Merge pull request #739 from RosettaCommons/vmullig/fix_swm_tests
Trying to fix failing stepwise Monte Carlo integration tests.
I think I've tracked the problem down to a point at which remove_variant_type_from_residue was being passed a residue and a pose that didn't contain that residue. This was an existing issue in protocols/stepwise/modeler/util.cc that was only revealed by my recent changes. Fixed now.