On (17 Jul 97) Sunir Shah wrote to Jerry Coffin...
 SS> I'm nowhere close.  Since when can you do this:
 SS> int foo( void ) const;
It's been quite a while.  This is valid only for member functions and
indicates that the member function isn't going to change the class of
which it's a member.
 JC> Sort of - C++ specifically allows for other implementation
 JC> defined forms of main, rather than merely making all other forms
 JC> undefined.  In the end, it makes little or no real difference in
 JC> how things work; if your implementation says you can use some
 JC> other form of main, you can.  Otherwise, you can't (at least
 JC> dependably.)
 SS> Yippee.  So main() is no longer a function.  It's a mini-
 SS> environment, oblivious of the OS, like in Pascal.  I thought C++
 SS> was supposed to be an improved C?  Pascal is going in the wrong
 SS> direction. :)
Maybe I'm missing something, but I don't see how being implementation
defined makes things much (or really ANY) different from how they've
been all along.  If a particular implemenation wants to accept a
different form of main, it's always been able to do so.
 SS> There are some very good reasons to return values from main()
 SS> and hard code that into the standard, such as when you want to
 SS> consider the rest of world and not just Win95...
Sure.  If you're interested in portability, you don't use things that
are implementation defined.  Can you explain how `void main()' being
undefined is ANY different from it being implementation defined?  If
there's a real difference, I'm missing it...
    Later,
    Jerry.
... The Universe is a figment of its own imagination.
--- PPoint 1.90
---------------