There is some cruft and extra coordination in the administration functions from the legacy multi-database implementation. This could probably be simplified a lot to assume a set only exists on one database at a time, and would likely improve other concepts (the forced update and removal path is pretty strange right now.)
There is some cruft and extra coordination in the administration functions from the legacy multi-database implementation. This could probably be simplified a lot to assume a set only exists on one database at a time, and would likely improve other concepts (the forced update and removal path is pretty strange right now.)