sure. First though, McGill is inferior to Queens in all ways ;)
The PersistenceBrokerBean is meant to be deployed into an application server to service requests from remote clients. As part of that setup, the dynamic proxy would be configured to check for the existence of a valid transaction context and use that, or if one doesn't exist, lookup the persistenceBrokerBean and interact with that. Once you have looked up the bean, it implements the "persistencebroker" interface so interactions are the same. in order for updates to work, you need to be using optimistic transactions, since we won't lock across tx. Not very complex, really, just haven't had time to finish it. The same approach could be used for a RMI based persistenceBroker. I did it with EJB, because that's what we use. m -----Original Message----- From: Phil Warrick [mailto:[EMAIL PROTECTED]] Sent: Friday, January 31, 2003 12:02 PM To: OJB Users List Subject: Re: RMI and Proxied collections and references Matthew, Matthew Baird wrote: > one problem will be that if you are running invm tomcat/jboss and your >proxy materialization happens AFTER the transaction is over, you'll need >a mechanism to know that the tx is over, because you were never serialized. > > I've thought a lot about this problem, and started to implement the > persistencebrokerbean to solve it, then got sidetracked. > Can you sketch out your approach with the persistencebrokerbean? Phil --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
