From: David Abrahams (dave_at_[hidden])
Date: 2006-10-06 12:24:58
"Robert Jones" <robertgbjones_at_[hidden]> writes:
> The boost documentation recommends implementing iterator_core_access
> methods as private static members
No, not really. It just recommends implementing private nonstatic
member functions with similar signatures to the static member
functions used by iterator_core_access.
Incidentally, you shouldn't be looking at those static member
functions. They're not documented for a reason: the interface to
iterator_core_access is an implementation detail of the library :)
> and then making the class a friend of classes derived from
> Why is this preferred to the approach of making
> iterator_core_access a private base of iterator_facade?
Your suggestion doesn't make any sense to me: the recommendation is
for users of iterator_facade, who don't get to control its base
Maybe you meant to suggest making iterator_core_access a private base
of the iterator class derived from iterator_facade? That wouldn't
help, as we're not overriding virtual functions.
-- Dave Abrahams Boost Consulting www.boost-consulting.com
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk