[Scilab-Dev] syredi broken?

[Arvid Rosén]

Hi!
I am having problems with the macro eqiir in scilab 5.4 on MacOS. I have traced it down to the call to syredi, which is a fortran function. When run the example from the help file I get different results from scilab 5.4 and scilab 4.


// Scilab 5.4
[fact,b2,b1,b0,c1,c0,zzeros,zpoles]=syredi(1,4,[2*%pi/10,4*%pi/10,0,0],0.02,0.001);
disp(b1)
--> [2 2 1]

// Scilab 4.1
[fact,b2,b1,b0,c1,c0,zzeros,zpoles]=syredi(1,4,[2*%pi/10,4*%pi/10,0,0],0.02,0.001);
disp(b1)
--> [ - 0.5354229    0.3489640    1.  ]

The poles end up correct, but the zeros are rounded or casted to int or something like that.

Does anyone know anything about this? Should I report it as a bug?

This turns out to break a lot of code for me, but in a very non-obvious way.

Best regards,
Arvid

--
Arvid Rosén

Softube AB
S:t Larsgatan 10c
SE-582 24 Linköping
Sweden
Mobile: (+46) 73 25 123 47
Office: (+46) 13 21 16 23
Fax: (+46) 13 10 00 23
Mail: arvid at softube.com
URL: www.softube.com

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/dev/attachments/20121010/acc94320/attachment.htm>