complex number

[Marco Atzeri]

--- Mer 1/4/09, Corinna Vinschen  ha scritto:

> Da: Corinna Vinschen 
> Oggetto: Re: complex number
> A: cygwin@cygwin.com
> Data: Mercoledì 1 Aprile 2009, 17:22
> On Apr  1 14:48, Marco Atzeri
> wrote:
> > 
> > Dear All,
> > I was trying to understand why this code
> > 
> > #include <iostream>
> > #include <complex>
> > 
> > int main()
> > {
> >   double a = 0;
> >   double b = 1. / a;
> >   a += 1;
> >   std::cout << std::abs
> (std::complex<double> (b, a)) << '\n';
> >  }
> > 
> > produce Inf on most platform and NaN on cygwin.
> 
> That's the result of the newlib function which is defined
> as
> 
>   double
>   cabs(z)
>   struct complex z;
>   {
>     return hypot(z.x, z.y);
>   }
> 

Hi Corinna,
that is not correct as hypot( 0, Inf) 
return Inf and not NaN.

There is a cabs backup implementation in libgcc that is probably
the one used when system cabs is not available.


> > I found that newlib have the function cabs (complex
> absolute)
> > but in cygwin the prototypes is
> > 
> >     extern double cabs();
> 
> The math.h header is from newlib as well.  I don't
> know why the
> math.h header is missing a normal prototype.  I
> *assume* that nobody
> ever made the effort to convert these old functions (cabs
> is from 1994)
> to the ANSI C style.  You should ask this question on
> the newlib list.
> 
> > so just a placeholder but not a useful function.
> > There is any reason why cygwin is missing functional
> > C99 complex functions ?
> 
> Sorry, it's all about newlib.
> 
> 
> Corinna

Ok,
I will follow up on newlib,
I was not sure it is a newlib issue or cygwin specific,
as newlib have cabs function and for linux also
the expected header:

http://sourceware.org/cgi-bin/cvsweb.cgi/src/newlib/libc/sys/linux/include/cmathcalls.h?cvsroot=src

Regards
Marco





--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/