I've checked in a version of the view-group-state-server work that uses gen_server and passes all but the purge test. Thanks Paul Davis for helping with the error handling stuff.
Here's the diff against trunk: http://friendpaste.com/w5Xks5Kd And here is the github branch: http://github.com/jchris/couchdb/tree/update-false Damien, I think this is ready for you to look at. It might be easiest to make purge work by crashing couchdb_view_updater when it encounters a purge that's too deep for it to repair. I think this is sort of what it's doing already. -- Chris Anderson http://jchris.mfdz.com
