Extension of compute group/group to kspace interactions (Ewald and PPPM)

I have some beta code that extends LAMMPS compute group/group to
include kspace interactions (for both Ewald and PPPM). This allows
one to calculate the total energy and force between any two arbitrary
groups of atoms in the simulation (and now includes the long-range
electrostatic interaction). See
http://lammps.sandia.gov/doc/compute_group_group.html for more

I have validated the code in a few ways and think it is right, but
could use another opinion. If anyone wants to try it out/help further
validate the code that would be great. Shoot me an email I will send
you a copy: [email protected]