I have identified what I consider a valid bug against Hibernate v.3.5.0 and posted the details of that bug on JIRA. To help with the re-production of the bug, I've provided a test case which clearly illustrates the bug and uploaded it along with the bug description.
Today, I got a notification from JIRA that the bug has been resolved as "Rejected". The only comment posted along with that rejection is "Please use the user forum (
https://forum.hibernate.org/) for help."
The bug is related to the fact that if one uses the new JPA 2.0 criteria API to create a query where one of the predicates in the where clause compares a numeric field (int or long) with a certain value, the generated query does not use a bind variable for that value, but rather hard-codes the value in the query, which can significantly impact the performance on Oracle databases. This can also have an impact on security in relation to SQL injection, although I can not verify that since I have not tested it or looked into the source code.
For me, this bug is a show-stopper and I would not switch to JPA 2.0 criteria until it's fixed. I would appreciate if someone from the Hibernate development team could highlight why such a bug has been rejected.