Wed, 3 Nov 1993 18:49:52 EST
|
On Wed, 3 Nov 1993 17:55:25 EST Gary Sponseller said:
>Has anyone else seen a duplicate posting problem that was solved solely
>by changing X-Tags? If not, how would you trace this problem from this
>Listserv site to that end site? Personal mail sent to the remote user
>arrives there in single copy...
Yes, this happened on TSNEWS-L a few months back. I was able to determine
that messages were bouncing off a certain site back to here and then being
sent out a second time. I contacted the postmaster at that site and got
the following response (edited here):
>This sounds like a known but esoteric bug involving use of UCLA/Mail and
>Interlink's SNS/TCPaccess.
>>but I suspect that the problem is due to an unfortunate interaction
>>between a poor LISTSERV design choice and a poor design choice in the
>>UCLA code upon which Acces/MVS, er, SNS/TCPaccess is based. I'm
>>specifically referring to the use of X-TO: and X-FROM: to pass the
>>recipient addresses from TCPaccess to UCLA/Mail. What happens is that
>>most LISTSERV lists have the default value for the "X-Tags=" keyword,
>>and under some conditions that I don't remember at the moment, this
>>causes mail to be distributed by LISTSERV with a header line like
>> X-TO: listpostingaddress
>>If you're running UCLA/Mail with the GATEWAY option, when this
>>arrives, it will be reflected back to the list. Eric Thomas, our VM
>>people, and I spent quite some time diagnosing this problem a few
>>years ago. Ironically, I think it may have been occurring on my
>>MVSTCPIP list!
>>
>>To fix it, you should do *both* of the following.
>> - Get the list header changed to include "X-Tags= Comment".
>> - Get the vendor to change TCPaccess to send BSMTP to UCLA/Mail
>> instead of using the in-band X-TO/X-FROM.
I changed the list to "X-Tags= Comment" and the duplication ceased.
------------------------------------------------------------------------
Kent Fiala <[log in to unmask]>
SAS Institute Inc., Cary NC 27513 USA 919-677-8000 x6646
|
|
|