I'm familiar with MAXWEIGHT and I'm using it.
It doesn't address this particular application.
Best Regards
*/Andy Schmidt/*/
/
Phone: +1 201 934-3414 x20 (Business)
Fax: +1 201 934-9206
-----Original Message-----
*From:* [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] *On
Behalf Of *John T (Lists)
*Sent:* Friday, November 17, 2006 05:52 PM
*To:* declude.junkmail@declude.com
*Subject:* RE: [Declude.JunkMail] Filter 'END' statement in 4.3.14
flushes WEIGHT?
OK, I understand that better but you will always be better off
grouping each intent into a different combo filter. Then, you can
even have a combo filter dependent upon another combo filter by
why of order of list and including the name of the combo filter as
an IF statement in the next one.
Combo filters need to be viewed as a different type of test rather
than a normal filter test. If you write down in groups want you
want to do, it will be easy to then create them.
Say if you want to add 12 if 4 or more rbl tests failed. You would
create a combo filter like this:
MINWEIGHTTOFAIL 12
MAXWEIGHT 12
TESTSFAILED 3 CONTAINS rbl1
TESTSFAILED 3 CONTAINS rbl2
TESTSFAILED 3 CONTAINS rbl3
TESTSFAILED 3 CONTAINS rbl4
TESTSFAILED 3 CONTAINS rbl5
TESTSFAILED 3 CONTAINS rbl6
TESTSFAILED 3 CONTAINS rbl7
TESTSFAILED 3 CONTAINS rbl8
That way, at least 4 have to hit to equal 12 before it will see
this test as failing, but it will only add 12 and not more.
**John T**
**eServices For You**
****
*"Life is a succession of lessons which must be lived to be
understood."*
*Ralph Waldo Emerson (1802-1882)*
**
-----Original Message-----
*From:* [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] *On
Behalf Of *Andy Schmidt
*Sent:* Friday, November 17, 2006 2:13 PM
*To:* declude.junkmail@declude.com
*Subject:* RE: [Declude.JunkMail] Filter 'END' statement in 4.3.14
flushes WEIGHT?
Hi John:
>> What is the logic of the second part anyways, to add weight for
test4 and test5 IF test1 and test2 failed? <<
If you have several blacklists of the same family (e.g., multiple
open-relay filters, or multiple open-proxy filters) I like to
group them together. I give a big weight to the entire group (the
filter itself) and then may add an increment for blacklists with
few false positives (each "contains" clause).
Simiarly with Sniffer or invURIBL. There is some overlap between
those two, and there is a potential overlap between Sniffer-IP and
blacklists of "recent spam sources" (e.g., SpamCop, MXRate-Block).
I have a filter that processes my various Sniffer types and
invURIBL returns. At some point, I'd like to stop and first look
if certain other Blacklist Tests had fired. If so, I'm done. If
not, I want to add a little extra for Sniffer-IP.
Best Regards
*/Andy Schmidt/*/
/
Phone: +1 201 934-3414 x20 (Business)
Fax: +1 201 934-9206
------------------------------------------------------------------------
*From:* [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] *On
Behalf Of *John T (Lists)
*Sent:* Friday, November 17, 2006 02:41 PM
*To:* declude.junkmail@declude.com
*Subject:* RE: [Declude.JunkMail] Filter 'END' statement in 4.3.14
flushes WEIGHT?
Any filter that I do not have as active is moved to
\declude\filters\notused from \declude\filtes so that my filters
folder only contains filters that I am currently using.
In your example, you are putting the IF statement after the THEN
statement. I am not a programmer, but IF (the test of what you
want to be quantified) always comes before THEN (the result you
want to occur depending upon whether the test of the
quantification failed or passed, meaning did it meat the criteria
which is the defined IF statement), so the first part of your
example does not make any sense. You only want weight added if
test3 failed, so you have to quantify that FIRST and then say what
occurs by adding weight.
So, you have to have to filters since both sections rely upon
testing IF test3 has failed or not.
What is the logic of the second part anyways, to add weight for
test4 and test5 IF test1 and test2 failed?
**John T**
**eServices For You**
*"Life is a succession of lessons which must be lived to be
understood."*
*Ralph Waldo Emerson (1802-1882)*
-----Original Message-----
*From:* [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] *On
Behalf Of *Andy Schmidt
*Sent:* Friday, November 17, 2006 10:48 AM
*To:* declude.junkmail@declude.com
*Subject:* RE: [Declude.JunkMail] Filter 'END' statement in 4.3.14
flushes WEIGHT?
>> Why the requirement of single filter? <<
Clarity? It's easier for me to follow a logic, if it's enclosed in
a SINGLE source document (= filter).
If the logical is spread over multiple source documents, I have to
first scour the Global.CFG to see which filters are active,
then inspect each one to see if by chance any one of them might
have any effect.
Best Regards
*/Andy Schmidt/*/
/
Phone: +1 201 934-3414 x20 (Business)
Fax: +1 201 934-9206
------------------------------------------------------------------------
*From:* [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] *On
Behalf Of *John T (Lists)
*Sent:* Friday, November 17, 2006 12:57 PM
*To:* declude.junkmail@declude.com
*Subject:* RE: [Declude.JunkMail] Filter 'END' statement in 4.3.14
flushes WEIGHT?
Why the requirement of single filter?
I have different combo filters created like this:
ComboFilterA
REM If testa and testb fail, and if testc or testd fail, add 10
ENDONFIRSTHIT
TESTSFAILED END NOTCONTAINS testa
TESTSFAILED END NOTCONTAINS testb
TESTSFAILED 10 CONTAINS testc
TESTSFAILED 10 CONTAINS testd
ComboFilterB
REM If testc and testd fail, and if teste or testf fail, add 20
ENDONFIRSTHIT
TESTSFAILED END NOTCONTAINS testc
TESTSFAILED END NOTCONTAINS testd
TESTSFAILED 10 CONTAINS teste
TESTSFAILED 10 CONTAINS testf
IMHO, that is a much cleaner and neater way to do it. You could
also use MAXWEIGHT instead of ENDONFIRSTHIT and then assign
different weights to different test.
**John T**
**eServices For You**
*"Life is a succession of lessons which must be lived to be
understood."*
*Ralph Waldo Emerson (1802-1882)*
-----Original Message-----
*From:* [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] *On
Behalf Of *Andy Schmidt
*Sent:* Friday, November 17, 2006 9:29 AM
*To:* declude.junkmail@declude.com
*Subject:* RE: [Declude.JunkMail] Filter 'END' statement in 4.3.14
flushes WEIGHT?
Hi John,
>> Was never changed. <<
Please read the URL I posted:
http://www.mail-archive.com/declude.junkmail@declude.com/msg14009.html
As you can tell, ORIGINALLY it did return the weight. He was
thinking of it even FAILING the test (if there was a weight).
>> What you should have been using is MAXWEIGHT at the top, or
STOPATFIRSTHIT. <<
Kindly, please read the sample I had posted (bottom of this
message). Your reply doesn't address the issue of trying to make
some sections of a test conditional. Example, the goal is to
return either 1 or 2 or 3 if test1 or test2 occur with test3 - and
to only add test4 and test5, if test3 is not true.
SKIPIFWEIGHT 20
MAXWEIGHT 3
TESTSFAILED 1 CONTAINS test1
TESTSFAILED 1 CONTAINS test2
TESTSFAILED END CONTAINS test3
TESTSFAILED 1 CONTAINS test4
TESTSFAILED 1 CONTAINS test5
etc etc
Please demonstrate how MAXWEIGHT or STOPATFIRSTHIT would do this
in a single filter?
Best Regards
*/Andy Schmidt/*/
/
Phone: +1 201 934-3414 x20 (Business)
Fax: +1 201 934-9206
---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail". The archives can be found
at http://www.mail-archive.com.
---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail". The archives can be found
at http://www.mail-archive.com.
---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail". The archives can be found
at http://www.mail-archive.com.
---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail". The archives can be found
at http://www.mail-archive.com.
---
This E-mail came from the Declude.JunkMail mailing list. To
unsubscribe, just send an E-mail to [EMAIL PROTECTED], and
type "unsubscribe Declude.JunkMail". The archives can be found
at http://www.mail-archive.com.
