[
https://issues.apache.org/jira/browse/QPID-6754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14998472#comment-14998472
]
Robbie Gemmell commented on QPID-6754:
--------------------------------------
Thanks for the updates Gordon, I tested them out (against the Proton 0.11.0 RC
patched using the fix for PROTON-1042) using the broker and JMS client. The
client picked up the advertised anonymous relay support and used it, with the
link attach response then containing a non-null target in the attach response,
and the message over the link being successfully delivered to the queue and
then consumed.
I did wonder about the 'queue patterns' option though and how it interacts with
anonymous senders, i.e links to an address in the queue pattern cause queue
creation, but anonymous messages to the same address don't. I think some folks
may find that unexpected. I can't see any specific docs for the queue-patterns
option (I think I only know about it from you mentioning it on a JIRA or
users@) other than the qpidd --help details, which say only "Pattern for
on-demand queues" for example.
The related 'queue policy' config (which I admit I haven't used) docs in the
AMQP_1.0 txt file do explicitly state it applies when attaching a link, so
there is a little less scope for confusion there, though it still seems like it
could happen.
Part of the reason I wonder about it is what happens when the queue isn't
there, with the message being accepted but dropped. That is fairly in keeping
with the earlier 0-10 model, but would doing a release/reject/modified/release
on the message be preferable to some? Its what Dispatch would do for example if
the address used didn't enable it to do anything with the message.
> Support anonymous relay
> -----------------------
>
> Key: QPID-6754
> URL: https://issues.apache.org/jira/browse/QPID-6754
> Project: Qpid
> Issue Type: Improvement
> Components: C++ Broker
> Reporter: Justin Ross
> Assignee: Gordon Sim
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]