December 2013 - Boost-users

one memory leak
by XIN LI 06 Dec '13

06 Dec '13

Hi there, I use boost version 1.49.0. I meet a memory leak for below sentences: boost::filesystem::path tmp = input.stem(); string strDir = tmp.normalize().string(); Valgrind report memory leak as below: ==18299== 56 bytes in 1 blocks are still reachable in loss record 26 of 61 ==18299== at 0x16DF68B8: malloc (vg_replace_malloc.c:270) ==18299== by 0x3F64FEC: operator new(unsigned long) (in /home/lixin/work/prj/hotfix/13q2_ehf1/src/prosim/dist/Debug_64/GNU-Linux-x86/nanospice) ==18299== by 0x3F5B8F1: stlp_std::priv::__acquire_category(char const*&, char*, _Locale_name_hint*, char const* (*)(char const*, char*, _Locale_name_hint*, int*), void* (*)(char const*, _Locale_name_hint*, int*), char const* (*)(char*), stlp_std::hash_map<stlp_std::basic_string<char, stlp_std::char_traits<char>, stlp_std::allocator<char> >, stlp_std::pair<void*, unsigned long>, stlp_std::hash<stlp_std::basic_string<char, stlp_std::char_traits<char>, stlp_std::allocator<char> > >, stlp_std::equal_to<stlp_std::basic_string<char, stlp_std::char_traits<char>, stlp_std::allocator<char> > >, stlp_std::allocator<stlp_std::pair<stlp_std::basic_string<char, stlp_std::char_traits<char>, stlp_std::allocator<char> > const, stlp_std::pair<void*, unsigned long> > > >**, int*) (in /home/lixin/work/prj/hotfix/13q2_ehf1/src/prosim/dist/Debug_64/GNU-Linux-x86/nanospice) ==18299== by 0x3F5BD04: stlp_std::priv::__acquire_ctype(char const*&, char*, _Locale_name_hint*, int*) (in /home/lixin/work/prj/hotfix/13q2_ehf1/src/prosim/dist/Debug_64/GNU-Linux-x86/nanospice) ==18299== by 0x3F3F502: stlp_std::_Locale_impl::insert_ctype_facets(char const*&, char*, _Locale_name_hint*) (in /home/lixin/work/prj/hotfix/13q2_ehf1/src/prosim/dist/Debug_64/GNU-Linux-x86/nanospice) ==18299== by 0x3F3DC3C: stlp_std::locale::locale(char const*) (in /home/lixin/work/prj/hotfix/13q2_ehf1/src/prosim/dist/Debug_64/GNU-Linux-x86/nanospice) ==18299== by 0x458FFFE: boost::filesystem3::path::wchar_t_codecvt_facet() (in /home/lixin/work/prj/hotfix/13q2_ehf1/src/prosim/dist/Debug_64/GNU-Linux-x86/nanospice) ==18299== by 0x4591806: boost::filesystem3::path::begin() const (in /home/lixin/work/prj/hotfix/13q2_ehf1/src/prosim/dist/Debug_64/GNU-Linux-x86/nanospice) ==18299== by 0x4592844: boost::filesystem3::path::m_normalize() (in /home/lixin/work/prj/hotfix/13q2_ehf1/src/prosim/dist/Debug_64/GNU-Linux-x86/nanospice) ==18299== by 0x21019EB: boost::filesystem3::path::normalize() (path.hpp:430) Because my program need run thousands of time and very sensitive for the memory leak. Could anyone help to suggest one way to avoid such memory leak? Thanks a lot! Best wishes, Andy Li

2 1

[log] Sink format is ignored
by Igor R 06 Dec '13

06 Dec '13

Hello, I'm setting up a sink like this: boost::shared_ptr<sinks::text_file_backend> backend = boost::make_shared<sinks::text_file_backend> ( keywords::file_name = fileName, keywords::rotation_size = maxSize, keywords::format = "[%TimeStamp%][%LineID%]: %Message%", keywords::auto_flush = true ); typedef sinks::synchronous_sink<sinks::text_file_backend> file_sink; boost::shared_ptr<file_sink> sink = boost::make_shared<file_sink>(backend); sink->set_filter(!expr::has_attr(serv_attr)); logging::core::get()->add_sink(sink); logging::add_common_attributes(); However, it seems that keywords::format parameter is just ignored. What am I doing wrong?

3 3

Re: [Boost-users] [graph] Compilation problem using depth_first_visit with a listS as a VertexListS
by Simon Bozonnet 06 Dec '13

06 Dec '13

Thanks Alex, I finally understood that using std::vec and std::list is different in the way that we need to specify an index for the edges/nodes while using such a routine like depth_first_visit. Thanks for your help! Simon > From that I isolated the following code which compiles without any problem: > > typedef adjacency_list < vecS, vecS, bidirectionalS, GraphvizVertexProperty,GraphvizEdgeProperty, GraphvizGraphProperty > Graph; > ... > depth_first_visit(_graph2, loop_head, default_dfs_visitor(), make_iterator_property_map(reachable_from_head.begin(),get(vertex_index, _graph2), c)); > So this works fine! > Then, I would like to change the type of graph, so more precisely, to use a std::list (or listS) instead of a std..vector (vecS) for the VertexListS > So the first line declaration becomes: > typedef adjacency_list < vecS, listS, bidirectionalS,GraphvizVertexProperty,GraphvizEdgeProperty, GraphvizGraphProperty > Graph; > //only the 2nd argument changes > But then when I compile the code, I get 1 error during the compilation as described hereafter. "If the VertexList of the graph is vecS, then the graph has a builtin vertex indices accessed via the property map for the vertex_index_t property." (http://www.boost.org/doc/libs/1_55_0/libs/graph/doc/adjacency_list.html <http://www.boost.org/doc/libs/1_55_0/libs/graph/doc/adjacency_list.html>) So when you use listS instead you have to supply your own vertex index map.

1 0

Fwd: support for bjam projects in IDE
by Witold E Wolski 06 Dec '13

06 Dec '13

Hi, I will have to work on bjam based project (rather big). I normally use QT Creator for programming and work with cmake based project. I could't find any hint how to set up an bjam project to work with QT neither information about any IDE which supports bjam. Do boost developer use vim and debug using gdb only? Actually the other guys on the project wrote some python scripts which parse bjam output to generated vcproj files :-@ I would be happy to hear from someone who are happy using bjam with qtcreator or any other IDE for larger projects and how to set this up. All my searches so far regarding bjam return almost exclusively FUD (http://en.wikipedia.org/wiki/Fear,_uncertainty_and_doubt) but looking at the Jamroot.jam file of the project (more than 1000 lines) I am inclinded to believe it. ;-( -- Witold Eryk Wolski -- Witold Eryk Wolski

2 2

Await the result of several futures to chain another task
by Kaspar Fischer 06 Dec '13

06 Dec '13

Hi, If I have two futures f1 and f2, can I compose them to a future that completes with the result of f1 and f2 when both f1 and f2 have finished? I'd like to have a future so that I can use the then()<http://www.boost.org/doc/libs/1_55_0/doc/html/thread/synchronization.html#t…>method to run another transformation. I see method wait_for_all()<http://www.boost.org/doc/libs/1_55_0/doc/html/thread/synchronization.html#t…>but this method blocks. Best, Kaspar

2 2

[serialization] How to deserialize only a part of an XML archive?
by luca 05 Dec '13

05 Dec '13

This is my scenario: I've got a shared library in which these classes are defined: class base { ... serialize() ... } class parent: public base { ... serialize(...) {...serialize base. and itselft too!} ... } the shared lib is used by a computational application that re-implements class parent in an extended version - this because the app is a Qt application that does visualization etc...: class grandParent: public parent { //it also defines some specific serialization capabilities: serialize (...) {serialize parent and also itself} } Now if I serialize into an XML archive an instance of grandParent I am clarly able to deserialize it in a grandParent object, but what about to deserialize it as a parent object? if I serialize from grandParent I would get something like: <grandParent> <parent> <base> </base> </parent> </grandParent> I would need to deserialize only the parent part: <parent> <base> </base> </parent> In practice I would like to be able to deserialize an XML archive to a base class instead that to the derived. But in this case the derived class is not accessible by boost (that thus complains about it) In this way I would get the better inter-operability between my application and the underlaying lib, that must be able to run without the app itself... Many thanks! Luca

1 0

Re: [Boost-users] [graph] Compilation problem using depth_first_visit with a listS as a VertexListS
by alex 05 Dec '13

05 Dec '13

> From that I isolated the following code which compiles without any problem: > > typedef adjacency_list < vecS, vecS, bidirectionalS,GraphvizVertexProperty,GraphvizEdgeProperty, GraphvizGraphProperty > Graph; > ... > depth_first_visit(_graph2, loop_head, default_dfs_visitor(), make_iterator_property_map(reachable_from_head.begin(),get(vertex_index, _graph2), c)); > So this works fine! > Then, I would like to change the type of graph, so more precisely, to use a std::list (or listS) instead of a std..vector (vecS) for the VertexListS > So the first line declaration becomes: > typedef adjacency_list < vecS, listS, bidirectionalS,GraphvizVertexProperty,GraphvizEdgeProperty, GraphvizGraphProperty > Graph; > //only the 2nd argument changes > But then when I compile the code, I get 1 error during the compilation as described hereafter. "If the VertexList of the graph is vecS, then the graph has a builtin vertex indices accessed via the property map for the vertex_index_t property." (http://www.boost.org/doc/libs/1_55_0/libs/graph/doc/adjacency_list.html) So when you use listS instead you have to supply your own vertex index map.

1 0

[log] ThreadID attribute is empty
by Igor R 05 Dec '13

05 Dec '13

Hello, I'm formatting a sink like this: sink->set_formatter ( expr::stream << expr::attr<boost::posix_time::ptime>("TimeStamp") << " T:" << expr::attr<boost::thread::id>("ThreadID") << ":" << expr::smessage ); logging::add_common_attributes(); For some reason ThreadID is empty. I tried to register the attribute explicitly: logging::core::get()->add_global_attribute("ThreadID", logging::attributes::current_thread_id()); but it doesn't help. The program is mutithreaded and linked with mt static libs. Thanks.

1 0

Having trouble compiling interval_map
by Daniel Rubi 05 Dec '13

05 Dec '13

Hi, I'm trying to implement an interval map where the keys are close_interval<long> an the values are set<string>. I modified the party example according to my needs, and came up with this short code: #include <boost/icl/interval.hpp> #include <boost/icl/interval_map.hpp> #include <boost/compressed_pair.hpp> using namespace std; using namespace boost::icl; typedef std::set<string> ids; int main(int argc, char* argv[]) { ids ids1; ids1.insert("T1"); ids ids2; ids2.insert("T2"); interval_map<closed_interval<long>, ids> mymap; closed_interval<long> i1 = closed_interval<long>(2, 7); closed_interval<long> i2 = closed_interval<long>(3, 8); mymap.insert(std::pair<closed_interval<long>, ids>(i1,ids1)); mymap.insert(std::pair<closed_interval<long>, ids>(i2,ids2)); return 0; } However, I get this compilation error: call of overloaded ‘insert(std::pair<boost::icl::closed_interval<long int>, std::set<std::basic_string<char> > >)’ is ambiguous mymap.insert(std::pair<closed_interval<long>, ids>(i2,ids2)); This is after getting compiation errors trying ti use: mymap += make_pair(i1,ids1); Any idea what's wrong?

1 0

[Any] inconsistent return types of any_cast overloads
by Arne Vogel 04 Dec '13

04 Dec '13

Hello Boost Users, Code like the following used to work with Boost 1.53 on x86-64 Linux using GCC 4.7 and -std=c++11: boost::any makeVec() { return std::vector<int>{ 1, 2, 3 }; } void foo() { const auto& vec = boost::any_cast<std::vector<int>>(makeVec()); for ( auto value : vec ) { // ... } } The same code is likely to cause a segmentation violation using Boost 1.55. According to the online documentation *1, any_cast has two overloads accepting reference arguments: template<typename ValueType> ValueType any_cast(any&); template<typename ValueType> ValueType any_cast(const any&); If these were indeed the only overloads accepting reference arguments, there is no reason why the example code shouldn't work. ValueType any_cast(const any&) should be the preferred overload and it returns a vector<int> temporary whose lifetime is then extended to the end of the block by being bound to a local const reference, as required by the C++ standard. However, in any.hpp, I discovered the following overload, which is only declared if the macro BOOST_NO_CXX11_RVALUE_REFERENCES is not set. I suppose this overload was added since 1.53, or that, previously, it was by default omitted. template<typename ValueType> ValueType&& any_cast(any&&); This has higher precedence binding to the temporary returned from makeVec, and returns an rvalue reference to the vector<int> contained in said temporary, which is then implicitly converted to a const lvalue reference as by the usual language rules. Since the lifetime of the temporary returned by makeVec is NOT extended - this only works for temporaries directly bound to a local const reference, whereas the result of makeVec is "hidden" inside the call to any_cast - the temporary is destroyed before the start of the for loop, which renders vec invalid and the code a denizen of undefined behavior land. So much for the analysis, now for my actual question. :-) Why does the new overload return by reference when the old ones return by value? The rationale for the old overloads was, I believe, "you should have to ask for a reference if you want one", which makes sense given precisely that using references tends to be more error prone if temporaries are involved. So this would be the very reason why Any does not declare: template<typename ValueType> const ValueType& any_cast(const any&); even though it could be implemented as easily as the actual overload. If I really wanted to get an const reference out of any_cast, I could write something like boost::any_cast<const ValueType&>(var); I can't think of a reason to be inconsistent in this regard and return a reference if a temporary is passed to any_cast, instead of requiring a user who really wants any_cast to return an rvalue reference to write any_cast<ValueType&&>(...). That the overload breaks existing code and (due to being undocumented and not checked at compile time *2), silently so, whereas using ValueType any_cast(any&&) would at least fail to do any harm, is, in my opinion, another reason to be sceptical of the current solution. Furthermore, there is a usability reason for preferring ValueType(any &&): If ValueType &&(any &&) is declared and I do NOT want a reference, I need to write more verbose and inelegant code than in the opposite case: // Name the temporary const auto &vecAny = makeVec(); // Now, ValueType&& any_cast(any&&) is eliminated from candidate set (no conversion from an lvalue to an rvalue reference) const auto& vec = boost::any_cast<std::vector<int>>(vecAny); Note that the const is accidental here, using a mutable lvalue would work as well. Looking forward to your opinions, Arne Vogel *1) http://www.boost.org/doc/libs/1_55_0/doc/html/boost/any_cast_idp25247624.ht… *2) A compiler may issue a warning. However, I found that GCC 4.8.1 does not warn, even with -Wall and the definition of any_cast being known at the point of use.

5 6

boost::regex Character set including braces
by Tobias 04 Dec '13

04 Dec '13

Dear boost-users, in the post http://stackoverflow.com/questions/20351369/regular-expression-cant-find-seq uence-of-numbers I discovered that if one uses an closing bracket as the first character of a character class the character class includes this bracket. This works with the standard setting of boost::regex (i.e., perl-regular expressions) but it is not documented in the manual page http://www.boost.org/doc/libs/1_55_0/libs/regex/doc/html/boost_regex/syntax/ perl_syntax.html#boost_regex.syntax.perl_syntax.character_sets Is this an undocumented feature, a bug or did I misinterpret something in the manual? The version information in version.hpp is: #define BOOST_LIB_VERSION "1_53" Info: I tried this specification of the bracket in the character set because I am used to it from emacs regular expressions. In the remainder there follows a program for reproduction of the phenomenon: #include <string> #include <boost/regex.hpp> #include <iostream> using namespace std; int main() { std::cout << "main()\n"; string s = "98[2]00"; string exp = "^[][0-9]*$"; const boost::regex e(exp); bool isSequence = boost::regex_match(s,e); //isSequence is boolean and should be equal to 1 cout << isSequence << endl; return 0; } /** Local Variables: compile-command: "g++ -g test.cc -o test.exe -lboost_regex-mt; ./test.exe" End: */ Best regards / Mit freundlichen Grüßen Tobias Zawada Engineering / Software Development ______________________________ ITI Events We say Thank you! for an outstanding ITI Symposium 2013! Make sure you save the date for next year: November 3-5, 2014 www.itisim.com ITI GmbH Schweriner Straße 1 01067 Dresden Germany T +49 (351) 26050 147 F +49 (351) 260 50 155 mailto:Zawada@itisim.com Managing Director/Geschäftsführer Jens O. Schindler · Dr. Andreas Uhlig Commercial Register/Handelsregister Amtsgericht Dresden · HRB 2331 USt.-IdNr.: DE140297717 Diese E-Mail ist streng vertraulich und nur für den Empfänger bestimmt. Sie kann Informationen enthalten, die rechtliche, geschäftliche oder andere Vergünstigungen bedeuten. Sollten Sie nicht der ausdrückliche Empfänger sein, dürfen Sie diese Informationen nicht benutzen, veröffentlichen oder vervielfältigen. The information in this e-mail is strictly confidential, is intended solely for the addressee, and may be privileged. If you are not the intended addressee you must not disclose, print or take action based upon this transmission and should delete it from your system.

2 1

[Filesystem] [Locale] Unicode filenames doc example error
by Egor Tensin 04 Dec '13

04 Dec '13

Why doesn't this simple code from the Boost.Locale docs work? #include <boost/filesystem.hpp> #include <boost/filesystem/fstream.hpp> #include <boost/locale.hpp> #include <exception> #include <iostream> int main() { try { std::locale::global(boost::locale::generator().generate("")); boost::filesystem::path::imbue(std::locale()); boost::filesystem::ofstream hello("привет.txt"); } catch (const std::exception& e) { std::cerr << e.what() << std::endl; } return 0; } Compiled with >cl /EHsc /D _UNICODE /D UNICODE b.cpp /I%BOOST_ROOT% /MD /link /LIBPATH:%BOOST_ROOT%\stage\lib it gives me >b.exe boost::filesystem::path codecvt to wstring: error Could someone please explain what's the problem there and how do I handle Unicode file names on Windows using Boost.Filesystem & Boost.Locale? I'm using Boost 1.55.0 & compiling with MS Visual C++ 2010 compiler v.16.00.30319.01.

1 0

[log] connecting a logger to a specific sink
by Igor R 04 Dec '13

04 Dec '13

Hello, I'm integrating Boost.Log to my project, and have a couple of questions. First of all, I'd like each instance of some class to write its log to a separate rotating file. I understand how to setup such a sink, but can't find a way to connect a particular source (logger) to the sink. Is it possible? Besides, when setting up the file, I configure it with: keywords::format = "[%TimeStamp%]: %Message%" but %TimeStamp% doesn't seem to get recognized, so I get empty [ ] in the file. Do I miss some header file? Thanks.

2 4

Print boost graph with custom vertex label
by Saha, Shambwaditya 04 Dec '13

04 Dec '13

Hi All, I want to print a Boost Graph having custom(string) vertex labels instead of the default vertex numbering label 0,1,2... I have initialized the graph as : typedef adjacency_list <vecS, vecS, directedS, property<vertex_name_t,string>> Graph; Graph g; set<string> names; map<string,Graph::vertex_descriptor> vertex ; for(auto it = names.begin() ; it != names.end(); ++it ) vertex[*it] = add_vertex(*it,g) ; Now how should I print this graph such that I get edges of the form abc -> xyz; instead of 1->2; Thanks and Regards, Shambwaditya

1 0

[phoenix][spirit] catch_all issues in V3
by Igor R 03 Dec '13

03 Dec '13

Hello, The following code doesn't compile (MSVC10, Boost 1.55). If I modify "catch_all" line like this: query = id[phx::try_[qi::_val = phx::bind(&string2test, qi::_1)].catch_all [ qi::_pass = false ], qi::_val = qi::_val ]; ...it compiles, but doesn't seem to invoke the semantic action in runtime. It there a workaround that would both compile and work? Thanks! #define BOOST_SPIRIT_USE_PHOENIX_V3 #include <string> #include <boost/spirit/include/qi.hpp> #include <boost/spirit/include/phoenix_core.hpp> #include <boost/spirit/include/phoenix_operator.hpp> #include <boost/phoenix/bind/bind_function.hpp> #include <boost/phoenix/statement/try_catch.hpp> namespace qi = boost::spirit::qi; namespace phx = boost::phoenix; struct test { }; test string2test(const std::string &str) { if (str == "throw") throw "exception!"; return test(); } template <typename Iterator = std::string::const_iterator> struct test_parser : qi::grammar<Iterator, test(), qi::blank_type> { test_parser() : test_parser::base_type(query) { query = id[phx::try_[qi::_val = phx::bind(&string2test, qi::_1)].catch_all [ qi::_pass = false ] ]; id = qi::lexeme[*qi::char_("a-zA-Z_0-9")]; } qi::rule<Iterator, test(), qi::blank_type> query; qi::rule<Iterator, std::string(), qi::blank_type> id; }; int main() { const std::string input = "qwerty"; test_parser<> p; test output; bool ok = phrase_parse(input.begin(), input.end(), p, qi::blank, output); } // 1> boost\boost\proto\traits.hpp(297): error C2338: 0 == Expr::proto_arity_c 1> boost\boost\mpl\eval_if.hpp(63) : see reference to class template instantiation 'boost::proto::result_of::value<Expr>' being compiled 1> with 1> [ 1> Expr=boost::phoenix::try_catch_actor<boost::proto::exprns_::basic_expr<boost::phoenix::tag::try_catch,boost::proto::argsns_::list2<boost::phoenix::try_catch_actor<boost::proto::exprns_::basic_expr<boost::phoenix::tag::try_catch,boost::proto::argsns_::list1<boost::phoenix::actor<boost::proto::exprns_::expr<boost::proto::tagns_::tag::assign,boost::proto::argsns_::list2<boost::proto::exprns_::expr<boost::proto::tagns_::tag::terminal,boost::proto::argsns_::term<boost::spirit::attribute<0>>,0>,boost::phoenix::actor<boost::proto::exprns_::basic_expr<boost::phoenix::detail::tag::function_eval,boost::proto::argsns_::list2<boost::proto::exprns_::basic_expr<boost::proto::tagns_::tag::terminal,boost::proto::argsns_::term<boost::phoenix::detail::function_ptr<1,test,test (__cdecl *)(const std::string &)>>,0>,boost::phoenix::actor<boost::spirit::argument<0>>>,2>>>,2>>>,1>>,boost::proto::exprns_::basic_expr<boost::phoenix::tag::catch_all,boost::proto::argsns_::list1<boost::phoenix::actor<boost::proto::exprns_::expr<boost::proto::tagns_::tag::assign,boost::proto::argsns_::list2<boost::proto::exprns_::basic_expr<boost::proto::tagns_::tag::terminal,boost::proto::argsns_::term<boost::phoenix::argument<3>>,0>,boost::phoenix::actor<boost::proto::exprns_::expr<boost::proto::tagns_::tag::terminal,boost::proto::argsns_::term<bool>,0>>>,2>>>,1>>,2>> 1> ]

2 3

[graph] Compilation problem using depth_first_visit with a listS as a VertexListS
by Simon Bozonnet 03 Dec '13

03 Dec '13

Hi There, I am struggling with BOOST while trying to link them with my code. Actually I start while using "loops_dfs.cpp" as an example. >From that I isolated the following code which compiles without any problem: typedef adjacency_list < vecS, vecS, bidirectionalS,GraphvizVertexProperty,GraphvizEdgeProperty, GraphvizGraphProperty > Graph; Graph _graph2; typedef color_traits < default_color_type > Color; typedef graph_traits < Graph >::vertex_descriptor VertexType; typedef graph_traits < Graph >::edge_descriptor EdgeType; EdgeType back_edge;//ok I know it is empty but it just to try the compilation!! VertexType loop_head, loop_tail; loop_tail = source(back_edge, _graph2); loop_head = target(back_edge, _graph2); std::vector < default_color_type > reachable_from_head(num_vertices(_graph2), Color::white()); default_color_type c; depth_first_visit(_graph2, loop_head, default_dfs_visitor(), make_iterator_property_map(reachable_from_head.begin(),get(vertex_index, _graph2), c)); So this works fine! Then, I would like to change the type of graph, so more precisely, to use a std::list (or listS) instead of a std..vector (vecS) for the VertexListS So the first line declaration becomes: typedef adjacency_list < vecS, listS, bidirectionalS,GraphvizVertexProperty,GraphvizEdgeProperty, GraphvizGraphProperty > Graph; //only the 2nd argument changes But then when I compile the code, I get 1 error during the compilation as described hereafter. Since I am fighting for more than one day now on this problem, I would **really** appreciate if someone can give me more explanations and/or advice to solve my problem... Thanks in advance, Simon 1>boost/property_map.hpp(351) : error C2679: binary '+' : no operator found which takes a right-hand operand of type 'boost::detail::error_property_not_found' 1>(or there is no acceptable conversion) 1> C:\Program Files (x86)\Microsoft Visual Studio 8\VC\include\vector(366): could be 'std::_Vector_iterator<_Ty,_Alloc> std::_Vector_iterator<_Ty,_Alloc>::operator +(__w64 int) const' 1> with 1> [ 1> _Ty=boost::default_color_type, 1> _Alloc=std::allocator<boost::default_color_type> 1> ] 1> while trying to match the argument list '(const std::_Vector_iterator<_Ty,_Alloc>, boost::detail::error_property_not_found)' 1> with 1> [ 1> _Ty=boost::default_color_type, 1> _Alloc=std::allocator<boost::default_color_type> 1> ] 1> \inc\boost/property_map.hpp(351) : while compiling class template member function 'boost::default_color_type boost::iterator_property_map<RandomAccessIterator,IndexMap,T,R>::operator [](void *) const' 1> with 1> [ 1> RandomAccessIterator=std::_Vector_iterator<boost::default_color_type,std::allocator<boost::default_color_type>>, 1> IndexMap=boost::adj_list_vertex_property_map<boost::adjacency_list<boost::listS,boost::listS,boost::bidirectionalS,LatticeNode,LatticeArc>,boost::detail::error_property_not_found,boost::detail::error_property_not_found &,boost::vertex_index_t>, 1> T=boost::default_color_type, 1> R=boost::default_color_type & 1> ] 1> ..\test.cpp(18) : see reference to class template instantiation 'boost::iterator_property_map<RandomAccessIterator,IndexMap,T,R>' being compiled 1> with 1> [ 1> RandomAccessIterator=std::_Vector_iterator<boost::default_color_type,std::allocator<boost::default_color_type>>, 1> 1> IndexMap=boost::adj_list_vertex_property_map<boost::adjacency_list<boost::listS,boost::listS,boost::bidirectionalS,LatticeNode,LatticeArc>,boost::detail::error_property_not_found,boost::detail::error_property_not_found &,boost::vertex_index_t>, 1> T=boost::default_color_type, 1> R=boost::default_color_type & 1> ]

1 0

[ANN] C++Now 2014: 5 Days to Submissions Deadline
by Boris Kolpackov 03 Dec '13

03 Dec '13

Hi, Only 5 days left before the submissions deadline for C++Now 2014! C++Now is a general C++ conference for C++ experts and enthusiasts. It is not specific to any library/framework or compiler vendor and has three tracks with presentations ranging from hands-on, practical tutorials to advanced C++ design and development techniques. For more information about C++Now, see the conference's website: http://cppnow.org/about/ Have you learned something interesting about C++ (e.g., a new technique possible in C++11)? Or maybe you have implemented something cool related to C++ (e.g., a C++ library)? If so, consider sharing it with other C++ enthusiasts by giving a talk at C++Now 2014. For more information on possible topics, formats, etc., see the call for submissions: http://cppnow.org/2013/10/21/2014-call-for-submissions/ Boris

1 0

Re: [Boost-users] [Thread] Using thread accross libraries
by Sébastien Gallou 02 Dec '13

02 Dec '13

> > You are linking to the static Boost.Thread library. This means that > you've got two distinct Boost.Thread libraries loaded into your process, > one into the executable and one in the DLL. Thanks Lars, you solve my problem. Effectively, I was linking with the static lib. When I was looking how to build shared libs, I found in the boost documentation that shared build should be the default. http://www.boost.org/boost-build2/doc/html/bbv2/tutorial/linkage.html I'm under Windows, and without add link=shared to the b2 command line, the build is static. Next was the problem to link my app with boost. It was necessary to define BOOST_ALL_DYN_LINK in boost\config\user.hpp, while the comment just above says that it should not be needed for boost_thread : "Forces all libraries [...] to be linked as dll's rather than static libraries on Microsoft Windows. [...] Note that there may be some libraries that can only [...] be dynamically linked (Boost.Threads for example), in these cases this macro has no effect." Is something I didn't well understand ? Anyway, it now works, Thank you again Lars, Sebastien Gallou

1 0

Re: [Boost-users] Boost-users Digest, Vol 3643, Issue 1
by Damien Kick 02 Dec '13

02 Dec '13

On Dec 1, 2013, at 11:00 AM, boost-users-request(a)lists.boost.org wrote: > Date: Sun, 1 Dec 2013 20:38:32 +0530 > From: Rahul Mathur <srivmuk(a)gmail.com> > Subject: [Boost-users] ZeroMQ integration with Boost-v1.54 > Message-ID: > <CALtf4sBn6Dh0WVOgAfaXvSj4obMiMw335w88xDj5vguPOoHMaw(a)mail.gmail.com> […] > BTW, How do I proceed to have ZeroMQ integrated with Boost-v1.54 with > GCC on Linux platform? Here is some code that I've been using to integrate Boost.Asio and 0MQ. The OP question asked for GNU/Linux only but this code has been working for me on both Windows and POSIX. Most of this comes from comments on various Internet forums but integrates together a few things which I personally found less than obvious on my first few attempts. For example, most of the code snippets I found only mentioned POSIX and I had to beat my head against a few walls to get something that worked on both, though YMMV. ==> config.hpp.in <== #ifndef INCLUDE_SHOR_DM_CONFIG_HPP #define INCLUDE_SHOR_DM_CONFIG_HPP // Boost C++ headers #include <boost/filesystem.hpp> // C++ headers #include <string> #ifndef _WIN32_WINNT # define _WIN32_WINNT 0x0601 #endif #ifdef _MSC_VER # define SHOR_DM_CONSTEXPR # define SHOR_DM_DEFAULT_CTOR(expr) expr { } # define SHOR_DM_DEFAULT_COPY(expr) # define SHOR_DM_DEFAULT_DTOR(expr) expr { } # define SHOR_DM_DEFAULT_ASSIGN(expr) # define SHOR_DM_DELETE(expr) expr # define SHOR_DM_NOEXCEPT throw() #else # define SHOR_DM_CONSTEXPR constexpr # define SHOR_DM_DEFAULT_CTOR(expr) expr = default # define SHOR_DM_DEFAULT_COPY(expr) expr = default # define SHOR_DM_DEFAULT_DTOR(expr) expr = default # define SHOR_DM_DEFAULT_ASSIGN(expr) expr = default # define SHOR_DM_DELETE(expr) expr = delete # define SHOR_DM_NOEXCEPT noexcept #endif #endif ==> zee/basic_zmq_fd.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_BASIC_ZMQ_FD_HPP #define INCLUDE_SHOR_DM_ZEE_BASIC_ZMQ_FD_HPP // 0MQ headers #include <zmq.hpp> namespace shor { namespace dm { namespace zee { template<typename Native_handle> struct Basic_zmq_fd { typedef Native_handle native_handle_type; typedef native_handle_type value_type; private: zmq::socket_t& zmq_socket_; Native_handle zmq_fd_; public: explicit Basic_zmq_fd(zmq::socket_t& zmq_socket) : zmq_socket_(zmq_socket), zmq_fd_() { std::size_t zmq_fd_size = sizeof(zmq_fd_); zmq_socket_.getsockopt(ZMQ_FD, &zmq_fd_, &zmq_fd_size); } value_type operator () () const { return zmq_fd_; } }; } // namespace zee } // namespace dm } // namespace shor #endif ==> zee/handler.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_HANDLER_HPP #define INCLUDE_SHOR_DM_ZEE_HANDLER_HPP // DM headers #include "dm/zee/msg.hpp" // Boost C++ headers #include <boost/system/error_code.hpp> #include <boost/function.hpp> // C++ headers #include <functional> namespace shor { namespace dm { namespace zee { typedef std::function< void (const boost::system::error_code& err, shor::dm::zee::Msg& msg) > Async_recv_handler; } // namespace zee } // namespace dm } // namespace shor #endif ==> zee/lossy_service.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_LOSSY_SERVICE_HPP #define INCLUDE_SHOR_DM_ZEE_LOSSY_SERVICE_HPP // Boost C++ headers #include <boost/asio.hpp> namespace shor { namespace dm { namespace zee { template<typename Service> struct Lossy_service : public Service { explicit Lossy_service(boost::asio::io_service& ios) : Service(ios) { } void destory(typename Service::implementation_type& impl) { } }; } // namespace zee } // namespace dm } // namespace shor #endif ==> zee/msg.cpp <== #include "dm/zee/msg.hpp" shor::dm::zee::detail::Msg::Frame::Frame( Frame&& that) : that_(std::move(that.that_)) { } shor::dm::zee::detail::Msg::Frame::Frame( zmq::message_t&& that) : that_(std::move(that)) { } shor::dm::zee::detail::Msg::Frame::Frame( const std::string& the_other) : that_(the_other.size()) { std::copy(the_other.begin(), the_other.end(), static_cast<char*>(that_.data())); } shor::dm::zee::detail::Msg::Frame& shor::dm::zee::detail::Msg::Frame::operator = ( Frame&& that) { that_ = std::move(that.that_); return *this; } shor::dm::zee::detail::Msg::Frame& shor::dm::zee::detail::Msg::Frame::operator = ( const std::string& the_other) { that_.rebuild(the_other.size()); std::copy(the_other.begin(), the_other.end(), static_cast<char*>(that_.data())); return *this; } shor::dm::zee::detail::Msg::Msg( Msg&& that) : frames_(std::move(that.frames_)) { } shor::dm::zee::detail::Msg& shor::dm::zee::detail::Msg::operator = ( Msg&& that) { frames_ = std::move(that.frames_); return *this; } ==> zee/msg.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_MSG_HPP #define INCLUDE_SHOR_DM_ZEE_MSG_HPP // My header #include "dm/zee/msg_fwd.hpp" // DM headers #include "dm/config.hpp" // 0MQ headers #include <zmq.hpp> // C++ headers #include <deque> namespace shor { namespace dm { namespace zee { namespace detail { struct Msg { class Frame { zmq::message_t that_; public: SHOR_DM_DEFAULT_CTOR(Frame()); Frame(Frame&& that); Frame(zmq::message_t&& that); Frame(const std::string& the_other); SHOR_DM_DEFAULT_DTOR(~Frame()); Frame& operator = (Frame&& that); Frame& operator = (const std::string& the_other); operator zmq::message_t& (); operator const zmq::message_t& () const; zmq::message_t& get(); const zmq::message_t& cget() const; const zmq::message_t& get() const; private: SHOR_DM_DELETE(Frame(const Frame& that)); SHOR_DM_DELETE(Frame& operator = (const Frame& that)); }; typedef std::deque<Frame> Frames; private: Frames frames_; public: SHOR_DM_DEFAULT_CTOR(Msg()); Msg(Msg&& that); SHOR_DM_DEFAULT_DTOR(~Msg()); Msg& operator = (Msg&& that); Frames& frames(); const Frames& frames() const; private: SHOR_DM_DELETE(Msg(const Msg& that)); SHOR_DM_DELETE(Msg& operator = (const Msg& that)); }; } // namespace detail typedef detail::Msg Msg; } // namespace zee } // namespace dm } // namespace shor #include "dm/zee/msg.ipp" #endif ==> zee/msg.ipp <== #ifndef INCLUDE_SHOR_DM_ZEE_MSG_IPP #define INCLUDE_SHOR_DM_ZEE_MSG_IPP // My header #include "dm/zee/msg.hpp" inline shor::dm::zee::detail::Msg::Frame::operator zmq::message_t& () { return that_; } inline shor::dm::zee::detail::Msg::Frame::operator const zmq::message_t& () const { return that_; } inline zmq::message_t& shor::dm::zee::detail::Msg::Frame::get() { return that_; } inline const zmq::message_t& shor::dm::zee::detail::Msg::Frame::cget() const { return that_; } inline const zmq::message_t& shor::dm::zee::detail::Msg::Frame::get() const { return that_; } inline shor::dm::zee::detail::Msg::Frames& shor::dm::zee::detail::Msg::frames() { return frames_; } inline const shor::dm::zee::detail::Msg::Frames& shor::dm::zee::detail::Msg::frames() const { return frames_; } #endif ==> zee/msg_fwd.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_MSG_FWD_HPP #define INCLUDE_SHOR_DM_ZEE_MSG_FWD_HPP namespace shor { namespace dm { namespace zee { namespace detail { struct Msg; } // namespace detail typedef detail::Msg Msg; } // namespace zee } // namespace dm } // namespace shor #endif ==> zee/posix_zmq_socket.cpp <== // My header #include "dm/zee/posix_zmq_socket.hpp" // DM headers #include "dm/zee/handler.hpp" // Boost C++ headers #include <boost/bind.hpp> struct shor::dm::zee::detail::Posix_zmq_socket::Read_handler { typedef Posix_zmq_socket::implementation_type Io_type; typedef Posix_zmq_socket::Base_read_handler Base_read_handler; private: Io_type& io_; Base_read_handler read_handler_; Async_recv_handler recv_handler_; public: Read_handler( Io_type& io, Base_read_handler read_handler, Async_recv_handler recv_handler) : io_(io), read_handler_(read_handler), recv_handler_(recv_handler) { } void operator () ( const boost::system::error_code& err, std::size_t /*bytes_transferred*/) { read_handler_(*this, err, recv_handler_); } void operator () () { io_.async_read_some(boost::asio::null_buffers(), *this); } }; shor::dm::zee::detail::Posix_zmq_socket::Posix_zmq_socket( boost::asio::io_service& ios, zmq::context_t& zenv, const int zmq_socket_type) : Zmq_socket_base(ios, zenv, zmq_socket_type), zmq_fd_(Zmq_socket_base::zmq_socket()), io_(ios, zmq_fd_()), io_release_(&io_, Release()) { } void shor::dm::zee::detail::Posix_zmq_socket::async_recv( Async_recv_handler recv_handler) { const auto base_read_handler_fn = boost::bind( &Posix_zmq_socket::read_handler, this, _1, _2, _3); Read_handler read_handler( std::ref(io_), base_read_handler_fn, recv_handler); read_handler(); } void shor::dm::zee::detail::Posix_zmq_socket::close_ext() { this->cancel(); // TODO: reset zmq_fd_ io_.release(); } ==> zee/posix_zmq_socket.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_POSIX_ZMQ_SOCKET_HPP #define INCLUDE_SHOR_DM_ZEE_POSIX_ZMQ_SOCKET_HPP // Base class header #include "dm/zee/zmq_socket_base.hpp" // DM headers #include "dm/zee/handler.hpp" #include "dm/zee/lossy_service.hpp" #include "dm/zee/basic_zmq_fd.hpp" // Boost C++ headers #include <boost/asio/posix/basic_stream_descriptor.hpp> #include <boost/asio/posix/stream_descriptor_service.hpp> #include <boost/asio.hpp> namespace shor { namespace dm { namespace zee { namespace detail { class Posix_zmq_socket : public Zmq_socket_base { using Stream_descriptor_service = boost::asio::posix::stream_descriptor_service; template<typename Service> using Basic_stream_descriptor = boost::asio::posix::basic_stream_descriptor<Service>; // Borrowed from <http://preview.tinyurl.com/n6jh2ne> but it // doesn't seem to work; i.e. I was still finding that the file // descriptor was being closed, causing a crash at shutdown. // Because of this, I (re)introduced Release (see below). typedef Lossy_service<Stream_descriptor_service> service_type; typedef Basic_stream_descriptor<service_type> implementation_type; struct Read_handler; struct Release { void operator () (implementation_type* ptr) { if (ptr->is_open()) { ptr->release(); } } }; Basic_zmq_fd<int> zmq_fd_; implementation_type io_; std::unique_ptr<implementation_type, Release> io_release_; public: Posix_zmq_socket( boost::asio::io_service& ios, zmq::context_t& zenv, int zmq_socket_type); SHOR_DM_DEFAULT_DTOR(~Posix_zmq_socket()); void async_recv(Async_recv_handler handler); void cancel(); private: SHOR_DM_DELETE(Posix_zmq_socket(const Posix_zmq_socket& that)); SHOR_DM_DELETE( Posix_zmq_socket& operator = (const Posix_zmq_socket& that)); virtual void close_ext() override; }; } // namespace detail typedef detail::Posix_zmq_socket Posix_zmq_socket; } // namespace zee } // namespace dm } // namespace shor #include "dm/zee/posix_zmq_socket.ipp" #endif ==> zee/posix_zmq_socket.ipp <== #ifndef INCLUDE_SHOR_DM_ZEE_POSIX_ZMQ_SOCKET_IPP #define INCLUDE_SHOR_DM_ZEE_POSIX_ZMQ_SOCKET_IPP // My header #include "dm/zee/posix_zmq_socket.hpp" inline void shor::dm::zee::detail::Posix_zmq_socket::cancel() { io_.cancel(); } #endif ==> zee/windows_zmq_socket.cpp <== // My header #include "dm/zee/windows_zmq_socket.hpp" // DM headers #include "dm/zee/handler.hpp" // Boost C++ headers #include <boost/system/error_code.hpp> #include <boost/system/system_error.hpp> #include <boost/asio/error.hpp> #include <boost/bind.hpp> struct shor::dm::zee::detail::Windows_zmq_socket::Read_handler { typedef Windows_zmq_socket::implementation_type Io_type; typedef Windows_zmq_socket::Base_read_handler Base_read_handler; private: Io_type& io_; Base_read_handler read_handler_; Async_recv_handler recv_handler_; public: Read_handler( Io_type& io, Base_read_handler read_handler, Async_recv_handler recv_handler) : io_(io), read_handler_(read_handler), recv_handler_(recv_handler) { } void operator () (const boost::system::error_code& err) { read_handler_(*this, err, recv_handler_); } void operator () () { io_.async_wait(*this); } }; shor::dm::zee::detail::Windows_zmq_socket::Windows_zmq_socket( boost::asio::io_service& ios, zmq::context_t& zenv, const int zmq_socket_type) : Zmq_socket_base(ios, zenv, zmq_socket_type), zmq_fd_(Zmq_socket_base::zmq_socket()), io_(ios) { auto zfd_event = CreateEvent( nullptr, // default security settings false, // auto reset false, // initial state: unsignalled nullptr); // anonymous object if (!zfd_event) { const boost::system::error_code an_error_code( GetLastError(), boost::asio::error::get_system_category()); throw boost::system::system_error(an_error_code, "CreateEvent"); } io_.assign(zfd_event); if (WSAEventSelect(zmq_fd_(), zfd_event, FD_READ) == SOCKET_ERROR) { const boost::system::error_code an_error_code( WSAGetLastError(), boost::asio::error::get_system_category()); throw boost::system::system_error( an_error_code, "WSAEventSelect"); } } void shor::dm::zee::detail::Windows_zmq_socket::async_recv( Async_recv_handler recv_handler) { const auto base_read_handler_fn = boost::bind( &Windows_zmq_socket::read_handler, this, _1, _2, _3); Read_handler read_handler( std::ref(io_), base_read_handler_fn, recv_handler); read_handler(); } void shor::dm::zee::detail::Windows_zmq_socket::close_ext() { this->cancel(); // TODO: reset zmq_fd_ io_.close(); } ==> zee/windows_zmq_socket.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_WINDOWS_ZMQ_SOCKET_HPP #define INCLUDE_SHOR_DM_ZEE_WINDOWS_ZMQ_SOCKET_HPP // Base class header #include "dm/zee/zmq_socket_base.hpp" // DM headers #include "dm/zee/basic_zmq_fd.hpp" #include "dm/zee/handler.hpp" // Boost C++ headers #include <boost/asio/windows/object_handle.hpp> #include <boost/asio.hpp> namespace shor { namespace dm { namespace zee { namespace detail { class Windows_zmq_socket : public Zmq_socket_base { typedef boost::asio::windows::object_handle implementation_type; struct Read_handler; Basic_zmq_fd<SOCKET> zmq_fd_; implementation_type io_; public: Windows_zmq_socket( boost::asio::io_service& ios, zmq::context_t& zenv, int zmq_socket_type); SHOR_DM_DEFAULT_DTOR(~Windows_zmq_socket()); void async_recv(Async_recv_handler handler); void cancel(); private: SHOR_DM_DELETE(Windows_zmq_socket(const Windows_zmq_socket& that)); SHOR_DM_DELETE( Windows_zmq_socket& operator = (const Windows_zmq_socket& that)); virtual void close_ext() override; }; } // namespace detail typedef detail::Windows_zmq_socket Windows_zmq_socket; } // namespace zee } // namespace dm } // namespace shor #include "dm/zee/windows_zmq_socket.ipp" #endif ==> zee/windows_zmq_socket.ipp <== #ifndef INCLUDE_SHOR_DM_ZEE_WINDOWS_ZMQ_SOCKET_IPP #define INCLUDE_SHOR_DM_ZEE_WINDOWS_ZMQ_SOCKET_IPP // My header #include "dm/zee/windows_zmq_socket.hpp" inline void shor::dm::zee::detail::Windows_zmq_socket::cancel() { io_.cancel(); } #endif ==> zee/zmq_socket.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_ZMQ_SOCKET_HPP #define INCLUDE_SHOR_DM_ZEE_ZMQ_SOCKET_HPP // My header #include "dm/zee/zmq_socket_fwd.hpp" // Conditional headers #ifndef _MSC_VER # include "dm/zee/posix_zmq_socket.hpp" #else # include "dm/zee/windows_zmq_socket.hpp" #endif // Boost C++ headers #include <boost/asio.hpp> namespace shor { namespace dm { namespace zee { #ifndef _MSC_VER typedef Posix_zmq_socket Zmq_socket; #else typedef Windows_zmq_socket Zmq_socket; #endif } // namespace zee } // namespace dm } // namespace shor #endif ==> zee/zmq_socket_base.cpp <== // My headers #include "dm/zee/zmq_socket_base.hpp" // C++ headers #include <cassert> shor::dm::zee::detail::Zmq_socket_base::Opts::Opts( zmq::socket_t& zmq_socket) : zmq_socket_(zmq_socket) { } shor::dm::zee::detail::Zmq_socket_base::Zmq_socket_base( boost::asio::io_service& ios, zmq::context_t& zenv, const int zmq_socket_type) : ios_(ios), zenv_(zenv), zmq_socket_(zenv, zmq_socket_type), opts_(zmq_socket_) { opts_.linger(0); } bool shor::dm::zee::detail::Zmq_socket_base::send( Msg msg) { if (msg.frames().empty()) { auto frame = Msg::Frame(); const auto send_p = zmq_socket_.send(frame); static_cast<void>(opts_.events()); return send_p; } const auto first = msg.frames().begin(); const auto last = std::prev(msg.frames().end()); for (auto pos = first; pos != last; ++pos) { const auto send_p = zmq_socket_.send(*pos, ZMQ_SNDMORE); static_cast<void>(opts_.events()); if (!send_p) { return false; } } const auto send_p = zmq_socket_.send(msg.frames().back(), 0); static_cast<void>(opts_.events()); return send_p; } shor::dm::zee::Msg shor::dm::zee::detail::Zmq_socket_base::recv() { Msg my_value; int more_p = 0; do { zmq::message_t frame; const int flags = 0; zmq_socket_.recv(&frame, flags); static_cast<void>(opts_.events()); my_value.frames().push_back(std::move(frame)); more_p = opts_.rcvmore(); } while (more_p); return my_value; } void shor::dm::zee::detail::Zmq_socket_base::read_handler( std::function<void ()> io_fn, const boost::system::error_code& err, const Async_recv_handler recv_handler) { if (err == boost::asio::error::operation_aborted) { auto msg = Msg(); recv_handler(err, msg); } else if (opts_.events() & ZMQ_POLLIN) { do { auto msg = this->recv(); recv_handler(err, msg); } while (opts_.events() & ZMQ_POLLIN); } else { io_fn(); } } ==> zee/zmq_socket_base.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_ZMQ_SOCKET_BASE_HPP #define INCLUDE_SHOR_DM_ZEE_ZMQ_SOCKET_BASE_HPP // DM headers #include "dm/zee/handler.hpp" #include "dm/zee/msg.hpp" // 0MQ headers #include <zmq.hpp> // Boost C++ headers #include <boost/asio.hpp> // C++ headers #include <string> namespace shor { namespace dm { namespace zee { namespace detail { struct Zmq_socket_base { class Opts { zmq::socket_t& zmq_socket_; public: explicit Opts(zmq::socket_t& zmq_socket); SHOR_DM_DEFAULT_DTOR(~Opts()); void linger(int value); void subscribe(const std::string& value); int events() const; int linger() const; int rcvmore() const; private: template<typename Opt_value> Opt_value getsockopt(int opt_name) const; void setsockopt(int opt_name, const std::string& opt_value); template<typename Opt_value> void setsockopt(int opt_name, const Opt_value& opt_value); SHOR_DM_DELETE(Opts(const Opts& that)); SHOR_DM_DELETE(Opts& operator = (const Opts& that)); }; private: boost::asio::io_service& ios_; zmq::context_t& zenv_; zmq::socket_t zmq_socket_; Opts opts_; public: void bind(const char* addr); void bind(const std::string& addr); void close(); void connect(const char* addr); void connect(const std::string& addr); boost::asio::io_service& get_io_service(); Opts& opts(); bool send(Msg msg); Msg recv(); bool connected() const; const boost::asio::io_service& get_const_io_service() const; const boost::asio::io_service& get_io_service() const; const Opts& opts() const; protected: typedef std::function<void ()> Void_fn; typedef std::function< void (Void_fn, const boost::system::error_code& err, Async_recv_handler recv_handler) > Base_read_handler; Zmq_socket_base( boost::asio::io_service& ios, zmq::context_t& zenv, int zmq_socket_type); SHOR_DM_DEFAULT_DTOR(virtual ~Zmq_socket_base()); void read_handler( std::function<void ()> io_fn, const boost::system::error_code& err, Async_recv_handler handler); zmq::context_t& zenv(); zmq::socket_t& zmq_socket(); const zmq::context_t& const_zenv() const; const zmq::socket_t& const_zmq_socket() const; const zmq::context_t& zenv() const; const zmq::socket_t& zmq_socket() const; private: SHOR_DM_DELETE(Zmq_socket_base(const Zmq_socket_base& that)); SHOR_DM_DELETE( Zmq_socket_base& operator = (const Zmq_socket_base& that)); virtual void close_ext() = 0; }; } // namespace detail typedef detail::Zmq_socket_base Zmq_socket_base; } // namespace zee } // namespace dm } // namespace shor #include "dm/zee/zmq_socket_base.ipp" #endif ==> zee/zmq_socket_base.ipp <== #ifndef INCLUDE_SHOR_DM_ZEE_ZMQ_SOCKET_BASE_IPP #define INCLUDE_SHOR_DM_ZEE_ZMQ_SOCKET_BASE_IPP // My header #include "dm/zee/zmq_socket_base.hpp" inline void shor::dm::zee::detail::Zmq_socket_base::Opts::linger( const int value) { this->setsockopt(ZMQ_LINGER, value); } inline void shor::dm::zee::detail::Zmq_socket_base::Opts::subscribe( const std::string& value) { this->setsockopt(ZMQ_SUBSCRIBE, value); } inline int shor::dm::zee::detail::Zmq_socket_base::Opts::events() const { return this->getsockopt<int>(ZMQ_EVENTS); } inline int shor::dm::zee::detail::Zmq_socket_base::Opts::linger() const { return this->getsockopt<int>(ZMQ_LINGER); } inline int shor::dm::zee::detail::Zmq_socket_base::Opts::rcvmore() const { return this->getsockopt<int>(ZMQ_RCVMORE); } template<typename Opt_value> Opt_value shor::dm::zee::detail::Zmq_socket_base::Opts::getsockopt( const int opt_name) const { Opt_value opt_value; std::size_t opt_size = sizeof(Opt_value); zmq_socket_.getsockopt(opt_name, &opt_value, &opt_size); return opt_value; } inline void shor::dm::zee::detail::Zmq_socket_base::Opts::setsockopt( const int opt_name, const std::string& opt_value) { zmq_socket_.setsockopt(opt_name, opt_value.c_str(), opt_value.size()); } template<typename Opt_value> inline void shor::dm::zee::detail::Zmq_socket_base::Opts::setsockopt( const int opt_name, const Opt_value& opt_value) { const std::size_t opt_size = sizeof(Opt_value); zmq_socket_.setsockopt(opt_name, &opt_value, opt_size); } inline void shor::dm::zee::detail::Zmq_socket_base::bind( const char* addr) { zmq_socket_.bind(addr); } inline void shor::dm::zee::detail::Zmq_socket_base::bind( const std::string& addr) { zmq_socket_.bind(addr.c_str()); } inline void shor::dm::zee::detail::Zmq_socket_base::close() { this->close_ext(); zmq_socket_.close(); } inline void shor::dm::zee::detail::Zmq_socket_base::connect( const char* addr) { zmq_socket_.connect(addr); } inline void shor::dm::zee::detail::Zmq_socket_base::connect( const std::string& addr) { zmq_socket_.connect(addr.c_str()); } inline boost::asio::io_service& shor::dm::zee::detail::Zmq_socket_base::get_io_service() { return ios_; } inline shor::dm::zee::detail::Zmq_socket_base::Opts& shor::dm::zee::detail::Zmq_socket_base::opts() { return opts_; } inline bool shor::dm::zee::detail::Zmq_socket_base::connected() const { // This function should have been const in zmq::socket_t return const_cast<zmq::socket_t&>(zmq_socket_).connected(); } inline const boost::asio::io_service& shor::dm::zee::detail::Zmq_socket_base::get_const_io_service() const { return ios_; } inline const boost::asio::io_service& shor::dm::zee::detail::Zmq_socket_base::get_io_service() const { return ios_; } inline zmq::context_t& shor::dm::zee::detail::Zmq_socket_base::zenv() { return zenv_; } inline zmq::socket_t& shor::dm::zee::detail::Zmq_socket_base::zmq_socket() { return zmq_socket_; } inline const zmq::context_t& shor::dm::zee::detail::Zmq_socket_base::const_zenv() const { return zenv_; } inline const zmq::socket_t& shor::dm::zee::detail::Zmq_socket_base::const_zmq_socket() const { return zmq_socket_; } inline const zmq::context_t& shor::dm::zee::detail::Zmq_socket_base::zenv() const { return zenv_; } inline const zmq::socket_t& shor::dm::zee::detail::Zmq_socket_base::zmq_socket() const { return zmq_socket_; } #endif ==> zee/zmq_socket_fwd.hpp <== #ifndef INCLUDE_SHOR_DM_ZEE_ZMQ_SOCKET_FWD_HPP #define INCLUDE_SHOR_DM_ZEE_ZMQ_SOCKET_FWD_HPP namespace shor { namespace dm { namespace zee { namespace detail { #ifndef _MSC_VER class Posix_zmq_socket; #else class Windows_zmq_socket; #endif } // namespace detail #ifndef _MSC_VER typedef detail::Posix_zmq_socket Posix_zmq_socket; typedef Posix_zmq_socket Zmq_socket; #else typedef detail::Windows_zmq_socket Windows_zmq_socket; typedef Windows_zmq_socket Zmq_socket; #endif } // namespace zee } // namespace dm } // namespace shor #endif ________________________________ This e-mail and any attachments are confidential. If it is not intended for you, please notify the sender, and please erase and ignore the contents.

1 0

ZeroMQ integration with Boost-v1.54
by Rahul Mathur 02 Dec '13

02 Dec '13

All, Looking to have ZeroMQ integration with Boost-v1.54. I am familiar with Boost.ASIO client and Server model. BTW, How do I proceed to have ZeroMQ integrated with Boost-v1.54 with GCC on Linux platform?

3 2

Ask for material about the algorithm of boost::polygon in detail
by rogy liu 02 Dec '13

02 Dec '13

Hi all, I want to optimize the performance of the boolean operation of boost::polygon with CUDA. I want to speedup it at the source code level, and I found it is hard to speedup the hot spot because of dependency & container. Then, I want to evaluate the feasibility from the algorithm. I have found the article "Industrial strength polygon clipping - A novel algorithm with applications in VLSI CAD", but I think it is hard to understand. Do you have more material for study? Thanks, Rogy

1 0

[MSM] Best place to execute an action with return value
by Daniele Barzotti 02 Dec '13

02 Dec '13

Hi, I'm diving into this very useful piece of code which is MSM and I'm modelling a state machine for a gsm modem. When I have to perform a call I need to move from the IDLE state to WAITING_RESPONSE and, in the transaction action I have to call the MakeCall() method on my GSM object. This method has a return value and a non 0 value means it is failed. In this situation the SM don't have to move on to the next state but remains in the IDLE one, so... Since the Actions has no return value, I have to use a Guard instead or I have to add another state called MAKE_CALL? Thanks in advance, Daniele.

2 4