Is it confused, or is it doing the right thing? As far as I can see "A() {}" defines a method named "A". It's unfortunate that this method hides the class named "A" in the outer scope, but it does. I think your example is equivalent to

class B {
  m a = new m();
  m() {}
}

In which case, I think the error that "m is not a type" is the correct error. (Where "correct" means "conforms to the specification".) It would be nice if we could always tell the difference between a method and a constructor, but we can't.

That said, I agree that we could do better. When we find an error of this type we could look in the outer scope to see if the name denotes a type somewhere else and provide a better error message, such as "The type 'A' cannot be used here because it is hidden by the method 'A'".

We might also want to provide a hint on the method 'A' that it is hiding the class A and provide a quick-fix to rename it to 'B', but that would have to be in addition to the error.

Removed this from the M6 milestone.
Added this to the Later milestone.

Removed Type-Defect label.
Added Type-Enhancement, Analyzer-UX labels.

Removed this from the Later milestone.
Added Oldschool-Milestone-Later label.

Removed Oldschool-Milestone-Later label.

I agree with @bwilkerson. Done?

We would like to improve the feedback when possible.