>> It was very interesting, though, to see how many people misspell
>> addresses or use the wrong format (such as people trying to send
>> mail to me at scott_perry@ or scottperry@ or sperry@ or scot.perry@,
>> etc., rather than [EMAIL PROTECTED]). Finally I wrote a
>> program to handle it automatically.
> details??
The details are quite dirty.
First, I have the "nobody" alias point to an account called "bouncer". That's the
extent of the setup on the IMail end.
The program I wrote then checks the \USERS\BOUNCER directory every minute or so, and
looks to see if the there is new mail (by checking the .UID file), and if so, it reads
the appropriate section from the .MBX file to get the E-mail address of the sender,
and the bogus address. It does a number of checks on the bogus address, comparing it
to real addresses, trying to find a match. It then generated a reject message through
imail1.exe. The only problem I haven't overcome is that I can't use imail1.exe to
send an E-mail with a null sender ("<>"), which the RFCs kind of require in this case.
It works well in a corporate setting, where all the E-mail addresses are in the same
format, but probably wouldn't work well for an ISP or other situation where users
choose their own names and don't know other people on that domain.
Of course, I'm sure that directly accessing the mailbox file voids any warrantees that
may have previously existed. It's not for the weak hearted.
I thought of using a program alias instead, which would have been a lot cleaner (no
going to the mailbox file), but you can't get the bogus address that way (you could
look at the "To:" and "Cc:" fields, but don't get the "Bcc" recipients). I also
thought of using the newly announced virus scanner interface, but that would also have
pitfalls (I would have to do the work to determine if the message was destined to
bounce, checking all user accounts and aliases and such).
-Scott
Please visit http://www.ipswitch.com/support/mailing-lists.html
to be removed from this list.
