Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!ut-sally!ut-ngp!infotel!pollux!bobkat!m5d
From: m5d@bobkat.UUCP (Mike McNally )
Newsgroups: comp.lang.c
Subject: Re: C and overflow anomolies
Message-ID: <881@bobkat.UUCP>
Date: Thu, 16-Apr-87 09:09:33 EST
Article-I.D.: bobkat.881
Posted: Thu Apr 16 09:09:33 1987
Date-Received: Sun, 19-Apr-87 10:27:57 EST
References: <15958@sun.uucp> <5716@brl-smoke.ARPA>
Reply-To: m5d@bobkat.UUCP (Mike McNally (Man Insane))
Distribution: world
Organization: Digital Lynx, Inc; Dallas, TX
Lines: 24

In article <827@xanth.UUCP> kent@xanth.UUCP (Kent Paul Dolan) writes:
>	                  . . .If I wrote (all "float" variables):
>		temp = a - b;
>		d = temp + c;
>	in a piece of existing code to avoid the above overflow; the
>	compiler writer is now free to elide temp from the code, then go
>	back and rearrange to achieve:
>		d = (a + c) - b;
>
>UUCP  :  kent@xanth.UUCP   or    ...{sun,cbosgd,harvard}!xanth!kent

Not a C compiler writer.  A C compiler (as I know it; I suppose my 
complete ignorance of ANSI C may show here) is not supposed to do
any non-expression-local optimizations, not even basic block type
stuff.  Thus while a Pascal compiler may do such work, or for that
matter an optimizing FORTRAN compiler, no C compiler will.



-- 
Mike McNally, mercifully employed at Digital Lynx ---
    Where Plano Road the Mighty Flood of Forest Lane doth meet,
    And Garland fair, whose perfumed air flows soft about my feet...
uucp: {texsun,killer,infotel}!pollux!bobkat!m5d (214) 238-7474