On 05/20/2006 08:04, David MacQuigg wrote:
At 10:14 AM 5/20/2006 +0000, Julian Mehnle wrote:
Scott Kitterman wrote:
IIRC, the DNS library it uses, pyDNS, doesn't support IPv6. That makes
it tough. The SPF specific part wouldn't be so hard. Not sure how I'd
test it though.
Retrofitting IPv6 support to libraries usually isn't all that hard. The
hardest part in retrofitting IPv6 support is address entry and output in
the user interface (command-line or graphical), but libraries usually
don't have that problem. So perhaps it would be an option to extend pyDNS
in cooperation with its authors?
Looks like pyDNS is dead. How hard would it be to switch to
dnspython? http://www.dnspython.org/ is very active, and there is support
for IPv6, DNSSEC, zone transfers, dynamic updates, TSIG authentication,
ENDS0, etc. The package is bigger (1.41MB vs 196KB installed), but maybe
we could strip out what isn't needed.
I would think supporting long records (more than 512 bytes) would be higher
priority than IpV6. IPv6 may never happen.
-- Dave
From the perspective of my validator, supporting long records is not
necessarily a major issue as I understand a reasonable fraction of the
internet still doesn't do EDNS0. If I switched to a DNS library that
supported it, I would also want to raise a warning that the record length was
risky.
From a production perspective, the size of the library is potentially a
concern. Twisted also has a DNS library. If I had infinite free time I'd
make pySPF library independent, but...
Thanks for the pointer. I didn't know about that one.
Scott K
-------
Sender Policy Framework: http://www.openspf.org/
Archives at http://archives.listbox.com/spf-discuss/current/
To unsubscribe, change your address, or temporarily deactivate your
subscription,
please go to
http://v2.listbox.com/member/?listname=spf-discuss(_at_)v2(_dot_)listbox(_dot_)com