Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site amdahl.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!decwrl!sun!amdahl!gam
From: gam@amdahl.UUCP (gam)
Newsgroups: net.bugs.4bsd,net.unix,net.wanted.sources
Subject: sum(1) for machine-independent chksum: bad
Message-ID: <969@amdahl.UUCP>
Date: Fri, 18-Jan-85 00:35:45 EST
Article-I.D.: amdahl.969
Posted: Fri Jan 18 00:35:45 1985
Date-Received: Sun, 20-Jan-85 05:12:42 EST
Distribution: net
Organization: Blue Mouse Trailer Resort, Hellmouth, CA
Lines: 15
Xref: watmath net.bugs.4bsd:1330 net.unix:3361 net.wanted.sources:337

Some version of a shar-type program out there is using
sum(1) to examine the unshar-ed files, comparing it to
the sum of the files at the machine of origin.

In 4.2 the sum is taken modulo BUFSIZ (from stdio.h) while
in SVR2 it is taken modulo BUFSIZE (defined in sum.c as 512).

Great, huh?

So let's all agree to do it with 512 or not use sum(1) for
machine-independent checksumming.

Thank you.
-- 
Gordon A. Moffett		...!{ihnp4,hplabs,sun}!amdahl!gam