Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!ames!think!barmar
From: barmar@think.COM (Barry Margolin)
Newsgroups: comp.lang.c
Subject: Re: C optimizer
Keywords: C pure function optimization
Message-ID: <36662@think.UUCP>
Date: 24 Feb 89 01:36:57 GMT
References: <515@larry.UUCP> <9648@smoke.BRL.MIL> <36034@bbn.COM> <3121@ficc.uu.net> <1028@frog.UUCP>
Sender: news@think.UUCP
Reply-To: barmar@kulla.think.com.UUCP (Barry Margolin)
Organization: Thinking Machines Corporation, Cambridge, MA
Lines: 22

In article <1028@frog.UUCP> john@frog.UUCP (John Woods) writes:
>I think that if
>	(getpid() != getpid())
>ever evaluated to 1, I would be severely astonished.

Well, how about

	(pid = getpid(), (void) fork(), pid != getpid())

?  It will evaluate to 0 in the parent process, 1 in the child
process.

Since fork() changes the value of getpid(), getpid() isn't pure.  But
it would be wrong to make fork() always prevent common subexpression
elimination, since it doesn't modify any of the program's own
variables.

Barry Margolin
Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar