Created an attachment (id=640) [details]
The script used to get the mentioned results of OLSR

In the OLSR Routing Protocol, the MPR set of a node is having more nodes than
necessary.

/* Topology (Constant Position)
 * A---D
 * | / |
 * C---E
 * |
 * F
 *
 * F - Node 0 - IP 10.1.1.1
 * C - Node 1 - IP 10.1.1.2
 * E - Node 2 - IP 10.1.1.3
 * A - Node 3 - IP 10.1.1.4
 * D - Node 4 - IP 10.1.1.5
 *
 * Trace File Line 47625 says MPR nodes of A are C & D. But it should be only
C.
 * Similarly  Line 47521 says MPR nodes of E are C & D. But it should be only
C.
 */

This doesn't follow the algorithm given in the comments of the
src/routing/olsr/olsr-routing-protocol.cc or the RFC. 
I mailed the developer Gustavo Carneiro
and quoting the mail:

 1. C is the only node that can reach the two-hop neighbor F, so it
should be among the first to be selected as MPR;

  2. Once C is selected as MPR, C can reach all the other nodes, so no
more MPR should be selected.

So I agree with you, the MPR set of A should be only C, not C+D.  If
this does not happen, sounds like a bug.