Refactoring Enum_rel using domains on class representatives only

This PR refactors the `Enum` relations in order to use a proper type
for the domains of enum semantic values.

Now, we only store domains for class representatives and I believe that the new
implementation is simpler to understand and to maintain.

I don't use the functor `Domains_make` of `Rel_utils` as domain's
propagations performed in `Enum_rel` are much simpler than the propagations in
the BV theory.

Author: Halbaroth

src/lib/reasoners/bitv_rel.ml

@@ -77,7 +77,7 @@ let is_bv_ty = function
 
 let is_bv_r r = is_bv_ty @@ X.type_info r
 
-module Domain : Rel_utils.Domain with type t = Bitlist.t = struct
+module Domain = struct
   (* Note: these functions are not in [Bitlist] proper in order to avoid a
      (direct) dependency from [Bitlist] to the [Shostak] module. *)
 
@@ -124,7 +124,8 @@ module Domain : Rel_utils.Domain with type t = Bitlist.t = struct
         | Ext (r, _r_size, i, j) -> extract (map_signed f r acc) i j
       ) empty (Shostak.Bitv.embed r)
 
-  let unknown = function
+  let unknown r =
+    match X.type_info r with
     | Ty.Tbitv n -> unknown n Ex.empty
     | _ ->
       (* Only bit-vector values can have bitlist domains. *)