Best practice for avoiding code duplication while implementing iterator and const_iterator classes

Question

What are the best practices to avoid code duplication when implementing class pairs such as iterator and const_iterator or similar?

• Does one usually implement iterator in terms of const_iterator using lots of const_casts?
• Does one employ some sort of traits class and end up defining both iterator and const_iterator as different instantiations of a common template?

This seems like a common enough problem to have a canonical solution but I have failed to find any articles dedicated to that.

Answer 1

I don't have any experience with implementing iterators, although I think this is similar to other projects. Refactor common code, etc.

Looking at GNU libstdc++'s implementation of std::vector::iterator

#include <bits/stl_iterator_base_funcs.h>
// ...
template ... class vector : ... {
    typedef __gnu_cxx::__normal_iterator<pointer, vector> iterator;
    typedef __gnu_cxx::__normal_iterator<const_pointer, vector> const_iterator;
};