Newsgroups: comp.unix.ultrix
Path: utzoo!utgpu!watserv1!watcgl!idallen
From: idallen@watcgl.waterloo.edu (Ian! D. Allen [CGL])
Subject: Re: unresolved symlinks in Ultrix 4.1 /usr/lib/cmplrs/2.[01]
Message-ID: <1991Feb28.050259.1751@watcgl.waterloo.edu>
Organization: Computer Graphics Laboratory, University of Waterloo, Ontario, Canada
References: <1991Feb27.214103.29826@watcgl.waterloo.edu} <2623@shodha.enet.dec.com>
Date: Thu, 28 Feb 1991 05:02:59 GMT
Lines: 23

Thanks for sending in the comment internally and saving me the grief
of finding a typewriter for the SPR form.

>} as2.1 is distributed as a symlink that points to itself; not very useful:
>} 
>} /usr/etc/subsets/COMU4BACKEND200.inv:0  5       00000   0       0 120755 3/2/90
>}     200     s       ./usr/lib/cmplrs/cc2.1/as2.1    as2.1 COMU4BACKEND200
>
>	Could be another hook for a layered product.  DEC C maybe?

But what good is a symlink that points to itself?

    # cat /usr/lib/cmplrs/cc2.1/as2.1
    /usr/lib/cmplrs/cc2.1/as2.1: Too many levels of symbolic links

I might believe you about the other links that point to things that
aren't there; yes, maybe, layered products.  But why can't those symlinks
be distributed with the layers, not left dangling into nowhere?  Having
stuff like that around makes the Ultrix 4.1 distribution look sloppy and
messes up simple things like "cp -r" or "size *".  Not tidy.  Yuk ugh etc.
-- 
-IAN! (Ian! D. Allen) idallen@watcgl.uwaterloo.ca idallen@watcgl.waterloo.edu
 [129.97.128.64]  Computer Graphics Lab/University of Waterloo/Ontario/Canada