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

[serialization] can't serialize to derived archive via base class ptr
by frankg 02 Sep '11

02 Sep '11

Hi, I try to serialize a derived object by its base class pointer to an archive I derived from xml_oarchive_impl exactly the same way as it is documented in the example code here: http://www.boost.org/doc/libs/1_47_0/libs/serialization/example/log_archive… http://www.boost.org/doc/libs/1_47_0/libs/serialization/example/log_archive… However, when I start the program (to output to my derived archive) I get the assertion (“BOOST_ASSERT(false)”) in the following method in “basic_serializer_map.cpp” (I checked with the debugger (VS2010) and found the m_map.size() is 0). BOOST_ARCHIVE_DECL(const basic_serializer *) basic_serializer_map::find( const boost::serialization::extended_type_info & eti ) const { const basic_serializer_arg bs(eti); map_type::const_iterator it; it = m_map.find(& bs); if(it == m_map.end()){ BOOST_ASSERT(false); return 0; } return *it; } If I use the DERIVED class pointer it works fine with my own derived archive. If I use a BASE class pointer and use the ORIGINAL xml_oarchive it works fine too. Only when I combine both (using my own derived archive AND try to serialize through a base class pointer) it fails with the assertion. Here is an excerpt of my code and 3 cases, where the first and second work fine and the third asserts: THIS WORKS FINE: int _tmain(int argc, _TCHAR* argv[]) { Base* bptr = new Derived; // using base class ptr boost::archive::xml_oarchive oa(std::cout, boost::archive::no_header); // using original archive from boost::serialization oa & BOOST_SERIALIZATION_NVP(bptr); return 0; } THIS WORKS FINE TOO: int _tmain(int argc, _TCHAR* argv[]) { Derived* dptr = new Derived; // using derived class ptr log_archive oa(std::cout, boost::archive::no_header); // using derived archive oa & BOOST_SERIALIZATION_NVP(dptr); return 0; } THIS ASSERTS: int _tmain(int argc, _TCHAR* argv[]) { Base* bptr = new Derived; // using base class ptr log_archive oa(std::cout, boost::archive::no_header); // using derived archive oa & BOOST_SERIALIZATION_NVP(bptr); return 0; } Is there something I can do to get my derived archive correctly working or doesn’t support the boost::serialization lib the combination of deriving archives and serializing via base class pointers at the same time? One more hint: using the debugger I saw that the m_map member in "basic_serializer_map.hpp" is statically initialized before main() is started. Some information for all my derived classes which contain serialization-methods where inserted into m_map. So m_map isn't actually empty but it seems to be a different object than the one where m_map has been initialized. Cheers, Frank -- View this message in context: http://boost.2283326.n4.nabble.com/serialization-can-t-serialize-to-derived… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

[MultiIndex] Iterator validity when modifying equal_range keys
by Ovanes Markarian 02 Sep '11

02 Sep '11

Hello *, reading this page states only that iterator validity is preserved for hashed indices. I have a multi-index container with ordered_non_unique entries. Is it safe to find a matching range and modify the keys there? Here an example: entry_data& entries = indexed; //get the view to the MI-Container for(iterator_range<entry_data::iterator> r = entries.equal_range(boost::make_tuple(false)) ; !r.empty() ; r.advance_begin(1) ) { index_data d = r.front(); d.masked = true; entries.replace(r.begin(), d); } Here updated d.masked field is a first member of the composite key. What happens if I do the update as shown upon? Would it be more efficient to iterate backwards and do the update, since than all the elements will not be rearranged? Is there also any way to use modify_key with composite index or are composite index key extractors always read-only? With Kind Regards, Ovanes

1 0

Re: [Boost-users] How to detect thread kill?
by Samvel Khalatyan 02 Sep '11

02 Sep '11

Thank you, Chris. On Fri, Sep 2, 2011 at 10:43 AM, Samvel Khalatyan <samvel(a)fnal.gov> wrote: > Thank you, Chris. > > > On Fri, Sep 2, 2011 at 10:36 AM, Chris Cleeland <chris.cleeland(a)gmail.com>wrote: > >> On Fri, Sep 2, 2011 at 10:24 AM, Samvel Khalatyan >> <samvel.khalatyan(a)gmail.com> wrote: >> > I guess you are right: the whole process is killed with all thread from >> > Linux command line. I suppose, there is no way to kill single thread. >> >> Not from the command line. pthread_kill() can direct a signal to a >> specific thread, but only from within the process that hosts the >> thread. >> >> The "pthreads primer" and "Programming Pthreads" books have some >> reasonable discussions on mixing signals and threads. Beware that >> they don't mix well. >> _______________________________________________ >> Boost-users mailing list >> Boost-users(a)lists.boost.org >> http://lists.boost.org/mailman/listinfo.cgi/boost-users >> > >

1 0

How to detect thread kill?
by Samvel Khalatyan 02 Sep '11

02 Sep '11

Hi, There is a number of standard ways to interrupt thread from within an application with library tools. However, is there a way to detect if one of the running threads was killed with system tools and thread is not available any more? thanks, Samvel.

3 7

[MultiIndex] cryptic gcc error
by Ovanes Markarian 02 Sep '11

02 Sep '11

Hello *, I try to use a multi_index container with a composite key. The container is defined as follows: struct index_data { entry_header header; //has various public fields one of them is timestamp_ fs::entry entry; bool masked; }; typedef size_t timestamp_t; //in reallity this type is a class, but the error can be reproduced with this typedef as well timestamp_t const& timestamp(index_data const& data) { return data.header.timestamp_; } bool mask_state(index_data const& data) { return data.masked; } struct entry_data_tag {}; typedef mi::multi_index_container < index_data, mi::indexed_by < mi::ordered_non_unique < mi::tag<entry_data_tag> , mi::composite_key < mi::global_fun<index_data const&, bool, &mask_state> //this could be a direct member reference, but I intially used fusion::map and reduced the problem to minimum , mi::global_fun<index_data const&, timestamp_t const&, &timestamp> > > > > entry_index; typedef entry_index::index<entry_data_tag>::type entry_data; void mask_all() { using namespace boost::lambda; entry_data& entries = pimpl_->indexed; entries.find(boost::make_tuple(false)); //!!!! this line results in an error novel by gcc 4.3 } The error description is: multi_index/composite_key.hpp:638: error: no match for 'operator*' in '*x' Does anyone have some suggestions why that happens? With Kind Regards, Ovanes

2 3

[serialization] Export macro doesnt work
by Jari 02 Sep '11

02 Sep '11

Hi I was trying to use BOOST_CLASS_EXPORT for all my classes in order to save them but I always got unregistered class error. So I used oa.register_type<object>(); for the class that was causing the error and problem solved. But why doesnt the BOOST_CLASS_EXPORT macro work? Why I need to use oa.register_type<object>(); ? Im using Boost 1.46.1 Thanks

3 10

complex matrix operations
by sairamesh raghuraman 02 Sep '11

02 Sep '11

Hello, I am looking for a C++ library that has routines for a variety of complex matrix decompositions (eigen, cholesky, singular value etc) and even complex matrix inverse, principal matrix logarthm of a complex matrix etc. And the matrices which I am talking are pretty huge..can go upto 1024 x 1024 x 32. I have already found a library that can do all of the above..used it..it works fine..but a disadvantage is that it cannot handle multidimensional arrays. Now the question is, does Boost C++ library have functions for the above? Looking forward to a reply and thanks in advance Ramesh

5 9

[Program Options] notifiers for options with no value
by Dmitriy Matison 01 Sep '11

01 Sep '11

Hello, I'm wondering if there is a possibility to make program_options library to notify me automatically when option with no value is passed? As far as I understand this feature exists only for options with defined value_semantic. But this means I must permit passing a value for such options. Say, I make an option with zero_tokens() semantics so it can be passed without value and throw an error if some value was provided. Unfortunately, this doesn't solve the problem if you have positional options. In that case a positional option following a no-value option will be parsed as this option's value. I can't set notifier for options with no value? This is something not understandable. I even think of some patch to be added to the lib. Regards, Dmitriy.

1 1

problem deserializing to base class pointer
by Shipa 01 Sep '11

01 Sep '11

I have a strange problem. I searched through the list and modified my initial code quite a bit but cannot seem to get this to work. Serialization seems fine, but deserialization to base pointer in particular doesn't seem to work correctly (I am getting instance of base class inside). The example below is canonical version of a much more complex case I was working with, involving shared_ptr and template classes, but it boiled down (for now) to this: A.h ---------------------------------------------------- #pragma once #include "boost/serialization/serialization.hpp" #include "boost/shared_ptr.hpp" class A { public: int t; A():t(0){}; //friend std::ostream& operator<<(std::ostream& os, const A& a); virtual ~A(){ }; friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & t; } }; std::ostream& operator<<(std::ostream& os, const A& a) { return os << "A: t = " << a.t << std::endl; } typedef boost::shared_ptr<A> APtr; ---------------------------------------------------- B.h ---------------------------------------------------- #pragma once #include "boost/serialization/serialization.hpp" #include "boost/shared_ptr.hpp" class B: public A { public: int m; B():A(),m(0){}; ~B(){}; //friend std::ostream& operator<<(std::ostream& os, const B& b); friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version) { ar & boost::serialization::base_object<A>(*this); ar & m; } }; std::ostream& operator<<(std::ostream& os, const B& b) { return os << "B: t = "<< b.t << "; m = "<< b.m << std::endl; } typedef boost::shared_ptr<B> BPtr; ---------------------------------------------------------------- Test.cpp ---------------------------------------------------------------- #include "stdafx.h" #include "boost/shared_ptr.hpp" #include "boost/serialization/shared_ptr.hpp" #include <boost/archive/text_oarchive.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/serialization/export.hpp> #include <istream> #include <ostream> #include <fstream> #include "A.h" #include "B.h" BOOST_CLASS_EXPORT(A) BOOST_CLASS_EXPORT(B) boost::archive::text_iarchive* GetiArchiveAndRegister(std::ifstream& f) { boost::archive::text_iarchive* ar= new boost::archive::text_iarchive(f); // ar->register_type(static_cast<A *>(NULL)); // ar->register_type(static_cast<B *>(NULL)); return ar; } boost::archive::text_oarchive* GetoArchiveAndRegister(std::ofstream& f) { boost::archive::text_oarchive* ar=new boost::archive::text_oarchive(f); // ar->register_type(static_cast<A*>(NULL)); // ar->register_type(static_cast<B*>(NULL)); return ar; } int _tmain(int argc, _TCHAR* argv[]) { A* a=new A(); a->t=1; B* b=new B(); b->t=2; b->m=1; A* c=new B(); c->t=2; std::ofstream f1("c:\\a"); boost::archive::text_oarchive* oa1=GetoArchiveAndRegister(f1); *oa1 & a; f1.close(); std::ofstream f2("c:\\b"); boost::archive::text_oarchive* oa2=GetoArchiveAndRegister(f2); *oa2 & b; f2.close(); std::ofstream f3("c:\\c"); boost::archive::text_oarchive* oa3=GetoArchiveAndRegister(f3); *oa3 & c; f3.close(); //deserializes fine std::ifstream f4("c:\\a"); boost::archive::text_iarchive* oa4=GetiArchiveAndRegister(f4); *oa4 & a; f4.close(); //deserializes an instance of A std::ifstream f5("c:\\b"); boost::archive::text_iarchive* oa5=GetiArchiveAndRegister(f5); *oa5 & a; f5.close(); //deserializes fine std::ifstream f6("c:\\c"); boost::archive::text_iarchive* oa6=GetiArchiveAndRegister(f6); *oa6 & a; f6.close(); //deserializes fine std::ifstream f7("c:\\b"); boost::archive::text_iarchive* oa7=GetiArchiveAndRegister(f7); *oa7 & b; f7.close(); std::ifstream f8("c:\\c"); boost::archive::text_iarchive* oa8=GetiArchiveAndRegister(f8); //runtime error: vector subscript out of range on the next line *oa8 & b; f8.close(); return 0; } ------------------------------------------------------------------ deserializations of f5 and f8 are controversial. I tried exchanging BOOST_CLASS_EXPORT calls with register_type call (commented out in GetiArchiveAndRegister and GetoArchiveAndRegister), and when that is done both serialization and deserialization finish correctly. I tried changing the order of includes, moving BOOST_CLASS_EXPORT around, etc. This version is what I'm settled for as the correct one. What am I doing wrong? Can anybody help?

2 1

Re: [Boost-users] [Threads] How to keep number of running thread constant, performing different operations
by Alessandro Candini 01 Sep '11

01 Sep '11

I've tryed implementing your suggestions with the following code: #include <iostream> #include </usr/local/include/boost/asio.hpp> #include </usr/local/include/boost/thread/thread.hpp> #include </usr/local/include/boost/bind.hpp> #include </usr/local/include/boost/date_time/posix_time/posix_time.hpp> #include </usr/local/include/boost/scoped_ptr.hpp> void handler1() { unsigned short int i; std::cout << "O--> Handler 1 <--O" << std::endl; for(i=0; i<65000; i++); std::cout << "X--> Handler 1 <--X" << std::endl; } void handler2() { unsigned short int i; std::cout << "O--> Handler 2 <--O" << std::endl; for(i=0; i<1000; i++); std::cout << "X--> Handler 2 <--X" << std::endl; } int main() { unsigned short nProc, i; boost::asio::io_service io; boost::scoped_ptr<boost::asio::io_service::work> work(new boost::asio::io_service::work(io)); boost::thread_group tg; nProc = boost::thread::hardware_concurrency(); for(i=0; i<nProc; ++i) tg.create_thread(boost::bind(&boost::asio::io_service::run, boost::ref(io))); io.post(handler2); io.post(handler1); io.post(handler2); io.post(handler2); io.post(handler2); io.post(handler2); io.post(handler2); io.post(handler2); io.post(handler2); io.post(handler1); io.post(handler2); work.reset(); tg.join_all(); return 0; } It seems to work, but it is correct, in your opinion? The only problem is that it seems to never stop if I increase the value of i in handler1 up to 70000. Have you any other suggestions? -- Alessandro Candini MEEO S.r.l. Via Saragat 9 I-44122 Ferrara, Italy Tel: +39 0532 1861501 Fax: +39 0532 1861637 http://www.meeo.it ======================================== "ATTENZIONE:le informazioni contenute in questo messaggio sono da considerarsi confidenziali ed il loro utilizzo è riservato unicamente al destinatario sopra indicato. Chi dovesse ricevere questo messaggio per errore è tenuto ad informare il mittente ed a rimuoverlo definitivamente da ogni supporto elettronico o cartaceo." "WARNING:This message contains confidential and/or proprietary information which may be subject to privilege or immunity and which is intended for use of its addressee only. Should you receive this message in error, you are kindly requested to inform the sender and to definitively remove it from any paper or electronic format."

3 2