On Oct 13, 2011, at 1:59 PM, Max Ogden wrote: > Hi Alexander, > > I think that the reason that force_update/delete_conflicts doesn't exist is > because it's assumed that since HTTP is a stateless protocol that > applications built using Couch will keep state on the client. For instance, > if you update a document Couch will hand you back the new revision so you > don't have to fetch it separately. > > 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?
Just an idea on top of Panop question. For example: user know about several conflicts and wanna save his version and purge all other conflicted revs. Some server side support would be useful in this case. Well, omitting revisions in an update request is worst, indeed, thus my proposal becomes offtopic. I'll rethink and probably post separate feature request later. > > Max > > On Thu, Oct 13, 2011 at 12:26 AM, Alexander Uvarov < > [email protected]> wrote: > >> I would prefer a parameter like "force_update=true" and >> "delete_conflicts=true" to purge all existing versions and conflicts and >> just save new one. Would be useful for cases when user need to overwrite >> existing versions or when conflicts are resolved. Thoughts?
