A while ago I reported problems with the output of Equation_Systems,
most notably zero solution vectors. Today I finally had time to play
with Libmesh again, and I finally found the reason. The problem was
that I still used ImplicitSystem in my code, instead of the newer
I'm glad that one was so simple, but it took me some hours to figure
KIRK, BENJAMIN (JSC-EG) (NASA) writes:
> Thanks. I changed the way EquationSystems handles iterating over the
> systems, and this must cause the problem.
> -----Original Message-----
> The Equation_Systems.build_solution_vector returns a wrong solution vector
> for multiple systems. If more than one Equation_Systems are defined, only
> the variables of the first are inserted into the solution vector. All other
> systems have just zero entries.
> I am using the newest CVS source.