Hm, cool. They must be doing some internal type inference, insofar that's at all possible, then, in order to add such refactoring features.

Thanks for pointing it out!

How does it work for you in practice? E.g. does it help you what operations are possible on a function argument? (inside the function definition, I mean)