rename the deprecated functions.

docs/user_guide.org

@@ -332,7 +332,7 @@ This is the easiest method, but cannot be parallelized efficiently (i.e., with a
 The idea here is to traverse the internal structure of the tree. We start to browse the groups of leaves, then for each group the leaves and finally for each leaf the particles. In this approach, we don't know a priori the storage of the different objects.
 #+begin_src c++
 // loop on the group of the tree
-    for(auto const pg: tree.group_of_leaves())
+    for(auto const pg: tree.vector_of_leaf_groups())
     {
         // Loop on the leaf of thz group
         for(auto const& leaf: pg->components())
@@ -352,7 +352,7 @@ This method is slightly more efficient than the classical method and can be para
 We will use our knowledge of particle storage to be more efficient. In the tree version with views, the particles of a group of leaves are stored continuously in the tree. Moreover, we need to know the type of the particles to get the right pointers (this can be automated) to the different components of the particles (position, inputs, outputs, variables).
 #+begin_src c++
 // loop on the group of the tree
-    for(auto pg: tree.group_of_leaves())
+    for(auto pg: tree.group_of_levector_of_leaf_groupsaves())
     {
         //  std::cout << pg->storage() << "\n";
         auto nb_particles = pg->storage().size();
@@ -367,7 +367,10 @@ We will use our knowledge of particle storage to be more efficient. In the tree
         auto ptr_fx = ptr_potential + nb_particles;
         auto ptr_fy = ptr_fx + nb_particles;
         // variables
-        auto ptr_var0 = reinterpret_cast<std::size_t*>(ptr_fz + nb_particles);
+        // tuple_ptr_variables is a tuple of pointer pointing on the first elemnt of each variables
+        auto tuple_ptr_variables = pg->storage().ptr_on_variables();
+        // To access the pointer on the first variable
+        const auto var0 = std::get<0>(variable_storage);
 
         for(auto i = 0; i < nb_particles; ++i)
         {