Re: gEDA-user: New beta version of gEDA Suite CD up -- please test!

[David Hart <dhart155@xxxxxxxxxxx>]

On Sun, 2005-03-27 at 17:08 -0500, Stuart Brorson wrote:
> > 
> > I just tried installing the gEDA/gaf Suite CD (i.e., geda-
> > install-20050325.iso) on my Fedora Core 3 installation and it fails
> > building ngspice.
> > 
> > Checking config.log for ngspice, I see that there are multiple undefined
> > references. Also, libncurses, libreadline, and libgc libraries were
> > configured (I have libgc installed from a previous install of TclSpice).
> 
> The install CD configures with readline enabled.  Libgc is the garbage
> collection library, which you apparently have on your system.  It is
> autosensed by ngspice's configure script.
> 
> > 
> > Uninstalling the latest version (version 6.4) of the Boehm-Demers-Weiser
> > conservative garbage collector (i.e., libgc) resolves this issue and the
> > build is successful.
> 
> Interesting report!  Thank you!
> 
> Just to make it clear -- you are saying that ngspice's configure
> script found that you had the Boehm-Weiser Conservative Garbage
> Collector installed on your system and configured it in.  Then,
> ngspice died during "make".  You fixed the problem by removing the
> garbage collection library libgc from your library of .sos.  Correct?

Exactly. I removed the libgc library.

> 
> I have seen other cases where the presence of a garbage collection
> library causes ngspice compilation to fail.  
> 
> If this is the case, all I need to do is get rid of checking for the
> GC library, or just force no garbage collector detection at configure
> time.  I will try to do that for the final release (coming any day
> now. . .)
> 
> > Will this be a problem with installation of TclSpice v0.2.17 - June 23,
> > 2004, which recommends installing the garbage collector?
> 
> IIRC, the garbage collector is optional.  Therefore, getting rid of it
> shouldn't cause a problem.  The only problem might arise when you run
> tclspice for a long time on a large design, and it uses up all your
> memory during a calculation.  You'll just have to see if this
> happens. 
> 
> Thanks again for the report!
> 
> Stuart
> 
> 
-- 
David Hart <dhart155@xxxxxxxxxxx>