RE: [octave:code] 3 new commits to Octave Forge Code

[Marco Vassallo]

> Subject: Re: [octave:code] 3 new commits to Octave Forge Code

> From: address@hidden
> Date: Tue, 11 Jun 2013 08:41:05 +0200
> To: address@hidden
> CC: address@hidden
>
>
> On 11 Jun 2013, at 05:47, Repository Octave Forge code <address@hidden> wrote:
>
> > Array doesn't work fine, we come back to Matrix
>
>
> what is the problem with Array?
> If you see incorrect behaviour, I doubt the proble is with the Array class itself, it is more likely with my code,
> it was written very quickly just an example, so I am not surprised it might be flawed.
> Rather than delete it, though, you should be describing what is not working and looking at a way to fix it.
>
Hi,

the problem arises when reading the content of the struct:

             1) Matrix p (a.contents (a.seek ("p"))(0).matrix_value ());

while it seems to fail with


             2) Array<octave_idx_type> p (a.contents (a.seek ("p"))(0).matrix_value ());

or with

            3) Array<octave_idx_type> p (a.contents (a.seek ("p"))(0).array_value ());

The problem was that I submitted the version with option number 2 without testing it extensively
(it passed the test and I was happy..) but when I use it later with more complicated data it failed. So I come back to option number 1) so that we have in the repository something working and I will investigate later if the problem is really about line 2) or 3) or if it is in the way we use it.

Marco
problem is option 2) or 3) 


>
> c.