"Daniel James" <daniel@calamity.org.uk> wrote
Still, his implementation is fantastic, considering that it runs on Visual C++ 6. Since my last post, I've made quite a lot of progress at getting my variation to cope with complicated expressions with less calls to 'foo'. But Peder's version takes about half the time to run my tests, and will scale a lot better.
Besides... I've just made a little improvement to my implementation: I added another function to determine the size of encoded expression: template<class T> char(&size(const T&))[ mpl::size<typename encode_type<mpl::vector0<>, T>::type>::value ]; Then I modified the nth item extraction to look like this: #define BOOST_TYPEOF_TYPEITEM(z, n, expr)\ boost::mpl::int_<sizeof(boost::type_of::foo<(n < sizeof(boost::type_of::size(expr)))?n:sizeof(boost::type_of::size(expr))>(ex pr))> Note that now no new templates gets instantiated for n larger than size. I tried it at work, and this improvement cuts almost in half the compile time of my example (with Spirit, Lambda, etc.), presumably having taken care of smaller types. I'll post the modified system later, when I am done with integral template parameters. Regards, Arkadiy