Thu, 29 Sep 1994 20:43:35 +0100
|
On Thu, 29 Sep 1994 15:01:00 EDT Shahrukh Merchant <[log in to unmask]>
said:
>
>1. The problem seems to be that one of the systems along the way is
> inserting an extra blank line in the headers, causing some of the
> headers to look like part of the message body. This seems to be
> "on or near" admin.fsu.edu, if you follow the "Received:" headers.
> (I have one user on that machine, and the problem started after he
> subscribed.) Has anyone seen this sort of thing before? Who is
> the culprit and who should fix it?
"Smail" is known to do that (mostly on Linux systems, but then Smail is
the default mailer for Linux). I have no first hand experience with that,
though. IBM's SMTP software is known to do that if it is configured to
use PUNCH format to pass mail to the system mailer, but this is a rare
case. You have to have an address with a very long local part, like an
X.400 address, for it to happen. IBM refuses to fix that bug, saying
people should use NETDATA between SMTP and the system mailer. People
should indeed definitely do that, for *other* reasons, and it does happen
to bypass the buggy code.
>2. Why did the message ever get back to listserv?? Seems like it got
> to listserv properly and got distributed properly to other members
> of the list. Even if admin.fsu.edu inserted an extra blank line,
> how did listserv@mitvma get to see the message again?
I guess some idiotic mailer along the way decided that the "To:" field
was an interesting address to send the message to, and it got back to the
list.
>3. Why does listserv care even if there is a "Sender:" "From:" or
> "Reply-To:" field in the message body that points to the list?
That's how it detects mailing loops caused by incidents such as this one.
If it hadn't filtered the message, you'd have 50 copies in your mailbox
by now.
Eric
|
|
|