TONGARI wrote:
Do you have codes? Something that just works may not be guaranteed to work...
Very true. I have a zip archive on dropbox: http://dl.dropbox.com/u/3512486/underlying.zip (7 kB) It's all quite immature and it would need rigorous boostification to eventually become a candidate for a review, but the implementation works and I believe it to be without errors (whether conceptual or practical). My own template metaprogramming solution for the underlying_type might be replaced by boost::aligned_storage. Thanks to Mathias Gaunard for bringing it to my attention.
It probably depends on what you mean by 'move semantics' and what you mean by 'implement'. What I'm sure of, is that my library can (unintrusively) simulate move constructors, move destructors and the returning of rvalue references from functions. Of course, it doesn't use *only* bitwise copies for that purpose and the type involved needs to be default-constructible.
What type needs to be default-constructible? Something like underlying_type_traits<T>::underlying_type or T itself?
T itself.