release 3.2.1

[Tatsuro MATSUOKA]

Hello

Perhaps　

--- Riccardo Corradini  wrote:

> % program bug.m
> 
> clear all;
> k=70
> rm = rand(k,k);
> im = i*rand(k,k);
> M =[rm+im];
> 
> eig(M) 
> 
> These are the results with octave3.0.5
> octave:2> bug
> ** On entry to ZGEEV parameter number 5 had an illegal value
> k = 70
> error: exception encountered in Fortran subroutine zgeev_
> error: unrecoverable error in zgeev
> error: near line 9 of file `/home/corradin/octbenchmark/bug.m'
> 
> These are the results with octave3.2.0
> octave:2> bug
> panic: Segmentation fault -- stopping myself...
> k = 70
> attempting to save variables to `octave-core'...
> save to `octave-core' complete
> Segmentation fault
> 
> I think that octave3.2.0 should give the same results of octave3.0.5. 
> Hope this helps to show the bug.

I do not think that it is bug of octave-3.2.0 or octace-3.2.1.

I have test bug.m  on octave-3.0.5 on mingw packaged by Benjamin and 0ctave-3.2.1-rc4 on mingw built
myself.  The script bug.m gives the same results.  Octave-3.2.0 mingw packaged by Benjamin hangs up by
clear all. However, it is removed. The script bug.m gives the same results.

Benjamin's Binary and my binaries use Lapack-lite 3.1.1 and ATLAS 3.8.3.

Please say your platform and package tha that you used.  Errors of this type are strongly depended on
how the maintainer carefully treats the lapack and Atlas libraries.

Regards

Tatsuro

--------------------------------------
Power up the Internet with Yahoo! Toolbar.
http://pr.mail.yahoo.co.jp/toolbar/