Just as a side note to all of this, I did some testing when this thread
first come up and sent the results on to Derek, perhaps he can dig up that
email and post it here...
I injected a 'raw' message into various MTA's ran by various companies
that I have accounts with or 'relay access' to, by using a standard
command prompt telnet session. The message I injected contained
a BCC header with multiple addresses in them, and to my thinking, this is
a sendmail does one thing, everyone else (or nearly everyone) does another
problem.
The two sendmail and one Lotus Notes 5 server either removed or striped
the BCC header, while Exim, OpenWave InterMail, MS SMTP, MS Exchange, and
qmail (I did not test postfix that I remember) all left the BCC header
intact and unmodified.
My personal opinion on this goes along these lines....
Sendmail has historically been a part of the MUA (or rather, part of the
MUA) for many *nix mailers and utilities, and therefore has been
defaulted (don't know if you can change this behavior, but searches on
Google news groups seems to say no) to strip the BCC header for the
client regardless of how the message is accepted, or from where it comes.
Since the other SMTP software have been primarily for the MTA parts of the
transactions only, they do as an MTA should, and leave the message headers
and body intact without modification.
So, outside of the Sendmail world (with a few exceptions, Lotus being one)
it seems that all other MTA software expects the MUA to decide what and
how the BCC header is to be handled. MS Outlook/Outlook Express for
example never send a BCC line ever, I do not know, nor have I tested
exactly what other MUA softwares do, but I would think that the majority
of them do as Outlook does, and just never includes a BCC header with
email address in it.
Just somethings that I found, and I thought you may want to take note of.
Cheers!
--
--EAL--
