Add a new method to "quickly jump" documents in a cursor (that has been skipped and already iterated).

This might be done either allowing to invoke skip() twice (in a way that the second time it starts skipping from current position) or providing a method similar to next() that does not actually return data (at least avoiding network traffic between MongoDB and the client).

The usage scenario is the following:

We have a huge collection and created a cursor skipping N documents (being N big, say 10 millions). We retrieve some documents, and now we need to jump another M documents.

We have two options:
1- Create another cursor and skipping it N+M times;
2- Invoke next() M times on the original cursor.

Being N big, option 1 results slow.

On the other side, next() is less efficient than skip() to "just jump" documents. Our tests suggest that using next() M times (option 2) is worthwhile only if M is less than 10% of N.

I'm from the RESTHeart team (http://restheart.org), and we met this need implementing a DBCursor Pool to speedup queries on collections (that already speedups things up to 1000% according to performance tests).

For some reference about the idea, have a look at: https://softinstigate.atlassian.net/wiki/x/h4CM

Performance test results:
https://softinstigate.atlassian.net/wiki/x/gICM

For some code:
https://github.com/SoftInstigate/restheart/blob/develop/src/main/java/org/restheart/db/DBCursorPool.java