Should consistent_generic_type include _, too, where it's accepted?

[paulo-ferraz-oliveira]

Consider the following code

-spec myfunc(_) -> _.
myfunc(_) ->

this is valid, but is it informative enough?

As-is, though, we could increase the list to also accept _ as a potential value.

[elbrujohalcon]

Personally, I distinguish between _ (this value is ignored) and term() | any() (this function can receive/produce elements of any type).

[paulo-ferraz-oliveira]

Sure, but I was wondering about it in a generic Erlang way. They represent the same thing under the hood, right?

Anyway, I'm Ok to not change it, but thought the discussion was start-able.