STOMP ACK error messages

[mbunzel]

When using the STOMP protocol, a completely unexpected ACK message causes a NullPointerException to be returned. The minimal test case consists of  connecting (with our without client-id header), followed by sending an ACK message with an arbitrary message-id. The returned ERROR message looks like this:

ERROR
message:null
content-length:1636

java.lang.NullPointerException
        at org.hornetq.core.protocol.stomp.StompSession.acknowledge(StompSession.java:174)
        at org.hornetq.core.protocol.stomp.StompProtocolManager.onAck(StompProtocolManager.java:402)
        at org.hornetq.core.protocol.stomp.StompProtocolManager.handleBuffer(StompProtocolManager.java:203)
        at org.hornetq.core.protocol.stomp.StompConnection.bufferReceived(StompConnection.java:269)
        at org.hornetq.core.remoting.server.impl.RemotingServiceImpl$DelegatingBufferHandler.bufferReceived(RemotingServiceImpl.java:459)
[...]

Is this an expected behavior or should I open a JIRA issue?

Another thing that I find curious is that when a large number of unexpected ACKs are sent, the exception (STOMP ERROR frame and logged to stdout) is shortened to:

ERROR
message:null
content-length:31

java.lang.NullPointerException

This appears to persist until HornetQ is restarted. Is this an intentional behavior to prevent log spamming? Is there any way I can turn it off during dev/testing?

Many thanks for such a great product,

Mike

[ataylor]

Have you tried with the latest version, if so, can you provide a test and open a jira

[mbunzel]

Sorry, I forgot to mention that I can reliably reproduce this on 2.2.2 as well as 2.2.5. I'll go ahead and open a jira. Should I also mention the shortened NPE or is intentional?

[ataylor]

yes, and if you could add a test that would be helpful

[mbunzel]

I've created HORNETQ-727 and included a test script. Please let me know if you need further information.