Re: Legalize access to file descriptors now! (So sue me.)

[Jonathan Wakely <cow at compsoc dot man dot ac dot uk>]

On Thu, Feb 17, 2005 at 11:27:05PM +0100, Richard B. Kreckel wrote:

> On Thu, 17 Feb 2005, Jonathan Wakely wrote:
> [...]
> >     "at least all the ugly stuff is completely hidden away inside a
> >     compiled function and the interface is clear and pristine"
> >
> > Is that valid?  Doesn't a definition of fileno() have to be visible
> > when it's used?
> 
> No.  The declaration is enough provided the correct instantiation is done
> in some translation unit.  And that's what the template<> int
> fileno<char>(const std::ios&) template specialization does.

Ah, yes, I see. Although 14/8 in the standard says:

   A non-exported template must be defined in every translation unit in
   which it is implicitly instantiated (14.7.1  temp.inst ), unless the
   corresponding specialization is explicitly instantiated (14.7.2
   temp.explicit ) in some translation unit; no diagnostic is required.
   [Note: See also 14.7.2  temp.explicit ]

template <C,T> fileno() is implicitly instantiated in any translation
unit that uses it.  Assuming it is only instantiated for std::ios and
std::wios the corresponding specialisations are defined in some other
translation unit, but they're _not_ explicitly instantiated.

To be completely conforming, do you need to add explicit instantiations
to the .cc file, like so?

template int fileno<char>(const std::ios& stream);

template int fileno<wchar_t>(const std::wios& stream);


jon

-- 
"There are only two industries that refer to their customers as 'users'."
	- Edward Tufte