Path: utzoo!attcan!uunet!lll-winken!ncis.llnl.gov!helios.ee.lbl.gov!pasteur!agate!bionet!csd4.milw.wisc.edu!nic.MR.NET!xanth!ukma!husc6!cmcl2!adm!smoke!gwyn
From: gwyn@smoke.BRL.MIL (Doug Gwyn )
Newsgroups: comp.lang.c
Subject: Re: Turbo C bug !
Message-ID: <9487@smoke.BRL.MIL>
Date: 23 Jan 89 21:08:27 GMT
References: <PK-TLE.89Jan21103247@shai.sics.se>
Reply-To: gwyn@brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>)
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 9

In article <PK-TLE.89Jan21103247@shai.sics.se> pk-tle@sics.se (Tommy Levitte) writes:
>(57600 = 240*240   !!!!!!!)

Not with 16-bit ints, it doesn't.  240*240 overflows a 16-bit int,
with generally unpredictable results.  Apparently Turbo C treated
the result as -7936, which is what you get on "silent" overflow
on a 2's-complement 16-bit machine.  That's permissible.  As you
noted, forcing the computation to be done using longs avoids the
overflow.