Retry loops for bucket update ops have no backoff and no max retry count

The update operations seen in bucket_gocb.go use a naive for loop, with breaks only for unrecoverable errors, or successful updates.

• func (bucket CouchbaseBucketGoCB) Update
• func (bucket CouchbaseBucketGoCB) WriteUpdate
• func (bucket CouchbaseBucketGoCB) WriteUpdateWithXattr

In the case of network timeout errors, these for loops can retry forever with no backoff mechanism, which has the potential to make these network timeout issues even more likely to occur.