Status: Won't Fix
Fix Version/s: None
Component/s: obs_cfht, obs_decam, obs_lsst, obs_subaru
This may be a "won't fix" because it only affects the gen2 CameraMappers, but lsst, subaru, decam, and cfht all use filter.getId() in their calculation of the singleFilter coadd exposure id. The new FilterLabel does not have a numeric Id, so these mappers will have to be changed. However, I do not know what the time ordering of gen2 removal vs. Filter removal is: if gen2 is removed first, then this won't matter at all.
I did not change this on
DM-27170 because I did not want to alter this calculation in gen2 without more discussion, and want to get that ticket finished up.
I'm about to merge
DM-34919 (as soon as the jira ticket is marked reviewed) and that will delete this code.
Confirmed, but I had to deal with something similar in obs_cfht. I took the hash approach; so far it's passing tests.
I'm trying to implement
DM-27177and have just run into this issue in obs.lsst.LsstCamBaseMapper._computeCoaddExposureId. How should I handle it? Should I just have _computeCoaddExposureId raise NotImplementedError? Should I remove the if singleFilter block? Hash the filter name?
I have no idea what will break if, for example, the IDs are no longer unique.