Respect MSRV (MAYBE_BOUND_IN_WHERE)

[sanxiyn]

Fix #12370.

changelog: [multiple_bound_locations]: respect MSRV for combining ?Sized bound

[rustbot]

r? @y21

rustbot has assigned @y21.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[y21]

+1 for having this be part of the trait_bounds pass, fits nicely with the other lints, and also nice seeing the cannot_combine_maybe_bound function get some use

[y21]

Can we have a test case for something like

fn f<'a, a: Sized>() where 'a: 'static {}

where the type parameter is named after a lifetime? I think this currently happens to work (ie not get linted) because the name of lifetimes include the quote, which differentiates them, but it would be good to have this as a guarantee in the testsuite.

[y21]

Since we now not only lint on functions, but also impls and a bunch of other items, I wonder how this interacts with derive macros such as Debug. Would be good to have a test, to make sure it doesn't lint here:

#[derive(Debug)]
struct S<T> where T: Sized { t: T }

[y21]

Hmm, looking more closely into the lints covered by this pass... Doesn't the type_repetition_in_bounds lint basically cover everything that this lint catches, but more thorough (as in, will also catch regular types and not just type parameters, though it is in the nursery category)?
Not necessarily related to what this PR addresses though, so that's for another time ^^

[bors]

☔ The latest upstream changes (presumably #12077) made this pull request unmergeable. Please resolve the merge conflicts.