AMDG Niels Dekker - mail address until 2008-12-31 wrote:
Hmmm... not such a good idea of mine! Doing so would lead to ambiguity when the type T would /not/ have its own swap overload, while it would have the std namespace as an associated namespace. For example, when T would be a pointer-to-std::string or an std::time_base::dateorder enum.
I wish this boost_swap_impl::swap could have been a little less "specialized"! (But still, it would need to be more specialized than boost::swap itself!)
There's a config macro |BOOST_FUNCTION_SCOPE_USING_DECLARATION_BREAKS_ADL |Does it help to a) using namespace std; or b) put the using in the namespace rather than in the function
Anyway, I think we should add unit tests for std types!
Definitely. In Christ, Steven Watanabe