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

Boost.Random - random_device error LNK2019 ()
by Valen 09 Apr '09

09 Apr '09

Hello there! Windows XP, Boost 1.38, VC++ 7.1 / Mingw 3.4.2 sample code: //----------------------------------------------------- #include <boost/random/uniform_int.hpp> #include <boost/random/variate_generator.hpp> #include <boost/nondet_random.hpp> int main() { boost::random_device m; typedef boost::uniform_int<> distribution_type; distribution_type generator2(m()); return 0; } //----------------------------------------------------- VC++ Output: error LNK2019: unresolved external symbol "public: __thiscall boost::random_device::~random_device(void)" (??1random_device@boost@@QAE@XZ) referenced in function "protected: void __thiscall CGraphDoc::OnBtnNewnode(void)" (?OnBtnNewnode@CGraphDoc@@IAEXXZ) error LNK2019: unresolved external symbol "public: unsigned int __thiscall boost::random_device::operator()(void)" (??Rrandom_device@boost@@QAEIXZ) referenced in function "protected: void __thiscall CGraphDoc::OnBtnNewnode(void)" (?OnBtnNewnode@CGraphDoc@@IAEXXZ) error LNK2019: unresolved external symbol "public: __thiscall boost::random_device::random_device(class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const &)" (??0random_device@boost@@QAE@ABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z) referenced in function "protected: void __thiscall CGraphDoc::OnBtnNewnode(void)" (?OnBtnNewnode@CGraphDoc@@IAEXXZ) error LNK2001: unresolved external symbol "private: static char const * const boost::random_device::default_token" (?default_token@random_device@boost@@0QBDB) fatal error LNK1120: 4 unresolved externals MinGW gave same output I checked the sources and find no implementation of ~random_device and etc... Can anyone help please? -- View this message in context: http://www.nabble.com/Boost.Random---random_device-error-LNK2019-%28%29-tp2… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

Passing reference to boost::bind object
by david.weber＠l-3com.com 08 Apr '09

08 Apr '09

So, we are implementing boost::signals/slots, and in several cases, we have a connect() function to allow us to hook up an arbitrary slot to the signal, without exposing the details of the signal for cleanliness issues. We would like to implement the signals::trackable interface, and have the objects which are passing in the bound slot manage their connections in this manner. However, we are passing them as boost::function objects, which (as documented) does not work properly with boost::function objects, only boost::bind objects. In all of the documentation that I've found, about assigning the results of a boost::bind call were to function objects. Is there an alternative available to not destroy the trackable functionality that we're trying to use? Thanks --dw

3 3

help building boost for iphone
by Kenny Riddile 08 Apr '09

08 Apr '09

I'm fairly inexperienced with boost.build and I'm having an issue building boost for the iphone. My user-config.jam (I put this in the same dir as the bjam executable right?) looks like this: using darwin : 4.0.1~iphone : /Developer/Platforms/iPhoneOS.platform/Developer/usr/bin/arm-apple-darwin9-g++-4.0.1 -arch arm : <striper> : <architecture>arm <target-os>iphone <macosx-version>iphone-2.2 ; using darwin : 4.0.1~iphonesim : /Developer/Platforms/iPhoneSimulator.platform/Developer/usr/bin/g++ : <striper> : <architecture>x86 <target-os>iphone <macosx-version>iphonesim-2.2 ; and I'm invoking bjam like this: ./bjam debug release toolset=darwin --layout=system link=static architecture=arm target-os=iphone macosx-version=iphone-2.2 My error is that it can't find g++ becuase it isn't looking in the right place: g++-4.0: installation problem, cannot exec '/usr/bin/arm-apple-darwin9-g++-4.0.1': No such file or directory

2 2

[graph] hash_set
by liquid 08 Apr '09

08 Apr '09

How to use hash_setS? If I define BOOST_HAS_HASH the compiler will generate "error C2977: 'stdext::hash_set' : too many template arguments". This is caused by using "BOOST_STD_EXTENSION_NAMESPACE::hash_set<Key,Eq,Hash,Alloc>" in definition of "template <class Key, class Eq, class Hash, class Alloc> struct container_traits< BOOST_STD_EXTENSION_NAMESPACE::hash_set<Key,Eq,Hash,Alloc> > { typedef set_tag category; typedef stable_tag iterator_stability; // is this right? };" Class hash_set is defined in STL as follows: template<class _Kty, class _Tr = hash_compare<_Kty, _STD less<_Kty> >, class _Alloc = _STD allocator<_Kty> > class hash_set : public _Hash<_Hset_traits<_Kty, _Tr, _Alloc, false> > {...}; which has only 3 params. Regards liquid

2 2

[serialization] appending elements
by Agustín K-ballo Bergé 08 Apr '09

08 Apr '09

Is it possible to append serialized elements to an existing archive? I need to keep a set of records on a file, and periodically add new entries to it (kind of like a database table). Reading the whole set, adding the new records, and serializing the set back is not an option since I expect the set to go pretty big. I have tried to keep several archives at the same stream (with no_header and no_tracking flags), but then I don't know how to read them. Reading them as if they were a single archive does not work, should it? Agustín K-ballo Bergé.-

4 5

[serialization] Test for end of archive
by Tom Aylesworth 08 Apr '09

08 Apr '09

Is there a way to test for the end of an archive, similar to testing for end of a stream? I have an indeterminate number of objects serialized to an archive and would like to loop until the end of the archive has been reached without reading past the end of the archive and getting a stream exception. Thanks! Thomas Aylesworth Senior Software Engineer Reality Mobile

1 0

Determining thread existence
by Jeremy Rottman 08 Apr '09

08 Apr '09

I am trying to setup my application so that it runs as a daemon. One aspect that I need to run a sanity check on is to determine if there are currently any parent threads running. Currently this is what I am testing with, when all threads are finish running, I then join all the threads to terminate them. Doing so should give me a thread group size of 0. However, this is now the case. Currently after I join the threads, the thread group size is still 0. Here is my current test code that I am working with. Any help with this is greatly appreciated. int main(int argc, char *argv[]) { InitConfig mconfig; if(mconfig.setConfig()){ boost::thread_group Tg; cout << "Thread Size Before: " << Tg.size() << endl; // DISPLAYS 0 for(int i = 1; i < 25; i++){ dmParse parse; Tg.create_thread(boost::bind(&dmParse::onInit,parse,i)); } cout << "Thread Size: " << Tg.size() << endl; // DISPLAYS 24 Tg.join_all(); cout << "Thread Count After:" << Tg.size() << endl; // DISPLAYS 24 }

3 8

a vector of subgraphs
by fabricio 08 Apr '09

08 Apr '09

Hello, I'm creating a program that will require a vector, of dynamic size, of subgraphs. so let's suppose I have something like this: typedef subgraph< adjacency_list > Graph; Graph G; what I want is to create a: vector< Graph& > subG; but the usage of & here is illegal. Do you guys have any suggestion on how to do it? regards,

4 8

Re: [Boost-users] [serialization] Export question
by Tom Aylesworth 08 Apr '09

08 Apr '09

I think I just answered my own question regarding whether I can use the EXPORT macro in the header that declares the class being exported. Since the EXPORT macro instantiates a static guid_initializer for the class, including it in multiple source files creates multiple static objects and results in link errors. So at this point, I wonder, why bother with the EXPORT macro at all? Given that the class has to be registered with the archive, why not just use register_type when the archive is created? When I thought the EXPORT macro allowed me to register the class in the same header where the class is created, it seemed very useful. Now I guess I don't see any advantage of using EXPORT. Am I missing something? Thomas Aylesworth Senior Software Engineer Reality Mobile

3 2

[mpl] newbie question about if_c et al
by Jonathan Leonard 08 Apr '09

08 Apr '09

Hi, I'm trying to make a class which is conditionally synchronized. What I have is like the following: template<bool SYNCH=true> class c { public: void f() { typename boost::mpl::if_c<SYNCH, ScopedCriticalSection<>, void*>::type scs(cs); // SCS == RAII wrapper around CS // do other stuff } private: typename boost::mpl::if_c<SYNCH, CriticalSection<>, void*>::type cs; }; This is not entirely optimal though as it requires 3 MOV instructions to create the scs void* copy when SYNCH is false (and I'd rather not lean on compiler optimizations even if they will help in this case). [Something like NullType, VoidType, or OptionType (a la F#) would be ideal but I can't seem to locate one (and I'm not sure the type system would allow such a thing-it doesn't seem to like a simple 'void']. Any better idiom or alternative approach anyone can offer for this purpose would be greatly appreciated. Many thanks, Jonathan

3 7

Re: [Boost-users] Linking with dynamic boost thread library
by Steven Watanabe 08 Apr '09

08 Apr '09

AMDG Green, Jason M NSWCDL, W33 wrote: > I'm having a similar problem with my project. I'd like to use the dlls but > no matter what, the compile warns me that it "cannot open file > 'boost_regex-vc80-mt-gd-1_38.lib'" > You /are/ linking to the dll then. > I'm trying to find where I can set the project (in vis studio 2005) to use > the dynamic link libraries as opposed to the static. Based on previous > entries in this thread, I assume that is the first step that I need to take. > Have you added the location of the libraries to the library paths? (Tools->Options->Projects and Solutions->VC++ Directories, I think) > Can someone tell me where to set this? > In Christ, Steven Watanabe

2 1

Re: [Boost-users] Linking with dynamic boost thread library
by Ovanes Markarian 08 Apr '09

08 Apr '09

Project Properties -> C/C++ -> Code Generation: Runtime Library: Multi-Threaded Debug DLL (/MDd) or Multi-Threaded DLL (/MD) for release version Good Luck, Ovanes On Wed, Apr 8, 2009 at 6:40 PM, Green, Jason M NSWCDL, W33 < jason.m.green3(a)navy.mil> wrote: > I’m having a similar problem with my project. I’d like to use the dlls > but no matter what, the compile warns me that it “cannot open file > ‘boost_regex-vc80-mt-gd-1_38.lib’” > > > > I’m trying to find where I can set the project (in vis studio 2005) to use > the dynamic link libraries as opposed to the static. Based on previous > entries in this thread, I assume that is the first step that I need to take. > > > > Can someone tell me where to set this? > > Jason > > > > *From:* boost-users-bounces(a)lists.boost.org [mailto: > boost-users-bounces(a)lists.boost.org] *On Behalf Of *Ovanes Markarian > *Sent:* Wednesday, April 08, 2009 11:01 AM > *To:* boost-users(a)lists.boost.org > *Subject:* Re: [Boost-users] Linking with dynamic boost thread library > > > > What I mean is the reversed order. You can not link the static runtime and > dynamic boost_thread or any other lib. > > On Wed, Apr 8, 2009 at 4:50 PM, Steven Watanabe <watanabesj(a)gmail.com> > wrote: > > AMDG > > > > Igor MI wrote: > > But now I wonder about what Ovanes said. Why did I linked successfully with > static thread library when my project is set to multi threaded debug dll? > > > > > You can link using the dll runtime and the static boost_thread. > > > > In Christ, > Steven Watanabe > > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users >

2 1

[ASIO] Synchronous write on serial_port
by Daniele Barzotti 08 Apr '09

08 Apr '09

Hi, I'm developing a library that use ASIO Serial Port and I need to implement a synchronous write. The structure is: EuroATLib.cpp : library main file Serial.cpp : serial port class that wrap asio::serial_port ATDispatcher.cpp : a dispatcher object ATModem.cpp : the consumer object The problem is that my code writes characters on serial port very slowly (1 char every 10 seconds) and I can't understand why. Also, when the program ends, the io_service thread seems to be already alive: it continues to write to serial port for a while! The implementation is like this...can someone give me some feedback? PS: In the following code TBUFFER is std:string. In EuroATLib.cpp : ---------------------------------------------------------- // Main io_service static boost::asio::io_service io_service; static boost::scoped_ptr<boost::thread> io_service_thread; /*! \brief Factory function for ATModem objects creation */ EXTERN_C EUROATLIB_API ATModem* CALL CreateATModemObj() { static bool thread_started; if (!thread_started) { try { // run the IO service as a separate thread io_service_thread.reset(new boost::thread( boost::bind(&boost::asio::io_service::run, &io_service) )); thread_started = !thread_started; } catch (boost::thread_resource_error e) { // Failed to create the new thread return 0; } } // create the new object return new ATModem(io_service); }; ---------------------------------------------------------- In Serial.hpp : ---------------------------------------------------------- // the main IO service that runs this connection boost::asio::io_service& _io_service; // the serial port this instance is connected to boost::asio::serial_port _serialPort; ---------------------------------------------------------- In Serial.cpp : ---------------------------------------------------------- SerialPort::SerialPort(boost::asio::io_service &_io) : m_buffer_size(100), _io_service(_io), _serialPort(_io_service), _rx_timer(_io_service), ReadTimeOut(READ_TIMEOUT) { //init_dcs(); EosString = '\n'; _buf = 0; }; BOOL SerialPort::Write(TBUFFER &buffer) { _io_service.post(boost::bind(&SerialPort::do_write, this, buffer)); return true; }; void SerialPort::do_write(TBUFFER &buffer) { // size_t byte_written = boost::asio::write(_serialPort, boost::asio::buffer(buffer.c_str(), buffer.size())); size_t byte_written = _serialPort.write_some(boost::asio::buffer(buffer.c_str(), buffer.size())); }; ---------------------------------------------------------- In ATDispatcher.cpp simply have a SerialPort member and a function SendCommand that start a timer and call Serial.Write(cmd) In ATModem.cpp I have an ATDispatcher Object on which it calls disp_obj.SendCommand("AT+CMGS"): ---------------------------------------------------------- RET_VAL ATModem::Connect(std::string DevName) { if (DevName.length() > 0) SetComPort(DevName); if (!_serial_port.Open()) return -1; //Connect the OnDataReceived event to the ATDispatcher object _dispatcher->OnData(boost::bind(&EuroATLib::ATModem::OnDataReceived, this)); //Register Voice Service if ( _dispatcher->SendCommand("AT+CTSP=2,0,0") != ATDispatcher::RET_OK ) return DEV_AT_CMD_ERROR; //Register STATUS Service if ( _dispatcher->SendCommand("AT+CTSP=2,2,20") != ATDispatcher::RET_OK ) return DEV_AT_CMD_ERROR; this->SetStatus(STATE_READY); }; ---------------------------------------------------------- Finally in the main client program: // Create a ATModem object boost::scoped_ptr<ATModem> _devATModem( CreateATModemObj() ); _devATModem->Connect("\\\\.\\COM1");

2 5

Linking with dynamic boost thread library
by Igor 08 Apr '09

08 Apr '09

I build the thread library with the following cmd command: bjam.exe toolset=msvc --with-thread --prefix=C:\boost\ variant=debug,release link=shared install after which I get the following libs: boost_thread-vc80-mt.lib boost_thread-vc80-mt-1_38.dll boost_thread-vc80-mt-1_38.lib boost_thread-vc80-mt-gd.lib boost_thread-vc80-mt-gd-1_38.dll boost_thread-vc80-mt-gd-1_38.lib In my application, I placed #define BOOST_THREAD_DYN_LINK in all the cpp files where thread library is included. I get the following error: LINK : fatal error LNK1104: cannot open file 'libboost_thread-vc80-mt-gd-1_38.lib' How many mistakes have I done?

5 7

Re: [Boost-users] difference b/w regex lib files
by John Maddock 08 Apr '09

08 Apr '09

>I have a number of boost libs that I built (not being sure which one I > really needed). I understand the difference b/w them, with the exception > of > two. libboost*.lib are static link libraries, and boost*.lib are import libraries for the dll's. HTH, John.

2 1

Re: [Boost-users] boost build interface
by Vladimir Prus 08 Apr '09

08 Apr '09

Green, Jason M NSWCDL, W33 wrote: > Not that I am volunteering, but it isn't that hard to come up with a simple > dialog based mfc project that would set up the commands to properly run the > boost build. I did not claim that coming up with an MFC project is hard per se. However: 1. It will require *me* quite some time, for I was not doing any real windows development in years. 2. It might be best to use some more lightweight solution, but I don't know what are scripting options on modern Windows. - Volodya

3 2

[date_time] local time & date arithmetics
by Filip Konvička 08 Apr '09

08 Apr '09

Hi all, first, thanks for the date_time library. Although I see some issues with it, I still think it is highly needed and does a lot of good stuff for me. Recently I stumbled upon a problem with date arithmetic that involves local_date_time. The following code fragment local_date_time t( ptime(boost::gregorian::date(2008, 4, 7), time_duration(22, 0, 0)), boost::shared_ptr<posix_time_zone>(new posix_time_zone("CET+1CEST,M4.1.0,M10.1.0"))); cout << t.utc_time() << " UTC = " << t << endl; t -= boost::gregorian::days(6); cout << t.utc_time() << " UTC = " << t << endl; yields the following output 2008-Apr-07 22:00:00 UTC = 2008-Apr-08 00:00:00 CEST 2008-Apr-01 22:00:00 UTC = 2008-Apr-01 23:00:00 CET This indicates that the date operation just subtracted 6*24 hours from the internal representation of the local time. It missed the DST change within the time interval. I would rather expect the operation to do the following: 2008-Apr-07 22:00:00 UTC = 2008-Apr-08 00:00:00 CEST 2008-Apr-01 23:00:00 UTC = 2008-Apr-02 00:00:00 CET If the current behavior is intentional, it is not clear to me why have such operations as "+= days(...)" available in local_date_time at all, as they don't seem to obey the general local time semantics (at least as I understand it) - we could just use an UTC ptime for that, right? Another thing related to this is that many functions in local_date_time are inherited from the base class, which has a different semantics. From the user point of view, this seems to be quite dangerous - one has always to verify whether the function being called is overloaded in local_date_time (and so should have local time semantics) or whether it is defined in ptime and so works with the internal UTC representation. Why is the base class not protected? Thanks for any comments, Filip

1 0

Re: [Boost-users] difference b/w regex lib files
by Steven Watanabe 08 Apr '09

08 Apr '09

AMDG Green, Jason M NSWCDL, W33 wrote: > I have a number of boost libs that I built (not being sure which one I > really needed). I understand the difference b/w them, with the exception of > two. > > > > All of the libs are libboost_regex..lib except 2 which are just > boost_regex...lib > > What is the difference between the 2? > The boost_regex*.lib files are import libraries for boost_regex*.dll In Christ, Steven Watanabe

1 0

difference b/w regex lib files
by Green, Jason M NSWCDL, W33 08 Apr '09

08 Apr '09

I have a number of boost libs that I built (not being sure which one I really needed). I understand the difference b/w them, with the exception of two. All of the libs are libboost_regex..lib except 2 which are just boost_regex...lib What is the difference between the 2? Jason

1 0

[Boost-Users][interprocess]ERROR_ACCESS_DENIED when removing a message_queue
by Laurence Chambenois 07 Apr '09

07 Apr '09

Hi, I'm a beginner with boost, so please don't throw things at me ... I'm using boost message_queue in a sample program, on Windows, that just does : - create the message_queue - remove the message_queue It works fine, if my program does this : try { message_queue mq1(create_only, "toto", 10, 100); } catch (interprocess_exception &err) { return -1; } message_queue::remove("toto"); but if it does : message_queue *mq1=0; try { mq1 = new message_queue(create_only, "toto", 10, 100); } catch (interprocess_exception &err) { return -1; } message_queue::remove("toto"); then the remove return ERROR_ACCESS_DENIED (on tthe DeleteFile call). I don't understand why. Is somebody can tell me what is wrong with my code ? Thanks for your reply.

4 3

interprocess windows_shared_memory implementation
by Brent Arias 07 Apr '09

07 Apr '09

The documentation repeatedly states that "windows_shared_memory has no kernel persistence and is destroyed when all processes using that shared memory are gone." But I'm trying to find out what exactly is used for the implementation of winows_shared_memory? Is it implemented internally as a Windows named pipe? Is it based on using ReadProcessMemory? Anyone know? Brent Arias Software Architect Lumension Security(tm) Tel ~ (602) 692-8088 Fax ~ (480) 970-6323 brent.arias(a)lumension.com<mailto:brent.arias@lumension.com>

3 2

[Serialize] Serialize just the base class of polymorphic pointers
by Brandon Aubie 07 Apr '09

07 Apr '09

I have a base class BASE that stores data. It has a virtual function (empty in base class) that determines how this data is calculated, but there's no data stored in the derived classes. In my program, I'll load a pluging that is based on BASE and implements the function to calculate the data. Right now I have BOOST_EXPORT_CLASS in the plugin header files and I serialize boost::serialization::base_object<BASE>(*this) in the derived class. This lets me save the serialization to file just fine. When I load it back in though, my application isn't aware of the plugins (I'm just analyzing the data). But I don't need the plugins, since I just need the data. So when I try to load in my serialization, I get an "unregistered class" exception. Here's how it works /*************** CODE ******************/ class BASE { public: vector<double> data; virtual calculate() {}; private: friend class boost::serialization::access; template<class Archive> void serialize(Archive &ar, const unsigned int version) { ar & data; } } class Plugin: public BASE { public: virtual calculate() { /* do stuff */ } private: friend class boost::serialization::access; template<class Archive> void serialize(Archive &ar, const unsigned int version) { ar & boost::serialization::base_object<BASE>(*this); } } BOOST_CLASS_EXPORT(Plugin) /*************** CODE ******************/ Class Plugin is loaded dynamically at runtime via dlopen() on linux. So I have a BASE* that points to an instance of A. When I load in the serialized data though, I don't really care about A and would rather just have a BASE* that points to an instance of BASE. Since there's no data described in A. Can I do this somehow? Thanks

2 1

Re: [Boost-users] [mpl]... is there an mpl::string
by raindog＠macrohmasheen.com 07 Apr '09

07 Apr '09

L$$m$Ll$l$l$ ------Original Message------ From: Eric Niebler Sender: boost-users-bounces(a)lists.boost.org To: boost-users(a)lists.boost.org ReplyTo: boost-users(a)lists.boost.org Sent: Apr 7, 2009 10:21 AM Subject: Re: [Boost-users] [mpl]... is there an mpl::string Noah Roberts wrote: > Eric Niebler wrote: >>>>> Andy Stevenson wrote: >>>>>> >>>>>> I recall some discussion of there being an mpl::string >>>>>> template..... Can't see it in the mpl library. Is it elsewhere? >> >> It has been added to trunk as of revision 52208: >> >> https://svn.boost.org/trac/boost/changeset/52208 >> >> No doubt the regression tests will reveal portability problems. Once >> they have been worked out, we can move this to release. > > Why is this better than a metafunction c_str< Sequence >? > > typedef mpl::vector_c<char, 'h', 'e', 'l', 'l', 'o'> str; > > template < char const* sz > > struct x > {}; > > x< c_str<str>::value > test; > > Is there some reason that's not possible or is prone to problems avoided > by mpl::string? That is certainly a valid design. It's somewhat subjective, but IMO multi-character literals give a nicer compile-time string interface. Consider: // With mpl::vector_c mpl::vector_c<char, 'h','e','l','l','o',' ','w','o','r','l','d'> // With mpl::string mpl::string<'hell','o wo','rld'> Neither will win a beauty contest, but my preference is strongly for the latter. -- Eric Niebler BoostPro Computing http://www.boostpro.com _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users Sent from my Verizon Wireless BlackBerry

1 0

[program_options] Specify the same flag multiple times
by KSpam 07 Apr '09

07 Apr '09

I would like to be able to specify a flag multiple times on the command line. For instance, imagine a program that has two levels of verbosity. The first level of verbosity would be activated by specifying "-v" on the command line, and the second level of verbosity would be activated by specifying "-vv" or "-v -v" on the command line. Display verbose print statements: myProgram -v Display very verbose print statements: myProgram -vv In digging through the documentation, it is not evident to me how I can accomplish this with Boost.Program_Options. Any help would be appreciated. Thanks, Justin

1 0

building Boost with -fPIC on x86_64 linux box
by Francois Mauger 07 Apr '09

07 Apr '09

Hi boosters, He is my problem: I have a AMD 64 Linux box with boost installed on it, particularly with the Boost.Python component. Boost version is 1.34.1 (an old one but the same problem occurs with new 1.38.0 as explained below). also gcc version is 3.4.6. I use the standard: ./configure --with-xxxx; make; make install; sequence to build/install the libs. Unfortunately I cannot figure out if the -fPIC is activated while compiling the code... and I don't know how to check the PIC status from the resulting .so files (it is possible and does it make sense? I'm not an expert with this concept). I only know that if I missed the -fPIC flag on this system, I have linker errors. Next step, I have my own 'datatools' library with some classes using Boost stuff. I use plain makefiles for it and I garantee the -fPIC flags is activated to generated my 'libdatatools.so' shared library. Finally, I try to build a Boost.Python wrapper for my classes, trying to link against the 'libdatatools.so' file and 'libboost-xxxx.so' stuff. [Note this procedure works perfectly under 32 bits ubuntu system with gcc 4 and boost 1.34.1] Unfortunately I get this nasty well known error: >>> g++ datatools.cc ... -I.../sw/boost/install-1_38_0-Linux-x86_64-gcc34/include/boost-1_38 -I/usr/local/include/python2.5 -fno-strict-aliasing -DN DEBUG -g -O3 -Wall -I.../sw/datatools/pro/include -I.-shared -fPIC -O -DBOOST_PYTHON_DYNAMIC_LIB -L../Linux-x86_64/lib -ldatatools -lboost_iostreams-gcc34-mt -lboost_filesystem-gcc34-mt -fPIC -L.../sw/boost/install-1_38_0-Linux-x86_64-gcc34/lib -lboost_python-gcc34-mt -lboost_serialization-gcc34-mt -L/usr/local/lib/python2.5/config -lpthread -ldl -lutil -lpython2.5 -lm -o datatools.so /usr/bin/ld: /usr/local/lib/python2.5/config/libpython2.5.a(abstract.o): relocat ion R_X86_64_32 against `a local symbol' can not be used when making a shared object; recompile with -fPIC /usr/local/lib/python2.5/config/libpython2.5.a: could not read symbols: Bad value collect2: ld returned 1 exit status make: *** [datatools.so] Error 1 <<< So what is the problem? I used -fPIC for my classes and I used -fPIC during the Boost-Python wrapping procedure. Does it mean the Boost library itself has not been compiled with -fPIC despite the fact that it is recommended/mandatory with 64 bits architecture. Or does it mean there is something wrong with 'libpython2.5.a' which is provided by the sysadmin and out of my privilege scope? It seems this kind of issue has already been addressed before but scanning the list archives left me rather confused with no clear solution/hint. I tried to enforce the -fPIC flags adding the following line in user-config.jam with boost 1.38.0: >>> using gcc : : : <cxxflags>-fPIC ; <<< But it did not help... I'd really appreciate any hints about this issue. Thanks a lot. regards frc -- François Mauger Département de Physique - Université de Caen Basse-Normandie courriel/e-mail: mauger(a)lpccaen.in2p3.fr tél./phone: 02 31 45 25 12 / (+33) 2 31 45 25 12 fax: 02 31 45 25 49 / (+33) 2 31 45 25 49 Adresse/address: Laboratoire de Physique Corpusculaire de Caen (UMR 6534) ENSICAEN 6, Boulevard du Marechal Juin 14050 CAEN Cedex FRANCE

4 4