Date: Thu, 17 Jul 2008 08:45:46 -0400
From: Andrew Sullivan <ajs@commandprompt.com>
To: pgsql-www@postgresql.org
Subject: Re: Spam filtering on the mailing lists
Message-ID: <20080717124546.GA6732@commandprompt.com>
References: <200807161509.m6GF9RO14179@momjian.us>
	<1216227673.28961.139.camel@jd-laptop>
	<20080716180738.GF5758@commandprompt.com>
	<8C1E1C3655705FB9706E2E10@ganymede.hub.org>
	<361E31A33E6E19B7844FA853@ganymede.hub.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <361E31A33E6E19B7844FA853@ganymede.hub.org>
User-Agent: Mutt/1.5.17 (2007-11-01)

On Thu, Jul 17, 2008 at 12:01:08AM -0300, Marc G. Fournier wrote:

> Does anyone know if there is some 'rule' that return mail about 'Undeliverable' 
> has to be in English?  

Nope, there isn't.  The bounce message is intended for human
consumption, and therefore localisation is approproate. 

Note, however, that many of the messages should have a Content-Type:
multipart/report; report-type=delivery-status; header.  See RFC 3462,
which obsoletes RFC 1892.

RFC 3462 requires a machine-parsable delivery report, and one such
report is defined in RFC 3464.  Many mail systems (most?) conform to
this these days, so just detecting thatt here is a delivery-status
header, then parsing the second body-part for "^Action: failed" oughta
work.  The message/delivery-status content-type MUST be 7 bit, so you
can be relatively certain it's not going to be anything but ASCII.
(AFAIK, there are no machines left on the Internet that leak 7-bit
charsets that are not ASCII).

A

-- 
Andrew Sullivan
ajs@commandprompt.com
+1 503 667 4564 x104
http://www.commandprompt.com/