From: Jonathan Turkanis (technews_at_[hidden])
Date: 2003-12-28 16:05:04
I have found a couple small problems with the documentation for
boost::rational, and I have a question about the design, but I haven't seen
any posts from Paul Moore in the last year. Who maintains boost rational
1. All the members of boost::rational are documented as private.
2. Missing semicolon at the end of the class definition.
Why is rational::int_type private? It seems like a fundamental type
associated with rational -- like the value_type of a container, the
value_type of std::complex, the first_type and second_type of std::pair. Why
not make it public?
I've written traits templates pair_traits an nary_traits to provide
compile-time access to the element types and runtime access to the elements
of pair-like and tuple-like types (pair, complex, boost::rational,
boost::interval, boost::tuple, ...) in a uniform manner. It seems strange
that the underlying element type of rational has to be extracted via partial
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk