Another inconsistence in the spec 21.13 Some methods require a non-null state manager. In these cases, if the jdoStateManager is null, then IllegalStateException is thrown.
e.g. 21.20 jdoReplaceField If there is no StateManager (the jdoStateManager field is null), then this method throws JDOFatalInternalException. Regards, Quoting Craig L Russell <[EMAIL PROTECTED]>: > Hi Erik, > > Sorry, you are right. The spec is inconsistent. I agree with you that > 21.20 is right and 7.5 is wrong. This will go on the "errata" list > for the next spec update. > > Craig > > On Mar 6, 2006, at 10:48 AM, Erik Bengtson wrote: > > > Craig, > > > > I thought it was a correction rather than an item to discuss. See > > spec 21.20 > > > > Quoting Craig L Russell <[EMAIL PROTECTED]>: > > > >> Hi Erik, > >> > >> If the spec were still open, I'd agree to discuss it. But it > >> isn't. :- > >> ( We can look at this for a future revision. > >> > >> By the way, the rationale was that it's less work in the enhancer to > >> generate the code, but I admit the rationale is weak. But either way, > >> it's a matter of the implementation checking for null or catching a > >> NullPointerException versus catching a JDOFatalInternalException. I'd > >> guess it's a "startup cost". > >> > >> Craig > >> > >> On Mar 6, 2006, at 7:28 AM, Erik Bengtson wrote: > >> > >>> > >>> 7.5 > >>> "If the class is abstract, null is returned." > >>> > >>> but should be "If the class is abstract, a > >>> JDOFatalInternalException is thrown." > >>> > >> > >> Craig Russell > >> Architect, Sun Java Enterprise System http://java.sun.com/products/ > >> jdo > >> 408 276-5638 mailto:[EMAIL PROTECTED] > >> P.S. A good JDO? O, Gasp! > >> > >> > > > > > > > > Craig Russell > Architect, Sun Java Enterprise System http://java.sun.com/products/jdo > 408 276-5638 mailto:[EMAIL PROTECTED] > P.S. A good JDO? O, Gasp! > >
