Nigel/all,
Thanks a lot for the update.
Although I have to use earlier JMS versions and currently we tend more
towards a synchronous sender/receiver pair, I also did some evaluations
with MDBs, and feel positive about
* Allowing multiple callback methods (though not everyone likes this)
E.g. one callback for "regular" messages, while others may receive messages
from a dead letter queue or similar queues/topics for errors and unusual
situations. The same MDB could then behave differently, but in a single
place to both regular and irregular messages.
Werner
On Fri, Jul 24, 2015 at 3:13 AM, <users-request_at_jms-spec.java.net> wrote:
> Table of contents:
>
> 1. [jms-spec users] Re: JMS_SPEC-116: Take advantage of EJB 3.2's RA
> improvement for MDBs - Nigel Deakin <nigel.deakin_at_oracle.com>
>
>
>
> ---------- Forwarded message ----------
> From: Nigel Deakin <nigel.deakin_at_oracle.com>
> To: users_at_jms-spec.java.net, jsr368-experts_at_jms-spec.java.net
> Cc:
> Date: Thu, 23 Jul 2015 13:02:29 +0100
> Subject: [jms-spec users] Re: JMS_SPEC-116: Take advantage of EJB 3.2's RA
> improvement for MDBs
> Thanks to those who commented on my proposals to improve JMS MDBs at
> https://java.net/projects/jms-spec/pages/JMSListener
>
> I've reviewed the comments and produced a second version, which can be
> seen at
> https://java.net/projects/jms-spec/pages/JMSListener2
>
> This incorporates the main suggestions made so far, notably
>
> * Allowing multiple callback methods (though not everyone likes this)
> * Requiring the return type to be void
> * Allowing these new anotations to be used with the onMessage method of a
> MessageListener
> * The earlier proposal to set a parameter to null if it cannot be set
> because its type is incompatible has been dropped. Instead, if this
> happens, then the message will not be delivered. (do we need to say
> anything else?)
> * New proposals on callback methods that throw exceptions (are these good
> enough?)
> * Dropping the ability to specify the non-portable queue or topic name (at
> least for now)
>
> A full list of changes is given on that wiki page at
>
> https://java.net/projects/jms-spec/pages/JMSListener2#Changes_from_version_1
>
> These changes raise new issues, and there are lots of existing issues
> still to work on, but I think this new version is a better basis for
> continued review.
>
> Meanwhile I have successfully implemented almost all of this in GlassFish
> and MQ, and continue to be pleased at how straightforward this is, with
> almost all the changes in the resource adapter. I am still looking at how
> to share this prototype, and the tests I have written, with the community
> (we really need a new version of GlassFish to be launched).
>
> Although I'm grateful for the feedback so far, I have had very few
> comments overall, even from members of the expert group. I appreciate
> everyone has other work to do, so is there anything I can do to make it
> easier for people to participate? One possibility would be for me to hold a
> conference call where we could chat about this. Would that help?
>
> In any case, please take a look at the updated proposals and let me know
> your comments.
>
> Thanks,
>
> Nigel
>
>
>
> End of digest for list users_at_jms-spec.java.net - Fri, 24 Jul 2015
>
>