From: Aleksey Gurtovoy (alexy_at_[hidden])
Date: 2001-12-17 15:42:27
Emily Winch wrote:
> for_each for tuples makes sense. for_each for typelists doesn't.
Actually it does :), although I am not talking about current (deprecated)
version of 'for_each' in boost::mpl - as I already said, it should be called
'accumulate'. For example, consider a compile-time computation with results
that need to be passed to a run-time part of the program:
// fixed-point algorithm input
typedef mpl::type_vector<
mpl::fixed_t<-1,2345678>
, mpl::fixed_t<9,0001>
// ..
, mpl::fixed_t<3,14159>
> input_data;
/*
complex compile-time algorithm skipped :)
*/
typedef /*...*/ result_data;
namespace aux {
struct push_back
{
template<typename T> struct apply
{
template<typename C> void operator()(C& c)
{
// in our case T() == fixed_t() == fixed_t().operator()()
c.push_back(T());
}
};
};
}
// passing the results to the run-time part of the program
std::vector<double> results;
typedef mpl::for_each<result_data, aux::push_back> copy_results;
copy_results::execute(results); // here! :)
// using results in run-time world
std::copy(
results.begin()
, results.end()
, std::ostream_iterator<double>(std::cout, "\n"));
);
There are other ways to generate an analogous code, but IMO they are much
less expressive.
Aleksey