Re: [boost] [BGL] depth_first_visitor improvement?

From: Vladimir Prus (ghost_at_[hidden])
Date: 2002-01-25 11:05:55

Jeremy Siek wrote:
> On Fri, 25 Jan 2002, Vladimir Prus wrote:
> ghost> BTW, I'm confused by docs to DFSVisitor::finish_vertex:
> ghost>
> ghost>     This invoked on a vertex after all of its out edges have been
> added to ghost>     the search tree and all of the adjacent vertices have
> been discovered ghost>    (but before their out-edges have been examined).
> ghost>
> ghost> I don't understand the statement about out-edges of the
> ghost> adjacent vertices. In fact, depth_first_visit_impl is called
> ghost> and all adjacent vertices and it will traverse out-edges. Am
> ghost> I wrong?
>
> Does the following wording make more sense to you:
>
> (but before the out-edges of the adjacent vertices have been examined)

Let's try once again: suppose we've discovered vertex u and the only out edge 
is (u->v). The following wording means: "when finish_vertex(u) is called, 
discover_vertes(v) is already called, but examine_edge for any of v's out 
edges is not. Looking at depth_first_search.hpp, line 80 and below, I see 
that immediately after discover_vertex(v) is called, all out edges are 
examined. Therefore, I don't understand this wording either.

- Volodya