Hi everyone

I'd like to propose an small adjustment to the way ruote-amqp works before I
push, package and release a new ruote-amqp version (2.2.1).

Currently whenever we register an AMQP participant, we implicitly setup the
AMQP stack by calling RuoteAMQP.start! inside the initializer. This used to
be fine and mostly suitable until I had the realization that we can register
participants in one process (engine-only setup) and only consume work items
in other processes (engine+storage setup).

In my specific case I register participants with ruote-kit in a passenger
stack, and run workers outside of the passenger stack with rake tasks. I'd
like to change the ParticipantProxy and Receiver classes to NOT call
RuoteAMQP.start!, and leave that up to the user. The user can then decide
where a realm AMQP connection is necessary or not, and probably handle AMQP
errors better. The connection is only required on consumption of a workitem,
or receiving replies from remote workers.

I'm testing this locally, and if there are no solid objections I'll push,
package and release the update on Friday.

Kind regards

-- 
Kenneth Kalmer
[email protected]
http://opensourcery.co.za
@kennethkalmer

-- 
you received this message because you are subscribed to the "ruote users" group.
to post : send email to [email protected]
to unsubscribe : send email to [email protected]
more options : http://groups.google.com/group/openwferu-users?hl=en

Reply via email to