Merge pull request #1958 from RosettaCommons/roccomoretti/chifix
ICOOR/CHI correspondence fixes
There's an implicit assumption in ResidueType that the two terminal bonds of a chi are situated along the ICOOR tree. We get that right most of the time, but there's a couple of examples where it doesn't quite work.
This PR adds a check during finalize() to make sure things are set up appropriately w/r/t CHI and ICOORD trees. It also fixes some of the existing params files which don't get things correct, as well as the SDF/CIF loading code which was ignoring the issue.
Trajectory changes to pna and vancomycin integration tests expected: these use params files that changed.