[Bug middle-end/17603] [4.0 Regression] cpowf and cpowl give wrong results

[pcarlini at suse dot de]

------- Additional Comments From pcarlini at suse dot de  2004-09-28 15:44 -------
In fact, builtins are not directly responsible.
This trivial, pure C, testcase:

#include <complex.h>
#include <stdio.h>

int main()
{
  long double complex b, a, r;

  b = -1;
  a = 0.5;
  r = cpowl(b, a);

  printf("%Lg, %Lg\n", creall(r), cimagl(r));
  
  return 0;
}

Suffices to show the problem on x86_64:
3.69315e-4664, 0

instead of the expected (3.3 and 3.4 are ok -> *not* a glibc problem!):
-2.71051e-20, 1

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[4.0 Regression] Cpowf and  |[4.0 Regression] cpowf and
                   |cpowl builtins broken       |cpowl give wrong results


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17603