[Mailman-Developers] Does mailman-cvs remove reply-to?

[Marc MERLIN]

On Mon, Mar 11, 2002 at 01:46:33PM -0500, Barry A. Warsaw wrote:
>     MM> Basically, I'm saying that if I post to a list without
>     MM> reply-to munging, if I set (as a poster) a reply-to, it
>     MM> doesn't make it to the list.  (I just checked on 2 other
>     MM> machines where I have mailman-cvs installed)
> 
> I think it's a documentation bug in the description of
> first_strip_reply_to.  The intent is to give the list owner a knob
> they can use to always strip an existing Reply-To: header, regardless
> of whether Mailman adds one back or not.

Mmmh,  I'm really  not  sure why  one would  want  that. Replacing a  user's
Reply-To with the list munged Reply-to  can make sense (and it's yet another
reason  why listwide  reply-to  munging is  bad, since  you  lose the  reply
address of the poster if he/she set a reply-to to a different account).
However, having  mailman unconditionally remove  poster set reply-tos  (to a
different account  or a different list),  doesn't seem to be  an option that
makes too much sense.

But yes, re-reading Handlers/CookHeaders.py I now see why this is happening.
It seems that the  code doesn't do what the option was meant  to do, per the
documentation, the  comment in Defaults.py,  and the  fact that it's  in the
reply-to munging section.`

# Before munging Reply-To: Mailman can be configured to strip any existing
# Reply-To: header first, or simply extend any existing Reply-To: with one
# based on the above setting.  This is a boolean variable.
DEFAULT_FIRST_STRIP_REPLY_TO = 1

Would you agree  that this setting was really meant  to select whether, only
in the case where you do listwide reply-to munging, you replace the reply-to
with the  list reply-to  or you add  the list reply-to  to the  sender's set
reply-to?
In  this case,  would you  also agree  that since  RFC 2822  does allow  two
addresses  or  more  in the  Reply-to  header,  DEFAULT_FIRST_STRIP_REPLY_TO
should really  default to  0, because  your code  that extends  the existing
reply-to  allows for  the reply  to go  to both  the list  and the  reply-to
address the sender specified?

(my patch  on per-user  reply-to munging doesn't  have this  switch, because
should the end user  have an MUA that barfs on  multiple Emails in Reply-To,
and  I have  yet to  meet one,  (s)he  always has  the option  of not  doing
munging. Leaving this, as a disabled option by default, for listwide munging
can't hurt though)

Marc
-- 
Microsoft is to operating systems & security ....
                                      .... what McDonalds is to gourmet cooking
  
Home page: http://marc.merlins.org/   |   Finger marc_f@merlins.org for PGP key