std::iter::DoubleEndedIterator — Rust trait

What `DoubleEndedIterator` does

DoubleEndedIterator is a trait exposed by the std::iter module of the Rust standard library. It is part of the public, stable API and you can rely on it from any edition of Rust.

As a trait, DoubleEndedIterator defines an interface that other types can implement. Code generic over T: DoubleEndedIterator can call any of its methods on values of any implementing type. The compiler monomorphises generic uses for zero-cost dispatch, while dyn DoubleEndedIterator opts into runtime dispatch with a vtable. The full canonical path is std::iter::DoubleEndedIterator; bring it into scope with use std::iter::DoubleEndedIterator; or refer to it by its full path.

When to use it

Implement DoubleEndedIterator on your own types to plug into the standard library's abstractions, or accept `T: DoubleEndedIterator` as a generic bound on functions that can operate uniformly across every DoubleEndedIterator-shaped type.

Annotated examples

Implementing DoubleEndedIterator for your own type

use std::iter::DoubleEndedIterator;

struct MyType;

// Implementing a std trait is a contract — the compiler enforces every
// required method and associated type. Hover the trait in your editor
// to see the full list of required and provided methods.
impl DoubleEndedIterator for MyType {
    // Required methods go here. Refer to the std::iter::DoubleEndedIterator docs
    // for the exact signatures.
}

Implementing a std trait integrates your type with every generic function and adapter that takes `T: DoubleEndedIterator`. That single impl block can unlock dozens of free helper methods.

For deeper background, see the canonical Rust patterns reference for the broader context behind this section.

Calling generic code that requires DoubleEndedIterator

use std::iter::DoubleEndedIterator;

// `T: DoubleEndedIterator` is a trait bound — only types that implement DoubleEndedIterator
// are accepted. The body can use any of DoubleEndedIterator's methods, and the
// compiler monomorphises the function for each concrete T at the
// call site, so dispatch is always direct.
fn use_with_trait<T: DoubleEndedIterator>(value: T) {
    let _ = value;
}

// Same idea with `impl Trait` for a single argument — equivalent
// when T appears only once in the signature.
fn observe(value: impl DoubleEndedIterator) {
    let _ = value;
}

Trait bounds are the language's zero-cost generics. Each call site is monomorphised, so there is no virtual-dispatch overhead — `dyn DoubleEndedIterator` is a separate, runtime-dispatch flavour you opt into explicitly when you need heterogeneous values.

Common pitfalls

Don't implement DoubleEndedIterator for types that don't naturally satisfy its contract — implementing a trait dishonestly is one of the few ways Rust will compile but produce surprising behaviour at runtime. If you need dynamic dispatch, use `Box<dyn DoubleEndedIterator>` or `&dyn DoubleEndedIterator` instead of a generic parameter.

For deeper background, see an in-depth Rust idioms cheat sheet for the broader context behind this section.

Performance & threading notes

Generic uses (`T: DoubleEndedIterator`) are monomorphised — each concrete T gets its own specialised function, with no virtual-dispatch overhead. `dyn DoubleEndedIterator` adds a single indirection through a vtable, which is fast but blocks some compiler optimisations.

What DoubleEndedIterator does