Clippy: Remove TyCtyt::super_traits_of public comment and test behavior

[xFrednet]

The TyCtyt::super_traits_of function is actually not needed. This PR removes the reference and adds a test as well :)

changelog: none

[rustbot]

r? @Jarcho

(rustbot has picked a reviewer for you, use r? to override)

[Jarcho]

That note is still accurate. while_let_on_iterator is working because the closure's kind is FnOnce.

I'm not sure if ty_is_fn_once_param is needed any more. It might be needed to handle cases where the closure is first assigned to a local, but that would have to be implemented first.

[xFrednet]

Could you give an example for code that wouldn't work? I tested this (Playground):

    // Lint if `FnOnce` is a super trait
    trait MySpecialFnMut: FnOnce() {}
    impl<T: FnOnce()> MySpecialFnMut for T {}
    fn f4(_: impl MySpecialFnMut) {}
    let mut it = 0..10;
    let x: Box<dyn MySpecialFnMut> = Box::new(|| {
        while let Some(x) = it.next() {
            if x % 2 == 0 {
                break;
            }
        }
    });
    f4(x);

and it lints just fine. I have the feeling I'm just overlooking something 😅

[Jarcho]

You're not missing anything. Looks like the closure's kind is now being set based on how it's used rather than how it's capable of being used. I guess the function can be removed then.

[bors]

☔ The latest upstream changes (presumably #11239) made this pull request unmergeable. Please resolve the merge conflicts.

[Jarcho]

Ping @xFrednet

ty_is_fn_once_param can just be removed and get_enclosing_loop_or_multi_call_closure can be simplified (is_once will always be false due to the closure kind check right before).

@rustbot author

[xFrednet]

Hey, year sorry, I wasn't sure if I could just remove the function and then I lost track of it 🙈

I'll close this in favor of #11812, since that includes the suggested change.