From: "Gennadiy Rozental" <gennadiy.rozental@thomson.com>
You seem to be arguing that
template<typename T, typename U = traits<T> > class Foo { };
is wrong because it blurs the distinction between traits and policies. I
I think Gennadiy would agree that this is fine because U is a policy template parameter defined to use at least a subset of traits' interface.
I agree. It may even theoretically be possible that Foo uses U interface that has nothing to do with trait<T> being the trait. But! It would make me very wary. Even though I admit above is possible, in most cases it's sign of incorrect design.
I'm not sure it warrants wariness. If Foo defines semantics for its U policy and those semantics just happen to be fulfilled by a traits class, what's the problem with using the traits class as the policy argument? It doesn't mean other policy classes aren't possible, just that there is a convenient one available in an existing traits class.
If U isn't well-defined and traits becomes the definition of U, even though Foo doesn't use all of traits' interface, then the design of Foo is poor. That would seem to be the case with std::basic_string and std::char_traits. I haven't looked at the implementation with this in mind, but I'm guessing that std::basic_string doesn't use all of std::char_traits interface, and yet the "traits" parameter of std::basic_string isn't defined separately.
I am not sure where you going with "doesn't use all of std::char_traits" IMO it does not matter which part if char_traits interface basic_string is using.. In any case it should've referred to it by name and not through template parameter.
I mean that std::basic_string's "traits" policy parameter may only require a subset of what is defined in std::char_traits. If that's the case, it is a mistake, in my mind, not to separately document the required semantics for that policy. Thus, while std::char_traits may fulfill the requirements of that policy, it is a superset that should be documented separately. Whether std::basic_string should have a policy or should just use a traits class is another matter entirely. -- Rob Stewart stewart@sig.com Software Engineer http://www.sig.com Susquehanna International Group, LLP using std::disclaimer;