Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!mcdchg!chinet!les
From: les@chinet.chi.il.us (Leslie Mikesell)
Newsgroups: comp.unix.wizards
Subject: Re: ln -f, corrections
Message-ID: <1990Aug01.165645.3434@chinet.chi.il.us>
Date: 1 Aug 90 16:56:45 GMT
References: <3770@auspex.auspex.com> <18475@rpp386.cactus.org> <3797@auspex.auspex.com>
Organization: Chinet - Public Access UNIX
Lines: 18

In article <3797@auspex.auspex.com> guy@auspex.auspex.com (Guy Harris) writes:

>Which means the ultimate problem isn't that one behavior is "good" and the
>other is "bad", but that they're *different*.  Standardizing on either
>one would have worked (modulo windows opened by having to implement one
>behavior with multiple commands on a system that provides the other).

On systems without ln -f (where ln defaults to removing the target if
it already exists) another program is required to perform a link which
fails if the target exists.  On SysVr[23] on 3B2's, this program exists
as /etc/link:
-r-x------   1 root     bin         1716 Mar 3  1988 /etc/link
Thus, unless someone changes things there is no atomic file action
that can be used by an ordinary user in a shell script.  I'd call
that "bad".

Les Mikesell
  les@chinet.chi.il.us