False positive with type_repetition_in_bounds and generics

[taiki-e]

type_repetition_in_bounds seems missed generics in types.

#![allow(dead_code)]
#![warn(clippy::type_repetition_in_bounds)]

pub struct Foo<A>(A);

pub struct Bar<A, B> {
    a: Foo<A>,
    b: Foo<B>,
}

impl<A, B> Unpin for Bar<A, B>
where
    Foo<A>: Unpin,
    Foo<B>: Unpin, //~ WARN this type has already been used as a bound predicate
{
}

[flip1995]

cc @xd009642 Any idea why this happens? Seems like the hashing of the types is missing the generics.

[xd009642]

I'll have a look tonight but that seems the most likely scenario

[xd009642]

So I can't see any parameters in Ty or TyKind for generic types associated with the type. There's 1rustc::hir::WhereBoundPredicate::bounded_generic_params` which maybe should be hashed? Provided that it's just the generics used in each bounded type. I'll have a play and see what I can do

[flip1995]

rustc::hir::WhereBoundPredicate::bounded_generic_params which maybe should be hashed

That sounds promising. Thanks for taking care of this!

[ThibsG]

I cannot reproduce this warning. Am I doing something wrong or the issue has gone?

[taiki-e]

I don't think this has been fixed: playground

[ThibsG]

Indeed, I cannot trigger this on my local Clippy test, I will figure out why

[taiki-e]

I cannot trigger this on my local Clippy test

I guess it's one of the following:

• compiletest may ignore the warning, so you might be able to reproduce this by using #![deny(...)] instead of #![warn(...)].
• This has been fixed in the master but not yet released.

[ThibsG]

Yes I just fixed it, something went wrong with deny/warn and my eyes.

Note: the problem is still here.