*BSD News Article 44983


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!munnari.oz.au!spool.mu.edu!sol.ctr.columbia.edu!news.mindlink.net!agate!violet.berkeley.edu!jkh
From: jkh@violet.berkeley.edu (Jordan K. Hubbard)
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: Re: Major strcmp bug under BSD 2.0?
Date: 6 Jun 1995 04:19:17 GMT
Organization: University of California, Berkeley
Lines: 22
Message-ID: <3r0l05$58c@agate.berkeley.edu>
References: <3qfn52$188j@troy.la.locus.com> <3qo3m8$aq7@park.uvsc.edu> <3qvdsn$ioo@helena.MT.net> <3qvs1d$oj6@park.uvsc.edu>
NNTP-Posting-Host: violet.berkeley.edu

In article <3qvs1d$oj6@park.uvsc.edu>,
Terry Lambert  <terry@cs.weber.edu> wrote:
>Several of us have argued for a long time that most system binaries
>could be dynamically linked, with the exception of init.
>
>Some of us have even proven it by running our systems that way.
>
>The argument that it aids recovery is bogus unless you statically
>link the entire tool chain needed to get a working ld.so.

Bogus?  Oh, I don't think so.  Assuming that the resurrection of the
tool chain is your only possible recourse, then sure - your argument
holds water.  However, thank heavens that Real Life(tm) is somewhat
more flexible!  I've resurrected my ld.so by discovering it in
my /usr/obj dir and copying it up, by tftp'ing it, by grabbing it off
of a DOS floppy, all kinds of ways!  Not once have I needed to reboot
the machine from a floppy or do anything really drastic to get ld.so back.
It was usually enough to go "woo sh*t!  did I really just do that??"
and use one of the many fine _statically linked_ tools to copy a new
one into place before things got too far out of hand! :-)

						Jordan