Hi Brian, The test case attached with OPENJPA-911 (you got the important JIRA number:) runs on trunk. I did not see any assert() commented out -- so am assuming that it meets expected behavior. OpenJPA issued 4 SQL to fetch the entire graph before and after delete. If it possible for you, run it against the latest nightly builds. If the test was failing with an earlier nightly build version, I may be responsible. Some changes I have made to reuse SQL for repeated JPQL execution *may* have caused it as your test does execute the same JPQL twice. With the recent changes, the second execution of JPQL will bypass the major code path and directly execute cached SQL.
Brian Dellert wrote: > > Just an FYI: I have created a Jira bug to hopefully keep this issue from > slipping through the cracks. The URL for the bug entry is: > > https://issues.apache.org/jira/browse/OPENJPA-911 > > And the entry contains the example that was also provided in this thread. > Thanks. > > -- View this message in context: http://n2.nabble.com/Issue-with-FetchAttribute-recursionDepth-tp2260751p2312511.html Sent from the OpenJPA Users mailing list archive at Nabble.com.
