March 2010 - Boost-users - lists.preview.boost.org

[Fusion] is mpl::insert supported for fusion::vector?
by Christian Holmquist 16 Mar '10

16 Mar '10

Hi, The following meta program compiles, except for the last (commented) line. I expected it to work since it does so for the mpl::vector case. Any ideas what I am doing wrong here? #include <boost/fusion/container/vector.hpp> #include <boost/fusion/adapted/mpl.hpp> #include <boost/mpl/insert.hpp> #include <boost/mpl/begin.hpp> #include <boost/mpl/advance.hpp> #include <boost/mpl/int.hpp> #include <boost/mpl/vector.hpp> #include <boost/mpl/equal.hpp> using namespace boost; template<class Sequence, class Pos, class T> struct insert_at { typedef typename mpl::advance<typename mpl::begin<Sequence>::type, Pos>::type itr; typedef typename mpl::insert<Sequence, itr, T>::type type; }; typedef insert_at<mpl::vector<>, mpl::int_<0>, int>::type vector1_int; BOOST_MPL_ASSERT(( mpl::equal<vector1_int, mpl::vector<int>>)); //Doesn't compile with VC2008 // typedef insert_at<fusion::vector<>, mpl::int_<0>, int>::type fusion_vector1_int; Code is using boost-trunk, rather recently updated. Thanks, Christian

2 1

Dijkstra Shortest path problem
by Kya Bey 16 Mar '10

16 Mar '10

Hi all, I'm trying to get Dijkstra's shortest path algorithm to work with my graph which has some bundled edge properties with weight as an attribute in the bundle struct edge_properties { int capacity ; double weight ; }; typedef boost::adjacency_list<listS, vecS, bidirectionalS, //directedS, property<vertex_name_t, std::string>, edge_properties> Graph; I read in the graph using the graph viz library and that works fine. (tested by printing the graph info) dynamic_properties dp; dp.property("weight", get(&edge_properties::weight, g)); dp.property("capacity", get(&edge_properties::capacity, g)); read_graphviz(in, g,dp , "name") ; I'm constantly seeing compiler errors when i use something from dijkstra-example.cpp. // Dijkstra SRC is set to 1. vertex_descriptor src = vertex(1, g); // Just as in the example std::vector<vertex_descriptor> p(num_vertices(g)); std::vector<double> d(num_vertices(g)); //RESULTS in compiler errors. See below. dijkstra_shortest_paths( g, src, predecessor_map(&p[0]).distance_map(&d[0]) ) ; //USING This call works. But have no way to get the predecessor/distance info // This fails if ***Graph*** is defined as listS,listS but works if Graph is defined as listS,vecS or vecS,vecS dijkstra_shortest_paths( g, src, weight_map(get(&edge_properties::weight,network.g)) ); The compiler error is copied. Any Pointers will be helpful. (Version of boost used = v1.34) Thanks in advance make errors /usr/include/boost/graph/dijkstra_shortest_paths.hpp: In member function 'void boost::detail::dijkstra_bfs_visitor<UniformCostVisitor, UpdatableQueue, WeightMap, PredecessorMap, DistanceMap, BinaryFunction, BinaryPredicate>::examine_edge(Edge, Graph&) [with Edge = boost::detail::edge_desc_impl<boost::bidirectional_tag, long unsigned int>, Graph = const boost::adjacency_list<boost::listS, boost::vecS, boost::bidirectionalS, boost::property<boost::vertex_name_t, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, boost::no_property>, edge_properties, boost::no_property, boost::listS>, UniformCostVisitor = boost::dijkstra_visitor<boost::null_visitor>, UpdatableQueue = boost::relaxed_heap<long unsigned int, boost::indirect_cmp<double*, std::less<double> >, boost::vec_adj_list_vertex_id_map<boost::property<boost::vertex_name_t, std::basic_string<char, std::char_traits<char>, std::allocator<char> >, boost::no_property>, long unsigned int> >, WeightMap = boost::adj_list_edge_property_map<boost::bidirectional_tag, boost::detail::error_property_not_found, const boost::detail::error_property_not_found&, long unsigned int, const boost::property<boost::edge_bundle_t, edge_properties, boost::no_property>, boost::edge_weight_t>, ..... main.cpp:34: instantiated from here /usr/include/boost/graph/dijkstra_shortest_paths.hpp:121: error: no match for call to '(std::less<double>) (const boost::detail::error_property_not_found&, double&)' /usr/include/c++/4.3/bits/stl_function.h:229: note: candidates are: bool std::less<_Tp>::operator()(const _Tp&, const _Tp&) const [with _Tp = double] make: *** [main.o] Error 1

2 2

Custom Edge Properties (boost-graph) - struct
by Kya Bey 16 Mar '10

16 Mar '10

Hi All , I've been trying to use a custom structure for describing the various edge properties of a graph. I'm defining a graph as the following struct edge_properties { int capacity ; double weight ; typedef struct edge_properties_tag kind ; // NOT SURE what this does ? }; typedef boost::adjacency_list<listS, vecS, bidirectionalS, property<vertex_name_t, std::string>, edge_properties > Graph; This allows me to create a Graph instance g. I've been trying to retrieve the edge properties via a get call, but have not been able to do so. To retrieve the vertex properties doing the following works property_map<Graph, vertex_name_t>::type v_names ; v_names = get(vertex_name,g) ; But on attempting to do something similar for retrieving the edge properties always fails with the following error message "test.cpp: 75: error: expected primary-expression before ',' token" test.cpp 71: property_map<Graph, edge_properties>::type epr; 72: property_map<Graph, vertex_name_t>::type v_names ; 74: v_names = get(vertex_name,g) ; 75: epr = get(edge_properties,g) ; I've tried various variants but seems to me i'm missing something fundamental in the way boost-graph keeps the edge_properties hidden. Any pointers will be useful. Thanks in advance

3 4

Review of Interval Template Library (ITL) for inclusion
by Jerry Jeremiah 16 Mar '10

16 Mar '10

> Author: Hartmut Kaiser > Date: 2010-02-19 02:22 +1300 > To: boost, boost-announce, boost-users > Subject: [boost] [Review] ITL review starts today, February 18th > > > Hi all, > > The formal review of the Interval Template Library (ITL) starts today, > February 18th, 2010 and will end February 27th, 2010. > ITL is being developed by Joachim Faulhaber. > > ------------------------------------------------ > > Everybody on this list is invited to participate in this formal review. I > hope to see your review of this library, including your vote, and I welcome > your participation in the discussions on the Boost mailing list. > > Please always state in your review, whether you think the library should be > accepted as a Boost library. > > Additionally, please consider giving feedback on the following general > topics: > > - What is your evaluation of the design? > - What is your evaluation of the implementation? > - What is your evaluation of the documentation? > - What is your evaluation of the potential usefulness of the library? > - Did you try to use the library? With what compiler? > Did you have any problems? > - How much effort did you put into your evaluation? A glance? > A quick reading? In-depth study? > - Are you knowledgeable about the problem domain? > > > Regards Hartmut > Review Manager > Firstly, I want to say that I have never participated in a review before so hopefully I can meet expectations for a good quality review. And let me also say, I am really glad that the review period was extended - I just don't have time to get onto something specific within a week. This review wasn't really an academic exercise for me. This library has a practical use for me so this review is a by-product of my attempt to see how using the library would compare with what I would have to hand write to get the same result. I work for a telecom and we have call detail records for millions of calls per day. What I want to know is how many simultaneous phone calls are there at any point in time. Traditionally, one would make an integer array of 7 days x 86400 seconds to hold the number of calls for each second in the week and then increment all the seconds between the start and end time of each call. With this library one simply inserts into an interval map the start and end times of the calls and then iterates through the map printing out the results. I read all the documentation twice. The first time didn't really make it clear how all the possible options related to each other and in what circumstances I would use each of them. But the second reading really helped. The examples in the documentation (like Table 1.3) were useful to my understanding. I thought the documentation was quite comprehensive and I really appreciated that - an example of that is Table 1.12 which I mention later on. The thing that differentiates a library from something I would write is that a library must cater for every single possible use. I thought the author went to a lot of effort to do this and the number of combinations of sets and maps that the library supports far exceeds the number of different use cases I have. The code samples helped clarify the differences between the different options. I didn't see an example where split_interval_map gave a result that was more useful than that of the interval_map. When would not joining touching intervals with the same value be the desired outcome? I guess the "time grid" could be such an example but it didn't use interval maps. I am not a meta-programming guru so looking at the source for Proto, MPL or Fusion just confuses me and I try to avoid that at all costs. I was a bit scared to look at the source for this library, but I put on a brave face and had a peek. It wasn't all that scary after all - it was quite readable. The source reminds me a lot like the STL headers which I guess isn't that surprising since it is implementing similar style containers (map and set). In the Party example a boost::itl::set<string> is used to store the names of the guests because it supports += as set union. In that example you can just as easily use: struct GuestSetT : public std::set<string> { std::set<string> & operator += (const std::set<string> & right) { insert(right.begin(), right.end()); return *this; }}; My question is, does this example really benefit from using itl::set? I do see value in itl::set and itl::map on their own but except this one use there aren't any examples that show why they would suit a particular application. I played with all the examples without any problem and I made a couple of my own. The library handled my example use cases elegantly although I certainly never did anything complicated enough to test the limits of the library. I used MSVC 2005 on Windows XP. There just wasn't time to try a full test with millions of phone calls spread over several days so I don't really have any idea how the performance compares to anything - I am depending on the table in the documentation (Table 1.12) for those answers. I was a bit concerned about the questions raised by http://comments.gmane.org/gmane.comp.lib.boost.devel/199902 - I did notice the disclaimer in the documentation saying that an interval_map of sets is a bad idea but maybe that should be explained more explicitly under the complexity topic. >From my point of view, the library has practical uses. I found the library easy to use because the documentation was comprehensive and the examples covered my use cases. I don't really understand how I think it would make a good addition to boost and so my vote is yes. Jerry Jeremiah

3 2

ASIO with SSL over UDP...
by Rick Meier 15 Mar '10

15 Mar '10

Is it possible to use the ASIO::SSL classes on a UDP connection? It appears that the SSL context will only work over TCP. Thanks Rick M

3 2

Problem with serialization!
by alev mutlu 15 Mar '10

15 Mar '10

Hi, I can not write serialization function for my structure. When I compile this I get very strange errors. Can someone help me please? #include <boost/mpi/environment.hpp> #include <boost/mpi/communicator.hpp> #include <boost/mpi/collectives.hpp> #include <boost/thread/barrier.hpp> #include <boost/thread/mutex.hpp> #include <boost/config.hpp> #include <string> #include <iostream> #include <vector> #include <boost/archive/text_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/serialization/access.hpp> #include <boost/serialization/string.hpp> #define MASTER 0 using namespace std; namespace mpi = boost::mpi; struct field { string field_name; string field_type; string field_ref; string field_colType; }; namespace boost { namespace serialization{ template<class Archive> void serialize(Archive & ar, field & f, unsigned int version){ ar & f.field_name; ar & f.field_type; ar & f.field_ref; ar & f.field_colType; } } } int main(int argc, char* argv[]) { mpi::environment env(argc, argv); mpi::communicator world; int m_mySize, w_mySize, kk,i; field *relAllValues; field *relMyValues; vector<field> C; w_mySize = arraySize / world.size(); if (world.rank() == MASTER){ m_mySize = arraySize / world.size() + arraySize % world.size(); relMyValues = (field *) malloc(m_mySize); allValues = (int *) malloc(arraySize); } else{ relMyValues = (field *) malloc(w_mySize); } if (world.rank() == MASTER){ scatter(world, C, relMyValues, w_mySize, 0); } else{ scatter(world, a, myValues, w_mySize, 0); } }

2 4

Problem building boost::filesystem on vxworks 6.8 using windriver's workbench
by Reger, Paul 15 Mar '10

15 Mar '10

Hi, We are trying to build version 1.41 of boost::filesystem on vxworks 6.8 using windriver's workbench, using gnu c++ compiler system version 4.1.2 and we are using the build configuration: -D__VXWORKS__ -DBOOST_ASSERT_CONFIG -DBOOST_PLATFORM_CONFIG=\"$(WS_ROOT_DIR)/ThirdParty/Boost/flir//vxworks- 68.hpp\" (We generated vxworks-68.hpp from 1.41's boost/config/platform/vxworks.hpp, and it appears below). But, we are getting the compiler error: C:/AccuRev/Boost/Boost_141_Vx_Port/ThirdParty/Boost/libs/filesystem/src/ operations.cpp:62:30: error: sys/statvfs.h: No such file or directory Anyone see this problem? Got a solution? Thanks, Paul Software Engineer FLIR Systems, Inc. 27700 SW Parkway Ave. Wilsonville, OR 97070 Ph: 503.498.3069 Fx: 503.498.3910 paul.reger(a)flir.com vxworks-68.hpp follows: // Taken from boost 1.41.0 version of the vxworks.hpp platform file. // Modified for use with VxWorks 6.8 // vxWorks specific config options: #define BOOST_PLATFORM "vxWorks 6.8" // disable wide char support // we don't use it anyway #define BOOST_NO_CWCHAR #define BOOST_NO_INTRINSIC_WCHAR_T #define BOOST_NO_STD_WSTRING // added for 6.8 #define BOOST_NO_CWCTYPE // added for 6.8 #define BOOST_NO_STD_WSTREAMBUF // added for 6.8 #define BOOST_NO_SWPRINTF // added for 6.8 // Strict ansi eliminates 64-bit integers - #if defined(__GNUC__) && defined(__STRICT_ANSI__) #define BOOST_NO_INT64_T #warning - 64-bit integer support disabled. #endif // Header file inclusions #define BOOST_HAS_UNISTD_H // these allow posix_features to work, since vxWorks doesn't // define them itself //#ifndef _POSIX_TIMERS //#define _POSIX_TIMERS 1 //#endif //#ifndef _POSIX_THREADS //#define _POSIX_THREADS 1 //#endif // We use the posix schduler in priority mode // #define _POSIX_PRIORITY_SCHEDULING // vxworks doesn't work with asio serial ports // TODO: verify this is still true in 6.8 #define BOOST_ASIO_DISABLE_SERIAL_PORT // boilerplate code: #include <boost/config/posix_features.hpp> Notice to recipient: This email is meant for only the intended recipient of the transmission, and may be a communication privileged by law, subject to export control restrictions or that otherwise contains proprietary information. If you receive this email by mistake, please notify us immediately by replying to this message and then destroy it and do not review, disclose, copy or distribute it. Thank you in advance for your cooperation.

1 0

IOService and socket CAN under Linux
by Andreas Wehrmann 15 Mar '10

15 Mar '10

Hello! I'm currently working with socket CAN under Linux 2.6.25 and was wondering how I can use these with the very handy boost ioservice facility. Socket CAN is simply a raw socket created like this: socket(PF_CAN, SOCK_RAW, CAN_RAW); Does someone maybe already have a socket CAN implementation for Boost? Is there a way I can add a socket descriptor to an IOService and get a callback when the socket becomes readable, so that I can read data from the socket? The peculiarity about raw CAN sockets is, that when they become readable, you don't read the payload of the received CAN frame(s) but a struct that represents the received CAN frame. So it behaves like a datagram socket: You only read full frames. What I actually need is a notification mechanism using the IOService facility. Best regards, Andreas

2 2

[iterator] Help needed understanding output_iterator requirements
by Thomas Klimpel 15 Mar '10

15 Mar '10

Dear iterator experts, Jesse Perla recently reported that boost::multi_array fails compilation for many basic tasks on MSVC 2010. I investigated the failures, because I had already investigated other failures on MSVC 2010 reported by Jesse Perla for boost::ublas. While I'm quite familiar with boost::ublas, I'm a complete novice with respect to boost::multi_array and boost::iterator. I could track the failures to a specific (intentional?) behavior of iterator_facade from Boost.Iterator (more specifically of facade_iterator_category.hpp): http://lists.boost.org/Archives/boost/2010/03/163039.php Thomas Klimpel wrote: > What effectively happens it that the iterator gets the category > "input_iterator", but std::copy wants to have an "output_iterator". I > think I understand why the iterator doesn't satisfies the requirement > of a "forward_iterator", but it's unclear to me why it didn't qualify > as an "output_iterator". There is an "input_output_iterator_tag" in > "boost/iterator/detail/facade_iterator_category.hpp", so it would have > been possible for the iterator to be both "input_iterator" and > "output_iterator". However, it seems to me that > "input_output_iterator_tag" is never used in the Boost.Iterator > library. I would have to read more about "output_iterator" to > understand whether Boost.Iterator is correct here. With my current > understanding of "output_iterator", I would say it's a bug in > "boost/iterator/detail/facade_iterator_category.hpp". > > Perhaps somebody with more knowledge about "output_iterator" and > Boost.Iterator can explain to me why an iterator using the > "boost::random_access_traversal_tag" for the "CategoryOrTraversal" > template parameter of "iterator_facade" will never get the category > "output_iterator" (or "input_output_iterator_tag"). Of course I could go on and read more about "output_iterator", or find out whether the Boost.Iterator documentation has to say something about this issue, but since I only came to this problem because I wanted to help Jesse Perla on a problem with boost::multi_array, I would much prefer if an "iterator expert" could "take over". Regards, Thomas

1 0

Re: [Boost-users] ranking of ublas vector
by Kraus Philipp 14 Mar '10

14 Mar '10

Hi, I think the problem is: we discuss about the "rank" and the "rank index". I my problem I only need the rank index and the value to calculate a new value like: exp( -rank_index) * vec(rank_index) So the smallest element in vec will be multiply with 1 and other elements with values < 1 Phil

2 4