Boost: applying dijkstra_shortest_path() to filter

I have a graph with bundled properties

//property
struct Edge
{
    int distance;
};

//graph
typedef boost::adjacency_list<boost::vecS, boost::vecS, boost::bidirectionalS, Node, Edge> graph_t;

//predicate
template <typename Graph>
struct Filter
{
    const Graph gr;
    Filter(const Graph &g): gr(g) {}
    bool operator() (const Edge &e) const {
        if(gr[e].distance < 5) return 0;
        return 1;
    }
};

Filter <graph_t> filter(g);
boost::filtered_graph <graph_t, Filter <graph_t> > fg(g, filter);

And I want to apply dijkstra_shortest_path algorithm to filtered graph.

std::vector<int> d(num_vertices(fg));
std::vector<v> p(boost::num_vertices(fg));
boost::dijkstra_shortest_paths(fg, nodes[0], boost::weight_map(get(&Edge::distance, fg))
        .distance_map(boost::make_iterator_property_map(d.begin(), get(boost::vertex_index, fg)))
        .predecessor_map(boost::make_iterator_property_map(p.begin(), get(boost::vertex_index, fg))));

The errors i get, while compiling are:

/Users/artem/Documents/boost_1_55_0/boost/concept_check.hpp:142: error: object of type 'boost::filter_iterator >, boost::keep_all, boost::filtered_graph, Filter >, boost::keep_all> >, boost::detail::out_edge_iter, void *>, Edge> *>, unsigned long, boost::detail::edge_desc_impl, long> >' cannot be assigned because its copy assignment operator is implicitly deleted a = b; // require assignment operator ^

What is incorrect? Can't really understand, how to solve this problem.

As referred in this post, your operator() should accept an edge descriptor and not an edge.

For example, you can try something like this:

bool operator() (const boost::graph_traits<graph_t>::edge_descriptor& e) const 
{
    ... get the variable edge of type Edge from the edge_descriptor e in some way ...

    if(gr[edge].distance < 5) return 0;
    return 1;
}

All these things make boost::graph very annoying to use. I hope the developers will improve the usability, which for the moment is very poor.

It may be useful to have a look to this post as well.