[Octave-bug-tracker] [bug #40467] cellfun ignores UniformOutput=false sometimes & needs doc update

[Markus Appel]

Follow-up Comment #1, bug #40467 (project octave):

Update:

I was a bit too fast jumping too conclusions with "isclass", the documentation
actually shows that a second argument is necessary which makes it work as
expected.

Looking into Matlab R2012b I saw that calls of the form cellfun('isempty',...)
are supported for backwards compatibility only and do not accept the
"UniformOutput" property at all.

Here is another example of "UniformOutput" failing, this time with function
handles:


octave-cli:3> cellfun(@isreal, {1 inf nan []}, "UniformOutput", false)
ans =

   1   1   1   1

octave-cli:4> cellfun(@iscomplex, {1 inf nan []}, "UniformOutput", false)
ans = 
{
  [1,1] = 0
  [1,2] = 0
  [1,3] = 0
  [1,4] = 0
}


    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?40467>

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