Peter, >>So the block listener developer instantiates and throws back? I'd go for : >> >>ApplicationStartupException extended by VetoApplicationStartupException >>(to please Stephen?). >> > >Im not sure - seems like extra complexity that we don't need? Is there any >going to be more than 1 type of exception? > OK. I am fond of trees of Exceptions. However this is not so much an application as a framework, one will do.
>>>Not sure what do you think? >>> >>For the sake of backCompat, I;d say ApplicationListener extends >>BlockListener then. >> > >ok. Fine by me. It feels a bit icky technically but right because it is >backwards compatible. > I'm quite happy with interface extension. If the JDBC team had done it for JDBC3 then AvalonDB would not have been so screwed trying to be fully compatible..... We'd have assembly XML segments ... <application-listener class="class" name="name" /> .... in additon to .... <block-listener class="class" name="name" /> ... to declare which is being used. >I suppose in the future we could add an >AbstractBlockListener and gradually migrate them to one interface (at same >time we go from alpha to beta). > Perhaps, but it could stay as is. I'd prefer to get it right now, despite pain rather than choose to migrate later. However, I'll never force a policy for Phoenix development :-) Regards, - Paul H -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
