On Oct 13, 2011, at 12:59 AM, Max Ogden wrote: In your case you seem to be somehow replacing documents without first knowing about them, which seems weird to me. Can you shed some more light on what application level features depend on the force_update/delete_conflicts API?
This does seem weird in many contexts, but I ran into a situation last month where it would have been useful: deleting a replication document from the _replicator database. The problem is that, while the replication is active, the server task is updating its document fairly frequently with new state info. Trying to stop the replication by deleting the document while this is going on can be frustrating because it’s a race condition — you get the latest revision, try to delete it, get a conflict, get the new revision, try to delete that, get another conflict… I finally had to write a loop that would keep doing the above over and over until it succeeded. Based on this, I think it can be useful in some contexts to have an override that says “I want to delete (or overwrite) this document regardless of what its current state is”. —Jens
