Changelog

Adjustments for hight availability and cluster failover:

Allow non-durable queues

Not replicated durable queues, if their node dies, can not be recreated
on another node until their owner node is back or removed from the
cluster.
This change allows to declare worker as worker with non-durable queues

self.durable_queues = false

Consumers will try automatically to recreate the queue and re-subscribe if they are not durable

Verbose logging

You can now inject a logger object. We've pretty verbose logging enabled to help debugging.

no-rspec support

We're checking whether you're using rspec, and avoid requiring it if absent.

restructured code

Restructured the library into smaller modules.