Expressions

You write expressions to determine values, to alter values stored in objects, and to call functions that perform input and output. In fact, you express all computations in the program by writing expressions. The translator must evaluate some of the expressions you write to determine properties of the program. The translator or the target environment must evaluate other expressions prior to program startup to determine the initial values stored in objects with static duration. The program evaluates the remaining expressions when it executes.

This document describes briefly just those aspect of expressions most relevant to the use of the Standard C library:

An address constant expression specifies a value that has a pointer type and that the translator or target environment can determine prior to program startup.

A constant expression specifies a value that the translator or target environment can determine prior to program startup.

An integer constant expression specifies a value that has an integer type and that the translator can determine at the point in the program where you write the expression. (You cannot write a function call, assigning operator, or comma operator except as part of the operand of a sizeof operator.) In addition, you must write only subexpressions that have integer type. You can, however, write a floating-point constant expression as the operand of an integer type cast operator.

A floating-point constant expression specifies a value that has a floating-point type and that the translator can determine at the point in the program where you write the expression. (You cannot write a function call, assigning operator, or comma operator except as part of the operand of a sizeof operator.) In addition, you must write only subexpressions that have integer or floating-point type.

An lvalue expression An lvalue expression designates an object that has an object type other than an array type. Hence, you can access the value stored in the object. A modifiable lvalue expression designates an object that has an object type other than an array type or a const type. Hence, you can alter the value stored in the object. You can also designate objects with an lvalue expression that has an array type or an incomplete type, but you can only take the address of such an expression.

Promoting occurs for an expression whose integer type is not one of the ``computational'' types. Except when it is the operand of the sizeof operator, an integer rvalue expression has one of four types: int, unsigned int, long, or unsigned long. When you write an expression in an rvalue context and the expression has an integer type that is not one of these types, the translator promotes its type to one of these. If all of the values representable in the original type are also representable as type int, then the promoted type is int. Otherwise, the promoted type is unsigned int. Thus, for signed char, short, and any signed bitfield type, the promoted type is int. For each of the remaining integer types (char, unsigned char, unsigned short, any plain bitfield type, or any unsigned bitfield type), the effect of these rules is to favor promoting to int wherever possible, but to promote to unsigned int if necessary to preserve the original value in all possible cases.

An rvalue expression is an expression whose value can be determined only when the program executes. The term also applies to expressions which need not be determined until program execution.

You use the sizeof operator, as in the expression sizeof X to determine the size in bytes of an object whose type is the type of X. The translator uses the expression you write for X only to determine a type; it is not evaluated.

A void expression has type void.

See also the Table of Contents and the Index.

Copyright © 1989-2002 by P.J. Plauger and Jim Brodie. All rights reserved.