[SURBL-Discuss] Re: multi.surbl.org combined list available for development

Jeff Chan jeffc at surbl.org
Wed Jun 16 01:06:24 CEST 2004


On Tuesday, June 15, 2004, 5:59:17 AM, Frank Ellermann wrote:
> Jeff Chan wrote:

>> I've set up a bitmasked, combined SURBL list with sc, ws,
>> and ph data:   multi.surbl.org

> Tnx, I replace sc.surbl.org by multi.surbl.org in my script.

OK, I should have mentioned that's not the preferred way to do
things.  :-)  The multi.surbl.org list has lists encoded with
bitmasks that SpamCopURI and urirhsbl won't work too well
with.  You *could* make many rules to handle every numeric
combination of the bits, but since this would require 2 ^ N
rules, it shows that the existing single-list programs aren't
really the right tool.  What's needed is some different or
added code to decode these back into their original lists.

Justin Mason is adding such code into a new program "urirhsblsub"
now.  Hopefully Eric Kolve will be able to roll similar code into
SpamCopURI in future also, or perhaps code up a differently
named program.

> I'm not sure how other combined lists handle this, but how
> about some "combined" test entries:

> Add 127.0.0.4  to ws,            127.0.0.6  to sc + ws,
>     127.0.0.8  to ph,            127.0.0.10 to sc + ph,
>     127.0.0.14 to sc + ws + ph.  127.0.0.12 to ws + ph.

> With these test entries we would get for multi.surbl.org :

>  0.0.0.127.multi => not found
>  1.0.0.127.multi => not found (= no localhost confusion)
>  2.0.0.127.multi => 127.0.0.2  => sc
>  3.0.0.127.multi => not found (same for all odd.0.0.127)
>  4.0.0.127.multi => 127.0.0.4  => sc + ws
> [...]
14.0.0.127.multi =>> 127.0.0.14 => sc + ws + ph

Yes, we could probably make some bitmasked tests.  I'll
consider it a task for later....  :-)

Jeff C.



More information about the Discuss mailing list