Path: utzoo!attcan!uunet!husc6!bbn!oberon!sdcrdcf!ism780c!haddock!karl
From: karl@haddock.ISC.COM (Karl Heuer)
Newsgroups: comp.lang.c
Subject: Re: Noalias trivia question
Message-ID: <4192@haddock.ISC.COM>
Date: 25 May 88 00:51:13 GMT
References: <14522@brl-adm.ARPA>
Reply-To: karl@haddock.ima.isc.com (Karl Heuer)
Organization: Interactive Systems, Boston
Lines: 16

In article <14522@brl-adm.ARPA> dsill@nswc-oas.arpa (Dave Sill) writes:
>[<bill@proxftl.uucp> (T. William Wells) says that the general aliasing
>problem is unsolvable]
>
>You are implying that programmers can solve unsolvable problems but that
>programs cannot, which is simply not true.  The algorithm the programmer
>applies to determine when "noalias" can be used could be used by an
>optimizer.

The programmer has more information.  In hand-compiling the memcpy routine, I
can look at the functional specification and know that it will never be
legally called with overlapping arrays.  The corresponding way for the
compiler to have this information is via a keyword or #pragma or equivalent
which is attached to the memcpy source code.

Karl W. Z. Heuer (ima!haddock!karl or karl@haddock.isc.com), The Walking Lint