Mailman 3 November 2001 - Mailman-Developers

Re: Problem with lockfiles and time in the future
by brandon＠roguetrader.com 12 Nov '01

12 Nov '01

Dan Mick said: > The lock date is set into the future on purpose, so that eventually > stale locks can be automatically reclaimed. > > Some process died holding a lock. Remove it and drive on if you're > sure that process is dead. I have tried that actually, and have nuked the lockfiles and re-run the processes, still no go; new locks appear and the processes just hang again, eventually printing something to the logfile saying that it cannot get a lock. I am totally at loss now, if you are saying the future timestamp is intentional. -Brandon Gillespie

1 0

Re: [Mailman-Developers] Problem with lockfiles and time in the future
by Dan Mick 12 Nov '01

12 Nov '01

The lock date is set into the future on purpose, so that eventually stale locks can be automatically reclaimed. Some process died holding a lock. Remove it and drive on if you're sure that process is dead. > I am having a very odd situation with lock files. My entire mailman > system suddenly stopped working a few days ago, and I have tracked it > down to the lockfiles. In the logs, nothing is able to get a lock. > What I have noticed is the timestamp on the lock file is exactly 10 > hours in the future from the current time, when it tried to get the > lock. This is not a problem with cron or the OS time setting, as I > setup another job in cron exactly like qrunner but through bourne > shell, and it reports time properly, so something from python on down > is causing it to do this oddly. Does anybody have any idea what this > could be? Where I can start digging? I am running mailman 2.0.6 on > Solaris 8, x86. > > -Brandon Gillespie > > (BTW, please CC me on any response as I am not on the mailing list) > > _______________________________________________ > Mailman-Developers mailing list > Mailman-Developers(a)python.org > http://mail.python.org/mailman/listinfo/mailman-developers

1 0

Problem with lockfiles and time in the future
by brandon＠roguetrader.com 12 Nov '01

12 Nov '01

I am having a very odd situation with lock files. My entire mailman system suddenly stopped working a few days ago, and I have tracked it down to the lockfiles. In the logs, nothing is able to get a lock. What I have noticed is the timestamp on the lock file is exactly 10 hours in the future from the current time, when it tried to get the lock. This is not a problem with cron or the OS time setting, as I setup another job in cron exactly like qrunner but through bourne shell, and it reports time properly, so something from python on down is causing it to do this oddly. Does anybody have any idea what this could be? Where I can start digging? I am running mailman 2.0.6 on Solaris 8, x86. -Brandon Gillespie (BTW, please CC me on any response as I am not on the mailing list)

1 0

Other Bounce issues with 2.0.6
by Bob Puff＠NLE 12 Nov '01

12 Nov '01

Hi Gang, Ok, here's an interesting situation: On 11/03, hotmail was down for a bit, bouncing messages. Mailman bounce logs confirm this. They came back up, and all was well. On 11/08, 5 days later, they had the same problem - bounced some messages. Later they fixed the problem. However, Mailman saw the bounces coming in 5 days later, and did its thing removing the people. It didn't realize that there were no bounces in between. How long is the bounce database held? In other words, what happens if a single bounce happens, then 7 days go by, and another single bounce occurs. Will it see this and disable the user? Or does it only look at the 5th day? What happens on a busy list? What about a list that gets a post every couple weeks? Bob

1 0

A few bugs and some thoughts
by Daniel Buchmann 12 Nov '01

12 Nov '01

[this is all about latest CVS, MM2.1a3+] 1. A bug: If digest-mode is disabled for a list, and a member tries to turn it on, a Mailman bug appears: Traceback (most recent call last): File "/home/mailman/scripts/driver", line 96, in run_main main() File "/home/mailman/Mailman/Cgi/options.py", line 456, in main mlist.setMemberOption(user, flag, newval) File "/home/mailman/Mailman/OldStyleMemberships.py", line 242, in setMemberOption raise Errors.CantDigestError CantDigestError And the same occurs the other way around (if only digest mode is enabled, and a member tries to turn it off): Traceback (most recent call last): File "/home/mailman/scripts/driver", line 96, in run_main main() File "/home/mailman/Mailman/Cgi/options.py", line 456, in main mlist.setMemberOption(user, flag, newval) File "/home/mailman/Mailman/OldStyleMemberships.py", line 254, in setMemberOption raise Errors.MustDigestError MustDigestError IMHO, I think a user should not be able to choose between normal and digest mode if only one of them are available. (The choice shouldn't appear at all on their personal memberpage.) 2. Another bug?: Held postings for a list are not deleted when the list is deleted (e.g. with the rmlist command.) A .pck file for each held posting is still around in the data/ directory. Maybe this is true for pending subscriptions as well? 3. A feature request: It would be really great if the text boxes on the [Sender filters] page in addition could accept domain names (or maybe regexps?) instead of email address only. Example: I have several lists where posters would only come from the .no domain. To reduce spam, non-member postings are held for approval. If I could set Mailman to auto-approve postings from .no, and auto-discard all other postings, the list moderators of these lists would be _very_ happy... ;) 4. I'm also experiencing the same problem as totschnig.michael(a)uqam.ca in hist post "mangling of subject header by outlook or mailman". I have Mailman 2.1a3+ on two different linux boxes. I have a test list with the same name and the same config on both. The problem only occurs on one of them! Both Mailman installations are the same version (latest CVS) with the same config. If anyone has any idea of where I might look for differences in the two installations, let me know... maybe I could find out what causes it.. :)

1 0

Mod to list_members to show nomail, digest types
by Bob Puff＠NLE 12 Nov '01

12 Nov '01

Hey gang, A while back someone requested a way to get the listing of nomail list members. I recently needed not only that, but also the ability to sort out digest mime users, and digest plain text users. Below is a replacement list_members program (that goes in $prefix/bin) that adds four new commands with which to accomplish this: -x = exclude nomail members -n = list only nomail members -m = list digest mime members -l = list plain text digest members This script was modified from the 2.0.6 Mailman code - NOT tested with 2.1!!! Enjoy! Bob ========cut here============= #! /usr/bin/env python # # Copyright (C) 1998,1999,2000 by the Free Software Foundation, Inc. # # This program is free software; you can redistribute it and/or # modify it under the terms of the GNU General Public License # as published by the Free Software Foundation; either version 2 # of the License, or (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License # along with this program; if not, write to the Free Software # Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. """List all the members of a mailing list. Usage: %(program)s [-o file] [-r] [-d] [-p] [-h] listname Where: -o file or --output file Write output to specified file instead of standard out. -r or --regular Print just the regular (non-digest) members. -d or --digest Print just the digest members. -x or --exclude Exclude all nomail members from list -n or --nomail Print only the nomail members -l or --plain Print only plain text digest members -m or --mime Print only mime digest members -p or --preserve Output member addresses case preserved the way they were added to the list. Otherwise, addresses are printed in all lowercase. -h or --help Print this help message and exit. listname is the name of the mailing list to use. Note that if neither -r or -d is supplied, both regular members are printed first, followed by digest members, but no indication is given as to address status. """ import sys import string import getopt import paths from Mailman import MailList from Mailman import Errors from Mailman import Defaults from Mailman import mm_cfg program = sys.argv[0] rmem = {} dmem = {} def usage(status, msg=''): print __doc__ % globals() if msg: print msg sys.exit(status) def main(): try: opts, args = getopt.getopt( sys.argv[1:], 'dprnmxlo:h', ['digest', 'regular', 'preserve', 'nomail', 'mime', 'xclude', 'plain', 'output=', 'help']) except getopt.error, msg: usage(1, msg) if len(args) <> 1: usage(1) listname = string.lower(args[0]) outfile = None regular = None digest = None preserve = None nomail = None mimef = None xclude = None plaint = None for opt, arg in opts: if opt in ('-h', '--help'): usage(0) elif opt in ('-o', '--output'): outfile = arg elif opt in ('-r', '--regular'): regular = 1 elif opt in ('-d', '--digest'): digest = 1 elif opt in ('-p', '--preserve'): preserve = 1 elif opt in ('-n', '--nomail'): nomail = 1 elif opt in ('-m', '--mime'): mimef = 1 elif opt in ('-x', '--exclude-nomail'): xclude = 1 elif opt in ('-l', '--plain'): plaint = 1 if plaint: digest = 1 if mimef: digest = 1 if regular is None and digest is None: regular = digest = 1 if xclude is None and nomail is None: xclude = nomail = 1 if mimef is None and plaint is None: mimef = plaint = 1 if outfile: try: fp = open(outfile, 'w') except IOError: print 'Could not open file for writing:', outfile sys.exit(1) else: fp = sys.stdout try: mlist = MailList.MailList(listname, lock=0) except Errors.MMListError, e: print 'No such list "%s"\n%s' % (listname, e) sys.exit(1) # filters for recipient calculation def delivery_enabled_p(x, s=mlist, v=mm_cfg.DisableDelivery): return not s.GetUserOption(x, v) def delivery_disabled_p(x, s=mlist, v=mm_cfg.DisableDelivery): return s.GetUserOption(x, v) def likes_mime_p(x, s=mlist, v=mm_cfg.DisableMime): return not s.GetUserOption(x, v) def hates_mime_p(x, s=mlist, v=mm_cfg.DisableMime): return s.GetUserOption(x, v) if preserve: for addr in mlist.GetDeliveryMembers(): rmem[addr] = addr for addr in mlist.GetDigestDeliveryMembers(): dmem[addr] = addr else: for addr in mlist.GetMembers(): rmem[addr] = addr for addr in mlist.GetDigestMembers(): dmem[addr] = addr rmembers = filter(delivery_enabled_p, rmem.keys()) rnomembers = filter(delivery_disabled_p, rmem.keys()) dmem2 = filter(delivery_enabled_p, dmem.keys()) dmembers = filter(hates_mime_p, dmem2) dmime = filter(likes_mime_p, dmem2) dmem2 = filter(delivery_disabled_p, dmem.keys()) dnomembers = filter(hates_mime_p, dmem2) dnomime = filter(likes_mime_p, dmem2) stdout = sys.stdout try: sys.stdout = fp if regular: if xclude: for addr in rmembers: print addr if nomail: for addr in rnomembers: print addr if digest: if xclude: if plaint: for addr in dmembers: print addr if mimef: for addr in dmime: print addr if nomail: if plaint: for addr in dnomembers: print addr if mimef: for addr in dnomime: print addr finally: sys.stdout = stdout if __name__ == '__main__': main()

1 0

Bounce handler problem - another one in 2.0.6
by Bob Puff＠NLE 11 Nov '01

11 Nov '01

Just discovered this one: A user was subscribed to the list using some capital letters in their email address. Emails to ther user were bouncing, and the bounce handler did properly detect the user's address. However, the bounce log shows "User not a member", and shows the email in lower case. Other than the case, the address -is- exactly the same. So, I assume that the mixed case is the reason the bounce handler is not taking the guy out. Ideas? Fixes? Bob

1 0

cron/checkdbs error
by Daniel Buchmann 10 Nov '01

10 Nov '01

If there are subscription/unsubscription requests waiting, but no moderated posts waiting for approval, cron/checkdbs gives me this error: Traceback (most recent call last): File "cron/checkdbs", line 104, in ? main() File "cron/checkdbs", line 67, in main text = text + '\n' + pending_requests(mlist) File "cron/checkdbs", line 82, in pending_requests when, addr, passwd, digest, lang = mlist.GetRecord(id) ValueError: unpack tuple of wrong size This only seems to happen on newly created lists (e.g. lists that never have had any moderated posts waiting for approval). How to produce the error: 1. Create a new list, set steps required for subscription to "Require approval" or "Confirm and approve". 2. Subscribe to the list. 3. run cron/checkdbs when the subscription request is waiting for moderator approval.

2 1

clobber_date bug
by Eric Seppanen 09 Nov '01

09 Nov '01

Hello nice Mailman folks, thanks for the great software. I've been using it for over a year to run a handful of small mailing lists and I've always been impressed with the quality. I just noticed a small bug. On an active list, I turned on the clobber_date setting (to apply the "resent" time to messages stored in the archives), and I noticed that it messed up the order of a bunch of messages in the archive. The reason is that clobber_date doesn't put the timezone information in the Date: field. So it looks as though the messages are dated GMT instead of GMT-6, and the messages from this morning appear after those from the afternoon. Example message without clobber_date: Date: Thu, 8 Nov 2001 10:48:20 -0600 Example message with clobber_date: Date: Thu Nov 8 13:33:00 2001 X-Original-Date: Thu, 08 Nov 2001 13:31:30 -0600 Thanks, Eric

2 1

FAQ-o-Matic
by barry＠zope.com 09 Nov '01

09 Nov '01

A while back it was suggested that we implement an on-line interactive FAQ, where the community can add and expand on questions. I've now adapted the Python FAQ wizard to add a Mailman FAQ wizard. See http://www.python.org/cgi-bin/faqw-mm.py There's only 3 questions and answers there so far, but you can help make it a better FAQ! I'm opening up access so that anybody can add or change FAQ entries. This is an honor system, and the password is "Mailman". Let me know if you think we need more, or different categories. Enjoy! -Barry

2 1