RE: [boost] in search of a technique

From: Gennadiy Rozental (rogeeff_at_[hidden])
Date: 2002-03-22 02:05:17

You prevent using my<int> ealer in translation unit either.

Gennadiy.

--- Aleksey Gurtovoy <agurtovoy_at_[hidden]> wrote:
>David Abrahams wrote:
>> I'm looking for a legal way to cause certain constructs to fail to
>> compile based on what templates have been instantiated at 
>> that point in the program. The catch is that I also want it to be 
>> legal to compile the same constructs in other translation units 
>> (and, if I have to settle for something permissive, earlier in the 
>> same translation unit) where the preventing templates have not 
>> been instantiated. Most approaches to this problem result in 
>> ODR violations, AFAIK. I'm thinking that there may be
>> some promise in the use of the Barton & Nackmann trick in 
>> combination with the unnamed namespace, but I'm not very 
>> clear about it. 
>> Has anyone already solved this puzzle?
>
>I am not quote sure that I understood your requirements completely, but may
>be something like this:
>
>namespace {
>template< typename T >
>struct disable { typedef T type; };
>}
>
>template< typename T >
>struct my
>{
>    my() { typedef typename disable<my>::type t; }
>};
>
>namespace {
>template<> struct disable< my<int> >;
>}
>
>int main()
>{
>    my<long> m1; // ok
>    my<int> m2; // error
>    return 0;
>}
>
>?
>
>Aleksey
>_______________________________________________
>Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

_____________________________________________________________
A free email account your friends will never forget!
Get YOURNAME_at_[hidden]  at http://www.emailaccount.com/

_____________________________________________________________
Run a small business? Then you need professional email like you_at_[hidden] from Everyone.net  http://www.everyone.net?tag