Re: can anybody review my patchset?

Am 20.04.2009 02:25 schrieb Matthew Dempsky:
> > 9. own Patch: let tinydns rotate multiple MX records
> 
> This patch is bad.  There's a buffer overflow if you serve MX records
> with names over 128 bytes.  If you have over 8 MX records for a name,
> it will always discard records after the first 8.  The shuffling
> algorithm can take much longer than necessary; in particular, it goes
> into an infinite loop if "." is used as a mail server name.
So I will remove the patch from the web.
I know, It's not optimal. also I habe only 6 Recordy which _are_ shorter then 128 byte.
But this is not a apology.


> What's the motivation for these two patches?
8. maxttl-patch

our internal systems may not query to the internet itself. They can only act as
forwarder to an other dnscache. So the are unable to resolv for instance MX-Records
for domains in transit in realtime. With this patch any internal system 
can query to an "dnsproxy" an get real, fresh an uncached results.

9. mx rotation patch
I have 6 MX-Servers. The first is permanent overloaded by stupid spam engines.
rotating MX-Entries seemes cheaper then a Loadbalancer.

-- 
########################################################################
#
# Andreas Schulze
# https://andreasschulze.de
# 
# GnuPG Key-ID: A7DBA67F, https://andreasschulze.de/sca.asc
# GnuPG Fingerprint: 14C1 39A8 CE6D 6BE0 28C6 5652 03B5 6793 A7DB A67F
#
# $Id: .signature,v 1.3 2007-12-27 21:13:36 sca Exp $
########################################################################

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iJwEAQECAAYFAknsRhwACgkQGXQ4Obkc43neEAP+Ke4fQ5owbDb3gpcNdX7dD3oI
GEzsifB98MJfTjwX23C05i6QFyA7J4bBev6yYPa4f7MoXqyan0MuGWVJLwj/fJb4
X1GcXWn0sFjlNmCaZwmohYIZrt/qYYz0PSxC8m/yFhyzN8okDX8WMjsYXZDXAsp/
8LUbWDh2rycg+/vtQY8=
=Bzt2
-----END PGP SIGNATURE-----