Suggest using Self::AssociatedType for trait implementations

What it does

Suggests replacing a concrete type like State with the associated type of the trait.

Categories (optional)

• Kind: clippy::style or clippy::pedantic

What is the advantage of the recommended code over the original code

• the name of the associated type can be changed, without having to change the code that uses the associated type

Drawbacks

Might make the code more difficult to read.

Example

pub struct Example;

pub enum State {
    A,
    B,
    C,
}

pub trait SimpleTrait {
    type Associated;

    fn do_something_with_associated(&self, associated: &Self::Associated);
}

impl SimpleTrait for Example {
    type Associated = State;

    fn do_something_with_associated(&self, associated: &Self::Associated) {
        match associated {
            State::A => println!("State::A"),
            State::B => println!("State::B"),
            State::C => println!("State::C"),
        }
    }
}

Could be written as:

impl SimpleTrait for Example {
    type Associated = State;

    fn do_something_with_associated(&self, associated: &Self::Associated) {
        match associated {
            Self::Associated::A => println!("State::A"),
            Self::Associated::B => println!("State::B"),
            Self::Associated::C => println!("State::C"),
        }
    }
}

playground

Start with the Rust playground example in the issue and verify whether the proposed Self::Associated syntax is valid and appropriate for a Clippy lint. Done means a decided lint design that detects concrete associated-type references in trait implementations, with its category and behavior documented and covered by appropriate Clippy tests.