Paging result sets is required in order to avoid failures
related to large result sets.

The Next() call on StateQueryIterator from Chaincode was failing
in peer due to a corner case. This fix addresses the corner case (where
the number of results is an exact multiple of shim batch size).

This is done by keeping the HasMore flag in iterator accurate by
pre-fetching the next batch of results when the last cached item
from shim is accessed.

This fix also adds the Namespace field to commonledger.KV type. The data
returned to peer has namespace information. All state iterators return
 QueryResult with commonledger.KV type elements.

Change-Id: Id872130db8a20ec5a59593b1fa004dce43ee36ac
Signed-off-by: Balaji Viswanathan <balaji.viswanathan@gmail.com>
Signed-off-by: denyeart <enyeart@us.ibm.com>