Comment 2 for bug 1560574

Perhaps, but at least it's deterministic. I would expect if the system was littered with various copies of both the object and tombstones, and I did a GET, that I would get undefined results. Hopefully the ticket I opened with regards to using X-Newest will sort out that particular issue.

However, why do we treat tombstones are normal 404s and then search 2x replicas? If I find 1 copy of the object, I return the results, no matter if it's the latest and greatest or not. But I can go through 5 tombstones, find a copy of the object, and still return it?

It feels like a tombstone presence should count the same as object presence, subject to normal eventual consistency, and let X-Newest sort out resolving conflicts.