i believe the app is working correctly -- even with the reported exceptions; i have obtained a debug version of derby/javadb and the sources and have stepped thru the code where the earliest client side indication of the "problem" occurs -- this code is outside of my EJB code directly -- that stack trace is a thread started/controlled by gfv2ur2 -- making use of the CMP tree; i have not tried to debug the javadb server code at this time.
the highest GF client side stack seems to go to Sqlca.getUpdateCount() in org.apache.derby.client.am directory; it return an int - value 0 -- which ends up being returned to the GF source proper in SQLStoreManager.executeUpdate where it reads,
affectedRows = s.executeUpdate(); and the returned value of int 0 is translated into the error that eventually gets returned as a concurrent access exception.
at this point -- i guess i'm convinced the Sun blueprint code is working -- but i still dont fully understand why the error occurs even if it eventually recovers???
thanks
[Message sent by forum member 'emiddio' (emiddio)]
http://forums.java.net/jive/thread.jspa?messageID=274724