Thank you Steven, 2009/11/2 Steven Watanabe

AMDG

...

<snip>

You can define two overloads of operator() and use enable_if to distinguish them.

I thought about this solution but did not know how to implement it. Could you please provide a small sample? My current solution is to build mpl::vector from sequences and boost::preprocessor. Then, with BOOST_PP_SEQ_FOR_EACH I can automagically generate overloads for all types. // image_types.hpp #define RGB_IMAGE_TYPES (boost::gil::rgb8_image_t)(boost::gil::rgb16_image_t)(boost::gil::rgb32_image_t) typedef boost::mpl::vector< BOOST_PP_SEQ_ENUM( RGB_IMAGE_TYPES ) > rgb_image_types; // ... // min_max_functor.hpp #define OVERLOAD_MIN_MAX_PARENTHESIS_OPERATOR( r , n , data ) template <> \ any_view_min_max::result_type any_view_min_max::operator()data::view_t(const data::view_t& v) const \ { \ // ... } BOOST_PP_SEQ_FOR_EACH( OVERLOAD_MIN_MAX_PARENTHESIS_OPERATOR , ~ , RGB_IMAGE_TYPES ) Regards, Olivier

In Christ, Steven Watanabe

_______________________________________________ Boost-users mailing list Boost-users@lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

Thank you Steven for your reply. It works great. Here is what I finally did: // struct histogram_functor typename boost::enable_if< boost::mpl::contains< boost::mpl::transformboost::mpl::_1

::type, ViewType>, result_type>::type operator()(const ViewType& v) const { // ... }

template<class ViewType> typename boost::enable_if< boost::mpl::or_< boost::mpl::contains< boost::mpl::transform< rgb_image_types, add_view_type >::type, ViewType>, boost::mpl::contains< boost::mpl::transform< rgba_image_types, add_view_type >::type, ViewType> >, result_type>::type operator()(const ViewType& v) const { // ... } // ... template <typename ImageType> struct add_view_type { typedef typename ImageType::view_t type; }; However, I have one other need: type reduction. For instance, all rgb image types have the same overload operator(). I do not know if the solution I came up instantiate the operator for all types, but if it is the case, it is a valid but not a good solution. What I need is to have the same operator() for all rgb types. I have read in a paper from L .Bourdev ( www.lubomir.org/academic/Minimizing*CodeBloat*.pdf) that I could use type reduction, but I need some help. Can you help ? Here is the definition of rgb_image_types: #define RGB_IMAGE_TYPES (boost::gil::rgb8_image_t)(boost::gil::rgb16_image_t)(boost::gil::rgb32_image_t) typedef boost::mpl::vector< BOOST_PP_SEQ_ENUM( RGB_IMAGE_TYPES ) > rgb_image_types; Best regards, Olivier