- Boost-users - lists.boost.org

compilation errors of Parallel BGL
by Da Zheng 07 Apr '11

07 Apr '11

Hello, I try to build Parallel BGL in Ubuntu. I use gcc version 4.4.5, boost v1.46.1, openmpi v1.4.1-3ubuntu1. I believe my environment fulfills your requirements. Then I download parallel-bgl-0.7.0.tar.gz, and try to compile it, but I get some compilation errors. Below shows all my steps of compiling the library. cmake works fine. zhengda@zhengda-desktop:~/graph/parallel-bgl-0.7.0$ BOOST_ROOT=/home/zhengda/graph/boost_1_46_1 cmake . -- The C compiler identification is GNU -- The CXX compiler identification is GNU -- Check for working C compiler: /usr/bin/gcc -- Check for working C compiler: /usr/bin/gcc -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Check for working CXX compiler: /usr/bin/c++ -- Check for working CXX compiler: /usr/bin/c++ -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Found MPI: /usr/lib/libmpichcxx.so -- Boost version: 1.46.1 -- Found the following Boost libraries: -- mpi -- serialization -- filesystem -- system -- Configuring done -- Generating done -- Build files have been written to: /home/zhengda/graph/parallel-bgl-0.7.0 When I compile the code, it shows the following errors. zhengda@zhengda-desktop:~/graph/parallel-bgl-0.7.0$ make Scanning dependencies of target boost_graph_mpi [ 2%] Building CXX object libs/graph_parallel/src/CMakeFiles/boost_graph_mpi.dir/mpi_process_group.cpp.o In file included from /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/mpi_process_group.hpp:807, from /home/zhengda/graph/parallel-bgl-0.7.0/libs/graph_parallel/src/mpi_process_group.cpp:13: /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp: In function ‘int boost::graph::distributed::receive(const boost::graph::distributed::mpi_process_group&, int, int, T&)’: /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp:722: error: ‘stderr’ was not declared in this scope /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp:724: error: there are no arguments to ‘fprintf’ that depend on a template parameter, so a declaration of ‘fprintf’ must be available /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp:724: note: (if you use ‘-fpermissive’, G++ will accept your code, but allowing the use of an undeclared name is deprecated) /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp: In function ‘typename boost::enable_if<boost::mpi::is_mpi_datatype<T>, std::pair<int, unsigned int> >::type boost::graph::distributed::receive(const boost::graph::distributed::mpi_process_group&, int, int, T*, size_t)’: /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp:743: error: ‘stderr’ was not declared in this scope /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp:745: error: there are no arguments to ‘fprintf’ that depend on a template parameter, so a declaration of ‘fprintf’ must be available /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp: In function ‘int boost::graph::distributed::receive(const boost::graph::distributed::mpi_process_group&, int, int, T&) [with T = int]’: /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp:609: instantiated from ‘void boost::graph::distributed::mpi_process_group::global_trigger_launcher<Type, Handler>::receive(const boost::graph::distributed::mpi_process_group&, int, int, boost::graph::parallel::trigger_receive_context, int) const [with Type = int, Handler = void (*)(const boost::graph::distributed::mpi_process_group&, int, int, int, bool)]’ /home/zhengda/graph/parallel-bgl-0.7.0/libs/graph_parallel/src/mpi_process_group.cpp:1110: instantiated from here /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp:722: error: ‘fprintf’ was not declared in this scope /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp: In function ‘int boost::graph::distributed::receive(const boost::graph::distributed::mpi_process_group&, int, int, T&) [with T = boost::graph::distributed::mpi_process_group::outgoing_messages]’: /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp:609: instantiated from ‘void boost::graph::distributed::mpi_process_group::global_trigger_launcher<Type, Handler>::receive(const boost::graph::distributed::mpi_process_group&, int, int, boost::graph::parallel::trigger_receive_context, int) const [with Type = boost::graph::distributed::mpi_process_group::outgoing_messages, Handler = void (*)(const boost::graph::distributed::mpi_process_group&, int, int, boost::graph::distributed::mpi_process_group::outgoing_messages&, bool)]’ /home/zhengda/graph/parallel-bgl-0.7.0/libs/graph_parallel/src/mpi_process_group.cpp:1110: instantiated from here /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/detail/mpi_process_group.tpp:722: error: ‘fprintf’ was not declared in this scope make[2]: *** [libs/graph_parallel/src/CMakeFiles/boost_graph_mpi.dir/mpi_process_group.cpp.o] Error 1 make[1]: *** [libs/graph_parallel/src/CMakeFiles/boost_graph_mpi.dir/all] Error 2 make: *** [all] Error 2 So I modify boost/graph/distributed/detail/mpi_process_group.tpp, and insert "#include <stdio.h>". It seems the error above is solved, but I get a new error. This time I don't know how to fix this error. zhengda@zhengda-desktop:~/graph/parallel-bgl-0.7.0$ make [ 2%] Building CXX object libs/graph_parallel/src/CMakeFiles/boost_graph_mpi.dir/mpi_process_group.cpp.o [ 5%] Building CXX object libs/graph_parallel/src/CMakeFiles/boost_graph_mpi.dir/tag_allocator.cpp.o Linking CXX static library ../../../lib/libboost_graph_mpi.a [ 5%] Built target boost_graph_mpi Scanning dependencies of target adjlist_build_test [ 8%] Building CXX object libs/graph_parallel/test/CMakeFiles/adjlist_build_test.dir/adjlist_build_test.cpp.o In file included from /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/adjacency_list.hpp:3961, from /home/zhengda/graph/parallel-bgl-0.7.0/libs/graph_parallel/test/adjlist_build_test.cpp:10: /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/adjlist/serialization.hpp: In function ‘std::vector<int, std::allocator<int> > boost::detail::parallel::available_process_files(const std::string&)’: /home/zhengda/graph/parallel-bgl-0.7.0/boost/graph/distributed/adjlist/serialization.hpp:97: error: conversion from ‘boost::filesystem3::path’ to non-scalar type ‘std::string’ requested make[2]: *** [libs/graph_parallel/test/CMakeFiles/adjlist_build_test.dir/adjlist_build_test.cpp.o] Error 1 make[1]: *** [libs/graph_parallel/test/CMakeFiles/adjlist_build_test.dir/all] Error 2 make: *** [all] Error 2 Did I do something wrong? Why can I not compile the library? Thanks, Da

3 2

Re: [Boost-users] (no subject)
by Gabriel Marchand 07 Apr '11

07 Apr '11

Thank you for this solution. It compiles. Nevertheless, I still don't achieve what I want to do and I don't understand a few things (I attached the code): 1/ I run the connected_component() function after defining the graph graph_t G(num_nodes) and I expect that there should be as many components as num_nodes as I didn't add the edges yet. However the return of the function tells me there is 1 component ?? After adding the edges between all vertices, the number of components is 1, as it should be. But when I use the function clear_vertex(), I expect to create a new component as the vertex cleaned contain no more edges. The number of components is still one. What I am doing wrong here? 2/I need to know the size in number of vertices of the components, therefore I need to use something like for (i = 0; i != num_nodes; ++i) cout << "Vertex " << i <<" is in component " << component[i] << endl; cout << endl; but of course with our changings done this can't be used. Do you have an idea how I could extract this information. Gabriel --- En date de : Jeu 7.4.11, Cedric Laczny <cedric.laczny(a)gmx.de> a écrit : De: Cedric Laczny <cedric.laczny(a)gmx.de> Objet: Re: [Boost-users] (no subject) À: boost-users(a)lists.boost.org Date: Jeudi 7 avril 2011, 6h50 Alright, now that was a somehow tricky one... Details s. below. On Wednesday, 6. April 2011 12:55:52 Gabriel Marchand wrote: > Hi, I am new to boost and I face the same problem mentioned here with > "connected_components()" using a mutable graph (hence using a list for the > vertex container). I try the detailed solution of Aaron Windsor in the > previous post but I still get some errors at compilation. Here is my > little code: > > > > #include <boost/graph/adjacency_list.hpp> > #include <boost/graph/connected_components.hpp> > > using namespace std; > using namespace boost; > > Box& b = conf.GetBox(); > > > if (!ai) { > ai = new AtomList(); > *ai = conf.GetAtoms().AtomSelector(selecti, b); > } > > typedef boost::adjacency_list< > listS, > listS, > undirectedS, > property<vertex_index_t, size_t>, > no_property> > > > graph_t; > > graph_t g(ai->size()); > property_map<graph_t, vertex_index_t>::type index = get(vertex_index, > g); graph_traits<graph_t>::vertex_iterator vi, vi_end; > graph_traits<graph_t>::vertices_size_type cnt = 0; > for(tie(vi,vi_end) = vertices(g); vi != vi_end; ++vi) > put(index, *vi, cnt++); > > for (unsigned int i = 0; i < ai->size(); ++i) { > Cartesian& crdi = ai->GetElement(i)->pos; > for (unsigned int j = i + 1; j < ai->size(); ++j) { > Cartesian& crdj = ai->GetElement(j)->pos; > double r = conf.GetBox().CartesianDistance(crdi, crdj); > if (r < rcut) { > add_edge(vertex(i,g), vertex(j,g), g); // Add edge to Map > if distance criteria is satisfied > > } > } > } > > vector<int> component(num_vertices(g)); > int num_clusters = connected_components(g, &component[0], > vertex_index_map(index)); > In the example for connected_components() you used, you probably found the "&component[0]" notation and simply used it. However, this does _not_ work for graphs with listS, only for vecS. So basically what you have to do is to define your own components map and _not_ use that wrapper around a vector. As boost seems to model vertices with listS as void* this can hardly work as it is missing the index that it would normally have when using vecS. Also, for vecS, vertices seem to be modeled as int and therefore can serve as indices for vectors directly. So, to make your code compiles, you should use: typedef graph_traits<graph_t>::vertex_descriptor Vertex; map< Vertex, unsigned int > index_std_map; typedef associative_property_map< map< Vertex, unsigned int > > IndexMap; IndexMap index( index_std_map ); // put() the values by iterating over all vertices instead of getting the vertex_index from the graph via "get(vertex_index, G)". With this, you can define your own components-map: vector<int> v_component(num_vertices(G)); iterator_property_map< vector<int>::iterator, IndexMap > component( v_component.begin(), index ); and call connected_components() via: int num_clusters = connected_components(G, component, vertex_index_map(index)); > > > > The compilation is successfull until I add the last line: > int num_clusters = connected_components(g, &component[0], > vertex_index_map(index)); > > > Hope somebody can give a hint > gabriel Hope that I could help and that this solution will be correct for you. Please, of course, test the whole approach in order to make sure it works as you expect it to. Generally it should, as long as the mapping from vertex to index/offset is correct. If there are further questions about the problem or my solution, feel free to mail again. Best, Cedric _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

2 1

Simple Lambda Problem
by Richard Howard 07 Apr '11

07 Apr '11

I am currently trying to get to grips with using Lambda, but I have hit a problem fairly early on. Can anyone explain what I am doing wrong. bool isCar(std::string const & val) { return val.find("car") != std::string::npos; } int main() { std::list<std::string> l; l.push_back("car"); l.push_back("blue car"); l.push_back("red van"); l.push_back("bike"); l.push_back("blue car"); // ... l.remove_if(&isCar); // this works and is how I have done things in the past but I was hoping to avoid the need for the separate function by using Lambda. // I have tried the following l.remove_if(boost::lambda::var(boost::lambda::bind(&std::string::find, boost::lambda::_1, "car")) != std::string::npos); // I am not sure why this does not work, can anyone point me in the right direction? }

2 1

[gil] Processing huge images
by Franz Alt 07 Apr '11

07 Apr '11

Hi all, I'm very new in Boost.GIL so I hope anyone could help me with my question. Imagine several images all with identical dimensions (e.g. 10000x10000). All images form one huge "virtual" image. Because every image is very big (~ 50M) I couldn't copy them together to one image in memory. I want to perform some algorithms (e.g. x-gradient) over the huge image with GIL in a (hopefully) elegant way. The user should use GIL on the huge image "as normal" as possible. Does anyone know how I could handle this problem? Regards Franz

4 6

from matrix to vectors
by Virili 07 Apr '11

07 Apr '11

Hi to all, I'm trying to have a vector (for example a row or a column) from a matrix. like in Matlab : my_column_vector = my_matrix(:,3) or my_row_vector = my_matrix(1,:); I know that it will be a stupid thing, but can you post a reference or a pseudo code?? Thanks!!

2 1

[range] for_each and std::map of std::list
by Mark Snelling 07 Apr '11

07 Apr '11

Hi, Consider a map defined as... std::map< int, std::list< std::string > > my_map; I'm wondering if/how it's possible to use Boost.Range to call a function on each std::string in each of the lists contained in the map. I have an implementation below which works using a C++0x lambda function but am curious if this can be done purely using Boost.Range? boost::for_each( boost::adaptors::values( my_map), []( const std::list< std::string >& my_list ) { boost::for_each( my_list, some_func ); } ); ideas anyone?

5 8

a question on boost::spirit with a quick program.
by Littlefield, Tyler 07 Apr '11

07 Apr '11

Hello all: Sorry for all the questions; I am kind of easing into boost and finding cool things i can do with it, as well as uses for a couple of projects. Eventually I would like to write a small scripting language using boost::spirit, just to do it and get the practice using it. As a starting point, I wrote a quick calculator (that just does small single operations like x+x, x*x, etc). Now i have two questions. 1) My rule seems to be off. the first number comes through clearly, but no matter what I put in, I always get the first number and then a 42. 2) How would I go about extending this to implement order of operations, parenthases, and longer operations (like 3+3*3)? I'm going to try to implement a small function table so I can do tan(3), so I assume I will need nultiple rules, how might that be accomplished? Any comments and suggestions would be welcome. I have pasted the code below. I was working from a tutorial, so early appologies if it feels messy; I plan to clean up the namespace aliases and the using keywords. /* A simple calculator, supports multiplication, division, addition and subtraction. */ #include <boost/spirit/include/qi.hpp> #include <boost/spirit/include/phoenix_core.hpp> #include <boost/spirit/include/phoenix_operator.hpp> #include <boost/spirit/include/phoenix_operator.hpp> #include <iostream> #include <string> #include <boost/fusion/include/adapt_struct.hpp> #include <boost/fusion/container/vector.hpp> #include <boost/fusion/container/vector/convert.hpp> #include <boost/fusion/include/io.hpp> namespace qi = boost::spirit::qi; namespace ascii = boost::spirit::ascii; namespace phoenix = boost::phoenix; using qi::double_; using qi::_1; using ascii::space; using phoenix::ref; using std::cout; using std::endl; typedef struct { int a, b; char op; } calculation; BOOST_FUSION_ADAPT_STRUCT(calculation, (int, a) (int, b) (char, op) ) template <typename iterator> class CParser:public qi::grammar<iterator, calculation(), ascii::space_type> { qi::rule <iterator, calculation(), ascii::space_type> crule; public: CParser():CParser::base_type(crule) { using ascii::char_; using qi::int_; crule %= (int_ >> (char_('+')|char_('-')|char_('*')|char_('/')) >> int_); } }; int main() { using std::cout; using std::cin; using std::endl; using boost::fusion::vector; CParser<std::string::iterator> p; std::string str; calculation result; cout << "Enter your calculation>"; cin >> str; bool r = phrase_parse(str.begin(), str.end(), p, ascii::space, result); if (r) { cout << result.a << " " << result.b << " " << result.op << endl; switch(result.op) { case '+': cout << "Result: " << (result.a + result.b) << endl; break; case '-': cout << "result: " << (result.a - result.b) << endl; break; case '*': cout << "Result: " << (result.a * result.b) << endl; break; case '/': if (result.a == 0) { cout << "Division by zero error." << endl; break; } cout << "result: " << (result.a / result.b) << endl; break; default: cout << "Invalid operation." << endl; } } else { cout << "Invalid input." << endl; } return 0; } -- Thanks, Ty

4 3

[Spirit 2] on_error to string
by Vincent Agnus 07 Apr '11

07 Apr '11

Hi, I have some trouble to retreive error parsing information using qi::on_error<fail> to work on a std::string. The documentation propose the following code : on_error<fail> ( start , std::cout << val("constituent: expected ") << _4 // what failed? << val(" here: \"") << construct<std::string>(_3, _2) // iterators to error-pos, end << val("\"") << std::endl ); * On my First attempt. I have define a class member (std::stringstream m_error) in my grammar class and replace std::cout by phx::ref(m_error). But this code generate a compilation error * For the second attempt, I declare class member m_error as std::string and use this error handler: namespace qi = boost::spirit::qi; namespace phx = boost::phoenix; on_error< qi::fail> ( start , phx::ref(m_error) = ( phx::construct<std::string>(phx::val("Error! Expecting ")) // + qi::_4 REMARK1 + phx::construct<std::string>( phx::val(" here: \"") ) + phx::construct<std::string>(qi::_3, qi::_2) + phx::construct<std::string>( phx::val("\"") ) ) ); that's work perfectly. On error my string is correctly filled. But my problem is that I can't get the expected token (_4) (REMARK1), If I uncomment line REMARK1, I get a compilation error, same result using the code phx::construct<std::string>(qi::_4). How can a get qi::_4 as string ? Can any one with kindness help me?

2 5

(no subject)
by Gabriel Marchand 07 Apr '11

07 Apr '11

Hi, I am new to boost and I face the same problem mentioned here with "connected_components()" using a mutable graph (hence using a list for the vertex container). I try the detailed solution of Aaron Windsor in the previous post but I still get some errors at compilation. Here is my little code: #include <boost/graph/adjacency_list.hpp> #include <boost/graph/connected_components.hpp> using namespace std; using namespace boost; Box& b = conf.GetBox(); if (!ai) { ai = new AtomList(); *ai = conf.GetAtoms().AtomSelector(selecti, b); } typedef boost::adjacency_list< listS, listS, undirectedS, property<vertex_index_t, size_t>, no_property> > graph_t; graph_t g(ai->size()); property_map<graph_t, vertex_index_t>::type index = get(vertex_index, g); graph_traits<graph_t>::vertex_iterator vi, vi_end; graph_traits<graph_t>::vertices_size_type cnt = 0; for(tie(vi,vi_end) = vertices(g); vi != vi_end; ++vi) put(index, *vi, cnt++); for (unsigned int i = 0; i < ai->size(); ++i) { Cartesian& crdi = ai->GetElement(i)->pos; for (unsigned int j = i + 1; j < ai->size(); ++j) { Cartesian& crdj = ai->GetElement(j)->pos; double r = conf.GetBox().CartesianDistance(crdi, crdj); if (r < rcut) { add_edge(vertex(i,g), vertex(j,g), g); // Add edge to Map if distance criteria is satisfied } } } vector<int> component(num_vertices(g)); int num_clusters = connected_components(g, &component[0], vertex_index_map(index)); The compilation is successfull until I add the last line: int num_clusters = connected_components(g, &component[0], vertex_index_map(index)); Hope somebody can give a hint gabriel

2 7

[phoenix] [range] I think I'm missing a couple things w/ these libs
by Jake Brownson 07 Apr '11

07 Apr '11

I love the ideas behind these libs, but I think I need to figure out a bit more before I can use them. Here's an example that models what I'm trying to do for real as an experiment w/ using phoenix and range. I can't quite get it to build. I think I'm really close. I tried it w/ lambda before I found phoenix and I had to do some ret<> things to get past some errors, but I'm not sure what the equivalent in phoenix is, or if I should really need them. Also, I think there should be a way to instantiate a vector directly from a range w/o having to use begin() and end() somehow, but I can't find it. I've tried various approaches with <something> including transformed_range<> and any_range<> and haven't gotten them to work. The transformed line gives me errors complaining about missing result_type, iterator_category, value_type, difference_type, pointer, reference. I'm using GCC 4.2 on OS X Thanks! ---- #include <boost/range/adaptor/transformed.hpp> #include <boost/spirit/include/phoenix_core.hpp> #include <boost/spirit/include/phoenix_bind.hpp> #include <vector> class A {}; class B { public: A a; }; class C { public: static B BFromA(const A& a) { return B(); } }; std::vector<C> CsFromBs(std::vector<B> bs) { <something> cs = bs | boost::adaptors::transformed(boost::phoenix::bind(&C::BFromA, boost::phoenix::bind(&B::a, boost::phoenix::arg_names::arg1))); return std::vector<C>(boost::begin(cs), boost::end(cs)); }

4 9