[oak-core] offset and count parameters in NodeState.getChildNodeEntries

[Michael Dürig]

Hi,
Is there a uses case for having offset and count parameters in 
NodeState.getChildNodeEntries (package 
org.apache.jackrabbit.oak.spi.state)? The only one I can see is to do 
lazy reading from the Microkernel. This can however also be done 
implicitly by the implementation of getChildNodeEntries() instead of 
explicitly by the API consumer.

These parameters are troublesome since they make decorating NodeState 
instances to filter out certain child node entries difficult to 
implement. In fact any such implementation would need to explicitly 
iterate through up to the offset supplied which make the whole thing 
pointless.

If no one objects, I'll remove these parameters and adapt 
KernelNodeState to take care of lazy reading.

Michael