RE: [boost] Re: Re: Using MPL on MSVC7

From: Aleksey Gurtovoy (agurtovoy_at_[hidden])
Date: 2003-01-04 04:08:22

Jaap Suter wrote:
> > Nifty!  You're using one of my favorite metaprogramming tricks, the
> > default template parameter which allows you to avoid creating a
> > separate implementation template.  I almost forgot about that one,
> > it's been so long since I've been able to use it.
> 
> Funny you mention that. I've been meaning to ask for its 
> appropriateness on this list. After all, we are exposing 
> implementation details in the interface which is not something 
> I would recommend on a daily basis. 

At least one has to be careful :). There are at least two situations when
default template parameters affect the observable library behavior:

1) adding default _non-type_ template parameters "disqualifies" a
metafunction from participating in lambda expressions:

    template<
          typename N
        , int one_before_previous_ = 0
        , int previous_ = 1
        >
    struct fibonacci
    {
        // ...
    };

    typedef fold<
          range_c<int,0,10>
        , int_c<0>
        , plus< _1, fibonacci<_2> >
        >::type sum; // compilation error here!

    BOOST_STATIC_ASSERT(sum::value == 88);

2) adding a default template parameter to a metafunction which number of
parameters is already equal to BOOST_MPL_METAFUNCTION_MAX_ARITY results in
the same negative effect:

    // BOOST_MPL_METAFUNCTION_MAX_ARITY == 5
    template<
          typename N1
        , typename N2
        , typename N3 = integral_c<int,0>
        , typename N4 = integral_c<int,0>
        , typename N5 = integral_c<int,0>
        , typename result_ = integral_c<int,0>
        >
    struct my_fun
    {
        // ...
    };

    typedef fold<
          range_c<int,0,10>
        , int_c<0>
        , my_fun<_1,_2>
        >::type sum; // error!

Otherwise, it's more or less safe :).

Aleksey