Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!mips!daver!tscs!tct!chip
From: chip@tct.uucp (Chip Salzenberg)
Newsgroups: comp.lang.c++
Subject: Re: Implementing LISP in C++ (type discrimination)
Message-ID: <27E181D5.72D5@tct.uucp>
Date: 16 Mar 91 02:24:21 GMT
References: <1991Mar12.221015.22144@aero.org> <1991Mar13.145107.19724@linus.mitre.org> <1991Mar13.214542.13779@aero.org>
Organization: Teltronics/TCT, Sarasota, FL
Lines: 30

According to jordan@aero.org (Larry M. Jordan):
>... a generic mark function 'void mark(Node*)' which knows how to
>traverse LISP structures, as I indicated before in a non recursive
>way, saving backpointer when chasing car's and cdr's and restoring
>pointers when unwinding.  Writing such a generic mark function seems
>to require doing a case analysis, doesn't it?

In general, an "isKindOf()" test can almost always be replaced by a
virtual function.  For example:

     switch (p->type()) {
	case Cons:    do_cons_thing((Cons *)p);  break;
        case String:  do_string_thing((String *)p); break;
     }

can be replaced with a virtual do_thing() function, with
implementations Cons::do_thing() and String::do_thing().  Then the
above code becomes simply:

     p->do_thing();

>there are a lot of calls to virtual methods during the marking phase.
>How can this be as efficient as an up front case analysis?!

A switch() is typically a table lookup.  So is a virtual function
lookup.  Any more questions?
-- 
Chip Salzenberg at Teltronics/TCT     <chip@tct.uucp>, <uunet!pdn!tct!chip>
 "Most of my code is written by myself.  That is why so little gets done."
                 -- Herman "HLLs will never fly" Rubin