February 2009 - Boost-users - lists.preview.boost.org

[Q] storing function pointers to member function
by V S P 20 Feb '09

20 Feb '09

I am trying to come up with a system that would let me have a function (getter) for every member of the class Using macros I can generate a 'get_<membername> function such that struct my { float mem1; } would be struct my { float mem1; static const float& get_mem1( const my& ) { return my.mem1; } }; Now, I would like to store pointer to the get_mem1 function in a staticaly declared std::map Using boost/function.hpp I can declare a type for float typedef boost::function<const float& (const my& ,const char* )> tFget__float ; and then I can store it in, for example std::set<tFget__float> The problem is that if I have another function that returns a different data type (something other than float, for example std::string) I get a compilation error when trying to insert that function pointer into the set. There is a check implemented, I think, via the template system that prevents compiler from making copies of one ponter type to another -- even though they only differ by return type. When I just do the assignment tFget__float x= <function returning float> or to function returning std::string -- I do not get the error. So I would like to ask -- if there is a good way to store function pointers in a static array (I later on plan to iterate ove the set of function pointers and get the value of each data member into a stream ) thanks in advance, Vlad -- V S P toreason(a)fastmail.fm -- http://www.fastmail.fm - The professional email service

1 0

[serialization][build][spirit] Problems with spirit when building 1.38 serialization on windows with Intel C++ 11.0
by Jesse Perla 20 Feb '09

20 Feb '09

I am on windows using the latest version of the intel compiler (set in my user-config.jam). I was able to do a debug build of everything using: bjam link=static debug stage But when I tried to do a release build, it is popping up a spirit error for serialization. (fails with or without mpi) Any ideas on a workaround? See below: C:\apps\boost\boost_1_38_0>bjam --with-mpi link=static release stage ...patience... ...found 1080 targets... ...updating 4 targets... compile-c-c++ bin.v2\libs\serialization\build\intel-win\release\link-static\thre ading-multi\xml_grammar.obj icl: command line remark #10010: option '/Op' is deprecated and will be removed in a future release. See '/help deprecated' xml_grammar.cpp Warning: This header is deprecated. Please use: boost/spirit/include/classic_rul e.hpp Warning: This header is deprecated. Please use: boost/spirit/include/classic_ope rators.hpp Warning: This header is deprecated. Please use: boost/spirit/include/classic_act ions.hpp Warning: This header is deprecated. Please use: boost/spirit/include/classic_num erics.hpp C:\apps\boost\boost_1_38_0\boost\spirit\home\classic\utility\impl\chset\range_ru n.ipp(142) (col. 21): internal error: 0_1504 compilation aborted for libs\serialization\src\xml_grammar.cpp (code 4) call "C:\apps\Intel\Compiler\11.0\066\cpp\bin\ia32\iclvars_ia32.bat" > nul icl @"bin.v2\libs\serialization\build\intel-win\release\link-static\threading-mu lti\xml_grammar.obj.rsp" ...failed compile-c-c++ bin.v2\libs\serialization\build\intel-win\release\link-s tatic\threading-multi\xml_grammar.obj... ...skipped <pbin.v2\libs\serialization\build\intel-win\release\link-static\threa ding-multi>libboost_serialization-iw-mt-1_38.lib for lack of <pbin.v2\libs\seria lization\build\intel-win\release\link-static\threading-multi>xml_grammar.obj... ...skipped <pstage\lib>libboost_serialization-iw-mt-1_38.lib for lack of <pbin.v 2\libs\serialization\build\intel-win\release\link-static\threading-multi>libboos t_serialization-iw-mt-1_38.lib... ...skipped <pstage\lib>libboost_serialization-iw-mt.lib for lack of <pstage\lib> libboost_serialization-iw-mt-1_38.lib... ...failed updating 1 target... ...skipped 3 targets...

1 0

[Test] Compilation inconsistency
by Greg Ravikovich 19 Feb '09

19 Feb '09

Hello, I posted a question regarding the following issue a few months back: Empty cpp file containing only: #include <boost/test/test_case_template.hpp> Fails to compile, however an empty cpp file containing #include <boost/test/unit_test.hpp> #include <boost/test/test_case_template.hpp> Compiles fine. The issue is still there in boost 1.38. Can someone address it please? Thanks. Greg

3 2

[Proto] Failure on VC9
by Joel Falcou 19 Feb '09

19 Feb '09

I've been tryign to port some of my proto code to VC9 but it seems I failed. The code compiles fine but when I try to use any of my newly defined operator on my terminal, VC9 complains it cna't find them. What can be the reason of this symptoms as the same code works on both gcc and icc ?

2 3

if_then_else_return (boost::lambda) question
by Jonathan Leonard 19 Feb '09

19 Feb '09

Greetings! I'm fairly new to boost and have a little problem with if_then_else_return. Can anyone tell me the problem with the following short program? [It seems that my 'thenBranch' and 'elseBranch' functions are never called and if_then_else_return' always returns true (as if the 'then' branch were called)]. Many thanks in advance for any help you may offer! P.S. (A more general C++ q: A little improvement I'd like to make is to drop the ITERATORTYPE type argument from IfNotEndThenElseReturn and use CONTAINERTYPE::iterator instead, but the compiler won't allow that. Anyone know why?). Thanks again! ============================================ #include <boost/lambda/lambda.hpp> #include <boost/lambda/if.hpp> #include <boost/lambda/bind.hpp> #include <boost/function.hpp> #include <functional> #include <list> using namespace boost::lambda; template<typename RET> RET InvokeReturn(boost::function<void ()> f, const RET& retVal) { printf("in InvokeReturn!\n"); f(); return retVal; } bool InvokeReturnTrue(boost::function<void ()> f) { return InvokeReturn(f, true); } bool InvokeReturnFalse(boost::function<void ()> f) { return InvokeReturn(f, false); } boost::function<bool ()> MakeReturnTrue(boost::function<void ()> f) { printf("in MakeReturnTrue!\n"); return boost::lambda::bind(&InvokeReturnTrue, f); } boost::function<bool ()> MakeReturnFalse(boost::function<void ()> f) { printf("in MakeReturnFalse!\n"); return boost::lambda::bind(&InvokeReturnFalse, f); } template<typename CONTAINERTYPE, typename ITERATORTYPE> bool IfNotEndThenElseReturn(const ITERATORTYPE& it, CONTAINERTYPE& container, boost::function<void ()> thenBranch, boost::function<void ()> elseBranch) { printf("in IfNotEndThenElseReturn!\n"); std::not_equal_to<ITERATORTYPE> ne; return boost::lambda::if_then_else_return( boost::lambda::bind(&std::not_equal_to<ITERATORTYPE>::operator(), ne, it, container.end()), MakeReturnTrue(thenBranch), MakeReturnFalse(elseBranch))(); } void thenBranch() { printf("in then branch!\n"); } void elseBranch() { printf("in else branch!\n"); } int _tmain(int argc, _TCHAR* argv[]) { printf("beginning execution!\n"); std::list<int> l; l.push_back(8); l.push_back(7); l.push_back(6); l.push_back(5); std::list<int>::iterator it = l.end(); std::list<int>::iterator it2 = l.begin(); bool b = IfNotEndThenElseReturn(it, l, bind(&thenBranch), bind(&elseBranch)); printf("b is %s\n", b ? "true" : "false"); bool b = IfNotEndThenElseReturn(it2, l, bind(&thenBranch), bind(&elseBranch)); printf("b is %s\n", b ? "true" : "false"); printf("finished!\n"); return 0; }

2 2

Order of composite units in the output
by Esben Mose Hansen 19 Feb '09

19 Feb '09

Hi, I think I am beginning to get the boost:units library, but I still have one problem. Consider: #include <iostream> #include <boost/units/io.hpp> #include <boost/units/systems/si/length.hpp> #include <boost/units/make_system.hpp> #include <boost/units/quantity.hpp> #include <boost/units/base_dimension.hpp> #include <boost/units/base_unit.hpp> #include <boost/units/derived_dimension.hpp> // Introduce slot (measured in e.g. ffe or tfe) as a basic dimension, unit and in a system struct slot_base_dimension : boost::units::base_dimension<slot_base_dimension, 1> { }; typedef slot_base_dimension::dimension_type slot_dimension; struct ffe_base_unit : public boost::units::base_unit<ffe_base_unit, slot_dimension, 1> { static std::string name() { return("40ft equivalent"); } static std::string symbol() { return("ffe"); } }; typedef ffe_base_unit::unit_type slot_t; BOOST_UNITS_STATIC_CONSTANT(ffe, slot_t); /// derived dimension (slot) flow typedef boost::units::derived_dimension<slot_base_dimension,1,boost::units::time_base_dimension,-1>::type flow_dimension; /// placeholder class defining si unit system typedef boost::units::make_system<boost::units::si::meter_base_unit, boost::units::si::second_base_unit, ffe_base_unit>::type unit_system; typedef boost::units::unit<flow_dimension,unit_system> flow_t; BOOST_UNITS_STATIC_CONSTANT(ffe_per_second, flow_t); int main() { boost::units::quantity<flow_t> flow(4*ffe_per_second); std::cout << "flow= " << flow << std::endl; } This program outputs flow= 4 s^-1 ffe which is correct, but a bit unusual... 4 ffe/s or even 4 ffe s^-1 would have been better. Is this possible? Thanks! -- kind regards, Esben

3 2

[serialization|TR1] serialization impementation for std::tr1::tuple?
by Markus Werle 19 Feb '09

19 Feb '09

Hi! Sorry, if this has been discussed before. I am too stupid (and tired) to find the equivalent for tuple_io.hpp for the boost tr1::tuple implementation. So is there already some serialization and/or I/O implementation for a tr1::tuple available somewhere? The proposal at http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2002/n1403.pdf mentions operator<<, but VC9 SP1 refuses to find one if I #include <boost/tr1/tuple.hpp> or <tuple>. Is this another VC9 bug? I could not find the compiler flag which hinders M$'s implementation of TR1 ... Any pointer appreciated. best regards, Markus

2 1

[optional] operator <<
by Igor R 19 Feb '09

19 Feb '09

Hello, I've just noticed that the following code might be missleading: boost::optional<int> opt; opt = 0; std::cout << opt; If boost/optional/optional_io.hpp is not included, the above code outputs "1" - because boost::optional is implicitly convertible to bool. Maybe it's worth either including boost/optional/optional_io.hpp in boost/optional.hpp or making some trick to generate compiler error if operator <<() used and optional_io.hpp is not included?

2 2

[Python] can't import boost modules in an embedded interpreter
by George Slavov 19 Feb '09

19 Feb '09

I am trying to create an executable which will start a python interpreter, import a module written in python, called entry.py, run its run() method and then quit. I have this part done, although I use only the C API. Now I define a boost python module like so struct World { void set(std::string msg) { this->msg = msg; } std::string greet() { return msg; } std::string msg; }; BOOST_PYTHON_MODULE(hello) { class_<World>("World") .def("greet", &World::greet) .def("set", &World::set) ; } In my entry.py I have the simple code import hello def run(): planet = hello.World() planet.set("howdy") print planet.greet() When I run my executable, it tells me that World is undefined in the entry module. However, if in C++ I use code like this PyRun_SimpleString("import hello"); PyRun_SimpleString("planet = hello.World()"); PyRun_SimpleString("planet.set('howdy')"); PyRun_SimpleString("print planet.greet()"); It works! What am I doing wrong? I am attaching my (very short) code. Be advised that I am using the funky platform of Windows XP, cygwin and GCC 3.4, boost 1.33.1. That might be the problem, but I have no other platform to try this on easily. What am I missing here? Best, George

1 0

wchar_t and boost::asio
by Cain Miley 19 Feb '09

19 Feb '09

Okay, i sorted out that you can't use boost::asio::basic_streambuf with wide chars, but if not that, how do I send wide chars with asio? Is it even possible? Thanks.

3 2