On Wed, 14 Jul 1993 17:00:00 PDT Melvin Klassen said:
>On Wed, 14 Jul 1993 14:52:43 PDT Tim Rushing <RUSHING@WSUVM1> wrote:
>>Here's the sequence of events:
<stuff deleted>
> 3A. Some subscriber grabs the message, and posts it back to 'TESL-L@CUNYVM',
> either maliciously, or, with the help of some "I'm on vacation" software.
> 3B. LISTSERV calculates the checksum for this posting,
> and finds that it matches the checksum from a recent posting.
> 3C. LISTSERV decides to **not** distribute the "duplicate" posting
> to the subscribers.
NO. sorry. see below.
<stuff delete>
> 5. List-owner breathes a sigh of relief,
> because LISTSERV's "loop-detection" code has prevented a loop
> which would have bombarded each subscriber with hundreds of copies
> of the same message. :-)
>
there still is the limit of the number of messages the list processes,
which would have generated user complaints, but I got those anyhow since
every posting to the list posted the message then gave the 'duplicate
posting' message to the sender of the message.
>
>It *MUST* be the fault of *ONE* of those users,
>given that the problem vanishes when those users
>don't receive postings from the list.
>
no.
I had the same problem with a list. Each posting would generate the
message, without listserv even receiving another file! You would see
multiple messages being processed by listserv in the listserv log.
the second one would follow the first right away. there would not
be sufficient time to send the mail out to a mailer, and have it bounce
back.
I was able to duplicate the problem at will at the time.
I included the log in my posting of the message some time ago, and
got no response.
I started manipulating userids (i.e. deleting suspicious userids, posting,
trying again). About that time, the listowner sent in a bunch of deletes,
and the problem disappeared. I didnt have the time to pursue it.
jr
>One by one, set each of them back to 'MAIL' status,
>make a posting to the list, and correlate the results.
|
