Grant Taylor wrote:
Any suggestions on how a non Python programmer might start looking for
said possible errors?
It sounds like you've already done a lot of the right stuff initially --
tracing packets with tcpdump, etc....
Beyond that, you kind of just have to get stuck into the Python.
Were you trying to act as an NNTP server so that another NNTP server
could ""feed you articles, or were you trying to act as an NNTP server
so that you could use server to server communications rather than client
to server communications?
I was going for bi-directionality.
Using the server interface to accept articles gives you a much easier
and more obvious way to handle re-posts and mail/news/mail loops, using
the exact same mechanisms that real news servers use. With this kind of
mechanism, you could even use Mailman to act as a robo-moderator for a
newsgroup, in addition to acting as the official gateway.
Using the server interface to send articles allows you to avoid some
extra labels in the path, and gives you more control over the other
headers, etc....
But no one else was interested, and I didn't have the skills necessary
to do the code myself. So, the simpler newsreader interface was kept.
Hum. I don't have a log file named "usenet". Below are the log files
that I do have.
Sorry, I meant "fromusenet". It's been a while since I hacked on the
mmdsr script, and I had forgotten what the proper name was for that
particular log.
Unfortunately, you're not having problems in the incoming direction, are
you?
bounce
error
fromusenet
gate_news.cron.txt
post
qrunner
smtp
subscribe
I would be curious to know if you're seeing anything related in the
error, post, or queuerunner logs.
Will you please give me a pointer in the direction to look at where
Mailman stores the history of what articles it has seen so I can explore
this further?
The Python code in /usr/local/mailman/cron/gate_news is actually pretty
readable. A quick scan confirms that the watermark is encoded in the
list configuration pickle, but without knowing more about the data
structures in the list configuration pickles, I couldn't tell you
anything more than that.
Sorry. I thought that I had posted the version. I am (and have
successfully been) running Mailman 2.1.7, and I am running Python 2.3.4
(nice numerical progression of numbers there).
Note that there have been some significant security fixes as well as a
number of other changes since 2.1.7. See the latest version of the
changes at <http://tinyurl.com/6mprx3>.
I'm not doubting that upgrading both Mailman and Python might solve the
problem, but that is (in my opinion) a rather drastic step to resolve
something that was working.
Agreed.
I would be more inclined to do so if I was
doing a new install and not able to get things to work. But, if it
comes down to that and I can save all my current information (archives,
user databases, etc, I'm willing to entertain that option down the road.
Upgrading is usually pretty painless. I've done it several times for
the python.org site where all the official mailman-* mailing lists are
hosted, as well as all of the other official Python-related mailing lists.
It is possible that you may be running into bugs that have already
been quashed.
Like bugs having to do with the history getting full or keeping track of
too many seen messages?
In looking at the list of changes, I did see two or three mentions of
things being fixed related to USENET newsgroups and gatewaying, but I
don't know the exact details of what all was fixed, and what other
things might potentially have been silently fixed while in the process
of fixing other things.
Is there a place that I can review change logs to see what bugs might
have been corrected in subsequent versions of Mailman and / or Python?
From what I can tell, everything is in the launchpad tracker.
The path from <https://launchpad.net/mailman> is not that clear, but if
you follow that TinyURL I posted above, that will take you straight to
the as-of-now current version of the "NEWS" file that should be updated
with every major change to the code, and you should be able to follow
the links up or down from there to the actual code in question.
--
Brad Knowles <[EMAIL PROTECTED]>
LinkedIn Profile: <http://tinyurl.com/y8kpxu>
------------------------------------------------------
Mailman-Users mailing list
Mailman-Users@python.org
http://mail.python.org/mailman/listinfo/mailman-users
Mailman FAQ: http://wiki.list.org/x/AgA3
Searchable Archives: http://www.mail-archive.com/mailman-users%40python.org/
Unsubscribe:
http://mail.python.org/mailman/options/mailman-users/archive%40jab.org
Security Policy: http://wiki.list.org/x/QIA9
