unixdev.net


Switch to SpeakEasy.net DSL

The Modular Manual Browser

Home Page
Manual: (SunOS-4.1.3)
Page:
Section:
Apropos / Subsearch:
optional field

FREXP(3M)                                                            FREXP(3M)



NAME
       frexp, modf, ldexp - traditional UNIX functions

SYNOPSIS
       #include <&lt;math.h>&gt;

       double frexp(value, eptr)
       double value;
       int *eptr;

       double ldexp(x,n)
       double x;
       int n;

       double modf(value, iptr)
       double value, *iptr;

DESCRIPTION
       These  functions  are provided for compatibility with other UNIX system
       implementations.  They are not used internally in libm or libc.  Better
       ways  to accomplish similar ends may be found in ieee_functions(3M) and
       rint(3M).

       ldexp(x,n) returns x * 2**n computed by  exponent  manipulation  rather
       than  by  actually  performing  an  exponentiation or a multiplication.
       Note:  ldexp(x,n)  differs  from  scalbn(x,n),  defined  in  ieee_func-
       tions(3M),  only  that  in  the  event  of IEEE overflow and underflow,
       ldexp(x,n) sets errno to ERANGE.

       Every non-zero number can be written uniquely as x *  2**n,  where  the
       significant x is in the range 0.5 &lt;= |x| &lt; 1.0 and the exponent n is an
       integer.  The function frexp() returns  the  significant  of  a  double
       value  as  a  double quantity, x, and stores the exponent n, indirectly
       through eptr.  If value == 0, both results returned by frexp() are 0.

       modf() returns the fractional part of value  and  stores  the  integral
       part indirectly through iptr.  Thus the argument value and the returned
       values modf() and *iptr satisfy

              (*iptr + modf) == value

       and both results have the same sign as value.  The definition of modf()
       varies  among  UNIX system implementations, so avoid modf() in portable
       code.

       The results of frexp() and modf() are not defined when value is an IEEE
       infinity or NaN.

SEE ALSO
       ieee_functions(3M), rint(3M)



                                21 January 1988                      FREXP(3M)