Add note when inherent impl for a alias type defined outside of the crate

[xizheyin]

Fixes #141679

r? compiler

[BoxyUwU]

r? compiler

[xizheyin]

@rustbot ready

[fee1-dead]

do you not get TyKind::Alias when you use type_of on the impl? coherence is probably fine to use middle queries

[xizheyin]

I've tried using type_of before, but I don't seem to be getting the information I'm looking for. When I try this code

            let ty = self.tcx.type_of(impl_def_id).skip_binder();
            if let ty::Adt(def, _) = ty.kind() {
                let def_id = def.did();
                let ty_name = self.tcx.def_path_str(def_id);
                let alias_ty_name = self.tcx.type_of(def_id).skip_binder().to_string();
                println!("ty_name: {:?}", ty_name);
                println!("alias_ty_name: {:?}", alias_ty_name);
            }

it prints:

ty_name: "Rc"
alias_ty_name: "Rc<T, A>"

It should be:

ty_name: "Function"
alias_ty_name: "Rc<Foo>"

Local variable ty is a adt of type Rc. Maybe the information we need only exists in HIR? Or am I using it wrong?

[xizheyin]

I merged the two ifs.

[compiler-errors]

r? compiler

[fmease]

Only under the experimental feature lazy_type_alias do we represent aliases in middle::ty IR, that's why type_of returns the expanded form (it gets expanded during HIR ty lowering for eager type aliases).

[estebank]

Let's deal with the structured suggestion later.

r=me after addressing nitpicks

View changes since this review

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[xizheyin]

@estebank I have rebased and fix the nits. CI green. Sorry for late response.