<<< Date Index >>>     <<< Thread Index >>>

Re: Dangers of discarding duplicated messages



On Fri, 18 Feb 2005, Maciej Soltysiak wrote:

> Developers: Consider using checksum of the email messages, not only the
>             Message-ID to distinguish between duplicated messages.

We have a product that needs to detect duplicate messages (retransmissions
in the face of temporary SMTP failures.)  Obviously, we'd never use
the Message-ID for this.  Less obviously, you can't always use a hash
because some mailers actually change the message body on each retransmission!

For example, Lotus Domino appears to store the e-mail in some internal
format, and generates a new MIME message on each retransmission.  Each
generated message has different MIME boundary delimiters.

So we have to calculate a "modified" hash and ignore bits of the message
that mutate.

--
David.