Re: [exim] bounce per lang

Top Page
Delete this message
Reply to this message
Author: Ian Eiloart
Date:  
To: Jeremy Harris, exim-users @ exim. org
CC: 
Subject: Re: [exim] bounce per lang


--On 8 September 2005 12:27:53 +0100 Jeremy Harris <jgh@???> wrote:

> Ian Eiloart wrote:
>> For multiple recipients, one could maybe 4xx the message, store a
>> message ID and charset in a database, and reject per recipient when we
>> see it next.
>
> Woah! You can see a message ID at rcpt time?


Ah, no. Good point. I can see a sender, an IP address and a recipient (or a
list of recipients). This isn't going to work in two cases:

1. A sender's host is a spambot. It spams me (and other recipients) in an
unwanted language a few minutes before the sender sends me a legitimate
email in a wanted language.

2. Vice-versa.

I guess the likelihood is increased in the event that the spambot is using
the sender's address book.

So, we'd have to implement this by storing the sender, recipient, IP
address triple. It's not error-proof, and it is quite complex. I guess
sticking to single recipient messages would fix 99% of the problem
(reliably) at a cost of 1% of the work.
>
> -Jeremy




--
Ian Eiloart
Servers Team
Sussex University ITS