There is a small inconsistency in the hardcoded JNDI name for JMS queues
when error messages are output.  I had an incorrect jndi name on one of my
MDB's and the error message I got was that "QUEUE.myQueue" was not found
instead of "queue/myQueue". 

A bit puzzled by the "QUEUE." name, I searched the Branch_3_2 code base and
found that the SpyQueue.java file in the org.jboss.mq package has the
following code in the constructor:

   public SpyQueue( String queueName ) {
      super( queueName );
      toStringStr = "QUEUE." + name;
      hash++;
   }


That toStringStr shows up in JMSException message when a queue is not found
in JNDI space by the JMSContainerInvoker.  I noticed that in many of the
other JMS related source files, developers hardcoded "queue/" not "QUEUE.".
I know it's a small thing, but for the sake of consistency and reduction of
confusion, shouldn't this be modified to be "queue/", not "QUEUE."?

I am also very suprised that in many places in the code, "queue/" +
queuename is hardcoded for JNDI lookups (The JMSContainerInvoker has an
example of this).  I find that quite odd that JBoss would enforce that
naming standard for JMS queues' JNDI names when they should be able to be
bound to any JNDI name without having to be forced into a having "queue/"
being prepended on the JNDI name whenever it is internally reference in
JBoss.

Dustin Barlow


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Jboss-development mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-development

Reply via email to