"Optimistic locking" or "how to catch org.hibernate.StaleOb

[bigm25]

Is there a way to know when the EntityManager is internally "discarding" a merge, so we could show the user that his last action failed and he should reload???

Right now there's just an exception trace getting logged, but we can't catch anything while doing "entityManager.merge(...)"

[bigm25]

Sorry, subject was truncated, so i'm posting the question again, with some clarifications:

Is there a way to know when the EntityManager is internally "discarding" a merge (due to a version conflict), so we could show the user that his last action failed and he should reload???

Right now there's just an exception trace getting logged, but catched internally (not in our code) so we can't catch anything while doing "entityManager.merge(...).

[epbernard]

If the exception is really catched internally, open a bug in JIRA please.

[pomeloverde]

If you are trying to catch a Hibernate exception and you have failed, remember to flush() the Session inside your method:

myMethod() {
try {
...
Session.saveorUpdate(...);
Session.flush();
} catch (StaleObjectException ex) {
...
}
}

If you don't make a explicit call to flush(), then Hibernate will flush the Session automatically, but after your method call ends, so you will cannot trap the exceptions. Normally this is the case when you use Hibernate from the EntityManager.

Best regards