Void atom created when merging a ligand

[FilomenoSanchez]

When a ligand is fit in Moorhen, a void atom is created just before HETATM records in mmCIF file. This void atom has a real residue name but no atom name, which breaks all things in subsequent programs.

[FilomenoSanchez]

Probably this issue originates in molecules_container_t::merge_molecules rather than fitting the ligand itself

[FilomenoSanchez]

Currently a void atom is added at the end of the chain when writing a mmcif (similar to a TER record). It seems that we need to edit residue::make_atom_cif to check if is_ter.