October 2008 - Boost-users - lists.preview.boost.org

Re: [Boost-users] boost::mpl::map and inserting the same pair twice
by peter_foelsche＠agilent.com 06 Oct '08

06 Oct '08

I added the ::type to every insert. It still does not compile -- not using g++ nor using visual C++. Error message is not helpful. Also is there some operation to insert all elements of one map into another one? This would be extremely helpful. #include <boost/mpl/map.hpp> #include <boost/mpl/insert.hpp> #include <boost/mpl/size.hpp> #include <boost/type_traits/is_same.hpp> #include <boost/mpl/assert.hpp> typedef boost::mpl::map< boost::mpl::pair< boost::mpl::int_<0>, boost::mpl::int_<1> >, boost::mpl::pair< boost::mpl::int_<1>, boost::mpl::int_<0> > > CMap2; typedef boost::mpl::insert< CMap2, boost::mpl::pair< boost::mpl::int_<2>, boost::mpl::int_<3> >::type > CMap3; typedef boost::mpl::insert< CMap3, boost::mpl::pair< boost::mpl::int_<2>, boost::mpl::int_<3> >::type > CMap31; int main(int, char**) { CMap2 s; CMap3 s1; CMap31 s2; //BOOST_MPL_ASSERT((boost::is_same<CMap31, CMap3>)); return 0; } "The first point (using an init() function in preference to a constructor) is bogus. Using constructors and exception handling is a more general and systematic way of dealing with resource acquisition and initialization errors. This style is a relic of pre-exception C++." -- Stroustrup

2 1

Problem linking to Boost Filesystem Library
by DDBeatty＠dstsystems.com 06 Oct '08

06 Oct '08

I am getting an unresolved external error when auto-linking to the Boost Filesystem Library Compiler: Visual Studio 2005 Boost Library Version: 1.36.0 The error: packagereport.obj : error LNK2001: unresolved external symbol "class boost::system::error_code boost::filesystem::detail::throws" (?throws@detail@filesystem@boost@@3Verror_code@system@3@A) The Boost Libraries were compiled using Visual Studio 2005 as well. I compiled with BOOST_LIB_DIAGNOSTIC and got the following two lines in the output: Linking to lib file: libboost_system-vc80-mt-gd-1_36.lib Linking to lib file: libboost_filesystem-vc80-mt-gd-1_36.lib I added /verbose to the linker and got the following: Processed /DEFAULTLIB:libboost_system-vc80-mt-gd-1_36.lib Processed /DEFAULTLIB:libboost_filesystem-vc80-mt-gd-1_36.lib Searching ..\3rdParty\Boost\lib\libboost_system-vc80-mt-gd-1_36.lib: Found "class boost::system::error_category const & __cdecl boost::system::get_system_category(void)" (?get_system_category@system@boost@@YAABVerror_category@12@XZ) Referenced in packagereport.obj Loaded libboost_system-vc80-mt-gd-1_36.lib(boost_system-vc80-mt-gd-1_36.dll) Found "class boost::system::error_category const & __cdecl boost::system::get_generic_category(void)" (?get_generic_category@system@boost@@YAABVerror_category@12@XZ) Referenced in packagereport.obj Loaded libboost_system-vc80-mt-gd-1_36.lib(boost_system-vc80-mt-gd-1_36.dll) Found __IMPORT_DESCRIPTOR_boost_system-vc80-mt-gd-1_36 Referenced in libboost_system-vc80-mt-gd-1_36.lib(boost_system-vc80-mt-gd-1_36.dll) Referenced in libboost_system-vc80-mt-gd-1_36.lib(boost_system-vc80-mt-gd-1_36.dll) Loaded libboost_system-vc80-mt-gd-1_36.lib(boost_system-vc80-mt-gd-1_36.dll) Found boost_system-vc80-mt-gd-1_36_NULL_THUNK_DATA Referenced in libboost_system-vc80-mt-gd-1_36.lib(boost_system-vc80-mt-gd-1_36.dll) Loaded libboost_system-vc80-mt-gd-1_36.lib(boost_system-vc80-mt-gd-1_36.dll) Searching ..\3rdParty\Boost\lib\libboost_filesystem-vc80-mt-gd-1_36.lib: Found "class boost::filesystem::file_status __cdecl boost::filesystem::detail::status_api(class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const &,class boost::system::error_code &)" (?status_api@detail@filesystem@boost@@YA?AVfile_status@23@ABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@AAVerror_code@system@3@@Z) Referenced in packagereport.obj Loaded libboost_filesystem-vc80-mt-gd-1_36.lib(boost_filesystem-vc80-mt-gd-1_36.dll) Found "class boost::system::error_code __cdecl boost::filesystem::detail::remove_api(class std::basic_string<char,struct std::char_traits<char>,class std::allocator<char> > const &)" (?remove_api@detail@filesystem@boost@@YA?AVerror_code@system@3@ABV?$basic_string@DU?$char_traits@D@std@@V?$allocator@D@2@@std@@@Z) Referenced in packagereport.obj Loaded libboost_filesystem-vc80-mt-gd-1_36.lib(boost_filesystem-vc80-mt-gd-1_36.dll) Found __IMPORT_DESCRIPTOR_boost_filesystem-vc80-mt-gd-1_36 Referenced in libboost_filesystem-vc80-mt-gd-1_36.lib(boost_filesystem-vc80-mt-gd-1_36.dll) Referenced in libboost_filesystem-vc80-mt-gd-1_36.lib(boost_filesystem-vc80-mt-gd-1_36.dll) Loaded libboost_filesystem-vc80-mt-gd-1_36.lib(boost_filesystem-vc80-mt-gd-1_36.dll) Found boost_filesystem-vc80-mt-gd-1_36_NULL_THUNK_DATA Referenced in libboost_filesystem-vc80-mt-gd-1_36.lib(boost_filesystem-vc80-mt-gd-1_36.dll) Loaded libboost_filesystem-vc80-mt-gd-1_36.lib(boost_filesystem-vc80-mt-gd-1_36.dll) Using Dumpbin.exe I get the following: >From the dll: 68 43 00011E04 ?throws@detail@filesystem@boost@@3Verror_code@system@3@A >From the lib: ?throws@detail@filesystem@boost@@3Verror_code@system@3@A (class boost::system::error_code boost::filesystem::detail::throws) It looks like the decorated names in the lib and dll match the name that is being searched for by the linker and it looks like the linker found the lib file. Any ideas why this would not be linking? Also, I have tried disabling the auto-linking (using BOOST_ALL_NO_LIB) and linking manually, but I get the same error. Thanks, Doug Beatty ----------------------------------------- This e-mail and any attachments are intended only for the individual or company to which it is addressed and may contain information which is privileged, confidential and prohibited from disclosure or unauthorized use under applicable law. If you are not the intended recipient of this e-mail, you are hereby notified that any use, dissemination, or copying of this e-mail or the information contained in this e-mail is strictly prohibited by the sender. If you have received this transmission in error, please return the material received to the sender and delete all copies from your system.

1 0

*NIX: Boost ASIO and file descriptor limits
by Etienne Philip Pretorius 06 Oct '08

06 Oct '08

Hello list, I have a question regarding the system limit on a *NIX system for the maximum open file descriptors that a process may have open. If you were using the boost asio library to design a server application how would you go about and extend the server to handle more than the 1024 file descriptor limit. The common approach on *nix systems is to fork(2) the process and have the child accept the client connections and have the parent process inform the child process when the listening socket has clients waiting to be accepted. My question is how would I achieve a similar design approach. Any comments or ideas are welcome. Kind Regards, Etienne Pretorius.

3 8

[MPL] Generate template instantiation ?
by Olivier Tournaire 06 Oct '08

06 Oct '08

Hi all, I have a class templated on 2 arguments, e.g. : template <typename PixelType , typename NbChannels> class myClass { // ... }; This class is in a lib, and I would like to instantiate it for all base type (signed char, unsigned char, int, unsigned int ...), and for a given number of channels. For example : template class myClass<unsigned char,1>; template class myClass<unsigned char,2>; // ... template class myClass<unsigned char,10>; template class myClass<signed char,1>; template class myClass<signed char,2>; // ... template class myClass<signed char,10>; And so on. Is there a way to do this without writing all possiblities. Maybe with the MPL ? Best regards, Olivier

2 1

Exception handling inside a dll
by Matthew LaCrosse 06 Oct '08

06 Oct '08

Hello, I'm writing a JNI dll that can be called from a Java GUI. The java program passes a regex and a string to search. The dll function does a boost::regex_search and returns an IntArray with the results. The problem that I'm having is that bad expressions result in a crash. For instance, if the user inputs \b as the regex and 'bad' as the string, the program crashes.

4 10

Mutex is always blocking
by Jean-Sebastien Stoezel 06 Oct '08

06 Oct '08

Hello, Is there a specific way of initializing the boost::mutex objects? Here is the sequence I have been using: Instantiation of the mutex boost::mutex m_CmdMutex; Then try to lock it m_CmdMutex.lock(); It looks like every time I try to lock a mutex I lock my thread. Anything else needs to be done? Thanks, Jean

3 3

filesystem::recursive_directory_iterator quits on unreadable directory
by Charles Karney 06 Oct '08

06 Oct '08

Version: boost 1.36 Library: filesystem Platform: Linux Compiler: g++ 4.3.0 The increment operator in filesystem::recursive_directory_iterator throws an exception when trying to enter an unreadable directory (permissions, no-such-device, etc.). Often (usually?), a user wants to be able to skip the unreadable directory and continue the iteration over the rest of the directory tree. Catching this exception doesn't help because the iterator hasn't been incremented and I couldn't figure a way to skip over the unreadble directory. The following modification of filesystem/convenience.hpp catches the exception at a low enough level to allow the directory to be skipped over. Is there a better way of handling this? Thanks. --- /usr/local/include/boost-1_36/boost/filesystem/convenience.hpp 2008-09-30 17:34:10.000000000 -0400 +++ boost/filesystem/convenience.hpp 2008-10-02 16:40:43.000000000 -0400 @@ -239,10 +239,15 @@ else if ( is_directory( m_imp->m_stack.top()->status() ) ) { system::error_code ec; - m_imp->m_stack.push( - m_imp->m_no_throw - ? basic_directory_iterator<Path>( *m_imp->m_stack.top(), ec ) - : basic_directory_iterator<Path>( *m_imp->m_stack.top() ) ); + const Path& top = *m_imp->m_stack.top(); + m_imp->m_stack.push(end_itr); + try { + m_imp->m_stack.top() = + m_imp->m_no_throw + ? basic_directory_iterator<Path>( top, ec ) + : basic_directory_iterator<Path>( top ); + } + catch ( ... ) { } if ( m_imp->m_stack.top() != end_itr ) { ++m_imp->m_level; -- Charles Karney <ckarney(a)sarnoff.com> Sarnoff Corporation, Princeton, NJ 08543-5300 URL: http://charles.karney.info Tel: +1 609 734 2312 Fax: +1 609 734 2662

3 3

[MPL] for_each help needed
by Manuel Jung 06 Oct '08

06 Oct '08

Hey, I have problems using mpl::for_each. Well, it works for simple functors, but i need a functor with some context. I'm trying to build a functor through boost::bind. mpl::_1 should be a type from the consensus sequence if evaluated, shouldn't it? You can see a small (hopefully not too small) example below. Kind Regards Manuel class filter{ //[..] // "In" and "Out" are fusion::maps template< typename T > inline const boost::promise_stream<T>& get_promise(){ return fusion::at_key< T >(Out); }; template< typename T > inline void connect(const boost::promise_stream<T> &promise){ fusion::at_key< T >(In).reset(new boost::future_stream< T >(promise)); }; template< typename producer_type > void connect(producer_type Producer){ //[..] // "consensus" is a mpl sequence. mpl::for_each< consensus >(boost::bind(&producer_type::connect, Producer, get_promise< mpl::_1 >())); };

3 2

Cross-compile Boost, python include path problem
by raoul 06 Oct '08

06 Oct '08

Hi everyone, I'm trying to cross compile boost 1.36, and having an issue regarding boost.python compilation. In fact the cross compile is getting right. I'm using this script to cross compile boost: echo "using gcc : i686 : $CC ;" > tools/build/v2/user-config.jam cd tools/jam/src ./build.sh cp bin.*/bjam $BASE_DIR/toolchain/$TOOLCHAIN_VER/$TOOLCHAIN/bin/ cd ../../../ bjam -d2 --toolset=gcc-i686 --prefix=${ROOTFS_DIR}/usr --with-python install Boost.python compiles right, but when I want to link to the created library (libboost_python...) it failes with some strange errors: libboost_python-gcc-mt-1_36.so: undefined reference to `PyUnicodeUCS4_FromEncodedObject' libboost_python-gcc-mt-1_36.so: undefined reference to `PyUnicodeUCS4_AsWideChar' I googled a bit, and it seems that this problem is because libboost_python was not compiled with the same python version than the one I've corsscompiled with. And this is the problem. When reading the boost compilation log: "i686-unknown-linux-gnu-gcc" -ftemplate-depth-128 -O3 -finline-functions -Wno-inline -Wall -fPIC -pthread -DBOOST_ALL_NO_LIB=1 -DBOOST_PYTHON_SOURCE -DNDEBUG -I"." -I"/usr/include/python2.5" -c -o "bin.v2/libs/python/build/gcc-calaos/release/threading-multi/dict.o" "libs/python/src/dict.cpp" bjam is using the host include path for python (the -I/usr/include/python2.5 option), and this is wrong. How can I force bjam to use ${ROOTFS_DIR}/usr/include/python2.5 for python include path, instead of the one of my host system? I don't know how boost is checking and detecting python cflags. Using python-config? Is there a way to tell bjam to use ${ROOTFS_DIR}/usr/bin/python-config? Thanks, -- ------------------------ Raoul Hecky

2 2

boost::iostream exception
by Lloyd 06 Oct '08

06 Oct '08

Hi, I am trying to decompress a file using boost::iostreams (1.35.0) library built with zlib-1.2.3 . I am able to decompress the file with other utilities but this program throws me an exception saying "gzip error". I am able to uncompress other gzipped files. What could be the reason for the exception? Any help is very much appreciated. Thanks Lloyd The decompress function is also added here ========================================== bool ParseHttp::Decompress(const char* inBuff, int inLen,char** outBuff, int* OutLen,int type) { enum { CHUNK=4096 }; if(inBuff==NULL) return false; filtering_streambuf<input> in; switch(type) { case GZIP: in.push(gzip_decompressor()); break; case ZLIB: in.push(zlib_decompressor()); break; } in.push(boost::make_iterator_range((const char*)inBuff,(const char*)inBuff+inLen)); vector<vUnzip*> unZipV; int totLen=0; while(1) { char* data=new char[CHUNK]; int xyz=0; try { xyz=boost::iostreams::read(in,data,CHUNK); } catch(exception& e) { if(debug) cout<<e.what()<<endl; } unZipV.push_back(new vUnzip(data,xyz)); totLen+=xyz; if(xyz!=CHUNK) break; } *outBuff=new char[totLen]; int g=0; for(int k=0;k<unZipV.size();++k) { memcpy((*outBuff)+g,unZipV[k]->ptr,unZipV[k]->len); g+=unZipV[k]->len; delete unZipV[k]; } *OutLen=totLen; return true; } ______________________________________ Scanned and protected by Email scanner

1 0