According to RFC 2822, the presence of the Bcc header
field or the absence of the recipient's address in the
recipient list are both indicators that the recipient
was sent a blind carbon copy.
During a reply, if the user was determined to be the
recipient of a blind copy, let's warn the user of the
implications of replying to the message, which will
reveal to the recipients that the blind recipient
received a copy of the message. That could get the
sender in deep doo-doo. The prompt should definitely
Because the MH behavior of including a Bcc warning in
the text of the message is neither condoned by RFC
2822, nor implemented by most agents, we should not
depend on it to inform the user.
Thanks to Sergey for pointing this out.