[boost] Re: [math constants] VC++ experts needed :)

From: Daniel Frey (d.frey_at_[hidden])
Date: 2004-03-11 02:52:18

Bronek Kozicki wrote:
> Daniel Frey <d.frey_at_[hidden]> wrote:
> [...]
> 
> I tried following code:
> 
> #include <iostream>
> template <typename T> class V {};
> template <template <typename> class T> struct B {};
> struct D : B<V> {};
> template <typename T>
> void f (const B<T>&)
> {
>  std::cout << typeid(T).name() << std::endl;
> }
> int main()
> {
>   D d;
>   f (d);
> }
> 
> According to Comeau compiler and GCC this program is invalid. I have

Yes, the code is invalid. The correct code (used in the constant 
library) looks like this:

template<template <typename> class T>
void f( const B<T>& ) {}

T is than deduced to V. V is not a valid class on itself, it's just a 
valid template for a template template parameter. However:

> .... however it compiles under VC71, and the output of this program is:
> class V
> 
> I think that's a surprise?

Nope. My skills in guessing VC-bugs just seem to get better every day :) 
Could you ask on some MS-support-groups about this, please? I think it's 
wrong to compile the code and deduce T to V, as V is not a class. The 
MS-folks should please either confirm the bug or kindly point me to the 
standard text that allows this behaviour.

Regards, Daniel