May 2011 - Boost-users - lists.preview.boost.org

Timer Callback
by asif saeed 30 May '11

30 May '11

Hi, Can I use Boost Timer or Boost DateTime to trigger callbacks? I just want to be able to have the timer/date-time class trigger my handler functions after every 10 seconds for example or trigger a callback handling function after 25 seconds and die. Is that possible with Boost libraries? I am asking this question because, after a cursory look, it appears to me that there is no such facility right now - if this is true, do you suggest any portable library offering this functionality? Thank you, Asif

3 3

program_options: commands-like style
by Matan Nassau 29 May '11

29 May '11

Hi, How would I implement a program options with a commands style? Something similar to git and other version control systems: ./a.out ls --pattern '[a-z][0-9]+\.txt' --count 7 Here I run a.out with command "ls". The "ls" command has a particular options description. ./a.out read --count Dracula Here I run a.out with command "read". This command also has options of its own, and just to emphasize it I show that the --count option has a different value type here. ./a.out --help Show me general help. ./a.out help ls Does the library support this style? How would I get it to work? Matan

1 0

[exception] Using Boost.Exception infrastructure without using boost::exception
by Mostafa 29 May '11

29 May '11

I'm interested in using the infrastructure in Boost.Exception without using using boost::exception or deriving from it. Is this possible, if so, how? From the documentation, it seems so, but the documentation is not totally clear on this. From what I deduced, one should do: class my_exception {}; //1 throw boost::enable_current_exception(my_exception()); //2 catch(my_exception & ex) { /do stuff .... boost::rethrow_exception(boost::current_exception()); } Though, http://www.boost.org/doc/libs/1_46_1/libs/exception/doc/enable_current_exce… states: Note: Instead of using the throw keyword directly, it is preferable to call boost::throw_exception. This is guaranteed to throw an exception that derives from boost::exception and supports the exception_ptr functionality. And it's not very clear what exactly are the requirements for supporting exception_ptr functionality if one doesn't derive from boost::exception. Thanks, Mostafa

2 11

Python 3.2 (wide-unicode) with Boost 1.46.1 on Ubuntu 10.04 64bit LTS
by Nicholas Yue 29 May '11

29 May '11

Hi, I have a local build of Python 3.2 which have have used to build Python extensions and loaded them successfully in Blender 2.5 (which uses Python 3.2) I am hoping to migrate my bindings to using Boost.Python. I have been been getting error messages from bootstrap.sh and bjam NOTE : I have no problem building boost with Python 2.6.5 which comes with Ubuntu 10.04 LTS e.g. bootstrap.sh ./bootstrap.sh --prefix=/opt/boost/vTrunk-Python-3.2 --without-icu --with-python=/opt/python/v3.2/bin/python3.2 --with-python-version=3.2 >>>>>>>>Building Boost.Jam with toolset gcc... tools/build/v2/engine/src/bin.linuxx86_64/bjam >>>>>>>>Detecting Python root... File "<string>", line 1 >>>>>>>> import sys; print sys.prefix >>>>>>>> ^ >>>>>>>>SyntaxError: invalid syntax e.g. bjam ./bjam --prefix=/opt/boost/vTrunk-Python-3.2 --without-mpi >>>>>>>>gcc.compile.c++ bin.v2/libs/python/build/gcc-4.4.3/release/threading-multi/numeric.o >>>>>>>>In file included from ./boost/python/detail/prefix.hpp:13, >>>>>>>> from ./boost/python/numeric.hpp:8, >>>>>>>> from libs/python/src/numeric.cpp:6: >>>>>>>>./boost/python/detail/wrap_python.hpp:50:23: error: pyconfig.h: No such file or directory >>>>>>>>./boost/python/detail/wrap_python.hpp:75:24: error: patchlevel.h: No such file or directory >>>>>>>>./boost/python/detail/wrap_python.hpp:78:2: error: #error Python 2.2 or higher is required for this version of Boost.Python. >>>>>>>>./boost/python/detail/wrap_python.hpp:142:21: error: Python.h: No such file or directory My local build of python is deployed as per the configure script >>>>>>>>/opt/python/v3.2/ >>>>>>>>|-- bin >>>>>>>>|-- include >>>>>>>>| `-- python3.2mu >>>>>>>>|-- lib >>>>>>>>| |-- pkgconfig >>>>>>>>| `-- python3.2 >>>>>>>>`-- share >>>>>>>> `-- man The boost build problem might be directory path naming convention related. If anyone knows the correct to build Boost or Python so they play nice together, please share. Regards

1 1

Boost filesystem library on AIX 5.3
by Manoj Jangid 28 May '11

28 May '11

Hi I am getting following error when I am trying to build filesystem library vacpp.compile.c++ /home/mjangid/boost_rel2/boost/bin.v2/libs/filesystem/build/vacpp/release/threading-multi/v2/src/v2_operations.o "./boost/filesystem/v2/operations.hpp", line 796.73: 1540-1118 (S) The declaration of "end_itr" uses the undefined class "boost::filesystem2::basic_director y_iterator<boost::filesystem2::basic_path<std::basic_string<char,std::char_traits<char>,std::allocator<char> >,boost::filesystem2::path_traits> >" when the class must be complete. "./boost/filesystem/v2/operations.hpp", line 794.21: 1540-0700 (I) The previous message was produced while processing "boost::filesystem2::detail::remove_al l_aux<boost::filesystem2::basic_path<std::basic_string<char,std::char_traits<char>,std::allocator<char> >,boost::filesystem2::path_traits> >(const basic_pat h<std::basic_string<char,std::char_traits<char>,std::allocator<char> >,boost::filesystem2::path_traits> &, file_status)". "./boost/filesystem/v2/operations.hpp", line 497.28: 1540-0700 (I) The previous message was produced while processing "boost::filesystem2::remove_all<boost: :filesystem2::basic_path<std::basic_string<char,std::char_traits<char>,std::allocator<char> >,boost::filesystem2::path_traits> >(const basic_path<std::basic _string<char,std::char_traits<char>,std::allocator<char> >,boost::filesystem2::path_traits> &)". "./boost/filesystem/v2/operations.hpp", line 734.16: 1540-0700 (I) The previous message was produced while processing "boost::filesystem2::remove_all(const path &)". "./boost/filesystem/v2/operations.hpp", line 796.73: 1540-1118 (S) The declaration of "end_itr" uses the undefined class "boost::filesystem2::basic_director y_iterator<boost::filesystem2::basic_path<std::basic_string<wchar_t,std::char_traits<wchar_t>,std::allocator<wchar_t> >,boost::filesystem2::wpath_traits> >" when the class must be complete. "./boost/filesystem/v2/operations.hpp", line 794.21: 1540-0700 (I) The previous message was produced while processing "boost::filesystem2::detail::remove_al l_aux<boost::filesystem2::basic_path<std::basic_string<wchar_t,std::char_traits<wchar_t>,std::allocator<wchar_t> >,boost::filesystem2::wpath_traits> >(const basic_path<std::basic_string<wchar_t,std::char_traits<wchar_t>,std::allocator<wchar_t> >,boost::filesystem2::wpath_traits> &, file_status)". "./boost/filesystem/v2/operations.hpp", line 497.28: 1540-0700 (I) The previous message was produced while processing "boost::filesystem2::remove_all<boost: :filesystem2::basic_path<std::basic_string<wchar_t,std::char_traits<wchar_t>,std::allocator<wchar_t> >,boost::filesystem2::wpath_traits> >(const basic_path< std::basic_string<wchar_t,std::char_traits<wchar_t>,std::allocator<wchar_t> >,boost::filesystem2::wpath_traits> &)". "./boost/filesystem/v2/operations.hpp", line 736.16: 1540-0700 (I) The previous message was produced while processing "boost::filesystem2::remove_all(const wpath &)". xlC_r -c -DBOOST_ALL_NO_LIB=1 -DBOOST_FILESYSTEM_DYN_LINK=1 -DBOOST_SYSTEM_DYN_LINK=1 -DNDEBUG -qcpluscmt -O3 -qstrict -qfuncsect -qeh -qrtti -I"." -o "/home/mjangid/boost_rel2/boost/bin.v2/libs/filesystem/build/vacpp/release/threading-multi/v2/src/v2_operations.o" "libs/filesystem/v2/src/v2_operations .cpp" ...failed vacpp.compile.c++ /home/mjangid/boost_rel2/boost/bin.v2/libs/filesystem/build/vacpp/release/threading-multi/v2/src/v2_operations.o... ...skipped <p/home/mjangid/boost_rel2/boost/bin.v2/libs/filesystem/build/vacpp/release/threading-multi>libboost_filesystem-mt.so for lack of <p/home/mjangid /boost_rel2/boost/bin.v2/libs/filesystem/build/vacpp/release/threading-multi>v2/src/v2_operations.o... ...skipped <p/home/mjangid/boost_rel2/lib>libboost_filesystem-mt.so for lack of <p/home/mjangid/boost_rel2/boost/bin.v2/libs/filesystem/build/vacpp/release/ threading-multi>libboost_filesystem-mt.so... ...failed updating 1 target... can you please tell me what is wrong or what I am missing? Thanks Manoj Jangid

2 3

Wave - idl parsing
by g w 28 May '11

28 May '11

Hi, I need to parse and generate code for forbade 2.3 idl. Is wave the way to go? Is there a demo our there that shows idl parsing? If yes could you direct me to it pls? Or maybe wave is not the way to go. I'm doing nothing tricky in code generation, logging and operator<< code. Thanks for any info

2 2

[1.47.0] Release schedule
by Beman Dawes 28 May '11

28 May '11

* Monday, June 6: branches/release closed for all changes, except by permission of a release manager. * Monday, June 13: branches/release frozen. Only release managers may commit. Once tests have cycled and the release managers are satisfied with stability, release candidates for the beta will be generated. * Monday, Jun 20: Beta release target date. See June in http://www.boost.org/development/index.html or https://www.google.com/calendar/embed?src=5rorfm42nvmpt77ac0vult9iig@group.… The final release date won't be set until later in the process. Thanks, --Beman

1 0

Boost Unit Testing of DLL's
by Johan Doré 28 May '11

28 May '11

Dear Boost Mailing List We at trying to apply the Boost Unit Testing Framework to our Visual C++ 2010 projects and so far we must conclude that the framework is really excellent and does what we expect. Now most of our modules are developed as DLL's and as we of course want to keep the number of projects and compilation configurations down, we have discussed if we should just distribute the binaries including the test code even though it includes a slight overhead. The thing that is bugging us with that approach is however that the boost_unit_test_framework-vc100-mt-gd-1_46_1.dll seems to be needed for the DLL's to load and we therefore wonder if there is any way around that problem or if you have any other suggestions for us? Best regards Johan Doré, M.Sc. Vice President, CTO Phone : +45 88 20 20 88 Direct : +45 88 20 20 80 Mobile: +45 40 53 93 21 Fax : +45 88 20 20 99 Agern Allé 3 DK-2970 Hørsholm Denmark www.visiopharm.com <http://www.visiopharm.com> This e-mail including attachments is intended for the addressee(s) only and may contain confidential information protected by law. Unauthorised reading, disclosure, copying or distribution of this e-mail or use of information contained herein is strictly prohibited and could violate rights to proprietary information. If you are not an intended recipient, please notify the sender and delete it immediately hereafter. In no event will the sender or Visiopharm be liable for any damages, claims or costs whatsoever or any consequential, indirect incidental damages, any lost profits or lost savings arising from the contents of this mail.

2 1

Re: [Boost-users] Boost Unit Testing of DLL's
by Johan Doré 27 May '11

27 May '11

Hi, Thanks for the quick response. The reason why we would like to have the test code compiled into the modules themselves is simply because it would allow us to test functions and classes only available to the modules without having to export them out of the DLL. But also because it would limit the number of projects we have to compile. Currently without having separate unit testing projects we have more than 100 individual DLL's to which we want to apply UTF. STATIC LIBRARY APPROACH I have tried to link against the static library but for some reason ( Probably my novice level on this ) the console_test_runner-d.exe just tells me that: 1>EXEC : Test setup error : test tree is empty The way I implemented the static link was by adding the following code to the stdafx.cpp #include <boost/test/included/unit_test.hpp> extern "C" { __declspec(dllexport) bool init_unit_test() { boost::unit_test::framework::master_test_suite().p_name.value = "Imaging_Utilities tests"; return true; } } __declspec(dllexport) boost::unit_test::test_suite * init_unit_test_suite(int,char * * const) { boost::unit_test::framework::master_test_suite().p_name.value = "Imaging_Utilities tests"; return(&boost::unit_test::framework::master_test_suite()); } In the cpp's implementing the test's it looks like: #define BOOST_TEST_INCLUDED #include <boost/test/unit_test.hpp> BOOST_AUTO_TEST_SUITE(Polygon_Tests) BOOST_AUTO_TEST_CASE(Polygon_Tests_Clockwize_001) { . . Maybe you can see if I am doing something wrong in the above? DELAY LOADING APPROACH This was actually my preferred approach but when I specify boost_unit_test_framework-vc100-mt-gd-1_46_1.dll to be delay loaded, I can't even load my DLL. I don't even get to the DLL_PROCESS_ATTACH in my DLL. It looks like it is because macros like BOOST_AUTO_TEST_SUITE is designed to call functions (at load time) which seems to be implemented in boost_unit_test_framework-vc100-mt-gd-1_46_1.dll. The way I implemented the dynamic load was by adding the following code to the stdafx.cpp #define BOOST_TEST_DYN_LINK #include <boost/test/unit_test.hpp> extern "C" { __declspec(dllexport) bool init_unit_test() { boost::unit_test::framework::master_test_suite().p_name.value = "Imaging_Utilities tests"; return true; } } In the cpp's implementing the test's it looks like: #define BOOST_TEST_DYN_LINK #include <boost/test/unit_test.hpp> BOOST_AUTO_TEST_SUITE(Polygon_Tests) BOOST_AUTO_TEST_CASE(Polygon_Tests_Clockwize_001) { . . When the boost_unit_test_framework-vc100-mt-gd-1_46_1.dll is available all the tests runs nicely but without the DLL will not load not even if boost_unit_test_framework-vc100-mt-gd-1_46_1.dll is specified to be delay loaded. If possible we would prefer a solution like this as it would minimize compile time and the size of the binaries. So I am most curious to suggestions I should try to get this running. Best regards Johan Doré

2 1

[Asio] Common method of converting boost::asio::ip::address to network for both v4() and v6() addresses ?
by Sean Ryan 27 May '11

27 May '11

Greetings, As a short background, I have a protocol which transfers IPs by converting to a uint32_t and then shifting out the bytes on the other end. For example: 127.0.1.1 = 2130706689 Clearly, this won't work for an ipv6 address, as requires more bytes. The address 127.0.1.1 represents as: v6(): endpoint.address().to_string() yields ::ffff:127.0.1.1 v4(): endpoint.address().to_string() yields 127.0.1.1 Is there a method within the boost implementation that would allow ( or assist ) an effort to a uniform interpretation of the boost::asio::ip::address ( I expect this will likely involve the utin64_t over the 32 ... which I can live with ) ? Many thanks, Sean

1 0