May 2007 - Boost-users - lists.preview.boost.org

Newbie: converting a fs::path to std::string?
by Alan Robert Clark 17 May '07

17 May '07

Morning all I am trying to do: <snip> #include "boost/filesystem/fstream.hpp" namespace fs=boost::filesystem; int produce_lists(fs::path homedir) { fs::path del_fn = homedir / ".wkhm/deletions"; fs::path cr_fn = homedir / ".wkhm/creations"; fs::ofstream deletionsfile (del_fn); fs::ofstream creationsfile (cr_fn); // ensure that the deletions are always transferred creationsfile << homedir << ".wkhm/deletions" << std::endl; } and the << operator does not try to convert fs::path to std:: string. Best way of handling this? Any suggestions welcomed Alan -- Alan Robert Clark, Pr Eng Computational Electromagnetics Associate Professor Wits University School of Electrical and Information Engineering Wits; Gauteng ``Bugs are later known as features'' 2050 South Africa Ps 111:10; Ps 37/150 Fax (+27 11)403-1929 a.clark(a)ee.wits.ac.za Tel (+27 11)717-7223(24hr) <http://YingTongDiddleiPo.ee.wits.ac.za/> GPG Public Key at <http://YingTongDiddleiPo.ee.wits.ac.za/arc.gpg> **Linux ---the choice of a GNU generation.**

2 1

Best way to create a class using the singleton pattern and shared pointers?
by Stephen Torri 17 May '07

17 May '07

Here is a function that I am trying to understand why its giving me a error via valgrind of "Invalid read of size 4". 16: Trace_State::ptr_t Trace_State::Instance() 17: { 18: // Lock the resource? 19: if ( m_instance.get() == 0 ) 20: { 21: m_instance = Trace_State::ptr_t ( new Trace_State() ); 22: } 23: 24: // Unlock the resource? 25: 26: return m_instance; 27: } When I look at the full error report I see: ==10721== Invalid read of size 4 ==10721== at 0x41682E3: boost::detail::atomic_increment(int*) (sp_counted_base_gcc_x86.hpp:66) ==10721== by 0x4168359: boost::detail::sp_counted_base::add_ref_copy() (sp_counted_base_gcc_x86.hpp:133) ==10721== by 0x41684B1: boost::detail::shared_count::shared_count(boost::detail::shared_count const&) (shared_count.hpp:170) ==10721== by 0x41775CF: boost::shared_ptr<libreverse::trace::Trace_State>::shared_ptr(boost::shared_ptr<libreverse::trace::Trace_State> const&) (shared_ptr.hpp:106) ==10721== by 0x41772F0: libreverse::trace::Trace_State::Instance() (Trace.cpp:26) ==10721== by 0x417731B: libreverse::trace::Trace::write_Trace(unsigned, std::string) (Trace.cpp:180) ==10721== by 0x425E394: libreverse::elf_module::Elf_Reader<32>::~Elf_Reader() (Elf_Reader_T.cpp:52) ==10721== by 0x423C8FD: void boost::checked_delete<libreverse::elf_module::Elf_Reader<32> >(libreverse::elf_module::Elf_Reader<32>*) (checked_delete.hpp:34) ==10721== by 0x42480B8: boost::detail::sp_counted_impl_p<libreverse::elf_module::Elf_Reader<32> >::dispose() (sp_counted_impl.hpp:76) ==10721== by 0x804C207: boost::detail::sp_counted_base::release() (sp_counted_base_gcc_x86.hpp:145) ==10721== by 0x804C231: boost::detail::shared_count::~shared_count() (shared_count.hpp:159) ==10721== by 0x423C171: boost::shared_ptr<libreverse::io::File_Reader>::~shared_ptr() (shared_ptr.hpp:106) ==10721== Address 0x43F7A44 is 12 bytes inside a block of size 24 free'd ==10721== at 0x40050FF: free (vg_replace_malloc.c:233) ==10721== by 0x4185AC5: operator delete(void*) (sp_debug_hooks.cpp:205) ==10721== by 0x417C9A4: boost::detail::sp_counted_impl_p<libreverse::trace::Trace_State>::~sp_counted_impl_p() (sp_counted_impl.hpp:52) ==10721== by 0x4168337: boost::detail::sp_counted_base::destroy() (sp_counted_base_gcc_x86.hpp:126) ==10721== by 0x804C1CD: boost::detail::sp_counted_base::weak_release() (sp_counted_base_gcc_x86.hpp:159) ==10721== by 0x804C212: boost::detail::sp_counted_base::release() (sp_counted_base_gcc_x86.hpp:146) ==10721== by 0x804C231: boost::detail::shared_count::~shared_count() (shared_count.hpp:159) ==10721== by 0x4168EBB: boost::shared_ptr<libreverse::trace::Trace_State>::~shared_ptr() (shared_ptr.hpp:106) ==10721== by 0x4176711: __tcf_1 (Trace.cpp:14) ==10721== by 0x47BDCBE8: __cxa_finalize (in /lib/libc-2.5.so) ==10721== by 0x4167192: (within /usr/local/lib/libreverse.so.0.0.0) ==10721== by 0x429EE3B: (within /usr/local/lib/libreverse.so.0.0.0) ==10721== The error is coming when the return of the Trace boost shared ptr. I am wondering if I have made an error in the creation of the class. Here is how the class is defined: #ifndef LIBREVERSE_TRACE_H #define LIBREVERSE_TRACE_H #include <boost/cstdint.hpp> #include <boost/shared_ptr.hpp> #include <fstream> #include "Reverse.h" namespace libreverse { namespace trace { /* Idea taken from http://www.codeproject.com/debug/xytrace.asp * (28 Jan 2002 - Xiangyang Liu) * * I have modified this so that we don't use varargs and use constant types */ class Trace_State { public: typedef boost::shared_ptr<Trace_State> ptr_t; virtual ~Trace_State(); static Trace_State::ptr_t Instance(); void set_Trace_File_Prefix ( std::string name ); void set_Trace_Level ( boost::uint32_t level ); void open_Trace_File ( void ); std::string get_ID_String ( void ); void close_Trace_File ( void ); boost::uint32_t get_Trace_Level ( void ) const; void write_Message ( boost::uint32_t level, std::string msg ); private: Trace_State(); static Trace_State::ptr_t m_instance; std::string m_file_prefix; boost::uint32_t m_trace_level; //boost::recursive_mutex m_mutex; std::ofstream m_log_stream; }; class Trace { public: static bool write_Trace ( boost::uint32_t level, std::string message ); }; } /* namespace trace */ } /* namespace libreverse */ #endif /* LIBREVERSE_TRACE_H */ Is this an appropriate way to create a class using the singleton pattern? Is the constructor assigning the boost shared_ptr? Stephen

3 7

Re: [Boost-users] fatal error C1001 with visual studio 8, boost 1.34, BOOST_AUTO() macro
by Trigve Siver 17 May '07

17 May '07

>----- Original Message ---- >From: Arkadiy Vertleyb <vertleyb(a)hotmail.com> >To: boost-users(a)lists.boost.org >Sent: Thursday, May 17, 2007 11:56:38 AM >Subject: Re: [Boost-users] fatal error C1001 with visual studio 8, boost 1.34, BOOST_AUTO() macro > >"Trigve Siver" <trigves(a)yahoo.com> wrote > >> The signature of exec() method is: >> >> --- >> result exec(const PGSTD::string &Query, >> const PGSTD::string &Desc=PGSTD::string()); >> --- >> >> >Is "pqxx::result" the exact type the expression T.exec("some select...") >> >returns? Can you do: >> > >> >pqxx::result r; >> >BOOST_AUTO(rr, r); >> > >> >? >> >> Yes I can do this without error: >> >> --- >> pqxx::result rs = T.exec(sql_stream.str()); >> BOOST_AUTO(r, rs); >> --- > >Can you paste the exact line of code that causes the error? --- BOOST_AUTO(rs, T.exec("SELECT 1")); --- > >Regards, >Arkadiy thanks Trigve ____________________________________________________________________________________ No need to miss a message. Get email on-the-go with Yahoo! Mail for Mobile. Get started. http://mobile.yahoo.com/mail

3 3

[1.33.1] Problem with fast_pool_allocator across DLL boundaries?
by Andrew Bancroft 17 May '07

17 May '07

Platform: Win32, VC 7 & VC 8.1. I have run into a problem using fast_pool_allocator across DLLs: it 'leaks' memory. In this example, DLL A: 1. declares a container type that uses fast_pool_allocator 2. provides a function to populate the container type DLL B creates instance(s) of the container type & calls DLL A to populate those instances. DLL A ----- struct S { /* A bunch of data */ }; typedef std::list<S, boost::fast_pool_allocator<S> > container_of_s; /* Fills s with a number of S instances */ void dll_populate_container_of_s(container_of_s &s); DLL B ----- void fun_a() { for (int nLoop=0; nLoop<100; ++nLoop) { container_of_s sc; dll_populate_container_of_s(sc); /* manipulate sc in some way */ /* sc is destroyed, but memory is not returned to the pool it was allocated from */ } } The problem is that the memory allocated by fast_pool_allocator inside DLL A (by dll_populate_container_of_s) is not returned to the same pool it was allocated from. I've tracked this down to fast_pool_allocator being backed by a singleton. It turns out that each DLL get's it's own copy of that singleton. So the memory allocated by fast_pool_allocator<S> inside DLL A is returned to a different fast_pool_allocator<S> inside DLL B. Technically not a leak, but the end result is that repeated calls to dll_populate_container_of_s cause more & more memory to be consumed. Does anyone have a solution or workaround for this? Or an alternate pool allocator? Thanks Andrew ---------------------------------------------------------------------- This e-mail, including any attached files, may contain confidential and privileged information for the sole use of the intended recipient. Any review, use, distribution, or disclosure by others is strictly prohibited. If you are not the intended recipient (or authorized to receive information for the intended recipient), please contact the sender by reply e-mail and delete all copies of this message.

1 0

[1.33.1][program_options] long option descriptions cause a crash in windows
by Scott Finley 17 May '07

17 May '07

I am using program_options on an app that I am compiling for linux and windows (vs2005). I have found that long option descriptions cause a fatal error on windows when they are long enough that they need to wrap to the next line. (I'm having no problem in linux). If I run the debug version I get an assertion failed in xstring. Release just gives a fatal error. The following code causes the problem: #include <iostream> #include <boost/program_options.hpp> using namespace std; namespace po = boost::program_options; using namespace boost; int main( int argc, char *argv[] ) { string tmp; // Declare a group of options that will be // allowed only on command line po::options_description cmdline_options("Usage: programOptionsTest [options] - Allowed Options blah blah blah blah blah blah blah blah"); cmdline_options.add_options() ("help,h", "produce this help message a b c d e f g h i j k l m n o p"); cout << cmdline_options << endl; cin >> tmp; return 0; } Note that the "blah blah blah..." part is wrapped ok, but the "help" line need to wrap after the "o" and crashes. If I remove the "p" from the string then it doesn't need to wrap and everything is ok. (You might need to adjust the string to cause wrapping if you console width is different then mine.) /************************************** * Scott Finley * Research Software Engineer * TomoTherapy Inc. * Email: sfinley(a)tomotherapy.com * Phone: 608.824.3381 ***/

2 2

How can I execute a system command in boost?
by Meryl Silverburgh 17 May '07

17 May '07

Hi, I am writing a C++ program under linux. Can you please tell me if there is a boost library which I execute a system command? Kind of like System.exec() in Java? Thank you.

2 1

Re: [Boost-users] fatal error C1001 with visual studio 8, boost 1.34, BOOST_AUTO() macro
by Trigve Siver 17 May '07

17 May '07

>----- Original Message ---- >From: Arkadiy Vertleyb <vertleyb(a)hotmail.com> >To: boost-users(a)lists.boost.org >Sent: Wednesday, May 16, 2007 10:50:54 PM >Subject: Re: [Boost-users] fatal error C1001 with visual studio 8, boost 1.34, BOOST_AUTO() macro > >"Trigve Siver" <trigves(a)yahoo.com> wrote > >> I'm using boost 1.34 with postgresql c++ binding (libpqxx). I've got this >> fatal error: >> >> fatal error C1001: An internal error has occurred in the compiler. >> (compiler file >> 'F:\SP\vctools\compiler\utc\src\P2\main.c[0x10C68EF3:0x00000008]', line >> 182) >> To work around this problem, try simplifying or changing the program near >> the locations listed above. >> Please choose the Technical Support command on the Visual C++ >> Help menu, or open the Technical Support help file for more information >> >> when compiling sources with BOOST_AUTO macro from boost/typeof library. >> The code that triggers this error is: >> >> --- >> pqxx::nontransaction T(*m_pSQLConnection, "GetCustomer"); >> BOOST_AUTO(rs, T.exec("some select...")); // pqxx::result rs = >> T.exec(""some select..."); >> --- >> >> pqxx is namespace for postgresql c++ binding. I have registered >> pqxx::result with >> BOOST_TYPEOF_REGISTER_TYPE (pqxx::result) > Hi, thanks for reply, The signature of exec() method is: --- result exec(const PGSTD::string &Query, const PGSTD::string &Desc=PGSTD::string()); --- >Is "pqxx::result" the exact type the expression T.exec("some select...") >returns? Can you do: > >pqxx::result r; >BOOST_AUTO(rr, r); > >? Yes I can do this without error: --- pqxx::result rs = T.exec(sql_stream.str()); BOOST_AUTO(r, rs); --- >Regards, >Arkadiy thanks Trigve ____________________________________________________________________________________ Moody friends. Drama queens. Your life? Nope! - their life, your story. Play Sims Stories at Yahoo! Games. http://sims.yahoo.com/

2 1

How can I wrap class functions that return a reference?
by Patrick Ye 17 May '07

17 May '07

Hi guys, I'm very new to boost.python, so my question might have been already answered. I'm having a lot of trouble creating interfaces for class member functions that return a reference to an object. For example if we have the following class: class btQuadWord { protected: btScalar m_x; btScalar m_y; btScalar m_z; btScalar m_unusedW; public: SIMD_FORCE_INLINE const btScalar& getX() const { return m_x; } }; and we have the following interface code: class_<btQuadWord>("btQuadWord") .def("getX", &btQuadWord::getX) ; I would get the following error messages (ignore the line numbers): /usr/include/boost/python/detail/invoke.hpp: In function 'PyObject* boost::python::detail::invoke(boost::python::detail::invoke_tag_<false, true>, const RC&, F&, TC&) [with RC = boost::python::detail::specify_a_return_value_policy_to_wrap_functions_returning<const btScalar&>, F = const btScalar& (btQuadWord::*)()const, TC = boost::python::arg_from_python<btQuadWord&>]': /usr/include/boost/python/detail/caller.hpp:199: instantiated from 'PyObject* boost::python::detail::caller_arity<1u>::impl<F, Policies, Sig>::operator()(PyObject*, PyObject*) [with F = const btScalar& (btQuadWord::*)()const, Policies = boost::python::default_call_policies, Sig = boost::mpl::vector2<const btScalar&, btQuadWord&>]' /usr/include/boost/python/object/py_function.hpp:38: instantiated from 'PyObject* boost::python::objects::caller_py_function_impl<Caller>::operator()(PyObject*, PyObject*) [with Caller = boost::python::detail::caller<const btScalar& (btQuadWord::*)()const, boost::python::default_call_policies, boost::mpl::vector2<const btScalar&, btQuadWord&> >]' interface.cpp:112: instantiated from here /usr/include/boost/python/detail/invoke.hpp:88: error: no match for call to '(const boost::python::detail::specify_a_return_value_policy_to_wrap_functions_returning<const btScalar&>) (const float&)' Now, btScalar here is just a #define for double, I've had similar problems when the return reference is to a class object. Is there anyway to get around this problem? Thanks a lot. Patrick P.S. How do you get the full list of past mails from the mailing list?

1 0

Re: [Boost-users] DeSerialization to update content of objectinsteadof construct new object
by Ng Pan 17 May '07

17 May '07

Thank you very much. I have successfully overloaded the serialization function. Here is the code as the reference for other ppl doing the similar modification. base.h #ifndef BASE_H #define BASE_H ... #include<boost\serialization\vector.hpp> ... class base { friend class boost::serialization::access; template<class Archive> void serialize(Archive &ar, const unsigned int /* file_version */){ ar & BOOST_SERIALIZATION_NVP(m_nID); } ..... } ... #if defined(__SGI_STL_PORT) || defined(_STLPORT_VERSION) #define STD _STLP_STD #else #define STD std #endif namespace boost { namespace serialization { //vector<Base *> template<class Archive, class Allocator> inline void save( Archive & ar, const STD::vector<base *, Allocator> &t, const unsigned int /* file_version */ ){ // record number of elements unsigned int count = t.size(); ar << BOOST_SERIALIZATION_NVP(count); for(unsigned int i=0; i< count; i++) ar<<boost::serialization::make_nvp("loaditem", t.at(i)); } template<class Archive, class Allocator> inline void load( Archive & ar, STD::vector<base *, Allocator> &t, const unsigned int /*file_version*/ ){ // // retrieve number of elements unsigned int count; ar >> BOOST_SERIALIZATION_NVP(count); for(unsigned int i=0; i<count; i++) ar >> boost::serialization::make_nvp("loaditem", *t.at(i)); } } } ... #endif Barr Ng >From: Matthias Troyer <troyer(a)phys.ethz.ch> >Reply-To: boost-users(a)lists.boost.org >To: boost-users(a)lists.boost.org >Subject: Re: [Boost-users] DeSerialization to update content of >objectinsteadof construct new object >Date: Wed, 16 May 2007 07:22:28 +1000 > >Another, maybe easier option, could be to overload the serialization >functions for std::vector<base *> . > >Matthias Troyer > >On 16 May 2007, at 01:02, Robert Ramey wrote: > > > The code for serialization of a vector is found in serialization/ > > vector.hpp. > > > > I implemented this code to reconstuct the vector items rather than > > just reload them. I did this for a number of reasons - which for > > this discussion are not important. > > > > If this doesn't suit your needs, I would suggest you make your own > > version of serialization.hpp which operates differently. Starting > > with the current version it wouldn't be at all hard to do. > > > > If you want to use both current ("standard") method in some > > places in addition to your "custom" method, you might consider > > making some sort of wrapper which tags your special > > vectors with a type which can be distinguished for special > > serialization treatment. > > > > Robert Ramey > > > > Ng Pan wrote: > >> Hi > >> > >> I am using boost serialization to synchronise objects between > >> different computers (network). The objects would be serializated as a > >> string instead of a file by using the class stringstream(instead of > >> fstream). The string would be transferred via TCP/IP layer. > >> > >> Let simpify the situation as only two computers (Master and Slave) > >> ,both of two computers have N objects and I wanna synchronise their > >> content. I create a vector of pointer ( std::vector<base *> m_vpBase > >> ) which points to the dirty objects. Both of two computers have this > >> vector. Let assume the vectors are already pointing to the > >> corresponding objects. For Tx side, it would serialize the objects as > >> a string. For Rx side, I want to update the content of the objects > >> that its vector is pointing to. > >> > >> Tx side: > >> oa << BOOST_SERIALIZATION_NVP(m_vpBase); > >> > >> Rx side: > >> oa >> BOOST_SERIALIZATION_NVP(m_vpBase ); > >> > >> However, the above code would construct new objects. > >> > >> What i want is just update the content of the objects , not construct > >> new objects. > >> > >> I know how to update the content for one object by using pointer ( > >> base * m_pBase). > >> Tx side: > >> oa << BOOST_SERIALIZATION_NVP(m_pBase); > >> > >> Rx side: > >> oa >> BOOST_SERIALIZATION_NVP(*m_pBase); > >> > >> > >> But, I don't know how to do for a vector of pointer. If you know, > >> please give me some hints. Thank you very much. > >> > >> Regards > >> Barr Ng > >> [I found Boost serialization is useful and easy to use. ] > >> > >> _________________________________________________________________ > >> No masks required! Use MSN Messenger to chat with friends and family. > >> http://go.msnserver.com/HK/25382.asp > > > > > > > > _______________________________________________ > > Boost-users mailing list > > Boost-users(a)lists.boost.org > > http://lists.boost.org/mailman/listinfo.cgi/boost-users > >_______________________________________________ >Boost-users mailing list >Boost-users(a)lists.boost.org >http://lists.boost.org/mailman/listinfo.cgi/boost-users _________________________________________________________________ Get 10Mb extra storage for MSN Hotmail. Subscribe Now! http://join.msn.com/?pgmarket=en-hk

3 4

[1.34][program_options] crash on release builds on windows using msvc-8.0
by Christopher Lux 17 May '07

17 May '07

hi, i compiled boost for windows using the msvc-8.0 and msvc-8.0_64 targets. the following code works perfectly on debug builds but crashes on the marked line in x64 mode and goes into an infinite loop in x86 mode. i used visual studio 2005 with and without service pack 1 and even tried the head version of boost. everytime the crash occurs. int main(int argc, char **argv) { int width; int height; try { boost::program_options::options_description cmd_options("program options"); cmd_options.add_options() ("help", "show this help message") ("width", boost::program_options::value<int>(&width)->default_value(1024), "output width") ("height", boost::program_options::value<int>(&height)->default_value(640), "output height"); boost::program_options::variables_map command_line; ***> boost::program_options::store(boost::program_options::parse_command_line(argc, argv, cmd_options), command_line); boost::program_options::notify(command_line); if (command_line.count("help")) { std::cout << "usage: " << std::endl; std::cout << cmd_options; return (0); } winx = width; winy = height; } catch (std::exception& e) { std::cout << e.what() << std::endl; return (-1); } } thanks for your help -chris -- Christopher Lux | | Bauhaus University Weimar | Faculty of Media - Virtual Reality Systems Group

1 0