Inside Macintosh: PowerPC Numerics / Part 2 - The PowerPC Numerics C Implementation

Chapter 7 - Numeric Data Types in C

Inquiries: Class and Sign

MathLib provides macros you can use to determine the class and sign of a floating-point value. All of these macros return type long int. They are listed in Table 7-3.

Table 7-3 Class and sign inquiry macros

Macro	Value returned	Condition
fpclassify(x)	FP_SNAN	x is a signaling NaN
	FP_QNAN	x is a quiet NaN
	FP_INFINITE	x is $-$ or +
	FP_ZERO	x is +0 or $-0$
	FP_NORMAL	x is a normalized number
	FP_SUBNORMAL	x is a denormalized (subnormal) number
isnormal(x)	TRUE	x is a normalized number
isfinite(x)	TRUE	x is not $-$ , + , or NaN
isnan(x)	TRUE	x is a NaN (quiet or signaling)
signbit(x)	1	The sign bit of x is 1 (x is negative)
	0	The sign bit of x is 0 (x is positive)