You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by "Robbie Gemmell (JIRA)" <ji...@apache.org> on 2015/11/10 15:26:10 UTC
[jira] [Resolved] (DISPATCH-192) a null target is used in attach
response, without subsequent detach, following attempted use of anonymous
relay node
[ https://issues.apache.org/jira/browse/DISPATCH-192?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Robbie Gemmell resolved DISPATCH-192.
-------------------------------------
Resolution: Done
Fix Version/s: (was: 0.6)
PROTON 1042 is now merged to the 0.11.x branch for inclusion in the second 0.11.0 RC, so using 0.11.0 once released should resolve this behaviour.
Since there weren't any changes needed in Dispatch itself I am resolving this JIRA without a fix version
> a null target is used in attach response, without subsequent detach, following attempted use of anonymous relay node
> --------------------------------------------------------------------------------------------------------------------
>
> Key: DISPATCH-192
> URL: https://issues.apache.org/jira/browse/DISPATCH-192
> Project: Qpid Dispatch
> Issue Type: Task
> Reporter: Robbie Gemmell
>
> As Dispatch advertises the ANONYMOUS-RELAY capability, signalling support for the anonymous relay node, when creating an anonymous JMS MessageProducer (using a null Destination) against it the JMS client sends a link attach with a non-null target representing a null address (i.e the anonymous relay node). Dispatch currently responds to this with an attach containing a null target field. The AMQP spec indicates that a null target is used in the attach response when the link is being refused, with a Detach frame then following to close the link and potentially give a cause. Dispatch doesnt send a Detach in this case and its unclear that it actually intends to refuse the link in this case (since it advertises ANONYMOUS-RELAY), but clients would have to assume it did (and may get themselves into a state when the detach doesnt follow).
> In this case I would expect the attach response to contain a non-null target with null address, but failing that a Detach frame should be sent after using the null target in the attach response.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
For additional commands, e-mail: dev-help@qpid.apache.org