[Mimedefang] AD: Roaring Penguin press release on Locked Addresses

Jan Pieter Cornet johnpc at xs4all.nl
Tue Oct 18 15:40:36 EDT 2005

On Tue, Oct 18, 2005 at 10:01:18AM -0700, Matthew.van.Eerde at hbinc.com wrote:
> David F. Skoll wrote:
> > Matthew.van.Eerde at hbinc.com wrote:
> >> ... for example, could I
> >> subscribe to a mailing list with the unique address and expect that
> >> "unique sender" algorithm will be able to uniquely identify list
> >> mail?
> > 
> > It depends on the mailing list software.  I know it works fine with
> > mailman.  It will fail for mailing list software that changes the
> > envelope sender with each posting, in which case the best you can do
> > is lock to the domain instead of the complete sender address.
> OK, so for this list the email address would be locked to sender:
> 	mimedefang-bounces at lists.roaringpenguin.com
> But for the SpamAssassin list (for example), the senders are all different:
> 	All the below are @spamassassin.apache.org:
> 	users-return-33376-Matthew.van.Eerde=hbinc.com
> 	users-return-33377-Matthew.van.Eerde=hbinc.com
> 	users-return-33378-Matthew.van.Eerde=hbinc.com
> 	...
> So the best I could do is lock to @spamassassin.apache.org

If you want more options, you might want to look at the options in
TMDA http://tmda.net/ . What David's invented is kind of a TMDA
"sender locked" email address, except that the sender is only locked
upon first arrival of an email. Clever idea, thanks for sharing.

David: your proposal will use the envelope return-path, which means
that it will fall down in case of VERP <http://cr.yp.to/proto/verp.txt>
which is the case Matthew noticed, and in case of BATV

VERP is only useful for mailinglists, and I haven't seen it anywhere
else. Unfortunately, there are a few standards to choose from here,
so it isn't always possible to extract the original address. Using
the domain, as you said, is probably best.

BATV will be somewhat more of a problem if it becomes widespread,
but from what it looks now you can anticipate this, and extract
the "original" address, or wildcard the variant bits.

#!perl -wpl # mmfppfmpmmpp mmpffm <pmmppfmfpppppfmmmf at fpffmm4mmmpmfpmf.ppppmf>
$_[2]}->(map{/p|f/i+/f/i}split//,$&)+97):qw(m p f)[map{((ord$&)%32-1)/$_%3}(9,
3,1)]),5,1)='`'lt$&;$f.eig;                                # Jan-Pieter Cornet

More information about the MIMEDefang mailing list