Debian Codegraph @

upstream/debian-unstable/man3/fpclassify.3

raw
1 '\" t
2 .\" Copyright 2002, Walter Harms <[email protected]>
3 .\" Copyright, the authors of the Linux man-pages project
4 .\"
5 .\" SPDX-License-Identifier: GPL-1.0-or-later
6 .\"
7 .TH fpclassify 3 2026-02-08 "Linux man-pages 6.18"
8 .SH NAME
9 fpclassify, isfinite, isnormal, isnan, isinf \- floating-point
10 classification macros
11 .SH LIBRARY
12 Math library
13 .RI ( libm ,\~ \-lm )
14 .SH SYNOPSIS
15 .nf
16 .B #include <math.h>
17 .P
18 .BI "int fpclassify(" x );
19 .BI "int isfinite(" x );
20 .BI "int isnormal(" x );
21 .BI "int isnan(" x );
22 .BI "int isinf(" x );
23 .fi
24 .P
25 .RS -4
26 Feature Test Macro Requirements for glibc (see
27 .BR feature_test_macros (7)):
28 .RE
29 .P
30 .\" I haven't fully grokked the source to determine the FTM requirements;
31 .\" in part, the following has been tested by experiment.
32 .BR fpclassify (),
33 .BR isfinite (),
34 .BR isnormal ():
35 .nf
36 _ISOC99_SOURCE || _POSIX_C_SOURCE >= 200112L
37 .fi
38 .P
39 .BR isnan ():
40 .nf
41 _ISOC99_SOURCE || _POSIX_C_SOURCE >= 200112L
42 || _XOPEN_SOURCE
43 || /* Since glibc 2.19: */ _DEFAULT_SOURCE
44 || /* glibc <= 2.19: */ _BSD_SOURCE || _SVID_SOURCE
45 .fi
46 .P
47 .BR isinf ():
48 .nf
49 _ISOC99_SOURCE || _POSIX_C_SOURCE >= 200112L
50 || /* Since glibc 2.19: */ _DEFAULT_SOURCE
51 || /* glibc <= 2.19: */ _BSD_SOURCE || _SVID_SOURCE
52 .fi
53 .SH DESCRIPTION
54 Floating point numbers can have special values, such as
55 infinite or NaN.
56 With the macro
57 .BI fpclassify( x )
58 you can find out what type
59 .I x
60 is.
61 The macro takes any floating-point expression as argument.
62 The result is one of the following values:
63 .TP 14
64 .B FP_NAN
65 .I x
66 is "Not a Number".
67 .TP
68 .B FP_INFINITE
69 .I x
70 is either positive infinity or negative infinity.
71 .TP
72 .B FP_ZERO
73 .I x
74 is zero.
75 .TP
76 .B FP_SUBNORMAL
77 .I x
78 is too small to be represented in normalized format.
79 .TP
80 .B FP_NORMAL
81 if nothing of the above is correct then it must be a
82 normal floating-point number.
83 .P
84 The other macros provide a short answer to some standard questions.
85 .TP 14
86 .BI isfinite( x )
87 returns a nonzero value if
88 .br
89 (fpclassify(x) != FP_NAN && fpclassify(x) != FP_INFINITE)
90 .TP
91 .BI isnormal( x )
92 returns a nonzero value if
93 (fpclassify(x) == FP_NORMAL)
94 .TP
95 .BI isnan( x )
96 returns a nonzero value if
97 (fpclassify(x) == FP_NAN)
98 .TP
99 .BI isinf( x )
100 returns 1 if
101 .I x
102 is positive infinity, and \-1 if
103 .I x
104 is negative infinity.
105 .SH ATTRIBUTES
106 For an explanation of the terms used in this section, see
107 .BR attributes (7).
108 .TS
109 allbox;
110 lbx lb lb
111 l l l.
112 Interface Attribute Value
113 T{
114 .na
115 .nh
116 .BR fpclassify (),
117 .BR isfinite (),
118 .BR isnormal (),
119 .BR isnan (),
120 .BR isinf ()
121 T} Thread safety MT-Safe
122 .TE
123 .SH STANDARDS
124 C11, POSIX.1-2008.
125 .SH HISTORY
126 POSIX.1-2001, C99.
127 .P
128 In glibc 2.01 and earlier,
129 .BR isinf ()
130 returns a nonzero value (actually: 1) if
131 .I x
132 is positive infinity or negative infinity.
133 (This is all that C99 requires.)
134 .SH NOTES
135 For
136 .BR isinf (),
137 the standards merely say that the return value is nonzero
138 if and only if the argument has an infinite value.
139 .SH SEE ALSO
140 .BR finite (3),
141 .BR INFINITY (3),
142 .BR isgreater (3),
143 .BR signbit (3)