Hi Gordon, Yes, you're right - jmsSession.rollback() does result in an AMQP message.release command.
Phil On 8 November 2012 08:44, Gordon Sim <gordon.r....@gmail.com> wrote: > On 11/08/2012 08:30 AM, Phil Harvey wrote: > >> The delivery count is incremented each time the message is rejected by the >> client, eg when it explicitly calls session.rollback() after your >> application code decides that its format is wrong. >> > > Does that actually reject (in 0-10 parlance) or just release the > message(s) (from my cursory examination of the code it seemed like it was > the latter). > > Of course, if the broker provides the means to handle that as desired, the > difference need not be relevant. However if the dead-lettering is based on > delivery count and the message can only be delivered once, then is it not > the case that a failure might also result in dead-lettering of in-doubt > messages? > > > ------------------------------**------------------------------**--------- > To unsubscribe, e-mail: > users-unsubscribe@qpid.apache.**org<users-unsubscr...@qpid.apache.org> > For additional commands, e-mail: users-h...@qpid.apache.org > >