Thank you first. I'm so glad :-) 2007/11/12, Hervé Brönnimann <hervebronnimann@mac.com>:
Ben: I'm shocked to discover such a "natural" extension of prev_ and next_permutation, and that one year later it still isn't in the standard. I have a commitment to algorithms and to C++. This should definitely make it in a TR, but IMHO it's simple enough and useful enough that it should even definitely be in C++0x if there's still time.
I't simple, but not very clear. I know how it works, but I nearly can't prove and explain why. The functions are like std::next_ and prev_permutation, but the classes are not. <algorithm> has no classes. The five classes are a little like STL containers, have the [begin(), end()), but dynamic generating. Another truth, it will exist a conflict between the gacap::next_permutation(first, middle, last) and std::next_permutation(first, last, comp).
I've had success in getting the boost.minmax library into C++0x (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2006/n1990.htm), I'm proud to say it took all of ten minutes of the committee's time in Mont-Tremblant for the making the decision. I'll see if I can work on your proposal. Get in touch later.
There's no (good) proposal yeat. I think it's a hard work. Should I draft a simple proposal?
-- HB