Thanks for your reply. I think I found the issue - shortly before the message that failed to be sent through was another message that had a null body (and was much smaller in the tcp dump so I missed it).
It seems to be a problem with any TextMessage that has a null body: empty string is ok. So the workaround is simple (for us): if the body is null, set the body to be empty string. If you consider this a bug I'd be happy to write a junit test for it. -- David On 25/02/2012, at 1:03 AM, tabish...@gmail.com [via ActiveMQ] wrote: > On Thu, 2012-02-23 at 19:38 -0800, davidw wrote: > > > OH - here's an example of what gets printed to the logs when it works > > > > > > > > and the message that seems to break it > > > > > > > > -- > > View this message in context: > > http://activemq.2283324.n4.nabble.com/JMS-goes-in-sometimes-STOMP-comes-out-but-not-always-tp4416098p4416114.html > > Sent from the ActiveMQ - User mailing list archive at Nabble.com. > > Your best bet here is to try and create a unit test that reproduces the > problem, there's nothing obvious in the code you've shown so far. Could > have something to do with the Stomp subscribers though. There are > several Stomp tests in the AMQ source you can work from if you want to > try and come up with a test case. > > > -- > Tim Bish > Sr Software Engineer | FuseSource Corp > [hidden email] | www.fusesource.com > skype: tabish121 | twitter: @tabish121 > blog: http://timbish.blogspot.com/ > > > > If you reply to this email, your message will be added to the discussion > below: > http://activemq.2283324.n4.nabble.com/JMS-goes-in-sometimes-STOMP-comes-out-but-not-always-tp4416098p4417270.html > To unsubscribe from JMS goes in, sometimes STOMP comes out (but not always), > click here. > NAML -- View this message in context: http://activemq.2283324.n4.nabble.com/JMS-goes-in-sometimes-STOMP-comes-out-but-not-always-tp4416098p4423787.html Sent from the ActiveMQ - User mailing list archive at Nabble.com.