[Mailman-Users] query re "message has implicit destination"

Bretton Vine bretton at hivemind.net
Wed Aug 30 20:12:32 CEST 2006

Setup: Debian AMD64 (mix testing/stable), Exim 4.62, Mailman 2.1.8 from
       source (with patch to allow @listname in allowed posters), multiple
       installations of Mailman (one instance of MM per virtual domain)

In the last 24 hours we've had the same situation occur with two corporate

[case 1]
List-member (and host of an INX) sent notification to a network-notices list
with a CC to one of their employees. The message was held for administrative
approval with the error: "message has implicit destination"
 + they are a member, have allowance to post to list
 + they claim no BCC in their message
 + virtualdomain1

[case 2]
List-member sends message to listname at virtualdomain2 and BCC's
listname2 at virtualdomain2. The message was held for administrative approval
with the error: "message has implicit destination"
 + they are a member, have allowance to post to list
 + they indicated BCC to second list in follow-up query, problem resolved
   and message approved for second list

Now in [case 1] we are being held accountable for the 24 hour delay in
responding by authorising the post. This is a critical matter as it impacts
on local Internet infrastructure, but due to the number of lists and
virtualdomains it's simply impossible to have immediate notification for
/every/ administrative request for every list. A once-a-day notification is
one of many as it is and I (or my boss) spend up to an hour a day each
morning just dealing with (mostly spam, non-member) administrative requests.

Now I've already notified client of the reasons (with links to FAQ etc) over
why the "message has implicit destination" error occurs, and switched it off
for that list in particular. What I need though is a concise and accurate
answer as to why this is the default setting in a Mailman installation.

I understand the reason, but I need a non-technical, max 4 line rationale.
Failing that I've been instructed to switch it off for /all/ the lists we
host for corporate clients. Mind you that instruction stands regardless.

What confuses me though is that in [case 1] the list-member claims to have
sent the message as:

 TO: list at virtualdomain1
 CC: some at other.address

In [case 2] it was clearly:

 TO:  list1 at virtualdomain2
 BCC: list2 at virtualdomain2

But both instances resulted in the same administrative delay.

So I'm a little confused. In terms of the info I can find online, there is
no reason [case 1] should have resulted in an error /unless/ the list member
had also included some BCC: addresses, which they claim not to have done. In
[case 2] the error was quickly explained and rectified and the client happy.
Obviously in [case 1] it's a slightly different story as due to the delay in
notification/approval our national peering infrastructure had problems and
someone needs to be held accountable -- in this case me :-)

i.e. What happened, don't let it happen again.

Now when I test the following I /don't/ get the error.

 TO: list1 at virtualdomain3
 CC: list2 at virtualdomain3

Which brings to mind an obvious possibility, but this being what it is that
isn't one that can be entertained. (heads must roll! <chuckles>)

The configuration for all MM virtual domains is the same (barring domain
specifics) and the configuration is highly tweaked to our environment based
on online documentation relating to Mailman and Exim.

Can anyone shed some light (and yes, I've googled and gone through numerous
FAQ answers which have bought me some time but not a reason why the default
installation behaviour is as it is for BCC to lists etc)

| Bretton Vine | 083 633 8475 | bretton at hivemind.net |
| GPG: http://bretton.hivemind.net/bretton_vine.asc  |

"All men who have turned out worth anything have had the chief hand in their
own education." - Sir Walter Scott

More information about the Mailman-Users mailing list