Jukka Zitting wrote:
> Hi,
> 
> On Mon, Jul 14, 2008 at 10:37 AM, Carsten Ziegeler <cziegeler@apache.org> wrote:
>> what is now the prefered way? Extending the event interface or
>> providing a marker interface for the event listener?
> 
> Whoever writes the patch gets to decide. :-)
> 
Hehe, right!

> The reason I prefer marking the event listener is that it simplifies
> the client code a bit and potentially (not currently, but perhaps in
> future) allows the repository to optimize event delivery as it has
> better knowledge of what events a listener is interested in. But I
> wouldn't object to the alternative if you think it's better.
> 
No, I agree with you and Marcel and think that the interface is the 
nicer way of handling this. The only drawback would be that the 
implementation is not able to check whether the feature is supported or 
not; but I guess this could be done by checking the jcr implementation 
on startup.

Carsten

-- 
Carsten Ziegeler
cziegeler@apache.org