You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Matt Kettler <mk...@verizon.net> on 2006/12/15 02:03:33 UTC
SURBL scored stronger than normal on the apache servers?
postmaster@verizon.net wrote:
> This report relates to a message you sent with the following header fields:
>
> Message-id: <45...@verizon.net>
> Date: Thu, 14 Dec 2006 11:37:35 -0500
> From: Matt Kettler <mk...@verizon.net>
> To: kquillen@wifi7.com
> Subject: Re: installing URIDNSBL
>
> Your message cannot be delivered to the following recipients:
>
> Recipient address: users@spamassassin.apache.org
> Reason: SMTP transmission failure has occurred
> Diagnostic code: smtp;552 spam score (21.0) exceeded threshold
> Remote system: dns;herse.apache.org (TCP|206.46.252.46|57572|140.211.11.133|25) (apache.org ESMTP qpsmtpd 0.29 ready; send us your mail, but not your spam.)
>
>
> ------------------------------------------------------------------------
<snip> email containing the surbl permanent test point, and no spam quotes.
The test-point URL used to only be listed in SC, although tests at
uribl.com and rulesemporium.com both just report it as listed as a "test
point" and don't list out any SURBL sub-lists it belongs to. ...
So has apache.org jumped up their score, or is there some change in the
listing here that's causing SA deployments to go nuts on this test point?
21 points seems absolutely *absurd* for just SC, or any test point.
(Actually 21 seems a little bit out-of-whack for any combination of
rules all looking at the same small attribute of the email, no matter
how strong a spam sign it is, except perhaps an end-user configured
explicit blacklist.)
Re: SURBL scored stronger than normal on the apache servers?
Posted by Matt Kettler <mk...@verizon.net>.
Jeff Chan wrote:
> The test points were changed from returning a value of 127.0.0.2
> to 127.0.0.126 as of about a year ago. I neglected to announce
> the change, though it was mentioned on the SURBL discussion list.
> Announcing now.
>
> 127.0.0.126 represents all ones for the bits of all existing
> lists. This was a suggestion from one of the SpamAssassin
> developers.
That at least explains the change.. It's also in some ways sensible in
that you can quickly see that all the SURBL checks are enabled.
Unfortunately it also means we can't even post messages to the list
containing the test point, which isn't so good, but I guess the same
goes for the GTUBE.
Re: SURBL scored stronger than normal on the apache servers?
Posted by Jeff Chan <je...@surbl.org>.
On Thursday, December 14, 2006, 5:03:33 PM, Matt Kettler wrote:
> <snip> email containing the surbl permanent test point, and no spam quotes.
> The test-point URL used to only be listed in SC, although tests at
> uribl.com and rulesemporium.com both just report it as listed as a "test
> point" and don't list out any SURBL sub-lists it belongs to. ...
> So has apache.org jumped up their score, or is there some change in the
> listing here that's causing SA deployments to go nuts on this test point?
> 21 points seems absolutely *absurd* for just SC, or any test point.
> (Actually 21 seems a little bit out-of-whack for any combination of
> rules all looking at the same small attribute of the email, no matter
> how strong a spam sign it is, except perhaps an end-user configured
> explicit blacklist.)
The test points were changed from returning a value of 127.0.0.2
to 127.0.0.126 as of about a year ago. I neglected to announce
the change, though it was mentioned on the SURBL discussion list.
Announcing now.
127.0.0.126 represents all ones for the bits of all existing
lists. This was a suggestion from one of the SpamAssassin
developers.
Jeff C.
--
Jeff Chan
mailto:jeffc@surbl.org
http://www.surbl.org/