- Boost-users - lists.preview.boost.org

boost/graph - induced subgraph
by Matthew Galati 24 Mar '04

24 Mar '04

Is it possible to find the minimum spanning tree of an induced subgraph? vector< graph_traits<Graph>::vertex_descriptor > vd; typedef subgraph< Graph > SubGraph; SubGraph & inducedG = subG.create_subgraph(vd.begin(), vd.end()); Here is the example I am looking at: induced verts: 1 2 3 4 5 7 Original Graph: 0 <--> 1 2 5 7 1 <--> 0 3 5 7 2 <--> 0 3 4 5 3 <--> 2 1 4 5 4 <--> 2 3 5 6 5 <--> 0 1 2 3 4 6 7 6 <--> 4 5 7 7 <--> 0 1 5 6 Induced Graph: 1 <--> 3 5 7 2 <--> 3 4 5 3 <--> 2 1 4 5 4 <--> 2 3 5 5 <--> 1 2 3 4 7 7 <--> 1 5 So, the "create_subgraph" seems to be working fine. But, when I run kruskals with: vector<graph_traits < SubGraph >::edge_descriptor> spanning_tree; kruskal_minimum_spanning_tree(inducedG, back_inserter(spanning_tree)); for (vei = spanning_tree.begin(); vei != spanning_tree.end(); ++vei) { pair<int,int> ij = incident(*vei, inducedG); int edge_index = e_index[*vei]; cout << "e_index: " << edge_index << " : " << ij.first << " " << ij.second << endl; I get a spanning tree of just 4 edges (while the subgraph has 6 vertices). Also, how does the edge_index of the subgraph correspond to the original graph? For example, edge (4,1) is not in the original graph, but it shows up in my spanning tree. Spanning Tree: e_index: 3 : 2 0 e_index: 7 : 4 0 e_index: 4 : 3 1 e_index: 8 : 4 1 Thanks, Matthew -- Matthew Galati ISE Lehigh University IBM Service Parts Solutions 610.758.4042 (Office) 610.882.0779 (Home) magh(a)lehigh.edu, magal11(a)us.ibm.com http://sagan.ie.lehigh.edu/mgalati/

4 3

boost-sandbox socket and boost 1.31.0?
by Robert Helgesson 23 Mar '04

23 Mar '04

Howdy, I'm wondering regarding the socket library found in boost-sandbox. Will it be updated to function nicely with the latest Boost release? The problem seems to be that boost::default_iterator_policies and boost::iterator_adaptor no longer work like the socket library expect. Regards -- / Robert Helgesson |-- rycee(a)home.se --\ \-- GPG Key ID: 41E95A5B --| http://www.rycee.cx/ /

3 2

Re: [Boost-users] obtaining vertex descriptor from property_map (newbie)
by Gustavo Sánchez 23 Mar '04

23 Mar '04

I tend to use a map<string, vertex_descriptor> for this. Doug Thanks for the idea, I was just thinking something like this, but I have a problem. I need the definition of the Graph typedef property<edge_name_t, std::string> EdgeProperty; typedef property<vertex_name_t, std::map<string,vertex_property??> > NameProperty; typedef property<vertex_index_t, std::size_t, NameProperty> VertexProperty; typedef adjacency_list<vecS, vecS, undirectedS, VertexProperty, EdgeProperty> UGraph; in order to get the vertex_descriptor, but I need the map<string,vertex_descriptor> for the VertexPorperty so I drop in a circular type dependency. Can you please be more verbose? thanks --------------------------------- Antivirus Filtros antispam 6 MB gratis ¿Todavía no tienes un correo inteligente?

2 1

Solaris and thread_resource_error
by Peter Jones 23 Mar '04

23 Mar '04

On Feb 1 of this year Baptiste Lepilleur posted a status report for Boost 1.31.0 Release candidate 2 on Solaris and AIX. It was mentioned that "boost.thread is now systematically failing with 'thread resource error'". I am experiencing the same thing on Solaris 8, with GCC 3.3.2 (GNU binutils) and Boost 1.31.0. Has anyone looked into this yet? If not, I am willing to track down what line in Boost.Threads is throwing thread_resource_error(). -- Peter Jones [unix;vi;crypto;c++;4wd] (http://pmade.org) PGP [5137 173E 5BFA 8C63 8982 F035 FDBB 1636 7D69 C63F]

2 3

linking to boost::thread
by Jerry 23 Mar '04

23 Mar '04

Hi, I've just built boost1.31.0 using bjam under Linux, apparently successfully. There don't appear to be any static libraries for boost::thread. Is this correct? If so, could someone please explain why I'm getting linker errors under g++ if I don't mention the _dynamic_ thread libs on the g++ command line, and "cannot open shared object file" at runtime if I do. Any pointers appreciated. TIA, Jerry.

2 1

shared_ptr help needed, please!
by Eric Damphousse 23 Mar '04

23 Mar '04

hi all, quick question concerning shared_ptr... not sure if things get deleted (destroyed) properly and its freaking me out. I hope this is not too long Basically, I use the BGL to store my data and that seems to be ok. The way the BGL stuff is defined is as such... Note that spObjectNode is defined as typedef boost::shared_ptr<CObject> spObjectNode // Set vertex properties typedef boost::property<boost::vertex_name_t, std::string> node_name_t; typedef boost::property<graph_node_t, spObjectNode, node_name_t> node_property_t; // Set edge properties typedef boost::property<boost::edge_name_t, std::string> connection_name_t; typedef boost::property<graph_connection_t, spObjectNode, connection_name_t> edge_property_t; // Set graph typedef adjacency_list<listS, vecS, bidirectionalS, node_property_t, edge_property_t> graph_t; // graph_t m_Graph; The function below is what I use to add a node to the BGL... template <typename T> void AddNode(const T& object) { // Create the vertex boost::shared_ptr<T> pNode(new T(object)); // Add the vertex + its name to the graph vertexdescriptor_t newNode = boost::add_vertex(node_property_t(pNode, node_name_t(pNode->GetObjectName())), m_Graph); }; The way I use the AddNode ... for example in a main.cpp: /// blablabla above CSphere sphere("sphere1"); CSceneDB::Instance().AddNode(sphere); return 0; For testing purposes, in the base class and derived classes (see below) I print a message in the destructor of the class, just to make sure things are ok. If I just have in my main the previous call then upon exit I expect and get the following message... CSceneDB cleanup CSphere cleanup CGeometry cleanup CObject cleanup Now for the source of my anxiety... I wrote a function to get back my object from the BGL and it goes like this... inline boost::shared_ptr<CObject> getNodeProperty(const vertexdescriptor_t& vertex) { boost::property_map<graph_t, graph_node_t>::type node_prop_map; node_prop_map = boost::get(graph_node_t(), m_Graph); return node_prop_map[vertex]; } boost::shared_ptr<CObject> CSceneDB::getNodebyName(const std::string& strName) { boost::shared_ptr<CObject> node_prop; boost::graph_traits<graph_t>::vertex_iterator vi, vi_end; boost::tie(vi, vi_end) = boost::vertices(m_Graph); for(; vi != vi_end; ++vi) { node_prop = getNodeProperty(*vi); if(node_prop != 0) { if(node_prop->GetObjectName() == strName) break; } } return node_prop; } template <typename T> boost::shared_ptr<T> getNode(const std::string& strName) { boost::shared_ptr<CObject> abc = getNodebyName(strName); shared_ptr<T> def = boost::dynamic_pointer_cast<T>(abc); return def; } The problem is that if I do something like this... in the main.cpp /// blablabla stuff above CSphere sphere("sphere1"); CSceneDB::Instance().AddNode(sphere); boost::shared_ptr<CSphere> abc = CSceneDB::Instance().getNode<CSphere>("sphere1"); return 0; I only get upon exit... CSceneDB cleanup what happened to CSphere cleanup, CGeometry cleanup and CObject cleanup, are they leaking? Now for a sanity check, if I do in the main.cpp ... i do NOT do the assignment boost::shared_ptr<CSphere> abc = /// blablabla stuff above CSphere sphere("sphere1"); CSceneDB::Instance().AddNode(sphere); CSceneDB::Instance().getNode<CSphere>("sphere1"); return 0; upon exit i get ... CSceneDB cleanup CSphere cleanup CGeometry cleanup CObject cleanup Can someone tell me why this is, is my assignment wrong? Any help would be greatly appreciated Eric Brief description of the classes used ofr my test // Object.h class CObject { protected: CObject(){}; protected: virtual ~CObject(){cout << "CObject cleanup" << endl;}; }; typedef boost::shared_ptr<CObject> spObjectNode; // Geometry.h class CGeometry : public CObject { protected: CGeometry(){}; protected: virtual ~CGeometry(){cout << "CGeometry cleanup" << endl;}; }; // Sphere.h class CSphere : public CGeometry { public: CSphere(){}; ~CSphere(){cout << "CSphere cleanup" << endl;}; };

3 4

dumb question: iterator dynamic_cast adapter?
by Paul Elliott 22 Mar '04

22 Mar '04

I am trying to figure out iterators and I have a dumb question: I have an iterator A_iter, that can traverse in both directions ( A_iter++ and A_iter-- ) and can do input (operator*()). Value_type = type of *A_iter is A; and A is static_cast able to B: A start; B fish = static_cast<B>(start); // OK How do I use adapters to create B_iter, which is the same as A_iter, but whose value_type= typeof *B_iter is B? I am sorry if this is a dumb question. -- Paul Elliott 1(512)837-1096 pelliott(a)io.com PMB 181, 11900 Metric Blvd Suite J http://www.io.com/~pelliott/pme/ Austin TX 78758-3117

2 1

obtaining vertex descriptor from property_map (newbie)
by Gustavo Sánchez 22 Mar '04

22 Mar '04

Dear all, We have a graph which has a property of type string associated with the vertexes. What we want to do is to obtain a vertex descriptor from a string. Could you please tell us how to implement it? Thank you in advance! --------------------------------- Antivirus Filtros antispam 6 MB gratis ¿Todavía no tienes un correo inteligente?

2 1

Controlling selection of runtime libraries
by Bob Kline 22 Mar '04

22 Mar '04

We're using Boost's regex package in a Windows environment for a server built with Microsoft Visual Studio 7.x. We'd like to be able to build our server with debugging symbols for our own code, but still use the release versions of the Microsoft runtime DLLs (because the debugging versions of the Microsoft runtime DLLs aren't redistributable, so we can't legally install them on our production server). The problem is, when we compile our own code for debugging (with -Zi and -D_DEBUG), the Boost headers detect this and do their behind-the-curtain magic to point to a version of the regex package which is linked to the debugging versions of the runtime libraries (msvcr71d and msvcp71d). We've tried using the BOOST_REGEX_NO_LIB trick, specifying the version of the regex library to which we want to link (boost_regex_vc7_mdi.lib) but this doesn't work, because although it compiles fine, the linker chokes complaining about the inability to find a bunch of severely mangled names. What's the *real* trick for using the version of the library we want? Do we need to hack the distributed makefile each time we upgrade regex? (Or worse, shuffle libraries around by renaming them?) Thanks. Bob Kline

3 4

Re: boost-sandbox socket and boost 1.31.0?
by Hugo Duncan 22 Mar '04

22 Mar '04

Robert Helgesson <rycee(a)home.se> wrote: > I'm wondering regarding the socket library found in boost-sandbox. Will > it be updated to function nicely with the latest Boost release? Hi, I have something similar that works with current boost CVS code. I have uploaded this to http://sourceforge.net/projects/giallo/ in an attempt to get this moving again. So far I have low level wrappers for sockets and named pipes. Documentation, tests are lacking. I will be adding the reactor code that I uploaded to the yahoo groups files section in due course. Hugo ps Thanks Jeff for pointing me at this post

1 0