[Mailman-Users] Problems with mailman header regexps

[Russell King]

Hi,

Please CC: me with any responses; I'm not subscribed to these mailman
lists.

I'm desperately trying to get mailman to filter HTML-encoded email
from my mailing lists.  The problem is no matter what I try it still
gets through.

It appears that some of my regexps work, but some don't; this has been
annoying me for some time now, and I'm now at the point of considering
dropping mailman unless a solution to this problem can be found.

The message in question had the following headers:

> Reply-To: guisee9 at lycos.com
> From: ³×Æ®¿öÅ© <guisee9 at lycos.com>
> To:  <linux-arm-kernel at lists.arm.linux.org.uk>
> Subject: [±¤°í]¼Õ½¬¿î »ç¾÷¹× ºÎ¾÷À» ¼Ò°³ÇÕ´Ï´Ù
> Mime-Version: 1.0
> Content-Type: text/html; charset="ks_c_5601-1987"
> X-User: 2.62-figjdliv-lilkhn-Ddigq
> Message-Id: <E16bPBd-0004Xc-00 at www.linux.org.uk>

We've had all of the following in "bounce matching headers":

Content-type: \(text\/html\|multipart\/alternative\)
Content-type: \(text/html\|multipart/alternative\)
Content-type \(text/html\|multipart/alternative\)

(the last is definitely wrong; mailman complains in the config log).
None of these have worked.  The full regexp list is currently:

Subject: \(xxx\|ndn:\|RCPT:\|Delivery Confirmation\|NON-DELIVERY of\|Undeliverable message\|Receipt Confirmation\|\(Failed\|Returned\|unable to deliver\|away from my\) mail\|autoreply\|millionaire\|money\|ADV\|=?ISO-20220JP?\|Snowhite and the Seven Dwarfs\|China-diesel\|invitation\|really works for me\|DVD Movies\|Tax Free\|Rated\|Credit\|\$\|MobiCam\)
Content-type: \(text\/html\|multipart\/alternative\)
Date: ..\s...\s..\s.*:..:..\s..\s*$
Date: .*=\?.*\?=
X-Mailer: eMerge
X-EM-Registration: .*
X-Mailman-Version: .*
To: linux-\(arm\|arm-announce\)\@lists.arm.linux.org.uk
CC: linux-\(arm\|arm-announce\)\@lists.arm.linux.org.uk

Note that the last two definitely have worked.

This is mailman 2.0.8.

Also, I'm pretty concerned that mailman developers don't take bug reports
seriously.  I have a bug report in the system against mailman that has sat
there since April 7th 2001 - 414502.  The nature of this bug is such that
mail content can cause mailman's queue runner to stop processing mail
(because it dies each time it hits this particular message).  If mail
content can crash mailman, maybe mail content can be used to exploit
mailman?  Well, I've had nothing back, apart from soneone else confirming
the problem.

Does anyone actually process any mailman bug reports?

-- 
Russell King (rmk at arm.linux.org.uk)                The developer of ARM Linux
             http://www.arm.linux.org.uk/personal/aboutme.html