There is no void onMessage(javax.jms.Message m) signature in any jms spec version so there has to be some bytecode manipulation to see this type of error.
Ain't no one here doing any bytecode manipulation sir.
Found the cause of the error, but not sure why it's complaining.
When I remove code in onMessage that accesses a HashMap, it works fine. The HashMap is within the class that implements the MessageListener.
Any clue why it would be complaining?
Moderated: What has not recompiling your classes correctly when you make changes
got to do with JMS?
I'm using eclipse, so you're saying eclipse is not compiling my classes correctly?
I'm using JDK 5.0.... All I know is when I start Jboss, if I have code that accesses a HashMap in my onMessage() implementing method of MessageListener, I get that VerifyError. When I take the code that does gets or puts on the HashMap (but leave the declaration and the new HashMap()), I don't get a VerifyError.
I don't see how I could be recompiling my classes incorrectly, I don't and have never seen this anywhere else using eclipse.
While your post "what has recompiling my classes incorrectly" was cute, it was very unhelpful. I POSTED a REPLY to see if doing something with a Map inside an onMessage was a bad thing..
So to answer your question, I have no clue what it has to do with JMS, but I only see this error, have only seen this error, with regard to my onMessage(javax.jms.Message) method, so I thought it might be somehow related to JMS.
Post a bug report with an example so someone can look at.
Include the full jdk info and javac options used to compile. This simply looks to be a compiler bug, nothing to do with jboss.