Hi
Am 04.06.2013 um 08:31 schrieb <stefane...@apache.org> <stefane...@apache.org>:
> sling/trunk/bundles/extensions/discovery/api/src/main/java/org/apache/sling/discovery/DiscoveryService.java
> (original)
> +++
> sling/trunk/bundles/extensions/discovery/api/src/main/java/org/apache/sling/discovery/DiscoveryService.java
> Tue Jun 4 14:31:14 2013
> @@ -24,6 +24,9 @@ package org.apache.sling.discovery;
> * The discovery service is in charge of managing live instances that
> * have announced themselves as being part of a topology view. The exact
> * details of how this announcement occurs is implementation dependent.
> + * <p>
> + * Note that the discovery service fails if it detects a situation
> + * where more than one instance with the same sling.id exists in a cluster.
> */
> public interface DiscoveryService {
This sounds wrong documentation here: There is nothing in the JavaDoc of the
DiscoveryService API that indicates that the sling.id is important. So an
application does not require the use of the sling.id. In fact why should it ?
So adding this sentence here makes no sentence. To the contrary it adds some
confusing information about a concrete implementation.
It really is an implementation thing to (a) use the sling.id and (b) to
(probably not register the service ? ) handle duplication situation.
Plus: what does "the discovery service fails" mean ?
I suggest to remove this.
Regards
Felix
