Constant expressions

From cppreference.com
< c‎ | language

Several varieties of expressions are known as constant expressions

Preprocessor constant expression

The expression following #if or #elif must expand to

Character constants, when evaluated in #if-expressions, may be interpreted in the source character set, the execution character set, or some other implementation-defined character set.

Integer arithmetic in #if-expressions is performed using the semantics of intmax_t for signed types and uintmax_t for unsigned types.

(since C99)

Integer constant expression

An integer constant expression is an expression that consists only of

(since C11)

Integer constant expressions are evaluated at compile time. The following contexts require expressions that are known as integer constant expressions:

(since C99)
(since C11)

Static initializer

Expressions that are used in the initializers of objects with static and thread_local storage duration must be expressions that may be one of the following

1) arithmetic constant expression, which is an expression of any arithmetic type that consists of
(since C11)
2) the null pointer constant NULL
3) address constant expression, which is
  • a null pointer
  • lvalue designating an object of static storage duration or a function designator, converted to a pointer either
  • by using the unary address-of operator
  • by casting an integer constant to a pointer
  • by array-to-pointer or function-to-pointer implicit conversion
4) address constant expression of some complete object type, plus or minus an integer constant expression
5) constant expression of one of the other forms accepted by the implementation.

Unlike with integer constant expressions, static initializer expressions are not required to be evaluated at compile time; the compiler is at liberty to turn such initializers into executable code which is invoked prior to program startup.

static int i = 2 || 1 / 0; // initializes i to value 1

The value of a floating-point static initializer is never less accurate than the value of the same expression executed at run time, but it may be better.

Floating-point constant expressions

Arithmetic constant expressions of floating-point types that are not used in static initializers are always evaluated as-if during run-time and are affected by the current rounding (if FENV_ACCESS is on) and report errors as specified in math_errhandling.

void f(void)
{
#pragma STDC FENV_ACCESS ON
    static float x = 0.0/0.0; // static initializer: does not raise an exception
    float w[] = { 0.0/0.0 }; // raises an exception
    float y = 0.0/0.0; // raises an exception
    double z = 0.0/0.0; // raises an exception
}

Notes

If an expression evaluates to a value that is not representable by its type, it cannot be used as a constant expression.

Implementations may accept other forms of constant expressions. However, these constant expressions are not considered as integer constant expressions, arithmetic constant expressions, or address constant expressions, and thus cannot be used in the contexts requiring these kinds of constant expressions. For example, int arr[(int)+1.0]; declares a VLA.

References

  • C17 standard (ISO/IEC 9899:2018):
  • 6.6 Constant expressions (p: 76-77)
  • C11 standard (ISO/IEC 9899:2011):
  • 6.6 Constant expressions (p: 106-107)
  • C99 standard (ISO/IEC 9899:1999):
  • 6.6 Constant expressions (p: 95-96)
  • C89/C90 standard (ISO/IEC 9899:1990):
  • 3.4 CONSTANT EXPRESSIONS

See also