.\" Copyright 2002 Walter Harms (walter.harms@informatik.uni-oldenburg.de)
.\"
.\" %%%LICENSE_START(GPL_NOVERSION_ONELINE)
.\" Distributed under GPL
.\" %%%LICENSE_END
.\"
.\" This was done with the help of the glibc manual.
.\"
.\" 2004-10-31, aeb, corrected
.TH FPCLASSIFY 3  2015-03-02 "" "Linux Programmer's Manual"
.SH NAME
fpclassify, isfinite, isnormal, isnan, isinf \- floating-point
classification macros
.SH SYNOPSIS
.nf
.B #include <math.h>
.sp
.BI "int fpclassify(" x );
.sp
.BI "int isfinite(" x );
.sp
.BI "int isnormal(" x );
.sp
.BI "int isnan(" x );
.sp
.BI "int isinf(" x );
.fi
.sp
Link with \fI\-lm\fP.
.sp
.in -4n
Feature Test Macro Requirements for glibc (see
.BR feature_test_macros (7)):
.in
.sp
.\" I haven't fully grokked the source to determine the FTM requirements;
.\" in part, the following has been tested by experiment.
.ad l
.BR fpclassify (),
.BR isfinite (),
.BR isnormal ():
.RS 4
_XOPEN_SOURCE\ >=\ 600 || _ISOC99_SOURCE ||
_POSIX_C_SOURCE\ >=\ 200112L;
.br
or
.I cc\ -std=c99
.RE
.BR isnan ():
.RS 4
_BSD_SOURCE || _SVID_SOURCE || _XOPEN_SOURCE || _ISOC99_SOURCE ||
_POSIX_C_SOURCE\ >=\ 200112L;
.br
or
.I cc\ -std=c99
.RE
.BR isinf ():
.RS 4
_BSD_SOURCE || _SVID_SOURCE || _XOPEN_SOURCE\ >=\ 600 || _ISOC99_SOURCE ||
_POSIX_C_SOURCE\ >=\ 200112L;
.br
or
.I cc\ -std=c99
.RE
.ad
.SH DESCRIPTION
Floating point numbers can have special values, such as
infinite or NaN.
With the macro
.BI fpclassify( x )
you can find out what type
.I x
is.
The macro takes any floating-point expression as argument.
The result is one of the following values:
.TP 14
.B FP_NAN
.I x
is "Not a Number".
.TP
.B FP_INFINITE
.I x
is either positive infinity or negative infinity.
.TP
.B FP_ZERO
.I x
is zero.
.TP
.B FP_SUBNORMAL
.I x
is too small to be represented in normalized format.
.TP
.B FP_NORMAL
if nothing of the above is correct then it must be a
normal floating-point number.
.LP
The other macros provide a short answer to some standard questions.
.TP 14
.BI isfinite( x )
returns a nonzero value if
.br
(fpclassify(x) != FP_NAN && fpclassify(x) != FP_INFINITE)
.TP
.BI isnormal( x )
returns a nonzero value if
(fpclassify(x) == FP_NORMAL)
.TP
.BI isnan( x )
returns a nonzero value if
(fpclassify(x) == FP_NAN)
.TP
.BI isinf( x )
returns 1 if
.I x
is positive infinity, and \-1 if
.I x
is negative infinity.
.SH ATTRIBUTES
For an explanation of the terms used in this section, see
.BR attributes (7).
.ad l
.TS
allbox;
lbw28 lb lb
l l l.
Interface	Attribute	Value
T{
.BR fpclassify (),
.BR isfinite (),
.BR isnormal (),
.BR isnan (),
.BR isinf ()
T}	Thread safety	MT-Safe
.TE
.ad
.SH CONFORMING TO
C99, POSIX.1.

For
.BR isinf (),
the standards merely say that the return value is nonzero
if and only if the argument has an infinite value.
.SH NOTES
In glibc 2.01 and earlier,
.BR isinf ()
returns a nonzero value (actually: 1) if
.I x
is positive infinity or negative infinity.
(This is all that C99 requires.)
.SH SEE ALSO
.BR finite (3),
.BR INFINITY (3),
.BR isgreater (3),
.BR signbit (3)