Re: [boost] operators v.2?

From: Daniel Frey (daniel.frey_at_[hidden])
Date: 2002-03-25 05:14:36

Aleksey Gurtovoy wrote:
> 
> How about this:
> 
> template< typename T >
> struct operators
> {
>  private:
>     T& self() { return static_cast<T&>(*this); }
>     T const& self() const { return static_cast<T const&>(*this); }
> 
>  private:
>     template<typename U> static T do_add(T x, U const& y) { return x += y; }
> 
>  protected:
>     // ...
>     T operator+(T const& other) { return T(self()) += other; }
> 
>     // ...
>     template<typename U> friend T operator+(T& x, U const& y) { return
> x.do_add(x,y); }
>     template<typename U> friend T operator+(U const& y, T& x) { return
> x.do_add(x,y); }
> };

Hm, what if the class T contains 'operator U'? For me, this doesn't seem
to work and is therefore dangerous as operator U is prefered over the
given operator+. Thus the code used is not what the user intended. For
all this operator-stuff, I have another version which uses #defines.
It's not what I'd call cool, but it works reliably and yields very few
surprises compared to the native C++-version. Should we consider such an
old solution?

Regards, Daniel

PS: I once intended to clarify a statement I made just a minute ago, but
if I reply to my own post, it doesn't appear in the mailing list. Is
this intendet behaviour of the list or is it a bug?

--
Daniel Frey
aixigo AG - financial training, research and technology
Schloß-Rahe-Straße 15, 52072 Aachen, Germany
fon: [phone], fax: [phone]
eMail: daniel.frey_at_[hidden], web: http://www.aixigo.de