From 4384064843952398074
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: f78e5,e94695845965b249
X-Google-Attributes: gidf78e5,public
X-Google-ArrivalTime: 1992-12-16 13:01:43 PST
Newsgroups: comp.std.c++
Path: sparky!uunet!utcsri!torn!nott!bnrgate!bcrka451!bcrki65!sjm
From: sjm@bcrki65.bnr.ca (Stuart MacMartin)
Subject: Re: Zero-length structures and pointer comparisons
Message-ID: <1992Dec16.202711.22367@bcrka451.bnr.ca>
Sender: 5E00 Corkstown News Server
Organization: Bell-Northern Research Ltd., Ottawa, Canada
References: <1992Dec9.133956.29659@lth.se> <1992Dec9.191606.5665@lucid.com> <BzCG7K.2sG@frumious.uucp>
Date: Wed, 16 Dec 1992 20:27:11 GMT
Lines: 40

In article <BzCG7K.2sG@frumious.uucp> pat@frumious.uucp (Patrick Smith) writes:
>
>jss@lucid.com (Jerry Schwarz) writes:
>|And
>|
>|       if p < q is defined,  ptrcmp(p,q) < 0   <=>  p < q
>
>
>This strikes me as nice to have, but not essential.
>
>On the other hand, this might be fairly cheap, given that
>we're already insisting on
>
>   p == q   =>   ptrcmp(p,q) == 0
>

It makes me nervous to disagree with you, but haven't we already 
hammered to death the concept that p < q might be defined (but 
not generate a total ordering) if some ps and qs are not in the
same segment?  Or do the PC compilers only permit p < q for ps
and qs in the same array?

Suppose we have a compiler that uses segment offset to determine
if p < q but memory address to determine if p == q.

If p1 < q1 are near the end of segment 1,
   p2 < q2 are near the beginning of segment 2
and q1 == p2 (because they refer to the same memory location),

then
   
   p2 < q2 < p1 < q1   (because the segment is being ignored)
but
   p2 == q1   and   p2 != q2.

Stuart
--
: Stuart MacMartin                                    email: sjm@bnr.ca      :
: Bell-Northern Research                              phone: (613) 763-5625  :
: PO Box 3511, Stn C, Ottawa, K1Y-4H7, CANADA    Standard disclaimers apply. :