Re: [boost] Re: compile time member detection

From: Daniel Wallin (dalwan01_at_[hidden])
Date: 2003-11-14 15:06:59

David Abrahams wrote:
> Daniel Wallin <dalwan01_at_[hidden]> writes:
> 
> 
>>Brian McNamara wrote:
>>
>>>There are a number of recent messages on this thread, and I've chosen to
>>>reply "in batch".
>>>Daniel Wallin:
>>>
>>>
>>>>Why complicate things? Just make it a generic function that uses copy
>>>>construction by default, and rely on ADL and overloading instead of
>>>>class specialization.
>>>
>>>This strategy uses structural conformance.  I want named
>>>conformance.
>>>(Named conformance is not as susceptible to accidental conformance.)
>>
>>What about tagging the clone function?
>>
>>   struct clone_factory_tag {};
>>
>>   template<class T>
>>   T* clone(clone_factory_tag, const T& x)
>>   {
>>       return new T(x);
>>   }
>>
>>   ...
>>
>>   Foo* clone(clone_factory_tag, const Foo& x)
>>   {
>>       return x.clone();
>>   }
> 
> 
> I had this idea some time ago, but Peter Dimov pointed out that it
> doesn't really prevent accidental ADL problems:
> 
> 
>       namespace my {
>       
>          class X {};
>          template <class T>
>          garbage clone(T, X& x)
>          {
>              // not the same meaning for "clone"
>          }
>       }
> 
>       lib::some_generic_algorithm(my::X());

I might be mistaken, but that shouldn't compile should it? This should
fail due to ambiguity, which isn't really an ADL problem in this case
since my::clone() wasn't meant to be found.

-- 
Daniel Wallin