[olsr-dev] malloc thrashing / duplicate data structures ?

[Hannes Gredler]

On Tue, Mar 27, 2007 at 12:33:36AM +0200, Thomas Lopatic wrote:
| Hey Hannes,
| 
| Yes, the data structures are quite redundant. The reason for this is
| that we first had an olsrd that was strictly RFC-compliant and stable.
| When implementing the LQ extensions we did not want to touch the stable
| code too much and implemented LQ mostly around the existing code. That's
| why translation between the data structures in the original
| RFC-compliant code and the LQ code happens, e.g. translation from the
| topology database (which is also used by the RFC-compliant code) into
| the graph for SPF/Dijkstra.
| 
| Yup, I fully agree that in a world that is ideal from an implementation
| point of view, we would drop RFC compliance, not perform any such
| translations, and save a lot of malloc()s at the same time.

ok thanks for the clarification - makes sense now to me.
looks like i have found now a good starting point for any olsr-ng efforts.

/hannes