failed asertion

[Ben Sapp]

"Richard B. Kreckel" wrote:
>
> There doesn't seem to be enough information to judge what's going on.  I
> don't know how you allocate expressions.  Can you strip this problem down
> to the bare GiNaC-C++-layer?  I would bet you'll find the problem
> anyways, while doing so, though...

I did this and you are right I did find the problem.  I ended up
changing a function from this: 

GiNaC::symbol octave_sym::sym_value () const
{
  return x;
}

to this:

GiNaC::symbol octave_sym::sym_value () const 
{
  return ex_to_symbol(GiNaC::ex(*(x.duplicate ())));
}

Octave classes overload the new operator and sometimes octave types get
created as a const value.  This was also causing me other problems
relating to the fact that GiNaC does reference counting.  An ex variable
might be a const and GiNaC might change the reference count.  This
caused my code to crash in unpredictable ways.  I can't say I completely
understand why the particular assertion failed, but it does not seem to
be a problem any more.  

Thanks.  

-- 
Ben Sapp                         Los Alamos National Laboratory
email: <mailto:bsapp at lanl.gov>   Phone: (505)667-3277
Fax:   (505)665-7920             URL:   http://www.neutrino.lanl.gov/
--
-
To UNSUBSCRIBE, email to ginac-list at ginac.de with a subject of "unsubscribe".