WorkImpl may leak an InboundJmsResource

This is a companion to #57 and was found by code inspection whilst investigating the former issue.

com.sun.genericra.inbound.async.WorkImpl.run() contains a code path which neglects to release the ServerSession (= InboundJmsResource) instance back to its pool, viz. when an exception is thrown whilst releasing the message endpoint. This would result in leaking one resource from the pool, leaving it unusable until the application server is restarted. If the MaxPoolSize is 1, or otherwise if enough resources have been leaked in this way, messages from the queue associated with this pool would no longer be processed.

I have never actually seen this happen, and it is unlikely to go unnoticed when it happens, since said exception would be logged at level SEVERE. But it is easy to fix.
#### Environment
GlassFish (2.1.1 or 3.1.2), Linux (SLES10SP2, SLES11SP2), x86_64
#### Affected Versions
[0.9, 1.0a, 1.0, v1.5, 1.5, v1.6, 1.6, 2.1a, 2.1b, current, 2.2a]