My bad, the jndi-name in jboss:resourceGroup is the good one.
Leaves only the question whether jndi-name deprecates mapped-name or should be treated as a separate element?
I'm working out an alternative assertJndiName where a mode argument can be passed. This way we can still make sure that a jndi-name overrides a mapped-name.
Bugrit, in the end it still gets to convoluted, so I'm not going to override your fix. I've committed the mode construct though, it might be useful in the future.