release 3.2.1

[Tatsuro MATSUOKA]

Hello

I have tried to build octave-3.2.0 on http://hg.tw-math.de/release-3-2-x.

# description	Preparation repository for releases of the 3.2 series
# owner	Jaroslav Hajek <highegg_at_gmail.com>
# last change	Fri, 03 Jul 2009 18:59:07 +0200

make check results
  c:\usr\Tatsu\mingwhome\octaves\hg\release-3-2-x-work\src\data.cc  PASS  506/509  FAIL 3
  test_string.m .......................................... PASS   78/79   FAIL 1

   PASS   5726
  FAIL      4

Errors the above is the same when I have tested with lapack 3.1.1  

http://www.nabble.com/Build-result-of-octave-3.2.1-rc4-by-GCC-4.4.0-MinGW-on-the-official-site-to24426125.html#a24426125

And the test
clear all;
k=70
rm = rand(k,k);
im = i*rand(k,k);
M =[rm+im];

eig(M) 

also went well without trouble.

I think that the neither lapack 3.2.x nor octave 3.2.x are origins of errors.

Lapack should be prepared using the blas you would like to use.

If the you are using the ATLAS as a blas library the liblapack.produced by ATLAS should be merged to
the liblapack.a from the netlilb.

http://math-atlas.sourceforge.net/atlas_install/atlas_install.html#SECTION00041000000000000000
3.1.3 Adding a full LAPACK library to an already-installed ATLAS

I have produced  lapack.dll from this lapack library using the ATLAS library. 

Regards

Tatsuro



--- Sergei Steshenko wrote:

> 
> 
> 
> --- On Wed, 7/15/09, John W. Eaton <jwe at octave.org> wrote:
> 
> > From: John W. Eaton <jwe at octave.org>
> > Subject: Re: release 3.2.1
> > To: "Sergei Steshenko" <sergstesh at yahoo.com>
> > Cc: "Riccardo Corradini" <riccardocorradini at yahoo.it>, bug at octave.org
> > Date: Wednesday, July 15, 2009, 7:09 PM
> > On 15-Jul-2009, Sergei Steshenko
> > wrote:
> > 
> > | 
> > | 
> > | --- On Wed, 7/15/09, John W. Eaton <jwe at octave.org>
> > wrote:
> > | 
> > | > From: John W. Eaton <jwe at octave.org>
> > | > Subject: Re: release 3.2.1
> > | > To: "Riccardo Corradini" <riccardocorradini at yahoo.it>
> > | > Cc: bug at octave.org
> > | > Date: Wednesday, July 15, 2009, 11:33 AM
> > | > On 7-Jul-2009, Riccardo
> > | > Corradini wrote:
> > | > 
> > | 
> > | [snip]
> > | 
> > | > I'm confused now. Does this mean that you claim
> > the
> > | > bug is fixed,
> > | > there is not really a bug, the bug is in ATLAS
> > instead of
> > | > Octave, or
> > | > what?
> > | > 
> > | 
> > | [snip]
> > | 
> > | It looks like the bug manifests itself with lapack-3.2
> > rather than
> > | lapack-3.1.1.
> > | 
> > | I have noticed similar behavior with octave-3.0.5, but I
> > haven't had a chance to rebuild stUff again with various
> > 'lapack' versions (3.1.1, 3.2)
> > | and various 'atlas' versions (3.8.2, 3.8.3) in order to
> > exactly tell which
> > | version of what causes the bug to show up.
> > | 
> > | Though it looks like the bug is in 'lapack'/'atlas', it
> > may also be that,
> > | 
> > | say, 'lapack' interface has changed, but 'octave' is
> > unaware of this.
> > 
> > What about lapack 3.2.1? Does that also show the
> > problem for you?
> > I'm using the Debian package for 3.2.1 and I don't see the
> > problem.
> > 
> > jwe
> > 
> 
> Yes - I actually meant lapack-3.2.1.
> 
> The point is that lapack-3.2* is significantly different from lapack-3.1*,
> that's why it takes more than just to change version in my environment to
> revert back to lapack-3.1.1.
> 
> Just to remind/clarify - I observe the problem in octave-3.0.5 - I am not
> yet building octave-3.2*.
> 
> Regards,
>   Sergei.
> 
> 
>       
> 
> _______________________________________________
> Bug-octave mailing list
> Bug-octave at octave.org
> https://www-old.cae.wisc.edu/mailman/listinfo/bug-octave
> 


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