Berkeley DB Family

Hello,

I'm implementing a simple datastore on top of Berkley DB. I'm going to add a cursor utility but I'm not sure about implications of using a cursor (even in RO mode) while reading/writing with simple gets and puts on the same db.

Suppose I opened a cursor on a db and I'm doing some aggregation (RO mode) over its elements while other threads are writing on the same db. What are the implications of this scenario ? Have I to implement some tx in there ???

Thank you for the help,
Ugo.