Re: numel(foo{:}) - feature or bug ?

[Sergei Steshenko]

--- On Fri, 8/5/11, Marco atzeri <address@hidden> wrote:

> From: Marco atzeri <address@hidden>
> Subject: Re: numel(foo{:}) - feature or bug ?
> To: "Sergei Steshenko" <address@hidden>
> Cc: address@hidden
> Date: Friday, August 5, 2011, 11:56 PM
> On 8/6/2011 8:03 AM, Sergei Steshenko
> wrote:
> :
> 
> >> I supect you are thinking as foo(:) and foo{:} as
> >> equivalent
> >> but they are not.
> >> 
> 
> > 
> > So, how does your example answer my question:
> > 
> > "
> > Why does adding an element causes non-monotonicity in
> numel(foo2d{:}) ?
> > "
> 
> because numel is not supposed to handle food2d{:} and
> it is producing only garbage-like output.
> It should refuse to work as size and class do.
> 

[snip]


Thanks for confirming this. I mean

"
numel is not supposed to handle food2d{:}
"

part.

So, if your statement is correct, then the fact that in octave-3.4.2
it "works", i.e. the construct is allowed, is a bug in octave-3.4.2.

Also, if your statement is correct, all those who gave justification to
the original '72' number (which became '96' on Windows according to a
recent post: 
http://octave.1599824.n4.nabble.com/numel-foo-feature-or-bug-td3720972.html#a3721796
 )
are wrong in their attempts to justify any number.

Again, if your statement is correct, could you please reopen

http://savannah.gnu.org/bugs/?33954

and state "numel is not supposed to handle food2d{:}" there ?

Thanks,
  Sergei.