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

[Serialization] creating an archive fails under GCC-4.6.0
by doug livesey 18 Apr '11

18 Apr '11

Hi -- I have some code that runs when compiled with GCC-4.2 (on OSX Snow Leopard), but that throws the following error with my fresh compile of 4.6.0: run_features(2872) malloc: *** error for object 0x100502060: pointer being freed was not allocated *** set a breakpoint in malloc_error_break to debug The line that creates this error is here, on line 24: https://gist.github.com/923262 (Apologies for any egregious code abuses in there, I'm just learning!) In the debugger, it lets me step from that line to the constructor of xml_oarchive (line 114 of xml_oarchive.hpp), and when I try to step any further, I get this: run_features(2885) malloc: *** error for object 0x100501b00: pointer being freed was not allocated *** set a breakpoint in malloc_error_break to debug Program received signal SIGABRT, Aborted. 0x00007fff871a95d6 in __kill () Could anyone advise me on what I could do to fix this? I really want to be working with GCC 4.6.0 and the boost serialisation libraries. Incidentally, I reinstalled boost after I installed the latest GCC (with homebrew), and I'd set cc, gcc, and g++ symlinks to GCC-4.6.0 before I did that, so boost should be compiled with the latest. Thanks very much for any & all assistance. Doug.

3 3

Boost asio lib question: async_accept and EBADF
by olegb＠it-forces.com 18 Apr '11

18 Apr '11

Cheers, I have a question regarding handling error bad file descriptor in asio async_accept. Standard example doesn't expose whether it's fatal or not. If I understand correctly bad_descriptor error is not fatal here, like it would be for usual POSIX accept() call, and we should continue accepting new incoming calls, no? void handle_accept(tcp_connection::pointer new_connection, const boost::system::error_code& error) { if (!error) { new_connection->start(); start_accept(); } } e.g. if (!error) { new_connection->start(); start_accept(); } else if (error == bad_descriptor) { start_accept(); } -- King regards, Oleh

1 0

Pimpl framework comments/suggestions
by Clinton Mead 18 Apr '11

18 Apr '11

Hi All I've been experimenting with pimpl recently. What I've seen written about the pimpl idiom is that you have an interface class and an implementation class, and in the interface class you put some kind of pointer to the implementation. However, I thought, sometimes you might want that to be a shared pointer. Or an intrusive pointer. Or unique pointer. Or maybe sometimes you might decide not to use pimpl at all. I thought it was a bit silly that you had to make these decisions at the time you create the interface class. I thought it would be nice if you could write the interface independent of how it was going to be connected to the implementation, and then hook them together however you like. So this is what I've done. Basically I've broke things up into three separate parts: (1) Interface (2) Implementation (3) What joins them together (i've created unique_pimpl and shared_pimpl as examples) I've implemented a unique_ptr and shared_ptr pimpl implementations as examples. The idea is that you can basically do this: unique_pimpl<interface, impl>::create(...) or shared_pimpl<interface, impl>::create(...) make_your_own<interface, impl>::create(...) for any interface and implementation you choose. I've included all the code in a question at http://stackoverflow.com/questions/5666442 but it doesn't seem to be getting much attention there, so I thought I'd bring it up on this list. If something like this isn't in Boost, and people like it, with a bit of help I'd be happy to work on it to get it into Boost. Alternatively I'm happy for someone else to Boostify this. I've used a bit of c++0x in the code, gcc 4.4.5 will compile it though. However if I'm just (badly) reinventing the wheel, if someone could direct me to where this has already been done that would be good to. Thanks Clinton

2 5

filesystem v3, unicode and std::codecvt linker error
by archie14 18 Apr '11

18 Apr '11

I am porting the code from (boost44, msvc2008, Qt, unicode support) to (boost46, msvc2010, Qt, unicode support). I am getting following linker error: error LNK2001: unresolved external symbol "private: static class std::codecvt<unsigned short,char,int> const * & __cdecl boost::filesystem3::path::wchar_t_codecvt_facet(void)" (? wchar_t_codecvt_facet@path@filesystem3@boost@@CAAAPBV?$codecvt@GDH@std@@XZ) I found that std::codecvt defined in <xlocale>. May be the filesystem lib should be built in certain way to support widestrings? Any advise/help will be greatly appreciated.

1 1

[system] static linking + static member in y a dynamic design
by Amoriello Hutti 17 Apr '11

17 Apr '11

Hi everyone, I am facing a "design" issue while sharing error_codes across modules boundaries. Has I understand the design, error_categories like "generic_error_category" are static (maybe to provide unique way to represent error_category, avoiding error_category ID confilct, and allowing users to extend with there own categories). The question is : How can we share error_code across modules (A and B), when these modules are both statically linked with there own boost::system library (same version)? Under Windows operating system, both static version of boost::system library co-exist in the same process address space (that contains module A and module B), they do not share code or static data members. Comparison between error_code is made by looking error_value and error_category, but error_categroy doesn't has the same "static" address, causing comparison to fail, even when the semantic is respected (because of the co-existing static error_category with different address). sharing the boost::system has a unique shared library used by both module A and B could solve this issue, but this solution is not suitable for certain needs. Thanks for your help! -- Amoriello Hutti

1 0

[phoenix][accumulators] Is is possible to use phoenix v3 and accumulators in parallel?
by alfC 17 Apr '11

17 Apr '11

Hi, Last week I started a thread called "[Boost-users] [phoenix][spirit] Is is possible to use phoenix v3 and spirit in parallel? " . While that issue seems under control, now I found a compatibility problem between accumulators and phoenix 3. #include <boost/phoenix/phoenix.hpp> //#include <boost/spirit/home/ phoenix.hpp> #include<boost/accumulators/accumulators.hpp> #include<boost/accumulators/statistics/min.hpp> int main(){ return 0; } gives an error boost/accumulators/numeric/detail/function_n.hpp:83:1: error: ‘apply12’ in namespace ‘boost::mpl’ does not name a type I tried predefining certain macros but it didn't work: #define BOOST_SPIRIT_USE_PHOENIX_V3 #define BOOST_PROTO_MAX_ARITY 11 #define BOOST_PROTO_MAX_LOGICAL_ARITY 11 #define BOOST_MPL_LIMIT_METAFUNCTION_ARITY 12 Is there a workaround? Thank you, Alfredo

1 0

iterator_facade and distance()
by Andrej van der Zee 17 Apr '11

17 Apr '11

Hi, I am trying to implement an iterator using boost::iterator_facade. It works fine but it is very slow, I think because the distance() function takes lineair time as it is seems to be implemented in terms of increment(). I would like to implement the distance() function myself but I am not sure how. I tried to implement distance_to() like this: class const_data_stream_iterator : public boost::iterator_facade<const_data_stream_iterator, char const, boost::forward_traversal_tag> { public: typedef boost::iterator_facade<const_data_stream_iterator, char const, boost::forward_traversal_tag> facade_t; private: friend class boost::iterator_core_access; facade_t::difference_type difference_to(const_data_stream_iterator & to) const { ... } }; But this does not work. How can I do this? Thank you, Andrej

2 6

[BGL] Modifying graph in BFS
by Michal Sekletar 16 Apr '11

16 Apr '11

Hello everyove, I have a question about BGL. I want to traverse through graph using BFS and before actual algorithm I want to remove from graph structure (adjacency_list) some vertices and all edges connected to these vertices, so I wrote my own visitor class and implemented initliaze_vertex method to do check on every vertex (it's property) if it's satisfied then I iterate through all edges in graph and i store iterators to edges - edge must be connected to current vertex. Then I want to remove these edges and vertex itself. I store edge iterators for edges connected to vertex I want to remove, because I understand that calling remove_edge directly, while iterating through edges will lead to undefined behavior (if edges are stored in vector - removing invalidates iterators). My problem is that I can't compile source file, compiler is complaining about template instantiation(error message is very long and I don't understand what is cause of the problem) Any ideas, suggestions, code snippets, how to solve this issue would be much appreciated. Thank you Michal Sekletar

2 1

[phoenix][spirit] Is is possible to use phoenix v3 and spirit in parallel?
by alfC 16 Apr '11

16 Apr '11

Hi, I have a program that uses Phoenix and Spirit. Spirit uses indirectly Phoenix2. With some effort I modified the part that use Phoenix to phoenix3 but now when I want to put all together it complains. The obivous warnings are: usr/include/boost/phoenix/version.hpp:16:1: warning: "BOOST_PHOENIX_VERSION" redefined usr/include/boost/phoenix/core/limits.hpp:23:1: warning: "PHOENIX_LIMIT" redefined and the most serious is: usr/include/boost/phoenix/core/limits.hpp:112:5: error: #error "BOOST_PROTO_MAX_LOGICAL_ARITY < BOOST_PHOENIX_COMPOSITE_LIMIT" The question is how can I use Phoenix3 and Spirit in the same code (in parallel). thank you, Alfredo

3 4

[phoenix] function object that is also lazy
by alfC 16 Apr '11

16 Apr '11

Hi, Is this possible at all? struct A : boost::phoenix::function<A const&>{ double operator()(double const& d) const{return 5.;} } such that an object function it is not only a function but also a lazy function of itself? std::clog << a(4.) << std::endl; std::clog << a(arg1 + 1.)(4.) << std::endl; the code above doesn't work because A is not complete when the type function<A> is instatiated. However after some experimentation I got the following. The question is, does it really work in more exotic context? Is there a possible simplification? Am I reinventing the wheel? template<class Self> struct lazy{ boost::phoenix::function<Self const&> const* const lazy_; lazy(Self const& self) : lazy_(new boost::phoenix::function<Self const&>(self)){} template<class Arg> const 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::tag::terminal, boost::proto::argsns_::term<Self>, 0l>, Arg >, 2l> > operator()(Arg const& arg) const{return (*lazy_)(arg);} ~lazy(){delete(lazy_);} }; struct A : lazy<A>{ using lazy<A>::operator(); A() : lazy<A>(*this){} A(A const& other) : lazy<A>(*this){} typedef double result_type; result_type operator()(double const& d) const{return d;} }; so, any function object B can be a lazy function of it self provided that we use the correct CRTP argument (lazy<B>) and manually add the code: using lazy<A>::operator(); B() : lazy<A>(*this){} B(A const& other) : lazy<A>(*this){} inside the class. This feature seems to make the usage of the function object b more flexible, since it can we used in many different contexts. For example, as integrate ( b, 1., 2.); or as integrate ( 2.*b(arg1 + 1.) , 1., 2.) Thanks, Alfredo

1 1