[Octave-bug-tracker] [bug #50553] eps with imaginary argument gives wrong value

[Rik]

Update of bug #50553 (project octave):

                Category:                    None => Octave Function        
              Item Group:                    None => Matlab Compatibility   
                  Status:                    None => Confirmed              
        Operating System:               GNU/Linux => Any                    

    _______________________________________________________

Follow-up Comment #1:

Confirmed.  The function eps is written in C++ and is located in
libinterp/corefcn/data.cc.  The bit of code which is the problem is


          Array<double> x = args(0).array_value ();

          Array<double> epsval (x.dims ());

          for (octave_idx_type i = 0; i < x.numel (); i++)
            {
              double val = ::fabs (x(i));
              if (octave::math::isnan (val) || octave::math::isinf (val))
                epsval(i) = lo_ieee_nan_value ();
              else if (val < std::numeric_limits<double>::min ())
                epsval(i) = pow (2.0, -1074e0);
              else
                {
                  int expon;
                  octave::math::frexp (val, &expon);
                  epsval(i) = std::pow (2.0,
                                        static_cast<double> (expon - 53));
                }

              retval = epsval;
            }


The first line asks for "array_value" which produces an array of doubles (and
thus only the real part).  It seems that this needs to be broken up in to two
sections.


  if (args(0).is_complex_type)
    {
      // Different calculation for complex
    }
  else
    {
       // Existing code
    }


I don't know if the best strategy is to declare a complex NDArray and then use
the abs() function to produce a real NDArray and then iterate through that. 
Or whether we should be using Array directly.




    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?50553>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/