[boost] Re: Canonical form for Type Traits specializations?

From: David Abrahams (dave_at_[hidden])
Date: 2003-11-28 00:48:14

Eric Friedman <ebf_at_[hidden]> writes:

> David Abrahams wrote:
>> Eric Friedman <ebf_at_[hidden]> writes:
>> 
>>>It seems to me the standard form of such a specialization would look
>>>as follows:
>>>
>>>// ...at file scope...
>>>namespace boost {
>>>
>>>   template <>
>>>   struct has_nothrow_copy< myUDT >
>>>     : mpl::true_
>>>   {
>>>   };
>>>
>>>}
>>>
>>>The problem I see here is this approach (needlessly) forces MPL upon
>>>the user.
>>
>> It's not needless.  You're going to force *something* upon the user.
>> Why not boost/mpl/bool.hpp?
>> 
>
> But why? The type traits library should have nothing to do with MPL
> (at least from a user's perspective).

Why?

>>> 1) Provide a typedef of boost::mpl::true_ named boost::true_, or
>>>    perhaps boost::tt_true to keep it Type Traits-specific. (I suppose
>>>    we could do the same for false_.)
>> What problem is this solving?  Just that they don't have to write
>> the
>> letters "mpl"?
>
> Yes. I believe it is confusing for a user who is seeking simply to
> specialize a type trait to have to think about an unrelated library.

It's clearly not completely unrelated.

> I think maybe we should have boost::integral_constant (a la TR1)
> instead of boost::mpl::integral_c, boost::mpl::bool_, etc.

Why?

-- 
Dave Abrahams
Boost Consulting
www.boost-consulting.com