Because when programming beans with bean-managed transaction demarcation you have two ways of getting hold of a UserTransaction object: 1. By calling getUserTransaction() on the context object. 2. By looking up in JNDI the UserTransaction object mapped to the java:comp/UserTransaction
My Linked In
posted 15 years ago
Oh I see.
I was not aware of the 2nd way of getting the UserTransaction object.
This makes me wonder why must there be 2 ways... kind of redundant.
Is this address (java:comp/UserTransaction) in the JNDI fixed ?