generic type parameters each time. When we use the next method on Counter, we would have to provide type annotations to indicate which implementation of Iterator we want to use.

With associated types, we don’t need to annotate types because we can’t implement a trait on a type multiple times. In Listing 19-12 with the definition that uses associated types, we can only choose what the type of Item will be once, because there can only be one impl Iterator for Counter. We don’t have to specify that we want an iterator of u32 values everywhere that we call next on Counter.

Associated types also become part of the trait’s contract: implementors of the trait must provide a type to stand in for the associated type placeholder. Associated types often have a name that describes how the type will be used, and documenting the associated type in the API documentation is good practice.

Default Generic Type Parameters and Operator Overloading

When we use generic type parameters, we can specify a default concrete type for the generic type. This eliminates the need for implementors of the trait to specify a concrete type if the default type works. You specify a default type when declaring a generic type with the syntax.

A great example of a situation where this technique is useful is with operator overloading, in which you customize the behavior of an operator (such as +) in particular situations.

Rust doesn’t allow you to create your own operators or overload arbitrary operators. But you can overload the operations and corresponding traits listed in std::ops by implementing the traits associated with the operator. For example, in Listing 19-14 we overload the + operator to add two Point instances together. We do this by implementing the Add trait on a Point struct:

Filename: src/main.rs

  1. use std::ops::Add;
  3. #[derive(Debug, Copy, Clone, PartialEq)]
  4. struct Point {
  5.     x: i32,
  6.     y: i32,
  7. }
  9. impl Add for Point {
  10.     type Output = Point;
  12.     fn add(self, other: Point) -> Point {
  13.         Point {
  14.             x: self.x + other.x,
  15.             y: self.y + other.y,
  16.         }
  17.     }
  18. }
  20. fn main() {
  21.     assert_eq!(
  22.         Point { x: 1, y: 0 } + Point { x: 2, y: 3 },
  23.         Point { x: 3, y: 3 }
  24.     );
  25. }

Listing 19-14: Implementing the Add trait to overload the + operator for Point instances

The add method adds the x values of two Point instances and the y values of two Point instances to create a new Point. The Add trait has an associated type named Output that determines the type returned from the add method.

The default generic type in this code is within the Add trait. Here is its definition:

  1. #![allow(unused)]
  2. fn main() {
  3. trait Add<Rhs=Self> {
  4.     type Output;
  6.     fn add(self, rhs: Rhs) -> Self::Output;
  7. }
  8. }

This code should look generally familiar: a trait with one method and an associated type. The new part is Rhs=Self: this syntax is called default type parameters. The Rhs generic type parameter (short for “right hand side”) defines the type of the rhs parameter in the add method. If we don’t specify a concrete type for Rhs when we implement the Add trait, the type of Rhs will default to Self, which will be the type we’re implementing Add on.

When we implemented Add for Point, we used the default for Rhs because we wanted to add two Point instances. Let’s look at an example of implementing the Add trait where we want to customize the Rhs type rather than using the default.

We have two structs, Millimeters and Meters, holding values in different units. This thin wrapping of an existing type in another struct is known as the newtype pattern, which we describe in more detail in the “Using the Newtype Pattern to Implement External Traits on External Types” section. We want to add values in millimeters to values in meters and have the implementation of Add do the conversion correctly. We can implement Add for Millimeters with Meters as the Rhs, as shown in Listing 19-15.

Filename: src/lib.rs

  1. use std::ops::Add;
  3. struct Millimeters(u32);
  4. struct Meters(u32);
  6. impl Add<Meters> for Millimeters {
  7.     type Output = Millimeters;
  9.     fn add(self, other: Meters) -> Millimeters {
  10.         Millimeters(self.0 + (other.0 * 1000))
  11.     }
  12. }

Listing 19-15: Implementing the Add trait on Millimeters to add Millimeters to Meters

To add Millimeters and Meters, we specify impl Add to set the value of the Rhs type parameter instead of using the default of Self.

You’ll use default type parameters in two main ways:

  • To extend a type without breaking existing code
  • To allow customization in specific cases most users won’t need

The standard library’s Add trait is an example of the second purpose: usually, you’ll add two like types, but the Add trait provides the ability to customize beyond that. Using a default type parameter in the Add trait definition means you don’t have to specify the extra parameter most of the time. In other words, a bit of implementation boilerplate isn’t needed, making it easier to use the trait.

The first purpose is similar to the second but in reverse: if you want to add a type parameter to an existing trait, you can give it a default to allow extension of the functionality of the trait without breaking the existing implementation code.

Fully Qualified Syntax for Disambiguation: Calling Methods with the Same Name

Nothing in Rust prevents a trait from having a method with the same name as another trait’s method, nor does Rust prevent you from implementing both traits on one type. It’s also possible to implement a method directly on the type with the same name as methods from traits.

When calling methods with the same name, you’ll need to tell Rust which one you want to use. Consider the code in Listing 19-16 where we’ve defined two traits, Pilot and Wizard, that both have a method called fly. We then implement both traits on a type Human that already has a method named fly implemented on it. Each fly method does something different.

Filename: src/main.rs

  1. trait Pilot {
  2.     fn fly(&self);
  3. }
  5. trait Wizard {
  6.     fn fly(&self);
  7. }
  9. struct Human;
  11. impl Pilot for Human {
  12.     fn fly(&self) {
  13.         println!("This is your captain speaking.");
  14.     }
  15. }
  17. impl Wizard for Human {
  18.     fn fly(&self) {
  19.         println!("Up!");
  20.     }
  21. }
  23. impl Human {
  24.     fn fly(&self) {
  25.         println!("*waving arms furiously*");
  26.     }
  27. }
  29. fn main() {}

Listing 19-16: Two traits are defined to have a fly method and are implemented on the Human type, and a fly method is implemented on Human directly

When we call fly on an instance of Human, the compiler defaults to calling the method that is directly implemented on the type, as shown in Listing 19-17.

Filename: src/main.rs

  1. trait Pilot {
  2.     fn fly(&self);
  3. }
  5. trait Wizard {
  6.     fn fly(&self);
  7. }
  9. struct Human;
  11. impl Pilot for Human {
  12.     fn fly(&self) {
  13.         println!("This is your captain speaking.");
  14.     }
  15. }
  17. impl Wizard for Human {
  18.     fn fly(&self) {
  19.         println!("Up!");
  20.     }
  21. }
  23. impl Human {
  24.     fn fly(&self) {
  25.         println!("*waving arms furiously*");
  26.     }
  27. }
  29. fn main() {
  30.     let person = Human;
  31.     person.fly();
  32. }

Listing 19-17: Calling fly on an instance of Human

Running this code will print *waving arms furiously*, showing that Rust called the fly method implemented on Human directly.

To call the fly methods from either the Pilot trait or the Wizard trait, we need to use more explicit syntax to specify which fly method we mean. Listing 19-18 demonstrates this syntax.

Filename: src/main.rs

  1. trait Pilot {
  2.     fn fly(&self);
  3. }
  5. trait Wizard {
  6.     fn fly(&self);
  7. }
  9. struct Human;
  11. impl Pilot for Human {
  12.     fn fly(&self) {
  13.         println!("This is your captain speaking.");
  14.     }
  15. }
  17. impl Wizard for Human {
  18.     fn fly(&self) {
  19.         println!("Up!");
  20.     }
  21. }
  23. impl Human {
  24.     fn fly(&self) {
  25.         println!("*waving arms furiously*");
  26.     }
  27. }
  29. fn main() {
  30.     let person = Human;
  31.     Pilot::fly(&person);
  32.     Wizard::fly(&person);
  33.     person.fly();
  34. }

Listing 19-18: Specifying which trait’s fly method we want to call

Specifying the trait name before the method name clarifies to Rust which implementation of fly we want to call. We could also write Human::fly(&person), which is equivalent to the person.fly() that we used in Listing 19-18, but this is a bit longer to write if we don’t need to disambiguate.

Running this code prints the following:

  1. $ cargo run
  2.    Compiling traits-example v0.1.0 (file:///projects/traits-example)
  3.     Finished dev [unoptimized + debuginfo] target(s) in 0.46s
  4.      Running `target/debug/traits-example`
  5. This is your captain speaking.
  6. Up!
  7. *waving arms furiously*

Because the fly method takes a self parameter, if we had two types that both implement one trait, Rust could figure out which implementation of a trait to use based on the type of self.

However, associated functions that are not methods don’t have a self parameter. When there are multiple types or traits that define non-method functions with the same function name, Rust doesn't always know which type you mean unless you use fully qualified syntax. For example, in Listing 19-19 we create a trait for an animal shelter that wants to name all baby dogs Spot. We make an Animal trait with an associated non-method function baby_name. The Animal trait is implemented for the struct Dog, on which we also provide an associated non-method function baby_name directly.

Filename: src/main.rs

  1. trait Animal {
  2.     fn baby_name() -> String;
  3. }
  5. struct Dog;
  7. impl Dog {
  8.     fn baby_name() -> String {
  9.         String::from("Spot")
  10.     }
  11. }
  13. impl Animal for Dog {
  14.     fn baby_name() -> String {
  15.         String::from("puppy")
  16.     }
  17. }
  19. fn main() {
  20.     println!("A baby dog is called a {}", Dog::baby_name());
  21. }

Listing 19-19: A trait with an associated function and a type with an associated function of the same name that also implements the trait

We implement the code for naming all puppies Spot in the baby_name associated function that is defined on Dog. The Dog type also implements the trait Animal, which describes characteristics that all animals have. Baby dogs are called puppies, and that is expressed in the implementation of the Animal trait on Dog in the baby_name function associated with the Animal trait.

In main, we call the Dog::baby_name function, which calls the associated function defined on Dog directly. This code prints the following:

  1. $ cargo run
  2.    Compiling traits-example v0.1.0 (file:///projects/traits-example)
  3.     Finished dev [unoptimized + debuginfo] target(s) in 0.54s
  4.      Running `target/debug/traits-example`
  5. A baby dog is called a Spot

This output isn’t what we wanted. We want to call the baby_name function that is part of the Animal trait that we implemented on Dog so the code prints A baby dog is called a puppy. The technique of specifying the trait name that we used in Listing 19-18 doesn’t help here; if we change main to the code in Listing 19-20, we’ll get a compilation error.

Filename: src/main.rs

  1. trait Animal {
  2.     fn baby_name() -> String;
  3. }
  5. struct Dog;
  7. impl Dog {
  8.     fn baby_name() -> String {
  9.         String::from("Spot")
  10.     }
  11. }
  13. impl Animal for Dog {
  14.     fn baby_name() -> String {
  15.         String::from("puppy")
  16.     }
  17. }
  19. fn main() {
  20.     println!("A baby dog is called a {}", Animal::baby_name());
  21. }

Listing 19-20: Attempting to call the baby_name function from the Animal trait, but Rust doesn’t know which implementation to use

Because Animal::baby_name doesn’t have a self parameter, and there could be other types that implement the Animal trait, Rust can’t figure out which implementation of Animal::baby_name we want. We’ll get this compiler error:

  1. $ cargo run
  2.    Compiling traits-example v0.1.0 (file:///projects/traits-example)
  3. error[E0283]: type annotations needed
  4.   --> src/main.rs:20:43
  5.    |
  6. 20 |     println!("A baby dog is called a {}", Animal::baby_name());
  7.    |                                           ^^^^^^^^^^^^^^^^^ cannot infer type
  8.    |
  9.    = note: cannot satisfy `_: Animal`
  11. For more information about this error, try `rustc --explain E0283`.
  12. error: could not compile `traits-example` due to previous error

To disambiguate and tell Rust that we want to use the implementation of Animal for Dog as opposed to the implementation of Animal for some other type, we need to use fully qualified syntax. Listing 19-21 demonstrates how to use fully qualified syntax.

Filename: src/main.rs

  1. trait Animal {
  2.     fn baby_name() -> String;
  3. }
  5. struct Dog;
  7. impl Dog {
  8.     fn baby_name() -> String {
  9.         String::from("Spot")
  10.     }
  11. }
  13. impl Animal for Dog {
  14.     fn baby_name() -> String {
  15.         String::from("puppy")
  16.     }
  17. }
  19. fn main() {
  20.     println!("A baby dog is called a {}", <Dog as Animal>::baby_name());
  21. }

Listing 19-21: Using fully qualified syntax to specify that we want to call the baby_name function from the Animal trait as implemented on Dog

We’re providing Rust with a type annotation within the angle brackets, which indicates we want to call the baby_name method from the Animal trait as implemented on Dog by saying that we want to treat the Dog type as an Animal for this function call. This code will now print what we want:

  1. $ cargo run
  2.    Compiling traits-example v0.1.0 (file:///projects/traits-example)
  3.     Finished dev [unoptimized + debuginfo] target(s) in 0.48s
  4.      Running `target/debug/traits-example`
  5. A baby dog is called a puppy

In general, fully qualified syntax is defined as follows:

  1. <Type as Trait>::function(receiver_if_method, next_arg, ...);

For associated functions that aren’t methods, there would not be a receiver: there would only be the list of other arguments. You could use fully qualified syntax everywhere that you call functions or methods. However, you’re allowed to omit any part of this syntax that Rust can figure out from other information in the program. You only need to use this more verbose syntax in cases where there are multiple implementations that use the same name and Rust needs help to identify which implementation you want to call.

Using Supertraits to Require One Trait’s Functionality Within Another Trait

Sometimes, you might write a trait definition that depends on another trait: for a type to implement the first trait, you want to require that type to also implement the second trait. You would do this so that your trait definition can make use of the associated items of the second trait. The trait your trait definition is relying on is called a supertrait of your trait.

For example, let’s say we want to make an OutlinePrint trait with an outline_print method that will print a given value formatted so that it's framed in asterisks. That is, given a Point struct that implements the standard library trait Display to result in (x, y), when we call outline_print on a Point instance that has 1 for x and 3 for y, it should print the following:

  1. **********
  2. *        *
  3. * (1, 3) *
  4. *        *
  5. **********

In the implementation of the outline_print method, we want to use the Display trait’s functionality. Therefore, we need to specify that the OutlinePrint trait will work only for types that also implement Display and provide the functionality that OutlinePrint needs. We can do that in the trait definition by specifying OutlinePrint: Display. This technique is similar to adding a trait bound to the trait. Listing 19-22 shows an implementation of the OutlinePrint trait.

Filename: src/main.rs

  1. use std::fmt;
  3. trait OutlinePrint: fmt::Display {
  4.     fn outline_print(&self) {
  5.         let output = self.to_string();
  6.         let len = output.len();
  7.         println!("{}", "*".repeat(len + 4));
  8.         println!("*{}*", " ".repeat(len + 2));
  9.         println!("* {} *", output);
  10.         println!("*{}*", " ".repeat(len + 2));
  11.         println!("{}", "*".repeat(len + 4));
  12.     }
  13. }
  15. fn main() {}

Listing 19-22: Implementing the OutlinePrint trait that requires the functionality from Display

Because we’ve specified that OutlinePrint requires the Display trait, we can use the to_string function that is automatically implemented for any type that implements Display. If we tried to use to_string without adding a colon and specifying the Display trait after the trait name, we’d get an error saying that no method named to_string was found for the type &Self in the current scope.

Let’s see what happens when we try to implement OutlinePrint on a type that doesn’t implement Display, such as the Point struct:

Filename: src/main.rs

  1. use std::fmt;
  3. trait OutlinePrint: fmt::Display {
  4.     fn outline_print(&self) {
  5.         let output = self.to_string();
  6.         let len = output.len();
  7.         println!("{}", "*".repeat(len + 4));
  8.         println!("*{}*", " ".repeat(len + 2));
  9.         println!("* {} *", output);
  10.         println!("*{}*", " ".repeat(len + 2));
  11.         println!("{}", "*".repeat(len + 4));
  12.     }
  13. }
  15. struct Point {
  16.     x: i32,
  17.     y: i32,
  18. }
  20. impl OutlinePrint for Point {}
  22. fn main() {
  23.     let p = Point { x: 1, y: 3 };
  24.     p.outline_print();
  25. }

We get an error saying that Display is required but not implemented:

  1. $ cargo run
  2.    Compiling traits-example v0.1.0 (file:///projects/traits-example)
  3. error[E0277]: `Point` doesn't implement `std::fmt::Display`
  4.   --> src/main.rs:20:6
  5.    |
  6. 20 | impl OutlinePrint for Point {}
  7.    |      ^^^^^^^^^^^^ `Point` cannot be formatted with the default formatter
  8.    |
  9.    = help: the trait `std::fmt::Display` is not implemented for `Point`
  10.    = note: in format strings you may be able to use `{:?}` (or {:#?} for pretty-print) instead
  11. note: required by a bound in `OutlinePrint`
  12.   --> src/main.rs:3:21
  13.    |
  14. 3  | trait OutlinePrint: fmt::Display {
  15.    |                     ^^^^^^^^^^^^ required by this bound in `OutlinePrint`
  17. For more information about this error, try `rustc --explain E0277`.
  18. error: could not compile `traits-example` due to previous error

To fix this, we implement Display on Point and satisfy the constraint that OutlinePrint requires, like so:

Filename: src/main.rs

  1. trait OutlinePrint: fmt::Display {
  2.     fn outline_print(&self) {
  3.         let output = self.to_string();
  4.         let len = output.len();
  5.         println!("{}", "*".repeat(len + 4));
  6.         println!("*{}*", " ".repeat(len + 2));
  7.         println!("* {} *", output);
  8.         println!("*{}*", " ".repeat(len + 2));
  9.         println!("{}", "*".repeat(len + 4));
  10.     }
  11. }
  13. struct Point {
  14.     x: i32,
  15.     y: i32,
  16. }
  18. impl OutlinePrint for Point {}
  20. use std::fmt;
  22. impl fmt::Display for Point {
  23.     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
  24.         write!(f, "({}, {})", self.x, self.y)
  25.     }
  26. }
  28. fn main() {
  29.     let p = Point { x: 1, y: 3 };
  30.     p.outline_print();
  31. }

Then implementing the OutlinePrint trait on Point will compile successfully, and we can call outline_print on a Point instance to display it within an outline of asterisks.

Using the Newtype Pattern to Implement External Traits on External Types

In Chapter 10 in the “Implementing a Trait on a Type” section, we mentioned the orphan rule that states we’re only allowed to implement a trait on a type if either the trait or the type are local to our crate. It’s possible to get around this restriction using the newtype pattern, which involves creating a new type in a tuple struct. (We covered tuple structs in the “Using Tuple Structs without Named Fields to Create Different Types” section of Chapter 5.) The tuple struct will have one field and be a thin wrapper around the type we want to implement a trait for. Then the wrapper type is local to our crate, and we can implement the trait on the wrapper. Newtype is a term that originates from the Haskell programming language. There is no runtime performance penalty for using this pattern, and the wrapper type is elided at compile time.

As an example, let’s say we want to implement Display on Vec, which the orphan rule prevents us from doing directly because the Display trait and the Vec type are defined outside our crate. We can make a Wrapper struct that holds an instance of Vec; then we can implement Display on Wrapper and use the Vec value, as shown in Listing 19-23.

Filename: src/main.rs

  1. use std::fmt;
  3. struct Wrapper(Vec<String>);
  5. impl fmt::Display for Wrapper {
  6.     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
  7.         write!(f, "[{}]", self.0.join(", "))
  8.     }
  9. }
  11. fn main() {
  12.     let w = Wrapper(vec![String::from("hello"), String::from("world")]);
  13.     println!("w = {}", w);
  14. }

Listing 19-23: Creating a Wrapper type around Vec to implement Display

The implementation of Display uses self.0 to access the inner Vec, because Wrapper is a tuple struct and Vec is the item at index 0 in the tuple. Then we can use the functionality of the Display type on Wrapper.

The downside of using this technique is that Wrapper is a new type, so it doesn’t have the methods of the value it’s holding. We would have to implement all the methods of Vec directly on Wrapper such that the methods delegate to self.0, which would allow us to treat Wrapper exactly like a Vec. If we wanted the new type to have every method the inner type has, implementing the Deref trait (discussed in Chapter 15 in the “Treating Smart Pointers Like Regular References with the Deref Trait” section) on the Wrapper to return the inner type would be a solution. If we don’t want the Wrapper type to have all the methods of the inner type—for example, to restrict the Wrapper type’s behavior—we would have to implement just the methods we do want manually.

This newtype pattern is also useful even when traits are not involved. Let’s switch focus and look at some advanced ways to interact with Rust’s type system.