At 11:42 PM 7/15/97 -0700, Joseph Georges wrote:
>At 01:37 AM 7/16/97 EDT, Bill Gruber wrote:
>>On Tue, 15 Jul 1997 22:25:07 -0700 Joseph Georges said:
>>>The message thus goes to subscribers despite listserv >>complaining to me
>that it is a duplicate. Of course, >>it is not a duplicate at all.
>>
>>If LISTSERV says it's a duplicate, I'd have absolutely >no doubt that it's
>a duplicate.  The first one is >distributed to the list, the second duplicate
>>message is sent to you as an error.
>
>Actually, it is not a duplicate. As I mentioned in my original post, "I
>also occasionally receive another type of delivery error saying that a
>message (always a reply quoting another message) has already been
>distributed to the list."
>
>Listserv sees the reply as duplicating the original message that was quoted
>in the reply. The error message returns the reply to me, saying that the
>reply is unprocessed. But, in fact, the reply is processed and distributed
>to the list anyway despite the error message. I would like to find a way to
>disable this rather useless error message. If the cost of that step is that
>subscribers do occasionally post genuinely duplicate messages, so be it.

I suspect that indeed two copies of the reply are indeed being sent to the
list, and one is distributed and the other generates the error.

There are two ways that am aware of that this can happen (I've seen both):
        1. The replier sent both "To:" the list and "Cc:" the list
           (or possibly Bcc: the list).
        2. The replier hit "send" twice.
The first of these should have identical "Message-Id:" headers on the
emails; the second should have different "Message-Id:" headers.

As an example of the first, using Eudora, I used the "Shift-Reply"
(which is for a "group" reply) to your post; if I had not then edited
the headers, it would have gone both To: LSTOWN-L and Cc: LSTOWN-L.
I presume this is because LSTOWN-L rewrites the "To:" header, and
is also set "Reply-To:" LSTOWN-L.  (I do this on purpose so that the
top line doesn't read "...you wrote..." but rather puts your name in.)

As for the second:  some mail clients make it easy to accidentally
send twice by providing poor feedback that the first send took effect.

You should be able to get a better clue what the problem is by
getting one of the posts with full headers to compare with the
error message.  The easiest way may be to set your own subscription
(or open another one if you have a second address) with SET IETFHDR,
which will preserve the sender's headers on your copy.

Hope that helps,
Stan