I can explain this one easily enough, Glyn. JAX-RPC's javadocs and
jaxrpc-APIs.jar show QName as a final class. So we made it final in our
version.
We can always fight this if you think it's appropriate. The JAX-RPC spec
itself doesn't say anything about QName. Probably because QName has been
defined in other locations by other specs and there's a stirring of people
to coalesce all these disparate QNames into one. I wonder whether the
other QNames are final?
Russell Butek
[EMAIL PROTECTED]
Glyn Normington/UK/IBM@IBMGB on 04/11/2002 04:48:26 AM
Please respond to [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
cc:
Subject: Re: cvs commit: xml-axis/java/src/org/apache/axis/message
PrefixedQName.java
Rich,
Although I can speculate about why you changed PrefixedQName so that it
delegates to QName instead of extending it, I'd be interested in *your*
rationale.
My rationale for using extension was that there could easily be cases where
it would be necessary to use a PrefixedQName where a QName was required,
e.g. as a parameter to a method. I applied the "is a" test and decided that
PrefixedQName is a QName.
I wouldn't mind your change if I understood and agreed with the rationale.
Thanks,
Glyn
