>Yes but as I said LISTSERV itself does not generate error delivery
>notices, as it doesn't actually deliver the message: it just passes it to
>the mailer, which performs the actual physical delivery and will complain
>by itself if there is a problem. And, the rejection notices we have been
>speaking about don't come from LISTSERV but from VMS nodes (well unless I
>have misunderstood something?)
Hmmm, maybe they have replaced CMS by VMS on this machine without
telling me. Recently I received two questions/complaints regarding
mailings from Listserv. From the user's point of view these mailings
are shitty error messages which are sent to the From: address (instead
to the list maintainer) by one of these braindamaged Bitnet mailers.
We all know that Bitnet is one the best things one could imagine,
but sometimes it is a bit difficult to explain this to aliens living
in Arpa- and UUCP-land :-)
1) If Listserv can't append a letter to the notebook for this list,
it sends a notification (Error xyz from EXECIO while logging ...)
to the sender of the letter. Looks not much better than these
cryptic VMS or UUCP error messages, at least for the innocent
user who is not able to fix the problem. Furthermore I'm not
sure whether this notice is sent to the From: or to the Sender:
address (if present). I have looked into LSVXMAIL, and would say
it will go to the From: address, but I don't understand the
code good enough to be sure. From what I have learned by the question
of an affected user, I tend to believe that the Sender: field is
ignored.
2) The moderator of a Usenet newsgroup sent me a complaint ("FIX YOUR
MAILER"). because he found a batch of "rejection notices"
from Listserv in his mailbox.
Date: Sat, 13 Feb 1988 03:01 MEZ
From: Revised List Processor (1.5m) <[log in to unmask]
Subject: Output of job "UNIX-SRC" from LISTSERV@HEARN
To: [log in to unmask]
Status: R
No recipients for this server, probable loop
symptom. Distribution for all the remaining
recipients has been carried out by this server.
Here is the story: This Usenet newsgroup is gatewayed to a Bitnet
list (maybe there is an Arpa list in between, but this doesn't
matter). I assume that it is set up as a Mail-via= Distribute
mailing list. But then there should be a Sender: field where
these notifications (if ever) should go. The distribute job came from
FINHUTC(?) via HEARN to DB0TUI11 to serve some recipients at DBSTU1
(which is located between HEARN and DB0TUI11). It is hard to say
whether this is the best choice as the EARN topology was changed
once per day recently :-) The "probable loop symptom" barking
vanished after I restarted Listserv, so I assume this warning was
caused because of an inconsistency problem with network tables.
Anyway, I can understand Rich Salz' complaint, he is really
not the right recipient for this kind of notices.
I don't ask for just another hack in Listserv. I would like to ask
Eric to make sure that error notices as the above are sent to the
Sender: and not to the From: address.
----------------------------------------------------------------
This doesn't belong here, but I need to say it:
I'm so tired of reading flames on network A regarding the behaviour
of software on network B, whose participants are eagerly writing
flames regarding the messy software on network C and so on ...
with so many flames based on half-baked knowledge (e.g. mixing
RFC 821 and 822 FROM, ignoring that not all the world is speaking
SMTP, ...). It's just like in real life, it's easier to flame each
other than trying to understand each other.
|
