Re: need test run on Mac

[c.]

On 27 Feb 2014, at 07:28, John W. Eaton <address@hidden> wrote:

> On 02/27/2014 01:15 AM, Lukas Reichlin wrote:
>> On 27.02.2014, at 06:52, "John W. Eaton" <address@hidden> wrote:
>> 
>> octave:1> y = zeros(1,25);
>> octave:2> u = ones(1,25);
>> octave:3> tsam = 0.1;
>> octave:4> data = iddata(y,u,tsam);
>> warning: iddata: more outputs than samples - matrice 'y' should probably be 
>> transposed
>> warning: iddata: more inputs than samples - matrice 'u' should probably be 
>> transposed
>> octave:5> n4sid(data,1);
>>  ** On entry to IB01AD parameter number  7 had an illegal value
>> 
>> warning: octave_chunk_buffer::clear: 2 active allocations remain!
>> ra:~ lukas$
>> 
>> I installed Octave through the Fink package manager on OS X 10.8.5.
> 
> Thanks.
> 
> OK, something funny is happening there, because what should happen
> after the error occurs in IB01AD is that an exception should be
> thrown, and you should see subsequent error messages that look like
> this:
> 
>  error: __sl_ib01ad__: exception encountered in Fortran subroutine ib01ad_
>  error: called from:
>  error: /usr/share/octave/packages/control-2.6.2/__slicot_identification__.m 
> at line 156, column 13
>  error:   /usr/share/octave/packages/control-2.6.2/n4sid.m at line 237, 
> column 5
> 
> Instead you're getting some warning message about memory allocation
> which should not be happening.  It works as I would expect on my
> Debian system, so I'm suspecting that either Octave is miscompiled on
> the Mac or our assumptions about this part of Octave's exception
> handling scheme are wrong and need to be revised to work properly.
> 
> jwe

can you suggest a simple example to test these assumptions?
c.