unconditional_recursion false positive in nightly

[zacknewman]

Summary

Using nightly, cargo clippy incorrectly reports an unconditional_recursion warning.

Lint Name

unconditional_recursion

Reproducer

I tried this code:

#[derive(PartialEq)]
pub struct Foo;
impl PartialEq<&Self> for Foo {
    #[deny(unconditional_recursion)]
    fn eq(&self, other: &&Self) -> bool {
        // This is _not_ recursive since this calls the derived `PartialEq` implementation.
        *self == **other
    }
}
// Uncommenting below will lead to a compilation failure since the compiler correctly detects
// an actual recursion.
//impl PartialEq<Foo> for &Foo {
//    #[deny(unconditional_recursion)]
//    fn eq(&self, other: &Foo) -> bool {
//        *self == *other
//    }
//}
#[cfg(test)]
mod tests {
    use super::Foo;
    #[test]
    fn not_recursive() {
        let x = Foo;
        let y = &x;
        // If the call were recursive, this won't ever finish but it does.
        assert!(x == y);
        // Same as above, but more obviously illustrates it's relying on the
        // "recursive" implementation.
        assert!(<Foo as PartialEq<&Foo>>::eq(&x, &y));
    }
}

I saw this happen:

[zack@laptop src]$ cargo clippy
    Checking bar v0.1.0 (/home/zack/projects/bar)
warning: function cannot return without recursing
 --> src/lib.rs:5:5
  |
5 | /     fn eq(&self, other: &&Self) -> bool {
6 | |         // This is _not_ recursive since this calls the derived `PartialEq` implementation.
7 | |         *self == **other
8 | |     }
  | |_____^
  |
note: recursive call site
 --> src/lib.rs:7:9
  |
7 |         *self == **other
  |         ^^^^^^^^^^^^^^^^
  = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#unconditional_recursion
  = note: `#[warn(clippy::unconditional_recursion)]` on by default

warning: `bar` (lib) generated 1 warning
    Finished dev [unoptimized + debuginfo] target(s) in 0.06s
[zack@laptop src]$ cargo t
    Finished test [unoptimized + debuginfo] target(s) in 0.03s
     Running unittests src/lib.rs (/home/zack/projects/bar/target/debug/deps/bar-5f4c34b6090a377c)

running 1 test
test tests::not_recursive ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s

   Doc-tests bar

running 0 tests

test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s

I expected to see this happen:
no warnings

Version

[zack@laptop src]$ rustc -Vv
rustc 1.78.0-nightly (8ace7ea1f 2024-02-07)
binary: rustc
commit-hash: 8ace7ea1f7cbba7b4f031e66c54ca237a0d65de6
commit-date: 2024-02-07
host: x86_64-unknown-linux-gnu
release: 1.78.0-nightly
LLVM version: 17.0.6
[zack@laptop src]$ uname -a
Linux laptop 6.7.4-arch1-1 #1 SMP PREEMPT_DYNAMIC Mon, 05 Feb 2024 22:07:49 +0000 x86_64 GNU/Linux

Additional Labels

No response

[y21]

Thanks for the report! This was fixed in #12177. One of the test cases it adds looks pretty much identical to your case :)

rust-clippy/tests/ui/unconditional_recursion.rs

Lines 310 to 320 in 62dcbd6

	// Not necessarily related to the issue but another FP from the http crate that was fixed with it:
	// https://docs.rs/http/latest/src/http/header/name.rs.html#1424
	// We used to simply peel refs from the LHS and RHS, so we couldn't differentiate
	// between `PartialEq<T> for &T` and `PartialEq<&T> for T` impls.
	#[derive(PartialEq)]
	struct HeaderName;
	impl<'a> PartialEq<&'a HeaderName> for HeaderName {
	fn eq(&self, other: &&'a HeaderName) -> bool {
	*self == **other
	}
	}

[zacknewman]

Apologize for the noise. I clearly did a poor job at looking at recently closed issues.