---

 

For a  x that is not a , the selector expression

x.f

denotes the field or method f of the value x (or sometimes *x; see below). The identifier f is called the (field or method) selector; it must not be the . The type of the selector expression is the type of f. If x is a package name, see the section on .

A selector f may denote a field or method f of a type T, or it may refer to a field or method f of a nested  of T. The number of anonymous fields traversed to reachf is called its depth in T. The depth of a field or method f declared in T is zero. The depth of a field or method f declared in an anonymous field A in T is the depth of f in A plus one.

The following rules apply to selectors:

1. For a value x of type T or *T where T is not an interface type, x.f denotes the field or method at the shallowest depth in T where there is such an f. If there is not exactly with shallowest depth, the selector expression is illegal.
2. For a variable x of type I where I is an interface type, x.f denotes the actual method with name f of the value assigned to x. If there is no method with name f in the  of I, the selector expression is illegal.
3. In all other cases, x.f is illegal.
4. If x is of pointer type and has the value nil and x.f denotes a struct field, assigning to or evaluating x.f causes a .
5. If x is of interface type and has the value nil,  or  the method x.f causes a .

Selectors automatically  pointers to structs. If x is a pointer to a struct, x.y is shorthand for (*x).y; if the field y is also a pointer to a struct, x.y.z is shorthand for (*(*x).y).z, and so on. If x contains an anonymous field of type *A, where A is also a struct type, x.f is shorthand for (*x.A).f.

原文 

---