December 2009 - Boost-users

Template metaprogramming technique for operator overloading
by Edward Diener 15 Dec '09

15 Dec '09

I could not find this in type_traits, or any other Boost TMP library, but I am wondering if there is any TMP technique which can determine if a user-defined type supports a particular operator overload for that type ?

3 2

[1.41][serialization] deserialization of weak_ptr fails
by Alexander Gutenev 15 Dec '09

15 Dec '09

The example below works in 1.40 and fails in 1.41 #include <string> #include <sstream> #include <vector> #include <boost/smart_ptr.hpp> #include <boost/serialization/serialization.hpp> #include <boost/serialization/shared_ptr.hpp> #include <boost/serialization/vector.hpp> #include <boost/archive/xml_oarchive.hpp> #include <boost/archive/xml_iarchive.hpp> struct X : boost::enable_shared_from_this<X> { X(){} virtual ~X() {} explicit X(int i) : i(i) {} template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & boost::serialization::make_nvp("i", i); (void) version; } int i; }; typedef boost::shared_ptr<X> PX; typedef std::vector<PX> VPX; int main() { std::ostringstream oss; VPX vpx; vpx.push_back(PX(new X(2))); vpx.push_back(PX(new X(5))); vpx.push_back(PX(new X(7))); boost::archive::xml_oarchive oa(oss); oa << boost::serialization::make_nvp("v", vpx); vpx.clear(); std::string s = oss.str(); std::istringstream iss(s); boost::archive::xml_iarchive ia(iss); ia >> boost::serialization::make_nvp("v", vpx); PX px = vpx[1].get()->shared_from_this(); return 0; }

2 3

[Boost Graph] newbie - dijkstra & external weightmap
by Fergal Dalton 14 Dec '09

14 Dec '09

Hi, I'm trying to write a program which considers some "agents" traversing a graph following dijkstra, from some source vertex to some destination. Each agent will have a different weightmap for the network. Therefore, I need a routine to which I pass the source & destination vertices, and a pointer or reference to the calling agents weightmap. This routine should return the next step for that agent. Alternatively, I could write a visitor which, everytime any agent calls dijkstra, overwrites the weight in the bundled edge structure, with that agent's weights array. Here below is the nuts of what I have, which works but only with the bundled edge_weight (called "transit_time"). Below that, is what I'd like to do. Anyway, the long & short of it is, I couldn't wrap my head around the syntax for either. I've looked in the books, in the lists, on the site, and I can't find any help on how to incorporate external weightmaps with dijkstra. Can anyone tell me how to proceed? Thanks, Fergal. --------------- This works, but only with the bundled weight property, identical for all agents: class road; typedef adjacency_list<setS,vecS,undirectedS,no_property,road> Graph; typedef Graph::vertex_descriptor vdest; vdest shortest_path(Graph& g, vdest src, vdest dst) { vector<vdest> parent(num_vertices(g)); dijkstra_shortest_paths(g,src,weight_map(get(&road::transit_time,g)) .predecessor_map(make_iterator_property_map(parent.begin(), get(vertex_index, g)))); vdest vd; for (vd=dst ; parent[vd]!=src ; vd=parent[vd]); // ATTN: isolated vertices return(vd); } ----------------- What I would like, where agent_weightmap is (a pointer to) the agent's weightmap: class road; typedef adjacency_list<setS,vecS,undirectedS,no_property,road> Graph; typedef Graph::vertex_descriptor vdest; vdest shortest_path(Graph& g, vdest src, vdest dst, vector<double> agent_weightmap) { vector<vdest> parent(num_vertices(g)); dijkstra_shortest_paths(g,src,weight_map(agent_weightmap)) .predecessor_map(make_iterator_property_map(parent.begin(), get(vertex_index, g)))); vdest vd; for (vd=dst ; parent[vd]!=src ; vd=parent[vd]); // ATTN: isolated vertices return(vd); }

2 1

Linker error with boost::filesystem
by seandr＠millisecond.com 14 Dec '09

14 Dec '09

Hi, I'm trying to use the boost::filesystem library, and I'm getting two linker errors that I can't figure out. Does anyone have any ideas why I'm seeing these and how I can fix them? 1. error LNK2001: unresolved external symbol "class boost::filesystem::file_status __cdecl boost::filesystem::detail::status_api(class std::basic_string<unsigned short,struct std::char_traits<unsigned short>,class std::allocator<unsigned short> > const &,class boost::system::error_code &)" 2. error LNK2001: unresolved external symbol "class boost::system::error_code __cdecl boost::filesystem::detail::set_current_path_api(class std::basic_string<unsigned short,struct std::char_traits<unsigned short>,class std::allocator<unsigned short> > const &)" I'm using the msvc8 toolset. Thanks, Sean

1 0

documentation build problem
by Stefan Strasser 14 Dec '09

14 Dec '09

I get the following error when I try to build any documentation with boost 1.41: cat "/tmp/jam112c3106.000" > "/bin/boostbook_catalog.xml" obviously it can't copy to /bin/, but BOOST_ROOT is set. not only is the root missing from the directory, but the bin directory is called "bin.v2", not "bin". I can't figure out where this "cat" is coming from or how to set the path correctly. can anyone help me with this?

2 1

io_service object pool X io_service::run thread pool
by Matheus Araújo Aguiar 14 Dec '09

14 Dec '09

Hello dear list, I am using boost for some months now and i have implemented a server which creates several threads of io_service::run in order to respond to async writes and reads. I am looking for some performance improvements and just read the http server 2 example and saw that it uses a pool of io_service objects. So my question is, what would be the pros and cons of using such a pool instead of a thread pool of io_service::run ? Thanks in avance and best regards, -- Matheus Araújo Aguiar Computer Scientist matheus.pit(a)gmail.com

1 0

[thread] 1 blocks are still reachable in loss record
by Sakharuk, Vladimir 14 Dec '09

14 Dec '09

Hi all, Valgrind reports follows memory leak. ==24567== at 0x4A1FE1B: malloc (vg_replace_malloc.c:207) ==24567== by 0x6AC7FDD: boost::detail::get_once_per_thread_epoch() (in /app/local/pkgs/boost/lib/libboost_thread-gcc41-mt-1_39.so.1.39.0) ==24567== by 0x6AC4A17: void boost::call_once<void (*)()>(boost::once_flag&, void (*)()) (in /app/local/pkgs/boost/lib/libboost_thread-gcc41-mt-1_39.so.1.39.0) ==24567== by 0x6AC1D96: boost::detail::get_current_thread_data() (in /app/local/pkgs/boost/lib/libboost_thread-gcc41-mt-1_39.so.1.39.0) ==24567== by 0x6AC36B4: boost::thread::join() (in /app/local/pkgs/boost/lib/libboost_thread-gcc41-mt-1_39.so.1.39.0) Code is looks like: thread_->join(); delete thread_; Is there a way to fix it? Regards \/lad. =============================================================================== Please access the attached hyperlink for an important electronic communications disclaimer: http://www.credit-suisse.com/legal/en/disclaimer_email_ib.html ===============================================================================

2 1

[date_time] Is there a more direct way to format time strings?
by Sajjan Kalle 14 Dec '09

14 Dec '09

In the examples all formatting of strings seems to use date_facet, which requires ( ? ) the use of streams and locales. Is there a more direct way to format a time?

1 0

[program-options] Modifying the variable_value of a variables_map...
by W Eryk Wolski 14 Dec '09

14 Dec '09

Hi, I am using program options to parse cmdline parameters. I would like to use the variable_map to pass parameters in my application and I would like to modify the values of the boost::program_options::variable_map. while variable_value provides a method value() allowing me to modify the content... the variable_map "overrides" the operator[] of the std::map const variable_value& operator[](const std::string& name) const; so something like this (where mmap is a variable_map) mmap["x"].value() = 2.; does not compile..... I tried this because the doc reads : The variables_map class can be used just like std::map, except that values stored there must be retrieved with the as method shown above. (If the type specified in the call to the asmethod is different from the actually stored type, an exception is thrown.) Using the iterator it is map_t::iterator it(mmap.find("x")); boost::program_options::variable_value & vx(it->second); vx.value() = 2.; still easily possible to modify the variable_value . I would like to better understand why the operator[] returns a const variable_value? regards Eryk Here is some sample code I wrote to test how to modify a variable_map. boost::any a(1.); boost::program_options::variable_value vv(a,false); //variable_value provides the method value() returning a reference via which I am able to modify the content. // i.e. vv.value() = 2.1; //I store the variable_value in a variables_map typedef boost::program_options::variables_map map_t; map_t mmap; mmap.insert( std::make_pair(std::string("x"), vv ) ); std::cout << mmap["x"].as<double>() << std::endl; //now I would like to modify the value.. mmap["x"].value() = 2.; //does not compile..... ///error C2678: binary '=' : no operator found which takes a left-hand operand of type 'const boost::any' (or there is no acceptable conversion) // thats because abstract_variables_map defines the operator[] as follows // const variable_value& operator[](const std::string& name) const; //my workaround... map_t::iterator it(mmap.find("x")); boost::program_options::variable_value & vx(it->second); vx.value() = 2.; std::cout << mmap["x"].as<double>() << std::endl; -- Witold Eryk Wolski Heidmark str 5 D-28329 Bremen tel.: 04215261837

1 0

[Interprocess] How to prevent collision of mapping addresses with DLL Loading on Windows?
by pheres 14 Dec '09

14 Dec '09

Hi, I do something like fixed_managed_shared_memory segment (open_only, "MyFixedAddressSharedMemory" //Shared memory object name ,(void*)0x30000000 //Mapping address to establish an shared memory segment which gets faded into the virtual address space of every process taking part on the IPC at the same offset (this is important for me). This works quite well for most processes. But problems arise with IPC-clients using a lot of delay loaded DLL's. This seems to be especially to be true for MFC windows programs. While normal loaded DLL's goes to 0x10000000, or their fixed address or are relocated to some there like 0x0..., some windows delay loaded DLL's seem to can be placed nearly anywhere inside the virtual address space, This makes the success of establishing the IPC into a lottery game. For example I used 0x40000000 for a shared memory range in the IPC clients. One client used windows MFC. Taking part on the IPC inside this client should happen after some pushbutton was pressed but it failed sometimes. I found out that in the moment of the button press windows delay loaded two additional DLL's which unfortunately faded in at the memory range I configured for the IPC (around 0x40000000). I was not able to find a 100% satisfying solution. The best thing I came up with was: -link everything you can static -Link the self produced DLL's with fixed base address -Allocate the memory range as the first call in main() using OS-specific functions like VirtualAlloc() -Free the allocated memory just before initializing the IPC (so it's sure that no delay loaded DLL can be faded in in the mean time) -for the rest (mainly windows or other 3rd party DLLs) hope the best This seem to work at least for now but I guess there must be a more elegant solution. I did google a lot about the MS loader/run time linker, but there seems to be no option to prevent it from selecting a certain address range. Could anybody help me out? Thank you in advance! Best Regards Joerg

3 8