|
From: Colin S. <col...@ex...> - 2003-10-17 19:13:23
|
The new Hibernate stuff is working ok for me (with JtaTransactionManager). jürgen höller [werk3AT] wrote: >I've already fixed the Hibernate one: It's what the discussion on commit/rollback was about, i.e. a flushing failure on Hibernate commit that left the database transaction in an unfinished state. It could be considered a Hibernate peculiarity, but fortunately it's quite easy to work around it by triggering an explict flush before commit (and turning the Hibernate session FlushMode.NEVER afterwards to avoid an unnecessary flush attempt at actual commit). > >I've generally revised the flushing behavior. There was a situation when combining HibernateTemplate and JtaTransactionManager that would not have flushed the latest changes if the user code did not trigger flush itself. It does flush now at JTA transaction commit too, via a new beforeCommit callback in the TransactionSynchronization interface. That took me quite some time to analyze and solve; I hope it behaves correct in all situations now. > >Please, everybody working with Hibernate either with HibernateTransactionManager and JtaTransactionManager, give your apps a retry with the current codebase. I'd even wait with the release till early next week to be able to double-check the current behavior. > >Regarding the com.interface21 definition in PetClinic: There is no such reference in the current codebase, and there was none in the 1.0 M1 snapshot. I guess the reporter of that bug had an old Petclinic lying around and tried to drop in the new org.springframework jar files. > >Juergen > > > > -----Ursprüngliche Nachricht----- > Von: Rod Johnson [mailto:rod...@in...] > Gesendet: Fr 17.10.2003 10:29 > An: spr...@li... > Cc: > Betreff: [Springframework-developer] Bugs > > > > Guys, > > I've added bug categories in the bug tracker and assignees. > > We currently have 4 open bugs. The MDB one is low priority, debatable and > doesn't break anyone's code. Assigned to me. > > The prototype/singleton bean factory one shouldn't be a showstopper for > anyone, and I'll try to look at it next week. I've assigned it to me. > > Can someone please try to resolve the PetClinic one and the Hibernate one? > > Regards, > > |