[
https://issues.apache.org/jira/browse/QPIDJMS-530?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17311453#comment-17311453
]
Robbie Gemmell commented on QPIDJMS-530:
----------------------------------------
There doesnt seem to be any way this change could cause that to me, as it only
comes into play at all during a remote consumer closure, and that particular
test has no consumers.
There dont look to have been any changes since 0.56.0 that I think should cause
any difference to that test either though.
The actual main objective of the test has actually passed by that point (that
commit threw), the bit that is failing is a secondary element, essentially
checking for giggles that an internal listener got told the connection ws done
as failover had failed as expected, as it was set up to fail. Thats the only
failure route I can see - something else on your box takes the port and
prevents the failover attempts being exhausted within the expected time.
> Potential NPE in JmsSession consumerClosed()
> --------------------------------------------
>
> Key: QPIDJMS-530
> URL: https://issues.apache.org/jira/browse/QPIDJMS-530
> Project: Qpid JMS
> Issue Type: Bug
> Affects Versions: 0.56.0
> Reporter: David Martin
> Assignee: Timothy A. Bish
> Priority: Major
> Fix For: 0.58.0
>
>
> In the following code (JMSSession line 416), if consumer is null this throws
> a NPE and the consumer never recovers:
> {noformat}
> JmsMessageConsumer consumer = consumers.get(resource.getId());
> if (consumer.hasMessageListener()) {
> connection.onAsyncException(JmsExceptionSupport.create(cause));
> }
> try {
> if (consumer != null) {
> consumer.shutdown(cause);
> }
> {noformat}
> Happy to raise a PR myself for this (just ask) but I have had to fix it
> locally.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
