It's hard to believe a single period on a line by itself could get
by as far as sending out a digest. It would take more than _one_
MTA along the way to not handle it correctly, it would take _every_
relay point not handling it correctly.
However, how about either a NULL or a CTL-Z? The behavior of individual
systems will vary on either of those, including the choices of:
1) being ignored as if they weren't there;
2) being translated to a printable char. sequence (my mail-reading
system does this);
3) not being handled correctly and causing premature truncation of
The NULL is probably the better guess, because taken as an end-of-string/
end-of-line indicator, it could cause the trailing text to be munged.
The CTL-Z would just indicate end-of-file, and you would _never_ see
any trailing text.
On the other hand, it's easier to imagine her using a text editor in
her mailer that inserts DOS-style CTL-Z end-of-file characters, than
a message-ending NULL.
Another guess heard from...