Beman Dawes <[EMAIL PROTECTED]> writes: > At 07:57 PM 6/2/2003, David Abrahams wrote: > > >I'm going to want to replace the old Boost iterator adaptors > >implementation with the new one in the Boost sandbox pretty soon, and > >while they are similar in intent and spirit, they have totally > >incompatible interfaces. In fact, the new one lives in a subdirectory > >boost/iterator, but I think it would be highly unwise to keep the old > >header around since it defines some of the same names, notably, um, > >"iterator_adaptor" ;-). It would be an invitation to ODR violations. > > Just to verify, the new interface corresponds to the proposal now > before the C++ committee?
Yes. > >This is going to break every boost library and every user which is > >depending on the current CVS version of iterator_adaptors. The new > >adaptors are so much better designed and easier to use that making the > >transition should be cake, but before I do anything drastic I want to > >hear how loud the pre-emptive screaming is ;-) > > Is the sandbox version ready? Yes. > I'd like to have the filesystem library ready to change > simultaneously so the regression tests themselves don't break. (They > depend on filesystem, and it depends on iterator_adaptors.) > > >Objections? > > Better now than later. Do you have a specific change-over date in mind? In the next week, or 2 at the outset. -- Dave Abrahams Boost Consulting www.boost-consulting.com _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
