[rabbitmq-discuss] 2.4.1 features
simon at rabbitmq.com
Mon May 23 11:46:27 BST 2011
On 21/05/11 12:14, Steven Taylor wrote:
> just upgrading from 2.2 to 2.4.1. Would noAck typically be set to
> "false" in prior versions for BasicConsume?
The overridden versions of BasicConsume that don't take noAck do set it
> BasicConsume(string queue, bool noAck, string consumerTag,
> IBasicConsumer consumer)
> Also, seems like you've snuck in some nice features into the patterns
> area like EventingBasicConsumer (thanks Simone for your example on
> this... otherwise I wouldn't have known about it). How are you Rabbit
> folk currently making announcements about new features?
The release notes are posted to the rabbitmq-announce mailing list:
and are available archived from our news page:
Highlights for each release are summarised at the changelog:
> "The consumer tag is local to a channel" -- so we're talking about
> identifying one consumer, of many listenning in to a single channel?
Yes, to allow you to:
* Know which messages arrived for which consumer.
* Cancel an individual consumer.
NB: the consumer tag always exists, the server will; generate one for
you if you don't provide one.
> If we've got a field/tag for a particular consumer (even one on the same
> channel), perhaps it'd be nice to have a publisher-tag as well. i.e.
> issolate a particular publisher for tracing purposes. reply-to merely
> distinguishes the queue.
The trouble is that AMQP 0-9-1 doesn't require you to declare a
"publisher" in the same way that you do a consumer (maybe it should) -
so it's not clear where this would live.
> Maybe I could use "IDictionary arguments" could be used for this
> instead? I've ignored "arguments" up until now.
If you want to identify a particular publisher, the canonical way to do
that is with the "app-id" property in BasicProperties.
Staff Engineer, RabbitMQ
SpringSource, a division of VMware
More information about the rabbitmq-discuss