Compiler indefinitely takes resources when HRTB is used in conjunction with lifetime in GAT

[dodomorandi]

I tried this code:

pub trait A {
    type Item<'a>
    where
        Self: 'a;

    fn run(&mut self) -> Self::Item<'_>;
}

#[derive(Debug)]
pub struct S<F>(F);

impl<F> A for S<F>
where
    for<'a> F: FnMut(&Self::Item<'a>),
{
    type Item<'a> = i32
    where
        Self: 'a;

    fn run(&mut self) -> Self::Item<'_> {
        todo!()
    }
}

I expected to see this happen: the code should compile or return a compilation error in a short time.

Instead, this happened: the compiler start working and taking resources, apparently indefinitely.

Notice that removing the bound for F stops the issue to occur.

---

EDIT: the code triggers infinite recursion, the bound should be FnMut(&I::Item<'a>). As suggested by others, the issue of infinite recursion is managed by the next solver (using -Znext-solver), and thanks to that is also easy to spot the problem in the code.

Meta

rustc --version --verbose:
Stable:

rustc 1.76.0 (07dca489a 2024-02-04)
binary: rustc
commit-hash: 07dca489ac2d933c78d3c5158e3f43beefeb02ce
commit-date: 2024-02-04
host: x86_64-unknown-linux-gnu
release: 1.76.0
LLVM version: 17.0.6

Nightly:

rustc 1.78.0-nightly (c475e2303 2024-02-28)
binary: rustc
commit-hash: c475e2303b551d726307c646181e0677af1e0069
commit-date: 2024-02-28
host: x86_64-unknown-linux-gnu
release: 1.78.0-nightly
LLVM version: 18.1.0

The issue occurs both on stable and nightly.

[lqd]

This overflow should be detected by the next solver IIUC.

[clubby789]

Confirmed