http://bugzilla.spamassassin.org/show_bug.cgi?id=4013
------- Additional Comments From [EMAIL PROTECTED] 2004-12-03 10:08 -------
Subject: Re: Please add IP address host checking to urirhsbl and urirhssub
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
ah, I remember this now. it's entirely to do with looking stuff up in
SBL-type "lookup domain NS in blocklist" zones. In the domain-name case
we do:
- resolve domain name's NS
- resolve those into A records
- lookup addresses in SBL
but if it's an IP address we can skip straight to the very last step,
hence the optimisation. However, that assumption fails when you're
looking it up in SURBL or any similar "lookup domain name in rhsbl" zone;
I guess I never checked that! (I should have added a test.)
The fix is to have that "optimization" conditional on !is_rhsbl.
BTW, that flowchart above is one of the reasons where there's two separate
rule types, and why there's a difference between them. "lookup domain NS
in blocklist" is entirely different behaviour from "lookup domain name in
rhsbl", due to the extra steps and different lookups required.
I'm -1 on the idea of changing them to a single rule type, differentiated
using a tflag -- throwing in random tflags is messy as we just wind up
cluttering the tflags space with a tflag that only makes sense for URIBL
rules. It makes more sense to keep "tflags" as a space for flags that work
for all, or at least most, rule types.
- --j.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Exmh CVS
iD8DBQFBsKujMJF5cimLx9ARAqKTAJ0ej7vWN/wg5iRCqK6hh5XlSm/VhACffdYn
glauborxGPxufSAD/PMrbEI=
=qz7E
-----END PGP SIGNATURE-----
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
