Archived posting to the Leica Users Group, 1999/06/05

[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]

Subject: [Leica] Resolution of "missing digests" problem
From: Brian Reid <reid@mejac.palo-alto.ca.us>
Date: Sat, 05 Jun 1999 11:07:21 -0700

Dear Leica-Users subscribers,

It took a couple of days, but I have found and repaired the problem
that was causing some digests not to be sent. Here is what happened.
For quite some time, the Leica digests had been festooned with an extra

Topics in this digest:
  _SUBJECTS_

at the beginning. This was an annoyance, but only a minor one, so I let
it go for a long time. On May 20, I finally figured out what was
causing those extra lines, and I changed it. The problem was that the
Leica mailing list was being routed through the Digest processor twice,
and the second pass through found only one message, the Digest itself,
and so did not generate a list of subjects to replace the token
_SUBJECTS_.

I changed the distribution table so that it routed through the digest
processor only once, and checked my notes to see if I had recorded a
reason for setting it up that way in the first place. I had not written
anything down, so I assumed it was a mistake.

But:  about a year ago, the Leica list was given a limit that no
individual message could be bigger than a certain size. But digests
must be immune to that size limit, because they are combinations of
many messages. In routing the messages past the digest processor, I had
activated the "this is not a digest, so it must be subject to a size
limit" logic. I think that is probably a latent bug in the majordomo
software and not anything I did wrong, but the end result is the same.
When a digest is generated by having the waiting-list overflow, during
the middle of the day, the software is smart enough to know that it is
not supposed to do this checking, but when a digest is generated by the
clock, at midnight California time, the software made the digest and
then checked to see if it was small enough to send. In many cases it
was not. I have modified the majordomo software so that it no longer
makes this check. I won't know for sure that this fix is working until
tonight, since only the clock-triggered digests are failing, but I'm
confident enough that this is the problem that I am announcing its
repair.

Those of you who subscribe to the digest will know that I have just
send the last 5 missing digests to all digest subscribers. I'm not
going to go back in time any farther than that. If you want older
digests, you will need to request them from the archives. 

Here is how to request a digest from the archives. Send an email message

	To: majordomo@mejac.palo-alto.ca.us
	Subject: 

	get leica-users-digest v09.n004
	get leica-users-digest v09.n005

and so forth. Here is the complete list of digests that were not sent,
and that I have not already retransmitted:

	get leica-users-digest v09.001
	get leica-users-digest v09.004
	get leica-users-digest v09.005
	get leica-users-digest v09.009
	get leica-users-digest v09.011
	get leica-users-digest v09.015
	get leica-users-digest v09.017

I am thinking about turning off the clock-generated issue of the Leica
Users Digest and let the list volume generate all of them. The original
purpose of the clock-generated digest was to make sure that digest
subscribers got at least one message a day, so that they would not
bother me with "am I still on the list?" questions. As I study the
logs, I see that on May 22, May 23, May 13, April 27, April 8, and
April 7 there would have been intervals of longer than 24 hours without
a digest being generated. 

Brian Reid
your Leica Users mailing list operator