At 12:21 PM 2/3/2001 +0100, Edwin Steiner wrote:
>Matthew Cline wrote:
> >
> > We might want to have a member/method by which we can tell what type of
> > variable we're dealing with, so that a function can be passed an arbitrary
> > XV, and treat it differently if it's an SV, AV, and so on.
> >
> > For SVs, we might want to have a method to determine what form of
> > representation the internal data is in: string, int, double, bigint,
> > reference, and so on.
>
>These methods common to all types sound to me like a special case of
>inheritance (like it is handled in C++ (single inheritance) or COM).
>
>derived vtable contains:
> - base vtable (possibly overwritten entries)
> - additional entries for derived class/interface
>
>Would it make sense to order the types in a tree?
No, not really. How the data is stored in a PMC is independent of how we
want it out. Plus we're not going to be deriving anything here--we're a
touch too low-level to be doing inheritance games of that sort.
>Isn't this the way perl5 already "thinks"?
Sort of, yes. Doesn't necessarily mean we need to think similarly, though.
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
[EMAIL PROTECTED] have teddy bears and even
teddy bears get drunk
