Xref: utzoo rec.humor:18668 rec.humor.d:1588 comp.misc:4997
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uflorida!gatech!bloom-beacon!athena.mit.edu!jik
From: jik@athena.mit.edu (Jonathan I. Kamens)
Newsgroups: rec.humor,rec.humor.d,comp.misc
Subject: Re: Looking for Computer Folklore
Message-ID: <9183@bloom-beacon.MIT.EDU>
Date: 9 Feb 89 06:04:51 GMT
References: <7143@pyr.gatech.EDU> <532@geovision.UUCP> <4575@tekgvs.GVS.TEK.COM> <6255@saturn.ucsc.edu> <1232@raspail.UUCP>
Sender: daemon@bloom-beacon.MIT.EDU
Reply-To: jik@athena.mit.edu (Jonathan I. Kamens)
Organization: Massachusetts Institute of Technology
Lines: 45

In article <1232@raspail.UUCP> bga@raspail.UUCP (Bruce Albrecht) writes:
>
>...
>
>seems that RSTS/E sends out a message informing the users that the system is
>on its way up, and when the message was sent, the loopback plug turned it
>into a user input, to which the system sent a message 'input ignored.', 
>which also became user input ..., and the system died because it ran out of
>free buffers.

Sounds like a bug I just fixed in the syslogd we use here at Athena.

Our syslogd is modified to allow syslog message to be sent out to
users over the network using the zephyr message delivery system.  One
of the requirements of all pieces of software using zephyr is that
they call the routine ZInitialize() before using any zephyr functions.

Well, one day I decided that I wanted to receive *all* syslog messages
on my workstation, so I placed "*.debug   jik" in my syslog.conf file.
Instead of receiving all messages, the next time I rebooted the
workstation I stopped receiving any messages at all!

It turns out that the person who put the zephyr code into the syslogd
sources placed it *after* the syslogd init call.  The init call
attempts to log a startup message on level syslog.warning.  Since
*.debug gets syslog.warning messages, it attempted to send me the
startup message over zephyr.  Well, ZInitialize() hadn't been run
yet, so this generated an error, which syslogd promptly attempted to
log to syslog.error, which therefore attempted to send a notice out to
me, which generated an error, which syslogd promptly attmpted.... you
get the idea.  It eventually died a cruel, horrible death and didn't
so much as leave a core file.

The fix was nothing more than moving two lines of code eight lines
down in the sources.

Of course, in the infinite wisdom of the release engineering team
here, that fix has *still* not been installed in the standard release,
although it's been several months since I submitted it :-)


Jonathan Kamens			              USnail:
MIT Project Athena				410 Memorial Drive, No. 223F
jik@Athena.MIT.EDU				Cambridge, MA 02139-4318
Office: 617-253-4261			      Home: 617-225-8218