[IMP-dev] Particle removal and deactivation

[Daniel Russel <drussel@salilab.org>]

I can't find the imp-commit email for some reason. Anyway, there was a 
recent patch (581) changing a
model::remove_particle to deactivating the particle. This change is 
wrong as the particles are currently reference counted in C++ and 
removing the particle from the model should do the right thing assuming 
there are no python references hanging around. If it does not, it is a 
bug in kernel code which should be fixed.
In fact, the "Particle::set_is_active" method should probably go away.