Re: Code questions...

[John Gardiner Myers <jgm+@CMU.EDU>]

Sorry, I got confused as to which side of the protocol was being
talked about.

In the server, the timeouts you mention are fine.  To avoid the
duplicate mail problem described in RFC 1047, the server should make
sure it doesn't dally (like trying to actually deliver the mail, do
DNS lookups, etc.)  between receiving the final "." and sending the
250 reply.  I don't have the current code handy, so I can't check on
what it's doing.

In the SMTP client, the timeout between sending the final "." and the
250 should be as large as reasonable--10 minutes is the absolute
minimum.  Unfortunately I don't have data on what kind of delays show
up in real practice, but unless there's reason to do otherwise I would
suggest using sendmail's default of a 1 hour timeout on that
particular protocol exchange.

-- 
_.John G. Myers		Internet: jgm+@CMU.EDU
			LoseNet:  ...!seismo!ihnp4!wiscvm.wisc.edu!give!up