Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!nuchat!uhnix1!sugar!peter
From: peter@sugar.UUCP (Peter da Silva)
Newsgroups: comp.lang.c
Subject: Re: another auto-increment problem
Message-ID: <868@sugar.UUCP>
Date: Thu, 8-Oct-87 10:41:54 EDT
Article-I.D.: sugar.868
Posted: Thu Oct  8 10:41:54 1987
Date-Received: Sun, 11-Oct-87 08:37:44 EDT
References: <9600@brl-adm.ARPA> <15298@topaz.rutgers.edu>
Organization: Sugar Land UNIX - Houston, TX
Lines: 25

In article <15298@topaz.rutgers.edu>, ron@topaz.rutgers.edu (Ron Natalie) writes:
> Certainly variable
> number of arguments is easier handled by the first, but they are
> forbidden in ANSI C anyway.

Not in the draft I have. In fact the draft I have has syntax for handling
variable numbers of arguments: printf(char *, ...);

> The order is undefined not just because stack arguments may be placed
> in either order, but because the call linkage is not defined by the
> C language at all.

The call linkage is effectively defined by "what is efficient to perform
on the target macgine". Since this may vary, it's undefined in the language.
In practice, it's more efficient to evaluate and stack the arguments in
reverse order for conventional hardware architectures.

> An implementor is free to place all the args in
> an envelope and mail them to the subroutine as far as the C standard
> is concerned.

Well put. Personally I prefer "call by long distance" to "call by USPS" :->.
-- 
-- Peter da Silva  `-_-'  ...!hoptoad!academ!uhnix1!sugar!peter
-- Disclaimer: These U aren't mere opinions... these are *values*.