Barry Finkel wrote: > > 3) When is the list-creation e-mail sent to the one owner e-mail > address? It appears to be done in the "newlist" code, but > at that time, I have not yet run the "config_list" command > that resets the parameters. When I ran my test, was I lucky > that by the time that the list-creation mail had been sent > to Postfix, the "config_list" command had finished re-populating > the owner field?
bin/newlist "sent" the mail to the specified owner address before it returned. Here's what happens in detail: 1) bin/newlist creates the mail with list-ow...@... as the recipient and queues it in the virgin queue. This is it's last act before exiting. 2) VirginRunner picks up the queue entry processes it and puts it in the out queue. 3) OutgoingRunner picks up the entry from the out queue and delivers it to the MTA with RCPT TO list-ow...@... 4) The MTA pipes the message to the wrapper which calls the scripts/owner script to queue the message "toowner" in the in queue. 5) IncomingRunner picks up the entry from the in queue, locks the list, builds the recipient list from the list's owner attribute, queues the message in the out queue and unlocks the list. 6) OutgoingRunner picks up the message and delivers it to the MTA with RCPT TO commands for the recipient list built in step 5. So if your config_list runs and gets the list lock before IncomingRunner gets it at step 5, the mail goes to your config_list owners. Otherwise it will be sent to LIST-owner and we are back to step 4. This is a loop. Loops of normal list posts are detected because Mailman adds an X-BeenThere: header to outgoing posts and detects it in an incoming post to detect a loop. In this case, we don't detect loops of the -owner email. There is no protection against this loop and it continues until sooner or later config_list or something changes the owner attribute to break the loop. I'm not actually clear if the initial owner you set with newlist was that list's -owner address, but if it was, you might be able to tell from your mail logs if the mail looped more than once. I actually think it probably didn't, because the expected delays in the runners waking up at steps 2, 3 and 5 total over 1 second, and the shell script should get to the next line way before that. On the other hand, if this config_list is not going to run immediately, I think there may be nothing to break the loop which is a problem. Also, I think I need to do something to detect and break this loop. >I know that there has been talk in the past weeks about list-creation >templates. I would like to be able to specify multiple e-mail >addresses at the list-creation prompt so that all of the list owners >would get the mail. I do not want to rely on specifying one list owner >and then insuring that he or she forwards a copy to the other >co-owners. The list creation templates are a Mailman 3 feature. At this point, it's up to Barry W. to speak to that. One thing you might consider is using the --quiet newlist option when you create the list and then have your script send a notice to LIST-owner after config_list runs. -- Mark Sapiro <m...@msapiro.net> The highway is for gamblers, San Francisco Bay Area, California better 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
