boost/proto/transform/pass_through.hpp
///////////////////////////////////////////////////////////////////////////////
/// \file pass_through.hpp
///
/// Definition of the pass_through transform, which is the default transform
/// of all of the expression generator metafunctions such as unary_plus<>, plus<>
/// and nary_expr<>.
//
// Copyright 2008 Eric Niebler. Distributed under the Boost
// Software License, Version 1.0. (See accompanying file
// LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)
#ifndef BOOST_PROTO_TRANSFORM_PASS_THROUGH_HPP_EAN_12_26_2006
#define BOOST_PROTO_TRANSFORM_PASS_THROUGH_HPP_EAN_12_26_2006
#include <boost/preprocessor/cat.hpp>
#include <boost/preprocessor/repetition/enum.hpp>
#include <boost/preprocessor/iteration/iterate.hpp>
#include <boost/mpl/bool.hpp>
#include <boost/type_traits/remove_reference.hpp>
#include <boost/proto/proto_fwd.hpp>
#include <boost/proto/args.hpp>
#include <boost/proto/transform/impl.hpp>
#include <boost/proto/detail/ignore_unused.hpp>
namespace boost { namespace proto
{
namespace detail
{
template<
typename Grammar
, typename Expr
, typename State
, typename Data
, long Arity = arity_of<Expr>::value
>
struct pass_through_impl
{};
#include <boost/proto/transform/detail/pass_through_impl.hpp>
template<typename Grammar, typename Expr, typename State, typename Data>
struct pass_through_impl<Grammar, Expr, State, Data, 0>
: transform_impl<Expr, State, Data>
{
typedef Expr result_type;
/// \param e An expression
/// \return \c e
/// \throw nothrow
#ifdef BOOST_PROTO_STRICT_RESULT_OF
result_type
#else
typename pass_through_impl::expr_param
#endif
operator()(
typename pass_through_impl::expr_param e
, typename pass_through_impl::state_param
, typename pass_through_impl::data_param
) const
{
return e;
}
};
} // namespace detail
/// \brief A PrimitiveTransform that transforms the child expressions
/// of an expression node according to the corresponding children of
/// a Grammar.
///
/// Given a Grammar such as <tt>plus\<T0, T1\></tt>, an expression type
/// that matches the grammar such as <tt>plus\<E0, E1\>::type</tt>, a
/// state \c S and a data \c V, the result of applying the
/// <tt>pass_through\<plus\<T0, T1\> \></tt> transform is:
///
/// \code
/// plus<
/// T0::result<T0(E0, S, V)>::type
/// , T1::result<T1(E1, S, V)>::type
/// >::type
/// \endcode
///
/// The above demonstrates how child transforms and child expressions
/// are applied pairwise, and how the results are reassembled into a new
/// expression node with the same tag type as the original.
///
/// The explicit use of <tt>pass_through\<\></tt> is not usually needed,
/// since the expression generator metafunctions such as
/// <tt>plus\<\></tt> have <tt>pass_through\<\></tt> as their default
/// transform. So, for instance, these are equivalent:
///
/// \code
/// // Within a grammar definition, these are equivalent:
/// when< plus<X, Y>, pass_through< plus<X, Y> > >
/// when< plus<X, Y>, plus<X, Y> >
/// when< plus<X, Y> > // because of when<class X, class Y=X>
/// plus<X, Y> // because plus<> is both a
/// // grammar and a transform
/// \endcode
///
/// For example, consider the following transform that promotes all
/// \c float terminals in an expression to \c double.
///
/// \code
/// // This transform finds all float terminals in an expression and promotes
/// // them to doubles.
/// struct Promote
/// : or_<
/// when<terminal<float>, terminal<double>::type(_value) >
/// // terminal<>'s default transform is a no-op:
/// , terminal<_>
/// // nary_expr<> has a pass_through<> transform:
/// , nary_expr<_, vararg<Promote> >
/// >
/// {};
/// \endcode
template<typename Grammar>
struct pass_through
: transform<pass_through<Grammar> >
{
template<typename Expr, typename State, typename Data>
struct impl
: detail::pass_through_impl<Grammar, Expr, State, Data>
{};
};
/// INTERNAL ONLY
///
template<typename Grammar>
struct is_callable<pass_through<Grammar> >
: mpl::true_
{};
}} // namespace boost::proto
#endif