Open jchris opened 11 years ago
Creating and destroying the bucket seems like a good solution. Any reason not to do that?
Good reason to assume our code owns the Couchbase Server is if we end up regularly targeting specific versions of it. Otherwise we'll just be running against whatever clusters happen to be provisioned.
Once this is closed we'll be able to avoid users discovering sync errors like this https://github.com/couchbaselabs/sync_gateway/issues/90
Duplicate or related to https://github.com/couchbase/sync_gateway/issues/353
We need a cluster we can test against, with test buckets set up to be flushable. Or to create a fresh cluster on every build?