spec: clarify rules for uniqueness of method names

[momchil-velikov]

The specification states that

In a method set, each method must have a unique non-blank method name.

It also specifies that promoted methods are included in the method set of a struct:

Given a struct type S and a type named T, promoted methods are included in the method set of the struct as follows:

• If S contains an anonymous field T, the method sets of S and *S both include promoted methods with receiver T. The method set of *S also includes promoted methods with receiver *T.
• If S contains an anonymous field *T, the method sets of S and *S both include promoted methods with receiver T or *T.

These two excerpts seem to imply that there exist a conflict between a promoted method and a non-promoted method with the same name, or between two promoted methods, like in these examples:

package main
type A struct{}
func (A) F() {}
type B struct{ A }
func (B) F() {}
func main() {}

package main
type A struct{}
func (A) F() {}
type B struct{}
func (B) F() {}
type C struct {
    A
    B
}
func (C) F() {}
func main() {}

However, all of go, gccgo, and gotypes allow the above programs, so perhaps the intent is for the method name to be unique only among the non-promoted methods?

[ianlancetaylor]

The rules are spelled out at https://golang.org/ref/spec#Selectors . But the spec doesn't seem to clearly state that those rules apply not just to selectors, but also to method sets. At least, if it does state that, I can't find it.

[ianlancetaylor]

Oh, wait, I see it now. A field or method is promoted if x.f is valid. I think all is correct, if terse.