On Tue, Jan 29, 2013 at 1:41 AM, Murali Reddy <murali.re...@citrix.com> wrote: > >>Murali, >> >>The use of Rabbit is still a question for me. It seems like you went >>with Rabbit, but the answer you gave as to "why" [1] didn't really >>answer the question or respond to the issue I raised about the >>practicalities of AMQP differences between versions and brokers [2]. >>Can you address these please? >> >>-chip > > Chip, > > First, there is NO hard dependency on the RabbitMQ AMQP server. My > intention was just to use a AMQP client which can work with any other AMQP > broker. When I looked at AMQP java clients, I found rabbitMQ client API > was more simple and closer to the AMQP protocol and did not have > additional dependencies (like in case of Spring-AMQP, JMS in Qpid Client) > and had active community. That made me to choose RabbitMQ client. > > I know you raised concern that RabbitQM AMQP client may not necessarily > compatible with QPID, also on points raised by David [2], I donĀ¹t have > knowledge of nuances of availability in distro's and AMQP version and > broker incompatibility. But what I have done is to keep things pluggable. > RabbitMQ client based plug-in I wrote can be treated as reference > implementation if does not fit their needs and they can write their own > plug-in (for JMS or other messaging protocols) implementing the event bus. > > From the framework design perspective I have kept it configurable and > pluggable and there is no enforcement on any particular broker or client. > Please let me know if there is still a concern.
I think we're good. I raised the point specifically because the decision would have been best discussed on this list prior to implementation. That being said, I don't have a technical concern (given how it's implemented). IMO we're good on this front. > > Thanks, > Murali > >> >>[1] http://markmail.org/message/smzzk5sjflfhj7j7 >>[2] http://markmail.org/message/k2unpaz5g6w54f7o >> > > >