On DM-37428 we found that the cause of the bug was that afw::image::positionToIndex does not use the exposure's XY0, but instead assumes it is (0, 0). In a follow-up discussion with Jim Bosch, he said to "be aware that most of the C++ interfaces to Image access don't respect XY0 at all either; a long while back we finally got pretty much all of the Python ones under control, but we never had the effort to go after the C++ ones."
This ticket is to discover and fix those places in afw::image that do not make proper use of XY0. It is not expected that we have effort to work on this at the moment, but we should have a ticket to cover this technical debt in case it comes up again in the future.