Name Shadowing

Member variables can be shadowed by local variables—in particular function arguments. This allows methods and constructors of the standard Java form:

1Public Class Shadow {
2  String s;
3  Shadow(String s) { this.s = s; } // Same name ok
4  setS(String s) { this.s = s; } // Same name ok
5}

Member variables in one class can shadow member variables with the same name in a parent classes. This can be useful if the two classes are in different top-level classes and written by different teams. For example, if one has a reference to a class C and wants to gain access to a member variable M in parent class P (with the same name as a member variable in C) the reference should be assigned to a reference to P first.

Static variables can be shadowed across the class hierarchy—so if P defines a static S, a subclass C can also declare a static S. References to S inside C refer to that static—in order to reference the one in P, the syntax P.S must be used.

Static class variables cannot be referenced through a class instance. They must be referenced using the raw variable name by itself (inside that top-level class file) or prefixed with the class name. For example:

1public class p1 { 
2  public static final Integer CLASS_INT = 1;
3  public class c { };
4}
5p1.c c = new p1.c();
6// This is illegal
7// Integer i = c.CLASS_INT;
8// This is correct
9Integer i = p1.CLASS_INT;