On Tue, 6 Nov 2007, Patrick Linskey wrote: > > If I use the javaagent as you described, the test passes. If I omit the > > javaagent entirely, the test fails (because the m:m field is null). > > Cool, thanks for testing that out. One more question: is this > happening with a new instance (i.e., one that you call persist() on), > or on an instance that you've looked up from the DB?
This is happening on an instance that I've looked up from the DB. I would expect the field to be null on a new object that had not yet been persisted, unless the collection were already initialized in the declaration. One thing that's also interesting is that if I define the field so that it's initialized in the declaration (e.g. Set<Bar> foo = new HashSet<Bar>()), it still ends up being null when retrieved from the database... so it would appear that it is being explicitly set to a null value at some point. (Though I don't discount the possibility that something else could be happening as well.) Nick -- "Courage isn't just a matter of not being frightened, you know. It's being afraid and doing what you have to do anyway." Doctor Who - Planet of the Daleks This message has been brought to you by Nick Johnson 2.3b1 and the number 6. http://healerNick.com/ http://morons.org/ http://spatula.net/
