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

date_time, g++-4.7.0, compile error
by List User 29 Apr '12

29 Apr '12

The following code gives compile error with g++-4.7.0 but compiles fine with g++-4.6. #include <iostream> #include <boost/date_time/local_time/local_time.hpp> using namespace std; int main(){ boost::posix_time::ptime time_t_epoch(boost::gregorian::date(1970,1,1)); cout << time_t_epoch << endl; return 0; } The following is the error message that is repeatedly seen (compiler outputs a lot of messages) /usr/include/boost/date_time/local_time/local_date_time.hpp:433:84: error: use of deleted function boost::shared_ptr<boost::date_time::time_zone_base<boost::posix_time::ptime, char> >::shared_ptr(const boost::shared_ptr<boost::date_time::time_zone_base<boost::posix_time::ptime, char> >&) I am using Ubuntu 12.04 and libboost-date-time1.46.1. Any suggestions? thanks suresh

2 2

[wserialization] serializing to a stringstream
by Sean Farrow 29 Apr '12

29 Apr '12

Hi: I'm using xml serialization and need to serialize to a string stream for encryption purposes. As I am using xml_woarchive, should I use wstringstream or just the normal stringstream? Does the wide version even exist? Essentially does the boost serialization library convert to narrow characters before serializing or not? Any help appreciated. Regards Sean.

2 1

How to send serialized data via boost::interprocess::message_queue
by Jens Saathoff 28 Apr '12

28 Apr '12

Hi! I'm working with message_queue and want to send a structure to receiver-app. How to serialize i correctly and how to receive it? Here's some sending code: Data data; > data.counter = 9; > data.money = 5.4; > //data.name = "Guten Tag"; > > std::stringstream os; // = new std::stringstream; > boost::archive::text_oarchive oa(os); > oa << data; > > ipc::message_queue::remove("mq_main"); > ipc::message_queue mq(ipc::create_only, "mq_main", 1, > sizeof(std::stringstream)); > mq.send(&os, sizeof(os), 0); I implemented the serializing code in the structure "Data" Add here's my code to receive: Data data; > std::stringstream is; // = new std::stringstream; > try > { > unsigned int priority; ///< The priority > ipc::message_queue::size_type recvd_size; > ipc::message_queue mq(ipc::open_only, "mq_main"); > > if(mq.try_receive(&is, sizeof(std::stringstream), recvd_size, > priority)) > { > boost::archive::text_iarchive ai(is); // CRASH > ai >> data; > } > //mq.receive(&is, sizeof(is), recvd_size, priority); > //mq.do_receive(ipc::message_queue_t::non_blocking, &is, > sizeof(is), recvd_size, priority); > > //boost::archive::text_iarchive * ai = new > boost::archive::text_iarchive(is); > > //*ai >> data; > > //data << ai; > //mq_recv->receive(&data, sizeof(data), recvd_size, priority); > > ipc::message_queue::remove("mq_main"); > //is.~basic_stringstream(); > } > catch(ipc::interprocess_exception &ex) > { > ipc::message_queue::remove("mq_main"); > std::cout << ex.what() << std::endl; > return; > } > The receiver-app crashes at boost::archive::text_iarchive ai(is). The error is: 0xC0000005: Access violation reading location 0x00f56a34. I really don't have a clue...pls help. Thank you very much :-)

1 0

serialize derived class object through boost::serialization
by breadbread1984 28 Apr '12

28 Apr '12

I want to serialize a derived class object through boost::serialization library. To test whether both the base and the derived class object have been serialized properly, I print message to the console when the serialize function is called. When the following code is compiled and runned, the message of serialize function for the base class object is not shown. Thus, the base class object is not serialized properly. Why the serialization is not done on the base class object? #include <cstdlib> #include <iostream> #include <fstream> #include <boost/serialization/version.hpp> #include <boost/serialization/export.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/archive/text_oarchive.hpp> #include <boost/shared_ptr.hpp> using namespace std; using namespace boost::archive; using boost::shared_ptr; class Base { friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version); protected: int a; public: Base():a(0){} virtual ~Base(){} int geta() const {return a;} void seta(int aa) {a = aa;} }; template<class Archive> void Base::serialize(Archive & ar,const unsigned int version) { cout<<"serializing base class object"<<endl; ar & a; } class Derived : public Base { friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int version); protected: int b; public: Derived():Base(),b(0){} virtual ~Derived(){} int getb() const {return b;} void setb(int bb) {b = bb;} }; template<class Archive> void Derived::serialize(Archive & ar,const unsigned int version) { boost::serialization::base_object<Base>(*this); boost::serialization::void_cast_register<Derived,Base> ( static_cast<Derived*>(NULL), static_cast<Base*>(NULL) ); cout<<"serializing derived class object"<<endl; ar & b; } BOOST_CLASS_EXPORT_GUID(Derived,"Derived"); int main(int argc,char ** argv) { shared_ptr<Base> ptr(new Derived()); ptr->seta(4); Base * p = ptr.get(); Derived * pp = dynamic_cast<Derived*>(p); assert(pp); pp->setb(4); ofstream out("test.dat"); text_oarchive oa(out); oa << p; return EXIT_SUCCESS; }

2 1

undefined symbol: PyUnicodeUCS4_FromEncodedObject
by Dylan Dandelion 28 Apr '12

28 Apr '12

Is there a way to specify boost.python to be built with UCS2 instead of USC4? I am trying to build some libraries and ran into the following error: ImportError: /home/user/boost_1_49_0/stage/lib/libboost_python.so.1.49.0: undefined symbol: PyUnicodeUCS4_FromEncodedObject I am using a version of Python that seems to use only UCS2. # python Python 2.7.2 (default, Aug 24 2011, 22:13:27) [GCC 4.2.4 (Ubuntu 4.2.4-1ubuntu4)] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import sys >>> print sys.maxunicode 65535 >From what I searched online so far, there are some suggestions to recompile Python with some flags to use UCS4, but that is not possible for me. Thanks!

1 0

Do boost run well on x64 platform with such warning?
by clive 27 Apr '12

27 Apr '12

Hi,my friends I have a stupid question Now i build boost 1.49 on x64 platform with vc2005 I find such many warnings: warning C4267: '=' : conversion from 'size_t' to 'unsigned int', possible loss of data or conversion from '_int64' to 'unsigned int', possible loss of data I want to make sure that these warnings really do lose data whether or not on x64 flatform. Put another way, could these warning be neglected or should i change the source code? Thanks

2 1

Re: [Boost-users] [xpressive][proto] Generates VS2010 C4714 in 1.49 but not in 1.48
by Lawry, Brian 27 Apr '12

27 Apr '12

Hi Eric. I replaced everything under ..\boost-1_49\boost\proto with the version from trunk yesterday, but I still get the warnings. "/DBOOST_FORCEINLINE=inline" works for me (performance is not an issue in my application). Thanks. Brian >Message: 1 >Date: Wed, 25 Apr 2012 12:44:46 -0700 >From: Eric Niebler <eric(a)boostpro.com> >To: boost-users(a)lists.boost.org >Subject: Re: [Boost-users] [xpressive][proto] Generates VS2010 C4714 > in 1.49 but not in 1.48 >Message-ID: <4F98542E.9080400(a)boostpro.com> >Content-Type: text/plain; charset=ISO-8859-1 >On 4/24/2012 11:33 PM, Eric Niebler wrote: >=> On 4/24/2012 2:48 PM, Lawry, Brian wrote: >>> I would be OK with suppressing the warning, but I haven't been able >>> to do so. Putting "#pragma warning(disable : 4714)" in my source >>> file and in boost/xpressive/detail/core/ sub_match_vector.hpp and in >>> boost/proto/transform/call.hpp does not suppress the warning. Using >>> the "disable specific warnings" compiler switch /wd4714 also does not >>> work. >I checked this and found that the warning can in fact be suppressed. I went through >proto's source files and suppressed it everywhere __forceinline is used. Your sample >code now compiles cleanly with /W4 with boost trunk. >Thanks for the report. >-- >Eric Niebler >BoostPro Computing >http://www.boostpro.com

1 0

Re: [Boost-users] Do boost run well on x64 platform with such warning?
by John Drescher 27 Apr '12

27 Apr '12

> I have a stupid question > Now i build boost 1.49 on x64 platform with vc2005 > I find such many warnings: warning C4267: '=' : conversion from 'size_t' to > 'unsigned int', possible loss of data or conversion from '_int64' to > 'unsigned int', possible loss of data > I want to make sure that these warnings really do lose data whether or not > on x64 flatform. > Put another way, could these warning be neglected or should i change the > source code? Just neglect these warnings. These would only be a problem if size_t had > 4 billion elements. I certainly do not put that many elements into my structures yet! John

1 0

[asio] Getting End of file on ssl::stream::shutdown
by Tak Sze 26 Apr '12

26 Apr '12

I have a very simple SSL client and server program that send messages back and forth using asio ssl stream. Both sides send a "no more message" message when they are done. Then they both call shutdown to shutdown the SSL connection. However, I kept getting "shutdown: End of file" from shutdown. Does anyone know why I am getting that? - Tak

1 0

error: partial specialization after instantiation when using boost.atomic
by niXman 26 Apr '12

26 Apr '12

Hello! I try to use boost.atomic library[1] and when compiling this trivial code: > #include <iostream> > #include <boost/atomic.hpp> > int main() {} http://liveworkspace.org/code/3c355fc61db67d52266a66103bce325a and I get the following error: > boost/atomic/detail/integral-casts.hpp:155:7: > error: partial specialization of 'boost::detail::atomic::platform_atomic<T, 4u>' > after instantiation of 'boost::detail::atomic::platform_atomic<int>' I try to find the 'boost::detail::atomic::platform_atomic<int>' specialization, but couldn`t find it. Maybe someone faced with a similar error. Or maybe someone has an idea what kind of error is this, in fact? Thanks! [1] http://www.chaoticmind.net/~hcb/projects/boost.atomic/ -- Regards, niXman

1 0