1FPCLASSIFY(3)              Linux Programmer's Manual             FPCLASSIFY(3)
2
3
4

NAME

6       fpclassify, isfinite, isnormal, isnan, isinf - floating-point classifi‐
7       cation macros
8

SYNOPSIS

10       #include <math.h>
11
12       int fpclassify(x);
13
14       int isfinite(x);
15
16       int isnormal(x);
17
18       int isnan(x);
19
20       int isinf(x);
21
22       Link with -lm.
23
24   Feature Test Macro Requirements for glibc (see feature_test_macros(7)):
25
26       fpclassify(), isfinite(), isnormal(): _XOPEN_SOURCE >= 600 ||
27       _ISOC99_SOURCE; or cc -std=c99
28       isnan(): _BSD_SOURCE || _SVID_SOURCE || _XOPEN_SOURCE ||
29       _ISOC99_SOURCE; or cc -std=c99
30       isinf(): _BSD_SOURCE || _SVID_SOURCE || _XOPEN_SOURCE >= 600 ||
31       _ISOC99_SOURCE; or cc -std=c99
32

DESCRIPTION

34       Floating  point  numbers  can  have special values, such as infinite or
35       NaN.  With the macro fpclassify(x) you can find out  what  type  x  is.
36       The  macro takes any floating-point expression as argument.  The result
37       is one of the following values:
38
39       FP_NAN        x is "Not a Number".
40
41       FP_INFINITE   x is either positive infinity or negative infinity.
42
43       FP_ZERO       x is zero.
44
45       FP_SUBNORMAL  x is too small to be represented in normalized format.
46
47       FP_NORMAL     if nothing of the above is correct then it must be a nor‐
48                     mal floating-point number.
49
50       The other macros provide a short answer to some standard questions.
51
52       isfinite(x)   returns a non-zero value if
53                     (fpclassify(x) != FP_NAN && fpclassify(x) != FP_INFINITE)
54
55       isnormal(x)   returns a non-zero value if (fpclassify(x) == FP_NORMAL)
56
57       isnan(x)      returns a non-zero value if (fpclassify(x) == FP_NAN)
58
59       isinf(x)      returns 1 if x is positive infinity, and -1 if x is nega‐
60                     tive infinity.
61

CONFORMING TO

63       C99, POSIX.1.
64
65       For isinf(), the standards merely say that the return value is non-zero
66       if and only if the argument has an infinite value.
67

NOTES

69       In  glibc 2.01 and earlier, isinf() returns a non-zero value (actually:
70       1) if x is positive infinity or negative infinity.  (This is  all  that
71       C99 requires.)
72

SEE ALSO

74       finite(3), INFINITY(3), isgreater(3), signbit(3)
75

COLOPHON

77       This  page  is  part of release 3.22 of the Linux man-pages project.  A
78       description of the project, and information about reporting  bugs,  can
79       be found at http://www.kernel.org/doc/man-pages/.
80
81
82
83                                  2008-08-07                     FPCLASSIFY(3)