On 26/06/2011 01:05, Phil Steitz wrote: > On 6/25/11 4:28 PM, Mark Thomas wrote: >> On 17/06/2011 09:02, Mark Thomas wrote: >>> On 17/06/2011 00:32, Gary Gregory wrote: >>>> I think 2.0 is the opportunity to do this right. Almost like we were >>>> designing this from scratch. >>>> >>>> Making the factory an invariant of the pool sounds good. >>>> >>>> Otoh If a setFactory method exists it should be implemented fully. The >>>> throw an exception impl is pretty "smelly". >>> I agree that it is smelly. I'm not sure we can change this without a >>> huge amount of work in DBCP. >> I think largely due to the clean-up Phil has already done in DBCP, this >> is now manageable. I have just done this for GOP and will do the same >> for GKOP next (exact timing TBD). > > +1 - definitely better. Will test and review and if I get to if > before you, handle GKOP. Thanks for doing this.
Looking at this, I can't see a way of doing this without introducing a new interface StatementPoolFactory that extends KeyedObjectPoolFactory and adds a setPoolingConnection() method (or something along those lines anyway). The names may well change as the refactoring evolves and we add generics support to DBCP but this should do for now. Mark --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
