Wrap entry iterator using efficient order.

Iterating over entries using their index may be inefficient as it generate a lot of cache miss in the internal cluster cache.

Efficient order is entries sorted by they cluster, and so, provides to most efficient way to use the cluster cache.

zim-bind introduce new classes EntryRangeEfficient and IterEfficient which wrap zim's EntryRange and iterator but specialized with EntryOrder::efficientOrder.