[Mailman-Users] Posting-Filter is broken in 1.1 !?

[Chris Fox]

On Mon, 03 Jan 2000 09:08:17 PST, J C Lawrence wrote:
>On Mon, 3 Jan 2000 10:30:03 +0100 
>Aaron Optimizer Digulla <digulla at hepe.com> wrote:
>
>> I have reported these bugs several *months* ago and still got no
>> feedback about them :-( Is anyone ever fixing bugs ?
>
>> Ok, here they are again:
>
>> 1. Some posts are held no matter if the poster is subscribed or in
>> the posters list (this is a different bug than #2 !!).
>
>I suspect that you still have MailMan configured to authenticate by
>envelope address and that that address is not in your membership
>list or authorised_poster list.
>
>> 2. Some posts are held because Mailman uses the "Sender:" header
>> item added by Netscape (Mailman should ignore it).
>
>MailMan pays no attention to the Sender: header.  It does by default
>read the message envelope and does ignore the From: header however,
>
>> 3. When I have a list with the policy "Only subscribed posters can
>> post" set to "Yes" and I set this option to "No", then all posts
>> must be approved manually ! No one can post anymore :-(
>
>Set "Must posts be approved by an administrator?" to "No" in Privacy 
>Options.


I also experience the reported problem, and requested help on this 
list--subject: How to make Mailman more trusting of subscriber 
supplied addresses--with no response (at least, none that reached 
me).  This is with USE_ENVELOPE_SENDER = 0, and it even when faking 
the envelope from by piping through formail.  At times, Mailman seems 
to behave as if it uses the return-path: header.

Am I correct in assuming that the existence of this bug is either 
being denied or ignored?  Although in our application we _may_ be 
able to find some workaround using an Exim recipe, we might have to 
reconsider if Mailman is going to be appropriate for our needs if 
such problems are not to be addressed.

If it really  is'nt a bug in Mailman itself, I wonder whether the 
problem arises due to some combination of mailman, MUA and MTA.  What 
evidence would be required to track it down (and convince people that 
it exists) and to fix it?

Chris


-- 
Dr. Chris Fox,  Dept. of Mathematical & Computing Sciences, Goldsmiths
College, New Cross, London SE14 6NW, UK.       Email: C.Fox at gold.ac.uk
http://homepages.gold.ac.uk/chris/    Tel/Fax: +44(0)20 7919 7856/7853
GnuPG ID 6026D072 (09DF C4DF EF12 B96A 7EDD  27D8 1C0C DC33 6026 D072)