Allow static regions in type_name.

[nnethercote]

Fixes #146249.

r? @lcnr

[rustbot]

Some changes occurred to the CTFE machinery

cc @RalfJung, @oli-obk, @lcnr

[nnethercote]

@lcnr: I don't know if this static region is expected or not, I just assumed it was ok.

[theemathas]

Another test case:

struct Foo<T>(T);

impl Foo<&'static i32> {
    fn bar() {
        let f = || {};
        dbg!(std::any::type_name_of_val(&f));
    }
}

fn main() {
    <Foo<_>>::bar();
}

[theemathas]

Another two test cases:

trait Trait {
    type Assoc;
}

impl<'a> dyn Trait<Assoc = &'static i32> + 'a {
    fn bar() {
        let f = || {};
        dbg!(std::any::type_name_of_val(&f));
    }
}

fn main() {
    <dyn Trait<Assoc = &'static i32>>::bar();
}

struct Thing<T, U>(T, U);

trait Trait {
    type Assoc<T>;
}
impl Trait for () {
    type Assoc<T> = ();
}

impl<T: Trait> Thing<T, T::Assoc<&'static i32>> {
    fn bar() {
        let f = || {};
        dbg!(std::any::type_name_of_val(&f));
    }
}

fn main() {
    Thing::<(), ()>::bar();
}

[nnethercote]

Thanks for the extra test cases. I checked; they are all fixed by this PR's change.

[rustbot]

⚠️ Warning ⚠️

• There are issue links (such as #123) in the commit messages of the following commits.
  Please move them to the PR description, to avoid spamming the issues with references to the commit, and so this bot can automatically canonicalize them to avoid issues with subtree.
  • 9a52a83

[lcnr]

@bors r+ rollup

[bors]

📌 Commit 9a52a83 has been approved by lcnr

It is now in the queue for this repository.

[lqd]

since kind now exists, maybe

Suggested change

	match region.kind() {
	match kind {

could be good?

[nnethercote]

Whoops, yes, it was meant to be like that.