Re: Octave.app 2.9.16 for MacOSX

[Thomas Treichl]

Ok, thanks Henry. So we can summarize that the messages are warnings and do not 
produce broken packages (ie. the packages are working - that's good). BTW, I 
never would request that you delete another binary (your 2.1.73), if Octave.app 
wouldn't work beside your 2.1.73 then I would have a big problem ;) pkg.m also 
does not exist in 2.1.73, it first appeared somewhen 2.9.9 to 2.9.10.
I'll have a look for these ugly warning messages, maybe I find the culprit that 
tries to link against two different/same math-libraries (oh, I forgot to tell 
you that this is the problem: -lm and -lmx, ie. g++ tries to link against 
libm.dylib and libmx.dylib - both seem to be math libraries - but only one 
library should be used).

Thanks for testing!

  Thomas

-------- Original-Nachricht --------
> Datum: Mon, 05 Nov 2007 15:08:57 -0800
> Von: "Henry F. Mollet" <address@hidden>
> An: Thomas Treichl <address@hidden>, Octave_post <address@hidden>
> Betreff: Re: Octave.app 2.9.16 for MacOSX

> on 11/5/07 12:37 PM, Thomas Treichl at address@hidden wrote:
> > David posted this to the thread 'Re: Problem with pkg install on 2.9.16
> cvs':
> > 
> > Remove your existing packages and the ~/.octave_packages file.. The
> > archprefix tag was added to the package database in 2.9.15 to allow
> > correct Fedora packaging and so the old package databases are invalid..
> > 
> > Can you try again, thanks ;)
> > 
> >    Thomas
> 
> Thomas, (above was offlist, applies to OS X 10.4.10 on PowerPC G4)
> 1. I have deleted the  ~/.octave_packages file but *not* the octave-2.1.71
> installation and tried again:
> octave-2.9.16:4> pkg install ~/Desktop/odepkg-0.3.3.tar.gz
> 
> 2. Apparently successful (see below for demo output and I also got the
> graph). The installation log still produced "tons" of warnings re multiple
> definitions but no errors at the end.
> 
> octave-2.9.16:5> demo ode5r
> ode5r example 1:
> 
>  A = odeset ('RelTol', 1e-1, 'AbsTol', 1e-2);
>  [vx, vy] = ode5r (@odepkg_equations_secondorderlag, [0 2.5], [0 0], ...
>     A, 5, 2, 0.02, 0.1);
> 
>  plot (vx, vy(:,1), '-ob;y, x2;', vx, vy(:,2), '-or;x1;', ...
>     vx, ones(length(vx),1)*5, '-og;u;');
> 
>  % ---------------------------------------------------------------------
>  % The figure window shows the state variables x1, x2 as well as the
>  % input signal u and the output signal y(=x2) of a second order lag
>  % implementation (cf. the control theory). The function ode5r was
>  % called with an option argument A that has been set before with the
>  % command "odeset" and with further parameters "5, 2, 0.02, 0.1" that
>  % are passed to the set of ordinary differential equations.
> 
> 3. octave-2.9.16:6> pkg install ~/Desktop/signal-1.0.4.tar.gz
> mkoctfile: stripping disabled on this platform
> 
> Apparently now successful also without any errors at the end. Again "tons"
> of warnings re multiple definitions and still the above stripping disabled
> comment.
> Thanks.
> Henry
> 

-- 
GMX FreeMail: 1 GB Postfach, 5 E-Mail-Adressen, 10 Free SMS.
Alle Infos und kostenlose Anmeldung: http://www.gmx.net/de/go/freemail