Fix a corner case in hydrogen suppression when a double bond has 2 hydrogens on the same end!
Fix the small cycle/ring algorithms and reuse them in the stereo element factory.
Fix a minor corner case in hydrogen normalisation.
Watch out for NaN when the offset of double bonds is computed.
Missed change to the BeamToCdk part
Allow coordinate bonds to be parsed from CXSMILES - https://github.com/cdk/depict/issues/102
Spotted in molwitch-cdk, this check should only be warning (and skipping) if hashed wedge is between two stereocenters!
Reuse the usmigen field when creating abbreviations, this allows hydrogen isotopes in the abbreviated structure.