Re: another effc++ spurious warning?

["Gabriel Dos Reis" <gdr at acm dot org>]

Giovanni Bajo:

> I think it makes sense in the average case that the operator++(int)
> returns a
> value of the class type rather than a reference. At least, for normal uses
> of
> operator overloading and C++. If you do something more advanced like funky
> overloading within complex template metaprogramming (thinking of
> Boost.Spirit
> or similar), you surely don't need -Weffc++, and you are already smart
> enough

I don't believe that you need advanced fu-bhar-programming.
The concept of InputIterator and OutputIterator is a very
very basic concept, for which you have examples for
in the real life.

When, you're limited to faily conventional C-style
programming, you most certainly would comply to -Weffc++.
But, as soon as you start putting abstractions in your code,
the set of things -Weffc+ complains about just tend not to
make much sense.

I strongly feel about these warnings against abstraction