December 2004 - Boost-users

Serialization: Cannot use name-value-pairs with indirectly serialized maps.
by Martin Proulx 10 Dec '04

10 Dec '04

Hello, While evaluating the serialization library, I've ran into a problem while modifying my code to add name value pairs to the serialized objects in order to be able to serialize them to an xml file. before doing this, everything worked fine, but now that I'm using NVPs, I get the following error: c:\projects\dev_tools\libs\c++\boost_1_32_0\boost\serialization\access.hpp(106): error C2039: 'serialize' : is not a member of 'std::map<_Kty,_Ty>' with [ _Kty=std::string, _Ty=std::string ] I'm using VC 7.1, and a stripped-down but compilable version of the code that causes the problem is included at the end of this mail. I'm really puzzled at this problem as the following modifications to the code will have things compile: -) Comment out only the 'oa << BOOST_SERIALIZATION_NVP(instanceOfB);' line. -) Comment out only the 'ia >> BOOST_SERIALIZATION_NVP(instanceOfB);' line. -) Serializing and unserializing directly the instanceOfA object instead. -) redefining BOOST_SERIALIZATION_NVP to 'BOOST_SERIALIZATION_NVP(x) x' is this really a bug or am I missing something ? Thanks! Martin ------------------------------------------------------------------------ #include <iostream> #include <fstream> #include <string> #include <boost/serialization/map.hpp> #include <boost/archive/text_iarchive.hpp> #include <boost/archive/text_oarchive.hpp> using namespace std; class A { private: friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int /* file_version */) { ar & BOOST_SERIALIZATION_NVP(mappedData_); } map<string, string> mappedData_; }; class B { private: friend class boost::serialization::access; template<class Archive> void serialize(Archive & ar, const unsigned int /* file_version */) { ar & BOOST_SERIALIZATION_NVP(privateInstanceOfA_); } A privateInstanceOfA_; }; int main(void) { A instanceOfA; B instanceOfB; std::ofstream ofs("test.txt"); boost::archive::text_oarchive oa(ofs); oa << BOOST_SERIALIZATION_NVP(instanceOfB); std::ifstream ifs("test.txt"); boost::archive::text_iarchive ia(ifs); ia >> BOOST_SERIALIZATION_NVP(instanceOfB); } -- Tel: (450) 681-1681, #271 ------------------------------------------------------------------------ OKIOK Enterprise and e-business security solutions Solutions de sécurité d'entreprise et d'affaires électronique Tel. : (450) 681.1681 http://www.okiok.com This e-mail message (including attachments, if any) is intended for the use of the individual or entity to which it is addressed and may contain information that is privileged, proprietary, confidential and exempt from disclosure. If you are not the intended recipient, you are notified that any dissemination, distribution or copying of this communication is strictly prohibited. If you have received this communication in error, please notify the sender and erase this e-mail message immediately. Le présent message électronique (y compris les pièces qui y sont annexées, le cas échéant) s'adresse au destinataire indiqué et peut contenir des renseignements de caractère privé ou confidentiel. Si vous n'êtes pas le destinataire de ce document, nous vous signalons qu'il est strictement interdit de le diffuser, de le distribuer ou de le reproduire. Si ce message vous a été transmis par erreur, veuillez en informer l'expéditeur et le supprimer immédiatement.

3 4

Some suggest to the author of Serialization Lib
by goodmen 10 Dec '04

10 Dec '04

Dear Robert Ramey , Here is a new boost user ( but an 'old' c++ programmer ) writting to you. Now , with several painful night , my test code of serialization lib has't run. OK , as some letter in the list , I meet the link problem. The libs has been compiled and linked with bjam correctly,and the lib files have been copied to the right position.But the compiler complained there are some symbol not resolved. I have inspected the source code,and find the serialization use some file in the [[archive]] dir. And in the docs, there are no any info about archive. It seems that archive is not an independent lib(bjam not compile the archive files,but copy the header file to the destination dir). Some classes in the archive are used by Serialization, but the implementation of these classes are not processed by bjam,so these class excite link errors(not compile errors!). Another feedback about your code is that There are not any C1204 compile error in MSVC71 if you build in the release mode. It seems that in debug mode, the compiler add addition info to the result binary and madee the scope of compiling became to too large to the compiler. Here are some questions about the serialization: 1, Is the archive written by you ? 2, Is the archive an independent lib? 3, how can I solve the link problem? 4, It seems that there is not a single header file that can be included by user who want to use the serialization(and the user can forget all other header files ). Do you have any plan to add this feature ? Here are my suggestions: 1, Take the archive to be an independent lib. 2, Move the whole serialization code to the head file,so there is no need to compile the lib with the bjam,and this can bring more protablity.( Since the io stream interface of stl is standard, I think it is no need to compile code into platform specific libs ) note: I am a Chinese , and my english is poor . If some word above offends you or seems to be impolite , forgive me please ^_^ 　　　　　　　　goodmen 　　　　　　　　goodmen(a)x263.net 　　　　　　　　　　2004-12-09

4 4

Re: [Boost-users] lu.hpp : permutation matrix error
by Daniel L Elliott 10 Dec '04

10 Dec '04

On Tuesday 07 December 2004 01:18 am, Toon Knapen wrote: > Daniel L Elliott wrote: > > Thank you for the response! I was unaware of this include-policy > > problem. Will you be merging this code soon? I will try to look for your > > email about this merge. > > Yes. I posted a mail on the ublas-dev just before the weekend about > this. Since no one objected I will go forward and do the merge (although > I'm not sure when exactly because there are a few bugs in the uBLAS_pure > branch that need to be resolved first) > I look forward to your notification that this has been done :) Thanks! > > Since adding > > #include<storage.hpp> to lu.hpp didn't fix the problem, should I expect > > to have to include additional files to my code to fix errors in files > > such as storage.hpp? > > Adding '#include <storage.hpp>' did not solve the problem ? Are you sure > you added this include before you include lu.hpp (or you can also add > the include to the top of you lu.hpp). The problem is that the default > value for the second template argument is an unbounded_array. This need > to be defined of course and the definition is in storage.hpp. I was able > to compile your testcase after adding this include. If this does not > work for you, send your testcase again that demonstrates the problem and > we'll have a look at it. OK, here is the code: #include<boost/numeric/ublas/storage.hpp> #include<boost/numeric/ublas/lu.hpp> using namespace boost::numeric::ublas; int main(){ // create a permutation matrix permutation_matrix P( 5 ); // default contructions gives the identity-permutation return 0; } The errors are: g++ -c -I. -I/s/chopin/a/grad/dane/thesis/code/boost-1.32/include/boost-1_32 -Wall -g3 -o permutationTest.o permutationTest.cpp In file included from permutationTest.cpp:5: /s/chopin/a/grad/dane/thesis/code/boost-1.32/include/boost-1_32/boost/numeric/ublas/lu.hpp: In function `typename M::size_type boost::numeric::ublas::axpy_lu_factorize(M&, PM&)': /s/chopin/a/grad/dane/thesis/code/boost-1.32/include/boost-1_32/boost/numeric/ublas/lu.hpp:181: error: parse error before `>' token /s/chopin/a/grad/dane/thesis/code/boost-1.32/include/boost-1_32/boost/numeric/ublas/lu.hpp:181: error: parse error before `)' token permutationTest.cpp: In function `int main()': permutationTest.cpp:15: error: `permutation_matrix' undeclared (first use this function) permutationTest.cpp:15: error: (Each undeclared identifier is reported only once for each function it appears in.) permutationTest.cpp:15: error: parse error before `(' token I thought that this error might signal the need for including vector.hpp (or vector_expression.hpp or something) but this did not fix it. Including operation.hpp (which I beleive contains the definition for axpy) causes several screen-fulls of errors! Again, I am at a loss and appreciate any help I might receive. Thank you, Dan elliott

2 1

Linker error when building regex (Boost 1.32.0) with Visual C++ 7.1
by Jochen Hammann 10 Dec '04

10 Dec '04

Hello, I tried to build the regex library from Boost 1.32.0 with the Microsoft Visual C++ 7.1 compiler and without native wchar_t support. Below is the call to bjam: bjam -sTOOLS=vc-7_1 "-sBUILD=<native-wchar_t>off" --with-regex The build results of the dynamic link library (DLL) shows the following linker errors: Creating library bin\boost\libs\regex\build\boost_regex.dll\vc-7_1\debug\native-wchar_t-off\threading-multi\boost_regex-vc71-mt-gd-1_32.lib and object bin\boost\libs\regex\build\boost_regex.dll\vc-7_1\debug\native-wchar_t-off\threading-multi\boost_regex-vc71-mt-gd-1_32.exp instances.obj : error LNK2001: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) winstances.obj : error LNK2001: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) regex.obj : error LNK2001: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) regex_debug.obj : error LNK2019: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) referenced in function "void __cdecl regex_lib_print_compiler_macros(void)" (?regex_lib_print_compiler_macros@@YAXXZ) w32_regex_traits.obj : error LNK2001: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) wide_posix_api.obj : error LNK2001: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) c_regex_traits.obj : error LNK2019: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) referenced in function "public: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::syntax_type(unsigned int)const " (?syntax_type@?$cpp_regex_traits@_W@boost@@QBAII@Z) c_regex_traits_common.obj : error LNK2001: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) cregex.obj : error LNK2001: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) posix_api.obj : error LNK2001: unresolved external symbol "private: unsigned int __cdecl boost::cpp_regex_traits<wchar_t>::do_syntax_type(unsigned int)const " (?do_syntax_type@?$cpp_regex_traits@_W@boost@@ABAII@Z) bin\boost\libs\regex\build\boost_regex.dll\vc-7_1\debug\native-wchar_t-off\threading-multi\boost_regex-vc71-mt-gd-1_32.dll : fatal error LNK1120: 1 unresolved externals ...failed vc-Link bin\boost\libs\regex\build\boost_regex.dll\vc-7_1\debug\native-wchar_t-off\threading-multi\boost_regex-vc71-mt-gd-1_32.dll bin\boost\libs\regex\build\boost_regex.dll\vc-7_1\debug\native-wchar_t-off\threading-multi\boost_regex-vc71-mt-gd-1_32.lib... ..skipped <@boost!lib>boost_regex-vc71-mt-gd-1_32.dll for lack of <@boost!libs!regex!build\boost_regex.dll\vc-7_1\debug\native-wchar_t-off\threading-multi>boost_regex-vc71-mt-gd-1_32.dll... ...skipped <@boost!lib>boost_regex-vc71-mt-gd-1_32.lib for lack of <@boost!libs!regex!build\boost_regex.dll\vc-7_1\debug\native-wchar_t-off\threading-multi>boost_regex-vc71-mt-gd-1_32.lib... The build of the static library succeeds. Do I have to define a special macro, if I want to build regex without native wchar_t support? Or do the regex library support both, native wchar_t support and the mapping to unsigned short? In the header regex_traits.hpp I saw some wrappers to "redirects calls to cpp_regex_traits<unsigned short> to cpp_regex_traits<__wchar_t>". Regards, Jochen.

2 2

Boost 1.32.0: lexcial_cast and wstring problem
by Jochen Hammann 10 Dec '04

10 Dec '04

Hello, I have tried to compile the following code with the Microsoft Visual C++ 7.1 compiler. #include <boost/lexical_cast.hpp> int main(int argc, char** argv) { using namespace std; using namespace boost; wstring wstrFoo = boost::lexical_cast<wstring>( 200 ); return 0; } If I compile this example with native wchar_t support (/Zc:wchar_t) everything works fine. Otherwise, if I compile the example without native wchar_t compiler support, I encounter the following compiler error: d:\Temp\Boost\boost\lexical_cast.hpp(150) : error C2679: binary '>>' : no operator found which takes a right-hand operand of type 'std::wstring' (or there is no acceptable conversion) d:\Temp\Boost\boost\lexical_cast.hpp(189) : see reference to function template instantiation 'bool boost::detail::lexical_stream<Target,Source>::operator >><Target>(InputStreamable &)' being compiled with [ Target=std::wstring, Source=int, InputStreamable=std::wstring ] d:\Temp\BoostConsole\BoostConsole.cpp(12) : see reference to function template instantiation 'Target boost::lexical_cast<std::wstring,int>(Source)' being compiled with [ Target=std::wstring, Source=int ] As I can see in the header file "lexical_cast.hpp" the macro BOOST_NO_INTRINSIC_WCHAR_T is defined, if the compiler option /Zc:wchar_t is not specified. This causes the macro DISABLE_WIDE_CHAR_SUPPORT to be defined as well and thus the lexical_stream::operator(std::wstring) will not be compiled. Do anyone know a workaround, so that I can use lexical_cast without native wchar_t compiler support and wstring? Is it really necessary to define DISABLE_WIDE_CHAR_SUPPORT, if BOOST_NO_INTRINSIC_WCHAR_T is defined only? For testing purposes I have commented out the two #ifndef DISABLE_WIDE_CHAR_SUPPORT statement. Now, I could compile the example without native wchar_t support. But I think, that this results in side effects, which I'm not able to see. Regards, Jochen.

1 0

Re: Re: Unresolved symbols in regex libs on tru64
by Mark Helzer 10 Dec '04

10 Dec '04

>> Sam results with static linking. >> I'm not good with reading mangled names but it looks to me like they >> are there. Am I perhaps missing a link option? > >I'm not familiar with the True64 linker so I can't say. What happens if you build the >examples with bjam - do those all build OK? > >John. OK I think I found it -- using the ansi object model $cxx -c ... -model ansi I have always just accepted the default - arm. It was in the man page for cxx and also in the documentation for the tru64cxx65 toolset. I should read more carefully. So it looks like all the code has to use ansi or arm, since I have a LOT of stuff compiled in the arm mode, it would be better to build boost using the "-sBUILD=<object-model>arm" switch, however now I'm back to compile errors: cxx: Error: /adp/local/src/boost_1_32_0/libs/program_options/build/../src/convert.cp p, line 62: #434 a reference of type "wchar_t *&" (not const-qualified) cannot be initialized with a value of type "wchar_t [32]" detected during instantiation of "std::basic_string<charT, std::char_traits<charT>, std::allocator<charT>> ...

3 2

Re: [Boost-users] simple regex compile error
by Brian Gaber 10 Dec '04

10 Dec '04

On AIX would it make a difference if I used the gcc compiler? If so, what would you recommend for the compiler command? Thanks. Brian

2 1

Newbie using multi_index
by Hammann, Jochen 10 Dec '04

10 Dec '04

Hello, I'm new in using the MultiIndex container of Boost 1.32.0. To get familiar with this library I wrote a small example. During the compilation with the Microsoft Visual C++ 7.1 compiler I encounter an internal compiler error. The error occurs, if I use the MultiIndex container with the following configuration: - Specification of tagged indices. - Specification of one composite_key template parameter. - Specification of at least three indices. Example: struct SFirst { }; struct SSecond { }; struct SBoth { }; typedef std::pair<std::wstring, std::wstring> Element_t; typedef multi_index_container<Element_t, indexed_by<ordered_non_unique<tag<SFirst>, member<Element_t, std::wstring, &Element_t::first> >, ordered_non_unique<tag<SSecond>, member<Element_t, std::wstring, &Element_t::second> >, ordered_unique<tag<SBoth>, composite_key<Element_t, member<Element_t, std::wstring, &Element_t::first>, member<Element_t, std::wstring, &Element_t::second> > > > > MultiIndex_t; //... int main(int argc, char** argv) { MultiIndex_t test; MultiIndex_t::index<SBoth>::type& both = test.get<SBoth>(); return 0; } If I compile this code I get the following internal compiler error: d:\Libraries\DLLs\Boost\Boost_1_32_0\inc\boost\tuple\detail\tuple_basic.hpp(612) : fatal error C1001: INTERNAL COMPILER ERROR (compiler file 'msc1.cpp', line 2701) Please choose the Technical Support command on the Visual C++ Help menu, or open the Technical Support help file for more information If I try to access the first (#0) and second (#1) index with their respective tags, everything works fine. Also, I'm able to access all three indices by their numerical value without an error. Is the declaration of the MultiIndex_t typedef wrong? Regards, Jochen.

4 5

[iostreams] Version of Iostreams compatible with 1.32 available
by Jonathan Turkanis 10 Dec '04

10 Dec '04

Hi All, A version of the Iostreams library designed to work with release 1.32 is now available: http://home.comcast.net/~jturkanis/iostreams/ This version does not incorporate all the changes decided on during the review, but most of the name changes have been applied. In particular, - The library is now contained entirely in the iostreams subdirectories of <boost> and <libs>. - Resources are now called Devices - The i/o mode Inout has been renamed Bidirectional - The concept Buffered has been renamed Multi-Character - Filters are now located in the directory boost/iostreams/filter - Devices are now located in the directory boost/iostreams/device The library has been tested on VC7.1, Intel 7.1-8.0 for Windows, Codewarrior 8.3 and GCC 3.2 (mingw). Unfortunately, I have noticed some regressions of GCC 3.3-4, but am working to correct them. I am confident that the library will eventually work on GCC 3.3+ and Borland 5.6.4. In addition, I may be able to provide limited support for VC6-7.0. I'm hoping to get a version into CVS soon. Also, several of the fixes that people sent me during the review have been removed because they caused regressions. They have not been forgotten -- they will be incorporated in to the final version in modified form. -- Jonathan Turkanis technews(a)kangaroologic.com

1 0

shared/weak_ptr conventions?
by Niklas Wiberg 10 Dec '04

10 Dec '04

Hi, I have a question that I believe is more a matter of taste than one of functional correctness... Say that I have a model (see Basket below) that formally owns the heap-objects (see Apple below). Basket has a function that is merely a query (hasApple) and involves no ownership-changes. Also assume that there is good reason that Apple:s are allocated on the heap... On one hand, I can see how using a weak_ptr as the parameter type would signal to users that hasApple has no intention of modifying ownership of the pointed-to object. On the other hand, in order to actually use the pointee (search for it in this case), hasApple must convert the weak_ptr to shared_ptr (lock() or shared_ptr constructor), so it would seem more pragmatic to just use shared_ptr as the parameter type. class Apple; class Basket { std::vector< boost::shared_ptr<Apple> > m_contents; bool hasApple( shared OR weak_ptr<Apple> a); // no transfer ...more functions }; Is there a "good style" concensus on how to choose smart pointer type for function arguments among boost smart pointer users? Best regards, Niklas Wiberg

2 1