[OLSR-users] Bellman-Ford instead of Dijkstra?

[Thomas Lopatic]

Hey Benjamin,

Thanks for your clarification. Now I get it. Looks like there really
isn't any way around implementing Bellman-Ford for your scenario. Hmmm.
But then again, maybe we could turn the edges with weight -2 into edges
with weight 0 and tell Dijkstra to always take the 0-weight edge from a
given node. To be unambiguous this would restrict us to one 0-weight
edge per node.

But then again, moving to Bellman-Ford shouldn't be that hard. We could
keep most of the code in lq_route.c, I guess. Anyone? :-)

-Thomas