Fixed a bug in the FosterBoys class wherein only the last optimization was returned and not the best one
Removed unnecessary const-ness of Fgamma functions
Added Fgamma-cached variant for repulsion integrals based on a suggestion as found here, yet no difference in execution speed is observed, concluding that this operation is far from rate-limiting.
const-ness of Fgamma functions