[boost] Re: is_base_and_derived

From: David Abrahams (dave_at_[hidden])
Date: 2003-10-27 14:08:45

Rene Rivera <grafik.list_at_[hidden]> writes:

> [2003-10-27] Brian McNamara wrote:
>
>>On Mon, Oct 27, 2003 at 07:33:16AM -0700, Dave Gomboc wrote:
>>> > Actually I'd prefer
>>> > 
>>> >    is_proper_subtype< D, B >
>>> >    is_subtype< D, B >      
>>> > 
>>> > but you have to be careful that you reverse the argument 
>>> > order when changing your old code over from is_base_and_derived.
>>> 
>>> Many different forms specifying one but not the other (ancestor,
>>> descendent, superclass, subclass, etc.) were rejected due to the ease
>>> with which a developer could nonetheless misinterpret which argument is
>>> supposed to go first.  (You could trawl the archives for this.)
>>
>>I can definitely imagine this.  I was thinking the same thing and was
>>going to cite the 'argument order clarity' as one of the advantages to
>>is_base_and_derived, until I had the epiphany that
>>
>>   is_subtype< D, B >      
>>
>>is just how C++ forces us to spell
>>
>>   D::is_subtype<B>
>>
>>non-intrusively.  
>>
>>
>>Oh, which means I actually prefer the name "is_subtype_of".
>>
>>
>>It's totally analogous to the non-meta case, e.g.
>>
>>   contains( aRegion, aPoint )
>>
>>means
>>
>>   aRegion.contains( aPoint )
>
> Doesn't that then argue for something like:
>
>     is_<D>::subtype_of<B>
>
> ??

only if you don't mind:

     typename is_<D>::template subtype_of<B>::type
                      ^^^^^^^^

Yikes.  How about 

    infix<D,is_subtype_of,B>::type

?? 

I kinda like that one. Probably better names than "infix" are possible.

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