A closure is a block of code that can be referenced (and passed around) with access to the variables of the enclosing scope.
Since Java 1.1, anonymous inner class have provided this facility in a highly verbose manner. They also have a restriction of only being able to use final (and definitely assigned) local variables. (Note, even non-final local variables are in scope, but cannot be used.)
Java SE 8 is intended to have a more concise version of this for single-method interfaces*, called "lambdas". Lambdas have much the same restrictions as anonymous inner classes, although some details vary randomly.
*Originally the design was more flexible allowing Single Abstract Methods (SAM) types. Unfortunately the new design is less flexible, but does attempt to justify allowing implementation within interfaces.
A closure is a block of code that can be referenced (and passed
around) with access to the variables of the enclosing scope.
Now I'm trying to emulate the JavaScript closure example on Wikipedia, with a "straigth" translation to Java, in the hope to be useful:
//ECMAScript
var f, g;
function foo() {
var x = 0;
f = function() { return ++x; };
g = function() { return --x; };
x = 1;
print('inside foo, call to f(): ' + f()); // "2"
}
foo();
print('call to g(): ' + g()); // "1"
print('call to f(): ' + f()); // "2"
Now the java part: Function1 is "Functor" interface with arity 1 (one argument). Closure is the class implementing the Function1, a concrete Functor that acts as function (int -> int).
In the main() method I just instantiate foo as a Closure object, replicating the calls from the JavaScript example. The IntBox class is just a simple container, it behave like an array of 1 int:
int a[1] = {0}
interface Function1 {
public final IntBag value = new IntBag();
public int apply();
}
class Closure implements Function1 {
private IntBag x = value;
Function1 f;
Function1 g;
@Override
public int apply() {
// print('inside foo, call to f(): ' + f()); // "2"
// inside apply, call to f.apply()
System.out.println("inside foo, call to f.apply(): " + f.apply());
return 0;
}
public Closure() {
f = new Function1() {
@Override
public int apply() {
x.add(1);
return x.get();
}
};
g = new Function1() {
@Override
public int apply() {
x.add(-1);
return x.get();
}
};
// x = 1;
x.set(1);
}
}
public class ClosureTest {
public static void main(String[] args) {
// foo()
Closure foo = new Closure();
foo.apply();
// print('call to g(): ' + g()); // "1"
System.out.println("call to foo.g.apply(): " + foo.g.apply());
// print('call to f(): ' + f()); // "2"
System.out.println("call to foo.f.apply(): " + foo.f.apply());
}
}
It prints:
inside foo, call to f.apply(): 2
call to foo.g.apply(): 1
call to foo.f.apply(): 2