Thinking in Java, 3rd ed. Revision 4.0


prcdentsommairesuivant

8. Interfaces & Inner Classes

Interfaces and inner classes provide more sophisticated ways to organize and control the objects in your system.

C++, for example, does not contain such mechanisms, although the clever programmer may simulate them. The fact that they exist in Java indicates that they were considered important enough to provide direct support through language keywords.

In Chapter 7 you learned about the abstract keyword, which allows you to create one or more methods in a class that have no definitions-you provide part of the interface without providing a corresponding implementation, which is created by inheritors. The interface keyword produces a completely abstract class, one that provides no implementation at all. You'll learn that the interface is more than just an abstract class taken to the extreme, since it allows you to perform a variation on C++'s "multiple inheritance" by creating a class that can be upcast to more than one base type.

At first, inner classes look like a simple code-hiding mechanism: you place classes inside other classes. You'll learn, however, that the inner class does more than that-it knows about and can communicate with the surrounding class-and that the kind of code you can write with inner classes is more elegant and clear, although it is a new concept to most. It takes some time to become comfortable with design using inner classes.

8.1. Interfaces

The interface keyword takes the abstract concept one step further. You could think of it as a "pure" abstract class. It allows the creator to establish the form for a class: method names, argument lists, and return types, but no method bodies. An interface can also contain fields, but these are implicitly static and final. An interface provides only a form, but no implementation.

An interface says, "This is what all classes that implement this particular interface will look like." Thus, any code that uses a particular interface knows what methods might be called for that interface, and that's all. So the interface is used to establish a "protocol" between classes. (Some object-oriented programming languages have a keyword called protocol to do the same thing.)

To create an interface, use the interface keyword instead of the class keyword. Like a class, you can add the public keyword before the interface keyword (but only if that interface is defined in a file of the same name) or leave it off to give package access, so that it is only usable within the same package.

To make a class that conforms to a particular interface (or group of interfaces), use the implements keyword, which says, "The interface is what it looks like, but now I'm going to say how it works." Other than that, it looks like inheritance. The diagram for the instrument example shows this:

TIJ322.png

You can see from the Woodwind and Brass classes that once you've implemented an interface, that implementation becomes an ordinary class that can be extended in the regular way.

You can choose to explicitly declare the method declarations in an interface as public, but they are public even if you don't say it. So when you implement an interface, the methods from the interface must be defined as public. Otherwise, they would default to package access, and you'd be reducing the accessibility of a method during inheritance, which is not allowed by the Java compiler.

You can see this in the modified version of the Instrument example. Note that every method in the interface is strictly a declaration, which is the only thing the compiler allows. In addition, none of the methods in Instrument are declared as public, but they're automatically public anyway:

 
Sélectionnez

//: c08:music5:Music5.java
// Interfaces.
package c08.music5;
import com.bruceeckel.simpletest.*;
import c07.music.Note;
 
interface Instrument {
	// Compile-time constant:
	int I = 5; // static & final
	// Cannot have method definitions:
	void play(Note n); // Automatically public
	String what();
	void adjust();
}
 
class Wind implements Instrument {
	public void play(Note n) {
		System.out.println("Wind.play() " + n);
	}
	public String what() { return "Wind"; }
	public void adjust() {}
}
 
class Percussion implements Instrument {
	public void play(Note n) {
		System.out.println("Percussion.play() " + n);
	}
	public String what() { return "Percussion"; }
	public void adjust() {}
}
 
class Stringed implements Instrument {
	public void play(Note n) {
		System.out.println("Stringed.play() " + n);
	}
	public String what() { return "Stringed"; }
	public void adjust() {}
}
 
class Brass extends Wind {
	public void play(Note n) {
		System.out.println("Brass.play() " + n);
	}
	public void adjust() {
 		System.out.println("Brass.adjust()");
	}
}
 
class Woodwind extends Wind {
	public void play(Note n) {
		System.out.println("Woodwind.play() " + n);
	}
	public String what() { return "Woodwind"; }
}
 
public class Music5 {
	private static Test monitor = new Test();
	// Doesn't care about type, so new types
	// added to the system still work right:
	static void tune(Instrument i) {
		// ...
		i.play(Note.MIDDLE_C);
	}
	static void tuneAll(Instrument[] e) {
		for(int i = 0; i < e.length; i++)
			tune(e[i]);
	}
	public static void main(String[] args) {
		// Upcasting during addition to the array:
		Instrument[] orchestra = {
			new Wind(),
			new Percussion(),
			new Stringed(),
			new Brass(),
			new Woodwind()
		};
		tuneAll(orchestra);
		monitor.expect(new String[] {
			"Wind.play() Middle C",
			"Percussion.play() Middle C",
			"Stringed.play() Middle C",
			"Brass.play() Middle C",
			"Woodwind.play() Middle C"
		});
	}
} ///:~


The rest of the code works the same. It doesn't matter if you are upcasting to a "regular" class called Instrument, an abstract class called Instrument, or to an interface called Instrument. The behavior is the same. In fact, you can see in the tune() method that there isn't any evidence about whether Instrument is a "regular" class, an abstract class, or an interface. This is the intent: Each approach gives the programmer different control over the way objects are created and used.

8.1.1. "Multiple inheritance" in Java

The interface isn't simply a "more pure" form of abstract class. It has a higher purpose than that. Because an interface has no implementation at all-that is, there is no storage associated with an interface-there's nothing to prevent many interfaces from being combined. This is valuable because there are times when you need to say "An x is an a and a b and a c." In C++, this act of combining multiple class interfaces is called multiple inheritance, and it carries some rather sticky baggage because each class can have an implementation. In Java, you can perform the same act, but only one of the classes can have an implementation, so the problems seen in C++ do not occur with Java when combining multiple interfaces:

TIJ323.png

In a derived class, you aren't forced to have a base class that is either an abstract or "concrete" (one with no abstract methods). If you do inherit from a non-interface, you can inherit from only one. All the rest of the base elements must be interfaces. You place all the interface names after the implements keyword and separate them with commas. You can have as many interfaces as you want; each one becomes an independent type that you can upcast to. The following example shows a concrete class combined with several interfaces to produce a new class:

 
Sélectionnez

//: c08:Adventure.java
// Multiple interfaces.
 
interface CanFight {
	void fight();
}
 
interface CanSwim {
	void swim();
}
 
interface CanFly {
	void fly();
}
 
class ActionCharacter {
	public void fight() {}
}
 
class Hero extends ActionCharacter
		implements CanFight, CanSwim, CanFly {
	public void swim() {}
	public void fly() {}
}
 
public class Adventure {
	public static void t(CanFight x) { x.fight(); }
	public static void u(CanSwim x) { x.swim(); }
	public static void v(CanFly x) { x.fly(); }
	public static void w(ActionCharacter x) { x.fight(); }
	public static void main(String[] args) {
		Hero h = new Hero();
		t(h); // Treat it as a CanFight
		u(h); // Treat it as a CanSwim
		v(h); // Treat it as a CanFly
		w(h); // Treat it as an ActionCharacter
	}
} ///:~


You can see that Hero combines the concrete class ActionCharacter with the interfaces CanFight, CanSwim, and CanFly. When you combine a concrete class with interfaces this way, the concrete class must come first, then the interfaces. (The compiler gives an error otherwise.)

Note that the signature for fight() is the same in the interface CanFight and the class ActionCharacter, and that fight() is not provided with a definition in Hero. The rule for an interface is that you can inherit from it (as you will see shortly), but then you've got another interface. If you want to create an object of the new type, it must be a class with all definitions provided. Even though Hero does not explicitly provide a definition for fight(), the definition comes along with ActionCharacter, so it is automatically provided and it's possible to create objects of Hero.

In class Adventure, you can see that there are four methods that take as arguments the various interfaces and the concrete class. When a Hero object is created, it can be passed to any of these methods, which means it is being upcast to each interface in turn. Because of the way interfaces are designed in Java, this works without any particular effort on the part of the programmer.

Keep in mind that the core reason for interfaces is shown in the preceding example: to be able to upcast to more than one base type. However, a second reason for using interfaces is the same as using an abstract base class: to prevent the client programmer from making an object of this class and to establish that it is only an interface. This brings up a question: Should you use an interface or an abstract class? An interface gives you the benefits of an abstract class and the benefits of an interface, so if it's possible to create your base class without any method definitions or member variables, you should always prefer interfaces to abstract classes. In fact, if you know something is going to be a base class, your first choice should be to make it an interface, and only if you're forced to have method definitions or member variables should you change to an abstract class, or if necessary a concrete class.

8.1.1.1. Name collisions when combining interfaces

You can encounter a small pitfall when implementing multiple interfaces. In the preceding example, both CanFight and ActionCharacter have an identical void fight() method. This is not a problem, because the method is identical in both cases. But what if it isn't? Here's an example:

 
Sélectionnez

//: c08:InterfaceCollision.java
 
interface I1 { void f(); }
interface I2 { int f(int i); }
interface I3 { int f(); }
class C { public int f() { return 1; } }
 
class C2 implements I1, I2 {
	public void f() {}
	public int f(int i) { return 1; } // overloaded
}
 
class C3 extends C implements I2 {
	public int f(int i) { return 1; } // overloaded
}
 
class C4 extends C implements I3 {
	// Identical, no problem:
	public int f() { return 1; }
}
 
// Methods differ only by return type:
//! class C5 extends C implements I1 {}
//! interface I4 extends I1, I3 {} ///:~


The difficulty occurs because overriding, implementation, and overloading get unpleasantly mixed together, and overloaded methods cannot differ only by return type. When the last two lines are uncommented, the error messages say it all:

InterfaceCollision.java:23: f() in C cannot implement f() in I1; attempting to use incompatible return type
found : int
required: void
InterfaceCollision.java:24: interfaces I3 and I1 are incompatible; both define f(), but with different return type

Using the same method names in different interfaces that are intended to be combined generally causes confusion in the readability of the code, as well. Strive to avoid it.

8.1.2. Extending an interface with inheritance

You can easily add new method declarations to an interface by using inheritance, and you can also combine several interfaces into a new interface with inheritance. In both cases you get a new interface, as seen in this example:

 
Sélectionnez

//: c08:HorrorShow.java
// Extending an interface with inheritance.
 
interface Monster {
	void menace();
}
 
interface DangerousMonster extends Monster {
	void destroy();
}
 
interface Lethal {
	void kill();
}
 
class DragonZilla implements DangerousMonster {
	public void menace() {}
	public void destroy() {}
}
 
interface Vampire extends DangerousMonster, Lethal {
	void drinkBlood();
}
 
class VeryBadVampire implements Vampire {
	public void menace() {}
	public void destroy() {}
	public void kill() {}
	public void drinkBlood() {}
}
 
public class HorrorShow {
	static void u(Monster b) { b.menace(); }
	static void v(DangerousMonster d) {
		d.menace();
		d.destroy();
	}
	static void w(Lethal l) { l.kill(); }
	public static void main(String[] args) {
		DangerousMonster barney = new DragonZilla();
		u(barney);
		v(barney);
		Vampire vlad = new VeryBadVampire();
		u(vlad);
		v(vlad);
		w(vlad);
	}
} ///:~


DangerousMonster is a simple extension to Monster that produces a new interface. This is implemented in DragonZilla.

The syntax used in Vampire works only when inheriting interfaces. Normally, you can use extends with only a single class, but since an interface can be made from multiple other interfaces, extends can refer to multiple base interfaces when building a new interface. As you can see, the interface names are simply separated with commas.

8.1.3. Grouping constants

Because any fields you put into an interface are automatically static and final, the interface is a convenient tool for creating groups of constant values, much as you would with an enum in C or C++. For example:

 
Sélectionnez

//: c08:Months.java
// Using interfaces to create groups of constants.
package c08;
 
public interface Months {
	int
		JANUARY = 1, FEBRUARY = 2, MARCH = 3,
		APRIL = 4, MAY = 5, JUNE = 6, JULY = 7,
		AUGUST = 8, SEPTEMBER = 9, OCTOBER = 10,
		NOVEMBER = 11, DECEMBER = 12;
} ///:~


Notice the Java style of using all uppercase letters (with underscores to separate multiple words in a single identifier) for static finals that have constant initializers.

The fields in an interface are automatically public, so it's unnecessary to specify that.

You can use the constants from outside the package by importing c08.* or c08.Months just as you would with any other package, and referencing the values with expressions like Months.JANUARY. Of course, what you get is just an int, so there isn't the extra type safety that C++'s enum has, but this (commonly used) technique is certainly an improvement over hard coding numbers into your programs. (That approach is often referred to as using "magic numbers," and it produces very difficult-to-maintain code.)

If you do want extra type safety, you can build a class like this: (33)

 
Sélectionnez

//: c08:Month.java
// A more robust enumeration system.
package c08;
import com.bruceeckel.simpletest.*;
 
public final class Month {
	private static Test monitor = new Test();
	private String name;
	private Month(String nm) { name = nm; }
	public String toString() { return name; }
	public static final Month
		JAN = new Month("January"),
		FEB = new Month("February"),
		MAR = new Month("March"),
		APR = new Month("April"),
		MAY = new Month("May"),
		JUN = new Month("June"),
		JUL = new Month("July"),
		AUG = new Month("August"),
		SEP = new Month("September"),
		OCT = new Month("October"),
		NOV = new Month("November"),
		DEC = new Month("December");
	public static final Month[] month =  {
		JAN, FEB, MAR, APR, MAY, JUN,
		JUL, AUG, SEP, OCT, NOV, DEC
	};
	public static final Month number(int ord) {
		return month[ord - 1];
	}
	public static void main(String[] args) {
		Month m = Month.JAN;
		System.out.println(m);
		m = Month.number(12);
		System.out.println(m);
		System.out.println(m == Month.DEC);
		System.out.println(m.equals(Month.DEC));
		System.out.println(Month.month[3]);
		monitor.expect(new String[] {
			"January",
			"December",
			"true",
			"true",
			"April"
		});
	}
} ///:~


Month is a final class with a private constructor, so no one can inherit from it or make any instances of it. The only instances are the final static ones created in the class itself: JAN, FEB, MAR, etc. These objects are also used in the array month, which lets you iterate through an array of Month2 objects. The number() method allows you to select a Month by giving its corresponding month number. In main() you can see the type safety; m is a Month object so it can be assigned only to a Month. The previous example Months.java provided only int values, so an int variable intended to represent a month could actually be given any integer value, which wasn't very safe.

This approach also allows you to use == or equals() interchangeably, as shown at the end of main(). This works because there can be only one instance of each value of Month. In Chapter 11 you'll learn about another way to set up classes so the objects can be compared to each other.

There's also a month field in java.util.Calendar.

Apache's Jakarta Commons project contains tools to create enumerations similar to what's shown in the preceding example, but with less effort. See http://jakarta.apache.org/commons, under "lang," in the package org.apache.commons.lang.enum. This project also has many other potentially useful libraries.

8.1.4. Initializing fields in interfaces

Fields defined in interfaces are automatically static and final. These cannot be "blank finals," but they can be initialized with nonconstant expressions. For example:

 
Sélectionnez

//: c08:RandVals.java
// Initializing interface fields with
// non-constant initializers.
import java.util.*;
 
public interface RandVals {
	Random rand = new Random();
	int randomInt = rand.nextInt(10);
	long randomLong = rand.nextLong() * 10;
	float randomFloat = rand.nextLong() * 10;
	double randomDouble = rand.nextDouble() * 10;
} ///:~


Since the fields are static, they are initialized when the class is first loaded, which happens when any of the fields are accessed for the first time. Here's a simple test:

 
Sélectionnez

//: c08:TestRandVals.java
import com.bruceeckel.simpletest.*;
 
public class TestRandVals {
	private static Test monitor = new Test();
	public static void main(String[] args) {
		System.out.println(RandVals.randomInt);
		System.out.println(RandVals.randomLong);
		System.out.println(RandVals.randomFloat);
		System.out.println(RandVals.randomDouble);
		monitor.expect(new String[] {
			"%% -?\\d+",
			"%% -?\\d+",
			"%% -?\\d\\.\\d+E?-?\\d+",
			"%% -?\\d\\.\\d+E?-?\\d+"
		});
	}
} ///:~


The fields, of course, are not part of the interface but instead are stored in the static storage area for that interface.

8.1.5. Nesting interfaces

Interfaces may be nested within classes and within other interfaces. (34) This reveals a number of very interesting features:

 
Sélectionnez

//: c08:nesting:NestingInterfaces.java
package c08.nesting;
 
class A {
	interface B {
		void f();
	}
	public class BImp implements B {
		public void f() {}
	}
	private class BImp2 implements B {
		public void f() {}
	}
	public interface C {
		void f();
	}
	class CImp implements C {
		public void f() {}
	}
	private class CImp2 implements C {
		public void f() {}
	}
	private interface D {
		void f();
	}
	private class DImp implements D {
		public void f() {}
	}
	public class DImp2 implements D {
		public void f() {}
	}
	public D getD() { return new DImp2(); }
	private D dRef;
	public void receiveD(D d) {
		dRef = d;
		dRef.f();
	}
}
 
interface E {
	interface G {
		void f();
	}
	// Redundant "public":
	public interface H {
		void f();
	}
	void g();
	// Cannot be private within an interface:
	//! private interface I {}
}
 
public class NestingInterfaces {
	public class BImp implements A.B {
		public void f() {}
	}
	class CImp implements A.C {
		public void f() {}
	}
	// Cannot implement a private interface except
	// within that interface's defining class:
	//! class DImp implements A.D {
	//!  public void f() {}
	//! }
	class EImp implements E {
		public void g() {}
	}
	class EGImp implements E.G {
		public void f() {}
	}
	class EImp2 implements E {
		public void g() {}
		class EG implements E.G {
			public void f() {}
		}
	}
	public static void main(String[] args) {
		A a = new A();
		// Can't access A.D:
		//! A.D ad = a.getD();
		// Doesn't return anything but A.D:
		//! A.DImp2 di2 = a.getD();
		// Cannot access a member of the interface:
		//! a.getD().f();
		// Only another A can do anything with getD():
		A a2 = new A();
		a2.receiveD(a.getD());
	}
} ///:~


The syntax for nesting an interface within a class is reasonably obvious, and just like non-nested interfaces, these can have public or package-access visibility. You can also see that both public and package-access nested interfaces can be implemented as public, package-access, and private nested classes.

As a new twist, interfaces can also be private, as seen in A.D (the same qualification syntax is used for nested interfaces as for nested classes). What good is a private nested interface? You might guess that it can only be implemented as a private inner class as in DImp, but A.DImp2 shows that it can also be implemented as a public class. However, A.DImp2 can only be used as itself. You are not allowed to mention the fact that it implements the private interface, so implementing a private interface is a way to force the definition of the methods in that interface without adding any type information (that is, without allowing any upcasting).

The method getD() produces a further quandary concerning the private interface: It's a public method that returns a reference to a private interface. What can you do with the return value of this method? In main(), you can see several attempts to use the return value, all of which fail. The only thing that works is if the return value is handed to an object that has permission to use it-in this case, another A, via the receiveD() method.

Interface E shows that interfaces can be nested within each other. However, the rules about interfaces-in particular, that all interface elements must be public-are strictly enforced here, so an interface nested within another interface is automatically public and cannot be made private.

NestingInterfaces shows the various ways that nested interfaces can be implemented. In particular, notice that when you implement an interface, you are not required to implement any interfaces nested within. Also, private interfaces cannot be implemented outside of their defining classes.

Initially, these features may seem like they are added strictly for syntactic consistency, but I generally find that once you know about a feature, you often discover places where it is useful.

8.2. Inner classes

It's possible to place a class definition within another class definition. This is called an inner class. The inner class is a valuable feature because it allows you to group classes that logically belong together and to control the visibility of one within the other. However, it's important to understand that inner classes are distinctly different from composition.

While you're learning about them, the need for inner classes isn't always obvious. At the end of this section, after all of the syntax and semantics of inner classes have been described, you'll find examples that should begin to make clear the benefits of inner classes.

You create an inner class just as you'd expect-by placing the class definition inside a surrounding class:

 
Sélectionnez

//: c08:Parcel1.java
// Creating inner classes.
 
public class Parcel1 {
	class Contents {
		private int i = 11;
		public int value() { return i; }
	}
	class Destination {
		private String label;
		Destination(String whereTo) {
			label = whereTo;
		}
		String readLabel() { return label; }
	}
	// Using inner classes looks just like
	// using any other class, within Parcel1:
	public void ship(String dest) {
		Contents c = new Contents();
		Destination d = new Destination(dest);
		System.out.println(d.readLabel());
	}
	public static void main(String[] args) {
		Parcel1 p = new Parcel1();
		p.ship("Tanzania");
	}
} ///:~


The inner classes, when used inside ship(), look just like the use of any other classes. Here, the only practical difference is that the names are nested within Parcel1. You'll see in a while that this isn't the only difference.

More typically, an outer class will have a method that returns a reference to an inner class, like this:

 
Sélectionnez

//: c08:Parcel2.java
// Returning a reference to an inner class.
 
public class Parcel2 {
	class Contents {
		private int i = 11;
		public int value() { return i; }
	}
	class Destination {
		private String label;
		Destination(String whereTo) {
			label = whereTo;
		}
		String readLabel() { return label; }
	}
	public Destination to(String s) {
		return new Destination(s);
	}
	public Contents cont() {
		return new Contents();
	}
	public void ship(String dest) {
		Contents c = cont();
		Destination d = to(dest);
		System.out.println(d.readLabel());
	}
	public static void main(String[] args) {
		Parcel2 p = new Parcel2();
		p.ship("Tanzania");
		Parcel2 q = new Parcel2();
		// Defining references to inner classes:
		Parcel2.Contents c = q.cont();
		Parcel2.Destination d = q.to("Borneo");
	}
} ///:~


If you want to make an object of the inner class anywhere except from within a non-static method of the outer class, you must specify the type of that object as OuterClassName.InnerClassName, as seen in main().

8.2.1. Inner classes and upcasting

So far, inner classes don't seem that dramatic. After all, if it's hiding you're after, Java already has a perfectly good hiding mechanism-just give the class package access (visible only within a package) rather than creating it as an inner class.

However, inner classes really come into their own when you start upcasting to a base class, and in particular to an interface. (The effect of producing an interface reference from an object that implements it is essentially the same as upcasting to a base class.) That's because the inner class-the implementation of the interface-can then be completely unseen and unavailable to anyone, which is convenient for hiding the implementation. All you get back is a reference to the base class or the interface.

First, the common interfaces will be defined in their own files so they can be used in all the examples:

 
Sélectionnez

//: c08:Destination.java
public interface Destination {
	String readLabel();
} ///:~


 
Sélectionnez

//: c08:Contents.java
public interface Contents {
	int value();
} ///:~


Now Contents and Destination represent interfaces available to the client programmer. (The interface, remember, automatically makes all of its members public.)

When you get back a reference to the base class or the interface, it's possible that you can't even find out the exact type, as shown here:

 
Sélectionnez

//: c08:TestParcel.java
// Returning a reference to an inner class.
 
class Parcel3 {
	private class PContents implements Contents {
		private int i = 11;
		public int value() { return i; }
	}
	protected class PDestination implements Destination {
		private String label;
		private PDestination(String whereTo) {
			label = whereTo;
		}
		public String readLabel() { return label; }
	}
	public Destination dest(String s) {
		return new PDestination(s);
	}
	public Contents cont() {
		return new PContents();
	}
}
 
public class TestParcel {
	public static void main(String[] args) {
		Parcel3 p = new Parcel3();
		Contents c = p.cont();
		Destination d = p.dest("Tanzania");
		// Illegal -- can't access private class:
		//! Parcel3.PContents pc = p.new PContents();
	}
} ///:~


In the example, main() must be in a separate class in order to demonstrate the privateness of the inner class PContents.

In Parcel3, something new has been added: The inner class PContents is private, so no one but Parcel3 can access it. PDestination is protected, so no one but Parcel3, classes in the same package (since protected also gives package access), and the inheritors of Parcel3 can access PDestination. This means that the client programmer has restricted knowledge and access to these members. In fact, you can't even downcast to a private inner class (or a protected inner class unless you're an inheritor), because you can't access the name, as you can see in class TestParcel. Thus, the private inner class provides a way for the class designer to completely prevent any type-coding dependencies and to completely hide details about implementation. In addition, extension of an interface is useless from the client programmer's perspective since the client programmer cannot access any additional methods that aren't part of the public interface. This also provides an opportunity for the Java compiler to generate more efficient code.

Normal (non-inner) classes cannot be made private or protected; they may only be given public or package access.

8.2.2. Inner classes in methods and scopes

What you've seen so far encompasses the typical use for inner classes. In general, the code that you'll write and read involving inner classes will be "plain" inner classes that are simple and easy to understand. However, the design for inner classes is quite complete, and there are a number of other, more obscure, ways that you can use them if you choose; inner classes can be created within a method or even an arbitrary scope. There are two reasons for doing this:

  • As shown previously, you're implementing an interface of some kind so that you can create and return a reference.
  • You're solving a complicated problem and you want to create a class to aid in your solution, but you don't want it publicly available.

In the following examples, the previous code will be modified to use:

  • A class defined within a method
  • A class defined within a scope inside a method
  • An anonymous class implementing an interface
  • An anonymous class extending a class that has a nondefault constructor
  • An anonymous class that performs field initialization
  • An anonymous class that performs construction using instance initialization (anonymous inner classes cannot have constructors)

Although it's an ordinary class with an implementation, Wrapping is also being used as a common "interface" to its derived classes:

 
Sélectionnez

//: c08:Wrapping.java
public class Wrapping {
	private int i;
	public Wrapping(int x) { i = x; }
	public int value() { return i; }
} ///:~


You'll notice that Wrapping has a constructor that requires an argument, to make things a bit more interesting.

The first example shows the creation of an entire class within the scope of a method (instead of the scope of another class). This is called a local inner class:

 
Sélectionnez

//: c08:Parcel4.java
// Nesting a class within a method.
 
public class Parcel4 {
	public Destination dest(String s) {
		class PDestination implements Destination {
			private String label;
			private PDestination(String whereTo) {
				label = whereTo;
			}
			public String readLabel() { return label; }
		}
		return new PDestination(s);
	}
	public static void main(String[] args) {
		Parcel4 p = new Parcel4();
		Destination d = p.dest("Tanzania");
	}
} ///:~


The class PDestination is part of dest() rather than being part of Parcel4. (Also notice that you could use the class identifier PDestination for an inner class inside each class in the same subdirectory without a name clash.) Therefore, PDestination cannot be accessed outside of dest(). Notice the upcasting that occurs in the return statement-nothing comes out of dest() except a reference to Destination, the base class. Of course, the fact that the name of the class PDestination is placed inside dest() doesn't mean that PDestination is not a valid object once dest() returns.

The next example shows how you can nest an inner class within any arbitrary scope:

 
Sélectionnez

//: c08:Parcel5.java
// Nesting a class within a scope.
 
public class Parcel5 {
	private void internalTracking(boolean b) {
		if(b) {
			class TrackingSlip {
				private String id;
				TrackingSlip(String s) {
					id = s;
				}
				String getSlip() { return id; }
			}
			TrackingSlip ts = new TrackingSlip("slip");
			String s = ts.getSlip();
		}
		// Can't use it here! Out of scope:
		//! TrackingSlip ts = new TrackingSlip("x");
	}
	public void track() { internalTracking(true); }
	public static void main(String[] args) {
		Parcel5 p = new Parcel5();
		p.track();
	}
} ///:~


The class TrackingSlip is nested inside the scope of an if statement. This does not mean that the class is conditionally created-it gets compiled along with everything else. However, it's not available outside the scope in which it is defined.Other than that, it looks just like an ordinary class.

8.2.3. Anonymous inner classes

The next example looks a little strange:

 
Sélectionnez

//: c08:Parcel6.java
// A method that returns an anonymous inner class.
 
public class Parcel6 {
	public Contents cont() {
		return new Contents() {
			private int i = 11;
			public int value() { return i; }
		}; // Semicolon required in this case
	}
	public static void main(String[] args) {
		Parcel6 p = new Parcel6();
		Contents c = p.cont();
	}
} ///:~


The cont() method combines the creation of the return value with the definition of the class that represents that return value! In addition, the class is anonymous; it has no name. To make matters a bit worse, it looks like you're starting out to create a Contents object:

 
Sélectionnez

return new Contents()


But then, before you get to the semicolon, you say, "But wait, I think I'll slip in a class definition":

 
Sélectionnez

return new Contents() {
	private int i = 11;
	public int value() { return i; }
};


What this strange syntax means is: "Create an object of an anonymous class that's inherited from Contents." The reference returned by the new expression is automatically upcast to a Contents reference. The anonymous inner-class syntax is a shorthand for:

 
Sélectionnez

class MyContents implements Contents {
	private int i = 11;
	public int value() { return i; }
}
return new MyContents();


In the anonymous inner class, Contents is created by using a default constructor. The following code shows what to do if your base class needs a constructor with an argument:

 
Sélectionnez

//: c08:Parcel7.java
// An anonymous inner class that calls
// the base-class constructor.
 
public class Parcel7 {
	public Wrapping wrap(int x) {
		// Base constructor call:
		return new Wrapping(x) { // Pass constructor argument.
			public int value() {
				return super.value() * 47;
			}
		}; // Semicolon required
	}
	public static void main(String[] args) {
		Parcel7 p = new Parcel7();
		Wrapping w = p.wrap(10);
	}
} ///:~


That is, you simply pass the appropriate argument to the base-class constructor, seen here as the x passed in new Wrapping(x).

The semicolon at the end of the anonymous inner class doesn't mark the end of the class body (as it does in C++). Instead, it marks the end of the expression that happens to contain the anonymous class. Thus, it's identical to the use of the semicolon everywhere else.

You can also perform initialization when you define fields in an anonymous class:

 
Sélectionnez

//: c08:Parcel8.java
// An anonymous inner class that performs
// initialization. A briefer version of Parcel4.java.
 
public class Parcel8 {
	// Argument must be final to use inside
	// anonymous inner class:
	public Destination dest(final String dest) {
		return new Destination() {
			private String label = dest;
			public String readLabel() { return label; }
		};
	}
	public static void main(String[] args) {
		Parcel8 p = new Parcel8();
		Destination d = p.dest("Tanzania");
	}
} ///:~


If you're defining an anonymous inner class and want to use an object that's defined outside the anonymous inner class, the compiler requires that the argument reference be final, like the argument to dest().If you forget, you'll get a compile-time error message.

As long as you're simply assigning a field, the approach in this example is fine. But what if you need to perform some constructor-like activity? You can't have a named constructor in an anonymous class (since there's no name!), but with instance initialization, you can, in effect, create a constructor for an anonymous inner class, like this:

 
Sélectionnez

//: c08:AnonymousConstructor.java
// Creating a constructor for an anonymous inner class.
import com.bruceeckel.simpletest.*;
 
abstract class Base {
	public Base(int i) {
		System.out.println("Base constructor, i = " + i);
	}
	public abstract void f();
}
 
public class AnonymousConstructor {
	private static Test monitor = new Test();
	public static Base getBase(int i) {
		return new Base(i) {
			{
				System.out.println("Inside instance initializer");
			}
			public void f() {
				System.out.println("In anonymous f()");
			}
		};
	}
	public static void main(String[] args) {
		Base base = getBase(47);
		base.f();
		monitor.expect(new String[] {
			"Base constructor, i = 47",
			"Inside instance initializer",
			"In anonymous f()"
		});
	}
} ///:~


In this case, the variable i did not have to be final. While i is passed to the base constructor of the anonymous class, it is never directly used inside the anonymous class.

Here's the "parcel" theme with instance initialization. Note that the arguments to dest() must be final since they are used within the anonymous class:

 
Sélectionnez

//: c08:Parcel9.java
// Using "instance initialization" to perform
// construction on an anonymous inner class.
import com.bruceeckel.simpletest.*;
 
public class Parcel9 {
	private static Test monitor = new Test();
	public Destination
	dest(final String dest, final float price) {
		return new Destination() {
			private int cost;
			// Instance initialization for each object:
			{
				cost = Math.round(price);
				if(cost > 100)
					System.out.println("Over budget!");
			}
			private String label = dest;
			public String readLabel() { return label; }
		};
	}
	public static void main(String[] args) {
		Parcel9 p = new Parcel9();
		Destination d = p.dest("Tanzania", 101.395F);
		monitor.expect(new String[] {
			"Over budget!"
		});
	}
} ///:~


Inside the instance initializer you can see code that couldn't be executed as part of a field initializer (that is, the if statement). So in effect, an instance initializer is the constructor for an anonymous inner class. Of course, it's limited; you can't overload instance initializers, so you can have only one of these constructors.

8.2.4. The link to the outer class

So far, it appears that inner classes are just a name-hiding and code-organization scheme, which is helpful but not totally compelling. However, there's another twist. When you create an inner class, an object of that inner class has a link to the enclosing object that made it, and so it can access the members of that enclosing object-without any special qualifications. In addition, inner classes have access rights to all the elements in the enclosing class. (35) The following example demonstrates this:

 
Sélectionnez

//: c08:Sequence.java
// Holds a sequence of Objects.
import com.bruceeckel.simpletest.*;
 
interface Selector {
	boolean end();
	Object current();
	void next();
}
 
public class Sequence {
	private static Test monitor = new Test();
	private Object[] objects;
	private int next = 0;
	public Sequence(int size) { objects = new Object[size]; }
	public void add(Object x) {
		if(next < objects.length)
			objects[next++] = x;
	}
	private class SSelector implements Selector {
		private int i = 0;
		public boolean end() { return i == objects.length; }
		public Object current() { return objects[i]; }
		public void next() { if(i < objects.length) i++; }
	}
	public Selector getSelector() { return new SSelector(); }
	public static void main(String[] args) {
		Sequence sequence = new Sequence(10);
		for(int i = 0; i < 10; i++)
			sequence.add(Integer.toString(i));
		Selector selector = sequence.getSelector();
		while(!selector.end()) {
			System.out.println(selector.current());
			selector.next();
		}
		monitor.expect(new String[] {
			"0",
			"1",
			"2",
			"3",
			"4",
			"5",
			"6",
			"7",
			"8",
			"9"
		});
	}
} ///:~


The Sequence is simply a fixed-sized array of Object with a class wrapped around it. You call add() to add a new Object to the end of the sequence (if there's room left). To fetch each of the objects in a Sequence, there's an interface called Selector, which allows you to see if you're at the end(), to look at the current() Object, and to move to the next() Object in the Sequence. Because Selector is an interface, many other classes can implement the interface in their own ways, and many methods can take the interface as an argument, in order to create generic code.

Here, the SSelector is a private class that provides Selector functionality. In main(), you can see the creation of a Sequence, followed by the addition of a number of String objects. Then, a Selector is produced with a call to getSelector(), and this is used to move through the Sequence and select each item.

At first, the creation of SSelector looks like just another inner class. But examine it closely. Note that each of the methods-end(), current(), and next()-refer to objects, which is a reference that isn't part of SSelector, but is instead a private field in the enclosing class. However, the inner class can access methods and fields from the enclosing class as if it owned them. This turns out to be very convenient, as you can see in the preceding example.

So an inner class has automatic access to the members of the enclosing class. How can this happen? The inner class must keep a reference to the particular object of the enclosing class that was responsible for creating it. Then, when you refer to a member of the enclosing class, that (hidden) reference is used to select that member. Fortunately, the compiler takes care of all these details for you, but you can also understand now that an object of an inner class can be created only in association with an object of the enclosing class. Construction of the inner class object requires the reference to the object of the enclosing class, and the compiler will complain if it cannot access that reference. Most of the time this occurs without any intervention on the part of the programmer.

8.2.5. Nested classes

If you don't need a connection between the inner class object and the outer class object, then you can make the inner class static. This is commonly called a nested class. (36) To understand the meaning of static when applied to inner classes, you must remember that the object of an ordinary inner class implicitly keeps a reference to the object of the enclosing class that created it. This is not true, however, when you say an inner class is static. A nested class means:

  • You don't need an outer-class object in order to create an object of a nested class.
  • You can't access a non-static outer-class object from an object of a nested class.

Nested classes are different from ordinary inner classes in another way, as well. Fields and methods in ordinary inner classes can only be at the outer level of a class, so ordinary inner classes cannot have static data, static fields, or nested classes. However, nested classes can have all of these:

 
Sélectionnez

//: c08:Parcel10.java
// Nested classes (static inner classes).
 
public class Parcel10 {
	private static class ParcelContents implements Contents {
		private int i = 11;
		public int value() { return i; }
	}
	protected static class ParcelDestination
	implements Destination {
		private String label;
		private ParcelDestination(String whereTo) {
			label = whereTo;
		}
		public String readLabel() { return label; }
		// Nested classes can contain other static elements:
		public static void f() {}
		static int x = 10;
		static class AnotherLevel {
			public static void f() {}
			static int x = 10;
		}
	}
	public static Destination dest(String s) {
		return new ParcelDestination(s);
	}
	public static Contents cont() {
		return new ParcelContents();
	}
	public static void main(String[] args) {
		Contents c = cont();
		Destination d = dest("Tanzania");
	}
} ///:~


In main(), no object of Parcel10 is necessary; instead, you use the normal syntax for selecting a static member to call the methods that return references to Contents and Destination.

As you will see shortly, in an ordinary (non-static) inner class, the link to the outer class object is achieved with a special this reference. A nested class does not have this special this reference, which makes it analogous to a static method.

Normally, you can't put any code inside an interface, but a nested class can be part of an interface. Since the class is static,it doesn't violate the rules for interfaces-the nested class is only placed inside the namespace of the interface:

 
Sélectionnez

//: c08:IInterface.java
// Nested classes inside interfaces.
 
public interface IInterface {
	static class Inner {
		int i, j, k;
		public Inner() {}
		void f() {}
	}
} ///:~


Earlier in this book I suggested putting a main() in every class to act as a test bed for that class. One drawback to this is the amount of extra compiled code you must carry around. If this is a problem, you can use a nested class to hold your test code:

 
Sélectionnez

//: c08:TestBed.java
// Putting test code in a nested class.
 
public class TestBed {
	public TestBed() {}
	public void f() { System.out.println("f()"); }
	public static class Tester {
		public static void main(String[] args) {
			TestBed t = new TestBed();
			t.f();
		}
	}
} ///:~


This generates a separate class called TestBed$Tester (to run the program, you say java TestBed$Tester). You can use this class for testing, but you don't need to include it in your shipping product; you can simply delete TestBed$Tester.class before packaging things up.

8.2.6. Referring to the outer class object

If you need to produce the reference to the outer class object, you name the outer class followed by a dot and this. For example, in the class Sequence.SSelector, any of its methods can produce the stored reference to the outer class Sequence by saying Sequence.this. The resulting reference is automatically the correct type. (This is known and checked at compile time, so there is no run-time overhead.)

Sometimes you want to tell some other object to create an object of one of its inner classes. To do this you must provide a reference to the other outer class object in the new expression, like this:

 
Sélectionnez

//: c08:Parcel11.java
// Creating instances of inner classes.
 
public class Parcel11 {
	class Contents {
		private int i = 11;
		public int value() { return i; }
	}
	class Destination {
		private String label;
		Destination(String whereTo) { label = whereTo; }
		String readLabel() { return label; }
	}
	public static void main(String[] args) {
		Parcel11 p = new Parcel11();
		// Must use instance of outer class
		// to create an instances of the inner class:
		Parcel11.Contents c = p.new Contents();
		Parcel11.Destination d = p.new Destination("Tanzania");
	}
} ///:~


To create an object of the inner class directly, you don't follow the same form and refer to the outer class name Parcel11 as you might expect, but instead you must use an object of the outer class to make an object of the inner class:

 
Sélectionnez

Parcel11.Contents c = p.new Contents();


Thus, it's not possible to create an object of the inner class unless you already have an object of the outer class. This is because the object of the inner class is quietly connected to the object of the outer class that it was made from. However, if you make a nested class (a static inner class), then it doesn't need a reference to the outer class object.

8.2.7. Reaching outward from a multiply-nested class

(37) It doesn't matter how deeply an inner class may be nested-it can transparently access all of the members of all the classes it is nested within, as seen here:

 
Sélectionnez

//: c08:MultiNestingAccess.java
// Nested classes can access all members of all
// levels of the classes they are nested within.
 
class MNA {
	private void f() {}
	class A {
		private void g() {}
		public class B {
			void h() {
				g();
				f();
			}
		}
	}
}
 
public class MultiNestingAccess {
	public static void main(String[] args) {
		MNA mna = new MNA();
		MNA.A mnaa = mna.new A();
		MNA.A.B mnaab = mnaa.new B();
		mnaab.h();
	}
} ///:~


You can see that in MNA.A.B, the methods g() and f() are callable without any qualification (despite the fact that they are private). This example also demonstrates the syntax necessary to create objects of multiply-nested inner classes when you create the objects in a different class. The ".new" syntax produces the correct scope, so you do not have to qualify the class name in the constructor call.

8.2.8. Inheriting from inner classes

Because the inner class constructor must attach to a reference of the enclosing class object, things are slightly complicated when you inherit from an inner class. The problem is that the "secret" reference to the enclosing class object must be initialized, and yet in the derived class there's no longer a default object to attach to. The answer is to use a syntax provided to make the association explicit:

 
Sélectionnez

//: c08:InheritInner.java
// Inheriting an inner class.
 
class WithInner {
	class Inner {}
}
 
public class InheritInner extends WithInner.Inner {
	//! InheritInner() {} // Won't compile
	InheritInner(WithInner wi) {
		wi.super();
	}
	public static void main(String[] args) {
		WithInner wi = new WithInner();
		InheritInner ii = new InheritInner(wi);
	}
} ///:~


You can see that InheritInner is extending only the inner class, not the outer one. But when it comes time to create a constructor, the default one is no good, and you can't just pass a reference to an enclosing object. In addition, you must use the syntax

 
Sélectionnez

enclosingClassReference.super();


inside the constructor. This provides the necessary reference, and the program will then compile.

8.2.9. Can inner classes be overridden?

What happens when you create an inner class, then inherit from the enclosing class and redefine the inner class? That is, is it possible to override the entire inner class? This seems like it would be a powerful concept, but "overriding" an inner class as if it were another method of the outer class doesn't really do anything:

 
Sélectionnez

//: c08:BigEgg.java
// An inner class cannot be overriden like a method.
import com.bruceeckel.simpletest.*;
 
class Egg {
	private Yolk y;
	protected class Yolk {
		public Yolk() { System.out.println("Egg.Yolk()"); }
	}
	public Egg() {
		System.out.println("New Egg()");
		y = new Yolk();
	}
}
 
public class BigEgg extends Egg {
	private static Test monitor = new Test();
	public class Yolk {
		public Yolk() { System.out.println("BigEgg.Yolk()"); }
	}
	public static void main(String[] args) {
		new BigEgg();
		monitor.expect(new String[] {
			"New Egg()",
			"Egg.Yolk()"
		});
	}
} ///:~


The default constructor is synthesized automatically by the compiler, and this calls the base-class default constructor. You might think that since a BigEgg is being created, the "overridden" version of Yolk would be used, but this is not the case, as you can see from the output.

This example shows that there isn't any extra inner class magic going on when you inherit from the outer class. The two inner classes are completely separate entities, each in their own namespace. However, it's still possible to explicitly inherit from the inner class:

 
Sélectionnez

//: c08:BigEgg2.java
// Proper inheritance of an inner class.
import com.bruceeckel.simpletest.*;
 
class Egg2 {
	protected class Yolk {
		public Yolk() { System.out.println("Egg2.Yolk()"); }
		public void f() { System.out.println("Egg2.Yolk.f()");}
	}
	private Yolk y = new Yolk();
	public Egg2() { System.out.println("New Egg2()"); }
	public void insertYolk(Yolk yy) { y = yy; }
	public void g() { y.f(); }
}
 
public class BigEgg2 extends Egg2 {
	private static Test monitor = new Test();
	public class Yolk extends Egg2.Yolk {
		public Yolk() { System.out.println("BigEgg2.Yolk()"); }
		public void f() {
			System.out.println("BigEgg2.Yolk.f()");
		}
	}
	public BigEgg2() { insertYolk(new Yolk()); }
	public static void main(String[] args) {
		Egg2 e2 = new BigEgg2();
		e2.g();
		monitor.expect(new String[] {
			"Egg2.Yolk()",
			"New Egg2()",
			"Egg2.Yolk()",
			"BigEgg2.Yolk()",
			"BigEgg2.Yolk.f()"
		});
	}
} ///:~


Now BigEgg2.Yolk explicitly extends Egg2.Yolk and overrides its methods. The method insertYolk() allows BigEgg2 to upcast one of its own Yolk objects into the y reference in Egg2, so when g() calls y.f(), the overridden version of f() is used. The second call to Egg2.Yolk() is the base-class constructor call of the BigEgg2.Yolk constructor. You can see that the overridden version of f() is used when g() is called.

8.2.10. Local inner classes

As noted earlier, inner classes can also be created inside code blocks, typically inside the body of a method. A local inner class cannot have an access specifier because it isn't part of the outer class, but it does have access to the final variables in the current code block and all the members of the enclosing class. Here's an example comparing the creation of a local inner class with an anonymous inner class:

 
Sélectionnez

//: c08:LocalInnerClass.java
// Holds a sequence of Objects.
import com.bruceeckel.simpletest.*;
 
interface Counter {
	int next();
}
 
public class LocalInnerClass {
	private static Test monitor = new Test();
	private int count = 0;
	Counter getCounter(final String name) {
		// A local inner class:
		class LocalCounter implements Counter {
			public LocalCounter() {
				// Local inner class can have a constructor
				System.out.println("LocalCounter()");
			}
			public int next() {
				System.out.print(name); // Access local final
				return count++;
			}
		}
		return new LocalCounter();
	}
	// The same thing with an anonymous inner class:
	Counter getCounter2(final String name) {
		return new Counter() {
			// Anonymous inner class cannot have a named
			// constructor, only an instance initializer:
			{
				System.out.println("Counter()");
			}
			public int next() {
				System.out.print(name); // Access local final
				return count++;
			}
		};
	}
	public static void main(String[] args) {
		LocalInnerClass lic = new LocalInnerClass();
		Counter
			c1 = lic.getCounter("Local inner "),
			c2 = lic.getCounter2("Anonymous inner ");
		for(int i = 0; i < 5; i++)
			System.out.println(c1.next());
		for(int i = 0; i < 5; i++)
			System.out.println(c2.next());
		monitor.expect(new String[] {
			"LocalCounter()",
			"Counter()",
			"Local inner 0",
			"Local inner 1",
			"Local inner 2",
			"Local inner 3",
			"Local inner 4",
			"Anonymous inner 5",
			"Anonymous inner 6",
			"Anonymous inner 7",
			"Anonymous inner 8",
			"Anonymous inner 9"
		});
	}
} ///:~


Counter returns the next value in a sequence. It is implemented as both a local class and an anonymous inner class, both of which have the same behaviors and capabilities. Since the name of the local inner class is not accessible outside the method, the only justification for using a local inner class instead of an anonymous inner class is if you need a named constructor and/or overloaded constructor, since an anonymous inner class can only use instance initialization.

The only reason to make a local inner class rather than an anonymous inner class is if you need to make more than one object of that class.

8.2.11. Inner class identifiers

Since every class produces a .class file that holds all the information about how to create objects of this type (this information produces a "meta-class" called the Class object), you might guess that inner classes must also produce .class files to contain the information for their Class objects. The names of these files/classes have a strict formula: the name of the enclosing class, followed by a '$', followed by the name of the inner class. For example, the .class files created by LocalInnerClass.java include:

 
Sélectionnez

Counter.class
LocalInnerClass$2.class
LocalInnerClass$1LocalCounter.class
LocalInnerClass.class


If inner classes are anonymous, the compiler simply starts generating numbers as inner class identifiers. If inner classes are nested within inner classes, their names are simply appended after a '$' and the outer class identifier(s).

Although this scheme of generating internal names is simple and straightforward, it's also robust and handles most situations. (38) Since it is the standard naming scheme for Java, the generated files are automatically platform-independent. (Note that the Java compiler is changing your inner classes in all sorts of other ways in order to make them work.)

8.3. Why inner classes?

At this point you've seen a lot of syntax and semantics describing the way inner classes work, but this doesn't answer the question of why they exist. Why did Sun go to so much trouble to add this fundamental language feature?

Typically, the inner class inherits from a class or implements an interface, and the code in the inner class manipulates the outer class object that it was created within. So you could say that an inner class provides a kind of window into the outer class.

A question that cuts to the heart of inner classes is this: If I just need a reference to an interface, why don't I just make the outer class implement that interface? The answer is "If that's all you need, then that's how you should do it." So what is it that distinguishes an inner class implementing an interface from an outer class implementing the same interface? The answer is that you can't always have the convenience of interfaces-sometimes you're working with implementations. So the most compelling reason for inner classes is:

Each inner class can independently inherit from an implementation. Thus, the inner class is not limited by whether the outer class is already inheriting from an implementation.

Without the ability that inner classes provide to inherit-in effect-from more than one concrete or abstract class, some design and programming problems would be intractable. So one way to look at the inner class is as the rest of the solution of the multiple-inheritance problem. Interfaces solve part of the problem, but inner classes effectively allow "multiple implementation inheritance." That is, inner classes effectively allow you to inherit from more than one non-interface.

To see this in more detail, consider a situation in which you have two interfaces that must somehow be implemented within a class. Because of the flexibility of interfaces, you have two choices: a single class or an inner class:

 
Sélectionnez

//: c08:MultiInterfaces.java
// Two ways that a class can implement multiple interfaces.
 
interface A {}
interface B {}
 
class X implements A, B {}
 
class Y implements A {
	B makeB() {
		// Anonymous inner class:
		return new B() {};
	}
}
 
public class MultiInterfaces {
	static void takesA(A a) {}
	static void takesB(B b) {}
	public static void main(String[] args) {
		X x = new X();
		Y y = new Y();
		takesA(x);
		takesA(y);
		takesB(x);
		takesB(y.makeB());
	}
} ///:~


Of course, this assumes that the structure of your code makes logical sense either way. However, you'll ordinarily have some kind of guidance from the nature of the problem about whether to use a single class or an inner class. But without any other constraints, the approach in the preceding example doesn't really make much difference from an implementation standpoint. Both of them work.

However, if you have abstract or concrete classes instead of interfaces, you are suddenly limited to using inner classes if your class must somehow implement both of the others:

 
Sélectionnez

//: c08:MultiImplementation.java
// With concrete or abstract classes, inner
// classes are the only way to produce the effect
// of "multiple implementation inheritance."
package c08;
 
class D {}
abstract class E {}
 
class Z extends D {
	E makeE() { return new E() {}; }
}
 
public class MultiImplementation {
	static void takesD(D d) {}
	static void takesE(E e) {}
	public static void main(String[] args) {
		Z z = new Z();
		takesD(z);
		takesE(z.makeE());
	}
} ///:~


If you didn't need to solve the "multiple implementation inheritance" problem, you could conceivably code around everything else without the need for inner classes. But with inner classes you have these additional features:

  • The inner class can have multiple instances, each with its own state information that is independent of the information in the outer class object.
  • In a single outer class you can have several inner classes, each of which implement the same interface or inherit from the same class in a different way. An example of this will be shown shortly.
  • The point of creation of the inner class object is not tied to the creation of the outer class object.
  • There is no potentially confusing "is-a" relationship with the inner class; it's a separate entity.

As an example, if Sequence.java did not use inner classes, you'd have to say "a Sequence is a Selector," and you'd only be able to have one Selector in existence for a particular Sequence. You can easily have a second method, getRSelector(), that produces a Selector that moves backward through the sequence. This kind of flexibility is only available with inner classes.

8.3.1. Closures & Callbacks

A closure is a callable object that retains information from the scope in which it was created. From this definition, you can see that an inner class is an object-oriented closure, because it doesn't just contain each piece of information from the outer class object ("the scope in which it was created"), but it automatically holds a reference back to the whole outer class object, where it has permission to manipulate all the members, even private ones.

One of the most compelling arguments made to include some kind of pointer mechanism in Java was to allow callbacks. With a callback, some other object is given a piece of information that allows it to call back into the originating object at some later point. This is a very powerful concept, as you will see later in the book. If a callback is implemented using a pointer, however, you must rely on the programmer to behave and not misuse the pointer. As you've seen by now, Java tends to be more careful than that, so pointers were not included in the language.

The closure provided by the inner class is a perfect solution-more flexible and far safer than a pointer. Here's an example:

 
Sélectionnez

//: c08:Callbacks.java
// Using inner classes for callbacks
import com.bruceeckel.simpletest.*;
 
interface Incrementable {
	void increment();
}
 
// Very simple to just implement the interface:
class Callee1 implements Incrementable {
	private int i = 0;
	public void increment() {
		i++;
		System.out.println(i);
	}
}
 
class MyIncrement {
	void increment() {
		System.out.println("Other operation");
	}
	static void f(MyIncrement mi) { mi.increment(); }
}
 
// If your class must implement increment() in
// some other way, you must use an inner class:
class Callee2 extends MyIncrement {
	private int i = 0;
	private void incr() {
		i++;
		System.out.println(i);
	}
	private class Closure implements Incrementable {
		public void increment() { incr(); }
	}
	Incrementable getCallbackReference() {
		return new Closure();
	}
}
 
class Caller {
	private Incrementable callbackReference;
	Caller(Incrementable cbh) { callbackReference = cbh; }
	void go() { callbackReference.increment(); }
}
 
public class Callbacks {
	private static Test monitor = new Test();
	public static void main(String[] args) {
		Callee1 c1 = new Callee1();
		Callee2 c2 = new Callee2();
		MyIncrement.f(c2);
		Caller caller1 = new Caller(c1);
		Caller caller2 = new Caller(c2.getCallbackReference());
		caller1.go();
		caller1.go();
		caller2.go();
		caller2.go();
		monitor.expect(new String[] {
			"Other operation",
			"1",
			"2",
			"1",
			"2"
		});
	}
} ///:~


This example also provides a further distinction between implementing an interface in an outer class versus doing so in an inner class. Callee1 is clearly the simpler solution in terms of the code. Callee2 inherits from MyIncrement, which already has a different increment() method that does something unrelated to the one expected by the Incrementable interface. When MyIncrement is inherited into Callee2, increment() can't be overridden for use by Incrementable, so you're forced to provide a separate implementation using an inner class. Also note that when you create an inner class, you do not add to or modify the interface of the outer class.

Notice that everything except getCallbackReference() in Callee2 is private. To allow any connection to the outside world, the interface Incrementable is essential. Here you can see how interfaces allow for a complete separation of interface from implementation.

The inner class Closure implements Incrementable to provide a hook back into Callee2-but a safe hook. Whoever gets the Incrementable reference can, of course, only call increment() and has no other abilities (unlike a pointer, which would allow you to run wild).

Caller takes an Incrementable reference in its constructor (although the capturing of the callback reference could happen at any time) and then, sometime later, uses the reference to "call back" into the Callee class.

The value of the callback is in its flexibility; you can dynamically decide what methods will be called at run time. The benefit of this will become more evident in Chapter 14, where callbacks are used everywhere to implement GUI functionality.

8.3.2. Inner classes & control frameworks

A more concrete example of the use of inner classes can be found in something that I will refer to here as a control framework.

An application framework is a class or a set of classes that's designed to solve a particular type of problem. To apply an application framework, you typically inherit from one or more classes and override some of the methods. The code that you write in the overridden methods customizes the general solution provided by that application framework in order to solve your specific problem (this is an example of the Template Method design pattern; see Thinking in Patterns (with Java) at www.BruceEckel.com). The control framework is a particular type of applicationframework dominated by the need to respond to events; a system that primarily responds to events is called an event-driven system. One of the most important problems in application programming is the graphical user interface (GUI), which is almost entirely event-driven. As you will see in Chapter 14, the Java Swing library is a control framework that elegantly solves the GUI problem and that heavily uses inner classes.

To see how inner classes allow the simple creation and use of control frameworks, consider a control framework whose job is to execute events whenever those events are "ready." Although "ready" could mean anything, in this case the default will be based on clock time. What follows is a control framework that contains no specific information about what it's controlling. That information is supplied during inheritance, when the "template method" is implemented.

First, here is the interface that describes any control event. It's an abstract class instead of an actual interface because the default behavior is to perform the control based on time. Thus, some of the implementation is included here:

 
Sélectionnez

//: c08:controller:Event.java
// The common methods for any control event.
package c08.controller;
 
public abstract class Event {
	private long eventTime;
	protected final long delayTime;
	public Event(long delayTime) {
		this.delayTime = delayTime;
		start();
	}
	public void start() { // Allows restarting
		eventTime = System.currentTimeMillis() + delayTime;
	}
	public boolean ready() {
		return System.currentTimeMillis() >= eventTime;
	}
	public abstract void action();
} ///:~


The constructor captures the time (from the time of creation of the object) when you want the Event to run and then calls start(), which takes the current time and adds the delay time to produce the time when the event will occur. Rather than being included in the constructor, start() is a separate method because this way, it allows you to restart the timer after the event has run out so the Event object can be reused. For example, if you want a repeating event, you can simply call start() inside your action() method.

ready() tells you when it's time to run the action() method. Of course, ready() could be overridden in a derived class to base the Event on something other than time.

The following file contains the actual control framework that manages and fires events. The Event objects are held inside a container object of type ArrayList, which you'll learn more about in Chapter 11. For now, all you need to know is that add() will append an Object to the end of the ArrayList, size() produces the number of entries in the ArrayList, get() will fetch an element from the ArrayList at a particular index, and remove() removes an element from the ArrayList, given the element number you want to remove.

 
Sélectionnez

//: c08:controller:Controller.java
// With Event, the generic framework for control systems.
package c08.controller;
import java.util.*;
 
public class Controller {
	// An object from java.util to hold Event objects:
	private List eventList = new ArrayList();
	public void addEvent(Event c) { eventList.add(c); }
	public void run() {
		while(eventList.size() > 0) {
			for(int i = 0; i < eventList.size(); i++) {
				Event e = (Event)eventList.get(i);
				if(e.ready()) {
					System.out.println(e);
					e.action();
					eventList.remove(i);
				}
			}
		}
	}
} ///:~


The run() method loops through eventList, hunting for an Event object that's ready() to run. For each one it finds ready(),it prints information using the object's toString() method, calls the action() method, and then removes the Event from the list.

Note that so far in this design you know nothing about exactly what an Event does. And this is the crux of the design-how it "separates the things that change from the things that stay the same." Or, to use my term, the "vector of change" is the different actions of the various kinds of Event objects, and you express different actions by creating different Event subclasses.

This is where inner classes come into play. They allow two things:

  • To create the entire implementation of a control framework in a single class, thereby encapsulating everything that's unique about that implementation. Inner classes are used to express the many different kinds of action() necessary to solve the problem.
  • Inner classes keep this implementation from becoming awkward, since you're able to easily access any of the members in the outer class. Without this ability the code might become unpleasant enough that you'd end up seeking an alternative.

Consider a particular implementation of the control framework designed to control greenhouse functions. (39) Each action is entirely different: turning lights, water, and thermostats on and off, ringing bells, and restarting the system. But the control framework is designed to easily isolate this different code. Inner classes allow you to have multiple derived versions of the same base class, Event, within a single class. For each type of action, you inherit a new Event inner class, and write the control code in the action() implementation.

As is typical with an application framework, the class GreenhouseControls is inherited from Controller:

 
Sélectionnez

//: c08:GreenhouseControls.java
// This produces a specific application of the
// control system, all in a single class. Inner
// classes allow you to encapsulate different
// functionality for each type of event.
import com.bruceeckel.simpletest.*;
import c08.controller.*;
 
public class GreenhouseControls extends Controller {
	private static Test monitor = new Test();
	private boolean light = false;
	public class LightOn extends Event {
		public LightOn(long delayTime) { super(delayTime); }
		public void action() {
			// Put hardware control code here to
			// physically turn on the light.
			light = true;
		}
		public String toString() { return "Light is on"; }
	}
	public class LightOff extends Event {
		public LightOff(long delayTime) { super(delayTime); }
		public void action() {
			// Put hardware control code here to
			// physically turn off the light.
			light = false;
		}
		public String toString() { return "Light is off"; }
	}
	private boolean water = false;
	public class WaterOn extends Event {
		public WaterOn(long delayTime) { super(delayTime); }
		public void action() {
			// Put hardware control code here.
			water = true;
		}
		public String toString() {
			return "Greenhouse water is on";
		}
	}
	public class WaterOff extends Event {
		public WaterOff(long delayTime) { super(delayTime); }
		public void action() {
			// Put hardware control code here.
			water = false;
		}
		public String toString() {
			return "Greenhouse water is off";
		}
	}
	private String thermostat = "Day";
	public class ThermostatNight extends Event {
		public ThermostatNight(long delayTime) {
			super(delayTime);
		}
		public void action() {
			// Put hardware control code here.
			thermostat = "Night";
		}
		public String toString() {
			return "Thermostat on night setting";
		}
	}
	public class ThermostatDay extends Event {
		public ThermostatDay(long delayTime) {
			super(delayTime);
		}
		public void action() {
			// Put hardware control code here.
			thermostat = "Day";
		}
		public String toString() {
			return "Thermostat on day setting";
		}
	}
	// An example of an action() that inserts a
	// new one of itself into the event list:
	public class Bell extends Event {
		public Bell(long delayTime) { super(delayTime); }
		public void action() {
			addEvent(new Bell(delayTime));
		}
		public String toString() { return "Bing!"; }
	}
	public class Restart extends Event {
		private Event[] eventList;
		public Restart(long delayTime, Event[] eventList) {
			super(delayTime);
			this.eventList = eventList;
			for(int i = 0; i < eventList.length; i++)
				addEvent(eventList[i]);
		}
		public void action() {
			for(int i = 0; i < eventList.length; i++) {
				eventList[i].start(); // Rerun each event
				addEvent(eventList[i]);
			}
			start(); // Rerun this Event
			addEvent(this);
		}
		public String toString() {
			return "Restarting system";
		}
	}
	public class Terminate extends Event {
		public Terminate(long delayTime) { super(delayTime); }
		public void action() { System.exit(0); }
		public String toString() { return "Terminating";  }
	}
} ///:~


Note that light, water, and thermostat belong to the outer class GreenhouseControls, and yet the inner classes can access those fields without qualification or special permission. Also, most of the action() methods involve some sort of hardware control.

Most of the Event classes look similar, but Bell and Restart are special. Bell rings and then adds a new Bell object to the event list, so it will ring again later. Notice how inner classes almost look like multiple inheritance: Bell and Restart have all the methods of Event and also appear to have all the methods of the outer class GreenhouseControls.

Restart is given an array of Event objects that it adds to the controller. Since Restart() is just another Event object, you can also add a Restart object within Restart.action() so that the system regularly restarts itself.

The following class configures the system by creating a GreenhouseControls object and adding various kinds of Event objects. This is an example of the Command design pattern:

 
Sélectionnez

//: c08:GreenhouseController.java
// Configure and execute the greenhouse system.
// {Args: 5000}
import c08.controller.*;
 
public class GreenhouseController {
	public static void main(String[] args) {
		GreenhouseControls gc = new GreenhouseControls();
		// Instead of hard-wiring, you could parse
		// configuration information from a text file here:
		gc.addEvent(gc.new Bell(900));
		Event[] eventList = {
			gc.new ThermostatNight(0),
			gc.new LightOn(200),
			gc.new LightOff(400),
			gc.new WaterOn(600),
			gc.new WaterOff(800),
			gc.new ThermostatDay(1400)
		};
		gc.addEvent(gc.new Restart(2000, eventList));
		if(args.length == 1)
			gc.addEvent(
		gc.new Terminate(Integer.parseInt(args[0])));
		gc.run();
	}
} ///:~


This class initializes the system, so it adds all the appropriate events. Of course, a more flexible way to accomplish this is to avoid hard-coding the events and instead read them from a file. (An exercise in Chapter 12 asks you to modify this example to do just that.) If you provide a command-line argument, it uses this to terminate the program after that many milliseconds (this is used for testing).

This example should move you toward an appreciation of the value of inner classes, especially when used within a control framework. However, in Chapter 14 you'll see how elegantly inner classes are used to describe the actions of a graphical user interface. By the time you finish that chapter, you should be fully convinced.

8.4. Summary

Interfaces and inner classes are more sophisticated concepts than what you'll find in many OOP languages; for example, there's nothing like them in C++. Together, they solve the same problem that C++ attempts to solve with its multiple inheritance (MI) feature. However, MI in C++ turns out to be rather difficult to use, whereas Java interfaces and inner classes are, by comparison, much more accessible.

Although the features themselves are reasonably straightforward, the use of these features is a design issue, much the same as polymorphism. Over time, you'll become better at recognizing situations where you should use an interface, or an inner class, or both. But at this point in this book, you should at least be comfortable with the syntax and semantics. As you see these language features in use, you'll eventually internalize them.

8.5. Exercises

Solutions to selected exercises can be found in the electronic document The Thinking in Java Annotated Solution Guide, available for a small fee from www.BruceEckel.com.

  1. Prove that the fields in an interface are implicitly static and final.
  2. Create an interface containing three methods, in its own package. Implement the interface in a different package.
  3. Prove that all the methods in an interface are automatically public.
  4. In c07:Sandwich.java, create an interface called FastFood (with appropriate methods) and change Sandwich so that it also implements FastFood.
  5. Create three interfaces, each with two methods. Inherit a new interface from the three, adding a new method. Create a class by implementing the new interface and also inheriting from a concrete class. Now write four methods, each of which takes one of the four interfaces as an argument. In main(), create an object of your class and pass it to each of the methods.
  6. Modify Exercise 5 by creating an abstract class and inheriting that into the derived class.
  7. Modify Music5.java by adding a Playable interface. Move the play() declaration from Instrument to Playable. Add Playable to the derived classes by including it in the implements list. Change tune() so that it takes a Playable instead of an Instrument.
  8. Change Exercise 6 in Chapter 7 so that Rodent is an interface.
  9. In Adventure.java, add an interface called CanClimb, following the form of the other interfaces.
  10. Write a program that imports and uses Month.java.
  11. Following the example given in Month.java, create an enumeration of days of the week.
  12. Create an interface with at least one method, in its own package. Create a class in a separate package. Add a protected inner class that implements the interface. In a third package, inherit from your class and, inside a method, return an object of the protected inner class, upcasting to the interface during the return.
  13. Create an interface with at least one method, and implement that interface by defining an inner class within a method, which returns a reference to your interface.
  14. Repeat Exercise 13 but define the inner class within a scope within a method.
  15. Repeat Exercise 13 using an anonymous inner class.
  16. Modify HorrorShow.java to implement DangerousMonster and Vampire using anonymous classes.
  17. Create a private inner class that implements a publicinterface. Write a method that returns a reference to an instance of the private inner class, upcast to the interface. Show that the inner class is completely hidden by trying to downcast to it.
  18. Create a class with a nondefault constructor (one with arguments) and no default constructor (no "no-arg" constructor). Create a second class that has a method that returns a reference to the first class. Create the object to return by making an anonymous inner class that inherits from the first class.
  19. Create a class with a private field and a private method. Create an inner class with a method that modifies the outer class field and calls the outer class method. In a second outer class method, create an object of the inner class and call its method, then show the effect on the outer class object.
  20. Repeat Exercise 19 using an anonymous inner class.
  21. Create a class containing a nested class. In main(), create an instance of the inner class.
  22. Create an interface containing a nested class. Implement this interface and create an instance of the nested class.
  23. Create a class containing an inner class that itself contains an inner class. Repeat this using nested classes. Note the names of the .class files produced by the compiler.
  24. Create a class with an inner class. In a separate class, make an instance of the inner class.
  25. Create a class with an inner class that has a nondefault constructor (one that takes arguments). Create a second class with an inner class that inherits from the first inner class.
  26. Repair the problem in WindError.java.
  27. Modify Sequence.java by adding a method getRSelector() that produces a different implementation of the Selector interface that moves backward through the sequence from the end to the beginning.
  28. Create an interface U with three methods. Create a class A with a method that produces a reference to a U by building an anonymous inner class. Create a second class B that contains an array of U. B should have one method that accepts and stores a reference to a U in the array, a second method that sets a reference in the array (specified by the method argument) to null, and a third method that moves through the array and calls the methods in U. In main(), create a group of A objects and a single B. Fill the B with U references produced by the A objects. Use the B to call back into all the A objects. Remove some of the U references from the B.
  29. In GreenhouseControls.java, add Event inner classes that turn fans on and off. Configure GreenhouseController.java to use these new Event objects.
  30. Inherit from GreenhouseControls in GreenhouseControls.java to add Event inner classes that turn water mist generators on and off. Write a new version of GreenhouseController.java to use these new Event objects.
  31. Show that an inner class has access to the private elements of its outer class. Determine whether the reverse is true.



prcdentsommairesuivant
This approach was inspired by an e-mail from Rich Hoffarth. Item 21 in Joshua Bloch's Effective Java (Addison-Wesley, 2001) covers the topic in much more detail.
Thanks to Martin Danner for asking this question during a seminar.
This is very different from the design of nested classes in C++, which is simply a name-hiding mechanism. There is no link to an enclosing object and no implied permissions in C++.
Roughly similar to nested classes in C++, except that those classes cannot access private members as they can in Java.
Thanks again to Martin Danner.
On the other hand, '$' is a meta-character to the Unix shell and so you'll sometimes have trouble when listing the .class files. This is a bit strange coming from Sun, a Unix-based company. My guess is that they weren't considering this issue, but instead thought you'd naturally focus on the source-code files.
For some reason this has always been a pleasing problem for me to solve; it came from my earlier book C++ Inside & Out, but Java allows a much more elegant solution.

Vous avez aim ce tutoriel? Alors partagez-le en cliquant sur les boutons suivants: Viadeo Twitter Facebook Share on Google+   

Ce document est issu de http://www.developpez.com et reste la proprit exclusive de son auteur. La copie, modification et/ou distribution par quelque moyen que ce soit est soumise l'obtention pralable de l'autorisation de l'auteur.