[Chapter 4] 4.13 Assignment Operators

4.13 Assignment Operators

Assignment operators set the values of variables and array elements. An assignment operator may appear in an assignment expression:

[Graphic: Figure from the text]

The actual assignment operator in an assignment expression can be the simple assignment operator = or one of the compound assignment operators shown below. All of the assignment operators are equal in precedence. Assignment operators are evaluated from right to left, so a=b=c is equivalent to a=(b=c).

The left operand of an assignment operator must be an expression that produces a variable or an array element. The left operand of an assignment operator cannot be an expression that evaluates to a pure value, or a compile-time error occurs. So, for example, the left operand cannot be a final variable, since a final variable evaluates to a pure value, not a variable.

The assignment operator itself produces a pure value, not a variable or an array element. The pure value produced by an assignment operator is the value of the variable or array element after it has been set by the assignment operation. The type of this pure value is the type of the variable or array element.

The simple assignment operator = just sets the value of a variable or array element. It does not imply any other computation. The right operand of the simple assignment operator can be of any type, as long as that type is assignment-compatible with the type of the left operand, as described in the next section. If the right operand is not assignment-compatible, a compile-time error occurs.

The compound assignment operators are:

`+=`	`-=`	`*=`
`/=`	`\|=`	`&=`
`^=`	`%=`	`<<=`
`>>=`	`>>>=`

Both of the operands of a compound assignment operator must be of primitive types, or a compile-time error occurs. The one exception is if the left operand of the += operator is of type String; in this case the right operand can be of any type.

A compound assignment operator combines a binary operator with the simple assignment operator =. Thus, to be assignment-compatible, the right operand of a compound assignment operator must be of a type that complies with the rules for the indicated binary operation. Otherwise, a compile-time error occurs. An assignment expression of the form:

e1 op = e2

is approximately equivalent to:

e1 = (type) ((e1) op (e2))

where type is the type of the expression e1. The only difference is that e1 is only evaluated once in the expression that uses the compound assignment operator.

For example, consider the following code fragment:

j = 0;
a[0] = 3;
a[1]=6;
a[j++] += 2;

After this code is executed, j equals 1 and a[0] is 5. Contrast this with the following code:

j = 0;
a[0] = 3;
a[1]=6;
a[j++] = a[j++] + 2;

After this code is executed, j equals 2 and a[0] is 8 because j++ is evaluated twice.

References Array Types; **UNKNOWN XREF**; Conditional Operator; Interface Variables; Local Variables; Order of Operations; Primitive Types; Reference Types; String; Unary Operators; Variables

Assignment Compatibility

Saying that one type of value is assignment-compatible with another type of value means that a value of the first type can be assigned to a variable of the second type. Here are the rules for assignment compatibility in Java:

Every type is assignment-compatible with itself.
The boolean type is not assignment-compatible with any other type.
A value of any integer type can be assigned to a variable of any other integer type if the variable is of a type that allows it to contain the value without any loss of information.
A value of any integer type can be assigned to a variable of any floating-point type, but a value of any floating-point type cannot be assigned to a variable of any integer type.
A float value can be assigned to a double variable, but a double value cannot be assigned to a float variable.
With a type cast, a value of any arithmetic type can be assigned to a variable of any other arithmetic type.
Any reference can be assigned to a variable that is declared of type Object.
A reference to an object can be assigned to a class-type reference variable if the class of the variable is the same class or a superclass of the class of the object.
A reference to an object can be assigned to an interface-type reference variable if the class of the object implements the interface.
A reference to an array can be assigned to an array variable if either of the following conditions is true:
- Both array types contain elements of the same type.
- Both array types contain object references and the type of reference contained in the elements of the array reference can be assigned to the type of reference contained in the elements of the variable.

Here's an example that illustrates the rules about assignment compatibility of arrays:

class Triangle extends Shape {...}
...
int[] i = new int[8];
int j[];
long l[];
short s[];
Triangle[] t;
Shape[] sh;
j = i;    // Okay
s = i;    // Error
l = i;    // Error
sh = t;   // Okay
t = sh;   // Error

Assigning i to j is fine because both variables are declared as references to arrays that contain int values. On the other hand, assigning i to s is an error because the variables are declared as references to arrays that contain different kinds of elements and these elements are not object references. Assigning i to l is an error for the same reason. Assigning t to sh is fine because the variables are declared as references to arrays that contain object references, and sh[0]=t[0] is legal. However, assigning sh to t is an error because t[0]=sh[0] is not legal.

It is not always possible for the compiler to determine if an assignment to an array element is legal; in these cases the assignment compatibility is checked at runtime. This situation can occur when a variable contains a reference to an array whose type of elements is specified by a class or interface name. In this case, it may not be possible to determine the actual type of the array elements until runtime. Consider the following example:

void foo (InputStream a[]) {
    a[0] = new FileInputStream("/dev/null");
}

Figure 4.1 shows the InputStream class and some of its subclasses in the java.io package.

Figure 4.1: InputStream and some of its classes

[Graphic: Figure 4-1]

Any array with elements that contain references to objects of class InputStream or any of its subclasses can be passed to the method foo() shown in the above example. For example:

FileInputStream f[] = new FileInputStream[3];
foo(f);

Since FileInputStream is a subclass of InputStream, the call to foo() does not cause any problems at runtime.

However, the following call to foo() is problematic:

DataInputStream f[] = new DataInputStream[3];
foo(f);

This call causes an ArrayStoreException to be thrown at runtime. Although DataInputStream is a subclass of InputStream, it is not a superclass of FileInputStream, so the array element assignment in foo() is not assignment-compatible.

References Arithmetic Types; Array Types; Boolean Type; Class Types; Interface Types