dupfilter works by calculating a hash value of the contents of every message it sees. Hash values of recent messages are kept in a circular FIFO queue. When dupfilter starts seeing messages with the same hash value, it will reject them. The hash value is calculated in such a way as to try to detect minor changes to messages' contents that are designed to avoid these kinds of filters, but it's still not a very sophisticated approach, and it can be easily fooled.

BUGS

This is not a very smart mail filter. The hash function is very simple and is not expected to be very useful.

Don't even think of using dupfilter on a machine that either runs a mailing list, or contains mailboxes that subscribe to mailing lists. dupfilter likes to bounce mailing list traffic.

FILES

dupfilter uses the following configuration files. Changes to the following files do not take effect until the filter has been stopped and restarted.

/etc/courier/filters/dupfilter-hashsize

This file contains a single numerical value that sets the size of dupfilter's circular FIFO queue.

/etc/courier/filters/dupfilter-duplevel

This file contains a single numerical value that sets how many identical messages dupfilter will tolerate before blocking them. When this many identical messages are seen, dupfilter will start rejecting all messages with the same content.

/etc/courier/filters/dupfilter-mode

If this file exists and contains the word "all", dupfilter will create its socket in /var/lib/courier/allfilters, otherwise the socket will be created in /var/lib/courier/filters, see m[blue]courierfilter(8)m[][1] for more information.

/etc/courier/filters/dupfilter-nthreads

This file contains a single numerical value that sets the number of threads created (each thread is used to calculate a message's hash value). The default number of threads is 4.

SEE ALSO

m[blue]courierfilter(8)m[][1].

AUTHOR

Sam Varshavchik

Author

NOTES

1.

courierfilter(8)

[set $man.base.url.for.relative.links]/courierfilter.html