Jaakko Jarvi wrote:
std::for_each(M.begin(), M.end(), std::cout << l::bind(select2nd(), l::_1) << " ");
Well, that's too complicated ;-) _binding_ to a function or functor is the approach I wanted to avoid. I think that select1st and select2nd (or - to be generic - select<n>) could be directly supported by boost::lambda. It's so common to have a list/vector of tuples or a std::map (yes, its a request for a feature ;-)) I tried to write something taking ll_static_cast as an example, but failed to quickly understand the requirements and semantics of lambda_functor* Could You correct/complete the code below? I could not figure out how select2nd_action::apply gets its RET defined :-( Is there some documentation on the functor conventions or on how to extend LL? class select2nd_action { public: template<class RET, class Arg1> static RET apply(Arg1 &a1) { return a1.second; } }; template <class Arg1> inline const lambda_functor< lambda_functor_base< action<1, select2nd_action >, tuple<typename const_copy_argument <const Arg1>::type>
ll_select2nd(const Arg1& a1) { return lambda_functor_base< action<1, select2nd_action >, tuple<typename const_copy_argument <const Arg1>::type> > ( tuple<typename const_copy_argument <const Arg1>::type>(a1)); } best regards, Markus -- Build your own Expression Template Library with Daixtrose! Visit http://daixtrose.sourceforge.net/