spf-discuss
[Top] [All Lists]

Re: SRS in .forward and reversing in /etc/aliases

2004-02-18 13:32:20
On Wed, Feb 18, 2004 at 09:25:32AM +0000, Shevek wrote:
On Tue, 17 Feb 2004, Meng Weng Wong wrote:

On Tue, Feb 17, 2004 at 09:26:57AM -0600, 
mw-list-spf-discuss(_at_)csi(_dot_)hu wrote:
| But why not focus on the really fundamental problems:  
| 
| 1) How do you propose to handle users' forwarding from .forward or
|    .qmail files (as opposed to relaying)?

.forward before: final(_at_)destination(_dot_)com
.forward after:  "|/usr/bin/srs --secret=/etc/srs.secret 
final(_at_)destination(_dot_)com"

| 2) How do you propose to handle SRS in case of multiple incoming and
|    outgoing servers?

Each server has to know the SRS secret(s).

/etc/aliases:

srs0:     "|/usr/bin/srs -reverse --secret=/etc/srs.secret"
srs1:     "|/usr/bin/srs -reverse --secret=/etc/srs.secret"

The above proposed "srs" syntax is just one possible form.

I have taken this email as an approximate specification, and have included 
such a script in version 0.23 of Mail::SRS. I am eager to include such 
glue in the standard distribution, since it helps the less technical 
implementors and keeps everyone speaking the same language for support 
questions.

The above /etc/alias lines seem to imply that not only envelope
address rewriting happens but forwarding as well.  It seems that your
script, on the other hand, just takes an address as an argument, and
transforms it.  Another problem with your script is that it seems to
exit with 0 even if the address does not check out.  (Bug: except in
the first place, the separator is always taken to be `=')

Do the above /etc/alias lines handle addresses of the form
srs0-blah...?

Do the above /etc/alias lines assume that srs is doing some checking
as well (like if this incoming message is really a bounce)?  If not,
where and how is it checked that the message is a bounce?  (Otherwise,
it seems I can just use the address to send spam back to the original
sender.)

Mate

-- 
---
Mate Wierdl | Dept. of Math. Sciences | University of Memphis  
Please avoid sending me Word or PowerPoint attachments.
See http://www.fsf.org/philosophy/no-word-attachments.html