Boost logo

Boost :

From: David Abrahams (abrahams_at_[hidden])
Date: 2001-01-26 12:34:40

Wow, that's a terriffic simplification if it works -- like proving Fermat's
theorem in terms of hyperbolic geometry (but far easier to check) ;-)
I'd love to hear from Peter Dimov whether he believes this fits the bill.

----- Original Message -----
From: "Dietmar Kuehl" <dietmar_kuehl_at_[hidden]>

> Hi,
> Gabriel Dos Reis wrote:
> > | template <typename T>
> > | struct swap_t {
> >
> > That is what I used to suggest. It provides a uniform framework hook
> > for standard function "specialisations". It doesn't have the
> > overloading surprise.
> I'm sorry for repeating a suggestion when jumping into this thread
> lately... However, after sending the previous message it occured to
> me that at the London meeting we convinced the Core group to
> keep template template arguments by clearly specifying the *exact*
> semantics on the basis of the wording without templates (well, it was
> not the only argument...): Basically, template template arguments
> should work like using 'helper' as template argument and using
> 'helper::foo<T>':
> struct helper {
> template <typename T> class foo { ... };
> };
> That is, the template template arguments didn't add a new feature
> to the language but merely simplified and existing one. Maybe we
> can convince them to do the same thing with the suggestion above
> for partial function template specialization...

Boost list run by bdawes at, gregod at, cpdaniel at, john at