Remailer .sig suppressio

Hal 74076.1041 at CompuServe.COM
Fri Jan 1 22:23:32 PST 1993


-----BEGIN PGP SIGNED MESSAGE-----

From: edgar at spectrx.Saigon.COM (Edgar W. Swank)

> Hugh Daniels said here on Dec 28:
>  
>     There are very good reasons to build remailers (and all mail
>     tools) to pass on all the bytes they can, trailing spaces and
>     .sigs included.
>  
> Hugh doesn't say what these reasons are.  They are not obvious to me,
> so I must disagree. I've already stated what I think are good reasons
> at least for remailers whose purpose is anonymity to remove automatic
> sigs which are likley to destroy anonymity.
>  
> I've said I would accept either a less ambiguous sig delimiter than
> "--" or a remailer option to remove the sig (default) or leave it in.

I'll just relate one story that happened to me today.  I wanted to
try an experiment in which I would use two non-cypherpunks remailers
to set up a chained anonymous address.  One is anon.penet.fi, which
doesn't do any encryption, but which will allow you to specify an
arbitrary destination address.  The other is pax.tpa.com.au, which
does PGP decryption (but you can't encrypt the remailer destination
address like you can with our remailers).

The Pax remailer lets you send them a PGP key which it saves.  Then,
any future messages to you are encrypted by the remailer using that
key.  That way message contents are always protected between Pax and
you.

I wanted to send Pax a key via the Penet remailer so that Pax wouldn't
know who I really was.  I tried this, and got a message back from
Pax saying:

Error: you didn't include a public key for us !
So we can't assign an alias or send you our public key.

But I _had_ sent them a public key.  After some head-scratching
I figured out the answer.  My public key had started with the string:
"-----BEGIN PGP MESSAGE-----".  But the Penet remailer strips sigs, which
it considers to be any line starting with "--".  It thought my PGP key
was a signature!  It had stripped it, so that Pax received only a
blank message.

I haven't thought of a way around this problem yet.

Now, Edgar may take as the moral of the story that remailers should
have smarter sig recognition.  But I take the moral to be that munging
mail messages may cause problems when people try to use it for something
which you didn't anticipate.

Hal

-----BEGIN PGP SIGNATURE-----
Version: 2.1

iQCVAgUBK0UJOqgTA69YIUw3AQHusQP/YuzvntMZ8XPpfLwwE5YElLjwfKGF0Q2e
Cjk1PMmvtcn/bjSCB49lagOs0cEjm9Vt4gsEkTxwVlOya0+WOTeY/zzZAYlf3z4R
9QY7uRSyPQYJlPH6rosifEREMNWksRMCNMlISp8PDh1duJf3BvdwY3nyXk/PABpS
LTp6NAFaFi4=
=j0Wl
-----END PGP SIGNATURE-----


Distribution:
  CYPHERPUNKS >INTERNET:CYPHERPUNKS at TOAD.COM







More information about the cypherpunks-legacy mailing list