Hello
I have to reduce the load from our mail server. We have really tons of
filters, which stresses the server sometimes really. You may say, use
skipifweight xx in every file and you are good. We use this on the servers
of our customers.
For my own server, I cant do it this way. I need to know, how many points a
certain filter file has generated. This data will show me, how good my
filter files are (or not). I tried testsfailed, which give me an indication,
if a filter was fired, but not the score.
I tried this:
# ==== If a different phrase file has given 2 digit points , don't continue
====================== #
# -> example: bl-<FilterName>-00-eng [00]
anywhere end pcre (?i:bl-.{4,15}-(00|01)-(pre|eng|fre|ger|source)
\[\d{2}\])
# -> example: bl-country-ar-00 [00]
anywhere end pcre (?i:bl-country-\w{2}-(00|01) \[\d{2}\])
I have a strict rule, how the names are of my filter files. Its easy, to
get them with RexEx.
Why I use anywhere? headers will never trigger. The reason for this might be
the fact, that headers will read the data before declude adds the line
X-Declude-Tests. Is this right?
Is the way I read the the test data okay and efficient? 200 filter files wil
have these checks.
Many thanks
Freundliche Grüsse
----------------------------------------------------------------------------
----------
netfusion GmbH | Martin Schaible
Mittelfeldstrasse 27 | CH-8700 Küsnacht | Switzerland
Tel.: +41 44 391 30 00
E-Mail: <mailto:[email protected]> [email protected]
Internet: <http://www.netfusion.ch/> www.netfusion.ch |
<http://hosting.netfusion.ch/> hosting.netfusion.ch
Helpdesk: <http://helpdesk.netfusion.ch/> helpdesk.netfusion.ch
Wird sind auch auf Facebook präsent:
<http://www.facebook.com/NetfusionGmbH> www.facebook.com/NetfusionGmbH
----------------------------------------------------------------------------
----------
nfqrcode150x150
