[Octave-bug-tracker] [bug #30521] Error in diag function for complex numbers

[Rik]

Update of bug #30521 (project octave):

                  Status:                    None => Confirmed              

    _______________________________________________________

Follow-up Comment #1:

The behavior is slightly better in the development version of Octave (3.3.51)
in that the first two parts work.  However, the third step is still broken.

m1 = [0.1+2i]
m2 = diag(m1)
m3 = diag(m2)

typeinfo(m1)
typeinfo(m2)
typeinfo(m3)
-------------------------
Results
m1 =    0.1000 + 2.0000i
m2 =  0.10000 + 2.00000i
m3 =  0.10000

ans = complex matrix
ans = complex scalar
ans = scalar

>From the documentation, diag is designed to work on a vector to produce a
matrix or on a matrix to produce a vector.  I'm guessing that there's no test
to see if the input is "complex scalar" and handle that case.  At least for a
scalar input, it returns a scalar output and that should probably be the
default for complex inputs as well.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?30521>

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