Merge pull request #2211 from RosettaCommons/aleaverfay/bugfix_chain_endings_in_pdb_reading
Bugfix in Pose construction logic
Fixing another chain-endings bug that is tripped when the Cterm of onechain (residue i) is missing some backbone atoms, AND the Nterm of the next chain
is also missing some backbone atoms, so that Rosetta won't put a Cterm at
residue i-1 nor an Nterm at i+2. As a result, a two chain PDB will get
read in as a single chain. Super simple fix for this.