On Sat, Feb 23, 2013 at 3:56 PM, Jeffrey Lee Hellrung, Jr. < jeffrey.hellrung@gmail.com> wrote:

On Sat, Feb 23, 2013 at 2:32 AM, Niitsuma Hirotaka < hirotaka.niitsuma@gmail.com> wrote:

...

I would like to use mpl::lambda in nest.

In this case works fine. --------- typedef boost::mpl::lambda<boost::mpl::max<boost::mpl::_1,boost::mpl::_2>

...

my_max; typedef boost::mpl::fold< boost::mpl::vector_c<int,1,7,0,-2,5,-1> , boost::mpl::int_<6> //, boost::mpl::max<boost::mpl::_1,boost::mpl::_2> , my_max >::type r2;

BOOST_MPL_ASSERT(( boost::is_same< r2, boost::mpl::integral_c<int, 7> > )); ---------

However

---------

typedef boost::mpl::lambda< boost::mpl::fold< boost::mpl::_1 ,boost::mpl::front<boost::mpl::_1> ,my_max >::type >::type my_max_types;

Also: remove the ::type from the fold<> expression. Works for me after that (and the fix with adding the ::type on the definition of my_max). I would think you would've gotten a compiler error at this point but, meh. typedef my_max_types::apply<boost::mpl::vector_c<int,1,7,0,-2,5,-1>

...

...

::type r3;

BOOST_MPL_ASSERT(( boost::is_same< r3, boost::mpl::integral_c<int, 7> > )); ///error => becomes boost::mpl::integral_c<int, 1> ---------

How to define such nest lambda in mpl?

Uh, I see a red flag in that you don't grab the nested type typedef from your mpl::lambda invocations, i.e., I think you should have

typedef mpl::lambda< mpl::max< mpl::_1, mpl::_2 > >::type my_max;

Incidentally, you could also do

typedef mpl::quote2< mpl::max > my_max

(Well, assuming mpl::max doesn't have some hidden default parameters...)

Without grabbing the type typedef in your mpl::lambda invocation, your use of mpl::_1 and mpl::_2 are "exposed" to premature substitution.

- Jeff

On Sat, Feb 23, 2013 at 9:38 PM, Niitsuma Hirotaka < hirotaka.niitsuma@gmail.com> wrote:

...

Also: remove the ::type from the fold<> expression. Works for me after that (and the fix with adding the ::type on the definition of my_max). I would think you would've gotten a compiler error at this point but, meh.

Thank you for your advice. However with gcc 4.6.3, that cause compiler error.

Sorry, I think I inadvertently fixed a 3rd error in your original post (accessing ::type on the mpl::max invocation) when I copied your code. The following works for me on MSVC9: -------- #include <boost/mpl/assert.hpp> #include <boost/mpl/fold.hpp> #include <boost/mpl/front.hpp> #include <boost/mpl/integral_c.hpp> #include <boost/mpl/lambda.hpp> #include <boost/mpl/max.hpp> #include <boost/mpl/placeholders.hpp> #include <boost/mpl/vector_c.hpp> #include <boost/type_traits/is_same.hpp> namespace mpl = boost::mpl; typedef mpl::lambda< mpl::max< mpl::_1, mpl::_2 > >::type my_max; typedef mpl::fold< mpl::vector_c<int,1,7,0,-2,5,-1> , mpl::int_<6> , my_max

::type r2; BOOST_MPL_ASSERT(( boost::is_same< r2, mpl::integral_c<int, 7> > ));

typedef mpl::lambda< mpl::fold< mpl::_1 , mpl::front< mpl::_1 > , my_max

::type my_max_types;

typedef my_max_types::apply< mpl::vector_c<int,1,7,0,-2,5,-1> >::type r3; BOOST_MPL_ASSERT(( boost::is_same< r3, mpl::integral_c<int, 7> > )); -------- If you can't figure out the reasons your original posting didn't work and you want to know, holler back. - Jeff