[
https://issues.apache.org/jira/browse/WSCOMMONS-468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12758360#action_12758360
]
Andreas Veithen commented on WSCOMMONS-468:
-------------------------------------------
Removing the code entirely would obviously solve the inconsistency :-)
However, the fact that the feature is implemented using a hack doesn't mean
that it is not useful. It allows the user (for some JMS providers) to expose
services over JMS without the need to explicitly create and/or map services to
destinations. The issue is that this should be done in a proper and consistent
way.
> The JMS transport doesn't use the ActiveMQ hack consistently
> ------------------------------------------------------------
>
> Key: WSCOMMONS-468
> URL: https://issues.apache.org/jira/browse/WSCOMMONS-468
> Project: WS-Commons
> Issue Type: Bug
> Components: Transport
> Reporter: Andreas Veithen
> Priority: Minor
> Fix For: Transports 1.0
>
>
> JMSOutTransportInfo#getDestination contains an ActiveMQ specific hack that
> prefixes the destination JNDI name with "dynamicQueues/" or "dynamicTopics/"
> if the initial lookup fails. On the other hand,
> JMSConnectionFactory#getDestination doesn't apply this hack. This means that
> the behavior in the JMSSender is inconsistent and depends on whether an
> existing connection factory definition is found in axis2.xml or not.
> In addition, ServiceTaskManager#getDestination uses yet another approach,
> namely to get the Destination object from Session#createQueue or
> Session#createTopic.
> Note that I believe that adding provider specific hacks is bad and I'm -1 for
> a solution that would simply add the hack to
> JMSConnectionFactory#getDestination.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
