…ry and JTATransactionFactory
… I was doing more than I needed to. Updated the LockTest to enable another nowait test for postgresql.
… thrown. This feels sooo wrong. I know dialects aren't mapping lock exceptions properly, and this test case exposes that if you fail the test when PersistenceException is thrown. So what do I do now? There are other tests in this very test class that don't fail the build when the wrong exception is thrown, even though they clearly should.
…tly fails for PostgreSQL matrix tests, unless HHH-7251 is also applied.
The code was a bit non-obvious (assignment in a method call argument) where it was being done, and the local variable (lockOptions) assignment was missing in the find() method implementation. I changed all instances of lockOptions assignment to assign before the method call where it's required as an argument, ensuring proper scope for exception mapping.
…or Deadlock / lock timeout (not available with NOWAIT query).
consider objects loaded from shared cache)
…Impl.notifySynchronizationsAfterTransactionCompletion due to SynchronizationRegistryImpl.clearSynchronizations clearing SynchronizationRegistryImpl.synchronizations
…ers on close
…observer on Session close and added some checks to ensure the session is still open when the listeners fire.
Adding unit tests for the transaction coordinator when using shared transaction context. One test to show that Sessions are leaked to the transaction observer, one to show that afterTransactionCompletion is not called on any session other than the main and finally one to show that the original session cannot be reused if child sessions were autoclose or flushBeforeCompletion.