html-stripper-0.1 patch and problems with MIME multipart messages

[Toomas Aas <toomas . aas @ raad . tartu . ee>]

Hello!

I seem to have a problem with html-stripper-0.1 patch and some MIME 
multipart messages.

Let's say the original message looks someting like this:

==========================================================
From: sender <sender @
hot .
ee>
Date: Mon, 12 Mar 2007 19:49:30 +0200
To: test-l @
mydomain .
com
Subject: =?utf-8?q?p=C3=A4iste=20test?=
MIME-Version: 1.0
Content-Type: multipart/alternative;
    boundary="--------=BoundaryPyDog1173721770.89------"
Message-Id: <20070312174934 .
4E0517B43C @
mh3-4 .
hot .
ee>


----------=BoundaryPyDog1173721770.89------
Content-Type: text/plain;
    charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Mis paistab kirja p=C3=A4istes? (katse nr 1)
----------=BoundaryPyDog1173721770.89------
Content-Type: text/html;
    charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Mis paistab kirja p=C3=A4istes? (katse nr 1)
----------=BoundaryPyDog1173721770.89--------

==========================================================

So it's an multipart/alternative message with text/plain part specifying 
charset UTF-8.

After passing through Majordomo with html-stripper enabled and 
html_policy set to 'strip', the message becomes something like this:

==========================================================
From: <sender @
hot .
ee>
Date: Mon, 12 Mar 2007 19:49:30 +0200
To: test-l @
mydomain .
com
Subject: =?utf-8?q?p=C3=A4iste=20test?=
MIME-Version: 1.0
Content-Type: text/plain
Message-Id: <20070312175506 .
88428207DB @
mh3-5 .
hot .
ee>
Sender: owner-test-l @
post .
raad .
tartu .
ee
Precedence: bulk

Mis paistab kirja päistes? (katse nr 1)
==========================================================

As you see, the content-type in message headers has changed from 
multipart/alternative to text/plain, but the charset info has been 
completely lost. As a result of this, the message body contains raw 
8-bit characters which remain untranslated in MUA.

If I set the list's html_policy to 'pass', the message passes through 
with all it's parts intact and is displayed correctly to users.

I tried to look at the code of html-stripper-0.1 patch, but I'm not a 
programmer so only thing I got was a headache ;) Maybe someone has a fix?

Thanks in advance,
--
Toomas Aas