Developers of Sympa

[Jeff Abbott <address@concealed>]

Folks,

So one of the other guys here, who used to run a departmental Mailman  
instance, had an interesting thought on my problem of users sending  
messages, due to historical use of such addresses here at Duke, to  
$LISTNAME-owner rather than Sympa's preferred $LISTNAME-request when  
trying to contact the list owner.  The specific issue I've been  
taking claim with is that bouncequeue is dropping non-DSN messages,  
so a user would send a message to the list owner at $LISTNAME-owner  
and neither end would know the message was sent but not received and  
seen.

His suggestion was to have $LISTNAME-owner send to both bouncequeue  
and redirect to $LISTNAME-request.  For instance, for a list named  
testlist1:

  #------------------------------ testlist1: list alias created 27  
Oct 2005
  testlist1: "| /etc/smrsh/queue address@concealed"
  testlist1-request: "| /etc/smrsh/queue address@concealed"
  testlist1-editor: "| /etc/smrsh/queue address@concealed"
  #testlist1-subscribe: "| /etc/smrsh/queue testlist1- 
address@concealed"
  testlist1-unsubscribe: "| /etc/smrsh/queue testlist1- 
address@concealed"
  testlist1-owner: "| /etc/smrsh/bouncequeue address@concealed",  
testlist1-request

This appears to be working quite well as bouncequeue is just dropping  
the message if it's not a DSN, but since it's also being sent to  
testlist1-request then the list owner still sees it.

My question is:  What is the harm in this?  Is this a valid  
workaround for Duke to use to help ease the transition for our  
users?  I'm seeing a couple of extraneous log entries, namely:

  Oct 27 11:43:11 peake bounced[29321]: error: no address found in  
message from address@concealed for list testlist1

when a non-bounce message is sent to testlist1-owner, and:

  Oct 27 11:46:13 peake sympa[29307]: Ignoring message which would  
cause a loop, sent by address@concealed

when the message is a bounce.  This would meet my needs assuming it's  
not a completely bad idea, so I'd like to hear from the people who  
know:  Is it a bad idea?

Thanks,
Jeff

Attachment: smime.p7s
Description: S/MIME cryptographic signature