[Spambayes] why doesn't the pop3 proxy filter all messages?

The Rolphs rolph at 110.net
Fri Nov 26 04:31:57 CET 2004

I can confirm that sending the CAPA causes the session to fail to close 
reliably from a telnet session through the proxy.  Not using the proxy 
results in correct receipt of mail by the client.  Using the proxy results 
in all mail being received, but many of the messages do not have the 
X-Spambayes headers.

what I think would resolve this is for the proxy to send an extra <CR><LF> 
after the QUIT command to ensure that the session has closed.

I have had two of these cases after changing the  check forr SSL setting, 
but the problem is greatly reduced.

At 03:26 PM 11/26/2004 +1300, Tony Meyer wrote:
> > Ok, it looks like the problem is the CAPA command from
> > eudora.  It looks like this is confusing the proxy server.
>This should not be the case.  CAPA should work, although it should strip out
>PIPELINING and STLS if they are present in the server's response.  You
>should be able to telnet to the proxy and give it a "CAPA" command and get a
>response.  Does that work?
> > My hypothesis is that if I can get Eudora to not send the
> > CAPA command, that we should work better.  Does this sound
> > reasonable?
>If there is a problem with the CAPA command (it's possible, although
>unlikely, since I can't see how this wouldn't have hit many people already),
>then it's a SpamBayes problem, so we need to fix something.
>[Next message]
> > Ok, when I set never for secure sockets when receiving
> > options, then the CAPA is not sent and it looks like
> > things work.
>I presume you weren't actually using SSL (TLS) in any case, since that
>wouldn't have worked through the proxy.
>I still don't understand:
>   (a) Why CAPA is causing problems
>   (b) How mail can be delivered if the proxy dies after a CAPA (is there a
>second, direct, connection to the mail server?)
>Please always include the list (spambayes at python.org) in your replies
>(reply-all), and please don't send me personal mail about SpamBayes.
>http://www.massey.ac.nz/~tameyer/writing/reply_all.html explains this.


The Rolphs rolph at 110.net

More information about the Spambayes mailing list