I have filed the bug in the bugtrackere (#3974 and #3987). Based on feedback from Nico it has been resolved and the fix has been commited to CVS.
In the original code, certain conditions could lead to connections being closed while a concurrent thread was preparing it to execute a statement. As a result both threads can get caught in a deadlock (at least on Oracle). This situation could arise when the threads waiting for a database connection vastly outnumber the available connections. Rob van Maris Developer Finalist IT Group Java Specialists ------------------------------------------------------------- Amsterdam, The Netherlands Office: +31 20 5962321 (Direct) Mobile: +31 651444006 Fax: +31 20 5962331 -------------------------------------------------------------
