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

[program_option] how to say only certain strings are allowed as arguments?
by Thorsten Ottosen 06 Mar '10

06 Mar '10

Hi, Is there some way to say bpo::options_description desc("Allowed options"); desc.add_options() ("method", bpo::value<std::string>()->allowed_values("foo")("bar"), "" ) so I donøt have to reject the arguments manually? Thanks -Thorsten

4 4

[proto] Deep copy question
by Manjunath Kudlur 06 Mar '10

06 Mar '10

Hello, I define the copy constructor for my class and print out out a message when it is called. See the following program. When I have "proto::terminal<Var<int> >::type a", and try to call BOOST_PROTO_AUTO(pr, a), I see that the copy constructor is called two times. Why is the object getting copied twice during the deep copy? When I have "proto::literal<Var<int> > a", I see that the copy constructor is getting called twice during the declaration of 'a' itself, and twice when I try to call BOOST_PROTO_AUTO(pr, a). I was under the impression that objects are held by reference inside expressions. Why are they copied twice in the proto::literal case? Actually, I am looking for some advice on the following: I know that when I have proto::terminal<Var<int> >, the underlying object is created once, and held by reference *everywhere* it is appears in a large complex expression. But as soon as I try to do a deep_copy of the expression (as in BOOST_PROTO_AUTO), the underlying object is copied once for *every* appearance in the expression. How do I make it such the underlying object doesn't copied many times? I want to have the copied expression also to have the property that the original expression had, i.e., everywhere an underlying object is referenced, they all refer to the same object. I will be grateful for any advice. #include <boost/proto/proto.hpp> #include <boost/proto/proto_typeof.hpp> #include <iostream> #include <sstream> #include <string> namespace proto=boost::proto; namespace mpl=boost::mpl; template<typename VT> struct Var { VT value; Var(const Var<VT> &that) { std::cout << "Copy\n"; value = that.value; } Var() { std::cout << "Default\n"; value = 0; } void assign(const VT v) { value = v; } }; typedef proto::literal<Var<int> > int32_; //typedef proto::terminal<Var<int> >::type int32_; int main() { int32_ a; std::cout << "Before\n"; BOOST_PROTO_AUTO(pr, (a)); std::cout << "After\n"; } Thanks, Manjunath

1 1

[mpl] defaults with associative sequence
by Manfred Doudar 06 Mar '10

06 Mar '10

Hi, Maybe I am misreading the documentation, but if I have an mpl::map and try and get an element associated with a key all's good, while if I assign a default - I get an error that mpl::at only takes 2 and not 3 template parameters. Simple example for sake of completness: using namespace boost::mpl; typedef map< pair<int, int_<0> > , pair<float, int_<1> > > m; at<m, float>::type::value; // Ok at<m, double, int_<2> >::type::value; // compiler error Indeed a look in boost/mpl/at.hpp even leads me to believe that mpl::at only takes two template parameters (maybe I mistaken), but documentation suggests otherwise. [Compiler: GCC-4.4.2 ; Boost-1.42] Am I doing something blatantly wrong? Cheers, -- Manfred

2 2

Problem with Boost Serialization
by Binh Luong 05 Mar '10

05 Mar '10

Hello, i have some problems with the serialization of boost. I am working now with a program and want to use boost to serialize an object to a xml archive. I have read the tutorials on the homepage and make the necessary steps but still get some errors while compiling: In file included from /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/xml_oarchive.hpp:31, from message/remote/RemoteMessagingSystem.h:30, from message/remote/RemoteMessagingSystem.cpp:1: /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/basic_xml_oarchive.hpp: In member function 'void boost::archive::basic_xml_oarchive<Archive>::save_override(T&, int) [with T = Message* const, Archive = boost::archive::xml_oarchive]': /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/detail/interface_oarchive.hpp:64: instantiated from 'Archive& boost::archive::detail::interface_oarchive<Archive>::operator<<(T&) [with T = Message* const, Archive = boost::archive::xml_oarchive]' /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/detail/interface_oarchive.hpp:72: instantiated from 'Archive& boost::archive::detail::interface_oarchive<Archive>::operator&(T&) [with T = Message*, Archive = boost::archive::xml_oarchive]' message/remote/RemoteMessagingSystem.cpp:147: instantiated from here /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/basic_xml_oarchive.hpp:92: error: no matching function for call to 'assertion_failed(mpl_::failed************ boost::serialization::is_wrapper<Message* const>::************)' In file included from /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/xml_iarchive.hpp:24, from message/remote/RemoteMessagingSystem.h:31, from message/remote/RemoteMessagingSystem.cpp:1: /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/basic_xml_iarchive.hpp: In member function 'void boost::archive::basic_xml_iarchive<Archive>::load_override(T&, int) [with T = Message*, Archive = boost::archive::xml_iarchive]': /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/xml_iarchive.hpp:80: instantiated from 'void boost::archive::xml_iarchive_impl<Archive>::load_override(T&, int) [with T = Message*, Archive = boost::archive::xml_iarchive]' /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/detail/interface_iarchive.hpp:61: instantiated from 'Archive& boost::archive::detail::interface_iarchive<Archive>::operator>>(T&) [with T = Message*, Archive = boost::archive::xml_iarchive]' message/remote/RemoteMessagingSystem.cpp:164: instantiated from here /home/v0d0i/libraries/boost_1_42_0/include/boost/archive/basic_xml_iarchive.hpp:70: error: no matching function for call to 'assertion_failed(mpl_::failed************ boost::serialization::is_wrapper<Message*>::************)' make[2]: *** [RemoteMessagingSystem.o] Error 1 make[2]: Leaving directory `/home/v0d0i/workspace/trunk1/src/base' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/v0d0i/workspace/trunk1/src' make: *** [all-recursive] Error 1 It seems like there is something wrong with the Wrapper although i have defined the following Wrappers: BOOST_CLASS_IS_WRAPPER(Message*); BOOST_CLASS_IS_WRAPPER(Message* const); BOOST_CLASS_IS_WRAPPER(Message); I also used the boost::serialization::make_nvp for the name-value pairs purpose of xml archive. What am i doing wrong? Thanks, Ngoc Mai

2 10

[utility] swap
by Robert Jones 05 Mar '10

05 Mar '10

I'm sure I'm being a bit dense here, but apart from handling native arrays what does boost::swap give you that std::swap doesn't? Thanks, - Rob.

4 5

BGL using the VertexIndexMap with VertexList=listS
by khaled 05 Mar '10

05 Mar '10

Hello I'm using the BGL and wandering to use dijkstra_short_path I this code struct vertex1 { int step; int node; int MPI; int OMP; int acc; }; struct edge1 { double cost; }; typedef boost::adjacency_list< boost::listS, boost::listS, boost::directedS, vertex1, edge1> Graph; Graph g; std::vector<double> distances(num_vertices(g)); dijkstra_shortest_paths(g, from, weight_map(get(&edge1::cost, g)) .distance_map(make_iterator_property_map(distances.begin(), get(vertex_index, g)))); I have a problem with the VertexIndexMap (I guess is this the problem :) ) as I'm using a adjacenty_list with VertexList=listS does not have an internal vertex_index property. So, Please can you help me to fix this ? what is the VertexIndexMap to be used with the listS VertexList ?? Thanks -- K.H

2 1

Waiting for a thread on another thread
by Daniel Lidström 05 Mar '10

05 Mar '10

Hello! I have created a simple file logger using the active object pattern. This is intended to offload logging operations to a separate thread, to not disturb the main thread with time- consuming file operations. It works very nicely with boost::thread, but now I have a new desired functionality. The file logger keeps an internal queue of strings that need to be written to file. I want to know, from the main thread, when this queue is empty. To illustrate I have created a minimal, hopefully compilable, sample that shows what I am trying to achieve. #include <boost/thread/thread.hpp> #include <boost/thread/xtime.hpp> #include <boost/thread/condition.hpp> #include <boost/foreach.hpp> #include <boost/tr1/functional.hpp> #include <vector> #include <fstream> #include <string> class FileLogger { typedef boost::unique_lock<boost::mutex> unique_lock; std::ofstream logfile; bool running; boost::thread running_thread; boost::mutex mutex; boost::condition_variable condition; std::vector<std::string> queue; public: FileLogger(const std::string& filename) : logfile(filename.c_str()) , running(true) , running_thread(std::tr1::bind(&FileLogger::Run, this)) { } ~FileLogger() { running = false; condition.notify_one(); running_thread.join(); } void Log(const std::string& s) { unique_lock lock(mutex); queue.push_back(s); } void WaitForQueueFlush() { // ?? } private: void Run() { while( running ) { std::vector<std::string> copyList; { // wait 0.1 s here, for queue to fill up boost::xtime sleepTime; boost::xtime_get(&sleepTime, boost::TIME_UTC); sleepTime.nsec += 100*1000*1000; unique_lock lock(mutex); condition.timed_wait(lock, sleepTime); // fetch the current message list copyList.swap(queue); } BOOST_FOREACH(const std::string& s, copyList) { logfile << s << '\n'; } logfile.flush(); } } }; int main() { FileLogger fileLogger("logfile.txt"); fileLogger.Log(std::string(1000*1000, '*')); // this call should wait, synchronously, until // current queue has been flushed to disk fileLogger.WaitForQueueFlush(); } So the problem is to implement FileLogger::WaitForQueueFlush. How can I signal the main thread when the queue has been emptied? Thanks in advance! Regards, Daniel Lidström Stockholm, Sweden

2 1

[interprocess] Sharing data in a peer-to-peer fashion
by Brett Gmoser 05 Mar '10

05 Mar '10

Hello, I'm working on what seems to be a fairly interesting problem, and I'm looking for any interprocess experts to lend any advice. After reading all of the documentation for Interprocess, it seems that all of the examples work with a "one parent, many children" type model. The problem that I am having is that I am working with a "many children, no parent" model. My processes are spawned by the web server (and FastCGI), so there is no easy way to modify that code to manage the interprocess data. At least one instance of my program is meant to stay alive forever, so I'm not very concerned about deleting the interprocess data at any time. My problem however is the mutexes, and the stale locks that result if one of the processes crashes or exits abnormally. It seems that once an instance of my application crashes, no other instance may ever get a lock on the shared mutex, because the original process still holds a lock. I've thought of a few different ideas to solve this problem: * Using an interprocess shared_ptr for the mutex and data, with a custom deleter to remove all instances once the last application is exiting and the use count is zero. However this suffers from the same problem - it seems the use count is never decremented when the application exits abnormally (for example, kill -9, crash, or CTRL+C). * Using a "heart beat" type system to keep track of processes. My idea was to do something like this: Keep an interprocess associative array of process ID's mapped to last heartbeat time, with each process updating it's own heartbeat. At the same time, every other process (since there is no one parent process) must keep track of the heartbeats, and remove processes from the array which have not responded with a heartbeat in awhile. The problem is this - what about the mutex for the associative array? And the locks that another application might have on them? I wind up back at the point that I started at. The stale locks are destroyed if the mutex is deleted (via boost::interprocess::named_mutex::delete), and I can detect the stale lock pretty reliably with a timed try lock (wait 30 seconds or so to aquire the lock, if it can never be acquired then there is obviously a problem). But what of thread safety, since I now have to delete the lock and re-create it, where other processes might wind up doing the same thing? And what if those other processes are also trying to obtain a lock on the mutex at the same time I'm deleting and recreating it? Those are pretty much the only ideas I've had. Does anybody have anything better? Maybe some of you have tackled a similar problem? Thanks!

1 0

question about a asio example
by pedro chaparro 05 Mar '10

05 Mar '10

Hi, inthe example of a chat server usin asio library, http://www.boost.org/doc/libs/1_42_0/doc/html/boost_asio/example/chat/chat_… there is a fuction socket() , that socket_ value the function returns i use to each new connection with the server right? so if i have n users , i'll have n opened sockets?? i'll aprecciate to get me clear on this thanks tcp::socket& socket() { return socket_; } -- Ing Pedro Alonso Chaparro Valero Ciudad Politecnica de la Innovación iTEAM - Mobile Communications Group Polytechnic University of Valencia C\ Camino de Vera S/N, Edificio 8G 46022 Valencia, Spain

2 1

lambda::bind and abstract class - how to?
by Archie14 05 Mar '10

05 Mar '10

Here is the sample that fails to compile. Complain is about foo being abstract in IA. Any advise of how to make it work will be greatly appreciated. #include <boost/lambda/lambda.hpp> #include <boost/lambda/bind.hpp> #include <boost/ptr_container/ptr_vector.hpp> class IA { public: virtual int foo() const = 0; }; class B : public IA { public: B (int a) : _a(a) {}; int foo() const {return _a;} private: int _a; }; int _tmain(int argc, _TCHAR* argv[]) { boost::ptr_vector<IA> lst; lst.push_back(new B(1)); lst.push_back(new B(2)); boost::ptr_vector<IA>::iterator it = std::find_if (lst.begin(), lst.end(), boost::lambda::bind(&IA::foo, boost::lambda::_1) == 1); return 0; }

2 2