Hello, Does anyone have a simple sieve web interface for creating a filter for spam-tagged messages? We'll be tagging spam messages with a "X-Spam-Status: Yes" header or somesuch and now I'm looking for an easy Sieve interface for creating simple sieve scripts to filter spam-tagged messages. At the moment some of our users are using Websieve but I would like to offer an easier web form. Does anyone already have one? If there is one publicly available, I'd be happy to see it. Otherwise I'll just have to make one. :-)
I've recently upgraded my Exim setup to include ExiScan-ACL and SpamAssassin. It is configured to add several headers for different levels of filtering capability. (The ones I chose were taken from the examples for ExiSan-ACL and SA; so they should be fairly common. They differ slightly from the set which SA can be configured to add.)
The X-Spam-Score header contains a floating-point number, which may include a leading sign, followed by a parenthesized string. In the string will be a single slash, or a series of plus or minus signs. The number of plus or minus signs equals the integer portion of the spam score. Thus:
X-Spam-Score: 12.1 (++++++++++++)
It is very easy for even fairly primitive mail filtering programs to recognize a number of plus or minus signs. In Sieve:
if header :contains [ "X-Spam-Score" ] [ "++++++++++" ] {
fileinto "INBOX.SPAM.score-10+" ;
}
elsif header :contains [ "X-Spam-Score" ] [ "+++++" ] {
fileinto "INBOX.SPAM.score-5+" ;
}
elsif header :contains [ "X-Spam-Score" ] [ "-----" ] {
fileinto "INBOX.HAM.score-5+" ;
}
I also provide an X-Spam-Score-Int; which is just the spam score multiplied by 10 to create an integer. (For filters that can deal with integers but not floating point.)
And there's an 'X-Spam-Flag: YES' header added if the spam score is over some site-specified limit. (But lower than the automatic rejection limit.)
(I also include the SA spam report as a header; which could be used by more sophisticated filters based on particular tests matched by SA.)
-Pat
