[boost] Re: is_convertible: rationale and wording

From: Gennaro Prota (gennaro_prota_at_[hidden])
Date: 2003-01-30 04:12:31

On Tue, 28 Jan 2003 16:21:07 +0100, Gennaro Prota
<gennaro_prota_at_[hidden]> wrote:

>On Tue, 28 Jan 2003 12:58:33 -0000, "John Maddock"
><jm_at_[hidden]> wrote:
>
>>> is there, among the uses of is_convertible that you have listed, any
>>> usage where no expression could be used (so that you would need to see
>>> if the "type" is convertible)?
>>
>>Yes - named template parameters form the main example of that, concept
>>checking another (although in that case you could probably get hold of a
>>variable if you really wanted to).
>
>Thanks. Can you provide a minimal example of the technique used to
>implement named template arguments?

I see that you haven't replied to this for long time now. So either
you are bored from the question, or it wasn't clear enough. To see if
it is the second case I thought to reformulate it:

Can you show, with an example, why the code used for named template
arguments can't reasonably use an expression?

Genny.