non-`Clone` type, with `Clone` associated type, as type param leads to confusing behavior in `#[derive(Clone)]` about rust HOT 7 CLOSED

If want to get rid of the extra Clone bound, you need to replace the #[derive(Clone)] with a manual Clone impl: Playground.

from rust.

#[derive] is not “perfect” (precise). The generated impl looks like:

impl<T: Clone/*<-- [!] */ + NotClone>Clone for ShouldBeClone<T>
where
    T::AssocClonable: Clone,
{ /* ... */ }

Meaning it generates a Clone impl that applies to all types T that impl Clone and NotClone etc.

I see. So this is why it is able to compile and only error when I try to call .clone()

from rust.

Apart from https://crates.io/crates/perfect-derive, there's also https://lib.rs/crates/derivative if you don't want to write the impl manually where you can specify if you want to omit certain bounds with attributes.

from rust.

You are basically looking for perfect derive:

• https://smallcultfollowing.com/babysteps/blog/2022/04/12/implied-bounds-and-perfect-derive/
• https://internals.rust-lang.org/t/overly-restrictive-trait-bounds-on-derives/17122
• https://crates.io/crates/perfect-derive
• #26925

from rust.

#[derive] is not “perfect” (precise). The generated impl looks like:

impl<T: Clone/*<-- [!] */ + NotClone>Clone for ShouldBeClone<T>
where
    T::AssocClonable: Clone,
{ /* ... */ }

Meaning it generates a Clone impl that applies to all types T that impl Clone and NotClone etc.

from rust.

Closing as duplicate of #26925.

from rust.

I see. So this is why it is able to compile and only error when I try to call .clone()

Exactly!

from rust.