knutsen@sri-unix.ARPA (10/29/85)
From: knutsen (Andrew Knutsen) Those huge messages were due to a mailer off in some remote part of the net sending error messages back to the "from" address. The new digesting software, in order to facilitate replys to the list, lists the list as the "from" address. So, when that mailer didnt recover from its "internal error"s in a reasonable amount of time.... What Ive done for now is remove the offending site from the list. Fortunately that site also sent out "date" lines in a format "rnews" couldnt parse, so none of the cruft leaked into Usenet (as far as I can tell). However, Im not too sure about an appropriate final solution. The choice seems to be to pray that very few sites have long-term problems and also send errors to the "from" address, or to list "physics-request" as the "from" address which will result in misdirected traffic to me on some replies. Any other ideas? Andrew, physics-request
MJackson.Wbst@Xerox.ARPA (10/30/85)
Would this have happened with the old, nondigesting software? If not, I'd vote for going back to the old stuff. I don't find the present grouping service particularly useful, given the fact that the range of response times is substantially greater than the grouping interval (24 hours). Mark
mikes@AMES-NAS.ARPA (10/30/85)
From: mikes@AMES-NAS.ARPA (Peter Mikes) Any other ideas (on handling mail problems)? Yes: How about reading the digests before you send them out? I always thought you do that . (?).