Aaron Windsor schrieb:
No, again, I don't think directly supporting directed graphs is the way to go. The Boyer-Myrvold algorithm is based on a depth-first traversal of an undirected graph, and algorithms such as connected components and biconnected components are used elsewhere as subroutines. All of these take undirected graphs as input, so I think the right thing to do is to create an undirected adapter class for a directed graph and pass in the adapted directed graph.
Sorry, forgot that (was quite busy the last month). So, I'll see if I can turn my "make bidirected" adapter into an undirected adapter. Should be quite the same, except that the both directions are equal, and edge indices don't need to get adapted, and probably some other small stuff ...