Tomcat problem found

[clark@tomcat.gsfc.nasa.gov]

Well, after much head-scratching the problem with tomcat barfing seems to
have been fixed.

What seems to have been the case is that ucsd.edu had a message queued to
send to tomcat that had in the text a line beginning with "..   " followed
by more text. The internal mailer in NOS took the first "." as the end-of-
text delimiter, but the code was bombarded by the rest of the message coming
in. Somehow this caused what I believe to be a stack overflow in the 8/28
version of NOS, causing instant death!

My fix was pretty trivial -- I replaced NOS with a copy of the old NET
(ver.31) which didn't croak, allowed the bogus message to be delivered,
and then brought NOS up again.

Tom