In [EMAIL PROTECTED], Jan Harkes <[EMAIL PROTECTED]> wrote: > I'm not sure whether I can really get away without any constructors or > overloaded allocators. If we really have to lose those it is probably > better to only access RVM objects through a well defined layer written > in C.
OK, wacky kludge alert: Would it be worthwhile to store the important data in C structs that live in the RVM, and then have C++ wrappers objects that have pointers to those structs and do the object oriented things you want?
