[Opensource] Upgrading from 4.02 to 5.02: DBObject question

[Vincent Gogan]

Has there been a way that DBObject's cache? I ran into the following
difference
in behaviour.

I seem to have a DBObject that contains stale information.

In Thread 1:

	DBSomeThing s = new DBSomeThing();
	s.setField(FIELD_ID, n);
	if (!s.find()) blowUp();

	display(s.getField(FIELD_INFO));

	performALengthyTask();

	display(s.getField(FIELD_INFO));

Meanwhile, in another thread, executed while the first thread is performing
the lengthy task,

	DBSomething s = new DBSomeThing();
	s.setField(FIELD_ID, same_index);
	if (!s.find()) blowUp());

	s.setField(FIELD_INFO, new_value);

In 4.02, this code worked just fine and the first thread picked up the
change. In 5.02,
this breaks and the first thread can only acces the old/stale value. One
possible
contributing factor is that thread 1 may be using a derived class:

public class DBExtendedSomeThing extends DBSomeThing {
	// Just overrides/adds some methods
}

and be actually using DBExtendedSomeThing in the above code.

Thanks,

Vincent