Serge Knystautas wrote:
>
> I started thinking of several far fetched examples as to why this won't
> work (notes scribbled below), but I have 2 really good reasons (IMHO)
> why this won't work. (I agree it'd be a nice hack if it allowed us to
> work against other mail servers, but consider these issues.)
>
> 1. There is delivery information that is not contained in the email
> message itself. You designate through SMTP's RCPT TO command who this
> message should be delivered to... these addresses are in no way required
> to be in the email address.
Agree. That's way you have a State object containig all SMTP info and a
different object Headers to access headers.
>
> 2. It restricts mail servlets as receivers only... a real life example
> is such as attaching a disclaimer on all email leaving the company (or
> logging it, whatever)... you need to process email messages as they are
> handled by the SMTP server and not wait for them to arrive at a POP3
> address.
Actually MailServletRequest and Response are just message wrappers so
basically the Response is returned to the reactor and will be passed as
a new Request to the next MailServlet.
This way they act like filters. At the end of the servlet pipe line
you'll have a DeliveryMailServlet which will take cares of deliver
remote messages, a localDeliveryMailServlet to store locally mails etc.
Only these last servlets are receivers.
>
> It also means you need a separate mail box for every address you're
> going to consider processing. This introduces much more overhead and
> administration it seems to the mail server... it also means you can't
> use dynamic email addresses (not sure whether it's a good practice or
> not).
Do not agree. All changes during servlet pipe are kept in memory so that
only when the message is successful delivered the original message is
deleted. All this is acts like a simple transaction db so that is the
server goes down during processig at reboot every unfinished work
restart from the original message and nothing is lost.
>
> Serge Knystautas
> Loki Technologies
> http://www.lokitech.com/
>
> ------------------------------------------------------------
> To subscribe: [EMAIL PROTECTED]
> To unsubscribe: [EMAIL PROTECTED]
> Archives and Other: <http://java.apache.org/>
> Problems?: [EMAIL PROTECTED]
--
Federico
*-----------------------------------*
"Troubles occurr when two people try
to be intelligent in the same time."
- Murphy
*-----------------------------------*
------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Archives and Other: <http://java.apache.org/>
Problems?: [EMAIL PROTECTED]
