[Octave-bug-tracker] [bug #53462] angle(0+xi) producing inaccurate result.

[John W. Eaton]

Follow-up Comment #2, bug #53462 (project octave):

For inf*i, Octave computes inf * complex (0, 1), which produces complex (NaN,
Inf), so I think angle is doing the right thing to return NaN for that value.

What happens for you if you do


angle (complex (0, Inf))
angle (complex (NaN, Inf))
angle (complex (Inf, NaN))
angle (complex (NaN, NaN))
angle (complex (Inf, Inf))


in both Octave and Matlab?

If these all agree, then I think the actual "problem" is that Octave doesn't
have a special case for Inf * i.  The real issue here is that Octave (and
Matlab) don't have pure imaginary numbers.  So given that, the question is
what (if any) special cases should be handled for operations when the real
part of a complex number is zero.  We've been over this issue a number of
times before, and it's not clear to me what the correct thing is.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?53462>

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