From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID
	autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,6a9844368dd0a842
X-Google-Attributes: gid103376,public
From: eachus@spectre.mitre.org (Robert I. Eachus)
Subject: Re: seperate keyword and seperate compilation with Gnat?
Date: 1996/07/17
Message-ID: <EACHUS.96Jul17113313@spectre.mitre.org>#1/1
X-Deja-AN: 169296438
distribution: world
references: <31D95D93.28D8D15B@jinx.sckans.edu>
 <4rckva$dj1@Starbase.NeoSoft.COM>
organization: The Mitre Corp., Bedford, MA.
newsgroups: comp.lang.ada
Date: 1996-07-17T00:00:00+00:00
List-Id: <comp.lang.ada>



   Someone asked:

   > For example, suppose someone starts worrying whether it is valid to
   > compute a + b as (4 * a + b - 3 * a), which could raise intermediate
   > overflow.

    Lets look at this from a practical point of view.  Assume that 4 *
a + b was computed earlier and is available in a register.  The
compiler can save one load of b this way, so it could be a
"worthwhile" optimization.  Multiplying a by 3 can't cause an overflow
when 4 * a won't, so if you are in the same exception handler scope,
it looks like a valid optimization, even if for other reasons, the
computation of 4 * a + b didn't overflow.  (Perhaps 2 * a + b was in a
register, and b is a large negative number.)

    So I think it is legal unless it could cause an exception in a
different handler scope, again assuming that there was a partial lying
around.
--

					Robert I. Eachus

with Standard_Disclaimer;
use  Standard_Disclaimer;
function Message (Text: in Clever_Ideas) return Better_Ideas is...