P3D Re: digest loops?

[Bob Wier <wier@xxxxxxxxxxxxxxxxx>]

At 3:57 AM -0600 on 8/25/98, Tom Deering wrote:


|On 8/24/98, Bob Wier wrote:
|>Here's a message which got diverted off. Note that you *cannot*
|>quote digests - listserv's anti-looping mechanisms are quite
|>emphatic on this as a digest loop will likely cause disk space
|>overflows on the server...
|
|Hey Bob!
|
|Could you explain this looping thing, technically?  I've never heard of
|this, and thought I had quoted digests before.  I'm probably
|misunderstanding you, but I don't know how.
|
|Cheers!
|
|Tom

Hi Tom - 

A mail loop is [a very bad thing (tm)]  :-)

There are generally two kinds - small ones and big ones. The big
ones are worse than the small ones (also :-)

The small ones occur if a person sends a message which for some
reason is malformed. Listserv bounces it back to the sender. Now,
if the sender's address is also malformed (unusual, but possible),
then the sender's machine also generates a bounce back to listserv.
Listserv receives the bounce, but since it didn't originate the
message, bounces is back to the sender - etc etc etc etc...

The big ones occur if an additional copy of the bounce message
is forwarded out to everyone on the list. What happens then is that
many of the usual subscribers will bounce the bounced message, returning
it to listserv, which ALSO generates another bounce. Etc etc etc etc.

Generally this continues until the disk space on one or the other
machine is filled and it crashes. 

Since digests are long anyway (typically 20k), you *really* don't want
to have a mail loop form with multiple digest bounces. Hence listserv's
extremely agressive action against anyone quoting a digest (since that
APPEARS to be a bounced message with a digest content). 

It's a balancing act between stopping mail loops, and being so picky about
addresses that lots of people can't get mail because of slight irregularities
such as the funky behavior of MSMAIL...

THANKS

--Bob Wier

 



------------------------------

End of PHOTO-3D Digest 2922
***************************