将类型投影应用于精炼类型

Question

Consider the following example:

trait T3

trait T2{
  type TT4
  type TT3 <: T3
}

trait T1{
  type TT2 <: T2
}

now I want to write a function the roughly speaking looks as

def test[T <: T1](t: T#TT2{type TT4 = Int}#TT3) = println(t)  //invalid syntax

which unfortunately is not a valid syntax. It is perfectly possible to write a function like this

def test[T <: T1](t: T#TT2#TT3) = println(t)

But I'd like to add a bit stricter restriction on T#TT2 making it to be a refined type T#TT2{ type TT4 = Int} .

Is there any workaround?

Answer 1

Try wrapping type T#TT2 { type TT4 = Int } in parenthesis before the final projection like so

def test[T <: T1](t: (T#TT2 { type TT4 = Int })#TT3) = ???

Types can always be wrapped in parentheses

SimpleType        ::=  SimpleType TypeArgs
                      |  SimpleType ‘#’ id
                      |  StableId
                      |  Path ‘.’ ‘type’
                      |  Literal
                      |  ‘(’ Types ‘)’      <======= note the parentheses

for example

scala> val xs: (List[(Int)]) = List(42)
val xs: List[Int] = List(42)