Hi,
Has anyone had any luck configuring SmartList as moderated and the moderated list having a digest component?
Thanks,
--Paul T.
--
At 8:02 PM -0400 4/4/01, CueMan is rumored to have typed:
Has anyone had any luck configuring SmartList as moderated and the moderated list having a digest component?
Sure. But I don't handle things the "blessed" way at all; I have a slightly different approval scheme (password in the subject), I drop _all_ mail through the interactive list first (my slist user is a _real_ user, with its own .procmailrc, instead of just an owner for sendmail aliases), and only pipe a copy of the mail to the digest in the interactive's rc.local.s20 (before adding footers and whatnot) - that's the _only_ way for mail to get to the digest list, since _nothing_ else calls flist with the digested listname. If it wasn't approved for the interactive, it never gets so far as the rc.local.s20...QED.
(*sigh*) Like I said, completely different than the "blessed" method, and I know it sounds more difficult, but it works much better for me than the whole add-digest-list-user-to-interactive-dist-list technique. Allows me to set up anti-spam recipies globally for _all_ the lists, require the list address in either To: or Cc: header (no blind carbons accepted here), lets digested lists remain _true_ slaves of the interactive, and I can do a bunch of other stuff not possible the "normal" way.
Some say I'm a control freak... ;)
Charlie
Thanks. I guess I was concerned with what seems to be the default behavior of SmartList breaking the digest list when one puts the main list into moderated mode.
As you know, when the main list sends an email to the digest list, the digest list bounces a copy of that email back to the main list (I guess that is to make sure the main list gets a copy of an email if someone mistakenly sends their post to the digest list instead of the main list). That copy's Msg ID shows up as a duplicate of the original message and is dumped by the main list.
When the main list goes into moderated mode, the bounce copy needs the approval of the moderator and so is sent to the moderator for approval instead of dumping it.
I wonder if there is a better way to handle this.
Thanks,
--Paul T.
On Wed, 4 Apr 2001, Charlie Summers wrote:
At 8:02 PM -0400 4/4/01, CueMan is rumored to have typed:
Has anyone had any luck configuring SmartList as moderated and the moderated list having a digest component?
Sure. But I don't handle things the "blessed" way at all; I have a slightly different approval scheme (password in the subject), I drop _all_ mail through the interactive list first (my slist user is a _real_ user, with its own .procmailrc, instead of just an owner for sendmail aliases), and only pipe a copy of the mail to the digest in the interactive's rc.local.s20 (before adding footers and whatnot) - that's the _only_ way for mail to get to the digest list, since _nothing_ else calls flist with the digested listname. If it wasn't approved for the interactive, it never gets so far as the rc.local.s20...QED.
(*sigh*) Like I said, completely different than the "blessed" method, and I know it sounds more difficult, but it works much better for me than the whole add-digest-list-user-to-interactive-dist-list technique. Allows me to set up anti-spam recipies globally for _all_ the lists, require the list address in either To: or Cc: header (no blind carbons accepted here), lets digested lists remain _true_ slaves of the interactive, and I can do a bunch of other stuff not possible the "normal" way.
Some say I'm a control freak... ;)
Charlie
Smartlist mailing list Smartlist@lists.RWTH-Aachen.DE http://MailMan.RWTH-Aachen.DE/mailman/listinfo/smartlist
-- A little quote from Nader: "More Colombians die each year from American cigarettes than Americans from Colombian cocaine. Imagine the Colombian Government deciding to spray defoliants on our tobacco crops!"
At 9:16 AM -0400 4/5/01, CueMan is rumored to have typed:
(I guess that is to make sure the main list gets a copy of an email if someone mistakenly sends their post to the digest list instead of the main list)
That's one of the reasons I send any mail sent to the digest list's address to the interactive one instead. If I were using sendmail aliases (I ain't, but never mind that), they would look something like:
list-l: "|exec /path/to/flist list-l" list-l-request: "|exec /path/to/flist list-l-request" list-d: "|exec /path/to/flist list-l" <--------- list-d-request: "|exec /path/to/flist list-d-request"
Note that mail to list-d is piping to list-l. Clean, simple, and avoids mismailed problems entirely.* This _requires_ you to then use a recipe in rc.local.s20 to pipe a copy of the mail directly to flist list-d (since if you subscribe list-d to list-l, the post just gets sent back to list-l by sendmail). Of course, THIS is a good thing, since if you add a footer in rc.local.s20, you can make the pipe BEFORE you add it, eliminating all the nonsense people go through to _remove_ it in the digested list's rc.local.s00.
(Gee, this is all much easier than I made it sound, honest.)
When the main list goes into moderated mode, the bounce copy needs the approval of the moderator and so is sent to the moderator for approval instead of dumping it.
Er...um...I don't see this behavior at all. Since the approved message is sent to the digest and then returned to the interactive WITH THE SAME messageID, the interactive doesn't send it _back_ to the moderators for re-approval, at least on my system.
I send an approved message to the interactive, which pipes it directly to the digest; the digest _does_ return the message to the interactive (in my situation this is completely unnecessary but harmless), which eats the copy since it's "been here" before.
Maybe you have a different problem? Is anything mucking around with messageID header fields?
Charlie
* Ok, ok, not entirely. There _is_ one anomoly with this...any -request email sent to the digest's list address will be processed by the interactive, not the digest. It's rare, but it does happen, and requires the listmaster to pay a little attention to misdirected unsubscribe requests. This is more a failure of the way SmartList kludges digested lists as completely seperate mailing lists, though, and to me is a completely acceptable problem for all the advantages it allows. --cfs3
On Thu, 5 Apr 2001, Charlie Summers wrote:
That's one of the reasons I send any mail sent to the digest list's address to the interactive one instead. If I were using sendmail aliases (I ain't, but never mind that), they would look something like:
list-l: "|exec /path/to/flist list-l" list-l-request: "|exec /path/to/flist list-l-request" list-d: "|exec /path/to/flist list-l" <--------- list-d-request: "|exec /path/to/flist list-d-request"
Note that mail to list-d is piping to list-l. Clean, simple, and avoids mismailed problems entirely.* This _requires_ you to then use a recipe in rc.local.s20 to pipe a copy of the mail directly to flist list-d (since if you subscribe list-d to list-l, the post just gets sent back to list-l by
So in rc.local.s20, you pipe a copy of the mail directly to flist list-d like the following?:
:0 c | /home/lists/.bin/flist list-d
I tried that some time ago and had problems with it. Perhaps it was because I did this to several lists I was running and my own custome pipeings in my /etc/aliases file. But eventually, with the above scheme, a SL related procmail would get stuck, then unprocessed SL procmail processes would start backing up. Deleting the .etc/rc.lock file would break the logjam and everything would run fine for awhile again. So I stopped doing things that way.
Thanks,
--Paul T.
--
At 7:29 AM -0400 4/14/01, CueMan is rumored to have typed:
So in rc.local.s20, you pipe a copy of the mail directly to flist list-d like the following?:
Yup.
I tried that some time ago and had problems with it.
A couple of years ago, with an older version of SmartList (and on a shared machine that I didn't root), I had that happen _once,_ but I haven't seen that happen at all otherwise. (TIA, I have my sendmail aliases set up to send all mail to the real user slist, and then use slist's .procmailrc file to send to flist. I do _not_ recommend that anyone else do it that way, but it gives me more control, allows me to apply my body-filter anti-spam rules, ignores Bcc mail, and some other stuff. Maybe this is what's "protecting" me from hanging lock files?)
Can anyone else share any experiences of direct pipings to flist causing a problem? I mean, I don't know what the difference would be between piping from an alias file and piping from rc.local.s20, but then what I don't know fills libraries...
Charlie
On Sat, 14 Apr 2001, Charlie Summers wrote:
A couple of years ago, with an older version of SmartList (and on a shared machine that I didn't root), I had that happen _once,_ but I haven't seen
An older version of SL...like do you thing the multigram was having the trouble or do you think it was some recipe?
that happen at all otherwise. (TIA, I have my sendmail aliases set up to send all mail to the real user slist, and then use slist's .procmailrc file to send to flist. I do _not_ recommend that anyone else do it that way, but it
I do a similar thing so I can filter for viruses, works fine but I have trouble with hanging lock files when piping mail to digests from rc.local.s20. But then I am using an older version of SL that has been incrementally upgraded as needed.
Thanks,
--Paul T.
gives me more control, allows me to apply my body-filter anti-spam rules, ignores Bcc mail, and some other stuff. Maybe this is what's "protecting" me from hanging lock files?)
Can anyone else share any experiences of direct pipings to flist causing a problem? I mean, I don't know what the difference would be between piping from an alias file and piping from rc.local.s20, but then what I don't know fills libraries...
Charlie
Smartlist mailing list Smartlist@lists.RWTH-Aachen.DE http://MailMan.RWTH-Aachen.DE/mailman/listinfo/smartlist
-- A little quote from Nader: "More Colombians die each year from American cigarettes than Americans from Colombian cocaine. Imagine the Colombian Government deciding to spray defoliants on our tobacco crops!"
smartlist@lists.rwth-aachen.de
-
Charlie Summers -
CueMan