Re: [boost] [MPL] Making Generators

From: David A. Greene (greened_at_[hidden])
Date: 2002-12-05 13:27:38

David Abrahams wrote:

>>>    template <class T, class U, class V>
>>>    struct my_type_generator
>>>    {
>>>        typedef my_type<T,U,V> type;
>>>    };
>>>lambda<my_type_generator> does it
>>
> 
> Oops, I meant
> 
>    lambda<my_type_generator<Bound,mpl::_1,mpl::_2> > 
> 
> of course!

Ok, that makes more sense now.  :)

>>, unless of course your compiler
>>
>>>needs BOOST_MPL_AUX_LAMDA_SUPPORT.  I don't think it's much of a
>>>savings, though.
>>
>>Not for one class, no, but when we're talking several classes
>>with several binding requirements, I think there's a significant
>>savings to be had.
> 
> So do you feel you need an additional library feature?

That's what I'm trying to find out.  It seems like most of the
stuff is there already in MPL placeholders and binders.

>>Plus your solution here doesn't bind T to a type.  :)
> 
> Are you just pointing out my error?

Well...yeah.  :)  It wasn't meant as an attack.  I honestly
was confused about what you presented.  Your correction above
makes everything clear to me now.

>>g++ 3.2.  The MPL paper and docs don't say anything about
>>using placeholders or binders with classes that aren't
>>metafunctions.  How would the binders know what to typedef
>>as apply::type?
> 
> I don't know... well, it could detect whether there was a ::type
> member, and if it were not present, it could just give you the outer
> class. I think that's a bit of a hack, though.

Agreed.  Urk...I'm not sure how to get around this problem without
requiring template template parameters (beyond what's used for
placeholders currently).

                            -Dave

-- 
"Some little people have music in them, but Fats, he was all music,
  and you know how big he was."  --  James P. Johnson