std::arg() bug : not repetitive ?

[Dave Korn]

Marco Atzeri wrote:
> Hi DaveK,
> the following test case on complex numbers 
> is producing, puzzling result on cygwin (both 1.5 and 1,7) 
> with gcc-4.3.2 (and also 3.4.4), while working on other platform:
> 
> #include <iostream>
> #include "oct-cmplx.h"
> 
> int main ()
> {
>   Complex z1 (1.0, 1.0), z2 (1.0, 1.0);
>   std::cout << (arg(z1)) << '\n';
>   std::cout << (arg(z2)) << '\n';
>   std::cout << (arg(z1)<arg(z2)) << '\n';
>   std::cout << (arg(z1)-arg(z2)) << '\n';
> }
> 
> 
> $ g++-4 comp_2.cc -o0 -o comp_2
> 
> $ ./comp_2
> 0.785398
> 0.785398
> 1
> -3.06287e-17    <<-- arg(1+i) is lower then arg(1+i) !!
> 
> Using different complex numbers is also possible to get 
> 
>   arg(-1-i) bigger then arg(-1-i)
> 
> Any idea what could cause it ? newlib ?

  Or maybe it's PR323 (excess precision) in some aspect.  Don't know yet, I'll
have to have a look into it.

    cheers,
      DaveK


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