CIS/TRANS for consecutive double bonds

[fgrunewald]

A curious problem arises for cis trans definition of consecutive double bonds. Should the smiles be C1=CC=C(C=C1)/C=C/C=C//C2=CC=CC=C2 or C1=CC=C(C=C1)/C=C/C=C/C2=CC=CC=C2. I think commonly one symbol is used. In either case our annotation fails since the middle atom only get's assigned one classification based on the first bond

[pckroon]

If I carefully read the opensmiles syntax definition C1=CC=C(C=C1)/C=C//C=C/C2=CC=CC=C2 is invalid. / (and \) is a 'bond' token, and it's not allowed to have 2 bonds directly following eachother.