Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!uakari.primate.wisc.edu!aplcen!haven!adm!smoke!gwyn
From: gwyn@smoke.BRL.MIL (Doug Gwyn)
Newsgroups: comp.lang.c
Subject: Re: pointer representation (was: Re: effect of free())
Message-ID: <11028@smoke.BRL.MIL>
Date: 12 Sep 89 02:53:13 GMT
References: <319@cubmol.BIO.COLUMBIA.EDU> <3756@buengc.BU.EDU> <11021@smoke.BRL.MIL> <2075@munnari.oz.au>
Reply-To: gwyn@brl.arpa (Doug Gwyn)
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 11

In article <2075@munnari.oz.au> ok@cs.mu.oz.au (Richard O'Keefe) writes:
>My point is that optimising compilers generally assume that "equals may
>be subsituted for equals" (that's what equality is all about), so that
>*correct* source code is very likely to be incorrectly *optimised*.

Any my point is that the code was NOT correct, because it relied on the
unwarranted assumption that C pointers' associated ring numbers were
preserved, while the C implementation specifically ignores the ring
numbers, which should be a tip-off that they should not be relied on.
I still maintain that the proper way to deal with this is via some
implementation-specific code that explicitly knows about ring numbers.