[Mailman-Users] Re: autorespond_requests in 2.1x

[Mark Geisinger]

Sunday, December 29, 2002, 2:06:27 PM, you wrote:

>>>>>> "MG" == Mark Geisinger <list-mailman at lanbytes.com> writes:

>     MG> Is autorespond_requests broken in 2.1x, or do I perhaps
>     MG> misunderstand its purpose? I'm using 2.1b4 and would like to
>     MG> avoid having a request autoresponse sent to a subscribe
>     MG> request, or for any other request.

> You (or I ;) might be misunderstanding the feature.  Any message sent
> to the -request, -join, or -leave address will get a confirmation
> message indicating the results of the request.  This isn't
> configurable.

> In addition, Mailman can send an autoresponse message to messages sent
> to the -request address, and this is configurable via the
> autorespond_requests variable.

Barry, thanks for the explanation. Such an autoresponse to -request makes sense,
given that such a message can be related to commands that require action, or
response, outside of the subscription process. Consider this behavior, though.
We provide a Web interface to a handful of topical lists. If someone comes by
and subscribes to four lists, they receive eight messages. What's the point?

I just un-sub'ed a test address from one of my own lists. Confirming the
un-subscribe, I first received the expected two line "Your address has been
removed from the mailing list." message. Cool. Then I get the autoresponse that
offers this,

> The results of your email command are provided below. Attached is your
> original message.
> 
> - Results:
>     Confirmation succeeded

along with the entire body of the original remove confirmation message.

My point is a) that traffic for subscription confirms is redundant, and b) it's
actually confusing subscribers. Since moving from v2.0.13, where this behavior
didn't exist, I've actually had a number of folks miss the purpose of the
subscription confirmation msg (and thus not confirming), believing that the
autoresponse is an affirmative response to their sub. request.

Is there no way to ditch this behavior for join/leave?

-- 
Best regards,

Mark
list-mailman at lanbytes.com