March 2012 - Boost-users - lists.boost.org

Re: [Boost-users] Graph: writing graph properties to graphml file
by Juerg Tschirren 15 Mar '12

15 Mar '12

I still have a problem getting graph properties written to a graphml file (vertex and edge properties get written just fine). Following Jeremiah's advise form two weeks ago I implemented the code below. The problem appears to be that the instance of ref_property_map contained in the dynamic_properties uses the key type Graph*, but the write_graphml function implemented in boost/graph/graphml.hpp expects it to be of type Graph. This of course results in a different typeid and the write_graphml function does not recognize the entry in the dynamic properties as graph properties. I tried to change the key type of ref_property_map to Graph, which resulted in a compilation error. Any insight in how to solve that problem would be greatly appreciated! Here is my current code: #include <boost/graph/graph_traits.hpp> #include <boost/graph/adjacency_list.hpp> #include <boost/graph/graphml.hpp> struct VertexProperties { std::string strName; }; struct EdgeProperties { std::string strName; }; struct GraphProperties { std::string strName; }; std::ostream& operator<< (std::ostream& strm, const VertexProperties& p) { strm << p.strName; return strm; } std::istream& operator>> (std::istream& strm, VertexProperties& p) { strm >> p.strName; return strm; } std::ostream& operator<< (std::ostream& strm, const EdgeProperties& p) { strm << p.strName; return strm; } std::istream& operator>> (std::istream& strm, EdgeProperties& p) { strm >> p.strName; return strm; } std::ostream& operator<< (std::ostream& strm, const GraphProperties& p) { strm << p.strName; return strm; } std::istream& operator>> (std::istream& strm, GraphProperties& p) { strm >> p.strName; return strm; } typedef boost::property<boost::vertex_name_t, VertexProperties> vertex_p; typedef boost::property<boost::edge_weight_t, EdgeProperties> edge_p; typedef boost::property<boost::graph_name_t, GraphProperties> graph_p; typedef boost::adjacency_list<boost::listS, boost::vecS, boost::directedS, vertex_p, edge_p, graph_p> Graph; typedef boost::graph_traits<Graph>::vertex_descriptor vertex_t; typedef boost::graph_traits<Graph>::edge_descriptor edge_t; typedef boost::property_map<Graph, boost::vertex_name_t>::type vertex_prop_t; typedef boost::property_map<Graph, boost::edge_weight_t>::type edge_prop_t; typedef boost::ref_property_map<Graph*, GraphProperties> graph_propt_t; int main() { //-- create a graph: Graph g; vertex_t u = boost::add_vertex(g); vertex_t v = boost::add_vertex(g); edge_t e; bool b; boost::tie(e,b) = boost::add_edge(u,v,g); //-- get the property maps: vertex_prop_t vertexPropMap = boost::get(boost::vertex_name, g); edge_prop_t edgePropMap = get(boost::edge_weight, g); GraphProperties& graphProp = boost::get_property(g, boost::graph_name); //-- set some vertex, edge, and graph properties: vertexPropMap[u].strName = "first vertex"; vertexPropMap[v].strName = "second vertex"; edgePropMap[e].strName = "an edge"; graphProp.strName = "the graph"; //-- create dynamic property map: boost::dynamic_properties dp(boost::ignore_other_properties); dp.property("vertex_properties", vertexPropMap); dp.property("edge_properties", edgePropMap); graph_propt_t graphPropMap(graphProp); dp.property("graph_properties", graphPropMap); //-- write graphml to file: std::ofstream graphmlOutFile("graph.xml"); boost::write_graphml(graphmlOutFile, g, dp, true); graphmlOutFile.close(); return 0; } On Thu, 1 Mar 2012, Jeremiah Willcock wrote: > On Tue, 28 Feb 2012, Juerg Tschirren wrote: > > I wrote the code below to construct a directed graph with internal properties (for vertices, edges, and the graph itself) and write everything to a graphml file. The code > > compiles and works as it is, but my two questions are: > > > > 1. What dynamic property do I have to add to write the graph properties out to the graphml file as well? So far I could only figure out how to write vertex and edge > > properties. > You should be able to add a graph property like in the following code > (from libs/graph/test/graphviz_test.cpp): > boost::ref_property_map<graph_t*,std::string> gname( > get_property(graph,graph_name)); > dp.property("name",gname); > > 2. Does everything else look good? Specifically, is it O.K. to define the properties the way I did in the first typedef line (without using boost::property<>), and is it > > O.K. to access the properties the way I do it? > It will work, but a better way would be to use boost::property or a class > containing your string which you can then access using g[v].mem or > g[e].mem. >-- Jeremiah Willcock

2 1

[boost::asio::serial_port] Calling cancel() causes Run-Time Check Failure #0 on Visual Studio 2010 SP1
by Dario Ramos 15 Mar '12

15 Mar '12

This is a fatal error, there's no way to recover from it, and in release builds, a nasty message box appears. It happens when I call serial_port::cancel and Boost tries to throw some exception. The exact error is: Run-Time Check Failure #0 - The value of ESP was not properly saved across a function call. This is usually a result of calling a function declared with one calling convention with a function pointer declared with a different calling convention. And it seems to happen in throw_error.ipp, inside the do_throw_error method. Here's a small sample program which reproduces this: #include <boost/asio/serial_port.hpp> #include <boost/asio/read.hpp> #include <boost/thread.hpp> #include <iostream> using namespace std; BYTE g_pBuffer[128]; void ReadCompletionHandler( boost::system::error_code ec, std::size_t bytesTransferred ){ if( !ec ){ cout << "Read " << bytesTransferred << " bytes successfully" << endl; }else if( ec == boost::asio::error::operation_aborted ){ cout << "Aborted async_read_some, " << bytesTransferred << " bytes transferred" << endl; }else{ cout << "Read finished with errors: " << ec.message() << endl; } } void ReaderThread( boost::asio::serial_port* pSerialPort ){ cout << "[ReaderThread] Started reading" << endl; boost::asio::async_read( *pSerialPort, boost::asio::buffer( g_pBuffer, 5 ), ReadCompletionHandler ); cout << "[ReaderThread] Done!" << endl; } void main(){ //Tell cout not to buffer its output so output better reflects multithreaded execution std::cout.setf(std::ios::unitbuf); boost::asio::io_service ioService; boost::asio::serial_port serialPort( ioService ); string portName = "COM1"; serialPort.open( portName ); if( !serialPort.is_open() ){ cout << "Failed to open " << portName << endl; return; } cout << "Launching reader thread and io_service" << endl; boost::thread readerThread( boost::bind( ReaderThread, &serialPort ) ); ioService.run(); cout << "Sleeping..." << endl; ::Sleep( 5000 ); cout << "Woke up, cancelling pending reads" << endl; serialPort.cancel(); cout << "Waiting for reader thread to finish..." << endl; readerThread.join(); cout << "Done!" << endl; cin.get(); } Maybe I'm doing something stupid. What do you think? -- *Darío Eduardo Ramos* Meditech S.A. www.meditech.com.ar (+54) 01147603300, Interno 31 Av. Julio A. Roca 3456 Florida Oeste, Bs.As. Argentina

3 4

GSoC 2012
by Devashish Tyagi 15 Mar '12

15 Mar '12

Hello, I am Devashish Tyagi, an Undergraduate Student of Computer Science and Engineering at Indian Institute of Technology, Delhi. I am currently in my Third Year. I have a considerable experience of coding in C++ and strong background in the field of Algorithms. I would like to be part of Boost C++ Library for GSoC 2012. I went through the Ideas page here<https://svn.boost.org/trac/boost/wiki/SoC2012>. I would like to extend Boost.Graph by implementing the requests mentioned on this page<http://www.boost.org/doc/libs/1_49_0/libs/graph/doc/challenge.html>. Among the Challenges mentioned, I am particularly interested in implementing Dynamic Graph Algorithms and a minimum cost maximum flow Algorithm. Could someone guide me as to from where should I begin my research and Is there a priority order among the requests/challanges mentioned on this page<http://www.boost.org/doc/libs/1_49_0/libs/graph/doc/challenge.html>. Regards -- Devashish Tyagi Junior Undergraduate Computer Science and Engineering Dept.

2 1

Boost Asio - VxWorks
by Julien PONNOU 15 Mar '12

15 Mar '12

Hello, I'm trying to compile Boost Asio under VxWorks OS, with GNU compiler, without any success. Right now, I've setted up a new project with Workbench 3.3, and included the <boost/asio.hpp> library. As a result, I got a tone of errors because many files it is including doesn't exist. Actually, it is trying to include Linux's files such as /sys/poll.h and /sys/uio.h. These files don't exist in VxWorks. Does somebody already compiled it under VxWorks ? Best regards, Julien.

2 3

iterator_adaptor usage question
by John M. Dlugosz 15 Mar '12

15 Mar '12

What is a good way to create both an iterator and a const iterator while only having to specify all my custom behavior once?

4 3

[MPL]...if_ and enable_if
by Robert Jones 15 Mar '12

15 Mar '12

I'm having a bit of punt into the murky world of metaprogramming, so this might be a bit naive...! I've been using Boost.enable_if for a while, and now I've just encountered Boost.MPL.if_ and wonder if enable_if is simply a special case of MPL.if_. Given struct empty { }; template <typename T> struct my_condition : /* boost true_ or false_ */; is template <typename T> typename enable_if<my_condition<T>, int>::type f( ) { /* .... */ } the same as template <typename T> typename eval_if<my_condition<T>, identity<int>, empty>::type f( ) { /* ... */ } Or is there some subtlety I'm missing (which is quite likely!) Thx - Rob.

3 2

[Graph] On Dijkstra SSSP performance
by Leo Hidd 15 Mar '12

15 Mar '12

Hello guys, I'm newbie on Boost, but thanks to some examples I found over the net I could make it work. Now I have a question on the performance of the dijkstra_shortest_paths function. The kind of graph I have is actually a 3D surface mesh, with each vertex being a node and each edge being an edge of the graph. So my graph is undirected, the number of edges is approximately 3 times the number of vertices (I typically have around 4 millions of vertices), the degree of the vertices is not constant, it varies between 3 and 12 (but 90% of the vertices have 5 or 6 neighbors) and the weight of the edges are the (3D Euclidean) distances between the adjacent vertices. As for the Boost lib, so far I'm using the classic adjacency_list<listS, vecS, undirectedS, no_property, < boost::edge_weight_t, Scalar > > graph_t type, with the dijkstra_shortest_paths function. But I feel the performance is very low compared to my implementation of Dijkstra algorithm. Dijkstra algorithm is very simple, so my implementation is a simple transcription of the text book in C language (I'm using Microsoft Visual Studio 2008). The data structure I'm using is a classic binary tree typedef struct tree_el { int idx; int source_idx; double dist; tree_el * right, * left; } node; and the functions I've implemented are insert (inserts a node with the smallest dist as the left most element or, in case of equality, it put the node with smallest idx as the left most element), remove_smallest (return the left most node and delete it from the tree) and remove_node (finds and return a given node and delete it from tree). The remove_node I use combined with insert whenever a given node has it's dist updated (in case new dist is smallest than the current one), so I remove the node, create a new one with the new dist (but same idx) and insert it into the tree (so that it will be always sorted). Well, since this is a real simple scholar case and that the performance is around 4 times faster than the dijkstra_shortest_paths function, which I call like this: std::vector<vertex_descriptor> p(num_vertices(g)); std::vector<Scalar> d(num_vertices(g)); vertex_descriptor s = vertex(A, g); dijkstra_shortest_paths (g, s, predecessor_map(make_iterator_property_map(p.begin(), get(vertex_index, g))). distance_map(make_iterator_property_map(d.begin(), get(vertex_index, g)))); I'm here to ask you how to make a better use of Boost in order to have performance at least in pair with what I have so far. I'm not concerned by the build time of the graph (add nodes/edges) since I do it only once and it remains the same all over my runtime. I'm only concerned by the SSSP itself, which I call thousands of times with different sources (I actually do multi source shortest path search) (all pairs is not a solution to me since I use way less sources than the total of nodes I have, and, for a graph with as much nodes as I have, I can't hold in memory all these distances). It seems to massively use STL containers (like vectors and lists) which maybe be the reason why it runs slower, but I'm not sure. Is there a way to use other graph type so that it can run faster? Thank you for your attention, leo

3 2

Re: [Boost-users] C++ committee calling for lib proposals
by Brian Wood 15 Mar '12

15 Mar '12

> The C++ committee is calling for library proposals for their inclusion in C++11: http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3370.html . > I think some of the boost libraries are good candidates, like the recently accepted boost::lockfree. Has anyone considered sending such proposal? I'm not sure how the committee feels about open code generation -- http://web.it.kth.se/~cschulte/projects/unison.php -- but I think moving in that direction could help revive library development. Regards, Brian Wood Ebenezer Enterprises http://webEbenezer.net "In a sort of *ghastly* *simplicity* we remove the organ and demand the function. ... We castrate and bid the geldings be *fruitful*." -- C.S. Lewis

2 1

Problems using boost serializing and custom constructors
by Allan Nielsen 14 Mar '12

14 Mar '12

Hi I'm trying to use boost::serialize for the first time, but I have some troubles figuring out how to serialize a class like this: #include <vector> struct Foo { struct Bar { std::vector<int> * data; // Must point to Foo::data Bar( std::vector<int> * d ) : data(d) { } }; std::vector<int> data; std::vector<Bar> elements; Foo() { // do very time consuming calculation to populate "data" and "elements" } }; Here is what I have tried: struct Foo { struct Bar { std::vector<int> * data; Bar( ) : data( 0 ) { } Bar( std::vector<int> * d ) : data(d) { } template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & data; // is this correct? } }; std::vector<int> data; std::vector<Bar> elements; Foo() { std::cerr << "Running default constructor" << std::endl; data.push_back(1); data.push_back(2); data.push_back(3); data.push_back(4); data.push_back(5); elements.push_back( Bar( &data ) ); elements.push_back( Bar( &data ) ); elements.push_back( Bar( &data ) ); } template<class Archive> Foo( Archive & ar ) { ar >> data; // is this corrent? ar >> elements; } private: BOOST_SERIALIZATION_SPLIT_MEMBER(); friend class boost::serialization::access; template<class Archive> void save(Archive & ar, const unsigned int version) const { const std::vector<int> * data_ptr = &data; // should data be seriliazed as pointer... // it is used as a pointer in Bar ar << data_ptr; ar << elements; } }; int main(int argc, const char *argv[]) { #if 0 // serialize Foo foo; boost::archive::text_oarchive oar(std::cout); oar << foo; #else // deserialize boost::archive::text_iarchive oar(std::cin); Foo foo(oar); #endif std::cerr << foo.data.size() << std::endl; std::cerr << foo.elements.size() << std::endl; std::cerr << (&foo.data) << std::endl; for( const auto& a : foo.data ) std::cerr << a << " "; std::cerr << std::endl; for( const auto& a : foo.elements) std::cerr << a.data << " "; std::cerr << std::endl; return 0; } The only problem is that it does not work. After I deserialize Foo, data is still empty. Please advise me on how to do this properly. Best regards Allan W. Nieslen

3 12

Przemyslaw SLIWA is out of the office.
by przemyslaw.sliwa＠uk.bnpparibas.com 14 Mar '12

14 Mar '12

I will be out of the office starting 14/03/2012 and will not return until 26/03/2012. I will respond to your message when I return. ___________________________________________________________ This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. Any unauthorised copying, disclosure or distribution of the material in this e-mail is prohibited. Please refer to http://www.bnpparibas.co.uk/en/information/legal_information.asp?Code=ECAS-… for additional disclosures. BNP Paribas is proud to sponsor David Hockney RA: A Bigger Picture at the Royal Academy, 21 January - 9 April 2012

1 0