Scaladoc: reported type of Map#map

[jmgimeno]

The type reported for Map#map in the docs accessible from https://scala-lang.org/api/3.x/scala/collection/immutable/Map.html#map-fffffcb7 is

def map[K2, V2](f: (K, V) => (K2, V2)): Map[K2, V2]

but, if we go to its definition in the linked source file https://github.com/scala/scala/blob/v2.13.6/src/library/scala/collection/Map.scala#L299 we find that the real type is

def map[K2, V2](f: ((K, V)) => (K2, V2)): CC[K2, V2]

which, at least, I find it very confusing.

If I understand untupling correctly, the magic that allows calling a function which receives a tuple with a function with various parameters occurs at the call-site, so the documented type of the method should not change (or maybe this issue in the documentation has nothing to do with it).

[pikinier20]

The map method is defined in trait MapOps which takes a type parameter named CC. The docs that you provided lead to scala.collection.immutable.Map which mixins MapOps and applies type Map[K, V] in place of CC. Therefore, I think the shown type is ok.

Scaladoc 2 does the same: Map.map

[jmgimeno]

I know there is another version of map but the problem I report is that the shown type in the documentation does not match the type in the implementation. The type of the parameter changes.

[griggt]

def map[K2, V2](f:  (K, V)  => (K2, V2)): Map[K2, V2]
def map[K2, V2](f: ((K, V)) => (K2, V2)): CC[K2, V2]
                   ^^^^^^^^

[pikinier20]

Oh, sorry. I missed that. Now I found a place in Scaladoc where we omit brackets if the type matches some criteria. Probably it's the source of bug.