Jeff, you are right. I wonder if non-linear hierarchy can be mapped to a GCD problem anyway. I guess it is possible for a finite set of hierarchies, but figuring out is a rather hard job todo. Therefore I'll switch to a min (or max) problem using consecutive integers. Another advantage of this is that the hierarchy can be supplied as an mpl sequence of types and the forward/backward mapping of numbers can happen implicitely inside widest_convertible: ------ template<class Sequence, class LinearHierarchySequence> struct widest_convertible { typedef typename .... type; }; typedef mpl::vector< boost::incrementable_traversal_tag boost::single_pass_traversal_tag, boost::forward_traversal_tag, boost::bidirectional_traversal_tag, boost::random_access_traversal_tag> linear_hierarchy; widest_convertible< mpl::vector<boost::forward_traversal_tag, boost::bidirectional_traversal_tag>, linear_hierarchy>::type t; ----- Best regards, Christoph On Mon, Sep 28, 2009 at 9:48 PM, Jeffrey Hellrung <jhellrung@ucla.edu> wrote:
Christoph Heindl wrote:
Clever! Though in this instance, the minimum is identical to the gcd, at which point you could just assign the traversal tags to consecutive integers ;)
I'm not sure if I understood you correctly: Assigning, for example, 1:readable, 2:single pass and 3:forward iterable would yield 1:readable for the gcd of 2:single pass and 3:foward which is not true (IMHO :).
Oh, I meant that you could replace gcd with min (in a linear hierarchy).
- Jeff
_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost