[Mailman-Users] Bounce Handling Question
I am running Mailman for two announcement only mailing list. It is running with no problems. I wanted to send out an announcement to all of my members without them getting a duplicate if they were on both lists. I created a new list called all yesterday. I put the two lists under the regular_include_lists setting. It ran perfect. But, I was wondering, how does the bounce handling work for that list? If a user bounces a message when I send to the all list, do they get removed from the other list they are actually on? The all list does not have any members associated with it. Thanks, Jeff -- 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
[Mailman-Users] remove_members program
I am finding myself using the remove_members program a good amount of time. I am part of the AOL and Yahoo feedback loops and I unsubscribe any member who classifies my mail as spam. Is there a way to either use the remove_members program of unsubscribe a member from all lists using the web? I would prefer this instead of having to connect to the server via ssh each time I want to use it. Thanks, Jeff -- 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
Re: [Mailman-Users] bouce processing problem
Dimitrios Karapiperis wrote: Is there anything I can do for the floating point issue? It probably isn't an issue. I'm just guessing with incomplete information. here is a chink form error file admin(28713): Traceback (most recent call last): 173 admin(28713): File /usr/lib/mailman/scripts/driver, line 101, in run_main 174 admin(28713): main() 175 admin(28713): File /usr/lib/mailman/Mailman/Cgi/options.py, line 178, in main 176 admin(28713): mlist.ConfirmUnsubscription(user, userlang) 177 admin(28713): File /usr/lib/mailman/Mailman/MailList.py, line 1286, in ConfirmUnsubscription 178 admin(28713): cookie = self.pend_new(Pending.UNSUBSCRIPTION, addr) 179 admin(28713): File /usr/lib/mailman/Mailman/Pending.py, line 65, in pend_new 180 admin(28713): db = self.__load() 181 admin(28713): File /usr/lib/mailman/Mailman/Pending.py, line 95, in __load 182 admin(28713): return cPickle.load(fp) 183 admin(28713): AttributeError: 'module' object has no attribute 'UserDesc' 184 admin(28713): [- Python Information -] 185 admin(28713): sys.version = 2.4.3 (#1, Sep 17 2008, 16:07:08) 186 [GCC 4.1.2 20071124 (Red Hat 4.1.2-41)] 187 admin(28713): sys.executable = /usr/bin/python 188 admin(28713): sys.prefix = /usr 189 admin(28713): sys.exec_prefix = /usr 190 admin(28713): sys.path= /usr 191 admin(28713): sys.platform= linux2 This specific error has nothing to do with bounce processing. It would have resulted in a we hit a bug message when someone tried to unsubscribe from the options login page for some list. It indicates that that list's lists/LISTNAME/pending.pck file is corrupt. This could also cause errors in bounce processing for that list which should also be logged. Try the following for each list (assuming normal RedHat paths) #! /bin/sh for list in `/usr/lib/mailman/bin/list_lists --bare`; do echo $list /usr/lib/mailman/bin/dumpdb /var/lib/mailman/lists/$list/pending.pck done If the attempt to dump the pending.pck for a list throws an exception with a traceback, rather than dumping the contents, that list's pending.pck is corrupt. Remove it, and a new one will be created when needed. This may fix bounce processing for any list with a corrupt pending.pck. Also note the following: The bounce log messages like: Sep 25 18:43:55 2009 (5775) test: city at domain.gr already scored a bounce for date 25-Sep-2009 are normal. Once a bounce is recorded for a particular user on a particular list, all subsequent bounces for that user/list on the same calendar day are ignored. This also means that if the threshold is, e.g., 5 and the user's score is 2 (from a bounce on this day), lowering the threshold to 2 or less will; have no effect for that user until a bounce is received on a subsequent day. -- Mark Sapiro m...@msapiro.netThe highway is for gamblers, San Francisco Bay Area, Californiabetter use your sense - B. Dylan -- 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
Re: [Mailman-Users] Bounce Handling Question
Jeff Grossman wrote: I am running Mailman for two announcement only mailing list. It is running with no problems. I wanted to send out an announcement to all of my members without them getting a duplicate if they were on both lists. I created a new list called all yesterday. I put the two lists under the regular_include_lists setting. It ran perfect. But, I was wondering, how does the bounce handling work for that list? If a user bounces a message when I send to the all list, do they get removed from the other list they are actually on? The all list does not have any members associated with it. Hmmm... You raise an interesting issue. The short answer is No. The message is sent with envelope from all-bounces... and the bounce is returned there and will be processed only for members of the all list, i.e. no one. So the issue is, is it worth enhancing bounce processing so that when a bounce is returned for a non-member of list X, we look at list X's regular_include_lists and score the bounce for any of those lists of which the bouncing address is a member? Thoughts anyone? It seems that in the situation above where there are (I think) two active lists that get their own posts, there would be enough regular bounces so that ignoring the additional bounces from the all list wouldn't be a problem, but there may be other situations where a list is used as an alternative umbrella where almost all the posts would be to the 'umbrella' and this would be a problem. I added a note about this to the FAQ at http://wiki.list.org/x/TIA9. -- Mark Sapiro m...@msapiro.netThe highway is for gamblers, San Francisco Bay Area, Californiabetter use your sense - B. Dylan -- 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
Re: [Mailman-Users] remove_members program
Jeff Grossman wrote: Is there a way to either use the remove_members program of unsubscribe a member from all lists using the web? I would prefer this instead of having to connect to the server via ssh each time I want to use it. There is currently no global web unsubscribe. What I do is allow the web server to run certain Mailman commands. I have fairly tight control of the web site, and only allow it to run find_member, list_members and a local invite_members command. You could do something similar. The way it works is I have a compiled wrapper in Mailman's bin/ directory. This wrapper is owned and executable by the web server and is group mailman and SETGID. The wrapper itself is invoked with an argument string which is simply the command line you want it to execute. It checks if the command is allowed and if so, invokes it. This wrapper is called by some php web applications. You could make a wrapper to fit your needs and make a php script with whatever security you want to receive some post data and invoke the command. However, you probably don't want to do this if untrusted users can create web pages on the server. I have attached a sample wrapper to the FAQ at http://wiki.list.org/x/uIA9. -- Mark Sapiro m...@msapiro.netThe highway is for gamblers, San Francisco Bay Area, Californiabetter use your sense - B. Dylan -- 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
[Mailman-Users] mmfold.py with MM 2.1.12
After having no luck finding Skip Montanaro's revised (as of late 2008) mmfold.py script on his website (appears to be unreachable at http://www.webfast.com/~skip/python/mmfold.py) I did find a version at http://smontanaro.dyndns.org/python/ mmfold.py . but it returns the following error with Mailman 2.1.12: r...@soho89-16-227-18 20:03 pts/1 /var/www/mailman/bin# ./mmfold.py http://path.to.mailman.list.admin Traceback (most recent call last): File ./mmfold.py, line 181, in ? sys.exit(main()) File ./mmfold.py, line 64, in main lst = x.path.split(/)[-1] My goal in using it is to have more command-line options in admin of a mailman list. Does anyone have the most recent version, or perhaps a clue why the version I have, as retrieved at http://smontanaro.dyndns.org/python/mmfold.py fails? thanks in advance. -- Rob Lingelbach r...@colorist.org -- 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
Re: [Mailman-Users] mmfold.py with MM 2.1.12
Rob Lingelbach wrote: I did find a version at http://smontanaro.dyndns.org/python/ mmfold.py . but it returns the following error with Mailman 2.1.12: r...@soho89-16-227-18 20:03 pts/1 /var/www/mailman/bin# ./mmfold.py http://path.to.mailman.list.admin Traceback (most recent call last): File ./mmfold.py, line 181, in ? sys.exit(main()) File ./mmfold.py, line 64, in main lst = x.path.split(/)[-1] Did it say what the exception was? My goal in using it is to have more command-line options in admin of a mailman list. Does anyone have the most recent version, or perhaps a clue why the version I have, as retrieved at http://smontanaro.dyndns.org/python/mmfold.py fails? It works for me with both Python 2.5.2 and 2.6.1. I did make one change at line 173 I dropped a slash so it reads webbrowser.open(file:// + htmlfile) Links didn't like file:tmp/... as a URL. Also if you supply the password in the command line, you probably need to escape the '?' for the shell as in $ ./mmfold.py http://www.grizz.org/mailman/admindb/list\?admpw=... -- Mark Sapiro m...@msapiro.netThe highway is for gamblers, San Francisco Bay Area, Californiabetter use your sense - B. Dylan -- 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
Re: [Mailman-Users] mmfold.py with MM 2.1.12
Rob Lingelbach wrote: On Sep 26, 2009, at 5:04 PM, Mark Sapiro wrote: webbrowser.open(file:// + htmlfile) Links didn't like file:tmp/... as a URL. Also if you supply the password in the command line, you probably need to escape the '?' for the shell as in $ ./mmfold.py http://www.grizz.org/mailman/admindb/list\?admpw=... did do that, escaping the ? from the first attempt. but now, That's all for stuff that comes later. This problem r...@soho89-16-227-18 21:10 pts/1 /var/www/mailman/bin# ./mmfold.py -p (password) http://tig.colorist.org/mailman/admindb/tig Traceback (most recent call last): File ./mmfold.py, line 181, in ? sys.exit(main()) File ./mmfold.py, line 64, in main lst = x.path.split(/)[-1] AttributeError: 'tuple' object has no attribute 'path' is because x is the return from urlparse.urlparse() which prior to Python 2.5 was a straight tuple rather than a subclass of tuple with attributes. You either need to run this with Python 2.5 or later or you need to modify mmfold.py replacing lst = x.path.split(/)[-1] with lst = x[2].split(/)[-1] for keyval in x.query.split(): with for keyval in x[4].split(): and url = urlparse.urlunparse((x.scheme, x.netloc, x.path, x.params, query, x.fragment)) with url = urlparse.urlunparse((x[0], x[1], x[2], x[3], query, x[5])) -- Mark Sapiro m...@msapiro.netThe highway is for gamblers, San Francisco Bay Area, Californiabetter use your sense - B. Dylan -- 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
Re: [Mailman-Users] bouce processing problem
Dimitrios wrote: Yes that was it. When I erased pending.pck , I submitted an e-mail to the list and everything worked smoothly. I got the warning and the account had been disabled. Why was pending.pck corrupt? Any ideas? No. If you still have the original, you could check it's last change time and see what else was going on at that time, but if it was corrupt because of a system failure while it was being updated, it's possible the timestamp wasn't updated. The basic possibilities are it got written by something that wasn't supposed to write to it or some process that was updating it got interrupted and wasn't allowed to finish. Beyond that, I can't begin to even guess. -- Mark Sapiro m...@msapiro.netThe highway is for gamblers, San Francisco Bay Area, Californiabetter use your sense - B. Dylan -- 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
Re: [Mailman-Users] Bounce Handling Question
On Sat, 26 Sep 2009 10:54:58 -0700, Mark Sapiro m...@msapiro.net wrote: Jeff Grossman wrote: I am running Mailman for two announcement only mailing list. It is running with no problems. I wanted to send out an announcement to all of my members without them getting a duplicate if they were on both lists. I created a new list called all yesterday. I put the two lists under the regular_include_lists setting. It ran perfect. But, I was wondering, how does the bounce handling work for that list? If a user bounces a message when I send to the all list, do they get removed from the other list they are actually on? The all list does not have any members associated with it. Hmmm... You raise an interesting issue. The short answer is No. The message is sent with envelope from all-bounces... and the bounce is returned there and will be processed only for members of the all list, i.e. no one. So the issue is, is it worth enhancing bounce processing so that when a bounce is returned for a non-member of list X, we look at list X's regular_include_lists and score the bounce for any of those lists of which the bouncing address is a member? Thoughts anyone? It seems that in the situation above where there are (I think) two active lists that get their own posts, there would be enough regular bounces so that ignoring the additional bounces from the all list wouldn't be a problem, but there may be other situations where a list is used as an alternative umbrella where almost all the posts would be to the 'umbrella' and this would be a problem. I added a note about this to the FAQ at http://wiki.list.org/x/TIA9. I would like to see the bounce processing enhanced to handle this situation. If I remember correctly I think I saw where version 3 will be able to have it set to handle bounces on a global basis instead of per list? If that is the case, it will be taken care of in version 3. I am not sure what the timetable is for version 3, but it might be worth just waiting. But, I would love to see it in version 2 if it is not that big of a change. Jeff -- 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
Re: [Mailman-Users] remove_members program
On Sat, 26 Sep 2009 11:56:52 -0700, Mark Sapiro m...@msapiro.net wrote: Jeff Grossman wrote: Is there a way to either use the remove_members program of unsubscribe a member from all lists using the web? I would prefer this instead of having to connect to the server via ssh each time I want to use it. There is currently no global web unsubscribe. What I do is allow the web server to run certain Mailman commands. I have fairly tight control of the web site, and only allow it to run find_member, list_members and a local invite_members command. You could do something similar. The way it works is I have a compiled wrapper in Mailman's bin/ directory. This wrapper is owned and executable by the web server and is group mailman and SETGID. The wrapper itself is invoked with an argument string which is simply the command line you want it to execute. It checks if the command is allowed and if so, invokes it. This wrapper is called by some php web applications. You could make a wrapper to fit your needs and make a php script with whatever security you want to receive some post data and invoke the command. However, you probably don't want to do this if untrusted users can create web pages on the server. I have attached a sample wrapper to the FAQ at http://wiki.list.org/x/uIA9. Mark, Thanks for the valuable information. Currently that whole process is a bit over my head. But, after a few days of reading I am sure I will have it figured out. Hopefully I can get started on learning how to do it tomorrow. Jeff -- 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
