March 2009 - Boost-users - lists.preview.boost.org

Making friend with boost::interprocess
by Oncaphillis 11 Mar '09

11 Mar '09

Hi, I'd like to implement a cross-process singleton which resides in a shared memory segment. So the basic idea is to provide a factory method that is the only way to construct the Singleton. e.g. class Foo { public: static Foo & GetInstance() { try { boost::interprocess::managed_shared_memory segment( boost::interprocess::open_or_create, "FooSegment", 65536); void_allocator_t alloc(segment.get_segment_manager()); return *(segment.find_or_construct<Foo>( "FooObject")(alloc)); } catch(std::exception & ex) { throw; } } private: Foo(const void_allocator_t & a) { ... ... } Foo(const Foo &); } Of course somewhere deep in the guts of boost I get an error since the Foo constructor is private. Who do I have to befriend in order to make this scheme work ? Thank you O.

1 0

pure virtual function call error
by e120281 11 Mar '09

11 Mar '09

class A { public: int i; A( int i ) { this->i = i; } ~A() {} template<class Archive> void serialize(Archive &ar, const unsigned int version) { ar & i; } }; class B { public: B() {} ~B() { std::string filename = "C:\\a2.txt"; const A a( 2 ); std::ofstream ofs( filename.c_str(), std::ios::out | std::ios::binary ); boost::archive::binary_oarchive boa( ofs ); boa << a; } static B* getInstance() { static B instance; return &instance; } void serialize() { std::string filename = "C:\\a1.txt"; const A a( 1 ); std::ofstream ofs( filename.c_str(), std::ios::out | std::ios::binary ); boost::archive::binary_oarchive boa( ofs ); boa << a; } }; int main() { B* b = B::getInstance(); b->serialize(); } When I run the above code in visual studio 2005 I get pure virtual function call error. If I comment out the line b->serialize(); then everything is ok. What is wrong with the above code??? -- View this message in context: http://www.nabble.com/pure-virtual-function-call-error-tp22434024p22434024.… Sent from the Boost - Users mailing list archive at Nabble.com.

1 0

How to build specified library for install?
by Dancefire 11 Mar '09

11 Mar '09

Hi, I am going to build boost::iostreams and install to some location. I tried following command: C:\boost_1_38_0>bjam toolset=msvc variant=debug,release link=static threading=multi runtime-link=shared -sZLIB_SOURCE=E:\zlib123 -sBZIP2_SOURCE=E:\bzip2-1.0.5 --prefix="E:\Program Files\boost\boost_1_38_0\" --with-iostreams install However, I got following error: error: wrong library name 'iostreams install' in the --with-<library> option. It looks like that bjam consider install as a part of the library name. Is anything wrong with me? or it is a bug of bjam? Tao Wang

1 1

[Q] boost class to handle SQL Decimal (19,6)
by V S P 11 Mar '09

11 Mar '09

Hello, I have been searching and searching but could not find a Boost (or Boost-friendly) extension that handles SQL Decimal datatype (in my case decimal (19,6) I did find in Boost's vault http://www.boostpro.com/vault/index.php mp_math_v04 by Baraclese but not clear if it supports decimal point (I did not see any documentation) thanks in advance for any help, Vlad -- V S P toreason(a)fastmail.fm -- http://www.fastmail.fm - Or how I learned to stop worrying and love email again

1 0

MPL design question
by Jesse Perla 11 Mar '09

11 Mar '09

I have a class that is getting a huge number of template parameters (mostly a bunch of size_t used for vector sizes). They have a natural organization into groups, but almost all of them have a default as 0. How do you suggest that I design the template parameter for this? Is there any way to use some kind of associative container so I could look these up by "name" and default to 0 if they aren't there? I can't see how to use mpl::map to do this kind of stuff. Simplified Pseudocode: -------------------- What I have: template<std::size_t IndividualSize = 0, std::size_t AggregateSize> //Many others class my_problem{...}; What I want: template<class StateSizes> class MyProblem{...} Called something like: typedef my_problem<map<AggregateSize, 2> > my_problem_type; where AggregateSize is an enum, placeholder struct, etc. And in this case I would default IndividualSize to 0. -Jesse

2 1

Re: [Boost-users] [wave] assertion failed - multithread issue?
by Ein Held 10 Mar '09

10 Mar '09

Hello Hartmut Using multiple threads, wave crashes randomly after a few seconds of running. The crash is actually an assert in test.exe!std::list<boost::wave::cpplexer::lex_token<boost::wave::util::file_position<boost::wave::util::flex_string<char,std::char_traits<char>,std::allocator<char>,boost::wave::util::CowString<boost::wave::util::AllocatorStringStorage<char,std::allocator<char> >,char *> > > >,boost::fast_pool_allocator<boost::wave::cpplexer::lex_token<boost::wave::util::file_position<boost::wave::util::flex_string<char,std::char_traits<char>,std::allocator<char>,boost::wave::util::CowString<boost::wave::util::AllocatorStringStorage<char,std::allocator<char> >,char *> > > >,boost::default_user_allocator_new_delete,boost::details::pool::win32_mutex,32> >::_Const_iterator<1>::operator*() Line 218 + 0x17 bytes C++ indicating that the iterator is not dereferencable. I am using an VC 9 x86 debug build on Vista x64. Here's the code: #include <fstream> #include <string> //Wave does not like the new interface :( //#define BOOST_FILESYSTEM_NO_DEPRECATED #include <boost/thread.hpp> #include <boost/wave.hpp> #include <boost/wave/cpplexer/cpp_lex_token.hpp> #include <boost/wave/cpplexer/cpp_lex_iterator.hpp> std::string file; const char* file_name="C:\\Program Files (x86)\\Microsoft SDKs\\Windows\\v5.0\\Include\\Windows.h"; void test() { for(;;) { typedef boost::wave::cpplexer::lex_token<> token_type; typedef boost::wave::cpplexer::lex_iterator<token_type> lex_iterator_type; typedef boost::wave::context<std::string::iterator, lex_iterator_type> context_type; context_type ctx (file.begin(), file.end(), file_name); ctx.set_language(static_cast<boost::wave::language_support>(boost::wave::support_cpp|boost::wave::support_option_variadics|boost::wave::support_option_include_guard_detection)); ctx.add_sysinclude_path("C:\\Program Files (x86)\\Microsoft SDKs\\Windows\\v5.0\\Include\\"); ctx.add_sysinclude_path("C:\\Program Files (x86)\\Microsoft Visual Studio 9.0\\VC\\include\\"); ctx.add_macro_definition("_CPPRTTI",true); ctx.add_macro_definition("_CPPUNWIND",true); ctx.add_macro_definition("_MT",true); ctx.add_macro_definition("_MSC_VER=1600",true); ctx.add_macro_definition("_M_IX86=600",true); ctx.add_macro_definition("_WIN32=600",true); context_type::iterator_type first = ctx.begin(); context_type::iterator_type last = ctx.end(); while (first != last) { ++first; } } } int main() { std::ifstream instream(file_name); file = std::string(std::istreambuf_iterator<char>(instream.rdbuf()),std::istreambuf_iterator<char>()); instream.close(); for(int i=0;i<10;++i) { boost::thread t(test); } test(); } Thank you

4 5

boost/math/special_functions/legendre.hpp causing compile errors in other headers
by Jeremy Conlin 10 Mar '09

10 Mar '09

I recently started using the Boost Math libraries because I need legendre polynomials. However I cam across a problem I was hoping someone on this list could help me with. My personal code compiles fine with no problems. When I add the line #include "boost/math/special_functions/legendre.hpp" then I get compiler errors attributed to one of my header files. Specifically I get: my_header.hh:109 error: ISO C++ forbids declaration of ‘tuple’ with no type As I mentioned before, I don't believe the problem is in my header file because it compiles just fine without the legendre.hpp file. Has anyone seen this problem before? Do you know how to fix it? I am running gcc 4.2.1 on Mac OS X 10.5.6. Thanks, Jeremy

2 2

[mpl] Trivial metafunctions and placeholders don't mix ?
by jean-louis.a.leroy＠fortis.com 10 Mar '09

10 Mar '09

Hello. I'm trying to use a pair<> to pass values inside a fold<> : template<typename Seq> struct remove_duplicates // even if not consecutive; also keep order { typedef typename mpl::second< mpl::fold< Seq, mpl::pair< mpl::set<>, mpl::vector<> >, typedef typename mpl::if_< mpl::has_key<mpl::first<mpl::_1>, mpl::_2>, mpl::_1, mpl::pair< typename mpl::insert< mpl::first<mpl::_1>, void, mpl::_2 >::type, // insert mpl::push_back< mpl::second<mpl::_1>, _2 >::type > // pair > >::type > type; }; It looks like first<> and second<> don't work on placeholders. My compiler - Visual Studio 2008 - complains that "'first' : is not a member of 'boost::mpl::arg<1>". When I look at the definition of first<> I see a "BOOST_MPL_AUX_LAMBDA_SUPPORT(1,first,(P))" that should take care of placeholders I guess. But when I look at the definition of the macro, alas : #if !defined(BOOST_MPL_CFG_NO_FULL_LAMBDA_SUPPORT) # define BOOST_MPL_AUX_LAMBDA_SUPPORT_SPEC(i, name, params) /**/ # define BOOST_MPL_AUX_LAMBDA_SUPPORT(i,name,params) /**/ What's the story here? My compiler is not good enough? Could be... While we're at it, should I put a BOOST_MPL_AUX_LAMBDA_SUPPORT(1,remove_duplicates,(P)) at the bottom of remove_duplicates<> ? Would it make my metafunction work with lambda placeholders? Thanks, Jean-Louis Leroy = = = = = = = = = = = = = = = = = = = = = = = = = Fortis disclaimer : http://www.fortis.be/legal/disclaimer.htm Privacy policy related to banking activities of Fortis: http://www.fortis.be/legal/privacy_policy.htm = = = = = = = = = = = = = = = = = = = = = = = = =

2 1

Trouble using boost::edge_range in Boost Graph library
by Artyom 10 Mar '09

10 Mar '09

Hi all, I am having trouble getting this function to work. I am using 'multisetS' for OutEdgeProperty as instructed in the documentation, however, my code refuses to compile with the following errors: http://rafb.net/p/LgT89332.html I am using Visual Studio 8.0 SP1 under Windows Vista x64. Regards, Artyom

2 2

[program_options] Getting access to the remaining argv
by Ruediger Berlich 10 Mar '09

10 Mar '09

Hi there, I have a program ("A") that calls other, external programs ("B"). A can get a number of command line options, which are parsed with the help of the program_options library. B may also accept options (which are possibly defined by other people and not known to me in advance). Thus I want to pass all those options that are not known to A on to B, in the hope that they are needed for that program. Is there a way to retrieve the remaining argv string, with all options "known" to A removed ? Or should I better invent a flag of the following kind for A --external-options="-a 1 -b 2 -c 3" whose value is passed on to the external program ? Thanks and Best Regards, Ruediger

3 2