Newsgroups: comp.std.c
Path: utzoo!henry
From: henry@utzoo.uucp (Henry Spencer)
Subject: Re: Semi constant expressions
Message-ID: <1989Sep6.170011.21564@utzoo.uucp>
Organization: U of Toronto Zoology
References: <1237@gmdzi.UUCP> <10885@smoke.BRL.MIL> <1989Sep5.180315.25627@utzoo.uucp> <10949@smoke.BRL.MIL>
Date: Wed, 6 Sep 89 17:00:11 GMT

In article <10949@smoke.BRL.MIL> gwyn@brl.arpa (Doug Gwyn) writes:
>>I believed then and now that anyone who depended on such things was
>>mad, and the recent examples have not convinced me otherwise.
>
>I think most of us would agree that it is "mad" to write such an
>expression.  However, C provides a preprocessor which is perfectly
>capable of rewriting an innocuous expression so that such a case
>results...

Dennis's objection may have been the same as the one I would have to such
an expression:  exploiting internal side effects at all is a lousy idea,
barring a few well-defined idioms (which don't run into this problem).
There are no "innocuous" expressions which can experience such rewriting,
although there are plenty of poorly-written ones which can.
-- 
V7 /bin/mail source: 554 lines.|     Henry Spencer at U of Toronto Zoology
1989 X.400 specs: 2200+ pages. | uunet!attcan!utzoo!henry henry@zoo.toronto.edu