> From: MX%"email@example.com" 17-FEB-1996 09:28:05.11
> Subj: Re: Horror Story - is there a defense?
> Listproc 6.0 (and probably later versions) do a few things to stop
> your problem of duplicate messages going to a list:
> - it keeps the last 500 (I think) Message-IDs sent to each list, and
> also sent to listproc itself (administrative requests). If a new
> message comes in with the same Message-ID as a previous message,
> listproc routes the message to the list owner or system manager.
> But some systems give the same Message-ID to more than one message
> (yes, this is wrong) and listproc rejects those too... I deal with
> that on a user-by-user basis; we have one user with that problem.
> - it computes a checksum of each message body sent to a list.
> When a new message comes in, it compares the checksum to the last 500
> checksums; if there's a match, the message goes to the list owner.
> By default, listproc uses the sum(1) utility, which sometimes gives
> the same checksum for different message bodies. I hacked listproc
> to use the MD5 algorithm; we haven't had any problems since.
> You could probably add code like that to your Majordomo. The checksum
> would probably be enough. I haven't kept up on the lasest Majordomo
> features; there are probably others like this in Majordomo already.
> --Jerry, firstname.lastname@example.org
This would assume, of course, that the rogue mailer kept
generating the same message-id for each message. If not, then there
would still be trouble...