In an MTA which I am developing, I dot-strip every message body which
comes in through SMTP. Later I dot-stuff every message body which
goes out through SMTP or POP3. (Reference RFC 2821 section 4.5.2 and
RFC 1939 section 3 paragraph 4.)
But I never use the message body between the times when I dot-strip
and dot-stuff. Unless I am mistaken, I could entirely skip the
dot-stripping and dot-stuffing, given these functions which I am
performing.
So, if this sort of optimization is worth considering, I think it
would entail rethinking the boundaries which surround where message
bodies need to be dot-stuffed. Perhaps dot-stuffing would occur when
a message first entered the SMTP system, dot-stripping would occur
when a message was handed to some local process where no further
Internet transport was expected.
Rich Hammer
mailscreen.net
Hillsborough, N.C.