Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!usc!orion.oac.uci.edu!cedman
From: cedman@golem.ps.uci.edu (Carl Edman)
Newsgroups: news.software.nntp
Subject: Re: fix for nntp 1.5.10 / GNUS problem
Message-ID: <CEDMAN.90Nov20192202@lynx.ps.uci.edu>
Date: 21 Nov 90 03:21:57 GMT
References: <1990Nov20.221152.10825@csn.org>
Organization: University of California, Irvine, USA.
Lines: 66
Nntp-Posting-Host: lynx.ps.uci.edu
In-reply-to: huntting@csn.org's message of 20 Nov 90 22:11:52 GMT

In article <1990Nov20.221152.10825@csn.org> huntting@csn.org (Brad Huntting) writes:
   I recently upgraded to nntp 1.5.10 so I could use the LOADADV feature.

   Unfortunately I was then swamped by emacs freaks (no offence intented)
   complaining that their readers (GNUS and GNEWS) broke....   *Sigh*...

   We have worked out a client side hack that fixes the problem, but the
   hate mail is still pretty bad.

   Does anyone know when/if this bug will be fixed, and/or if anyone has a
   patch for it?

Client side hack ? I can imagine: Set nntp-maximum-request to 1 and wait
several minutes for a small newsgroup to load. You deserve your hate
mail :-) !

About the hack to fix it. I usually post it once a week in
gnu.emacs.gnus (that is how often the problem comes up :-), but here
it is:

In article <9011091453.AA09160@mandrill.cv.nrao.edu> bglenden@MANDRILL.CV.NRAO.EDU (Brian Glendenning) writes:

   Yesterday GNUS stopped working - the symptom is that when I try to
   enter a newsgroup with more than 2 unread articles it just hangs
   forever. I am trying to find out what software the newshost has
   changed, I suspect nntp but I don't know yet.

   I vaguely recall people discussing something like this. Unfortunately
   I didn't pay enough attention, and now I can't read the old
   discussion. If this is a known problem could someone send me the fix.
   My recollection is that this involves a performance loss. True?

OK, this is about the 30th time people have mailed me or posted here and
asked for the solution. I hope the net excuses me for sending a short
summary of the solution (again).

There are many solutions and "solutions" to this problem, some of
them are quite primitive and make using GNUS almost unbearable,
others are so much work that nobody has yet troubled himself
to do it.

I understand the most popular solution is the "Disable-Timeout-Kludge".
It works like this:

1. Grep the header files of NNTP 1.5.10.
2. Find where TIMEOUT is defined
3. Comment it out
4. Recompile
5. Et voila: Everything works just fine
6. (If it doesn't check that you removed the "(setq nntp-maximum-request 1)" and
"(setq nntp-buggy-select t)" lines from your "~/.emacs" file, if you ever
had them there)

Negative consequence: There used to be a timeout for idle NNTP connections
of about 2 hours. This is gone.

I (in fact this university) can live(and post) with this just fine,
but your mileage may vary.

	Carl Edman


Theorectical Physicist,N.:A physicist whose  | Send mail
existence is postulated, to make the numbers |  to
balance but who is never actually observed   | cedman@golem.ps.uci.edu
in the laboratory.                           | edmanc@uciph0.ps.uci.edu