Again review/comments from Alex would be good

@johnmay, did you get feedback from Alex?

Nope, @aclarkxyz any thoughts?

Changing the default is a total dealbreaker: fingerprints must not change. If people want to make sub-variants based on different inputs that's fine, but they have to be non-default, labelled as something else, and difficult to confuse with the originals.

The fingerprints do not change.

OK, well as long as that's guaranteed, all good then.

@johnmay, do you have, perhaps, some test results that should for, say ChEMBL, that the FPs are identical before and after your patch?

There is a built in regression test on a set of structures. No changes there but there can be differences when the structure is drawn poorly - e.g. #377. I use my own ECFP algorithm so take the patch or leave it - make no difference to me :-).