-  add support for discrete=1 loading of CIF files 
-  improve multi-model loading (inhomogenious models, pdb_honor_model_number support) 
-  support loading multiple objects (“data_” blocks) from one CIF file (multiplex != 0) 
-  support _symmetry_equiv.pos_as_xyz, _space_group_symop.operation_xyz 
-  set cartoon_trace_atoms and ribbon_trace_atoms for CA-only models 
-  load biological unit if assembly setting is set 
-  Ignore “auth” identifiers with cif_use_auth=off (default: on) 
-  Reimplement reading full sequence from mmCIF (fill in missing CA atoms for polypeptides). Only with cif_use_auth=off. Uses _entity_poly_seq. 
-  Improve connect_mode=4 support and performance: Bonding for most common residue types provided in $PYMOL_DATA/chem_comp_bond-top100.cif. Other types are downloaded on the fly from PDBeChem. 
-  read metalc and disulf bonds