core: allows template specialization of EigenToPy

include/eigenpy/eigen-to-python.hpp

@@ -50,7 +50,7 @@ namespace eigenpy
 {
   namespace bp = boost::python;
 
-  template<typename MatType>
+  template<typename MatType, typename _Scalar>
   struct EigenToPy
   {
     static PyObject* convert(typename boost::add_reference<typename boost::add_const<MatType>::type>::type mat)
@@ -82,8 +82,8 @@ namespace eigenpy
     }
   };
 
-  template<typename MatType, int Options, typename Stride>
-  struct EigenToPy< Eigen::Ref<MatType,Options,Stride> >
+  template<typename MatType, int Options, typename Stride, typename _Scalar>
+  struct EigenToPy< Eigen::Ref<MatType,Options,Stride>,_Scalar >
   {
     static PyObject* convert(const Eigen::Ref<MatType,Options,Stride> & mat)
     {

include/eigenpy/fwd.hpp

@@ -28,8 +28,8 @@
 
 namespace eigenpy
 {
-  template<typename MatType> struct EigenToPy;
   template<typename MatType> struct EigenFromPy;
+  template<typename MatType, typename Scalar = typename boost::remove_reference<MatType>::type::Scalar> struct EigenToPy;
 }
 
 #endif // ifndef __eigenpy_fwd_hpp__