Is anyone aware of any bugs in the graph library 'transitive_closure' algorithm? I believe I have a concrete example that exhibits an obvious bug (in particular, missing vertices that are reachable from certain source vertices), but (1) wasn't sure if there was a work-around or some well-known avoidance tactic and (2) whether anyone was actively supporting the graph library components.

Thanks!

Jean-Paul Watson                                              E-mail: jwatson@sandia.gov
Discrete Math Dept (9215), MS 1110                   Phone: (505) 845-8887
P.O. Box 5800, Sandia National Labs                  FAX:     (505) 845-7442
Albuquerque, NM 87185-1110 USA                     (Old) URL: http://www.cs.colostate.edu/~watsonj