on Tue Nov 11 2008, "Tim Keitt"
I have a question about the requirements of the MultiPassInputIterator concept as used in the Boost Graph Library. Does the ++first_iter == ++second_iter requirement hold across calls to out_edges(vertex, g)? Or put another way, does out_edges always have to return an iterator with same ordering of returned edges?
The difference is between:
graph_traits<graph>::out_edge_iterator i = out_edges(vertex, g); graph_traits<graph>::out_edge_iterator j = i; ++i == ++j; // clearly required, but sufficient?
and
graph_traits<graph>::out_edge_iterator i = out_edges(vertex, g); graph_traits<graph>::out_edge_iterator j = out_edges(vertex, g); ++i == ++j; // also required?
Strangely enough, I have an application where the second case is (currently) not guaranteed (its a custom graph class where the out_edge_iterator is constructed on-the-fly).
I don't think any code is relying on the 2nd case, but it's hard to be sure. You could check for it by adding a generation counter in the iterator and assert if two iterators from different generations are compared. After all, they are effectively iterating different sequences. -- Dave Abrahams BoostPro Computing http://www.boostpro.com