Unfortunately, we don't have a good enough way to discern this possibility. Technically we could add yet another attribute, but that does not seem to be worth the cost at this point, in terms of complexity and possibly even performance. I take the blame for this one. I identified the need for the second merge commit in review but didn't require it in the first merge. These changes really should have been done in a single merge. That would have prevented this situation.
72 KiB
72 KiB